Liin SI , Larsson JE , Barro-Soria R , Bentzen BH , Larsson HP .

R01 GM109762 NIGMS NIH HHS , R01 HL131461 NHLBI NIH HHS

	Figure 1. Biophysical properties of LQTS and LQTS-like KV7.1+KCNE1 channel mutants expressed in Xenopus oocytes.(a) Topology of KV7.1 and KCNE1, and position of tested LQTS and LQTS-like mutants. (b) G(V) midpoints (V50) from the Boltzmann fits for mutants co-expressed with KCNE1. n = 5–11. Data as mean ± SEM. The statistics represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare the mutants to wild-type KV7.1+KCNE1; **p<0.01; ns is p≥0.05. # denotes lowest estimate. Dashed line denotes wild-type V50. (c) Representative example of KV7.1/S225L+KCNE1 G(V) (black line and symbols) compared to wild-type KV7.1+KCNE1 (blue line and symbols, mean ± SEM, n = 5). (d–e) Representative example of KV7.1/S225L+KCNE1 opening kinetics and KV7.1+KCNE1/K70N closing kinetics (black lines) compared to wild-type KV7.1+KCNE1 (blue lines).DOI: http://dx.doi.org/10.7554/eLife.20272.003Figure 1—figure supplement 1. KV7.1/F351S mutant expressed in Xenopus oocytes.The KV7.1/F351S mutant does not generate currents when expressed in Xenopus oocytes. The holding voltage is –80 mV, and test voltages range between –80 and +60 mV for 3 s in 10 mV increments. The tail voltage is –20 mV.DOI: http://dx.doi.org/10.7554/eLife.20272.004
	Figure 1—figure supplement 2. V50 of LQTS and LQTS-like KV7.1 mutants expressed in Xenopus oocytes.G(V) midpoints (V50) for LQTS and LQTS-like mutants without co-expression of KCNE1. Mean ± SEM. n = 5–12. The statistics represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare V50 of mutant to V50 of wild-type KV7.1; **p<0.01; ns p≥0.05. Dashed line denotes wild-type V50.DOI: http://dx.doi.org/10.7554/eLife.20272.005
	Figure 1—figure supplement 3. KV7.1/R583C mutant expressed in Xenopus oocytes.(a) The KV7.1/R583C mutant generates currents that inactivate at positive voltages. The holding voltage is –80 mV, and test voltages range between –80 and +40 mV for 3 s in 20 mV increments. The tail voltage is –20 mV. Tail currents are measured at the arrow. Inset: representative current trace at +40 mV for wild-type KV7.1. (b) Representative example of G(V) curves generated using the protocol in panel a (filled circles) or a triple pulse protocol (open circles) with a brief hyperpolarizing pulse (–140 mV for 20 ms) between the test pulse and the tail pulse to release a fraction of channels from inactivation. The triple pulse protocol generates a G(V) that is shifted ~9 mV towards positive voltages (V50 = ~ –39 mV for the regular protocol and –30 mV for the triple pulse protocol), which matches the G(V) of the wild-type KV7.1 fairly well (V50 = –29.4 mV).DOI: http://dx.doi.org/10.7554/eLife.20272.006
	Figure 1—figure supplement 4. Comparison of current amplitude of wild-type KV7.1+KCNE1 and LQTS and LQTS-like mutants when expressed in Xenopus oocytes.KV7.1 and KCNE1 were co-injected in oocytes for homozygous (a) and heterozygous (b) expression, as described in Materials and Methods. Current were recorded after two days of incubation at 16°C. The holding voltage is –80 mV, and test voltages range between 0 and +60 mV for 5 s in 20 mV increments. The tail voltage is –20 mV. Current amplitudes at the end of the 5 s test pulse are normalized to the wild-type KV7.1+KCNE1 current amplitude at +60 mV recorded in the same batch of oocytes. Dashed line in (a) is the wild-type curve shifted +25 mV. (c–d) Detailed comparison of current amplitudes at +20 mV (c) and +40 mV (d). Mutant current amplitudes are normalized to the wild-type KV7.1+KCNE1 current amplitude at indicated voltage. Dashed lines denote relative wild-type KV7.1+KCNE1 current amplitude (=1). The statistics represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare the current amplitude of mutants to wild-type current amplitudes. *p<0.05; **p<0.01; ***p<0.001; ns is p≥0.05. Mean ± SEM. n = 4–12.DOI: http://dx.doi.org/10.7554/eLife.20272.007
