| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1The Bobby R. Alford Department of Otorhinolaryngology and Communicative Sciences, Baylor College of Medicine, Houston, Texas 77030; and 2Department of Neurobiology, Pharmacology, and Physiology, University of Chicago, Chicago, Illinois 60637
Submitted 3 April 2002; accepted in final form 13 March 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
subunit being CaV1.3 (1D). Rat vestibular
epithelia and ganglia were probed for L-type -subunit expression with
the reverse transcription-polymerase chain reaction. The epithelia expressed
CaV1.3 and the ganglia expressed CaV1.2
(1C). |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
;
Sewell 1996). In auditory hair
cells, most of the voltage-gated Ca2+ current appears to
be carried by L-type channels (Platzer et
al. 2000; Schnee and Ricci
2002; Zidanic and Fuchs
1995). Such channels are selectively affected by dihydropyridines
with sensitivity that varies across L-channel subtypes. Dihydropyridine (DHP)
antagonists partly block both Ca2+ currents and
exocytosis in mouse cochlear hair cells
(Moser and Beutner 2000) and
partly block the Ca2+ currents but completely block
exocytosis in chick cochlear hair cells
(Spassova et al. 2001). The
spontaneous and sound-evoked discharges of guinea pig cochlear afferents are
depressed by DHP antagonists but not by a blocker of N-type
Ca2+ channels
(Robertson and Paki 2002). The
pore-forming subunit that predominates in the chick cochlea is the
CaV1.3 (1D) subunit
(Kollmar et al. 1997), which,
among L-type channels, is relatively insensitive to DHP antagonists
(Koschak et al. 2002;
Xu and Lipscombe 2001). Mice
that are null for CaV1.3 lack most of the cochlear inner hair cell
Ca2+ current, have no discernable auditory brain stem
response, and are deaf (Platzer et al.
2000). Remarkably, there are no obvious vestibular deficits in the
CaV1.3-null mice (Platzer et
al. 2000), raising the possibility that other
Ca2+ channel types operate in mammalian vestibular hair
cells. Here we have investigated whether hair cells from the sensory
epithelium (crista) of the rat anterior semicircular canal have an L-type
component and whether that component is CaV1.3.
In auditory hair cells of all vertebrates and in all hair cells of fish and
amphibians, the primary afferent neuron forms bouton endings on hair cells,
opposite presynaptic ribbons. In the vestibular organs of mammals, birds, and
reptiles, however, there are two kinds of afferent terminals: bouton endings
on type II hair cells and unusual cup-shaped endings, called calyces, on type
I hair cells (Wersäll and
Bagger-Sjöbäck 1974). The calyx ending differs from the
much-studied calyces of Held and of the superior cervical ganglia (reviewed in
Catterall 1999;
von Gersdorff and Borst 2002)
in being postsynaptic rather than presynaptic. It envelopes much of the
basolateral membrane of the hair cell. Although this unusual morphology has
long excited speculation about the nature of transmission at this synapse, the
presence of numerous synaptic dense bodies and vesicles in type I hair cells
of the mature rodent crista (Lysakowski
and Goldberg 1997) implies that quantal transmission takes place,
possibly in combination with other forms of transmission.
According to the prevailing view of quantal transmission at the hair cell
synapse, some hair cell Ca2+ channels must be open near
the resting potential, VR, to account for background
firing levels in afferent fibers. For most hair cells, VR
is –60 mV or more positive and therefore reasonably close to the
activation range of Ca2+ channels in hair cells of the
frog saccule and chick and mouse cochleas (positive to –60 or –50
mV) (reviewed in Zidanic and Fuchs
1995). Type I hair cells, however, have unusually negative resting
potentials (–70 to –85 mV) and low input resistances (10–100
M
) because they express a large, negatively activating K+
conductance (gK,L or gKI)
(Brichta et al. 2002;
Chen and Eatock 2000;
Correia and Lang 1990;
Rennie and Correia 1994;
Rennie et al. 1996;
Rüsch and Eatock 1996b).
This is true for diverse preparations (semicircular canals and utricles from
reptiles, birds, and mammals) and recording conditions (isolated hair cells
vs. excised, intact epithelia; ruptured-patch vs. perforated-patch modes of
whole cell recording). If these properties hold in vivo, then there should be
little quantal transmitter release because even large transduction currents
would not depolarize the hair cell into the activation range of the
Ca2+ channels. Yet "calyx afferents," which
innervate type I cells, have both background and evoked discharges
(Baird et al. 1988;
Goldberg et al. 1990;
Schessel et al. 1991).
Two kinds of solution have been proposed. Transmission at the type I/calyx
synapse may be partly mediated by other mechanisms, such as ephaptic
transmission (Gulley and
Bagger-Sjöbäck 1979;
Hamilton 1968;
Trussell 2000; Yamashita and
Ohmori 1990,
1991) or depolarization by
extracellular K+ accumulation
(Chen 1995;
Goldberg 1996). A second kind
of proposal is that in vivo, modulation of gK,L reduces
its activation at VR
(Behrend et al. 1997;
Chen and Eatock 2000), thereby
depolarizing the cells and increasing their input resistances. For this study,
we considered a third possibility: that the activation range of
Ca2+ channels in type I hair cells is unusually negative
and therefore overlaps the membrane potential for background levels of
stimulation. We report that in type I cells of the rat crista, the
Ca2+ channel activation range is only slightly negative
to that in type II cells, not enough to compensate for the type I cell's much
more negative resting potential and low input resistance.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Hair cells were dissociated from the sensory epithelia of anterior cristas
from young Long-Evans rats (postnatal days, P, 4–20) as previously
described (Chen and Eatock
2000). All procedures for handling animals were approved by the
animal care review committee at Baylor College of Medicine. All dissections
were done in L-15 (GIBCO BRL; additionally buffered with 10 mM HEPES, pH 7.3,
330 mmol/kg). The ampulla was excised and placed in L-15 medium, supplemented
with 1.2 mM EGTA, to lower Ca2+ to 100 µM, and the
following enzymes: protease XXVII (Sigma, St. Louis MO; 500 µg/ml) for 10
min at room temperature and papain (crude papain, Sigma, 500 µg/ml) and
L-cysteine (300 µg/ml) for 45 min at 37°C. All subsequent
procedures, including recording, were at room temperature (22–25°C).
