I.V. overdose of sodium pentobarbitone (120 mg kg-1) and the portal vein was removed. The tissue was dissected free of connective tissue and fat in normal physiological salt solution (PSS), before being cut into strips and placed in 'Ca2+-free' PSS for 10 min. The cell dispersion procedure involved two sequential enzyme steps. The first step involved incubation of the tissue strips in 'Ca2+-free' PSS with 0·2-0·4 mg ml-1 of either protease type I (Sigma) or protease type XIV (Sigma) for 5 min, and then washed with 'Ca2+-free' PSS alone. In the second step, 0·5-1 mg ml-1 collagenase type XI (Sigma) was added to the tissue in 'Ca2+-free' PSS for 10 min before a final wash with 'Ca2+-free' PSS. All enzyme and wash procedures were performed at 37°C. After the enzyme treatment, cells were released into 'Ca2+-free' PSS by mechanical agitation of the tissue using a wide-bore Pasteur pipette. The solution containing dissociated cells was then centrifuged (100 g) to form a loose pellet which was re-suspended in 0·75 mM Ca2+ PSS. The cells were then plated onto glass coverslips and stored at 4°C before use (1-6 h).
The normal PSS contained (mM): NaCl, 126; KCl, 6; CaCl2, 1·5; MgCl2, 1·2; glucose, 10; and Hepes, 11; pH adjusted to 7·2 with 10 M NaOH. 'Ca2+-free' PSS and 0·75 mM Ca2+ PSS had the same composition except that Ca2+ was omitted and 0·75 mM CaCl2 was added, respectively.
Electrophysiology
Cell membrane currents were recorded with a List L/M-PC patch clamp amplifier at room temperature (20-25°C) using the standard whole-cell recording configuration. Patch pipettes were fabricated from borosilicate glass and had resistances of approximately 6 M
. After achieving the whole-cell configuration the series resistance was typically between 15 and 20 M (see Helliwell & Large, 1996, for further details). For voltage-jump relaxation analysis voltage steps were applied and currents recorded on-line using a Pentium personal computer (Gateway, Ireland), and CED hardware (1401plus) and software (Patch). Currents were filtered at 1 kHz (-3 dB, Bessel) and sampled at 3 kHz. For noise analysis and subsequent play-back of long-term records, data were also stored on a CDAT4 digital recorder (Cygnus Technology Inc., Delaware Water Gap, PA, USA).
Noise analysis
Data preparation. In order to calculate current variance (
I2) the signal was first high-pass filtered (0·3 Hz, -3 dB) to remove the DC component before being low-pass filtered (800 Hz, -3 dB), using 8-pole Butterworth filters (Barr & Stroud Ltd, London, UK). This band-pass signal was then amplified (× 10) to minimize digitization errors (0·0244 pA unit-1 ADC resolution) prior to sampling at 2 kHz using the 1401plus and Spike 2 software (CED). This signal was used to calculate I2 and simultaneously another current signal was sampled which was low-pass filtered (0·3 Hz, -3 dB) to calculate the mean current (I(barover)).
Calculation of the single channel conductance from the relationship between I2 and I. In a number of cells the decay of Icat was sufficiently slow to examine the current variance at different current levels, which provided a method for calculating the single channel conductance (
) using the assumptions outlined below.
The band-pass signal (see above) was split into 5 s data blocks containing 10 000 sample points, where the corresponding mean current data block did not change significantly in amplitude over this time period. The I2 can be calculated as follows:
(1)
where Ii is the current amplitude of the ith sample point, and N is the number of sample points (10 000). Since the band-pass signal has no DC component, Ii = 0 so that the I2 was found simply by squaring each sample point and calculating the mean (N = 10 000), which was then plotted against the corresponding I. The background variance obtained prior to activation of Icat was subtracted from the variance recorded during Icat. In order to estimate , a number of assumptions were made. First, underlying events can exist in either a conducting (open) or non-conducting (closed) state; second, gating of each channel is independent of other channels; and third, the population of channels is homogeneous (Sigworth, 1980).
Based on the above and at constant voltage, if the relationship between I2 and I is linear (i.e. the probability of a channel opening (Po) is 
1), which was found in the majority of cells, then the single channel current amplitude (i) can be calculated from the gradient since:
However, occasionally (Fig. 2C) where the relationship between I2 and I deviated from linearity, the data were fitted using a quadratic equation of the form:
(3)
where N is the number of channels. In these cases the parameter N could not be accurately determined although this did not affect estimates of i.
Assuming i has an ohmic current-voltage (I-V) relationship, was calculated using the equation:
(4)
= i / (V - Vrev), (4)
where V is the membrane potential and Vrev is the reversal potential.
Calculation of the spectral density function of Icat. A section of data where I during Icat remained constant was selected and a fast Fourier transform (FFT) of the band-pass signal, using a raised cosine (Hanning) window, was then performed (Spike 2 software). The FFT used data blocks consisting of 2048 points, which corresponded to a sample length of approximately 1·024 s, so that the lowest frequency that could be resolved was approximately 0·976 Hz. The resulting power spectra therefore contained 1024 points, where the lowest frequency point was 0·976 Hz and the highest frequency point in the spectrum was 1 kHz (i.e. half the sampling rate). The same procedure was used on sections of data before activation of Icat to obtain background spectra. The FFT mean of approximately ten data blocks was then computed and subtracted from the FFT mean of fifteen to twenty data blocks during Icat to yield the spectral density function of Icat alone.
