Pedigree, audiological data, genetic data, and locus mapping. a The consanguineous family of Iranian origin with hearing loss and segregation of the CLRN2 c.494C > A variant. Linked haplotypes harbouring the pathogenic variant coloured in red with meiotic recombination SNP markers underlined. SNP positions are annotated using the GRCh37 human genome assembly. b Pure-tone audiograms from affected individuals IV-1 (red) and IV-6 (blue), as well as an unaffected heterozygous individual V-1 (green). Air-conduction thresholds in dB HL for the right and left ears are represented by circles and crosses, respectively. Bone-conduction thresholds are represented by < and > for right and left ears, respectively, and confirm a sensorineural hearing loss in the affected individuals. c Linkage mapping reveals a 14.96 Mb locus on chromosome 4 containing CLRN2. d Sequence electropherograms showing the homozygous, heterozygous and WT images of the CLRN2 c.494C > A; pThr165Lys pathogenic variants

Conservation of the p.Thr165 residue, and clarin 1/clarin 2 alignment. a Overview of clarin 2 protein and modular structure of the PMP-22/EMP/EP20/Claudin superfamily, with amino acid residue coordinates and position of the p.(Thr165Lys) substitution shown (upper panel). An alignment of the amino acid sequences from the segment of clarin 2 (represented by dashed lines) from vertebrate species shows the Thr165 position (asterisk) is well conserved among vertebrates. b Alignment of clarin 2 (UniProtKB: A0PK11, upper alignment) and clarin 1 (UniProtKB: P58418, lower alignment) amino acid residues. Transmembrane domains are marked in grey, conservation is shown in yellow, and consensus sequences are shown below for the 232 amino acid proteins. Missense and nonsense variants in clarin 1 (Deafness Variation Database v8.2) and clarin 2 (present study, asterisk) are marked in red. c The predicted secondary structure of human clarin 2 (NP_001073296.1) wild-type (Thr165) and mutated (Thr165Lys) protein. H represents alpha-helix, S represents beta-strand and C represents coil

Analysis of the CLRN2 c.494C > A variant on splicing. a Schematic illustration of the mini-gene splice construct design. Genomic representation of CLRN2, including the position of the missense variant c.494C > A (arrow) on exon 3 with 3′ UTR (green), and the 5′ UTR, as well as exons 1 and 2 (grey) (upper panel). Regions captured by mini-gene PCR primers are represented in brown. Schematic illustration of the mini-gene splice construct including exon 3 and its flanking sequence (green) cloned into multiple cloning sites (SalI and SacII sites) of pET01 backbone vector (lower panel). Blue boxes represent native exons of the pET01 vector. b RT-PCR of transcripts from post-mini-gene transfected COS-7 cells. Amplicons derived from the transcripts of WT (CLRN2), a benign CLRN2 polymorphism (rs117875715, chr4(GRCh37):g.17,528,480G > A), the CLRN2 c.494C > A variant and a negative control, were visualized on a 1.5% agarose gel. The SNP, rs117875715, was used to test and validate the designed WT and mutant mini-gene assay. The ~ 650 bp amplicon was associated with the WT and validation control rs117875715. The amplicon derived from the CLRN2 c.494C > A transcripts showed two bands: a 650 bp band and a larger ~ 1360 bp band, indicating retention of intron separating the donor site of the 5′ exon and the acceptor site of CLRN2 exon 3. c Retention of intron in CLRN2 c.494C > A mini-gene results in a stop codon (TGA) after CLRN2 exon 2

Clarin 2 is required for the inner ear function in zebrafish. a RT-qPCR of clrn2 mRNA expression from 1 to 120 hpf of WT embryos/larvae. clarin 2 mRNA expression can be detected starting from 18 hpf and then increased throughout development. Data shown are mean ± SD and compared to 18 hpf. b RT-qPCR of clrn2 mRNA expression in different adult tissues. Data shown are mean ± SD and compared to muscle. c–d Whole-mount in situ hybridization (WISH) using antisense clrn2 probe reveals the inner ear expression of clrn2 mRNA (relative dark purple color, black arrowhead) at 3 (c) and 5 (d) dpf embryos. Sense clrn2 probe was used as negative control and relative light purple color is considered as background. clrn2 mRNA was consistently expressed in hair cells within inner ear macula (c–d) with lined and arrayed structure. A known hair cell marker pvalb9 was used as an indicator for hair cells in the inner ear of 3 dpf embryos (c). Cryosection was performed after clrn2 WISH at 5 dpf to confirm the small patch of signal on the macula is from hair cells rather than supporting cells (d, black arrow lower panel). Scale bar = 100 µm, except lower panel in D (20 µm). e RT-qPCR of clrn2 mRNA expression level was decreased 70% in clrn2 crispants compared to uninjected larvae, indicating clrn2 was successfully knocked out (p = 2.06E-06). Data shown are mean ± SD. ***p < 0.001, two-tailed unpaired Student’s t test. f Acoustically evoked behavioral responses (AEBR) in clrn2 wild type and crispants reveal significant reduction of sound induced responses. g Phalloidin staining on clrn2 crispants show that the hair cells in the inner ear anterior and posterior maculae display splayed, thin and split structures (purple arrowheads). A, anterior to the left. D, dorsal to the top. V, ventral to the top. Scale bar = 10 µm

Clarin 2 is required for hearing function in mouse. a The genomic structure of mouse Clrn2 (ENSMUST00000053250), and domains of the encoded tetraspan-like glycoprotein (232 amino acids). The positions of the transmembrane (TM) domains (dark green) and the structures of the WT Clrn2 and Clrn2^del629 alleles are indicated. Deletion of exon 2 leads to a shortened clarin 2 lacking the two central transmembrane domains. b ABR threshold measurements at P21 (± 1 day) show that Clrn2^{del629/del629} mice (red) exhibit a severe-to-profound hearing loss affecting all frequencies tested, with thresholds at 80 dB SPL and beyond. Age-matched Clrn2^+/+ (black) and Clrn2^del629/+ (grey) controls display thresholds within the expected range (15–40 dB SPL). Averaged DPOAE responses at P28 (± 1 day), showing significantly reduced responses in Clrn2 ^{del629/del629} mice. Data shown are mean ± SD. **p < 0.001, one-way ANOVA. c Pseudo-colored scanning electron micrographs illustrate the three full rows, tallest (red), middle (blue) and short (yellow), of P28 (± 1 day) stereocilia in IHC and OHC hair bundles. Unlike the fragmented hair bundle in Clrn1^−/− mice, lack of clarin 2 does not affect the shape of IHC or OHC hair bundles. However, all the short row stereocilia have completely or partially regressed in the absence of either clarin protein. Scale bar = 1 µm