Missense mutations in POU4F3 cause autosomal dominant hearing impairment DFNA15 and affect subcellular localization and DNA binding

Abstract

In a Dutch pedigree suffering from autosomal dominant nonsyndromic hearing impairment (ADNSHI), linkage was found to the locus for DFNA15, with a two-point logarithm of the odds (LOD) score of 5.1. Sequence analysis of the POU4F3 gene that is involved in DFNA15 revealed the presence of a missense mutation (c.865C>T), segregating with the deafness in this family. The mutation is predicted to result in the substitution of a phenylalanine residue for a leucine residue (p.L289F) in the POU homeodomain of the transcription factor POU4F3. Mutation analysis of the POU4F3 gene in 30 patients suffering from dominantly inherited hearing impairment revealed a second novel missense mutation (c.668T>C), resulting in the substitution of a proline for a leucine residue (p.L223P) within the POU-specific DNA-binding domain of the protein. In a computer model describing the structure of the two DNA-binding domains, the alterations are predicted to affect the tertiary structure of these domains. Transient transfection studies showed that whereas the wild-type POU4F3 is located almost exclusively in the nucleus, part of the mutant proteins was also present in the cytoplasm. In addition, both mutant proteins showed greatly reduced capability for binding to DNA as well as transcriptionally activating reporter gene expression. Together, our results describe the identification of the first missense mutations in POU4F3 causing DFNA15. Furthermore, mutations in this gene do not seem to be a rare cause of hearing impairment in the Dutch population, and the POU4F3 gene may thus be suitable for implementation in diagnostic testing.

FIGURE 1

A: Pedigree of Family W05−549. Two individuals for which audiograms are shown (B) are indicated by 1 and 2, respectively. Two phenocopies that were affected but did not carry the mutation are indicated by PC. B: Serial audiograms of two of the affected members from Family W05−549 (left panel 1, right panel 2). Shown are binaural mean thresholds in decibels (dB) for each frequency in kilohertz (kHz).The legend on the right shows the age in years (y) at which the audiometric tests were performed.

FIGURE 2

A: Schematic representation of the human POU4F3 protein.The POU-specific and the POU homeodomain are depicted in red and blue, respectively. B: Sequence comparison of the POU-specific and the POU homeodomain of several vertebrate POU4F3 proteins. Identical residues in all sequences are black on a white background, whereas amino acids that are identical in at least four species are gray on a black background. Conserved changes are black on a light gray background, whereas similar amino acids are depicted in white on a black background. The two leucine residues (L223 and L289) that were found to be substituted by a proline and phenylalanine in the two families with hearing impairment are indicated by an arrow. The positions of the various α-helices in the two DNA-binding domains are underlined. Accession numbers of the POU4F3 protein sequences: human: Q15319; mouse Q63955; rat XP 344676; chicken NP 990090; frog AAG17008; and zebrafish NP571353. C: Molecular model of the human POU4F3 protein.The part of the wild-type POU4F3 protein containing the DNA-binding domains is depicted, with the various α-helices represented as cylinders. The target DNA is indicated in purple. D: Graphic representation of the predicted effect of the p.L223P mutation. The leucine residue at position 223 is replaced by a proline (depicted in green), resulting in a clash with leucine residue 215 (in yellow) that is present in the second α-helix of the POU-specific domain. E: Graphic representation of the predicted effect of the p.L289F mutation. The leucine residue 289 is replaced by a phenylalanine. Two possible rotamer orientations of the aromatic side chain are depicted in green. In one case (upper orientation), the phenylalanine side chain clashes with the aromatic ring of a highly conserved tryptophan residue (W321) in the third α-helix of the POU homeodomain, whereas the other (less favored) orientation clashes mainly with the side-chains of two adjacent leucine residues. All residues that may somehow clash with the phenylalanine at position 289 are depicted in yellow. Images as presented in panels C–E were made using the YASARA NOVA program [Krieger et al., 2002]. [Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com.]

FIGURE 3

A: Immunofluorescence analysis of transiently transfected HA-tagged wild-type and mutant POU4F3 proteins. Representative examples are shown, with for the two mutant POU4F3 proteins a number of cells expressing POU4F3 outside the cell nuclei indicated by a white arrow. Images are presented as DAPI (staining cell nuclei), anti-HA (detecting the transfected HA-tagged POU4F3 proteins), and merged pictures (DAPI in blue, HA-tagged POU4F3 in green). B: Quantification of the number of transfected cells expressing HA-tagged POU4F3 outside the nucleus. Slides were blinded and for each transfected construct in duplicate, 100 cells were evaluated. C: Western blot analysis of lysates of transfected cells, stained with an anti-HA antibody. Beta-actin levels are used for comparison. [Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com.]

FIGURE 4

A: EMSA using a labeled Pou4f3 binding site as probe. Reactions contained untreated lysate (lane 2), in vitro translated wild-type Pou4f3 (lanes 3, 4, 9, and 10), and/or translated mutant L223P (lanes 5, 6, and 9) or L289F (lanes 7, 8, and 10) in the presence (lanes 4, 6, and 8) or absence of a 500-fold excess amount of cold specific competitor oligonucleotides. Wild-type Pou4f3 formed a strong protein-DNA complex that could be abolished by excess amounts of cold competitor, whereas p.L223P and p.L289F exhibited no detectable DNA binding activity and did not appear to significantly interfere with DNA binding of the wild-type protein. B: Relative luciferase activities following cotransfection of wild-type and mutant Pou4f3 expression constructs with the Prox or Prox3 luciferase reporter plasmids. Cotransfection was conducted in 293T cells and luciferase activities were measured 48 hr after transfection. Transfection efficiency was controlled by measuring Renilla luciferase activities following cotransfection with the pRL-TK Renilla luciferase reporter plasmid. Results represent means ±SD of triplicate assays in a single experiment. Statistical significance: *P<0.05; **P<0.001.