Partial expression defect for the SCN5A missense mutation G1406R depends on splice variant background Q1077 and rescue by mexiletine

Bi-Hua Tan, Carmen R. Valdivia, Chunhua Song, Jonathan C. Makielski

Abstract

Mutations in the cardiac Na+ channel gene SCN5A cause loss of function and underlie arrhythmia syndromes. SCN5A in humans has two splice variants, one lacking a glutamine at position 1077 (Q1077del) and one containing Q1077. We investigated the effect of splice variant background on loss of function and rescue for G1406R, a mutation reported to cause Brugada syndrome. Mutant and wild-type (WT) channels in both backgrounds were transfected into HEK-293 cells and incubated for up to 72 h with and without mexiletine. At 8 h, neither current nor cell surface expression was observed for the mutant in either background, but both were present in WT channels. At 24 h, small (<10% compared with WT) currents were noted and accompanied by cell surface expression. At 48 h, current density was ∼40% of WT channels for the mutant in the Q1077del variant background but remained at <10% of WT channels in Q1077. Current levels were stable by 72 h. Coexpression with β1- or β3-subunits or insertion of the polymorphism H558R in the background did not significantly affect current expression. Mexiletine restored current density of the mutant channel in both backgrounds to nearly WT levels. The mutant channels also showed a negative shift in inactivation, slower recovery, and enhanced slow inactivation, consistent with a loss of function phenotype. These data show that a trafficking defect may be partial and time dependent and may differ with the splice variant background. Also, expression defects and gating abnormalities may contribute to loss of function for the same mutation.

  • sodium channels
  • mexiletine rescue

the voltage-dependent cardiac Na+ channel α-subunit pore-forming protein in human heart is encoded by SCN5A. This channel is responsible for a large peak inward Na+ current (INa), which underlies excitability and conduction in working myocardium. Alternative splicing of a single amino acid at the beginning of exon 18 causes insertion of glutamine at position 1077 (Q1077), resulting in two splice variants, one that forms a 2,016-amino acid protein designated Q1077 and a 2,015-amino acid protein that is designated Q1077del (11). Messenger RNA for these two splice variants was present in every human heart examined in a ratio of ∼2:1, with a 65% predominance of the shorter 2,015-amino acid variant Q1077del (11). The longer and less abundant Q1077 background was used in most studies of mutations in SCN5A (6).

Abnormalities in INa associated with loss of function underlie heritable arrhythmia syndromes such as Brugada syndrome, progressive cardiac conduction disease, and congenital sick sinus syndrome (9). Loss of function can occur because of altered channel gating (5) or decreased expression of channels at the cell surface (2, 3). Mechanisms for loss of expression are likely to be complex. The full spectrum of expression defects for ion channels in general and for SCN5A in particular is just beginning to be described. Previously we showed that the common polymorphisms H558R and S524Y show decreased surface expression only when expressed in the Q1077 background (15). We hypothesized that splice variant background might also affect other mutations, causing loss of function. The mutation G1406R was previously reported as a Brugada syndrome mutation (10); we made this mutation in the Q1077del and Q1077 backgrounds and assessed the time dependence of cell surface expression by voltage clamp and by in situ immunocytochemistry and confocal microscopy. These results show that, in HEK-293 cells, expression defects in SCN5A are not all or none, they can be time dependent, and they can be modulated by splice variant background and drugs.

MATERIALS AND METHODS

Site-directed mutagenesis, heterologous expression, and drug treatment.

