Fragment Screening Hit Draws Attention to a Novel Transient Pocket Adjacent to the Recognition Site of the tRNA-Modifying Enzyme TGT

Article abstract of DOI:10.1021/acs.jmedchem.0c00115

Fragment-based lead discovery was applied to tRNA-guanine transglycosylase, an enzyme modifying post-transcriptionally tRNAs in Shigella, the causative agent of shigellosis. TGT inhibition prevents translation of Shigella's virulence factor VirF, hence reducing pathogenicity. One discovered fragment opens a transient subpocket in the preQ1-recognition site by pushing back an aspartate residue. This step is associated with reorganization of further amino acids structurally transforming a loop adjacent to the recognition site by duplicating the volume of the preQ1-recognition pocket. We synthesized 6-carboxamido-, 6-hydrazido-, and 4-guanidino-benzimidazoles to target the opened pocket, including a dihydro-imidazoquinazoline with a propyn-1-yl exit vector pointing into the transient pocket and displacing a conserved water network. MD simulations and hydration-site analysis suggest water displacement to contribute favorably to ligand binding. A cysteine residue, exclusively present in bacterial TGTs, serves as gatekeeper of the transient subpocket. It becomes accessible upon pocket opening for selective covalent attachment of electrophilic ligands in eubacterial TGTs.

Full text of DOI:10.1021/acs.jmedchem.0c00115

Subscriber access provided by the University of Exeter
Article
Fragment Screening Hit Draws Attention to a Novel Transient Pocket
Adjacent to the Recognition Site of the tRNA-Modifying Enzyme TGT
Engi Hassaan, Christoph Hohn, Frederik R. Ehrmann, Friedrich Wieland
Goetzke, Levon Movsisyan, Tobias Hüfner-Wulsdorf, Maurice Sebastiani,
Adrian Hartsch, Klaus Reuter, François Diederich, and Gerhard Klebe
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00115 • Publication Date (Web): 09 Jun 2020
Downloaded from pubs.acs.org on June 19, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 39
Journal of Medicinal Chemistry
1
1
2
3
4
5
6
Fragment Screening Hit Draws Attention to a Novel Transient Pocket
Adjacent to the Recognition Site of the tRNA-Modifying Enzyme TGT
7
8
9
Engi Hassaan,†,# Christoph Hohn,‡,# Frederik R. Ehrmann,† F. Wieland Goetzke,‡ Levon Movsisyan,‡
Tobias Hüfner-Wulsdorf,† Maurice Sebastiani†, Adrian Härtsch,‡ Klaus Reuter†, François Diederich,‡
Gerhard Klebe†*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[†] Institute of Pharmaceutical Chemistry, University of Marburg, Marbacher Weg 6, 35032
Marburg, Germany
[‡] Laboratorium für Organische Chemie, ETH Zurich, Vladimir-Prelog-Weg 3, 8093 Zurich,
Switzerland
[#] These authors contributed equally to this work.
ABSTRACT
Fragment-based lead discovery was applied to tRNA-guanine transglycosylase, an enzyme modifying
post-transcriptionally tRNAs in Shigella, the causative agent of Shigellosis. TGT inhibition prevents
translation of Shigella’s virulence factor, hence reducing pathogenicity. One discovered fragment
opens a transient sub-pocket in the preQ1-recognition site by pushing back an aspartate residue. This
step is associated with reorganization of further amino acids structurally transforming a loop adjacent
to the recognition site by duplicating the volume of the preQ1-recognition pocket. We synthesized 6-
carboxamido-, 6-hydrazido-, and 4-guanidino-benzimidazoles to target the opened pocket, including a
dihydro-imidazoquinazoline with a propyn-1-yl exit vector pointing into the transient pocket, and
displacing a conserved water network. MD simulations and hydration-site analysis suggest water
displacement to contribute favorably to ligand binding. A cysteine residue, exclusively present in
bacterial TGTs, serves as gatekeeper of the transient sub-pocket. It becomes accessible upon pocket
opening for selective covalent attachment of electrophilic ligands in eubacterial TGTs.
Keywords: Transient binding pocket, tRNA-modifying enzyme, Fragment-based lead discovery,
Crystallography, Ligand synthesis, Hydration-site analysis
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 39
2
1
2
3
4
INTRODUCTION
5
6
7
8
Fragment-based lead discovery is an increasingly popular approach to generate first hits primarily as
promising starting points for drug development projects. In order to generate these starting points,
general-purpose fragment libraries have emerged that usually compile a broad range of chemotypes
and can therefore probe protein-binding sites efficiently. As opposed to tailored fragment libraries,
general-purpose libraries easily allow for the exploration of unprecedented binding motifs or scaffold
hopping. They can provide surprising insights into unexpected features of the studied protein.[1]
Amongst the recent famous success stories of fragment-derived drug candidates progressing in clinic
trials is the discovery of the Bcr-Abl Kinase allosteric inhibitor ABL001 by Novartis.[2]
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We recently performed a crystallographic fragment screening on tRNA-guanine transglycosylase (TGT),
an enzyme well studied in our laboratory.[3] It catalyzes the exchange of the genetically encoded
guanine at the wobble position 34 of the anticodon loop of tRNAs specific for Asn, Asp, His and Tyr
with the modified base preQ1 via breaking and regenerating the covalent bond between the ribose
sugar and the nucleobase.[4] To properly recognize the nucleobases, the TGT enzyme features two
adjacent aspartates (Asp102, Asp156 in the Z. mobilis enzyme, Figure 1a) in its active center. These
aspartates along with Asp280 are responsible for carrying out the catalysis of the base-exchange
reaction.[5] In addition, the carbonyl group of the substrate’s pyrimidone ring is hydrogen-bonded to
the backbone NH of Gly230 and the terminal carboxamide of Gln203.
Gln203
O
O
Leu231
a)
b)
Gly230
N
H
Leu231
N
H
H
N
N
Gly230
Gln203
H
H
H
O
H
O-
Asp156
N
Ala232
O
O
H
O
O
N+
H
H
O-
H
Asp156
HN
N+
Ala232
N
H
Met260
O
H
Gly261
Gly261
O
R
N
O-
O
Asp102
Met260
H
Asp102
HN
d)
c)
Gln203
O
Leu231
O
Gly230
Gln203
Leu231
Gly230
N
H
N
H
N
N
H
H
H
Asp156
O
N
O
Ala232
O
N
Ala232
H
H
O
H
O-
N
H
N
H
N+
H
Asp156
N
N
H
O-
O
Asp102
Asp102
ACS Paragon Plus Environment

Page 3 of 39
Journal of Medicinal Chemistry
3
1
2
3
4
5
6
7
8
9
Figure 1. a) Schematic binding mode of the Arg-related fragment as bound to Zymomonas (Z.) mobilis
TGT; b) Crystal structure of this fragment in the catalytic center of TGT (carbon atoms magenta, PDB
code: 5UTI, 1.36 Å resolution). The fragment succeeds in opening a transient binding pocket next to
Gly230 and Gln203, which is solvated by water molecules; c) Schematic binding mode of lin-
benzoguanine (1.1) as bound to Z. mobilis TGT; d) Crystal structure of 1.1 (carbon atom cyan, PDB code:
4PUK, 1.49 Å resolution) in the catalytic center of TGT.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our screen identified a small fragment that is chemically related to the amino acid arginine, which
opens a new, transient sub-pocket when bound to the recognition pocket of the target protein (PDB
code: 5UTI). [3] Via its guanidinium group, the discovered fragment binds to the Asp102/Asp156 motif
forming two bidentate hydrogen-bond interactions and perturbs the spatial arrangement of the
remaining recognition pattern next to Gly230 and Gln203 (Figure 1b). As a result, a small transient
pocket opens, which increases the volume of the recognition site significantly (for further illustration
s. Supporting Information, Figure S49). Two water molecules, stabilized by the nearby carboxylate
group of the fragment, are found in the transient pocket. The amino nitrogen of the fragment also
forms a hydrogen bond with the amide oxygens of Leu231 and Met260. The observed structure
immediately provokes some interesting questions. Is the opening of the transient pocket just an
uncommon and reinforced but functionally useless adaptation of the protein or does it indicate a
functionally essential feature of the enzyme? Furthermore, can a ligand successfully accommodate this
transient pocket and does this support the development of more potent or even more selective
ligands?
The studied enzyme TGT is a potential target to fight Shigellosis.[6,7] This infection is a world-wide
endemic with 165 million cases reported per year, including 1.1 million deaths globally, mostly in
children under five years of age.[8] It is an acute bacterial infection of the intestine caused by the gram-
negative bacterium Shigella. Symptoms of the disease include sudden abdominal cramping, nausea,
fever, vomiting, and bloody stool. Sources of infection include contaminated food or water and
therefore the disease predominantly affects low income countries where poor sanitation and
overcrowding are prevalent.[8] Furthermore, this disease can also rapidly spread in countries of higher
life standards.[9] The usual treatment of shigellosis is an antibiotics therapy; however, resistance is
rapidly developed. Shigella pathogenicity depends on virulence factors, which are required to invade
epithelial cells. The expression of these virulence factors is modulated by the enzyme TGT (EC 2.4.2.29).
This enzyme occurs in all domains of life, however, it differs in substrate specificity and the pre-
modified base to be inserted.[10] Eubacterial TGTs catalyze the exchange of the base guanine by preQ1
which is, once incorporated, further transformed to queuine by other enzymes.[4] Eukaryotic TGTs,
which show a high sequence identity to eubacterial TGTs (e.g. 43.2% between Homo sapiens and
Shigella flexneri) directly incorporate queuine at the wobble position of tRNAs. These enzymes use a
subpocket to accommodate the additional substituent of queuine, which is located in a similar region
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 39
4
1
2
3
4
5
6
7
as the transient pocket opened by the above-mentioned fragment.[5] Archaebacterial TGTs on the
other hand, show only about 20–25% sequence identity to eubacterial TGTs and address guanine at a
different site, namely position 15 of the D-arm in the majority of archaeal tRNAs.[10]
8
9
In the present contribution, we study parameters which feature the above-described opening of the
transient recognition pocket and elucidate whether the observed adaptations relate to any functional
properties of the enzyme family that could be exploited for the development of potent and selective
anti-infective drugs. Starting with the well-characterized lin-benzoguanine scaffold 1,[11] appropriate
to inhibit eubacterial TGT, we first designed some ligands capable to perturb the recognition pattern
next to the residues Gly230/Gln203 as suggested by our fragment structure. By molecular design and
subsequent chemical synthesis of suitable model compounds, determination of their binding affinity
to the enzyme and crystal structure analyses along with a molecular dynamics-based hydration site
analysis, we succeeded to trace the role and trigger of the opening of the transient pocket and its
potential impact on the development of selective inhibitors of eubacterial TGTs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 5 of 39
Journal of Medicinal Chemistry
5
1
2
3
4
RESULTS AND DISCUSSION
5
6
7
8
Ligand Design. In the past, the design of our ligands, aimed at inhibiting bacterial TGT, was focused on
lin-benzoguanine derivatives (Scheme 1, scaffold 1) as this parent scaffold perfectly occupies the
recognition pocket (Schematic binding mode Figure 1c). Without further decoration, it already shows
an inhibition of Ki = 58 ± 36 nM[11] and provides multiple options for exit vectors to efficiently address
the various sub-pockets surrounding the central recognition site. Thus, sub-nanomolar potency could
be achieved.[12] Stimulated by the binding geometry of the above-mentioned fragment complex and
the surprising opening of a transient pocket, we decided to either open the aminopyrimidone ring of
the lin-benzoguanine scaffold (scaffolds 2 and 3) or to reduce the carbonyl group in the pyrimidone
ring of the parent scaffold (scaffold 4). The resulting 7,8-dihydro-imidazoquinazolines in 4 allow the
attachment of substituents at C8 that may penetrate into the opened transient pocket (s. Scheme 1).
Thereby, the carbonyl function of the pyrimidone ring, interacting with the backbone NH of Gly230
and the carboxamide nitrogen of Gln203, would be lost. As indicated by the fragment complex, some
inherent flexibility with respect to the latter residues might be given. In order to start our ring opening
exercise conservatively and to study the impact of the abandoned hydrogen-bonding interactions of
the core scaffold to Gly230 and Gln203, we decided to also investigate benzimidazole derivatives
(scaffolds 2 and 3) with substituents attached to positions 5 or 6. For this purpose, we considered
amides, hydrazides, and guanidines, respectively. With respect to R1 in Scheme 1, we studied all ligands
with an attached N-methyl group at position 2. For some selected examples, also the morpholino ethyl
derivatives were investigated.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Design of pyrimidone ring-opened benzimidazole (2 and 3) derivatives and tricyclic 7,8-
dihydroimidazoquinazoline analogues (4) of the parent lin-benzoguanine scaffold (1).
Synthesis. The synthesis of 2-substituted lin-benzoguanine 1[13] and lin-benzohypoxanthines 5[14] (for
the structure, see Table 1) was previously reported by our group. The 6-substituted benzimidazole
ligands 2 were synthesized in 6 steps from 5-benzimidazole carboxylic acid (6) (Scheme 2).[13,15] The
synthesis commenced with an esterification to give 7. The benzimidazole nitrogen was sulfamoyl-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 39
6
1
2
3
4
5
6
7
8
9
protected to give an inseparable, tautomeric mixture of 8 and 8’ (for the structure of 8’, see Table S1
in the Supporting Information), followed by selective bromination at the 2-position. Partial separation
of the tautomers was possible by flash column chromatography, and the enriched tautomer 9 was
subjected to nucleophilic aromatic substitution with primary amines to afford 2-aminobenzimidazoles
10.1 and 10.2. Functional group interconversion of the ester with aqueous ammonia, methanolic
methylamine, or aq. hydrazine under microwave irradiation afforded the protected carboxamides and
hydrazines 11.1-11.6, respectively. Final deprotection with HCl (4 M solution in dioxane) and reverse
phase flash chromatography purification yielded the ligands 2.1-2.6.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of 6-Substituted Benzimidazole Ligands 2.a
aReaction conditions: a) H2SO4, MeOH, 65 °C, 18 h, quant.; b) Me2NSO2Cl, NEt3, toluene, 23 °C, 24 h,
66 % (1/0.6 mixture with tautomer 8’; 8’ not shown ); c) LiN(SiMe3)2, THF, –78 °C, 1h, then NBS, THF, –
78 °C to 23 °C, 15 min, 56 % (1/0.05 mixture with tautomer 9’; 9’ not shown); d) for 10.1: MeNH2,
EtOAc/MeOH, 23 °C, 14 h, quant.; for 10.2: 2-morpholinoethylamine, iPr2NEt, EtOAc/MeOH, 23 °C,
15 h, 54%; e) for 11.1 and 11.2: aq. NH4OH, MeOH, 100 °C (MW), 15-30 min, 11.1: 37%; 11.2: 31%; for
11.3 and 11.4: MeNH2, MeOH 140 °C (MW), 1.5-2.5 h, 11.3: 67%; 11.4: 75%; for 11.5 and 11.6: N2H4,
H2O, 100 °C (MW), 30 min, 11.5: 80%; 11.6: 47%; f) HCl (4 M solution in dioxane), dioxane, reflux, 14-
45 h, 2.1: 23%; 2.2: 64%; 2.3: 14 %; 2.4: 19%; 2.5: 9%; 2.6: 52%. THF = tetrahydrofuran; NBS = N-
bromosuccinimide; MW = microwave.
For the 5-substituted benzimidazole ligands 3.1 and 3.2, we pursued an alternative synthetic strategy
(Scheme 3).[15] 4-Nitroaniline 12 was converted into 5-nitrobenzimidazoles 13.1 and 13.2 via a two-
step protocol with isothiocyanates, followed by treatment with EDC.[16] The nitro group was reduced
using PtO2 and Pd/C under H2, and the obtained aniline was guanylated with N,N′-di-Boc-1H-pyrazole-
1-carboxamidine to afford 14.1 and 14.2, respectively. The N-Boc protecting groups were cleaved with
HCl (4 M ethereal solution), and recrystallization afforded the ligands 3.1 and 3.2.
ACS Paragon Plus Environment

