Expression of cadherin 23 isoforms is not conserved: implications for a mouse model of Usher syndrome type 1D

RNA isolation and cDNA synthesis

Total RNA was isolated from 14 to 20 inner ears and retinas of wild-type (B10.A-H2^h4/(4R)SgDvEg×C57BL/6)F1 of embryonic day 16.5 (E16.5) and post natal day 0 (P0), P4, P7, P14, P21 as well as RNA from pooled P30 to P90 inner ears and retinas. Total RNA was also isolated from P0 and adult homozygous v^6J (B10.A-H2^h4/(4R)SgDvEg×C57BL/6)F1 mice. Mice were housed in the NIDCD 5 Research Court, animal facility, and were fed NIH 07 ad libitum, and triple filtered chlorinated water was provided. Light cycles are set to be 12 h of light followed by 12 h of darkness. Animals were euthanized by a combination of CO₂ inhalation, cervical dislocation and decapitation. All animal procedures were approved and conducted according to the NIH Animal Care and Use Committee guidelines and Animal Protocol 1263–06 to T.B.F. RNA was extracted using TRIzol^® (Invitrogen, Carlsbad, CA) followed by poly(A)⁺ RNA isolation using Oligotex^® mRNA spin-columns (Qiagen, Valencia, CA). First strand cDNA was synthesized from 1 µg of poly(A)⁺ RNA using SuperScript II Reverse Transcriptase and oligo dT priming (Clontech, Palo Alto, CA).

Semiquantitative real-time PCR

TaqMan assays were designed using Assay by Design (Applied Biosystems, Foster City, CA). Probes hybridized to the junctions of Cdh23 exons 47–48 (NM_023370), the junction of exons 47a-48 (GenBank AY563163, AY563164), and the junction of exons 44–48 (GenBank AK039126; Table 1) are illustrated in Figure 1A. Semiquantitative real-time PCR was performed in triplicates on an ABI 7500 real-time PCR system (Applied Biosystems). PCR reactions were performed in a 50 µl volume containing 1.5 µl cDNA, 25 µl TaqMan Universal PCR Master Mix (Applied Biosystems), 2 µl target primers and probe mix (Applied Biosystems), 1 µl control primers and probe mix (Applied Biosystems) and 20.5 µl of distilled water. Cycling conditions were 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, and 60 °C for 1 min. A probe to Gapdh (Applied Biosystems) was used as the endogenous control. All ΔC_t values are expressed relative to that of E16.5 inner ear cDNA sample using probe 44–48.

Statistical analysis

The data were analyzed by a two-way ANOVA test using GraphPad Prism software version 5 for Windows (GraphPad Software, La Jolla, CA). P values <0.05 were considered statistically significant.

Expression construct

A cDNA encoding the complete sequence of Cdh23 variant Cdh23_v3a (Figure 1A; GenBank AY563159) was PCR amplified from mouse P5 inner ear cDNA [39] using LA Taq DNA Polymerase (Takara Mirus, Madison, WI), cloned into the XhoI and EcoRI sites of GFP-N2 vector (Clontech) and sequence verified. Primers used for cloning: forward (XhoI): 5'-CCG GAT CTC GAG CCT ATG CAG CCC TGC TGA AGG TTC T-3'; reverse (EcoRI): 5'-CCG GAT GAA TTC CAG CTC CGT GAT TTC CAG AGG GC-3'.

