|Home | About | Journals | Submit | Contact Us | Français|
Type 2 long QT syndrome (LQT2) involves mutations in the human ether a-go-go-related gene (hERG or KCNH2). T421M, a S1 domain mutation in the Kv11.1 channel protein, was identified in a resuscitated patient. We assessed its biophysical, protein trafficking and pharmacological mechanisms in adult rat ventricular myocytes (ARVMs).
Isolated ARVMs were infected with WT- and T421M-Kv11.1 expressing adenovirus and analyzed using patch clamp, Western blot and confocal imaging techniques. Expression of WT- or T421M-Kv11.1 produced peak tail current (IKv11.1) of 8.78±1.18 and 1.91±0.22 pA/pF, respectively. Loss of mutant IKv11.1 resulted from, 1) a partially trafficking-deficient channel protein with reduced cell surface expression, and 2) altered channel gating with a positive shift in the voltage-dependence of activation and altered kinetics of activation and deactivation. Co-expression of WT+T421M-Kv11.1 resulted in heterotetrameric channels that remained partially trafficking-deficient with only a minimal increase in peak IKv11.1 density, whereas the voltage-dependence of channel gating became “WT-like”. In the ARVM model, both WT- and T421M-Kv11.1 channels responded to β-adrenergic stimulation by increasing IKv11.1.
The T421M-Kv11.1 mutation caused a loss of IKv11.1 through interactions of abnormal protein trafficking and channel gating. Furthermore, for co-expressed WT+T421M-Kv11.1 channels, different dominant negative interactions govern protein trafficking and ion channel gating, and these are likely to be reflected in the clinical phenotype. Our results also show that WT and mutant Kv11.1 channels responded to β-adrenergic stimulation.
The human ether-a-go-go related gene (hERG or KCNH2) encodes the α-subunits that co-assemble as tetramers to form Kv11.1 K+ channels.1,2 Kv11.1 channels mediate the rapidly activating delayed rectifier K+ current (IKr) that is critical for normal repolarization of human heart. Mutations in KCNH2 are associated with chromosome 7-linked congenital long QT syndrome (LQT2, 3). To date, >300 KCNH2 mutations have been identified with most being missense mutations that decrease IKr either by disrupting intracellular trafficking of Kv11.1 channel protein to the cell surface or by altering Kv11.1 channel gating properties.4–6
T421M-Kv11.1 is a missense mutation located in the S1 transmembrane domain of the Kv11.1 protein and was identified in a patient with LQT2 resuscitated from ventricular fibrillation. Very few LQT2 mutations localize to the S1 domain. Therefore, we characterized the T421M-Kv11.1 mutation to test the hypothesis that a S1 domain mutation should alter both protein trafficking and gating properties of Kv11.1 channels. We overexpressed channels in a native adult rat ventricular myocyte (ARVM) model, and we studied WT- or T421M-Kv11.1 homotetrameric channels as well as co-expressing genes for WT+T421M-Kv11.1 subunits to create heterotetrameric channels that more closely simulate human heart. Because the patient experienced cardiac arrest following an acute auditory trigger, we also studied whether there was a link between β-adrenergic stimulation and Kv11.1 channel activity. These studies demonstrate the feasibility of using ARVMs as a native myocyte model for Kv11.1 expression and regulation studies, and the complex mechanisms we found are likely to contribute to determining the cellular phenotype and disease severity in some LQT2 patients.
T421M-Kv11.1 mutation was identified from a 33 yo female who presented with a near sudden death episode during the night after startle by a ringing telephone.7 There were no other precipitating causes (e.g., drugs, other diseases, etc). The patient suffered chronic anoxic brain injury. A post-event ECG showed sinus rhythm at 71 bpm with a QTc of 520 msec with a prominent U wave. Her father and a paternal uncle died suddenly at age 44 yrs and in his 30s, respectively, with no ECGs or autopsies. Two of her three children are mutation positive but with normal QTc intervals.
