Mutations in the voltage sensor domain (VSD) of CaV1.1, the α1S subunit of the L-type calcium channel in skeletal muscle, are an established cause of hypokalemic periodic paralysis (HypoPP). Of the 10 reported mutations, 9 are missense substitutions of outer arginine residues (R1 or R2) in the S4 transmembrane segments of the homologous domain II, III (DIII), or IV. The prevailing view is that R/X mutations create an anomalous ion conduction pathway in the VSD, and this so-called gating pore current is the basis for paradoxical depolarization of the resting potential and weakness in low potassium for HypoPP fibers. Gating pore currents have been observed for four of the five CaV1.1 HypoPP mutant channels studied to date, the one exception being the charge-conserving R897K in R1 of DIII. We tested whether gating pore currents are detectable for the other three HypoPP CaV1.1 mutations in DIII. For the less conserved R1 mutation, R897S, gating pore currents with exceptionally large amplitude were observed, correlating with the severe clinical phenotype of these patients. At the R2 residue, gating pore currents were detected for R900G but not R900S. These findings show that gating pore currents may occur with missense mutations at R1 or R2 in S4 of DIII and that the magnitude of this anomalous inward current is mutation specific.

Hypokalemic periodic paralysis (HypoPP) is an inherited disorder of skeletal muscle in which recurrent episodes of weakness occur in association with low extracellular [K+], <3 mM (normal range 3.5–5.5 mM; Cannon, 2015). During an attack, the resting potential is paradoxically depolarized, and this sustained depolarization inactivates sodium channels, thereby reducing fiber excitability and causing weakness. The molecular basis for HypoPP is heterogeneous, with missense mutations in the pore-forming α1S subunit of the voltage-gated L-type Ca2+ channel (CaV1.1) accounting for ∼60% of cases and missense mutations in the α subunit of NaV1.4 occurring in another 20% (Sternberg et al., 2001). Remarkably, 9 of 10 established HypoPP mutations in CaV1.1 and all 13 in NaV1.4 are missense mutations at arginine residues in the outer positions (R1 or R2) of S4 transmembrane segments in the voltage sensor domains (VSDs; Cannon, 2017; Matthews et al., 2009).

The conserved structural motif of HypoPP mutations has an accompanying commonality of functional defects. Scanning mutagenesis studies, originally designed to map voltage-dependent movement of S4 segments in voltage-gated K+ (KV) channels, revealed the surprising finding that R/X mutations may create an anomalous leakage current in the VSD (Starace and Bezanilla, 2001; Tombola et al., 2005). When R1 or R2 is mutated, ion conduction through this “gating pore” is permissive at hyperpolarized potentials when the S4 is in the inward or “down” conformation and the R1 or R2 residue lies near the hydrophobic charge transfer center (CTC) of the VSD. Depolarization favors outward movement of S4, which removes the R/X mutation from the critical hydrophobic CTC, and the anomalous gating pore becomes nonconducting (Monteleone et al., 2017; Moreau et al., 2014). This insight directly led to the hypothesis that R/X mutations in HypoPP might also give rise to gating pore currents and that this would produce an anomalous inward current at the resting potential of muscle fibers. This notion was quickly confirmed for HypoPP mutations in NaV1.4 (Sokolov et al., 2007; Struyk and Cannon, 2007) and was subsequently verified for 12 of 13 R/X mutations in NaV1.4, with the one exception being the only known charge-conserving mutation, R219K (Kubota et al., 2020). Moreover, model simulations show that the small anomalous conductance of the gating pore current (only ∼1% of the total resting conductance in a muscle fiber) is sufficient to cause paradoxical depolarization of the resting potential in low K+ (Jurkat-Rott et al., 2012; Struyk and Cannon, 2008).

Confirmation that HypoPP R/X mutations in S4 segments of CaV1.1 are also permissive for gating pore currents, $Igp,$ lagged behind the studies of NaV1.4. This delay occurred because in heterologous systems CaV1.1 does not express well at the plasma membrane, and high levels of expression are required for sufficient signal to noise to detect $Igp.$ In a knock-in mutant mouse model of CaV1.1-R528H HypoPP, but not in WT, we observed an aberrant inward current that was not blocked by dihydropyridines, tetrodotoxin, TEA, 4-aminopyridine, or 9-anthracenecarboxylic acid, consistent with a gating pore current (Wu et al., 2012). Expression of HypoPP CaV1.1 mutant constructs by electroporation of muscle fibers in vivo also showed evidence of an anomalous inward current consistent with $Igp$ for R1239H (Fuster et al., 2017a) and the atypical mutation V876E in domain III (DIII) S3 (Fuster et al., 2017b), which is the only well-established example of a HypoPP mutation that is not R/X in S4. The detection of $Igp$ was technically challenging for all of these studies in mouse muscle fibers because it is difficult to sufficiently block all of the endogenous conductances to clearly observe the gating pore current. For example, the evidence for $Igp$ in the electroporation studies was based on a 20% increase of the inward current compared with WT and required a large sample size of ∼20 fibers to detect this difference. Currently, no specific blockers of $Igp$ with sufficient potency have been identified that would enable a pharmacologic leak subtraction strategy to isolate the gating pore current.

