((S)-3-Mercapto-2-methylpropanamido)acetic acid derivatives as metallo-β-lactamase inhibitors: Synthesis, kinetic and crystallographic studies

Article abstract of DOI:10.1016/j.ejmech.2018.01.032

The emergence and global spread of metallo-β-lactamase (MBL) mediated resistance to almost all β-lactam antibacterials poses a serious threat to public health. Since no clinically useful MBL inhibitors have been reported, there is an urgent need to develop new potent broad-spectrum MBL inhibitors effective against antibacterial resistance. Herein, we synthesized a set of 2-substituted ((S)-3-mercapto-2-methylpropanamido) acetic acid derivatives, some of which displayed potent inhibition with high ligand efficiency to the clinically relevant MBL subtypes, Verona Integron-encoded MBL (VIM)-2 and New Delhi MBL (NDM)-1. Kinetic studies revealed that the inhibitors are not strong zinc chelators in solution, and they bind reversibly to VIM-2 but dissociate very slowly. Crystallographic analyses revealed that they inhibit VIM-2 via chelating the active site zinc ions and interacting with catalytically important residues. Further cell- and zebrafish-based assays revealed that the inhibitors slightly increase susceptibility of E. coli cells expressing VIM-2 to meropenem, and they have no apparent toxicity to the viability of HEK293T cells and the zebrafish embryogenesis.

Full text of DOI:10.1016/j.ejmech.2018.01.032

Accepted Manuscript
((S)-3-Mercapto-2-Methylpropanamido)acetic acid derivatives as metallo-β-lactamase
inhibitors: Synthesis, kinetic and crystallographic studies
Sha Liu, Li Jing, Zhu-Jun Yu, Chengyong Wu, Yongxiang Zheng, En Zhang, Qiang
Chen, Yamei Yu, Li Guo, Yong Wu, Guo-Bo Li
PII:
S0223-5234(18)30045-X
10.1016/j.ejmech.2018.01.032
EJMECH 10109
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 18 October 2017
Revised Date: 9 January 2018
Accepted Date: 10 January 2018
Please cite this article as: S. Liu, L. Jing, Z.-J. Yu, C. Wu, Y. Zheng, E. Zhang, Q. Chen, Y. Yu, L. Guo,
Y. Wu, G.-B. Li, ((S)-3-Mercapto-2-Methylpropanamido)acetic acid derivatives as metallo-β-lactamase
inhibitors: Synthesis, kinetic and crystallographic studies, European Journal of Medicinal Chemistry
(2018), doi: 10.1016/j.ejmech.2018.01.032.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract

ACCEPTED MANUSCRIPT
((
S)-3-Mercapto-2-Methylpropanamido)Acetic Acid Derivatives as
Metallo-β-Lactamase Inhibitors: Synthesis, Kinetic and Crystallographic
Studies
Sha Liu,a Li Jing,a Zhu-Jun Yu,a Chengyong Wu,b Yongxiang Zheng,a En
Zhang,c Qiang Chen,b Yamei Yu,b Li Guo,a Yong Wu,a* and Guo-Bo Lia*
a Key Laboratory of Drug Targeting and Drug Delivery System of Ministry of Education, West
China School of Pharmacy, Sichuan University, Sichuan 610041, China.
b Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical
School, Sichuan University, Sichuan 610041, China.
c School of Pharmaceutical Sciences, Zhengzhou University, Henan 450001, China
* Correspondence: liguobo@scu.edu.cn (G.-B. Li); wyong@scu.edu.cn (Y. Wu)
1

ACCEPTED MANUSCRIPT
Abstract
The emergence and global spread of metallo-β-lactamase (MBL) mediated resistance
to almost all β-lactam antibacterials poses a serious threat to public health. Since no
clinically useful MBL inhibitors have been reported, there is an urgent need to
develop new potent broad-spectrum MBL inhibitors effective against antibacterial
resistance.
Herein,
we
synthesized
a
set
of
2-substituted
((S)-3-mercapto-2-methylpropanamido) acetic acid derivatives, some of which
displayed potent inhibition with high ligand efficiency to the clinically relevant MBL
subtypes, Verona Integron-encoded MBL (VIM)-2 and New Delhi MBL (NDM)-1.
Kinetic studies revealed that the inhibitors are not strong zinc chelators in solution,
and they bind reversibly to VIM-2 but dissociate very slowly. Crystallographic
analyses revealed that they inhibit VIM-2 via chelating the active site zinc ions and
interacting with catalytically important residues. Further cell- and zebrafish-based
assays revealed that the inhibitors slightly increase susceptibility of E. coli cells
expressing VIM-2 to meropenem, and they have no apparent toxicity to the viability
of HEK293T cells and the zebrafish embryogenesis.
Keywords: Antibacterial resistance; Metallo-β-lactamase (MBL); VIM-2;
Crystallography; Inhibitor design
2

