Strategy in oligosaccharide synthesis: An application to a concise total synthesis of the KH-1(adenocarcinoma) Antigen

Article abstract of DOI:10.1021/ja9725864

A concise and potentially practical synthesis of the title compound has been achieved. The route features a high degree of convergence and economy of synthetic operations. A key step is the concurrent introductory addition of three α-1-fucosyl residues at

Full text of DOI:10.1021/ja9725864

1600
J. Am. Chem. Soc. 1998, 120, 1600-1614
Strategy in Oligosaccharide Synthesis: An Application to a Concise
Total Synthesis of the KH-1(adenocarcinoma) Antigen
Prashant P. Deshpande,†,∇ Hyunjin M. Kim,†,‡ Andrzej Zatorski,† Tae-Kyo Park,†
,|
Govindaswami Ragupathi,§ Philip O. Livingston,§ David Live, and
Samuel J. Danishefsky*,†,‡
Contribution from the Laboratories for Bioorganic Chemistry, Tumor Vaccinology, and
Nucleic Acid and Protein Structure, Sloan-Kettering Institute For Cancer Research,
1275 York AVenue, New York, New York 10021, and The Department of Chemistry,
Columbia UniVersity, New York, New York 10027
ReceiVed July 29, 1997
Abstract: A concise and potentially practical synthesis of the title compound has been achieved. The route
features a high degree of convergence and economy of synthetic operations. A key step is the concurrent
introductory addition of three R-L-fucosyl residues at required hydroxyl acceptor sites (see 37 f 39).
Conjugation to carrier protein was achieved, and a route to include truncated structures for investigations for
antibody specificity was accomplished.
Background
associated with isolating small amounts of elaborate glycocon-
jugate structures from tissue collections of patients. Moreover,
retrieval of the intact carbohydrate lattice by severing its covalent
attachment to bioconjugating molecules (proteins or lipids) is
only rarely possible.
Many tumors are characterized by the appearance of large
and unusual oligosaccharide subtypes.1,2 These structures tend
to emerge covalently bound to proteins in the form of cell
surface glycoproteins. Alternatively, the anomalous carbohy-
drates can be encountered as glycolipids, adhering to cell
surfaces through attractive molecular forces rather than via
classical covalent bonds. The isolation, immunocharacterization,
and structural identification of large carbohydrate-based tumor
antigens constitutes a highly complex and challenging undertak-
ing.
The possibility of exploiting these tumor-associated carbo-
hydrate ensembles to provoke productive immune responses to
the transformed state has occurred to glycobiologists, im-
munologists, and clinicians for some time.3,4 Progress along
these lines had to await breakthroughs in isolation, and
purification techniques, as well as in the ability to assign
connectivity and stereochemistry of carbohydrate domains
through spectroscopic methods. Fortunately, major advances
in these areas have provided a series of interesting structures
(proposed with varying degrees of rigor) that might form the
basis of strategies to achieve active immunity. However, a
complicating feature, not easily overcome, is the difficulty
These complexities pose large challenges as well as important
opportunities for organic synthesis. There is, first, a great need
for synthesizing the epitope structures themselves. However,
superimposed on this goal is the challenge of delivering the
tumor antigens in a favorable molecular framework for eliciting
therapeutically useful, active immunity. The ultimate success
of the chemistry end of the enterprise, then, is measured not
only by the attainment of the total synthesis of the epitope sector,
but also by the incorporation of this epitope into carrier protein.
The total construct becomes the subject of immunological
evaluation as part of a coherent vaccinology program.
The strategy and methodology which we have been employ-
ing for synthesizing the oligosaccharide epitopes are conve-
niently grouped under the term “glycal assembly”.5 The logic
of the method has been amply reviewed. The employment of
glycal assembly in the construction of epitope structures
including the follow-up bioconjugation and mouse immunization
studies have resulted in a vaccine (Globo-H)6,7 which is currently
undergoing advanced clinical evaluation. Two other fully
synthetically derived vaccines are at the stage of advanced,
preclinical processing.
† Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute For
Cancer Research.
‡ Columbia University.
§ Laboratory for Tumor Vaccinology, Sloan-Kettering Institute For
Cancer Research.
In this paper, we deal with what is perhaps the most
formidable carbohydrate-based tumor antigen thus far character-
Laboratory for Nucleic Acid and Protein Structure, Sloan-Kettering
Institute For Cancer Research.
(5) Bilodeau, M. T.; Danishefsky, S. J. Angew. Chem., Int. Ed. Engl.
1996, 35, 1380.
∇ Current address: Bristol Meyers Squibb, Princeton, NJ 08543.
| Current address: The Department of Biochemistry, University of
Minnesota, Minneapolis, MN 55455.
(6) (a) Park, T.-K.; Kim, I. J.; Hu, S.; Bilodeau, M. T.; Randolph, J. T.;
Kwon, O.; Danishefsky, S. J. J. Am. Chem. Soc. 1996, 118, 11488. (b)
Bildoeau, M. T.; Park, T.-K.; Hu, S.; Randolph, J. T.; Danishefsky, S. J.;
Livingston, P. O.; Zhang, S. J. Am. Chem. Soc. 1995, 117, 7840. (c) Kim,
I. J.; Park, T.-K.; Hu, S.; Abrampah, K.; Zeng, S.; Livingston, P. O.;
Danishefsky, S. J. J. Org. Chem. 1995, 60, 7716.
(7) Ragupathi, G.; Park, T.-K.; Zhang, S.; Kim, I. J.; Graber, L.; Adluri,
S.; Lloyd, K. O.; Danishefsky, S. J.; Livingston, P. O. Angew. Chem., Int.
Ed. Engl. 1997, 36, 125.
(1) Hakomori. S. Cancer Res. 1985, 45, 2405. Feizi. T. Cancer SurVeys
1985, 4, 245; Lloyd, K. O. Am. J. Clin. Pathol. 1987, 87, 129. Lloyd, K.
O. Cancer Biol. 1991, 2, 421.
(2) (a) Hakomori, S. Cancer Cells 1991, 3, 461. (b) Hakomori, S. AdV.
Cancer Res. 1989, 52, 257.
(3) Toyokuni, T.; Singhal, A. K. Chem. Soc. ReV. 1995, 24, 231.
(4) Hakomori, S.; Zhang, Y. Chem. Biol. 1997, 4, 97.
S0002-7863(97)02586-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/04/1998

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1601
Figure 1.
ized. The KH-1 antigen (1; see Figure 1)8 was isolated from
human colonic adenocarcinoma cells by using antibodies
generated against the classical Ley determinant. System 1, in
the form of glycolipid conjugates, was found to be present on
the cell surface of all adenocarcinoma cells thus far studied.
Furthermore, its presence has never been detected in normal
colonic extracts.
product synthesis, the basic building blocks are rather restricted
and tend to bear obvious homology with the readily recognized
components of the target system. Nonetheless, there are
considerable opportunities for the realization of formats, which
might lead to major synthetic economies. Thus, strategy in
oligosaccharide synthesis tends to focus on programs for optimal
conciseness in identifying specific centers for glycosylation, and
Monoclonal antibodies9a were raised against this antigen and
found to bind specifically to compound 1. On the basis of these
studies, Hakomori et al.9b postulated that the KH-1 antigen is a
highly specific marker for malignancy and premalignancy
involving colonic adenocarcinoma. Recently, an X-ray crystal
structure of an antitumor antibody BR96 in complex with the
nonanoate ester derivative of Ley tetrasaccharide was reported
by Jeffrey et al.10 The view of the antibody-Ley complex
provided by this determination suggested that the BR96 antibody
has unused binding capacity which might also recognize
structures larger than the Ley tetrasaccharide (such as the KH-1
antigen).
The difficulties associated with isolation and separation of
complex carbohydrates from human colonic cancer tissue have
been such that compound 1, either as a glycolipid or as a protein
conjugate, has not been available for evaluation. Thus, chemical
synthesis could offer a viable alternative to produce workable
quantities of complex systems such as the KH-1 antigen (1).
Our interests were not limited to KH-1 (1), but included
congeners11 which would be bioconjugated to the appropriate
protein carrier systems. Below, we describe (vide infra) the
total synthesis of the KH-1 antigen 1.12 We also describe a
protocol for reaching a suitable bioconjugatable analogue, as
well as a truncated heptasaccharide congener. In addition, we
also relate the upgrading of the KH-1 epitope for conjugation
to carrier protein.
for the attainment of stereocontrol in these couplings.
A
cardinal strategic objectiVe in oligosaccharide assembly is that
of gaining maximum relief from blocking group manipulations.
From these perspectives, we came to favor a plan that would
build a hexasaccharide (cf. structure 4), so differentiated in terms
of its protecting patterns as to allow for the simultaneous
unveiling of the three free hydroxyls destined for fucosylation
at a strategic point of our choosing. The three fucosylations
would then be conducted concurrently.
Broadly speaking, our plan called for inclusion of three kinds
of blocking groups in structure 4 (see Figure 2). The two
nitrogen centers carry special R functions. The three oxygen
centers to be fucosylated carry unique blocking groups (R*).
The remaining hydroxyls are protected by a general blocking
group (P). The three unique protecting groups would be cleaved
in a single operation (4 f 3). Hopefully, the three im-
munologically defining R-L-fucose entities could be introduced
in one concurrent synthetic operation (see structure 2). The
terminal glycal in 2 would be used to provide access to the
native KH-1 antigen (1) or to bioconjugates, en route to
evaluatable antiadenocarcinoma vaccines.
Considerable thought was also directed to assembling the
hexasaccharide with minimal protection-deprotection maneu-
vers which would still be consistent with sound management
of the complex network of hydroxyl groups. Toward this end,
we found advantages in drawing from principles which have
come forth from the logic of glycal assembly.5 Thus, differenti-
ated glycal types 6 and 7 (see Figure 3) were to be derived
from D-glucal by exploiting reliable reactivity preferences of
the C6, C3, and C4 hydroxyls (C6 > C3 > C4).13 Moreover,
the clean fashioning of an R-epoxide from appropriate galactal
derivatives (cf. 5) is well-known. Also well-known is the
excellent â-galactosyl-donating capacity of properly chosen
epoxides.14
Coupling of such a galactal-derived epoxide (5) to acceptors
6 and 7 would give rise to lactal-related disaccharides 8 and 9,
respectively. The C2′ hydroxyl of the lactal derivative 8 could
be uniquely protected with R* in anticipation of future fuco-
sylation at this site (see 10 bearing two unique R* blocking
groups). In another arm of the effort, the same hydroxyl groups
of 8 could be protected with a standard blocking (P) group to
Synthetic Planning
In conducting this project, we also hoped to address the
important issue of “strategy” in oligosaccharide synthesis. Of
course, in this field, as opposed to “conventional” natural
(8) Nudelman, E.; Levery, S. B.; Kaizu, T.; Hakomori, S.-I. J. Biol. Chem.
1986, 261, 11247.
(9) (a) Kaizu, T.; Levery, S. B.; Nudelman, E.; Stenkamp, R. E.;
Hakomori, S.-I. J. Biol. Chem. 1986, 261, 11254. (b) Kim, S. Y.; Yuan,
M.; Itzkowitz, S. H.; Sun, Q.; Kaizu, T.; Palekar, A.; Trump, B. F.;
Hakamori, S.-I. Cancer Res. 1986, 46, 5985.
(10) Jeffery, P. D.; Bajorath, J.; Chang, C. Y.; Dale, Y.; Hellstrom, I.;
Hellstrom, E. K.; Sheriff, S. Nat. Struct. Biol. 1995, 2, 456.
(11) (a) Helland, A.-C.; Nilsson, M.; Norberg, T. J. Carbohdyr. Chem.
1992, 11, 77. (b) Windmuller, R.; Schimidt, R. R. Tetrahedron Lett. 1994,
35, 7927.
(12) (a) For a communication of parts of this work see: Deshpande, P.
P.; Danishefsky, S. J. Nature 1997, 387, 164. (b) Another synthesis of
nonconjugated KH-1 antigen was reported while our manuscript (ref 12a)
was being reviewed: Hummel, G.; Schmidt, R. R. Tetrahedron Lett. 1997,
38, 1173.
(13) (a) Danishefsky, S. J.; Gervay, J.; Peterson, J. M.; McDonald, F.
E.; Koseki, K.; Griffith, D. A.; Oriyama, T.; Marsden, S. P. J. Am. Chem.
Soc. 1995, 117, 1940. (b) Danishefsky, S. J.; Behar, V.; Randolph, J. T.;
Lloyd, K. O. J. Am. Chem. Soc. 1995, 117, 5701.

