- Email: [email protected]
Characterization of a delayed rectifier potassium channel in the slowly adapting stretch receptor neuron of crayfish
Brain Research 913 (2001) 1–9 www.elsevier.com / locate / bres
Research report
Characterization of a delayed rectifier potassium channel in the slowly adapting stretch receptor neuron of crayfish Jia-Hui Lin, Bo Rydqvist* Department of Physiology and Pharmacology, Karolinska Institutet, S-171 77, Stockholm, Sweden Accepted 19 June 2001
Abstract Single channel recordings were performed on enzyme-cleaned slowly adapting sensory neurons of crayfish, in cell-attached configuration, with a physiological K 1 gradient across the neuronal membrane. An outward rectifying, voltage-gated K 1 channel with a slope conductance of 13 pS and a K 1 ion permeability of PK 56.5310 214 cm 3 / s was characterized. This 13 pS K 1 channel started to be activated at around 20 mV depolarization. Its open probability increased upon depolarization with V0.5 5225.3 mV and Pmax 50.83. The averaged currents showed a delay following the onset of depolarization. The activation time constant was voltage-dependent. The maximal value was 17.0 ms at 225 mV and at 135 mV the time constant was 1.7 ms. Little inactivation was observed throughout the 80or 1500-ms long depolarization pulses. A sum of two exponentials provided the optimal fit for open time and closed time distribution. At 80-mV depolarization, the open time constants were 0.4 and 10.4 ms; the close time constants were 0.4 and 2.3 ms. The first-latency distribution suggested that at least two closed states preceded two open states. This 13 pS delayed rectifier plays a minor role in the maintenance of the resting membrane potential but contributes to the action potential repolarization. It may also modify the stretch-induced receptor potential and affect the adaptation behaviours in this neuron. 2001 Elsevier Science B.V. All rights reserved. Theme: Sensory systems Topic: Invertebrate sensory systems Keywords: Delayed rectifier; Voltage-gated K 1 channel; Single channel recording; Patch clamp; Slowly adapting receptor neuron
1. Introduction Potassium channels are ubiquitous in biological membranes. A large amount of different K 1 channels are found with important biological functions, such as encoding and modifying trains of action potentials in excitable cells. Among them, the voltage-gated delayed rectifier K 1 channels (Kv) play an important role in regulation of the repolarization phase of action potentials in excitable cells [8]. In the crustacean abdominal stretch receptor organ, which consists of a slowly adapting (SA) and rapidly adapting (RA) receptor neuron, the voltage-dependent potassium permeability system has been suggested to be one of the factors determining the difference in action potentials and adaptive behaviours between these two *Corresponding author. Tel.: 146-8-728-7267; fax: 146-8-32-7026. E-mail address: [email protected] (B. Rydqvist).
mechanoreceptors [2,19,20]. The outward currents in response to depolarizing potential steps have been shown to be carried by K 1 ions and to be delayed rectifying. At least two different voltage-gated K 1 channels have been suggested to be present in the neurons, according to their different activation–inactivation kinetics and sensitivities to extracellular blockers such as tetraethylammonium (TEA) and 4-aminopyridine (4-AP) [13,18]. This knowledge was derived from studies of macroscopic ionic currents using the two-electrode intracellular voltage clamp and macropatch clamp techniques. To determine in more detail the properties of the macroscopic K 1 currents, it is necessary to characterize individual K 1 channels and the contributions of the different K 1 channels to the macrocurrents. In addition to the single stretchactivated channels in SA and RA neurons [6], a Ca 21 activated K 1 channel (K 21 Ca ) with single channel conduct21 ance of 87 pS has been suggested [7]. This K Ca channel 21 was found to be activated by Ca influx through stretch-
0006-8993 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 01 )02737-8
2
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
activated channels. Due to the considerable difficulty in exposing the neuronal membrane of the stretch sensory neuron to patch clamping, little is know about the properties of voltage-gated single K 1 channels in this preparation, which are supposed to account for the delayed rectifying whole cell currents. We therefore performed single channel recordings of the voltage-gated K 1 channels in SA neuronal membrane, which could be patched after enzymatic treatment. Several different kinds of K 1 channels were observed during the study, and in this paper we focused on characterizing a 13 pS delayed rectifier K 1 channel on the basis of its singlechannel properties such as permeability / conductance, kinetics, open probability and its physiological role when compared to the macrocurrents obtained from two-electrode voltage clamp recordings and macropatch recordings [13,19].
2. Material and methods The slowly adapting (SA) stretch receptor neuron of crayfish (Pacifastacus leniusculus) was dissected free and mounted with silver clips on the bottom of a 35-mm plastic Petri dish covered with a thin layer of Sylgard 184 (Dow Corning, Seneffe, Belgium). The chamber was placed on an inverted microscope (IM-35, Zeiss, Oberkochen, Germany), and perfused with the normal astacus saline (NAS) at room temperature (20–248C). The NAS solution contained (in mM): 200 NaCl, 5.4 KCl, 13.5 CaCl 2 , 2.6 MgCl 2 , 10 Hepes, pH adjusted to 7.4 with NaOH. The osmolality of NAS was 420615 mOsm / kg, which is similar to the cytosolic osmolality. In order to expose the neuronal membrane, the glial cells and connective tissues which cover the soma were removed enzymatically by lightly pressing the tip of a fire-polished pipette (having a diameter of 30–50 mm) over the soma for 30–70 min, depending on the extent of glial cells and connective tissue. The pipettes contained an enzyme mixture consisting of: protease (type: XIV), 20 mg / ml; collagenase / dispase (EC 3.4.24.3), 10 mg / ml; hyaluronidase (EC 3.2.1.35), 7 mg / ml and elastase (EC 3.4.21.36), 1 mg / ml [13]. Additional mechanical agitation was applied after the enzymatic treatment to remove the debris by blowing and sucking the bath solution over the enzyme-treated area. All enzymes were purchased from Sigma (St Louis, MO, USA).
2.1. Single channel recording Single channel K 1 currents were recorded in the cellattached configuration of the patch clamp technique. Patch pipettes were pulled from aluminosilicate glass (Hilgenberg, Malsfeld, Germany; o.d. 1.6 mm, i.d. 1.28 mm), coated with Sylgard 184 and fire polished. Pipettes were
filled with NAS and had a resistance of 2–5 MV. During some experiments, 0.3 mM tetrodotoxin (TTX, Sigma) was included in pipette solution to block voltage-gated Na 1 channels. The patch pipette was positioned on the enzymetreated part of the neuronal soma. A gigaohm seal of 10–50 GV was formed by suction applied to the pipette. As was previously discussed [13] only cells which showed action potentials, were used. We also consider this a strong support for an intact membrane including channel proteins. If action potentials can be seen the resting membrane must be normal and consequently the K 1 leak channels intact. If the leak channels were damaged over the entire enzymetreated area (30–40 mm in diameter) the resting membrane would have deteriorated. This also implies that other channels are not affected. Currents were recorded using an Axopatch 200 amplifier and pClamp 6 programs (Axon Instruments, Foster City, CA, USA), and stored in a computer. Either ramps or steps were applied as voltage clamp commands. Data were filtered at 1–2 kHz by a 23 dB, eight-pole low-pass Bessel filter and sampled at 10 kHz. All analysis was made with pClamp 6 programs. When constructing averaged currents, leakage and capacitative currents were digitally removed by subtraction of blank sweeps with no channel activities. Single channel amplitudes were determined either by onscreen measurement of 5–10 events via cursors, or by forming an all-points histogram of the baseline and openlevel data points, and then fitting with Gaussians by a Marquardt least-squares algorithm (Marquardt-LSQ) [15]. Transitions between open and closed states were detected using a half-amplitude threshold criterion. The dwell time and first-latency histograms were formed by reconstructed current traces resulting from 80-ms long steps to different potentials. The number of channels present in a patch was assumed to be the observed maximal number of open channels within the patch. It was further assumed that when the potential was stepped to a new value, the rate constants for the channel kinetics immediately attained new values, which resulted in time constants in the open and closed time histograms. Since no inactivation was observed within 1500 ms, we considered the 80-ms pulses adequate for the present analysis. Potentials were expressed as the potentials of the interior face relative to the exterior face of the membrane. Outward currents are shown as upward deflections. Vh is referred to as the holding potential (2Vpipette in cell-attached configuration), and Vm the membrane potential across the patch. A resting membrane potential of 265 mV was assumed when analysing the data from cell-attached recordings, as determined in earlier results obtained either by two-electrode intracellular current clamp or by whole-cell patch clamp recordings [13]. The reference potential for all measurements was corrected to be zero before establishing the seal. Junction potential calculated by the Axoscope program (Axon Instruments) was less than 2 mV, and no correction was made.
