Functional Analysis of the Single Nucleotide Polymorphism (787T>C) in the Tissue-Nonspecific Alkaline Phosphatase Gene Associated With BMD


  • The authors have no conflict of interest.


Polymorphisms of the TNSALP gene have not previously been studied as a possible determinant for variations in BMD or as a predisposing genetic factor for osteoporosis. This study showed a significantly higher association between the 787T>C (Tyr246His) TNSALP gene and BMD among 501 postmenopausal women. Furthermore, the effects of amino acid substitution on the catalytic property of the protein translated from the 787T>C gene were examined.

Introduction: Alkaline phosphatase (ALP) is present mainly on the cell membrane in various tissues and hydrolyzes a variety of monophosphate esters into inorganic phosphoric acid and alcohol. Human ALPs are classified into four types: tissue-nonspecific, intestinal, placental, and germ cell types. Based on studies of hypophosphatasia, which is a systemic skeletal disorder resulting from a tissue-nonspecific ALP (TNSALP) deficiency, TNSALP was suggested to be indispensable for bone mineralization.

Materials and Methods: We explored the possibility that the TNSALP gene may contribute to age-related bone loss in humans by examining the association between TNSALP gene polymorphisms and BMD in 501 Japanese postmenopausal women. To analyze the protein translated from the TNSALP gene associated with BMD, we constructed a TNSALP cDNA expression plasmid.

Results: We genotyped two single nucleotide polymorphisms (787T>C[Tyr246His] and 876A>G[Pro275Pro]), which proved to be in complete linkage disequilibrium. There was a significant difference in BMD and the BMD score adjusted for age and body weight (Z score) among haplotypes (p = 0.041), which was lowest among 787T/876A homozygotes, highest among 787T>C/876A>G homozygotes, and intermediate among heterozygotes. In subgroups divided by age, haplotypes were significantly associated with BMD in older postmenopausal women (>74 years; p = 0.001), but not in younger postmenopausal women (<74 years; p = 0.964). Expression of the 787T>C TNSALP gene using COS-1 cells showed that the protein translated from 787T>C had ALP-specific activity similar to that of 787T. Interestingly, the Km value for TNSALP in cells transfected with the 787T>C TNSALP gene was decreased significantly compared with that of cells bearing the 787T gene, reflecting the higher affinity.

Conclusions: These results suggest that variation in TNSALP may be an important determinant of age-related bone loss in humans and that the phosphate metabolism pathway may provide a novel target for the prevention and treatment of osteoporosis.


ALKALINE PHOSPHATASE (ALP; EC is present mainly on the cell membrane in a variety of tissues and hydrolyzes a variety of monophosphate esters into inorganic phosphoric acid and alcohol at a high optimal pH (pH 8–10). Human ALPs are classified into four types: tissue-nonspecific, intestinal, placental, and germ cell types.(1–3) Tissue-nonspecific ALP (TNSALP) is found in the bone, liver, kidney, and other tissues. The TNSALP gene is located on chromosome 1 and consists of 12 exons and 11 introns, with the coding sequence beginning in the second exon.(1) Concerning the message for TNSALP, a noncoding first exon was identified in the liver message (liver type),(4) which differed from that of the previously known osteoblast-derived cDNA sequence (bone type).(5) TNSALP is expressed at high levels in cell mineralizing tissues such as osteoblasts and odontoblasts. Previously, we identified TNSALP in human periodontal ligament cells, translated from mRNA that assesses the first exon of the ALP mRNA, as in osteoblastic cells.(6) The physiological role of ALP is unknown, but strong evidence is provided by the rare genetic disease hypophosphatasia (HOPS). HOPS is an inherited disorder characterized by a defect in skeletal mineralization caused by TNSALP deficiency. We have analyzed various mutations in the TNSALP gene(7–17) and have shown that deletion of T at a position (1559delT) located in exon 12 causes the loss of ALP activity and results in the synthesis of an abnormal TNSALP molecule, which is not expressed on the plasma membrane.(9) Clinical symptoms vary greatly among the four different subforms of the disease (perinatal, infantile, childhood, and adult), which are distinguished from each other primarily by the age at which the skeletal lesions develop. Patients with premature loss of deciduous teeth, but without bone involvement, are regarded as having odontohypophosphatasia.(18) The analysis of HOPS mutations will help elucidate the molecular and cellular functions of TNSALP.

