Challenges in the Development of a Thiol-Based Broad-Spectrum Inhibitor for Metallo-β-Lactamases

Article abstract of DOI:10.1021/acsinfecdis.7b00129

Pathogens, expressing metallo-β-lactamases (MBLs), become resistant against most β-lactam antibiotics. Besides the dragging search for new antibiotics, development of MBL inhibitors would be an alternative weapon against resistant bacterial pathogens. Inhibition of resistance enzymes could restore the antibacterial activity of β-lactams. Various approaches to MBL inhibitors are described; among others, the promising motif of a zinc coordinating thiol moiety is very popular. Nevertheless, since the first report of a thiol-based MBL inhibitor (thiomandelic acid) in 2001, no steps in development of thiol based MBL inhibitors were reported that go beyond clinical isolate testing. In this study, we report on the synthesis and biochemical characterization of thiol-based MBL inhibitors and highlight the challenges behind the development of thiol-based compounds, which exhibit good in vitro activity toward a broad spectrum of MBLs, selectivity against human off-targets, and reasonable activity against clinical isolates.

Full text of DOI:10.1021/acsinfecdis.7b00129

Subscriber access provided by READING UNIV
Article
Challenges in the Development of a Thiol-Based
Broad-Spectrum Inhibitor for Metallo-#-Lactamases
Dominik Büttner, Jan S Kramer, Franca Maria Klingler, Sandra K. Wittmann, Markus
Hartmann, Christian Kurz, Daniel Kohnhäuser, Lilia Weizel, Astrid Brüggerhoff, Denia
Frank, Dieter Steinhilber, Thomas A Wichelhaus, Denys Pogoryelov, and Ewgenij Proschak
ACS Infect. Dis., Just Accepted Manuscript • DOI: 10.1021/acsinfecdis.7b00129 • Publication Date (Web): 26 Nov 2017
Downloaded from http://pubs.acs.org on November 28, 2017
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 free 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 accessible to all
readers and 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.
ACS Infectious Diseases 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 44
ACS Infectious Diseases
1
2
3
4
5
6
Challenges in the Development of a ThiolꢀBased BroadꢀSpectrum
Inhibitor for MetalloꢀβꢀLactamases
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
Dominik Büttner†, Jan S. Kramer†, FrancaꢀM. Klingler†, Sandra K. Wittmann†,
Markus R. Hartmann†, Christian G. Kurz†, Daniel Kohnhäuser†, Lilia Weizel†,
Astrid Brüggerhoff†, Denia Frank‡, Dieter Steinhilber†, Thomas A. Wichelhaus‡,
Denys Pogoryelov*§, Ewgenij Proschak*†
†
§
Institute of Pharmaceutical Chemistry, and Institute of Biochemistry, Goethe University
Frankfurt, MaxꢀvonꢀLaueꢀStraße 9, 60438 Frankfurt, Germany; Email: pogoryelov@em.uniꢀ
frankfurt.de or proschak@pharmchem.uniꢀfrankfurt.de
‡Institute of Medical Microbiology and Infection Control, Goethe University Hospital, Paulꢀ
EhrlichꢀStraße 40, 60596 Frankfurt, Germany
Abstract: Pathogens, expressing metalloꢀβꢀlactamases (MBLs), become resistant against
most βꢀlactam antibiotics. Besides the dragging search for new antibiotics, development
of MBL inhibitors would be an alternative weapon against resistant bacterial pathogens.
Inhibition of resistance enzymes could restore the antibacterial activity of βꢀlactams.
Various approaches to MBL inhibitors are described, among others, the promising motif
of a zinc coordinating thiol moiety is very popular. Nevertheless, since the first report of a
thiolꢀbased MBLꢀInhibitor (thiomandelic acid) in 2001, no steps in development of thiol
based MBL inhibitors were reported that go beyond clinical isolate testing. In this study
we report on the synthesis and biochemical characterization of thiolꢀbased MBL inhibitors
and highlight the challenges behind the development of thiolꢀbased compounds, which
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 2 of 44
1
2
3
exhibits good in vitro activity towards a broad spectrum of MBLs, selectivity against
4
5
6
human offꢀtargets, and reasonable activity against clinical isolates.
7
8
9
Keywords: Bacterial resistance, metalloꢀβꢀlactamases, metalloꢀβꢀlactamase inhibitor,
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
synthesis, thiolꢀbased, biochemical characterization
Today, it is usual to treat a bacterial infection with βꢀlactam antibiotics, which are the
most used antibacterial drugs.1 However, the efficacy of this drug class is restricted more
and more by the increasing spread of bacterial resistance worldwide.2 βꢀlactamases are the
most common way of bacterial resistance against βꢀlactam antibiotics. These enzymes
hydrolyze the βꢀlactam ring which is crucial for antibacterial activity and release the
inactive hydrolyzed product.3 βꢀLactamase expressing, Gramꢀnegative bacteria among the
ESKAPE4 pathogens become resistant to most classes of antibiotics. Thus, nosocomial
infections caused by these pathogens become very poorly treatable.5
βꢀLactamases can be categorized in two groups based on their catalytic mechanism: the
serineꢀβꢀlactamases (SBLs) which cleave the βꢀlactam ring by a nucleophilic attack of the
serine residue6 and the metalloꢀβꢀlactamases (MBLs) that contain one or two zinc ions
inside the active center.7 MBLs belong to class B, which is further divided in three subꢀ
classes (B1, B2 and B3).8 According to the proposed catalytic mechanism, MBLs
coordinate the βꢀlactam scaffold inside the active center, and hydrolyze it by a
nucleophilic water molecule, which is polarized by the zinc ions. In contrast to SBLs,
there is no covalent interaction between enzyme and ligand. Thus, MBLs work much
faster than SBLs, and more like chemical metalꢀcatalysts.9,10
Since we are confronted with a drying pipeline of new antibiotics, it would be useful to
have an alternative defense against multiꢀresistant bugs. One approach is the development
ACS Paragon Plus Environment

