Molecular characterization of two founder mutations causing long QT syndrome and identification of compound heterozygous patients

Abstract

Background. Mutations of at least six different genes have been found to cause long QT syndrome (LQTS), an inherited arrhythmic disorder characterized by a prolonged QT interval on the electrocardiogram (ECG), ventricular arrhythmias and risk of sudden death.

Aim. The aims were to define the yet undetermined phenotypic characteristics of two founder mutations and to study clinical features in compound heterozygotes identified during the course of the study.

Methods. To maximize identification of the compound heterozygotes, we used an extended group of LQTS patients comprising 700 documented or suspected cases. Functional studies were carried out upon transient expression in COS‐7 or HEK293 cells.

Results. The KCNQ1 IVS7‐2A>G (KCNQ1‐FinB) mutation associated with a mean QTc interval of 464 ms and a complete loss‐of‐channel function. The HERG R176W (HERG‐FinB) mutation caused a reduction in current density as well as slight acceleration of the deactivation kinetics in vitro, and its carriers had a mean QTc of 448 ms. The HERG R176W mutation was also present in 3 (0.9%) out of 317 blood donors. A total of six compound heterozygotes were identified who had the HERG R176W mutation in combination with a previously reported LQTS mutation (KCNQ1 G589D or IVS7‐2A>G). When present simultaneously with an apparent LQTS‐causing mutation, the HERG R176W mutation may exert an additional in vivo phenotypic effect.

Conclusions. The HERG R176W mutation represents a population‐prevalent mutation predisposing to LQTS. Compound heterozygosity for mutant LQTS genes may modify the clinical picture in LQTS.

• Compound heterozygous carriers
• founder mutations
• ion channel
• long QT syndrome
• patch‐clamp

Long QT syndrome (LQTS) is a familial arrhythmic disease, characterized by prolonged ventricular repolarization, manifesting usually as a prolonged QT interval on the electrocardiogram (ECG). Clinical manifestations include ventricular arrhythmias, particularly torsades de pointes, syncope, seizures and risk of sudden death 1,2. Acquired forms of LQTS may be caused by metabolic disturbances (hypokalemia, hypomagnesemia) or by exposure to certain QT‐prolonging drugs 3. Until now, molecular genetic studies have disclosed six forms of inherited LQTS caused by mutations in five voltage‐gated cardiac ion channels (KCNQ1, HERG, KCNE1, KCNE2, and SCN5A), or the adaptor protein ankyrin‐Β 4,5. Characteristically, the disease‐causing mutation is family specific, but recently several founder mutations have been identified in Finland 6–8, Norway 9 and South Africa 10,11. However, considerable variation in phenotypic expression is present between and within families carrying the same mutation, suggesting that some additional genetic and/or non‐genetic factors add to the outcome of the disease 12.

Patients simultaneously carrying two different LQTS mutations, i.e. compound heterozygotes, appear to present a more severe phenotype compared to patients with only one mutation 13–17. The study by Schwartz et al. 13 detected compound mutations in approximately 5% of LQTS probands studied, and demonstrated that the compound mutation carriers had longer QTc intervals and a higher risk for syncope and cardiac arrest than patients with only one mutation.

We have previously identified four common mutations, two (G589D and IVS7‐2A>G) in the KCNQ1 gene and two (R176W and L552S) in the KCNH2 gene, together accounting for 73% of the known genetic spectrum underlying LQTS in Finland 6–8,18. The in vitro electrophysiological characteristics of the G589D and L552S mutations were described in our previous papers 6,7. While expanding our previous material of LQTS probands 8, we became aware that several LQTS probands in fact carried two different mutations simultaneously, with the HERG R176W mutation always as a partner. The fact that the latter mutation was subsequently also identified in occasional blood donors suggested that it would represent a ‘mild’ but relatively common LQTS gene. Therefore, we decided to carry out a detailed examination of the electrophysiological characteristics of the R176W mutated HERG channel. In addition, we screened for its occurrence in an extended cohort of 700 consecutive cases with documented or suspected LQTS referred to our unit, in order to identify sufficient numbers of R176W carriers to provide assessment of its QT‐prolonging effect in vivo, alone or in combination with other LQTS genes.

