First total synthesis of caminoside A, an antimicrobial glycolipid from sponge

Article abstract of DOI:10.1055/s-2004-837221

Caminoside A, a novel antimicrobial tetrasaccharide glycolipid from the marine sponge Caminus sphaeroconia, which represents the first bacterial type III secretion inhibitor, is synthesized in a total of 57 steps starting from D-glucose, D-galactose, L-rh

Full text of DOI:10.1055/s-2004-837221

LETTER
437
First Total Synthesis of Caminoside A, an Antimicrobial Glycolipid from
Sponge
First
T
otal
i
S
ynthe
a
sis of Cam
n
inoside
A
song Sun,a Xiuwen Han,a Biao Yu*b
a
State Key Laboratory of Catalyst, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, P. R. China
b
State Key Laboratory of Bioorganic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry,
Chinese Academy of Sciences, Shanghai 200032, P. R. China
E-mail: byu@mail.sioc.ac.cn
Received 18 November 2004
Abstract: Caminoside A, a novel antimicrobial tetrasaccharide
glycolipid from the marine sponge Caminus sphaeroconia, which
represents the first bacterial type III secretion inhibitor, is syn-
thesized in a total of 57 steps starting from D-glucose, D-galactose,
L-rhamnose, and 9-decenal.
Key words: caminoside A, D-glucose, D-galactose
Enteropathogenic Escherichia coli (EPEC) and enterohe-
morragic E. coli 0157:H7 (EHEC) are deadly pathogens
toward children and the elderly. Infection of these patho-
genic E. coli requires the bacterial secreted protein (Esps)
and a type III secretory apparatus, which translocates se-
creted proteins across bacterial membranes, out of EPEC
and EHEC, into the host epithelial cells. Remarkably, the
type III secretory system, which is essential for the patho-
genicity of EPEC and EHEC, is absent in nonpathogenic
E. coli. Thus, selective inhibition of the type III secretory
system might specifically attenuate pathogenic EPEC and
EHEC without affecting the commensal E. coli flora.1 Re-
cently, a high throughput assay for inhibitors of the type
III secretion of EPEC was developed. Activity-guided iso-
lation of marine invertebrate extracts led to the discovery
of caminoside A (1) from the marine sponge Caminus
sphaeroconia.1 This type III secretion inhibitor (IC50 = 20
mM) also displayed reasonably potent in vitro inhibition
against the growth of methicillin resistant Staphylococcus
aureus (MIC = 12 mg/mL) and vancomycin resistant En-
terococcus (MIC = 12 mg/mL).1 Equally attractive is the
structure of caminoside A (1): its middle glucose residue
(sugar B) is fully substituted; the terminal 6-deoxy-D-
talose (sugar C) and L-quinovose (sugar D) are rare in na-
ture; and the methyl ketone lipid aglycone is without pre-
cedent in sponge metabolites. Here we report the first total
synthesis of this structurally and biologically interesting
marine natural product.
Scheme 1
assembling 3, stereocontrolled construction of the four
glycosidic linkages turns out to be the key. Formation of
the 1,2-trans-b-glucopyranoside linkages (of sugars A
and B) would be ensured by glycosylation with donors
installed with a neighboring participating group, i.e., tri-
fluoroacetimidate 4 bearing 2-O-Ac and a donor elaborat-
ed from p-methoxyphenyl (MP) glucoside 5 bearing 2-O-
Azmb [2-(azidomethyl)benzoyl] group.2 Glycosylation
with the L-glucopyranose-type donor 7, which bears a
non-participating 2-O-Bn group, would produce the de-
sired thermodynamically favored 1,2-cis-a-quinovopyra-
noside linkage (of sugar D) predominantly.3 However,
construction of the 1,2-cis-b-mannopyranoside-type
linkage of the 6-deoxy-talose (sugar C) is not a trivial
problem.4 An indirect approach was thus planned: glyco-
sylation with 1-thio-fucopyranoside 6 installed with a di-
recting 2-O-Lev (levulinyl) group should afford the b-
glycosidic linkage stereoselectively;5 then selective re-
moval of the 2-O-Lev group followed by an inversion of
the 2-OH configuration would furnish the desired 6-
deoxy-1,2-cis-b-talopyranosidic linkage.4,6
Removal of the benzyl protective groups (Bn on 3) under
neutral hydrogenolysis conditions was planed as the final
step to elaborate the target molecule 1, to avoid probable
cleavage or migration of the acetyl (Ac) and butyryl (Bt)
moieties under acidic/basic conditions (Scheme 1). In
SYNLETT 2005, No. 3, pp 0437–0440
1
6.
0
2.
2
0
0
5
Advanced online publication: 22.12.2004
DOI: 10.1055/s-2004-837221; Art ID: U32204ST
© Georg Thieme Verlag Stuttgart · New York

438
J. Sun et al.
LETTER
Preparation of the designed monosaccharide building Ethyl 3,4-di-O-Bn-2-O-Lev-1-thio-b-D-fucopyranoside
blocks 4–7 was straightforward (Scheme 2, Scheme 3, (6) was synthesized from D-galactose (Scheme 4). Known
Scheme 4, and Scheme 5). Modifying literature proce- procedures were employed in the preparation of peracetyl
dures,7 2-O-acetyl-3,4,6-tri-O-benzyl-D-glucose (9) was D-fucopyranoside (16, 5 steps, 44% yield from D-galac-
conveniently synthesized in 6 steps and 42% overall yield tose)12 and the conversion of 16 into thioglycoside 17 (3
from D-glucose. Under our reported conditions,8 com- steps, 70%).13 After temporary protection of the 2-OH
pound 9 was converted into the trifluoroacetimidate 4 in with p-methoxybenzyl (PMB) group, the 3,4-O-iso-
high yield (90%, Scheme 2).
propylidene was taken off, and the resulting 3,4-OH
groups were blocked with the persistent Bn groups.
Oxidative removal of the 2-O-PMB group followed by
protection of the resulting -OH with Lev group afforded 6
in excellent yield.
Scheme 2 (a) CF3C(Cl)=NPh, K2CO3, acetone, r.t., 90%.
Scheme 3 shows the synthesis of p-methoxyphenyl 4-O-
Ac-2-O-Azmb-3-O-Bt-b-D-glucopyranoside (5). 1,2,4,6-
Tetra-O-Ac-3-O-Allyl-D-glucose (10), readily prepared
from D-glucose (4 steps, 70%),9 was converted into the
p-methoxyphenyl b-glucoside (11) under glycosylation
conditions (70%).10 Removal of the 2,4,6-acetates fol-
lowed by selective protection of the 4,6-OH groups with
benzylidene provided 12. The remaining 2-OH was then
protected with Azmb group. Cleavage of the 4,6-ben-
zylidene followed by selective protection of the resulting
6-OH with Lev group and the remaining 4-OH with ace-
tate afforded orthogonally fully protected glucopyrano-
side 14. Selective cleavage of the 3-O-All group was then
achieved under the action of PdCl2 in methanol,11 albeit in
a moderate yield (66%). Substitution of the resulting 3-
OH with Bt group led to 15. Finally, selective removal of
the 6-O-Lev group using hydrazine acetate furnished 5,
which is a ready acceptor for coupling with the glycosyl
donor of sugar C (6).
Scheme 4 (a) PMBCl, NaH, DMF, 0 °C to r.t., 80%; (b) 70%
HOAc, 40 °C; (c) BnBr, NaH, DMF, 0 °C to r.t., 80% (for 2 steps);
(d) DDQ, CH2Cl2–H2O, r.t., 88%; (e) levulinic acid, DCC, DMAP,
CH2Cl2, r.t., 90%.
The perbenzyl L-quinovopyranosyl trifluoroacetimidate
donor 7 was prepared starting from L-rhamnose
(Scheme 5). The 2-OH configuration of L-rhamnose was
conversed using a literature method,14 thus leading to
2,3,4-tri-O-benzoyl-L-quinovopyranosyl bromide 21 in 3
steps and 36% yield. Glycosylation of 21 with p-methoxy-
benzoic alcohol under the promotion of AgOTf gave the
b-O-PMB quinovoside 22 in quantitative yield. After
changing the Bz protection into Bn protection, the ano-
meric PMB group was then removed under the action of
trifluoroacetic acid (TFA) to give 24, which was readily
converted into the desired trifluoroacetimidate donor 7.
The C-B disaccharide fragment was elaborated first.
(Scheme 6) Coupling of thioglycoside 6 with 6-OH-glu-
coside 5 under the promotion of NIS/TMSOTf15 afforded
the expected (1→6)-b-disaccharide 25 in 76% yield (d =
4.46 ppm, J = 7.8 Hz for the anomeric H-1¢). Highly
selective removal of the 2-O-Lev group, in the presence of
Ac, Azmb, and Bt groups, was achieved with hydrazine
acetate, giving 26 (90%). The configuration of the equa-
torial 2-OH was then successfully conversed into the axial
one following an oxidation–reduction sequence,6 pro-
viding 27 in a satisfactory 56% yield (for 2 steps, no
equatorial isomer was detected in the reduction step). The
Scheme 3 (a) p-MeOPhOH, TMSOTf, CH2Cl2, 0 °C, 70%; (b)
MeONa, MeOH, r.t.; (c) PhCH(OMe)2, TsOH, DMF, 40 °C, 72% (for
2 steps); (d) AzmbCl, DMAP, CH2Cl2, r.t., 85%; (e) TsOH, CH2Cl2–
MeOH, r.t., 76%; (f) levulinic acid, DCC, DMAP, CH2Cl2, 0 °C to
r.t.; (g) Ac2O, pyridine, r.t., 84% (for 2 steps); (h) PdCl2, CH2Cl2–
MeOH, r.t., 66%; (i) BtCl, pyridine, r.t., 100%; (j) H2NNH2–HOAc,
CH2Cl2–HOMe, r.t., 96%.
Synlett 2005, No. 3, 437–440 © Thieme Stuttgart · New York

LETTER
First Total Synthesis of Caminoside A
439
elaborated by a Wacker oxidation18 of the terminal C=C
double bond, providing ketone 30 in 80% yield. Glycosyl-
ation of the equatorial 2-OH of 30 with trifluoroacetimi-
date 28 (under conditions similar to those for 4) afforded
the desired Talb-(1→6)-Glub-(1→2)-Glub-trisaccharide
31 in 76% yield (d = 4.28 ppm, J = 7.41 Hz for the ano-
meric H-1¢¢), whereupon no glycosylation took place on
the free axial 2-OH of the talose moiety. Blocking this 2-
OH with Bn group was found difficult: treatment of 31
with benzyl trichloroacetimidate in the presence of triflic
acid provided 32 in moderate yield (61%), with 30% of
the starting 32 being recovered.19 Selective removal of the
2-O-Azmb group (on 32), in the presence of Ac and Bt
groups, was achieved under the action of tributylphos-
phine, affording 33 in 70% yield. The resulting 2-OH was
then glycosylated with the newly prepared perbenzyl L-
quinovopyranosyl trifluoroacetimidate 7 (under condi-
Scheme 5 (a) p-MeOBnOH, AgOTf, MS 4 Å, CH2Cl2, –10 °C,
100%; (b) MeONa–MeOH, r.t.; (c) BnBr, NaH, DMF, 0 °C to r.t.,
86% (for 2 steps); (d) TFA, CH2Cl2, r.t., 80%; (e) CF3C(Cl)=NPh,
K2CO3, acetone, r.t.
1H NMR signal of the anomeric H-1¢ then showed up as a tions similar to those used for 4 and 28) to provide the de-
singlet (at d = 4.29 ppm). Oxidative removal of the ano- sired tetrasacharide 3 as the major product (53%; d = 5.23
meric MP group gave the hemiacetal, which was convert- ppm, singlet for the anomeric H of sugar D). No b-isomer
ed into the disaccharide trifluoroacetimidate 28 in was detected. Hydrogenolysis of the Bn groups in the
excellent yield (90% for 2 steps). No addition of the tri- presence of Pd/C met with no difficulty, providing the
fluoroacetimidoyl chloride with the sterically hindered
axial 2-OH was detected.
Scheme 6 (a) NIS, TMSOTf, MeCN, –60 °C to r.t., 76%; (b)
H2NNH2–HOAc, CH2Cl2–MeOH, r.t., 90%; (c) Dess–Martin per-
iodane, CH2Cl2, r.t.; (d) NaBH4, CH2Cl2–MeOH, 0 °C, 56% (for 2
steps); (e) CAN, toluene–MeCN–H2O, r.t.; (f) CF3C(Cl)=NPh,
K2CO3, acetone, r.t., 90% (for 2 steps).
With trifluoroacetimidate donors 4, 7, and 28 at hand, we
practiced the final assembly of the target molecule 1
(Scheme 7). 1-Nonadecen-10-ol (8), prepared by addition
of 9-decenal16 with nonylmagnesium bromide,17 was gly-
cosylated with glucopyranosyl trifluoroacetimidate (4)
under the catalysis of TMSOTf8 to provide the desired b-
glucoside 29 in 92% yield (d = 4.38 ppm, J = 7.8 Hz for
the anomeric H). The 2-O-Ac group was then removed for
coupling with disaccharide trifluoroacetimidate 28. At
this stage, the methyl ketone function in the aglycone was
Scheme 7 (a) TMSOTf, MS 4 Å, CH2Cl2, 0 °C, 92%; (b) K2CO3,
THF–MeOH, r.t., 80%; (c) PdCl2, Cu(OAc)2, O2 (1 atm), DMA–H2O,
50 °C, 80%; (d) TMSOTf, MS 4 Å, CH2Cl2, 0 °C, 76%; (e)
BnOC(=NH)CCl3, TfOH, MS 4 Å, CH2Cl2–cyclohexane, r.t., 61%;
(f) Bu3P, THF–H2O, 20 °C, 70%; (g) TMSOTf, MS 4 Å, Et2O, 0 °C
to r.t., 53%; (h) H2 (1 atm), Pd/C, MeOH, r.t.; (i) Ac2O, pyridine,
40 °C, 62% (for 2 steps).
Synlett 2005, No. 3, 437–440 © Thieme Stuttgart · New York

440
J. Sun et al.
LETTER
target glycolipid 1, which was transformed into peracetate
References
1
2 directly for comparison of the analytical data. The H
(1) Linington, R. G.; Robertson, M.; Gauthier, A.; Finlay, B. B.;
van Soest, R.; Andersen, R. J. Org. Lett. 2002, 4, 4089.
(2) (a) Wada, T.; Ohkabso, A.; Mochizuki, A.; Sekine, M.
Tetrahedron Lett. 2001, 42, 1069. (b) Love, K. R.;
Andradego, R. B.; Seeberger, P. H. J. Org. Chem. 2001, 66,
8165.
(3) (a) Schmidt, R. R.; Toepfer, A. Tetrahedron Lett. 1991, 32,
3353. (b) Han, X.; Schmidt, R. R. Liebigs Ann. Chem. 1992,
817.
(4) (a) Fairbanks, A. J. Synlett 2003, 1945. (b) Crich, D.; Sun,
S. J. Am. Chem. Soc. 1997, 119, 11217.
(5) Zhu, T.; Boons, G.-J. Org. Lett. 2001, 3, 4201.
(6) (a) Liu, K. K.-C.; Danishefsky, S. J. J. Org. Chem. 1994, 59,
1892. (b) Lichtenthaler, F. W.; Schneider-Adams, T. J. Org.
Chem. 1994, 59, 6728.
NMR and 13C NMR spectra of 2 are identical to those
from the natural compound.1
In a failed attempt to assemble the tetrasaccharide target,
we found that glycosylation of the trisaccharide diol 34
with perbenzyl L-quinovopyranosyl trifluoroacetimidate
7 (under similar conditions as those for coupling of 7 with
33) provided the tetrasaccharide 35 as the major product
(72%, Scheme 8), whereupon the glycosylation took
place unexpectedly on the axial 2-OH of the talose moiety
(vs the equatorial 2-OH of the glucose moiety). While the
glycosylation with glucopyranosyl trifluoroacetimidate
28 selected the equatorial 2-OH of the glucose moiety (on
30) rather than the axial 2-OH of the talose moiety (of 28,
Scheme 7). These results reflect the importance of
‘match’ of the donor and acceptor for the selective glyco-
sidic coupling.20
(7) (a) Banoub, J.; Boullanger, P.; Potier, M.; Descotes, G.
