A concise synthesis of (3S,4S,5R)-1-(α-d-galactopyranosyl)-3- tetracosanoylamino-4,5-decanediol, a C-glycoside analogue of immunomodulating α-galactosylceramide OCH

Article abstract of DOI:10.1016/j.tetlet.2005.05.069

A concise and convergent synthesis of the C-glycoside analogue 2b of immunomodulating α-galactosylceramide OCH 1b starting from readily available 2,3,4,6-tetra-O-benzyl-d-galactose 3 and l-arabinose 6 is described. The synthesis features the nucleophilic addition of an α-ethynyl sugar 5 to the phytosphingosine-precursor aldehyde 9 and would be applicable to a variety of C-glycoside analogues of interest.

Full text of DOI:10.1016/j.tetlet.2005.05.069

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 5043–5047
A concise synthesis of (3S,4S,5R)-1-
(a-D-galactopyranosyl)-3-tetracosanoylamino-4,5-decanediol,
a C-glycoside analogue of immunomodulating
a-galactosylceramide OCH
Tetsuya Toba,a Kenji Murata,a Takashi Yamamura,b
Sachiko Miyakeb and Hirokazu Annouraa,*
aBiomedical Research Laboratories, Daiichi Suntory Pharma Co., Ltd., 1-1-1, Wakayamadai, Shimamoto-cho, Mishima-gun,
Osaka 618-8513, Japan
bDepartment of Immunology, National Institute of Neuroscience, NCNP, Tokyo 187-8502, Japan
Received 1 February 2005; revised 13 May 2005; accepted 18 May 2005
Available online 9 June 2005
Abstract—A concise and convergent synthesis of the C-glycoside analogue 2b of immunomodulating a-galactosylceramide OCH 1b
starting from readily available 2,3,4,6-tetra-O-benzyl-D-galactose 3 and L-arabinose 6 is described. The synthesis features the
nucleophilic addition of an a-ethynyl sugar 5 to the phytosphingosine-precursor aldehyde 9 and would be applicable to a variety
of C-glycoside analogues of interest.
Ó 2005 Elsevier Ltd. All rights reserved.
Natural killer (NK) T cells are potent producers of
immunoregulatory cytokines and specific for glycolipid
antigens bound by a nonclassical major histocompatibil-
ity complex (MHC) class I-like molecule, CD1d.1 The
glycolipids, an a-galactosylceramide named KRN7000
1a2 and an altered analogue termed OCH 1b possessing
a shorter C5 sphingosine side chain,3 have been identi-
fied as NKT cell ligands (Fig. 1). Compound 1b was
shown to induce a predominant production of interleu-
kin (IL)-4, a key Th2cytokine engaged in autoimmunity
control, over Th1 cytokine interferon (IFN)-c, while 1a
induced both Th1/Th2cytokines. Only compound 1b
but not 1a is significantly effective in animal models of
Th1-mediated autoimmune diseases such as experimen-
tal autoimmune encephalomyelitis (EAE) and collagen
induced arthritis (CIA).3,4 Quite recently, we have re-
OH
HO
HO
O
O
HN
HO
R1
HO
A
OH
R2
1a (KRN7000; A = O, R1 = n-C25H51, R2 = n-C14H29
1b (OCH; A = O, R1 = n-C23H47, R2 = n-C5H11
)
)
(A = CH2, R1 = n-C25H51, R2 = n-C14H29
)
2a
2b
(A = CH2, R1 = n-C23H47, R2 = n-C5H11
)
Figure 1. Structure of a-galactosylceramides 1a,b and their C-glyco-
side analogues 2a,b.
ported a practical and efficient synthesis of 1b in 12steps
and 19% overall yield from commercially available
D-arabitol.5 It is often a conventional strategy to synthe-
size the C-glycoside analogue of biologically active
O-glycosides, since C-glycosides are in general resistant
to enzymatic degradation by glycosidases and may exhi-
bit longer duration of action. It has recently been dem-
onstrated that conversion of 1a to its C-glycoside
analogue 2a leads to striking enhancement of activity
on in vivo animal models of malaria and lung cancer
Keywords: C-glycoside; OCH; Ceramide; CD1d; NKT cell ligand.
Abbreviations: TMSOTf, trimethylsilyl trifluoromethanesulfonate;
PTSA, p-toluenesufonic acid; MTPA, 2-methoxy-2-phenyl-2-(tri-
fluoromethyl)acetic acid; DME, 1,2-dimethoxyethane; EDCIÆHCl,
1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride; HOAt,
1-hydroxy-7-azabenzotriazole.
*
Corresponding author. Tel.: +81 75 9628188; fax: +81 75 962
6448; e-mail: hirokazu_annoura@dsup.co.jp
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.05.069

5044
T. Toba et al. / Tetrahedron Letters 46 (2005) 5043–5047
O
R1
X
R3
OH
OH
R3
β
γ
OH
OBn
O
OH
O
O
α
O
HO
OH
BnO
OBn
OH
OBn
OBn
O
2,3,4,6-tetra-O-benzyl-
D-galactose
BnO
OBn
OBn
+
O
O
OH
R3
L-arabinose
O
O
O
O
Figure 2. Retrosynthetic analysis of C-glycolipids.
by inducing prolonged production of the Th1 cytokines
IFN-c and IL-12.6
OBn
OBn
TMS
O
OH
O
a, b
J = 5.7Hz
OBn
These results prompted us to investigate a short and ver-
satile synthetic pathway applicable to the C-glycoside
analogues of 1b and related compounds. There have
been only a few reports on the synthesis of 2a to date,
that utilize Wittig reaction,7 Ramberg–Ba¨cklund reac-
tion followed by b-selective hydrogenation from the
diisopropylsilyl protecting group,6b or olefin cross-
metathesis for the installation of the C-glycoside link-
age.8 We present herein a concise and short synthesis
of 2b, which is independent of previous methodologies
for 2a and would be applicable to a variety of phyto-
sphingosine derivatives in terms of chain length and sub-
stitution, including aromatic groups and heteroatoms.
BnO
OBn
BnO
OBn
OBn
3
4
OBn
O
c
BnO
OBn
OBn
5
Scheme 1. Reagents and conditions: (a) Ac2O, pyridine, CH2Cl2, 0 °C
to rt (quant.); (b) tri-n-butyl(trimethylsilylethynyl)tin, TMSOTf,
MS4A, CH2Cl2, rt (60%); (c) 1 N NaOH, MeOH, CH2Cl2, rt (quant.).
Our retrosynthetic analysis for 2b revealed a straightfor-
ward and versatile strategy based on the nucleophilic
addition of an a-ethynyl sugar to the phytosphingo-
sine-precursor aldehyde synthesized from L-arabinose
(Fig. 2). The critical feature of our strategy is the dissect-
ing position, which former retrosyntheses of 2a all
cleaved between the a- and b-carbons of the anomeric
center, while we chose between the b-and c-carbons.
This allowed the shorter and expansive synthesis at the
expense of a need to construct a new stereocenter.
substrate, was detected by TLC. Desilylation of 4 was
accomplished by NaOH–MeOH to give the desired
compound 5 quantitatively.
The counterpart 9 was efficiently synthesized from
L-arabinose 6 endowed with suitable stereochemistry
corresponding to the vicinal hydroxyl groups in the
phytosphingosine moiety (Scheme 2). Thus, compound
6 was protected as a 3,4-O-acetonide, reduced to a
triol and subjected to NaIO4 degradation to produce
2,3-O-isopropylidene-L-erythrose 710 in 71% overall
yield. Four carbon homologation by Wittig reaction
followed by hydrogenation and standard Swern oxida-
tion of the primary alcohol produced the requisite alde-
hyde 9 in 61% overall yield. No epimerization of 9 was
observed neither during the reaction nor after storing
overnight at ꢀ20 °C.
Perbenzylated 1-deoxy-1-a-ethynyl-galactopyranose 5
was prepared from commercially available 2,3,4,6-tetra-
O-benzyl-D-galactose 3 in three steps (Scheme 1).9 Thus,
compound 3 was quantitatively acetylated and a-face
selectively coupled with tri-n-butyl(trimethylsilylethyn-
yl)tin in the presence of TMSOTf and MS-4A to give
4 (3JanomericHH = 5.7 Hz) in 60% yield. The correspond-
ing b-isomer of 4 was not detected but a small amount
of 3, presumably due to hydrolysis of the acetylated

T. Toba et al. / Tetrahedron Letters 46 (2005) 5043–5047
5045
OH
OH
δ
∆δ
δ
R
(ppm) =
-
O
S
OH
- 0.07905
OH
OH
MTPA
O
a
b
O
- 0.0357
O
O
OBn
O
O
O
+ 0.0492
+ 0.0226
BnO
+ 0.0306
+ 0.01855
OBn
- 0.0139
- 0.0603
6
7
OH
O
+ 0.06005
OBn
c, d
Figure 3. Dd values for the MTPA esters of 10a.
O
O
O
O
8
9
meric propargylic hydroxyl group. When the lithium
acetylide derived from 5 was employed, the reaction
gave a 3.2:2 mixture of diastereomers, 10a and 10b, in
47% and 30% yields, respectively, on the basis of the
recovered starting materials11 (Scheme 3). Compounds
10a and 10b were easily separated by column chroma-
tography over silica gel using n-hexane/EtOAc (10:1).
The stereochemistry of the propargylic hydroxyl group
was determined for the more polar isomer 10a to be
the R-isomer by applying modified Mosherꢀs protocol
(Fig. 3).12,13 Other attempts of chelation-controlled
addition reaction utilizing Zn or Mg species14 did not
improve the diastereoselectivity and decreased the chem-
ical yields.15
Scheme 2. Reagents and conditions: (a) (i) Me2C(OMe)2, PTSA,
DMF, rt, (ii) NaBH4, EtOH, rt, (iii) NaIO4, H2O, rt (71%; three steps);
(b) n-BuPh3PBr, n-BuLi, THF, ꢀ78 °C to rt (79%); (c) H2, Pd/C,
EtOH, rt (96%); (d) DMSO, (COCl)2, CH2Cl2, ꢀ78 to 0 °C (80%).
Next, the coupling reaction of 5 and 9 was examined
using various metal acetylide ions. This key step requires
the generation of an adequate stereochemistry at the epi-
O
OBn
O
a
O
O
+
BnO
OBn
Once the key intermediate 10a was available, the synthe-
sis of 2b was completed in a straightforward and efficient
manner (Scheme 4). Thus, reduction of the triple bond
in 10a with TsNHNH2 followed by mesylation of the
hydroxyl group gave 11 in 86% yield. Compound 11
was azidated, reduced to an amine by hydrogenation
and consecutively acylated with n-tetracosanoic acid to
afford amide 12 in 48% overall yield. Finally, deprotec-
tion of the isopropylidene acetal of 12 under acidic
conditions and subsequent removal of the benzyl groups
by hydrogenation furnished 2b in 88% yield. The syn-
OBn
5
9
Y
X
OBn
O
O
10a (X=OH, Y=H)
10b (X=H, Y=OH)
O
BnO
OBn
OBn
1
thetic sample displayed satisfactory H NMR spectrum
and was confirmed by high-resolution mass spectrum
Scheme 3. Reagents and conditions: (a) n-BuLi, THF, ꢀ48 to ꢀ30 °C
(47% yield for 10a, 30% yield for 10b).
OMs
OH
OBn
OBn
O
O
O
O
a, b
O
O
BnO
OBn
BnO
OBn
OBn
OBn
10a
11
O
CH3(CH2)22
OBn
NH
O
O
O
f, g
c, d, e
2b
BnO
OBn
OBn
12
Scheme 4. Reagents and conditions: (a) TsNHNH2, DME, NaOAc aq, reflux (91%); (b) MsCl, Pyridine, CH2Cl2, 0 °C to rt (94%); (c) NaN3, DMF,
90 °C; (d) H2, Pd/CaCO3, EtOH, rt; (e) tetracosanoic acid, EDCIÆHCl, HOAt, Et3N, DMF–CH2Cl2, rt (48%; three steps); (f) 80% AcOH, 60 °C
(88%); (g) H2, Pd(OH)2/C, MeOH–CH2Cl2, rt (quant.).

5046
T. Toba et al. / Tetrahedron Letters 46 (2005) 5043–5047
(HR-MS).16 Compound 2b did not show in vitro IL-4
and IFN-c production in splenocytes but increased ser-
um level of IL-4 in C57BL/6 mice in vivo. Thus, com-
pound 2b proved to possess the pharmacological
profiles distinctively different from that of 1b from the
preliminary results of biological testing.17 Similar phar-
macological differences have been reported between 1a
and its C-glycolipid 2a,6 which support our findings.
E.; Araki, M.; Miyake, S. Curr. Top. Med. Chem. 2004, 4,
561–567; (d) Oki, S.; Chiba, A.; Yamamura, T.; Miyake,
S. J. Clin. Invest. 2004, 113, 1631–1640.
4. Chiba, A.; Oki, S.; Miyamoto, K.; Hashimoto, H.;
Yamamura, T.; Miyake, S. Arthritis Rheum. 2004, 50,
305–313.
5. Murata, K.; Toba, T.; Nakanishi, K.; Takahashi, B.;
Yamamura, T.; Miyake, S.; Annoura, H. J. Org. Chem.
2005, 70, 2398–2401.
6. (a) Schmieg, J.; Yang, G.; Frank, R. W.; Tsuji, M. J. Exp.
Med. 2003, 198, 1631–1641; (b) Yang, G.; Schmieg, J.;
Tsuji, M.; Frank, R. W. Angew. Chem., Int. Ed. 2004, 43,
3818–3822. This route consists of over 20 steps for the
synthesis of 2a upon counting on all the substep sequence.
7. Tomiyama, H.; Yanagisawa, T.; Nimura, M.; Noda, A.;
Tomiyama, T. JP 2001/354666, 2002; Chem. Abstr. 2002,
136, 37901x. This route was unreproducible and the final
product was unable to be isolated due to considerable
amounts of inseparable contaminants.
8. Chen, G.; Schmieg, J.; Tsuji, M.; Franck, R. W. Org. Lett.
2004, 6, 4077–4080. This route has been improved
compared with that reported in Ref. 6b and more practical
for 2a which completes in only 11 steps, however, it
requires the commercially available phytosphingosine and
is not expandable enough for analogue syntheses including
2b. Preparing commercially unavailable phytosphingosine
derivatives for this route must require several additional
steps, which makes this route less attractive for SAR
purposes; For examples of the syntheses of phytosphin-
gosine derivatives, see: Lin, C.-C.; Fan, G.-T.; Fang, J.-M.
Tetrahedron Lett. 2003, 44, 5281–5283; Chiu, H.-Y.; Tzou,
D.-L. M.; Patkar, L. N.; Lin, C.-C. J. Org. Chem. 2003,
68, 5788–5791, and references cited therein.
In conclusion, we have developed a concise protocol for
the synthesis of 2b involving only 12steps starting from
commercially available 2,3,4,6-tetra-O-benzyl-D-galac-
tose 3 and L-arabinose 6. Although the coupling reac-
tion of 5 and 9 proceeded but not yet with sufficient
stereoselectivity, the total sequence of this method is
more convergent and versatile compared with those pre-
viously reported for 2a,6b,7,8 since the two obtained dia-
stereomers 10a and 10b can be easily separated by
conventional purification procedures. Consequently,
this new synthetic route would enable the synthesis of
a variety of C-glycoside analogues of phytosphingoli-
pids related to 1a and 1b, especially those which vary
in the sphingosine side chain length or substituents other
than aliphatic alkyl groups. In addition, it should be
noted that this route promises to contribute significantly
for clarifying the structure–activity relationships (SARs)
of this series of C-glycosides of interest. The SAR study
in this series of compounds will be reported in due
course.
