CIB2 function is distinct from that of whirlin in the organization of sterocilia architecture

Hearing depends upon hair cells, the polarized epithelial cells of the inner ear that have mechano-sensitive hair bundles located at their apical pole. The hair bundle is composed of numerous stereocilia that are organized in a graded staircase pattern. The staircase architecture of the stereocilia bundle is conserved across all vertebrate hair cells and essential for hearing function, because it allows effective pulling of the tip links between stereocilia of neighboring rows. The organization, elongation and row identity of stereocilia in cochlear hair cells are highly regulated and involve several protein complexes, including myosin (MYO)3A, MYO3B (Ebrahim et al., 2016), MYO15A (Belyantseva et al., 2003), whirlin (WHRN) (Belyantseva et al., 2005), epidermal growth factor receptor pathway substrate 8 (EPS8) (Manor et al., 2011; Zampini et al., 2011), EPS8-like 2 (EPS8L2) (Furness et al., 2013) and G protein signaling modulator 2 (GPSM2)/G protein subunit alpha i3 (GNAI3) (Mauriac et al., 2017; Tadenev et al., 2019; Tarchini et al., 2016). For instance, a long isoform of WHRN is localized at the very tips of stereocilia and, together with its carrier – MYO15A, is essential for the normal elongation of stereocilia and formation of the characteristic staircase shape of the hair bundle (Belyantseva et al., 2003, 2005). Furthermore, during development, WHRN targets GPSM2/GNAI3 to the tips of the first-row stereocilia, leading to the accumulation of the elongation protein complex at this site. In the absence of the GPSM2/GNAI3 complex, the first-row stereocilia fail to grow to the correct height, resulting in profound deafness (Mauriac et al., 2017; Tadenev et al., 2019; Tarchini et al., 2016).

We previously reported that, in calcium and integrin binding protein 2 (CIB2) homozygous mutant mice, the overall architecture of the cochlear stereociliary bundle is affected. CIB2 deficiency results in overgrowth of transducing shorter-row stereocilia (rows 2 and 3) in the auditory hair cells, suggesting a direct role of CIB2 in stereocilia staircase patterning (Giese et al., 2017). We have also demonstrated that CIB2 physically interacts with WHRN (Riazuddin et al., 2012). However, our prior studies in Whrn^wi mutant mice revealed that WHRN is not necessary for the localization of CIB2 in mouse inner ear hair cell stereocilia (Riazuddin et al., 2012). Given the role of both CIB2 and WHRN in the normal staircase patterning of stereocilia, and their binding with each other, we sought to determine (1) their interacting domains; (2) whether CIB2 and WHRN have functional overlap with each other; and (3) whether there is genetic interaction between the genes encoding CIB2 and WHRN. To map the interacting regions, we generated a series of CIB2 and WHRN deletion and point mutation fluorescently tagged constructs, and performed nanoscale pulldown (NanoSPD) assays, co-immunoprecipitation studies and molecular modeling. For functional interactions, we adapted classical genetic approaches, and crossed Cib2^KO mice with Whrn^wi (knockout) or Whrn^BAC279 (overexpressing) (Mburu et al., 2003) mice and analyzed their first- and second-generation offspring. Analysis of double heterozygotes and double homozygous mutants allowed us to determine whether loss of CIB2 and WHRN affects viability, as both proteins are expressed in many tissues besides inner ear (Riazuddin et al., 2012; Mburu et al., 2003), and produces a more severe inner ear phenotype or a superimposition of pathologies. Analysis of double heterozygotes also allowed us to determine digenic interaction between Cib2 and Whrn. Finally, analysis of offspring from Cib2^KO/KO, Whrn^BAC279 crosses was used to determine whether overexpressing WHRN could rescue the stereocilia staircase pathology. Our data suggest that CIB2 has a role that is distinct from that of WHRN in development and/or organization of the stereocilia staircase patterning in the cochlear hair cells.

