Hair-cell mechanotransduction — CDH23/PCDH15 tip link, TMC1/TMC2 MET channel with LHFPL5/TMIE/CIB2, USH1 upper tip-link density tensioning, MYO1C adaptation and inner hair cell VGLUT3/otoferlin ribbon-synapse output; CDH23/PCDH15/TMC1/TMC2/LHFPL5/TMIE/CIB2/USH1C/USH1G/MYO7A/MYO1C/SLC17A8/OTOF

Hair cells of the cochlea and of the vestibular organs convert deflection of their apical hair bundle into an electrical signal. The hair bundle is a staircase of actin-filled stereocilia. The tip of each shorter stereocilium is joined to the side of its taller neighbour by a tip link, a fine extracellular filament made of two cadherins: a parallel homodimer of cadherin 23 (CDH23) forms the upper part and a homodimer of protocadherin 15 (PCDH15) the lower part, and the two meet tip to tip. At the lower end of the tip link, PCDH15 is bound to the mechanoelectrical transduction (MET) channel complex in the stereocilium membrane. The pore of that channel is formed by the transmembrane channel-like protein TMC1, or by its paralog TMC2, which is expressed in vestibular hair cells and only transiently in neonatal cochlear hair cells. The tetraspan LHFPL5 (TMHS) and the two-pass TMIE couple PCDH15 to the pore, and the EF-hand protein CIB2 (with CIB3 in vestibular hair cells) binds the cytoplasmic side of TMC1/TMC2 and is needed for transduction. At the upper end of the tip link, CDH23 is anchored in the upper tip-link density, where the PDZ scaffold harmonin (USH1C), the scaffold SANS (USH1G) and the motor myosin VIIa (MYO7A) form a complex thought to hold the tip link under resting tension. Deflection of the bundle towards its tallest row raises tip-link tension and increases the open probability of the channels, and the resulting cation influx depolarises the cell. Sustained deflection is followed by adaptation; the slow component in vestibular hair cells is mediated by myosin-1c (MYO1C) slipping and climbing along the stereocilium actin core to reset tension. In inner hair cells of the cochlea, the receptor potential drives glutamate release at ribbon synapses onto type I spiral ganglion neurons. Synaptic vesicles are loaded with glutamate by the vesicular glutamate transporter VGLUT3 (SLC17A8), and Ca2+-triggered fusion requires otoferlin (OTOF), a multi-C2-domain protein that binds Ca2+ and the SNARE proteins and is the main Ca2+ sensor for exocytosis at this synapse. Outer hair cells use the same transduction apparatus to drive electromotile amplification. Loss-of-function variants in CDH23, PCDH15, USH1C, USH1G, MYO7A and CIB2 cause Usher syndrome type 1 or nonsyndromic deafness; variants in TMC1, LHFPL5 and TMIE cause nonsyndromic deafness; OTOF variants cause auditory neuropathy (DFNB9); and SLC17A8 variants cause DFNA25.

