Biochimica et Biophysics Acfa, 1166 (1993) 177-182 0 1993 Elsevier Science Publishers B.V. All rights reserved 0005-2760/93/$06.00

177

BBALIP 54091

24-, 25 and 27-hydroxylation of cholesterol by a purified preparation of 27-hydroxylase from pig liver

Erik Lund, Ingemar Bjijrkhem, Catrin Furster and Kjell Wikvall

Department of Clinical Chemistry, Karolinska Institutet, Huddinge Hospital, Huddinge (Sweden) and Department of Pharmaceutical Biochemistry, University of Uppsala, Uppsala (Sweden)

(Received 14 July 1992)

Key words: Hydroxylation; Cholesterol; 27-Hydroxylase; Cytochrome P-450; (Pig liver mitochondria)

Pig liver mitochondria were found to catalyze 27-, 25 and 24-hydroxylation of cholesterol at relative rates of about 1: 0.2: 0.04. An apparently homogeneous preparation of pig liver mitochondrial cytochrome P-450-27 was found to catalyze the same three hydroxylations at about the same relative rates when reconstituted with adrenodoxin and adrenodoxin reductase. The 24-hydroxycholesterol formed was shown to consist of one of the two possible stereoisomers. When using specifically deuterium- labeled substrates a significant isotope effect was observed in the case of 24-hydroxylation (KH/KD > lo>, but not 25hydroxyl- ation (KH/KD = 1.0, or 27-hydroxylation (KJK,, = 1.1). The difference between the 24-hydroxylation and the other two hydroxylations may be due to different interactions between cholesterol and the same enzyme, with a resulting difference with respect to the rate-limiting step in the reaction. The physiological significance of the mitochondrial 24-hydroxylation is discussed.

Introduction

Liver mitochondria contain a cytochrome P-450 species, P-450-27, that catalyzes 27-hydroxylation of cholesterol and a number of other C,, steroids (for reviews, see Refs. l-3). Cytochrome P-450-27 species from liver mitochondria of rabbit [4], rat [5], pig [6] and human [71 have been extensively characterized, some of them also at the gene level. This enzyme is involved in bile acid biosynthesis from cholesterol. Since 27-hy- droxycholesterol is a potent suppressor of cholesterol synthesis in various cell systems, it has been suggested that the enzyme may also be of importance in connec- tion with regulation of cholesterol homeostasis [l]. The enzyme is present not only in the liver but also in various other organs and tissues [l-31. Most interesting was that the abundance of mRNA for the enzyme was found to parallel the cholesterol biosynthesis capacity of the tissues [81.

In addition to 27-hydroxylase activity, liver mito- chondria also contain 25-hydroxylase activity towards both cholesterol and vitamin D, [9,10]. The same activ- ities are also found in purified fractions of cytochrome P-450-27 [ll-131. With the present state of knowledge it is not yet known with certainty if there is only one enzyme responsible for sterol 27-hydroxylase activity,

Correspondence to: I. Bjiirkhem, Department of Clinical Chemistry, Huddinge Hospital, S-141 86 Huddinge, Sweden.

sterol 25-hydroxylase activity and vitamin D, 25-hy- droxylase activity. COS-cells transfected with cy- tochrome P-450-27 cDNA express both sterol 27- and vitamin D, 25-hydroxylase activities [14], suggesting that the activites are derived from the same gene. Experiments with antibodies towards the sterol 27-hy- droxylase [121 and some kinetic experiments with crude mitochondrial fractions [ll] suggest, however, that dif- ferent enzymes may be involved. The apparent contra- diction may be due to posttranslational modifications of the cytochrome P-450-27 enzyme, resulting in differ- ent isozymes with different substrate specificities [3,14]. Another alternative is that in addition to a cytochrome P-450-27 enzyme, possessing both 27- and 25-hydroxyl- ase activities, there may be another very similar cy- tochrome P-450 enzyme that catalyzes 25-hydroxyl- ation.

In the present work we report that pig liver mito- chondria in addition to sterol 27- and 25-hydroxylase activity also contain 24-hydroxylase activity. This activ- ity could also be found in a purified preparation of cytochrome P-450-27 from pig liver. The ratio between the three activites was about the same in the purified enzyme fraction as in the crude mitochondrial fraction.

