Pathophysiology of human hearing loss associated with variants in myosins

Abstract

Deleterious variants of more than one hundred genes are associated with hearing loss including MYO3A, MYO6, MYO7A and MYO15A and two conventional myosins MYH9 and MYH14. Variants of MYO7A also manifest as Usher syndrome associated with dysfunction of the retina and vestibule as well as hearing loss. While the functions of MYH9 and MYH14 in the inner ear are debated, MYO3A, MYO6, MYO7A and MYO15A are expressed in inner ear hair cells along with class-I myosin MYO1C and are essential for developing and maintaining functional stereocilia on the apical surface of hair cells. Stereocilia are large, cylindrical, actin-rich protrusions functioning as biological mechanosensors to detect sound, acceleration and posture. The rigidity of stereocilia is sustained by highly crosslinked unidirectionally-oriented F-actin, which also provides a scaffold for various proteins including unconventional myosins and their cargo. Typical myosin molecules consist of an ATPase head motor domain to transmit forces to F-actin, a neck containing IQ-motifs that bind regulatory light chains and a tail region with motifs recognizing partners. Instead of long coiled-coil domains characterizing conventional myosins, the tails of unconventional myosins have various motifs to anchor or transport proteins and phospholipids along the F-actin core of a stereocilium. For these myosins, decades of studies have elucidated their biochemical properties, interacting partners in hair cells and variants associated with hearing loss. However, less is known about how myosins traffic in a stereocilium using their motor function, and how each variant correlates with a clinical condition including the severity and onset of hearing loss, mode of inheritance and presence of symptoms other than hearing loss. Here, we cover the domain structures and functions of myosins associated with hearing loss together with advances, open questions about trafficking of myosins in stereocilia and correlations between hundreds of variants in myosins annotated in ClinVar and the corresponding deafness phenotypes.

Keywords: cargo transport; hearing; hereditary deafness; myosin; stereocilia.

FIGURE 1

Functions of myosins in stereocilia of inner ear hair cells (A) Sensory epithelia in the inner ear. The cochlear sensory epithelium forms a spiral and has one row of IHCs and three rows of OHCs to detect sound stimuli (top right). Vestibular sensory epithelia are in the utricle, the saccule and three cristae ampullaris. Epithelia in the utricle and saccule detect linear acceleration in different orientations and gravity (bottom right). Epithelia in the cristae ampullaris cover angular accelerations in three-dimensional space using the fluid motion in semicircular canals perpendicular to each other (bottom left). (B) Inner hair cell schematically showing the mechanotransduction event during sound stimulation. Ion concentrations are previously reviewed (Wangemann, 2006; McPherson, 2018). The plasma membrane of cochlear hair cells is negatively charged between −45 mV and −60 mV against the perilymph, which have a gradient along the cochlear length (Purves, 2018), and has a steeper gradient against the positively charged endolymph, which is approximately +90 mV against the perilymph (Li et al., 2020). Deflection of stereocilia allows positive K⁺ and Ca²⁺ ions in the endolymph to enter the cell body through MET channels at stereocilia tips and depolarizes the plasma membrane from the resting potential. This depolarization triggers synaptic vesicle release at the base of hair cell initiating signal transmission through the afferent fiber synapses (yellow) to the cochlear nerve. Regulatory efferent fiber endings (dark blue) are connected to the afferent nerve endings in sound transducing mature inner hair cells. (C) Architecture of stereocilia. The MET channels are localized at the tips of stereocilia and connected by tip-links to the side of adjacent stereocilia of the longer row, which is referred to as the upper tip-link density (UTLD) based on the high scattering of electrons in transmission electron microscopy. Each stereocilium contains a core of tightly-packed unidirectional F-actin which narrows down and is connected to the cuticular plate. Connection between the F-actin core and the cuticular plate is supported by rootlets consisting of more tightly packed F-actin. (D) Major functions of myosins in a stereocilium. MYO7A is localized at the UTLD and involved in localization of components of the tip-link and the MET channel (and also formation of ankle-links, right panel). Some class-I myosins are reported to play a role in adaptation during sustained sound stimulation. MYO3A and MYO15A accumulate at stereocilia tips and both transport cargo essential for elongating the F-actin core. MYO6 is thought to connect the plasma membrane to the F-actin core at stereocilia tapered base and also function in the cuticular plate.

FIGURE 2

Structure of myosins and associations with human hearing loss (A) Four (sub)domains of the motor domain (underlined) illustrated using the structure of myosin subfragment-1 from Gallus gallus (PDB: 2MYS) (Rayment et al., 1993). There is a large cleft between the upper and lower 50-kDa domains. Essential and regulatory light chains are also included. (B) Magnified image of the motor domain in (A) showing three amino-acid loops crucial for ATP hydrolysis (underlined). Loop 2 connects the upper and lower 50-kDa domains. (C) Schematic cycle of ATP hydrolysis illustrated based on previous reviews (Houdusse and Sweeney, 2016; Sweeney and Holzbaur, 2018). Conformational changes causing F-actin binding, unbound from F-actin and a power stroke are tightly linked to each step of ATP hydrolysis. (D) Domain structure of conventional myosins associated with human hearing loss. MYH9 and MYH14 have a long coiled-coil to dimerize and then form actomyosin structures. The assembly competence domains are necessary for bundling in an antiparallel manner and forming a bipolar thick filament. (E) Domain structure of unconventional myosins associated (or potentially associated) with human hearing loss. Each myosin has unique motifs in the tail. MYO6 has a unique insertion of 53 residues called a “reverse gear” between the motor domain and the neck allowing MYO6 to move toward the pointed end of F-actin (Preller and Manstein, 2017). (F) Domain structure of unconventional myosins with a coiled-coil shown for comparison with (E). These myosins can walk independently as a parallel dimer (MYO5A) or an antiparallel dimer (MYO10) or form a filament resembling conventional myosins (MYO18A). MYO18A has a motor domain that lacks the ATPase activity (Taft and Latham, 2020).