The G1406R mutation was created by site-direct mutagenesis (mutagenesis kit from Stratagene) using a PCR technique. The appropriate nucleotide changes for G1406R were engineered into two common splice variants of human cardiac voltage-dependent Na+ channel SCN5A/hNav1.5 [one lacks glutamine at position 1077 (Q1077del; GenBank accession no. AY148488) and the other contains Q1077 (Q1077; GenBank accession no. AC1377587)] in the pcDNA3 vector (Invitrogen, Carlsbad, CA). The integrity of the constructs was verified by DNA sequencing. Wild-type (WT) and mutant channels in these two splice variants of SCN5A were transiently transfected in HEK-293 cells as described previously (13, 18). After 3–5 h of incubation, Super Fect (Qiagen, Valencia, CA) transfection reagent-DNA mixture was replaced with 3 ml of normal culture medium with or without 500 μM mexiletine, and the cells were incubated at 37°C for 1 or 2 days. Before the electrophysiological recording, the plates containing the cells were removed from the 37°C incubator, growth medium was aspirated off the plates with or without drug, and the cells were treated with a 0.25% trypsin-1 mM EDTA solution (GIBCO-BRL) and transferred to a fresh tube along with 2 ml of normal culture medium and directly to the experimental chamber for electrophysiological recording.

Standard electrophysiological measurements for functional characterization.

Macroscopic INa was measured using a standard whole cell patch-clamp method at 22–24°C, as described previously (13). The extracellular (bath) solution contained (mmol/l) 140 NaCl, 4 KCl, 1.8 CaCl2, 0.75 MgCl2, and 5 HEPES (with pH adjusted to 7.4 with NaOH). The pipette solution contained (mmol/l) 120 CsF, 20 CsCl, 2 EGTA, and 5 HEPES (with pH adjusted to 7.4 with CsOH). Resistance of pipettes filled with recording solution was 1.0–2.0 MΩ. The data were acquired using pClamp 8.2 (Axon Instruments, Union City, CA) and analyzed using Clampfit (Axon Instruments). The standard voltage-clamp protocols are presented with the data and as described previously (15).

Immunocytochemistry.

The FLAG epitope was introduced between S1 and S2 in domain I of SCN5A for the immunocytochemistry experiments. Transfected and nontransfected HEK-293 cells were fixed with 4% paraformaldehyde for 15 min, permeated with 0.1% Triton X-100 for 10 min, and quenched with 0.75% glycine for 10 min at room temperature. Then the cells were blocked with blocking buffer (10% goat serum and 5 mM NaN3-PBS) and incubated at 37°C for 30 min. After the blocking procedure, the cells were incubated for 1 h at 37°C with a 1:200 dilution of mouse anti-FLAG M2 primary antibody (Stratagene) against the FLAG-tagged α-subunit of the Na+ channel and rabbit anti-calnexin NH2-terminal polyclonal antibody (StressGen Bioreagents, Victoria, BC, Canada) for endoplasmic reticulum (ER) labeling at a ratio of 1:250. The cells were washed with blocking buffer before 1:250 fluorescein-conjugated goat anti-mouse and 1:300 fluorescein-conjugated goat anti-rabbit antibodies (Molecular Probes, Eugene, OR) were applied as the secondary antibody and allowed to react for 1 h at 37°C.

Confocal laser microscopy.

Fluorescent probe-labeled HEK-293 cells were viewed using a confocal imaging system (model MRC-1024, Bio-Rad) equipped with a krypton-argon laser beam and mounted on an inverted microscope (Diaphot 200, Nikon). A ×60 oil immersion lens with a 1.4 NA was used. The microscope was controlled by the Lasersharp 2000 program (Bio-Rad). A Kalman collection filter with four frames per image was applied to record the image. Z series were created by sequential scanning of green and red channels at 0.2-mm steps. Single sections were exported to Adobe Photoshop 6.0 for final image processing.

Statistical analysis.

Values are means ± SE. Determinations of statistical significance were performed using Student's t-test for comparisons of two means or analysis of variance for comparisons of multiple means. P < 0.05 was considered statistically significant. Curve fits were performed using pClamp 8.2 (Axon Instruments). Nonlinear curve fitting was performed with Origin 6.0 (Microcal Software).

RESULTS

Time-dependent expression and mexiletine rescue for G1406R in the two backgrounds.

