p19(Ink4d) and p21(Cip1) collaborate to maintain the postmitotic state of auditory hair cells, their codeletion leading to DNA damage and p53-mediated apoptosis

Sensory hair cells of the auditory organ are generated during embryogenesis and remain postmitotic throughout life. Previous work has shown that inactivation of the cyclin-dependent kinase inhibitor (CKI) p19(Ink4d) leads to progressive hearing loss attributable to inappropriate DNA replication and subsequent apoptosis of hair cells. Here we show the synergistic action of another CKI, p21(Cip1), on cell cycle reactivation. The codeletion of p19(Ink4d) and p21(Cip1) triggered profuse S-phase entry of auditory hair cells during a restricted period in early postnatal life, leading to the transient appearance of supernumerary hair cells. In addition, we show that aberrant cell cycle reentry leads to activation of a DNA damage response pathway in these cells, followed by p53-mediated apoptosis. The majority of hair cells were absent in adult cochleas. These data, together with the demonstration of changing expression patterns of multiple CKIs in auditory hair cells during the stages of early postnatal maturation, show that the maintenance of the postmitotic state is an active, tissue-specific process, cooperatively regulated by several CKIs, and is critical for the lifelong survival of these sensory cells.

Mammalian cochlear supporting cells can divide and trans-differentiate into hair cells

Sensory hair cells of the mammalian organ of Corti in the inner ear do not regenerate when lost as a consequence of injury, disease, or age-related deafness. This contrasts with other vertebrates such as birds, where the death of hair cells causes surrounding supporting cells to re-enter the cell cycle and give rise to both new hair cells and supporting cells. It is not clear whether the lack of mammalian hair cell regeneration is due to an intrinsic inability of supporting cells to divide and differentiate or to an absence or blockade of regenerative signals. Here we show that post-mitotic supporting cells purified from the postnatal mouse cochlea retain the ability to divide and trans-differentiate into new hair cells in culture. Furthermore, we show that age-dependent changes in supporting cell proliferative capacity are due in part to changes in the ability to downregulate the cyclin-dependent kinase inhibitor p27(Kip1) (also known as Cdkn1b). These results indicate that postnatal mammalian supporting cells are potential targets for therapeutic manipulation.

A morphogenetic wave of p27Kip1 transcription directs cell cycle exit during organ of Corti development

The molecular mechanisms coordinating cell cycle exit with cell differentiation and organogenesis are a crucial, yet poorly understood, aspect of normal development. The mammalian cyclin-dependent kinase inhibitor p27(Kip1) is required for the correct timing of cell cycle exit in developing tissues, and thus plays a crucial role in this process. Although studies of p27(Kip1) regulation have revealed important posttranscriptional mechanisms regulating p27(Kip1) abundance, little is known about how developmental patterns of p27(Kip1) expression, and thus cell cycle exit, are achieved. Here, we show that during inner ear development transcriptional regulation of p27(Kip1) is the primary determinant of a wave of cell cycle exit that dictates the number of postmitotic progenitors destined to give rise to the hair cells and supporting cells of the organ of Corti. Interestingly, transcriptional induction from the p27(Kip1) gene occurs normally in p27(Kip1)-null mice, indicating that developmental regulation of p27(Kip1) transcription is independent of the timing of cell cycle exit. In addition, cell-type-specific patterns of p27(Kip1) transcriptional regulation are observed in the mature organ of Corti and retina, suggesting that this mechanism is important in differential regulation of the postmitotic state. This report establishes a link between the spatial and temporal pattern of p27(Kip1) transcription and the control of cell number during sensory organ morphogenesis.

