The synthesis and activity in vitro of 25-masked-1α-hydroxylated vitamin D3 analogs

453

2322

THE SYNTHESIS AND ACTIVITY IN VITRO OF 25-MASKED-la-HYDHOXYLATED VITAMIN D3 ANALOGS Joseph L. Napoli, Mary A. Fivizzani, Alan H. Hamstra Heinrich K. S&noes, Hector I. DeLuca, and Paula H. Stern Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison Madison, Wisconsin 53706 and Department of Pharmacology, Northwestern University Chicago, Illinois 60611 Received 5-26-78 ABSTRACT la-Hydroxylated-25-masked-vitamin D3 analogs were synthesized as probes to help waluate the role of 25-hydroxyl group in hormone-receptor interactions of la,25_dihydroxyvitaminD (13a). Synthetic work on model systems showed that the steroidal z5-wroxyl group could be easily fluorinated in high yield with diethylaminosulfurtrifluoride. Treatment of 25-fluoro compound with acetic acid resulted in both elimination and displacement of fluorine by acetate. The desired la-hydroxy-25fluoro-vitaminD (14b) was obtained efficiently by fluorination and subsequent deace2ylxon of la,25_dihydroxyvitaminD 1,3-diacetate (13b). Also obtained was a mixture of la-hydroxyvit in D3-24- and 25&rl e= (15b). Both 14b and 15b were 300-400 times less active than 13a in the chick intestinal cytosrprotein binding assay, making these axogs similar in potency to la-hydroxyvitaminD in vitro. The essentially equivalent activity of 14b and 15b with l&Fdroxyvitamin D indicates that in the absence of a 25-hydroxyl group some alterations3to the side chain carbons of 13a may be tolerated without further weakening analogprotein interactions. The fluoroanalog 14b was also about 250 times less potent than 13a in stimulating bone resorption -in vitro. These compounds should prove to be valuable tools in aiding understanding of the salient structural features of the vitamin D3 metabolites. INTRODUCTION Vitamin D3 is a prohormone that undergoes two sequential hydroxylations -in vivo prior to manifesting optimal physfological activity. The first occurs in the liver to produce 25-hydroxyvitaminD3 (25-OH-D3), the major circulating metabolite. The second occurs in the kidney and is controlled directly or indirectly by blood calcium and phosphate levels. The final product, la,25_dihydroxyvitaminD3 (1,25-(OH)2D3),

VooZwne32, Number 4

S

W?BEOXtDI

November,

1s

2978

a hormone that mediates calcium and phosphate metabolism in at least bone and intestine [l-3]. Many analogs of vitamin D3 and its metabolites have been synthesized to examine structure-activity relationships and to search for agents with augmented or differential activity [3].

These studies have not yet

resulted in compounds more potent than 1,25-(OH) D 23

or which demonstrate

true selectivity; but they have outlined structure-function relationships. Thus, 3-deoxy-la-hydroxylated

vitamins retain some potency [4-61, which

suggests that the 3-hydroxyl is not essential in the expression of 1,25(0H)2D31~ final responses.

Conversely, a la-hydroxyl group is apparently

of greater importance for target-tissue responses to 1,25-(OH)2D3 as indicated by the fact that nephrectomized animals do not respond to physiological amounts of vitamin D3 compounds that are not la-hydroxylated 11,71.

However, the precise role of the 25-hydroxyl group in target-

tissue response -in vivo is unclear.

Conceivably, it could be important

only to intermediary metabolism since a 25-hydroxyl function is essential for l- and 24-hydroxylation

[8].

Alternately, it may function in target tissue

responses or may control the further metabolism of 1,25-(OH)2D3 [9,10]. To investigate these possibilities, we synthesized la-hydroxylated vitamin D 3 analogs blocked at the 25 position to prevent hydroxylation in vivo. -D3, c,

Our primary synthetic objective was la-hydroxy-25-fluorovitamin since the carbon-fluorine bond is more stable than the carbon-

hydrogen bond, particularly in vivo [U-13].