WT and G1406R mutant channels in the two common splice variant backgrounds Q1077del and Q1077 were voltage clamped after 24 and 48 h of incubation with and without 500 μM mexiletine. Representative current traces for WT and G1406R channels in the Q1077del background are shown in Fig. 1A, with summary data in Fig. 1B; representative current traces for WT and G1406R channels in the Q1077 background are shown in Fig. 2A, with summary data in Fig. 2B. The current levels for WT-Q1077del (Fig. 1B) and WT-Q1077 (Fig. 2B) tended to increase modestly with time and with exposure to mexiletine, but this did not reach statistical significance. At 24 h, the expression defect for the G1406R mutant is profound in both backgrounds, with 3.4% of WT levels for Q1077del (Fig. 1B) and 9.4% of WT levels for Q1077 (Fig. 2B). At 48 h, the current levels for G1406R in Q1077del had increased significantly to −133 ± 27 pA/pF compared with 24-h levels of −11 ± 4 pA/pF (Figs. 1B and 3), whereas in the Q1077 background even with >48 h of incubation, the current levels of G1406R remained severely depressed at only −34 ± 8 pA/pF, which was significantly lower than in the Q1077del background (Figs. 2B and 3). Incubation in mexiletine increased currents dramatically and significantly for the mutant channel in both backgrounds, with levels reaching nearly WT levels. Currents were measured after washout of mexiletine, which was present only during the incubation period. Expression of the G1406R mutants in either background did not increase at 72 h (Fig. 3) and remained at ∼33% in Q1077del and 10% in Q1077.

Fig. 1.

Partial expression defect and mexiletine rescue for G1406R channel in the Q1077del background. A: whole cell current traces from representative G1406R and wild-type (WT) channels without (left) and with (right) mexiletine treatment. B: summary of Na+ current (INa) density in G1406R and WT after 24 and 48 h of incubation with and without mexiletine. Values are means ± SE of number of experiments shown in parentheses above bars. *Significant difference between G1406R after 24 h of incubation without mexiletine and with mexiletine or WT after 24 h incubation (P < 0.0001). †Significant difference between G1406R after 48 h of incubation without mexiletine and with mexiletine or WT after 24 h of incubation (P < 0.03). §Significant difference between 24 and 48 h of incubation without mexiletine (P < 0.0001).

Fig. 2.

Partial expression defect and mexiletine rescue for G1406R channel in the Q1077 background. A: whole cell current traces as in Fig. 1, except in Q1077 background. B: summary of INa density in G1406R and WT after 24 and 48 h of incubation with and without mexiletine. Values are means ± SE of number of experiments shown in parentheses above bars. *Significant difference between G1406R after 24 h of incubation without mexiletine and with mexiletine or WT after 24 h of incubation (P < 0.0001). †Significant difference between G1406R after 48 h of incubation without mexiletine and with mexiletine or WT after 24 h of incubation (P < 0.001).

Fig. 3.

Time-dependent expression of INa density of G1406R channel normalized to WT levels in Q1077del and Q1077 backgrounds showing no increase in current densities after 48 h. Data are normalized to WT densities at the same time point. Values are means ± SE of number of experiments shown in parentheses above bars. *Significant increase after 48 h of incubation (P < 0.0001). †Significant difference between Q1077 and Q1077del background after 48 h, even at 72 h, of incubation (P < 0.05).

Voltage-dependent gating properties of “rescued” G1406R in the two splice variants of SCN5A.

