Membrane traffic in outer hair cells of the adult mammalian cochlea

Membrane traffic in outer hair cells of the adult mammaliancochlea

Toshihiko Kaneko,1 Csaba Harasztosi,1 Andreas F. Mack2 and Anthony W. Gummer11Department of Otolaryngology, Section of Physiological Acoustics and Communication, University of Tubingen, Elfriede-Aulhorn-Str. 5, 72076, Tubingen, Germany2Institute of Anatomy, Department of Cellular Neurobiology, University of Tubingen, Osterbergstr. 3, 72074, Tubingen, Germany

Keywords: Ca2+ ⁄ calmodulin, electromotility, endocytosis, FM1-43, guinea-pig, transcytosis

Abstract

Outer hair cells (OHCs), the sensory-motor cells responsible for the extraordinary frequency selectivity and dynamic range of thecochlea, rapidly endocytose membrane and protein at their apical surface. Endocytosis and transcytosis in isolated OHCs from themature guinea-pig cochlea were investigated using the amphiphatic membrane probe FM1-43. We observed membrane transportfrom the apical surface to both the basolateral wall and the subnuclear pole. By double-labelling with DiOC6, a stain for endoplasmicreticulum, and aspiration of the plasma membrane, we showed that the basolateral target was the subsurface cisternae. Thefluorescent signal was about three times weaker at the basal than at the apical pole. The speed of vesicle transport to the subnuclearpole was approximately 0.4 lm ⁄ s. Changing extracellular Ca2+ concentration from 25 lm to 2 mm accelerated rapid endocytosis.Extracellular application of BAPTA-AM (25 lm), an intracellular Ca2+ chelator, and TFP (20 lm), a specific inhibitor of calmodulin,reduced endocytic activity, as did depolarization of the whole cell. The presence of extracellular Cd2+ (200 lm), a Ca2+-channelblocker, had no effect on the voltage dependence of endocytosis at the apical pole, and inhibited the voltage dependence at thesubnuclear pole. These results suggest that rapid endocytosis is a Ca2+ ⁄ calmodulin-dependent process, with extracellular Ca2+

entering through voltage-gated Ca2+ channels at the basal pole. The two distinct destinations of endocytosed membrane areconsistent with the functional polarization of the OHC, with the basolateral wall being dedicated to electromechanical transductionand the subnuclear pole being dedicated to electrochemical transduction processes.

Introduction

Acoustic information in the mammalian cochlea is encoded by theinner hair cells (IHCs). A second type of hair cell, called the outer haircell (OHC), is essential for the exquisite sensitivity of the IHC afferentneurons (Dallos & Harris, 1978). The basis of this sensitivity appearsto be the electromechanical action of the soma (Brownell et al., 1985;Kalinec et al., 1992; Liberman et al., 2002; Cheatham et al., 2004).OHCs are capable of producing mechanical force in response to achange of transmembrane potential (Ashmore, 1987; Dallos & Evans,1995), up to frequencies of at least 70 kHz (Frank et al., 1999).Prestin, member A5 of solute carrier family SLC26, is the motorprotein (Zheng et al., 2000). The motor molecules are locatedexclusively in the basolateral plasma membrane (Kalinec et al., 1992),and are of exceedingly high density (at least 7.500 ⁄ lm2, Huang &Santos-Sacchi, 1993), perhaps to the exclusion of channels notassociated with the motor complex (Ashmore, 1992).Very little is known about the molecular constitution, maintenance

and function of the motor complex. Here we examine mechanisms thatmight be associated with the maintenance ) sorting, recycling orturnover ) of the motor complex. It has been suggested that rapidendocytosis might underlie such mechanisms (Meyer et al., 2001;Griesinger et al., 2004). Hair cells are known to have endocytic

activity at their apical surface (Forge & Richardson, 1993; Hassonet al., 1997; Kachar et al., 1997; Richardson et al., 1997; Meyer et al.,2001; Griesinger et al., 2004). Using the amphipathic membrane dyeFM 1-43 as a marker for endocytosed membrane, evidence for rapidendocytosis has been found in hair cells of the mature guinea-pigcochlea (Meyer et al., 2001; Griesinger et al., 2002, 2004) and sensoryhair cells of the zebrafish lateral line (Seiler & Nicolson, 1999). InOHCs, endocytosed vesicles are found in the infracuticular zone(Hensen’s body) and also along a central strand extending from theinfracuticular zone down to the nucleus, as well as along thebasolateral membrane (Meyer et al., 2001; Griesinger et al., 2004).There is evidence that rapid, apical endocytosis is part of a system-trafficking membrane to the lysosomal system (Meyer et al., 2001;Griesinger et al., 2004) and protein to the basolateral wall (Griesingeret al., 2004). However, there is apparently very little evidence ofendocytosed vesicles at the subnuclear pole (Griesinger et al., 2004),where the OHC is both afferently and efferently innervated (Saito,1980). Ca2+ is a trigger for rapid endocytosis in hair cells (Seiler &Nicolson, 1999; Meyer et al., 2001; Griesinger et al., 2002, 2004).Moreover, Seiler & Nicolson (1999) showed Ca2+ ⁄ calmodulindependence of rapid endocytosis in zebrafish lateral line.Because little is known about endocytosis and transcytosis in

OHCs, the aim of this study was to identify destinations ofendocytosed vesicles, and examine the effects of Ca2+, calmodulinand membrane potential on the dynamics of endocytosis andtranscytosis.

Correspondence: Dr A. W. Gummer, as above.E-mail: [email protected]

Received 23 December 2005, revised 2 March 2006, accepted 13 March 2006

European Journal of Neuroscience, Vol. 23, pp. 2712–2722, 2006 doi:10.1111/j.1460-9568.2006.04796.x

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd

Materials and methods

Cell preparation

OHCs were mechanically isolated from the cochlea of maturepigmented guinea-pigs, weighing 200–400 g. After killing by cervicaldislocation, the temporal bone was dissected from the skull and storedin cooled (4 �C) Hanks’ balanced salt solution (HBSS; Biochrom KG,Berlin, Germany), containing (in mm): NaCl, 137; KCl, 5.4; CaCl2,1.25; NaHCO3, 4.2; MgSO47H2O, 0.81; KH2PO4, 0.44; Na2HPO42-H2O, 0.34; glucose, 5.0; HEPES, 10; with osmolarity of 310 mOsm ⁄ Ladjusted with glucose and pH 7.25. Most chemicals were from Sigma(Taufkirchen, Germany). KCl and HEPES were obtained fromMERCK (Darmstadt, Germany). The bullae were carefully removedand the cochleae exposed. The bony wall of the cochlea was openedwith fine forceps in cooled HBSS to expose the modiolus and the organof Corti. Strips of the organ of Corti were carefully separated from themodiolus using a fine needle and placed in an experimental chamber.OHCs were isolated mechanically using a 100-lL pipette (Eppendorf,Hamburg, Germany) and allowed 10 min to settle onto a poly-l-lysine-coated coverslip (0.1%) in 1.0 mL HBSS. When investigating theextracellular calcium dependence of endocytosis, CaCl2 was dissolvedin Ca2+- and Mg2+-free HBSS to give a Ca2+ concentration of 25 lm.Only healthy-looking OHCs were used; namely, those with cylindricalshape, bifringent and smooth basolateral membrane, nucleus located atthe subnuclear pole of the cell, no Brownian motion of the organellesand (visually) intact hair bundle orientated orthogonal to the apicalsurface of the cell. All experiments were conducted at controlled roomtemperature of 21.5 ± 0.5 �C. All animal procedures were performed inaccordance with approved protocols at the University of Tubingen,complying with legal requirements of animal care in Germany(Deutsches Tierschutzgesetz).

