Design and efficient synthesis of new stable 1alpha,25-dihydroxy-19-norvitamin D3 analogues containing amide bond.

Article abstract of DOI:10.1016/S0960-894X(02)00800-4

The design and synthesis of new 1alpha,25-dihydroxy-19-norvitamin D(3) analogues 3a-c, which have an amide bond in the molecule instead of the diene, are described. The A-ring moiety was constructed by a (3S,5S)-3,5-dihydroxypiperidine derivative (9, 11,

Full text of DOI:10.1016/S0960-894X(02)00800-4

Bioorganic & Medicinal Chemistry Letters 12 (2002) 3533–3536
Design and Efficient Synthesis of New Stable 1ꢀ,25-Dihydroxy-
19-norvitamin D3 Analogues Containing Amide Bond
Yoshitomo Suhara,a,y Atsushi Kittaka,a Keiichiro Ono,a Masaaki Kurihara,b
Toshie Fujishima,a Akihiro Yoshidaa and Hiroaki Takayamaa,*
aFaculty of Pharmaceutical Sciences, Teikyo University, Sagamiko, Kanagawa 199-0195, Japan
bNational Institute of Health Sciences, Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan
Received 5 August 2002; accepted 17 September 2002
Abstract—The design and synthesis of new 1a,25-dihydroxy-19-norvitamin D3 analogues 3a–c, which have an amide bond in the
molecule instead of the diene, are described. The A-ringmoiety was constructed by a (3 S,5S)-3,5-dihydroxypiperidine derivative (9,
11, or 13) prepared from d-mannose, and a CD-ringcarboxylic acid 16 was synthesized from Grundmann’s ketone. Couplingthose
parts gave desired 3a–c in good yield. This strategy can be applied in combinatorial chemistry; therefore, those compounds would
be applicable as useful tools in the development of new drugs.
# 2002 Elsevier Science Ltd. All rights reserved.
1a,25-Dihydroxyvitamin D3 (1) is the major active
metabolite of vitamin D3, which exerts hormonal con-
trol over important physiological processes, including
calcium and phosphorus metabolism, cellular differ-
entiation and immune reactions.1 A number of vitamin
D analogues have been synthesized in order to investi-
gate the biological roles of vitamin D and to develop
potential therapeutic agents.2 Since effective doses of 1
may cause hypercalcemia, the therapeutic use of 1 in the
treatment of certain cancers and skin disorders is lim-
ited. Dissociation of the calcemic activity from the other
activities of 1 is one of the important issues in searching
for novel drugs based on vitamin D analogues. DeLuca
et al. synthesized a 19-nor derivative 2, which is a non-
calcemic analogue of 1 that still preserves its cell differ-
entiatingactivities ( Fig. 1).3
(1a),4 2a-hydroxypropyl (1b),5 and 2a-hydroxypropoxyl
(1c)6 analogues showed higher potency than the natural
hormone 1 in terms of bovine thymus vitamin D recep-
tor (VDR) bindingaffinity, elevation of rat serum cal-
cium concentration, and induction of HL-60 cell
differentiation (Fig. 1). For example, compound 1b
exhibited 3-fold higher VDR binding affinity and more
than double potency in inducingdifferentiation of
HL-60 cells as compared to 1.5a
Previously, we synthesized several 1a,25-dihydroxy-
vitamin D3 derivatives, which systematically introduced
a 2a-alkyl, 2a-hydroxyalkyl, or 2a-hydroxyalkoxyl
group into the parent hormone 1, in order to investigate
the A-ringconformation– and structure–activity
relationships.4ꢀ6 Some of these analogues exhibited
interestingbiological activities, in particular, 2 a-methyl
*Correspondingauthor. Tel.: +81-426-85-3713; fax: +81-426-85-
3714; e-mail: hi-takay@pharm.teikyo-u.ac.jp
yPresent address: Department of Hygienic Sciences, Kobe Pharma-
ceutical University, Kobe 658–8558, Japan.
Figure 1. Structures of the natural hormone 1, and its 2a-substituted
analogues 1a–c with higher binding affinity for VDR than that of 1,
DeLuca’s 19-nor derivative 2, and amide derivatives 3a–c.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00800-4

3534
Y. Suhara et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3533–3536
Based on these results, we then focused on modification
of the diene moiety in the 19-nor skeleton of the B-seco
steroid compound.7 We anticipated that if the diene was
replaced by an amide bond, the resultinganalouges
would retain the original biological activities of 19-nor-
vitamin D3 (2), since they would possibly take on a
similar ‘flat’ conformation.
Our designed analogues are shown in Figure 1. The
A-ringmoiety was substituted with a piperidine ring
havingtwo hydroxyls at C1 a and C3b (seco-steroidal
numbering) with the appropriate stereochemistry, which
play a crucial role in VDR bindingand bioloigcal
actions.8 The diene part was converted to an amide
bond to connect with the normal CD-ringmoiety. We
also planned modification at the C2 position with the
substituent of the 3-hydroxypropoxyl group, because
this motif can strengthen binding affinity to VDR.5,9
The amide structures 3a–c would allow the N–C6 bond
to rotate, whose original structure of the C5–C6 bond is
fixed in the natural hormone 1. The relative positions of
the two hydroxyls at C1a and C3b toward the VDR
ligand binding domain (LBD) are retained with the
rotation, and the artificial C2 substituent may settle
down in an a or a b-orientation in preference when 3b
or 3c complexes with LBD. In Figure 2, the X-ray crys-
Scheme 1. Reagents and conditions: (a) (i) anhydrous CuSO4, ace-
tone, concd H2SO4, (ii) BzCl, pyridine (86% in two steps); (b) 70%
AcOH aq (quant); (c) (i) NaIO4, Et2O, H2O, (ii) NaBH4, MeOH (64%
in two steps); (d) TsCl, pyridine (75%); (e) NaN3, DMF (72%); (f) 1 N
NaOH aq, MeOH (97%); (g) (i) H2, cat. Pd(OH)2, MeOH, (ii) CbzCl,
Et2O, sat. NaHCO3 aq (87% in two steps); (h) 80% AcOH aq, 90 ꢁC
(92%); (i) TBSCl, Et3N, DMF (87%); (j) H2, cat. Pd/C, MeOH (71%).
duced d-mannofuranoside 4. Oxidative diol cleavage of
4 with sodium periodate, and reduction of the resultant
aldehyde afforded the primary alcohol. Azide 5 was
prepared from the alcohol through tosylate in good
yield. After removal of the benzoyl group of 5, reductive
amination with H2/cat. Pd(OH)2 in methanol followed
by protection of the amino group with Cbz gave the
piperidine derivative 6 in 87% yield. Deprotection of
the acetonide of 6, then introduction of the TBS group
to the C1 and C3 hydroxyl groups afforded 8 in 92 and
87% yields, respectively. Finally, the Cbz group was
hydrogenated with cat. Pd/C to give the desired A-ring
part 9 for 3b in 71% yield.
10
tal structure of 1 (red) dockingin LBD of VDR and
the modeled structure of the amide D3 3a (blue) in the
same cavity of the VDR are shown. The latter structure
constructed by molecular dynamics calculations
(AMBER* force field) usingMacroModel ver. 6.5, and
3a could interact with the VDR through six hydrogen
bonds (dotted lines) from the C1a, C3b, and
C25-hydroxyls as in 1. This modelingalso encouraged
us to synthesize the amide D3 analogues.
We synthesized two additional piperidine derivatives 11
and 13 from 8 as shown in Scheme 2. The A-ringpor-
tion 11 was synthesized by Williamson ether synthesis
using(3-bromopropoxy)- tert-butyldimethylsilane and
sodium hydride, followed by hydrogenolysis of the Cbz
group. On the other hand, the natural type of A-ring
portion 13 was obtained by radical reduction of thio-
carbonate 12 with tributyltin hydride and AIBN in
good yield.
Scheme 1 shows the synthesis of the A-ringportions 8
(an intermediate) and 9. Treatment of d-mannose with
anhydrous CuSO4 and a catalytic amount of concd
H2SO4 in acetone followed by benzoylation of the
anomeric hydroxyl group gave a d-mannofuranose
derivative in 86% yield. Selective deprotection of the
O4–O6 acetonide using70% aqueous acetic acid pro-
The CD-ringportion was prepared from Grundmann’s
ketone 1411 as shown in Scheme 3. After protectingthe
C25 hydroxyl group with MOM, Horner–Wadsworth–
Emmons reaction usingtriethyl phosphonoacetate and
sodium hydride in THF gave the ester 15 in an excellent
Figure 2. Crystal structure of VDR bound to 1 (red) by Moras et al.10
and computer modelingof 3a (blue) in the VDR ligand binding
domain.
