Identification of Gating Modes in Single Native Na+ Channels From Human Atrium and Ventricle
The aim of the present study was to investigate the single-channel properties of different gating modes in the native human cardiac Na+ channel. Patch-clamp experiments were performed at low noise using ultrathick-walled pipettes. In 17 cell-attached patches containing only one channel, fast back and forth switching between five different Na+-channel gating modes (F-mode, M1-mode, M2-mode, S-mode, and P-mode) was identified, but no difference in the gating properties was found between normal and diseased cardiomyocytes from atrium or ventricle, respectively. Hodgkin-Huxley fits to the ensemble-averaged currents yielded the activation-time (τm) and inactivation-time (τh) constants. τm was comparably fast in the F-mode, M1-mode, M2-mode, and S-mode (0.15 ms) and slow in the P-mode (0.3 ms). τh ranged from 0.35 ms (F-mode) to 4.5 ms (S-mode and P-mode). The mean open-channel lifetime (τo) was shortest in the F-mode and P-mode (0.15 ms) and longest in the S-mode (1.25 ms). The time before which half of the first channel openings occurred (t0.5) was comparably short in the F-mode, M1-mode, M2-mode, and S-mode (0.3 ms) and long in the P-mode (0.9 ms). It is concluded that (1) a single native human cardiac Na+ channel can be recorded at low noise, (2) this channel can change its gating properties at a time scale of milliseconds, (3) lifetimes of the observed gating modes are short ranging from milliseconds to seconds only, and (4) the gating modes are characterized by specific activation and inactivation kinetics and differ at least in their mean open time and first latency.
Although Na+ channels have been subject to intense electrophysiological investigation in various tissues, 1 little is known about the native human cardiac Na+ channel. Previously published studies have been conducted in heterologous expression systems, where channel proteins are removed from their physiological environment.2–4⇓⇓ In these experiments, the Na+ channel α-subunit (hH1) was expressed alone, or coexpression with the β1-subunit was performed. Other methods used the reconstituted Na+-channel protein fused into planar lipid bilayers.5 A major disadvantage of these methods is the artificial environment, in which electrophysiological experiments on the channel were performed. In the present study, results on single native human cardiac Na+ channels are presented. Our approach may lead to important insights into the physiological function of this channel protein.
Using this approach, arrhythmias may be understood in more detail. It has been shown that the LQT3 syndrome and the Brugada syndrome are based on different mutations in the gene (SCN5A, 3p21) encoding for hH1.6–8⇓⇓ In heterologous expression systems, a distinct mutation (ΔKPQ) underlying the LQT3 syndrome is associated with late Na+ currents.9 From fits of the results of these experiments to different kinetic models, it was postulated that modal gating of the channel protein may be the reason for the late currents.9,10⇓ Hence, genetic defects may underlie different rates of switching between gating modes. Therefore, the analysis of modal gating in native human cardiac Na+ channels may also help to better understand pathophysiological processes.
For the characterization of different gating modes, it is necessary to record at low noise from a patch, which contains no more than one channel. By using thick-walled patch pipettes11 with ultraclean tips,12 five distinct gating modes of Na+-channel action (F-mode, M1-mode, M2-mode, S-mode, and P-mode) have been described in ventricular cells from adult white mice.13–15⇓⇓ They differ at least in activation and/or inactivation kinetics, open-channel lifetime, and first latency. In the present report, we identified these gating modes in atrial and ventricular cells of normal and diseased human myocardium. Durations of these modes were observed to be short.
Materials and Methods
Right atrial appendage specimens were obtained from patients suffering from coronary heart disease or aortic-valve disease undergoing heart surgery. The procedure for isolation of cardiomyocytes from human atrial appendages has been described in detail by Brandt et al.16 Single cells of nonfailing (not suitable for transplantation) and terminally failing (ischemic/dilated cardiomyopathy) ventricles were isolated according to the procedure described by Beuckelmann et al.17 The experimental protocol was designed to conform with the recommendations from the Declaration of Helsinki and approved by the ethical committee of the University of Cologne.
