Diastereotopic and deuterium effects in gemini

Article abstract of DOI:10.1021/jm400032t

Changing the geminal methyl groups on 1α,25-dihydroxyvitamin D ₃ and its analogues to the deuterio versions generally improves the bioactivity. Derivatives of 1α,25-dihydroxyvitamin D₃ with two chains emanating at C20, commonly referred to as gemini, are subject to the same phenomenon. Additionally, gemini with different side chains are susceptible to bioactivity differentials where the C17-C20 threo configuration usually imparts higher activity than the corresponding erythro arrangement. In an effort to analyze the deuterium effect on gemini with minimal diastereotopic distortion, we synthesized gemini with equal side chains but introduced deuterium diastereospecifically on either chain. We solved the crystal structures of these compounds in the zebra fish zVDR ligand binding domain as complexes with NCoA-2 coactivator peptide and correlated the findings with growth inhibition in a breast cancer cell line.

Full text of DOI:10.1021/jm400032t

Article
pubs.acs.org/jmc
Diastereotopic and Deuterium Effects in Gemini
Hubert Maehr,*,† Natacha Rochel,‡ Hong Jin Lee,†,∥ Nanjoo Suh,†,§ and Milan R. Uskokovic†
†Department of Chemical Biology, Ernest Mario School of Pharmacy, Rutgers, The State University of New Jersey, 164 Frelinghuysen
Road, Piscataway, New Jersey 08854, United States
‡Dep
́
artement de Biologie et de Gen
́
omique Structurales, Institut de Gen
́
et
́
ique et de Biologie Molec
́
ulaire et Cellulaire (IGBMC),
de Strasbourg, 1 rue
Centre National de la Recherche Scientifique, Institut National de la Sante
Laurent Fries, 67404 Illkirch, France
§The Cancer Institute of New Jersey, 195 Little Albany Street, New Brunswick, New Jersey 08903, United States
́
de la Recherche Med
́
icale, Universite
́
S
* Supporting Information
ABSTRACT: Changing the geminal methyl groups on 1α,25-
dihydroxyvitamin D3 and its analogues to the deuterio versions
generally improves the bioactivity. Derivatives of 1α,25-dihydrox-
yvitamin D3 with two chains emanating at C20, commonly
referred to as gemini, are subject to the same phenomenon.
Additionally, gemini with different side chains are susceptible to
bioactivity differentials where the C17−C20 threo configuration
usually imparts higher activity than the corresponding erythro
arrangement. In an effort to analyze the deuterium effect on
gemini with minimal diastereotopic distortion, we synthesized
gemini with equal side chains but introduced deuterium
diastereospecifically on either chain. We solved the crystal
structures of these compounds in the zebra fish zVDR ligand
binding domain as complexes with NCoA-2 coactivator peptide and correlated the findings with growth inhibition in a breast
cancer cell line.
complex of the vitamin D receptor (VDR), the retinoid X
receptor (RXR), and a 1α,25-(OH)2D3 response element
(VDRE). It functions through the LBD of the VDR that is
ligand-stabilized in either agonistic, antagonistic, or non-
agonistic conformation,4 all depending on ligand constitution
and configuration. Similarly, DRIP (vitamin D receptor
interacting protein) mediated interactions have been invoked
as ligand-dependent phenomena and potentially creating
aberrant biological responses.5,6 Extending C21 of 1α,25-
(OH)2D3 and its derivatives to a second side chain as in 3, 4,
and 5 gave rise to a new class of derivatives, known as gemini,
that significantly augment the biological spectrum of 1α,25-
(OH)2D3 on various fronts.7−12 These compounds double the
opportunity for chain modifications and offer exploitation of
newly generated stereochemical consequences.
With a focus on metabolic stability, we studied, some 30
years ago, the stimulation of cartilage growth in vitro by 1 in
two growing-cartilage models, the avian pelvic cartilage, and the
mammalian scapular growth-plate cartilage, known to be
susceptive to vitamin D regulation.13 In that study, we
discovered that 1, whose geminal methyl groups in the side
chain were replaced with deuteriomethyl moieties, exhibited
superior activity in the growth-plate cartilage model. This
INTRODUCTION
■
Modification of the chemical structure of drugs to prevent or
retard metabolic degradation is an important activity in drug
optimization and requires an alteration of the sites of metabolic
activity so that it is no longer a suitable substrate for the
enzyme that initiates the metabolism. Addition of substituents
and protective groups and changes to groups of different size
and polarity are commonly employed for this purpose.1
Although isotopology is an extremely useful tool in metabolism
studies, its application as a modifier of drug metabolism is
restricted to deuterium2,3 which has the renown of rendering
the intrinsic drug properties invariant.
In the vitamin D arena, molecular transmutations of 1α,25-
dihydroxyvitamin D3 (1, 1α,25-(OH)2D3, or calcitriol) have
been pursued not only with the aim of metabolic stabilization
but also to modify the biological properties that reach beyond
calcium and phosphorus metabolism and act as agonists in
transcriptional activities involving carcinogenesis, immune
disorders, and skin diseases. Tailored modifications, intended
to reach specifically but one of those goals, have proved to be
nontrivial as small chemical changes in the ligand enable
conformational changes in the VDR ligand-binding domain
(LBD) with notoriously far-reaching consequences by affecting
the central role of the hormone in calcium and bone
metabolism and the molecular switch mechanism of nuclear
1α,25-(OH)2D3 signaling. This molecular switch comprises the
Received: January 7, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
which is not only a function of deuterioisotopology alone but
further obfuscated by diastereotopicity in molecules with two
unequal chains.
To shed some light onto these stereochemical and deuterium
effects of 1α,25-(OH)2D3 and its congeners, it is appropriate to
review briefly the observed biological consequences resulting
from diastereotopic and isotopological changes. The examples
that permit such comparisons are few but are indicative of
certain trends. Current emphasis is placed exclusively on the
chirality at C20 and the effect of deuteriomethyl groups at the
chain termini corresponding to C26 and C27 in 1.
Commencing with the C20 chirality issue, it is known that
epi-1α,25-(OH)2D3 (2) shows exceptionally high affinity to the
vitamin D receptor due to altered protein binding;25 it is a more
potent inducer of cell differentiation than 1, it is transcription-
ally more active by some 3 powers of magnitude, it is
metabolized 36 times slower than 1, it showed decreased DBP
(rat plasma) binding affinity but is bound to the bovine thymus
VDR with 5 times greater affinity than 1,26 it is a more powerful
regulator of cell cycle progression, and it was concluded that 2
protected the VDR against catabolism more efficiently than 1
by preventing, or retarding, the binding of factors mediating
VDR degradation.27 In a comparison of 1 and 2, one can now
aver that 2 is more active due to its 20S configuration. The
recognition of stereochemical details in crystallographic images
is facilitated by viewing the stereocenter C20 as a unit with the
contiguous and configurationally defined C17. Thus rephrasing,
the compound where the C17−C20 bond is of the threo
configuration is more active.
“deuterium effect” in biological systems has since been verified
by us in a number of 1α,25-(OH)2D3 derivatives and is now
being exploited by commercial enterprises that dwell
specifically on the deuteration of existing drugs in search of
novelty and modified biological properties and activities.14
Subsequent to our original observation, we also synthesized a
number of 1α,25-(OH)2D3 derivatives in the gemini series,
wherein geminal methyl groups were transmuted to perdeuterio
equivalents.15−17 In all instances, these derivatives were
biologically more active than the protio parents. The reason
for this elevation of activity is not entirely clear. Because side
chain degradation of 1α,25-(OH)2D3 is recognized as the initial
step toward drug inactivation and excretion, and positions 23,
24, and 26 are known sites of oxidative degradation, one could
indeed infer, a priori, magnified activity of the deuterio versions
as the direct result of increased metabolic stability. The
metabolic events, however, point to calcitroic acid as the major
excretory form of the hormone, which is produced by the
inducible, mitochondrial cytochrome P450 (CYP24) by initial
C24 hydroxylation and oxidation, followed by hydroxylation at
C23 and cleavage of the C23−C24 bond.18,19 This mechanism
also appears to hold for gemini and its derivatives, as it has been
shown that gemini suffers diasterospecific C24 hydroxylation as
the first step in the metabolic process.20 Although the
deuterium−carbon bond is some 10 times stronger than the
hydrogen−carbon bond, the argument of bond strength in
support of metabolic stability enhancement as primary isotope
effect is untenable because the deuterio modification occurs at
the geminal methyl groups and cleavage of a C−D bond at
either C26 or C27 cannot be implicated in the rate-determining
step in the major metabolic pathway. Formation and further
metabolism of a lactone, corresponding to 1α,25-dihydrox-
yvitamin D3-26,23-lactone21 and initiated by C23 and C26
oxidations, is also an unsatisfactory explanation of the
deuterium effect because the expected oxidation intermedi-
ates22 remain undetected.20 A similar argument can be made
against other minor pathways23 including a conceivable
metabolism via initial C20- and C22-hydroxylations.24
It has been shown that 1α,25-(OH)2D3 and its derivatives
assume very similar positions of their A-seco-B, C, and D-rings
in the VDR LBD where locus and configuration are enforced by
well-defined sets of interactions.28−31 The ease of side chain
accommodation of either 1 or 2 in the same channel apparently
renders irrelevant the rotational barrier that the required
change of the torsion angle at C17−C20 entails. Figure1 shows
Figure 1.
simulated views of 1 and 2 from the proximal C17 to the distal
C20 with synclinal and antiperiplanar dihedral angles of the
hydrogen atoms, respectively, suggesting lower ground state
energy of the threo configuration. These simplified representa-
tion convey the similarity with the corresponding arrangements
discernible in the crystal structures of VDR LBD complexes of
compounds derived from either 1 or 2.32 Although the same
channel is occupied by 1 or 2, the chains line opposite sides in
the pocket and consequently engage in different protein
interactions.
