Design of an oligosaccharide scaffold that binds in the minor groove of DNA

Article abstract of DOI:10.1021/ja992513f

Many biological recognition events involve extensive interactions between macromolecules. Strategies to design compounds that mimic large peptides or other elements of secondary structure could be useful for blocking interactions between large surfaces. I

Full text of DOI:10.1021/ja992513f

J. Am. Chem. Soc. 2000, 122, 1883-1890
1883
Design of an Oligosaccharide Scaffold That Binds in the Minor
Groove of DNA
Hayley Xuereb, Milana Maletic, Jeff Gildersleeve, Istva´n Pelczer, and Daniel Kahne*
Contribution from the Department of Chemistry, Princeton UniVersity, Princeton, New Jersey, 08544
ReceiVed July 16, 1999
Abstract: Many biological recognition events involve extensive interactions between macromolecules. Strategies
to design compounds that mimic large peptides or other elements of secondary structure could be useful for
blocking interactions between large surfaces. In this paper, the use of oligosaccharides as scaffolds for the
design of peptidomimetics is addressed. A functionalized oligosaccharide modeled after a basic region peptide
helix has been designed and synthesized. This oligosaccharide binds to duplex DNA with micromolar affinity.
The mode of binding has been established by 2D NMR, and shows that the oligosaccharide binds DNA in the
minor groove with one surface of the oligosaccharide contacting the floor of the DNA.
Introduction
The use of functionalized sugar scaffolds as peptide mimetics
has also received some attention.2a,b,c,i,j,3 For example, in 1993,
Hirschmann and co-workers demonstrated that an appropriately
functionalized glucose residue could serve as a mimetic of a
biologically active cyclic peptide.4 Hirschmann noted that
monosaccharides are good scaffolds for the presentation of
amino acid side chains because they can be derivatized easily,
come in a variety of different stereoisomers, and have well-
defined conformations, making it easy to predict the position
of the functional groups. Oligosaccharides may make good
scaffolds for mimicking larger elements of peptide structure
because they too come in a wide variety of well-defined
conformations and can be readily functionalized.5
In this paper, we present the synthesis and evaluation of a
functionalized pentasaccharide 1 (Figure 1) that was designed
to bind to duplex DNA. The design of the oligosaccharide
scaffold and the development of an efficient synthetic route to
the scaffold are described. The solution structure of the
pentasaccharide has been determined by NMR and is consistent
with the predicted structure. The functionalized oligosaccharide
binds to DNA, and the mode of binding, determined by 2D
NMR, shows that in several important respects the scaffold binds
to DNA as designed.
Biomolecular recognition events such as protein-protein and
protein-DNA interactions typically involve contacts over a large
surface area. For example, it is quite common for an R helix in
a protein to contact another molecule over its entire length. There
are no clear prescriptions for how to design compounds that
block interactions between large surfaces, and this is one of
the most challenging problems facing researchers interested in
problems of molecular recognition. The obvious strategy is to
design compounds that mimic peptide R helices or other
elements of secondary structure. Some of the most exciting work
in the area involves the use of polymers composed of â amino
acids that fold into stable secondary structures.1,2 Gellman and
others have been exploring strategies to stabilize helical â
peptides in water, and it seems likely that these structures can
be functionalized to serve as mimetics of more standard
polypeptides.2m,n,p
(1) For a review of engineered biopolymers with well-defined folding
properties, see: Gellman, S. Acc. Chem. Res. 1998, 31, 173-180.
(2) For recent work in the area of biopolymer design, see: (a) Nicolaou,
K. C.; Florke, H.; Egan, M. G.; Barth, T.; Estevez, V. A. Tetrahedron Lett.
1995, 36, 1775-1778. (b) Suhara, Y.; Ichikawa, M.; Hildreth, J. E. K.;
Ichikawa, Y. Tetrahedron Lett. 1996, 37, 2549-2552. (c) Suhara, Y.; Izumi,
M.; Ichikawa, M.; Penno, M. B.; Ichikawa, Y. Tetrahedron Lett. 1997, 38,
7167-7170. (d) Abele, S.; Guichard, G.; Seebach, D. HelV. Chim. Acta
1998, 81, 2141-2156. (e) Armand, P.; Kirshenbaum, K.; Goldsmith, R.
A.; Farr-Jones, S.; Barron, A. E.; Truong, K. T.; Dill, K. A.; Mierke, D. F.;
Cohen, F. E.; Zuckermann, R. N.; Bradley, E. K. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 4309-4314. (f) Hintermann, T.; Gademann, K.; Jaun, B.;
Seebach, D. HelV. Chim. Acta 1998, 81, 983-1002. (g) Hanessian, T.; Luo,
X.; Schaum, R.; Michnick, S. J. Am. Chem. Soc. 1998, 120, 8569-8570.
(h) Kishenbaum, K.; Barron, A. E.; Goldsmith, R. A.; Armand, P.; Bradley,
E. K.; Troung, K. T.; Dill, K. A.; Cohen, F. E.; Zuckerman, R. N. Proc.
Natl. Acad. Sci. U.S.A. 1998, 4303-4308. (i) Smith, M. D.; Claridge, T.
D. W.; Trnater, G. E.; Sansom, M. S.; Fleet, G. W. J. Chem. Soc., Chem.
Commun. 1998, 2041-2042. (j) Szabo, L.; Smith, B. L.; McReynolds, K.
D.; Parill, A. L.; Morris, E. R.; Gervay, J. J. Org. Chem. 1998, 63, 1074-
1078. (k) Yang, D.; Qu, J.; Li, B.; Ng, F.; Wang, X.; Cheung, K.; Wang,
D.; Wu, Y. J. Am. Chem. Soc. 1999, 121, 589-590. (l) Prince, R. B.; Okada,
T.; Moore, J. S. Angew. Chem., Int. Ed. Engl. 1999, 38, 233-236. (m)
Appella, D. H.; Christianson, L. A.; Karle, I. L.; Powell, D. R.; Gellman,
S. H. J. Am. Chem. Soc. 1999, 121, 6206-6212. (n) Appella, D. A.; Barchi,
J. J.; Durell, S. R.; Gellman, S. H. J. Am. Chem. Soc. 1999, 121, 2309-
2310. (o) Gung, B. W.; Zou, D.; Stalcup, A. M.; Cottrell, C. E. J. Org.
Chem. 1999, 64, 2176-2177. (p) Appella, D. H.; Christianson, L. A.; Klein,
D. A.; Richards, M. R.; Powell, D. R.; Gellman, S. H. J. Am. Chem. Soc.
1999, 121, 7574-7581.
Results and Discussion
Design. We have utilized structural motifs found in DNA
binding proteins to design a DNA-binding oligosaccharide.
Many proteins that bind to DNA utilize R helical recognition
peptides. Most recognition helices are part of a larger folded
(3) Sofia, M. J.; Hunter, R.; Chan, T. Y.; Vaughan, A.; Dulina, R.; Wang,
H.; Gange, D. J. Org. Chem. 1998, 63, 2802-2803.
(4) (a) Hirschman, R.; Nicolaou, K. C.; Pietranico, S.; Leahy, E. M.;
Salvino, J.; Arison, B.; Cichy, M. A.; Spoors, P. G.; Shakespeare, W. C.;
Sprengeler, P. A.; Hamley, P.; Smith, A. B., III; Reisine, T.; Raynor, K.;
Maechler, L.; Donaldson, C.; Vale, W.; Friedinger, R. M.; Cascieri, M. R.;
Strader, C. D. J. Am. Chem. Soc. 1993, 115, 12550-12568. (b) Hirschmann,
R.; Haynes, J., J.; Cichy-Knight, M. A.; Van Rijn, R. D.; Sprengeler, P.
A.; Spoors, P. G.; Shakespeare, W. C.; Pietranico-Cole, S.; Barbosa, J.;
Liu, J.; Yao, W.; Rohrer, S.; Smith, A. B., III J. Med. Chem. 1998, 41,
1382-1391.
(5) Studies done with VSG proteins have shown that a trisaccharide can
substitute for an R helix: Blum, M. L.; Down, J. A.; Gurnett, A. M.;
Carrington, M.; Turner, M. J.; Wiley, D. C. Nature 1993, 362, 603-609.
10.1021/ja992513f CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/23/2000

1884 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Xuereb et al.
Figure 1. (A) The target pentasaccharide 1. (B) The target pentasaccharide 1 shown in its preferred conformation. (C) Pentasaccharide 1 (purple)
superimposed on residues 231-244 of GCN4 from a crystal structure of the DNA complex (green).7 Only arginine side chains within this region
are shown.
unit that stabilizes the recognition helix while also making some
contacts to the DNA. The basic regions of bZIP proteins,
however, contact DNA as isolated helices that are not stabilized
by tertiary contacts to other parts of the protein.6 Therefore,
basic regions are the simplest type of recognition helices to try
to mimic. Structural studies done with GCN4, a bZIP protein,
have shown that it binds to DNA with one surface of the helix
buried in the major groove.7 A number of arginine and lysine
side chains emerge from the flanks of the helix and contact the
phosphate backbone on both strands of the DNA duplex. The
peptide backbone itself does not make any direct contacts to
the DNA. In an abstract sense, a basic region peptide can be
viewed as a cylinder from which positively charged groups
emanate; the spacing between the charged groups is such that
the helix can make at least two contacts to each strand of the
phosphate backbone simultaneously.
Figure 2. Steric analysis of ψ angle conformers showing analogous
With an idealized basic region helix as a blueprint for design,
diisopropyl methane conformers.
we searched for a rodlike oligosaccharide scaffold that could
conformations around the φ angle in which the lone pair of the
exo oxygen is anti to the CO bond of the sugar ring (the so-
called exo-anomeric effect). Only two of the three possible
staggered conformations satisfy this criterion. For an equatorial
glycosidic linkage, both conformers are populated to a signifi-
cant extent. For axial glycosidic linkages, however, one of the
two conformers is highly disfavored as a result of severe steric
interactions. Therefore, axial glycosides adopt a single confor-
mation around the φ angle. The preferred conformation around
the ψ angle can be predicted by analyzing the nonbonded steric
interactions. By holding the φ angle constant and rotating around
the ψ angle, one can determine the conformation(s) that
minimize unfavorable steric interactions (Figure 2). The non-
bonded steric interactions that determine the preferred confor-
mation for the nonanomeric C-O bond are analogous to those
found in a diisopropylmethane unit. For example, just as is seen
for diisopropylmethane, there is only one ψ angle conformer
for the corresponding glycoside that does not have unfavorable
1,3-diaxial-like steric interactions.
be easily functionalized to permit electrostatic contacts to both
strands of the phosphate backbone. There is structural informa-
tion showing that R-1,4-linked oligosaccharides display corre-
sponding substituents on opposite sides of a central axis.8 We
desired an oligosaccharide in which positive charges emerge
from alternating sides of the molecule, a condition that appeared
to be met in these oligosaccharides. A conformational analysis
of R-1,4 glycosidic linkages shows that the conformation
observed in X-ray structures and by NMR for this type of
linkage is dictated by a combination of steric and electronic
preferences.
Conformational Analysis. The overall shape of an oligosac-
charide can be predicted by analyzing the low-energy conforma-
tions around the glycosidic linkages.9 In a glycosidic linkage
to a secondary alcohol, there is free rotation around two bonds,
the bond from the exocyclic oxygen to the anomeric carbon
and the bond from the exocyclic oxygen to the carbon of the
aglycone. The conformations around these two bonds are defined
by two angles, φ and ψ, respectively. Sugars preferentially adopt
On the basis of the preceding analysis, we chose to make a
homopolymer of R-1,4-linked fucose residues functionalized at
C3 (1, Figure 1A). This R-1,4-homopolymer was predicted to
have the rodlike conformation shown in Figure 1B. Molecular
mechanics calculations also identify this structure as the low-
energy conformation for 1.10 Because of the preferred confor-
mation around the glycosidic linkages, the C3 substituents on
(6) Hurst, H. C. Protein Profile 1994, 1, 123-168.
(7) Ellenberger, T. E.; Brandl, C. J.; Struhl, K.; Harrison, S. C. Cell 1992,
71, 1223-1237.
(8) (a) Faham, S.; Hileman, R. E.; Fromm, J. R.; Linhardt, R. J.; Rees,
D. C. Science 1996, 271, 1116-1120. (b) The conformation of the
trisaccharide portion of Aclacinomycin A is also consistent with this
observation: Yang, D.; Wang, A. H. J. Biochemistry 1994, 33, 6595-6604.
(9) Lemieux, R. U.; Koto, S. Tetrahedron 1974, 30, 1933-1944.

Design of an Oligosaccharide That Binds to DNA
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1885
Scheme 1. Synthesis of Pentasaccharide 1a
a Reaction conditions: (a) Tf2O, 2,6-di-tert-butyl-4-methylpyridine, CH2Cl2, -78 °C, 44% of 4, 9% of 4a, 21% of 5; (b) mCPBA, CH2Cl2, -78
°C, 85%; (c) HF‚pyr/THF, 95%; (d) Tf2O, 2,6-di-tert-butyl-4-methylpyridine, 4-allyl-3,4-dimethoxybenzene, CH2Cl2, -78 °C, 74%; (e) HF‚pyr/
THF, 90%; (f) (1) NaH, DMF and (2) CH3I, DMF, 97%; (g) Hg(OCOCF3)2, MeOH, CH2Cl2, 0°C (4:1 â:R), 94%; (h) H2, PtO2, MeOH, 90%; (i)
H-Gly-NHCBz, HOBT, EDIC, DIEA, 85%; (j) H2, Pd/C, MeOH, 74%; (k) (1) CBzNHC(SMe)NCBz,33 DMF, Et3N, 45 °C, 24 h, 85% and (2) H2,
Pd/C, MeOH, 90%.
consecutive sugars are displayed from opposite sides of the
scaffold. In 1, these C3 substituents have been functionalized
with guanidinium groups to mimic arginine side chains. Figure
1C shows an overlay of the low-energy conformation of
pentasaccharide 1 and the GCN4 basic region helix from a
crystal structure. There is reasonable correspondence between
the position of the charged groups in the two structures.
It is worth making some additional points about the design
of the oligosaccharide scaffold. 2-Deoxyfucose was chosen as
a monomer building block because it is more hydrophobic than
fully oxygenated sugars. The hydrophobicity was expected to
help facilitate groove binding.11,12 An oligosaccharide composed
of five monomers was chosen because the length is similar to
that of the DNA binding portion of a basic region helix.
Synthesis of Pentasaccharide 1. One of the critical steps
for the use of an oligosaccharide as a scaffold is the development
of an efficient synthesis. Choosing a homopolymer as our initial
target simplifies the synthesis of pentasaccharide 1 for two
reasons. First, it is possible to obtain the requisite monomers 2
and 3 from a single precursor (Scheme 1). Second, it is possible
to devise a highly convergent approach for coupling the
monomers together. This approach involves the use of a
polymerization reaction of the sulfoxide glycosylation method
to form disaccharide 4 and trisaccharide 5 in a single glycosy-
lation reaction. Following isolation, these two components could
then be coupled together to form a pentasaccharide using
conditions which did not promote polymerization. Using this
approach, only two glycosylation reactions would be required
to assemble the pentasaccharide core 8.
The synthetic strategy described above grew out of mecha-
nistic investigations on the sulfoxide glycosylation reaction.13,14
We have identified conditions to permit multiple glycosylation
reactions to occur to form a mixture of products of different
lengths. Mechanistic studies in this laboratory have shown that
there are two pathways in which an acceptor such as 2, which
contains an anomeric sulfide, and a sulfoxide donor such as 3
can react to give products longer than a disaccharide (Scheme
1). One way involves the loss of the TMS group on the donor
during the reaction to expose an acceptor site that can react
with another molecule of donor. The other way involves
activation of the anomeric sulfide at the reducing end of the
glycosyl acceptor by phenyl sulfenyl triflate, which is generated
during the course of the sulfoxide reaction. Both of these
processes work to our advantage in the first glycosylation
reaction, in which we want to obtain both di- and trisaccharide
(13) (a) Raghavan, S.; Kahne, D. J. Am. Chem. Soc. 1993, 115, 1580-
1581. (b) Gildersleeve, J.; Pascal, R. A.; Kahne, D. J. Am. Chem. Soc. 1998,
120, 5961-5969. (c) Gildersleeve, J.; Smith, A.; Sakurai, K.; Raghavan,
S.; Kahne, D. J. Am. Chem. Soc. 1999, 121, 6176-6182.
(14) For other mechanistic studies on the sulfoxide reaction, see: (a)
Crich, D.; Sun, S. J. Am. Chem. Soc. 1997, 119, 11217-11223. (b) Crich,
D.; Sun, S. J. Am. Chem. Soc. 1998, 120, 435-436.
(10) Molecular modeling studies were done on pentasaccharide 1 using
published methods, see: Senderowitz, H.; Parish, C.; Still, W. C. J. Am.
Chem. Soc. 1996, 118, 2078-2086.
(11) Kahne, D. Chem., Biol. 1995, 2, 7-12.
(12) Ding, W.; Ellestad, G. A. J. Am. Chem. Soc. 1991, 113, 6617-
6620.

1886 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Xuereb et al.
Scheme 2. Potential Mechanisms for Trisaccharide 5 Formation
products. However, it is important to suppress processes that
lead to polymerization in the second glycosylation reaction to
ensure that only pentasaccharide is formed. We have identified
conditions to suppress unwanted activation of anomeric sulfides
during glycosylations using anomeric sulfoxide donors. Hence,
as described below, our synthesis involved two successive
glycosylation reactions run under different conditions (Scheme
1).
Table 1. 1H Chemical Shifts (ppm) of the Nonexchangeable
Pentasaccharide Resonances in the Free Pentasaccharide 1
(Chemical Shift Differences are also Reported for Pentasaccharide 1
in the Pentasaccharide 1-d[GGAATTCC]2 Complex35)
A
B, C, D35
E
H1
4.61(+0.01) 5.04(+0.01)
5.01(+0.03)
1.89(+0.02)
1.84(+0.02)
4.35(+0.02)
H2ax 1.64(+0.02) 2.02-2.03(+0.01)
H2eq 1.84(+0.02) 1.86-1.88(+0.01)
H3
H4
H5
4.04(+0.02) 4.32-4.33(+0.01)
Monomer 2 was prepared from commercially available
rhamnose.15 Monomer 3 was synthesized from 2 via silylation
and oxidation. The monomers were then coupled under condi-
tions that permit sulfide activation to form 4 and 5 in a combined
yield of 65%. Disaccharide 4 and trisaccharide 5 were converted
to the requisite donor disaccharide 7 and acceptor trisaccharide
6, respectively, as shown in Scheme 1. The glycosylation
reaction was carried out using the inverse addition procedure,
which involves the slow addition of sulfoxide to a pre-cooled
solution of acceptor, triflic anhydride, and base. This inverse
addition method of activation prevents anomeric sulfenate
formation, which is a facile side reaction with activated 2-deoxy
donors. Trisaccharide 6 and disaccharide 7 were then coupled
to form pentasaccharide 8. The sulfoxide glycosylation condi-
tions in this case included the addition of a phenyl sulfenyl
triflate scavenger to suppress sulfide activation and prevent the
formation of undesired higher order oligomers. The pentasac-
charide was produced in 74% yield as a single stereoisomer,
and no higher order oligomers were detected.16 Using this
approach, the pentasaccharide core was obtained rapidly and
efficiently in an overall yield of 26% from monomers 2 and 3.
3.76(+0.01) 3.83(+0.01), 3.87-3.88(+0.01) 3.41(+0.02)
3.76(+0.01) 4.22-4.23(+0.02)
1.23(+0.02) 1.17-1.18(+0.02)
4.19(+0.02)
1.20(+0.02)
3.48(+0.02)
CH3
OCH3 3.47(-0.04)
The final stage of the synthesis of the target pentasaccharide
involved the installation of the guanidinium groups. This was
effected by simultaneously reducing the five azides to give the
pentamine 12. The amines were reacted with the activated
glycine derivative followed by removal of the protecting groups
to generate 14. In the final stage, the guanidine moiety was
installed to provide 1. The synthesis of pentasaccharide 1 is
versatile and efficient, allowing for many analogues with
different functionality to be made and tested.
Solution Conformation of 1. With the pentasaccharide in
hand, we conducted a series of NMR experiments to determine
whether 1 adopts the predicted conformation in solution. The
nonexchangeable proton resonances in D2O were assigned using
a combination of 2Q-HoMQC, TOCSY, ROESY, and NOESY
experiments. Assignments were readily made for the terminal
sugars (A and E). Assignments were slightly more difficult for
the internal sugars (B, C, and D) due to overlap of some of the
resonances. However, the individual spin systems could be
assigned and the chemical shifts of at least some of the protons
were sufficiently different to permit sequence-specific assign-
ments to be made (Table 1). Interresidue NOE cross-peaks
between CH3 and H1′, H4 and H1′, and H2 and H5′ were
observed around all four glycosidic linkages as shown in Figure
3. These NOEs are consistent with the predicted glycosidic
linkage conformation shown in Figure 1B.
(15) Compounds 2 and 3 were synthesized from readily available
2-deoxy-3-azido-4-O-acetyl-L-rhamnopyranoside (Florent, J. F.;, Monneret,
C. J. Chem. Soc., Chem. Commun. 1987, 1171-1172). (1) Ac2O, pyridine,
DMAP, room temperature, 20 min; (2) PhSH, BF3‚Et2O, CH2Cl2, 20 min,
room temperature, 93% over two steps; (3) K2CO3, MeOH, room temper-
ature, 2 h; (4) Tf2O, pyridine, CH2Cl2, 1 h, 75% over two steps; (5) KOBz,
18-crown-6, DMF, 4 h; (6) LiOH, MeOH, 99% over two steps to give 2;
(7) TMSOTf, TEA, CH2Cl2, -78 °C, 20 min; (8) mCPBA, CH2Cl2, -42
°C, 92% over two steps to give 3. For experimental details, see: Smith, A.
Ph.D. Thesis, Princeton University, 1997.
(16) This result suggests that the primary pathway for polymerization
of these substrates involves activation of the anomeric sulfide with
phenylsulfenyl triflate (Pathway 2, Scheme 2).

Design of an Oligosaccharide That Binds to DNA
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1887
Table 2. Energetics of DNA Binding for (A) Pentasaccharide 1
and (B) Pentalysine
Table 3. 1H Chemical Shifts (ppm) of the Nonexchangeable DNA
Resonances in the Free DNA and the Pentasaccharide
1-d[GGAATTCC]2 Complex
[Na+]
(M)
Ka(obs)
(M-1
∆G(obs)
(kcal/mol)
∆Gpe
(kcal/mol)
∆GT
(kcal/mol)
H8/H6
H5
H2
CH3
H1′
)
G1 7.80(+0.01)
G2 7.80(+0.01)
A3 8.15(+0.02)
A4 8.15(+0.02)
T5 7.14(+0.01)
T6 7.14(+0.01)
5.55(+0.03)
5.40(0.00)
6.01(+0.02)
6.18(+0.01)
(A) pentasaccharide 1
0.015
0.040
0.090
3.0 × 106
2.1 × 105
2.6 × 104
-8.83
-7.25
-6.01
-6.59
-5.05
-3.78
-2.24
-2.20
-2.23
7.26(+0.02)
7.62(+0.01)
1.24(+0.03) 5.94(0.00)
1.52(+0.02) 6.12(0.00)
6.01(-0.01)
(B) pentalysine
-7.85
0.015
0.040
0.090
5.8 × 105
5.1 × 104
5.0 × 103
-6.59
-5.05
-3.78
-1.27
-1.37
-1.26
C7 7.56(+0.01) 5.67(+0.01)
C8 7.62(-0.02) 5.75(-0.05)
-6.41
6.21(0.00)
-5.04
Figure 4. NOESY cross-peaks between (A) A4H2 and a hexose
methyl, (B) A3H2 and a hexose methyl, and (C) A4H2 and a hexose
H2′/H2′′ pair.
Figure 3. (A) The estimated distances corresponding to NOE cross-
peaks observed between H1′ and CH3, H1′ and H4, and H5′ and H2.
These distances were consistent with the preferred conformation of 1
shown in Figure 1B. (B) Interresidue cross-peaks between H1′ and CH3,
and H1′ and H4.
double stranded regions of yeast tRNA even at concentrations
2-fold higher than were used to displace ethidium from duplex
DNA.
These preliminary studies indicated that pentasaccharide 1
discriminates between duplex DNA and duplex RNA, raising
the possibility that there might be a specific binding mode.
Consequently, we undertook an investigation of DNA binding
using NMR. A 0.5:1 and 1:1 complex of the pentasaccharide
and the self-complementary DNA duplex d-(G1G2A3A4T5T6C7C8)2
were prepared in buffered water.20,21 There were small chemical
shift differences between the free DNA and pentasaccharide
controls and the 1:1 complex, consistent with an interaction
between the DNA and 1 (Tables 1 and 3). A single set of
resonance lines were observed for the 0.5:1 pentasaccharide/
DNA complex, indicating that the ligand is in fast exchange
between the free and bound states.
The NOESY spectra of the 1:1 complex confirms that the
pentasaccharide binds to DNA and also provides evidence about
the mode of binding. Although the ligand is in fast exchange,
we were able to identify several intermolecular NOEs. The
strongest resolved NOE cross-peaks were between the adenine
H2 protons and two of the pentasaccharide methyl groups as
well as one pair of 2-deoxy protons (Figure 4). All three of
these groups of protons were assigned to the internal sugars in
the pentasaccharide. Although sequence-specific assignments
for the three internal sugars in the complex could not be made
with certainty, only one set of assignments can account for these
NOEs. In the low-energy conformation of the pentasaccharide,
the C5 methyl groups of residues B and D, and the 2-deoxy
protons on residue C are located on the same side of the
DNA Binding. As a preliminary evaluation of the pentasac-
charide, the DNA binding properties were measured using an
ethidium bromide assay.17 The pentasaccharide was found to
bind to calf thymus DNA with a dissociation constant of ∼10-6
M. Virtually all positively charged compounds interact with
DNA under appropriate conditions, and the estimated affinity
of the pentasaccharide is only slightly higher than that of
pentalysine, a flexible peptide containing the same net charge.
Pentalysine binds to DNA almost entirely by electrostatic
interactions. To determine whether the binding of pentasaccha-
ride 1 is also due primarily to electrostatic interactions, we
measured the salt dependence of the binding constant.18 The
results show that the polyelectrolyte contribution (∆Gpe) to
binding is virtually identical for both pentasaccharide 1 and
pentalysine (Table 2). However, the nonpolyelectrolyte contri-
bution to binding (∆GT) is 1 kcal/mol higher for pentasaccharide
1 than for pentalysine. Hence, the increase in binding affinity
for pentasaccharide 1 relative to pentalysine is due to something
other than electrostatic interactions.
The next step was to try to determine if pentasaccharide 1
discriminates between duplex DNA and other types of nucleic
acids. To address this question, we investigated the ability of 1
to displace ethidium bromide from yeast RNA.19 RNA also has
a negatively charged phosphate backbone and double stranded
regions, like DNA. The conformation of an RNA duplex,
however, is quite different from the conformation of a typical
B form DNA duplex. For example, an RNA duplex has a wider
and shallower minor groove and contains phosphates that are
spaced differently than in a DNA duplex. The assay results
showed that pentasaccharide 1 did not displace ethidium from
(20) This DNA sequence has been well-characterized by NMR: (a) Patel,
D. J. Eur. J. Biochem. 1979, 99, 369-378. (b) Patel, D. J.; Canuel, L. L.
Eur. J. Biochem. 1979, 96, 267-276. (c) Broido, M. S.; Zon, G.; James, T.
L. Biochem. Biophys. Res. Commun. 1984, 119, 663-670.
(17) LePecq, J. B.; Paoletti, C. J. Mol. Biol. 1967, 27, 87-106.
(18) Chaires, J. B. Anti-Cancer Drug Des. 1996, 569-580.
(19) Sakai, T. T.; Torget, R.; I, J.; Freda, C. E.; Cohen, S. S. Nucleic
Acids Res. 1975, 2, 1005-1022.
(21) An internally AT-rich DNA sequence was chosen since it has been
shown that arginine residues in peptides prefer AT regions in DNA:
Geierstanger, B. H.; Volkman, B. F.; Kremer, W.; Wemmer, D. E.
Biochemistry 1994, 33, 5347-5355.

1888 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Xuereb et al.
using silica gel 60 (230-400 mesh) from EM Science/Bodman. CM-
Sephadex ion-exchange resin was purchased from Sigma.
All reactions were carried out under argon atmosphere with dry,
freshly distilled solvents under anhydrous conditions unless otherwise
noted. All reagents were purchased from commercial suppliers and used
without further purification unless otherwise noted.
DNA and RNA Binding Assay. Sonicated calf thymus DNA was
purchased from Pharmacia Biotech. Yeast-soluble RNA was obtained
from Calbiochem. An absorbance was measured and the concentration
of RNA was determined using a millimolar extinction coefficient of
570 mM-1 cm-1 at 260 nm.22 DNA binding was conducted at 15 mM
NaCl, 25 mM Tris, pH 7.5. RNA binding was conducted at 50 mM
KCl, 50 mM potassium cacodylate, pH 7.0. Fluorescence measurements
were made using a Perkin-Elmer LS 50 fluorescence spectrophotometer.
Binding constants were determined by carrying out fluorescence
titrations in which calf thymus DNA or yeast RNA concentrations were
fixed at 1 µM each. Stock solutions of pentasaccharide 1 were made
in the respective buffer for titration. The excitation wavelength for DNA
studies was set to 546 nm and all emitted light passing through a 600
nm cutoff filter was collected. For RNA, the excitation wavelength
was set to 520 nm and all emitted light passing through a 590 nm
cutoff filter was collected. All titrations were carried out at 25 °C. The
salt dependence of the equilibrium binding constant (Ka(obs)) was
determined using the same DNA binding buffer at 15, 40, and 90 mM
NaCl. Fluorescence titrations were directly fit to obtain binding
constants using Kaleidagraph (Synergy Software, Reading, PA).
NMR. Purified DNA was synthesized on a 10 µmol scale at the
Princeton Synthesis Facility. Following dialysis to remove TEA salts,
the strands were lyophilized and dissolved in 1 mL of NMR buffer
(25 mM phosphate, pH 7). An absorbance was measured and the
concentration was determined from the calculated extinction coefficient
at 260 nm (54 × 103 M-1 cm-1).23 After annealing and repeated
lyophilization from D2O, a sample was prepared containing 2.4 mM
of the DNA duplex in 0.5 mL of D2O. A 0.5 or 1.0 equiv sample of
pentasaccharide 1 in 100 µL of D2O was added to the DNA. The sample
was lyophilized and redissolved in 0.5 mL of D2O. No precipitate was
observed.
Figure 5. Pentasaccharide 1 highlighting the hydrophobic contacts
made to DNA in the minor groove by (A) the C5 methyl group on
residue B, (B) the 2-deoxy protons on residue C, and (C) the C5 methyl
group on residue D.
molecule, creating a hydrophobic cluster (Figure 5). The same
interresidue NOEs are observed in the pentasaccharide in the
absence and presence of DNA, suggesting that the conformation
of the oligosaccharide does not change upon binding to DNA.
Therefore, we infer that the face of the pentasaccharide
containing the hydrophobic cluster, formed by methyl protons
on residues B and D and the H2 protons on residue C, is buried
in the minor groove giving rise to the observed NOESY cross-
peaks to the adenine H2 protons. It is worth noting that when
the oligosaccharide is docked into the minor groove with the
NOESY contacts satisfied, the basic side chains on the pen-
tasaccharide are positioned to contact both strands of the DNA
duplex, as designed.
Conclusion
All one-dimensional and two-dimensional NMR experiments were
run on a Varian Unity/INOVA NMR spectrometer at 600 MHz and at
10 °C, unless indicated otherwise. Through bond connectivities were
mapped by gradient-selected 2Q-HoMQC24 and TOCSY using a
z-filtered MLEV-16 based spin-lock period,25 while ROESY25 and
NOESY26 were used to determine through space interactions. The water
signal in H2O samples was suppressed using the WATERGATE
method27 with selective pulses of 1.7 ms. For 2D experiments in H2O
additional gradient pulses were applied during the t1 incremented delay
and for most of the NOESY mixing time to suppress radiation
damping.28 Typical 2D acquisition parameters include a 8 kHz spectral
window in both dimensions, t1max of 64-100 ms, 16-32 scans for each
increment, and States-Redfield sign discrimination29 in the remote
dimension.
We have designed a groove binding oligosaccharide that
interacts with duplex DNA. Based on a self-consistent set of
NOEs, we have shown that the oligosaccharide binds to DNA
in the minor groove with one surface of the oligosaccharide
contacting the floor of the DNA. The major driving force for
binding involves electrostatic contacts. Nevertheless, the mol-
ecule is able to discriminate between duplex DNA and duplex
RNA, showing that it has shape selectivity. Pentasaccharide 1
differs from the basic region helices that inspired its design in
that it binds in the minor groove rather than the major groove.
It is possible that the preference for major or minor groove
binding is related to the dimensions of the molecules. Basic
region peptides are larger in diameter than the functionalized
oligosaccharide. If so, it might be possible to facilitate a switch
in grooves by increasing the size of some of the substituents. It
may also be possible to manipulate sequence selectivity by
altering the groups that point in toward the floor of the groove.
One-dimensional spectra were processed by the Vnmr software
(Varian, Inc., Palo Alto, CA). All 2D spectra were processed off-line
with NMRPipe30 using standard procedures,29 and subsequently visual-
ized and analyzed in NMRView.31 Typical 2D processing parameters
(22) Lindahl, T.; Henley, D. D.; Fresco, J. R. J. Am. Chem. Soc. 1965,
87, 4961-4963.
(23) Patel, D. J.; Canuel, L. L. Eur. J. Biochem. 1979, 96, 267-276.
(24) Pelczer, I.; Bishop, K. D. Methods for Structure Elucidation by High-
Resolution NMR; Kover, K. E., Batta, G., Szantay, J. Cs., Eds.; Elsevier:
1997; pp 187-207.
(25) (a) Rance, M. J. Magn. Reson. 1987, 74, 557-564. (b) Bax, A. Isr.
J. Chem. 1988, 309-317.
(26) Neuhaus, D.; Williamson, M. P. The Nuclear OVerhauser Effect in
Structural and Conformational Analysis; VCH Publishers: New York, 1987.
(27) Piotto, M.; Suader, V.; Sklenar, V. J. Biomol. NMR 1992, 2, 661-
666.
Experimental Section
Synthesis of Pentasaccharide 1. One-dimensional NMR spectra
were recorded on a Varian UNITY/ Inova 500-MHz Fourier transform
NMR spectrometer. Chemical shifts (δ) are reported in parts per million
(ppm) downfield from tetramethylsilane (TMS) unless otherwise noted.
Coupling constants (J) are reported in hertz (Hz). Multiplicities are
abbreviated as follows: singlet (s), doublet (d), triplet (t), quartet (q),
multiplet (m), and broadened (br). Mass spectra were obtained on a
VG ZAB, VG 7070, HP 5989A (University of California, Riverside,
Mass Spectrometry Facility).
(28) Sklenar, V. J. Magn. Reson., Ser. A 1995, 114, 132-135.
(29) Pelczer, I.; Carter, B. G. Methods in Molecular Biology; Reid, D.
G., Eds.; Humana Press: Totowa, NJ, 1997; pp 71-156.
(30) Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeifer, J.; Bax,
A. J. Biomol. NMR 1995, 6, 277-293.
Analytical thin-layer chromatography (TLC) was performed using
silica gel 60 F254 precoated plates (0.25 mm thickness) with a
fluorescent indicator. Flash column chromatography was performed
(31) Johnson, B. A.; Blevins, R. A. J. Biomol. NMR 1994, 4, 603-614.

Design of an Oligosaccharide That Binds to DNA
J. Am. Chem. Soc., Vol. 122, No. 9, 2000 1889
were as follows: time domain filtering of the residual solvent signal if
necessary, apodization with a slightly shifted cosine function in
combination with additional Gaussian broadening in both dimensions,
removal of the axial peaks using time domain filtering again, baseline
correction with 4th-order polynomial as necessary, and zero filling in
both dimensions. Final size of the frequency domain spectra was 8K
× 2K data points.
(m, 2H), 3.86-3.71 (m, 4H), 2.41-2.33 (m, 2H), 2.14-1.96 (m, 4H),
1.37-1.20 (m, 9H); 13C NMR (CDCl3, 125.8 MHz) δ 134.6, 131.3,
129.2, 127.5, 98.9, 83.7, 75.8, 75.5, 70.0, 67.9, 67.7, 67.1, 58.2, 57.1,
56.9, 30.9, 29.9, 28.7, 17.7, 17.6, 17.0; HRFABMS calcd for C24H33O6N9S
(M + Na+) 598.2172, found 598.2204.
Phenyl [[[[(3-Azido-2,3-dideoxy-4-trimethylsilyl-r-L-fucopyran-
osyl)]-(1-4)-(3-azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-
azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-azido-2,3-
dideoxy-r-L-fucopyranosyl)]-(1-4)-3-azido-2,3-dideoxy-1-thio-r-L-
fucopyranoside (8). To disaccharide 4 (150 mg, 0.30 mmol) in
methylene chloride (6 mL) at -42 °C was added 3-chloroperoxybenzoic
acid (114 mg of 65% dispersion, 1.03 mmol). The reaction was allowed
to warm to -20 °C, and dimethyl sulfide (30 µL) was added. The
reaction was diluted with methylene chloride (15 mL) and extracted
with saturated aqueous NaHCO3 (2 × 15 mL). The organic layer was
dried over Na2SO4 and concentrated in vacuo. The products were
purified by flash chromatography (35% EtOAc/petroleum ether) to
afford 7 (128 mg, 85%).
Phenyl (3-Azido-2,3-dideoxy-4-trimethylsilyl-r-L-fucopyranosyl)-
(1-4)-3-azido-2,3-dideoxy-1-thio-r-L-fucopyranoside (4). Alcohol 2
(518 mg, 1.95 mmol) and 2,6-di-tert-butyl-4-methylpyridine (1.06 g,
5.2 mmol) were combined and residual water was removed by
azeotropic distillation with toluene (3 × 20 mL). To the residue in
methylene chloride (26 mL) was added 4 Å molecular sieves (500 mg),
and the resulting suspension was stirred at room temperature. The
suspension was cooled to -78 °C, and triflic anhydride (219 µL, 1.3
mmol) was added over 1-2 min. A solution of sulfoxide 3 (460 mg,
1.3 mmol) in methylene chloride (17 mL) was added via syringe over
10-15 min. After 15 min at -78 °C, the reaction was quenched with
diethylamine (1 mL), filtered into saturated aqueous NaHCO3 (100 mL),
and extracted with methylene chloride (3 × 80 mL). The organic layers
were combined, dried over Na2SO4, filtered, and concentrated in vacuo.
The product was purified by flash chromatography (10% EtOAc/
petroleum ether) to afford disaccharide 4 (280 mg, 44%): Rf ) 0.44
(10% EtOAc/petroleum ether); 1H NMR (CDCl3) δ 7.47-7.46 (m, 2H),
7.32-7.24 (m, 3H), 5.71 (d, J ) 5.5 Hz, 1H), 5.08 (d, J ) 3.1 Hz,
1H), 4.33 (q, J ) 6.4 Hz, 1H), 4.25 (q, J ) 6.4 Hz, 1H), 3.96 (ddd, J
) 12.21, 3.7, 3.36 Hz, 2 H), 3.80-3.68 (m, 2H), 2.38 (ddd, J ) 13.0,
13.0, 5.8, 1H), 2.22 (ddd, J ) 12.5, 12.5, 3.7 Hz, 1H), 2.05-1.97 (m,
2H), 1.23 (d, J ) 6.7 Hz, 3H), 1.21 (d, J ) 6.4 Hz, 3H), 0.22 (2, 3H);
13C NMR (CDCl3, 125.8 MHz) δ 134.6, 131.1, 129.0, 127.2, 98.8, 83.5,
75.1, 71.1, 67.8, 67.7, 58.0, 56.6, 30.8, 28.2, 17.7, 17.4, 0.5; HRFABMS
calcd for C21H32O4N6SSi (M + Na+) 515.1873, found 515.1891.
Alcohol 6 (73 mg, 0.127 mmol), 4-allyl-1,2-dimethoxybenzene (218
µL, 1.27 mmol), and 2,6-di-tert-butyl-4-methylpyridine (130 mg, 0.63
mmol) were combined, and residual water was removed by azeotropic
distillation with toluene (3 × 10 mL). To the residue in methylene
chloride (2.3 mL) and diethyl ether (2.3 mL) was added 4 Å molecular
sieves (500 mg), and the resulting suspension was stirred at room
temperature. The suspension was cooled to -78 °C, and triflic anhydride
was added (43 µL, 0.25 mmol) over 1-2 min. A solution of sulfoxide
7 (128 mg, 0.25 mmol) in methylene chloride (1 mL) was added via
syringe over 10-15 min. After 15 min at -78 °C, the reaction was
quenched with diethylamine (150 µL), filtered into saturated aqueous
NaHCO3 (45 mL), and extracted with methylene chloride (3 × 35 mL).
The organic layers were combined, dried over Na2SO4, filtered, and
concentrated in vacuo. The product was purified by flash chromatog-
raphy (20% EtOAc/petroleum ether) to afford pentasaccharide 8 (90
mg, 74%): Rf ) 0.41 (20% EtOAc/petroleum ether); 1H NMR (CDCl3)
δ 7.45-7.44 (m, 2H), 7.32-7.25 (m, 3H), 5.71-5.70 (d, J ) 5.5 Hz,
1H), 5.08-5.05 (m, 4H), 4.35-4.23 (m, 5H), 4.01-3.96 (m, 5H), 3.77-
3.67 (m, 5H), 2.36 (ddd, J ) 12.9, 12.8, 5.5 Hz, 1H), 2.20 (ddd, J )
12.4, 12.4, 3.7 Hz, 1H), 2.04-1.97 (m, 8H), 1.29-1.18 (m, 15 H),
0.19 (s, 9H); 13C NMR (CDCl3, 125.8 MHz) δ 134.6, 131.3, 129.2,
127.5, 98.9, 98.8, 98.7, 83.7, 75.5, 75.3, 71.3, 67.8, 67.7, 60.6, 58.1,
57.0, 56.8, 30.9, 29.8, 28.3, 21.2, 17.8, 17.7, 17.5, 14.4, 0.6; HRFABMS
calcd for C39H59O10N15SSi (M + Na+) 980.3957, found 980.3951.
Methyl [[[[(3-Azido-2,3-dideoxy-4-O-methyl-r-L-fucopyranosyl)]-
(1-4)-(3-azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-azido-2,3-
dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-azido-2,3-dideoxy-r-L-fu-
copyranosyl)]-(1-4)-3-azido-2,3-dideoxy-r-L-fucopyranoside (11).
A solution (1.4 mL) of freshly prepared, buffered pyridinium hydro-
fluoride (10 mL of THF, 5.7 mL of pyridine, and 2.1 g of pyridinium
hydrofluoride)32 was added to a plastic vial containing 8 (90 mg, 0.094
mmol) in THF (3 mL). The reaction was quenched after 1 h with 5
mL of saturated aqueous NaHCO3. The aqueous layer was extracted
with methylene chloride (3 × 5 mL), and the organic layers were
combined, dried over Na2SO4, concentrated in vacuo, and purified by
flash chromatography (10% MeOH/CHCl3) to give 75 mg (90%) of 9:
Phenyl (3-Azido-2,3-dideoxy-r-L-fucopyranosyl)-(1-4)-3-azido-
2,3-dideoxy-1-thio-r-L-fucopyranoside (4a) (49 mg, 9%): Rf ) 0.17
(20% EtOAc/petroleum ether); 1H NMR (CDCl3) δ 7.46-7.44 (m, 2H),
7.32-7.25 (m, 3H), 5.69 (d, J ) 5.2 Hz, 1H), 5.04 (d, J ) 2.6 Hz,
1H), 4.34 (q, J ) 6.3 Hz, 2H), 3.96 (ddd, J ) 9.7, 4.2, 3.3 Hz, 2H),
3.84-3.71 (m, 2H), 2.37 (ddd, J ) 13.0, 13.0, 5.8 Hz, 1H), 2.15-
2.01 (m, 3H), 1.29 (d, J ) 6.7 Hz, 3H), 1.23 (d, J ) 6.7 Hz, 3H); 13
C
NMR (CDCl3, 125.8 MHz) δ 134.6, 131.3, 129.2, 127.5, 99.0, 83.7,
75.7, 69.9, 67.7, 67.2, 58.1, 57.0, 30.9, 28.7, 17.5, 17.2; HRFABMS
calcd for C21H24O4N6S (M + Na+) 443.1477, found 443.1497.
Phenyl [[(3-Azido-2,3-dideoxy-4-trimethylsilyl-r-L-fucopyrano-
syl)]-(1-4)-(3-azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-3-azido-
2,3-dideoxy-1-thio-r-L-fucopyranoside (5) (90 mg, 21%): Rf ) 0.30
(10% EtOAc/petroleum ether); 1H NMR (CDCl3) δ 7.45-7.44 (m, 2H),
7.32-7.24 (m, 3H), 5.70 (d, J ) 5.2 Hz, 1H), 5.06 (d, J ) 16.5 Hz,
2H), 4.34 (q, J ) 6.6 Hz, 1H), 4.29 (q, J ) 6.5 Hz, 1H), 4.24 (q, J )
6.3 Hz, 1H), 4.00-3.96 (m, 3H), 3.80-3.67 (m, 3H), 2.37 (ddd, J )
13.1, 13.0, 5.5 Hz, 1H), 2.20 (ddd, J ) 12.5, 12.5, 3.4 Hz, 1H), 2.05-
1.97 (m, 4H), 1.24 (d, J ) 6.7 Hz, 3H), 1.22 (d, J ) 6.7 Hz, 3H), 1.18
(d, J ) 6.4 Hz, 3H), 0.19 (s, 9H); 13C NMR (CDCl3, 125.8 MHz) δ
134.6, 131.3, 129.2, 127.4, 98.9, 98.8, 83.7, 75.4, 75.2, 71.3, 67.9, 67.8,
67.7, 58.1, 56.9, 56.8, 30.9, 29.8, 28.3, 17.8, 17.7, 17.5, 0.6; HRFABMS
calcd for C27H41O6N9SSi (M + Na+) 670.2567, found 670.2557.
1
Rf ) 0.31 (10% MeOH/CHCl3); H NMR (CDCl3) δ 7.45-7.44 (m,
2H), 7.32-7.24 (m, 3H), 5.70 (d, J ) 5.5 Hz, 1H), 5.05 (s, 4H), 4.37-
4.29 (m, 5H), 4.02-3.97 (m, 5H), 3.82-3.71 (m, 5H), 2.37 (ddd, J )
13.1, 13.0, 5.5 Hz, 1H), 2.14-1.98 (m, 9H), 1.28 (d, J ) 6.7 Hz, 3H),
1.27-1.22 (m, 12 H); 13C NMR (CDCl3, 125.8 MHz) δ 134.6, 131.3,
129.2, 127.4, 98.8, 98.7, 83.6, 75.8, 75.5, 70.0, 67.8, 67.7, 67.6, 67.1,
60.6, 58.1, 57.0, 56.9, 30.9, 29.8, 28.7, 21.2, 17.7, 17.5, 17.0, 14.4;
HRFABMS calcd for C36H51O10N15S (M + Na+) 908.3562, found
908.3610.
To a solution of 9 (75 mg, 0.085 mmol) in DMF (1.5 mL) was added
sodium hydride (4 mg of 95% in mineral oil, 0.127 mmol). The mixture
was allowed to stir for 5 min before addition of methyl iodide (11 µL,
169 µmol). After 20 min of stirring, the mixture was diluted with
methylene chloride and poured into saturated aqueous NaHCO3. The
organic layer was washed with saturated aqueous NaHCO3 (2 × 10
mL), dried over Na2SO4, concentrated, and purified by flash chroma-
tography (20% EtOAc/petroleum ether) to afford 10 (76 mg, 97%). Rf
Phenyl [[(3-Azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-
azido-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-3-azido-2,3-dideoxy-
1-thio-r-L-fucopyranoside (6). A solution (1.5 mL) of freshly prepared,
buffered pyridinium hydrofluoride (10 mL of THF, 5.7 mL of pyridine,
and 2.1 g of pyridinium hydrofluoride)32 was added to a plastic vial
containing 5 (90 mg, 0.139 mmol) in THF (1.5 mL). The reaction was
quenched after 0.5 h with 5 mL of saturated aqueous NaHCO3. The
aqueous layer was extracted with methylene chloride (3 × 5 mL), and
the organic layers were combined, dried over Na2SO4, concentrated in
vacuo, and purified by flash chromatography (20% EtOAc/petroleum
ether) to give 76 mg (95%) of 6: Rf ) 0.14 (20% EtOAc/petroleum
1
ether); H NMR (CDCl3) δ 7.45-7.44 (m, 2H), 7.32-7.24 (m, 3H),
5.70 (d, J ) 5.5 Hz, 1H), 5.05 (s, 2H), 4.40-4.29 (m, 3H), 4.03-3.96
(32) Trost, B. M.; Caldwell, C. G.; Murayama, E.; Heissler, D. J. Org.
Chem. 1983, 48, 3252-3265.

1890 J. Am. Chem. Soc., Vol. 122, No. 9, 2000
Xuereb et al.
) 0.27 (20%EtOAc/ petroleum ether); 1H NMR (CDCl3) δ 7.45-7.44
(m, 2H), 7.32-7.24 (m, 3H), 5.70 (d, J ) 5.5 Hz, 1H), 5.05 (s, 4H),
4.36-4.29 (m, 5H), 4.01-3.97 (m, 5H), 3.77-3.74 (m, 4H), 3.62 (s,
3H), 3.26 (s, 1H), 2.37 (ddd, J ) 13.1, 13.0, 5.8 Hz, 1H), 2.18 (ddd,
J ) 12.6, 12.5, 3.4 Hz, 1H), 2.04-1.91 (m, 8H), 1.28 (d, J ) 6.4 Hz,
3H), 1.26-1.22 (m, 12H); 13C NMR (CDCl3, 125.8 MHz) δ 134.6,
131.3, 129.2, 127.2, 98.9, 98.7, 83.7, 80.4, 75.5, 67.8, 67.7., 67.6, 61.8,
60.6, 58.1, 57.0, 56.9, 55.8, 30.9, 29.8, 29.0, 21.2, 17.7, 17.5, 17.1,
14.1; HRFABMS calcd for C37H53O10N15S (M + Na+) 922.3718, found
922.3722.
HRFABMS calcd for C82H106O26N10 (M + Na+) 1669.7177, found
1669.7158.
A solution of 13 was dissolved in methanol (46 mg, 0.028 mmol))
and 10% Pd/C (90 mg) was added. The reaction mixture was stirred
vigorously under hydrogen (1 atm) and the reaction was complete after
1.5 h. The catalyst was filtered off and the filtrate was concentrated.
An aqueous solution of the crude mixture was then passed over a 5
mL CM-Sephadex column equilibrated in 1 M NH4OH and washed
with water. The amine 14 was eluted with a NH4OH gradient and
washed off at 30 mM NH4OH. Fractions containing 14 were then
lyophilized overnight to provide 14 (20 mg, 74%). Rf ) 0.27 (5:4:4:1
EtOAc/iPrOH/H2O/NH4OH); 1H NMR (CD3OD) δ 5.04-5.00 (m, 4H),
4.77 (br, 1H), 4.52-4.38 (m, 5H), 4.23-4.19 (m, 4H), 4.08-4.06 (m,
1H), 3.82-3.79 (m, 2H), 3.70-3.68 (m, 2H), 3.53-3.49 (m, 3H), 3.37-
3.36 (m, 4H), 3.34-3.31 (m, 10H), 2.10-1.91 (m, 4H), 1.87-1.69
(m, 6H), 1.30-1.20 (m, 15H); 13C NMR (CD3OD, 125.8 MHz) δ 173.8,
108.1, 192.1, 99.7, 99.6, 79.9, 77.7, 77.4, 76.7, 72.5, 68.3, 61.8, 56.0,
49.2, 45.7, 45.5, 44.5, 44.2, 32.5, 30.6, 30.4, 17.7, 17.2, 17.0; HRFAB-
MS calcd for C42H76O16N10 (M + Na+) 999.5338, found 999.5375.
Methyl [[[[(3-N-((2-Gaunidino)acetamido)-(2,3-dideoxy-4-O-meth-
yl-r-L-fucopyranosyl)]-(1-4)-(3-N-((2-guanidino)acetamido))-2,3-dideoxy-r-L-
fucopyranosyl)]-(1-4)-(3-N-((2-guanidino)acetamido))-2,3-dideoxy-
r-L-fucopyranosyl)]-(1-4)-(3-N-((2-guanidino)acetamido))-2,3-
dideoxy-r-L-fucopyranosyl)]-(1-4)-3-N-((2-guanidino)acetamido))-
2,3-dideoxy-r-L-fucopyranoside (1). Compound 14 (20 mg, 0.02
mmol) was dissolved in DMF (0.5 mL). To this solution was added
triethylamine (57 µL, 0.41 mmol) followed by N,N′-di-CBz-S-meth-
ylisothiourea (147 mg, 0.41 mmol).33 The reaction mixture was stirred
at 45 °C for 2 days. The solvent was evaporated in vacuo, redissolved
in methylene chloride, and washed with saturated NaHCO3 (2 × 10
mL). The solvent was evaporated and the crude material was purified
by flash chromatography (10% MeOH/CHCl3) to give the final product
with CBz protecting groups (44 mg, 85%). Rf ) 0.20 (5% MeOH/
CHCl3); 1H NMR (CDCl3) δ 11.62 (br, 5H), 8.91-8.87 (br, 5H), 7.54-
7.52 (m, 1H), 7.35-7.26 (m, 50H), 7.20-7.18 (m, 3H), 6.21-6.20
(m, 1H), 5.18-5.16 (m, 10H), 5.10-5.04 (m, 10H), 4.88-4.81 (m,
4H), 4.51-4.50 (m, 4H), 4.39-4.37 (d, J ) 9.5 Hz, 1H), 4.16-3.88
(m, 15H), 3.57-3.56 (m, 1H), 3.48-3.37 (m, 10H), 3.20 (br, 1H),
1.80-1.51 (m, 10H), 1.25 (d, J ) 6.1 Hz, 3H), 1.19 (d, J ) 6.4 Hz,
3H), 1.13 (d, J ) 6.4 Hz, 3H), 1.10-1.08 (m, 6H); FABMS (relative
intensity) m/e 2552 (M+ + Na, 2.1), 2529 (M+ + 1, 2.2).
The CBz-protected final product (44 mg, 0.017 mmol) was dissolved
in MeOH (0.6 mL) and 44 mg of 10% Pd/C was added. The reaction
mixture was stirred vigorously under hydrogen (1 atm) and the reaction
was complete after 1.5 h. The catalyst was filtered off and the filtrate
was concentrated. An aqueous solution of the crude mixture was then
passed over a 5 mL CM-Sephadex column equilibrated in 1 M NH4-
HCO3 and washed with water. The pentasaccharide 1 was eluted with
a NH4HCO3 gradient and washed off at 1 M NH4HCO3. Fractions
containing 1 were then lyophilized overnight to provide 1 as the HCO3
salt (23 mg, 90%). Rf ) 0.23 (5:4:4:3 EtOAc/iPrOH/H2O/NH4OH);
1H NMR (D2O) δ 4.97-4.90 (m, 4H), 4.53 (d, J ) 9.5 Hz, 1H), 4.28-
4.26 (m, 4H), 4.16-4.11 (m, 4H), 3.99-3.75 (m, 14H), 3.71-3.68
(m, 2H), 3.42 (s, 3H), 3.41 (s, 3H), 3.39 (br, 1H), 1.98-1.93 (m, 3H),
1.83-1.77 (m, 6H), 1.61-1.54 (m, 1H), 1.21 (d, J ) 7.0 Hz, 3H),
1.18 (d, J ) 10.4 Hz, 3H), 1.16-1.10 (m, 9H); 13C NMR (D2O, 125.8
MHz)34 δ 172.2, 172.1, 171.8, 167.1, 160.3, 103.9, 101.9, 101.7, 81.6,
80.0, 79.8, 78.8, 74.6, 70.6, 70.3, 64.4, 59.2, 53.2, 51.7, 48.4, 47.9,
46.6, 46.4, 33.6, 32.1, 26.0, 19.5, 19.1, 19.0; MS (rel intensity), m/e
1188 (M+ + 1, 2.4), 616 (M+/2 + Na, 13), 594 (M+/2 + 1, 99).
A solution of 10 (68 mg, 0.076 mmol) in dry methylene chloride
(2.7 mL) and methanol (4 mL) was cooled to 4 °C for 10 min. Mercury-
(II) trifluoroacetate (97 mg, 0.227 mmol) was then added and allowed
to stir for 10 min. The mixture was poured into 1 N aqueous HCl (10
mL). The organic layer was washed with saturated NaHCO3 (2 × 10
mL), dried over Na2SO4, concentrated in vacuo, and purified by flash
chromatography (20% EtOAc in petroleum ether) to afford a mixture
of anomers of 11 (62 mg, 94%; 4:1 â:R). Rf(R anomer) ) 0.12, Rf(â
1
anomer) ) 0.10 (20% EtOAc/ petroleum ether); H NMR (CDCl3) δ
5.04 (d, J ) 8.2 Hz, 4H), 4.40 (d, J ) 1.8 Hz, 1H), 4.38-4.28 (m,
4H), 4.08-3.97 (m, 3H), 3.73 (s, 3H), 3.64-3.58 (m, 6H), 3.53-3.49
(m, 4H), 3.26 (s, 1H), 2.19 (ddd, J ) 12.4, 12.3, 9.8 Hz, 1H), 1.31 (d,
J ) 6.4 Hz, 3H), 1.28 (d, J ) 6.4 Hz, 3H), 1.24-1.23 (m, 9H); 13C
NMR (CDCl3, 125.8 MHz) δ 101.3, 99.1, 98.9, 98.7, 80.4, 75.5, 74.7,
71.6, 67.8, 61.8, 60.6, 60.0, 57.0, 56.8, 55.8, 31.5, 29.8, 29.0, 21.3,
17.7, 17.6, 17.1, 14.4; HRFABMS calcd for C32H51O11N15 (M + Na+)
844.3790, found 844.3826.
Methyl [[[[(3-Amino-2,3-dideoxy-4-O-methyl-r-L-fucopyranosyl)]-
(1-4)-(3-amino-2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-amino-
2,3-dideoxy-r-L-fucopyranosyl)]-(1-4)-(3-amino-2,3-dideoxy-r-L-
fucopyranosyl)]-(1-4)-3-amino-2,3-dideoxy-r-L-fucopyranoside (12).
A solution of 11 (30 mg, 0.0365 mmol) was dissolved in methanol (2
mL) and PtO2‚5H2O (18 mg) was added. The reaction mixture was
stirred vigorously under hydrogen (1 atm) and the reaction was complete
after 1.5 h. The catalyst was filtered off and the filtrate was concentrated.
An aqueous solution of the crude mixture was then passed over a 5
mL CM-Sephadex column equilibrated in 1 M NH4OH and washed
with water. The amine 12 was eluted with a NH4OH gradient and
washed off at 30 mM NH4OH. Fractions containing 12 were then
lyophilized overnight to provide 12 (23 mg, 90%). Rf ) 0.23 (5:4:4:1
EtOAc/iPrOH/H2O/NH4OH); 1H NMR (CD3OD) δ 4.94-4.87 (m, 4H),
4.42 (d, J ) 9.5 Hz, 1H), 4.15-4.10 (m, 4H), 3.62-3.60 (m, 7H),
3.48-3.46 (m, 4H), 3.30-3.22 (m, 5H), 2.87-2.84 (m, 1H), 1.90-
1.87 (m, 4H), 1.82-1.76 (m, 5H), 1.50-1.43 (m, 1H), 1.33-1.28 (m,
6H), 1.26-1.21 (m, 9H); 13C NMR (CD3OD, 125.8 MHz) δ105.4,
103.3, 102.1, 102.0, 101.9, 83.6, 83.5, 83.0, 82.6, 73.4, 69.6, 69.4, 62.8,
56.8, 51.6, 47.7, 47.5, 47.4, 36.7, 35.2, 34.8, 18.0, 17.8; HRFABMS
calcd for C32H61O11N5 (M + H+) 692.4446, found 692.4418.
Methyl [[[[(3-N-((2-Aminoacetamido))-(2,3-dideoxy-4-O-methyl-
r-L-fucopyranosyl)]-(1-4)-(3-N-((2-aminoacetamido))-2,3-dideoxy-
r-L-fucopyranosyl)]-(1-4)-(3-N-((2-aminoacetamido))-2,3-dideoxy-
r-L-fucopyranosyl)]-(1-4)-(3-N-((2-aminoacetamido))-2,3-dideoxy-
r-L-fucopyranosyl)]-(1-4)-(3-N-((2-aminoacetamido))-2,3-dideoxy-
r-L-fucopyranoside (14). Compound 12 (23 mg, 0.033 mmol) was
dissolved in DMF (1.7 mL). To this solution was first added N,N′-
diisopropylethylamine (90 µL, 0.5 mmol) followed by 1-hydroxyben-
zotriazole (68 mg, 0.5 mmol), 1-(3-dimethylaminopropyl)-3-ethylcar-
bodiimide hydrochloride (68 mg, 0.5 mmol), and carbobenzyloxyglycine
(105 mg, 0.5 mmol). The reaction was stirred for 2 h. The solvent was
evaporated in vacuo, redissolved in methylene chloride, and washed
with saturated NaHCO3 (2 × 5 mL). The solvent was evaporated and
the crude material was purified by flash chromatography (10% MeOH
in CHCl3) to give 13 (46 mg, 85%). Rf ) 0.28 (10% MeOH/ CHCl3);
1H NMR (CDCl3) δ 7.46-7.40 (br, 1H), 7.35-7.30 (m, 15 H), 7.17-
7.15 (br, 3H), 6.30 (br, 1H), 5.80-5.72 (br, 1H), 5.52 (br, 2H), 5.16-
5.06 (m, 10H), 4.89-4.83 (m, 4H), 4.46-4.39 (m, 5H), 4.21-4.11
(m, 5H), 3.88-3.82 (m, 10H), 3.59-3.54 (m, 1H), 3.48-3.42 (m, 10H),
3.22 (bs, 1H), 1.88-1.77 (m, 4H), 1.57-1.55 (m, 6H), 1.30-1.27 (m,
6H), 1.20-1.16 (m, 9H); 13C NMR (CDCl3, 125.8 MHz) δ 168.8, 156.7,
136.6, 128.8, 128.7, 128.6, 128.4, 128.3, 101.6, 99.9, 79.2, 71.8, 68.3,
67.5, 67.3, 67.1, 62.3, 56.7, 48.3, 44.6, 32.9, 31.3, 30.9, 17.5;
Acknowledgment. This work was supported by grants
GM53066 and GM42733 from the National Institutes of Health.
JA992513F
(33) Tian, Z.; Edwards, P.; Roeske, R. W. Int. J. Peptide Protein Res.
1992, 40, 119-126.
(34) Indirect chemical referencing was done using chemical shift
referencing ratios obtained from the following: www.bmrb.wisc.edu/ref_info/
cshift.html
(35) Assignments are reported to (0.005 ppm accuracy. For the internal
sugars B, C, and D overlap of resonances upon complexation to DNA made
it difficult to measure chemical shift changes.