Solutions and Temperature
The cardioplegic solution used for transport of the tissue was composed of (in mmol/L) NaCl 15, KCl 9, MgCl2 4, histidine hydrochloride 18, histidine 180, tryptophan 2, mannitol 30, CaCl2 0.015, glutaric acid 1, and KOH 2. The modified Tyrode’s solution for cell isolation and storage contained (in mmol/L) NaCl 135, KCl 4, MgCl2 1, CaCl2 0.1, NaH2PO4 0.33, HEPES 10, and glucose 10 (pH 7.3). The experiments were performed at 23±1°C in a bath solution that contained (in mmol/L) KCl 230, CsCl 20, MgCl2 1, EGTA 10, and HEPES 5 (pH 7.3) and a pipette solution that contained (in mmol/L) NaCl 255, CaCl2 2.5, KCl 4, and HEPES 5 (pH 7.3). The elevated Na+ concentration was used to enhance unitary current amplitudes.18
For low-noise recordings,11 patch pipettes were prepared from thick-walled borosilicate glass tubing (Hilgenberg) with an external diameter of 2.0 mm and an internal diameter of 0.25 mm. The patch pipettes were inserted into a pipette holder of small dimension made completely of silver.11 Pipette tips were prepared only seconds before starting gigaohm seal formation by breaking off the final tip region at the glass bottom of the bath chamber.12 Gigaohm seals were formed after touching the cell membrane by application of slight suction.
Data Acquisition and Analysis
Unitary Na+-channel currents were recorded at a sampling rate of 100 kHz in the cell-attached patch configuration with an Axopatch 200B amplifier (integrating head stage, intrinsic noise 0.045 pA rms at 5 kHz, Axon Instruments, Inc). Analog filtering was performed with an intrinsic filter at a bandwidth of 10 kHz. When data needed further filtering for the analysis, an offline gaussian filter algorithm was used. The holding potential was −120 mV, duration of prepulses was 20 ms, and 4-ms pulses were applied at a rate of 10 Hz. Capacitive transients were compensated for carefully via compensation circuits containing two exponentials. Leak and residual capacitive currents were removed by subtracting averaged blank traces, which were formed exclusively from traces in the neighborhood of the actual sweep. None of the patches in the present study showed any sign of containing more than one active Na+ channel. This was verified by inspecting several thousands of consecutive traces at pulses to −40 mV or more positive potentials to exclude the existence of any superimposition of opening events. Histograms of the open-channel lifetime were constructed by using the baseline method.11 Herein single-channel open times were calculated from the distribution of the opening-induced gaps in the middle of the baseline noise. In contrast to usual open-time histograms, the distribution of dwell times measured at the baseline contains both channel events (to) and noise events (tn). For separating the to from the tn distribution, an exponential can be fitted to the tn distribution, and the corresponding tn bins can be clipped for fitting the to distribution. To improve resolution, amplitude histograms were formed by eliminating the transition points with the variance-mean technique.19 For curve-fitting, a derivative-free Levenberg-Marquardt routine20 was used. Recording was performed on a PC-80486 with P-clamp software (Axon Instruments, Inc). Analysis was performed on a Pentium personal computer with ISO2 software (MFK Computer) after converting the data with PCV software (MFK Computer).
Gating modes were selected by reidentifying blocks of traces from the average-of-interval plots. In these diagrams, the time course of an experiment is illustrated by plotting the averaged current of each individual sweep as a dot. The method of separating different gating modes is illustrated in Figure 1.
Figure 1A depicts a narrow part of an average-of-interval plot, in which switching between two different gating modes is detected. In Figure 1B, 16 consecutive sweeps in the F-mode (left) and 16 sweeps after switching to the S-mode (right) are illustrated. Both before and after the switch, empty sweeps, ie, traces without openings were excluded. Next (Figure 1C), ensemble-averaged currents of each suspected gating mode were formed from the traces in panel B.
Different blocks of traces in each suspected gating mode were combined to increase the number of openings and to yield smooth ensemble-averaged currents. This is demonstrated later in Results (see Figure 3). The mode-specific kinetics of the currents were then fitted (see also Figure 3) with a Hodgkin-Huxley model of the type Am3h.21 In this model, A represents a scaling factor, and m and h denote parameters describing activation and inactivation, respectively, of the Na+ channel. m is raised to the third power as activation is assumed to appear in several steps, ie, several conformational changes of the Na+-channel protein are necessary for activation. This fitting procedure yields the activation-time constant (τm) and the inactivation-time constant (τh). We have chosen only one inactivation gate (h) rather than two (h, j), as often used in the more common Am3hj scheme, because we believe that h and j may represent two different gating modes. In contrast, fitting one gate to our ensemble-averaged current represents fitting of a single gating mode. The gating modes were further differentiated by the mean open-time constant (τo) and the time before which half of the first channel openings occurred (t0.5), as is also illustrated in Results (see Figures 4 and 5⇓).
Figure 2 compares single Na+-channel currents of cells isolated from different regions of human hearts. These organs were either normal or had various diseases.
