Highly stereoselective and iterative synthesis of α-(1→4)-linked polysaccharides composed of 3-O-methyl-D-mannose

Article abstract of DOI:10.1021/ol0713335

A second-generation synthesis of synthetic 3-O-methyl-D-mannose-containing polysaccharides (sMMPs) is reported. The glycosidation of donor B and acceptor C, prepared from a common precursor A in two and one steps, respectively, is effected by t-butyldimethylsilyl trifluoromethanesulfonate to furnish only the desired α-anomer D in high yields. Unlike the first-generation synthesis, this synthesis gives the desired product free from contamination of scrambling products. A three-step protocol is used to deprotect D to furnish sMMPs.

Full text of DOI:10.1021/ol0713335

ORGANIC
LETTERS
2007
Vol. 9, No. 17
3323-3326
Highly Stereoselective and Iterative
Synthesis of -(1 4)-Linked
Polysaccharides Composed of
3-O-Methyl- -mannose
r
f
D
Hwan-Sung Cheon, Yiqian Lian, and Yoshito Kishi*
Department of Chemistry and Chemical Biology, HarVard UniVersity,
12 Oxford Street, Cambridge, Massachusetts 02138
kishi@chemistry.harVard.edu
Received June 5, 2007
ABSTRACT
A second-generation synthesis of synthetic 3-O-methyl-
and acceptor C, prepared from common precursor
trifluoromethanesulfonate to furnish only the desired -anomer D in high yields. Unlike the first-generation synthesis, this synthesis gives the
desired product free from contamination of scrambling products. A three-step protocol is used to deprotect D to furnish sMMPs.
D-mannose-containing polysaccharides (sMMPs) is reported. The glycosidation of donor
B
a
A
in two and one steps, respectively, is effected by t-butyldimethylsilyl
r
Fatty acid (FA) biosynthesis in Mycobacterium smegmatis
is known to show a bimodal product distribution, with
palmitic and tetracosanoic acids being the two dominant
products. Bloch and co-workers discovered that two endog-
enous M. smegmatis polysaccharides, 3-O-methyl-D-man-
nose-containing polysaccharides (MMPs) and 6-O-methyl-
D-glucose-containing lipopolysaccharides (MGLPs)/6-O-
methyl-D-glucose-containing polysaccharides (MGPs), have
profound effects on the FA biosynthesis catalyzed by M.
smegmatis fatty acid synthase I (FAS I), most noticeably on
the product distribution and stimulation.1 We are interested
in gaining mechanistic insight into the intriguing biological
role(s) of MMP and MG(L)P. However, we felt that naturally
occurring MMP and MG(L)P are not necessarily ideal
substrates for our study, as they are isolated as complex
mixtures of closely related polysaccharides.1 Thus, we
designed and used synthetic polysaccharides structurally
related to natural MMP and MG(L)P for two reasons: (1)
synthetic polysaccharides can be available as structurally
well-defined and chemically homogeneous materials and (2)
synthetic polysaccharides can be structurally tunable for the
needs of our systematic investigation. Obviously, the most
unique structural feature of natural MMP and MG(L)P is
the polymeric form of 3-O-methyl mannose and 6-O-methyl
glucose, respectively. Therefore, we incorporated this struc-
(1) For isolation and characterization of MMPs, see (a) Gary, G. R.;
Ballou, C. E. J. Biol. Chem. 1971, 246, 6835. (b) Ilton, M.; Jevans, A. W.;
McCarthy, E. D.; Vance, D.; White, H. B., III; Bloch, K. Proc. Natl. Acad.
Sci. U.S.A. 1971, 68, 87. (c) Maitra, S. K.; Ballou, C. E. J. Biol. Chem.
1977, 252, 2459. For isolation and characterization of MGLPs, see (d) Lee,
Y. C.; Ballou, C. E. J. Biol. Chem. 1964, 239, PC3602. (e) Forsberg, L. S.;
Dell, A.; Walton, D. J.; Ballou, C. E. J. Biol. Chem. 1982, 257, 3555. For
the revised structure of MG(L)P, see (f) Tuffal, G.; Albigot, R.; Rivie`re,
M.; Puzo, G. Glycobiology 1998, 8, 675.
10.1021/ol0713335 CCC: $37.00
© 2007 American Chemical Society
Published on Web 07/21/2007

tural feature in the synthetic polysaccharides, that is, synthetic
MMPs (sMMPs) and synthetic MGPs (sMGPs) (see Figure
1).2
synthesis. First, as mentioned, the exceptionally high ste-
reoselectivity of glycosidation was due to the stereospecific
anomerization of the undesired â-anomers to the desired
R-anomers. This beneficial anomerization was compromised
by the fact that the glycosidic bonds were susceptible to
random cleavages, thereby resulting in the product being
contaminated with scrambled polysaccharides.2a Second, the
glycosidation in the presence of the Mukaiyama catalyst
presented technical difficulties in terms of scalability and
reproducibility, particularly for the synthesis of larger
oligosaccharides.2
Clearly, our first priority was to identify a glycosidation
without scrambling. Using the monomeric substrates shown
in Table 1, we screened a variety of the glycosidation
Figure 1. Structures of natural and synthetic MMPs.
We recently reported a first-generation synthesis of
sMMPs.2a With slight modifications of the Mukaiyama
glycosidation,3 high R-selectivity (>50:1 to >20:1) and
yields (74-79%) were achieved for the key glycosidation
steps (Scheme 1). The observed, exceptionally high R-se-
Table 1. Glycosidation of R- and â-Anomeric Phosphates
Scheme 1. First Generation of Iterative sMMP Synthesis
donor (1)
glycosidation
selectivity
stereochemistry
(C1)
entry
R
yield
(R:â)
1
2
3
4
Bn
Bn
Bz
Bz
R
â
R
â
92%
93%
90%
91%
1.3:1
1.2:1
>20:1
>20:1
lectivity was due to the stereospecific anomerization of â-
to R-anomer under the glycosidation conditions. This gly-
cosidation was well suited for a highly convergent oligosac-
charide synthesis, particularly because of excellent chemical
yields even when using equal- or similar-sized donors and
acceptors in an approximately 1:1 molar ratio. Thus, an
iterative sequence allowed the growing oligosaccharide to
double in size after each cycle and led to an efficient
synthesis of sMMP 8-, 12-, and 16-mers. With the use of
these sMMPs, we were able to demonstrate that sMMPs
mimic the chemical and biological roles of natural MMP.4,5
With this exciting result in hand, we then realized that it
had become critically important to secure the supply of a
relatively large quantity of sMMPs for further studies. In
this connection, we were anxious to address two specific
issues regarding the glycosidation in the first-generation
methods reported in the literature.6 Through this screening,
the glycosidation via an anomeric phosphate, originally
reported by Hashimoto, Honda, and Ikegami,7,8 emerged as
the most promising candidate; in particular, we were encour-
aged with the mildness of the glycosidation conditions
(activator, reaction temperature, and reaction time). With the
donor bearing a benzoate at C2, the desired R-anomer was
obtained in high yields from both R- and â-anomeric
phosphates. A 1H NMR analysis indicated that no undesired
â-anomer was formed in the glycosidation.9 This high
(6) For comprehensive monographs, general reviews, and examples
relevant to this work, see references 9, 10, and 11 cited in reference 2a.
(7) Hashimoto, S.; Honda, T.; Ikegami, S. J. Chem. Soc., Chem. Commun.
1989, 685.
(8) Seeberger has extensively used anomeric phosphates for both solution-
and solid-phase syntheses. For examples of solution-phase synthesis, see
(a) Plante, O. J.; Andrade, R. B.; Seeberger, P. H. Org. Lett. 1999, 1, 211.
(b) Plante, O. J.; Palmacci, E. R.; Andrade, R. B.; Seeberger, P. H. J. Am.
Chem. Soc. 2001, 123, 9545. (c) Codee, J. D. C.; Seeberger, P. H. ACS
Symp. Ser. 2007, 960, 150-164 and references cited therein. For recent
examples of solid-phase synthesis, see (d) Plante, O. J.; Palmacci, E. R.;
Seeberger, P. H. Science 2001, 291, 1523. (e) Werz, D. B.; Castagner, B.;
Seeberger, P. H. J. Am. Chem. Soc. 2007, 129, 2770 and references cited
therein.
(2) (a) Hsu, M. C.; Lee, J.; Kishi, Y. J. Org. Chem. 2007, 72, 1931. (b)
Meppen, M.; Wang, Y.; Cheon, H.-S.; Kishi, Y. J. Org. Chem. 2007, 72,
1941.
(3) (a) Mukaiyama, T.; Takashima, T.; Katsurada, M.; Aizawa, H. Chem.
Lett. 1991, 533. (b) Mukaiyama, T.; Katsurada, M.; Takashima, T. Chem.
Lett. 1991, 985. (c) Mukaiyama, T.; Matsubara, K.; Sasaki, T.; Mukaiyama,
T. Chem. Lett. 1993, 1373.
(4) Cheon, H.-S.; Wang, Y.; Ma, J.; Kishi, Y. ChemBioChem 2007, 8,
353.
(5) Papaioannou, N.; Cheon, H.-S.; Lian, Y.; Kishi, Y. ChemBioChem,
in press.
(9) The stereoselectivity was determined from the 1H NMR spectrum of
the crude product; only the desired R-anomer and unreacted acceptor were
detected. The stereochemistry of the coupled product was assigned by
nuclear Overhauser effect studies.
3324
Org. Lett., Vol. 9, No. 17, 2007