More recently, an accessory protein, Stac3, was identified (Horstick et al., 2013; Nelson et al., 2013) that is essential for skeletal muscle–type excitation–contraction coupling and that greatly enhances expression of CaV1.1 at the plasma membrane (Polster et al., 2015). We showed that Stac3 coexpression with L-type Ca2+ channel subunits in Xenopus laevis oocytes dramatically increases Ca2+ current by 100-fold (Wu et al., 2018) and achieves sufficient channel levels to detect gating pore currents for the CaV1.1 HypoPP mutations in DII (R528H and R528G). Four HypoPP mutations have been reported in DIII of CaV1.1 (R897S, R897K, R900G, and R900S). In a collaborative study, we recently showed that the atypical charge-conserving mutation at R1 in DIII, R897K, does not lead to a gating pore current detectable in the oocyte expression study (Kubota et al., 2020). In the present study, we ascertained whether gating pore currents are detectable for the other three CaV1.1 HypoPP mutations in DIII. Of note, the R897S mutation is associated with a severe clinical phenotype (Chabrier et al., 2008; Hanchard et al., 2013), and we found the largest amplitude $Igp$ encountered for any HypoPP mutant channel (NaV1.4 or CaV1.1), even though the charge-conserving mutation at the same residue, R897K, had no detectable $Igp.$ Similarly, at the R2 position in DIII, the gating pore current was detected for R900G but not for substitution of the larger residue R900S.

### Expression of human CaV1.1 (hCaV1.1) channels in oocytes

The human α1S subunit of the skeletal muscle Ca2+ channel was expressed in Xenopus oocytes as previously described (Wu et al., 2018). Briefly, the hCaV1.1, rabbit β1a, and mouse Stac3 cDNAs were subcloned into an oocyte-optimized expression vector, pGEMHE (Liman et al., 1992), whereas the rat α21b cDNA was inserted into pcDNA3 (Invitrogen). Site-directed mutagenesis of hCaV1.1 was performed using the QuikChange II Mutagenesis Kit (Agilent) to create the HypoPP missense mutations R897S, R900G, and R900S, verified by sequencing the entire cDNA insert and flanking regions. Complementary RNA was synthesized by in vitro transcription (∼2 ng/nl) using the mMESSAGE mMACHINE kit (Ambion).

Oocytes were harvested by partial oophorectomy and defolliculated in collagenase type I at room temperature for 2 h. Equal volumes of all four transcripts (hCaV1.1, β1b, α21b, and Stac3) were combined in a master mix, and 50 nl was injected into the animal pole of each oocyte. Oocytes were maintained at 18°C in 0.5× Leibovitz’s L-15 medium (HyClone) supplemented with 1% horse serum, 100 U/ml penicillin, 100 µg/ml streptomycin, and 100 µg/ml amikacin. All experiments were performed within the guidelines established by the University of California, Los Angeles, Institutional Animal Care and Use Committee.

### Electrophysiology

Currents were recorded 3–6 d after RNA injections using the cut-open oocyte voltage clamp in headstage mode (Dagan CA-1B) as previously described (Wu et al., 2018). After mounting the oocyte in the recording chamber, the membrane exposed to the bottom compartment was permeabilized by brief application of 0.1% saponin. A voltage-sensing microelectrode was inserted into the dome of the oocyte protruding into the upper compartment. The upper and middle compartments were actively clamped using platinum wire-threaded salt bridges filled with Cl free agar (1 M sodium methanesulfonate and 10 mM HEPES, pH 7.0). The linear membrane capacitance was compensated with the analogue circuitry of the amplifier, but no offset or linear leak subtraction was performed. Currents were low-pass filtered at 10 kHz to measure gating charge displacement and at 1 kHz for measuring ionic currents. Signals were sampled at 20 kHz with 16-bit resolution.

Chloride-free solutions were used for all recordings to minimize the contribution from endogenous chloride conductances in the oocyte. Prior to mounting the oocyte in the recording chamber, oocytes were injected with 100 nl of a BAPTA solution (100 mM BAPTA-K4 and 10 HEPES, pH 7.0 with Mes) to buffer intracellular Ca2+ and further reduce endogenous Ca2+-activated Cl currents. The internal solution (bottom compartment) contained 96 mM KOH and 10 mM HEPES adjusted to pH 7.0 with methanesulfonic acid. The standard external solution used for the middle (guard) and upper compartments contained (in mM): 96 NaOH, 6 calcium acetate, 10 HEPES, and 0.1 ouabain titrated to pH 7.0 with methanesulfonic acid. For measurement of ionic current conducted by the main pore of the CaV1.1 subunit, the Ca2+ of the external solution was replaced by 10 mM barium acetate. Gating charge-displacement currents were measured by substituting 2 mM Co2+ for Ca2+ in the external solution to block conduction by the pore. For studies of the permeation properties of the gating pore current, the external Na+ was replaced by 110 mM NMDG (see Fig. 7) or a mixture of 60 mM NMDG and 60 mM guanidinium (see Fig. 9). Block by divalent cations (see Fig. 8) was assessed by bath exchange of the top compartment from a control with 6 mM Ca2+ to test solutions with 4 mM Ca2+ and 2 mM of either Ba2+ or Co2+.

### Data analysis

The peak I-V relation for the Ba2+ currents was fit to an ohmic conductance that was scaled by a Boltzmann function to model the voltage dependence of channel activation:
where $Gmax$ is the maximal conductance, $Erev$ is the fitted reversal potential, $V1/2$ is the voltage midpoint for activation, and $k$ is the steepness factor. The inward Ba2+ current increased monotonically with time (see Figs. 1 and 3 A), and although the time course was more complex than a single exponential, we used a single exponential fit to compute an apparent time constant because this approach of using a single number facilitates the comparison between WT and HypoPP mutant channels. A more exact fit using a double exponential showed that ∼80% of the response can be accounted for by a single exponential component. A qualitative comparison was also performed by amplitude scaling of the currents to the maximum value and superimposing the traces (see Fig. 3 A).
The “on” gating charge, $Q(V),$ was calculated as the area under the charge-displacement current elicited by a voltage step from −100 mV. Most of the linear capacitance of the membrane was compensated by the analogue circuitry of the Dagan CA-1B, and the residual linear capacitance and nonspecific leakage current were compensated by subtracting a scaled version of the passive linear response for a depolarization from −100 mV to −80 mV. Secondary offset subtraction of the current 5 ms after the voltage step was used to compensate for potential effects from a nonlinear gating pore current, as previously described (Mi et al., 2014). The $Q(V)$ relation was fit to a Boltzmann distribution:
where $Qmax$ is the maximal $Qon,$$V1/2$ is the voltage at which half the charge has moved, and $k$ is the steepness factor.