1602 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
Figure 2.
afford 11. Thus, segment 11 bears one unique blocking group
(R*) anticipating fucosylation amidst more robust blocking
groups (P). In some fashion (vide infra), lactal derivative 9
would be so arranged as to function as a glycosyl acceptor at
C3′ (see system 12).
benzylation of the previously reported 15.17 As has been
demonstrated many times,5 the cyclic carbonate blocking group,
engaging C3 and C4 of a 6-monoprotected galactal, has a very
powerful R-directing effect in the epoxidation reaction, and a
strong â-directing effect in the use of such an epoxide for
galactosylation. These findings proved to be applicable to this
synthesis (vide infra).
Azaglycosylative coupling6b,15 of 12 with 11 would provide
tetrasaccharide 13. In a manner to be discussed, the C3
hydroxyl of the remote galactose of 13 would be identified as
an azaglycosyl acceptor site (see 14). Coupling of 14 at this
site with an azaglycosyl donor derived from 10 would afford
4, thus feeding back in to the prospectus adumbrated in Figure
2.
Needless to say, considerable trial and error was to be
necessary before translating this program to the realm of
practice. An account of the manner in which implementation
was accomplished in a highly concise fashion is described
below.
The specific glucals corresponding to formal structures 6 and
7 were 18 and 20, respectively. The dibenzyl derivative 18
had been reported from our laboratory by controlled dibenzy-
lation of D-glucal.13b The preparation of 20 involved a readily
achieved mono-triethylsilylation of 19. Compound 19 had been
reported in the literature in very low yield by the mono-
benzylation of D-glucal at C6.18 An improved but still unop-
timized preparation of 19 was accomplished as part of our
studies. With the required monosaccharide structures well in
hand, glycal assembly could commence. Thus, coupling of 17
with 20 occurred under mediation by zinc chloride, affording
21 in 65% yield.
Discussion and Results
The specific galactal derivative corresponding to formal
structure 5 was the epoxide 17. This compound was readily
and stereospecifically obtained by action of dimethyldioxirane16
on 16 (see Scheme 1). Compound 16 was synthesized through
The lone hydroxyl group at C2′ of lactal product 21 was
protected as its acetate (see compound 22). In terms of our
overall logic, this acetyl group serves as a general protecting
(14) Gervay, J.; Peterson, J. M.; Oriyama, T.; Danishefsky, S. J. J. Org.
Chem. 1993, 58, 5465.
(15) Griffith, D. A.; Danishefsky, S. J. J. Am. Chem. Soc. 1990, 112,
5811.
(17) Randolph, J. T.; McClure, K. F.; Danishefsky, S. J. J. Am. Chem.
Soc. 1995, 117, 5712.
(18) (a) Blackburne, I. D.; Burlitt, A. I. R.; Fredericks, P. F.; Guthrie,
R. D. Aust. J. Chem. 1976, 29, 381. (b) Fischer, S.; Hamann, C. H. J.
Carbohydr. Chem. 1995, 14 (3), 327.
(16) Murray, R. W.; Jeyaraman, R. J. J. Org. Chem., 1985, 50, 2847.

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1603
Figure 3.
group, P. In parallel fashion, and according to plan, the same
C2′ hydroxyl in compound 21 could be converted as its
triethylsilyl ether (see compound 23). It will thus be recognized
that these simply deriVed lactal deriVatiVes 22 and 23 already
carried the three “unique” (triethylsilyl) functions at the sites
anticipated for 3-fold fucosylation.
in the field of sialic acid acceptor sites19 to suggest that a
nondifferentiated vicinal triol of this type might be reasonably
selective or even specific for reaction at C3′. To investigate
this scenario for conciseness, compound 24 was subjected to
basic conditions which served to cleave the cyclic carbonate,
generating the potential triol acceptor 25 (see Scheme 2).
Addition of N-iodobenzenesulfonamide to compound 22, under
the usual conditions, gave rise to the iodosulfonamide derivative
26. Rearrangement of the sulfonamido function with concurrent
thiolation was accomplished through the action of lithium
hexamethyldisilazane in the presence of ethanethiol on com-
pound 26. This treatment was followed by acetylation (to
restore any C3′ alcohol to the acetate state), thereby affording
compound 27. Thus, we had in hand the two components
necessary to produce a tetrasaccharide.
In the event, glycosylation did occur at primarily C3′. Union
of 25 and 27, under the agency of methyl triflate in the presence
of di-tert-butylpyridine, afforded compound 28 as the major
product (see Scheme 3). Although this route constituted an
extremely concise synthesis of tetrasaccharide 28, the yield of
this coupling event was problematic. The maximum yield
obtained was 55%. More typical yields were in the range of
35-45%. Closer examination of the highly polar region of the
chromatogram revealed an isomeric tetrasaccharide from a
structure we tentatively assign as 29 (isolated in a ratio of 3:1-
Similarly, couplings of dibenzyl glucal acceptor 18 with
galactosyl donor 17 afforded a 55% yield of 24. With
disaccharide building blocks 22, 23, and 24 in hand, the program
for reaching a hexasaccharide system, generalized as 4, was
carried forward.
We first turned to the merger of systems derived from 22
and 24. In the latter case, it would be necessary to identify a
specific glycosylation acceptor site at C3′ of a substituted
galactal related to 24. It will be recalled (see generalized
structure 12 in Figure 3) that we had deliberately not specified
the status of the neighboring oxygens at C2′ and C4′ in the
galactose segment of the lactal (for 12, R ) protecting group
or R ) H). At first glance, it would seem to be necessary to
develop specific protection at C2′ and C4′ of this acceptor
moiety in order to pinpoint C3′ as the azaglycosyl acceptor site,
and we were certainly prepared to pursue possibilities along
these lines. A more exciting alternative view of the problem
presented itself. Perhaps eVen if C2′, C3′, and C4′ were all
unprotected hydroxyl groups, the effectiVe glycosyl acceptor site
would be at C3′. Certainly, there was encouraging precedent
(19) (a) Kameyama, A.; Ishida, H.; Kiso, M.; Hasegawa, A. J. Carbohydr.
Chem. 1991, 10, 549. (b) Kondo, T.; Tomoo, T.; Abe, H.; Isobe, M.; Goto,
T. Chem. Lett. 1996, 5, 337.

1604 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
Scheme 1a
a Reagents: (a) NaH, DMF, BnBr (85%); (b) 3,3-dimethyldioxirane, CH2Cl2; (c) LHMDS, BnBr, DMF (30-40%); (d) TESCl, imidazole, CH2Cl2
(76%); (e) 20, ZnCl2, THF (65%); (f) 18, ZnCl2, THF (55%); (g) TESOTf, Et3N, CH2Cl2 (92%); (h) Ac2O, Et3N, DMAP, CH2Cl2 (85%).
Scheme 2a
a Reagents: (a) K2CO3, MeOH (80%); (b) I(coll)2ClO4, PhSO2NH2, 4 Å MS, CH2Cl2 (81% for 26, 92% for 32); (c) LHMDS, EtSH, DMF (91%
for 33); (d) (only for 27) Ac2O, py, CH2Cl2 (95% over 2 steps).
2:1 favoring 28). The massive polarity differences (Rf ) 0.3
for 28 and 0.03 for 29 in 1:1 EtOAc-hexanes) between 28 and
29 presumably reflects different accessibilities of the two alcohol
linkages to the silica surface in the two compounds. Acetylation
of presumed 29 provided a triacetate assigned as 30. Thus, the
presumption that the C3′ hydroxyl in 27 would function as a
fully regiospecific acceptor site turned out to be optimistic.
However, synthetically useful selectivity in this reaction was
achieved. While this route was certainly very concise in that it
avoided the need for specific blocking functions at carbons 2′
and 4′ of the lactal derived from 24, the formation of significant
amounts of isomeric tetrasaccharide was cause for concern.
We hoped to improve upon the regiochemistry of this reaction
by employing another glycosylation protocol that we had first
demonstrated in a context of the synthesis of sialyl Lex
derivatives.20 To implement this idea, we returned to the
iodosulfonamide 26 with the hope that it itself might function
as the eventual lactosamine donor. The glycosyl acceptor
molecule would once again be compound 25. However, here,
the reaction would be mediated by bis(tri-n-butyltin)oxide in
the presence of silver tetrafluoroborate. This reaction did,
indeed, lead to an 84% yield of the previously encountered diol
(20) Danishefsky, S. J.; Koseki, K.; Griffith, D. A.; Gervay, J.; Peterson,
J. M.; McDonald, F. E.; Oriyama, T. J. Am. Chem. Soc. 1992, 114, 8331.

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1605
Scheme 3a
a Reagents: (a) MeOTf, di-tert-butylpyridine, Et2O/CH2Cl2 (2:1), 4 Å MS (3:1-2:1 28/29, 35-55% for 28); (b) Ac2O, Py, DMAP, CH2Cl2
(95%); (c) (i) (Bu3Sn)2O (2.2 equiv), PhH, 25 (4 equiv), reflux (ii) AgBF4, THF, 4 Å MS (84%).
Scheme 4a
a Reagents: (a) K2CO3, MeOH (85%); (b) (Bu3Sn)2O, benzene, reflux; (c) 33, MeOTf, di-tert-butylpyridine, Et2O/CH2Cl2 (2:1), 4 Å MS (60%);
(d) 32, AgBF4, THF, 4 Å MS (62%); (e) Ac2O, Pyr, DMAP, CH2Cl2 (>95%); (f) TBAF/AcOH (93%).
tetrasaccharide 28. While this direct “rollover” constituted a
very pleasing and attractive solution to the assembly of 28, this
method is also not without its shortcomings in that it required
recourse to excesses (4 equiv) of glycosyl acceptor 25 for high-
yield coupling. In the absence of excess acceptor, it was
difficult to drive the process to completion. On the other hand,
the recovery of unused glycosyl acceptor, after the tin-mediated
methodology, was very high (ca. 60-90%).
In summary on this point, the formation of tetrasaccharide
28 can be conducted through two methods. The reaction of
donor 27 with acceptor 25, in near stoichiometric equivalency,
results in regioselective, but nonspecific, glycosylation at the
desired C3′ site, giving rise to ca. 35-55% yields of tetrasac-
charide. By contrast, recourse to “direct rollover” methodology,
employing the tributylstannylated derivative of acceptor 25,
incurs the disadvantage of requiring significant excesses of
acceptor. Since the recovery of the unreacted acceptor was high,
we are currently favoring the direct method.
It was hoped that the glycosyl acceptor site of this pentaol
would occur at the C3 hydroxyl group of the terminal galactose
residue. The rationale was that the cluster of three neighboring
and unprotected hydroxyl groups in the D ring in 31 would
provide a greater locus of activity than the C2′ and C4′
hydroxyls of the B ring, each of which is flanked by two
extensively substituted oxygens. Thus, the C2′ hydroxyl of the
B ring is surrounded by the glycosidic bond to the A ring glucal
as well as the glycosidic bond to the C ring glucosamine
analogue. Similarly, the C4′ hydroxyl of the B ring abuts a
benzyl ether protecting group at C6′ and is vicinal to a
disaccharide group projecting from the oxygen from C3′.
Once again, we elected to explore the use of a thioethyl donor
type to be derived from compound 23. Toward this end,
iodosulfonamidation of 23 afforded 32 which, under thiolate-
mediated rearrangement of the sulfonamido group, gave rise to
33 (see Scheme 2). The coupling reaction between 28 and 33
was performed by the mediation of methyl triflate in the
presence of di-tert-butylpyridine. Fortunately, in this case, the
coupling could well be executed in acceptable yield (40-60%)
without detectable formation of side product arising from
glycosylation at alternative acceptor sites. In that sense, the
situation was far better than in the 2 + 2 coupling of 25 + 27,
conducted through the same experimental protocol. However,
we still experienced difficulties in obtaining reproducible yields.
The next plateau to be reached was seen to be a hexasac-
charide corresponding to a type 4 compound (see Figure 2). To
further maximize the opportunities for conciseness, we were
prepared to assume another significant risk in the use of
undifferentiated hydroxyl groups in the acceptor site. The
thought was to convert cyclic carbonate acetate 28 to pentaol
31. This transformation was readily accomplished by treatment
of 28 with potassium carbonate in methanol as shown in Scheme
4.
For this reason, we elected again to examine the direct
rollover method, this time using the tributylstannylated acceptor