TNSALP expression is modulated by growth conditions and various factors, including BMP, glucocorticoid, 1,25(OH)2 vitamin D3, and PTH. BMP is a bone-inductive protein and can initiate a process that begins with cartilage formation and ends in de novo bone formation. Human BMP stimulated the TNSALP of mouse osteoblast-like cells in vitro.(19–21) The level of TNSALP (bone-type) activity in serum is very high in childhood for the promotion of bone formation. In type I osteoporosis, which occurs within 20 years after menopause and is associated with excessive cancellous bone loss, the level of TNSALP activity is also elevated and is used as a marker of bone metabolism.

Osteoporosis is characterized by low BMD and increasing risk for fractures in the elderly.(22) BMD seems to be determined by genetic as well as environmental factors. If genetic markers can indicate the risk for osteoporosis, they would be useful for its prevention and early effective treatment. Several genes have been implicated as genetic determinants of osteoporosis.(23–30) However, any conclusive results have been obtained from the association studies of candidate genes and the panel of genes examined should be expanded further. TNSALP may play an important role in active bone metabolism by hydrolyzing phospho-compounds and supplying free inorganic phosphate; however, the polymorphisms of the TNSALP gene have not been studied as a possible determinant for variation in BMD or as a genetic predisposing factor for osteoporosis. This is the first report on the age-related association of TNSALP gene polymorphisms.


Subjects and measurement of radial BMD

DNA samples were obtained from peripheral blood of 501 postmenopausal women living in southern Japan. The mean age was 73.6 ± 5.8 (SD) years (range, 65-93 years). All subjects were unrelated volunteers and provided informed consent before the study. No participant had medical complications or was undergoing treatment for conditions known to affect bone metabolism, such as pituitary diseases, hyperthyroidism, primary hyperparathyroidism, renal failure, adrenal diseases, or rheumatic diseases, and none were receiving estrogen replacement therapy. The BMD of the radial bone (expressed in g/cm2) of each participant was measured by DXA using a DTX-200 (Osteometer Meditech, Hawthorne CA, USA). The BMD was normalized for differences in age, height, and weight, using a multiple regression analysis program. Z scores and T scores were calculated using installed software (DTX-200) based on Japanese women.

Detection of polymorphisms by DHPLC

To detect polymorphisms in the TNSALP gene, DNA samples (n = 20) were selected randomly from the 501 postmenopausal women described above. PCR products were obtained using 13 sets of primers, which covered all exons (from exon 1 to exon 12). Information on the primers and PCR conditions is available on request.(7) The detection of single nucleotide polymorphisms (SNPs) in PCR products was performed by denaturing high-performance liquid chromatography (DHPLC) on a WAVE DNA Fragment Analysis System (Transgenomic, Omaha, NE, USA). PCR products were introduced into the mobile phase at an injection volume of 5 μl by an auto sampler. A DNA Sep analytical column (Transgenomic) was used as the solid phase, and the products were eluted from the column with a binary gradient of 0.1 M triethylamine acetate buffer (TEAA)/25% acetonitrile as the mobile phase at a flow rate of 0.9 ml/min. Temperatures for analyses were automatically predicted from the target sequence by WAVE Maker 4.0 software (Transgenomic). The SNPs detected in PCR products were confirmed by sequencing using the thermo sequence Cy 5.5 dye terminator cycle sequencing kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) with a Gene Rapid sequencer (Amersham Pharmacia Biotech).

Genotyping for molecular variants in the TNSALP gene

All 501 participants were genotyped for the two polymorphisms (rs3200254 and rs3200255) archived in the dbSNP at ( The first of these, TNSALP 787T>C, corresponded to Tyr246His (Y246H), but the second, TNSALP 876A>G, was not accompanied by amino acid substitutions. The amino acid sequence was numbered from the N terminus of the mature protein (i.e., Met at the translation initiation site was −17).

