1521-0111/88/5/866–879$25.00 http://dx.doi.org/10.1124/mol.115.099580 MOLECULAR PHARMACOLOGY Mol Pharmacol 88:866–879, November 2015 Copyright ª 2015 by The American Society for Pharmacology and Experimental Therapeutics

Mechanism of Modification, by Lidocaine, of Fast and Slow Recovery from Inactivation of Voltage-Gated Na1 Channels

Vaibhavkumar S. Gawali, Peter Lukacs, Rene Cervenka, Xaver Koenig, Lena Rubi, Karlheinz Hilber, Walter Sandtner, and Hannes Todt Center for Physiology and Pharmacology, Department of Neurophysiology and Neuropharmacology (V.S.G., P.L., R.C., X.K., L.R., K.H., H.T.) and Center for Physiology and Pharmacology (W.S.), Medical University of Vienna, Vienna, Austria Received April 17, 2015; accepted September 9, 2015 Downloaded from

ABSTRACT

The clinically important suppression of high-frequency dis- mutationsofdomainIVsegment6inrNav1.4 that resulted in charges of excitable cells by local anesthetics (LA) is largely constructs with varying propensities to enter fast- and slow- determined by drug-induced prolongation of the time course of inactivated states. We tested the effect of the LA lidocaine on repriming (recovery from inactivation) of voltage-gated Na1 the time course of recovery from short and long depolarizing channels. This prolongation may result from periodic drug- prepulses, which, under drug-free conditions, recruited mainly molpharm.aspetjournals.org binding to a high-affinity binding site during the action poten- fast- and slow-inactivated states, respectively. Among the tials and subsequent slow dissociation from the site between tested constructs the mutation-induced changes in native slow action potentials (“dissociation hypothesis”). For many drugs it recovery induced by long depolarizations were not correlated has been suggested that the fast inactivated state represents with the respective lidocaine-induced slow recovery after short the high-affinity binding state. Alternatively, LAs may bind with depolarizations. On the other hand, for long depolarizations the high affinity to a native slow-inactivated state, thereby accel- mutation-induced alterations in native slow recovery were erating the development of this state during action potentials significantly correlated with the kinetics of lidocaine-induced (“stabilization hypothesis”). In this case, slow recovery between slow recovery. These results favor the “dissociation hypothe- action potentials occurs from enhanced native slow inactiva- sis” for short depolarizations but the “stabilization hypothesis” tion. To test these two hypotheses we produced serial cysteine for long depolarizations. at ASPET Journals on September 28, 2021

Introduction which underlies their analgesic, anticonvulsive, and antiar- 1 rhythmic efficacy (Macdonald and Kelly, 1993; Pisani et al., Blockers of voltage-gated Na channels are used as local 1995; Weirich and Antoni, 1998; Dupere et al., 1999). anesthetics, antiarrhythmics, antiepileptics, analgesics, and The use-dependent effect of Na1 channel blockers arises in the treatment of specific skeletal muscle diseases. Potential from high-affinity binding to specific states that are periodi- “noncanonical” roles of Na1 channels in nonexcitable cells 1 cally populated during the firing of action potentials. Thus, may open new drug targets for Na channel blockers, such as local anesthetic drugs bind preferentially to activated (Hille, cancer treatment (Black and Waxman, 2013). Recently, there 1977; Yeh and Tanguy, 1985; Wang et al., 1987; McDonald has been an increasing interest in developing blockers with 1 et al., 1989; Vedantham and Cannon, 1999) and/or inactivated high specificity for certain Na channel isoforms. For exam- states (Hille, 1977; Cahalan, 1978; Yeh, 1978; Bean et al., ple, targeting the neuronal isoforms NaV1.7 and NaV1.8 may 1983; Bennett et al., 1995; Balser et al., 1996a). As a result, the lead to the development of potent analgesic drugs with few restoration of excitability between action potentials is pro- side effects (Leffler et al., 2007; Dib-Hajj et al., 2009; Waxman, longed. This latter effect can result from two processes: Drug- 2013; Lampert et al., 2014). Furthermore, targeting the late associated slowing of recovery of excitability could result from: sodium current in the heart may be used in the treatment 1) a slow rate of dissociation of the drug, which slows recovery of arrhythmias, ischemic heart disease, and heart failure from inactivation, (“dissociation-hypothesis”), or 2) an in- (Coppini et al., 2013; Antzelevitch et al., 2014). Functional 1 creased stability of an intrinsic, slow-inactivated state, on specificity of Na channel blockers is thought to arise from the basis of an increased rate of entry into the slow-inactivated their use-dependent effect, which is responsible for the state in the presence of drug (“stabilization hypothesis”; Fig. suppression of high-frequency discharges in excitable cells, 1). These are two fundamentally different mechanisms that have qualitatively similar effects on excitability. With regard This study was funded by the Austrian Science Fund FWF [Grants P210006- B11 and W1232-B11]. to the dissociation hypothesis, drugs are thought to bind with dx.doi.org/10.1124/mol.115.099580. high affinity to a specific state-dependent conformation of the

1 ABBREVIATIONS: DIV, Domain IV; IF, fast inactivation; IFM, intermediate fast inactivation; INa,Na current; IM, intermediate inactivation; IM-MUT, modified intermediate inactivation; IS, slow inactivation; QX-222, N-(2,6-Dimethylphenylcarbamoylmethyl) trimethylammonium chloride; rNaV1.4, rat muscle NaV1.4; S6, segment 6; VGSC, voltage-gated sodium channels.

866 Lidocaine Modifies Fast and Slow Repriming of Nav Channels 867

dissociation from fast-inactivated states and stabilization of slow-inactivated states need not be mutually exclusive, but both mechanisms may operate in concert, as noted by Nuss et al. (2000). This study was designed to systematically investigate drug-induced modification of the time course of recovery from both fast- and slow-inactivated states. There- Fig. 1. Kinetic scheme of proposed state-dependent binding of lidocaine. A three-affinity modulated receptor model in which NI, IF,IM,andIS fore, we applied the local anesthetic lidocaine to channels represent noninactivated states (i.e., pooled open and closed states), fast, exhibiting varying properties of fast and slow inactivation. a a intermediate, and slow-inactivated states, respectively. F and M are To this end we introduced serial cysteine replacements in the the voltage-dependent forward rates into I and I ; b and b are the F M F M segment 6 (S6) of domain IV (DIV) of Na 1.4. This region of respective voltage-dependent backward rates (a9F, a9 M, b9 F, b9 M refer to v the respective rates among lidocaine-bound states). NI-L, IF-L, and IM-L the channel has previously been shown to modulate both fast are the lidocaine-bound NI, IF,andIM states. kon-NI,kon-F and kon-M are and slow inactivation (Chahine et al., 1994; McPhee et al., the on-rates, koff-NI,koff-F,andkoff-M the off-rates for lidocaine binding to 1995, 1998; Chen et al., 1996; Lerche et al., 1997; Sheets NI, IF,andIM. During short depolarizations channels move from NI to IF, during repolarization the transit is from IF to NI (recovery from IF). et al., 1999; Vedantham and Cannon, 2000; Chanda and During long depolarizations channels move from NI to IF and then Bezanilla, 2002; Bosmans et al., 2008; Goldschen-Ohm et al., a b a b further to IM.Sinceboth M and M are smaller than F and F,both 2013) as well as the binding properties of local anesthetic development of and recovery from I require a longer time course than Downloaded from M drugs (Ragsdale et al., 1994, 1996; Yarov-Yarovoy et al., the respective transitions into and out of IF.The“dissociation hypoth- esis” assumes high affinity binding of lidocaine to IF such that during 2002). In addition, we investigated the effect of lidocaine on short depolarizations channels mainly enter the IF-L state. Upon recovery from inactivation in the mutation W1531G. This repolarization channels recover slowly via the small koff-F rate (drug- dissociation), which determines slow dissociation from the high affinity mutation has been shown to generate a pathway for rapid state (IF-L) and via b9 F.The“stabilization hypothesis” assumes high external access and egress of the permanently charged affinity binding of lidocaine to IM (small value for koff-M). Thus, with derivative of lidocaine QX-222 (Lukacs et al., 2014). This lidocaine, even during short depolarizations, channels move from IF-L mutation can be used to study the effects of untrapping of molpharm.aspetjournals.org (low affinity) to IM-L (high affinity). This occurs because, owing to the high affinity of lidocaine to IM-L, the value for koff-M is small (relative to drugs bound to the internal vestibule. We found that most koff-F and koff-NI). Consequently, from detailed balances, a9M has to be probably slow drug dissociation prolongs recovery from brief large, resulting in an acceleration of development of IM-L.Hence,with depolarizations, whereas drug-induced stabilization of slow- the “stabilization hypothesis” recovery from short depolarizations is inactivated states prolongs recovery from long depolariza- prolonged by lidocaine, because channels are moving slowly from IM-L to NI (Chen et al., 2000). The states IFM and IM-MUT are not considered in tions. Furthermore, we found evidence for lidocaine binding this scheme but would be located between IF and IM. to a state closely associated with the fast-inactivated state. binding site (“modulated receptor theory”) (Hille, 1977; Materials and Methods at ASPET Journals on September 28, 2021 Hondeghem and Katzung, 1977) or drugs may be trapped by the channel gates during specific gating states (“guarded Mutagenesis and Electrophysiology receptor theory”) (Starmer, 1986). The stabilization hypothe- Mutagenesis and electrophysiology were performed as reported sis suggests that drugs bind to native “slow” inactivated states (Zarrabi et al., 2010; Lukacs et al., 2014). that under drug-free condition require a long time to develop Mutagenesis of rat muscle NaV1.4 (rNaV1.4-A) vector consisting of (Khodorov et al., 1976; Zilberter Yu et al., 1991; Balser et al., the rNaV1.4 coding sequence flanked by Xenopus globin 59 and 39 1996b; Kambouris et al., 1998; Chen et al., 2000). However, untranslated regions was provided as a gift by R. Moorman (Univer- drug-binding to a slow-inactivated state accelerates the time sity of Virginia, Charlottesville, VA). This was used as the template for course of development of this state during a depolarization inserting oligonucleotide-directed point mutations by either four- primer polymerase chain reaction and subsequent subcloning into (“stabilization”), thereby increasing the fraction of channels the template using directional ligations or the QuikChange Site- recovering from that state during a subsequent hyperpolar- Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Oligonucleotide ization. The characterization of such multiple binding mech- primers containing a mutation were designed with a change in a silent anisms is of utmost importance for drug development as all restriction site to allow rapid identification of the mutation. In- binding modes of a drug have to be considered in order to corporation of the mutation into the template was then confirmed by predict the overall biologic treatment response. Slow drug DNA sequencing.