9. Dondoni, A.; Mariotti, G.; Marra, A. J. Org. Chem. 2002,
67, 4475–4486.
Acknowledgements
10. Sabino, A. A.; Pilli, R. A. Tetrahedron Lett. 2002, 43,
2819–2821.
11. In this reaction, the starting a-ethynyl-galactopyranose 5
was usually recovered unchanged in over 15% yield.
12. Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J.
Am. Chem. Soc. 1991, 113, 4092–4096.
We thank Ms. K. Nakanishi for her initial contribution
to this synthetic work. We also thank Drs. M. Masu-
mura, Y. Tanaka, and Ms. M. Goto for carrying out
biological testing. Drs. T. Nishihara and G. Nakayama
are acknowledged for their supports and encouragement
throughout this study.
13. 1H NMR of (S)-MTPA ester of 10a (400 MHz, CDCl3): d
7.6–7.55 (m, 2H), 7.4–7.2 (m, 23H), 5.65 (dd, 1H, J = 7.8
Hz, 1.7 Hz), 4.90 (d, 1H, J = 11.4 Hz), 4.83 (dd, 1H,
J = 5.8 Hz, 1.7 Hz), 4.77 (d, 1H, J = 11.7 Hz), 4.73 (d, 1H,
J = 11.8 Hz), 4.69 (d, 1H, J = 11.7 Hz), 4.64 (d, 1H, J =
11.8 Hz), 4.54 (d, 1H, J = 11.4 Hz), 4.44 (d, 1H, J = 11.9
Hz), 4.37 (d, 1H, J = 11.9 Hz), 4.18 (dd, 1H, J = 7.6 Hz,
5.8 Hz), 4.08 (dd, 1H, J = 9.7 Hz, 5.7 Hz), 4.15–4.05 (m,
1H), 4.01 (t, 3H, J = 6.2Hz), 3.95–3.93 (m, 1H), 3.78 (dd,
1H, J = 9.8 Hz, 2.7 Hz), 3.49 (s, 3H), 3.5–3.4 (m, 2H), 1.6–
1.5 (m, 2H), 1.55–1.4 (m, 1H), 1.39 (s, 3H), 1.27 (s, 3H), 1.3–
1.1 (m, 5H), 0.82(t, 3H, J = 6.5 Hz); 1H NMR of (R)-
MTPA ester of 10a (400 MHz, CDCl3): d 7.6–7.57 (m, 2H),
7.5–7.2(m, 23H), 5.66 (dd, 1H, J = 8.9 Hz, 1.8 Hz), 4.90 (d,
1H, J = 11.4 Hz), 4.80 (dd, 1H, J = 6.0 Hz, 1.8 Hz), 4.77 (d,
1H, J = 11.7 Hz), 4.69 (d, 1H, J = 11.7 Hz), 4.68 (d, 1H,
J = 11.7 Hz), 4.61 (d, 1H, J = 11.7 Hz), 4.55 (d, 1H, J =
11.4 Hz), 4.45 (d, 1H, J = 11.9 Hz), 4.37 (d, 1H,
J = 11.9 Hz), 4.26 (dd, 1H, J = 8.9 Hz, 5.7 Hz), 4.11 (ddd,
1H, J = 13.6 Hz, 5.7 Hz, 3.2Hz), 4.06 (dd, 1H, J = 9.8 Hz,
5.9 Hz), 3.96 (t, 3H, J = 6.4 Hz), 3.93–3.91 (m, 1H), 3.72
(dd, 1H, J = 9.8 Hz, 2.8 Hz), 3.57 (s, 3H), 3.5–3.4 (m, 2H),
1.75–1.55 (m, 2H), 1.55–1.5 (m, 1H), 1.40 (s, 3H), 1.32 (s,
3H), 1.3–1.1 (m, 5H), 0.81 (t, 3H, J = 6.8 Hz).
