TMIE Defines Pore and Gating Properties of the Mechanotransduction Channel of Mammalian Cochlear Hair Cells

Abstract

TMC1 and TMC2 (TMC1/2) have been proposed to form the pore of the mechanotransduction channel of cochlear hair cells. Here, we show that TMC1/2 cannot form mechanotransduction channels in cochlear hair cells without TMIE. TMIE binds to TMC1/2, and a TMIE mutation that perturbs TMC1/2 binding abolishes mechanotransduction. N-terminal TMIE deletions affect the response of the mechanotransduction channel to mechanical force. Similar to mechanically gated TREK channels, the C-terminal cytoplasmic TMIE domain contains charged amino acids that mediate binding to phospholipids, including PIP₂. TMIE point mutations in the C terminus that are linked to deafness disrupt phospholipid binding, sensitize the channel to PIP₂ depletion from hair cells, and alter the channel's unitary conductance and ion selectivity. We conclude that TMIE is a subunit of the cochlear mechanotransduction channel and that channel function is regulated by a phospholipid-sensing domain in TMIE with similarity to those in other mechanically gated ion channels.

Figure 1.. Effects of TMIE on TMC1-HA and TMC2-MYC localization.

(A) Diagram of a hair cell and the mechanotransduction complex at tip links. (B) Schematic of generation of mouse strains with epitope-tag endogenous TMC1 and TMC2. (C) OHCs and IHCs from P7 Tmc1^HA/HA mice immunostained in whole mounts for HA (green) and phalloidin (red) to label stereocilia in the presence (Tmie^+/+;Tmc1^HA/HA, left panels) or absence (Tmie^−/−;Tmc1^HA/HA, right panels) of Tmie. (D) OHCs and IHCs from P3 Tmc2^MYC/MYC mice immunostained in whole mounts for MYC (green) and phalloidin (red). (E) Cryosection from P7 Tmc1^HA/HA cochlea immunostained for HA (green) and DAPI (blue). Scale bar in C upper panel: 5 μm; applies to top two rows. Scale bars: 5 μm. See also Fig. S1.

Figure 2.. TMC1 and TMC2 do not affect TMIE localization.

(A) Schematic of generation of mouse strains with epitope-tag endogenous TMIE. (B) Auditory brainstem response (ABR) thresholds in response to click and pure tone stimuli for Tmie^HA/+ (n=3) and Tmie^HA/HA (n=3) mice at ~5 weeks of age. (C) Mechanotransduction currents in OHCs from wild-type and Tmie^HA/HA mice at P6–7 in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm (100 nm steps). (D) Current/displacement plot from similar data as shown in (C) (mean ± SEM). (E) Cochlear whole mounts from P4 Tmie^HA/HA mice immunostained for HA (green) and phalloidin (red) to label OHCs. (F) Cochlear whole mounts from P3 Tmie^3XMYC/3XMYC mice immunostained for MYC (green) and phalloidin (red). (G) Cochlear whole mounts from P3 Tmc1^+/+; Tmc2^+/−;Tmie^HA/HA mice immunostained for HA (green) and phalloidin (red). (H) Cochlear whole mounts from P3 Tmc1^dn/dn;Tmc2^−/−;Tmie^HA/HA mice immunostained for HA (green) and phalloidin (red). Scale bar in F: 5 μm.

Figure 3.. TMIE is essential for TMC1/2 function in hair cells.

(A) Diagram of injectoporation procedure to express genes in hair cells (Xiong et al 2014). (B) Epitope-tagged TMC1/2 and TMIE constructs. (C) Mechanotransduction currents in OHCs from Tmc1^dn/dn;Tmc2^−/− mice at P3 + 1 day in vitro (DIV) after injectoporation of various constructs. Currents are in response to 10 ms hair bundle deflections from −400 nm to 1000 nm. (D) Current/displacement plot from data as in (C) (mean ± SEM; *, p<0.05 (TMC1 vs TMC2)). (E) Mechanotransduction currents in OHCs from Tmie^−/− mice at P3 + 1 DIV after injectoporation of various constructs. Currents are in response to 10 ms hair bundle deflections from −400 nm to 1000 nm. Inset shows P3 + 1 DIV injectoporated OHCs immunostained for HA (green, upper row) or MYC (green, lower row) and phalloidin (red). (F) Current/displacement plot from data as in (E) (mean ± SEM). (G) Mechanotransduction currents in OHCs from Tmie^−/− (left) and Tmc1^dn/dn;Tmc2^−/− mice (right) at P3 + 1 DIV after injectoporation of TMIE-HA. Inset shows P3 + 1 DIV injectoporated hair cells immunostained for HA (green) and phalloidin (red). (H) Current/displacement plot from data as in (G) (mean ± SEM). Scale bars: 5 μm. See also Fig. S2.

Figure 4.. TMIE binds to TMC1 and TMC2 and N-terminal deletions affect TMIE function.

