Cochlear specific yeast two-hybrid cDNA library constructed from mRNA isolated from P3-P7 mouse organ of Corti was constructed previously (Zhao et al., 2014) and yeast two-hybrid screening was performed following the manufacturer’s instruction (Dualsystems Biotech). In brief, Clic5 cDNA (NM_172621) was cloned into a yeast two-hybrid bait vector and introduced into yeast by transformation. Self-activation of the bait was tested and screening stringency was optimized. Yeast libraries were then transformed into yeast expressing the bait. Putative interactors were selected from among ∼1 million transformants. Positive clones were picked and re-tested for bait dependency. Plasmids from positive clones were isolated and the identity of interacting partners was determined by DNA sequencing.

HEK293 cell line was obtained from ATCC. Cells were maintained in the DMEM medium (Cat# 11965118, Thermo Fisher Scientific) supplemented with 10% heat-inactivated fetal bovine serum (Fisher Scientific), 100 U/ml penicillin and 100 μg/ml streptomycin (Fisher Scientific). Cells were grown at 37°C in a 5% CO₂ humidified atmosphere.

The coding sequence of Clic5, Clic4, Grxcr2, and taperin were amplified from mouse cochlear cDNA library. Expression of the constructs, immunoprecipitations, and western blots were carried out as described (Senften et al., 2006; Zhao et al., 2016; Liu et al., 2018). Immunoprecipitation experiments were carried out at least three times to verify the reproducibility of the data. The following antibodies were used for the experiments: anti-HA (Cat# 2367S, Cell signaling); anti-Myc (Cat# 2278S, Cell signaling); anti-Myc (Cat# sc-40, Santa Cruz); anti-GFP (Cat# sc-9996, Santa Cruz).

Grxcr2^–/– mouse (previously named as Grxcr2^D46/D46 mouse) has been described previously (Liu et al., 2018). Clic5^–/– mouse (also named as Clic5^jbg/jbg mouse) was purchased from Jackson lab (Gagnon et al., 2006). Using clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technology, 180 bp nucleotides in exon 1 of Grxcr2 were deleted from the genome of C57BL/6J mouse, resulting in a change of amino acid sequence after residue 35, a loss of 60 amino acids and remains in frame. In brief, two sgRNAs (5′-GGATGGCGTTTATGGGTCTGGGG-3′ and 5′-CAGCGGCGCCTACACTCTGGCGG-3′) were synthesized by in vitro transcription and microinjected into one-cell embryos. Genomic DNA was then collected from the offspring obtained by the embryo injections, screened using PCR and then sequenced to confirm in-frame deletion. The founder mice were then back-crossed with C57BL/6J mice for two generations. To genotype the Grxcr2^D180/D180 mice, the following primers were used: 5′-TCTTCCTACAGTGGCCGAGT-3′ and 5′-TGAATGTGAGCGAGATACCG-3′. All animal experiments were approved by Institutional Animal Care and Use Committee of The Scripps Research Institute and Indiana University School of Medicine. Both male and female mice were used in our experiment, and we did not find any sex-based differences.

Cochlear whole mount staining was carried out as described (Zhao et al., 2016; Liu et al., 2018). In brief, organ of Corti tissue was dissected and fixed in 4% PFA in Hank’s Balanced Salt Solution (HBSS) for 20 min. Samples were blocked for 20 min at room temperature in HBSS containing 5% bovine serum albumin (BSA), 1% goat serum and 0.5% Triton X-100, and then incubated overnight at 4°C with primary antibodies in HBSS containing 1% BSA and 0.1% Triton X-100. Samples were washed in HBSS and incubated 2 h at room temperature with secondary antibodies. Tissues were mounted in ProLong^® Antifade Reagents (Invitrogen). Stacked images (Z step, ∼0.17 μm; pixel size, 0.04 μm) were then captured by deconvolution microscope (Leica) using a 100 X objective (HCX PL APO 100×/1.40–0.70 OIL). Images were then deconvoluted using blind deconvolution method. Primary antibodies were as follows: anti-GRXCR2 (Cat# HPA059421, Sigma); anti-taperin (Cat# HPA020899, Sigma); anti-CLIC5 (Cat# ACL-025, Alomone labs). Additional reagents were: Alexa Fluor 488-phalloidin (Thermo Fisher Scientific), Alexa Fluor 568-phalloidin (Thermo Fisher Scientific), Alexa Fluor 647-phalloidin (Thermo Fisher Scientific), Alexa Fluor 488 goat anti-rabbit (Thermo Fisher Scientific), and Alexa Fluor 546 goat anti-rabbit (Thermo Fisher Scientific).

The experiments were performed as described (Zhao et al., 2016; Liu et al., 2018). In brief, inner ears were dissected in fixative (2.5% glutaraldehyde; 4% formaldehyde; 0.05 mM Hepes Buffer pH 7.2; 10 mM CaCl₂; 5 mM MgCl₂; 0.9% NaCl) and fixed for 1 h at RT. Samples were then dissected to remove the stria vascularis, Reissner’s membrane and tectorial membrane. Samples were post-fixed by immersion in for 1 day in the same fixative at 4°C. After fixation in 1% OsO₄ for 1 h, samples were serially dehydrated in ethanol, dried in a critical point drier (Autosamdri-815A, Tousimis), finely dissected and mounted on aluminum stubs. Samples were then coated by gold and viewed on a JEOL 7800F scanning electron microscope. At least three animals representative of each experimental paradigm were analyzed.

Auditory brainstem responses (ABRs) of mice were recorded as described (Zhao et al., 2016; Liu et al., 2018) using TDT Bioacoustic system 3 and BioSigRZ software. In brief, mice were anesthetized using the mixture of 100 mg/kg ketamine and 10 mg/kg xylazine. Electrodes were inserted under the skin at the vertex and ipsilateral ear, while a ground was inserted under the skin near the tail. The speaker was placed 5 cm away from the mouse ear. Tone stimulus is presented 21 times per second. Band-pass filtered from 300 to 3000 Hz. Averaging window was 10 ms. A total of 512 responses were averaged at each frequency and level combination. The intensity of sound stimulus was started at 90 dB SPL and decreased in 10 dB SPL stepwise to a sub-threshold level. ABR thresholds were analyzed for both ears and for a range of frequencies (for Pure Tone, 4–28 kHz). If no ABR wave was detected at maximum intensity stimulation, a nominal threshold of 90 dB was assigned.

All data are mean ± standard error of the mean. Student’s two-tailed unpaired t test or Two-way ANOVA were used to determine statistical significance (^∗, p < 0.05, ^∗∗, p < 0.01, ^∗∗∗, p < 0.001).

Chloride intracellular channel protein 5 has been linked to sensorineural hearing loss in humans and mice (Gagnon et al., 2006; Salles et al., 2014; Seco et al., 2015). However, the molecular functions of CLIC5 in hair cells are still unknown. To identify binding partners for CLIC5, an unbiased yeast-two-hybrid screening was performed using full length CLIC5 as bait. By screening a yeast-two-hybrid library constructed using RNA extracted from organ of Corti, we identified 30 positive clones from ∼1 million transformants. One positive clone expresses taperin, consistent with the previous finding that CLIC5 interacts with taperin (Salles et al., 2014; Bird et al., 2017). Interestingly, 23 positive clones express GRXCR2, suggesting a strong interaction between CLIC5 and GRXCR2. To confirm the yeast-two-hybrid data, we carried out co-immunoprecipitation (co-IP) experiments with extracts from HEK293 cells that were transfected with GFP-tagged CLIC5 and Myc-tagged GRXCR2. CLIC5-GFP was co-immunoprecipitated with Myc-GRXCR2 (Figure 1C). Correspondingly, GFP tagged GRXCR2 was co-immunoprecipitated with HA-tagged CLIC5 (Figure 1D).