Materials and Methods

Incubations were made using either pure unlabeled cholesterol, or a mixture of [6,7,7- ‘H Jcholesterol

178

(synthesized as described in Ref. 15) and [25,26,26,26, 27,27,27- 2H,]cholesterol (from Larodan AB, Malmii, Sweden), or a mixture of [26,26,26,27,27,27- 2H,lcholesterol (from Medical Isotopes, Inc. NH), and [23,23,24,24,25-2H,]cholesterol (synthesized as de- scribed in Ref. 16). The mixtures were nearly equimo- lar and the exact composition was determined using gas chromatography-mass spectrometry. An equimolar mixture of (24R)- and (24S)-24-hydroxycholesterol was synthesized as described [ 161.

Cytochrome P-450-27 was prepared from pig liver mitochondria as described in Ref. 6. The preparation showed a single major protein band with an apparent M, = 53 000 upon SDS-polyacrylamide gel elec- trophoresis. The specific cytochrome P-450 content was 7.5 nmol per mg protein. These properties were the same as those reported in Ref. 6. Adrenodoxin and adrenodoxin reductase were prepared as described previously [4]. Incubations with intact mitochondria were performed for 30 min with 25 pg unlabeled cholesterol dissolved in 25 ~1 acetone, 5 mg mitochon- drial protein, 20 pmol isocitrate and 10 mM MgCl, in a total volume of 3 ml 100 mM Tris-HCI buffer, pH 7.4, containing 20% glycerol. Boiled mitochondria in- cubated in the same way served as controls. Incuba- tions with purified cytochrome P-450 were performed for 30 min with 10 pg unlabeled or deuterium-labeled cholesterol dissolved in 25 ~1 acetone, 0.2 nmol cy- tochrome P-450, 2 nmol adrenodoxin, 0.2 nmol adren- odoxin reductase and 1 Frnol NADPH in total volume of 1 ml of 50 mM Tris-acetate buffer, pH 7.4, contain- ing 20% glycerol and 0.1 mM EDTA. Incubations without adrenodoxin and adrenodoxin reductase incu- bated in the same way and incubations without en- zymes containing only substrate and buffer served as controls. The incubations were terminated by addition of 5 ml trichloroethane/methanol (2: 1, v/v> and the organic phases obtained after centrifugation were dried with N, at room temperature.

The extracts ,from incubations with purified enzyme or isolated mitochondria were dissolved in chloroform, 0.5 ml, and subjected to solid phase chromatography using Bond-Elut NH, cartridges as described by Kaluzny et al. [17]. Only the neutral lipid fraction was collected. The solvent was removed under an argon stream and the residue was dissolved in 100 ~1 of methanol, of which 50 ~1 were purified on HPLC. The system consisted of two type 420 pumps, a type 491 dynamic gradient mixer, and a type 432 UV detector (all from Kontron Instruments SpA, Italy). The system was controlled by a Kontron 457 software controller. Reversed phase HPLC was performed using a Nova- Pak C,, column (8 x 100 mm, 4 pm particle size, from Waters Inc.) with a methanol-water gradient (O-1 min: 90: 10 isocratic, l-30 min: linear gradient from 90: 10 to 100: 0, 30-35 min: 100: 0 isocratic, v/v) at a

flow rate of 1 ml/min. Retention times were (min): 24-hydroxycholesterol, 10.9; 25-hydroxycholesterol, 11.7; 27-hydroxycholesterol, 12.6; cholesterol, 31.3. The effluent was collected between 10 and 14 min for a hydroxysterol fraction and between 30 and 33 min for a cholesterol fraction. The hydroxysterol fraction was divided into two fractions of equal size and a small sample was drawn from the cholesterol fraction. The solvent was evaporated under an argon stream and one of the hydroxysterol samples and the cholesterol sam- ple was trimethylsilylated by dissolving it in 50 ~1 of pyridine/ hexamethyldisilazane/ trimethylchlorosilane (3: 2: 1, v/v> and heating the mixture at 60°C for 30 min. After evaporating the derivatization mixture, the residue was dissolved in 100 ~1 of hexane and sub- jected to CC-MS analysis. The second half of the hydroxysterol fraction was first converted into terf-