The detection of gating pore currents was optimized by measuring the isochronal current over the interval 5–8 ms, as done in our previous study of VSDII HypoPP mutant channels (Wu et al., 2018). This delay allows the gating charge-displacement current to decay back to 0 nA and at the same time is fast compared with the opening of the channel pore such that the majority of the ionic current is conducted by the gaiting pore (plus nonspecific leakage).

In all figures, the error bars indicate the SEM. Statistical tests for the difference between the mean values for parameter estimates of four constructs (WT and three HypoPP mutants) were assessed by two-tailed ANOVA with Bonferroni correction.

A simulation is provided in the supplemental text at the end of the PDF to show the feasibility of and potential sources of error for measuring charge displacement in the presence of a large-amplitude gating pore current.

### Online supplemental material

Fig. S1 shows simulated currents produced by the sum of leak $Ileak,$ gating pore $Igp,$ and charge-displacement components $Iq.$Fig. S2 shows simulated $Q(V)$ curves. Table S1 lists model parameters. Supplemental text provides a numerical simulation to demonstrate the feasibility of estimating the gating charge displacement, based on conventional P/N leak-subtracted responses, despite the presence of a contribution from the nonlinear, voltage-dependent gating pore current.

### HypoPP DIII mutant channels are expressed in the plasma membrane and conduct Ba2+

Ionic currents were recorded in the cut-open oocyte configuration with 10 mM external Ba2+ as the charge carrier and no added Ca2+. Large-amplitude, slowly activating inward currents were observed at test potentials more positive than 0 mV for oocytes expressing WT and all three HypoPP mutant constructs (Fig. 1). The linear capacitive transient has been suppressed by analogue compensation of the amplifier, but otherwise these traces show currents without linear leak or offset subtraction. At the holding potential of −100 mV, an inward holding current was consistently larger for the R897S and R900G constructs than for WT or R900S, as shown in Fig. 1. The current amplitudes for test potentials from −50 to −10 mV spanned a larger range for R897S and R900G (e.g., curly bracket in Fig. 1, B and C) than for WT or R900S. These differences reflect a larger conductance in a voltage range below the activation of hCaV1.1 channels for oocytes expressing R897S and R900G. This aberrant conductance in R897S and R900G mutant channels is the gating pore, and the anomalous currents are large enough to be detected compared with the Ba2+ current conducted by the central pore of hCaV1.1.

Current conducted by the hCaV1.1 pore was isolated as the component of the total current that was blocked by a combination of 10 µM nifedipine plus 2 mM Co2+. Fig. 2 A shows an example of the nifedipine/Co2+-sensitive current elicited by a test pulse to +25 mV in an oocyte expressing the R900S construct. Most of the total current was conducted by the hCaV1.1 pore, as shown in the inset by the comparison of the current without blockers (solid line) and with blockers (dashed line). The I-V relation, measured as the nifedipine/Co2+-sensitive current at the end of a 250-ms pulse, shows the current amplitude was smaller for HypoPP mutant constructs than for WT (Fig. 2 B). On average, the peak inward current amplitude was 42%, 41%, and 67% of WT for R897S, R900G, and R900S, respectively (Table 1, $Gmax$ values).

The voltage dependence of activation for $IBa$ was shifted by 8 mV in the depolarized direction for R900G, whereas the voltage midpoint was not statistically different from WT for R897S or R900S (Fig. 2 C and Table 1). Although all three HypoPP mutations neutralize a gating charge in the DIII VSD, only the R897S mutation in the outermost R1 position decreased the slope of the G-V relation $(k$ = 7.9 ± 0.34 mV compared with 6.8 ± 0.34 in WT).

The Ba2+ currents for the representative cells shown in Fig. 1 have slower activation kinetics for all three HypoPP mutant channels than for WT. This difference was consistently observed, as shown by superposition for the averages of amplitude-normalized Ba2+ currents elicited by a voltage step to 25 mV (Fig. 3 A). To quantitatively compare this difference, we fit the rising phase of the Ba2+ current with a single exponential, which accounts for ∼80% of this monotonically increasing transient. This approximation facilitates the comparison with a single activation time constant $(τact).$ The rate of activation was approximately twofold slower for R897S and R900G mutant channels than for WT (20–40-mV range; ANOVA P < 0.01), whereas for R900S, the 1.5-fold slower trend was not statistically different from WT (Fig. 3 B).

We measured the gating charge displacement as an index of channel expression level at the membrane and to assess the voltage dependence for displacement of the voltage sensor. The inward rectification of gating pore currents produces a nonlinearity that may corrupt the standard P/N linear leak subtraction method used to isolate the charge-displacement current. We previously showed that secondary offset subtraction of the steady-state current (after the charge-displacement transient has decayed) can be used to compensate for this distortion (Mi et al., 2014). For those data, the amplitude of the gating pore current was ∼5% of the peak charge-displacement transient, and the estimated uncertainty for $Qmax$ was ±7.5%. For R897S, however, the amplitude of the gating pore current was much larger (Fig. 1 B), and consequently the effect on P/N leak subtraction was much greater. Fig. 4 A shows representative current transients, without leak or offset subtraction, that were used to determine $Q(V)$ for R897S. The corresponding “leak trace” for a small voltage step from −100 mV to −80 mV is shown in Fig. 4 B. The conductance is rather large (150 nA/20 mV) because of contributions from both the “open” gating pore and the nonspecific leakage current. When strongly depolarized voltage steps were applied to measure charge displacement, the gating pore closed, so the linearly scaled P/N subtraction erroneously overcompensated and caused a substantial negative current in steady state (Fig. 4 C, at 5 ms). A secondary offset subtraction of this steady-state current was performed to derive the transient response that was integrated to calculate $Q$ (Fig. 4 D).

