Fimbrin is a homologue of the cytoplasmic phosphoprotein plastin and has domains homologous with calmodulin and actin gelation proteins

Fimbrin is an actin-bundling protein found in intestinal microvilli, hair cell stereocilia, and fibroblast filopodia. The complete protein sequence (630 residues) of chicken intestine fimbrin has been determined from two full-length cDNA clones. The sequence encodes a small amino-terminal domain (115 residues) that is homologous with two calcium-binding sites of calmodulin and a large carboxy-terminal domain (500 residues) consisting of a fourfold-repeated 125-residue sequence. This repeat is homologous with the actin-binding domain of alpha-actinin and the amino-terminal domains of dystrophin, actin-gelation protein, and beta-spectrin. The presence of this duplicated domain in fimbrin links actin bundling proteins and gelation proteins into a common family of actin cross-linking proteins. Fimbrin is also homologous in sequence with human L-plastin and T-plastin. L-plastin is found in only normal or transformed leukocytes where it becomes phosphorylated in response to IL 1 or phorbol myristate acetate. T-plastin is found in cells of solid tissues where it does not become phosphorylated. Neoplastic cells derived from solid tissues express both isoforms. The differences in expression, sequence, and phosphorylation suggest possible functional differences between fimbrin isoforms.     

Functional differences between L- and T-plastin isoforms

Fimbrins/plastins are a family of highly conserved actin-bundling proteins. They are present in all eukaryotic cells including yeast, but each isoform displays a remarkable tissue specificity. T-plastin is normally found in epithelial and mesenchymal cells while L-plastin is present in hematopoietic cells. However, L-plastin has been also found in tumor cells of non-hematopoietic origin (Lin, C.-S., R. H. Aebersold, S. B. Kent, M. Varma, and J. Leavitt. 1988. Mol. Cell. Biol. 8:4659-4668; Lin, C.-S., R. H. Aebersold, and J. Leavitt. 1990. Mol. Cell. Biol. 10: 1818-1821). To learn more about the biological significance of their tissue specificity, we have overproduced the T- and L-plastin isoforms in a fibroblast-like cell line, CV-1, and in a polarized epithelial cell line, LLC-PK1. In CV-1 cells, overproduction of T- and L-plastins induces cell rounding and a concomitant reorganization of actin stress fibers into geodesic structures. L- plastin remains associated with microfilaments while T-plastin is almost completely extracted after treatment of the cells with non-ionic detergent. In LLC-PK1 cells, T-plastin induces shape changes in microvilli and remains associated with microvillar actin filaments after detergent extraction while L-plastin has no effect on these structures and is completely extracted. The effect of T-plastin on the organization of microvilli differs from that of villin, another actin- bundling protein. Our experiments indicate that these two isoforms play differing roles in actin filament organization, and do so in a cell type-specific fashion. Thus it is likely that these plastin isoforms play fundamentally different roles in cell function.     

Preliminary biochemical characterization of the stereocilia and cuticular plate of hair cells of the chick cochlea

The sensory epithelium of the chick cochlea contains only two cell types, hair cells and supporting cells. We developed methods to rapidly dissect out the sensory epithelium and to prepare a detergent-extracted cytoskeleton. High salt treatment of the cytoskeleton leaves a "hair border", containing actin filament bundles of the stereocilia still attached to the cuticular plate. On SDS-PAGE stained with silver the intact epithelium is seen to contain a large number of bands, the most prominent of which are calbindin and actin. Detergent extraction solubilizes most of the proteins including calbindin. On immunoblots antibodies prepared against fimbrin from chicken intestinal epithelial cells cross react with the 57- and 65-kD bands present in the sensory epithelium and the cytoskeleton. It is probable that the 57-kD is a proteolytic fragment of the 65-kD protein. Preparations of stereocilia attached to the overlying tectorial membrane contain the 57- and 65-kD bands. A 400-kD band is present in the cuticular plate. By immunofluorescence, fimbrin is detected in stereocilia but not in the hair borders after salt extraction. The prominent 125 A transverse stripping pattern characteristic of the actin cross-bridges in a bundle is also absent in hair borders suggesting fimbrin as the component that gives rise to the transverse stripes. Because the actin filaments in the stereocilia of hair borders still remain as compact bundles, albeit very disordered, there must be an additional uncharacterized protein besides fimbrin that cross-links the actin filaments together.     

Identification of I-plastin, a human fimbrin isoform expressed in intestine and kidney.

The complete cDNA sequence of human intestine-specific plastin (I-plastin) was determined from a clone derived by PCR. It consists of a 97-bp 5' untranslated region, a 1,887-bp coding region, and a 1,655-bp 3' untranslated region. The coding region predicts a 629-residue polypeptide whose sequence displays 86, 75, and 73% identities with chicken intestine fimbrin, human T-plastin, and human L-plastin, respectively. Recombinant I-plastin cross-linked actin filaments into bundles in the absence but not in the presence of calcium. The I-plastin gene was mapped by PCR to human chromosome 3; the L- and T-plastin genes were previously mapped to chromosomes 13 and X, respectively. I-plastin mRNA was detected in the small intestine, colon, and kidneys; relatively lower levels of expression were detected in the lungs and stomach. In contrast, L-plastin expression was restricted to the spleen and other lymph node-containing organs, while T-plastin was expressed in a variety of organs, including muscle, brain, uterus, and esophagus. In contrast to the situation for the intestine, high levels of L- and T-plastin mRNAs were detected in Caco-2, a human colon-derived cell line. Immunofluorescence microscopy detected I-plastin in the brush border of the small intestine and colon. These results identify I-plastin as the human homolog of chicken intestine fimbrin and as a third plastin isoform in humans.     

Ectopic expression of L-plastin in human tumor cells: Diagnostic and therapeutic implications

Modulation of the actin cytoskeleton is critical for tumor cell migration and invasion. Therefore, actin-binding proteins which regulate this modulation may be valuable targets to inhibit the metastatic properties of tumor cells. Changes in the actin cytoskeleton are accomplished by a variety of actin-binding proteins such as cofilin, α-actinin, filamin, fascin and the plastins. Interestingly, the hematopoetic isoform of the plastins, L-plastin, is not only expressed by hematopoetic cells, but also by most human cancer cell lines. Yet, data regarding the functional importance of L-plastin expression in tumor tissues are controversial: in colon carcinomas, the expression level of L-plastin correlated with tumor progression, whereas no such correlation could be seen in breast carcinomas. We therefore systematically investigated whether expression of L-plastin influences the adhesiveness, the motility and invasiveness of human tumor cells. An siRNA mediated knock-down of L-plastin in an L-plastin positive melanoma cell line inhibited migration of these cells. Accordingly, expression of L-plastin in L-plastin negative melanoma cells led to enhanced cell migration towards extracellular matrix components. However, mere expression of L-plastin did not promote tumor cell invasion into basement membranes. Only, if L-plastin was phosphorylated, tumor cell invasion was promoted. Therefore, in clinical studies, not only the expression of L-plastin but also the phosphorylation status of L-plastin should be compared with regard to tumor progression.Besides the potential prognostic relevance of L-plastin expression and phosphorylation in human cancer cells, L-plastin may represent a novel target for cancer therapy. Moreover, the constitutive activity of the L-plastin promotor in non-hematopoetic tumors opens up novel perspectives for gene therapy of cancer using L-plastin-promotor driven viral vectors.     

The actin filament content of hair cells of the bird cochlea is nearly constant even though the length, width, and number of stereocilia vary depending on the hair cell location

By direct counts off scanning electron micrographs, we determined the number of stereocilia per hair cell of the chicken cochlea as a function of the position of the hair cell on the cochlea. Micrographs of thin cross sections of stereociliary bundles located at known positions on the cochlea were enlarged and the total number of actin filaments per stereocilium was counted and recorded. By comparing the counts of filament number with measurements of actin filament bundle width of the same stereocilium, we were able to relate actin filament bundle width to filament number with an error margin (r2) of 16%. Combining this data with data already published or in the process of publication from our laboratory on the length and width of stereocilia, we were able to calculate the total length of actin filaments present in stereociliary bundles of hair cells located at a variety of positions on the cochlea. We found that stereociliary bundles of hair cells contain 80,000-98,000 micron of actin filament, i.e., the concentration of actin is constant in all hair cells with a range of values that is less than our error in measurement and/or biological variation, the greatest variation being in relating the diameters of the stereocilia to filament number. We also calculated the membrane surface needed to cover the stereocilia of hair cells located throughout the cochlea. The values (172-192 micron 2) are also constant. The implications of our observation that the total amount of actin is constant even though the length, width, and number of stereocilia per hair cell vary are discussed.     

Molecular basis for dissimilar nuclear trafficking of the actin-bundling protein isoforms T- and L-plastin

T- and L-plastin are highly similar actin-bundling proteins implicated in the regulation of cell morphology, lamellipodium protrusion, bacterial invasion and tumor progression. We show that T-plastin localizes predominantly to the cytoplasm, whereas L-plastin distributes between nucleus and cytoplasm in HeLa or Cos cells. T-plastin shows nuclear accumulation upon incubation of cells with the CRM1 antagonist leptomycin B (LMB). We identified a Rev-like nuclear export sequence (NES) in T-plastin that is able to export an otherwise nuclear protein in an LMB-dependent manner. Deletion of the NES promotes nuclear accumulation of T-plastin. Mutation of residues L17, F21 or L26 in the T-plastin NES inhibits nuclear efflux. L-plastin harbors a less conserved NES and lacks the F21 T-plastin residue. Insertion of a Phe residue in the L-plastin NES specifically enhances its export activity. These findings explain why both isoforms exhibit specific distribution patterns in eukaryotic cells.     

Actin filaments, stereocilia, and hair cells of the bird cochlea. II. Packing of actin filaments in the stereocilia and in the cuticular plate and what happens to the organization when the stereocilia are bent

A comparison of hair cells from different parts of the cochlea reveals the same organization of actin filaments; the elements that vary are the length and number of the filaments. Thin sections of stereocilia reveal that the actin filaments are hexagonally packed and from diffraction patterns of these sections we found that the actin filaments are aligned such that the crossover points of adjacent actin filaments are in register. As a result, the cross-bridges that connect adjacent actin filaments are easily seen in longitudinal sections. The cross-bridges appear as regularly spaced bands that are perpendicular to the axis of the stereocilium. Particularly interesting is that, unlike what one might predict, when a stereocilium is bent or displaced, as might occur during stimulation by sound, the actin filaments are not compressed or stretched but slide past one another so that the bridges become tilted relative to the long axis of the actin filament bundle. In the images of bent bundles, the bands of cross- bridges are then tilted off perpendicular to the stereocilium axis. When the stereocilium is bent at its base, all cross-bridges in the stereocilium are affected. Thus, resistance to bending or displacement must be property of the number of bridges present, which in turn is a function of the number of actin filaments present and their respective lengths. Since hair cells in different parts of the cochlea have stereocilia of different, yet predictable lengths and widths, this means that the force needed to displace the stereocilia of hair cells located at different regions of the cochlea will not be the same. This suggests that fine tuning of the hair cells must be a built-in property of the stereocilia. Perhaps its physiological vulnerability may result from changes of stereociliary structure.     

Actin filaments, stereocilia, and hair cells of the bird cochlea. I. Length, number, width, and distribution of stereocilia of each hair cell are related to the position of the hair cell on the cochlea

Located on the sensory epithelium of the sickle-shaped cochlea of a 7- to 10-d-old chick are approximately 5,000 hair cells. When the apical surface of these cell is examined by scanning microscopy, we find that the length, number, width, and distribution of the stereocilia on each hair cell are predetermined. Thus, a hair cell located at the distal end of the cochlea has 50 stereocilia, the longest of which are 5.5 microns in length and 0.12 microns in width, while those at the proximal end number 300 and are maximally 1.5 microns in length and 0.2 micron in width. In fact, if we travel along the cochlea from its distal to proximal end, we see that the stereocilia on successive hair cells gradually increase in number and width, yet decrease in length. Also, if we look transversely across the cochlea where adjacent hair cells have the same length and number of stereocilia (they are the same distance from the distal end of the cochlea), we find that the stereocilia of successive hair cells become thinner and that the apical surface area of the hair cell proper, not including the stereocilia, decreases from a maximum of 80 microns2 to 15 microns2. Thus, if we are told the length of the longest stereocilium on a hair cell and the width of that stereocilium, we can pinpoint the position of that hair cell on the cochlea in two axes. Likewise, if we are told the number of stereocilia and the apical surface of a hair cell, we can pinpoint the location of that cell in two axes. The distribution of the stereocilia on the apical surface of the cell is also precisely determined. More specifically, the stereocilia are hexagonally packed and this hexagonal lattice is precisely positioned relative to the kinocilium. Because of the precision with which individual hair cells regulate the length, width, number, and distribution of their cell extensions, we have a magnificent object with which to ask questions about how actin filaments that are present within the cell are regulated. Equally interesting is that the gradient in stereociliary length, number, width, and distribution may play an important role in frequency discrimination in the cochlea. This conclusion is amplified by the information presented in the accompanying paper (Tilney, L.G., E.H. Egelman, D.J. DeRosier, and J.C. Saunders, 1983, J. Cell Biol., 96:822- 834) on the packing of actin filaments in this stereocilia.     

Three different actin filament assemblies occur in every hair cell: each contains a specific actin crosslinking protein

The apex of hair cells of the chicken auditory organ contains three different kinds of assemblies of actin filaments in close spatial proximity. These are (a) paracrystals of actin filaments with identical polarity in stereocilia, (b) a dense gellike meshwork of actin filaments forming the cuticular plate, and (c) a bundle of parallel actin filaments with mixed polarities that constitute the circumferential filament belt attached to the cytoplasmic aspect of the zonula adhaerens (ZA). Each different supramolecular assembly of actin filaments contains a specific actin filament cross-linking protein which is unique to that particular assembly. Thus fimbrin appears to be responsible for paracrystallin packing of actin filaments in stereocillia; an isoform of spectrin resides in the cuticular plate where it forms the whisker-like crossbridges, and alpha actinin is the actin crosslinking protein of the circumferential ZA bundle. Tropomyosin, which stabilizes actin filaments, is present in all the actin filament assemblies except for the stereocilia. Another striking finding was that myosin appears to be absent from the ZA ring and cuticular plate of hair cells although present in the ZA ring of supporting cells. The abundance of myosin in the ZA ring of the surrounding supporting cells means that it may be important in forming a supporting tensile cellular framework in which the hair cells are inserted.     

Human T cell L-plastin bundles actin filaments in a calcium-dependent manner.

The amino acid sequences deduced from cDNA analyses revealed that human leucocyte L-plastin phosphorylated in response to interleukin 1, 2 closely resembles a chicken intestinal microvilli protein, fimbrin, that bundles actin filaments [de Arruda et al. (1990) J. Cell Biol. 111, 1069-1079]. In the present work, it was observed that unphosphorylated L-plastin isolated from human T cells bundled F-actin just as fimbrin does. L-Plastin acted on T cell beta-actin, but hardly acted on muscle alpha-actin or chicken gizzard gamma-actin, whereas fimbrin bundled muscle alpha-actin. Unlike fimbrin, L-plastin's actin-bundling action was strictly calcium-dependent: the bundles were formed at pCa 7, but not at pCa 6. Under suitable conditions, approximately one molecule of L-plastin bound to 8 molecules of actin monomer in the actin filament.     

The deaf mouse mutant whirler suggests a role for whirlin in actin filament dynamics and stereocilia development

Stereocilia, finger-like projections forming the hair bundle on the apical surface of sensory hair cells in the cochlea, are responsible for mechanosensation and ultimately the perception of sound. The actin cytoskeleton of the stereocilia contains hundreds of tightly cross-linked parallel actin filaments in a paracrystalline array and it is vital for their function. Although several genes have been identified and associated with stereocilia development, the molecular mechanisms responsible for stereocilia growth, maintenance and organisation of the hair bundle have not been fully resolved. Here we provide further characterisation of the stereocilia of the whirler mouse mutant. We found that a lack of whirlin protein in whirler mutants results in short stereocilia with larger diameters without a corresponding increase in the number of actin filaments in inner hair cells. However, a decrease in the actin filament packing density was evident in the whirler mutant. The electron-density at the tip of each stereocilium was markedly patchy and irregular in the whirler mutants compared with a uniform band in controls. The outer hair cell stereocilia of the whirler homozygote also showed an increase in diameter and variable heights within bundles. The number of outer hair cell stereocilia was significantly reduced and the centre-to-centre spacing between the stereocilia was greater than in the wildtype. Our findings suggest that whirlin plays an important role in actin filament packing and dynamics during postnatal stereocilium elongation.     

AtFim1 is an actin filament crosslinking protein from Arabidopsis thaliana

ATFIM1 is a widely expressed gene in Arabidopsis thaliana that encodes a putative actin filament-crosslinking protein, AtFim1, belonging to the fimbrin/plastin class of actin-binding proteins. In this report we have used bacterially expressed AtFim1 and actin isolated from Zea mays pollen to demonstrate that AtFim1 functions as an actin filament-crosslinking protein. AtFim1 binds pollen actin filaments (F-actin) in a calcium-independent manner, with an average dissociation constant (Kd) of 0.55+/-0.21 microM and with a stoichiometry at saturation of 1:4 (mol AtFim1 : mol actin monomer). AtFim1 also crosslinks pollen F-actin by a calcium-independent mechanism, in contrast to crosslinking of plant actin by human T-plastin, a known calcium-sensitive actin-crosslinking protein. When micro-injected at high concentration into living Tradescantia virginiana stamen hair cells, AtFim1 caused cessation of both cytoplasmic streaming and transvacuolar strand dynamics within 2-4 min. Using the 'nuclear displacement assay' as a measure of the integrity of the actin cytoskeleton in living stamen hair cells, we demonstrated that AtFim1 protects actin filaments in these cells from Z. mays profilin (ZmPRO5)-induced depolymerization, in a dose-dependent manner. The apparent ability of AtFim1 to protect actin filaments in vivo from profilin-mediated depolymerization was confirmed by in vitro sedimentation assays. Our results indicate that AtFim1 is a calcium-independent, actin filament-crosslinking protein that interacts with the actin cytoskeleton in living plant cells.     

Gelsolin Plays a Role in the Actin Polymerization Complex of Hair Cell Stereocilia

A complex of proteins scaffolded by the PDZ protein, whirlin, reside at the stereocilia tip and are critical for stereocilia development and elongation. We have shown that in outer hair cells (OHCs) whirlin is part of a larger complex involving the MAGUK protein, p55, and protein 4.1R. Whirlin interacts with p55 which is expressed exclusively in outer hair cells (OHC) in both the long stereocilia that make up the stereocilia bundle proper as well as surrounding shorter microvilli that will eventually regress. In erythrocytes, p55 forms a tripartite complex with protein 4.1R and glycophorin C promoting the assembly of actin filaments and the interaction of whirlin with p55 indicates that it plays a similar role in OHC stereocilia. However, the components directly involved in actin filament regulation in stereocilia are unknown. We have investigated additional components of the whirlin interactome by identifying interacting partners to p55. We show that the actin capping and severing protein, gelsolin, is a part of the whirlin complex. Gelsolin is detected in OHC where it localizes to the tips of the shorter rows but not to the longest row of stereocilia and the pattern of localisation at the apical hair cell surface is strikingly similar to p55. Like p55, gelsolin is ablated in the whirler and shaker2 mutants. Moreover, in a gelsolin mutant, stereocilia in the apex of the cochlea become long and straggly indicating defects in the regulation of stereocilia elongation. The identification of gelsolin provides for the first time a link between the whirlin scaffolding protein complex involved in stereocilia elongation and a known actin regulatory molecule.     

Generation and Characterization of Monoclonal Antibodies That Specifically Recognize p65/L-Plastin Isoform but Not T-Plastin Isoform

The 65-kDa protein (p65) was previously identified as a phosphorylated protein in activated macrophages, and has turned out to be a member of a plastin protein family characterized by a series of Ca²⁺-, calmodulin-, and β-actin-binding domains. In mice, two isoforms, p65/L-plastin and T-plastin, have so far been identified; p65/L-plastin is expressed in hemopoietic cells and cancer cells, and T-plastin in solid tissue cells. We generated monoclonal antibodies to p65/L-plastin, examined the isoform-specificity by using recombinant (r) T-plastin, and found that the antibodies were specific for rp65/L-plastin, whereas immune sera to rp65/L-plastin showed cross-reactions to rT-plastin. One of the antibodies, p65-7B5, was demonstrated to react to native p65/L-plastin by Western blot, flow cytometric, and immunohistochemical analysis. Furthermore, p65-7B5 has made it possible to detect p65/L-plastin-expressing cells in tissues where T-plastin is abundantly expressed. These reagents and procedures should provide specific tools to investigate the role of p65/L-plastin in leukocytes.     

Fluorescently-labeled fimbrin decorates a dynamic actin filament network in live plant cells

Recently it has been established, through a detailed biochemical analysis, that recombinant Arabidopsis thaliana fimbrin 1 (AtFim1) is a member of the fimbrin/plastin family of actin filament bundling or cross-linking proteins [D.R. Kovar et al. (2000) Plant J 24:625-636]. To determine whether AtFim1 can function as an F-actin-binding protein in the complex environment of the plant cell cytoplasm, we created a fluorescent protein analog and introduced it by microinjection into live Tradescantia virginiana L. stamen hair cells. AtFim1 derivatized with Oregon Green 488 had biochemical properties similar to unlabeled fimbrin, including the Kd value for binding to plant F-actin and the ability to cross-link filaments into higher-order structures. Fluorescent-fimbrin decorated an array of fine actin filaments in the cortical cytoplasm of stamen hair cells, which were shown with time-course studies to be highly dynamic. These data establish AtFim1 as a bona fide member of the fimbrin/plastin family, and represent the first use of a plant actin-binding protein as a powerful cytological tool for tracking the spatial and temporal redistribution of actin filaments in individual cells.     

Eps8 regulates hair bundle length and functional maturation of mammalian auditory hair cells

Hair cells of the mammalian cochlea are specialized for the dynamic coding of sound stimuli. The transduction of sound waves into electrical signals depends upon mechanosensitive hair bundles that project from the cell's apical surface. Each stereocilium within a hair bundle is composed of uniformly polarized and tightly packed actin filaments. Several stereociliary proteins have been shown to be associated with hair bundle development and function and are known to cause deafness in mice and humans when mutated. The growth of the stereociliar actin core is dynamically regulated at the actin filament barbed ends in the stereociliary tip. We show that Eps8, a protein with actin binding, bundling, and barbed-end capping activities in other systems, is a novel component of the hair bundle. Eps8 is localized predominantly at the tip of the stereocilia and is essential for their normal elongation and function. Moreover, we have found that Eps8 knockout mice are profoundly deaf and that IHCs, but not OHCs, fail to mature into fully functional sensory receptors. We propose that Eps8 directly regulates stereocilia growth in hair cells and also plays a crucial role in the physiological maturation of mammalian cochlear IHCs. Together, our results indicate that Eps8 is critical in coordinating the development and functionality of mammalian auditory hair cells.     

Generation and characterization of monoclonal antibodies that specifically recognize p65/L-plastin isoform but not T-plastin isoform

The 65-kDa protein (p65) was previously identified as a phosphorylated protein in activated macrophages, and has turned out to be a member of a plastin protein family characterized by a series of Ca(2+)-, calmodulin-, and beta-actin-binding domains. In mice, two isoforms, p65/L-plastin and T-plastin, have so far been identified; p65/L-plastin is expressed in hemopoietic cells and cancer cells, and T-plastin in solid tissue cells. We generated monoclonal antibodies to p65/L-plastin, examined the isoform-specificity by using recombinant (r) T-plastin, and found that the antibodies were specific for rp65/L-plastin, whereas immune sera to rp65/L-plastin showed cross-reactions to rT-plastin. One of the antibodies, p65-7B5, was demonstrated to react to native p65/L-plastin by Western blot, flow cytometric, and immunohistochemical analysis. Furthermore, p65-7B5 has made it possible to detect p65/L-plastin-expressing cells in tissues where T-plastin is abundantly expressed. These reagents and procedures should provide specific tools to investigate the role of p65/L-plastin in leukocytes.     

Differential expression of espin isoforms during epithelial morphogenesis, stereociliogenesis and postnatal maturation in the developing inner ear

The espins are a family of multifunctional actin cytoskeletal proteins. They are present in hair cell stereocilia and are the target of mutations that cause deafness and vestibular dysfunction. Here, we demonstrate that the different espin isoforms are expressed in complex spatiotemporal patterns during inner ear development. Espin 3 isoforms were prevalent in the epithelium of the otic pit, otocyst and membranous labyrinth as they underwent morphogenesis. This espin was down-regulated ahead of hair cell differentiation and during neuroblast delamination. Espin also accumulated in the epithelium of branchial clefts and pharyngeal pouches and during branching morphogenesis in other embryonic epithelial tissues, suggesting general roles for espins in epithelial morphogenesis. Espin reappeared later in inner ear development in differentiating hair cells. Its levels and compartmentalization to stereocilia increased during the formation and maturation of stereociliary bundles. Late in embryonic development, espin was also present in a tail-like process that emanated from the hair cell base. Increases in the levels of espin 1 and espin 4 isoforms correlated with stereocilium elongation and maturation in the vestibular system and cochlea, respectively. Our results suggest that the different espin isoforms play specific roles in actin cytoskeletal regulation during epithelial morphogenesis and hair cell differentiation.