Synthesis and biological evaluations of C-23-modified 26,26,26,27,27,27-F6-vitamin D3 analogues

Article abstract of DOI:10.1016/S0968-0896(00)00106-1

A convenient synthetic method which could allow flexible modification at C-23 of 26,26,26,27,27,27-hexafluoro-lα,25-dihydroxyvitamin D₃ (3) has been developed. An effective construction of hexafluoroacetone (HFA) aldol part on the side chain of 10 was achieved by aldol reaction with HFA gas. This route is also attractive as an approach to diverse 26,27-modified vitamin D₃ analogues. The preliminary biological activities of 23-modifed 26,27-F₆ vitamin D₃ analogues are evaluated. The potency of VDR affinities of the C-23-substituted analogues (keto group (4); OH group (5a,5b); fluorine atom (6a,6b); and oxetane ring (7a,7b)) was found to vary depending upon both the nature and stereochemistry of the substituents. In contrast, the HL-60 cell differentiation property was less varied than VDR affinity, and depended upon the nature rather than the stereochemistry of the substituents. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00106-1

Bioorganic & Medicinal Chemistry 8 (2000) 1809±1817
Synthesis and Biological Evaluations of C-23-Modi®ed
26,26,26,27,27,27-F6-Vitamin D3 Analogues
Masahiko Ikeda,a,* Haruki Matsumura,b Nobuyuki Sawada,a Katsuhiro Hashimoto,a
Tomoyuki Tanaka,a Toshihiro Noguchia and Masaji Hayashia
aSumitomo Pharmaceuticals Co., Ltd, 3-1-98, Kasugadenaka, Konohana-ku, Osaka 554-0022, Japan
bSumitomo Chemical Co., Ltd, 4-5-33, Kitahama, Chuo-ku, Osaka 541-0041, Japan
Received 3 February 2000; accepted 24 March 2000
AbstractÐA convenient synthetic method which could allow ¯exible modi®cation at C-23 of 26,26,26,27,27,27-hexa¯uoro-la,25-
dihydroxyvitamin D3 (3) has been developed. An e€ective construction of hexa¯uoroacetone (HFA) aldol part on the side chain of
10 was achieved by aldol reaction with HFA gas. This route is also attractive as an approach to diverse 26,27-modi®ed vitamin D3
analogues. The preliminary biological activities of 23-modifed 26,27-F6 vitamin D3 analogues are evaluated. The potency of VDR
anities of the C-23-substituted analogues (keto group (4); OH group (5a,5b); ¯uorine atom (6a,6b); and oxetane ring (7a,7b)) was
found to vary depending upon both the nature and stereochemistry of the substituents. In contrast, the HL-60 cell di€erentiation
property was less varied than VDR anity, and depended upon the nature rather than the stereochemistry of the substituents.
# 2000 Elsevier Science Ltd. All rights reserved.
Introduction
OH group (5a,5b); (iii) ¯uorine atom (6a,6b); and (iv)
oxetane ring (7a,7b). In particular, modi®cations includ-
ing (iii) and (iv) at C-23 in vitamin D3 have not been
attempted previously. This report concerns a novel syn-
thetic method which enables ¯exible modi®cations at
C-23 of 3 via the key-intermediate 10 and the biological
activities of the resulting analogues.
1a,25-Dihydroxyvitamin D3 (1; 1a,25(OH)2D3), the horm-
onally active form of vitamin D3, plays an important
role in regulating cell proliferation and di€erentiation
in a variety of cell types.1 A large number of side
chain modi®ed analogues have been reported during the
last decade.2 Among them, 26,26,26,27,27,27-hexa¯uoro-
1a,25-dihydroxyvitamin D3 (3; 26,27-F6-1a,25(OH)2D3)
exhibits a characteristic potency with some biological
activities to 1.3 5 Our laboratory has already reported
that one analogue of 3, 26,27-F6-1a,23(S),25(OH)3D3
(5a) showed a binding anity similar to 3, to the vita-
min D3 receptor (VDR),6 whereas 1a,23(S),25(OH)3D3
(2) had a lower binding anity than 1 (2: 12% compared
to 1).2 Data suggested that modi®cation at C-23 of 3
would produce an unexpected biological e€ect, com-
pared to 1. Therefore we planned to prepare C-23-mod-
i®ed analogues of 3 and evaluate their biological
activities. At present, no convenient synthetic methods to
introduce the substituents at C-23 of 3 have been report-
ed, so we worked to develop a novel synthetic method.
Our intended modi®cations at C-23 can be categorized
by introducing the following groups: (i) ketone (4); (ii)
Results and Discussion
Synthesis
As shown in Scheme 1, the synthetic plan relied upon
construction of a side chain of the key-intermediate 10
using an aldol reaction of 13 with hexa¯uoroacetone
(HFA). Tosylate 14, which was prepared from readily
available vitamin D2 in six steps,7,8 could be transformed
into methyl ketone 13. In practice, methyl ketone 13 was
prepared as described in Scheme 2. Displacement of the
tosyl group in 14 with KCN and subsequent reduction
by diisobutylalminium hydride (DIBALH) a€orded
aldehyde 16. The reaction of 16 with methylmagnesium
bromide followed by oxidation gave the desired methyl
ketone 17. Following the 1a-hydroxylation procedure
developed by Andrews et al.,8,9 18 was obtained in 41%
yield. Silylation of 18 and successive photoisomerization
gave methyl ketone 13. As previously reported in the
literature, HFA aldols can be prepared from enol silyl
*Corresponding author at mailing address; Exploratory Research
Group, Sumitomo Pharmaceuticals Research Center 2-1, Takatsukasa
4-chome, Takarazuka Hyogo 665-0051, Japan. Tel.: +81-797-74-
2067; fax: +81-797-74-2142; e-mail: ikeda@sumitomopharm.co.jp
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00106-1

1810
M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
Figure 1.
Scheme 1.
ethers and HFA in the presence of SnCl4,10 and O-sily-
lated HFA aldols can be obtained in the absence of
SnCl4.11 To prevent the risk of degrading the labile
vitamin D triene unit by Lewis acids, and to simplify the
reaction procedure from the methyl ketone to the HFA
aldol, we examined basic conditions using lithium bases;
Lithiation of 13 with LDA (1.1±3.0 equiv) in THF at
70 ꢀC followed by the addition of an excess amount of
HFA gas gave the undesired 24-dehydrate compound 20
(15%), along with recovery of starting material 13
(80%).12 Switching the base to lithium bis(trimethylsilyl)
amide [LiN(TMS)2] (1.4 equiv)13 gave the desired HFA
aldol adduct 10 (82%) as a major product with a small
amount of recovered 13 (4%). In this case, 20 was not
obtained.
Scheme 2. (i) KCN, DMSO, 90 ꢀC, 30 min; (ii) DIBALH, THF,
70 ꢀC!rt; (iii) MeMgBr, Et2O, rt, 10 min; (iv) TPAP, NMO, MS4A,
CH2Cl2, rt, 15 min; (v) SeO2, NMO, MeOH-CH2Cl2, re¯ux, 2 h; (vi)
TBSCl, imidazole, DMF, rt, 3 h; (vii) Hg lamp, Anthracene, NEt3, rt,
1 h; (viii) CF3COCF3, LiN(TMS)2, THF, 70 ꢀC, 30 min *Considering
the recovery of 13.