Page 7 of 39
Journal of Medicinal Chemistry
7
1
2
3
4
Scheme 3. Synthesis of 5-Substituted Benzimidazole Ligands 3.a
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a Reaction conditions: a) for 13.1 and 13.2: R1NCS, DMF, 90 °C, 2 h, then EDC, 90 °C, 2 h; 13.1: 69%;
13.2: 63% b) for 14.1 and 14.2: H2, Pd/C, PtO2, EtOH, 4 h, 23 °C, c) N,N'-di-Boc-1H-pyrazole-1-
carboxamidine, iPr2NEt, DMF, 23 °C, 16 h; 14.1: 55%; 14.2: 50%; d) for 3.1 and 3.2: HCl (4 M ethereal
solution), Et2O/CH2Cl2, 23 °C, 10 min, 3.1: 59%; 3.2: 52%. DMF = N,N-dimethylformamide; EDC = 1-
ethyl-3-(3-dimethylaminopropyl)carbodiimide; Boc = tert-butyloxycarbonyl.
Our synthetic strategy for the tricyclic dihydro-imidazoquinazoline analogues 4.1 and 4.2 relied on an
intramolecular cyclization of N-Boc protected guanidines on activated benzylic alcohols (Schemes 4
and 5). The synthesis of precursor 15 was previously described by our group.[17] The methyl ester of 15
was reduced with LiAlH4 to afford benzyl alcohol 16. Chemoselective N-guanylation gave N-Boc
protected 17. Tosyl chloride was used to activate the benzylic alcohol for cyclization to 18. Final
cleavage of the SEM and the N-Boc protecting groups of 18 was achieved with HBF4, and
recrystallization yielded ligand 4.1.
Scheme 4. Synthesis of the Tricyclic Dihydro-imidazoquinazoline Ligand 4.1.a
a Reaction conditions: a) LiAlH4, THF, 0 °C, 1 h, 58%; b) N'-di-Boc-1H-pyrazole-1-carboxamidine, iPr2NEt,
DMF, 23 °C, 2 h, 44%; c) TsCl, KOH, THF, reflux, 1 h, 52%; d) HBF4Et2O, CH2Cl2, 23 °C, 30 min, 57%. SEM
= 2-(trimethylsilyl)ethoxymethyl; Ts = p-toluenesulfonyl.
We applied a similar synthetic strategy to construct the 2-amino-benzimidazole core of 4.2. A three-
step protocol starting from 4-bromobenzene-1,2-diamine, (19)[18] gave nitro derivative 20, and
benzimidazole formation afforded 21. Introduction of the SEM protecting group yielded an inseparable
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 39
8
1
2
3
4
5
6
7
8
9
regioisomeric mixture of 22 and 22´ (for the structure of 22´, see Table S1 in the Supporting
Information), and carbonylation via a sequence of a palladium catalyzed Heck reaction with ethyl
acrylate, followed by oxidative cleavage with catalytic OsO4, afforded a mixture of N-tautomeric
nitrobenzaldehydes 23 and 23’ (for the structure of 23’, see Table S1 in the Supporting Information),
which was separated by careful column flash chromatography. Alkynylation with in situ generated
propynyllithium gave the racemic propargylic alcohol (±)-24. The nitro group was reduced selectively
with zinc and aq. NH4Cl, and subsequently the activated guanylation reagent N,N’-di-Boc-1H-pyrazole-
1-carboxamidine was used to prepare (±)-25. Intramolecular cyclization of the guanidine in (±)-25 on
the propargylic alcohol was again achieved by activation with tosyl chloride under basic conditions to
yield (±)-26. Final deprotection with SnCl4 and subsequent purification by reverse-phase HPLC afforded
the bis-hydrochloride salt of racemic (±)-4.2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. Synthesis of the Tricyclic Dihydro-imidazoquinazoline Ligand 4.2.a
a Reaction conditions: a) TsCl, pyridine, 65 °C, 17 h, 83%; b) HNO3, AcOH, 50 °C, 1 h, 67%; c) H2SO4, H2O,
85 °C, 2 h, 87%; d) MeNCS, (CH3)2SO, 70 °C, 21 h, then EDC, 70 °C, 1 h, 56%; e) SEM chloride, NaH, DMF,
0 °C, 1.5 h, 77% (1.2/1 mixture with tautomer 22’; 22’ not shown); f) CH2=CHCO2Et, [Pd(OAc)2], PPh3,
toluene, 110 °C, 48 h; g) OsO4, NaIO4, THF/H2O, 23 °C, 36 h, 52% of 23 and 32% of 23’ (not shown) (over
2 steps). h) H3CCH=CHBr, n–BuLi, THF/n–hexane, –78 °C to 23 °C, 3 h, 83%; i) Zn, aq. NH4Cl, MeOH, 0 °C,
2.5 h; j) N'-di-Boc-1H-pyrazole-1-carboxamidine, iPr2NEt, CH2Cl2, 21 h, 50 °C, 57% (over 2 steps); k) TsCl,
Cs2CO3, 23 °C, 4 h, 58%; l) SnCl4, CH2Cl2, 0 °C to 23 °C, 6 h, 34%.
ACS Paragon Plus Environment

Page 9 of 39
Journal of Medicinal Chemistry
9
1
2
3
4
5
Table 1. Chemical Formulas, Inhibition (Ki) or Dissociation (KD) Constants in µM, PDB codes, and
resolution (Å) of the Complexes Formed by the Studied Compounds with Z. mobilis TGT.a
6
7
8
Ligand
Parent scaffold
R
PDB
code
Resolution
Å
KI [M]
9
1.1
0.058 ± 0.036[8]
4PUK[6]
4PUJ[6]
1.49
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0.006 ± 0.006[8]
1.42
1.2
2.1
2.2
2.3
2.4
2.5
2.6
3.1
3.2
300 ± 37
544 ± 43
270 ± 50
264 ± 40
283 ± 40
282 ± 18
19 ± 2
5J9M
5JT5
---
1.33
1.21
---
---
---
5J9N
5JT6
5J9O
5JT7
1.64
1.54
1.41
1.70
58 ± 2
4.1
(±)-4.2
5.1
17.2 ± 0.2
47.9 ± 6.6
6.5 ± 2.9[9]
6RKT
6RKQ
1.75
1.67
1.82
R =
3S1G[9]
4.1[9]
4Q4R
1.45
5.2
a Binding constants were determined via a radioactive assay.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 39
10
1
2
3
4
5
6
7
8
Affinity Determination. The affinities of the studied compounds were determined using a radioactive
assay. This assay is a functional biochemical assay that measures the incorporation of radioactively
labeled guanine into tRNA. The potency of an inhibitor for TGT is quantified by the amount of
decreasing the rate of guanine incorporation into tRNA resulting in a Ki value.[19]
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Crystal Structure Determinations. The crystal structures were obtained by co-crystallizing the
corresponding ligands with the enzyme. In all cases, structures with the common twofold dimer
packing in space group C2 could be determined at a resolution between 1.21 – 1.82 Å. In some cases,
both a structure with R1= Me and R1= morpholino ethyl could be resolved. However, as previously
observed, the morpholino substituent showed pronounced positional disorder in all examples apart
1.2[13] and no difference electron density could be attributed to this part of the ligand. The structural
comparison of the corresponding complexes with R1 = Me and morpholino ethyl showed identical
binding modes, which are adopted in all cases. Accordingly, in the following section only crystal
structures of the R1 = Me derivatives will be described. The crystallographic tables of all here
investigated complexes can be found in the Supporting Information and the coordinates of all
complexes have been deposited with the Protein Data Bank (PDB; www.rcsb.org).
Co-Crystal Structures with the Benzimidazoles of 6-Carboxamide 2.1 and 6-Hydrazide 2.5. As
observed in previous studies, the bicyclic benzimidazole core scaffold binds to the nucleobase
exchange pocket and forms the expected H-bonding pattern to the backbone atoms of the amino acids
Ala232, Leu231 and Gly230 (Figure 2a,b).[12,13,20] The carboxamide terminus of Gln203 acts as H-bond
donor to the carbonyl groups of the attached 6-amide or 6-hydrazide substituents of 2.1 and 2.5.
Furthermore, the bicyclic core structure is sandwiched between the amino acids Tyr106 and Met260,
establishing π-stacking interactions (not shown). The side chain of the catalytically active Asp102
adopts two alternative, equally populated conformations in the complex with 2.1 (Figure 2a), whereas
in the complex with the hydrazide substituent 2.5, only the conformer with the carboxylate group
oriented into the recognition pocket is observed (Figure 2c). The geometry with the side chain rotated
off the catalytic center is also found in the structure of the apo protein[21] and complexes
accommodating lin-benzohypoxanthines (5.1, 5.2).[14] The alternative orientation towards the ligand-
binding site is generally found for complexes with lin-benzoguanines. In TGT·2.1, the carboxylate group
of Asp102 binds to the primary amide function of the ligand (2.9 and 3.0 Å) via the interstitial water
molecule W1 (Figure 2b). The second oxygen of this carboxylate is solvated by additional water
molecules found in the ribose-34 pocket (2.9 - 3.1 Å). The carboxylate group of Asp156 also forms an
H-bond to the terminal amide nitrogen of 2.1 (3.0 Å) and water W1 (Figure 2b) additionally solvates
ACS Paragon Plus Environment

Page 11 of 39
Journal of Medicinal Chemistry
11
1
2
3
4
5
the neighboring oxygen of the Asp156 carboxylate (3.1 Å). Apart from these changes in the local water
structure, the complex with 2.1 is very similar to that with 5.1 (Figure 2b).
6
7
8
b)
a)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
c)
d)
Figure 2: a) Crystallographically determined binding mode of 2.1 (carbon atoms green, PDB code:
5J9M) in the recognition pocket of TGT, heteroatoms type-coded, water molecules as small spheres,
all distances in Å, hydrogen-bonds as dashed lines; b) Superposition of 2.1 (carbon atoms green PDB
code: 5J9M) and 5.1 (carbon atoms orange, PDB code: 3S1G [14]); c) Crystallographically determined
binding mode of 2.5 (carbon atoms dark red, PDB code: 5J9N); d) Superposition of 2.5 (carbon atoms
dark red, PDB code:5J9N), 2.6 (carbon atoms cyan, PDB code: 5JT6) and 5.1 (carbon atoms orange, PDB
code: 3S1G [14]), the push back of Asp156 and the conformational transformation of the protein
opening the transient pocket in the complexes with 2.5 and 2.6 in comparison to 5.1 is indicated by
arrows. For electron density around the bound ligands and water molecules, s. Figure S50, Supporting
Information).
In the complex with the hydrazide-based benzimidazole 2.5 (Figure 2c), the terminal nitrogen of the
hydrazide protrudes directly toward the carboxylic acid carbon of the Asp156 side chain and forms two
H-bonds to its oxygens (2.6 - 2.8 Å). The hydrazide group adopts the trans geometry (179°) and pushes
Asp156 by approx. 1.2 Å (Cγ---Cγ) out of the recognition pocket compared to the complex with 5.1. This
push back of Asp156 is associated with a reorganization of further neighboring amino acids and leads
to a conformational change of the loop between the amino acids Glu157 and Tyr161 (Figure 2d and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 39
12
1
2
3
4
5
6
7
8
9
Supporting Information Figure S49). As mentioned above, the carboxylate group of Asp102 in 2.5
refines to one fully populated orientation toward the ligand. Nevertheless, its oxygens do not build a
water-mediated interaction network with the ligand, as found for 2.1. No sufficiently well-defined
difference electron density for a water molecule could be detected. In contrast, the interactions to the
water molecules in the ribose-34 pocket are also established here. The affinities of the carboxamide
2.1 (Ki = 300 ± 37 µM) and the morpholino analog 2.2 (Ki = 544 ± 43 µM) decrease by a factor of 46 and
132, respectively, compared to the analog lin-benzohypoxanthines 5.1 (Ki = 6.5 ± 2.9 µM) and 5.2 (Ki =
4.1 µM). In the crystal structures with the latter ligands a well-defined water network between the
ligand and Asp102 is formed (Figure 2b).[14] A similar network is impossible to establish with 2.1, which
likely explains the reduced affinity of the 5-amide derivatives 2.1 and 2.2. The affinity loss compared
to the parent lin-benzoguanine scaffold (1.1, 1.2) is even more dramatic and accounts for nearly five
orders of magnitude with respect to 2.1 (Table 1). The structural rearrangement of the loop extending
from Glu157 to Tyr161 along with the loss of interactions and the reorganization of water molecules
makes it difficult to compare the hydrazide-based ligands 2.5 and 2.6 with the 6-amide analogs 2.1 and
2.2. Nevertheless, it appears surprising that for 2.1 (Ki = 300 ± 37 µM) and 2.5 (Ki = 283 ± 40 µM) the
affinity falls into the same range. This suggests that only minimal energetic costs are involved in the
structural rearrangement of the protein, which may represent the transformation to a conformation
linked to a protein function that stabilizes the bound ligands.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Crystal structures with the ring opened 5-guanidino-benzimidazole 3.1. Compared to the lin-
benzoguanine scaffold, the 5-guanidino-benzimindazole 3.1 lacks the endocyclic carbonyl function to
establish direct H-bonding interactions with Gly230 and Gln203. The corresponding crystal structure
shows (Figure 3a) that these amino acids are now solvated and stabilized (3.3 - 3.5 Å) via hydrogen
bonds to an interstitial water molecule (W1, Figure 3a). The carboxylate groups of Asp102 and Asp156
each adopt an orientation which is analogously observed with the bound lin-benzoguanine scaffold
1.1.[11] The basic guanidino group (pKa >12) likely binds in charged state. Consequently, salt bridges are
formed, as expected, with both aspartates. Furthermore, the 2-aminoimidazole interacts with the
carbonyl groups of Ala232 and Leu231 (2.9 Å) via H-bonds.
b)
a)
ACS Paragon Plus Environment

Page 13 of 39
Journal of Medicinal Chemistry
13
1
2
3
4
5
6
7
8
9
Figure 3: a) Crystallographically determined binding mode of 3.1 (PDB code: 5J9O) in the recognition
pocket of TGT, heteroatoms type-coded, water molecules small spheres, all distances in Å, hydrogen-
bonds as dashed lines; b) Superposition of 3.1 (carbon atoms cyan, PDB code: 5J9O), 3.2 (carbon
atoms light blue, PDB code: 5JT7 and 5.1 (carbon atoms orange, PDB code: 3S1G)[14] shows the push
back of Asp156 and the triggered conformational transformation of the protein opening the transient
pocket in the complexes with 3.1 and 3.2 in comparison to 5.1 (indicated by the arrows). The latter
ligand has only one H-bond donor to interact with Asp156, therefore a slightly tilted geometry is
found, however, no opening of the transient pocket is initiated. For electron density around the
bound ligands and water molecules, s. Figure S50, Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
At first sight, the loss of the H-bonds to Gly230 and Gln203 also leads to a drop in affinity as observed
for the 6-amide analogs 2.1 and 2.5 with respect to the corresponding lin-benzoguanine 1.1 or lin-
benzohypoxanthine 5.1. With Ki = 19 ± 2 µM for 3.1 and Ki = 58 ± 2 µM for 3.2, the 5-guanidine-
substituted ligands bind slightly stronger to TGT than the studied 6-benzimidazoles. This suggests that
the gained charge-assisted interactions to Asp102 and Asp156 can only partially compensate for the
lost interactions to the nitrogen atoms of Gly230 and Gln203. However, a perhaps even more
important effect must be considered. By opening the pyrimidone ring of the lin-benzoguanine scaffold,
the ligand scaffold loses its correct preorganization. Instead, the guanidine portion, attached to the
benzimidazole, must adopt one single conformation, which supposedly is not the most favored one.
This, along with the loss of conformational degrees of freedom, will have a detrimental effect on
binding affinity. In other studies we saw that this effect can reduce the binding constant by two to
three orders of magnitude.[22]
Interestingly, as with hydrazide-substituted benzimidazoles 2.5 and 2.6, the structural transformation
of the secondary structural element (Glu157 to Tyr161) is also observed with the guanidine-substituted
ligands 3.1 and 3.2 (Figure 3b). Analogously, a spatial shift of Asp156 is observed, as seen in TGT·2.5
(Cγ---Cγ = 0.9 Å) compared to the complex with 1.1. However, a second aspect may be of importance
in this complex to transfer the enzyme into the alternative state with the opened transient pocket. The
above-mentioned water W1 (Figure 3a), which binds in the vicinity of the original position of the
carbonyl O-atom of the lin-benzoguanine scaffold, solvates the polar groups of Gly230 and Gln203, but
in addition, it fills the created void and exerts steric pressure onto the flexible loop of Glu157 to Tyr161
(Figure 3a). Accordingly, W1 supports the expansion of the opening pocket.
To study the impact of steric pressure in this region of the binding pocket by a bound ligand in more
detail, we designed chemically reduced tricyclic analogs of the lin-benzoguanine scaffold bearing an
b)
c)
a)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 39
14
1
2
3
4
5
6
7
8
9
sp3-carbon atom at the former spatial position of the carbonyl group in pyrimidone ring (4 in Scheme
1). Through the attachment of substituents at this carbon atom, a stereogenic center is created.
Therefore, we decided to first study the achiral unsubstituted derivative 4.1 and subsequently the
propyn-1-yl analog ((±)-4.2). Modeling considerations strongly supported the hypothesis that only the
R-enantiomer will orient the attached substituents in a way to fit into the possibly opened transient
pocket, as modeling the S-enantiomer would create major steric conflicts with the protein. The
compounds were synthesized (Schemes 4 and 5) and subsequently subjected to co-crystallization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: a) Crystallographically determined binding mode of 4.1 (carbon atoms dark red, PDB code:
6RKT) in the recognition pocket of TGT, heteroatoms type-coded, all distances in Å, hydrogen-bonds
as dashed lines; b) 4.1 difference electron density (|Fo|-|Fc| omit map) contoured at 3σ, protein
displayed with the solvent accessible surface showing the recognition site together with the opened
transient pocket; c) Conformational transformation of the protein induced by the binding of 4.1
(carbon atoms dark red, PDBcode: 6RKT). For comparison, the complex with 4.1 is superimposed with
the crystal structure of the apo protein (carbon atoms in wheat color). For electron density around the
bound ligands and water molecules, s. Figure S50, Supporting Information).
Co-Crystal Structures with Dihydro-imidazoquinazolines 4.1 and (±)-4.2 featuring the Tricyclic lin-
Benzoguanine Scaffold. Ligand 4.1 binds with an occupancy of 87% to the recognition pocket. The
guanidine moiety in the six-membered ring forms two bidentate hydrogen bonds with the carboxylate
groups of Asp156 and Asp102 (Figure 4a). The second guanidine moiety in the imidazole ring interacts
via a hydrogen bond with the carbonyl oxygen of Leu231, while the exocyclic amino group hydrogen-
bonds the carbonyl oxygen of Ala232. Furthermore, 4.1 establishes a π-π stacking by being sandwiched
between the side chains of Tyr106 and Met260. Thus, a very similar recognition pattern as seen for lin-
benzoguanine 1.1[11] is found in this area. Nevertheless, the ligand also induces the structural
transformation of the secondary structural element (loop Glu157 to Tyr161), which leads to an opening
of the transient sub-pocket (Figure 4b, c) and the carboxamide group of Gln203 forms a van der Waals
contact to C8 of the dihydro-imidazoquinazoline scaffold. Due to the pocket opening, the distance
between Val233 Cα and Cys158 Sγ expands in TGT·4.1 to 10.2 Å, which amounts in the complex with
the lin-benzoguanine parent scaffold 1.1 to 6.5 Å.
The complex formed with (±)-4.2 shows a very similar geometry to TGT·4.1 with nearly identical
recognition pattern. In contrast to the latter complex, Gln203 adopts an altered conformation
providing additional space to accommodate the needle-like propyn-1-yl side chain. Even though the
c)
b)
a)
ACS Paragon Plus Environment