ACCEPTED MANUSCRIPT
Introduction
The selection pressure exerted by global overuse of β-lactam antibiotics,
including penicillins, cephalosporins, carbapenems, and monobactams, has led
to the rapid emergence and spread of antibacterial resistance in a growing
number of pathogens [1,2]. These antibiotic-resistant pathogens are increasing
threats to public health [3]. One of the most important mechanisms of
resistance to β-lactam antibiotics involves their hydrolysis as catalysed by
β-lactamases [4,5]. According to catalytic mechanism, β-lactamases are
grouped into two categories: serine β-lactamases (SBLs; Amber class A, C, and
D) which use a nucleophilic serine in catalysis, and metallo-β-lactamases
(MBLs; Amber class B) which use one or two zinc ions in catalysis (Figure 1a)
[4,6]. With the recent approval of avibactam, four SBL inhibitors (including
clavulanic acid, sulbactam, and tazobactam) are clinically available in
combination of appropriate β-lactam antibiotics to overcome bacterial
resistance [7-9]. In contrast, there are no reports of clinical useful MBL
inhibitors.
As the MBLs, which are structurally and mechanistically distinct from the
SBLs, restricted distribution in less clinically relevant pathogens, they were
long not considered as a significant threat to the clinical effectiveness of
β-lactam antibiotics [9]. Alarmingly, this situation has been changed. The
emergence of horizontally acquired MBLs has resulted in their extensive
dissemination across geographical and species boundaries [9]. This has become
3

ACCEPTED MANUSCRIPT
a major public health concern, because i) MBLs can hydrolytically degrade
almost all β-lactams (with one current exception, i.e. the monobactam
aztreonam) including carbapenems, which are considered as the ‘last resort’
antibiotics [10,11]; ii) MBLs are produced by many clinically relevant
Gram-negative pathogens such as Klebsiella pneumoniae, Pseudomonas
aeruginosa Aeromonas Acinetobacter Enterobacteriaceae, and Serratia
,
,
,
marcescens [12-14]; iii) many MBL genes are present in environmental
species, constituting frightening reservoirs of β-lactam resistance genes [15-17].
Therefore, there is an urgent need to develop new MBL inhibitors effective
against resistance to β-lactam antibacterials.
The MBLs are subdivided into three subclasses (i.e., B1, B2 and B3) based
on their sequence homology, metal ion feature, and substrate selectivity;
notably, recently, an additional MBL subclass (B4) has been described, of
which the MBLs (such as SPR-1 and CSA-1) have unique active site and may
employ a catalytic mechanism distinct from other MBLs [18-21]. The B1
subclass MBLs are most clinically relevant, such as Verona integron-encoded
MBL (VIM), the New Delhi MBL (NDM), and Imipenemase (IMP), which are
di-Zn(II) utilizing enzymes, with both Zn1 and Zn2 ions having crucial roles
for β-lactam hydrolysis (Figure 1a). To date, a number of classes of MBL
inhibitors have been reported, such as thiols [22-25], carboxylic acids [26,27],
rhodanines [11], triazolylthioacetamide [28], cyclic boronates [29,30], and
others [31-36]; most of them act via active site metal chelation.
4

ACCEPTED MANUSCRIPT
We recently identified the 3-oxoisoindoline-4-carboxylate derivatives as a
new type of MBL inhibitors using combined virtual and biophysical screening,
which do not chelate the active site zinc ions [37]. However, these derivatives
have narrow selectivity for VIM-2 MBL, which may limit their use for
potentiating β-lactam antibiotics, since ideal MBL inhibitors should achieve
broad-spectrum inhibition against multiple MBL types, at least the B1 subclass
[31,37]. L-captopril, which is an old drug used in clinic for several decades to
control high blood pressure, has been reported as a broad-spectrum MBL
inhibitor by chelating the active site zinc ions via its thiol group [23]. The
captopril stereoisomers (Figure 1b) and analogues (e.g., 1 and 2, Figure 1c) also
show broad-spectrum, potent inhibition to multiple B1 MBLs (e.g., NDM-1,
VIM-2, VIM-1, VIM-5, SPM-1, BcII, and IMP-1) [23,38,39], and have the
ability to potentiate meropenem against MBL-producing bacteria [23,38].
Consequently, we think that ((S)-3-mercapto-2-methylpropanamido)acetic acid,
the core scaffold of the captopril stereoisomers and analogues (color in Red,
Figure 1b-c), might be a privileged scaffold for developing potential clinically
useful MBL inhibitors. Herein, we report a set of 2-substituted
((S)-3-mercapto-2-methylpropanamido)acetic acid derivatives, which show
potent inhibition to the clinically relevant MBL subtypes, VIM-2 and NDM-1.
Some of these derivatives inhibit VIM-2 with high ligand efficiency (LE) (LE >
0.6) values; LE is a measurement of the binding energy per atom of a ligand to
its target, which is particularly useful during hit to lead optimization [40].
5

ACCEPTED MANUSCRIPT
Further kinetic, crystallographic, cell-based, and zebrafish-based assays were
carried out to investigate their inhibition mechanism, cell susceptibility, and
potential toxicity.
Figure 1. (a) Outline of the mechanism for MBL catalysed β-lactam hydrolysis; note,
for enzyme-substrate combinations, variations of this mechanism are possible. (b) The
inhibition modes of L-captopril (PDB code: 4C1D) [23] and D-captopril (PDB code:
4C1E) [23] with VIM-2 MBL; (c) The inhibition modes of captopril analogues 1
(PDB code: 5N4T) [38] and 2 (PDB code: 5N4S) [38] with VIM-2 MBL.
2. Results and Discussion
2.1 Chemistry
The 2-substituted ((
S)-3-mercapto-2-methylpropanamido)acetic acid
derivatives ( 12) were synthesized via the synthetic routes shown in Scheme 1.
3
-
The commercially available D-/L-alpha-phenylglycines or D-/L-phenylalanines
6

ACCEPTED MANUSCRIPT
were first reacted with thionyl chloride (SOCl2) and methanol (MeOH) at 75
to afford the corresponding esters (II, Scheme 1), which were taken up for the
next step without any purification. Then, intermediates II and
℃
(
S
)-3-(acetylthio)-2-methylpropanoic acid (III, Scheme 1) were reacted in the
presence of 2.0 equiv of hydroxybenzotriazole (HOBT), 2.0 equiv of
carbodiimide hydrochloride (EDCI) and 3.0 equiv of
-diisopropylethylamine (DIPEA) to yield the corresponding amides (IV),
respectively. Finally, treatment of the compounds (IV, Scheme 1) with 1M
aqueous NaOH in methanol (1:1) gave the desirable target compounds 12 at
N,N
3
-
room temperature (r.t.); the yields over the two steps (b-c, Scheme 1) are
66%-72%.
Scheme 1. Reagents and conditions: (a) SOCl2, MeOH, 0-75℃, 2h, 96-100% yields;
(b) HOBT, EDCI, DIPEA, DMF, r.t. 12h, 81-88% yields; (c) NaOH, MeOH, H2O, r.t.
2 h, 78-85% yields.
7

ACCEPTED MANUSCRIPT
2.2 Structure-Activity Relationship and Kinetic Studies with VIM-2 and
NDM-1.
All of the synthesized ((
derivatives ( 12) were tested against VIM-2 and NDM-1, which are clinically
important targets for combating antibacterial resistance, using
S)-3-mercapto-2-methylpropanamido)acetic acid
3
-
a
fluorescence-based assay as previously described [9,37,38]. The inhibitory
activity (IC50/pIC50), LE, clogP and clogS values are given in Table 1;
L-/D-captopril and their analogues
in Table 1, compounds 12 have similar liposolubility and water-solubility
properties (clogP and clogS), as predicted by the ALOGPS 2.1 program,[41] to
L-/D-captopril and ; all of these derivatives have clogP in the range of 1~3
1 and 2 are used for comparison. As shown
3
-
1/2
and clogD ranging from -2 to -3 (Table 1), similar with many marked
antibacterials [42,43]. In general, the D-amino acid-containing compounds
(
8-12, Table 1) displayed more potent inhibition to VIM-2 and NDM-1 than the
corresponding L-amino acid-containing compounds (3-7, Table 1). We
observed that the slopes of the dose-inhibition curves of compounds 12 with
3
-
VIM-2 and NDM-1 (Figure S1-2 in Supporting Information) are close to 1,
which may reflect that the inhibition mode of these compounds does not
involve multisite binding or colloidal aggregation [44].
Table 1. The inhibition potency (IC50/pIC50), LE, clogP, and clogS of the
((S)-3-mercapto-2-methylpropanamido)acetic acid derivatives.
a
ID
Chemical structure
IC50 (µM) / pIC50 / s.e. pIC50
LEb
clogPc
clogSc
8

