Design, synthesis, and evaluation of N-acyl modified sialic acids as inhibitors of adenoviruses causing epidemic keratoconjunctivitis

Article abstract of DOI:10.1021/jm801609s

The adenovirus serotype Ad37 binds to and infects human corneal epithelial (HCE) cells through attachment to cellular glycoproteins carrying terminal sialic acids. By use of the crystallographic structure of the sialic acid-interacting domain of the Ad37

Full text of DOI:10.1021/jm801609s

3666
J. Med. Chem. 2009, 52, 3666–3678
Design, Synthesis, and Evaluation of N-Acyl Modified Sialic Acids as Inhibitors of Adenoviruses
Causing Epidemic Keratoconjunctivitis†
Susanne Johansson,‡ Emma Nilsson,§ Weixing Qian,‡ Delphine Guilligay,| Thibaut Crepin,| Stephen Cusack,| Niklas Arnberg,§
and Mikael Elofsson*,‡
Department of Chemistry, Umeå UniVersity, SE-901 87 Umeå, Sweden, Department of Clinical Microbiology, DiVision of Virology, Umeå
UniVersity, SE-901 87 Umeå, Sweden, and European Molecular Biology Laboratory, EMBL Grenoble, BP 181, 6 Rue Jules Horowitz,
38042 Grenoble Cedex 9, France
ReceiVed December 19, 2008
The adenovirus serotype Ad37 binds to and infects human corneal epithelial (HCE) cells through attachment
to cellular glycoproteins carrying terminal sialic acids. By use of the crystallographic structure of the sialic
acid-interacting domain of the Ad37 fiber protein in complex with sialyllactose, a set of N-acyl modified
sialic acids were designed to improve binding affinity through increased hydrophobic interactions. These
N-acyl modified sialic acids and their corresponding multivalent human serum albumin (HSA) conjugates
were synthesized and tested in Ad37 cell binding and cell infectivity assays. Compounds bearing small
substituents were as effective inhibitors as sialic acid. X-ray crystallography and overlays with the
Ad37-sialyllactose complex showed that the N-acyl modified sialic acids were positioned in the same
orientation as sialic acid. Their multivalent counterparts achieved a strong multivalency effect and were
more effective to prevent infection than the monomers. Unfortunately, they were less active as inhibitors
than multivalent sialic acid.
Introduction
drugs might be to target virus adhesion by using antiviral agents
that are themselves multivalent.
Despite 50 years of research, the development of antiviral
drugs is far behind the development of antibacterial agents.
Viruses are obligate intracellular parasites; they are totally
dependent on the cellular mechanisms of their host and are
thereby difficult to selectively target without disturbing some
of the essential functions of the host cell.1,2 Today, most antiviral
drugs on the market target the synthesis of new viral components
by deactivating viral enzymes. The need for new antiviral drugs
is growing rapidly as more virus diseases are recognized.3
During the past decade, nonenzymatic processes of the virus
life cycle such as attachment to host cells and fusion to and/or
penetration through the cell membrane have come into focus
for research aimed at development of new antiviral agents.4 Cell
surface oligosaccharides are crucial for the attachment of
microbes to and infection of host cells. Despite this central role
of oligosaccharides, there are relatively few carbohydrate-based
therapeutic agents. The reasons for this are their low absorption
after oral administration and sensitivity to metabolic breakdown,
leading to poor oral bioavailability and rapid clearance. Another
drawback is the weak, often millimolar affinity between
monovalent carbohydrates and proteins. Nature overcomes this
problem elegantly and improves the binding affinity enormously
by gathering carbohydrate conjugates closely together to produce
multivalent binding epitopes that simultaneously associate with
multiple protein receptors on the attaching cell or organism.
This multivalency phenomenon has been extensively discussed
and reviewed.5-7 Since viruses in general attach to host cells
in a multivalent manner, one strategy for finding new antiviral
Our aim was to find antivirals that target adhesion of
adenoviruses of types Ad8, Ad19, and Ad37 to human corneal
epithelial (HCEa) cells. These viruses are the causative agents
of the severe ocular condition epidemic keratoconjuctivitis
(EKC).8,9 At present, there are no licensed antiviral agents for
treatment of this disease. The viruses are transferred by contact,
such as via towels, handshakes, and contaminated tonometers
at ophthalmic departments. Thus, these infections are most
common in densely populated areas, predominantly in Asia but
also in North America and Europe.10-12 For example, between
half a million and one million individuals fall ill from EKC
every year in Japan alone.13 As a result of the infection, the
patient is handicapped for weeks, resulting in substantial
suffering and economic losses. In some cases the infection leads
to permanent impairment of sight.8
Adenoviruses interact with their cellular receptors through
the fiber proteins that extend from the virus particle. Each
particle has 12 homotrimeric fiber proteins, and thus, the receptor
binding domain of the fiber (the knob) presents three separate
binding sites.14 Recently, the cellular receptor of Ad37 was
found to be a glycoprotein carrying at least one terminal sialic
acid moiety linked through an R(2-3) glycosidic bond to the
neighboring saccharide chain. Thus, the ability of adenoviruses
Ad8, Ad19, and Ad37 to cause EKC is strongly associated with
their ability to bind to sialic acid on the cell surface.15,16 An
antiviral drug based on sialic acid might therefore have the
potential to block viral attachment and therefore be used
successfully to treat patients suffering from EKC. As discussed
above, weak interactions between carbohydrates and proteins
limit the use of carbohydrates as drugs. However, since EKC-
†
Compound 17a: PDB code 2WGT, structure factors r2wgtsf. Compound
17g: PDB code 2WGU, structure factors r2wgusf.
a Abbreviations: Ad, adenovirus; EKC, epidemic keratoconjunctivitis;
FITC, fluorescein isothiocyanate; FFU, fluorescent focus units; HATU, O-(7-
azabenzotriazole-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate;
HCE cells, human corneal epithelial cells; HSA, human serum albumin;
SA, sialic acid; SHEM, supplemented hormone epithelial medium.
* To whom correspondence should be addressed. Telephone: +46-786
9328. Fax: +46-90 786 7655. E-mail: mikael.elofsson@chem.umu.se.
‡
Department of Chemistry, Umeå University.
Department of Clinical Microbiology, Umeå University.
§
|
European Molecular Biology Laboratory.
10.1021/jm801609s CCC: $40.75
2009 American Chemical Society
Published on Web 05/20/2009

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3667
Chart 1. HSA Conjugates of 3′-Sialyllactose and Sialic Acid
Figure 1. With the crystallographic structure of the sialic acid
interacting knob domain of Ad37 in complex with 3′-sialyllactose,
molecular modeling was used to design a library of 10 functionalized
sialic acids (A-J).
causing adenoviruses attach to the host cells via multiple fiber
proteins, it is possible that a drug could be prepared by using
the multivalency approach, for example, using a glycoconjugate
with several sialic acid derivatives linked to a protein carrier.
The pharmacokinetic drawbacks of carbohydrate-based drugs
are likely to be less severe in this case, since the potential drug
would be administered directly to the eye, using a cream or
eye drops. Naturally occurring, multivalent carbohydrates are
often structurally complex and are generally only found in
limited amounts in nature; they are therefore not preferred in
drug development. By use of synthetic multivalent neoglyco-
conjugates instead, it is possible to control important structural
properties such as the valency number, size, and flexibility of
the conjugate.17
Recently, we reported two synthetic multivalent inhibitors
(F1, F2, Chart 1) of the EKC-causing adenovirus Ad37.18,19
The carbohydrate epitopes, coupled to an aminoalkyl spacer,
were conjugated to human serum albumin (HSA) with varying
degrees of valency by using squaric acid chemistry.20-22
Biological evaluation of these compounds as inhibitors of Ad37
indicated an increased degree of inhibition with higher orders
of multivalency and showed that multivalent conjugates ef-
fectively inhibit the binding of Ad37 to host cells and thereby
prevent infection. The synthesis of the sialic acid-based inhibitor
F2 is more efficient than synthesis of the sialyllactose inhibitor
F1. This is an important advantage in terms of time and cost of
producing an antiviral drug. In addition, a smaller and less
complex structure is generally easier to modify than a large
structure.
In an attempt to optimize the inhibitory power of conjugate
F2, we have now investigated the possibility of improving the
binding affinity of sialic acid for the Ad37 fiber protein. On
the basis of the crystal structure of the fiber knob protein of
Ad37 in complex with sialyllactose,23 we used structure-based
design and docking to design a small library of 10 sialic acid
derivatives in which the acetamide was replaced by different
amides, urethanes, or ureas (Figure 1, structures A-J). We have
developed a straightforward and efficient synthetic route to these
compounds and their multivalent counterparts, i.e., correspond-
ing HSA conjugates. The potency of these monovalent and
multivalent structures as inhibitors of EKC-causing adenoviruses
was evaluated in cell-binding assays, and the binding mode was
studied by crystallization of sialic acid derivatives with the fiber
knob protein of Ad37.
in order to achieve a higher affinity of binding to the fiber knob
of EKC-causing adenoviruses. From the crystallographic struc-
ture of the sialic acid-interacting knob domain of the Ad37 fiber
protein in complex with sialyllactose, four main interactions
between sialic acid and the protein were identified: one salt
bridge, two hydrogen bond interactions, and one hydrophobic
interaction (Figure 2). The axial carboxylic acid forms a salt
bridge with the amino group of Lys345, the hydroxyl group in
position 4 forms a hydrogen bond with the hydroxyl group of
Tyr312, the amide group binds to the carbonyl group of Pro317
through a hydrogen bond, and the N-acetyl group of sialic acid
is directed into a deep hydrophobic pocket formed by Tyr312,
Tyr308 of another protein monomer, Pro317, and Val322.23 We
hypothesized that a bigger and more hydrophobic group, instead
of the acetamide of sialic acid, would position itself in the
hydrophobic cavity of the binding pocket and thereby improve
the affinity because of increased hydrophobic interactions. Thus,
we concentrated on optimizing this interaction between the
hydrophobic pocket and the 5′-amide and keeping the rest of
the sialic acid unchanged to ensure that the three important
interactions mentioned above would be maintained. In order to
obtain information about the limitations of the pocket in terms
of size and shape, we docked a set of sialic acids where the
acetamide was replaced with different amides, urethanes, or
ureas (Figure 1, structures A-E) into the sialic acid binding
site.24 To validate the reliability of the dockings of the
derivatives A-E, an initial docking of a new conformation of
sialic acid into the binding site was performed. The calculated
binding modes for sialic acid were compared with the experi-
mental binding conformation of the sialic acid residue in
sialyllactose. All docking poses were clustered within a 1.5 Å
root-mean-square deviation (rmsd), indicating a strong consensus
for a single binding mode similar to the experimental pose.25
The results from the docking study of the functionalized sialic
acids indicated that the deep hydrophobic cavity of the pocket
was not as easy to enter as initially expected. A shelf at the
mouth of the hydrophobic pocket forces derivatives with larger
N-acyl substituents into major conformational changes, and the
N-acyl substituents are not turned down into the large cavity as
expected. The binding modes of these derivatives that are
actually afforded are not likely to find the critical interactions
discussed above. Except for sialic acid, the natural ligand, the
best binding poses were observed for the propanamide derivative.
Figure 3A shows an overlay of the 10 best binding poses of
the propanamide (displayed as line structures) with the sialic
acid residue in the crystallographic structure of sialyllactose in
complex with the Ad37 fiber protein (displayed as a capped
stick structure). The majority of calculated poses were clustered
Results and Discussion
Structure-Based Design and Docking. Structure-based
design was used to identify parts of sialic acid that can be varied

3668 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Johansson et al.
Figure 2. View of the sialic acid-interacting knob domain of the Ad37 fiber protein in complex with 3′-sialyllactose. The molecular surface of the
fiber knob is colored blue for polar parts, brown for nonpolar parts, and green for intermediate polarity. Four main interactions between the sialic
acid residue and the protein are highlighted:23 one salt bridge (COO--Lys345), two hydrogen bonds (4-OH-Tyr312, 5-NH-Pro317), and one
hydrophobic interaction between the N-acetyl group of sialic acid and the hydrophobic pocket formed by Tyr312, Tyr308 of another monomer,
Pro317, and Val322.
Figure 3. Overlays of the calculated binding poses for the functionalized sialic acids (displayed as lines) and the sialic acid residue of the experimental
ligand (displayed as capped sticks). (A) The poses of the propanamide are clustered in a single binding mode close to the experimental ligand,
indicating an affinity similar to that of sialic acid. (B) The poses of the pentanamide are spread out in the docked region of the binding site; thus,
that derivative was predicted to be a poor binder. The poses of isobutanamide (C) and benzamide (D) are clustered in new binding modes with
major differences compared to the experimental ligand, indicating other binding possibilities inside the binding site.
close to that of sialic acid and in a plausible conformation for
finding the important hydrogen bond interactions discussed
above. This strongly suggests that propanamide has a single
binding mode, and this structure was therefore predicted to be
a good binder with the ability to increase the binding affinity.
With the extended derivatives, i.e., the butanamide and pen-
tanamide, the docked poses were not clustered at all. They were
found to be spread out in the entire docking region of the binding
site, indicating that these structures have no specific binding
mode. In addition, by overlaying of the docked positions on
the experimental sialic acid residue, it could be seen clearly
that these compounds would hardly find the important hydrogen
bonding interactions. Thus, these compounds could be predicted
to be poor binders. In Figure 3B, 10 representative docking poses
of the pentanamide (lines) have been overlaid on the experi-
mental structure (capped sticks). The poses of the benzamide
and the isobutanamide differ from those of the other derivatives.
According to the results with butanamide and pentanamide, we
had expected that a larger substituent on the amide would cause
a poor fit to the binding site, resulting in an extensive spread of
the docking poses of the isobutanamide and the benzamide.
Indeed, these amide substituents would be too big to enter the
hydrophobic cavity and they are also positioned without any
chance of finding the three important hydrogen bonds found
for sialic acid, but interestingly, they are closely clusteredsin
one single binding pose for the benzamide and two binding
poses for the isobutanamide (Figure 3C and Figure 3D). This
might indicate a second possibility of binding inside the binding
site.

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3669
Chart 2. Primary Synthetic Targets
The deprotection of 2a-h to their corresponding sialic acids
was planned to go through three steps: hydrolysis of the SPh
group by NIS, H2O, and MeCN, deacetylation of the O-acetates
with methanolic sodium methoxide, and finally hydrolysis of
the methyl ester by LiOH. This strategy was initially tried out
for the derivatives 2a and 2b. The first step was quite
straightforward, and 7a and 7b were obtained in 60% and 84%
yields, respectively. After deacetylation and hydrolysis of 7a
and 7b with LiOH, however, the sialic acids expected (3a and
3b) could not be isolated. Instead, a product with a completely
different 1H NMR spectrum was observed. By careful investiga-
1
tion of LCMS chromatograms and H NMR spectra of the
products obtained and by extensive search of the literature, we
found a feasible explanation. According to published investiga-
tions of the thermal and physical stability of sodium N-
acetylneuraminate,30 degradation of sialic acid can occur under
thermal alkaline conditions whereby the sialic acid ring structure
is degraded. Although the hydrolysis was performed at room
temperature, all the chromatographic and spectroscopic data
indicate that deprotection of 7a and 7b resulted in an inseparable
mixture of the target sialic acids 3a and 3b and the degraded
derivatives 9a and 9b. Despite several attempts to avoid this
transformation, it was not possible to synthesize fully depro-
tected N-acyl modified sialic acids. However, for future pur-
poses the deprotected thiophenyl derivatives 10a-i have been
synthesized in 44-94% yield from 2a-h and 6 by convenient
deacetylation and hydrolysis.
Although the propanamide was the only derivative predicted
to have the ability to increase the affinity for the fiber protein
of Ad37, we did not discount the possibility that any of the
other derivatives might be active in a natural biological system.
The predictions above are based on computer-simulated
protein-ligand interactions, and the protein structure is a static
model of a dynamic structure. As with most docking approaches,
the docking algorithm used in this study only allows the ligand
and the hydrogens of the protein to be flexible and considers
the protein to be rigid. Thus, features such as induced fit or
other conformational changes that could occur upon binding
are not taken into account and structures other than the
propanamide could actually fit inside the binding site.26 In order
to get as much information as possible about the effect when
the acetamide is replaced, we decided to synthesize a set of
sialic acid derivatives that were predicted to range from poor
to good binders compared to sialic acid. The set contains all
the hydrophobic amides from the docking study and also small
and hydrophilic functionalities that were not docked (Figure 1).
In order to find a new strategy for the synthesis, it was decided
to base the library on the methyl glycoside of sialic acid 11
(Scheme 2) that was synthesized from sialic acid by published
procedures.31 From the crystallographic structure of the sialic
acid-interacting domain of Ad37 in complex with sialyllactose,
we have already shown that the lactose residue of 3′-sialyllactose
is directed out of the binding site toward the environment.23
Thus, we hypothesized that an R-methyl glycoside instead of
an R-hydroxyl group in the anomeric position would not change
the binding properties of the derivatives. To verify this
hypothesis before we started the synthesis of this second library,
we ran a cell-binding assay to compare the methyl glycoside
of sialic acid with sialic acid regarding their potential to inhibit
adenoviral adhesion to HCE cells. As expected, the results
showed no significant difference in inhibitory power (data not
shown).
Synthesis. In our previous work with the synthetic multivalent
inhibitor F2 (Chart 1), we used sialic acid derivative 1 (Chart
2) that was conjugated to a protein carrier by the aminopentyl
spacer.19 In the present study we intended to use the same
synthetic strategy to afford multivalent conjugates of selected
parts of the library, and therefore, the sialic acid derivative 2
(Chart 2) was chosen as a target structure. From 2, we planned
to synthesize both functionalized sialic acids 3 (Chart 2) and
also the aminopentyl derivative 4 (Chart 2) that can be
conjugated to HSA.
Synthesis of N-Acyl Functionalized Sialic Acids. Com-
mercial sialic acid was used as starting material, and synthesis
of the sialic acid thiophenyl derivative 5 (Scheme 1) was
performed according to published procedures.27,28 In order to
functionalize 5, the amine was first Boc-protected and then
deacetylated; this was followed by O-reacetylation to get 6.29
Transformation to the amides 2a-h was then performed by
removal of the Boc group with TFA in CH2Cl2, directly followed
by acylation of the amine. For the derivatives 2a-e, we used
the corresponding anhydride in the presence of a base, and to
get derivatives 2f, 2g, and 2h, we used benzoic acid, methyl
chloroformate, and methylisocyanate, respectively. The coupling
reactions were followed by LCMS. The larger anhydrides
required longer coupling times, and they also resulted in lower
yields due to O- to N-acetyl migration. In order to minimize
the amount of side products, the reaction was stopped when
the formation of the main product ceased, despite there being
some remaining starting material in the mixture. The crude
products were purified by column chromatography and also
preparative HPLC when necessary. The target derivatives 2a-h
were obtained in yields ranging from 52% to 89%, calculated
from 6.
Functionalization of the methyl glycoside 11 was performed
in four steps. The amine 15 was obtained by a procedure similar
to that discussed for the thiophenyl derivatives above. A
t-butoxycarbonyl group (Boc) was added to the acetamide, and
deacetylation of 12 was performed with methanolic sodium
methoxide. Column chromatography of the crude residue gave
the pure R isomer 13 in 65% yield from 11. The R-anomeric
configuration was established by determination of the coupling
constant between C1 and H3ax (J ) 6.63 Hz).32 Compound 14
was obtained in 92% yield after reacetylation of the hydroxyl
groups by Ac2O in pyridine.
Prior to the amide-functionalizing step, the Boc group was
removed with TFA in CH2Cl2, resulting in the free amine as
TFA salt (15), which was used immediately without any
purification. The synthesis of 16a-h was generally performed
using traditional amide coupling synthesis with activated
carboxylic acids. The carboxylic acid, HATU, and DIPEA were
added to the amine 15 in CH2Cl2 and stirred at room temperature
until no further reaction was observed with LCMS. The use of
activated carboxylic acids instead of anhydrides (Scheme 1)
shortened the coupling reaction time and minimized acetyl

3670 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Scheme 1. Synthesis of Amide-Functionalized Sialic Acidsa
Johansson et al.
a Reagents and conditions: (a) 90% TFA (aq), CH2Cl2, room temperature, 1-2 h. (b) 2a-e: (RCO)2O, DMAP, CH2Cl2, room temperature, 5-11 h. 2f:
PhCOOH, HATU, DIPEA, CH2Cl2, room temperature, 3 h. 2g: CH3OCOCl, DIPEA, CH2Cl2, room temperature, 1 h. 2h: CH3NCO, DIPEA, CH2Cl2, room
temperature, 4 h. (c) NIS, MeCN/H2O 10:1, room temperature, 3 h; (d) (1) NaOMe, MeOH, room temperature, (2) LiOH (aq) MeOH, room temperature.
Scheme 2. Synthesis of Amide-Functionalized Sialic Acid Methyl Glycosidesa
a Reagents and conditions: (a) Boc2O, DMAP, THF, reflux, 2 h; (b) NaOMe, MeOH, room temperature, 3 h; (c) Ac2O, Py, room temperature, overnight;
(d) 90% TFA (aq), CH2Cl2, room temperature, 1-2 h. (e) 16a-e: RCOOH, HATU, DMAP, CH2Cl2, room temperature, 3-5 h. 16f: (CF3CO)2O, DMAP,
CH2Cl2, room temperature, 0.5 h. 16g: CH3OCOCl, DIPEA, CH2Cl2, room temperature, 1.5 h. 16h: CH3NCO, DIPEA, CH2Cl2, room temperature, 4 h. (f)
(1) NaOMe, MeOH, room temperature, 2 h, (2) 2 M LiOH (aq), MeOH, room temperature, overnight; (g) 2 M LiOH (aq), MeOH, room temperature,
overnight.
migration. We nevertheless obtained low yields and acetate
migration for the most hydrophobic amides 16c and 16d. To
get 16f, 16g, and 16h, we used methyl chloroformate, methyl
isocyanate, and trifluoroacetanhydride, respectively. In order to
obtain pure products, extensive purification by repeated column
chromatography was necessary. All derivatives were obtained
with reasonable yields (26-92%). With the methyl glycoside
instead of the free hydroxyl group at the anomeric position,
there were no difficulties with the deprotection. The fully
deprotected compounds 17a-h were obtained in 74-100%
yield after O-deacetylation with methanolic sodium methoxide
and methyl ester hydrolysis with LiOH. Unfortunately, 17f was
obtained as an inseparable mixture with the corresponding free
amine and was therefore excluded from the library. The
unprotected Boc derivative 18 was easily obtained by hydrolysis
of 13 in quantitative yield.
terparts of 17a-c were synthesized and evaluated. It was
assumed that the sialic acid derivatives would display differences
in affinity for the fiber knob protein and that the differences
would be enhanced by multivalency. The multivalent com-
pounds 22a-c were synthesized from sialic acid derivative 619
(Scheme 3). In order to minimize the amount of synthetic work,
we decided to elongate 6 with the aminopentyl spacer needed
for further conjugation to HSA, before diversifying the sialic
acid amide. Commercially available Fmoc-aminopentanol was
glycosylated with compound 619 to give the pure R-isomer 19
in 68% yield after extensive purification. The R-anomeric
configuration of 19 was established by determination of the
coupling constant between C1 and H3ax (J ) 5.9 Hz). The amide
functionalizing step to 20a-c followed the carboxylic acid/
HATU procedure described above, and the amides were
obtained in 18-64% yield. In addition, it was found that methyl
ester hydrolysis and partial deacetylation occurred during Boc
removal according to LCMS analysis. This in part explains the
modest yield of the amide couplings, both in this case and for
Synthesis of Multivalent N-Acyl Functionalized Sialic
Acids. In order to investigate the differences in inhibitory power
of N-acyl modified sialic acids further, the multivalent coun-

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3671
Scheme 3a
a Reagents and conditions: (a) Fmoc-aminopentanol, molecular sieves (3 Å), AgOTf, 1 M IBr in CH2Cl2, CH3CN/CH2Cl2 (3:2), -72 °C, 4 h; (b) 90%
TFA (aq), CH2Cl2, room temperature, 1 h; (c) RCOOH, HATU, DIPEA, CH2Cl2, room temperature, 9-24 h; (d) NaOMe, MeOH, room temperature, 1 h;
(e) LiOH (aq), MeOH, room temperature overnight.; (f) didecyl squarate, DMF, Et3N, room temperature, 11 h; (g) NaHCO3, HSA (human serum albumin),
room temperature, 27 h.
Table 1. Evaluation of N-Acyl Modified Sialic Acids 17a-h (17f
Excluded), 18, and Commercial NeuGc as Inhibitors of Ad37
Attachment to HCE Cellsa
17g and the benzamide 17e were slightly efficient as inhibitors,
whereas all of the other sialic acids with large N-acyl substituents
had no effect at all. An assay that shows Ad37 infected cells as
fluorescent focus units (FFU) was used to investigate whether
the compound could prevent infection (Figure 4).15,35 Unlabeled
Ad37 virons were preincubated with a dilution series of sialic
acid or the N-acyl modified sialic acids 17a,b,c,e,g before being
incubated with HCE cells. Unbound virions were washed away,
and cell-associated virions were allowed to infect the cells.
Thereafter, the cells were incubated with rabbit polyclonal anti-
Ad37 serum followed by FITC (fluorescein isothiocyanate)
conjugated swine antirabbit IgG antibodies and examined with
an immunofluorescence microscope. The results confirmed the
activity of the propanamide 17a and the methyl carbamate 17g.
Both prevented Ad37 from infecting the cells to the same extent
as sialic acid, or more. The benzamide 17e, on the other hand,
had no effect in this assay. The N-acyl modified sialic acids
17b and 17c had no effect in the cell-binding assay, and as
expected, they were inactive in the infection assay. For several
of the compounds, most notably 17b, 17c, and 17e, infection
was enhanced when the concentration of compound was
increased from 1 to 2 mM. The underlying reason for this is,
however, unknown. The multivalent HSA conjugates 22a-c
were also evaluated in the cell infectivity assay (Figure 5). As
expected, the multivalent N-acyl modified sialic acids followed
the behavior of their monovalent counterparts. The conjugates
with large N-acyl substituents, 22b and 22c, did not prevent
Ad37 from infecting HCE cells. The multivalent propanamide
22a was active as an inhibitor, with a clear multivalency effect.
However, 22a was much less efficient than multivalent sialic
acid (17-valent), and thus, sialic acid remains the most efficient
inhibitor of EKC-causing adenoviruses.
compd
run 1
run 2
run 3
SA
+ +
+ +
-
+ +
no data
+
+ +
+ +
-
17a
17b
17c
17d
17e
17g
17h
18
-
-
no data
-
-
+
+ +
+
-
-
+
+ +
+
+
-
-
+
-
-
NeuGc
-
-
a Compounds are denoted + + if they were effective as inhibitors to
the same extent as sialic acid or more, + if they were less potent than
sialic acid, and - if they had no effect at all. Each experimental run consisted
of a duplicate dilution series of each compound (2, 4, 8, and 16 mM), and
the data represent the average of duplicate determinations (for experimental
details, see Supporting Information).
the monovalent library described above. The amides 20a-c were
fully deprotected under standard conditions, and the crude
products 4a-c were used in the following conjugation step
without further purification. Conjugation of 4a-c to HSA was
performed by the squaric acid strategy,20-22 and the intermediate
squaric decyl ester sialosides 21a-c were obtained in 15-81%
yield. The three neoglycoproteins 22a-c were subsequently
obtained with 14, 16, and 17 saccharides incorporated per HSA,
respectively. The average degree of incorporation was deter-
mined by MALDI-TOF MS using the center of the single-
charged neoglycoprotein peak. This was confirmed with gel
electrophoresis.
Biological Evaluation. In order to investigate whether the
modified sialic acids synthesized above could prevent attachment
of Ad37 virions to HCE cells, a cell-binding assay based on
35S-labeled virions was used (Table 1).33-35 S-Labeled virions
were preincubated with commercial sialic acid or the synthesized
N-acyl modified R-methylsialosides 17a-h (17f excluded), 18,
or NeuGc (N-glycolylneuraminic acid) before adding the virions
to HCE cells. After incubation, unbound virions were washed
away and cell-associated radioactivity was measured. The
experiment was repeated three times. Because of large standard
deviations and variation in results, it was only possible to make
a rough estimate of the activity of the compounds. The results
are presented in Table 1 and show that the propanamide 17a,
which was predicted to be a good inhibitor according to the
docking study, was as efficient as sialic acid in preventing
attachment of Ad37 to cells. In addition, the methylcarbamate
Crystallography. X-ray crystallography was used to inves-
tigate the exact binding of 17a, 17e, and 17g to the fiber protein
of Ad37 at the molecular level. From the results of the docking
study and the biological evaluation of these compounds, 17a
and 17g bind to the Ad37 fiber knob in a fashion similar to that
of sialic acid, whereas the benzamide 17e may interact with
the fiber knob in a different way.
Cocrystals of the Ad37 fiber knob in complex with each
ligand were grown by the hanging drop method.23,45 Data were
collected using synchrotron radiation and the crystals diffracted
to 1.8 Å resolution. The structures were solved by molecular
replacement using the previous structure of the complex of
Ad37-sialyllactose45 (Table 2). As control, crystals with
sialyllactose were reproduced. The ligand 17e was not visible

3672 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Johansson et al.
Figure 5. Dose-dependent inhibition of EKC-causing adenoviruses
by 17-valent sialic acid and the multivalent N-acyl modified sialic acids
22a-c (14- to 17-valent). The experiment was repeated three times
with reproducible results. The data represent the average of dupli-
cates for each dilution series (for experimental details, see Materials
and Methods).
Table 2. Data Collection and Refinement Statistics for Structures of
Ad37 Fiber Protein in Complex with the Modified Sialic Acids 17a and
17g
17a
17g
space group
unit-cell dimensions
a (Å)
b (Å)
c (Å)
P21
P21
61.27
69.99
75.00
94.42
30-1.8 (1.80-1.90)
4.5 (46.7)
61.14
69.93
74.45
94.35
30-1.80 (1.80-1.86)
5.3 (60.8)
ꢀ (deg)
resolution rangea (Å)
a
Rmerge
I/σ(I)a
18.39 (2.88)
99.8 (99.8)
13.85 (1.84)
99.3 (98.2)
completeness (%)a
no. of unique reflections 58545
Refinement
30-1.80 (1.80-1.85) 30-1.80 (1.80-1.85)
57656
resolution rangea (Å)
total no. reflections/free 55584/2953
54725/2931
0.183 (0.304)
0.223 (0.387)
19.2
a
Rwork
0.183 (0.261)
0.226 (0.284)
19.4
a
Rfree
average B-factor (Å2)
no. water molecules
721
674
Root-Mean-Squared Deviation
bond lengths (Å)
bond angles (deg)
0.010
1.225
0.012
1.339
Ramachandran Plotb
favored (%)
allowed (%)
96.5
99.3
96.5
99.3
a Values in parentheses are for highest-resolution shell. b Molprobity
http://molprobity.biochem.duke.edu/.
in the structure of the known binding site of sialic acid or
anywhere else in the fiber protein, indicating no binding. The
potential of 17e as an inhibitor of Ad37 attachment (Table 1)
most likely results from unspecific binding to the fiber protein.
In contrast, 17a and 17g were both clearly observed within the
sialic acid binding site and with well-defined electron density
(Figure 6). Overlays with the Ad37-sialyllactose complex
showed that both 17a and 17g were positioned in the same
orientation as sialic acid. All of the important interactions were
fulfilled, and as predicted, the propanoyl and methoxycarbonyl
groups were directed into the hydrophobic part of the binding
site. From the crystallographic structures, it was also evident
that 17a and 17g were slightly shifted in comparison to sialic
acid while the amino acid residues in the binding pocket were
unaffected by the variation in structure between the different
ligands. In addition, the crystallographic structures suggest that
Figure 4. Dose-dependent inhibition of Ad37 infection of HCE cells
by N-acyl modified sialic acids 17a, 17b, 17c, 17e, and 17g (black
lines), compared to sialic acid (SA, dashed lines). The derivatives 17a
and 17g were as effective at inhibition as sialic acid, while 17b, 17c,
and 17e were inactive. The data represent the average of duplicates
within one experiment, and the experiment was repeated twice with
reproducible results (for experimental details, see Materials and
Methods).

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3673
Figure 6. Crystallographic structures of the N-acyl modified sialic acids 17a (A, C, E) and 17g (B, D, F), in complex with the Ad37 fiber knob
protein were solved to 1.8 Å resolution by molecular replacement. Both ligands were seen with well-defined electron density (A, B) within the
binding site of sialic acid (C, D). Superposition with the sialyllactose-Ad37 structure (E, F) shows that the N-acyl modified sialic acids bind to the
binding site in a similar manner as sialic acid, with the extra atoms on the N-acyl group directed down to the hydrophobic part of the binding site:
(A, B) black electron density map, unbiased difference density (2.50σ); blue, final refined map (0.95σ); (C-F) white and yellow residues, chain C
and chain B, respectively, of the homotrimer of the Ad37 fiber protein in presence of the N-acyl modified sialic acids; orange sticks, N-acyl
modified sialic acids; (E, F) green, sialic acid and surrounding residues from the sialyllactose-Ad37 structure.
larger groups such as a butanoyl moiety could readily be
accommodated in the hydrophobic pocket. The experimental
data on virus attachment and infection for the butanoyl derivative
17b do, however, indicate that this pocket is not fully accessible.
pocket. A set of 10 N-acyl modified sialic acids was designed,
synthesized, biologically evaluated, and compared to sialic acid.
The N-acyl modified sialic acids were synthesized as their R-O-
methyl glycosides (17a-h and 18). They were synthesized in
seven steps from the methyl glycoside of sialic acid in 12-65%
total yield. Biological evaluation of these compounds in a cell
binding assay and a cell infection assay revealed that only sialic
acids with small N-acyl groups were effective as inhibitors of
attachment of Ad37 to and infection of HCE cells. The
propanamide 17a and the methyl carbamate 17g were almost
as efficient as sialic acid in this respect. In order to investigate
the capabilities of N-acyl substituted sialic acids further, the
multivalent HSA conjugates 22a-c were synthesized and
Conclusions
In previous studies, we have shown that multivalent sialic
acid conjugates are effective inhibitors of EKC-causing
adenoviruses.18,19 In this study, structure-based design was used
to increase our understanding of the sialic acid-Ad37 interac-
tion. Using the crystallographic structure of sialyllactose in
complex with the fiber knob protein of Ad37, we found that
the acetamide of sialic acid is directed into a deep hydrophobic

3674 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Johansson et al.
docking region was enclosed by a sphere of 10 Å centered on the
center of gravity of the sialic acid residue of the experimental ligand.
Each ligand was docked into the binding site in 20 GA runs,
resulting in 20 binding poses per ligand. To test the reliability of
the poses generated by GOLD, the ability to reproduce the
experimental binding mode of the sialic acid residue of the
experimental ligand was tested. The experimental ligand was taken
out of the X-ray structure, and a new conformation of sialic acid
was docked back into the binding site using the settings above. An
average rmsd value of 1.5 Å between the 20 calculated poses and
the experimental pose indicates a reliable prediction of the binding
mode.25
evaluated. An efficient synthetic strategy to obtain the multi-
valent HSA conjugates was developed, and they were obtained
in seven steps from the 2-thiophenyl donor of sialic acid in
10-32% yield. The degree of multivalency was determined to
be 14-17 saccharides incorporated per HSA. The multivalent
butanamide 22b and the isobutanamide 22c showed no activity.
The multivalent propanamide 22a, which was as potent as sialic
acid in the monovalent case (17a), was much less effective than
multivalent (17-valent) sialic acid. From these results, it is
obvious that the hydrophobic cavity of the binding site is not
as easy to occupy as initially expected. Crystallography
confirmed that 17a and 17g bind in the same orientation as sialic
acid. It seems likely that the sialic acid interacting site of Ad37
is very rigid and not prone to any conformational changes to fit
the larger substituents.
Future efforts to obtain a high-affinity structure as an inhibitor
of EKC-causing adenoviruses will require more extensive
structure-activity investigations. Because of the complicated
synthesis of functionalizing sialic acid, an alternative strategy
for development of successful inhibitors might be replacement
of the entire sialic acid core with a non-carbohydrate sialyl
mimetic presenting the critical attributes for viral recognition.
A non-carbohydrate sialyl mimetic would be more advantageous
than functionalized sialic acids in terms of less complicated
synthesis and a greater possibility of varying the functional
groups. The results presented here would be valuable for future
efforts to develop an antiviral drug for treatment of EKC,
regardless of whether it is via the sialic acid strategy or by using
sialyl mimetics.
Evaluation of the Docking. All docking poses for all ligands
were evaluated visually. The docked poses were overlaid on the
experimental ligand in the binding site using the software SYBYL.42
Compounds with a majority of the calculated poses were clustered
close to the experimental binding mode of sialic acid and in a
reliable conformation allowing the critical hydrogen bonds, and the
salt bridge was defined as good binders with potential to increase
the binding affinity. Compounds with the docked poses spread out
in the binding site and not clustered at all were assumed to have
unspecific binding modes and were predicted to be poor binders.
Biological Evaluation. The binding assay33,34 and the infectivity
assay15,35 were carried out essentially according to the indicated
literature procedures. See Supporting Information for details.
Protein Expression and Purification. Ad37 fiber knob DNA
corresponding to residues 177-365 of the full-length fiber was
cloned into the pPROEX HTb plasmid. The protein was expressed
in the BL21 Star (DE3) bacterial strain in LB medium at 37 °C for
4 h after induction with 1 mM IPTG. The pellet was resuspended
in buffer (150 mM NaCl, 30 mM Tris, pH 7.5) and incubated with
lysozyme (1 mg/mL) on ice for 30 min and sonicated 12 times for
10 s each time. After centrifugation at 45000g for 30 min at 4 °C,
the cleared lysate was loaded onto 10 mL of Ni Sepharose resin in
a column in the presence of 10 mM imidazole, washed first with
50 mL of 150 mM NaCl, 30 mM Tris, pH 7.5, 10 mM imidazole
and then with 40 mL of 150 mM NaCl, 30 mM Tris, pH 7.5, 20
mM imidazole and then eluted in several 5 mL fractions with buffer
(150 mM NaCl, 30 mM Tris, pH 7.5, 0.5 M imidazole). The His
tag was cleaved off by an overnight incubation with a 1/100 dilution
of His-tagged TEV protease at 10 °C. Imidazole was eliminated
by dialysis against the buffer 150 mM NaCl, 30 mM Tris, pH 7.5,
20 mM imidazole. Cleaved protein was then incubated with 5 mL
of Ni Sepharose for 1 h at 4 °C and recovered in the flow-through
and washing fractions ((5-6) × 5 mL). These fractions were
pooled, concentrated to 7-8 mg/mL, and purified over a Superdex
200 column, and the peak collected was concentrated to 15 mg/
mL for crystallization.
Crystallization and Structure Determination. Cocrystals of
Ad37 fiber knobs (at 13 mg/mL) with each ligand were grown by
the hanging drop method23,45 using a reservoir solution of 25-27%
polyethylene glycol 8000, 50 mM zinc acetate, and 100 mM HEPES
(pH 7.15, 7.3, or 7.5). Crystals were fast-frozen in nitrogen prior
to data collection without cryoprotection. Data were collected on
beamline ID14-2 and ID14-4 at the European Synchroton Radiation
Facility (Grenoble, France). As a control, crystals with 20 mM
sialyllactose were reproduced.23 The best diffraction was obtained
with crystals grown under conditions of 25% PEG 8000, 50 mM
zinc acetate, and 100 mM HEPES, pH 7.5. Data were collected on
ID14.4 beamline at 1.8 Å resolution. For 17a, 17e, and 17g the
best diffracting crystals were grown in the presence of 20 mM
ligand in the same crystallization condition as for sialyllactose. Data
were collected on ID14-eh2 beamline at 1.8 Å resolution. The
structure was solved by molecular replacement using the Ad37-sialic
acid structure. Data collection for 17e at 1.8 Å resolution and
structure resolution were performed in the same way, but the ligand
was not visible in the structure at the known site for sialic acid.
The best diffracting crystals were grown in the presence of 10 mM
ligand in crystallization solution 27% PEG 8000, 50 mM zinc
acetate, and 100 mM HEPES, pH 7.5. Data collection at 1.8 Å
resolution and structure resolution were performed in the same way
Materials and Methods
Computational Methodology. Pretreatment of Protein. The
three-dimensional coordinates of the fiber knob protein of Ad37
were obtained from the Protein Data Bank (code 1UXA). All
hydrogens were first removed from the PDB file and then added
again using the software Reduce,36 after which the orientations of
the side chains of ASN, GLN, and HIS were optimized. Protonation
states were set according to the program, based on the likelihood
that the residues would be charged at physiological pH, together
with their opportunities to form hydrogen bonds with the surround-
ing environment. All water coordinates and metal complexes were
removed from the protein file. Important residues, especially in the
binding site, were visually inspected and corrected according to
charge and according to atom and bond type.
Pretreatment of Ligand. The ligands that were selected for the
docking studies were drawn using CS Chemdraw Ultra37 software
and represented as SMILES.38 Their 3D structures were generated
with CORINA.39 All ligands were inspected and refined to correct
stereochemistry and atom and bond type according to GOLD
(genetic optimization for ligand docking)40 specifications using
MOE (Molecular Operating Environment)41 and SYBYL42 soft-
ware. In order to obtain low-energy conformations for all ligands,
they were energy-minimized using the MMFF94 force field
implemented in MOE.
Energy Minimization of the Binding Site. The protein-ligand
complex of one of the three binding sites presented by the Ad37
fiber protein was energy-minimized using the Amber94 force field
on the 3′-sialyllactose ligand and its surroundings within a distance
of 3 × 4.5 Å (corresponding residues were also included). The
binding site and the ligand were then visually examined to confirm
that the conformations were plausible. Before docking, the ligand
was removed from the binding site.
Docking in GOLD. In GOLD, the ligands are placed in the
protein binding site by a mechanism based on fitting points. A
genetic algorithm (GA) is used to explore possible binding modes,
and a scoring function is used to rank the calculated binding
modes.25,43,44 In this study, the default settings of the GA
parameters were used with the Goldscore scoring function. The

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3675
as before, but the ligand was not visible in the structure at the known
site for sialic acid.
3.26 (s, 3H, OCH3), 3.75 (s, 3H, COOCH3), 4.02-4.10 (m, 3H,
H5, H6, H9a), 4.25 (dd, JH8 ) 12.40 Hz, JH9a ) 2.75 Hz, 1H, H9b),
4.76-4.88 (m, 1H, H4), 5.26 (dd, J ) 1.94 Hz, J ) 8.31 Hz, 1H,
H7), 5.34-5.39 (m, 2H, NH, H8).
Methyl (O-Methyl-5-N-isobutanoyl-4,7,8,9-tetra-O-acetyl-3,5-
dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16c).
Yield, 65 mg (26%). 1H NMR (CDCl3): δ 1.02-1.1 (m, 6H,
CH(CH3)2), 1.95 (bt, 1H, J ) 12.7 Hz, H3ax), 2.00, 2.03, 2.13, 2.15
(4s, 3H each, 4Ac), 2.21-2.32 (m, 1H, CH(CH3)2), 2.56 (dd, 1H,
JH3ax ) 12.8 Hz, JH4 ) 4.6 Hz, H3eq), 3.32 (s, 3H, OCH3), 3.81 (s,
Chemical Synthesis. Methyl (O-Methyl-5-(N-tert-butoxycar-
bonylacetamido)-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-r-
D-galacto-2-nonulopyranosyl)onate (12). A solution of 1131 (1.81
g, 3.6 mmol, 4:1 R/ꢀ), Boc2O (1.8 mg, 8.2 mmol), and 4-dimethy-
laminopyridine (DMAP) (267 mg, 2.2 mmol) in dry THF (60 mL)
was refluxed for 2 h under nitrogen. The reaction was followed by
TLC (toluene/ethanol 10:1). The reaction mixture was cooled to
room temperature, diluted with CH2Cl2, washed with 0.5 M aqueous
HCl, water, saturated NaHCO3 (aq), and water, dried, and concen-
trated. Column chromatography (toluene/ethanol 10:1) gave 2.1 g
of product 12 as a yellow oil, still as a 4:1 R/ꢀ mixture and with
additional 10% of impurities. Since the impurities as well as the ꢀ
isomer were removed in later steps, this product was used despite
having impurities. Thus, no yield and no NMR data are presented.
Methyl (O-Methyl-5-(N-tert-butoxycarbonyl)-3,5-dideoxy-D-
glycero-r-D-galacto-2-nonulopyranosyl)onate (13). Partly purified
compound 12 (2.1 g, 3.5 mmol) was stirred at room temperature
in methanolic sodium methoxide (0.03 M, 180 mL) for 3 h. The
solution was neutralized with 10% aqueous HCl and concentrated.
Column chromatography (CH2Cl2/MeOH 20:1) afforded 927 mg
of pure 1332 (65%). 1H NMR (CD3OD): δ 1.45 (s, 9H, -C(CH3)3),
1.70 (bt, 1H, J ) 12.6 Hz, H3ax), 2.63 (m, 1H, JH3ax ) 12.9 Hz,
JH4 ) 4.5 Hz, H3eq), 3.33 (s, 3H, OCH3), 3.43-3.66, 3.83-3.88
(2m, 7H, H4, H5, H6, H7, H8, H9a, H9b), 3.83 (s, 3H, COOCH3).
Methyl (O-Methyl-5-(N-tert-butoxycarbonyl)-4,7,8,9-tetra-O-
acetyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)-
onate (14). To a 0 °C solution of 13 (1.21 g, 3.1 mmol) in pyridine
(100 mL), Ac2O (60 mL) was added in small portions. The reaction
mixture was stirred at room temperature overnight, concentrated,
and coevaporated with toluene. Column chromatography (toluene/
3H, COOCH3), 4.03-4.19 (m, 3H, H5, H6, H9a), 4.31 (dd, JH8
)
12.44 Hz, JH9a ) 2.72 Hz, 1H, H9b), 4.85-4.95 (m, 1H, H4),
5.25-5.35 (m, 2H, NH, H7), 5.37-5.44 (m, 1H, H8).
Methyl (O-Methyl-5-N-pentanoyl-4,7,8,9-tetra-O-acetyl-3,5-
dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16d).
1
Yield, 137 mg (64%). H NMR (CDCl3): δ 0.85 (t, JCH2 ) 7.32,
3H, CH2CH3), 1.23-1.37 (m, 2H, -CH2CH3), 1.45-1.59 (m, 2H,
CH2CH2CH2-), 1.93 (m, 1H, J ) 12.6 Hz, H3ax), 2.01, 2.04, 2.14,
2.15 (4s, 3H each, 4Ac), 2.04-2.14 (m, 2H, CH2CH2CH2-), 2.57
(dd, 1H, JH3ax ) 12.9 Hz, JH4 ) 4.6 Hz, H3eq), 3.32 (s, 3H, OCH3),
3.81 (s, 3H, COOCH3), 3.99-4.18 (m, 3H, H5, H6, H9a), 4.31 (dd,
JH8 ) 15.14 Hz, JH9a ) 4.31 Hz, 1H, H9b), 4.84-4.94 (m, 1H,
H4), 5.24-5.33 (m, 2H, NH, H7), 5.38-5.45 (m, 1H, H8).
Methyl (O-Methyl-5-N-benzoyl-4,7,8,9-tetra-O-acetyl-3,5-
dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16e).
1
Yield, 89 mg (39%). H NMR (CDCl3): δ 1.95, 1.98, 2.13, 2.20
(4s, 3H each, 4Ac), 1.97-2.03 (m, 1H, H3ax), 2.63 (dd, 1H, JH3ax
) 12.8 Hz, JH4 ) 4.6 Hz, H3eq), 3.34 (s, 3H, OCH3), 3.79 (s, 3H,
COOCH3), 4.08 (dd, JH8 ) 12.41 Hz, JH9a ) 5.82 Hz, 1H, H9b),
4.23-4.33 (m, 3H, H5, H6, H9a), 5.00-5.09 (m, 1H, H4),
5.30-5.35 (m, 1H, H7), 5.42-5.47 (m, 1H, H8), 5.93 (d, JH5
8.87, 1H, NH), 7.36-7.50 (m, 5H, Ar).
)
1
ethanol 8:1) afforded 1.56 g of 14 (92%). H NMR (CDCl3): δ
Methyl (O-Methyl-5-N-trifluoroacetamido-4,7,8,9-tetra-O-
acetyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)-
onate (16f). To a solution of 14 (220 mg, 0.39 mmol) in CH2Cl2
(6 mL), 90% aqueous TFA (2 mL) was added. The reaction mixture
was stirred at room temperature for 1 h and monitored with TLC
(toluene/acetone 1:1). The reaction mixture was concentrated,
coevaporated with toluene, and dried extensively in vacuo to give
the ammonium salt 15. To a solution of 15 in dry CH2Cl2 (4 mL)
at 0 °C, trifluoroacetic anhydride (108 µL, 0.78 mmol, 2 equiv)
and DMAP (95 mg, 0.78 mmol, 2 equiv) were added in the given
order. The reaction mixture was stirred at room temperature under
nitrogen and monitored by LCMS. After being stirred for 30 min,
the reaction mixture was diluted with CH2Cl2, washed with water
and saturated NaHCO3 (aq) and water, dried, and concentrated.
Repeated column chromatography (toluene/ethanol 10:1) gave 114
1.39 (s, 9H, (CH3)3), 1.84-1.95 (m, 1H, H3ax), 2.03, 2.04, 2.13,
2.14 (4s, 3H each, 4Ac), 2.58 (dd, 1H, JH3ax ) 12.8 Hz, JH4 ) 4.5
Hz, H3eq), 3.31 (s, 3H, -OCH3), 3.74-3.80 (m, 1H, H5), 3.80 (s,
3H, -COOCH3), 4.0-4.14 (m, 2H, H6, H9a), 4.18 (d, JH5 ) 10.6
Hz, 1H, NH), 4.3 (d, JH8 ) 11.53 Hz, 1H, H9b), 4.7-4.8 (m, 1H,
H4), 5.4-5.5 (m, 2H, H7, H8).
General Procedure for Preparation of Methyl (O-Methyl-5-
N-acyl-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-r-D-galacto-
2-nonulopyranosyl)onates (16a-e). To a solution of 14 (0.38-0.40
mmol) in CH2Cl2 (6 mL), 90% aqueous TFA (2 mL) was added.
The reaction mixture was stirred at room temperature for 1-2 h
and monitored with TLC (toluene/acetone 1:1). The reaction mixture
was concentrated, coevaporated with toluene, and dried extensively
in vacuo to give the ammonium salt 15. To a solution of 15 in dry
CH2Cl2 (2 mL), RCOOH (1.2 equiv), HATU (1.2 equiv according
to the acid), and DIPEA (3 eq, 0.2 mL) were added in the given
order. The reaction mixture was stirred at room temperature under
nitrogen and monitored by LCMS. After being stirred for 3-5 h,
the reaction mixture was diluted with CH2Cl2, washed with water,
saturated NaHCO3, and water, dried, and concentrated. Column
chromatography (toluene/ethanol 10:1) followed by preparative
HPLC yielded 16a-e.
1
mg 16f (56%). H NMR (CDCl3): δ 1.94 (bt, 1H, J ) 12.6 Hz,
H3ax), 2.02, 2.04, 2.14, 2.17 (4s, 3H each, 4Ac), 2.62 (m, 1H, JH3ax
) 12.9 Hz, JH4 ) 4.7 Hz H3eq), 3.33 (s, 3H, OCH3), 3.82 (s, 3H,
COOCH3), 3.93-4.05 (m, 1H, H5), 4.13 (dd, JH8 ) 18.01 Hz, JH9a
) 5.50 Hz, 1H, H9b), 4.26-4.37 (m, 2H, H6, H9a), 4.94-5.05 (m,
1H, H4), 5.30 (dd, 1H, J ) 2.08, J ) 10.18, H7), 5.38-5.45 (m,
1H, H8), 6.52 (d, JH5 ) 9.98, 1H, NH). 19F NMR (CDCl3): -76.7.
Methyl (O-Methy-5-N-methoxycarbonyl-4,7,8,9-tetra-O-acetyl-
3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16g).
To a solution of 14 (201 mg, 0.36 mmol) in CH2Cl2 (6 mL), 90%
aqueous TFA (2 mL) was added. The reaction mixture was stirred
at room temperature for 1 h and monitored with TLC (toluene/
acetone 1:1). The reaction mixture was concentrated, coevaporated
with toluene, and dried extensively in vacuo to give the ammonium
salt 15. To a solution of 15 (0.26 mmol) in dry CH2Cl2 (3 mL),
DIPEA was added (132 µL, 0.78 mmol, 3 equiv). The mixture was
cooled to 0 °C before adding methyl chloroformate (60 µL, 0.78
mmol, 3 equiv) dropwise. The reaction mixture was stirred at room
temperature under nitrogen, which was followed by LCMS. After
being stirred for 1.5 h, the reaction mixture was diluted with CH2Cl2,
washed with water, 1 M aqueous HCl, and water, dried, and
concentrated. Column chromatography (toluene/acetone 4;1) af-
Methyl (O-Methyl-5-N-propionyl-4,7,8,9-tetra-O-acetyl-3,5-
dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16a).
Yield, 114 mg (56%). 1H NMR (CDCl3): δ 1.07 (t, 3H, CH2CH3),
1.94 (bt, 1H, J ) 12.5 Hz, H3ax), 2.00, 2.03, 2.13, 2.15 (4s, 3H
each, 4Ac), 2.05-2.12 (m, 2H, CH2CH3), 2.56 (dd, 1H, JH3ax
)
12.9 Hz, JH4 ) 4.7 Hz, H3eq), 3.32 (s, 3H, OCH3), 3.81 (s, 3H,
COOCH3), 4.05-4.16 (m, 3H, H5, H6, H9a), 4.29 (dd, JH8 ) 15.13
Hz, JH9a ) 2.7 Hz, 1H, H9b), 4.80-4.93 (m, 1H, H4), 5.10 (d, JH5
) 9.3 Hz, 1H, NH), 5.3 (dd, J ) 1.66 Hz, J ) 10.13 Hz, 1H, H7),
5.38-5.46 (m, 1H, H8).
Methyl (O-Methyl-5-N-butanoyl-4,7,8,9-tetra-O-acetyl-3,5-
dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16b).
1
Yield, 186 mg (92%). H NMR (CDCl3): δ 0.85, (t, J ) 7.37 Hz,
3H, CH3), 1.47-1.57 (m, 2H, CH2CH3), 1.87 (bt, 1H, J ) 12.6,
H3ax), 1.95, 1.98, 2.08, 2.09 (4s, 3H each, 4Ac), 1.92-2.08 (m,
2H, CH2CH2-), 2.52 (dd, 1H, JH3ax ) 12.7 Hz, JH4 ) 4.6 Hz, H3eq),
1
forded 90 mg of 16g (66%). H NMR (CDCl3): δ 1.89 (bt, 1H, J
) 12.6 Hz, H3ax), 2.03, 2.04, 2.15, 2.15 (4s, 3H each, 4Ac), 2.61