Our previous studies revealed that the row identity of the cochlear stereociliary bundles was not maintained in Cib2 mutant mice (Giese et al., 2017). Here, we investigated whether the observed loss of row identity is due to a role of CIB2 in the regulation of proteins essential for the staircase pattern (e.g. MYO15A, WHRN, EPS8, etc.) (Belyantseva et al., 2003, 2005; Manor et al., 2011; Zampini et al., 2011). To test this hypothesis, we immunostained organs of Corti from Cib2 mutant mice (Cib2^KO/KO), along with those from controls, at postnatal day (P)12, for the proteins MYO15A, WHRN, EPS8 and EPS8L2. Immunolabeling using PB48 antibody revealed aberrant staining pattern for MYO15A in Cib2^KO/KO mutants, with overaccumulation at the tips of first-row stereocilia of inner hair cell (IHC) bundles (Fig. 1A). The overaccumulation of MYO15A at the tips of the tallest row of stereocilia was further confirmed by the quantification of fluorescent signal measured using confocal microscopy (Fig. 1B). Furthermore, in the IHCs of Cib2^KO/KO mice, WHRN immunostaining, detected using antibodies specific to the long isoform of WHRN (Belyantseva et al., 2005), was slightly weaker, but not statistically significant, at the tips of stereocilia (Fig. 1A,B). However, EPS8 and EPS8L2 immunostaining in the IHCs of Cib2^KO/KO mice persisted at levels similar to those observed at the tips of stereocilia of control hair cells (Fig. 1A,B).

We previously documented that CIB2 directly interacts with WHRN, and forms a tri-partite complex with WHRN and MYO15A (Riazuddin et al., 2012). To further confirm these findings, and to characterize the specific domains required for the CIB-WHRN interaction, we performed NanoSPD assays (Bird et al., 2017). For these studies, COS-7 cells were co-transfected with GFP-WHRN and various mCherry-MYO10-CIB2 deletion constructs (Fig. 2A,B; Fig. S1). The assay, using mCherry-MYO10-CIB2 and full-length GFP-WHRN constructs, confirmed interaction of these proteins. Further, both proteins significantly accumulated at the tip of filopodia in COS-7 cells compared to negative control cells transfected with either mCherry-MYO10 or GFP-WHRN only (Fig. 2B,C).

Next, we performed NanoSPD assays in COS-7 cells using various mCherry-MYO10-CIB2 deafness-causing missense variants (p.Glu64Asp, p.Arg66Trp, p.Phe91Ser, p.Cys99Trp, p.Ile123Thr, p.Arg186Trp) and GFP-WHRN full-length constructs (Fig. 2B,C; Fig. S1). Except the p.Phe91Ser variant, none of the other tested missense pathogenic variants affect the CIB2-WHRN complex (Fig. 2C). Although the interaction between CIB2, harboring p.Phe91Ser variant, and WHRN was maintained, it was significantly weaker (*P=0.0247) than that between WHRN and wild-type (WT) CIB2 (Fig. 2C). Next, as bait, we used several truncated CIB2 constructs deleting the EF hand domains (p.Pro103*, p.Gln139*, p.Phe183*). The p.Pro103* truncation completely abolished the WHRN interaction, whereas p.Gln139* and p.Phe183* truncated CIB2 proteins retained the ability to bind WHRN (Fig. 2B,C).

We next investigated the specific domain region within WHRN required for interaction with CIB2. We performed NanoSPD assay in COS-7 cells using full-length mCherry-MYO10-CIB2 and GFP-WHRN truncated constructs (Fig. 3; Fig. S2). All the GFP-WHRN truncated constructs that have PDZ2 and HDD2 domains in them co-accumulated at the tips of filopodia, suggesting that the PDZ1 and PDZ3 domains are not necessary for CIB2-WHRN interaction (Fig. 3B,C; Fig. S2).

To gain detailed insight into the interacting regions of CIB2-WHRN, we generated an AlphaFold2 multimer (Jumper et al., 2021; Tunyasuvunakool et al., 2021) model of the CIB2 in complex with WHRN (Fig. 4A,B; Fig. S3). Significant complex formation was predicted between CIB2 and well-defined regions of WHRN, involving 83 and 74 residues, respectively, burying 2693 Å (Belyantseva et al., 2003). The predicted interaction is between the CIB2 C-terminal EF2 domain and the HDD2 domain of WHRN (Fig. 4A). The involvement of the CIB2 C-terminal domain is consistent with deletion experiments, discussed above. Further, analysis of the complex indicated that it was strongly centered on the HDD2 domain of WHRN, with eight of ten potential salt bridges, and 12 of 16 potential hydrogen bonds, contributed by the HDD2 domain of WHRN (Fig. S3). Strikingly, this potential interaction is highly structurally homologous to that formed between the high-affinity transmembrane channel-like protein 1 (TMC1)-binding region and CIB2 (Liang et al., 2021; Giese et al., 2025) (Fig. 4B). Indeed, the core helical CIB2-binding regions of both TMC1 and WHRN have shared highly hydrophobic faces bound to the EF2 domain of CIB2, suggesting a competitive interaction.