1606 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
Scheme 5a
The locations of the four acetate protecting groups were
determined unequivocally through the use of several NMR
experiments. The four multiplets between 5.0 and 5.45 ppm
fall in the distinctive region for the protons on O-acetylated
sugar carbons. The multiplets at 5.06 and 5.16 ppm were
assigned to the H-2 protons of either sugar ring B or ring D, on
the basis of their coupling patterns, and their cross peaks in the
double quantum COSY experiment (DQCOSY) to protons
assumed to be at anomeric carbons on the basis of chemical
shifts.
The confirmation that these signals indeed arise from ano-
1
meric protons followed from the H-13C correlation experi-
ments, where these resonances show cross peaks to carbons with
the distinct anomeric carbon shift.22 These H-2 protons, which
had now been identified, show DQCOSY cross peaks to one
other vicinal partner each, which must be the respective H-3’s
for the two pyranose rings. The peaks at 5.36 and 5.40 ppm
arise from the two other sites of O-acetylation, i.e., at the axial
hydroxyls at C4 of the rings B and D. That these centers carry
the remaining acetoxy groups can be surmised from the
correlation of these protons to carbonyl resonances in long-
a Reagents: (a) Sn(OTf)2, di-tert-butylpyridine, Tol/THF (10:1), 4
Å MS (60%).
1
range (HMBC) H-13C correlation experiments.
34 derived from pentaol 31 reacting with donor 32. Using a
4-fold excess of 31 (once again unused acceptor was recovered
in very high yield), a 62% yield of 35 was obtained. As before,
the direct method lent itself to much more reproducible
chemistry which compensated for the need to recover unused
acceptor. This is the method we currently follow.
We were now ready to address the culminating stage of the
plan, namely, the construction of the nonasaccharide. The
preparation for this purpose simply involved cleavage of the
unique (silyl) protecting groups (see system 3, Figure 2). We
had carefully prepared for the possibility of such a step by
carrying relatively dischargable triethylsilyl protecting groups
at the oxygens destined for fucosylation. Indeed, treatment of
compound 36 with buffered tetrabutylammonium fluoride
provided triol 37. As our L-fucosyl donor, we employed
compound 3821 which had been used several times before5 in
our program to synthesize blood group determinants and tumor
antigens.
The critical step of the synthesis, 3-fold fucosylation, was
indeed accomplished through the reaction of 37 and 38. A 60%
yield of nonasaccharide 39 was obtained (see Scheme 5).
Needless to say, we did not have available to us any substance
for direct comparison, or any downstream product which might
be used in a strategy to confirm the structure of 39. The
structure assignment would not be fully clarified until the
synthesis was concluded and even then without benefit of a
reference sample. Therefore, it was important to marshal
convincing evidence that we had indeed generated the compound
we were claiming, i.e., structure 39. Our strategy in this regard
was to return to compound 37. If its structure could be
established, we would be confident in concluding that the three
fucosyl monosaccharides had been introduced at the proper
positions. NMR analysis at the eight anomeric centers in the
nonasaccharide, in conjunction with the information we had
already accumulated at the hexa stage, would provide acceptable
support of the structure of the trifucosylated product 39.
The examination of the NMR and mass spectra of compound
37 indicated the presence of the expected four acetate groups.
Fortunately NMR experiments could be used to assign the
positions of these acetates. The sites of the free hydroxyl groups
would thus have been established by difference.
Analysis of the DQCOSY experiment showed that three
additional sites have cross peaks to respective H-3 protons
identified from connection to the H-2 resonances. These data
in the aggregate prove that the acetylation sites on sugar rings
B and D are on carbons 2 and 4. The data confirm that the
glycosylation linkages formed during the reaction between
components 25 and 27 (or 25 and 26) as well as between 31
and 33 (or 32 and 34) have occurred through acceptor hydroxyl
sites at C-3 of the respective residues. Using the markers for
rings B and D that come from the acetates, along with the
characteristic 1H and 13C shifts of the glycal, and data from the
long-range 1H-13C correlation (HMBC) experiment, it was
possible to trace the through bond connectivities around the
respective sugar rings and across the glycosidic linkages to
confirm the desired assembly sequence of the compound 37.
Also, all of the glycosidic linkages in compound 37 were
confirmed as â-linkages on the basis of the splitting in the proton
dimension of the individually resolved anomeric cross peaks
in the HMQC experiment. These exhibited a coupling constant
for the anomeric H-1 protons of 6.00 Hz or higher. It is on
these bases that we were confident of the assignments for
compounds 37 and 39. The latter would now be advanced to
our goal structures.
At this point, our remaining objectives were (i) the total
synthesis of the KH-1 antigen itself, (ii) the synthesis of a form
of the epitope which would be suitable for conjugation to carrier
program, and (iii) the synthesis of a truncated construct which
would help in ascertaining antibody specificities. In terms of
reaching goals (i) and (ii) we were now in a position starting
with the nonasaccharide glycal 39 to take advantage of previous
experiences in the syntheses of glycolipids, particularly those
in the Globo-H (breast tumor antigen) construct.6
Our first step in the proposed introduction of the ceramide
side chain to reach the natural KH-1 antigen (1) was the
epoxidation of glycal 39. While this reaction seemed to occur
smoothly, attempts to use the epoxide directly as a glycosyl
donor with respect to the well-known preceramide acceptor 4023
led to low yields of coupling. These difficulties perhaps arise
from dominant conformers of a nonasaccharide epoxide in which
required access to the anomeric oxido carbon by glycosyl
(21) Danishefsky, S. J.; Gervay, J.; Peterson, J. M.; McDonald, F. E.;
Koseki, K.; Oriyama, T.; Griffith, D. A.; Wong, C.-H.; Dumas, D. P. J.
Am. Chem. Soc. 1992, 114, 8329.
(22) Lerner, L.; Bax, A. Carbohydr. Res. 1987, 166, 35.
(23) Schmidt, R. R.; Zimmermann, P. Tetrahedron Lett. 1986, 27, 481.

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1607
Scheme 6a
a Reagents: (a) (i) 3,3-dimethyldioxirane, CH2Cl2; (ii) EtSH, CH2Cl2, H+(cat); (iii) Ac2O, Py, CH2Cl2 (60% 3 steps); (b) 40, MeOTf, Et2O/
CH2Cl2 (2:1), 4 Å MS (55%); (c) Lindlar’s catalyst, H2, palmitic anhydride, EtOAc (85%); (d) (i) Na0, NH3, THF; MeOH quench; (ii) Ac2O, Et2N,
DMAP, DMF, THF; (iii) MeONa, MeOH (70% 3 steps).
acceptor 40 was difficult. Given the chemical instability of
glycal epoxides to Lewis acids, relatively slow glycosylation
would result in low yields of glycoside, through loss of
competence of the donor moiety.
consistent NMR analysis which had been conducted on inter-
mediates en route to, and including, 39 (particularly compound
37). Also, a fully rigorous NMR-based verification was possible
at the stage of allyl glycoside 44, whose synthesis was
accomplished from the same 39 (vide infra).
Accordingly, we turned to the application of a recently
developed variation of the glycal epoxy donor method.24 This
protocol started with epoxidation, followed by thiolation of the
resulting epoxide of 39 and further acetylation in the usual way,
leading to acetate 41 (see Scheme 6). With expectation of
effective neighboring group participation of the C2 acetoxyl
function available to guide a â-glycosidation, compound 41 was
treated with 40 under the agency of methyl triflate. This process
indeed led to the formation of glycoside 42, in 55% yield.
From this point, the required methodology was quite familiar
to us.6 Reduction of the azide 42 was coordinated with
palmitoylation of the resultant amine generated, in situ (see
compound 43). Cleavage of all aromatic groups was apparently
accomplished by treatment of 43 with sodium in liquid
ammonia. The resultant product was quenched with methanol.
Under the basic conditions thus employed, all of the esters
(including the cyclic carbonate) were cleaved. To regularize
the blocking groups (vide infra), the resultant material was
peracetylated. Finally, cleavage of all acetates was accom-
plished with sodium methoxide in methanol to give rise to the
pure KH-1 antigen (1).
For our bioconjugation studies, we identified allyl glycoside
44 as our goal system. For this purpose we returned to the
glycal 39. Again, discharge of all the aromatic groups was
accomplished by treatment of compound 39 with sodium in
liquid ammonia (see Scheme 7). The remaining esters and the
cyclic carbonate were cleaved after the reaction was quenched
with methanol. The crude product was subjected to exhaustive
acetylation, followed by epoxidation in the usual way with
dimethyldioxirane. The resultant epoxide was opened by
solvolysis with allyl alcohol. The crude product was fully
deacylated through the action of sodium methoxide in methanol
to afford compound 44 in 60% overall yield.
Conjugation of 44 to carrier keyhole lympet hemocyanin
(KLH) protein was accomplished through ozonolysis of the
double bond of the allyl group. This treatment was followed
by reductive amination of the resultant aldehyde to KLH. The
procedure for accomplishing conjugation to KLH is detailed in
the Experimental Section.
In preparing for immunological investigations, it would be
helpful to determine the specificities of various antibodies to
the structural features of the KH-1 antigen. Toward that end,
it was of interest to generate truncated structures in which
segments of the molecule would be deleted. In our first such
effort, we directed our attention to a construct in which the three
fucose residues as well as the N-acetyl function on the E ring
would be retained (see compound 50). However, the terminal
N-acetylactosamine substructure would be deleted. For this
purpose, a modified synthesis was initiated. We returned to
The structure assignment of the fully synthetic material is
supported by a mass spectral measurement which confirmed
the molecular formula. More telling was the very close
correspondence of the NMR data measured on fully synthetic
1, with the fragmentary data previously reported for the naturally
occurring material. Most compelling was the extensive and self-
(24) Seeberger, P. H.; Eckhardt, M.; Gutteridge, C. E.; Danishefsky, S.
J. J. Am. Chem. Soc. 1997, 119, 10064.

1608 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
Scheme 7a
a Reagents: (a) (i) Na0, NH3, THF; (MeOH quench); (ii) Ac2O, Et2N, DMAP, DMF, THF; (b) (i) 3,3-dimethyldioxirane, CH2Cl2; (ii) allyl
alcohol; (c) MeONa, MeOH (60% 3 steps).
Scheme 8a
a Reagents: (a) K2CO3, MeOH (quant); (b) MeOTf, di-tert-butylpyridine, Et2O/CH2Cl2 (2:1), 4 Å MS (61%); (c) Ac2O, Py, DMAP, CH2Cl2
(93%); (d) 1:1 TBAF/AcOH (72%); (e) Sn(OTf)2, Tol/THF (10:1), 4 Å MS (39%); (f) (i) Na0, NH3, THF; MeOH quench; (ii) Ac2O, Et2N, DMAP,
THF, DMF; (iii) 3,3-dimethyldioxirane, CH2Cl2; (iv) allyl alcohol; (v) MeONa, MeOH (60% 3 steps).
compound 21 which would first serve as an acceptor substruc-
ture. Cleavage of the cyclic carbonate linkage gave rise to
compound 45 (see Scheme 8).
again, it was possible to conduct 3-fold fucosylation. This
chemistry was not optimized, and we accepted a 39% yield of
compound 49. The remaining steps for conversion of 49 to
allyl glycoside 50 ran parallel to those used with related
structures in the synthesis of compound 44.
With serviceable routes to the various tetrasaccharide, dis-
accharide, and monosaccharide moieties contained in the KH-1
construct, it seems likely that other permuted structures can be
readily assembled. These can be used to ascertain epitope
specificities for monoclonal antibodies which are being elicited
as part of this vaccine-oriented program.
The donor, to be coupled to compound 45, was the C1
thioethyl C2 phenylsulfonamide system 33, containing the two
unique TES protecting groups in the E and F rings. Coupling
of 45 with 33 occurred under the usual conditions to give an
acceptable 60% yield of the tetrasaccharide 46. Acetylation of
46 afforded a diacetate, 47. This system now contained the
three uniquely identified triethylsilyl ether functions which were
cleaved through the action of TBAF (see compound 48). Once

