Novel synthesis of 2-substituted 19-norvitamin D A-ring phosphine oxide from D-glucose as a building block

Article abstract of DOI:10.1016/S0960-894X(03)00005-2

19-Norvitamin D A-ring phosphine oxide 5 was synthesized by a new sequence mode starting from D-glucose as a chiral template. Transformation of the pyranoside ring into the A-ring carbocycle was achieved by the Pd-catalyzed Ferrier rearrangement. The phosphine oxide 5 was obtained in an 18% overall yield by this novel cost-effective method.

Full text of DOI:10.1016/S0960-894X(03)00005-2

Bioorganic & Medicinal Chemistry Letters 13 (2003) 809–812
Novel Synthesis of 2-Substituted 19-Norvitamin D A-Ring
Phosphine Oxide from D-Glucose as a Building Block
Masato Shimizu,a,* Yukiko Iwasaki,a Yoshinori Shibamoto,a Miki Sato,a
H. F. DeLucab and Sachiko Yamadaa,*
aInstitute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, 2-3-10 Kandasurugadai, Chiyoda-ku,
Tokyo 101-0062, Japan
bDepartment of Biochemistry, University of Wisconsin-Madison, 433 Babcock Drive, Madison, WI 53706-1544, USA
Received 21November 2002; accepted 27 December 2002
Abstract—19-Norvitamin D A-ring phosphine oxide 5 was synthesized by a new sequence mode starting from d-glucose as a chiral
template. Transformation of the pyranoside ring into the A-ring carbocycle was achieved by the Pd-catalyzed Ferrier rearrange-
ment. The phosphine oxide 5 was obtained in an 18% overall yield by this novel cost-effective method.
# 2003 Elsevier Science Ltd. All rights reserved.
Active vitamin D3, 1a,25-dihydroxyvitamin D3 1 elicits
its biological functions such as calcium and phosphorous
homeostasis, cell differentiation, and immunoregulation
by binding to the vitamin D receptor (VDR) and reg-
ulating the transcription of target genes.1 Until recently,
attention has been focused mostly on the side-chain
derivatives of vitamin D, and we have developed a
structure–function theory for vitamin D focusing on the
side chain.2 Recently, however, interest in vitamin D
drug development has shifted to the A-ring.3 1a,25-
Dihydroxy-19-norvitamin D3 2, which lacks the 10(19)-
exomethylene group of 1, has reduced VDR affinity
(30% of 1) and shows a non-calcemic activity profile.4,5
Interestingly, introduction of a C(2)-substituent to 19-
norvitamin D analogues restores the VDR affinity as
well as the calcemic activity.6,7 Thus, 2-substituted 19-
norvitamin D analogues open a new horizon in the field
of synthetic vitamin D drugs.
from expensive (À)-quinic acid is both convenient and
versatile.5 We have envisioned a new method for synthe-
sizing the A-ring phosphine oxide 5, a key intermediate
for the 2-substituted 19-norvitamin D analogue. Here
we report the synthesis of 5 starting from low-cost
d-glucose as a chiral building block and its application to
novel 2-substituted 19-norvitamin D analogue synthesis.
Our retrosynthetic analysis of the A-ring synthon is
outlined in Figure 2: phosphine oxide 5 can be derived
from the cyclohexanone derivative 6, which is obtained
from 7. The key step of the synthesis, conversion of the
pyranoside ring to a carbocyclic system, can be achieved
by Ferrier rearrangement of the hex-5-enopyranoside 8,
which is readily available from the 2-deoxypyranoside
derivative 9. Ultimately, we envisioned d-glucose to be
a suitable chiral template for the 2-substituted 19-nor-
A-ring synthon 5.
The 19-norvitamin D skeleton is constructed using the
standard convergent method based on coupling of the
A-ring phosphine oxide with the C/D-ring 25-hydroxy
Grundmann’s ketone. Although at least three methods
for synthesis of the 19-nor-A-ring phosphine oxide have
been reported,5,8,9 for the synthesis of 2-substituted 19-
norvitamin D analogues (Fig. 1), the method starting
d-Glucose was easily transformed to an anomeric mixture
of 2-deoxypyranoside 10 (ca. a/b=10:1)10 by a 4-step, one-
column procedure with a 92% overall yield. Alkaline
hydrolysis of 11, followed by benzylideneacetalization
afforded the acetal 11 (89%). After protection of a
hydroxyl group at C(3) as a benzyl ether, the 4,6-O-
benzylidene acetal ring was cleaved by LiAlH4 in the
presence of AlCl3 to give the dibenzyl ether 12 (76%).
Iodination of the alcohol 12 gave the iodide, which
upon treatment with AgF (or DBU), afforded the hex-5-
enopyranoside 13 (78%). Among various methods used
*Corresponding author. Tel.: +81-352-808037; fax: +81-352-808005;
e-mail: shimizu.mr@tmd.ac.jp
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00005-2