As described above, this implementation of a FFT gave a spectrum up to 1 kHz. This showed a steep fall off in power at around 800 Hz, the corner frequency of the low-pass filter, since frequencies above this are effectively attenuated (48 dB octave-1). Furthermore, under our conditions the whole-cell configuration low-pass filtered the current signal with a corner frequency within the range 400-500 Hz, since this arrangement simulates a 1-pole RC filter. Consequently only data between the lowest frequency point up to 500 Hz were analysed. In our experiments these spectra could be described by the sum of two Lorentzians according to:
(5)
where G(f) is the spectral density at a given frequency (f), and fc1 and fc2 are the corner frequencies associated with G1(0) and G2(0), the zero frequency asymptotes for the first and second components in the spectrum, respectively. Curve fitting was performed with Origin software (MicroCal Software Inc., Northampton, MA, USA) using a least squares method where each parameter was unconstrained.
Corner frequencies were converted to time constants (
, Table 1) using the relation:
(6)
= 1 /2
fc,
for direct comparison with time constants obtained from relaxation analysis. Combining eqns (2) and (4) we have:
(7)
From the spectral density function for a single component, I 2 is the total integral of the single Lorentzian:
(8)
Thus,
(9)
Since the data could be described by a double Lorentzian (see above and Fig. 3) we have further assumed that the states involved have the same conductance, so that:
(10)
Values of calculated using this equation are shown in Table 1.
Voltage-jump relaxation analysis
The standard voltage protocol, used to examine the kinetic properties of noradrenaline-evoked Icat in various external calcium concentrations, involved steps to +50 mV for 500 ms from a holding potential (Vh) of -50 mV, and then stepping back to -30 mV for 500 ms before returning to Vh. We used a test negative potential of -30 mV, instead of Vh, since the driving force for inward current flow through cation channels at -30 mV is roughly equal to the outward current flow through cation channels at +50 mV, assuming a reversal potential of about +10 mV (Helliwell & Large, 1996). This enabled direct comparison of the amplitude of the current relaxation at +50 mV with the relaxation at -30 mV without the necessity to correct for changes in driving force. Leak-subtracted cation currents were obtained by subtracting the mean of five control currents recorded before the addition of noradrenaline from the mean of three to five currents recorded at the peak of noradrenaline-evoked Icat. Current relaxations were fitted by single exponentials using Origin software. The fitting was constrained between two time points; the first time point was selected at the start of the relaxation (determined by eye) and the second point was selected at an arbitary time after the relaxation had reached steady state. The fit was extrapolated back to the onset of the voltage pulse at +50 or -30 mV and the amplitude of the relaxation measured as the difference (in pA) between the amplitude at this time point and the amplitude after steady state had been achieved. The amplitude of the instantaneous current at -30 mV was then measured as the difference (in pA) between 0 mV and the extrapolated fit (i.e. the remaining net inward current). Similarly the amplitude of the instantaneous current at +50 mV was measured as the difference (in pA) between 0 mV and the extrapolated fit at +50 mV.
Solutions and drugs
The standard K+-free external solution used for experimentation contained (mM): NaCl, 126; CaCl2, X; glucose, 10; and Hepes, 11; buffered to pH 7·2 with 10 M NaOH, where the value of X was in the range < 10 nM to 5 mM. For X < 50 µM, 1 mM BAPTA was also added. At these lower [Ca2+]o, in the presence of BAPTA, the amount of CaCl2 required to obtain the desired free [Ca2+]o was calculated using EQCAL software (Biosoft, Ferguson, MO, USA). Nicardipine (3 µM) was also present in external solutions to block voltage-dependent Ca2+ currents. The internal pipette solution contained (mM); CsCl, 18; caesium aspartate, 108; MgCl2, 1·2; Hepes, 10; glucose, 11; BAPTA, 10; and CaCl2, 1 (free [Ca2+]i approximately 14 nM, calculated using EQCAL software); pH adjusted to 7·2 using Tris. Under these conditions voltage-gated Ca2+ currents, K+ currents and calcium-activated chloride currents were abolished and Icat was recorded in isolation without contamination from other currents. In some experiments (see Fig. 6) 1·2 mM MgCl2 was omitted from the pipette solutions and 10 mM EDTA was used instead of 10 mM BAPTA to buffer [Ca2+]i at 14 nM (calculated using EQCAL software) and chelate contaminating Mg2+. All drugs were obtained from Sigma.
The values in the text are means ± S.E.M. of n cells.
