Total synthesis of 25-hydroxy-16,23E-diene vitamin D3 and 1α,25-dihydroxy-16,23E-diene vitamin D3: Separation of genomic and nongenomic vitamin D activities

Article abstract of DOI:10.1016/S0968-0896(98)00164-3

Separation of genomic and nongenomic vitamin D activities was achieved by structural modification of 1,25-dihydroxy vitamin D₃ by introduction of 16 and 23E double bonds. The modified compound 3, lacking a 1α-hydroxy group, exhibits only nongen

Full text of DOI:10.1016/S0968-0896(98)00164-3

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 2051±2059
Total Synthesis of 25-Hydroxy-16,23E-diene Vitamin D3 and
1ꢀ,25-Dihydroxy-16,23E-diene Vitamin D3: Separation of
Genomic and Nongenomic Vitamin D Activities
Marek M. Kabat,a,* Walter Burger,a Sandra Guggino,b Bernard Hennessy,a
Jerome A. Iacobelli,a Kazuhiro Takeuchi,b and Milan R. Uskokovica
aRoche Research Center, Ho€mann-La Roche Inc., 340 Kingsland Street, Nutley, NJ 07110, USA
bThe Johns Hopkins University School of Medicine, 720 Rutland Avenue, Baltimore, MD 21205, USA
Received 22 April 1998; accepted 15 June 1998
AbstractÐSeparation of genomic and nongenomic vitamin D activities was achieved by structural modi®cation of
1,25-dihydroxy vitamin D3 by introduction of 16 and 23E double bonds. The modi®ed compound 3, lacking a 1a-
hydroxy group, exhibits only nongenomic activity. Its 1a-hydroxy relative 4 expresses fully both genomic and non-
genomic activities. A total synthesis of analogues 3 and 4 is described. # 1998 Elsevier Science Ltd. All rights reserved.
Introduction
Although 1 has long been recognized as a calcitropic
hormone that potentiates long-term changes resulting
from gene activation, recent studies have shown that
physiological concentrations of 1 also cause rapid e€ects
such as acute increases of calcium in¯ux into osteoblasts
cells.10 The rapid increases in intracellular calcium sti-
mulated by 1 in mouse osteoblast primary cultures is
blocked by nitrendipine and varapamil, indicating that a
speci®c class of channels called `l-type calcium chan-
nels' is involved.10,11 1 rapidly activates a similar cal-
cium in¯ux in cardiac muscle12 and myoblasts13 as well
as both calcium in¯ux and internal calcium release in
liver.14 Multiple signal transduction pathways are sti-
mulated in the CaCo2 intestinal cell line including cal-
cium in¯ux, internal calcium release, phosphoinositol
turnover,15 and tyrosine kinase phosphorylation.16
The nuclear 1,25-dihydroxy vitamin D3 (1) receptor
(VDR) is a member of the steroid±thyroid receptor
superfamily that regulates various gene transcriptions.
These gene transcriptions are usually mediated by a
heterodimer of VDR and a 9-cis-retinoic acid receptor
(RXR) which targets sequence-speci®c DNA binding
domains, vitamin
D
receptor-responsive elements
(VDRE).1,2 1 causes numerous transcriptional respon-
ses3 including the di€erentiation of intestine, bone, and
skin cells, and the di€erentiation and inhibition of pro-
liferation of various tumor cells. In target organs
involved in calcium homeostasis, such as intestine, 1
initiates gene transcription of proteins (such as calbini-
din D28 and calbinidin D9) which are thought to be
involved in transepithelial calcium absorption. In bone,
1 is responsible for increased gene transcriptions leading
to bone matrix proteins such as osteocalcin,4,5 osteopo-
nin,6,7 and collagen8,9 type 1 produced by osteoblasts in
the process of bone formation. The 1 receptor (VDR) in
osteoblasts a€ects both the process of bone formation
and bone resorption thus altering bone turnover.
Results
Separation of genomic and nongenomic vitamin D
activities
After the original observation of the rapid e€ects of 1 on
perfused chick duodenum,17 multiple vitamin D3 analo-
gues have also been tested for transepithelial calcium
transport (transcaltachia).18,19 Several compounds,
which have a low potency (less than 1%) for activation
Key words: Vitamin D3 analogues; genomic e€ect; nongenomic
e€ect; calcium channel; transcriptional e€ect.
*Corresponding author.
0968-0896/98/$ - see front matter # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00164-3

2052
M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
Comparison of calcium channel potentiation by 3 and
its 1a-hydroxy analogue 4 indicates that the 1a-hydroxy
function diminishes anity for calcium channel poten-
tiation (Table 1) only slightly. In comparison to the
activity of 25(OH)D3 (2) and 1, the 16,23E-diene struc-
tural modi®cation does contribute signi®cantly to cal-
cium channel potentiation.
Osteocalcin biosynthesis by 1 and 4 was almost iden-
tical, but 3 was much less active, indicating the sig-
ni®cance of the 1a-hydroxyl group in¯uence on this
genomic transcriptional event (Table 2). Thus, at phy-
siological concentrations (10 10 M), 3 exhibits only the
nongenomic rapid e€ects of 1 cleanly without any sig-
ni®cant genomic e€ects mediated by VDR.