Page 15 of 39
Journal of Medicinal Chemistry
15
1
2
3
4
5
6
7
8
9
complex was co-crystallized from a racemic mixture, the enzyme only picks the R-enantiomer. During
structure determination, placement of the S-enantiomer into the electron density lead to negative and
positive sigma values in the m|Fo|-|Fc|map after refinement, which indicates an incorrect solution.
Thus, in agreement with our modeling considerations described above, the S-enantiomer does not
bind as it would have to poke into the protein and create substantial steric clashes. The difference
density allows placement of all atoms of the propyn-1-yl side chain, even though the less-well defined
density next to the terminal methyl group indicates enhanced residual mobility of the side chain
terminus. As for the other complex, (±)-4.2 induces a spatial shift of the loop comprising Glu157 to
Tyr161. Asp156 is, compared to TGT·1.1, pushed by 0.6 Å out of the recognition pocket and helps to
trigger the structural transformation of the loop (Figure 5c). The opening of the transient pocket results
in a mutual distance between Val233 Cα and Cys158 Sγ of 10.8 Å.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5: a) Crystallographically determined binding mode of (R)-4.2 (carbon atoms orange, PDB code:
6RKQ) in the recognition pocket of TGT, heteroatoms type-coded, all distances in Å, hydrogen-bonds
as dashed lines; b) (R)-4.2 together with the difference electron density (|Fo|-|Fc| omit map)
contoured at 3σ, protein displayed with the solvent accessible surface showing the recognition site
together with the opened transient pocket; c) Conformational transformation of the protein induced
by the binding of (R)-4.2 (carbon atoms orange, PDB code: 6RKQ). For comparison, the complex with
4.2 is superimposed with the crystal structure of the apo protein (carbon atoms in wheat color). For
electron density around the bound ligands and water molecules, s. Figure S50, Supporting
Information).
Expansion of the Recognition Pocket by a Transient Sub-Pocket. The enzyme TGT can perform large
conformational adaptations upon ligand binding. [21] The opening of the small transient pocket is not
only observed in our initially described fragment structure [3] (Figure 1) but also in the here described
subset of complexes. These complexes push Asp156 slightly out of the recognition pocket and create
sufficient steric pressure onto residues Gly230 and Gln203, thus initiating the opening of the transient
binding cleft. Transient pockets may populate opened or closed states within the conformational
dynamics of the protein under thermodynamic equilibrium conditions. Alternatively, transient pockets
may open only in the presence of ligating species such as a drug or substrate molecule, a situation that
we could observed for the specificity pocket found in the enzyme aldose reductase.[23] Under either
scenario, conformational selection or induced fit, energetically favorable ligand binding is possible.
However, in both cases differing thermodynamic consequences accompanied by fundamentally
different binding mechanisms are at play. We therefore performed MD simulations in order to obtain
some insights into the distribution of opened and closed states of the binding pocket. We found that
the binding pocket of the apo enzyme is transiently adopting opened and closed states during the MD
simulation (3x 200 ns simulations). We first used the mutual Cys158Sγ -Val233Cα distance (see analysis
of the crystal structures above) but furthermore the adopted volume of the binding pocket to separate
both states. The distance descriptor shows a broad distribution suggesting multiple states for the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 39
16
1
2
3
4
5
6
7
8
9
transition between both states (see Figure S1). Nonetheless, a large portion of conformations during
the MD simulation adopts distances between 6 - 8 Å, thus indicating a closed binding pocket. The
volume descriptor records a bimodal distribution with maxima at 150 Å3 and 650 Å3. The ratio of the
integrated densities of the two volume distributions suggests that 13% and 86% of the time the lower
and upper distributions are populated (see Figure S1). The mutual correlation of both descriptors (see
Figure S2) reveals that the small volumes are only present at lower distance range. This is in agreement
with the crystal structures and we assign distances between 6 – 8 Å to the closed state. Using the
volume distribution as molecular descriptor to characterize the opened state is somewhat
exaggerating as it also contains states that depict an expanded volume but do not properly correspond
to the protein in opened states of the transient binding pocket. In summary, the MDs support the
transient nature of the binding pocket. However, they suggest a more complex distribution of
conformations along the Cys158Sγ -Val233Cα distance descriptor of the apo protein under dynamic
conditions than being observed in the ligand-bound crystal structures. We therefore also performed
simulations starting with TGT and its bound natural substrate preQ1. This ligand is hydrogen-bonded
to Asp156 within 95% of all MD frames. However, the interactions hardly trigger any of the above-
described steric pressure on Asp156 since the binding pocket is exclusively observed in closed state
(see Figures S1 and S4). This finding thus indicates that once the interaction to Asp156 is stably formed
by the substrate, the transient behavior of the binding pocket is suppressed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Mutual Cys158Sγ -Val233Cα distances and backwards push of Asp156 Cγ in different
TGT·ligand complexes with respect to TGT·1.1.
complex
Cys158Sγ -Val233Cα
Asp156 push compared
pdb-code
distance [Å]
to TGT·1.1 [Å]a)
TGT·1.1
TGT·frag
TGT·2.1
TGT·2.2
TGT·2.5
TGT·2.6
TGT·3.1
TGT·3.2
TGT·5.1
TGT·4.1
TGT·(R)-4.2
6.5
9.8
6.4
6.2
10.2
10.2
10.5
10.6
6.2
0.0
0.4
0.2
0.1
0.9
0.6
0.7
0.7
0.3
0.3
0.6
4PUK
5UTI
5J9M
5JT5
5J9N
5JT6
5J9O
5JT7
3S1G
6RKT
6RKQ
10.2
10.8
a)
The amount by which Asp156 is pushed out of the recognition pocket was determined by performing an overall RMSD Cα
atom superposition of the studied complex with TGT·1.1 and measuring the mutual distances of Asp156Cγs in both
structures.
The transformation of the protein, leading to the expansion of the binding pocket, correlates with the
properties of the substituents in 2.1 – 2.6, 3.1, and 3.2 attached to the benzimidazole or the dihydro-
imidazoquinazoline scaffold (4.1 and (R)-4.2). The primary amide stabilizes the closed conformation
found in most TGT-ligand complexes. The extension of the amide by an additional amino group to a
ACS Paragon Plus Environment

Page 17 of 39
Journal of Medicinal Chemistry
17
1
2
3
4
5
6
7
8
9
hydrazide leads to the structural transformation. The attached terminal amino group of the hydrazide
exerts steric pressure on Asp156 and thus induces the protein rearrangement leading to its
stabilization. The guanidine-substituted ligands show the same reorganization as a combined effect of
steric pressure and altered solvation of the amino acids Gly230 and Gln203. The water molecule W1
(Figure 3a) and the slight spatial shift of Asp156 are responsible for this effect. In the complexes with
the dihydro-imidazoquinazolines 4.1 and (R)-4.2 the steric pressure is exerted by the attachment at
C8. Obviously, solely a hydrogen as substituent is sufficient to trigger the pocket opening.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The newly formed cavity is occupied and stabilized by several water molecules, which indicates the
hydrophilic character of the created pocket (Figure 6). A complex water network with short distances
between 2.5 and 3.5 Å is detected involving the carbonyl groups of the Glu157, Thr159, Ala168, and
Gly230 and the backbone amide nitrogens of Gly204, Ser205, and Thr159, which protrude toward the
expanded pocket. In addition, the side chains of Gln203, Ser171, and Ser205 form H-bonds to the
accommodated water molecules.
c)
a)
b)
Figure 6: a) Binding pose of 3.1 together with the opened transient pocket, which is filled with a cluster
of water molecules. The protein is displayed by its solvent accessible surface, and the residues involved
in interactions with the water cluster are indicated, hydrogen bonds as dashed lines. b) Binding mode
of 4.1 (PDB code: 6RKT) together with the difference electron density (|Fo|-|Fc| omit map) contoured
at 3σ, protein displayed with the solvent accessible surface showing the recognition site together with
the opened transient pocket which is filled by water molecules; c) Comparison of the spatial
arrangement of computed hydration sites (numbered pink balls, Table 3) together with the
crystallographically determined waters in TGT·4.1 (orange spheres, cf. b)).
Depending on the achieved resolution, the amount of water molecules captured by crystallography
will differ.[25] Furthermore, crystal structures only provide implicit information about the
thermodynamic properties of the water molecules found in such a transient pocket. However, upon
ligand binding to the protein, the solvent molecules occupying the pocket will be displaced to the
surrounding bulk solvent phase. To estimate the contribution of the water displacement to the overall
inventory of ligand binding, the contributions resulting from the displacement of these waters are
important. We usually assume that the hydration of the transient pocket occurs instantly as soon as
the pocket undergoes the transition into the opened state. However, the solvation thermodynamic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 39
18
1
2
3
4
5
6
7
8
9
properties of the water molecules (i.e. free energy, enthalpy and entropy of solvation) in the opened
pocket must ultimately be a contribution to the thermodynamic properties of the opened transient
pocket itself. Thus, the distribution of opened and closed pocket conformations depends on the water
molecules as well as the internal molecular interactions of the protein. Waters displaced from
hydration sites that are energetically less favorable than in the bulk, will contribute favorably to ligand
binding. While displacement from hydration sites that are more favorable in the protein than in the
bulk, will counteract binding. The energetic contributions are usually counterbalanced by a favorable
entropic contribution resulting from the displacement of bound water molecules into bulk water
phase. In order to investigate the thermodynamic inventory of the water molecules in the transient
binding pocket in conjunction with possible consequences for ligand binding, we performed MD
simulations and hydration site analysis (HSA).[26] We carried out 30 ns MD simulations based on the
bound protein structure of TGT·4.1 but with the ligand removed from the binding site. Then, we
processed the MD trajectories using the HSA implementation of the SSTMap package in order to
determine the enthalpic and entropic contributions of the bound water molecules with respect to pure
bulk water (Table 3). Our calculations suggest that individual water molecules cluster at representative
sites, which reasonably well match with the crystallographically determined water molecules (Figure
6b, c). Thus, the crystal structure of TGT·4.1 reveals the presence of eight water molecules in the
transient binding pocket, while MD simulations suggests a similar cluster of nine hydration sites
populating this region. The distances between the assigned crystallographic water molecules and the
centers of the computed hydration sites are listed in Table 3.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3: Hydration site analysis of the 9 most prominent hydration sites found in the transient
recognition pocket by HSA, ranked by ΔGsolv.
Hydration
Site
Distance to
closest
ΔHsolv
TΔSsolv
ΔGsolv
[kJ·mol-1]
[kJ·mol-1]
[kJ·mol-1]
crystal
water [Å]
9
1
4
0
2
5
7
8
6
0.5
12.51
5.23
1.42
-1.17
-3.18
-2.85
-3.31
-5.27
0.63
-16.40
-22.64
-19.00
-21.34
-22.38
-20.29
-16.95
-18.24
-12.30
28.90
27.86
20.43
20.13
19.18
17.47
13.65
12.97
12.92
0.89
1.37
0.68
0.72
0.73
1.08
2.3
1.55
ACS Paragon Plus Environment

Page 19 of 39
Journal of Medicinal Chemistry
19
1
2
3
4
5
6
7
8
The HSA calculations suggest that all the waters in the transient pocket have unfavorable Gibbs free
binding energies with respect to the bulk phase. Hydration sites 9 and 1 are energetically highly
depleted, suggesting unfavorable interactions in the transient binding pocket compared to bulk
solvent. We therefore expect that their displacement favorably contributes to ligand binding. This
hypothesis is supported by the observation that the ligands, which open the transient pocket and
displace some of the water molecules in the pocket (e.g. 4.1 and (R)-4.2, note, the affinity was
determined for the racemic mixture, likely only the R-enantiomer shows inhibitory potency), are by a
factor of 10 more potent compared to the members of series 2 and equipotent to those of series 3.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Functional Role of the Transient Pocket and Outlook for Further Inhibitor Design. From the
crystallographic as well as the computational perspective, the solvation structure in the studied
binding pocket has quite a critical role for the understanding of the binding behavior as well as the
design of successive ligand molecules. For the transient pocket studied in the present contribution, it
is most remarkable that the hydration sites computed with our HSA calculations are quite unfavorable
in free energy compared to their bulk solvent state. We already observed similar behavior in the case
of human Aldose Reductase,[23] in which a transient pocket was occupied with an energetically
depleted solvent pattern. Although the present case of TGT most likely demonstrates a different
mechanism of pocket opening and closing, we reason that the very unfavorable solvation patterns are
key for the understanding of the stability of transient binding pockets. From the crystallographic point
of view, the present case of TGT does not indicate any obviously unfavorable interactions between
protein atoms at this pocket, such that the role of the solvation thermodynamics must be a critical
factor for the population of the different states of the pocket.
We inspected the Z. mobilis TGT complex structures, deposited in the Protein Data Bank (PDB,
www.rcsb.org).[27] The multiple structures show that even though the loop between Glu157 and
Tyr161 is structurally in most complexes highly conserved, the side chain of Cys158 occurred already
in a few structures with alternative conformations. In these examples, the Cys158Sγ - Val233Cα
distances vary from 6.6 Å to 9.8 Å. However, the examples with expanded distances are predominantly
mutated variants of TGT in the guanine/preQ1 recognition pocket. These variants were produced to
investigate the differences of the eubacterial and eukaryotic TGTs.[5] The latter enzymes recognize
queuine as significantly expanded substrate. In the human enzyme as in most other eukaryotic species,
Cys158 is replaced by Val and the opposing Val233 is exchanged to Gly.[28] Remarkably, the crystal
structure of queuine (PDB code: 3BLO)[5] with this variant shows a similar expansion of the recognition
pocket as observed in the complexes studied here. Possibly, the latter complexes indicate an intrinsic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 39
20
1
2
3
4
5
conformational flexibility the architecture of TGT enzymes exhibits in this area that is functionally
important for the eukaryotic species to recognize its extended substrate.
6
7
8
9
The pronounced similarity of the recognition site in the eubacterial variant, which hosts the
nucleobases guanine and preQ1, and the eukaryotic one means a real challenge developing species-
selective inhibitors. However, one observation with our expanded inhibitors might provide a promising
opportunity to approach this problem. Remarkably, the Cys158 residue is highly conserved across all
eubacterial enzymes but lacking in eukaryotic species. The thiol group of a cysteine residue is a
potential anchor to covalently attach an inhibitor, e.g. via a Michael acceptor or a potentially alkylating
group.[29–31] For the development of a potent anti-infective, irreversible inhibitors may constitute a
desired strategy to follow. The ligands studied here based on the 4-guanidino benzimidazoles or the
C(8)-functionalized dihydro-imidazoquinazoline scaffold induce the opening of the transient pocket
and make the thiol group of the crucial cysteine accessible for chemical modification. Via appropriate
design, it may be possible to develop inhibitors, preferentially based on the latter likely more potent
dihydro-imidazoquinazoline scaffold, that selectively and irreversibly block the bacterial TGT isoforms
via an electrophilic warhead capable to covalently bind to the thiol group of Cys158.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSION
Crystallographic fragment screening not only provides structural information useful for lead
development but can also unveil surprising hits pointing towards intrinsic functional properties of, e.g.,
enzymes important for some members of a given family. The screening of a 96-fragment library against
the shigellosis target TGT led to the discovery of a fragment molecule that opens a transient sub-pocket
in the preQ1 recognition site of the enzyme by pushing back Asp156 in the catalytic center. The latter
movement is associated with a reorganization of further neighboring amino acids and leads to a
conformational change of the loop between the amino acids Glu157 and Tyr161. The transient pocket,
which had been predicted before by MD simulations, causes an expansion of the preQ1 recognition
pocket volume between 250 and 750 Å3. Inspired by these observations, we designed and synthesized
a series of model compounds to target the opening of the pocket, including a dihydro-
imidazoquinazoline derivative with a propyn-1-yl exit vector that points directly into the transient
pocket and causes the displacement of a conserved water network. The latter compound, which is a
reduced analog of our initial parent ligand 1.1 is still far from being optimized and shows reduced
potency compared to 1.1. Our MD simulations and hydration site analysis used to determine the
thermodynamic contribution of the water network to the binding of the reduced ligand scaffold
showed that the displacement of the involved water molecules will contribute favorably to ligand
binding. Additionally, the Cys158 residue of the TGT binding pocket was traced to be the gatekeeper
of the transient sub-pocket. The conservation of this Cys158 in 99% of eubacterial TGT and the
ACS Paragon Plus Environment

Page 21 of 39
Journal of Medicinal Chemistry
21
1
2
3
4
5
6
7
8
9
presence of Val145 in its eukaryotic counterpart opens the possibility of developing inhibitors based
on the dihydro-imidazoquinazoline scaffold, that selectively and irreversibly block the bacterial TGT
isoforms via a warhead capable to covalently bind to the thiol group of Cys158 as the latter functional
group becomes accessible upon pocket opening.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Expression and Purification of the Z. mobilis TGT. The expression and purification protocol relates to
work published earlier by us.[32] The E. coli cells BL21-CodonPlus (DE3) -RIPL (Cam r ), transformed with
the plasmid vector pPR-IBA2-ZM10 (Amp r ), were incubated in a pre-culture of 100 mL LB medium
containing 100 mg · L-1 ampicillin and 34 mg · L-1 chloramphenicol for 17 h at 37°C and 220 rpm. In
addition to ampicillin and chloramphenicol resistance, the plasmid contains a sequence encoding an
N-terminal Strep-tag II® separated from the tgt start codon by a spacer sequence and a sequence
encoding a thrombin cleavage site. The pre-culture was added to 2 × 2 L main culture (LB medium
including 100 mg × L-1 ampicillin and 34 mg. L-1 chloramphenicol) which is incubated at 37°C and 220
rpm until the OD600 = 0.7. This main culture was then cooled to 15°C and protein expression induced
by addition of IPTG (final concentration = 1 mM). The main culture was then incubated at 15°C and 220
rpm for a further 16-18 h. Afterwards, the cell pellets were harvested through centrifugation (10,000
rpm at 4°C). The cell pellets were re-suspended in 100 mL lysis buffer (20 mM TRIS pH 7.8, 10 mM
EDTA, 1mM DTT and 2 cOmplete™-Protease Inhibitor Cocktail Tablets (Roche) per 4 L of bacterial
culture) and cell disruption was achieved via 3 rounds of sonification using Branson Sonifier 250, with
90 seconds intervals (Duty cycle 30%, output control 7). Alternatively, cell disruption was achieved via
an EmulsiFlex-C5™ high-pressure homogenizer (Avestin Europe GmbH). The soluble protein in the
supernatant was then separated by centrifugation from the insoluble cell constituents in the pellet
(centrifugation speed 19,000 rpm, 45 min, 4°C).
Purification of the protein was achieved at room temperature by two FPLC steps using an ÄKTA Purifier
LC system. A Q-Sepharose Fast Flow Anion Exchange Column (XK 26/15; GE Healthcare) was
equilibriated with buffer A (10 mM TRIS pH 7.8, 1 mM EDTA, 1mM DTT) and the clear cell lysate was
passed through the column. The protein was then eluted by buffer B (10 mM TRIS pH 7.8, 1 mM EDTA,
1mM DTT, 1M NaCl) through gradient elution (a linear increase in the proportion of buffer B (from 0 -
100% at 4 mL · min-1 column flow) and fractionation of the eluted portion. Fractions containing
the target protein Z. mobilis TGT with the Strep tag II were determined by SDS-PAGE. A Strep-Tactin®
Superflow® column (XK 16/10, IBA) was equilibrated with buffer W (100 mM TRIS pH 7.8, 1MNaCl, 1
mM EDTA) and the corresponding fractions were passed through the column. The target protein was
eluted by buffer E (100 mM TRIS pH 7.8, 1MNaCl, 1 mM EDTA, 2.5 mM D-desthiobiotin).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 39
22
1
2
3
4
5
6
7
8
9
The fractions containing TGT protein were then concentrated in a VIVASPIN®20 centrifugal
concentrator (Sartorius, MWCO = 30,000) to a concentration of approximately 2 mg · mL -1 in a high
salt buffer (10 mM TRIS pH 7.8, 2MNaCl, 1mM EDTA). Subsequently, the Strep-tag® II was cleaved off
and separated from the TGT protein via a Thrombin Cleavage Capture Kit (Novagen®) following the
manufacturer's instructions where 2.5 U of biotinylated thrombin per mg TGT protein was incubated
with the TGT protein for 16-18 h at 20°C. The cleaved TGT protein was separated from the Strep-tag®
II, the biotinylated thrombin from the kit, and the streptavidin-agarose beads by filtration using the
filters of the kit. The separated TGT protein was then dialyzed against high salt buffer and concentrated
via VIVASPIN®20 centrifugal concentrator until a final protein concentration of 12 mg · mL-1. Finally,
the protein was flash frozen into aliquots of 70 μL and stored at -80°C.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Crystallization and Structure Determination of Z. mobilis TGT
Crystallization. Z. mobilis TGT co-crystals were grown at 18°C using the sitting-drop vapor diffusion
method. A solution of 0.23 mM of Z. mobilis TGT (stored in high salt buffer) was mixed with 6.6 mM of
the respective inhibitor (stock solution in 100% DMSO) and incubated at 18°C for 1 hour. The solution
was then centrifuged and 1.5 µL of this mixture was mixed with 1.5 µL of reservoir solution (100 mM
MES pH 5.5, 1 mM DTT, 13% (w/v) PEG8000, 10% (v/v) DMSO) in the wells of a crystallization plate
containing 650 μL reservoir solution. Co-crystals could be seen within three days up to one week. They
were transferred into cryo-buffer (50 mM MES pH 5.5, 0.5mM DTT, 300mM NaCl, 2% (v/v) DMSO, 4%
(w/v) PEG 8000, 30% (v/v) glycerol) for several seconds and flash-frozen in liquid nitrogen.
Data Collection and Structural Refinement. Crystallographic data and data collection and refinement
results are summarized in Tables S1 in the Supporting Information. The diffraction data were indexed,
scaled, and merged using XDS[33] and XDSAPP.[34] Molecular replacement was performed with the
program PHASER MR[35] from the CCP4 suite[36] to determine all crystal structures. The structure 4LBU
was used as a search model. During refinement, a 5% subset of all reflections was omitted to be used
for Rfree calculation. Model building was achieved in COOT[37] and refinement using PHENIX.refine
version 1.10.1-2155.[38] Cartesian simulated annealing with default parameters was used as a first
refinement step for all the structures. This was followed by the refinement of XYZ coordinates and
occupancies of protein residues and fragments (with the exception of water molecules whose
occupancies were fixed). In the case of protein residues that gave additional density, they were refined
in double conformation and kept if their refined occupancy was ≥ 20%. Chemicalize[39] developed by
ChemAxon[40] was used for name to structure generation and SMILES code generation. The ligand PDB
and restraint files were generated with the Grade Web Server.[41]
Enzyme kinetic investigation. The inhibition constants (Ki) of the ligands with respect to the Z. mobilis
[19]
TGT were determined according to the method described by Meyer et al. (2006)
with minor
ACS Paragon Plus Environment

Page 23 of 39
Journal of Medicinal Chemistry
23
1
2
3
4
5
6
7
8
9
modifications. The concentration of Z. mobilis TGT used depended on the expected ligand affinity and
amounted to 9 – 150 nM. The concentration of tRNATyr was kept at 1.5 μM, while the concentration of
[8-3H]-guanine used was a (saturating) concentration of 10 μM. The assay buffer consisted of 100 mM
HEPES pH 7.3, 20 mM MgCl2 and 0.037% (v/v) Tween 20. A total of four aliquots of 15 μL were
withdrawn every 60 minutes from the sample and scintillation counting gave for the liquid scintillation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
counting. The final Ki value reported is the average of three measurements. The preparation of E. coli
[42]
tRNATyr (ECYC2)
was carried out via in vitro transcription using the RiboMAX™ Large Scale RNA
Production System-T7 (Promega) following the manufacturer’s instructions.
MD Simulations. We first performed MD simulations of the apo and preQ1-bound protein (using the
crystal structures 1PUD and 1P0E) to record the transient opening and closing of the binding pocket.
Overall, 3 x 200 ns and 3 x 100 ns were simulated for the apo and the preQ1-bound protein,
respectively. The applied simulation protocol is described in the Supporting Information (section S2.2
and S2.3).
The hydration site analysis also afforded molecular dynamics simulations and we used the protein-
ligand complex of TGT·4.1 (PDB code 6RKT) of this investigation with the following protocol. The
TGT·4.1 structure was loaded into MOE[43] and the ligand molecule was removed. Then, missing atoms
were built and protonation states were assigned using the protonate3D utility of MOE. The cysteine
residues coordinating the Zn2+ ion, were modelled in their deprotonated form. Then, the protonated
protein structure, crystallographic water molecules and the zinc ion were loaded into tLEaP from the
AmberTools17 package.[44] In tLEaP, parameters from the amber FF99SB[45] force field were assigned
to the protein atoms and TIP4P-Ew[46] parameters were assigned to the water molecules. Two sodium
counter ions were added using the addions2 utility in order ensure net neutrality through the MD
simulation. Finally, the protein was embedded into a truncated octahedron simulation box filled with
TIP4P-Ew water molecules. The box was build such that the minimum distance between each solute
or crystallographically determined water molecule and any box edge was at minimum 10 Å, resulting
in a simulation box with 12756 water molecules in total.
In the following, all minimization operations were carried out using pmemd and all MD simulations
were calculated using the pmemd.cuda[47–49] for use with GPUs. Both programs were, as well as all
other programs from the Amber program package, used from Amber16 together with
AmberTools17.[44]
Initially, the system energy was minimized with 2500 steps of steepest descent minimization followed
by 2500 steps of conjugate gradient minimization. During this minimization operation, all solute heavy
atoms were restrained to their crystallographically determined positions using a harmonic restraining
potential with a force constant of 25 kcal·mol-1·Å-2. Subsequently, the force constant was set to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 39
24
1
2
3
4
5
6
7
8
9
2 kcal·mol-1·Å-2 and the minimization was repeated analogously to the first one. In the next step, the
system was heat to 300 K within 25 ps using an integration time step of 1 fs. At this point, the solute
heavy atoms are again positionally restrained with a force constant of 25 kcal·mol-1·Å-2. Now, the
integration time step was switched to 2 fs and the system equilibrated under NPT conditions for 5 ns
using the Berendsen barostat[50] and a target pressure of 1 bar. The temperature was regulated at
300 K using a Langevin dynamics thermostat with a collision frequency value of  = 2 ps-1. In the final
equilibration step, the system is run under NVT conditions for an additional 5 ns.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A final production run was carried out for 30 ns and coordinates were saved to disk every 2 ps.
During all MD runs, the system was treated using periodic boundary conditions and the particle-mesh
Ewald technique together with a 9.0 Å real-space distance cutoff for the electrostatic interactions.
Furthermore, all bonds involving hydrogen atoms were constrained using the SHAKE algorithm.[50]
Hydration Site Calculations. For the HSA calculations, we used the program SSTMap[51,52] (Version
1.1.1). We used ligand molecule TGT·4.1 in its bound pose in order to define the binding site for the
HSA analysis. After the HSA calculation, the hydration site energy values were manually referenced to
the TIP4P-Ew bulk energy value.
Synthesis of the Ligands
For SMILES codes, general methods and 1H and 13C NMR spectra, see Supporting Information.
Chemical Synthesis. Materials and suppliers, the analytical instrumentation used, and the
chromatographic equipment are described in the General Methods in the Supporting Information. The
main manuscript contains the General Procedures, the complete synthesis of carboxamide 2.1,
hydrazide 2.5, guanidine 3.1, and dihydro-imidazoquinazoline (±)-4.2 and the characterization of
intermediates and target products. The complete synthesis of all other target compounds is reported
in the Supporting Information. The purity of the compounds was achieved by chromatography and was
found to be above 95% as confirmed by HPLC and elementary analysis.
General Procedure 1 (GP1). The tautomerically enriched ester (1.0 equiv) was suspended in 25% aq.
ammonia (6–18 mL), and the mixture was treated in a microwave at 100 °C for 15–30 min. An aq. sat.
NaCl solution (10 mL) was added, and the aq. layer was extracted with EtOAc or CH2Cl2 (4x 20 mL). The
combined organic layers were dried over Na2SO4, and the solvent was evaporated. FC yielded the
desired amide.
General Procedure 2 (GP2). A solution of tautomerically enriched ester (1.0 equiv) in 1 M methanolic
MeNH2 (2.4–12 mL) was treated in a microwave at 140 °C for 1.5–2.5 h. The solvent was evaporated,
the solid was dissolved in sat. NaHCO3 (20 mL) and CH2Cl2 (20 mL), and the aq. layer was extracted with
ACS Paragon Plus Environment

Page 25 of 39
Journal of Medicinal Chemistry
25
1
2
3
4
5
CH2Cl2 (3x 20 mL). The combined organic layers were dried over Na2SO4, and the solvent was
evaporated. FC yielded the desired secondary amide.
6
7
8
9
General Procedure 3 (GP3). The tautomerically enriched ester (1.0 equiv) was suspended in H2O (6 mL)
and hydrazine monohydrate (60–65%, 6 mL) or hydrazine hydrate (50–60%, 6 mL). The mixture was
treated in the microwave at 100 °C for 0.5–1 h. A white solid was filtered off, and the aq. layer was
extracted with CH2Cl2 (4x 20 mL). The combined organic layers were dried over Na2SO4, and the solvent
was evaporated. FC yielded the desired hydrazide.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General Procedure 4 (GP4). A suspension of tautomerically enriched primary amide (1.0 eq),
secondary amide (1.0 equiv), or hydrazide (1.0 equiv) in 4 M HCl in 1,4-dioxane (4–5 mL) was heated
to reflux for 16–22 h. The solvent was evaporated. FC and/or reversed-phase FC followed by
lyophilization yielded the desired product.
General Procedure 5 (GP5). A solution of the aminoaniline (1.0 equiv) in DMF was treated with the
thioisocyanate (1.1–2.4 equiv) at 90 °C and stirred at this temperature for 1–2 h (monitored by TLC).
After addition of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (1.2–1.3 equiv), the solution
was stirred at this temperature for 1–2 h (monitored by LC/MS). The mixture was cooled down to 23 °C,
filtered, and washed with EtOAc (3 x). Evaporation of the solvent, and FC gave the desired 2-
aminobenzimiadzole.
General Procedure 6 (GP6). A suspension of the nitrobenzimidazole (1.0 equiv), Pd/C (10% w/w), and
PtO2 (10% w/w) in EtOH was stirred under hydrogen atmosphere at 23 °C for 2–4 h (monitored by TLC).
The mixture was filtered over Celite® and washed with CH2Cl2/MeOH. The solvent was evaporated,
and the crude amine was directly dissolved in dry DMF. N,N'-Di-Boc-1H-pyrazole-1-carboxamidine
(1.2 eq) and N,N-diisopropylethylamine (1.5 eq) were added, and the mixture was stirred at 23 °C for
16 h (monitored by LC/MS). The solution was diluted with H2O and EtOAc, and the organic layer was
washed with H2O. The organic solvent was evaporated, and the desired protected guanidine was
obtained after FC.
General Procedure 7 (GP7). The protected guanidine (1.0 eq) was dissolved in CH2Cl2 (2 mL) and 4 M
ethereal HCl (1 mL) was added. The solution was stirred at 23 °C for 10 min. The solvent was
evaporated, and the solid was recrystallized from Et2O/MeOH.
Methyl 1H-Benzimidazole-5-carboxylate (7). A solution of 5-benzimidazole carboxylic acid (6) (19.3 g,
119 mmol) in anhydrous MeOH (500 mL) was treated dropwise with conc. H2SO4 (10 mL), and the
mixture was stirred for 18 h at 65 °C. The solvent was evaporated, and the residue was suspended in
EtOAc (50 mL). A sat. aq. NaHCO3 solution was added until pH ≈ 8 was reached. The aq. layer was
extracted with EtOAc (3x 250 mL), the combined organic layers were dried over Na2SO4, and the
solvent was evaporated. 7 (21.0 g, assumed quant.) was obtained as a brown crystalline solid. The
1
product was used in the next reaction without further purification. M.p. 135 °C; H NMR (400 MHz,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 39
26
1
2
3
4
5
6
7
8
9
(CD3)2SO): δ = 3.86 (s, 3H; COOCH3), 7.67 (s, 1H; H–C(6)), 7.84 (d, J = 8.5 Hz, 1H; H–C(7)), 8.22 (s, 1 H;
H–C(4)), 8.40 (s, 1 H; H–C(2)), 12.79 ppm (s, 1H; NH); 13C NMR (101 MHz, (CD3)2SO; broad peaks due to
tautomeric N=C-NH): δ = 52.0 (COOCH3), 111.7 (C(7)), 120.8 (C(4)), 123.3 (C(6)), 132.8 (C(5)), 136.8
(C(7a)), 142.6 (C(3a)), 144.3 (C(2)), 166.8 ppm (COOCH3); IR (ATR): ṽ = 3095 (w), 2811 (w), 1711 (s),
1625 (m), 1434 (m), 1412 (m), 1300 (s), 1223 (s), 1198 (s), 1128 (m), 1080 (s), 954 (m), 743 cm–1 (s);
+
HR-ESI-MS: m/z (%): 177.0674 (100, [M + H]+ calcd. for C9H9N2O2 : 177.0659).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Methyl 1-(N,N-Dimethylsulfamoyl)benzimidazole-5-carboxylate (8) and Methyl 1-(N,N-
Dimethylsulfamoyl)benzimidazole-6-carboxylate (8’). A solution of ester 7 (20.6 g, 117 mmol) in
toluene (300 mL) was treated with NEt3 (18.5 mL, 131 mmol) and Me2NSO2Cl (14.2 mL, 131 mmol) at
23 °C. The solution was stirred for additional 24 h at 23 °C. After evaporation of the solvent, FC (SiO2;
CH2Cl2/EtOAc 4:1) yielded a mixture of both regioisomers 8 and 8’ (21.8 g, 66%; 1:0.6 ratio) as a white
1
crystalline solid. Rf = 0.56 (SiO2; EtOAc/CH2Cl2 4:1); m.p. 118 °C; for 8: H NMR (400 MHz, (CD3)2SO):
δ = 2.91 (s, 6H; N(CH3)2), 3.89 (s, 3H; COOCH3), 7.97 (dd, J = 8.1, 0.6 Hz, 1H; H–C(7)), 8.04 (dd, J = 8.7,
1
1.6 Hz, 1H; H–C(6)), 8.33 (dd, J = 1.6, 0.6 Hz, 1H; H–C(4)), 8.79 ppm (s, 1H; H–C(2)); for 8’ H NMR
(400 MHz, (CD3)2SO): δ = 2.91 (s, 4H; N(CH3)2), 3.90 (s, 1.8H; COOCH3), 7.92 (dd, J = 8.5, 0.6 Hz, 0.6H;
H–C(7)), 8.00 (dd, J = 6.9, 1.6 Hz, 0.7H; H–C(6)), 8.41 (dd, J = 1.6, 0.7 Hz, 0.6H; H–C(4)), 8.87 ppm (s,
13
0.6H; H–C(2)); C NMR (101 MHz, (CD3)2SO): δ = 37.86, 37.87, 52.3, 52.4, 113.1, 114.0, 120.6, 121.7,
125.1, 125.8, 126.0, 126.3, 131.4, 134.7, 142.8, 144.4, 145.4, 146.2, 165.98, 166.02 ppm; IR (ATR):
ṽ = 3130 (w), 2953 (w), 1709 (s), 1618 (m), 1493 (m), 1430 (m), 1386 (s), 1317 (m), 1281 (s), 1268 (s),
1239 (m), 1166 (s), 1147 (s), 960 (s), 732 cm–1 (s); HR-ESI-MS: m/z (%): 284.0704 (100, [M + H]+, calcd.
for C11H14N3O4S+: 284.0700).
Methyl 2-Bromo-1-(N,N-dimethylsulfamoyl)benzimidazole-5-carboxylate (9). A solution of 8 and 8’
(21.4 g, 75.5 mmol) in anhydrous THF (500 mL) was treated dropwise with a solution of LiN(SiMe3)2 in
THF (1 M, 112 mL, 112 mmol) at –70 °C under a N2 atmosphere. The mixture was stirred at –70 °C for
1 h until a solution of NBS (20.3 g, 114 mmol) in THF (250 mL) was added dropwise. The solution was
stirred 15 min at –70 °C, before it was allowed to warm to 23 °C. The solvent was evaporated, and the
residue was dissolved in H2O (200 mL) and CH2Cl2 (200 mL) and the aq. layer was extracted with CH2Cl2
(3x 200 mL). The combined organic layers were dried over Na2SO4, and evaporated. FC (SiO2;
CH2Cl2/EtOAc 10:0 to 10:1) afforded 9 (15.3 g, 56%) as a pale yellow solid with traces of 9’. Rf = 0.76
1
(SiO2; EtOAc/CH2Cl2 4:1); m.p. 151 °C; H NMR (400 MHz, (CD3)2SO): δ = 3.01 (s, 6H; N(CH3)2), 3.89 (s,
3H; COOCH3), 8.01 (2 s, 2H; H–C(7), H–C(6)), 8.23 ppm (dd, J = 1.2, 1.2 Hz, 1H; H–C(4)); 13C NMR
(101 MHz, (CD3)2SO; C(2) not visible): δ = 38.2 (2C; N(CH3)2), 52.4 (COOCH3), 114.3 (C(7)), 120.3 (C(4)),
126.0 (C(6)), 128.8, 137.9 (C(7a)), 141.5, 165.8 ppm (COOCH3); IR (ATR): ṽ = 2954 (w), 1708 (s), 1464
(m), 1443 (m), 1418 (m), 1395 (m), 1300 (s), 1197 (s), 1165 (m), 1132 (m), 1028 (m), 968 (s), 719 cm–1
(s); HR-ESI-MS: m/z (%): 363.9788 (44, [M + H]+, calcd. for C11H13N3O4S81Br+: 363.97), 361.9806 (40,
[M + H]+, calcd. for C11H13N3O4S79Br+: 361.9805).
Methyl 1-(N,N-Dimethylsulfamoyl)-2-(methylamino)benzimidazole-5-carboxylate (10.1). A solution of
tautomerically enriched bromide 9 (3.00 g, 8.28 mmol) in EtOAc/MeOH (1:1; 120 mL) was treated with
1 M methanolic MeNH2 (12.4 mL, 124 mmol), and the mixture was stirred at 23 °C for 14 h. The
solvents were evaporated, and the residue was dissolved in a mixture of H2O (150 mL) and CH2Cl2
(150 mL). The aq. layer was extracted with CH2Cl2 (2x 150 mL). The combined organic layers were dried
over Na2SO4, and the solvent was evaporated to afford 10.1 (2.76 g, assumed quant.) as a crystalline
1
white solid with traces of 10.1’. Rf = 0.61 (SiO2; CH2Cl2/MeOH 10:1); m.p. 160 °C; H NMR (400 MHz,
ACS Paragon Plus Environment