Page 3 of 44
ACS Infectious Diseases
1
2
3
4
5
of βꢀlactamase inhibitors, which have usually no intrinsic bactericidal effect, but could
restore the activity of βꢀlactam antibiotics by inhibiting βꢀlactamases.
6
7
8
9
Whereas numerous SBL inhibitors are already in clinical use, this does not apply to
inhibitors of MBLs.11 Literature provides various approaches, and an even higher number
of chemical scaffolds for the design of MBL inhibitors. These can be categorized by their
mode of inhibition. Many of them contain dicarboxylate structures (1), which are
coordinating the zinc ions.12 Cyclic boronates (2) represent an elegant transition state
mimetic approach.13 Multivalent chelators like EDTA (3) remove the zinc ions from the
active site of MBLs.14,15 Some MBLs can be inhibited by a covalent adduct to Lys211
which is generated in situ during the hydrolysis of cefaclor (4)16 (Figure 1).
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 1. Examples for known MBL inhibitors: A maleic acid derivate (1), a cyclic boronate (2),
EDTA (3) and cefaclor(4).
Thiol containing compounds were found as an interesting scaffold for the inhibition of
MBLs; as thiomandelic acid (5) was discovered as a broadꢀspectrum inhibitor for MBLs.
CdꢀNMR studies demonstrated that the sulfur atom of the compound intercalates the metal
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 4 of 44
1
2
3
4
5
6
7
8
9
ions within the active site and displaces the bridging water molecule that is essential for
the MBL hydrolase activity.17,18 This binding mode was confirmed by crystal structures of
Bacillus cereus MBL (BcII) and 519 and also of VIMꢀ2 (Verona integrinꢀencoded metalloꢀ
βꢀlactamase) and a mercaptocarboxylate inhibitor (6).20
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
Skagseth et al. published investigations concerning opportunities to introduce bioisosteric
groups to mercaptocarboxylic acid derived compounds. In summary, replacement of the
thiol group by a hydroxyl function was not tolerated by any MBL. Substitution of the
carboxylic acid by phosphonic acid or phosphonate was both accepted in principal, but
only the combination of phosphonate and unprotected thiol (7) led to subꢀmicromolar
inhibitory activity.21
A different approach to thiolꢀbased MBL inhibitors was followed by González et al who
characterized bisthiazolidines as a novel scaffold with enhanced activity on all three
subclasses of MBLs.22–24 Due to their bicyclic structure and the tetrahedral oriented
tertiary nitrogen, bisthiazolidines, represented by LꢀCS319 (8), are substrateꢀmimicking
MBL inhibitors (Figure 2).
Figure 2. Chemical structures of thiomandelic acid (5), 2ꢀmercaptoꢀ5ꢀphenylpentanoic acid (6),
diethyl (1ꢀmercaptoꢀ5ꢀphenylpentanꢀ2ꢀyl)phosphonate (7) and the bisthiazolidine LꢀCS319 (8).
ACS Paragon Plus Environment

Page 5 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
Besides the search for new chemical structures with inhibitory activity against MBLs,
already approved drugs were investigated as MBL inhibitors. First of all, Captopril (9)
was found as potent inhibitor for New Delhi metalloꢀβꢀlactamase (NDMꢀ1, 7.9 µM),25 and
successfully coꢀcrystalized with NDMꢀ1 (PDB: 4EXS). Based on this crystal structure, the
binding mode of 9 and NDMꢀ1 could be determined. Similar to the binding mode of 5, the
sulfur atom of 9 intercalates between the zinc ions and additionally, the acid group
interacts with Asn220.26 Recent investigations demonstrated that several approved drugs,
like Tiopronin (10) and Thiorphan (11), which contain thiols, are able to inhibit MBLs in
a lower micromolar range (Figure 3). Furthermore, the coꢀcrystallization of NDMꢀ1 and
10, a drug for the treatment of heavy metal poisoning, indicates that the described binding
mode for 9 is true for other thiol containing structures.27
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 3. Thiol containing drugs with inhibitory activity in MBLs: Captopril (9), Tiopronin (10) and
Thiorphan (11).
The general validity of the binding mode of 9 on other B1 MBLs was proved by further
crystal structures, including VIMꢀ2 (Verona integrinꢀencoded metalloꢀβꢀlactamase, PDB:
4C1D), and IMPꢀ1 (Imipenem hydrolyzing metalloꢀβꢀlactamase, PDB: 4C1F). Attention
must be paid to the fact that the residue, which interacts with the carboxylic group, differs,
depending on the respective MBL and on the stereochemistry of the inhibitors carboxylic
group.28
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 6 of 44
1
2
3
4
5
6
7
8
9
The structure of 9 seemed to be a promising chemical landmark for the further
development of MBL inhibitors. Li et al. synthesized a library of simplified Captopril
analogues, where they substituted the pyrrolidine ring by various amideꢀcoupled residues,
and tested them on their inhibitory potency against NDMꢀ1. Li et al. could demonstrate
that the carboxylic acid, located at the pyrrolidine ring of 9, is highly important for
activity. Removal of the carboxylic acid goes along with a dramatically decreased
potency, right up to complete loss of activity.29
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
Lienard et al. synthesized thiol derivatives of diverse amino acids. Compound 12
displayed good Ki values on distinct members of each MBLꢀsubclass,30 what was
supplemented by inhibitory activities within a bacterial imipenemꢀinduced growth
inhibition assay for the structurally related compound 13.31 Latest research yielded in
compound 14 which combined tryptophan and a (S)ꢀ3ꢀmercaptoꢀ2ꢀmethylpropanyl moiety
of Captopril and showed good IC50 values on IMPꢀ1 and several VIM mutants32 (Figure
4).
Figure 4. Chemical structures of 12, 13 and 14.
Taken together, there are strong reasons in favor of development of thiolꢀbased MBL
inhibitors and a large number of potent compounds was described in literature. However,
the aim to develop a clinically useful thiolꢀbased MBL inhibitor seems to be far away.
Although most of the presented inhibitors showed good in vitro activity, they often lacked
in broadꢀspectrum activity, or were tested on only a minor number of MBLs. Furthermore,
ACS Paragon Plus Environment

