Fig 1.
Grxcr2 is paralogous to the previously identified deafness-associated gene Grxcr1.
(A) Regions of mouse chromosomes 5 and 18 that contain the Grxcr1 and Grxcr2, respectively, also contain additional conserved gene pairs (Kctd8 and Kctd16; Yipf7 and Yipf5) that are indicative of an ancestral duplication. Cen, centromere; tel, telomere. (B) Amino acid sequence comparison of GRXCR2 orthologs in a range of vertebrate species and of mouse GRXCR1 indicates strong conservation of the N- and C- termini, while the central domains are more divergent. Gray bar, thioredoxin-like domain of GRXCR1; bracket, putative GRXCR1 active site. Asterisks, conserved bipartite arrangement of cysteine residues. Black or gray shading indicates positions at which residues from at least four species are identical or biochemically similar, respectively. (C) Predicted secondary structure of the C-terminal Cys-rich domain of GRXCR2 exhibits significant similarity to Cys-rich domains present in E. coli dnaJ/Hsp40 (P08622) and human DNAJ3 (NP_005138) proteins. PSI-BLAST alignments are shown along with secondary structural information (H, alpha helix; E, beta strand; and C, coil) and associated confidence scores for the GRXCR2 predictions derived from analysis with Phyre2 (9, highest; 0, lowest). E values supporting significance of the PSI-BLAST alignments (three iterations) were 4x10-10 (dnaJ/Hsp40) and 2x10-8 (DNAJ3). (D) Structure of the dnaJ/Hsp40 Cys-rich domain bound to two zinc ions is similar to the top scoring model of the GRXCR2 C-terminal domain predicted by I-TASSER using the MUSTER threading algorithm [36]. The confidence score supporting this GRXCR2 model was -0.41, consistent with an RMSD of 3.7 ± 2.5 angstroms with respect to dnaJ/Hsp40. Secondary structure is indicated by color: red, alpha helix; yellow, beta strand; and green, coil. Zinc ions are indicated in blue and Cys residues comprising zinc binding domain I (ZBDI) and ZBDII are indicated in orange. (E) RT-PCR products corresponding to full-length Grxcr2 transcripts were amplified from cochlear RNA, but not from RNA derived from a variety of other adult tissues. Molecular weight markers are indicated at right, in kilobases. RT, reverse transcriptase.
Fig 2.
Mutational targeting of exon 1 of Grxcr2.
(A) The wild type Grxcr2 locus, the recombinant flox-neo allele after homologous integration of the targeting construct, and the deleted (Δ) allele after Cre recombination are shown. Red arrows indicate the positions of genotyping primers to detect the endogenous (primer pair A and B) and deleted (primer pair A & C) alleles of Grxcr2. (B) Triplex PCR of genomic DNA using primers A, B, and C generated a 532 bp product from wild type mice (+/+), a 490 bp product from homozygous mutants (Δ/Δ), and both products from heterozygotes (+/Δ). (C) RT-PCR of cochlear RNA from heterozygous +/Δ mice generated a 719 bp product, indicating expression of normal Grxcr2 transcripts. The absence of this product from Δ/Δ mutants demonstrated loss of normal Grxcr2 expression. Amplification of Tfrc transcripts from cochlear RNA of both genotypes demonstrated similar RNA quantity and quality. Tfrc, transferrin receptor; RT, reverse transcriptase. (D) GRXCR2 immunoreactivity was demonstrated in stereocilia bundles in the cochlea of Grxcr2 heterozygotes at postnatal day 3 but was absent from Δ/Δ mutants. Phalloidin reactivity indicates actin filament content in the stereocilia of mice of both genotypes. Scale bars indicate 5 μm.
Fig 3.
Grxcr2 mutants exhibit auditory defects including outer hair cell dysfunction.
(A) At 4 weeks of age, Grxcr2 homozygous mutants (n = 7) exhibited increased ABR thresholds in response to pure tones at 4, 12, and 24 kHz relative to wild type (n = 2) and heterozygous (n = 7) littermates (p < 0.05). Vertical bars here and in (B) and (C) represent standard deviations. Asterisks represent probabilities of statistically significant differences in threshold means (*, p < 0.05). (B) Similar ABR threshold increases were demonstrated in Grxcr2 homozygous mutants (n = 7) relative to wild type (n = 2) and heterozygous (n = 6) littermates at 12 weeks of age (p < 0.05). (C) At 4 weeks of age, Grxcr2 homozygotes (n = 5) exhibited significantly reduced DPOAE in response to 12 kHz stimuli, relative to DPOAE of heterozygous littermates (n = 5) (p < 0.05).
Fig 4.
Grxcr2 mutant mice exhibit stereocilia defects at early postnatal ages.
(A) and (C) Scanning electron micrographs of the cochlear sensory epithelium from Grxcr2 heterozygotes demonstrated an organized mosaic of inner and outer hair cells and normal stereocilia bundle morphology at P0 and P7. (B) The cochlear sensory epithelium from Grxcr2 mutant homozygotes exhibited normal organization of sensory cells. Bundle morphology was also relatively normal, but bundles on outer hair cells exhibited slight deviations in orientation at P0. (D) At P7, outer hair cell bundles of Grxcr2 homozygotes exhibited more severe orientation defects and were severely disorganized, including flattened, splayed, and distorted bundles (arrowheads). All scale bars indicate 5 μm. All images are taken from positions in the cochlea approximately one-half turn from the apex. Similar bundle defects were observed in more basal regions of the mutant cochleae.