Fluorescent dyes

The stryl pryridinium dye FM1-43 (Molecular Probes, Leiden, theNetherlands) was used to study endocytosis and transcytosis inisolated OHCs. A stock solution (10 mm) of the dye was prepared indimethylsulphoxide and kept at )20 �C. Directly before the experi-ment, the stock solution was diluted in HBSS to the final concentrationof 10 lm. The dye solution was applied locally to the cell usingpressure application (Transjector 5246, Eppendorf) through a borosil-icate glass capillary (GC150F-10, Harvard Apparatus, Edenbridge,UK), drawn to a tip diameter of about 2.5 lm (DMZ Universal Puller,Zeitz-Instrumente, Augsburg, Germany). However, in some of theexperiments the OHCs were incubated with the dye in the bathsolution. The application capillary was positioned with an electricallydriven micromanipulator (Mini 25, Luigs and Neumann, Ratingen,Germany); its tip was located near the middle of the cell at a distanceof about 40 lm from the cell. This configuration allowed the dye tosurround the cell, as viewed with the built-in, high-pressure mercurylamp. Control experiments in which the tip was located at differentaxial positions (apex, middle, base) showed that the intracellularfluorescence signals were independent of tip position.

DiOC6, a label of the endoplasmic reticulum (ER) (Terasaki et al.,1984; Ikeda & Takasaka, 1993), was used in double-labellingexperiments to identify morphological structures stained by FM1-43(Meyer et al., 2001; Griesinger et al., 2004). DiOC6 (MolecularProbes) was prepared in 100% ethanol as a stock solution (1 mg ⁄mL).The stock solution was diluted with HBSS to the final concentration of0.5 lg ⁄mL. After incubation in DiOC6 for 1 min at room temperature,OHCs were thoroughly washed with fresh HBSS using a perfusionpump (Ismatec, Glattbrugg, Switzerland).

Drug application

Dihydrostreptomycin (DHSM), a blocker of open mechanoelectricaltransduction (MET) channels (Hacohen et al., 1989; Kroese et al.,1989; Kimitsuki & Ohmori, 1993; Meyer et al., 1998; Ricci, 2002;Schulte et al., 2002; Marcotti et al., 2005), was used to test forpermeation of MET channels by FM1-43 (Seiler & Nicolson, 1999;Meyer et al., 2001; Griesinger et al., 2002, 2004). Permeation has beenreported for MET channels in the immature inner ear in mouse (Galeet al., 2001; Geleoc & Holt, 2003; Meyers et al., 2003; Cheathamet al., 2004), in chick hair cells (Si et al., 2003) and in the adultbullfrog sacculus (Meyers et al., 2003), as well as in the cochlea ofmouse using AM1-43, the fixable analogue of FM1-43 (Cheathamet al., 2004). We used two protocols. In the first protocol ) bathapplication ) DHSM (1 lL of 100 mm stock solution) was added tothe experimental chamber during the fluorescence measurement; thefinal drug concentration was 100 lm. In the second protocol, OHCswere incubated in DHSM (100 lm) for about 2 min and then FM1-43was applied locally in the presence of DHSM.PPADS, a P2X antagonist (McLaren et al., 1994), was used to test

for permeation by FM1-43 of P2X receptors, known to be abundant atthe apical surface of OHCs (Raybould & Housley, 1997). Permeationof these receptors expressed in HEK 293T cells has been reported(Meyers et al., 2003). Cells were preincubated for 5 min in PPADS(10 or 100 lm) and then FM1-43 applied locally in the presence ofPPADS.To examine the effects of extracellular Ca2+, fluorescence was first

measured with the cell immersed in HBSS of low Ca2+ concentration(25 lm), as described above. CaCl2 (2 lL of 1 m stock solution) wasthen added to the experimental chamber during the fluorescencemeasurement; the final extracellular Ca2+ concentration was 2 mm.Rapid endocytosis has been reported to be dependent on cytosolic

Ca2+ (Neher & Zucker, 1993; Artalejo et al., 1995; Seiler & Nicolson,1999; Meyer et al., 2001; Griesinger et al., 2002, 2004). To examineCa2+ dependence, cells were preincubated for 30 min in BAPTA-AM(25 lm), a chelator of cytosolic Ca2+, and then washed with freshHBSS before applying FM1-43. In a further set of experiments, cellswere preincubated for 15 min in trifluoperazine (TFP; 20 lm), aninhibitor of calmodulin-dependent phosphorylation (Levin & Weiss,1978; Cook et al., 1994; Vandonselaar et al., 1994), and then FM1-43applied in the presence of TFP.

Micropipette aspiration

To distinguish possible endocytic properties of the plasma membranefrom the subsurface cisternae (SSC), we physically separated the twostructures by aspiration of the basolateral wall into a micropipette,using a technique originally described by Sit et al. (1997). Aspirationmicropipettes were fabricated from borosilicate glass capillaries(GC150F-10, Harvard Apparatus) by drawing to an internal tipdiameter of about 3 lm using a DMZ Universal Puller (Zeitz-Instrumente, Augsburg, Germany). The tip was placed on thebasolateral wall of the OHC, near its centre, and negative pressure(£ 10 cm H2O) applied to produce a membrane tongue into thecapillary. NBD-cholesterol (N-1148, Molecular Probes), a fluorescentderivative of cholesterol (Chattopadhyay, 1990) with a high affinityfor the SSC (Oghalai et al., 1999), was used to label the SSC as acontrol. A stock solution (100 mm) of the dye was prepared indimethylsulphoxide and stored at )20 �C. OHCs were incubated inthe NBD-cholesterol (100 mm) for 5 min before the experiments. Toterminate the incubation, the extracellular solution was replaced bydye-free HBSS using a perfusion pump.

Intracellular transport in OHCs 2713

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing LtdEuropean Journal of Neuroscience, 23, 2712–2722

Confocal microscopy

Confocal images were collected using a Zeiss LSM 510 confocallaser-scanning microscope equipped with a META detector (Oberko-chen, Germany). The experimental chamber was mounted on the stageof an upright microscope (Zeiss Axioskop2 FS mot). The objectivewas a Zeiss 40 ·, IR-achroplan, water-immersion lens with NA 0.8and WD 3.61 mm. FM1-43-loaded cells were excited at 488 nm withargon laser light; the emission light was collected above 505 nm usinga long-pass filter. The META detector was used for double-stainingexperiments. The META detector has 32 separate channels to improveseparation of overlapping fluorescence emission spectra. Emissionlight was detected at 493–504 nm for DiOC6 and at 590–643 nm forFM1-43. Pictures were stored with a resolution of either256 · 512 pixels or 512 · 512 pixels, with pixel depth set to 12 bits.The excitation energy of the laser light was reduced as much aspossible to avoid bleaching, saturation and phototoxic effects.Fluorescence signals were quantified and analysed with Zeiss LSM510 software, Photoshop 6.0 (Adobe) and Origin 7 (OriginLab).Quantitative data of fluorescence intensity were derived from regionsof interest (ROIs), with background intensity prior to application ofFM1-43 subtracted.

Patch-clamp experiments

For whole-cell patch-clamping experiments, we used an EPC-9amplifier equipped with the Pulse v8.66 software (Heka Elektronik,Lambrecht ⁄ Pfalz, Germany). Patch pipettes were fabricated fromborosilicate glass capillaries (GC150F-10, Harvard Apparatus) and hadseries resistance of 3–5 MW in the bath, when filled with artificialintracellular solution (in mm): KCl, 140; MgCl2, 2; EGTA, 11;CaCl2, 1; HEPES, 10; with an osmolarity of 315 mOsm ⁄ L adjustedwith glucose and pH 7.2. In order to reduce the electromechanicalmovement artefact caused by stepping the holding potential, the cellwas patched just above the nucleus for fluorescence measurements atthe subnuclear pole, but patched in the apical third for fluorescencemeasurements at the apical pole. After establishing the whole-cellconfiguration, the holding potential was set to )60 mV. Seriesresistance in situ was 7–10 MW and was compensated on-line by atleast 40%. In these experiments, the application of the dye through theperfusion capillary, positioned near the cell, and the collection of thefluorescence signals were the same as described above. CdCl2 was usedto block voltage-gated Ca2+ channels (Lansman et al., 1986). It wasstored in 200 mm stock solution and diluted on the day of experiment tothe final concentration (200 lm) in the perfusion pipette solutioncontaining the FM1-43 dye.