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1609
(film) 3030, 2875, 1797, 1647, 1496, 1453, 1371, 1244, 1164, 1110,
1010, 837, 699 cm-1; 1H NMR (400 MHz, CDCl3) δ 3.7-3.9 (m, 2H,
H-6), 4.08 (br t, 1H, J ) 7.4 Hz, H-5), 4.58 (s, 2H, -CH2Ar), 4.90 (d,
1H, J ) 7.8 Hz, H-4), 4.93 (br m, 1H, H-3), 5.14 (dd, 1H, J ) 3.2, 7.7
Hz, H-2), 6.66 (d, 1H, J ) 6.2 Hz, H-1), 7.28-7.45 (m, 5H, Ar-H);
13C NMR (400 MHz, CDCl3) δ 67.97, 68.74, 72.41, 73.14, 73.66, 97.97,
127.77, 127.93, 128.44, 137.18, 149.06, 153.98; HRMS (FAB) calcd
for C14H14O5‚Na (M + Na)+ 285.084, found 285.072.
6-O-Benzylglucal (19). To a solution of D-glucal (10.0 g, 68.4
mmol) in dry DMF (200 mL) was added LHMDS (1.0 M solution in
THF, 75.3 mL, 1.1 equiv) dropwise at -40 °C, followed by BnBr (8.18
mL, 68.4 mmol). The solution was stirred mechanically for 6 h,
allowing the temperature to rise to 0 °C. Saturated solution of NH4Cl
(50 mL) was added to the reaction mixture, followed by EtOAc (200
mL). The organic layer was separated, and the aqueous layer was
extracted with EtOAc (3 × 50 mL). Combined organic layers were
washed with brine (50 mL) and water (50 mL), dried with MgSO4,
filtered, concentrated, and submitted for column chromatography (1:1
EtOAc/hexanes) to obtain compound 19 as a syrup (4.80 g, 30%):
[R]23D ) +11.0° (c 1.0, CHCl3); FTIR (film) 3342, 2871, 1642, 1656,
1231, 1101, 1027, 851, 738 cm-1; 1H NMR (400 MHz, CDCl3) δ 3.6-
3.85 (m, 5H,), 4.06 (d, 1H, J ) 4.0 Hz, -OH), 4.11 (br t, 1H, H-3),
4.46 (d, 1H, J ) 12.0 Hz, -CH2Ar), 4.52 (d, 1H, J ) 12.0 Hz, -CH2-
Ar), 4.57 (dd, 1H, J ) 1.8, 6.0 Hz, H-2), 6.21 (d, 1H, J ) 6.0 Hz,
H-1), 7.15-7.35 (m, 5H, Ar-H); 13C NMR (400 MHz, CDCl3) δ 69.06,
69.63, 70.28, 73.427, 76.95, 102.72, 127.29, 127.55, 128.21, 128.24,
137.60, 137.75, 143.86.
Summary
All of the chemical goals identified at the outset of this project
have been realized. The KH-1 antigen (1) has been synthesized.
It has also been synthesized in conjugatable form for delivery
to carrier protein (see compound 44). Conjugation to KLH has,
in fact, been achieved (see the Experimental Section). A format
for synthesizing truncated versions of the KH-1 antigen in order
to establish epitope specificities for various monoclonal antibod-
ies being harvested has also been achieved (see compound 50).
From a strategic standpoint, the strategy of 3-fold fucosylation
to achieve conciseness in the synthesis was certainly vindicated.
Also, the synthesis speaks forcefully to the advantages of glycal
assembly and to the minimization of protecting group manipula-
tions for purposes of differentiating functional groups. Thus,
in three instances (see compounds 25, 31, and 45) we took
recourse to glycosyl acceptors with more than one potential
glycosylation site. In the latter two cases, coupling was
accomplished under nearly stoichiometric conditions utilizing
1-â-thioethyl-2R-phenylsulfonamido donors under the agency
of methyl triflate. In the case of coupling of acceptors 25 and
31, these delicate reaction conditions led to some difficulties
in reproducing the best case yield. However, excellent yields
in a highly reproducible fashion were obtained by treatment of
the 1R-sulfonamido-2â-iodo donor type (see compounds 26 and
32), utilizing a stannylated version of acceptor (25 and 31).
These reactions, however, require a significant excess of
glycosyl acceptor. Since the recovery yields are very high, our
current technology favors this practice.
6-O-Benzyl-3-O-(triethylsilyl)glucal (20). To a solution of com-
pound 19 (4.00 g, 16.8 mmol) in dry CH2Cl2 (70 mL) were added
imidazole (1.26 g, 18.5 mmol) and DMAP (0.01 g), and the solution
was cooled to -78 °C. At -78 °C, to the reaction mixture was added
TESCl (3.10 mL, 18.5 mmol) dropwise, and the resulting solution was
allowed to warm to -40 °C. The reaction mixture was stirred for 4 h,
diluted with EtOAc (200 mL), and washed with water (2 × 100 mL),
a saturated solution of NaHCO3 (3 × 100 mL), and brine (10 mL).
The organic layer was dried (NaSO4), filtered, concentrated, and
submitted for column chromatography (1:9 EtOAc/hexanes) to provide
The primary focus of the KH-1 antigen problem has now
shifted to issues of immunology and vaccinology. Data on these
matters will soon be forthcoming.
Experimental Section
20 (4.46 g,76%) as a syrup: [R]22 ) +44.0° (c 1.0, CHCl3); FTIR
D
(film) 3468, 3030, 2953, 2875, 1644, 1453, 1237, 1086, 871, 737 cm-1
;
General Methods. All commercial materials were used without
further purifications unless otherwise noted. The following solvents
were distilled under positive pressure of dry nitrogen immediately before
use: THF and ether from sodium-benzophenone ketyl, and CH2Cl2,
toluene, and benzene from CaH2. All the reactions were performed
under N2 atmosphere. NMR (1H, 13C) spectra were recorded on a
Bruker AMX-400 MHz spectrometer, a Bruker Avance DRX-500 MHz
spectrometer, and a Varian 600 Unity plus 600 MHz spectrometer,
referenced to TMS (1H NMR, δ 0.00) or CDCl3 (13C NMR, δ 77.0)
and CD3OD( 13C NMR, δ 49.05) peaks unless otherwise stated. LB
) 1.0 Hz was used before Fourier transformation for all of the 13C
NMR. IR spectra were recorded with a Perkin-Elmer 1600 series FTIR
spectrometer, and optical rotations were measured with a JASCO DIP-
370 digital polarimeter using a 10 cm path length cell. Low- and high-
resolution mass spectral analyses were performed with a JEOL JMS-
DX-303 HF mass spectrometer. Analytical thin-layer chromatography
was performed on E. Merck silica gel 60 F254 plates (0.25 mm).
Compounds were visualized by dipping the plates in a cerium sulfate-
ammonium molybdate solution followed by heating. Flash column
chromatography was performed using the indicated solvent on E. Merck
silica gel 60 (40-63 µm) or Sigma H-Type silica gel (10-40 µm) for
normal phase and EM Science Lichroprep RP-18 (15-25 µm) for
reverse phase. Melting points are obtained with an Electrothermal
melting point apparatus (series no. 9100) and are uncorrected.
1H NMR (400 MHz, CDCl3) δ 0.55 (q, 6H, J ) 7.9 Hz, -SiCH2CH3),
0.88 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 2.47 (d, 1H, J ) 4.1 Hz, -OH),
3.6-3.75 (m, 3H, 2H-6, H-4), 4.13 (br d, 1H, J ) 6.4 Hz, H-3), 4.47
and 4.52 (2d, 2H, J ) 12.0 Hz, -CH2Ar), 4.55 (dd, 1H, J ) 2.2, 6.2
Hz, H-2), 6.21 (d, 1H, J ) 6.0 Hz, H-1), 7.10-7.40 (m, 5H, Ar-H);
13C NMR (400 MHz, CDCl3) δ 4.84, 6.66, 69.05, 69.64, 70.56, 73.47,
76.97, 103.44, 127.59, 127.64, 128.27, 137.78, 143.33; HRMS (FAB)
calcd for C19H30O4Si‚Na (M + Na)+ 373.180, found 373.181.
6,6′-Di-O-benzyl 3′,4′-Carbonate 3-O-Triethylsilyl Lactal Deriva-
tive 21. To a solution of compound 16 (2.99 g, 11.4 mmol) in dry
CH2Cl2 (20 mL) at 0 °C was added 3,3-dimethyldioxirane (300 mL,
0.0800 M solution in acetone). The reaction mixture was stirred at 0
°C for 1 h. The solvent was evaporated by N2, and further dried in
vacuo for 10 min. The resulting 1,2-anhydro sugar 17 was dissolved
in a solution of compound 20 (6.00 g, 17.1 mmol) in dry THF (30
mL), and at 0 °C to the resulting solution was added a 1.0 M solution
of ZnCl2 in ether (5.70 mL, 0.5 equiv). The reaction mixture was stirred
at room temperature for 24 h, diluted with EtOAc (50 mL), and washed
with a saturated solution of NaHCO3 (2 × 10 mL). The organic layer
was separated, dried (MgSO4), filtered, concentrated, and submitted
for chromatography (2:3 EtOAc/hexanes) to provide compound 21 (4.76
g, 66%) as a syrup: [R]22D ) -25.0° (c 1.0, CHCl3); FTIR (film) 3439,
3030, 2910, 1804, 1725, 1647, 1453, 1371, 1243, 1074, 847, 741 cm-1
;
6-O-Benzyl 3,4-O-Carbonate 16. To a solution of the galactal
carbonate derivative 15 (5.36 g, 34.4 mmol) in dry DMF (50 mL) at 0
°C was added benzyl bromide (12.3 mL, 103 mmol), followed by NaH
(60% oil dispersion, 1.50 g, 1.1 equiv). The reaction mixture was stirred
for 1 h, diluted with CHCl3 (50 mL), treated with brine solution (20
mL), and again stirred for 5 min. The organic layer was separated,
dried (MgSO4), and filtered. The filtrate was concentrated to afford a
syrup. Chromatography with 1:1 EtOAc/hexanes afforded 7.66 g of
compound 16 (85%) as a syrup: [R]23D ) -92.0° (c 1.0, CHCl3); FTIR
1H NMR (CDCl3, 400 MHz) δ 0.58 (q, 6H, J ) 8.0 Hz, -SiCH2CH3),
0.92 (t, 9H, J ) 8.0 Hz, -SiCH2CH3), 3.51 (d, 1H, J ) 2.8 Hz, -OH),
3.62 (ddd, 1H, J ) 2.8, 7.2, 7.2 Hz, H-2′), 3.65-3.75 (m, 3H), 3.85
(m, 1H), 3.93 (dd, 1H, J ) 4.9, 11.2 Hz), 3.99 (br t, 1H, J ) 5.3, 6.48
Hz), 4.09 (br m, 1H), 4.27 (br t, 1H, J ) 4.2 Hz), 4.48-4.68 (m, 6H,
-CH2Ar), 4.70 (dd, 1H, J ) 3.4, 6.2 Hz, H-2), 4.74 (dd, 1H, J ) 1.8,
7.2 Hz, H-4), 6.32 (d, 1H, J ) 6.0 Hz, H-1), 7.2-7.4 (m, 10H, Ar-
H); 13C NMR (500 MHz, CDCl3) δ 4.75, 6.67, 65.65, 67.79, 67.93,
70.42, 71.49, 73.43, 73.58, 74.46, 75.27, 75.42, 78.05, 99.94, 102.61,

