|Home | About | Journals | Submit | Contact Us | Français|
Sensorineural hearing loss, which stems primarily from the failure of mechanosensory hair cells, changes the traveling waves that transmit acoustic signals along the cochlea. However, the connection between cochlear mechanics and the amplificatory function of hair cells remains unclear. Using an optical technique that permits the targeted inactivation of prestin, a protein of outer hair cells that generates forces on the basilar membrane, we demonstrate that these forces interact locally with cochlear traveling waves to achieve enormous mechanical amplification. By perturbing amplification in narrow segments of the basilar membrane, we further show that a cochlear traveling wave accumulates gain as it approaches its peak. Analysis of these results indicates that cochlear amplification produces negative damping that counters the viscous drag impeding traveling waves; targeted photoinactivation locally interrupts this compensation. These results reveal the locus of amplification in cochlear traveling waves and connect the characteristics of normal hearing to molecular forces.
Human hearing is extraordinarily sensitive and discriminating. We can hear sounds down to the level of thermal fluctuations in the ear. Our ability to detect subtle differences in tones over a frequency span of three decades allows us to distinguish human voices of nearly identical timbre. We additionally perceive sounds of vastly differing intensities, enabling us to discern the strumming of nylon strings on a classical guitar playing in concert with a full orchestra. It is remarkable that the ear can achieve such sensitivity despite the viscous damping that impedes the oscillation of structures within the cochlea. Indeed, the frequency resolution of human hearing inferred from psychophysics is too great to be explained by passive resonance (Gold, 1948). Measurements of amplified, compressive vibrations within the cochlea (Rhode, 1971; Le Page and Johnstone, 1980) as well as the discovery that healthy ears produce sounds (Kemp, 1978)—so-called otoacoustic emissions—have established that the inner ear possesses an active amplification mechanism.
The qualities of active hearing can be observed in the inner ear's mechanical response to sound. A pure-tone stimulus elicits a traveling wave along the cochlear partition (von Békésy, 1960), a flexible complex of membranes that divides the spiral cochlea into three fluid-filled chambers. Increasing in amplitude as it propagates, the traveling wave peaks at a characteristic place for each specific frequency of stimulation, thereby delivering most of its energy to a select population of mechanosensory hair cells. This frequency-dependent wave profile is brought about by an interplay between longitudinal fluid coupling and local displacement of the cochlear partition, whose stiffness and mass are graded (Lighthill, 1981). In a normal ear, an active process in outer hair cells amplifies and sharpens the traveling wave, thereby fostering the remarkable frequency resolution and dynamic range that characterize healthy hearing (Rhode, 1971; Le Page and Johnstone, 1980; Sellick et al., 1982). The traveling wave of a compromised cochlea, in contrast, is diminished and broadened.
Where along the cochlear partition do active forces impart mechanical energy? A passive traveling wave conveys energy up to a resonant position that is dictated by the cochlear partition's gradient of mass and stiffness. Outer hair cells can locally inject energy that is thought to counter viscous damping and thus to augment the vibration of each segment of the partition. Because the resulting active wave can then accumulate gain by traversing the region in which amplification occurs, the cumulative gain at the wave's peak, or the integral of gain as a function of distance, is thought to dramatically exceed the local gain provided by outer hair cells (de Boer, 1983; Reichenbach and Hudspeth, 2010). Although a logical way of testing this hypothesis would be to inactivate amplification at specific positions basal to a traveling wave's peak, this has heretofore been possible only by focal ablation of hair cells (Cody, 1992). This approach reduces amplification, but at the cost of significantly altering the passive mechanical properties that transmit energy to the characteristic place.