Page 27 of 39
Journal of Medicinal Chemistry
27
1
2
3
4
5
6
7
8
9
(CD3)2SO): δ = 2.88 (s, 6H; N(CH3)2), 2.99 (d, J = 4.7 Hz, 3H; CH3NH), 3.84 (s, 3H; COOCH3), 6.87 (q,
J = 4.6 Hz, 1H; CH3NH), 7.63 (dd, J = 8.4, 0.6 Hz, 1H; H–C(7)), 7.68 (dd, J = 8.4, 1.6 Hz, 1H; H–C(6)),
7.79 ppm (d, J = 1.1 Hz, 1H; H–C(4)); 13C NMR (101 MHz, (CD3)2SO): δ = 30.0 (CH3NH), 38.2 (2C; N(CH3)2),
52.0 (COOCH3), 111.8 (C(7)), 116.5 (C(4)), 121.8 (C(6)), 125.3, 135.5 (C(7a)), 142.1, 154.0 (C(2)),
166.4 ppm (COOCH3); IR (ATR): ṽ = 3355 (m), 2960 (w), 1712 (s), 1626 (s), 1582 (m), 1435 (m), 1409
(m), 1379 (s), 1294 (s), 1216 (m), 1163 (m), 1144 (m), 1123 (s), 1085 (m), 1025 (m), 968 (m), 884 (m),
758 cm–1 (s); HR-ESI-MS: m/z (%): 313.0972 (100, [M + H]+, calcd. for C12H17N4O4S+: 313.0965).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1-(N,N-Dimethylsulfamoyl)-2-(methylamino)benzimidazole-5-carboxamide (11.1). A suspension of
tautomerically enriched ester 10.1 (501 mg, 1.60 mmol) was treated in 25% aq. ammonia (18 mL)
according to GP1 at 100 °C for 15 min. FC (SiO2; CH2Cl2/MeOH 10:0 to 9:1) afforded 11.1 (154 mg, 37%)
as a white crystalline solid with traces of 11.1’. Unreacted 10.1 (59 mg, 12%) was reisolated. Rf = 0.34
(SiO2; CH2Cl2/MeOH 10:1); m.p. 175 °C; 1H NMR (400 MHz, (CD3)2SO): δ = 2.87 (s, 6H; N(CH3)2), 2.98 (d,
J = 4.7 Hz, 3H; CH3NH), 6.80 (q, J = 4.6 Hz, 1H; CH3NH), 7.26 (s, 1H; CONHH), 7.55 (dd, J = 8.4, 0.6 Hz,
1H; H–C(7)), 7.58 (dd, J = 8.4, 1.6 Hz, 1 H; H–C(6)), 7.80 (dd, J = 1.6, 0.8 Hz, 1H; H–C(4)), 7.89 ppm (s,
13
1H; CONHH); C NMR (101 MHz, (CD3)2SO): δ = 30.0 (CH3NH), 38.2 (2C; N(CH3)2), 111.3 (C(7)), 115.1
(C(4)), 120.3 (C(6)), 130.2, 133.9 (C(7a)), 141.8, 153.7 (C(2)), 168.0 ppm (CONH2); IR (ATR): ṽ = 3418
(m), 1629 (s), 1576 (s), 1434 (m), 1413 (m), 1376 (s), 1260 (m), 1224 (w), 1155 (m), 1027 (m), 936 (m),
830 (w), 715 cm–1 (m); HR-ESI-MS: m/z (%): 298.0973 (100, [M + H]+, calcd. for C11H16N5O3S+: 298.0968).
1-(N,N-Dimethylsulfamoyl)-2-(methylamino)benzimidazole-5-hydrazide (11.5). A suspension of
tautomerically enriched ester 10.1 (505 mg, 1.62 mmol) in 50–60% hydrazine hydrate (6 mL) was
treated according to GP3 for 0.5 h. The organic layer was evaporated to obtain 11.5 (403 mg, 80%) as
a white crystalline powder with traces of 11.5’ and was used without further purification. Rf = 0.41
(SiO2; CH2Cl2/MeOH 10:1); m.p. 190 °C; 1H NMR (400 MHz, (CD3)2SO): δ = 2.87 (s, 6H; N(CH3)2), 2.98 (d,
J = 4.6 Hz, 3H; CH3NH), 4.46 (s, 2H; CONHNH2), 6.80 (q, J = 4.6 Hz, 1H; CH3NH), 7.53 (dd, J = 8.4, 1.6 Hz,
1H; H–C(6)), 7.56 (dd, J = 8.4, 0.7 Hz, 1H; H–C(7)), 7.72 (dd, J = 1.6, 0.7 Hz, 1H; H–C(4)), 9.68 ppm (s,
1H; CONHNH2); 13C NMR (101 MHz, (CD3)2SO): δ = 30.0 (CH3NH), 38.2 (2C; N(CH3)2), 111.5 (C(7)), 114.4
(C(4)), 119.8 (C(6)), 129.2, 133.8 (C(7a)), 141.9, 153.7 (C(2)), 166.1 ppm (CONHNH2); IR (ATR): ṽ = 3324
(m), 2911 (w), 1657 (s), 1621 (s), 1592 (s), 1577 (s), 1493 (m), 1458 (m), 1408 (m), 1376 (s), 1308 (m),
1253 (m), 1160 (s), 1122 (s), 1051 (w), 1025 (s), 967 (s), 867 (m), 715 cm–1 (s); HR-ESI-MS: m/z (%):
313.1078 (100, [M + H]+, calcd. for C11H17N6O3S+: 313.1077).
2-(Methylamino)benzimidazole-6-carboxamide (2.1). A suspension of tautomerically enriched 11.1
(96 mg, 320 µmol) in a solution of 4 M HCl in 1,4-dioxane (5 mL) was treated according to GP4 for 45 h.
FC (SiO2; CH2Cl2/7 M NH3 in MeOH 10:1) followed by reversed-phase FC (C18; H2O/THF 10:0 to 10:1)
yielded 2.1 (14 mg, 23%) as a white powder. Rf = 0.06 (SiO2; CH2Cl2/MeOH/NH4OH 10:1:0.1); m.p.
1
261 °C; H NMR (600 MHz, (CD3)2SO/CF3COOD 5:1; four exchanging NH protons not visible): δ = 3.00
(s, 3H; CH3NH), 7.37 (dd, J = 8.4, 0.6 Hz, 1 H; H–C(4)), 7.79 (dd, J = 8.4, 1.6 Hz, 1 H; H–C(5)), 7.84 ppm
13
(dd, J = 1.6, 0.6 Hz, 1H; H–C(7)); C NMR (151 MHz, (CD3)2SO/CF3COOD 5:1): δ = 29.4 (CH3NH), 110.9
(C(4)), 111.0 (C(7)), 123.4 (C(5)), 129.8, 130.1, 132.5 (C(3a)), 152.0 (C(2)), 167.8 ppm (CONH2); IR (ATR):
ṽ = 3177 (m), 1638 (s), 1587 (s), 1388 (s), 1283 (s), 1104 (w), 991 (w), 881 (w), 772 cm–1 (w); HR-ESI-
MS: m/z (%): 191.0942 (100, [M + H]+, calcd. for C9H11N4O+: 191.0855); elemental analysis calcd (%) for
C9H10N4O·0.7H2O (202.82): C 53.30, H 5.67, N 27.62; found: C 53.36, H 5.83, N 27.79.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 39
28
1
2
3
4
5
6
7
8
9
2-(Methylamino)benzimidazole-6-hydrazide (2.5). A suspension of tautomerically enriched 11.5
(101 mg, 340 µmol) in a solution of 4 M HCl in 1,4-dioxane (4 mL) was treated according to GP4 for
16 h. FC (SiO2; CH2Cl2/7 M NH3 in MeOH 10:1) followed by reversed-phase FC (C18; H2O/THF 10:0 to
10:1) yielded 2.5 (6 mg, 9%) as a white powder. Rf = 0.11 (SiO2; CH2Cl2/MeOH/NH4OH 10:1:0.1); m.p.
276 °C; 1H NMR (600 MHz, (CD3)2SO/CF3COOD 5:1; three exchanging NH protons not visible): δ = 3.03–
3.05 (m, 3H; CH3NH), 7.51 (dd, J = 8.3, 1.4 Hz, 1H; H–C(4)), 7.79 (dd, J = 8.4, 1.6 Hz, 1H; H–C(5)), 7.85
(d, J = 1.6 Hz, 1H; H–C(7)), 9.22 (q, J = 4.8 Hz, 0.9H; CH3NH), 11.53 ppm (s, 0.7H; NH); 13C NMR
((CD3)2SO/CF3COOD 5:1): δ = 29.3 (CH3NH), 110.6 (C(4)), 111.2 (C(7)), 123.2 (C(6)), 125.1 (C(3a)), 130.1
(C(5)), 133.5 (C(7a)), 152.2 (C(2)), 165.7 ppm (CONHNH2); IR (ATR): ṽ = 3124 (m), 1602 (s), 1519 (m),
1464 (m), 1420 (m), 1283 (s), 1153 (m), 992 (m), 816 cm–1 (w); HR-ESI-MS: m/z (%): 206.1041 (25,
[M + H]+, calcd. for C9H12N5O+: 206.1036); elemental analysis calcd. (%) for C9H11N5O·1.3H2O (228.64):
C 47.28, H 6.00, N 30.63; found: C 47.23, H 6.23, N 30.56.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2-(Methylamino)-5-nitrobenzimidazole (13.1). A solution of 4-nitro-1,2-diaminobenzene (12) (1.50 g,
9.80 mmol) in dry DMF (10 mL) was treated with 2-(4-morpholino)ethyl isothiocyanate (788 mg,
10.8 mmol) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (2.3 mL, 13 mmol) according to GP5.
FC (SiO2; EtOAc/MeOH 100:0 to 90:10) yielded the desired benzimidazole 13.1 (1.30 g, 69%) as an
orange solid. Rf = 0.21 (SiO2; EtOAc/MeOH 97:3); m.p. 192–193 °C; 1H NMR (400 MHz,
(CD3)2SO/CF3COOD): δ = 3.04 (s, 3H; CH3NH), 7.53 (d, J = 8.6 Hz, 1H; H–C(7)), 8.11–8.14 (m, 2H; H–C(4),
H–C(6)), 9.47 ppm (q, J = 5.0 Hz, 0.4H; CH3NH); 13C NMR (101 MHz, (CD3)2SO/CF3COOD): δ = 29.3 (CH3),
107.0 (C(4) or C(6)), 111.3 (C(7)), 119.5 (C(4) or C(6)), 139.3, 135.4, 143.1, 153.0, 153.1 ppm (C(2)); IR
(ATR): ṽ = 3294 (w), 3127 (w), 2875 (m), 1682 (m), 1615 (m), 1587 (m), 1544 (w), 1491 (m), 1476 (m),
1420 (w), 1366 (w), 1315 (m), 1263 (s), 1223 (s), 1152 (m), 1119 (m), 1064 (s), 978 (m), 944 (m), 878
(m), 856 (w), 808 (m), 757 (m), 728 (m), 701 (w), 688 (m), 622 cm–1 (w); HR-ESI-MS: m/z (%): 193.0725
+
(100, [M + H]+, calcd. for C8H9N4O2 : 193.0720).
N,N'-Bis(tert-butoxycarbonyl)-N''-[(2-methylamino)benzimidaz-5-yl]guanidine (14.1). A solution of
13.1 (200 mg, 1.04 mmol) in EtOH (10 mL) was treated with Pd/C (10% w/w, 20 mg) and PtO2 (10%
w/w, 20 mg) under a H2 atmosphere, followed by N,N-diisopropylethylamine (260 µl, 1.6 mmol) and
N,N’-di-Boc-1H-pyrazole-1-carboxamidine (388 mg, 1.25 mmol) in anhydrous DMF (10 mL) according
to GP6. FC (SiO2; EtOAc/MeOH/NH4OH 100:0:0 to 89:10:1) yielded the desired protected guanidine
14.1 (230 mg, 55%) as a white solid. Rf = 0.15 (SiO2; EtOAc/MeOH/NH4OH 89:10:1); m.p. > 280 °C
(decomp.); 1H NMR (400 MHz, CD2Cl2;two exchanging NH not visible): δ = 1.36 (s, 9H; C(CH3)3), 1.57 (s,
9H; C(CH3)3), 2.89 (s, 3H; CH3NH), 6.85 (dd, J = 8.2, 2.0 Hz, 1H; H–C(6)), 6.94 (d, J = 2.0 Hz, 1H; H–C(4)),
7.08 (d, J = 8.2 Hz, 1H; H–C(7)), 10.03 (br. s, 0.8H; NH), 11.71 ppm (br. s, 0.8H; NH); 13C NMR (101 MHz,
CD2Cl2; one signal are hidden by noise): δ = 28.36 (3C; C(CH3)3), 28.37 (3C; C(CH3)3), 29.3 (CH3NH), 80.4
(C(CH3)3), 84.5 (C(CH3)3), 109.7 (C(4)), 112.9 (C(7)), 119.4 (C(6)), 135.0, 138.7 153.7, 156.8, 157.8,
163.6 ppm; IR (ATR): ṽ = 3231 (w), 3044 (m) 2919 (m), 2786 (m) 2664 (m), 2573 (m), 1683 (s), 1582 (w),
1543 (w), 1506 (m), 1484 (m), 1441 (m), 1355 (m), 1307 (m), 1283 (w), 1245 (m), 1226 (m), 1150 (m),
1122 (w) 1094 (m), 988 (m), 954 (w), 869 (m), 852 (w), 816 (m), 780 (w), 740 (m), 678 (m), 620 cm–1
+
(w); HR-MALDI-MS: m/z (%): 405.2245 (32, [M + H]+, calcd. for C19H29N6O4 : 405.2245), 293.0998 (100,
+
[M – C8H16]+, calcd. for C11H13N6O4 : 293.1022).
N-[(2-Methylamino)-benzimidaz-5-yl]guandinium chloride (3.1). A solution of 14.1 (30 mg, 74 µmol) in
CH2Cl2 (2 mL) was treated with 4 M ethereal HCl (1 mL) according to GP7. Recrystallization from
MeOH/Et2O afforded 3.1·2HCl (12 mg, 59%) as white solid. M.p. > 250 °C (decomp.); 1H NMR (600 MHz,
ACS Paragon Plus Environment

Page 29 of 39
Journal of Medicinal Chemistry
29
1
2
3
4
5
6
7
8
9
(CD3)2SO; two exchanging NH protons not visible): δ = 3.03 (d, J = 4.8 Hz, 3H; CH3NH), 7.07 (dd, J = 8.4,
2.0 Hz, 1H; H–C(6)), 7.25 (d, J = 2.0 Hz, 1 H; H–C(4)), 7.41 (dd, J = 8.4, 0.5 Hz, 1H; H–C(7)), 7.50 (br. s,
+
4H; C(NH2 )NH2), 9.17 (s, 1H; NH), 10.07 ppm (s, 1H; NH); 13C NMR (151 MHz, (CD3)2SO): δ = 29.5
(CH3NH), 108.8 (C(4)), 111.9 (C(7)), 120.6 (C(6)), 128.8 (C(7a)), 130.1 and 130.7 (2C; C(3a), C(5)), 151.4
+
(C(NH2 )NH2), 156.6 ppm (C(2)); IR (ATR): ṽ = 2933 (br. m), 1662 (s), 1627 (s), 1598 (s), 1507 (m), 1471
(m), 1262 (m), 1187 (w), 1133 (m), 1095 (m), 1015 (w), 906 (w), 870 (w), 806 (w), 692 cm–1 (w); HR-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MALDI-MS: m/z (%): 205.1196 (51, [M + H]+, calcd. for C9H13N6 : 205.1196).
+
N,N’-(4-Bromo-1,2-phenylene)bis(4-methylbenzenesulfonamide) (19a). A solution of 4-bromo-
benzene-1,2-diamine (19) (25.4 g, 136 mmol) in pyridine (30 mL) was treated dropwise with a solution
of TsCl (54.7 g, 271 mmol) in pyridine (60 mL) over 30 min at 23 °C. After stirring the mixture for 17 h
at 65 °C, the reaction mixture was poured into a mixture of ice-water (400 mL) and aq. HCl (37 wt%,
100 mL). The resulting solid was filtered and washed with water. The crude product was suspended in
a mixture of acetic acid (630 mL) and water (70 mL) and was heated to reflux, and after filtration 19a
1
(56.2 g, 83%) was collected as a white solid. Rf = 0.54 (SiO2; cy–hexane:EtOAc 50:50); H NMR
(400 MHz, (CD3)2SO): δ = 2.36 (s, 3H; CH3), 2.37 (s, 3H; CH3), 6.93 (d, J = 8.7 Hz, 1H; H–C(6)), 7.13 (d,
J = 2.3 Hz, 1H; H–C(3)), 7.19 (dd, J = 8.7, 2.3 Hz, 1H; H–C(5)), 7.37 (dd, J = 8.4, 8.4 Hz, 4H; H–C(3’)), 7.60
13
(dd, J = 8.4, 6.8 Hz, 4H; H–C(2’)), 9.41 (br. s, 1H; NH), 9.48 ppm (br. s, 1H; NH); C NMR (101 MHz,
(CD3)2SO): δ = 21.0 (2C; CH3), 117.4 (C(4)), 125.0 (2C; C(3), C(6)), 126.89 (2C; C(2’)), 126.91 (2C; C(2’)),
128.3 (C(5)), 128.6 (C(2)), 129.82 (2C; C(3’)), 129.87 (2C; C(3’)), 131.3 (C(1)), 135.8 (C(1’)), 135.9 (C(1’)),
143.8 (C(4’)), 144.0 ppm (C(4’)); IR (ATR): ṽ = 3289 (w), 3241 (w), 2922 (w), 1910 (w), 1591 (w), 1583
(w), 1492 (m), 1432 (w), 1389 (m), 1336 (s), 1311 (w), 1293 (w), 1260 (m), 1184 (m), 1161 (s), 1120 (m),
1092 (s), 1024 (w), 955 (w), 930 (m), 894 (w), 873 (m), 856 (m), 810 (m), 721 (m), 704 (w), 666 (s),
+
629 cm–1 (m); HR-EI-MS: m/z (%): 497.0022 (72, [M + H]+, calcd. for C20H2081BrN2O4S2 : 497.0023),
+
495.0041 (64, [M + H]+, calcd. for C20H2079BrN2O4S2 : 495.0042), 369.9684 (100, calcd. for
C13H1181BrN3O3S+: 369.9679), 367.9703 (100, calcd. for C13H1179BrN3O3S+: 367.9699).
N,N’-(4-Bromo-5-nitro-1,2-phenylene)bis-4-methylbenzene-sulfonamide (19b). A solution of 19a
(30.0 g, 60.6 mmol) in acetic acid (120 mL) was treated with a mixture of acetic acid (2.7 mL) and conc.
nitric acid (2.0 mL). After stirring the mixture for 1 h at 50 °C, the reaction mixture solidified. The solid
was filtered and washed with water (10 mL). Recrystallization from acetic acid (1.1 L) gave the nitro
compound 19b (21.8 g, 67%) as a beige solid. Unreacted starting material 19a (7.20 g, 24%) was
reisolated by evaporation of the filtrate of the recrystallization. Rf = 0.11 (SiO2; cy–hexane:EtOAc
50:50); 1H NMR (400 MHz, (CD3)2SO): δ = 2.37 (s, 6H; CH3), 7.39 (dd, J = 8.1, 8.1 Hz, 4H; H–C(3’)), 7.47
(s, 1H; H–C(3)), 7.63 (d, J = 8.4 Hz, 2H; H–C(2’)), 7.67 (d, J = 8.4 Hz, 2H; H–C(2’)), 7.74 (s, 1H; H–C(6)),
13
9.82 ppm (br. s, 2H; NH); C NMR (101 MHz, (CD3)2SO): δ = 21.1 (2C; CH3), 110.8 (C(2)), 120.6 (C(3)),
124.2 (C(6)), 126.89 (2C; C(2’)), 126.91 (2C; C(2’)), 127.1 (C(1)), 130.0 (2C; C(3’)), 130.1 (2C; C(3’)), 135.6
(C(1’)), 135.7 (C(1’)), 135.8 (C(4)), 143.2 (C(5)), 144.2 (C(4’)), 144.3 ppm (C(4’)); IR (ATR): ṽ = 3253 (m),
1599 (w), 1579 (m), 1526 (m), 1488 (m), 1434 (m), 1410 (w), 1376 (m) 1327 (s), 1289 (m), 1272 (m),
1185 (w), 1162 (s), 1089 (s), 1019 (w), 985 (w), 908 (m), 866 (m), 835 (m), 812 (s), 750 (w), 723 (m),
689 (w), 666 (s), 636 (w), 609 cm–1 (w); HR-EI-MS: m/z (%): 563.9692 (51, [M + Na]+, calcd.
81
+
79
+
C20H9 BrN3NaO6S2 : 563.9693), 561.9715 (45, [M + Na]+, calcd. for C20H9 BrN3NaO6S2 : 561.9713),
541.9872 (26, [M + H]+, calcd. for C20H1081BrN3O6S2 : 541.9874), 539.9893 (24, [M + H]+, calcd. for
C20H1079BrN3O6S2 : 539.9893).
+
+
ACS Paragon Plus Environment