Sodium channel slow inactivation interferes with open channel block

Mutations in the voltage-gated sodium channel Nav1.7 are linked to inherited pain syndromes such as erythromelalgia (IEM) and paroxysmal extreme pain disorder (PEPD). PEPD mutations impair Nav1.7 fast inactivation and increase persistent currents. PEPD mutations also increase resurgent currents, which involve the voltage-dependent release of an open channel blocker. In contrast, IEM mutations, whenever tested, leave resurgent currents unchanged. Accordingly, the IEM deletion mutation L955 (ΔL955) fails to produce resurgent currents despite enhanced persistent currents, which have hitherto been considered a prerequisite for resurgent currents. Additionally, ΔL955 exhibits a prominent enhancement of slow inactivation (SI). We introduced mutations into Nav1.7 and Nav1.6 that either enhance or impair SI in order to investigate their effects on resurgent currents. Our results show that enhanced SI is accompanied by impaired resurgent currents, which suggests that SI may interfere with open-channel block.

have been assumed to support resurgent currents 9 . A characteristic of many IEM mutations is an enhanced slow inactivation (SI).
Here we set out to investigate the effect of enhanced SI on the formation or enhancement of resurgent currents in Navs, thereby furthering our understanding of mutation-induced clinical symptoms (e.g. IEM or PEPD 1 ). We selectively enhanced or inhibited SI by introducing residue substitutions into Nav1.7 and also the Nav1.6 channel, which has a distinctly different biophysical profile. We demonstrate for both subtypes that there is a correlation between enhanced SI and reduced resurgent currents, independent of the amount of persistent currents. Our results suggest a new regulatory mechanism for resurgent currents, namely that SI interferes with open channel block.

Results
The erythromelalgia mutation Nav1.7/ΔL955 fails to produce resurgent currents despite enhanced persistent currents. Many IEM mutations display enhanced SI but none of these show resurgent currents 1,8 . SI of the IEM mutation Nav1.7/Δ L955 is strongly enhanced (Fig. 1A,B). In addition to the pronounced hyperpolarizing shift of the midpoint (V 1/2 ) of activation (Fig. 1C, − 26.8 ± 1.6 mV for wildtype (WT) and − 44.5 ± 1.3 mV for Nav1.7/Δ L955, Δ V 1/2 = − 17.7 mV), Nav1.7/Δ L955 produces enhanced persistent currents (Fig. 1D). In addition to persistent currents, Nav1.7/Δ L955 was reported to have a slower current decay 10 , a combination that would predispose it to generate resurgent currents. We therefore tested this mutation under conditions that would enhance any occurring open channel block. HEK cells lack the endogenous blocker and therefore resurgent currents are normally absent (Fig. S1). In an experimental setting with 100 μ M β 4-peptide in the pipette and 50 nM ATX-II in the bath solution as described previously 11 (see Methods section) we observed resurgent currents in Nav1.7 WT ( Fig. 2A,B) in about 50% of the tested cells (Fig. 2C). Despite optimizing all conditions to generate resurgent currents only one out of 15 cells expressing the Nav1.7/Δ L955 mutation displayed resurgent currents, which were very small (Fig. 2). This clearly contrasts with the enhanced persistent currents that we (Fig. 1D) and others 10 described for the Nav1.7/Δ L955 mutation. Nav1.7/V948C impairs SI but enhances resurgent currents. As SI is negatively shifted in the Nav1.7/ Δ L955 mutation and resulted in diminished resurgent currents, we questioned whether a depolarizing shift in SI would have the opposite effect on the generation of resurgent currents. For this reason we mutated residues in the pore region of Nav1.7 that were shown to affect SI of Nav1.4. In the Nav1.7/V948C mutation (corresponding to Nav1.4/V787C 12 ) SI is strongly reduced with a depolarizing shift of the midpoint (V 1/2 ) of SI of + 16.2 mV (Fig. 1B, − 63.5 mV ± 2.1 mV for WT and − 47.3 mV ± 2.8 mV for V948C) whereas activation and steady-state fast inactivation were affected to a much lesser extent (activation Δ V 1/2 = + 8.1 mV, steady-state fast inactivation Δ V 1/2 = − 7.8 mV, Fig. 1C,E, supplementary Table S1). Surprisingly Nav1.7/V948C displays no persistent currents, which would typically be considered to render the channel unlikely to produce resurgent currents (Fig. 1D).