The ampulla was transferred from the papain solution to L-15 containing bovine
serum albumen (500 µg/ml; 10 min) and then to the experimental chamber.
Cells were brushed off the sensory epithelium with an eyelash and allowed to
settle on the glass floor of the chamber. The cells were viewed on an inverted
microscope at x400 or x600 with differential interference contrast
optics (Olympus IMT-2, Olympus, Lake Success NY). The chamber was superfused
at a low rate with "standard external solution"
(Table 1).
|
Solutions
Whole cell recordings were made with the perforated-patch method
(Horn and Marty 1988). The
internal solution contained (in mM) 75 Cs2SO4, 35 CsCl,
5 MgSO4, 0.1 CaCl2, 5 HEPES, and 5 EGTA and 480 µg/ml
amphotericin B; pH 7.4 and osmolality 285 mmol/kg. The external solutions used
during recordings are given in Table
1. The channel blockers tetraethylammonium chloride (TEA-Cl),
4-aminopyridine (4-AP), and tetrodotoxin (TTX) were all obtained from Sigma.
Seal formation was always made in the standard external solution. We
ascertained that recordings were in perforated-patch mode rather than
ruptured-patch mode by the gradual reduction in access resistance and by the
persistence of the Ca2+ channel currents, which washed
out rapidly in ruptured-patch mode because of the lack of ATP in the internal
solution (Forscher and Oxford
1985). Final series resistance, Rs, was
between 5 and 25 M and was electronically compensated with the
amplifier circuitry by 50–80%.
Drug- and blocker-containing solutions were applied by local perfusion. Separate lines containing the standard external solution and the test solution were fed through a peristaltic pump into needles, which were lowered into the bath after whole cell currents were recorded in the bath solution (standard external solution). The cell was then moved into the flow of solution from each needle.
Junction potentials were calculated with JPCalc software
(Barry 1994). The calculated
liquid junction potential in standard external solution
(Table 1) was 9 mV. Local
perfusion with the other solutions, in which TEA-Cl replaced NaCl as the
dominant salt, added 4 mV to the total junction potential correction. No
corrections were made for Donnan potentials that might arise across the
perforated patch if there were differences between the impermeant ion
concentrations in the hair cell's endogenous solution and the pipette
solution. Such potentials are likely to be small based on other experiments on
delayed rectifier currents in type II hair cells. The pipette solution was
similar to that used here except that K+ replaced Cs+.
Half-maximal activation voltages were –29 ± 2 (SE) mV (n
= 8 cells) in the ruptured-patch mode and –30 ± 1 mV (20 cells)
in the perforated-patch mode, suggesting that Donnan potentials across the
perforated patch were negligible (K. M. Hurley and R. A. Eatock, unpublished
observations).
Recording
Recordings were made with a patch-clamp amplifier (L/M EPC-7, Adams and
List Associates, Great Neck, NY) and a 12-bit acquisition board (Digidata
1200, Axon Instruments, Foster City, CA), controlled by pClamp 6.1 software
(Axon Instruments). The clamp rise time was on the order of ≤50 µs
[(Rs ≤ 10 M) x (Cm
< 5 pF)]. The amplifier output was low-pass filtered at 5 kHz with an
8-pole Bessel filter (Model 901, Frequency Devices, Haverhill, MA). In most
protocols, test steps were 10 ms, and the filtered
Ca2+-channel data were sampled at 10-µs intervals.
For longer test steps to study Ca2+-channel
inactivation, sampling intervals were 1.5 ms, and the data during this portion
of the sweep were filtered off-line at 200 Hz with digital filtering
algorithms in Clampfit (v. 8, Axon Instruments).
LEAK SUBTRACTION. We usually used a voltage protocol, called
"P, –P/4", which provides on-line leak subtraction
(Armstrong and Bezanilla 1974).
Before each trial, the entire test waveform (P) was divided by –4 and
the resulting –P/4 waveform was presented four times from a potential of
–103 mV (to be out of the range of activation of
gK,L). The method assumes that the –P/4 waveform
evokes only linear leak current and that the total current summed over the
four –P/4 steps is equal and opposite to the linear leak current evoked
by P. The sum of the currents evoked by the –P/4 waveform was summed
with the current evoked by P, leaving only nonlinear current.
For our usual protocol (10-ms test steps with an inter-trial interval of 100 ms), each trial comprised, in order, four small 10-ms steps (–P/4; from –103 mV), the test step (P; from –73 mV), and a final 40-ms interval at –73 mV.
Data analysis
Voltages were corrected off-line for liquid junction potentials but not for
uncompensated series resistance. Maximum voltage errors were ≤5 mV
(corresponding to total currents ≤500 pA and residual Rs ≤ 10
M) and most often on the order of 1 mV.