Page 7 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
promising inhibitors failed or were not evaluated against clinically relevant isolates.
Finally, most of the inhibitors were more or less structurally related to approved drugs and
therefore addressing human enzymes as well. Nevertheless, almost none of the showed
compounds were investigated in the context of the relevant human targets.
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 aim of this study is to perform a broad investigation of a series of thiolꢀbased MBL
inhibitors, the latter mentioned factors, and to highlight opportunities how to handle them.
Therefore, we designed a series of thiol containing inhibitors, based on the chemical
structure of 9. We modified both, the pyrrolidine ring, including the position of the acidic
group, and the methyl group. Interactions, between inhibitors and MBLs (NDMꢀ1, VIMꢀ
1/2 and IMPꢀ7) were investigated by several biochemical and biophysical methods.
Furthermore, we examined the logD7.4 values and activity in clinical isolates as well as
offꢀtarget activity of selected compounds in order to investigate the challenges and
opportunities in development of a thiolꢀbased MBL inhibitor.
RESULTS AND DISCUSSION
The synthetic procedure to substitute the methyl group by different aromatic residues is
outlined in Scheme 1. Frost et al. described an elegant method to prepare acrylic esters,
starting from an aldehyde.33 Therefore, benzaldehyde derivatives 15b and 15c or 2ꢀ
thenaldehyde (15d) reacted with Meldrum´s acid (16) in an acidꢀcatalyzed, Knoevenagelꢀ
like reaction at 75 °C for 2 h. The condensation product (17bꢀd) was reduced towards
18bꢀd using in situ generated NaHB(OAc)3 in ice bath cooled DCM. The corresponding
acrylic ester derivatives 19bꢀd was obtained by the reaction of 18bꢀd with Eschenmoser´s
salt (20) in methanol at 70 °C for 18 h.
Ester hydrolysis of 19bꢀd was performed using aqueous NaOH and heating to 100 °C for
1 h under microwave irradiation. The resulting acrylic acid (21bꢀd) was activated through
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 8 of 44
1
2
3
4
5
6
7
8
9
by thionyl chloride in methylene chloride at 40 °C for 3 h. The following amide coupling
reaction with 22 and DIPEA was carried out in 1,2ꢀdichloro ethane at 90 °C for 1 h under
microwave irradiation. The obtained acrylic amides (23bꢀd) reacted with thioacetic acid in
a Michael addition at room temperature for three days to yield 24bꢀd. Synthesis of 25a
was accomplished as shown in Scheme 2. Final ester hydrolysis were carried out with
TFA in methylene chloride at 40 °C overnight and subsequently treated with aqueous
ammonia for 3 h at room temperature to obtain 25aꢀd.
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: Synthesis route for derivatives 25aꢀd.a
aReaction conditions: (a) 15bꢀd, AcOH, iPrOH, 75 °C, 2 h; (b) NaBH4, AcOH, DCM, 0 °C, 1 h, r.t., 1
h; (c) 20, MeOH, 70 °C, 18 h; (d) NaOH, H2O, MW, 100 °C, 1 h; (e) SOCl2, 40 °C, 3 h and 22,
DIPEA, 1,2ꢀDCM, MW, 90 °C, 1 h; (f) AcSH, r.t., 3 d; (g): TFA, DCM, 40 °C, o.n. and NH3 (aq), r.t.,
3 h.
ACS Paragon Plus Environment

Page 9 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
The synthesis routes for the 2ꢀbenzylꢀ3ꢀmercaptopropionic acid derivatives are shown in
Scheme 2. Amino acids (26a,b,dꢀf) were suspended in tertꢀbutyl acetate and treated with
perchloric acid to obtain the corresponding tertꢀbutyl esters (27a,b,dꢀf). The carboxylic
acid group of the commercial available starting compound 3ꢀ(acetylthio)ꢀ2ꢀ
benzylpropanoic acid (28) was activated by thionyl chloride in dichloromethane at 40 °C.
The corresponding acid chloride (29) was directly subjected to amide coupling with the
protected amines (27a,b,dꢀf) in ACN (or ACN/DMF) using potassium carbonate (or
DIPEA) as base at room temperature overnight. The resulting amides (30a,b,dꢀf) were
heated with TFA in methylene chloride overnight and subsequently treated with aqueous
ammonia for 3 h at room temperature to obtain the final 1ꢀ(2ꢀbenzylꢀ3ꢀ
mercaptopropanoyl)amino acid derivatives (31a,b,dꢀf). For unprotected amino acids
(32c,gꢀi), an alternate amide coupling reaction (33c,gꢀi) was performed in a mixture of
ACN and DMF (or water/DMF) using DIPEA (or NaHCO3) as base at room temperature
overnight. The final cleavage of the thioester was carried out in aqueous ammonia (31c,gꢀ
i).
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

ACS Infectious Diseases
Page 10 of 44
1
2
3
Scheme 2: Synthesis routes for introduction of different amino acids, starting from 28.a
4
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
aCompounds 31a,b,dꢀf and 25a were synthesized via route A, and 31c,gꢀi via route B. Reaction
conditions (h1/h2) TBAC, HClO4, 0 °C, r.t., o.n.; (i) SOCl2, DMF, DCM, 40 °C, 3 h; (j1) K2CO3,
ACN, r.t., o.n.; (j2) DIPEA, ACN/DMF, r.t., o.n.; (k1) TFA, DCM, 40 °C, o.n. and NH3 (aq), r.t., 3 h;
(k2) HCl (37 %) reflux, 3 h; (k3) NH3 (aq), r.t., 3 h; (l1) DIPEA, ACN/DMF, r.t., o.n. (l2) NaHCO3,
H2O/ACN, r.t., o.n.
The preparation of 4ꢀNꢀsubstituted piperazineꢀ2ꢀcarboxylic acid building blocks is shown
in Scheme 3. 1ꢀNꢀbocꢀpiperazineꢀ2ꢀcarboxylic acid (34) served as a starting material for
both derivatives 35a and 35b. Benzoic acid was preꢀstirred with EDC and HOBt in DCM
for 1 h and subsequently reacted with 34 at room temperature for 72 h. The crude
intermediate was treated with trifluoroacetic acid in DCM at 40 °C for 3 h to yield 35a.
Phenyl isocyanate and 34 were mixed in dioxane and kept at 110 °C for 30 minutes under
microwave irradiation to obtain the tertiary urea 36. The benzyl ester 37 was introduced
by the reaction of 36 with benzyl bromide and K2CO3 in ACN at room temperature for
ACS Paragon Plus Environment