 |
RESULTS |
Effect of micromolar concentrations of Ca
on the peak amplitude of Icat
The first series of experiments was designed to confirm the facilitatory action of micromolar concentrations of Ca on Icat reported previously (Helliwell & Large, 1996). In the previous study these effects were quantified by first activating Icat in 1·5 mM Ca and then changing to the test [Ca2+]o before re-addition of 1·5 mM Ca. However, in the present work the facilitatory effect was quantified by simply measuring the peak amplitude of Icat evoked by bath-applied noradrenaline (100 µM) in the presence of various [Ca2+]o, which was also the protocol used in both noise and voltage-jump experiments. It is evident that even in negligible [Ca2+]o (< 10 nM) a relatively small Icat can be activated by noradrenaline (Fig. 1A, top trace). This indicates that Ca is not obligatory for activation of Icat although it is clear that as [Ca2+]o is increased into the micromolar range a marked potentiation in the amplitude of Icat is observed (Fig. 1A, middle and bottom traces). This concentration-dependent effect of Ca is quantified in Fig. 1B where the peak response of Icat is plotted against log[Ca2+]o. The relationship can be empirically described by a Hill equation (see legend of Fig. 1), which gave an estimated half-maximal [Ca2+]o (EC50) of 3 µM. Thus the facilitatory effect of Ca reported previously (Helliwell & Large, 1996) is the same as the phenomenon reported here since the EC50 values are similar (EC50 by the previous method was about 6 µM Ca; Helliwell & Large, 1996).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 1. The facilitatory action of Ca2+o on the amplitude of noradrenaline-activated Icat
A, bath application of noradrenaline (100 µM, open bar) evoked a relatively small Icat in < 10 nM Ca2+o (top trace), whereas addition of 3 µM (middle trace) or 100 µM (bottom trace) Ca2+o led to a marked increase in the amplitude of Icat. Standard K+-free conditions were used and Vh was -50 mV. The vertical calibration bar is the same for all records. B, the relationship between the mean peak amplitude of Icat and [Ca2+]o could be empirically described by a Hill equation of the form:
y = ymax (xnH) / (EC50nH + xnH)
which gave an EC50 of 3 µM and a coefficient, nH, of 2. Each data point presents the mean of between 5 and 10 replicates.
|
Table 1. Characteristics of Icat obtained from the spectral density function in the presence of various [Ca2+]o
| [Ca2+]o | 1 (ms) | 2 (ms) | G2(0)/G1(0) | n | (pS) | n |
| < 10 nM | 11·3 ± 1·3 | 1·2 ± 0·08 | 0·6 ± 0·09 | 6 | 10·8 ± 1·5 | 6 |
| 1 µM | 43 ± 4 | 1·4 ± 0·15 | 0·16 ± 0·03 | 5 | 9 ± 1·4 | 5 |
| 3 µM | 48·6 ± 3 | 1·2 ± 0·19 | 0·01 ± 0·002 | 8 | 22·2 ± 1·5 | 5 |
| 100 µM | 44·8 ± 4 | 1·6 ± 0·4 | 0·01 ± 0·003 | 6 | 24·1 ± 1·7 | 5 |
The corner frequencies fc1 and fc2 are shown as time constants 1 and 2, respectively.
Noise analysis of Icat in various [Ca2+]o
It was conceivable that the facilitatory action of Ca resulted from an overall increase in the unitary conductance of the underlying cation channels. Since Icat desensitized sufficiently slowly in the presence of noradrenaline (Fig. 1A), this possibility was tested by calculating the variance of the current, I2, (Fig. 2A, band-pass trace) at different mean current values, I, (Fig. 2A, low-pass trace) to obtain estimates of the single channel conductance, (see Methods). In the majority of cases, in any given [Ca2+]o the relationship between I2 and I was linear as Po was low even at the peak of Icat (e.g. Fig. 2B and D) and therefore could be calculated using eqns (2) and (4). However, in 3 µM Ca this relationship occasionally deviated from linearity at the peak of Icat (4 out of 8 cells, Fig. 2C), so that was calculated using eqn (3) instead of eqn (2). In negligible [Ca2+]o (< 10 nM) was 13 ± 0·4 pS which was similar to the value in 1 µM Ca ( = 11 ± 1 pS, n = 4). However in 3 and 100 µM Ca the single channel conductance was, respectively, 21 ± 2 pS (n = 8) and 23 ± 2 pS (n = 5), i.e. almost double the estimate in < 10 nM and 1 µM Ca. Therefore, these data suggest that there is an overall increase in as [Ca2+]o is increased into the micromolar range, which is likely to contribute to the facilitatory action of Ca on Icat.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 2. The effect of [Ca2+]o on the single channel conductance of Icat
A, noradrenaline-activated Icat in the presence of various [Ca2+]o was simultaneously low-pass filtered (0·3 Hz, upper trace) and band-pass filtered (0·3-800 Hz, lower trace) to calculate the mean current (I) and the variance (I2), respectively, during the decay of Icat (see Methods for further details). The traces illustrated were obtained in the presence of 100 µM Ca2+o. The presence of noradrenaline is indicated by the open bar. B, in < 10 nM Ca2+o the relationship between I and I2 was always linear so that was calculated using eqns (2) and (4). In the example shown the unitary current was 0·7 pA which gave a value of 12 pS. C, in 3 µM Ca2+o the relationship occasionally deviated from linearity at the peak of Icat, as in this example, suggesting that Po was significant at these current values. Thus the data were fitted with a quadratic function (eqn (3)). In this case i was 1·5 pA which gave a value of 25 pS. D, in 100 µM Ca2+o the relationship was linear although the values at the peak of Icat were much greater than in < 10 nM Ca2+o in B. The i value was 1·4 pA and the value was 23 pS in the example shown. All currents were recorded in standard K+-free conditions at a Vh of -50 mV.
|
It is also possible to obtain an estimate of by analysing the spectral density function of Icat, and therefore we carried out noise analysis to confirm the above data. In addition, noise analysis provides information regarding the number of kinetically resolvable states non-selective cation channels can adopt. Therefore we analysed the spectral density function of Icat to estimate the number of resolvable states and to assess if the equilibrium between these states is affected by [Ca2+]o.
