Isolation and Characterization of a Cytochrome P-450 from Rat Kidney Mitochondria That Catalyzes the 24-Hydroxylation of 25-Hydroxyvitamin D3*

A cytochrome P-450 that catalyzes the 24-hydrox- ylation of 25-hydroxyvitamin D3 (P-45Occ24: P-450~hole~alcifero124) was purified to electrophoretic homogeneity from the kidney mitochondria of female rats treated with vitamin D3 (Ohyama, Y., Hayashi, S., and Okuda, K. (1989) FEBSLett. 255,405-408). The molecular weight was 53,000, and its absorption spectrum showed peaks characteristic of cytochrome P- 450. The turnover number was 22 min“ and the specific content was 2.8 nmol/mg protein. The N-terminal amino acid sequence, Arg-Ala-Pro-Lys-Glu-Val-Pro-Leu-, is different from the N-terminal sequence of any other cytochrome P-450s so far reported. Upon reconstitution with the electron-transferring system of the adrenal mitochondria, the enzyme showed a high activ- ity in hydroxylating 25-hydroxyvitamin D3 as well as la,25-dihydroxyvitamin DS at position 24. However, the purified enzyme hydroxylated neither vitamin D3 nor la-hydroxyvitamin D3. The enzyme was also in- active toward xenobiotics. The enzyme hydroxylated 25-hydroxyvitamin D3 at position 24 but not at la, indicating that the enzyme is distinct from


Yoshihiko Ohyama and Kyuichiro Okuda
Iq'rom the DeDartment of Biochemistrv. Hiroshima University School of Dentistry, 1-2-3 Kasumi, Minumi-ku, -. Hiroshima 734, Japan ' A cytochrome P-450 that catalyzes the 24-hydroxylation of 25-hydroxyvitamin D3 (P-45Occ24: P-450~hole~alcifero124) was purified to electrophoretic homogeneity from the kidney mitochondria of female rats treated with vitamin D3 (Ohyama, Y., Hayashi, S., and Okuda, K. (1989) FEBSLett. 255,[405][406][407][408]. The molecular weight was 53,000, and its absorption spectrum showed peaks characteristic of cytochrome P- 450. The turnover number was 22 min" and the specific content was 2.8 nmol/mg protein. The N-terminal amino acid sequence, Arg-Ala-Pro-Lys-Glu-Val-Pro-Leu-, is different from the N-terminal sequence of any other cytochrome P-450s so far reported. Upon reconstitution with the electron-transferring system of the adrenal mitochondria, the enzyme showed a high activity in hydroxylating 25-hydroxyvitamin D3 as well as la,25-dihydroxyvitamin DS at position 24. However, the purified enzyme hydroxylated neither vitamin D3 nor la-hydroxyvitamin D3. The enzyme was also inactive toward xenobiotics. The enzyme hydroxylated 25-hydroxyvitamin D3 at position 24 but not at l a , indicating that the enzyme is distinct from that catalyzing la-hydroxylation. The reaction followed Michaelis-Menten kinetics, and the K,,, value for 25-hydroxyvitamin D3 was 2.8 PM. Both vitamin D3 and lahydroxyvitamin D3 inhibited the 24-hydroxylation of 25-hydroxyvitamin D3 in a competitive, concentration-dependent manner. 25-Hydroxyvitamin D3 24hydroxylase activity was significantly inhibited by 7,8-benzoflavone, ketoconazole, and CO, whereas it was only slightly inhibited by aminoglutethimide, metyrapone, and SKF-525A. Mouse antibodies raised against the cytochrome P-450 inhibited the reaction about 70% and reacted with the P-45Occ24 in immunoblotting but did not react with other kinds of cytochrome P-450 in rat liver microsomes and mitochondria. In 1972Omdahl et al. (1972 found that a metabolite of 25hydroxyvitamin D3 other than la,25-&hydroxyvitamin D3, named peak Va, could be synthesized in the kidney, depending on the concentration of calcium in the diet, i.e. mitochondria isolated from high calcium-fed animals produced peak Va, whereas those isolated from low calcium-fed animals metabolized 25-hydroxyvitamin D3 to la,25-dihydroxyvitamin D3.
