The patient was a 5-year-old girl with premature loss of the deciduous teeth and severe mobility of the remaining teeth. Medical history reported an absence of previous dental treatment, recurrent infection, systemic disease, trauma or surgery. No history of HP or consanguineous marriage was identified in the family of the patient. X-ray examination, bone mineral densitometry, pathological examinations and laboratory tests were conducted. In addition, bidirectional sequencing of the ALPL gene was performed. All the examinations were performed with permission from the patient and parents in the form of full informed written consent.

Blood samples (2 ml) were collected from the patient and parents. Gene sequencing was conducted using an ABI PRISM^® 3730 DNA Analyzer (Shanghai South Gene Technology, Co., Ltd., Shanghai, China). Sites of variation were identified as described previously (10). The ALPL sequence derived from GenBank (accession no. NM_000478.4) was referred to. Altered nucleotides were confirmed by bidirectional sequencing. The novelty of the two variants were determined from the National Center for Biotechnology Information (NCBI) human SNP database (dbSNP, https://www.ncbi.nlm.nih.gov/snp/), the 1000 Genomes Project database (https://www.ncbi.nlm.nih.gov/variation/tools/1000genomes/), and the Exome Sequencing Project (ESP, http//evs.gs.washington.edu/EVS/) (10).

The ALPL sequences of other species were obtained from GenBank, and conservation analysis was performed with Clustalw2 software (ebi.ac.uk/Tools/msa/clustalw2/). The SIFT (sift.jcvi.org/) and PolyPhen-2 (genetics.bwh.harvard.edu/pph2/) algorithms were used to predict whether the missense mutations may affect the protein function.

The plasmid construction and site-directed mutagenesis was performed by GenScript (Nanjing, China). Briefly, a full-length cDNA for wild-type human ALPL (NM_000478.3) was synthesized and confirmed by DNA sequencing using ABI PRISM 3700 DNA Sequencer (Perkin-Elmer Applied Biosystems, USA). Amplification was performed using polymerase chain reaction (PCR) with primers ALPL forward (F) 5′-AGCTCATGCATAACATCAGG-3′ and reverse (R) 5′-GTGGCAACTCTATCTTTGGT-3′. PCR was conducted using Phusion high fidelity DNA polymerase (New England Biolabs, Inc., Ipswich, MA, USA) according to the manufacturer's protocol.

Site directed mutagenesis was performed following the overlap extension PCR method as previously described (11). Mutagenic primer W29R 5′-CCTTAGTGCCAGAGAAAGAGAAAGACCCCAAGTACCGGCGAGACCAAGCGCAAGAG-3′ was designed containing the c.85T>C mutation. Primers for I395V F 5′-CGTGGCAACTCTGTCTTTGGTCTGGCCCCCATGCTGAGTGACACAGAC-3′ and R 5′-AGCATGGGGGCCAGACCAAAGACAGAGTTGCCACGGGGGGTGTATCCAC-3′ were used to generate the c.1183A>G mutation. With amplified the ALPL gene as templates, the fragments containing overlapping regions were amplified through the first two PCR with mutagenic primers (W29R, ALPL-mF, ALPL-mR for the c.85T>C mutation; I395V-F, I395V-R, ALPL-mF, ALPL-mR for c.1183A>G mutation). Following purification using a DNA purification kit (Beyotime Institute of Biotechnology, Haimen, China), the products of the two reactions in the first PCR were mixed and subjected to the second PCR using a 5′-end primer ALPL-TBF (5′-GCTAGCGCTACCGGACTCAGATCTCGAGATGATTTCACCATTCTTAGTACTGGCCATTGGCA-3′; the XhoI site is underlined) with a XhoI restriction site and a 3′ end primer ALPL-TBR (5′-TTAGGGGGGGGGGAGGGAGAGGGGCGGTACCTCAGAACAGGACGCTCAGGGGGTAGA-3′; the KpnI site is underlined) with a KpnI restriction site to acquire full-length cDNAs encoding mutated ALPL.

For construction of ALPL expression vectors, we used a CMV-driven plasmid pIRES2-EGFP (5.3 kb; BD Biosciences, San Jose, CA, USA) containing enhanced green fluorescent protein (EGFP) gene. In this vector, the gene of interest and the EGFP gene are translated respectively, which means the ALPL would be expressed as an unfused protein (Fig. 1). Recombinant of the plasmids were constructed as follows, after purification using the DNA purification kit (Beyotime Institute of Biotechnology), double restriction enzymes (XhoI and KpnI) digested the wild type and mutated full-length ALPL gene and the pIRES2-EGFP plasmid. Subsequently, the digested products were isolated using 0.8% agarose gel electrophoresis and purified with a DNA clear up kit (Beyotime Institute of Biotechnology). The digested fragments were inserted into pIRES2-EGFP plasmid according to the manufacturer's protocol of the DNA Ligation kit (Takara Biotechnology Co., Ltd., Dalian, China). Next, the recombinant vectors were transformed into competent Escherichia coli cells. The transformation procedures and colony PCR were performed as previously described (12). The recombinant plasmids were extracted using Plasmid Preparation kit (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. Primers were designed using the Primer Designer^™ Tool (Thermo Fisher Scientific, Inc., Waltham, MA, USA) (Table I). The insertion was verified by sequencing and restriction-enzyme digestion.

The HEK293 human embryonic kidney cell line was purchased from the Cell Bank of Chinese Academy of Sciences (Shanghai, China). These cells were cultured as described previously (13). Cells were seeded into a 6-well plate at a density of 2×10⁵ cells per well for 24 h before transfection. A 0.6 µg aliquot of plasmid of wild type and mutated ALPL were transiently transfected into HEK293 cells. The Attractene Transfection Reagent (Qiagen GmbH, Hilden, Germany) was used according to the manufacturer's protocol. Cells were observed under a fluorescence microscope after 48 h. Each assay was duplicated or triplicated in independent experiments.

At 48 h after transfection, the medium was removed. The cells were lysed with 0.2 ml lysis buffer (Beyotime Institute of Biotechnology). A protease inhibitor mixture without any phosphatase inhibitor was added to the cell lysate as described previously (14). The protease inhibitor mixture, which consisted of phenylmethylsulfonyl fluoride (100 mM), aprotinin (15 µM), leupeptin (100 µM), bestatin (100 µM), pepstatin A (100 µM) and E-64 (80 µM), was purchased from Shanghaibocai (Shanghai, China). The lysate and media were centrifuged at 15,000 × g for 10 min at 4°C to remove insoluble material. ALP activity was determined using the Alkaline Phosphatase Assay kit (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. In this kit, para-nitrophenyl phosphatase is adopted as the substrate of ALP. This reaction generates para-nitrophenol, which is yellow in an alkaline environment. The absorbance at a wavelength of 405 nm is proportional to the concentration of para-nitrophenol. A six-point standard curve is recommended in the range from 0.02–0.2 mM. Sample (1 µl) was added to each system. The absorbance at a wavelength of 405 nm was detected every minute automatically by a SpectraMax M3 spectrophotometer (Molecular Technologies, Sunnyvale, CA, USA). Each ALP assay was repeated at least twice in independent experiments. Protein concentrations were determined using a Bicinchoninic Acid Protein assay kit (Beyotime Institute of Biotechnology, China). ALP activities were calculated relative to protein concentration.

Western blotting was performed to detect the protein expression levels of ALPL in transfected cells. Briefly, the remaining cell lysate was prepared in 5X SDS loading buffer. Equal amounts of protein (20 µg per lane) were separated by 4–12% SDS-PAGE (GenScript, Nanjing, China) and transferred onto a polyvinylidene difluoride membrane. Following blocking in 5% bovine serum albumin (BSA, Beyotime Institute of Biotechnology, Haimen, China) at room temperature for 2 h, the membrane was incubated overnight at 4°C with primary antibody. ALPL antibody purchased from Bioworld Technology, Inc., Nanjing, China (cat. no. BS6134), was diluted to 1:500 in 5% BSA. GAPDH antibody (cat. no. AP0063), which was also purchased from Bioworld Technology, was diluted at 1:10,000 in 5% BSA. The next day, the blots were washed 3 times with TBS and incubated with a horseradish peroxidase-conjugated goat anti-rabbit IgG antibody (cat. no. BS13278; Bioworld Technology, Inc., Nanjing, China) at a dilution of 1:50,000 for 2 h at room temperature. The blots were washed again and subsequently visualized with an ImageQuant LAS 4000 mini system (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA). This experiment was performed three times.

To investigate the effect of the two mutations p.Trp29Arg (c.85T>C) and p.Ile395Val (c.1183A>G) on ALPL dynamics, all-atom MD simulations were performed on wild type and mutated ALPL in monomer form. The Protein Data Bank structure (1EW2) (15) was selected as the initial structure of wild-type ALPL. The mutated structure was built by Modeller version 9.13 software (16). All the MD simulations were carried out using Gromacs package software version 5.0.4 (17) with Amber99sb-ildn force field (18). The TIP3P water model was used for the solvate (19). The monomer was solvated in a 9.893×9.893×9.893 nm water box. The Zn²⁺ and Mg²⁺ in the crystal structure were explicitly included in the simulations but without consideration of the effects of charge transfer and protonation/deprotonation, since the current simulations do not involve the binding/unbinding of Zn²⁺ and Mg²⁺ ions. To achieve charge neutral and an equivalent metal ion concentration of 150 mM/l, 85 Na⁺ and 85 Cl⁻ were added to the water box of wild-type ALPL, and 85 Na⁺ and 86 Cl⁻ to the water box of mutated ALPL. The electrostatic interaction was treated using Particle Mesh Ewald with a cutoff of 10 nm. The same cutoff was used in the calculation of the van der Waals interactions. All bonds were constrained using the LINCS algorithm and the MD time step was set to 2 fsec (20). V-rescale and Parrinello-Rahman algorithms were used for temperature and pressure coupling (21,22). The initial structure of each system was first subjected to a minimization of 50,000 steps. Following minimization, a 500 fsec constant number of atoms, volume and temperature equilibration at 300 K was performed. Subsequently, a 500 fsec number of particles, pressure and temperature (NPT) equilibration at 1 atm and 300 K was conducted, followed by a 100 nsec production simulation under NPT conditions at 1 atm and 300 K.

Data were statistically analyzed using GraphPad Prism version 5.0 (GraphPad Software Inc., La Jolla, CA, USA). Results are presented as the mean ± standard derivation. A one-way analysis of variance and Student-Newman-Keuls test were conducted to determine the significant differences in mean values among groups. P<0.05 was considered to indicate a statistically significant difference.

A physical examination demonstrated that the body weight and height of the patient was just above the norm. She had a waddling gait without limb or skin abnormalities. Extensive premature loss of deciduous teeth was observed, with only 55, 63, 64, 65, 75, 84 and 85 remaining (Fig. 2A). In addition, constricted neck and denudation of dentin on the occlusal surface was observed in several teeth (Fig. 2B). Furthermore, although abnormal mobility was detected, the patient had a good oral hygiene, demonstrating no sign of attachment loss. Panoramic radiographs demonstrated malformed deciduous teeth without congenital loss of permanent teeth. No considerable alveolar bone resorption was observed during the 3-year follow-up (Fig. 3A-C).

Slight dislocation of the hip and disrupted Shenton's lines were observed with mild osteodystrophy in the knee. In addition, a wide epiphyseal plate with rough surface was observed (Fig. 3D and E). Bone mineral densitometry velocity of sound was 3698 m/s, and the Z-value was 1.96 m with a relative risk of fracture of 1.000.

The naturally lost primary tooth exhibited enamel defects, peculiar short roots and widened root canals (Fig. 4A-C). Hematoxylin and eosin staining of the histological slices demonstrated that calcification of the dentin was abnormal. An unusually distinct incremental line of dentin and an uneven width of the predentin were observed (Fig. 4D and E) and patches of cementum were found to be embedded in dentin (Fig. 4F).

The patient (5-years-old) had decreased serum ALP, parathyroid hormone and serum osteocalcin compared with the normal levels at 5-years-old (Table II). Additionally, the level of ALP was also decreased compared with earlier time points: 83 U/l at 3 years old, and 74 U/l at 4 years old.

Direct sequencing of the 12 exons of ALPL demonstrated that the patient was a compound heterozygote with two missense mutations in the ALPL gene: p.Trp29Arg (c.85T>C) and p.Ile395Val (c.1183A>G). p.Trp29Arg (c.85T>C) is located on exon 3 and p.Ile395Val (c.1183A>G) is on exon 10. The p.Trp29Arg (c.85T>C) mutation was also present on exon 3 of the mother's ALPL gene, while the p.Ile395Val (c.1183A>G) mutation was present on exon 10 of the father's ALPL gene (Fig. 5).

SIFT and PolyPhen-2 were used to predict the impact of the substitution on ALPL function. The prediction by SIFT indicated that p.Trp29Arg and p.Ile395Val are ‘tolerable’. The prediction conducted by PolyPhen-2 demonstrated that p.Trp29Arg and p.Ile395Val are ‘possibly damaging’ in terms of conservation, with scores of 0.554 and 0.817, respectively. However, p.Trp29Arg and p.Ile395Val were ‘benign’ in term of variation, with scores of 0.238 and 0.422, respectively.

The correct construction of the plasmids was verified through sequencing and restriction-enzyme digestion (Fig. 6). All six groups transfected with wild type and/or mutated ALPL demonstrated similar ALPL protein expression levels (Fig. 7). Loss of protein function was observed in p.Trp29Arg and p.Ile395Val. When Trp29 of all the monomers were substituted by Arg, only 4.1% ALP activity remained relative to wild type ALP. When half of the monomers were wild type, the ALP activity increased to 33.7% relative to wild type ALP. The corresponding data for p.Ile395Val, wild type + p.Ile395Val and p.Trp29Arg + p.Ile395Val were 19.1, 50.1 and 7.6%, respectively (Fig. 8A and B). Furthermore, the activity of wild type ALP was reduced when co-expressed with p.Trp29Arg and reduced to a lesser degree by p.Ile395Val, suggesting a dominant negative effect of these two mutants on wild type ALPL (Fig. 8C and D).

The simulations revealed that the N-terminal helix of wild type and mutated ALPL exhibit varying behaviors. Several snapshots were extracted at different stages along the simulations of wild type and mutated monomer ALPL, and these were aligned to the crystal structure of monomer ALPL (Fig. 9). After 30 nsec, the N-terminal helix of mutated ALPL separated from the main body of the protein, and moved freely. However, in the wild type ALPL, this part remained close to the rest of monomer at all times. Except for the N-terminal helix, the overall structure of the remaining part of ALPL was minimally affected by the mutation. Taking into consideration that Trp29 is located at the N-terminal helix, it was speculated that the mutation (p.Trp29Arg) may affect the dynamics of ALPL significantly, whereas the substitution (p.Ile395Val) may have little effect on ALPL.