INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

MUTATIONS IN GENES coding for potassium and sodium channels that cause a prolongation of the QT interval, which may lead to fatal ventricular arrhythmias, were recently identified (15). Abnormalities in ion channel expression were also reported in congestive heart failure and atrial fibrillation (6). These abnormalities can cause marked spatial and temporal dispersion of repolarization and may lead to sustained reentrant arrhythmias that can then deteriorate to ventricular fibrillation (2).

Recent advances in mice genetically modified at various potassium channels have fostered a better understanding of the electrophysiological roles of ionic currents in this species. Functional knockout of the rapidly activating (I_Kr) or slowly activating (I_Ks) components of the delayed rectifier potassium current, the two important repolarization currents responsible for the congenital long QT syndrome, did not lead to either QT prolongation or arrhythmogenic substrate (3, 13). In contrast, attenuation of Kv1.5-, Kv4-, or Kv2-encoded currents in mice with a dominant-negative (DN) approach resulted in prolongation of the action potential duration (APD) in vitro and the QT interval of the electrocardiogram in vivo (5, 14, 16). These results confirmed the substantial difference in ventricular ion channel expression in adult mice from that in humans. Previous studies examined the electrophysiological properties of Kv1.5- and Kv4-like currents in ventricular myocytes isolated from the adult mouse heart, which demonstrated two distinct outward potassium currents, i.e., a rapidly activating, slowly inactivating current (I_K,slow1) and a fast component transient outward current (I_to,_f) (5, 10, 18). However, the properties of the Kv2-like current have not been illustrated.

The transgenic mouse model (Kv1DN) of long QT syndrome was created by overexpression of the NH₂ terminus and the first transmembrane segment of Kv1.1 in the heart (Kv1.1N206Tag) (14). This truncated channel forms homo- and heteromultimeric complexes in vitro and coassembles with wild-type Kv1.x channels to form nonfunctional complexes that are trapped in the endoplasmic reticulum (9). Mice overexpressing Kv1.1N206Tag in the heart are characterized by a prolonged QT interval and a spontaneous monomorphic ventricular tachycardia (VT) (14). Transvenous programmed electrical stimulation induced polymorphic VT in 50% of these mice (12). Cardiac myocytes derived from these mice had a prolonged APD caused by the loss of a rapidly activating, slowly inactivating current, I_K,slow1, which correlated with a marked decrease in the level of Kv1.5 polypeptide (14, 18). Programmed electrical stimulation at the apex, but not at the base, of Kv1DN hearts perfused on a Langendorff system resulted in a long-lasting VT (4). Optical mapping experiments revealed substantially longer effective refractory periods (ERPs) at the base of Kv1DN left ventricles, which resulted in increased spatial dispersion of repolarization and reentrant arrhythmia (4). Interestingly, none of the other mouse models, in which the I_Kr (3), I_Ks (13), Kv4 (5), or Kv2 (16) channel was functionally knocked out, was reported to have proarrhythmic activity. Therefore, uncovering the ionic basis of these discrepancies may be very helpful for a better understanding of how these channels are functionally integrated to maintain the electrical stability of the heart.

In the present study, we characterize an electrical remodeling response of the cardiocytes derived from the Kv1DN mouse hearts. A current component (I_K,slow2) is selectively upregulated in the myocytes isolated from the apex of the left ventricle but not in those from the base. This spatially restricted electrical remodeling may underlie the enhanced dispersion of repolarization observed in Kv1DN hearts and the spontaneous and inducible arrhythmias in these animals. The upregulation of Kv2.1 transcripts in Kv1DN mouse hearts revealed by Northern blot and RT-PCR and the elimination of I_K,slow2 in cardiocytes by crossbreeding of Kv1DN with Kv2DN mice suggest that Kv2.1 underlies the inducible component, I_K,slow2.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolation of adult mouse ventricular myocytes. Mouse ventricular myocytes were obtained as described previously (14, 18). Briefly, control and Kv1DN adult FVB mouse (3-6 mo) hearts were perfused retrogradely with a Langendorff apparatus. A 5-min perfusion with oxygenated, nominally calcium-free solution A was followed by an 8- to 12-min perfusion with solution B. The ventricles were then chopped into small pieces, incubated at 37°C in solution B supplemented with 0.02 mg/ml protease (type XXIV; Sigma) for 5-10 min, and mechanically dispersed. Myocytes were obtained by repeated washing with a series of centrifugations at 500 rpm for 2-3 min and resuspension in 0.25, 0.5, and 1 mM calcium-containing solution A supplemented with 2% FCS. Calcium-tolerant, rod-shaped ventricular myocytes with clear striations were selected randomly for electrophysiological studies at room temperature. For the regional ion channel expression experiments, the apical and basal segments of the left ventricle free wall were obtained by cutting the left ventricle free wall, after removing the septum, horizontally into three parts and discarding the middle segment. We then proceeded as described above.