Using this method, the maximal charge displacement, $Qmax,$ was half that of WT for all three HypoPP mutant constructs (Fig. 5 A; P < 0.001). A commensurate decrease was observed in the peak Ba2+ current (Fig. 2), consistent with the notion that the expression of functional channels in the membrane for all three HypoPP mutant constructs was about half of that for WT. A plot of the relative charge displacement as a function of membrane potential showed a 21-mV leftward shift for oocytes expressing R897S compared with all other constructs (Fig. 5 B and Table 2). Neutralization of a gating charge (R1 or R2) in DIII did not reduce the slope of the $Q(V)$ curve compared with WT.

We performed model simulations (see supplemental text and Table S1) to estimate the possible error in determining $Q(V)$ that may have been introduced by residual effects of the nonlinear gating pore current, especially for R897S. In the limit, if the rate of gating pore closure is instantaneous, then there is no error. On the other hand, the slowest possible rate for closure of the gating pore would be with the same kinetics as the decay for the total charge displacement (movement of S4 is expected to directly close the gating pore, with no delay from allosteric coupling). In this slowest possible case, the contribution to the transient from the gating pore current would cause the amplitude of the apparent $Q(V)$ curve to be overestimated and to increase linearly at strongly depolarized potentials (Fig. S1 and Fig. S2). However, the $Q(V)$ for R897S in Fig. 5 B showed a clear plateau, suggesting that the gating pore does indeed close faster than the rate of decay for the charge displacement, and therefore our estimate for is reasonable. Importantly, over the entire range of simulated gating pore closing rates, the apparent voltage midpoint of $Q(V)$ did not have a leftward shift (Fig. S2 B). Therefore, we interpret the 21-mV leftward shift in Fig. 5 B to be an intrinsic feature of R897S gating and not an artifact of the method used to calculate charge displacement.

### Gating pore currents were detected for R897S and R900G

The protocol we previously developed to detect gating pore currents for hCaV1.1 DII HypoPP mutant channels (Wu et al., 2018) was used to test for detectable $IGP$ in the DIII HypoPP mutants. Voltage steps were applied from a holding potential of −100 mV, and most of the capacitive current was compensated by the analogue circuitry of the amplifier. No offset or linear leak subtraction was performed. The representative responses in Fig. 6 A show a larger holding current of −200 to −400 nA at $Vhold$ = −100 mV for R900G and R897S compared with WT or R900S hCaV1.1 (approximately −100 nA). These larger holding currents for R897S and R900G were associated with a higher conductance (i.e., wider spread of ionic current levels) over the voltage range from −140 mV to −40 mV, which is hyperpolarized relative to activation of the Ba2+ current conducted by the pore of hCaV1.1 (Fig. 2). Moreover, the response for R897S shows clear evidence for inward rectification (Fig. 6 A). These features are consistent with an anomalous gating pore current, $IGP,$ for R897S and R900G HypoPP mutant channels.

To quantify $IGP,$ we determined the time-averaged current over the interval 5–8 ms, which is late enough for the residual capacitive and charge-displacement currents to settle and is too early for activation of the Ba2+ current conducted by the main pore of hCaV1.1 for $Vstep$ <10 mV. Average values for this isochronal current are plotted as a function of $Vstep$ in Fig. 6 B. The plots diverge at hyperpolarized potentials, with strong inward rectification for R897S and a clear separation from WT for R900G (P < 0.05 for $Vstep$ less than or equal to −85 mV). The measured current is a combination of nonspecific leakage current and $IGP.$ Because effective blockers have not been identified for $IGP$ in S4 R/X missense mutants of CaV, NaV, or Kv channels, it is not possible to pharmacologically separate $IGP$ from the total current. Moreover, a post hoc linear subtraction of the leak component, used to extract the inwardly rectifying $IGP$ for S4 R/X mutations in NaV (Sokolov et al., 2007; Struyk and Cannon, 2007), is not feasible for hCaV1.1, because $IGP$ is usually linear (e.g., R528H and R528G in DII [Wu et al., 2018] and R900G herein). The gating pore current observed for R897S is remarkable because it is the only example of an hCaV1.1 HypoPP mutation where $IGP$ shows strong inward rectification and because the amplitude is unusually large. The magnitude of the gating pore effect shown for R900G in Fig. 6 B is more typical for HypoPP hCaV1.1 mutant channels studied in the oocyte expression system. For example, in our prior study of R528H and R528G in DII (Wu et al., 2018), the total inward current at hyperpolarized potentials (nonspecific leak plus $IGP)$ was approximately twice the amplitude of that observed for WT hCaV1.1. For a DIV HypoPP mutation (R1239H) and the atypical mutation in S3 of DIII (V876E), the total inward current was only ∼40% larger than for WT when studied by plasmid electroporation into muscle of normal mice (Fuster et al., 2017a; Fuster et al., 2017b).

### Permeation of the gating pore in DIII HypoPP hCaV1.1

Permeation through an anomalous gating pore is variable and mutation specific. Gating pore currents may be either nonselective for monovalent cations or proton selective (especially for R/H missense mutations), may be enhanced in extracellular guanidinium, and may be weakly blocked by multivalent cations (Sokolov et al., 2010; Starace and Bezanilla, 2004; Struyk and Cannon, 2007). Our initial measurements were performed in the context of the ion gradient for a normal mammalian cell, with Na+ as the predominant extracellular cation and K+ as the intracellular cation. Under these conditions, the reversal potential was approximately −20 mV (Fig. 6 B), consistent with a gating pore that is permeable to both Na+ and K+. The resting potential of skeletal muscle is substantially hyperpolarized (−90 mV) compared with the reversal potential of the gating pore, so, under physiological conditions, the predominant charge carrier would be an influx of Na+.