The appropriate nucleotide change in wild type (WT) KCNH2 cDNA (1348 C>T), resulting in the threonine (T) to methionine (M) mutation at amino acid 421 (T421M-Kv11.1) was done by site directed mutagenesis and verified by DNA sequencing. The WT- and T421M-KCNH2 cDNA were subcloned into pcDNA3.0 vector (Invitrogen, Carlsbad, CA).5 Adenoviral vector construction is described in the online supplement.
In one series of experiments, HEK293 cells were transiently transfected with WT- or T421M-KCNH2 pcDNA 3.0 constructs and studied 48 hrs after transfection.
Ventricular myocytes were isolated from 12 week old female Sprague Dawley rats by a modified method of Powell and Twist.8 See online supplement for expanded methods. ARVMs were cultured for 4 days. They remained rod shaped and retained a striated appearance at the time of experiments. The ARVMs were infected with 1 × 106 adenoviral particles (AdWT-Kv11.1 or AdT421M-Kv11.1 or Ad-GFP), and for co-expression studies 0.5 × 106 of AdWT- and 0.5 × 106 of AdT421M-Kv11.1 viral particles, on the day of isolation for 24 hrs, and experiments were conducted after 4 days of infection once the expression of GFP was confirmed.
Patch clamp electrodes had resistances of 1.5 to 1.8 MΩ. Whole cell currents were recorded at room temperature with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA), Series resistance compensation was 70–80% and whole cell capacitance compensation was used. The pipette solution contained (in mM) 130 KCl, I MgCl2, 5 EGTA, 5 MgATP, 10 HEPES (pH 7.2 with KOH). External solution contained (in mM) 137 NaCl, 4 KCl, 1.8 CaCl2, 1 MgCl2, 10 glucose, and 10 HEPES (pH 7.4 with NaOH). In some experiments the ARVMs were superfused with 1 μM E-4031 or 10 μM Isoproterenol (Sigma-Aldrich Corporation, St. Louis, MO). Voltage protocols and data analysis were done with pCLAMP 10 (Molecular Devices, Sunnyvale, CA) and Origin v7.5 or v8.6 software (Origin, Northampton, MA).
Adenoviral infected ARVMs were lysed with NP-40 lysis buffer (1% NP-40, 150mM NaCl, 10% glycerol, 5mM EDTA, and 50mM Tris-HCl, pH 7.4). The whole cell lysates were centrifuged at 15000 g for 10 min and the supernatant used for Western blot analysis as previously described.9 Approximately equal numbers of ARVMs were used for each sample to control the protein loading.
ARVMs plated on coverslips were subjected to immunocytochemistry and confocal imaging as described in the online supplement.
Data are expressed as mean±SEM. Normality was tested by using Shapiro-Wilk test. Statistical significance of normally distributed data was then determined by ANOVA for multiple group comparisons with Turkey post-hoc test. Comparisons between two groups performed by Students’ t-test (two-tailed). Statistical analysis was performed using Microcal Origin, v7.5 or v8.6, and p<0.05 was considered significant.
We confirmed robust expression of channel protein in ARVMs infected with AdWT-Kv11.1 virus by confocal imaging as shown in Figure 1A. WT-Kv11.1 channel protein was detected by anti-Kv11.1 antibody staining along with GFP expression indicated by the red and green imaging patterns, respectively. In uninfected ARVMs, no staining was observed, indicating a lack of detectable endogenous Kv11.1 protein. In freshly isolated canine myocytes, native IKr channels localize to t-tubules.10 Previous studies in cultured ARVMs have shown the time-dependent loss of t-tubules along with membrane depolarization11,12, therefore we did not study t-tubular localization of overexpressed Kv11.1 channels or record action potentials. Figure 1B shows patch clamp records from control and AdWT-Kv11.1 infected ARVMs. The ARVMs were depolarized from a holding potential of −80 mV to −70 to 60 mV in 10 mV increments for 3 sec followed by a step to −50 mV, with each pulse repeated every 10 sec. The protocol evoked large amplitude inward current transients (data not shown). The control ARVM record shows a voltage-dependent rapidly activating outward current with depolarization that gradually decayed, and is consistent with transient outward K+ current (Ito).13 Following repolarization, no tail current was detected consistent with the cultured ARVMs lacking the delayed rectifier K+ channels, IKr and IKs. Although a small amplitude IKr was reported in freshly isolated ARVMs14,15, the cultured ARVMs have no detectable endogenous IKr.