3676 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Johansson et al.
(dd, 1H, JH3ax ) 12.8 Hz, JH4 ) 4.6 Hz H3eq), 3.32 (s, 3H, OCH3),
3.62 (s, 3H, NHCOOCH3), 3.67-3.78 (m, 1H, H5), 3.81 (s, 3H,
COOCH3), 4.05-4.15 (m, 2H, H6, H9a), 4.30-4.35 (m, 1H, H9b),
4.42 (d, JH5 ) 9.76, 1H, NH), 4.79-4.90 (m, 1H, H4), 5.36-5.47
(m, 2H, H7, H8).
3H, CH2CH3), 1.33-1.45 (m, 2H, CH2CH3), 1.57-1.68 (m, 2H,
CH2CH2CH3), 1.68-1.78 (m, 1H, H3ax), 2.21-2.31 (m, 2H,
COCH2-), 2.71 (dd, 1H, JH3ax ) 12.8 Hz, JH4 ) 4.1 Hz, H3eq),
3.39 (s, 3H, OCH3), 3.53-3.91 (m, 7H, H4, H5, H6, H7, H8, H9a,
H9b). HRMS (FAB) calcd for C15H27NO9: 410.1404 [M + 2Na -
2H]. Found: 410.1399.
Methyl (O-Methyl-5-N-(3-methylureido)-4,7,8,9-tetra-O-acetyl-
3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosyl)onate (16h).
To a solution of 14 (100 mg, 0.18 mmol) in CH2Cl2 (3 mL), 90%
aqueous TFA (1 mL) was added. The reaction mixture was stirred
at room temperature for 1 h and monitored with TLC (toluene/
acetone 1:1). The reaction mixture was concentrated, coevaporated
with toluene, and dried extensively in vacuo to give the ammonium
salt 15. To a solution of 15 in dry CH2Cl2 (2 mL), DIPEA was
added (125 µL, 0.72 mmol, 4 equiv). The mixture was cooled to 0
°C before methyl isocyanate was added dropwise (106 µL, 1.8
mmol, 10 equiv). The reaction mixture was stirred at room
temperature under nitrogen, which was followed by LCMS. After
being stirred for 4 h, the reaction mixture was diluted with CH2Cl2,
washed with water, 1 M aqueous HCl, and water, dried, and
concentrated. Column chromatography (toluene/ethanol 10:1) af-
O-Methyl-5-N-benzoyl-3,5-dideoxy-D-glycero-r-D-galacto-2-
nonulopyranosylonic Acid (17e). Yield, 38 mg (78%). [R]2D0 6.3
(c 0.83, MeOH). 1H NMR (CD3OD): δ 1.78 (bt, 1H, J ) 12.2 Hz,
H3ax), 2.75 (dd, 1H, JH3ax ) 12.9 Hz, JH4 ) 4.4 Hz H3eq), 3.40 (s,
3H, -OCH3), 3.56-4.07 (m, 7H, H4, H5, H6, H7, H8, H9a, H9b),
7.39-7.58 (m, 3H, Ar), 7.87-7.89 (m, 2H, Ar). HRMS (FAB)
calcd for C17H23NO9: 430.1091 [M + 2Na - 2H]. Found: 430.1095.
O-Methyl-5-N-methoxycarbonyl-3,5-dideoxy-D-glycero-r-D-
galacto-2-nonulopyranosylonic Acid (17g). Yield, 47 mg (100%).
[R]2D0 -2.9 (c 0.58, MeOH). 1H NMR (CD3OD): δ 1.67 (bt, 1H, J
) 12.2 Hz, H3ax), 2.69 (dd, 1H, JH ) 4.5 Hz, JH3ax ) 12.69 Hz,
H3eq), 3.36 (s, 3H, OCH3), 3.46-3.93 (m, 10H, H4, H5, H6, H7,
H8, H9a, H9b, NHCO2CH3). HRMS (FAB) calcd for C12H21NO10:
384.0883 [M + 2Na - 2H]. Found: 384.0882.
1
forded 62 mg of 16h (66%). H NMR (CDCl3): δ 1.95 (bt, 1H, J
O-Methyl-5-N-(3-methylureido)-3,5-dideoxy-D-glycero-r-D-
galacto-2-nonulopyranosylonic Acid (17h). Yield, 65 mg (100%).
) 12.6 Hz, H3ax), 2.027, 2.034, 2.14, 2.16 (4s, 3H each, 4Ac),
2.56 (dd, 1H, JH3ax ) 12.9 Hz, JH4 ) 4.6 Hz H3eq), 2.70 (s, 3H,
NHCH3), 3.33 (s, 3H, OCH3), 3.81 (s, 3H, COOCH3), 3.88-3.98
(m, 1H, H5), 4.03-4.22 (m, 3H, H6, H9a, NH), 4.36 (dd, 1H, JH9a
) 12.5 Hz, JH8 ) 2.4 Hz H9b), 4.80-4.91 (m, 1H, H4), 5.35-5.47
(m, 2H, H7, H8).
1
[R]2D0 10.2 (c 0.84, MeOH). H NMR (CD3OD): δ 1.67 (bt, 1H, J
) 12.1 Hz, H3ax), 2.60-2.81 (m, 4H, H3eq, NHCH3), 3.36 (s, 3H,
OCH3), 3.44-3.97 (m, 7H, H4, H5, H6, H7, H8, H9a, H9b). HRMS
(FAB) calcd for C12H22N2O9: 361.1223 [M + Na - H]. Found:
361.1222.
General Procedure for Preparation of O-Methyl-5-N-acyl-
3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosylonic Acids
(17a-h). The N-acyl modified methylsialosides 16a-h (0.10-0.30
mmol) were stirred in methanolic sodium methoxide (0.03 M, 7-15
mL) for 2 h. Complete O-deacetylation was confirmed with LCMS.
The solutions were neutralized with Amberlite IR-120, filtered, and
concentrated. The residues were diluted in methanol (4-8 mL),
and LiOH (2 M (aq), 6 equiv) was added and stirred at room
temperature overnight. The hydrolysis of the methyl esters was
monitored by LCMS. The solutions were carefully neutralized with
Amberlite IR-120, filtered, and concentrated. Purification by HPLC
and lyophilization gave the products 17a-h. Compounds 17a-e
and 17g,h were estimated to be >95% pure according to hydrophilic
interaction liquid chromatography (HILIC) and detection with MS
(see Supporting Information for details). No UV-active impurities
were observed with analytical reversed-phase HPLC and UV
detection at 214 nm.
O-Methyl-5-N-propaonyl-3,5-dideoxy-D-glycero-r-D-galacto-
2-nonulopyranosylonic Acid (17a). Yield, 73 mg (100%). [R]D20
10 (c 0.67, MeOH). 1H NMR (CD3OD): δ 1.14 (t, JCH2 ) 7.60 Hz,
3H, CH2CH3), 1.59-1.69 (m, 1H, H3ax), 2.27 (q, JCH3 ) 7.60, 2H,
CH2CH3), 2.73 (dd, 1H, JH3ax ) 12.8 Hz, JH4 ) 4.1 Hz H3eq), 3.36
(s, 3H, OCH3), 3.47-3.91 (m, 7H, H4, H5, H6, H7, H8, H9a, H9b).
HRMS (FAB) calcd for C13H23NO9: 382.1091 [M + 2Na - 2H].
Found: 382.1111.
O-Methyl-5-N-tert-butoxycarbonyl-3,5-dideoxy-D-glycero-r-
D-galacto-2-nonulopyranosylonic Acid (18). To a solution of 13
(28 mg, 0.071 mmol) in MeOH (2 mL), LiOH (0.2 mL, 2 M aq, 6
equiv) was added. The solution was stirred at room temperature
overnight and then carefully neutralized with Amberlite IR-120,
filtered, and concentrated at reduced pressure. Lyophilization
afforded 27 mg of 18 (quant). Compound 18 was estimated to be
>95% pure according to hydrophilic interaction liquid chromatog-
raphy (HILIC) and detection with MS (see Supporting Information
for details). No UV-active impurities were observed with analytical
reversed-phase HPLC and UV detection at 214 nm. [R]2D0 14.4 (c
0.32, MeOH). 1H NMR (CD3OD): δ 1.45 (s, 9H, C(CH3)3),
1.53-1.60 (m, 1H, H3ax), 2.75-2.79 (m, 1H, H3eq), 3.35 (s, 3H,
OCH3), 3.36-3.43 (m, 1H, H5), 3.53-3.65, 3.83-3.92 (2m, 6H,
H4, H6, H7, H8, H9a, H9b), 6.75 (d, JH5 ) 6.74 Hz, 1H, NH).
HRMS (FAB) calcd for C15H27NO10: 426.1353 [M + 2Na - 2H].
Found: 426.1353.
Methyl (5-Fmoc-aminopentyl[5-(N-tert-butoxycarbonylamido)-
4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonu-
lopyranosyl])onate (19). Compound 6 (50 mg, 0.078 mmol), Fmoc-
aminopentanol (127 mg, 0.39 mmol), and powdered molecular
sieves (3 Å, 0.17 g) were stirred in a mixture of CH3CN and CH2Cl2
(3:2, 3 mL) for 1 h under nitrogen. Silver trifluoromethanesulfonate
(40 mg, 0.16 mmol) was added, and the reaction mixture was cooled
to -72 °C. Iodine monobromide in CH2Cl2 (0.5 M, 240 µL, 0.12
mmol) was added, and the mixture was stirred for 4 h. DIPEA
(127 µL, 0.51 mmol) was added, and stirring was continued for 30
min. The mixture was allowed to come to room temperature before
being filtered and concentrated in vacuo. Repeated column chro-
matography of the residue (silica gel, 30-70 µm, toluene/acetone,
4:1, petroleum ether/ethyl acetate, 1.5:1) afforded 46 mg of the
title product 19 as a white powder (68%). [R]2D0 -7.5 (c 1.0, MeOH).
1H NMR (CDCl3): δ 1.39 (s, 9H, C(CH3)3, 2H, OCH2-
CH2CH2CH2CH2NH-), 1.51-1.60 (m, 4H, OCH2CH2CH2CH2-
CH2NH-), 1.91 (bt, 1H, J ) 12.4 Hz, H3ax), 2.02 (s, 6H, 2Ac),
2.13 (s, 3H, Ac), 2.14 (s, 3H, Ac), 2.58 (dd, 1H, JH3ax ) 12.4 Hz,
JH4 ) 4.4 Hz, H3eq), 3.15-3.27 (m, 2H, OCH2CH2-, 1H,
CH2CH2NH-), 3.72-3.81 (m, 2H, H5, CH2CH2NH-), 3.78 (s,
O-Methyl-5-N-butanoyl-3,5-dideoxy-D-glycero-r-D-galacto-2-
nonulopyranosylonic Acid (17b). Yield, 92 mg (97%). [R]2D0 8.9
1
(c 0.27, MeOH). H NMR (CD3OD): δ 0.96 (t, JCH2 ) 7.40 Hz,
3H, CH2CH3), 1.61-1.70 (m, 3H, CH2CH3, H3ax), 2.19-2.27 (m,
2H, CH2CH2CH3), 2.74 (dd, 1H, JH3ax ) 12.8 Hz, JH4 ) 4.1 Hz,
H3eq), 3.36 (s, 3H, OCH3), 3.49-3.91 (m, 7H, H4, H5, H6, H7,
H8, H9a, H9b). HRMS (FAB) calcd for C14H25NO9: 396.1247 [M
+ 2Na - 2H]. Found: 396.1236.
O-Methyl-5-N-isobutanoyl-3,5-dideoxy-D-glycero-r-D-galacto-
2-nonulopyranosylonic Acid (17c). Yield, 28.4 mg (74%). [R]D20
1
-16.5 (c 0.26, H2O). H NMR (CD3OD): δ 1.10-1.17 (m, 6H,
CH(CH3)2) 1.65-1.77 (m, 1H, H3ax), 2.45-2.56 (m, 1H,
CH(CH3)2), 2.70 (m, 1H, JH3ax ) 12.7 Hz, JH4 ) 4.1 Hz, H3eq),
3.37 (s, 3H, OCH3), 3.46-3.91 (m, 7H, H4, H5, H6, H7, H8, H9a,
H9b). HRMS (FAB) calcd for C14H25NO9: 396.1247 [M + 2Na -
2H]. Found: 396.1251.
3H, CO2CH3), 3.99-4.10 (m, 2H, H6, H9a), 4.19 (d, 1H, JH5
)
10.4 Hz, CHNH), 4.21(t, 1H, JOCH2CH ) 6.8 Hz, OCH2CH-), 4.33
(dd, 1H, JH8 ) 2.4 Hz, JH9a ) 12.0 Hz, H9b), 4.39 (d, 2H, JOCH2CH
) 6.8 Hz, OCH2CH-), 4.72-4.81 (m, 1H, H4), 5.12(bs, 1H,
-CH2NH), 5.37-5.46 (m, 2H, H7, H8).
O-Methyl-5-N-pentanoyl-3,5-dideoxy-D-glycero-r-D-galacto-
2-nonulopyranosylonic Acid (17d). Yield, 68 mg (85%). [R]2D0 7.7
1
(c 0.85, MeOH). H NMR (CD3OD): δ 0.96 (t, JCH2 ) 7.32 Hz,