* This study was supported in part by Grants-in-Aid for Scientific Research 61771449 (to Y. 0.) and 63635004 (to K. 0.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Subsequently, the structure of peak Va was identified as 24R,25-dihydroxyvitamin D3 by Holick et al. (1972), Lam et al. (1973) and Tanaka et al. (1975). Holick et al. (1972 speculated that when the animal is hypocalcemic, the need for calcium is interpreted in some way by the kidney, resulting in the "turning on" of the 25-hydroxyvitamin D3 la-hydroxylation in order to increase the flux of calcium concentration, while when the animal is normocalcemic or hypercalcemic, there is little need for a supply of calcium and therefore the kidney "shuts off" the 25-hydroxyvitamin D3 la-hydroxylase system and produces instead another metabolite, 24,25-dihydroxyvitamin D3. Accordingly, clarification of how this switchover of the two enzyme activities occurs may lead to elucidation of the mechanism of calcium homeostasis in the animal body, and for that matter purification and characterization of the individual enzymes are of pivotal importance. However, owing to their extremely low contents and lability, it was difficult to purify these enzymes. As to the properties of 25-hydroxyvitamin D3 24-hydroxylase, several reports have been published to date. Knutson and DeLuca (1974) observed that almost all of the recovered 24hydroxylation activity was located in the crude nuclear debris and heavy mitochondrial fraction of chickens raised on a high calcium, vitamin Da-supplemented diet. They also showed that the reaction required NADPH and oxygen, suggesting that the enzyme is a mixed function oxidase. When NADPH was supplied through the oxidation of succinate and malate, the electron transport inhibitors, cyanide, antimycin, and carbon monoxide, inhibited the reaction. However, when NADPH was supplied directly, these inhibitors had no effect on the 24-hydroxylase activity. More direct evidence that 25hydroxyvitamin D3 24-hydroxylase is a mixed function oxidase was obtained by the work of Madhok et al. (1977), who observed that the oxygen enzymatically inserted as a hydroxyl group by chick kidney homogenates into the 24-position of 25-hydroxyvitamin D3 to give 24,25-dihydroxyvitamin D3 is derived exclusively from Kulkowski et al. (1979) observed that the enzyme is inhibited by carbon monoxide and metyrapone . Pedersen et al. (1983) also observed that 24-hydroxylation in rat kidney is inhibited by metyrapone. Subsequently, solubilization and reconstitution of the chick renal enzyme were achieved by Burgos-Trinidad et al. (1986). Although these results are highly suggestive that the enzyme is a member of the cytochrome P-450 family, no conclusive evidence has been published so far either by taking photochemical action spectra or by purifying the enzyme to a homogenous state.
Recently, the cytochrome P-450 involved in this reaction was highly purified in this laboratory based on the catalytic activity, and the enzyme activity was reconstituted from the 25-Hydroxyvitamin D3 24-Hydroxyhe cytochrome P-450, adrenodoxin, and NADPH-adrenodoxin reductase (Ohyama et al., 1989). In this paper the details of purification and some properties of this cytochrome P-450 (tentatively named cytochrome P-45Occ24) are described.

RESULTS
Induction of 25-Hydroxyuitamin D3 24-Hydroxyhe-The activity of 25-hydroxyvitamin D3 24-hydroxylation could be increased by intraperitoneal injection of vitamin D3 into rats. In this experiment, daily injection of 50,000 IU of vitamin D3 elicited about a 10-fold increase of enzyme activity.
Purification of 25-Hydroxyuitamin D3 24-Hydroxylase-25-Hydroxyvitamin D3 24-hydroxylase was solubilized from kidney mitochondria of female rats with cholate and Lubrol (Ohyama et al., 1989). The solubilized fraction was then subjected to a pentyl-Sepharose column. Of the several derivatives of Sepharose tested, including w-hexyl-, propyl-, and pentyl-Sepharose, the last one was found to be the most suitable for the present purpose. In our experience, preliminary separation of the enzyme at this stage seems to be important for further purification. Furthermore, owing to the very low content of cytochrome P-450 and relative abundance of other kinds of hemoprotein in kidney mitochondria, it was important to pursue the enzyme protein by measuring the catlytic activity throughout the preparation. A sodium dodecyl sulfate-polyacrylamide gel electrophoretogram of the eluate from high performance liquid chromatography with the activity is shown in Fig. 6 (lane 7). The active fraction revealed a single major band. Molecular weight of the enzyme calculated from the electrophoretogram was 53,000. The specific activity was 54.6 nmol/min/mg protein corresponding to 316-fold purification from the cholate-solubilized fraction, and specific content was 2.8 nmol of cytochrome P-450/mg of protein. That the major band on the electrophoretogram is the cytochrome P-450 responsible for 25-hydroxyvitamin D3 24-hydroxylase was confirmed by Western blotting analysis with antibody raised against the purified enzyme, which inhibited the enzyme activity.