810
M. Shimizu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 809–812
benzyl ether 16 (89%). Swern oxidation of 16, followed
by Pd/C-catalyzed hydrogenolysis of the benzyl ether
proceeded quantitatively to give the 3,4,5-triol, which
was regioselectively silylated with bulky t-butyldi-
methylsilyl chloride to give the cyclohexanone deriva-
tive 17 (72%) as a single product. Peterson olefination
of 17 and transformation to the final phosphine oxide 5
were accomplished by a method analogous to that
developed by DeLuca and co-workers.4 After protection
of the hydroxyl group of 17 with TMS-imidazole, reac-
tion with methyl (trimethylsilyl)acetate was carried out
to give an allylic ester quantitatively, which on reduc-
tion with (iso-Bu)2AlH afforded the corresponding
alcohol 18. This allylic alcohol was converted to the A-
ring phosphine oxide 5 via allylic tosylate and the
diphenylphosphine derivative with a 78% yield. The
overall yield of 5 from glucose was about 18% and bet-
ter than that via the (À)-quinic acid route (about 13%
Figure 1. Structures of the active vitamin D3 and its 19-norvitamin D
analogue.
Figure 2. Retrosynthetic analysis of the A-ring phosphine oxide.
for the rearrangement of pyranosides into carbocyclic
compounds, we investigated the Pd (II)-mediated Fer-
rier reaction.11 Under mild and neutral conditions (cat-
alytic amount of PdCl2 in an aqueous dioxane, 60 ꢀC),
13 was converted to the hydroxy ketone derivative,
which was silylated to yield the t-butyldimethylsilyl
ether 14 as a separable diastereoisomeric mixture (ca.
1a-isomer/1b-isomer=6:1, 89%). The reduction of the
major 1a-isomer of 14 with l-Selectride1 proceeded
stereoselectively with an excellent yield (95%) to give
the 5a-hydroxy isomer 15 as a sole product. A similar
result was obtained using the 1b-isomer of 14 to give the
corresponding 5a-hydroxy compound. The stereo-
chemistry at C(1) has little effect on the stereoselectivity
of the reduction of the 5-ketone 14, probably because
the equatorially oriented 4a-benzyloxy group hinders
reduction from the a side. Protection of the hydroxyl
group and subsequent desilylation afforded the tri-O-
Figure 3. Synthesis of the A-ring phosphine oxide from d-glucose: (a)
(1) Ac2O, HClO4, then HBr; (2) Zn-aq AcOH; (3) MeOH, pTsOH
(92%); (b) (i) NaOMe, MeOH; (ii) PhCH(OMe)2, pTsOH (89%); (c)
BnBr, DMF (90%); (d) LiAlH4, AlCl3, Et2O (84%); (e) I2, PPh3,
THF (82%); (f) AgF, Py. (95%); (g) PdCl2, dioxane, H2O (92%); (h)
t-BuSi(Me)2Cl, DMF (97%); (i) l-Selectride, THF (95%); (j) BnBr,
DMF (94%); (k) Bu4NF, THF (95%); (l) DMSO, (COCl)2, CH2Cl2
(98%); (m) (1) H2/Pd/C, EtOH, (2) t-BuSi(Me)2Cl, DMF (73%); (n)
TMS-imidazole, CH2Cl2 (99%); (o) TMS–CH2CO2Me, BuLi, THF
(99%); (p) (iso-Bu)2AlH, Tol (98%); (q) (1) TsCl, BuLi, THF, then
Ph2PH; (2) 10% H2O2, CH2Cl2 (78%).