Figure 3A illustrates a typical spectral density function for Icat in < 10 nM Ca, which can be described by the sum of two Lorentzians and thus indicates the presence of three resolvable states. In this example the corner frequency of the lower frequency Lorentzian component (fc1) was 14 Hz and the corner frequency of the higher frequency Lorentzian component (fc2) was 107 Hz. The mean values of fc1 and fc2 are shown as equivalent time constants 1 and 2, respectively, in Table 1. When [Ca2+]o was increased to 1 µM (Fig. 3B) fc1 was much lower whereas fc2 remained unchanged; 4 and 98 Hz, respectively, in the example shown in Fig. 3B. Increasing [Ca2+]o still further (3 and 100 µM) produced no significant further changes in the corner frequency of either component (Fig. 3C and see Table 1 for the mean values of the equivalent time constants). Consequently [Ca2+]o had marked effects on the kinetics of Icat and furthermore [Ca2+]o had a significant effect on the relative contributions of the two Lorentzian components to the total I2. This is quantified as a ratio (G(0) ratio) by dividing G2(0), the zero frequency asymptote of the higher frequency Lorentzian component, by G1(0), the zero frequency asymptote of the lower frequency Lorentzian component. It can be seen that in < 10 nM Ca the G(0) ratio is 0·6 ± 0·09 (n = 6, Table 1 and Fig. 3A), indicating that the higher frequency Lorentzian component contributes to just over half the total variance. However, increasing [Ca2+]o to 1 µM led to a significant fall in the G(0) ratio to 0·16 ± 0·03 (n = 5, Table 1 and Fig. 3B), reflecting a greater contribution of the lower frequency Lorentzian component to the total variance under these conditions. Increasing [Ca2+]o to either 3 or 100 µM Ca lead to a further reduction in the G(0) ratio (Table 1 and Fig. 3C). These changes in the relative amplitudes of the two Lorentzian components suggest that [Ca2+]o affects the equilibrium of the underlying channel states. Estimates of from the spectral density function in various [Ca2+]o gave similar values to those obtained above where total I2 was plotted as a function of I. From the spectral densities was 10·8 ± 1·5 pS (n = 6), 22·2 ± 1·5 pS (n = 5) and 24·1 ± 1·7 pS (n = 6) in < 10 nM, 3 µM and 100 µM Ca, respectively (Table 1). Clearly the similar values of obtained from the spectra and the relationship between I2 and I indicate that the assumption that Po 1 is valid (eqn (7)). Interestingly, in 1 µM Ca (9 ± 1·4 pS, n = 5) is similar to that obtained in < 10 nM Ca, although there is a marked change in the kinetics, which is manifest as an increase in 1 and a reduction in the G(0) ratio (see Table 1). This could mean that the [Ca2+]o-dependent effect on the kinetics of Icat is distinct from its effect on . We propose that both these effects contribute to the facilitatory action of Ca on Icat (see Discussion).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 3. The effect of [Ca2+]o on the spectral density function of Icat
A, a typical example of the spectral density function of Icat in < 10 nM Ca2+o, which was described by the sum of two Lorentzians (continuous line). The corner frequencies of the lower frequency (fc1) and higher frequency (fc2) Lorentzian components (dotted lines) were 14 and 107 Hz, respectively. The zero frequency asymptotes for the lower (G1(0)) and higher frequency (G2(0)) Lorentzian components were 8 × 10-26 and 4·6 × 10-26 A2 Hz-1, respectively, which gave a G(0) ratio of about 0·6. B, a typical example of the spectral density function of Icat in the presence of 1 µM Ca2+o, where fc1 and fc2 were 4 and 98 Hz, respectively. G1(0) and G2(0) were 2 × 10-25 and 4·5 × 10-26 A2 Hz-1, which gave a G(0) ratio of about 0·2. C, a typical example of the spectral density function of Icat in the presence of 100 µM Ca2+o, where fc1 and fc2 were 3 and 125 Hz, respectively. Under these conditions G1(0) was 2·14 × 10-23 A2 Hz-1 and G2(0) was 1·6 × 10-25 A2 Hz-1, which gave a G(0) ratio of 0·007.
|
Relaxation analysis of Icat in various [Ca2+]o
We have shown previously that Icat is voltage dependent so that at potentials positive to Vrev (Vrev is approximately +10 mV; Helliwell & Large, 1996) there is little outward current because the majority of channels are in a non-conducting state, whereas at potentials negative to Vrev a significant proportion of channels are in a conducting state (i.e. the steady-state current-voltage relationship exhibits marked inward rectification; see Wang & Large, 1991; Helliwell & Large, 1996). Therefore it was possible to shift rapidly the equilibrium of conducting to non-conducting states by using a standard voltage protocol where the membrane voltage was stepped from -50 to +50 mV and then back to -30 mV (see Methods for further details). Using this approach it was possible to investigate whether the kinetically distinct states revealed using noise analysis could be observed in voltage-jump relaxations. Of particular interest was to test the hypothesis that the increase in the relative amplitude (G1(0)) and time constant (1) of the lower frequency Lorentzian component in micromolar concentrations of Ca contributed to the facilitatory effect of Ca on Icat. If this was the case then the equivalent relaxation should contribute to the increase in the amplitude of Icat on stepping from +50 to -30 mV.
