Expression density and functional characteristics of the outer hair cell motor protein are regulated during postnatal development in rat

Abstract

1. The non-linear capacitance (C_non-lin) of postnatal outer hair cells (OHCs) of the rat was measured by a patch-clamp lock-in technique. C_non-lin is thought to result from a membrane protein that provides the molecular basis for the unique electromotility of OHCs by undergoing conformational changes in response to changes in membrane potential (V_m). Protein conformation is coupled to V_m by a charged voltage sensor, which imposes C_non-lin on the OHC. C_non-lin was investigated in order to characterize the surface expression and voltage dependence of this motor protein during postnatal development.

2. On the day of birth (P0), C_non-lin was not detected in OHCs of the basal turn of the cochlea, whilst it was 89 fF in apical OHCs. C_non-lin increased gradually during postnatal development and reached 2.3 pF (basal turn, P9) and 7.5 pF (apical turn, P14) at the oldest developmental stages covered by our measurements. The density of the protein in the plasma membrane, deduced from non-linear charge movement per membrane area, increased steeply between P6 and P11 and reached steady state (4200 e⁻μm⁻²) at about P12.

3. Voltage at peak capacitance (V_½) shifted with development from hyperpolarized potentials shortly after birth (-88.3 mV, P2) to the depolarized potential characteristic of mature OHCs (-40.8 mV, P14). This developmental difference in V_½ was also observed in outside-out patches immediately after patch excision. During subsequent wash-out V_½ shifted towards the depolarized value found in the adult state, suggesting a direct modulation of the molecular motor.

4. Thus, the density of the motor protein in the plasma membrane and also its voltage dependence change concomitantly in the postnatal period and reach adult characteristics right at the onset of hearing.

Outer hair cells (OHCs) of the mammalian cochlea change their cell length reversibly upon electrical stimulation (Brownell et al. 1985). Contraction-elongation cycles can be evoked by stimuli of at least 25-100 kHz (Dallos & Evans, 1995; Gale & Ashmore, 1997b; Frank et al. 1999). This fast electromotility is thought to constitute the effector arm of an active feedback mechanism, the cochlear amplifier (Davis, 1983; Dallos & Corey, 1991), which enables the exquisite sensitivity and frequency selectivity of the mammalian inner ear.

The electromotility of OHCs is directly driven by changes in membrane potential (Ashmore, 1987; Santos-Sacchi & Dilger, 1988; Dallos et al. 1991; Kalinec et al. 1992) and is accompanied by translocation of electrical charge across the basolateral membrane (Ashmore, 1989; Santos-Sacchi, 1991), similar to the gating currents of voltage-gated ion channels (Armstrong & Bezanilla, 1974). Based on the tight coupling of electromotility and charge transfer and the localization of both effects in the lateral membrane, a transmembrane protein has been proposed as the molecular basis of OHC electromotility (Dallos et al. 1991; Kalinec et al. 1992). This protein, termed motor protein (Gale & Ashmore, 1997a), is thought to constitute a charged voltage sensor that moves through the membrane electrical field, thereby inducing a conformational change in the protein, which in turn alters its area in the membrane plane (Dallos et al. 1993; Iwasa, 1994). The molecular identity of the OHC motor protein is as yet unknown.

He et al. (1994) have shown that postnatal OHCs of gerbils begin to acquire electromotile properties 7 days after birth, and reach adult motility levels shortly after the onset of hearing at about P12. During these first 2 weeks, OHCs change from small epithelial cells to the elongated cylindrical cells characteristic of the adult mammalian organ of Corti. Maturation includes formation of the lateral cortex of the OHC, which is composed of the cytoskeletal lattice and one to several layers of subsurface cisternae (Pujol et al. 1991; Weaver & Schweitzer, 1994). It is, therefore, unclear whether the onset of motile properties reflects the expression pattern of the putative motor or rather results from the acquisition of the prerequisites of motility, such as the elastic properties of the OHC or cell turgor (Holley & Ashmore, 1988).