One of the desired analogues, 26,27-F6-1a,25(OH)2-23-
oxo-D3 (4) was obtained from 10 by removing protecting
groups. Reduction of 10 with NaBH4 and separation of
the resulting two isomers by column chromatography
readily a€orded the more polar isomer 21a and the less
polar isomer 21b. Removal of the protecting groups of
21a and 21b gave 26,27-F6-1a,23(S),25(OH)3D3 (5a) and
26,27-F6-1a,23(R),25(OH)3D3 (5b) respectively. The

M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
1811
stereochemical assignment at C-23 of 5a was con®rmed
Biological evaluations
by X-ray crystallographic analysis of monohydrate
form of 5a. The assignment at C-23 of 5b could therefore
be informed from this data. Introduction of ¯uorine at
C-23 of 21b was performed by treatment with diethyla-
minosulfur tri¯uoride (DAST)14 via SN2 displacement. In
this step, the formation of signi®cant amount of undesired
elimination product 22 was observed by 1H NMR analy-
sis of crude product (the molar ratio of 23S-F derivative
and 22 (2:1)).15 The following deprotection and removal
of the elimination product gave 6b. The epimer 6a was
derived from 21a by the same procedure. For the synthesis
of the respective oxetane analogues, mesylation of 21a
and treatment with KOH in MeOH followed by depro-
tection using tetrabutylammonium ¯uoride (TBAF)
gave the corresponding oxetane 7a. The epimer 7b was
also taken through the same procedure, from 21b.
Biological evaluations of the 26,27-F6 analogues descri-
bed above, together with 3 (for comparison) are sum-
marized in Table 1. The binding anity to VDR was
determined by a competitive binding assay using the
cytosolic receptor from chick intestine.18 The activity in
di€erentiating HL-60 cells was examined by nitro blue
tetrazolium (NBT) reduction.19 The potency of VDR
anities of the C-23-substituted analogues (5a, 5b, 6a,
6b) was found to vary depending upon both the nature
and stereochemistry of the substituents (30±1040% of 3).
Remarkably, S-isomers (5a,6b) were more potent than
R-isomers (5b,6a), especially when ¯uorine was sub-
stituted. In contrast, the HL-60 cell di€erentiation
property was less varied than VDR anity (47±200% of
3), and this depended upon the nature rather than the
This method via an aldol reaction is also attractive as an
approach to new types of 26,27-modi®ed analogues for
biological screening purposes. In a preliminary study, we
have already synthesized two compounds (816 and 9) using
acetone and dimethoxyacetone17 with satisfactory yields
(Schemes 1 and 4). Introduction of two extra methoxy
groups at C-26 and C-27 has not been previously reported.
Scheme 4. (i) LiN(TMS)2, THF, 78 ꢀC, (ii) MsOH, MeOH, rt, 1 h
*Considering the recovery of 13.
Scheme 3. (i) TsOH, MeOH, rt, 2 h; (ii) NaBH4, THF-MeOH, 10 ꢀC, 10 min; (iii) MsOH, MeOH, rt, 1 h; (iv) DAST, CH2Cl2, 50 ꢀC!rt, 1 h; (v)
TBAF-THF, rt, overnight; (vi) MsCl, pyridine, 0 ꢀC, 1 h; (vii) 5% KOH-MeOH; rt, 10 min. *Considering the recovery of starting material.

1812
M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
Table 1. Comparison of biological activities of 23-modi®ed 26,27-F6
analogues and 3a
Table 2. Conformational distribution for C-23-modi®ed vitamin D3
analogues (2000 iterations)
Compound
VDR
HL-60
Compound
Number of conformers
ꢁ50 kJ/mola
ꢁ10 kJ/mola
3
4 [23-keto]
(100)
130
100b
30
160
1040
30
(100)
Ðc
47
3
4 [23-keto]
815
262
480
534
700
436
44
64
12
4
13
27
9
5a [23(S)-OH]
5b [23(R)-OH]
6a [23(R)-F]
6b [23(S)-F]
7a [23(R)-oxetane]
7b [23(S)-oxetane]
54
5a [23(S)-OH]
5b [23(R)-OH]
6a [23(R)-F]
6b [23(S)-F]
7a [23(R)-oxetane]
7b [23(S)-oxetane]
200
140
54
7
64
12
5
aThe results for 3 are normalized to 100.
bThe data was already reported in ref 6.
c4 was 40-fold more active than 1. Comparison of HL-60 cell di€er-
entiating activity between 1 and 3 was described in ref 4 (3: 10-fold
compared to 1). The data suggested 4 is more active than 3.
29
aRelated energy above global minimum energy conformer.
Table 3. Conformers of 23-modi®ed analogues at low energy level
(0±6 kJ/mol)a
stereochemistry of the substituents. For the oxetane
compounds, the S-isomer (7b) leads to a drop in VDR
anity (7% of 3) without signi®cantly decreased HL-60
cell di€erentiating property, whereas the R-isomer (7a)
has decreased binding to both VDR and HL-60 cell
di€erentiation. As described in the literature, introduc-
tion of S/R-OH, keto group at C-23 of 1 gave signi®cant
decreased binding anity to VDR compared to 1.2,16
On the other hand, among 26,27-F6 analogues, S-OH
and keto-modi®cations exhibited a similar and more
potent binding anity than 3 respectively. And S-OH
modi®cation showed decreased binding anity, which
was more moderate than that of 1.
Compound
Conformation type
ÁE (kJ/mol)b
Mol%
4 [23-keto]
A
B
C
B
C
C
A
A
B
C
B
A
0
57
9
7
6
5
86
67
60
20
32
27
21
3.1
3.8
4.2
4.5
0
0
0
2.8
0
0.5
1.1
5a [23(S)-OH]
5b [23(R)-OH]
6a [23(R)-F]
6b [23(S)-F]
aHigh proportioned conformers (>5 mol% (Boltzmann distribution at
298 K)).
bRelative energy above global minimum energy conformer.
Conformational analysis
To comprehend the modi®cational e€ects of the sub-
stituents at C-23 of 3 on VDR anity, we focused on the
side chain conformers possessing the biologically impor-
tant functional groups (25-OH group). Conformational
searches of side chains of 3 and C-23-modi®ed analogues
were subjected to a Monte Carlo simulation (2000
iterations) using the MM3 force ®eld20 and the GB/SA
Solvation Model (in a water matrix)21 including Mac-
roModel (v4.5).22 It was carried out on model com-
pounds in which the A ring and diene system up to C-5
were replaced by a H atom. Rotations were applied to
the rotatable C C bonds of the side chain and the 23-
OH bond (in the case of 5a and 5b). All conformers
within 50 kJ/mol of the global minimum were saved and
ranked by relative energy. The data suggested that the
side chains of all C-23-modi®ed compounds clearly have
fewer degrees of freedom than that of 3 (Table 2).
