Novel human mutation and CRISPR/Cas genome-edited mice reveal the importance of C-terminal domain of MSX1 in tooth and palate development

Several mutations, located mainly in the MSX1 homeodomain, have been identified in non-syndromic tooth agenesis predominantly affecting premolars and third molars. We identified a novel frameshift mutation of the highly conserved C-terminal domain of MSX1, known as Msx homology domain 6 (MH6), in a Japanese family with non-syndromic tooth agenesis. To investigate the importance of MH6 in tooth development, Msx1 was targeted in mice with CRISPR/Cas system. Although heterozygous MH6 disruption did not alter craniofacial development, homozygous mice exhibited agenesis of lower incisors with or without cleft palate at E16.5. In addition, agenesis of the upper third molars and the lower second and third molars were observed in 4-week-old mutant mice. Although the upper second molars were present, they were abnormally small. These results suggest that the C-terminal domain of MSX1 is important for tooth and palate development, and demonstrate that that CRISPR/Cas system can be used as a tool to assess causality of human disorders in vivo and to study the importance of conserved domains in genes.

Mutation analyses. The proband was screened for mutations of genes previously related to tooth agenesis (i.e., PAX9, AXIN2, EDA, WNT10A, etc.) by targeted exome sequencing (TES), with next generation technology. A heterozygous nucleotide substitution in exon 4 of WNT10A (c.874 A> G, p.S292G) was detected, but not observed in family members by Sanger sequencing. This substitution had a 0.4% frequency in 1,208 sampled Japanese individuals [Human Genetic Variation Database (HGVD), http://www.genome.med.kyoto-u.ac.jp/ SnpDB/], suggesting that it was an unrelated de novo mutation.
No other single nucleotide variations, small insertions and/or deletions (indels), or gross gains or losses were observed around other candidate genes. However, a segment of exon 2 in MSX1 was covered by only a few TES reads. Sanger sequencing of this region revealed a heterozygous guanine deletion in exon 2 of MSX1 (NM_002448.3: c.844delG) in the proband (II:2) and her affected father and brother (I:1, II:1), but not in her unaffected mother (Fig. 1d, data not shown). This deletion was predicted to cause a frameshift (p.A282Rfs*21) that alters 21 C-terminal amino acids starting from residue 282, creating a premature stop codon that results in a 301-, rather than 303-, residue truncated protein lacking MH6 domain (Fig. 1e). The sequence around MH6 is highly conserved in mammals (Fig. 1e).
Generation of MH6-disrupted mice. Most MSX1 variants isolated from patients with tooth agenesis to date involved single amino acid substitutions in the highly conserved homeodomain/MH4 sequence; few frameshift, nonsense, and splice-site mutations that lead to premature termination have been examined. To verify genotype-phenotype correlation in vivo, MH6 was targeted using CRISPR/Cas system in mice. A gRNA targeting upstream region of MH6 was designed and synthesized in vitro (Fig. 2a), and then co-injected with Cas9 mRNA into one-cell zygotes. Among the mosaic mice produced (17 of 20 F 0 ), three mutant alleles were selected for phenotype analysis: in-frame 21-nucleotide deletion (Msx1 −21 ) and frameshift 28-nucleotide deletion (Msx1 −28 ) causing premature stop codon upstream of MH6; and single-nucleotide insertion (Msx1 +1 ) causing a frameshift mutation affecting MH6 (Fig. 2b,c). The selected mice were mated to wild-type BDF1 mice.

Secondary cleft palate and tooth agenesis in MH6 deficient mice. Because conventional
Msx1-deficient mice is neonatally lethal 25 , phenotypes were first analyzed at E16.5. Coronal sections of Msx WT/−21 and Msx1 −21/−21 embryos showed normal tooth and palate development ( Fig. 3a-f, data not shown). Although Msx WT/−28 mice showed normal palate and tooth development, two phenotypes were observed in Msx1 −28/−28 mice: secondary cleft palate with agenesis of lower incisors; and agenesis of the lower incisors with a thin palate ( Fig. 3g-o). Because one-third of the Msx1 −28/−28 pups (5/15) exhibited palate development, phenotypes of 4-week-old mice were analyzed. All newborn homozygotes with cleft palates died neonatally, whereas those with a thin palate were viable. Micro-CT at 4 weeks old revealed no dentition anomalies in mice with in-frame homozygous mutations (Fig. 4e-h), but in homozygous Msx1 −28/−28 mice, it revealed agenesis of the upper third molar and lower second and third molars as well as undersized upper second molars, in addition to the missing lower incisors observed at E16.5 ( Fig. 4m-p). Similar phenotypes were observed in mice with compound heterozygous mutations affecting MH6 (Figs 3p-u and 4q-t).

Discussion
In this study, we have reported a novel frameshift mutation in exon 2 of MSX1 in a family with autosomal-dominant inheritance of non-syndromic oligodontia. The proband and her father had tooth agenesis phenotypes that were similar to those observed with previously identified MSX1 mutations 26,27 . The guanine deletion identified (p.A282Rfs*21) is upstream of the MH6 domain and produced a frameshift that replaces the highly conserved Scientific RepoRts | 6:38398 | DOI: 10.1038/srep38398 MH6 sequence with an unrelated peptide. A heterozygous mutation disrupting the original stop codon and opening the reading frame was reported previously in a familial case of non-syndromic tooth agenesis 28 , but the presently identified mutation is the first to affect the entire MH6 domain.
Heterozygous MH6 disruption in model mice did not alter craniofacial development, whereas homozygous disruption resulted in agenesis of lower incisors with or without cleft palate at E16.5 as well as agenesis of the upper third molars and the lower second and third molars in 4-week-old mice. It is noteworthy that although the MSX1 mutation in the proband family showed autosomal dominant inheritance, tooth agenesis and cleft palate were observed only in the bialellic disruption of MH6. Our findings are in line with previous studies reporting several nonsense oligodontia-causing mutations located in MH4 or its upstream regions with autosomal dominant trait 29,30 , although failure in tooth and palate development was inherited in a autosomal recessive manner in conventional Msx1-deficient mice with disrupted MH4 domain 25 .
The MH6 is the most C-terminal domain of MSX1, known as a PIAS binding domain 31 . It is necessary for localization of the MSX1 protein in the nuclear periphery, which enables repression of MyoD in myoblasts 31 . In addition, it has been reported that Msx1 and Msx2 act as potent transcriptional activators of the promoter of the Heat shock 70 kDa protein 1B gene HSPA1B through their C-terminal domains 32 . Although the MH6 domain has not been reported to interact with genes involved in craniofacial development, our results suggest that it is important in MSX1 function during tooth and palate development.
Most of the MSX1 mutations that have been identified previously as causing non-syndromic tooth agenesis are located in MH4 domain, or in upstream regions that affect MH4 1,27,33 . The MH4 domain is involved in DNA binding and protein-protein interactions 34,35 . Functions of MH4-containing proteins are not mediated solely through protein-DNA interactions. It has been demonstrated that transcriptional repression by MSX1 occurs in the absence of DNA-binding sites; the repressor function is attributed to multiple domains in the N-and C-terminal regions of MSX1 36 . In silico analysis studies have indicated that missense mutations affecting MH4 in MSX1 alter the encoded 3D structures of the proteins, specifically by destabilizing the helix-turn-helix motif or altering its protein-DNA interactions 37,38 . Moreover in vitro studies have demonstrated that nuclear localization of MSX1 protein is mediated by the homeodomain 37 .
Finnerty et al. proposed that the non-random distribution of mutations in MH domains may be related to the yet unexplained genotype-phenotype correlation 12 . Specifically, they suggested that mutations that disrupt the MH1C or MH6 domain is associated with cleft disorders and may act through a dominant negative mechanism. On the other hand, tooth agenesis phenotypes attributed to MH1N and MH4 domain mutations have been explained by functional redundancy from the MSX2 domains, which are also highly conserved. Our findings contradict these observations, because the MH6 domain-affected members showed non-syndromic tooth agenesis phenotype.
This study has some limitations. First, we did not analyze possible off-target effects of the CRISPR/Cas9 modification. However, it is important to note that the genotype-phenotype correlations were maintained across mouse generations. Moreover, the phenotypes in MH6-disrupted mice were consistently observed in Msx1 −28/−28 and Msx −1/+28 . Second, although mouse models have been extensively used to understand tooth morphogenesis and to establish disease models, there are differences from humans in terms of number of the teeth. Mice have monophyodont dentition, presenting a single continuously growing incisor separated from three molars in each quadrant. In contrast, humans have dyphyodont dentitions with all four types of teeth. Therefore, further studies in non-human primates, such as marmosets in which premolars are observed, should be conducted. In conclusion, we identified a novel mutation in a familial case of non-syndromic tooth agenesis affecting the MH6 domain of MSX1. Genotype-phenotype correlation was corroborated through CRISPR/Cas system-mediated gene targeting in mice. Thus, the present findings demonstrate that MH6 is functionally required in MSX1 for tooth and palate development.

Subjects.
A Japanese family with tooth agenesis was enrolled. The pedigree was made based on clinical examination of the proband and interviews with all available family members. The diagnosis of oligodontia in the proband was verified by panoramic radiographs. Written informed consent was obtained from all participants. This study complied with the tenets of the Declaration of Helsinki and was approved by the ethics committee of Tokushima University Hospital (H24-8, H26-40).
TES and Sanger sequencing. Genomic DNA was extracted from saliva samples with the Oragene ® DNA collection kit (OG-500, DNA Gentotec Inc., Ottawa, Canada), according to the manufacturer's instructions. TES was performed with TruSight One sequencing panel (Illunima, San Diego, CA) and MiSeq benchtop sequencer (Illumina). Alignment of sequencing reads to a human reference genome (hg19), duplicate read removal, local realignment around indels, base quality score recalibration, variant calling, and detailed annotation were performed as described elsewhere 41,42 . Copy-number variations were detected relative to NGS data as described elsewhere 41,42 . Direct Sanger sequencing using polymerase-chain reaction products and BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) were conducted with ABI 3500xL Genetic Analyzer (Applied Biosystems). Sequencing results were compared by BLAST (http://blast.ddbj.nig.ac.jp/top-j. html). Putative functional consequences of mutations were predicted in silico with MutationTaster.

Animals. All animal experiments were approved by the Ethical Committee of Tokushima University for
Animal Research (Approval number: 12064, T27-16). All procedures were conducted in accordance with the Guidelines for Animal Experiments of Tokushima University.
Production of sgRNA and Cas9 mRNA. Two oligonucleotides (Forward, TAGGCGCGCTGGAAAG GGCCAG; Reverse, AAACCTGGCCCTTTCCAGCGCG) including the target sequence in Msx1 were annealed and cloned into the BsaI site of the pDR274 plasmid (Addgene, Cambridge, MA). The Cas9 plasmid (pMLM3613, Addgene) and target sequence-inserted pDR274 were digested with PmeI and DraI, respectively. Linearized templates were transcribed with mMessage mMachine T7 ULTRA kits (Ambion, Austin, TX), and then treated with DNase I, according to the manufacturer's instructions. Cas9-encoding mRNA and sgRNA were suspended in appropriate volumes of RNase-free water, after purification by phenol-chloroform-isoamylalcohol extraction and isopropanol precipitation. RNA microinjection into embryos. Cas9-encoding mRNA (10 ng/μ l) and sgRNA (1 ng/μ l) quantified by a NanoDrop 2000 UV-Vis spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE) were co-injected into the cytoplasm of fertilized eggs, obtained after mating BDF1 (C57BL/6 × DBA2 F 1 ) male with superovulated female mice. Injected eggs were cultured overnight in M16 medium (Sigma, St. Louis, MO) at 37 °C in 5% CO 2 . Two-cell embryos were then transferred into the oviducts of pseudopregnant MCH(ICR) mice.

Mouse genotyping.
To verify CRISPR/Cas9-mediated mutations, genomic DNA was extracted from tail biopsies. The genomic regions flanking the gRNA target were PCR amplified with KOD-Plus-Neo (Toyobo, Osaka, Japan) and an Msx1-specific primer pair (Forward, 5′ -CGCAAGCACAAGACTCTCTTT-3′ ; and Reverse 5′ -AGGGGTCAGATGAGGAAGGT-3′ ), according to the manufacturer′ s instructions. PCR products were purified for direct or cloned sequencing using Wizard SV Gel and PCR Clean-up System (Promega). For mosaic F 0 genotyping, purified PCR amplicons were cloned into plasmids using DynaExpress TA PCR Cloning Kits (BioDynamics Laboratory, Tokyo, Japan). Isolated plasmids from each sample were sequenced with a BigDye Terminator Sequence Kit ver. 3.1 and an ABI 3500xL Genetic Analyzer (Applied Bioystems). After targeting verification, F 0 mice were mated with wild-type BDF1 mice to propagate alleles of interest. F 2 to F 4 generations were analyzed.
Maxillae and mandibles of 4-week-old mice were resected and fixed overnight in 70% ethanol after removal of soft tissues. Molars and incisors were analyzed by high-resolution micro-CT (Skyscan 1176, operated at 50 kV and 200 μ A; Bruker-microCT, Kontich, Belgium). Two-dimensional images were used to generate three-dimensional (3D) renderings with CTVox 3D Creator software (version 3.0, Bruker). The resolution of the micro-CT images was 9 μ m per pixel.