Novel side chain analogs of 1α,25-dihydroxyvitamin D3: Design and synthesis of the 21,24-methano derivatives

Article abstract of DOI:10.1016/S0039-128X(00)00156-2

The syntheses of the new 21,24-methano derivatives of 1α,25-dihydroxyvitamin D₃ [viz. 1(S),3(R)-dihydroxy-17(R)-(1′,4′-cis-(4′-(1′-hydroxy-1′-methylethyl)-cyclo-hexyl))-9,10-seco-androsta-5(Z),7(E),10(19)-triene (MC 2108) and its (1′,4′-trans)-

Full text of DOI:10.1016/S0039-128X(00)00156-2

Steroids 66 (2001) 249–255
Novel side chain analogs of 1␣,25-dihydroxyvitamin D3: design and
synthesis of the 21,24-methano derivatives
Martin J. Calverley*
Department of Medicinal Chemistry, Leo Pharmaceutical Products, 55 Industriparken, DK-2750 Ballerup, Denmark
Abstract
The syntheses of the new 21,24-methano derivatives of 1␣,25-dihydroxyvitamin D3 [viz. 1(S),3(R)-dihydroxy-17(R)-(1Ј,4Ј-cis-(4Ј-(1Ј-
hydroxy-1Ј-methylethyl)-cyclo-hexyl))-9,10-seco-androsta-5(Z),7(E),10(19)-triene (MC 2108) and its (1Ј,4Ј-trans)-isomer (MC 2110)] are
described. The key step is the establishment, by Diels-Alder reaction on a CD-ring side chain diene intermediate prepared from vitamin D2,
of a 1,4-disubstituted cyclohexene moiety in the side chain. Hydrogenation to a 1:1 mixture of cis and trans cyclohexane derivatives and
separation of the two series at a stage prior to the standard Horner-Wittig coupling with the (Hoffmann-La Roche) ring-A building block
were other important steps in the syntheses of the target analogs. The relative configurations of intermediates were assigned by NMR
spectroscopy. MC 2108 and MC 2110 are of interest as conformationally locked side chain derivatives to probe the receptor interactions
of not only the parent vitamin D hormone but also its biologically active symmetrical ‘double side chain’ analog [21-(3Ј-hydroxy-3Ј-
methylbutyl)-9,10-seco-cholesta-5(Z),7(E),10(19)-triene-1(S),3(R),25-triol (MC 2100)], ‘both’ side chains of which can formally be traced
out in the new analogs. The preferred conformations, inferred from an analysis of 13C-NMR characteristics, notably the chemical shift of
C-17 in a series of analogs, to have the tertiary alcohol (1Ј-hydroxy-1Ј-methylethyl) substituent equatorial on the cyclohexane chair, are
confirmed by molecular modeling. © 2001 Elsevier Science Inc. All rights reserved.
Keywords: Sterol; Vitamin D analogs; Side chain conformation; Diels-Alder reaction; 13C-NMR spectroscopy; Molecular modeling
1. Introduction
of other analogs [2–4] prompted both theoretical studies
using molecular mechanics based conformational analysis
[5] and an X-ray study [6], revealing that the side chain
backbone shows a different directional preference in the
unnatural epi-series. Other series have been similarly ana-
lyzed by molecular modeling, and a correlation with bio-
logic results was achieved [7]. Furthermore, analogs have
been designed in which particular regions of space (for the
isolated molecule) can be energetically excluded or made
accessible to the side chain hydroxyl, and the correlation
with biologic potency was remarkable [8]. Thus the concept
of an ‘active’ conformation for the side chain of the vitamin
D hormone is emerging. On the other hand, limited protease
digestion studies [9,10] (in which the bound ligands stabi-
lize different fragments of the receptor) suggest that the side
chains of the 20-normal and 20-epi compounds have differ-
ent contact sites within the LBD, and that a unique location
for the hydroxyl group is therefore not a requirement for
receptor activation. Indeed there is evidence that the ‘double
side chain’ analog (3) interacts differently from either 1 or
2 [11].
