Replacement of water molecules in a phosphate binding site by furanoside-appended lin-benzoguanine ligands of tRNA-guanine transglycosylase (TGT)

Article abstract of DOI:10.1002/chem.201405764

The enzyme tRNA-guanine transglycosylase has been identified as a drug target for the foodborne illness shigellosis. A key challenge in structure-based design for this enzyme is the filling of the polar ribose-34 pocket. Herein, we describe a novel series of ligands consisting of furanoside-appended lin-benzoguanines. They were designed to replace a conserved water cluster and differ by the functional groups at C(2) and C(3) of the furanosyl moiety being either OH or OMe.The unfavorable desolvation of Asp102 and Asp280, which are located close to the ribose-34 pocket, had a significant impact on binding affinity. While the enzyme has tRNA as its natural substrate, X-ray co-crystal structures revealed that the furanosyl moieties of the ligands are not accommodated in the tRNA ribose-34 site, but at the location of the adjacent phosphate group. A remarkable similarity of the position of the oxygen atoms in these two structures suggests furanosides as a potential phosphate isoster.

Full text of DOI:10.1002/chem.201405764

DOI: 10.1002/chem.201405764
Full Paper
&
Molecular Recognition |Hot Paper|
Replacement of Water Molecules in a Phosphate Binding Site by
Furanoside-Appended lin-Benzoguanine Ligands of tRNA-Guanine
Transglycosylase (TGT)
Luzi J. Barandun,[a] Frederik R. Ehrmann,[b] Daniel Zimmerli,[c] Florian Immekus,[b]
Maude Giroud,[a] Claudio Grꢀnenfelder,[a] W. Bernd Schweizer,[a] Bruno Bernet,[a]
Michael Betz,[b] Andreas Heine,[b] Gerhard Klebe,*[b] and FranÅois Diederich*[a]
Dedicated to the 20th anniversary of Chemistry—A European Journal
Chem. Eur. J. 2015, 21, 126 – 135
126
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Abstract: The enzyme tRNA-guanine transglycosylase has
been identified as a drug target for the foodborne illness
shigellosis. A key challenge in structure-based design for this
enzyme is the filling of the polar ribose-34 pocket. Herein,
we describe a novel series of ligands consisting of furano-
side-appended lin-benzoguanines. They were designed to
replace a conserved water cluster and differ by the function-
al groups at C(2) and C(3) of the furanosyl moiety being
either OH or OMe. The unfavorable desolvation of Asp102
and Asp280, which are located close to the ribose-34
pocket, had a significant impact on binding affinity. While
the enzyme has tRNA as its natural substrate, X-ray co-crystal
structures revealed that the furanosyl moieties of the ligands
are not accommodated in the tRNA ribose-34 site, but at the
location of the adjacent phosphate group. A remarkable sim-
ilarity of the position of the oxygen atoms in these two
structures suggests furanosides as a potential phosphate
isoster.
that binds to the guanine/preQ1 binding site (Figure 1b,c).[12,13]
Derivatization at H2N-C(2) and/or C(4) allowed us to target the
ribose-33 pocket,[14,15] the ribose-34 pocket,[16,17] or both ribose
pockets simultaneously.[18] Furthermore, we have used ligands
based on this scaffold for the modulation of the protein–pro-
tein interaction in homo-dimeric TGT,[19] and studied the ther-
modynamic profile of ligand binding by isothermal titration
calorimetry (ITC).[20–22]
Introduction
Filling polar binding pockets of an enzyme by a ligand is a
difficult task in medicinal chemistry and structure-based drug
design.[1] In particular, phosphate binding sites,[2] which are
present in many classes of enzymes, such as kinases and
phosphatases,[3] are frequently targeted for improving potency
and/or selectivity of a drug lead compound. Thereby, the
understanding of replacing water molecules in the pocket of
the apoenzyme is a key challenge,[4] and several approaches
towards the consideration of water replacement in
structure-based drug design and molecular docking have been
developed.[5]
In the presence of a ligand in the preQ1/guanine binding
site, the uncomplexed polar ribose-34 pocket is solvated by
a highly conserved water cluster, which consists of five water
molecules (W1–W5; see X-ray crystal structure in Figure 1c and
Supporting Information Figure S1), with water molecule W6
connecting the cluster to the solvent-exposed ribose-33
pocket.[17] Displacing these water molecules, which solvate the
side chains of the aspartates Asp102 and Asp280, by an apolar
substituent (compound 2) led to an unfavorable desolvation of
the ribose-34 pocket, which was manifested in the weaker
binding affinity (Ki =235ꢀ50 nm) compared to the 4-unsubsti-
tuted analogue 1 (Ki =58ꢀ36 nm). Parts of this water cluster
were displaced by the polar substituents of ethanol 3 (Ki =
97ꢀ5 nm; PDB code: 3EOU[17]) or protonated amine 4 (Ki =
55ꢀ3 nm; PDB code 3GC5[17]), but without gaining binding af-
finity. Only by further expanding the substituent into an apolar
groove, shaped by Val45, Leu68, and Val282, was it possible to
improve the binding affinity (e.g. compound 5; Ki =4ꢀ2 nm;
PDB code: 3EOS[17]).[16] These findings were supported by mo-
lecular dynamics simulations, which identified nine water bind-
ing sites in the ribose-34 pocket, whereby only two of them
were found to be favorable for displacement by an apolar
ligand.[23] Although the lin-benzoguanine derivatives are very
potent ligands, their physicochemical properties are not opti-
mal—in particular their low solubility in both water and organ-
ic solvents is a problem for synthesis, biological assays, and po-
tential administration as a drug.[13] This prompted us to investi-
gate in the current study furanoside-based substituents for the
displacement of the water cluster in the ribose-34 pocket
(compounds 6a–c; Table 1). The hydroxy groups may replace
water molecules solvating the polar residues Asp102 and
Asp280, but do not introduce an additional charge into the in-
hibitors, unlike the ethylammonium linker in compounds 4
and 5. Since the furanosides are not linked by a N-glycosidic
bond to the lin-benzoguanine scaffold, they are expected to
be stable towards acidic or enzymatic depurination. Although
Herein, we present a new approach to the filling of a polar
binding pocket in the enzyme tRNA–guanine transglycosylase
(TGT; EC 2.4.2.29). It has been shown that TGT is essential in
the development of the pathogenicity of Shigella bacteria,[6,7]
which cause the severe inflammatory bowel disease shigellosis
and over one million lethal cases per year.[8,9] Bacterial TGT is
involved in the modification of tRNA and catalyzes the ex-
change of guanine by preQ1 (7-aminomethyl-7-deazagua-
nine).[10] Three major pockets are found in the active site of
Zymomonas mobilis TGT: the central guanine/preQ1 binding
site, where the base-exchange reaction takes place, and the
ribose-33 and ribose-34 pockets, where the tRNA backbone is
accommodated (Figure 1a).[11] We have introduced 2-amino-lin-
benzoguanines (such as 1; IUPAC: 6-amino-2-(methylamino)-
imidazo[4,5g]quinazolin-8(7H)-one) as central ligand scaffold
[a] Dr. L. J. Barandun,+ M. Giroud, C. Grꢀnenfelder, Dr. W. B. Schweizer,
Dr. B. Bernet, Prof. Dr. F. Diederich
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich
Vladimir-Prelog-Weg 3, HCI, 8093 Zꢀrich (Switzerland)
Fax: (+41)44-632-1109
E-mail: diederich@org.chem.ethz.ch
[b] F. R. Ehrmann,+ Dr. F. Immekus, M. Betz, Prof. Dr. A. Heine, Prof. Dr. G. Klebe
Institut fꢀr Pharmazeutische Chemie, Philipps-Universitꢁt Marburg
Marbacher Weg 6, 35032 Marburg (Germany)
Fax: (+49)6421-282-8994
E-mail: klebe@mailer.uni-marburg.de
[c] D. Zimmerli
F. Hoffmann-La Roche Ltd, Discovery Technologies, Bldg 92/5.64C
4070 Basel (Switzerland)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201405764.
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
127
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Binding affinities, clogD7.4, and clogP values.