The inward gating pore current is much larger in external guanidinium than for Na+ with many S4 R/X HypoPP mutations in NaV1.4 (Sokolov et al., 2010; Thor et al., 2019) and with the HypoPP CaV1.1-R528G mutation in DII (Wu et al., 2018). We assessed guanidinium permeability by initially recording the nonspecific leakage current in external NMDG and then repeated the measurement after exchanging the bath with a mixture of 60 mM guanidinium and 60 mM NMDG. Fig. 7 A (left panel) shows representative currents recorded from R897S-expressing oocytes in NMDG. Compared with the response in an Na+ bath (Fig. 6, different R897S oocytes), the holding current at −100 mV is smaller, and the currents elicited over a range of voltage steps are more tightly clustered (i.e., lower conductance) with little apparent inward rectification. These differences show that the R897S gating pore is much less permeable to NMDG than to Na+. Moreover, this result also shows that R897S is not a proton-selective gating pore, because the currents have the same amplitude in Na+ or NMDG for those HypoPP mutant channels (e.g., R667H and R669H in NaV1.4; Struyk and Cannon, 2007; Struyk et al., 2008).

When the bath was exchanged from NMDG to a mixture of NMDG and guanidinium, the small ionic currents became very large and inwardly rectifying, as shown for the representative oocyte in Fig. 7 A (right panel). This dramatic change is strong supportive evidence for the presence of an anomalous gating pore conductance (Sokolov et al., 2010; Tombola et al., 2005), although not all gating pore conductances show this behavior (Wu et al., 2018). The average I-V response over the 5–8-ms isochronal interval is shown in Fig. 7 B for oocytes expressing WT and all three mutant constructs. Currents were recorded from each oocyte in NMDG, and then the same oocyte was tested with NMDG plus guanidinium. Only R897S oocytes had enhancement of the inwardly rectifying current in guanidinium (Fig. 7 B, upper left panel, red symbols). Although R900G showed clear evidence of a (nonrectifying) gating pore current in a Na+ bath (Fig. 6), guanidium did not enhance the conductance (Fig. 7 B, upper right, blue symbols). In all oocytes, the reversal potential depolarized after the bath exchange (Fig. 7 B, open to closed symbol transition; see inset in upper left for $Erev$ of R897S), which implies that the small nonspecific leak was more permeable to guanidinium than to NMDG.

We calculated the slope of the isochronal I-V plots over the voltage range from −140 mV to −120 mV to quantitatively compare the membrane conductance for each CaV1.1 construct and how this changed for the different ionic conditions describe above. The slope conductances in Na+, NMDG, and guanidinium are shown as a cluster of three bars for each construct in Fig. 8. In Na+, the conductance was larger for the two HypoPP mutants with gating pore currents (R897S and R900G; P < 0.0001 and P < 0.01, respectively) than for WT or R900S. These larger conductances for R897S and R900G were both markedly reduced in external NMDG (lower by 9.8 μS for R897S and by 2.7 μS for R900G). By comparison, the conductance in NMDG was 1.1 μS lower for WT. The larger reductions of conductance in mutant channels (R897S and R900G) than in WT shows that the lower conductance in NDMG is caused primarily by an effect on the gating pore component of the current rather than by the nonspecific leakage current. These observations show that the gating pores are more permeable to Na+ than to NMDG and imply that Na+ influx is the major charge carrier of the gating pore current at the resting potential of skeletal muscle. The conductance in guanidinium is fourfold higher than in Na+ for R897S. The other CaV1.1 constructs had either a conductance decrease (R900G and WT) or no change (R900S) in guanidinium.

Divalent block of the NaV1.4 gating pore for mutant channels found in HypoPP is modest (Kd values of ∼0.5 to 3 mM) with block by Ca2+ < Ba2+ < Ni2+$≈$ Zn2+ (Francis et al., 2011; Sokolov et al., 2007). We assessed sensitivity to divalent block by comparison of the current in 6 mM Ca2+ (control) with the response in the same oocyte after bath exchange with 2 mM test divalent cation plus 4 mM Ca2+. A total divalent of 6 mM was maintained to keep the effects of screening membrane surface charge relatively constant and to avoid the development of nonspecific leakage current in low divalent cation conditions. From the holding potential of −100 mV, 2 mM Ba2+ reduced the current at $Vstep$ −120 mV by 290 ± 41 nA in oocytes expressing R897S and by 83 ± 11 nA for R900G (Fig. 9, left column). Because the measured current is a combination of the gating pore and nonspecific leakage components, it is not possible to quantitatively estimate the potency of block. These data, however, are consistent with block of $Igp$ because, in oocytes expressing WT channels, 2 mM Ba2+ reduced the current at −120 mV by only 62 ± 5.1 nA. We also tested Co2+ because this divalent cation blocks the Ca2+-conducting pore of CaV1.1 and thereby is potentially a good condition in which to record $Igp$ without interference from main pore currents. Application of 2 mM Co2+ to oocytes expressing R897S or R900G did not produce a detectable block of $Igp,$ as shown by the absence of an effect on the total current over the range −120 mV to 10 mV (Fig. 9, right column).

This study is the first opportunity to ascertain whether the canonical HypoPP mutations in VSDIII of CaV1.1 (i.e., R/X substitutions in S4 that do not conserve charge) give rise to gating pore currents. These anomalous currents were clearly detected for two mutations (R897S and R900G), but not a third (R900S). The gating pore current is widely accepted as the primary defect that causes susceptibility to recurrent episodes of depolarization-induced loss of excitability and weakness in HypoPP (Cannon, 2015; Cannon, 2017). This anomalous current has been observed for six of eight HypoPP mutations that have been functionally studied with sufficient resolution to detect these small-amplitude currents (Table 3). In further support of this proposed pathomechanism, a gating pore current was detected for V876E in DIII-S3, which is the only established HypoPP mutation that is not an R/X substitution in an S4 segment (Fuster et al., 2017b). Failure to detect a gating pore current for HypoPP mutant CaV1.1 channels has occurred twice: for an unusual mutation that preserves charge, R897K (Kubota et al., 2020), and for R900S in this study. Interestingly, gating pore currents are unequivocally detected for alternative missense changes at both of these same residues: R897S and R900G.