Compounds (4, 5a, 5b, 6a, 6b) possess three families of
conformers at low energy levels (0±6 kJ/mol)(Figure 2
and Table 3)23 (The torsion angle at C17±C20±C22±C23,
and C20±C22±C23±C24 anti-anti (conformer A; black),
gauche (+)-anti (conformer B; white), anti-gauche(+)
(conformer C; gray)). The stereochemistry of the sub-
stituents at C-23 of 3 in¯uenced the conformational
behavior of the side chains signi®cantly. First, with the
respect to hydroxy compounds, the R-isomer (5b) and
S-isomer (5a) adopt the characteristic conformers A and
C respectively in high population at a low energy level
(0±6 kJ/mol). Next, with the respect to the ¯uorinated
analogues (6a,6b), R- and S-isomers have higher con-
Figure 2. Superposition of three families of conformers of 5a, 5b, 6a
and 6b at 0±6 kJ/mol (stereoview) as their deprotonated and de¯uori-
nated forms. The substituents at C-23 were omitted to see the orien-
tation of side chains clearly.
formational ¯exibility than hydroxy compounds, because
of the reduced steric repulsion between the 23-substituents
and the 18-methyl group. At an energy level (0±6kJ/mol),
conformer C was adopted by the S-isomer (6b), not by the
R-isomer (6a). In the case of 23-keto compound (4) the
side chain adopted more expanded mobility than those of
the other compounds (5a,5b,6a,6b), because of reduced
steric repulsion between carbonyl group and the 18-
methyl group. As the result, in the case of hydroxy and
¯uorinated analogues, one explanation was provided by
conformational analysis and biological data. Conformer

M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
1813
signal as an internal standard: d 7.26. Dichloromethane,
tetrahydrofuran (THF), and diethyl ether (Et2O) was
dried over molecular sieves 4A (MS-4A). The purity of
products (4, 5a, 5b, 6a, 6b, 7a, 7b) was judged to be at
least 98% by HPLC analysis (UV, 265 nm). ( )* The
yields were considered with the recovery of starting
materials.
3ꢀ-[(t-Butyldimethylsilyl)oxy]-20(S)-(cyanomethyl)-9,10-
secopregna-5(E),7(E),10(19)-triene (15). To a suspension
of 14 (24.5 g, 40.9 mmol) in DMSO (250 mL) was added
potassium cyanide (13.0 g, 265 mmol) and heated for
30 mm at 90 ꢀC. The reaction mixture was cooled in an
ice bath, and then diluted with water, and extracted with
ethyl acetate. The organic layer was washed with water,
brine, and dried over MgSO4. The ®ltrate was con-
centrated in vacuo. The residue was puri®ed by silica gel
chromatography (EtOAc:hexane, 1:10) to give 15 (13.0 g,
Figure 3. Superposition of four families of conformers of two oxetane
compound (7a; left, 7b; right) at 0±8 kJ/mol as their deprotonated and
de¯uorinated forms.
C adopted by S-isomers (5a,6b) might be more appro-
priate to keep the size and lipophilicity of the molecule
favorable to ®t into the VDR cavity than R-isomers
(5b,6a), especially when ¯uorine is substituted. On the
other hand, oxetane analogues (7a,7b) adopt four famil-
ies of conformers within 8 kJ/mol from the lowest energy
conformer.23 As shown in Figure 3, the four conformers
of 7a were nearly superimposable on those of 7b, except
for the positions of the oxygen atom in the oxetane ring.
The positional change of the oxygen atom in the oxetane
ring may explain the di€erence of binding to the VDR.
70%) as an amorphous solid. IR (neat): 2932, 2858, 2247,
1
1467, 1254, 1098, 836 cm
.
1H NMR (270 MHz) d:
0.06±0.07 (6H, 2s), 0.58 (3H, s), 0.88 (9H, s), 1.18 (1H,
d, J=6.5 Hz), 3.85 (1H, m), 4.64 (1H, br s), 4.92 (1H, br
s), 5.85 (1H, d, J=11.5 Hz), 6.46 (1H, d, J=11.5 Hz).
3ꢀ-[(t-Butyldimethylsilyl)oxy]-20(S)-(formylmethyl)-9,10-
secopregna-5(E),7(E),10(19)-triene (16). Diisobutylalu-
minium hydride (DIBALH, 0.93 M in hexane, 2.27 mL,
2.11 mmol) was added to a solution of 15 (640 mg,
1.41 mmol) in THF (12 mL) at 70 ꢀC. After warming
to room temperature, the reaction mixture was stirred
for 30 min. When the starting material was detected by
TLC, DIBALH (0.93 M in hexane solution, 2.27 mL,
2.11 mmol) was added to the reaction mixture at 70 ꢀC.
After the starting material was consumed completely,
5% aqueous HCl was added. The mixture was extracted
with ethyl acetate. The organic layer was washed with
5% aqueous HCl, saturated NaHCO3, brine, and dried
over MgSO4. The ®ltrate was concentrated in vacuo. The
residue was puri®ed by silica gel chromatography
(EtOAc:hexane, 1:10) to give 16 (350 mg, 54%) as an
amorphous solid. IR (neat): 2929, 2856, 1728 1471, 1251,
Conclusion
We developed a convenient synthetic method which
enables ¯exible modi®cation at C-23 of 3. An e€ective
construction of the side chain of 10 was achieved by aldol
reaction with HFA. As the result of preliminary biologi-
cal evaluations of 26,27-F6 analogues, with respect to
VDR anities, the modi®cation at C-23 of 3 in¯uenced
biological activities signi®cantly. It depends on both the
nature and stereochemistry of the substituents. In con-
trast, the HL-60 cell di€erentiation property was less
varied than VDR anity. It was also found that R/S-
OH and keto-modi®cational e€ects of 3 to VDR anity
provided a di€erent biological pro®le compared to 1.
1
1098, 836 cm 1. H NMR (270 MHz) d: 0.06±0.07 (6H,
2s), 0.61 (3H, s), 0.88 (9H, s), 1.03 (1H, d, J=6.5 Hz),
3.86 (1H, m), 4.64 (1H, br s), 4.93 (1H, br s), 5.86 (1H, d,
J=11.5 Hz), 6.49 (1H, d, J=11.5 Hz), 9.77 (1H, d,
J=2.5 Hz).
Experimental
All reactions involving oxygen- or moisture-sensitive
compounds were carried out under a dry nitrogen atmo-
sphere. Reaction temperatures refer to external bath
temperatures. Reactions were monitored by thin layer
chromatography (TLC) using Merck TLC plates (silica
gel 60 F254, 0.25 mm thickness). After ultraviolet illumi-
nation at 254 nm, the plates were visualized by immer-
sion in a solution of phosphomolybdic acid in MeOH
followed by heating. Column chromatography puri®ca-
tion was performed with Merck silica gel (silica gel 60,
70-230 mesh), and PLC puri®cation was performed with
Merck PLC plates (silica gel 60 F254, 1 mm or 0.5 mm
thickness). 1H NMR data were measured in CDCl3.