Antibodies

Antisera TF7 was generated against the cadherin 23 cytoplasmic domain (residues 2973–3215 RefSeq NM_023370) [28]. The peptide sequence CMLLPNYRAN derived from the seven unique N-terminus amino acids in CDH23_V3 (residues 1–7; MLLPNYR GenBank AY563159; Figure 1A) with the exception of the N-terminus cysteine (Princeton BioMolecules, Langhorne, PA) was used by Covance (Denver, PA) to immunize three New Zealand white rabbits producing antibodies TF647, TF648, and TF649. We purified antisera using an AminoLink Plus Immobilization Kit (Pierce, Rockford, IL). Antiserum 1D1 was generated against an epitope in the cytoplasmic domain of human CDH23 as previously described [42]. A mouse monoclonal antibody to centrin isoforms Cen1p-Cen4p was used as a molecular marker for the connecting cilium and the adjacent centriole of photoreceptor cells [43]. SNAP25 antibody was used as a marker for hair cell synaptic region (Sternberger Monoclonals Inc., Lutherville, MD). PCDH15 antiserum PB303 was raised in rabbit against a synthetic peptide (CGAEPHRHPKGILRHVKNLAELEK; corresponding to residues 1860–1882 of the mouse PCDH15 sequence, GenBank AAG53891) as previously described by Ahmed et al. [44]. Since PCDH15 was shown to be expressed in the retina, antiserum PB303 was used as a positive control in our retinal western blot analyses.

Transfection of GFP-Cdh23_v3a into HeLa cells

We evaluated the specificity of our antibodies raised against CDH23_V3 (Figure 1A) by performing colocalization assays. Lipofectamine 2000 (Invitrogen) was used to transfect a GFP-Cdh23_v3a expression vector into HeLa cells (American Type Culture Collection, Manassas, VA), which do not endogenously express CDH23. After growth for 24 h at 37 °C in 5% CO₂ in Dulbecco’s Modified eagle medium (DMEM; Invitrogen) supplemented with 10% fetal calf serum (FCS; Invitrogen), transfected cells were fixed with 4% paraformaldehyde and processed for immunostaining [45] using a 1:200 dilution (roughly 0.5 mg/ml stock) of antibodies TF648 and TF649. Anti-GFP antibody (Sigma-Aldrich, St. Louis MO) was used in western blots of protein extracts of transfected and untransfected HeLa cells.

Western blot analyses

Western blot analyses were performed as previously described [28,44]. Fresh P0 to P90 inner ears and retinas of wild-type C57BL/6J and homozygous Cdh23^v-6J mice were dissected. Proteins were extracted and denatured by boiling for 5 min in SDS–PAGE sample buffer (0.125 M Tris–HCl, 20% glycerol, 4% SDS, 0.005% bromo-phenol blue). A 20 µg protein sample was separated on 4%–20% Tris-glycine gels (Invitrogen) for blots incubated with antibody TF648. We used 3%–8% Tris-acetate gels (Invitrogen) for blots incubated with antibody TF7. SeeBlue Plus2 pre-stained standard (Invitrogen) was used as size markers. Gels were transferred to polyvinyldifluoride (PVDF) membranes (Millipore, Billerica, MA), and blocked overnight with 5% dry milk in TBST (10mM Tris-HCl pH 7.5, 150 mM NaCl, 0.05% Tween-20). Primary antibodies were diluted 1:1,000 from a roughly 0.5 mg/ml stock. Alkaline phosphatase-conjugated anti-rabbit secondary antibody was used at a 1:10,000 dilution (Promega, Madison, WI). ECL Plex Cy5 and Cy3 secondary antibodies were used in a 1:2,500 dilution. Fluorescent signals were detected and images captured on a Typhoon Trio Plus imaging system (GE Healthcare Bio-Science Corp., Piscataway, NJ). As a control to be sure that large retinal proteins were transferred to the membrane used in this western blot, we used validated antisera PB303 generated against PCDH15 [44] that, as expected, recognized a band of approximately 220 kDa (data not shown).