Statistics

Averaged data are expressed as mean ± SEM. Differences betweenaveraged datawere evaluated by the Student’s t-test andwere consideredsignificant if P < 0.05. Least-mean-square curve fitting was performedwith the Levenberg–Marquardt algorithm using SigmaPlot� 2.0 or 5.0software; fitted parameters are given as mean ± SD.

Results

Data presented here are from 65 OHCs from 30 cochleae. Another 485cells were used either to establish the experimental techniques, or wereexcluded because of a detectable movement artefact. All cells wereisolated from the apical half of the cochlea, and had lengths of 65–80 lm.

No detectable effect of DHSM and PPADS on the FM1-43 signal

Because it has been reported that FM1-43 can rapidly pass throughMET channels (Nishikawa & Sasaki, 1996; Gale et al., 2001; Geleoc& Holt, 2003; Meyers et al., 2003; Si et al., 2003; Cheatham et al.,2004) and P2X receptors (Meyers et al., 2003), it was important todetermine whether these routes contribute significantly to the fluor-escent signals.FM1-43 was locally applied to OHCs for 600 s. Fluorescence

signals were taken from the infracuticular zone containing Hensen’sbody (Fig. 1A). Data collated in Fig. 1B are normalized to thefluorescence signal of FM1-43 measured at 600 s (control, n ¼ 6).DHSM is a well-known voltage-dependent blocker of open METchannels, blocking them completely and rapidly at )60 mV and aconcentration of 100 lm. In the first set of DHSM experiments,DHSM (100 lm) was added to the bath solution in the middle of thefluorescence measurement. There was no detectable differencebetween fluorescence intensity before and after DHSM application(n ¼ 4) (Fig. 1B). In the second set of DHSM experiments, OHCswere preincubated in DHSM (100 lm) for about 2 min, and thenFM1-43 was locally applied for 600 s in the continued presence ofDHSM. No significant difference was seen between control cells(1.00 ± 0.13, n ¼ 6) and DHSM-preincubated cells (1.01 ± 0.11,n ¼ 7) (Fig. 1C). Therefore, MET channels are not a significantpathway for uptake of FM1-43 over a period of minutes, the timecourse of our experiments.In the next set of experiments, OHCs were preincubated in a P2X

antagonist, PPADS (100 lm), for 5 min. FM1-43 was then locallyapplied in the presence of PPADS. There was no significant differencebetween control (1.00 ± 0.06, n ¼ 5) and PPADS-preincubated cells(0.97 ± 0.10, n ¼ 4) (Fig. 1C). Therefore, P2X receptors are not asignificant pathway for uptake of FM1-43 over a period of minutes.

Intracellular transport

FM1-43 fluorescent staining was found: (i) at the apical surface; (ii) inthe infracuticular zone, which includes the circular structure, calledHensen’s body; (iii) the central strand; and (iv) the subnuclear area(Figs 1A and 2A), as also reported by Meyer et al. (2001) andGriesinger et al. (2004). The time course of staining is shown inFig. 2B and C for ROI as colour-coded in Fig. 2A. FM1-43 wasapplied for 30 s through the glass capillary at about every 300 s. Thefluorescence signal rose most rapidly in the ROI at the apical surfacewith stereocilia bundle (black curve) and returned exponentially(s ¼ 58 ± 8 s) back to approximately its initial value after terminationof FM1-43 application. However, the signal from the infracuticularzone (red curve) continued to increase after termination of theapplication. Not until about 30 s after terminating the application didthe signal in this region begin to decrease; the decay was exponentialwith s¼ 97 ± 9 s. This is significantly longer than the time constantfor the apical surface. Fluorescence did not recover to its initial levelbefore the next dye application. In contrast to the situation at the apicalsurface and infracuticular zone, the fluorescence signal from thesubnuclear area (green curve) continuously increased during the entiremeasurement period. After about 400 s, the intensity of the basalsignal was, on average, about three times smaller than the infracu-ticular signal. The staining in the subnuclear region was fairlyhomogeneous, without evidence of hot spots (Fig. 2D). It is unlikelythat endocytosis occurred in the basal region, because fluorescenceintensity would be expected to decrease after terminating the dyeapplication, as in the apical region, not to increase as found here. Thus,we conclude that the endocytosed vesicles observed in the subnuclear

2714 T. Kaneko et al.


area derive from vesicles trapped originally at the apical pole, whichwere then transported to the subnuclear pole.

The speed of this intracellular transport can be quantified from thedelays of the onset responses. Referring to the expanded time scale inFig. 2C, this delay (arrows) increases progressively from apex to baseof the cell. This indicates intracellular vesicle transport in thisdirection. Evaluated from linear regression of the onset delays, thespeed amounts to 0.37 ± 0.01 lm ⁄ s. For 15 cells, the mean speed was0.36 ± 0.03 lm ⁄ s.

Endocytic activity in the basolateral wall

There was also evidence for endocytic activity in the basolateral wall(Fig. 3). The initial response to FM1-43 in all three basolateral regionswas extremely rapid (< 1 bin ¼ 6 s) compared with that for theinfracuticular region (red ROI). This initial fluorescence is almostcertainly due to FM1-43 attached to the outer leaflet of the plasmamembrane. Likewise, the initial rapid decay of the signal afterterminating the FM1-43 application (s ¼ 15.1 ± 0.7 s for the firstapplication) is probably due to release of dye attached to the outermembrane leaflet. However, there is evidence for a second time constantof exponential decay, particularly for themost apical basolateral location(yellowROI); (s ¼ 396 ± 56 s for the first application).Moreover, withsuccessive applications of FM1-43 there was a systematic accumulationof fluorescence in the basolateral wall, such that the fluorescence waslargest at the apical location (yellow ROI) and smallest at the basallocation (light blueROI). This suggests that: (i) endocytosed vesicles arelocated at the basolateral wall, possibly in the SSC; and (ii) there isvesicle transport to the more basal locations (assuming that thebasolateral wall is structurally homogeneous along its length).

Double-staining of the subsurface cisternae with DiOC6

Because in OHCs the SSC (and Hensen’s body) consists of severallayers of ER,we looked for evidence of endocytic vesicles in the SSC bydouble-staining with FM1-43 andDiOC6, a stain for ER. Isolated OHCs

were incubated with DiOC6 for 1 min, followed by wash out. FM1-43was applied to the DiOC6-treated cells for 100 s, by local pressureapplication through the glass capillary, as in other experiments in thisstudy. An example is given in Fig. 4. The entire basolateral wall wasstained during application of FM1-43.Upon terminating the application,the fluorescence intensity began to decrease at the apical surface (dashedline in Fig. 4A). However, in the basolateral wall, the fluorescenceintensity continued to increase for about 40 s after terminating theapplication (full line in Fig. 4A). Non-linear fitting from the maximumsignal at t ¼ 150 s up to the maximum observation point at t ¼ 264 ssuggested that the signal exponentially decayed (s ¼ 25.2 ± 5.9 s) to anelevated value, which was as much as 77 ± 2% of the maximal signal.The delayed post-stimulus decrease and the accumulation of endocyticvesicles, implied by the elevated post-stimulus signal, was similar to thatfound for all control cells (e.g. Fig. 3). Co-localized staining by DiOC6

and FM1-43 of SSC in the basolateral wall and in Hensen’s body isclearly seen (Fig. 4B). The two stains were co-localized both during andafter FM1-43 application (Fig. 4B, columns marked 75 s and 250 s,respectively). The longitudinal extent of the two stains coincided.Although it is difficult to distinguish between plasma membrane andSSCs with this double-staining experiment, FM1-43 attached to theouter leaflet of the basolateral plasma membrane is almost certainlyreleased after terminating the application, as it was from the apicalsurface (Fig. 4A). That is, FM1-43 fluorescence intensity in thebasolateral wall after termination of the application appears to derivemainly from intracellular structures in the basolateral wall.The co-localized staining suggests that the SSC is the intracellular

location of the endocytic vesicles. However, although unlikely, strictlythe experiment does not allow us to discount the possibility that the innersurface of the plasma membrane is the main site of endocytic vesicles.