In ARVMs expressing WT-Kv11.1, depolarization activated additional time- and voltage-dependent outward current, and with repolarization a large amplitude tail current was present (lower panel, Fig. 1B). Tail current amplitude was stable over the course of experiments. Figure 1C shows currents from a different WT-Kv11.1 expressing ARVM recorded using the same voltage protocol. We first recorded the total current followed by superfusion with E4031 (1 μM), a specific blocker of Kv11.1 channels and IKr to obtain the E4031-insensitive current. E4031 blocked a component of outward current during depolarization and abolished the tail current. The E4031-insensitive currents were subtracted from the total currents to obtain the E4031-sensitive current. The E4031-insensitive currents are similar to those found in control ARVMs (Fig. 1B) and the E4031-sensitive current is similar to IKv11.1 found in other expression models1–4. Figure 1D shows I-V plots of membrane currents (total, E4031-insenitive and E4031-sensitive) in ARVMs measured at the end of the depolarizing voltage clamp steps. The E4031-sensitive current began to activate at −40 mV, peaked at 0 mV and undergoes inward rectification at more depolarized voltages as channels inactivate. The peak tail current amplitudes are plotted in Figure 1E as a function of prepulse voltage. The total and E4031-sensitive current plots are indistinguishable and represent IKv11.1. When fit with a Boltzmann function, the voltage at half-maximal activation (V1/2) of IKv11.1 and slope factor (k) were −13.5±3.0 mV and 6.9±0.7, respectively, for the E4031-sensitive current and these data are similar to previously published data from multiple expression systems from many laboratories.
T421M is located in the S1 transmembrane spanning domain of Kv11.1 protein (Fig. 2A). Uninfected or WT- or T421M- Kv11.1 infected ARVMs were lysed and subjected to Western blot analyses (Fig 2B). WT- and T421M-Kv11.1 expressing ARVMs showed bands consistent with core glycosylated (135 kDa, immature) and complexly glycosylated (155 kDa, mature) Kv11.1 protein, however the 155 kDa band was less dense in the T421M expressing ARVMs compared to the 135 kDa protein band (n=2). Uninfected ARVMs did not show protein bands specific for Kv11.1. We confirmed expression of T421M-Kv11.1 channel protein in ARVMs by confocal imaging (Fig. 2C). After 4 days of infection, expression of WT- and T421M-Kv11.1 channel protein is indicated by the red staining patterns, however, the T421M-Kv11.1 staining was less intense at the cell edge where functional surface membrane channels are located, whereas GFP images were similar in intensity. We then performed Kv11.1 and GFP fluorescence intensity line scans across ARVMs expressing WT- or T421M-Kv11.1, and representative data are shown in Figure 2D (left). For WT Kv11.1, the fluorescence intensity was greatest at the cell edge compared to T421M Kv11.1 where the fluorescence intensity had a nearly uniform distribution. GFP, a cytosolic protein, had a similar nearly uniform distribution in all ARVMs studied. We also compared the averaged peak fluorescence intensity at the cell edges to that at the middle of the cell (Fig. 2D, right). For WT- and T421M-Kv11.1 the fluorescence intensity ratios (mean cell edge/cell middle) were 6.2 ± 1.8 and 1.8 ± 0.5 (p<0.05), respectively. Taken together, the Western blot and confocal imaging data suggest that T421M-Kv11.1 is partially trafficking-deficient compared with WT-Kv11.1 in ARVMs. Figure 2E shows patch clamp recordings from AdWT- or AdT421M-Kv11.1 infected ARVMs. The WT-Kv11.1 expressing ARVMs showed depolarization activated outward current followed by a large amplitude slowly decaying tail current. T421M-Kv11.1 expressing ARVMs showed a smaller depolarization activated outward current. More importantly, with repolarization to −50 mV there was smaller amplitude tail current that rapidly activated and decayed (see inset). The I-V relations for peak tail current