M. Shimizu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 809–812
811
overall yield). It should be noted that compounds 11
and 14 serve as useful intermediates for preparing other
modified A-ring derivatives, expanding the method for
synthesis of diverse 19-norvitamin D analogue: for
example, deoxygenation of the C(3)-hydroxyl group of
11 can afford the A-ring having 1,2- or 2,3-vicinal
hydroxyl groups, and the A-ring synthon with 1,3,4- or
1,3,10-trihydroxyl groups can be synthesized directly
from the Ferrier-ketone 14 (Figs. 3 and 4).
2b-isomers.3,7,13 In the present case of 2-hydroxyethoxy-
19-norvitamin D derivatives 3b and 4b, the situation is
reversed. The natural ligand 1 is harbored in the VDR
adopting the b-conformation at the A-ring, where the
1a-hydroxyl group takes an equatorial conformation
and the 10(19)-exocyclic methylene group is oriented
towards the b-face.14 In a preliminary docking study of
compounds 3b and 4b, we assume that both can be
accommodated in the ligand binding pocket with the
b-form, Arg274 and Asp144 playing a key role in
anchoring the ligands at the C(2) substituent. However,
more detailed studies such as site directed mutation
analyses are necessary to identify the key interactions
responsible for their potency differences. These studies
are under investigation and will be reported elsewhere.
We synthesized six 2-substituted 19-norvitamin D ana-
logues 3a,b,c and 4a,b,c using the phosphine oxide 5.15
We obtained 1a,2a,25- and 1a,2b,25-trihydroxy-19-
norvitamin D 3a and 4a,6 by reaction with 25-hydroxy
Grundmann’s ketone 19.12 2-Hydroxyethoxy- (3b, 4b)
and 2-diethylcarbamoylmethoxy-19-norvitamin D ana-
logues (3c, 4c) were obtained from 2-hydroxy deriva-
tives 20 and 21 by the Williamson ether synthesis.
In conclusion, we have accomplished a new and
easy method for efficient synthesis of the 2-substituted
19-norvitamin D A-ring building block starting from
d-glucose. We synthesized six 2-substituted 19-norvita-
min D derivatives using this A-ring synthon and their
VDR affinity was evaluated. We are continuing the
synthesis of a variety of 2-substituted 19-norvitamin D
analogues using this methodology and a full account
will be published in due course.
The binding affinity of 3 and 4 for the bovine thymus
VDR was evaluated. It should be noted that the 2b-
hydroxyethoxy analogue 4b has equivalent affinity with
the natural ligand 1, while the 2a-isomer 3b and the
carbamoyl derivative 3c were 10-fold and 2000-fold less
potent than 1, respectively. It has been reported that
among the 2-substituted vitamin D analogues, 2a-iso-
mers have higher VDR affinity than the corresponding
References and Notes
1. Feldman, D., Glorrieux, F. H., Pike, J. W. Eds., Vitamin D;
Academic; New York, 1997.
2. (a) Yamada, S.; Yamamoto, K.; Masuno, H.; Ohta, M. J.
Med. Chem. 1998, 41, 1467. (b) Yamada, S.; Yamamoto, K.;
Masuno, H.; Choi, M. Steroids 2001, 66, 177.
3. Fujishima, T.; Konno, K.; Nakagawa, K.; Tanaka, M.;
Okano, T.; Kurihara, M.; Miyata, N.; Takayama, H. Chem.
Biol. 2001, 8, 1011.
4. Perlman, K. L.; Sicinski, R.; Schnoes, H. K.; DeLuca, H. F.
Tetrahedron Lett. 1990, 31, 1823.
5. Perlman, K. L.; Swenson, R. E.; Paaren, H. E.; Schnoes,
H. K.; DeLuca, H. F. Tetrahedron Lett. 1991, 32, 7663.
6. Sicinski, R. R.; Perlman, K. L.; DeLuca, H. F. J. Med.
Chem. 1994, 37, 3730.
7. (a) Sicinski, R. R.; Prahl, J. M.; Smith, C. M.; DeLuca,
H. F. J. Med. Chem. 1998, 41, 4662. (b) Sicinski, R. R.; Rot-
kiewicz, P.; Kolinski, A.; Sicinska, W.; Prahl, J. M.; Smith,
C. M.; DeLuca, H. F. J. Med. Chem. 2002, 45, 3366.
8. (a) Mikami, K.; Osawa, A.; Isaka, A.; Eiji, S.; Shimizu, M.;
Terada, M.; Kubodera, N.; Nakagawa, K.; Tsugawa, N.;
Okano, T. Tetrahedron Lett. 1998, 39, 3359. (b) Courtney,
L. F.; Lange, M.; Uskokovic, M. R.; Wovkulich, P. M.
Tetrahedron Lett. 1998, 39, 3363.
9. Hilpert, H.; Wirz, B. Tetrahedron 2001, 57, 681.
10. (a) Kozikowski, A. P.; Lee, J. J. Org. Chem. 1990, 55, 863.
(b) Inglis, G. R.; Schwarz, J. C. P.; McLaren, L. J. Chem. Soc.
1962, 1014.
11. (a) Ferrier, R. J. J. Chem. Soc., Perkin Trans. 1 1979,
1455. (b) Adam, S. Tetrahedron Lett. 1988, 29, 6589. (c)
Iimori, T.; Takahashi, H.; Ikegami, S. Tetrahedron Lett. 1996,
37, 649.
12. Kiegiel, J.; Wovkulich, P. M.; Uskokovic, M. R. Tetra-
hedron Lett. 1991, 32, 6057.
13. (a) Tsugawa, N.; Nakagawa, K.; Kurobe, M.; Ono, Y.;
Kubodera, N.; Ozono, K.; Okano, T. Biol. Pharm. Bull. 2000,
23, 66. (b) Suhara, Y.; Nihei, K.; Tanigawa, H.; Fujishima, T.;
Konno, K.; Nakagawa, K.; Okano, T.; Takayama, H. Bioorg.
Figure 4. Synthesis of 2-substituted 19-norvitamin D analogues: (a)
PhLi, THF (51%); (b) aq AcOH, THF (82%); (c) BrCH2CH2OTBS,
NaH, DMF, THF (41% for 3b, 88% for 4b); (d) BrCH2CONEt2,
NaH, DMF, THF (67% for 3c, 67% for 4c); (e) (À)-camphorsulfonic
acid, MeOH (42% for 3b, 51% for 4b, 84% for 3c, 67% for 4c).