In < 10 nM Ca there were negligible current relaxations in response to voltage jumps to +50 mV or on stepping back to -30 mV (Fig. 4A). However, in 2 µM Ca there was an obvious relaxation on stepping the voltage to +50 mV which was described by a single exponential (relax,+50 mV, Fig. 4B) of 52 ms. This relaxation represents the switching off of Icat and reflects a shift in the equilibrium from open to closed channel states, since the channels spend a significant amount of time in the conducting state at -50 mV but are primarily in the non-conducting state at +50 mV. On stepping back to -30 mV there was a significant instantaneous current (Iinst,-30 mV) followed by an inward current relaxation (Irelax,-30 mV) which again could be described by a single exponential (relax,-30 mV), where relax,-30 mV was 58 ms (Fig. 4B). This relaxation represents an increase in the amplitude of Icat due to net channel opening because of a shift in the equilibrium of non-conducting to conducting states when the voltage is stepped from +50 to -30 mV. It should be noted that even in 1 µM Ca relaxations were evident although they were too small to obtain reliable estimates of relax,+50 mV and relax,-30 mV. In 3 µM Ca and 100 µM Ca the time course of the relaxations at both +50 and -30 mV was similar to the values observed in 2 µM Ca (Fig. 4C and D; mean data are shown in Table 2). Therefore the time constant of shutting off Icat at positive potentials and switching on Icat on subsequent steps to negative potentials was not dependent on [Ca2+]o. Moreover, the time constant of the inward relaxation at positive potentials was not dependent on the test voltage as the values were 40 ± 3 ms (n = 4), 53 ± 4 ms (n = 12) and 52 ± 3 ms (n = 4) at, respectively, +30, +50 and +70 mV. In contrast, the amplitude of Irelax,-30 mV progressively increased relative to Iinst,-30 mV as [Ca2+]o was increased in the micromolar range. This is evident in Fig. 4B -D, and is shown quantitively in Fig. 5 where it is evident that in < 10 nM Ca, Iinst,-30 mV accounts for all of the inward current. However, as [Ca2+]o is increased the relative contribution of Iinst,-30 mV decreases and there is an increase in the relative contribution of Irelax,-30 mV to the total inward current. These relationships could be empirically described by identical Hill curves (EC50 = 2 µM), although they were of opposite co-operativity, since Irelax,-30 mV increased at the expense of Iinst,-30 mV as [Ca2+]o was increased. Since the EC50 value obtained for the effect of [Ca2+]o on the absolute amplitude of Icat (Fig. 1B) is approximately 3 µM it is likely that Irelax,-30 mV contributes to the facilitatory effect of Ca on the steady-state amplitude of Icat. Furthermore, consistent with our hypothesis from noise analysis, Irelax,-30 mV principally reflects the same kinetic process as the first Lorentzian component, since they have similar time constants in micromolar concentrations of Ca, although 1 from noise analysis (about 45 ms) is slightly faster than relax,-30 mV (about 60 ms). Moreover, the relative amplitudes of both the first Lorentzian component and relaxation at -30 mV increase similarly as [Ca2+]o is increased in the micromolar range, which also supports this conclusion.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 4. The effect of [Ca2+]o on the amplitude of current relaxations
A, in < 10 nM Ca2+o current relaxations were negligible in response to voltage steps. B, in 2 µM Ca2+o relaxations were clearly evident at +50 and -30 mV and were described by single exponentials, relax,+50 mV and relax,-30 mV, which were 52 and 58 ms, respectively, in the example shown. Under these conditions the amplitudes of Iinst,-30 mV and Irelax,-30 mV were similar. C, in 3 µM Ca2+o the time course of the relaxations at +50 and -30 mV were similar to those in B, which were 50 and 60 ms for relax,+50 mV and relax,-30 mV, respectively, in the example shown. However, the amplitude of Irelax,-30 mV was greater than Iinst,-30 mV under these conditions. D, in 100 µM Ca2+o relax,+50 mV and relax,-30 mV were similar to B and C, which were 52 and 62 ms in the example shown. Under these conditions Irelax,-30 mV was responsible for most of the steady-state inward current at -30 mV. The dashed line in all current records is the zero current level and the dotted line indicates the current level from which measurements of Iinst,-30 mV and Irelax,-30 mV were obtained (see Methods for further details). The voltage protocol at the bottom of the figure applies to all current traces. All current records were leak subtracted and obtained in standard K+-free conditions.
|
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 5. Quantification of the effect of [Ca2+]o on the relative amplitudes of Iinst,-30 mV and Irelax,-30 mV
The ratio of either Iinst,-30 mV ( ) or Irelax,-30 mV ( ) to the total steady-state inward current at -30 mV is plotted as a function of [Ca2+]o. Irelax,-30 mV increases at the expense of Iinst,-30 mV as [Ca2+]o is increased. These data can be described by identical Hill curves where the EC50 was 2 µM and nH was 2.
|
Table 2. Time constants of current relaxations to voltage steps in various [Ca2+]o
| [Ca2+]o | relax,+50 mV (ms) | n | relax,-30 mV (ms) | n |
| < 10 nM | * | 6 | * | 6 |
| 2 µM | 46 ± 4 | 4 | 60 ± 8 | 6 |
| 3 µM | 47 ± 3 | 5 | 65 ± 3 | 6 |
| 100 µM | 45 ± 3 | 6 | 60 ± 4 | 5 |
* Negligible relaxation in < 10 nM Ca2+o.