Key messages

• Two founder mutations, KCNQ1 IVS7‐2A>G and HERG R176W, associate with a prolonged QT interval and altered channel function in vitro.

• HERG R176W may represent a population‐prevalent mutation causing long QT syndrome (LQTS), and it may exert an additional in vivo phenotypic effect when present simultaneously with an apparent LQTS‐causing mutation.

Patients and methods

Study subjects

A total of 700 consecutive unrelated probands, with suspected (n = 550) or DNA‐verified diagnosis (n = 150, 8) of LQTS, were examined (Figure 1). Each proband was screened for the presence of four putative Finnish LQTS founder gene mutations, including KCNQ1 G589D (KCNQ1‐FinA), KCNQ1 IVS7‐2A>G (KCNQ1‐FinB), HERG L552S (HERG‐FinA) and HERG R176W (HERG‐FinB). The clinical indications for DNA studies were as follows: QT interval prolongation (with or without QT prolonging drug) according to ECG data, syncopal spell of unknown etiology, aborted cardiac arrest, documented torsades de pointes or sudden death.

Figure 1. Flowchart describing the screening process and the selection of patients and controls.

In addition, we have previously identified the SCN5A variant R190G, which appeared in both LQTS patients and blood donors 8. This variant was also included in the present study in order to gain further insights into its electrophysiological characteristics.

Available family members of the probands with established mutations were invited to clinical examinations and a blood sample was taken for DNA analysis. As controls, we studied DNA samples from 317 Finnish blood donors (mean age 45 years, equal numbers of men and women) kindly provided by Dr Tom Krusius (The Finnish Red Cross Blood Service, Helsinki). The original study protocol did not permit disclosure of any health information of these blood donors. This study was approved by the Ethics Review Committee of the Department of Medicine, University of Helsinki, and all patients have given their written informed consent for the genetic studies. The investigation conforms with the principles outlined in the Declaration of Helsinki 19.

Statistical methods

A t test was used to compare the QTc intervals between carriers and non‐carriers (Table II). Different types of t tests were applied depending on whether the QT duration showed a normal distribution or not, and depending on whether the variances were equal or not. The Kolmogorov‐Smirnov test for different distributions was used for comparing QTc intervals between carriers and non‐carriers of KCNQ1 G589D and HERG L552S. The Aspin‐Welch unequal‐variance test and the Mann‐Whitney U test for difference in medians was used for comparing QTc intervals between carriers and non‐carriers of KCNQ1 IVS7‐2A>G and HERG R176W, respectively. ANOVA or Mann‐Whitney test was used to compare electrophysiological recordings between wild‐type and mutant channels. A P‐value <0.05 was considered statistically significant.

DNA analysis

The KCNQ1 G589D and HERG L552S mutations were detected essentially as described earlier 6,7. The HERG R176W and KCNQ1 IVS7‐2A>G mutations were assayed by restriction enzyme analysis. The regions of interest were amplified by polymerase chain reaction (PCR) with the primers: 5′‐acgaccacgtgcctctcctctc‐3′ and 5′‐ggctggggcggaacgggtcc‐3′ for R176W, and 5′‐ggggagctgtagcttccata‐3′ and 5′‐agccaaatgcatggtgagat‐3′for IVS7‐2A>G. For HERG R176W assay, the PCR product was digested with 1 U of SmaI (New England Biolabs, Beverly, MA, USA) at 30°C for 4 h. This digestion produces fragments of the following sizes: 196, 182, 82 and 46 bp in wild‐type, and 196, 182, 128, 82 and 46 bp in R176W heterozygotes. The PCR product containing KCNQ1 IVS7‐2A>G was digested at 37°C for 4 h with 1 U of DdeI (New England Biolabs). This digestion produces fragments of the following sizes: 228, 39, and 33 bp in wild‐type, and 261, 228, 39 and 33 bp in IVS7‐2A>G heterozygotes.