amplitudes for WT- and T421M-Kv11.1 are plotted in Figure 2F as a function of prepulse voltage. In contrast to WT-Kv11.1, T421M-Kv11.1 current activation was markedly shifted positively and current amplitudes were small. The maximal peak tail current density for T421M-Kv11.1 was reduced by 78.3% (WT-Kv11.1 = 8.78±1.18 pA/pF, T421M-Kv11.1 = 1.91±0.22 pA/pF, p<0.05, Fig. 2G). For T421M-Kv11.1 the V1/2 (28.3±3.3 mV) was shifted positively by nearly 42 mV (WT −13.5±3.0 mV, p<0.05) without altering the slope factor (k=6.4±0.7 mV/e-fold change) compared to WT-Kv11.1 (6.9±0.4 mV/e-fold change, p>0.05). These data suggest that homotetrameric T421M-Kv11.1 channels should contribute minimal repolarizing current at physiological voltages due to abnormalities in both protein trafficking and channel gating.
The rate of channel activation was studied using an envelope of tails protocol where from a holding potential of −80 mV, a prepulse was applied to voltages from −40 to 60 mV in 20 mV increments for different durations ranging from 0.1 to 2 sec followed by repolarization to −50 mV to generate tail current as channels rapidly reopen from the inactivated state before slowly closing. Each pulse sequence was applied at 10 sec intervals. The protocol and representative envelope of tail current traces are shown in Figure 3A for WT-Kv11.1 (prepulse to 20 mV) and T421M-Kv11.1 (prepulse to 40 mV). Figure 3B shows the mean peak envelope of tail currents plotted for WT- and T421M-Kv11.1 at the different prepulse voltages. The peak tail currents were fitted as a single exponential process at each prepulse voltage and the time constants for WT- and T421M-Kv11.1 channel activation are given in Table 1. At comparable prepulse voltages, the time constants for activation of T421M-Kv11.1 are larger compared to the WT-Kv11.1 channels showing a reduced rate of activation.
The rate of deactivation was studied for WT- or T421M-Kv11.1 channels expressed in ARVMs by depolarizing cells to 50 mV for 2 sec to activate and inactivate the channels, followed by repolarizing steps to test voltages between −100 and −20 mV in 10 mV increments for 3 sec. The protocol and representative current traces for the repolarizing steps are shown in Figure 4A for WT- and T421M-Kv11.1. Each current decay trace was fitted as a double exponential process.4 The mean fast (tfast) and slow (tslow) time constants for deactivation were plotted against the test voltages (Fig. 4B). The upper plots show tslow and the lower plots show tfast for WT- and T421M-Kv11.1 with the T421M-Kv11.1 plots expanded (right panels). At all the test voltages, the apparent mean rates of deactivation for T421M-Kv11.1 channels were more rapid than for WT-Kv11.1 channels. In fact, we cannot exclude the possibility that peak tail current for T421M-Kv11.1 channels is reduced by its rapid deactivation which could overlap recovery from inactivation.
The rate of inactivation was studied using a protocol where cells were first depolarized to 50 mV for 1.5 sec to inactivate the channels followed by a hyperpolarizing step to −100 mV for 10 msec to recover channels to the open state, followed by depolarizing test steps to −40 to 40 mV for 1.5 sec in 10 mV increments to inactivate the channels.4 Each pulse protocol was repeated every 10 sec. Because this protocol activates large amplitude inward Na+ and Ca2+ current transients in ARVMs that interfere with the measurement of the rate of inactivation of Kv11.1 channels (data not shown) we used HEK293 cells transiently transfected with WT- or T421M-Kv11.1 cDNA. In Figure 5A representative families of current traces for the depolarizing test steps show for WT and mutant channels large amplitude currents that rapidly inactivate. The time course of each current was fit as a single exponential process and mean time constants were plotted against the test voltages. Our results show that at most voltages, the inactivation time constants were similar between WT and mutant channels.