Total synthesis of 3 and 4
Recently, we described a convergent synthesis of 1 and 2
in which 25-hydroxy Grundmann±Windaus ketone was
coupled with a cyclopropane±acetylenic precursor of the
A ring 5.24,25 Two asymmetric routes elaborated in our
laboratory, starting from cyclopentenone (7), provided 5
in multigram quantities. For the synthesis of title com-
pounds 3 and 4, we investigated a similar coupling of
the A-ring precursor 5 (Scheme 1) with a suitably func-
tionalized 25-hydroxy-16-ene-23-yne C/D-ring 6. We
anticipated that the synthetic precursor of compound 6
would be the ole®n 8, which was obtained from Hajos
ketone 9.26
Chart 1.
of genomic events, are found to stimulate this process.19
These compounds also stimulate rapid uptake of
[45Ca2+] into ROS cells20 or activate plasma membrane
calcium in¯ux through Ca2+ channels21,22 suggesting
the same rapidly acting nongenomic receptor may be
present in both bone and intestine. It is generally
thought that the 1a-hydroxy group is critical for anity
on VDR. Alterations in the CD-ring side chain in the
absence of the 1a-hydroxy group, however, contribute
to an increased stimulation of calcium current via cal-
cium channels, despite that their binding to VDR is
insigni®cant.
The ene reaction between ole®n 827 and aldehyde 1028
produced, in the presence of dimethylaluminum chlor-
ide, an epimeric 4:1 mixture of propargylic alcohols 11
in quantitative yield (Scheme 2).29 The natural con®g-
uration at C-20 in product 11 was secured by the Z-
con®guration of the ethylidene double bond of 8. The
stereogenic center at C-22 was inconsequential since the
hydroxyl group was removed by two-step procedure
developed by Barton et al.30 Toward this end, com-
pounds 11 were transformed into the respective mixture
of phenylthiocarbonyl derivatives 12 that on treatment
with tri-n-butyltin hydride yielded compound 13.
Hydrolysis of the acetate group and deprotection of the
25-hydroxyl group provided diol 14, which was oxidized
with PCC to give hydroxy ketone 15. Protection of the
25-hydroxyl group of 15 with TMS-imidazole produced
the C/D ring precursor 6.
For example, the analogue 25(OH)-16,23E-diene D3 (3)
(Chart) potently stimulates calcium in¯ux with a K 1/2
greater than 0.05 nM, compared to 1 which has K 1/2 of
0.5 nM, but the analogue 3 does not stimulate tran-
scription by VDR.23 From the data for 1 and 3, it is not
clear what the e€ect of 1a-hydroxy group is with respect
to nongenomic activity. We have examined, therefore,
the nongenomic e€ects of the addition of a 1a-hydroxy
group to 3, i.e., 1,25(OH)2-16,23E-diene D3 (4). At the
same time, we looked at the e€ects of the insertion of
16- and 23E-double bonds in 1 on the genomic function,
by measuring the production of osteocalcin in ROS 17/
2.8 cells.
In the next stage, we performed the coupling of A-ring
precursor 5 and the C/D fragment 6 (Scheme 3).
Treatment of compound 5 with n-BuLi produced its
acetylide, which reacted with ketone 6 to give the main
framework 16. Deprotection of the 25-hydroxyl group
with TBAF provided diol 17. Compound 17, the struc-
ture of which consists of two propargylic alcohols,

M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
2053
Table 1. Potentiation of calcium channels by vitamin D3 in ROS cells. The agent 25(OH)-16,23E-diene D3 (3) shows very potent e€ect
(mean mV‹SE) on calcium channels at 0.05 nM and maximal e€ects at 0.5 nM. Likewise, the agent 1,25(OH)2-16,23E-diene D3 (4) is
nearly as potent at 0.05 nM, suggesting that addition of the 1a group does not inhibit anity for calcium channel activation
Vitamin D analogue
Concn
0.5 nM
0.05 nM
0.1 nM
5 nM
50 nM
25(OH)D3
1,25(OH)2D3
1.6‹1.16a
5.73‹1.33a
10.53‹0.3a
10.42‹1.07b
4.7‹3.7a
14.0‹3.6a
10.17‹0.54c
1.78‹0.82c
6.05‹0.4c
5.17‹2.98a
9.79‹0.19c
25(OH)-16,23E-diene D3
1,25(OH)2-16,23E-diene
D3
9.31‹0.4b
8.36‹0.23b
10.78‹1.34b
a
Number of patches equals 3.
b Number of patches equals 4.
c Number of patches equals 5.
served as an essential intermediate for the synthesis of
both vitamin D3 analogs 3 and 4. In the synthesis of 3,
compound 17 was reduced with LiAlH4 in THF to give
the product 18 bearing two allylic alcohols. To gen-
erate the vitamin D3 triene, solvolysis of the cyclopro-
pane-allylic alcohol moiety in 18 was carried out
analogously to the way we described earlier,24,25 i.e., by
the treatment with a catalytic amount of acid (p-TsOH)
in dioxane±water. The obtained mixture of 3 and the
corresponding 5,6-cis/trans-isomer was photolysed with
a mercury lamp in the presence of 9-acetoxyanthracene
as photosensitizer, to give the desired compound 3 as a
single product.
For the synthesis of compound 4, we needed to intro-
duce a hydroxyl group with the alpha con®guration at
C-1 carbon. For this purpose, a two-step procedure was
applied which consisted of the oxidation of 17 at the
allylic position with selenium dioxide. The obtained
mixture of 1-ketone and 1a-alcohol was then reduced
with Super-Hydride1. The reduction of 1-ketone was
fully stereospeci®c (none of the epimeric alcohol was
detected). Subsequently, the reduction of 19 with
LiAlH4 provided bis allylic alcohols 20. This product
was subjected to solvolysis and photolytic isomerization
of thus-obtained 5,6-cis/trans mixture to produce 4 as a
single compound.
Table 2. Concentration dependence of osteocalcin secretion
by vitamin D3 analogs. As previously reported, 1,25(OH)2D3
(1) stimulates osteocalcin secretion above the control value. The
analogue 1,25(OH)2-16,23E-diene D3 (4) shows an equally
potent secretion but 25(OH)-16,23E-diene D3 (3) does not show
secretion above control until concentration of 1 nM
Conclusion
Vitamin D analogue
Concn (nM) Osteocalcin
secretiona
Introduction of the 16- and 23E-bonds into 25-hydroxy
vitamin D3 structure produced 3 which at physiological
concentrations exhibits only nongenomic rapid e€ects
characteristic of 1, cleanly, without signi®cant genomic
e€ect mediated by VDR. In addition, the 1a group
added to this molecule does not alter the sensitivity of
the nongenomic e€ect of 3.