butyldimethylsilyl ether according to the method of Corey and Venkateswarlu [18]. The sterically hindered 25-hydroxy group is not derivatized under these condi- tions. After conversion, the sample was diluted with water and extracted with ethyl acetate. The solvent was removed under an argon stream and the sample was trimethylsilylated as above and subjected to GC-MS. An HP 5970 MSD instrument (HP: Hewlett-Packard Co.) operating at 70 eV was used. Instrumental condi- tions: DB-5 column, 13.3 m, 0.18 mm i.d., 0.4 pm phase thickness. Oven program: initial temp. 180°C 1.5 min, 19”C/min to 270°C lO”C/min to 300°C where the oven was kept for 10 min. The analysis of recovered cholesterol was performed in order to check the possi- ble occurrence of contaminating unlabeled cholesterol. The compounds were identified by taking mass spectra, and isotope effects were determined by the selective monitoring of the molecular ions of the differently labeled trimethylsilylated 24- and 27-hydroxycholester- 01 species. With regard to 25-hydroxycholesterol, the M-57 ion of the TBDMS-TMS derivative was used for the determination of isotope effects. The calculation was made by taking the intensity ratio of unlabeled to labeled product after correction for natural isotope abundances and for the initial ratio of labeled to unlabeled cholesterol in the mixture subjected to the enzyme or the mitochondria. Note: by ‘unlabeled’ is meant undeuterated at the position of hydroxylation here. The following ions were monitored: 24- and 27-hydroxycholesterol formed from pure unlabeled cholesterol, m/z 546; 24-hydroxycholesterol formed from [26,26,26,27,27,27-2H,]cholesterol, m/z 552; 24- hydroxy-cholesterol formed from [23,23,24,24,25- 2Hs]cholesterol, m/z 550; 25-hydroxycholesterol formed from pure unlabeled cholesterol, m/z 531; 25-hydroxycholesterol formed from [25,26,26,26,27,27, 27-2H,]cholesterol, m/z 537; 25-hydroxycholesterol formed from [6,7,7-‘H,lcholesterol, m/z 534; 27-hy- droxycholesterol formed from [25,26,26,26,27,27,27-

179

27-OH- A 25-OH-

eholerterot

cholesterol I

B 25-OH-

cholesterol

I

ch OH-

erol

24-OH- cholesterol

cholesterol

x 20

A , , A

11.5 12.0 12.5 13.0 11.5 12.0 12.5

Time (min) Time (min)

L-

13.0

Fig. 1. (A) GC-MS ion chromatogram (m/z 546) obtained from a typical incubation of cholesterol with intact mitochondria from pig liver. The

peak eluting at 11.7 min corresponds to 24-hydroxycholesterol, the peak at 11.9 min corresponds to 2Shydroxycholesterol and that at 12.7 min

corresponds to 27-hydroxycholesterol. (B) Ion chromatogram (m/z 549) obtained from a typical incubation of [6,7,7-*H,]cholesterol with a

purified preparation of cytochrome P-450-27 from pig liver. Otherwise as above. The conditions of incubation are described in Materials and Methods.

‘H,lcholesterol, m/z 552; 27-hydroxy-cholesterol formed from [6,7,7-2H,]cholesterol, m/z 549.

Results

Gas chromatographic separation of (240 and Incubation of pig liver mitochondria with isocitrate (24R)-24-hydroxycholesterol (TMS ether) was obtained and unlabeled cholesterol under the conditions de- using a DB-210 column (30m X 0.32 mm, 0.25 pm) scribed in Materials and Methods resulted in an overall coupled to the HP GC-MS. Temperature program: conversion of about 0.01 nmol/min per mg protein 140°C 1 min; 20”C/min to 198°C where the oven was into more polar products. The products were purified kept for 70 min. The ions with m/z 145 and 413 were by preparative HPLC and were then subjected to com- detected. Retention times: (24S)-24-hydroxycholester- bined gas chromatography-mass spectrometry (as 01, 65.3 min; (24R)-24-hydroxycholesterol, 66.1 min. trimethylsilyl ether). As shown in Fig. lA, the major The absolute configurations were tentatively deter- product was identified as 27-hydroxycholesterol. 25- mined as described in Results. Hydroxycholesterol and 24-hydroxycholesterol were

(htcso-43)

loo- b

% 00. r 145

2 al ‘; 60. ._

: : ._ z 40- 129

5 L

20.