Recovery from inactivation was measured in ARVMs utilizing the same protocol used to measure deactivation. The ARVMs were depolarized to 60 mV for 2 sec to activate and inactivate the channels. Following repolarization to different test voltages, recovery from inactivation was measured as the rising phase (“hook”) of the tail current.4 This was fit as a single exponential process and the mean data are plotted as a function of test voltages in Figure 5B. At most voltages, recovery from inactivation was slightly faster for WT-Kv11.1 channels but the differences were small. These data suggest that the T421M-Kv11.1 mutation compared with WT-Kv11.1 exerts its predominant effects on closed-open state gating transitions with only mild effects on open-inactivated state transitions.
LQT2 follows an autosomal dominant inheritance pattern, therefore the patient harbors a normal KCNH2 allele and a mutant KCNH2 allele that encodes for T421M-Kv11.1. Providing the α-subunits co-assemble, the patient would be expected to express different allelic populations of Kv11.1 channels including small numbers of homotetrameric WT- and homotetrameric T421M-Kv11.1 channels, with most being heterotetramers of the two allelic products at various combinations (for review of heterotetrameric co-assembly, see 16). In an attempt to recapitulate this combinatorial environment, we studied co-expressed WT- and T421M-KCNH2 genes in ARVMs. For these studies, we infected the ARVMs with equal titers of AdWT- and AdT421M-KCNH2 viral particles and studied the distributions of Kv11.1 channel protein (red staining) and the endogenous Endoplasmic Reticulum (ER) specific protein calnexin (cyan staining), using double immunoflourescence and confocal microscopy. WT-Kv11.1 expression shows its greatest intensity at the cell edges as shown by the sharp peaks in line scan plot (red, Fig. 6j), whereas with T421M-and WT+T421M- expression this pattern is absent. These patterns are consistent with WT-Kv11.1 channel protein trafficking into the cell membrane whereas T421M and WT+T421M-Kv11.1 channel proteins were predominantly restricted to intracellular compartments suggesting that both mutant and WT+mutant channels are trafficking-deficient. In contrast, calnexin (cyan) staining in these three experimental groups lacks the sharp peaks at the cell edge as expected of its intracellular location. Images also show similar sarcomeric pattern to suggest possible co-localization of mutant Kv11.1 and calnexin proteins.
We next characterized the electrophysiology of WT+T421M-Kv11.1 channels. Activation and deactivation properties of the channels were measured as described earlier. Figure 7A shows the voltage-dependence of activation with data normalized to the maximal peak tail IKv11.1. Figure 7B shows summarized data for co-expressed channels compared with WT and T421M channels alone. For co-expressed channels, the V1/2 (−1.5 ±0.9 mV)value is shifted ~13 mV positively compared to WT-Kv11.1 channels and ~29 mV negatively compared to T421M-Kv11.1 channels to give an intermediate phenotype (p<0.05), although closer to the WT-Kv11.1 value. The slope factor, k (8.4±0.5), is slightly greater than that found for WT-Kv11.1 (6.9±0.4, p<0.05) but is not significantly different from T421-Kv11.1 (6.4±0.7, p>0.05). The voltage-dependence relation also supports co-assembly of WT and mutant allele α-subunits into heterotetrameric channels to generate the intermediate electrophysiological phenotype, and not the generation of two separate populations of channels (homotetrameric WT and T421M) which should generate a differently shaped (bi-modal) activation plot reflecting the markedly different activation properties. Peak tail IKv11.1 density was measured for a test step from 60 to −50 mV and was 3.88±0.42 pA/pF, which is a 59.1% reduction compared to the WT-Kv11.1 peak tail IKv11.1 density (p<0.05). Thus, the co-expressed channel peak tail IKv11.1 density is close to T421M-Kv11.1 channels and this is consistent with mutant α-subunit containing channels retaining a partially trafficking-deficient