Fig. 2. Original traces showing recovery from inactiva- tion in wild-type Nav1.4 channels and in the mutation M1585C. As shown in the inset, cells were first depolar- ized to –20 mV for 10 seconds to induce inactivation (conditioning prepulse). Then cells were repolarized to – 140 mV for variable durations. Thereafter a 5-millisec- ond test pulse to –20 mV was introduced to test for the fraction of available channels. Test-pulse currents are superimposed for the indicated repolarization times after the conditioning prepulse. The time course of recovery is substantially faster in M1585C than in wild-type Nav1.4 channels. 868 Gawali et al.

penicillin and streptomycin (Life Technologies, Carlsbad, CA). Cells

were maintained at 37°C in a humid atmosphere containing 5% CO2.A mixture of plasmids coding 1.5 mg of rNaV1.4 a subunit, 0.2 mgof voltage-gated sodium channel (VGSC) b1 subunit, and 0.02 mgof eGFP were transiently transfected into tsA201 cells in 35-mm dish (Nunc, Roskilde, Denmark) using TurboFect Transfection Reagent (Thermo Fisher Scientific Baltics UAB, Vilnius, Lithuania) according to the manufacturer’s instructions. Sixteen hours later, cells were dissociated from the dish surface by treatment with a 0.25% trypsin solution (Life Technologies) for approximately 2 minutes, pelleted, resuspended in growth medium, and allowed to settle to the bottom of the recording chamber. Prior to recording, the growth medium was removed and changed to bath solution.

Whole-Cell Patch-Clamp Recording Experiments were performed using the whole-cell patch-clamp recording technique. Na1 currents were recorded using an Axo- patch 700B patch-clamp amplifier (Molecular Devices, Sunnyvale, Downloaded from CA). Pipettes were formed from aluminosilicate glass (AF150-100- 10; Science Products, Hofheim, Germany) with a P-97 horizontal puller (Sutter Instruments, Novato, CA), and had resistances between 1.5 and 2.5 MV when filled with the intracellular pipette solution consisting of (mM): 105 CsF, 10 NaCl, 10 EGTA, and 10HEPES.ThepHwassetto7.3withCsOH.Thebathsolution molpharm.aspetjournals.org consisted of (mM): 140 NaCl, 2.5 KCl, 1 CaCl2,1MgCl2,and 10 HEPES. The pH was set to 7.4 with NaOH. Voltage-clamp protocols and data acquisition were performed with pCLAMP 10.2 software (Molecular Devices, Sunnyvale, CA) through a 16-bit A–D/D–A interface (Digidata 1440A; Molecular Devices). Data were low-pass filtered at 10 kHz (–3 dB) and digitized at 100 kHz. Fig. 3. Time course of recovery from inactivation produced by conditioning Lidocaine dissolved in the extracellular solution was applied via an prepulses of varying durations in wild-type Nav1.4. Shown are test pulse OCTAFLOW drug application device (ALA Scientific Instruments currents recorded with electrophysiologic protocols similar to the protocol Inc., Farmingdale, NY). Recording was begun about 5 minutes after shown in Fig. 2. Channels were driven into inactivation by conditioning – whole-cell access was attained to minimize time-dependent shifts in prepulses to 20 mV for the indicated durations. Test-pulse currents were at ASPET Journals on September 28, 2021 normalized to the respective maximum value attained at full recovery and gating. plotted as a function of the recovery interval. The time courses of recovery were assessed for rNav1.4 channels under drug-free conditions (A) and Electrophysiologic Protocols during superfusion with 500 mM lidocaine (B, solid symbols). For analysis the data points were fit with exponential functions (eqs.1–3; connecting The holding potential was –140 mV. The time course of recovery lines). In (B) some fitted lines are reproduced from (A) for comparison. from inactivation was assessed by application of a conditioning pulse See Table 1 for the parameter estimates. Prolonging the duration of the followed by a 5-millisecond test pulse to –20 mV after a varying conditioning prepulse induces a slow phase of recovery. This effect is interval at the holding potential. Unless specified otherwise, the enhanced by lidocaine. duration of the conditioning prepulse was 50 milliseconds and 10 seconds for the assessment of recovery from faster and slower Transfection Procedure inactivated states,. In most experiments the time course of recovery tsA201 cells were grown in Dulbecco’s modified Eagle’s medium from prepulse potentials of both –50 mV and –20 mV was examined as supplemented with 10% fetal bovine serum and 20 IU/ml each of an internal test for the accuracy of the estimation of the respective

TABLE 1 Biophysical parameters for wild type recovery from fast and slow inactivation Results of fitting of eqs. 1, 2, and 3 to the data in Fig. 3. Parameters presented without S.E.M. were fixed during the curve fitting procedure. Statistical comparisons of the control values are versus 10 ms with 50-ms prepulse duration and versus 5 s. with 10-s and 20-s prepulse durations. Prepulse voltage was –20 mV. Values during lidocaine exposure are compared with the same protocol under control conditions. As a result of the negative recovery potential of –140 mV the first phase of recovery was extremely fast. Hence, with short depolarizations, a fraction of channels recovered prior to the first test pulse, resulting in total amplitudes ,1.

Prepulse Duration t1 t2 t3 A1 A2 A3 n

ms ms ms WT (Control) 10 ms 0.57 6 0.03 0.59 6 0.01 4 50 ms 0.71 6 0.03* 100 0.72 6 0.01 0.05 6 0.007 8 5 s 0.71 76.53 6 9.63 1755 0.12 6 0.04 0.75 6 0.03 0.11 6 0.03 4 10 s 0.71 98.39 6 6.45 3044 6 993 0.07 6 0.02 0.79 6 0.02 0.14 6 0.02 5 20 s 0.71 104.6 6 15.79 1755 6 729.7 0.10 6 0.03 0.65 6 0.06 0.24 6 0.05 4 WT (Lidocaine) 10 ms 0.81 6 0.05 ** 185.4 6 10.83 0.38 6 0.01 0.42 6 0.006 4 50 ms 2.34 6 0.36 *** 176.8 6 10.6 0.30 6 0.01*** 0.69 6 0.01 7 10 s 2.34 154.8 6 6.46*** 3044 0.08 6 0.01 0.80 6 0.01** 0.09 6 0.01* 8

*P , 0.05, **P , 0.01, ***P , 0.001. Lidocaine Modifies Fast and Slow Repriming of Nav Channels 869

