| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Journal of Physiology (2001), 536.1, pp. 49-65
© Copyright 2001 The Physiological Society
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
2-3 ms) and occurred within the physiological range of membrane potentials (V1/2 = -63 mV). Recovery from inactivation was also rapid ( 10 ms).
2) thought to mediate BK channel inactivation in rat chromaffin cells. We hypothesize that Itransient results from the association of a similar subunit with some of the BK channels and that papain removes inactivation by cleaving extracellular sites required for this association.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Perhaps more so than any other ion channel potassium channels play an integral role in shaping the electrical behaviour of cells. A striking example of this is seen in the auditory and vestibular systems of non-mammalian vertebrates where each hair cell is tuned to a particular frequency of sound through, at least in part, an electrical resonance. Resonant frequency, and hence the frequency of sound to which each hair cell is maximally sensitive, is determined primarily by the class(es) and kinetics of the K+ channels that each hair cell expresses (see Fettiplace & Fuchs, 1999). In frog saccular hair cells, it was originally reported that non-inactivating large-conductance calcium-activated (BKN) channels were the only class of K+ channel involved in generating resonant frequencies of 80-160 Hz (Lewis & Hudspeth, 1983; Hudspeth & Lewis, 1988a,b). These conclusions, however, came from studies of enzymatically dissociated hair cells, the properties of which are greatly altered by the enzyme (papain) used in the dissociation procedure. Using a semi-intact epithelial preparation that avoided enzymatic dissociation, we have shown that both BK channels and voltage-dependent K+ (KV) channels play a role in generating resonance of 35-75 Hz (Armstrong & Roberts, 1998). Not only is a second class of K+ channels involved in generating resonance in these cells, but as we show here, the BK current is more complex than previously appreciated.
Patch-clamp recordings from hair cells in semi-intact epithelial preparations bathed in 4-aminopyridine (4-AP), an agent effective at blocking KV, but not BK, channels in hair cells (Goodman & Art, 1996), reveals a rapidly activating, partially inactivating outward current. Both the transient (Itransient) and steady-state (Isteady) components of this current were Ca2+ dependent, sensitive to a peptide toxin specific for BK channels, and had a large single-channel conductance. Thus, frog saccular hair cells contain both inactivating (BKI) and non-inactivating (BKN) populations of BK channels.
Although BK channels are found in myriad organisms and cell types (for a review see Latorre et al. 1989; McManus, 1991), only rarely do they exhibit inactivation (Pallotta, 1985; Ikemoto et al. 1989; McLarnon, 1995; Lingle et al. 1996; Hicks & Marrion, 1998; Li et al. 1999). The only known mechanism of BK channel inactivation is that conferred by auxiliary () subunits, which mediate inactivation through an intracellular 'ball-and-chain' mechanism (Wallner et al. 1999; Xia et al. 1999; Xia et al. 2000). Our finding that inactivation of BKI in frog saccular hair cells can be removed by extracellular application of papain and subsequently restored by intracellular application of the 'ball' domain of the 2 subunit (Wallner et al. 1999; same as 3 in Xia et al. 1999) suggests that in these cells BKI inactivation also occurs via a 'ball-and-chain' mechanism. Because subunits have a large extracellular domain and are likely to interact with the BK channel at an extracellular site these two results can be consolidated by hypothesizing that a subunit similar to 2 mediates inactivation of BKI in frog saccular hair cells.
A preliminary report of some of these results has appeared in abstract form (Armstrong & Roberts, 1999).
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Hair cell preparations
Grassfrogs (Rana pipiens) were decapitated and pithed according to guidelines established by the Institutional Animal Care and Use Committee at the University of Oregon. All experiments were carried out at room temperature (20-23 °C). Recordings were made in either a semi-intact epithelial preparation or after enzymatic dissociation. Both preparations are described in Armstrong & Roberts (1998). In this study no attempt was made to correlate hair cell position or morphology with channel properties.
Recording procedures
Hair cells were visualized on an upright microscope (Zeiss, Thornwood, NY, USA) equipped with a 
40 water immersion objective and differential interference contrast optics. Whole-cell (Hamill et al. 1981) and perforated-patch (Horn & Marty, 1988) recordings were made using a Dagan patch-clamp amplifier (model 3900 or model 3900A; Dagan Corp., Minneapolis, MN, USA), with active series resistance compensation. Recording pipettes were pulled from borosilicate glass (Sutter Instrument Co., Novato, CA, USA), coated with Sylgard (type 182; Dow Chemical Co., Midland, MI, USA), and heat-polished to have a tip diameter of ~1 µm. Pipettes were filled with a standard intracellular solution, consisting of (mM): 122 K+, 114 aspartate, 0.08 Ca2+, 4 Cl-, 2 Mg2+, 5 Hepes, 1 EGTA and 1 ATP. Pipette resistances were generally 2-6 M
. After formation of a tight (>1 G) seal on the cell's membrane, electrical access to the cell's interior was achieved either by rupturing the membrane inside the pipette, thereby replacing the soluble components of the cytoplasm with the intracellular solution (whole-cell), or through channels formed by the perforating agent nystatin (perforated-patch). Stock solutions of nystatin (50 mg ml-1 in DMSO) were stored in the dark at -20 °C for up to 8 h. Working solutions of nystatin (4 µl of stock per millilitre of intracellular solution) were prepared approximately every 20 min and were kept on ice in the dark. In all experiments, membrane potentials were corrected for the liquid junction potential between the intracellular and bath solutions (Vjp = -13 mV).
pCLAMP 6 (Axon Instruments, Foster City, CA, USA) was used to generate voltage commands and record the resulting currents. Data were filtered at either 5 or 10 kHz and digitized at either 10 or 20 kHz. Except where noted, all voltage-clamp steps were applied from a holding potential of -70 mV, averaged five times, and leak subtracted using a standard P/4 protocol (Armstrong & Bezanilla, 1977). In most experiments the stimulus consisted of a series of depolarizing steps to potentials between +90 and -60 mV in 10 mV increments. The stimulus protocol used for ensemble-variance analysis is described below and those used for other experiments are noted in the text.
Solutions and pharmacological agents
All solutions were adjusted to have a pH of 7.25 and an osmotic strength of ~220 mosmol l-1. The recording chamber was continuously perfused with a perilymph-like solution (normal extracellular solution; Armstrong & Roberts, 1998) containing (mM): 112 Na+, 2 K+, 1.8 Ca2+, 0.7 Mg2+, 119 Cl-, 3 D-glucose and 5 Hepes. Pharmacological agents were applied to hair cells either via a local perfusion system, or by exchanging the entire bath (~0.5 ml of solution). Local perfusion was achieved through two- or five-barrelled pipettes with openings approximately 30 µm in diameter. Flow through this system was driven by a peristaltic pump. Solutions containing >1 mM of pharmacological agents were prepared by equimolar substitution for NaCl. In extracellular solutions lacking Ca2+, Mg2+ was substituted to keep the concentration of divalent cations constant. Potentially unstable chemicals, such as ATP, 4-AP, peptides and enzymes, were added to solutions on the day of recording. Iberiotoxin (IbTX) was obtained from Bachem (King of Prussia, PA, USA) and apamin was obtained from Calbiochem (La Jolla, CA, USA). Purified papain (twice crystallized; Sigma, cat no. P3125) was used at 8 units ml-1 of normal extracellular solution and was activated by 2.5 mM L-cysteine. The 19 amino acid 'ball' peptide (MFIWTSGRTSSSYRHDEKR), which constitutes the amino terminus of the BK channel 2 subunit (Wallner et al. 1999; Xia et al. 1999), was synthesized by Genemed Synthesis, Inc. (South San Francisco, CA, USA).