phenotype. Figure 7C shows the time and voltage-dependence of activation of co-expressed (WT+T421M-Kv11.1) channels measured using an envelope of tails method (see Fig. 3) along with plots at 0 and 20 mV comparing co-expressed channels with WT and T421M alone. In contrast to WT-Kv11.1, where activation began at ~−40 mV (Fig. 3), activation for the co-expressed channels began at ~−20 mV. Also the co-expressed channels were nearly fully activated by 20 mV, whereas full activation of T421M-Kv11.1 required depolarization to ~60 mV (Fig. 3). We fitted the envelope of peak tail currents for WT+T421M channels at each prepulse voltage as a single exponential process and these time constants are given in Table 1. Activation properties of WT+T421M channels were intermediate between those of homotetrameric WT- and homotetrameric T421M-Kv11.1 channels although much closer to WT channel properties. We also studied the rates of deactivation in the co-expressed channels compared to homotetrameric WT- and T421M-Kv11.1 channels. The rate of deactivation was measured as described earlier (see Fig. 4) and the voltage-dependence of the fast and slow time constants of deactivation for WT, T421M and WT+T421M channels are shown in Figure 7D for the voltage range −60 to −20 mV as deactivation at more negative voltages was not different (see Fig. 4). At most voltages studied, deactivation of co-expressed channels had an intermediate phenotype closer to that of WT-Kv11.1 channels. We conclude that gating properties of co-assembled WT+T421M-Kv11.1 channels have an intermediate phenotype that is more “WT-like” to suggest that there is not a prominent dominant negative effect of mutant α-subunits on heterotetrameric channel gating properties.
Because the patient experienced cardiac arrest triggered by auditory startle, we studied the sensitivity of WT and mutant Kv11.1 channels expressed in ARVMs to β-adrenergic receptor stimulation. Figure 8 shows example tail current traces (upper panels) for WT- and T421M-Kv11.1. The tail currents were recorded at −50 mV following a 3 sec long prepulse to 20 mV (WT-Kv11.1) or 60 mV (T421M-Kv11.1) sufficient to fully activate the channels with the protocol repeated every 10 sec. After recording the control tail current, the ARVMs were superfused with isoproterenol (10 μM) for 5 min and then the drug was washed out for 5 min. ARVMs were then superfused with E4031 (1 μM) to confirm that the tail current remained E4031-sensitive. The example traces show that isoproterenol increased the peak tail current amplitude and E4031 completely blocked the tail currents. The normalized tail current amplitudes are plotted in the lower panels. For the WT-Kv11.1, peak tail currents were increased by 32% (p<0.05) whereas for the T421M-Kv11.1, the peak tail currents were increased by 22% (p<0.05) following superfusion with isoproterenol. The effect of isoproterenol was partially reversed by washout for WT-channels but not for T421M-Kv11.1 channels.
In the present work, we characterized a LQT2 missense mutation in the S1 transmembrane spanning domain of the Kv11.1 channel expressed in a native adult cardiomyocyte model. The cultured ARVMs we studied lack detectable endogenous IKr, as well as IKs, making this model useful for adenoviral expression of delayed rectifier K+ channel genes. There are several important findings. 1) This is the first study of a LQT2 mutation expressed in a native adult cardiomyocyte model to evaluate biophysical, biochemical and pharmacological properties of homotetrameric and heterotetrameric channels. 2) In ARVMs, the T421M-Kv11.1 mutation showed both altered protein trafficking and channel gating properties. Thus, the diminished IKv11.1 occurs by mixed Class 2 (altered protein trafficking) and Class 3 (altered channel gating) mechanisms.6,17 3) The voltage-dependence of gating properties of heterotetrameric channels does not follow a dominant negative pattern, whereas the trafficking abnormality does.5 4) WT- and T421M-Kv11.1 channels responded to β-adrenergic stimulation by increasing IKv11.1 amplitude.
