Squaric Acid Monoamide Mannosides as Ligands for the Bacterial Lectin FimH: Covalent Inhibition or Not?

Article abstract of DOI:10.1002/cbic.201000774

Bacteria use long proteinaceous appendages, called fimbriae or pili, to adhere to the surfaces of their host cells. Widely distributed among the Enterobacteriacae are type 1 fimbriae that mediate mannose-specific bacterial adhesion through the lectin FimH, located at the fimbrial tips. It is possible to design synthetic mannosides such that they show high affinity for FimH and can thus inhibit mannose-specific bacterial adhesion in a competitive manner. It has been found that mannosidic squaric acid monoamides serve especially well as inhibitors of type 1 fimbriae-mediated bacterial adhesion, but it has remained unclear whether this effect is due to specific inhibition of the bacterial lectin FimH or to unspecific bioconjugation between the lectin's carbohydrate binding site and a squaric acid monoamide. A bioconjugation reaction would result in a covalently crosslinked squaric acid diamide. Here it is shown that covalent inhibition of FimH by mannosidic squaric acid derivatives is very unlikely and that compounds of this type serve rather as excellent specific candidates for low-molecular-weight inhibitors of bacterial adhesion. This has been verified by testing the properties of glycosidic squaric acid monoamides in diamide formation, by two different adhesion assays with a series of selected control compounds, and by molecular docking studies that further support the results obtained in the bioassays. Candidates for specific antiadhesives: Squaric acid (SA) monoamides have the ability to crosslink unspecifically to amines. It has been shown that mannosidic SA monoamides serve as specific inhibitors of the bacterial lectin FimH and that no covalent bioconjugation within the lectin's carbohydrate binding site (CRD) occurs.

Full text of DOI:10.1002/cbic.201000774

DOI: 10.1002/cbic.201000774
Squaric Acid Monoamide Mannosides as Ligands for the
Bacterial Lectin FimH: Covalent Inhibition or Not?
Carsten Grabosch, Mirja Hartmann, Jçrn Schmidt-Lassen, and Thisbe K. Lindhorst*[a]
Bacteria use long proteinaceous appendages, called fimbriae
or pili, to adhere to the surfaces of their host cells. Widely dis-
tributed among the Enterobacteriacae are type 1 fimbriae that
mediate mannose-specific bacterial adhesion through the
lectin FimH, located at the fimbrial tips. It is possible to design
synthetic mannosides such that they show high affinity for
FimH and can thus inhibit mannose-specific bacterial adhesion
in a competitive manner. It has been found that mannosidic
squaric acid monoamides serve especially well as inhibitors of
type 1 fimbriae-mediated bacterial adhesion, but it has re-
mained unclear whether this effect is due to specific inhibition
of the bacterial lectin FimH or to unspecific bioconjugation be-
tween the lectin’s carbohydrate binding site and a squaric acid
monoamide. A bioconjugation reaction would result in a cova-
lently crosslinked squaric acid diamide. Here it is shown that
covalent inhibition of FimH by mannosidic squaric acid deriva-
tives is very unlikely and that compounds of this type serve
rather as excellent specific candidates for low-molecular-
weight inhibitors of bacterial adhesion. This has been verified
by testing the properties of glycosidic squaric acid mono-
amides in diamide formation, by two different adhesion assays
with a series of selected control compounds, and by molecular
docking studies that further support the results obtained in
the bioassays.
Introduction
Bacterial infections constitute a major global health problem,
especially threatening the health of young children.[1] The most
common serious neonatal infections involve bacteremia, men-
ingitis, and respiratory-tract infections.[2] Key pathogens in
these infections are Escherichia coli, Klebsiella spp., Staphylococ-
cus aureus, and Streptococcus pyogenes. It is therefore of great
interest to understand the mechanisms that facilitate pathoge-
nicity of bacteria: that is, carbohydrate-specific adhesion to
glycosylated surfaces (Figure 1).
To cause infection, bacteria usually need to adhere to their
target cells. For their attachment to cell surfaces and coloniza-
tion most bacteria depend on the expression of specialized ad-
hesive organelles, called fimbriae or pili.[3] Fimbriae are hair-like
(1–2 mm long and ~7 nm wide) protein structures produced on
the bacterial surface. Much studied examples include P fim-
briae and type 1 fimbriae that provide uropathogenic E. coli
(UPEC) with the ability to attach to specific niches in the urina-
ry tract.[4] There, a wide range of high-mannose type glycopro-
teins serve as receptor molecules for type 1 fimbriae. Specifici-
ty for terminal a-d-mannosyl residues is mediated by the fimb-
rial lectin FimH, which is located at the type 1 fimbrial tip, as
revealed by studies on fimbriae assembly.[5] FimH-mediated
bacterial adhesion can be prevented to a greater or lesser
extent by competitive mannosidic inhibitors, such as natural
oligosaccharides[6] or synthetic mannosides and cluster manno-
sides.[7] Studies on inhibition of bacterial adhesion have been
of importance for functional investigations[8] as well as for
applications in medicine, such as in the development of anti-
adhesives and antiadhesive surfaces.[9] For any application,
however, it is essential to distinguish specific inhibition from
unspecific inhibition, as well as competitive inhibition from ir-
Figure 1. Adhesion of E. coli to glycosylated surfaces: the drawing (not to
scale) shows type 1 fimbriae-mediated bacterial adhesion that is enabled by
the a-d-mannoside-specific lectin FimH located at the fimbrial tips. Adhesion
can be inhibited by addition of appropriate mannosides (specific binding).
With reactive mannosides it is possible that unspecific covalent binding to
the FimH CRD could also occur, leading to covalent inhibition.
[a] C. Grabosch,+ M. Hartmann,+ J. Schmidt-Lassen,+ Prof. Dr. T. K. Lindhorst
Otto Diels Institute of Organic Chemistry
Christiana Albertina University of Kiel
Otto-Hahn-Platz 3-4, 24098 Kiel (Germany)
Fax: (+49)431-880-7410
E-mail: tklind@oc.uni-kiel.de
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000774.