RT‐PCR of the IVS7‐2A>G splice‐site mutation in the KCNQ1 gene

The KCNQ1 IVS7‐2A>G mutation 8 was further studied by reverse transcription (RT)‐PCR. Samples were analyzed from three controls and two mutation carriers. Lymphocytes were isolated from 8 mL of fresh blood using BD Vacutainer CPT tubes (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). RNA was isolated with the RNeasy kit from Qiagen Inc. (Valencia, CA, USA) with slight modifications of the protocol. The cDNA was prepared with the Superscript First‐Strand synthesis system (Invitrogen, Carlsbad, CA, USA) using oligo(dT) primers provided by the kit, and thereafter amplified with KCNQ1 specific primers: 5′‐tttgccatctccttctttgc‐3′ and 5′‐gtctccccttccaggtcc‐3′ resulting in a 482 bp product. After running the PCR product on 1.5% agarose gel, a longer fragment, in addition to the 482 bp fragment, could be detected in the patients' samples only. This PCR product was subcloned with a PCR cloning kit (Qiagen Inc.), and several clones were sequenced to identify the different fragments detected on the agarose gel.

Mutagenesis and in vitro transcription

The human KCNH2 cDNA (U04270) in the pcDNA3.1 vector (Invitrogen, Carlsbad, CA, USA) was kindly provided by Dr Gail Robertson (University of Wisconsin, Madison, WI, USA). The human SCN5A (NM_198056; hH1C‐Q1077; the 2016aa transcript variant 1) in the pcDNA3.1 vector was a kind gift from Dr J. Makielski (University of Wisconsin Hospital and Clinics, Madison, WI, USA). Site‐directed mutagenesis was performed using the QuickChange site‐directed mutagenesis kit (Stratagene, CA, USA). The mutagenic primers were 5′ cgttcactttccttggggacccatggaactgg 3′ and 5′ ccagttccatgggtccccaaggaaagtgaacg 3′ for the R190G variant, and 5′ ggcgctgacggcctgggagtcgtcggtg 3′ and 5′ caccgacgactcccaggccgtcagcgcc 3′ for the R176W mutation. A fragment of 160 bp containing the 30 bp insert, ultimately detected in the KCNQ1 IVS7‐2A>G mutant allele, was cloned into the KCNQ1 human cDNA (NM_000218 in the pIRES‐CD8 plasmid) with the restriction enzymes BsgI and BspEI (New England Biolabs). The mutagenic clones for all the constructs were sequenced to confirm the presence of the mutations and to exclude possible errors.

Heterologous expression of the HERG R176W and KCNQ1 IVS7‐2A>G mutations in COS‐7 cells

COS‐7 cells were maintained as described 20. Cells were transiently transfected by the DEAE (diethylaminoethyl)‐dextran precipitate method with wild‐type (WT) KCNQ1, KCNQ1 IVS7‐2A>G, KCNQ1+KCNQ1 IVS/‐2A>G, KCNQ1+KCNE1, KCNQ1 IVS7‐2A>G+KCNE1, or KCNQ1+KCNQ1 IVS7‐2A>G+KCNE1. To study the functional consequences of HERG R176W, the following combinations were used in COS‐7 cells: HERG‐WT, HERG R176W, and HERG‐WT+HERG R176W. Current recordings were performed 48 hours after transfection in the whole‐cell configuration at room temperature (∼24°C), using EPC‐10 amplifier (HEKA Instruments Inc.), and as previously described 21. Deactivating current traces were fitted using a two‐exponential function: I(t) = a₀+a₁*exp(−t/τ₁)+a₂*exp(−t/τ₂), where I(t) is the current, a₀, a₁, and a₂ the current amplitude, t is the time, and τ₁ and τ₂ the time constants of deactivation. Data acquisition and analysis were performed using Pulse and Pulsefit (HEKA Electronic, Germany), and IgorPro (WaveMetrics Inc., OR) softwares.