Under conditions that would support open channel block, the SI deficient Nav1.7/V948C mutation still showed no persistent currents, whereas its resurgent currents were clearly detectable in a substantial subset of patched cells, especially when compared to Nav1.7 WT (Fig. 2). Six out of eleven cells (55%) showed resurgent currents for the Nav1.7/V948C mutation, whereas in Nav1.7 WT transfected cells only eight out of 19 cells (42%) displayed this current.
Subtype Specificity: SI of Nav1.7 WT is more pronounced than that of Nav1.6 WT. The pain-linked Nav1.7 WT channel produces little-to-no resurgent currents 13 . In contrast, Nav1.6 is believed to be the predominant carrier of resurgent currents in Purkinje cells 14 and also DRGs 15 . Interestingly Nav1.6 is described as relatively resistant to SI 16 and both features promote high-frequency firing of action potentials.
The voltage-dependence of steady-state SI is more than 20 mV more hyperpolarized in Nav1.7 (V 1/2 = −63.5 ± 2.1 mV, n = 10) compared to Nav1.6 (V 1/2 = − 43.4 ± 1.2 mV, n = 7) (Fig. 3). The maximum amount of channels undergoing SI revealed by Boltzmann fits was significantly different for the two channels (90.0% for Nav1.7 and 66.9% for Nav1.6, p < 0.001). Even an increased prepulse length of 60 s did not enhance SI of Nav1.6 to the extent observed in Nav1.7 (supplemental Table S1). As our data on Nav1.7 suggest that strong SI may impair the formation of resurgent currents, we hypothesized that a mutation that enhances SI in Nav1.6 may decrease or abolish resurgent currents.
Thus, mutations in the pore of Nav1.6 are capable of strongly enhancing (Nav1.6/N1455A) or reducing (Nav1.6/V966C) SI with only mild effects on activation or steady-state fast inactivation. Nav1.6 WT and Nav1.6/V966C display persistent currents and large resurgent currents. Under conditions that favor open channel block, Nav1.6 WT and the V966C mutation displayed small persistent currents. This contrasts with the corresponding V948C substitution in Nav1.7 that did not show any persistent currents (see Fig. 1D). Persistent currents were completely abolished in the Nav1.6/N1455A mutation (Fig. 4E).
Scientific RepoRts | 6:25974 | DOI: 10.1038/srep25974 well as in Nav1.7 reduced SI may lead to enhanced resurgent currents, whereas persistent currents do not seem to be necessarily required (summarized in Table 1).
Our results suggest that SI in Nav1.7 as well as in Nav1. 6 interferes with open channel block and thus may present a new mechanism to suppress the generation of resurgent currents.

Discussion
Our studies disclose a relationship between SI and resurgent currents. Nav1.7 and Nav1.6 differ considerably in their propensity for SI and resurgent currents and we mutated both channels to further study these and other biophysical properties (Table 1). Nav1.7/Δ L955 and Nav1.6/N1455A enhanced SI but lead to an attenuated capability to produce resurgent currents. In contrast Nav1.7/V948C and Nav1.6/V966C mutations impaired SI but generated larger resurgent currents. Thus, there appears to be a negative correlation between the extent of SI and the capability to generate resurgent currents.
The current mechanism proposed for resurgent currents is that an endogenous blocker -most likely the positively-charged C-terminus of the Nav β 4 subunit 6,20,21 -binds to and blocks the open pore of an activated Nav channel. Upon repolarization the blocker unbinds to allow current flow, thus generating resurgent currents. Fast inactivation and resurgent currents have been described as mutually exclusive as binding of the fast inactivation particle on the DIII-DIV linker precludes binding of the resurgent-current blocker 22 . Instead, resurgent currents have to date been considered strongly associated with persistent currents.
Besides the so-called "window" current observed at voltages where channels activate but inactivation is incomplete, persistent current is mainly explained by two hypotheses: (1) channels may have an incomplete equilibrium occupancy of inactivated states 23 , and (2) persistent current is a short lived closed-state distinct from the inactivated state, which occurs when inactivation is slow or does not take place while channels are able to reopen by bursting channel behavior. Probably even more than one "persistent current state" exists 24,25 . Bant and Raman 20 also hypothesize that the blocking particle could be involved in both currents: Persistent currents could occur by rapid binding and unbinding with low affinity at negative voltages and resurgent currents require strong de-and repolarizations to expel the blocking particle bound with high affinity that inhibits persistent currents.
