Conformationally restricted hybrid analogues of the hormone 1α,25-dihydroxyvitamin D3: Design, synthesis, and biological evaluation

Article abstract of DOI:10.1016/S0968-0896(01)00087-6

Four new conformationally restricted hybrid analogues of the hormone 1α-25-dihydroxyvitamin D₃ (1,25D3) have been synthesized in a convergent manner by combining enantiomerically pure C,D-ring ketones ( - )-15 and ( - )-17 with racemic 1-hydrox

Full text of DOI:10.1016/S0968-0896(01)00087-6

Bioorganic & Medicinal Chemistry 9 (2001) 1691–1699
Conformationally Restricted Hybrid Analogues of the Hormone
1ꢀ,25-Dihydroxyvitamin D3:Design, Synthesis, and Biological
Evaluation
M. Christina White,a,* Martin D. Burke,b,* Sara Peleg,c Henry Bremb
and Gary H. Posnera
aDepartment of Chemistry, The Johns Hopkins University, Baltimore, MD 21218, USA
bDepartment of Neurosurgery, The Johns Hopkins School of Medicine, Baltimore, MD 21218, USA
cDepartment of Medical Specialities, The University of Texas, M.D. Anderson Cancer Center, Houston, TX, 77030, USA
Received 21 November 2000; accepted 22 February 2001
Abstract—Four new conformationally restricted hybrid analogues of the hormone 1a-25-dihydroxyvitamin D3 (1,25D3) have been
synthesized in a convergent manner by combining enantiomerically pure C,D-ring ketones (À)-15 and (À)-17 with racemic 1-
hydroxymethyl A-ring phosphine oxide (Æ)-18. Parent hybrid analogue 6, which combines the calcemia-inactivating 1b-hydro-
xymethyl A-ring modification with the antiproliferation- activating 20-epi-22-oxa-25-hydroxydiethyl C,D-ring side chain modifica-
tion, is comparable in potency to 1,25D3 at the low nM level in inhibiting proliferation in a wide assortment of malignant cell lines
in vitro with extremely low calcemic activity in vivo. Surprisingly, both conformationally restricted analogues of 6 (8b and 9b),
which incorporate rigidifying units at their 25-hydroxyl side chain termini, retained the desirable antiproliferative, transcriptional,
and calcemic activities of the parent compound. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
Since the molecular modes of action mediating 1,25D3’s
In addition to its ‘classical’ physiological role in the
regulation of calcium and phosphorus homeostasis,1 the
hormonally active metabolite of vitamin D3, 1a,25-
dihydroxyvitamin D3 (1,25D3, 1) is known to con-
tribute significantly to the growth and differentiation of
a variety of cell types (e.g., keratinocytes from epi-
dermal precursors,2 osteoclasts from mononuclear pre-
cursors3). At supraphysiological concentrations, the
growth regulatory abilities of 1,25D3 have proven
effective at suppressing proliferation and inducing dif-
ferentiation in psoriatic cells4 and in a variety of malig-
nant cell lines both in vitro and in vivo.5 The
therapeutic usefulness of 1,25D3 when administered at
pharmacological doses to treat such hyperproliferative
diseases as cancer or psoriasis is limited, however, by
the causation of toxic hypercalcemia that results from
its potent effects on calcium metabolism.6
diverse physiological effects are not yet well defined,7
rationale for the design of analogues of 1,25D3 in which
the calcium regulating activity is effectively separated
from the regulation of cell growth has been largely
based on information obtained from structure–activity
relationships (SARs).8 These studies attempt to define
the influence of specific structural unit modifications on
1,25D3’s various biological activities. The high con-
formational flexibility of 1,25D3, however, suggests that
its diverse spectrum of biological activities may in part
be induced by different topologies (shapes) of the hor-
mone.8,9 One reasonable hypothesis is that different
forms of 1,25D3 interact selectively with different pro-
teins (e.g., receptors, transcription factors) ultimately
resulting in the variety of biological responses
observed.9 Many of 1,25D3’s biological activities result
from transcriptional modifications initiated by the hor-
mone binding to the nuclear vitamin D receptor (VDR).
The selectivity in transcriptional responses observed
with many of the structural analogues is thought to be
the result of ligand induction of distinct conformational
changes in the VDR dimerization interfaces that reg-
ulate the selection of dimerization partners such as
*Corresponding authors at current address: Department of Chemistry
and Chemical Biology, Harvard University, 12 Oxford Street, Cam-
bridge, MA 02138, USA. Tel.: +1-617-496-1806; fax: +1-617-496-1880;
Tel.: +1-617-495-8394; fax: +1-617-495-0751; e-mail: white@ascat.
harvard.edu; burke@slsiris.harvard.edu
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00087-6

1692
M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
nuclear vitamin D response elements (VDRE), co-acti-
vators, and other transcriptional effectors.7 If this is the
case, conformation–activity relationships, particularly
of 1,25D3 analogues that have already demonstrated a
promising separation between antiproliferative and cal-
cemic activities, would strongly augment the informa-
tion obtained from SARs for the design of analogues
with selective physiological action (Fig. 1).
lines B-16 (malignant melanoma), EMT-6 (breast cell
carcinoma), and RENCA (renal cell carcinoma).18
Conversely, hybrid analogue 5, differing only in relative
stereochemistry at the 1 and 3 positions [i.e., 5 (1a,3b)
versus 6 (1b,3a)], showed no significant antiproliferative
activity, even in the mM range. Curiously, hybrid ana-
logue 6 also demonstrated a marked increase in vitamin
D receptor (VDR) mediated transcriptional activity
relative to 4 which was not associated with increased
binding affinity to the VDR.19
The majority of structural modifications capable of
substantially enhancing the antiproliferative potency of
1,25D3 appear at the steriodal D-ring side chain. Out-
standing examples of beneficial individual modifications
include the following: (1) methyl homologation at C-26
and C-27, (2) homologation at C-24, (3) incorporation
of an oxygen atom at C-22, (4) inversion of stereo-
chemistry at C-20.8 Side-chain analogues of 1,25D3 that
incorporate rigid, conformationally restricting struc-
tural units tend to show moderate antiproliferative
activities with low calcemic activities. Outstanding
examples include side-chain aromatic analogues (aro-
calciferols),10 23-yne analogues (e.g., Hoffmann-La
Roche’s Ro-23-755311) and Leo Pharmaceutical Com-
pany’s 22,24-diene analogue EB 1089 3.12,13 Leo Phar-
maceutical’s analogue KH 1060 2,14,15 which
incorporates many of the aforementioned structural
side-chain modifications, has extraordinarily enhanced
abilities to induce differentiation and inhibit prolifera-
tion relative to 1,25D3 in a variety of skin and malig-
nant cell lines, but retains undesirable calcemic activity.