In experiments in which we applied trypsin (bovine pancreatic; Worthington, Freehold, NJ, USA) intracellularly, approximately 0.5 mm of the pipette's tip was front filled with standard intracellular solution that did not contain trypsin before filling the back of the pipette with intracellular solution containing trypsin. To ensure that trypsin had sufficient time to diffuse to the pipette's tip and would therefore be introduced into the hair cell during whole-cell recording, we waited 7-9 min after establishing a giga-seal before rupturing the membrane beneath the pipette. To pharmacologically isolate the Ca2+ current (ICa), Cs+ was substituted for K+ in the internal solution and 5 mM Cs+ was added to the extracellular solution to block an inwardly rectifying current (IKIR) which was often activated during the P/4 leak subtraction procedure. In recordings made using K+ internal and normal extracellular saline solutions IKIR was sometimes significantly activated during the leak subtraction pulses for the largest voltage steps. The leak subtractions pulses were always examined for such non-linear responses, and when necessary contaminated leak pulses were replaced by appropriately scaling pulses elicited by smaller voltage steps.
Data analysis
Most data analysis was performed using pCLAMP 6. Additional analysis and plotting were carried out using Microsoft Excel (Microsoft Corp., Redmond, WA, USA) and IgorPro (WaveMetrics, Lake Oswego, OR, USA).
Offline series resistance compensation
Despite using low-resistance recording pipettes and active series resistance compensation, a significant uncompensated series resistance (RS ~4 M for whole-cell, ~15 M for perforated-patch) remained. One effect of this uncompensated RS was to cause the membrane voltage (VM) to deviate from the command voltage (VC). All voltage amplitudes reported here were corrected for this error using the following procedure: To estimate RS the area (Q) and decay time constant (C) of the charging transient evoked by a small step applied during the P/4 leak subtraction procedure were measured. RS was then calculated according to: RS = (CVS)/Q, where VS is the amplitude of the small voltage step. We then calculated VM using the formula: VM = VC - IMRS, where IM is the peak membrane current (before leak subtraction). Because the corrected values of VM deviated from VC by different amounts in each recording, we interpolated linearly between adjacent voltage steps to compare currents recorded at the same voltage in different cells.
Deconvolution to restore ICa activation
An additional effect of the residual RS was to low-pass filter IM at a frequency that varied inversely with C: fcutoff = 1/(2
C). For the cells used in this study, fcutoff was between 1 and 3 kHz (C = 50-150 µs), a degree of filtering sufficient to significantly slow the apparent activation rate of ICa. To recover the unfiltered ICa, recorded currents were deconvolved as described in Armstrong & Roberts (1998). Uncompensated RS also slowed changes in VM, an effect that deconvolution does not correct. For this reason, 4 of 79 cells were omitted from the data set because their C exceeded 150 µs.
Separating ICa, Itransient and Isteady
The 4-AP-insensitive current was separated into three components, a non-inactivating inward current (ICa), a non-inactivating outward current (Isteady) and a transient outward current (Itransient), having peak amplitudes aCa, asteady and atransient, respectively (Fig. 1A). To calculate aCa, currents measured in response to depolarizing voltage steps from the holding potential of -70 mV were corrected for voltage errors due to RS, interpolated to VM = -20 mV and deconvolved, as described previously. To estimate aCa the initial inward current at the beginning of the voltage step was then fitted using the equation:
 |
View larger version [in this window] [in a new window] |
|
|
Figure 1. Separating currents into ICa, Itransient and Isteady A, ICa (thin trace) was estimated by fitting the initial inward component of the recorded current using the known activation kinetics (eqn (1) with | ||
|
ICa(t) = aCa(1 - e-t/m)3,
| (1) |
where m is the time constant for one of three theoretical gating particles (m = 0.13 ms at -20 mV; Armstrong & Roberts, 1998) and t is the time. Using this procedure, the average aCa at -20 mV in our population of cells was -330 ± 191 pA (mean ± S.D., n = 75), in agreement with more direct measurements (-358 ± 145 pA, n = 10; Armstrong & Roberts, 1998). To calculate asteady, aCa was subtracted from the steady-state current at the end of the voltage step. atransient was determined as the difference between the peak outward current and the steady-state current. This estimation procedure may have slightly overestimated asteady due to a 'sag' in ICa (see Fig. 10A of Armstrong & Roberts, 1998). This error, however, is expected to be small (<50 pA, corresponding to <10 % error in asteady).
Ensemble-variance analysis
Single-channel current (i) was estimated using ensemble-variance analysis (Sigworth, 1980). All of these experiments were performed in whole-cell configuration in saline containing 10 mM 4-AP to block the KV current. Currents were elicited by a train of 500 identical stimuli, consisting of two 10 ms depolarizations to command potentials of either 0 or +40 mV (yielding VM values between -4 and +31 mV) separated by a 2 ms repolarization to the holding potential (-70 mV; Fig. 1B). Ensemble-variance analysis was performed on the current elicited during the first depolarization of each pair; the second depolarization was used only to determine steady-state values of the mean current, m(t), and variance, 
2(t). Data were acquired with no leak subtraction, and with a wide bandwidth (10 kHz low-pass cutoff) to minimize errors contributed by filtering (Silberberg & Magleby, 1993). Cells in which the peak outward current exhibited more than 10 % run-down during the train of 500 stimuli were excluded from analysis.
We calculated m(t) and 2(t) as:
| (2) |
| (3) |
where xj(t) is the current elicited by the jth stimulus, N is the number of stimuli in the train (500) and 02 is the baseline variance measured before the first voltage step. This formula for 2(t) is based on the squared differences between consecutive traces rather than the squared deviations from the global mean, a procedure which minimizes the effects of run-down (Roberts et al. 1990). To allow time for the capacitive transients to settle and ICa to become essentially fully activated, analysis of m(t) and 2(t) began at t = 0.4 ms after the beginning of the voltage step, by which time ICa is >95 % activated (Fig. 1C; m = 0.093 ms at -5 mV and is faster at more depolarized potentials; Armstrong & Roberts, 1998).