Chloride intracellular channel protein 5 shows an overall 75% similarity to its paralog CLIC4. CLIC5 and CLIC4 are both highly expressed in the cochlear hair cells and concentrated at the base of stereocilia (Shen et al., 2015; Bird et al., 2016). Although both of them interact with taperin (Figure 1E and Bird et al., 2017), only CLIC5 is known to be essential for the morphogenesis of stereocilia and auditory perception (Bird et al., 2016), suggesting that in addition to TPRN, CLIC5, and CLIC4 have different binding partners and functions. Interestingly, our data shows that only CLIC5 binds strongly to GRXCR2 (Figure 1D), suggesting that the specific interaction between CLIC5 and GRXCR2 might be important for the auditory perception. To identify region(s) in GRXCR2 critical for the interaction with CLIC5, we next generated several truncated GRXCR2 constructs and their interactions with CLIC5 were evaluated by co-IP (Supplementary Figure 1A). Finally, we found that 60 amino acids from 36 to 95 in GRXCR2, which are highly conserved in mammals, are essential for its interaction with CLIC5 (Figure 1F). A previous study reported that amino acids from 121 to 140 in GRXCR2 mediate the interaction with taperin (Liu et al., 2018). Deleting 60 amino acids from 36 to 95 in GRXCR2 did not affect its binding with taperin (Supplementary Figure 1B), further confirming that CLIC5 and taperin bind to different regions of GRXCR2. To identify region(s) in CLIC5 mediating the interaction with GRXCR2, we generated domain-swapped CLIC5 and CLIC4 proteins by exchanging their N terminal 94 aa motifs (Figure 1B and Supplementary Figure 1C). Interesting, both chimeric proteins interacted with GRXCR2 (Figure 1G), suggesting that at least two regions in CLIC5 are essential for its interaction with GRXCR2.

Previous studies found that GRXCR2 is essential for the localization of taperin to the basal region of stereocilia (Liu et al., 2018). CLIC5 is also localized at the base of stereocilia and forms a complex with taperin (Gagnon et al., 2006; Salles et al., 2014). To investigate the extent to which the interaction between GRXCR2 and CLIC5 is required for their proper localization in the stereocilia, immunohistochemistry experiments were performed. To confirm the immunostaining signals obtained are specific, antibodies used in this study were validated using samples from null mutant mice. In the wild-type hair cells, we observed expression in outer hair cells (OHCs) and inner hair cells (IHCs) where CLIC5 and GRXCR2 immunostaining signals concentrated at the base of each stereocilium, outlining the shape of stereociliary bundle (Figures 2A,B). In the hair cells from Clic5^jbg/jbg mouse (referred as Clic5^–/– mouse hereafter), in which a 97 bp deletion introduces a translational frameshift and premature stop codon in Clic5 (Gagnon et al., 2006), no immunostaining signal of CLIC5 was observed, suggesting that the immunostaining signals obtained using CLIC5 antibody are specific (Figure 2A). Similarly, the GRXCR2 antibody used in this study is also specific as no immunostaining signal was observed in the hair cells from Grxcr2^D46/D46 mouse (Liu et al., 2018, referred as Grxcr2^–/– mouse hereafter), in which a 46 bp frameshift deletion creates a premature stop codon in Grxcr2 (Figure 2B).

At higher magnifications, CLIC5 localized at the base of stereocilia in wild-type OHCs and IHCs, consistent with previous results (Gagnon et al., 2006; Salles et al., 2014). In Grxcr2^–/– hair cells, CLIC5 was still concentrated at the base of stereocilia, suggesting that GRXCR2 is not essential for the localization of CLIC5 to the basal region of stereocilia during development (Figure 2C). Our previous study found that taperin is diffused along the stereocilia length and sometimes accumulated toward the distal end of the stereocilia in the Grxcr2^–/– hair cells (Liu et al., 2018). Although taperin interacts with CLIC5 at the base of stereocilia (Salles et al., 2014; Bird et al., 2016, 2017), mislocalized taperin in the Grxcr2^–/– hair cells did not affect the localization of CLIC5 in stereocilia. Similarly, GRXCR2 localized at the base of stereocilia in both wild-type and Clic5^–/– hair cells, suggesting that CLIC5 is not required for the GRXCR2 localization in hair cells during development either (Figure 2D).

The 60 amino acids from 36 to 95 in GRXCR2, the binding site with CLIC5, are highly conserved in mammals (Figure 3A). To investigate the extent to which the interaction between GRXCR2 and CLIC5 is involved in auditory perception, 180 bp nucleotides coding these 60 amino acids were deleted from the exon 1 of Grxcr2 using the CRISPR/Cas9 system. Founder mouse was back-crossed with wild-type C57BL/6J mice for two generations to reduce potential off-target effects of CRISPR. Finally, we obtained a new Grxcr2-mutant mouse line, which will be referred as Grxcr2^D180/D180 mouse (Figure 3B). To confirm whether the 180 bp is deleted from the mRNA of Grxcr2, mRNA was extracted from the inner ear of Grxcr2^D180/D180 mice. Then cDNA of Grxcr2 was amplified and sequenced. Indeed, the 180 bp nucleotides were deleted from Grxcr2. The GRXCR2 antibody used for immunostaining does not recognize GRXCR2 (Δ36–95) as the immunogen of the antibody is located within the first 80 amino acids of GRXCR2. To investigate whether GRXCR2 (Δ36–95) is also concentrated at the base of stereocilia, injectoporation (Xiong et al., 2014) was performed to express Myc-tagged GRXCR2 (Δ36–95) in P3 hair cells. Two days after injectoporation, hair cells were fixed and Myc-antibody was used to detect Myc-GRXCR2 (Δ36–95). Similar to the full-length GRXCR2, GRXCR2 (Δ36–95) was also concentrated near the base of the stereocilia (Figure 3C).

To analyze the stereocilia morphology in Grxcr2^D180/D180 mice, scanning electron microscopy (SEM) was performed. In Grxcr2^–/– mice, most of the hair cell bundles were disorganized and had lost their characteristic V-shapes, which is consistent with the previous results (Figure 3D; Liu et al., 2018). Interestingly, in Grxcr2^D180/D180 mice, both IHC and OHC bundles were minimally affected at P14, suggesting that the interaction between GRXCR2 and CLIC5 is not essential for the stereocilia morphogenesis (Figure 3D).

To investigate the localization of CLIC5 in the Grxcr2^D180/D180 hair cells, whole mount immunostaining was performed. Similar to that in wild-type and Grxcr2^–/– hair cells, CLIC5 was also concentrated at the base of stereocilia in Grxcr2^D180/D180 hair cells at P7 (Figure 3E). In the 6-week-old wild-type hair cells, some CLIC5 entered stereocilia shaft. The immunostaining signals of CLIC5 had no significant difference between wild-type and Grxcr2^D180/D180 hair cells (Figure 3F). These results further suggest that GRXCR2 is not required for the localization of CLIC5 to the stereociliary base. Deleting 60 amino acids from 36 to 95 in GRXCR2 does not affect the binding of GRXCR2 with taperin. Consistently, in the Grxcr2^D180/D180 hair cells, taperin was still concentrated at the base of stereocilia in both P7 and adult hair cells (Figures 3G,H).

To investigate whether the interaction between GRXCR2 and CLIC5 is required for auditory perception, brain stem response (ABR) to broadband click stimuli in 6-week-old animals was measured. The Grxcr2^–/– mice had profound hearing loss. Interestingly, although the Grxcr2^D180/D180 has fairly normal V-shaped hair bundles, they had a moderate hearing loss with ∼20 dB hearing threshold elevation compared with wild-type mice (Figures 4A,B). Measurements of responses to pure tones revealed that Grxcr2^D180/D180 had moderate hearing loss at lower frequencies (∼10–15 dB hearing thresholds elevation at 8 and 12 kHz) and more severe hearing loss at higher frequencies (∼40 dB hearing thresholds elevation at 24 and 28 kHz) (Figure 4C). These results suggest that the interaction between GRXCR2 and CLIC5 is required for normal hearing especially for hearing at high frequencies (Figure 4D).