Staining of plasma membrane and subsurface cisternaedistinguished by aspiration

In order to distinguish between the SSC and the plasma membraneas sites of endocytic vesicles, these two structures were physically

Fig. 1. Endocytosis as the main pathway for uptake of FM1-43. FM1-43 was locally applied to OHCs for 600 s. The tip of the application capillary was positionednear the middle of the cell, about 40 lm from the cell (and therefore not visible in the pictures). (A) Fluorescence signals of FM1-43 from the infracuticular zone(rectangular region in top) containing the circular structure, called Hensen’s body (HB in bottom). Scale bar, 5 lm. (B) Average time course for control cells(dashed lines, mean ± SEM, n ¼ 6) and a cell for which dihydrostreptomycin (DHSM, 100 lm) was applied at 300 s during continuous fluorescence measurement(full line). Data from each cell were normalized to their fluorescence at 600 s. There was no significant difference before and after DHSM application (n ¼ 4).(C) Average data for OHCs preincubated with DHSM (100 lm, 2 min) or with PPADS (100 lm, 5 min). FM1-43 was then locally applied for 600 s in thecontinued presence of the drug. Fluorescence intensities were measured at 600 s. No significant difference was seen between control 1 (1.00 ± 0.13, n ¼ 6) andDHSM-treated (1.01 ± 0.11, n ¼ 7) cells, or between control 2 (1.00 ± 0.06, n ¼ 5) and PPADS-treated (0.97 ± 0.10, n ¼ 4) cells.



separated by the aspiration technique. Application of appropriatenegative pressure (£ 10 cm H2O) through a glass capillary placedon the external basolateral surface of the cell separated plasmamembrane from SSC, as shown for a control cell for which SSCwas stained with NBD-cholesterol (Fig. 5A). For a short localapplication of FM1-43 followed by aspiration, fluorescence wasonly observed from the plasma membrane layer (Fig. 5B). How-ever, when incubated in FM1-43 for 3 min, followed by aspiration,there was heavy staining of the SSC compared with staining of theplasma membrane (Fig. 5C). Together with the results of theER-staining experiments (Fig. 4), the results of the aspirationexperiments suggest that the main site of the endocytic vesicles inthe basolateral wall is the SSC.

Calcium-dependent rapid endocytosis

Because it is believed that Ca2+ influx is a trigger for rapid endocytosis(see Introduction), we examined Ca2+ dependence by extracellular(Fig. 6) and intracellular (Fig. 7) manipulation. The extracellular Ca2+

concentration was changed from a low (25 lm) to a high (2 mm)value whilst continuously measuring FM1-43-induced fluorescence. Inall experiments (n ¼ 3), the fluorescence signal more than doubled inboth the basolateral wall and the infracuticular region, as a result of thetwo orders of magnitude increase of extracellular Ca2+ concentration(Fig. 6). There was no detectable change of cell length or width.Chelation of intracellular Ca2+ by incubation in BAPTA-AM (25 lm)caused a significant decrease of the fluorescence signal, as illustrated

Fig. 2. Time course of FM1-43-induced fluorescence along the axis of the cell. (A) Rectangular ROIs (upper panel) and fluorescent staining during dyeapplication at the 610 s point (lower panel). Scale bar, 10 lm. (B) Time course at the apical surface with stereocilia bundle (black), infracuticular zone (red) andsubnuclear pole (green). FM1-43 was applied for 30 s (solid bar) through the glass capillary about every 300 s. The tip of the application capillary was located nearthe middle of the cell, about 40 lm from the cell. Notice: (i) near to complete return to background at the apical surface (s ¼ 58 ± 8 s); (ii) slower off-response(s ¼ 97 ± 9 s) for the infracuticular zone; (iii) monotonic increase of fluorescence at the subnuclear pole. (C) Onset responses examined with greater temporalresolution than in (B) (marked as a dashed rectangle) for the five intracellular ROIs in (A). The onset delay of fluorescence in cytosolic areas increased progressivelyfrom the Hensen’s body area (infracuticular zone) to the subnuclear pole (arrows). The speed of the intracellular vesicle movement, estimated from linear regression,is 0.37 ± 0.01 lm ⁄ s. (D) Fluorescent staining after cessation of dye at the 900 s point. Notice the staining in the subnuclear region (arrow). The same contrastsettings were used for the fluorescent pictures in (A) and (D).



in Fig. 7A for the time course of fluorescence averaged over fourOHCs in Hensen’s body. At 600 s, the fluorescence was, on average,69 ± 6% of the control value (Fig. 7B). These two sets of experimentssuggest that intracellular Ca2+ is required for rapid endocytosis andthat Ca2+ influx is a trigger for this process.

Calmodulin-dependent rapid endocytosis

Calmodulin is a Ca2+-binding protein involved in the regulation ofa number of cellular functions by modulating in a Ca2+-dependent

way. Calmodulin has important roles also for rapid endocytosis andits function is thought to be as a receptor of Ca2+ (Artalejo et al.,1996; Seiler & Nicolson, 1999). Therefore, the effect of TFP, aspecific calmodulin inhibitor, was examined. The protocol wassimilar to that used to describe the dependence on intracellular Ca2+

concentration. Fluorescence intensity was significantly less forTFP-treated cells (Fig. 7A), amounting to about 36 ± 7% ofcontrol at 600 s. Together with the Ca2+-dependence data, thisfinding suggests that Ca2+ ⁄ calmodulin is a regulator of rapidendocytosis.

Fig. 3. Time course of FM1-43-induced fluorescence at the basolateral wall. (A) Colour-coded ROIs show the origin of the fluorescence intensities plotted in (B).Notice: (i) extremely rapid onset (delay < 1 bin ¼ 6 s) and offset responses at the three basolateral wall locations compared with the infracuticular region; (ii) gradualaccumulation of the dye at all locations. The initial fast onset and offset responses are probably due simply to attachment and detachment, respectively, of FM1-43 atthe outer membrane leaflet; later, apically endocytosed vesicles are transported to the wall. Data are from the same cell as in Fig. 2. Scale bar, 10 lm.

Fig. 4. Co-localization of DiOC6 and FM1-43 in ER. FM1-43 was applied to the DiOC6-treated OHC for 100 s. (A) Fluorescence intensities of FM1-43 at theapical surface (dashed line) and at the basolateral wall (full line). Notice that the off-signal from the basolateral wall is delayed and elevated after termination of theFM1-43 application. (B) Staining by DiOC6 (top row) and FM1-43 (middle row) of Hensen’s body (HB) in the infracuticular zone and of subsurface cisternae in thebasolateral wall. Staining by DiOC6 and FM1-43 is co-localized in the basolateral wall and HB, both during (middle bottom) and after (right bottom) FM1-43application. This suggests the presence of endocytic vesicles in the ER of these structures. Scale bars, 5 lm.