Electrophysiological study. A whole cell patch-clamp technique (11) was used to investigate changes in the ionic currents and action potentials of the myocytes isolated from control and Kv1DN mouse ventricles. In brief, myocytes were superfused with Tyrode solution at 1-2 ml/min, and the electrophysiological data were obtained with electrodes having a resistance of 0.5-2 M when filled with a standard pipette solution. The electrode was connected to an Axopatch 200A amplifier (Axon Instruments), and the signals were digitized through a DigiData 1200 interface and stored in the computer. pCLAMP 6.0.4 software (Axon Instruments) was used to generate command pulses and acquire data. After the membrane was ruptured, the cell capacitance was evaluated (18). In this study, cell capacitances in the control and Kv1DN myocytes were 160.4 ± 68 pF (n = 34) and 172.4 ± 48.1 pF (n = 39), respectively (P > 0.05). To obtain correct voltage clamp and minimize the voltage error that might result from the series resistance, a technical adjustment similar to that in a previous study (18) was made. Recordings were started after 5 min of membrane rupture at room temperature (22-24°C). Linear "leakage" currents were corrected off-line only when the input resistances were 1 (but 0.3) G (n = 5). Junction potentials (~10 mV) resulting from the low-chloride pipette solution were corrected off-line.

Solutions. Solution A contained (in mM) 137 NaCl, 4.7 KCl, 1.2 MgCl₂, 1.0 KH₂PO₄, 11 glucose, 10 HEPES, and 0.125 Na₂EDTA, pH 7.35. Solution B was made from solution A with 0.5 mg/ml collagenase (type I; Worthington), 0.2 mg/ml hyaluronidase (type II; Sigma), and 1% FCS added. Tyrode solution contained (in mM) 137 NaCl, 5.4 KCl, 1.2 MgCl₂, 1 CaCl₂, 10 glucose, and 10 HEPES, pH 7.35 with NaOH. CoCl₂ was added (2 mM) when potassium currents were being recorded. For the pharmacological experiment involving tetraethylammonium (TEA), extracellular KCl was substituted with equimolar TEA · Cl when the concentration of TEA was 2.5 mM to maintain a constant ionic strength. Other blocking reagents were added directly into Tyrode solution. Standard pipette solution was composed of (in mM) 110 K-aspartate, 20 KCl, 1 MgCl₂, 0.5 CaCl₂, 10 HEPES, 5 EGTA, 5 Mg₂ATP, 5 Na-creatine phosphate, and 0.5 GTP-Tris, pH 7.2 with KOH (pCa 7.6). For calcium current (I_Ca) recording, potassium in both the Tyrode and pipette solutions was substituted with equimolar cesium.

Data analysis. Electrophysiological data were analyzed with Clampfit 6.0, Microsoft Excel, and Microcal Origin 5.0. The current inactivation kinetics was described by a sum of exponentials with the following function: y(t) = A_i · exp(t/_i) + C, where A_i and _i represent the amplitude(s) and time constant(s), respectively, for the inactivating current component(s), and C is the offset, which includes the remaining portion of those components, other noninactivating currents, and the leakage current. Goodness of fit was judged by visual inspection and the correlation coefficients (R), which were normally 0.975 in this study. The number of exponentials was determined by F-test (18), and P < 0.05 was taken to indicate a better fit. Concentration effects were quantified by fitting the data with a Hill equation: I_drug/I_control = 1/[1 + (D/IC₅₀)^n_H], where D is the drug concentration, IC₅₀ is the concentration for 50% inhibition, and n_H is the Hill coefficient. Data are expressed as means ± SD unless indicated. ANOVA was applied for analyzing the multigroup data. Student's t-test was used to compare unpaired data between two groups, and a two-tailed P < 0.05 was taken to indicate statistical significance.