When these antiproliferation activating structural mod-
ifications were incorporated into the side chain of anti-
The side-chain terminus, via the 25-hydroxyl moiety, is
thought to play a crucial role in the binding of 1,25D3
to the VDR and in eliciting its biological effects.20
Inversion of stereochemistry at C-20 is thought to bias
the conformation of the side chain such that the term-
inal 25-hydroxyl moiety resides primarily on the
‘northwest’ portion of the molecule (i.e., above the C-
ring).21,22 In a conformation–activity study performed
by the researchers at Leo Pharmaceuticals, compound 7,
in which the KH 1060 side chain was sterically restricted
over the ‘northwest’ portion of the molecule, showed
diminished antiproliferative activity with respect to the
parent compound but enhanced activity with respect to
the natural hormone 1,25D3.23 The biological activity
was thought to suggest the presence of a 25-hydroxyl
moiety binding site to the northwest of the C,D-ring
system. As previously indicated, several structural
modifications made directly at the side chain terminus
encasing the critical 25-hydroxyl group (i.e., one carbon
homologation at C-24, C-26, and C-27) in KH 1060 and
hybrid analogues 5 and 6 are known to potentiate the
antiproliferative effects of 1,25D3. Although it is evident
that these structural modifications increase the length
and hydrophobicity of the side chain, it is not clear what
role, if any, the conformation of this extended terminus
plays on the biological profile of analogues that incor-
porate it. To evaluate this, we have designed, synthe-
sized, and biologically evaluated two sets of
calcemic 1b-(hydroxymethyl) A ring analogue 4,16
a
dramatic enhancement of antiproliferative activity was
observed with no increase in calcemic activity.17 Extra-
ordinarily, hybrid analogue 6 was equipotent to 1,25D3
at the low nM level in inhibiting proliferation in human
malignant cell lines SK Br3 (breast cell carcinoma), HL-
60 (myeloid leukemia cells);17 murine malignant cell
Figure 1. Structure of 1a, 25-dihydroxyvitamin D3 and its analogues.

M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
1693
conformationally restricted analogues of 5 and 6, ana-
logues 8 and 9. Analogues 8 and 9 both incorporate the
floppy terminal C-25 diethyl groups into a con-
formationally rigid cyclohexyl ring. Analogue 9 further
restricts conformational flexibility at the side-chain ter-
minus by incorporating a rigidifying alkyne unit
between C-24 and C-24a.
by fluoride promoted desilylation of the C-8 alcohol in
85% yield. Formation of the bis-lithiated derivative of
(+)-11 (2.5 eq n-BuLi, THF, À50 ꢀC) followed by con-
densation with cyclohexanone and an aqueous work up
gave the desired key intermediate diol (+)-12 in 90%
yield. When the reaction was run at À78 ꢀC, the bis-
lithiated species precipitated out of solution and the
condensation yields dropped to 60%. Attempts to desi-
lylate the C-8 secondary alcohol after condensation
gave very poor yields (20–30%) of diol (+)-12. Initial
attempts to couple a pre-assembled primary tosylate
side chain to the 20-epi alcohol (+)-10 via a Williamson
ether coupling reaction gave no desired product with
70% recovery of (+)-10 and complete consumption of
side chain.
Results
Chemistry
We have prepared enantiomerically pure C,D-ring
ketones (À)-15 and (À)-17 for Horner–Wadsworth–
Emmons (HWE) coupling with racemic 1-(hydroxy-
methyl) A-ring phosphine oxide (Æ)-18 in a convergent
approach to conformationally restricted hybrid analo-
gues 8 and 9. A retrosynthetic analysis for ketones (À)-
15 and (À)-17 suggests diol (+)-12 as a suitable inter-
mediate for both. As shown in Scheme 1, 20-epi alcohol
(+)-10, prepared according to literature methods from
commercial vitamin D2,17 was converted into its pro-
pargyl ether derivative (+)-11 by alkylation of the
potassium alkoxide with propargyl bromide (13% yield,
80% recovered alcohol, ca. 54% over 11 runs) followed
Transformation of diol (+)-12 into saturated O-sily-
lated C,D-ring ketone (À)-15 proceeded smoothly in
three steps. Catalytic hydrogenation (Pd/C) of (+)-12 in
ethanol provided fully saturated diol (À)-13 in 50%
yield. Oxidation of the C-8 hydroxyl group by tetra-
propylammonium perruthenate (TPAP) and N-methyl-
morpholine N-oxide (NMO) in the presence of
molecular sieves (MS) provided ketone (À)-14 in 90%
yield. Silyl protection of the tertiary alcohol in prepara-
tion for HWE coupling was effected with 1-(tri-
methylsilyl)-imidazole (TMS-imid.) in CH2Cl2 to afford
the O-silylated product (À)-15 in 97% yield (Scheme 2).
Convergent HWE coupling of enantiomerically pure
(À)-15 with racemic 1-(hydroxymethyl) A ring phos-
phine oxide (Æ)-18 (64%), followed by fluoride-pro-
moted desilylation (75%), gave diastereomeric seco-
steriods (À)-8a (1a,3b) and (À)-8b (1b,3a) that were
easily separated by C-18 reverse-phase HPLC. The
absolute stereochemistry at C-1 and C-3 in these dia-
stereomers was tentatively assigned primarily by com-
paring characteristic portions of their 400 MHz 1H
NMR spectra and optical rotation data with those of
the corresponding 1-(hydroxymethyl) analogues (see
Table 1).16,17
Scheme 1.
ð1Þ
Elaboration of diol (+)-12 into O-silylated unsaturated
C,D-ring ketone (À)-17 proceeded smoothly in two
steps. Diol (+)-12 underwent TPAP oxidation at the C-
8 position to afford unsaturated ketone (À)-16 in 90%
yield which was silylated at the tertiary C-25 alcohol to
provide saturated, O-silylated ketone (À)-17 in 90%
yield (Scheme 3).
