Arabinofuranosyl oligosaccharides from mycobacteria: Synthesis and effect of glycosylation on ring conformation and hydroxymethyl group rotamer populations

Article abstract of DOI:10.1021/ja993543l

A series of αD-arabinofuranosyl oligosaccharides (2-8) that are fragments of the arabinan portions of two polysaccharides present in the cell wall of Mycobacterium tuberculosis have been synthesized. Preparation of the oligosaccharides involved the sequen

Full text of DOI:10.1021/ja993543l

VOLUME 122, NUMBER 7
FEBRUARY 23, 2000
© Copyright 2000 by the
American Chemical Society
Arabinofuranosyl Oligosaccharides from Mycobacteria: Synthesis and
Effect of Glycosylation on Ring Conformation and Hydroxymethyl
Group Rotamer Populations
Francis W. D’Souza, Joseph D. Ayers, Patrick R. McCarren, and Todd L. Lowary*
Contribution from the Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue,
Columbus, Ohio 43210
ReceiVed October 4, 1999
Abstract: A series of R-D-arabinofuranosyl oligosaccharides (2-8) that are fragments of the arabinan portions
of two polysaccharides present in the cell wall of Mycobacterium tuberculosis have been synthesized. Preparation
of the oligosaccharides involved the sequential addition of arabinofuranosyl residues from thioglycoside donors
3
to methyl glycoside acceptors. High-resolution NMR studies on the final products provided all JH,H values,
which were in turn used in PSEUROT 6.2 calculations to determine both the identity and equilibrium populations
of preferred conformers for each furanose ring in these glycans. Comparison of the ring conformers present
in 2-8 with those available in the parent monosaccharide, methyl R-D-arabinofuranose (16), allowed the
determination of the effect of glycosylation upon ring conformation. At equilibrium, 16 exists as an
approximately equimolar mixture of OT4 (North, N) and 2T3 (South, S) conformers. These studies showed that
glycosylation of 16 at OH5 resulted in no significant change in conformer identity or population relative to 16.
O
However, glycosylation of OH3 resulted in a change in the identity of the N species (to E) and a significant
favoring of this conformer at equilibrium. These trends were seen in all of the oligosaccharides. The populations
of the three possible staggered rotamers (gg, gt, tg) about the C4-C5 bond were essentially the same for all
residues in 2-8, and thus this equilibrium does not appear to be sensitive to glycosylation.
Introduction
complex) and consequently play critical roles in the survival
and pathogenicity of these microorganisms. A well-established
method for the treatment of bacterial disease is the use of
antibiotics that act by inhibiting cell wall biosynthesis.4 Given
the xenobiotic nature of furanose polysaccharides to humans,
the biosynthetic pathways leading to their formation are
particularly attractive targets for drug action. However, the
processes by which these polysaccharides are assembled in
nature is not well understood, and much additional research in
this area is needed before this goal can be achieved.
Although unknown in human biology, polysaccharides con-
taining furanosidic residues are important constituents of
glycoconjugates from many lower organisms including bacteria,1
parasites,2 and fungi.3 Glycans containing furanose moieties are
generally found in the extracellular glycocalyx (cell wall
(1) Brennan, P. J.; Nikaido, H. Annu. ReV. Biochem. 1995, 64, 29.
(2) (a) Gerold, P.; Eckert, V.; Schwarz, R. T. Trends Glycosci. Glyco-
technol. 1996, 8, 265. (b) de Lederkremer, R. M.; Colli, W. Glycobiology
1995, 5, 547.
(3) (a) Unkefer, C. J.; Gander, J. J. Biol. Chem. 1990, 265, 685. (b)
Mendonc¸a-Previato, L.; Gorin, P. A. J.; Travassos, L. R. Infect. Immun.
1980, 29, 934. (c). Groisman, J. F.; de Lederkremer, R. M. Eur. J. Biochem.
1987, 165, 327.
While a number of microorganisms produce polysaccharides
(4) (a) Williams, D. H. Nat. Prod. Rep. 1996, 13, 469. (b) Gadebusch,
H. H.; Stapley, E. O.; Zimmerman, S. B. Crit. ReV. Biotechnol. 1992, 12,
225. (c) Woodin, K. A.; Morrison, S. H. Pediatrics ReV. 1994, 15, 440.
10.1021/ja993543l CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/02/2000

1252 J. Am. Chem. Soc., Vol. 122, No. 7, 2000
D’Souza et al.
containing furanose residues,1-3 among the most important are
the mycobacteria.1 This genus of bacteria contains a number of
species, including the well-known human pathogens Mycobac-
terium tuberculosis and Mycobacterium leprae, respectively the
causes of tuberculosis and leprosy. In recent years infection by
other members of this genus have become health threats. Most
notable among these are Mycobacterium aVium infections, which
are now commonplace in AIDS patients.5 Although mycobac-
terial diseases have attracted renewed attention in recent years
because of their increasing incidence in the western world6 and
the emergence of drug-resistant strains,7 these diseases have been
a constant health threat worldwide for decades. For example,
over a third of the world’s population is estimated to be infected
with M. tuberculosis, and tuberculosis remains the single most
lethal bacterial disease, resulting in over 3 million deaths each
year.8
The treatment of mycobacterial diseases is difficult, requiring
adherence to a several-month regimen of antibiotics.7,9,10 Such
long treatments are necessary, in large part, as a result of the
incredible thickness and impermeability of the cell wall complex,
which prevents the effective passage of drugs into the organ-
ism.1,7,11 The morphology of the cell wall is unique to Myco-
bacteria and other members of the Actinomycetes family, and
the two major polysaccharide components are an arabinogalactan
(AG) and a lipoarabinomannan (LAM) in which all of the
galactose and arabinose residues are present in the furanose
form.1 The organism’s ability to make these polymers is critical
to its survival, and it has been shown that one of the drugs often
used to treat tuberculosis (ethambutol) acts by inhibiting the
biosynthesis of the arabinan portions of the AG and LAM.12
Figure 1. Hexaarabinoside motif (1) found at the nonreducing termini
of mycobacterial arabinogalactan (AG) and lipoarabinomannan (LAM).
binding proteins.15 In the AG, these same hydroxyl groups are
esterified with mycolic acids, branched, long-chain fatty ac-
ids.1,11,14 Through the tight packing of the alkyl chains, the
mycolic acids form a protective hydrophobic facade that in some
cases is nearly crystalline.11,16
It is unknown why mycobacteria synthesize cell wall polysac-
charides containing predominantly furanose residues. However,
it has been suggested that the protection afforded to the organism
by the tightly packed mycolic acids in the AG is one of the
reasons why mycobacteria have evolved to produce polyfura-
nosides and not polypyranosides.11 Polysaccharides containing
furanose rings are expected to be more conformationally flexible
than their counterparts composed of pyranose residues,17 and
thus the former would better allow the mycolic acids to align
in the proper orientation for side-by-side packing. This proposal
is, however, without any experimental support.
Identifying new antibiotics for the treatment of mycobacterial
diseases is an area of current interest,10,18 and inhibitors of
mycobacterial arabinosyltransferases are ideal synthetic targets.19
However, progress in this field has been hampered by the limited
amount of detailed information that is available concerning the
biosynthesis of mycobacterial arabinan.20 To some degree, this
is due to the lack of synthetic substrates that can be used for
fundamental biochemical studies leading to the isolation and
A detailed structural model of the mycobacterial cell wall is
now available.1,13,14 The terminal ends of both AG and LAM
are capped with the hexasaccharide motif 1 (Figure 1), which
is linked to the remainder of the polymer via an R-(1f5)-linked
linear chain of arabinofuranose residues. Hexasaccharide 1 in
turn serves as the attachment site for other functionalities present
in the cell wall. These groups are located at the periphery of
the cell wall complex and are therefore the interface between
the microorganism and its environment. In LAM, the primary
hydroxyl groups in 1 are often substituted with mannopyranosyl
oligosaccharides, which have been implicated in the initial stages
of infection through their interaction with human mannose-
(5) (a) Horsburgh, C. R., Jr. J. Infect. Dis. 1999, 179, S461. (b) Wahl,
S. M.; Greenwell-Wild, T.; Peng, G.; Hale-Donze, H.; Orenstein, J. M. J.
Infect. Dis. 1999, 179, S457. (c) Armstrong, W. S.; Katz, J. T.; Kazanjian,
P. H. Clin. Infect. Dis. 1999, 28, 341.
(6) (a) Opravil, M. Infection 1997, 25, 56. (b) Millard, F. J. J. Royal
Soc. Med. 1996, 89, 497. (c) Efferen, L. S.; Hyman, C. L. Curr. Opin.
Pulm. Med. 1996, 2, 236.
(7) Blanchard, J. S. Annu. ReV. Biochem. 1996, 65, 215.
(8) Dolin, P. J.; Raviglione, M. C.; Kochi, A. Bull. World Health Org.
1994, 72, 213.
(15) (a) Schlesinger, L. S. Curr. Topics Microbiol. Immunol. 1996, 215,
71. (b) Schlesinger, L. S.; Hull, S. R.; Kaufman, T. M. J. Immunol. 1994,
152, 4070. (c) Schlesinger, L. S.; Kaufman, T. M.; Iyer, S.; Hull, S. R.;
Marchiando, L. K. J. Immunol. 1996, 157, 4568. (d) Kang, B. K.;
Schlesinger, L. S. Infect. Immun. 1998, 66, 2769.
(16) (a) Jarlier, V.; Nikaido, H. FEMS Microbiol. Lett. 1994, 123, 11.
(b) Liu, J.; Barry, C. E., III; Besra, G. S.; Nikaido, H. J. Biol. Chem. 1996,
271, 29545.
(17) Westhof, E.; Sundaralingam, M. J. Am. Chem. Soc. 1983, 105, 970.
(18) Barry, C. E., III. Biochem. Pharm. 1997, 54, 1165.
(9) Mitchison, D. A. Tubercle 1985, 66, 219.
(10) Young, D. B.; Duncan, K. Annu. ReV. Microbiol. 1995, 49, 641.
(11) Connell, N. D.; Nikaido, H. Membrane permeability and transport
in Mycobacterium tuberculosis. In Tuberculosis: Pathogenesis, Protection
and Control; Bloom, B. R., Ed.; American Society for Microbiology,
Washington, D.C., 1994; p 333.
(12) (a) Mikusova´, K.; Slayden, R. A.; Besra, G. S.; Brennan, P. J.
Antimicrob. Agents Chemother. 1995, 39, 2484. (b) Deng, L.; Mikusova´,
K.; Robuck, K. G.; Scherman, M.; Brennan, P. J.; McNeil, M. R. Antimicrob.
Agents Chemother. 1995, 39, 694. (c) Khoo, K.-H.; Douglas, E.; Azadi, P.;
Inamine, J. M.; Besra, G. S.; Mikusova´, K.; Brennan, P. J.; Chatterjee, D.
J. Biol. Chem. 1996, 271, 28682.
(19) (a) Bouix, C. B.; Bisseret, P.; Eustache, J. Tetrahedron Lett. 1998,
39, 825. (b) Maddry, J. A.; Bansal, N.; Bermudez, L. E.; Comber, R. N.;
Orme, I. M.; Suling, W. J.; Wilson, L. N.; Reynolds, R. C. Bioorg. Med.
Chem. Lett. 1998, 8, 237. (c) Lee, R. E.; Smith, M. D.; Nash, R. J.; Griffiths,
R. C.; McNeil, M.; Grewal, R. K.; Yan, W.; Besra, G. S.; Brennan, P. J.;
Fleet, G. W. J. Tetrahedron Lett. 1997, 38, 6733.
(20) For some recent studies on arabinan biosynthesis in mycobacteria,
see: (a) Lee, R. E.; Brennan, P. J.; Besra, G. S. Glycobiology 1997, 7,
1121. (b) Woluka, B.; McNeil, M. R.; deHoffman, E.; Chojnacki, T.;
Brennan, P. J. J. Biol. Chem. 1994, 269, 23328. (c) Scherman, M. S.; Kalbe-
Bournonville, L.; Bush, D.; Xin, Y.; Deng, L.; McNeil, M. J. Biol. Chem.
1996, 271, 29652. (d) Lee, R. E.; Mikusova´, K.; Brennan, P. J.; Besra, G.
S. J. Am. Chem. Soc. 1995, 117, 11829. (e) Xin, Y.; Lee, R. E.; Scherman,
M. S.; Khoo, K. H.; Besra, G. S.; Brennan, P. J.; McNeil, M. Biochim.
Biophys. Acta 1997, 1335, 231.
(13) Besra, G. S.; Khoo, K.-H.; McNeil, M. R.; Dell, A.; Morris, H. R.;
Brennan, P. J. Biochemistry 1995, 34, 4257.
(14) Lee, R. E.; Brennan, P. J.; Besra, G. S. Curr. Topics Microbiol.
Immunol. 1996, 215, 1.

Arabinofuranosyl Oligosaccharides
J. Am. Chem. Soc., Vol. 122, No. 7, 2000 1253
Figure 2. Synthetic targets; rings have been lettered to facilitate comparison with 1.
purification of the appropriate biosynthetic enzymes. Further-
more, the rational design of inhibitors would be facilitated by
a greater appreciation of the conformational preferences of these
polysaccharides, which is currently unavailable. Such confor-
mational investigations could also provide experimental support
for the aforementioned hypothesis that the flexibility of these
molecules is a key factor as to why mycobacteria have evolved
to synthesize polysaccharides containing furanose, not pyranose,
rings.
furanosidic oligosaccharides are far less so. A relatively small
number of reports have described syntheses of glycoconjugates
containing D-galactofuranosyl,26 L-arabinofuranosyl,27 D-ribo-
furanosyl,28 D-arabinofuranosyl21,29,34 and D-fructofuranosyl30
residues. A range of glycosylating agents have been used in
(22) (a) Podlasek, C. A.; Stripe, W. A.; Carmichael, I.; Shang, M.; Basu,
B.; Serianni, A. S. J. Am. Chem. Soc. 1996, 118, 1413. (b) Church, T. J.;
Carmichael, I.; Serianni, A. S. J. Am. Chem. Soc. 1997, 119, 8946. (c)
Serianni, A. S.; Wu, J.; Carmichael, I. J. Am. Chem. Soc. 1995, 117, 8645.
(d) Serianni, A. S.; Chipman, D. M. J. Am. Chem. Soc. 1987, 109, 5297.
(e) Bandyopadhyay, T.; Wu, J.; Stripe, W. A.; Carmichael, I.; Serianni, A.
S. J. Am. Chem. Soc. 1997, 119, 1737. (f) French, A. D.; Dowd, M. K.;
Reilly, P. J. J. Mol. Struct. (THEOCHEM) 1997, 395-396, 271. (g) French,
A.; Schafer, L.; Newton, S. Q. Carbohydr. Res. 1993, 239, 51. (h)
Waterhouse, A. L.; Horva´th, K.; Liu, J. Carbohydr. Res. 1992, 235, 1. (i)
Hoffmann, R. A.; van Wijk, J.; Leeflang, B. R.; Kamerling, J. P.; Altona,
C.; Vliegenthart, J. F. G. J. Am. Chem. Soc. 1992, 114, 3710. (j) Serianni,
A. S.; Barker, R. J. Org. Chem. 1984, 49, 3292. (k) Cros, S.; Herve´ du
Penhoat, C.; Pe´rez, S.; Imberty, A. Carbohydr. Res. 1993, 248, 81. (l) Plavec,
J.; Tong, W.; Chattopadhyaya, J. J. Am. Chem. Soc. 1993, 115, 9734. (m)
Thibaudeau, C.; Plavec, J.; Chattopadhyaya, J. J. Am. Chem. Soc. 1994,
116, 8033. (n) Thibaudeau, C.; Plavec, J.; Garg, N.; Papchikhin, A.;
Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 4038. (o) Plavec, J.;
Thibaudeau, C.; Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 6558.
(p) Plavec, J.; Thibaudeau, C.; Chattopadhyaya, J. Pure Appl. Chem. 1996,
68, 2137. (q) Thibaudeau, C.; Chattopadhyaya, J. Nucleosides Nucleotides
1997, 16, 523.
In a previous communication,21 we have described prelimi-
nary synthetic details on the preparation of six oligosaccharide
fragments (2-7, Figure 2) of hexasaccharide 1 and showed that
these glycans are substrates for the arabinosyltransferases that
are involved in the biosynthesis of the mycobacterial cell wall.
In this paper we report a full account of the synthesis of
oligosaccharides 2-7 and the related disaccharide, 8. These
oligosaccharides are composed of 1,2-trans-arabinofuranosyl
residues; the preparation of glycans containing 1,2-cis-arabino-
furanosyl linkages, which are also found in these polysaccha-
rides, is underway and will be reported in the future. Addition-
ally, to probe the conformation of these molecules, we describe
the effect that glycosylation has upon not only the conformers
available to each ring in these oligosaccharides but also the
rotamer populations about the exocyclic C-C bonds. Although
a number of reports have described the conformation of furanose
monosaccharides and nucleotides,22 conformational studies of
oligosaccharides containing furanose rings have been more
limited, with the majority of studies being carried out on an
R-L-arabinofuranosyl disaccharide,23 sucrose and related fructo-
furanosyl oligomers,24 or glycans containing â-D-galacto-
furanosyl residues.25 This report represents the first study of
the effect of glycosylation on ring conformer populations in
oligosaccharides containing solely furanose rings.
(23) Cros, S.; Imberty, A.; Bouchemal, N.; Herve´ du Penhoat, C.; Pe´rez,
S. Biopolymers 1994, 34, 1433.
(24) (a) Duker, J. M.; Serianni, A. S. Carbohydr. Res. 1993, 249, 281-
303. (b) Casset, F.; Hamelryck, T.; Loris, R.; Brisson, J.-R.; Tellier, C.;
Dao-Thi, M.-H.; Wyns, L.; Poortmans, F.; Pe´rez, S.; Imberty, A. J. Biol.
Chem. 1995, 270, 25619. (c) Casset, F.; Imberty, A.; Herve´ du Penhoat,
C.; Koca, J.; Pe´rez, S. J. Mol. Struct. (THEOCHEM) 1997, 395-396, 211.
11. (d) Tran, V. H.; Brady, J. W. Biopolymers 1990, 29, 961. (e) Tran, V.
H.; Brady, J. W. Biopolymers 1990, 29, 977. (f) Herve´ du Penhoat, C.;
Imberty, A.; Roques, N.; Michon, V.; Mentech, J.; Descotes, G.; Pe´rez, S.
J. Am. Chem. Soc. 1991, 113, 3720. (g) Timmermans, J. W.; de Wit, D.;
Tournois, H.; Leeflang, B. R.; Vliegenthart, J. F. G. J. Carbohydr. Chem.
1993, 12, 969. (h) Liu, J.; Waterhouse, A. L.; Chatterton, N. J. J. Carbohydr.
Chem. 1994, 13, 859. (i) Waterhouse, A. L.; Calub, T. M.; French, A. D.
Carbohydr. Res. 1991, 217, 29. (j) Liu, J.; Waterhouse, A. L.; Chatterton,
N. J. Carbohydr. Res. 1991, 217, 43. (k) Immel, S.; Schmitt, G. E.;
Lichtenthaler, F. W. Carbohydr. Res. 1998, 313, 91.
Results and Discussion
Synthesis of Oligosaccharides. Although syntheses of pyra-
nosidic oligosaccharides are now common, similar studies of
(25) (a) Martin-Pastor, M.; Bush, C. A. Biochemistry 1999, 38, 8045.
(b) Gohle, H.; Immel, S.; Lichtenthaler, F. W. Carbohydr. Res. 1999, 321,
96.
(21) Ayers, J. D.; Lowary, T. L.; Morehouse, C. B.; Besra, G. S. Bioorg.
Med. Chem. Lett. 1998, 8, 437.

1254 J. Am. Chem. Soc., Vol. 122, No. 7, 2000
D’Souza et al.
these investigations, including thioglycosides, glycosyl imidates,
selenoglycosides, and glycosyl halides.
In designing the synthesis of oligosaccharides 2-8 we
endeavored to develop a route that would (1) be amenable to
the future preparation of analogues (potential arabinosyltrans-
ferase inhibitors) and (2) employ glycosyl donors that could be
stored for extended periods. We thus chose a linear approach
by which the glycans would be synthesized by the sequential
addition of thioglycoside donors to methyl glycoside acceptors.
We envisioned that the oligosaccharide targets could be prepared
from the seven building blocks (9-15) shown in Figure 3. The
linear approach would allow the easy synthesis of analogues
by substitution of the appropriately modified (e.g., fluorinated)
building blocks for one of those shown in Figure 3. Thio-
glycosides were chosen as the glycosyl donors because of their
hydrolytic stability and range of activation methods.31
The preparation of 9-15 from methyl R-D-arabinofuranoside
1632 and tetraacetate 1733 is described in the Supporting
Information. With these building blocks in hand, all glycosy-
lation reactions were carried out by reacting the appropriate
combination of thioglycoside donor and methyl glycoside
acceptor in the presence of N-iodosuccinimide and silver
triflate.31 In all cases, this promoter system provided good to
excellent yields of the oligosaccharide products, with excellent
stereoselectivities. The stereochemical outcome of the glyco-
sylations was straightforwardly determined by standard one-
dimensional 13C and 1H NMR experiments. The anomeric
carbons of R-arabinofuranosides resonate in the range of 107-
110 ppm, and the â-anomers appear between 97 and 104 ppm.
3
In addition, JH1,H2 is small (0-2 Hz) for the R-anomers and
Figure 3. Building blocks required for the synthesis of 2-8.
larger (4-5 Hz) for the â-anomers.27c,34
Disaccharides 2, 3, and 8 and trisaccharide 4 were readily
assembled in 2-3 steps as illustrated in Scheme 1. The R-(1f5)
linked disaccharide 2 and its R-(1f3) linked isomer 3 were
prepared, respectively, in 77% and 56% overall yields as
described previously34 via the protected derivatives 18 and 19.
Glycosylation of methyl glycoside 12 with thioglycoside 13
provided disaccharide 20 in 86% yield. The R-(1f2) linked
disaccharide 8 was obtained in 79% yield upon treatment of 20
with tetra-n-butylammonium fluoride (TBAF) followed by
Zemple´n deacetylation. Finally, reaction of diol 10 with 2.5
equiv of 13 provided trisaccharide 21 in 95% yield. Reaction
of 21 with sodium methoxide afforded a 72% yield of 4.
The assembly of the other three targets (5-7), although
requiring more extensive transformations, proceeded without
incident. The linear trisaccharide 5 was prepared as outlined in
Scheme 2. Reaction of alcohol 9 with thioglycoside 14 provided
the protected disaccharide 22 in 82% yield. The silyl protecting
group was cleaved (95% yield), and the resulting disaccharide
product 23 was glycosylated with 13 to provide trisaccharide
24 in 80% yield. Zemple´n deacetylation proceeded in 89% yield
to give 5. The remaining glycans, trisaccharide 6 and tetrasac-
charide 7, were synthesized from a common intermediate, diol
26, as illustrated in Schemes 3 and 4. Glycosylation of 9 with
thioglycoside 15 afforded, in 78% yield, disaccharide 25
(Scheme 3). Subsequent treatment with TBAF provided a 72%
yield of diol 26. A portion of 26 was reacted with tert-
butyldiphenylsilyl chloride in pyridine to give 27 in 78% yield.
Glycosylation of 27 with 13 afforded the protected trisaccharide
28 (76%), which was deprotected in two steps and 73% yield
to give 6. Alternatively (Scheme 4), glycosylation of 26 with
an excess of 13 provided tetrasaccharide 29 in 65% yield.
Treatment with sodium methoxide afforded an 85% yield of 7.
NMR Studies. Although NMR-based conformational inves-
(26) (a) Kochetkov, N. K.; Nepogod’ev, S. A.; Backinowsky, L. V.
Tetrahedron 1990, 46, 139. (b) Arasappan, A.; Fraser-Reid, B. Tetrahedron
Lett. 1995, 36, 7967. (c) Veeneman, G. H.; Notermans, S.; Hoogerhout, P.;
van Boom, J. H. Recl. TraV. Chim. Pays-Bas, 1989, 108, 344. (d) Gallo-
Rodriquez, C.; Varela, O.; de Lederkremer, R. M. J. Org. Chem. 1996, 61,
1886. (e) Gallo-Rodriquez, C.; Gandolfi, L.; de Lederkremer, R. M. Org.
Lett. 1999, 1, 245. (f) Johnston, B. D.; Pinto, B. M. Carbohydr. Res. 1998,
305, 289. (g) Choudhury, A. K.; Roy, N. Carbohydr. Res. 1998, 308, 207.
(h) Sugawara, F.; Nakayama, H.; Ogawa, T. Agric. Biol. Chem. 1986, 50,
1557. (i) van Heeswijk, W. A. R.; Visser, H. G. J.; Vliegenthart, J. F. G.
Carbohydr. Res. 1977, 59, 81.
(27) (a) Kawabata, Y.; Kaneko, S.; Kusakabe, I.; Gama, Y. Carbohydr.
Res. 1995, 267, 39-47. (b) Kaneko, S.; Kawabata, Y.; Ishii, T.; Gama, Y.;
Kusakabe, I. Carbohydr. Res. 1995, 268, 307. (c) Mizutani, K.; Kasai, R.;
Nakamura, M.; Tanaka, O.; Matsuura, H. Carbohydr. Res. 1989, 185, 27.
(d) Helm, R. F.; Ralph, J.; Anderson, L. J. Org. Chem. 1991, 56, 7015.
(28) (a) Shimomura, N.; Mukaiyama, T. Bull. Chem. Soc. Jpn. 1994,
67, 2532. (b) Uchiro, H.; Mukaiyama, T. Chem. Lett. 1996, 79. (c) Uchiro,
H.; Mukaiyama, T. Chem. Lett. 1996, 271. (d) Mukaiyama, T.; Matsubara,
K.; Hora, M. Synthesis 1994, 1368. (e) Chan, L.; Just. G. Tetrahedron 1990,
46, 151. (f) Chiu-Machado, I.; Castro-Palomino, J. C.; Madrazo-Alonso,
O.; Loipetgui-Palacios, C.; Verez-Benecomo, V. J. Carbohydr. Chem. 1995,
14, 551.
(29) (a) Backinowsky, L. V.; Nepogod’ev, S. A.; Kochetkov, N. K.
Carbohydr. Res. 1985, 137, C1. (b) Nepogod’ev, S. A.; Backinowsky, L.
V.; Kochetkov, N. K. Bioorg. Khim. 1986, 12, 940. (c) Nepogod’ev, S. A.;
Backinowsky, L. V.; Kochetkov, N. K. Bioorg. Khim. 1986, 12, 803. (d)
Kochetkov, N. K.; Bochkov, A. F.; Yazlovetsky, I. G. Carbohydr. Res.
1969, 9, 49. (e) Mereyala, H. B.; Hotha, S.; Gurjar, M. K. J. Chem. Soc.,
Chem. Commun. 1998, 685. (f) Pathak, A. K.; El-Kattan, Y. A.; Bansal,
N.; Maddry, J. A.; Reynolds, R. C. Tetrahedron Lett. 1998, 39, 1497.
(30) (a) Muller, T.; Schneider, R.; Schmidt, R. R. Tetrahedron Lett. 1994,
35, 4763. (b) Krog-Jensen, C.; Oscarson, S. J. Org. Chem. 1996, 61, 1234.
(c) Krog-Jensen, C.; Oscarson, S. J. Org. Chem. 1996, 61, 4512. (d) Krog-
Jensen, C.; Oscarson, S. J. Org. Chem. 1998, 63, 1780.
(31) Garegg, P. J. AdV. Carbohydr. Chem. Biochem. 1997, 52, 179.
(32) Martin, M. G.; Ganem, B.; Rasmussen, J. R. Carbohydr. Res. 1983,
123, 332.
(33) Kam, B. L.; Barascut, J.-L.; Imbach, J.-L. Carbohydr. Res. 1979,
69, 135.
(34) D’Souza, F. W.; Cheshev, P. E.; Ayers, J. D.; Lowary, T. L. J.
Org. Chem. 1998, 63, 9037.

Arabinofuranosyl Oligosaccharides
J. Am. Chem. Soc., Vol. 122, No. 7, 2000 1255
Scheme 1a
Scheme 2a
a (a) 14, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 82%; (b)
n-Bu4NF, THF, rt, 95%; (c) 13, N-iodosuccinimide, AgOSO2CF3,
CH2Cl2, 0 °C, 80%; (d) NaOCH3, CH3OH, rt, 89%.
and P, which defines the part of the ring that is most puckered,
is indicated along with the associated envelope or twist
conformer. For a given furanose ring, the standard model used
for assessing conformation in solution assumes that there are
two conformers existing in equilibrium. Typically, one of these
species is present in the northern hemisphere of the pseudo-
rotational wheel and the other in the southern hemisphere,
respectively termed the North (N) and South (S) conformers
(Figure 4).36 Interconversion between these puckered conformers
generally occurs through pseudorotation rather than inversion
via the planar ring form, in which the ring substituents are
eclipsed.37 The pathway by which the two interconvert (via the
West or East) depends on the identity and orientation of the
substituents on the ring.
a (a) 13, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 86%; (b)
NaOCH3, CH3OH, rt, 90%; (c) 13, N-iodosuccinimide, AgOSO2CF3,
CH2Cl2, 0 °C, 79%; (d) n-Bu4NF, THF, rt; (e) NaOCH3, CH3OH, rt,
two steps, 71%; (f) 13, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0
°C, 86%; (g) n-Bu4NF, THF, rt; (h) NaOCH3, CH3OH, rt, two steps,
79%; (i) 13, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 94%; (j)
NaOCH3, CH3OH, rt, 72%.
tigations of oligosaccharides are increasingly routine,35 the vast
majority of these studies have dealt with oligopyranosides,
which, in comparison with their furanosidic counterparts, are
relatively inflexible species. Most pyranose rings are rigid,
existing in one of two possible chair conformations. Therefore,
the major degrees of freedom available to an oligopyranoside
are rotation about glycosidic bonds (torsion angles Φ,Ψ), and
the exocyclic C-C bond (torsion angle ω).35a Understanding
the conformational properties of oligofuranosides is a more
complicated problem, because in addition to rotation about
glycosidic and exocyclic C-C bonds, the rings themselves are
flexible.36 Consequently, in addition to Φ, Ψ, and ω torsion
angles, furanoses also possess a ring torsion angle.
Furanose rings, like substituted cyclopentanes, exist in either
envelope (E) or twist (T) conformations. A particular ring
conformer can be described by two parameters, the puckering
amplitude (τm) and the pseudorotational phase angle (P), which
can be illustrated by the pseudorotational wheel36 shown in
Figure 4 for a D-furanose ring. The radius of the circle is τm
Figure 4. Pseudorotational itinerary for a D-aldofuranose ring.
In cases such as these where two conformers are equilibrating
rapidly on the NMR time scale, conformational analysis
becomes complicated because averaging of coupling constants
occurs. Fortunately, for furanose rings, a least-squares minimi-
zation program (PSEUROT 6.2)38 is available that, given the
3
observed intracyclic ring JH,H and the τm for each conformer,
will calculate the N/S ratio and provide P values for both
conformers from which the ring forms (e.g., 3T2) can be
determined.
Prior to the development of the PSEUROT method, Angyal
reported a more qualitative study on the conformations adopted

1256 J. Am. Chem. Soc., Vol. 122, No. 7, 2000
D’Souza et al.
Scheme 3a
a (a)15, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 78%; (b) n-Bu4NF, THF, rt, 72%; (c) t-BuPh2SiCl, pyridine, 0 °C, 78%; (d) 13,
N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 76%; (e) n-Bu4NF, THF, rt; (f) NaOCH3, CH3OH, rt, two steps, 73%.
Table 1. PSEUROT Analysis of 2-8 and 16a
compound
16
Ab
2
3
8e
4
C
5
6
7
A
B
B
D
H
B
D
A
B
C
A
B
D
A
B
C
D
PN (deg)c
72
77
75
94
70
OT4 OT4 OE
49 73 94
78
95
73
71
74
74
73
74
93
70
OT4 OT4 OE
49 53 92
75
97
74
OT4 OT4
57 53
73
N conformerd OT4 OT4 OT4 OE
OT4 OT4 OT4 OT4 OT4 OT4 OE
XN (%)
56
59
53
86
55
52
55
54
55
62
79
PS (deg)c
183 184 183 184 183 178 185 183 183 183 183 183 183 184 183 183 185 183 183
S conformerd 2T3
2T3
41
2T3
47
2T3
14
2T3
51
2T3
27
2T3
6
2T3 2T3
45 48
2T3
45
2T3
46
2T3
45
2T3
38
2T3
21
2T3
51
2T3
47
2T3
8
2T3
43
2T3
47
XS (%)
44
RMS (Hz)
0.02 0.01 0.02 0.02 0.11 0.14 0.12 0.6 0.10 0.00 0.04 0.06 0.02 0.07 0.11 0.02 0.06 0.06 0.07
a Calculated using a constant τm ) 39°. b See Figure 2 for assignment of ring letters. c P ) pseudorotational phase angle.36 d See Figures 4, 6,
and 8 for conformer definitions. e Analysis could not be done for ring G because of non-first-order coupling between ring protons.
Scheme 4a
Figure 5. Conformational equilibrium of 16.
the conformations of the molecules as a whole and the results,
which are presented in Table 1, are discussed in greater detail
below.
a (a) 13, N-iodosuccinimide, AgOSO2CF3, CH2Cl2, 0 °C, 65%; (b)
Ring Conformers Present in Monosaccharide 16. The
NaOCH3, CH3OH, rt, 85%.
parent monosaccharide, methyl R-D-arabinofuranoside, 16, was
investigated first and used as a basis of comparison with the
by methyl furanosides in solution.39 Using JH,H data, three
3
oligosaccharides. PSEUROT analysis of 16 identifies the OT4
criteria were developed for predicting the region of the pseu-
2
(N) and T3 (S) conformers as those that are in equilibrium in
dorotational itinerary in which the favored conformers of a
solution. Both conformations are present in roughly equimolar
furanose ring would be found. It was proposed that the preferred
amounts with the N form being favored slightly. In these two
conformations will be ones in which (1) the anomeric methoxy
conformers, the methoxy group occupies a pseudoaxial position,
group is placed pseudoaxial, (2) the hydroxymethyl group at
and the C4 hydroxymethyl group is pseudoequatorial (Figure
C4 is oriented in a pseudoequatorial fashion, and (3) the ring
5). However, in the S conformer, the orientation of this
substituents are staggered as much as possible. The first of these
hydroxymethyl group is oriented slightly less equatorially than
criteria is a consequence of the endo anomeric effect, which
necessitates an approximately antiperiplanar orientation between
in its N counterpart. These results compare favorably with those
one of the lone pairs on the ring oxygen and the C1-O1 bond.40
(35) (a) Peters, T.; Pinto, B. M. Curr. Opin. Struct. Biol. 1996, 6, 710.
(b) Homans, S. W. New Compr. Biochem. 1995, 29a, 67. (c) Rao, V. S. R.;
The second and third criteria would arise from steric effects.
Qasba, P. K.; Balaji, P. V.; Chandrasekaran, R. Conformation of Carbo-
From this discussion, it is clear that the first step in assessing
the conformation of an oligosaccharide containing furanose
residues is necessarily the determination of the ring conformers
that are present for each constituent monosaccharide. Conse-
quently, we have measured all 3JH,H in monosaccharide 16 and
oligosaccharides 2-8 and then used these couplings in PSEUROT
6.2 calculations to determine the conformers that each ring
adopts. Table S1 in the Supporting Information contains the
couplings used in these calculations. Through these investiga-
tions we have been able to more fully appreciate the effect of
substitution (glycosylation) on the equilibrium populations of
ring conformers. This work represents the first steps in probing
hydrates; Harwood Academic Publishers: Amsterdam, 1998.
(36) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94, 8205.
(37) Westhof, E.; Sundaralingam, M. J. Am. Chem. Soc. 1980, 102, 1493.
(38) (a) PSEUROT 6.2; Gorlaeus Laboratories, University of Leiden. (b)
de Leeuw, F. A. A. M.; Altona, C. J. Comput. Chem. 1983, 4, 428. (c)
Altona, C. Recl. TraV. Chem. Pays-Bas 1982, 101, 413.
(39) Angyal, S. J. Carbohydr. Res. 1979, 77, 37.
(40) (a) Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019. (b) Thatcher,
G. R. J. The Anomeric Effect and Associated Stereoelectronic Effects;
American Chemical Society: Washington, D.C., 1993. For papers specif-
ically addressing the anomeric effect in furanose rings, see: (c) Ellervik,
U.; Magnusson, G. J. Am. Chem. Soc. 1994, 116, 2340. (d) Cosse´-Barbi,
A.; Dubois, J. E. J. Am. Chem. Soc. 1987, 109, 1503. (e) Cosse´-Barbi, A.;
Watson, D. G.; Dubois, J. E. Tetrahedron Lett. 1989, 30, 163. (f) O’Leary,
D. J.; Kishi, Y. J. Org. Chem. 1994, 59, 6629.

Arabinofuranosyl Oligosaccharides
J. Am. Chem. Soc., Vol. 122, No. 7, 2000 1257
of Angyal39 and Serianni and Barker22j which predicted, without
the use of PSEUROT analysis, that the most favored solution
conformation of 16 would be in the OE/E1 region of the
pseudorotational itinerary.
In a previous investigation, we have carried out gas-phase
ab initio (Hartree-Fock, HF) and density functional theory
(DFT) calculations on 16.41 In that study, the gas-phase energies
of each of the 10 envelope conformers were calculated by first
fixing four atoms in a plane and then allowing the rest of the
molecule to minimize at HF and B3LYP levels with the 6-31G*
basis set. In the gas phase, the N and S conformers were
3
2
determined to be the E and E conformers, respectively, with
the latter being the global minima. In addition to the previously
mentioned steric and stereoelectronic factors, intramolecular
hydrogen bonding, particularly in the 2E conformer, was a key
stabilization factor. The OT4 and 2T3 solution conformers are in
reasonable agreement with the computationally derived geom-
etries, especially when considering that in solution solvation
will, without question, profoundly affect the energies of each
conformer. The agreement between S conformers is the best,
with the predicted solution structure (2T3) being immediately
adjacent to the gas-phase minimum (2E).
The two solution conformers presumably equilibrate via
pseudorotation through the east (OT4 T OE T OT1 T E1 T 2T1
Figure 6. Arabinofuranose derivatives that have been the subject of
previous conformational investigations.
2
2
T E T T3) as there is a favorable orientation of both the C1
and C4 substituents in all of these conformers. Pseudorotation
via the west would require proceeding through the EO conformer
in which the group at C4 is pseudoaxial and the anomeric
methoxy group is pseudoequatorial. Both of these factors would
significantly destabilize this conformer, in turn making
pseudorotation through the west a higher energy pathway than
the one involving the southeast quadrant. Our computational
studies support this proposal as the gas-phase barrier for
pseudorotation through the west (via EO) is approximately 4
in aqueous solution, they nevertheless provide a basis for
rationalizing our results.
In the gas phase, the global minima of 16 is the 2E conformer,
which is located in the southeast quadrant of the itinerary and
between the two conformations observed in solution. Given that
in the OT4 (N) conformation all of the ring substituents, except
the anomeric methoxyl group, are pseudoequatorial, it is not
surprising that this is one of the low-energy conformations in
solution. What is perhaps more surprising is the identity of the
S species, because in this conformer (2T3) both secondary OH
groups are pseudoaxial. Inspection of the gas-phase interatomic
distances41 show that the distance between O3 and O1 moves
from 4.41 Å in the OT4 conformer to 3.01 Å in the 2T3 form. A
similar diminution in the interatomic distance between O2 and
C5 is observed: 4.68 Å (OT4) to 3.48 Å (2T3).41 However,
concomitant with these changes is the formation of two
intramolecular hydrogen bonds, one between OH2‚‚‚O5 and the
other between OH3‚‚‚O1. Although the formation of these
intramolecular hydrogen bonds will certainly be diminished in
solution, the ability of these groups to interact in this manner
may be a factor contributing to the stabilization of the S
conformer.
O
kcal/mol higher in energy than through the east (via E).41
To further corroborate the structures of solution conformers
of 16 proposed to be in equilibrium, comparison was made with
two available crystal structures. Both X-ray and neutron
diffraction structures of 16 have recently been reported.42 In
the crystal, the unit cell is composed of two distinct structures
that are almost identical; the ring in both adopts an E4
conformation with a τm of 39°. In addition to having crystal
data for 16, we have also recently obtained an X-ray structure
of octyl R-D-arabinofuranoside (30, Figure 6).43 Similar to the
methyl glycoside, the unit cell has two distinct forms of the
molecule, with each furanose ring existing in an E4 conformation
with a τm of 39°. Thus, the structure of the N conformer of 16
observed in solution (OT4) is very similar to the conformation
of the rings present in two different R-D-arabinofuranosyl
glycosides in the solid state.
In addition to providing us with a potential energy surface
across the pseudorotational itinerary, our previous computational
studies41 allow the determination of all bond lengths, bond
angles, and dihedral angles of each envelope conformer in 16.
For a given parameter, taking an average of the values from
two adjacent envelopes provides an estimate of this parameter
in the intermediate twist conformer. Although clearly the values
obtained from these gas-phase calculations must be interpreted
with some caution when applied to experimental data obtained
A more compelling explanation is that this conformer is
stabilized by the gauche interactions44 that both OH2 and OH3
make with the ring oxygen. From earlier structural work done
in nucleic acids (Figure 7),22m-q,45 it is known that in double
stranded RNA, the ribofuranose rings exist in the 3T2 (C3′-endo)
conformation. This ring form is stabilized both by the anomeric
effect of the base and by the gauche relationship of the 2′
hydroxy group with the ring oxygen. In contrast, the furanose
rings in double stranded DNA, which lack this hydroxyl group,
can adopt either 3T2 (C3′-endo) or 2T3 (C2′-endo) conformations.
The former is stabilized by the anomeric effect of the base, and
the latter by the gauche effect between OH3 and the ring oxygen.
In 16, the inherent stereochemistry of the ring substituents is
such that in the 2T3 conformer stabilization is achieved via both
the anomeric effect of the aglycone and the gauche effects
between the two secondary hydroxyl group and the ring oxygen.
(41) Gordon, M. T.; Lowary, T. L.; Hadad, C. M. J. Am. Chem. Soc.
1999, 121, 9682.
(42) Evdokimov, A. G.; Kalb, A. J.; Koetzle, T. F.; Klooster, W. T.;
Martin, J. M. L. J. Phys. Chem. A 1999, 103, 744.
(43) Gallucci, J. C.; McCarren, P. R.; Lowary, T. L. Manuscript in
preparation.
(44) Wolfe, S. Acc. Chem. Res. 1972, 5, 102.

1258 J. Am. Chem. Soc., Vol. 122, No. 7, 2000
D’Souza et al.
that as a result of the exo-anomeric effect,46 the methyl group
in 16 would be expected to be oriented gauche to the ring
oxygen and trans to the C1-C2 bond and hence on the opposite
side of the molecule from OH3. Therefore, the effective steric
bulk that is presented to OH3 by the aglycone is minimal, and
the identity of this group should make little difference to the
conformation of the ring.
Our results for the reducing-end residue of 2 are in agreement
with previous studies on 5-O-methyl R-D-arabinofuranose (31),47
5-deoxy R-D-arabinofuranose (32),47 and the L-arabinofuranosyl
3
disaccharide 3322k,23 (Figure 6). The JH,H values for the C5
modified monosaccharides 31 and 32 are very similar to those
of 16, thus indicating that the identity of the group at C5 appears
to not dramatically affect conformation.47 In the case of
disaccharide 33,22k,23 each ring was shown to adopt conforma-
tions similar to those of the monosaccharide parent, methyl R-L-
arabinofuranoside, the enantiomer of 16. Molecular mechanics
calculations on 33 suggested that although the conformations
of the rings were unchanged, the energy barrier to pseudorotation
for each ring was higher than in the monosaccharide.22k,23
Figure 7.
The data presented here for the nonreducing residue in 2 also
compare favorably with previous conformational investigations
of L-arabinofuranosyl residues present in oligosaccharide frag-
ments of wheat arabinoxylans.22i In this earlier investigation,
two oligosaccharides, 34 and 35 (Figure 6), containing a total
of three different L-arabinofuranosyl residues were studied.
Although there were small differences between the three
furanose residues, all were shown to adopt approximately 1:1
2
mixtures of E4/OT4 (N) and T1/2E (S) conformers in solution.
Figure 8. (A) Conformational equilibrium in an unsubstituted or O5
substituted R-D-arabinofuranosyl ring. (B) Conformational equilibrium
in an O3 substituted R-D-arabinofuranosyl ring.
The best correlation with the results reported here is for the
L-arabinofuranosyl residue in 34, which exists as a 54:46 mixture
of E4/OT4 (N):2E (S) conformers. For the furanose residues in
35, poorer agreement is observed. This may arise from the
attachment of two arabinofuranose moieties on a single xylo-
pyranosyl residue, which may result in conformational changes
not observed in less sterically congested systems.
Ring Conformers Present in the Disaccharides. Armed
with an understanding of the conformation of the monosaccha-
ride, we next investigated disaccharides 2, 3, and 8 to determine
the effect of glycosylation on the conformer populations of each
ring. These results are presented in Table 1 and discussed in
detail below.
r-(1f3)-Linked Disaccharide. Although the identity of the
conformers and their equilibrium populations of the individual
rings in 2 do not differ appreciably from 16, the situation in
the R-(1f3) linked disaccharide 3 is rather different. The
nonreducing end residue (ring D) behaves as does the monosac-
charide, adopting the same two twist conformations in roughly
equal proportions. However, the conformational equilibria of
the reducing-end ring (ring B), which is glycosylated at OH3,
is significantly different from the monosaccharide. The N
conformer changes slightly from OT4 to OE and predominates
r-(1f5)-Linked Disaccharide. Both rings in the R-(1f5)-
linked disaccharide 2 adopt conformations identical to those of
the monosaccharide, namely the OT4 (N) and 2T3 (S) twist forms.
Clearly, glycosylation of an arabinofuranose ring at the primary
position does not alter the conformation of the ring. Similarly,
the replacement of a methyl group at the anomeric center with
another sugar does not change the conformer equilibrium either.
2
(86%) over the S form, T3 (14%). Clearly, glycosylation of
this secondary alcohol profoundly changes the conformation of
the ring.
That the conformation of the reducing-end residue (ring A)
of 2 does not differ from that of 16 is perhaps not surprising.
As mentioned previously, the N/S conformers of 16 both place
the C4 hydroxymethyl group in approximately the same,
pseudoequatorial position (Figure 8A). Given this sterically
favorable orientation, it could be expected that the glycosylation
of the primary hydroxyl group in 16 to give 2 would not
dramatically alter the identity of the conformers or their
equilibrium populations. When considering the conformations
available to the nonreducing arabinofuranose residue (ring B),
it might be expected that the replacement of the methoxy group
at the anomeric center with a furanose moiety would result in
a significant change in the conformation of the ring, due to
increased steric congestion between OH3 and the aglycone,
especially in the S conformer. However, it should be appreciated
To understand the origin of these changes two issues must
be addressed: one is the favoring of the N conformation over
the S; the other is the change in the identity of the N conformer.
The equilibrium preponderance of the OE (N) form can be
rationalized on the basis that in this conformer, the large glycosyl
group at O3 is in a more pseudoequatorial orientation than in
2
the T3 (S) conformer in which this group is pseudoaxial. Our
previous gas-phase calculations41 show that the O1/O3 distance
changes from 4.22 (OE) to 3.01 Å (2T3). Although in the
(45) (a) Saenger, W. Principles of Nucleic Acid Structure; Spring-
Verlag: Berlin, 1988. (b) Birnbaum, G. I.; Shugar, D. Biologically Active
Nucleosides and Nucleotides: Conformational Features and Interaction with
Enzymes. In Topics In Nucleic Acid Structure; Neidle, S., Ed.; MacMillan
Press: London, 1987; p 1.
(46) Lemieux, R. U.; Koto, S. Tetrahedron 1974, 30, 1933.
(47) Snyder, J. R.; Serianni, A. S. Carbohydr. Res. 1987, 163, 169.

Arabinofuranosyl Oligosaccharides
J. Am. Chem. Soc., Vol. 122, No. 7, 2000 1259
conformers relative to the monosaccharide. Furthermore, in most
cases the conformer populations of the O5 substituted rings are
essentially the same as the monosaccharide. Only in ring A of
6 is there a noticeable, albeit moderate, favoring (62%) of the
N conformer. Also unique to 6 is that, in comparison to 4 and
7, the N conformer of the O3 substituted ring (ring B) has a
lower population. In 6, the N:S ratio for ring B is 79:21, whereas
for 4 and 7, the ratios are, respectively, 94:6 and 92:8. The
reasons why 6 behaves slightly differently than the other
oligosaccharides are currently unknown.
These results suggest the conformers available to a particular
R-D-arabinofuranosyl ring can be predicted on the basis of their
substitution pattern and that, in general, these populations are
independent of the size of the molecule. That substitution of
either O5 or O1 with a variety of groups appears to not
significantly alter ring conformer populations is of particular
importance. Linear polymers of R-(1f5)-linked L-arabinofura-
nose are found in plants and have been suggested to adopt helical
conformations both in solution23 and in the solid state.48 The
arabinan core to which 1 is attached (Figure 1) is a similar
R-(1f5)-linked polymer of D-arabinofuranose. It would there-
fore be expected that these regions of the polymer would also
form helices and that this type of structural organization would
undoubtedly be of importance in the overall organization of the
cell wall complex. This could be probed through fluorescence
energy transfer experiments,49 through the synthesis of R-(1f5)-
linked oligosaccharides appropriately modified with fluorescent
groups. Our studies suggest that the attachment of the (typically
large) groups used for this purpose will not dramatically change
the overall conformation of the molecule.
Figure 9. O2-C2-C3-O3 and O3-C3-C4-C5 dihedral angles in OT4
and OE conformers of an R-D-arabinofuranosyl ring.
monosaccharide the S conformer can potentially be stabilized
by the formation of an OH3‚‚‚O1 hydrogen bond, in the
disaccharide this is not possible. More importantly, because of
unfavorable steric interactions in the S conformer, it would be
expected that its population will be reduced (Figure 8B). Clearly
these steric effects are enough to override, at least in part, the
stabilization provided by the gauche relationship between the
ring oxygen and both OH2 and O3.22m-q
Our previous computational studies41 on 16 are again useful
in explaining why the identity of the N conformer of an R-D-
arabinofuranosyl ring changes upon substitution of O3. An
analysis of the bond dihedral angles in 16 reveals that the O2-
C2-C3-O3 torsion angle increases from 108° in OT4 to 120°
in OE (Figure 9). A similar, though less dramatic, trend is
observed for the O3-C3-C4-C5 dihedral angle, which increases
from 91° (OT4) to 98° (OE). When taken together, these results
suggest that when 16 is glycosylated at O3, the ring adopts a
conformation that places the substituent group not only as far
away as possible from those adjacent to it but also pseudoequa-
torially.
C4-C5 Rotamer Populations. Rotamer populations about
the C4-C5 bond in each ring of 16 and 2-8 were calculated as
described previously50 using coupling constants measured from
1
r-(1f2)-Linked Disaccharide. For the nonreducing end
residue in the R-(1f2)-linked disaccharide 8 (ring H), the same
1D H NMR spectra. The populations of each rotamer (Table
2) are essentially the same in all of the compounds investigated;
the major one is gg (Figure 10) with a population of ap-
proximately 50%. The next most populated rotamer is gt
(∼38%) followed by tg (∼14%). In both the gg and gt rotamers
O5 is gauche to the ring oxygen and would be stabilized by the
gauche effect.44 Hence, it would be expected that these two
conformations would predominate over the tg rotamer, in which
there is no gauche effect stabilization. In the crystal structures
of both 1642 and 30,43 the C4-C5 bond adopts a gg orientation.
Furthermore, it is well-known that in nucleosides both in the
solid state51a and in solution,51b the preferred conformation about
this bond is gg.
2
two conformers present in 16, OT4 and T3, are found. This
behavior is therefore similar to the two other disaccharides.
However, in contrast to 2 and 3, there is a distinct favoring of
the N conformer (73%) in this residue in 8. Unfortunately, the
coupling between H1, H2, and H3 on the reducing-end residue
in 8 (ring G) is not first-order, and this makes the measurement
of accurate 3JH,H and subsequent PSEUROT analysis impossible.
The coupling between H3 and H4 on ring G (J3,4 ) 5.8 Hz) is
similar to the magnitude of this value in ring H (J3,4 ) 6.1 Hz).
Although this observation may suggest that the two rings adopt
similar conformations, in the absence of additional data this
proposal must be regarded as highly speculative. Indeed, one
might expect to see changes similar to those observed for ring
B in 3.
Ring Conformers Present in the Tri- and Tetrasaccha-
rides. We next investigated the conformer populations present
in each ring of oligosaccharides 4-7. The trends observed in
the disaccharides are also seen in the larger oligomers (Table
1).
In all cases, the residues at the nonreducing termini of the
oligosaccharides (rings C and D) behave as does 16, existing
as a nearly equimolar mixture of 2T3 and OT4 conformers.
Similar to 3, in the larger oligosaccharides when a residue is
glycosylated at OH3 (ring B in 4, 6, and 7), the N conformer
changes to OE and becomes favored at equilibrium. When this
residue is also glycosylated at OH5, as in 4 and 7, the population
of the N conformer exceeds 90%, indicating significant rigidity
in this ring relative to the monosaccharide parent, 16. In all
cases, the attachment of a single glycosyl residue to O5 (ring A
in 5-7 and ring B in 5) does not alter the identity of the
Conclusions
In this paper we report the synthesis of a series of arabino-
furanosyl oligosaccharides that are fragments of two mycobac-
terial cell wall polysaccharides. These compounds have previ-
ously been shown to be substrates for the enzymes involved in
the biosynthesis of these polysaccharides in mycobacteria.21 In
addition to the utility of these compounds in biochemical
investigations, the synthetic routes developed will be invaluable
in the synthesis of additional analogues. We have also, for the
(48) (a) Radha, A.; Chandrasekaran, R. Carbohydr. Res. 1997, 298, 105.
(b) Chandrasekaran, R.; Radha, A.; Lee, E. J.; Zhang, M. Carbohydr. Polym.
1994, 25, 235.
(49) (a) Stubbs H. J.; Lih, J. J.; Gustafson, T. L.; Rice, K. G. Biochemistry
1996, 35, 937. (b) Rice, K. G. Methods Enzymol. 1994, 247, 30.
(50) (a) Wu, G. D.; Serianni, A. S.; Barker, R. J. Org. Chem. 1983, 48,
1750. (b) Serianni, A. S.; Barker, R. Can. J. Chem. 1979, 57, 3160.
(51) (a) de Leeuw, H. P. M.; Haasnoot, C. A. G.; Altona, C. Isr. J. Chem.
1980, 20, 108. (b) Davies, D. B. Prog. Nucl. Magn. Reson. Spectrosc. 1978,
12, 135.

1260 J. Am. Chem. Soc., Vol. 122, No. 7, 2000
Table 2. C4-C5 Rotamer Populations in 2-8 and 16a
D’Souza et al.
compound
16
2
3
8
4
5
6
7
Ab
A
B
B
D
G
H
B
C
D
A
B
C
A
B
D
A
B
C
D
X
tg (%)c
Xgg (%)
gt (%)
14
48
38
15
47
38
16
50
34
15
47
38
15
47
38
14
48
38
15
47
38
12
54
34
15
47
38
14
46
40
13
48
39
13
50
37
14
48
38
13
52
35
15
48
37
12
48
40
13
52
35
12
53
35
13
50
37
13
50
37
X
3
3
a Calculated using the following equations:50a 1.3Xgg + 2.7Xgt + 11.7Xtg ) JH4,H5S; 1.3Xgg + 11.5Xgt + 5.8Xtg ) JH4,H5R; Xgg + Xgt + Xtg ) 1
b See Figure 2 for assignment of ring letters. c See Figure 10 for rotamer definitions.
identity of the N conformation changes from OT4 to OE, and
this conformation is favored at equilibrium. In cases in which
a ring is glycosylated at both OH3 and OH5, as in 4 and 7, the
favoring of the N conformation becomes even more pronounced.
(5) Rotamer populations about the C4-C5 bond are essentially
the same for each ring, regardless of the substitution pattern,
indicating that this parameter is insensitive to glycosylation.
The studies reported here will serve as the foundation for
additional studies on both the synthesis of furanose oligosac-
charides and their solution conformations. The synthesis of
additional analogues, including oligosaccharides containing 1,2-
cis-linked residues as found in the native polysaccharides, is
currently in progress. We are also carrying out further confor-
mational studies including the measurement of interresidue
NOEs, which will lead to an even greater appreciation of the
solution conformation of molecules of this type.
Figure 10.
first time, investigated the role that glycosylation has on
conformer populations in furanose rings. This has led to a greater
understanding of substituent effects in these molecules and to
the identification of some underlying principles that appear to
dictate the conformers available to both unsubstituted and
substituted R-D-arabinofuranosyl residues. Notable conclusions
from these conformational investigations are as follows:
(1) The monosaccharide parent, 16, exists in solution as an
approximately equimolar mixture of two twist conformers, OT4
and 2T3. These conformers are similar to those observed in both
the solid state42 and the gas phase.41
(2) Glycosylation and likely any form of substitution of a
primary hydroxyl group on an arabinofuranose ring does not
significantly alter the conformers present or their populations.
(3) Replacement of a methyl group at the anomeric center
with a monosaccharide residue does not alter the identity of
the conformers present. Hence, all terminal Araf residues exist
as equilibrium mixtures of the same two conformers present in
16. Furthermore, although in the terminal Araf residue in 8 there
is a favoring of the N (OT4) conformer, for the other oligosac-
charides the conformer populations for these residues are
essentially the same as in the monosaccharide.
Experimental Section
See Supporting Information.
Acknowledgment. This research has been supported by The
Ohio State University and the National Science Foundation
(CHE-9875163). P.R.M. is the recipient of a NSF-REU Fel-
lowship. We thank Dr. Charles Cottrell for the high field NMR
spectra and Dr. Christopher M. Hadad for helpful discussions.
Supporting Information Available: Full Experimental
Section and analytical data for all new compounds, discussion
of the preparation of 9-15, NMR conditions used for the
conformational studies, and table containing all 3JH,H values used
in the PSEUROT calculations (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
(4) Glycosylation of OH3 on an arabinofuranose ring does
alter the conformational equilibria of the ring significantly. The
JA993543L