To study the voltage dependence of activation, we delivered a series of
10-ms voltage steps from the holding potential and plotted the current values
at 9 ms after the step, well after activation was complete and before
significant inactivation occurred. The I-V curves are nonmonotonic;
at large depolarizations, when the channels are fully activated, current
decreases as driving force decreases. The increasing part of the I-V
relation, from –83 mV to the voltage corresponding to the peak current,
was fitted with a first-order Boltzmann function
 | (1) |
The activation time course of Ca2+ channel currents
was fitted with a Hodgkin-Huxley equation
 | (2) |
is a time constant of
activation, and is the power, usually 3.
The decay of the current (inactivation) during a 500-ms pulse was fitted by
a single exponential function
 | (3) |
is the time constant of
decay. Results are presented as means ± SE.
RT-PCR
We did reverse transcription-polymerase chain reaction (RT-PCR) to look for
expression of the pore-forming subunits of known L-type
Ca2+ channels. To prepare vestibular epithelia for
RT-PCR, we dissected ampullas, utricles, and vestibular ganglia using methods
similar to those described in the preceding text. We then exposed the ampullas
and utricles to protease XXIV (Sigma; 100 µg/ml in standard external
solution) for 10 min at room temperature to loosen connections to overlying
accessory structures, which were then removed. To loosen the epithelia from
the underlying stroma, we then treated the ampullas and utricles with protease
X (thermolysin; 500 µg/ml in standard external solution) for 1 h at
37°C (Saffer et al. 1996).
The epithelia were then peeled off and frozen on dry ice. The frozen tissue
was disrupted with a pestle, lysed, and homogenized on a QiaShredder column
(Qiagen, Valencia CA). The resulting lysate was spun, and the supernatant
applied to a silica-gel membrane column (RNAeasy; Qiagen). To eliminate
contamination by genomic DNA, we did oncolumn digestion by RNase-free DNase I
(Qiagen). The DNase was removed by washing and the RNA was eluted.
Reverse transcription of the eluted RNA to cDNA and subsequent PCR were
done with a PTC-100 thermocycler (M-J Research, Incline Village, NV). The RNA
was reverse transcribed with random hexamer primers and MMLV reverse
transcriptase (Clontech Laboratories, Palo Alto CA). cDNA was amplified with
TAQ polymerase (AmpliTaq, Applied Biosystems, Foster City CA) and the
following primer pairs directed against the alpha subunits of L-type channels
(5'-3'): 1) Cav1.2 (1C):
forward primer: AAGATGACTCCAACGCCACC; reverse primer: GATGATGACGAAGAGCACGAGG;
2) Cav1.3 (1D): forward:
TGAGACACAGACGAAGCGAAGC; reverse: GTTGTCACTGTTGGCTATCTGG; 3)
Cav1.4 (1F): forward: GGAGGAGGTCACTGTGGGAA;
reverse: GTTGGGATCCAGCCTGTAGC.
As a positive control, we tested for calmodulin expression with the following primers: forward: CTGAAGAGCAGATYGCAGAATTCA; reverse: TCACTTTGCTGTCATCATTTGTAC.
The primer pairs were directed against a region in the first repeat domain
for CaV1.2 (Snutch et al.
1991), against part of the second repeat domain for
CaV1.3 (Hui et al.
1991), and against the C-terminus for CaV1.4
(Morgans et al. 2001). We used
the following PCR program: hot start at 94°C for 4 min; cycle 40 times
between 94°C for 30 s, 59°C for 30 s, 72°C for 35 s; final
extension at 72°C for 7 min. Primer pairs for CaV1.2 and
CaV1.3 were designed by García-Palomero et al.
(2000).
For each tissue and primer pair, we ran two kinds of negative control. In the negative control for contamination, water replaced the template made from the inner ear organs. In the negative control for genomic DNA ("–RT" control), reverse transcriptase was omitted to prevent transcription of cDNA from mRNA. For positive controls for the primers, we used the same primer pairs on cDNA made from rat brain poly-A RNA (Clontech) for CaV1.3 and 1.4 and from rat heart poly-A RNA (Clontech) for CaV1.2. For positive controls for the quality of the inner ear tissues, we used the calmodulin primers on each batch of cDNA and also ran all three CaV1.x primer sets on the same batch of cDNA. PCR products were resolved on a 1.2% agarose gel and visualized with ethidium bromide. PCR product identity was confirmed by restriction digest and direct sequencing (SeqWright, Houston TX).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Hair cells were classified according to whether they had
gK,L, an outwardly rectifying K+ conductance
that activates at unusually negative potentials. This conductance is strongly
correlated with the distinctive amphora shape of type I hair cells in various
vestibular organs (Eatock et al.
1998; Ricci et al.
1996). In a previous study of hair cells enzymatically dissociated
from rat cristas (Chen and Eatock
2000), 86% (31/36) of cells that looked like type I cells had
gK,L, compared with 11% (4 /29) of cells that looked like
type II cells and 48% (11/23) of cells that could not be classified
morphologically. The data in the present study are from 15 cells with
gK,L and 19 cells without gK,L. We
have pooled data according to whether the cells had gK,L,
rather than by hair cell shape, because 1) gK,L
expression is less sensitive to dissociation procedures than is hair cell
shape and 2) gK,L is the relevant attribute, as
we are interested in the activation ranges of Ca2+
channel currents in cells with gK,L (see
INTRODUCTION).