1610 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
127.79, 127.85, 128.14, 128.33, 137.36, 137.53, 143.00, 153.96; HRMS
3,6,6′-Tri-O-benzyllactal (25). To a solution of compound 24 (3.00
g, 4.96 mmol) in MeOH (100 mL) was added a solution of sodium
methoxide (1.00 mL, 25% by weight in MeOH) dropwise. The reaction
mixture was stirred for 1 h. The volatiles were removed in vacuo.
The residual syrup obtained was quickly purified by column chroma-
tography (2.5% MeOH in EtOAc) to provide 2.68 g (91%) of 25 as a
syrup: [R]22D ) -14.0° (c 1.0, CHCl3); FTIR (film) 3415, 3029, 2867,
calcd for C33H44O10Si‚Na (M + Na)+ 651.170, found 651.368.
2′-O-Acetyl-6,6′-di-O-benzyl 3′,4′-Carbonate 3-O-Triethylsilyl
Lactal (22). To a solution of compound 21 (3.50 g, 5.50 mmol) in
CH2Cl2 (30 mL) were added Et3N (3 mL), Ac2O (3 mL), and a catalytic
amount of DMAP. The reaction mixture was stirred overnight, diluted
with EtOAc (50 mL), and washed with saturated solutions of CuSO4
(3 × 10 mL), water (1 × 10 mL), NaHCO3 (2 × 10 mL), and brine (1
× 10 mL) sequentially. The organic layer was separated, dried
(MgSO4), filtered, concentrated, and submitted to chromatography (1:1
EtOAc/hexanes) to provide 22 in quantitative yield: [R]23D ) -42.0°
(c 1.0, CHCl3); FTIR (film) 2954, 2875, 1809, 1755, 1646, 1454, 1222,
1060, 743 cm-1; 1H NMR (CDCl3) δ 0.58 (q, 6H, J ) 7.9 Hz, -SiCH2-
CH3), 0.92 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 2.06 (s, 3H, -COCH3),
3.63 (dd, 1H, J ) 2.9, 10.9 Hz, H-5), 3.70 (br d, 2H, J ) 10.5 Hz,
2H-6), 3.85 (dd, 1H, J ) 6.1 Hz, J ) 10.92, H-5′), 3.9-4.0 (m, 2H,
2H-6′), 4.10-4.2 (m, 2H), 4.5-4.6 (m, 4H, 2-CH2Ar), 4.64 (dd, 1H,
J ) 4.0, 8.0 Hz, H-3′), 4.71 (dd, 1H, J ) 4.2, 5.8 Hz, H-2), 4.84 (dd,
1H, J ) 1.0, 8.1 Hz, H-4′), 4.90 (d, 1H, J ) 4.6 Hz, H-1′), 4.99 (t, 1H,
J ) 4.2 Hz, H-4′), 6.30 (d, 1H, J ) 6.2 Hz, H-1), 7.15-7.40 (m, 10H,
Ar-H); 13C NMR (CDCl3) δ 4.78, 6.73, 20.58, 64.78, 67.94, 67.99,
69.38, 69.50, 73.19, 73.35, 73.79, 73.92, 74.56, 74.86, 96.82, 102.27,
127.60, 127.75, 127.78, 127.93, 128.29, 128.44, 137.35, 138.01, 142.99,
153.27, 168.54; HRMS calcd for C35H46O11Si‚Na (M + Na)+ 693.270,
found 693.273.
1
1647, 1453, 1246, 1068, 735 cm-1; H NMR (500 MHz, CDCl3) δ
3.48-3.56 (m, 2H), 3.62 (dd, 1H, J ) 4.8, 8.0 Hz), 3.66-3.78 (m,
3H), 3.91 (d, 1H, J ) 4.4 Hz), 3.97 (dd, 1H, J ) 4.0, 8.8 Hz), 4.18-
4.28 (m, 4H), 4.47 (s, 2H, -CH2Ar), 4.52 (d, 1H, J ) 8.0 Hz), 5.59 (s,
1H, -CH2Ar), 4.57-4.65 (m, 2H, -CH2Ar), 4.85 (dd, 1H, J ) 2.4,
4.8 Hz, H-2), 6.41 (d, 1H, J ) 4.8 Hz, H-1), 7.20-7.45 (m, 15H, Ar-
H); 13C NMR (400 MHz, CDCl3) δ 67.92, 68.86, 69.19, 69.82, 71.53,
73.35 (2C), 73.39, 73.42, 73.87, 76.26, 100.01, 103.30, 127.35, 127.42,
127.59, 127.74, 128.19, 128.29, 137.72, 137.81, 138.52, 144.57; HRMS
(FAB) calcd for C33H38O9Na (M + Na)+ 601.240, found 601.242.
Iodo Sulfonamide Disaccharide 26. To a solution of compound
22 (2.50 g, 3.72 mmol) with 4 Å molecular sieves (3.00 g) and
benzenesulfonamide (2.92 g, 18.6 mmol) in CH2Cl2 at 0 °C was added
a solution of I(sym-coll)2ClO4 [freshly prepared from Ag(coll)2ClO4
(8.36 g, 18.6 mmol) and I2 (4.53 g, 18.5 mmol)] in CH2Cl2 (40 mL)
via cannula. The ice bath was removed, and the reaction mixture was
stirred at room temperature for 1 h. The resulting suspension was
filtered through a pad of silica gel, and the filtrate was washed with a
saturated solution of Na2S2O3 (3 × 25 mL), followed by a saturated
solution of CuSO4 (5 × 25 mL) and H2O (2 × 10 mL). The separated
organic layer was dried (MgSO4), filtered, concentrated, and submitted
for column chromatography (5% EtOAc in CH2Cl2, start gradient) to
6,6′-Di-O-benzyl 3′,4′-Carbonate 2′,3-Bis(O-triethylsilyl) Lactal
(23). To a solution of the lactal 21 (3.00 g, 4.77 mmol) in dry CH2Cl2
(50 mL) was added Et3N (3.34 mL) followed by dropwise addition of
TESOTf (1.61 mL, 7.15 mmol) at 0 °C. The reaction mixture was
stirred for 3 h, and washed with a saturated solution of NaHCO3 (2 ×
15 mL). The organic layer was separated, dried (MgSO4), filtered,
concentrated, and submitted for chromatography (1:4 EtOAc/hexanes)
to provide 23 (3.27 g, 92%) as a syrup: [R]23D ) -38.0° (c 1.0, CHCl3);
FTIR (film) 3087, 2953, 2875, 1819, 1647, 1647, 1454, 1365, 1240,
obtain 26 (2.87 g, 81%) as a syrup: [R]23 ) -30.0° (c 1.0, CHCl3);
D
FTIR (film) 3267, 2954, 1806, 1755, 1495, 1458, 1370, 1342, 1090,
813, 750 cm-1; 1H NMR (CDCl3) δ 0.66 (m, 6H, J ) 7.8 Hz, -SiCH2-
CH3) 0.95 (t, 9H, J ) 7.8 Hz, -SiCH2CH3), 2.04 (s, 3H, -COCH3),
3.44 (dd, 1H, J ) 5.6, 10.2 Hz, H-5), 3.55-3.72 (m, 4H), 3.86 (br s,
1H), 4.11 (t, 1H, J ) 7.0 Hz), 4.23 (br s, 1H), 4.35 (dd, 1H, J ) 2.3,
10.0 Hz), 4.44 and 4.50 (2d, 2H, J ) 11.9 Hz, -CH2Ar), 4.57 (s, 2H,
-CH2Ar), 4.70 (br d, 1H, J ) 8.3 Hz), 4.89 (br s, 1H), 4.95-5.0 (m,
2H), 5.25 (t, 1H, J ) 9.6 Hz), 5.60 (d, 1H, J ) 9.9 Hz), 7.2-7.5 (m,
13H, Ar-H), 7.88 (d, 2H, J ) 7.7 Hz, Ar-H); 13C NMR (CDCl3) δ
4.93, 6.95, 20.64, 67.49, 67.86, 68.40, 68.46, 71.91, 72.57, 73.33, 73.94,
75.20, 79.30, 126.39, 127.35, 127.67, 127.85, 127.98, 128.09, 128.36,
128.54, 128.58, 129.10, 132.35, 132.68, 137.14, 137.91, 141.36, 153.60,
168.69; HRMS calcd for C41H52INO13SSiNa 976.190, found 976.187.
1
1101, 854, 739 cm -1; H NMR (CDCl3, 400 MHz) δ 0.57 and 0.617
(2q, 12H, J ) 8.0 Hz, -SiCH2CH3), 0.92 and 0.94 (2t, 18H, J ) 8.0
Hz, -SiCH2CH3), 3.5-3.75 (m, 4H), 3.8-4.0 (m, 3H), 4.05-4.20 (m,
2H), 4.49 (dd, 1H, J ) 4.4, 7.2 Hz), 4.50-4.62 (m, 4H, -CH2Ar),
4.64 (d, 1H, J ) 5.2 Hz, H-1′), 4.70 (dd, 1H, J ) 4.0, 5.6 Hz, H-4′),
4.76 (br d, 1H, J ) 7.5 Hz, H-2), 6.32 (d, 1H, J ) 6.0 Hz, H-1); 13C
NMR (CDCl3, 400 MHz) δ 4.56, 4.79, 6.58, 6.76, 65.24, 67.99, 68.02,
69.48, 71.06, 73.37, 73.76, 74.24, 74.37, 75.10, 78.21, 99.21, 99.34,
102.56, 127.63, 127.77, 127.79, 127.88, 128.32, 128.43, 137.53, 138.09,
143.08, 153.87; HRMS calcd for C39H58O10Si2‚Na 765.360 (M + Na)+,
found 765.347.
Ethylthio Sulfonamide Disaccharide 27. To a solution of iodo
sulfonamide 26 (2.80 g, 2.93 mmol) in dry DMF (40 mL) at -40 °C
was added EtSH (1.08 mL, 14.7 mmol), followed by dropwise addition
of a solution of LHMDS (1.0 M solution in THF, 8.80 mL). The
reaction mixture was stirred for 1 h, while it was allowed to warm to
room temperature, quenched with a saturated solution of NH4Cl (10
mL), and extracted with EtOAc (5 × 20 mL). The organic layer was
washed with brine (15 mL), separated, dried (MgSO4), filtered, and
concentrated. The resulting crude material was redissolved in CH2Cl2
(50 mL) and treated with pyridine (1.0 mL), Ac2O (1.0 mL), and the
reaction mixture was stirred overnight. The organic layer was washed
with a saturated solution of CuSO4 (3 × 15 mL) and water (1 × 10
mL), and finally with a saturated solution of NaHCO3 (2 × 15 mL).
The organic layer was separated, dried (MgSO4), filtered, concentrated,
and submitted for chromatography (1:1 EtOAc/hexanes) to provide 27
(2.38 g, 91%) as a syrup: [R]23D ) -4.0° (c 1.0, CHCl3); FTIR (film)
3316, 2955, 2875, 1815, 1745, 1448, 1371, 1330, 1227, 1092, 897,
740 cm-1; 1H NMR (CDCl3) δ 0.51 (q, 6H, J ) 8.0 Hz, -SiCH2CH3),
0.88 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 1.09 (t, 3H, J ) 7.2 Hz,
-SCH2CH3), 2.09 (s, 3H, -COCH3), 2.44 (m, 2H, -SCH2CH3), 3.48
(br m, 1H, H-2), 3.83-3.70 (m, 7H), 3.89 (br t, 1H), 3.95 (br s, 1H),
4.43 (d, 1H, J ) 5.4 Hz, H-1), 4.48 (br d, 2H, -CH2Ar), 4.53 (d, 1H,
J ) 6.3 Hz, H-1′), 4.57 (s, 2H, -CH2Ar), 4.75 (br t, 1H, J ) 5.7 Hz,
H-2′), 4.84 (br d, 1H, 9.9 Hz, -NHSO2Ph), 7.20-7.40 and 7.40-7.60
(m, 13H, Ar-H), 7.97 (d, 2H, J ) 7.2 Hz, Ar-H); 13C NMR (CDCl3)
δ 4.28, 6.65, 14.56, 20.64, 20.89, 56.96, 67.64, 70.49, 70.52, 70.57,
71.14, 73.26, 73.72, 73.96, 74.99, 75.02, 76.88, 82.48, 97.83, 126.21,
127.30, 127.61, 127.78, 127.87, 128.29, 128.38, 128.62, 128.94, 132.17,
3,6,6′-Tri-O-benzyl 3′,4′-Carbonate Lactal (24). To a solution of
compound 16 (2.99 g, 11.4 mmol) in dry CH2Cl2 (20 mL) at 0 °C was
added 3,3-dimethyldioxirane (300 mL, 0.0800 M solution in acetone).
The reaction mixture was stirred at 0 °C for 1 h. The organic solvent
was evaporated by N2 stream, and further dried in vacuo for 10 min.
The resulting 1,2-anhydro sugar 17 was dissolved in a solution of known
3,6-di-O-benzylglucal (18) (5.29 g, 17.2 mmol) in dry THF (30 mL).
To the solution at 0 °C was added a 1.0 M solution of ZnCl2 in ether
(5.71 mL, 0.5 equiv). The reaction mixture was stirred at room
temperature for 24 h, diluted with EtOAc (50 mL), and washed with a
saturated solution of NaHCO3 (2 × 10 mL). The organic layer was
separated, dried (MgSO4), filtered, concentrated, and submitted for
chromatography (1:1 EtOAc/hexanes) to provide compound 24 (3.31
g, 48%) as a syrup: [R]22D ) -38.0° (c 1.0, CHCl3); FTIR (film) 3437,
3029, 2871, 1804, 1648, 1453, 1367, 1166, 1097, 1027, 739, 697 cm-1
;
1H NMR (400 MHz, CDCl3) δ 3.55-3.62 (m, 2H), 3.62-3.70 (m,
2H), 3.70-3.78 (m, 2H), 3.95-4.11 (m, 2H), 3.95-4.11 (m, 2H), 4.17
(dd, 1H, J ) 5.4, 7.0 Hz), 4.27 (ddd, 1H, J ) 1.1, 1.7, 5.3 Hz), 4.44
(s, 2H, -CH2Ar), 4.77 (dd, 1H, J ) 2.5, 6.1 Hz, H-2), 6.28 (d, 1H, J
) 6.0 Hz, H-1), 7.10-7.40 (m, 15H, Ar-H); 13C NMR (400 MHz,
CDCl3) δ 68.00, 68.09, 70.55, 70.63, 72.20, 73.58, 73.81, 74.58, 74.82,
75.26, 76.18, 78.47, 100.17, 101.32, 127.43 (2C), 127.56, 127.72 (2C),
127.83, 127.90, 128.00 (2C), 128.31 (2C), 128.37 (2C), 128.44 (2C),
137.28, 137.43, 138.29, 144.59, 153.97; HRMS calcd for C34H36O10‚-
Na (M + Na)+ 627.220, found 627.220.