812
M. Shimizu et al. / Bioorg. Med. Chem. Lett. 13 (2003) 809–812
Med. Chem. Lett. 2000, 10, 1129. (c) Kittaka, A.; Suhara, Y.;
Takayanagi, H.; Fujishima, T.; Kurihara, H.; Takayama, H.
Org. Lett. 2000, 2, 2619.
14. Rochel, N.; Wurtz, J. M.; Mitschler, A.; Klaholz, B.;
Moras, D. Mol. Cell 2000, 5, 173.
15. Satisfactory spectral characterization of all intermediates
was obtained. Spectral data of 3a, 4a, 18 and 5 were in accor-
dance with those reported in literature6). Data for 3b: 1H
NMR (400 MHz, CDCl3) d: 0.55 (3H, s, 18-H), 0.93 (3H, d,
J=6.4 Hz, 21-H), 1.22 (6H, s, 26, 27-H), 3.37 (1H, dd, J=7.8,
3.0 Hz, 2-H), 3.71–3.83 (4H, m, OCH2CH2O), 3.96, 4.14 (each
1H, m, 3, 1-H), 5.84, 6.34 (each 1H, J=11.2 Hz, 7, 6-H). UV
lmax (MeOH): 242, 251, 261 nm. MS m/z (%): 464 (M+, 61),
446 (100), 428 (63). HR-EI–MS m/z: 464.3505 (calcd for
C28H48O5: 464.3502). 4b: d: 0.54 (3H, s, 18-H), 0.94 (3H, d,
J=6.4 Hz, 21-H), 1.22 (6H, s, 26, 27-H), 3.32 (1H, dd, J=8.7,
2.9 Hz, 2-H), 3.65–3.89 (4H, m, OCH2CH2O), 3.84, 4.17 (each
1H, m, 3, 1-H), 5.84, 6.29 (each 1H, J=11.2 Hz, 7, 6-H). MS
m/z (%): 464 (M+, 63), 446 (100), 428 (58). 3c: d: 0.55 (3H, s,
18-H), 0.93 (3H, d, J=6.4 Hz, 21-H), 1.14, 1.19 (each 3H, t,
J=7.1 Hz), 1.22 (6H, s, 26, 27-H), 3.19, 3.41 (each 1H, q,
J=7.1Hz), 3.32 (1H, dd, J=7.5, 1.9 Hz, 2-H), 4.01 (3H, m, 3,
1-H, OH), 4.30, 4.39 (each 1H, d, J=15.8 Hz), 5.85, 6.33 (each
1H, J=11.2 Hz, 7, 6-H). UV lmax (EtOH): 243, 251, 261 nm.
MS m/z (%): no M+, 497 (14), 479 (12), 383 (32), 365 (100). 4c: d:
0.54 (3H, s, 18-H), 0.93 (3H, d, J=6.4 Hz, 21-H), 1.15, 1.19 (each
3H, t, J=7.1Hz), 1.22 (6H, s, 26, 27-H), 3.27 (1H, dd, J=8.7,
2.8 Hz, 2-H), 3.83, 4.05 (each H, m, 3, 1-H), 4.32, 4.36 (each 1H,
d, J=15.9 Hz), 5.87, 6.29 (each 1H, J=11.2 Hz, 7, 6-H). MS m/z
(%): no M+, 497 (12), 479 (12), 383 (8), 365 (100).