Rapid progression of hearing loss in DBA/2J mice. Average 16 kHz ABR thresholds are shown for five DBA-related strains tested from 4-28 weeks of age. The data points for DBA/2J and those for DBA/2NCrl, the most closely related strain, are connected by lines to contrast the difference in their hearing loss progression. Bars represent the standard deviations of the means to illustrate the degree of variation observed at each time point. ABR thresholds for all auditory test stimuli, standard errors, and numbers of mice tested for each strain and test age combination are given in Supplementary Table 1.

FSCN2 in mouse cochlea and vestibular system. a-c. Developmental appearance of FSCN2 around P10 in bundles of inner hair cells and by P30 in bundles of outer hair cells. d. FSCN2 in P10 cochlea; some labeling is seen in bundles of outer hair cells. Image is maximum projection of ten images, taken at 0.3 μm z-intervals. e. Magnified view of two inner hair cell (IHC) bundles from panel d. FSCN2 is primarily in the tallest stereocilia, near tips (asterisk). f. Similar distribution of FSCN2 in P30 DBA/2J mice. g. FSCN2 in mouse utricle hair bundles; note localization towards tips. Image is a z-projected (maximum) reslice image from stack acquired axially through the utricular epithelium. h. ACTG1 developmental appearance around P10 in mouse cochlea. i. Quantitative RT-PCR showing elevation of FSCN2 transcripts around P10. No increase in Actg1 or Actb transcript levels was seen. Panel heights: a-c, f, h, 70 μm; d, 60 μm; e, g, 19 μm.

Rapid degeneration of cochlear hair cell bundles in DBA/2J mice. In the basal turn of the cochlea, bundles of outer hair cells (OHCs) and inner hair cells (IHCs) of DBA/2J mice (2J) appeared normal at two weeks of age (a), but at 1 month (b) and 3 months (c) of age most OHC bundles already showed signs of degeneration and some were missing (indicated by asterisks). By 6 months of age in the base (d), all OHCs were missing and the few remaining IHCs had degenerated bundles (indicated by arrow); in the middle turn (e), bundles were severely disheveled (arrowhead) and often had lost most rows of stereocilia (arrow). In contrast, all OHC and IHC bundles of the control DBA/2NCrl mice (2NCrl) at 3 months (g) and 6 months (f, h) of age appeared normal, and only a few OHC bundles were missing (asterisks in h); the bundles of the remaining OHCs appeared normal. A fourth row of OHCs was seen at frequent intervals throughout the length of the cochlea in both DBA/2J and DBA/2NCrl mice (indicated by arrow in h). Panel widths: a-d, g, h, 29 μm; e, f, 7.3 μm.

FSCN2 in chick and Xenopus hair bundles. a. Low-power view of actin and FSCN2 in E20 chick utricle. Areas magnified in (b) and (c) are indicated. b. High power view of single bundle. Note high levels of FSCN2 in rows of tallest stereocilia, adjacent to kinocilium notch. c. Reslice profile image of bundles showing localization of FSCN2 near stereocilia tips; tallest stereocilia are crushed against the coverslip. Note FSCN2 absence from a short, presumably immature bundle (asterisk). d. Reslice profile image showing FSCN2 localization near tips of adult Xenopus saccule bundles. e. Immunogold electron microscopy localization of FSCN2 in E20 chick utricle. All gold particles in micrograph are circled. Inset, magnification of region indicated with dashed lines. f. Differential localization of crosslinkers in E20 chick utricle. Note concentration of FSCN2 towards tips and PLS1 in middle of bundles. Neither crosslinker is abundant at bundle base. g. Quantitation of normalized actin signal, FSCN2/actin ratio, and PLS1/actin ratio in 14 cells from three utricle preparations. Cartoon behind data depicts typical bundle morphology; gray box indicates region chosen for profile analysis. h. Scanning electron micrograph showing typical chick utricle hair bundles; note the length of the tallest stereocilia. Abbreviations: KC, kinocilium. SC, stereocilia. Panel heights: a, 52 μm; b, 8 μm; c, 12 μm; d, 12 μm; f, 32 μm. Scale bars in panel e: main panel, 500 nm; inset, 100 nm. Scale bar in h: 1 μm.

Identification and quantitation of FSCN2 in chicken vestibular stereocilia. a. Structure of FSCN2 protein. Colors indicate identity of residues in three-way comparison of chicken, mouse, and Xenopus FSCN2. Peptides used for antibody generation are indicated by black bars and antibody numbers; the position of the mutation in DBA/2J is indicated with an asterisk. The positions of the four beta-trefoil domains are indicated below. b. Representative MS2 spectrum of a FSCN2 peptide identified by tandem mass spectrometry from a chicken hair bundle preparation. The intensity of peptide fragments, typically fragmented at peptide bonds during collision-induced dissociation, is plotted against the mass-to-charge ratio (m/z). Note the nearly-complete y- and b-ion ladders, representing fragmentation sequentially at successive peptide bonds from the N- and C-termini. c. Distribution of FSCN2 peptides identified by X! Tandem algorithm from hair bundles. The width of each bar indicates the length of the identified peptide, the position of the bar along the x-axis corresponds to its position in the FSCN2 sequence, the height of the bar indicates the number of identical peptides identified, and the color of the bar corresponds to log(e), the statistical significance for peptide identification, averaged over all identical peptides. Note that the peptide analyzed in (b) was one of those represented by the black bar at ∼380 residues. d. Crosslinker/actin abundance ratios in E20 chick and P5 rat vestibular hair bundles. e. Immunoblot detection of FSCN2 in purified hair bundles with antibody 36. Bu, 100 ng total hair-bundle protein; Ag, amount of agarose equivalent to that in bundles lane; Ut, 2 μg peeled utricular epithelium; BP, 2 μg peeled basilar papilla epithelium; Re, 1.5 μg retina; and Br, 2 μg total brain protein. f,g. Profile scans of bundle and utricular epithelium lanes from e.

The wildtype C57BL/6J (B6) allele of Fscn2 rescues hearing loss in congenic and transgenic lines of DBA/2J (D2) mice. a. Markers defining the B6-derived distal Chr 11 regions for each of the two D2.B6 congenic lines, designated Long (congenic region ∼ 26 Mb) and Short (congenic region ∼ 3 Mb). b, c. ABR threshold means and standard errors for heterozygous D2.B6 congenic line mice (D2/B6; Long n = 8, Short n = 8) and littermate D2 mice (D2/D2; Long n = 6, Short n = 14), tested at two months of age. Statistically significant ABR threshold differences between B6/D2 and D2/D2 mice are marked by asterisks: * p<0.05; ** p<0.001. ABR thresholds of age-matched B6 mice (B6/B6; n = 20) are shown for comparison. d. Fscn2 transgene integration. The D2 mutation of Fscn2 ablates an Eag I restriction site that is present in the wildtype B6 allele, allowing detection of B6-derived Fscn2 transgenes. After treatment with Eag I, Fscn2 PCR products of genomic DNA from mice with B6-derived transgenes include a diagnostic 121 bp fragment. The abundance of this B6-specific fragment, relative to the 216 bp endogenous D2 fragment, is much lower in Line 1 than in Line 2 transgenic mice. e. Fscn2 transgene expression. PCR primers specific to exons 1 and 2 of Fscn2 were used to amplify a 671 bp product from cDNAs derived from brain and inner ear RNA. After digestion of this product with Eag I, diagnostic B6-specific fragments of 353 and 95 bp could be detected in Line 2 (L2) transgenic mice, but not in Line 1 (L1) transgenic mice, which showed only the 448 and 223 bp fragments of D2 mice. f, g. ABR threshold means and standard errors of DBA/2J mice with B6-derived Fscn2 transgenes (D2+Tg; Line 1 n = 12; Line 2 n = 19) and their non-transgenic littermates (D2; Line 1 n = 10, Line 2 n = 20), tested at one month of age. ABR thresholds of age-matched DBA/2NCrl control mice (2NCrl; n = 3) are shown for comparisons. Statistically significant ABR threshold differences between D2+Tg and D2 mice are marked by asterisks: * p<0.05; ** p<0.001.

A SNP variant within the Fscn2 gene is unique to the DBA/2J strain and changes a highly conserved amino acid. a. Nucleotide variation among 20 inbred mouse strains for the rs26996001 SNP, which is located within exon 1 of the Fscn2 gene. DBA/2J is the only strain examined with an adenine (A) at this position; all other strains have guanine (G). The G to A nucleotide substitution causes an amino acid change from arginine (R) to histidine (H) at position 109 of the mouse FSCN2 protein (NP_766390). b. Evolutionary conservation of the FSCN2 arginine (R) residue that is mutated in DBA/2J mice. This residue (highlighted in blue) is conserved in fascin proteins of all mammalian species examined, as well as fascins of other vertebrate species. c. Genealogy of DBA-related inbred mouse strains and their association with the Fscn2 SNP. All are direct descendants of the original dba strain developed by C.C. Little in 1909. Shown are the approximate dates when these DBA-related inbred strains were separated from one another (Bailey, 1978; Zylstra et al., 2003). DBA/2J is the only strain with the G>A missense mutation in Fscn2; this mutation occurred sometime between 1951 and 1975, as deduced from archived DNA samples of DBA/2J strain mice. d. Molecular model of FSCN1 (Protein Databank Structure 1DFC) showing the beta-trefoil domains (1-4), the regulatory phosphorylation target S39, and the predicted location of the R109 residue of FSCN2 (R110 in FSCN1) that is mutated in DBA/2J mice. e. Hearing loss in F2 intercross progeny of (DBA/2J × DBA/2NCrl) F1 hybrids strongly associates with homozygosity for the DBA/2J allele of Fscn2. Shown are the average 16 kHz ABR thresholds and standard errors for each of the three Fscn2 genotypes of F2 mice: D2J/D2J (homozygous for the DBA/2J mutation, n=9), D2J/D2NCrl (heterozygous for the mutation, n=7), and D2NCrl/D2NCrl (homozygous for the wildtype DBA/2NCrl allele, n=5), all tested at 6 weeks of age. The DBA/2J mutant allele appears recessive to the DBA/2NCrl wildtype allele.