Neuroscience

Complexes of vertebrate TMC1/2 and CIB2/3 proteins form hair-cell mechanotransduction cation channels

Arnaud PJ Giese
Wei-Hsiang Weng
Katie S Kindt
Hui Ho Vanessa Chang
Jonathan S Montgomery
Evan M Ratzan
Alisha J Beirl
Roberto Aponte Rivera
Jeffrey M Lotthammer
Sanket Walujkar
Mark P Foster
Omid A Zobeiri
Jeffrey R Holt
Saima Riazuddin
Kathleen E Cullen
Marcos Sotomayor author has email address
Zubair M Ahmed author has email address

Department of Otorhinolaryngology – Head & Neck Surgery, University of Maryland School of Medicine, Baltimore, MD, USA
Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH, USA
Biophysics Graduate Program, The Ohio State University, Columbus, OH, USA
Section on Sensory Cell Development and Function, National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Bethesda, MD, USA
Department of Physiology, McGill University, Montreal, QC, Canada
Ohio State Biochemistry Program, The Ohio State University, Columbus, OH, USA
Departments of Otolaryngology and Neurology, F.M. Kirby Neurobiology Center, Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA
Department of Biomedical Engineering, McGill University, Montreal, QC, Canada
Departments of Biomedical Engineering, Neuroscience, and Otolaryngology and Head and Neck Surgery, Johns Hopkins University School of Medicine, Baltimore, MD, USA
Department of Biochemistry and Molecular Biology, University of Maryland School of Medicine, Baltimore, MD, USA
Department of Ophthalmology and Visual Sciences, University of Maryland School of Medicine, Baltimore, MD, USA

https://doi.org/10.7554/eLife.89719.2

Open access
Copyright information

Figures and data

Expression of Cib genes in the mouse inner ear and generation of Cib3 knockout mice.
(A) Maximum intensity projections of Z-stacks of confocal fluorescent images and corresponding DIC images of Cib2^KO/+ and Cib2^KO/KOcultured vestibular end organ explants imaged after exposure to 3 μM of FM 1-43 for 10 s. The samples were dissected at P5 and kept 2 days in vitro (P5 + 2div). Scale bar: 20 µm. (B) Real-time quantitative RT–PCR analysis of mRNA levels in the cochlear and vestibular sensory epithelia at P12 revealed different expression of Cib1, Cib2, Cib3, and Cib4 in the auditory and vestibular end organs. (C) Violin plots generated by gEAR portal showing the expression levels of Cib2, Cib3, Tmc1, and Tmc2 by scRNA-seq in immature, type 1, and type 2 utricular vestibular hair cells. TPM: transcript per million. (D) Gene structure of the wild-type and Cib3^KO alleles. Exons are shown in yellow. Indel in exon 4 leads to a nonsense mutation and a premature stop at the protein level (p. Asp54*) (E) Real-time quantitative RT–PCR analysis using cDNA from brain tissue of Cib2;Cib3 mutant mice was used to analyze the expression of Cib2 and Cib3. Contrary to Cib2 mRNA, Cib3 mRNA appears to be stable in the Cib2^KO/KO;Cib3^KO/KO mice and does not go through nonsense mediated decay process. The relative expression of Cib2 and Cib3 members were normalized against hprt. Asterisks indicate statistical significance: ***, p<0.001 (Student’s t-test). n.s.: not significant. Error bars represent SEM. (F) Western Blot analysis on Cib2;Cib3 mouse heart tissue was performed using CIB3 antibody. No CIB3 protein was detected in the Cib2^KO/KO;Cib3^KO/KO mice, validating our mouse model. GAPDH detection was used as a loading control.

CIB2/3 double mutant mice have profound hearing loss.
(A) ABR thresholds to broadband clicks and tone-pips with frequencies of 8 kHz, 16 kHz, 24 kHz and 32 kHz in Cib2^+/+;Cib3^KO/KO (black; n = 5) and Cib2^KO/KO;Cib3^KO/KO(red, n = 5) mice at P16. The same animals were tested with clicks and tone pips. (B) ABR pure tone traces of P32 Cib2^+/+;Cib3^KO/KO(black) and Cib2^KO/KO;Cib3^KO/KO (red) mice measured at 32 kHz. Hearing thresholds are in Decibels (dB SPL). (C) Maximum intensity projections of Z-stacks of confocal fluorescent images (right) and corresponding DIC images (left) of Cib2^KO/+;Cib3^+/+, Cib2^KO/KO;Cib3^+/+, Cib2^+/+;Cib3^KO/KOand Cib2^KO/KO;Cib3^KO/KO cultured organ of Corti explants imaged after exposure to 3 μM of FM 1-43 for 10 s. The samples were dissected at P5 and kept 2 days in vitro (P5 + 2div). Scale bar: 20 µm. (D) Maximum intensity projections of confocal Z-stacks of apical, medial and basal turns of organ of Corti of control and Cib2^KO/KO;Cib3^KO/KO mice immunostained with an anti-myosin VIIa antibody (green) and counterstained with phalloidin (red) at P18. Asterisks indicate hair cell loss. Scale bar: 10 µm.