We began each recording with our standard internal solution, which
contained Cs+ instead of K+, and standard external
solution, which lacked channel blockers
(Table 1). To determine whether
gK,L was present, we used the voltage protocol shown in
Fig. 1, A and
B. gK,L was recognized by its
substantial permeability to Cs+, slow activation kinetics, and, in
many cells, its relatively negative activation range. Cs+ is poorly
permeant in many K+ channels, but for gK,L, the
Cs+ permeability was calculated from reversal potential
measurements to be about one-third that of K+
(Rüsch and Eatock 1996a).
gK,L is also unusual in that in many cases (10/15 in this
data set) it is significantly activated at the holding potential of –69
mV. In such cells, there was a significant inward current at –69 mV,
reflecting K+ influx through gK,L. As voltage
was stepped to –129 mV, the current transiently increased because of the
increased driving force, then decayed to zero as gK,L
deactivated (Fig. 1A).
Subsequent voltage steps to between –89 and –9 mV re-activated
gK,L. At the most negative voltages, the inward
K+ currents activated slowly (see trace at –79 mV),
consistent with the activation kinetics of gK,L in this
voltage range (Rüsch and Eatock
1996a). Positive to –49 mV, the current became outward and
was presumably carried by Cs+.
|
In 5 of 15 cells with substantial Cs+ outward currents, there
was relatively little current activated at –69 mV. Four of the five
cells were from relatively young animals (3 from P5 animals, 1 from P9, and
the 5th from P18). Data from rodent utricles show that the very negative
activation range is acquired during or after the first postnatal week: after
P2 in semi-intact mouse utricles
(Rüsch et al. 1998) and
after P7 in isolated rat utricular hair cells (K. M. Hurley and R. A. Eatock,
unpublished observations). We therefore think that this is more likely to be
an immature, more positively activating version of gK,L
than a completely different Cs+-permeant conductance.
Ca2+ channel currents from these 5 cells were inspected
for differences from the 10 cells with negatively activating
gK,L; none were found and these cells are therefore
included in the category "cells with
gK,L."
In cells without gK,L (Fig. 1B), little current flowed during the depolarizing voltage steps, showing that the influx of K+ was largely blocked by the intracellular Cs+ ions. The small currents that did flow had fast kinetics, unlike gK,L. The step from –69 to –129 mV evoked little inward current as few channels were open at –69 mV, and no change in the steady current level, indicating that no major conductance was deactivated by the step.
To study current through Ca2+ channels, we applied
external solutions containing 1.3 or 5 mM Ca2+ or 5 mM
Ba2+, 0 K+, and the following K+
channel blockers: 130 mM TEA-Cl and 5 mM 4-AP to block outward rectifiers and
5 mM CsCl to block inward rectifiers (Table
1). TTX was also present at 1 µM to block a voltage-sensitive
Na+ current that was present in every cell studied. In these
solutions, depolarizing steps from –73 mV evoked linear leak currents
(subtracted out by the "P/–P/4" routine) and small,
voltage-dependent, fast-activating, sustained inward currents in cells with
gK,L (Fig.
1C) and in cells without gK,L
(Fig. 1D). Several
observations showed that the inward current flowed through
Ca2+ channels. First, it could be carried by
Ba2+ (e.g., Figs.
2C and
3). Second, removal of
extracellular Ca2+ eliminated the current
(Fig. 2, A and
B). Note that in our 0-Ca2+
external solution, Ca2+ was present at trace levels
(
10 µM) because we did not add a Ca2+ chelating
agent; at lower Ca2+ levels, Ca2+
channels become nonselective (Almers and
McCleskey 1984; Art and
Fettiplace 1987). Third, 25 µM Cd2+
blocked peak current by 82 ± 5% (n = 4 cells; 5
Ba2+ solution; Fig.
2, C and D), consistent with reported
IC50 values for Cd2+ block of L-type and
N-type channels (range: 10–300 µM across preparations) (reviewed in
Brammar 1999).
|
|
|
It has been suggested that guinea pig vestibular hair cells have T-type
Ca2+ currents, based on the presence of a
Cd2+-sensitive transient inward current
(Rennie and Ashmore 1991)
andaNi2+-sensitive depolarization-evoked
Ca2+ influx (Boyer
et al. 1998). Ni2+ sensitivity has been
considered a hallmark of T-type channels
(Fox et al. 1987), although
more recent studies suggest that this may not be generally true
(Zamponi et al. 1996). We
found no evidence for T-type Ca2+ currents in rat crista
cells. When TTX was not present and a hyperpolarizing prepulse was used,
depolarizing steps evoked inward currents with a rapidly inactivating
component (data not shown). This component was entirely through voltage-gated
Na+ channels, however, as it was eliminated in TTX or when external
Na+ was replaced by choline+.
Ca2+channel current has a dihydropyridine-sensitive component
PHARMACOLOGY. To test for an L-type component in the Ca2+ channel current, we added the DHP agonist, Bay K 8644, or the DHP antagonist, nimodipine, to the external solution. Bay K 8644 favors long channel openings (Hess et al. 1984; Nowycky et al. 1985) and therefore enhances and slows the current (Fig. 3, A and B). In four cells (2 of each type), the maximum current (Imax; Eq. 1) in 40 µM Bay K 8644 was 214 ± 17% of that in control. Nimodipine reduced the current but only weakly even at high doses (Fig. 3, C and D). Imax was blocked 34.1 ± 8.6% by 40 µM nimodipine [range: 20–59%, n = 4 (2 cells of each type), P5–P12]. This relatively small block raises the possibility that the Ca2+ channel current includes a non-L-type component, but for reasons discussed later (Subunit composition) is not strong evidence for it.