1.1. Background and design
Hydrogen bonding interaction of the of both the ring-A
(1␣-OH) and the side chain (25-OH) hydroxyl groups of
1␣,25-dihydroxyvitamin D3 (1) with their respective sites in
the ligand binding domain (LBD) of the vitamin D receptor
protein (VDR) is the paradigm for hormonal activation (by
the ligand-induced conformational change) of the VDR
prior to its initiating the cascade of events leading to regu-
lation of vitamin D related gene expression and protein
synthesis [1]. A feature characteristic to the vitamin D
hormone (as compared to other members of the steroid
hormone superfamily) is the long, conformationally flexible
sterol side chain (the C-17 substituent on ring-D) to which
the liganding atom is attached. The discovery of the in-
creased biologic potency associated with the 20-epimer (2,
MC 1288) of 1␣,25-dihydroxyvitamin D3 or the 20-epimers
The double side chain modification to 1␣,25-dihy-
droxyvitamin D3 was presented independently by the Hoff-
* Tel.: ϩ45-4492-3800; fax: ϩ45-4466-0720.
E-mail address: martin.calverley@leo-pharma.com (M.J. Calverley).
0039-128X/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S0039-128X(00)00156-2

250
M.J. Calverley / Steroids 66 (2001) 249–255
Fig. 1.
mann-La Roche (Nutley, NJ) and the Leo (Denmark)/Polish
Academy of Sciences, Warsaw, groups at the 10th Work-
shop on Vitamin D (Strasbourg, France) [11,12]. This com-
pound (3, Leo code MC 2100) binds to the VDR, is active
in functional assays, and furthermore is a substrate for
vitamin D side chain metabolism enzymes [13]. In order to
investigate the effects of restricting the conformational free-
dom of the side chains, we have formally tied them back
into a ring. The synthesis and properties of the simplest
‘re-converging’ double side chain analogs (4, MC 2108 and
5, MC 2110),1 containing, as a first attempt essentially to
lock the locus of the 25-O, the strain-free cyclohexane
moiety, form the subject of this communication. Of course,
these compounds can also be considered as the 21,24-
methano bridged (Fig. 1) derivatives of 1␣,25-dihydroxyvi-
tamin D3 (and its 20-epimer) and thus represent two con-
formationally locked, ‘snapshot’ versions of side chain
geometries accessible to both the normal and epi parents (1
and 2) as probes for receptor interactions.
the requisite side chain diene with methyl acrylate. Since
neither the cyclo-addition nor the reduction was anticipated
to be feasible on the intact triene system of vitamin D
analog key intermediates [2,3], we elected to perform these
reactions on CD-ring intermediates. Thus, the monosily-
lated diol (6) [14] [prepared from the corresponding diol,
obtained [15] by degradation of vitamin D2, by selective
desilylation (Bu4Nϩ FϪ/THF; room temperature) of the
bis-TBS ether] was converted via the mesylate (7) to the
iodide (8), which (unlike 7) smoothly underwent elimina-
tion of HI by treatment with t-butoxide to the olefin (9)
(Scheme 1). A small amount (5%) of the t-butyl ether
by-product formed in this reaction was easily removed by
chromatography. Essentially quantitative epoxidation with
m-chloroperbenzoic acid to a 2:1 20S:20R2 mixture of the
epoxides (10) followed by rearrangement of these with
magnesium isopropyl-cyclohexylamide (MICA) afforded a
single allylic alcohol (11). Oxidation with the Dess-Martin
periodinane produced the conjugated aldehyde (12), which
on Wittig olefination with methylenetriphenylphosphorane
gave the key intermediate (13).
2. Synthesis and discussion
Diels-Alder reaction of 13 with methyl acrylate (7 molar
equiv.) in the presence of anhydrous aluminum chloride (0.5
molar equiv.) cleanly generated the 1,4-adducts (14a and
14b), epimeric at C-24, with none of the 1,3-regio-isomers
The synthetic strategy involves generation of the cyclo-
hexane ring via hydrogenation of the Diels-Alder adduct of
1 Note that the projection (up or down) of the C-20 hydrogen is
arbitrary in drawing 4 or 5; it is the relative projections of the 20-H and the
24-H (neither C-20 nor C-24 is chiral) that distinguish the compounds. The
carbon atoms of the ‘normal’ and ‘epi’ side chain residues in both double
and reconverging side chains are conveniently numbered for identification
as indicated on structures 3 and 4. For convenience the numbering system
of Fig. 1 is retained in the discussion of synthetic intermediates even when
they lack the full C27 9,10-seco-cholestane framework of vitamin D3.