[c]
R1
H
R2
Ki/nm[a]
Kd/nm[b]
clogD7.4
clogP[d]
b-6a
H
H
217ꢀ81
286ꢀ16
288ꢀ55 ꢁ1.5
276ꢀ69 ꢁ1.1
ꢁ1.3
ꢁ0.9
ꢁ1.2
a/b-6b 15:85 Me
b-6c
H
Me 353ꢀ106 n.d.[e]
ꢁ1.5
[a] Inhibition constant measured by a radioactive assay. Ki values are
mean values over at least four measurements. [b] Dissociation constant
measured by ITC (see Supporting Information Section S3 for details of
the assays). Kd values are mean values over at least five measurements.
[c] Logarithmic distribution coefficient for octanol/water at pH 7.4.
[d] Logarithmic partition coefficient for octanol/water. Both values were
calculated using ACD/Labs software.[29] [e] Not determined due to shape
of curve.
Results and Discussion
Synthesis
Iodobenzimidazole 7 was prepared according to our previously
published procedure (Scheme 1).[14,16] The furanosyl moieties
were introduced by Sonogashira cross-coupling reaction of 7
and alkynes 8a–c. The synthesis of the latter is described in
the Supporting Information (Section S2). The resulting alkyn-
ated benzimidazoles 9a–c were obtained in yields of 30–86%.
Subsequent hydrogenation of the alkynes to the alkanes 10a–
c (74–84% yield) required a large excess of Raney nickel. Grati-
fyingly, the benzyl protecting groups were stable under these
conditions. Cyclization of the benzimidazoles was achieved by
adapting a literature procedure, which used HgCl2 for the acti-
vation of thiourea 11.[25] After guanylation, treatment with
NaOMe cleaved one methoxycarbonyl group and triggered
cyclization to the lin-benzoguanines 12a–c (52–58% yield).
Hydrogenation of the benzyl ethers 12b and 12c needed
a large excess of Pd/C and long reaction times, but afforded
quantitatively the corresponding alcohols 13b and 13c. The
methoxycarbonyl group of 12a, 13b, and 13c was cleaved by
hydrolysis with KOH. The removal of the sulfamoyl group re-
quired strong acids. Under these conditions, the isopropyli-
dene group of 12a was cleaved and partial anomerization
took place. The resulting a/b-anomers of 6a–c were separated
by HPLC on a chiral stationary phase (Phenomenex Lux 5 mm
Cellulose-2 AXIA packed column) giving pure b-6a and b-6c,
and an enriched sample of a/b-6b (15:85).
Figure 1. a) Active site of Z. mobilis TGT with a preQ1–tRNA substrate (PDB
code: 1Q2S, 3.20 ꢂ resolution[11]). b) Structure and inhibition constants Ki of
1,[14] and 2–5.[16] c) X-ray crystal structure of Z. mobilis TGT soaked with
1 (1.49 ꢂ resolution, PDB code: 4PUK[21]). Color code: CtRNA yellow, Cligand
green, Cenzyme gray, O red, N blue, P orange, enzyme surface gray. Selected
water molecules are shown as red spheres and labeled as W. Hydrogen
bonds are shown as dashed lines (heavy atom distances between 2.6 and
3.6 ꢂ). These characteristics apply to all figures unless otherwise stated.
Conformational analysis
several groups have analyzed the role of water molecules in
carbohydrate–protein interactions,[24] furanosides and pyrano-
sides are rarely used in structure-based drug design.
The conformation of the furanose ring of benzimidazole
10b (Figure 2a) and of 2-O-methyl-b-d-ribofuranoside 14
(Figure 2b) in the solid state was determined by X-ray crystall-
ography (for details, see Supporting Information Section S2.5).
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
128
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. Synthesis and numbering of the furanoside-appended lin-benzoguanine ligands for TGT. In contrast to nucleotides, the furanoside has a higher pri-
ority in the numbering than the purine moiety. Reagents and conditions: a) [PdCl2(PPh3)2], CuI, DIPA, THF, 508C, 4.5–14 h, 9a: 86%, 9b: 42%, 9c: 30%;
b) Raney-Ni, H2, MeOH, 258C, 4.5–24 h, 10a: 83%, 10b: 74%, 10c: 84%; c) 11, HgCl2, Et3N, DMF, 508C, 3.5–16 h; then NaOMe, MeOH, 508C, 1.5–2 h, 12a:
53%, 12b: 58%, 12c: 52%; d) KOH, MeOH, 608C, 4–24 h; e) conc. HCl, THF, MeOH, 608C, 20–24 h, b-6a: 11% (from 12a), a/b-6b: 11% (from 12b), b-6c:
11% (from 12c); f) Pd/C (10%), H2, EtOH, 258C, 2–4 d. DMF=N,N-dimethylformamide, DIPA=N,N-diisopropyl-amine, THF=tetrahydrofuran.
Both furanosides crystallized in a northern conformation, but
showed different puckering states according to the Altona–
Sundaralingam model.[26] In 10b, the ribo-hexofuranoside
c adopt in (CD3)2SO a northern conformation as evidenced by
1,2 <1 Hz. The ethanediyl moiety of b-6a and b-6c adopts
a gt/tg equilibrium of ca. 1:1 indicated by J4,5a =J4,5b of about
6.5 Hz. Broad signals prevented the determination of the
rotameric equilibrium of a/b-6b.
J
1
moiety is present in a T2 twist conformation with a pseudoro-
tational phase P of 3328 and a puckering amplitude fm of 388.
Ribofuranoside 14 shows an E2 envelope conformation with
P=3398 and fm =378 (for a schematic overview of the ring
puckering, see Figure 2c,d). These conformations are in good
agreement with a CSD search by Taha et al. who found a prefer-
ence for b-ribofuranosides for P=313–3608 and fm =34–
438.[27] The conformations of all furanosides in solution was
studied by NMR spectroscopy: the ratio J1,2/J3,4 revealed the
preference of ring conformations (northern vs. southern con-
formation)[26] and the size of J4,5-pro-S and J4,5-pro-R evidenced the
preferred orientation of the substituent at C(5) (for details, see
Supporting Information Section S2.6).[28] The furanosides b-6a–
Binding affinities
Compounds 6a–c are readily water-soluble, which facilitated
the assays of their biological activities. The binding affinities of
the inhibitors were measured by a radioactive assay, giving Ki,
and by ITC, giving Kd (for a description of both assays, see Sec-
tion S3).[7,12,20,22] Both assays gave comparable Ki and Kd values
in the range of 217–353 nm (Table 1). We recently found for
other lin-benzoguanine-derived ligands that both Ki and Kd
values are in very good agreement, in contrast to the findings
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
129
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. X-ray co-crystal structures: data collection and refinement
statistics.