Effect of membrane-potential change

Although it is known that potassium depolarization has an influenceon endocytosis of IHCs (Griesinger et al., 2002), the mechanisms arestill unclear. The patch-clamp technique in whole-cell configurationwas used to investigate the effect of membrane-potential change onmembrane transcytosis (Fig. 8). ROIs were at the apical (infracuticu-lar) and the basal (subnuclear) regions of the cell. Holding themembrane potential at )60 mV, continuous application of FM1-43dye evoked a monotonic increase of the fluorescence signal with time,in both the apical (Fig. 8A) and the basal (Fig. 8B) areas of the OHC.Depolarization reversibly reduced the slope of the fluorescence–timecurve. To quantify this change of slope, linear regression was used tofit each phase. Normalized to control, depolarization to 0 mV reducedthe slope to 0.29 ± 0.06 (n ¼ 5) and 0.20 ± 0.05 (n ¼ 7) in the apical(Fig. 8A and E) and basal (Fig. 8B and F) areas, respectively; thesevalues are not significantly different. This effect was reversible:repolarization to )60 mV increased the slope of the signal in the apicaland basal areas to 1.02 ± 0.14 (n ¼ 5) and 0.96 ± 0.11 (n ¼ 5),respectively. Extracellular application of 200 lm CdCl2, a potentblocker of voltage-gated Ca2+ channels, inhibited the effect of

depolarization in the basal area (0.92 ± 0.09, n ¼ 6) (Fig. 8Dand F) but not in the apical area (0.40 ± 0.06, n ¼ 5) (Fig. 8C andE). These data indicate that depolarization facilitates exocytosis at thebase of the cell, and that exocytosis requires Ca2+ entry throughvoltage-gated Ca2+ channels at the base.

Discussion

These experiments demonstrated intracellular membrane transportfrom the apical end to both the basolateral and the subnuclear domainsof mature OHCs. We showed that the dynamics of apical endocytosisand transcytosis are dependent on Ca2+ ⁄ calmodulin and membranepotential, and that Ca2+enters at the basal pole of the cell.

Endocytosis as the main route for uptake of FM1-43

There has been much discussion as to whether FM1-43 can passthrough MET channels (Nishikawa & Sasaki, 1996; Seiler &Nicolson, 1999; Gale et al., 2001; Griesinger et al., 2002, 2004;Geleoc & Holt, 2003; Meyer et al., 2001; Meyers et al., 2003; Siet al., 2003). Here, no obvious blocking effects were observed foreither DHSM treatment during FM1-43 application or DHSM-pretreated OHCs (Fig. 1). DHSM in this concentration (100 lm) isknown to block MET channels completely and rapidly (Hacohen et al.,1989; Kroese et al., 1989; Kimitsuki & Ohmori, 1993; Meyer et al.,1998; Ricci, 2002; Schulte et al., 2002; Marcotti et al., 2005). Inprevious studies, using guinea-pig OHCs (Meyer et al., 2001) andIHCs (Griesinger et al., 2002, 2004), there was also no evidence insupport of passage via MET channels. On the other hand, there arereports of FM1-43 uptake via MET channels in cultured mouse haircells (Gale et al., 2001; Geleoc & Holt, 2003; Meyers et al., 2003),cultured chick auditory papilla hair cells (Si et al., 2003), culturedzebrafish larvae lateral line organs (Seiler & Nicolson, 1999) andXenopus larvae lateral line organs (Nishikawa & Sasaki, 1996). Inaddition, FM1-43 is known to rapidly block transducer currents incultured mouse hair cells (Gale et al., 2001). However, the time

Fig. 5. Staining of the plasma membrane and the SSC distinguished bymembrane aspiration. Non-fluorescence (left) and fluorescence (right) imagesfrom the middle region of three OHCs. (A) Control, NBD-cholesterol stainscholesterol located in the SSC, consistent with the findings of Oghalai et al.(1999). (B) After a 30 s (short), local application of FM1-43, staining wasdetected only in the plasma membrane. (C) After incubation in FM1-43 for3 min (long), both plasma membrane and SSC were stained. However, thestaining of the SSC was heavy; this is consistent with this structure being themain target of endocytic vesicles in the basolateral wall. Black arrows, extent ofthe aspirated SSC. White arrows, extent of the aspirated plasma membrane.

Fig. 6. Increase of FM1-43-induced fluorescence with extracellular Ca2+

concentration. The extracellular concentration was increased from 25 lm to2 mm by the addition of CaCl2 to the experimental chamber. ROIs were fromHensen’s body (full line) and from the basolateral wall (broken line). Thedotted curve gives the predicted (exponential) time course at Hensen’s bodywere the Ca2+ concentration to have remained unchanged at 25 lm. Notice thatthe fluorescence is more than doubled as a result of the concentration increase.Scale bar, 5 lm.



constants for these MET-related processes are much shorter (< < 1 s)than for the endocytic processes observed here: Using a temporalresolution of 6 s and observation times up to 600 s, it is unlikely thatentry through MET channels would be detectable on the long timescale of our experiments. In addition, MET channels might be morepermeant in developmentally immature systems. Likewise, we did notfind evidence for entry through P2X receptors (Fig. 1C), which areexpressed in high density in the apical membrane of OHCs (Raybould& Housley, 1997). Moreover, this latter finding is consistent with theresults of Meyers et al. (2003) when cells are not stimulated with ATP.The most parsimonious conclusion is that the main pathway for uptakeof FM1-43 was via endocytosis, at least in these mature cochleae.

Rapid endocytosis and intracellular transport

We observed that vesicles trapped at the apical surface weretransported to Hensen’s body, to the basolateral wall and to thesubnuclear pole (Figs 2 and 3). The co-localized staining with FM1-43and DiOC6 (Fig. 4) and the micropipette aspiration experiments(Fig. 5) suggest that the SSC is the main location of endocytosedvesicles at the basolateral wall.

Kachar et al. (1997) have shown that two different types of vesiclesare endocytosed at the apical surface in bullfrog vestibular haircells ) electron microscopically they found both clathrin-coated andnon-clathrin-coated pits. Although clathrin-independent pathways arenot well characterized (Johannes & Lamaze, 2002), rapid endocytosis(in adrenal chromaffin cells) does not require clathrin activity (Artalejoet al., 1995). In general, vesicle formation with rapid endocytosisrequires less than 1 s, but hundreds of seconds for clathrin-dependentendocytosis (Henkel & Almers, 1996). Griesinger et al. (2004)showed a fast membrane and a slow protein internalization in OHCs,which are likely to derive from the clathrin-independent and clathrin-dependent pathways, respectively. Although we did not investigateclathrin dependence, the kinetics of our signals suggests that theendocytic activity is that of clathrin-independent, rapid endocytosis.

The observed transport speed of 0.36 ± 0.03 lm ⁄ s (n ¼ 15) isconsistent with values obtained from other studies for vesicle

movement along microtubules, which range from 0.02 to 2 lm ⁄ s(Nakata et al., 1998; Toomre et al., 1999; Kipp & Arias, 2002; Mundyet al., 2002), and shows similarities to the values presented byGriesinger et al. (2004) for signal delays between the apical andintracellular compartments. In general, microtubules are responsiblefor intracellular vesicle transport (Kachar et al., 1997). In OHCs,microtubules are found throughout the cell, but are especiallyconcentrated in the infracuticular region and the region around andbelow the nucleus (Furness et al., 1990). It is therefore likely that thetransport observed here was along the microtubules.