RT-PCR. The subunit of Ca2+
channels forms the pore and in L-type channels includes the DHP binding site.
L-type subunits that are expressed in the brain are CaV1.2
(1C) and CaV1.3 (1D);
CaV1.4 (1F) is expressed in retina. In chick and
mouse cochlear hair cells, the CaV1.3 (1D)
subunit predominates (Kollmar et al.
1997; Platzer et al.
2000). We reverse transcribed mRNA from vestibular ganglia and
from the epithelia of cristas and utricles of P13 rats. The epithelia were
peeled from the underlying stroma after treatment with thermolysin (see
METHODS) and presumably include hair cells, support cells, and
nerve terminals. The cDNA was probed with primers for CaV1.2 and
CaV1.3 (Fig. 4), as
well as CaV1.4 (data not shown) and, as a positive control,
calmodulin. All tissues were positive for calmodulin.
|
In the vestibular epithelia, products of the appropriate size and sequence
were obtained for CaV1.3 but not for CaV1.2 or
CaV1.4; the primer sets were tested on the same batch of cDNA. For
both the utricle and the cristas, the sequenced product (285 nucleotides,
excluding the first 20) was 100% identical to the published sequence for
CaV1.3 from rat brain (Hui et
al. 1991). It is possible that the PCR product includes a
contribution from the supporting cells
(Mori et al. 1998) or nerve
terminals. Nevertheless, given the evidence for an L-type current of low DHP
sensitivity in the hair cells and the lack of PCR products for
CaV1.2 and CaV1.4 in the epithelium, we conclude that
the hair cell Ca2+ current is at least partly carried by
CaV1.3 subunits.
In the vestibular ganglia, a PCR product was only obtained with
CaV1.2 (1C) primers. This argues that
CaV1.3 expression found in the epithelia does not originate in the
nerve terminals. The product obtained with CaV1.2 primers was
identical to the published sequence for CaV1.2 from rat brain
(Snutch et al. 1991). PCR
products corresponding to CaV1.2 were previously obtained after
reverse transcription of total RNA from the mouse cochlea
(Green et al. 1996).
DHP-sensitive voltagegated Ca2+ currents have been
recorded in dissociated mouse vestibular neurons
(Chambard et al. 1999).
A PCR product of the correct size (442 bp) and sequence was obtained with
primers for the CaV1.4 (1F) subunit in brain but
not in vestibular organs or ganglia (data not shown). This subunit is
predominantly expressed in retina
(Bech-Hansen et al. 1998;
Naylor et al. 2000).
In summary, our pharmacological and RT-PCR data indicate that mammalian vestibular hair cells express an L-type Ca2+ current carried by CaV1.3 subunits. It is possible that there are additional contributions from other Ca2+ channel types (see Subunit composition).
Current-voltage relations
Figure 5A compares the mean isochronal I-V relations for cells with and without gK,L in 5 mM external Ca2+. The current of the mean relations peaked at approximately –60 pA for cells with gK,L (range: –45 to –85 pA in individual cells) and approximately –80 pA for cells without gK,L (range: –45 to –125 pA). Boltzmann fits to the mean curves yielded V1/2 values that were on average slightly but significantly more negative in cells with gK,L (–41.1 ± 0.5 mV) than in cells without gK,L (–37.2 ± 0.2 mV) and S values that were similar (6.8 ± 0.4 vs. 6.6 ± 0.2 mV). The differences in peak currents and activation ranges have complementary effects between –70 and –40 mV, with the result that the two cell types have similar mean currents over much of the physiological range of voltages.
Figure 5B shows the
I-V relation for a cell with gK,L in 1.3 mM
Ca2+. The mean current in 1.3 mM
Ca2+ for eight cells with gK,L (not
shown) was 1% of its peak value at –72 mV, half-maximal at –46 mV,
and 90% of its peak value at –34 mV. Similar results were obtained in
mouse cochlear inner hair cells in 1.3 mM Ca2+
(Platzer et al. 2000).
Time course
Activation kinetics were fitted with a Hodgkin-Huxley scheme (Eq.
2). A similar scheme has been used to fit Ca2+
channel activation in other hair cells, with the exponent, , set to 2
(Art and Fettiplace 1987;
Perin et al. 2001;
Zidanic and Fuchs 1995) and 3
(Lewis and Hudspeth 1983). For
our cells, assumed values between 2 and 4 when it was allowed to vary.
To compare activation time constants () across cells, we set at
3. Examples of fits are shown in Fig.
6A. Figure
6B shows the voltage dependence of mean values
averaged from four type I and five type II cells, all in 5
Ca2+ medium. Values ranged from 400 µs at –13
mV to 730 µs at –48 mV. On average, the time constants were 100
µs slower for the type I cells. Even the values from type II cells
(100–500 µs across the voltage range) are somewhat slower than those
reported from other hair cells (Art and
Fettiplace 1987; Hudspeth and
Lewis 1988a; Perin et al.
2001; Zidanic and Fuchs
1995).
|
INACTIVATION. In 5 mM external Ca2+, no
inactivation was seen during small depolarizations lasting hundreds of
milliseconds (Fig. 6C,
top). For large depolarizations, however, inward currents decayed
with time constants of several hundred milliseconds. The example in
Fig. 6C decayed during
a step to –3 mV with a time constant of 350 ms to a steady-state
value 50% of its peak value. Similar results were obtained in five other
cells without gK,L. Inward currents during long
depolarizing steps also decayed in cells with gK,L. In the
latter, however, slow kinetic components in the tail currents at the offset of
long steps suggest that the Ca2+ current recorded during
long steps was contaminated by currents flowing through imperfectly blocked,
slowly activating channels. Such currents did not contaminate the responses to
our usual 10-ms protocols.
Currents through heterologously expressed human CaV1.3 channels
show inactivation with Ca2+-dependent and -independent
components (Bell et al. 2001).
In the rat crista cells, Ba2+ currents did not decay
(Fig. 6C,
bottom), consistent with Ca2+-dependent
inactivation being responsible for the decay in 5
Ca2+.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
In vitro measurements of resting potential from type I hair cells are
generally negative to –70 mV because of the large K+
conductance at negative potentials; in the rat crista, ruptured-patch
recordings with standard solutions yielded a mean resting potential of
–81.3 ± 0.2 mV (n = 144)
(Chen and Eatock 2000). At
–70 mV and in physiological Ca2+,
Ca2+ current is 1% maximal
(Fig. 5B). One would
predict on this basis that there would be little background release of quanta.
Furthermore, the observation that Ca2+ current is not
appreciable negative to –60 mV (Fig.
5B) together with other information from in vitro studies
argues that there would be little stimulated quantal release, as follows. For
a cell with a large gK,L, the input resistance is
frequently as low as 20 M. The largest transduction currents that have
been recorded from rodent vestibular hair cells are 400 pA (M. A.
Vollrath and R. A. Eatock, unpublished results). Such a current would
depolarize such a cell by just 8 mV, barely into the voltage range at which
significant Ca2+ current flows. In vivo, however, rodent
calyx afferents do have both background and evoked discharges
(Baird et al. 1988;
Goldberg et al. 1990). The
background rates and gains (evoked rates per unit stimulus) can be lower than
for other afferents but are nevertheless substantial. Although other
transmission mechanisms may operate at this synapse, the presence of the
presynaptic machinery makes it likely that chemical transmission operates at
least some of the time. In that case, type I cells must have, at least some of
the time, more positive resting potentials, larger input resistances, or both.
The responsible mechanisms may include down-modulation of
gK,L, which would have both effects
(Behrend et al. 1997;
Chen and Eatock 2000), and
accumulation of K+ in the synaptic cleft, which would depolarize
the hair cells (Chen 1995;
Goldberg 1996).
Comparison with Ca2+channel currents in other hair cells
BIOPHYSICAL PROPERTIES. We were interested in the possibility that the Ca2+ channels of type I cells have distinctive properties that explain how they control transmitter release in these low-resistance, negatively resting hair cells. But the differences that we saw were modest: type I Ca2+ channel currents were slightly slower and smaller and activated at slightly more negative potentials relative to type II Ca2+ channel currents. More striking are the similarities between the currents of both hair cell types and those described for hair cells from other inner ear organs. All are fast to activate and deactivate, show little inactivation for small depolarizations, and activate at fairly similar voltages.
The aspect in which we were most interested, the voltage dependence of type
I Ca2+ currents, is a few millivolts more negative than
that of type II Ca2+ currents
(Fig. 5A). The source
of this modest difference and the small difference in activation kinetics is
not clear. If there are multiple Ca2+ channel
types—reflecting differences in or 
subunits or
posttranslational modifications—the differences in biophysical
properties between the two cell types might reflect a difference in the
proportions of channel types. The slightly more negative activation range in
cells with gK,L is in the right direction to function with
the more negative resting potentials encountered in these cells. On the other
hand, because the average peak current was smaller in cells with
gK,L, the average current levels in the voltage range of
the greatest physiological significance, between –70 and –40 mV,
were indistinguishable (Fig.
5A).
The voltage ranges of activation for both cell types span the range
reported in other hair cells in comparable concentrations of charge carrier.
In 2.8 mM Ca2+, Ca2+ currents in
isolated turtle cochlear cells were half-maximal at approximately –40 mV
(Art et al. 1993). In 4–5
mM external Ca2+, Ca2+ currents in
hair cells from the frog amphibian papilla (an auditory organ) and frog crista
were half-maximal at –40 to –45 mV
(Martini et al. 2000;
Perin et al. 2001;
Smotherman and Narins 1999).
In 1.8 mM Ca2+, Ca2+ current is
half-maximal at –38 mV in enzymatically dissociated cells from the frog
saccule (Armstrong and Roberts
1998). In the latter study, the papain dissociation shifted the
activation range by +7 mV. Our dissociation procedure includes a lengthy
exposure to papain and a –5- to –10-mV shift would go a long way
to aligning the Ca2+ channel activation range with type
I membrane potentials at background levels of stimulation: the current might
be 1% of maximal at –80 mV (rather than –72 mV,
Fig. 5B) and 10% of
maximal at –65 mV. Any papain effect would act on both cell types,
however, so that there still would not be a major difference in their
Ca2+ channel activation ranges.
The activation kinetics of the Ca2+ currents in rat
crista hair cells were slightly slower than reported in other hair cells.
Assuming a thermal Q10 of 2–3, correction for mammalian body
temperature brings the time constants into line with time constants measured
at room temperature in poikilotherms: turtle cochlea
(Art and Fettiplace 1987), frog
saccule (Hudspeth and Lewis
1988a), and frog crista (Perin
et al. 2001), but they are still slower than those in chick
cochlea (Zidanic and Fuchs
1995) after correction for chick body temperature. The kinetics of
rat crista channels may have been slowed by the papain dissociation
(Armstrong and Roberts 1998).