Page 11 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
8 h. The tertꢀbutyloxycarbonyl protecting group was hydrolyzed with TFA in DCM at 40
°C for 3 h to receive the free amine 35b. The respective amide coupling reactions (36a,b)
were performed as described above (Scheme 2). To obtain 37b, as distinct from the
remaining derivatives, final ester cleavage was carried out with aqueous LiOH at room
temperature overnight. (For further synthetic details see 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
Scheme 3: Synthesis of 35a and 35b.a
O
O
NH
OH
o
n
N
N
N
N
N
O
H
H
Boc
N
O
N
O
Boc
Boc
OH
OBn
34
36
37
O
O
N
N
N
p
m
H
HN
O
HN
O
OH
OBn
35a
35b
aReaction conditions: (m) benzoic acid, EDC, HOBt, DCM, r.t., 1 h and 34, r.t., 72 h; (n) PhCNO,
dioxan, MW, 110 °C, 0.5 h; (o) K2CO,3, BnBr, ACN, r.t., 8 h; (p) TFA, DCM, 40 °C, 3 h. 35a was
further reacted according route B and 35b according route A (Scheme 2).
The MBL inhibitory potential of the synthesized compounds was investigated in a kinetic
fluorescenceꢀbased activity assay using Fluorocillin™ as substrate. The IC50 values of
D/LꢀCaptopril (9) and compounds 31aꢀd are listed in Table 1. Captopril showed similar
activities towards all tested MBLs with slight preference towards IMPꢀ7. In comparison,
the introduction of the unsubstituted phenyl ring (25a) led to comparable values on
NDMꢀ1 and IMPꢀ7, but resulted in a slight loss of activity on VIMꢀ1. Additional
substituents on the phenyl ring led to further decrease in activity. While the 2ꢀ
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 12 of 44
1
2
3
4
5
6
7
8
9
methoxyphenol derivative (25c) was not well tolerated by all three MBLs, the brominated
compound 25b showed only slightly decreased IC50 values. On the contrary, introduction
of the thiophene ring (25d) led to a gap in activity on NDMꢀ1 and IMPꢀ7, but in
comparison to the derivatives containing a phenyl ring, an increased activity on VIMꢀ1.
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 1. In vitro activity of compounds 9 and 25aꢀda
IC50 [µM]
R1
NDM-1
VIM-1
IMP-7
9
H
6.4 ± 2.50
6.8 ± 1.20
2.9 ± 0.90
25a
3.2 ± 0.29
6.2 ± 0.25
16.3 ± 2.63
16.1 ± 4.2
4.6 ± 0.17
7.7 ± 1.65
25b
25c
25d
25.2 ± 4.66
17.0 ± 1.09
23.1 ± 2.37
6.8 ± 0.74
30.4 ± 5.43
8.7 ± 1.73
aResults are mean ± SD of three different experiments.
The IC50 values of derivatives 31aꢀi, containing different amino acids, are shown in Table
2. All substitution patterns were well tolerated on NDMꢀ1 and showed a trend towards
improved activity in comparison with Captopril. Compound 31d was characterized by its
subꢀmicromolar activity, especially in comparison with its enantiomer 31e. The Sꢀ
ACS Paragon Plus Environment

Page 13 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
configured alanine (31a), the thiazolidine ring (31h) and likewise the four membered ring
(31c) were not well tolerated by VIMꢀ1. The sarcosine substructure (31b) and all
piperidine derivatives (31dꢀg, i) showed slightly increased activities. On IMPꢀ7, the
piperidine motive displayed the best inhibitory activities, independent from position and
absolute configuration of the carboxylate. Compound 31i, containing a sulfone group,
which is bioisosteric to carboxylic acids, showed reasonable activity on NDMꢀ1 and VIMꢀ
1, but not on IMPꢀ7.
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. In vitro activity of compounds 31aꢀia
IC50 [µM]
R2
NDM-1
VIM-1
IMP-7
31a
31b
1.7 ± 0.15
12.1 ± 1.48
6.1 ± 0.90
2.6 ± 0.71
2.2 ± 0.48
4.9 ± 0.70
n.d.b
4.7 ± 0.79
7.4 ± 1.06
31c
31d
N
O
0.7 ± 0.05
4.5 ± 0.95
4.6 ± 0.85
4.9 ± 0.62
0.9 ± 0.06
1.1 ± 0.06
OH
31e
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 14 of 44
1
2
3
4
5
6
7
8
9
31f
1.6 ± 0.25
1.8 ± 0.51
1.4 ± 0.16
2.2 ± 0.03
1.6 ± 0.09
31g
2.6 ± 0.32
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
31h
31i
6.7 ± 0.57
2.4 ± 0.51
16.2 ± 2.81
2.9 ± 0.27
2.9 ± 0.27
7.9 ± 0.48
aResults are mean ± SD of three different experiments; bnot determinable.
Both piperazine derivatives 37a and 37b together with their IC50 values are presented in
Table 3. While the activities of 37a, what carries a benzoyl group in position 4, were
comparable with that of 31d and 31e on NDMꢀ1 and VIMꢀ1, the tertiary urea moiety of
37b was not tolerated on NDMꢀ1. However, both compounds provided a further improved
activity on IMPꢀ7 and were well tolerated by VIMꢀ1.
Table 3. In vitro activity of compounds 37a and 37ba
IC50 [µM]
R3
NDM-1
VIM-1
IMP-7
ACS Paragon Plus Environment