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
Results throughout this paper are mean6S.E.M., unless otherwise indicated.
given
3
as
3. Results The single channel current described in this paper was investigated in cell-attached configuration with normal astacus saline (NAS, [K 1 ] o 55.4 mM) in the patch pipette. Cell-attached configuration was used to keep the K 1 ionic gradient and other cytosolic components as physiological as possible to mimic the normal channel environment. These arrangements also made it easier to relate the singlechannel data with results of macrocurrents recorded under similar experimental conditions by two-electrode voltage clamp or macropatch recordings [13,19]. Whether the patches were from the stretch receptor neuron or from the glial cells was decided based on the electrical excitability in cell-attached configuration or whole-cell configuration after the patch was ruptured [6,13]. Patches from neuronal membrane showed spontaneous action potentials, or initiated action potentials in response to bath application of 16 mM K 1 -NAS (the [K 1 ] in NAS was increased to 16 mM by the addition of an appropriate amount of 3 M KCl), which depolarized the cell by up to 25 mV, sufficient for the initiation of action potentials. Only those neurons that showed action potentials were used for single-channel studies [13]. After experiments some patches were ruptured into whole-cell configuration. The observation of intracellular action potentials further confirmed that the recordings had been obtained from the neuronal membrane.
3.1. Single channel conductance Applying depolarizing voltage steps to a cell-attached patch on neuronal membrane resulted in the activation of a voltage-gated ion channel, as illustrated in Fig. 1a. When the patch was hyperpolarized or at its resting membrane potential (Vh 50 mV), no single channel events could be discerned. This channel was activated by at least 20 mV depolarization, and distinct single channel events could then be observed. With further depolarization, the amplitude of the channel current increased and the channel became more active. At highly depolarized potentials the channel remained active throughout the depolarizing pulse (1500 ms) and little inactivation was seen. It also exhibited long openings with some short closings during channel opening. This kind of channel was frequently observed during the experiments. It was usually possible to observe channel activity immediately or soon after seal formation when depolarization was applied. The channel activity persisted for several minutes with no discernible changes in channel behaviour. In all patches analysed, only one or two of these channels were present. However, we also observed
Fig. 1. Single channel currents recorded from a cell-attached patch on neuronal membrane with NAS (5.4 mM K 1 ) in the pipette. (a) Representative currents at different holding potentials (Vh ) as indicated on the left. The patch was held at resting potential and activated by step depolarization. Each of the illustrated recordings is part of a 1500-ms episode. Data were acquired at the sampling frequency of 10 kHz and filtered at 1 kHz. The dashed lines mark the closed state of the channel. (b) Single channel current–voltage relationship. The unitary conductance from a linear regression between 240 and 140 mV is 15.5 pS. The smooth curve is obtained by the best fit with GHK equation; Pk 55.43 10 214 cm 3 / s. Same patch as in (a).
patches with more than two conductance levels of this channel type being simultaneously active, or patches containing several conductance levels, indicating the presence of different types of channels in addition to the channel discussed in this paper. Fig. 1b illustrates the I–V plot of single channel current amplitudes as a function of membrane potentials (Vm ) when a resting membrane potential of 265 mV was assumed. The I–V curve shows the voltage dependence of the current amplitude and its outward rectification. With the physiological K 1 gradient across the membrane, the unitary conductance for this patch was 15.5 pS between
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
4
membrane potentials of 240 and 140 mV. The average of nine such patches yielded a unitary conductance of 13.260.8 pS (n59). The Goldman–Hodgkin–Katz (GHK) equation which is valid under the assumption of a constant field within the membrane and independence of ion movements in both direction, was used to fit the I–V curve: IK 5 PK (EF 2 /RT )([K 1 ] o 2 [K 1 ] i exp(EF /(RT ))) / (1 2 exp(EF /(RT )))
possible inactivation of this K 1 channel. The channel exhibited some long openings and closings, as well as some fast events. The bottom trace illustrates the ensemble-averaged current. Following the onset of the depolarization, the current showed a delay in activation and did not inactivate much during the 80-ms depolarizing pulse, properties similar in time course to the whole-cell delayed 1 rectifier K current defined in giant squid axon [9].
(1)
3.2. Single channel kinetics where PK is the K 1 permeability (in cm 3 / s); E is the membrane potential; F, R, T have their usual meanings. [K 1 ] o and [K 1 ] i are the extracellular K 1 (5.4 mM) and the intracellular K 1 concentration (assumed to be 180 mM [19]), respectively. The GHK equation (smooth line in Fig. 1b) provided a reasonable description of the I–V relationship for the data, suggesting that this channel was K 1 selective. The PK value obtained from the fit was 5.43 214 3 10 cm / s in this patch, and the average PK value was 6.560.3310 214 cm 3 / s (n59). Fig. 2 illustrates five representative sweeps for the first opening elicited by a 80-mV depolarization, which was preceded by a 40-mV hyperpolarizing step to remove
Kinetic analysis of single channel activity was performed at different depolarized potential levels using the repetitive voltage step paradigm of Fig. 2. Only patches that contained one active channel were analysed for single channel kinetics. Fig. 3a,b illustrates the open time and closed time histograms of single channel events at a 80-mV depolarization (Vh 5 180 mV). Both the open time and closed time distributions were optimally fitted by two exponentials (Marquardt-LSQ), yielding one fast and one slow time constant. The open time constants at 80-mV depolarization for this patch were 0.460.2 and 9.660.0 ms (mean6variance, 2131 events); the averaged values from three patches were 0.460.1 ms (n53) and 10.460.7 ms (n53), respectively. At the same depolarization, the slow time constant for closed time averaged 2.360.2 ms (n53) and the fast time constant was 0.460.0 ms (n53). Fitting with two exponentials provided a good description to all the dwell time distributions at holding potentials from 120 to 1100 mV. Because the number of exponentials reflects the minimal number of channel states [10], these data suggested that this kind of K 1 channel has at least two open states and two closed states. For voltage-gated channels, the latency from the onset of a depolarization to the first channel opening gives information about the number of closed states preceding the opening of the channel. Fig. 3c,d shows the firstlatency histogram of the opening of this K 1 channel in response to a 80-mV depolarization. In Fig. 3c, the distribution peaks at a time later than zero, suggesting that the channel had to go through more than one closed state to reach the first open states [10]. Since at least two closed states and two open states have been derived from dwell time histograms (Fig. 3a,b), the simplest kinetic scheme to describe the behaviour of this delayed rectifier could be: k2
k1
k 22
k 21
C2 ⇔C1 ⇔O1 ⇔O2
Fig. 2. Representative sweeps initiated by the voltage step (top) after capacitative and leak currents were subtracted. In this patch, two openings of this K 1 channel with similar behaviour were observed. The sweeps illustrated here display the first open level only for clarity (more explanation in Fig. 4c). The ensemble-averaged current (187 sweeps, bottom) had a delay in activation and did not inactivate much during the depolarization. The sample frequency was 10 kHz; filter frequency was 1 kHz; pulses were applied every 0.6 s.
in which two closed states (C1 , C2 ) precede two open states (O1 , O2 ), and k represents the rate constant. The probability density function for the distribution of latency may be obtained as the time derivative of the expression [23]: f(t) 5 (R 1 R 2 )[exp(2R 2 t) 2 exp(2R 1 t)] /(R 1 2 R 2 ) where R 1 and R 2 are related to the rate constants as:
(2)
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
5
Fig. 3. Kinetic analysis of single channel activity. (a, b) Dwell time distribution of channel opening and closing upon a 80-mV depolarization. In total, 2131 events were examined. Both open time and closed time histograms were fitted by two exponentials, yielding 0.460.2 and 9.660.0 ms (mean6variance) for open time constants, and 0.360.3 and 2.260.0 ms for closed time constants, respectively. (c, d) The first-latency distribution at 180 mV depolarization. The curve in (c) was drawn to Eq. (2), scaled to the same area as the histogram. The fit gave R 1 50.63 ms 21 and R 2 50.46 ms 21 . The curve in the cumulative latency histogram of (d) is the integral to Eq. (2). Data are from the same patch.