Fig. 4. Serial cysteine replacements of the DIV-S6 segment modulate the time course of recovery from a short depo- larization and the effect of lidocaine. Normalized time course of recovery from inactivation produced by a short prepulse (50-millisecond) in wild-type Nav1.4 and in mutations carrying cyste- ine replacements of amino acids in DIV-S6 under drug-free conditions (A, B; n =4–7) and during superfusion with 500 mM lidocaine (C, D; n =4–8). The voltage of the conditioning prepulse was either –50 mV (A, C) or –20 mV (B, D). The insets in (A) and (B) reproduce only the late slow phase of recovery. Connect- ing lines are fits of eq. 2 to the data points. Because of the low amplitude of the late phase of recovery, the time

constant was constrained to 100 milli- Downloaded from seconds and only the amplitude was allowed to float during the fitting pro- cedure. The results of the fits are shown in Figs. 5–7. Arrows in (C, D) indicate the transition between the fast and the slow phase of recovery, which is lacking in the mutations I1576C and M1585C. molpharm.aspetjournals.org time constants, which should be independent of the prepulse poten- indicated intervals at –140 mV following a 10-second inacti- tial. Lidocaine hydrochloride was obtained from Sigma-Aldrich. Un- vating prepulse to –20 mV. Shown are currents from wild-type less stated otherwise lidocaine was applied at a concentration of Na 1.4 channels and from channels carrying a mutation in m V 500 M. the DIV-S6 segment at site 1585, which has been shown Data Evaluation. Nonlinear least squares fit of the normalized time course of recovery was performed using following equations: À Á y 5 A1 1- expð-t=t1Þ 1 C (1) at ASPET Journals on September 28, 2021 À Á À Á y 5 A1 1- expð-t=t1Þ 1 A2 1- expð-t=t2Þ 1 C (2)

À Á À À Á y 5 A1 1- expð-t=t1Þ 1 A2 1- expð-t=t2Þ 1 A3 1- expð-t=t3Þ 1 C (3) where t1, t2, and t3 are the time constants of distinct components of recovery; A1, A2, and A3 are the respective amplitudes of these components; and C is the final level of recovery. The order (i.e., first order, eq. 1; second order, eq. 2) of exponential fit was chosen after an increase in the order of the fit produced no significantly better fit as reported by the “extra sum-of-squares F test”. Graphic representation and data analysis was performed using GraphPad Prism, version 5.00 for Windows (GraphPad Software, LaJolla, CA). Statistics-data are expressed as means 6 S.E.M. Statistical comparisons were made using the two-tailed Student’s t tests. P , 0.05 was considered significant.

Results Protocol of Testing Time Course of Recovery from Inactivation. To examine the time course of recovery in wild- type rNaV1.4 channels, a conditioning prepulse to –20 mV was introduced from a holding potential of –140 mV. Thereafter, the potential was returned to –140 mV for varying intervals followed by a 5-millisecond test pulse to –20 mV, which opened Fig. 5. Modulation of the time constant of the first phase of recovery from the fraction of channels that had recovered during the hyper- a short depolarization (50-millisecond) by mutations in DIV-S6 and by polarized interval after the conditioning prepulse. The inward lidocaine. Shown are the time constants of the first rapid phase of recovery currents elicited by the test pulse were normalized to the as derived from fitting of eq. 2 to the data of Fig. 4. *P , 0.05; **P , 0.01; , respective maximum value attained at full recovery and ***P 0.001 compared with the respective value for wild-type channels. P , 0.05; P , 0.01; P , 0.001 for the effect of lidocaine compared plotted as a function of the recovery interval. Figure 2 shows with drug-free condition in a given construct. In most constructs lidocaine the original traces of test pulse currents obtained after the increased the time constants of this phase of recovery. 870 Gawali et al. previously to destabilize slow inactivation (Hayward et al., 5 seconds gave rise to a time course of recovery that could be 1997; Takahashi and Cannon, 2001). Clearly, the time course fitted by three exponentials (eq. 3) (Cummins and Sigworth, of recovery is substantially slower in wild-type Nav1.4 1996; Nuss et al., 1996; Hayward et al., 1997; Kambouris et al., channels than in the mutation M1585C. 1998). The time constants of the three phases were 0.71 Long Prepulse Durations Induce up to Three Inacti- milliseconds, 76.53 6 9.63 milliseconds, and 1755 milliseconds, vated States. The normalized time courses of recovery of the respective amplitudes were 0.12 6 0.04, 0.75 6 0.03, 0.11 6 wild-type NaV1.4 channels following conditioning prepulses of 0.03. Most probably the first phase corresponds to recovery varying durations to –20 mV are shown in Fig. 3; the estimated from IF (Kambouris et al., 1998), although we cannot exclude a fitting parameters are presented in Table 1. The time course of contribution by recovery from IFM. The second phase may recovery following a 10-millisecond prepulse could be fitted by reflect recovery from a state previously termed “intermediate a single exponential equation (eq. 1) with a time constant of inactivation” (IM), characterized by time constants on the order 0.57 6 0.03 milliseconds, indicating that all inactivated channels of several hundred milliseconds (Kambouris et al., 1998). The are recovering from a single inactivated state that we refer to as third phase is characterized by time constants on the order of fast-inactivation (IF; Balser et al., 1996b; Kambouris et al., 1998), seconds and has been referred to as slow inactivation (IS) defined by time constants of recovery on the order of several (Kambouris et al., 1998). Increasing the prepulse duration to milliseconds. If the duration of the conditioning prepulse was 20 seconds increased the fraction of channels recovering from Downloaded from prolonged to 50 milliseconds, the time course of recovery could be IM and IS, as indicated by the effect on the amplitudes A2 and A3. fitted by the double exponential eq. 2, yielding time constants of Effect of Lidocaine on Time Course of Recovery in 0.71 6 0.03 milliseconds and 100 milliseconds and amplitudes Wild-Type Nav1.4 Channels. During superfusion with of 0.72 6 0.01 and 0.05 6 0.007, respectively. This indicates 500 mM lidocaine the time course of recovery from a 50-millisecond that this conditioning prepulse recruited a minimum of two prepulse (further on referred to as “short depolarization”) had inactivated states. The faster phase of recovery most probably two exponential phases with time constants of 2.34 6 0.36 molpharm.aspetjournals.org corresponds to recovery from IF as observed following a milliseconds and 176.8 6 10.6 milliseconds and amplitudes of 10-millisecond prepulse. Because of the low amplitude of the 0.30 6 0.01 and 0.69 6 0.01, respectively (Fig. 3B). Thus, slower phase the time constant was constrained to 100 milliseconds. lidocaine substantially increased the time constant of the first We operationally refer to this phase as recovery from intermediate rapid phase of recovery and the amplitude of the second phase, fast inactivation (IFM). Increasing the prepulse duration to whose time constant is similar to recovery from the native at ASPET Journals on September 28, 2021

Fig. 6. Modulation of amplitudes of two phases of recovery from a short depolarization (50-millisecond) by mutations in DIV-S6 and by lidocaine. Shown are the amplitudes of the two phases of recovery as derived from fitting eq. 2 to the data of Fig. 4. The voltage of the conditioning prepulse was either –50 mV (A, C) or –20 mV (B, D) In most constructs lidocaine increased the amplitude of the second phase of recovery (A2). *P , 0.05; **P , 0.01; ***P , 0.001 compared with the respective value for wild-type channels. P , 0.05; P , 0.01; P , 0.001 for the effect of lidocaine compared with drug-free condition in a given construct. Lidocaine Modifies Fast and Slow Repriming of Nav Channels 871