Chemical shifts are reported as d units (ppm) down®eld
from tetramethylsilane (d 0.00) using residual solvent
3ꢀ-[(t-Butyldimethylsilyl)oxy]-20(S)-(propyl-23-oxo)-9,10-
secopregna-5(E),7(E),10(19)-triene (17). Methylmagne-
sium bromide (0.99 M in Et2O, 1.04 mL, 1.03 mmol)
was added to a solution of 16 (350 mg, 0.766 mmol) in
Et2O (7 mL) at room temperature. After stirring for
10 min, 10% aqueous NH4Cl was added, and the mix-
ture was extracted with ethyl acetate. The organic layer
was washed with brine, and dried over MgSO4. The ®l-
trate was concentrated in vacuo. The residue was pur-
i®ed by silica gel chromatography (EtOAc:hexane, 1:5)
to give the corresponding alcohol as an amorphous solid
(195mg, 54%). IR (neat): 3370, 2930, 2856, 1471, 1253,
1
1098, 836 cm 1. H NMR (270MHz) d: 0.06±0.07 (6H,
2s), 0.57 and 0.59 (total 3H, s), 0.88 (9H, s), 1.60 (3H, s),

1814
M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
3.86 (2H, m), 4.64 (1H, br s), 4.92 (1H, br s), 5.85 (1H, d,
J=11.5Hz), 6.47 (1H, d, J=11.5 Hz). Tetrapropyl-
ammonium perruthenate (TPAP, 26 mg, 0.074 mmol)
was added in one portion to a stirred mixture of the
alcohol (175 mg, 0.37mmol), 4-methylmorphorine N-
oxide (NMO, 65mg, 0.55mmol) and powdered 4 A
molecular sieves (125mg) in dichloromethane (4 mL) at
room temperature. After 15min, the reaction mixture was
passed through silica gel chromatography (EtOAc:hex-
ane, 1:10) to give 17 (136 mg, 78%) as an amorphous
solid. IR (neat): 2949, 2856, 1717 1470, 1252, 1098,
puri®cation (1 mm, EtOAc:hexane, 1:5), 13 (308 mg,
72%) as an amorphous solid. IR (neat): 2927, 2857,
1718, 1472, 1256, 1075, 829cm 1. 1H NMR (270 MHz) d:
0.05±0.06 (12H, 2s), 0.57 (3H, s), 0.88 (18H, s), 0.93 (3H, d,
J=6.5 Hz), 2.12 (3H, s), 4.18 (1H, m), 4.36 (1H, m), 4.86
(1H, d, J=2.5 Hz), 5.17 (1H, br s), 6.01 (1H, d,
J=11.0Hz), 6.23 (1H, d, J=11.0 Hz). MS(EI) m/z: 600
(M+), 585 (M+-CH3), 543 (M+-tBu), 486 (M+-2tBu), 468
(M+-2tBu-H2O), 453 (M+-2tBu-H2O-CH3). HRMS(EI)
m/z: 600.4421 (M+) (calcd for C36H64O3Si2 600.4394).
1
836 cm 1. H NMR (300 MHz) d: 0.05±0.07 (6H, 2s),
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-26,26,26,27,27,27-
hexa¯uoro-23-oxo-9,10-secocholesta-5(Z),7(E),10(19)-
trien-25-ol (10). n-Butyllithium (1.64 M solution hexane,
9.1 mL, 14.9 mmol) was added dropwise to a solution of
1,1,1,3,3,3-hexamethyldisilazane (3.33 mL, 15.8 mmol) in
THF (30 mL) at 70 ꢀC. The resulting mixture was
stirred for 10 min. A solution of 13 (6.30 g, 10.5mmol) in
THF (35mL) was added to the reaction mixture. After
stirring for 10min, hexa¯uoroacetone gas was blown
through the mixture for 30min. The reaction mixture was
quenched with 10% aqueous NH4Cl and extracted with
ethyl acetate. The organic layer was washed with 1N HCl,
saturated aqueous NaHCO3, brine, and dried over
MgSO4. The ®ltrate was concentrated in vacuo. The resi-
due was puri®ed by silica gel chromatography (EtOAc:
hexane, 1:30) to give 10 (6.56 g 82%, (85%)*) as an
0.59 (3H, s), 0.88 (9H, s), 0.94 (3H, d, J=6.5 Hz), 2.12
(3H, s), 3.85 (1H, m), 4.64 (1H, br s), 4.92 (1H, br s), 5.85
(1H, d, J=11.5 Hz), 6.46 (1H, d, J=11.5 Hz).
1ꢁ-Hydroxy-3ꢀ-[(t-butyldimethylsilyl)oxy]-20(S)-(n-pro-
pyl-23-oxo)-9,10-secopregna-5(E),7(E),10(19)-triene (18).
NMO (151 mg, 1.29 mmol) was added to a solution of
17 (117 mg, 0.249 mmol) in dichloromethane (1.5 mL),
and then a solution of selenium dioxide (SeO2, 28 mg,
0.252 mmol) in methanol (1.5 mL) was added, and the
mixture was heated under re¯ux for 2 h. After cooling to
room temperature, water was added and extracted with
dichloromethane. The organic layer was washed with
saturated aqueous NaHCO3, brine, and dried over
MgSO4. The ®ltrate was concentrated in vacuo. The resi-
due was puri®ed by silica gel chromatography (EtOAc:
hexane, 1:10) and the further HPLC puri®cation (Zorbax
BP SIL, 10 mmÂ25 cm, EtOAc:hexane, 1:10) to give 1a-
hydoxy compound 18 (50 mg, 41%) as an amorphous
solid. IR (neat): 3436, 2950, 2856, 1713, 1472, 1253, 1078,
836 cm 1. 1H NMR (270 MHz) d: 0.07 (6H, s), 0.59 (3H,
s), 0.88 (9H, s), 0.94 (3H, d, J=6.5 Hz), 2.13 (3H, s),
4.20 (1H, m), 4.50 (1H, br s), 4.94 (1H, br s), 5.06 (1H,
br s), 5.85 (1H, d, J=11.5 Hz), 6.50 (1H, d, J=11.5 Hz).
amorphous solid with the recovery of 13 (0.25 g, 4%).
1
IR (neat): 3306, 2929, 1709, 1462, 1258, 1089, 834 cm
.
1H NMR (270MHz) d: 0.04±0.06 (12H, 2s), 0.57 (3H, s),
0.88 (18H, s), 0.96 (3H, d, J=6.2 Hz), 2.88 (2H, br s), 4.20
(1H, m), 4.37 (1H, m), 4.85 (1H, d, J=2.3 Hz), 5.18 (1H,
br s), 6.01 (1H, d, J=11.0 Hz), 6.23 (1H, d, J=11.0 Hz),
6.88 (1H, br s). MS(EI) m/z: 766 (M+), 751 (M+-CH3),
709 (M+-tBu), 697 (M+-CF3), 652 (M+-2tBu), 634 (M+-
2tBu-H2O), 619 (M+-2tBu-H2O-CH3). HRMS(EI) m/z:
766.4299 (M+) (calcd for C39H64O4F6Si2 766.4247).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-20(S)-(n-propyl-23-
oxo) - 9,10 - secopregna - 5(E),7(E),10(19) - triene (19). A
solution of 18 (50 mg, 0.103 mmol) in DMF (0.5 mL)
containing imidazole (20 mg, 0.294 mmol) and t-butyldi-
methylsilyl chloride (TBSCl, 25mg, 0.166 mmol) was stir-
red at room temperature for 3 h. The reaction mixture was
quenched with water and extracted with ethyl acetate. The
organic layer was washed with 1N HCl, saturated aqu-
eous NaHCO3, brine, and dried over MgSO4. The ®ltrate
was concentrated in vacuo. The residue was puri®ed by
silica gel chromatography (EtOAc:hexane, 1:10) to give 19
(55 mg, 89%) as an amorphous solid. IR (neat): 2952,
26,26,26,27,27,27-Hexa¯uoro-23-oxo-9,10-secocholesta-
5(Z),7(E),10(19)-trien-1ꢁ,3ꢀ,25-triol (4). p-Toluene-
.
sulfonic acid monohydrate (TsOH H2O, 35 mg) was
added to a solution of 10 (185 mg, 0.241 mmol) in
methanol (2.5 mL). After stirring for 2 h at room tem-
perature, the reaction mixture was quenched with water
and extracted with ethyl acetate. The organic layer was
washed with saturated aqueous NaHCO3, brine, dried
over MgSO4 and concentrated in vacuo. The residue was
puri®ed by silica gel chromatography (EtOAc:hexane,
1:1) to give 4 (108 mg, 83%) as an amorphous powder.
