Electrical and Immunohistochemical Properties of Cochlear Fibrocytes in 3D Cell Culture and in the Excised Spiral Ligament of Mice.

Fibrocyte degeneration in the cochlear lateral wall is one possible pathology of age-related metabolic hearing loss (presbycusis). Within the lateral wall fibrocytes play a role in potassium recycling and maintenance of the endocochlear potential. It has been proposed that cell replacement therapy could prevent fibrocyte degeneration in the CD/1 mouse model of hearing loss. For this to work, the replacement fibrocytes would need to take over the structural and physiological role of those lost. We have grown lateral wall fibrocytes from neonatal CD/1 mice in a 3D-collagen gel culture with the aim of assessing their functional similarity to native lateral wall fibrocytes, the latter in a slice preparation and in excised spiral ligament pieces. We have compared cultured and native fibrocytes using both immuno-labelling of characteristic proteins and single cell electrophysiology. Cultured fibrocytes exhibited rounded cell bodies with extending processes. They labelled with marker antibodies targeting aquaporin 1 and calcium-binding protein S-100, precluding an unambiguous identification of fibrocyte type. In whole-cell voltage clamp, both native and cultured fibrocytes exhibited non-specific currents and voltage-dependent K+ currents. The non-specific currents from gel-cultured and excised spiral ligament fibrocytes were partially and reversibly blocked by external TEA (10 mM). The TEA-sensitive current had a mean reversal potential of + 26 mV, suggesting a permeability sequence of Na+ > K+. These findings indicate that 3D-cultured fibrocytes share a number of characteristics with native spiral ligament fibrocytes and thus might represent a suitable population for transplantation therapy aimed at treating age-related hearing loss.

Auditory hair cell-afferent fiber synapses are specialized to operate at their best frequencies.

Auditory afferent fiber activity is driven by high-fidelity information transfer from the sensory hair cell. Presynaptic specializations, posited to maintain fidelity, are investigated at synapses with characteristic frequencies of 120 Hz and 320 Hz. Morphological data indicate that high-frequency cells have more synapses and higher vesicle density near dense bodies (DBs). Tracking vesicular release via capacitance changes identified three overlapping kinetic components of release corresponding to morphologically identified vesicle pools. High-frequency cells released faster; however, when normalized to release site number, low-frequency cells released faster, likely due to a greater Ca2+ load per synapse. The Ca(2+)-dependence of release was nonsaturating and independent of frequency, suggesting that release, not refilling, was rate limiting. A model of release derived from vesicle equilibration between morphologically defined pools reproduced the capacitance data, supporting a critical role in vesicle trafficking for DBs. The model suggests that presynaptic specializations enable synapses to operate most efficiently at their characteristic frequencies.

The dimensions and structural attachments of tip links in mammalian cochlear hair cells and the effects of exposure to different levels of extracellular calcium.

The tip links between stereocilia of acousticolateral hair cells have been suggested to contain cadherin 23 (CDH23) comprising an upper branched portion that is bound to a lower portion composed of protocadherin 15 (PCDH15). The molecular conformation of CDH23, its binding to PCDH15, the tip links, and mechanoelectrical transduction have all been shown previously to be sensitive to exposure to low levels of calcium. The aim of this study was to compare the characteristics of tip links in guinea-pig cochlear hair cells with reported features of the CDH23-PCDH15 complex. Tip links were examined using field emission scanning electron microscopy and transmission electron microscopy in conventional preparations and after treatment with the detergent Triton-X-100 or varying calcium concentrations in the extracellular solution. The results showed that tip links have a twisted double-stranded appearance with a branched upper region. They survived demembranation of the stereocilia by detergent suggesting that they have transmembrane domains at both ends. Their lengths, when fixed in the presence of 2 mM extracellular calcium, were approximately 150 nm. With prior exposure to 1 mM calcium their lengths were approximately 164 nm. The lengths in 50 microM calcium are similar ( approximately 185 nm) to those reported for CDH23-PCDH15 complexes in 100 microM calcium ( approximately 180 nm). Exposure to 1 microM calcium caused loss of tip links and an increased distance between the residual attachment sites. The data indicate that extracellular calcium concentration affects tip-link length. One model compatible with the recently proposed tip-link structure is that the CDH23 double strand undergoes calcium-dependent unfolding, changing the length of the links. The bundle may also tilt in the direction of the tallest row of stereocilia as the tip link lengthens and then is lost. Overall, our data are consistent with a tip link composed of complexes of CDH23 and PCDH15 but do not rule out other possibilities.

A quantitative assessment of glutamate uptake into hippocampal synaptic terminals and astrocytes: new insights into a neuronal role for excitatory amino acid transporter 2 (EAAT2).

The relative distribution of the excitatory amino acid transporter 2 (EAAT2) between synaptic terminals and astroglia, and the importance of EAAT2 for the uptake into terminals is still unresolved. Here we have used antibodies to glutaraldehyde-fixed d-aspartate to identify electron microscopically the sites of d-aspartate accumulation in hippocampal slices. About 3/4 of all terminals in the stratum radiatum CA1 accumulated d-aspartate-immunoreactivity by an active dihydrokainate-sensitive mechanism which was absent in EAAT2 glutamate transporter knockout mice. These terminals were responsible for more than half of all d-aspartate uptake of external substrate in the slices. This is unexpected as EAAT2-immunoreactivity observed in intact brain tissue is mainly associated with astroglia. However, when examining synaptosomes and slice preparations where the extracellular space is larger than in perfusion fixed tissue, it was confirmed that most EAAT2 is in astroglia (about 80%). Neither d-aspartate uptake nor EAAT2 protein was detected in dendritic spines. About 6% of the EAAT2-immunoreactivity was detected in the plasma membrane of synaptic terminals (both within and outside of the synaptic cleft). Most of the remaining immunoreactivity (8%) was found in axons where it was distributed in a plasma membrane surface area several times larger than that of astroglia. This explains why the densities of neuronal EAAT2 are low despite high levels of mRNA in CA3 pyramidal cell bodies, but not why EAAT2 in terminals account for more than half of the uptake of exogenous substrate by hippocampal slice preparations. This and the relative amount of terminal versus glial uptake in the intact brain remain to be discovered.

Neuronal vs glial glutamate uptake: Resolving the conundrum.

Neither normal brain function nor the pathological processes involved in neurological diseases can be adequately understood without knowledge of the release, uptake and metabolism of glutamate. The reason for this is that glutamate (a) is the most abundant amino acid in the brain, (b) is at the cross-roads between several metabolic pathways, and (c) serves as the major excitatory neurotransmitter. In fact most brain cells express glutamate receptors and are thereby influenced by extracellular glutamate. In agreement, brain cells have powerful uptake systems that constantly remove glutamate from the extracellular fluid and thereby limit receptor activation. It has been clear since the 1970s that both astrocytes and neurons express glutamate transporters. However the relative contribution of neuronal and glial transporters to the total glutamate uptake activity, however, as well as their functional importance, has been hotly debated ever since. The present short review provides (a) an overview of what we know about neuronal glutamate uptake as well as an historical description of how we got there, and (b) a hypothesis reconciling apparently contradicting observations thereby possibly resolving the paradox.

Specificity of antibodies: unexpected cross-reactivity of antibodies directed against the excitatory amino acid transporter 3 (EAAT3).

UNLABELLED: Specific antibodies are essential tools for identifying individual proteins in biological samples. While generation of antibodies is often straightforward, determination of the antibody specificity is not. Here we illustrate this by describing the production and characterization of antibodies to excitatory amino acid transporter 3 (EAAT3). We synthesized 13 peptides corresponding to parts of the EAAT3 sequence and immunized 6 sheep and 30 rabbits. All sera were affinity purified against the relevant immobilized peptide. Antibodies to the peptides were obtained in almost all cases. Immunoblotting with tissue extracts from wild type and EAAT3 knockout animals revealed that most of the antibodies did not recognize the native EAAT3 protein, and that some recognized other proteins. Several immunization protocols were tried, but strong reactions with EAAT3 were only seen with antibodies to the C-terminal peptides. In contrast, good antibodies were obtained to several parts of EAAT2. EAAT3 was only detected in neurons. However, rabbits immunized with an EAAT3-peptide corresponding to residues 479-498 produced antibodies that labeled axoplasm and microtubules therein particularly strongly. On blots, these antibodies recognized both EAAT3 and a slightly smaller, but far more abundant protein that turned out to be tubulin. The antibodies were fractionated on columns with immobilized tubulin. One fraction contained antibodies apparently specific for EAAT3 while another fraction contained antibodies recognizing both EAAT3 and tubulin despite the lack of primary sequence identity between the two proteins. Addition of free peptide to the incubation solution blocked immunostaining of both EAAT3 and tubulin. CONCLUSIONS: Not all antibodies to synthetic peptides recognize the native protein. The peptide sequence is more important than immunization protocol. The specificity of an antibody is hard to predict because cross-reactivity can be specific and to unrelated molecules. The antigen preabsorption test is of little value in testing the specificity of affinity purified antibodies.

Differential distribution of beta- and gamma-actin in guinea-pig cochlear sensory and supporting cells.

Sensory and supporting cells of the mammalian organ of Corti have cytoskeletons containing beta- and gamma-actin isoforms which have been described as having differing intracellular distributions in chick cochlear hair cells. Here, we have used post-embedding immunogold labelling for beta- and gamma-actin to investigate semiquantitatively how they are distributed in the guinea-pig cochlea and to compare different frequency locations. Amounts of beta-actin decrease and gamma-actin increase in the order, outer pillar cells, inner pillar cells, Deiters' cells and hair cells. There is also more beta-actin and less gamma-actin in outer pillar cells in higher than lower frequency regions. In hair cells, beta-actin is present in the cuticular plate but is more concentrated in the stereocilia, especially in the rootlets and towards the periphery of their shafts; labelling densities for gamma-actin differ less between these locations and it is the predominant isoform of the hair-cell lateral wall. Alignments of immunogold particles suggest beta-actin and gamma-actin form homomeric filaments. These data confirm differential distribution of these actin isoforms in the mammalian cochlea and reveal systematic differences between sensory and supporting cells. Increased expression of beta-actin in outer pillar cells towards the cochlear base may contribute to the greater stiffness of this region.

Putative immunolocalization of the mechanoelectrical transduction channels in mammalian cochlear hair cells.

Hair cells bear an apical bundle of stereocilia arranged in serried rows. Deflection of the bundle controls the opening and closing of mechanoelectrical transduction channels, thereby altering the conductance across the apical plasma membrane. Two locations for these channels have been proposed in the bundle, either near the bases of the stereocilia or towards their tips. One hypothesis that is consistent with the latter possibility suggests that fine extracellular filaments, which run between the tips of the shorter stereocilia and the sides of the taller stereocilia behind, operate the channels. Determining the precise position of the channels is essential to test this hypothesis. We have therefore attempted to localize them immunocytochemically. Because hair-cell transduction is amiloride sensitive, the channels may have an amiloride-binding site associated with them. We have therefore used a polyclonal antibody raised against another amiloride-sensitive ion channel to hunt for them. This antibody recognizes a 62-64 kDa band in immunoblots of cochlear tissue, and produces discrete labelling in the hair bundle. This is most concentrated just below the tips of the shorter stereocilia, coinciding with a region of specialization in the closely apposed membranes of the short and tall stereocilia but not with either end of the tip link.

Kinematic analysis of shear displacement as a means for operating mechanotransduction channels in the contact region between adjacent stereocilia of mammalian cochlear hair cells.

In sensory hair cells of the cochlea, deflection of the stereociliary bundle results in direct mechanical gating of mechanoelectrical transduction channels, a function generally attributed to the tip link running between the tips of short stereocilia and the sides of adjacent taller ones. However, immunocytochemical experiments indicate that the channels may not be associated with the tip link but occur just below it in a region of contact between the stereocilia. To determine whether transduction channels in this location could be operated during physiologically appropriate deflections as effectively by shear displacement as if they were associated with the tip link, a two dimensional kinematic analysis of relative motion between stereocilia has been performed assuming contact between stereocilia is maintained during deflection. Bundle geometry and dimensions were determined from transmission electron micrographs of hair cells from several frequency locations between 0.27 and 13.00 kHz in the guinea-pig cochlea. The analysis indicates that for a 10 nm deflection of the tallest stereocilia of both inner and outer hair cells, i.e. within the range of the maximum sensitivity of mammalian hair bundles, the average shear displacement in the contact region would be 1.6 nm, but that it increases systematically towards higher frequency regions for outer hair cells. This displacement is comparable in magnitude to tip-link elongation for individual stereociliary pairs.

High-resolution scanning-electron microscopy of stereocilia using the osmium-thiocarbohydrazide coating technique.

Further observations on the detailed morphology of stereocilia have been made using high-resolution scanning-electron microscopy of osmium-thiocarbohydrazide-coated guinea pig cochleae. Three types of cross-link have been observed between stereocilia. Side-to-side and row-to-row linkages are composed of a filamentous network whilst upward-pointing links are a fine single strand, often with a terminal widening. The stereocilia have rough surfaces. These features are observed on both inner and outer hair cells despite reported sensitivity to long periods of osmium fixation. We suggest that osmium sensitivity may be altered by the buffering conditions used during preparation. The observations on osmium-coated material correspond more closely with those made using transmission-electron microscopy than those made using other scanning-electron microscopical preparation techniques, since gold-coating artefacts are absent and the degree of specimen collapse is less. This has enabled us to observe fine details of the links and their attachments which have not been reported previously in SEM.

Cross-links between stereocilia in the guinea pig cochlea.

Cross-links between stereocilia on guinea pig cochlear hair cells have been examined using high resolution scanning (SEM) and transmission electron microscopy (TEM), confirming recent descriptions of these structures. Links from the tips of shorter stereocilia to the sides of the adjacent taller stereocilia (upward-pointing links), between stereocilia of the same row (side-to-side links) and between adjacent rows (row-to-row links), have been observed on inner and outer hair cells. These links have been seen in material fixed using (1) glutaraldehyde only, (2) glutaraldehyde/osmium and (3) glutaraldehyde/osmium/thiocarbohydrazide (a technique which makes gold coating unnecessary). Upward-pointing links were seen less frequently, and the surfaces of stereocilia and microvilli were smoother after fixation (3) compared with fixations (1) and (2) in which they were usually roughened in appearance. In TEM, side-to-side and row-to-row links form a regular lattice between stereocilia, and consist of a number of strands. Upward-pointing links consist of a single strand, the ends of which are associated with electron-dense material. This lies between the stereociliary membrane and the actin filament bundle, at the tip of the shorter stereocilium and the side of the taller stereocilium.

The functional morphology of stereociliary bundles on turtle cochlear hair cells.

The stereociliary bundles of hair cells from the basilar papilla of the red-eared turtle were examined with transmission and high resolution scanning electron microscopy to provide a description of their morphology, orientation and inter-ciliary connections for comparison with physiological observations. Bundles on hair cells in the basilar membrane region are of a uniform shape and orientation, but bundles on the apical and basal limbus are distinct in having elongated kinocilia more than twice the length of the tallest stereocilia. Bundles in the basilar membrane region show a roughly two-fold increase in height from 5 to 9 microns from base to apex. Electrical recordings from isolated hair cells indicate that the bundle height is inversely proportional to the cell's characteristic frequency. It is argued that the change in dimensions is insufficient to contribute significantly to the cochlea's frequency selectivity. The cytoplasm adjacent to the kinocilium is filled with microtubules and large vesicles, and there are coated pits in the apical membrane which, it is suggested, may be indicative of rapid turnover of the membrane in this region.

Ultrastructural localisation of spectrin in sensory and supporting cells of guinea-pig organ of Corti.

Spectrin is a cytoskeletal protein found in the cortex of many cell types. It is known to occur in cochlear outer hair cells (OHCs) with previous immunoelectron microscopical studies showing that it is located in the cuticular plate and the cortical lattice. The latter is a network of filaments associated with the lateral plasma membrane that is thought to play a role in OHC motility. Spectrin has also been found in inner hair cells (IHCs) and supporting cells using immunofluorescent techniques, but its ultrastructural distribution in these cells has not yet been described. This has, therefore, been investigated using a monoclonal antibody to alpha-spectrin in conjunction with pre- and post-embedding immunogold labelling for transmission electron microscopy. Labelling was found in a meshwork of filaments beneath the plasma membranes of both IHCs and supporting cells and, in pillar cells, close to microtubule/microfilament arrays. It was also found in association with the stereocilia of OHCs and IHCs and, as expected, in the cortical lattice and cuticular plate of OHCs. Thus, spectrin is a general component of cytoskeletal structures involved in maintaining the specialised cell shapes in the organ of Corti and may contribute to the mechanical properties of all the cell types examined.

The effect of explantation and neomycin on hair cells and supporting cells in organotypic cultures of the adult guinea-pig utricle.

Recent reports suggest that immature hair bundles are observed following aminoglycoside-induced hair-cell loss in the mammalian utricle in vitro as well as in vivo. It is therefore important to document the initial morphological changes associated with both culturing and aminoglycoside application so that degeneration can be clearly distinguished from regeneration. In this study, utricles from adult guinea pigs were maintained in culture for either 3 or 8 days, half being exposed to neomycin for days 2 and 3. They were then processed for microscopical examination and compared with control utricles from animals of the same age. The numbers of hair-cell and supporting-cell nuclei were counted and hair-cell morphology assessed. Bundles were classified as having either stepped (SHB) or unstepped (UHB) stereocilia, and their density determined. The numbers of hair-cell, but not supporting-cell, nuclei declined significantly compared with controls in both untreated and treated explants, the greatest reduction occurring 5 days after neomycin administration. The density of SHBs also declined but there was no significant change in UHB density, resulting in a residual population of hair bundles of more immature appearance in both untreated and treated utricles in vitro than in vivo. Although degenerative events such as hair-cell ejection from, or retraction into, the sensory epithelium were observed, no evidence of regeneration was found.

The possible relationship between transient evoked otoacoustic emissions and organ of Corti irregularities in the guinea pig.

Otoacoustic emissions are believed to arise from an active process associated with the outer hair cells in the mammalian organ of Corti. They have been attributed to the presence of impedance discontinuities on the basilar membrane which might be caused by hair cell irregularities. To test this hypothesis we have investigated the possible relationship between transient evoked otoacoustic emissions (TEOAEs) and anatomical integrity in the organ of Corti. Click-evoked TEOAEs have been measured from the ear canals of normal, pigmented guinea pigs using an Otodynamics ILO88 analyser. Emissions were present in 18 out of 19 animals tested and the major frequencies observed were consistently present in different measurements over periods of up to ten weeks provided recording conditions were satisfactory. The frequency spectra of the TEOAEs resembled those measured in humans but the latencies of the responses were considerably shorter. In one acute experiment, the TEOAEs were shown to be dependent on metabolic energy as they were lost rapidly following termination with an overdose of anaesthetic. In another case, evoked emissions of long duration (sustained) at about 1 kHz were obtained from both ears. All cochleae examined showed irregularities, especially patches of mainly apical outer hair cell loss of differing extents. However, there was no evidence that substantial lesions coincided consistently with the frequency regions corresponding to the major emissions. Nevertheless, it was noted that the total energy level of emissions was proportional to the total outer hair cell loss, except in one case, where the outer hair cell loss was substantial and the energy level of TEOAEs was considerably lower. Although there is no clear relationship between TEOAEs of specific frequencies and abnormalities at the corresponding cochleotopic location in the organ of Corti which could represent impedance discontinuities, the degree of irregularity may determine the overall emission level. This finding is consistent with the idea that emissions arise as a result of irregularity producing variations in the reflection coefficient.

Stereociliary cross-links between adjacent inner hair cells.

An extensive network of intracellular cross-links occurs between the stereocilia of each cochlear hair cell bundle. These links fall into two main categories; lateral links which run roughly horizontal with respect to the reticular lamina and which join stereocilia of the same or adjacent rows, and tip-to-side links which run at a more vertical angle from the tip of each shorter stereocilium to the side of the adjacent longer stereocilium in the row behind. It has been proposed that deformation of the tip-to-side links causes alteration of the rate of opening of ion channels, producing transduction. Lateral linkages also occur between the stereocilia of adjacent hair cells. Now, intercellular links which resemble the tip-to-side links have been observed. Some of these occur in positions inappropriate to their proposed role in transduction. Several hypotheses are proposed to account for their presence e.g., the links could represent the remnants of a glycocalyx which is best preserved in areas where stereocilia are closely opposed.

Stereociliary bundle morphology in organotypic cultures of the mouse cochlea.

Organotypic cultures of the neonatal mouse cochlea have a band of hair cells consisting of 3-5 rows of outer hair cells and a single row of inner hair cells. The outer hair cell stereociliary bundles show progressive differentiation from the apical to the basal ends of the band. Undifferentiated apical bundles have a disk-like array of short stereocilia resembling microvilli. Partially differentiated bundles are hemispherical with poorly organized rows of thickly clustered stereocilia, which gradually increase in height in the direction of the kinocilium. More differentiated bundles remain hemispherical with many microvilli-like stereocilia, but have highly organized rows of sterocilia along the side nearest to the kinocilium, and well-defined height increments between the rows. Highly-differentiated, basal bundles usually have a 'V' or 'W' shape, although some can be almost polygonal. The basal bundles have 4-5 regular rows of stereocilia with a well-defined gradient in height across the rows, and very few microvilli-like stereocilia. Cross-links are only consistently observed in more differentiated bundles, where the rows of stereocilia are regular and have significant height increments across them. The links show a wide variety of forms and orientations not previously observed in other preparations. Spoke-like arrays of links project from the upper regions of many stereocilia and other stereocilia appear to bear distinct tip-to-side links, although with a variety of orientations. A similar variety of cross-links is observed in early postnatal cochleae in vivo, but not in the cochleae of adult mice, indicating that this variety may be a transient feature of sterociliary bundle development. In vitro, inner hair cell stereociliary bundles are often covered by overlying material from the developing tectorial membrane. The variations in morphology of inner hair cell bundles and their cross-links are similar to those of the outer hair cell bundles.

Ultrastructural localization of cadherin in the adult guinea-pig organ of Corti.

The apices of the majority of cells of the organ of Corti are connected together by junctional complexes to form the reticular lamina, a barrier that prevents the mixing of endolymph and perilymph. These complexes include tight junctions, adherens junctions and desmosomes. Further information is required about the identity and distribution of the molecules involved in these connections if the function and organization of the reticular lamina are to be well understood. One major category of molecules occurring in adherens junctions and desmosomes, and involved in the maintenance of tissue integrity, is the cadherins. However, although cadherin has been identified in junctions between supporting cells in the adult mammalian organ of Corti at the light microscopic level, its ultrastructural distribution has not so far been described. A post-embedding immunogold labelling technique has therefore been used in conjunction with a monoclonal antibody to cadherin to investigate its ultrastructural distribution in the adult guinea-pig reticular lamina. Immunolabelling is observed in hair cell-supporting cell junctions and in supporting cell-supporting cell junctions. In addition, there is more labelling associated with inner hair cell-supporting cell junctions than with outer hair cell-supporting cell junctions. This may indicate that the junctions associated with the two types of hair cell have different functional properties.

The binding site on cochlear stereocilia for antisera raised against renal Na+ channels is blocked by amiloride and dihydrostreptomycin.

The mechanoelectrical transduction channels on hair cells have been suggested to be operated by tip links that are stretched when the hair bundle is deflected in the direction of the tallest row of stereocilia. Localising these channels is therefore an important test of this hypothesis. The transduction channels are known to be amiloride-sensitive and immunogold labelling with antibodies raised against the amiloride-sensitive epithelial Na+ channel from kidney (alpha NaCh), has suggested that sites with similar characteristics are located in the region where the tips of the shorter stereocilia appear to come into contact with the sides of the adjacent taller stereocilia rather than being associated directly with the tip links. Now, further immunocytochemical experiments have been performed to determine if amiloride and dihydrostreptomycin, both of which can block transduction, can affect this labelling. Immunofluorescent labelling of the stereocilia is obtained when surface preparations of the organ of Corti are fixed and incubated with alpha NaCh followed by an appropriate secondary antibody. This labelling is abolished by trypsinization prior to fixation but retained if the tissue is pretreated with amiloride and then trypsinized in its presence. Because amiloride is known to protect amiloride-binding sites from degradation by trypsin, these results suggest that alpha NaCh is revealing amiloride-binding sites on the stereocilia. Similarly, immunofluorescent labelling of the stereocilia is abolished if cochlear tissue is pretreated with dihydrostreptomycin (DHS) and fixed in its presence prior to incubation with alpha NaCh. Quantitative analysis of colloidal gold labelling using transmission electron microscopy shows that DHS treatment produces a significant reduction in the number of gold particles on stereocilia, especially in the region of contact between them. These results suggest that anti-Na+ recognises a site with characteristics similar to the mechanoelectrical transduction channels.

An immunogold investigation of the distribution of calmodulin in the apex of cochlear hair cells.

Calmodulin is found in the mechanosensitive stereociliary bundle of hair cells where it plays a role in various calcium-sensitive events associated with mechanoelectrical transduction. In this study, we have investigated the ultrastructural distribution of calmodulin in the apex of guinea-pig cochlear hair cells, using post-embedding immunogold labelling, in order to determine in more detail where calmodulin-dependent processes may be occurring. Labelling was found in the cuticular plate as well as the hair bundle, the rootlets of the stereocilia being more densely labelled than the surrounding filamentous matrix. In the bundle, labelling was found almost exclusively at the periphery rather than over the centre of the actin core of the stereocilia, and was clearly associated with the attachments of the lateral links that connect them to their nearest neighbours. It was also found to be denser towards the tips of stereocilia compared to other stereociliary regions and occurred consistently at either end of the tip link that connects stereocilia of adjacent rows. The contact region between stereocilia that is found just below the tip link was also clearly labelled. These concentrations of labelling in the bundle are likely to indicate sites where calmodulin is associated with calcium/calmodulin-sensitive proteins such as the various myosin isoforms and the plasma membrane ATPase (PMCA2a) that are known to occur there, and possibly with the transduction channels themselves. At least one of the myosin isoforms, myosin 1c, is thought to be associated with slow adaptation, and PMCA2a with control of calcium levels in the bundle. The concentration of calmodulin in the contact region further supports the suggestion that this is a functionally distinct region rather than a simple geometrical association between adjacent stereocilia.