IM state (Table 1). This could indicate that lidocaine accelerates the entry into the IM state (Kambouris et al., 1998). Alternatively, lidocaine could bind only to the IF state during the conditioning prepulses, but, upon repolarization, dissociate from this state with a time constant similar to the time constant of recovery from the native IM state (Fig. 1). With longer prepulse durations (10 seconds) lidocaine in- creased both the time constant and the amplitude of re- covery from IM. Mutations in DIV-S6 Modulate Fast Inactivation. As reported previously, all tested mutations of the DIV-S6 segment gave rise to robust inward currents with the excep- tion of N1584C (Sunami et al., 2004). The time course of recovery from fast inactivation was examined by a prepulse- test pulse protocol as described before, whereby the duration of the prepulse was set to 50 milliseconds, which has Downloaded from previously been shown to be sufficient to cause NaV1.4 channels to populate mainly fast-inactivated states (Wang et al., 2003). Figure 4, A and B, show the time courses of recovery of the tested constructs for conditioning pulse voltages of –20 mV and –50 mV, respectively. Clearly, mutations in DIV-S6 produced alterations in the time course of recovery. In some mutations the time course of recovery was molpharm.aspetjournals.org monoexponential, whereas in others a second phase of low amplitude was observed (inset of Fig. 4, A and B). For example in the mutation F1579C this second slow phase had amplitude Fig. 7. Time constants of the lidocaine-induced slow phase of recovery of 0.04 6 0.007. In Figs. 5 and 6 the time constants and from a short depolarization (50-millisecond) are modulated by muta- amplitudes of the time course of recovery are presented. The tions in the DIV-S6 segment. Shown are the time constants of the t second phase of recovery as derived from fitting of eq. 2 to the time time constants of recovery from IF ( 1) were substantially courses of recovery of lidocaine-modified channels in Fig. 4 C, D. *P , modulated by the mutations. Thus, in wild-type t1 was 0.71 6 0.05; **P , 0.01; ***P , 0.001 compared with the respective value for 0.03 milliseconds, in F1579C t1 was decreased to 0.38 6 0.02 wild type milliseconds (P , 0.001), and in M1585C t1 was increased to at ASPET Journals on September 28, 2021 1.31 6 0.04 milliseconds (P , 0.001). In general, with conditioning voltages of –20 mV t1 values (further on referred to as “long depolarization”) was substan- were significantly different compared with wild type for all tially altered (Fig. 8). Figure 9 shows the effect of serial S6 constructs, with the exception of S1578C and V1582C (Fig. 5). mutagenesis on the time constants of recovery from IM. Clearly, With conditioning voltages of –50 mV the observed changes most cysteine replacements of DIV-S6 residues modulated the were similar to those observed with –20 mV, with the exception time constant of recovery from IM, with the greatest reduction of higher amplitude of noninactivating currents. As shown in produced by M1585C and the greatest increase generated by Fig. 6 the amplitudes of the first phase of recovery were Y1586C. substantially modified by mutations in DIV-S6. This most According to the classic Hodgkin-Huxley paradigm, chang- probably reflects mutation-induced shifts of the half-point of ing the voltage of the prepulse should alter the time constant steady state inactivation, which is highly modulated by muta- of entry into inactivation and, thus, the distribution of tions in DIV-S6 (Yarov-Yarovoy et al., 2002; Wang et al., 2003). channels among different channel states at the end of the Effect of Lidocaine on Recovery from Short Depola- prepulse, but not the time constants of recovery from these rizations. Superfusion with 500 mM lidocaine resulted in states [e.g., see Takahashi and Cannon (1999)]. If the serial significant increases in the time constant of recovery from IF cysteine replacements in DIV-S6 per se altered the time (t1) in all constructs with the exception of F1579C and Y1586C course of recovery from IM, then we would expect a strong (Fig. 5). The prolongation of t1 was greatest in the mutations correlation between the t2 values measured at inactivating I1576C and M1585C (see below for detailed analysis). voltages of –50 mV and –20 mV. As shown in Fig. 10 there is The most dramatic effect of lidocaine was to increase the indeed a strong positive correlation between the time con- amplitude of the second phase of recovery (A2) in all constructs stants measured at these two voltages. This supports the with the exception of F1579C. This increase in A2 occurred at conclusion that the differences in t2 values are the result of the the expense of decrease in A1 (Fig. 6) or by a decrease in the mutation itself rather than a result of protocol-dependent amplitude of the noninactivating fraction (Fig. 4). The time variabilities. These data also demonstrate that the introduced constant of this late phase (t2, Fig. 7) showed some variation mutations produce variations of time constants of recovery between constructs but, in general, was on the order of 100– from IM that are sufficiently large to give rise to a significant 200 milliseconds, as reported previously (Vedantham and correlation (see Fig. 12) Cannon, 1999; Kondratiev and Tomaselli, 2003). Effect of Lidocaine on Time Constants of Recovery Mutations in DIV-S6 Modulate Slower Forms of from IM. As shown in Fig. 9 lidocaine tended to increase the Inactivation. In some mutations the time course of recovery time constants of recovery from IM in most constructs with the from inactivation induced by 10-second depolarizing prepulses exception of I1575C, F1579C, and Y1586C, which have 872 Gawali et al. Downloaded from molpharm.aspetjournals.org at ASPET Journals on September 28, 2021

Fig. 8. Serial cysteine replacements of the DIV-S6 segment modulate the time course of recovery from a long depolarization (10-second) and the effect of lidocaine. Normalized time course of recovery from inactivation produced by long prepulses (10-second) in wild-type Nav1.4 and in mutations carrying cysteine replacements of amino acids in DIV-S6 under drug-free conditions (A, B; n =4–7) and during superfusion with 500 mM lidocaine (C, D; n =4–7). The voltage of the conditioning prepulse was either –50 mV (A, C) or –20 mV (B, D). Connecting lines are fits of eq. 3 to the data points (for each construct the time constant of the first phase of recovery, most probably representing recovery from residual IF, was constrained to the respective time constant of the fast phase of recovery shown in Fig. 5. The results of the fits are shown in Figs. 9 and 11. Arrows indicate a very slow phase of recovery (IS)thatis substantially reduced by lidocaine.

previously been shown to per se reduce binding of local koff-F; Fig. 1). Upon repolarization lidocaine would dissociate anesthetics (Ragsdale et al., 1994). from IF with a time constant similar to recovery from the Lidocaine Reduces Amplitude of Recovery from IS native IM. during Long Depolarizations. As shown in Fig. 3, recovery Lidocaine-Induced Slow Recovery following Short from long depolarizations occurred mostly from the IM state. Depolarizations Does Not Reflect Recovery from There is also a small component of even slower recovery, with Endogenous IM. As mentioned above, the most dramatic time constants on the order of several seconds, which most effect of lidocaine when tested with short depolarizations was probably results from recovery from a very slow-inactivated to increase the amplitude of a slower phase of recovery (Fig. 7) state, IS (Kambouris et al., 1998). The values of the time with a time constant similar to that of recovery from the native constants of this phase are not reported, because the small IM (Fig. 9). As shown in Fig. 9 the cysteine replacements of amplitude of this phase precluded their precise estimation. As amino acids in DIV-S6 produced alterations in the time course shown in Fig. 11, lidocaine tended to increase the amplitude of of recovery from IM. If the lidocaine-induced slower phase of recovery from IM (A2), mainly at the expense of a reduction in recovery from short depolarizations represented recovery the amplitude of recovery from IS (A3, arrows in Fig. 8) and the from lidocaine-induced IM (i.e., IM-L in Fig. 1) rather than noninactivating component. This could result from the bind- slow dissociation from IF (IF-L in Fig. 1), then we would expect ing of lidocaine to and stabilization of IM, thereby enhancing a strong correlation between the time constant of this the rate of entry into IM (a9 M in Fig. 1) and reducing the rate of lidocaine-induced IM state and the mutation-specific value of entry into IS during the conditioning pulse (via small koff-M, recovery from the native IM. As shown in Fig. 12 the kinetics of Fig. 1). Alternatively, lidocaine could bind to IF, thereby the lidocaine-induced slower phase of recovery from short reducing the time constant of entry into IM and IS (via a small depolarizations did not correlate with the kinetics of either the Lidocaine Modifies Fast and Slow Repriming of Nav Channels 873

Fig. 10. The time constants of recovery from intermediate inactivation produced by inactivating prepulses to –20 mV and to –50 mV are correlated with each other. Data points are t2 values of control measurements replotted from Fig. 9. The line is the best-fit linear regression of the form y = 0.89x + 3.72 (r2 = 0.63, P = 0.006). Hence, the time constants of recovery are independent from prepulse potential as predicted by the classic Downloaded from Hodgkin-Huxley paradigm.

(Hayward et al., 1997). Indeed, the time constant of recovery from IM was substantially shorter in M1585C compared with wild-type channels (32.78 6 4.27 milliseconds versus 98.39 6 6.45 milliseconds, respectively P , 0.0001; Fig. 9, Table 2). molpharm.aspetjournals.org We wondered whether this modified IM state, which we refer to as modified intermediate inactivation (IM-MUT), repre- sented an additional new IM state, or a modification of the native IM. Therefore, we explored whether prolongation of the conditioning prepulse could drive M1585C channels into the native IM. As shown in Fig. 13A prolonging the duration of the conditioning prepulse from 10 to 30 seconds resulted in a substantial additional slowing of the time course of re-

t at ASPET Journals on September 28, 2021 Fig. 9. Lidocaine tends to increase the time constant 2 for the covery of M1585C channels, reflected by an increase of the intermediate phase of recovery, whereas several of the mutations dominant time constant (t2) to a value similar to the time decrease t2 with respect to the wild-type construct under the same conditions. The time constants of recovery from intermediate inactiva- constant of recovery from IM in wild-type channels (73.61 6 tion (IM) were derived by fitting of eq. 3 to the time courses of recovery 14.11 milliseconds; Table 1). This suggests that prolonging , , , shown in Fig. 8. P 0.05; **P 0.01; ***P 0.001 compared with the the duration of the prepulse caused M1585C channels to respective value for wild-type channels (n =5–8). P , 0.05; P , 0.01; P , 0.001 for the effect of lidocaine compared with drug- move from IM-MUT to IM. Exposure of M1585C channels to freeconditioninagivenconstruct.Prepulsevoltagewas220 mV (A) lidocaine had an effect similar to prolongation of the prepulse and 250 mV (B). duration, suggesting that lidocaine accelerated the transi- tion of IM-MUT to IM during the conditioning prepulse (Fig. 13A, Table 2). native IM state (Fig. 12A) or the lidocaine-modified IM state Figure 13B shows the response of lidocaine-modified (Fig. 12B). This suggests that the lidocaine-induced slow M1585C to conditioning prepulses of the following durations: recovery from short depolarizations (Fig. 7) occurs by a mech- 10 milliseconds, 50 milliseconds, and 10 seconds. With anism different from recovery from IM, perhaps by slow lidocaine, recovery from inactivation after a 10-millisecond dissociation of lidocaine from IF (koff-F in Fig. 1). prepulse (–20 mV) occurred by two distinct phases (filled Lidocaine-Induced Slow Recovery following Long triangles in Fig. 13B, Table 1). However, when the prepulse Depolarizations Reflects Recovery from Endogenous duration was extended to 50 milliseconds, the clear separation IM. Conversely, one could ask whether the modulation of of these two phases was lost (filled circles in Fig. 13B). We the slow phase of recovery by lidocaine with long depolari- suspected that this was the result of the emergence of a third zations (Fig. 9) represents slow dissociation from fast phase of recovery that “bridged” the fast and slow phases inactivation rather than recovery from the stabilized native observed with the 10-millisecond prepulse. This third phase IM state.Inthelattercasewewouldexpectacorrelation had a time constant 22.9 6 8.18 milliseconds, i.e., a value within the tested constructs between the time constant of similar to the time constant of recovery from the native IM-MUT recovery from IM under control conditions and the time state (see Table 2, t2 M1585C control, 10-second versus constant of recovery during lidocaine exposure. As shown in M1585C lidocaine 50 milliseconds). A similar trend toward Fig. 12C this is indeed the case. Hence, lidocaine-induced the lidocaine-induced emergence of a third phase of recovery slowing of recovery from fast inactivation appears to be with short depolarizations was observed in the mutation mechanistically distinct from lidocaine modification of slow I1576C (Fig. 13, C and D, arrow). Interestingly, as with recovery (IM). M1585C, this mutation also shortened the time constant of The Mutations M1585C and I1576C Induce a Modified recovery from IM with long depolarizations under drug-free IM State. Mutations at site 1585 destabilize slow inactivation conditions (Fig. 9). These results suggest that with I1576C and 874 Gawali et al. Downloaded from molpharm.aspetjournals.org

