Identification of Binding Partners of Deafness-Related Protein PDZD7

PDZD7 is an important deafness gene, whose mutations are associated with syndromic and nonsyndromic hearing loss. PDZD7 contains multiple PDZ domains that are essential for organizing various proteins into protein complex. Several PDZD7-binding proteins have been identified, including usherin, ADGRV1, whirlin, harmonin, SANS, and MYO7A, all belonging to USH proteins. Here, we report the identification of novel PDZD7-binding partners through yeast two-hybrid screening using the first two PDZ domains of PDZD7 as bait. Eleven proteins were identified, most of which have not been reported as PDZD7-binding partners before. Among the identified proteins, ADGRV1, gelsolin, and β-catenin have been shown to play important roles in hearing, whereas the functions of other proteins in the inner ear remain elusive. We confirmed the expression of one candidate PDZD7-binding protein, CADM1, in the mouse inner ear and evaluated the auditory function of Cadm1 knockout mice by performing auditory brainstem response (ABR) measurement. Unexpectedly, Cadm1 knockout mice show normal hearing threshold, which might be explained by the possible compensation by its homologs that are also expressed in the inner ear. Taken together, our work identified several novel PDZD7-binding proteins, which will help us to further understand the role of PDZD7 in hearing transduction.


Introduction
Usher syndrome (USH) is the most frequent form of inherited sensory deaf-blindness that is characterized by hearing loss and vision defect [1,2]. According to the severity of hearing loss as well as the presence or absence of balancing problems, USH is clinically classified into three subtypes, namely, USH1, USH2, and USH3, with USH1 as the most severe one. At present, ten genes have been associated with USH, including MYO7A, USH1C, CDH23, PCDH15, USH1G, CIB2, USH2A, ADGRV1, WHRN, and CLRN-1 [3][4][5][6][7][8][9][10][11][12][13][14][15]. Mutations of USH genes are also responsible for nonsyndromic hearing loss. USH proteins have been shown to interact with one another and form multiprotein complexes and play important roles in the development, maintenance, and function of stereocilia and synapses in the inner ear sensory hair cells [16].
Recently, PDZD7 was suggested to be a USH modifier and a contributor to digenic USH [17]. Meanwhile, mutations in human PDZD7 gene are also associated with nonsyndromic hearing loss DFNB57 [18][19][20]. Similar to harmonin (USH1C) and whirlin (USH2D), full-length PDZD7 contains three PDZ domains, a harmonin-N like (HNL) domain, and a proline-rich (PR) region. Shorter PDZD7 isoforms containing the first two PDZ domains were also detected in the inner ear [17,18,21]. In mice, loss of PDZD7 was shown to result in stereocilia disorganization as well as mechanotransduction deficits [21].
At present, little is known about other non-USH PDZD7binding partners. In the present work, yeast two-hybrid screening was performed using the first two PDZ domains as bait to identify new PDZD7-binding partners that are expressed in the inner ear. Identification of PDZD7-binding proteins will help us to further understand the role of PDZD7 in hearing transduction.

Materials and Methods
2.1. DNA Constructs. Mouse cDNA encoding PDZD7 short isoform (amino acids 1-557) was inserted into pBD-GAL4 Cam vector (Stratagene) to express the bait protein for yeast two-hybrid screen. The same cDNA was inserted into pmCherry-N1 or pMYC-C2 (modified pEGFP-C2 with EGFP-coding sequence replaced by Myc-coding sequence) to express PDZD7-mCherry or Myc-PDZD7 fusion protein.

Yeast
Two-Hybrid Screen. The yeast two-hybrid screen was performed as previously described [26][27][28]. Briefly, yeast strain AH109 (Clontech) was sequentially transformed with the bait plasmid and a chicken cochlear cDNA library in the HybriZAP two-hybrid vector [29]. A total of 2.4 × 10 6 transformants were screened using HIS3 as the primary reporter gene with the presence of 2.5 mM of 3-amino-1,2,4-triazole (3-AT). The positive colonies were further examined using two other reporter genes ADE2 and lacZ. The prey vectors in triple-positive yeast colonies were recovered, and the sequence of cDNA inserts was determined by sequencing.