Based on these predictions, we generated a GFP-WHRN-HDD2 domain-only construct (NM_001008791.2; residues 415-561) and tested its interaction with full-length CIB2 through NanoSPD assay (Fig. 4C,D). Consistent with the AlphaFold2 prediction, CIB2 interaction with the HDD2 region was comparable to that with full-length WHRN protein (Fig. 4C,D). Taken together, these data establish that the CIB2-WHRN interaction is mediated through the EF2 domain of CIB2 and the HDD2 domain of WHRN.

Given the interaction between CIB2 and WHRN, increased expression of MYO15A (WHRN transporter) in Cib2 mutant mice, and the role of both proteins in orchestrating the staircase pattern of stereocilia bundle, we next sought to determine whether altered WHRN levels are responsible for impaired stereocilia architecture in Cib2^KO/KO mice using a classical genetic approach. To test this, we crossed Cib2^KO/KO and Whrn^wi/wi mutants (Fig. 5A) and analyzed first- and second-generation offspring. Analysis of double heterozygous (Cib2^KO/+;Whrn^wi/+) and double mutants (Cib2^KO/KO;Whrn^wi/wi) allowed us to determine whether deficiencies in, or loss of, CIB2 and WHRN produces a more severe phenotype or a superimposition of pathologies.

First, WT, double heterozygous and double homozygous mutants from these crosses were subjected to auditory brainstem response (ABR) measurements at 12-16 weeks of age (Fig. 5B). At this age, the double heterozygous Cib2^KO/+;Whrn^wi/+ mice had ABR thresholds similar to those of their WT littermates (Fig. 5B). In contrast to WT and double heterozygous mice, as anticipated from the reported phenotypes of Cib2^KO/KO and Whrn^wi/wi mice (Giese et al., 2017), the double mutants Cib2^KO/KO;Whrn^wi/wi had no response to any sound stimuli (Fig. 5B).

Next, we examined by scanning electron microscopy (SEM) the inner ears of double mutants and controls at 2 weeks of age. SEM images from all the double mutant mice exhibited an apparent phenotype (Fig. 5C and Fig. 6A) of Whrn^wi/wi mutants (Holme et al., 2002) with some superimposition of features of Cib2^KO/KO mice (Giese et al., 2017). Whrn mutants had extremely short stereocilia in IHCs and outer hair cells (OHCs) (Holme et al., 2002). In contrast, the OHCs in Cib2^KO/KO mutant mice often had overgrowth of the second row of stereocilia and horseshoe-shape bundle, whereas the IHCs had abnormally thick third- and fourth-row stereocilia and persistent kinocilia (Giese et al., 2017). The Cib2^KO/KO;Whrn^wi/wi double mutant mice had shorter stereocilia bundle, with kinocilia failing to regress properly (Fig. 5C and Fig. 6A). Moreover, reducing the levels of either protein (CIB2 or WHRN) (Wang et al., 2017) on the genetic background of the other mutant strain (Cib2^KO/KO;Whrn^wi/+ or Cib2^KO/+;Whrn^wi/wi) neither worsened nor rescued the normal staircase pattern in either situation (Fig. 5C and Fig. 6A). A prior study reported reduced levels of WHRN in the shaft of stereocilia of auditory hair cells in Cib2^KO/KO mice (Michel et al., 2017). Therefore, we also investigated the impact of WHRN overexpression in Cib2^KO/KO mice. For these studies, we crossed and generated mice that were homozygous for the Cib2^KO allele and were also positive for the Whrn^BAC279 transgene (Mburu et al., 2003). Overexpressing WHRN, using the Whrn^BAC279 strain, also failed to restore the stereocilia staircase pattern in Cib2^KO/KO mice (Fig. 6B). Collectively, these results support the notion that CIB2 and WHRN proteins have coordinated, but non-overlapping, functions in orchestrating the stereocilia staircase pattern and bundle shape.