The gating pore currents observed for R897S are exceptional for two reasons. First, the magnitude of the current is larger than for $Igp$ recorded from any other HypoPP mutation of CaV1.1 or NaV1.4 that has been studied in the oocyte expression system. The gating pore current for HypoPP mutant channels is typically 50–100 nA at −100 mV in 96 mM external Na+ (Sokolov et al., 2007; Struyk and Cannon, 2007; Struyk et al., 2008; Wu et al., 2018). The gating pore current in R897S is approximately three times larger (Fig. 6). This pronounced biophysical defect correlates with the severe clinical phenotype in R897S patients and thereby supports the notion that the gating pore current is the critical functional defect underlying susceptibility to HypoPP. Two unrelated individuals with de novo R897S have been described in the literature, and both had severe HypoPP with hypotonia at birth, early onset of episodic weakness in infancy, daily attacks of weakness, secondary joint deformities, and progressive myopathy with difficulty walking distances independently as teenagers (Chabrier et al., 2008; Hanchard et al., 2013).

The other remarkable feature of $Igp$ for R897S is the pronounced inward rectification (Figs. 6 and 7). Inward rectification is usually a defining feature of the gating pore currents for NaV or KV channels with R/X missense mutations of the outer R1 and R2 arginines (Sokolov et al., 2007; Starace and Bezanilla, 2004; Struyk and Cannon, 2007; Tombola et al., 2005). This rectification, often “strong,” occurs in symmetrical solutions of the permeant ion and has a voltage dependence similar to the gating charge displacement $Q(V),$ consistent with the model that rectification occurs as a result of outward motion of the S4 segment that closes the conduction pathway at depolarized potentials. The gating pore currents observed for CaV1.1 mutant channels are atypical in this regard. The $Igp−V$ relation is approximately linear over a voltage range from −140 mV to 0 mV. This linear behavior for $Igp$ in HypoPP mutant CaV1.1 has been reported for R/X mutations in S4 of DII or DIV and for the atypical V876E in S3 expressed in oocytes (Wu et al., 2018) or mouse muscle fibers (Fuster et al., 2017a; Fuster et al., 2017b; Wu et al., 2012). R900G in DIII also has a linear $Igp$ (Fig. 6 B). We propose that $Igp$ is also inwardly rectifying for CaV1.1 channels with R/X mutations in the R1 and R2 positions, but the opportunity to observe this effect is obscured by the rightward shift of voltage-dependent activation of S4 segments (compared with NaV1.4 channels) in combination with the interference from incompletely blocked ionic currents for mV. For the cut-open oocyte preparation, these interfering currents include incomplete block of the CaV1.1 pore and outwardly rectifying endogenous currents activated for mV. Recent evidence from voltage-clamp fluorometry shows the outward movement of the DIII-S4 is markedly leftward shifted compared with the DII-S4 (V1/2 = −27 ± 1.1 mV compared with 4.6 ± 1.1 mV; Savalli et al., 2020). This leftward shift and R897S being in the R1 position could account for the observed rectification for this HypoPP mutation in CaV1.1, whereas $Igp−V$ is approximately linear over the observable range of −140 mV to 0 mV for other HypoPP mutations because the voltage-dependent closure of the gating pore conductance is more depolarized for R2 (R900G) or for S4 in DII (R528H and R528G).

How can the failure to detect gating pore currents in the cut-open oocyte preparation for two HypoPP CaV1.1 mutations, R897K and R900S, be reconciled with the strong consensus that $Igp$ is the critical defect that predisposes a muscle fiber to paradoxical depolarization with hypokalemia, loss of excitability, and weakness? The first issue to be addressed is the confidence level with which these variants have been classified as pathogenic mutations that cause HypoPP. R897K was identified in three separate HypoPP pedigrees (one Japanese de novo; two French, with one inherited and the other de novo); the variant was not found in unaffected family members, and affected individuals had classical symptoms with recurrent episodes of weakness with low K+ (<3 mM), episodes triggered by carbohydrate ingestion or rest after exercise, and improvement with oral K+ supplementation and acetazolamide (Kubota et al., 2020). The R900S variant has been reported for an individual in the UK (Matthews et al., 2009) and for a family in Hangzhou, China (Ke et al., 2015). The clinical presentation was typical for HypoPP and included recurrent episodes of weakness associated with low K+, attacks often occurring at night or upon awakening, attacks provoked by carbohydrate meals or rest after exercise, and onset in the second decade of life. In the Chinese family, two male cousins and their uncle were symptomatic, whereas three female carriers of R900S did not have any muscle symptoms. There is precedent for reduced penetrance of the phenotype with absence of transient attacks of weakness in females for the most commonly occurring HypoPP mutation, CaV1.1-R528H (Links et al., 1994). In these cases, however, the females usually develop slowly progressive permanent weakness that is detected during clinical examination by middle age. A grandmother in the R900S pedigree from China remains asymptomatic with no fixed weakness. Neither variant is present in the Genome Aggregation Database (gnomAD version 2.1) of 141,456 individuals. Taken together, these data support the interpretation that R897K and R900S are both likely to be pathogenic mutations for HypoPP.