R 1 , R 2 5 hk 1 1 k 2 1 k 22 6[(k 1 1 k 2 1 k 22 )2 2 4k 1 k 2 ] 0.5 j / 2
Fig. 3c shows the least-squares fit obtained with Eq. (2), and Fig. 3d shows the cumulative latency histogram together with a line drawn to the integral of Eq. (2). Curve in Fig. 3c displays a fast rise phase with a slow decay, and it peaks at a time of 1.9 ms (t m ) when R 1 50.63 ms 21 and R 2 50.46 ms 21 . The mean latency was 11.3 ms for the patch shown; the averaged mean latency from three patches was 10.262.2 ms. These results support the notion that the first open state was actually preceded by at least two closed states. Although Eq. (2) gave a reasonable description to the data, the predicted peak occurred earlier than the experimental peak and some events with long latency have not been included (Fig. 3c) [1,14].
and Fig. 4b is the plot of these averaged current amplitudes versus membrane potentials. At the resting membrane potential or hyperpolarized potentials no outward currents could be detected. Averaged currents were observed only when the patch was depolarized by about 20 mV, and were highly dependent on the voltage. Following the onset of depolarization some delay in activation was present at all potentials, as illustrated in Fig. 2. No inactivation during the 80-ms-long pulses could be seen at any potentials. In this patch, two openings of the K 1 channel with similar channel behaviour were observed, as illustrated in the inset of Fig. 4c. The all-points histogram (Fig. 4c) could be optimally fitted by a third-order Gaussian distribution. Peak 1 and peak 2, corresponding to the two open levels, were evenly located, and had similar amplitude of 1.50 and 1.44 pA. No additional humps that would indicate the presence of another population of channels could be detected. The value of the open probability (Po ) was estimated using Eq. (5) and plotted as shown in Fig. 4d
3.3. Channel open probability
Po 5 I /(i 3 N)
Fig. 4a shows the ensemble-averaged currents initiated by depolarization of 20–100 mV obtained from one patch,
where I is the averaged current, i is the single channel current, and N is the observed maximal number of open
(3) This function also assumes a maximum at time t m [14] t m 5 (R 1 2 R 2 )21 ln(R 1 /R 2 )
(4)
(5)
6
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
Fig. 4. Analysis of ensemble-averaged currents. (a) Averaged currents initiated by depolarization of 20–100 mV from the resting membrane potential as indicated on the left. The number of sweeps included in each trace was: 165 sweeps for 0 mV, five sweeps for 20 mV, 111 sweeps for 40 mV, 194 sweeps for 60 mV, 105 sweeps for 80 mV and 196 sweeps for 100 mV. (b) The voltage dependence of the averaged currents plotted against the membrane potentials. (c) All-points amplitude histogram constructed from the recording at 80-mV depolarization, and optimally fitted by a third-order Gaussian distribution. The peaks corresponding to the closed level (0), first opening (1) and second opening (2) are indicated. Inset is a representative trace, showing the presence of two openings of this K 1 channel type in the patch. (d) The plot of the open probability (Po ) versus the membrane potentials. The smooth line is a fit of the Boltzmann equation to the experimental points, yielding the following values: Pmax 50.83, V0.5 5 225.3 mV and k514.2 mV. All data are from the same patch.
channels within the patch, which was two for this patch. The open probability was voltage-dependent. The relationship between voltage and open probability was sigmoidal with the voltage dependence being steepest with moderate depolarizations. Fitting the data by the Boltzmann equation: Po 5 Pmax /(1 1 exp((V0.5 2Vm ) /k))
(6)
where Pmax is the maximal open probability, V0.5 is the potential at half-maximal probability, Vm is the membrane potential, and k is the slope factor reflecting the steepness of the voltage dependence, we found Pmax 50.83, V0.5 5 2 25.3 mV and k514.2 mV when 265 mV was taken as the resting membrane potential.
3.4. Activation time constants of the K 1 current The activation time constants for this K 1 channel type were analysed by fitting the ensemble-averaged currents at different potentials to a second-order exponential function: I 5 b 3 (1 2 exp(2t /tn ))2 1 a
(7)
where a and b are the scaling parameter and tn is the activation time constant. Fig. 5a shows that such fitting with a power of two (superimposed dotted line) gave an adequate description to the averaged current (cf. [19]) in response to a 100-mV depolarization, yielding an activation time constant tn of 1.75 ms. The voltage dependence of the
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
7
sive. Further studies using excised patches are needed to pharmacologically characterize this K 1 channel.
4. Discussion
Fig. 5. Activation time constants of averaged currents for this 13 pS K 1 channel. (a) A second-order exponential function was used to fit the activation phase of the averaged current initiated by a 100-mV depolarization, producing b53.0 pA and tn 51.75 ms. Same data as in Fig. 4a. (b) The voltage dependence of the activation time constants tn versus the membrane potentials. Data were pooled from the patch shown in Fig. 4a (d) and four other patches (m, ., 쏆, j). The smooth line was computed according to the mean values at each potential.
activation time constants was illustrated in Fig. 5b. The constant approached the maximal value in the potential range of 240 to 210 mV, which was about 17.0 ms at 225 mV. It then decreased on either side. Towards positive potentials the activation time constant became much smaller; being 8.4 ms at 25 mV, 4.0 ms at 115 mV and 1.7 ms at 135 mV respectively. At potentials more positive than 135 mV, it is likely that this activation time constant would reach a stable level below 1 ms. The classical K 1 channel blocker TEA can act from both the internal and external side of the membrane but does not permeate the membrane. In cell-attached configuration, application of TEA to the pipette solution after control experiments with the same patch cannot be easily established. 4-AP is membrane permeable and can block variable K 1 channels from internal as well as external sites [8]. It could thus be applied to the bath while recording in cell-attached configuration. However the effects of the bath application of 5 mM 4-AP on this channel were inconclu-
In the present study which was performed on enzymetreated slowly adapting stretch receptor neurons, a delayed rectifier K 1 channel was characterized in cell-attached configuration with a physiological K 1 gradient across the membrane. Both its single channel amplitude (Fig. 1) and open probability (Fig. 4d) increased upon depolarization with a selectivity for K 1 ions. A sum of two exponentials provided the best fit for closed time and open time histograms. The analysis of the first-latency distribution indicated that at least two closed states (C1 and C2 ) preceded two open states (O1 and O2 ) (Fig. 3). The ensemble-averaged currents showed a delay following the onset of depolarization with little inactivation (Figs. 2 and 4a). The activation time constant for the K 1 averaged currents was highly voltage-dependent (Fig. 5). This delayed rectified K 1 channel had a conductance of 13 pS, in agreement with those reported in other preparations. Conti and Neher [4] showed a delayed rectifier K 1 channel in squid axon having a conductance of 17 pS, while Standen et al. [23] recorded a 15 pS delayed rectifier in frog skeletal muscle cells. Similar conductance for this delayed rectifier was also described in rat hippocampal neurons (9.5 pS [16]), mouse macrophages (16 pS [25]), embryonal carcinoma cells (10 pS [5]), PC 12 cells (7 pS [11]) and neocortical pyramidal neurons (17 pS [12]). However, in some preparations, a much higher conductance for delayed rectifier K 1 channel has been reported, such as that of rat dorsal ganglion neurons (55 pS [21]), LNCaP human prostate cancer cell (78 pS [22]) and rabbit portal vein smooth muscle (42 pS [17]). Therefore, single channel conductance less than 20 pS for the delayed rectifier K 1 channel seems to be more common in most preparations including the stretch receptor neuron, even though high conductance channels are also present in certain species. The 13 pS delayed rectifier described in this paper was activated by |20 mV depolarization, and no current was detected at the resting or hyperpolarized potentials. Its open probability increased with depolarization (V0.5 5 2 25.3 mV) and approached the maximal value (Pmax 50.83) at high positive potentials, suggesting that this channel played a minor role in the maintenance of the resting membrane potential but probably contributes to the repolarization of action potentials. Furthermore, no inactivation was observed within 80 ms for this 13 pS K 1 channel whereas the whole cell current in this preparation inactivates significantly [19]. This channel was also frequently observed during the experiments, suggesting that its distribution over the neuronal soma may be widespread and