1066
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1066 – 1074

Squaric Acid Monoamide Mannosides
Figure 2. A) The bottom site of the FimH CRD with docked p-nitrophenyl a-d-mannoside (pNPMan), depicted in ball and stick. The CRD contains the N-termi-
nal Phe1 amino acid of the protein. Prominent hydrogen bonds between the carbohydrate binding site and the ligand are depicted as dashed lines. B) The
amino acid residues at the entrance of the FimH CRD comprise a “hydrophobic ridge”. The aromatic side chains of Tyr48 and Tyr137 enter into p–p interac-
tions with docked pNPMan. Both graphics show the amino acid residues oriented as in the crystallized protein (PDB ID: 1KLF).[11] Pictures were generated with
VMD[14] and rendered in Pov-Ray.[15]
reversible covalent blocking of the lectin’s carbohydrate bind-
ing site (Figure 1). The latter mechanism has parallels with af-
finity labeling of proteins.[10] In this account, the details of
lectin inhibition have been investigated with the aid of a par-
ticularly interesting example of ligand binding to the bacterial
lectin FimH.
fimbriae-mediated bacterial adhesion according to an ELISA
(enzyme-linked immunosorbent assay). This finding was inter-
FimH is a mannose-specific, two-domain adhesin with a
“lectin domain”, FimHL, comprising the mannose-binding
pocket, and a “pilin domain”, FimHP, that performs an organelle
integration function, needed for fimbriae assembly. The man-
nose-binding pocket (carbohydrate recognition domain, CRD)
of FimH perfectly accommodates one a-mannosidic glycon
moiety. Its bottom site consists mainly of hydrophilic amino
acid side chains: those of Asn46, Asp47, Asp54, Gln133,
Asn135, Asp140 and the N-terminal Phe1 (Figure 2A).[11] These
amino acid residues form a stabilizing hydrogen-bond network
with the glycon moiety of a complexed a-d-mannoside ligand.
The periphery of the FimH CRD, on the other hand, is char-
acterized by amino acids with rather lipophilic side chains,
defining a “hydrophobic ridge” at the entrance of the CRD
(Figure 2B).[11,12] The aromatic side chains of Tyr48 and Tyr137
form what has been called the “tyrosine gate” at the entrance
of the mannose binding pocket.[9a,11–13] Mannosides with an
aromatic aglycon, such as p-nitrophenyl a-d-mannoside
(pNPMan), can establish favorable p–p interactions with these
two residues, thus leading to significantly improved affinities
relative to methyl a-d-mannoside (MeMan).
Scheme 1. Squarates such as diethyl squarate (DES) form squaric acid (SA)
monoamides in a fast reaction and the corresponding SA diamides in a
much slower subsequent reaction step. This feature can be exploited for
conjugation of two amines R1ÀNH2 and R2ÀNH2. Accordingly, the mannosidic
SA monoamide 1 can be formed and might, in the second reaction step, co-
valently attach to the N terminus of the bacterial lectin FimH (Phe1) located
within the CRD to yield a crosslinked ligand–lectin conjugate.
We have recently published a systematic study on carbohy-
drate binding of type 1-fimbriated E. coli in which we utilized a
series of a-d-mannosidic squaric acid monoamides.[16] There, it
was shown that the squaric acid monoamide 1 (Scheme 1)
clearly exceeds pNPMan in potency as an inhibitor of type 1
ChemBioChem 2011, 12, 1066 – 1074
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1067

T. K. Lindhorst et al.
preted in terms of additional interactions of the extended agly-
con moiety in 1, relative to pNPMan, at the entrance of the
FimH CRD. It has been asked, however, whether the reason for
the high inhibitory potency of the squaric acid derivative 1
might be found in the formation of a covalent bond within the
FimH CRD (Scheme 1). This hypothesis is justified by the spe-
cial reactivity of squaric acid monoesters, which are frequently
exploited for bioconjugation.[17]
the same reaction was carried out in citric acid buffer at pH 4.5
no reaction occurred over 24 h.
This result indicates that the SA monoamide 1 can undergo
the predicted reaction to form the corresponding diamide 3
(Scheme 2), and so it had to be investigated whether the anal-
ogous reaction can also occur within the lectin’s CRD. A
number of control compounds were therefore synthesized and
tested in parallel with 1 as inhibitors of type 1 fimbriae-mediat-
ed bacterial adhesion (Scheme 3). As the most obvious control
Sequential treatment of a squarate such as diethyl squarate
(DES) with two amines results in their conjugation through
two vinylogous amides. In these reactions, the formation of
the first amide bond is much faster than the reaction between
the squaric acid (SA) monoamide formed in the first process
and a second amine. Nevertheless, once a mannosidic squaric
acid monoamide such as 1 is complexed within the FimH CRD,
its reaction with the N-terminal amino group of Phe1 might
lead to a crosslinked ligand–lectin SA diamide conjugate
(Scheme 1). It could be this crosslinking reaction that accounts
for the high inhibitory potency of the squaric acid monoamide
1.[16]
Results and Discussion
Synthesis of squaric acid derivatives
In order to test the ability of the SA monoamide 1 to form a
diamide under the conditions used for the bioassay, it was
treated with the l-phenylalanine ester 2 (Scheme 2), as a
model for the N-terminal Phe1 of FimH. When this reaction
was carried out in PBS buffer under physiological conditions
(pH 7.2), the squaric acid diamide 3 was formed in quantitative
yield. Conversion of the reaction partners was complete after
45 min and the reaction outcome did not change further for
the next 24 h according to MADLI-TOF-MS monitoring. When
Scheme 3. Control compounds for the bioassay: Squaric acid diamide 4
lacks the crosslinking ability of SA monoamide 1; glucosides 5 and 6 were
required to test carbohydrate specificity of ligand binding and squaric acid
derivatives DES, 7 and 8 are unspecific control compounds.
mannoside, the SA diamide 4, lacking the potential for cova-
lent crosslinking within the FimH CRD, was required. The man-
noside 4 was synthesized by treatment of 1 with ethylamine
under triethylamine catalysis. In addition, the glucosides 5 and
6 were prepared as analogues of the SA monoamide 1 and its
diamide 4, respectively, in order to test the carbohydrate spe-
cificity of the inhibition. The synthesis of 5 and 6 started from
p-aminophenyl a-d-glucoside and proceeded analogously to
the preparation of 1[16] and 4. Finally, diethyl squarate (DES)
and its known monoamide and diamide derivatives, 7 and 8,
respectively,[18] were synthesized and included in the biological
study as unspecific control compounds.