Another possibility is that R897K and R900S both support gating pore currents, but the amplitude was below the detection limit of our measurements (two SD = ± 51 nA from the mean holding current of −77 nA at −100 mV for WT). It is reasonable to expect gating pore currents to be smaller for R900S than for R900G because of the greater steric hindrance from serine than from glycine. Similarly, the charge-conserving R897K would likely be accommodated better than R897S at the critical hydrophobic plug of the CTC in the VSD (Monteleone et al., 2017). On the other hand, simulations of the K+ sensitivity for the resting potential of skeletal muscle show that a gating pore current with a magnitude less than half of that detected for R900G herein is not likely to cause susceptibility to paradoxical depolarization and weakness in low K+ (Kubota et al., 2020). Alternatively, the gating pore current might be more prominent when these mutant channels are expressed in the context of a human muscle fiber instead of a frog oocyte. Finally, we should stay open-minded to the possibility of other mechanisms by which mutant CaV1.1 subunits may cause susceptibility to HypoPP. For example, the level of expression at the membrane is lower (smaller $Qmax$ and $Gmax;$Tables 1 and 2), and the kinetics of channel activation are slower (Fig. 3).

Jeanne M. Nerbonne served as editor.

The authors thank the members of the Olcese laboratory at the University of California, Los Angeles, for assistance with oocytes and M. DiFranco for technical assistance with the cut-open clamp.

This work was supported by the National Institutes of Health/National Institute of Arthritis and Musculoskeletal and Skin Diseases (R01-AR063182).

The authors declare no competing financial interests.

Author contributions: F. Wu and S. Cannon designed the experiments and analyzed the data. F. Wu performed the oocyte recordings, and M. Quinonez assisted with the molecular biology. S. Cannon wrote the paper, and all authors made revisions and approved the final version.