Crystal data
PDB code
b-6a
b-6b
b-6c
4LEQ
4LBU
4KWO
A) data collection and processing
collection site
No. crystal used
l [ꢂ]
BESSY 14.1
1
0.91841
BESSY14.2
1
0.85507
BESSY14.2
1
0.91841
space group
unit cell parameters
a [ꢂ]
b [ꢂ]
c [ꢂ]
C2
C2
C2
90.6
65.1
70.4
96.3
90.3
65.2
70.6
96.3
85.2
65.1
71.4
94.0
b [8]
B) diffraction data
resolution range [ꢂ]
50–1.41
50–1.17
50–1.32
(1.43–1.41)
(1.19–1.17)
(1.33–1.32)
unique reflections
R (I)sym [%]
completeness [%]
redundancy
I/s (I)
76331 (3734)[a] 136269 (6654)[a] 90020 (4448)[a]
[b]
6.5 (49.7)[a]
96.9 (94.7)[a]
3.8 (3.7)[a]
16.6 (2.8)[a]
2.4
4.3 (48.0)[a]
99.4 (97.6)[a]
3.1 (2.9)[a]
23.5 (2.1)[a]
2.4
4.1 (43.1)[a]
98.3 (96.9)[a]
4.2 (4.0)[a]
28.0 (3.0)[a]
2.3
Matthews
coefficient [ꢂ3Daꢁ1
Figure 2. a) ORTEP plot of 10b (CCDC code: 1016974). b) Structure and
ORTEP plot of 14 (CCDC code: 1016992). Schematic representation of the
ring puckering of c) 10b and d) 14. Atomic displacement parameters
obtained at 100 K are shown at the 50% probability level.
]
C) refinement
program used
resolution range [ꢂ]
reflns used
Phenix[32]
40.9–1.41
72493
Phenix[32]
14.9–1.17
129428
Phenix[32]
27.0–1.32
85496
final R values
[c]
for lin-benzohypoxanthines, lacking the exocyclic NH2 group,
for which the thermodynamic dissociation constants are much
lower than the inhibitory constants (for a detailed discussion,
see reference [22]). Compared to ethanol 3 (Figure 1b), the
furanoside-based inhibitors have weaker affinities by a factor
of about 2–3. While losing some affinity compared to the pre-
vious substituted ligands, the ribose derivatives 6a–c feature
the advantage of much enhanced water solubility, which is es-
sential for achieving solubility of the drug in blood. The hydro-
philicity of the inhibitors was estimated using ACD/Labs soft-
ware,[29] which was shown to predict clogP values with a root
mean squared error of 0.50 up to 1.28.[30] Although the predict-
ed values for the lin-benzoguanines are close to each other
within the error range, a plausible trend is observed comparing
their relative hydrophilicities. The ribose derivatives 6a–c had
a reduced hydrophilicity (clogD7.4 of ꢁ1.1 to ꢁ1.5) as com-
pared to the protonated ammonium compound derived from
4 (ꢁ2.5), but a higher hydrophilicity than the 4-unsubstituted
Rfree
Rwork
14.9 (19.7)
12.3 (16.8)
15.2 (22.9)
13.6 (21.8)
16.3 (25.0)
14.6 (19.9)
[d]
No. of atoms (non-hydrogen)
protein atoms
water molecules
ligand atoms
RMSD, angle [8]
RMSD, bond [ꢂ]
Ramachandran plot[e]
most favored
regions [%]
additionally allowed 5.0
regions [%]
generously allowed
regions [%]
mean B-factors [ꢂ2]
protein atoms
water molecules
ligand atoms
2984
428
28
1.1
0.006
2941
435
29
1.1
0.006
2847
291
29
1.1
0.006
94.7
93.4
6.3
95.3
4.4
0.3
0.3
0.3
14.1
30.5
11.8
14.1
30.1
13.4
17.1
30.9
15.3
[a] Values in parenthesis are statistics for the highest resolution shell.
[b] R(I)sym =[ꢀhꢀi jIi(h)ꢁhI(h)ij/ꢀhꢀiIi(h)]ꢃ100, in which hI(h)i is the mean
of the I(h) observation of reflection h. [c] Rwork =ꢀhkl jFoꢁFc j/ꢀhkl jFo j.
[d] Rfree was calculated as shown for Rwork, but on refinement-excluded 5%
of data. [e] Calculated with PROCHECK.[33]
analogue
1
(ꢁ0.3) or the cyclohexylmethylammonium
derivative from 5 (ꢁ0.2). Maintaining a low hydrophilicity of
a drug compound is crucial for membrane permeability and,
consequently, for its efficacy and good retention in the
body.[31]
only the b-anomer was bound suggesting a high preference
over the a-anomer. The structures of the enzyme complexes
with b-6a and b-6c are shown in Figure 3a,b, respectively,
while the structure with b-6b is depicted in Figure S2 in the
Supporting Information. In all three structures, the lin-benzo-
guanine core is accommodated in the guanine/preQ1 binding
site forming the same interactions as previously observed for
X-ray co-crystal structures
For all ligands b-6a, b-6b, and b-6c, highly resolved co-crystal
structures with Z. mobilis TGT (resolutions 1.17–1.41 ꢂ) were
obtained (Table 2).[32,33] Although a 15:85 anomeric mixture of
a/b-6a was applied for co-crystallization with Z. mobilis TGT,
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
130
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. X-ray co-crystal structure of Z. mobilis TGT in complex with a) b-6a (Cligand green, 1.41 ꢂ resolution, PDB code: 4 LEQ) and b) b-6c (Cligand salmon,
1.32 ꢂ resolution, PDB code: 4 KWO). c) Positions of the ethanediyl linkers in 5 (C yellow, PDB code: 3EOS[17]) and b-6a. A close unfavorable contact between
the linker and Asp102 is shown as a red dashed line. d) Comparison of the water clusters in the structure with b-6a (W1’, W2’, and W3’, shown as red spheres)
and 1 (W1–W6, shown as blue spheres; PDB code: 4PUK[21]). Distances are given in ꢂ.
1.[15,17–22] The ligands form an array of hydrogen bonds to the
side chains of Asp102, Asp156, and Gln203 and to the peptide
backbone of Gly230, Leu231, and Ala232 (Figure 3a,b,
Figure S2 in the Supporting Information).