1
1
2857, 1718, 1470, 1253, 1083, 835 cm 1. H NMR (270
UV (EtOH): lmax 264 nm. H NMR (270 MHz) d: 0.59
MHz) d: 0.05±0.06 (12H, 2s), 0.58 (3H, s), 2.13 (3H, s),
4.22 (1H, m), 4.53 (1H, m), 4.94 (1H, br s), 4.98 (1H, br s),
5.82 (1H, d, J=11.5 Hz), 6.45 (1H, d, J=11.5 Hz).
MS(EI) m/z: 600 (M+), 585 (M+-CH3), 543 (M+-tBu),
486 (M+-2tBu), 468 (M+-2tBu-H2O), 453 (M+-2tBu-
H2O-CH3). HRMS(EI) m/z: 600.4409 (M+) (calcd for
C36H64O3Si2 600.4394).
(3H, s), 0.97 (3H, d, J=6.0 Hz), 2.88 (2H, br s), 4.24
(1H, m), 4.43 (1H, m), 4.99 (1H, br s), 5.33 (1H, br s),
6.01 (1H, d, J=11.5 Hz), 6.37 (1H, d, J=11.5 Hz), 6.86
(1H, s). MS(EI) m/z: 538 (M+), 520 (M+-H2O), 502
(M+-2H2O). HRMS(EI) m/z: 538.2555 (M+) (calcd for
C27H36O4F6 538.2518).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxyl]-26,26,26,27,27,27-
hexa¯uoro-9,10-secocholesta-5(Z),7(E),10(19)-trien-23-
(S),25- diol (21a) and 1ꢁ,3ꢀ -bis[(t- butyldimethylsilyl)-
oxy] - 26,26,26,27,27,27 - hexa¯uoro - 9,10 - secocholesta -
5(Z),7(E),10(19)-trien-23(R),25-diol (21b). Sodium bor-
ohydride (514 mg, 13.6 mmol) was added to a solution
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-20(S)-(n-propyl-23-
oxo)-9,10-secopregna-5(Z),7(E),10(19)-triene (13). A
solution of 19 (430 mg, 0.715 mmol) in toluene (200 mL)
containing triethylamine (3 drop), anthracene (400 mg,
2.24 mmol) was irradiated for 1 h to give, after PLC

M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
1815
of 10 (7.0 g, 9.13 mmol) in THF (60 mL) and methanol
(60 mL) at 10 ꢀC. After stirring for 10 min, the reaction
mixture was quenched with 1N HCl and extracted with
ethyl acetate. The organic layer was washed with satu-
rated aqueous NaHCO3, brine, and dried over MgSO4.
The ®ltrate was concentrated in vacuo. The residue was
separated by silica gel chromatography (EtOAc:hexane,
1:20) to give 21a [2.27 g, 32%; Rf=0.25 (EtOAc:hexane,
1:10)] as a white powder and 21b [3.65 g, 52%, Rf=0.5
(EtOAc:hexane, 1:10)] as a white powder.
CH3), 504 (M+-2H2O), 489 (M+-2H2O-CH3). HRMS(EI)
m/z: 540.2618 (M+) (calcd for C27H38O4F6 540.2674).
23(R),26,26,26,27,27,27-Hepta¯uoro-9,10-secocholesta-
5(Z),7(E),10(19)-trien-1ꢁ,3ꢀ,25-triol (6a). DAST (35 mL,
0.26mmol) was added to a solution of 21a (50 mg,
0.065 mmol) in dichloromethane (1.0 mL) at 50ꢀC. The
mixture was stirred for 10 min at the same temperature,
warmed to room temperature and stirred for 1 h. The
reaction mixture was quenched with water and extrac-
ted with ethyl acetate. The organic layer was washed
with brine, and dried over MgSO4. The ®ltrate was
concentrated in vacuo. The residue was puri®ed by silica
gel chromatography (EtOAc:hexane, 1:10) to give crude
mixture (30 mg) with the recovery of 21a (10 mg, 20%).
To a solution of the crude mixture (30 mg) in THF
(1.0 mL) was added tetrabutylammonium ¯uoride
(TBAF, 1M in THF, 0.5 mL) at room temperature and
stirred overnight. The reaction mixture was quenched
with water and extracted with ethyl acetate. The organic
layer was washed with brine and dried over MgSO4. The
®ltrate was concentrated in vacuo. The residue was pur-
i®ed by silica gel chromatography (ethanol:hexane, 1:5)
and the further HPLC puri®cation (Zorbax BP SIL,
10 mmÂ25 cm, ethanol:hexane, 1:10) to give 6a (15 mg,
42% (53%)*) as an amorphous solid. UV (EtOH): lmax
21a: IR (neat): 3308, 2953, 2857, 1472, 1207, 1075,
835 cm 1. 1H NMR (270MHz) d: 0.06 (12H, s), 0.56 (3H,
s), 0.88 (18H, s), 0.98 (3H, d, J=5.9Hz), 4.20 (1H, m),
4.38 (2H, m), 4.86 (1H, d, J=2.3Hz), 5.18 (1H, br s), 6.02
(1H, d, J=11.0 Hz), 6.23 (1H, d, J=11.0 Hz), 6.31 (1H, br
s). MS(EI) m/z: 768 (M+), 753 (M+-CH3), 711 (M+-
tBu), 636 (M+-2tBu-H2O), 621 (M+-2tBu-CH3-H2O).
HRMS(EI) m/z: 768.4432 (M+) (calcd for C39H66O4
F6Si2 768.4404).
21b: IR (neat): 3326, 2953, 2857, 1472, 1211, 1075,
1
835 cm 1. H NMR (270 MHz) d: 0.06 (12H, s), 0.56
(3H, s), 0.88 (18H, s), 1.00 (3H, d, J=6.3 Hz), 4.20 (1H,
m), 4.37 (2H, m), 4.86 (1H, d, J=2.3 Hz), 5.18 (1H, br
s), 6.02 (1H, d, J=11.2 Hz), 6.24 (1H, d, J=11.2 Hz),
6.28 (1H, br s). MS(EI) m/z: 768 (M+), 753 (M+-CH3),
711 (M+-tBu), 636 (M+-2tBu-H2O), 621 (M+-2tBu-
CH3-H2O). HRMS(EI) m/z: 768.4402 (M+) (calcd for
C39H66O4F6Si2 768.4404).
1
265 nm. H NMR (270 MHz) d: 0.58 (3H, s), 1.00 (3H,
d, J=6.6 Hz), 4.23 (1H, br s), 4.43 (1H, m), 5.00 (1H,
br s), 5.0±5.3 (1H, m), 5.32 (1H, br s), 6.02 (1H, d,
J=11.5 Hz), 6.38 (1H, d, J=11.5 Hz). MS(FAB+) m/z:
542 (M+), 525 (M+-H2O+H), 523 (M+-F), 505 (M+-
F-H2O). HRMS(FAB+) m/z: 542.2597 (M+) (calcd for
C27H37O3F7 542.2619).
26,26,26,27,27,27-Hexa¯uoro-9,10-secocholesta-5(Z),7(E),
10(19)-trien-1ꢁ,3ꢀ,23(S),25-tetraol (5a). Methanesulfonic
acid (MsOH, 0.06mL) was added to a solution of 21a
(1.0 g, 1.30 mmol) in methanol (15 mL). After stirring
for 1 h at room temperature, the reaction mixture was
quenched with water and extracted with ethyl acetate.
The organic layer was washed with saturated aqueous
NaHCO3, brine, and dried over MgSO4. The ®ltrate was
concentrated in vacuo. The residue was puri®ed by silica
gel chromatography (EtOAc:hexane, 3:2) to give 5a
(624 mg, 89%) as an amorphous powder. UV (EtOH):
23(S),26,26,26,27,27,27-Hepta¯uoro-9,10-secocholesta-5
(Z),7(E),10(19)-trien-1ꢁ,3ꢀ,25-triol (6b). DAST (35 mL,
0.26mmol) was added to a solution of 21a (50 mg,
0.065 mmol) in dichloromethane (1.0 mL) at 50ꢀC. The
mixture was stirred for 10min at the same temperature,
warmed to room temperature and stirred for 1 h. The
reaction mixture was treated in the same manner described
in the preparation of 6a to give 6b (13 mg, 37%, (46%)*) as
an amorphous solid with the recovery of 21b (10 mg,
20%). UV (EtOH): lmax 265 nm. 1H NMR (270MHz) d:
0.57 (3H, s), 1.01 (3H, d, J=6.3 Hz), 4.23 (1H, br s), 4.43
(1H, m), 5.00 (1H, br s), 5.0±5.3 (1H, m), 5.33 (1H, br s),
6.02 (1H, d, J=11.0 Hz), 6.38 (1H, d, J=11.0 Hz). MS
(FAB+) m/z: 542 (M+), 525 (M+-H2O+H), 523 (M+-F),
505 (M+-F-H2O). HRMS(FAB+) m/z: 542.2595 (M+)
(calcd for C27H37O3F7 542.2619).
1
lmax 264 nm. H NMR (270 MHz) d: 0.57 (3H, s), 0.98
(3H, d, J=5.6 Hz), 4.23 (1H, br s), 4.38 (1H, m), 4.42 (1H,
m), 4.99 (1H, br s), 5.33 (1H, br s), 6.02 (1H, d, J=11.5 Hz),
6.29 (1H, br s), 6.37 (1H, d, J=11.5 Hz). MS(EI) m/z: 540
(M+), 522 (M+-H2O), 507 (M+-H2O-CH3), 504 (M+-
2H2O), 489 (M+-2H2O-CH3). HRMS(EI) m/z: 540.2631
(M+) (calcd for C27H3804F6 540.2674).
26,26,26,27,27,27-Hexa¯uoro-9,10-secocholesta-5(Z),7(E),
10(19)-trien-1ꢁ,3ꢀ,23(R),25-tetraol (5b). MsOH (0.06 mL)
was added to a solution of 21b (1.64 g, 2.13 mmol) in
methanol (15 mL). After stirring for 1 h at room tem-
perature, the mixture was treated in the same manner
described in the preparation of 4. The residue was pur-
i®ed by silica gel chromatography (EtOAc:hexane, 3:2)
to give 5b (1.00 g, 87%) as an amorphous powder. UV
(EtOH): lmax 264 nm. 1H NMR (270 MHz) d: 0.58 (3H,
s), 1.00 (3H, d, J=6.3 Hz), 4.23 (1H, m), 4.33 (1H, m),
4.42 (1H, m), 5.00 (1H, br s), 5.33 (1H, br s), 6.01 (1H, d,
J=10.5 Hz), 6.37 (1H, d, J=10.5 Hz), 6.41 (1H, br s).
MS(EI) m/z: 540 (M+), 522 (M+-H2O), 507 (M+-H2O-
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-23(S)-[(methanesulfo-
nyl)oxy]-26,26,26,27,27,27-hexa¯uoro-9,10-secocholesta-
5(Z),7(E),10(19)-trien-25-ol (23a). Methanesulfonyl chlo-
ride (MsCl, 0.25 mL, 3.23 mmol) was added to a solu-
tion of 21a (100 mg, 0.130 mmol) in pyridine (1.0 mL).
The reaction mixture was stirred at 0 ꢀC for 1 h. Water
was added to the mixture and extracted with ethyl ace-
tate. The organic layer was washed with 1N HCl, satu-
rated aqueous NaHCO3, brine, and dried over MgSO4.
The ®ltrate was concentrated in vacuo. The residue was
puri®ed by silica gel chromatography (EtOAc:hexane,
1:10) to give 23a (86 mg, 78%) as an amorphous solid.

1816
M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
1H NMR (300MHz) d: 0.06 (12H, s), 0.55 (3H, s), 0.88
(18H, s), 1.03 (3H, d, J=5.9Hz), 2.83 (1H, m), 3.07 (3H,
s), 4.19 (1H, m), 4.38 (1H, m), 4.59 (1H, br s), 4.85 (1H, br
s), 5.19 (2H, m), 6.02 (1H, d, J=11.3 Hz), 6.23 (1H, d,
J=11.3 Hz).
24b (118 mg, 0.157 mmol) in THF (1.0mL) was added
TBAF (1M in THF, 0.5 mL) at room temperature and
stirred overnight. The same treatment as described in the
preparation of 7a gave 7b (67 mg, 82%) as an amorphous
1
solid. H NMR (300MHz) d: 0.53 (3H, s), 0.96 (3H, d,
J=6.2Hz), 2.83 (1H, m), 4.23 (1H, m), 4.44 (1H, m), 5.00
(2H, m), 5.33 (1H, br s), 6.01 (1H, d, J=11.1 Hz), 6.37 (1H,
d, J=11.1 Hz). MS(FAB+) m/z: 522 (M+), 505 (M+-
H2O+H+), 487 (M+-2H2O+H+). HRMS(FAB+) m/z:
522.2549 (M+) (calcd for C27H36O3F6 522.2568).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-23(R),25-epoxy-26,
26,26,27,27,27 - hexa¯uoro - 9,10 - secocholesta - 5(Z),7(E),
10(19)-trliene (24a). To a solution of 23a (65 mg, 0.077
mmol) in methanol (0.5 mL) was added 5% KOH-
MeOH (1.0 mL) at room temperature. After stirring for
10 min, the reaction mixture was concentrated and pur-
i®ed by PLC (0.5 mm, EtOAc:hexane, 3:2) to give 24a
(40 mg, 69%). 1H NMR (300MHz) d: 0.06 (12H, s), 0.54
(3H, s), 0.88 (18H, s), 0.93 (3H, d, J=6.6 Hz), 2.82 (1H,
m), 4.19 (1H, m), 4.37 (1H, m), 4.86 (1H, br s), 5.01 (1H,
m), 5.18 (1H, br s), 6.02 (1H, d, J=11.3 Hz), 6.23 (1H, d,
J=11.3 Hz).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-23-oxo-9,10-seco-
cholesta-5(Z),7(E),10(19)-trien-25-ol (11). To a solution
of LiN(TMS)2 (1.04 M in hexane, 0.63 mL, 0.66 mmol) in
THF (1.0 mL) was added dropwise a solution of 13
(300 mg, 0.50 mmol) in THF (1.0 mL) at 78 ꢀC. After
stirring for 10 min, acetone (60 mL, 0.82 mmol) was
added. After stirring for 10 min, the reaction mixture was
quenched by 1N HCl and extracted with ethyl acetate.
