25-Hydroxyvitamin D3 1α-Hydroxylase and Vitamin D Synthesis

See allHide authors and affiliations

Science  19 Sep 1997:
Vol. 277, Issue 5333, pp. 1827-1830
DOI: 10.1126/science.277.5333.1827


Renal 25–hydroxyvitamin D3 1α-hydroxylase [1α(OH)ase] catalyzes metabolic activation of 25–hydroxyvitamin D3 into 1α,25–dihydroxyvitamin D3[1α,25(OH)2D3], an active form of vitamin D, and is inhibited by 1α,25(OH)2D3. 1α(OH)ase, which was cloned from the kidney of mice lacking the vitamin D receptor (VDR / mice), is a member of the P450 family of enzymes (P450VD1 α). Expression of 1α(OH)ase was suppressed by 1α,25(OH)2D3 in VDR+/+ and VDR+/ mice but not in VDR / mice. These results indicate that the negative feedback regulation of active vitamin D synthesis is mediated by 1α(OH)ase through liganded VDR.

Vitamin D is metabolized by sequential hydroxylations in the liver and kidney to a family of seco-steroids. The two most biologically active forms of vitamin D are 1α,25(OH)2D3 and 24R,25–dihydroxyvitamin D3 [24R,25(OH)2D3] (1,2). The binding of 1α,25(OH)2D3 to the nuclear receptor for the hormonally active form of vitamin D (VDR) activates the VDR (3), with subsequent regulation of physiological events such as calcium homeostasis and cellular differentiation and proliferation (4). Hydroxylation of 25–hydroxyvitamin D3 [25(OH)D3] is mediated by 25(OH)D3 1α-hydroxylase [1α(OH)ase] in the proximal tubule of the kidney. 1α(OH)ase is inhibited by its end product, 1α,25(OH)2D3 (5), and activated by calciotropic peptide hormones such as calcitonin and parathyroid hormone (6, 7). Thus, serum concentrations of 1α,25- (OH)2D are kept constant. Vitamin D–dependent rickets type I (8) may be caused by mutations in the 1α(OH)ase gene. Biochemical analysis of semipurified 1α- (OH)ase protein has suggested that 1α(OH)ase belongs to the P450 family of enzymes (9).

We developed a nuclear receptor–mediated expression system to clone the cDNA encoding 1α(OH)ase. This system is based on the fact that a precursor of 1α,25(OH)2D3, 25(OH)D3, can activate the transactivation function of the VDR only in the presence of 1α(OH)ase activity. Mice lacking the VDR (VDR / mice) developed an abnormally high serum concentration of 1α,25(OH)2D at 7 weeks, suggesting excessive 1α(OH)ase activity (Fig.1A) (10). Kidneys from these mice were used to prepare an expression library with the expression vector pcDNA3 in COS-1 cells. These cells were then transfected with another vector expressing a chimeric protein that includes the VDR ligand-binding domain [VDR(DEF)] fused to the yeast GAL4 DNA-binding domain [GAL4-VDR(DEF)] (11) and with a reporter plasmid bearing lacZ regulated by the GAL4 binding site (17M2-G–lacZ). Vectors expressing adrenodoxin (ADX) and adrenodoxin reductase (ADR) were also included to support efficient hydroxylation (12). When supplied with 25- (OH)D3, cells expressing 1α(OH)ase would produce ligands that are able to activate GAL4-VDR(DEF) and would be identifiable by expression of β-galactosidase (β-Gal) (Fig. 1Bc) (13). Transfected plasmids were extracted from eight β-Gal–positive cells and subjected to polymerase chain reaction (PCR) for amplification of the inserted cDNAs. PCR products of 2.0 to 2.5 kbp (the expected size for P450 family gene transcripts) (14) were recovered and subcloned. Sequence analysis of the isolated cDNAs from 64 random clones revealed that 13 clones encoded an identical, complete open reading frame (ORF). Reintroduction of this single cDNA sequence rendered the cells positively stained (Fig. 1Bd).