Colocalization
Assay. COS-7 cells were grown on gelatincoated glass cover slips and transfected with vectors that express target proteins fused to EGFP or mCherry. Twentyfour hours after transfection, cells were fixed with 4% paraformaldehyde (PFA) in PBS for 15 minutes, then permeabilized and blocked with PBT1 (0.1% Triton X-100, 1% BSA, 5% heat-inactivated donkey serum in PBS, pH 7.3) for 30 minutes. For nuclei staining, cells were incubated with DAPI (Gen-View Scientific Inc.) for 15 minutes, then mounted in glycerol/PBS (1 : 1). The subcellular localization of target proteins was examined with a confocal microscope (LSM 700, Zeiss).

Coimmunoprecipitation (co-IP) and Western Blot.
HEK293T cells were transfected with vectors that express target proteins fused to EGFP or Myc epitope. Twenty-four hours after transfection, cells were washed with PBS and lysed in ice-cold lysis buffer consisting of 150 mM NaCl, 50 mM Tris at pH 7.5, 1% (vol/vol) Triton X-100, 1 mM PMSF, and 1x protease inhibitor cocktail (Roche). After centrifugation at 4°C, the supernatant was incubated with immobilized anti-Myc antibody (Sigma-Aldrich, Cat. number E6654) at 4°C for 2 hours. Immunoprecipitated proteins were separated by polyacrylamide gel electrophoresis (PAGE), then transferred to PVDF membrane. After blocking in PBS containing 5% BSA and 0.1% Tween-20, the membrane was incubated with anti-Myc (Abmart, Cat. number M20002) or anti-GFP (Abmart, Cat. number M20004) antibody at 4°C overnight, followed by incubation with HRP-conjugated secondary antibody (Bio-Rad, Cat. number 170-6516) at room temperature for an hour. The signals were detected with the ECL system (Cell Signaling Technology, Danvers, MA). . To obtain the optimal sensitivity and specificity, cycle lengths for different PCR reaction sets were adjusted between 23 and 38 cycles, and annealing temperatures were adjusted between 56 and 64°C. The PCR products were separated by electrophoresis on agarose gel.
2.6. Quantitative Real-Time PCR (Q-PCR). Q-PCR was carried out using SYBR® Premix Ex Taq™ system (Perfect Real Time, Takara) according to the manufacturer's protocol. Amplification and detection were run in a Sequence Detection System SLA-3296 (Bio-Rad) in triplicate with an initial cycle of 95°C for 10 seconds followed by 40 cycles of 95°C for 5 seconds, 60°C for 30 seconds, and 72°C for 20 seconds. Negative control samples (without template) were processed in the same way. The specificity of the amplifications was verified by melting curve analysis. The sequences of primers are as follows: Cadm1: forward primer GTG ATC CAG CTC CTG AAC CC, reverse primer CGT GTA GAG CTG GCA GAA GT and β-actin: forward primer CTC CAT CCT GGC CTC GCT GT, reverse primer GCT GTC ACC TTC ACC GTT CC. Relative quantization of Cadm1 expression normalized to β-actin was calculated according to the 2 −ΔΔ CT method.

Animal Maintenance and Auditory Brainstem Response
(ABR) Measurement. Cadm1 knockout mice (number RBRC04063) were obtained from RIKEN BioResource Center. Generation and characterization of Cadm1 knockout mice have been described elsewhere [30,31]. All animal experiments were approved by the Ethics Committee of Shandong University School of Life Sciences and conducted accordingly. For ABR measurement, mice were anesthetized with 5% chloral hydrate (0.5 ml/100 g body weight). Electrodes were inserted subcutaneously at the vertex, pinna, and near the tail. A RZ6 workstation and BioSig software (Tucker Davis Technologies Inc.) were used for the stimulus generation, presentation, ABR acquisition, and data management. Specific acoustic stimuli were generated using highfrequency transducers, and ABR thresholds were obtained by reducing the stimulus intensity in 10 dB SPL steps to identify the lowest intensity at which all ABR waves were detectable. For noise exposure, mice were exposed to 2-8 kHz noise at 96 dB SPL (Crown, CD i1000) for 2 hours, and ABR thresholds were measured preexposure and at various postexposure time points. For each genotype, at least three animals were used, and data were shown as means ± standard errors. Student's t-test was used for statistical analysis, and p < 0 05 was considered statistically significant.