We fitted our variance vs. mean data with a model in which two types of BK channels, inactivating (BKI) and non-inactivating (BKN) channels, contribute to m(t) and 2(t). The two populations of channels were assumed to be identical in all respects, except that BKI channels underwent complete inactivation, while BKN did not inactivate at all. Using these assumptions the total current, m(t), was separated into the sum of transient and steady-state components:
| m(t) = Itrans(t) + Is-s(t), | (4) |
as shown in Fig. 1B and C. These definitions of Itrans and Is-s differ from the definitions of Itransient and Isteady because ICa and a leak current (Ileak) also contributed to m(t). During the time interval analysed, however, ICa and Ileak are only expected to add a constant offset (xoffset) to m(t), which was subtracted during the fitting procedure described below. Is-s(t) was described as the product of an activation time course, f(t), and a steady-state level (Ks-s):
| Is-s(t) = Ks-sf(t). | (5) |
Itrans(t) contained an additional exponential component to account for inactivation:
|
Itrans(t) = Ktransf(t) e-t/D.
| (6) |
Note that the same f(t) was used to calculate both Itrans(t) and Is-s(t) (see Results section). Ks-s, Ktrans and D, the time constant of decay of Itrans(t), were determined as shown in Fig. 1B and f(t) was calculated as:
| (7) |
Ensemble-variance data were then fitted to an equation that contained two sets of terms, one for the variance of Itrans and one for the variance of Is-s:
| (8) |
where the four fitted parameters were single-channel conductance (i), the number of channels contributing to Itransient (Ntransient), the number of channels contributing to Isteady (Nsteady) and the current offset (xoffset). This equation assumes that BKI and BKN channels gate independently, that ICa and Ileak do not contribute significantly to the variance and that i was the same for both channels. For statistical reasons (Roberts et al. 1990), averaged values of Nsteady and Ntransient were computed as the reciprocal of the means of 1/Nsteady and 1/Ntransient. Although what we believe to be a contaminating KV current may have contributed to our ensemble-variance analysis data (see Results), in the majority of cells this current was much smaller than Itransient and Isteady and the unitary conductance of KV currents are much smaller than BK channels (Hille, 1992). Thus, it is unlikely that this current contributed significantly to either m(t) or 2(t).
Estimates of i for Itrans alone were obtained by fitting only the falling phase of the outward current (Fig. 8B and D). To do this, steady-state values of 2 and m measured at the end of the second voltage step were subtracted before the data were fitted with an equation for variance that contained just the Itrans terms with an added xoffset.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The current revealed by 4-AP
We have previously shown that when 4-AP is applied to frog saccular hair cells in a semi-intact epithelial preparation the presence of a rapidly activating, partially inactivating outward current is revealed (Armstrong & Roberts, 1998; Fig. 2A). Block of this 4-AP-insensitive current by external TEA and its absence in recordings in which all of the K+ in the internal solution was replaced by Cs+ (Armstrong & Roberts, 1998) suggests that it is carried through K+ channels. This current was activated when the membrane potential was depolarized from the holding potential of -70 mV to levels more positive than -60 mV, becoming both larger and faster between -50 and -30 mV. Between -30 and +25 mV the time to peak remained constant at ~1.5 ms while the amplitude continued to grow with increasing depolarization. With depolarization above +25 mV activation slowed again, often becoming so slow that the current did not peak before the 10-15 ms voltage step ended (Fig. 2A). Inactivation of the transient component was always well approximated by a single exponential function (Fig. 3). Within the range of membrane potentials where the rate of activation remained constant (-30 to +25 mV), the time constant of decay (D) of this component was also constant. Above or below this voltage range, where the current activated more slowly, inactivation was also slower. Such a correlation between activation and inactivation kinetics suggests that channel opening might be required before inactivation can occur (Hille, 1992). This type of rapidly activating, partially inactivating current was found in all hair cells in this preparation (75 cells).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 2. Blocking KV with 4-AP reveals a rapidly activating, partially inactivating outward current A, a perforated-patch recording from a hair cell bathed in 10 mM 4-AP. Currents were elicited by depolarizing voltage steps (command potentials between +90 and -60 mV, in 10 mV increments) from a holding potential of -70 mV. The largest amplitude voltage reached (RS corrected) is indicated next to the step protocol. B, plot of the peak I-V relationship for the cell in A. | ||
 |
View larger version [in this window] [in a new window] |
|
|
Figure 3. Variability in the ratio of atransient to asteady Perforated-patch recordings from hair cells bathed in 10 mM 4-AP that exhibited large (A), intermediate (B), and small (C) amounts of Itransient relative to Isteady. The thick grey lines are exponential fits showing the time constant of decay ( | ||
The 4-AP-insensitive K+ current has two components
After estimating and subtracting ICa, which at -20 mV contributed on average ~330 pA of steady inward current, we found that all 75 hair cells in our population had a measurable steady-state outward current (Isteady) in addition to the more visually prominent transient component (Itransient). In fact, the mean amplitudes of Itransient and Isteady at -20 mV were roughly equal (atransient = 623 ± 370 pA, asteady = 707 ± 515 pA; mean ± S.D., n = 75), although their relative contributions to the recorded current varied considerably from cell to cell (Fig. 3). For example, in one hair cell Itransient contributed only 14 % of the peak outward current, while in another it contributed 90 %. On average, however, they contributed nearly equally to the peak outward current (Itransient was 51 ± 17 %; mean ± S.D., n = 75). In hair cells in which a large percentage of the peak outward current was due to Itransient (75-90 %) the average D was slightly faster (~2.0 ms) compared to hair cells in which Itransient made a small contribution (15-20 %, ~2.8 ms). This trend was found to be statistically significant using regression analysis (r = -0.48; n = 75; P < 0.001).
Between -60 and +30 mV the peak amplitude of this 4-AP-insensitive current rose monotonically. At potentials above approximately +30 mV, however, the peak amplitude fell precipitously (Fig. 2B), despite an increased driving force on K+. Such a decline was observed regardless of whether Itransient or Isteady was the dominant current. This decrease in current amplitude at large positive potentials was observed in 63 of the 65 hair cells that were sufficiently depolarized. Such a decrease at very depolarized potentials is expected for Ca2+-dependent currents, due to the suppression of Ca2+ influx as the membrane potential approached the equilibrium potential for Ca2+ (~100 mV). Thus, the data suggest that both Itransient and Isteady are KCa currents. Further support for the Ca2+ dependence of these currents is given next.
Ca2+ dependence may also explain why the activation kinetics became very slow during large depolarizations. One can imagine that as the membrane potential approaches the Ca2+ suppression potential, Ca2+ trickles, rather than flows into the hair cell (Hille, 1992). Under these conditions it might take longer for the Ca2+ concentration required for channel opening to be reached, particularly if the cytoplasm contains a substantial amount of immobile Ca2+ buffer that resists the increase in intracellular Ca2+ until the buffer saturates.
In several hair cells the amplitude of the outward current rose again at potentials greater than +75 mV, yielding an 'N'-shaped rather than an upside-down 'V'-shaped I-V relationship (for an example, see right panel of Fig. 6B). In two cells, the amplitude of the peak outward current never declined above +75 mV and the I-V relationships from these cells looked as if they belonged to KV currents. As will be discussed further later, we believe that such an increase in current above +75 mV is indicative of the presence of a contaminating Ca2+-independent outward conductance that was not blocked by 10 mM 4-AP. Alternatively, this increase in outward current at very depolarized potentials could reflect Ca2+-independent activation of a conductance that at more physiological potentials is Ca2+ dependent (Barrett et al. 1982). In either case, little if any of this contaminating outward current appeared to be present at -20 mV and for this reason much of the Itransient and Isteady characterization was carried out at this potential.