Scheme 2.

1694
M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
cellular proliferation than their diastereomeric coun-
terparts bearing the 1b,3a stereochemistry (i.e., 6, 8b,
and 9b). Surprisingly, despite the conformation restric-
tions imposed on the side-chain termini of both 8b and
9b, the desirable antiproliferative activities of these
analogues were similar to those of the parent hybrid
compound 6 at drug concentrations ranging from 1 to
1000 nM (Fig. 2). Even at physiologically relevant 1 and
10 nM concentrations, conformationally restricted
hybrids 8b and 9b, like hybrid 6, were at least as potent
antiproliferative agents as 1,25D3 (1). The average
standard deviation for the data points shown in Figure
1 (n=3 for each) was Æ10% of control.
To further characterize their biological profiles, the in
vitro VDR-mediated transcriptional activities of con-
formationally restricted analogues 8b and 9b were tested
in rat osteosarcoma ROS 17/2.8 cells using our pre-
viously described protocol.19 Reminiscent of the anti-
proliferative activity trends, the transcriptional
potencies of both 8b and 9b paralleled those of parent
compound 6 at drug concentrations ranging from 0.01
to 100 nM (Fig. 3). Both the saturated analogue 8b
(ED50=2Â10À10 M) and unsaturated analogue 9b
(ED50=1Â10À10 M), like 6 (ED50=1Â10À10 M) were at
least as transcriptionally potent as 1,25D3
(ED50=3Â10À10 M).
Scheme 3.
Convergent HWE coupling of enantiomerically pure
(À)-17 with racemic 1-(hydroxymethyl) A ring phos-
phine oxide (Æ)-18 (68%), followed by fluoride-pro-
moted desilylation (70%) gave diastereomeric seco-
steriods (À)-9a (1a,3b) and (À)-9b (1b,3a) that were
easily separated by C-18 reverse-phase HPLC and
characterized as described above (see Table 1).
Having established that conformationally restricted
hybrid analogues 8b and 9b have antiproliferation and
VDR-mediated transcriptional potencies paralleling
those of parent hydrid 6, we were interested in evaluat-
ing their calcemic activities in vivo. In contrast to the
natural hormone 1,25D3, which produced marked ele-
vations in blood ionized calcium levels in female Bl-6
mice after 7 days of treatment at doses of 1 and 10 mg/
kg/day, both 8b and 9b, like 6, produced no calcium
elevation above controls even at 100 mg/kg/day, a dose
which was lethal for 1,25D3 (Fig. 4). In addition, the
dose dependent weight loss seen with 1,25D3 was not
observed with any of these hybrid analogues (data not
shown).
ð2Þ
Table 1. Characteristics of hybrid analogues
1H NMR (d)
Analogue
C-18
C-19a
C-19b
[a]2D5(ꢀ)
5
6
8a
8b
9a
9b
0.55
0.53
0.55
0.52
0.62
0.59
5.16
5.15
5.16
5.14
5.17
5.15
5.01
4.99
5.01
4.98
5.02
4.99
À140
+4.0
À149
À1.3
À182
À5.0
Biology
Conformationally restricted analogues 8 and 9, hybrid
analogue 6, and the natural hormone 1,25D3 (1) were
all evaluated for in vitro antiproliferative activity in
murine B16 malignant melanoma cells using our pre-
viously described protocol.19 As predicted based on
previous studies of diastereomeric sets of hybrid analo-
gues, the diastereomers with the 1a,3b stereochemistry
(i.e., 8a and 9a) were much less potent inhibitors of
Figure 2. Dose–response effects of 1,25D3 and its analogues on mur-
ine B16 malignant melanoma cell proliferation. The average standard
deviation for data points (n=3 for each) was Æ10% of control.

M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
1695
increase an analogue’s in vivo half-life without, as we
have demonstrated herein, diminishing the analogue’s
desired biological activities.
Conclusion
Four new conformationally restricted hybrid analogues
of the hormone 1a-25-dihydroxyvitamin D3 (1,25D3)
have been synthesized in a convergent manner by com-
bining enantiomerically pure C,D-ring ketones (À)-15
and (À)-17 with racemic 1-hydroxymethyl A-ring phos-
phine oxide (Æ)-18. Both conformationally restricted
analogues of 6 (8b and 9b), which incorporate rigidify-
ing units at their 25-hydroxyl side-chain termini,
retained the desirable antiproliferative, transcriptional,
and calcemic activities of the parent compound. These
findings may have pharmacologically favorable meta-
bolic implications for the future design of clinically
relevant 1,25D3 analogues.
Figure 3. Transcriptional activation in rat osteosarcoma ROS, 17/2.8
cells in reponse to treatment with 1,25D3, 6, and conformationally
restricted hybrid analogues 8b and 9b.
Experimental
General
Tetrahydrofuran (THF) and diethyl ether (Et2O) were
distilled from benzophenone ketyl prior to use. Methyl-
ene chloride (CH2Cl2) and triethylamine (NEt3) were
distilled from calcium hydride prior to use. Commer-
cially available anhydrous solvents were used in other
instances. All reagents were purchased from Aldrich
Chemical Co. (Milwaukee, WI, USA) and were used as
received without further purification. FT-IR and UV
spectra were recorded using a Perkin-Elmer Model 1600
FT-IR spectrophotometer and a Beckman Du-70 spec-
trophotometer, respectively. The 1H and 13C NMR
spectra were recorded on a Varian XL-400 spectrometer
operating at 400 and 100 MHz, respectively. Chemical
shifts are expressed in parts per million downfield from
tetramethylsilane. Unless otherwise indicated, all reac-
tions were run under an Ar atmosphere. The purity of
products was judged to be at least 95% on the basis of
their chromatographic homogeneity.
Figure 4. Ionized calcium levels in murine blood following 7 days of
intraperitoneal injection of 1,25D3, 6, and conformationally restricted
hybrid analogues 8b and 9b at 1, 10, and 100 ug/kg/day. Values repre-
sent the meanÆstandard deviation for three animals in each group.