selectivity can be attributed to a neighboring group participa-
tion of the C2 benzoate during glycosidation.10 Thus, unlike
the previous synthesis, it is unnecessary to rely on the
anomerization to enhance the overall stereoselectivity of
glycosidation.
pated TBDPS to meet our needs as well, but did not see an
apparent advantage in using TBDPS over a TIPS group. With
these analyses, we decided to use the substrate A with R3Si
) TIPS as the starting material, which was prepared in six
straightforward steps from D-mannose.13
In order to adopt the phosphate-based glycosidation for
the iterative synthesis of sMMPs, we needed to use orthogo-
nal protecting groups for the C2, C4, and C6 hydroxyl
groups. To achieve a highly stereoselective glycosidation,
we chose to use a benzoyl (or an acyl) group for the C2
protecting group of the donor. We also recognized two
benefits of protecting the C2 hydroxyl group of the acceptor
with the same acyl group. First, because of the electronic
effect, we anticipated that the C2 acyl group should suppress
the formation of an oxonium ion, leading to the undesired
process of scrambling. Second, except for n(m), the structure
of the product after each iteration is identical to the structure
of the starting material, and therefore it is not necessary to
adjust the protecting groups after each iteration. Because of
its stability and accessibility, we decided to use the benzyl
group for the protection at the C6 hydroxyl. Considering that
these protecting groups should be orthogonal, we focused
on a silyl ether for the C4 protecting group and conducted a
stability test for t-butyldimethylsilyl (TBS), triisopropylsilyl
(TIPS), and t-butyldiphenylsilyl (TBDPS) under the de-
allylation and glycosidation conditions. As seen from the
summary given in Table 2, a TIPS group meets our needs
Table 3. Synthesis of Protected sMMPs via Iterative
Elongationa
glycosidationb
n
m
n + m
isolated yield (R only)
1
2
4
4
8
8
8
8
12
12
1
2
2
4
2
4
6
8
6
8
2
4
6
90%
87%
85%
83%
80%
78%
83%
75%
80%
76%
8
10
12
14
16
18
20
Table 2. Screening for C4-OH Protecting Group
a Reagents and conditions: (a) (1) (Ph3P)3RhCl, dabco, EtOH-toluene,
reflux, followed by treatment with I2, THF-CH2Cl2-H2O, 80-88%. (2)
Et2NP(OBn)2, 1H-1,2,4-triazole, NaHCO3, CH2Cl2, followed by workup
with 30% H2O2, THF-CH2Cl2, 83-89%. (b) TBAF, THF, 83∼92%; (c)
TBSOTf, -40 f -30 °C, 30 min, CH2Cl2. b For all the cases, the R/â-
selectivity of glycosidation was found to be >20:1.
R3Si
tBuMe2Si
(TBS)
iPr3Si
tBuPh2Si
(TBDPS)
reaction conditions
(TIPS)
Table 3 summarizes the iterative synthesis of various sizes
of the protected sMMPs. To our delight, transformations to
donors14 and acceptors, and their glycosidation at each
iteration, proceeded smoothly, thereby demonstrating the
generality of this synthetic approach. All the protecting
groups worked as designed and behaved orthogonally during
functional group manipulations. Donor B and acceptor C,
deallylationa
H3O+, heat desilylation desilylation desilylation
HgCl2-HgO stable
stable
stable
n.d.b
I2
stable
n.d.b
glycosidation
TMSOTf
desilylation desilylation desilylation
(CH2Cl2, -40 °C) TESOTf
desilylation desilylation n.d.b
TBSOTf
desilylation stable
n.d.b
a Deallylation was performed in two steps, i.e., (1) Rh-catalyzed double-
bond isomerization and (2) hydrolysis of the resulting enol ether. The
reaction conditions indicated are concerned with the second step. b n.d.:
not determined.
(12) Seeberger used TBSOTf as an activator. For examples, see (a)
reference 8b. (b) Bosse, F.; Marcaurelle, L. A.; Seeberger, P. H. J. Org.
Chem. 2002, 67, 6659.
(13) This substance was prepared as shown below.
well; it is stable under the deallylation (Rh-catalyzed double-
bond isomerization, followed by I2 treatment11) and the
glycosidation (using t-butyldimethylsilyl trifluoromethane-
sulfonate (TBSOTf),12 instead of trimethylsilyl trifluo-
romethanesulfonate (TMSOTf), as the activator). We antici-
(10) There are a number of examples for showing the high glycosidation
selectivity of mannosyl donors with C2 participating groups. For examples,
see (a) reference 8b. (b) Lemanski, G.; Ziegler, T. HelV. Chim. Acta 2000,
83, 2655. (c) Ning, J.; Kong, F. Tetrahedron Lett. 1999, 40, 1357.
(11) For deallylation, see (a) Corey, E. J.; Suggs, W. J. J. Org. Chem.
1973, 38, 3224. (b) Nashed, M. A.; Anderson, L. J. Chem. Soc., Chem.
Commun. 1982, 1274.
(14) There are numerous examples known for the preparation of anomeric
phosphates. With two slight modificationss(1) NaHCO3 added to accelerate
the phosphitylation reaction and (2) 1-H 1,2,4-triazole used instead of 1-H
tetrazoleswe adopted the protocol reported by Wong; see (a) Pederson, R.
L.; Esker, J.; Wong, C.-H. Tetrahedron 1991, 47, 2643. (b) Sim, M. M.;
Kondo, H.; Wong, C.-H. J. Am. Chem. Soc. 1993, 115, 2260.
Org. Lett., Vol. 9, No. 17, 2007
3325