No obvious differences were found in the characteristics of single-channel currents (Figure 2A), obtained from an atrial cell of a patient suffering from aortic valve incompetence (left) and a ventricular cell from a nonfailing human heart (right). In both patches, long and short openings were observed. The amplitude of short openings may be reduced due to filtering.
Ensemble-averaged currents (Figure 2B; illustrated for the F-mode; left, atrial cell of a patient with coronary heart disease; right, ventricular cell from a nonfailing human heart), which have been constructed according to the method illustrated in Figure 1, reveal no obvious differences. Subsequently, the currents were fitted with a Hodgkin-Huxley model of the type Am3h (see Materials and Methods for details). The activation-time constant (τm) and the inactivation-time constant (τh) are indicated in both fits.
To define the mean single-channel current amplitude at −40 mV, amplitude histograms were built. Figure 2C illustrates graphics from recordings of single Na+-channel currents from an atrial cell of a patient with coronary heart disease (left) and from a ventricular cell from a nonfailing human heart (right). Each distribution could be fitted with the sum of two gaussian curves, and the peak of the baseline noise is truncated. In both cell types, the mean amplitude of the open-channel current is essentially the same, −2.368 pA in the atrial cell and −2.309 pA in the ventricular cell. Furthermore, the single-channel conductance was not altered by mode switches (not illustrated).
Figure 3 shows ensemble-averaged currents of single Na+-channel openings at test pulses to −40 mV in the F-mode, M1-mode, M2-mode, S-mode, and P-mode, respectively. Gating modes were reidentified according to the method illustrated in Figure 1. Macroscopic kinetics of the Na+ current are characteristic in each mode.13,15⇓ In the F-mode, fast activation is followed by fast inactivation; in the M1-mode, fast activation is followed by fast to intermediate inactivation; in the M2-mode, fast activation is followed by intermediate to slow inactivation; in the S-mode, fast activation is followed by slow inactivation; and in the P-mode, slow activation is followed by slow inactivation.
The mode-specific kinetics of the currents were fitted with a Hodgkin-Huxley model of the type Am3h (see Materials and Methods for details). The magnitude of the parameters, activation-time constant (τm) and inactivation-time constant (τh), is indicated for each mode. The kinetics of activation in the F-mode, M1-mode, M2-mode, and S-mode are not fully resolved, but the data show that in the P-mode τm is slower by a factor of about two. Comparison of the first latency yields a respective factor of about three (see Figure 5). More accurate comparison is possible for τh. Compared with the F-mode, the inactivation-time constant is nearly 2 times slower in the M1-mode, more than 4 times slower in the M2-mode, and ≈13 times slower in the S-mode and the P-mode, respectively.
Figure 4 shows histograms of the open-channel lifetime in single-channel patches at test pulses to −40 mV in the F-mode, M1-mode, M2-mode, S-mode, and P-mode, respectively. Gating modes were reidentified from the average-of-interval plots and represent the same traces as those of Figure 3.
All histograms of the open-channel lifetime were formed by making use of the baseline method (see Materials and Methods for details).11 In the F-mode, M1-mode, M2-mode, and P-mode, respectively, only one mean open time was found. The F-mode (τo=0.15 ms) and the P-mode (τo=0.14 ms) are dominated by a short mean open time. The M1-mode (τo=0.28 ms) and the M2-mode (τo=0.40 ms) are dominated by intermediate mean open times. In the S-mode, two mean open times were found. One was long (τo1=1.24 ms) and the other was short (τo2=0.12 ms). These are addressed later (see Discussion).
Figure 5 shows histograms of the cumulative first latency in one-channel patches at test pulses to −40 mV in the F-mode, M1-mode, M2-mode, S-mode, and P-mode, respectively. Gating modes represent the same traces as those in Figures 3 and 4⇑.
Cumulative first-latency distributions for the F-mode, M1-mode, M2-mode, and S-mode differ only slightly. The time before which half of the first channel openings had occurred (t0.5) in each of these modes is ≈0.3 ms. The time course of the cumulative first latency in the P-mode was much slower, yielding t0.5 of ≈0.9 ms. In this mode, t0.5 may be even larger, because late first channel openings may have been cut off by the pulse duration of only 4 ms.
Because only traces with openings were reidentified from the average-of-interval plots, the probability of nonempty traces is zero in all plots.