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1611
137.32, 137.94, 141.38, 153.27, 170.99; HRMS calcd for C43H57NO13S2-
Si‚Na (M + Na)+ 910.290, found 910.294.
Tetrasaccharide Pentaol 31. To a solution of tetrasaccharide 28
(0.37 g, 0.277 mmol) in MeOH (5 mL) was added K2CO (0.37 g),
3
and the reaction mixture was stirred for 15 min. Then the solution
was diluted with CH2Cl2 (100 mL), and filtered through a pad of silica
gel, followed by a washing of the pad with EtOAc (100 mL). The
filtrate was concentrated, and dried in vacuo without further purification
to afford 31 (0.295 g, 85%) as a syrup: [R]23D ) -18.0° (c 1.0, CHCl3);
FTIR (film) 3469, 3030, 2873, 1648, 1496, 1452, 1328, 1092, 909,
737 cm-1; 1H NMR (CDCl3) δ 0.31 (q, 6H, J ) 6.4 Hz, -SiCH2CH3),
0.70 (t, 9H, J ) 6.4 Hz, -SiCH2CH3), 2.49 (br s, 1H, -OH), 2.82 (br
s, 1H, -OH), 3.16 (m, 1H, -CHNHSO2Ph), 3.3-3.6 (m, 12H), 3.6-
3.78 (m, 6H), 3.79 (br s, 2H), 3.8-3.85 (m, 3H), 3.92 (br d, 1H, J )
4.2 Hz), 4.0 (br t, 1H), 4.05-4.10 (m, 2H), 4.10-4.25 (m, 3H), 4.30-
4.40 (m, 6H), 4.4-4.55 (m, 7H), 4.75 (dd, 1H, J ) 2.7, 4.9 Hz), 4.9
(d, 1H, J ) 4.2 Hz), 6.17 (d, 1H, 6.6 Hz, -HNSO2Ph), 6.31 (d, 1H, J
) 4.9 Hz, H-1), 7.0-7.4 (m, 23H, Ar-H), 7.80 (d, 2H, J ) 6.0 Hz,
Ar-H); 13C NMR (CDCl3) δ 4.30, 6.72, 57.88, 68.00, 68.78, 68.84,
69.20, 70.46, 70.86, 71.39, 71.99, 73.04, 73.14, 73.31, 73.40, 73.54,
73.79, 75.72, 76.01, 76.16, 81.44, 100.15, 101.85, 102.32, 102.60,
127.30, 127.45, 127.58, 127.62, 127.65, 127.73, 128.17, 128.21, 128.30,
128.33, 128.90, 132.52, 137.71, 137.87, 137.91, 138.10, 138.62, 140.18,
144.27; HRMS calcd for C71H89NO20SSiNa 1358.536 (M + Na)+, found
1358.536.
Tetrasaccharide Diol 28. To a solution of a disaccharide 25 (0.100
g, 0.174 mmol) and thio donor 27 (0.308 g, 0.347 mmol) with 4 Å
molecular sieves (1.0 g) in dry CH2Cl2 (8 mL) was added di-tert-
butylpyridine (0.311 mL, 0.694 mmol). The suspension was cooled
to -10 °C, treated with MeOTf (0.156 mL, 0.694 mmol), and stirred
for 2 h. The reaction mixture was warmed to 0 °C, and stirred for 24
h. The reaction mixture was quenched with Et3N (0.1 mL), stirred for
an additional 3-5 min, diluted with EtOAc (25 mL), and filtered
through a pad of silica gel. The filtrate was washed with a saturated
solution of NaHCO3 (2 × 10 mL), and the organic layer was separated,
dried (MgSO4), filtered, concentrated, and submitted for chromatog-
raphy (35% EtOAc in hexanes) to provide 0.134 g (55%) of 28
tetrasaccharide as a syrup. Further elution (60% EtOAc in hexane)
provided the regioisomer 29 in 3:1-2:1 ratio favoring the desired
product: [R]23 ) -28.0° (c 1.0, CHCl3); FTIR (film) 3491, 3029,
D
3874, 1815, 1753, 1647, 1453, 1370, 1221, 1160, 1064, 738 cm-1; 1H
NMR (CDCl3) δ 0.38 (q, 6H, J ) 8.0 Hz, -SiCH2CH3), 0.76 (t, 9H,
J ) 8.0 Hz, -SiCH2CH3), 1.97 (s, 3H, -COCH3), 3.2-3.32 (m, 2H),
3.35-3.55 (m, 5H), 3.55-3.7 (m, 7H), 3.7-3.8 (m, 4H), 3.95 (dd,
1H, J ) 4.6, 11.3 Hz), 4.0-4.12 (m, 2H), 4.18 (br s, 1H), 4.3-4.65
(m, 15H), 4.7-4.8 (m, 2H), 4.89 (t, 1H, J ) 5.2 Hz), 5.31 (d, 1H, J )
8.4 Hz), 6.32 (d, 1H, J ) 6.0 Hz, H-1), 7.1-7.5 (m, 28H, Ar-H),
7.85 (d, 2H, J ) 7.4 Hz, Ar-H); 13C NMR (CDCl3) δ 4.47, 6.74,
20.62, 58.45, 67.89, 68.16, 68.91, 69.63, 70.01, 70.23, 70.70, 70.74,
72.95, 73.26, 73.32, 73.40, 73.43, 73.79, 74.38, 74.75, 74.81, 75.34,
76.60, 77.19, 82.21, 97.41, 100.41, 102.53, 102.84, 127.30, 127.44,
127.50, 127.56, 127.59, 127.62, 127.76, 127.82, 127.84, 127.96, 128.18,
128.24, 128.33, 128.38, 128.46, 128.83, 132.49, 137.32, 137.89, 138.70,
140.76, 144.51, 153.43, 168.94; HRMS calcd for C74H89O22SSiNa
1426.530, found 1426.526.
Di-triethylsilylated Iodo Sulfonamide Disaccharide 32. To a
solution of lactal 23 (2.50 g, 3.36 mmol) with 4 Å molecular sieves
(3.0 g) and benzenesulfonamide (2.64 g, 3.36 mmol) was added at 0
°C a freshly prepared solution of I(sym-coll)2ClO4 (5 equiv) in CH2-
Cl2. The reaction mixture was stirred at room temperature for 1 h,
filtered through a pad of silica gel, and washed with a saturated solution
of Na2S2O3 (3 × 25 mL), CuSO4 (5 × 25 mL), and water (2 × 10
mL). The organic layer was separated, dried (MgSO4), filtered,
concentrated, and submitted for chromatography (5% EtOAc in CH2-
Tetrasaccharide 3′,4′-Diol 29: [R]23 ) -40.0° (c 1.0, CHCl3);
D
Cl2) to provide 32 (3.20 g, 92%) as a syrup: [R]23 ) -19.0° (c 1.0,
D
FTIR (film) 3469, 3030, 2873, 1814, 1254, 1451, 1369, 1222, 1063,
737 cm-1; 1H NMR (CDCl3) δ 0.33 (q, 6H, J ) 7.7 Hz, -SiCH2CH3),
0.73 (t, 9H, J ) 7.5 Hz, -SiCH2CH3), 2.05 (s, 3H, -COCH3), 3.11 (s,
1H, -OH), 3.30-3.50 (m, 8H), 3.60 (br m, 6H), 3.68-3.81 (m, 5H),
3.80 (dd, 1H, J ) 4.9, 11.3 Hz), 4.02-4.12 (m, 2H), 4.14 (br s, 1H),
4.38-4.25 (m, 4H), 4.38-4.56 (m, 10H), 4.61 (dd, 2H, J ) 6.0, 10.4
Hz), 4.61 (m, 1H), 4.70-4.77 (m, 2H), 4.89 (br t, 1H, J ) 5.2 Hz),
5.49 (d, 1H, J ) 8.5 Hz, -NHSO2Ph), 6.31 (d, 1H, J ) 6.4 Hz,m
H-1), 7.10-7.30 (m, 25H, Ar-H), 7.32-7.45 (m, 3H, Ar-H), 7.87
(d, 2H, J ) 6.8 Hz, Ar-H); 13C NMR (CDCl3) δ 4.43, 6.74, 20.74,
57.53, 67.68, 68.11, 69.54, 70.26, 70.30, 70.51, 70.70, 72.07, 72.72,
73.18, 73.34, 73.50, 73.72, 73.81, 73.85, 74.04, 74.16, 74.26, 74.38,
74.89, 70.05, 76.11, 76.59, 77.24, 96.95, 100.38, 103.09, 103.52, 127.37,
127.41, 127.46, 127.68, 127.82, 127.85, 127.99, 128.24, 128.29, 128.39,
128.41, 128.49, 128.97, 132.55, 137.37, 137.81, 137.89, 138.37, 138.65,
139.47, 144.53, 153.47, 169.69; LRMS calcd for C74H89NO22SSi‚Na
1426.5 (M + Na)+, found 1426.6.
Alternative Synthesis of Tetrasaccharide Diol 28. To a solution
of compound 25 (0.0983 g, 0.170 mmol) in dry benzene (90 mL) was
added bis(tributyltin) oxide (0.0500 mL, 0.0935 mmol), and the resulting
solution was distilled overnight with removal of water with a Dean-
Stark trap. Thus formed tin ether was concentrated with a stream of
dry N2 and then further dried in vacuo. To a mixture of azeotropically
dried (3 × 5 mL of benzene) compound 26 (0.0405 g, 0.0425 mmol)
and freshly flame dried 4 Å molecular sieves (0.8 g) was added a
solution of the tin ether in 1.8 mL of THF via cannula. Then the
resulting suspension was cooled to -60 °C, and was treated with a
solution of AgBF4 (0.0337 g, 0.170 mmol) in 0.6 mL of THF via
cannula. The reaction mixture was stirred for 2 days with exclusion
of light while slowly being allowed to warm to room temperature. The
reaction mixture was diluted with EtOAc (100 mL) and filtered through
a pad of silica gel. The filtrate was washed with a saturated solution
of NaHCO3 (3 × 60 mL), and brine (1 × 60 mL). The organic layer
was separated, dried (Na2SO4), filtered, concentrated, and submitted
for chromatography (45% EtOAc in hexanes) to afford 0.0498 g (84%)
of tetrasaccharide 28 as the only product. Further chromatographic
separation (80% EtOAc in hexanes) afforded 0.0480 g of unused
acceptor 25.
CHCl3); FTIR (film) 3258, 2953, 2875, 1806, 1788, 1453, 1331, 1105,
849, 745 cm-1; 1H NMR (CDCl3, 400 MHz) δ 0.57 and 0.64 (2q, 12H,
J ) 8.0 Hz, -SiCH2CH3), 0.90 and 0.95 (2t, 18H, J ) 8.0 Hz,
-SiCH2CH3), 3.39 (br m, 1H, H-2), 3.60-3.70 (m, 4H), 3.78-3.83
(br m, 2H), 4.05-4.17 (m, 3H), 4.34 (dd, 1H, J ) 2.4, 8.7 Hz), 4.45
and 4.52 (2d, 2H, J ) 12.0 Hz, -CH2Ar), 4.55 (s, 2H, -CH2Ar), 4.68
(d, 1H, J ) 3.0 Hz), 4.89 (d, 1H, J ) 8.6 Hz), 5.29 (t, 1H, J ) 8.4
Hz), 5.47 (d, 1H, J ) 9.6 Hz, -NHSO2Ph), 7.2-7.5 (m, 13H, Ar-H),
7.89 (d, 2H, J ) 7.6 Hz, Ar-H); 13C NMR (CDCl3, 400 MHz) δ 4.54,
4.94, 6.59, 6.95, 67.94, 68.12, 68.39, 68.63, 73.12, 73.31, 73.36, 73.90,
75.26, 75.33, 76.86, 79.66, 100.04, 127.40, 127.67, 127.76, 127.93,
128.01, 128.36, 128.51, 128.60, 132.39, 137.34, 138.02, 141.31, 154.01;
HRMS calcd for C45H64INO12SSi2‚Na 1848.260 (M + Na)+, found
1848.263.
Di-triethysilylated Ethylthio Sulfonamide Disaccharide 33. To
a solution of iodo sulfonamide 32 (2.70 g, 2.63 mmol) in dry DMF
(40 mL) was added at -40 °C EtSH (0.584 mL, 7.89 mmol), followed
by dropwise addition of a solution of LHMDS (1.0 M solution in THF,
7.89 mL). The reaction mixture was stirred for 1 h while being allowed
to warm to room temperature. Then the solution was neutralized with
a saturated solution of NH4Cl (10 mL), and diluted with EtOAc (50
mL). The organic layer was washed with brine (5 mL), separated, dried
(MgSO4), filtered, concentrated, and submitted for chromatography (3:7
EtOAc/hexanes) to provide 33 (2.30 g, 91%) as a syrup: [R]23
)
D
-64.0° (c 1.0, CHCl3); FTIR (film) 3314, 2954, 2875, 1807, 1453,
1
1330, 1181, 1104, 739 cm-1; H NMR (CDCl3, 400 MHz) δ 0.50 (q,
6H, J ) 7.9 Hz, -SiCH2CH3), 0.624 (q, 6H, J ) 7.6 Hz, -SiCH2-
CH3), 0.87 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 0.94 (t, 9H, J) 8.0 Hz,
-SiCH2CH3), 1.11 (t, 3H, J ) 7.4 Hz, -SCH2CH3), 2.48 (m, 2H,-
SCH2CH3), 3.35 (m, 1H, H-2), 3.85-3.68 (m, 6H), 3.86 (br m, 1H),
3.97 (br t, 1H), 4.06 (br t, 1H, J ) 6.6 Hz), 4.49 (s, 2H, -CH2Ar),
4.57 (s, 2H, -CH2Ar), 4.55 (m, 1H), 4.61 (d, 1H, J ) 6.3 Hz, H-1),
4.67 (d, 1H, J ) 4.0 Hz, H′-4), 4.87 (d, 1H, J ) 8.8 Hz, H-1), 5.50 (d,
1H, J ) 8.8 Hz, -NHSO2Ph), 7.2-7.4 (m, 10H, Ar-H), 7.45-7.55
(m, 3H, Ar-H), 7.94 (d, 2H, J ) 7.2 Hz, Ar-H); 13C NMR (CDCl3)
δ 4.28, 4.45, 6.56, 6.71, 14.58, 25.64, 57.42, 67.82, 69.05, 69.68, 70.37,
71.80, 73.13, 73.66, 73.75, 76.45, 76.64, 77.05, 82.68, 100.94, 127.52,