Discussion
Hybrid analogues 8b and 9b, in which conformationally
rigidifying units have been incorporated into their side-
chain terminus, displayed a highly favorable biological
profile similar to that of parent hybrid analogue 6. The
similarity in antiproliferative, transcriptional, and cal-
cemic activities of 9b and 6 are especially noteworthy
since the presence of an acetylenic unit in 9b’s side-chain
terminus severely distorts the topology of the terminus
away from one that could be adopted by 6 via s bond
rotation. On the basis of these preliminary studies, it
appears as if the topology of the side-chain terminus,
thought to be crucial for the analogue’s binding to the
VDR and eliciting a biological response, does not mea-
surably influence the selective biological activities of
hybrid analogues of 1,25D3 described in this study.
These findings may prove significant in the future design
of clinically relevant 1,25D3 analogues. For example,
the bioavailabilty of compounds like 6 has been shown
to be limited by metabolic hydroxylation at C-24 by the
renal P-450 enzyme 24-hydroxylase. Incorporation of
the acetylenic unit between C-24 and C-24a might
interfere with this metabolic hydroxylation and thus
Acetylenic alcohol (+)-11. To a solution of 1.22 g (3.91
mmol) of alcohol (+)-10 dissolved in 15.0 mL of anhy-
drous THF was added 5.3 mL (5.3 mmol) of potassium
t-butoxide (1.0 M solution in THF). The mixture was
allowed to stir for 1 h at which time 1.4 g (5.3 mmol) of
18-crown-6 dissolved in 5.0 mL of anhydrous THF was
added via cannula. The mixture was allowed to stir for
an additional hour at which time 0.9 mL of propargyl
bromide (80 wt.% solution in toluene) was added
dropwise via syringe. Immediately upon addition of the
propargyl bromide, a deep purple color developed and
persisted. The mixture was allowed to stir for 12 h at
which time it was quenched with 10.0 mL of water,
extracted with EtOAc (3Â, 50 mL), dried over MgSO4,
filtered, concentrated, and purified by silica gel chro-
matography (10% EtOAc/hexanes) to afford 182.0 mg
(0.52 mmol) of propargylic ether in 13% yield and 972.0
mg (3.11 mmol) of recovered alcohol (+)-10 in 80%

1696
M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
yield. This procedure was repeated 11 times to afford
589 mg (1.68 mmol) of the desired propargylic ether in a
calculated 54% yield. An oven-dried 50.0 mL round
bottomed flask equipped with a magnetic stirring bar
was charged with 695 mg (1.98 mmol) of propargylic
ether, 30.0 mL of anhydrous THF, 700.0 mg of acti-
vated 4 A powdered molecular sieves (MS), and 1.04 g
(4.0 mmol) of solid tetrabutylammonium fluoride
hydrate (n-Bu4NF, TBAF). The reaction mixture was
refluxed for 2 days during which time additional TBAF
(2.44g, 9.3 mmol) and molecular sieves (1.1 g) were
added. The reaction mixture was cooled to room tem-
perature, concentrated by rotary evaporation, and pur-
ified by silica gel chromatography (10% EtOAc/
hexanes) to afford 368.0 mg (1.6 mmol) of acetylenic
alcohol (+)-11 in 80% yield: [a]2D8 +1.1ꢀ (c 2.7, EtOAc);
1H NMR (CDCl3) d 4.17 (dd, J=2.4, 15.6 Hz, 1H), 4.07
(m, 1H), 4.06 (dd, J=2.4, 15.6 Hz, 1H), 3.54 (m, 1H),
2.37 (t, J=2.4 Hz), 2.13 (m, 1H), 1.07 (d, J=5.6 Hz,
3H), 0.98 (s, 3H); 13C NMR (CDCl3) d 80.46, 76.21,
73.67, 69.18, 56.74, 54.64, 52.05, 41.69, 40.01, 33.78,
24.66, 22.55, 17.63, 17.34, 14.05, IR (CHCl3, cmÀ1) 3619,
3307, 3008, 2935, 2872, 1456; HRMS m/z (M+NH4+)
filtrate was evaporated and the residue was purified by
silica gel chromatography (30% EtOAc/hexanes) to
afford 47.0 mg (0.14 mmol) of saturated diol (À)-13 in
52% yield: [a]2D8 À9.3ꢀ (c 1.2, EtOAc); 1H NMR
(CDCl3) d 4.08 (m, 1H), 3.56 (m, 1H), 3.26 (m, 2H),
2.20 (bs, 1H), 2.12 (m, 1H), 1.07 (d, J=6.0 Hz, 3H),
0.94 (s, 3H); 13C NMR (CDCl3) d 77.61, 70.72, 69.22,
68.83, 56.84, 52.06, 41.67, 40.21, 37.98, 37.17, 33.76, 25.89,
24.72, 23.78, 22.63, 22.34, 22.25, 18.10, 17.35, 14.27; IR
(CHCl3, cmÀ1) 3378, 2935, 2870, 1455, 1374; HRMS m/z
(M+) calcd 338.2821 for C21H38O3, found 338.2826.
Saturated ketone (À)-14. Solid tetrapropylammonium
perruthenate (TPAP, 5.0 mg, 0.01 mmol) was added in
one portion to a stirring mixture of diol (À)-13 (47.0
mg, 0.14 mmol), 4-methyl morpholine N-oxide (NMO,
41.0 mg, 0.35 mmol), and 4 A MS (100.0 mg) in 10.0
mL of anhyd CH2Cl2. Upon completion, the reaction
mixture was diluted with a 50% EtOAc/hexanes solu-
tion and filtered through a silica gel plug. The filtrate
was evaporated and the residue was purified by silica gel
chromatography (30% EtOAc/hexanes) to afford 45.0
mg (0.13 mmol) of saturated ketone (À)-14 in 93%
yield: [a]2D8 À42.0ꢀ (c 0.81, EtOAc); 1H NMR (CDCl3) d
3.57 (m, 1H), 3.24 (m, 2H), 2.46 (m, 1H), 1.1 (d, J=6.0
Hz, 3H), 0.64 (s, 3H); 13C NMR (CDCl3) d 212.09,
77.55, 70.81, 68.90, 61.43, 56.76, 49.86, 41.15, 38.88,
37.85, 37.32, 25.85, 25.02, 24.16, 23.75, 22.31, 22.25,
19.33, 18.18, 13.04; IR (CHCl3, cmÀ1) 3392, 2937, 2876,
2859, 1705, 1451, 1381; HRMS m/z (M+) calcd
336.2664 for C21H36O3, found 336.2669.