CIB2/3 double mutant mice have vestibular dysfunction.
(A) Traces showing the open-field exploratory behavior of P60 Cib2^+/+;Cib3^+/+, Cib2^KO/+;Cib3^+/+, Cib2^+/+;Cib3^KO/KO, Cib2^KO/+;Cib3^KO/KOand Cib2^KO/KO;Cib3^KO/KO mutant mice. (B) Quantification of the number of rotations in 120 seconds (mean ± SEM), showing that, unlike Cib2^+/+;Cib3^KO/KO, and Cib2^KO/+;Cib3^KO/KOmice, Cib2^KO/KO;Cib3^KO/KO mutant mice display a circling behavior and a vestibular defect (t-test) (n = 4 for each genotype). (C) Examples of head-velocity (grey) and resultant eye-velocity traces evoked during VORd testing. Note, the eye movement is compensatory, and the trace has been inverted to facilitate comparison with head velocity. (D) VORd Gain and phase (mean ± SD) plotted as a function of frequency for Cib2^+/+;Cib3^KO/KO (n = 4), Cib2^KO/+;Cib3^KO/KO (n = 5), Cib2^KO/KO;Cib3^KO/KO (n = 4). (E) OKN gain and phase (mean ± SD) plotted as a function of frequency for Cib2^+/+; Cib3^KO/KO (n = 4), Cib2^KO/+;Cib3^KO/KO (n = 5), Cib2^KO/KO;Cib3^KO/KO(n = 4). (F) VORl gain and phase (mean ± SD) plotted as a function of frequency for Cib2^+/+;Cib3^KO/KO (n = 4), Cib2^KO/+;Cib3^KO/KO (n = 5), Cib2^KO/KO;Cib3^KO/KO (n = 4). Comparisons made with two-way ANOVA followed by Bonferroni post-hoc test for (A-D); *p < 0.05. (G) Quantification of the time mice remained on the rotating rod with increasing acceleration (mean ± SEM). Comparisons made with two-way ANOVA followed by Bonferroni post-hoc test for (F-K); *P<0.002. (Cib2^+/+;Cib3^KO/KO (n = 6), Cib2^KO/+;Cib3^KO/KO (n = 6), Cib2^KO/KO;Cib3^KO/KO (n = 6)). (H) Comparison of power spectra density of head movements in translational axes (top) and rotational axes (bottom) between wild type (Cib2^+/+;Cib3^KO/KO, black), heterozygous mutants (Cib2^KO/+;Cib3^KO/KO, blue), and homozygous mutants (Cib2^KO/KO;Cib3^KO/KO, red). The double knockout Cib2^KO/KO;Cib3^KO/KO exhibit significantly higher power than Cib2^+/+;Cib3^KO/KO and Cib2^KO/+;Cib3^KO/KO across all frequencies (0–30 Hz), in all six translational and rotational axes.

Vestibular hair cells do not have functional MET channels at rest in CIB2/3 double mutant mice.
(A) Maximum intensity projections of confocal Z-stacks of the vestibular end organs of P18 Cib2^KO/+;Cib3^KO/KOand Cib2^KO/KO;Cib3^KO/KO mutant mice immunostained with myosin VIIa antibody (green) and counterstained with phalloidin (red) and DAPI (blue). Scale bar = 50 µm. (B) Adult vestibular end organs of Cib2^+/+;Cib3^KO/KO, Cib2^KO/+;Cib3^KO/KO and Cib2^KO/KO;Cib3^KO/KO mutant mice imaged after exposure to 3 μM of FM 1-43 for 10 s. Vestibular hair cell bodies (green doted lines), and stereocilia bundles (red arrows) are shown. Scale bar: 100 µm. (C) FM 1-43FX (fixable form) labeling of saccules and utricles by IP injection at P60. Obvious reduction in signal in Cib2^KO/KO;Cib3^KO/KO, and subtle extrastriolar reduction in Cib2^+/+;Cib3^KO/KO mice was observed. Extrastriolar regions are delimited with white lines. Scale bar = 100 µm.