The ethanediyl linker, which connects the furanosyl moiety
to the lin-benzoguanine core of b-6a–c, is oriented away from
the bottom of the ribose-34 pocket. This is in contrast to our
previous C(4) ethylamino-substituted lin-benzoguanines, in
which the ethanediyl linker is oriented towards the bottom of
the pocket allowing formation of a hydrogen-bond from the
ammonium center (compounds 4 and 5) to Asp280 (Figure 3c;
PDB code: 3EOS[17]). In addition, the orientation towards the
bottom of the pocket avoids repulsive interactions with
Asp102 (e.g. for 5, d(Cethyl···OAsp102)=3.9 ꢂ), whereas the orienta-
tion towards the top in b-6a brings the ethanediyl linker into
unfavorable proximity of Asp102 (d(Cethyl···OAsp102)=3.4 ꢂ). Fur-
thermore, this conformation has also an effect on the orienta-
tion of side chain of Tyr106 (not shown in Figure 3; see Fig-
ure S4), which together with the side chain of Met260 forms
a sandwich incorporating the tricyclic lin-benzoguanine core.
When the ethanediyl linker is oriented towards the top of the
pocket, Tyr106 is no longer parallel to the lin-benzoguanine
The co-crystal structure with b-6a (PDB code: 4 LEQ, Fig-
ure 3a) shows the furanosyl moiety in the ribose-34 pocket in
an E2 envelope conformation with P=3448 and fm =448. It un-
dergoes hydrogen bonding with Asn70, Gln107, and Asp280.
Only one hydrogen bond from HOꢁC(3) of the ribofuranosyl
unit is formed to Asp280. From the original five-water cluster
(Figure 1d), three water molecules (W3, W4, W5) are displaced
by b-6a, two water molecules (W1’ and W3’; red) are at a slight-
ly altered position (Figure 3d and Figure S3 in the Supporting
Information), and water molecule W2’, connecting the cluster
to the solvent-exposed ribose-33 pocket, remained at the
same position. The methoxy group at C(2) accepts a hydrogen
bond from W3’.
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
131
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
core, but tilted by 508 and undergoes a CꢁH···p-interaction
with the linker.
ribose-34 pocket by displacing three water molecules. Ethanol
3 and ammonium derivative 4 are better binders, as they leave
a better solvated Asp side chain (c.f. Figure S5 in the Support-
ing Information). The furanosyl moiety of b-6c cannot form
a hydrogen bond to the side chain of Asp280 due to the
methyl group at O-C(3).
The superimposition of the crystal structures of TGT in
complex with 5 (PDB code: 3EOS[17]) and b-6a shows that the
furanosyl moiety does not extend into the hydrophobic
patch formed by Val45, Leu68, and Val282 (Figure 3c). The
OꢁC(2) bond of b-6a, however, is ideally positioned to direct
a lipophilic ether substituent into this patch for additional gain
in binding affinity (compare 4 and 5, Figure 1b).
A second negative influence on the binding affinity of the
furanosides is the repulsive interaction with Asp102, which
comes into unfavorable proximity of the ethanediyl linker of
b-6a and a/b-6b. Inhibitor b-6c avoids this penalty, but the
furanose moiety has to adopt a ring conformation that is most
likely higher in energy.[27]
A similar situation is found for the X-ray co-crystal structure
with b-6b (Figure S2 in the Supporting Information; PDB code:
4 LBU). The enzyme and the ligand, in particular the furanosyl
moiety (E2 envelope conformation with P=3508 and fm =44),
adopt a similar geometry to that seen in the structure with b-
6a. The additional methyl group at OꢁC(2) points towards the
hydrophobic groove of the ribose-34 pocket, consisting of
Val45, Leu68, and Val282, confirming the opportunity for filling
this pocket with larger, more complementary lipophilic groups
and concomitant gain in binding affinity. The positions of the
two residual water molecules of the initial cluster also comply
to the structure with b-6a, although W2’ is not observed.
The X-ray co-crystal structure of Z. mobilis TGT with b-6c
(PDB code: 4 KWO, Figure 3b) shows a different orientation
and conformation of the furanosyl moiety than b-6a. As
a result of the additional methyl group at OꢁC(3), a hydrogen
bond to Asp280 is no longer possible. Hence, the furanose
ring moves away as compared to the complexes with b-6a
(Figure 3b) and b-6b (Figure S2 in the Supporting Informa-
tion), avoiding a clash between the highly polar Asp280 side
chain and the methoxy group. The furanose ring is no longer
in the usual E2 envelope conformation but in a less stable OT4
twist conformation with P=798 and fm =398. In this position,
C(1)ꢁOꢁC(4) forms a hydrogen bond to Gln107. In contrast to
the co-crystal structures with b-6a and b-6b, the ethanediyl
linker of b-6c is oriented towards the bottom of the pocket. In
this position, Tyr106 does not form a CꢁH···p interaction with
the linker, but is oriented parallel to the lin-benzoguanine core.
Furthermore, the downwards orientation of the ethanediyl
group avoids repulsive interactions with Asp102. The water
cluster in the co-crystal structure with b-6c has an altered pat-
tern and consists of four molecules. Water molecule W2’,
which corresponds to W6 in the original five-water cluster (Fig-
ure 1c) is not displaced by the ligand as in b-6a and b-6b. The
water network comprises W1’, which is also seen in the two
other co-crystal structures with b-6a and b-6b, in addition to
W4’ and W5’.
Although all three furanosides form additional interactions
to the enzyme—such as hydrogen bonds of HOꢁC(2) to Asn70
and from the cyclic ether O-atom to Gln107—the penalty of
desolvation of Asp102 and Asp280 is not compensated.
Phosphate mimic
The natural substrate of TGT is tRNA, the phosphate backbone
of which is recognized in the ribose pockets of the enzyme.