References and notes
1. For reviews: (a) Porcelli, S. A.; Modlin, R. L. Annu. Rev.
Immunol. 1999, 17, 297–329; (b) Hong, S.; Scherer, D. C.;
Singh, N.; Mendiratta, S. K.; Serizawa, I.; Koezuka, Y.;
Van Kaer, L. Immunol. Rev. 1999, 169, 31–44; (c) Sidobre,
S.; Kronenberg, M. J. Immunol. Methods 2002, 268, 107–
121; (d) Wilson, M. T.; Van Kaer, L. Curr. Pharm. Des.
2003, 9, 201–220; (e) Cox, T. Phil. Trans. R. Soc. Lond. B
2003, 358, 967–973; (f) Brigl, M.; Brenner, M. B. Annu.
Rev. Immunol. 2004, 22, 817–890.
2. (a) Morita, M.; Motoki, K.; Akimoto, K.; Natori, T.;
Sakai, T.; Sawa, E.; Yamaji, K.; Koezuka, Y.; Kobayashi,
E.; Fukushima, H. J. Med. Chem. 1995, 38, 2176–2187; (b)
Morita, M.; Sawa, E.; Yamaji, K.; Sakai, T.; Natori, T.;
Koezuka, Y.; Fukushima, H.; Akimoto, K. Biosci.
Biotechnol. Biochem. 1996, 60, 288–292; (c) Takikawa,
H.; Muto, S.; Mori, K. Tetrahedron 1998, 54, 3141–3150;
(d) Graziani, A.; Passacantilli, P.; Piancatelli, G.; Tani, S.
Tetrahedron: Asymmetry 2000, 11, 3921–3937.
14. (a) Soucy, R. L.; Kozhinov, D.; Behar, V. J. Org. Chem.
2002, 67, 1947–1952; (b) Shimada, K.; Kaburagi, Y.;
Fukuyama, T. J. Am. Chem. Soc. 2003, 125, 4048–4049;
(c) Tamura, Y.; Annoura, H.; Fuji, M.; Yoshida, T.;
3. (a) Miyamoto, K.; Miyake, S.; Yamamura, T. Nature
2001, 413, 531–534; (b) Yamamura, T.; Miyake, S. PCT
Int. Appl. WO 2003/016326, 2003; Chem. Abstr. 2003, 138,
205292m; (c) Yamamura, T.; Miyamoto, K.; Illes, Z.; Pal,

T. Toba et al. / Tetrahedron Letters 46 (2005) 5043–5047
5047
Takeuchi, R.; Fujioka, H. Chem. Pharm. Bull. 1987, 35,
4736–4746; (d) Reez, M. T. Angew. Chem., Int. Ed. Engl.
1984, 23, 556–569.
(m, 3H), 4.37 (dd, 1H, J = 11.2Hz, 4.6 Hz), 4.25 (dd, 1H,
J = 8.9 Hz, 3.4 Hz), 4.25–4.15 (m, 3H), 2.8–2.65 (m, 1H),
2.65–2.52 (m, 1H), 2.52–2.38 (m, 2H), 2.38–2.15 (m, 3H),
1.95–1.8 (m, 4H), 1.75–1.55 (m, 1H), 1.45–1.1 (m, 44H),
0.87 (t, 3H, J = 6.6 Hz), 0.81 (t, 3H, J = 7.0 Hz); HR-
FABMS (m/z) [M+H]+ calculated for C40H80NO8,
702.5884, found 702.5853.
15. The total chemical yield of 10a,b was decreased to 39%
with a similar diastereomeric ratio when the magnesium
acetylide of 5 was used in this reaction. The zinc acetylide
derived from 5 produced only a trace of 10a,b.
28
16. Data for 2b: ½aꢁD +9.3 (c 0.10, pyridine); 1H NMR
17. The effect of 2b in several animal models of autoimmune
disease is currently under investigation and will be
reported elsewhere.
(400 MHz, pyridine-d5): d 8.45 (d, 1H J = 7.3 Hz), 5.2–
5.05 (m, 1H), 4.74 (dd, 1H, J = 8.9 Hz, 5.4 Hz), 4.6–4.45