Absorption Spectra of 25-Hydroxyvitamin D3 24-Hydroxylme-As shown in Fig. 1, the oxidized form exhibited an intense Soret absorption peak at 417 nm, typical of a low spin state of P-450, and CY and p bands at 570 and 535 nm, respectively. On addition of CO to the reduced form, a Soret peak characteristic of heme proteins of the cytochrome P-450 family was observed at 453 nm (inset). N-terminal Amino Acid Sequence-The N terminus of 25hydroxyvitamin DB 24-hydroxylase was sequenced by Dr. Yoshiyasu Yabusaki (Biotechnology Laboratory, Takarazuka Research Center, Sumitomo Chemical Co., Ltd.) as follows:

Arg-Ala-Pro-Lys-Glu-Val-Pro-Leu-.
The sequence is different from that of any other cytochrome P-450 so far reported, suggesting that 25-hydroxyvitamin D3 constitutes a unique family of cytochrome P-450.
Identification of the Reaction Product-Identification of the product obtained by incubating 25-hydroxyvitamin D3 with the purified P-45Occ24 together with adrenodoxin and NADPH-adrenodoxin reductase was performed by HPLC using two regular and one reversed phase system. A peak was observed at a retention time apparently corresponding to 24,25-dihydroxyvitamin D3, whereas no peak was observed at the retention time corresponding to la,25-dihydroxyvitamin D3. When the product eluted from HPLC was treated with periodate, the product could no longer be detected. From these results the product was identified as 24,25-dihydroxyvitamin D3 (Ohyama et al., 1989). Although the absolute configuration at position 24 is not known at this time, it is presumed to be 24R since it was established that the natural form of 24,25dihydroxyvitamin D3 has the 24R-configuration (Tanaka et al., 1975).
Kinetic Properties of 25-Hydroxyuitamin D3 24-Hydroxylme-The enzyme reaction proceeded in a time-linear fashion up to 10 min under the assay conditions described. The pH activity curve showed a broad peak at pH 7.7. When the substrate concentration was varied, the reaction followed Michaelis type kinetics, as shown in Fig. 2. The Michaelis constant calculated from the Lineweaver-Burk plot was 2.8 p~. The turnover number calculated from the Vmax value was 22 min". Interestingly, dilauroylglyceryl-3-phosphorylcholine, which is required for reconstitution of liver microsomal cytochromes P-450, enhanced the turnover number about 50% 25-Hydroxyuitamin D3 24-Hydroxylase (Table I). This effect was only observed in the purified preparation. Substrate Specificity-Substrate specificity is shown in Table I. As shown in the table the enzyme showed the highest activity toward 25-hydroxyvitamin D3. However, when the enzyme was incubated with vitamin D3 and la-hydroxyvitamin D3, no peak was observed in the region corresponding to monohydroxyvitamin D3 and dihydroxyvitamin D3, respectively, being consistent with the postulation of Tanaka et al. (1977) that a 25-hydroxyl group is required for kidney mitochondrial 24-hydroxylation of vitamin D3 derivatives. In fact, the enzyme hydroxylates la,25-dihydroxyvitamin D3 to give la,24,25-trihydroxyvitamin D3. Further kinetic studies with this substrate were, however, not attempted at this point. The enzyme did not show any activity toward xenobiotics such as benzphetamine, 7-ethoxycoumarin, and benzo[a]pyrene, suggesting the unique nature of P-45Occ24.
Inhibition of Enzyme Actiuity- Table I1 shows the effect of various P450 inhibitors on 25-hydroxyvitamin D3 24-hydrox- The incubation mixture contained 2.4 pmol of cytochrome P-450, 2 nmol of adrenodoxin, 0.05 unit of NADPH-adrenodoxin reductase, 0.5 pmol of EDTA, 0.5 pmol of NADPH, substrate to give a final concentration as described, and 50 pmol of Tris-HC1 (pH 7.7) in a final volume of 0.5 ml. Incubation was conducted at 37 "C for 7 min.