+
ꢁ
calcd 254.2120 for C15H24O2 NH4 , found 254.2123.
Acetylenic diol (+)-12. An oven dried 25 mL round
bottomed flask equipped with a magnetic stirring bar
was charged with 198 mg (0.84 mmol) of acetylenic
alcohol (+)-11 and 10 mL of anhydrous THF. The
solution was cooled to À50 ꢀC and 1.3 mL (2.08 mmol)
of n-BuLi (1.6 M solution in hexanes) was added. The
solution was allowed to stir for 30 min at which time 0.2
mL (1.9 mmol) of cyclohexanone was added via syringe.
Additional n-BuLi (1.2 mL, 1.9 mmol) followed by
additional cyclohexanone (0.2 mL, 1.9 mmol) was
added to the solution until the disappearance of all
starting material was observed by thin layer chromato-
graphy (TLC). The reaction mixture was quenched at
À50 ꢀC with water, allowed to slowly warm up to room
temperature, extracted with EtOAc (3Â, 50 mL), dried
over MgSO4, filtered, concentrated, and purified by
silica gel chromatography (30% EtOAc/hexanes) to
yield 250 mg (0.75 mmol) of the desired acetylenic diol
O-Silylated saturated ketone (À)-15. An oven dried 25.0
mL round bottomed flask was charged with 45.0 mg
(0.13 mmol) of tertiary alcohol (À)-14 and 10.0 mL of
anhydrous CH2Cl2. The reagent 1-(trimethylsilyl)-imi-
dazole (TMS-imid., 0.11 mL, 0.71 mmol) was added
dropwise via syringe. The mixture was stirred at room
temperature for 12 h, quenched with 2.5 mL of H2O,
extracted with EtOAc (3Â, 25.0 mL), dried over
MgSO4, filtered, concentrated and purified by silica gel
chromatography (15% EtOAc/hexanes) to afford 51.0
mg (0.12 mmol) of O-silylated saturated ketone (À)-15
(+)-12 in 90% yield: [a]2D8 +1.6ꢀ (c 0.88, EtOAc); H
in 92% yield: [a]D28 À37.0ꢀ (c 0.87, CHCl3); H NMR
1
1
NMR (CDCl3) d, 4.24 (d, J=15.6 Hz, 1H), 4.11 (d,
J=16.0 Hz, 1H), 4.08 (m, 1H), 3.56 (m, 1H), 2.16 (m,
1H), 1.07 (d, J=6.0 Hz, 3H), 1.01 (s, 3H); 13C NMR
(CDCl3) d 89.26, 80.76, 76.03, 69.21, 68.62, 56.74, 54.96,
52.04, 41.71, 40.05, 39.81, 39.74, 33.78, 25.12, 24.71,
(CDCl3) d 3.58 (m, 1H), 3.24 (m, 1H), 3.14 (m, 1H),
2.46 (m, 1H), 1.09 (d, J=6.0 Hz, 3H), 0.65 (s, 3H), 0.10
(s, 9H); 13C NMR (CDCl3) d 212.17, 77.50, 75.40, 68.97,
61.52, 56.84, 49.93, 41.18, 38.94, 38.21, 38.17, 25.91,
25.07, 24.22, 24.15, 22.71, 19.39, 18.26, 12.95, 2.72; IR
(CHCl3, cmÀ1) 2938, 2876, 2859, 1705; HRMS m/z
(M+H+) calcd 409.3138 for C24H44O3Siꢁ H+, found
409.3141.
23.14, 22.59, 17.70, 17.34, 14.15; IR (CHCl3, cmÀ1
3597, 3007, 2938, 2863, 1451, 1375; HRMS m/z
)
+
+
+
ꢁ
(M NH4 ) calcd 352.2852 for C21H34O3 NH4 , found
352.2847.
Acetylenic ketone (À)-16. Solid tetrapropylammonium
perruthenate (TPAP, 10.0 mg, 0.03 mmol) was added in
one portion to a stirring mixture of diol (+)-12 (90.0
mg, 0.27 mmol), 4-methylmorpholine N-oxide (NMO,
63.0 mg, 0.54 mmol), and activated 4 A MS (120.0 mg)
in 14.0 mL of anhyd CH2Cl2. Upon completion, the
reaction mixture was diluted with a solution of 50%
EtOAc/hexanes and filtered through a silica gel plug.
The filtrate was evaporated and the residue was purified
by silica gel chromatography (30% EtOAc/hexanes) to
Saturated diol (À)-13. To a 50.0 mL oven dried flask
charged with 90.0 mg (0.27 mmol) of acetylenic diol
(+)-12 was added 10.0 mL of EtOH. The solution was
purged with argon under reduced pressure (0.01 mm
Hg) followed by rapid addition of 40.0 mg of Pd/C (Pd
10%) in 2.0 mL of EtOH via pippette. The hetroge-
neous mixture was stirred over a hydrogen atmosphere
at room temperature for 12 h at which time it was fil-
tered through a silica gel plug (100% EtOH). The

M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
1697
afford 83.0 mg (0.25 mmol) of acetylenic ketone (À)-16
reaction mixture was stirred at room temperature for 12
h in the dark. The solvent was evaporated and the mix-
ture was purified by silica gel chromatography (1%
NEt3/EtOAc) to afford 28.0 mg (0.06 mmol) of a mix-
ture of two desilylated diastereomers (À)-8a and (À)-8b
in 75% yield. The mixture of diastereomers was seper-
ated using reverse-phase HPLC (C-18 semipreparative
column, 27% H2O/CH3CN, 2.0 mL/min) to give (À)-8a
(1a,3b, tR 25.43 min, 27% yield), (À)-8b (1b,3a, tR 27.96
min, 25% yield) and 20% was left as a mixture: (À)-8a:
[a]2D8 À149.0ꢀ (c 0.47, MeOH); 1H NMR (CDCl3) d 6.32
(d, J=10.8 Hz, 1H), 5.93 (d, J=11.6 Hz, 1H), 5.16 (m,
1H), 5.01 (m, 1H), 3.95 (m, 1H), 3.55 (m, 3H), 3.25 (m,
2H), 2.81 (m, 1H), 2.62 (m, 2H), 2.25 (m, 1H), 1.08 (d,
J=6.0 Hz, 3H), 0.55 (s, 3H); 13C NMR (CD3OD) d
147.79, 142.69, 136.71, 124.09, 119.14, 114.22, 79.87,
72.12, 70.24, 67.54, 64.82, 58.42, 57.24, 47.55, 47.12,
46.65, 41.96, 38.58, 38.32, 37.75, 30.20, 27.21, 26.28,
25.09, 24.81, 23.71, 23.49, 18.85, 13.13; UV (MeOH)
lmax 264 nm (e 14 250); HRMS m/z (M+) calcd
486.3709 for C31H50O4, found 486.3718. (À)-8b: [a]D28
in 92% yield: [a]2D8 À28.3ꢀ (c 1.47, EtOAc); H NMR
1
(CDCl3) d 4.26 (d, J=15.6 Hz, 1H), 4.12 (d, J=15.6 Hz,
1H), 3.54 (m, 1H), 2.96 (m, 1H), 1.10 (d, J=6.0 Hz,
3H), 0.70 (s, 3H); 13C NMR (CDCl3) d 212.04, 89.49,
80.44, 75.83, 68.60, 61.38, 56.62, 54.98, 49.86, 41.14,
39.80, 39.74, 38.73, 25.08, 24.99, 24.17, 23.12, 19.28,
17.68, 12.90; IR (CHCl3, cmÀ1) 3595, 3009, 2938, 2876,
2859, 1705, 1452, 1378; HRMS m/z (M+) calcd
332.2351 for C21H32O3, found 332.2356.
O-Silylated acetylenic C-8-ketone (À)-17. An oven dried
25.0 mL round bottomed flask was charged with 68.0
mg (0.20 mmol) of alcohol (À)-16 and 10.0 mL of anhyd
CH2Cl2. The reagent 1-(trimethylsilyl)-imidazole (TMS-
imid., 0.2 mL, 1.4 mmol) was added dropwise via syr-
inge. The mixture was stirred at room temperature for
12 h, quenched with 5.0 mL water, extracted with
EtOAc (3Â, 25.0 mL), dried over MgSO4, filtered, con-
centrated, and purified by silica gel chromatography
(10% EtOAc/hexanes) to afford 74.0 mg (0.18 mmol) of
the desiredꢀ O-silylated product (À)-17 in 90% yield:
À1.3ꢀ (c 0.47, MeOH); H NMR (CDCl3) d 6.32 (d,
1
1
[a]2D8 À28.6 (c .58, EtOAc); H NMR (CDCl3) d 4.27
J=11.2 Hz, 1H), 5.92 (d, J=11.2 Hz, 1H), 5.14 (m,
1H), 4.98 (m, 1H), 4.00 (m, 1H), 3.58 (m, 3H), 3.24 (m,
2H), 2.82 (m, 1H), 2.61 (m, 2H), 2.26 (m, 1H), 1.08 (d,
J=6.0 Hz, 3H), 0.52 (s, 3H); 13C NMR (CD3OD) d
147.85, 142.87, 136.77, 124.03, 119.09, 113.95, 79.86,
72.12, 70.23, 67.58, 64.79, 58.37, 57.19, 47.54, 47.07,
46.53, 41.98, 38.58, 38.32, 37.70, 30.21, 27.20, 26.34,
25.08, 24.68, 23.57, 23.48, 18.84, 13.10; UV (MeOH)
lmax 263 nm (e 16 371). HRMS m/z (M+) calcd
486.3709 for C31H50O4, found 486.3714.
(d, J=16.0 Hz, 1H), 4.11 (d, J=15.6 Hz, 1H), 3.54 (m,
1H), 2.46 (m, 1H), 1.1 (d, J=5.6 Hz, 3H), 0.70 (s, 3H),
0.16 (s, 9H); 13C NMR (CDCl3) d 212.02, 89.88, 81.38,
75.80, 69.89, 61.40, 56.62, 55.04, 49.88, 41.16, 41.01,
38.75, 25.19, 25.03, 24.15, 23.02, 19.30, 17.58, 12.88,
1.99; IR (CHCl3, cmÀ1) 3003, 2938, 2858, 1705, 1451,
1378; HRMS m/z (M+) calcd 404.2747 for C24H40O3Si,
found 404.2756.
Saturated hybrid analogues (À)-8a and (À)-8b. Racemic
phosphine oxide A-ring (Æ)-18 and C,D-ring ketone
(À)-15 were each azeotropically dried three times with
anhydrous benzene and held under vacuum for 24 h
immediately prior to use. Racemic phosphine oxide
(Æ)-18 (74.0 mg, 0.12 mmol) was dissolved in 1.0 mL of
anhydrous THF (freshly distilled over benzophenone
ketyl) and cooled to À78 ꢀC under an argon atmo-
sphere. To this was added 97.0 mL (0.17 mmol) of PhLi
(1.8 M soln in cyclohexanes) dropwise during which
time a deep red/orange color developed and persisted.
The mixture was allowed to stir an additional 5 min at
À78 ꢀC at which time a pre-cooled (À78 ꢀC) solution of
C,D-ring ketone (À)-15 (51.0 mg, 0.12 mmol) dissolved
in 1.0 mL of anhydrous THF was transfered dropwise
via cannula. The deep-red/orange solution was stirred in
the dark for approximately 5 h during which time the
color faded to light yellow. The reaction mixture was
quenched at À78 ꢀC with 2.0 mL of 2 N sodium potas-
sium tartrate solution followed by addition of 1.0 mL of
dilute aqueous potassium carbonate. The reaction mix-
ture was allowed to warm to room temperature at which
time it was extracted with EtOAc (3Â, 25.0 mL), dried
over MgSO4, filtered, concentrated, and purified by
silica gel chromatography (0.1% NEt3, 10% EtOAc,
hexanes) to afford 63.0 mg (0.08 mmol) of the coupled
product in 64% yield. The coupled product was imme-
diately placed in an oven dried 25.0 mL flask and dis-
solved in 5.0 mL of anhyd THF with 20.0 mL of NEt3
under Ar. To this solution was added 364.0 mg (1.4
mmol) of TBAF and 100.0 mg of activated 4 A MS. The
Unsaturated hybrid analogues (À)-9a and (À)-9b. Race-
mic phosphine oxide A-ring (Æ)-18 and C,D-ring
ketone (À)-17 were each azeotropically dried three times
with anhydrous benzene and held under vacuum for 24
h immediately prior to use. Racemic phosphine oxide
(Æ)-18 (74.0 mg, 0.12 mmol) was dissolved in 1.0 mL of
anhyd THF (freshly distilled over benzophenone ketyl)
and cooled to À78 ꢀC under an argon atmosphere. To
this was added 88.0 mL (0.15 mmol) of PhLi (1.8 M soln
in cyclohexanes) dropwise during which time a deep
red/orange color developed and persisted. The mixture
was allowed to stir an additional 5 min at À78 ꢀC at
which time a pre-cooled (À78 ꢀC) solution of C,D-ring
ketone (À)-17 (61.4 mg, 0.15 mmol) dissolved in 1.0 mL
of anhydrous THF was transferred dropwise via can-
nula. The deep red/orange solution was stirred in the
dark for approximately 6 h during which time the color
faded to light yellow. The reaction mixture was quen-
ched at À78 ꢀC with 2.0 mL of 2 N sodium potassium
tartrate solution followed by addition of 1.0 mL of
dilute aqueous potassium carbonate. The reaction mix-
ture was allowed to warm to room temperature at which
time it was extracted with EtOAc (3Â, 25.0 mL), dried
over MgSO4, filtered, concentrated, and purified by
silica gel chromatography (0.1% NEt3, 10% EtOAc,
hexanes) to afford 67.0 mg (0.08 mmol) of the coupled
product in 68% yield. The coupled product was imme-
diately placed in an oven dried 25.0 mL flask and
dissolved in 5.0 mL of anhyd THF with 20.0 mL of NEt3
under Ar. To this solution was added 374.0 mg (1.43

1698
M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
mmol) of TBAF and 100.0 mg of activated 4 A MS. The
reaction mixture was stirred at room temperature for 12
h in the dark. The solvent was evaporated and the mix-
ture was purified by silica gel chromatography (1%
NEt3/EtOAc) to afford 29.0 mg (0.06 mmol) of a mix-
ture of two desilylated diastereomers (À)-9a and (À)-9b
in 70% yield. The mixture of diastereomers was seper-
ated using reverse-phase HPLC (C-18 semipreparative
column, 27% H2O/CH3CN, 2.0 mL/min) to give (À)-9a
(1a,3b, tR 16.49 min, 20% yield), and (À)-9b (1b,3a, tR
18.22 min, 35% yield) and 20% was left as a mixture:
(À)-9a: [a]2D8 À182.0ꢀ (c 0.56, MeOH); 1H NMR
(CDCl3) d 6.33 (d, J=11.2 Hz, 1H), 5.94 (d, J=11.2 Hz,
1H), 5.17 (m, 1H), 5.02 (m, 1H), 4.25 (d, J=15.6 Hz,
1H), 4.11 (d, J=15.6 Hz, 1H), 3.96 (m, 1H), 3.55 (m,
3H), 2.82 (m, 1H), 2.62 (m, 2H), 1.09 (d, J=6.0 Hz,
3H), 0.62 (s, 3H); 13C NMR (CD3OD) d 147.81, 142.72,
136.72, 124.08, 119.12, 114.19, 90.75, 81.55, 78.35,
69.24, 67.56, 64.82, 58.30, 57.17, 56.06, 47.54, 47.14,
46.62, 41.72, 40.97, 40.93, 37.77, 30.22, 26.54, 26.19,
24.86, 24.47, 23.62, 18.39, 13.19; UV (MeOH) lmax 264
nm (e 18 905); HRMS m/z (M+) calcd 482.3396 for
C31H46O4, found 482.3404. (À)-9b: [a]2D8 À5.0ꢀ (c 0.43,
Acknowledgements
This work was supported by grants from the National
Institutes of Health [NIH (CA 44530 for G.H.P. and
M.C.W.] and [NIH (DK 50583 for S.P.)], and the
National Cancer Institute (NCI (NCDDG CA52857 for
HB and MDB). The American Chemical Society Divi-
sion of Medicinal Chemistry and Parke-Davis are
gratefully acknowledged for support of a predoctoral
fellowship to M.C.W.
References and Notes
1. Norman, A. W. Vitamin D; The Calcium Homeostatic
Steriod Hormone; Academic: New York, 1979.
2. Bikle, D. D.; Pillai, S. Endocr. Rev. 1993, 15, 3.
3. Takahashi, N.; Akatsu, T.; Sasaki, T.; Nicholson, G. C.;
Moseley, J. M.; Martin, T. J.; Suda, T. Endocrinology 1988,
123, 1504.
4. (a) De Jong, E. M. G. J.; van de Kerkhof, P. C. M. Br. J.
Dermatol. 1991, 124, 221. (b) Kragballe, K. In Molecular
Biology to Therapeutics; Bernard, B. A., Schroot, B., Eds.;
Birkhaeuser: Basal, 1993; pp 174–181.
5. (a) Abe, E.; Miyaura, C.; Sakagami, H.; Takeda, M.;
Konno, K.; Yamazaki, T.; Yoshiki, S.; Suda, T. Proc. Natl.
Acad. Sci. U.S.A. 1981, 78, 4990. (b) Ostrem, V. K.; Lau,
W. F.; Lee, S. H.; Perlman, K.; Prahl, J.; Schnoes, H. K.;
DeLuca, H. F. J. Biol. Chem. 1987, 262, 14164. (c) Kane,
K.; Langman, M. J. S.; Williams, G. R. Cancer Res. 1996,
56, 623. (d) Colston, K. W.; Berger, U.; Coombes, R. C.
Lancet 1989, 1, 188. (e) Love-Schimenti, C. D.; Gibson,
D. F. C.; Ratnam, A. V.; Bikle, D. D. Cancer Res. 1996, 56,
2789. (f) Walters, M. R. Endocr. Rev. 1992, 13, 719.
6. Abe, J.; Nakano, T.; Nishi, Y.; Matsumoto, T.; Ogata, E.;
Ikeda, K. Endocrinology 1991, 129, 832.
1
MeOH); H NMR (CDCl3) d 6.32 (d, J=11.2 Hz, 1H),
5.92 (d, J=11.6 Hz, 1H), 5.15 (m, 1H), 4.99 (m, 1H),
4.25 (d, J=16.0 Hz, 1H), 4.11 (d, J=15.6 Hz, 1H), 4.00
(m, 1H), 3.56 (m, 3H), 2.82 (m, 1H), 2.61 (m, 2H), 1.09
(d, J=6.0 Hz, 3H), 0.59 (s, 3H); 13C NMR (CD3OD) d
147.86, 142.91, 136.77, 124.04, 119.08, 113.96, 90.74,
81.55, 78.34, 69.23, 67.58, 64.80, 58.25, 57.12, 56.05,
47.54, 47.09, 46.53, 41.75, 40.97, 40.92, 37.71, 30.23,
26.53, 26.25, 24.73, 24.46, 23.50, 18.40, 13.17. UV
(MeOH) lmax 263 nm (e 19 143). HRMS m/z (M+)
calcd 482.3396 for C31H46O4, found 482.3399.
Determination of serum calcium levels
7. Peleg, S. In Vitamin D; Feldman, D., Glorieux, F. H.; Pike,
J. W., Eds.; Academic: New York, 1997; pp 1011–1025.
8. Bouillon, R.; Oklamura, W. H.; Norman, A. W. Endocr.
Rev. 1995, 16, 200.
9. (a) Okamura, W. H.; Zhu, G. D. In Vitamin D; Feldman,
D., Glorieux, F. H., Pike, J. W., Eds.; Academic: New York,
1997; pp 939–971. (b) Fall, Y.; Fernandez, C; Gonzalez, V.;
Vitale, C.; Mourino, A. In Vitamin D Endocrine System:
Structure, Biological, Genetic, and Clinical Aspects; Norman,
A. W., Bouillon, R., Thomasset, M., Eds.; University of Cali-
fornia: Riverside, 2000; pp 99–101.
10. Figadere, B.; Norman, A. W.; Henry, H. L.; Koeffler,
H. P.; Zhou, J.-Y.; Okamura, W. H. J. Med. Chem. 1991, 34,
2452.
11. Chen, T. C.; Parsons, K.; Uskokovic, M. R.; Horst, R. L.;
Holick, M. F. J. Nutr. Biochem. 1993, 4, 49.
12. Binderup, E.; Calverley, M. J.; Binderup, L. In Vitamin D:
Gene Regulation, Structure–Function Analysis and Clinical
Applications; Norman, A. W., Bouillon, R., Thomasset, M.,
Eds.; De Gruyter: Berlin, 1991; pp 192–193.
13. Mathiasen, I. S.; Colston, K. W.; Binderup, L. J. Steriod
Biochem. Molec. Biol. 1993, 46, 365.
14. Hansen, K.; Calverley, M. J.; Binderup, L. In Vitamin D:
Gene Regulation, Structure–Function Analysis and Clinical
Applications; Norman, A. W., Bouillon, R., Thomasset, M.,
Eds.; De Gruyter: Berlin, 1991; pp 161–162.
15. Elstner, E.; Linker-Israeli, M.; Umiel, T.; Le, J.; Grillier,
I.; Said, J.; Shintaku, I. P.; Krajewski, S.; Reed, J. C.;
Binderup, L.; Koeffler, H. P. Cancer Res. 1996, 56, 3570.
Female C57 Bl-6 mice (n=3) received daily intraper-
itoneal injections of one of four compounds (1,25D3, 6,
8b, and 9b) at 1, 10, or 100 mg/kg in 100 mL of vehicle
(80% propylene glycol/20% 0.05 M phosphate buffered
saline). Mice were monitored daily for weight loss as an
indicator of drug toxicity; 1,25D3 at 100 mg/kg/day was
fatally toxic by day 4. At day 7, mice were anesthetized
with intraperitoneal injection of 200 mL of a stock solu-
tion containing 25 mg/mL ketamine hydrochloride, 2.5
mg/mL xylazine, and 14.25% ethanol all diluted 1:3 in
0.9% sodium chloride solution. Each blood sample was
collected via cardiac puncture into a heparanized gas-
tight syringe which was immediately sealed and placed
on ice. Samples were then quantitatively analyzed
within 2.5 h for ionized calcium content using a Chiron
865 instrument in the Critical Care Lab at Johns Hop-
kins Hospital. Results are expressed as the mean ÆSD
from three animals in each group. The pH values of all
samples were determined concomitantly with the same
instrument and were within normal limits.
Supplementary Material
Proton and 13C NMR spectra for compounds 11–17
and 8–9 are available.

M. C. White et al. / Bioorg. Med. Chem. 9 (2001) 1691–1699
1699
16. Posner, G. H.; Nelson, T. D.; Guyton, K. Z.; Kensler,
T. W. J. Med. Chem. 1992, 35, 3280.
17. Posner, G. H.; White, M. C.; Dolan, P.; Kensler, T. W.;
Yukihiro, S.; Guggino, S. E. Bioorg. Med. Chem. Lett. 1994, 4,
2919.
18. Burke, M.; White, M.; Watts, M.; Lee, J.; Tyler, B.; Pos-
ner, G.; Brem, H. In Proceedings of the Tenth Workshop on
Vitamin D, Strasbourg, France, May 24–29, 1997.
19. Posner, G. H.; Lee, J. K.; Wang, Q.; Peleg, S.; Burke, M.;
Brem, H.; Dolan, P.; Kensler, T. W. J. Med. Chem. 1998, 41,
3008.
20. Peleg, S.; Liu, Y. Y.; Reddy, S.; Horst, R. L.; White,
M. C.; Posner, G. H. J. Cell. Biochem. 1996, 63, 149.
21. Okamura, W. H.; Palenzuela, J. A.; Plumet, J.; Midland,
M. M. J. Cell. Biochem. 1992, 49, 10.
22. Midland, M. M.; Plumet, J.; Okamura, W. H. Bioorg.
Med. Chem. Lett. 1993, 3, 1799.
23. Grue-Sorensen, G.; Hansen, C. M. In Vitamin D: Chem-
istry, Biology, and Clinical Applications of the Steriod Hor-
mone; Norman, A. W., Bouillon, R., Thomasset, M., Eds.;
Printing and Reprographics University of California: River-
side, CA, 1997; pp 32–33.