Calcium ⁄ calmodulin regulates rapid endocytosis

Ca2+ influx is generally believed to trigger rapid endocytosis (Neher &Zucker, 1993; Artalejo et al., 1995; Griesinger et al., 2002, 2004). Dyeuptake in the infracuticular zone was clearly accelerated withincreased extracellular Ca2+ concentration (Fig. 6), as also found byGriesinger et al. (2002, 2004, 2005) for hair cells of the guinea-pigcochlea. This contrasts to the situation for permeation of METchannels by FM1-43, which is independent of external Ca2+

concentration (below 10 mm) (Meyers et al., 2003). Although wedemonstrated here that the P2X receptor blocker PPADS had no effecton the endocytic activity (Fig. 1C), we can not rule out the possibilitythat in vivo Ca2+ influx through ATP-activated receptors can alsoinfluence rapid endocytosis. We also observed inhibition of rapidendocytosis by chelating intracellular Ca2+ (Fig. 7). In several celltypes, the ubiquitous Ca2+ receptor protein calmodulin has severalimportant roles. Unbound calmodulin can be activated on a millisec-ond time scale and Ca2+-binding is limited only by Ca2+ diffusion(Falke et al., 1994). In rapid endocytosis, calmodulin may act as aCa2+ receptor (Artalejo et al., 1996). Here we present evidence that thecalmodulin inhibitor, TFP, reduced rapid endocytic activity (Fig. 7).Although one should bear in mind that TFP might exert its influenceby blocking the MET channels (Seiler & Nicolson, 1999), we can ruleout this possibility in our experiments. First, in control experiments wedid not observe an effect of the open channel blocker DHSM onendocytosis (Fig. 1). Second, under voltage-clamp, the Ca2+-channel

Fig. 7. Decrease of FM1-43-induced fluorescence with Ca2+ chelation and calmodulin inhibition. Intracellular Ca2+ was chelated by incubating in BAPTA-AM(25 lm) and calmodulin was inhibited by incubating in trifluoperazine (TFP, 20 lm). (A) Time course of fluorescence, averaged for all cells.(B) Fluorescence values at 600 s. Fluorescence intensity was significantly less for both BAPTA-AM- and TFP-treated cells compared with control (control,n ¼ 6; BAPTA-AM, n ¼ 4; TFP, n ¼ 4). All fluorescence signals were taken from the Hensen’s body area.



blocker Cd2+ had no effect on the slopes of the fluorescence signal inthe apical region (Fig. 8C and E), suggesting that there was no Ca2+-permeable channel at the apical pole involved in the TFP-inducedinhibition.In conclusion, intracellular Ca2+ and calmodulin modulate rapid

endocytosis in OHCs. It has been shown in sensory hair cells of thezebrafish lateral line that rapid endocytosis is Ca2+ ⁄ calmodulindependent (Seiler & Nicolson, 1999). The present results are the firstdemonstration of Ca2+ ⁄ calmodulin dependence in mammalian OHCs.

Effect of membrane-potential changes on endocytosis

The reduced basal fluorescence upon increasing the whole-cellmembrane potential indicates that depolarization facilitates exocytosisof OHCs at the subnuclear pole, and that there it requires Ca2+ entrythrough voltage-gated Ca2+ channels (Fig. 8). In contrast to IHCs(Griesinger et al., 2002, 2005), in OHCs the rate of apical endocytosisappears not to be tightly regulated by synaptic exocytosis. Althoughdepolarization also decreased the slope of the fluorescence signal atthe apical pole, the same slope reduction was found when exocytosiswas blocked by Cd2+ (Fig. 8). Although the data show that the rate of

apical endocytosis is Ca2+ dependent (Figs 6 and 7), the Cd2+ data(Fig. 8), as well as the DHSM and PPADS data (Fig. 1), suggest thatCa2+ did not enter through the apical region of the cell.

Roles of rapid endocytosis and intracellular transport

A large variety of transcytic cargos have been examined in other cells(Tuma & Hubbard, 2003). Membrane proteins are known to have alimited lifetime and therefore need permanent replacement. Endocy-tosis and exocytosis are common mechanisms for the renewal ofmembrane proteins. Therefore, previous reports (Meyer et al., 2001;Griesinger et al., 2004) suggested that rapid, apical endocytosis andtrafficking to the basolateral wall in OHCs might be related to sorting,recycling and turnover of proteins required for the motor complex,because the motors are present in tremendously high density in thebasolateral wall (Huang & Santos-Sacchi, 1993), perhaps to theexclusion of channels not associated with the motor complex(Ashmore, 1992). Here, we observed endocytosed membrane at theapical pole being transported to Hensen’s body and to the lateral wall,mainly to the SSC (Figs 4 and 5). Hensen’s body and SSC are knownto be parts of the ER (Saito, 1983), which is important for proteinrecycling.Our finding of a fluorescent signal in the subnuclear region, albeit at

least three times weaker than in the SSC (Fig. 2), is certainlyconsistent with OHC synapses being located exclusively in that region(Saito, 1980). Ion channels, which are not directly associated with themotor complex, appear to be also located in that region. In IHCs, thereis convincing evidence that rapid endocytosis and transcytosis oftransmitter vesicles from the apex is important for maintainingadequate vesicle populations in the synaptic region (Griesinger et al.,2002, 2005). They have shown that exocytosis at the base upregulatesendocytic activity at the apex. However, our data do not provideevidence for such upregulation in OHCs: depolarization of the whole-cell membrane potential reduced the fluorescent signal in both thebasal and apical regions (Fig. 8). Nor did we find that the fluorescentvesicles were aggregated at the subnuclear pole, as was found in IHCs(Griesinger et al., 2002, 2005), but rather the signal appeared to befairly homogeneous in the subnuclear region (Fig. 2D). Thesedifferences between IHC and OHC findings are perhaps not surprisingbecause, unlike the IHC, the OHC is not dedicated to afferenttransmission. Only 5% of the afferent fibres innervate the OHCs, andthese fibres are small and unmyelinated, so that they are unlikely to berequired to support high firing rates (Spoendlin, 1986). The IHCs arethe true sensory cells of the cochlea, and their large, myelinatedafferents are required to support exceedingly high firing rates (Palmer& Russell, 1986). The speed of IHC afferent activity makes hugedemands on replenishment of vesicles (Moser & Beutner, 2000),which are satisfied by vesicle-generating compartments in the apex(Griesinger et al., 2002, 2005). In other words, because it is notexpected that OHC afferents are required to support high firing rates,apical stores of transmitter might not be necessary. Conversely, ourdata can not exclude the existence of such apical stores ) it might bethat the fluorescence signal from apical vesicles destined for thesubnuclear region was overshadowed by the sheer numbers of vesiclesdestined for the basolateral wall.As another possibility, consider that the neurotransmitters glutamate

and aspartate are not only found in hair cells (Altschuler et al., 1989;Usami & Ottersen, 1996), but have also been localized in other non-neural cells, including the mesothelial cells in Reissner’s membrane(Usami & Ottersen, 1996). Moreover, the glutamate-aspartatetransporter (GLAST) has been reported in the supporting Deiterscells (Furness et al., 2002). Also, the concentration of glutamate and

Fig. 8. Decrease of FM1-43-induced fluorescence upon whole-cell depolar-ization. FM1-43 was applied continuously from t ¼ 0 s, without washout. Thehorizontal line above the time axes, labelled ‘Depol.’, indicates the duration ofthe 0 mV resting membrane potential; otherwise it was )60 mV. Broken linesare the linear regression lines estimated for the interval of the depolarization.(A and C) Time courses of the infracuticular (apical) fluorescent signal incontrol and in the presence of CdCl2. (B and D) Time courses of the subnuclear(basal) fluorescent signal in control and in the presence of CdCl2. (A–D) Fromdifferent cells. (E) Regression slope for 0 mV, relative to )60 mV, forfluorescence in the apical area. (F) As for (E), but for the basal area. Asteriskmeans that the given slope was significantly and reversibly changed bydepolarization. Notice that 200 lm CdCl2 inhibited the effect of depolarizationin the basal region (D and F), but not in the apical region (C and E) of the cell.



aspartate is much higher in endolymph than in perilymph (Thalmannet al., 1981). In hair cells, glutamate and aspartate have been proposedto have important roles not only as neurotransmitters, but also asenergy sources (Erecinska & Silver, 1990; Usami & Ottersen, 1996;Ottersen et al., 1998). Perhaps apical endocytosis and transcytosissubserve such a glutamate or aspartate system and, in general, providea route for molecules from endolymph for metabolic or nutritionalrequirements.

In conclusion, the two distinct destinations of endocytosed mem-brane are consistent with the functional polarization of the OHC, withthe basolateral wall being dedicated to electromechanical transductionand the subnuclear pole being dedicated to electrochemical transduc-tion processes. Identifying transported molecules remains one of themost important goals for the future.