	Figure 2. Comparison of homozygous and heterozygous expression of LQTS and LQTS-like mutants.(a–b) Representative example of kinetics (middle panel) and G(V) (right panel) for homozygous expression and heterozygous expression of S225L (a) and K70N (b). Currents in response to steps from –80 mV to +40 mV (a, middle pane) and from +40 mV to –20 mV (b, middle panel). Homozygous expression (black), heterozygous expression (gray), and KV7.1+KCNE1 wild-type (blue). n = 7–13. (c–d) Summary of V50 (c) and T50 for closing (d) for homozygous and heterozygous expression. Data as mean ± SEM. n = 5–13. The statistics represent one-way ANOVA with pair-wise Bonferroni’s Test to compare homozygous and heterozygous expression; **p<0.01; ***p<0.001; ns is p≥0.05. # denotes lowest estimate. Not determined (nd). The statistics was not calculated for F351A. Dashed lines denote corresponding values for wild-type KV7.1+KCNE1.DOI: http://dx.doi.org/10.7554/eLife.20272.008
	Figure 3. Voltage-clamp fluorometry recordings of wild-type and mutated KV7.1+KCNE1 channels.(a-c) Representative fluorescence traces and mean F(V)/G(V) curves for KV7.1/G219C* (a), S225L (b), and F351A (c). Left panels without KCNE1 and right panels with KCNE1. The holding voltage is –80 mV, the pre-pulse –120 mV for 2 s (left panels) and –160 mV for 5 s (right panels), and test voltages between –140 and +80 mV for 3 s (left panels) and between –160 and +80 mV for 5 s (right panels) in 20 mV increments. The tail voltage is –80 mV (left panels) and −40 mV (right panels). For KV7.1/G219C*/F351A+KCNE1, the pre-pulse is –120 mV for 3 s, and test voltages ranging between –160 and +100 mV. The bottom of the fit of the KV7.1/G219C*/S225L+KCNE1 F(V) curve (which saturates fairly well at negative voltages) is set to 0 in the normalized F(V) curves in the right panels. The F1 amplitude of KV7.1/G219C*/F351A+KCNE1 is normalized to the F1 amplitude of wild-type. Data as mean ± SEM. n = 4–14. The dashed lines in (b) and (c) denote F(V) (red) and G(V) (black) for wild-type (from a).DOI: http://dx.doi.org/10.7554/eLife.20272.009Figure 3—figure supplement 1. Kinetic models for KV7.1 and KV7.1+KCNE1 channel gating.(a) A 10-state allosteric gating scheme for KV7.1 channels. Horizontal transitions represent independent S4 movements that increase the fluorescence to an intermediate level (which generates the F1 component). The vertical transition represents concerted channel opening with a concomitant additional fluorescence increase (which generates the F2 component). Cartoon shows KV7.1 channel labeled with a fluorophore on S3-S4 with all four voltage sensors in the resting state (C0), with one (C1), or four (C4) voltage sensor activated in the closed channel (top) or with all four voltage sensors in the resting state (O0), with one (O1), or four (O4) voltage sensor activated with the channel opened (bottom). (b) A 6-state allosteric gating scheme for KV7.1+KCNE1 channels. Horizontal transitions represent independent S4 movements that increase the fluorescence to an intermediate level (which generates the F1 component). The vertical transition represents concerted channel opening with a concomitant additional fluorescence increase (which generates the F2 component). Cartoon shows KV7.1 channel labeled with a fluorophore on S3-S4 with all four voltage sensors in the resting state (C0), with one (C1), or four (C4) voltage sensor activated without channel opening (top) that is followed by a concerted conformational change of all four S4s associated with channel opening (O4) (bottom).DOI: http://dx.doi.org/10.7554/eLife.20272.010
