Ethics statements

All mouse lines were maintained and expanded in the Specific Pathogenic Free System of the Institute of Genome Engineered Animal Models for Human Diseases at Dalian Medical University. All the mouse utilization and procedures in this study were approved by the Institutional Animal Care and Use Committee at Dalian Medical University with the protocol number AEE17038. All the research performance is accorded with the NIH Guide for the Care and Use of Laboratory Animals.

Constructing conditional Fam20C D446N knock-in allele

To construct the mouse Fam20cD446N knock-in allele, the targeting fragment spanning from exons 7 to 10 of mouse Fam20c was subcloned between two loxP sequences in targeting vector. An Frt flanked PKG-Neo cassette was inserted between the exon 10 and the 3’ loxP. The other fragment containing the exon 7–10 sequence fused with an IRES-eGFP element was ligated downstream to the 3’ loxP. The 5’ homologous arm of the targeting vector was 5.4Kb long and subcloned into the vector upstream of the 5’ loxP, while the 3’ homologous arm was 4.1Kb long and subcloned into the vector downstream of the IRES-eGFP element. Finally, a negative selection MC1-HSC-TK cassette was subcloned into the downstream of the 3’ homologous arm in the targeting vector. The final targeting construct was linearized with NotI and electroporated into W4 ES cells. After G418 selection, the genomic DNA was extracted from the ES clones and a 30 PCR screen was performed to screen the targeted clones. The 5’ screen was performed with Fam20c-ScrF2 as the forward primers (5’-TTTCTGTCCTAGGTAAGGGTGA AG-3’) and Lox-ScR2 as backward primer (5’-TGAAGTTCCTATTCTATAACTTCG − 3’) to expand the 5.4Kb fragment. The 3’ screen was performed with the forward primer (eGFP-ScF3:5’-AACGGC CACAATTCAGCGTGTCC-3’) and backward primer (Fam20c-ScrR2: 5’-ACAGCT TCTTGAATTGGGATAAAG-3’) to produce the 4.1 Kb fragment. The 5’ and 3’ screening PCR products were sequenced to confirm the presence of 5’ and 3’ loxP sites, the exons 7–10 containing D446N point mutation and IRES-eGFP element. After the neo cassette was removed from the targeted ES clones by transient transfection of pCAGGS-flpEpuro vector (Addgene, Watertown, MA, USA), two correctly targeted ES cell clones (Clone 5 A and 10 C) were selected and injected into the blastocysts from C57BL/6 mice in the Gene targeting and Transgenic Facility, University of Connecticut Health Center, Farmington, CT, USA. The male chimeras were crossbred with C57BL/6 females to produce F1 agouti offspring. The floxed alleles of F1 agouti mice (Fam20cD446N-flox/+) were genotyped by PCR analyses. With forward (LGF: 5’-CTATCTAAGATGTAGAAGAGGTG-3’) and backward primers (FrtR: 5’-CTCACAGTAATCAGGGTCGAC-3’), the WT allele produced 432 bp segment, and the Fam20cD446N-flox allele gave a 527 bp product.

Generating conditional and conventional Fam20c D446N knock-in mice

To generate the conditional Fam20c knock-in mice, the Fam20cD446N-flox/+ mice were first crossbred with 3.6 Kb Col1a1-Cre transgenic mice to gave rise to 3.6 Kb Col1a1-Cre; Fam20cD446N-flox/+ mice, which were mated with the Fam20cD446N-flox/+ mice to get the 3.6 Kb Col1a1-Cre; Fam20cD446N-flox/D446N-flox mice (referred as the conditional Fam20cD446N knock-in mice). When the Fam20cD446N-flox/+ male mice were crossbred with the Hprt-cre female mice (Stock NO. 004302, Jackson Laboratory), the loxP flanked sequence in the conditional Fam20cD446N-flox allele would be eliminated by Cre in the fertilized eggs or blastocysts. Thus, the conditional Fam20cD446N-flox allele would be transformed into the conventional Fam20cD446N-KI allele. We mated the Fam20cD446N-KI males with the Fam20cD446N-KI females to get Fam20cD446N-KI/D446N-KI mice (referred as the conventional Fam20cD446N knock-in mice). The gentyping PCT with the primers of LGF and FrtR produced a 374 bp fragment for the conventional Fam20cD446N knock-in allele, which distinguished from the 527 bp fragment of the conditional Fam20cD446N knock-in allele.

Plain X-ray and micro-CT examination

The X-ray radiography was performed by the inspector of Faxitron MX-20DC12 (Faxitron Bioptics, Tucson, Arizona, USA). The micro-CT scanning was conducted by the SCANCO Medical µCT system with a medium resolution of 7.0 mm slice increment, and the scanning data were analyzed and reconstructed by NRecon v1.6 and CTAn v1.13.8.1 software. The male WT and conditional or conventional Fam20cD446N knock-in mice from a litter were euthanized by carbon dioxide (followed by cervical dislocation), and then, dissected to get ribs, femurs and tibia (including the knee) for X-ray radiography. For the conventional Fam20cD446N knock-in mice die before weaning, we compared them to their littermates with same gender.

Histology and immunohistochemistry

The male P4W WT and conditional or conventional Fam20cD446N knock-in mice from a litter mice were euthanized and dissected for the femurs. The femurs were fixed in 4% PFA overnight, separated from soft tissues, and decalcified in 15% trypsin- ethylenediaminetetraacetic acid (EDTA) for 2 weeks. The decalcified femurs were dehydrated in gradient alcohols and embedded in paraffin for the 10 μm thick slices. Masson staining for the slices was described previously. Immunohistochemical staining was performed as previously described with the antibodies against Fam20c, Fgf23 and Dmp1 [7, 17, 18]. ABC kit and the DAB kit (Vector Laboratories, Burlingame, CA) were applied to color development with the methyl green counter-staining according to the manufacturer’s instructions. The negative controls of immunohistochemistry were conducted by above procedures without primary antibodies.

Serum biochemistry

The sera of male P4W WT and conditional or conventional Fam20cD446N knock-in mice were collected as previous described [14]. Fgf23 concentration were measured with a full-length FGF23 ELISA kit (Kainos Laboratories, Tokyo, Japan), phosphorus concentration with a kit using the phosphomolybdate ascorbic acid method (Stanbio Laboratory, Boerne, TX, USA), and calcium concentration with a colorimetric calcium kit (Stanbio Laboratory, Boerne, TX, USA). All the concentrations of Fgf23, phosphorus and calcium were quantified by a SPECTRA max 250 microplate spectrophotometer (Molecular Device Corporation, Sunnyvale, California, USA).

Statistical analysis

Statistical analysis were performed using a Student t test with SPSS 21.0 software, and stated as the Mean ± SD (standard derivation). P value less than 0.05 was regarded as the significant difference between two individual groups. To eliminate the discrepancy between genders, only the the body weights of the male WT and conditional or conventional Fam20cD446N knock-in mice were collected for statistical assay.