Control
1,25(OH)2D3
11‹0.6b
20‹2
51‹7
0.001
0.01
0.1
71‹10.5
78‹11
70‹17
1
10
25(OH)-16,23E-diene D3
0.001
0.01
0.1
12‹0.5
13‹0.4
12‹0.2
19‹0.4
21‹0.8
Experimental
1
10
Materials and methods
1,25(OH)2-16,23E-diene D3
0.001
0.01
0.1
18‹1.6
49‹4.6
78‹9.4
79‹7.4
72‹6.3
Cell culture. ROS 17/2.8 cells were grown in 5% CO2 at
37 ꢀC in Hamm's F12 medium (Life Technologies, Inc.)
containing 5% heat-inactivated fetal calf serum (Life
Technologies) and 0.1 mg/mL kanamycin bu€ered with
25 mM NaHCO3 and 14 mM Hepes using NaOH to
adjust pH to 7.4. For patch clamp experiments, cells
1
10
a
Mean value‹SE
b ng/mg of protein/24 h

2054
M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
Scheme 1.
Scheme 2.
were seeded at low density as single cells onto pieces of
coverslips. Patch clamp experiments were performed
within 24 h after plating, on single cells after overnight
incubation in 1% charcoal-treated serum. The data in
each table are compiled from one seeding of cells from
which control and test experiments were alternated on
the same day.
Solutions. The perforated-patch recording technique
was used for measuring inward barium currents under

M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
2055
Scheme 3.
voltage clamp conditions. The composition of the pip-
ette solution was 100 mM KOH, 150 mM Hepes, 20 mM
EGTA, 2 mM CaCl2, 2 mM MgCl2, 10 mM K2HPO4
bu€ered to pH 7.4 with KOH and osmolarity adjusted
to 290 mosm/kg using K-Hepes. Amphotericin B, an
antibiotic that created small nonselective pores in mem-
brane to allow ion ¯ow, was added to this solution at a
®nal concentration of 240 mg/mL, and then the mixture
was added to the tip of the pipette. The pipette was then
back®lled with pipette saline solution onto the antibiotic
saline mixture. The composition of the initial external
solution was 140 mM NaCl, 5 mM KCl, 20 mM Hepes
bu€ered to pH 7.4 with NaOH, and the osmolarity
adjusted to 290 mosm with Na-Hepes. After whole cell
currents were established, a solution which consisted
of 115 mM BaCl2 and 20 mM Hepes bu€ered to pH
7.4 with tetraethylammonium hydroxide, was added
to the initial external solution so the ®nal concentra-
tion of barium was 20 mM. Barium was used as a
current carrier for two reasons. Barium current through
l-type channels is known to be larger than calcium
currents, and barium inhibits the activation of potas-
sium channels. External tetraethylammonium hydroxide
was used as a potassium channel blocker. Hepes, a
nonpermeant anion, was used to eliminate inward cur-
rents via anion conductance. Using these conditions the
inward barium current is stimulated by BayK 8644,
completely blocked by nitrendipine, and displays the
voltage-gating characteristics indicative of l-type cal-
cium channels.
Current measurement. A pipette was placed to the sur-
face of cell, and then gentle suction was applied until a
tight seal of about 10 Gohm was formed. After about 3±
10 min, the amphotericin B di€used into the cell mem-
brane under the patch pipette causing the capacitance to
increase, at which time the experiment was initiated.
The cell membrane potential was held at 70 mV;
10 mV step pulses were applied for 300 ms between 60
and +50 mV with 2 s interval between pulses. Currents
were monitored on an EPC-7 patch clamp ampli®er
(List Electronics, Darmstadt, Germany) and visualized
on a Nicolet digital oscilloscope (Nicolet Instrument,
Madison, WI) after ®ltering at 1 kHz through a Bessel
®lter (Frequency Devices, Haverhill, MA). Test sub-
stances (vitamin D3 metabolites) were dissolved in a
bath solution with 10% ethanol then added to the bath
solution by pipette. The test substances were diluted

2056
M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
into the bath in a 1 to 100 ratio giving a ®nal dilution of
ethanol of 1/1000. Controls for addition of ethanol did
not cause left shift or changes in the amplitude of cal-
cium channel currents. It took about 30 s to add test
substances. After the initiation of drug addition, the ®rst
currents were recorded within 30 s, and the data were
sampled once per min afterwards. The data, which were
recorded just after adding the samples, showed ca. 80%
left shift. The data at 1.5 min showed a maximal steady-
state value. Data were stored and analyzed on computer
(IBM-PC compatible 386, AST) using the P-clamp soft-
ware version 5.5 (Axon instruments, Foster City, CA).
At the end of each experiment BayK 8644 (1 mM) was
added to determine the maximal current through the L-
type calcium channels and to standardize the data
between experiments. This amount of BayK 8644 was
previously shown in control cells to cause a maximal
increase in barium current and maximal left shift in
barium current. This concentration of BayK 8644
usually caused large currents in the range of 20 mV;
thus, because of the rather high resistance of the
amphotericin patch, the clamp was not totally e€ective
in this voltage range. BayK 8644 was used at the end of
each experiment to ensure that after three sequential
steroid additions the maximal barium currents could be
stimulated.
General
Melting points were measured in open capillary tubes
1
and are uncorrected. H NMR spectra were obtained in
CDCl3 at 200 or 400 MHz. Unless otherwise noted,
reactions were conducted under an atmospheric pres-
sure of dry argon. Organic extracts were dried over
anhydrous MgSO4 and ®ltered prior to removal of the
solvent under reduced pressure on rotary evaporator
(bath temperature 35 ꢀC). Chromatography was per-
formed on 230±400 mesh EM silica gel 60.