Following the discovery that the parental gemini 5 exhibited
enhanced antiproliferative activities compared to 1,7 it was
found that the structures of zVDR-5 and zVDR-1 were similar
but the accommodation of the second side chain in 5 induced a
backbone shift and side chain reorientation. This ligand-
dictated structural rearrangement of the protein core preserved
the original space used by the 1α,25-(OH)2D3 chain in 1 and 2
In the absence of metabolic involvement, the site of isotopic
changes exclusively at the geminal methyl groups at the
crowded termini should, per expectations, only infinitesimally
affect the physicochemical properties and be insufficient to
induce conformational changes in the LBD. In this account, we
intend to illuminate the “biological deuterium effect” in gemini,
B
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(referred to as the parental channel). This core is also able to
accommodate one of the gemini chains, but in addition, the
change of the protein core made possible the creation of a new
pocket (referred to as the “secondary channel”) to accept the
second chain.32
Given the choice between the aliphatic or alkyne chains
present in 3 and 4, it was shown that it is the larger alkyne side
chain that prefers the parental channel. The smaller and more
flexible aliphatic chain pushed for accommodation by creating
the new pocket resulting in superagonist activity due to
increased stabilization of helix H12.31,33 Considering that 2 is
more potent in a biological sense than 1, and considering only
the size of the side chains, one can surmise that the threo pair
3a/3b should elicit higher bioactivity than the erythro pair 4a/
4b. The gemini pairs 3a/4a and 3b/4b are thus examined with
respect to C20 epimerism and the pairs 3a/3b and 4a/4b with
respect to the deuterium effect.
Probing first into this “threo effect,” we observed varying
degrees of differences in biological responses among the
available compounds. The threo compound 3a was three times
more calcemic than either of the erythro compounds 1 or 4a
but some 500 times more potent in interferon-γ release. In a
tumor-growth model in mice, both 1 and 4a showed no
significant effect at 0.02 μg, whereas 3a exhibited a dose-
dependent response with tumor volume reduction of 41% at
the same dose. In a study of estrogen receptor (ER)-positive
carcinomas in rats, however, 3b and 4b were equally active in
the inhibition of NMU-induced breast cancer carcinogenesis in
disregard of the threo effect.15,17
Investigating the deuterium effect, we found 3a and 3b in the
tumor-growth model in mice with dose-dependent differences,
but the deuterio compound 3b was 10 times as active as the
protio version 3a. These observations raised new questions. In
the absence of a diastereotopic bias, would a change to the
geminal trideuteriomethyl groups in gemini alone suffice to
change the biological properties of the ligand; i.e., would there
be a difference between the parental gemini 5 and its
dodecadeuterio version 6? And if it does, could the ligand
properties be further refined through the reintroduction of a
diastereochemical bias in 6 by replacing just one set of the
geminal deuteriomethyl groups with the corresponding protio
versions? If the deuterium effect would still be observed, would
it then matter which of the side chains carries the
deuteriomethyl groups? In other words, would the threo-effect
still be palpable in molecules whose chemical constitutions
differ so little from each other?
diastereomeric mixture, and a subsequent treatment with
iodine/triphenylphosphorane/imidazole in DCM afforded the
diastereomeric iodo compound 12 using protocols previously
established for the preparation of 18 and 23.20 The second
chain was constructed from 12 by the Sustmann addition
process to ethyl acrylate.37 The resulting diastereomeric
mixture of ethyl esters 13 was treated with deuteriomethyl-
magnesium iodide in ether to afford the diol 14a as a single
stereoisomer. Cleavage of the t-butyldimethylsilyloxy group
with fluorosilicic acid38 gave the triol 14b, which was oxidized
with pyridinium dichromate in DCM to lead to the ketone 15a.
This ketone was converted to the bis-trimethylsilyl ether 15b
because it was found that it gave better yields in the subsequent
coupling reaction than the diol 15a. This additional operation is
insignificant because the required deprotection proceeds
concurrently with the compulsory desilylation of the two
protective groups in the A-ring after the coupling reaction.
On the basis of the successful construction of the coupling
partner 15b, we prepared the diastereomeric ketones 21b and
26b analogously, as illustrated in Scheme 1. Commencing with
the iodo-pairs 18 and 23 derived from the epimeric alcohols 17
and 22 as described previously, we repeated the chain extension
leading to the esters 19 and 24. Subsequent alkylations with
deuteriomethylmagnesium iodide gave 20a and 25a, depro-
tections with fluorosilicic acid advanced the intermediates to
the triols 20b and 25b, oxidations with pyridinium dichromate
gave ketones 21a and 26a, and reprotections with trimethylsilyl
imidazole furnished the coupling partners 21b and 26b.
The ketones 21b and 26b were coupled each with 30a,
prepared from carvone,39,40 and 15b, 21b, and 26b were
coupled each with 30b, prepared by total synthesis.41 These
coupling reactions were performed according to the established
protocols propagated by Lythgoe42 and illustrated in Scheme 2.
The resulting tetrasilylethers 27 and the pairs 28a/28b and
29a/29b were desilylated with tetrabutylammonium fluoride in
THF during disparate lengths of time to furnish the targets 6
and the pairs 7a/7b and 8a/8b, respectively.
RESULTS AND DISCUSSION
■
To provide the tools for the answers to these intriguing
questions, we synthesized dodecadeuteriogemini 6, the “threo”
hexadeuteriogemini 7a, and the 19-nor analogue 7b, as well as
the corresponding erythro pair 8a and 8b as described
hereunder. The synthetic concept used for the generation of
6 differs from the published methods employed for the protio
analogue 5b.34,35 According to those procedures, both side
chains were introduced concurrently; the construction of 7 and
8, however, required a stepwise process. Thus, we chose 6 as a
test candidate for the consecutive side chain assembly starting
with 10, which was then paradigmatic for the preparation of 7
and 8. The synthetic steps are illustrated in Scheme 1.
With these compounds in hand, we investigated the
orientations of the two epimeric compounds 7 and 8 within
the LBD of the zVDR. If the LBD of the zVDR is given the
choice between these C20 epimers, and if the zVDR is
discriminating toward their chain constitutions, we should
observe two distinct epimers in zVDR-7 and zVDR-8
The syntheses commenced with the previously prepared
alkenes 1016 and 1636 derived from methyl ester 9. Hydro-
boration of 10, with oxidative workup, gave 11 as a
C
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1
corresponding to the C17,20-threo and C17,20-erythro
configurations depicted schematically in Figure 2, wherein C
represents the side chain with geminal methyl groups as in 1
and D symbolizes the chain with deuteriomethyl groups
corresponding to C21 in 1α,25-(OH)2D3. The “natural” chain
of 1α,25-(OH)2D3 is allowed arbitrary occupancy in the
parental channel. The view was changed from that in Figure
1, with C20 now as the proximal carbon, to facilitate a
comparison with the crystal structure in Figure 3C discussed
below. The illustration allows a perception of the conforma-
tional task which the side chains would have to undergo to
secure accommodation in the same channel. Conversely, if the
zVDR is promiscuous toward accommodating these side chain
constitutions, the diastereotopicity of C20 with respect to the
VDR vanishes and only one of the two configurations portrayed
in Figure 2 would be discernible as the two chains are now
indistinguishable.
We accomplished the crystal structure elucidations of the
zebrafish zVDR LBD in complexes with 7a, 7b, 8a, and 8b,
together with the NCoA-2 coactivator peptide. Similar to 5 or
its derivatives 3a/b and 4a/b, the ligands 7a, 7b, 8a, and 8b
induce, upon binding, the reorientation of the side chain of a
Leu residue in helix 7 [Leu337] that allows the accommodation
of the second side chain of such ligands (Figure 3A).28,33 Both
ligands make similar interactions with the protein and the
hydroxyl groups of the analogues form similar hydrogen bonds
as zVDR LBD bound to the parental gemini. For the 19-nor
analogues, 7b and 8b, interactions with Leu261 (Helix3),
Ser265 (Helix3), and Ile299 (Helix5) are lost compared to the
natural A-ring compounds 7a and 8a (Figure 3B). These lost
D
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 2
between hydrogen and deuterium atoms. Most significantly, the
medium resolutions of the crystal structures reveal no
significant differences in the interactions, except for the one
with Ala331 which is the result of the A-ring modification. An
isotope-induced change in molecular conformation in one of
the ligands, either the “erythro ligand” or the threo equivalent,
to force the same specific side chain into the same specific
channel of the LBD did not materialize, as evidenced by Figure
3C. Both ligands are nearly superimposable, i.e., only one
conformation for both diastereotops is perceptible, which is
diagrammatically shown in Figure 2 as the threo isomer with
the positive dihedral angle between the hydrogen atoms. This
arrangement is consistent with the crystal structure shown in
Figure 3C.
As we have now shown that the zVDR does not distinguish
between the closely related deuterio and protio side chains,
compounds 7 and 8 maintain equal populations of the deuterio
chains in either parental and secondary channel; i.e., the LBD is
indifferent toward the diastereotopic relationship between 7
and 8, rendering them not only equally active, but, thanks to
the presence of one set of deuteriomethyl groups, distinctly
more potent than 5. Following this trend, one can predict that a
compound with two sets of deuteriomethyl groups should be
even more active than either 7b and 8b as both channels are
Figure 2.
interactions, notably with Ser265 that forms an anchoring H-
bond interaction with the ligands, may explain their lower
biological potency in the cell growth inhibition study presented
below.
Looking now to the interactions of the side chains of the
threo pair 7a/b and the erythro pair 8a/8b (Figure 3C), we
find similar contacts with the protein for both pairs. The only
observed difference within the threo pair 7a and 7b is a
stronger interaction of C26 of 7a with Ala331 at 3.7 Å
compared to 4.0 Å for 7b, again, in agreement with the
difference in cell growth inhibition.
Because the configuration at C20, and hence the spacial
arrangements of the protio- and deuterio-substituted side
chains, are firmly defined by synthesis, it is inconsequential that
the electron density maps of 7 and 8 do not allow a distinction
Figure 3. Crystal structures of zVDR LBD complexes. (A) Superposition of 7a (green) with parental gemini (yellow) and 1α,25-(OH)2D3 (pink).
Leu337 in zVDR complexes with gemini analogues adopts another conformation than in zVDR/1α,25-(OH)2D3 upon binding of the second side
chain. (B) The 19-nor compound 7b (pink) lost interactions with Leu261, Ser265, and Ile299 compared to the normal A-ring compound 7a (green).