Page 15 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
37a
37b
2.5 ± 0.33
4.5 ± 0.69
0.3 ± 0.01
0.3 ± 0.25
47.7 ± 9.48
7.0 ± 0.78
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
aResults are mean ± SD of three different experiments.
In summary, derivatives containing the piperidine substructure provided the best overall
activity on all three tested MBLs. On NDMꢀ1, the absolute configuration of the carboxylic
acid seemed to be of special importance regarding the inhibitory potency, as 31d provided
a sixꢀtime better activity compared to 31e. On VIMꢀ1 and IMPꢀ7, this effect was not
found. Molecular docking analysis of 31d in complex with NDMꢀ1 reveals the possible
reason for this behavior – an ionic interaction of the carboxylate moiety towards Lys211
which is unique for NDMꢀ1 (see Supporting Information, Figure S2). The introduction of
additional substituents to the piperazine ring improved the inhibition of IMPꢀ7. According
to the activities of the r and s configured 31d and 31e on NDMꢀ1, it could be worthwhile
to prepare the piperazine ring with an absolute orientation of the acid group. Likewise, a
shift of the acid group towards positon 3 could be a promising approach for further
studies.
Compounds 9, 25a, 31d and 37b were further investigated by differential scanning
fluorimetry (DSF) (Table 4). Neither of the synthesized compounds was able to stabilize
VIMꢀ1 in a significant manner, except 9, which showed a weak stabilization. On IMPꢀ7,
the shifts of melting points correlated well with the corresponding IC50 values. As
expected, the highest stabilization was found for the most active compound 37b. In
contrast, the thermal shifts on NDMꢀ1 did not correlate with the determined IC50 values.
The enzyme was stabilized most by 37b, which exhibited the weakest potency in the
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 16 of 44
1
2
3
4
5
activity assay; in contrast to this 31d, which had 68ꢀfold better inhibition values on NDMꢀ
1 compared to 37b, thermal shift was determined as half as strong.
6
7
8
9
Table 4. Thermal shifts (in ∆ °C)a
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
NDM-1 b
VIM-1 b
IMP-7 b
9
1.9 ± 0.6
0.9 ± 0.5
1.3 ± 0.2
2.6 ± 0.3
1.5 ± 0.5
0.2 ± 0.4
0.7 ± 0.6
0.7 ± 0.6
2.5 ± 0.2
1.8 ± 0.2
2.7 ± 0.3
6.0 ± 0.2
25a
31d
37b
aResults are mean ± SD of three different
b
experiments. Melting points without inhibitor
[in °C]: 56.3 ± 0.1 (NDM-1), 66.1 ± 0.6 (VIM-1)
and 72.5 ± 0.3 (IMP-7). Results are mean ± SD of
three different experiments.
The outcomes from DSF experiments, concerning the stabilization of NDMꢀ1 by 31d and
37b, required a more closely look at this enzymeꢀligand interactions by isothermal
titration calorimetry (ITC) experiments (Figure 5 and Supporting Information, Figure S1).
9, 31d and 37b exhibit comparable Kd values in low micromolar range (Table 5).
However, while 9 displays a fully enthalpically driven binding, the binding event between
NDMꢀ1 and 31d was mainly entropically driven.34 This could have been an explanation
for the lower thermal shift of NDMꢀ1 caused by 31d. On the other hand, binding of 37b
was found as distinctly more enthalpically driven, which led to an increased thermal shift.
Although 37b provided a poorer IC50 value towards NDMꢀ1 in the activity assay, the 4ꢀ
benzoyl piperazine substructure seemed to offer additional interaction points towards the
ACS Paragon Plus Environment

Page 17 of 44
ACS Infectious Diseases
1
2
3
4
5
enzyme, what could than again lead, with the right substitution pattern, to an increased
activity.
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
Figure 5. Titration curves of ITC experiments. A: NDMꢀ1: 130 µM, 31d: 700 µM, 10.98 µL/titration
step; B: NDMꢀ1: 100 µM, 37b: 1000 µM, 10.98 µL/titration step.
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 18 of 44
1
2
3
4
5
6
7
8
9
Table 5. Dissociation constant, binding enthalpies and entropies of 9, 31d and 37b at
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
25°C
Kd [µM]
∆G [kJ/mol] ∆H [kJ/mol] -T∆S [kJ/mol]
n
9
17.51
-27.15
-40.16
13.01
1.21
31d
37b
7.84
-29.14
-27.78
-4.86
-24.28
-12.99
0.85
0.89
13.16
-14.79
The direct binding of 31d to MBLs was confirmed by coꢀcrystallization and Xꢀray
structure determination of VIMꢀ2 from Pseudomonas aeruginosa in complex with 31d.
To avoid bias in model interpretation and ligand identification, the structure of VIMꢀ2 at
1.5 Å resolution was obtained ab initio, by single wavelength anomalous dispersion
(SAD) using zinc atoms as anomalous scatterers. The obtained crystals in the space group
C2 contained two VIMꢀ2 molecules in the asymmetric unit (Supporting Information). In
the initial mFoꢀDFc electon density map, obtained after SAD phasing and automatic
model building, the pronounced bulky positive electon densities were identified in the
vicinity of dizincꢀcontaining active sites of both VIMꢀ2 molecules of the asymmetric unit.
The properties of the obtained 2mFoꢀDFc (Figure 6a) maps allowed unambiguous
placement of 31d molecule in the unique orientation in the active sites of both VIMꢀ2
molecules. The correctness of the ligand placement in the final model was further
corroborated by the polder omit map (Figure 6a). 35 In its binding pose the thiol moiety of
31d intercalates directly between zinc atoms of the VIMꢀ2 active. We found that
ACS Paragon Plus Environment

Page 19 of 44
ACS Infectious Diseases
1
2
3
orientation of 31d in the active site is very similar to that one of the structurally related
4
5
6
28
inhibitor Lꢀcaptopril in the relevant structure of VIMꢀ2 (PDB ID:4C1E ). As shown in
7
8
9
Figure 6b, similarly to 31d, thiol moiety of Lꢀcaptopril is coordinated between the both
zinc cations inside the active center of VIMꢀ2 while replacing the polarized water
molecule required for catalysis. Furthermore, 31d and Lꢀcaptopril displays similarities of
coordination network with the protein backbone. Thus, the carbonyl oxygens of both
amide groups formed a hydrogen bonds with Asn210, the nonpolar parts of the
heterocycles fit well into the lipophilic subpocket of the active site, and the acid groups
directed towards the polar surface of the active site (Figure 6b). The additional phenyl
ring of 31b pointed in the direction of the solvent, apparently, due to the observed weak
electron density around this ring (Figure 6a), orients itself in a flexible manner, without
any steady interactions towards the enzyme.
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 6. A: Ligand interactions in the Xꢀray structure of VIMꢀ2 active site in complex with
31d. The positive 2mFoꢀDFc difference map (blue mesh contoured at 1σ) and mFoꢀDFc
polder omit map (green mesh) with solvent exclusion radius of 3Å and contoured at 3σ
around the 31d ligand were calculated from the final model (PDB: 5O7N). 31d molecule.
Ligand and neighboring protein environment are shown as stick models. The dashed lines
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 20 of 44
1
2
3
4
5
6
7
8
indicate the interactions of 31d with both Zinc ions (light blue spheres) and the key residues
of the active site. B: Superposition of the bound ligands 31d (cyan) and DꢀCaptopril (orange,
PDB: 4C1E) within VIMꢀ2 active site.
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
For 25a, 31d, and 37b the inhibitory potential on VIMꢀ2 was evaluated (table 6). Studies
concerning activity of Captopril on VIMꢀ2 were published elsewhere.28 In brief, activity
depended on the exact stereochemistry, whereby only one of the four possible
stereoisomers offered a nanomolar IC50. The other three showed activities in a low
micromolar range, comparable to VIMꢀ1. In the case of 25a, IC50 values were found only
as slightly lower than those on VIMꢀ1. For 31d and 37b the difference in the obtained
results was more obvious, as activity of 31d was increased by factor 15 and for 37b even
by factor 35. All tested compounds provided a significantly higher stabilization of VIMꢀ2
(Table 6). These variabilities between the single subtypes could be an additional difficulty
in the development of an “pan” MBL inhibitor.
Table 6. IC50 values and thermal shifts on VIMꢀ2 a
IC50 [µM]b
∆ °C b
9
0.07 / 4.4c
6.3 ± 0.2d
2.4 ± 0.1
6.3 ± 0.2
10.1 ± 0.2
25a 13.7 ± 8.30
31d
37b
0.3 ± 0.09
0.2 ± 0.12
aMelting point without inhibitor [in °C]: 61.9 ± 0.1. bResults are mean ± SD of three different
experiments. cD-/L-Captopril, values from the literature28; drac-Captopril
ACS Paragon Plus Environment

Page 21 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
Next, influence on cell viability of a human cell line was evaluated for 9, 25a, 31d, and
37b. In a LDH release assay, compounds 9, 25a and 31d showed only slight in vitro
toxicity to HeLa cells at a concentration of 100 µM (Figure 7a). Cytotoxicity of 37b was
found as marginally higher. In a WSTꢀ1 assay, compounds 9, 25a and 37b had only little
influence on cell proliferation at 100 µM. A more intensified antiꢀproliferative effect was
found for 31d (Figure 7b).
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 7. Cytotoxicity data of 9, 25a, 31d and 37b. A: LDH release assay; B: WSTꢀ1 assay.
The angiotensinꢀconverting enzyme (ACE) is an important player within the regulation of
blood pressure, and its inhibition leads to hypotension. In the case of an acute sepsis,
which goes along with an uncontrolled hypotension, a further blood pressure lowering
induced by application of an ACE inhibiting drug could have fatal consequences. Since all
synthesized compounds were derived from the molecular structure of Captopril, an
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 22 of 44
1
2
3
4
5
6
7
8
9
approved ACE inhibitor, 9, 25a, 31d and 37a were analyzed concerning their inhibitory
activity on ACE (Figure 8a). As expected, Captopril and 25a showed full inhibition, even
at 10 µM. For compound 31d, which differed in ring size of the heterocycle, a slightly
decreased inhibition was measured (10 µM: 85.8 ± 2.1 %, 30 µM: 95.0 ± 4.1). For 37a,
which contains the enlarged piperazine substructure, no inhibition could be detected at
concentrations up to 30 µM. Thus, the piperazine derivatives might serve best as a
scaffold for further investigations.
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
Orning et al. had demonstrated, that Captopril is also an inhibitor for LTA4H,36 which is
part of the biosynthesis of proꢀinflammatory mediator LTB4. In these studies a percentage
inhibition of 73.9 ± 4.5 % (10 µM) and 86.3 ± 3.1 % (30 µM) was found for 9 (Figure 8b).
Only minor inhibitory potentials were determined for the further evolved derivatives 25a
(10 µM: 12.4 ± 7.2 %, 30 µM: 17.6 ± 8.5), 31d (10 µM: 0.3 ± 2.8 %, 30 µM: 8.8 ± 8.0)
and 37a (10 µM: 12.1 ± 9.3 %, 30 µM: 17.0 ± 9.4).
ACS Paragon Plus Environment

Page 23 of 44
ACS Infectious Diseases
1
2
3
Figure 8. Offꢀtarget activity of 9, 25a, 31d and 37a. A: percentage inhibition of ACE; B: percentage
4
5
inhibition of LTA4H.
6
7
8
9
Compounds 25a and 31d were evaluated for their potential to inhibit bacterial MBLs
inside bacteria and their ability to restore the antibacterial activity of imipenem within an
antimicrobial susceptibility testing. Results are presented in table 7. In summary,
according to its IC50 values, 25a was less active in most E. coli strains, compared with 9,
which showed the best activity values. The increased activity of 31d in comparison to 25a
was reflected in the MIC experiments, nevertheless, without reaching the activity of 9. In
the case of clinical isolates, neither of the compounds, except 9 on an NDMꢀ1 carrying K.
pneumoniae T2301 strain, was able to restore activity of imipenem in a significant
manner.
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 7. MBL inhibitor (9, 25a, 31d) induced reduction of imipenem MIC in resistant
clinical isolates.
Imipenem MIC in mg/Lb
(fold change5 in MIC)
Isolatesa
+ 9,c,d
+ 25a c,d
+ 31d c,d
E. coli TOP10 T2359 (NDM-1)
E. coli TOP10 T2360 (IMP-7)
E. coli TOP10 T2361 (VIM-1)
E. coli pET24a T2377 (NDM-1)
E. coli pET24a T2378 (IMP-7)
E. coli pET24a T2379 (VIM-1)
K. pneumoniae T2301 (NDM-1)
P. aeruginosa T2226 (IMP-7)
128
2
4 (32)
0.5 (4)
0.5 (2)
1 (32)
0.5 (8)
2 (32)
4 (4)
32 (4)
1 (2)
16 (8)
1 (2)
1
0.5 (2)
2 (16)
1 (4)
1 (1)
32
4
2 (16)
0.5 (8)
2 (32)
8 (2)
64
16
64
32 (2)
16 (1)
32 (2)
32 (2)
32 (2)
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 24 of 44
1
2
3
K. pneumoniae T2216 (VIM-1)
8
4 (2)
8 (1)
4 (2)
4
5
6
7
a Transformants with respective MBL in TOPO-vector, pET24a-vector and clinical isolates; b 9, 25a and
31d in combination with imipenem was tested at a constant concentration of 128 mg/L; c 9, 25a and
8
9
d
31d did not exhibit any intrinsic antibacterial activity at the given concentration; Fold change was
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
calculated by MIC imipenem / MIC imipenem + inhibitor. Significant fold change (≥ 4) is indicated in
bold.
The results of MIC testing were surprising at the first glance but could be explained by the
capability of compounds to penetrate the outer membrane of Gramꢀnegative pathogens.
Brown et al. could demonstrate that antibacterial agents should have low logD7.4 values to
enable the permeation through the outer membrane of Gramꢀnegative pathogens. In the
case of E. coli, a logD7.4 of <0.5 should be sufficient, depending on the exact bacterial
strain. For the remaining Gramꢀnegative pathogens a significant lower logD7.4 value and
with that a higher polarity is needed.37 Polarity of 31d and other derivatives was measured
by a HPLCꢀbased test system with a logD7.4 of 0.49 for this structure (see Supporting
Information, Table S3). This was obviously higher than the logD7.4 of Captopril (ꢀ1.63).38
For efficient inhibition of Gramꢀnegatives it will be necessary to incorporate more polar
chemical groups, e.g. amines into the presented molecules. This could be realized by the
introduction of various residues at the piperazine ring, or by the substitution of the phenyl
ring.
CONCLUSIONS
We presented an overview about the challenges to develop thiolꢀbased MBL inhibitors.
Starting from the molecular structure of Captopril (9) we synthesized a small library of
differently substituted compounds and evaluated their activity on three MBL enzymes in a
ACS Paragon Plus Environment

Page 25 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
fluorescenceꢀbased assay system. We identified the piperidine substructure as a promising
motif, since the respective derivatives have shown good overall inhibition on all tested
MBLs. Furthermore, we found that the substitution of piperidine by piperazine provided
selectivity and a further increased activity on IMPꢀ7. DSF and ITC experiments
demonstrated that IC50 values do not always correlate with the enthalpically driven
binding behavior. Especially the increased activity of a promising piperidineꢀcontaining
derivative 31d on NDMꢀ1, within the activity assay, could not be confirmed by DSF and
ITC. In contrast, the less active piperazine (37b) provided an improved stabilization of the
enzyme, what was accounted for an increased enthalpic contribution. These results
underline the importance of biophysical methods for orthogonal compound
characterization in the field of thiolꢀbased MBL inhibitors. Furthermore, the piperazine
motive (37b) showed no or only minor activity on critical offꢀtargets like ACE or LTA4H,
whereas the piperidine derivative (31d) potently inhibited ACE.
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
To conclude, the piperazine scaffold is a promising approach for the further development
of thiolꢀbased MBLꢀinhibitors. We developed two novel synthesis routes, which allow to
extend the piperazine structure as well as to substitute the phenyl ring by various residues.
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 26 of 44
1
2
3
4
MATERIAL AND METHODS
5
6
Synthesis. For further synthetic details, see Supporting Information.
7
8
9
Protein Expression and Purification. The expression of IMPꢀ7, VIMꢀ1 and NDMꢀ1 was
performed according to the protocol, published by Klingler et al., with slight
modifications. In brief, 1L LB medium, supplemented with kanamycin was inoculated
with 10 mL overnight culture. The cultures were grown at 37 °C till an OD600 of 0.8 was
reached. Cultures were induced by adding IPTG to a final concentration of 400 µM,
before transferred to a 21 °C shaker for overnight incubation. The cells were harvested by
centrifugation (10866 x g, 20 min, 4 °C) the next day and cell pellets were stored at ꢀ20
°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
The frozen bacterial cell pellets were reꢀsuspended in buffer A (50 mM TRIS/HCl pH 8,
500 mM NaCl, 5 % Glycerol, 20 mM Imidazol/HCl pH 7, 100 µM ZnCl2) supplemented
with one tablet of Complete EDTA free protease inhibitor mix (Roche, Basel,
Switzerland) and a small amount of DNAse I (Applichem, Darmstadt, Germany). The cell
suspension was lysed and cell debris where separated by centrifugation (51.632 x g,
60 min, 4 °C). The supernatant was further purified by nickel affinity chromatograph,
using buffer A as running buffer and buffer B as elution buffer (50 mM TRIS/HCl pH 8,
500 mM NaCl, 5 % Glycerol, 400 mM Imidazol/HCl pH 7, 100 µM ZnCl2). The proteins
were purified using a step gradient protocol and eluted at 115 mM imidazole
concentration. The fractions containing the MBLs were pooled and concentrated. Then,
the concentrated fractions were loaded onto a Superdex 200 HiLoad 16/600 column (GE
Healthcare, Germany) equilibrated and run at 1 ml/min buffer C or D. For protein
samples used in the activity assay, buffer C (50 mM HEPES, pH 7.5) was used while
buffer D (47.6 mM TRIS, 476 mM NaCl, 4.7 % Glycerol, 0.95 µM ZnCl2 , pH 8) was
used for protein used in ITC experiments.
ACS Paragon Plus Environment

Page 27 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
FluorescenceꢀBased Activity Assay. Activity assays for MBLs were carried out, as
described by Klingler et al. Final protein concentrations were 0.5 nM (NDMꢀ1), 4.0 nM
(VIMꢀ1, ꢀ2) and 0.1 nM (IMPꢀ7) in a 50 mM HEPES buffer (pH 7.5, 0.01 % Triton Xꢀ
100). Substrate (Fluorocillin™ (Invitrogen Darmstadt, Germany) was dissolved in assay
buffer to a final concentration of 888 nM. Test compounds were dissolved and preꢀdiluted
in DMSO (final concentration: 1 %).
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
In a black polystyrol 96ꢀwell plate (Corning) an amount of 1 µL of the respective inhibitor
solution at different concentrations was incubated with 89 µL of MBL containing buffer
for 30 minutes at room temperature. 10 µL of substrate solution was added. The readout
of the emitted fluorescence was started immediately (45 s for 30 cycles) using a Tecan
Infinite F200Pro (Tecan Group Ltd; excitation at 495 nm and emission at 525 nm). Blank
controls were performed without enzyme. Positive controls were performed with enzyme
but without inhibitor. The inhibitory activity of each test compound was measured in three
independent experiments. For calculation of IC50 values, data obtained from
measurements with at least six different inhibitor concentrations were used. For the
evaluation of the sigmoidal dose response equation (variable slope with four parameters)
GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA) was used.
Differential Scanning Fluorimetry. Differential scanning fluorimetry (DSF) was
performed in transparent 96ꢀwell PCRꢀplates (MicroAmp, Applied Biosystems) with a
final assay volume of 40 µL. Test compounds were solved in DMSO an diluted in assay
buffer (47.6 mM TRIS, 476 mM NaCl, 4.7 % glycerol and 0.95 µM ZnCl2, pH 8) to a
final concentration of 100 µM. Final concentration of DMSO was 0.42 %. An amount of
32 µL enzyme containing assay buffer (final concentration of MBLs was 5 µM) was
mixed with 4 µL inhibitor solution and 4 µL SYPRO orange in assay buffer (2.5 x final
concentration). Melting points of untreated enzymes were measured without inhibitor.
ACS Paragon Plus Environment

ACS Infectious Diseases
Page 28 of 44
1
2
3
4
5
6
7
8
9
Background fluorescent was measured without enzyme. Temperatureꢀdependent increase
of fluorescent was measured from 25.0 °C to 79.8 °C, in steps of 0.2 °C every 24 seconds,
using an Icycler IQ single color real time PCR (BioRad). Excitation was set to 490 nm
and emission wavelength to 570 nm. All measurements were performed as triplicates in
three independent experiments. The first derivatives of melting curves were given by
MyIQ 1.0 software and the respective maxima, which were considered as melting point,
were determined in Microsoft Excel.
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
Isothermal Titration Calorimetry. The ITC experiments were performed on a Nano ITC
(TA Instruments, Waters GmbH), with the inhibitor solution in the 250 µL syringe and the
protein sample in the 940 µL sample cell. All measurements were performed at 25 °C
with a stirring speed of 350 rpm. The delay between the injections was set to 200 s. For all
measurements, except the blanks, frozen protein stocks have been used. The concentration
of the protein, the number of titration steps (16ꢀ25), and the volume (4ꢀ10 µL) were
slightly adjusted according to the specific measurement. The protein as well as inhibitor
samples were prepared in assay buffer D supplied with DMSO and diluted to a final
concentration of 2 %. All tested inhibitors have been prepared as 50 mM or 100 mM
DMSO stocks. The inhibitor samples were prepared with buffer D adjusting the final
DMSO concentration in the sample to 2 %, while the concentration of inhibitor was
chosen to optimize the shape of the curve. The data was analyzed using “NanoAnalyze
Data Analysis Version 3.5.0, TA Instruments”.
WSTꢀ1 Assay. Cell proliferation assay with WSTꢀ1 (Roche Diagnostics, Mannheim,
Germany) was performed to manufacturer´s protocol. HeLa cells were seeded in a 96ꢀwell
plate (5 x 103 cells/well) in DMEM supplemented with 1 mM PSP, 10 % FCS, 5 %
sodium pyruvate, 5 % penicillin and 5 % streptomycin. The medium was replaced by
DMEM, supplemented with 10 % FCS and test compounds (final concentrations 1 µM,
ACS Paragon Plus Environment

Page 29 of 44
ACS Infectious Diseases
1
2
3
4
5
6
7
8
9
10 µM and 100 µM), Revlotron (100 µM) as positive control and Zileuton (100 µM) and
DMEM / 1 % DMSO as negative control, after 24 h. After 48 h, WSTꢀ1 reagent (Roche)
was added to each well, according to manufacturer´s protocol and incubated for 45
minutes. Absorption was measured (450 nm, reference: 620 nm) using a Tecan Infinite
M200 (Tecan Group Ltd.). All experiments were carried out as triplicates in three
independent measurements.
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
LDH Release. Assay (Roche Diagnostics, Mannheim, Germany) was performed
according to manufacturer´s protocol. HeLa cells were seeded in a 96ꢀwell plate (1 x 104
cells/well) in DMEM supplemented with 1 mM PSP, 10 % FCS, 5 % sodium pyruvate,
5 % penicillin and 5 % streptomycin. After 24 h, the medium was replaced by DMEM,
supplemented with 10 % FCS and test compounds (final concentrations 1 µM, 10 µM and
100 µM), and incubated for 48 h. Triton Xꢀ100 (2 %) was added as positive control, 2 h
before measurement. After the incubation time, 100 µL of the supernatant was transferred
into a new 96ꢀwell plate and incubated with 100 µL of LDH substrate for 20 minutes.
Absorption was measured (490 nm, reference: 620 nm) using a Tecan Infinite M200
(Tecan Group Ltd.). All experiments were carried out as triplicates in three independent
measurements.
Crystallization of VIMꢀ2 from Pseudomonas aeruginosa in complex with 31d. VIMꢀ2
protein production and purification was performed accordingly to methods as published
39
by Christopeit et al. and as described in Protein Expression and Purification section
above and with some following modifications. TEV protease was added in a 1:10 molar
ratio to the pooled fractions of the protein eluted form the Ni2+ꢀNTA affinity column. The
resulting mixture was dialyzed for 18ꢀ24 hours at 4°C against 100fold excess of buffer E
(50 mM TRIS/HCl pH 8, 500 mM NaCl, 5 % Glycerol, 10 mM BME, 100 µM ZnCl2)
using a 3500 kDa membrane. The dialysis buffer was exchanged twice during the dialysis
ACS Paragon Plus Environment