Generation of Kv1/Kv2DN mice. Adult female FVB mice heterozygous for the DN transgene Kv1.1N206 were crossbred with male C57/BL6 mice carrying the DN transgene Kv2.1N216 (16). The resulting animals were screened for the presence of one or two transgenes by PCR analyses of tail DNA. Kv1/Kv2DN mice expressed both transgenes.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

I_K,slow2 exists in both control and Kv1DN mouse ventricular myocytes. We previously reported (18) that the outward current of adult mouse ventricular myocytes was composed of at least three components, each of which displayed distinct inactivation kinetics at 37°C. Here we confirm that, at room temperature, the decay of the outward currents elicited in control myocytes by 5-s pulses could be best fitted by three exponentials with the following time constants: ₁ 40 ms, ₂ 350 ms, and ₃ = 1.6-2.0 s (Fig. 1, A and C). These components represent I_to, I_K,slow1 (18), and I_K,slow2, respectively. By contrast, the outward currents in almost all the myocytes (86 of 89 cells in this study) derived from Kv1DN mouse hearts had only two exponential components (Fig. 1, B and C). The time constants were indistinguishable from those of I_to and I_K,slow2 in the control myocytes, confirming that I_K,slow1 was completely eliminated. Only at 20 mV did the time constants of the fastest components (I_to) differ between the two groups. However, it is unlikely that I_to inactivates any faster at this particular membrane potential in the Kv1DN myocytes than in control myocytes. In fact, I_to occupies a relatively small segment (~200 ms) during the whole recording course (5 s) because of its fast inactivation kinetics, and its amplitude was relatively small at 20 mV, making clear separation of this current from the overlapping I_K,slow1 in control myocytes somewhat difficult. In Fig. 1D, the relative amplitudes obtained from the exponential fittings showed a larger proportion of I_to in the total outward current in Kv1DN myocytes than in that of the controls (e.g., 53.4 ± 14.3% vs. 27.1 ± 8.2% at 20 mV; P < 0.01). However, the absolute values of the amplitudes of I_to, after normalization to individual cell capacitances, did not significantly differ between the two groups (e.g., 16.4 ± 9.4 vs. 11.6 ± 5.6 pA/pF at 20 mV; P > 0.05).

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Existence of IK,slow2 in both control and Kv1DN mouse ventricular myocytes. A and B: outward current traces from a representative control and a Kv1DN myocyte, respectively. Currents were elicited by a series of 5-s depolarizations ranging from 60 to 30 mV, with 10-mV increments. Holding potential (HP) was 80 mV. C and D: time constants and relative amplitudes, respectively, of 3 components of the outward current as a function of the membrane potential (MP). Tri- or biexponential fittings were performed on the current decays of control or Kv1DN myocytes. Opened and closed symbols represent the data from control (n = 13) and Kv1DN (n = 11) myocytes, respectively; squares, circles, and triangles show the data from 3 kinetically different components (Ito, IK,slow1, and IK,slow2, respectively; *P < 0.05). For clarity, inset in D shows the relative amplitudes at 20 mV obtained from control (open bars) and Kv1DN (shaded bars) myocytes.