Figure 1

Isolation of 1α(OH)ase cDNA from the kidney of VDR−/− mice by an expression cloning method (28). (A) Serum concentrations of 1α, 25(OH)2D in VDR+/+, VDR+/−, and VDR−/− mice at 3 and 7 weeks (3w and 7w, respectively). (B) Twelve hours after transfection, 10−8 M 25(OH)D3 was added to the media. After 36 hours, cells were fixed with 0.05% glutaraldehyde and incubated with X-Gal for 4 hours at 37°C (13). (a) Nontransfected cells. (b) Effect of active ligand 1α,25(OH)2D3 on cells with the expression system but lacking the kidney cDNA library. (c) Detection of 1α(OH)ase-expressing cells transfected with the kidney cDNA library. Positively stained cells were harvested by micromanipulation (29) and analyzed by PCR. The PCR products were run on a 1% agarose gel, and fragments of about 2.0 to 2.5 kbp [the predicted size for full-length 1α(OH)ase cDNA] were purified and subcloned into pcDNA3. (d) Cells transfected with the cDNA encoding 1α(OH)ase.

Using this cDNA as a probe, we obtained the full-length cDNA by colony hybridization screening of the same library. The amino acid sequence derived from the ORF predicts a protein with 507 amino acids (Fig. 2A). The in vitro–translated protein is 55 kD (Fig. 2B), which is similar in size to semipurified 1α(OH)ase (9). This protein (hereafter designated P450VD1 α) has a mitochondrial target signal and is homologous to members of the P450 family (14), particularly to rat vitamin D3 25–hydroxylase (41.7%) and mouse 25(OH)D324–hydroxylase [24(OH)ase] (31.6%) (15, 16). The putative sterol-binding domain (93% and 60% for rat and mouse, respectively) (17) and the heme-binding domain (70% and 80%, respectively) (18) show the greatest sequence similarities.

Figure 2

Predicted amino acid sequence of P450VD1α. (A) Amino acid sequence. The putative mitochondrial target signal is boxed. The putative sterol-binding domain is underlined. The heme-binding domain is indicated by a dashed line. GenBank–European Molecular Biology Laboratory–DNA Data Bank of Japan accession number, AB006034(30). (B) In vitro–translated P450VD1α protein. The P450VD1α protein that had been in vitro–translated in the presence of35S-labeled methionine with the reticulocyte lysate system (Promega, Madison, Wisconsin) was analyzed by a 10% SDS–polyacrylamide gel electrophoresis (31).

To confirm that P450VD1 α can convert 25(OH)D3 into 1α,25(OH)2D3, which in turn activates the VDR, we transfected COS-1 cells with GAL4-VDR(DEF), chloramphenicol acetyltransferase (CAT) reporter plasmid (17M2-G–CAT) (19), ADX and ADR expression vectors (12), and theP450VD1 α expression vector. Activation by 25(OH)D3 was observed only in the presence of P450VD1 α , ADX, and ADR (Fig. 3A). These results indicate that P450VD1 α is 1α(OH)ase that converts 25(OH)D3 into 1α,25(OH)2D3.

Figure 3

Conversion of 25(OH)D3 into an active vitamin D3 serving a VDR ligand by P450VD1α (A) and identification of the converted 25(OH)D3 by HPLC analysis (B). (A) COS-1 cells were cotransfected with 0.5 μg of GAL4-VDR(DEF), 1 μg of 17M2-G–CAT, 0.5 μg each of the ADX and ADR expression vectors, and 1 μg (+) or 3 μg (++) of theP450VD1α expression vector with or without the indicated ligands (19). One representative CAT assay (lower panel) and relative CAT activities (upper panel) corresponding to means ± SEM for three independent experiments are shown. (B) Normal- and reversed-phase HPLC analysis of 25(OH)D3metabolite converted by P450VD1α. 3H-Labeled 25(OH)D3 (105 dpm; 6.66 terabecquerels/mmol) (Amersham International) was incubated with COS-1 cells transfected with (b and e) or without (c andf) the P450VD1α expression vector together with ADX and ADR expression vectors for 6 hours at 37°C. The cultured media were extracted with chloroform-methanol and analyzed on normal-phase HPLC (a to c) with TSK gel silica 150 column (4.6 mm by 250 mm) (Tosoh) by means of established solvent systems (9, 32). Eluent fractions were collected, and radioactivity was estimated by liquid scintillation counting (22). Authentic vitamin D derivatives [1α(OH)D3, 25(OH)D3, 24R,25(OH)2D3, 1α,25(OH)2D3, and 1α,24,25(OH)3D3] were chromatographed, and the retention times of these vitamin D derivatives were determined by ultraviolet absorption at 264 nm [normal phase (a) and reverse phase (d)]. Reversed-phase HPLC (d to f) was run to confirm the presence of3H-labeled 1α,25(OH)2D3 with the use of a Cosmosil 5C18-AR packed column (4.6 mm by 150 mm) (Nacalai Tesque) (32). A 264, absorbance.