Cannon
,
S.C.
2015
.
Channelopathies of skeletal muscle excitability
.
Compr. Physiol.
5
:
761
790
.
Cannon
,
S.C.
2017
.
Sodium channelopathies of skeletal muscle
.
Handb. Exp. Pharmacol.
246
:
309
330
.
Chabrier
,
S.
,
N.
Monnier
, and
J.
Lunardi
.
2008
.
Early onset of hypokalaemic periodic paralysis caused by a novel mutation of the CACNA1S gene
.
J. Med. Genet.
45
:
686
688
.
Francis
,
D.G.
,
V.
Rybalchenko
,
A.
Struyk
, and
S.C.
Cannon
.
2011
.
Leaky sodium channels from voltage sensor mutations in periodic paralysis, but not paramyotonia
.
Neurology.
76
:
1635
1641
.
Fuster
,
C.
,
J.
Perrot
,
C.
Berthier
,
V.
Jacquemond
, and
B.
Allard
.
2017
a
.
Elevated resting H+ current in the R1239H type 1 hypokalaemic periodic paralysis mutated Ca2+ channel
.
J. Physiol.
595
:
6417
6428
.
Fuster
,
C.
,
J.
Perrot
,
C.
Berthier
,
V.
Jacquemond
,
P.
Charnet
, and
B.
Allard
.
2017
b
.
Na leak with gating pore properties in hypokalemic periodic paralysis V876E mutant muscle Ca channel
.
J. Gen. Physiol.
149
:
1139
1148
.
Hanchard
,
N.A.
,
D.R.
Murdock
,
P.L.
Magoulas
,
M.
Bainbridge
,
D.
Muzny
,
Y.
Wu
,
M.
Wang
,
J.R.
Lupski
,
R.A.
Gibbs
, and
C.W.
Brown
.
2013
.
Exploring the utility of whole-exome sequencing as a diagnostic tool in a child with atypical episodic muscle weakness
.
Clin. Genet.
83
:
457
461
.
Horstick
,
E.J.
,
J.W.
Linsley
,
J.J.
Dowling
,
M.A.
Hauser
,
K.K.
McDonald
,
A.
Ashley-Koch
,
L.
Saint-Amant
,
A.
Satish
,
W.W.
Cui
,
W.
Zhou
, et al
.
2013
.
Stac3 is a component of the excitation-contraction coupling machinery and mutated in Native American myopathy
.
Nat. Commun.
4
:
1952
.
Jurkat-Rott
,
K.
,
M.A.
Weber
,
M.
Fauler
,
X.H.
Guo
,
B.D.
Holzherr
,
A.
Paczulla
,
N.
Nordsborg
,
W.
Joechle
, and
F.
Lehmann-Horn
.
2009
.
K+-dependent paradoxical membrane depolarization and Na+ overload, major and reversible contributors to weakness by ion channel leaks
.
106
:
4036
4041
.
Jurkat-Rott
,
K.
,
J.
Groome
, and
F.
Lehmann-Horn
.
2012
.
Pathophysiological role of omega pore current in channelopathies
.
Front. Pharmacol.
3
:
112
.
Ke
,
Q.
,
F.
He
,
L.
Lu
,
P.
Yu
,
Y.
Jiang
,
C.
Weng
,
H.
Huang
,
X.
Yi
, and
M.
Qi
.
2015
.
The R900S mutation in CACNA1S associated with hypokalemic periodic paralysis
.
Neuromuscul. Disord.
25
:
955
958
.
Kubota
,
T.
,
F.
Wu
,
S.
Vicart
,
M.
Nakaza
,
D.
Sternberg
,
D.
Watanabe
,
M.
Furuta
,
Y.
Kokunai
,
T.
Abe
,
N.
Kokubun
, et al
.
2020
.
Hypokalaemic periodic paralysis with a charge-retaining substitution in the voltage sensor
.
Brain Commun.
2
:fcaa103.
Liman
,
E.R.
,
J.
Tytgat
, and
P.
Hess
.
1992
.
Subunit stoichiometry of a mammalian K+ channel determined by construction of multimeric cDNAs
.
Neuron.
9
:
861
871
.
,
T.P.
,
A.J.
Smit
,
W.M.
Molenaar
,
M.J.
Zwarts
, and
H.J.
Oosterhuis
.
1994
.
Familial hypokalemic periodic paralysis. Clinical, diagnostic and therapeutic aspects
.
J. Neurol. Sci.
122
:
33
43
.
Matthews
,
E.
,
R.
Labrum
,
M.G.
Sweeney
,
R.
Sud
,
A.
Haworth
,
P.F.
Chinnery
,
G.
Meola
,
S.
Schorge
,
D.M.
Kullmann
,
M.B.
Davis
, and
M.G.
Hanna
.
2009
.
Voltage sensor charge loss accounts for most cases of hypokalemic periodic paralysis
.
Neurology.
72
:
1544
1547
.
Mi
,
W.
,
V.
Rybalchenko
, and
S.C.
Cannon
.
2014
.
Disrupted coupling of gating charge displacement to Na+ current activation for DIIS4 mutations in hypokalemic periodic paralysis
.
J. Gen. Physiol.
144
:
137
145
.
Monteleone
,
S.
,
A.
Lieb
,
A.
Pinggera
,
G.
Negro
,
J.E.
Fuchs
,
F.
Hofer
,
J.
Striessnig
,
P.
Tuluc
, and
K.R.
Liedl
.
2017
.
Mechanisms responsible for ω-pore currents in Cav calcium channel voltage-sensing domains
.
Biophys. J.
113
:
1485
1495
.
Moreau
,
A.
,
P.
, and
M.
Chahine
.
2014
.
Molecular biology and biophysical properties of ion channel gating pores
.
Q. Rev. Biophys.
47
:
364
388
.
Nelson
,
B.R.
,
F.
Wu
,
Y.
Liu
,
D.M.
Anderson
,
J.
McAnally
,
W.
Lin
,
S.C.
Cannon
,
R.
Bassel-Duby
, and
E.N.
Olson
.
2013
.
Skeletal muscle-specific T-tubule protein STAC3 mediates voltage-induced Ca2+ release and contractility
.
110
:
11881
11886
.
Polster
,
A.
,
S.
Perni
,
H.
Bichraoui
, and
K.G.
Beam
.
2015
.
Stac adaptor proteins regulate trafficking and function of muscle and neuronal L-type Ca2+ channels
.
112
:
602
606
.
Savalli
,
N.
,
M.
Angelini
,
F.
Steccanella
,
F.
Wu
,
M.
Quinonez
,
S.C.
Cannon
, and
R.
Olcese
.
2020
.
The contribution of the individual voltage sensors to the activation of skeletal CaV1.1 channels [abstract]
.
Biophys. J.
118
(
3
,
Suppl 1
):
105a
.
Sokolov
,
S.
,
T.
Scheuer
, and
W.A.
Catterall
.
2007
.
Gating pore current in an inherited ion channelopathy
.
Nature.
446
:
76
78
.
Sokolov
,
S.
,
T.
Scheuer
, and
W.A.
Catterall
.
2010
.
Ion permeation and block of the gating pore in the voltage sensor of NaV1.4 channels with hypokalemic periodic paralysis mutations
.
J. Gen. Physiol.
136
:
225
236
.
Starace
,
D.M.
, and
F.
Bezanilla
.
2001
.
Histidine scanning mutagenesis of basic residues of the S4 segment of the shaker k+ channel
.
J. Gen. Physiol.
117
:
469
490
.
Starace
,
D.M.
, and
F.
Bezanilla
.
2004
.
A proton pore in a potassium channel voltage sensor reveals a focused electric field
.
Nature.
427
:
548
553
.
Sternberg
,
D.
,
T.
Maisonobe
,
K.
Jurkat-Rott
,
S.
Nicole
,
E.
Launay
,
D.
Chauveau
,
N.
Tabti
,
F.
Lehmann-Horn
,
B.
Hainque
, and
B.
Fontaine
.
2001
.
Hypokalaemic periodic paralysis type 2 caused by mutations at codon 672 in the muscle sodium channel gene SCN4A
.
Brain.
124
:
1091
1099
.
Struyk
,
A.F.
, and
S.C.
Cannon
.
2007
.
A Na+ channel mutation linked to hypokalemic periodic paralysis exposes a proton-selective gating pore
.
J. Gen. Physiol.
130
:
11
20
.
Struyk
,
A.F.
, and
S.C.
Cannon
.
2008
.
Paradoxical depolarization of BA2+- treated muscle exposed to low extracellular K+: insights into resting potential abnormalities in hypokalemic paralysis
.
Muscle Nerve.
37
:
326
337
.
Struyk
,
A.F.
,
V.S.
Markin
,
D.
Francis
, and
S.C.
Cannon
.
2008
.
Gating pore currents in DIIS4 mutations of NaV1.4 associated with periodic paralysis: saturation of ion flux and implications for disease pathogenesis
.
J. Gen. Physiol.
132
:
447
464
.
Thor
,
M.G.
,
V.
Vivekanandam
,
M.
Sampedro-Castañeda
,
S.V.
Tan
,
K.
Suetterlin
,
R.
Sud
,
S.
Durran
,
S.
Schorge
,
D.M.
Kullmann
,
M.G.
Hanna
, et al
.
2019
.
Myotonia in a patient with a mutation in an S4 arginine residue associated with hypokalaemic periodic paralysis and a concomitant synonymous CLCN1 mutation
.
Sci. Rep.
9
:
17560
.
Tombola
,
F.
,
M.M.
Pathak
, and
E.Y.
Isacoff
.
2005
.
Voltage-sensing arginines in a potassium channel permeate and occlude cation-selective pores
.
Neuron.
45
:
379
388
.
Wu
,
F.
,
W.
Mi
,
E.O.
Hernández-Ochoa
,
D.K.
Burns
,
Y.
Fu
,
H.F.
Gray
,
A.F.
Struyk
,
M.F.
Schneider
, and
S.C.
Cannon
.
2012
.
A calcium channel mutant mouse model of hypokalemic periodic paralysis
.
J. Clin. Invest.
122
:
4580
4591
.
Wu
,
F.
,
M.
Quinonez
,
M.
DiFranco
, and
S.C.
Cannon
.
2018
.
Stac3 enhances expression of human CaV1.1 in Xenopus oocytes and reveals gating pore currents in HypoPP mutant channels
.
J. Gen. Physiol.
150
:
475
489
.