TEA-sensitive I_K,slow2 is upregulated in Kv1DN myocytes. Although the exponential fitting permits the assessment of the amplitude values of each component, the accuracy of this method is limited. We therefore applied TEA to better evaluate I_K,slow2. Our results showed that 5 mM TEA inhibited most of the I_K,slow2 without significantly affecting the other two components (Fig. 2A). The time constants of the inactivation kinetics (obtained by single-exponential fittings to the current decays) of the TEA-sensitive current were indistinguishable from those of I_K,slow2 before the administration of TEA (data not shown). The TEA-sensitive current started to activate at about 30 mV, and the activation process was relatively slow. The current-voltage relationship (Fig. 2B) showed an outward rectification. Interestingly, the current density of the TEA-sensitive component in the Kv1DN myocytes was increased by ~90% at potentials positive to 30 mV compared with the control myocytes (P < 0.05). Thus cardiocytes derived from Kv1DN hearts underwent electrical remodeling with a significant enhancement of the expression of a TEA-sensitive current, I_K,slow2, to compensate for the loss of I_K,slow1.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Upregulation of IK,slow2 in Kv1DN myocytes. A: current traces from a representative control and a Kv1DN myocyte before (a) and after (b) the administration of 5 mM tetraethylammonium (TEA). The voltage protocol was the same as described in Fig. 1. Digital subtraction (a  b) was performed off-line to demonstrate the TEA-sensitive current (c). B: current-voltage (I-V) relationship of the TEA-sensitive current in control and Kv1DN myocytes. Current amplitudes measured at the peak of the TEA-sensitive current were normalized to individual cell capacitances. Values are means ± SE (n = 14 for control; n = 9 for Kv1DN). *P < 0.05 vs. control.

Regional remodeling: enhancement of I_K,slow2 at apex of Kv1DN hearts. A number of studies have emphasized the electrical heterogeneity that exists within the heart, pointing to regional differences both in electrical properties of ventricular myocardium and in the response of distinct regions to pharmacological agents and pathological states (2). The current density of I_K,slow2 in randomly studied Kv1DN myocytes (Fig. 2) displayed marked variation, suggesting varying densities of this current in the Kv1DN mouse cardiocytes. Moreover, previous optical mapping studies revealed markedly longer refractory periods at the base than at the apex of Kv1DN hearts (4). A premature impulse applied to the apex of Kv1DN hearts induced reentry after encountering a functional line of conduction block (4). To further evaluate the molecular basis of the spatial dispersion of the repolarization, we compared the three major outward current components expressed in the myocytes isolated from the apical segments of the left ventricle free wall with those from the base segments of either control or Kv1DN mouse hearts. Here, I_K,slow1 and I_K,slow2 were determined by measuring the 25 µM 4-aminopyridine (4-AP)-sensitive (14) and 500 µM TEA-sensitive currents, respectively, in response to a 5-s pulse from HP of 80 mV to 40 mV. I_to was determined by fitting the current traces in the presence of 25 µM 4-AP by two exponentials. The current was elicited by a 1-s pulse from 80 to 40 mV. These studies revealed that cardiac myocytes derived from the apex of Kv1DN hearts expressed a twofold higher density of I_K,slow2 current compared with cardiocytes derived from the base (P < 0.05; Fig. 3C). By contrast, I_to density did not differ in cardiocytes derived from either the apex or base. Moreover, in control myocytes, no significant difference was found between the current densities of I_to, I_K,slow1, or I_K,slow2 expressed in cardiocytes derived from the apex and those from the base. However, the current density of I_K,slow2 in the Kv1DN apical cells was significantly greater than that in the control apical myocytes (P < 0.01), whereas in the basal myocytes it remained similar to the level of I_K,slow2 in the control basal cells (P > 0.05). Together, these data indicate a selective upregulation of I_K,slow2 in the apical segments of Kv1DN hearts.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Current densities of Ito (A), IK,slow1 (B), and IK,slow2 (C) in apical and basal myocytes isolated from control and Kv1DN mouse ventricles. Ito was obtained by fitting the current trace in response to a 1,000-ms depolarization pulse from 80 to 40 mV in the presence of 25 µM 4-aminopyridine (4-AP) by 2 exponentials. IK,slow1 and IK,slow2 were determined as the 25 µM 4-AP- and 500 µM TEA-sensitive currents, respectively. Sample numbers of each group are shown in the bars. Ap and Bs, data from apical and basal myocytes, respectively. Values are means ± SE. *P < 0.05, **P < 0.01 vs. Kv1DN apex.