Alternatively, cochlear Ca2+ channels may really have
faster kinetics to reduce low-pass filtering of the afferent signal at
acoustic frequencies. A single exponential fit of activation in the rat crista
cells at –28 mV yields time constants at room temperature of 1 ms,
for a half-power low-pass frequency of 160 Hz—in the acoustic frequency
range and well above the frequencies of head movements. We therefore know of
no functional significance for the small difference between the activation
time constants of type I and type II hair cells. Inactivation was modest in
overall extent and quite slow; even for large depolarizations, the time
constant at room temperature was several hundred milliseconds, corresponding
to a high-pass corner frequency of 0.4 Hz, and there was a substantial
steady-state component.
Maximal Ca2+ current amplitudes and estimated
conductances in hair cells vary over an order of magnitude, from 50 pA to
1 nA. In the frog saccule, voltage-gated Ca2+ channels
are localized at presynaptic active zones
(Issa and Hudspeth 1994;
Roberts et al. 1990;
Rodriguez-Contreras and Yamoah
2001). In chick cochlear hair cells, Martinez-Dunst et al.
(1997) found that
Ca2+ channel number varies with "presynaptic
release area" [(the area of hair cell membrane adjacent to each
presynaptic dense body) x (the number of dense bodies)] and showed that
data from the turtle cochlea and frog saccule lie along the same continuum.
This relationship accounts for the increase in Ca2+
current amplitude with best frequency in the chick cochlea and possibly in the
auditory organs of other nonmammals (Art et
al. 1993; Smotherman and
Narins 1999). In a vestibular organ, the frog crista,
Ca2+ channel current amplitude and presynaptic release
area may also co-vary. Perin et al.
(2001) found that
Ba2+ current amplitude varies considerably with region
of origin, and noted that the larger Ba2+ currents occur
in cells with larger synaptic bodies
(Lysakowski 1996).
The presynaptic release areas, numbers of such areas per hair cell, and
current amplitudes of rodent crista hair cells appear to be comparable to
those of chick tall hair cells. Inspection of Figs. 15 and 16 in Lysakowski
and Goldberg (1997) suggests
that presynaptic release areas are similar to those in chick cells
(Martinez-Dunst et al. 1997).
There are 17–18 dense bodies in mature rodent crista hair cells
(Lysakowski and Goldberg 1997)
and 15 in chick tall hair cells
(Martinez-Dunst et al. 1997).
Peak Ca2+ channel currents are also consistent with each
other given that they were recorded in different concentrations of
Ba2+. For rat crista type II hair cells, currents
doubled when we increased the Ca2+ or
Ba2+ concentration fourfold, from 1.3 to 5 mM (data not
shown). Thus the 80-pA currents that we got in 5 mM
Ba2+ are comparable to the 100- to 300-pA currents
recorded at fourfold higher Ba2+ concentration in chick
hair cells (Martinez-Dunst et al.
1997; Zidanic and Fuchs
1995). Presynaptic release areas can be larger in type II hair
cells than in type I hair cells
(Lysakowski and Goldberg
1997), which might contribute to the difference in peak current
size between type I and type II cells (Fig.
5).
Another possible factor in the relative sizes of Ca2+
currents in hair cells from different inner ear organs is whether the currents
are used to drive the activation of Ca2+-gated
K+ channels in addition to synaptic transmission. Those hair cells
with the largest Ca2+ currents also have sizeable
Ca2+-gated K+ currents, and both channel
types participate in electrical tuning (Art
and Fettiplace 1987; Hudspeth and Lewis
1988a,b;
Lewis and Hudspeth 1983). In
rodent vestibular hair cells, in contrast, K(Ca) conductances are not the
dominant outward rectifiers and electrical tuning is broad
(Rennie and Correia 1994;
Rennie et al. 1996; Rüsch
and Eatock
1996a,b).
SUBUNIT COMPOSITION. Our pharmacological and RT-PCR data suggest
that rat crista hair cells express L-type channels made up of
CaV1.3 (1D) -subunits as do other hair
cells (Green et al. 1996;
Kollmar et al. 1997;
Platzer et al. 2000).
Are there other components to the whole cell Ca2+
current? The evidence so far is equivocal. By itself, modest block by high
doses of nimodipine does not constitute strong evidence for non-L-type
components. DHP block is strongly enhanced by depolarization
(Bean 1984; Koschak et al.
2001); thus the efficacy of the blocker in our experiments was reduced by
holding at relatively negative potentials and using short depolarizations.
Heterologously expressed CaV1.3 subunits form channels that are
less sensitive to DHP block at any voltage than are those channels formed by
other L-type subunits (Koschak et al. 2001;
Xu and Lipscombe 2001). The
DHP-sensitive component of the mouse inner hair cell current is even less
sensitive to DHP block than are heterologous CaV1.3 channels
(Koschak et al. 2001). For a voltage protocol similar to ours, the
Ca2+ channel currents of mouse inner hair cells are
blocked 40% by 10 µM nimodipine (Platzer et al. 2001), similar to the
effect that we observed at 40 µM nimodipine. The DHP sensitivities of
heterologously expressed CaV1.2 and CaV1.3 channels and
hair cell Ca2+ channels follow the same progression as
the amounts by which the channels inactivate (CaV1.2 >
CaV1.3 > hair cell Ca2+ channels),
consistent with the proposal that DHP blockers preferentially bind inactivated
channels (Bean 1984).
Nevertheless, a number of observations make it likely that there is a
non-L-type component. First, there is evidence from CaV1.3 null
mutants. The inner hair cells in the CaV1.3-null mice have a small
residual current (<10% of wild-type)
(Platzer et al. 2000). This
small current may not sustain synaptic transmission; the animals are deaf.
They lack obvious vestibular deficits, however, suggesting that
non-CaV1.3 components may be more important in vestibular hair
cells (Platzer et al. 2000).
Second, an R-type current component has been proposed in frog crista cells,
based on its resistance to L- and N-type blockers
(Martini et al. 2000). Third,
there is evidence for N-type subunits in hair cells. An antibody to N-type
subunits (CaV2.2, 1B) labeled the basolateral
membranes of type I and type II hair cells in the chinchilla crista
(Lopez et al. 1999). In frog
saccular hair cells, single-channel recordings reveal both L-type channels and
a second type with some N-type properties
(Rodriguez-Contreras and Yamoah
2001; Su et al.
1995), and antibody to N-type channels labels discrete spots on
the basolateral membrane
(Rodriguez-Contreras and Yamoah
2001). Further experiments with different pharmacological agents
and/or single-channel recordings are required to determine whether there are
multiple voltage-gated Ca2+ channels in mammalian
vestibular hair cells.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This work was supported by National Institute on Deafness and Other Communication Disorders Grant DC-02058.
Present address of H. Bao: Section of Neurobiology, University of Texas, Austin, TX 78712.
|
FOOTNOTES |
|---|
Address for reprint requests: R. A. Eatock, Dept. of Otolaryngology, Rm. NA-511, Baylor College of Medicine, One Baylor Plaza, Houston TX 77030 (E-mail: eatock{at}bcm.tmc.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Armstrong CE and Roberts WM. Electrical properties of frog saccular hair cells:
distortion by enzymatic dissociation. J Neurosci
18: 2962–2973,
1998.
Armstrong CM and Bezanilla F. Charge movement associated with the opening and closing
of the activation gates of the Na channels. J Gen
Physiol 63:
533–552, 1974.
Art JJ and
Fettiplace R. Variation of membrane properties in hair cells isolated from
the turtle cochlea. J Physiol
385: 207–242,
1987.
Art JJ,
Fettiplace R, and Wu YC. The effects of low calcium on the
voltage-dependent conductances involved in tuning of turtle hair cells.
J Physiol 470:
109–126, 1993.
Baird RA,
Desmadryl G, Fernández C, and Goldberg JM. The vestibular nerve of
the chinchilla. II. Relation between afferent response properties and
peripheral innervation patterns in the semicircular canals. J
Neurophysiol 60:
182–203, 1988.
Barry PH. JPCalc, a software package for calculating liquid junction potential corrections in patch-clamp, intracellular, epithelial and bilayer measurements and for correcting junction potential measurements. J Neurosci Methods 51: 107–116, 1994.[ISI][Medline]
Bean BP.
Nitrendipine block of cardiac calcium channels: high-affinity binding to the
inactivated state. Proc Natl Acad Sci USA
81: 6388–6392,
1984.
Bean BP and Mintz IM. Pharmacology of different types of calcium channels in rat neurons. In: Handbook of Membrane Channels: Molecular and Cellular Physiology, edited by Peraccio C. San Diego: Academic, 1994, p. 199–210.
Bech-Hansen NT, Naylor MJ, Maybaum TA, Pearce WG, Koop B, Fishman GA, Mets M, Musarella MA, and Boycott KM. Loss-of-function mutations in a calcium-channel alpha1-subunit gene in Xp11.23 cause incomplete X-linked congenital stationary night blindness. Nat Genet 19: 264–267, 1998.[ISI][Medline]
Behrend O, Schwark C, Kunihiro T, and Strupp M. Cyclic GMP inhibits and shifts the activation curve of the delayed-rectifier (IK1) of type I mammalian vestibular hair cells. Neuroreport 8: 2687–2690, 1997.[ISI][Medline]
Bell DC,
Butcher AJ, Berrow NS, Page KM, Brust PF, Nesterova A, Stauderman KA, Seabrook
GR, Nurnberg B, and Dolphin AC. Biophysical properties, pharmacology, and
modulation of human, neuronal L-type (alpha(1D), Ca(V)1.3) voltage-dependent
calcium currents. J Neurophysiol
85: 816–827,
2001.
Boyer C, Lehouelleur J, and Sans A. Potassium depolarization of mammalian vestibular sensory cells increases [Ca2+]i through voltage-sensitive calcium channels. Eur J Neurosci 10: 971–975, 1998.[ISI][Medline]
Brammar WJ. Voltage-gated calcium channels. In: The Ion Channel Facts Book. IV. Voltage-Gated Channels, edited by Conley EC and Brammar WJ. London, UK: Academic, 1999, p. 22–153.
Brichta AM,
Aubert A, Eatock RA, and Goldberg JM. Regional analysis of whole cell
currents from hair cells of the turtle posterior crista. J
Neurophysiol 88:
3259–3278, 2002.
Catterall WA. Interactions of presynaptic Ca2+ channels and snare proteins in neurotransmitter release. Ann NY Acad Sci 868: 144–159, 1999.[ISI][Medline]
Chambard JM,
Chabbert C, Sans A, and Desmadryl G. Developmental changes in low and high
voltage-activated calcium currents in acutely isolated mouse vestibular
neurons. J Physiol 518:
141–149, 1999.
= –37.2 ± 0. 2 mV; S = 6.6
± 0.2 mV. B: isochronal I-V relation in physiological
Ca2+ (1.3 mM). Morphological type I cell with
gK,L, P11. Fitted between –83 and –23 mV with
Eq. 1 (—): Imax = –59.4 pA,
Imin = –0.4 pA, V