Moreover, because fluorine

and hydrogen are approximately similar in atomic dimensions, a fluoroanalog would be least likely to disrupt hormone-protein interactions.

We

were also interested in determining the effect of fluorination on vitamin

S D's spectrum of activity. vitamin D

During the synthesis of la-hydroxy-25-fluoro-

a mixture of la-hydroxylated 24- and 25-enes, G,

3'

isolated.

TDEOXDd

was

Since these compounds also represent analogs modified at C-

25, they are of interest.

This paper discusses the synthesis and -in

vitro evaluation of 14b and 15b. RESULTS AND DISCUSSION Synthesis.

Replacement of the 25-hydroxyl function in i-ether 1 by

a fluorine atom was attempted with different reagents.

None was entirely

satisfactory, except diethylaminosulfur trifluoride (DAST) 1141 which reacted with 1 to provide its 25-fluorinated analog 2 in high yield. The reaction was rapid and mild.

The mass spectrum of J_ demonstrated a

molecular ion at m/e 418 and a prominent fragment at m/e 398 (loss of HF).

Also

present is the usual fragmentation pattern of steroidal i-

ethers, i.e. loss of 15, 32 and 55 mass units from the molecular ion. The NMR spectrum of -2 is consistent with substitution of the 25-hydroxyl by fluorine.

Normally, 26 and 27 methyls in 25-hydroxylated steroids

resonate as a singlet at 6 1.21. spectrum of -2.

This intense singlet is absent in the

Instead, a doublet at 6 1.36 (.I= 22 Hz) is seen.

Both

the deshielding effect and the large coupling constant are characteristic of protons vicinal to fluorine. Attempts to rearrange -i-ether -2 in warm glacial acetic acid to fluorosteroid 1 resulted in a mixture of products, which were separated by preparative layer chromatography.

The faster moving band (2, 28 mg)

consisted of 5u-cholesta-7,24-dien-38-01 3-acetate and 5a-cholesta-7,25dien-38-01 3-acetate in 1.9:l.O ratio, as estimated by EMR.

The mixture's

456

e,

R=H

140.

R=Ac

l50,

E.

R*Ac

14,

RsH

e,R=H

R :Ac

NMR spectrum is essentially a superimposition of the previously reported NMRspectra of the individual olefins 2 115,161. supported the structure assignments.

Mass spectral data

The slower migrating band contained

39 mg of pure 5a-cholest-5-en-3S,25-diol 3,25-diacetate, 5.

Some dehydro-

fluorination might have been expected; but nucleophilic displacement of a tertiary fluorine hy acetate certainly was not. product 2 was ohtained.

None of the desired

When the reaction of 2 with glacial acetic acid

S

TIJEOXDDI

was done on a larger scale (350 mg), 113 mg of the 24-dehydro compound could be isolated in 90% purity (NMR) by recrystallization(acetone) of the solid contained after trituration (methanol)of the crude reaction mixture. Preparative-layerchromatography showed that the other products were in the mother liquor. The 25-fluoro steroid -5 was obtained in 92% yield by first rearranging 1 to 5 [16] and then fluorinating4 with DAST. Synthesis of :he target compound -14b was first attempted starting from la,25-dihydroxycholesterol, I_ 1171. Acetylation with acetic anhydride-pyridineat 50' for 24 hr gave a mixture of the monoacetate 8 and the diacetate 2 which was easily separated by preparative-layer chromatography. More vigorous acetylation conditions resulted in formation of triacetylatedcompound. Fluorfnation of 2 went smoothly to give -10 in 97% yield, which was converted, following bromination and dehydrobrominationthrough the Diels-Alder adduct 11 to the prohormone analog -12.

Tntroduction of the double bond at carbon swen into -10 was

accomplished by a modified Hunziker-MUllnerprocedure 1181 that resulted in a mixture of 4,6- and 5,7-dienes. The dienophile, 4-phenyl-1,2,4triazoline-3,5-dionereacts essentially instantly and quantitatively with 5,7-diene, but not at all with 4,6-diene 1191. As a result, production of adduct -11 is one way to simplify chromatographicseparation of the dienes. Regeneration of the 5,7-diene system by reaction of -11 with sodium bis-(2-methoxyethoxy)-aluminum hydride gave only 28% of desired product. Direct fluorination of lo,25-(OQ2D3 appeared attractive as an alternate route to 14b. The model studies discussed above showed fluorination of a

S

458

TEEOIDB

25-hydroxysteroidwith DAST to be mild and rapid. We reasoned, therefore, that this might succeed on the vitamin D compounds themselves. This approach proved fruitful as demonstrated with 25-hydroxyvitaminD3 1201 and the present report. Acetylation of X,2.5-(OR)2Djz(a)

tith acetic

anhydride and pyridine gave 1,3-protected13b. Allowing DAST to react with 13b followed by deacetylation produced a crude reation mixture which displayed a vitamin D triene system DV spectrum (X

265).

Chromatographyof the reaction products on a microparticulate silicic acid column resulted in separation of 14b from 15b. These compounds each exhibited Xmax of 265 nm and Xmin of 228 nm In their UV spectra. Likewise, their NMR spectra confirmed the presence of the triene systems, viz by the 6, 7, and 19 proton signals. Their mass spectrum supported the UV and NMR data. The compounds had a major fragment at m/e 152 resulting from schism between carbons 7 and 8 in the triene system; and a base peak of m/e 134 (m/e 152 minus water). The NMR spectrum of 14b 15h was was also diagnostic for a 25-fluoro substituent. Hormone analog actually a mixture of 24- and 25-enes in a 65:35 ratio, respectively as shown by NMR.

Both 14b and 15b were judged to be pure by the quality of

their spectra. They were also homogeneous on HFLC analysis. Biological activity. Analogs 14b and 15b were evaluated in the competitive protein binding assay along with 13a (Figure 1).

The data

for 13a were subjected to linear regression analysis in the range of a.5 to 9.0 x lO--I1M; those of 14b and 15b underwent analysis in the 1.0 to 8.0 x 10B8 Pirange. Correlation coefficients for all curves were 0,99 and curves were parallel. In this assay, a 3.7 x lO-'l M solution of L3a caused 50% displacement of radlolaheled 13a from the chick intestinal

S

459

TIIBEOSD=

% SPECIFIC

BINDING

w 0

Displacement of 1,25-(OH) -[23,24-"HID from the chick intes.~~~~c~;osol binding by @ (II),4 (o), and & (X). Values are the mean + SSM of two to four determinations. cytosol binding protein. Analogs 14b and 15b required concentrationsof about 1.3 x 10-8 M and 1.8 x 10-8 M, respectively, to cause 50% displacement of radiolabeled 13a.

In comparison, it has been reported that about 1.9 x

10-S M la-hydroxyvitaminD3 (la-OH-D3) achieves the same effect 1211. Thus, within the error limits of the competitive binding protein assay, e, 9,

and la-OH-D3 are approximately equipotent; and all are about 300-400

times less active than 13a.

S

460

60-l.)

TBEOXDb

l,25-(OH), (xl

I-OH-D,

(0)

I -OH-

-D,

25-F-D,

I

50-

I 6”M

lOoeM

45 Figure 2. Ca release from cultured fetal rat bones treated with 1,25-(OH)2D3 (O), la-OH-D3 (X), or la-OH-25-Fluoro-D3 (0). Values are the mean + SEM of four to nine bone pairs per point. The control (vehicle only) showed a release of 22% + 2. In its ability to stimulate bone resportion --*in vitro 250 times less active than 13a (Figure 2). potent than la-OH-D?.

14b was about

It was also somewhat less

The results of both assays indicate that the fluoro

substituent in 14b behaves as the 25-proton in la-OH-D3 and does not enhance potency of la-OH-D3. intestine or bone selectively.

Nor does 14b seem able to interact with These properties, along with its blocked

25-position, make 14b a potentially valuable tool for studying the

importance of the 25-hydroxyl group to the spectrum of activity exhibited by --13a in vivo.

The fluoro analog 14b may also be useful for outlining

the role of a 25-hydroxyl group in sfde chain cleavage of 13a. It is interesting that 15b was similar to 14b and la-OK-D3 in ability to displace radiolabeled 13a from the binding protein. Apparently, in the absence of a 25-hydroxyl group, alterations to the terminal side chain carbons of lo-hydroxylatedD analogs do not signfficantlyfurther disrupt binding intereactions. Like _I&, analog 15b should be helpful in elucidating the metabolic transformationsobligatory for a D analog to mediate calcium metabolism. Both 14b and 15b are now undergoing -in vivo testing. MATERIALS AND METHODS IR spectra were taken with a Perkin-Elmer 567 grating spectropbotometer. Samples were recorded as films on KaCl plates. Ultraviolet spectra were measured in 95% EtOH with a Beckman DBG. NMR spectra @DC1 as internal standard) were recorded on a Varian T-60 or Bruker w;t12;;4= spectrometer. Optical rotations were obtained in CHCl with a Perkin-Elmer 141 polarimeter. High-pressure liquid chromatographyiKPLC) was done with a Waters Associates ALC/GPC 204 liquid chromatographusing a microparticulate silica gel column (0.7 x 25 cm). Mass spectra (7Q eV) were obtained at 110-120' above ambient on an AEI MS9 coupled with a DS-50 Data System. Preparative layer chromatographywas done on a 20 x 20 cm silica gel plates with a bed thickness of 0.75 cm. Benzene, xylene, and hexane were dried over sodium. Pyridfne was distilled from BaO. DAST was prepared by the method of Middleton 1143. Crystalline 7 and 13a were gifts from Hoffmann-LaRocheCo., Nutley, N.J. Crystaliine1 was a gift from the Upjohn Co., Kalamazoo, Mich. Fluorinations. Fluorinationswere achieved by a standard procedure. The alcohol was allowed to react with an excess of DAST initially at -78' in a chlorinated hydrocarbon. The cooling bath was removed, and after 5-10 min the reaction was quenched with 5% K CO . Chloroform or ether was added; the phases separated; and the org&& phase uasbed with water, brine, and dried (Ra2S04). In tE&s manner the following compounds were obtained.

3a,5-Cyclo-25-fluoro-5a-cholest-6B_ol 6-Methyl Ether (2). To a solution of DAST (50 mg, 0.31 mmol) CH Cl (1.5 ml) initially at -78', 1 was added (100 mg, 0.24 mmol) with goo 2 s&-ring. As the mixture warme;, 1 dissolved. The residue obtained from the ether phase after the usual work-up was chromatographed on a preparative layer developed with 20% ethyl acetate-hexane to yield 2 (86 mg): NMR (270 MHz) 6 0.43, 0.65 (2 m, cyclopropyl protons), 0.72 (s, 18-CH3), 0.92 (d, J = 7.0 Hz, Zl-CH3), 1.02 (s, 19-CH ), 1.34 (d, J = 21 Hz, 26,27-methyls), 2.77+(m, 6a-H), 3.33 (s, 66-OCH 3; mass spectrum m/e (rel. intensity) 418 (M , l.OO), 403 (M+-CH 0.45), 40 a (M+-H~O, 0.14), 398 (M+-HF, 0.47), 386 (M+MeOH, 0.91 3: 383 (M+-HF-CH 0.28), 366 (M+-MeOH-HF, 0.46), 363 (M+-55, 0.96), 360 (M+-58, 0.25) 32; (M+-HF-55, 0.47), 283 (0.22), 255 (0.20), 253 (0.24). Anal. Calcd for C28H47FO: 418.3610. Found: 418.3619. 25-Fluorocholest-5-en-38-01 3-Acetate (5). To a well stirred solution of DAST (50 mg, 0.31 mmol) in CH2C12 (2 ml) initially at -78" was added 4 (100 mg, 0.23 mmol). The residue obtained from the organic phase after the usual procedure was chromatographed on a preparative layer developed with 30% ethyl acetate-hexane to yield 5 (92 mg): NMR (270 MHz) 6 0.68 (s, 18-CH3), 0.93 (d, J = 7 Hz, Zl-CH3), 1.02 (s, 19-CH ), 1.34 (d, J = 22 Hz, 26,27_methyls), 2.03 (s, acetate methyl), 4.58 trn, 3a-H), 5.38 (m, 6-H); mass spectrum m/e (rel. intensity) 446 (M+, 0.03), 386 (M+-AcOH, l.OO), 371 (M+-AcOH-CH3, O.lZ), 366 (M+-AcOH-HF, 0.22), 255 (M+AcOH-side chain, 0.12). 25-Fluorocholest-5-en-la,38_diol 1,3-Diacetate (10). To a solution of DAST (150 mg, 0.93 mmol) in trichlorofluoromethaneT5 ml) at -78" was added 2 (265 mg, 0.53 mmol) in trichlorofluoromethane (5 ml). The residue obtained after the usual procedure was chromatographed on 15 g silica gel eluted with 5-107 ethyl acetate-benzene in 30 ml stepwise pz$tions to give 25 10 (257 mg): [a] -16.6 (5 0.018, chloroform); IR, 1735 cm ; NMR (270 MHz) 6 0.67 (s, ? 8-CH ), 0.91 (d, J = 7 Hz, 21-CH3), 1.08 (s, 19-CH ), 1.33 (d, J - 22 Hz, 22,271nethyls) , 2.02 (s, 3B-acetate methyl), 2.35 (s, la-acetate methyl), 4.92 (m, 3a-H), 5.06 (m, 18-H), 5.53 (m, 6-H); mass spectrym m/e (rel. intensity) 50$
s

TIBIEOSDI

463

265 nm; NME (270 MHz) 6 0.55 (8, 18-(X3), 0.94 15b (0.28 mg, 4%): W h 6 HZ, A24 21-CH3y:Q.93 (d, J = 6 HE, 625 21-ct5), 1.68, 1.60 (d,J= (25, A24 26,27-methyl.s), 1.71 (s, A25 27-CH ), 4.45 (m, l&H), 4.24 (m, 3o-H), 5.09 (t, A24 24-H), 4.69, 4.67 (25, 325 26-protons), 5.33, 5.01 @m, 19-protons), 6.39, 6.02 (AB quartet,+J = 11 Hz, 6 ant 7-protons);mass spectrum m/e (rel. intensityi 398 (14, 0.15), 380 (M -%O, 0.09), 362 (M+ - 2 x H20, O.QQ), 362 (M - 2 x R20, O.Q6r, 152 (0.45), 134 (l.QO). Anal. Calcd for C2,H4302: 398.3184. Found: 398.3156. Further elution provided I4b (3.3 mg, 44%): UV km 265 nm; NME (270 MHz) 6 0.55 (s, 18-CH ), 0.94 (d, J - 6 Hz, 21-q. 1.34 Cd, J = 22 Hz, 26,27-methyls),4.23 (m, 3o-R), 4.43 (m, 18-H), 5.33, 5.01 (2m, 19-protons), 5.85, 6.38 (AB quartet, J - 11 Hz, 6 and 7 hydrggens); mass spectrum m/e (rel. intensity) 418 (M+ -&LO-RF, 0.18), 362 (M -2 x H20-HE, 0.18), 15nO.42), 134 (1.00:).Anal. Calcd for Cz7H44F02: 418.3247. Found: 418.3222. Attempted Solvolysis ot 3a,5-cyclo-25-fluoro-5a-cholest-6B-oi 6-Methyl Ether (2). A solution of 2 (85 mg, 0.20 mmol) in glacial acetic acid (1 ml) -was heated at 70' for 24 hy. The reaction mixture was poured over ice, and extracted with ether. The ether phase was ~~::~:~~~"~~~~~d"~~'wster, brine, and dried (Na2S0 ). The residue obtained upon evaporation of solvent was chromatograp ked on a preparative layer. Elntion with IO% ethyl acetate-hexane provided two major bands. Band 1 (Ef = 0.52) contained 28 mg of Sa-cholesta-7,24-dien-38-01 3-acetate and ficc-cholesta-7,25-diene-38-01 3-acetate, 2, in a 65:35 ratio, respectively 115,161: NMR (270 MHz) 6 0.68 (s, f8-CH3), Q.94 (d, J = 6 Hz, A24 21-CH3), 0.93 (d, J = 6 Hz, 625 27-CH3), 2.03 (38-acetatemethyl), 4.61 (3a-H), 4.68, 4.66 (2s, A25 26-protons), 5.08 (t, 624 24-H), 5.38 (m, 6-H); mass spectrum m/e (rel. intensity) 426 (M+, 0.03), 366 (M' -AcOH, l.OO), 351 (M+ -AcOH-CH3, 0.11 ), 253 (M" -AcOH-CH3, O.ll), 253 (M+ -AcOH-side &al-n 0.17). Band 2 (Rf = 0.36) contained 39 mg of So-cholest-S-ene-3$,25-dial 3,25_diacetate,5_,NMB (270 MHz) 6 Q.68 (s, 18-C ), 0.92 (d, J - 6 Hz, 21-W ), 1.02 (s, I.$-CH3),1.42 (s, 26,27-methyls H? , 1.96 (s, 25-acetate methy;c ), 2.03 (3B-acetatemethyl), 4.64 (m, 3a-H), 5,4Q (m, 6-H). The NME spectrum was identical with that of an authentic sample. Cholestan-5-ene-la,38,25-trio1 1,3-Diacetate (9): A mixture of 7 (500 mg, 1.2 mmol), acetic anhydrfde (0.3 mll, and Fyridine (0.5 ml) & benzene (5 ml) was heated at 50" for 24 hr. The reaction mtiture was cooled, diluted with 1 N HCl (50 ml), and extracted with ether (50 ml,). The organic phase was separated and washed successivelywith dilute NaHC03, water, brine, and dried (Na2S04). Evaporatfon of the solvent yielded an oil which was chromatographedon silica gel (30 g). Stepwise elution with

25 5-25% ethyl acetate-benzeneproduced 2 (390 mg): [:]D -14.4" (c 25, chloroform); IR 3520, 3450, 1735, 1025, and 760 cm ; NMR (60 MHz) 6 0.68 (s, 18-CH3), 1.09 (s, 19-CH3), 0.95 (d, J = 6 Hz, 21-CH3), 1.20 (s, 26,27methyls), 2.00 (s, la-acetate methyl), 2.03 (s, 38-acetate methyl), 4.8 (m, 3a-H), 5.0 (m, la-Hi, 5.5 (m, 6-H); mass spectrum m/e (rel. intensity) 502 (M+, absent, 442 (M -AcOH,
(s, 18-Ct5), 0.95 (a, 19-CH31r 0.96 Cd, J J 5 Hz, 21+X3), 1.34 (d, J = 22 Hz, 26,27-methyls),3.78 (m, IS-H), 4.06 On, 3a-H), 5.40, 5.75 Urn, 6 and 7 protons); mass spectrum m/e Crel. intensity) 418 CM', loo), 400 (M+ -H o, 0.51), 398 (q+ -HF, 0.40). 382 (M+ -2 x H20, 0,541, 380 (M+ -H,O-Hi, O-14), 251 (M -2 x H20-side chain, 0.75). &x&. Calcd for C27H43F02: 418.3247. Found: 418.3237. lo,25-D~hydro~~~tamin D3 1,3-Diacetate (l3b). lu,25-~OH~2D3 (9.4 m8, 0.023 mmol) was heated at 50" for 6 hr under argon tith acetic anhydride (0.05 ml) and pyridine CO.3 ml) in benzene CO.3 ml). The remtion mixture was cooled; ether and water were added; and the organic phase was separated, washed with 1 N HCl, 5% K2C03, water, and 'Errtie.TLC (silica gel, 60% ethyl acetate-hexane) indicated the presence of one compound (Rf = 0.48): W Imax 265 nm. Competitive bindiug assay fFigure 1). The cfiickintestinal cytosof protein bindmng assays were done by publQ&ed methods f22,23]. In vitro bags resorption assay (Figure 2). Fetal rat bones prelabeled with ';)Cawere cultured by previously reported procedures [24-261. ACICNOWLEDGMENl’S

This work was supported by a Program-ProjectGrant from NIH No. AM-14881, Postdocotral Training Grant from NIH No. DE-07031, and the Harry Steenbock Research Fund. We thank Mr. Sean Hehir for the 270 MHz NMR spectra, Sir. Melvin Hicke for several mass spectra, and Dr. John Chu for his assistance in the preparation of the DAST reagent. REFEEENCES 1. 2.

DeLuea, 8. F., PED. PROC. 33_,2211 (1974). Kodicek, E., LANCET 1, 325 (1974). DeLuca, H. F., end Schnoea, H. K., ANN. REV, BIOCHEM. 45, 631 fl976). Lam, H.-Y., Onisko, B. L., S&noes, Ii.K-, and DeLuca, B. F. BIOCHEM. BIOPHYS. RES. CGMMIJN.2, 845 (1974f. 5, Okamura, W. H., Mftra, M. N., Proscaf, D. A., and Norman, A. W. BIOCHEM. BIOPHYS. RES. COMMUN. 2, 24 (1975). 6. Onisko, 33.L., Lam, H.-Y., Reeve, L. E., S&noes, If,K., and DeLuca, B. I?., BIOORGANIC CHEM. 2, 203 (1977). 7, Pavlovitch, H., Garabedian, M,, and Balsan, S., J. CLIN. INFEST. 52, 2656 (1973). 8. Tanaka, Y., Castillo, L., and DeLuca, H. F., J. BILL. CHEM. 252, 1421 (1977). 9. Harnden, D-, Kumar, R., Holick, M. F., and DeLuca, Ii.F., SCIENCE f93, 493 (1976). 10. Kumar, R., Harnden, D., and DeLuca, IL.F., BIOCHFMISIRY 15, 2420 (1976). 11. Peters, R. A., ADV. ENZYMOL. 3, 113 (1957).

2:

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

Heidelberger, C., Griesbach, L., Montag, B. J., Mooren, D., and Cruz, O., CANCER RES. 18, 305 (1958). Fried, J., and Box-man, A., VITAMINS AND HORMONES 16, 303 (1958). Middleton, W. J., J. ORG. CHEM. 60, 574 (1975). Dasgupta, S. K., Crump, D. R., and Gut, M., J. ORG. CHEM. 39, 1658 (1974). Moreau, J. P., Aberhart, D. J., and Caspi, E., J. ORG. CHEM. 2, 2018 (1974). Partridge, J. J., Faber, S., and Uskokovic, M. R., HELV. CHIM. ACTA 57_, 764 (1974). Hunziker, F., and Mullner, F. X., HELV. CHIM. ACTA 2, 70 (1958). Barton, D. H. R., Shioiri, T., and Widdowson, D. A., .T. CHEM. SOC. (C) 1971 (1968). Onisko, B. L., Schnoes, H. K., and DeLuca, H. F., TETRAHEDRON LETTERS (No. 13) 1107 (1977). Eisman, J. A., and DeLuca, H. F., STEROIDS 30, 245 (1977). Eisman, J. A., Hamstra, A. H., Kream, B. E., and DeLuca, H. F., SCIENCE 193, 1021 (1976). Eisman, J. A., Hamstra, A. H., Kream, B. E., and DeLuca, H. F. ARCH. BIOCHEM. BIOPHYS. 176, 235 (1976). Stern, P. H., Trummel, C. L., S&noes, H. K., and DeLuca, H. F., ENDOCRINOLOGY 9;1, 1552 (1975). Raisz, L. G., J. CLIN. INVEST. 64, 103 (1965). Raisz, L. G., Trummel, C. L., Holick, M. F., and DeLuca, H. F., SCIENCE 175, 768 (1972).