Biophysical properties of I_K,slow2. Because the TEA-sensitive current provided us with a fairly pure I_K,slow2, a single-exponential fit to the rising phases of the current was performed to determine the activation kinetics of I_K,slow2. The resulting time constants were then plotted as a function of the membrane potential. These data showed no significant differences in activation time constants between the control and Kv1DN groups (Fig. 4A), indicating that the activation kinetics of I_K,slow2 was not altered in the transgenic mouse myocytes. The activation rate of I_K,slow2 was much slower than that of the other two major components of the outward current, I_to and I_K,slow1, which were described previously (18).

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Biophysical properties of IK,slow2. A: voltage dependence of activation time constants () of IK,slow2. Time constants were obtained by fitting the TEA-sensitive current as shown in inset by a single-exponential function. Opened and closed circles indicate the data from control and Kv1DN myocytes (n = 5 for each group), respectively. Solid lines are the single-exponential fitting curves; and the dotted line shows the best fit to the mean values when the data from the control and Kv1DN myocytes are combined. B: voltage dependence of the steady-state activation of IK,slow2. Data were obtained from 7 observations by measuring the tail currents as shown in the inset (arrow). The solid line shows the best fit to the mean values by a single Boltzmann function with a half-activation voltage (V1/2) of 13 mV and a slope factor (S) of 8.7 mV. Inset: current traces from a representative Kv1DN cardiocyte elicited by a series of test potentials ranging from 50 to 40 mV. A 200-ms prepulse from HP of 50 mV to 50 mV was applied 3 ms before the test pulses to inactivate Ito. The inward sodium current (INa) was also evoked in this cell. Horizontal and vertical bars represent 50 ms and 500 pA, respectively. C: voltage dependence of steady-state inactivation of IK,slow2. Current (inset) was elicited by a double-pulse protocol. The voltages were held at 80 mV and stepped to 30 mV in increments of 10 mV for 10 s, followed by a 10-s depolarization step to 40 mV. Averaged relative currents (I/Imax, n = 8) were fitted by a single Boltzmann function with an averaged V1/2 of 32.0 mV and a slope factor S of 6.5 mV. D: recovery kinetics of IK,slow2. Current recordings elicited by a standard voltage protocol from a Kv1DN ventricular myocyte are shown in the inset. Currents were measured at 200 ms after depolarization to minimize the contamination of Ito. Mean data from 5 observations were fitted by double exponentials that resulted in fast (1) and slow (2) time constants of 129.4 and 1,720.4 ms, respectively, with relative amplitudes of 0.38 and 0.59, respectively.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effects of TEA and 4-AP on IK,slow2. Experiments were conducted in Kv1DN myocytes exclusively. A: current traces before (a) and after administration of 0.5 (b) and 5 (c) mM TEA. Currents were elicited by an 8-s step pulse from 80 mV (HP) to 40 mV at 0.03 Hz. B: summary of the effect of TEA. Currents were measured as the difference between the level at 200-ms depolarization and that at the end of pulse. Data were fitted by the Hill equation (n = 6). C: current recordings in the absence (a) and presence of 4-AP (b = 1, c = 5 mM). D: concentration dependence of block by 4-AP on IK,slow2; n = 5.

Pharmacological profile of I_K,slow2. To study the sensitivity of I_K,slow2 to the blockade of TEA, cells were superfused with Tyrode solution containing different concentrations of this agent (Fig. 5A). The amplitude of currents, measured as the difference between the level at 200-ms depolarization and that at the end of the pulse, were normalized to the respective control values (before drug application) and plotted as a function of the concentration (Fig. 5B). The concentration-dependent effect was evaluated by fitting the data to the Hill equation. The IC₅₀ of TEA necessary to block I_K,slow2 was 638 ± 231 µM (mean ± SE; n = 7), and the n_H of 1.0 indicates that TEA binds to a single site of the I_K,slow2 channel. These data suggest that I_K,slow2 is much more sensitive to TEA than the two major current components of the outward current, I_to and I_K,slow1, which are relatively insensitive to TEA (up to 10 mM) (18).