Itransient is not a 4-AP-induced artefact
Initially we were concerned that the observed transient current might be a pharmacological effect of 4-AP. In the squid giant axon, internally applied quaternary ammonium (QA) compounds, such as TEA, can act as 'open-channel blockers', blocking the KV conductance in these cells by lodging in its pore after it has opened (Armstrong, 1966, 1969). Because QA compounds block only after the channel has opened, the current recorded in the presence of these compounds can be transient, although the channel itself does not inactivate (French & Shoukimas, 1981). Although in our experiments 4-AP was applied extracellularly, it is a membrane-permeant compound and thought to block K+ channels at an intracellular site (Kirsch & Narahashi, 1983; Kirsch & Drewe, 1993). Although 4-AP is not known to act as an 'open-channel blocker', we considered the possibility.
Several pieces of evidence show that 4-AP was not having such an effect. First, increasing the 4-AP concentration by 20-fold had no effect on D. In 1 mM 4-AP, D was measured to be 2.1 ± 0.5 ms at -20 mV (mean ± S.D.), while in 20 mM 4-AP, D was 2.2 ± 0.4 ms (n = 6, each cell was tested at both concentrations). If 4-AP was behaving as an open-channel blocker, increasing its concentration would be expected to speed up the apparent inactivation. Second, evidence of a transient current was seen even in the absence of 4-AP. Although in recordings made in normal extracellular saline the slower activating KV current partially obscures the transient current, a 'hump' could sometimes be seen (Fig. 4A, left panel; for other examples, see Armstrong & Roberts, 1998). Finally, the transient current could be isolated from KV without using 4-AP. When the TEA-sensitive current (recall that both Itransient and Isteady are TEA sensitive) was reconstructed by subtracting the current recorded in 6 mM TEA from that recorded in normal extracellular saline (Fig. 4B) a transient current very similar to the current recorded from the same cell in 10 mM 4-AP was found. Similar results were obtained in eight other cells. From these three lines of evidence, we conclude that the transient current revealed in the presence of 4-AP is due to channel inactivation.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 4. Rapid inactivation is not an odd effect of 4-AP A, perforated-patch recordings from a single hair cell bathed sequentially in normal extracellular saline, 6 mM TEA and 10 mM 4-AP. The thin trace below the recording in 6 mM TEA shows the estimated ICa. The currents at -30 mV were estimated by linear interpolation of the current from adjacent steps. B, the TEA-sensitive current was obtained by subtracting the current recorded in 6 mM TEA from the current recorded in normal extracellular saline. For comparison with the recordings in A, the estimated ICa was added to the trace in B. The larger size of the TEA-sensitive current compared to the current recorded in 4-AP is likely to be the result of current run-down during the time course of these recordings. | ||
Steady-state inactivation and recovery from inactivation
The resting membrane potentials of frog saccular hair cells are around -70 mV, and in response to depolarizing current injection they exhibit resonance of frequencies between 35 and 75 Hz at potentials of -50 to -60 mV (Ashmore, 1983; Armstrong & Roberts, 1998). To gauge whether Itransient has appropriate voltage-dependence and kinetics to contribute in shaping the electrical behaviour of these cells, we examined the time course of recovery of Itransient from inactivation and the steady-state dependence of the amplitude of Itransient on the membrane potential (steady-state inactivation). Recovery from inactivation was examined using a standard paired-pulse protocol (Fig. 5A). Recovery of Itransient was rapid and followed a time course that was well approximated by a single exponential function. For the 14 hair cells tested, the mean recovery time constant (R) was 9.7 ± 4.7 ms (mean ± S.D.).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 5. Recovery from inactivation and steady-state inactivation of Itransient A, Itransient recovers rapidly from inactivation. Itransient inactivated during the first 10 ms voltage step to -11 mV (RS corrected; command potential was 0 mV). After stepping back to the holding potential (-70 mV) for between 2 and 30 ms (tested in 2 ms intervals), a second 10 ms voltage step was applied. The dotted line is an exponential fit of the recovery from inactivation, with time constant | ||
The stimulus protocol used to examine steady-state inactivation is described in the legend to Fig. 5B. The amplitude of Itransient following each prepulse was normalized by dividing it by the amplitude of Itransient recorded following the prepulse to -100 mV. The data were then fitted with an equation for steady-state inactivation (h
):
|
h = 1/(1 + e(V-V1/2)/k),
| (9) |
where V is the prepulse voltage, V1/2 is the voltage at which atransient was half-maximal and k determines how steeply h depends upon the prepulse potential (Hodgkin & Huxley, 1952). As the steady-state inactivation curve shows (Fig. 5B), the transient current goes from fully available to completely inactivated within the range of membrane potentials exhibited by these cells (-50 to -70 mV).
Both the onset and recovery of Itransient from inactivation are sufficiently rapid to occur during each cycle of electrical resonance, particularly in cells that oscillate near the low end of the 35-75 Hz range. Figure 5B shows that during a ±5 mV oscillation around a baseline of -55 mV, Itransient is expected to approach a steady-state level of ~80 % inactivation at the most hyperpolarized excursion of each cycle, leaving ~20 % available to undergo rapid activation and inactivation during the next depolarizing phase. Thus, Itransient has the potential to influence the electrical resonance in these low-frequency cells in a cycle-by-cycle manner.
In further support of Ca2+ dependence
The I-V relationships of Itransient and Isteady (Fig. 2B) suggest that these currents are Ca2+ dependent. To test this further, we examined the effects of blocking Ca2+ influx on these currents by either applying Cd2+, a non-specific Ca2+ channel blocker, or removing Ca2+ from the extracellular saline. In all eight hair cells tested, the majority of the 4-AP-insensitive outward current present at +30 mV (Fig. 6A), the potential at which Itransient and Isteady are maximal, was blocked by applying 100 µM Cd2+. In six out of seven of these cells, a small current remained in Cd2+. Unlike Itransient and Isteady, however, this Cd2+-insensitive current had slow activation kinetics and increased monotonically in amplitude with depolarization, suggesting the presence of a contaminating Ca2+-independent, voltage-dependent K+ current distinct from Itransient and Isteady. The I-V relationship of the seventh cell was slightly 'N'-shaped in Cd2+, suggesting that Itransient and Isteady were not fully blocked and that the contaminating current was present. Upon removal of Cd2+ recovery of Itransient and Isteady was nearly complete in six of the seven hair cells tested (Fig. 6A, right panel).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 6. Ca2+ dependence of Itransient and Isteady A, application of 100 µM Cd2+ reversibly eliminated both Itransient and Isteady, leaving a small residual outward current. B, the effect of removing external Ca2+ was similar to that of Cd2+, although a larger outward current remained. The panels on the right show the peak I-V relationship for each cell before ( | ||
The effect of removing extracellular Ca2+ from hair cells bathed in 10 mM 4-AP was examined in six cells (Fig. 6B). In five of these cells the majority of Itransient and Isteady were eliminated. For two of these cells the I-V relationships were monotonic in the absence of Ca2+, suggesting that Itransient and Isteady were fully blocked and that the same contaminating current as in the Cd2+ experiments was present. The I-V relationships of the three other cells were slightly 'N'-shaped (Fig. 6B, right panel), again suggesting that Itransient and Isteady were not fully blocked. In all five hair cells, Itransient and Isteady recovered fully when extracellular Ca2+ was restored. In the sixth cell, Itransient and Isteady became slightly larger when Ca2+ was removed, a conflicting result that we are at a loss to explain. Nonetheless, the sensitivity of Itransient and Isteady in the majority of cells to treatments that block Ca2+ influx indicates that, as their I-V relationships suggest, these currents are Ca2+ dependent.
