Glycal scavenging in the synthesis of disaccharides using mannosyl iodide donors

Article abstract of DOI:10.1021/jo0478609

(Chemical Equation Presented) High mannose glycans composed of α (1→2) and α (1→6) branched sugars are important components of the HIV-associated envelope glycoprotein, gp120. These substructures can be efficiently prepared in solution from glycosyl iodid

Full text of DOI:10.1021/jo0478609

Glycal Scavenging in the Synthesis of
Disaccharides Using Mannosyl Iodide
Donors1
Son N. Lam† and Jacquelyn Gervay-Hague*
University of California, Davis, Department of Chemistry,
One Shields Avenue, Davis, California 95616
gervay@chem.ucdavis.edu
Received December 2, 2004
FIGURE 1. Man9: representative high-mannose oligosaccha-
ride of HIV-1 rgp120.
these are substructures of HIV-associated high mannose
N-glycans (1,2) that are implicated in the pathogenicity
of this disease (Figure 1). Establishing efficient method-
ologies for the construction of structural subsets in
addition to the parent compounds is an important step
toward having materials for biological evaluation. Herein,
we report our initial investigations using in situ ano-
merization4,5 of mannosyl iodides en route to the targeted
dimers and the development of a scavenging protocol for
removing glycal byproducts.
The three monomer units required for the disaccharide
syntheses include 6-O-acetyl-2,3,4-tri-O-benzyl mannopy-
ranosyl iodide (5) along with the differentially protected
acceptors 6 and 7, Figure 2. A phenyl moiety was
introduced at the reducing end to facilitate HPLC puri-
fication of fully deprotected disaccharides.
High mannose glycans composed of R (1f2) and R (1f6)
branched sugars are important components of the HIV-
associated envelope glycoprotein, gp120. These substructures
can be efficiently prepared in solution from glycosyl iodide
precursors requiring only a slight excess of the iodide donor,
which offers advantages over solid-phase methods that
require more than 5 equiv of donor. During the reaction,
excess iodide is converted to a glycal that is not easily
separated from the desired disaccharide. To overcome this
difficulty, a phase-trafficking methodology that relies upon
nucleophilic interception of the 1,2 anhydrosugar resulting
from oxidation of the glycal has been developed.
Preparation of 5 began with formation of methyl R-D-
mannopyranoside (8),6 which was subsequently subjected
to per-benzylation using NaH and BnBr with catalytic
TBAI, to give 9. Acid-catalyzed acetolysis of 9 using H2-
SO4 or ZnCl2 furnished diacetate 10, which was con-
verted to 5 using TMSI. Alternatively, 10 could be reacted
with BF3‚OEt2 in the presence of phenol to afford 11,
which upon deacetylation provided acceptor 6, (Scheme
1).
The primary alcohol of 6 was glycosidated under
standard in situ anomerization conditions (TBAI and
Hu¨nig’s base in PhH at 80 °C), with mannosyl iodide
donor 5, Scheme 2. Though the reaction provided the
R-anomer 13, we were unable to chromatographically
(flash chromatography and HPLC) separate glycal 14
from the disaccharide, as both possessed very similar Rf
values in numerous solvent systems. However, following
deacetylation of the crude mixture using an excess of
NaOMe in anhydrous MeOH, the pure R-linked disac-
charide 15, verified by a large coupling 1JC.H ) 169.0 Hz,9
The utility of glycosyl iodides in organic synthesis has
been widely demonstrated in recent years. The reactions
can be conducted under neutral conditions giving rise to
highly â-selective glycosidations2 or alternatively, in situ
anomerization can be employed in R-glycosylations.3
Although we have shown that glycosyl iodides can be
effectively employed in solid-phase oligosaccharide syn-
thesis, typically 5 equiv of the iodide is required. The
preparation of appropriately protected monosaccharide
building blocks is laborious and often the slow step in
oligosaccharide synthesis prompting us to develop more
efficient methodologies. Solution-phase reactions look
promising as only a slight excess of the iodide donor is
required and it is not unusual for crude reaction mixtures
to have only the desired product along with glycal, which
is produced by HI elimination from the excess iodide.
As a part of our program focused on developing
solution-phase oligosaccharide synthesis, we have ex-
tended the glycosyl iodide studies to include the synthesis
of branched mannose containing sugars. We have tar-
geted 1f6 (3) and 1f2 (4) R-linked mannosides because
7
8
(4) Lemieux, R. U.; Hendricks, K. B.; Stick, R. V.; James, K. J. Am.
Chem. Soc. 1975, 97, 4056-4062.
(5) Lam, S. N.; Gervay-Hague, J. Carbohydr. Res. 2002, 337, 1953-
1965.
* Corresponding author. Tel: +1-530-754-9577. Fax: +1-530-752-
8995.
† Current address: National Institutes of Health, 9000 Rockville
Pike, Bldg 8, Rm 1A02, Bethesda, MD 20892.
(1) Taken in part from the dissertation of S.L., University of
California, Davis, 2003.
(6) Whistler, R. A. Methods Carbohydr. Chem. 1987, 2, 328-329.
(7) Tennant-Eyles, R. J.; Davis, B. G.; Fairbanks, A. J. Tetrahe-
dron: Asymmetry 2000, 11 (1), 231-243.
(8) Yang, G.; Ding, X.; Kong, F. Tetrahedron Lett. 1997, 38 (38),
6725-6728.
(2) Gervay, J.; Hadd, M. J. J. Org. Chem. 1997, 62 (20), 6961-6967.
(3) Hadd, M. J.; Gervay, J. Carbohydr. Res. 1999, 320 (1-2), 61-
69.
(9) Bock, K.; Lundt, I.; Pedersen, C. Tetrahedron Lett. 1973, (13),
1037-1040.
10.1021/jo0478609 CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/18/2005
J. Org. Chem. 2005, 70, 2387-2390
2387