Prospective identification and purification of hair cell and supporting cell progenitors from the embryonic cochlea

Expression of the cyclin-dependent kinase inhibitor p27(Kip1) defines a post-mitotic population of cells in the embryonic mammalian cochlea that constitutes the nascent organ of Corti. Here, we describe techniques to purify these precursors using a transgenic p27/GFP reporter and fluorescence activated cell sorting (FACS). We demonstrate that these cells express other markers of the sensory lineage, such as Sox2, and when placed in dissociated cell culture differentiate as hair cells and supporting cells. The purified sensory progenitors thus obtained provide a means of studying the process of hair cell and supporting cell differentiation in vitro, as well as providing a means of analyzing the molecular and physiological properties of this unique population of cells.

Dishevelled genes mediate a conserved mammalian PCP pathway to regulate convergent extension during neurulation

The planar cell polarity (PCP) pathway is conserved throughout evolution, but it mediates distinct developmental processes. In Drosophila, members of the PCP pathway localize in a polarized fashion to specify the cellular polarity within the plane of the epithelium, perpendicular to the apicobasal axis of the cell. In Xenopus and zebrafish, several homologs of the components of the fly PCP pathway control convergent extension. We have shown previously that mammalian PCP homologs regulate both cell polarity and polarized extension in the cochlea in the mouse. Here we show, using mice with null mutations in two mammalian Dishevelled homologs, Dvl1 and Dvl2, that during neurulation a homologous mammalian PCP pathway regulates concomitant lengthening and narrowing of the neural plate, a morphogenetic process defined as convergent extension. Dvl2 genetically interacts with Loop-tail, a point mutation in the mammalian PCP gene Vangl2, during neurulation. By generating Dvl2 BAC (bacterial artificial chromosome) transgenes and introducing different domain deletions and a point mutation identical to the dsh1 allele in fly, we further demonstrated a high degree of conservation between Dvl function in mammalian convergent extension and the PCP pathway in fly. In the neuroepithelium of neurulating embryos, Dvl2 shows DEP domain-dependent membrane localization, a pre-requisite for its involvement in convergent extension. Intriguing, the Loop-tail mutation that disrupts both convergent extension in the neuroepithelium and PCP in the cochlea does not disrupt Dvl2 membrane distribution in the neuroepithelium, in contrast to its drastic effect on Dvl2 localization in the cochlea. These results are discussed in light of recent models on PCP and convergent extension.

In vitro growth and differentiation of mammalian sensory hair cell progenitors: a requirement for EGF and periotic mesenchyme

The sensory hair cells and supporting cells of the organ of Corti are generated by a precise program of coordinated cell division and differentiation. Since no regeneration occurs in the mature organ of Corti, loss of hair cells leads to deafness. To investigate the molecular basis of hair cell differentiation and their lack of regeneration, we have established a dissociated cell culture system in which sensory hair cells and supporting cells can be generated from mitotic precursors. By incorporating a Math1-GFP transgene expressed exclusively in hair cells, we have used this system to characterize the conditions required for the growth and differentiation of hair cells in culture. These conditions include a requirement for epidermal growth factor, as well as the presence of periotic mesenchymal cells. Lastly, we show that early postnatal cochlear tissue also contains cells that can divide and generate new sensory hair cells in vitro.

Math1-driven GFP expression in the developing nervous system of transgenic mice

Math1 is a bHLH transcription factor expressed in neural progenitor cells in multiple regions of the nervous system. Previously we identified a Math1 enhancer that directs expression of reporter genes in a Math1 specific pattern [Development 127 (2000) 1185]. We have used a portion of this enhancer to drive expression of a nuclear GFP reporter in the Math1 lineage in transgenic mice. In this transgenic mouse strain, GFP is expressed in Math1 domains in the (1). developing spinal cord in progenitors to dI1 dorsal interneurons, (2). granule-cell progenitors in the developing cerebellum, (3). Merkel cells in the skin, and (4). hair cells in the developing vestibular and auditory systems. Furthermore, non-Math1 related expression is detected that is likely due to the absence of inhibitory regulatory sequences from the transgene. These expression domains include (1). the apical ectodermal ridge in developing limbs, (2). post-mitotic cells in the developing cortex and spinal cord, (3). the dentate gyrus, (4). retina, and (5). olfactory epithelium. Because GFP marks specific neuronal cell types in living tissue, this transgenic strain is a powerful tool for future studies on the development and electrophysiological properties of distinct cell types in the central nervous system and in sensory systems.