The overlay of the X-ray crystal structures of Z. mobilis TGT in
complex with a tRNA substrate and preQ1 introduced at the
wobble position 34 (PDB code: 1Q2S[11]) and in complex with
b-6a is shown in Figure 4. The pyrimidone moiety of the lin-
benzoguanine penetrates deeper into the guanine/preQ1 bind-
ing site than the same moiety of tRNA-bound preQ1 (about
0.8 ꢂ based on the pyrimidone core). A striking difference is
that the complexed tRNA does not form a hydrogen bond to
Asp280, which is involved in the catalytic cycle of the
enzyme.[11,34] The ethanediyl linker of the synthetic ligand
spans the ribose-34 ring of the tRNA to which preQ1 is at-
tached. This places the furanosyl moiety of b-6a into the posi-
tion of the phosphate-34 group of the tRNA. For a detailed
analysis of the two structures, it must be taken into consider-
ation that the resolution of the X-ray crystal structure with the
tRNA substrate is significantly lower (3.20 ꢂ)[11] compared to
the one with b-6a (1.41 ꢂ). In addition, the measured atomic
distances depend on the overlay of the structures (alignment
was based on the amino acid residues in the guanine/preQ1
binding site). Nevertheless, a qualitative comparison shows
that the oxygen atom of the anomeric methoxy group of b-6a
is at a similar position to O(d) of the phosphate group of the
tRNA ligand (d[C(1)ꢁOb-6a···O(d)tRNA]=1.2 ꢂ; see Figure 4b for
numbering of the atoms). The HOꢁC(2) moiety of b-6a is next
to O(g) (d[C(2)ꢁOb-6a···O(g)tRNA]=1.6 ꢂ), while O(a) is located in
between of HOꢁC(2) (d[C(2)ꢁOb-6a···O(a)tRNA]=1.5 ꢂ) and HOꢁ
C(3) (d[C(3)ꢁOb-6a···O(a)tRNA]=1.7 ꢂ). The ether oxygen C(4)ꢁO is
near O(b) of the phosphate group (d[C(4)ꢁOb-6a···O(b)tRNA]=
1.9 ꢂ). In addition, the water molecules W1’ and W2’ in the co-
crystal structure with b-6a are in proximity of the oxygen
atoms of the second phosphate group pointing into the
The structural insights from the three co-crystal structures
explain the weaker binding affinity compared to the 4-unsub-
stituted lin-benzoguanine 1 (Ki =58ꢀ36 nm) and inhibitor 5
(Ki =4ꢀ2 nm; Figure 1b). In the case of the uncomplexed
ribose-34 pocket in the co-crystal structure with 1, Asp280 is
optimally solvated by a water cluster. In the crystal structure in
complex with high-affinity ligand 5, Asp280 interacts with the
ammonium center in the linker by forming two charge-assisted
hydrogen bonds. In contrast, the furanosyl moieties of b-6a
and b-6b only form one hydrogen bond from HOꢁC(3) to the
anionic Asp280 side chain and reduce the solvation of the
ribose-33
pocket
(d[OW1’···O(g’)tRNA]=1.3 ꢂ
and
d[OW2’···O(d’)tRNA]=0.5 ꢂ). A similar situation is found for com-
pound b-6b (Figure S6a in the Supporting Information). In con-
trast, no similarity to the phosphate group is apparent for
ligand b-6c (Figure S6b in the Supporting Information). Given
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
132
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Conclusion
Herein, we describe a novel series of lin-benzoguanines that
target the polar ribose-34 pocket of Z. mobilis TGT with a fura-
nosyl moiety. The preparation involved a new cyclization strat-
egy for the lin-benzoguanine core and the highly challenging
separation of the anomeric mixtures by HPLC on a chiral sta-
tionary phase. While our previous lin-benzoguanine-based in-
hibitors suffered from poor water solubility, the furanosyl
moiety renders 6a–c freely water soluble. The new compounds
had Ki values in the range of 217–353 nm, which makes them
weaker inhibitors than ethanol 3 and ethylamine 4 by a factor
of 4–6. Based on the analysis of X-ray co-crystal structures, the
decrease in binding affinity is due to unfavorable desolvation
of Asp102 and Asp280. Although the furanosyl moiety estab-
lishes several additional interactions to the enzyme, the remov-
al of three of the five water molecules in the original five-
water cluster cannot be compensated. This cluster solvates the
ribose-34 pocket in the presence of ligand 1, which does not
penetrate into the pocket. However, the crystal structures also
point to the opportunity for regaining binding affinity through
extension of the ligands by etherification of HOꢁC(2) with resi-
dues complementary to the hydrophobic patch shaped by
Val45, Leu68, and Val282. The comparison of the binding affini-
ties of 4 and 5 shows that much affinity can be gained from
proper occupation of this hydrophobic subsite near the ribose-
34 pocket. The comparison of the X-ray co-crystal structures in
complex with the furanosyl lin-benzoguanines with a structure
with the tRNA-preQ1 substrate shows an intriguing finding.
The furanosyl moiety is not occupying the ribose-34-recogniz-
ing area of the enzyme active pocket, but the phosphate-34
binding site. Thereby, the positions of the oxygen atoms of the
furanoside resemble the positions of the oxygen atoms in the
phosphodiester. This gives rise to the hypothesis that furano-
sides might be potential surrogates for targeting phosphate
binding sites. The further decoration of the furanoside to prop-
erly fill the hydrophobic patch and reach the low nanomolar
activity range and the validation of furanosides and related
sugar derivatives as phosphate isosters are currently pursued
in our laboratories.
Figure 4. a) Comparison of the X-ray crystal structures of Z. mobilis TGT in
complex with a preQ1-tRNA substrate (C yellow, 3.20 ꢂ, PDB code: 1Q2S[11]
)
and with b-6a (PDB code: 4 LEQ, C green). b) Comparison of the positions
of the oxygen atoms of the phosphate group with the oxygen atoms of
furanosyl moiety.
the binding affinities of inhibitors b-6a and b-6b in the nano-
molar range, the overlays in Figure 4 and Figure S6a (Support-
ing Information) suggest that furanosides might be suitable
groups to target phosphate binding sites. The filling of these
polar pockets is a very challenging task in medicinal chemistry
and is often only achieved by charged functional groups that
impair membrane permeability.[35] In this work, we show the
potential of furanoside-based ligands to fill the phosphate
binding site in Z. mobilis TGT. Such substituents have no
charge, and the furanosyl moiety increases the water solubility
and should not suffer from fast metabolism as it is CꢁC linked
and not CꢁN linked as in natural nucleotides. Potent neutral
phosphate isosters are rare, and this has hampered in
particular the development of ligands for the active site of
phosphatases; therefore, our future work will focus on vali-
dating the hypothesis that furanosides, such as introduced
here, have more general phosphate isosteric character.
Experimental Section
All graphics of crystal structures were generated with the program
Pymol.[36] In the following, the experimental details for the synthe-
ses of compounds 6a–c are described. All other synthetic details
and experimental data, conformational analysis of the furanosides,
NMR spectra, and description of the radioactive assay, ITC, and
crystallization protocols are in the Supporting Information.
General procedure 1 (GP 1) for the cleavage of the methyl car-
bamates: A solution of the protected aminopyrimidinone (1 equiv)
and KOH (10 equiv) in MeOH was stirred at 608C for 4–24 h, neu-
tralized (pHꢂ7ꢀ1) by addition of 1m methanolic HCl solution,
and evaporated. The residue was suspended in EtOAc, filtered, and
evaporated. The crude product was used in the following step.
General procedure 2 (GP 2) for the cleavage of the N,N-di-
methylsulfamoates: A solution of the protected lin-benzoguanine
(1 equiv) in THF/MeOH/aq. conc. HCl solution 1:1:1 was stirred at
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
133
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
608C for 24 h. Under these conditions, partial anomerization took
place. Reaction control was only possible by analytical HPLC ([Nu-
cleosil-100 NH2, 250ꢃ4 mm, 5 mm]; H2O/MeCN 0:100 for 5 min,
0:100 to 40:60 within 25 min, flow rate 1 mLminꢁ1). The mixture
was basified (pH>7) with a solution of NaOMe in MeOH (20% w/
w) and evaporated. The anomers were separated by chiral HPLC
([Phenomenex Lux 5 mm Cellulose-2 Axia packed column, 250ꢃ
21.2 mm]; heptane/EtOH + 0.01% NH4OAc 30:70, isocratic for
60 min, flow rate 20 mLminꢁ1). The fractions of the b-d-anomer
were further purified by FC (MCI gel; H2O/MeCN 100:0 to 0:100) or
HPLC ([Nucleosil-100 NH2, 250ꢃ21 mm, 5 mm]; H2O/MeCN 0:100
for 5 min, 0:100 to 30:70 within 25 min, 30:70 for 10 min, flow rate
10 mLminꢁ1). Evaporation and lyophilization gave pure b-d-anom-
ers b-6a and b-6c and an enriched sample of a/b-6b (15:85).