Dark dashed lines indicate van der Waals interactions, and the red dashed line corresponds to hydrogen-bond interaction. (C) Superposition of 7a,
threo analogue (in green) with the erythro one 8a (in cyan).
E
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
nitrogen flux after cryoprotection with oil (Hampton research). Data
collection from a single frozen crystal was performed at 100 K at ESRF
(Grenoble, France), beamline ID-14-1. The crystals are isomorphous
and belong to the space group P6522, with one LBD complex per
asymmetric unit. The four structures were solved by molecular
replacement and refined to 2.95, 2.44, 2.70, and 2.70 Å for zVDR/7a,
zVDR/7b, zVDR/8a, and zVDR/8b, respectively. Diffraction data
collection and structure refinement were obtained as described.33
Cell Growth Inhibition Studies. MCF10A1a human estrogen
receptor (ER)-negative breast cancer cells were cultured in DMEM/F-
12 medium with 5% heat-inactivated horse serum (Hyclone, Logan,
UT) at 37 °C with 5% CO2. The MCF10CA1a cells were passaged 2
times a week to maintain log phase growth. For testing growth
inhibitory effects of gemini, MCF10CA1a cells were distributed in 24-
well plates (5000 cells/well) and treated with gemini for 3 days. One
μCi of [3H] thymidine was added 3 h before the harvest. The cells
were analyzed on a Beckman LS 6500 liquid scintillation counter.
Chemistry. Chemical reactions were monitored by TLC using
silica gel 60 F254 plates (Merck), typically 5 cm high, and visualized
with UV light or by treatment with phosphomolybdic acid (PMA).
Solvents were of commercial sources but THF was freshly distilled
from sodium benzophenone ketyl. The pH 7 phosphate buffer used
occasionally during work up comprised 139.4 g of dipotassium
phosphate in 400 mL of water plus 100 mL of 2 M phosphoric acid.
Column chromatography was carried out on silica gel Merck 230−400
mesh ASTM, and fractions were pooled according to TLC profiles.
Solutions were dried with sodium sulfate unless stated otherwise, and
chemical operations were conducted under nitrogen or argon.
Products were typically dried at ambient temperature at 1 Torr or
below. Organolithium solutions were purchased from Aldrich and used
at nominal molarities. Mass spectra were recorded on Waters ZQ (low
resolution) and Bruker APEX II (high resolution) spectrometers.
NMR spectra were obtained on a Varian INOVA 400 spectrometer at
occupied exclusively by deuterated chains. This prediction is
upheld in the breast cancer MCF10CA1 growth inhibition
study shown in Figure 4, where the dodcecadeuterio gemini 6
Figure 4.
appears to be more active than 7b and 8b, all of which
containing the 19-nor A-ring. The activity differentials are
highlighted by the IC50 values. Ligands 7b and 8b are nearly
equipotent and so is the pair 7a/8a which, in turn, is more
potent by virtue of the natural A-ring as predicted by the
binding anchors between Leu261, Ser265, and Ile299, which
are absent in 7b, as evident in Figure 3B.
CONCLUSIONS
■
The bioactivities of all vitamin D analogues tested so far benefit
from a replacement of the geminal methyl groups with deuterio
equivalents at the side chains. Gemini with significantly
different side chains usually exhibit the “threo effect”, i.e., the
20S epimers tend to be more active than the R-counterparts.
This bias in biological activity appears to parallel the ability of
the LBD to recognize C20 topographically. In the borderline
scenario where the side chains differ only by deuteration, the
topographic differential is minimal and no longer recognized by
the LBD thus rendering the members of the pairs 7a/8a and
7b/8b equipotent.
1
25 C. The chemical shifts for H spectra are referenced to CHCl3 at
7.26 ppm and DMSO-d5 at 2.50 ppm. For 13C NMR spectra, the
chemical shifts are referenced to CDCl3 at 77.0 ppm and DMSO-d6 at
39.5 ppm.
Ethyl 5-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]oxy-7a-
methyl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-10,10,10-trideu-
terio-9-hydroxy-9-(trideuteriomethyl)decanoate (13). To a
vigorously stirred mixture comprising Zn powder (0.297 mg, 4.54
mmol), pyridine (5.6 mL), and ethyl acrylate (0.455 g, 4.54 mmol)
was added nickel chloride hexahydrate (0.2682 g, 1.128 mmol). The
gray suspension was warmed to 50 °C and after 10 min to 65 °C and
stirred at that temperature for 30 min. During this time, the mixture
quickly assumed a greenish and then a brown−red color. This mixture
was then cooled in an ice bath, and a solution of 12 (0.517 g, 0.978
mmol) in pyridine (3.5 mL) was added. The ice bath was removed,
and stirring continued at ambient temperature. Reaction progress was
monitored by TLC (1:3 EtOAc−hexane), which showed the gradual
conversion of 12 (Rf 0.70) to the ethyl ester 13. The reaction was ca.
90% complete after 30 min, and after 3 h 12 was no longer detectable.
The mixture was then poured into EtOAc (50 mL); the resulting
suspension was filtered through Celite, the filter cake was washed with
EtOAc (40 mL), filtrate and washings were equilibrated with 1 M HCl
(2 × 50 mL) and 0.2 M phosphoric acid (40 mL) followed by brine (5
mL), an aqueous solution (25 mL) containing EDTA disodium salt
dihydrate (0.60 g) and sodium hydrogen sulfate (0.60 g), and then
with brine (5 mL). The solution was dried (MgSO4) and evaporated.
TLC of the residue showed two spots with very similar mobilities,
commensurate with the diastereomeric nature of ester 13. This
material was further purified by flash chromatography, without intent
to separate the mixture, by using a stepwise gradient from 1:19, 1:9,
and 1:6 (EtOAc−hexane). The appropriate fractions were pooled and
evaporated, and the resulting 13 (0.404 g, 82%) was used directly in
Binding affinities between biomolecules can be subject to
deuterium isotope effects on noncovalent interactions. More
specifically, isotopic ligand modifications can affect van der
Waals interactions with proteins and, in some cases, depending
on the position of deuterium, stabilize or destabilize the binding
of the ligand.42 We are consonant with the hypothesis that the
higher bioactivities of the “deuterio ligands” are the result of
secondary kinetic deuterium effects eliciting a certain
stabilization of the LBD of the VDR with enhanced coactivator
interactions.
The findings presented herein provide additional guidelines
for drug research, especially for future phenotypic drug
discovery efforts in the area related to vitamin D.
EXPERIMENTAL SECTION
■
Crystallography. Purification and crystallization. The zebrafish
VDR LBD (residues 156−453) was produced and purified as
described.29,33 The protein was concentrated using Amicon ultra-30
(Millipore) to 3−7 mg/mL and incubated with a 2-fold excess of
ligand and a 3-fold excess of the coactivator NCoA-2 peptide (686-
KHKILHRLLQDSS-698). Crystals were obtained in 50 mM Bis-Tris
pH 6.5, 1.6 M lithium sulfate and 50 mM magnesium sulfate.
1
the next step. TLC 1:3 EtOAc−hexane Rf 0.52. H NMR (CDCl3) δ
−0.01 (s 3 H), 0.00 (s, 3 H), 0.88 (s, 9 H), 0.89 (s, 3 H), 1.06−1.93
(m, 24 H), 1.25 (t, J = 7.2 Hz, 3 H), 2.25 (t, J = 7.4 Hz, 2 H), 3.99 (br
m, 1 H), 4.12 (q, J = 7.2 Hz, 2 H). LR-MS-ES(+) m/z 566 (86, M +
X-Ray Data Collection and Structure Determination. Protein
crystals were mounted in a fiber loop and flash-cooled under a
F
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Na + CH3CN), 525 (32, M + Na), 485 (38, M + H − H2O), 353
(100, 485-TBSOH).
15b as an oily substance (0.22 g, 85% from triol 14b). TLC 1:19
1
MeOH−DCM, Rf 0.43. H NMR (CDCl3) δ 0.10 (s, 9 H), 0.11 (s, 9
6-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]oxy-7a-meth-
yl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-1,1,1-trideuterio-10-
methyl-2-(trideuteriomethyl)undecane-2,10-diol (14a). The
epimeric 13 (0.404 g, 0.804 mmol) was dissolved in ether (6 mL),
the solution was cooled in an ice-bath, and a 1 M solution of
deuteriomethylmagnesium iodide in ether (4 mL) was added via
syringe at ≤5 °C. The ice bath was then removed, and stirring
continued at room temperature. TLC (1:3 EtOAc−hexane) showed a
gradual conversion of the ester (Rf 0.56) to the diol 14a, which was
complete after 3 h. The solution was cooled in an ice bath, and a satd
ammonium chloride solution (10 mL) was added dropwise, followed
by water (5 mL) to dissolve the white precipitate. The resulting
aqueous layer was re-extracted with EtOAc (30 mL), and the
combined organic layers were washed with brine (10 mL containing 3
drops of 1 M HCl) then with brine (8 mL), dried, and evaporated to a
colorless oil (0.4 g), which was flash-chromatographed using a
stepwise gradient of 1:9, 1:6, 1:4, 1:3, and 1:2 EtOAc−hexane.
Fractions were pooled and evaporated then redissolved in pentane,
filtered, and evaporated to afford 14a as a colorless, amorphous
powder (0.31 g, 78%). TLC 1:3 EtOAc−hexane Rf 0.16, 1:1 EtOAc−
hexane Rf 0.64. 1H NMR (CDCl3) δ −0.01 (s, 3H), 0.00 (s, 3H), 0.88
(s, 9H), 0.90 (s, 3H), 1.06−1.93 (m, 27H), 3.99 (br m, 1H). LR-MS-
ES(+) m/z 989 (64, 2M + 1), 536 (18, M + CH3CN + H), 495 (31, M
+ H), 477 (33, 495-H2O), 459 (50, 477-H2O), 327 (100, 459-
TBSOH).
H), 0.64 (s, 3 H), 1.10−2.11 (m, 22 H), 2.16−2.34 (m, 2 H), 2.46 (dd,
J = 7.4 and 11.4 Hz, 1H). LR-MS-ES(+) m/z 546 (18, M + Na), 343
(100, M − 2TMSOH).