Sensitivity to iberiotoxin and apamin
To date, IbTX (Galvez et al. 1990) is known to block only BK channels. For this reason, we chose to test whether IbTX blocks Itransient and Isteady. In the majority of hair cells (6/9) application of 100 nM IbTX eliminated essentially all of both currents, often revealing the underlying ICa (Fig. 7A). As above, a small amount of contaminating Ca2+-independent outward current was often also revealed when IbTX was applied. In the three other hair cells, IbTX blocked all of Itransient, but blocked only ~60 % of Isteady (the I-V relationships of these cells were slightly 'N'-shaped in IbTX), possibly indicating a differential sensitivity of Itransient and Isteady to IbTX. Block by IbTX was irreversible during the time course of these experiments. In contrast to IbTX, application of 1 µM apamin, a bee toxin that blocks small-conductance KCa channels (SK; Vergera et al. 1998), had no obvious effect on the outward current (Fig. 7B; n = 6), suggesting that not only are Itransient and Isteady apamin insensitive, but that under these conditions the contribution of SK to the outward current is minimal. Thus, Itransient and Isteady are sensitive to a toxin specific for BK channels, but not to a toxin which blocks a different class of KCa channels.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 7. The effects of IbTX and apamin on Itransient and Isteady Application of 100 nM IbTX eliminated both Itransient and Isteady in this hair cell (A), while in another hair cell 1 µM apamin had no effect (B). The right-hand panel in A shows the peak I-V relationship for this cell before ( | ||
Ensemble-variance analysis
BK channels are well known for having large single-channel conductances (between 150 and 300 pS in symmetrical K+; Latorre et al. 1989; McManus, 1991; Vergara et al. 1998). To determine whether the ion channels through which Itransient and Isteady flow can be classified as BK based upon this criterion, we performed ensemble-variance analysis on these currents (Sigworth, 1980). This method is based upon measuring the fluctuations in current amplitude during repeated presentations of a depolarizing voltage step. In addition to providing an estimate of the single-channel current (i), ensemble-variance analysis also provided estimates of the number of channels (Ntransient and Nsteady) that contributed to the mean current.
The left panels of Fig. 8 show two examples of mean current (m(t); eqn (2)) and variance (2(t); eqn (3)) calculated from whole-cell currents recorded in response to 500 voltage steps applied to hair cells bathed in 10 mM 4-AP. Ntransient, Nsteady and i were estimated by fitting the plots of 2(t) vs. m(t) with an equation for the sum of variances contributed by the two populations of channels (one of which undergoes inactivation and one of which does not). The equation used assumes that the time course of activation and i are the same for the two populations of channels (Fig. 8B and D; eqn (8)). The assumption that the activation kinetics of Itransient and Isteady were the same was tested by comparing the rate of activation of the outward current in recordings containing different proportions of each current. After subtracting ICa from deconvolved current traces (see Methods section), a sigmoidal curve was fit to the rising phase of the outward current. No difference in the activation time course was found in cells that contained predominantly Itransient or predominantly Isteady.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 8. Ensemble-variance analysis of Itransient and Isteady in two hair cells A, mean current (m(t)) and variance ( | ||
The variance model used to evaluate our ensemble-variance analysis data provided a good fit to our 2(t) vs. m(t) plots, both during the times when Itransient and Isteady were activating (rising phase of the 2(t) vs. m(t) plot) and while Itransient was inactivating (falling phase; Fig. 8B and D). Because Itransient and Isteady activate much faster than Itransient inactivates, information about the channels that carry Itransient alone could be obtained by analysing the data collected near the end of the voltage step, when Isteady had reached steady-state levels and Itransient was still inactivating. Once Isteady reaches steady-state it contributes only a constant offset to m(t) and 2(t) and therefore the changes in m and 2 seen at these later times can be attributed entirely to the channels carrying Itransient. By fitting our 2(t) vs. m(t) plots at these times with a variance model in which just one population of channels contributes to m and 2 we obtained an estimate of i for the channels that carry Itransient. For 11 of the 13 recordings, the estimate of i obtained from fitting Itransient alone was quite similar (<15 % difference) to that obtained from fitting all of the data points with the variance model in which two populations of channels contributed to m and 2. The similarity in the estimates of i and the excellence of our fits corroborates our assumption that i was the same for both Itransient and Isteady.
To determine i at 0 mV and the single-channel conductance (g), the estimates of i obtained from fitting all of the data points were plotted against the respective voltages at which they were made. Fitting these points with a straight line yielded i 3.6 pA at 0 mV and g 97 pS (Fig. 9). The estimated g from the one hair cell in which we were able to obtain data at two different voltages was slightly smaller (79 pS). Because these recordings were made under non-symmetrical ionic conditions (i.e. under typical whole-cell recording conditions, with an intracellular K+ concentration of 120 mM and an extracellular K+ concentration of 2 mM), we expect g to be roughly 2 times greater under symmetrical ionic conditions at the potentials tested here (Pallotta et al. 1981; Methfessel & Boheim, 1982). Thus, the single-channel conductance of Itransient and Isteady is certainly large enough to classify them as BK channels. The estimates of Ntransient and Nsteady obtained from this analysis were 636 ± 203 and 394 ± 268 (mean ± S.D.; n = 13), respectively.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 9. Single-channel currents estimated from ensemble-variance analysis vs. membrane potential The line fitted to the data indicates that the g for Itransient and Isteady is 97 pS and that i = 3.6 pA at 0 mV. The filled squares ( | ||
In addition to providing an estimate for the size of the channels that carried Itransient and Isteady, this analysis also supports the hypothesis that these currents were carried through two separate populations of channels, one of which inactivates and one that does not. The 2(t) vs. m(t) plot for a single population of channels should follow a single parabolic trajectory, with the maximum variance occurring when the open probability (Popen) equals 0.5. During inactivation of this population of channels the 2(t) vs. m(t) plot should double back on top of the rising phase: points having the same m(t) would also have the same 2(t). Instead of following a single parabolic trajectory, the 2(t) vs. m(t) plots from 10 of 13 cells that we studied were distinctly double-parabolic: 2(t) was larger during the rising phase than at the corresponding m(t) during the falling phase (Fig. 8D). A single population of channels cannot account for this behaviour. Our model explains the 2(t) vs. m(t) plot shown in Fig. 8D as follows. Depolarization from -70 to -5 mV caused both populations of channels to activate over a similarly rapid time course. As for the single population of channels, the 2(t) vs. m(t) plot follows a parabolic trajectory during activation (open squares), with 2(t) reaching a maximum (~10 nA2) when Popen = 0.5 (corresponding to m(t) = 4 nA). Popen for the two channels reaches its maximum (~0.75) when the current peaks (~6 nA). During inactivation (filled triangles), Popen for the two populations diverges, remaining high (P > 0.75) for the channels contributing to Isteady, but falling to zero for the Itransient channels. When Popen for the Itransient channels falls back to 0.5 the variance peaks a second time (~7 nA2). Because only a portion of the channels contributing to the variance seen during activation contribute to inactivation the peak of this second parabola is smaller. Thus, the double-parabolic shapes of our 2(t) vs. m(t) plots suggest that Itransient and Isteady are carried through two separate populations of channels.