2 The configuration assignments were made by comparison with the
product of a stereoselective (but less efficient) synthesis of either isomer by
(a) reaction of the 20-ketone obtained by ozonolysis of 9 with dimethyl-
sulphonium methylide (S-isomer) or (b) ␣-bromination of the same 20-
ketone followed by reaction with MeMgBr (R-isomer), invoking in both
cases the well known Felkin-Anh selectivity of kinetic nucleophilic addi-
tion to the 20-ketopregnane system.

M.J. Calverley / Steroids 66 (2001) 249–255
251
Scheme 1.
detectable by NMR spectroscopy (Scheme 2). The 1H-NMR
spectrum of the product showed two sharp singlets (␦ 0.76
and ␦ 0.74 ppm) for the angular methyls (C-18) of 14 in a
9:11 ratio (height), but it has not been possible to assign the
structures at this point. Recrystallization of the mixture only
inverted the ratio, but provided a 1:4 mixture in the mother
liquor. This was however of no consequence, since catalytic
hydrogenation (over palladium) of the cyclohexenes gave a
precisely 1:1 mixture of the 1,4-cis and 1,4-trans disubsti-
tuted cyclohexanes 15cis and 15trans (the only two possible
stereoisomers), irrespective of the ratio of starting materi-
als. The ratio and relative configurations in these two prod-
ucts were readily determined by the integration and shape of
the 24-H signals in the proton-NMR spectrum, in particular
the diagnostic trans-diaxial proton-proton [24-H,23(23Ј)-H]
coupling constants (12.3 Hz) which can only be associated
with the 1,4-diequatorial substitution pattern of 15trans
(Fig. 2). Partial equilibration of the mixture by treatment
with KOBut in methanol under reflux resulted in an in-
creased proportion of the more stable 15trans, thus con-
firming the assignment, but since both isomers of the title
analogs were required for biologic study, the 1:1 mixture (if
separable) was ideal. It transpired that separation of the
series was best deferred to a later stage in the synthesis
(15cis and 15trans ran very closely on TLC), so this mixture
was carried through the sequence 15 3 163 17 3 18,
involving reaction with methyl Grignard, desilylation with
HF, and Dess-Martin oxidation, as shown in Scheme 3. A
slight kinetic resolution of the mixture of esters was ob-
served under the first step of this sequence: A very small
Scheme 2.

252
M.J. Calverley / Steroids 66 (2001) 249–255
col. The relatively low yield (59%) of 21(cis) encountered
in the cis series was partly due to an adventitious excess of
n-butyl-lithium, which resulted in the consumption of some
of the ketone to give a substantial amount of the n-butyl
carbinol as a single by-product. The final step was depro-
tection of the three alcohol groups by desilylation of 21 with
HF to give 21,24-methano-1␣,25-dihydroxyvitamin D3 (4
and 5).3
A useful feature noticed in the 13C-NMR spectra of the
cyclohexyl-containing compounds described in this com-
munication (Scheme 3) is the sensitivity of the chemical
shift (␦C) of C-17 to the conformation of the side chain ring
whilst at the same time being relatively insensitive to the
nature of ring-C substitution (in contradistinction to the
C-14 signal, which shows the opposite dependence). Rele-
vant data are collected in Table 1. It is apparent that all
compounds in the trans series show a C-17 ␦-value corre-
sponding closely, though being slightly higher, to that ob-
served for the 20-normal side chain compounds such as 1.
The value for the 20-epi isomer (2) (and cognate com-
pounds) is on the other hand slightly, but consistently,
lower. The truncated side chain compound 22 [18] has the
highest ␦C value, providing a comparison standard in which
␥-substituents, with their associated magnetic shielding ef-
fect, are not present. Taken alone, these results would not
have attracted attention, but the comparison with the cis
series is dramatic and merits comment. Thus, whereas there
is a rather small deviation from the trans compound C-17 ␦
value (0.7 ppm to high field or 1.7 ppm to low field for the
vitamin D analog series) in the examples discussed so far,
the deviation between cis and trans isomers is 9 ppm (for
the tertiary alcohol derivatives). These data are consistent
Fig. 2. Partial proton-NMR spectrum (300 MHz, CDCl3, Me4Si) of the
mixture of compounds 15. The respective protons on C-24 are indicated in
partial structures showing the preferred chair conformation of 15trans and
the two chair conformations of 15cis.
3MC 2108 (4): M.p. 191–192°C (from methyl formate); UV (EtOH):
amount (about 1%) of ‘unreacted starting material’ was
recovered during the purification process and shown by
NMR to consist of essentially pure cis isomer 15cis. Resub-
mission of this compound to the Grignard reaction under the
original conditions then gave a reference sample of 16cis.
The ketones 18 (providing the optimum difference in TLC
Rf-values) were isolated in isomerically pure form from the
final mixture by chromatography. Samples of 18cis (low
Rf-isomer) and 18trans (high Rf-isomer) were individually
reduced back to reference samples of their respective pre-
cursor alcohols 17 with borohydride as shown in order to
correlate their configurations with 15. Thus, the NMR spec-
tra of 17cis and the previously obtained sample of 16cis
showed the expected correspondence (see Table 1, dis-
cussed below), and desilylation of 16cis did indeed give
17cis.
␭
264 nm (⑀ 17 500); NMR (CDCl3, Me4Si): ␦H (500 MHz) (J in Hz)
max
0.55 (3H, s, 18-H3), 1.18 (6H, s, 26-H3, 27-H3), 2.32 (H, dd, J 6.5 13.4,
4␤-H), 2.60 (H, dd, J 3 13.4; 4␣-H), 2.84 (H, bd, J 12.4, 9␤-H), 4.23 (H,
m, 3-H), 4.43 (H, m, 1-H), 5.01 (H, bs, 19E-H), 5.33 (H, bs, 19Z-H), 6.02
(H, d, J 11.3, 7-H), 6.38 (H, d, J 11.3, 6-H) ppm; ␦C (125.8 MHz) (CDCl3
ϭ 76.8) 11.5 (C-18), 21.5 and 21.7 (C-23, C-23Ј), 21.9 (C-11), 23.3 (C-15),
26.6 (C-27 and C-26), 27.3 (C-16), 28.7 (C-9), 29.2 and 29.8 (C-22,
C-22Ј), 35.3 (C-20), 40.0 (C-12), 42.4 (C-2), 44.8 (C-4), 45.6 (C-13), 47.7
(C-17), 49.0 (C-24), 56.3 (C-14), 66.2 (C-3), 70.3 (C-1), 72.7 (C-25), 111.5
(C-19), 116.9 (C-7), 124.35 (C-6), 133.1 (C-5), 142.5 (C-8), 147.35 (C-10)
ppm; MS: Calcd. for C28H44O3 (Mϩ) 428.3290. Found 410.3320, Calcd.
for C28H42O2 (M ϩ-H2O) 410.3185. Found 410.3193.
MC 2110 (5): M.p. 122–124°C (from methyl formate); UV (EtOH): ␭
max
264 nm (⑀ 17 500); NMR (CDCl3, Me4Si): ␦H 0.54 (3H, s, 18-H3), 0.82–
1.05 (4H, m, 22-H, 22Ј-H, 23-H, 23Ј-H), 1.14 (6H, s, 26-H3, 27-H3), 2.31
(1H, dd, J 6.5 13.4, 4␤-H), 2.56 (1H, dd, J 3.1 13.5, 4␣-H), 2.83 (1H, bd,
J 13.3, 9␤-H), 4.21 (1H, m, 3-H), 4.42 (1H, m, 1-H), 4.99 (1H, bs, 19E-H),
5.32 (1H, bs, 19Z-H), 6.04 (1H, d, J 11.3, 7-H), 6.35 (1H, d, J 11.3, 6-H)
ppm; ␦C (125.8 MHz) (CDCl3 ϭ 76.8) 11.8 (C-18), 21.9 (C-11), 23.3
(C-15), 26.4 (C-26 and C-27), 26.95 and 27.0 (C-23,C-23Ј), 27.1 (C-16),
28.8 (C-9), 32.2 and 32.5 (C-22 and C-22Ј), 40.3 (C-12), 40.8 (C-20), 42.3
(C-2), 44.8 (C-4), 45.6 (C-13), 48.6 (C-24), 56.1 (C-14), 56.7 (C-17), 66.2
(C-3), 70.3 (C-1), 72.6 (C-25), 111.6 (C-19), 116.85 (C-7), 124.4 (C-6),
133.0 (C-5), 142.6 (C-8), 147.3 (C-10); MS: Calcd. for C28H44O3 (Mϩ)
428.3290. Found 428.3289.
The syntheses of the target compounds 4 and 5 were
completed in separate sequences by protection of the tert-
alcohol function in 18 as the trimethylsilyl ether prior to the
Horner-Wittig coupling of 19 with the A-ring building
block (20) [16,17] used in the Hoffmann-La Roche proto-

M.J. Calverley / Steroids 66 (2001) 249–255
253
Scheme 3.
with the C20-C17 bond essentially adopting the axial con-
formation in the cis compounds, C-17 experiencing thereby
two ␥-gauche interactions [19], while the bulky tertiary
alcohol substituent claims the equatorial conformation on
the cyclohexane chair. In the 1,4-diequatorial situation of
the trans series, these gauche interactions are removed, and
Table 1
Selected 13C-NMR data (75.5 or 125.8 MHz, solvent CDCl3) in ppm, relative to solvent signal set to 76.8 ppm
Compound
Side
chain
␦
15,16
␦
Vitamin D
analog
C
C
C-17
C-14
17
18
15cis
(cis-ester)
52.9
52.9
53.1
16cis
17cis
18cis
4
47.8
47.7
48.1
47.7
52.6
62.0
56.3
(All)
(47.8 Ϯ 0.3)
15trans
(trans-ester)
56.9
52.9
52.9
16trans
17trans
18trans
5
57.1
56.9
56.9
56.7
52.4
61.8
56.1
(All)
(56.9 Ϯ 0.2)
1
2
3
normal
epi
double
56.4
56.0
52.8
56.2
56.2
56.1
22
58.4
56.1
Note that for compounds 1, 2, and 3, the corresponding ␦-values are identical in their respective 25-desoxy (25,25Ј-didesoxy) derivatives (formulae not
shown). Compound 22: 23,24,25,26,27-pentanor-1␣-hydroxyvitamin D3.

254
M.J. Calverley / Steroids 66 (2001) 249–255
relative to 1 is thus tentatively explained by a proposing a
higher proportion (albeit still very minor) of the gauche side
chain conformer in solution, reflecting the more congested
epi situation.
Preliminary in vitro biologic tests indicate that the cis
and trans isomers of 21,24-methano-1␣,25-dihydroxyvita-
min D3 (4 and 5) are markedly less potent than the parent
compounds 1, 2, and 3 indicating that an agonistic LBD
interaction of the side chain hydroxyl group is essentially
‘locked out’ in the new analogs,5 rendering them uninter-
esting for further biologic study. On the other hand, we can
conclude that the NMR spectroscopic data for these locked,
extreme, side chain conformations in the cyclohexane ana-
logs are useful in so far as they throw new light on the
solution conformations of flexible, open conformation, ac-
tive ligands, including 1␣,25-dihydroxyvitamin D3 itself.
Fig. 3. Energy minimized (PCMODEL) partial structures of MC 2108 (4)
(upper) and MC 2110 (5), wherein the C-7 appendage of the intact mole-
cules has been replaced by a hydrogen atom in order to facilitate the
computations (cf. Ref. [5]). For clarity only hydrogens (white atoms) on
carbons 17, 20, and 24 are depicted (black atoms: oxygens on C-25).
the chemical shift approaches that observed in 22. Molec-
ular modeling studies nicely confirm these features in the
preferred conformations of the molecules (see Fig. 3).4 In
the case of the ester derivative, 15cis, the lesser steric
requirement of the methoxycarbonyl substituent makes for a
more equal time-averaged distribution of conformers, in
accord with the intermediate coupling constant (ca. 5.5 Hz)
Acknowledgments
The important contribution of Ms. Lae¨titia Boue´rat, who
helped with much of the experimental work while on an
IAESTE student exchange visit during the summer of 1997
is appreciated. I thank Ms. C. Mørk Hansen of the Biochem-
istry Department for the biologic screening, the technical
staff of the Spectroscopic Department (under Mr. N.R.
Andersen) for spectroscopic services, and Mrs. Priscila
Madsen for her continuing excellent technical assistance.
Drs. M.D. Sørensen and L.K.A. Blaehr (Medicinal Chem-
istry Department) are thanked for their guidance with the
computer modeling. Finally the encouragement given by
Dr. F. Bjo¨rkling (Head of Medicinal Chemistry) in allowing
me to report on these studies is gratefully acknowledged.
1
noted in the H-NMR spectrum (Fig. 2), and here the C-17
␦ value is also intermediate (Table 1). In 15trans, the
substituents do not have to vie for the equatorial position,
consistent with the top-of-the-range trans-diaxial value for
the coupling constants 24-H,23(23Ј)-H already mentioned.
As a final observation, it is interesting to note that the
C-17 ␦ value typical for the double side chain analogues
(e.g. 3) is also intermediate between the two extremes
presented in 4 and 5. This suggests that the gauche 20,22-
conformation is significantly populated by one of the two
side chains at any given time in the dynamic situation.
Modeling studies on 3 [or rather on its 25,25Ј-didesoxy-
derivative (NB comment in the Table), a prodrug form of 3,
where intramolecular hydrogen bonding is ruled out] (un-
published) actually point to the global energy minimum
with one side chain extended (in fact the normal side chain)
and the other (epi) gauche (with only a slight energy dif-
ference to the switched situation). Indeed exactly this mixed
conformation scenario is dramatically demonstrated in the
X-ray crystal structure [12,20] of a 23Ј-oxa double side
chain derivative [12,21]. In the same context, it is interest-
ing to recall the molecular modeling studies [5] on 1 and 2
that reveal the gauche conformation of the (single) side
chain in each case as the global minimum. The slightly
lower C-17 ␦-value drawn attention to in Table 1 for 2
References
[1] For the definitive source literature, the reader is referred to the proceed-
ings of the triennial international workshops on Vitamin D: Norman AW,
Bouillon R, Thomasset M, editors. Vitamin D: Chemistry, biology and
clinical applications of the steroid hormone. Proceedings of the 10th
workshop on vitamin D. Riverside (Calif.): University of California,
1997:1–960; and previous volumes in the series.
[2] Calverley MJ, Binderup E, Binderup L. The 20-Epi Modification in
the Vitamin D Series: Selective enhancement of “non-classical” re-
ceptor mediated effects. In: Norman AW, Bouillon R, Thomasset M,
editors. Vitamin D: Gene Regulation, Structure-Function Analysis
and Clinical Application. Berlin: Walter de Gruyter, 1991. pp. 163–4.
[3] Hansen K, Calverley MJ, Binderup L. Synthesis and biological ac-
tivity of 22-oxa vitamin D analogues. In: Norman AW, Bouillon R,
Thomasset M, editors. Vitamin D: Gene Regulation, Structure-Func-
tion Analysis and Clinical Application. Berlin: Walter de Gruyter,
1991. pp. 161–2.
4 PCMODEL: Serena Software, Bloomington, Indiana, USA; Website:
http://serenasoft.com
[4] Binderup L, Latini S, Binderup E, Bretting C, Calverley M, Hansen K.
20-Epi-vitamin D3 analogs: a novel class of potent regulators of cell growth
and immune responses. Biochem Pharmacol 1991;42:1569–75.
[5] Midland MM, Plumet J, Okamura WH. Effect of C20 Stereochem-
istry on the conformational profile of the side chains of vitamin-D
analogs. BioMed Chem Lett 1993;3:1799–804.
5 This negative result is with hindsight perhaps not surprising in the
light of the recently published crystal structure of the LBD with its bound
natural ligand [22], revealing an active side chain conformation of 1
seemingly not readily accessible to 4 or 5.

M.J. Calverley / Steroids 66 (2001) 249–255
255
[6] Wilson SR, Zhao H, Dewan J. Chemistry of the new immunosup-
pressant KH-1060: new synthons and x-ray evidence for unique
side-chain geometry. BioMed Chem Lett 1993;3:341–4.
EG, Hennessy BM, Uskokovic MR. Stereoselective total synthesis
of 1␣,25S,26-trihydroxycholecalciferol. Tetrahedron 1984;40:
2283–96.
[7] Yamada S, Yamamoto K, Masuno H, Ohta M. Conformation-func-
tion relationship of vitamin D: Conformational analysis predicts po-
tential side-chain structure. J Med Chem 1998;41:1467–75.
[8] Yamamoto K, Ooizumi H, Umesono K, Verstuyf A, Bouillon R,
DeLuca HF, Shinki T, Suda T, Yamada S. Three-dimensional struc-
ture-function relationship of vitamin D: side chain location and var-
ious activities. Bioorg Med Chem Lett 1999;9:1041–6.
[15] Sardina FJ, Mourin˜o A, Castedo L. Studies on the synthesis of
side-chain hydroxylated metabolites of vitamin D. 2. Stereocontrolled
synthesis of 25-hydroxyvitamin D2. J Org Chem 1986;51:1264–9.
[16] Baggiolini EG, Iacobelli JA, Hennessy BM, Batcho AD, Sereno JF,
Uskokovic MR. Stereocontrolled total synthesis of 1␣,25-dihydroxy-
cholecalciferol and 1␣,25-dihydroxy-ergocalciferol. J Org Chem
1986;51:3098–108.
[9] Liu YY, Collins ED, Norman AW, Peleg S. Differential interaction of
1␣,25-dihydroxyvitamin D3 analogues and their 20-epi homologues
(sic) with the vitamin D receptor. J Biol Chem 1997;272:3336–45.
[10] Peleg S, Sastry M, Collins ED, Bishop JE, Norman AW. Distinct
conformational changes induced by 20-epi analogues of 1␣,25-dihy-
droxyvitamin D3 are associated with enhanced activation of the vi-
tamin D receptor. J Biol Chem 1995;270:10551–8.
[11] Uskokovic MR, Manchand PS, Peleg S, Norman AW. Synthesis and
preliminary evaluation of the biological properties of a 1␣,25-dihy-
droxyvitamin D3 analog with two side chains. In: Norman AW,
Bouillon R, Thomasset M, editors. Vitamin D. Riverside (Calif.):
University of California, 1997. pp. 19–21.
[12] Kurek-Tyrlik A, Makaev FZ, Wicha J, Zhabinskii V, Calverley MJ.
Pivoting 20-normal and 20-epi calcitriols: Synthesis and crystal struc-
ture of a “double side chain” analogue. In: Norman AW, Bouillon R,
Thomasset M, editors. Vitamin D. Riverside (Calif.): University of
California, 1997. pp. 30–1.
[13] Moore B, Makin HLJ, Schroeder NJ, Calverley MJ, Jones G. Metab-
olism of double side-chain calcitriol analogs by a human keratinocyte
cell line. J Bone Miner Res 1998;23(suppl 5):S451.
[17] Kutner A, Zhao H, Fitak H, Wilson SR. Synthesis of retiferol RAD1
and RAD2, the lead representatives of a new class of des-CD analogs
of cholecalciferol. Bioorg Chem 1995;23:22–32.
[18] DeLuca HF, Schnoes HK, Lau WF, inventors. Seco-sterol compounds. PCT
Int Appl WO 86 06,255. Chem Abs 1986;106:131710 m.
[19] Dalling DK, Grant DM. Carbon-13 magnetic resonance. XXI. Steric
interactions in the methylcyclohexanes. J Am Chem Soc 1972;94:
5318–24.
[20] Galdecka E, Galdecki Z, Gorkiewicz Z, Kurek-Tyrlik A, Makaev FZ,
Wicha J, Calverley MJ. Crystal and molecular structure of 3␤-ace-
toxy-21-(2-hydroxy-2-methylpropoxy)-cholest-5-en-25-ol, the first
“double side chain” cholesterol analog investigated by X-ray diffrac-
tion. Pol J Chem 1999;73:547–55.
[21] Kurek-Tyrlik A, Makaev FZ, Wicha J, Calverley MJ. Synthesis of
21-(3Ј-methylbutyl)cholest-5-ene-3␤,3Ј,25-triol and 21-(2Ј-methyl-
propoxy)cholest-5-ene-2Ј,3␤,25-triol, the first “double side chain”
cholesterol analogs. Liebigs Ann Chem 1997;1997–2001.
[22] Rochel N, Wurtz JM, Mitschler A, Klaholz B, Moras D. The crystal
structure of the nuclear receptor for vitamin D bound to its natural
ligand. Mol Cell 2000;5:173–9.
[14] Wovkulich PM, Barcelos F, Batcho AD, Sereno JF, Baggiolini