%ti3)

456 I

100 200 300 400 500

m/z m/z

Fig. 2. Mass spectrum of the trimethylsilyl ether derivative of (a) authentic 24-hydroxycholesterol, (b) compound eluting at the same retention

time on GC obtained after incubation of cholesterol with pig liver mitochondria.

minor products. All the products were identified by means of mass spectrometry and the mass spectrum of the trimethylsilyl ether of 24-hydroxycholesterol is shown in Fig. 2. The three products were formed in a ratio of about 1:0.2: 0.04. It was ascertained that no side-chain oxidized products were obtained in a blank incubation with boiled mitochondria or buffer only.

27-OH-Cholesterol

z 2 0)

; n

., A! 552 549

13.9 14.0 14.1

Time (min)

25-OH-Cholesterol-TMS-TBDMS

h

A% 537 534

b‘ 15.8 15.9 16.0

Time (min)

24-OH-Cholesterol

_A 552

550

13.00 13.04 13.08

Time (min)

A fraction of 24-hydroxycholesterol obtained from enzymatic incubation of cholesterol with purified cy- tochrome P-450-27 contained only one of the two possible stereoisomers of 24-hydroxycholesterol. This isomer was identical to the major isomer obtained from liver homogenate (both were found in different amounts) from mice fed 3% cholesterol for 4 days. This compound has previously been shown to have the 24Sconfiguration [231. We therefore tentatively con- clude that the sole isomer of 24-hydroxycholesterol formed by action of cytochrome P-450-27 on choles- terol has the 24S-configuration.

Incubation of [6,7a,7/C2H,]cholesterol with cy- tochrome P-450-27 reconstituted with adrenodoxin and adrenodoxin reductase gave the same three products as above in about the same proportions (Fig. 1B). The overall conversion was 9 nmol/min per mg protein. All three products were obtained in about the same rela- tive yield as that obtained in the mitochondrial incuba- tions. There were no traces of these compounds in the blank incubations with only substrate and buffer, or in incubations where the mitochondrial reductase compo- nents adrenodoxin and adrenodoxin reductase had been omitted.

In order to obtain information about the mechanism and the rate-limiting step in the three hydroxylations, the reconstituted system was incubated with a mixture of [25,26,26,26,27,27,27-2H,]cholesterol plus [6,7~7@- 2HJcholesterol as well as with a mixture of [23,23,24, 24,25-2H5]cholesterol plus [26,26,26,27,27,27-2H,]cho- lesterol (in each case an equimolar amount of the two specifically deuterated steroid). As shown in Figs. 3-5 there was no significant isotope effect in the 25- and the 27-hydroxylation of the 25- and 27-deuterated sub- strate (KH/KD = 1.1). There was however a highly significant isotope effect in the 24-hydroxylation of the 24-deuterated substrate (K,,/K, > 10).

Discussion

It was conclusively shown that pig liver mito- chondria contain 24-hydroxylase activity in addition to

Fig. 3. Ion chromatograms obtained after typical incubations of

differently labeled cholesterol with purified preparations of cy-

tochrome P-450-27 from pig liver corresponding to (a) the molecular

ion (m/z 549 and 552) of the trimethylsilyl ether derivative of

27-hydroxycholesterol obtained after incubation of an approximately

2 : 1 mixture of 6,7,7- *H,]cholesterol and [25,26,26,26,27,27,27-

‘H,]cholesterol; (b) the M-57 ion (m/z 534 and 537) of the

trimethylsilyl ether/ ferr-butyldimethylsilyl ether (TMS-TBDMS)

derivative of 25-hydroxycholesterol obtained after incubation of an

equimolar mixture of [6,7,7-2H,]cholesterol and [25,26,26,26,27,27, 27-‘H,]cholesterol; (c) the molecular ion (m/z 550 and 552) of the

TMS derivative of 24-hydroxycholesterol obtained after incubation of

an equimolar mixture of [23,23,24,24,25-2H,]cholesterol and [25,26,26,26,27,27,27-2H,]cholesterol.

the previously demonstrated 25 and 27-hydroxylase activities. According to a previous report by Aringer et al., small amounts of 24-hydro~cholesterol are also formed after incubation of cholesterol with rat liver mitochondria [ 191. 24-Hydroxylation of cholesterol has also been shown to occur in murine and bovine brain microsomes [20]. Liver microsomes from rat and hu- man contain a 24-hydro~lase active towards 5p- cholestan~-3~,7~,12~-trial, but this enzyme system seems to be inactive towards cholesterol 121,221. Saucier et al. reported the presence of significant amounts of (24S)-hydroxycholesterol in mouse liver [231. The levels increased markedly after feeding the mice with 3% cholesterol. In view of the resuits of the present work, it seems likely that the (24~)-hydro~cholesterol identi- fied by Saucier et al. had been formed by the action of a mitochondrial 24-hydroxylase. In some preliminary experiments we have shown that mouse liver mito- chondria but not mouse liver microsomes can catalyze 24-hydro~lation of cholesterol (Lund and Bjiirkhem, unpublished observation). In view of the fact that 24” hydroxycholesterol is a potent suppressor of the HMG- CoA reductase in various cell systems, the possibility has been suggested that enzymatic 24-hydroxylation may be of some regtdatory importance for cholesterol homeostasis [23].

The 24-hydroxylase activity was found also to be present in a cytochrome P-450-27 preparation from pig liver mitochondria. The enzyme preparation was ap- parently homogeneous judging from SDS-polyacryl- amide gel electrophoresis (see Ref. 6). The ratio be- tween 24-, 25 and 27-hydroxylase activity was the same in the purified fraction as in the crude mitochondrial fraction used as starting material in connection with the purification. It should perhaps be mentioned that one additional cytochrome P-450-27 preparation puri- fied from a separate animal was found to show the same ratio between the three activities as in the mito- chondria. These results support the contention that only one enzyme is responsible for all three activities. With the present state of knowledge we can not com- pletely exclude, however, that there may be a specific mitochondrial 24-hydroxylase in pig liver mitochondria with properties similar to those of the mitochondrial 27-hydroxylase. This problem may be solved for exam- ple, by an assay of 24-hydroxylase in COS-cells trans- fected with a cDNA corresponding to the cytochrome P-450-27. Attempts to use such an experimental model are in progress. In experiments with COS-cells trans- fected with cDNA encoding the rabbit cytochrome P-450-27 it has not been possible for us to accurately measure 27-hydroxylase activity towards cholesterol in view of the very low activity. Hence, it was not consid- ered meaningful to attempt to measure 24- and 25-hy- droxylase activities in these cells.

With regard to the very small amounts of 24-hy-

181

droxycholesterol formed, the possibility must be con- sidered that the compound was formed by autoxida- tion. This possibility was ruled out by showing that the 24-hydroxycholesterol formed consisted of only one of the two stereoisomers.

Regardless of whether or not a specific enzyme is involved in 24-hydroxylation of cholesterol, the mecha- nism of this hydro~lation may differ from that in- volved in 27-hydro~lation of cholesterol. We have previously shown that the mechanism of hydroxylation of a fatty acid by liver microsomes at the ultimate and penultimate carbon atom, respectively, are different and that there may be different rate-limiting steps in the two hydro~lations. Thus, a substitution of the specific hydrogen to be removed in the hydroxylation with deuterium is associated with a significant isotope effect in case of w - 1 hydroxylation but not in case of w-hydroxylation [24,25]. It was shown here that o - 2 hydroxylation (24-hydro~lation~ of cholesterol by the reconstituted system was associated with a significant isotope effect, whereas o-hydroxylation. (27-hydroxyl- ation) and w - 1 hydroxylation (25-hydroxylation) oc- curred without isotope effect. The magnitude of the isotope effect observed for the 24-hydroxylation is such that it is likely that the abstraction of the hydrogen is the rate-limiting step in the reaction [26]. It is thus evident that the mechanism of interaction between the substrate and the enzyme in connection with 24-hy- droxylation of cholesterol must be different from that occurring in connection with 25- and 27-hydroxylation of the same substrate. This does not exclude, however, the possibility that the same enzyme is involved in all three hydroxylations. It should be emphasized that the present results were obtained with pig liver mito- chondria and that the situation may be different with liver mitochondrial enzymes from other mammalian species.