Together, these data indicate that Itransient and Isteady flow through two separate, but similar, populations of channels: inactivating BK channels (BKI) and non-inactivating BK channels (BKN).
Inactivation can be removed by external papain
Rapid inactivation of ion channels is generally thought to be mediated by a 'ball-and-chain'-type mechanism, similar to that first proposed for Na+ channels (Armstrong & Bezanilla, 1977). According to this model, a cytosolic structure (the 'ball') tethered (via a 'chain') near the pore of the ion channel blocks conduction by occluding the pore after the channel has opened. One test for such a mechanism is to transiently apply a protease (commonly trypsin) to the intracellular face of the channel, to cleave the inactivation ball from the channel and thereby remove inactivation. In rat chromaffin cells brief exposure of the intracellular face of the BKI channel to trypsin removes inactivation, suggesting that such a mechanism mediates inactivation of these channels (Solaro & Lingle, 1992; Ding et al. 1998). Unfortunately, rather than selectively removing inactivation of Itransient, addition of trypsin (0.25 or 0.5 mg ml-1) to the intracellular solution eliminated essentially all of both Itransient and Isteady (in six of seven cells). This was not observed in control experiments of similar duration but in which trypsin was not added (two cells). Considering that a significant portion of the BK channel including the Ca2+-sensing domain resides in the cytosol (Wei et al. 1994; Schreiber & Salkoff 1997), elimination of Itransient and Isteady currents by chronic application of protease is not surprising.
Our previous data (Armstrong & Roberts, 1998) showed that exposure of frog saccular hair cells to external papain dramatically transforms their electrical properties. In addition to eliminating KV and having several apparently unrelated effects on other channels, papain eliminates Itransient leaving a BKN current (i.e. Isteady) as the only major outward current (Lewis & Hudspeth, 1983; Hudspeth & Lewis, 1988a; Roberts et al. 1990; Armstrong & Roberts, 1998). Because the known pharmacological properties and the single-channel conductance of Itransient and Isteady are similar to the BKN current of enzymatically dissociated hair cells (Hudspeth & Lewis, 1988a; Roberts et al. 1990), we considered the possibility that papain transforms Itransient into Isteady by removing inactivation. To test this possibility, and to determine whether papain itself rather than one of the other constituents of the crude papain that is used in the dissociation procedure (see Armstrong & Roberts, 1998) is the active agent, we applied a solution containing purified (twice crystallized) papain (8 units ml-1) onto hair cells bathed in 10 mM 4-AP. In three hair cells exposed to this solution for approximately 10-14 min, the partially inactivating outward current was transformed into a larger outward current with minimal inactivation (Fig. 10). In a fourth hair cell, exposed to papain for only ~4 min, ~40 % of inactivation was eliminated and a larger steady-state current was found. No changes in the amount of inactivation were observed in recordings of similar duration in hair cells not exposed to papain. The increase in the size of the outward current following exposure to papain argues that rather than eliminating Itransient, papain transforms Itransient into Isteady.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 10. External application of papain removes inactivation Perforated-patch recording from a single hair cell bathed in 10 mM 4-AP before (thick trace) and after (thin trace) bath application of 8 units ml-1 of twice crystallized papain for approximately 10 min. The current at -20 mV was estimated by linear interpolation of currents recorded from adjacent steps. | ||
Inactivation can be restored by an inactivation 'ball'
Recently, two BK channel auxiliary subunits that can confer inactivation on BK channels have been identified: 2 (Wallner et al. 1999; Xia et al. 1999) and 3 (Uebele et al. 2000; Xia et al. 2000). Interaction of the BK channel 
subunit with one of these subunits is the only mechanism known to confer rapid inactivation on BK channels. Although it is unknown whether either subunit is present in frog saccular hair cells, we tested whether the first 19 amino acids of the N-terminus of 2 (the putative 'inactivation ball'), which are sufficient to confer inactivation on heterologously expressed BK channels (Wallner et al. 1999), could restore inactivation in enzymatically dissociated hair cells. Frog saccular hair cells were dissociated in papain to remove inactivation and the 2 inactivation ball peptide was introduced into the cytoplasm by adding it to the recording pipette. Low concentrations (3 or 10 µM) of peptide did not cause appreciable inactivation during a 200 ms depolarization (n = 4). Higher concentrations (100 and 300 µM) did induce inactivation that, like native BKI inactivation, followed a single exponential time course (Fig. 11A). The time constant of inactivation, D, became progressively faster over time after the whole-cell configuration was established (Fig. 11A), reflecting the time needed to dialyse the cell with the contents of the pipette. Often the rate of inactivation did not completely reach a maximal rate before the end of the experiment, so the expected maximal rate of inactivation (D-MAX) was deduced by fitting plots of D vs. time after the whole-cell configuration was established with a single exponential function (Fig. 11B). With 300 µM of the 'ball' peptide applied internally, D-MAX was 8.4 ± 2.9 ms (mean ± S.D.; n = 6) at -40 mV, a rate approaching that of native BKI channels (~2.5 ms). Although we cannot be certain whether the peptide reached the same concentration in the cytoplasm as in the pipette, a similar concentration (100 µM) of the Shaker inactivation ball peptide is required to restore inactivation to mutant Shaker potassium channels that lack the N-terminus (Zagotta et al. 1990).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 11. Internally applied 'ball' peptide confers inactivation to the non-inactivating BK current in enzymatically dissociated hair cells A, whole-cell voltage-clamp recordings made over time from an enzymatically dissociated hair cell to which a 300 µM 'ball' peptide was added internally (via the pipette). Voltage steps to -39 mV (RS corrected) were presented every 10 s. For clarity only the response to every other step is shown. The first recording (arrow) was made roughly 30 s after achieving the whole-cell configuration. B, plot of the time constant of decay ( | ||
'Ball'-induced inactivation was essentially complete (Fig. 11A), suggesting that all of the BK channels can be blocked by the 'ball' domain and that if BK channel inactivation in hair cells is conferred by 2, or a similar subunit, not all BK channels have an associated subunit. The effect of the 'ball' peptide appeared to be specific for BK channels; no discernible difference was seen in either the ICa or the inward rectifier current that is present in these cells (Holt & Eatock, 1995) when the peptide was present. In contrast to native inactivation, the inactivation conferred by the 'ball' domain was slightly voltage dependent: between -30 and +25 mV D-MAX became slightly (~15 %) faster.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The outward current in frog saccular hair cells that is revealed by applying 4-AP contains two components: Itransient and Isteady (Fig. 2A and Fig. 3). Although other compounds that, like 4-AP, block ion channels at intracellular sites can produce transient K+ currents by acting as 'open-channel blockers', the appearance of Itransient is not a 4-AP-induced artefact (Fig. 4). From previous work (Armstrong & Roberts, 1998) we know that these currents are carried through K+ channels. Suppression of Itransient and Isteady by large depolarizations (Fig. 2B) and their susceptibility to treatments that block Ca2+ influx (Fig. 6) show that these currents are Ca2+ dependent. Sensitivity to IbTX but not apamin (Fig. 7) suggests that both currents are carried through the BK class of KCa channels. A large single-channel conductance (~100 pS in non-symmetrical K+; Fig. 9), as determined by ensemble-variance analysis (Fig. 8), also points to the conclusion that both Itransient and Isteady flow through BK channels. These data also suggest that these currents are carried through two separate populations of BK channels, one of which inactivates (BKI), while the other (BKN) does not.