Bioassay: Inhibition of bacterial adhesion
All of the squaric acid derivatives under investigation
(Scheme 3) were tested as inhibitors of type 1 fimbriae-mediat-
ed bacterial adhesion to mannan-coated 96-well microtiter
plates by using uropathogenic Escherichia coli cells. Serial dilu-
tions of each inhibitor were incubated with fluorescent E. coli
cells.[19] If possible, inhibition curves were determined for each
tested inhibitor, from which IC50 values were deduced. The IC50
values in this case reflect the inhibitor concentrations that
Scheme 2. Treatment of the squaric acid monoamide 1 with the phenylala-
nine derivative 2 under bioassay conditions led to quantitative formation of
the SA diamide 3. a) PBS buffer (pH 7.2), 378C, 45 min–24 h, quant. (no reac-
tion at pH 4.5).
1068
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1066 – 1074

Squaric Acid Monoamide Mannosides
cause 50% inhibition of bacterial binding to mannan. On each
individual test plate pNPMan was tested in parallel to allow ref-
erencing of the IC50 value obtained for each tested ligand to
the IC50 of pNPMan. This procedure leads to relative inhibitory
potencies (RIPs) for every tested compound, which are consis-
tently referenced and can therefore be compared even if they
were not examined in the same experiment.
Computer docking with FimH
In order to deepen our understanding of binding of the differ-
ent squaric acid derivatives to FimH, docking studies with the
FimH lectin domain were performed. Carbohydrate ligands
were docked into the CRD of FimH by using FlexX[20] flexible
docking and consensus scoring[21,22] as implemented in
Sybyl6.9.[23] During the docking process the FimH CRD was
held fixed, whereas the ligand was allowed to change its con-
formation. To allow for false-positive solutions a two-stage
strategy was employed. The conformations delivered by FlexX
were regarded as “unrelaxed”. These conformations were mini-
mized to obtain “relaxed” conformations and ranked.[24] All sol-
utions were then docked and screened for the most reasona-
ble structures. For each docked conformation a FlexX scoring
value was obtained, which is correlated to its free binding
energy, a more negative score correlating with higher binding
affinity. For validation of the obtained results, consensus scor-
ing was employed. In order to address the conformational flex-
ibility of the tyrosine gate at the entrance of the FimH CRD,
docking was based on two different FimH X-ray structures,
reflecting two extreme relative conformations of Tyr48 and
Tyr137: the tyrosine gate of FimH in complexation with man-
nose was found to be in an “open” conformation,[11] whereas a
“closed-gate” structure was seen in complexation with n-butyl
a-d-mannoside.[13]
This bioassay revealed no inhibitory potency for the gluco-
sides 5 and 6, or for the squaric acid derivatives DES, 7, and 8.
From this finding it can be concluded that the presence of a
squaric acid moiety is not sufficient for preventing adhesion of
bacteria, even though compounds such as 5, DES, and 7 pos-
sess the capacity to crosslink to amines. The mannosides 1 and
4, on the other hand, showed good inhibitory potency, as
expected (Table 1). Interestingly, though, the squaric acid dia-
Table 1. Inhibition of bacterial adhesion to a mannoside-coated surface:
IC50 values for the squaric acid-functionalized mannosides 1 and 4 were
compared to the reference mannoside pNPMan in three independent
bacterial adhesion assays.[a][19]
1
4
pNPMan
IC50 [mm]
RIP[b]
17.3Æ6.5
16Æ1.5
6.38Æ3.7
50Æ19
274Æ110
1
[a] All substances were tested as duplicates in all assays. [b] Relative in-
hibitory potencies are based on the IC50 value for pNPMan measured on
the same microtiter plate.
Docking of mannosides 1 and 4 into the FimH CRD showed
that the sugar glycon portion is always complexed within the
binding pocket, with the aglycon moiety sticking out of the
binding site (Figure 3). To enable crosslinking of the SA monoa-
mide ligand 1 to the N-terminal end of FimH (Phe1), “upside-
down” complexation of the mannoside would be required.
This is not seen in the docking studies. Instead, computer
modeling reveals favorable hydrophobic interactions estab-
lished by the aglycon moiety of 1 at the entrance of the FimH
CRD (Figure 3A and B) as reported earlier.[16] Complexation of 1
within the CRD leads to typical scoring values of À29.6 to
À33.1 depending on the X-ray structure used for the docking.
mide 4, without any crosslinking capability, exceeds the inhibi-
tory potency of the corresponding SA monoamide 1. Whereas
the mannoside 1 showed an inhibitory potency approximately
16 times higher than that of pNPMan, 4 works 50 times better.
This result makes the initial hypothesis—that 1 performs espe-
cially well as an inhibitor of type 1 fimbriae-mediated bacterial
adhesion because it undergoes crosslinking within the CRD of
the bacterial lectin, thus irreversibly blocking the carbohydrate
binding site—rather unlikely.
Figure 3. Representative conformations of the docked SA monoamide 1 and SA diamide 4. A) Fit of the top scoring conformation of the SA monoamide 1
with a FlexX score of À33.1; B) docking solution No. 8 for the SA monoamide 1 with a FlexX score of À32.8; C) fit of the top scoring conformation of the SA
diamide 4 with a FlexX score of À35.3; D) docking solution No. 5 for the SA diamide 4 with a FlexX score of À34.3. The SA diamide 4 can establish hydrogen
bonds (indicated in yellow) to the amino acid side chains of Asp47 (in C) or Tyr48 (in D), which are not seen in the case of 1. The depicted docking results are
based on the closed-gate crystal structure of FimH.[13]
ChemBioChem 2011, 12, 1066 – 1074
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1069

T. K. Lindhorst et al.
Analogous docking of the squaric acid diamide 4 led to scoring
values of À30.9 for the open-gate structure of FimH[11] (see the
Supporting Information) and up to À35.3 for the closed-gate
structure[13] (Table 2). The docking results thus predict a higher
Table 2. FlexX scoring values for seven different ligands obtained from
docking based on two different crystal structures.