1α,25-Dihydroxy-21-(3-hydroxy-3-trideuteriomethyl-4,4,4-
trideuteriobut-1-yl)-26,27-hexadeuterio-19-nor-cholecalcifer-
ol (6). A magnetically stirred 2-neck round-bottom flask equipped with
thermometer and Claisen adapter containing a nitrogen sweep and
rubber septum was charged with diphenylphosphine oxide 30b
(0.4209 g, 0.737 mmol) previously dried under high vacuum for 2
h. To this material was added THF (3.5 mL). The solution was stirred
and cooled to −70 °C, and butyl lithium in hexane (1.6 M, 0.46 mL,
0.736 mmol) was added dropwise within 10 min. The deep cherry-red
solution was stirred at −70 °C for 5 min, and then a solution of the
ketone 15b (0.2267 g, 0.433 mmol), dissolved in THF (2.5 mL), was
added dropwise below −65 °C. The flask containing the ketone was
rinsed with THF (1 mL), which was also added to the reactor. In
process control (TLC, EtOAc−hexane 1:39 and 1:19) revealed only
traces of starting material after 3 h. After an additional 45 min, the
mixture was allowed to warm to −30 °C and pH 7 phosphate buffer (5
mL) was added dropwise. The mixture was stirred vigorously for 25
min. At that time, all ice had melted and the solution was transferred
to a separatory funnel with the aid of hexane (40 mL). The aqueous
phase was re-extracted with hexane (20 mL), and the combined
hexane layers were washed with brine (5 mL), dried, and evaporated to
give crude 27 as a colorless oil. This material was purified by flash
chromatography using hexane and 1:79 EtOAc−hexane as mobile
phases. Fractions were pooled according to TLC and evaporated to
give the tetrasilyl ether 27 as a colorless oil, 0.34 g. Continued elution
with EtOAc recovered the unreacted diphenylphosphine oxide 30b.
The tetrasilyl ether 27 was dissolved in a 1 M solution of
tetrabutylammonium fluoride in THF (5 mL) and kept at ambient
temperature for 3 h and then in the refrigerator for 48 h. The solution
was then diluted with brine (10 mL), stirred for 10 min, and water (20
mL) was added to dissolve precipitated salt. This mixture was
equilibrated with EtOAc (40 mL), and the aqueous phase was re-
extracted with EtOAc (20 mL); the combined extracts were washed
with water (5 × 10 mL) and then with brine (5 mL), dried, and
evaporated to a colorless oil that was taken up in 2:1 EtOAc−hexane
and charged to a flash column, using stepwise gradients of 2:1 and 4:1
EtOAc−hexane, EtOAc, and 1:9 MeOH−EtOAc as mobile phases.
Fractions containing a minor quantity of partially desilylated material
preceded the product band. These fractions were pooled and
evaporated to give the tetraol 6 as white solid residue. This material
was dissolved in methyl formate to which a few drops of MeOH were
added to achieve complete dissolution. The resulting solution was
filtered and concentrated, then flash evaporated and dried under high
vacuum for 3 h to furnish 6 as an amorphous powder (183 mg, 84%
from 15b); [α]D +51.0 (MeOH, c 0.253). 1H NMR (400 MHz,
CD3Cl) δ 0.53 (s, 3 H), 1.14−2.08 (m, 30 H), 2.14−2.26 (m, 2 H),
2.47 (d, J = 10.9 Hz, 1 H), 2.68−2.85 (m, 2 H), 3.96−4.16 (m, 2H),
5.85 (d, J = 11.1 Hz, 1 H), 6.30 (d, J = 11.3 Hz, 1 H) ppm. 13C NMR
12.2, 19.8, 20.1, 22.18, 23.5, 27.1, 29.0, 31.3, 37.2, 39.4, 40.2, 42.2,
44.3, 44.4, 44.6, 45.8, 53.0, 56.3, 67.1, 67.3, 70.8, 115.2, 123.6, 131.2,
142.8 ppm. UVmax (ε) 243 (32529), 251 (37925), 261 (25718) nm.
LR-ES(+) m/z 566 (55, M + Na + CH3CN), 525 (29, M + Na), 503
(13, M + H), 467 (55, 503-2H2O), 449 (16, 467-H2O), 431 (449-
H2O). HR-ES(+) calcd for M + Na 531.4291, found 531.4289.
Ethyl (5S)-5-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]-
oxy-7a-methyl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-9-hy-
droxy-9-methyl-decanoate (19). A mixture of zinc powder (0.600
g, 9.18 mmol), pyridine (11.2 mL), and ethyl acrylate (1.0 mL, 0.918
g, 9.17 mmol) was stirred vigorously, and nickel chloride hexahydrate
(0.5364 g, 2.26 mmol) was added. The gray suspension was processed
as described for 13, and then a solution of the iodo compound 18
(1.02 g, 1.95 mmol) in pyridine (7 mL) was added dropwise over a
period of 20 min, followed by pyridine rinses (2 × 2 mL). The ice bath
was removed, and stirring continued at ambient temperature for 2 h.
At that time, 18 was no longer detectable (TLC 1:3 EtOAc−hexane, Rf
6-[(1R,3aR,4S,7aR)-4-Hydroxy-7a-methyl-1,2,3,3a,4,5,6,7-oc-
tahydroinden-1-yl]-1,1,1,11,11,11-haxadeuterio-2,10-bis-
(trideuteriomethyl)undecane-2,10-diol (14b). Diol 14a (0.31 g,
0.626 mmol) was dissolved in acetonitrile (10 mL), and a
hydrofluorosilicic acid solution (35%) was added. The mixture was
stirred, and the deprotection progress was monitored by TLC. After 3
h, the reaction was ca. 30% complete, another 1.5 mL of
hydrofluorosilicic acid solution was added, and the mixture was stirred
for a total of 24 h, at which time the reaction was complete. The
mixture was equilibrated with EtOAc (60 mL) and water (20 mL), the
resulting aqueous phase was re-extracted with EtOAc (20 mL), and the
combined organic layers were washed with water (5 mL), dried, and
concentrated. The residue was diluted with toluene and concentrated
to remove acetonitrile then chromatographed on silica gel using 1:3,
1:2, and 1:1 EtOAc−hexane as mobile phases. The appropriate
fractions were pooled and evaporated to furnish the triol 14b a white,
solid foam, 0.21 g, 88%. TLC 1:1 EtOAc−hexane Rf 0.34, 1:19
1
MeOH−DCM, Rf 0.55. H NMR (CDCl3) δ 0.91 (s, 3H), 1.09−1.63
(m, 23H), 1.70−1.88 (m, 4H), 1.92 (2H, br d), 4.05 (1H, br s). 13C
NMR (CDCl3) δ 13.7, 17.6, 19.8, 20.0, 22.5, 26.8, 31.1, 31.2, 33.6,
38.6, 40.1, 41.9, 44.3, 44.4, 52.6, 53.0, 69.3, 70.7. LR-MS-ES(+) m/z
444 (80, M + Na + CH3CN), 403 (48, M + Na), 363 (15, M + H −
H2O).
(1R,3aR,7aR)-7a-Methyl-1-[6,6,6-trideuterio-5-(trideuterio-
methyl)-1-[5,5,5-trideuterio-4-(trideuteriomethyl)-4-trimethyl-
silyloxy-pentyl]-5-trimethylsilyloxy-hexyl]-2,3,3a,5,6,7-hexa-
hydro-1H-inden-4-one (15b). Celite (0.5 g) and pyridinium
dichromate (1.13 g, 3 mmol) was added to a stirred solution of triol
14b (0.189 g, 0.496 mmol) in DCM (10 mL). The oxidation progress
was followed by TLC. The reaction mixture was poured onto a
column of silica gel G60 after 8 h, the column was rinsed with 1:4
EtOAc−ether (300 mL), and the residue, obtained after evaporation of
solvents, was taken up in DCM and evaporated again to yield 15a as a
sticky mass. LR-MS-ES(+) m/z 442 (45, M + Na + CH3CN), 402 (51,
M + H + Na), 401 (25, M + Na), 384 (343 + CH3CN), 361 (10, M +
H − H2O), 343 (95, M + H − H2O-CD3), 325 (58, 343 − H2O). This
material was directly taken up in cyclohexane (3 mL), and
trimethylsilyl imidazole (0.3 mL, 2 mmol) was added. The mixture
was stirred for a period of 6.5 h employing TLC 1:19 and 1:39
MeOH−DCM as process control, then diluted with hexane (10 mL)
and chromatographed on a silica gel column in hexane using a stepwise
gradient of hexane, 1:39, 1:19, and 1:9 EtOAc−hexane. The peak
fractions were pooled and evaporated to afford the protected ketone
G
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
0.70). The mixture was then poured into EtOAc (50 mL), the mixture
was filtered through Celite, filter cake was washed with EtOAc (40
mL), and the combined filtrate and washings were sequentially
equilibrated with 1 M HCl (50 mL), 0.5 M HCl (50 mL), and 2 M
phosphoric acid (40 mL). The solution was then washed with brine (2
× 15 mL), dried (MgSO4), and evaporated. The resulting residue
(1.15 g) was flash chromatographed using 1:19, 1:9, and 1:6 EtOAc−
hexane as mobile phases. The homogeneous fractions were pooled and
evaporated. The residue was evaporated twice from hexane then dried
to give 19 (0.82 g, 85%). TLC (1:3 EtOAc−hexane) Rf 0.52; [α]D
(1R,3aR,7aR)-7a-Methyl-1-[(1S)-6,6,6-trideuterio-1-(4-meth-
yl-4-trimethylsilyloxy-pentyl)-5-(trideuteriomethyl)-5-trime-
thylsilyloxy-hexyl]-2,3,3a,5,6,7-hexahydro-1H-inden-4-one
(21b). Trimethylsilyl imidazole (0.6 mL, 4 mmol) was added to the
suspension of 21a (0.32 g, 0.859 mmol) in cyclohexane (6 mL). The
starting material dissolved, and a new suspension formed within
minutes which was stirred overnight, diluted with hexane (10 mL), and
charged to a silica gel column. Elution with 1:39 and 1:19 EtOAc−
hexane, pooling of appropriate fractions, and evaporation gave a
residue that was coevaporated from hexane and dried to give 21b as an
oily substance. TLC 1:19 MeOH−DCM (Rf 0.91) and 1:39 EtOAc−
hexane (Rf 0.09). This material was divided into two nearly equal parts
and used directly for the syntheses of 7a and 7b as described below.
1,25-Dihydroxy-21-(3-hydroxy-3-trideuteriomethyl-4,4,4-tri-
deuteriobut-1-yl)-cholecalciferol (7a). A 25 mL 2-neck flask
equipped with thermometer, Claisen adapter containing a nitrogen
sweep, and rubber septum was charged with the diphenylphosphine
oxide 30a (0.5001 g, 0.858 mmol) previously dried for 3 h. This
material was dissolved in THF and cooled to −70 °C, and 2 M
phenyllithium in hexane (0.43 mL) was added within 10 min. The
deep orange-colored solution was stirred at −70 °C for 5 min, then a
solution of the ketone 21b (0.227 g, 0.439 mmol), dissolved in THF
(3 mL), was added dropwise below −65 °C during 30 min. The flask
containing the ketone was rinsed with THF (1 mL), and this rinse was
also added to the reactor after 2 h. The mixture was stirred for an
additional 30 min. At that time, 21b (TLC 1:39, Rf 0.10 and 1:19
EtOAc−hexane, Rf 0.37) was converted to 28a (TLC 1:39, Rf 0.30 and
1:19 EtOAc−hexane, Rf 0.73). Thus, the mixture was allowed to warm
to −30 °C, and pH 7 phosphate buffer (5 mL) was added dropwise.
The mixture was stirred vigorously for 25 min. At that time, all ice had
melted and the solution was equilibrated with hexane (40 mL). The
aqueous phase was re-extracted with hexane (20 mL). The combined
extracts were washed with brine (5 mL), dried, and evaporated to give
the crude tetrasilyl ether 28a as a colorless oil. This material was
dissolved in hexane and charged onto a flash column, which was eluted
with hexane and 1:79 EtOAc−hexane. Fractions were pooled and
evaporated, and the residue coevaporated once from hexane to give
28a (0.36 g) as a colorless oily residue. This material was dissolved in a
1 M solution of tetrabutylammonium fluoride in THF (6 mL) and
kept at room temperature for 29 h. This mixture was then equilibrated
with EtOAc (40 mL). The aqueous phase was extracted with EtOAc
(20 mL), and the combined extracts were washed with water (5 × 10
mL) and then with brine (5 mL), dried, and evaporated to leave a
colorless oil that was taken up in 2:1 EtOAc−hexane and charged to a
flash column using 2:1, 4:1, and EtOAc as stepwise gradients. Fractions
were pooled and evaporated, the residue was dissolved in methyl
formate, the solution filtered and concentrated, and the residue was
flash evaporated and dried for 3 h to afford 7a as amorphous solids
1
+31.5 (chloroform, c 0.308). IR (CHCl3) ν CO (1724 cm−1). H
NMR (CDCl3) δ −0.01 (s, 3 H), 0.00 (s, 3 H), 0.88 (s 9 H), 0.89 (s 3
H), 1.06−1.93 (m, 24 H), 1.21 (s, 6H), 1.25 (t, J = 7.2 Hz, 3 H), 2.25
(t, J = 7.4 Hz, 2H), 3.98 (br m, 1 H), 4.12 (q, J = 7.2 Hz). 13C NMR
(CDCl3) δ −5.0, −4.7, 14.0, 14.4, 17.8, 18.1, 20.0, 20.9, 23.0, 25.9,
26.8, 29.3, 29.4, 30.1, 31.2, 34.5, 34.8, 38.4, 40.4, 42.2, 44.5, 53,0, 53.1,
60.2, 69.4, 71.0, 173.8. LR-ES(+) m/z 560 (14, M + Na + CH3CN),
519 (10, M + Na), 479 (19, M + H − H2O), 347 (100, 479-TBSOH).
(6S)-6-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]oxy-7a-
methyl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-1,1,1-trideuterio-
10-methyl-2-(trideuteriomethyl)undecane-2,10-diol (20a). A
solution of monoester 19 (0.69 g, 1.39 mmol) in ether (10 mL)
was treated with a 1 M solution of deuteriomethylmagnesium iodide in
ether (6.8 mL) and worked up as described for 14a to afford 20a as a
colorless oil (0.94 g), which was chromatographed on a flash column
using 1:9, 1:6, 1:4, 1:3, and 1:2 EtOAc−hexane as stepwise gradients.
Fractions were pooled and evaporated to leave a residue that was taken
up in pentane and then filtered and evaporated to give 20a as a
colorless amorphous solid (0.57 g, 84%). TLC (1:3 EtOAc−hexane)
Rf 0.11. 1H NMR (CDCl3) δ −0.01 (s, 3 H), 0.00 (s, 3 H), 0.88 (s, 9
H), 0.90 (s 3 H), 1.07−1.84 (m, 26 H), 1.21 (s 6 H), 1.89 (m, 1 H),
3.99 (br m, 1 H). LR-ES(+) m/z 511 (4, M + Na), 471 (10, M + H −
H2O), 453 (28, 471-H2O), 357 (5, M + H − TBSOH), 339 (12, 471-
TBSOH), 321 (100, 453-TBSOH).
(6S)-6-[(1R, 3aR, 4S, 7aR)-4-Hydroxy-7a-methyl-
1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-1,1,1-trideuterio-10-
methyl-2-(trideuteriomethyl)undecane-2,10-diol (20b). A sol-
ution of 20a (0.54 g, 0.110 mmol) in acetonitrile (20 mL) was stirred,
and a 35% hydrofluorosilicic acid solution (3 mL) was added. The
hydrolysis was complete after 15 h, as evidenced by the disappearance
of 20a (TLC 1:1 EtOAc−hexane, Rf 0.64). The solution was
equilibrated with EtOAc (60 mL) and water (20 mL), the aqueous
layer was extracted with EtOAc (25 mL), and the combined organic
layers were washed with water (8 mL). The resulting organic layer was
dried and concentrated, and the residue was diluted with toluene and
concentrated to remove acetonitrile and then chromatographed using
1:3, 1:2, and 1:1 EtOAc−hexane as mobile phases. Pooled fractions
were evaporated, and the residue was coevaporated twice from hexane,
taken up in 1:1 DCM−hexane, filtered, and evaporated again and dried
to leave the triol 20b as an amorphous powder (0.36 g, 87%). TLC
(1:1 EtOAc−hexane) Rf 0.34; [α]D +25.2 (MeOH, c 0.306). 1H NMR
(CDCl3) δ 0.92 (s, 3 H), 1.10−1.66 (m, 24H), 1.21 (s, 6 H), 1.67−
1.90 (m, 3 H), 1.94 (br m, 1H), 4.07 (br m, 1 H). 13C NMR (CDCl3)
δ 13.7, 17.6, 19.8, 20.1, 22.5, 26.8, 29.3, 29.4, 31.1, 31.3, 33.7, 38.7,
40.1, 44.3, 44.6, 52.6, 53.0, 69.4, 71.05. LR-MS-ES(+) m/z 438 (18, M
+ Na + CH3CN), 397 (34, M + Na), 357 (20, M + H − H2O), 339
(54, 357-H2O), 321 (100, 339-H2O).
1
(0.1909 g, 85%), [α]D +16.0 (c 0.32, MeOH). H NMR (CDCl3) δ
0.50 (s, 3H), 1.05 (s, 6H), 1.10−1.54 (m, 19H, 1.56−1.71 (m, 3H),
1.72−1.84 (m, 2H), 1.84−2.00 (m, 2H), 2.16 (dd, J = 13.5, 5.6 Hz,
1H), 2.36 (d, J = 11.1 Hz, 1H), 2.72−2.87 (m, 1H), 3.90−4.08 (m,
3H), 4.12−4.27 (m, 1H), 4.55 (d, J = 3 Hz), 1H), 4.76 (d, J = 2.1 Hz,
1H), 4.86 (d, J = 4.7 Hz, 1H), 5.22 (d, J = 1.3 Hz), 5.99 (d, J = 11.1
Hz, 1H), 6.19 (d, J = 11.3 Hz, 1H). 13C NMR (CDCl3) δ 11.9, 19.5,
19.7, 21.8, 23.2, 26.7, 28.4, 29.2, 29.4, 31.1, 31.2, 38.7, 38.9, 39.0, 39.1,
39.3, 39.5, 39.6, 39.7, 39.9, 40.1, 43.1, 44.1, 44.3, 44.9, 45.3, 50.5, 52.5,
55.6, 65.0, 68.4, 68.5, 68.7, 109.9, 117.4, 122.4, 135.7, 140.0, 149.3.
UVmax (ε) 211 (14747), 265 (15897). HR-ES(+) calcd for
C32H48D6O4 + Na 531.4291, found 531.4290.
1,25-Dihydroxy-21-(3-hydroxy-3-trideuteriomethyl-4,4,4-tri-
deuteriobut-1-yl)-19-nor-cholecalciferol (7b). The protocol
described above was repeated using diphenylphosphine oxide 30b
(0.4289 g, 0.755 mmol), 1.6 M butyl lithium in hexane (0.47 mL 0.755
mmol), and 21b (0.2055 g, 0.397 mmol). The resulting 28b was taken
up in a 1 M solution of tetrabutylammonium fluoride in THF (5 mL),
and the solution was kept at room temperature for 47 h. Following an
extractive isolation as described above, 21a was obtained as a colorless
oil that was taken up in 2:1 EtOAc−hexane and charged to a flash
column which was eluted with 1:9, 2:1, 4:1 EtOAc−hexane, EtOAc,
(1R,3aR,7aR)-7a-Methyl-1-[(1S)-6,6,6-trideuterio-5-hydroxy-
1-(4-hydroxy-4-methyl-pentyl)-5-(trideuteriomethyl)hexyl]-
2,3,3a,5,6,7-hexahydro-1H-inden-4-one (21a). A solution of 20b
(0.35 g, 0.934 mmol) in DCM (20 mL) was stirred vigorously, and
Celite (1 g) was added followed by pyridinium dichromate (2.26 g, 6
mmol). The oxidation was complete after 4 h. The reaction mixture
was poured onto a column of silica gel G60, and the column was
rinsed with 1:4 EtOAc−ether (300 mL). The residue obtained after
evaporation of the column effluent was taken up in DCM, evaporated
to a foam, and dried to afford the ketone 21a (0.32 g, 92%). TLC
(1:19 MeOH−DCM) Rf 0.32. IR (CHCl3) ν CO (1705 cm−1), LR-
ES(+) m/z 396 (20, 355 + CH3CN), 378 (337 + CH3CN), 355 (8, M
+ H − H2O), 337 (50, 355-H2O), 319 (48, 337-H2O).