	Figure 3—figure supplement 2. Simulations of wild-type and mutant KV7.1 and KV7.1+KCNE1 channels reproduce currents and fluorescence.Simulated G(V) (black) and F(V) (red) curves for (a) wild-type, (b) S225L, and (c) F351A KV7.1 (left) and KV7.1+KCNE1 (right) channels using the KV7.1 and KV7.1+KCNE1 models in Figure 3—figure supplement 1. Parameters for the wild-type models were determined in earlier studies (see Supplementary file 4 for all rate constants). Current and fluorescence traces were simulated using Berkeley Madonna (Berkeley, CA).DOI: http://dx.doi.org/10.7554/eLife.20272.011
	Figure 3—figure supplement 3. Voltage-clamp fluorometry recordings of the KV7.1/G219C*/F351L mutant with and without KCNE1 co-expressed.Mean F(V)/G(V) curves for KV7.1/G219C*/F351L (mean ± SEM) are shown together with corresponding mean F(V)/G(V) curves for WT KV7.1/G219C* (blue lines) and KV7.1/G219C*/F351A (dashed red/black lines). Experiments are performed and data normalized as described in Figure 3. Note that all data presented in this graph are done on constructs with a Kv7.1/C214A/C331A background (Barro-Soria et al., 2014; Barro-Soria and Perez, 2015). The F(V)/G(V) curves are therefore shifted towards negative voltages compared the data presented in Figure 3 (which are done in WT background). n = 4–6.DOI: http://dx.doi.org/10.7554/eLife.20272.012
	Figure 4. Effect of N-AT on LQTS and LQTS-like mutants.All these experiments are done in the presence of KCNE1. Structure of N-AT is shown. (a–b) Representative effect of 7–70 µM N-AT on current amplitude (a) and G(V) (b) of KV7.1/S225L+KCNE1. Dashed line in (a) denotes 0 µA. (c) Representative currents generated by KV7.1/S225L+KCNE1 during pulsing at 1 Hz and +28°C in control solution (black) and after the cell had been bathed continuously in 70 µM N-AT (light to dark green, # denotes sweep order). Inset: corresponding currents from wild-type KV7.1+KCNE1 scaled similarly as KV7.1/S225L+KCNE1. Light grey trace denotes sweep #1, grey trace denotes sweep #2, and dark grey trace denotes sweep #20. (d) Summary of V50 for LQTS and LQTS-like mutants before and after 70 µM N-AT application. Dashed line denotes V50 for wild-type KV7.1+KCNE1. (e–f) Summary of ΔV50 (e) and ΔΔGo (f) for LQTS and LQTS-like mutants induced by 70 µM N-AT. # denotes an approximation. Dashed lines denote corresponding ΔV50 and ΔΔGo induced by 70 µM N-AT for wild-type KV7.1+KCNE1. The statistics in (f) represent one-way ANOVA with Dunnett’s Multiple Comparison Test to compare the N-AT-induced change in ΔΔGo of mutants to N-AT-induced change in ΔΔGo of wild-type KV7.1+KCNE1; *p≤0.05. Only significant differences shown in (f), other comparisons have p>0.05. (g) Estimate of the ability of 70 µM N-AT to restore LQTS and LQTS-like mutant current amplitude at +40 mV. The mean N-AT induced increase in current amplitude for each mutant (from Figure 4—figure supplement 4b) is multiplied with the control amplitude for each mutant (from Figure 1—figure supplement 4d). Not determined (nd). Data as mean ± SEM. n = 5–12. Dashed line denotes relative wild-type KV7.1+KCNE1 current amplitude in control solution (i.e. without N-AT).DOI: http://dx.doi.org/10.7554/eLife.20272.013Figure 4—figure supplement 1. N-AT effect on wild-type KV7.1+KCNE1 expressed in Xenopus oocytes.Representative effect of 70 µM N-AT on current amplitude (a) and G(V) (b) of wild-type KV7.1+KCNE1. The holding voltage is –80 mV and the tail current amplitude in (b) measured at –20 mV after 5 s test pulses. Dashed line in (a) denotes 0 µA current. G(V) curves in (b) are normalized to the fitted Gmax(as described in Materials and methods).DOI: http://dx.doi.org/10.7554/eLife.20272.014
	Figure 4—figure supplement 2. The time course of N-AT wash-in on KV7.1/S225L+KCNE1 expressed in Xenopus oocytes.Representative example showing that N-AT effects on current amplitude reaches steady state for each concentration within minutes. The holding voltage is –80 mV and current amplitude measured at the end of a 5 s test pulse to +20 mV. Dashed line denotes baseline (control amplitude).DOI: http://dx.doi.org/10.7554/eLife.20272.015