tion, and lyophilization yielded a/b-6b 15:85 (9 mg, 11% from
12b) as a white foam. M.p. >1808C; H NMR (600 MHz, (CD3)2SO;
1
a/b 15:85): signals of b-6b: d=1.78–1.82 (m, 2H; H2C(5)), 2.89 (d,
J=4.9 Hz, 3H; NMe), 3.06–3.16 (m, 2H; H2C(6)), 3.27 (s, 3H; MeO-
C(2)), 3.38 (s, 3H; MeO-C(1)), 3.47 (d, J=4.3 Hz, 1H; H-C(2)), 3.69–
3.75 (m, 1H; H-C(3)), 3.95–4.00 (m, 1H; H-C(4)), 4.73 (brs, 1H; H-
C(1)), 5.77–5.85 (brs, 2H; NH2), 7.00 (brq, J=4.9 Hz, 1H; NH), 7.47
(s, 1H; H-C(9’)), 10.53–10.62 (brs, 1H; NH), 10.87–10.98 ppm (brs,
1H; NH); 13C NMR (150 MHz, (CD3)2SO; a/b 15:85): signals of b-6b:
d=21.49 (C(6)), 28.79 (NMe), 35.44 (C(5)), 53.99 (MeO-C(1)), 57.73
(MeO-C(2)), 75.03 (C(3)), 81.90 (C(2)), 84.12 (C(4)), 101.39 (C(4’)),
104.78 (C(1)), 109.51 (C(9’)), 119.32 (C(8’a)), 130.13 (C(3’a)), 143.61
(C(9’a)), 148.41 (C(6’)), 148.71 (C(4’a)), 158.50 (C(2’)), 162.76 ppm
(C(8’)); signals of a-6b: d=22.11 (C(6)), 28.89 (NMe), 34.68 (C(5)),
54.13 (MeO-C(1)), 74.26 (C(3)), 82.08 (C(2)), 105.02 (C(4’)), 107.46
(C(1)), 111.34 (C(9’)), 142.70 (C(9’a)), 157.71 (C(2’)), 165.85 ppm
(C(8’)); IR (ATR): n˜ =3600–2880 (br, w), 2830 (w), 2778 (w), 1690 (w),
1619 (m), 1584 (m), 1520 (w), 1435 (w), 1356 (m), 1186 (w), 1030
(s), 962 (m), 929 (m), 772 cmꢁ1 (m); HR-MALDI-MS: m/z: calcd (%)
for C18H25N6O5+: 405.1881; found: 405.1881 (100) [M+H]+.
General procedure 3 (GP 3) for the cleavage of the benzyl
ethers: A solution of the benzyl ether (1 equiv) in EtOH was treat-
ed with 10% Pd/C (200–300% w/w), stirred at 258C for 2–4 d
under H2 atmosphere (balloon), centrifuged, and decanted. The re-
sidual slurry was suspended in EtOAc, ultrasonicated, centrifuged,
and decanted (3ꢃ). The combined organic layers were filtered over
Celite and evaporated. The crude product was used without fur-
ther purification in the following step.
Methyl 6-[6-amino-2-(methylamino)-8-oxo-7,8-dihydro-1H-imida-
zo[4,5g]quinazolin-4-yl]-5,6-dideoxy-3-O-methyl-b-d-ribo-hexo-
furanoside (b-6c): According to GP 3, starting from 12c (190 mg,
0.29 mmol) and 10% Pd/C (400 mg) in EtOH (20 mL). The obtained
alcohol 13c was deprotected according to GP 1, using KOH
(162 mg, 2.88 mmol) in MeOH (10 mL) to the quinoxalinylamine,
which was transformed according to GP 2 in THF/MeOH/aq. conc.
HCl solution 1:1:1 (9 mL) to an anomeric mixture of a/b-6c. Chiral
HPLC ([Phenomenex Lux 5 mm Cellulose-2 AXIA packed column,
250ꢃ21.2 mm]; heptane/EtOH + 0.01% NH4OAc 30:70 isocratic for
60 min, flow rate 20 mLminꢁ1, detected by 263 nm UV), FC (MCI
gel; MeCN/H2O 0:100 to 100:0), and lyophilization yielded b-6c
(15 mg, 11% from 12c) as a white foam. M.p. >1808C (decomp);
1H NMR (600 MHz, (CD3)2SO): d=1.67–1.76 (m, 1H; Ha-C(5)), 1.82–
1.88 (m, 1H; Hb-C(5)), 2.89 (d, J=4.7 Hz, 3H; NMe), 2.96–3.06 (m,
1H; Ha-C(6)), 3.10–3.20 (m, 1H; Hb-C(6)), 3.30 (s, 3H; MeO-C(3)),
3.32 (s, 3H; MeO-C(1)), 3.59–3.66 (m, 1H; H-C(2)), 3.89 (brq, J
ꢂ6.3 Hz, 1H; H-C(4)), 3.97 (brt, Jꢂ4.2 Hz, 1H; H-C(3)), 4.67 (s, 1H;
H-C(1)), 4.91 (brs, 1H; HO-C(2)), 5.77 (brs, 2H; NH2), 6.90 (brs, 1H;
NH), 7.46 (s, 1H; H-C(9’)), 10.51 (brs, 1H; NH), 10.81 ppm (brs, 1H;
NH); 13C NMR (150 MHz, (CD3)2SO): d=21.91 (C(6)), 28.87 (NMe),
35.58 (C(5)), 53.92 (MeO-C(1)), 57.20 (MeO-C(3)), 71.88 (C(2)), 80.48
(C(3)), 84.49 (C(4)), 101.17 C(4’)), 108.16 (C(1)), 109.30 (C(9’)), 119.77
(C(3’a)), 130.36 (C(8’a)), 143.40 (C(9’a)), 148.24 (C(6’)), 149.07 (C(4’a)),
158.53 (C(2’)), 162.81 ppm (C(8’)); IR (ATR): n˜ =3630–2649 (w), 1702
(m), 1587 (s), 1540 (s), 1439 (m), 1344 (m), 1180 (m), 1073 (s), 982
(m), 790 (w), 757 cmꢁ1 (w); HR-MALDI-MS: m/z: calcd (%) for
C18H25N6O5+: 405.1881; found: 405.1880 (100) [M+H]+.
Methyl 6-[6-amino-2-(methylamino)-8-oxo-7,8-dihydro-1H-imida-
zo[4,5g]quinazolin-4-yl]-5,6-dideoxy-b-d-ribo-hexofuranoside (b-
6a): According to GP 1, starting from 12a (431 mg, 0.72 mmol)
and KOH (403 mg, 7.24 mmol) in MeOH (20 mL). The crude quinox-
alinylamine was directly transformed according to GP 2 in THF/
MeOH/aq. conc. HCl solution 1:1:1 (9 mL) to an anomeric mixture
of a/b-6a. Chiral HPLC ([Phenomenex Lux 5 mm Cellulose-2 AXIA
packed column, 250ꢃ21.2 mm]; heptane/EtOH + 0.01% NH4OAc
30:70 isocratic for 60 min, flow rate 20 mLminꢁ1, detected at
264 nm UV), HPLC ([Nucleosil-100 NH2, 250ꢃ21 mm, 5 mm]; H2O/
MeCN 0:100 for 5 min, 0:100 to 30:70 within 25 min, 30:70 for
10 min, flow rate 10 mLminꢁ1; in 5 portions), evaporation, and lyo-
philization yielded b-6a (30 mg, 11% from 12a) as a white foam.
M.p. >2008C (decomp); [a]2D5 =ꢁ8.5 (c=0.2 in H2O); 1H NMR
(600 MHz, (CD3)2SO): d=1.76–1.87 (m, 2H; H2C(5)), 2.90 (brs, 3H;
NMe), 3.04–3.13 (m, 2H; H2C(6)), 3.25 (s, 3H; OMe), 3.75 (brq, J
ꢂ6.7 Hz, 1H; H-C(4)), 3.76 (d, J=4.9 Hz, 1H; H-C(2)), 3.92 (dd, J=
6.5, 4.9 Hz, 1H; H-C(3)), 4.61 (s, 1H; H-C(1)), 5.95 (brs, 2H; NH2),
7.01 (brs, 1H; NH), 7.48 (s, 1H; H-C(9’)), 8.30 ppm (brs, 1H; NH);
13C NMR (150 MHz, (CD3)2SO): d=20.88 (Me of AcOH), 21.60 (C(6)),
28.85 (NMe), 35.37 (C(5)), 53.90 (OMe), 74.67 (C(3)), 75.00 (C(2)),
81.74 (C(4)), 103.21 (C(4’)), 107.68 (C(1)), 110.06 (C(9’)), 117.37
(C(8’a)), 132.64 (C(3’a)), 142.79 (C(9’a)), 145.78 (C(4’a)), 148.86 (C(6’)),
158.27 (C(2’)), 163.18 (COOH of AcOH), 164.88 ppm (C(8’)); IR (ATR):
n˜ =3306 (m), 3170 (m), 2937 (m), 2799 (m), 2718 (m), 1697 (m),
1646 (s), 1586 (s), 1525 (m), 1441 (m), 1411 (m), 1376 (m), 1347 (m),
1240 (w), 1198 (w), 1126 (w), 1103 (w), 1087 (w), 1032 (w), 986 (w),
776 (w), 764 (w), 701 cmꢁ1 (w); HR-MALDI-MS: m/z: calcd (%) for
C17H23N6O5+: 391.1724; found: 391.1724 (100) [M+H]+.
Acknowledgements
Methyl 6-[6-amino-2-(methylamino)-8-oxo-7,8-dihydro-1H-imida-
zo[4,5g]quinazolin-4-yl]-5,6-dideoxy-2-O-methyl-a/b-d-ribo-hexo-
furanoside (a/b 15:85; a/b-6b): According to GP 3, starting from
12b (130 mg, 0.20 mmol), 10% Pd/C (300 mg), and H2 (balloon) in
EtOH (20 mL). The obtained alcohol 13b was deprotected accord-
ing to GP 1, using KOH (116 mg, 2.07 mmol) in MeOH (10 mL) to
the quinoxalinylamine, which was transformed according to GP 2
in THF/MeOH/aq. conc. HCl solution 1:1:1 (9 mL) to an anomeric
mixture of a/b-6b. Chiral HPLC ([Phenomenex Lux 5 mm Cellulose-
2 AXIA packed column, 250ꢃ21.2 mm]; heptane/EtOH + 0.01%
NH4OAc 30:70 isocratic for 60 min, flow rate 20 mLminꢁ1, detected
by 262 nm UV), FC (MCI gel; MeCN/H2O 0:100 to 100:0), evapora-
This work was supported by the ETH Research Council and by
F. Hoffmann-La Roche, Basel. We acknowledge the beamline
support of Bessy II in Berlin for practical help and the HZB for
travel grants. We thank Dr. Nils Trapp for help with the small
molecule crystal structures and Oliver Dumele for DFT
calculations.
Keywords: molecular recognition · phosphate binding sites ·
shigellosis · structure-based design · water cluster
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
134
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[1] C. Bissantz, B. Kuhn, M. Stahl, J. Med. Chem. 2010, 53, 5061–5084.
[2] For a general overview of molecular recognition of phosphate groups,
see: A. K. H. Hirsch, F. R. Fischer, F. Diederich, Angew. Chem. Int. Ed.
2007, 46, 338–352; Angew. Chem. 2007, 119, 342–357.
[18] L. J. Barandun, F. Immekus, P. C. Kohler, T. Ritschel, A. Heine, P. Orlando,
G. Klebe, F. Diederich, Acta Crystallogr. Sect. D 2013, 69, 1798–1807.
[19] a) T. Ritschel, C. Atmanene, K. Reuter, A. Van Dorsselaer, S. Sanglier-Cian-
ferani, G. Klebe, J. Mol. Biol. 2009, 393, 833–847; b) F. Immekus, L. J. Bar-
andun, M. Betz, F. Debaene, S. Petiot, S. Sanglier-Cianferani, K. Reuter, F.
Diederich, G. Klebe, ACS Chem. Biol. 2013, 8, 1163–1178.
[20] S. Grꢀner, M. Neeb, L. J. Barandun, F. Sielaff, C. Hohn, S. Kojima, T. Stein-
metzer, F. Diederich, G. Klebe, Biochim. Biophys. Acta 2014, 1840, 2843–
2850.
[21] M. Neeb, P. Czodrowski, A. Heine, L. J. Barandun, C. Hohn, F. Diederich,
G. Klebe, J. Med. Chem. 2014, 57, 5554–5565.
[22] M. Neeb, M. Betz, A. Heine, L. J. Barandun, C. Hohn, F. Diederich, G.
Klebe, J. Med. Chem. 2014, 57, 5566–5578.
[23] S. Riniker, L. J. Barandun, F. Diederich, O. Krꢅmer, A. Steffen, W. F. van
Gunsteren, J. Comput.-Aided Mol. Des. 2012, 26, 1293–1309.
[24] a) W. E. Minke, D. J. Diller, W. G. J. Hol, C. L. M. J. Verlinde, J. Med. Chem.
1999, 42, 1778–1788; b) C. Clarke, R. J. Woods, J. Gluska, A. Cooper,
M. A. Nutley, G.-J. Boons, J. Am. Chem. Soc. 2001, 123, 12238–12247;
c) S. M. Tschampel, R. J. Woods, J. Phys. Chem. A 2003, 107, 9175–9181;
d) A. H. Daranas, H. Shimizu, S. W. Homans, J. Am. Chem. Soc. 2004, 126,
11870–11876; e) Z. Li, T. Lazaridis, J. Phys. Chem. B 2005, 109, 662–670;
f) K. Saraboji, M. Hꢆkansson, S. Genheden, C. Diehl, J. Qvist, U. Weining-
er, U. J. Nilsson, H. Leffler, U. Ryde, M. Akke, D. T. Logan, Biochemistry
2012, 51, 296–306.
[25] P. E. Morris, Jr., A. J. Elliott, J. A. Montgomery, J. Heterocycl. Chem. 1999,
36, 423–427.
[26] C. Altona, M. Sundaralingam, J. Am. Chem. Soc. 1972, 94, 8205–8212.
[27] H. A. Taha, M. R. Richards, T. L. Lowary, Chem. Rev. 2013, 113, 1851–
1876.
[3] T. Hunter, Cell 1995, 80, 225–236.