The organic layer was washed with saturated aqueous
NaHCO3, brine, and dried over MgSO4. The ®ltrate was
concentrated in vacuo. The residue was puri®ed by silica
gel chromatography (EtOAc:hexane, 1:20!1:10) to give
11 (224 mg, 68% (87%) as an amorphous solid with
23(R),25-Epoxy-26,26,26,27,27,27-hexa¯uoro-9,10-seco-
cholesta-5(Z),7(E),10(19)-triene (7a). To a solution of
24a (40 mg, 0.053 mmol) in THF (1.0 mL) was added
TBAF (1M in THF, 0.5 mL) at room temperature and
stirred overnight. The reaction mixture was quenched
with water and extracted with ethyl acetate. The organic
layer was washed with brine and dried over MgSO4. The
®ltrate was concentrated in vacuo. The residue was
puri®ed by silica gel chromatography (EtOAc:hexane,
3:2) to give 7a (19 mg, 68%) as an amorphous solid. UV
(EtOH): lmax 265 nm. 1H NMR (300 MHz) d: 0.55 (3H,
s), 0.93 (3H, d, J=6.6 Hz), 2.60 (2H, m), 2.83 (1H, m),
4.24 (1H, m), 4.43 (1H, m), 5.00 (2H, m), 5.32 (1H, br
s), 6.02 (1H, d, J=11.5 Hz), 6.37 (1H, d, J=11.5 Hz).
MS(FAB+) m/z: 522 (M+), 505 (M+-H2O+H+), 487
(M+-2H2O+H+). HRMS(FAB+) m/z: 522.2574 (M+)
(calcd for C27H36O3F6 522.2568).
1
recovery of 13 (66 mg, 22%). H NMR (300 MHz) d:
0.05 (12H, s), 0.56 (3H, s), 0.86 (18H, s), 0.93 (3H, d,
J=6.4 Hz), 2.55±2.58 (2H, m), 3.98 (1H, s), 4.18 (1H,
m), 4.36 (1H, m), 4.85 (1H, br s), 5.16 (1H, br s), 6.00
(1H, d, J=11.2 Hz), 6.22 (1H, d, J=11.2 Hz).
23-Oxo-9,10-secocholesta-5(Z),7(E),10(19)-trien-1ꢁ,3ꢀ,
25-triol (8). MsOH (20 mL) was added to a solution of 11
(86 mg) in methanol (5.0 mL). The same treatment as
described in the preparation of 5a and puri®cation by silica
gel chromatography (EtOAc:hexane, 2:1) gave 8 (49 mg,
87%) as an amorphous solid. 1H NMR (300MHz) d: 0.58
(3H, s), 0.94 (3H, d, J=6.4 Hz), 1.24 (6H, s), 2.56±2.58
(2H, m), 3.97 (1H, s), 4.22 (1H, m), 4.42 (1H, m), 5.00 (1H,
br s), 5.33 (1H, br s), 6.01 (1H, d, J=11.4 Hz), 6.37 (1H, d,
J=11.4 Hz). MS(FAB+) m/z: 431 (M++1), 430 (M+),
413 (M+-H2O+H+). HRMS(FAB+) m/z: 430.3042
(M+) (calcd for C27H42O4 430.3083).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-23(R)-[(methanesul-
fonyl)oxy] - 26,26,26,27,27,27 - hexa¯uoro - 9,10 - secocho-
lesta-5(Z),7(E),10(19)-trien-25-ol (23b). MsCl (0.5 mL,
6.46mmol) was added to a solution of 21a (200mg,
0.260 mmol) in pyridine (2.0 mL). The same treatment as
described in the preparation of 23a gave 23b (220mg,
1
100%) as an amorphous solid. H NMR (300MHz) d:
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-26,27-dimethoxy-23-
oxo-9,10-secocholesta-5(Z),7(E),10(19)-trien-25-ol (12).
To a solution of LiN(TMS)2 (1.04 M in hexane, 1.26 mL,
1.31 mmol) in THF (5.0 mL) was added dropwise a solu-
tion of 13 (600mg, 1.0mmol) in THF (3.0 mL) at 78 ꢀC.
After stirring for 10 min, dimethoxyacetone (0.35 g,
3.0 mmol) was added. The same treatment as described in
the preparation of 11 and puri®cation by silica gel chro-
matography (EtOAc:hexane, 1:3)gave 12(660 mg, 92%) as
an amorphous solid. 1H NMR (300 MHz) d: 0.04 (12H, s),
0.55 (3H, s), 0.85 (18H, s), 0.91 (3H, d, J=6.4 Hz), 2.65
(2H, s), 3.32 (10H, s), 4.04 (1H, s), 4.17 (1H, m), 4.34 (1H,
m), 4.83 (1H, br s), 5.15(1H, br s), 5.99(1H, d, J=11.2 Hz),
6.20 (1H, d, J=11.2Hz).
0.06 (12H, s), 0.55 (3H, s), 0.88 (18H, s), 1.01 (3H, d,
J=6.3Hz), 3.08 (3H, s), 4.19 (1H, m), 4.38 (1H, m), 4.86
(1H, br s), 4.90 (1H, br s), 5.19 (2H, m), 6.02 (1H, d,
J=11.2 Hz), 6.25 (1H, d, J=11.2 Hz).
1ꢁ,3ꢀ-Bis[(t-butyldimethylsilyl)oxy]-23(R),25-epoxy-26,
26,26,27,27,27-hexa¯uoro-9,10-secocholesta-5(Z),7(E),
10(19)-triene (24b). To a solution of 23b (580 mg, 0.685
mmol) in MeOH (3.0 mL) was added 5% KOH-MeOH
(5.0 mL). The same treatment as described in the pre-
paration of 24a gave 24b (283 mg, 55%) as an amorphous
solid. 1H NMR (300MHz) d: 0.06 (12H, s), 0.52 (3H, s),
0.88 (18H, s), 0.96 (3H, d, J=6.2Hz), 2.82 (1H, m), 4.19
(1H, m), 4.38 (1H, m), 4.86 (1H, br s), 5.00 (1H, quint.,
J=7.2Hz), 5.18 (1H, br s), 6.01 (1H, d, J=11.2 Hz), 6.23
(1H, d, J=11.2 Hz).