N-Acyl Modified Sialic Acids
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12 3677
General Procedure for Preparation of Methyl (5-Fmoc-
aminopentyl[5-N-acyl-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glyc-
ero-r-galacto-2-nonylopyranosyl])onates (20a-c). To a solution
of 19 (0.10-0.20 mmol) in CH2Cl2 (1.0-2.0 mL), 90% aqueous
TFA (0.3-0.6 mL) was added. The reaction mixture was stirred at
room temperature for 1 h, followed by TLC (toluene/acetone 2:1).
When the Boc removal was complete, the mixtures were coevapo-
rated with toluene and dried extensively in vacuo to give the
corresponding ammonium salt. To a solution of the salt in dry
CH2Cl2 (0.8-1.5 mL), RCOOH (1.2 equiv) and HATU (1.2 equiv
according to the acid) and DIPEA (3.5 equiv) were added in the
given order. The reaction was monitored by TLC (toluene/acetone
2:1) and LCMS, and after being stirred for 9-24 h at room
temperature, the reaction mixture was diluted with CH2Cl2 and
washed with 0.5 M aq HCl and water before being dried over
Na2SO4. Repeated column chromatography (toluene/acetone; 2:1)
of the residue afforded 20a-c.
overnight, carefully neutralized with Amberlite IR-120, and con-
centrated. The fully deprotected compounds 4a-c were used
without further purification. The compounds were dissolved in DMF
(3.5 mL), and didecyl squarate (0.16-0.18 mmol) and Et3N (1.3
equiv according to 20a-c) were added. After 11 h, the solutions
were coevaporated with toluene and CH2Cl2. The residues were
purified by column chromatography (methanol/ethyl acetate, 1:3
f 1:2) to give the title products 21a-c as white powders.
Compounds 21a-c were >95% pure according to analytical
reversed-phase HPLC with UV detection (see Supporting Informa-
tion for details).
5-(2-Decyloxy-3,4-dioxocylobut-1-enyl)aminopenthyl-[5-N-
propanoyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyrano-
sylonic Acid (21a). Yield, 31 mg (73%) containing 4% of the
corresponding ꢀ-anomer. [R]2D0 3.1 (c 1.0, MeOH). 1H NMR
(CD3OD): δ 0.90 (t, 3H, JCH2CH3 ) 6.8 Hz, CH2CH3), 1.14 (t, 3H,
JCH2CH3 ) 7.6 Hz, CH2CH3), 1.23-1.50 (m, 16H), 1.56-1.69 (m,
5H), 1.78-1.86 (m, 2H, C(sp2)OCH2CH2), 2.28 (q, 2H, JCH2CH3
)
Methyl (5-Fmoc-aminopentyl[5-N-propanoyl-4,7,8,9-tetra-O-
acetyl-3,5-dideoxy-D-glycero-r-galacto-2-nonylopyranosyl])-
onate (20a). Yield, 53.0 mg (64%). [R]2D0 -9.8 (c 1.0, MeOH). 1H
NMR (CDCl3): δ 1.08 (t, 3H, JCH2CH3 ) 7.6 Hz, CH2CH3),
1.33-1.44 (m, 2H, OCH2CH2CH2CH2CH2NH-), 1.48-1.59 (m,
4H, OCH2CH2CH2CH2CH2NH-), 1.96 (bt, 1H, J ) 12.8 Hz, H3ax),
2.00 (s, 3H, Ac), 2.02 (s, 3H, Ac), 2.07 (q, 2H, JCH2CH3 ) 7.6 Hz,
7.6 Hz, -CH2CH3), 2.83 (dd, 1H, JH4 ) 4.0 Hz, JH3ax ) 12.0 Hz,
H3eq), 3.42 (t, 1H, J ) 6.8 Hz), 3.45-3.53 (m, 2H), 3.55-3.65
(m, 3H), 3.66-3.74 (m, 2H), 3.77-3.90 (m, 3H), 4.65-4.74 (m,
2H). HRMS (FAB) calcd for C31H52N2O12: 689.3239 [M + 2Na -
2H]. Found: 689.3222.
5-(2-Decyloxy-3,4-dioxocylobut-1-enyl)aminopenthyl-[5-N-bu-
tanoyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyranosylon-
ic Acid (21b). Yield, 23 mg (61%) containing 2% of the
corresponding ꢀ-anomer. [R]2D0 3.2 (c 1.0, MeOH). 1H NMR
(CD3OD): δ 0.90 (t, 3H, JCH2CH3 ) 6.8 Hz, CH2CH3), 0.96 (t, 3H,
JCH2CH3 ) 7.6 Hz, -CH2CH3), 1.27-1.50 (m, 16H), 1.53-1.71
(m, 7H), 1.80-1.90 (m, 2H, C(sp2)OCH2CH2), 2.25 (t, 2H, JCH2CH2
) 6.8 Hz, CH2CH2), 2.83 (dd, 1H, JH3ax ) 12.0 Hz, JH4 ) 4.8 Hz,
H3eq), 3.42 (t, 1H, J ) 7.2 Hz), 3.46-3.53 (m, 2H), 3.55-3.65
(m, 3H), 3.66-3.75 (m, 2H), 3.76-3.92 (m, 3H), 4.64-4.75 (m,
2H). HRMS (FAB) calcd for C32H54N2O12: 703.3395 [M + 2Na -
2H]. Found: 703.3395.
CH2CH3), 2.13 (s, 3H, Ac), 2.15 (s, 3H, Ac), 2.57 (dd, 1H, JH3ax
)
12.8 Hz, JH4 ) 4.8 Hz, H3eq), 3.15-3.29 (m, 2H, OCH2CH2, 1H,
CH2CH2NH), 3.71-3.81 (m, 1H, CH2CH2NH), 3.79 (s, 3H,
CO2CH3), 4.02-4.14 (m, 3H, H5, H6, H9a), 4.25 (t, 1H, JOCH2CH
) 7.2 Hz, OCH2CH), 4.34 (dd, 1H, JH8 ) 2.4 Hz, JH9a ) 12.8 Hz,
H9b), 4.39 (d, 2H, JOCH2CH ) 7.2 Hz, OCH2CH), 4.81-4.90 (m,
1H, H4), 5.05-5.17 (m, 2H, NH), 5.29 (d, 1H, JH8 ) 8.8 Hz, H7),
5.39-5.45 (m, 1H, H8).
Methyl (5-Fmoc-aminopentyl[5-N-butanoyl-4,7,8,9-tetra-O-
acetyl-3,5-dideoxy-D-glycero-r-galacto-2-nonylopyranosyl])-
1
onate (20b). Yield, 68 mg (66%). [R]2D0 -8.9 (c 1.0, MeOH). H
NMR (CDCl3): δ 0.91 (t, 3H, JCH2CH3) 7.2 Hz, CH2CH3),
1.32-1.45 (m, 2H, OCH2CH2CH2CH2CH2NH), 1.49-1.61 (m, 6H,
OCH2CH2CH2CH2CH2NH, CH2CH2CH3), 1.95 (bt, 1H, J ) 12.8
5-(2-Decyloxy-3,4-dioxocylobut-1-enyl)aminopenthyl-[5-N-
isobutanoyl-3,5-dideoxy-D-glycero-r-D-galacto-2-nonulopyrano-
sylonic Acid (21c). Yield, 31 mg (80%). [R]2D0 4.7 (c 1.0, MeOH).
1H NMR (CD3OD): δ 0.90 (t, 3H, JCH2CH3 ) 6.8 Hz, CH2CH3),
1.13 (d, 6H, JCH(CH3)2 ) 6.8 Hz, -CH(CH3)2), 1.25-1.50 (m, 16H),
1.52-1.71 (m, 5H), 1.75-1.89 (m, 2H, C(sp2)OCH2CH2), 2.53 (sep,
1H, JCH(CH3)2 ) 6.8 Hz, CH(CH3)2), 2.83 (dd, 1H, JH4 ) 4.8 Hz,
JH3ax ) 16.4 Hz, H3eq), 3.38-3.53 (m, 3H), 3.54-3.77 (m, 5H),
3.77-3.95 (m, 3H), 4.65-4.76 (m, 2H). HRMS (FAB) calcd for
C32H54N2O12: 703.3395 [M + 2Na - 2H]. Found: 703.3401.
Conjugation of 21a-c to HSA. Each of the compounds 21a-c
(14.5 mg, 0.0225 mmol) was added to HSA (25 mg, 0.38 µmol) in
1.0 mL of NaHCO3 buffer (pH 9.1). The mixtures were stirred at
room temperature for 27 h, dialyzed against water (3 × 1000 mL)
for 24 h, and lyophilized. The glycoproteins 22a-c were obtained
as white powders with 14, 17, and 16 saccharides per protein
according to MALDI TOF MS and as confirmed by gel electro-
phoresis.
Hz, H3ax), 2.00 (s, 3H, Ac), 2.01 (s, 3H, Ac), 2.08 (t, 2H, JCH2CH2
)
7.6 Hz, CH2CH2), 2.14 (s, 3H, Ac), 2.14 (s, 3H, Ac), 2.58 (dd, 1H,
JH3ax ) 12.8 Hz, JH4 ) 4.4 Hz, H3eq), 3.16-3.28 (m, 2H, OCH2CH2,
1H, CH2CH2NH), 3.73-3.82 (m, 1H, CH2CH2NH), 3.79 (s, 3H,
CO2CH3), 4.01-4.06 (m, 1H, H9a), 4.08-4.13 (m, 2H, H5, H6),
4.21(t, 1H, JOCH2CH ) 6.8 Hz, OCH2CH), 4.33 (dd, 1H, JH8 ) 2.8
Hz, JH9a ) 12.0 Hz, H9b), 4.39 (d, 2H, JOCH2CH ) 6.8 Hz,
OCH2CH-), 4.81-4.91 (m, 1H, H4), 5.06-5.16 (m, 2H, NH), 5.28
(d, 1H, JH8 ) 8.4 Hz, H7), 5.37-5.44 (m, 1H, H8).
Methyl (5-Fmoc-aminopentyl[5-N-isobutanoyl-4,7,8,9-tetra-
O-acetyl-3,5-dideoxy-D-glycero-r-galacto-2-nonylopyranosyl])-
1
onate (20c). Yield, 29 mg (18%). [R]2D0 -10.6 (c 1.0, MeOH). H
NMR (CDCl3): δ 1.06 (d, 3H, JCH(CH3)2 ) 7.2 Hz, CH(CH3)2), 1.08
(d, 3H, JCH(CH3)2 ) 7.2 Hz, CH(CH3)2), 1.32-1.44 (m, 2H,
OCH2CH2CH2CH2CH2NH), 1.48-1.61 (m, 4H, OCH2CH2-
CH2CH2CH2NH), 1.96 (bt, 1H, J ) 12.0 Hz, H3ax), 1.99 (s, 3H,
Ac), 2.01 (s, 3H, Ac), 2.13 (s, 3H, Ac), 2.15 (s, 3H, Ac), 2.18-2.26
(m, 1H, CH(CH3)2), 2.57 (dd, 1H, JH3ax ) 12.0 Hz, JH4 ) 4.8 Hz,
H3eq), 3.14-3.29 (m, 2H, OCH2CH2, 1H, CH2CH2NH), 3.72-3.78
(m, 1H, CH2CH2NH), 3.79 (s, 3H, CO2CH3), 4.01-4.08 (m, 1H,
H9a), 4.09-4.14 (m, 2H, H5, H6), 4.21 (t, 1H, JOCH2CH ) 7.2 Hz,
OCH2CH-), 4.33 (dd, 1H, JH8 ) 2.8 Hz, JH9a ) 12.4 Hz, H9b),
4.40 (d, 2H, JOCH2CH ) 7.2 Hz, OCH2CH), 4.83-4.92 (m, 1H, H4),
5.08-5.16 (m, 2H, NH), 5.26 (d, 1H, JH8 ) 8.4 Hz, H7), 5.34-5.43
(m, 1H, H8).
Acknowledgment. This work was supported by the Swedish
Research Council, the J. C. Kempe Foundation, and the Jeansson
Foundation. We thank Jan Kihlberg for valuable support
throughout the project. Gunnar Gro¨nberg of AstraZeneca R&D,
Mo¨lndal, Sweden, is gratefully acknowledged for help with
structure determination of the degraded sialic acids 9a,b. We
are grateful to Anna Linusson and David Andersson for their
support in computational chemistry. We thank Yvonne Nygren
for valuable help with HPLC-MS analysis.
General Procedure for Preparation of 5-(2-Decyloxy-3,4-
dioxocylobut-1-enyl)aminopenthyl-[5-N-acyl-3,5-dideoxy-D-glyc-
ero-r-D-galacto-2-nonulopyranosylonic Acids (21a-c). The com-
pounds 20a-c (0.06-0.07 mmol) were stirred in methanolic sodium
methoxide (6 mL, 0.03 M) for 1 h. The solutions were neutralized
with Amberlite IR-120 and concentrated under reduced pressure.
The residues were dissolved in MeOH (5.0 mL) before LiOH (0.4
mL, 1 M) was added. The solutions were stirred at room temperature
Supporting Information Available: General chemical methods
and materials; experimental procedures and data for compounds
2a-e, 7a,b, 3a,b, and 10a-i; 13C NMR data for compounds 13,
14, 16a-h, 17a-h, 18, 19, 20a-c, and 21a-c; experimental
procedures for the binding and infectivity assays. This material is
available free of charge via the Internet at http://pubs.acs.org.

3678 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 12
Johansson et al.
(23) Burmeister, W. P.; Guilligay, D.; Cusack, S.; Wadell, G.; Arnberg,
N. Crystal structure of species D adenovirus fiber knobs and their
sialic acid binding sites. J. Virol. 2004, 78, 7727–7736.
(24) Taylor, R. D.; Jewsbury, P. J.; Essex, J. W. A review of protein-
small molecule docking methods. J. Comput.-Aided Mol. Des. 2002,
16, 151–166.
(25) Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor,
R. D. Improved protein-ligand docking using gold. Proteins: Struct.,
Funct., Genet. 2003, 52, 609–623.
(26) Gohlke, H.; Klebe, G. Approaches to the description and prediction
of the binding affinity of small-molecule ligands to macromolecular
receptors. Angew. Chem., Int. Ed. 2002, 41, 2645–2676.
(27) Marra, A.; Sinay, P. Stereoselective synthesis of 2-thioglycosides of
N-acetylneuraminic acid. Carbohydr. Res. 1989, 187, 35–42.
(28) Waglund, T.; Claesson, A. Stereoselective synthesis of the alpha-allyl
C-glycoside of 3-deoxy-D-manno-2-octulosonic acid (KDO) by use
of radical chemistry. Acta Chem. Scand. 1992, 46, 73–76.
(29) Sherman, A. A.; Yudina, O. N.; Shashkov, A. S.; Menshov, V. M.;
Nifantiev, N. E. Preparative route to N-glycolylneuraminic acid phenyl
2-thioglycoside donor and synthesis of Neu5Gc-alpha-(2 f 3′)-
lactosamine 3-aminopropyl glycoside. Carbohydr. Res. 2002, 337,
451–457.
(30) Saito, K.; Sugai, K.; Fujikura, K.; Yamada, N.; Goto, M.; et al.
Thermal-degradation of sodium N-acetylneuraminate. Carbohydr. Res.
1989, 185, 307–314.
(31) Byramova, N. E.; Tuzikov, A. B.; Bovin, N. V. Studies on the synthesis
of sialosides and sialic-acid analogs. 2. A simple procedure for the
synthesis of the methyl and benzyl glycosides of Neu5Ac and 4-epi-
Neu5Acsconversion of the benzyl and methyl glycosides of Neu5Ac
into N-trifluoroacetylneuraminic acid benzyl glycosides. Carbohydr.
Res. 1992, 237, 161–175.
(32) Hori, H.; Nakajima, T.; Nishida, Y.; Ohrui, H.; Meguro, H. A simple
method to determine the anomeric configuration of sialic-acid and its
derivatives by C13-NMR. Tetrahedron Lett. 1988, 29, 6317–6320.
(33) Mei, Y. F.; Lindman, K.; Wadell, G. Two closely related adenovirus
genome types with kidney or respiratory tropism differ in their binding
to epithelial cells of various origins. Virology 1998, 240, 254–266.
(34) Segerman, A.; Arnberg, N.; Eriksson, A.; Lindman, K.; Wadell, G.
There are two different species B adenovirus receptors: sBAR,
common to species BB and B2 adenoviruses, and SB2AR, excusively
used by species B2 adenoviruses. J. Virol. 2003, 77, 1157–1162.
(35) Wadell, G.; Allard, A.; Hierholzer, J. C. Manual of Clinical Micro-
biology; ASM Press: Washington, DC, 1999; 970-982.
(36) Reduce 2.21 for Linux; The Richardson Laboratory, Duke University:
Durham, NC; http://kinemage.biochem.duke.edu (accessed Feb 2005).
(37) CS Chemdraw Ultra, version 7.0.1 for Macintoch; CambridgeSoft
Corp.: Cambridge, MA, 2002.
References
(1) Flint, S. J.; Enquist, L. W.; Racaniello, V. R.; Skalka, A. M. Principles
of Virology, 2nd ed.; ASM Press: Washington, D.C., 2004.
(2) Uhnoo, I. Antivirala la¨kemedels mo¨jligheter och begra¨nsningar. Tidskr.
Medikament 2002, 2, 24–29.
(3) Jones, P. S. Strategies for antiviral drug discovery. AntiViral Chem.
Chemother. 1998, 9, 283–302.
(4) De Clercq, E. Strategies in the design of antiviral drugs. Nat. ReV.
Drug DiscoVery 2002, 1, 13–25.
(5) Mammen, M.; Choi, S. K.; Whitesides, G. M. Polyvalent interactions
in biological systems: Implications for design and use of multivalent
ligands and inhibitors. Angew. Chem., Int. Ed. 1998, 37, 2755–2794.
(6) Kiessling, L. L.; Pohl, N. L. Strength in numbers: Non-natural
polyvalent carbohydrate derivatives. Chem. Biol. 1996, 3, 71–77.
(7) Borman, S. Multivalency: Strength in numbers. Chem. Eng. News 2000,
78, 48–53.
(8) Ford, E.; Nelson, K. E.; Warren, D. Epidemiology of epidemic
keratoconjunctivitis. Epidemiol. ReV. 1987, 9, 244–261.
(9) Gordon, Y. J.; Aoki, K.; Kinchington, P. R. Adenovirus Keratocon-
junctivitis. In Ocular Infection and Immunity; Pepose, J. S., Holland,
G. N., Wilhelmus, K. R., Eds.; C.V. Mosby: St. Louis, MO, 1996.
(10) Aoki, K.; Kawana, R.; Matsumoto, I.; Wadell, G.; Dejong, J. C. Viral
conjunctivitis with special reference to adenovirus type-37 and
enterovirus-70 infection. Jpn. J. Ophthalmol. 1986, 30, 158–164.
(11) Paparello, S. F.; Rickman, L. S.; Mesbahi, H. N.; Ward, J. B.; Siojo,
L. G.; et al. Epidemic keratoconjunctivitis at a United-States military
basesRepublic of the Philippines. Mil. Med. 1991, 156, 256–259.
(12) Kemp, M. C.; Hierholzer, J. C.; Cabradilla, C. P.; Obijeski, J. F. The
changing etiology of epidemic keratokonjunctivitis: Antigenic and
restriction enzyme analyses of adenovirus types 19 and 37 isolated
over a 10-year period. J. Infect. Dis. 1983, 148, 24–33.
(13) Aoki, K.; Yoshitugu, T. A twenty-one year surveillance of adenoviral
conjunctivitis in sapporo, japan. Int. Ophthamol. Clin. 2002, 42, 49–
54.
(14) Henry, L. J.; Xia, D.; Wilke, M. E.; Deisenhofer, J.; Gerhard, R. D.
Characterisation of the knob domain of the adenovirustype 5 fiber
protein expressed in Escherichia coli. J. Virol. 1994, 68, 5239–5246.
(15) Arnberg, N.; Edlund, K.; Kidd, A. H.; Wadell, G. Adenovirus type
37 uses sialic acid as a cellular receptor. J. Virol. 2000, 74, 42–48.
(16) Arnberg, N.; Kidd, A. H.; Edlund, K.; Olfat, F.; Wadell, G. Initial
interactions of subgenus D adenoviruses with A549 cellular receptors:
Sialic acid versus alpha(v) integrins. J. Virol 2000, 74, 7691–7693.
(17) Kiessling, L. L.; Gestwicki, J. E.; Strong, L. E. Synthetic multivalent
ligands in the exploration of cell-surface interactions. Curr. Opin.
Chem. Biol. 2000, 4, 696–703.
(18) Johansson, S. M. C.; Arnberg, N.; Elofsson, M.; Wadell, G.; Kihlberg,
J. Multivalent HSA conjugates of 3′-siallyllactose are potent inhibitors
of adenoviral cell attachment and infection. ChemBioChem 2005, 6,
358–364.
(19) Johansson, S. M. C.; Nilsson, E. C.; Elofsson, M.; Ahlskog, N.;
Kihlberg, J.; et al. Multivalent sialic acid conjugates inhibit adenovirus
type 37 from binding to and infecting human corneal epithelial cells.
AntiViral Res. 2007, 73, 92–100.
(20) Tietze, L. F.; Arlt, M.; Beller, M.; Glusenkamp, K. H.; Jahde, E.; et
al. Anticancer agents. 15. Squaric acid diethyl ester, a new coupling
reagent for the formation of drug biopolymer conjugatesssynthesis
of squaric acid ester amides and diamides. Chem. Ber. 1991, 124,
1215–1221.
(21) Tietze, L. F.; Schroter, C.; Gabius, S.; Brinck, U.; Goerlachgraw, A.;
et al. Conjugation of para-aminophenyl glycosides with squaric acid
diester to a carrier protein and the use of neoglycoprotein in the
histochemical detection of lectins. Bioconjugate Chem. 1991, 2, 148–
153.
(38) Weininger, D. SMILES, a chemical language and information-system.
1. Introduction to methodology and encoding rules. J. Chem. Inf.
Comput. Sci. 1988, 28, 31–36.
(39) CORINA; Molecular Networks GmbH: Erlangen, Germany; http://
www2.chemie.uni-erlangen.de/sofware/corina/free_struct.html (ac-
cessed Feb 2005).
(40) GOLD, version 3.1.1; The Cambridge Crystallographic Data Center:
Cambridge, U.K., 2004.
(41) MOEsMolecular Operating EnVironment, version 2004.03; Chemical
Computing Group Inc.: Montreal, Canada, 2004.
(42) SYBYL, version 7.2; Tripos Inc.: St. Louis, MO, 2004.
(43) Jones, G.; Willett, P.; Glen, R. C. Molecular recognition of receptor-
sites using a genetic algorithm with a description of desolvation. J.
Mol. Biol. 1995, 245, 43–53.
(44) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible docking.
J. Mol. Biol. 1997, 267, 727–748.
(45) Rhodes, G. Crystallography Made Crystal Clear, 2nd ed.; Academic
Press: San Diego, CA, 2000.
(22) Bergh, A.; Magnusson, B. G.; Ohlsson, J.; Wellmar, U.; Nilsson, U. J.
Didecyl squarate, a practical amino-reactive cross-linking reagent for
neoglycoconjugate synthesis. Glycoconjugate J. 2001, 18, 615–621.
JM801609S