Different degrees of loss of function between GEFS+ and SMEI Nav1.1 missense mutants at the same residue induced by rescuable folding defects

Authors


Address correspondence to Yoshihiro Sugiura, Department of Neurology, School of Medicine, Fukushima Medical University, 1 Hikarigaoka, Fukushima 960-1295, Japan. E-mail: y-sugiura@umin.ac.jp

Summary

Generalized epilepsy with febrile seizures plus (GEFS+) and severe myoclonic epilepsy of infancy (SMEI) differ in their clinical severity and prognosis even though mutations of the Nav1.1 sodium channel are responsible for both disorders. We compared the electrophysiologic properties of two mutant Nav1.1 channels characterized by distinct amino acid substitutions at the same residue position: GEFS+ (A1685V) and SMEI (A1685D). Both the mutants showed complete loss of function when expressed alone. However, the function of A1685V can be partly rescued by the β1 subunit, consistently with a folding defect, whereas that of A1685D was not rescued. These electrophysiologic differences are consistent with the divergence in clinical severity between GEFS+ and SMEI.

Generalized epilepsy with febrile seizures plus (GEFS+) is an autosomal dominant epileptic disorder characterized by childhood febrile seizures and various afebrile seizures after the age of 6 years (Scheffer & Berkovic, 1997). Severe myoclonic epilepsy of infancy (SMEI) is a severe epileptic encephalopathy characterized by febrile seizures with an onset before the age of 1 year, which usually develops into various intractable seizures and psychomotor impairments between 1 and 4 years of age (Dravet et al., 2005). These two markedly different clinical disorders are caused by mutations of the same voltage-dependent sodium channel α subunit gene (SCN1A) that codes for Nav1.1. Several articles have reported the electrophysiologic properties of the mutant Nav1.1 in GEFS+ and SMEI. It has not been clarified, however, what determines the divergence in severity between these two diseases, although it has been proposed that folding defects of Nav1.1 mutants may be common and could generate different phenotypes according to the degree of rescue (Rusconi et al., 2009).

Two distinct amino acid substitutions of the same residue cause two different clinical features. An A1685V missense mutation of Nav1.1 caused GEFS+ in a Japanese family (Sugawara et al., 2001), and an A1685D missense mutation was found in a patient with SMEI (Fujiwara et al., 2003). These mutations are situated at the S5 transmembrane region in domain IV of Nav1.1. To address the above question, we studied the properties of these two sodium channel mutants.

Methods

Plasmid construction and transfection

The cDNA of the SCN1A gene, encoding the human sodium channel α1 subunit (Nav1.1), was inserted into the NotI-ApaI site of a pcDNA3.1(+) plasmid vector (Invitrogen, Carlsbad, CA, U.S.A.) (Sugawara et al., 2003). Two mutations of SCN1A, corresponding to A1685V and A1685D, were introduced into the wild-type (WT) construct with a QuickChange mutagenesis kit (Stratagene, La Jolla, CA, U.S.A.). All constructs were verified by sequencing. These plasmid DNAs were isolated for transfection with a Plasmid Maxi kit (Qiagen, Hilden, Germany). Human embryonic kidney (HEK) 293 cells were then transiently transfected with pcDNA3.1-SCN1A (WT), pcDNA3.1-SCN1A (A1685V), or pcDNA3.1-SCN1A (A1685D) using PolyFect Transfection Reagent (Qiagen). Plasmids encoding the human sodium channel β1 subunit and CD8 (pIRES-CD8-hβ1) were also employed, so that we could identify cells expressing Nav1.1 visually with Dynabeads M-450 CD8 (Dynal, Oslo, Norway). In addition, HEK293 cells were transfected with β1 subunit–lacking plasmids to investigate the rescuing effects of the β1 subunit on the sodium currents of channels with the A1685V or A1685D mutations.

Patch-clamp analysis

Thirty-six hours to 40 h after transfection, sodium currents were recorded from HEK293 cells at room temperature (20–22°C) using by the whole cell patch clamp technique with a pipette resistance of 1–2 MΩ. The pipette solution contained (in mm): NaCl 10, CsF 110, CsCl 20, EGTA 2, and HEPES 10. The pH was adjusted to 7.35 using CsOH. The extracellular solution contained (in mm): NaCl 145, KCl 4, CaCl2 1.8, MgCl2 1, and HEPES 10. The pH was adjusted to 7.35 using NaOH. Sodium currents were recorded with a patch-clamp amplifier (AXOPATCH 200B, Axon Instruments, Foster City, CA, U.S.A.) and pCLAMP v.9.2 (Axon Instruments, Foster City, CA, U.S.A.). Our data are presented as means and standard errors with estimated fit parameters. Statistical analyses of variance were performed using analysis of variance (ANOVA) and post hoc comparisons with Fisher’s PLSD test using the ORIGIN Pro 8 (OriginLab, Northampton, MA, U.S.A.) program. The significance level was set at p = 0.05 for all comparisons.

Immunocytochemistry

HEK293 cells with pcDNA3.1-SCN1A (WT), pcDNA3.1-SCN1A (A1685V), or pcDNA3.1-SCN1A (A1685D) were cultured on glass coverslips for 2 days after transfection. Cells were fixed with 4% paraformaldehyde for 20 min and permeabilized in phosphate-buffered saline (PBS). Nonspecific binding was blocked by application of 2.5% normal goat serum in PBS for 30 min at room temperature. Immunolabeling was performed using rabbit polyclonal anti-Scn1a antibodies (1:200, ab24820; Abcam, Cambridge, U.K.) in PBS overnight at 4°C. Cells were rinsed three times with PBS and incubated with FITC conjugated anti-rabbit antibodies (1:100, FI-1000; VECTOR Lab., Burlingame, CA, U.S.A.) for 4 h at 4°C. Immunostained cells on glass coverslips were analyzed using a microscope digital camera system (DP 70; Olympus, Tokyo, Japan) with a fluorescence filter (U-MNIBA; Olympus).

Results

Figure 1A shows three representative sets of sodium currents recorded from HEK293 cells expressing WT, A1685V, or A1685D Nav1.1 channels, all with the β1 subunit present. The maximum current density at −20 mV was 152.6 ± 10.8 pA/pF in the WT Nav1.1 channel with the β1 subunit present (n = 5), 118.6 ± 17.2 pA/pF (n = 5) in the WT Nav1.1 channel in the absence of the β1 subunit, and 31.6 ± 9.8 pA/pF in the A1685V mutant Nav1.1 channel with the β1 subunit present (n = 5). These currents were suppressed by tetrodotoxin. The A1685V mutant Nav1.1 channel lacking the β1 subunit and the A1685D mutant Nav1.1 channels with and without β1 subunits did not conduct any sodium currents. Significant differences in current densities were found between WT and A1685V mutant Nav1.1 channels with the β1 subunit (p < 0.001, Fisher’s PLSD test), and between A1685V and A1685D mutant channels with the β1 subunit (3.6 ± 1.4 pA/pF, n = 5) (p = 0.04469, Fisher’s PLSD test) (Fig. 1B). There were no significant differences between WT channels with the β1 subunit present and those without it.

Figure 1.

 Maximum sodium currents were recorded from HEK293 cells transfected with the wild-type (WT), A1685V, or A1685D mutant Nav1.1 channels (A). Maximum current densities at −20 mV of WT Nav1.1 channel with β1 subunit (WTB), without β1 subunit (WTnB), A1685V Nav1.1 with β1 subunit (A1685V), or A1685D Nav1.1 with β1 subunit (A1685D) (n = 5) (B). Voltage dependency of activation (C) and steady-state inactivation (D) curves of WT and A1685V sodium channels. Activation curves were obtained using a series of 20 msec depolarizing test pulses within a voltage range of −100 mV and +20 mV from a holding potential of −120 mV, applied every 5 s using the whole cell patch-clamp method (C). Steady-state inactivation curves were obtained using a series of 300 msec depolarizing prepulses to potentials ranging from −130 to −20 mV from a holding potential of −120 mV, followed by a test step to −20 mV for 20 ms, applied every 5 s (D). WT and A1685V are represented by blue and red lines, respectively.

Next, we investigated the voltage dependencies of the activation and steady-state inactivation of WT versus A1685V mutant Nav1.1 channels with β1 subunits. The activation curve of the A1685V conductance showed a negative shift of −8 mV compared with the WT (Fig. 1C). The membrane potential at which 50% of the current was activated (V1/2) was −22.1 ± 4.2 mV in the WT (n = 5) and −30.3 ± 3.8 mV in A1685V (n = 5). The steady-state inactivation curve of A1685V (V1/2 = −79.0 ± 2.9 mV, n = 5) showed a negative shift of −15 mV compared with the WT (V1/2 = −63.6 ± 7.2 mV, n = 5) (Fig. 1D). There were no significant differences in the voltage dependencies of the activation or steady-state inactivation between the WT channels with β1 subunits present and those without. The activation V1/2 was −22.3 ± 3.3 mV and the steady-state inactivation V1/2 was −64.7 ± 3.6 mV in the WT channels without β1 subunits (n = 5).

Because no sodium currents were evoked in the A1685D mutant Nav1.1 channel, Nav1.1 immunostaining was carried out to confirm the expression of the Nav1.1 channel in HEK293 cells. As shown in Fig. 2, in any groups, anti-Nav1.1 immunostaining labeled HEK293 cells revealed that the channel proteins had been successfully transfected in all experimental groups.

Figure 2.

 Immunocytochemical staining with antibody against Nav1.1. HEK293 cells transfected with WT, A1685V, and A1685D Nav1.1 cDNA were stained. Nav1.1 did not appear in a negative control culture (NC). The expression of A1685V and A1685D mutant Nav1.1 channels was similar in level to that of the WT channels.

Discussion

We have shown that both A1685V (GEFS+) and A1685D (SMEI) mutant Nav1.1 channels are characterized by complete loss of function when they are expressed alone. In addition, we showed that β1 subunit association partially rescued sodium currents in A1685V mutant channels, although it had no effect on A1685D mutant channels. This functional difference is consistent with the differences in clinical severity between these two mutations, although they occur at the same residue position.

Mice that are heterozygous for mutations causing a complete loss of function of the SCN1A gene show a severe phenotype comprising frequent spontaneous seizures and increased mortality (Yu et al., 2006; Ogiwara et al., 2007). Mice heterozygous for a GEFS+ mutation causing a partial loss of function show a much milder phenotype (Martin et al., 2010). These data suggest that complete loss of function of Nav1.1 causes SMEI, whereas impaired Nav1.1 function causes GEFS+. Our present results are also consistent with this idea. A1685V may cause GEFS+ because of partial restoration of currents via association with the β1 subunit, and A1685D may cause SMEI because of its complete lack of functionality, even when associated with the β1 subunit. Rhodes et al. (2004) have reported that two mutations of the same residue cause GEFS+ (R1648H) and SMEI (R1648C), respectively. Both R1648H and R1648C mutant Nav1.1 showed persistent sodium currents, but the currents were larger in R1648C than in R1648H. However, the gain of function has not been observed by other authors who have expressed R1648H in neuronal cells (Tang et al., 2009). Therefore both of these mutations may induce a loss of function, although R1648C has not yet been tested in neurons.

Why do the mutations at the same amino acid position cause such a different degree of sodium current reductions? Both the mutants show a complete loss of function when expressed alone. However, function of the GEFS+ mutant A1685V was partially restored by the addition of the β1 subunit, similarly to the folding-defective Nav1.1 already described (Rusconi et al., 2007, 2009). Therefore, A1685V is a folding-defective mutant that can be restored by the β1 subunit and possibly by other interacting proteins in vivo, which may induce a stronger rescue than that observed in our in vitro experiments. This is consistent with the milder GEFS+ phenotype. However, if interactions with associated proteins are ineffective, A1685V may exhibit a complete loss of function, consistent with the severe phenotypes that can occur within some GEFS+ families. Notably, A1685D function was not restored by β1, which is consistent with the more severe phenotype of SMEI. It may be that this mutation causes Nav1.1 dysfunction through the disruption of the pore or interference with its gating properties. Both alanine and valine are nonpolar, neutral amino acids, whereas aspartic acid is a polar, negatively charged amino acid. Therefore A1685D in the S5 transmembrane region may produce more marked conformational changes and a more severe degree of dysfunction than those produced by A1685V.

The next question to address is how sodium current reduction results in epilepsy, a hyperexcitability of neuronal circuits. Nav1.1 has been localized to parvalbumin-positive inhibitory interneurons in the hippocampus, whereas pyramidal cells expressed Nav1.1 at negligible levels (Ogiwara et al., 2007). Moreover, action potentials were abnormally reduced in γ-aminobutyric acid (GABA)ergic inhibitory interneurons in SCN1A mutant mice (Yu et al., 2006; Ogiwara et al., 2007). These reports suggest that the dysfunction of Nav1.1 causes impairment of the GABAergic inhibitory system, which likely finally leads to hyperexcitability in the cerebral cortices, thereby causing epilepsy.

Acknowledgment

The authors would like to thank Dr. Edward H. Bertram (Department of Neurology, University of Virginia, U.S.A.) for his insightful comments and Dr. Naomasa Makita (Department of Molecular Physiology-1, Nagasaki University Graduate School of Biomedical Sciences, Japan) for his generous gift of the CD8-hβ1 plasmid.

Disclosure

None of the authors have any conflicts of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Ancillary