It seems intuitive that an open pore that conducts persistent currents would also increase the probability for the blocking particle to bind, thus producing more pronounced resurgent currents 6,8,26,27 and indeed our data for the Nav1.6/N1455A mutation fit with this model. However, this does not hold true for the other mutations we investigated. Despite pronounced persistent currents in the IEM mutation Nav1.7/Δ L955 10 , no resurgent currents were observed. Surprisingly, the Nav1.7/V948C mutation with no persistent currents exhibited increased resurgent currents. Furthermore, all cells expressing Nav1.6/V966C displayed more resurgent currents compared to WT, even though persistent currents were twice the size in Nav1.6/WT compared with Nav1.6/V966C.
It is conspicuous that the voltage range in which persistent and resurgent currents arise is largely the same for both channel isoforms and their mutants tested here. Therefore one can assume that both current states can arise at the same voltage and thus might be alternative gating modes that both become more likely if for example fast inactivation is slow. Interestingly, there was a strong negative correlation between SI and resurgent currents that has also been described for persistent currents 24 . SI could therefore represent another gating mode competing with short lived closed states. Based on the knowledge we have up to now on persistent and resurgent currents our findings suggest that persistent currents and resurgent currents could both resemble two competing short-lived closed states that might be two alternatives that can be favored when inactivation is slowed and SI did not yet occur. Figure 3. SI is more pronounced in Nav1.7 WT than in Nav1.6 WT. SI of Nav1.7 WT develops at more than 20 mV more hyperpolarized potentials than SI of Nav1.6 WT (V 1/2 = − 63.5 ± 2.1 mV for Nav1.7 (filled black squares, n = 8) and − 43.4 ± 1.2 mV for Nav1.6 WT (filled black circles, n = 7). Note that maximum amount of SI at depolarized potentials is also different (90.0% for Nav1.7 WT and 66.9% for Nav1.6 WT). For the data shown in this figure the identical protocol was used for Nav1.7 and for Nav1.6 (see inset).
The poor correlation we observed between persistent currents and resurgent currents contrasts with the strong relationship we discovered between SI and resurgent currents. The structural basis underlying SI remains poorly understood but is believed to involve structural rearrangements of the pore, which is mechanistically analogous to C-type inactivation in potassium channels 28 . It is a separate process that does not involve the inactivation particle on the DIII-DIV linker 29 (Fig. 7). Numerous site-directed mutagenesis studies have identified residues in the S6 segments that affect SI 18,30-33 , supporting the hypothesis that SI involves structural changes and possible collapse of the pore region. The link between SI and the pore has also been demonstrated pharmacologically. Local anesthetics (LA) that bind to their receptor site in the pore can affect SI e.g. lidocaine reduces transition to SI 34 and stabilizes the SI state 35 .
The binding site for the resurgent-current blocking particle overlaps with the local anesthetic (LA) receptor 20 . We therefore propose that SI-associated structural rearrangements of the pore may disrupt the receptor for the resurgent-current blocking particle or could obstruct access to the pore for this blocker. This could explain the relationship we observed between enhanced SI and attenuated resurgent currents. Accordingly, the impairment of SI could increase the availability of the pore receptor for the blocking particle, leading to greater resurgent currents. While we cannot rule out that mutagenesis might have knocked out a direct binding contact for the resurgent-current blocker, the congruity of our findings across six constructs involving three different mutations For each channel the maximal inward current and corresponding potential is colored in green. (B) Steady-state SI is enhanced by the Nav1.6/N1455A mutation (open blue circles, V 1/2 = − 81.9 ± 2.5 mV, n = 4) and decreased by the Nav1.6/V966C mutation (crossed red circles, V 1/2 = − 33.8 ± 5.1 mV, n = 5) compared to Nav1.6 WT (filled black circles, V 1/2 = − 59.6 ± 1.5 mV, n = 13). (C) Normalized conductance-voltage relationship is shifted to more depolarized potentials for both Nav1.6 mutants whereas (D) the V 1/2 of steady-state fast inactivation is slightly hyperpolarized. (E) Nav1.6 WT (filled black circles) and the mutant Nav1.6/V996C (crossed red circles) display persistent currents but the mutant Nav1.6/N1455A (open blue circles) does not. Data shown in E were recorded in the presence of β 4-peptide and 50 nM ATX-II and were significantly different for Nav1.6 WT and Nav1.6/N1455A (blue*), for Nav1.6 WT and Nav1.6/V966C (red*) and for Nav1.6/N1455A and Nav1.6/V966C (black # ), p < 0.05. and two distinct Nav subtypes indicates that a general correlation exists between the extent of SI in a channel and its capability to generate resurgent currents: a negative shift in the voltage-dependence of SI is associated with decreased resurgent currents.