Instantaneous rectification of Icat at +50 mV
If the same processes are involved in controlling the amplitude and time course of the relaxations at both +50 and -30 mV, then one would expect that the amplitude of the relaxations at +50 mV (net closure of channels) and at -30 mV (net opening of channels) to be proportional in a given [Ca2+]o. Indeed, if this is the case Irelax,-30 mV and Irelax,+50 mV should be the same in a given [Ca2+]o, since the driving force at both potentials is the same, assuming a Vrev of +10 mV (see Methods, and Helliwell & Large, 1996). The relationship between Irelax,+50 mV and Irelax,-30 mV in various [Ca2+]o is shown in Fig. 6A, where it can be seen that although there is a marked correlation between the relaxations at +50 and -30 mV, Irelax,+50 mV is approximately half Irelax,-30 mV in a given [Ca2+]o (see Fig. 6A and B). Therefore on stepping from Vh to +50 mV there must be a significant amount of 'instantaneous' rectification. Such rapid processes on stepping to depolarized voltages are generally ascribed to blocking phenomena rather than gating of the channel. One possibility for the [Ca2+]o-independent 'instantaneous' rectification observed in the present study was fast voltage-dependent block by internal Mg2+ (Mg
) as reported previously in, for example, neuronal nicotinic receptors (Sands & Barish, 1992) and K+ channels (Horie et al. 1987). In order to examine this possibility we removed Mg2+ from the pipette solution and chelated contaminating concentrations of Mg2+ with EDTA instead of BAPTA (see Methods). Under these conditions, in the presence of 100 µM Ca, the amplitude of Irelax,+50 mV was very similar to Irelax,-30 mV (Fig. 6C), which can be accounted for by the removal of 'instantaneous' rectification since the ratio of the chord conductance at -50 mV to the instantaneous chord conductance at +50 mV (conductance ratio) was 1·04 ± 0·12 (n = 5). Therefore, in the absence of Mg the instantaneous conductance behaves ohmically (the mean conductance at -50 mV was 1·5 ± 0·37 nS and the mean conductance at +50 mV was 1·74 ± 0·56 nS). This contrasts with the conductance ratio in normal Mg-containing pipette solutions which was 0·64 ± 0·06 (n = 5; the mean conductance at -50 mV was 4·8 ± 1 nS and at +50 mV was 2·8 ± 0·5 nS). However, Mg2+-containing pipette solutions were used in the majority of the present experiments since the amplitude of Icat was generally much smaller in the absence of Mg (e.g. Fig. 6B and C). Furthermore, it should be noted that the removal of Mg had no effect on the relaxation time constants relax,+50 mV or relax,-30 mV, which were 40 ± 4 and 63 ± 10 ms (n = 5), respectively, in 100 µM Ca and thus similar to the time constants obtained in the presence of Mg in the same [Ca2+]o (see above).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 6. The effect of [Mg2+]i on the amplitude of the relaxation at +50 mV (Irelax,+50 mV)
A, correlation plot of Irelax,+50 mV and Irelax,-30 mV at different [Ca2+]o (shown next to the appropriate data point). The straight line through the data was obtained by linear regression (r = 0·99). B, a current record obtained in 100 µM Ca2+o and 1·2 mM Mg2+i (upper trace) in response to voltage steps (lower trace). Under these conditions Irelax,+50 mV is approximately half the amplitude of Irelax,-30 mV. C, a current record obtained (upper trace) using the same voltage protocol (lower trace) in 100 µM Ca2+o and in the absence of Mg2+i. Under these conditions Irelax,+50 mV is of a similar amplitude to Irelax,-30 mV. The current records in B and C have different vertical calibration bars, where Irelax,-30 mV in B has been normalized to the same amplitude as Irelax,-30 mV in C for comparison. The arrows indicate the zero current level and the dashed lines indicate the current level from which current amplitudes were measured (see Methods for further details).
|
|
DISCUSSION |
In the present study we have used noise and voltage-jump analysis to investigate the mechanisms that may be responsible for the facilitatory effect of Ca on noradrenaline-activated Icat in isolated rabbit portal vein smooth muscle cells. The data suggest that external Ca increases the single channel conductance and also produces a change in the kinetics of the underlying channel states.
In < 10 nM and 1 µM Ca the single channel conductance () was about 10 pS but approximately doubled when [Ca2+]o was 
3 µM (about 23 pS). It is likely that this mechanism contributes to the facilitatory action of Ca, since the increase in occurs in a similar [Ca2+]o range as the potentiating effect on the peak amplitude of Icat (EC50 = 3 µM). Our estimates of in [Ca2+]o of 3 µM and above are very similar to that (about 25 pS) of Inoue & Kuriyama (1993) who measured directly from phenylephrine-activated cation channels in the same tissue in the presence of 2 mM Ca. Consequently, the full conductance of about 20-25 pS seems to have a lower limit requirement for [Ca2+]o of around 3 µM but then remains constant as [Ca2+]o is increased above this value.
Experiments where the kinetics of Icat were studied revealed other effects of Ca which might contribute to its facilitatory action. In voltage-jump experiments only instantaneous currents were observed in < 10 nM Ca, both on stepping from the holding potential to +50 mV and on returning to -30 mV. However, in 1-2 µM Ca current relaxations were observed subsequent to the instantaneous current. The relaxations at -30 mV followed an exponential time course with a time constant (relax,-30 mV) of about 60 ms, and the amplitude of these relaxations (Irelax,-30 mV) was increased and there was a concomitant decrease of the instantaneous current as [Ca2+]o was increased. The EC50 value for this relationship between the relative contribution of Irelax,-30 mV to the total inward current and [Ca2+]o was approximately 2 µM. Since the [Ca2+]o dependence of the increase of Irelax,-30 mV had a similar EC50 value to that obtained for the ability of Ca to increase the peak amplitude of Icat (Fig. 1), it is evident that this relaxation is associated with the facilitatory action of Ca on Icat. This [Ca2+]o-dependent increase in the amplitude of Irelax,-30 mV is probably due to an increase in Po in addition to the effect on described above. Even though the effect of [Ca2+]o on Po was not investigated systematically, since it was typically very low in any given [Ca2+]o (i.e the relationship between variance and mean current was linear in most cells), there were indications that this may be the case. Thus, in 3 µM Ca the relationship between variance and mean current occasionally deviated from linearity, suggesting that Po approached 0·5 at the peak of Icat. Furthermore, in both 3 and 100 µM Ca the larger mean currents recorded under these conditions compared with < 10 nM Ca were associated with larger variance values. These data also indicated that Po was greater in micromolar concentrations of Ca than in < 10 nM Ca. Surprisingly, Po was typically low even when high concentrations of noradrenaline were used (100 µM in this study) and when the facilitatory effect of Ca on Icat was maximal. The reason for this is unknown although these observations are consistent with a previous report which showed that Po for phenylephrine-activated cation channels in rabbit portal vein cells was normally less than 0·2 (Inoue & Kuriyama, 1993).
In all [Ca2+]o studied (< 10 nM to 100 µM) the spectral density function of Icat could be described by the sum of two Lorentzians. In < 10 nM Ca the time constants, 1 and 2, associated with these two Lorentzians were about 11 ms and 1 ms, respectively. In 1-100 µM Ca, 1 was increased to 40-50 ms but there was no detectable change in 2 (Table 1). The value of 1 is similar to the time constant of the [Ca2+]o-dependent current relaxation on stepping the voltage from +50 to -30 mV (relax,-30 mV). This indicates that the lower frequency Lorentzian component from noise analysis is kinetically equivalent to Irelax,-30 mV and since this relaxation leads to an increase in the steady-state amplitude of Icat, this Lorentzian component is similarly associated with the facilitatory action of Ca on Icat. Further evidence to support this hypothesis is that the contribution of the first Lorentzian component to the overall power spectrum was much greater when [Ca2+]o was changed from < 10 nM to micromolar concentrations (see Table 1). Clearly this trend parallels the increase in Irelax,-30 mV seen in voltage-jump experiments (Figs 4 and 5).
To obtain initial estimates of the underlying rate constants and to propose a more explicit hypothesis to explain the effects of Ca on the kinetics of Icat would require the formulation of a simple mechanistic model. Since the presence of two Lorentzians in the spectral density function of Icat indicates that the channels exist in at least three states, a simple linear scheme where all three states are inter-related is:
State 1 
State 2 State 3
Scheme 1
A classical interpretation may be that state 1 is the closed conformation and states 2 and 3 are open conformations. In this model the transitions involving states 1 and 2 and those involving states 2 and 3 correspond to the two Lorentzian components associated with the lower and higher frequencies, respectively. Since changes of [Ca2+]o produced marked alterations of the spectral density function, it is evident that at least one of the rate constants in Scheme 1 is Ca2+ dependent so that changes in [Ca2+]o will shift the equilibrium between the proposed states. However, some of the observations are inconsistent with this scheme. First, increasing [Ca2+]o from < 10 nM into the micromolar range significantly increased the time constant (1) and the relative amplitude of the lower frequency Lorentzian component (G(0) ratio) but the time constant (2) of the higher frequency Lorentzian component appeared to be unaffected (approximately 1-2 ms in all [Ca2+]o). Clearly Scheme 1 predicts that both 1 and 2 should be affected since these macroscopic time constants depend on all the rate constants in the sequence (see e.g. Colquhoun, 1995). Second, increasing [Ca2+]o from 1 to 3 µM led to a marked increase in the relative amplitude of the first Lorentzian component (G(0) ratio was reduced from 0·12 to 0·01) without any detectable change in either 1 or 2. This observation would also be difficult to envisage in terms of Scheme 1 where the relative amplitudes of the two Lorentzians are predicted to change according to an [Ca2+]o-dependent change in the rate constants. Thus at present the data cannot be quantitatively described by a simple linear scheme. It should be noted that since in 1 µM Ca is similar to that obtained in < 10 nM Ca, although there is an increase in both 1 and the relative amplitude of the lower frequency Lorentzian component, it is possible that the [Ca2+]o-dependent effect on the kinetics of Icat is distinct from its effect on .