10 pg per incubation mixture. ND, not detectable. ylase. As shown in the table neither metyrapone, which is known to inhibit beef adrenal mitochondrial cytochrome P-450118 (Harding et al., 1969), nor aminoglutethimide, which is known to inhibit beef adrenal mitochondrial cytochrome P-450,,, (Dexter et al., 1967), inhibited the reaction appreciably. SKF-525A also showed very little inhibitory action. By contrast, 7,8-benzoflavone, which is known to inhibit arylhydrocarbon hydroxylase (Wiebel et al. 1971), and ketoconazole, a synthetic fungicide known to inhibit aromatase (Mason et al., 1985), inhibited the enzyme activity significantly. Some vitamin D, analogs also inhibited the enzyme activity, of which la-hydroxyvitamin D, showed the highest inhibition. The type of inhibition was competitive (Fig. 3), Ki value being 0.27 p M (inset). Vitamin D3 inhibited less strongly than la-hydroxyvitamin D,, and vitamin Dz much less strongly. When the 3P-hydroxyl group of vitamin DB was tosylated, the derivative revealed no inhibition, suggesting that the 3P-hydroxyl group is important for binding of substrates or inhibitors to the cytochrome.
The activity of 25-hydroxyvitamin D3 24-hydroxylase was also inhibited by carbon monoxide (Fig. 4). The Warburg partition constant, K , was determined since the sensitivity of the 25-hydroxyvitamin D3 to CO is one of a critical point to differentiate this enzyme from 25-hydroxyvitamin D3 l ahydroxylase as described under "Discussion" and was found to be 0.46, a value similar to that of other mixed function oxidases (Triilzsch et al., 1973).

Effect of Antibodies on the Reconstituted 25-Hydroxyuitamin
D3 24-Hydroxylase-Mice were immunized by injecting the purified P-45Occ24, and their serum was collected at the time when their spleens were removed for preparation of monoclonal antibodies. Fig. 5 shows that the antiserum inhibited the 24-hydroxylation by about 70% at a concentration of 400 pg of serum protein/0.5 ml of reaction mixture. However, three different monoclonal antibodies inhibited the reaction

FIG. 5. Effect of anti-P-450cc24 on the 25-hydroxyvitamin
Da 24-hydroxylation activity. The reaction was conducted as described under "Experimental Procedures" using 17.5 pg of protein; a partially purified sample was used (activity of 1.6 nmol/min/mg protein). ., control serum; 0 , anti-P-450cc24 serum.
only slightly, although they reacted with P-45Occ24 on immunoblotting analysis (data not shown).
Immunoblotting Analysis- Fig. 6 shows SDS-polyacrylamide gel electrophoresis and immunoblotting analysis with antibodies raised against P-45Occ24 in mice. The antibody reacted with purified P-45Occ24. However, it did not react with P-450b, P-450c, P-450rh7,,, or the P-450 catalyzing vitamin Da 25-hydroxylation in rat liver mitochondria.

DISCUSSION
The content of 25-hydroxyvitamin DR 24-hydroxylase in normal rat kidney mitochondria is extremely low. T o purify the enzyme, therefore, it was important to enrich the enzyme concentration of the starting material. For this purpose the enzyme was induced by injecting vitamin Ds into animals before they were killed. By this procedure the activity was increased about 10-fold. Even after this treatment, however, the concentration of the enzyme was not as high as that of most cytochrome P-450s in liver microsomes. It was therefore necessary to select a highly efficient column for hydrophobic chromatography. None of the columns commonly used for hydrophobic chromatography, eg. w-aminohexyl-Sepharose immunoblotting of various P-450 preparations. Anti-P-45Occ24 was used as a probe. Panel A , silver staining;panel H, immunoblotting with anti-P-450cc24. I , P-450cc24 (crude preparation, 0.4 p g ) ; 2, P-45OCcr5 (Hayashi et al., 1986); 3, P-450b; 4, P-450c;5, P-450chTn (Ogishima et al., 1987); 6, P -4 5 0 1 ,~~, (Masumoto ~t al., 1988); 7, the purified P-45Occ24 (0.05 pg); 8, molecular weight standards (phos- or w-aminooctyl-Sepharose, was effective for this purpose. We therefore prepared and tested various kinds of hydrophobic columns. As a result it was found that pentyl-Sepharose was the most effective one for the purpose of this study. Furthermore, owing to its extremely low concentration in the tissue, the P-450 was too precious to use by measuring its concentration by the CO difference spectrum, as is commonly employed in purification of most cytochrome P-450s acting on xenobiotics or carcinogens. Other kinds of hemoproteins existing in kidney mitochondria also interfered with this measurement. The present purification was therefore performed based on enzymatic activity. It was recently shown in this laboratory that such a method is important in purifying endogenous cytochrome P-450s that are minor components (Ogishima et al., 1987).