Acknowledgements

We thank J. Meyer and S. Preyer for early helpful suggestions, and N. Bayer,R. Lauf and A. Seeger for excellent technical assistance. This work wassupported by the Deutscher Akademischer Austausch Dienst (DAAD, T.K.)and the HBFG programme 127-545 (A.W.G.). Parts of this work werepresented at the 27th ARO Midwinter Research Meeting, Daytona Beach,Florida, USA (21–26 February 2004).

Abbreviations

BAPTA-AM, 1.2-bis-(O-aminophenoxy)-ethane-N,N,N¢,N¢-tetraacetic acid;DHSM, dihydrostreptomycin; DiOC6, 3,3¢-dihexyloxacarbocyanine iodide;EGTA, ethylene glycol-bis(b-aminoethyl ether)-N,N,N¢,N¢-tetraacetic acid; ER,endoplasmic reticulum; FM1-43, N-(3-triethylammoniumpropyl)-4-(4-(dibutyl-amino)styryl)pyridinium dibromide; HBSS, Hanks’ balanced salt solution; HE-PES, 2-[4-(2-hydroxyethyl)-1-piperazinyl]-ethanesulphonic acid; IHC, inner haircell; MET, mechanoelectrical transduction; NBD-cholesterol, 22-(N-(7-nitro-benz-2-oxa-1,3-diazol-4-yl)amino)-23,24-bisnor-5-cholen-3b-ol; OHC, outerhair cell; PPADS, 4-[[4-formyl-5-hydroxy-6-methyl-3-[(phosphonooxy)-methyl]-2-pyridinyl]azo]-1,3-benzenedisulphonic acid tetrasodium salt; ROI,region of interest; SSC, subsurface cisternae; TFP, trifluoperazine.

References

Altschuler, R.A., Sheridan, C.E., Horn, J.W. & Wenthold, R.J. (1989)Immunocytochemical localization of glutamate immunoreactivity in theguinea pig cochlea. Hear. Res., 42, 167–173.

Artalejo, C.R., Elhamdani, A. & Palfrey, H.C. (1996) Calmodulin is thedivalent cation receptor for rapid endocytosis, but not exocytosis, in adrenalchromaffin cells. Neuron, 16, 195–205.

Artalejo, C.R., Henley, J.R., McNiven, M.A. & Palfrey, H.C. (1995) Rapidendocytosis coupled to exocytosis in adrenal chromaffin cells involves Ca2+,GTP, and dynamin but not clathrin.Proc. Natl. Acad. Sci. USA, 92, 8328–8332.

Ashmore, J.F. (1987) A fast motile response in guinea-pig outer hair cells: thecellular basis of the cochlear amplifier. J. Physiol., 388, 323–347.

Ashmore, J.F. (1992) Mammalian hearing and the cellular mechanisms of thecochlear amplifier. Sensory Transduction. The Rockefeller University Press,New York, pp. 396–412.

Brownell, W.E., Bader, C.R., Bertrand, D. & de Ribaupierre, Y. (1985) Evokedmechanical responses of isolated cochlear outer hair cells. Science, 227, 194–196.

Chattopadhyay, A. (1990) Chemistry and biology of N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-labeled lipids: fluorescent probes of biological and modelmembranes. Chem. Phys. Lipids, 53, 1–15.

Cheatham, M.A., Huynh, K.H., Gao, J., Zuo, J. & Dallos, P. (2004) Cochlearfunction in Prestin knockout mice. J. Physiol., 560, 821–830.

Cook, W.J., Walter, L.J. & Walter, M.R. (1994) Drug binding by calmodulin:crystal structure of a calmodulin-trifluoperazine complex. Biochemistry, 33,15259–15265.

Dallos, P. & Evans, B.N. (1995) High-frequency motility of outer hair cells andthe cochlear amplifier. Science, 267, 2006–2009.

Dallos, P. & Harris, D. (1978) Properties of auditory nerve responses in absenceof outer hair cells. J. Neurophysiol., 41, 365–383.

Erecinska, M. & Silver, I.A. (1990) Metabolism and role of glutamate inmammalian brain. Prog. Neurobiol., 35, 245–296.

Falke, J.J., Drake, S.K., Hazard, A.L. & Peersen, O.B. (1994) Molecular tuningof ion binding to calcium signaling proteins. Q. Rev. Biophys., 27, 219–290.

Forge, A. & Richardson, G. (1993) Freeze fracture analysis of apicalmembranes in cochlear cultures: differences between basal and apical-coilouter hair cells and effects of neomycin. J. Neurocytol., 22, 854–867.

Frank, G., Hemmert, W. & Gummer, A.W. (1999) Limiting dynamics of high-frequency electromechanical transduction of outer hair cells. Proc. Natl.Acad. Sci. USA, 96, 4420–4425.

Furness, D.N., Hackney, C.M. & Steyger, P.S. (1990) Organization ofmicrotubules in cochlear hair cells. J. Electron Microsc. Techn., 15, 261–279.

Furness, D.N., Hulme, J.A., Lawton, D.M. & Hackney, C.M. (2002)Distribution of the glutamate ⁄ aspartate transporter GLAST in relation tothe afferent synapses of outer hair cells in the guinea pig cochlea. J. Assoc.Res. Otolaryngol., 3, 234–247.

Gale, J.E., Marcotti, W., Kennedy, H.J., Kros, C.J. & Richardson, G.P. (2001)FM1-43 dye behaves as a permeant blocker of the hair-cell mechano-transducer channel. J. Neurosci., 21, 7013–7025.

Geleoc, G.S.G. & Holt, J.R. (2003) Developmental acquisition of sensorytransduction in hair cells of the mouse inner ear. Nat. Neurosci., 6, 1019–1020.

Griesinger, C.B., Richards, C.D. & Ashmore, J.F. (2002) FM1-43 revealsmembrane recycling in adult inner hair cells of the mammalian cochlea.J. Neurosci., 22, 3939–3952.

Griesinger, C.B., Richards, C.D. & Ashmore, J.F. (2004) Apical endocytosis inouter hair cells of the mammalian cochlea. Eur. J. Neurosci., 20, 41–50.

Griesinger, C.B., Richards, C.D. & Ashmore, J.F. (2005) Fast vesiclereplenishment allows indefatigable signalling at the first auditory synapse.Nature, 435, 212–215.

Hacohen, N., Assad, J.A., Smith, W.J. & Corey, D.P. (1989) Regulation oftension on hair-cell transduction channels: displacement and calciumdependence. J. Neurosci., 9, 3988–3997.

Hasson, T., Gillespie, P.G., Garcia, J.A., MacDonald, R.B., Zhao, Y., Yee, A.G.,Mooseker, M.S. & Corey, D.P. (1997) Unconventional myosins in inner-earsensory epithelia. J. Cell Biol., 137, 1287–1307.

Henkel, A.W. & Almers, W. (1996) Fast steps in exocytosis and endocytosisstudied by capacitance measurements in endocrine cells. Curr. Opin.Neurobiol., 6, 350–357.

Huang, G. & Santos-Sacchi, J. (1993) Mapping the distribution of the outer haircell motility voltage sensor by electrical amputation. Biophys. J., 65, 2228–2236.

Ikeda, K. & Takasaka, T. (1993) Confocal laser microscopical images ofcalcium distribution and intracellular organelles in the outer hair cell isolatedfrom the guinea pig cochlea. Hear. Res., 66, 169–176.

Johannes, L. & Lamaze, C. (2002) Clathrin-dependent or not: is it still thequestion? Traffic, 3, 443–451.

Kachar, B., Battaglia, A. & Fex, J. (1997) Compartmentalized vesicular trafficaround the hair cell cuticular plate. Hear. Res., 107, 102–112.

Kalinec, F., Holley, M.C., Iwasa, K.H., Lim, D.J. & Kachar, B. (1992) Amembrane-based force generation mechanism in auditory sensory cells.Proc. Natl. Acad. Sci. USA, 89, 8671–8675.

Kimitsuki, T. & Ohmori, H. (1993) Dihydrostreptomycin modifies adaptationand blocks the mechano-electric transducer in chick cochlear hair cells.Brain Res., 624, 143–150.