prepared from the common intermediate A in two and one
steps, respectively, were coupled via TBSOTf-promoted,
phos-
From the protected polysaccharides D, sMMPs were
obtained by a three-step sequence of deprotection (Table 4).
phate-based glycosidation to give the larger oligosaccharide
D, which, in turn, served as the starting material for the next
round of iteration. In practice, we chose donors as limiting
substrates, because the acceptors were available from the
common intermediate A more readily than the donors. For
the case of n + m ) 2 glycosidation, the optimal donor/
acceptor ratio was found to be around 1:1.2. Although fine-
tunings were required for each case, the optimal ratio was
found to always be less than 1:1.4.
Table 4. Deprotection to Form sMMPs
n
step 1
step 2
6
92%
90%
91%
88%
86%
84%
89%
83%
87%
82%
78%
81%
66%
71%
69%
73%
8
10
12
14
16
18
20
It is worth mentioning several aspects of the improved
glycosidation. First, regardless of the size of the oligosac-
charides, the glycosidation between donors and acceptors was
highly stereoselective to give only the R-linked isomer. No
undesired â-anomer was detected in either the crude or
isolated product of each glycosidation (1H NMR).15 Second,
no scrambled product was detected in the crude product of
each glycosidation. This was unambiguously demonstrated
by a high-performance liquid chromatography (HPLC)
analysis; namely, each oligosaccharide has a characteristic
retention time and is readily distinguishable from the other
oligosaccharides (Figure 2). Third, the glycosidation of
Overall, this improved synthesis enabled us to secure sMMPs
ranging from 6-mer through 20-mer in relatively large
quantities.
In conclusion, we have developed a second-generation
synthesis of sMMPs. Glycosidation donor B and acceptor
C were prepared from a common precursor A in 2 and 1
steps, respectively (Table 3). The glycosidation of B with C
was effected by TBSOTf in CH2Cl2 to furnish only the
desired R-anomer D in high yields. Except for the number
of hexoses, the product D is structurally identical to the
starting material A, thereby allowing one to iterate the
synthetic sequence. Unlike the first generation synthesis, this
method yielded the desired product free from contamination
of the scrambling products. The three-step protocol was then
used to deprotect D to furnish sMMP 6-, 8-, 10-, 12-, 14-,
16-, 18-, and 20-mers. This synthesis is easy to scale and
applicable to the synthesis of analogs of sMMPs. For
example, this iterative synthesis has been used for the
synthesis of synthetic 3-O-ethyl-D-mannose-containing polysac-
charides.5
Figure 2. HPLC traces of the polysaccharides D. Conditions:
column: Hypersil (Keystone Scientific); eluent: 30% EtOAc-
hexanes (isocratic); detector: UV at 254 nm; rt.
Acknowledgment. Financial support from Eisai Research
Institute, Amgen, and Eli Lilly is gratefully acknowledged.
oligosaccharides was achieved without significant reduction
in yield upon scale-up. In addition, the phosphate-based
approach was found to be equally effective for larger
substrates;16 indeed, sMMP 18-mer and 20-mer were syn-
thesized by this method.
Supporting Information Available: Experimental details
1
and H NMR spectra (29 pages). This material is available
free of charge via the Internet at http://pubs.acs.org
OL0713335
(16) Although reaction conditions for each glycosidation were similar
for all the substrates, it was found that the amount of TBSOTf required to
complete the glycosidation largely depends on both the size of oligosac-
charides and the concentration of the reaction. Generally, for larger
oligomers and reactions at lower concentrations, it was required to use
slightly more TBSOTf. For details, see Supporting Information.
(15) The stereochemistry of the product was further confirmed after the
deprotection to sMMPs. No â-linked anomeric proton signals were detected
in their 1H NMR spectra; â-linked anomeric protons are known to give
resonances shifted to upfields compared to those of the corresponding
R-linked anomeric protons.
3326
Org. Lett., Vol. 9, No. 17, 2007