[3aS-[3(1R*),3aꢀ,7ꢀ,7aꢁ]]-7-(acetoxy)-[3-[[(1,1-dimethyl-
ethyl)dimethylsilyl]oxy]-3-methyl-1-butynyl]-3a,4,5,6,7,7a-
hexahydro-ꢁ,3a-dimethyl-1H-indene-3-ethanol (mixture
of epimers) (11). To the solution of [3aR-(1Z,3aa,
4b,7ab]-1-ethylidene-octahydro-7a-methyl-1H-inden-4-
ol acetate (8) (2.0 g, 9.0mmol) and 4-methyl-4-tert-butyl-
dimethylsilyloxy-butynal (10) (2.25 g, 9.0 mmol) in anhy-
drous dichloromethane (5 mL), cooled at 78 ꢀC, was
added over a 10 min period dimethylaluminum chloride
(20 mL, 1 M solution, 20 mmol). After 2 h, the reaction
by TLC was only half complete; therefore, additional
aldehyde 10 (2.25 g, 9 mmol) and dimethylaluminum
chloride (20 mL, 1 M solution, 20 mmol) were added.
After 1 h additional stirring, the reaction mixture was
poured into potassium bicarbonate (600 mL, 2 N solu-
tion) and extracted with ethyl acetate (4Â100 mL). The
extracts were combined, washed with water and brine,
dried, ®ltered, and evaporated to dryness. Flash chro-
matography on a silica gel column, eluting ®rst with
methylene chloride and then with hexanes/ethyl acetate
(3/1), gave the epimeric mixture 11 (4.16 g, 100% yield)
which was separated by HPLC with hexanes/ethyl acet-
ate (10/1). Major epimer (3.25 g, 78% yield); mp 47±
The peak of the current±voltage relationship was esti-
mated by drawing a line joining the three largest cur-
rent amplitudes. Using an equation that weights the
slope of the line on either side of the maximal measured
negative current, the peak current was estimated. Thus,
the peak current was estimated by equation
fꢁa b†=ꢁa b ‡ a c†g  10 mV ‡ e, where a is the
maximal current measured by the depolarization proto-
col; b is the current measured 10 mV more positive and
c is the current measured 10 mV more negative than
maximal current. e is the di€erence between the voltage
depolarization causing the maximal measured current
and 0 mV. All data are represented as mean‹SE. For
statistical analyses, the Student's test was used, and
p<0.05 was recognized as statistically signi®cant.
48 ꢀC; [a]2d5+7.2 ꢀ (c 0.56, EtOH); H NMR (CDCl3) d
1
0.15 (s, 3H), 0.16 (s, 3H), 0.86 (s, 9H), 1.07 (s, 3H), 1.14
(d, J 6.9Hz, 3H), 1.44 (s, 6H), 2.05 (s, 3H), 2.70 (p, J
6.4 Hz, 1H), 4.43 (t, J 6.3Hz, 1H), 5.21 (brs, 1H), 5.68
(brs, 1H); Anal. calcd for C26H44O4Si: C, 69.59; H, 9.88.
Found: C, 69.85; H, 9.74. Minor epimer (0.754 g, 18%
yield); mp 77±79 ꢀC; [a]2d5+18.1 ꢀ (c 0.52, EtOH); 1H
NMR (CDCl3) d 0.17 (2s, 6H), 0.87 (s, 9H), 1.03 (s, 3H),
1.14 (d, J 6.9 Hz, 3H), 1.47 (s, 6H), 2.05 (s, 3H), 2.37 (p,
J 6.4Hz, 1H), 4.40 (dd, J 3.7 and 7.8Hz, 1H), 5.21 (brs,
1H), 5.54 (brs, 1H); Anal. calcd for C26H44O4Si: C,
69.59; H, 9.88. Found: C, 69.42; H, 10.15.
Osteocalcin secretion. Hanks' bu€ered salt solution (pH
7.4) twice and incubated with a mixture of Ham's F12
medium/DMEM (1/1) containing 2% fetal bovine
serum and 50 mg of ascorbic acid per mL for 24 h. The
incubation was begun by replacing the medium with the
same medium plus 0.01 mM menadione, vitamin K3
(Sigma), and vitamin D metabolites. The incubation
was stopped at the appropriate time by removing the
medium and storing it at 20 ꢀC. Osteocalcin in the
medium was measured by radioimmunoassay. Speci®c
antibodies to rat osteocalcin or bovine osteocalcin that
cross-react with human material were purchased from
Biomedical Technologies (Stoughton, MA).
[3aS-[3(1R*),3aꢀ,7ꢀ,7aꢁ]]-thiocarbonic acid O-[1[1-[7-
(acetoxy)-3a,4,5,6,7,7a-hexahydro-3a-methyl-1H-inden-
3-yl]ethyl]-4-[[[(1,1-dimethylethyl)dimethylsilyl]oxy]4-
methyl-2-pentynyl] O-phenyl ester (mixture of epimers)
(12). To the solution of major epimer 11 (448 mg,
1.0 mmol) in anhydrous dichloromethane (10 mL) was
added pyridine (0.5 mL) and phenyl chlorothionoform-
ate (0.533 mL, 4.0 mmol). The reaction mixture was then

M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
2057
stirred at rt for 3 h, quenched with methanol, and after
15 min diluted with water (200 mL). After extraction
with ethyl acetate, the extracts were washed with 2 N
HCl, water, brine, 2 N potassium bicarbonate, water,
and brine, dried, ®ltered and evaporated to dryness.