FIGURE 2. Retrosynthesis of targeted disaccharides.
SCHEME 1
SCHEME 2
was recovered in 89% yield over two steps. Through a
combination of chemical and chromatographic means, the
glycal was satisfactorily separated from the synthetic
target but this protocol was far from ideal and not
amenable to automated synthesis.
Having synthesized one of two targets, our attention
turned toward the preparation of acceptor 7, Scheme 3.
The synthesis began with glycosidation of acetobromo-
mannose 1610 with ethanol under basic conditions to give
1,2-O-(1-ethoxyethylidene)-3,4,6-tri-O-acetyl-â-D-man-
nopyranose 17, in high yields (>90% over 2 steps) as an
epimeric mixture. Removal of the acetyl groups using
excess NaOMe in MeOH (to give 18) with successive
benzylation provided large quantities of the tri-O-benzyl
ortho ester 19 in 30 min with high yields. Treatment of
19 with phenol in the presence of camphorsulfonic acid
provided R-phenyl glycoside 20 (Scheme 4), which after
quantitative deacetylation using catalytic NaOMe in
anhydrous MeOH afforded 7. As anticipated, glycosida-
tion of 7 with mannosyl iodide 5 under in situ anomer-
ization conditions produced disaccharide 22. After aque-
ous workup and passage of the crude mixture through a
decolorizing silica gel plug, the crude 1H NMR spectrum
illustrated the clean conversion of the monosaccharide
to the disaccharide (86% conversion to 22 based upon 1H
NMR integrations and the amount of isolated crude
material). Two major materials dominated the spectrum,
the glycal 14 (H1 6.30 ppm) and the disaccharide 22 (H1′
OPh 6.10 ppm, H1′′ OMan 5.43 ppm) with relative
integrations of 1.7:1:1, respectively.
The 1JC1,H ) 170.5 Hz coupling confirmed the R-glyco-
sidic linkage;9 however, again we encountered a situation
whereby the glycal was not readily separated from the
disaccharide. Because chromatographic separation of
deacetylated 14 and 22 is not an ideal strategy for
automated procedures, we decided to explore methods for
scavenging the glycal out of the mixture.
(10) Millar, A.; Kim, K. H.; Minster, D. K.; Ohgi, T.; Hecht, S. M. J.
Org. Chem. 1986, 51 (2), 189-196.
2388 J. Org. Chem., Vol. 70, No. 6, 2005