Fig. 11. Modulation of amplitudes of two late phases of recovery from a long depolarization (10-second) by mutations in DIV-S6 and by lidocaine. Shown are the amplitudes of the two late phases of recovery, representing recovery from IM and IS. The values were calculated by fitting eq. 3 to each of the time courses of recovery presented in Fig. 8. The voltage of the conditioning prepulse was either –50 mV (A, C) or –20 mV (B, D). *P , 0.05; **P , 0.01; ***P , 0.001 (n =4–10) compared with the respective value for wild-type channels (n =5–8). P , 0.05; P , 0.01; P , 0.001 (n =5–7) for the effect of lidocaine compared with drug-free condition in a given construct. at ASPET Journals on September 28, 2021

M1585C lidocaine appears to stabilize IM-MUT during support the idea that the lidocaine-induced increase in t1 (Fig. 50-millisecond depolarizing pulses. 5) is a result of binding to IFM, i.e., a native state with recovery The assumption of an increased entry into IM-MUT upon kinetics slightly slower than IF.Thus,IFM may be a state lidocaine exposure also explains the substantial drug-induced similar to IM-MUT in M1585C and I1576C. increase in t1 with I1576C and M1585C shown in Fig. 5. Most The Mutation W1531G Accelerates Recovery from probably, with these constructs, t1 reports a mixture of time Lidocaine Block. As mentioned above, the lidocaine- constants of recovery from fast inactivation and from IM-MUT. induced delay in recovery can arise from drug-dissociation or Hence, the enhanced recovery from IM-MUT during lidocaine from drug-induced stabilization of native slow recovery. These exposure may artifactually produce a substantial increase in t1. two processes are difficult to distinguish because, with lido- Can a similar mechanism be responsible for the lidocaine- caine treatment, the drug-induced slow phase of recovery from induced increase in the time constant of the first phase of short depolarizations (Fig. 7) has a time constant similar to that recovery in other constructs (Fig. 5)? Indeed, as shown in the for recovery from unmodified IM (Fig. 9). Such separation would insets to Fig. 3, A and B, in a number of constructs the fast be possible if a mutation would either accelerate recovery from phase of recovery from short depolarizations was followed by lidocaine-modified IF or substantially slow recovery from native a slower phase of low amplitude, which we operationally refer IM. We recently reported that mutation of a tryptophan in the to as recovery from IFM to indicate its kinetic similarity to both outer vestibule of DIV produced a rapid access and egress IF and IM. (Although IFM may be kinetically indistinguishable pathway for external QX-222, a membrane impermeant de- from IM-MUT, only the latter state was observed with long rivative of lidocaine (Lukacs et al., 2014). As shown in Fig. 14 prepulses in mutations I1576C and M1585C.) If the lidocaine- this mutation, W1531G, substantially shortened the time induced increase in the time constant of fast phase was a result constant of recovery from lidocaine-modified IF. This phase of enhanced recovery from IFM, then we would expect a corre- most probably reflects rapid egress of lidocaine from its binding lation between the amplitudes of IFM under drug-free condi- site. However, with W1531G, lidocaine modified the time tions (Fig. 12D, A2-drug-free) and the time constant of the first course of recovery from 10-second prepulses in a way similar phase during lidocaine exposure (Fig. 12D, t1-Lido). Figure 12D to wild-type Nav1.4 channels, i.e., lidocaine substantially re- shows that, in most constructs, there is a significant positive duced the amplitude of recovery from IS (A3 in Fig. 14) correlation between A2-drug-free and t1-lidocaine. Outliers consistent with binding to and stabilizing IM. These data from this trend are I1576C and M1585C, which have strongly support the notion that delayed recovery from inactivation altered kinetics, and F1579C and Y1586C, in which lidocaine produced by short depolarizations represents dissociation binding is defective (Ragsdale et al., 1994). Hence, these data from IF, whereas prolonged recovery from inactivation Lidocaine Modifies Fast and Slow Repriming of Nav Channels 875 Downloaded from molpharm.aspetjournals.org Fig. 12. Lidocaine modifications of recovery from short and from long depolarizations are distinct processes. Correlation analysis is performed to explore the effects of lidocaine on the time constants of recovery. The mutations I1575C, F1579C, and Y1586C are excluded from this analysis because these mutations abolished the effect of lidocaine on the time constants of recovery from long depolarizations at –20 mV, consistent with earlier reports implicating these sites in lidocaine binding (Fig. 9) (Ragsdale et al., 1994). M1585C is also excluded because of mutation-induced complete destabilization of IM.(A)Thetimeconstantofthe second phase of recovery from inactivation produced by 50-millisecond prepulses to –20 mV during lidocaine exposure (Fig. 7A) is plotted as a function of the time constant of the second phase of recovery from 10-second prepulses to –20 mV under drug-free conditions (open bars in Fig. 9A). The line is the best-fit linear regression of the form y = 0.13x + 160.3 (R2 = 0.04, P = 0.63). (B) The time constant of the second phase of recovery from inactivation produced by 50-millisecond prepulses to –20 mV during lidocaine exposure (Fig. 7A) is plotted as a function of the time constant of the second phase of recovery from 10-second prepulses to – 20 mV during lidocaine exposure (solid bars in Fig. 9A). The line is the best-fit linear regression of the form y = 0.27x + 125.9 (R2 =0.07,P =0.48).Inparts(A)and (B), the absence of significant correlations suggests that lidocaine-induced modifications of recovery from short and long repolarizations are mechanistically, as well as kinetically, distinct processes. (C) The time constant of the second phase of recovery from inactivation produced by 10-second prepulses to –20 mV during lidocaine exposure (Fig. 9A, solid bars) is plotted as a function of the time constant of the second phase of recovery from 10-second prepulses to –20 mV during at ASPET Journals on September 28, 2021 drug-free conditions (Fig. 9A, open bars). The line is the best-fit linear regression of the form y = 0.54x + 113.7 (R2 =0.61,P = 0.02). This analysis suggests the presence of a link between the mechanism of generation of IM and the lidocaine-induced modification of the second phase of recovery from long depolarizations. (D) Lidocaine-induced changes in the time constant of the first rapid phase of recovery from short depolarizations (i.e., the difference among the values denoted by the solid and open bars in Fig. 5A) are plotted as a function of the amplitude of the second phase of recovery from short depolarizations (inset of Fig. 4B). The line is the best-fit linear regression of the form y = 29.60x – 0.022 (R2 =0.81,P = 0.0022). Constructs with red labels were excluded from the analysis. This analysis suggests that the lidocaine-induced increase in the time constant of the first phase of recovery from short depolarizations results from a modification of the second slow phase of recovery observed under drug-free conditions (inset of Fig. 4 B, i.e., recovery from IFM). produced by long depolarizations reflects recovery from drug- time course of recovery from inactivation from both short and modified IM. long depolarization. Such protocols allow for judgment of the Discussion limiting number of inactivated states populated during a given conditioning prepulse but provide only limited information on Electrophysiological Assessment of Inactivation. In the voltage dependency of inactivation. We did not apply the present study we examined the effect of lidocaine on the classic availability protocols for different inactivated states

TABLE 2 Biophysical parameters for M1585C and I1576C recovery from fast and slow inactivation Results of fitting of eqs. 1, 2, and 3 to the data presented in Fig. 13. Parameters presented without S.E.M. were fixed during the curve-fitting procedure. Values during lidocaine exposure are compared to the same protocol under control conditions.

Prepulse Duration t1 t2 t3 A1 A2 A3 n

ms ms ms M1585C (Control) 10 ms 1.15 6 0.08 1 6 0.04 5 50 ms 1.13 6 0.04 100 0.84 6 0.01 0.04 6 0.008 7 10 s 1.13 32.78 6 4.27 1035 6 349.3 0.12 6 0.03 0.57 6 0.02 0.15 6 0.02 5 30 s 1.13 73.61 6 14.11 1119 6 299 0.04 6 0.02 0.61 6 0.05 0.33 6 0.05 8 M1585C (Lidocaine) 10 ms 1.9 6 0.2** 290 6 34.3 0.49 6 0.02 0.5 6 0.01 4 50 ms 1.9 22.9 6 8.18 290 0.09 6 0.03 0.22 6 0.03 0.65 6 0.02 7 10 s 1.9 141.8 6 18.81*** 1010 6 353..8 0.04 6 0.02 0.69 6 0.07** 0.24 6 0.07 5 I1576C (Lidocaine) 50 ms 1.9 78.95 6 10.32 290 6 34.3 0.16 6 0.01 0.51 6 0.04 0.31 6 0.04 7

**P , 0.01, ***P , 0.001. 876 Gawali et al. Downloaded from

Fig. 13. The time course of recovery in I1576C and M1585C. (A) Time course of recovery from prepulses of the indicated durations in wild-type and M1585C channels. Lines are the result of fits of eq. 3 to the data points. The fitting parameters are presented in Table 2. The time constant of recovery of the dominating phase (t2) is significantly shorter than in wild-type channels. Increasing the prepulse duration to 30 seconds increases t2 to a value similar molpharm.aspetjournals.org to wild-type channels following 10-second depolarization. A similar effect is produced by exposure to lidocaine (500 mM). Hence the native IM state is not abolished in this construct but rather supplemented by the mutation-induced IM-MUT state. (B) Effect of lidocaine on time course of recovery from conditioning prepulses of the indicated duration in M1585C. The broken and solid lines are fits of eqs. 2 and 3 to the data points, respectively. Recovery from a 10-millisecond depolarization in lidocaine-modified channels occurs by two clearly separated (arrow) phases (solid triangles). However, the separation between these two phases is lost if channels are depolarized for 50 milliseconds (filled circles). These data points were fit with both eqs. 2 and 3. (dotted and solid lines, respectively) and demonstrate a better fit with eq. 3, suggesting the emergence of a third inactivated state with this protocol. This state has a similar time constant of recovery as IM-MUT under drug-free conditions (Table 2; t2 M1585C control, 10 seconds versus M1585C lidocaine 50-milliseconds). Hence, with 50-millisecond prepulse lidocaine may bind to and stabilize IM-MUT, resulting in the loss of separation between the fast and the slower phase of recovery. (C, D) Time course of recovery of the indicated constructs from 50-millisecond depolarizations to the indicated potentials during lidocaine exposure. As in (B) the broken and solid lines are fits of eqs. 2 and 3 to the data points, respectively. The fitted lines for drug-free recoveries are reproduced from Fig. 4, A, B. The arrows indicate the transition between the fast and the slow phases of recovery in wild-type channels, which is not present in M1585C and I1576C. The parameters of the fits are presented in Table 2. The data suggest that lidocaine may stabilize I

M-MUT at ASPET Journals on September 28, 2021 both in M1585C and in I1576C.

because of the difficulty in adjusting these protocols to the 4. An increase in the amplitude of recovery from IM complex mutation-induced alterations in the time course of following long depolarizations (Fig. 11). recovery (Karoly et al., 2010). 5. A decrease in the amplitude of recovery from IS Mutation-Induced Changes in Recovery from following long depolarizations (Fig. 11). Inactivation. To the best of our knowledge this study is Recovery from short depolarizations in lidocaine-exposed the first one to report the effect of serial cysteine replacements channels occurs by two phases, a fast phase commonly in DIV-S6 on the time course of recovery from fast- and slow- interpreted as recovery of drug-unbound channels (Ragsdale inactivated states. For recovery from 50-millisecond pre- et al., 1994) and an additional slow phase representing pulses, both the time constants and the amplitudes of the recovery of drug-bound channels (Vedantham and Cannon, dominating phase varied substantially between constructs 1999). We find that the first rapid phase is also substantially (Figs. 5 and 6). For recovery from long depolarizations the slowed by lidocaine in most tested constructs (Fig. 5), as has greatest changes were generated by M1585C, which sub- been reported previously by numerous studies but not sys- stantially accelerated recovery, and by Y1586C, which caused tematically investigated. Crumb and Clarkson (1990) sug- a delay in recovery (Fig. 8). These changes with mutations at gested that this effect is caused by open-channel unblock sites 1585 and 1586 have been reported previously (Ragsdale during the test pulse. The most dramatic slowing of this first et al., 1994; Hayward et al., 1997; Bai et al., 2003). In gen- phase was observed in I1576C and M1585C. eral these results confirm previous reports implementing Notably, for both mutations, the intermediate time con- the DIV-S6 segment as an important player in the control of stant of recovery was also substantially shortened (Fig. 9). inactivation. This probably indicates destabilization of I by the muta- Effect of Lidocaine on Recovery from Inactivation. M tions [I ; Hayward et al. (1997)]. As shown in Fig. 13B In most constructs we observed five major effects of lidocaine: M-MUT lidocaine-modified M1585C channels recovered from a 1. An increase in the time constant of the first phase of 10-millisecond depolarizing prepulse with two distinct phases, recovery from short depolarizations (Figs. 4, 5, 13). as observed in most other constructs. However, if the prepulse 2. The emergence of a second, slow phase of recovery from duration was increased to 50 milliseconds, a third phase short depolarizations (Figs. 4, 6, 7, 11, 13, 14). of recovery appeared to emerge. This phase had a time con- 3. An increase in the time constant of recovery from IM stant of 22.9 6 8.18 milliseconds and thus “bridged” the two following long depolarizations (Fig. 9). distinct phases of recovery observed with a 10-millisecond Lidocaine Modifies Fast and Slow Repriming of Nav Channels 877

in Fig. 9 the time constants of recovery from IF were sub- stantially modulated by serial cysteine mutagenesis of DIV-S6 residues. Figure 10 demonstrates that the mutation-induced alterations of the time constants of recovery from IF are independent of the prepulse voltage, as would be expected from a Markovian kinetic scheme, and that the changes are suffi- ciently large to be explored by linear correlation analysis. Figure 12, A and B, demonstrates that there is no correlation between the mutation-induced alterations of the time constant of recovery from I (unmodified and drug-modified) and the time Fig. 14. Modification of the time course of recovery in W1531G. The M mutation W1531G creates a pathway for rapid external drug access and constant of the lidocaine-induced slow phase of recovery from egress (Lukacs et al., 2014), thereby allowing for assessment of gating- short depolarization. Hence, this latter phase of recovery independent drug dissociation. Here, we make use of this mutation to appears to be mechanistically different from recovery from I examine the contribution of drug-dissociation to the drug-induced pro- M longation of channel repriming. Shown here are the time courses of and probably does not result from drug-induced acceleration of recovery from prepulse durations of the indicated durations (prepulse entry into IM (a9M in Fig. 1). Most probably this phase represents voltage = –20 mV). Open and filled symbols denote data recorded under drug dissociation from the IF state (koff-F in Fig. 1), as proposed drug-free conditions and during application of 1 mM lidocaine, respec- previously (Hille, 1977; Ragsdale et al., 1994). This notion is Downloaded from tively. Connecting broken lines are fits of eqs. 2 and 3 to the data points of recovery from 50-millisecond and 10-second prepulses to –20 mV, re- supported by the effects of lidocaine on the mutation W1531G spectively. Compared with wild-type, recovery from 50-millisecond pulses (Fig. 14). W1531G is located in the P-loop of DIV and is part of in lidocaine-exposed channels is substantially accelerated in W1531G. a highly conserved “ring of tryptophanes” forming a portion of However, the time courses for lidocaine-modified recovery from 10-second depolarizations are similar for the two different constructs. For compar- the outer vestibule (Lipkind and Fozzard, 2000; Durell and Guy, ison, fitted lines to the indicated wild-type data are reproduced from Fig. 3. 2001; Tikhonov and Zhorov, 2011). W1531G opens an external The parameters of the fits are: access/egress pathway for local anesthetics (Lukacs et al., 2014). molpharm.aspetjournals.org t 6 t 6 50 milliseconds (control, n = 9), 1 = 1.42 0.12 milliseconds, 2 = 40.42 The lidocaine-induced slow phase of recovery from short depola- 22.91 milliseconds, A1 = 0.99 6 0.03, A2 = 0.10 6 0.002; 50 milliseconds (lidocaine, 500 mM, n =5)t1 = 2.13 6 0.55 milliseconds, t2 = rizations was substantially accelerated in W1531G compared 6 6 6 31.87 5.91 milliseconds, A1 = 0.49 0.05, A2 = 0.52 0.06; with wild-type channels, whereas lidocaine-modified recovery t 6 t 6 10 seconds (control, n =4), 1 1 = 1.42 0.12 milliseconds, 2 = 166.5 from inactivation produced by long depolarizations occurred at 14.59 milliseconds, t = 1801 6 255.3 milliseconds, A = 0.01 6 0.01, 3 1 a much slower rate. Thus, upon recovery from short depolari- A2 = 0.61 6 0.03, A3 = 0.37 6 0.03; 10 seconds (lidocaine, 1 mM, n =3),t1 1 = 2.13 6 0.55 milliseconds, t2 2= zations lidocaine most probably dissociates from fast- 6 t 6 6 140.2 11.02 milliseconds, 3 3 = 1801 255.3 milliseconds, A1 = 0.01 inactivated channels. This dissociation is accelerated by the 0.01, A = 0.83 6 0.02, A = 0.15 6 0.02. 2 3 external egress pathway created by W1531G. Although the

high-affinity state for lidocaine with short depolarizations at ASPET Journals on September 28, 2021 depolarizing prepulse (Fig. 13, B–D). This effect may have appears to be the fast-inactivated state, it has to be noted that generated artifactually a dramatic increase in the time there is increasing experimental evidence for local anesthetic constant of the first phase if this phase was fitted with a single drug-binding along the activation pathway that precedes entry exponential equation, as performed in Fig. 5. A similar into inactivation (Vedantham and Cannon, 1999; Muroi and phenomenon occurred with I1576C (Fig. 13, C and D). Thus, Chanda, 2009). it appears that in these constructs, with short depolarizations, On the other hand, after long depolarizations recovery lidocaine stabilizes IM-MUT (Fig. 9). A similar mechanism may appears to occur mainly from the lidocaine-modified IM state. account for the lidocaine-induced increase in the time constant Lidocaine significantly increased both the time constant and of the fast phase of recovery from short depolarizations, as the amplitude of recovery from IM in most examined muta- observed in most constructs (Fig. 5). As shown in Fig. 12D this tions (Fig. 9 and Fig. 11, A and B). The modification of recovery increase significantly correlates with the mutation-induced from IM by lidocaine has been reported previously (Kambouris changes in the amplitude of IFM, i.e., the slow phase of et al., 1998). The authors noted that “lidocaine enhanced recovery from short depolarizations under drug-free condi- occupancy of a nonconducting state with intermediate re- tions (insets of Fig. 4, A and B). Hence, lidocaine may bind to covery kinetics similar (but not identical) to IM.” In this study IFM, which is populated during short depolarizations and lidocaine consistently increased the time constant of IM with recovers with a time course slightly slower than that for IF. prepulses to –20 mV (Fig. 9). With prepulses to –50 mV the Further studies will be needed to elucidate whether IFM is effect was less dramatic, mainly because of the lower ampli- identical to IM-MUT, which was observed in the mutations tude of recovery at this voltage, which reduces the accuracy of I1576C and M1585C. estimation of the time constants by nonlinear curve fitting. The most dramatic effect of lidocaine on recovery from short The lidocaine-induced increase in the time constant of re- depolarizations was the emergence of a slower phase of recovery covery from IM may reflect slow drug dissociation from that with a time constant of ∼170 milliseconds. This time constant state at hyperpolarized voltages. was similar to the time constant of recovery from IM observed Figure 12C shows that there is a significant correlation with long depolarizations (Fig. 9). Hence, according to the between the mutation-induced changes in the time constant of “stabilization hypothesis,” during short depolarizations lidocaine recovery from IM and the lidocaine-modified time constant of may accelerate entry into the native IM state. Alternatively, IM, suggesting a link between recovery from the native IM according to the “dissociation hypothesis,” lidocaine may bind to state and lidocaine-induced slowing of recovery. IF during the short depolarization and, upon repolarization, The significant reduction of the amplitude of recovery from dissociate with a time constant that, by coincidence, has a value IS by lidocaine (Fig. 11, C and D) most probably results from similar to the time constant of recovery from native IM.Asshown the drug-induced stabilization of IM, which prevents channels 878 Gawali et al. from entering into other connected states (Kambouris et al., Hondeghem LM and Katzung BG (1977) Time- and voltage-dependent interactions of antiarrhythmic drugs with cardiac sodium channels. Biochim Biophys Acta 472: 1998). This effect also explains the substantial acceleration of 373–398. recovery from long depolarizations in W1531G (Fig. 14). With Kambouris NG, Hastings LA, Stepanovic S, Marban E, Tomaselli GF, and Balser JR (1998) Mechanistic link between lidocaine block and inactivation probed by outer this mutation lidocaine reduced the fraction of channels that pore mutations in the rat micro1 skeletal muscle sodium channel. J Physiol 512: – recovered from IS with long depolarizations from 37% to 15% 693 705. Karoly R, Lenkey N, Juhasz AO, Vizi ES, and Mike A (2010) Fast- or slow-inactivated (Fig. 14). A similar phenomenon has been observed in NaV1.7 state preference of Na1 channel inhibitors: a simulation and experimental study. channels (Sheets et al., 2011) and may allow for development PLOS Comput Biol 6:e1000818. Khodorov B, Shishkova L, Peganov E, and Revenko S (1976) Inhibition of sodium of a new generation of analgesic drugs. currents in frog Ranvier node treated with local anesthetics. Role of slow sodium inactivation. Biochim Biophys Acta 433:409–435. Acknowledgments 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- The authors thank Jarmila Uhrinova for her excellent technical dependent Na1 channels. Mol Pharmacol 64:741–752. assistance. Lampert A, Eberhardt M, and Waxman SG (2014) Altered sodium channel gating as molecular basis for pain: contribution of activation, inactivation, and resurgent – Authorship Contributions currents. Handbook Exp Pharmacol 221:91 110. Leffler A, Reiprich A, Mohapatra DP, and Nau C (2007) Use-dependent block by Participated in research design: Gawali, Lukacs, Todt. lidocaine but not amitriptyline is more pronounced in tetrodotoxin (TTX)- 1 Conducted experiments: Gawali, Lukacs. Resistant Nav1.8 than in TTX-sensitive Na channels. J Pharmacol Exp Ther 320:354–364.

Performed data analysis: Gawali, Lukacs, Cervenka, Koenig, Rubi, Lerche H, Peter W, Fleischhauer R, Pika-Hartlaub U, Malina T, Mitrovic N, Downloaded from Sandtner, Todt., and Lehmann-Horn F (1997) Role in fast inactivation of the IV/S4-S5 loop of the 1 Wrote or contributed to the writing of the manuscript: Gawali, human muscle Na channel probed by cysteine mutagenesis. J Physiol 505: 345–352. Cervenka, Koenig, Hilber, Sandtner, Todt. Lipkind GM and Fozzard HA (2000) KcsA crystal structure as framework for a mo- lecular model of the Na(1) channel pore. Biochemistry 39:8161–8170. References Lukacs P, Gawali VS, Cervenka R, Ke S, Koenig X, Rubi L, Zarrabi T, Hilber K, Antzelevitch C, Nesterenko V, Shryock JC, Rajamani S, Song Y, and Belardinelli L Stary-Weinzinger A, and Todt H (2014) Exploring the structure of the voltage- (2014) The role of late I Na in development of cardiac arrhythmias. Handbook Exp gated Na1 channel by an engineered drug access pathway to the receptor site for

Pharmacol 221:137–168. local anesthetics. J Biol Chem 289:21770–21781. molpharm.aspetjournals.org Bai CX, Glaaser IW, Sawanobori T, and Sunami A (2003) Involvement of local an- Macdonald RL and Kelly KM (1993) Antiepileptic drug mechanisms of action. esthetic binding sites on IVS6 of sodium channels in fast and slow inactivation. Epilepsia 34 (Suppl 5):S1–S8. Neurosci Lett 337:41–45. McDonald TV, Courtney KR, and Clusin WT (1989) Use-dependent block of single Balser JR, Nuss HB, Orias DW, Johns DC, Marban E, Tomaselli GF, and Lawrence sodium channels by lidocaine in guinea pig ventricular myocytes. Biophys J 55: JH (1996a) Local anesthetics as effectors of allosteric gating. Lidocaine effects on 1261–1266. inactivation-deficient rat skeletal muscle Na channels. J Clin Invest 98:2874–2886. McPhee JC, Ragsdale DS, Scheuer T, and Catterall WA (1995) A critical role for Balser JR, Nuss HB, Romashko DN, Marban E, and Tomaselli GF (1996b) Functional transmembrane segment IVS6 of the sodium channel alpha subunit in fast in- consequences of lidocaine binding to slow-inactivated sodium channels. JGen activation. J Biol Chem 270:12025–12034. Physiol 107:643–658. McPhee JC, Ragsdale DS, Scheuer T, and Catterall WA (1998) A critical role for the Bean BP, Cohen CJ, and Tsien RW (1983) Lidocaine block of cardiac sodium chan- S4-S5 intracellular loop in domain IV of the sodium channel alpha-subunit in fast nels. J Gen Physiol 81:613–642. inactivation. J Biol Chem 273:1121–1129. Bennett PB, Valenzuela C, Chen LQ, and Kallen RG (1995) On the molecular nature Muroi Y and Chanda B (2009) Local anesthetics disrupt energetic coupling between of the lidocaine receptor of cardiac Na1 channels. Modification of block by alter- the voltage-sensing segments of a sodium channel. J Gen Physiol 133:1–15. ations in the alpha-subunit III-IV interdomain. Circ Res 77:584–592. Nuss HB, Balser JR, Orias DW, Lawrence JH, Tomaselli GF, and Marban E (1996) at ASPET Journals on September 28, 2021 Black JA and Waxman SG (2013) Noncanonical roles of voltage-gated sodium Coupling between fast and slow inactivation revealed by analysis of a point mu- channels. Neuron 80:280–291. tation (F1304Q) in mu 1 rat skeletal muscle sodium channels. J Physiol 494: Bosmans F, Martin-Eauclaire MF, and Swartz KJ (2008) Deconstructing voltage 411–429. sensor function and pharmacology in sodium channels. Nature 456:202–208. Nuss HB, Kambouris NG, Marbán E, Tomaselli GF, and Balser JR (2000) Isoform- Cahalan MD (1978) Local anesthetic block of sodium channels in normal and specific lidocaine block of sodium channels explained by differences in gating. pronase-treated squid giant axons. Biophys J 23:285–311. Biophys J 78:200–210. Chahine M, George AL, Jr, Zhou M, Ji S, Sun W, Barchi RL, and Horn R (1994) Pisani A, Stefani A, Siniscalchi A, Mercuri NB, Bernardi G, and Calabresi P (1995) Sodium channel mutations in paramyotonia congenita uncouple inactivation from Electrophysiological actions of felbamate on rat striatal neurones. Br J Pharmacol activation. Neuron 12:281–294. 116:2053–2061. Chanda B and Bezanilla F (2002) Tracking voltage-dependent conformational Ragsdale DS, McPhee JC, Scheuer T, and Catterall WA (1994) Molecular determi- changes in skeletal muscle sodium channel during activation. J Gen Physiol 120: nants of state-dependent block of Na1 channels by local anesthetics. Science 265: 629–645. 1724–1728. Chen LQ, Santarelli V, Horn R, and Kallen RG (1996) A unique role for the S4 Ragsdale DS, McPhee JC, Scheuer T, and Catterall WA (1996) Common molecular segment of domain 4 in the inactivation of sodium channels. J Gen Physiol 108: determinants of local anesthetic, antiarrhythmic, and anticonvulsant block of 549–556. voltage-gated Na1 channels. Proc Natl Acad Sci USA 93:9270–9275. Chen Z, Ong BH, Kambouris NG, Marban E, Tomaselli GF, and Balser JR (2000) Sheets MF, Kyle JW, Kallen RG, and Hanck DA (1999) The Na channel voltage Lidocaine induces a slow inactivated state in rat skeletal muscle sodium channels. sensor associated with inactivation is localized to the external charged residues of J Physiol 524 (Pt 1):37–49. domain IV, S4. Biophys J 77:747–757. Coppini R, Ferrantini C, Yao L, Fan P, Del Lungo M, Stillitano F, Sartiani L, Tosi B, Sheets PL, Jarecki BW, and Cummins TR (2011) Lidocaine reduces the transition to Suffredini S, and Tesi C et al. (2013) Late sodium current inhibition reverses slow inactivation in Na(v)1.7 voltage-gated sodium channels. Br J Pharmacol 164 electromechanical dysfunction in human hypertrophic cardiomyopathy. Circula- (2b):719–730. tion 127:575–584. Starmer CF (1986) Theoretical characterization of ion channel blockade: ligand Crumb WJ, Jr and Clarkson CW (1990) Characterization of cocaine-induced block of binding to periodically accessible receptors. J Theor Biol 119:235–249. cardiac sodium channels. Biophys J 57:589–599. Sunami A, Tracey A, Glaaser IW, Lipkind GM, Hanck DA, and Fozzard HA (2004) Cummins TR and Sigworth FJ (1996) Impaired slow inactivation in mutant sodium Accessibility of mid-segment domain IV S6 residues of the voltage-gated Na1 channels. Biophys J 71:227–236. channel to methanethiosulfonate reagents. J Physiol 561:403–413. Dib-Hajj SD, Binshtok AM, Cummins TR, Jarvis MF, Samad T, and Zimmermann K Takahashi MP and Cannon SC (1999) Enhanced slow inactivation by V445M: a so- (2009) Voltage-gated sodium channels in pain states: role in pathophysiology and dium channel mutation associated with myotonia. Biophys J 76:861–868. targets for treatment. Brain Res Brain Res Rev 60:65–83. Takahashi MP and Cannon SC (2001) Mexiletine block of disease-associated mutations Dupere JR, Dale TJ, Starkey SJ, and Xie X (1999) The anticonvulsant BW534U87 in S6 segments of the human skeletal muscle Na(1)channel.JPhysiol537:701–714. depresses epileptiform activity in rat hippocampal slices by an adenosine- Tikhonov DB and Zhorov BS (2011) Possible roles of exceptionally conserved residues dependent mechanism and through inhibition of voltage-gated Na1 channels. Br around the selectivity filters of sodium and calcium channels. J Biol Chem 286: J Pharmacol 128:1011–1020. 2998–3006. Durell SR and Guy HR (2001) A putative prokaryote voltage-gated Ca(21) channel Vedantham V and Cannon SC (1999) The position of the fast-inactivation gate during with only one 6TM motif per subunit. Biochem Biophys Res Commun 281:741–746. lidocaine block of voltage-gated Na1 channels. J Gen Physiol 113:7–16. Goldschen-Ohm MP, Capes DL, Oelstrom KM, and Chanda B (2013) Multiple pore Vedantham V and Cannon SC (2000) Rapid and slow voltage-dependent conforma- conformations driven by asynchronous movements of voltage sensors in a eukary- tional changes in segment IVS6 of voltage-gated Na(1) channels. Biophys J 78: otic sodium channel. Nat Commun 4:1350. 2943–2958. Hayward LJ, Brown RH, Jr, and Cannon SC (1997) Slow inactivation differs among Wang GK, Brodwick MS, Eaton DC, and Strichartz GR (1987) Inhibition of sodium mutant Na channels associated with myotonia and periodic paralysis. Biophys J currents by local anesthetics in chloramine-T-treated squid axons. The role of 72:1204–1219. channel activation. J Gen Physiol 89:645–667. Hille B (1977) Local anesthetics: hydrophilic and hydrophobic pathways for the drug- Wang SY, Bonner K, Russell C, and Wang GK (2003) Tryptophan scanning of D1S6 receptor reaction. J Gen Physiol 69:497–515. and D4S6 C-termini in voltage-gated sodium channels. Biophys J 85:911–920. Lidocaine Modifies Fast and Slow Repriming of Nav Channels 879

Waxman SG (2013) Painful Na-channelopathies: an expanding universe. Trends Mol Zarrabi T, Cervenka R, Sandtner W, Lukacs P, Koenig X, Hilber K, Mille M, Lipkind Med 19:406–409. GM, Fozzard HA, and Todt H (2010) A molecular switch between the outer Weirich J and Antoni H (1998) Rate-dependence of antiarrhythmic and proar- and the inner vestibules of the voltage-gated Na1 channel. J Biol Chem 285: rhythmic properties of class I and class III antiarrhythmic drugs. Basic Res Cardiol 39458–39470. 93 (Suppl 1):125–132. Zilberter Yu, Motin L, Sokolova S, Papin A, and Khodorov B (1991) Ca-sensitive slow Yarov-Yarovoy V, McPhee JC, Idsvoog D, Pate C, Scheuer T, and Catterall WA (2002) inactivation and lidocaine-induced block of sodium channels in rat cardiac cells. J Role of amino acid residues in transmembrane segments IS6 and IIS6 of the Na1 Mol Cell Cardiol 23 (Suppl 1):61–72. channel alpha subunit in voltage-dependent gating and drug block. J Biol Chem 277:35393–35401. Yeh JZ (1978) Sodium inactivation mechanism modulates QX-314 block of sodium Address correspondence to: Dr. Hannes Todt, Center for Physiology and channels in squid axons. Biophys J 24:569–574. Pharmacology, Department of Neurophysiology and Neuropharmacology, Yeh JZ and Tanguy J (1985) Na channel activation gate modulates slow recovery Medical University of Vienna, 1090 Vienna, Austria. E-mail: hannes.todt@ from use-dependent block by local anesthetics in squid giant axons. Biophys J 47: meduniwien.ac.at 685–694. Downloaded from molpharm.aspetjournals.org at ASPET Journals on September 28, 2021