1α,25-Dihydroxyvitamin D3-24-Hydroxylase (CYP24) Hydroxylates the Carbon at the End of the Side Chain (C-26) of the C-24-fluorinated Analog of 1α,25-Dihydroxyvitamin D3 *

The sequential oxidation and cleavage of the side chain of 1α,25-dihydroxyvitamin D3(1α,25(OH)2D3) initiated by the hydroxylation at C-24 is considered to be the major pathway of this hormone in the target cell metabolism. In this study, we examined renal metabolism of a synthetic analog of 1α,25(OH)2D3, 24,24-difluoro-1α,25-dihydroxyvitamin D3(F2-1α,25(OH)2D3), C-24 of which was designed to resist metabolic hydroxylation. When kidney homogenates prepared from 1α,25(OH)2D3-supplemented rats were incubated with F2-1α,25(OH)2D3, it was mainly converted to a more polar metabolite. We isolated and unequivocally identified the metabolite as 24,24-difluoro-1α,25,26-trihydroxyvitamin D3(F2-1α,25,26(OH)3D3) by ultraviolet absorption spectrometry, frit-fast atom bombardment liquid chromatography/mass spectroscopy analysis, and direct comparison with chemically synthesized F2-1α,25,26(OH)3D3. Metabolism of F2-1α,25(OH)2D3into F2-1α,25,26(OH)3D3 by kidney homogenates was induced by the prior administration of 1α,25(OH)2D3 into rats. The C-24 oxidation of 1α,25(OH)2D3 in renal homogenates was inhibited by F2-1α,25(OH)2D3 in a concentration-dependent manner. Moreover, F2-1α,25,26(OH)3D3 was formed in ROS17/2.8 cells transfected with a plasmid expressing 1α,25(OH)2D3-24-hydroxylase (CYP24) but not in the cells transfected with that expressing vitamin D3-25-hydroxylase (CYP27) or containing inverted CYP27 cDNA. These results show that CYP24 catalyzes not only hydroxylation at C-24 and C-23 of 1α,25(OH)2D3 but also at C-26 of F2-1α,25(OH)2D3, indicating that this enzyme has a broader substrate specificity of the hydroxylation sites than previously considered.

D 3 , in its target cells is initiated by side chain hydroxylation at C-23, C-24, and C-26 (1)(2)(3)(4). Of these hydroxylation sites, it is now accepted that the sequential oxidation and cleavage of the side chain initiated by the hydroxylation at C-24 catalyzed by mitochondrial 1␣,25(OH) 2 D 3 -24-hydroxylase (CYP24) is the major pathway by which the hormone is inactivated (5). Because transcription of the CYP24 gene is highly up-regulated by 1␣,25(OH) 2 D 3 in its target cells (6 -8), CYP24 is regarded as the key enzyme for the breakdown of the hormone (9). Metabolism of 1␣,25(OH) 2 D 3 initiated by C-23 hydroxylation is induced by 1␣,25(OH) 2 D 3 itself (10), and recombinant human CYP24 also catalyzes C-23 hydroxylation of 25-hydroxyvitamin D 3 (25(OH)D 3 ) to yield 23S,25-dihydroxyvitamin D 3 (23S,25(OH) 2 D 3 ) (11). Therefore, it is likely that CYP24 initiates both C-24 and C-23 hydroxylation pathways of 1␣,25(OH) 2 D 3 . In contrast, mitochondrial vitamin D-25-hydroxylase (CYP27) catalyzes the hydroxylation at C-25 and C-26 of vitamin D 3 and 1␣-hydroxyvitamin D 3 (1␣(OH)D 3 ) (12), but it is not clear whether or not this enzyme hydroxylates C-26 of 1␣,25(OH) 2 D 3 . Recently, a model for the mechanism of the hydroxylation site selection by CYP24 and CYP27 was proposed. This model postulates that CYP24 directs its hydroxylation site(s) by the distance of C-24 and C-23 from the vitamin D ring structure and that CYP27 does so by the distance between the hydroxylation sites and the end of the side chain (13).

Metabolism of 1␣,25(OH) 2 D 3 by Rat Renal
Homogenates-The renal homogenate prepared as described above was incubated with 0.25 nM 1␣,25(OH) 2 [1␤-3 H]D 3 (specific activity, 3.4 GBq/mmol) for 15 min at 37°C under oxygen gas in the presence or the absence of graded amounts of F 2 -1␣,25(OH) 2 D 3 . The radioactivity in the aqueous phase after extraction by the method of Bligh and Dyer (23) was measured using a scintillation counter. The lipid fraction of the reaction mixture was analyzed by reverse phase HPLC under the conditions described above after adding the chemically synthesized 1␣,25(OH) 2 D 3 and 1␣,24R,25(OH) 3
Mass spectra were recorded on a JEOL AX505 frit-FAB LC/MS spectrometer (Tokyo, Japan) equipped with a JASCO 880-PU HPLC system (Tokyo, Japan) (HPLC column, Develosil ODS-HG-5, 2.0 ϫ 150 mm (Nomura Chemical Co. Ltd., Aichi, Japan)) and operated at an accelerating voltage of 3 kV in a positive ion mode. A portion of the isolated metabolite ( Fig. 1, peak X, about 500 ng) was analyzed by frit-FAB LC/MS using 45% acetonitrile/water containing 0.7% glycerol matrix as the mobile phase. An aliquot of the same metabolite (500 ng) was oxidized with periodate as follows: 5% NaIO 4 (20 l) was added to a solution of the metabolite in ethanol (25 l), and the mixture was left at room temperature for 1 h (24). The mixture was extracted with methylene chloride and analyzed by frit-FAB LC/MS using a mobile phase of 65% acetonitrile/water containing 0.7% glycerol.
Experiments to assess the kinetics of lowering the substrate in rat kidney homogenates indicated that the K m value for F 2 -1␣,25(OH) 2 D 3 was 2.0 M, which was similar to that for 1␣,25(OH) 2 D 3 (1.7 M). On the other hand, the V max for the former was 75 pmol/min/g tissue, whereas that for the latter was 510 pmol/min/g tissue.
Purification and Identification of the Major Metabolite (Peak X) of F 2 -1␣,25(OH) 2 D 3 -To isolate the major metabolite (peak X), 30 g of F 2 -1␣,25(OH) 2 D 3 was incubated with 150 ml of 10% (w/v) kidney homogenates obtained from rats given 1␣,25(OH) 2 D 3 . The lipid extract of the metabolite was purified on a Sep-pak cartridge followed by two HPLC separations (straight and reverse phase) to yield 1.7 g of the homogeneous metabolite (peak X). The UV spectrum of peak X was typical of the vitamin D triene system ( max at 265 nm, min at 228 nm in ethanol). The frit-FAB LC mass spectrum ( Fig. 2A)  To examine whether a vicinal diol function is present in the metabolite, peak X was treated with NaIO 4 , and the product was analyzed by frit-FAB LC/MS. HPLC showed a single major peak, and the MS spectrum of the peak showed a parent ion at 437 (M ϩ H) ϩ (relative intensity, 11.5%) and fragment ions at m/z 419 (M ϩ H Ϫ H 2 O) ϩ (relative intensity, 13.3%), and m/z 401 (M ϩ H Ϫ 2H 2 O) ϩ (relative intensity; 10.2%). These results indicate that a formyl group was removed by NaIO 4 oxidation and that there is a terminal vicinal diol function in the side chain of peak X. Thus, we identified peak X as one of the 25-epimers of F 2 -1␣,25,26(OH) 3 D 3 . Finally the structure of peak X was confirmed by direct comparison with two chemically synthesized 25-epimers of F 2 -1␣,25,26(OH) 3 D 3 by HPLC and frit-FAB LC/MS. The metabolite comigrated with the two synthesized compounds (epimers were not separated under the HPLC conditions) (Fig. 3), and the mass spectra of the isolated and synthetic epimers were identical (Fig. 2).  (5). Therefore we examined the ability of F 2 -1␣,25(OH) 2 D 3 to compete for the metabolism of 1␣,25(OH) 2 D 3 in rat kidney homogenates. F 2 -1␣,25(OH) 2 D 3 inhibited the conversion of 1␣,25(OH) 2 D 3 into water-soluble metabolite(s) in a concentration-dependent manner (Fig. 4, left). Furthermore, HPLC analysis of the lipid-soluble fraction revealed that F 2 -1␣,25(OH) 2 D 3 inhibited the 24-hydroxylation of 1␣,25(OH) 2 D 3 (Fig. 4, right). These results indicated that F 2 -1␣,25(OH) 2 D 3 competes with the interaction of CYP24 and 1␣,25(OH) 2 D 3 .