(A, E) Schematic depicting constructs used for CoIP experiments. (B–D) HEK293 cells were transfected with the constructs indicated on top of each panel. Immunoprecipitations (IP) were carried out with HA- (B) or Flag- (C,D) conjugated agarose beads, followed by Western blotting to detect epitope-tagged proteins (upper rows, CoIP; middle rows, IP; lower rows, input). Molecular weights of proteins (in kDa) indicated left of blots. (F) CoIP experiment from HEK293 cells using constructs indicated at top of panel. IP carried out with HA-conjugated agarose beads, followed by Western blotting to detect tagged proteins (upper panels, CoIP; middle row, IP; lower rows, input). (G) Quantification of CoIP results from 4 independent experiments. Binding of TMIE constructs to MYC-TMC1 was normalized to full-length TMIE-HA values (mean ± SEM; n.s., not significant). (H) Examples of OHCs from P3+1 DIV wild-type mice injectoporated with TMIE constructs and immunostained for HA (green) and phalloidin (red). (I) Quantification of stereocilia expression for TMIE constructs using fluorescent intensities from data as in (H). (mean ± SEM; expressed as a ratio of stereocilia intensity to cell body intensity; n.s., not significant). Number of cells: TMIE-HA, n=13; TMIE-11–153-HA, n=16; TMIE-28–153-HA, n=11. (J) Mechanotransduction currents in OHCs from Tmie^−/− mice at P3 + 1 DIV after injectoporation with indicated constructs. Currents are in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm. (K, L) Current/displacement plots and probability/displacement plots from similar data as in (J) (mean ± SEM; **, p<0.01; TMIE-28153-HA vs TMIE-HA).

Figure 5.. A region in the TMIE C-terminal cytoplasmic domain is required for binding to TMC1 and for mechanotransduction.

(A) Schematic depicting epitope-tagged constructs. (B–C) HEK293 cells were transfected with the constructs indicated on top of each panel. IPs were carried out with HA-conjugated agarose beads, followed by Western blotting to detect tagged proteins (upper panels, CoIP; middle row, IP; lower rows, input). Molecular weights of proteins (in kDa) are indicated. (D) Quantification of CoIP results from a minimum of 3 independent experiments each. Binding of TMIE constructs to MYC-TMC1 normalized to wild-type TMIE-HA values (mean ± SEM; n.s., not significant, **, p<0.01, ****, p<0.0001). (E) Examples of OHCs from P3+1 DIV wild-type mice injectoporated with TMIEdel80–100-HA and immunostained for HA (green) and phalloidin (red). (F) Quantification of stereocilia expression for TMIEdel80–100-HA using fluorescent intensities from data as in (E) (mean ± SEM; expressed as a ratio of stereocilia intensity to cell body intensity). Numbers of cells: TMIE-HA, n=5; TMIEdel80–100-HA, n=4. (G) Mechanotransduction currents in OHCs from Tmie^−/− mice at P3 + 1 DIV after injectoporation of indicated constructs. Currents are in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm. (H) Current/displacement plot from similar data as in (G) (mean ± SEM). Numbers of cells: TMIE-HA, n=5; TMIEdel80–100-HA, n=7. Scale bar in E: 5 μm.

Figure 6.. Analysis of deafness-associated point mutations in TMIE.

(A) Schematic depicting constructs containing TMIE point mutations associated with deafness (Naz et al., 2002; Zhao et al., 2014). (B) HEK293 cells were transfected with the constructs indicated on top to perform CoIP experiments. IPs were carried out with HA-conjugated agarose beads, followed by Western blotting to detect tagged proteins (upper panels, CoIP; middle row, IP; lower rows, input). Molecular weights of proteins (in kDa) are indicated. (C) Examples of OHCs from P3+1 DIV wild-type mice injectoporated with TMIE-R82C-HA (top row) and TMIE-R85W-HA (bottom row) and immunostained for HA (green) and phalloidin (red). (D) Quantification of stereocilia expression using fluorescent intensities from data as in (C) (mean ± SEM; expressed as a ratio of stereocilia intensity to cell body intensity). Number of cells: TMIE-HA, n=13; TMIE-R82C-HA, n=7; TMIE-R85W-HA, n=17. (E) Quantification of CoIP results from (B). Quantification from a minimum of 5 independent experiments each. Binding of TMIE constructs to MYC-TMC1 normalized to wild-type TMIE-HA values (mean ± SEM; n.s., not significant) (F) Cochlear whole mounts from P5 Tmie^+/+;Tmc1^HA/HA (left) and Tmie^R82C/R82C;Tmc1^HA/HA mice (right) stained for HA (green) and phalloidin (red) to reveal localization of TMC1 in OHCs. (G) Cochlear whole mounts from P3 Tmie^+/+;Tmc2^Myc/Myc (left) and Tmie^R82C/R82C;Tmc2^Myc/Myc mice (right) stained for MYC (green) and phalloidin (red) to reveal localization of TMC2 in OHCs or IHCs. Scale bars: 5 μm. See also Fig. S3.