Derivation. Derived from the RGD Pathway Ontology record PW:0001503 through the RGD_PATHWAY_ONTOLOGY project's diagram->module route. No GO-CAM in gocams/index.tsv contains any of the RGD genes or the MET-channel subunits, and none uses GO:0050910 or GO:0050973, so the module has no gocam_associations. Of the 21 RGD human genes, 8 are modelled as participants (CDH23, PCDH15, USH1C, USH1G, MYO7A, MYO1C, SLC17A8, OTOF), 12 are excluded as bundle-building or bundle-maintaining proteins (decision 3), and POTEI is excluded as a mapping artefact (decision 6). Five MET-channel proteins that are not on the RGD list (TMC1, TMC2, LHFPL5, TMIE, CIB2) are added (decision 2). Boundary decisions. (1) Participant rule. CLAUDE.md requires that a product does some of the work of the process, not merely be needed for it. Nearly every hair-bundle protein is needed for transduction, because without a correctly built bundle no tip links form and no current flows; that is necessity evidence. The question asked for each RGD gene was which entity performs a step of detection: transmitting force (tip link), conducting the current (MET channel), setting or resetting tension on the channel (upper tip-link density and adaptation motor), or, in the output arm, loading and releasing transmitter. Comparator check (QuickGO, GO:0050910, human and mouse, October 2026): the term is carried by TMC1 and TMC2 (IMP involved_in, plus IBA), and by Pcdh15, Lhfpl5, Whrn, Ptprq, Adgrv1, Pdzd7, Strc, Atp2b2, Kcnq1, Slc12a2, Sox2, Myc, Chrna9/Chrna10 and others, mostly as acts_upstream_of_or_within from mouse phenotypes. The phenotype-based rows include transcription factors and the efferent nicotinic receptor, so the term in GOA does not mark direct participation, and its presence on a gene was not taken as a reason to add it. (2) MET channel. The RGD record calls the channel's identity unknown. Later work cached here identifies TMC1/TMC2 as the pore (PMID:30138589 "TMC1 forms the entire pore"; PMID:22105175; PMID:23871232), with LHFPL5 (PMID:23217710), TMIE (PMID:25467981) and CIB2/CIB3 (PMID:28663585, PMID:34089643) as required auxiliary subunits. Without them the module would have no transducer, so they are added. The channel is modelled as a protein complex with mechanosensitive monoatomic cation channel activity GO:0140135, following the TMC1 gene review, which records TMC1 as contributes_to GO:0140135 in cation channel complex GO:0034703 and proposes a new term "mechanoelectrical transduction channel complex" (not yet in GO). The pore subunit is a variant set: TMC1 in mature cochlear hair cells, TMC2 in vestibular hair cells and transiently in neonatal cochlear hair cells (PMID:22105175). LHFPL5 and TMIE carry no molecular function in GOA and are modelled only as active units of the complex. CIB2 is an active unit too; its gene-review core function ion channel regulator activity GO:0099106 is noted but not given a separate annoton, because the channel activity is a complex-level claim. CIB3 substitutes for or adds to CIB2 in vestibular hair cells (PMID:34089643); it is not grounded here because no CIB3 review or cached CIB3-specific paper beyond that one was read. TOMT, which is needed to traffic TMC proteins to the bundle, is a trafficking factor rather than a channel subunit and is not added. The candidate channel proteins TMEM63B, PIEZO1 (the reverse-polarity current) and TMC paralogs other than TMC1/TMC2 are not added. (3) Bundle architecture excluded. ACTB and ACTG1 (stereocilium actin core), ESPN, PLS1 and PLS3 (actin cross-linkers), MYO15A, WHRN and MYO3A (stereocilium length regulation and tip-complex delivery), PTPRQ (shaft connectors and phosphoinositide phosphatase at the stereocilium base), USH2A and ADGRV1 (transient ankle links) and MYO6 (cuticular plate and stereocilium rootlet anchoring) build or maintain the bundle in which transduction happens. Their process is inner ear receptor cell stereocilium organization GO:0060122 / auditory receptor cell stereocilium organization GO:0060088, as recorded in the WHRN, USH1C and ESPN reviews. They are context for this module, not steps, and a future stereocilium-organization module is the right home for them. Two disagreements are recorded rather than overridden. The ADGRV1 gene review lists GO:0050910 as a process of its ankle-link core function, citing reduced MET currents in