Kv11.1 channel current, because of its unique slow activation and deactivation gating kinetics relative to its rapid inactivation and recovery from inactivation kinetics, plays an important role late in cardiac repolarization in mammalian heart.1, 2, 18 The T421M mutation markedly altered its voltage-dependence and kinetics of both channel activation and deactivation but had minimal effects on the rates of inactivation and recovery from inactivation. At membrane voltages where WT-Kv11.1 channels activate, T421M-Kv11.1 channels either remain closed or at very positive voltages T421M channels activate slowly and deactivate rapidly, thus the probability of channels being in the open state (Po) is reduced at physiologically relevant voltages. It is interesting that a single uncharged amino acid substitution in the S1 transmembrane domain exerts profound changes on the voltage-dependence of channel availability and the rates of activation and deactivation. Although the Kv11.1 S1-S3 domains are thought to modulate the principle voltage sensor in S4, there are few studies specifically probing the role of the S1 transmembrane domain on the regulation of channel gating. Liu and colleagues studied mutating negatively charged amino acid residues in S1–S3 of Kv11.119 Neutralizing the aspartate to a cysteine (D411C) in the S1 transmembrane domain resulted in channels activating in a more negative voltage range than WT-Kv11.1 channels, and the rates of activation and deactivation were accelerated. Our findings show that T421M-Kv11.1, unlike D411C-Kv11.1, caused a marked positive shift in the voltage-dependence of activation. For T421M-Kv11.1 the slope factor was not significantly different from WT-Kv11.1, which may suggest that the charges involved in sensing the transmembrane voltage for channel opening are not affected, rather the mutation acts to destabilize the open state of the channel.19 In T421M-Kv11.1, it is possible that the methionine (a hydrophilic amino acid) in replacing the threonine (a non-polar hydrophobic amino acid) interacts with water filled crevices surrounding the S1 and S2 transmembrane domains to alter its interaction with the S4 domain to affect activation, and possibly the S4–S5 linker to affect deactivation.20 Our results support recent findings by Lee and colleagues.21 that a molecular “interface” exists between S1 and the pore region of K+ channels that allows for efficient transmission of conformational changes to affect channel gating.
Most LQTS follow an autosomal dominant inheritance pattern with both WT and mutant copies of the gene present in affected individuals. For LQT2, these alleles generate α-subunits that may co-assemble to form both homotetrameric and heterotetrameric channels. For LQT2 missense mutations studied in heterologous expression models (e.g., HEK293 cells, CHO cells, neonatal mouse myocytes, etc) homotetrameric channels commonly have a trafficking-deficient phenotype thought to result from channel protein misfolding and its retention in intracellular compartments including the ER.5,12 Furthermore, co-expression of WT with mutant α-subunits does not correct the trafficking abnormality, rather this results in heterotetrameric channels that retain a dominant negative trafficking-deficient phenotype of variable severity with only a small increase in IKv11.1.5,22 Less is known about the electrophysiological effect of WT α-subunit co-assembly with mutant α-subunits, and whether this follows a dominant negative pattern. Our findings show that the mutant α-subunits do not exert a prominent dominant negative effect, rather the voltage dependence of gating phenotype is close to that of the WT-Kv11.1 channel. This “WT-like” electrophysiological phenotype potentially ameliorates the impact of abnormal gating associated with mutant channels. We conclude that dominant negative interactions of mutant with WT α-subunits on channel trafficking and channel gating appear to be fundamentally different. An intermediate electrophysiological phenotype has been observed for the R56Q LQT2 mutation in the Kv11.1 PAS domain when co-expressed with WT-Kv11.1 in Xenopus oocytes23 and for a KCNJ2 SQT3 mutation co-expressed with WT-KCNJ2 in CHO cells.24
Because the patient experienced arrhythmia triggered by auditory startle, which is a “classic” arrhythmogenic trigger in LQT2 that has been associated with sympathetic activation25–27, we studied the effects of β-adrenergic stimulation on IKv11.1 as cultured ARVMs have intact β-adrenergic signaling.28 In native heart cells controversial findings have been reported with IKr amplitude increased (guinea pig ventricular myocytes and rabbit SA node myocytes)29,30 or decreased (guinea pig ventricular myocytes)31 following β-agonist treatment (for review, see 32). For both WT- and T421M-Kv11.1 channels we found that isoproterenol increased peak tail IKv11.1. This is the first study of β-adrenergic responses using heterologously expressed WT and mutant Kv11.1 channels in adult cardiac myocytes, and this work demonstrates the feasibility of using ARVMs as a native model for IKv11.1 expression and regulation studies. Stimulation of IKv11.1 alone is not likely to explain the patient’s cardiac arrest. Rather it is the interplay of multiple β-adrenergically regulated cardiac ion channels, including L-type Ca2+ channels and IKs, as well as altered cell Ca2+, that likely alter repolarization properties and trigger cardiac arrhythmia (for review, see 33). However, our findings do suggest that Kv11.1 channels may be a direct substrate for β-adrenergic receptor blocker drugs that are commonly used in patients with congenital LQTS.