1612 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
127.57, 127.79, 127.83, 128.22, 128.35, 128.84, 132.25, 137.49, 138.01,
140.70; HRMS (FAB) calcd for C47H69NO12S2Si2Na 982.380, found
982.370.
(s, 3H, -COCH3), 1.90 (s, 3H, -COCH3), 2.08 (s, 2H, -COCH3),
2.15 (s, 3H, -COCH3), 3.03 (br d, 1H, J ) 7.7 Hz, -CHNHSO2Ph),
3.2-3.4 (m, 8H), 3.4-3.85 (m, 30H), 3.85-4.2 (m, 8H), 4.20-4.6
(m, 29H), 4.75 (q, 1H, J ) 3.1, 6.0 Hz), 4.8 (br d, 1H, J ) 8.2 Hz),
4.88 (d, 1H, J ) 3.5 Hz), 5.10 (m, 2H, J ) 8.8 Hz), 5.26 (d, 1H, J )
2.5 Hz), 5.33 (d, 1H, J ) 8.7 Hz, -NHSO2Ph), 5.42 (d, 1H, J ) 2.6
Hz), 5.90 (d, 1H, J ) 10.8 Hz, -NHSO2Ph), 6.31 (d, 1H, J ) 6.0 Hz,
H-1), 7.1-7.5 (m, 41H, Ar-H), 7.82 and 7.89 (2br m, 4H, Ar-H);
13C NMR (CDCl3) δ 4.10, 4.14, 4.49, 6.52, 6.60, 6.64, 20.75, 20.81,
20.09, 21.46, 55.97, 56.73, 67.83, 68.41, 68.63, 68.80, 69.35, 69.82,
69.88, 70.12, 70.49, 71.09, 71.20, 71.71, 72.84, 72.95, 73.11, 73.38,
73.53, 73.60, 73.67, 73.74, 73.79, 74.10, 74.33, 74.40, 75.32, 75.78,
75.89, 76.18, 76.77, 77.20, 99.75, 100.15, 100.38, 100.53, 101.55,
102.17, 127.26, 127.34, 127.42, 127.47, 127.52, 127.58, 127.61, 127.62,
127.66, 127.73, 127.73, 127.80, 127.85, 128.14, 128.21, 128.26, 128.39,
128.41, 128.66, 128.99, 131.93, 132.60, 137.47, 137.66, 137.77, 137.92,
138.31, 138.43, 138.77, 139.96, 141.74, 144.48, 154.07, 169.44, 169.60,
169.64, 171.34; LRMS (ES) calcd for C124H160N2O36S2Si3Na 2424.1
(M + Na)+, found 2424.1.
Hexasaccharide Tetraol 35. To a solution of disaccharide 33 (0.200
g, 0.208 mmol) and tetrasaccharide 31 (0.278 g, 0.208 mmol) in CH2-
Cl2/Et2O (1:2, 15 mL), with 4 Å molecular sieves (1.20 g) and di-tert-
butylpyridine (0.180 mL, 0.778 mmol) at -10 °C, was added MeOTf
(0.0880 mL, 0.778 mmol). The reaction mixture was stirred for 2 h,
allowed to warm to 0 °C, and stirred for 24 h. Then the suspension
was diluted with EtOAc (15 mL) and filtered through a pad of silica
gel, and the filtrate was washed with a saturated solution of NaHCO3
(2 × 10 mL). The organic layer was separated, dried (MgSO4), filtered,
concentrated, and submitted for chromatography (1:1 EtOAc/hexanes)
to provide 35 (0.276 g, 60%) as a syrup: [R]23 ) -23.0° (c, 1.0,
D
CHCl3); FTIR (film) 3490, 3030, 2875, 1807, 1649, 1453, 1330, 1093,
1
909, 743 cm-1; H NMR (CDCl3) δ 0.25 (m, 6H, -SiCH2CH3), 0.37
(q, 6H, J ) 7.9 Hz, -SiCH2CH3), 0.71 (t, 9H, J ) 7.9 Hz,
-SiCH2CH3), 0.76 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 0.86 (t, 9H, J )
7.9 Hz, -SiCH2CH3), 2.46 (s, 1H, -OH), 2.52 (s, 1H, -OH), 3.15
(m, 1H, -CHNHSO2Ph), 3.21 (m, 1H, -CHNHSO2Ph), 3.28 (dd, 1H,
J ) 3.0, 9.2 Hz), 3.37-3.55 (m, 7H), 3.55-3.79 (m, 14H), 3.82 (br s,
2H), 3.89 (br s, 1H), 3.94-4.11 (m, 4H), 4.18 (br s, 1H), 4.28 (m,
1H), 4.33-4.40 (m, 3H), 4.41 (s, 2H, -CH2Ar), 4.44-4.47 (m, 3H),
4.49 (s, 2H, -CH2Ar), 4.52 (m, 1H), 4.54 (s, 2H, -CH2Ar), 4.55-
4.63 (m, 2H), 4.66 (dd, 2H, J ) 3.9, 6.0 Hz), 4.74 (dd, 1H, J ) 2.8,
6.1 Hz), 5.28 (d, 1H, J ) 7.5 Hz, -NHSO2Ph), 5.51 (d, 1H, J ) 8.3
Hz, -NHSO2Ph), 6.32 (d, 1H, J ) 6.0 Hz, H-1), 7.10-7.55 (m, 41H,
Ar-H), 7.83 (d, 2H, J ) 7.4 Hz, Ar-H), 7.89 (d, 2H, J ) 7.5 Hz,
Ar-H); 13C NMR (CDCl3) δ 4.36, 4.44 (2C), 6.55, 6.67, 6.87, 58.52,
58.82, 67.61, 67.76, 67.82, 68.11, 68.71, 68.94, 69.07, 69.49, 69.75,
69.78, 69.92, 70.47, 70.73, 72.51, 72.92, 73.26, 73.31, 73.34, 73.37,
73.68, 73.85, 74.26, 74.61, 75.21, 75.27, 75.75, 75.90, 76.40, 77.10,
82.97, 83.60, 99.93, 100.49, 101.64, 102.74, 102.82, 103.08, 127.20,
127.38, 127.42, 127.49, 127.54, 127.62, 127.72, 127.76, 127.88, 128.10,
128.17, 128.26, 128.30, 128.40, 128.84, 128.97, 132.38, 132.71, 137.37,
137.57, 137.87, 137.90, 138.12, 138.18, 138.76, 139.9, 140.62, 144.45,
154.16; HRMS (FAB) calcd C116H152N2O32S2Si3Na 2255.900 (M +
Na)+, found 2255.898.
Alternative Synthesis of Hexasaccharide Tetraol 35. To a solution
of compound 31 (0.125 g, 0.0932 mmol) in dry benzene (90 mL) was
added bis(tributyltin) oxide (0.0272 mL, 0.0513 mmol). The resulting
solution was distilled overnight with removal of water with a Dean-
Stark trap. Thus formed tin ether 34 was concentrated with a stream
of dry N2 and then further dried in vacuo. To a mixture of
azeotropically dried (4 × 5 mL of benzene) compound 32 (0.0224 g,
0.0233 mmol) and freshly flame dried 4 Å molecular sieves (0.76 g)
was added a solution of the tin ether 34 in 2.0 mL of THF via cannula.
The resulting suspension was cooled to -60 °C, and was treated with
a solution of AgBF4 (0.0185 g, 0.0932 mmol) in 1.0 mL of THF via
cannula. The reaction mixture was stirred for 4 days with exclusion
of light while slowly being allowed to warm to room temperature. The
reaction mixture was diluted with EtOAc (100 mL) and filtered through
a pad of silica gel. The filtrate was washed with a saturated solution
of NaHCO3 (3 × 60 mL) and brine (1 × 60 mL), and the organic
layer was separated, dried (Na2SO4), filtered, concentrated, and
submitted for chromatography (40% EtOAc in hexanes) to provide
0.0321 g (62%) of hexasaccharide 35 as the only product. Further
elution (80% EtOAc in hexanes) provided 0.0801 g of unused acceptor
31.
Hexasaccharide Triol 37. To a solution of hexasaccharide 36
(0.175 g, 0.0725 mmol) in dry THF (5 mL) was added a solution of
TBAF (1.0 M THF) and AcOH (1:1, 0.725 mL, 10 equiv). The solution
was stirred at 35 °C for 24 h, diluted with EtOAc (10 mL), and washed
with a saturated solution of NaHCO3 (2 × 5 mL), and the organic layer
was separated, dried (MgSO4), filtered, concentrated, and submitted
for chromatography (4:1 EtOAc/hexanes) to provide 37 (0.143 g, 93%)
as a white glassy substance: [R]23 ) -6.0° (c 1.0, CHCl3); FTIR
D
(film) 3472, 3028, 2870, 1805, 1745, 1369, 1225, 1161, 1069, 752
cm-1 1H NMR (CDCl3) δ 1.88, 1.92, 2.01, 2.02 (4s, 3H each,
;
-COCH3), 2.85 (br t, 1H, J ) 8.2 Hz, -CHNHSO2Ph), 3.02 (br q,
1H, J ) 7.0 Hz, -CHNHSO2Ph), 3.20 (dd, 1H, J) 7.6, 8.0 Hz), 3.27
(dd, 2H, J ) 4.7, 10.0 Hz), 3.3-3.8 (m, 36H), 3.87 (br s, 2H), 4.03
(br d, 3H), 4.10 (br s, 1H), 4.2-4.65 (m, 33H), 4.66 (d, 1H, 5.1 Hz),
4.77 (q, 1H, J ) 3.2 Hz), 5.01 (dd, 1H, J) 8.3, 9.7 Hz), 5.12 (dd, 1H,
J ) 8.2, 9.8 Hz), 5.25 (d, 1H, J ) 3.2 Hz), 5.39 (d, 1H, J ) 3.1 Hz),
6.32 (d, 1H, J ) 6.1 Hz, H-1), 7.10-7.45 (m, 41H, Ar-H), 7.78 (m,
4H, Ar-H); 13C NMR (CDCl3) δ 20.69, 20.78, 21.18 (2C), 59.66 (2C),
67.79, 67.88, 68.12, 68.47, 69.42, 69.64, 70.13, 70.48, 71.14, 72.25,
72.70, 72.86, 73.11, 73.22, 73.38, 73.44, 73.50, 73.56, 73.71, 73.93,
74.29, 75.75, 77.20, 79.97, 80.25, 99.66, 100.99, 101.07, 101.18, 101.45,
101.55, 127.22, 127.29, 127.34, 127.63, 127.67, 127.73, 127.75, 127.80,
127.84, 128.00, 128.15, 128.26, 128.36, 128.41, 128.49, 128.61, 132.07,
132.20, 137.13, 137.34, 137.70, 137.71, 137.92, 138.21, 138.66, 140.96,
141.23, 144.43, 145.90, 170.04, 170.09, 170.15, 170.16; HRMS calcd
for C106H118N2O36S2Na 2081.860, found 2081.676.
Nonasaccharide 39. To a solution of hexasaccharide 37 (0.140 g,
0.0680 mmol) and the known â-fluorofucose 38 (0.239 g, 0.530 mmol)
in dry toluene (10 mL) with 4 Å molecular sieves at 0 °C was added
2,6-di-tert-butylpyridine (0.157 mL, 0.680 mmol) followed by a solution
of Sn(OTf)2 (0.228 g, 0.53 mmol) in THF (1 mL). The suspension
was stirred for 36 h at room temperature and quenched with Et3N (0.5
mL). The reaction mixture was stirred for an additional 3 min, diluted
with EtOAc (25 mL), and filtered through a pad of silica gel. The
organic layer was washed with a saturated solution of NaHCO3 (2 ×
10 mL), and separated, dried (MgSO4), filtered, concentrated, and
submitted for chromatography (1:1 EtOAc/hexanes) to provide 39
(0.135 g, 60%) as a syrup: [R]23 ) -88.0° (c 1.0, CHCl3); FTIR
D
(film) 3029, 2869, 1817, 1745, 1722, 1452, 1270, 1096, 697 (cm-1);
1H NMR (CDCl3) δ 0.87 (d, 3H, J ) 6.2, -CH3), 0.93 (d, 3H, J ) 6.3
Hz, -CH3), 1.07 (d, 3H, J ) 6.4 Hz, -CH3), 1.61 (s, 3H, -COCH3),
1.80 (s, 3H, -COCH3), 1.93 (s, 3H, -COCH3), 1.98 (s, 3H, -COCH3),
3.2-3.4 (m, 6H), 3.4-3.9 (m, 27H), 3.9-4.0 (m, 2H), 4.0-4.2 (m,
8H), 4.2-4.65 (m, 34H), 4.65-4.8 (m, 7H), 4.84 (d, 1H, J ) 5.3 Hz),
4.89 (br t, 1H, J ) 8.4 Hz), 5.05-5.15 (m, 2H), 5.28 (br s, 1H), 5.35
(d, 1H, J ) 2.8 Hz), 5.40 (br s, 1H), 5.53 (br s, 3H), 5.65 (d, 1H, J)
5.6 Hz), 6.33 (d, 1H, J ) 6.0 Hz, H-1), 7.0-7.3 (m, 68H, Ar-H),
7.3-7.45 (m, 9H, Ar-H), 7.45-7.57 (m, 3H), 7.65 (d, 2H, J ) 7.6
Hz, Ar-H), 7.75 (d, 2H, J ) 7.5 Hz, Ar-H), 7.99, 7.96, 7.94 (3d, 6H,
J ) 7.5 Hz, Ar-H); 13C NMR (CDCl3) δ 16.09, 16.13, 16.17, 20.38,
20.60, 21.16, 21.23, 56.42, 58.41, 64.97, 65.24, 65.59, 67.39, 67.40,
67.45, 67.86, 68.53, 68.71, 69.16, 69.35, 70.17, 70.25, 70.83, 70.99,
Tetraacetylated Hexasaccharide 36. To a solution of hexasac-
charide 35 (0.175 g, 0.0784 mmol) in dry CH2Cl2 (20 mL) were added
pyridine (2 mL), Ac2O (2 mL) and DMAP (catalytic). The reaction
mixture was stirred for 24 h and washed with a saturated solution of
CuSO4 (3 × 10 mL) and NaHCO3 (3 × 10 mL), and the organic layer
was separated, dried (MgSO4), filtered, concentrated, and submitted
for chromatography (25% EtOAc in hexanes) to give rise to 36 (0.179
g, 95%) as a syrup: [R]23D ) -30.0° (c 1.0, CHCl3); FTIR (film) 3030,
2953, 1809, 1748, 1453, 1369, 1221, 1094, 738 (cm-1); 1H NMR
(CDCl3) δ 0.24 (m, 12H, -SiCH2CH3), 0.54 (q, 6H, J ) 8.1 Hz,
-SiCH2CH3), 0.68 (br t, 9H, J ) 7.7 Hz, -SiCH2CH3), 0.70 (br t, 9H,
J ) 7.8 Hz, -SiCH2CH3), 0.87 (t, 9H, J ) 7.9 Hz, -SiCH2CH3), 1.86