[4] a) F. A. Quiocho, D. K. Wilson, N. K. Vyas, Nature 1989, 340, 404–407;
b) J. D. Dunitz, Science 1994, 264, 670; c) J. E. Ladbury, Chem. Biol. 1996,
3, 973–980; d) C. Barillari, J. Taylor, R. Viner, J. W. Essex, J. Am. Chem.
Soc. 2007, 129, 2577–2587; e) P. Ball, Chem. Rev. 2008, 108, 74–108; f) J.
Michel, J. Tirado-Rives, W. L. Jorgensen, J. Am. Chem. Soc. 2009, 131,
15403–15411; g) M. Ellermann, R. Jakob-Roetne, C. Lerner, E. Borroni, D.
Schlatter, D. Roth, A. Ehler, M. G. Rudolph, F. Diederich, Angew. Chem.
Int. Ed. 2009, 48, 9092–9096; Angew. Chem. 2009, 121, 9256–9260;
h) B. Breiten, M. R. Lockett, W. Sherman, S. Fujita, M. Al-Sayah, H. Lange,
C. M. Bowers, A. Heroux, G. Krilov, G. M. Whitesides, J. Am. Chem. Soc.
2013, 135, 15579–15584; i) S. G. Krimmer, M. Betz, A. Heine, G. Klebe,
ChemMedChem 2014, 9, 833–846.
[5] a) A. T. Garcꢄa-Sosa, R. L. Mancera, P. M. Dean, J. Mol. Model. 2003, 9,
172–182; b) M. L. Verdonk, G. Chessari, J. C. Cole, M. J. Hartshorn, C. W.
Murray, J. W. M. Nissink, R. D. Taylor, R. Taylor, J. Med. Chem. 2005, 48,
6504–6515; c) B. C. Roberts, R. L. Mancera, J. Chem. Inf. Model. 2008, 48,
397–408; d) A. Amadasi, J. A. Surface, F. Spyrakis, P. Cozzini, A. Mozzar-
elli, G. E. Kellogg, J. Med. Chem. 2008, 51, 1063–1067; e) T. Beuming, R.
Farid, W. Sherman, Protein Sci. 2009, 18, 1609–1619; f) S. B. A. de Beer,
N. P. E. Vermeulen, C. Oostenbrink, Curr. Top. Med. Chem. 2010, 10, 55–
66; g) D. D. Robinson, W. Sherman, R. Farid, ChemMedChem 2010, 5,
618–627; h) L. Wang, B. J. Berne, R. A. Friesner, Proc. Natl. Acad. Sci. USA
2011, 108, 1326–1330.
[6] J. M. Durand, N. Okada, T. Tobe, M. Watarai, I. Fukuda, T. Suzuki, N.
Nakata, K. Komatsu, M. Yoshikawa, C. Sasakawa, J. Bacteriol. 1994, 176,
4627–4634.
[28] A. Ritter, D. Egli, B. Bernet, A. Vasella, Helv. Chim. Acta 2008, 91, 673–
714.
[7] U. Grꢅdler, H.-D. Gerber, D. M. Goodenough-Lashua, G. A. Garcia, R.
Ficner, K. Reuter, M. T. Stubbs, G. Klebe, J. Mol. Biol. 2001, 306, 455–467.
[8] For an overview of Shigella, see: a) P. J. Sansonetti, FEMS Microbiol. Rev.
2001, 25, 3–14; b) S. K. Niyogi, J. Microbiol. 2005, 43, 133–143; c) B. S.
Marteyn, A. D. Gazi, P. J. Sansonetti, Gut. Microbes 2012, 3, 104–120.
[9] a) K. L. Kotloff, J. P. Winickoff, B. Ivanoff, J. D. Clemens, D. L. Swerdlow,
P. J. Sansonetti, G. K. Adak, M. M. Levine, WHO Bull. 1999, 77, 651–666;
b) P. J. Sansonetti, PLoS Med. 2006, 3, e354.
[29] ACD/Labs, Version 12.01, I. Advanced Chemistry Development, Toronto,
2009.
[30] R. Mannhold, G. I. Poda, C. Ostermann, I. V. Tetko, J. Pharm. Sci. 2009,
98, 861–893.
[31] H. Van de Waterbeemd, H. Lennernas, P. Artursson, Drug Bioavailability:
Estimation of Solubility, Permeability, Absorption and Bioavailability,
Wiley-VCH, Weinheim, 2003.
[32] P. D. Adams, P. V. Afonine, G. Bunkꢇczi, V. B. Chen, I. W. Davis, N. Echols,
J. J. Headd, L.-W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy,
N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson,
T. C. Terwilliger, P. H. Zwart, Acta Crystallogr. Sect. D 2010, 66, 213–221.
[33] R. A. Laskowski, M. W. MacArthur, D. S. Moss, J. M. Thornton, J. Appl.
Crystallogr. 1993, 26, 283–291.
[10] a) D. Iwata-Reuyl, Bioorg. Chem. 2003, 31, 24–43; b) R. M. McCarty, V.
Bandarian, Bioorg. Chem. 2012, 43, 15–25.
[11] W. Xie, X. Liu, R. H. Huang, Nat. Struct. Biol. 2003, 10, 781–788.
[12] E. A. Meyer, N. Donati, M. Guillot, W. B. Schweizer, F. Diederich, B. Stengl,
R. Brenk, K. Reuter, G. Klebe, Helv. Chim. Acta 2006, 89, 573–597.
[13] L. J. Barandun, F. Immekus, P. C. Kohler, S. Tonazzi, B. Wagner, S. Wendel-
spiess, T. Ritschel, A. Heine, M. Kansy, G. Klebe, F. Diederich, Chem. Eur.
J. 2012, 18, 9246–9257.
[14] S. R. Hçrtner, T. Ritschel, B. Stengl, C. Kramer, W. B. Schweizer, B. Wagner,
M. Kansy, G. Klebe, F. Diederich, Angew. Chem. Int. Ed. 2007, 46, 8266–
8269; Angew. Chem. 2007, 119, 8414–8417.
[34] D. M. Goodenough-Lashua, G. A. Garcia, Bioorg. Chem. 2003, 31, 331–
344.
[35] a) C. S. Rye, J. B. Baell, Curr. Med. Chem. 2005, 12, 3127–3141; b) N. A.
Meanwell, J. Med. Chem. 2011, 54, 2529–2591; c) T. S. Elliott, A. Slowey,
Y. Ye, S. J. Conway, MedChemComm 2012, 3, 735–751.
[36] The PyMOL Molecular Graphics System, Version 1.6.0.0; Schrçdinger,
LLC., 2012.
[15] T. Ritschel, S. Hoertner, A. Heine, F. Diederich, G. Klebe, ChemBioChem
2009, 10, 716–727.
[16] P. C. Kohler, T. Ritschel, W. B. Schweizer, G. Klebe, F. Diederich, Chem. Eur.
J. 2009, 15, 10809–10817.
Received: October 21, 2014
[17] T. Ritschel, P. C. Kohler, G. Neudert, A. Heine, F. Diederich, G. Klebe,
ChemMedChem 2009, 4, 2012–2023.
Published online on December 5, 2014
Chem. Eur. J. 2015, 21, 126 – 135
www.chemeurj.org
135
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim