Quantitative studies of the binding of the class II PapG adhesin from uropathogenic Escherichia coli to oligosaccharides

Article abstract of DOI:10.1016/S0968-0896(03)00114-7

Binding of the class II PapG adhesin, found at the tip of filamentous pili on Escherichia coli, to the carbohydrate moiety of globoseries glycolipids in the human kidney is a key step in development of pyelonephritis, a severe form of urinary tract infection. An assay based on surface plasmon resonance for quantification of the binding of the class II PapG adhesin to oligosaccharides has been developed. Using this assay dissociation constants ranging from 80 to 540 μM were determined for binding of the PapG adhesin to di-pentasaccharide fragments from the globoseries of glycolipids. A series of galabiose derivatives, modified at the anomeric position, O-2′ or O-3′, was also investigated. The anomeric position appeared to be the most promising for development of improved inhibitors of PapG-mediated adhesion of E. coli. p-Methoxyphenyl galabioside was found to be most potent (K_d=140 μM), and binds to PapG almost as well as the Forssman pentasaccharide.

Full text of DOI:10.1016/S0968-0896(03)00114-7

Bioorganic & Medicinal Chemistry 11 (2003) 2255–2261
Quantitative Studies of the Binding of the Class II PapG Adhesin
from Uropathogenic Escherichia coli to Oligosaccharides
Andreas Larsson,a Jorgen Ohlsson,b Karen W. Dodson,c Scott J. Hultgren,c
Ulf Nilssonb and Jan Kihlberga,*
a
˚
˚
Organic Chemistry, Department of Chemistry, Umea University, SE-901 87 Umea, Sweden
bBioorganic Chemistry, Lund University, PO Box 124, SE-221 00 Lund, Sweden
cDepartment of Molecular Microbiology, Washington University School of Medicine, St. Louis, MO 63110, USA
Received 13 September 2002; accepted 7 February 2003
Abstract—Binding of the class II PapG adhesin, found at the tip of filamentous pili on Escherichia coli, to the carbohydrate moiety
of globoseries glycolipids in the human kidney is a key step in development of pyelonephritis, a severe form of urinary tract infection.
An assay based on surface plasmon resonance for quantification of the binding of the class II PapG adhesin to oligosaccharides has
been developed. Using this assay dissociation constants ranging from 80 to 540 mM were determined for binding of the PapG
adhesin to di-pentasaccharide fragments from the globoseries of glycolipids. A series of galabiose derivatives, modified at the
anomeric position, O-20 or O-30, was also investigated. The anomeric position appeared to be the most promising for development
of improved inhibitors of PapG-mediated adhesion of E. coli. p-Methoxyphenyl galabioside was found to be most potent
(Kd=140 mM), and binds to PapG almost as well as the Forssman pentasaccharide.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
pyelonephritis adhere to the globoseries of glycolipids
found in the upper urinary tract. Several studies have
shown that the disaccharide galabiose (Gala1-4Gal),
which constitutes a part of these glycolipids, is required
for binding of pyelonephritic E. coli.6À10 Bacterial bind-
ing to the galabiose receptor is mediated by supramole-
cular protein appendages, termed P pili, which extend
from the outer cell membrane of the E. coli.11 P pili
consist of a thin tip fibrillum that is joined to the distal
end of a thicker pilus rod. Altogether the P pilus con-
sists of a large number of proteins of six different types,
with the galabiose-specific adhesin PapG located at the
very tip of the pilus. Three different variants exist of the
PapG adhesin (classes I–III), with the class II adhesin
being particularly associated with development of pye-
lonephritis in humans.12,13
Urinary tract infections (UTIs) belong to the most
common bacterial infections, second in occurrence only
to infections of the respiratory tract.1 UTIs are more
frequent in women than in men and one-third of the
women in the United States have suffered from at least
one UTI before the age of 65, with recurrent infections
being common.2 Recurrent bladder infections are asso-
ciated with an increased risk of developing a more severe
infection in the kidneys (pyelonephritis). Escherichia coli
is the predominant causative agent of pyelonephritis,3
and the ability of the bacterium to adhere to epithelial
cells in the human kidney has been shown to be critical
for development of disease.4
Adhesion to host tissue is an important virulence factor
in many infections caused by bacteria, and glycoconju-
gates on the host cell surface often function as receptors
for adhesive proteins expressed by bacteria and viru-
ses.5 An overwhelming majority of E. coli which cause
Most proteins with lectin-like functions bind naturally
occurring carbohydrate ligands with low affinity (Kd
ꢀ0.1–1 mM). This constitutes a significant obstacle in
attempts to develop carbohydrate based drugs which
interfere with microbial attachment to host cell-surface
glycoconjugates. However, some successful strategies to
overcome this problem have been described recently.
When microbial attachment occurs by use of multiple
*Corresponding author. Tel.: +46-90-7866890; fax: +46-90-138885;
e-mail: jan.kihlberg@chem.umu.se
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00114-7

2256
A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
copies of an adhesin, multivalent carbohydrate ligands
have proven to be efficient inhibitors of attachment.14,15
Another attractive approach to enhance the affinity of a
ligand for a lectin is to use a small saccharide, which
fulfils key polar interactions with the lectin, as a scaffold
onto which substituents that enhance the affinity are
added.16,17
(cf. 1 vs 2, 3 vs 4, and 5 vs 6). In contrast, extension of
the galabiose part with a mono- or disaccharide unit at
O-30 did not influence the dissociation constant (com-
pare 5 with 3 and 1). These results agree well with inhib-
itory powers determined for saccharides 1–8, when used
as inhibitors of haemagglutination caused by E. coli
expressing the class II PapG adhesin, as well as adher-
ence of the same bacteria to kidney tissue sections.9
As part of a program aimed at development of carbo-
hydrate based inhibitors of pyelonephritic E. coli we
now describe the development of an assay based on
surface plasmon resonances for direct studies of the
binding of the class II PapG adhesin to saccharide
ligands. The assay was used to determine the dissocia-
tion constants for binding of PapG to the saccharide
moieties of the globoseries of glycolipids and fragments
thereof, as well as to a panel of substituted derivatives
of the disaccharide galabiose.
The crystal structure of the complex between globoside
3 and the oligosaccharide binding domain of the class II
PapG adhesin was recently determined at 1.8 A resolu-
tion (Fig. 2).18 In the complex a large number of
hydrogen bonds are formed between the adhesin and 3,
including both direct and water mediated hydrogen
bonds. Four hydrogen bonds are formed between the
adhesin and HO-4, O-5 and HO-6 of the GalNAc resi-
due. The Gala moiety participates in the most extensive
Results and Discussion
Binding of the class II PapG adhesin to oligosaccharides
The oligosaccharide binding domain of the class II
PapG adhesin, which consists of the N-terminal 196
amino acids, was recently cloned and expressed.18 This
PapGII truncate has now been covalently bound to the
surface of a dextrane-coated sensor chip in order to
investigate binding to receptor-active saccharides. Cou-
pling was accomplished by using a standard BIACORE
amine coupling procedure. Binding of saccharide moi-
eties from the globoseries of glycolipids to the adhesin
was then determined using surface plasmon resonance
(Table 1, Fig. 1). The adhesin was found to bind to
saccharides that contain a galabiose moiety (i.e., sac-
charides 1–6) with Kd-values ranging from 80 to
540 mM, that is, in the range usually found for lectin-
carbohydrate interactions. In agreement with previous
studies using piliated bacteria, which carried different
classes of PapG adhesins, saccharides which did not
contain galabiose were not bound at all (cf. 7 and 8).6À10
The presence of a glucose moiety at the reducing end of
the galabiose moiety led to a significant decrease in Kd,
that is, it increased the affinity for the PapG adhesin
Figure 1. Binding isotherms with concentration-dependent responses
at equilibrium for oligosaccharides 1 (*), 6 (&) and 11 (~) binding to
immobilized PapGII adhesin on a CM5 surface plotted versus oligo-
saccharide concentration in triplicates. The binding responses at equi-
librium were normalized against the maximum binding (Rmax).
Dissociation constants (Kd) were determined as the concentration of a
saccharide that elicited half of the maximal response (Rmax), that is,
when 50% of the PapGII adhesin bound a ligand.
Table 1. Dissociation constants for binding of the class II PapG
adhesin to the saccharide moieties of the globoseries of glycolipids,
and fragments thereofa
Compd
Kd (mM)
1
2
3
4
5
6
7
8
GalNAca3GalNAcb3Galꢀ4Galꢁ4GlcbOTMSEtb
GalNAca3GalNAcb3Galꢀ4GalꢁOTMSEtb
GalNAcb3Galꢀ4Galꢁ4GlcbOTMSEtb
GalNAcb3Galꢀ4GalꢁOTMSEtb
Galꢀ4Galꢁ4GlcbOTMSEtb
91
250
84
180
78
Galꢀ4GalꢁOTMSEtb
540
—
c
GalNAca3GalNAcb3GalaOMe
Galb4GlcbOTMSEtb
c
—
Figure 2. Structure of the crystalline complex between globoside 3 and
the class II PapG adhesin.18 The groove adjacent to HO-20 of the
galabiose moiety is indicated by an arrow. Polar parts of the PapG
surface are blue, nonpolar parts are brown, and green represents
intermediate polarity. The figure was generated with the program
Sybyl.
aDetermined by surface plasmon resonance using a Biacore 3000
instrument (cf. Experimental).
bTMSEt=2-trimethylsilylethyl.
cInactive, that is, no binding detected.

A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
2257
hydrogen-bonding network, involving seven hydrogen
bonds to HO-2, O-3, HO-4 and HO-6. For Galb four
hydrogen bonds to HO-3, O-5 and HO-6 are found,
whereas Glc takes part in three hydrogen bonds to HO-2
and HO-3. In addition, the complex contains several
hydrophobic contacts including H-1, H-2 and H-6 of
Gala, H-1, H-3, H-4, H-5 and H-6 of Galb, as well as
H-2 and H-4 of Glc. The dissociation constants deter-
mined for oligosaccharides 1–8 provide additional
insight into the importance of these interactions for the
stability of the complex. Since attachment of further
saccharides at O-3 of Gala in 5 does not influence the
Kd-values (compare 5, 3 and 1) it can be concluded that
the interactions between the GalNAc residue and the
adhesin do not contribute to the stability of the com-
plex. The critical role of the central galabiose moiety for
binding is easily understood in view of the extensive
contacts, including both hydrogen bonds and hydro-
phobic interactions, with the adhesin. The additional
stabilization provided by the Glc residue could be due
either to the hydrogen bonds formed with HO-2 and
HO-3, or to hydrophobic contacts with Trp107, which is
sandwiched below a non-polar region of the Glc and
Galb residues. Since attachment of aromatic aglycones
to galabiose gives potent binders (cf. below) it is con-
cluded that hydrophobic contacts play a significant role.
Table 2. Dissociation constants for binding of the class II PapG
adhesin to substituted derivatives of galabiose (Gala4Galb)a
Compd
R1
R2
R3
Kd (mM)
6
9
TMSEtOb
MeO
H
H
H
H
540
340
10
11
12
13
14
15
16
17
18
19
20
Cyclohexyl-O
pMeOPhO
2-Naphthyl-O
Ph(CO)NH
pMeOPhS
pMeOPhO
pMeOPhO
pMeOPhO
pMeOPhO
pMeOPhO
pMeOPhO
H
H
H
H
H
H
H
H
H
H
H
H
260
140
200
>1000
490
>1500
>1000
490
H
Me
nPr
MeOCH2
H
H
H
Allyl
Xc
mNO2Bn
140
290
120
aDetermined by surface plasmon resonance using a Biacore 3000
instrument (cf. Experimental).
bTMSEt=2-trimethylsilylethyl.
cX=
.
and 16. Substituents at O-20 thus have large, and differ-
ing, influences on the binding strength, which leads to
the speculation that selection of a suitable substituent at
this position might yield more potent ligands for the
adhesin. At O-30, the presence of an allyl group (18), or
a substituent obtained by addition of a protected cystein
to the allyl group (19), either did not affect or led to an
increase in the Kd-value. However, attachment of a
m-nitrobenzyl group at O-30 (cf. 20) resulted in a slight
lowering of the Kd value, indicating the possibility that
appropriate substituents at this position could have a
beneficial influence on binding.
Binding of galabiose derivatives to the class II PapG
adhesin. The binding studies performed with saccharides
1–8, and the crystal structure of the complex between 3
and the class II adhesin,18 suggest that substituents
could be added at several positions of a galabiose core
in efforts to develop improved PapG binders. The role
of the anomeric position is highlighted by the additional
stabilization provided by the presence of a glucose resi-
due (cf. 6 vs 5). Since the GalNAc residue is tolerated at
O-30 of the galabiose moiety it could well be possible to
increase the affinity by attachment of substituents at this
position. In addition, the crystal structure of the com-
plex reveals the presence of a groove adjacent to HO-20
(Fig. 2), which potentially could be filled by addition of
substituents to O-20 of galabiose.
In an attempt to understand the reasons for the
improved binding of p-methoxyphenyl galabioside 11 to
PapG, as compared to for example, alkyl galabiosides 6
and 9, galabioside 11 was docked into the active site of
the class II PapG adhesin. This was done by manual
superimposition of 11 onto the galabiose moiety of tet-
rasaccharide 3 in the crystalline complex with PapG,18
followed by energy minimization using Macromodel. In
the resulting complex (Fig. 3) all interactions between
the galabiose moiety and PapG that were found in the
crystalline complex of PapG and tetrasaccharide 3 are
maintained. In addition, the p-methoxyphenyl group is
located in a pocket formed by Trp107 and Arg170 of
PapG. Most likely this stabilizes the complex by hydro-
phobic interactions with the aromatic side chain of
Trp107. Additional stabilization appears to be obtained
through a p-cation interaction between the p-methoxy-
phenyl group and Arg170, thus explaining why galabio-
sides with aromatic aglycons bind better than those
having aliphatic aglycones.
The role of substituents at the anomeric position of
galabiose was probed with different O-alkyl (6, 9, and
10), O-aryl (11 and 12), benzamido (13), and S-aryl (14)
groups (Table 2, Fig. 1). The 2-trimethylsilylethyl
glycoside 6, as well as benzamide 13 and thioglycoside
14, bound with lower affinity than the simple methyl
glycoside 9. More potent ligands were obtained when
O-aryl groups were located at the anomeric position,
with the p-methoxyphenyl glycoside 11 being the most
potent binder (Kd=140 mM). It should be pointed out
that 1119 binds almost as well as pentasaccharide 1 (Fig.
1),20 which is considerably more demanding to prepare.
In view of the high potency of 11, the p-methoxyphenyl
group was maintained at the anomeric position during
preliminary studies with substituents at O-20 and O-30.
The extended pocket in the adhesin adjacent to O-20 was
probed with compounds 15–17. Attachment of a methyl
(15) or propyl (16) group at O-20 gave very poor bin-
ders. In contrast, use of a methoxymethyl group (17)
gave a significantly improved binder, as compared to 15
Synthesis of oligosaccharides and galabiose derivatives.
The synthesis of the saccharides used in the present
study, except 15–17 and 19, has been described previously

2258
A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
Scheme 2. (a) (i) N-acetyl-l-cystein methyl ester, AIBN, hu, EtOAc,
48 h; (ii) NaOMe, MeOH.
in 90% yield. As an alternative, a galactosyl donor with
an allyl group instead of a p-methoxybenzyl group at
HO-2 was prepared from 21 by a similar reaction
sequence as for 23. However, in the following a-glyco-
sylation a modest yield of 68% was obtained and this
route was not explored further. Conventional, base-cat-
alyzed debenzoylation of 25 followed by benzylation
gave 26 in 83% yield. Treatment of 26 with 2% tri-
fluoroacetic acid in dichloromethane25,26 then afforded
galabioside 27, which has HO-2 unprotected. The 20-O-
modified galabiosides (15–17) were prepared in 41–81%
yield by treatment of 27 with NaH and either of methyl
iodide, allyl bromide, or bromomethyl methyl ether, fol-
lowed by hydrogenolys over Pd/C in glacial acetic acid.
Galabioside 19 was prepared in two steps from the known
2819 (Scheme 2). UV-mediated radical addition of N-ace-
tyl-l-cystein methyl ester to the allyl group at O-30 of 28,
followed by conventional O-deacylation gave 21 in 65%
overall yield.
Figure 3. An energy-minimized model of the complex between gala-
bioside 11 and the class II PapG adhesin. Polar parts of the PapG
surface are blue, nonpolar parts are brown, and green represents
intermediate polarity. The figure was generated with the program
Sybyl.
(cf. references in the Experimental). For the preparation
of the 20-O-modified p-methoxyphenyl galabiosides
(15–17), a benzylated galabioside with HO-20 unpro-
tected was required (Scheme 1). Synthesis of galactosyl
donor 23 was achieved by protection of HO-2 of 2121 as
a p-methoxybenzyl ether, followed by removal of the
isopropylidene and 1-methoxyisopropyl groups (! 22,
79% yield), and subsequent benzylation (! 23, 89%
yield). a-Galactosylation of acceptor 2422 with 23, using
N-iodosuccinimide and trimethylsilyl trifluoromethane-
sulfonate as promotor system,23,24 gave galabioside 25
Conclusions
An assay based on surface plasmon resonance that
allows determination of dissociation constants for
binding of the class II PapG adhesin of uropathogenic
E. coli to oligosaccharides has been developed. The
adhesin was found to bind to saccharides from the
globoseries of glycolipids, which function as ligands for
E. coli in the upper urinary tract, with Kd-values ranging
from 80 to 540 mM. These values agree well with those
usually found for interactions between lectins and car-
bohydrates. A series of galabiose derivatives, modified
at the anomeric position, O-20 or O-30, was also investi-
gated. The anomeric position appeared to be the most
promising from the point of view of development of
improved inhibitors of adhesion of E. coli to host tissue.
p-Methoxyphenyl galabioside (11) was found to be a
most potent galabioside (Kd=140 mM), and binds to
PapG almost as well as the Forssman pentasaccharide,
which is considerably more difficult to synthesize.
Experimental
Scheme 1. (a) (i) NaH, pMeOBnBr, DMF; (ii) 80% aqueous AcOH,
60 ꢁC. (b) NaH, BnBr, DMF. (c) NIS, TMSOTf, CH2Cl2–Et2O (1:1),
À50 ꢁC. (d) (i) NaOMe, MeOH; (ii) NaH, BnBr, DMF. (e) 2% TFA,
CH2Cl2, 0 ꢁC. (f) (i) NaH, alkyl halide, DMF; (ii) H2 (1 bar), 10% Pd/
C, MeOH.
Synthetic saccharides (1–20). The saccharides used in
the present study, except 15–17 and 19, were synthesized
as described: 1,20 2,20 3 and 4;27 5,28 6,29 7,20 8,29 9,30
10,31 11,19 12–14,31 18,19 20.19

A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
2259
General synthetic procedures. All non-aqueous reactions
were run in septum-capped, oven-dried flasks under Ar
(1 atm). CH2Cl2 was dried by distillation from CaH2
and Et2O was distilled from Na. Concentration of
organic solutions were made using rotary evaporation
with a bath temperature at or below 40 ꢁC. Flash
chromatography was performed on Grace Amicon
Silica gel 60 (35–70 mm) and TLC was performed on
Kieselgel 60 F254 plates (Merck). NMRspectra were
recorded with a Bruker DRX-400 or a Bruker ARX-300
instrument for solutions in CDCl3 or MeOH-d4 (resi-
dual CHCl3 or CD2HOD were used as internal refer-
ences at 7.27 and 3.31 ppm, respectively). 1H NMR
spectral assignments were made based on COSY spectra.
J=11.7 Hz, OCH2Ph), 4.01 (d, 1H, J=2.4 Hz, H-4),
3.94 (t, 1H, J=9.4 Hz, H-2), 3.85 (s, 3H, OMe), 3.70 (m,
2H, H-6), 3.64–3.61 (m, 2H, H-3, H-5), 2.33 (s, 3H,
Me); 13C NMR(CDCl ): d 159.7, 139.2, 138.8, 138.3,
3
137.6, 132.6, 131.0, 130.7, 130.4, 130.0, 128.9, 128.6,
128.4, 128.3, 128.2, 128.1, 128.0, 127.9, 114.2, 88.5, 84.7,
75.7, 74.8, 74.0, 74.00, 73.2, 69.2, 55.7, 21.5; HRMS
calcd for C42H44O6SNa (M+Na): 699.2756, found:
699.2773.
4-Methoxyphenyl 3,4,6-tri-O-benzyl-2-O-p-methoxybenzyl-
ꢁ-D-galactopyranosyl-(1!4)-2,3,6-tri-O-benzoyl-ꢀ-D-
galactopyranoside (25). CH2Cl2 (1.5 mL) and Et2O
(3.0 mL) were added to a mixture of 2422 (63 mg,
0.106 mmol), 23 (100 mg, 0.148 mmol) and N-iodosucci-
nimide (38 mg, 0.170 mmol), and the mixture was cooled
to À55 ꢁC. Trimethylsilyl trifluoromethanesulfonate
(4 mL, 20 mmol) was added and the mixture was stirred
for 1 h. Triethylamine (0.1 mL) was then added and the
mixture was stirred for 0.5 h at À55ꢁC. The mixture was
allowed to obtain room temperature, and was diluted with
CH2Cl2 (10 mL), washed with 10% aqueous Na2S2O3
(2 mL) and saturated aqueous NaHCO3 (2 mL), dried
(MgSO4), and concentrated. The residue was purified
by flash chromatography (SiO2, 3:1 heptane–EtOAc) to
4-Methylphenyl 2-O-(p-methoxybenzyl)-1-thio-ꢀ-D-galacto-
pyranoside (22). Sodium hydride (55 mg, 1.34 mmol,
60% in mineral oil) was added to a solution of 2121
(410 mg, 1.03 mmol) in DMF (8 mL). The mixture was
stirred for 15 min, cooled to 0 ꢁC, and then p-methoxy-
benzyl chloride (193 mL 1.43 mmol) was added drop-
wise. The mixture was stirred at room temperature for
4 h then methanol (1 mL) was added. The mixture was
diluted with CH2Cl2 (30 mL), washed with saturated
aqueous NaHCO3 (10 mL), dried (Na2SO4), filtered,
and concentrated. The residue was purified by flash
chromatography (SiO2, 3:1 heptane–EtOAc, 2% Et3N)
to give a fully protected galactoside as intermediate. To
the fully protected galactoside was added 80% aqueous
acetic acid and the solution was stirred at 60 ꢁC for 2 h.
The mixture was concentrated, co-concentrated with
toluene, and the residue was purified by flash chroma-
tography (SiO2, 3:1 toluene–acetone, 2% MeOH) to
23
1
give 25 (110 mg, 90%): ½aꢂD +69 (c 1.0, CHCl3); H
NMR(CDCl ): d 8.11–7.97 (m, 6H, Ar–H), 7.67–7.27
3
(m, 26H, Ar–H), 7.01 (m, 2H, OPhOMe), 6.72 (m, 4H,
OPhOMe, OCH2PhOMe), 6.04 (dd, 1H, J=7.8,
10.5 Hz, H-2), 5.32 (dd, 1H, J=2.9, 10.6 Hz, H-3), 5.19
(d, 1H, J=7.8 Hz, H-1), 4.94–4.80 (m, 7H), 4.68 (A-part
of ABq, 1H, J=11.4 Hz,), 4.53 (B-part of ABq, 1H,
J=11.1 Hz,), 4.49 (d, 1H, J=2.6 Hz, H-4), 4.42 (dd, 1H,
J=5.0, 9.0 Hz, H-50), 4.25–4.14 (m, 5H), 4.10 (dd, 1H,
J=3.4, 10.2 Hz, H-6), 3,76 (s, 3H, OMe), 3.68 (s, 3H,
OMe), 3.45 (t, 1H, J=8.8 Hz, H-60), 3.01 (dd, 1H,
J=5.0, 8.4 Hz, H-60); 13C NMR(CDCl 3): d 166.9,
166.5, 165.8, 159.5, 155.9, 151.7, 139.3, 139.2, 138.8,
133.8, 133.6, 133.6, 130.9, 130.4, 130.4, 130.3, 130.2,
130.1, 130.0, 129.5, 128.9, 128.9, 128.8, 128.8, 128.7, 128.5,
128.4, 128.0, 127.9, 127.8, 127.8, 119.2, 114.8, 114.1,
101.8, 101.4, 79.4, 76.1, 76.0, 75.4, 75.2, 74.6, 74.1, 73.6,
73.4, 73.0, 70.3, 70.0, 67.9, 63.3, 56.0, 55.5; HRMS calcd
for C69H66O16Na (M+Na): 1173.4249, found:
1173.4250.
23
give 22 (327 mg, 79%): ½aꢂD À3 (c 1.0, MeOH); 1H
NMR(CD OD): d 7.44 (m, 2H, SPhMe), 7.37 (m, 2H,
3
OCH2PhOMe), 7.11 (m, 2H, SPhMe), 6.87 (m, 2H,
OCH2PhOMe), 4.72 (s, 2H, OCH2PhOMe), 4.56 (d, 1H,
J=9.2 Hz, H-1), 3.88 (dd, 1H, J=0.8, 3.3 Hz, H-4),
3.78–3.68 (m, 5H, H-6, OMe), 3.64–3.55 (m, 2H, H-2,
H-3), 3.50 (dt, 1H, J=0.8, 5.4 Hz, H-5), 2.30 (s, 3H,
CH3); 13C NMR(CDCl ): d 160.9, 138.5, 132.7, 132.6,
3
132.1, 131.1, 130.7, 114.7, 89.7, 80.6, 79.5, 76.6, 76.1,
70.9, 62.7, 55.8, 21.2; HRMS calcd for C21H26O6SNa
(M+Na): 429.1348, found: 429.1351.
4-Methylphenyl 3,4,6-tri-O-benzyl-2-O-(p-methoxybenzyl)-
1-thio-ꢀ-D-galactopyranoside (23). Sodium hydride
(24 mg, 0.60 mmol, 60% in mineral oil) was added to a
solution of 22 (68 mg, 0.17 mmol) in DMF (2 mL). The
mixture was stirred for 15 min and then cooled to 0 ꢁC.
Benzyl bromide (77 mL, 0.65 mmol) in DMF (1 mL) was
added dropwise and the resulting mixture was stirred at
room temperature for 2 h after which methanol (0.5 mL)
was added. The mixture was diluted with CH2Cl2
(10 mL), washed with saturated aqueous NaHCO3
(5 mL), dried (MgSO4), and concentrated. The residue
was purified by flash chromatography (SiO2, 5:1 hep-
4-Methoxyphenyl 3,4,6-tri-O-benzyl-2-O-p-methoxybenzyl-
ꢁ-D -galactopyranosyl-(1!4)-2,3,6-tri-O-benzyl-ꢀ-D-
galactopyranoside (26). NaOMe (50 mL, 1 M in MeOH)
was added to a solution of 25 (110 mg, 96 mmol) in a
mixture of methanol (4 mL) and CHCl3 (1 mL), and the
solution was stirred over night. Methanolic acetic acid
(10%) was added until a neutral reaction on moist
pH-paper was obtained. The resulting mixture was
concentrated and the residue was purified by flash
chromatography (SiO2, 2:1 toluene–acetone). Sodium
hydride (15 mg, 0.37 mmol, 60% in mineral oil) was
added to the debenzoylated product in DMF (3 mL).
The mixture was stirred for 15 min and then cooled to
0 ꢁC. Benzyl bromide (51 mL, 0.43 mmol) was added
dropwise and the resulting mixture was stirred at room
temperature for 2 h, then methanol (0.5 mL) was added.
Concentration and purification of the residue by flash
23
tane–EtoAc) to give 23 (101 mg, 89%): ½aꢂD +4 (c 1.0,
CHCl3); 1H NMR(CDCl 3): d 7.53–7.50 (m, 2H, SPhMe),
7.41–7.30 (m, 17H, Ar–H), 7.04 (m, 2H, SPhMe), 6.90
(m, 2H, OCH2PhOMe), 5.01 and 4.65 (ABq, 1H each,
J=11.5 Hz, OCH2Ph), 4.80–4.69 (m, 4H), 4.63 (d, 1H,
J=9.5 Hz, H-1), 4.51 and 4.47 (ABq, 1H each,

2260
A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
chromatography (SiO2, 3:1 heptane–EtOAc) gave 26
J=7.4 Hz, H-1), 4.20 (t, 1H, J=5.6 Hz, H-50), 4.02 (d,
1H, J=2.7 Hz, H-4), 3.93–3.71 (m, 11H), 3.62–3.52 (m,
23
(88 mg, 83%): ½aꢂD +39 (c 1.0, CHCl3); 1H NMR
(CDCl3): d 7.23–7.00 (m, 32H, Ar–H), 6.88 (m, 2H,
OPhOMe), 6.62 (m, 2H, OPhOMe), 6.56 (m, 2H,
OCH2PhOMe), 4.83–4.62 (m, 9H), 4.45–4.37 (m, 3H),
4.28 (dd, 1H, J=4.4, 8.8 Hz, H-50), 4.07–3.91 (m, 7H),
3.88 (d, 1H, J=2.8 Hz, H-4), 3.82–3.73 (m, 2H, H-2, H-
6), 3.58 (s, 3H, OMe), 3.53 (s, 3H, OMe), 3.48–3.38 (m,
3H, H-5, H-6, H-60), 3.29 (dd, 1H, J=2.8, 9.9 Hz, H-3),
5H); 13C NMR(CD OD): d 156.8, 153.2, 119.5, 115.6,
3
104.2, 100.7, 80.7, 80.6, 76.5, 75.0, 73.2, 72.7, 71.3, 70.6,
62.8, 61.9, 59.7, 56.2; HRMS calcd for C20H30O12Na
(M+Na): 485.1635, found: 485.1645.
4 - Methoxyphenyl 2 - O - propyl - ꢁ - D - galactopyranosyl-
(1!4)-ꢀ-D-galactopyranoside (16). Compound 16 was
prepared from 27 and allyl bromide as described for 15
in 81% yield. Compound 16 had: 1H NMR(CD 3OD): d
7.05 (m, 2H, Ar–H), 6.88 (m, 2H, Ar–H), 5.13 (d, 1H,
J=3.7 Hz, H-10), 4.78 (d, 1H, J=7.5 Hz, H-1), 4.19 (t,
1H, J=5.4 Hz, H-50), 4.01 (d, 1H, J=2.2 Hz, H-4),
3.93–3.58 (m, 15H), 1.65 (m, 2H, CH2CH3), 0.94 (t, 3H,
3.10 (dd, 1H, J=4.9, 8.4 Hz, H-60); 13C NMR(CDCl ):
3
d 159.4, 155.5, 152.1, 139.4, 139.2, 139.0, 139.0, 138.6,
138.5, 131.4, 130.1, 128.8, 128.8, 128.7, 128.6, 128.5,
128.3, 128.0, 128.0, 127.9, 127.9, 127.8, 127.8, 118.8,
114.9, 114.0, 103.5, 101.1, 81.3, 79.6, 79.0, 77.2, 76.7,
75.7, 75.4, 75.2, 75.1, 74.5, 73.9, 73.7, 73.6, 72.8, 72.7,
69.9, 68.8, 68.5, 56.1, 55.6; HRMS calcd for
C69H72O13Na (M+Na): 1131.4871, found: 1131.4860.
J=7.4 Hz, CH3); 13C NMR(CD OD): d 156.7, 153.1,
3
119.4, 115.4, 104.0, 101.2, 80.8, 79.0, 76.4, 75.0, 73.2,
72.6, 71.4, 70.6, 68.1, 62.8, 62.0, 56.0, 24.2, 10.7; HRMS
calcd for C22H34O12Na (M+Na): 513.1948, found:
513.1941.
4-Methoxyphenyl 3,4,6-tri-O-benzyl-ꢁ-D-galactopyranosyl-
(1!4)-2,3,6-tri-O-benzyl-ꢀ-D-galactopyranoside (27). Tri-
flouroacetic acid (0.36 mL) was added to a solution of
26 (365 mg, 0.355 mmol) in CH2Cl2 (18 mL) at 0 ꢁC and
the resulting solution was stirred for 40 min. n-Propyl-
acetate (4 mL) was added and the solution was con-
centrated and then co-concentrated with toluene. The
residue was purified by flash chromatography (SiO2, 3:1
4-Methoxyphenyl 2-O-methoxymethyl-ꢁ-D-galactopyra-
nosyl-(1!4)-ꢀ-D-galactopyranoside (17). Compound 17
was prepared from 27 and bromomethyl methyl ether as
1
described for 15 in 41% yield. Compound 17 had: H
NMR(CD OD): d 7.05 (m, 2H, Ar–H), 6.82 (m, 2H,
3
23
heptane–EtOAc) to give 27 (308 mg, 95%): ½aꢂD +53 (c
Ar–H), 5.09 (d, 1H, J=3.3 Hz, H-10), 4.82–4.74 (m, 3H,
H-1, CH2), 4.20 (t, 1H, J=6.5 Hz, H-50), 4.02 (d, 1H,
J=2.8 Hz, H-4), 3.95–3.68 (m, 12H), 3.60 (dd, 1H,
J=2.9, 10.1, H-3), 3.43 (s, 3H, CH2OCH3); 13C NMR
(CD3OD): d 156.7, 153.1, 119.4, 115.4, 104.1, 101.9,
98.5, 80.7, 76.9, 76.6, 74.9, 73.2, 72.6, 71.2, 70.3, 62.7,
61.9, 56.2, 56.0; HRMS calcd for C21H32O13Na
(M+Na): 515.1741, found: 515.1740.
1
1.0, CHCl3); H NMR(CDCl ): d 7.30–7.06 (m, 30H,
3
Ar–H), 6.88 (m, 2H, OPhOMe), 6.67 (m, 2H,
OPhOMe), 4.95 (d, 1H, J=3.9 Hz, H-10), 4.87–4.59 (m,
7H), 4.48–4.42 (m, 3H), 4.35–4.29 (m, 2H), 4.10–4.03
(m, 4H), 3.99 (m, 1H, H-40), 3.81 (t, 1H, J=8.4 Hz, H-
6), 3.71 (dd, J=7.6, 9.8 Hz, H-2), 3.66 (dd, J=2.6,
10.1 Hz, H-30), 3.21 (s, 3H, OMe), 3.55–3.46 (m, 3H, H-
5, H-6, H-60), 3.35 (dd, 1H, J=2.8, 9.9 Hz, H-3), 3.18
(dd, 1H, J=4.9, 8.4 Hz, H-6), 13C NMR(CDCl 3): d
155.3, 151.6, 138.9, 138.5, 138.5, 138.4, 138.4, 138.0,
137.6, 128.5, 128.5, 128.5, 128.5, 128.4, 128.4, 128.3,
128.2, 128.2, 128.1, 127.9, 127.7, 127.7, 127.7, 127.6,
127.5, 127.4, 127.0, 118.5, 114.6, 103.3, 100.5, 80.4, 79.3,
78.7, 75.3, 74.9, 74.1, 73.5, 73.4, 73.3, 72.8, 72.3, 72.1,
69.7, 68.1, 67.3, 55.6; HRMS calcd for C61H64O12Na
(M+Na): 1011.4295, found: 1011.4277.
4-Methoxyphenyl 3-O-[3-(S-2-acetamido-2-methoxy-
carbonyl-ethylthio)-propyl]-ꢁ-D-galactopyranosyl-(1!4)-
ꢀ-D-galactopyranoside (19). N-Acetyl-l-cystein methyl
ester (29 mg, 0.16 mmol) and AIBN (cat.) were added to
a solution of 2819 (30 mg, 32 mmol) in EtOAc (0.5 mL),
and Ar was bubbled through the mixture for 15 min.
The reaction flask was sealed with a septum and irra-
diated with UV light for 48 h using a water-cooled Ori-
ginal Hanau 70 W mercury high-pressure lamp.
Concentration and purification of the residue with flash
chromatography (SiO2, 2:1!1:2, heptane–EtOAc gra-
dient) gave protected 19 which was de-O-acylated in
methanolic NaOMe (10 mM, 1.5 mL) for 5 h. Methanolic
acetic acid (10%) was added until a neutral reaction on
moist pH paper was obtained. The resulting mixture
was concentrated and the residue was purified by flash
chromatography (SiO2, 10:1:0!4:1:0.2, CH2Cl2–MeOH–
H2O gradient) to give 21 (14 mg, 65%): 1H NMR
(CD3OD): d 7.07 (m, 2H, Ar–H), 6.85 (m, 2H, Ar–H),
5.01 (d, 1H, J=4.0 Hz, H-10), 4.82 (d, 1H, J=7.4 Hz, H-
1), 4.62 (dd, 1H, J=5.5, 8.0 Hz, CHNHAc), 4.31 (t, 1H,
J=6.3 Hz, H-50), 4.14 (d, 1H, J=2.1 Hz, H-40), 4.07 (d,
1H, J=2.9 Hz, H-4), 3.91–3.55 (m, 15H), 3.00 (ddd, 1H,
J=1.3, 5.2, 13.8 Hz, SCH2CHNHAc), 2.85 (ddd, 1H,
J=1.1, 8.1, 13.8 Hz, SCH2CHNHAc), 2.70 (m, 2H,
SCH2CH2CH2O), 1.99 (s, 3H, NHAc), 1.89 (m, 2H,
SCH2CH2CH2O); HRMS calcd for C28H43O15NSSiNa
(M+Na): 688.2251, found: 688.2246.
4-Methoxyphenyl 2-O-methyl-ꢁ -D -galactopyranosyl-
(1!4)-ꢀ-D-galactopyranoside (15). NaH (11 mg,
0.26 mmol, 60% in mineral oil) was added to a solution
of 27 (125 mg, 0.132 mmol) in DMF (2 mL). The result-
ing mixture was stirred at room temperature for 15 min
and was then cooled to 0 ꢁC. MeI (16 mL, 0.26 mmol)
was added and the resulting mixture was allowed to
obtain room temperature and was stirred overnight.
MeOH (1 mL) was added and the mixture was con-
centrated, co-concentrated with toluene, and the residue
was purified by flash chromatography (SiO2, 4:1 hep-
tane–EtOAc) to give 20-O-methylated galabioside. This
was dissolved in AcOH (2 mL) and hydrogenolyzed (H2,
1 bar, 10% Pd/C, 20 mg) for 2 h. The solution was fil-
tered through Celite, concentrated, and the residue was
purified by flash chromatography (CH2Cl2—MeOH–
H2O, 6:1:0.1!4:1:0.1) to give 15 (36 mg, 64%): 1H
NMR(CD OD): d 7.05 (m, 2H, Ar–H), 6.83 (m, 2H,
3
Ar–H), 5.17 (d, 1H, J=3.7 Hz, H-10), 4.77 (d, 1H,

A. Larsson et al. / Bioorg. Med. Chem. 11 (2003) 2255–2261
2. Foxman, B. Am. J. Public Health 1990, 80, 331.
2261
Surface plasmon resonance studies of binding of com-
pounds 1–20 to the class II PapG adhesin. The binding
of carbohydrates 1–20 to the class II PapG adhesin was
determined by surface plasmon resonance using a BIA-
CORE 3000 instrument. The adhesin was immobilized
on a sensor chip CM5 using a standard BIACORE
amine coupling procedure. This involved activation of
the carboxylic acid moieties of the dextrane surface of
the sensor chip with N-ethyl-N-(dimethylaminopropyl)-
carbodiimide hydrochloride (EDC) and N-hydroxy-
succinimide (NHS). The PapGII 196 truncate18 (1.2 mg/
mL in 20 mM MES buffer, pH 5.8) was then reacted
with the derivatized dextrane surface. Immobilization
levels of 4000–6000 RU were obtained. Derivatized
dextrane reacted with ethanolamine in one channel was
used as reference surface. Each of carbohydrates 1–20
were diluted in BIA certified HBS-EP running buffer
(10 mM HEPES pH 7.4, 0.15 M NaCl, 3 mM EDTA,
0.005% Surfactant P20) in concentration series ranging
from 0.001 to 10 mM. The compounds were injected
over the sensor chip (flow rate: 10 mL/min at 25 ꢁC) in
triplicates and binding to immobilized PapGII was
observed in real time. After injection of each compound
the surface of the sensor chip was regenerated by dis-
sociation in running buffer for 18 min. Dissociation
constants (Kd) were determined as the concentration of
a saccharide that elicited half of the maximal response,
that is, when 50% of the PapGII adhesin bound a
ligand.
3. Svanborg, C.; Godaly, G. Infect. Dis. Clin. N. Am. 1997,
11, 513.
4. Roberts, J. A.; Marklund, B.-I.; Ilver, D.; Haslam, D.;
Kaack, M. B.; Baskin, G.; Louis, M.; Mollby, R.; Winberg,
J.; Normark, S. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
11889.
5. Karlsson, K.-A. Curr. Opin. Struct. Biol. 1995, 5, 622.
6. Leffler, H.; Svanborg Eden, C. FEMS Microbiol. Lett.
´
1980, 8, 127.
7. Kallenius, G.; Mollby, R.; Svenson, S. B.; Winberg, J.;
Lundblad, A.; Svensson, S.; Cedergren, B. FEMS Microbiol.
Lett. 1980, 7, 297.
8. Bock, K.; Breimer, M. E.; Brignole, A.; Hansson, G. C.;
Karlsson, K.-A.; Larsson, G.; Leffler, H.; Samuelsson, B. E.;
´
Stromberg, N.; Svanborg-Eden, C.; Thurin, J. J. Biol. Chem.
1985, 260, 8545.
9. Striker, R.; Nilsson, U.; Stonecipher, A.; Magnusson, G.;
Hultgren, S. J. Mol. Microbiol. 1995, 16, 1021.
10. Nilsson, U.; Striker, R. T.; Hultgren, S. J.; Magnusson, G.
Bioorg. Med. Chem. 1996, 4, 1809.
11. Hultgren, S. J.; Abraham, S.; Caparon, M.; Falk, P.; St.
Geme, J. W., III; Normark, S. Cell 1993, 73, 887.
12. Johanson, I.-M.; Plos, K.; Marklund, B.-I.; Svanborg, C.
Microb. Path. 1993, 15, 121.
13. Otto, G.; Sandberg, T.; Marklund, B.-I.; Ulleryd, P.;
Svanborg, C. Clin. Infect. Dis. 1993, 17, 448.
14. Hansen, H. C.; Haataja, S.; Finne, J.; Magnusson, G.
J. Am. Chem. Soc. 1997, 119, 6974.
15. Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong,
G. D.; Ling, H.; Pannu, N. S.; Read, R. J.; Bundle, D. R.
Nature 2000, 403, 669.
16. Nilsson, U. J.; Fournier, E. J.-L.; Hindsgaul, O. Bioorg.
Med. Chem. 1998, 6, 1563.
Prediction of the binding of galabioside 11 to the class II
PapG adhesin. para-Methoxyphenyl galabioside 11 was
docked manually into the active site of the class II PapG
adhesin using Sybyl 6.3.32 Docking was performed by
superimposition of 11 on the galabiose moiety of the
tetrasaccaride GalNacb3Gala4Galb4GlcbOTMSET (3)
in the crystalline complex with PapGII.18 The complex
was energy minimized to relieve strain caused by addi-
tion of hydrogen atoms to the complex, and to allow 11
to assume a favorable conformation. This was per-
formed on an SGI R12000 Octane computer using the
batchmin program from Macromodel 6.5,33 together
with the Amber all-atom force field. The values for shell
and core were set to 10 and 8 A, respectively. The gra-
dient convergence was 0.1 and an iteration limit of 10
000 was used. No solvation model was employed.
17. Barkley, A.; Arya, P. Chem. Eur. J. 2001, 7, 555.
18. Dodson, K. W.; Pinkner, J. S.; Rose, T.; Magnusson, G.;
Hultgren, S. J.; Waksman, G. Cell 2001, 105, 733.
19. Ohlsson, J.; Jass, J.; Uhlin, B. E.; Kihlberg, J.; Nilsson,
U. J. ChemBioChem. 2002, 3, 772.
20. Nilsson, U.; Ray, A. K.; Magnusson, G. Carbohydr. Res.
1994, 252, 137.
21. Wilstermann, M.; Balogh, J.; Magnusson, G. J. Org.
Chem. 1997, 62, 3659.
22. Ohlsson, J.; Magnusson, G. Carbohydr. Res. 2000, 329,
49.
23. Konradsson, P.; Udodong, U. E.; Fraser-Reid, B. Tetra-
hedron Lett. 1990, 31, 4313.
24. Veeneman, G. H.; van Leeuwen, S. H.; van Boom, J. H.
Tetrahedron Lett. 1990, 31, 1331.
25. De Medeiros, E. F.; Herbert, J. M.; Taylor, R. J. K.
J. Chem. Soc., Perkin Trans. 1 1991, 2725.
26. Yan, L.; Kahne, D. Synlett 1995, 523.
27. Nilsson, U.; Ray, A. K.; Magnusson, G. Carbohydr. Res.
1994, 252, 117.
Acknowledgements
28. Kihlberg, J.; Hultgren, S. J.; Normark, S.; Magnusson, G.
J. Am. Chem. Soc. 1989, 111, 6364.
29. Jansson, K.; Ahlfors, S.; Frejd, T.; Kihlberg, J.; Magnus-
´
son, G.; Dahmen, J.; Noori, G.; Stenvall, K. J. Org. Chem.
1988, 53, 5629.
30. Garegg, P. J.; Oscarson, S. Carbohydr. Res. 1985, 137,
270.
31. Ohlsson, J.; Larsson, A.; Kihlberg, J.; Nilsson, U. J.
manuscript.
This work was funded by grants from the Swedish
Research Council, the Goran Gustafsson Foundation
for Research in Natural Sciences and Medicine, the
Kempe Foundation, and the program ‘Glycoconjugates
in Biological Systems’, sponsored by the Swedish
Foundation for Strategic Research.
32. SYBYL: Tripos Associates, Inc., 1699 S. Hanley Rd., St.
Louis, MO 63144, USA.
33. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Lis-
kamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson,
T.; Still, C. W. J. Comp. Chem. 1990, 11, 440.
References and Notes
1. Hooton, T. M.; Stamm, W. E. Infect. Dis. Clin. N. Am.
1997, 11, 551.