Strategy in Oligosaccharide Synthesis
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1613
71.23, 71.25, 71.47, 71.50, 71.90, 72.55, 74.60, 73.14, 73.22, 73.39,
73.44, 73.49, 73.58, 73.66, 73.79, 73.81, 73.86, 74.72, 74.98, 75.37,
75.53, 75.71, 75.85, 76.10, 76.64, 76.94, 77.18, 96.01, 78.77, 79.40,
96.01, 97.38, 97.72, 98.67, 99.74, 100.00, 100.16, 100.24, 100.66,
127.11, 127.19, 127.25, 127.30, 127.37, 127.50, 127.60, 127.69, 127.79,
127.82, 127.88, 127.92, 128.02, 128.10, 128.12, 128.15, 128.29, 128.31,
128.39, 128.43, 128.99, 129.76, 129.84, 130.07, 132.27, 132.65, 132.90,
132.91, 133.18, 137.45, 137.48, 137.70, 137.73, 137.86, 137.90, 137.95,
138.03, 138.14, 138.17, 138.20, 138.26; LRMS calcd for
(m, 6H, -CH3), 0.85 (d, 3H, J ) 6.0 Hz, -CH3), 0.93 (d, 3H, J ) 6.4
Hz, -CH3), 1.07 (d, 3H, J ) 6.4 Hz, -CH3), 1.18 (br s, 48H, aliphatic
-CH2), 1.6 (m, 4H), 1.84 (s, 6H, 2-COCH3), 2.0 (s, 3H, -COCH3),
2.08 (s, 6H, -COCH3) 3.2 (br m, 1H), 3.3 (br m, 6H), 3.3-3.75 (m,
27H), 3.78 (m, 3H), 3.85 (m, 2H), 3.95 (m, 3H), 4.2 (dd,4H), 4.15-
4.5 (m, 23H), 4.5-4.75 (m, 9H), 4.75-4.8 (m, 4H), 4.8-4.95 (m, 3H),
5.1 (br s, 2H), 5.26 (m, 2H), 5.57 (m, 2H), 5.59 (m, 2H), 7.00-7.35
(m, 79H, Ar-H), 7.4 (m, 8H, Ar-H), 7.5 (m, 3H, Ar-H), 7.65 (d,
2H), 7.75 (2d, 2H), 7.96 (m, 6H); 13C NMR (anomeric carbons) δ 98.06,
100.11 (2C), 100.47, 101.28, 102.74, 102.89, 102.97, 103.16; LRMS
calcd for C230H271N3O56S2‚NH4 4052.8 (M + NH4)+, found 4052.0.
C
187H196N2O51S2‚2Na 3395.2 (M + 2Na)2+, found 3395.2.
Nonasaccharide Thioglycoside 41. To a solution of a nonasac-
charide 39 (0.0502 g, 0.0150 mmol) in dry CH2Cl2 (1 mL) with 4 Å
molecular sieves (0.10 g) was added a solution of dimethyldioxirane
in acetone (ca. 0.08 M, 3 mL). The solution was stirred for 45 min,
and the solvent was removed with a stream of dry N2, and further dried
in vacuo (10 min). The resultant crude epoxide was dissolved in CH2-
Cl2 (1 mL), and was treated with EtSH (1 mL) and TFAA (0.005 mL)
at -78 °C. The reaction mixture was stirred for 6 h while warming to
room temperature. Volatiles were evaporated with a stream of dry N2
and in vacuo. The crude mixture was dissolved in CH2Cl2 (1 mL) and
then treated with acetic anhydride (0.5 mL) and pyridine (0.5 mL).
After 24 h of stirring, the reaction mixture was concentrated under
reduced pressure and purified by flash chromatography (2:3 EtOAc/
KH-1 Antigen (1). To a blue solution of sodium (0.018 g) in liquid
ammonia (5 mL) under N2 at -78 °C was added a solution of protected
KH-1 derivative 43 (0.0202 g, 0.00500 mmol) in dry THF (1 mL).
After 45 min of reflux at -78 °C, the reaction mixture was quenched
with absolute MeOH (5 mL). Most of the ammonia was removed with
a stream of dry nitrogen, and the solution was diluted with MeOH (5
mL) and stirred overnight. The solution was neutralized with Et3N‚-
HCl, stirred for an additional 15 min, and dried with a stream of dry
nitrogen. The crude material obtained was then suspended in DMF
(1.0 mL), THF (1.0 mL), and Et3N (1.0 mL) and treated with Ac2O (1
mL) and a catalytic amount of DMAP. The reaction mixture was stirred
for 24 h. The solution was concentrated, passed through a pad of silica
gel with the help of EtOAc, and again concentrated. The syrup obtained
was dissolved in MeOH (5 mL), treated with MeONa (0.005 mL),
neutralized with Dowex 50-X8 after 24 h, filtered, and concentrated to
give 0.00711 g (70%) of KH-1 antigen (1). An analytical sample was
prepared by RP column chromatography, eluting with water-5%
methanolic water, followed by lyophilization to deliver 1 as a white
foam: 1H NMR (DMSO) δ 0.95 (m, 3H), 1.1-1.35 (3d, 9H, -CH3),
1.38 (br m, multiple protons, aliphatic -CH2), 1.5 (m, 9H), 1.85 (s,
6H, -CHNHCOCH3), 1.9 (m, 2H), 2.0-2.20 (m, 6 H), 3.0-4.0 (m,
multiple protons), 4.01 (q, 1H, J ) 6.5 Hz), 4.17 (d, 1H, J ) 7.5 Hz),
4.27 (br d, 1H), 4.34 (br m, 1H), 4.40 (d, 1H, J ) 7.0 Hz), 4.59 (m,
1H), 4.65 (d, 1H, J ) 8.0 Hz), 4.67 (d, 1H, J ) 4.0 Hz), 4.72 (d, 1H,
J ) 7.0 Hz), 4.73 (d, 1H, J ) 7.5 Hz), 4.87 (d, 1H, J ) 3.5 Hz), 4.97
(d, 1H, J ) 3.5 Hz), 5.37 (dd, 1H, J ) 7.0 Hz, 15.5 Hz), 5.55 (dt, 1H,
6.5 Hz, 15.0 Hz); 13C NMR (DMSO, anomeric carbons) δ 101.79,
102.40, 103.60, 103.78, 104.51 (2C), 105.18, 106.70, 106.77; HRMS
(FAB) calcd for C92H163N3O45‚Na 2053.046 (M + Na)+, found
2053.049.
hexanes) to afford thioglycoside 41 (0.0302 g, 60%) as a syrup: [R]23
D
) -0.030° (c 1.0, CHCl3); FTIR (film) 3064, 2875, 1808, 1744, 1720,
1
1453, 1369, 1250, 1090, 730 cm-1; H NMR (CDCl3) δ 0.85 (d, 3H,
J ) 6.0 Hz, -CH3), 0.93 (d, 3H, J ) 6.1 Hz, -CH3), 1.07 (d, 3H, J
) 6.2 Hz, -CH3), 1.18 (dt, 3H, J ) 1.5 Hz, -SCH2CH3), 1.67 (s, 3H,
-COCH3), 1.81 (s, 3H, -COCH3), 1.92 (s, 3H, -COCH3), 1.97 (s,
3H, -COCH3), 2.05 (s, 3H, -COCH3), 2.62 (m, 2H, -SCH2CH3),
3.10 (br m, 1H), 3.2-3.4 (m, 7H), 3.4-3.75 (m, 24H) 3.8 (m, 3H),
3.82 (dd, 2H), 3.85-4.0 (m, 4H), 4.0-4.2 (m, 4H), 4.2-4.5 (m, 26H),
4.5-4.63 (m, 1H), 4.63-4.8 (m, 8H), 4.8-5.0 (m, 5H), 5.11 (d&m,
2H), 5.21 (d, 1H), 5.35 (d, 1H), 5.40 (br s, 1H), 5.54 (br s, 2H), 5.64
(br s, 1H), 7.0-8.2 (m, 90H, Ar-H); LRMS calcd for C191H204N2O53S3‚-
NH4 3369.2 (M + NH4)+, found 3469.0.
Azide 42. To a solution of thioglycoside 41 (0.0290 g, 0.00860
mmol), azidohyrin 40 (0.0357 g, 10.0 equiv) at 0 °C in dry Et2O/CH2-
Cl2 (2:1, 1.5 mL), and 4 Å molecular sieves (0.10 g) was added MeOTf
(0.00389 mL, 4 equiv). The reaction mixture was stirred while warming
to room temperature. After 24 h of stirring, the suspension was
quenched with Et3N (0.020 mL), diluted with EtOAc (5 mL), filtered
through a pad of silica gel, and washed with a saturated solution of
NaHCO3 (2 × 5 mL), and the organic layer was separated, dried
(MgSO4), filtered, concentrated, and submitted for column chroma-
tography (1:1 EtOAc/hexanes) to provide azide 42 (0.0181 g, 55%) as
Allyl Glycoside 44. To a solution of sodium (0.060 g) in liquid
ammonia (8 mL) under N2 at -78 °C was added a solution of
nonasaccharide glycal 39 (0.0469 g, 0.0140 mmol) in dry THF (2 mL).
After 45 min of reflux at -78 °C, the reaction mixture was quenched
with absolute MeOH (5 mL). Most of the ammonia was removed with
a stream of dry nitrogen and then diluted with MeOH (5 mL) and stirred
overnight. The basic solution was neutralized with Dowex 50-X8
(0.557 g), and the mixture was stirred for an additional 15 min, and
filtered. The resins were washed with a saturated solution of NH3 in
MeOH (4 × 50 mL), and all filtrates were combined and dried with a
stream of dry nitrogen. The crude material obtained was then suspended
in DMF (1.0 mL), THF (1.0 mL), and Et3N (1.0 mL), treated with
Ac2O (1 mL) and a catalytic amount of DMAP, and stirred for 24 h.
The solution was concentrated, passed through a pad of silica gel with
the help of EtOAc, and concentrated. The syrup obtained was dissolved
in CH2Cl2 and then at 0 °C under N2 treated with dimethyldioxirane
solution in acetone (ca. 0.08 M, 6 mL) and stirred for 45 min. Solvent
was removed with a stream of N2. The resultant syrup (0.0400 g) was
redissolved in allyl alcohol (5 mL). After 24 h, excess allyl alcohol
was evaporated, and the crude syrup was dissolved in MeOH (5 mL)
and treated with MeONa (25% in MeOH, 0.060 mL). After 24 h, the
mixture was neutralized with Dowex 50-X8, filtered, and concentrated
to provide allyl glycoside 44 (0.0130 g, 60%). An analytical sample
was prepared by RP column chromatography, eluting with water-5%
methanolic water, followed by lyophilization to obtain a white foam:
[R]23D ) -58.0° (c 0.6, H2O); 1H NMR (D2O) δ 1.12 (d, 3H, J ) 6.5
Hz, -CH3), 1.21 (d, 3H, J ) 7.0 Hz, -CH3), 1.24 (d, 3H, J ) 6.5 Hz,
-CH3), 2.00 (1.35 (3d, 9H, -CH3), 2.0 (s, 6H, -COCH3), 3.31 (br t,
1H, -CHNHAc), 3.4-4.0 (m, multiple protons), 4.06 (d, 1H, J ) 3.0
Hz), 4.13 (d, 1H, J ) 3.0 Hz), 4.22 (m, 2H), 4.41 (d, 1H, J ) 8.0 Hz,
a syrup: [R]23 ) -17.4° (c 0.7, CHCl3); FTIR (film) 3030, 2923,
D
2100, 1814, 1744, 1602, 1451, 1368, 1269, 1222, 1096, 743 cm-1; 1H
NMR (CDCl3) δ 0.80 (br t, 3H, -CH3), 0.85 (d, 3H, J ) 6.3 Hz,
-CH3), 0.93 (d, 3H, J ) 6.3 Hz, -CH3), 1.07 (d, 3H, J ) 6.4 Hz,
-CH3), 1.18 (br s, 23H, aliphatic -CH2), 1.33 (br m, 2H) 1.52 (s, 3H,
-COCH3), 1.81 (s, 3H, -COCH3), 1.94 (s, 3H, -COCH3), 2.02 (s,
6H, 2-COCH3), 3.20 (m, 2H), 3.3 (m, 5H), 3.38-3.43 (m, 5H), 3.5-
3.75 (m, 20H), 3.75-3.83 (br d, 2H), 3.85 (br q, 2H, J ) 5.8, 8.8 Hz),
3.9-4.3 (m,4H), 4.3-4.54 (m, 3H), 4.54-4.63 (m, 7H), 4.63-4.82
(m,7H), 4.83 (d, 1H, J ) 5.0 Hz), 4.89 (t, 1H, J ) 9.6 Hz), 5.10 (d,
1H, J ) 3.5 Hz), 5.12 (m, 1H), 5.26 (br s, 1H), 5.34 (m, 1H), 5.39 (br
d, 2H, J ) 8.2 Hz), 5.47 (d, 1H, J ) 8.2 Hz), 5.53 (br s, 2H), 5.60 (d,
1H, J ) 5.5 Hz), 5.69 (dt, 1H, J ) 6.6, 14.7 Hz), 7.0-7.3 (m, 80H,
Ar-H), 7.33-7.43 (m, 8H), 7.5 (m, 4H), 7.65 (d, 2H, J ) 7.6 Hz,
Ar-H), 7.78 (d, 2H, J ) 7.5 Ar-H), 7.97 (m, 6H, Ar-H); LMRS calcd
for C214H239N5O55S2‚2NH4 3858.5 (M + 2NH4)2+, found 3859.0.
Protected KH-1 Antigen 43. To a solution of azide 42 (0.0181 g,
0.00473 mmol) in EtOAc (3 mL) were added Lindlar’s catalyst (0.050
g) and palmitic anhydride (0.0102 g, 0.0200 mmol), and the reaction
mixture was stirred at room temperature under a H2 atmosphere for 24
h. The reaction mixture was filtered through a pad of silica gel, rinsed
with EtOAc (20 mL), concentrated, and submitted for chromatography
(1:1 EtOAc/hexanes) to provide amide 43 (0.0165 g, 85%) as a syrup:
[R]23D ) -36.0° (c 0.5, CHCl3); FTIR (film) 3025, 2923, 1814, 1745,
1
1657, 1451, 1367, 1266, 1095, 742 cm-1; H NMR (CDCl3) δ 0.80

1614 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Deshpande et al.
anomeric), 4.49 (d, 1H, J ) 8.0 Hz, anomeric), 4.51 (d, 1H, J ) 8.5
Hz, anomeric), 4.69 (d, 2H, J ) 8.5 Hz, 2 anomeric), 4.79 (q, 1H, J )
7.0 Hz), 4.86 (q, 1H, J ) 7.0 Hz), 5.09 (d, 1H, J ) 4.0 Hz, anomeric),
5.10 (d, 1H, J ) 4.0 Hz, anomeric), 5.25 (d, 1H, J ) 3.0 Hz, anomeric)
5.36 (d, 1H, J ) 17 Hz, -CHdCH2), 5.95 (m, 1H, -CHdCH2); 13C
NMR (D2O, anomeric carbons) δ 101.31 (2C), 102.20, 102.91, 103.79,
104.50, 105.23 (2C), 105.70; HRMS (FAB) calcd for C61H102N2O43‚-
Na 1573.575 (M + Na)+, found 1573.568.
Synthesis of KH-1 Aldehyde. Ozone gas was bubbled through a
solution of 0.004 g of KH-1 allyl glycoside 50 in 5 mL of methanol
for 10 min while the solution was stirred vigorously at -78 °C. The
excess ozone was then displaced with nitrogen over a period of 5 min.
To the solution was added 0.10 mL of methyl sulfide. The reaction
mixture was stirred for 2 h at room temperature.
molecular weight cutoff value of 30 000, with 6-7 changes of PBS at
4 °C. The epitope ratio was determined by estimating protein content
by BioRad dye binding protein assay and carbohydrate by a HPAEC-
PAD assay. The epitope ratio was 141:1.
Acknowledgment. This research was supported by the
National Institutes of Health (Grant Numbers AI 16943 and CA-
61422). An Army Graduate Fellowship is gratefully acknowl-
edged by H.M.K. We gratefully acknowledge Dr. George
Sukenick (Spectral Core Lab, SKI) for NMR and mass spectral
contributions and The Synthesis Core Lab (SKI) for preparative
synthesis work.
Supporting Information Available: Experimental proce-
dures for 30, 45, 46, 47, 48, 49, and 50 and NMR spectra (1H
for 1, 16, 19, 20-33, 35-37, 39, and 41-50, 13C for 16, 19,
20-33, 35-37, 39, and 45-49, and HMQC for 1, 43, 44, and
50) (67 pages). See any current masthead page for ordering
information and Web access instructions.
Conjugation of KH-1 Aldehyde with KLH. To a solution of 0.002
g of KH-1 aldehyde and 0.004 g of keyhole lympet hemocyanin (KLH)
in 1.0 mL of 0.1 M phosphate-buffered saline (PBS), pH 7.2, was added
0.0020 g of sodium cyanoborohydride, and the mixture was incubated
under gentle agitation at 37 °C. After 16 h, an additional 0.001 g of
sodium cyanoborohydride was added, and the incubation was continued
for another 22 h. The unreacted KH-1 aldehyde was removed
completely with multiple washes using a Amicon Centriprep with
JA9725864