Figure 7.. Point mutations in the TMIE cytoplasmic domain affect pore properties of the mechanotransduction channel in OHCs.

(A) Mechanotransduction currents in OHCs from Tmie^R82C/+ and Tmie^R82C/R82C mice at P5 in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm. (B,C) Current/displacement plot and open probability/displacement plot from similar data as in (A) (mean ± SEM). (D) Single channel events elicited by 300 nm stereocilia deflections in OHCs from P3–4 Tmie^+/+ (left) and Tmie^R82C/R82C (right) mice (C = closed state, O = open state). Ensemble averages of 15 traces at bottom. (E) Amplitude histograms generated from the fourth trace in (D). Gaussian fits of the two peaks in the histograms determine a single-channel current, which is indicated in the panel. (F) Histograms of open-time for single-channel events for Tmie^+/+ (left) and Tmie^R82C/R82C (right) from data as in (D). Curves are single-exponential fits with time constant Tau of 2.08 ms for Tmie^+/+ (298 events from 12 cells, left) and 1.94 for Tmie^R82C/R82C (335 events from 13 cells, right). (G) Summary plot of single-channel currents for Tmie^+/+ and Tmie^R82/R82C cells. Average single-channel current was reduced from 4.32 ± 0.12 pA (n=72) to 3.40 ± 0.13 pA (n=50) (mean ± SEM; ***, p<0.001). (H) Experiments to examine reversal potential and relative Ca²⁺ permeability (P_Ca/P_Cs). Hair cell membrane was held at various potentials (in 20 mV increments from −89 mV to +111 mV), and the hair bundle was mechanically stimulated to elicit currents at each holding potential. Top, mechanical stimulus paradigm; bottom, mechanotransduction currents in OHCs from Tmie^R82C/+; Tmc1^dn/dn and Tmie^R82C/R82C; Tmc1^dn/dn mice at P5. (I). Left: Current/voltage plot averaged at various membrane potentials for data as in (H). Center: Magnified current/voltage plot indicates reversal potential as the traces cross the X-axis. Right: Current/voltage relations were normalized to maximal currents at +111mV to account for differences in current amplitude in controls and mutants. Current/voltage plots were fitted with a single-site binding model. (J) Reversal potential (R_v) plots for OHCs calculated from (H,I) (29.37±1.48 mV vs 23.71 ± 2.03 mV, Mean ± SEM, *, p<0.05) (K) Relative Ca²⁺ permeability plots calculated using the Goldman-Hodgkin-Katz equation (6.71 ± 0.70 vs 4.69 ± 0.63, Mean ± SEM, *, p<0.05). See also Figs. S4 and S5.

Figure 8.. Domains in the C-terminal cytoplasmic domain of TMIE critical for lipid binding and mechanotransduction.

(A) Schematic depicting the TMIE C-terminal cytoplasmic domain, and HA-tagged synthetic peptides used in lipid binding experiments. Mutations in red. (B) Diagram of a Membrane Lipid Strip used in (C, D) and (J). (C) Representative results from Membrane Lipid Strip assays. Membranes were incubated with proteins/peptides listed above membranes, followed by antibody detection using TMIE antibodies (full length TMIE) or HA antibodies (peptides). (D) Quantification of PIP₂ binding intensity, normalized to Phosphatidic Acid binding intensity (mean ± SEM from 3 independent experiments). (E) Examples of OHCs from P3+1 DIV wild-type mice injectoporated with TMIEdel122–142-HA and immunostained for HA (green) and phalloidin (red). (F) Quantification of stereocilia expression for TMIEdel122–142-HA using fluorescent intensities from data as in (E) (mean ± SEM; expressed as a ratio of stereocilia intensity to cell body intensity; n.s., not significant). Number of cells: TMIE-HA, n=4; TMIEdel122–142-HA, n=4. (G) Mechanotransduction currents in OHCs from Tmie^−/− mice at P3 + 1 DIV after injectoporation of TMIE constructs. Currents are in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm. (H,I) Current/displacement plots and open probability/displacement plots from similar data as in (G) (mean ± SEM; *, p<0.05). (J) Representative results from Membrane Lipid Strip assays using wildtype TMIE 80–100 and deafness-associated mutant peptides. Top: legend identifying lipids corresponding to each lipid spot. Lower panels show representative examples of binding results. Peptides listed above images. (K) Mechanotransduction currents in OHCs from Tmie^R82C/+ (left) and Tmie^R82C/R82C (right) mice at P5 in response to a set of 10 ms hair bundle deflections ranging from −400 nm to 1000 nm. Example traces are for each genotype before (control) and after 12 minute PAO treatment (Effertz et al., 2017). (L) Summary graph showing time-course of changes to mechanotransduction currents during PAO treatment for Tmie^R82C/+ and Tmie^R82C/R82C from data as in (K) (mean ± SEM; **, p<0.01). Scale bar in E: 5 μm. See also Figs. S6 and S7.