Electrical resonance in these cells typically involves oscillations of the membrane potential at frequencies between 35 and 75 Hz, around a baseline of ~55 mV (Ashmore, 1983; Armstrong & Roberts, 1998). Examination of the time constants for inactivation (D 2.5 ms) and recovery from inactivation (R 10 ms at -70 mV; Fig. 5A), and of steady-state inactivation of Itransient (Fig. 5B), suggests that this current may undergo significant activation, inactivation and recovery during each cycle of resonance. Thus, Itransient may contribute significantly to electrical tuning in these cells. The contribution of this current in shaping the electrical behaviour of these cells could be further enhanced by efferent-induced hyperpolarization (Art et al. 1984). In order to elucidate what role this current plays in shaping resonance in frog saccular hair cells further studies will be necessary to explore the detailed behaviour of Itransient throughout the physiological range of membrane potentials.
Inactivation of BKI was removed by extracellular application of papain (Fig. 10), and could be restored by intracellular application of the 2 'ball' peptide (Fig. 11). Although not all of the characteristics of the restored inactivation were identical to native BKI inactivation, the ability of this peptide to induce inactivation supports the hypothesis that a 'ball-and-chain' mechanism, perhaps conferred by another subunit such as 3, mediates BKI channel inactivation in frog saccular hair cells. Although our hypothesis that inactivation involves an intracellular 'ball-and-chain' mechanism may seem to conflict with elimination of inactivation by extracellular protease a plausible model for how these two might be consolidated is presented below.
BK channel structure
All BK channels cloned thus far, including those in hair cells (Jiang et al. 1997; Navaratnam et al. 1997; Rosenblatt et al. 1997; Jones et al. 1998), are homologues of slowpoke (Slo), which was first cloned from Drosophila melanogaster (Atkinson et al. 1991). Nevertheless, properties of BK channels, such as Ca2+ and voltage dependence and kinetics vary significantly from cell to cell. Such diversity is likely to arise from alternative splicing of Slo mRNA (Adelman et al. 1992; Tseng-Crank et al. 1994; Navaratnam et al. 1997; Rosenblatt et al. 1997; Saito et al. 1997; Jones et al. 1998), interaction of BK with auxiliary subunits (McManus et al. 1995; Tseng-Crank et al. 1996; Schopperle et al. 1998; Xia et al. 1998; Wallner et al. 1999; Xia et al. 1999; Uebele et al. 2000; Xia et al. 2000) and post-translational modulatory processes (Lechleiter et al. 1988; Reinhart & Levitan, 1995; Wang et al. 1997).
The core of the BK channel is similar to that of KV channels, with six membrane-spanning segments (S1-S6), complete with a positively charged voltage sensor (S4) and a pore-forming region (P-loop) between segments S5 and S6. In addition to this KV-like core, BK channels possess four hydrophobic segments on their C-terminus (S7-S10) and an additional membrane-spanning segment on their N-terminus (S0; Wallner et al. 1996; Meera et al. 1997). Segments S7-S10 of the BK channel reside in the cytosol and contain the Ca2+-sensing domain (Wei et al. 1994; Schreiber & Salkoff, 1997). The additional membrane-spanning segment (S0) on the N-terminus leaves the N-terminal of the BK channel on the extracellular side of the membrane (Wallner et al. 1996; Meera et al. 1997), creating potential cleavage sites for an extracellular protease. As for other K+ channels, four of these pore-forming subunits are required to form a functional BK channel.
Inactivating BK channels
Although BKI channels have been reported in several cell types including skeletal muscle (Pallotta, 1985), hippocampal neurons (Ikemoto et al. 1989; McLarnon, 1995; Hicks & Marrion, 1998) and pancreatic cells (Li et al. 1999), the BKI channel of rat chromaffin cells serves as the only well-characterized example (for a review see Lingle et al. 1996). Like Shaker B (ShB) K+ channels (Hoshi et al. 1990; Zagotta et al. 1990), inactivation of the rat chromaffin BKI channel seems to involve a 'ball-and-chain' mechanism: inactivation is conferred by multiple cytosolic domains which are susceptible to intracellular trypsin (Solaro & Lingle, 1992; Ding et al. 1998). Failure to identify a Slo splice variant in rat chromaffin cells (Saito et al. 1997) implies that unlike for ShB channels, the inactivating 'ball' domain is not contained on the subunit of BK channels.
Coexpression of the BK channel subunit with 2, a BK channel auxiliary subunit, produces a BKI current with characteristics similar to native rat chromaffin BKI channels (Xia et al. 1999). 2 was discovered by virtue of its sequence similarity to 1 (Knaus et al. 1994a,b), an auxiliary subunit that increases the Ca2+ sensitivity of BK channels but does not bestow inactivation (McManus et al. 1995). Like 1, 2 has two transmembrane domains which are connected via a large extracellular loop. Both the N- and C-termini of 2 are intracellular, and the N-terminus contains the 'ball' domain that confers inactivation (Wallner et al. 1999; Xia et al. 1999). Recently, another subunit, 3 (also identified by its similarity to 1), that can also confer inactivation to heterologously expressed BK channels was identified (Uebele et al. 2000; Xia et al. 2000).
Although the sites of interaction between 2 or 3 and the BK channel have not been examined, it is known that the first 31 amino acids at the extracellular N-terminus of the BK channel are required for 1 interaction (Wallner et al. 1996). Given the extensive sequence similarity among subunits (Xia et al. 2000) it is reasonable to suppose that 2 and 3 interact with the same sites on the BK channel. If true, an extracellular enzyme could disrupt inactivation by cleaving the large extracellular loop of the subunit or the N-terminus of the BK channel itself. In either case, interaction between the subunit and the BK channel could be disrupted and inactivation could be eliminated. Thus, association of an auxiliary subunit like 2 or 3 with some of the BK channels in frog saccular hair cells would provide a nice explanation for the data presented here.
BK channels in non-mammalian vertebrate hair cells
Recently, the BK channels of non-mammalian vertebrate hair cells have received much attention. The endogenous electrical resonant behaviour exhibited by these cells is thought to play a major role in tuning them to specific frequencies of sound (for a review see Fettiplace & Fuchs, 1999). Generation of resonance involves the interplay of currents flowing through voltage-dependent Ca2+ channels and K+ channels. In higher frequency hair cells (>60 Hz), BKN channel kinetics alone seem to determine resonant frequency (Art & Fettiplace, 1987; Art et al. 1995): in the turtle cochlea, where hair cell tuning varies systematically (tonotopically; Crawford & Fettiplace, 1980) BKN channel kinetics also vary systematically. Both alternative splicing of the subunit (Navaratnam et al. 1997; Rosenblatt et al. 1997; Jones et al. 1998) and expression of a 1 subunit (Ramanathan et al. 1999) may be involved in generating the observed range of BK channel kinetics and Ca2+ dependence (Jones et al. 1999; Ramanathan et al. 2000) in tonotopically organized sensory organs.