Progressive hearing loss in mice lacking the cyclin-dependent kinase inhibitor Ink4d

Maintenance of the post-mitotic state in the post-natal mammalian brain is an active process that requires the cyclin-dependent kinase inhibitors (CKIs) p19Ink4d (Ink4d) and p27Kip1 (Kip1). In animals with targeted deletions of both Ink4d and Kip1, terminally differentiated, post-mitotic neurons are observed to re-enter the cell cycle, divide and undergo apoptosis. However, when either Ink4d or Kip1 alone are deleted, the post-mitotic state is maintained, suggesting a redundant role for these genes in mature neurons. In the organ of Corti--the auditory sensory epithelium of mammals--sensory hair cells and supporting cells become post-mitotic during embryogenesis and remain quiescent for the life of the animal. When lost as a result of environmental insult or genetic abnormality, hair cells do not regenerate, and this loss is a common cause of deafness in humans. Here, we report that targeted deletion of Ink4d alone is sufficient to disrupt the maintenance of the post-mitotic state of sensory hair cells in post-natal mice. In Ink4d-/- animals, hair cells are observed to aberrantly re-enter the cell cycle and subsequently undergo apoptosis, resulting in progressive hearing loss. Our results identify a novel mechanism underlying a non-syndromic form of progressive hearing loss in mice.

The role of Math1 in inner ear development: Uncoupling the establishment of the sensory primordium from hair cell fate determination

During embryonic development of the inner ear, the sensory primordium that gives rise to the organ of Corti from within the cochlear epithelium is patterned into a stereotyped array of inner and outer sensory hair cells separated from each other by non-sensory supporting cells. Math1, a close homolog of the Drosophila proneural gene atonal, has been found to be both necessary and sufficient for the production of hair cells in the mouse inner ear. Our results indicate that Math1 is not required to establish the postmitotic sensory primordium from which the cells of the organ of Corti arise, but instead is limited to a role in the selection and/or differentiation of sensory hair cells from within the established primordium. This is based on the observation that Math1 is only expressed after the appearance of a zone of non-proliferating cells that delineates the sensory primordium within the cochlear anlage. The expression of Math1 is limited to a subpopulation of cells within the sensory primordium that appear to differentiate exclusively into hair cells as the sensory epithelium matures and elongates through a process that probably involves radial intercalation of cells. Furthermore, mutation of Math1 does not affect the establishment of this postmitotic sensory primordium, even though the subsequent generation of hair cells is blocked in these mutants. Finally, in Math1 mutant embryos, a subpopulation of the cells within the sensory epithelium undergo apoptosis in a temporal gradient similar to the basal-to-apical gradient of hair cell differentiation that occurs in the cochlea of wild-type animals.

Deafness Due to Degeneration of Cochlear Neurons in Caspase-3-Deficient Mice

Mice that lack caspase-3, which functions in apoptosis, were generated by gene targeting and shown to undergo hearing loss. The ABR threshold of the caspase-3(-/-) mice was significantly elevated compared to that of caspase-3(+/+) mice at 15 days of age and was progressively elevated further by 30 days. Distortion product otoacoustic emissions were not detectable in caspase-3(-/-) mice at 15 days of age. Caspase-3(-/-) mice exhibited marked degeneration of spiral ganglion neurons and a loss of inner and outer hair cells in the cochlea at 30 days of age, although no such changes were apparent at 15 days. The degenerating neurons manifested features, including cytoplasmic vacuolization, distinct from those characteristic of apoptosis. Spiral ganglion neurons and cochlear hair cells thus appear to require caspase-3 for survival but not for initial development. The mapping of both the human caspase-3 gene and the locus responsible for an autosomal dominant, nonsyndromic form of hearing loss (DFNA24) to chromosome 4q35 suggests that the caspase-3(-/-) mice may represent a model of this human condition.

p27(Kip1) links cell proliferation to morphogenesis in the developing organ of Corti

Strict control of cellular proliferation is required to shape the complex structures of the developing embryo. The organ of Corti, the auditory neuroepithelium of the inner ear in mammals, consists of two types of terminally differentiated mechanosensory hair cells and at least four types of supporting cells arrayed precisely along the length of the spiral cochlea. In mice, the progenitors of greater than 80% of both hair cells and supporting cells undergo their terminal division between embryonic day 13 (E13) and E14. As in humans, these cells persist in a non-proliferative state throughout the adult life of the animal. Here we report that the correct timing of cell cycle withdrawal in the developing organ of Corti requires p27(Kip1), a cyclin-dependent kinase inhibitor that functions as an inhibitor of cell cycle progression. p27(Kip1) expression is induced in the primordial organ of Corti between E12 and E14, correlating with the cessation of cell division of the progenitors of the hair cells and supporting cells. In wild-type animals, p27(Kip1) expression is downregulated during subsequent hair cell differentiation, but it persists at high levels in differentiated supporting cells of the mature organ of Corti. In mice with a targeted deletion of the p27(Kip1) gene, proliferation of the sensory cell progenitors continues after E14, leading to the appearance of supernumerary hair cells and supporting cells. In the absence of p27(Kip1), mitotically active cells are still observed in the organ of Corti of postnatal day 6 animals, suggesting that the persistence of p27(Kip1) expression in mature supporting cells may contribute to the maintenance of quiescence in this tissue and, possibly, to its inability to regenerate. Homozygous mutant mice are severely hearing impaired. Thus, p27(Kip1) provides a link between developmental control of cell proliferation and the morphological development of the inner ear.

The cyclin-dependent kinase-activating kinase (CAK) assembly factor, MAT1, targets and enhances CAK activity on the POU domains of octamer transcription factors

Octamer binding transcription factors (Oct factors) play important roles in activation of transcription of various genes but, in some cases, require cofactors that interact with the DNA binding (POU) domain. In the present study, a yeast two-hybrid screen with the Oct-1 POU domain as a bait identified MAT1 as a POU domain-binding protein. MAT1 is known to be required for the assembly of cyclin-dependent kinase (CDK)-activating kinase (CAK), which is functionally associated with the general transcription factor IIH (TFIIH). Further analyses showed that MAT1 interacts with POU domains of Oct-1, Oct-2, and Oct-3 in vitro in a DNA-independent manner. MAT1-containing TFIIH was also shown to interact with POU domains of Oct-1 and Oct-2. MAT1 is shown to enhance the ability of a recombinant CDK7-cyclin H complex (bipartite CAK) to phosphorylate isolated POU domains, intact Oct-1, and the C-terminal domain of RNA polymerase II, but not the originally defined substrate, CDK2. Phosphopeptide mapping indicates that the site (Ser385) of a mitosis-specific phosphorylation that inhibits Oct-1 binding to DNA is not phosphorylated by CAK. However, one CAK-phosphorylated phosphopeptide comigrates with a Cdc2-phosphorylated phosphopeptide previously shown to be mitosis-specific, suggesting that, in vitro, CAK is able to phosphorylate at least one site that is also phosphorylated in vivo. These results suggest (i) that interactions between POU domains and MAT1 can target CAK to Oct factors and result in their phosphorylation, (ii) that MAT1 not only functions as a CAK assembly factor but also acts to alter the spectrum of CAK substrates, and (iii) that a POU-MAT1 interaction may play a role in the recruitment of TFIIH to the preinitiation complex or in subsequent initiation and elongation reactions.

Mitotic regulation of TFIID: inhibition of activator-dependent transcription and changes in subcellular localization

Mitosis in higher eukaryotes is accompanied by a general inhibition of transcription. To begin to understand the mechanisms underlying this inhibition we have examined the behavior of the general transcription factor TFIID during mitosis. Immunocytochemistry and subcellular fractionation studies indicate that the majority of TFIID is displaced from the disassembling prophase nucleus to the mitotic cytoplasm around the time of nuclear envelope breakdown. However, a subpopulation of TFIID remains associated tightly with the condensed mitotic chromosomes. Metabolic labeling of mitotic cells revealed that several subunits of TFIID undergo mitosis-specific phosphorylation, but in spite of these changes, the TFIID complex remains intact. Functional analysis of purified TFIID from mitotic cells shows that phosphorylated forms are unable to direct activator-dependent transcription, but that this activity is restored upon dephosphorylation. These results demonstrate that TFIID regulation by phosphorylation is likely to have an important role in mitotic inhibition of RNA polymerase II transcription. In addition, they suggest a mechanism for regulating gene expression through the selective disruption of polymerase II promoter structures during mitosis.

The lamin B receptor of the inner nuclear membrane undergoes mitosis-specific phosphorylation and is a substrate for p34cdc2-type protein kinase

The lamin B receptor (LBR) is an integral protein of the inner nuclear membrane that interacts with lamin B in vitro. If contains a 204-amino acid nucleoplasmic amino-terminal domain and a hydrophobic carboxyl-terminal domain with eight putative transmembrane segments. We found cell cycle-dependent phosphorylation of LBR using phosphoamino acid analysis and phosphopeptide mapping of in vivo 32P-labeled LBR immunoprecipitated from chicken cells in interphase and arrested in mitosis. LBR was phosphorylated only on serine residues in interphase and on serine and threonine residues in mitosis. Some serine residues phosphorylated in interphase were not phosphorylated in mitosis. To identify a threonine residue specifically phosphorylated in mitosis and the responsible protein kinase, wild-type and mutant LBR nucleoplasmic domain fusion proteins were phosphorylated in vitro by p34cdc2-type protein kinase. Comparisons of phosphopeptide maps to those of in vivo 32P-labeled mitotic LBR showed that Thr188 is likely to be phosphorylated by this enzyme during mitosis. These phosphorylation/dephosphorylation events may be responsible for some of the changes in the interaction between the nuclear lamina and the inner nuclear membrane that occur during mitosis.

Mitotic phosphorylation of the Oct-1 homeodomain and regulation of Oct-1 DNA binding activity

Oct-1 is a transcription factor involved in the cell cycle regulation of histone H2B gene transcription and in the transcription of other cellular housekeeping genes. Oct-1 is hyperphosphorylated as cells enter mitosis, and mitosis-specific phosphorylation is reversed as cells exit mitosis. A mitosis-specific phosphorylation site in the homeodomain of Oct-1 was phosphorylated in vitro by protein kinase A. Phosphorylation of this site correlated with inhibition of Oct-1 DNA binding activity in vivo and in vitro. The inhibition of Oct-1 DNA binding during mitosis suggests a mechanism by which the general inhibition of transcription during mitosis might occur.

Differential phosphorylation of the transcription factor Oct1 during the cell cycle

Orderly progression through the somatic cell division cycle is accompanied by phase-specific transcription of a variety of different genes. During S phase, transcription of mammalian histone H2B genes requires a specific promoter element and its cognate transcription factor Oct1 (OTF1). A possible mechanism for regulating histone H2B transcription during the cell cycle is direct modulation of Oct1 activity by phase-specific posttranslational modifications. Analysis of Oct1 during progression through the cell cycle revealed a complex temporal program of phosphorylation. A p34cdc2-related protein kinase that is active during mitosis may be responsible for one mitotic phosphorylation of Oct1. However, the temporally controlled appearance of Oct1 phosphopeptides suggests the involvement of multiple kinases and phosphatases. These results support the idea that cell cycle-regulated transcription factors may be direct substrates for phase-specific regulatory enzymes.

Development and hormone regulation of androgen receptor levels in the sexually dimorphic larynx of Xenopus laevis

Development of the sexually dimorphic larynx in African clawed frogs is controlled by secretion of androgenic steroids (D. Sassoon and D. Kelley, 1986, Amer. J. Anat. 177, 457-472). Adult laryngeal muscle shows high levels of androgen binding relative to other skeletal muscles and binding activity in males is three times that in females (N. Segil, L. Silverman, and D. Kelley, 1987, Gen. Comp. Endocrinol. 66, 95-101). To determine when androgen sensitivity and sex differences arise, we assayed [3H]dihydrotestosterone (DHT) binding activity in larynges from metamorphic and postmetamorphic male and female frogs. Scatchard analyses indicate that DHT binds to a saturable component with high affinity. At metamorphosis, male and female juveniles have average binding levels of 262 and 269 fmoles/mg protein, respectively, approximately 7 to 20 times their adult values. At 3 months postmetamorphosis (PM), sexually dimorphic binding levels are observed. Binding activity declines gradually in females from metamorphosis to 9 months PM. In males, levels of binding activity remain high throughout the first 6 months PM and then decrease to near adult levels by 9 months PM. Administration of exogenous DHT to 3 months PM juveniles decreases average binding activity from 180 (male) or 74 fmoles/mg (female) to 33.5 fmoles/mg in both sexes. Testosterone has a less pronounced effect on binding activity in males than DHT and is ineffective in females. We conclude that sexually dimorphic adult levels of androgen binding in larynx arise by differential decrease from initially high, sexually monomorphic levels and that high titers of circulating androgens normally present by 6 months PM in males are responsible for the marked decrease in binding activity observed during laryngeal development.

Enolase isoenzymes in adult and developing Xenopus laevis and characterization of a cloned enolase sequence

As part of a study of glycolysis during early development we have examined the pattern of expression of enolase isoenzymes in Xenopus laevis. In addition, the nucleotide sequence of a cDNA clone coding for the complete amino acid sequence of one enolase gene (ENO1) in X. laevis was determined. X. laevis ENO1 shows highest homology to mammalian non-neuronal enolase. Analysis of enolase isoenzymes in X. laevis by non-denaturing electrophoresis on cellulose acetate strips revealed five isoenzymes. One form was present in all tissues tested, two additional forms were expressed in oocytes, embryos, adult liver and adult brain, and two further forms were restricted to larval and adult muscle. Since enolase is a dimer, three different monomers (gene products) could account for the observed number of isoenzymes. This pattern of enolase isoenzyme expression in X. laevis differs from that of birds and mammals. In birds and mammals the most acidic form is neuron-specific and there is only one major isoenzyme expressed in the liver. RNAase protection experiments showed the presence of ENO1 mRNA in oocytes, liver and muscle, suggesting that it codes for a non-tissue-restricted isoenzyme. ENO1 mRNA concentrations are high in early oocytes, decrease during oogenesis and decrease further after fertilization. Enolase protein, however, is maintained at high concentrations throughout this period.

The nervus terminalis ganglion in Anguilla rostrata: an immunocytochemical and HRP histochemical analysis

Immunocytochemistry and retrograde horseradish peroxidase (HRP) transport were used to study the ganglion of the nervus terminalis in the American eel, Anguilla rostrata. Luteinizing hormone releasing hormone (LHRH) like immunoreactivity was found in large, ganglion-like cells located ventromedially at the junction of the telencephalon and olfactory bulb and in fibers within the retina and olfactory epithelium. HRP transport from the retina demonstrated direct connections with both the ipsi- and contralateral populations of these ganglion-like cells. Given the well-documented role of both olfaction and vision during migratory and reproductive phases of the life cycle of eels, the robust nature of a nervus terminalis system in these fish may present a unique opportunity to study the behavioral correlates of structure-function organization in a discrete population of ganglion-like cells.

Androgen-binding levels in a sexually dimorphic muscle of Xenopus laevis

The larynx of adult South African clawed frogs, Xenopus laevis, is larger in males than in females and hypertrophies in adult females and juveniles in response to androgen. Sexual dimorphism and androgen sensitivity suggest that the larynx is a testosterone target tissue. Saturation analysis of androgen (R1881) binding in laryngeal cytosol revealed an approximately threefold quantitative difference between male and female androgen-binding levels (36.4 vs 11.5 fm/mg protein). By contrast, as measured by one-point assay, androgen-binding levels in thigh muscle of either males or females were between 0 and 4 fm/mg protein with no apparent sex difference. Competition studies indicated that dihydrotestosterone was the most effective competitor for R1881 binding activity in the larynx. Saturation analysis showed the binding activity to be saturable and of high affinity (apparent Kd 0.46 nM in the male and 0.38 nM in the female). After 1 month of testosterone treatment, female binding levels averaged 16.6 fm/mg protein with a Kd of 0.49 nM, within the range for normal females. In males castrated for 4 months, binding levels were 52 fm/mg protein. After 1 year of castration, binding levels were 25 fm/mg protein. We conclude that laryngeal muscle is an androgen target tissue with sexually dimorphic levels of binding in adults.

Pyruvate kinase isozymes in oocytes and embryos from the frog Xenopus laevis

1. The kinetic characteristics of pyruvate kinase isozymes from oocytes, embryos, liver and skeletal muscle from the clawed frog Xenopus laevis were measured in cell extracts. 2. The muscle and liver isozymes display Michaelis-Menten kinetics with Kms for phosphoenolpyruvate (PEP) of 0.02 and 0.05 mM, respectively. 3. Pyruvate kinase from oocytes and embryos displays cooperative kinetics for PEP with a Km of about 0.15 mM; the kinetics become hyperbolic and the Km for PEP is reduced to 0.05 mM in the presence of microM concentrations of fructose-1,6-bisphosphate. 4. These data serve to characterize pyruvate kinase activity in oocytes and embryos and the kinetics are compared to mammalian pyruvate kinase isozymes.

Androgen binding in the brain and electric organ of a mormyrid fish

The mormyrid fish of Africa produce a weak electric pulse called an Electric Organ Discharge (EOD) that functions in electrical guidance and communication. The EOD waveform describes the appearance of a single pulse which is produced by the electric organ's excitable cells, the electrocytes. For some species, there is a sex difference in the appearance and duration of the EOD waveform, which is under the control of gonadal steroid hormones. We now show, using biochemical techniques, that the steroid-sensitivity of the myogenic electric organ correlates with the presence of comparatively high levels of androgen-binding activity in the cytosol of electrocytes. The EOD rhythm describes the rate at which the electric organ fires and is under the control of a central electromotor pathway. Sex differences have also been described for the EOD rhythm. Using steroid autoradiographic techniques, we found uptake of tritium-labelled dihydrotestosterone (3H-DHT) by cells within the reticular formation that lie adjacent to the medullary 'relay nucleus' which innervates the spinal electromotoneurons that excite the electric organ. However, no DHT-binding was observed in the relay or electromotor nuclei. Steroid-concentrating cells were also found in several other brainstem regions, the hypothalamus, and the thalamus. In particular, a group of DHT-concentrating, motoneuron-like cells were observed in the caudal medulla and were identified as a swimbladder or sonic motor nucleus. The biochemical data suggest that the electric organ has evolved a sensitivity to gonadal steroid hormones that may underlie the development of known sex differences in the EOD waveform. The autoradiographic results suggest that if steroids do affect the development of sex differences in the EOD rhythm, it is at some level removed from known spinal and medullary electromotor nuclei.