Vlgr1-/- outer hair cells (PMID:17567809); that defect accompanies disorganised bundles and was read here as a developmental effect. Human PTPRQ carries GO:0050910 by IBA; mouse Ptprq mutants lose bundle integrity, which again is upstream of transduction. Both are worth a second look by the gene reviewers. USH1C harmonin-b also bundles F-actin (PMID:12485990); that role belongs with bundle architecture and is not modelled; only its upper-tip-link adaptor role is. (4) Upper tip-link density. Harmonin, SANS and MYO7A are kept as participants because the harmonin dfcr mutant loses the UTLD and has shifted channel sensitivity with a normal bundle (PMID:19447093), and MYO7A at the UTLD is proposed to generate the resting tip-link tension (PMID:21709241). Harmonin and SANS take the gene-review core function protein-macromolecule adaptor activity GO:0030674; MYO7A takes microfilament motor activity GO:0000146 (IDA, PMID:21687988; MYO7A review core). Whether MYO7A or MYO1C is the tensioning motor in cochlear hair cells is not settled (knowledge gap). (5) Adaptation. MYO1C is kept, as a separate part, because the Y61G chemical-genetic experiment blocked adaptation within seconds of inhibitor delivery in utricular hair cells (PMID:11853671), which is a test of participation, not of necessity. Its molecular function is microfilament motor activity GO:0000146 (mouse EXP, human IBA). The evidence comes from vestibular hair cells, and adaptation in cochlear hair cells is an order of magnitude faster and may use different molecules (PMID:19447093), so the part is marked optional and its cochlear role is a knowledge gap. Fast adaptation by Ca2+ entering through the channel itself is not given a separate participant. The RGD diagram edge Myo1c-PIP2 and the PTPRQ phosphoinositide gradient are noted but not modelled. (6) POTEI excluded. POTEI (UniProtKB:P0CG38) is a primate POTE ankyrin-domain protein whose C-terminal region belongs to the actin family. The rat diagram has no Pote gene, so POTEI most likely entered the human gene list through orthology mapping of rat Actb/Actg1 to an actin-like human sequence. Nothing links it to hair cells. (7) Synaptic output. The RGD record includes the inner hair cell ribbon synapse, so the two RGD genes that act there are modelled: SLC17A8 with L-glutamate uniporter activity GO:0140788 (IDA PMID:12151341 in GOA; a child of GO:0005313 used for the VGLUT family in modules/synaptic_vesicle_neurotransmitter_loading.yaml, whose glutamate_loading variant grounds only VGLUT1/VGLUT2), and OTOF with calcium ion sensor activity GO:0061891, the term the synaptic_vesicle_exocytosis module uses for synaptotagmins (OTOF GOA has calcium ion binding GO:0005509 IDA PMID:17055430; the UniProt function text calls it the key Ca2+ sensor of IHC ribbon-synapse fusion). The generic SNARE, priming and active-zone machinery is in modules/synaptic_vesicle_exocytosis.yaml and is not repeated; the IHC synapse is atypical (otoferlin rather than synaptotagmin-1/2 as sensor), so a module_ref to that module was not used as a part. The CaV1.3 channel (CACNA1D) that couples the receptor potential to release, the ribbon protein RIBEYE (CTBP2) and the postsynaptic AMPA receptors of spiral ganglion neurons are not on the RGD list and are not modelled; the connection from the MET channel to exocytosis is therefore indirect. The MYO6-OTOF interaction drawn by RGD is noted but not modelled. (8) Out of scope: outer hair cell electromotility (prestin, SLC26A5; GO:0099129), the medial olivocochlear efferent synapse (CHRNA9/CHRNA10), the tectorial membrane and otolithic membrane, hair-cell development and planar polarity, and the endolymph potassium recycling that supplies the transduction current. Cochlear versus vestibular. The transduction parts are shared by cochlear and vestibular hair cells; the differences captured are the TMC1/TMC2 pore variants and the vestibular provenance of the MYO1C evidence. The ribbon-synapse output is modelled for cochlear inner hair cells only; vestibular hair cells also have ribbon synapses, but VGLUT3 is the IHC transporter named by RGD. Gene reviews used: CDH23, TMC1, CIB2, MYO7A, USH1C and USH1G core functions. PCDH15, TMC2, LHFPL5, TMIE, MYO1C, OTOF and SLC17A8 have no review yet; their functions come from GOA and the cited literature. PANTHER family ids and PTNs are deliberately not asserted.