The voltage-dependent gating properties of G1406R channels with and without mexiletine in the Q1077del splice variant background are compared with those of WT channels in Fig. 4; summary data for the fitted kinetic parameters are shown in Table 1. The comparable data for G1406R in the Q1077 splice variant background are shown in Fig. 5 and Table 2. Activation of the channel was the same for both WT backgrounds and for the mutation in both backgrounds and was unaffected by previous incubation in mexiletine (Figs. 4A and 5A). For steady-state inactivation, however, the midpoints were significantly shifted by −7 mV for G1406R compared with WT in the Q1077del variant and by −10 mV for G1406R compared with WT in the Q1077 background (Figs. 4B and 5B). Slower time constants of recovery were observed for G1406R than for WT in both backgrounds (Figs. 4C and 5C). Finally, a protocol was designed to measure the component of current inactivated in a prepulse that cannot recover quickly. This inactivation, which is sometimes called intermediate, slow, and ultraslow inactivation, depending on the length of the prepulse, will be referred to collectively as slow inactivation. Slow inactivation was significantly greater in mutant than in WT channels, with statistical significance demonstrated for prepulse durations as long as 6 s (Figs. 4D and 5D). Late INa, measured 720 ms after the initial depolarization, was 0.4–1.2% of peak current and was not significantly different from that of WT channels (data not shown). In all cases, previous incubation with mexiletine did not increase the relative late current, nor did it alter voltage-dependent gating.

Fig. 4.

Voltage-dependent gating for G1406R and WT channels with and without prior incubation in mexiletine (MEX) in Q1077del variant background. A: voltage dependence of activation for G1406R and WT with and without mexiletine. Line represents a fit to the Boltzmann function: Na+ conductance (GNa) = [1 + exp(V1/2V)/k]−1, where V1/2 and k are midpoint and slope factor (as an index of voltage control, all factors were >4), respectively, and GNa = INa(norm)/(VVrev), where Vrev is reversal potential and V is membrane potential. B: steady-state availability from inactivation for G1406R and WT with and without mexiletine. Line represents a fit to the Boltzmann function: INa = INa-max[1 + exp(VcV1/2)/k]−1, where V1/2 and k represent midpoint and slope factor, respectively, and Vc is membrane potential. C: recovery from inactivation for G1406R and WT with time on a logarithmic scale to better show early time course of recovery. Recovery time course was best fit with 2 exponentials: normalized INa = [Af exp(−tf)] + [As exp(−ts)], where t is recovery time interval, τf and τs represent fast and slow time constants, respectively, and Af and As represent fractional amplitude of fast and slow recovery components, respectively. D: slow inactivation for G1406R and WT with prepulse duration on a logarithmic time scale. Insets: diagrams of voltage protocols. Values are means ± SE; n and fit parameters are given in Table 1.

Fig. 5.

Voltage-dependent gating for G1406R and WT channels with and without mexiletine as in Fig. 4, except in Q1077 variant background. A: voltage dependence of activation for G1406R and WT with and without mexiletine. B: steady-state availability from inactivation for G1406R and WT with and without mexiletine. C: recovery from inactivation for G1406R and WT with time on a logarithmic scale to better show the early time course of recovery. D: slow inactivation for G1406R and WT with prepulse duration on a logarithmic time scale. Lines are fits to equations as described in Fig. 4 legend. Insets: diagrams of voltage protocols. Values are means ± SE; n and fit parameters are given in Table 2.

View this table:
Table 1.

Voltage-dependent gating parameters of G1406R and WT channels with and without mexiletine in Q1077del background

View this table:
Table 2.

Voltage-dependent gating parameters of G1406R and WT with and without mexiletine in Q1077 background

Na+ channel protein expression with and without drug treatment.

WT and mutant Na+ channels were labeled using the FLAG epitope, expressed in the HEK-293 cells, and examined by voltage clamp and confocal microscopy (Figs. 6 and 7). As shown by the representative current traces in Figs. 6A and 7A, the currents were unaffected by the presence of the FLAG epitope. The cells were labeled with anti-FLAG antibody to mark the location of channels and with anti-calnexin antibody (Figs. 6B and 7B). Calnexin, a chaperon protein present in the ER, was used as an ER marker (8). Expression of the protein at the cell surface is demonstrated by a rim of fluorescence, as seen in the WT controls (Figs. 6B and 7B). These results show little expression at the cell surface for the mutation in either background at 8 h, whereas expression of the WT at the surface is robust at 8 h (Figs. 6B and 7B). Cell surface expression of the mutations for both backgrounds is also noted after 48 h and in the presence of mexiletine, consistent with the current expression data. These results show that the current expression defect is caused by a decrease in channel expression at the cell surface and that the defect appears to be a slowed and incomplete channel expression at the cell surface. Other attempts to restore the current density in G1406R, such as coexpression with the β1- and β3-subunits and inclusion of the common polymorphism H558R, which rescued the expression of M1766L (18), failed to restore the current densities for G1406R in the two splice variants (data not shown).