The specific activity of the purified 25-hydroxyvitamin D:, 24-hydroxylase was several thousand-fold higher than those of solubilized preparations so far reported (Rurgos-Trinidad et al., 1986;Gray and Ghazarian, 1989). However, one of the problems of the present method is its relatively low yield. The loss of activity occurs at the final stage of high performance liquid chromatography on DEAE-5PW. However, to obtain a homogeneous preparation it was the most important step and could not be omitted. The loss of enzyme activity seemed to be due to removal of heme during this chromatography. The yield of the enzyme may be raised by further improvement of this stage. Cytochrome P-45Occ24 may constitute a novel family of cytochrome P-450, since the N-terminal amino acid sequence, as so far determined, is different from that of any such cytochrome P-450s so far reported. 24-Hydroxylation activity was inhibited by vitamin D:% and la-hydroxyvitamin Ds and at least the latter was a competitive inhibitor from kinetic studies. These results indicate that these compounds could occupy the active site of the enzyme, while they are not hydroxylated a t all. The requirement for the 25-hydroxyl group as postulated by Tanaka et al. (1975) was thus confirmed from kinetic studies as well.
It is well known that in kidney mitochondria an enzyme exists which catalyzes the la-hydroxylation of 25-hydroxyvitamin Ds, an active metabolite of vitamin D:,. A question has been raised whether 25-hydroxyvitamin D, 24-hydroxylation and la-hydroxylation are catalyzed by a single enzyme. Knutson and DeLuca (1974) have shown several reasons that 24hydroxylase is distinct from la-hydroxylase, one of which is the failure of inhibition of 24-hydroxylation by a carbon monoxide to oxygen ration of 3:l. However, the value of the

25-Hydroxyuitamin 0 3 24-Hydroxylase
Warburg constant, K , for CO inhibition of the 24-hydroxylation presently determined was 0.46. This value is similar to those reported for la-hydroxylation of 25-hydroxyvitamin D3 in chick kidney mitochondria (Ghazarian and DeLuca, 1974;Henry and Norman, 1974), suggesting that the CO inhibition study is not suitable for differentiation of these reactions. Inhibition study using other cytochrome P450 inhibitors such as metyrapone and ketoconazole is not suitable for this differentiation either, since these inhibitors do inhibit both 24hydroxylation and la-hydroxylation more or less (Paulson and DeLuca, 1985;Henry, 1985). In contrast, the direct evidence obtained in this study that the purified enzyme did not catalyze the la-hydroxylation highly suggests that 25-hydroxyvitamin D3 24-hydroxylase is a distinct cytochrome P-450 from 25-hydroxyvitamin D3 la-hydroxylase. Conclusive evidence would be obtained by expression of a cDNA clone encoding this enzyme.
Holick et al. (1973) isolated 1,24,25-trihydroxyvitamin D3 from intact rats treated with 25-hydroxyvitamin D3 and from 24,25-dihydroxyvitamin D3 incubated with chicken kidney homogenates. The compound has 60% of the activity of vitamin D3 in curing rickets. It is less active on a weight for weight basis than la,25-dihydroxyvitamin D3 in stimulating and sustaining intestinal calcium transport and bone calcium mobilization. It may therefore be surmised that the enzyme may also function to produce 1,24,25-trihydroxyvitamin D3 from 1,25-dihydroxyvitamin D3 so that the action of surplus la,25-dihydroxyvitamin D3 can be alleviated. In conclusion, we have purified and characterized 25-hydroxyvitamin D3 24hydroxylase from rat kidney mitochondria that showed a high turnover number but did not catalyze la-hydroxylation of 25hydroxyvitamin D3. Determination of N-terminal amino acid sequence and preparation of antibodies will make it possible to isolate a cDNA encoding this enzyme and will pave the way for study of the mechanism of calcium homeostasis in animals at a molecular biological level.