A de novo silencer causes elimination of MITF-M expression and profound hearing loss in pigs

Abstract

Background: Genesis of novel gene regulatory modules is largely responsible for morphological and functional evolution. De novo generation of novel cis-regulatory elements (CREs) is much rarer than genomic events that alter existing CREs such as transposition, promoter switching or co-option. Only one case of de novo generation has been reported to date, in fish and without involvement of phenotype alteration. Yet, this event likely occurs in other animals and helps drive genetic/phenotypic variation.

Results: Using a porcine model of spontaneous hearing loss not previously characterized we performed gene mapping and mutation screening to determine the genetic foundation of the phenotype. We identified a mutation in the non-regulatory region of the melanocyte-specific promoter of microphthalmia-associated transcription factor (MITF) gene that generated a novel silencer. The consequent elimination of expression of the MITF-M isoform led to early degeneration of the intermediate cells of the cochlear stria vascularis and profound hearing loss, as well as depigmentation, all of which resemble the typical phenotype of Waardenburg syndrome in humans. The mutation exclusively affected MITF-M and no other isoforms. The essential function of Mitf-m in hearing development was further validated using a knock-out mouse model.

Conclusions: Elimination of the MITF-M isoform alone is sufficient to cause deafness and depigmentation. To our knowledge, this study provides the first evidence of a de novo CRE in mammals that produces a systemic functional effect.

Keywords: De novo silencer; Hearing loss; MITF-M; Pig; Waardenburg syndrome; cis-regulatory element.

Fig. 1

Cochlear morphology and auditory electrophysiology defects of albino pigs. a Gross image of a normal pig and an albino pig. b Results of auditory brainstem response tests showing profound hearing loss of albino pigs. The raw data is provided in Additional file 5: sheet 1 Data of ABR tests (pigs). c Scanning electron microscopy images showing missing or fused (star) stereocilias of inner (arrow) and outer (arrowhead) hair cells in albino pigs. d Images showing that the stria vascularis (SV) of albino pig are remarkably thinner than that of normal pig. e Image showing lack of intermediate cells in the SV of albino pigs. Marginal cell layer, arrowheads; intermediate cell, stars; basal cell, arrows; spiral ligament, Spl. f and g The average values of endolymphatic potential and scala media potassium concentration in albino pigs were significantly lower than in normal pigs (raw data in Additional file 5: sheet 2 Data of EP and sheet 3 Data of K+ concentration). Error bars indicate the standard deviations. Scale bars in c = 100 μm, in d = 50 μm, and in e = 5 μm

Fig. 2

Genome-wide association mapping of the hearing loss trait in albino pigs. a The strongest association was identified on chromosome 13 by case-control association (P _genome = 0.00242) using the whole-genome data analysis toolset PLINK. b Haplotype sharing analysis showed a perfect concordance of a haplotype with hearing loss phenotype in the mapped population. A 767-kb associated interval (from INRA0040190 to ALGA0070147) was defined by five single nucleotide polymorphism markers in that haplotype and proximal recombinant markers. c Mitf is the only known gene located in that region

Fig. 3

Expression analysis of Mitf transcriptional variants and isoforms. a Schematics of splicing structure in the porcine Mitf transcript variants (leading to MITF-A, MITF-H and MITF-M) detected in the cochlea. The specific fragments of transcript variants used for quantitative PCR in this study are indicated by blue lines. b Expression profiles of MITF-A, MITF-H and MITF-M during cochlear development were examined by reverse transcription-PCR. MITF-M expression was present at detectable levels in the Mitf  ^R/R cochlea, but not in the Mitf  ^r/r cochlea. c Immunoblotting analysis of MITF isoforms in the cochlea and skin. MITF-M expression was detectable in the Mitf  ^R/R cochlea and skin, but not in samples from the Mitf  ^r/r pigs. d Differential level of expression of the Mitf exons in Mitf  ^R/r and Mitf  ^r/r stria vascularis (SV). The M-exon showed a 11.5-fold decrease in the Mitf  ^r/r SV. The fold change of each exon is estimated by comparing the normalized read count of each exon between Mitf  ^R/r and Mitf  ^r/r SV in the RNA-seq assay. Raw data for this is provided in Additional file 5; sheet 4, Data of Mitf exon fold change

Fig. 4

A new generated silencer in the M-promoter eliminated Mitf-m transcription. a Transcriptional activity analysis of the Mitf-m promoter from the R and r alleles. The reporter constructs are shown on the left, and the corresponding relative luciferase activity measured in transient transfection assays is shown on the right. The luciferase activity of pGL3-r-7.8 k was significantly lower than that of pGL3-R-7.8 k. There was no significant difference between the R and r alleles when constructs were shorter than 7.8 k. Error bars indicate the standard deviations. The results shown are for one experiment with four technical replicates. Raw data for this and three additional experiments with similar results are provided in Additional file 5; sheet 5, Data of reporter assay. b Schematics of the M-promoter. The sequence variations between the R and r alleles are labeled. INS, insertion; DEL, deletion. c Schematics of the M-promoter from –7513 bp to –7609 bp relative to the transcription start site of the M-exon. Sequence differences between R and r allele are indicated with a red box. The new sites showing consensus sequence for SOX protein binding are underlined in red. The oligonucleotide probes designed for electrophoretic mobility shift assay (EMSA) are highlighted. d EMSA shows the specific binding of the nuclear proteins to the r2 probe, and absence of binding to the R2 probe. In vitro incubation was performed using the indicated nuclear extracts, probe and unlabeled oligonucleotides (cold probes). C1, complex 1; C2, complex 2; R, R2 probe; r, r2 probe; N, random (negative control) probe

Fig. 5

Phenotypes of the Mitf-m knock-out mice. a Schematic of the Mitf-m targeting technical process. The region of the Mitf gene containing exons M, 2, 3, and 4 are shown at the top. The targeting vector with a floxed-neomycin cassette in the M-promoter/M-exon region is shown in the middle. The resultant Mitf gene portion after targeting (Mitf  ^mi-ΔM allele) is shown at the bottom. b Mitf  ^+/+ had a black coat color, and Mitf  ^{mi-ΔM/mi-ΔM} had a white coat color and black eyes. c The auditory brainstem response thresholds were 20–30 dB SPL for the Mitf  ^+/+ mice, and 100–110 dB SPL for the Mitf  ^{mi-ΔM/mi-ΔM} mice (from 4 to 32 kHz). The raw data is provided in Additional file 5: sheet 6 Data of ABR tests (mice). d Mitf-m was not expressed at detectable levels in the Mitf  ^{mi-ΔM/mi-ΔM} cochlea (red arrow), but was expressed at detectable levels in the Mitf  ^+/+ cochlea. There was no difference observed between the expression levels of Mitf-a and Mitf-h in Mitf  ^{mi-ΔM/mi-ΔM} and Mitf  ^+/+ mice. Error bars in c and d indicate the standard deviations. Raw data in Additional file 5: sheet 7 Data of mouse Mitf qPCR (Ct) and sheet 8 Data of mouse Mitf qPCR (FC). e In the Mitf  ^{mi-ΔM/mi-ΔM} cochlea, most of the stereocilias of inner hair cells (arrows) and outer hair cells (arrowheads) were fused or missing (stars). f The stria vascularis of Mitf  ^{mi-ΔM/mi-ΔM} cochlea are significantly thinner and shorter than that of Mitf  ^+/+ cochlea

Fig. 6

Schematics showing the genetic effect of the causative mutation. In Mitf  ^R/R stria vascularis (SVs) (before duplication). SOX proteins cannot recognize and bind to the M-promoter and MITF-M was normally transcribed. In Mitf  ^r/r SVs, a new consensus site for SOX protein binding, which resulted from the 14-bp duplication, created a de novo silencer in the M-promoter. SOX proteins ectopically binding to that silencer may repress the transcription of Mitf-m (after duplication)