Selectively perturbing amplification requires an understanding of the underlying active process in outer hair cells. Experiments involving isolated hair cells have identified two force-generating mechanisms. The mechanoreceptive hair bundles of many tetrapods are capable of generating forces that can be entrained by an external stimulus (Martin and Hudspeth, 1999; Kennedy et al., 2003, 2005). These forces have been observed in the form of spontaneous hair-bundle oscillations and as negative stiffness that can increase a bundle's response to low-amplitude mechanical stimulation (Martin et al., 2000, 2003). Active hair-bundle motility also contributes to nonlinear amplification in an in vitro preparation of the mammalian cochlea (Chan and Hudspeth, 2005). Another force-generating mechanism specific to the outer hair cell of mammals is somatic motility or electromotility: changes in membrane potential rapidly alter the cylindrical cell's length (Brownell et al., 1985). This behavior is mediated by voltage-dependent conformational changes in the membrane protein prestin (Zheng et al., 2000), which is expressed at high levels in the basolateral plasmalemma (Huang and Santos-Sacchi, 1993). An extensive body of research on both isolated hair cells and mammalian cochleas in vivo has demonstrated the importance of functional prestin in healthy hearing (Ashmore, 2008). Mutant mice that lack prestin display substantially degraded hearing and auditory tuning curves that resemble those of a damaged cochlea (Cheatham et al., 2004). A knockin mutant mouse that expresses immotile prestin displays knockout-like hearing thresholds (Dallos et al., 2008). Because the mutant's outer hair cells appear to maintain stiffnesses similar to those in a wild-type mouse, the passive mechanical properties of the mutant's cochlea are unlikely to be significantly altered.
Although these studies demonstrate that somatic motility plays a critical role in cochlear amplification, prestin's role in shaping the cochlear traveling wave remains unclear. The interpretation of mutant studies is complicated by the fact that these modifications affect hair cells throughout the cochlea. Cochlear amplification is highly tuned; even subtle changes in the cochlea's passive properties might therefore interfere with the transmission of an acoustic signal to its characteristic place of maximal amplification. Furthermore, prestin serves as a transporter of chloride (Bai et al., 2009) and sugars (Chambard and Ashmore, 2003) and may play a significant role in maintaining physiological homeostasis in hair cells. Mutations of prestin might compromise these transporter functions.
Although a wealth of evidence suggests that hair-cell forces influence movement of the cochlear partition and vice versa, it has not been possible heretofore to locally separate the contributions of active hair-bundle motility and somatic motility. Because it affects the membrane potential, hair-bundle motility regulates somatic motility, confounding the interpretation of experiments that perturb the former mechanism. In contrast, inactivating prestin should not significantly affect hair-bundle motility. We therefore developed an optical technique that permits the targeted inactivation of somatic motility without significantly altering passive power transmission.
Active traveling waves, which are characterized by peaks that are amplified and grow nonlinearly with increasing stimulation, are notoriously difficult to visualize in vivo because surgery readily disrupts the delicate cochlea (Ren, 2002; Ren et al., 2011). By applying a scanned-beam laser interferometer in the chinchilla's cochlea, we recorded the two-dimensional profiles of the traveling waves elicited by pure tones (Figure S1A available online; Movie S1; Supplemental Experimental Procedures, Section 1). Because the chinchilla's hearing, like that of humans, is most sensitive to frequencies of a few kilohertz (Figure S1B; Heffner and Heffner, 1991), we were able to study traveling waves at frequencies relevant to human hearing.
We measured sensitive and highly compressive traveling waves on the basilar membrane, the lower surface of the cochlear partition, through a 1 mm window in the cochlea's bony wall about 2 mm apical to the round window. The basilar membrane at this position was maximally sensitive to acoustic stimulation at about 9 kHz (Figures 1A, 1B, and S1C). As the sound-pressure level (SPL) increased from about 40 dB to over 90 dB, the responses grew less steeply, a manifestation of compressive nonlinearity (Figures 1C and 1D). The peaks of active traveling waves also shifted basally as the stimulus level increased (Figures S1D and S1E). The high gain and nonlinearity were completely abolished when the active process was interrupted by anoxia (Figure 1E), which additionally displaced the wave's peak toward the cochlear base. The phase profiles of traveling waves displayed slopes that were dependent on stimulus level in healthy cochleas but not following anoxia (Figure S1F). These phenomena reflect the loss after anoxia of a tuned, tonotopically distributed amplification mechanism that enhances a traveling wave as it approaches the characteristic place at which it peaks.
To investigate the interplay between active cellular forces and the spatial shaping of an active traveling wave, we developed an optical technique that locally and significantly perturbs electromotility. Small carboxylic acids inhibit prestin-based motility; salicylate is the most effective of these blockers (Tunstall et al., 1995; Oliver et al., 2001). Our technique uses 4-azidosalicylate, the azide group of which forms covalent bonds upon activation by ultraviolet (UV) light (Figures 2A and 2B). The compound is therefore an inhibitor that forms an irreversible complex with prestin, effectively disabling it. We initially characterized the effect of 4-azidosalicylate on somatic motility in HEK293T cells transfected with prestin-eGFP. Motility was deduced from measurements of a cell's voltage-dependent capacitance, which reflects the gating currents that accompany conformational changes in large ensembles of prestin molecules. Capacitance was measured from phase changes in the currents elicited by sinusoidal membrane-potential perturbations at different holding potentials (Fidler and Fernandez, 1989). When washed onto prestin-transfected HEK293T cells, 4-azidosalicylate largely abolished somatic motility as inferred from the linearization of the voltage-capacitance relation, an effect that was reversible upon washout (Figure 2C). UV irradiation had no effect on motility in control medium (Figure 2D). If a cell incubated in 4-azidosalicylate was exposed to UV light, however, motility did not return after washout (Figures 2E and 2F). The cell nonetheless remained healthy as assessed by visual appearance and by the absence of leakage currents.
Because the nonlinear capacitance measured in prestin-expressing cells cannot be dissociated from mechanical motility (Santos-Sacchi, 1991), photoinactivation presumably elicits a concurrent attenuation of the latter. We nonetheless confirmed that photoinactivation affects the somatic motility of isolated outer hair cells. As observed in our capacitance measurements, electromotility was permanently inactivated by the combination of exposure to 4-azidosalicylate and UV irradiation, but not by either procedure alone (Figures 2G and 2I; Supplemental Experimental Procedures, Section 2). We obtained similar results from experiments on the chinchilla's cochlea in vivo (Figure 3): although UV irradiation alone did not perturb the traveling wave, 4-azidosalicylate diminished the basilar membrane's movement reversibly and irradiation in the drug's presence produced a permanent deficit.
Salicylate interacts directly with prestin; the irreversible blockage of somatic motility therefore presumably reflects the covalent binding of 4-azidosalicylate to a binding site. To obtain evidence for such a direct interaction, we immunoprecipitated prestin from prestin-transfected HEK293T cells that had been incubated in 4-azidosalicylate and irradiated with UV light. Using tandem mass spectrometry, we confirmed that the final eluate contained prestin. Compared with a control sample, the prestin precipitated from photolyzed cells was predominantly oligomeric, which suggests that 4-azidosalicylate facilitates interactions between prestin protomers (Figure S2; Supplemental Experimental Procedures, Section 3). We surmise that washing 4-azidosalicylate into the scala tympani temporarily blocks motility in a large number of outer hair cells; after targeted photoinactivation and washout of the free compound, all the cells recover motility except for those that have been irradiated.
We used focal photoinactivation to probe the region at which gain occurs in active traveling waves. To guide our experiments, we computed a spatial map of cochlear-partition impedance based on measurements of active traveling waves. The local impedance Z(x,ω) at a distance x from the cochlear base describes how a segment of the partition responds to a periodic pressure difference across it. Acoustic stimulation at an angular frequency ω produces an oscillating pressure difference
in which c.c. denotes the complex conjugate. In response, the basilar membrane oscillates at the same frequency,
The Fourier coefficient (x, ω) follows from the pressure amplitude (x, ω) through the local impedance:
in which A(x) denotes the area of a thin radial strip of the basilar membrane.
The partition's local impedance can be represented as Z(x, ω) = ξ(x) + i[ωm(x) – k(x)/ω], with a local mass m(x), drag coefficient ξ(x), and stiffness k(x). The real part of the impedance therefore represents viscous damping; it is positive when viscous force impedes the partition's vibration, whereas a negative value signifies an active force that augments vibration and hence produces gain. The imaginary part of the impedance reflects stiffness, which makes a negative contribution, and inertia, whose influence has a positive sign.
We devised a mathematical technique for computing the basilar-membrane impedance, and therefore gain, based on our traveling-wave measurements. To this end, we have employed a model that depicts the cochlea as two fluid-filled chambers that are separated by a partition of graded impedance. The fluid pressure can additionally vary with both the length and the height of the cochlea's chambers (Reichenbach and Hudspeth, 2010). The Wentzel-Kramers-Brillouin approximation yields an estimate of the velocity profile (x, ω) that follows from a spatially varying impedance Z(x, ω) (Steele and Taber, 1979; Reichenbach and Hudspeth, 2010). This approximation can conversely be used to compute the local impedance from a measured velocity profile (Figure S3; Supplemental Experimental Procedures, Section 4).
Applied to velocity measurements of active, nonlinear traveling waves, this technique revealed a region of negative damping basal to the stimulus frequency's characteristic place. In contrast, damping was everywhere positive for measurements from anoxic preparations (Figures S3A and S3B). The presence of negative damping and the spatial profile of the calculated impedance support earlier theoretical predictions (de Boer, 1983). The imaginary components of the impedance were negative for all measured waves (Figure S3C), an indication that the effect of stiffness dominated that of inertia. Furthermore, the predicted values for stiffness, 1.5–3.5 N·m–1, were similar to those measured by using compliant fibers to induce point deflections (Olson and Mountain, 1991).
Informed by the locus of amplification provided by the impedance analysis, we next sought to determine the contribution of somatic motility to local amplification. A 500 μm-long segment of the cochlear partition that extended roughly one cycle basal from a wave's peak, depending on the location of the hole, encompassed most of the expected region of gain. Photoinactivating prestin over this broad segment reduced the sensitivity dramatically throughout the traveling wave (Figure 4A). This result was confirmed in six additional experiments; the average sensitivity along a 50 μm segment at the traveling wave's peak fell to 8% ± 2% (mean ± SEM) of the control level. That amplification was largely eliminated by irradiation encompassing a full cycle basal to the traveling wave's peak accords with indications from studies of noise damage and compressive nonlinearity that amplification occurs primarily within a region 1–2 mm before the wave's peak (Cody, 1992). Photoinactivation significantly attenuated the local gain—the amount of gain accrued per unit length along the basilar membrane—near the wave's characteristic place (Figure S3E). Photoinactivation additionally altered the frequency tuning of the cochlear partition; after irradiation, the characteristic place for the same stimulus frequency shifted basally (Figure 4A).
Two-dimensional maps of the traveling wave in a control cochlea revealed a lag in the phase of the basilar membrane approximately beneath the outer hair cells relative to that near the spiral lamina or spiral ligament (Figure S4). The phase lag diminished during intense stimulation and vanished in the presence of 4-azidosalicylate or after anoxia. After photoinactivation, the phase pattern was radially homogeneous. Although the presence and implication of a radially varying phase profile remain controversial (Nilsen and Russell, 1999, 2000; Rhode and Recio, 2000; Homer et al., 2004), this result provides further evidence that our technique diminished the cellular forces underlying the active process.
Having established that photoinactivation of somatic motility dramatically reduces local amplification in the cochlea, we next used this tool to gauge the spatial extent of amplification and to observe how focal perturbation affects the accumulation of gain. We probed two narrow segments that extended roughly 50 μm along the cochlear partition: one region lying a full cycle basal to the traveling wave's peak, and another situated just an eighth of a cycle before the peak. Inactivation of the more basal segment elicited a more gradual accumulation of gain; this caused a small decrement in gain that persisted, but did not increase, up to the wave's peak (Figure 4B). The modification did not significantly shift the wave's peak, suggesting that the inactivated segment lay near the beginning of the region of active amplification. This effect was confirmed in two additional experiments; the average sensitivity at the wave's peak remained 79% ± 12% of the control value.
Perturbation in a narrow segment near the active wave's peak, in contrast, significantly reshaped the wave, indicating that local amplification is spatially nonuniform and increases near the peak (Figure 4C). In this instance, the traveling wave initially accumulated gain at a rate similar to that under control conditions. The cumulative gain ceased to grow in the inactivated region, over which some viscous loss was evident. Finally, gain began to accumulate again just beyond the affected region. As before, the accumulation of local gain was abolished only in the segment of photoinactivation. This effect was confirmed in three additional experiments; the average sensitivity at the wave's peak was reduced to 18% ± 4% of the control value. After washout of 4-azidosalicylate, there were occasionally slight offset changes in the overall sensitivity, but these were not consistent (Figures 4B and 4C). However, the elimination of local gain—the slope of the cumulative gain as a function of position—occurred consistently in the photoinactivated region. In addition, focal perturbation in narrow regions locally eliminated the radial phase lag at the outer hair cells (Figure S4).
Impedance reconstructions based on these experiments indicate that inactivating the active process locally reduced negative damping and thus reveal the extent of the intrinsic positive damping by viscous forces (Figure 5). The imaginary part of the impedance did not change significantly after photoinactivation of somatic motility, suggesting that the active process has little influence on the stiffness and mass of the cochlear partition (Figure S3). This finding is consistent with observations that the intact cochlear partition is roughly three orders of magnitude stiffer than an individual outer hair cell (He and Dallos, 1999; Olson and Mountain, 1991). A portion of the active process remained after photoinactivation of regions in which gain occurred, however, for damping increased even further after anoxia (Figures 4A and 4B). Although the residual active process might reflect incomplete blockage of prestin, we ascertained that repeated exposure to UV light in the presence of 4-azidosalicylate did not further diminish the response (Figure 3D).
Understanding the degree to which photoinactivation eliminates electromotility in vivo could yield a more quantitative assessment of prestin's contribution to amplification. Obtaining a clearer picture of the intact active process would also necessitate an appreciation of the specificity with which photoinactivation affects electromotility; are other cellular processes affected? If, for example, active hair-bundle motility were entirely spared the effects of photoinactivation, then bundle forces might account for the balance of amplification. It is possible, however, that photoinactivating prestin affected hair-bundle forces as well. Changes in an outer hair cell's membrane potential can elicit hair-bundle deflections (Jia and He, 2005). Moreover, hair bundles can be displaced by somatic length changes of outer hair cells through the mechanical coupling in an intact organ of Corti. Although adaptation in hair bundles restores the set point of nonlinear amplification even for large static deflections (Martin et al., 2003), photoinactivation might force all prestin molecules into a conformation normally elicited only by extreme depolarization. In this circumstance, active hair-bundle motility could be compromised. Finally, salicylate might affect other aspects of hair-cell physiology. Although we failed to detect photolabeling of other proteins during our biochemical investigation, it remains possible that photoinactivation modifies proteins in addition to prestin.
These results demonstrate that an active process overcomes viscous damping to locally amplify the cochlear traveling wave and that this locally accrued gain accumulates spatially up to the wave's peak. The results further indicate that prestin plays a crucial role in establishing this gain. It remains an open question, however, how this active process is locally tuned to yield a tonotopic map of amplification. Despite its critical contribution, there is no evidence yet that somatic motility exhibits resonance. Under physiological conditions, sound-evoked receptor potentials in mammalian outer hair cells modulate a resting potential of –40 mV by less than 10 mV at moderate stimulus levels (Johnson et al., 2011; Kössl and Russell, 1992). Prestin's voltage-length relationship is almost linear in this region (Santos-Sacchi, 1991) and thus cannot alone explain nonlinear amplification. Active hair-bundle motility, in contrast, can be highly tuned (Martin and Hudspeth, 2001) and may account for the frequency selectivity and nonlinearity associated with amplification (O Maoiléidigh and Jülicher, 2010). In vivo experiments that selectively interfere with active hair-bundle motility while leaving transduction currents unperturbed might resolve this issue.
Human embryonic kidney (HEK) 293T cells were cultured at 37°C in humidified air containing 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin (Invitrogen). The cells were transfected (Lipofectamine 2000, Invitrogen) according to the manufacturer's protocol with pEGFP-N2-prestin (Zheng et al., 2000). Fusion of GFP to either the amino or the carboxy terminus of prestin does not affect prestin's function (Ludwig et al., 2001). Cells were harvested after 24 hr of incubation.
The extracellular saline solution for electrophysiological recordings comprised 120 mM NaCl, 20 mM tetraethylammonium chloride, 2 mM MgCl2, 10 mM HEPES, and 5 mM D-glucose. The internal solution with which tight-seal pipettes were filled included 135 mM KCl, 3.5 mM MgCl2, 0.1 mM CaCl2, 5 mM K2EGTA, 2.5 mM Na2ATP, and 5 mM HEPES. Both solutions were adjusted to an osmolality of 300 mOsmol·kg–1 and a pH of 7.3. In experiments that involved isolated outer hair cells, the extracellular solution was supplemented with 2 mM CoCl2 to eliminate voltage-dependent ionic conductances. Solution containing 4-azidosalicylate was added to the recording chamber at a rate of 0.5–1 ml/min through a gravity-feed perfusion system controlled by a solenoid-gated pinch valve (VC-66MCS, Warner Instruments).
Whole-cell voltage-clamp recording was conducted at room temperature with borosilicate-glass microelectrodes 2–3 MΩ in resistance when filled with internal solution. Nonlinear capacitance was measured by the phase-tracking technique, which involves analysis of the phase of the current elicited by a high-frequency sinusoidal command voltage (Fidler and Fernandez, 1989). The holding potential was sinusoidally modulated at 2.6 kHz with an amplitude of 5 mV. The series resistance and phase angle at which the current was most sensitive to capacitance changes were identified by dithering the series resistance by 500 kΩ (DR-1, Axon Instruments). The proportionality between phase change and capacitance was obtained through dithering by 100 fF the capacitance compensation of the amplifier (Axopatch 200B, Axon Instruments). Electrophysiological measurements were sampled at 12 μs intervals and analyzed with MATLAB.
HEK293T cells transfected to express prestin-eGFP were incubated with 4-azidosalicylate and exposed to UV light. Prestin-eGFP was immunoprecipitated with agarose beads coated with anti-GFP and resolved by electrophoresis through a linear-gradient polyacrylamide gel. Following in-gel digestion with trypsin, the peptide extracts were analyzed by liquid chromatography-tandem mass spectrometry. Peptide identification and analysis of modified residues were conducted with the Mascot algorithm (Matrix Science).
Animal procedures were performed with approval from the Institutional Animal Care and Use Committee at The Rockefeller University. In each experiment, a chinchilla (Chinchilla lanigera) weighing 300–500 g was anesthetized with intraperitoneally injected ketamine hydrochloride (30 mg/kg) and xylazine hydrochloride (5 mg/kg). The animal's body temperature was maintained at 37°C with a homeothermic heating pad (Stoelting). The trachea and neck musculature were exposed and a tracheotomy was performed. The pinna was then removed, the bulla opened widely through lateral and ventral approaches, and the tendons of the middle-ear muscles sectioned. A 500–700 μm hole was drilled in the basal turn of the otic capsule 1–2 mm apical to the round window, exposing a segment of the basilar membrane and permitting access for the probe beam of a laser interferometer. Through the tip of a 30G needle placed next to the hole, the scala tympani was perfused with artificial perilymph consisting of 137 mM NaCl, 5 mM KCl, 12 mM NaHCO3, 2 mM CaCl2, 1 mM MgCl2, 1 mM NaH2PO4, and 11 mM D-glucose. The solution was added at a rate of 0.5 ml/min for 1–2 min.
Two-dimensional profiles of traveling waves were measured by serially scanning the beam of a heterodyne Doppler interferometer (OFV-501, Polytec) over the basilar membrane and reconstructing the spatial patterns of vibration through analysis of each scan point's complex Fourier coefficient at the stimulus frequency. No beads or other reflective elements were deposited on the basilar membrane. Heterodyne interferometric measurements of poorly reflective surfaces such as the basilar membrane are sometimes contaminated by signals from deeper surfaces of the cochlear partition (de La Rochefoucauld et al., 2005). Although the use of reflective beads can increase the signal-to-noise ratio of vibration measurements of the basilar membrane, we found that depositing beads on the basilar membrane resulted in severe spatial inhomogeneities. Because our experiments required smooth, two-dimensional measurements of a traveling wave, we avoided the use of beads. The only two other studies that have reported two-dimensional measurements of traveling waves in vivo have similarly omitted beads (Ren, 2002; Ren et al., 2011). It is nonetheless possible that these surface measurements are contaminated by internal modes of motion within the cochlear partition, an effect that could obscure the exact range and magnitude of local amplification.
Pure-tone stimuli were delivered by a calibrated sound source and the measurements were phase-locked to the stimulus waveform. Examining data in the time domain revealed no significant low-frequency modulation onto which high-frequency vibrations were superimposed. This criterion ensured that the active process was not unintentionally removed from the range of small displacements over which nonlinear amplification is significant.
Interferometric measurement and photoinactivation were performed with a custom-built optical apparatus that consisted of an upright fluorescence microscope (BX51WI, Olympus) into the trinocular port of which were directed both the probe laser beam from the interferometer and the beam of a helium-cadmium laser operating at 325 nm (IK3202R-D, Kimmon Electrical).
We locally photoinactivated electromotility in vivo by scanning the beam of the 325 nm UV laser over select segments of the basilar membrane. Because the beam was loosely focused to a diameter of 10 μm, we were able to photolyze large areas at single-cell resolution by irradiating a relatively coarse grid of scan points. A custom program (LabVIEW, National Instruments) was used to define a photolysis region and control the relevant devices. After a polygonal region was selected for photolysis on the basis of a background image of the basilar membrane, an electronic shutter (VS25S2T0-10, UniBlitz) opened long enough to permit the galvanometric mirrors to scan the UV laser beam over points on a Cartesian grid.
We thank B. Fabella for technical assistance; M. Vologodskaia for assistance in molecular-biological techniques; Y. Castellanos and L. Kowalik for assistance with transfection and mammalian cell culture; D. Z.-Z. He, S. Jia, and X. Tan for training on electrophysiological measurements from outer hair cells; T. Ren for discussions of traveling-wave preparations; J. Ashmore, N. Cooper, R. Fettiplace, D. Navaratnam, and M. Ruggero for comments on the experimental approach; S. Ye for discussions of azide photochemistry; N. Chandramouli for comments on photoaffinity labeling; C. Bergevin and E. Olson for discussions of sound calibration; K. Leitch for assistance with illustrations; and members of our research group for comments on the manuscript. This investigation was supported by a Bristol-Myers Squibb Postdoctoral Fellowship in Basic Neurosciences and a research grant from the American Hearing Research Foundation (to J.A.N.F.), a Career Award at the Scientific Interface from the Burroughs Wellcome Fund (to T.R.), and a Postdoctoral Fellowship for Research Abroad from the Japan Society for the Promotion of Science (to F.N.). A.J.H. is an Investigator of Howard Hughes Medical Institute.
Supplemental Information includes four figures, one movie, and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10.1016/j.neuron.2012.09.031.