26,27-Dimethoxy-23-oxo-9,10-secocholesta-5(Z),7(E),
10(19)-trien-1ꢁ,3ꢀ,25-triol (9). MsOH (16 mL) was
added to a solution of 11 (69 mg, 0.096 mmol) in metha-
nol (4.0 mL). The same treatment as described in the
preparation of 5a silica gel chromatography (EtOAc:
23(S),25-Epoxy-26,26,26,27,27,27-hexa¯uoro-9,10-seco-
cholesta-5(Z),7(E),10(19)-triene (7b). To a solution of

M. Ikeda et al. / Bioorg. Med. Chem. 8 (2000) 1809±1817
1817
hexane, 3:1) gave 9 (41 mg, 87%) as an amorphous
1
4. Inaba, M.; Okuno, S.; Nishizawa, Y.; Yukioka, K.; Otani,
S.; Matsui-Yuasa, I.; Morisawa, S.; Deluca, H. F.; Morii, H.
Arch. Biochem. Biophys. 1987, 258, 421.
5. Honda, A.; Nakashima, N.; Mori, Y.; Katsumata, T.; Ishi-
zuka, S. J. Steroid Biochem. Molec. Biol. 1992, 41, 109.
6. Katsumata, T.; Ishibashi, E.; Tanaka, T.; Nakatsuka, M.
KISO TO RINSHO 1996, 30, 2955.
solid. H NMR (300 MHz) d: 0.56 (3H, s), 0.92 (3H, d,
J=6.4 Hz), 2.65 (2H, s), 3.33 (10H, s), 4.04 (1H, s), 4.20
(1H, m), 4.41 (1H, m), 4.97 (1H, br s), 5.31 (1H, br s),
6.00 (1H, d, J=11.4 Hz), 6.34 (1H, d, J=11.4 Hz).
MS(FAB+) m/z: 491 (M++1), 490 (M+), 473 (M+-
H2O+H+). HRMS(FAB+) m/z: 490.3333 (M+) (calcd
for C29H46O6 490.3297).
7. Ando, K.; Kondo, F.; Koike, F.; Takayama, H. Chem.
Pharm. Bull. 1992, 40, 1662.
8. Andrews, D. R.; Barton, D. H. R.; Cheng, K. P.; Finet, J.
P.; Hesse, R. H.; Johnson, G.; Pechet, M. M. J. Org. Chem.
1986, 51, 4819.
9. Andrews, D. R.; Barton, D. H. R.; Hesse, R. H.; Pechet, M.
M. J. Org. Chem. 1986, 51, 1637.
10. Ishihara, T.; Shinjo, H.; Inoue, Y.; Ando, T. J. Fluorine
Chem. 1983, 22, 1.
11. Burger, K.; Helmreich, B. J. Prakt. Chem./Chem.-Ztg
1992, 334, 219.
12. The reaction mixture was quenched with 10% aqueous
NH4Cl solution.
13. Lithium 2,2,6,6-tetramethylpiperidide, lithium bis(di-
methylphenysilyl)amide (each 1.4 equiv) also gave similar
results as LiN(TMS)2.
14. Middleton, W. J. J. Org. Chem. 1975, 40, 574.
15. The treatment of mesylate 23b with TBAF in THF at room
temperature was also investigated. However, oxetane 24b was
obtained as an undesired product.
16. 1a,25(OH)2-23-oxo-D3 has a lower binding anity to
VDR (50% compared to 1); Horst, R. L.; Reinhardt, T. A.;
Napoli, J. L. Biochem. Biophys. Res. Cormmun. 1982, 107,
1319.
Measurement of vitamin D3 receptor binding anity.
Chick intestinal 1a,25-dihydroxyvitamin D3 receptor
was obtained from Yamasa Biochemical (Chiba, Japan)
and dissolved in 50 mM Tris-HCl bu€er (pH 7.4) con-
taining 0.25 M sucrose, 25 mM KCl, 5 mM MgCl2, 1 mM
EDTA and 12 mM thioglycerol just before use. The
receptor solution (0.2 mL, 0.08 mg protein) was incubated
with 1.8 nM [3H]-1a,25-dihydroxyvitamin D3 (180 Ci/
nmol) and 1a,25-dihydroxyvitamin D3 or analogue at
various concentration for 24 h at 4 ꢀC. The bound and
free [3H]-1a,25-dihydroxyvitamin D3 were separated by
treatment with dextran-coated charcoal and centrifuged
at 3000 rpm for 15 min at 4 ꢀC. The radioactivity of
supernatant with ACS-II (Amersham, UK) was counted.
Measurement of cellular di€erentiation. Human leuke-
mia HL-60 cells were incubated as described previously.19
The incubation was carried out for 4 days and at the end
of the fourth day, superoxide production was measured by
nitro blue tetrazolium (NBT) reduction. The number of
cells containing intracellular black-blue formazan depos-
its was determined by light microscopy. At least 200 cells
were counted in duplicate per determination.
17. Henze, H. R.; Rogers, B. G. J. Am. Chem. Soc. 1939, 61,
434.
18. Seino, Y.; Yamaoka, K.; Ishida, M.; Yabuuchi, H.; Ichi-
kawa, M.; Ishige, H.; Yoshino, H.; Avioli, LV. Calci®ed Tissue
International 1982, 34, 265.
19. Ostrem, V. K.; Lau, W. F.; Lee, S. H.; Penman, K.; Prahl,
J.; Schnoes, H. K.; Deluca, H. F.; Ikekawa, N. J. Biol. Chem.
1987, 262, 14164.
Acknowledgements
20. Allinger, N. L.; Yuh, Y. H.; Lii, J.-H. J. Am. Chem. Soc.
1989, 111, 8551.
21. Qiu, D.; Shenkin, P. S.; Hollinger, F. P.; Still, W. C. J.
Phys. Chem. A 1997, 101, 3005.
We acknowledge the assistance of Ms. Maki Tamura in
conformational analysis.
22. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Lis-
kamp, R.; Lipton, M.; Cau®eld, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J. Comput. Chem. 1990, 11, 440.
23. At higher than indicated energy level, the spatial region
occupied by side chains of 3±7b are expanded and almost
overlapped with each other. The side chain of 3 tends to have
a similar densely populated region when compared to that of
1a,25(OH)2D3. About the conformational study of
1a,25(OH)2D3 and its analogues, see ref 2 and the following
references: (a) Yamada, S.; Yamamoto, K.; Masuno, H.;
Ohta, M. J. Med. Chem. 1998, 41, 1467. (b) Yamamoto, K.;
Ohta, M.; Deluca, H. F.; Yamada, S. Bioorg. Med. Chem.
Lett. 1995, 5, 979.
References and Notes
1. Ettinger, R. A.; DeLuca, H. F. Advances in Drug Research
1996, 28, 269.
2. Competitive biological activities of 1, 2, 1a,23(R),25 (OH)3D3
and other analogues were summarized in the following refer-
ence. The OH modi®cation lead to decrease VDR anity com-
pared to 1, and the stereochemical e€ect of substituents was not
observed. (23S-OH: 12%, 23R-OH: 11%); Bouillon, R.; Oka-
mura, W. H.; Norman, A. W. Endocr. Rev. 1995, 16, 200.
3. Tanaka, Y.; Deluca, H. F.; Kobayashi, Y.; Ikekawa, N.
Arch. Biochem. Biophys 1984, 229, 348.