Ligand
Scores (hit no.)[a] based on
open-gate structure[11]
closed-gate structure[13]
1
4
5
6
DES
7
À29.6 (1)
À30.9 (1)
À15.1 (3)
À12.8 (4)
À10.1 (1)
À17.5 (1)
À18.5 (1)
À33.1 (1)
À35.3 (1)
[b]
[c]
[c]
À17.1 (3)
À21.2 (1)
8
[a] In each case scoring values are listed for the most reasonable ligand
conformation. [b] Only 7 of 30 hits are complexed within the CRD, so no
representative conformation can be deduced. [c] None of 30 hits is com-
plexed within the CRD.
affinity to FimH for 4 than for 1 and this is in accordance with
the experimental findings (Table 1). The observed higher affini-
ty of the SA diamide 4 relative to the SA monoamide 1 can be
explained by the formation of additional hydrogen bonds at
the entrance of the CRD, especially when the closed-gate
structure of FimH is used for docking. Hydrogen-bridging of 4
is due to the NH-ethyl hydrogen bond donor function, which
is absent in case of the SA monoamide 1. As hydrogen bond
acceptors the side chains of Asp47 (Figure 3C) and also of
Tyr48 (Figure 3D) can be identified.
Figure 4. Comparison of “upside-down” complexation of the glucosides 5
and 6 within the FimH CRD. A) Representative conformations (here hit No. 3
with a FlexX score of À15.1) of the SA monoamide 5 are rather extended;
B) representative conformations (here hit No. 4 with a FlexX score of À12.8)
of the SA diamide 6 are more likely to be bent. The depicted docking results
are based on the open-gate crystal structure of FimH.[11]
predicted with scoring values of À17.5 (open-gate) and À17.1
(closed-gate structure of FimH). For the SA diamide 8, scores of
À18.5 for the open-gate and À21.2 for the closed-gate struc-
ture were obtained, most likely resulting from further hydro-
gen bond interactions with the second hydroxyethyl moiety
(see the Supporting Information). Even docking of DES to the
open-gate structure of FimH revealed hits, albeit with a very
poor FlexX score of À10.1 (Table 2).
Docking studies with glucosides 5 and 6, on the other hand,
suggested some “upside-down” complexation, in which the
squaric acid aglycon moiety is complexed within the mannose
binding pocket of the open-gate structure (17 hits out of 30 in
the case of 5 and 14 hits out of 30 in case of 6). Docking
based on the closed-gate structure of FimH hardly showed any
complexation of these two glucosides (Table 2). For complexa-
tion of the SA monoamide 5, docking predicts an extended
conformation for the complexed ligand, whereas for complexa-
tion of 6 a more bent binding mode is indicated (Figure 4),
possibly due to a hydrogen bond between the NH-ethyl donor
function in 6 and the side chain of Asp140 (not shown), which
is not possible in the case of the glucoside 5. The scoring
values for all conformations complexed in an upside-down
mode, however, are extremely poor, with values between
À15.3 (5, hit No. 3 and 16 other solutions) and À12.8 (6, hit
No. 4 and 13 other solutions). FlexX docking with the low-affin-
ity ligand MeMan, for example, gives scoring values of around
À23.
Preincubation of bacteria with ligands
The obtained computer-aided docking results are in good
accordance with the experimental testing results. However,
some—though small—affinity for FimH was even predicted for
the glucosides 5 and 6, as well as for 7, 8, and DES, a finding
not reflected in the inhibition adhesion assay employed so far.
This might be due to the fact that in this assay inhibitors have
to compete against a highly mannosylated surface presenting
a high concentration of specific ligands for the fimbrial lectin
FimH. To test crosslinking ability in a noncompetitive setup,
serial dilutions of all three classes of ligands—the manno-
sides 1 and pNPMan, the glucosides 5 and pNPGlc, and the
non-glycosidic DES—were preincubated with type 1-fimbriated
E. coli bacteria. After 1 h of incubation under physiological con-
ditions, the mixtures were transferred into mannan-coated test
plates and the preincubated bacteria were allowed to bind to
the mannosylated surface. In this case, bacterial binding to the
It can thus be concluded that glucosides 5 and 6 do not
serve as ideally specific ligands for FimH. Likewise, docking
studies with the SA derivatives DES, 7, and 8 suggested ex-
tremely weak affinities for FimH in all cases (Table 2). However,
for the SA monoamide 7, possessing a hydroxyethyl moiety,
hydrogen-bond-supported complexation within the CRD was
1070
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1066 – 1074

Squaric Acid Monoamide Mannosides
Figure 5. Preincubation assay with GFP-expressing E. coli: after preincubation with the indicated inhibitors at the given concentrations, the ligand/E. coli mix-
tures were transferred to mannan-coated plates, to allow adhesion to the mannan surface. Non-adhered bacteria were washed away and adhered bacteria
were detected by fluorescence. Lower fluorescence intensity thus reflects more effective inhibition of bacterial adhesion to the microtiter surface due to pre-
incubation with the tested ligand. Right bar (0 control): bacterial adhesion without added inhibitor.
mannan surface is only possible if the fimbrial carbohydrate
binding sites are not ligand-saturated.
extended aglycon moiety that enhances affinity of this manno-
side to the lectin. This is seen by comparison of the testing
results obtained with 1 and pNPMan, on one hand, and the SA
monoamide 5 and pNPGlc on the other.
This assay showed that when ligand concentrations between
1.25 mm and 5 mm were applied for preincubation of bacteria,
concentration-dependent inhibition of bacterial adhesion was
seen in the cases of the mannosides 1, pNPMan, and to some
extent with the glucoside 5 (Figure 5). The mannosidic SA
monoamide 1 prohibits the attachment of bacteria very effec-
tively at all concentrations applied, whereas inhibition with
pNPMan follows a typical sigmoidal concentration dependence
curve (see the Supporting Information). With the glucoside 5,
low, but also concentration-dependent inhibition was detect-
ed. In contrast, a concentration gradient of the reference glu-
coside pNPGlc or of DES had no effect on the extent of bacteri-
al adhesion to mannan (Figure 5).
Conclusions
The mannosidic SA monoamide 1 can covalently crosslink to
amines such as the phenylalanine ester 2, leading to the corre-
sponding SA diamide. Such a crosslinking reaction might also
occur with the N terminus (Phe1) of the bacterial lectin FimH,
located within the lectin’s mannose binding site (CRD). Accord-
ing to the results presented here, however, it can be conclud-
ed that the relatively high affinity of the SA monoamide 1 for
the bacterial lectin FimH[16] must be due to its extended agly-
con moiety and not to covalent crosslinking within the bacteri-
al binding site.
Because the applied concentrations of the SA monoamides 1
and 5 are high enough to saturate all lectin binding sites of
the employed type 1-fimbriated bacteria, the observed concen-
tration dependency of inhibition once again proves that no co-
valent blocking of the FimH binding site occurred. Even DES
showed no inhibitory effect, although it should fit into the
FimH carbohydrate binding site. It can thus be concluded that
neither DES nor the SA monoamides 1 and 5 lead to covalent
binding within the FimH CRD. Probably protonation of the N-
terminal Phe1 amino group prevents crosslinking in these
cases. Our initial finding that the mannosidic SA monoamide 1
serves as an especially potent inhibitor of type 1 fimbriae-
mediated bacterial adhesion[16] can in fact be attributed to the
This conclusion was confirmed by an adhesion inhibition
assay in which the affinity of the SA diamide 4 towards FimH
was found to exceed that of its monoamide analogue 1. The
SA diamide 4 reproducibly showed an inhibitory potency three
times greater than that of the SA monoamide 1 even though it
lacks the potential to form a covalent bond to a free amino
function. A covalently crosslinked ligand–inhibitor complex
after incubation with 1 is hence very unlikely. Furthermore, co-
valent crosslinking of 1 within the FimH CRD could only occur
if the mannoside were complexed in an “upside-down” mode,
an option that is not supported by the performed docking
ChemBioChem 2011, 12, 1066 – 1074
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1071

T. K. Lindhorst et al.
0.063 mm). NMR spectra were recorded with 500 or 600 MHz
Bruker DRX 500 or AV 600 instruments. Chemical shifts (d) are cali-
brated relative to internal solvent. Full assignment was achieved
with 2D NMR techniques (1H,1H COSY and 1H,13C HSQC). ESI-MS
measurements were performed with a Mariner instrument, MALDI-
TOF mass spectra were recorded with a Bruker Biflex III instrument
with 19 kV acceleration voltage and an ionization laser at 337 nm.
As matrices, 2,5-dihydroxybenzoic acid and a-cyano-4-hydroxycin-
namic acid were used. For measurement of optical rotations a
Perkin–Elmer 241 polarimeter was used (10 cm cells, Na D-line:
589 nm). Purities of employed products were checked by HPLC.
Analytical HPLC was performed with a Merck–Hitachi LaChrom in-
strument with D-7000 interface and L-7455 diode array detector
and a LiChrosorb RP-8 silica column. Preparative HPLC was per-
formed with a Shimadzu system, an SPD-M10A diode array detec-
tor, and a Merck Hibar RT250–25 mm column with LiChrosorb RP-8
silica. (For HPLC chromatograms see the Supporting Information.)
For bacterial adhesion studies, a TECAN infinite 200 multifunction
microplate reader was employed.
studies. Modeling instead provides an explanation for the
higher affinity of the SA diamide 4 in relation to the monoa-
mide 1. For the glucosides 5 and 6, on the other hand, molecu-
lar docking studies suggested some probability for “upside-
down” complexation of these ligands into the FimH CRD.
Therefore, an assay was performed in which type 1-fimbriated
E. coli were preincubated with different ligands prior to adhe-
sion. If covalent crosslinking were to occur within the FimH
CRD, no bacterial adhesion should be possible after preincuba-
tion with ligands at appropriate concentrations. This was not
found; rather, concentration-dependent inhibition of bacterial
adhesion was seen in the case of the mannosides 1 and
pNPMan and to some extent with the glucoside 5. These re-
sults once again support the affinity-promoting properties of
the aglycon moiety in 1 and 5 and at the same time confirm
that no covalent crosslinking occurs within the FimH CRD with
SA monoamides. Interestingly, the strict mannoside-specificity
of FimH-mediated bacterial adhesion might be put into per-
spective under the conditions of the preincubation assay. Here,
the a-d-glucoside 5 had an effect on bacterial adhesion, lower
than that of the analogous mannoside 1, but in a concentra-
tion-dependent manner.
See the Supporting Information for additional procedures and for
supplementary analytical and graphical material.
p-[N-(4-Ethylamino-2,3-dioxocyclobut-1-enyl)amino]phenyl a-d-
mannopyranoside (4): A methanolic solution of ethylamine (2.0m,
240 mL, 480 mmol) and NEt3 (134 mL, 960 mmol) were added to a so-
lution of the monoamide 1 (95.0 mg, 240 mmol) in dry MeOH
(10 mL). The reaction mixture was stirred at room temperature for
12 h, followed by neutralization with Amberlite IR120 ion-exchange
resin, filtration, and concentration in vacuo. The crude product was
purified by silica gel chromatography (MeOH/AcOEt 1:1) to provide
the diamide 4 (60 mg, 63%) as a colorless lyophilizate. [a]D =+100
Finally, crosslinking of SA monoamides to the N-terminal
Phe1 residue of FimH is completely unlikely if the N terminus
is protonated. In this case, no diamide formation can occur (cf.
Scheme 2).
In summary, it has been shown that the high inhibitory po-
tency of 1 is not the result of covalent linkage of the SA mono-
amide to the N terminus of the FimH CRD but is due to the
specific structure of this synthetic mannoside. The affinity for
FimH found with 1 is further enhanced in its SA diamide ana-
logue 4. Mannosides such as 1 and 4 and similar derivatives
that can be obtained from them[16] thus constitute promising
candidates for a new class of low-molecular-weight antiadhe-
sives for type 1-fimbriated bacteria, exceeding the inhibitory
potencies of many other mannosides, in particular those of
longer-chain alkyl mannosides.[25] The squaric acid inhibitor 4
was shown to perform ~50 times better than pNPMan, where-
as n-heptyl a-d-mannoside has been reported to be a ~1.6-
times better inhibitor.[25c] Competitive binding of mannosiolic
SA diamides to type 1-fimbriated E. coli will therefore be fur-
ther investigated in our laboratory.
1
(c=0.10 in DMSO); H NMR (500 MHz, [D6]DMSO, 298 K): d=9.70
(brs, 1H; NH), 7.75 (brs, 1H; NH), 7.35 (d, J=9.0 Hz, 2H; 2HaAr),
7.06 (d, J=9.0 Hz, 2H; 2HbAr), 5.28 (d, J1,2 =1.8 Hz, 1H; H-1), 4.91–
4.42 (m, 4H; 4OH), 3.82 (dd, J1,2 =1.8 Hz, J2,3 =3.3 Hz, 1H; H-2), 3.67
(dd, J2,3 =3.3 Hz, J3,4 =9.4 Hz, 1H; H-3), 3.63–3.59 (m, 3H; H-6a, SA-
NHCH2CH3), 3.51–3.46 (m, 2H; H-4, H-6b), 3.43 (ddd, J4,5 =9.6 Hz,
J
5,6a =1.8 Hz, J5,6b =5.8 Hz, 1H; H-5), 1.21 ppm (t, J=7.1 Hz, 2H; SA-
NHCH2CH3); 13C NMR (150 MHz, [D6]DMSO, 298 K): d=183.6, 180.2,
168.8, 163.6 (CSA), 152.2 (man-O-CAr), 133.6 (CAr-NHSA), 119.4 (CaAr),
117.9 (CbAr), 99.5 (C-1), 74.8 (C-5), 70.6 (C-3), 70.1 (C-2), 66.8 (C-4),
61.1 (C-6), 38.7 (CH2), 16.4 ppm (CH3); ESI MS: m/z calcd for
C18H22N2O8Na: 417.1268 [M+Na]+; found: 417.1246.
p-[N-(4-Ethoxy-2,3-dioxocyclobut-1-enyl)amino]phenyl a-d-glu-
copyranoside (5): p-Aminophenyl a-d-glucopyranoside[26] (400 mg,
1.48 mmol) was dissolved in dry MeOH (20 mL), DES (432 mL,
2.95 mmol) was added, and the reaction mixture was stirred at
room temperature for 12 h. The solvent was then removed under
reduced pressure and the resulting syrup was subjected to purifi-
cation by column chromatography (MeOH/AcOEt 1:3) to provide
the title compound (308 mg, 65%) as a colorless lyophilizate.
[a]D =+126 (c=0.19 in DMSO); 1H NMR (600 MHz, [D6]DMSO,
300 K): d=10.67 (brs, 1H; NH), 7.29 (brs, 2H; 2HaAr), 7.08 (d, J=
8.9 Hz, 2H; 2HbAr), 5.33 (d, J1,2 =3.6 Hz, 1H; H-1), 4.77 (q, J=7.1 Hz,
2H; SA-OCH2CH3), 4.41–3.78 (m, 4H; 4OH), 3.63–3.56 (m, 2H; H-3,
Experimental Section
Reagents and methods: Commercially available starting materials
(phenylalanine tert-butyl ester from Fluka, DES and p-nitrophenyl
a-d-glucopyranoside from Aldrich) were used without further pu-
rification. p-Aminophenyl a-d-mannopyranoside and p-aminophen-
yl a-d-glucopyranoside were prepared by catalytic hydrogenation
of the corresponding p-nitrophenyl glycosides.[19,26] All solvents
used were purified by distillation. Methanol was dried over magne-
sium turnings with subsequent distillation. Monitoring of reactions
was performed by TLC on silica gel F254 (Merck) with detection by
UV light and/or by charring with ethanolic sulfuric acid (10%) or
ninhydrin solution [ninhydrin (300 mg) in butanol (100 mL) and
glacial acetic acid (3.00 mL)] and subsequent heating. Flash chro-
matography was performed on Merck silica gel 60 (0.040–
H-6a), 3.49–3.45 (m, 2H; H-5, H-6b), 3.36 (dd, J1,2 =3.6 Hz, J2,3
=
9.6 Hz, 1H; H-2), 3.18 (t, J3,4 =J4,5 =9.0 Hz, 1H; H-4), 1.40 ppm (t, J=
7.1 Hz, 3H; SA-OCH2CH3); 13C NMR (125 MHz, [D6]DMSO, 323 K):
d=187.9, 183.2, 177.7, 169.3 (CSA), 154.0 (man-O-CAr), 132.1 (CAr-
NHSA), 121.0 (CaAr), 117.5 (CbAr), 98.3 (C-1), 73.6 (C-5), 73.0 (C-3),
71.5 (C-2), 70.0 (C-4), 69.2 (CH2), 60.7 (C-6), 15.4 ppm (CH3); ESI MS:
m/z calcd for C18H21NO9Na: 418.1109 [M+Na]+; found: 418.1135.
1072
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1066 – 1074

Squaric Acid Monoamide Mannosides
p-[N-(4-Ethylamino-2,3-dioxocyclobut-1-enyl)amino]phenyl a-d-
glucopyranoside (6): A methanolic solution of ethylamine (2.0m,
709 mL, 354 mmol) and NEt3 (199 mL, 142 mmol) were added to a
solution of the SA monoamide 5 (140 mg, 345 mmol) in a mixture
of MeOH (10 mL) and DMSO (2 mL). After having been stirred at
room temperature for 12 h the reaction mixture was neutralized
with Amberlite IR120 ion-exchange resin and filtered, the solvent
was removed under reduced pressure, and the crude product was
purified by silica gel chromatography (MeOH/AcOEt 1:2) to provide
the title squaric acid diamide (132 mg, 95%) as a colorless lyophili-
for 1 h. Then, 100 mL of the preincubated mixture was transferred
into each well of a mannoside-coated, BSA-blocked 96-well micro-
titer plate and the plates were agitated (120 rpm) and incubated
for 1 h at 378C. After washing with PBS (3ꢁ150 mL), the wells were
filled with PBS (100 mL per well) and the fluorescence intensity
(485 nm/535 nm) was determined.
Docking: Two different crystal structures were used for docking
studies: FimH in complexation with a-d-mannose (PDB ID: 1KLF;
open-gate structure)[11] and FimH in complexation with n-butyl a-
d-mannoside (PDB ID: 1UWF; closed-gate structure).[13] For docking,
flexible FlexX 1.11.1L[20] as implemented in Sybyl6.9[23] was em-
ployed. The mannose binding pocket (CRD) was specified as a
sphere around the carboxyl C atom of Asp54 with a radius of 10 ꢂ.
This procedure results in a CRD consisting of 26 amino acids avail-
able for interactions with a docked ligand.
1
zate. [a]D =+155 (c=0.10 in DMSO); H NMR (600 MHz, [D6]DMSO,
298 K): d=7.34 (d, J=8.4 Hz, 2H; 2HaAr), 7.06 (d, J=8.4 Hz, 2H;
2HbAr), 5.29 (d, J1,2 =3.6 Hz, 1H; H-1), 3.63–3.55 (m, 4H; H-3, H-6a,
SA-NHCH2), 3.51–3.46 (m, 2H; H-5, H-6b), 3.36 (dd, J1,2 =3.6 Hz,
J
2,3 =9.7 Hz, 1H; H-2), 3.18 (t, J3,4 =J4,5 =9.2 Hz, 1H; H-4), 1.21 ppm
(t, J=6.9 Hz, 3H; SA-OCH2CH3); 13C NMR (150 MHz, [D6]DMSO,
298 K): d=183.4, 180.6, 168.9, 163.8 (CSA), 152.9 (man-O-CAr), 134.0
(CAr-NHSA), 119.5 (CaAr), 117.9 (CbAr), 98.6 (C-1), 73.6 (C-5), 73.0 (C-
3), 71.6 (C-2), 70.0 (C-4), 60.6 (C-6), 38.7 (CH2), 16.5 ppm (CH3);
MALDI-TOF MS: m/z calcd C18H22N2O8Na: 417.13 [M+Na]+; found:
417.79.
For every docking run 30 conformations (hits) were scored by
FlexX. Consensus scoring was used for validation of docked confor-
mations. Default parameters for formal charges and CScore calcula-
tions[21,22] were employed for consensus scoring, including the fol-
lowing scoring functions: FlexX original score, ChemScore,[27] Dock-
Score,[28] GoldScore,[29] PMFScore.[30] The conformational hits were
relaxed with the aid of the Tripos Force Field with default parame-
ters and scored.
Bioassay: Media and buffer solutions: LB medium (+AMP, +CAM):
tryptone (10.0 g), sodium chloride (10.0 g), and yeast extract
(5.00 g) were dissolved in doubly distilled water (1.00 L); after steri-
lization, ampicillin (100 mg) and chloramphenicol (50.0 mg) were
added. Carbonate buffer solution (pH 9.5): sodium carbonate
(1.59 g) and sodium hydrogen carbonate (2.52 g) were dissolved in
doubly distilled water (1.00 L) with subsequent pH adjustment. PBS
buffer solution (pH 7.2): sodium chloride (8.00 g), potassium chlo-
ride (200 mg), sodium hydrogen phosphate dihydrate (1.44 g), and
potassium dihydrogen phosphate (200 mg) were dissolved in
doubly distilled water (1.00 L). PBST buffer solution (pH 7.2): PBS
buffer+Tween(R) 20 (0.05%, v/v); pH values were adjusted with
HCl (0.1m) or NaOH (0.1m).
For graphical representation of docked ligand conformations (Fig-
ures 3 and 4), the protein FimH is depicted as a Connolly surface,
with a color scale reflecting the lipophilic potential of the surface
(brown correlates with high lipophilicity, blue with high hydrophi-
licity).
Acknowledgements
This work was supported by the Christiana Albertina University.
We are thankful to Thermo Fisher Scientific for Nunc microtiter
plates.
Liquid bacterial culture: E. coli of the strain PKL1162 were grown
in LB medium (+AMP, +CAM) overnight, washed with PBS buffer
and suspended to a concentration of 2 mgmLÀ1 in PBS buffer.
Mannan coating: Black 96-well plates (Nunc Maxisorp) were filled
Keywords: cell adhesion
·
FimH lectin
·
glycosides
·
with
a solution of mannan from Saccharomyces cerevisiae
(1.2 mgmLÀ1 in carbonate buffer, pH 9.5; 120 mL solution per well)
and allowed to dry in at 378C overnight. The plates were washed
with PBST (3ꢁ150 mL per well) and stored at 48C. Before use the
wells were blocked with BSA (5% in PBS, 150 mL per well) for 2 h
at 378C and then washed with PBST (3ꢁ150 mL per well).
inhibition · squaric acid conjugation
[1] E. K. Mulholland, R. A. Adegbola, N. Engl. J. Med. 2005, 352, 75–77.
[2] D. Osrin, S. Vergnano, A. Costello, Curr. Opin. Infect. Dis. 2004, 17, 217–
224.
Ligand solutions: All tested ligands were dissolved in PBS buffer.
The glucosides 5 and 6 and the SA derivatives DES, 7, and 8 were
tested at the highest possible concentrations with respect to their
solubility.
[3] a) P. Klemm, M. Schembri, Int. J. Med. Microbiol. 2000, 290, 27–35; b) K.
Ohlsen, T. A. Oelschlaeger, J. Hacker, A. S. Khan, Top. Curr. Chem. 2009,
288, 109–120.
[4] A. L. Kau, D. A. Hunstad, S. J. Hultgren, Curr. Opin. Microbiol. 2005, 8,
54–59.
[5] H. Remaut, C. Tang, N. S. Henderson, J. S. Pinkner, T. Wang, S. J. Hultg-
ren, D. G. Thanassi, G. Waksman, H. Li, Cell 2008, 133, 640–652.
[6] a) N. Firon, I. Ofek, N. Sharon, Carbohydr. Res. 1983, 120, 235–249; b) N.
Firon, I. Ofek, N. Sharon, Infect. Immun. 1984, 43, 1088–1090; c) J.-R.
Neeser, B. Koellreutter, P. Wuersch, Infect. Immun. 1986, 52, 428–436;
d) N. Sharon, FEBS Lett. 1987, 217, 145–157.
[7] a) M. Dubber, O. Sperling, T. K. Lindhorst, Org. Biomol. Chem. 2006, 4,
3901–3912; b) J. Bouckaert, J. Mackenzie, J. L. de Paz, B. Chipwaza, D.
Choudhury, A. Zavialov, K. Mannerstedt, J. Anderson, D. Pierard, L.
Wyns, P. H. Seeberger, S. Oscarson, H. De Greve, S. D. Knight, Mol. Micro-
biol. 2006, 61, 1556–1568; c) M. Lahmann, Top. Curr. Chem. 2009, 288,
17–65; d) B. Ernst, J. L. Magnani, Nat. Rev. Drug Discovery 2009, 8, 661–
677; e) Y. M. Chabre, R. Roy, Adv. Carbohydr. Chem. Biochem. 2010, 63,
165–393.
GFP-based bacterial adhesion assay: A serial dilution of the ex-
amined inhibitor was prepared and 50 mL was transferred into
each well of a mannoside-coated, BSA-blocked test plate. The bac-
terial suspension in PBS buffer (2 mgmLÀ1) was added (50 mL per
well) and the plates were agitated (120 rpm) and incubated for 1 h
at 378C. After washing with PBS (3ꢁ150 mL), the wells were filled
with PBS (100 mL per well) and the fluorescence intensity (485 nm/
535 nm) was determined.
GFP-based bacterial adhesion assay with preincubation: A serial
dilution of the examined inhibitor was prepared and mixed with
an equal volume of the bacterial suspension in PBS (2 mgmLÀ1).
