INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

HEREDITARY LONG QT SYNDROME (LQTS) is a heterogeneous genetic disease that predisposes affected individuals to life-threatening arrhythmias (25, 26). The disease is caused by mutations in five known genes, of which four encode potassium channels (1, 19, 19a, 19b). The second most common form, LQT2, is linked to mutations in the cardiac potassium channel HERG on human chromosome 7 (22). Mutations in HERG have been described in both cytoplasmic termini (2, 5) as well as transmembrane domains (e.g., Refs. 10, 21, 30) and result in three classes of defects: 1) channels with abnormal gating and reduced currents (21, 30), 2) channels that reach the plasma membrane but do not conduct current (30), and 3) channels that are retained in the endoplasmic reticulum (ER) (10, 30, 31). Mutations in classes 1 and 2 reduce expression levels even more if there is dominant-negative suppression of functional channels produced by the wild-type (WT) allele (21). Some mutations in class 3 are hypomorphic, i.e., they show severely impaired trafficking where the majority of channel protein is retained in the ER and few channels leak out to the plasma membrane producing small currents at 37°C (10, 31).

Mutations defective in trafficking have been reported for other membrane proteins including minK (3), P-glycoprotein (14), aquaporins (27) and the cystic fibrosis transmembrane conductance regulator (CFTR) (11, 12). Loss of function associated with trafficking mutations in CFTR (8), P-glycoprotein (14), and aquaporins (27) may be rescued by synthesis at reduced temperature (8, 14, 31), by overexpression (6), by chemical chaperones such as glycerol and dimethylsulfoxide (4, 24, 27), or by specific drugs binding to misprocessed membrane proteins (15, 31). For the hypomorphic HERG mutation N470D, it was shown more recently that trafficking can be improved not only by incubation at low temperature but also in the presence of osmolytes or the specific channel ligand E-4031 (31).

In a large LQT2 family we found a novel COOH-terminal missense mutation, HERG R752W, which was retained in the ER. The mutant protein was "rescued" and delivered to the plasma membrane after incubation at reduced temperatures in mammalian cells or after expression in Xenopus oocytes. However, trafficking could not be restored by incubation with chemical chaperones or E-4031, a specific blocker of HERG channels. When trafficking was corrected, HERG R752W channels conducted enhanced currents that slightly shortened the cardiac action potential simulated by the Luo-Rudy model of a mammalian ventricular myocyte.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Molecular biology and electrophysiology. HERG R752W was prepared by overlap extension PCR, verified by sequencing, and subcloned as a Xho I-BamH I fragment into either full-length HERG-pSP64 or HERG-pCDNA3. In vitro RNA synthesis and oocyte injections were performed as described (9). Production and size of full-length cRNA was verified by denaturing agarose gel electrophoresis, and cRNA concentrations were determined with Ribogreen RNA quantitation kit (Molecular Probes). In Xenopus oocytes, currents were recorded 2-7 days after injection using a Dagan 8500 two-electrode voltage-clamp amplifier. Current and voltage electrodes were filled with 3 M KCl and had resistances of ~1 M. The composition of the bath solution used to perfuse oocytes was (in mM) 96 NaCl, 5 KCl, 1.8 CaCl₂, 1.0 MgCl₂, and 5 HEPES (K⁺ Ringer solution, pH 7.4). Cell-attached macropatch recordings in oocytes were performed as described previously (9). Patch pipettes (resistance, 0.2-0.6 M) were filled with K⁺ Ringer solution, and the depolarizing bath solution had the following composition (in mM): 100 KCl, 10 EGTA, 1 EDTA, and 10 HEPES (pH 7.4). For electrophysiological experiments in HEK293 cells, transient transfections were carried out with the Ca²⁺ phosphate method. After transfection, cells were incubated for 48 h at either 26 or 37°C. For visual identification, HEK cells were cotransfected with CD8 antigen (1:5 ratio) and incubated before recordings with beads coated with CD8 antibodies (Dynal). Patch pipettes were filled with (in mM) 100 K-aspartate, 20 KCl, 2 MgCl₂, 1 CaCl₂, 10 EGTA, and 10 HEPES (pH 7.2). The extracellular solution had the following composition (in mM): 140 NaCl, 5 KCl, 1 MgCl_2, 1.8 CaCl₂, 10 HEPES, and 10 glucose (pH 7.4). No leak subtraction was applied. All current recordings were performed at room temperature (20-22°C). pClamp software (Axon Instruments) was used for the generation of voltage-clamp protocols and data acquisition. Data are presented as means ± SE of n experiments.

Simulations. The simulations in this study were conducted using the Luo-Rudy model, a theoretical dynamic model of a mammalian ventricular action potential. The model includes membrane ionic channel currents that are formulated mathematically using the Hodgkin-Huxley approach as well as ionic pumps and exchangers. It also includes processes that regulate intracellular concentration changes of Na⁺, K⁺, and Ca²⁺ (7, 16, 29). For the purpose of this study, the formulation of the fast delayed rectifier potassium current, I_Kr, was modified to fit our experimental voltage clamp data obtained from HERG-WT and HERG-R752W channels. All kinetic parameters were normalized to 37°C with a Q₁₀ of 3. The basic protocol for action potential simulations involved stimulation of the model cell 10 times at a constant cycle length (CL) of 1,000 ms.

Antibodies. The anti-HERG polyclonal antibody used in Western blots and for immunostaining was generated in rabbits against a glutathione-S-transferase (GST) fusion protein containing the last 112 amino acids of HERG (residues 1048-1159). HERG antiserum was either purified on an affinity column consisting of a short COOH-terminal peptide corresponding to HERG residues 1102-1121 (TLTLDSLSQVSQFMACEELP) or the entire fusion protein.

Western blot analysis and immunostaining. For Western blotting, HEK 293 cells were transfected with Lipofectamine/Plus as recommended by manufacturer (Life Technologies). Forty-eight hours after transfection, cells were solubilized for 1 h at 4°C in lysis buffer [50 mM Tris · HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100], containing protease inhibitor mix (Complete, Roche Biochemicals). To isolate a crude membrane fraction, oocytes were homogenized in 0.3 M sucrose and 10 mM NaPO₄ (pH 7.4) supplemented with protease inhibitor mix (Complete). After removal of debris and nuclei (3,000 g, 10 min), the supernatant was spun at 50,000 g for 1 h at 4°C to pellet membranes. Protein concentrations were determined by the bicinchoninic acid method (Pierce). For Western blotting, proteins were separated on 8% SDS polyacrylamide gels and transferred to polyvinylidene difluoride membranes. Membranes were blocked overnight with 5% nonfat dry milk in PBS plus 0.1% Tween and immunoblotted with polyclonal rabbit anti-HERG antibody (1:200 dilution; 1 h at RT) followed by horseradish peroxidase-conjugated secondary antibody (1:3,000; 1 h at RT; Amersham Pharmacia Biotech). Western blots were developed with the ECL-Plus detection system (Amersham Pharmacia). For immunostaining, transiently transfected COS cells were grown on coverslips and fixed with 4% paraformaldehyde in PBS 48 h after transfection. After fixing, cells were permeabilized with 0.1% Triton X-100 in PBS, blocked with 5% nonfat dry milk in PBS, and stained at 4°C overnight with anti-HERG antibody (1:100). The tetramethylrhodamine isothiocyanate-conjugated secondary antibody (1:100; Jackson Labs) was added for 2 h at room temperature. For observation, coverslips were mounted on slides with Vectashield (Vector Labs).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

HERG R752W was identified in a 31-yr-old woman with a history of syncope who died suddenly. Twenty-six of 101 genotyped family members were found to be carriers. Of these, 17 had prolonged QT intervals (QTc > 0.44ms) and 6 had a history of syncope.

Characterization of HERG R752W currents in Xenopus oocytes. The cellular phenotype of LQT mutations is often studied by measuring currents expressed in Xenopus oocytes. After injection of equivalent amounts of cRNA, we were surprised to find that mutant HERG R752W channels produced larger outward currents than WT channels (Fig. 1, A and B). Both WT and R752W produced typical bell-shaped isochronal current-voltage (I-V) relationships (Fig. 1C). For R752W, the amplitudes were larger and the I-V curve was shifted to more hyperpolarized potentials. To evaluate the shift, maximum tail currents elicited at 90 mV were used to calculate a correction factor to account for small differences in expression levels (Fig. 1D). Compared with the corrected WT, I-V curve threshold and peak for R752W remained shifted (Fig. 1C), suggesting that altered gating properties contributed importantly to the larger currents of R752W. Coinjection of equal amounts of WT and R752W cRNA produced a summation of current amplitudes at potentials more positive than 20 mV. At more hyperpolarized potentials, current amplitudes from the combined injection were not significantly changed from R752W alone despite the increased amount of cRNA injected. One explanation might be that below 20 mV, activation rates for R752W/WT heteromultimers are very different from homomultimeric channels (see Fig. 2C), thereby producing a complex interaction with current amplitudes as measured in Fig. 1C. At more positive potentials, however, activation rates converge, and consequently current amplitudes increase as expected for coinjection experiments. Nonetheless, our results show clearly that mutant channels do not suppress WT currents in a dominant-negative manner.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   HERG R752W is located in the COOH-terminal linker between transmembrane domain S6 and the cyclic-nucleotide binding fold common to all members of the ether-a-go-go gene family with the arginine at this position being highly conserved. Macroscopic current recordings from Xenopus oocytes injected with 0.5 ng each of HERG wild-type (WT; A) or HERG R752W cRNA (B; RW). Inset: voltage protocol; holding potential, 85 mV, test pulses, from 100 to +60 with 10-mV increments; return potential 90 mV. C: isochronal current-voltage (I-V) relationship for HERG WT (0.5 ng cRNA, mean values from 5 oocytes, ), HERG RW (0.5 ng, n = 6, ), and 1:1 coinjection HERG WT/HERG RW (0.5 + 0.5 ng, n = 5, ). Currents were elicited with voltage protocol shown in inset to A. Amplitudes were analyzed at the end of the depolarizing 650-ms voltage commands. A single batch of oocytes was used for C. All recordings were made 2 days after injection. D: maximal tail current amplitudes elicited on return to 90 mV after a test pulse to 60 mV were back-extrapolated and used to assess the average number of functional channels expressed in oocytes injected with WT or R752W cRNA. The fractional difference in tail current expression between HERG WT and RW was used to obtain the normalized WTcor I-V relationship in (C, ). E and F: steady-state activation and inactivation properties of HERG WT (, ) and HERG RW channels (, ). Steady-state activation was analyzed in cell-attached macropatches by plotting tail current amplitudes at 110 mV against 30-s depolarizing test pulses between 80 and +40 mV. Steady-state inactivation was assessed with brief hyperpolarizing steps from +20 mV (E); plotted are normalized peak currents on return to +20 mV. Dashed line indicates zero current level. Data were fitted to Boltzmann equations of the form: y = 1/{1+exp[(VhV)/k]}. Half-maximal inactivation Vk was 63.3 ± 1.1 mV for WT and 55.5 ± 1.6 mV for HERG R752W (n = 4). Half-maximal activation Vh was 49.6 ± 0.7 mV (n = 4) and 37.7 ± 0. 2 mV (n = 7) for HERG WT and HERG RW, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   HERG R752W accelerates current activation (A). Activation time constants were measured using an envelope of tails protocol to separate current activation from inactivation. A test pulse of variable duration (in this case to 0 mV) was given to open and/or inactivate channels before stepping back to 100 mV. B: peak tail current amplitudes at 100 mV were back extrapolated for time 0 and plotted against test pulse duration. The resulting activation time course could be fitted to a single exponential function. C: activation time constants plotted for different voltages (n = 8-10). Note that activation time constants are significantly different between HERG WT () and HERG RW () even at +20 mV [WT, 105.3 ± 8.1 ms (n = 8); RW, 67.9 ± 3.7 ms (n = 10)] and at +40 mV [WT, 71 ± 4.8 ms (n = 8); RW, 45.8 ± 2.9 ms (n = 10)]. HERG R752W speeds channel deactivation (D). Currents were activated during a 750-ms depolarizing test pulse to 20 mV before stepping back to potentials ranging from 40 to 120 mV to measure current deactivation. Deactivation kinetics were fitted to double exponential functions. Fast (E, n = 16-19) and slow deactivation time constants (F, n = 5 or 6) are shown as a function of membrane potential. , Time constants measured on coinjection of equal amounts of WT and R752W cRNA.

Simulations of the effects of WT HERG R752W current on the cardiac action potential. Because the kinetic differences were ambiguous with respect to changes of the QT interval, the effects were simulated with the Luo-Rudy model of the cardiac ventricular action potential (7, 16, 29). The mathematical formulation of the rapid delayed rectifier, I_Kr, reproduced the kinetic data of Figs. 1 and 2 for WT and mutant channels. Figure 3 shows action potentials and I_Kr currents computed from WT and mutant cells. Due to faster activation kinetics, mutant I_Kr was larger than WT I_Kr and produced a minor shortening of action potential duration.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Action potentials (A) and corresponding delayed rectifier current IKr (B) simulated using the Luo-Rudy model. Thick lines correspond to WT cells; thin lines correspond to mutant cells. A: the last AP from a train of 10 action potentials. Because of the faster activation kinetics of HERG RW, IKr reaches a higher magnitude, thereby causing minor shortening of the mutant cell action potential duration. However, quantitatively the action potential duration change is negligible and should not affect electrophysiological function.

Reconciling mutant HERG currents with QT prolongation. The discordance between the cellular phenotype in Xenopus oocytes and the prolonged QT interval observed clinically recalled the temperature sensitivity of the CFTR F508 mutant (8). Therefore, we expressed HERG R752W in mammalian HEK293 cells at a physiological temperature of 37°C and used whole cell patch clamp recordings to characterize mutant currents. We were not able to detect any HERG-like outward current in HEK293 cells transfected with R752W, although WT current was obtained routinely under identical experimental conditions (Fig. 4, A and B). Because mutant currents were elicited in Xenopus oocytes and because oocytes are maintained at lower temperatures than mammalian cells, we tested the effect of reduced incubation temperature on expression of R752W. After incubation for 2 days at 26°C, large whole cell HERG currents were recorded in cells transfected with R752W cDNA and had an electrophysiological profile similar to that described in Xenopus oocytes (Fig. 4, C and D).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Processing of HERG RW is temperature sensitive. Current recordings from HEK293 cells transfected with either HERG WT (A) or HERG RW cDNA (B). Inset: voltage protocol; holding potential, 80 mV, test pulses, from 120 to +60 mV, return potential 100 mV. After transfection, cells were grown for 2 days at 37°C before recording. Current recordings from WT (C) and RW channels (D) after incubation for 2 days at 26°C. E: isochronal I-V relationship for HERG WT () and HERG RW () expressed in HEK293 cells at 26°C. Amplitudes are analyzed at the end of depolarizing voltage commands shown in C and D and normalized to maximal outward current amplitudes. F: activation properties of HERG WT () and HERG RW () analyzed at the end of depolarizing 650-ms voltage commands. Data were fitted to Boltzmann equations of the form: y = 1/{1 + exp[(Vh  V)/k]}. More negative half-maximal activation of HERG RW (4.5 vs. 3.8 mV for HERG WT) reflects faster current activation of HERG R752W. G: effect of incubation temperature (WT37, RW37, WT26, RW26) and cotransfection (coWT37, WT/RW37) on expression of HERG currents. To evaluate effects of incubation temperature on current expression, HEK293 cells were transfected with either 2 µg HERG WT or 2 µg HERG R752W cDNA and incubated for 2 days at 26 or 37°C as indicated. In coexpression experiments, HEK cells were transfected with either 1 µg HERG WT + 1 µg vector cDNA (coWT37) or 1 µg HERG WT + 1 µg HERG R752W (WT/RW37) cDNA and incubated for 2 days at 37°C. Current densities were calculated from tail currents elicited on return to 100 mV after a test pulse to 60 mV. Note that in coexpression experiments current densities measured in cells transfected with WT cDNA alone proved not to be different from cells expressing equivalent amounts of WT and HERG RW cDNA (Student's t-test, P > 0.05). Therefore, dominant-negative suppression of HERG WT currents by HERG RW at 37°C is unlikely.

View larger version (56K): [in this window] [in a new window]   Fig. 5.   Immunostaining of COS-7 cells transiently transfected with WT (A) and HERG RW mutant channels (B). After transfection, cells were incubated for 2 days at 37°C. Transfected cells were immunostained with anti-HERG antibody showing strong perinuclear staining of HERG RW reflecting endoplasmic reticulum localization. In contrast, the entire cytoplasm was stained in WT cells. Calibration bar is 10 µm. C: Western blot analysis of HERG RW and HERG WT channels. Left shows the effect of temperature, glycerol, 4-phenylbutyrate, and E-4031 on the expression and maturation of WT and HERG RW. Cells were grown for 2 days either at 37 or 26°C (WT37, 26; RW37, 26). Glycerol treated mutant cells (RW7.5G, 10G and 12.5G) were incubated for 48 h at 37°C in media supplemented with 7.5, 10, and 12.5% glycerol. 4-Phenylbutyrate (2.5 mM) or 5 µM E-4031 was added for 48 h at 37°C to cells transfected with HERG RW (RW4PB2.5, RWE4031). Right shows membrane proteins isolated from Xenopus oocytes injected with WT or HERG RW. Control lane represents uninjected oocytes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The electrophysiological phenotype of the processed R752W mutant is characterized by mild changes in gating properties that do not translate into prolongation of the cardiac action potential and hence the QT interval as judged by our simulations. The accelerated activation and deactivation kinetics of R752W resembles mutations described previously in the NH₂-terminal PAS (Per, Arnt, and Sim) domain of the channel protein, especially HERG R56Q (5). In prokaryotes, PAS domains serve as sensors for light and redox potential (32) and in HERG channels it is thought that the PAS domain is attached to the intracellular mouth of the channel protein, thereby regulating channel gating (17, 28). Accordingly, structure-function studies suggested that mutations in the PAS domain disrupted interactions with amino acid residues in the cytoplasmic S4-S5 linker (23). We speculate that R752 may interact in a similar manner with the activation machinery of HERG or, alternatively, may stabilize the interaction of the PAS domain with the S4-S5 linker.

Our electrophysiological studies show that mutant channels function with slightly altered kinetics. The Luo-Rudy simulations suggest that these channels generate a normal action potential once they reach the plasma membrane, indicating that the clinically observed prolonged QT interval does not reflect altered channel kinetics. At physiological temperatures, however, mutant channels cannot pass quality control processes of the ER and do not reach the plasma membrane. At reduced temperatures where protein folding might be altered, HERG R752W attained a conformation that ultimately permitted exit from the ER. The molecular mechanism responsible for this conformational stabilization is not understood and remains to be analyzed in the future.

The coexpression experiments suggest that at 37°C, mutant channels are arrested in conformations that do not allow for the assembly of HERG tetramers, because neither dominant-negative current suppression nor a significant increase in current amplitudes has been observed. This interpretation is in line with previous reports that the COOH terminus is critical for HERG assembly and expression (13).

We explored whether chemical chaperones known to stabilize temperature-sensitive mutants were effective in correcting the trafficking defect associated with HERG R752W. Glycerol has been described as the most effective chemical chaperone to correct misfolding of CFTR, P-glycoprotein, or aquaporin mutant channels (14, 24, 27), and at molar concentrations we found no increase in the delivery of mutant protein to the cell membrane. In contrast, glycerol seemed to decrease the overall synthesis of mutant protein. An explanation might be that the efficacy of chemical chaperones not only depends on the particular cell line used for heterologous expression but also on the protein domain affected by a mutation and on the chemical nature of the amino acid side chain introduced (24). In CFTR it was shown that glycerol promoted the maturation of mutants K464R and K464A but not of K464W, a mutation with an amino acid substitution similar to the one analyzed in HERG R752W.

The failure of glycerol treatment to facilitate trafficking of HERG R752W was surprising because trafficking of HERG N470D, a hypomorphic mutation in the transmembrane domain of HERG, was improved by incubation in 10% glycerol (31). Similarly, the antiarrhythmic drug E-4031 facilitated trafficking of HERG N470D, but was ineffective on the R752W mutant. These differences suggest that HERG R752W represents a new and different class of misprocessed mutants with distinct mechanisms responsible for retention of mutations localized to different functional domains of the HERG channel protein. This observation is important because it makes a unifying approach for altering HERG trafficking unlikely. On the contrary, functionally different classes of mutations will have to be targeted separately in any attempt to discover novel and useful chaperones for misprocessed channel proteins.

Because chemical chaperones were not effective in HERG R752W, we complemented our experiments by treating HEK cells expressing HERG R752W with 4-PB, an analog of the transcriptional regulator butyrate. In analogy to experiments in CFTR, we tried to boost expression of HERG R752W to force the mutant by mass action past the ER to the plasma membrane (20). Although synthesis of HERG R752W was substantially enhanced, we were not able to detect fully glycosylated mature protein. This might indicate that at 37°C, only a small portion, if any, of the mutant protein was exportable and that the quality control system was not saturated.

In summary, HERG R752W is the first member of a novel class of temperature-sensitive trafficking mutants that do not express any current at 37°C and are resistant to rescue by osmolytes or specific channel ligands. The clinically observed prolonged QT interval is explained best by a reduction of the delayed rectifier current I_Kr caused by the inability of the mutant protein to reach the plasma membrane at body temperature. Thus, although HERG R752W behaves as a loss-of-function mutation, there is one major difference from hLQT mutations reported heretofore: HERG R752W encodes a fully functional channel protein if it can be made to reach the plasma membrane. For this reason, further efforts should aim at discovering novel and useful chemical chaperones.