H
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and MeOAc. Fractions were pooled and evaporated. The white residue
obtained was evaporated twice with hexane and the resulting residue
(0.1 g) dissolved in methyl formate to which a few drops of MeOH
were added to ensure complete dissolution. This solution was filtered
and concentrated then evaporated and dried to give 7b as an
29.3, 29.4, 31.1, 31.3, 33.7, 38.7, 40.1, 42.0, 44.4, 52.6, 53.0, 69.4, 70.8,
71.0. m/z 497 (22, M + CH3CN − H2O), 369 (20, M + Na − H), 357
(39, M + H − H2O), 339 (58, 357 − H2O). LR-ES(+) m/z 397 (15,
M + Na), 357 (33, M + H − H2O), 339 (357-H2O). LR-ES(−) m/z
409 (80, M + Cl).
(1R,3aR,7aR)-7a-Methyl-1-[(1R)-6,6,6-trideuterio-5-hydroxy-
1-(4-hydroxy-4-methyl-pentyl)-5-(trideuteriomethyl)hexyl]-
2,3,3a,5,6,7-hexahydro-1H-inden-4-one (26a). A mixture of the
triol 25b (0.45 g. 1.20 mmol), DCM (25 mL), and Celite (1.2 g) was
treated with pyridinium dichromate (2.71 g, 7.3 mmol) as described
for 21a to yield ketone 26a (0.44 g, 98%) as an amorphous powder,
TLC 1:19 MeOH−DCM 0.36, IR (CHCl3) ν CO (1705 cm−1),
LR-ES(+) m/z 396 (40, M + H + CH3CN − H2O), 378 (24, 337 +
CH3CN), 355 (8, M + H − H2O), 337 (100, 355-H2O), 319 (59, 337-
H2O).
1
amorphous solid (0.118 g, 60%); [α]D +53.8 (c 0.31, MeOH). H
NMR δ 0.50 (s, 3H), 1.06 (s, 6H), 1.13−1.73 (m, 23H), 1.71−2.10
(m, 5H), 2.26 (d, J = 10.7 Hz, 1H), 2.38−2.46 (m, 1H), 2.62−2.86
(m, 1H), 3.75−3.91 (m, 2H), 3.94−4.18 (m, 2H), 4.23−4.69 (m, 2H),
5.79 (d, J = 11.1 Hz), 6.07 (s, J = 11.1 Hz). 13C NMR δ 12.1, 19.5,
19.8, 21.8, 23.1, 26.8, 28.3, 29.2, 29.4, 31.1, 31.2, 37.0, 38.9, 39.1, 39.3,
39.5, 39.7, 39.9, 40.1, 42.2, 44.1, 44.4, 44.6, 45.2, 52.6, 55.6, 65.3, 65.6,
68.5, 68.8, 115.9, 120.9, 134.5, 139.4. UVmax (ε) 243 (31661), 252
(36406), 261 (24892). LR-ES(+) m/z 461 (50, M + H − H2O), 443
(28, 461-H2O), 275 (100, M − C13H22D6O2 + H). HR-ES(+) calcd
for (C31H48D6O4 + Na]+ 519.4291, found 519.4294.
(1R,3aR,7aR)-7a-Methyl-1-[(1R)-6,6,6-trideuterio-1-(4-meth-
yl-4-trimethylsilyloxy-pentyl)-5-(trideuteriomethyl)-5-trime-
thylsilyloxy-hexyl]-2,3,3a,5,6,7-hexahydro-1H-inden-4-one
(26b). Trimethylsilyl imidazole (0.75 mL, 4 mmol) was added to a
suspension of 26a (0.44 g, 1.18 mmol) in cyclohexane (8 mL) and
processed as described for 21b to furnish the disilyl ether 26b (0.59 g,
96%) as a colorless syrup. TLC 1:19 MeOH−DCM, Rf 0.91, and 1:19
EtOAc−hexane, Rf 0.30. 1H NMR (CDCl3) δ 0.10 (s, 9 H), 0.11 (s, 9
H), 0.64 (s, 3 H), 1.10−2.11 (m, 22 H), 1.21 (s, 6 H), 2.16−2.34 (m,
2 H), 2.46 (dd, J = 7.4 and 11.4 Hz, 1 H). This material was split into
two nearly equal parts and used for the synthesis of 8a and 8b.
[(1S,3Z,5R)-3-[(2E)-2-[(1R,3aS,7aR)-7a-Methyl-1-[(1R)-6,6,6-
trideuterio-1-(4-methyl-4-trimethylsilyloxy-pentyl)-5-(trideu-
teriomethyl)-5-trimethylsilyloxy-hexyl]-2,3,3a,5,6,7-hexahy-
dro-1H-inden-4-ylidene]ethylidene]-5-[tert-butyl(dimethyl)-
silyl]oxy-2-methylene-cyclohexoxy]-tert-butyl-dimethyl-silane
(29a). Diphenylphosphine oxide 30a (0.5238 g, 0.899 mmol) was
deprotonated and coupled to the ketone 26b (0.2942 g, 0.569 mmol)
as described for 28a to yield crude 29a. This material was dissolved in
hexane and flash-chromatographed using a stepwise gradient of hexane
and 1:79 EtOAc−hexane to remove some unreacted ketone (TLC,
1:19 EtOAc−hexane, Rf 0.33). Homogeneous fractions were pooled
and evaporated. The colorless residue was coevaporated once from
hexane, to give 29a (0.30 g, 60%) as a colorless oily residue. TLC
(1:19 EtOAc−hexane) Rf 0.73, (1:39 EtOAc−hexane) Rf 0.27. LR-
ES(+) m/z 881 (15, M + H), 659 (12, M + H − TMSOH −
TBSOH).
Ethyl (5R)-5-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]-
oxy-7a-methyl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-9-hy-
droxy-9-methyl-decanoate (24). Zn powder (1.12 g, 17.13 mmol),
pyridine (I20 g), ethyl acrylate (1.87 mL, 17.13 mmol, 1.715 g), nickel
chloride hexahydrate (1.00 g, 4.22 mmol), and iodide 23 (1.91 g, 3.65
mmol) in pyridine (13 mL) was processed as described for 13 TLC
(1:3 ethyl acetate−hexane) to achieve a complete conversion of 23 (Rf
0.57) to the ester 24 (Rf 0.52). The mixture was then poured into
EtOAc (80 mL), the resulting suspension was filtered through Celite,
and the filter cake was washed with EtOAc (40 mL). Filtrate and
washings were combined and equilibrated with 2 M phosphoric acid,
and the slightly acidic extract was washed with brine (5 mL), dried,
and evaporated. The residue was flash chromatographed using hexane
and 1:19 and 1:9 EtOAc−hexane as stepwise gradient. Homogeneous
fractions were pooled and evaporated (0.64 g), tale fractions were also
pooled and evaporated and purifies by preparative HPLC on a 2 in. ×
18 in. 15−20 μ silica YMC HPLC column using 1:3 EtOAc−hexane as
mobile phase and running at 100 mL/min to afford additional pure 24
(peak volume: 1.85 L, 1.01 g of residue), total yield 1.65 g, 86%. TLC
1:2 EtOAc−hexane, Rf 0.70; [α]D +12.2 (chloroform, c 0.262). IR
(CHCl3) ν CO (1725 cm−1). 1H NMR (CDCl3) δ −0.01 (s, 3 H),
0.0 (s 3 H), 0.88 (s, 9 H), 0.89 (s, 3 H), 1.06−1.93 (m, 24 H), 1.21 (s,
6 H), 1.25 (t, J = 7.2 Hz, 3 H), 2.26 (m, 2 H), 3.98 (br m, 1 H), 4.12
(q, J = 7.2 Hz, 2H). 13C NMR (CDCl3) δ 13.9, 14.4, 17.8, 19.9, 21.0,
23.0, 25.9, 26.8, 29.3, 29.4, 30.2, 31.1, 34.5, 34.9, 38.5, 40.5, 42.2, 44.4,
53.0, 53.1, 60.2, 69.4, 71.0, 76.5, 173.8. LR-ES(+) m/z 993 (32 2M +
1), 560 (18, M + Na + CH3CN), 479 (54, M + H − H2O), 347 (72, M
− H2O − THSOH).
(6R)-6-[(1R,3aR,4S,7aR)-4-[tert-Butyl(dimethyl)silyl]oxy-7a-
methyl-1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-1,1,1-trideuterio-
10-methyl-2-(trideuteriomethyl)undecane-2,10-diol (25a). A
solution of the ester 24 (1.50 g, 3.02 mmol) in ether (20 mL) was
treated with a 1 M solution of deuteriomethylmagnesium iodide in
ether (16 mL) as described previously. The resulting oily 25a (1.63 g)
was chromatographed as described to afford a residue (1.43 g) which
was taken up in pentane, filtered, concentrated, flash evaporated, and
dried to yield 25a as a colorless powder (1.32 g, 89%). TLC 1:2
EtOAc−hexane, Rf 0.37. LR-ES(+) m/z 453 (35, M + H − 2H2O). 1H
NMR (CDCl3) δ −0.01 (s, 3 H), 0.00 (s, 3 H), 0.88 (s, 9 H), 0.90 (s,
3 H), 1.07−1.61 (m, 23H), 1.21 (s, 6 H), 1.62−1.85 (m, 3 H), 1.89
(br m, 1 H), 3.99 (br m, 1 H).