SCHEME 3
SCHEME 4
We felt that phase trafficking techniques11,12 were
ideally suited for purification of the glycal from the
desired disaccharide due to their orthogonal reactivities.
Phase trafficking requires the use of resins or crystalliz-
able polymers to scavenge unwanted molecules from
reactions. Whether the support is present throughout the
reaction or added post-completion is dependent upon the
nature of the support and the reaction. For example,
divinylbenzene (DVB) cross-linked polystyrene (PS) beads,
represented by a shaded ball and abbreviated as PS, offer
the advantage of insolubility in most organic solvents,
which facilitates removal via filtration. Various function-
alities attached to PS such as nucleophiles (PS-X) allow
for chemical elaboration.
Taking advantage of the electron density about the enol
ether in 14, we envisioned implementing dimethyl diox-
irane (DMDO)13 to provide a 1,2-anhydro sugar, which
would be susceptible to nucleophilic attack. Adam and
co-workers had previously shown that 2,3,4,6-tetra-O-
benzyl-D-glucal 23 undergoes electrophilic epoxidation
with DMDO at -20 °C in 2 h.14 However upon warming
to room temperature, only 2,3,4,6-tetra-O-benzyl-D-glu-
cofuranose (25) with an aldehydic C-1 was seen in the
1H NMR spectrum. It was suggested that in order to
relieve ring strain, the C-2 benzyl ether donates electrons
forming the C-2 oxacarbenium (24), which also opens the
oxirane and breaks the C-1-O-5 bond, Figure 3.
oxirane could be intercepted. The nucleophile would
attack the highly strained and electrophilic anomeric
center of 26 to immobilize the epoxidized sugar giving
27 and leaving behind only the disaccharide 22. As 22
does not possess functionalities that react with either
DMDO or the anhydro sugar, it would remain silent
throughout the sequence. This strategy was reduced to
practice upon treatment of the crude mixture with a 1.5
molar excess of DMDO relative to the glycal in the
presence of PS-OH (1.5 molar excess to the glycal) at
-50 °C for 3 h (Scheme 4), which removed 60% of the
1
glycal as revealed by H NMR integrations.
Signs of unreacted 1,2-anhydro sugar were disclosed
by the presence of ring-contracted furanosyl C-glycoside
1
revealed through H and 13C NMR (H1 9.41 ppm, C1
196.17 ppm). The aldehydic resonance had an integration
of 0.3 relative to the disaccharide (1.0). The resonance
corresponding to the glycal (6.30 ppm) was also apparent
1
in the H NMR spectrum with a relative integration of
0.4. The experiment was repeated with larger quantities
of the PS-OH, in an effort to increase the concentration
of nucleophilic groups capable of trapping the anhydro-
sugar intermediate. However, increasing the molar ratios
of PS-OH to glycal to 10:1 as well as DMDO (4.5 molar
excess), did not drive the reaction to 100% completion
nor did it dramatically increase glycal removal (65%
arrested glycal). The aldehyde and glycal were still
We reasoned that a similar strategy would be effective
in removing the glycal from solution so long as the
1
apparent in the H NMR spectrum. Nonetheless, based
1
upon H NMR integrations (1.7:1 to 0.4:1 glycal:disac-
(11) Barrett, A. G.; Hopkins, B. T.; Kobberling, J. Chem. Rev. 2002,
102, 3301-3324.
charide), ∼76% of the glycal had reacted with DMDO.
Moreover, the disaccharide remained untouched by the
mild reaction conditions as evidenced by the relative
(12) Harned, A. M.; Mukherjee, S.; Flynn, D. L.; Hanson, P. R. Org.
Lett. 2003, 5 (1), 15-18.
1
(13) Murray, R. W.; Jeyaraman, R. J. Org. Chem. 1980, 50, 2847-
2853.
cleanliness of the TLC plate and the H NMR spectra.
This initial attempt demonstrated the viability of the
phase-trafficking glycal removal process.
(14) Adam, W.; Hadjiarapoglou, L.; Wang. X. Tetrahedron Lett. 1991,
32 (10), 1295-1298.
J. Org. Chem, Vol. 70, No. 6, 2005 2389

SCHEME 6
FIGURE 3. Rearrangement of 1,2-anhydro sugar to furanose.
SCHEME 5
in the presence of the solution-phase nucleophile 28 for
2 h at -30 °C, Scheme 6. After rotoevaporation, 1H NMR
analysis of the crude mixture revealed the absence of
resonances corresponding to either 14 (6.30 ppm) or the
aldehyde (9.41 ppm). The spectrum was much more
complicated than before, which was primarily due to the
fact that eight new norbornene methyl glycosides 30 had
been produced as three new stereogenic centers resulted.
The presence of 30 was also indicated by the appearance
of new resonances between 5.5 and 6.2 ppm. Most
importantly, resonances corresponding to the anomeric
protons of 22 were also observed. This experiment clearly
demonstrated that a solution-phase alcohol was much
more efficient at scavenging 26 than PS-alcohols.
After copolymerization of the norbornene methyl gly-
cosides 30 with 28 using Grubbs generation I catalyst, a
gelatinous film was deposited. Heterogeneous extraction
of the gelatinous mass with boiling heptane provided in
three 10 mL fractions 22 tainted with Grubbs catalyst
and small impurities. Following flash chromatography,
22 was obtained in 80% yield from the crude phase-
trafficked material.
In summary, mannosyl iodide 5 has been efficiently
utilized in the construction of disaccharides 15 and 22.
The glycosidations are highly stereoselective cleanly
presenting the dissacharides in excellent yields; but these
reactions are plagued by the formation of glycal 14, which
is not readily separated. To overcome this difficulty, a
scavenging protocol that depends on selective epoxidation
of the glycal followed by nucleophilic attack was devel-
oped. A solution-phase nucleophilic scavenger 28 proved
more effective toward scavenging epoxidized glycal than
a comparable polymer-supported nucleophile PS-OH. The
phase trafficking strategy, coupled with highly stereo-
selective glycosidations provides a foundation for future
studies directed toward automated solution-phase syn-
thesis of branched-chain mannose oligomers.
Departing from the polymer-bound nucleophile, we
sought to employ norbornene-2-methanol 28 as a more
viable solution-phase scavenger, Scheme 5. Compound
28 and its derivatives have been successfully subjected
to ring opening metatheses polymerization (ROMP) in
the presence of Grubbs or Schrock’s catalyst to form
ROMPgels.11,15 ROMPgels are fluidic supports11,16,17 that
can be precipitated from various solvent mixtures de-
pending upon the nature of the polymer. Numerous
reagents and reactants immobilized onto ROMPgels have
been employed in a myriad of chemical transformations
including synthesis of oxazoles18 and homoallylic alco-
hols,19 scavenging of amines,20 and Horner-Emmons
olefinations.11,21 Mindful of these developments, we en-
visioned the immediate trapping of 26 by 28 and subse-
quent ROMP to provide a polymer (29) that could be
precipitated leaving behind only 22.
We first treated 28 with a 10-fold molar excess of
DMDO and no reaction was seen as monitored by TLC
after 5 h at -30 °C. The norbornene olefin was left
untouched even following reaction for an additional 10
1
h at room temperature as determined by TLC and H
NMR analysis. This result alleviated concerns that 28
might interfere with the formation of 26. We next
subjected a mixture of 22 and 14 to oxidation with DMDO
Acknowledgment. This work was supported by
National Science Foundation CHE-0196482. NSF CRIF
program (CHE-9808183), NSF Grant OSTI 97-24412,
and NIH Grant RR11973 provided funding for the NMR
spectrometers used on this project.
(15) Buchmeiser, M. R. Chem. Rev. 2000, 100 (4), 1565-1604.
(16) Bazan, G. C.; Schrock, R. R.; Cho, H. N. Gibson, V. C.
Macromolecules 1991, 24, 4495-4502.
(17) Buchmeiser, M. R.; Wurst, K. J. Am. Chem. Soc. 1999, 121 (48),
11101-11107.
(18) Barrett, A. G. M.; Cramp, S. M.; Hennessy, A. J.; Procopiou, P.
A.; Roberts, R. S. Org. Lett. 2001, 3 (2), 271-273.
(19) Arnauld, T.; Barrett, A. G. M.; Seifried, R. Tetrahedron Lett.
2001, 42 (44), 7899-7901.
(20) Arnauld, T.; Barrett, A. G. M.; Cramp, S. M.; Roberts, R. S.;
Zecri, F. J. Org. Lett. 2000, 2 (17), 2663-2666.
(21) Barrett, A. G. M.; Cramp, S. M.; Roberts, R. S.; Zecri, F. J. Org.
Lett. 1999, 1 (4), 579-582.
Supporting Information Available: Experimental pro-
cedures for the synthesis of 5, 6, 15, 20, and 22. This material
is available free of charge via the Internet at http://pubs.ac-
s.org.
JO0478609
2390 J. Org. Chem., Vol. 70, No. 6, 2005