DISCUSSION
CYP24 was discovered as the enzyme responsible for the hydroxylation at C-24 in the metabolism of 1␣,25(OH) 2 D 3 and 25(OH)D 3 (25). Recently, it was found that human recombinant CYP24 also catalyzes the C-23 hydroxylation of 25(OH)D 3 (11), indicating that this enzyme has multicatalytic functions. However, there is no evidence that CYP24 hydroxylates any other carbons than C-24 or C-23 of vitamin D compounds. In this study, we showed that F 2 -1␣,25(OH) 2 D 3 , C-24 of which is protected from the hydroxylation by fluorination, is metabolized into F 2 -1␣,25,26(OH) 3 D 3 by a 1␣,25(OH) 2 D 3 -induced enzyme in the rat kidney (Figs. 1-3). The enzyme involved in this hydroxylation was CYP24 (Figs. 4 and 5A). This is the first report to describe that CYP24 hydroxylates a carbon other than C-24 and C-23 of vitamin D compounds. It is generally accepted that the fluorine atom mimics the hydrogen atom. A computer analysis confirmed that 24,24-F 2 -1␣,25(OH) 2 D 3 was very similar to that of 1␣,25(OH) 2 D 3 , though the electronegativity and hydrophobicity of the fluorine atom were stronger than those of the hydrogen atom (data not shown). Thus the possibility cannot be ruled out at present that the fluorine atoms at C-24 influence the susceptibility of the neighboring carbons to CYP24. Hydroxylation at C-26 of F 2 -1␣,25(OH) 2 D 3 made C-25 asymmetric. The stereochemical configuration at C-25 of this metabolite has yet to be determined. According to the model proposed by Dilworth et al., CYP24 selects its hydroxylation site(s) by the distance from the vitamin D ring structure (13). The results of the present study, however, suggest that the hydroxylation site selected by the enzyme is not necessarily strict and that CYP24 can hydroxylate a carbon other than C-24 and C-23 when C-24 is protected from metabolic hydroxylation. Therefore, it is highly likely that CYP24 is also responsible for the C-26 hydroxylation of the vitamin D 3 metabolites in vivo. At present, the possibility cannot be excluded that enzymes other than CYP24, such as CYP27, hydroxylate C-26 of 1␣,25(OH) 2 D 3 in vivo. In fact, CYP27 reportedly hydroxylates C-25 and C-26 of vitamin D 3 and 1␣(OH)D 3 (12). Under our conditions, however, kidney homogenates obtained from rats given either 1␣,25(OH) 2 D 3 or vehicle did not metabolize vitamin D 3 into 25(OH)D 3 or other metabolites (data not shown), indicating that no CYP27 is present in the kidney. In addition, CYP27-transfected ROS17/ 2.8 cells did not metabolize F 2 -1␣,25(OH) 2 D 3 into F 2 -1␣,25,26(OH) 3 D 3 (Fig. 5B). Therefore, the C-26 hydroxylation of F 2 -1␣,25(OH) 2 D 3 in kidney homogenates does not appear to be mediated by CYP27.
Two methyl groups at C-25 of 1␣,25(OH) 2 D 3 (or 25(OH)D 3 ) are heterotopic. Hydroxylation of one of the methyls yields a new chiral center at C-25. Hydroxylation of the pro-S-methyl group produces 25R configuration and pro-R-methyl group 25S configuration. Two types of 26-oxygenated vitamin D 3 metabolites have been found; one is the metabolites with 25S configuration such as 1␣,25S,26-trihydroxyvitamin D 3 (26) and 25S,26-dihydroxyvitamin D 3 (27), and the other is those with 25R configuration such as 25R-hydroxyvitamin D 3 -26,23S-lactone (28), 1␣,25R-dihydroxyvitamin D 3 -26,23S-lactone (29), and their precursors. It has also been reported that natural 25,26-dihydroxyvitamin D 3 is a mixture of 25R-and 25S-isomers (30). These results suggest that there are two C-26 hydroxylation enzymes; one catalyzes the hydroxylation of the pro-S-methyl and the other catalyzes the pro-R-methyl. It may be likely that these two types of hydroxylation at C-26 are catalyzed by CYP24 and CYP27, respectively. The stereochemical configuration at C-25 of the metabolite X, F 2 -1␣,25,26(OH) 3 D 3 , is now under investigation.
In conclusion, CYP24 appears to solely regulate the intracellular concentration of the VDR ligand and hence the VDRmediated transactivation in the target cells of vitamin D. It is highly likely that CYP24 catalyzes all three known catabolic pathways of 1␣,25(OH) 2 D 3 , namely the C-23, C-24, and C-26 hydroxylation pathways, further emphasizing the importance of this enzyme in regulating vitamin D metabolism and function.