Both Cib2 and Cib3 are required in zebrafish for acoustic startle and MET function.
(A) Compared to controls (siblings) cib3 mutants have a normal acoustic startle response. Cib2;cib3 double mutants completely lack an acoustic startle response, while cib2 mutants have a reduced probability to startle compared to controls (n = 43 sibling, 20 cib2, 19 cib3 and 7 cib2;cib3 animals). (B) Compared to controls cib3 mutants have a normal % of hair cells that label with FM 1-43 per neuromast. In cib2;cib3 mutants no hair cells label with FM 1-43, and in cib2 mutants a significantly reduced % of hair cells label with FM 1-43 per neuromast. (C) The average intensity of FM 1-43 labeling in cib3 mutant hair cells is similar to controls. In contrast, in cib2 mutants, hair cells that label with FM 1-43, have a significantly reduced intensity compared to controls (n = 8 neuromasts per genotype in B and C). (D-G) Representative neuromasts labeled with FM 1-43 for each genotype. FM 1-43 label is overlaid onto laser scanning DIC images. (H) Schematic showing the localization of membrane-localized GCaMP6s (memGCaMP6s); the indicator used to measure Ca²⁺ influx into lateral-line hair bundles. The plane used for Ca²⁺ imaging and to determine if a hair cell is mechanosensitive is indicated by the dashed line. (I-I’) Representative example of a hair bundle imaging plane (I), along with the resulting Ca²⁺ responses (I’) from hair bundles from a control neuromast. The ROIs used to measure Ca²⁺ signals in each hair bundle are indicated in (I). (J-L) Representative examples of Ca²⁺ responses from neuromasts in cib2 (J), cib3 (K) and cib2;cib3 mutants (L). FM 1-43 imaging and behavior were performed at 5 dpf. Ca²⁺ imaging was performed at 5 or 6 dpf. A Kruskal-Wallis test was used in A; A one-way ANOVA was used in B and C. The scale bar in D and I = 5 µm. The online version of this article includes source data and the following figure supplement(s) for Figure 5. The online version of this article includes source data and the following figure supplement(s) for Figure 5. Figure supplement 1. Summary of zebrafish cib2 and cib3 alleles used in this study. Figure supplement 2. Summary of mechanosensitive Ca²⁺ responses in zebrafish.

Cib2 is required in specific subsets of hair cells in the zebrafish pLL and inner ear.
(A) Overview of Tmc requirements in pLL hair cells. Hair cells that sense posterior flow (orange) rely on Tmc2b, while hair cells that sense anterior flow (blue) can rely on Tmc2b and/or Tmc2a for mechanosensitive function. (B-C’) FM 1-43FX labeling in pLL neuromasts from sibling control (B) and cib2 mutants (C). The residual cells in cib2 mutants rely on Cib3 for mechanosensitive function. Phalloidin label can be used to link FM 1-43FX label to the orientation of pLL hair cells (C,C’; see colored asterisks). In B’, orange and blue arrows rest above hair cells that are oriented to respond to anterior and posterior flow, respectively. (D) Quantification reveals that a similar percentage of anterior and posterior responsive pLL hair cells label with FM 1-43FX in sibling controls (posterior flow: 49.7 %, anterior flow: 49.2 %, n = 9 neuromasts). In contrast, in cib2 mutants, a significantly higher percentage of posterior responsive pLL hair cells label with FM 1-43FX in cib2 mutants (posterior flow: 60.5 %, anterior flow: 39.5 %, n = 26 neuromasts). (E) Overview of Tmc and Cib requirements in the hair cells of zebrafish cristae. Short teardrop-shaped cells rely primarily on Tmc2a, while tall gourd-shaped cells rely primarily on Tmc2b and Tmc1. (F-I’) Medial cristae for each genotype. MemGCaMP6s labels all hair cells, while FM 4-64 labels hair cells with intact mechanosensitive function. In siblings and cib3 mutants, both teardrop and gourd cells label with FM 4-64 (F-F’, G-G’). In cib2;cib3 double mutants, no hair cells label with FM 4-64 (asterisks) (I-I’). In cib2 mutants many short cells fail to label with FM 4-64 (asterisks) (H-H’), (n = 13 double het sibling, 7 cib3, 14 cib2 and 4 double mutant crista). All images were acquired at 5 dpf. A Kruskal-Wallis test was used in D; P > 0.01. The scale bars in B, B’, and F’ = 5 µm.

AF2 predictions and NMR data support a clamp-like model for the human TMC1 and CIB2 complex.
(A) AF2 model of hs TMC1 in complex with hs CIB2. Side view of monomeric hs TMC1 (light purple) in complex with hs CIB2 (light sea green). N-terminal residues of TMC1 known to interact with CIB2 are in red. (B) Top left panel shows residues at the hs TMC1-NT fragment (115-117) shown to weaken interaction with CIB2 upon mutation (Liang et al., 2021) at the interface with CIB2. Additional panels show details of contacts between TMC1 and CIB2 with formation of K124:D14 and K112:E80 salt bridges, as well as hydrophobic interactions. (C) Overlay of ¹H-¹⁵N TROSY-HSQC spectra of hs [U-¹⁵N]-CIB2 (black) alone and bound to either only hs TMC1-IL (red) or both hs TMC1-IL and hs TMC1-NT (blue). NMR data were obtained in the presence of 3 mM CaCl₂. These data are also shown in Figure 7-figure supplement 7. The online version of this article includes source data and the following figure supplement(s) for Figure 7. Figure supplement 1. Overview of TMC1/2 and CIB2/3 AF2 models. Figure supplement 2. TMC sequence alignment. Figure supplement 3. CIB2 sequence alignment. Figure supplement 4. CIB3 sequence alignment. Figure supplement 5. CIB sequence alignment. Figure supplement 6. SEC of hs CIB2 and hs CIB3 either refolded alone or co-refolded with hs TMC1-IL1, and hs TMC1-NT. Figure supplement 7. ¹H-¹⁵N TROSY-HSQC spectra. Figure 7–video 1. Overview of hs TMC1 (light and dark purple) in complex with hs CIB2 (light sea green)-Ca²⁺ (green spheres). K⁺ ions (pink) were shown to be permeating the pore region between α4 and α6. Water molecules, lipids, and protein side chains were omitted for clarity.

Simulations of AF2 predictions show that human TMC1 and CIB2 complexes form cation channels.
(A) Left: AF2 model of the dimeric hs TMC1 + hs CIB2 complex. Right: A complete simulation system of hs TMC1 in complex with hs CIB2-Ca²⁺ is shown in surface representation. hs TMC1 is shown in light (monomer A) and dark (monomer B) purple, hs CIB2 in light sea green, water in transparent light blue, lipid membrane in gray lines, and K⁺ and Cl^- ions are shown as pink and green spheres, respectively. (B) The hs TMC1-NT domain explores diverse conformational space in the absence of CIB during equilibrium simulations (S3a). (C) Salt bridges K124:D14 and K112:E80 remain throughout equilibrium simulations. (D) Snapshot of amphipathic helix α0 inserting into the lipid bilayer (S4c). (E) Number of K⁺ crossings as a function of time for pores A and B of hs TMC1, hs TMC1 + hs CIB2 + Ca²⁺, and hs TMC1 + hs CIB2 in a POPC bilayer at −0.5 V, −0.25 V, and −0.125 V. Insets show top views of pore region with TMC1 (dark purple) and lipids (light gray) in surface representation (protein and lipids partially excluded for clarity). Representative images of K⁺ permeating the pore region between α4 and α6 of hs TMC1 are shown with protein in surface representation and ions as pink spheres. Conductance values calculated for K⁺ crossings. The online version of this article includes source data and the following figure supplement(s) for Figure 8. Figure supplement 1. Stability of TMC/CIB complexes, monomers, and subdomains. Figure supplement 2. BSA in hs TMC1/2 and hs CIB2/3 complexes. Figure supplement 3. Number of K⁺ crossings as a function of time for pores A and B of hs TMC1 and hs TMC2. Figure supplement 4. Conduction events in long timescale simulations and summary of predicted conductance values. Figure supplement 5. Possible combinations of TMC1, TMC2, CIB2, and CIB3 complexes. Figure 8–video 1. Side view of the hs TMC1 channel pore during a trajectory at −0.5 V (monomer B). K⁺ (pink) are seen permeating the pore from one side (S1b); hs TMC1 (monomer B) is shown in dark purple and hs CIB2 is shown in light sea green. Water molecules, lipids, hs TMC1 (monomer A), and protein side chains were omitted for clarity. Figure 8–table supplement 1. Summary of simulations. Figure 8–table supplement 2. Ion conduction.

Sign up for email alerts