Kipp, H. & Arias, I.M. (2002) Trafficking of canalicular ABC transporters inhepatocytes. Annu. Rev. Physiol., 64, 595–608.

Kroese, A.B., Das, A. & Hudspeth, A.J. (1989) Blockage of the transductionchannels of hair cells in the bullfrog’s sacculus by aminoglycosideantibiotics. Hear. Res., 37, 203–217.

Lansman, J.B., Hess, P. & Tsien, R.W. (1986) Blockade of current throughsingle calcium channels by Cd2+, Mg2+, and Ca2+. Voltage and concentrationdependence of calcium entry into the pore. J. Gen. Physiol., 88, 321–347.

Levin, R.M. & Weiss, B. (1978) Specificity of the binding of trifluoperazine tothe calcium-dependent activator of phosphodiesterase and to a series of othercalcium-binding proteins. Biochim. Biophys. Acta, 540, 197–204.

Liberman, M.C., Gao, J., He, D.Z.Z., Wu, X., Jia, S. & Zuo, J. (2002) Prestin isrequired for electromotility of the outer hair cell and for the cochlearamplifier. Nature, 419, 300–304.

Marcotti, W., van Netten, S.M. & Kros, C.J. (2005) The aminoglycosideantibiotic dihydrostreptomycin rapidly enters mouse outer hair cellsthrough the mechano-electrical transducer channels. J. Physiol., 567, 505–521.

McLaren, G.J., Lambrecht, G., Mutschler, E., Baumert, H.G., Sneddon, P. &Kennedy, C. (1994) Investigation of the actions of PPADS, a novel



P2x-purinoceptor antagonist, in the guinea-pig isolated vas deferens. Br. J.Pharmacol., 111, 913–917.

Meyer, J., Furness, D.N., Zenner, H.P., Hackney, C.M. & Gummer, A.W.(1998) Evidence for opening of hair-cell transducer channels after tip-linkloss. J. Neurosci., 18, 6748–6756.

Meyer, J., Mack, A.F. & Gummer, A.W. (2001) Pronounced infracuticularendocytosis in mammalian outer hair cells. Hear. Res., 161, 10–22.

Meyers, J.R.,MacDonald, R.B., Duggan, A., Lenzi, D., Standaert, D.G., Corwin,J.T. & Corey, D.P. (2003) Lighting up the senses: FM1-43 loading of sensorycells through nonselective ion channels. J. Neurosci., 23, 4054–4065.

Moser, T. & Beutner, D. (2000) Kinetics of exocytosis and endocytosis at thecochlear inner hair cell afferent synapse of the mouse. Proc. Natl. Acad. Sci.USA, 97, 883–888.

Mundy, D.I., Machleidt, T., Ying, Y.S., Anderson, R.G. & Bloom, G.S. (2002)Dual control of caveolar membrane traffic by microtubules and the actincytoskeleton. J. Cell Sci., 115, 4327–4339.

Nakata, T., Terada, S. & Hirokawa, N. (1998) Visualization of the dynamics ofsynaptic vesicle and plasma membrane proteins in living axons. J. Cell Biol.,140, 659–674.

Neher, E. & Zucker, R.S. (1993) Multiple calcium-dependent processes relatedto secretion in bovine chromaffin cells. Neuron, 10, 21–30.

Nishikawa, S. & Sasaki, F. (1996) Internalization of styryl dye FM1-43 in thehair cells of lateral line organs in Xenopus larvae. J. Histochem. Cytochem.,44, 733–741.

Oghalai, J.S., Tran, T.D., Raphael, R.M., Nakagawa, T. & Brownell, W.E.(1999) Transverse and lateral mobility in outer hair cell lateral wallmembranes. Hear. Res., 135, 19–28.

Ottersen, O.P., Takumi, Y., Matsubara, A., Landsend, A.S., Laake, J.H. &Usami, S. (1998) Molecular organization of a type of peripheral glutamatesynapse: the afferent synapses of hair cells in the inner ear. Prog. Neurobiol.,54, 127–148.

Palmer, A.R. & Russell, I.J. (1986) Phase-locking in the cochlear nerve of theguinea-pig and its relation to the receptor potential of inner hair-cells. Hear.Res., 24, 1–15.

Raybould, N.P. & Housley, G.D. (1997) Variation in expression of the outerhair cell P2X receptor conductance along the guinea-pig cochlea. J. Physiol.,498, 717–727.

Ricci, A. (2002) Differences in mechano-transducer channel kinetics underlietonotopic distribution of fast adaptation in auditory hair cells. J. Neurophys-iol., 87, 1738–1748.

Richardson, G.P., Forge, A., Kros, C.J., Fleming, J., Brown, S.D. & Steel, K.P.(1997) Myosin VIIA is required for aminoglycoside accumulation incochlear hair cells. J. Neurosci., 17, 9506–9519.

Saito, K. (1980) Fine structure of the sensory epithelium of the guinea pigorgan of Corti: afferent and efferent synapses of hair cells. J. Ultrastruct.Res., 71, 222–232.

Saito, K. (1983) Fine structure of the sensory epithelium of guinea-pig organ ofCorti: subsurface cisternae and lamellar bodies in the outer hair cells. CellTissue Res., 229, 467–481.

Schulte, C.C., Meyer, J., Furness, D.N., Hackney, C.M., Kleyman, T.R. &Gummer, A.W. (2002) Functional effects of a monoclonal antibody onmechanoelectrical transduction in outer hair cells. Hear. Res., 164, 190–205.

Seiler, C. & Nicolson, T. (1999) Defective calmodulin-dependent rapid apicalendocytosis in zebrafish sensory hair cell mutants. J. Neurobiol., 41, 424–434.

Si, F., Brodie, H., Gillespie, P.G., Vazquez, A.E. & Yamoah, E.N. (2003)Developmental assembly of transduction apparatus in chick basilar papilla.J. Neurosci., 23, 10815–10826.

Sit, P.S., Spector, A.A., Lue, A.J., Popel, A.S. & Brownell, W.E. (1997)Micropipette aspiration on the outer hair cell lateral wall. Biophys. J., 72,2812–2819.

Spoendlin, H. (1986) Receptoneural and innervation aspects of the inner earanatomy with respect to cochlear mechanics. Scand. Audiol. Suppl., 25,27–34.

Terasaki, M., Song, J., Wong, J.R., Weiss, M.J. & Chen, L.B. (1984)Localization of endoplasmic reticulum in living and glutaraldehyde-fixedcells with fluorescent dyes. Cell, 38, 101–108.

Thalmann, R., Comegys, T.H., DeMott, J.E. & Thalmann, I. (1981) Steepgradients of amino acids between cochlear endolymph and perilymph.Laryngoscope, 91, 1785–1791.

Toomre, D., Keller, P., White, J., Olivo, J.C. & Simons, K. (1999) Dual-colorvisualization of trans-Golgi network to plasma membrane traffic alongmicrotubules in living cells. J. Cell Sci., 112, 21–33.

Tuma, P.L. & Hubbard, A.L. (2003) Transcytosis: crossing cellular barriers.Physiol. Rev., 83, 871–932.

Usami, S. & Ottersen, O.P. (1996) Aspartate is enriched in sensory cells andsubpopulations of non-neuronal cells in the guinea pig inner ear:a quantitative immunoelectron microscopic analysis. Brain Res., 742,43–49.

Vandonselaar, M., Hickie, R.A., Quail, J.W. & Delbaere, L.T. (1994)Trifluoperazine-induced conformational change in Ca2+-calmodulin. Nat.Struct. Biol., 1, 795–801.

Zheng, J., Shen, W., He, D.Z., Long, K.B., Madison, L.D. & Dallos, P. (2000)Prestin is the motor protein of cochlear outer hair cells. Nature, 405, 149–155.



Documents

Membrane traffic in outer hair cells of the adult mammalian cochlea