Fig. 6.

Confocal imaging of immunolabeled channels in Q1077del background. WT (right) and G1406R mutant (left) channels after transient transfection (from left to right) by 8 h, 48 h, and 48 h + mexiletine incubation and WT channels (right) after transient transfection (from left to right) by 8 and 48 h of incubation at 37°C. A: whole cell current traces from representative G1406R mutant and WT channels. Voltage clamp was a 24-ms step depolarization from a holding potential of −140 mV to different potentials in 10-mV increments, as shown Fig. 4A, inset. B: confocal microscopic images of HEK-293 cell expressing G1406R mutant and WT channels. Top: cell labeled with anti-FLAG (green). Middle: the same cell labeled with anti-calnexin (red). Bottom: superimposition of both images to show colocalization.

Fig. 7.

Confocal imaging of immunolabeled channels in Q1077 background (layout as described in Fig. 6 legend). G1406R mutant channels (left) after transient transfection (from left to right) by 8 and 48 h of incubation and 48 h of incubation + mexiletine and WT channels (right) after transient transfection (from left to right) by 8 and 48 h of incubation at 37°C in Q1077 splice variant.

DISCUSSION

This study shows that the SCN5A mutation G1406R causes a partial trafficking defect with slowed and incomplete expression relative to WT channels and that the severity of the defect depends on the background splice variant in which it is expressed and is worse in the Q1077 variant. The antiarrhythmic drug mexiletine can improve expression of the mutant channel and restore the current density of G1406R after 48 h of incubation to WT levels. Despite this rescue, the mutant channel still has kinetic abnormalities in inactivation, slow inactivation, and recovery from inactivation, which would tend to cause a loss of function.

Previously, most of the SCN5A arrhythmia mutations causing loss of function were studied in only one of the splice variant backgrounds. The first clone of SCN5A, denoted hH1 (6), is a 2,016-amino acid protein that contains Q1077 and a rare variant R1027Q, which is not found in >1,600 control alleles (1). The second clone, denoted hH1a (7), is a 2,015-amino acid protein that lacks Q1077 and also contains the rare variant T559A. In this study and previously, we used a background WT clone designated hH1c (11), which does not contain the rare variants found in hH1 and hH1a. The effects of the rare variants R1027Q and T559A are unknown. We previously characterized eight common [>0.5% allelic frequency in ≥1 control population (1)] polymorphisms in the long Q1077 and short Q1077del variants (15). Two of these common polymorphisms (H558R and S524Y) caused a profound expression defect in the Q1077, but not the Q1077del, background (11, 15). We hypothesized that the splice variant background might also be important for arrhythmia mutations with loss of function. This study shows that the loss of function for G1406R is more severe in the Q1077 background. Although it is not always stated, most previous studies appear to have used the hH1 background WT clone containing Q1077. The possible effects of the rare variants on the function of WT and in the presence of other mutations found in hH1 and hH1a are unknown. We have studied all three WT clones, hH1, hH1a, and hH1c (11, 13), and, except for a tendency for reduced current densities in the Q1077-containing clone hH1 (11), we have not seen any obvious functional differences in the WT currents. This does not exclude, however, the possibility of new interactions when additional mutations are introduced into the backgrounds. Moreover, the results of the present study again show that the background WT clone can be important in determining the functional properties of a mutation in heterologous expression systems, and this could be one explanation for the clinical observations of the same single SCN5A missense mutation leading to different phenotypes, depending on the different genetic background.