Genotyping was performed for both SNPs by TaqMan allelic discrimination assay using the following primers and probes (the polymorphic base in each probe is in italic): 787T>C-forward primer, 5′-GACGGCCTGGACCTCGTT-3′; reverse primer, 5′-CCACCTCTGCAGCCACATG-3′; FAM-labeled probe, 5′-ACCGAGACACAAGGT3′; VIC-labeled probe, 5′-CAAACCGAGATACAAGG-3′; 876A>G-forward primer, 5′-CACCACCATCTCGGAGAGTGA-3′; reverse primer, 5′-TCACAGCCTCTCAGCATCCA-3′; VIC-labeled probe, 5′-CTGCATGTCCCCTGG-3′; FAM-labeled probe, 5′-CTGCATGTCCCCCGG-3′. Primer and probe concentrations were optimized according to the manufacturer's recommendations, so that each reaction contained 100 nM of FAM-labeled and VIC-labeled probe, respectively. The Taqman reactions were thermocycled as follows: 95°C for 10 minutes and then 40 cycles at 95°C for 15 s, followed by 60°C for 1 minute for 787T>C, and 95°C for 10 minutes and 40 cycles at 95°C for 15 s, followed by 58°C for 1 minute for 876A>G. The completed reactions were analyzed on an ABI Prism 7200 sequence detection platform.

Construction and analysis of the mutant TNSALP cDNA expression plasmid

Normal TNSALP cDNA was obtained from human periodontal ligament cells(6) and was inserted downstream of the cytomegalovirus gene promoter (pCMV) of the expression plasmid vector pcDNA3 (Invitrogen, Carlsbad, CA, USA) as described previously.(9) Site-directed mutagenesis was performed with a Unique Site Elimination (USE) Mutagenesis Kit (Amersham Pharmacia Biotech). Normal TNSALP cDNA inserted into pcDNA3 (wildtype, 787T) was templated by mutagenesis. An oligonucleotide of the mutagenesis primer 5′-TCAAACCGAGACACAAGCACT-3′ (the italic nucleotide is the target mutagenic nucleotide) was used for the mutation of 787T>C. The resultant mutant 787T>C was confirmed by sequencing using the thermo sequence Cy 5.5 dye terminator cycle sequencing kit (Amersham Pharmacia Biotech) with a Gene Rapid sequencer (Amersham Pharmacia Biotech).

Transfection of the mutant plasmid

To examine the effects of 787T>C on the properties of TNSALP, COS-1 cells were transfected with circular plasmid DNA (5 μg/35-mm-diameter dish) by a lipofection technique (Gibco BRL, Grand Island, NY, USA). After a 48- or 72-h transfection, cells were collected and homogenized using a Polytron homogenizer (Kinematica) with 10 mM Tris-buffered saline (TBS; pH 7.4) containing 1% Triton X-100 and 1 mM phenylmethylsufonyl fluoride (PMSF). After centrifugation at 10,000g for 5 minutes, the supernatant was assayed.

Enzyme activity and protein assays

The ALP activity was determined with 10 mM p-nitrophenylphosphate (p-NPP) as a substrate in 100 mM 2-amino-2-methyl-1,3-propanediol-HCl buffer containing 5 mM MgCl2, pH 10.0, at 37°C. The enzyme activity was determined by the rate of hydrolysis of p-NPP and expressed in units of U = μmol p-nitrophenol formed per minute. The enzyme activity was also assayed in the presence of an inhibitor: LEV or L-phe. To study thermostability, we pretreated the enzyme preparation at 56°C and allowed it to react with the substrate. Protein concentrations were determined using bicinchoninic acid (BCA) protein assay reagent (Pierce, Rockford, IL, USA). For the kinetic studies, measurements of Km were performed using p-NPP in 100 mM 2-amino-2-methyl-1,3-propanediol-HCl buffer (pH 10.0) containing 5 mM MgCl2. Michaelis-Menten constants (Km) were determined graphically from Lineweaver-Burk Plots.

Enzyme histochemistry

COS-1 cells expressing TNSALP mutants on coverslips were stained for ALP activity.(31) Cells were fixed with 4% (wt/vol) paraformaldehyde in PBS (pH 7.2) for 10 minutes on ice and washed three times with 0.1 M TBS. The cells were incubated with a mixture of 0.1 mg/ml of naphthol AS-MX phosphate and 0.4 M Tris-HCl buffer containing 5 mM MgCl2 as a substrate in 20 ml of 0.4 M Tris-HCl buffer (pH 8.8) at room temperature for 5 minutes. The cells were counterstained with Fast red violet salt.