Immunohistology of mouse retina

Some eyes of adult wild-type mice were cryofixed in melting isopentane without chemical prefixation and cryosectioned as described elsewhere [46]. Cryosections were placed on poly-L-lysine-precoated coverslips and subsequently incubated with 0.01% Tween-20 in phosphate buffered saline (PBS, Quality Biologic, Inc., Gaithersburg, MD; 90 g/l NaCl, 1.44 g/l KH₂PO₄, 7.95 g/l Na₂HPO₄ anhydrous pH 7.4 [10×]) for 20 min. After PBS washing steps (2×5 min at room temperature-RT), sections were, sections were covered with blocking solution that contained 0.5% cold-water fish gelatin plus 0.1% ovalbumin in PBS. Next, sections were incubated for a minimum of 30 min followed by overnight incubation with primary antibodies, diluted in blocking solution at 4 °C. Washed cryosections were incubated for a minimum of 1.5 h at room temperature in the dark with secondary antibodies conjugated to Alexa 488 or Alexa 568 (Invitrogen, Karlsruhe, Germany) in PBS with 4’,6-Diamidin-2’-phenylindoldihydrochlorid (DAPI; Sigma-Aldrich, Deisenhofen, Germany) to stain the DNA of the cell nuclei. After 3x10 min washings with PBS at RT, sections were mounted in Mowiol 4.88 (Hoechst, Frankfurt, Germany). Mounted retina cryosections were analyzed by microscopy (DMRB; Leica microsystems, Bensheim, Germany). Images were obtained with an ORCA-ER camera (Hamamatsu, Herrsching, Germany) and processed with Adobe Photoshop CS (Adobe Systems, San Jose, CA).

Immunofluorescence of mouse inner ear

Mouse inner ears were removed by dissection and cochleae were perfused through the round window with 4% paraformaldehyde (PFA) and incubated in PFA for 2 h at RT. The organ of Corti and vestibular tissues were dissected from the cochlear spiral in PBS. Samples were permeabilized in 0.5% Triton X-100 (Acros Organics, Morris Plains, NJ) for 30 min and then washed in PBS. Non-specific binding sites were blocked using 5% normal goat serum (Gibco-Invitrogen) and 2% BSA (BSA, Gibco-Invitrogen) in PBS for 1 h. Samples were incubated for 2 h in primary antibodies at a concentration of ~0.5 μg/µl in blocking solution and rinsed in PBS. Samples were then incubated for 30 min in a 1:200 dilution of anti-rabbit IgG fluorescein 488 (Amersham Biosciences, Piscataway, NJ) and Alexa Fluor 594 goat anti-rabbit (Molecular Probes-Invitrogen Eugene, OR) secondary antibodies. Rhodamine-phalloidin and Alexa fluor 633 (Molecular Probes-Invotrogen) were diluted 1:100 and used to visualize F-actin. Samples were mounted using ProLong Antifade Kit (Molecular Probes-Invitrogen). Images were captured on a Carl Zeiss LSM-510 laser scanning confocal microscope.

Immunoelectron microscopy

For immunoelectron microscopy a recently introduced protocol was applied [47]. Eyes were perforated and lenses were removed during fixation of isolated mouse eyes in 4% PFA in Soerensen buffer that contained 0.1 M disodiumhydrogenphophate and 0.1 M potassiumdihydrogenphosphate, pH 7.3. After washing (4×15 min at RT) retinas were dissected from eye cups and infiltrated with 10% and 20% sucrose in Soerensen buffer, followed by incubation in buffered 30% sucrose overnight. After four cycles of freezing in liquid nitrogen and thawing at 37 °C retinas were washed in PBS and embedded in buffered 2% agar (Sigma-Aldrich). Agar blocks were sectioned with a vibratome (Leica, Wetzlar, Germany) in 50 µm slices. Vibratome sections were blocked in 10% normal goat serum, 1% BSA in PBS and subsequently incubated with primary antibodies against CDH23 (TF7) for four days at 4 °C. After washing with PBS (4×15 min at RT) the appropriate biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA) were applied to the sections. After washing with PBS (4×15 min at RT) the appropriate were visualized by a Vectastain ABC-Kit (Vector Laboratories) according to manufactor’s protocol. Subsequently, stained retinas fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) and diaminobenzidine precipitates were silver enhanced [48,49], followed by postfixation in 0.5% OsO₄ in 0.1 M cacodylate buffer (pH 7.3) on ice. Dehydrated specimens were embedded in araldite. Ultrathin sections (< 60 nm) were analyzed in a Tecnai 12 BioTwin transmission electron microscope (FEI, Eindhoven, The Netherlands).