Scheme 2. Reagents and conditions: (a) TBSO(CH2)3Br, NaH, DMF
(20%); (b) H2, Pd/C, MeOH (quant); (c) PhOC(S)Cl, DMAP, CH3CN
(36%); (d) Bu3SnH, AIBN, benzene (95%).

Y. Suhara et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3533–3536
3535
superimposed on the modeled structure of 3a (yellow)
bindingto the VDR. Conformation of the A-ringis
quite different from each other. However, the X-ray
structure of 1 (yellow) in the VDR and the energy
minimized structure of 1 (green) are shown in Figure 3b,
in which the latter modeled structure is very similar to the
X-ray structure in VDR. It may be difficult for VDR to
recognize the amide analogue 3a as a ligand in a solution.
In summary, we have developed an efficient systematic
route for synthesizingnew stable amide analogues 3a–c
utilizing d-mannose. Takahashi et al. recently reported
the synthesis of 1a,25-dihydroxyvitamin D3 derivatives
by the combinatorial procedure.17 Our amide analogues
can also be applied to this synthetic method and would
easily provide many kinds of this class of D3 analogues.
We consider that these syntheses and analogues would
contribute to the elucidation of vitamin D3 action
mechanisms with medicinal actions. Further biological
studies of the amide analogues are currently in progress
in our laboratory.
Scheme 3. Reagents and conditions: (a) MOMCl, DIEA, CH2Cl2
(quant); (b) triethyl phosphonoacetate, NaH, THF (98%); (c) 1 N
NaOH, MeOH (86%); (d) 9, 11, or 13, BOP, DIEA, DMF (68–88%);
(e) p-TsOH, MeOH (40–85%).
yield.12 The ethyl ester of 15 was subjected to hydrolysis
by 1 N aqueous NaOH in a MeOH solution to afford
the desired carboxylic acid 16 in 86% yield. Subsequent
Acknowledgements
condensation of 16 with the piperidine derivatives 9, 11,
13
We are grateful to Miss Junko Shimode and Miss Mar-
oka Kitsukawa for the spectroscopic measurements.
This work was supported in part by Kanagawa Acad-
emy of Science and Technology Research Grants, which
we gratefully acknowledge.
and 13 usinga BOP reagent
and DIEA provided the
amides 17a–c in good yields, respectively. Finally,
deprotection under acidic conditions furnished the tar-
get amide analogues 3a–c14 in considerable yield.
We tested VDR bindingaffinity of the amide analogues
15
3a–c usingbovine thymus VDR
though, unfortu-
References and Notes
nately, we could not recognize the affinity for the VDR
(3ꢂ10ꢀ6, 2ꢂ10ꢀ6, and 3ꢂ10ꢀ6 times that of 1 for 3a, 3b,
and 3c, respectively).16
1. (a) Holick, M. F.; Schnoes, H. K.; DeLuca, H. K.; Suda,
T.; Cousins, R. J. Biochemistry 1971, 10, 2799. (b) Abe, E.;
Ishimi, Y.; Takahashi, N.; Akatsu, T.; Ozawa, H.; Yamana,
H.; Yoshiki, S.; Suda, T. Proc. Natl. Acad. Sci. U.S.A. 1981,
78, 4990. (c) Walters, M. R. Endocr. Rev. 1992, 41, 231. (d)
Christakos, S.; Raval-Pandya, M.; Wernyj, R. P.; Yang, W.
Biochem. J. 1996, 316, 361.
2. (a) Kobat, M. K.; Burger, W.; Guggino, S.; Hennessy, B.;
Iacobelli, J. A.; Takeuchi, K.; Uskokovic, M. R. Bioorg. Med.
Chem. 1998, 6, 2051. (b) Sicinski, R. F.; DeLuca, H. F. Bioorg.
Med. Chem. 1999, 7, 2877. (c) Okano, T.; Nakagawa, K.;
Kubodera, N.; Ozono, K.; Isaka, A.; Osawa, A.; Terada, M.;
Mikami, K. Chem. Biol. 2000, 7, 173. (d) Gao, L.-J.; Zhao,
X.-Y.; Vandewalle, M.; Clercq, P. D. Eur. J. Org. Chem. 2000,
2755.
Two reasons were considered. One would be the elec-
trostatic difference between the amide moiety and the
diene part. The polar amide bond in 3a–c could not be
well suited for the hydrophobic region surrounded by
amino acid residues Phe-150, Leu-233, Tyr-295, and
Trp-286 of VDR, which would originally accept the
diene part of 2. The other would be the difference in steric
energy (9.8 kcal/mol; ab initio MO method, HF/6–31G*)
between the active structure in the VDR and the energy
minimized structure. Figure 3a shows the energy mini-
mized structure of 3a (green) lacking the side chain, as
calculated by Monte Carlo conformational search and
3. (a) Perlman, K. L.; Sicinski, R. R.; Schnoes, H. K.;
DeLuca, H. F. Tetrahedron Lett. 1990, 31, 1823. (b) Perlman,
K. L.; Swenson, R. E.; Paaren, H. E.; Schnoes, H. K.;
DeLuca, H. F. Tetrahedron Lett. 1991, 32, 7663.
4. (a) Konno, K.; Maki, S.; Fujishima, T.; Liu, Z.-P.; Miura,
D.; Chokki, M.; Takayama, H. Bioorg. Med. Chem. Lett.
1998, 8, 151. (b) Konno, K.; Fujishima, T.; Maki, S.; Liu,
Z.-P.; Miura, D.; Chokki, M.; Ishizuka, S.; Yamaguchi, K.;
Kan, Y.; Kurihara, M.; Miyata, N.; Smith, C; DeLuca, F. H.;
Takayama, H. J. Med. Chem. 2000, 43, 4247.
5. (a) Suhara, Y.; Nihei, K.; Tanigawa, H.; Fujishima, T.;
Konno, K.; Nakagawa, K.; Okano, T.; Takayama, H. Bioorg.
Med. Chem. Lett. 2000, 10, 1129. (b) Suhara, Y.; Nihei, K.;
Kurihara, M.; Kittaka, A.; Yamaguchi, K.; Fujishima, T.;
Konno, K.; Miyata, N.; Takayama, H. J. Org. Chem. 2001,
66, 8760.
Figure 3. (a) The energy minimized structure of 3a (green) and the
modeled structure of 3a (yellow) dockingin VDR; (b) the X-ray
structure of 1 (yellow) in VDR10 and the energy minimized structure
of 1 (green).

3536
Y. Suhara et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3533–3536
6. Kittaka, A.; Suhara, Y.; Takayanagi, H.; Fujishima, T.;
Kurihara, M.; Takayama, H. Org. Lett. 2000, 2, 2619.
7. Trials to modify the triene moiety in the D3 structure to
induce the differentiation of HL-60 cells were interesting, but
not very productive to date. For singlet oxygen adducts: (a)
Yamada, S.; Yamamoto, K.; Naito, H.; Suzuki, T.; Ohmori,
M.; Takayama, H.; Shiina, Y.; Miyaura, C.; Tanaka, H.; Abe,
E.; Suda, T.; Matsunaga, I.; Nishii, Y. J. Med. Chem. 1985,
28, 1148. (b) Shiina, Y.; Miyaura, C.; Tanaka, H.; Abe, E.;
Yamada, S.; Yamamoto, K.; Ino, E.; Takayama, H.; Matsu-
naga, I.; Nishi, Y.; Suda, T. J. Med. Chem. 1985, 28, 1153.
8. All A-ringdiastereomers of 1, see: Fujishima, T.; Konno,
K.; Nakagawa, K.; Kurobe, M.; Okano, T.; Takayama, H.
Bioorg. Med. Chem. 2000, 8, 123.
9. Examples of 2-substituted 1a,25-dihydroxy-19-norvitamin
D3 derivatives, see: (a) Sicinski, R. R.; Perlman, K. L.;
DeLuca, H. F. J. Med. Chem. 1994, 37, 3730. (b) Sicinski,
R. R.; Prahl, J. M.; Smith, C. M.; DeLuca, H. F. J. Med.
Chem. 1998, 41, 4662. (c) Sicinski, R. R.; Prahl, J. M.; Smith,
C. M.; DeLuca, H. F. Steroids 2002, 67, 247. (d) Mikami, K.;
Koizumi, Y.; Osawa, A.; Terada, M.; Takayama, H.; Naka-
gawa, K.; Okano, T. Synlett 1999, 1899.
13. Anisfeld, S. T.; Lansbury, P., Jr. J. Org. Chem. 1990, 55,
5560.
14. Data for 3a: 1H NMR (600 MHz, CD3OD) d 0.66 (3H, s),
0.97 (1H, d, J=6.6 Hz), 1.05–1.10 (1H, m), 1.16–1.18 (7H, m),
1.23–1.29 (2H, m), 1.31–1.48 (9H, m), 1.54–1.59 (2H, m),
1.62–1.67 (2H, m), 1.73–1.80 (2H, m), 1.85–1.89 (1H, m),
1.92–1.98 (1H, m), 2.04 (1H, dt, J=3.0, 12.9 Hz), 2.04–2.10
(1H, m), 2.72 (1H, dt, J=2.9, 13.6 Hz), 3.27 (1H, dd, J=6.9,
12.9 Hz), 3.38 (1H, dd, J=6.0, 13.2 Hz), 3.53 (1H, dd, J=3.0,
13.2 Hz), 3.84 (1H, dd, J=3.3, 12.9 Hz), 3.95–4.00 (2H, m),
5.61 (1H, s). HRMS (EI) m/z: calcd for C25H43O4N 421.3192.
Found 421.3181 (M+). Data for 3b: 1H NMR (600 MHz,
CD2Cl2, selected) d 0.61 (3H, s), 0.95 (1H, d, J=6.6 Hz), 1.07
(1H, t, J=8.1 Hz), 1.92 (1H, m), 2.03–2.06 (2H, m), 2.74–2.83
(2H, m), 2.95–3.02 (1H, m), 3.23 (1H, br d, J=13.2 Hz), 3.50
(1H, br d, J=6.0 Hz), 3.57 (1H, br d, J=6.0 Hz), 3.78 (1H, br
s), 3.85 (0.6H, br d, J=13.2 Hz), 3.94–3.97 (1H, m), 4.04
(0.4H, br s), 4.34 (1H, br d, J=10.8 Hz), 5.58 (1H, s). HRMS
(EI) m/z: calcd for C25H43O5N 437.3144. Found 437.3148
1
(M+). Data for 3c: H NMR (600 MHz, CD2Cl2, selected) d
0.61 (3H, s), 0.95 (1H, d, J=6.6 Hz), 3.63–3.67 (4H,
m,-OCH2CH2CH2OH), 5.56 (1H, s). HRMS (EI) m/z: calcd
for C28H49O6N 495.3560. Found 495.3558 (M+).
15. Imae, Y.; Manaka, A.; Yoshida, N.; Ishimi, Y.; Shinki, T.;
Abe, E.; Suda, T.; Konno, K.; Takayama, H.; Yamada, S.
Biochim. Biophys. Acta 1994, 1213, 302.
10. Rochel, N.; Wultz, J. M.; Mitschler, A.; Klaholz, B.;
Moras, D. Mol. Cell 2000, 5, 173.
11. Trost, B. M.; Dumas, J.; Villa, M. J. Am. Chem. Soc.
1992, 114, 9836.
12. The E/Z ratio of this reaction was 92:8 with 98% chemical
yield. Chemical shifts of the vinylic proton of the E- and Z-
isomers are 5.45 and 5.65 ppm, respectively. Trost et al.
reported that stereochemistry on bromoolefins of the corre-
spondingCD-ringpart in ref 11, in which the vinylic protone
of the E-isomer (d 5.63 ppm) appears in a higher field than
that of the Z-isomer (5.93 ppm) due to the anisotropy of the
C–C single bond of the five-membered D-ring.
16. The bindingaffinity of 2 for VDR (pig) was reported as
30% of the natural hormone 1, see: Zhou, X.; Zhu, G.-D.;
Van Haver, D.; Vandewalle, M.; De Clercq, P. J.; Verstuyf,
A.; Bouillon, R. J. Med. Chem. 1999, 42, 3539.
17. (a) Doi, T.; Hijikuro, I.; Takahashi, T. J. Am. Chem. Soc.
2001, 123, 3716. (b) Hijikuro, I.; Doi, T.; Takahashi, T. In
Symposium Papers, The 43rd Symposium on the Chemistry of
Natural Products, Osaka, 2001; p 79.