Most IEM mutations not only enhance SI but also increase its voltage-sensitivity 4,10,36-43 . We also observed a steeper SI-voltage relationship in the Nav1.7/Δ L955 and Nav1.6/N1455A mutants (Figs 1 and 4), which may The maximal resurgent current obtained by a repolarizing pulse to − 10 mV is colored green. Cells expressing Nav1.6 WT and the Nav1.6/V966C mutant show robust resurgent currents (B) in 90 to 100% of cells (black bars) (C) whereas the N1455A mutant was not able to produce any resurgent currents. Significant differences between Nav1.6 WT and Nav1.6/N1455A are marked with blue*, between Nav1.6 WT and Nav1.6/V966C with red* and between Nav1.6/N1455A and Nav1.6/V966C with # p < 0.05. 1 mV for Nav1.7 control (filled black squares, n = 10) and − 99.8 ± 1.2 mV for Nav1.7 500 μ M LCM (filled green squares, n = 6)). The data were recorded with the protocol depicted in Fig. 3 (inset). (B) Representative resurgent currents traces of Nav1.7 recorded at the indicated test pulses before (black) and after application of 500μ M LCM (green) via perfusion to the same cell. (C) Resurgent current amplitude (black squares) decreased after the application of 500 μ M LCM (filled green squares, n = 6). Transient peak current was reduced by LCM, therefore resurgent currents are displayed as percentage of transient peak currents of each trace and plotted as a function of voltage. *p < 0.05 and **p < 0.005 (paired t-test).
Scientific RepoRts | 6:25974 | DOI: 10.1038/srep25974 influence the overall conformation of the pore, rendering it more likely to undergo SI and thus impairing the binding of an open channel blocker.
Resurgent currents have not been found in any IEM-associated mutations of Nav1.7 investigated to date 1 . In light of our current findings, we suggest that the hyperpolarized shift of SI that is common to IEM mutants could account for this absence of resurgent currents. Indeed, this relationship between enhanced SI and absent resurgent currents may be a key determinant of IEM symptoms. For example, in a recent study Emery et al. 3 characterized the Nav1.7/L245V mutation that exhibits depolarized steady-state fast inactivation with no change of activation. Despite large persistent currents that have up-to-now been characteristic of PEPD, the clinical phenotype of the Nav1.7/L245V patient was IEM. In agreement with our results and in support of our hypothesis, the mutant displayed no resurgent currents and SI was enhanced.
Resurgent currents appear to be decisive for the PEPD phenotype, whereas persistent currents may not be as important as previously suggested 2,8,9,27 , especially considering the Nav1.7/Δ L955 and L245V IEM-associated mutations that both exhibit large persistent currents. Every PEPD mutant tested so far has been shown to produce resurgent currents and none has been reported with enhanced SI; in these channels SI remains unmodified or is impaired. Therefore it may be the interplay between SI and resurgent currents that ultimately determines whether the PEPD or IEM phenotype manifests. LCM is an antiepileptic drug in clinical use that enhances SI of Nav1.7 19 . Our experiments show a strong reduction of resurgent currents by this compound, suggesting that using drugs that selectively target SI of Nav1.7 may be helpful in the future to treat patients suffering from PEPD or other forms of pain.
In this study we determined a negative correlation between SI and resurgent currents. Our findings may have clinical relevance, as enhancement of SI by LCM, a drug already in clinical use, reduced resurgent currents and therefore has potential as treatments for PEPD patients. Beside LCM other SI-enhancing drugs are known, such as eslicarbazepine 44 . In general SI-modifying compounds might offer new opportunities in the treatment not only of hereditary pain disorders but for the therapy of sporadic chronic and acute pain.
HEK293 cells were maintained in DMEM medium (Gibco-Life technologies) including 10% FBS, 1.0 g/l Glucose and 1% Penicillin/Streptomycin (PAA Laboratories GmbH) whereas neuronal cell lines required 4.5 g/l Glucose. If not indicated otherwise 0.5 μ g EGFP-C1 (Clontech Laboratories, Inc., Mountain View, USA) was cotransfected in order to detect transfected cells via green fluorescence. Except for measurement of resurgent currents mNav1.6 and its mutants were coexpressed with a GPF-tagged mouse β 4-subunit in a 4:1 ratio to enhance current density. Cells were recorded 1 to 2 days after transfection.
Capacitive transients were cancelled and series resistances (< 5 MΩ) were compensated by at least 65%. Leak currents were subtracted digitally online using the P/4 procedure following the test pulses. Signals were digitized at sampling rates between 20 kHz to 100 kHz adapted to the voltage protocol length and purpose. Voltage protocols were carried out after current stabilization and equilibration was established using a holding potential of − 120 mV. For SI measurements of hNav1.7 and its mutations a holding potential of − 140 mV was used. Patchmaster/Fitmaster software (HEKA Elektronik, Lamprecht, Germany) was used for acquisition and off-line analysis.