In this study, we demonstrate that CIB2 is multifunctional, with key independent functions in development and/or maintenance of stereocilia staircase pattern in auditory hair cells. Although, we cannot rule out the possibility of lack of interaction due to the misfolding of the deletion constructs (e.g. GFP-WHRN HHD1-PDZ1), both our NanoSPD-based interaction assays and AlphaFold2 multimer (Jumper et al., 2021) prediction model support the interaction between the CIB2 C-terminal EF2 domain and the HDD2 domain of WHRN (Fig. 4A).

Loss of CIB2 caused a slight, but not statistically significant, reduction in WHRN levels at the tips of the tallest rows of IHC stereocilia. However, double heterozygous mice from a Cib2^KO×Whrn^wi cross exhibited normal startle responses to sound (Fig. 5B). The double homozygous mutants of Cib2 and Whrn exhibited profound hearing loss. The morphology of cochlear hair cell stereocilia in double homozygous mutant mice suggests a superimposition of the phenotypes generated by each of the single homozygotes. Non-overlapping functions would be expected to generate a more pronounced phenotype. Furthermore, overexpression of WHRN in Cib2^KO mice did not restore normal staircase architecture of stereocilia in cochlear hair cells. Taken together, our studies indicate that CIB2 performs a distinct function in regulating the staircase architecture of cochlear hair cell stereocilia that does not obviously overlap with the function of WHRN. The superimposition of phenotypes in the double homozygous mutant mice indicates that CIB2 and WHRN have unique and specific functions in stereocilia bundle development and patterning. This is consistent with molecular analysis, suggesting overlapping mutually exclusive binding interfaces on CIB2 for WHRN and TMC1. The nature of the formed complexes will be governed by inherent binding affinities, the stoichiometry of components, and the complex nature of interactions involving multiple dynamic molecular components. These factors warrant further investigation.

Based upon the present study, and upon the individual differences in the stereocilia bundle phenotypes of Cib2 and Whrn single homozygous mutant mice (Giese et al., 2017), both Cib2 and Whrn appear to play distinct roles in establishing the correct architecture of stereocilia bundles. Recent studies have also demonstrated a critical role of MET activity in regulating the staircase pattern of stereocilia bundles in developing cochlear hair cells (Krey et al., 2020). In Usher mutant mouse models, such as whirler (Whrn^wi/wi), shaker (Myo7a^sh1/sh1), Ush1g (Ush1g^js/js) and Ush1c (Ush1c^dfcr/dfcr), in which MET is abolished in sensory hair cells, it has been reported that the stereocilia staircase pattern is altered, and that stereocilia are dramatically reduced in length, suggesting that the MET machinery has a positive effect on F-actin polymerization (Krey et al., 2020; Velez-Ortega et al., 2017).

Several studies have reported loss of MET function in Cib2 mutant mice (Giese et al., 2017; Liang et al., 2021; Michel et al., 2017). However, in Cib2 mutants, the second- and third-row stereocilia are elongated, which is opposite to the expected retraction of transducing stereocilia that occurs after loss of MET current (Krey et al., 2020; Velez-Ortega et al., 2017). Thus, CIB2 likely has a role in stereocilia growth, unrelated to MET. Recent studies demonstrated that GPSM2 and inhibitory G proteins of the alpha family (GNAI1, GNAI2 and GNAI3) form a complex, which is essential for stereocilia elongation and organization into a staircase pattern (Mauriac et al., 2017; Tadenev et al., 2019). GPSM2-GNAI binds to WHRN, and the whole complex relies on MYO15A to be transported to the tips of stereocilia (Mauriac et al., 2017; Tadenev et al., 2019). As hair cells mature, the GPSM2-GNAI complex and its partners are trafficked to the tips of stereocilia adjacent to the bare zone by the MYO15A motor, thereby establishing the ‘identity’ of the first, tallest row of stereocilia (Tadenev et al., 2019). Bearing in mind the abnormal stereocilia heights in Cib2 mutants, it could be speculated that CIB2 has a role in the GPSM2-GNAI stereocilia elongation complex.

All animal procedures were approved by the Institutional Animal Care and Use Committees (IACUCs) of the participating institutes. Animal strains used in this study have been previously reported (Giese et al., 2017; Mburu et al., 2003).