	Figure 4—figure supplement 3. ‘Leak’ component of KV7.1/S225L+KCNE1.Currents generated by the KV7.1/S225L+KCNE1 mutant have a small instantaneous ‘leak’ component at positive voltages. Dashed line denotes 0 µA.DOI: http://dx.doi.org/10.7554/eLife.20272.016
	Figure 4—figure supplement 4. Effect of N-AT on current amplitude of LQTS and LQTS-like mutants.(a-b) Current amplitudes in the presence of 70 µM N-AT measured at the end of a 5 s test pulse to +20 mV (a) or +40 mV (b). The currents are normalized to the current amplitude in control solution in the same oocyte. Dashed lines denote N-AT effects on wild-type KV7.1+KCNE1 current amplitude (a factor 2.9 ± 0.4 and 1.9 ± 0.3 (n = 5), respectively). (c) Ability of 70 µM N-AT to restore LQTS and LQTS-like mutant current amplitude at +20 mV. The mean N-AT-induced fold increase in current amplitude for each mutant (data from panel a) is multiplied by the relative current amplitude for each mutant compared to wild-type KV7.1+KCNE1 in control solution (from Figure 1—figure supplement 4). Dashed line denotes relative wild-type KV7.1+KCNE1 current amplitude in control solution (i.e. without N-AT). Mean ± SEM. n = 4–12. nd = not determined.DOI: http://dx.doi.org/10.7554/eLife.20272.017
	Figure 5. Effect of 70 µM N-AT on S4 movement and gate opening in S225L and F351A mutants.(a–c) Representative example of the effect of 70 µM N-AT on F1 kinetics (a), current opening kinetics (b), and F2 kinetics (c) in KV7.1/G219C*/S225L+KCNE1. Control fluorescence (red) and current (black). N-AT fluorescence (magenta) and current (green). Top in (c) shows an overlay of the later part of the fluorescence (after most of F1 has occurred) and the later part of the currents (after the initial delay) before and after application of N-AT. Middle and lower (c) show that there is not a great overlap of the fluorescence in the presence of N-AT and the current in control solution (middle) or the fluorescence in control solution and the current in the presence of N-AT (lower). (d–e) Representative example of effect of 70 µM N-AT on F1 kinetics (d) and current opening kinetics (e) in KV7.1/G219C*/F351A+KCNE1. Same colouring as in (a–b). Dashed line in (b) and (e) denotes 0 µA. Fluorescence traces and all traces in (c) have been normalized to better allow temporal comparison. (f) Summary of the effect of 70 µM N-AT on the kinetic parameters of KV7.1/G219C*/S225L+KCNE1 and KV7.1/G219C*/F351A+KCNE1. Kinetics of the fast (F1) and slow (F2) fluorescence components were deduced from a double-exponential function fitted to the fluorescence traces. The kinetics of currents were deduced from a single-exponential function fitted to current traces. Ratios of time constants (τN-AT/τCtrl) were calculated pair-wise (control compared to N-AT) in each oocyte and analysed using two-tailed one sample t-test where ratios were compared with a hypothetical value of 1. Data as mean ± SEM. n = 4 (3for fluorescence kinetics for KV7.1/G219C*/F351A+KCNE1). *p<0.05; **p<0.01. nd = not determined.DOI: http://dx.doi.org/10.7554/eLife.20272.018Figure 5—figure supplement 1. Effect of N-AT on the F(V) of KV7.1/G219C*/S225L mutant co-expressed with KCNE1 in Xenopus oocytes.(a) Representative example of the time course of the reduction in fluorescence intensity upon N-AT application. The fluorescence intensities shown is the fluorescence measured at +80 mV (during repeated applications of the voltage protocol used to measure the complete F(V) as in panel c), normalized to the fluorescence intensity at +80 mV recorded in the first F(V) in control solution. Red symbols denote control (without N-AT) and purple symbols denote in the presence of N-AT. The fluorescence signal reduces with time in the presence of N-AT. In contrast, the fluorescence signal is preserved in the absence of N-AT (red symbol, recorded in another cell). (b) Summary of fluorescence emission monitored from unbound Alexa488 in control solution and in taurine-supplemented control solution (0.25 or 0.5 M taurine). In these experiments, no oocytes or channels were present. A.U. denotes arbitrary units. Data as mean ± SEM. n = 3. (c) Mean F(V) curve for KV7.1/G219C*/S225L+KCNE1 in the absence (red symbols, data from Figure 3B) or presence of 70 µM N-AT. The holding voltage is –80 mV, the pre-pulse –160 mV for 5 s, and test voltages between –160 and +100 mV for 5 s in 20 mV increments. The tail voltage is –40 mV. Each F(V) curve is normalized between 0 and 1 based on the bottom and top deduced from the double Boltzmann fits for each curve (see Materials and methods). Data as mean ± SEM. n = 3 for N-AT.DOI: http://dx.doi.org/10.7554/eLife.20272.019

Anderson, Large-scale mutational analysis of Kv11.1 reveals molecular insights into type 2 long QT syndrome. 2014, Pubmed

Anderson, Large-scale mutational analysis of Kv11.1 reveals molecular insights into type 2 long QT syndrome. 2014, Pubmed
Barhanin, K(V)LQT1 and lsK (minK) proteins associate to form the I(Ks) cardiac potassium current. 1996, Pubmed , Xenbase
Barro-Soria, KCNE3 acts by promoting voltage sensor activation in KCNQ1. 2015, Pubmed , Xenbase
Barro-Soria, KCNE1 divides the voltage sensor movement in KCNQ1/KCNE1 channels into two steps. 2014, Pubmed
Bianchi, Mechanisms of I(Ks) suppression in LQT1 mutants. 2000, Pubmed , Xenbase
Busch, Effects of [Ca2+]i and temperature on minK channels expressed in Xenopus oocytes. 1993, Pubmed , Xenbase
Börjesson, Electrostatic tuning of cellular excitability. 2010, Pubmed , Xenbase
Börjesson, An electrostatic potassium channel opener targeting the final voltage sensor transition. 2011, Pubmed , Xenbase
Cerrone, A clinical approach to inherited arrhythmias. 2012, Pubmed
Chowdhury, Estimating the voltage-dependent free energy change of ion channels using the median voltage for activation. 2012, Pubmed , Xenbase
DeCaen, Disulfide locking a sodium channel voltage sensor reveals ion pair formation during activation. 2008, Pubmed
Delisle, Biology of cardiac arrhythmias: ion channel protein trafficking. 2004, Pubmed
Deschênes, Biophysical characteristics of a new mutation on the KCNQ1 potassium channel (L251P) causing long QT syndrome. 2003, Pubmed
Eldstrom, Mechanistic basis for LQT1 caused by S3 mutations in the KCNQ1 subunit of IKs. 2010, Pubmed
Harmer, Mechanisms of disease pathogenesis in long QT syndrome type 5. 2010, Pubmed
Hedley, The genetic basis of long QT and short QT syndromes: a mutation update. 2009, Pubmed
Henrion, Long QT syndrome-associated mutations in the voltage sensor of I(Ks) channels. 2009, Pubmed , Xenbase
Jackson, LQTS in Northern BC: homozygosity for KCNQ1 V205M presents with a more severe cardiac phenotype but with minimal impact on auditory function. 2014, Pubmed
Jespersen, Dual-function vector for protein expression in both mammalian cells and Xenopus laevis oocytes. 2002, Pubmed , Xenbase
Lai, Denaturing high-performance liquid chromatography screening of the long QT syndrome-related cardiac sodium and potassium channel genes and identification of novel mutations and single nucleotide polymorphisms. 2005, Pubmed
Leitner, Restoration of ion channel function in deafness-causing KCNQ4 mutants by synthetic channel openers. 2012, Pubmed
Li-Smerin, Helical structure of the COOH terminus of S3 and its contribution to the gating modifier toxin receptor in voltage-gated ion channels. 2001, Pubmed
Liin, The KCNQ1 channel - remarkable flexibility in gating allows for functional versatility. 2015, Pubmed
Liin, Polyunsaturated fatty acid analogs act antiarrhythmically on the cardiac IKs channel. 2015, Pubmed , Xenbase
Milescu, Opening the shaker K+ channel with hanatoxin. 2013, Pubmed , Xenbase