Our demonstration that extracellular papain eliminates BKI inactivation explains why this current was not described in previous studies of frog saccular hair cells, and raises the possibility that BKI channels have gone undetected in the hair cells of other species because papain was used to dissociate the cells. It has been reported that turtle cochlear hair cells dissociated without papain have electrical properties similar to papain-treated cells (Goodman & Art, 1996), but it is unclear whether the currents were examined in sufficient detail to rule out the presence of BKI channels in these cells. We are aware of only one other study in which the potassium currents in hair cells from a non-mammalian species were systematically examined without enzymatic dissociation (a slice preparation of the frog crista ampullaris; Masetto et al. 1994), and in that study a transient current similar to the BKI current we describe here was reported. Thus, our finding that BKI inactivation is susceptible to extracellular papain, an enzyme common to dissociation procedures, opens up the possibility that BKI channels are present in many other cell types. This might explain why reports of BKN channels are commonplace, while reports of BKI channels are rare.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| ADELMAN, J. P., SHEN, K. Z., KAVANAUGH, M. P., WARREN, R. A., WU, Y. N., LAGRUTTA, A., BOND, C. T. & NORTH, R. A. (1992). Calcium-activated potassium channels expressed from cloned complementary DNAs. Neuron 9, 209-216 | [Medline] |
| ARMSTRONG, C. E. & ROBERTS, W. M. (1998). Electrical properties of frog saccular hair cells: distortion by enzymatic dissociation. Journal of Neuroscience 18, 2962-2963 | [Abstract/Full Text] |
| ARMSTRONG, C. E. & ROBERTS, W. M. (1999). A very rapidly inactivating calcium-dependent potassium current in frog saccular hair cells. Biophysical Journal 76, A74 | |
| ARMSTRONG, C. M. (1966). Time course of TEA+-induced anomalous rectification in squid giant axons. Journal of General Physiology 50, 491-503 | [Medline] |
| ARMSTRONG, C. M. (1969). Inactivation of the potassium conductance and related phenomena caused by quaternary ammonium ion injection in squid axons. Journal of General Physiology 58, 413-437 | |
| ARMSTRONG, C. M. & BEZANILLA, F. (1977). Inactivation of the sodium channel II: gating current experiments. Journal of General Physiology 70, 567-590 | [Abstract] |
| ART, J. J. & FETTIPLACE, R. (1987). Variation of membrane properties in hair cells isolated from the turtle cochlea. Journal of Physiology 385, 207-242 | [Abstract] |
| ART, J. J., FETTIPLACE, R. & FUCHS, P. A. (1984). Synaptic hyperpolarization and inhibition of turtle cochlear hair cells. Journal of Physiology 356, 525-550 | [Abstract] |
| ART, J. J., WU, Y.-C. & FETTIPLACE, R. (1995). The calcium-activated potassium channels of turtle hair cells. Journal of General Physiology 105, 49-72 | [Abstract] |
| ASHMORE, J. F. (1983). Frequency tuning in a frog vestibular organ. Nature 304, 536-538 | [Medline] |
| ATKINSON, N. S., ROBERTSON, G. A. & GANETZKY, B. (1991). A component of calcium-activated potassium channels encoded by the Drosophila slo locus. Science 253, 551-554 | |
| BARRETT, J. N., MAGLEBY, K. L. & PALLOTTA, B. S. (1982). Properties of single calcium-activated potassium channels in cultured rat muscle. Journal of Physiology 331, 211-230 | [Medline] |
| CRAWFORD, A. C. & FETTIPLACE, R. (1980). The frequency selectivity of auditory nerve fibres and hair cells in the cochlea of the turtle. Journal of Physiology 306, 79-125 | [Medline] |
| DING, J. P., LI, Z. W. & LINGLE, C. J. (1998). Inactivating BK channels in rat chromaffin cells may arise from heteromultimeric assembly of distinct inactivation-competent and noninactivating subunits. Biophysical Journal 74, 268-289 | [Abstract/Full Text] |
| FETTIPLACE, R. & FUCHS, P. A. (1999). Mechanisms of hair cell tuning. Annual Review of Physiology 61, 809-834 | [Abstract/Full Text] |
| FRENCH, R. J. & SHOUKIMAS, J. J. (1981). Blockage of squid axon potassium conductance by internal tetra-N-alkylammonium ions of various sizes. Biophysical Journal 34, 271-291 | [Abstract] |
| GALVEZ, A., GIMENEZ-GALLEGO, G., REUBEN, J. P., ROY-CONTANCIN, L., FEIGENBAUM, P., KACZOROWSKI, G. J. & GARCIA, M. L. (1990). Purification and characterization of a unique, potent, peptidyl probe for the high conductance calcium-activated potassium channel from venom of the scorpion Buthus tamulus. Journal of Biological Chemistry 265, 11083-11090 | [Abstract] |
| GOODMAN, M. B. & ART, J. J. (1996). Variations in the ensemble of potassium currents underlying resonance in turtle hair cells. Journal of Physiology 497, 395-412 | [Abstract] |
| HAMILL, O. P., MARTY, A., NEHER, E., SAKMANN, B. & SIGWORTH, F. J. (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflügers Archiv 391, 85-100 | [Medline] |
| HICKS, G. A. & MARRION, N. V. (1998). Ca2+-dependent inactivation of large conductance Ca2+-activated K+ (BK) channels in rat hippocampal neurones produced by pore block from an associated particle. Journal of Physiology 508, 721-734 | [Abstract/Full Text] |
| HILLE, B. (1992). Ionic Channels of Excitable Membranes, 2nd edn. Sinauer Associates, Sunderland, MA, USA | |
| HODGKIN, A. L. & HUXLEY, A. F. (1952). A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of Physiology 117, 500-544 | |
| HOLT, J. R. & EATOCK, R. A. (1995). Inwardly rectifying currents of saccular hair cells from the leopard frog. Journal of Neurophysiology 73, 1484-1502 | [Medline] |
| HORN, R. & MARTY, A. (1988). Muscarinic activation of ionic currents measured by a new whole-cell recording method. Journal of General Physiology 92, 145-159 | [Abstract] |
1 mV during the prepulse step and were therefore not considered in the making of this figure.
), during (
) and after (
) these treatments. The stimulus protocol was the same as in Fig. 2. The largest amplitude voltage reached (RS corrected) for each family of traces is indicated on the right. Recordings were made using the whole-cell configuration and 10 mM 4-AP was present throughout.
). Intermediate points are marked with open circles (
). Leak subtraction was not used in this protocol. Recordings were made using the whole-cell configuration and in the presence of 10 mM 4-AP.
) are estimates of i obtained from a single cell, tested at two potentials. The g for this cell is 79 pS. The data in this plot are from 12 hair cells.