Heterologous expression of the SCN5A R190G variant in HEK293 cells

Wild‐type and R190G sodium currents were recorded as described previously 22. Persistent sodium currents were obtained by subtracting to control currents evoked by a 300‐ms long pulse to −25 mV traces recorded in presence of 30 µM tetrodotoxin (Alomone, Israel).

Results

Identification of Finnish LQTS compound heterozygous patients

Upon screening a total of 700 Finnish probands known or suspected to have LQTS, we identified four families who had the HERG R176W mutation in combination with another LQTS gene mutation (KCNQ1 G589D or IVS7‐2A>G) (Table I). In family 139 the mother and uncle of the proband were also compound heterozygotes (Table I). As the only mutation the KCNQ1 G589D was identified in 66, KCNQ1 IVS7‐2A>G in 13, HERG R176W in 12, and HERG L552S in 13 of the 700 probands.

Table I. Compound heterozygote carriers with mutations in the KCNQ1 and KCNH2 genes.

Mutation combination	Family code	Age/sex	QTc	Symptoms^c	Event and trigger description
KCNQ1 G589D+HERG R176W	129^a	4^b/M	457	Yes	Episodes of convulsions and lethargy during fever. Sudden death at 4 yr.
KCNQ1 G589D+HERG R176W	080^a	22/M	508	Yes	Near‐drowning
KCNQ1 G589D+HERG R176W	139^a	5/M	451	No	‐
KCNQ1 G589D+HERG R176W	139	27/M	447	Yes	Syncope while standing up quickly
KCNQ1 G589D+HERG R176W	139	30/F	502	No	‐
KCNQ1 IVS7‐2A>G+HERG R176W	373^a	38/F	486	Yes	Palpitations, syncope

^a Proband, ^b this individual deceased at the age of 4 years, ^c sudden death, documented torsades de pointes arrhythmia, or any loss of consciousness.

Clinical data of the compound heterozygous patients

The proband of family 129 had experienced episodes of flaccidity and apparent fever convulsions, and died suddenly at the age of 4 years (Table I). The proband of family 080 had been resuscitated from ventricular fibrillation after near‐drowning in a swimming‐pool. One of the compound heterozygotes in family 139 was symptomatic. In family 373, the patient carrying simultaneously the KCNQ1 IVS7‐2A>G and the HERG R176W mutation had frequently experienced cardiac palpitations, convulsions and loss of consciousness (Table I).

Patients carrying any of the four founder mutations alone had longer QTc than non‐carriers (Table II). Although the number of the compound heterozygotes is too low for exact statistical calculations, an attempt was made to compare their QTc data (Table I) to those obtained in patients carrying single mutations only (Table II). The HERG R176W mutation occurred in five KCNQ1 G589D carriers and one KCNQ1 IVS7‐2A>G carrier, with a mean QTc value of 473±30 and 486 ms (Table I); four (67%) out of six were symptomatic. Since not all of the compound heterozygote carriers were probands (Table I), their clinical findings were compared to all patients carrying either the KCNQ1 G589D or IVS7‐2A>G mutation (Table II). These patients had mean QTc values of 457±33 ms and 464±32 ms (Table II), and the corresponding symptomaticity rates were 30% and 25%, respectively 8. These data support the assumption that the HERG R176W mutation may exert an additional phenotypic effect when present along with a KCNQ1 mutation, but additional studies with extended patient collections are needed to verify this.

Table II. Carriers (all carriers and probands) versus non‐carriers of four common mutations identified in Finnish LQTS families. Only carriers and non‐carriers who did not use any medication at the time of ECG recording are included.

Gene	Mutation	Status	Mean age in years	M/F	QTc (ms)	P‐value^*
KCNQ1	FinA (G589D)	all carriers probands	36±19 (1–84)	163/208	457±33 (376–579)
			44±20 (2–84)	16/50	492±51 (415–691)
		non‐carriers	37±19 (1–91)	224/247	414±24 (349–488)	<0.0001
	FinB (IVS7‐2A>G)	all carriers probands	36±21 (1–88)	21/29	464±32 (398–540)
			37±19 (4–61)	5/8	482±32 (416–516)
		non‐carriers	26±19 (1–69)	19/19	411±20 (375–463)	<0.0001
HERG	FinA (L552S)	all carriers probands	32±18 (2–74)	20/42	459±40 (402–618)
			45±21 (3–77)	2/11	496±56 (441–618)
		non‐carriers	36±20 (2–81)	41/49	417±22 (365–476)	<0.0001
	FinB (R176W)	all carriers probands	34±21 (1–80)	23/39	448±30 (386–569)
			41±20 (9–79)	3/9	462±62 (366–569)
		non‐carriers	40±18 (2–74)	29/36	416±26 (369–477)	<0.0001

^* The P‐value given is for the groups all carriers versus non‐carriers

Functional studies of the KCNQ1 IVS7‐2A>G founder mutation

The IVS7‐2A>G mutation is proposed to alter the splicing acceptor site between exons 7 and 8 of the KCNQ1 gene and to result in insertion of 10 (or 11) amino acids into the C‐terminal part of the KCNQ1 protein (Figure 2C). To verify that the IVS‐2A>G acceptor site mutation indeed affects splicing, we performed RT‐PCR followed by PCR from two patient and three control RNA (isolated from peripheral lymphocytes) samples. As shown in Figure 2A, a 482 bp long fragment was produced in all samples corresponding to the wild‐type allele, but in addition a longer fragment was seen in the patients' samples only, thus representing the mutant allele. After subcloning of the PCR product several clones were sequenced. In the samples from the patients, the longer fragment showed an insertion of 30 bp (PCR product 512 bp, in 12 clones) or 33 bp (PCR product 515 bp, in one clone) corresponding to the sequence at the end of intron 7, causing an in‐frame addition of 10 or 11 amino acids (Figure 2B). We chose the cDNA construct corresponding to the 10 residues longer KCNQ1 protein product for in vitro electrophysiological studies.

Figure 2. A) Reverse transcription polymerase chain reaction (RT‐PCR) of lymphocytic RNA to characterize the KCNQ1 IVS7‐2A>G mutation. Lane 1: molecular weight marker; Lanes 2–4: RT‐PCR products from control blood samples; Lanes 5–6: RT‐PCR products from individuals carrying the KCNQ1 IVS7‐2A>G mutation. The size of the fragment corresponding to the wild‐type allele is 482 bp, while the corresponding sizes of the mutant alleles (as determined by DNA sequencing) are 512 and 515 bp. B) DNA and amino acid sequences of the in‐frame insertions of 30 or 33 bp (marked in bold) corresponding to the mutant allele in KCNQ1 IVS7‐2A>G carriers. The IVS7‐2A>G mutation at the end of intron 7 is indicated with the capital letter G. C) A schematic view of the KCNQ1 protein with the KCNQ1 IVS7‐2A>G (KCNQ1‐FinB) mutation located in the C‐terminal part.

The electrophysiological studies of the KCNQ1 IVS7‐2A>G mutation in COS‐7 cells, in the absence or presence of KCNE1, revealed a complete loss‐of‐channel function (Figure 3A, 3B). Upon co‐expressing the IVS7‐2A>G mutant with the wild‐type channel the mutant exerted a dominant‐negative effect on the whole channel complex. For example, at +40 mV, the current density for the combination of KCNQ1+KCNE1 was 26.7±5.5 pA/pF (n = 7), for KCNQ1 IVS7‐2A>G+KCNE1 2.2±0.4 pA/pF (n = 8), and for KCNQ1+KCNQ1 IVS7‐2A>G+KCNE1 7.44±2.2 pA/pF (n = 11) (Figure 3C).

Figure 3. A, B) Current density was obtained from whole‐cell current recordings of COS‐7 cells that were transfected with KCNQ1 (1.5 µg, n = 7), KCNQ1‐IVS7‐2A>G (1.5 µg, n = 8); KCNQ1 (0.75 µg)+KCNQ1 IVS7‐2A>G (0.75 µg each, n = 14), KCNQ1+KCNE1 (1.5 µg each, n = 7), KCNQ1 IVS/‐2A>G+KCNE1 (1.5 µg each, n = 8), or KCNQ1+KCNQ1 IVS/‐2A>G+KCNE1 (0.75 µg, 0.75 µg, and 1.5 µg respectively, n = 11). Cells were held to –80 mV then depolarized to various potentials ranging from –100 mV to+50 mV. C) Current density at +40 mV. Each value represents Mean ± SEM.

Functional expression of the HERG R176W mutation

The R176W mutation substitutes a basic arginine to a non‐polar tryptophan residue in the cytoplasmic N‐terminal region of the HERG channel (Figure 4A). The expressed current of the HERG R176W mutant was lower compared to the wild‐type (Figure 4B). Current density and tail current for HERG R176W represent only about 25% of the WT (Figure 4C, 4D). In general the main properties of the R176W current were very similar to the WT, however, with a slight acceleration of the deactivation kinetics for both the slow and the fast components (Figure 5). This gating alteration was probably not the only cause for the smaller current observed. Another possible reason is a reduced trafficking of the mutant channel to the membrane 23. Upon co‐expression with HERG‐WT, the R176W mutant did not display a dominant‐negative effect. Rather it behaved more or less as a WT subunit, being able to associate normally with the WT, as well as traffic normally to the membrane. Furthermore, no differences in current densities existed between HERG‐WT and HERG‐WT+HERG R176W channel complexes. However, the presence of the mutated subunit was evident as a slight acceleration of the deactivation kinetics upon co‐expressing HERG‐WT+HERG R176W (Figure 5).

Figure 4. A) A schematic view of the HERG channel protein, indicating the location of the HERG R176W (HERG‐FinB) mutation. B) Current recordings obtained from HERG‐ or HERG R176W‐transfected cells according to the pulse protocol shown. C) Comparison of current density obtained at the end of a 1‐second test pulse from cells transiently transfected with DNA encoding for HERG (1.5 µg, open circle, n = 17), HERG R176W (1.5 µg, solid circle, n = 23), and HERG+HERG R176W (0.75 µg each, solid square, n = 17). Currents were elicited by application of potentials ranging from −80 mV to +50 mV for 1 s, then depolarized to −40 mV for 1 s. D) Current‐voltage relationships were obtained from the same pulse protocol by plotting the maximum current obtained at −40 mV versus test potential, for HERG (n = 31), HERG R176W (n = 21), and HERG+HERG R176W (n = 19). Symbols are as in C.

Figure 5. A) Representative tail current traces for HERG and HERG R176W using the following pulse protocol: cells were held at −80 mV, depolarized to +40 mV for 200 ms, and a test pulse was applied at hyperpolarized potentials where HERG channels are silent (−120 mV and −100 mV) for 1 s. B) Traces were normalized to the maximum value and only the highlighted area in A is scaled and shown. C, D) Fast and slow deactivation time constants respectively for HERG (open circle, n = 17), HERG R176W (solid circle, n = 16), and HERG+HERG R176W (solid square, n = 36). * denotes statistical significance (P<0.05) using ANOVA.

Functional data for the SCN5A R190G variant

As our previous study suggested a slight overrepresentation of the SCN5A R190G variant in LQTS patients compared to controls 8, its in vitro electrophysiological characteristics were also examined. None of the R190G sodium current biophysical properties tested were significantly different from WT (data not shown). In particular, since in most cases prolongation of the cardiac action potential has been shown to be caused by a small tetrodotoxin‐sensitive persistent current Iss generated by LQT3 mutant channels 24, we recorded this current and did not find any difference between WT and R190G Iss. Altogether, these findings suggest that the R190G SCN5A variant represents a polymorphism without a pathophysiological significance.

Discussion

We here describe the electrophysiologic characteristics of the products of two mutant potassium channel genes, KCNQ1 IVS7‐2A>G and HERG R176W, that both are relatively prevalent among Finnish probands with LQTS. Our previous study using a clinical and molecularly defined group of unrelated patients showed prevalences of 7.3% for KCNQ1 IVS7‐2A>G and 8.0% for HERG R176W in LQTS 8. These figures are significantly higher than the corresponding carrier frequencies among blood donors, i.e. 0/200 and 3/317 (0.9%), respectively. Their role as disease‐causing mutations is further substantiated by data from in vitro electrophysiological studies that show an altered channel function in both cases. In contrast, the R190G SCN5A variant channel did not show any significant electrophysiological alterations in vitro.

The KCNQ1 IVS7‐2A>G mutation abolishes a splice acceptor site between exons 7 and 8, and appears to cause an in‐frame addition of 30 bases to the mRNA, corresponding to the sequence at the end of intron 7 (Figure 2B). To find out whether the predicted insertion of 10 amino acids (Val‐Thr‐Ala‐Cys‐Pro‐Pro‐Ala‐Arg‐Pro‐Arg) in the C‐terminal part of the KCNQ1 protein would affect its function, we performed electrophysiological studies. Indeed, upon expressing the IVS7‐2A>G mutant in vitro no current was present, and it exerted a dominant‐negative effect when co‐expressed with the wild‐type (Figure 3). The in vitro results are consistent with the clinical findings. Until now, the KCNQ1 IVS7‐2A>G has been identified in 13 families and altogether 75 patients, and is associated with significant QTc prolongation (Table II). Approximately 25% of its carriers were symptomatic 8. In addition to the KCNQ1 IVS7‐2A>G mutation, a few other mutations affecting mRNA splicing have been identified in the KCNQ1 gene 25–29. Mutations at codon 344 altering the last base of exon 6 in the KCNQ1 gene have been identified in several LQTS families, and are considered to represent a mutation hot spot 28. The codon 344 mutation was studied with RT‐PCR, which revealed exon skipping in the mutation carriers 28.

The HERG R176W mutation has been identified in 16 Finnish LQTS families and 92 carriers by now. This mutation seems to prolong the QTc interval as assessed in patients receiving no drugs: the QTc interval in 62 mutation carriers was 448±30 ms versus 416±26 ms in 65 non‐carriers (P<0.0001, Table II). About 35% of a total of 92 HERG R176W carriers have suffered from symptoms. Recently, Ackerman et al. 30 identified the HERG R176W mutation in one (0.5%) out of 187 apparently healthy white subjects but none in 305 black individuals. We failed to detect this mutation in our previous sample of 120 Finnish blood donors 18; however, while now extending this population of blood donors to 317 samples we identified the HERG R176W mutation in three (0.9%) out of the 317 subjects studied. It will be of major interest to study whether the HERG R176W mutation is present in other populations and whether it tends to cluster among LQTS patients.

The R176W mutation resides in close vicinity of the PAS (Per‐Arnt‐Sim) domain in the N‐terminus of the HERG channel, corresponding to its first 135 amino acids. This domain is highly conserved among species and is known to be important in protein‐protein interactions in sensing and signal transduction pathways 31,32. Removal of the N‐terminus, including the PAS domain, markedly increases the rate of HERG channel deactivation probably by impaired interaction with the S4‐S5 linker region 33. Many mutations in the PAS domain in HERG have been described, and all of them affect gating by accelerating channel deactivation 32,34. Deactivation or return to the closed state is normally very slow in the HERG channel, and it can be observed as a large inward tail K⁺ current during voltage clamp experiments 35. Interestingly, also the HERG R176W mutation caused acceleration of channel deactivation (Figure 5) like the mutations in the PAS domain. Additionally, reduced current density of the R176W was observed, which could be due to abnormal trafficking of the mutant subunits to the membrane; however, this defect was corrected upon co‐expression with the wild‐type. In conclusion, it most likely that the prolonged QT interval seen in R176W carriers is due to reduced outward current during the repolarization phase of the action potential caused by altered gating of deactivation.

When using strict diagnostic criteria (QTc⩾450 ms or 460 ms for males and females, respectively) for LQTS, 252 of our 700 probands had LQTS, and 4 (1.6%) of these 252 probands were compound heterozygotes for two different founder mutations, with HERG R176W always as the second component of the pair. Our data provides some evidence that the R176W mutation may confer an additional QTc‐prolonging and symptom‐provoking risk for these types of patients (Table I, Table II), but this assumption requires studies in additional patients, if identified in the future.

Recently, Westenskow et al. 16 screened 252 LQTS probands for mutations in the KCNQ1, KCNH2, KCNE1, KCNE2 and SCN5A genes, and found two variants simultaneously in 20 (7.9%) probands. In 10 out of these 20 probands one of the variants detected was the KCNE1 D85N polymorphism, and it was also identified in 4.7% of control individuals. In other recent studies, 1.1% to 3% of Caucasians were reported to be heterozygotes for the same KCNE1 D85N polymorphism 30,36. Furthermore, according to expression studies the KCNE1 D85N polymorphism alone causes reduction of the I_Ks current 16.

The HERG K897T polymorphism is of special interest as it has tentatively been shown to correlate with the QT interval 18,37–41. Further, this polymorphism has been associated with changes in HERG channel activation, deactivation and inactivation kinetics 38,39,42, but does not seem to associate with an increased risk for blockade by drug or drug‐induced LQTS 3,42,43. However, the results from these in vivo and in vitro studies provide inconsistent data and therefore do not permit definitive conclusions on the functional importance of the K897T polymorphism, although they imply that it might act to modify the phenotype in LQTS.

Relatively rare variants of the cardiac ion channel genes, such as the KCNE1 D85N, HERG K897T and SCN5A S1102Y 24, as well as the HERG R176W mutation characterized in detail in the present study, may cause form fruste—types of LQTS that may attain a symptomatic status when notorious extrinsic factors come into the play, or may exert additional phenotypic effects upon those of ‘classical’ LQTS genes in the compound heterozygous individuals.

There are limitations of the present study. First, no health data or ECG recordings were available from the 317 blood donors, whose DNA samples served as a reference for allele frequency calculations. Thus, it is possible that asymptomatic LQTS patients were included among this group. Second, the number of compound heterozygotes was too small for making definitive conclusions on the phenotype of the patients.

In conclusion, we have demonstrated that the KCNQ1 IVS7‐2A>G and HERG R176W mutations represent a common cause of LQTS in Finland and are associated with significant defects of channel function in vitro. The HERG R176W may exert weaker phenotypic effects than most other LQTS‐causing mutations. Most interestingly, it is present among blood donors with a frequency of about 1%. Further studies are needed to clarify whether its carrier status, present alone or combined to another QT interval‐prolonging gene, signals an increased risk of life‐threatening arrhythmias at the population level.

Acknowledgements

The authors wish to thank Dr Kirsi Paukku, Ms Saara Nyqvist, Tuula Soppela, Susanna Tverin and Eija Hämäläinen for excellent technical help. This work was supported by grants from The Academy of Finland, The Special State Share (EVO) for the Helsinki University Hospital, The Sigrid Juselius Foundation, The Finnish Foundation for Cardiovascular Research, the Centre National de la Recherche Scientifique, and the Association Française contre les myopathies (JB, SB), the Swiss National Science Foundation (632‐66149.01 to HA).