Polyacrylamide gel (7.5%) electrophoresis in the presence of SDS was performed according to the method of Weber and Osborn.(32) The osteosarcoma cell line Saos-2 (Riken, RCB0428) was used as a control. After electrophoresis, ALP isozymes separated in the gel were stained by the coupling method of β-naphthyl-phosphoric acid monosodium salt with Fast violet B salt.(31)

Western blotting

Saos-2 and COS-1 cells expressing the wildtype TNSALP (787T) or TNSALP (787T>C) were homogenized in 10 mM TBS (pH 7.4) containing 1 mM PMSF. Samples of the cell homogenates were subjected to SDS-PAGE (7.5%) according to the method of Weber and Osborn.(32) After electrophoresis, proteins were transferred to Immobilon P (Millipore, Bedford, MA, USA). The membrane was incubated first with 400-fold diluted antibody against TNSALP(33) and then with 1000-fold diluted anti-rabbit IgG conjugated with horseradish peroxidase (Sigma). Proteins were detected with ECL Western blotting detection reagents (Amersham Biosciences).

Statistical analysis of linkage disequilibrium and tests for association with BMD

Linkage disequilibrium was studied for all two-way comparisons of the polymorphisms according to Thompson's method (D, D′, and r2).(34,35) The coefficient of disequilibrium, D, is the difference between the observed haplotype frequency and the frequency expected under statistical independence = pAB × pab − pAb × paB. Dividing D by its maximum possible value gives the normalized disequilibrium coefficient (D′ = D/Dmax); Dmax = min(pA × pB, pa × pb) if D < 0, and Dmax = min(pA × pb, pa × pB) if D > 0, r2 = D2/(pA × pB × pa × pb). Levels of significance were determined by χ2 statistics for the corresponding 2 × 2 table. BMD, Z score, height, weight, and body mass index (BMI) were compared between genotypic (haplotypic) categories using ANOVA. Statistical significance regression was determined by ANOVA and the posthoc test. The significance of enzyme activity and parameters of the COS-1 cells was determined by Mann-Whitney U-test. The level of statistical significance was set at p < 0.05. All statistical analyses were performed using the statistical package SPSS 10.07J (SPSS, Chicago, IL, USA).

3D structure of human TNSALP

The modeled structures of human TNSALP were constructed using the program Discovery STUDIO, Modeling 1.1 (Accelrys, San Diego, CA, USA) based on human placental ALP (PALP).(36) The amino acid substitutions on the 3D modeled structures of PLAP and TNSALPs (787T, 787T>C) were studied.


Detection of polymorphisms in the TNSALP gene in a Japanese population

In total, eight SNPs detected in the TNSALP gene (sequence accession ID, NM 000478) of the Japanese population (n = 20) are summarized in Table 1. We found the polymorphism 534C>T first in this study, and it was not accompanied by amino acid substitutions. 862+20 (G>T), 862+51 (G>A), 862+58 (C>T), and 1309+46 (G>T) have been identified in a Japanese population, and 330T>C has been reported in a German population by Orimo et al.(15,37)

Table Table 1.. SNPs Detected in the TNSALP Gene (nm 00478) of a Japanese Population
original image

Among these eight polymorphisms, only the T>C at position 787 resulted in an amino acid substitution. In addition, 787 (T>C) and 876 (A>G) have been detected as frequent polymorphisms in a North American population by Henthorn et al.(38) They reported that 787T>C and 876A>G were found together in one allele and that these two polymorphisms are in linkage disequilibrium.(38) Therefore, we selected these two polymorphisms (787T>C and 876A>G) for further analysis.

Identification and association of 787T/876G with low BMD in a Japanese population

We genotyped 501 postmenopausal women and revealed that the two SNPs selected were closely associated with BMD: the first of these, 787T>C, corresponded to Tyr246His (Y246H), and the other, 876A>G, was not accompanied by amino acid substitutions (Pro275Pro). An obvious consistency led us to analyze linkage disequilibrium for possible two-way comparisons of the variations using Thompson's method (D, D′, and r2). The results confirmed complete linkage disequilibrium of 787T>C with 876A>G (D′ = 1.0000, r2 = 1.0000, χ2 = 1002.00, p < 0.001). We observed haplotypes and identified only two distinct haplotypes among the test subjects. The haplotype 787T/876A constituted 50.3%, and the other haplotype, 787T>C/876A>G, represented 49.7% in our test population. No significant difference in height, weight, or BMI was seen among the genotypes of the SNPs (787T and 787T>C) in either subpopulation (data not shown). These results indicate that these SNPs are not associated with age, height, or weight in this population.