Acknowledgments

This work was supported by the Swedish Medical Research Council (Projects 03X-3141 and 03X-218).

References

1 Javitt, N.B. (1990) J. Lipid Res. 31, 1X27-1533. 2 Bjorkhem, 1. (1992) J. Lipid Res. 33, 455-472. 3 W&all, K. (1991) in Frontiers in Biotransformation (Ruckpaul,

K. and Rein, H., eds.), Vol. 6, pp. 147-182, Akademie Verlag, Berlin.

4 Wikvall, K. (1984) J. Biol. Chem. 259, 3800-3804. 5 Okuda, K., Masumoto, 0. and Ohyama, Y. (1988) J. Biol. Chem.

263, 18138-18142. 6 Bergman, T. and Posthnd, H. (1991) Biochem. J. ‘276, 427-432. 7 Cah, J.J. and Russell, D.W. (1991) J. Biol. Chem. 266, 7774-7778. 8 Andersson, S., Davis, D.L., Dahlback, H., Jornvall, H. and Rus-

sell, D.W. (1989) J. Biol. Chem. 264, 8222-8229.

182

9 Bjorkhem, I. and Gustafsson, J. (1974) J. Biol. Chem. 249, 2528-

2535.

10 Bjorkhem, I. and Holmberg, I. (1978) J. Biol. Chem. 253, 842-849.

11 Bjiirkhem, I., Holmberg, I., Oftebro, H. and Pedersen, J.I. (1980)

J. Biol. Chem. 255, 5244-5249.

12 Dahlback, H. (1988) Biochem. Biophys. Res. Commun. 157, 30-

36.

13 Ohyama, Y., Masumoto, O., Usui, E. and Okuda, K. (19911 J.

Biochem. (Tokyochem) 109, 389-393.

14 Dahlblck, H. (1989) Acta Univ. Upps., Compr. Sum. Upps. Diss.

Fat. Pharm. 60 (Dissertation).

15 Lund, E., Diczfalusy, U. and Bjorkhem, I. (1992) J. Biol. Chem.

267, 12462-12467.

16 Lund, E., Breuer, 0. and Bjorkhem, I. (1992) J. Biol. Chem., in

press.

17 Kaluzny, M.A., Duncan, L.A., Merritt, M.V. and Epps, D.E.

(1985) J. Lipid Res. 26, 135-140.

18 Corey, E.J. and Venkateswarlu, A. (1972) J. Am. Chem. Sot. 94,

6190-6191.

19 Aringer, L., Eneroth, P. and Nordstrom, L. (1976) J. Lipid Res.

17, 263-272.

20 Lin, Y.Y. and Smith, L.L. (1974) Biochim. Biophys. Acta 348,

189-196.

21 Cronholm, T. and Johansson, G. (1970) Eur. J. Biochem. 16,

373-381.

22 Bjdrkhem, I., Gustafsson, J., Johansson, G. and Persson, B.

(1975) J. Clin. Invest. 55, 478-486.

23 Saucier, SE., Kandutsch, A.A., Gayen, A.K., Swahn, D.K. and

Spencer, T.A. (1989) J. Biol. Chem. 264, 6863-6869.

24 Hamberg, M. and Bjorkhem, I. (1971) J. Biol. Chem. 246, 7411-

7416.

25 Bjorkhem, I. (1977) Pharm. Ther. 1, 327-348.

26 Northrop, D.B. (1975) Biochemistry 14, 2644-2651.