Recently, LQT2 mutations in α-helices, particularly in the S5–S6 pore domain, were shown to predict a greater risk for cardiac events.34 Our findings may provide new mechanistic insight into the disease phenotype and severity. For mutations having altered gating properties (Class 3 mutations)6,17, the lack of a severe dominant negative effect of mutant α-subunits on the cellular electrophysiology may result in heterotetrameric channels that are more “WT-like” to potentially result in a milder clinical phenotype. In contrast, mutations causing trafficking-deficient channels (Class 2 mutations) 6,17, which usually lie in α-helices, remain trafficking-deficient when co-assembled with WT α-subunits5, thus they retain a more “mutant-like” phenotype to potentially invoke a more severe loss of function and greater risk of cardiac events. LQT2 mutations such as T421M-Kv11.1 that involve both altered gating and trafficking abnormalities form heterotetrameric channels when co-expressed with WT α-subunits may have complex cellular and clinical phenotypes reflecting interactions of protein trafficking and channel gating mechanisms. Studies of additional mutations, including in cardiomyocyte models, are needed to evaluate further our findings.
Congenital Long QT syndrome type 2 (LQT2) is an inherited arrhythmia disorder characterized by prolonged QT intervals on ECGs in the absence of structural heart disease. LQT2 is caused by mutations in the KCNH2 (hERG) gene encoding Kv11.1 potassium channels resulting in reduced repolarizing current (IKr). LQT2 mutations affect channel function mainly by defective channel protein trafficking to cell membrane or abnormal channel gating. We studied the T421M-Kv11.1 missense mutation identified in a multi-generational LQT2 family with variable penetrance of the mutation phenotype. T421M is localized in the Kv11.1 S1-transmembrane domain. We investigated the disease mechanism by heterologous adenoviral expression of mutant and WT channels in adult rat ventricular myocytes (ARVMs). Compared to WT channels, the T421M-Kv11.1 homotetrameric channels have both altered protein trafficking and channel gating properties. When co-expressed, heterotetrameric channels remained trafficking deficient but had “WT-like” gating properties. Thus, the gating properties of heterotetrameric channels do not follow a dominant negative pattern, whereas the trafficking abnormality does. The lack of a severe dominant negative effect of mutant α-subunits on the cellular electrophysiology may result in heterotetrameric channels that potentially result in a milder clinical phenotype. Finally, because the proband experienced cardiac arrest triggered by auditory startle we studied β-adrenergic stimulation of WT- and T421M-Kv11.1 channels in ARVMs. WT- and T421M-Kv11.1 channels responded to isoproterenol by increasing current amplitude suggesting that Kv11.1 channels could be a direct substrate for β-blocker therapy. Our findings may provide new mechanistic insights into disease phenotype and severity.
Funding Sources: This study was supported, in part by, National Heart Lung and Blood Institute R01 HL60723 (CTJ), National Institute of Child Health & Development R01 HD42569 (MJA) and National Heart Lung and Blood Institute R01 HL105713 and American Heart Association-Grant In Aid 11GRNT7610094 (RCB)
Conflict of Interest Disclosures: None.