Results
3.1. Identification of Potential PDZD7-Binding Partners through Yeast Two-Hybrid Screening. In order to identify new PDZD7-binding partners, we performed yeast twohybrid screening of a chicken cochlear cDNA library using PDZD7 short isoform as bait. This isoform contains the first two PDZ domains of PDZD7. Around thirty positive clones were obtained that activate all the three reporter genes, representing eleven candidate PDZD7-binding proteins (Table 1). Among the proteins identified, ADGRV1 (USH2C) is a known PDZD7-binding partner, whereas the interactions between PDZD7 and the other proteins have not been reported. The most frequently encountered two proteins are β-catenin and ADGRV1, both of which contain a type I PDZ-binding interface (PBI) at their C-termini. Six candidate PDZD7-binding proteins (gelsolin, TRIM35, CADM1, AMOT, Golgin45, and Numb) contain a type II PBI at their C-termini. Three candidates (KCTD10, CCDC27, and TRIP11) do not have a predictable C-terminal PBI.

PDZD7 Colocalizes with β-Catenin, AMOT, and CADM1
When Overexpressed in COS-7 Cells. Next, we examined the subcellular localization of PDZD7 in the presence of these candidate binding partners in cultured cells. When overexpressed in COS-7 cells, PDZD7-mCherry localized in the cytoplasm as well as on the plasma membrane (Figure 2(a)), whereas EGFP-β-catenin mainly localized in the nuclei in a punctate pattern (Figure 2(b)). Noticeably, when expressed together with EGFP-β-catenin, PDZD7-mCherry translocated into the nuclei and colocalized with EGFP-β-catenin (Figure 2(c)), in consistent with the potential interaction between these two proteins.

Cadm1
Expression in the Mouse Inner Ear. Among the identified candidate PDZD7-binding partners, CADM1 attracted our most attention. The interaction between CADM1 and PDZD7 was further confirmed by co-IP of epitope-tagged proteins (Figures 4(c)). RT-PCR results showed that Cadm1 is highly expressed in the spiral ganglion and weakly expressed in the basilar membrane ( Figure 5(a)). The expression of Cadm1 in the developing inner ear was examined by performing quantitative real-time PCR (Q-PCR), which showed that Cadm1 was detected in all developmental stages examined, peaking at around postnatal day 9 (P9) (Figure 6(a)). The expression pattern of Cadm1 in the cochlea was further examined using a mouse model whose exon 1 of Cadm1 gene was replaced by lacZ reporter gene cassette [30,31]. X-gal staining of P7 Cadm +/− inner ear suggested that Cadm1 is abundantly expressed in the spiral ganglion. At this stage, the expression of Cadm1 in the basilar membrane was relatively weak and mainly enriched in supporting cells (Figures 6(b)-6(e)).

Cadm1 Knockout Mice Have Normal Hearing Threshold.
We then evaluated the effect of Cadm1 disruption on mouse auditory function by performing ABR measurement. The result showed that hearing thresholds of 1-month-old to 4-month-old Cadm1 −/− mice were comparable to those of wild-type or Cadm1 +/− mice, suggesting that CADM1 is not indispensable for hearing transduction (Figure 7(a)). To investigate whether Cadm1 −/− mice show increased acoustic vulnerability, we exposed P45 mice to 2-8 kHz noise at 96 dB SPL for 2 hours. ABR thresholds were measured before and after the noise exposure, which did not reveal any significant difference between Cadm1 −/− and Cadm1 +/− or wild-type mice (Figure 7(b)). Taken together, our results suggested that the auditory function of Cadm1 −/− mice is normal.
The normal hearing threshold of Cadm1 knockout mice promoted us to look for possible explanations. It has been suggested that the loss of specular protein might be compensated for by its homologous protein(s). As an immunoglobulin-(Ig-) like cell adhesion molecule (CAM), CADM1 belongs to nectin-like molecule (Necl) family, which contains five members (CADM1, CADM2, CADM3, CADM4, and Necl5) [32,33]. We examined the expression of Necl family members in mouse inner ear by performing RT-PCR. The results showed that all members are expressed in the mouse inner ear (Figure 5(a)), whereas none of them is upregulated in Cadm1 knockout mice ( Figure 5(b)).