To our knowledge, this is the first description of the gating characteristics of single native human cardiac Na+ channels in atrium and ventricle. We looked for distinct Na+-channel gating modes (F-mode, M1-mode, M2-mode, S-mode, and P-mode), which have previously been characterized in myocardial cells from adult white mice.13 For this purpose, patches containing only a single Na+ channel were investigated. In previous studies,13–15⇓⇓ gating modes were identified from the long time course of the averaged current per trace (average-of-interval plots). There it was found that modal gating may happen in at least two different ways: (1) The lifetime of a specific mode may be rather long, ie, about some seconds to several minutes or (2) the lifetime of a specific mode may be rather short, ie, less than about a few seconds. The second possibility was derived from the fact that in some experiments, from only one large block of traces of the average-of-interval plots, two different mean open-time constants were found. We suggested that fast back and forth switching between two different gating modes was present in these recordings. Both in (1) and (2), the time course of switching itself was not resolved. We observed that it appeared from one pulse to the next, which means that at a pulse rate of 10 Hz, switching may have happened faster than 100 ms. Sometimes switching even appeared within one trace, indicating a time course of less than a few milliseconds. The gating modes in the present report were also identified from the average-of-interval plots. Because only fast back and forth switching between gating modes according to type 2 was observed, a new method had to be found to resolve different gating modes. The results indicate that identification of relatively small blocks of sweeps inherent in individual gating modes and their subsequent combination facilitates the analysis of characteristics of human Na+-channel gating modes.
We found that the mean open time was short and nearly identical in the F-mode (τo=0.15 ms) and in the P-mode (τo=0.14 ms). It was prolonged in the M1-mode (τo=0.28 ms) and in the M2-mode (τo=0.40 ms). In the S-mode, two different mean open-time constants were found. The first was long (τo1=1.25 ms) and the second was short (τo2=0.12 ms). Because we were not able to separate these two time constants by selecting different traces, we conclude that only in this case switching between two different gating modes happened within the same trace. We assume that the long open time represents the S-mode, and the short open time represents the P-mode.
Cumulative first-latency distributions of different gating modes allowed determination of the time before which half of the first channel openings occurred (t0.5). This was fast and nearly identical in the F-mode (t0.5=0.26 ms), the M1-mode (t0.5=0.31 ms), the M2-mode (t0.5=0.36 ms), and the S-mode (t0.5=0.30 ms). In contrast, it was slow in the P-mode (t0.5=0.92 ms).
From the Hodgkin-Huxley fits of the ensemble-averaged currents, activation and inactivation kinetics of the different gating modes were calculated. The activation-time constant (τm) was fast and similar in the F-mode (τm=0.14 ms), the M1-mode (τm=0.17 ms), the M2-mode (τm=0.17 ms), and the S-mode (τm=0.14 ms), and it was slow in the P-mode (τm=0.29 ms). The inactivation-time constant (τh) was fast in the F-mode (τh=0.35 ms) and successively slower in the M1-mode (τh=0.62 ms), the M2-mode (τh=1.56 ms), and the S-mode (τh=4.57 ms). In the P-mode (τh=4.50 ms), it was identical to the S-mode.
The mean single Na+-channel current amplitude was identical in atrial and ventricular myocytes, in normal and diseased tissue, and in different gating modes, ie, it was not altered by mode switching.
When the five gating modes of native human cardiac Na+ channels are compared with those of adult white mice (see Table 1 in Böhle and Benndorf13; see Figure 3 in Böhle et al15), only slight differences can be found. The mean open times (τo), the activation-time constants (τm), and the inactivation-time constants (τh) are almost identical. Whereas in human cardiac Na+ channels the time before which half of the first channel openings occurred (t0.5) was slightly slower in the F-mode, M1-mode, M2-mode, and S-mode, it was slightly faster in the P-mode. In tissue from adult white mice, fast back and forth switching was observed only between the S-mode and the P-mode. In contrast in human tissue, we found fast back and forth switching between all five gating modes. Thus, the main difference in modal gating of myocardial cells from adult white mice compared with human tissue is the respective lifetime of the five gating modes. In human tissue, in all five modes it was found to be only short. In tissue from adult white mice, a long lifetime could be shown in the F-mode, M1-mode, M2-mode, and the P-mode, but in the S-mode it was found to be only short. In the P-mode, it could be short, too.
All gating modes were found in both normal and diseased human atrium and ventricle, but no regulation site is yet detected. Modal gating may be regulated by (1) phosphorylation through protein kinase C22 (brain Na+ channels),23–24⇓ (cardiac Na+ channels), and/or by protein kinase A24–26⇓⇓ (cardiac Na+ channels),27–30⇓⇓⇓ (cardiac L-type Ca2+ channels), and (2) mechanisms involving intracellular Ca2+,31–33⇓⇓ or (3) mechanical stress on the plasma membrane during contraction of the heart muscle cell. In this context, it is important that the contraction-relaxation cycle of single heart-muscle cells is altered in terminal heart failure. Compared with nonfailing cells, the intracellular Ca2+ concentration during relaxation, ie, during the diastolic phase, is increased, and during contraction systole, it is depressed. Additionally, the diastolic decay of the intracellular Ca2+ concentration is markedly slowed.17 When regulation sites (2) or (3) are discussed, it might be of further importance that different types of Na+ current, presumably reflecting different gating modes, have been identified in single beating cells of chick embryo heart and have been correlated to different phases of the action potential by direct measurement.34–36⇓⇓ From these results, we suspect that synchronous appearance of gating modes may be triggered by contraction of the muscle cell.
This research was supported by grants from the Bundesministerium für Bildung, Forschung und Technologie (BMBF 01 KS 9502, ZMMK, Projekt 4), and the M. and W. Boll-Stiftung. We thank Nadine Henn and Iris Berg for their excellent technical assistance, Harald Metzner and Jürgen Staszewski (Center of Physiology and Pathophysiology, University of Cologne) and their colleagues for their continual technical advice and support, and Prof DeVivie (Department of Cardiac Surgery, University of Cologne) and his colleagues for providing the myocardial tissue.
Original received March 11, 2002; revision received July 15, 2002; accepted August 2, 2002.
- ↵Nuss HB, Chiamvimonvat N, Perez-Garcia MT, Tomaselli GF, Marban E. Functional association of the β1 subunit with human cardiac (hH1) and rat skeletal muscle (μ1) sodium channel α subunits expressed in Xenopus oocytes. J Gen Physiol. 1995; 106: 1171–1191.
- ↵Makita N, Shirai N, Wang DW, Sasaki K, George AL Jr, Kanno M, Kitabatake A. Cardiac Na+ channel dysfunction in Brugada syndrome is aggravated by β1-subunit. Circulation. 2000; 101: 54–60.
- ↵Wartenberg HC, Wartenberg JP, Urban BW. Single sodium channels from human ventricular muscle in planar lipid bilayers. Basic Res Cardiol. 2001; 96: 6645–651.
- ↵Kambouris NG, Nuss HB, Johns DC, Tomaselli GF, Marban E, Balser JR. Phenotypic characterization of a novel long-QT syndrome mutation (R1623Q) in the cardiac sodium channel. Circulation. 1998; 97: 640–644.
- ↵Chen Q, Kirsch GE, Zhang D, Brugada R, Brugada J, Brugada P, Potenza D, Moya A, Borggrefe M, Breithardt G, Ortiz-Lopez R, Wang Z, Antzelevitch C, O’Brien RE, Schulze-Bahr E, Keating MT, Towbin JA, Wang Q. Genetic basis and molecular mechanism for idiopathic ventricular fibrillation. Nature. 1998; 292: 293–296.
- ↵Benndorf K. Low noise recording.In: Sakmann B, Neher E, eds. Single Channel Recording. 2nd ed. New York, NY: Plenum Press; 1995: 129–145.
- ↵Beuckelmann DJ, Näbauer M, Erdmann E. Intracellular calcium handling in isolated ventricular myocytes from patients with terminal heart failure. Circulation. 1992; 85: 1046–1055.
- ↵Yue DT, Lawrence JH, Marban E. Two molecular transitions influence cardiac sodium channel gating. Science. 1989; 244: 349–352.
- ↵Patlak JB. Sodium channel subconductance levels measured with a new variance-mean analysis. J Gen Physiol. 1988; 92: 413–430.
- ↵Numann R, Catterall WA, Scheuer T. Functional modulation of brain sodium channels by protein kinase C phosphorylation. Science. 1991; 254: 115–118.
- ↵Tateyama M, Rivolta I, Rao JJ, Kass RS. PKC-dependent phosphorylation of Ser1503 in the III-IV linker of the human heart Na+ channel α subunit (SCN5A) stabilizes open-state inactivation in a long QT (LQT-3) mutation. Biophys J. 2002; 82: 353a.Abstract.
- ↵Rivolta I, Tateyama M, Rao JJ, Kass RS. Regulation of cardiac Na+ channel bursting by PKA and PKC. Biophys J. 2002; 82: 180a.Abstract.
- ↵Ono K, Fozzard HA, Hanck DA. Mechanism of cAMP-dependent modulation of cardiac sodium channel current kinetics.2 Circ Res. 1993; 72: 807–815.
- ↵Egger M, Greef NG. Modulation of the number of activatable Na+ channels by [Ca2+]i and a phosphatase blocker. Biophys J. 1994; 66: 203a.Abstract.