1α,25-Dihydroxy-21-(3-hydroxy-3-methyl-but-1-yl)-26,27-
hexadeuterio-cholecalciferol (8a). The residue obtained above was
dissolved in a 1 M solution of tetrabutylammonium fluoride in THF (6
mL) and kept at room temperature for 30 h. The solution was diluted
with brine (10 mL) and stirred for 10 min, and then water (5 mL) was
added to dissolve precipitated salt. This mixture was equilibrated with
EtOAc (40 mL) and then worked up and purified as described for 7a
to yield 8a as amorphous solids (0.151 g, 87%). TLC (EtOAc) Rf 0.64;
[α]D +11.9° (MeOH, c 0.16). UVmax (MeOH) ε 211 (16548), 266 nm
1
(16932). H NMR (DMSO-d6) δ 0.50 (s, 3H), 1.06 (s, 9H), 1.10−
1.53 (m, 19H), 1.57−1.69 (m, 3H), 1.71−1.84 (m, 2H), 1.89 (d, J =
10.7 Hz, 1H), 1.93−2.02 (m, 1H), 2.16 (dd, J = 14.0, 6.1 Hz), 2.36 (d,
11.1 Hz, 1H), 2.79 (d, J = 8.7 Hz, 1H), 3.99 (br s, 1H), 4.01 (s, 1H),
4.04 (s, 1H), 4.19 (br s, 1H), 4.54 (d, J = 3.6 Hz, 1H), 4.85 (d, J = 4.9
Hz, 1H), 5.22 (br s, 1H), 5.99 (d J = 11.5 Hz, 1H), 6.19 (d, J = 11.7
Hz). 13C NMR (DMSO-d6) 11.9, 19.5, 19.7, 21.8, 23.2, 26.7, 28.4,
29.2, 29.4, 31.1, 31.2, 38.7, 38.9, 39.1, 39.3, 39.5, 39.7, 39.9, 40.1, 43.1,
44.2, 44.9, 45.3, 52.5, 55.6, 65.0, 68.4, 68.5, 68.2, 109.9, 117.4, 122.4,
135.7, 140.0, 149.3. LR-ES(+) m/z 509 (15, M + H), 491 (34, M + H
− H2O), 473 (100, 491-H2O), 455 (60, 473-H2O), 437 (15, 455-
H2O).
(6R)-6-[(1R, 3aR, 4S, 7aR)-4-Hydroxy-7a-methyl-
1,2,3,3a,4,5,6,7-octahydroinden-1-yl]-1,1,1-trideuterio-10-
methyl-2-(trideuteriomethyl)undecane-2,10-diol (25b). A sol-
ution of the tetrasilyl ether 25a (1.32 g, 0.110 mmol) in a mixture of
THF (5 mL) and acetonitrile (45 mL) was further diluted with a
hydrofluorosilicic acid solution (35%, 7.5 mL). The conversion of 25a
(TLC 1:1 EtOAc−hexane, Rf 0.64) to 25b (Rf 0.34) was complete
after 14.5 h. The solution was diluted with EtOAc (120 mL) and water
(40 mL) and processed as described for 20b to afford 25b as an
amorphous powder, 0.87 g (86%). TLC 1:19 MeOH−DCM Rf 0.30;
[(1R,5R)-3-[(2E)-2-[(1R,3aS,7aR)-7a-Methyl-1-[(1R)-6,6,6-tri-
deuterio-1-(4-methyl-4-trimethylsilyloxy-pentyl)-5-(trideuter-
iomethyl)-5-trimethylsilyloxy-hexyl]-2,3,3a,5,6,7-hexahydro-
1H-inden-4-ylidene]ethylidene]-5-[tert-butyl(dimethyl)silyl]-
oxy-cyclohexoxy]-tert-butyl-dimethyl-silane (29b). The protocol
for 7b was repeated using diphenylphosphine oxide 30b (0.5844 g,
1.02 mmol, 1.82 equiv) and ketone 26b (0.2833 g, 0.548 mmol) to
furnish 29b (0.41 g, 77%) as a colorless oil. TLC (1:19 EtOAc−
1
[α]D +23.9 (MeOH), c 0.426). H NMR (CDCl3) δ 0.93 (s, 3H),
1.12−1.55 (m, 28H), 1.73−1.86 (3H, m), 1.90−1.98 (1H, m), 4.07
(1H, m) ppm. 13C NMR (CDCl3) δ 13.7, 17.6, 19.9, 20.1, 22.5, 26.8,
I
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1
hexane) Rf 0.75. H NMR (CDCl3) δ 0.4 (s, 3H), 0.05 (s, 6H), 0.06
(3) Kushner, D. J.; Baker, A.; Dunstall, T. G. Pharmacological uses
and perspectives of heavy water and deuterated compounds. Can. J.
Physiol. Pharmacol. 1999, 77, 79−88.
(s, 3H), 0.10 (s, 18H), 0.53 (s, 3H), 0.86 (s, 9H), 0.87 (s, 9H), 1.11−
1.73 (m, 21H), 1.20 (s, 6H), 1.74−1.89 (m, 3H), 1.94 (br d, J ∼ 12.1
and 1 Hz, 1H), 2.00 (brt, J ∼ 9.6 and 1 Hz), 2.10 (dd, J = 7.4 and 13.1
Hz, 1H), 2.26 (dd, J ∼ 2.7 and 13.4 Hz, 1H), 2.34−2.42 (m, 2H), 2.81
(br m, 1H), 4.02−4.12 (m, 2H), 5.82 (d, J = 11.2 Hz, 1H), 6.17 (d, J =
11.2 Hz, 1H).
1,25-Dihydroxy-21-(3-hydroxy-3-methyl-but-1-yl)-19-nor-
26,27-hexadeuterio-cholecalciferol (8b). The tetrasilyl ether 29b
obtained above (0.41 g, 0.471 mmol) was dissolved in a 1 M solution
of tetrabutylammonium fluoride in THF (8 mL). The deprotection
was complete after 62 h at room temperature. The solution was diluted
with brine (15 mL), stirred for 10 min, and then equilibrated with the
EtOAc (50 mL) and treated as described for deprotection of 28b to
afford 8b as an amorphous powder (0.202 g, 86%). TLC (EtOAc) Rf
0.36; [α]D +54.5 (MeOH, c 0.398). UVmax (ε) 243 (25005), 251
(4) Carlberg, C. Molecular basis of the selective activity of vitamin D
analogues. J. Cell. Biochem. 2003, 88, 274−281.
(5) Yang, W.; Freedman, L. P. 20-Epi analogues of 1,25-dihydroxy-
vitamin D3 are highly potent inducers of DRIP coactivator complex
binding to the vitamin D3 receptor. J. Biol. Chem. 1999, 274, 16836−
16845.
(6) Rachez, C.; Suldan, Z.; Ward, J.; Chang, C-P; Burakov, D.;
Erdjument-Bromage, H.; Tempst, P.; Freedman, L. P. A novel protein
complex that interacts with the vitamin D3 receptor in a ligand-
dependent manner and enhances VDR transactivation in a cell-free
system. Genes Dev. 1998, 12, 1787−1800.
(7) Norman, A. W.; Manchand, P. S.; Uskokovic, M. R.; Okamura,
W. H.; Takeuchi, J. A.; Bishop, J. E.; Hisatake, J-I; Koeffler, H. P.;
Peleg, S. Characterization of a novel analog of 1α,25(OH)2-vitamin D3
with two side chains: interaction with its nuclear receptor and cellular
action. J. Med. Chem. 2000, 43, 2719−2730.
(8) Uskokovic, M. R; Maehr, H.; Reddy, S. G; Li, Y. C.; Adorini, L.;
Holick, M. F. Gemini: the 1,25-dihydroxy vitamin D analogs with two
side-chains. Vitamin D, 2nd ed.; Feldman, D., Pike, J. W., Glorieux, F.
H., Eds.; Elsevier Academic Press: London, 2005; pp 1511−1524.
(9) Lee, H. J.; Ji, Y.; Paul, S.; Maehr, H.; Uskokovic, M.; Suh, N.
Activation of bone morphogenetic protein signaling by a gemini
vitamin D3 analogue is mediated by Ras/protein kinase Cα. Cancer
Res. 2007, 67, 11840−11847.
1
(29167), 261 (19777) nm (MeOH). H NMR (DMSO-d6) δ 0.50 (s,
3H), 1.05 (s, 6H), 1.10−1.72 (m, 23H), 1.73−1.85 (m, 1H), 1.90 (d, J
= 11.3 Hz), 1.94−2.11 (m, 3H), 2.19−2.32 (m, 1H), 2.38−2.47 (m,
1H), 2.63−2.84 (m, 1H), 3.73−3.93 (m, 2H), 4.04 (d, 8.5 Hz, 2H),
4.38 (d, J = 3.4 Hz), 4.49 (d, J = 3.6 Hz), 5.80 (d, J = 11.1 Hz), 6.08
(d, J = 11.1 Hz, 1H). 13C NMR δ 12.0, 19.5, 19.7, 21.7, 23.1, 26.8,
28.3, 29.2, 29.4, 31.1, 31.2, 37.00, 38.7, 38.9, 39.1, 39.3, 39.5, 39.7,
39.9, 40.1, 42.2, 44.2, 44.6, 45.2, 52.5, 55.6, 65.3, 65.5, 68.4, 68.7,
115.9, 129.9, 134.5, 139.4. LR-ES(+) m/z 497 (6, M + H − H2O), 461
(78, 479-H2O), 443 (20, 461-H2O), 425 (6, 443-H2O). HR-ES(+)
calcd for C31H48D6O4 + Na 519.4291, found 519.4290.
(10) Lee, H. J.; Paul, S.; Atalla, N.; Thomas, P. E.; Lin, X.; Yang, I.;
Buckley, B.; Lu, G.; Zheng, X.; Lou, Y. R.; Conney, A. H.; Maehr, H.;
Adorini, L.; Uskokovic, M.; Suh, N. Gemini vitamin D analogues
inhibit estrogen receptor-positive and estrogen receptor-negative
mammary tumorigenesis without hypercalcemic toxicity. Cancer Prev.
Res. 2008, 1960, 1−9.
(11) So, J. Y.; Lee, H. J.; Smolarek, A. K.; Paul, S.; Wang, C.-X.;
Maehr, H.; Uskokovic, M.; Zheng, X.; Conney, A.; Cai, L.; Liu, F.; Suh,
N. A novel vitamin D analog represses the expression of stem cell
marker CD44 in breast cancer. Mol. Pharmacol. 2011, 79, 360−367.