Immunohistology of human and monkey retinas

Eyes from three-year-old monkeys were obtained by S.L.B. from John Cogan (Bureau of Biologics, FDA, Bethesda, MD). Tissue was fixed in 4% PFA-PBS and frozen in OCT. Frozen sections of 10 µm were blocked for 1 h with 5% normal goat serum containing 2% bovine serum albumen. Tissue sections were reacted overnight at 4 °C with primary antibody (TF7) at a concentration of 5 µg/ml containing 2% BSA. Treated sections were incubated with alkaline phosphatase secondary antibody (Vector Laboratories, Burlingame, CA). The slides were not counterstained. Human donor eyes were obtained from a Caucasian male (74 years old) who died of congestive heart failure. Human eyes were obtained 6 h post-mortem by S.L.B. from the Maryland State Anatomy Board under IRB exemption SB-019701. The anterior segment eye tissues (cornea, iris) and posterior non-retina tissues (lens, vitreous, ciliary body) were removed. Proteins from human retina were extracted and processed for western analyses.

Cdh23 expression in wild-type and waltzer inner ear and retina

Three semiquantitative real-time PCR assays that could distinguish among some but not all of the Cdh23 isoforms were used to examine the expression of Cdh23 transcripts. Real-time PCR assay 47–48 was designed to detect a class of Cdh23 transcripts, which includes, but is not limited to, full-length Cdh23_v1 (Figure 1A). Expression of this class of transcripts varied significantly both by age and tissue source (Figure 1B, top panel). Two way ANOVA reveals that the expression of this class of transcripts varies significantly both by age and between tissue types, with a p-value <0.0001 for interaction of age and tissue type. Transcripts with the junction of exons 47–48 were present at all ages in both tissues, but were most abundant around age P14. The sustained expression detected in the adult inner ear was consistent with previous published reports describing expression of the longest CDH23 isoform in wild-type mouse inner ear, which appeared to be a component of the tip link [32,37]. The sustained expression may also be due to the expression of still unknown Cdh23 splice variants.

In the homozygous mutant Cdh23^v-6J inner ear and retina, the transcripts detected by assays 47–48 maintained a stable level of expression throughout development. Age and tissue have no significant effect on transcript expression levels in Cdh23^v-6J mice (two-way ANOVA, p=0.05 and p=0.08 respectively). However, compared to the corresponding wild-type transcripts expression levels, the Cdh23^v-6J mutation contributed to a higher level of expression in the adult inner ear and retina (two-way ANOVA; p<0.0001). The waltzer Cdh23^v-6J allele is a G to T transversion occurring in exon 9 and was predicted to truncate the protein in EC3 of the longest Cdh23 transcript Cdh23_v1 [14,15]. Because this nonsense mutation is close to the 5′ end of Cdh23, transcripts that include exons 47 and 48, but are initiated from promoters downstream of exon 9, would be unaffected and detected in real-time PCR assay 47–48.

Real-time TaqMan assay 47a-48 detects transcripts Cdh23_v2a and Cdh23_v2b that contain exons 47a and 48 (Figure 1A). These transcripts use an alternate promoter upstream of exon 47a and they lack 20 ECs encoded by exons 2–46 (RefSeq NM_023370) [28]. In wild-type mice these transcripts were expressed in the retina at all time points and only transiently in the inner ear (two-way ANOVA for tissue and age; p<0.0001; Figure 1B). Expression level in the retina was unchanged in the homozygous Cdh23^v-6J mutant, and was upregulated in the adult inner ear of Cdh23^v-6J mutants (Figure 1B).

Assay 44–48 recognizes the class of transcripts with exon 44 spliced directly to exon 48, which we designated as isoform Cdh23_v4 (Figure 1A). The existence of Cdh23_v4 was based on a partial cDNA (GenBank AK039126), whose promoter and protein products are unknown. There may be many different cDNAs that omit exons 45–47, so the transcript in Figure 1A was meant to represent a class of potential alternatively spliced transcripts of Cdh23 that can be identified by our real-time PCR assay.