To address this question, we aimed to derive information on the expression of the motor protein during postnatal development by measuring the voltage-dependent charge transfer that is associated with the voltage sensor. Fast charge translocation imposes a voltage-dependent capacitance on the OHC (Santos-Sacchi, 1991; Gale & Ashmore, 1997a), which was measured with a high resolution patch-clamp lock-in technique (Neher & Marty, 1982; Fidler & Fernandez, 1989).

METHODS

Preparation

Organs of Corti were dissected from the cochleae of postnatal Wistar rats as described previously (Oliver et al. 1997). Briefly, animals were killed by decapitation, cochleae were dissected and the organ of Corti was separated from the modiolus and stria vascularis. The basal and apical turns were mounted onto coverslips with glass filaments. The preparation was performed in a solution containing (mM): 144 NaCl, 5.8 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 10 Hepes, 0.7 Na₂HPO₄, 5.6 glucose, adjusted to pH 7.3 with NaOH. Osmolality was 305 ± 2 mosmol kg⁻¹. For the recordings, OHCs located approximately half a turn from either the apex or the base of the cochlea were selected. If necessary, supporting cells were removed with gentle suction from a cleaning pipette. All experiments were performed at room temperature (21-24°C) within 4 h of dissection. Animal care and experiments followed approved institutional guidelines at the University of Tübingen.

Patch-clamp recording

Whole-cell patch-clamp recordings were performed with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA, USA). Electrodes were pulled from quartz glass, had resistances of 2-3 MΩ and were coated with Sylgard. Electrodes were filled with intracellular solutions, composition (mM): Cs⁺-based solution: 100 CsCl, 20 TEA-Cl, 3.5 MgCl₂, 0.1 CaCl₂, 6 BAPTA, 2.5 Na₂ATP, 5 Hepes, pH 7.3 with CsOH; K⁺-based solution: 125 KCl, 20 TEA-Cl, 3.5 MgCl₂, 0.1 CaCl₂, 5 K₂EGTA, 2.5 Na₂ATP, 5 Hepes, pH 7.3 with KOH. Osmolality was adjusted to 305 ± 1 mosmol kg⁻¹ with glucose. These solutions blocked most of the outwardly rectifying K⁺ currents of postnatal OHCs. After seal formation on the basolateral membrane the whole-cell configuration was achieved by rupturing the patch with electrical zapping to avoid disturbance of the cell turgor by suction. Whole-cell series resistances ranged from 3 to 10 MΩ. Cells were voltage clamped at -60 or -80 mV, without correction for liquid junction potentials (6 mV and 4 mV for Cs⁺- and K⁺-based pipette solutions, respectively).

During recordings the bath was continuously perfused with dissection solution (see above). For one set of experiments, OHCs were locally perfused with extracellular solutions of modified osmolarity via a glass capillary. These solutions had the following composition (mM): 125 NaCl, 5.8 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 10 Hepes, 0.7 Na₂HPO₄, 5.6 glucose, adjusted to pH 7.3 with NaOH. Osmolality was adjusted to 285, 305, 365 or 405 mosmol kg⁻¹ with sucrose.

Outside-out patches were excised from the basolateral membrane of OHCs after a quick test for non-linear capacitance in the whole-cell mode. K⁺-based pipette solution was used for outside-out patch experiments.

Capacitance measurements

Voltage-dependent capacitance was measured by two different methods. First, the phase-tracking technique (Neher & Marty, 1982; Fidler & Fernandez, 1989), which employs a software lock-in amplifier and was implemented as an IgorPro XOP (Herrington et al. 1995), was used. After manual compensation of membrane capacitance (C_m), lock-in phase angles yielding signals proportional to changes in C_m and conductance were calculated by dithering the series resistance by 0.5 MΩ. The capacitance output was calibrated by a 100 fF change in the C_m compensation setting. Command sinusoid (2.6 kHz, amplitude 10 mV, sampling rate 83 kHz) was filtered at 5 kHz with an 8-pole Bessel filter, and 16 periods were integrated to generate each capacitance point. To obtain the voltage dependence of membrane capacitance, voltage ramps ranging from -120 to +70 mV or -160 to +30 mV (slope, 0.16 V s⁻¹ or -0.16 V s⁻¹) were summed to the sinusoid command during capacitance measurements. Scaled capacitance traces as well as the conductance signal and DC current induced by the voltage ramp were plotted versus the membrane voltage.

The second method employed a stairstep stimulus (Huang & Santos-Sacchi, 1993a), which increased membrane potential in 5 mV steps, each lasting 5 ms. C_m compensation was omitted in these measurements. Current output filter was set to 10 kHz (sampling rate, 100 kHz). For each voltage step, the decay time constant of the current, τ, was obtained from a monoexponential fit to the transient. Input resistance (R_in) was calculated from the steady-state current and charge (Q) was obtained by integration. Capacitance C_m(i) at each voltage step i was derived by an equation given by Huang & Santos-Sacchi (1993a), taking into account the contribution of the relatively low R_m of mature OHCs:

(1)

where

(2)

(3)

R_s is the series resistance, R_m is the membrane resistance and V_c is the command voltage step amplitude. C_m(i) was plotted versus (V(i) +V(i+ 1))/2, after correction of all voltages for series resistance errors.

Data evaluation

Processing and fitting of data were performed with IgorPro (Wavemetrics, Lake Oswego, OR, USA) on a Macintosh PowerPC.

The capacitance was fitted with a derivative of the Boltzmann function:

(4)

where C_lin is the residual linear membrane capacitance, not compensated for by the C_m compensation circuit of the amplifier, V is the membrane potential, Q_max is the maximum voltage sensor charge moved through the membrane electrical field, V_½ is the voltage at half-maximum charge transfer and α is the slope factor of the voltage dependence. Maximum voltage-dependent capacitance (C_peak) occurs at V_½.

The whole-cell membrane capacitance (C_m) of OHCs was determined from the current transient evoked by a 10 mV step depolarization with C_m compensation disabled, as described above for the stairstep method (eqn (1)). Linear capacitance was calculated as the difference between C_m and non-linear capacitance. When the stairstep method was applied, linear capacitance was directly derived as C_lin from eqn (4). As a measure of the motor protein density in the OHC membrane, the maximum charge transfer, Q_max, reflecting the number of voltage-sensing proteins, was divided by the linear capacitance, which is proportional to the membrane area. This normalized charge transfer will be termed charge density.

All data are given as means ± standard deviation.

RESULTS

Non-linear capacitance in postnatal OHCs

Non-linear capacitance (C_non-lin) as measured in rat OHCs at three stages of postnatal development is shown in Fig. 1A. At any stage, C_non-lin showed a bell-shaped voltage dependence that could be well fitted with the derivative of a first-order Boltzmann function (eqn (4)). At P6, for example, these fits yielded values for C_peak, V_½ and α of 0.79 ± 0.33 pF, -73.0 ± 7.6 mV and 30.3 ± 1.6 mV, respectively (n = 10). C_non-lin was not detected in supporting cells (Deiters cells) at any developmental stage (data not shown).

Figure 1. C_non-lin from OHCs of postnatal rats measured with the phase-tracking technique.

A, C_non-lin from apical OHCs of the developmental stages indicated. Residual linear capacitance was subtracted for display. Inset: C_non-lin (P0) on an enlarged scale. The line is the fit of eqn (4) to the data. Fit parameters were: V_½= -88.7 mV, α= 27.8 mV, Q_max= 7.2 fC at P0; V_½= -66.3 mV, α= 30.0 mV, Q_max= 116.3 fC at P6; and V_½= -51.8 mV, α= 30.4 mV, Q_max= 522.1 fC at P11. B, capacitance (upper trace), conductance (middle trace), and DC current (lower trace) as obtained with the phase-tracking technique during a voltage ramp (1.2 s) from -120 to +70 mV (P7, apical). Squares overlaid onto the upper trace are capacitance data from the same cell obtained with the stairstep method. Fits to the capacitance (not shown) yielded V_½= -60.1 mV, α= 27.6 mV, Q_max= 192.3 fC (phase-tracking) and V_½= -62.2 mV, α= 26.9 mV, Q_max= 195.4 fC (stairstep). Cs⁺-based pipette solution was used throughout the experiments shown in A and B.

As shown in Fig. 1A, the amplitude of C_non-lin varied considerably between developmental stages, necessitating an experimental approach that is both highly sensitive to detect small C_non-lin and robust for evaluation of large changes in capacitance. Figure 1B shows capacitance, conductance, and DC current traces recorded with the phase-tracking technique in response to a voltage ramp from -120 to +70 mV. No cross-talk between capacitance and conductance signals occurred, indicating that this capacitance technique can be applied with negligible errors, even though some residual potassium current was activated by depolarization. Phase tracking allowed the measurement of voltage-dependent capacitances of C_peak < 50 fF in neonatal OHCs (Fig. 1A, inset). However, for C_peak > 4 pF, significant cross-talk between capacitance and conductance traces became obvious in the phase tracking data. Therefore, quantitative measurements of capacitances > 4 pF were achieved with the stairstep method. For intermediate peak capacitances, data obtained by the two methods were in good agreement (Fig. 1B), while small capacitance changes (C_peak < 1pF) could not be resolved sufficiently with the stairstep method due to its lower signal-to-noise ratio.

Development of motor protein expression

The increase in C_non-lin was investigated in more detail by determining its growth from P0 to P14 (Fig. 2A). At P0, C_non-lin was not observed in OHCs from the basal turn of the cochlea (n = 6 cells), while a small amount was found in apical OHCs (C_peak= 89 ± 36 fF, n = 8). During subsequent development, C_peak gradually increased to 2.3 ± 0.7 pF (n = 6) in basal and to 2.4 ± 0.5 pF (n = 7) in apical OHCs at P9. Since basal OHCs from older animals died rapidly after dissection of the cochlea, experiments on basal OHCs were restricted to cells younger than P10. The C_peak of apical OHCs further increased to reach 7.5 ± 0.8 pF at P14 (n = 12).

Figure 2. Increase in C_non-lin during postnatal development.

Peak C_non-lin (A), linear capacitance (B) and charge transfer density (C, see Methods), plotted as a function of time after birth for OHCs from the basal turn (○) and the apical turn (•) of the rat cochlea. The numbers of cells measured at the various ages are 6, 6, 7, 5, 6, 5, 5, 6, 1 and 6 for basal cells and 8, 4, 8, 5, 7, 6, 5, 8, 9, 7, 1, 6, 6 and 12 for apical OHCs. Cells were obtained from individuals of 4 litters. Error bars indicate standard deviation.

Simultaneously with C_non-lin, the size of OHCs changed, as reflected by changes in the linear capacitance, which is proportional to membrane area. In apical OHCs, membrane area increased as the cells elongated to establish their adult appearance, whereas it decreased slightly in basal OHCs (Fig. 2B). Thus, to derive information about protein density in the membrane the charge, Q_max, was normalized to the linear membrane capacitance. Figure 2C shows that charge density increased steeply between postnatal days 5 and 10. In contrast to the absolute value of C_non-lin, however, charge density displayed saturation after P11, with a steady-state value of 67.5 ± 12.5 fC pF⁻¹ at P14. This indicated that mature motor protein density is established at this stage, and further increase in non-linear capacitance is due to an increase in membrane area containing the same density of motor proteins.

Voltage dependence changes during postnatal development

In addition to its density in the plasma membrane, the voltage dependence of the motor protein changed during postnatal development. As shown in Fig. 3B, V_½ shifted from -88.3 ± 2.7 mV (P0) to -40.8 ± 6.0 mV at P14. This shift might be caused by biochemical modulation of the motor protein or it might reflect the influence of cellular properties, such as membrane tension (Iwasa, 1993; Gale & Ashmore, 1994; Kakehata & Santos-Sacchi, 1995) or structural elements, on functional characteristics of the protein. To test for alteration of membrane tension as a mechanism underlying the shift of V_½, we abolished tension by deflating the cells via perfusion with hyperosmotic (365 or 405 mosmol kg⁻¹) extracellular solution. Fast cell collapse and reinflation of the cells following solution exchange were monitored through the experimental microscope. As found previously (Kakehata & Santos-Sacchi, 1995), removal of cell turgor (and thereby membrane tension) led to a reversible shift of V_½ to more negative potentials (Fig. 4). The osmotic collapse of 15 OHCs (P12-14) yielded a V_½ at zero tension of -54.9 ± 7.4 mV, a value significantly different from that observed in neonatal OHCs (P < 0.001, Student's t test). Furthermore, perfusion of OHCs at P3 and P4 with hyper- and hyposmotic solution also shifted V_½ reversibly (data not shown). This indicated that a mechanism different from increase in membrane tension must account for the observed developmental shift of V_½.

Figure 3. Voltage dependence of the motor protein changes during postnatal development.

A, C_non-lin from P4 and P10 apical OHCs normalized to the respective C_peak. Note the developmental shift of V_½ at constant slope. Parameters yielded by fits (not shown) as in Fig. 1 were V_½= -79.1 mV, α= 31.1 mV for P4 and V_½= -36.2 mV, α= 30.5 mV for P10. B, V_½ of apical OHCs plotted versus time after birth (n = 3, 4, 8, 5, 7, 6, 8, 10, 10 and 12, respectively). Measurements were obtained within 30 s of establishment of the whole-cell configuration. Error bars indicate standard deviation.

Figure 4. Release of cell turgor reversibly shifts V_½.

A P13 OHC was perfused with hyposmotic solution. Upon switching to hyperosmotic solution, the cell quickly collapsed as viewed through the experimental microscope, but recovered after return to control solution. V_½ was monitored continuously by taking a phase-tracking measurement of C_non-lin every 3.15 s. Note that V_½ does not reach the hyperpolarized values of neonatal OHCs.

More information about the mechanism underlying the V_½ shift was gained from experiments with excised patches. C_non-lin was consistently detected with the phase-tracking technique in outside-out patches (Fig. 5), excised from the basolateral membrane of postnatal OHCs. At stages P13-14 C_peak was 30.2 ± 14.1 fF (n = 6), while at P2 only a small capacitance signal was detected in 3 out of 7 patches (C_peak= 2.1 ± 0.7 fF). These values of C_non-lin in patches correspond to the developmental differences found in the whole-cell configuration. Moreover, the developmental shift in V_½ that was observed in the whole-cell configuration was reproduced under cell-free conditions: V_½ was -83.4 ± 12.6 mV at P2 and -39.1 ± 16.5 mV at P13/14 when measured immediately after excision of the patches (Fig. 5A). Following excision, however, V_½ slowly shifted to more depolarized potentials, finally yielding V_½≥ -20 mV. This was observed in patches from both P2 and P13/14 (Fig. 5B). These findings indicate that the developmental shift of V_½ can be mimicked in excised patches by wash-out of the cytoplasm and suggest modulation of the motor protein rather than an interaction with structural elements, such as the cortical lattice or subsurface cisternae.

Figure 5. Changes in V_½ in outside-out patches during development and wash-out.

A, normalized C_non-lin measured in outside-out patches from OHCs at P2 and P13. Lines represent fits to the data as in Fig. 1. Parameters were V_½= -70.3 mV, α= 24.2 mV, Q_max= 0.28 fC at P2, and V_½= -37.9 mV, α= 25.5 mV, Q_max= 5.41 fC at P13. B, V_½ shifts to depolarized membrane potentials following patch excision in both P2 and P13 patches. Measurements are from the same patches as in A. The pipette solution was K⁺ based.

In contrast to V_½, the slope (α) of voltage dependence remained constant throughout the postnatal development. At P1, α was 31.7 ± 2.0 mV and at P11 it was 29.4 ± 0.8 mV. Calculation of the sensor charge that moves through the membrane field yields a charge valency of 0.80 e⁻ and 0.87 e⁻, respectively, closely matching data from adult guinea-pig OHCs (Santos-Sacchi, 1991, 1998; Tunstall et al. 1995).

DISCUSSION

Here we describe the occurrence and developmental changes of C_non-lin in postnatal OHCs from rat, using high-resolution lock-in techniques. This capacitance is generated by a transmembrane protein, which presumably underlies fast electromotility (Huang & Santos-Sacchi, 1993a, 1994; Iwasa, 1994). Capacitance measurements suggest the onset of surface expression of the motor protein shortly after birth and a subsequent steep increase in abundance as well as a change in its voltage dependence. Both surface expression and voltage dependence change with a similar timecourse and reach adult characteristics at the onset of hearing prior to the subsequent refinement of auditory function (Henley et al. 1989; Rybak et al. 1992). The developmental expression pattern shown here might be used for molecular identification of the motor protein in the future.

Onset and development of expression of the motor protein

C_non-lin, reflecting functional expression of the motor protein, can first be detected in rat OHCs at P0 to P1, and protein expression increases by more than two orders of magnitude during postnatal development. Expression density saturates concomitantly with the onset of hearing, which occurs at about P12 in rat, as indicated by the onset of the compound action potential (Rybak et al. 1992) and otoacoustic emissions (Henley et al. 1989). Assuming a specific membrane capacitance of 10 fF μm⁻², charge transfer is then about 4200 e⁻μm⁻², similar to values obtained from adult guinea-pig OHCs (Santos-Sacchi, 1991; Huang & Santos-Sacchi, 1993a; Gale & Ashmore, 1997a). No attempts have been made here to account for the apical and basal pole membrane areas, which are known to be devoid of non-linear capacitance (Huang & Santos-Sacchi, 1993a). The development of electromotility has been studied in gerbil OHCs (He et al. 1994). Though inter-species differences have to be considered, the functional maturation of the auditory system with the onset of hearing around P12-14 and the subsequent refinement of auditory function is very similar in rat (Henley et al. 1989; Rybak et al. 1992) and gerbil (Woolf & Ryan, 1984; Norton et al. 1991), allowing useful comparison of OHC maturation between the two species. Using the microchamber technique, He et al. (1994) reported the first appearance of motility at P7 in basal and at P8 in apical OHCs. The threshhold of the motile reaction improved up to P12-14, when adult sensitivity was reached. This increase in sensitivity correlates well with the postnatal period in which we detected a steep increase in charge transfer density, and may, therefore, be linked to an increase in surface expression level of the motor protein. More specifically, the relative length change of OHCs, which should be proportional to the number of independent motor units per cell length unit, saturates at about P14 in the gerbil, reasonably matching our data on protein density. However, it should be stressed that we detect a significant amount of functional motor protein prior to the onset of electromotility. An explanation for this divergence might be that the motile amplitudes provided by a small number of motor units may not exceed the noise floor of the photodiode measurements (He et al. 1994).

There is substantial experimental evidence that OHC motility originates in the plasma membrane and that the cytoskeletal cortical lattice and the subsurface cisternae are not essential for electromotility (Kalinec et al. 1992; Huang & Santos-Sacchi, 1994; Gale & Ashmore, 1997b), although the cytoskeleton appears to be necessary to shape the macroscopic movement of the cell, i.e. elongation and contraction in the longitudinal axis (Holley et al. 1992). In addition, cell turgor is required to establish the macroscopic movement, though voltage-dependent transitions of the motor protein occur in the absence of positive pressure (Santos-Sacchi, 1991). It is, therefore, conceivable that the maturation of structural features of the OHC determines the onset of electromotility. Indeed, Pujol et al. (1991) observed a developmental correlation between the onset of electromotility and the maturation of the cytoskeleton and subsurface cisternae. This might also explain the base-to-apex gradient in the onset of motility found in the gerbil (He et al. 1994) and guinea-pig (Pujol et al. 1991), which is not obvious in the expression pattern of the motor protein. The assembly of motor units into the tightly packed orthogonal array of particles seen in the adult OHC membrane (Forge et al. 1991; Kalinec et al. 1992) might also be a prerequisite for the generation of macroscopic movements and might lag behind the appearance of the protein in the membrane.

Freeze-fracture replicas of the OHC lateral membrane show a large increase in the number of intramembrane protein particles during postnatal development in the gerbil (Souter et al. 1995). Though different populations of particles were observed, the agreement with the onset of non-linear capacitance, e.g. occurrence prior to the onset of motility, indicates that at least a subset of these particles might indeed represent the motor protein.

Changes in voltage dependence

In addition to the deduced upregulation of protein density, voltage dependence of the motor protein undergoes a significant change during the first two weeks after birth. V_½ shifts by as much as 50 mV, reaching the value reported for adult guinea-pig OHCs (Ashmore, 1989; Santos-Sacchi, 1991; Gale & Ashmore, 1997a). In contrast, the slope of the voltage dependence did not change. Although the mechanism underlying this change remains to be elucidated, we ruled out several mechanisms previously shown to affect V_½. The influence of membrane tension, which has been shown to shift V_½ towards more positive voltages (Iwasa, 1993; Gale & Ashmore, 1994; Kakehata & Santos-Sacchi, 1995), can be excluded since the V_½ shift persists at zero membrane tension as well as in outside-out patches. Equally, dependence of V_½ on membrane potential (Santos-Sacchi et al. 1998) does not change during development (D. Oliver, unpublished observation). Moreover, the V_½ difference remains unaltered by patch excision, suggesting that membrane-associated structures such as the cortical lattice and the subsurface cisternae are not directly involved in modulating voltage dependence. Another intriguing explanation for the observed voltage shift would be the assembly of the voltage sensor with an accessory β-subunit, as would be expected to occur if the voltage sensor and molecular motor were located on separate subunits. In isolated patches, however, a slow shift in V_½, mimicking the developmental shift, was consistently observed. This argues strongly against an association-induced shift, but rather resembles the wash-out phenomena well known from a large number of ion channel types (e.g. Fakler et al. 1994). We therefore conclude that it is most likely that the motor protein is subject to a regulatory process, such as phosphorylation/dephosphorylation (Hunter, 1995) or interaction with specific membrane phospholipids, which have recently been shown to regulate a variety of transporters and ion channels (Hilgemann & Ball, 1996; Baukrowitz et al. 1998). In line with this interpretation are recent results indicating that reagents affecting calmodulin-dependent protein phosphorylation may shift non-linear capacitance (Huang & Santos-Sacchi, 1993b; Frolenkov et al. 1999).

The V_½ shift of the voltage sensor explains nicely an observation by He et al. (1994) on isolated OHCs from the gerbil. These authors showed an elongation asymmetry in P7 cells, i.e. the elongation observed upon hyperpolarization is larger than the contraction upon depolarization. A minor population of cells exhibited the reverse behaviour, contraction asymmetry. The relation between cells displaying elongation versus cells displaying contraction asymmetry was shown to reverse during the course of the next week. Elongation asymmetry means that at the resting potential (V_R), more motor units are in the ‘contracted state’ than in the ‘elongated state’, i.e. the resting potential is positive to V_½. Assuming a developmentally constant V_R of around -60 mV (Russell & Richardson, 1987; Oliver et al. 1997), our data show a V_½ shift from negative to V_R, yielding an elongation asymmetry, to positive to V_R, yielding a contraction asymmetry.

In addition to the asymmetry of OHC motility, the gain of the cochlear amplifier, which is defined by the slope of the length change versus voltage function, δL(V), also changes. It is directly proportional to C_non-lin, since this function is the derivative of the voltage dependence of the motor protein. Maximum gain corresponds to peak capacitance, i.e. it occurs at V_½. At the onset of OHC electromotility around P7 (He et al. 1994), V_½ equals the assumed membrane potential of -60 mV, providing maximum gain of the OHC motor. During subsequent development, V_½ shifts away from the resting potential, resulting in decreased motor gain. Nevertheless, the amplitude and sensitivity of motility increase during this period (He et al. 1994), suggesting that the level of motility in maturing OHCs may be explained by the density of functional motor units, as outlined above.