(12) Burch, W. M.; Lopez-Claros, M.; Uskokovic, M. R.; Drezner, M.
K. 1,25-Dihydroxyvitamin D3 stimulates avian and mammalian
cartilage growth in vitro. J. Bone Mineral Res. 1988, 3, 87−91.
(13) Sanderson, K. Big interest in heavy drugs. Nature 2009, 458,
269.
ASSOCIATED CONTENT
* Supporting Information
Analytical data for final compounds and crystal data, including
refinement characteristics. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
Accession Codes
The PDB ID codes shown below are also included in the
Supporting Information. PDB ID: 4IA2 (7a); 4IA1 (7b); 4IA3
(8a); 4IA7 (8b).
AUTHOR INFORMATION
Corresponding Author
*Phone: (973) 831-1591. E-mail: hubert.maehr@verizon.net.
■
(14) Maehr, H.; Uskokovic, M. R.; Adorini, L.; Penna, G.; Mariani,
R.; Panina, P.; Passini, N.; Bono, E.; Perego, S.; Biffi, M.; Holick, M.;
Spina, C.; Suh, N. Calcitriol derivatives with two different side chains
at C-20. III. An epimeric pair of the gemini family with unprecedented
antiproliferative effects on tumor cells and renin mRNA expression
inhibition. J. Steroid Biochem. Mol. Biol. 2007, 103, 277−281.
(15) Maehr, H.; Lee, H. J.; Perry, B.; Suh, N.; Uskokovic, M. R.
Calcitriol derivatives with two different side chains at C20. V. Potent
inhibitors of mammary carcinogenesis and inducers of leukemia
differentiation. J. Med. Chem. 2009, 52, 5505−5519.
(16) Spina, C.; Tangpricha, V.; Min, Y.; Zhou, W.; Wolfe, M.; Maehr,
H.; Uskokovic, M.; Adorini, L.; Holick, M. F. Colon cancer and
ultraviolet B radiation and prevention and treatment of colon cancer in
mice with vitamin D and its gemini analogs. J. Steroid Biochem. Mol.
Biol. 2005, 97, 111−120.
(17) Reddy, G. S.; Tserng, K. Y. Calcitroic acid, end product of renal
metabolism of 1,25-dihydroxyvitamin D3 through C-24 oxidation
pathway. Biochemistry 1989, 28, 1763−1769.
Present Address
∥For H.J.L.: Department of Food Science and Technology,
Chung-Ang University, Anseong Si, Gyeonggi-Do, 456−756
South Korea.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Alastair McEwen and the staff of the beamline ID-14-
1 at ESRF for experimental assistance during data collection.
We are also indebted to Bioxell Inc for chemical and
Hoffmann-La Roche Inc, Nutley, NJ, for analytical support.
Cell growth inhibition studies were supported in part by NIH
R01 CA127645, NIEHS P30 ES005022, and by the Board of
Trustees Fellowship for Scholarly Excellence at Rutgers
University.
(18) Makin, G.; Lohnes, D.; Byford, V.; Ray, R.; Jones, G. Target cell
metabolism of 1,25-dihydroxyvitamin D3 to calcitroic acid. Evidence
for a pathway in kidney and bone involving 24-oxidation. Biochem. J.
1989, 262, 173−180.
(19) Maehr, H.; Uskokovic, M. R.; Adorini, L.; Reddy, S. D. Calcitriol
derivatives with two different side chains at C-20. II. Diastereoselective
syntheses of the metabolically produced 24(R) hydroxygemini. J. Med.
Chem. 2004, 47, 6478−6484.
REFERENCES
■
(1) Foster, A. B. Deuterium isotope effects in the metabolism of
drugs and xenobiotics: implications for drug design. Adv. Drug Res.
1985, 1, 2.
(2) Mutlib, A. E. Application of stable labeled compounds in
metabolism and in metabolism-mediated toxicity studies. Chem. Res.
Toxicol. 2008, 1672−1689.
J
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(20) Yamada, S.; Nakayama, K.; Takayama, H.; Shinki, T.; Takasaki,
Y.; Suda, T. Isolation, identification, and metabolism of (23S,25R)-
hydroxyvitamin D3-26,23-lactone. J. Biol. Chem. 1984, 259, 884−889.
(21) Horst, R. L.; Wovkulich, P. M.; Baggiolini, E. G.; Uskokovic, M.
R.; Engstrom, G. W.; Napoli, J. L. (23S)-1,23,25-trihydroxyvitamin D3:
Its biological activity and role in 1α,25-dihydroxyvitamin D3-26,23-
lactone biosynthesis. Biochemistry 1984, 23, 3973−3976.
(22) Bouillon, R. O., W. H.; Norman, A. W. Structure−function
relationships in the vitamin D endocrine system. Endocr. Rev. 1995, 16,
200−257.
synthesis of the A-ring phosphine oxide building block useful for
1α,25-dihydroxy vitamin D3 and analogues. J. Org. Chem. 2002, 67,
1580−1587.
(40) Courtney, L. F.; Lange, M.; Uskokovic, M. R.; Wovkulich, P. M.
Asymmetric synthesis of a key ring A synthon for 1α-hydroxy-19-nor
vitamin D. Tetrahedron Lett. 1998, 39, 3363−3366.
(41) Zuh, G.-D.; Okamura, W. H. Synthesis of vitamin D (calciferol).
Chem. Rev. 1995, 95, 1877−1952.
(42) Wade, D. Deuterium isotope effects on noncovalent interactions
between molecules. Chem.−Biol. Interact. 1999, 117, 191−217.
(23) Guryev, O.; Carvalho, R. A.; Usanov, S.; Gilep, A.; Estabrook, R.
W. A pathway for the metabolism of vitamin D3: unique hydroxylated
metaboles formed during catalysis with cytochrome P450scc
(CYP11A1). Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 13754−14759.
(24) Binderup, L.; Latini, S.; Binderup, E.; Bertting, C.; Calverley, M.;
Hansen, K. 20-Epivitamin D3 analogues: a novel class of potent
regulators of cell growth and immune responses. Biochem. Pharmacol.
1991, 42, 1569−1575.
(25) Dilworth, F. J.; Calverly, M. J.; Makin, H.; Jones, G. Increased
biological activity of 20-epi-1,25-dihydroxyvitamin D3 is due to
reduced catabolism and altered protein binding. Biochem. Pharmacol.
1994, 47, 987−993.
(26) Maenpaa, P. H.; Vaisanen, S.; Jaaskelainen, T.; Ryhanen, S.;
̈
̈
̈
̈
̈
̈
̈
̈
̈
Rouvinen, J.; Duchier, C.; Mahonen, A. Vitamin D3 analogs (MC
1288, KH 1060, EB 1089, GS 1558, and CB 1093): studies on their
mechanism of action. Steroids 2001, 66, 223−225.
(27) Rochel, N.; Wurtz, J. M.; 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.
(28) Ciesielski, F.; Rochel, N.; Mitschler, A.; Kouzmenko, A.; Moras,
D. Structural investigation of the ligand binding domain of the
zebrafish VDR in complex with 1a25(OH2D3 and gemini:
purification, crystallization and preliminary X-ray diffraction analysis.
J. Steroid Biochem. Mol. Biol. 2004, 89−90, 55−59.
(29) Ciesielski, F.; Rochel, N.; Moras, D. Adaptability of the vitamin
D nuclear receptor to the synthetic ligand gemini: Remodeling the
LPB with one side chain rotation. J. Steroid Biochem. Mol. Biol. 2007,
103, 235−242.
(30) Rochel, N.; Moras, D. Ligand binding domain of vitamin D
receptors. Curr. Top. Med. Chem. 2006, 6, 1229−41.
(31) Rochel, N.; Moras, D. Vitamin D Nuclear Receptor Ligand-
Binding Domain Crystal Structure. In Vitamin D, 2nd ed.; Feldman,
D., Pike, J. W, Glorieux, F. H., Eds.; Elsevier Academic Press:
Burlington, MA, 2005; pp 279−289.
(32) Huet, T.; Maehr, H.; Lee, H. J.; Uskokovic, M. R.; Suh, N.;
Moras, D.; Rochel, N. Structure−function study of gemini derivatives
with two different side chains at C-20, Gemini-0072 and Gemini-0097.
Med. Chem. Commun. 2011, 2, 424−429.
(33) Manchand, P. S.; Uskokovic M. R. Preparation of vitamin D3
analogs with bis C-20 side chains as therapeutic agents. Patent
WO9849138, 1998, 35 pp.
(34) Manchand P. S.; Uskokovic, M. R. Method of using vitamin D
analogs with bis C-20 side chains for treating or preventing neoplastic
diseases. U.S. Patent US 6008209, 1999, 13 pp.
(35) Maehr, H.; Uskokovic, M. R. Formal desymmetrization of the
diastereotopic chains in gemini calcitriol derivatives with two different
side chains at C-20. Eur. J. Org. Chem. 2004, 1703−1713.
(36) Manchand, P. S.; Yiannikouros, G. P.; Belica, P. S.; Madan, P.
Nickel-mediated conjugate addition. Elaboration of calcitriol from
ergocalciferol. J. Org. Chem. 1995, 60, 6574−6581.
(37) Pilcher, A. S; DeShong, P. Improved protocols for the selective
deprotection of trialkylsilyl ethers using fluorosilicic acid. J. Org. Chem.
1993, 58, 5130−5134.
(38) Baggiolin, E. G; Iacobelli, J. A.; Hennessy, B. M.; Batcho, A. D.;
Sereno, J. F.; Uskokovic, M. R. Stereocontrolled total synthesis of
1.alpha.,25-dihydroxycholecalciferol and 1.alpha.,25-dihydroxyergocal-
ciferol. J. Org. Chem. 1986, 51, 3098−3108.
(39) Daniewski, A. R.; Garofalo, L. M.; Hutchings, S. D.; Kabat, M.
M.; Liu, W.; Okabe, M.; Radinov, R.; Yannikouros, G. P. Efficient
K
dx.doi.org/10.1021/jm400032t | J. Med. Chem. XXXX, XXX, XXX−XXX