The expression pattern of Cdh23_v4 is almost the converse of Cdh23_v2 in that the expression was maximal at earlier stages in the wild-type retina and was downregulated in adult retina, while moderate expression was observed at all time points in the wild-type inner ear (Figure 1B, lower panel). As with the other assays, transcripts with this structure were detected in both the inner ear and retina of Cdh23^v-6J mutant mice and were upregulated in the retina relative to the wild-type adult retina (two-way ANOVA; p<0.0001).

Western blot analyses using CDH23_V3 antibodies

Cdh23_v3a and Cdh23_v3b are short transcripts (708 and 600 nucleotides, respectively) initiated from a promoter in intron 65 and encode 235 and 199 amino acid proteins lacking ECs and transmembrane domains. CDH23_V3a/b expression was evaluated with antibodies TF648 and TF649, directed to the seven unique amino acids encoded at the N-terminus that differentiate this isoform from other CDH23 isoforms (Figure 1A). The antibodies were validated by transfecting a GFP-Cdh23_v3a (Figure 1A) expression vector into HeLa cells. Twenty-four hours after transfection the staining pattern obtained with the GFP-CDH23_V3a overlapped with the staining pattern obtained with antibody TF648. Nontransfected cells did not show antibody staining (Figure 2A-C).

Western blot analyses of transfected and untransfected cells revealed a band of approximately 50 kDa corresponding in size to the predicted size of the GFP-CDH23_V3a fusion protein (approximately 53 kDa; GFP approximately 27 kDa and CDH23_V3a approximately 26 kDa; Figure 2D). Western blot analyses of mouse inner ear and retina protein extracts revealed that antibody TF648 recognizes a band corresponding to the predicted size of CDH23_V3 (roughly 26 kDa) both in the inner ear and the retina (Figure 2E). Similar results were obtained with antibody TF649 (data not shown).

Photoreceptor subcellular distribution of cadherin 23 isoforms

The subcellular distribution of CDH23 in the retina was determined by analyzing cryosections through adult mice eyes by immunofluorescence with antibody TF7, which recognizes the CDH23 cytoplasmic domain, and therefore multiple Cdh23 isoforms (Figure 1A, lower panel). Cryosections of mouse retinas were double-labeled either with anti-centrin as a marker for the ciliary apparatus of photoreceptor cells [43], or anti- PSD95 as a marker for the pre- and post-synaptic terminals of photoreceptor cells located in the outer plexiform layer [50]. Double labeling with TF7 and anti-centrin (Figure 3A-D) revealed CDH23 in the ciliary region, the outer nuclear layer, and in the outer plexiform layer. Double-labeling with TF7 and anti-PSD95 (Figure 3E-H) confirmed localization of CDH23 in the outer plexiform layer. Higher resolution imaging of double-labeling immunofluorescence with TF7 and anti-centrin revealed partial colocalization of CDH23 protein isoforms and centrins in the basal bodies of the ciliary apparatus of photoreceptor cells (Figure 4). Pre-embedding high resolution immunoelectron microscopy, using antibody TF7, localized CDH23 in the basal body of the cilium and its adjacent centriole and confirmed CDH23 as a component of the ciliary apparatus (Figure 5) [42].

CDH23_V3 localizes to the synaptic region of photoreceptor cells

Cryosections through adult mice eyes were double-labeled with the CDH23_V3 isoform specific antibodies TF648 and TF649, and an anti-PSD95, respectively. Immunofluorescence double-labeling showed colocalization of CDH23_V3 and PSD95 at the synaptic terminals of photoreceptor cells (Figure 3I-P). CDH23 immunoreactivity was also observed in this region using antibody TF7 directed against CDH23 cytoplasmic domain (Figure 3E-H). This overlap was expected since antibody TF7 should recognize CDH23 protein isoforms, which include the cytoplasmic domain or portions of it, such as CDH23_V3. No signal was detected in the photoreceptor inner segment and outer nuclear layer with CDH23_V3 specific antibodies TF648 and TF649 (Figure 3I-P).

CDH23_3 is expressed in the synaptic region of auditory and vestibular hair cells

Mouse inner ears hair cells were double-labeled with anti-SNAP25 and CDH23_V3 antibody TF648. SNAP25 localizes to the synaptic region of hair cells [51]. CDH23_V3 immunoreactivity was detected as large punctae in the basal end of vestibular hair cells, partially colocalizing with SNAP-25 in the synaptic region of vestibular (Figure 6A-C) and auditory (Figure 6D-F) hair cells. Similar results were seen with for the CDH23 cytoplasmic domain antibody TF7 and anti-SNAP25 in vestibular hair cells (data not shown) and in the synaptic fibers in the region below auditory hair cell nuclei (Figure 6G-I).

Expression of CDH23 in the monkey retina

Using CDH23 cytoplasmic domain antibody TF7, we detected CDH23 immunoreactivity in the region between the inner and outer segments of photoreceptor cells of monkey retina (where the connecting cilium and basal bodies are present), in the outer plexiform, and in the inner and outer nuclear layers, similar to the pattern observed in the mouse retina (Figure 3). Prominent CDH23 immunoreactivity was observed in the base of the cone inner segments (Figure 7A, C), a pattern of staining not observed in the mouse retina (Figure 3). Some CDH23 immunoreactivity was also detected in rod photoreceptors although less prominent than in the cone photoreceptors, possibly due to their structure and because cone photoreceptor staining is more obvious in frozen sections of monkey retina.

CDH23_V1 is expressed in primate retina but not in mouse retina

We performed western blot analyses of mouse protein extracts from P0, P4, and adult wild-type, and P0 and adult mutant inner ear and retinas using TF7 antibody. Results were obtained from 3 to 4 independent protein extractions replicated 2 to 9 times. Representative western blots (Figure 8) revealed a large band of ~350 kDa corresponding in size to CDH23 longest protein isoform (CDH23_V1) only in P0 and P4 wild-type inner ear (Figure 8A-C). This large band was not detected in wild-type P0, P4, P60, or P90 retinas and was not detected in the waltzer Cdh23^v-6J inner ear or retinal samples (Figure 8A-C). A trace of this high molecular cadherin 23 isoform was detected with antibody TF7 in the P60 inner ear sample (Figure 8B, lane 2), which was consistent with the reported reduction in cadherin 23 in the mature mouse inner ear [28,29,31,32]. Various small bands were also detected both in inner ear and retina that corresponded to the predicted sizes of CDH23_V2, roughly 120 kDa or CDH23_V3, approximately 26 kDa. Some variability in the appearance of the small bands was detected between inner ear or retina, the age of the mouse and the strain, either wild-type or mutant Cdh23^v-6J. A high molecular weight band of roughly 350 kDa was detected in human and monkey retinas (Figure 8C) corresponding in size to the high molecular weight band detected in P0 and P4 inner ear (Figure 8A-C).

High molecular weight CDH23 detected by western analysis was previously reported in the rat, mouse, and human retina [42]. In the effort to reconcile our data with this report we performed western blot analysis of mouse inner ear and retinas using antibodies TF7 and 1D1 [42] directed against mouse and human CDH23 (respectively, Figure 9). We found that antiserum 1D1 recognized protein bands of roughly 210 kDa and roughly 70 kDa in the mouse inner ear and retina. Antibody TF7 also recognized bands of approximately 210 kDa and approximately 70 kDa, corresponding to the bands detected by antibody 1D1. But TF7 recognizes a larger band (roughly 350 kDa), not recognized by 1D1. This band may represent the longest CDH23 protein isoform with a deduced kDa of 370 that is present in the mouse inner ear at early developmental stages.