Discussion
PDZD7 is an important deafness gene, whose mutations contribute to syndromic as well as nonsyndromic hearing loss [17][18][19][20]. PDZD7 is a scaffold protein containing three PDZ domains, a HNL domain, and a PR region. Scaffold proteins are important for organizing multiple proteins into protein complex. At present, only a few PDZD7-binding proteins have been reported, including usherin, ADGRV1, whirlin, harmonin, SANS, and MYO7A [17,18,[21][22][23][24][25]. In this work, we used yeast two-hybrid screening to identify new PDZD7binding proteins, which will help us to learn more about the role of PDZD7 in hearing transduction.
Among the potential PDZD7-binding partners identified in this work, β-catenin is the most frequently encountered    one. Wnt/β-catenin signaling pathway plays pivotal roles in development, tissue homeostasis, and so on [34]. It has been suggested that Wnt/β-catenin signaling regulates proliferation of sensory precursors in the postnatal mouse cochlea [35,36]. β-Catenin could upregulate the expression of Atoh1, a transcription factor that is critical for hair-cell differentiation [37]. Consistently, loss of β-catenin inhibited hair-cell differentiation from sensory progenitors [38], whereas forced stabilization of β-catenin in supporting cells resulted in proliferation of supporting cells and generation of hair cells [39]. Our data show that PDZD7 interacted with β-catenin and that PDZD7 translocated into the nuclei together with β-catenin in transfected cells, suggesting a potential role of PDZD7 in regulating β-catenin pathway. Further investigation is needed to fully understand the significance and the mechanism of this interaction.
Gelsolin is a calcium-activated actin-binding protein and plays important roles in F-actin severing, capping, and nucleation [40,41]. It has been shown that gelsolin binds p55 and localizes to the tips of shorter stereocilia of outer hair cells (OHCs) [42]. In mice lacking gelsolin, stereocilia in the apex of the cochlea became long and straggly, suggesting that gelsolin is involved in the regulation of stereocilia elongation [42,43]. Our data suggested that PDZD7 might interact with gelsolin, hence might play a role in stereocilia development and/or maintenance. Consistent with this hypothesis, OHC stereocilia disorganization has been observed in Pdzd7 knockout mice [21].
Numb is an evolutionary conserved protein with multiple functions such as asymmetric cell division control, cell fate determination, endocytosis, cell adhesion, cell migration, ubiquitination of specific substrates, and a number of signaling pathways [44]. It has been reported that Numb was expressed in rat cochlear sensory epithelium, and overexpression of Numb upregulated the expression of Atoh1 in cochlear whole mount cultures [45]. The potential interaction of PDZD7 with Numb raises the possibility that PDZD7 might regulate the function of Numb, which awaits further investigation.
Unlike ADGRV1, β-catenin, gelsolin, and Numb, the other PDZD7-binding proteins identified in the present work have not been reported to function in the inner ear. Genes encoding some of the proteins including CADM1, AMOT, Golgin45, and KCTD10 have been detected in mouse cochlea by RNA transcriptome sequencing (SHIELD, https://shield. hms.harvard.edu) [46]. Among these proteins, CADM1 attracted most our attention. CADM1 is an immunoglobulin (Ig) superfamily protein that contains extracellular Ig-like domains, a single transmembrane domain, and a small intracellular C-terminal tail. CADM1 can bind either transhomophilically or transheterophilically with other nectins or Necls [47,48]. CADM1 plays important roles in modulating synapse development and plasticity, and mutations in CADM1/ Cadm1 gene have been associated with autism spectrum disorder [49][50][51]. We show here that CADM1 interacts with PDZD7 and Cadm1 is abundantly expressed in mouse inner ear. However, our data did not reveal any auditory deficit in Cadm1 knockout mice, suggesting that CADM1 is dispensable for hearing function in mice.
Alternatively, other Necl family members might compensate for the loss of CADM1 in the inner ear. Similar scenario has been observed in the neuromuscular junction (NMJ) of Cadm1 knockout mice, where the loss of CADM1 was compensated for by CADM4 [52].
In conclusion, our present work identified several novel inner ear-expressed PDZD7-binding partners, which will help us to learn more about the role of PDZD7 in hearing. Further investigation is needed to fully understand the biological significance of these interactions.

Conflicts of Interest
The authors declare that there is no conflict of interests regarding the publication of this paper.