Fig 5.
Loss of Grxcr2 function results in stereocilia orientation defects.
(A) Schematic representations of the orientation of stereocilia bundles in Grxcr2 heterozygotes and homozygous mutants are shown. Bundle orientation was calculated by identifying the kinocilium at the vertex of the bundle and the mid-point (solid black dot). The angle between a line extended through these two landmarks and a longitudinal reference line through pillar cells was measured. (B) Cochlea from a Grxcr2 heterozygote at P0 was stained with phalloidin (red) to mark the stereocilia and acetylated tubulin (green) to mark the kinocilium. The bundles appeared morphologically normal with a kinocilium at the vertex of each bundle. The angles measured show a tight distribution around 90 degrees with respect to the longitudinal axis of the cochlea. (C) Kinocilia were present at the vertex of bundles from the cochlea of a Grxcr2 homozygote at P0, but bundles were misoriented with significant deviation from 90 degrees.
Fig 6.
Mechanosensitivity is preserved in neonatal hair cells of homozygous Grxcr2 mutant mice.
(A) Differential Interference contrast (DIC) images of the mid-apical turn at P6 shows disorganized hair bundles in Δ/Δ mutants. (B) FM1-43 uptake into hair cells in the same section appears normal, suggesting functional mechanotransduction channels are present. (C) and (D) Comparable transduction currents are evoked in mid-basal outer hair cells of heterozygous mice (C) and homozygous mutants (D) at P2. Bundle displacements ranged from -0.2 to 1.2 μm. Corresponding DIC images depict hair bundles before application of the stimulus pipette. Scale bars indicate 10 μm (A and B) and 3 μm (C and D).
Fig 7.
Grxcr2 mutants exhibit altered responses to linear acceleration.
(A) At 5 months of age, P1-N1 response amplitudes of /Δ mutants (n = 4) were decreased relative to +/Δ (n = 5) and +/+ (n = 3) mice, consistent with significantly elevated thresholds observed in the Δ/Δ mutants (ANOVA, F(2,11) = 11.442, p = 0.003). Post hoc analysis revealed that Δ/Δ mutant thresholds were significantly higher than +/Δ (p = 0.004) and +/+ (p = 0.002) mice. Thresholds for +/Δ and +/+ mice were statistically equivalent. (B) P1 response latencies of Δ/Δ mutants were increased at +6 dB (ANOVA, F(2,11) = 5.23, p = 0.03), indicating altered response timing in the mutants. Post hoc analysis revealed that latencies for Δ/Δ mutants were significantly longer than +/Δ mice (p = 0.012).
Fig 8.
Grxcr2 transcript and protein exhibit peak levels in the cochlea at early postnatal time points.
(A) Grxcr2 transcript levels in the cochleae of C57BL/6J mice at early postnatal time points and in adults were determined by qRT-PCR and normalized to those present at birth (P0). Grxcr2 expression peaked at P3 and declined through the first postnatal week and young adulthood (P27). Vertical bars represent standard deviations. (B) The sensory epithelium from the cochleae of C57BL/6J mice was immunostained using an anti-GRXCR2 antibody and phalloidin to mark the stereocilia bundles. Outer hair cell bundles of mice at P0 and P3 exhibited robust GRXCR2 reactivity while bundles from P7 mice exhibited lower reactivity. Images are derived from the middle regions of the cochlea. Scale bars indicate 5 μm. (C) Stereocilia of sensory hair cells in the utricle at P0 were also reactive with the anti-GRXCR2 antibody (left, anti-GRXCR2; right panel, anti-GRXCR2 signal merged with phalloidin-rhodamine signal). Scale bars indicate 5 μm.
Fig 9.
Hypomorphic expression of Grxcr2 is associated with progressive hearing loss and stereocilia bundle defects.
(A) Grxcr2 transcript levels in the cochleae of +/+ mice at 3 months of age were compared by qRT-PCR to those in mice carrying either the deleted or the flox-neo alleles in heterozygous or homozygous forms (N = 2 for Δ/Δ mice; N = 3 for all other genotypes). Flox-neo/flox-neo mice expressed less than half of the level of Grxcr2 transcripts observed in +/Δ mice. The mean fold change for Δ/Δ mice was 0.00075, SD = 0.0017. Vertical bars represent standard deviations. Asterisks represent probabilities of statistically significant differences in mean fold changes (*, p < 0.02; **, p < 0.0001). (B) At 2 months, ABR thresholds in response to 12 and 24 kHz stimuli were similar in the two genotypes whereas at 4 kHz Grxcr2 flox-neo homozygotes had elevated hearing thresholds (p < 0.05). N = 5 mice of each genotype. (C) At 5 months, flox-neo homozygotes exhibited significant ABR threshold shifts in response to both 4 and 12 kHz stimuli (p < 0.05), indicating progression of hearing loss relative to heterozygote controls. N = 6 (+/flox) and 9 (flox-neo/flox-neo) mice. (D) Scanning electron microscopy of cochlear sensory epithelia demonstrated slight abnormalities in outer hair cell bundle orientation and organization of the flox-neo homozygotes at 2 months. (E) By 7 months, stereocilia bundles of flox-neo homozygotes exhibited extensive defects in orientation and organization. All images indicate cochlear epithelia approximately one-half turn from the apex. Similar bundle defects were observed in more basal regions of the mutant cochleae. All scale bars indicate 10 μm.