Tetrahedron Lett. 1986, 27, 4145. (b) Trumtel, M.;
Tavecchia, P.; Veyrieres, A.; Sinaÿ, P. Carbohydr. Res.
1989, 191, 29. (c) Wotovic, A.; Jacquinet, J.-C.; Sinaÿ, P.
Carbohydr. Res. 1990, 205, 235.
(8) (a) Yu, B.; Tao, H. Tetrahedron Lett. 2001, 42, 2405.
(b) Yu, B.; Tao, H. J. Org. Chem. 2002, 67, 9099.
(c) Adinolfi, M.; Iadonisi, A.; Ravida, A.; Schiattarella, M.
Synlett 2004, 275.
(9) Takeo, K.; Nakaji, T.; Shinmitsu, K. Carbohydr. Res. 1984,
133, 275.
(10) Lee, Y. S.; Rho, E. S.; Min, Y. K.; Kim, B. T.; Kim, K. H. J.
Carbohydr. Chem. 2001, 20, 503.
(11) Ogawa, T.; Yamamoto, H. Agric. Biol. Chem. 1985, 49, 475.
(12) Lerner, L. M. Carbohydr. Res. 1993, 241, 291.
(13) (a) Zegelaar-Jaarsveld, K.; van der Plas, S. C.; van der
Marel, G. A.; van Boom, J. H. J. Carbohydr. Chem. 1996,
15, 591. (b) Zhang, Z.; Magnusson, G. Carbohydr. Res.
1994, 262, 79.
(14) (a) Bilik, V.; Voelter, W.; Bayer, E. Angew. Chem., Int. Ed.
Engl. 1971, 10, 909. (b) Lichtenthaler, F. W.; Metz, T. Eur.
J. Org. Chem. 2003, 3081.
(15) (a) Konradsson, P.; Udodong, U. E.; Fraser-Reid, B.
Tetrahedron Lett. 1990, 31, 4313. (b) Hunt, D. K.;
Seeberger, P. H. Org. Lett. 2002, 4, 2751.
(16) Banwell, M. G.; McRae, K. J. Org. Lett. 2000, 2, 3583.
(17) Holm, T. J. Chem. Soc., Perkin Trans. 2 1981, 464.
(18) (a) Smith, A. B.; Cho, Y. S.; Friestad, G. K. Tetrahedron
Lett. 1998, 39, 8765. (b) Nicolaou, K. C.; Gray, D.; Tea, J.
Angew. Chem. Int. Ed. 2001, 3675.
(19) Wessel, H.-P.; Ivesen, T.; Bundle, D. R. J. Chem. Soc.,
Perkin Trans. 1 1985, 2247.
(20) For a recent account, see: Fraser-Raid, B.; López, J. C.;
Gómez, A. M.; Uriel, C. Eur. J. Org. Chem. 2004, 1387.
(21) The stereochemistry of the aglycone was determined
recently, see: MacMillan, J. B.; Linington, R. G.; Andersen,
R. J.; Molinski, T. F. Angew. Chem. Int. Ed. 2004, 43, 5946.
Scheme 8 (a) TMSOTf, MS 4 Å, CH2Cl2, 0 °C, 72%.
It should also be reported that a pair of the diastereomeric
products were observed on TLC since the attachment of
the sugar to the racemic alcohol aglycone (for compounds
29–33 and 2, 3, respectively). And each pair of the isolat-
ed diastereoisomers has shown little difference on their 1H
NMR and 13C NMR signals.21
Acknowledgment
Financial support from the National Natural Science Foundation of
China (20172068, 20321202, and 29925203) is gratefully acknow-
ledged.
Synlett 2005, No. 3, 437–440 © Thieme Stuttgart · New York