Current-voltage (I-V) relations were obtained using 100 ms pulses to a range of test potentials in 10 mV steps with an interpulse interval of 5 seconds. The voltage-dependent sodium channel conductance G Na was calculated using the following equation: G Na = I Na /(V m − E rev ) where I Na is the amplitude of the current at the voltage V m , and E rev is the reversal potential for sodium, which was determined for each cell individually.
Activation curves were derived by plotting normalized G Na as a function of test potential and fitted with the Boltzmann equation: G Na = G Na, max /(1 + exp [(V m − V 1/2 )/k]) where G Na, Max is the maximum sodium conductance, V 1/2 is the membrane potential at half-maximal activation, V m is the membrane voltage and k is the slope factor.
Persistent sodium currents are given as the mean remaining sodium current between 66 to 99 ms for each test pulse in the presence of 50 nM ATX-II in the bath and 100 μ M β 4-peptide in the pipette.
Resurgent currents were assessed using a 20 ms depolarizing voltage step to + 30 mV to allow channel opening followed by a 500 ms test pulse in steps of 10 mV (ranging from − 80 mV to 20 mV) with ATX-II in the bath and the β 4-peptide in the internal solution. Peak resurgent currents were analyzed between 0.6 ms and 5 ms of the repolarizing test pulse and only currents showing a slowly activating component which was clearly distinguishable from tail currents were counted as resurgent current positive. Activation traces with the respective voltage step were subtracted from recorded resurgent current traces to minimize the artifact of persistent currents which occurs after the first depolarizing voltage pulse in the resurgent current protocol, especially in the hNav1.7/Δ L955 mutant. The resulting resurgent currents are given in absolute values measured as the peak inward current following the prepulse and are plotted against the voltage of the test pulse.
Voltage-dependence of steady-state fast inactivation was measured using a series of 500 ms prepulses (−130 mV or − 150 mV for mNav1.6 or hNav1.7 respectively to − 10 mV in each case), followed by a 40 ms supramaximal depolarization to + 20 mV (mNav1.6) or 0 mV (hNav1.7) that served as a test pulse to assess the available non-inactivated channels. Normalized peak inward current amplitude (I Na /I Na, Max ) at each test pulse is displayed Scientific RepoRts | 6:25974 | DOI: 10.1038/srep25974 as a function of prepulse potential (V m ) and fitted using the following (Boltzmann) equation I Na /I Na, max (V) = 1/(1+ exp[(V m − V 1/2 )/k]), resulting in V 1/2 (the potential of half maximal inactivation) and the slope factor k.
Steady-state SI of wildtype and mutant hNav1.7 was induced by a 30 s conditioning pulse to various potentials in steps of 20 mV followed by a 100 ms step to − 120 mV to remove fast inactivation. Channel availability was assessed by a 40 ms test pulse to 0 mV. To guarantee full recovery of inactivation a 90 s sweep interval was used. To correct for channel rundown during the protocol an additional 10 ms test pulse to 0 mV was implemented before the conditioning pulse and served as reference. The same voltage protocol was used to compare SI of hNav1.7 and mNav1.6r WT seen in Fig. 3.
To examine steady-state SI of mNav1.6 and its mutants a conditioning prepulse of 60 s ranging from − 130 to + 10 mV in steps of 10 mV was followed by a 100 ms step to − 120 mV to remove fast inactivation and a test pulse to + 20 mV for 40 ms, at which available channels were assessed. Normalized current amplitude was plotted against prepulse potential and fitted by a Boltzmann equation as described for steady-state fast inactivation.
SI, activation and steady-state fast inactivation were determined in the presence of 500 μ M LCM in the external bath solution (preincubation), where indicated. The effect of LCM on resurgent currents was obtained by perfusion of external bath solution containing 500 μ M LCM to each recorded cell. Conductance data obtained from activation protocols were normalized to the maximum conductance value. Normalized peak inward currents obtained from steady-state fast inactivation and SI protocols were fitted with a single Boltzmann equation.

Chemicals.
For fitting of SI the data points at − 100 and − 80 mV for Nav1.7/V948C mutation were neglected, to assume a single exponential fit for better comparison.
For statistical testing groups were compared by an ANOVA or a Kruskal-Wallis test in case of non-parametric testing followed by a Tukey post-hoc analysis (significance at least p < 0.05 for*, p < 0.005 for** and p < 0.001 for***).