A spontaneous mouse deletion in Mctp1 uncovers a long-range cis-regulatory region crucial for NR2F1 function during inner ear development

Abstract

Hundreds of thousands of cis-regulatory DNA sequences are predicted in vertebrate genomes, but unlike genes themselves, few have been characterized at the functional level or even unambiguously paired with a target gene. Here we serendipitously identified and started investigating the first reported long-range regulatory region for the Nr2f1 (Coup-TFI) transcription factor gene. NR2F1 is temporally and spatially regulated during development and required for patterning and regionalization in the nervous system, including sensory hair cell organization in the auditory epithelium of the cochlea. Analyzing the deaf wanderer (dwnd) spontaneous mouse mutation, we traced back the cause of its associated circling behavior to a 53 kb deletion removing five exons and adjacent intronic regions of the poorly characterized Mctp1 gene. Interestingly, loss of Mctp1 function cannot account for the hearing loss, inner ear dysmorphology and sensory hair cell disorganization observed in dwnd mutants. Instead, we found that the Mctp1^dwnd deletion affects the Nr2f1 gene located 1.4 Mb away, downregulating transcription and protein expression in the embryonic cochlea. Remarkably, the Mctp1^dwnd allele failed to complement a targeted inactivation allele of Nr2f1, and transheterozygotes or Mctp1^dwnd homozygotes exhibit the same morphological defects observed in inner ears of Nr2f1 mutants without sharing their early life lethality. Defects include improper separation of the utricle and saccule in the vestibule not described previously, which can explain the circling behavior that first brought the spontaneous mutation to attention. By contrast, mice homozygous for a targeted inactivation of Mctp1 have normal hearing and inner ear structures. We conclude that the 53 kb Mctp1^dwnd deletion encompasses a long-range cis-regulatory region essential for proper Nr2f1 expression in the embryonic inner ear, providing a first opportunity to investigate Nr2f1 function in postnatal inner ears. This work adds to the short list of long-range regulatory regions characterized as essential to drive expression of key developmental control genes.

Keywords: Cis-regulation; Deafness; Inner ear development; Long-range enhancer; Mctp1; Nr2f1.

Figure 1.. Auditory dysfunction and inner ear abnormalities of dwnd/dwnd mice.

(A) Mice homozygous for the dwnd mutation are hearing impaired. Means (± SEM) of ABR thresholds of 16 dwnd/dwnd mice tested at one month of age and 12 tested at six months of age were compared with those of 7 +/dwnd control mice tested at six months of age for 8, 16, and 32 kHz test stimuli. Thresholds of mutant mice were about 20-30 dB higher than those of controls at one month of age and about 25-45 dB higher at six months of age. (B-E) Cochleae of dwnd/dwnd mice have regions with supernumerary hair cells. SEM images of organ of Corti surface preparations from cochleae of P8 mice show extra inner hair cells (IHC) at the cochlear base (C) and an extra row of outer hair cells (OHC) at the apex (E) of dwnd/dwnd mice as compared with corresponding +/+ controls (B, D). Scale bar for B-E (shown in B), 10 μm. (F) The overall size of the cochlea (indicated by arrows) in dwnd/dwnd mice is smaller than that of +/+ controls. Dissected inner ears of adult mice that also carry a copy of an Lhfpl5-LacZ reporter gene were cleared and the cochlear spiral exposed, showing rows of hair cells stained with LacZ (see Methods). Scale bar, 0.5 mm. (G) Inner ear paintfills of P1 (newborn) mice. In addition to a shorter cochlear duct (cd), the dwnd/dwnd inner ear has a smaller saccule (s) that is not separated from the utricle (u) and a much larger cochleosaccular duct (csd) than the +/+ control. Scale bar, 0.5 mm. (H-J) Cochlear lengths and hair cell densities in dwnd/dwnd and control mice. The density of inner hair cells near the apex and near the base of the cochlea is greater in dwnd/dwnd mutants than controls (H), and the density of outer hair cells near the cochlear apex is much greater in dwnd/dwnd mice than controls (I). Although hair cell densities are greater, the length of the cochlear duct is shorter in dwnd/dwnd mutants than controls (J), suggesting that total hair cell counts are similar. Means and standard error bars are shown for the average measurements of individual mice of each genotype (ten ears of +/+ mice and nine ears of dwnd/dwnd mice). Statistical significance of mean differences was determined using a student’s t-test. * p<0.01; **** p<0.0001.

Figure 2.. The dwnd mutation of the Mctp1 gene.

(A) The structure of the Mctp1 gene is shown as numbered vertical lines representing exons and horizontal connecting lines representing introns. The dwnd mutation (represented by the blue-gray bar beneath the gene diagram) is a 53 kb intragenic deletion of Mctp1, which includes exons 11-15 and adjoining intronic regions. The red arrows indicate positions and directions of PCR primers designed to distinguish wildtype and dwnd alleles. (B) Alternating pink and purple line segments indicate the exon-encoded regions of the protein-coding portion of the Mctp1 transcript, with corresponding protein domains shown below. The out-of-frame dwnd mutation is expected to create a premature stop codon and truncate the protein after exon 10, thereby eliminating the third C2 domain and the two transmembrane helices of the encoded MCTP1 protein, as shown by the horizontal red arrow. (C) The dwnd mutation in the Mctp1 gene was first identified by RT-PCR. Amplification of cDNA from brain tissue of mutant mice with primers in exons 9 and 17 of the Mctp1 transcript (NM_03174) produced a product that was 503 bp smaller than that of the +/+ control, corresponding to the loss of exons 11-15. (D) A simple PCR assay of genomic DNA was designed to genotype mice with the dwnd mutation. The PCR assay (using three primers positioned as shown in A) amplifies a 450 bp product from the wildtype + allele and a 375 bp product from the mutant dwnd allele. Genotypes can be determined by the presence or absence of these PCR products as shown in the gel image for +/+, +/dwnd, and dwnd/dwnd individuals. (E) The site of the Mctp1^dwnd deletion at the 76.8 Mb position on mouse Chromosome 13 is 1.4 Mb from the Nr2f1 gene at the 78.2 Mb position. The Mctp1 (21 exons) and Nr2f1 (3 exons) genes are shaded in gray, and a blue horizontal bar beneath the Mctp1 gene designates the position and extent of the dwnd deletion. Genome coordinates are shown above. Panels A, B, and E were derived from the mouse Ensembl Genome Browser (GRCm38).

Figure 3.. Nr2f1 transcript and protein downregulation in Mctp1^dwnd cochlea.

(A) RT-qPCR on E16.5 cochleae pooled by genotype. Mctp1 expression is reduced by about 75% in Mctp^dwnd/dwnd mutants, and Nr2f1 expression is reduced by about 50%. Data points represent samples pooled by genotype for each of two litters collected (see Methods). Numbers inside the bars indicate the average ΔCT = Ct_target - Ct_Gapdh, and reveal that Nr2f1 is much more abundantly expressed at E16.5 than is Mctp1. (B-C) E16.5 immunodetection of NRF1 (red) and SOX2 (green) proteins in cochlear cryosections in the basal turn (B) and in half-retina at the optic nerve (C). In the cochlear floor (B), NR2F1 protein is enriched at much lower levels in Mctp1^dwnd/dwnd compared to heterozygote controls, whereas SOX2 is comparable. Arrowheads point to abneural cells in the sensory domain that interestingly appear to retain normal NR2F1 protein in homozygotes. In the retina (C), NR2F1 is by contrast enriched in a comparable manner in Mctp1^dwnd/dwnd and controls. (D) Quantification of the immunostaining intensity ratio between colocalized NR2F1 and SOX2 proteins in the E16.5 cochlea and retina. Using SOX2 as an internal control, NR2F1 is significantly downregulated in the cochlea but not in the retina. Each point represents one section (10 cochlea or 6 retina sections per genotype), two animals represented per genotype; average +/− SD; unpaired t-test with Welch’s correction, ns, not significant; **** p<0.0001). Scale bars 50 μm.

Figure 4.. Hearing thresholds and inner ear paintfills of Mctp1^dwnd/dmid, Mctp1^−/−, Nr2f1^−/− and Mctp1^+/dwnd Nr2f1^+/− mice.

(A) Mctp1^dwnd/dwnd mice and Mctp1^+/dwnd Nr2f1^+/− transheterozygotes are hearing impaired. Mean (± SEM) ABR thresholds of 11 Mctp1^dwnd/dwnd mice and 10 Mctp1^+/dwnd Nr2f1^+/− transheterozygotes and 12 Mctp1^−/− mice tested at one month of age were compared with those of 12 age-matched Mctp1^+/dwnd heterozygous control mice for 8, 16, and 32 kHz test stimuli Thresholds of Mctp1^dwnd/dwnd and Mctp1^+/dwnd Nr2f1^+/− mice are not statistically significantly different from each other but are about 25-50 dB higher than those of Mctp1^+/dwnd control mice. In contrast, thresholds of Mctp1^−/− mice are not statistically significantly different than those of control mice. Nr2f1^−/− mice are not included in the hearing test comparisons because they die after birth. (B) Inner ear paintfills of E15.5 embryos clearly show the fusion of the utricle (u) and saccule (s) in inner ears of Mctp1^dwnd/dwnd, Nr2f1^−/−, and Mctp1^+/− Nr2f1^+/− mice (u/s) but not in the Mctp1^−/− mice. A smaller and shortened cochlear duct (cd) can also be seen in inner ears of the three mutant strains with u/s fusions. Scale bar, 0.5 mm.

Figure 5.. Hair cell patterning in inner ears of Mctp1^dwnd/dwnd, Mctp1^+/dwnd Nr2f1^+/−, and Mctp1^−/− mice.

All panels show phalloidin-stained flat mounts. (A) P0 organ of Corti for the genotypes and positions indicated. Mctp1^+/dwnd Nr2f1^+/𢈒 transheterozygotes show hair cell disorganization at the base (top) and extra OHC rows at the apex (bottom, arrowheads), as seen in the Mctp1^dwnd/dwnd mutants (see Fig. 1B-E). In contrast, hair cell organization in targeted Mctp1 mutants (Mctp1^−/−) looks normal. (B) P0 utricle with the sensory region (macula) outlined. Transheterozygotes show a dysmorphic macula, unlike Mctp1^−/− mutants. (C) P0 cochlear base showing failed separation of the sensory epithelia of the cochlea and the saccular macula in transheterozygotes. Arrowheads indicate the basal most position of the organ of Corti. Mctp1^−/− have a normal saccule (not shown). Mctp1^+/dwnd is used as control. Scale bars, 10 μm (A), 100 μm (B) and 20 μm (C).

Figure 6.. Supplementary information accessed from public database browsers.

(A) Mctp1^dwnd and Nr2f1 are in the same TAD. The Mctp1^dwnd deletion and the Nr2f1 gene (shown by horizontal red bars above the gene diagrams enclosed by red rectangles) are located in the same topologically associating domain (TAD), supporting the possibility of long-range chromatin interactions. The Hi-C heatmap was obtained from the 3D Genome Browser: a web-based browser for visualizing 3D genome organization and long-range chromatin interactions (http://promoter.bx.psu.edu/hi-c/). Although the heatmap shown in this figure was generated from Hi-C analysis of mouse B-lymphoblast cells (Rao et al., 2014), TADs are relatively constant between different cell types and are highly conserved across species, and so should be similar in the mouse embryonic inner ear. (B) Conserved DNA sequences suggest possible enhancer elements in the Mctp1^dwnd deleted region. Multiple phylogenetically conserved DNA sequences are present in the 53 kb Mctp1^dwnd deletion (Chr13: 76,815,703-76,868,709; GRCm38). The ECR Browser (https://ecrbrowser.dcode.org/) was used to compare DNA sequence similarities of mouse with other vertebrate species (human, opossum, chicken, frog, and fish). In addition to the sequences corresponding to exons 11-15 (indicated by downward pointing arrows), five non-coding sequences are highly conserved in all of the vertebrates except fish (indicated by upward pointing arrows).