Mutations in the SCN5A coding sequence can cause loss of function by several general mechanisms: 1) A stop codon is introduced, causing incomplete transcription and no channel proteins. 2) Channel protein is produced but traffics poorly to the surface. 3) Channel protein is produced and reaches the surface but malfunctions because of altered kinetic or pore properties. We observed a partial and time-dependent trafficking defect in G1406R, where the defect was profound at 8 h, with no measurable channels at the surface by imaging and no currents (Figs. 6 and 7), but with definite channels at the surface at 24 and 48 h but reduced currents (Figs. 6 and 7). These results emphasize that trafficking defects can represent a spectrum in severity and timing of expression.

Our results also show that more than one mechanism may contribute to loss of function. The mutant channels that reach to the surface on their own, or with the help of mexiletine, demonstrate a leftward shift in inactivation (Figs. 4B and 5B) and enhanced slow inactivation (Figs. 4D and 5D), which is a mechanism for the loss of function seen in other arrhythmia mutations (14). The rescue by mexiletine is the most complete yet demonstrated, in that it is more robust than that demonstrated previously for M1766L (16) and G1743R (17). The therapeutic potential of mexiletine to rescue trafficking defective mutations, however, is in doubt for several reasons. Despite this robust rescue, the doses required for rescue in these heterologous systems are far above therapeutic levels. Also, in this case, as well as in previous examples of drug rescue (16, 17), the rescued channel retains arrhythmogenic kinetic abnormalities. Although, at first thought, the Na+ channel-blocking action of mexiletine would also appear to limit usefulness, in that it causes a loss of function, this is may not be the case in vivo. As a Vaughan Williams class 1b drug, it has very rapid off kinetics for the Na+ channel and does not block peak or early current in therapeutic concentrations; therefore, it is unlikely to cause proarrhythmia in theory; also, in clinical practice, mexiletine does not have the proarrhythmia potential of class Ia and Ic drugs, which have slower kinetics (4). Mexiletine rescue, however, should be considered a “proof of principle,” and perhaps other rescue drugs might be found to rescue in concentrations nontoxic to patients.

Other factors that may affect current densities when expressing SCN5A in heterologous systems include β-subunits and the polymorphism H558R (18). Many, but not all, previous studies of SCN5A mutations routinely coexpress the β1-subunit with the channel, and the effect of the presence or absence of the β1-subunit has not been routinely assessed. Moreover, additional β-subunits have been described in the heart, including a β3-subunit (12). When we coexpressed the β1- and β3-subunits with G1406R, we saw no significant increase in current density beyond WT. Also, when we included the common polymorphism H558R in the background, we saw no significant change in current density for G1406R compared with WT. These factors are likely to be specific for each mutation, because H558R did increase expression levels for the trafficking defective mutation M1766L (18). For this and other reasons, the limitations of the heterologous expression system for extrapolating these results to arrhythmia causation and amelioration must be acknowledged. These cells are not heart muscle cells, and they do not have the full panoply of subunits and regulator machinery found in the native heart cells. β-Subunits and other interacting proteins may have important and unknown effects on the trafficking and kinetics of the WT and mutated channels. This limitation is of course shared by all previous studies of arrhythymia mutations using heterologous expression systems.

In conclusion, this study characterized the electrophysiological function of the SCN5A missense mutation G1406R and showed that the molecular phenotype depends on the splice variant background. These findings emphasize again the importance of studying arrhythmia mutations to characterize the electrophysiological phenotype using at least two common background channel sequences and may have implications for expression defect mechanisms and potential treatment of SCN5A mutations causing loss of function.

GRANTS

This work was supported by an American Heart Association, Greater Midwest Affiliate, Postdoctoral Fellowship to B.-H. Tan and National Heart, Lung, and Blood Institute Grant HL-71092 to J. C. Makielski.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract