Synthesis, biological activity, and conformational analysis of CD-ring modified trans-decalin 1α,25-dihydroxyvitamin D analogs

Article abstract of DOI:10.1039/b209147j

A novel series of analogs of 1.25-dihydroxyvitamin D₃, the hormonally active metabolite of vitamin D₃, characterised by the presence of a trans-fused decalin CD-ring system, possesses surprising biological activities in combination w

Full text of DOI:10.1039/b209147j

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis, biological activity, and conformational analysis of CD-
ring modified trans-decalin 1ꢀ,25-dihydroxyvitamin D analogs†
Yong-Jun Chen,a Ling-Jie Gao,a Ibrahim Murad,a Annemieke Verstuyf,b Lieve Verlinden,b
Christel Verboven,b Roger Bouillon,b Davide Viterbo,c Marco Milanesio,c Dirk Van Haver,a
Maurits Vandewallea and Pierre J. De Clercq*a
a Laboratory for Organic Synthesis, Department of Organic Chemistry, Ghent University,
Krijgslaan 281 (S4), B-9000 Gent, Belgium. E-mail: pierre.declercq@rug.ac.be
b Laboratory of Experimental Medicine and Endocrinology, Catholic University of Leuven,
Gasthuisberg, Herestraat 49, B-3000, Leuven, Belgium
c Dipartimento di Scienze e Tecnologie Avanzate, Università del Piemonte Orientale
“A. Avogadro”, Corso T. Borsalino 54, I-15100 Alessandria, Italy
Received 18th September 2002, Accepted 2nd December 2002
First published as an Advance Article on the web 19th December 2002
A novel series of analogs of 1,25-dihydroxyvitamin D3, the hormonally active metabolite of vitamin D3,
characterised by the presence of a trans-fused decalin CD-ring system, possesses surprising biological activities
in combination with specific structural modifications in the flexible parts of the molecule, when compared with
the natural hydrindane derivatives. (1) A large difference in biological activity is observed between the 20-epimeric
trans-decalin analogs that follows a pattern opposite to what is usually observed for the natural ring size. (2)
Several trans-decalin analogs that are modified in the seco-B-ring region, including previtamin derivatives, possess
a pronounced vitamin D-like activity, whereas the corresponding hydrindane derivatives are inactive. The molecular
origin of this behavior is still under study.
Introduction
For chemists, biochemists, medicinal chemists, and physicians
alike, vitamin D represents a fascinating molecule.1 Its chemi-
cal history started in the 1930’s when Windaus proposed a
surprising secosteroid structure with an intact cholesterol
side chain for the substance,2 a deficiency in which was
known to be responsible for bone diseases such as endemic
rickets and osteomalacia. Since the importance of calcium in
bone formation was known at that time, vitamin D3 was also
called cholecalciferol (D3). Many years later, it was discovered,
in particular through the seminal work of DeLuca, Kodicek,
and Norman that the biological action of the vitamin origin-
ates from the dihydroxylated metabolite 1α,25-dihydroxy-
vitamin D3 (1,25-D3), also called calcitriol, which functions
as a hormone.3–5 As was subsequently assessed, these actions
not only involve regulation of calcium homeostasis, but also
immunomodulation, selected cell differentiation, and anti-
proliferation.6 At the molecular level 1,25-D3 stimulates
biological responses via signal transduction pathways that
utilize either a nuclear receptor (n-VDR) for regulating gene
transcription (slow genomic pathway) or a putative membrane
Scheme 1 Biosynthetic pathway to vitamin D.
receptor (m-VDR) for its rapid actions such as transcaltachia
(non-genomic pathway).7
The characteristic seco-structure of vitamin D originates
from the photolytic cleavage of the C-9,C-10 bond upon UV
irradiation of the provitamin 7-dehydrocholesterol (Scheme 1).
The so generated previtamin triene (pre-D3) then undergoes a
reversible thermal [1,7]-sigmatropic hydrogen shift leading to
the vitamin form in its less stable s-cis conformation. Rapid
rotation around the C-6,C-7 single bond eventually yields D3 in
its more stable extended conformation.8,9 It is believed now that
the transoid and cisoid conformations represent the active
geometries of the hormone when binding to the n-VDR and m-
VDR, respectively.10 In this context the previtamin form is
unique in that it corresponds to a vitamin tautomer whose
geometry is locked in the cisoid form.
The vast therapeutic potential of 1,25-D3 that is associated
with the inhibition of cell growth and with the stimulation of
cell differentiation is however hampered by the observation that
† Electronic supplementary information (ESI) available: Further
experimental details. See http://www.rsc.org/suppdata/ob/b2/b209147j/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
257

View Article Online
above the required physiological doses the hormonally active
metabolite leads to calcemic side effects such as hypercalcemia,
hypercalciuria and bone decalcification. This has originated in a
search for structural analogs of 1,25-D3 that show a separation
in calcemic and antiproliferative–prodifferentiating activities. A
turning point in this research constitutes the recent disclosure
by Moras of the detailed structure of the ligand binding
domain (LBD) of the n-VDR, obtained via X-ray diffraction
analysis of the complex between 1,25-D3 and a truncated
mutant of the n-VDR.11
In the context of analog development it is appropriate to
distinguish three different parts in the structure of 1,25-D3: a
central rigid CD-bicyclic ring portion to which are connected,
at C-17 and at C-8, two flexible entities, the side chain and
the seco-B,A-ring system, respectively (Chart 1). Although
modifications in the synthetically less accessible CD-ring part
of the molecule.15,16 The search has led to the development of
different series of non-steroidal analogs featuring, instead
of the natural trans-fused hydrindane‡ bicyclic system, a
mono-cyclic central moiety such as a six-membered C- or a
five-membered D-ring, an unnatural ringsize such as a six-
membered D-ring, even an extraneous ring such as a five-
membered E-ring (Chart 1).17–20 In each of these series,
derivatives were found showing a clear dissociation in activities.
For the purpose of illustration, the biological activity of a
selected set of analogs possessing structural modifications in
the side chain or in the CD-ring system and that show a dissoci-
ation in activities is given in Table 1. Compared to 1,25-D3 these
analogs possess both a higher prodifferentiative and a lower
calcemic activity, except KH 1060 and KS 176, which have
a higher calcemic and a somewhat lower prodifferentiative
activity, respectively.
In the same context we wish to describe here the synthesis
and biological activity of a series of trans-decalin§ CD-ring
modified 1,25-D3 analogs in combination with classical modifi-
cations such as epimerisation at C-20 and removal of C-19
(Chart 2).21 In addition, the isomerisation between vitamin and
previtamin forms is also investigated (Scheme 2). Indeed, the
Chart 2 Symbolic designation used for the stereochemistry at C-20
and for the nature of the A-ring.
Chart 1 Selected set of examples of modifications in the side chain
(top) and in the central CD-ring system (bottom) of 1,25-D3.
Scheme 2 Vitamin–previtamin equilibrium.
structural modifications have been introduced to each of these
three parts that successfully led to derivatives possessing the
desired separation of activities, the side chain has been the
favourite location for analog development.6 Moreover molecu-
lar modeling studies have indicated that there exists a link
between the biological activities and the preferred spatial
orientation of the side chain.12 Seminal in this respect was the
finding that 20-epi-1,25-D3 is several orders of magnitude
more potent than the natural hormone both in prodifferentiat-
ing and calcemic activity.13 Another interesting modification,
now in the A-ring of the molecule, consists in the deletion of
C-19. Indeed, it was reported by DeLuca that 19-nor-1,25-D3
possesses a biological profile with the same cell differentiative
and antiproliferative activity as the natural hormone, yet with a
10-fold reduction in calcemic activity.14 For one decade now our
laboratories in Ghent and Leuven have focussed on structural
composition of this equilibrium is known to be determined by
the nature of the CD-ring system.22 In comparison with the
equilibrating natural forms, it was expected that in the trans-
decalin homologous series the corresponding previtamin deriv-
ative would be less prone to undergo a [1,7]-sigmatropic
rearrangement, hence representing a system in which the bio-
logical profile of a genuine previtamin form could be assessed.
In what follows 16 different compounds corresponding to
structures 1–8 will be compared (Scheme 3). They represent
structures not only in the vitamin (1, 2), and the previtamin
(3, 4) series, but also analogs (5, 6, 7, 8) that were obtained in
the context of the synthesis of the previtamin derivatives. In
‡ The IUPAC name for hydrindane is hexahydroindane.
§ The IUPAC name for decalin is decahydronapthalene.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
258

View Article Online
Table 1 Biological activities of selected known 1,25-D3 analogs shown in Chart 1a
Binding affinity
Calcemic effect
Cell differentiation and proliferation
VDR
(Pig)
DBP
(Human)
Serum
(Mice)
Code
HL-60
MCF-7
Keratinocytes
OCTb
20
85
85
45
60
80
125
10
0
7
0
200
150
1000
4000
1000
85
100
—
2500
17000
6000
85
300
200
8000
30000
6000
90
0.3
1
8
500
50
0.3
4
MC 903c
Ro 23–7553d
KH 1060e
ZG 1368f
SL 117g
0
20
10
80
19
BL 269h
KS 176i
90
20
300
30
300
10
<0.1
a All data resulted from evaluation in the Legendo laboratory at the KULeuven. The activities are presented as relative values, the reference value of
1,25-D3 (1at) being defined as 100. Further details about the methodology are given in the Experimental Section. b 22-oxa-1,25-D3, ref. 52. c 22-ene-
26,27-dehydro-1,24(S)-D3, ref. 53. d 16-ene-23-yne-1,25-D3, ref. 54. e 20-epi-22-oxa-24,26,27-trihomo-1,25-D3, ref. 13 and 55. f Ref. 18. g Ref. 19.
h Ref. 20a. i Ref. 17.
Results and discussion
Synthesis
The general strategy for the synthesis of the discussed com-
pounds is outlined in Scheme 3.24,25 Central in the approach is
the side-chain hydroxylated Grundmann’s ketone 9 or the
corresponding homologous decalone¶ 10. Whereas the
Horner–Wittig olefination, originally introduced by Lythgoe,26
leads directly to the vitamins 1 and 2 (both in the natural and
in the 19-nor series),27,28 the synthesis of the corresponding
previtamins 3 and 4 involves the semi-hydrogenation of the
ynedienes 5 and 6, respectively.29 The latter are obtained now by
cross-coupling of the enoltriflates derived from 9 and 10, with a
terminal alkynyl substituted A-ring.30 Finally, derivatives 7 and
8 have been obtained as products of overreduction.
The synthesis of decalones 10a and 10b requires a multi-step
sequence that is worked out in Schemes 4 and 5. Reaction of the
known enantiomerically pure unsaturated decalone 1131 with
the anion derived from diethyl cyanomethylphosphonate32
gave the conjugated nitrile 12 as an unseparable mixture of E-
and Z-isomers (ratio 6 : 1, respectively).21 Subsequent selective
reduction of the conjugated double bond with magnesium in
methanol33 afforded a mixture of the diastereomeric nitriles 13
and 14 (ratio 10 : 1), which were separated by flash chromato-
graphy.21 The assignment of the major isomer to 13, in which
the cyanomethyl substituent at C-17 occupies the expected
more stable equatorial position,34 rests on the following NOE
experiment: irradiation of the C-18 methyl group (δ 0.96)
showed a 4.2% enhancement of H-20 (δ 2.49) in the case of the
major isomer 13, whereas in the case of the minor isomer 14 no
such enhancement was observed. Our previously reported syn-
thetic sequence towards trans-decalin analogs 2at and 2bt
involved alkylation of the enolate derived from cyanomethyl
derivative 13 with the complete C22–C27 side chain segment,
followed by conversion of the cyano function into the C-21
methyl group.21 The configuration at C-20 in the major isomer
was found to correspond to the natural (20R)-configuration.35
In analogy with this result the alkylation of 13 with iodo-
methane leads to a mixture of 15a and 15b in which 15a pre-
dominates (ratio 5 : 1). This assigment was confirmed by the 1H
NMR spectral data: comparison of the observed and calculated
vicinal coupling constant values for J(H17,H20) (15a: obsd <1.0
Scheme 3 General synthetic strategy.
Hz, calcd 1.0 Hz; 15b: obsd 3.1 Hz, calcd 4.0 Hz) indicates 15a,
in which C-20 possesses the natural configuration, to be the
order to increase the readability of the paper the following
symbols are used for structural identification: a and b refer to
the natural 20R- and epimeric 20S-configuration, respectively;
t and r to the natural and 19-nor A-ring skeleton, respectively;
the odd Arabic numerals 1, 3, 5 and 7 refer to products possess-
ing the natural ring system with a five-membered D-ring, and
the even numerals 2, 4, 6 and 8 to the trans-decalin analogs with
a six-membered D-ring.23
major stereoisomer.36
After conversion of the nitrile function into the correspond-
ing primary alcohol (diisobutylaluminium hydride, 2 times), the
two diastereomers 16a and 16b were separated by chromato-
¶ The IUPAC name for decalone is octahydronapthalenone. The
IUPAC name for triflate is trifluoromethanesulfonate.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
259

View Article Online
Scheme 4 (a) (EtO)2P(O)CH2CN, NaNH2, THF, rt, 6 h; 11, THF, rt,
12 h (96%); (b) Mg, MeOH, rt, 18 h (92%); (c) n-BuLi, i-Pr2NH, THF,
Ϫ20 ЊC, 30 min; 13, THF, Ϫ78 ЊC, 1 h; MeI, THF, Ϫ78
Ϫ20 ЊC,
3 h (93%); (d) DIBAL-H, n-hexane–Et2O, Ϫ78 0 ЊC, 4 h (2 times)
(64%); (e) (S)-(ϩ)-C10H7CH(CH3)NCO, CF3SO3Si(CH3)3, CH2Cl2, rt,
40 min (97%).
graphy. In view of the applied reaction conditions for trans-
forming the mixture of nitriles 15a and 15b into the correspond-
ing alcohols 16a and 16b, substantial isomerisation was not
expected. Hence the major isomer that was obtained after
reduction of the nitrile function was anticipated to correspond
to 16a, possessing the “natural” configuration at C-20. A fur-
ther indication that this assignment is correct follows again
from the 1H NMR spectral data and from the result of molecu-
lar modeling. Fig. 1 shows the predominant orientation of the
side chain as calculated for 16a and 16b. This is in accordance
with (i) the large shift difference of 0.52 ppm between the two
hydrogen atoms on C-22 in 16a as compared to 16b (0.02 ppm);
(ii) the NOE enhancement (3.8%) of the signal of the low-field
hydrogen atom on C-22 (δ 3.81) upon irradiation of the C-18
methyl group (δ 1.06), which is only observed in the case of 16a,
and (iii) the experimental vicinal coupling constants of H-20
with the hydrogens on C-22 (16a: obsd 3.5, 9.9 Hz, calcd 2.9, 9.6
Hz; 16b: obsd 6.8, 7.5 Hz, calcd 6.1, 6.5 Hz). The structural
assignment within the b-series also follows from the X-ray dif-
fraction analysis of crystalline 17 (Fig. 2).37 The absolute struc-
ture could not be determined from the X-ray data, but was
established from the synthesis pathway. Indeed, the latter
carbamate was obtained by reaction of 16b with (S)-(ϩ)-1-
(1-naphthyl)ethyl isocyanate in the presence of trimethylsilyl
triflate. Obviously, both the expected relative configuration at
C-13, C-17 and C-20, and the absolute configuration as shown
in 16 is thereby unambiguously proven.
Scheme 5 (a) TsCl, pyridine, 0 ЊC, 12 h (18a: 92%; 18b: 91%);
(b) BH3ؒTHF, THF, Ϫ20 ЊC, 12 h; 2 M NaOH, 35% H2O2, Ϫ30 ЊC
rt, 30 min (19a ϩ 20a: 82%; 19b ϩ 20b: 82%); (c) NaI, acetone, 60 ЊC,
12 h; (d) NiCl2, Zn, CH2᎐᎐CHCO2Et, pyridine, 65 ЊC, 45 min; mixture of
iodides from (c), pyridine, rt, 1.5 h (21a ϩ 22a: 80%; 21b ϩ 22b: 72%);
(e) MeMgBr, THF, Ϫ5 ЊC
rt, 1 h (23a ϩ 24a: 93%; 23b ϩ 24b: 91%);
(f ) PDC, CH2Cl2, rt, 12 h (25a ϩ 10a: 88%; 25b ϩ 10b: 86%).
configurational series). The sequence involves two basic oper-
ations consisting of the establishment of a trans-fused C-8
decalone and the introduction of the C-25 hydroxylated side
chain. The oxidative hydroboration procedure applied to
tosylates 18a and 18b, obtained in the classical way from the
corresponding alcohols 16a and 16b, respectively, led in both
cases to a 2 : 1 mixture of cis- and trans-fused isomers (19a and
20a from 18a, and 19b and 20b from 18b, respectively). On mere
steric grounds the preferred addition to the convex face of the
bicyclic molecule resulting in the preferred formation of the
cis-fused isomer (i.e., 19a and 19b) was anticipated.38 Indeed,
the major isomer in the obtained 2 : 1 mixture is the cis-fused
isomer in which the C-18 methyl group in the 1H NMR reson-
ates at lower field. The same trend is also observed when com-
paring subsequent cis- and trans-fused isomers; the average
The further conversion of unsaturated alcohols 16a and 16b
into the corresponding decalones 10a and 10b, respectively, is
outlined in Scheme 5 (the same scheme is used for the two
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
260

View Article Online
and stereoselective Wittig–Horner type olefination involves
reaction with the A-ring phosphine oxide 27 (natural A-ring)44
and 28 (19-nor A-ring),28,45 followed by deprotection of the
different trialkylsilyl ether groups (Scheme 6).
Fig. 1 Predominant orientation of the side chain as calculated for 16a
and 16b, and the experimental and calculated 1H NMR vicinal coupling
constants J for the hydrogens on C-22.
Scheme 6 (a) TESCl, imidazole, CH2Cl2, rt, 3 h (26a: 88%; 26b: 84%);
(b) 27 (28), n-BuLi, THF, Ϫ78 ЊC, 1 h; 26a (26b), THF, Ϫ78 ЊC
2 h; (c) TBAF, THF, rt, 12 h.
rt,
The synthesis of the previtamin derivatives 3 and 4 is high-
lighted in Scheme 7. The chosen strategy involves coupling of
vinyl triflates 29a and 30a with the A-ring terminal alkynes 31
(natural A-ring)46 and 32 (19-nor A-ring)47 following the pro-
cedure developed by Castedo and Mouriño and co-workers.48
The sequences leading to the previtamin derivatives in the
natural hydrindane series have been described before, both for
the preparation of the 19-natural (3at; pre-D3)9 and the 19-nor
(3ar)47 derivative. The required hydrindanone intermediate 9a
was obtained following the known procedure involving ozon-
olysis of vitamin D3, followed by direct hydroxylation at the 25-
position.49 The conversion of both ketones 9a and 10a into the
reactive vinyl triflate intermediates 29a and 30a proceeded via
prior protection (triethylsilyl chloride, dimethylaminopyridine,
dichloromethane), and treatment of the enolate anions (LDA,
THF) with N-phenyltriflimide in THF (>90% yield).50 The sub-
sequent coupling reactions followed by classical deprotection
led to the four yne–diene type analogs 5at, 5ar (natural CD-
series), and 6at, 6ar (trans-decalin series). For the synthesis of
the previtamin derivatives 3 and 4, the coupling process is
immediately followed by the semi-hydrogenation of the C-6,C-7
yne-moiety with Lindlar catalyst in the presence of quinoline.
Careful reaction control is required since overreduction to yield
the corresponding saturated derivatives at C-6,C-7, 7 and 8
readily occurs. In the context of analog development, however,
this side reaction was gratefully acknowledged as a novel series
of analogs became available for biological screening. Due to the
known facile conversion of the natural previtamin D3 (3at) to
the corresponding vitamin form (1at, calcitriol),9 the semi-
hydrogenation of compound 33at was not undertaken in this
work.
Fig. 2 Perspective view of the crystal structure of two (molecules A
and B) of the four independent molecules in the unit cell of crystalline
17, with the adopted atom numbering.
shift value for 10, 20, 22, and 24: 0.70–0.80 ppm, and for 19, 21,
23, and 25: 0.85–1.13 ppm.39 For the introduction of the side
chain we used an improved procedure developed by Mouriño
which involves a zinc–copper mediated conjugate addition of
an iodide (obtained from the corresponding tosylates 19 and 20
via Finkelstein’s protocol) to ethyl acrylate under aqueous
sonochemical conditions (85% yield).40,41 The mixture of esters
21a and 22a (and of 21b and 22b in the epimeric series) was
not separated but treated successively with methylmagnesium
bromide and pyridinium dichromate. At this stage the cis- and
trans-fused decalones 25a and 10a (and 25b and 10b in the
20-epimeric series) were separated and spectroscopically identi-
fied. In particular the resonance of the angular methyl group
has a diagnostic value: a characteristic high field resonance is
observed for the trans-fused derivatives (δ 0.74 for 10a and 0.75
for 10b) when compared to the corresponding cis-fused com-
pounds (δ 1.07 for 25a and 1.13 for 25b).39 Conversion of the
cis-fused decalones into the more stable desired trans-fused
isomers was readily realised by treatment in base.42
The preparation of the vitamin analogs 2at, 2ar (and 2bt,
2br) proceeded via the classical Lythgoe coupling of the tri-
ethylsilyl protected decalone 26a (and 26b).43 This very reliable
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
261

View Article Online
their differentiation into monocytes–macrophages.51 After the
1981 report, analogs were developed on the basis of the idea
that by modifying the chemical structure of 1,25-D3 one could
separate the differentiation inducing activities from the calcium
regulating activities. As mentioned earlier the side-chain por-
tion has been the preferred location for introducing structural
modifications in the development of vitamin D analogs. The
first analog in which the desired separation of activities was
reported is 22-oxa-1α,25(OH)2–D3 (OCT).52 Other important
analogs in this respect are shown in Chart 1 and include
calcipotriol (MC 903),53 16-ene-23-yne-1α,25(OH)2–D3 (Ro
23–7553)54 and 20-epi-22-oxa-24,26,27-trihomo-1α,25(OH)2–
D3 (KH 1060).13,55 The latter compound exhibits a 20,000 fold
increased activity on cell differentiation and immune responses
but with increased calcemic activity. Its structure accumulates
several side-chain modifications that have been reported to
enhance the cell-differentiation activity: 20-epimerisation,
22-oxygen insertion, and side-chain elongation.6 Interesting
modifications in the A-ring include 19-nor-1α,25(OH)2–D3,14
1α-(hydroxymethyl)-25-OH-D3,56
2β-(3-hydroxypropoxy)-
1α,25-(OH)-D3 (ED-71)57 and 2-methyl alkylated derivatives.58
The 19-nor modification in particular is crucial in the context
of vitamin–previtamin equilibration (vide infra). Analogs that
have been developed in the period 1985–1995 have been
reviewed in a comprehensive manner by Bouillon, Okamura
and Norman: 280 analogs were listed involving in total 820
structural variations of 1,25-D3.6 Following this review many
more analogs have been described, in particular also novel
non-steroidal compounds.59,15
The 20-epi-modification
In the context of the study of structure–function relationships
the 20-epimerisation modification has taken a unique position.
Following the 1991 report by Binderup about the increased
overall activity of 20-epi-1,25-D3 it became clear that the orien-
tation in space of the side chain plays an important role.13
Okamura in Riverside developed a dot-map approach in order
to visualise the portions in space that were accessible for the
side-chain, with focus on the 25-OH group.60 This approach
was further used by Yamada in a comprehensive structure–
function relationship study.61 In particular, it was found that
analogs with preferred side-chain orientation in the region
represented as EA in Fig. 3 consistently possess a higher bio-
logical activity (except the affinity for DBP).12 We developed a
similar procedure for defining a volume in space, the preferred
occupation of which would correspond to high prodifferentiat-
ing (HL-60) and antiproliferative (MCF-7) activity.62 This
“active” volume is represented by a red sphere and coincides
with the above mentioned EA region.
Inspired by the observation that severe structural modifi-
cations in the central CD-ring part are coupled with a reduction
in calcemic activity (Table 1), we became interested in the
development of analogs in which the applied modifications
would direct the side chain in the presumed active region. This
has previously led to the development of 6D-analogs, charac-
terised by a six-membered 1,3-dimethyl substituted (C-16,C-14)
cyclohexane D-ring with reduced conformational mobility of
the side chain.20b Fig. 3 shows the result of the conformational
analysis procedure (see Experimental Section) for the side-
chain orientation of 1a, 1b, 2a and 2b. Inspection of the figure
learns that substantial occupation of the active volume is real-
ised by 2a, but not by the 20-epi analog 2b.63 This is particularly
well reflected in the results of the biological evaluation: the cell
differentiative, antiproliferative and calcemic activities of 2at
are substantially higher than of 2bt, whereas the affinity for the
VDR is almost identical. The same trend, although somewhat
less pronounced, is also observed for the corresponding 19-nor
derivatives 2ar and 2br. This trans-decalin pair represents one
of the rare examples where the 20-epimer shows a reduced
Scheme 7 (a) TESCl, DMAP or imidazole, CH2Cl2, rt, 3–12 h (from
9a: 96%; from 10a: 88%); (b) n-BuLi, i-Pr2NH, THF, 0 ЊC, 20 min;
ketone–silyl ether from (a), THF, Ϫ78 ЊC, 45 min; then
rt over 2 h;
(CF3SO2)2NC6H5, THF, Ϫ78 ЊC, 10 min; then rt over 4 h (29a: 88%;
30a: 71%); (c) 31 (32), CuI, Pd(PPh3)2(OAc)2, Et2NH, MeCN, rt, 1 h
(33at: 86%; 34at: 92%; 34ar: 91%); (d) H2, Lindlar catalyst, quinoline,
EtOAc, rt (protected 3ar: 86%; 4at: 54% from 6at); (e) H2, Lindlar
catalyst, EtOAc, rt (8at: 34% from 6at under conditions (d)); (f ) TBAF,
THF, rt, 12 h (3ar: 80% from 33ar; 4ar: 83% from 34ar; 5at: 90%; 5ar:
92% from 29a; 6at: 91%; 6ar: 96%; 7at: 66% from 33at; 7ar: 81% from
protected 3ar; 8ar: 81% from 34ar).
Biological evaluation
The biological evaluation of the parent compound (1at) and the
different analogs has included the determination of (1) the
affinity for the pig intestinal mucosa vitamin D receptor (VDR)
and for the human vitamin D binding protein (hDBP), (2) the
prodifferentiating activity, tested on promyelocytic leukemia
HL60 cells, (3) the in vitro antiproliferative effects, determined
on breast cancer MCF-7 cells, and (4) the in vivo calcemic activ-
ity, tested in vitamin D-replete NMRI mice. Results are shown
in Tables 2–3.
Discussion
The search for analogs showing a separation in prodifferentiat-
ing and calcemic activities followed the 1981 report by Suda
and his associates who indicated for the first time that 1,25-D3
inhibits the proliferation of myeloid leukemic cells and induces
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
262

View Article Online
Table 2 Biological activities of 1,25-D3 analogs 1–2a
Binding affinity
Calcemic effect
Cell differentiation and proliferation
VDR
(Pig)
DBP
(Human)
Serum
(Mice)
Code
HL-60
MCF-7
Kerat.b
1,25-D3 (1at)
100
90
90
—
90
80
75
8
100
0.2
40
—
7
1
2
100
3000
150
—
1000
70
100
5000
90
—
2250
85
100
2200
—
—
1400
70
100
800
2
20-epi-1,25-D3 (1bt)
19-nor-1,25-D3 (1ar)
19-nor-20-epi- (1br)c
CY 10012 (2at)
CY 943 (2bt)
—
200
<0.1
100
<0.1
CY 10010 (2ar)
CY 941 (2br)
1000
10
1000
30
1000
30
0
a All data resulted from evaluation in the Legendo laboratory at the KULeuven. The activities are presented as relative values, the reference value of
1,25-D3 (1at) being defined as 100. Further details about the methodology are given in the Experimental Section. b Keratinocytes. c The correspond-
ing activities of this analog have not yet been described in the current literature.
Table 3 Biological activities of 1,25-D3 analogs 3–8a
Binding affinity
Cell differentiation
and proliferation
Calcemic effect
VDR
(Pig)
DBP
(Human)
Serum
(Mice)
Code
HL-60
MCF-7
Pre-1,25-D3 (3at)b
GAO 306 (3ar)
IM 9053 (4at)
GAO 182 (4ar)
GAO 315 (5at)
GAO 205 (5ar)
IM 902 (6at)
GAO 183 (6ar)
GAO 320 (7at)
GAO 305 (7ar)
IM 9102 (8at)
GAO 181 (8ar)
4
5
0.5
8
7
9
7
<5
8
700
85
7
—
30
2000
350
30
—
<0.25
20
2
—
<0.1
10
10
—
—
4
9
60
20
80
20
75
30
20
30
50
85
80
3
10
70
100
9
20
100
300
4
8
85
600
100
2300
30
30
100
2200
0.5
10
a All data resulted from evaluation in the Legendo laboratory at the KULeuven. The activities are presented as relative values, the reference value of
1,25-D3 (1at) being defined as 100. Further details about the methodology are given in the Experimental Section. b Pentadeuterio d5 derivative, ref. 66.
biological activity when compared to the derivative with the
natural configuration. In spite of the usefulness of the active
volume concept as illustrated above with 2a and 2b, it should be
stressed that it does not provide a solution for the discovery of
analogs with dissociation in activity since virtually without
exception the more potent derivatives in cell differentiation and
antiproliferation also possess a higher calcemic activity.
The behavior of the classical 20-epi analogs, such as 20-epi-
1,25-D3 (1bt, MC 1288) and KH 1060, is controversial. Indeed,
whereas the nuclear receptor activation mechanism somehow
implies that the affinity of the ligand for the receptor is pro-
portional to the activation, in the case of these 20-epi analogs
the affinity for the n-VDR is similar or even lower than that of
1,25-D3, whereas their transcriptional activity is several orders
of magnitude higher. The molecular origin of this behavior is
still under discussion.64,65
involving the biological evaluation of analogs locked into the
s-cis geometry have substantiated this model.70
Pre-1,25-D3 (3at) is unique since it can be considered as a
tautomer of 1,25-D3 in the cisoid conformation.71 Whereas the
direct screening of the natural previtamin form is problematic,
due to its propensity (even though slow) to convert into the
active vitamin form, this problem was circumvented in an ele-
gant way by Okamura and co-workers through the study of the
kinetically more stable 9,14,19,19,19-pentadeuterio previtamin
derivative.9,66 An alternative but more drastic solution resides in
the study of the 19-nor derivative (3ar), which obviously is
unable to isomerise to the vitamin form.72 Both studies have
shown that as far as the transcriptional activity is concerned the
previtamin form has lost most of its biological activity in vitro
and in vivo. In the same context the previtamin form of the
trans-decalin analog (4at) is an attractive derivative to study.
Indeed, whereas in the natural hydrindane series the pre-
vitamin–vitamin equilibrium composition between 1at and 3at
is largely in favor of the vitamin derivative (90 : 10), in the case
of the trans-decalin analogs this is less pronounced, i.e. the ratio
of 2at : 4at equals 75 : 25 and the latter equilibrium is very
slowly installed (5 h at 85ЊC in CD3CN). In both cases subtle
conformational preferences related to the unfavourable position
of the previtamin ∆8 bond in the trans-fused system lead to the
preferred formation of the vitamin form,73 but this is less pro-
nounced in the trans-decalin case. The results of the biological
screening were quite unexpected since they showed 4at to pos-
sess a remarkably high prodifferentiating (HL-60) and anti-
proliferative (MCF-7) activity, with no affinity for the n-VDR;
the calcemic activity was moderately reduced. In order to
rule out any possible prior conversion to the vitamin form, the
The previtamin form
Pre-1,25-D3 (3at, as a pentadeuterio derivative) has been
shown to stimulate nongenomic but not genomic biological
responses.66 Since the rapid nongenomic responses have been
attributed to the interaction of 1,25-D3 and a putative mem-
brane receptor protein,67 the interesting hypothesis was raised
that different conformations of 1,25-D3, readily accessible
through rotation about the C-6,C-7 bond, would be involved in
binding to the nuclear and membrane receptor proteins.68
Whereas the more stable extended transoid geometry has been
shown experimentally to be involved in the binding in the LBD
of the n-VDR,11 the compact steroid-like cisoid geometry
would be involved in the binding to the m-VDR.69 Studies
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
263

View Article Online
Fig. 3 Volume maps representing the conformational behavior of the side chain of 1,25-D3 (1a), 20-epi-1,25-D3 (1b), and trans-decalin analogues
2a and 2b (truncated model); colored spheres indicate positions of O-25: yellow to brown represent conformations within 0–5, 5–10, 10–15, and
15–20 kJ molϪ1, respectively, of the minimum energy form (grey). Top: 1,25-D3 (1a); front view (left); top view (right); indicated EA region following
ref. 61; active volume (red) according to ref. 62; position of O-25 in receptor-bound conformation (green) taken from ref. 11. Middle: front view of
1,25-D3 (1a) (left) and 20-epi-1,25-D3 (1b) (right). Bottom: front view of trans-decalin analogues 2a (left) and 2b (right).
corresponding 19-nor-previtamin 4ar was also synthesised
and its biological activity assessed. This further modification
results in an approximate 10-fold reduction in activity com-
pared to 4at (HL-60, MCF-7, and calcemic effects), but 4ar
still possesses a substantial prodifferentiative and antiprolifer-
ative activity. Does this unexpected result originate from a
genuine previtamin triene effect or rather from a trans-decalin
effect? In order to settle this unambiguously 3ar was resynthe-
sised and fully evaluated.74 Comparison of the results obtained
for 3ar and 4ar confirm that the trans-decalin structure is
responsible.
counterparts with the trans-fused decalin system. Hence again
the origin of the biological behaviour may be attributed to the
particularity of the trans-decalin system.
In conclusion, several analogs have been developed, charac-
terised by a trans-decalin modified central CD-ring system,
which show surprising biological activity when compared
to the corresponding derivatives with a natural hydrindane
structure. It is expected that these derivatives will be useful
in the discovery of the detailed molecular mechanism of the
transcriptional activity of 1,25-D3 and related compounds.
Experimental
The seco-B-ring modification
Detailed information about the preparation of all derivatives,
except a few general procedures of key experiments, is
deposited as Electronic Supplementary Information (ESI).
The unexpected biological activity of a previtamin derivative
prompted us to also screen the yne–diene derivatives 6 that were
originally synthesised as precursor molecules for the previtamin
derivatives.75 As the hydrogenation procedure easily leads to full
reduction of the C-6,C-7 triple bond, the corresponding satur-
ated analogs 8 were also prepared and screened. Again the
results were unexpected. Both 6at and 8at possess an activity
quite similar to 1,25-D3, except for a reduced calcemic activity.
The corresponding 19-nor derivatives 6ar and 8ar have an
almost identical profile with, however, a much more pro-
nounced prodifferentiative (HL-60) and antiproliferative activ-
ity (MCF-7) compared to 6at and 8at. Again, the origin of this
behaviour was examined through the evaluation of the corre-
sponding natural CD-ring derivatives, i.e. 5at, 5ar, 7at, and 7ar.
These compounds, however, have only a marginal vitamin D
like activity profile and are 10 to 100 times less potent than their
Coupling reaction of ketones 26a and 26b with phosphine oxides
27 and 28 – general procedure
To a solution of A-ring phosphine oxide 27 (28) (0.67 mmol) in
dry THF (7.5 cm3) was added dropwise n-BuLi (2.5 M solution
in hexanes, 0.24 cm3, 0.60 mmol) at Ϫ78 ЊC under Ar. The
formed dark red solution was stirred at Ϫ78 ЊC for 1 h and a
solution of ketone 26a (26b) (0.17 mmol) in dry THF (2.5 cm3)
was added dropwise. The red solution was stirred at Ϫ78 ЊC for
2 h and was then allowed to warm to rt. The reaction mixture
was loaded onto a silica gel column, the reaction product was
eluted (n-hexane–EtOAc, 5 : 1) and further purified by HPLC
to give the protected 1,25-D3 analog.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
264

View Article Online
Desilylation of the protected 1,25-D3 analogs using TBAF –
general procedure
L. Binderup); Ro 23–7553 was a gift from Hoffmann-La Roche
Inc. (USA).
To a solution of the protected 1,25-D3 analog (0.14 mmol) in
THF (4 cm3) was added tetrabutylammonium fluoride (TBAF,
1 M solution in THF, 2.1 cm3, 2.1 mmol). The reaction mixture
was stirred at rt for 12 h and then loaded onto a silica gel
column. The reaction product was eluted (n-pentane–acetone,
1 : 1) and further purified by HPLC to give the 1,25-D3 analog.
Binding studies
The affinity of the decalin analogs of 1,25-D3 to the vitamin
D receptor was evaluated by their ability to compete with
[3H]1,25-D3 for binding to high speed supernatant from intes-
tinal mucosa homogenates obtained from normal pigs as
described previously.17b The relative affinity of the analogs was
calculated from their concentration needed to displace 50% of
[3H]1,25-D3 from its receptor compared with the activity of
1,25-D3. Binding to hDBP was performed at 4 ЊC essentially as
described previously.76
Coupling reaction of enol triflates 29a and 30a with A-ring
intermediates 31 and 32 – general procedure
To a solution of enol triflate 29a (30a) (0.5 mmol) and A-ring
synthon 31 (32) (0.5 mmol) in a mixture of Et2NH (2.5 cm3)
and CH3CN (5 cm3) were added CuI (10 mg, 0.05 mmol)
and bis(triphenylphosphine)palladium() acetate (Pd(PPh3)2-
(OAc)2, 37.5 mg, 0.05 mmol). The reaction mixture was stirred
at rt for 1 h and diluted with n-pentane–EtOAc 99 : 1 (25 cm3).
The solution was passed through a short pad of silica gel, which
was rinsed with n-pentane–EtOAc 99 : 1 (250 cm3). Concen-
tration under reduced pressure left a residue, which was purified
by HPLC to give the protected 1,25-D3 analog.
MCF-7 proliferation assay
The human breast carcinoma (MCF-7) cell line was obtained
from the American Tissue Culture Company (Rockville, MD).
The antiproliferative activity on MCF-7 cells was assessed by
evaluating [3H]thymidine incorporation. Cells were seeded in
96-well plates (7500 cells per well) and 1 µCi [methyl-3H]thymid-
ine (ICN Biomedicals, Costa Mesa, CA) was added 72 h after
the initiation of treatment. Cells were semi-automatically har-
vested after an additional 6 h of incubation on filter plates only
retaining incorporated thymidine (GF/C Filter and Filtermate
Universal Harvester, Packard Instrument, Meriden, CT).
Counting was performed using a microplate scintillation
counter (Topcount, Packard).
Semi-hydrogenation and hydrogenation of the yne–diene type
derivatives – general procedure
A mixture of the ynediene 33 or 34 (0.029 mmol), Lindlar
catalyst (0.02 g), and quinoline (2% solution in n-pentane,
0.1 cm3) in EtOAc (2 cm3) was stirred under a H2 atmosphere at
rt. The reaction was carefully monitored by TLC for dis-
appearance of the starting material. The reaction mixture was
then passed through a pad of silica gel and concentrated under
reduced pressure to yield the semi-hydrogenated derivative. In
the absence of quinoline in the reaction mixture the procedure
gave the corresponding derivative with a fully hydrogenated
triple bond.
Differentiation of HL-60 cells
The human promyelocytic leukemia cell line (HL-60) was
obtained from the American Tissue Culture Company
(Rockville, MD). HL-60 cells were seeded at 4 × 104 cells cmϪ3
in 25 cm2 Falcon tissue chambers using RPMI 1640 medium
supplemented with 20% fetal calf serum (Sera-Lab, W. Sussex,
UK) and gentamycin (50 µg cmϪ3; Gibco, Roskilde, Denmark).
1,25-D3, analogs or vehicle were added the day after plating.
After 4 days of culture, differentiation was measured by using
the NBT reduction assay as described previously.77
Molecular modelling and conformational analysis
All modelling was performed by molecular mechanics calcu-
lations using the MacroModel V6.0 molecular modelling
programme of Still36 run on a Silicon Graphics Octane work-
station and using the programme’s standard parameters. Con-
formational analysis of the side chain of compounds 1a, 1b,
2a and 2b was carried out on model compounds in which the
A-ring and the diene system up to C6 were replaced by a H
atom. Rotations with 60Њ increments were applied to the rotat-
able C–C bonds of the side chain, while the 25-OH was rotated
with increments of 120Њ. The so generated 23,328 starting
conformations were minimised using the MM2* force field
implementation of the programme and the conformations
within 20 kJ molϪ1 of the minimum energy form were retained.
All conformations of each compound were then overlaid using
C-13 as common origin (x, y, z = 0), C-14 was positioned in the
yz-plane (x = 0) and the position of C-18 was made to coincide
with the positive y-axis (x, z = 0). A ball-and-stick model was
generated of the minimum energy conformation and the pos-
ition of O-25 in each of the local energy minima within the
given energy window was represented by a yellow–brown sphere
to obtain the views shown in Fig. 3. The procedure for the
calculation of the active volume shown as a red sphere has been
reported earlier.62 The position of O-25 of 1,25-D3 as bound to
the n-VDR and shown as a green sphere was taken from the
work of Moras et al.11
In vivo studies
NMRI mice were obtained from the Proefdierencentrum
of Leuven (Belgium) and fed on a vitamin D-replete diet
(0.2% calcium, 1% phosphate, 2000 U vitamin D kgϪ1; Hope
Farms, Woerden, The Netherlands). The calcemic effect of the
decalin analogs was tested in NMRI mice by daily subcutane-
ous injection of 1,25-D3, its analogs or the solvent during 7
consecutive days, using serum calcium concentration as a
parameter.
X-Ray crystal structure analysis of compound 1778
Single crystals suitable for X-ray analysis were obtained by re-
crystallisation from n-pentane.
Crystal data
C27H35NO2; M = 405.56 g molϪ1; monoclinic, space group P21,
a = 12.368(3), b = 17.711(4), c = 21.575(4)Å, β = 92.382(4)Њ,
U = 4722(2)Å3, T = 293(2), Z = 8, dcalcd = 1.141 Mg/m3,
µ(Mo–Kα) = 0.071 mmϪ1, 32,705 reflections collected, 13,128
unique (Rint = 0.032), 7,859 observed reflections [I > 2σ(I )].
Structure solution and refinement
Biological evaluation
Solution by direct methods,79 refinement by full matrix least-
squares80 on F 2, non-H atoms with anisotropic displacement
parameters, H atoms geometrically positioned and treated as
riding atoms; data : parameters ratio: 13128 : 1081; R = 0.0420
(observed data), Rw(F 2) = 0.1120 (all data).
1,25-D3 (1at) was a gift from Duphar (Weesp, The Netherlands;
J. P. van de Velde); OCT was a gift from Chugai Pharmaceutical
Co (Japan); MC 1288 (1bt), MC 903, and KH 1060 were
kind gifts from Leo Pharmaceuticals (Ballerup, Denmark;
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
265

View Article Online
20 For 6D-ring analogs, see: (a) B. Linclau, P. De Clercq and
M. Vandewalle, Bioorg. Med. Chem. Lett., 1997, 7, 1461–1468;
(b) S. Gabriëls, D. Van Haver, M. Vandewalle, P. De Clercq,
A. Verstuyf and R. Bouillon, Chem. Eur. J., 2001, 7, 520–532.
21 For a preliminary report, see: Y.-J. Chen, P. De Clercq and
M. Vandewalle, Tetrahedron Lett., 1996, 37, 9361–9364.
22 E. Havinga, Experientia, 1973, 29, 1181–1316.
23 For the sake of consistency the atom numbering used through-
out the paper follows the conventional steroid nomenclature; see:
I. A. Rose, K. R. Hanson, K. D. Wilkinsons and M.-J. Wimmer,
Proc. Natl. Acad. Sci. USA, 1980, 77, 2439–2441.
24 For an authoritative and comprehensive review, see: G.-D. Zhu and
W. H. Okamura, Chem. Rev., 1995, 95, 1877–1952.
25 For a review of the major synthetic approaches to the triene unit of
vitamin D, see: H. Dai and G. H. Posner, Synthesis, 1994, 1383–
1398.
Description of the structure
The crystal structure of 17 contains four independent molecules
in the unit cell, which have been indicated as molecules A, B, C
and D. They differ in conformation: molecules A and B (Fig. 2)
have very different conformations of their side chains, while
molecule C is similar to B and molecule D to A. The crystal
packing is governed by rather weak intermolecular hydrogen
bonds. More details are provided in Table 4 (relevant torsion
angles of the side chain of 17) and in Table 5 (geometry of the
intermolecular hydrogen bonds in the crystal structure of 17),
and are found in the Electronic Supplementary Information
(ESI).
26 B. Lythgoe, T. A. Moran, M. E. N. Nambudiry, J. Tideswell and
P. W. Write, J. Chem. Soc., Perkin Trans. 1, 1978, 590–595.
27 E. G. Baggiolini, J. A. Iacobelli, B. M. Hennesy, A. D. Batcho,
J. F. Sereno and M. R. Uskokovic, J. Org. Chem., 1986, 51, 3098–
3108.
28 K. L. Perlman, R. E. Swenson, H. E. Paaren, H. K. Schnoes and
H. F. DeLuca, Tetrahedron Lett., 1991, 32, 7663–7666.
29 T. M. Dawson, J. Dixon, P. S. Littlewood, B. Lythgoe and A. K.
Saksena, J. Chem. Soc. (C), 1971, 2960–2966.
30 L. Castedo, A. Mouriño and L. A. Sarandeses, Tetrahedron Lett.,
1986, 27, 1523–1526.
31 Obtained from readily available (ϩ)-Wieland–Miescher ketone in
3 steps (ref. 21). See also: (a) M. Hirayama, S. Fukatsu and
N. Ikekawa, J. Chem. Soc., Perkin Trans. 1, 1981, 88–93;
(b) M. Ando, S. Sayama and K. Takese, J. Org. Chem., 1985, 50,
251–264.
Acknowledgements
Financial assistance of the “Fonds voor Wetenschappelijk
Onderzoek Vlaanderen” (project G.0150.02, G.0036.00) is
gratefully acknowledged.
References
1 (a) Vitamin D, eds. D. Feldman and F. H. Glorieux, Academic Press,
San Diego, USA, 1997 (1285 pages); (b) Proceedings of the Eleventh
Workshop on Vitamin D, eds. A. W. Norman, R. Bouillon and
M. Thomasset, University of California, Riverside, USA, 2000,
p. 1–992, and the previous 10 volumes in this series.
2 A. Windaus and E. Auhagen, Z. Physiol. Chem., 1931, 196, 108–120.
3 M. F. Holick, H. K. Schnoes and H. F. DeLuca, Proc. Natl. Acad.
Sci. USA, 1971, 68, 803–804.
32 A. K. Bose and R. T. Dahille Jr., J. Org. Chem., 1965, 30, 505–
509.
4 D. E. M. Lawson, D. R. Fraser, E. Kodicek, H. R. Morris and
D. H. Williams, Nature, 1971, 230, 228–230.
5 A. W. Norman, J. F. Myrtle, R. J. Midgett, M. G. Nowicki,
V. Williams and G. Popjak, Science, 1971, 173, 51–54.
6 R. Bouillon, W. H. Okamura and A. W. Norman, Endocr. Rev.,
1995, 16, 200–257.
7 A. W. Norman, H. L. Henry, J. E. Bishop, X.-D. Song, C. Bula and
W. H. Okamura, Steroids, 2001, 66, 147–158.
8 L. Velluz, G. Amiard and B. Goffinet, Bull. Soc. Chim. Fr., 1955, 22,
1341–1348.
33 (a) J. A. Profitt, D. S. Watt and E. J. Corey, J. Org. Chem., 1975, 40,
127–128; (b) P. J. Garratt, C. W. Doecke, J. C. Weber and L. A.
Paquette, J. Org. Chem., 1986, 51, 449–452.
34 A. Zarecki and J. Wicha, Synthesis, 1996, 455–456.
35 It is interesting to note that the stereochemical outcome of the
alkylation of the analogous alkoxycarbonylmethyl derivative is
different: see ref. 21.
36 F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp,
M. Lipton, C. Caufield, G. Chang, T. Hendrickson and W. C. Still,
J. Comput. Chem., 1990, 11, 440–467.
37 The main difference between the four molecules (A, B, C, and D)
found in the unit cell of crystalline 17 is that A and D possess a
gauche conformation at C(17)–C(20)–C(22)–O(23), and C and B an
anti conformation. More details are given in the Experimental
Section.
38 (a) M. Van Gool and M. Vandewalle, Eur. J. Org. Chem., 2000,
3427–3431.; See also: (b) M. P. Bosch, F. Camps, J. Coll,
A. Guerrero, T. Tatsuoka and J. Meinwald, J. Org. Chem., 1986, 51,
773–784.
39 L. M. Jackman and S. Sternell, in Application of NMR Spectroscopy
in Organic Chemistry, 2nd edn, Pergamon, Oxford and New York,
1969, p. 243.
9 M. L. Curtin and W. H. Okamura, J. Am. Chem. Soc., 1991, 113,
6958–6966.
10 A. W. Norman, W. H. Okamura, M. W. Hammond, J. E. Bishop,
M. C. Dormanen, R. Bouillon, H. Van Baelen, A. L. Ridall,
E. Daane, R. Khoury and M. C. Farach–Carson, Mol. Endocrinol.,
1997, 11, 1518–1531.
11 N. Rochel, J. M. Wurtz, A. Mitschler, B. Klaholtz and D. Moras,
Mol. Cell, 2000, 5, 173–179.
12 S. Yamada, K. Yamamoto, H. Masuno and M. Choi, Steroids, 2001,
66, 177–187.
13 L. Binderup, S. Latini, E. Binderup, C. Bretting, M. Calverley and
K. Hansen, Biochem. Pharmacol., 1991, 42, 1569–1575.
14 K. L. Perlman, R. R. Sicinski, H. K. Schnoes and H. F. DeLuca,
Tetrahedron Lett., 1990, 31, 1823–1824.
40 J. L. Mascareñas, J. Pérez–Sestelo, L. Castedo and A. Mouriño,
Tetrahedron Lett., 1991, 32, 2813–2816.
15 R. Bouillon, P. De Clercq, P. Pirson and M. Vandewalle, Novel
Structural Analogues of Vitamin D, Patent PCT/EP 93.202037.3,
1993 (Chem. Abstr., 1996, 122, 314933m).
41 J. Petrier, C. Dupuy and J. L. Luche, Tetrahedron Lett., 1986, 27,
3149–3152.
42 The difference in steric energy between the cis- and trans-fused
decalones 25 and 10 (side chain truncated to an isopropyl group)
calculated using MacroModel V6.0 (ref. 36) was found to be 3.4 kJ
moleϪ1 in favor of the trans-fused decalones 10.
16 For the 14-epi modification, see: L. Verlinden, A. Verstuyf,
M. Van Camp, S. Marcelis, K. Sabbe, X.-Y. Zhao, P. De Clercq,
M. Vandewalle and R. Bouillon, Cancer Res., 2000, 60, 2673–2679.
17 For 5E-ring analogs, see: (a) K. Sabbe, C. D’Halleweyn, P. De
Clercq, M. Vandewalle, R. Bouillon and A. Verstuyf, Bioorg. Med.
Chem. Lett., 1996, 6, 1697–1702; (b) A. Verstuyf, L. Verlinden,
H. Van Baelen, K. Sabbe, C. D’Halleweyn, P. De Clercq,
M. Vandewalle and R. Bouillon, J. Bone Miner. Res., 1998, 13, 549–
558.
18 For 6C-ring analogs, see: (a) G.-D. Zhu, Y. Chen, X. Zhou,
M. Vandewalle, P. J. De Clercq, R. Bouillon and A. Verstuyf,
Bioorg. Med. Chem. Lett., 1996, 6, 1703–1708; (b) X. Zhou,
G.-D. Zhu, D. Van Haver, M. Vandewalle, P. J. De Clercq,
A. Verstuyf and R. Bouillon, J. Med. Chem., 1999, 42, 3539–3556.
19 For 5D-ring analogs, see: (a) Y. Wu, S. Ling, C. D’Halleweyn,
D. Van Haver, P. De Clercq, M. Vandewalle, R. Bouillon
and A. Verstuyf, Bioorg. Med. Chem. Lett., 1997, 7, 923–928;
(b) A. Verstuyf, L. Verlinden, E. Van Etten, S. Ling, Y. Wu,
C. D’Halleweyn, D. Van Haver, G.-D. Zhu, Y.-J. Chen, X. Zhou,
M. R. Haussler, P. De Clercq, M. Vandewalle, H. Van Baelen,
C. Mathieu and R. Bouillon, J. Bone Miner. Res., 2000, 15, 237–252.
43 P. J. Kociensky, B. Lythgoe and I. Waterhouse, J. Chem. Soc.,
Perkin Trans. 1, 1980, 1045–1050.
44 (a) E. G. Baggiolini, J. A. Iacobelli, B. M. Hennesy and M. R.
Uskokovic, J. Am. Chem. Soc., 1982, 104, 2945–2948; . For a recent,
optimised and very efficient synthesis of 27, see: (b) A. R.
Daniewski, L. M. Garofalo, S. D. Hutchinys, M. M. Kabat, W. Liu,
M. Okabe, R. Radinov and G. P. Yiannikouros, J. Org. Chem., 2002,
67, 1580–1587.; For an interesting review on the practical synthesis
of vitamin D analogs from the industrial point of view, including a
review of syntheses of 27, see: (c) M. M. Kabat and R. Radinov,
Curr. Opin. Drug Dis. Dev., 2001, 4, 808–833.
45 For an alternative synthesis of 28, see: S.-Z. Zhou, S. Anné and
M. Vandewalle, Tetrahedron Lett., 1996, 37, 7637–7640.
46 (a) J. M. Aurrecoechea and W. H. Okamura, Tetrahedron Lett., 1987,
28, 4947–4950; (b) W. H. Okamura, J. M. Aurrecoechea, R. A. Gibbs
and A. W. Norman, J. Org. Chem., 1989, 54, 4072–4083.
47 L. A. Sarandeses, J. L. Mascareñas, L. Castedo and A. Mouriño,
Tetrahedron Lett., 1992, 33, 5445–5448.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
266

View Article Online
48 J. L. Mascareñas, L. A. Sarandeses, L. Castedo and A. Mouriño,
Tetrahedron, 1991, 47, 3485–3498.
49 R. R. Sicinski and H. F. DeLuca, Bioorg. Med. Chem. Lett., 1995, 5,
159–162.
50 (a) J. B. Hendrickson and R. Bergeron, Tetrahedron Lett., 1973,
4607–4610; (b) J. E. McMurry and W. J. Scott, Tetrahedron Lett.,
1983, 24, 979–982.
51 E. Abe, C. Miyaura, A. Sakagami, M. Takida, K. Komo,
T. Yamazaki, S. Yoshiki and T. Suda, Proc. Natl. Acad. Sci. USA,
1981, 78, 4990–4994.
52 (a) E. Murayama, K. Miyamoto, M. Kubodera, T. Mori and
I. Matsunaga, Chem. Pharm. Bull. (Tokyo), 1986, 34, 4410–4413;
(b) J. Abe, M. Morikawa, K. Miyamoto, S. Kaiko, M. Fukushima,
C. Miyaura, E. Abe, T. Suda and Y. Nishii, FEBS Lett., 1987, 226,
58–62; (c) Y. Nishii and T. Okano, Steroids, 2001, 66, 137–146.
53 (a) M. J. Calverley, Tetrahedron, 1987, 43, 4609–4619; (b) L.
Binderup and E. Bramm, Biochem. Pharmacol., 1988, 37, 889–895.
54 J. Y. Zhou, A. W. Norman, D. Chen, G. W. Sun, M. R. Uskokovic
and H. P. Koeffler, Proc. Natl. Acad. Sci. USA, 1990, 87, 3929–3932.
55 P. H. Mäenpää, S. Väisänen, T. Jääskeläinen, S. Ryhänen,
J. Rouvinen, C. Duchier and A. Mahonen, Steroids, 2001, 66,
223–226.
56 G. H. Posner, J. K. Lee, Q. Wang, S. Peleg, M. Burke, H. Brem,
P. Dolan and T. W. Kernler, J. Med. Chem., 1998, 41, 3008–3014.
57 T. Okano, N. Tsugawa, S. Masuda, A. Takeuchi, T. Kobayashi,
Y. Takita and Y. Nishii, Biochem. Biophys. Res. Commun., 1989, 163,
1444–1449.
65 G. Tocchini–Valentini, N. Rochel, J. M. Wurtz, A. Mitschler and
D. Moras, Proc. Natl. Acad. Sci. USA, 2001, 98, 5491–5496.
66 A. W. Norman, W. H. Okamura, M. C. Farach–Carson,
K. Allewaert, D. Branisteanu, I. Nemere, K. R. Madalidharan and
R. Bouillon, J. Biol. Chem., 1993, 268, 13811–13819.
67 I. Nemere, M. C. Dormanen, M. W. Hammond, W. H.
Okamura and A. W. Norman, J. Biol. Chem., 1994, 269, 23750–
23756.
68 M. C. Dormanen, J. E. Bishop, M. W. Hammond, W. H. Okamura,
I. Nemere and A. W. Norman, Biochem. Biophys. Res. Commun.,
1994, 201, 394–401.
69 W. H. Okamura, M. M. Midland, A. W. Norman, M. W. Hammond,
M. C. Dormanen and I. Nemere, Ann. N.Y. Acad. Sci., 1995, 761,
344–348.
70 A. W. Norman, W. H. Okamura, M. W. Hammond, J. E. Bishop,
M. C. Dormanen, R. Bouillon, H. Van Baelen, A. L. Ridall,
E. Daane, R. Khoury and M. C. Farach–Carson, Mol. Endocrinol.,
1997, 11, 1518–1531.
71 For a recent interesting study aiming at the synthesis of a stable
transition-state mimic of the vitamin–previtamin equilibrium, see:.
(a) R. Hayashi, S. Fernández and W. H. Okamura, Org. Lett., 2002,
4, 851–854.; See also: (b) E. M. Cosdesido, J. R. Rodríguez,
L. Castedo and J. R. Granja, Org. Lett., 2002, 4, 1651–1654.
72 R. Bouillon, L. A. Sarandeses, K. Allewaert, J. Zhao,
J. L. Mascareñas, A. Mouriño, H. Van Baelen, S. Vrielynck,
P. De Clercq and M. Vandewalle, J. Bone Miner. Res., 1993, 8, 1009–
1015.
58 H. Takayama, K. Konno, T. Fujishima, S. Maki, L. Liu, D. Miura,
M. Chokki, S. Ishizuka, C. Smith, H. F. DeLuca, K. Nakagawa,
M. Kurobe and T. Okano, Steroids, 2001, 66, 277–286.
59 (a) A. Kutner, H. Zhao, H. Fitak and S. R. Wilson, Bioorg. Chem.,
1995, 23, 21–32; (b) M. F. Boehm, P. Fitzgerald, A. Zou,
M. G. Elgort, E. D. Bischoff, L. Mere, D. E. Mais, R. P. Bissonnette,
R. A. Heyman, A. M. Nadzan, M. Reichman and E. A. Allegretto,
Chem. Biol., 1999, 6, 265–275.
60 (a) W. H. Okamura, J. A. Palenzuela, J. Plumet and M. M. Midland,
J. Cell. Biochem., 1992, 49, 10–18; (b) M. M. Midland, J. Plumet and
W. H. Okamura, Bioorg. Med. Chem. Lett., 1993, 3, 1799–1804.
61 (a) K. Yamamoto, W. Y. Jun, M. Ohta, K. Hamada, H. F. DeLuca
and S. Yamada, J. Med. Chem., 1996, 39, 2727–2737; (b) S. Yamada,
K. Yamamoto, H. Masuno and M. Ohta, J. Med. Chem., 1998, 41,
1467–1475.
62 D. Van Haver and P. J. De Clercq, Bioorg. Med. Chem. Lett., 1998, 8,
1029–1034. For applications of the procedure, see ref. 18b and 20b.
63 According to a Boltzmann distribution at 298 K, the occupation of
the active volume by the 25-hydroxy group of the side chain was
calculated to correspond to 69 mole% in the case of 2a and to 0.1
mole% in the case of 2b.
73 J. L. M. A. Schlatmann, J. Pot and E. Havinga, Rec. Trav. Chim.,
1964, 83, 1173–1184.
74 Although 3ar has been described previously (ref. 72) some screening
tests had not been performed at that time: MCF-7, calcemic effects
in mice measured by calcium levels in serum.
75 For syntheses of yne-diene intermediates, see: (a) ref. 9 and 47.;
(b) M. Díaz, M. Ferrero, S. Fernández and V. Gotor, Tetrahedron
Lett., 2000, 41, 775–779; (c) M. Díaz, M. Ferrero, S. Fernández and
V. Gotor, J. Org. Chem., 2000, 65, 5647–5652.
76 R. Bouillon, H. Van Baelen and P. De Moor, J. Steroid Biochem.,
1980, 13, 1029–1034.
77 V. K. Ostrem, Y. Tanaka, J. Prahl, H. F. DeLuca and N. Ikekawa,
Proc. Natl. Acad. Sci. USA, 1987, 84, 2610–2614.
78 Crystallographic data (excluding structure factors) for the structure
reported in this paper have been deposited with the Cambridge
Crystallographic Data Centre. CCDC reference number 193947. See
http://www.rsc.org/suppdata/ob/b2/b209147j/ for crystallographic
files in .cif or other electronic format.
79 M. C. Burla, M. Camalli, B. Carrozzini, G. L. Cascarano,
C. Giacovazzo, G. Polidori and R. Spagna, Acta. Crystallogr. Sect.
A, 2000, 56, 451–457.
64 S. Peleg, M. Sastry, E. D. Collins, J. E. Bishop and A. W. Norman,
J. Biol. Chem., 1995, 270, 10551–10558.
80 G. M. Sheldrick, SHELXL-97, 1997, University of Göttingen,
Germany (web site: http://shelx.uni-ac.gwdg.de/SHELX/).
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 7 – 2 6 7
267