The cochlear and vestibular sensory epithelia were isolated, fine dissected and permeabilized in 0.25% Triton X-100 for 1 h, and blocked with 10% normal goat serum in PBS for 1 h. The tissue samples were probed with primary antibodies against MYO15A, WHRN, EPS8 or EPS8L2 overnight, and, after three washes, they were incubated with the secondary antibody for 45 min at room temperature (RT). Rhodamine phalloidin or Alexa Fluor phalloidin 488 were used at a 1:250 dilution for F-actin labeling. Nuclei were stained with DAPI (Molecular Probes). Images were acquired using either a LSM 700 laser scanning confocal microscope (Zeiss) with a 63×/1.4 NA or 100×/1.4 NA oil immersion objectives or a Leica SP8 laser scanning confocal microscope with a 100×/1.44 NA objective. Stacks of confocal images were acquired with a z-step of 0.05-0.5 µm and processed using ImageJ software [National Institutes of Health (NIH)]. Experiments were repeated at least three times, using at least three different animals.

Human full-length CIB2 and WHRN cDNA constructs were generated as previously described (Riazuddin et al., 2012). Site-directed mutagenesis was performed on the full-length constructs using QuickChange PCR (Stratagene) to generate specific truncated or mutated versions. All constructs were sequence verified before use in the experiments.

For NanoSPD assays, we followed the instructions reported previously (Bird et al., 2017; Sethna et al., 2021). Briefly, 60-70% confluent COS-7 cells in six-well plates for nanoTRAP [a construct consisting of GFP nanobody fused with the heavy meromyosin (HMM domain) of myosin 10 (myo10HMM-GFP nanobody); this allows preferential migration of any GFP-tagged protein to the filopodia tips in the transfected cells] were transfected with Lipofectamine 2000 (3:1 ratio) with 1 µg plasmid construct each (nanoTRAP, GFP-tagged bait and prey), and, 24 h post-transfection, cells were split 1:10 ratio on glass coverslips to allow for filopodia formation. The following day, cells were fixed with 4% paraformaldehyde for 15 min at RT and permeabilized with 0.2% Triton X-100 in PBS for 15 min at RT, followed by blocking with 10% normal goat serum (NGS) in PBS for at least 30 min at RT. Primary antibodies were diluted (1:200) in 3% NGS-PBS and incubated overnight at 4°C, followed by the incubations with the indicated goat secondary antibodies (diluted 1:1000). A Zeiss 710 laser scanning confocal microscope or Nikon W1 spinning disk microscope was used for image acquisition.

Quantification of fluorescence intensities at the tips of the filopodia was performed using ImageJ software (NIH). Each value is an average over a square area of 1 μm², with a center at the tip of each individual filopodium (n≥60). Data are expressed as a ratio of GFP/mCherry intensities. Each value was then normalized against the values measured in the cells transfected with the bait constructs only. All data represent the mean±s.e.m. One-way ANOVA with Tukey's multiple comparisons test was used to compare the different groups of independent samples.

An AlphaFold2 (Jumper et al., 2021) multimer model (Evans et al., 2021 preprint) of the CIB2/WHRN complex was generated using the Colab server without template constraint (https://github.com/sokrypton/ColabFold). Interaction interfaces were analyzed using PDBePISA (Krissinel and Henrick, 2007). Structural models were analyzed, and figures were prepared, using PyMOL (Version 2.4.1, Schrödinger).

Hearing thresholds of mice at 12-16 weeks of age (n=4/genotype) were evaluated by recording ABR. All ABR recordings, including broadband clicks and tone-burst stimuli at three frequencies (8, 16 and 32 kHz), were performed using an auditory-evoked potential RZ6-based auditory workstation (Tucker-Davis Technologies) with high-frequency transducer RA4PA Medusa PreAmps. Maximum sound intensity tested was 100 dB sound pressure level. TDT system III hardware and BioSigRZ software (Tucker Davis Technology) were used for stimulus presentation and response averaging.

Cochleae were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4 (Electron Microscopy Sciences, Hatfield, PA, USA) supplemented with 2 mM CaCl₂ (Sigma-Aldrich) for 1-2 h at RT. Then, the sensory epithelia were dissected in distilled water, dehydrated through a graded series of ethanol, critical point dried from liquid CO₂ (Leica EM CPD300), sputter coated with 5 nm platinum (Q150T, Quorum Technologies, Guelph, Canada), and imaged with a field-emission scanning electron microscope (Helios Nanolab 660, FEI, Hillsboro, OR, USA).

We thank the University of Maryland School of Medicine Microscopy Core.