To chemically confirm the enzymatic product of P450VD1 α, we used normal and reversed phases of high-performance liquid chromatography (HPLC) (20). When3H-labeled 25(OH)D3 was added to cells transfected with theP450VD1 α expression vector, a metabolite was detected in the incubated medium. The retention times of the metabolite matched those of authentic 1α,25(OH)2D3 in both the normal and reversed phases of HPLC (Fig. 3B), even when different HPLC systems were used on both phases (21, 22).

In both normal and VDR / mice, the 2.4-kbpP450VD1 α transcript was detected only in the kidney at 7 weeks (Fig.4A). We did not detect 1α(OH)ase transcript in the other tissues, although 1α(OH)ase activity has been reported in extrarenal tissues such as placenta and macrophages (1, 22, 23).P450VD1 α was overexpressed in the VDR / mice, by about 2.5-fold at 3 weeks and 50-fold at 7 weeks (Fig. 4, B and C). Administration of 1α,25(OH)2D3 repressed 1α(OH)ase gene expression in the VDR+/+ and VDR+/ mice, but not in the VDR / mice, at 3 and 7 weeks. The unusually high concentrations of serum 1α,25(OH)2D3 in the VDR / mice at 7 weeks (Fig. 1A) (10) are thus probably due to the overexpression of 1α(OH)ase. Thus, it is likely that liganded VDR inhibits 1α(OH)ase gene expression. In the VDR / mice, expression of 24(OH)ase, which converts 25(OH)D3 into 24R,25(OH)2D3, was reduced to an undetectable amount, and the normal response (24) to 1α,25(OH)2D3 was not observed (Fig. 4, B and C). Thus, our results indicate that liganded VDR also regulates expression of 24(OH)ase, normally activating it (25).

Figure 4

Liganded VDR is involved in the negative regulation of the 1α(OH)ase gene and the positive regulation of the 24(OH)ase gene by 1α,25(OH)2D3. (A) Kidney-specific expression of theP450VD1α gene. We extracted and analyzed poly(A)+ RNA from various vitamin D target tissues, including kidney and liver, of the wild-type (+/+) and VDR-knockout (−/−) mice at 7 weeks by Northern (RNA) blot analysis using theP450VD1α and the β-actin cDNA as probes (26). No transcript was detected in the other tissues, such as placenta (21). (B and C) Lack of response to 1α,25(OH)2D3 in the expression of the 1α(OH)ase and 24(OH)ase genes in the VDR / mice. Northern blot analysis of the P450VD1α and 24(OH)ase (P450cc24) gene expressions was performed in the VDR+/+, VDR+/−, and VDR−/− mice at 3 and 7 weeks with (+) or without (−) an excess dose of 1α,25(OH)2D3 (50 ng per mouse) (33). A representative Northern blot analysis is shown in (B), and the relative abundance of the hydroxylase gene transcripts normalized with the β-actin transcript from more than five mice for one group is calculated (C). ND, not detected.

  • * To whom correspondence should be addressed. E-mail: uskato{at}


View Abstract

Stay Connected to Science

Navigate This Article