Functional role of I_K,slow2. To evaluate the functional importance of the upregulation of I_K,slow2, we applied 5 mM TEA to control and Kv1DN myocytes and investigated its effect on their action potentials. As shown in Fig. 6A, the APD of a representative Kv1DN cardiac cell was significantly prolonged by the application of 5 mM TEA, whereas only a mild effect was seen in a control myocyte. The prolongation was more profound in the late phase of repolarization. Similar effects were observed in several other Kv1DN and control myocytes. At this concentration, TEA prolonged the APD at 95% repolarization by 70.4 ± 51.6% (from 61.1 ± 1.1 to 105.0 ± 3.9 ms; n = 7) in Kv1DN cardiocytes and by 17.4 ± 16.7% (from 48.3 ± 39.4 to 56.2 ± 45.7 ms; n = 10) in control myocytes (P < 0.01), whereas the prolongation of APD at 30%, 50%, and 70% repolarization in both groups did not differ significantly (Fig. 6B). The prolongation of the APD by TEA in Kv1DN myocytes showed a large variation, reflecting the uneven upregulation of the TEA-sensitive current, I_K,slow2, in the whole ventricle and its different functional contributions in myocytes that have different lengths of repolarization. These results indicate that I_K,slow2 participates in the last stage of repolarization in the formation of the action potential, which is supported by the relatively slow activation kinetics of this current. Upregulated I_K,slow2 contributes more to the repolarization of the action potential in Kv1DN myocytes.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Functional role of IK,slow2. A: effect of 5 mM TEA on the action potential (AP) recorded from a control and a Kv1DN myocyte. APs were elicited by suprathreshold currents injected through the recording electrode at 2 Hz under the current-clamp mode. Arrows indicate the recordings after the application of 5 mM TEA. Horizontal and vertical bars represent 20 ms and 20 mV, respectively. B: summary of the TEA effect on the action potential duration. APD30, APD50, APD70, APD90, and APD95, action potential duration at 30%, 50%, 70%, 90%, and 95% repolarization, respectively. Values are means ± SD (n = 10 for control; n = 7 for Kv1DN). *P < 0.05, **P < 0.01 vs. control.

I_K,slow2 is eliminated by crossbreeding with Kv2DN mice. Our data indicate that the biophysical and pharmacological properties of I_K,slow2 are similar to those of the Kv2 family of voltage-gated potassium channels (see DISCUSSION). To show the correlation between the expression of the Kv2 gene and this current, we first checked the steady-state levels of Kv2.1 transcript in either control or Kv1DN hearts. The results revealed that there was a twofold increase in the steady-state level of Kv2.1 transcript (Fig. 7A). Real-time PCR also showed a twofold increase in Kv2.1 transcript (data not shown). To further determine the gene that codes for the upregulated I_K,slow2, we crossbred Kv1DN mice with mice expressing a truncated Kv2 polypeptide in the heart (Kv2DN). Kv1/Kv2DN cardiocytes exhibited outward currents (Fig. 7B) that lacked both the 4-AP- (Fig. 7E) and the TEA (Fig. 7F)-sensitive components of I_K,slow. Thus overexpression of Kv2N216 eliminated I_K,slow2 in Kv1DN mice

View larger version (33K): [in this window] [in a new window]   Fig. 7.   A: steady-state levels of Kv2.1 transcript in Kv1DN mice. Representative Northern blot analysis of wild-type (WT) and Kv1DN mice. Each lane represents a sample from 1 mouse. Bar graph depicts densitometric analyses of the ratio of Kv to GAPDH compared with that in WT mice (n = 5; P < 0.05). B: outward currents expressed in Kv1/Kv2DN myocytes. Current traces under control conditions (B) and in the presence of 50 µM 4-AP (C) or 5 mM TEA (D) from a representative cell are shown. E and F represent 4-AP- and TEA-sensitive currents by the subtraction method. Scale bars are identical for all traces.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

I_K,slow2 is encoded by Kv2.1. Our recent studies (14, 18) suggest that, besides I_to, a rapidly activating, slowly inactivating 4-AP-sensitive delayed rectifier current plays an important role in the cardiac repolarization of murine hearts. This current, first termed I_slow (now referred to as I_K,slow1), has an inactivation constant of ~400 ms at room temperature. The continued decline of the outward current of mouse ventricular myocytes after 3-4 s of depolarization, when I_K,slow1 is already fully inactivated, pointed toward the existence of another component (I_K,slow2) with slower inactivation kinetics than I_K,slow1. In the present paper, we have described the electrophysiological features and functional significance of I_K,slow2. To separate the existing three components in the total outward current of the control (wild-type) myocytes, we applied three exponentials to best describe the current decay. Although the curve-fitting method to determine the number of current components is relatively crude and may introduce bias, we have found that the utility of three-exponential fitting in control myocytes is not only statistically valid (examined by F-test, which indicated that goodness of fit by 3 exponentials is superior to the 2-exponential model at potentials more positive than 10 mV when all current components are well activated; Ref. 18) but also well supported by the functional data: 1) The 4-AP-sensitive current (I_K,slow1) had only one inactivation component with time constants (400-500 ms) similar to ₂ obtained from the control outward current, and it was completely inactivated within 3-4 min, as could be calculated from ₂; 2) the TEA-sensitive (at <1 mM) current (I_K,slow2) also had single-exponential inactivation kinetics with time constants of 1.6-2 s indistinguishable from ₃; 3) in Kv1DN mouse myocytes, I_K,slow1 was selectively eliminated, leaving I_to and I_K,slow2 with unchanged time constants compared with ₁ and ₃ in the control current. Obviously, the use of two exponentials to fit the current decay of control myocytes will give different kinetic parameters, and therefore the time constants of ~80 ms for inactivation of I_to and >1 s for "I_K,slow" reported by Xu et al. (17) actually reflect an incomplete separation of I_to, I_K,slow1, and I_K,slow2. In fact, in the studies on Kv2DN mice (16), these authors described an "accelerated" inactivation kinetics of I_K,slow, so they suspected that it might have two components and that the slower component was likely encoded by the Kv2 -subunits. The present study refines and extends these observations and fully characterizes I_K,slow2.

Prolongation of APD and QT interval-induced electrical remodeling. We created a mouse with a marked reduction in I_K,slow1, a significant prolongation of the APD, and a prolonged QT interval (14). The mouse heart exhibited an enhanced spatial dispersion of repolarization and a highly arrhythmogenic substrate (4, 12). Here we showed that the suppression of I_K,slow1 was associated with an upregulation of I_K,slow2. The gating properties and pharmacological features of I_K,slow2 in the transgenic mice were not altered. Therefore, it is likely that the increase of I_K,slow2 was due to either an increase in the number of the functional channels or an increase of open probability, although single-channel recordings were not conducted to rule out the possibility of altered single-channel conductance. Thus the prolongation of APD and QT interval likely triggered the induction of I_K,slow2 to partially compensate for the loss of I_K,slow1. Interestingly, a similar electrical remodeling phenomenon was also described in transgenic mice (Kv4DN) overexpressing Kv4.2W362F (5), a nonconducting mutant Kv4.2 -subunit, in the heart. The selective suppression of I_to,f in these mice resulted in the induction of I_to,s, a slowly inactivating transient outward current (10). Together, these observations suggest that the electrical remodeling induced by the inhibition of a repolarizing current is gene specific and triggers a compensatory response of a current most similar to the suppressed current.

Limitations of this study. The outward current of mammalian cardiac myocytes is normally composed of several different depolarization-activated ionic current components. At least five repolarization currents (10, 14, 16) have been described in adult murine ventricular myocytes: I_to,f, I_to,s, I_K,slow1, I_K,slow2, and sustained current (I_sus). Although these currents have distinct gating and pharmacological properties, it is still difficult to efficiently separate and accurately assess each component. In the present paper, we have used low concentrations of 4-AP or TEA to investigate I_K,slow1 or I_K,slow2 because, at these concentrations, the two blockers had little effect on the other current. However, the 4-AP- or TEA-sensitive currents represent only part of the I_K,slow1 or I_K,slow2 (tested at ~IC₅₀ concentrations). Thus the densities of these currents may have been underestimated in our studies. For the measurement of I_to, a curve fitting to a sum of two exponentials was used to resolve the contamination problems and the current traces were elicited by a short (1-s) depolarization pulse in the presence of 25 µM 4-AP to suppress I_K,slow1.