These mixtures were incubated at 378C with agitation (120 rpm)
ChemBioChem 2011, 12, 1066 – 1074
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1073

T. K. Lindhorst et al.
[8] a) T. K. Lindhorst, K. Bruegge, A. Fuchs, O. Sperling, Beilstein J. Org.
Chem. 2010, 6, 801–809; b) T. K. Lindhorst, M. Mꢃrten, A. Fuchs, S. D.
Knight, Beilstein J. Org. Chem. 2010, 6, 810–822.
[9] a) S. D. Knight, J. Bouckaert, Top. Curr. Chem. 2009, 288, 67–107; b) R. J.
Pieters, Med. Res. Rev. 2007, 27, 796–816; c) I. Ofek, D. L. Hasty, N.
Sharon, FEMS Immunol. Med. Microbiol. 2003, 38, 181–191; d) K. E.
Sivick, H. L. T. Mobley, Infect. Immun. 2010, 78, 568–585; e) M. Durka, K.
Buffet, J. Lehl, M. Holler, J.-F. Nierengarten, J. Taganna, J. Bouckaert, S. P.
Vincent, Chem. Commun. 2011, 47, 1321–1323.
[20] a) M. Rarey, B. Kramer, T. Lengauer, G. Klebe, J. Mol. Biol. 1996, 261,
470–489; b) M. Rarey, B. Kramer, T. Lengauer, J. Comput. Aid. Mol. Des.
1997, 11, 369–384; c) B. Kramer, M. Rarey, T. Lengauer, Proteins Struct.
Funct. Genet. 1999, 37, 228–241; d) M. Rarey, S. Wefing, T. Lengauer, J.
Comput. Aid. Mol. Des. 1996, 10, 41–54.
[21] R. D. Clark, A. Strizhev, J. M. Leonard, J. F. Blake, J. B. Matthew, J. Mol.
Graphics Modell. 2002, 20, 281–295.
[22] P. S. Charifson, J. J. Corkery , M. A. Murcko, W. P. Walters, J. Med. Chem.
1999, 42, 5100–5109.
[10] a) Y. Takaoka, H. Tsutsumi, N. Kasagi, E. Nakata, I. Hamachi, J. Am. Chem.
Soc. 2006, 128, 3273–3280; b) J. H. Harvey, D. Trauner, ChemBioChem
2008, 9, 191–193.
[11] C. S. Hung, J. Bouckaert, D. Hung, J. Pinkner, C. Widberg, A. Defusco,
C. G. Auguste, R. Strouse, S. Langermann, G. Waksman, S. J. Hultgren,
Mol. Microbiol. 2002, 44, 903–918.
[12] a) D. Choudhury, A. Thompson, V. Stojanoff, S. Langermann, J. Pinkner,
S. J. Hultgren, S. D. Knight, Science 1999, 285, 1061–1066; b) A. Wellens,
C. Garofalo, H. Nguyen, N. Van Gerven, R. Slꢃttegꢄrd, J.-P. Hernalsteens,
L. Wyns, S. Oscarson, H. De Greve, S. Hultgren, J. Bouckaert, PloS ONE
2008, 3, 1–13.
[13] J. Bouckaert, J. Berglund, M. Schembri, E. De Genst, L. Cools, M. Wuhrer,
C. S. Hung, J. Pinkner, R. Slattegard, A. Zavialov, D. Choudhury, S. Lang-
ermann, S. J. Hultgren, L. Wyns, P. Klemm, S. Oscarson, S. D. Knight, H.
De Greve, Mol. Microbiol. 2005, 55, 441–455.
[23] Tripos, Inc., SYBYL 6.9, 1699 South Hanley Road, St. Louis, MO 63144–
2319.
[24] D. Hoffmann, B. Kramer, T. Washio, T. Steinmetzer, M. Rarey, T. Lengauer,
J. Med. Chem. 1999, 42, 4422–4433.
[25] a) Z. Han, J. S. Pinkner, B. Ford, R. Obermann, W. Nolan, S. A. Wildman,
D. Hobbs, T. Ellenberger, C. K. Cusumano, S. J. Hultgren, J. W. Janetka, J.
Med. Chem. 2010, 53, 4779–4792; b) T. Klein, D. Abgottspon, M. Witt-
wer, S. Rabbani, J. Herold, X. Jiang, S. Kleeb, C. Lꢅthi, M. Scharenberg, J.
Bezencon, E. Gubler, L. Pang, M. Smiesko, B. Cutting, O. Schwardt, B.
Ernst, J. Med. Chem. 2010, 53, 8627–8641; c) S. Rabbani, X. Jiang, O.
Schwardt, B. Ernst, Anal. Biochem. 2010, 407, 188–195.
[26] J. C. Briggs, A. H. Haines, Carbohydr. Res. 1996, 282, 293–298.
[27] M. D. Eldridge, C. W. Murray, T. R. Auton, G. V. Paolini, R. P. Mee, J.
Comput. Aid. Mol. Des. 1997, 11, 425–445.
[28] I. D. Kuntz, J. M. Blaney, S. J. Oatley, R. Langridge, T. E. Ferrin, J. Mol. Biol.
1982, 161, 269–288.
[29] a) G. Jones, P. Willett, R. C. Glen, A. R. Leach, R. Taylor, J. Mol. Biol. 1997,
267, 727–748; b) G. Jones, P. Willett, R. C. Glen, J. Mol. Biol. 1995, 245,
43–53.
[14] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graphics 1996, 14, 33–38.
[15] Persistence of Vision Pty. Ltd. (2004), persistence of vision raytracer (Ver-
sion 3.6), retrieved from http://www.povray.org/download.
[16] O. Sperling, A. Fuchs, T. K. Lindhorst, Org. Biomol. Chem. 2006, 4, 3913–
3922.
[30] I. Muegge, Y. C. Martin, J. Med. Chem. 1999, 42, 791–804.
[17] J. Kalia, R. T. Raines, Curr. Org. Chem. 2010, 14, 138–147.
[18] G. Maahs, P. Hegenberg, Angew. Chem. 1966, 78, 927–931; Angew.
Chem. Int. Ed. Engl. 1966, 5, 888–893.
Received: December 21, 2010
[19] M. Hartmann, A. K. Horst, P. Klemm, T. K. Lindhorst, Chem. Commun.
2010, 46, 330–332.
Published online on April 5, 2011
1074
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1066 – 1074