As shown in Fig. 1, ANOVA and posthoc test in these subjects revealed a significant correlation between the TNSALP haplotypes and BMD, indicating a co-dominant effect; that is, BMD was lowest among 787T/787T (homozygotes, TT-type; 0.308 ± 0.003 g/cm2), highest among 787T>C/787T>C (homozygotes, CC-type; 0.327 ± 0.006 g/cm2), and intermediate among heterozygotes (TC-type; 0.310 ± 0.004 g/cm2).

Figure FIG. 1..

Association of two common SNPs (787T and 787T>C) with BMD in Japanese postmenopausal women (n = 501). Data are expressed as means (bars) ± SE (error bars). The number above the bars is the p value of the difference in the mean BMD (ANOVA).

There was also a significant difference in Z score and T score among the haplotypes; the lowest was in 787T/787T (TT-type) and the highest in 787T>C/787T>C (CC-type; Fig. 2; Table 2). Aging is known to be one of the major factors for bone loss in women, so we divided the population into two subgroups according to their age: the <74 years group and the >74 years group. The dividing point, 74 years, was near to the average age of the population. In the younger group (<74 years), there was no significant difference in BMD (Fig. 1), Z score (Fig. 2), and T score (Table 2). In the older group (>74 years), there were significant differences; that is, BMD, Z score, and T score were lowest among 787T/787T, highest among 787T>C/787T>C, and intermediate among heterozygotes. These results indicated that the effect of the haplotypes on BMD depended on the age.

Table Table 2.. T Scores of Haplotype of 787T and 787T>C in Japanese Postmenopausal Women
original image
Figure FIG. 2..

Association of two common SNPs (787T and 787T>C) with Z score in Japanese postmenopausal women (n = 499). Data are expressed as means (bars) ± SE (error bars). The number above the bars is the p value of the difference in the mean Z score (ANOVA).

Expression of the investigated SNPs (787T or 787T>C) in TNSALP cDNA

To examine the effects of amino acid substitutions on the catalytic properties of COS-1 cells transiently expressing SNPs of the human TNSALP (787T or 787T>C) gene, the cells were homogenized and assayed for ALP activity. As shown in Fig. 3, there was no significant difference in the levels of ALP-specific activity between 787T and 787T>C. When co-transfected with the same quantity of 787T and 787T>C plasmids (5:5), the cells exhibited a level of ALP-specific activity almost the same as that of 787T or 787T>C, and no interaction in ALP activity between 787T and 787T>C was recognized. In addition, there was no significant difference between 787T and 787T>C in the activity remaining after the treatment with inhibitors (levamisole and L-phenylalanine) or heat inactivation (56°C, 10 minutes; Table 3).

Table Table 3.. Biochemical Properties of 787T and 787T>C TNSALP Proteins Transiently Expressed in COS-1 Cells
original image
Figure FIG. 3..

COS-1 cells were transfected with pcDNA3-tnsalps (787T, 787T>C, 1559delT) separately or in combination. Mock indicates cells transfected with pcDNA3 plasmid. The transfected cells were homogenized and assayed for ALP activity. Data are expressed as means ± SE (n = 5∼7). n indicates the number of experiments. Each experiment was performed in triplicate. We previously reported that a deletion of T at position 1559 (1559delT) in the TNSALP gene caused a frame shift downstream from codon 503 and that cells transfected with the 1559delT mutant plasmid had undetectable ALP activity.(9) Thus, we compared the expression of ALP activity of the 1559delT mutant as a negative control.

Enzyme histochemistry

Cytohistochemical staining confirmed the presence of strong ALP activity at the surface of cells transfected with 787T, 787T+787T>C (5:5) or 787T>C (Fig. 4). No significant differences among these cells were observed.

Figure FIG. 4..

COS-1 cells transfected with pcDNA3-tnsalps (787T, 787T:787T>C = 1:1, 787T>C) were stained for ALP for 10 minutes at room temperature.