FIGURE 3

Interactome and pathogenic variants of MYO3A (A) Scheme showing interacting partners of MYO3A. The N-terminal kinase domain of MYO3A can phosphorylate the kinase domain and the motor domain of MYO3A (autoinhibition). The tail has region/domains that interacts with MORN4, ESPN isoform 1 (or ESPNL) and F-actin. (B) Pathogenic or likely pathogenic MYO3A variants curated by ClinVar as of 3 December 2023. Variants are plotted on the protein domains using three categories, AD: autosomal dominant nonsyndromic hearing loss, AR: autosomal recessive nonsyndromic hearing loss, HL: hearing loss with no description on the mode of inheritance. A few studies associate three variants with autosomal dominant hearing loss although the precise mechanism of the dominant inheritance is unknown (shown with asterisk). Variants in the 3′- and 5′- untranslated regions or without clinical information are not included. Most MYO3A variants occur in the kinase domain, the motor domain and the neck region (yellow rectangle). Two noncoding variants exist in the THDI domain that interacts with ESPN isoform 1 or ESPNL (arrowheads). Variants that belong to more than one category are plotted by crosses.

FIGURE 4

Pathogenic and likely pathogenic MYO6 variants mapped on the domain structure. Most variants affect the motor domain and the neck region (yellow rectangle) except for some frameshift variants in the CBD (arrowheads). See the text for discussion of variants indicated by an arrow (p.Thr13fs) and an open arrowhead (p.Glu299Asp). Variants are obtained and classified as described for Figure 3.

FIGURE 5

Interactome, N-terminal splicing variations and pathogenic variants of MYO7A (A) Interacting partners of MYO7A. SANS and harmonin isoform b (harmonin b) bridge interactions with other partners. Serine and threonine-rich (PST) sequence of harmonin b can bind to F-actin (Grillet et al., 2009). SAH domain of MYO7A has a weak dimerization activity (Sakai et al., 2011; Liu et al., 2021). SANS, PCDH15 and CDH23 can dimerize with each other (Adato et al., 2005; Dionne et al., 2018; Jaiganesh et al., 2018). The tail of MYO7A can inhibit the motor function (autoinhibition). (B) Different N-termini of two MYO7A isoforms. The canonical isoform has an eleven amino-acid extension at the N-terminus (MYO7A-C), while the short isoform does not (MYO7A-S). (C) Mapping of pathogenic and likely pathogenic MYO7A variants. Obtained and classified as described for Figure 3 adding a category, VIS, to indicate variants associated with retinal dysfunction but not with hearing loss. Nonsense and frameshift variants in MYO7A usually result in autosomal recessive Usher syndrome or nonsyndromic hearing loss (yellow rectangles), but some are additionally associated with autosomal dominant nonsyndromic hearing loss (pink rectangle). Overlapping phenotypic categories are observed also for missense mutations (examples shown by open and closed arrowheads). Missense variants are reported for the first methionine codon of the N-terminal extension (p.Met1Val and p. Met1Ile, arrows).

FIGURE 6

Interactome, N-terminal splicing variations and pathogenic variants of MYO15A (A) Known interacting partners of MYO15A. A ternary complex of MYO15A, WHRN and EPS8 interacts with each other. Another ternary of WHRN, GPSM2 and GNAI3 interacts with MYO15A. EPS8 can bundle F-actin and also cap barbed ends. Interacting partners for the first MyTH4-FERM domain and the SH3 domain of MYO15A have not been identified. (B) Difference in the N-termini between three MYO15A isoforms. MYO15A-1 has a large 133-kDa N-terminal domain encoded by a single exon, while a short MYO15A-2 isoform does not include this sequence. A novel MYO15A-3 isoform has a small 6-kDa N-terminal extension. (C) Mapping of pathogenic and likely pathogenic variants of MYO15A. Variants are obtained and classified as described for Figure 3. Nonsense and frameshift mutations distribute along the entire length of MYO15A. Some missense variants are observed in the N-terminal domain (yellow rectangle with asterisk) and in the MyTH4-FERM and SH3 domains whose binding partners are unknown (yellow rectangle with double asterisk). One variant frameshifts the C-terminal PDZ ligand (arrowhead).

FIGURE 7

Pathogenic variants of MYH9 and MYH14. (A, B) Mapping of pathogenic and likely pathogenic variants in MYH9 and MYH14, correspondingly. Variants are obtained and classified as described for Figure 3 with additional categories for MYH9. Nonsense or frameshift mutations are only found close to the N-terminus or the C-terminus (arrows). Almost all MYH14 variants are associated with autosomal dominant hearing loss. In contrast, most variants of MYH9 are associated with MYH9-related disorders accompanying platelet disorders and nephritis (MYH9-REL). There is one cluster of inframe mutations (p.Gln1068_Leu1074dup, p. Gln1068_Leu1074del and p. Glu1084del) in the coiled-coil domain (open arrow) for MYH9-REL. One variant is associated with a complex of symptoms including abnormal platelets and facial dysmorphology (p.Asp1424Gly, arrowhead, categorized as FACE). All MYH9 missense variants associated with autosomal dominant hearing loss occur in the motor domain near the neck or in the neck or tail (yellow rectangle) and are associated also with MYH9-related disorders.