Neuroscience Letters, 133 ( 1991 ) 17 l-174 (C" 1991 Elsevier Scientific Publishers Ireland Ltd. All rights reserved 0304-3940/91/$ 03.50

171

NSL 08215

Stretch sensitivity of the lateral wall of the auditory outer hair cell from the guinea pig K.H. Iwasa 1, M i n x u Li 1, M i n Jia ~ and B. K a c h a r z ~Biophysics Laboratory, N1NDS and 2Molecular Otology Laboratory, NIDCD, National Institutes of Health, Bethesda, MD 20892 (U.S.A.) (Received 24 May 1991; Revised version received 28 August 1991; Accepted 30 August 1991) Key words." Mechano-electric transduction; Stretch-activated channel; Outer hair cell; Potassium current; Guinea pig; Cytoskeletal structure The inner and outer hair cells of the mammalian hearing organ are mechano-transducer cells. Here we report evidence that the lateral wall of outer hair cells (OHCs) is a mechano-receptor. This mechano-sensitivity appears to complement that of the stereocilia. Patch clamping studies showed that stretching of the membrane patches by suction at the pipette activated potassium channels with 130 pS unit conductance specifically localized in the lateral wall. Application of an osmotic tension to the entire cell membrane under whole-cell recording produced a 10 mV hyperpolarization. The reversal potential and the magnitude of the macroscopic current under voltage clamp were consistent with the single-channel properties of stretch-activated potassium channels. The elongated cylindrical cell body of the OHC is optimally positioned in the cochlea to sense axial force due to the vibrations of the basilar membrane during sound stimulation. This sensitivity can explain the production of a predominantly hyperpolarizing response to sound stimuli, unique to the OHC. Coupled with voltage-dependent OHC motility, the stretch-activated channels may play an important role in producing a mechanical feedback, an indispensable element in cochlear tuning.

Non-selective channels associated with stereocilia have been considered to be the basis of mechano-sensitivity of hair cells [7]. These channels are usually more efficient in causing depolarizing than hyperpolarizing responses in these cells. There are, however, recent reports suggestive that the outer hair cell (OHC) has an additional mechano-receptor: Cody and Russell [2] reported that in vivo response of OHC membrane potential to sound stimuli is larger in the hyperpolarizing direction, a non-linearity difficult to explain by non-selective channels alone. It is also reported that the length of this cell changes when mechanical stimuli at audio frequencies are applied to the cell body of isolated OHCs [1]. The response is greatest when the middle of the lateral wall was stimulated. This unique property of the OHC could be associated with its physiological significance in the fine tuning of the ear [12]. To address the question as to whether the OHC has an additional mechano-sensitivity, we examined the stretch sensitivity of the lateral wall with the patch clamping method. To clarify the significance of these channels, we also examined the macroscopic current due to a membrane stress.

Correspondence: K.H. lwasa, National Institutes of Health Bldg. 9, Rm. IE124, Bethesda, MD 20892, U.S.A.

Single-channel current was recorded at voltageclamped on-cell patches located at the lateral wall of guinea pig OHCs. An application of suction at the recording pipette reversibly increased the fraction of time a channel is open (Fig. la). The open probability of the channels is an increasing function of the suction applied to the pipette (Fig. lb). The open time histograms of these channels show that the closing rate does not change significantly during a 5-fold change in open probability with stretching (Fig. ld). Thus, a stretch increases the rate of the rate-limiting step in the opening process. The open probability of these channels increased when the membrane was depolarized (Fig. lc) and when the internal Ca 2+ concentration was higher (seen in detached patches). The current-voltage relationship of the channels in oncell and detached patches show that they are potassium channels (Fig. 2). The slope of the plot shows a 130 pS conductance, and the reversal potentials indicate the permeability ratio of K ÷ to Na + between 15:1 and 30:1. Of the 184 tight-sealed patches that we formed on the lateral wall, 23 patches showed these channels. If we assume that the pipettes (thin glass, 3 Mg2 resistance) form membrane patches of about 2/tm 2 area, the entire lateral wall, which is on average a cylinder 10/tm in diameter and about 50 /tm in length, would contain about 95

172

a) lxP=O

(kPa)

20ms

b)

[4pA

c) 06-

_~'

_.

,a

~ 0.4

20.

o

el C

12. C 02

o

o

1 pressure

d)

2 (-kPa)

o

-50 30 membrane

-10 10 potential

30 (mY)

channels. While the total number of channels may appear small, it is comparable to the estimated total number of mechano-transducer channels associated with the stereocilia [7]. The relatively low density of these channels can explain why we did not see multiple channel openings in any of the patches. The importance of these channels can be examined by subjecting the entire cell membrane, instead of a small part of the membrane in on-cell patch experiments, to an externally applied tension and monitoring the response with whole-cell recording electrodes. For this experiment, we controlled the osmolarity of the external medium with a non-electrolyte. This was accomplished by using the same low concentration of NaC1 (100 mM) for both control and hypo-osmotic media (Fig. 3). When the cell was externally perfused with a hypo-osmotic medium, a 10 mV hyperpolarization was recorded (Fig. 3a). An increase of conductance, monitored by the response to current pulses, accompanied the hyperpolarizing response to the osmotic stress, indicating the opening

300.

e 2oc

~oo0

tdP=O x=2.7

20-

& p = -1 k P a 3

Irls

x=28ms

~oo

ID o

~'0~

.....

2'o

open time (ms)

o %'o

,

open tim

~90 000

[]

A

ms)

Fig. I. Stretch activated channels in lateral wall of the OHC. The membrane potential of the cell was measured in whole-cell recording mode after taking these records. The recording pipette contained high-K medium. The bath medium was high-Na medium (see below), a: current record. Pipette potential was clamped to 0 mV (membrane potential - 30 mV). Upward and downward deflections correspond to opening and closing of the channel, respectively. The stretch sensitivity was observed in all (23) 130 pS channels observed in the lateral wall. b: dependence of open probability on the negative pressure at the pipette. On-cell patch. Pipette potential 0 mV (membrane potential - 30 mV). [] and * represent different patches. Saturation of the response was not seen, because we could not maintain on cell-recording mode at negative pressures beyond 2 kPa. c: dependence of open probability of the channel on the membrane potential. Inside-out patch. No suction at the recording pipette, d: open time histogram of the channel, with and without - I kPa pressure at the pipette, r represents the relaxation time of the open state. The scales of the ordinates reflect the difference in open probability. Closed time histograms have 3 exponential components. All relaxation times for these components decreased on application of negative pressure at the pipette. Isolated OHCs were prepared from the cochlea of guinea pigs (200-250 g) by mechanical dissociation with or without pretreatment with 0.5 mg/ml dispase (Boehringer Mannheim) for 20 min at 22°C and maintained in L-15 culture medium. Media used: high-K medium (in mM, KC1 145, MgCI2 2, CaC12 1.5, HEPES 5, pH 7.4), and high-Na medium (NaCI 145, KC1 5, MgC12 2, CaC12 1.5, HEPES 5, pH. 7.4). Open probability and open time histogram were evaluated from each record of 22 s duration. Technical details of data acquisition and analysis are described previously [8].

DID

v

"1o

10[]

E

t~

•[i] m

t--

I-1

[] •" ,

0

0

O 0

/

X

o

x

4-

-10 -80

'

'

'

O I

-40

'

'

'

'

0

'

'

I

'

'

40

'

I

80

membrane potential (mY) Fig. 2. L V plot of single channel current, x + : on-cell patches with high-K medium in the pipette and high-Na medium in the bath; ©: inside~ut patch with high-K medium in the pipette and high-Na medium in the bath; D, ~: inside-out patches with high-K medium in the bath and high-Na medium in the pipette. For plotting L-V relationship for on-cell patch, resting membrane potential of - 3 0 mV was used. The slopes give approximately 130 pS conductance in all three conditions. Media used are described in Fig. I.

173

a) -50 mV --60 mV -

-

lOS

b) -20 mV

~

11111111111111111111111111111!1111!!/1111111111111111111111111111

-- " ~ - -

- ....

i, 7-1--i

.....

"

,L p , . . i , , i . . l | , . l r

I~l'llrl~ ~

-4o ItlIIIIIIlIIIIIIIIIIIIIIIIIIIIIIIIIUlIIItlIIIIIIIIIIIIIIIIIIIIIIIIIIIIII lllillllllJllllllllill fllllllllllllllllllllllllllllllllMllllllllll,lllllll i =v

-=

°v

-80

mV

111111111111111111111111t1111111111111111111tl lt11111111111tlltl1111111 11111111"llI I I I irlrlTl! q r, rqll ~111111111 q IIl11ITIIIi I [lllrl iI ITTITfi ITII~

o.I nA 105

c) XxXX%:~

xXxx xxx

xxx

xXxx __j2~m los

d)

_= =v ]

-LIJJJJL

U_

90=V _ _

.25

nA

2s Fig. 3. Whole-cell response to a hypo-osmotic medium. With a perfusion pipette, a hypo-osmotic medium was applied to the cell for durations indicated by bars under individual records, a: membrane potential record under current clamp. Spikes represent current injections of 0.2 nA. A reduction in spike length corresponds to a conductance increase. The amplitude of the change was 10+3 mV (n= 14). b: current record under voltage clamp. Holding potentials are indicated on the left. Spikes are responses to 10 mV voltage pulses. The half times for response and for recovery were 4.5 + 0.5 s and 8.5 ___0.7 s, respectively (n=4). c: time course of the cell length under voltage clamp (corresponding to b, - 4 0 mV). A calibration, using amplitude of the length change (reduction), showed that the osmolarity of the medium around the cell during the perfusion was 15% lower. The half times for response and for recovery were 3.8+0.5 s and 9.0+0.7 s, respectively (n = 4). d: sections of current record under voltage clamp before (left), during (middle), and after (right) osmotic stress. Voltage pulses 20 mV were applied during the experiment to monitor conductance of the patch. An increase of current pulse corresponds to an increase in patch conductance. Perforated patch whole-cell recording method [6] was used to prevent deflating the cell. Media used were: hypo-osmotic bath medium (HOM) (in mM: NaC1 100, KC1 5, MgCI2, 2, CaC12 1.5, HEPES 5, pH 7.4), regular-osmolarity bath medium (90 mM mannitol added to HOM) and pipette medium (KCI 145, MgC12 2, CaC12 2.81, EGTA 5 and HEPES 5, pH 7.4), which had 0.1 aM free Ca 2+.

Fig. 4. Light micrograph of an isolated outer hair cell with a patch pipette at the lateral wall (a). Fluorescence image of the OHC labelled with rhodamine-phalloidin, specific to actin (b). The distribution of actin shows that the relatively rigid cytoskeletal structure along the cell body is limited to a central core and to the lateral wall structure. Electron micrograph (c) of an area corresponding to a rectangle (in a) shows the lateral wall of the OHC visualized by the freeze-etching electron microscopy technique. Water has been etched away to expose the true outer surface of the lateral plasma membrane (M), the stack of membranes of the subsurface laminated cisternae (S), and the filament components of the cortical lattice (L). The cortical lattice is structurally coupled to the plasma membrane, particularly to the dense population of intramembrane particles, by a framework of pillar structures (arrowheads). Bar = 10/an (a,b), 100 nm (c).

174 o f potassium channels. A control perfusion with the equi-osmolar medium did not cause any response. The corresponding current was observed in voltage clamp experiments (Fig. 3b). The time courses o f the current and cell length were similar (Fig. 3c). Details of the current response show a conductance increase while the cell was subjected to the stress (Fig. 3d). They also show the reversal o f the current between the holding potentials o f - 6 0 mV and - 9 0 inV. A n interpolation gives a reversal potential of - 8 0 mV. The amplitude o f the current response was 30 p A at -- 20 mV. Since the unit current for a potassium channel is 4 p A in the current-membrance potential plot (Fig. 2), the whole-cell current corresponds to 7.5 open channels at any time. A value 0.1 (cf. Fig. lb) for open probability gives 75 channels, about equal to the estimate based on the singel channel recording. C o m p a r i s o n o f the whole-cell response with the single channel current indicates that the stretch sensitivity o f the cell is based on the stretch-activated potassium channels in the lateral wall. The whole-cell response did not show a contribution o f non-selective channels from either the stereocilia [13] or the cell b o d y [3]. Moreover, this response is not mediated by a rise o f Ca 2+ from the external source, since it remained unchanged when we used media with low Ca 2+ concentration (1/tM) both in the bath and in the external perfusing pipette. In the cochlea, the elongated cylindrical cell b o d y of the O H C is optimally positioned to sense axial forces due to vibrations o f the basilar m e m b r a n e during sound stimulation. The cylindrical cell b o d y o f the individual O H C is primarily supported by a cortical lattice containing actin (Fig. 4) [5]. Freeze-etching further shows a direct association of the pillar filaments o f this cortical lattice with the dense population o f large intra-membrane particles o f the plasma membrane. Structural coupling of the cortical lattice to the plasma m e m b r a n e by the framework o f micropillar structures could convey axial forces applied to the cell b o d y to the lateral plasma membrane. This structural efficiency o f the O H C may explain why a macroscopic current attributable to stretch-sensitive channels has not previously been observed in other cells [4, 11]. This stretch sensitivity o f the O H C could be associated with a unique non-linear response o f the cell to acoustic stimuli [2], i.e. greater voltage response in hyperpolarizing direction than in depolarizing. This property is not attributed to different transducer channels at the stereo-

cilia since stimuli specifically applied to O H C stereocilia produced responses similar to that o f the inner hair cell [13]. We suggest that acoustic stimuli stretches the O H C besides bending stereocilia, producing the unique response. The voltage response of the cell, in turn, is associated with movements, fast and slow, o f the O H C . The fast mechanical motility [9], is voltage dependent and hyperpolarization elicits elongation. The O H C also has slower length changes associated with a train o f acoustic stimulation [1, 10], where these stretch-activated channels may be receptors. In association with these motilities o f the cell, the stretch-sensitive channel could be an important element of the feedback mechanisms that have been proposed for the fine tuning o f the m a m m a lian hearing organ [12]. Brundin, L., Flock, A. and Canlon, B., Sound-induced motility of isolated cochlear outer cell is frequency specific, Nature, 342 (1989) 814 816. Cody. A.R. and Russell, l.J., Outer hair cells in the mammalian cochlea and noise induced hearing loss, J. Physiol., 383 (1987) 551 569. Ding, J.P., Salvi, R.J. and Sachs, F., (Abstract) ARO Midwinter Meeting, 13 (1990) 226. Guharay. F. and Sachs. F.. Stretch-activated single ion channel currents in tissue-cultured embryonic chick skeletal muscle, J. Physiol., 352 (1984) 685 7(11. Holley, M.C. and Ashmore, J.F., Spectrin, actin and the structure of the cortical lattice in mammalian cochlear outer hair cells, J. Cell Sci., 96 (1990) 283 291. Horn, R. and Marty, A., Muscarimc activation of ionic currents measured by a new whole-cell recording method, J. Gen. Physiol., 92 (1988) 145 159. Hudspeth, A.J., How the ear's works work, Nature, 342 (1989) 814 816. Iwasa, K., Ehrenstein, G., Moran, M. and Jia, M., Evidence for interactions between two batrachotoxin-modified sodium channels, Biophys. J., 50 (1986) 531 537. Kachar, B., Brownell, W.E., Altschuler, R. and Fex, J., Electrokinetic shape changes of cochlear outer hair cells, Nature, 322 (1986) 365 367. 10 LePage, E.C. Functional role of the olivo-cochlear bundle: a motor unit control system in the mammalian cochlea, Hearing Res., 38 (1989) 177 198. Morris, C.E. and Horn, R.. Failure to elicit neuronal macroscopic mechanosensitive currents anticipated by single-channel studies, Science, 251 (1991) 1246 1249. 12 Mountain, D.C., Changes in endolymphatic potential and crossed olivocochlear bundle stimulation alter cochlear mechanics, Science, 210(1980) 71 72. 13 Russell, I.J., Richardson, G.P. and Cody, A.R., Mechanosensitivity of mammalian auditory hair cells in vitro, Nature, 321 (1986) 517 5[9.