SDS-PAGE analysis

The further identification of enzymatic properties of TNSALP isozymes was performed using SDS-PAGE (Figs. 5A and 5B). In both groups (lane 1, 787T; lane 2, 787T>C), we detected a main band of ALP activity, and the molecular weight was 110 kDa as calculated from the positions of the standard proteins (Fig. 5A, lanes 1 and 2). This molecular weight was identical to that of the Saos-2 cell lysate (Fig. 5A, lane 3). We also obtained the same results from the immunohistochemical staining of COS-1 cells transfected with 787T or 787T>C for TNSALP using a specific purified antibody against TNSALP (PCA-SWKD; Fig. 5B, lanes 1 and 2).

Figure FIG. 5..

SDS-PAGE analysis of TNSALPs. Expression of the 787C or 787T>C TNSALP gene using COS-1 cells. COS-1 cells were cultured for 72 h after transfection and were homogenized in 10 mM TBS (pH 7.4) containing 1 mM PMSF. Each cell lysate was subjected to 7.5% SDS-PAGE according to the method of Weber and Osborn.(32) Std, molecular-size standard (Precision Plus Protein Standards; Bio-Rad); lane 1, COS-1 cells were transfected with pcDNA3-tnsalp 787T; lane 2, COS-1 cells were transfected with pcDNA3-tnsalp 787T>C; lane 3, Saos2 cells were not transfected with DNA. (A) ALP activity in the gel was detected by staining with the coupling method of β-naphthyl-phosphoric acid monosodium salt with Fast violet B salt.(31) (B) After electrophoresis, proteins were transferred to an Immobilon P. The membrane was incubated first with 400-fold diluted antibody against TNSALP (PCA-SWKD)(33) and then with 1000-fold diluted anti-rabbit IgG conjugated with horseradish peroxidase. Proteins were detected with ECL Western blotting detection reagents.

Kinetic study and the location of the SNPs (787T or 787T>C) in the modeled 3D structure of human TNSALP

As shown in Table 4, the Km (Michaelis constant) value for the protein translated from the TNSALP (787T>C) gene was significantly higher than that from 787T (p < 0.01). The substitution of Tyr-246 at this position for His-246 produced a 1.7-fold reduction in Km. To compare the substitutions of the amino acid (Tyr-246 or His-246), which may have led to the difference in the Km values, we constructed a homodimeric model of TNSALP based on the crystal structure of human PALP,(36) as shown in Fig. 6. The position of Tyr-246 or His-246 in TNSALP corresponds to that of Arg-241 in PALP,(36) and we compared the effect of the substitutions of these amino acids.

Table Table 4.. Values of Km of 787T and 787T>C TNSALP Proteins Transiently Expressed in COS-1 Cells Using p-NPP as Substrate
original image
Figure FIG. 6..

Location of the investigated SNPs (787T and 787T>C) in the 3D modeled structures of human TNSALPs. (A) Overall dimeric structure of TNSALP. Monomer II is shown in blue. (B) Comparison of the amino acid substitutions investigated in the 3D modeled structure of human PALP and TNSALPs (787T, 787T>C), shown in different colors: PALP (Arg-241), green; TNSALP787T (Tyr-246), yellow; TNSALP 787T>C (His-246), purple.

The overall structure of human TNSALP is that of a dimer; each monomer contains 507 amino acids and the residues involved in the active site of the enzyme (Ser-93, Asp-92, Arg-167), and the ligands coordinating the two zinc atoms and the magnesium ion are conserved.(5,36) The model of the 3D structure of Tyr-246 and His-246 in TNSALP based on PALP revealed that both amino acids are positioned at the surface of the dimeric TNSALP and are located near the metal binding domains (Fig. 6A). Arg-241 in PALP faces the inside of the dimeric PALP structure, whereas Tyr-246 and His-246 face the outside of the dimeric TNSALP (Fig. 6B).


This study investigated the contribution of SNPs at the human TNSALP locus to BMD in Japanese women. As shown in Figs. 1 and 2, SNPs were associated with BMD, especially in the older population (>74 years). A significantly higher association of BMD with these variants was detected; the Z score was lowest in 787T/876A homozygotes (TT-type), intermediate among heterozygotes (TC-type), and highest among 787C/876G homozygotes (CC-type). The data implied that variation in the coding region of the TNSALP gene might have affected bone metabolism in postmenopausal Japanese women, eventually introducing variation in BMD. The decreased BMD in postmenopausal women could be the result of accelerated bone loss and/or lesser acquisition of bone mass before maturation. This result provides the first evidence that the TNSALP gene may be involved in the etiology of osteoporosis. The finding that the association of the TNSALP gene with this disorder is dependent on age cannot be explained by differences in the distribution of allele types among generations, because the allelic frequency was similar in all subpopulations. As shown in Figs. 1 and 2 and Table 2, in the younger subgroup (>74 years), there were no significant differences in BMD, Z score, or T score. These results suggested that the haplotype 787T>C/876A>G (CC-type) could be associated with prevention of bone loss, depending on age.

Various mutations in the TNSALP gene discovered with HOPS have been analyzed.(9,10,12,14,16,17,37,39) The deletion of T at nucleotide 1559 (1559delT) seems to be a mutational hot spot in Japanese HOPS patients.(9,15,40) In 1559delT cases, a frame shift is observed downstream from codon L503, and we reported that the 1559delT mutant protein in COS-1 cells exhibited no enzymatic activity; its molecular size was larger than that of the wildtype protein.(9) HOPS is highly variable in its clinical expression, which ranges from stillbirth, with no radiographically apparent mineralized bone, to pathological fractures that occur only during late adulthood. In general, infantile-type HOPS is considered a disease with autosomal recessive inheritance. However, the inheritance pattern of milder forms of HOPS is less clear, and both autosomal dominant and autosomal recessive patterns have been proposed. We previously suggested the presence of dominant mutations in milder cases (odonto-type and adult-type HOPS patients).(8,13) In the recessive model, it is assumed that the enzymatic activity of cells co-transfected with equal quantities of wildtype and mutant plasmids is 50% or more, depending on the residual activity of the mutant, and is linearly correlated with the portion of wildtype cDNA.(41) The enzymatic activity of COS-1 cells co-transfected with equal quantities of 787T and 787T>C was similar to that of 787T, and a correlation between 787T and 787T>C was not observed (Fig. 3).

The polymorphic 787T>C base change causes an amino acid substitution of histidine for tyrosine at position 246 in TNSALP. Henthorn et al.(38) reported that the base changes of 787T>C and 876A>G, found together in one allele from each of four HOPS fibroblast lines, seem to be polymorphic in both normal and HOPS populations. They suggested that the presence of both tyrosine and histidine residues at this position in mammalian TNSALP supports the conclusion that the 787T>C base change is a functionally silent polymorphism. Moreover, they indicated that the frequencies of 787T>C in the normal and affected populations was 0.33 (24/73) and 0.43 (21/49), respectively, whereas 876A>G was found at frequencies of 0.31 (22/71) and 0.47 (23/49), respectively. They reported that these frequencies do not differ significantly between the normal and HOPS populations.(38) In this report, we obtained additional valuable data suggesting that this variation in SNPs in the TNSALP gene may be an important determinant for older postmenopausal osteoporosis, and the effects of this genetic variation on age-related bone loss may accumulate through aging.

TNSALPs in human bone, liver, kidney, and dental tissues show similar enzymatic properties, such as susceptibility to various inhibitors and thermostability, and are clearly different from human intestinal and placental type ALPs.(6,33) Previously, we reported that the human intestinal and placental type ALPs showed heat stability, but TNSALPs lost >50% of their activity within 10 minutes at 56°C.(6,33) Levamisole (Lev) is an effective inhibitor of TNSALP, but it is not effective against intestinal ALP and PALP.(6,33) The biochemical properties of these TNSALP preparations, such as the inhibitory effects of Lev and L-phenylalanine (L-Phe) and the heat stability (56°C, 10 minutes), were identical between 787T and 787T>C, and typical of bone-type ALP (Table 3).

By SDS-PAGE analysis, COS-1 cells transfected with the cDNA plasmid (787T or 787T>C) were found to express high levels of ALP activity (Fig. 5A, lane 1 or 2). Previously, we detected normal nascent protein translated from the wildtype (787T) gene using COS-1 cells and calculated the molecular size to be ∼57.2 kDa under conditions of dissociation of TNSALP into a primarily monomeric form.(9) In this study, the apparent molecular weight of TNSALP under nonreducing conditions from cells transfected with the cDNA plasmid (787T or 787T>C) was estimated to be about 110 kDa, which is consistent with the dimeric form of native mammalian TNSALPs. TNSALP in Saos-2 cells derived from osteosarcoma expressed a high level of ALP activity and showed the same mobility as 787T or 787T>C (Fig. 5A, lane 3). Their protein bands had a similar mobility in the Western blotting analysis, as shown in Fig. 5B (lanes 1-3). These results suggested that the ALP isoenzyme from cells transfected with the cDNA plasmid of 787T or 787T>C was in a native dimeric form, as in other mammalian TNSALPs.

In humans, TNSALP is about 50% identical with the other three isozymes (intestinal, placental, and germ cell). The other isozymes are tissue-specific, and their genes are 90∼98% homologous and clustered on chromosome 2. The core structures are largely conserved and have the same metal ions and glycosylation site in all mammalian ALPs. The residue His-246 in TNSALP corresponds to Arg-241 in PALP. Arg-241 is unique to PALP and the residue corresponding to it is histidine in both human intestinal ALP (His-241) and germ cell ALP (His-241).(36) Although the ancestral allele of TNSALP is not available, we consider His-246 (787T>C) to possibly be the wildtype residue in the human TNSALP gene given these results. These three residues, arginine (Arg), histidine (His), and tyrosine (Tyr), are all hydrophilic. Arg-241 and His-246 are basic amino acids, whereas Tyr-246 is a neutral amino acid residue. The differences among these residues may reflect the enzyme's behavior in affinity for the substrate.

Interestingly, expression of the 787T>C TNSALP gene using COS-1 cells showed that the protein translated from 787T>C (His-246) had a lower Km value than 787T (Tyr-246), as shown in Table 4. The Km value indicates the concentration of the substrate at the 1/2 Vmax (maximum velocity), and the kinetic affinity for the substrate. The kinetic affinity may affect the mediation of the specificity and modulation of activity and may contribute to regulatory effects on bone metabolism.

As a result of studies on cDNA encoding ALP isozymes, it is known that the primary structure in the catalytic region is well conserved in the ALPs of humans, animals, and E. coli, suggesting that TNSALP plays important roles in active metabolism by hydrolyzing phospho-compounds, supplying free inorganic phosphate (Pi). In E. coli, a limitation of extracellular phosphate induces expression of the ALP gene, indicating a role for extracellular Pi in ALP gene regulation. Previously, we reported that Pi starvation increased TNSALP activity and regulated its expression in the mouse stromal cell line ST2, derived from mouse bone marrow.(20,42) However, little is known about how the hydrolysis of phospho-compounds works to increase the absorption of phosphate. Kawano et al.(28) reported that klotho gene polymorphisms are associated with BMD of older postmenopausal women. Mice lacking a functional klotho gene exhibit multiple aging phenotypes including osteopenia, and the klotho gene is closely related to Pi homeostasis and dietary Pi intake.(43)

Because elevated extracellular concentrations of inorganic pyrophosphate (PPi), phosphoethanolamine (PEA), and pyridoxal-5′-phosphate (PLP) were observed in HOPS, these phospho-compounds seem to be the physiological substrates of TNSALP.(18) In mineralizing tissues, the formation of the hydroxyapatite crystal begins in matrix vesicles (MVs). TNSALP in MVs may play an important role in the mineralizing process and it may regulate extracellular PPi concentrations by hydrolyzing PPi, which is an inhibitor of the formation of hydroxyapatite crystal.(44) The production of PPi is controlled by the nucleoside triphosphate pyrophosphohydrolase (NTPPPH) isozymes, such as plasma cell membrane glycoprotein-1 (PC-1). Recently, Hessle et al.(45) showed that bone mineralization in double-KO mice lacking both TNSALP and PC-1 is essentially normal, providing evidence that TNSALP and PC-1 are key regulators of the extracellular PPi concentrations required for bone mineralization. As aging associates with elevated levels of cartilage extracellular PPi and PPi-generating NTPPPH,(46) the differences of enzyme affinity between 787T and 787T>C in TNSALP genes may cause the variation in the regulation of mineralization in the elderly.

In conclusion, we analyzed the novel candidate osteoporosis-susceptibility gene TNSALP and found the substitution of amino acids to be significantly associated with the BMD and Z score among postmenopausal Japanese women. Further study on bone metabolism including the phosphate metabolism pathway may provide novel targets for the prevention and treatment of osteoporosis.