Monks, Helical structure and packing orientation of the S2 segment in the Shaker K+ channel. 1999, Pubmed , Xenbase
Morita, The QT syndromes: long and short. 2008, Pubmed
Murray, Unnatural amino acid photo-crosslinking of the IKs channel complex demonstrates a KCNE1:KCNQ1 stoichiometry of up to 4:4. 2016, Pubmed
Nakajo, Stoichiometry of the KCNQ1 - KCNE1 ion channel complex. 2010, Pubmed , Xenbase
Nakajo, KCNQ1 channel modulation by KCNE proteins via the voltage-sensing domain. 2015, Pubmed
Napolitano, Genetic testing in the long QT syndrome: development and validation of an efficient approach to genotyping in clinical practice. 2005, Pubmed
Nerbonne, Molecular physiology of cardiac repolarization. 2005, Pubmed
O'Hara, Simulation of the undiseased human cardiac ventricular action potential: model formulation and experimental validation. 2011, Pubmed
Osteen, KCNE1 alters the voltage sensor movements necessary to open the KCNQ1 channel gate. 2010, Pubmed , Xenbase
Osteen, Allosteric gating mechanism underlies the flexible gating of KCNQ1 potassium channels. 2012, Pubmed , Xenbase
Pathak, The cooperative voltage sensor motion that gates a potassium channel. 2005, Pubmed , Xenbase
Perry, Rescue of protein expression defects may not be enough to abolish the pro-arrhythmic phenotype of long QT type 2 mutations. 2016, Pubmed , Xenbase
Phillips, Position and motions of the S4 helix during opening of the Shaker potassium channel. 2010, Pubmed
Piacentino, Cellular basis of abnormal calcium transients of failing human ventricular myocytes. 2003, Pubmed
Plant, Individual IKs channels at the surface of mammalian cells contain two KCNE1 accessory subunits. 2014, Pubmed
Priori, Low penetrance in the long-QT syndrome: clinical impact. 1999, Pubmed
Priori, A recessive variant of the Romano-Ward long-QT syndrome? 1998, Pubmed , Xenbase
Ruscic, IKs channels open slowly because KCNE1 accessory subunits slow the movement of S4 voltage sensors in KCNQ1 pore-forming subunits. 2013, Pubmed
Sanguinetti, Coassembly of K(V)LQT1 and minK (IsK) proteins to form cardiac I(Ks) potassium channel. 1996, Pubmed , Xenbase
Sanguinetti, HERG1 channelopathies. 2010, Pubmed
Sato, Glycerol reverses the misfolding phenotype of the most common cystic fibrosis mutation. 1996, Pubmed
Schmitt, The novel C-terminal KCNQ1 mutation M520R alters protein trafficking. 2007, Pubmed
Seebohm, Dependence of I(Ks) biophysical properties on the expression system. 2001, Pubmed , Xenbase
Seebohm, Pharmacological activation of normal and arrhythmia-associated mutant KCNQ1 potassium channels. 2003, Pubmed , Xenbase
Silva, Subunit interaction determines IKs participation in cardiac repolarization and repolarization reserve. 2005, Pubmed
Splawski, Mutations in the hminK gene cause long QT syndrome and suppress IKs function. 1997, Pubmed , Xenbase
Vyas, Phenotype guided characterization and molecular analysis of Indian patients with long QT syndromes. 2016, Pubmed
Wilson, Abnormal KCNQ1 trafficking influences disease pathogenesis in hereditary long QT syndromes (LQT1). 2005, Pubmed
Xiong, Zinc pyrithione-mediated activation of voltage-gated KCNQ potassium channels rescues epileptogenic mutants. 2007, Pubmed
Yamaguchi, Clinical and electrophysiological characterization of a novel mutation (F193L) in the KCNQ1 gene associated with long QT syndrome. 2003, Pubmed , Xenbase
Yang, Allelic variants in long-QT disease genes in patients with drug-associated torsades de pointes. 2002, Pubmed
Yang, An allosteric mechanism for drug block of the human cardiac potassium channel KCNQ1. 2013, Pubmed
Zaydman, Domain-domain interactions determine the gating, permeation, pharmacology, and subunit modulation of the IKs ion channel. 2014, Pubmed , Xenbase
Zhang, Identification of a novel KCNQ1 mutation associated with both Jervell and Lange-Nielsen and Romano-Ward forms of long QT syndrome in a Chinese family. 2008, Pubmed