In conclusion our data suggest that the facilitatory effect of Ca on Icat results from an increase in , and a change in the kinetics of the underlying cation channels, which may increase Po. Clearly these ideas will require testing at the single channel level when this becomes routinely possible.
|
REFERENCES |
| Amédée, T. & Large, W. A. (1989). Microelectrode study on the ionic mechanisms which contribute to the noradrenaline-induced depolarization in isolated cells of rabbit portal vein. British Journal of Pharmacology 97, 1331-1337 |
[Medline] |
| Byrne, N. G. & Large, W. A. (1988). Membrane ionic mechanisms activated by noradrenaline in cells isolated from the rabbit portal vein. The Journal of Physiology 404, 557-573 |
[Abstract] |
| Colquhoun, D. (1995). The principles of the stochastic interpretation of ion-channel mechanisms. In Single Channel Recording, 2nd edn, ed. Sakmann, B. & Neher, E., pp. 397-482. Plenum Press, New York. |
|
| Helliwell, R. M. & Large, W. A. (1996). Dual effect of external Ca2+ on noradrenaline-activated cation current in rabbit portal vein smooth muscle cells. The Journal of Physiology 492, 75-88 |
[Abstract] |
Helliwell, R. M. & Large, W. A. (1997).  -Adrenoceptor activation of a non-selective cation current in rabbit portal vein by 1,2-diacyl-sn-glycerol. The Journal of Physiology 499, 417-428 |
[Abstract] |
| Horie, M., Irisawa, H. & Noma, A. (1987). Voltage-dependent magnesium block of adenosine-triphosphate-sensitive potassium channel in guinea-pig ventricular cells. The Journal of Physiology 387, 251-272 |
[Abstract] |
| Inoue, R. & Kuriyama, H. (1993). Dual regulation of cation-selective channels by muscarinic and 1-adrenergic receptors in the rabbit portal vein. The Journal of Physiology 465, 427-448 |
[Abstract] |
| Sands, S. B. & Barish, M. E. (1992). Neuronal nicotinic acetylcholine receptor currents in phaeochromocytoma (PC12) cells: dual mechanisms of rectification. The Journal of Physiology 447, 467-487 |
[Abstract] |
| Sigworth, F. J. (1980). The variance of sodium current fluctuations at the node of Ranvier. The Journal of Physiology 307, 97-129 |
[Medline] |
| Wang, Q. & Large, W. A. (1991). Noradrenaline-evoked cation conductance recorded with the nystatin whole-cell method in rabbit portal vein cells. The Journal of Physiology 435, 21-39 |
[Abstract] |
|
|
Acknowledgements
This work was supported by The Wellcome Trust. We are grateful to Ade Aromolaran for carrying out a few of the experiments.
Corresponding author
W. A. Large: Department of Pharmacology, St George's Hospital Medical School, Cranmer Terrace, London SW17 0RE, UK.
Email: w.large{at}sghms.ac.uk
This article has been cited by other articles:
|
|
|
|
 
H.-W. Cheng, A.F. James, R.R. Foster, J.C. Hancox, and D.O. Bates
VEGF Activates Receptor-Operated Cation Channels in Human Microvascular Endothelial Cells
Arterioscler. Thromb. Vasc. Biol.,
August 1, 2006;
26(8):
1768 - 1776.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
Y. Maruyama, Y. Nakanishi, E. J. Walsh, D. P. Wilson, D. G. Welsh, and W. C. Cole
Heteromultimeric TRPC6-TRPC7 Channels Contribute to Arginine Vasopressin-Induced Cation Current of A7r5 Vascular Smooth Muscle Cells
Circ. Res.,
June 23, 2006;
98(12):
1520 - 1527.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. P. Albert, V. Pucovsky, S. A. Prestwich, and W. A. Large
TRPC3 properties of a native constitutively active Ca2+-permeable cation channel in rabbit ear artery myocytes
J. Physiol.,
March 1, 2006;
571(2):
361 - 369.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. Shi, E. Mori, Y. Mori, M. Mori, J. Li, Y. Ito, and R. Inoue
Multiple regulation by calcium of murine homologues of transient receptor potential proteins TRPC6 and TRPC7 expressed in HEK293 cells
J. Physiol.,
December 1, 2004;
561(2):
415 - 432.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. J. Beech, K. Muraki, and R. Flemming
Non-selective cationic channels of smooth muscle and the mammalian homologues of Drosophila TRP
J. Physiol.,
September 15, 2004;
559(3):
685 - 706.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. C. M. Choisy, J. C. Hancox, L. A. Arberry, A. M. Reynolds, M. J. Shattock, and A. F. James
Evidence for a Novel K+ Channel Modulated by {alpha}1A-Adrenoceptors in Cardiac Myocytes
Mol. Pharmacol.,
September 1, 2004;
66(3):
735 - 748.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. Inoue, T. Okada, H. Onoue, Y. Hara, S. Shimizu, S. Naitoh, Y. Ito, and Y. Mori
The Transient Receptor Potential Protein Homologue TRP6 Is the Essential Component of Vascular {{alpha}}1-Adrenoceptor-Activated Ca2+-Permeable Cation Channel
Circ. Res.,
February 16, 2001;
88(3):
325 - 332.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Okada, R. Inoue, K. Yamazaki, A. Maeda, T. Kurosaki, T. Yamakuni, I. Tanaka, S. Shimizu, K. Ikenaka, K. Imoto, et al.
Molecular and Functional Characterization of a Novel Mouse Transient Receptor Potential Protein Homologue TRP7. Ca2+-PERMEABLE CATION CHANNEL THAT IS CONSTITUTIVELY ACTIVATED AND ENHANCED BY STIMULATION OF G PROTEIN-COUPLED RECEPTOR
J. Biol. Chem.,
September 24, 1999;
274(39):
27359 - 27370.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Jung, R. Strotmann, G. Schultz, and T. D. Plant
TRPC6 is a candidate channel involved in receptor-stimulated cation currents in A7r5 smooth muscle cells
Am J Physiol Cell Physiol,
February 1, 2002;
282(2):
C347 - C359.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Introduction