Puri®cation was performed by ¯ash chromatography on
a silica gel column with dichloromethane/hexanes (3/1)
to give compound 12 (538 mg, 92% yield), a colorless
oil. [a]2d5+69.6 ꢀ (c 0.23, EtOH); 1H NMR (CDCl3) d
0.17 (s, 3H), 0.18 (s, 3H), 0.86 (s, 9H), 1.06 (s, 3H), 1.20
(d, J 7.1 Hz, 3H), 1.75 (s, 6H), 2.05 (s, 3H), 2.68 (p, J
6.6 Hz, 1H), 5.22 (brs, 1H), 5.62 (brs, 1H), 5.89 (d, J
6.8.0 Hz, 1H), 7.09 (d, J 8.0 Hz, 2H), 7.30 (t, J 8.0 Hz,
1H), 7.42 (t, J 8.0 Hz, 2H); Anal. calcd for C33H48O5SSi:
C, 67.77; H, 8.27. Found: C, 67.81; H, 8.24.
2.6 mmol) in dry methylene chloride (44 mL) was added
anhydrous sodium acetate (1.59 g, 19.4 mmol) and pyr-
idium chlorochromate (PCC) (3.18 g, 10.5 mmol). The
mixture was stirred at rt for 2 h. Additional PCC (1.59 g,
5.27 mmol) was then added and stirring was continued
for another 2 h. After this time, 2-propanol (6 mL) was
introduced and 15 min later, the mixture was diluted
with water and extracted with ether/ethyl acetate (1/1).
The organic phase was washed with water, 1 N aqueous
H2SO4, saturated aqueous NaHCO3, and brine. After
drying, the solution was evaporated and the residue
chromatographed on silica gel. Eluting with ethyl acet-
ate/hexanes (1/1) yielded compound 15 (560 mg, 78%
yield), a colorless glass. [a]2d5+35.3 ꢀ (c 0.36, CHCl3); IR
1
(CHCl3) 3600, 2225, 1709 cm
;
1H NMR (CDCl3) d
0.84 (s, 3H), 1.15 (d, J 8.0 Hz, 3H), 1.46 (s, 6H), 2.44 (m,
1H), 2.85 (m, 1H), 5.36 (brs, 1H); MS (m/e) 274 (M+).
[3aS-[3(S*),3aꢀ,7ꢀ,7aꢁ]]-3-[5-[[[(1,1-dimethylethyl)di-
methylsilyl]oxy]-1,5-dimethyl-3-hexyl]-3a,4,5,6,7,7a-hexa-
hydro-3a-methyl-1H-inden-7-ol acetate (13). A solution
of compound 12 (223 mg, 0.38 mmol) in toluene (15 mL)
was added dropwise over 1.5 h period to a preheated
(120 ꢀC) mixture of AIBN (5mg) and tri-n-butyltin
hydride (5g) under argon. The temperature was main-
tained at 120 ꢀ for 15 min. The reaction mixture was
cooled to rt and the toluene was removed under reduced
pressure. The residue was subjected to ¯ash chromato-
graphy with hexane/ethyl acetate (10/1) and the title pro-
duct was further puri®ed by HPLC with hexane/ethyl
acetate (10/1) to a€ord amorphous 13 (146 mg, 88%
yield). IR (CHCl3) 1730cm 1; 1H NMR (CDCl3) d 0.13 (s,
6H), 0.85 (s, 9H), 1.00 (s, 3H), 1.09 (m, 3H), 1.41 (s, 6H),
2.03 (s, 3H), 5.20 (brs, 1H), 5.36 (brs, 1H); Anal. calcd for
C27H48O3Si: C, 72.26; H, 10.78. Found C, 72.45; H, 10.91.
[3aR-[1(R*),3aꢀ,7aꢁ]]-1-[1,5-dimethyl-5-[(trimethylsilyl)-
oxy]-3-hexynyl]-3,3a,5,6,7,7a-hexahydro-7a-methyl-4H-
inden-4-one (6). To a solution of compound 15 (552 mg,
2.01 mmol) in dry dichloromethane (70 mL) was added
1-(trimethylsilyl)imidazole (2.00 g, 14.2 mmol). The
mixture was stirred at rt for 17 h and then was quenched
with water (22 mL). After stirring for an additional
20 min, the mixture was extracted with ethyl acetate and
the organic phase was washed with water and saturated
brine, dried, and evaporated to dryness. The residue was
chromatographed on silica gel eluting with ethyl acet-
ate/hexanes (1/4) to give compound 6 (693 mg, 99%
yield): [a]2d5+29.5 ꢀ (c 0.20, CHCl3); IR (CHCl3) 1710,
842 cm 1; 1H NMR (CDCl3) d 0.16 (s, 9H), 0.84 (s, 3H),
1.14 (d, J 8.0 Hz, 3H), 1.43 (s, 6H), 1.77 (m, 1H), 1.91
(m, 1H), 2.84 (m, 1H), 5.34 (brs, 1H); MS (m/e) 346
(M+).
[3aS-[3(S*),3aꢀ,7ꢀ,7aꢁ]]-3a,4,5,6,7,7a-hexahydro-3-(5-
hydroxy-1,5-dimethy-3-hexynyl)-3a-methyl-1H-inden-7-ol
(14). To a solution of compound 13 (283 mg, 0.65 mmol)
in THF (5 mL) was added tetrabutylammonium ¯uoride
(2 mL, 1 M in THF, 2 mmol) and the reaction mixture
was stirred overnight at rt. Then potassium hydroxide
(700 mg), ethanol (2 mL), and water (2 mL) were added
and the reaction mixture was heated 3 h at 80 ꢀC. After
dilution with water, the reaction mixture was extracted
with ethyl acetate. The extract was washed with water,
dried, ®ltered, and evaporated to dryness to give crude
product (245 mg). Further puri®cation by HPLC eluting
with hexanes/ethyl acetate (1/1) a€orded 14 (160 mg,
88% yield). Mp 107±109 ꢀC (dichloromethane±hexane);
[3aS-[3(S*),3aꢀ,7ꢀ,7(1S*,5S*),7aꢁ]]-3-[1,5-dimethyl-5-
[(trimethylsilyl)oxy]3-hexynyl]-3a,4,5,6,7,7a-hexahydro-
3a-methyl-7-[(2-methylenebicyclo[3.1.0]hex-1-yl)ethynyl]-
1H-inden-7-ol (16). A cold solution ( 30 ꢀC) of com-
pound 5 (306 mg, 1.61 mmol) in THF (15 mL) was trea-
ted dropwise over 10 min with n-BuLi (1.0 mL, 1.6 M in
hexane, 1.6 mmol). After the addition was complete, the
cooling bath was removed and the mixture was heated
at 45 ꢀC for 15 min. Then the solution of acetylide was
cooled to 78 ꢀC and a solution of ketone 6 (372 mg,
1.08 mmol) in THF (3 mL) was added dropwise. The
reaction mixture was stirred for 15 min and then was
allowed to warm to rt over ca. 30 min. Water was added,
followed by extraction with ethyl acetate. The extract
was washed with water and brine, dried, ®ltered, and the
solvent was removed under reduced pressure. The crude
product was puri®ed by column chromatography on
silica gel eluting with hexanes/ether (96/4) to give crude
[a]2d5+25.0 ꢀ (c 0.2, EtOH); H NMR (CDCl3) d 1.08 (s,
1
3H), 1.09 (d, J 8.0 Hz, 3H), 1.79 (s, 6H), 4.19 (brs, 1H),
5.39 (brs, 1H); Anal. calcd for C18H28O2: C, 78.21; H,
10.21. Found: C, 78.07; H, 10.22.
[3aR-[1(R*),3aꢀ,7aꢁ]]-3,3a,5,6,7,7a-hexahydro-1-(5-
hydroxy-1,5-dimethy-3-hexynyl)-7a-methyl-4H-inden-4-
one (15). To a solution of compound 14 (720 mg,
16 (453 mg, 92% yield), as a colorless oil, con-
taminated with unreacted ketone 6 (ca. 7%). Without

2058
M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
further puri®cation, crude compound 16 was used in the
next step. 1H NMR (CDCl3) d 0.17 (s, 9H), 1.01 (s, 3H),
1.08 (d, J 6.0 Hz, 3H), 1.44 (s, 6H), 4.87 (s, 1H), 5.11 (s,
1H), 5.39 (s, 1H).
t-BuOMe (3 mL) and photolyzed in a quartz tube under
argon with 450 W low-pressure mercury lamp and ura-
nium ®lter in the presence of 9-acetoxyanthracene
(1 mg). After 45 min, the solvent was removed and the
residue was puri®ed by column chromatography eluting
with hexanes/ethyl acetate (8/2) to give compound 3
(36 mg, 84% yield) as a glass. [a]2d5 62.5 (c 0.2, MeOH);
[3aS-[3(S*),3aꢀ,7ꢀ,7(1S*,5S*),7aꢁ]]-3a,4,5,6,7,7a-hexa-
hydro-3-(5-hydroxy-1,5-dimethyl-3-hexynyl)-3a-methyl-7-
[(2-methylenebicyclo[3.1.0]hex-1-yl)ethynyl]-1H-inden-7-ol
(17). To a solution of crude compound 16 (453 mg,
0.97 mmol) in THF (12 mL), tetra-n-butylammonium
¯uoride (1.5 mL, 1 M in THF, 1.50 mmol) was added.
The solution was stirred for 1 h at rt, whereupon it was
taken up by ethyl acetate, washed with water and brine,
then dried. The solvent was removed under reduced
pressure. The residue was puri®ed by chromatography
on silica gel while eluting with hexanes/ethyl acetate (9/
1) to yield compound 17 (363 mg, 95% yield) as a foam.
[a]2d5 48.1 ꢀ (c 0.32, CH2Cl2); IR (CHCl3) 3600, 2231,
1
IR (CHCl3) 3605, 1637 cm 1; H NMR (CDCl3) d 0.67
(s, 3H), 1.00 (s, 3H), 1.28 (s, 6H), 2.56 (m, 1H), 2.80 (m,
1H), 3.95 (brs, 1H), 4.82 (s, 1H), 4.98 (s, 1H), 5.24 (s,
1H), 5.41±5.56 (m, 2H), 6.04 (d, J 12.0 Hz, 1H), 6.16 (d,
J 12.0 Hz, 1H); MS (m/e) 396 (M+, 4%), 378 (12), 245
(5), 158 (19), 136 (19), 118 (23), 105 (14), 91 (27);
HRMS: for C27H40O2 (M+) calcd, 396.3028; found,
396.3020.
(1ꢀ,3ꢀ,5ꢀ)-3,5-cyclo-9,10-secocholesta-10(19),16-dien-6,
23-diyne-1,8,25-triol (19). A solution of compound 17
(42 mg, 0.10 mmol) in dichloromethane (5 mL) was
treated at +4 ꢀC with t-butyl hydroperoxide (107 mL,
3 M in isooctane, 0.32 mmol) and selenium dioxide
(4 mg, 0.03 mmol). The reaction mixture was monitored
by TLC (hexanes/ethyl acetate (1/1)). After 48 h, all the
starting material disappeared and the mixture of pro-
ducts (1-alcohol and 1-ketone) was extracted with ethyl
ether and the extracts washed with water. The extract
was dried and the solvents were removed under reduced
pressure. The residue was dissolved in THF (3 mL) and
treated at 60 ꢀC with Super-Hydride1 (0.2 mL, 1 M in
THF, 0.2 mmol). After 15 min, the icebath was removed
and the solution was extracted with ethyl acetate, the
extracts was washed with water and dried, and the sol-
vent was evaporated under reduced pressure. The resi-
due was puri®ed by chromatography on silica gel eluting
with hexanes/ethyl acetate (7/3) to give compound 19
1
1653 cm 1; H NMR (CDCl3) d 1.02 (s, 3H), 1.08 (d, J
6.1 Hz, 3H), 1.48 (s, 6H), 4.88 (s, 1H), 5.12 (s, 1H), 5.40
(s, 1H); MS (m/e) 375 (9%), 357 (12), 154 (100), 136
(84), 115 (91); HRMS for C27H36O2 (M+) calcd,
392.2715; found, 392.2707. Anal. calcd for C27H36O2: C,
82.61; H, 9.24. Found: C, 82.96; H, 9.40.
[3aS-[3(1S*,3E),3aꢀ,7ꢀ,7[E(1R*,5S*),7aꢁ,]]-3a,4,5,6,7,
7a-hexahydro-3-(5-hydroxy-1,5-dimethyl-3-hexenyl)-3a-
methyl-7-[2-(2-methylenebicyclo[3.1.0]hex-1-yl)ethenyl]-
1H-inden-7-ol (18). To a solution of compound 17
(60 mg, 0.15 mmol) in THF (3 mL) was added LiAlH4
(20 mg, 0.53 mmol) and the reaction mixture was stirred
at 55 ꢀC for 5 h. After cooling to rt, saturated NH4Cl
(0.2 mL) was carefully added, and the reaction mixture
was extracted with ethyl acetate. The extract was
washed with water and brine, then dried. After eva-
poration of solvent the residue was puri®ed by chroma-
tography on silica gel eluting with hexanes/ethyl acetate
(8/2) to a€ord compound 18 (52 mg, 86% yield) as a
colorless oil. [a]d25 16.1 (c 0.41, CH2Cl2); IR (CHCl3)
1
(32 mg, 76% yield) as a colorless oil. H NMR (CDCl3)
d 0.90 (t, J 5.0 Hz, 1H), 1.02 (s, 3H), 1.08 (d, J 6.0 Hz,
3H), 1.48 (s, 6H), 4.17 (t, J 8.0 Hz, 1H), 5.17 (d, J
2.0 Hz, 1H), 5.30 (d, J 2.0 Hz, 1H), 5.40 (s, 1H); MS (m/
e) 408 (M+, 1%), 390 (1), 375 (4), 347 (3), 265 (7), 199
(35), 159 (34), 115 (43), 105 (58), 91 (88), 55 (57), 43
(100); HRMS: for C27H36O3 calcd, 408.2664; found
408.73.
1
3601 cm 1; H NMR (CDCl3) d 0.83 (t, J 4.6 Hz, 1H),
0.98 (d, J 6.4 Hz, 1H), 1.03 (s, 3H), 1.48 (s, 6H), 4.76 (s,
1H), 4.81 (s, 1H), 5.30 (s, 1H), 5.45 (d, J 14.8 Hz, 1H),
5.54±5.66 (m, 2H), 6.02 (d, J 14.8 Hz, 1H).
(3ꢁ,5Z,7E,23E)-9,10-Secocholesta-5,7,10(19),16,23-pen-
aene-3,25-diol (3). To a solution of compound 18
(43 mg, 0.11 mmol) in dioxane/water (8/2, 3 mL), p-
TsOH (catalytic amount, 1 mg) was added. The reaction
mixture was stirred at 60 ꢀC under argon for 2 h. The
solution was extracted with ethyl acetate and the
organic phase was washed with water and brine; then
dried. The solvent was evaporated to dryness and the
residue was passed through silica gel while eluting with
hexanes/ethyl acetate (8/2). Fractions containing mix-
tures of cis/trans-isomers were combined and the sol-
vents were removed. The residue was then dissolved in
(1ꢀ,3ꢀ,5ꢀ,6E,23E)-3,5-cyclo-9,10-secocholesta-6,10(19),
16,23-tetraene-1,8,25-triol (20). A stirred solution of
compound 19 (22 mg, 0.054 mmol) in THF (4 mL) was
treated with LiAlH4 (15 mg, 0.40 mmol) and the reac-
tion mixture was heated at 65 ꢀC for 16 h. Then it was
cooled to 0 ꢀC and carefully treated with saturated
NH4Cl solution, whereupon it was dissolved in ethyl
acetate, washed with water, dried and ®ltered. The sol-
vent was removed under reduced pressure and the oily
residue was puri®ed by passing through a short path of
silica gel eluting with hexanes/ethyl acetate (6/4) to
a€ord compound 20 (16 mg, 72% yield) as a colorless

M. M. Kabat et al./Bioorg. Med. Chem. 6 (1998) 2051±2059
2059
oil. 1H NMR (CDCl3) d 0.72 (t, J 4.8 Hz, 1H), 0.99 (d, J
6.2.0 Hz, 3H), 1.03 (s, 3H), 1. 29 (s, 6H), 4.26 (q, J
8.2 Hz, 1H), 4.98 (d, J 2.0 Hz, 1H), 5.10 (d, J 2.0 Hz,
1H), 5.31 (s, 1H), 5.48 (d, J 15.6 Hz, 1H), 5.55±5.63 (m,
2H), 6.04 (d, J 15.6 Hz, 1H).
7. Prince, C. W.; Butler, W. T. Collagen Relat. Res. 1987, 7,
305.
8. Franceschi, R. T.; Roman, P. R.; Park, K.-Y. J. Biol. Chem.
1988, 263, 18938.
9. Row, D. W.; Kream, B. E. J. Biol. Chem. 1982, 257, 8009.
10. Lieberherr, M. J. Biol. Chem. 1987, 262, 13168.
(1ꢀ,3ꢁ,5Z,7E,23E)-9,10-secocholesta-5,7,10(19),16,23-
pentaene-1,3,25-triol (4). A solution of compound 20
(12 mg, 0.012 mmol) and pyridinum p-toluenesulfonate
(PPTS) (1 mg) in acetonitrile/water (7/3, 3 mL) was stir-
red under argon at 65 ꢀC for 3 h. Then the reaction
mixture was dissolved in ethyl acetate and washed with
sodium bicarbonate solution, then dried. After eva-
poration of the solvent, an isomeric mixture of 5,6-cis/
trans compounds (12 mg) was obtained. 1H NMR
(CDCl3) d 0.68 and 0.70 (2s, 3H), 1.02 (d, J 6.3 Hz, 3H),
1.30 (s, 6 H), 4.25 (m, 1 H), 4.49 (m, 1 H), 4.99, 5.02,
5.14, 5.34 (4s), 5.55±5.59 (m), 5.99 and 6.58 (2d, J
11.7 Hz, =CH± of 5,6-trans-compound), 6.11 and 6.38
(2d, J 10.8 Hz, =CH± of 5,6- cis compound). The mix-
ture of 5,6-cis/trans compounds was subsequently irra-
diated for 30 min in t-BuOMe (3 mL), in the presence of
9-acetoxyanthracene (1 mg), with a 450 W low-pressure
mercury lamp through a uranium ®lter. The solvent was
removed and the residue was puri®ed by passing
through a short path of silica gel. Elution with hexanes/
ethyl acetate (7/3) a€orded pure compound 4 (10 mg,
83% yield) as a colorless glass. IR (CHCl3) 3605,
11. Guggino, S. E.; Lajeunesse, D.; Wagner, J. A.; Snyder,
S. H. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 2957.
12. Bidmon, H.-J.; Gutkowska, J.; Murakami, R.; Stumpf,
W. E. Experientia 1991, 47, 958.
13. Morelli, S.; DeBoland, A. R.; Boland, L. J. Biochem. 1993,
289, 675.
14. Baran, D. T.; Sorensen, A. M.; Shalhoub, V.; Owen, T.;
Stein, G.; Lian, J. J. Cell Biochem. 1992, 50, 124.
15. Tien, X.-Y.; Brasitus, T. A.; Qasawa, B. M.; Norman,
A. W.; Sitrin, M. D. Am. J. Physiol. 1993, 265, 143.
16. Roy, H. K.; Bissonnette, M.; Wali, R. Digestive Disease
Week; New Orleans, LA, 1994; Abstract 1213.
17. Nemere, I.; Yoshimoto, Y.; Norman, A. W. Endo. Soc.
1984, 115, 1476.
18. Yoshimoto, Y.; Norman, A. W. Steroid Biochem. 1986, 25,
905.
19. Zhou, L.-X.; Nemere, I.; Norman, A. W. Bone Mineral
Res. 1992, 7, 457.
20. Farach-Carson, M. C.; Sergeev, I.; Norman, A. W. Endo-
crine 1991, 129, 1876.
21. Ca€rey, J. M.; Farach-Carson, M. C. J. Biol. Chem. 1989,
264, 20265.
1
1652 cm 1; H NMR (CDCl3) d 0.68 (s, 3H), 1.02 (d, J
22. Yakihiro, S.; Posner, G. H.; Guggino, S. E. J. Biol. Chem.
1994, 269, 23889.
6.3 Hz, 3H), 1.30 (s, 6H), 2.54±2.68 (m, 1H), 2.76±2.88
(m, 1H), 4.25 (m, 1H,), 4.45 (m, 1H), 5.02 (s, 1H), 5.33
and 5.34 (two overlapped s, 2H), 5.55±5.68 (m, 2H),
6.11 (d, J 11.1 Hz, 1H), 6.38 (d, J 11.1 Hz, 1H); MS (m/e)
394 (8%), 376 (9), 358 (14), 343 (7), 105 (100).
23. Ferrara, J.; McCuaig, K.; Hendy, G. N.; Uskokovic, M.;
White, J. H. J. Biol. Chem. 1994, 269, 2971.
24. Kabat, M. M.; Kiegiel, J.; Cohen, N.; Toth, K.; Wovkulich,
P. M.; Uskokovic, M. R. Tetrahedron Lett. 1991, 32, 2343.
25. Kabat, M. M.; Kiegiel, J.; Cohen, N.; Toth, K.; Wovkulich,
P. M.; Uskokovic, M. R. J. Org. Chem. 1996, 61, 118.
References and Notes
26. Hajos, Z. G.; Parrish, D. R. Org. Synth. 1985, 63, 26.
27. (a) Baggiolini, E. G.; Iacobelli, J. A.; Hennessy, B. M.;
Uskokovic, M. R. J. Am. Chem. Soc. 1982, 104, 2945. (b)
Daniewski, A. R.; Kiegiel, J. J. Org. Chem. 1988, 53, 5534.
1. Darwish, H.; DeLuca, H. F. Crit. Rev. Eukarot. Gene. Exp.
1993, 3, 89.
2. Lian, J. F.; Stein, G. S. J. Cell Biochem. 1992, 49, 37.
3. Minghetti, P. P.; Norman, A. W. FASEB J. 1988, 2, 3034.
28. Baggiolini, E. G.; Hennesy, B. M.; Shiuey, S. J.; Truitt,
G. A.; Uskokovic, M. R. Eur. Pat. Appl. EP 398, 217; Chem.
Abstr. 1991, 115, 150403.
4. Noda, M.; Yoon, K.; Rodan, G. A. J. Biol. Chem. 1988,
263, 18574.
29. Okabe, M.; Sun, R.-C.; Scalone, M.; Jibilian, C.; Hutchings,
S. D. J. Org. Chem. 1995, 60, 767.
5. Price, P. A.; Baukol, S. A. J. Biol. Chem. 1980, 255, 11660.
6. Noda, M.; Yoon, K.; Prince, C. W.; Butler, W. T.; Rodan,
G. A. J. Biol. Chem. 1988, 263, 13916.
30. For reviews see: (a) Ramaiah, M. Tetrahedron 1987, 43,
3541. (b) Hartwig, W. Tetrahedron 1983, 39, 2609.