8
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
its channel density high. It is therefore likely that its contribution to the repolarization of action potentials is considerable and that it accelerates membrane repolarization at high positive potentials. Further, since this 13 pS delayed rectifier could be activated by |20 mV depolarization, it may also interact with the stretch-induced permeability change by making the repolarization of the receptor potential faster. Such modification in stretchinduced receptor potential may affect the adaptation behaviour in the SA sensory neuron [24]. Both the activation phase of the averaged currents for this 13 pS K 1 channels and that of the whole cell K 1 currents recorded by two-electrode voltage clamp, could be best fitted by a second-order exponential function (Fig. 5a; [19]). The activation time constant for the 13 pS K 1 current was about 6 ms at 0 mV (Fig. 5b), much slower in comparison to 1 ms for the whole cell K 1 current at 0 mV. At high positive potentials the difference becomes less. The whole cell K 1 currents also exhibit a marked inactivation within 50 ms, whereas the 13 pS K 1 averaged currents were not inactivated within 80 ms. However, K 1 macrocurrents obtained by macropatch recordings show both fast inactivation in some patches and little inactivation in other patches [13]. Since other K 1 channels with different properties were actually observed during this study, it is conceivable that this 13 pS delayed rectifier is one of several types of K 1 channels contributing to the whole cell K 1 currents. Other K 1 channels with faster activation as well as faster inactivation could be expected. The distribution of these 13 pS K 1 channels over neuronal soma may be widespread but uneven; they might aggregate with each other in some parts of the soma whereas in other locations they might cluster with other different types of K 1 channels, as suggested by the macropatch recordings [13]. Whether the 13 pS K 1 channel is the dominant one or how much it contributes to the total K 1 current cannot be determined at present. Whole cell currents recorded using intracellular micropipettes [18] showed that ca. 25% of the peak outward current and a major part of the outward current at 500 ms was TEA sensitive indicating that an outward rectifier constitutes a dominant part of the K 1 channel population. Whether the 13 pS channel is the only channel contributing to this outward rectifier cannot be concluded at present. Further work is needed to define other K 1 channels and their contribution to the total macroscopic K 1 current. Among voltage-gated K 1 channel proteins (Kv family), eight subfamilies (Kv1–Kv6 and Kv8–Kv9) have been cloned in heterologous expression systems [3]. In comparison with these cloned K 1 channels, the 13 pS K 1 channel found in crayfish stretch sensory neuron displays properties similar to Kv1.2 channel in activation threshold (Von ), k, inactivation process and single channel conductance, which has Von of 240 mV, k of 13 mV, single channel conductance of 9.2–17 pS and very slow inactivation [3]. However, there is some discrepancy in the half-activation
potential (V0.5 ) with a V0.5 of 25 to 5 mV for Kv1.2 channel and a V0.5 of 225 mV for this 13 pS delayed rectifier found in crayfish SA receptor neuron. Therefore, a detailed molecular characterization of the channel proteins underlying the 13 pS delayed rectifier will be helpful in providing more insights about this channel.
Acknowledgements We are grateful to Drs Fredrik Elinder and Joe Bruton for reading the manuscript and for fruitful suggestions. This work was supported by grants from the Swedish Medical Research Council (Project No. 6838) and Karolinska Institutet.
References [1] K. Bokvist, P. Rorsman, P.A. Smith, Effects of external tetraethylammonium ions and quinine on delayed rectifying K 1 channels inmousepancreaticb-cell,J.Physiol.423(1990)311–325. [2] N.M. Brown, D. Ottoson, B. Rydqvist, Crayfish stretch receptor: an investigation with voltage-clamp and ion sensitive electrodes, J. Physiol. 284 (1978) 155–179. [3] W.A. Coetzee, Y. Amarillo, J. Chiu, A. Chow, D. Lau et al., Molecular diversity of K 1 channels, Ann. NY Acad. Sci. 868 (1999) 233–285. [4] F. Conti, E. Neher, Single channel recordings of K 1 currents in squid axons, Nature 288 (1980) 140–143. [5] L. Ebihara, W.C. Speers, Ionic channels in a line of embryonal carcinoma cells induced to undergo neuronal differentiation, Biophys. J. 46 (1984) 827–830. [6] C. Erxleben, Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish, J. Gen. Physiol. 94 (1989) 1071–1083. [7] C. Erxleben, Calcium influx through stretch-activated cation channels mediates adaptation by potassium current activation, Neuroreport 4 (1993) 616–618. [8] B. Hille, Ionic Channels of Excitable Membrane, Sinauer Associates Inc., Sunderland, MA, 1992. [9] A.L. Hodgkin, A.F. Huxley, The component of membrane conductance in the giant axon of Loligo, J. Physiol. 116 (1952) 473–496. [10] R. Horn, N.B. Standen, Counting kinetic states: the single channel approach, in: A. Grinell, W. Moody (Eds.), The Physiology of Excitable Cells, Alan R. Liss Inc., New York, 1983, pp. 181–189. [11] T. Hoshi, R.W. Aldrich, Voltage-dependent K 1 currents and underlying single K 1 channels in phenochromocytoma cells, J. Gen. Physiol. 91 (1988) 73–106. [12] J. Kang, R. Huguenard, D.A. Prince, Voltage-gated potassium channels activated during action potentials in layer V neocortical pyramidal neurons, J. Neurophysiol. 83 (2000) 70–80. [13] J.H. Lin, P. Sand, B. Rydqvist, Macrocurrents of voltage gated Na 1 and K 1 channels from the neuronal soma of the stretch receptor neuron in crayfish, Neuroreport 10 (1999) 2503–2507. [14] H.D. Lux, Observations on single calcium channels, in: B. Sakmann, E. Neher (Eds.), Single-channel Recording, 1st Edition, Plenum Press, New York, 1985, pp. 437–449. [15] D.L. Marquardt, An algorithm for least-squares estimation of nonlinear parameters, J. Soc. Ind. Appl. Math. 11 (1963) 431–441. [16] L.M. Masukawa, A.J. Hansen, G. Shepherd, Distribution of singlechannel conductances in cultured rat hippocampal neurons, Cell. Mol. Neurobiol. 11 (1991) 231–243.
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9 [17] R. Ogata, K. Kitamura, Y. Ito, H. Nakano, Inhibitory effects of genistein on ATP-sensitive K 1 channels in rabbit portal vein smooth muscle, Br. J. Pharmacol. 122 (1997) 1395–1404. [18] N. Purali, B. Rydqvist, Block of potassium outward currents in the crayfish stretch receptor neurons by 4-aminopyridine, tetraethylammonium chloride and some other chemical substances, Acta Physiol. Scand. 146 (1992) 67–77. [19] B. Rydqvist, J. Zhou, Potential-dependent potassium currents in the slowly adapting stretch receptor of the crayfish, Acta Physiol. Scand. 317 (1989) 409–419. [20] B. Rydqvist, N. Purali, Potential-dependent potassium currents in the rapidly adapting stretch receptor neuron of the crayfish, Acta Physiol. Scand. 142 (1991) 67–76. [21] B.U. Safronov, U. Bischoff, W. Vogel, Single voltage-gated K 1
[22]
[23]
[24] [25]
9
channels and their functions in small dorsal root ganglion neurons of rat, J. Physiol. 493 (1996) 393–408. R. Skryma, F. Van Coppenolle, L. Dufy-Barbe, B. Dufy, N. Prevarskaya, Characterization of Ca 21 -inhibited potassium channels in the LNCaP human prostate cancer cell line, Receptors Channels 6 (1999) 241–253. N.B. Standen, P.R. Stanfield, T.A. Ward, Properties of single potassium channels in vesicles formed from the sarrolemma of frog skeletal muscle, J. Physiol. 364 (1985) 339–358. C. Swerup, B. Rydqvist, The abdominal stretch receptor organ of the crayfish, Comp. Biochem. Physiol. 103A (1992) 423–431. D.L. Ypel, D.E. Clapham, Development of a delayed outwardrectifying K 1 conductance in cultured mouse peritoneal macrophages, Proc. Natl. Acad. Sci. USA 81 (1984) 3083–3087.
Research report
Characterization of a delayed rectifier potassium channel in the slowly adapting stretch receptor neuron of crayfish Jia-Hui Lin, Bo Rydqvist* Department of Physiology and Pharmacology, Karolinska Institutet, S-171 77, Stockholm, Sweden Accepted 19 June 2001
Abstract Single channel recordings were performed on enzyme-cleaned slowly adapting sensory neurons of crayfish, in cell-attached configuration, with a physiological K 1 gradient across the neuronal membrane. An outward rectifying, voltage-gated K 1 channel with a slope conductance of 13 pS and a K 1 ion permeability of PK 56.5310 214 cm 3 / s was characterized. This 13 pS K 1 channel started to be activated at around 20 mV depolarization. Its open probability increased upon depolarization with V0.5 5225.3 mV and Pmax 50.83. The averaged currents showed a delay following the onset of depolarization. The activation time constant was voltage-dependent. The maximal value was 17.0 ms at 225 mV and at 135 mV the time constant was 1.7 ms. Little inactivation was observed throughout the 80or 1500-ms long depolarization pulses. A sum of two exponentials provided the optimal fit for open time and closed time distribution. At 80-mV depolarization, the open time constants were 0.4 and 10.4 ms; the close time constants were 0.4 and 2.3 ms. The first-latency distribution suggested that at least two closed states preceded two open states. This 13 pS delayed rectifier plays a minor role in the maintenance of the resting membrane potential but contributes to the action potential repolarization. It may also modify the stretch-induced receptor potential and affect the adaptation behaviours in this neuron. 2001 Elsevier Science B.V. All rights reserved. Theme: Sensory systems Topic: Invertebrate sensory systems Keywords: Delayed rectifier; Voltage-gated K 1 channel; Single channel recording; Patch clamp; Slowly adapting receptor neuron
1. Introduction Potassium channels are ubiquitous in biological membranes. A large amount of different K 1 channels are found with important biological functions, such as encoding and modifying trains of action potentials in excitable cells. Among them, the voltage-gated delayed rectifier K 1 channels (Kv) play an important role in regulation of the repolarization phase of action potentials in excitable cells [8]. In the crustacean abdominal stretch receptor organ, which consists of a slowly adapting (SA) and rapidly adapting (RA) receptor neuron, the voltage-dependent potassium permeability system has been suggested to be one of the factors determining the difference in action potentials and adaptive behaviours between these two *Corresponding author. Tel.: 146-8-728-7267; fax: 146-8-32-7026. E-mail address: [email protected] (B. Rydqvist).
mechanoreceptors [2,19,20]. The outward currents in response to depolarizing potential steps have been shown to be carried by K 1 ions and to be delayed rectifying. At least two different voltage-gated K 1 channels have been suggested to be present in the neurons, according to their different activation–inactivation kinetics and sensitivities to extracellular blockers such as tetraethylammonium (TEA) and 4-aminopyridine (4-AP) [13,18]. This knowledge was derived from studies of macroscopic ionic currents using the two-electrode intracellular voltage clamp and macropatch clamp techniques. To determine in more detail the properties of the macroscopic K 1 currents, it is necessary to characterize individual K 1 channels and the contributions of the different K 1 channels to the macrocurrents. In addition to the single stretchactivated channels in SA and RA neurons [6], a Ca 21 activated K 1 channel (K 21 Ca ) with single channel conduct21 ance of 87 pS has been suggested [7]. This K Ca channel 21 was found to be activated by Ca influx through stretch-
0006-8993 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 01 )02737-8
2
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
activated channels. Due to the considerable difficulty in exposing the neuronal membrane of the stretch sensory neuron to patch clamping, little is know about the properties of voltage-gated single K 1 channels in this preparation, which are supposed to account for the delayed rectifying whole cell currents. We therefore performed single channel recordings of the voltage-gated K 1 channels in SA neuronal membrane, which could be patched after enzymatic treatment. Several different kinds of K 1 channels were observed during the study, and in this paper we focused on characterizing a 13 pS delayed rectifier K 1 channel on the basis of its singlechannel properties such as permeability / conductance, kinetics, open probability and its physiological role when compared to the macrocurrents obtained from two-electrode voltage clamp recordings and macropatch recordings [13,19].
2. Material and methods The slowly adapting (SA) stretch receptor neuron of crayfish (Pacifastacus leniusculus) was dissected free and mounted with silver clips on the bottom of a 35-mm plastic Petri dish covered with a thin layer of Sylgard 184 (Dow Corning, Seneffe, Belgium). The chamber was placed on an inverted microscope (IM-35, Zeiss, Oberkochen, Germany), and perfused with the normal astacus saline (NAS) at room temperature (20–248C). The NAS solution contained (in mM): 200 NaCl, 5.4 KCl, 13.5 CaCl 2 , 2.6 MgCl 2 , 10 Hepes, pH adjusted to 7.4 with NaOH. The osmolality of NAS was 420615 mOsm / kg, which is similar to the cytosolic osmolality. In order to expose the neuronal membrane, the glial cells and connective tissues which cover the soma were removed enzymatically by lightly pressing the tip of a fire-polished pipette (having a diameter of 30–50 mm) over the soma for 30–70 min, depending on the extent of glial cells and connective tissue. The pipettes contained an enzyme mixture consisting of: protease (type: XIV), 20 mg / ml; collagenase / dispase (EC 3.4.24.3), 10 mg / ml; hyaluronidase (EC 3.2.1.35), 7 mg / ml and elastase (EC 3.4.21.36), 1 mg / ml [13]. Additional mechanical agitation was applied after the enzymatic treatment to remove the debris by blowing and sucking the bath solution over the enzyme-treated area. All enzymes were purchased from Sigma (St Louis, MO, USA).
2.1. Single channel recording Single channel K 1 currents were recorded in the cellattached configuration of the patch clamp technique. Patch pipettes were pulled from aluminosilicate glass (Hilgenberg, Malsfeld, Germany; o.d. 1.6 mm, i.d. 1.28 mm), coated with Sylgard 184 and fire polished. Pipettes were
filled with NAS and had a resistance of 2–5 MV. During some experiments, 0.3 mM tetrodotoxin (TTX, Sigma) was included in pipette solution to block voltage-gated Na 1 channels. The patch pipette was positioned on the enzymetreated part of the neuronal soma. A gigaohm seal of 10–50 GV was formed by suction applied to the pipette. As was previously discussed [13] only cells which showed action potentials, were used. We also consider this a strong support for an intact membrane including channel proteins. If action potentials can be seen the resting membrane must be normal and consequently the K 1 leak channels intact. If the leak channels were damaged over the entire enzymetreated area (30–40 mm in diameter) the resting membrane would have deteriorated. This also implies that other channels are not affected. Currents were recorded using an Axopatch 200 amplifier and pClamp 6 programs (Axon Instruments, Foster City, CA, USA), and stored in a computer. Either ramps or steps were applied as voltage clamp commands. Data were filtered at 1–2 kHz by a 23 dB, eight-pole low-pass Bessel filter and sampled at 10 kHz. All analysis was made with pClamp 6 programs. When constructing averaged currents, leakage and capacitative currents were digitally removed by subtraction of blank sweeps with no channel activities. Single channel amplitudes were determined either by onscreen measurement of 5–10 events via cursors, or by forming an all-points histogram of the baseline and openlevel data points, and then fitting with Gaussians by a Marquardt least-squares algorithm (Marquardt-LSQ) [15]. Transitions between open and closed states were detected using a half-amplitude threshold criterion. The dwell time and first-latency histograms were formed by reconstructed current traces resulting from 80-ms long steps to different potentials. The number of channels present in a patch was assumed to be the observed maximal number of open channels within the patch. It was further assumed that when the potential was stepped to a new value, the rate constants for the channel kinetics immediately attained new values, which resulted in time constants in the open and closed time histograms. Since no inactivation was observed within 1500 ms, we considered the 80-ms pulses adequate for the present analysis. Potentials were expressed as the potentials of the interior face relative to the exterior face of the membrane. Outward currents are shown as upward deflections. Vh is referred to as the holding potential (2Vpipette in cell-attached configuration), and Vm the membrane potential across the patch. A resting membrane potential of 265 mV was assumed when analysing the data from cell-attached recordings, as determined in earlier results obtained either by two-electrode intracellular current clamp or by whole-cell patch clamp recordings [13]. The reference potential for all measurements was corrected to be zero before establishing the seal. Junction potential calculated by the Axoscope program (Axon Instruments) was less than 2 mV, and no correction was made.
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
Results throughout this paper are mean6S.E.M., unless otherwise indicated.
given
3
as
3. Results The single channel current described in this paper was investigated in cell-attached configuration with normal astacus saline (NAS, [K 1 ] o 55.4 mM) in the patch pipette. Cell-attached configuration was used to keep the K 1 ionic gradient and other cytosolic components as physiological as possible to mimic the normal channel environment. These arrangements also made it easier to relate the singlechannel data with results of macrocurrents recorded under similar experimental conditions by two-electrode voltage clamp or macropatch recordings [13,19]. Whether the patches were from the stretch receptor neuron or from the glial cells was decided based on the electrical excitability in cell-attached configuration or whole-cell configuration after the patch was ruptured [6,13]. Patches from neuronal membrane showed spontaneous action potentials, or initiated action potentials in response to bath application of 16 mM K 1 -NAS (the [K 1 ] in NAS was increased to 16 mM by the addition of an appropriate amount of 3 M KCl), which depolarized the cell by up to 25 mV, sufficient for the initiation of action potentials. Only those neurons that showed action potentials were used for single-channel studies [13]. After experiments some patches were ruptured into whole-cell configuration. The observation of intracellular action potentials further confirmed that the recordings had been obtained from the neuronal membrane.
3.1. Single channel conductance Applying depolarizing voltage steps to a cell-attached patch on neuronal membrane resulted in the activation of a voltage-gated ion channel, as illustrated in Fig. 1a. When the patch was hyperpolarized or at its resting membrane potential (Vh 50 mV), no single channel events could be discerned. This channel was activated by at least 20 mV depolarization, and distinct single channel events could then be observed. With further depolarization, the amplitude of the channel current increased and the channel became more active. At highly depolarized potentials the channel remained active throughout the depolarizing pulse (1500 ms) and little inactivation was seen. It also exhibited long openings with some short closings during channel opening. This kind of channel was frequently observed during the experiments. It was usually possible to observe channel activity immediately or soon after seal formation when depolarization was applied. The channel activity persisted for several minutes with no discernible changes in channel behaviour. In all patches analysed, only one or two of these channels were present. However, we also observed
Fig. 1. Single channel currents recorded from a cell-attached patch on neuronal membrane with NAS (5.4 mM K 1 ) in the pipette. (a) Representative currents at different holding potentials (Vh ) as indicated on the left. The patch was held at resting potential and activated by step depolarization. Each of the illustrated recordings is part of a 1500-ms episode. Data were acquired at the sampling frequency of 10 kHz and filtered at 1 kHz. The dashed lines mark the closed state of the channel. (b) Single channel current–voltage relationship. The unitary conductance from a linear regression between 240 and 140 mV is 15.5 pS. The smooth curve is obtained by the best fit with GHK equation; Pk 55.43 10 214 cm 3 / s. Same patch as in (a).
patches with more than two conductance levels of this channel type being simultaneously active, or patches containing several conductance levels, indicating the presence of different types of channels in addition to the channel discussed in this paper. Fig. 1b illustrates the I–V plot of single channel current amplitudes as a function of membrane potentials (Vm ) when a resting membrane potential of 265 mV was assumed. The I–V curve shows the voltage dependence of the current amplitude and its outward rectification. With the physiological K 1 gradient across the membrane, the unitary conductance for this patch was 15.5 pS between
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
4
membrane potentials of 240 and 140 mV. The average of nine such patches yielded a unitary conductance of 13.260.8 pS (n59). The Goldman–Hodgkin–Katz (GHK) equation which is valid under the assumption of a constant field within the membrane and independence of ion movements in both direction, was used to fit the I–V curve: IK 5 PK (EF 2 /RT )([K 1 ] o 2 [K 1 ] i exp(EF /(RT ))) / (1 2 exp(EF /(RT )))
possible inactivation of this K 1 channel. The channel exhibited some long openings and closings, as well as some fast events. The bottom trace illustrates the ensemble-averaged current. Following the onset of the depolarization, the current showed a delay in activation and did not inactivate much during the 80-ms depolarizing pulse, properties similar in time course to the whole-cell delayed 1 rectifier K current defined in giant squid axon [9].
(1)
3.2. Single channel kinetics where PK is the K 1 permeability (in cm 3 / s); E is the membrane potential; F, R, T have their usual meanings. [K 1 ] o and [K 1 ] i are the extracellular K 1 (5.4 mM) and the intracellular K 1 concentration (assumed to be 180 mM [19]), respectively. The GHK equation (smooth line in Fig. 1b) provided a reasonable description of the I–V relationship for the data, suggesting that this channel was K 1 selective. The PK value obtained from the fit was 5.43 214 3 10 cm / s in this patch, and the average PK value was 6.560.3310 214 cm 3 / s (n59). Fig. 2 illustrates five representative sweeps for the first opening elicited by a 80-mV depolarization, which was preceded by a 40-mV hyperpolarizing step to remove
Kinetic analysis of single channel activity was performed at different depolarized potential levels using the repetitive voltage step paradigm of Fig. 2. Only patches that contained one active channel were analysed for single channel kinetics. Fig. 3a,b illustrates the open time and closed time histograms of single channel events at a 80-mV depolarization (Vh 5 180 mV). Both the open time and closed time distributions were optimally fitted by two exponentials (Marquardt-LSQ), yielding one fast and one slow time constant. The open time constants at 80-mV depolarization for this patch were 0.460.2 and 9.660.0 ms (mean6variance, 2131 events); the averaged values from three patches were 0.460.1 ms (n53) and 10.460.7 ms (n53), respectively. At the same depolarization, the slow time constant for closed time averaged 2.360.2 ms (n53) and the fast time constant was 0.460.0 ms (n53). Fitting with two exponentials provided a good description to all the dwell time distributions at holding potentials from 120 to 1100 mV. Because the number of exponentials reflects the minimal number of channel states [10], these data suggested that this kind of K 1 channel has at least two open states and two closed states. For voltage-gated channels, the latency from the onset of a depolarization to the first channel opening gives information about the number of closed states preceding the opening of the channel. Fig. 3c,d shows the firstlatency histogram of the opening of this K 1 channel in response to a 80-mV depolarization. In Fig. 3c, the distribution peaks at a time later than zero, suggesting that the channel had to go through more than one closed state to reach the first open states [10]. Since at least two closed states and two open states have been derived from dwell time histograms (Fig. 3a,b), the simplest kinetic scheme to describe the behaviour of this delayed rectifier could be: k2
k1
k 22
k 21
C2 ⇔C1 ⇔O1 ⇔O2
Fig. 2. Representative sweeps initiated by the voltage step (top) after capacitative and leak currents were subtracted. In this patch, two openings of this K 1 channel with similar behaviour were observed. The sweeps illustrated here display the first open level only for clarity (more explanation in Fig. 4c). The ensemble-averaged current (187 sweeps, bottom) had a delay in activation and did not inactivate much during the depolarization. The sample frequency was 10 kHz; filter frequency was 1 kHz; pulses were applied every 0.6 s.
in which two closed states (C1 , C2 ) precede two open states (O1 , O2 ), and k represents the rate constant. The probability density function for the distribution of latency may be obtained as the time derivative of the expression [23]: f(t) 5 (R 1 R 2 )[exp(2R 2 t) 2 exp(2R 1 t)] /(R 1 2 R 2 ) where R 1 and R 2 are related to the rate constants as:
(2)
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
5
Fig. 3. Kinetic analysis of single channel activity. (a, b) Dwell time distribution of channel opening and closing upon a 80-mV depolarization. In total, 2131 events were examined. Both open time and closed time histograms were fitted by two exponentials, yielding 0.460.2 and 9.660.0 ms (mean6variance) for open time constants, and 0.360.3 and 2.260.0 ms for closed time constants, respectively. (c, d) The first-latency distribution at 180 mV depolarization. The curve in (c) was drawn to Eq. (2), scaled to the same area as the histogram. The fit gave R 1 50.63 ms 21 and R 2 50.46 ms 21 . The curve in the cumulative latency histogram of (d) is the integral to Eq. (2). Data are from the same patch.
R 1 , R 2 5 hk 1 1 k 2 1 k 22 6[(k 1 1 k 2 1 k 22 )2 2 4k 1 k 2 ] 0.5 j / 2
Fig. 3c shows the least-squares fit obtained with Eq. (2), and Fig. 3d shows the cumulative latency histogram together with a line drawn to the integral of Eq. (2). Curve in Fig. 3c displays a fast rise phase with a slow decay, and it peaks at a time of 1.9 ms (t m ) when R 1 50.63 ms 21 and R 2 50.46 ms 21 . The mean latency was 11.3 ms for the patch shown; the averaged mean latency from three patches was 10.262.2 ms. These results support the notion that the first open state was actually preceded by at least two closed states. Although Eq. (2) gave a reasonable description to the data, the predicted peak occurred earlier than the experimental peak and some events with long latency have not been included (Fig. 3c) [1,14].
and Fig. 4b is the plot of these averaged current amplitudes versus membrane potentials. At the resting membrane potential or hyperpolarized potentials no outward currents could be detected. Averaged currents were observed only when the patch was depolarized by about 20 mV, and were highly dependent on the voltage. Following the onset of depolarization some delay in activation was present at all potentials, as illustrated in Fig. 2. No inactivation during the 80-ms-long pulses could be seen at any potentials. In this patch, two openings of the K 1 channel with similar channel behaviour were observed, as illustrated in the inset of Fig. 4c. The all-points histogram (Fig. 4c) could be optimally fitted by a third-order Gaussian distribution. Peak 1 and peak 2, corresponding to the two open levels, were evenly located, and had similar amplitude of 1.50 and 1.44 pA. No additional humps that would indicate the presence of another population of channels could be detected. The value of the open probability (Po ) was estimated using Eq. (5) and plotted as shown in Fig. 4d
3.3. Channel open probability
Po 5 I /(i 3 N)
Fig. 4a shows the ensemble-averaged currents initiated by depolarization of 20–100 mV obtained from one patch,
where I is the averaged current, i is the single channel current, and N is the observed maximal number of open
(3) This function also assumes a maximum at time t m [14] t m 5 (R 1 2 R 2 )21 ln(R 1 /R 2 )
(4)
(5)
6
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
Fig. 4. Analysis of ensemble-averaged currents. (a) Averaged currents initiated by depolarization of 20–100 mV from the resting membrane potential as indicated on the left. The number of sweeps included in each trace was: 165 sweeps for 0 mV, five sweeps for 20 mV, 111 sweeps for 40 mV, 194 sweeps for 60 mV, 105 sweeps for 80 mV and 196 sweeps for 100 mV. (b) The voltage dependence of the averaged currents plotted against the membrane potentials. (c) All-points amplitude histogram constructed from the recording at 80-mV depolarization, and optimally fitted by a third-order Gaussian distribution. The peaks corresponding to the closed level (0), first opening (1) and second opening (2) are indicated. Inset is a representative trace, showing the presence of two openings of this K 1 channel type in the patch. (d) The plot of the open probability (Po ) versus the membrane potentials. The smooth line is a fit of the Boltzmann equation to the experimental points, yielding the following values: Pmax 50.83, V0.5 5 225.3 mV and k514.2 mV. All data are from the same patch.
channels within the patch, which was two for this patch. The open probability was voltage-dependent. The relationship between voltage and open probability was sigmoidal with the voltage dependence being steepest with moderate depolarizations. Fitting the data by the Boltzmann equation: Po 5 Pmax /(1 1 exp((V0.5 2Vm ) /k))
(6)
where Pmax is the maximal open probability, V0.5 is the potential at half-maximal probability, Vm is the membrane potential, and k is the slope factor reflecting the steepness of the voltage dependence, we found Pmax 50.83, V0.5 5 2 25.3 mV and k514.2 mV when 265 mV was taken as the resting membrane potential.
3.4. Activation time constants of the K 1 current The activation time constants for this K 1 channel type were analysed by fitting the ensemble-averaged currents at different potentials to a second-order exponential function: I 5 b 3 (1 2 exp(2t /tn ))2 1 a
(7)
where a and b are the scaling parameter and tn is the activation time constant. Fig. 5a shows that such fitting with a power of two (superimposed dotted line) gave an adequate description to the averaged current (cf. [19]) in response to a 100-mV depolarization, yielding an activation time constant tn of 1.75 ms. The voltage dependence of the
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
7
sive. Further studies using excised patches are needed to pharmacologically characterize this K 1 channel.
4. Discussion
Fig. 5. Activation time constants of averaged currents for this 13 pS K 1 channel. (a) A second-order exponential function was used to fit the activation phase of the averaged current initiated by a 100-mV depolarization, producing b53.0 pA and tn 51.75 ms. Same data as in Fig. 4a. (b) The voltage dependence of the activation time constants tn versus the membrane potentials. Data were pooled from the patch shown in Fig. 4a (d) and four other patches (m, ., 쏆, j). The smooth line was computed according to the mean values at each potential.
activation time constants was illustrated in Fig. 5b. The constant approached the maximal value in the potential range of 240 to 210 mV, which was about 17.0 ms at 225 mV. It then decreased on either side. Towards positive potentials the activation time constant became much smaller; being 8.4 ms at 25 mV, 4.0 ms at 115 mV and 1.7 ms at 135 mV respectively. At potentials more positive than 135 mV, it is likely that this activation time constant would reach a stable level below 1 ms. The classical K 1 channel blocker TEA can act from both the internal and external side of the membrane but does not permeate the membrane. In cell-attached configuration, application of TEA to the pipette solution after control experiments with the same patch cannot be easily established. 4-AP is membrane permeable and can block variable K 1 channels from internal as well as external sites [8]. It could thus be applied to the bath while recording in cell-attached configuration. However the effects of the bath application of 5 mM 4-AP on this channel were inconclu-
In the present study which was performed on enzymetreated slowly adapting stretch receptor neurons, a delayed rectifier K 1 channel was characterized in cell-attached configuration with a physiological K 1 gradient across the membrane. Both its single channel amplitude (Fig. 1) and open probability (Fig. 4d) increased upon depolarization with a selectivity for K 1 ions. A sum of two exponentials provided the best fit for closed time and open time histograms. The analysis of the first-latency distribution indicated that at least two closed states (C1 and C2 ) preceded two open states (O1 and O2 ) (Fig. 3). The ensemble-averaged currents showed a delay following the onset of depolarization with little inactivation (Figs. 2 and 4a). The activation time constant for the K 1 averaged currents was highly voltage-dependent (Fig. 5). This delayed rectified K 1 channel had a conductance of 13 pS, in agreement with those reported in other preparations. Conti and Neher [4] showed a delayed rectifier K 1 channel in squid axon having a conductance of 17 pS, while Standen et al. [23] recorded a 15 pS delayed rectifier in frog skeletal muscle cells. Similar conductance for this delayed rectifier was also described in rat hippocampal neurons (9.5 pS [16]), mouse macrophages (16 pS [25]), embryonal carcinoma cells (10 pS [5]), PC 12 cells (7 pS [11]) and neocortical pyramidal neurons (17 pS [12]). However, in some preparations, a much higher conductance for delayed rectifier K 1 channel has been reported, such as that of rat dorsal ganglion neurons (55 pS [21]), LNCaP human prostate cancer cell (78 pS [22]) and rabbit portal vein smooth muscle (42 pS [17]). Therefore, single channel conductance less than 20 pS for the delayed rectifier K 1 channel seems to be more common in most preparations including the stretch receptor neuron, even though high conductance channels are also present in certain species. The 13 pS delayed rectifier described in this paper was activated by |20 mV depolarization, and no current was detected at the resting or hyperpolarized potentials. Its open probability increased with depolarization (V0.5 5 2 25.3 mV) and approached the maximal value (Pmax 50.83) at high positive potentials, suggesting that this channel played a minor role in the maintenance of the resting membrane potential but probably contributes to the repolarization of action potentials. Furthermore, no inactivation was observed within 80 ms for this 13 pS K 1 channel whereas the whole cell current in this preparation inactivates significantly [19]. This channel was also frequently observed during the experiments, suggesting that its distribution over the neuronal soma may be widespread and
8
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9
its channel density high. It is therefore likely that its contribution to the repolarization of action potentials is considerable and that it accelerates membrane repolarization at high positive potentials. Further, since this 13 pS delayed rectifier could be activated by |20 mV depolarization, it may also interact with the stretch-induced permeability change by making the repolarization of the receptor potential faster. Such modification in stretchinduced receptor potential may affect the adaptation behaviour in the SA sensory neuron [24]. Both the activation phase of the averaged currents for this 13 pS K 1 channels and that of the whole cell K 1 currents recorded by two-electrode voltage clamp, could be best fitted by a second-order exponential function (Fig. 5a; [19]). The activation time constant for the 13 pS K 1 current was about 6 ms at 0 mV (Fig. 5b), much slower in comparison to 1 ms for the whole cell K 1 current at 0 mV. At high positive potentials the difference becomes less. The whole cell K 1 currents also exhibit a marked inactivation within 50 ms, whereas the 13 pS K 1 averaged currents were not inactivated within 80 ms. However, K 1 macrocurrents obtained by macropatch recordings show both fast inactivation in some patches and little inactivation in other patches [13]. Since other K 1 channels with different properties were actually observed during this study, it is conceivable that this 13 pS delayed rectifier is one of several types of K 1 channels contributing to the whole cell K 1 currents. Other K 1 channels with faster activation as well as faster inactivation could be expected. The distribution of these 13 pS K 1 channels over neuronal soma may be widespread but uneven; they might aggregate with each other in some parts of the soma whereas in other locations they might cluster with other different types of K 1 channels, as suggested by the macropatch recordings [13]. Whether the 13 pS K 1 channel is the dominant one or how much it contributes to the total K 1 current cannot be determined at present. Whole cell currents recorded using intracellular micropipettes [18] showed that ca. 25% of the peak outward current and a major part of the outward current at 500 ms was TEA sensitive indicating that an outward rectifier constitutes a dominant part of the K 1 channel population. Whether the 13 pS channel is the only channel contributing to this outward rectifier cannot be concluded at present. Further work is needed to define other K 1 channels and their contribution to the total macroscopic K 1 current. Among voltage-gated K 1 channel proteins (Kv family), eight subfamilies (Kv1–Kv6 and Kv8–Kv9) have been cloned in heterologous expression systems [3]. In comparison with these cloned K 1 channels, the 13 pS K 1 channel found in crayfish stretch sensory neuron displays properties similar to Kv1.2 channel in activation threshold (Von ), k, inactivation process and single channel conductance, which has Von of 240 mV, k of 13 mV, single channel conductance of 9.2–17 pS and very slow inactivation [3]. However, there is some discrepancy in the half-activation
potential (V0.5 ) with a V0.5 of 25 to 5 mV for Kv1.2 channel and a V0.5 of 225 mV for this 13 pS delayed rectifier found in crayfish SA receptor neuron. Therefore, a detailed molecular characterization of the channel proteins underlying the 13 pS delayed rectifier will be helpful in providing more insights about this channel.
Acknowledgements We are grateful to Drs Fredrik Elinder and Joe Bruton for reading the manuscript and for fruitful suggestions. This work was supported by grants from the Swedish Medical Research Council (Project No. 6838) and Karolinska Institutet.
References [1] K. Bokvist, P. Rorsman, P.A. Smith, Effects of external tetraethylammonium ions and quinine on delayed rectifying K 1 channels inmousepancreaticb-cell,J.Physiol.423(1990)311–325. [2] N.M. Brown, D. Ottoson, B. Rydqvist, Crayfish stretch receptor: an investigation with voltage-clamp and ion sensitive electrodes, J. Physiol. 284 (1978) 155–179. [3] W.A. Coetzee, Y. Amarillo, J. Chiu, A. Chow, D. Lau et al., Molecular diversity of K 1 channels, Ann. NY Acad. Sci. 868 (1999) 233–285. [4] F. Conti, E. Neher, Single channel recordings of K 1 currents in squid axons, Nature 288 (1980) 140–143. [5] L. Ebihara, W.C. Speers, Ionic channels in a line of embryonal carcinoma cells induced to undergo neuronal differentiation, Biophys. J. 46 (1984) 827–830. [6] C. Erxleben, Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish, J. Gen. Physiol. 94 (1989) 1071–1083. [7] C. Erxleben, Calcium influx through stretch-activated cation channels mediates adaptation by potassium current activation, Neuroreport 4 (1993) 616–618. [8] B. Hille, Ionic Channels of Excitable Membrane, Sinauer Associates Inc., Sunderland, MA, 1992. [9] A.L. Hodgkin, A.F. Huxley, The component of membrane conductance in the giant axon of Loligo, J. Physiol. 116 (1952) 473–496. [10] R. Horn, N.B. Standen, Counting kinetic states: the single channel approach, in: A. Grinell, W. Moody (Eds.), The Physiology of Excitable Cells, Alan R. Liss Inc., New York, 1983, pp. 181–189. [11] T. Hoshi, R.W. Aldrich, Voltage-dependent K 1 currents and underlying single K 1 channels in phenochromocytoma cells, J. Gen. Physiol. 91 (1988) 73–106. [12] J. Kang, R. Huguenard, D.A. Prince, Voltage-gated potassium channels activated during action potentials in layer V neocortical pyramidal neurons, J. Neurophysiol. 83 (2000) 70–80. [13] J.H. Lin, P. Sand, B. Rydqvist, Macrocurrents of voltage gated Na 1 and K 1 channels from the neuronal soma of the stretch receptor neuron in crayfish, Neuroreport 10 (1999) 2503–2507. [14] H.D. Lux, Observations on single calcium channels, in: B. Sakmann, E. Neher (Eds.), Single-channel Recording, 1st Edition, Plenum Press, New York, 1985, pp. 437–449. [15] D.L. Marquardt, An algorithm for least-squares estimation of nonlinear parameters, J. Soc. Ind. Appl. Math. 11 (1963) 431–441. [16] L.M. Masukawa, A.J. Hansen, G. Shepherd, Distribution of singlechannel conductances in cultured rat hippocampal neurons, Cell. Mol. Neurobiol. 11 (1991) 231–243.
J.-H. Lin, B. Rydqvist / Brain Research 913 (2001) 1 – 9 [17] R. Ogata, K. Kitamura, Y. Ito, H. Nakano, Inhibitory effects of genistein on ATP-sensitive K 1 channels in rabbit portal vein smooth muscle, Br. J. Pharmacol. 122 (1997) 1395–1404. [18] N. Purali, B. Rydqvist, Block of potassium outward currents in the crayfish stretch receptor neurons by 4-aminopyridine, tetraethylammonium chloride and some other chemical substances, Acta Physiol. Scand. 146 (1992) 67–77. [19] B. Rydqvist, J. Zhou, Potential-dependent potassium currents in the slowly adapting stretch receptor of the crayfish, Acta Physiol. Scand. 317 (1989) 409–419. [20] B. Rydqvist, N. Purali, Potential-dependent potassium currents in the rapidly adapting stretch receptor neuron of the crayfish, Acta Physiol. Scand. 142 (1991) 67–76. [21] B.U. Safronov, U. Bischoff, W. Vogel, Single voltage-gated K 1
[22]
[23]
[24] [25]
9
channels and their functions in small dorsal root ganglion neurons of rat, J. Physiol. 493 (1996) 393–408. R. Skryma, F. Van Coppenolle, L. Dufy-Barbe, B. Dufy, N. Prevarskaya, Characterization of Ca 21 -inhibited potassium channels in the LNCaP human prostate cancer cell line, Receptors Channels 6 (1999) 241–253. N.B. Standen, P.R. Stanfield, T.A. Ward, Properties of single potassium channels in vesicles formed from the sarrolemma of frog skeletal muscle, J. Physiol. 364 (1985) 339–358. C. Swerup, B. Rydqvist, The abdominal stretch receptor organ of the crayfish, Comp. Biochem. Physiol. 103A (1992) 423–431. D.L. Ypel, D.E. Clapham, Development of a delayed outwardrectifying K 1 conductance in cultured mouse peritoneal macrophages, Proc. Natl. Acad. Sci. USA 81 (1984) 3083–3087.