ACCEPTED MANUSCRIPT
VIM-2
NDM-1
VIM-2
NDM-1
0.37
2.47 / 5.61 /
0.038
196.8 / 3.71 /
0.173
L-capd
0.56
1.02
1.02
2.58
2.58
2.07
1.85
1.91
2.03
1.86
2.07
1.85
1.91
-1.68
-1.68
-4.35
-4.35
-2.84
-3.1
4.55 / 5.34 /
0.045
62.15 / 4.21 /
0.107
D-capd
0.53
0.40
0.48
0.43
0.46
0.61
0.54
0.55
0.54
0.48
0.60
0.42
0.31
0.28
0.28
0.26
0.42
0.38
0.36
0.40
0.31
0.39
1.06 / 5.97 /
0.020
24.81 / 4.61 /
0.097
1d
2d
3
0.05 / 7.26 /
0.053
64.77 / 4.19 /
0.333
2.85 / 5.55 /
0.084
233.30 / 3.63 /
0.356
0.49 / 6.31 /
0.043
326.00 / 3.49 /
1.008
4
0.039 / 7.41 /
0.068
8.77 / 5.06 /
0.047
5
-2.66
-3.27
-2.68
-2.84
-3.1
0.121 / 6.92 /
0.034
11.66 / 4.93 /
0.170
6
O
OH
O
0.077 / 7.11 /
0.058
24.02 / 4.62 /
0.053
(
S
)
7
(
)
S
HS
N
H
OH
0.121 / 6.92 /
0.055
7.00 / 5.16 /
0.111
8
0.276 / 6.56 /
0.052
73.12 / 4.14 /
0.160
9
0.057 / 7.25 /
0.058
20.46 / 4.69 /
0.1275
10
-2.66
9

ACCEPTED MANUSCRIPT
0.055 / 7.26 /
0.034
5.59 / 5.25 /
0.119
11
12
0.56
0.61
0.41
0.42
2.03
1.86
-3.27
-2.68
0.0137 / 7.86 /
0.038
3.57 / 5.45 /
0.037
a The method for measuring IC50/pIC50 (n=3) values is described in Experimental Section.
b
The ligand efficiency (LE) is calculated using the formula: LE = 1.4*pIC50/N, where N is
the number of non-hydrogen atoms.[40]
c
The clogP and clogS values were calculated using the ALOGPS 2.1 program
(http://www.vcclab.org/lab/alogps/; available August 5, 2017).[41]
d The IC50/pIC50 values from literature.[38]
For VIM-2,
moieties, respectively, showed slightly better or comparable inhibitory activity
to L-captopril and the L-tryptophan-containing compound . By contrast,
and which bear L-phenylglycine, L-4-fluorophenylglycine, and
L-4-hydroxyphenylglycine moieties, respectively, manifested IC50 of 39~121
nM (Table 1), more potent than L-captopril (IC50 = 2.47 µM) and (IC50 = 1.06
µM); the ligand efficiency (LE) values of compounds (LE = 0.54 ~ 0.61)
are close to that of L-captopril (LE = 0.56), and higher than that of (LE = 0.40)
(Table 1). In comparison, D-amino acid containing compounds ( 12
displayed substantially improved inhibitory potency to VIM-2 MBL comparing
with D-/L-captopril, , and (Table 1). Particularly, 10 11, and 12, which are
3 and 4, which contain L-phenylalanine and L-tyrosine
1
5, 6,
7
,
1
5-7
1
8
-
)
1
2
,
derived from phenylglycines, manifested IC50 values of 56.6 nM, 54.8 nM, and
10

ACCEPTED MANUSCRIPT
13.7 nM, respectively, and their LE values are 0.60, 0.56, and 0.61,
respectively.
For NDM-1, we observed that the synthesized derivatives (
3
-
12) only
, and
showed slightly improved potency compared with L-/D-captopril,
1
2
(Table 1); the most potent compounds 11 and 12 manifested inhibitory activity
(IC50) of 5.59 µM and 3.57 µM, respectively, and had LE values of 0.41 and
0.42, respectively (Table 1). No obvious inhibitory activity of 3-12 for the
serine-β-lactamase TEM-1 were observed at 100 µM.
Since these compounds (3-12) possess potential metal-chelating motifs (e.g.,
thiol and carboxylate), further VIM-2 inhibition assays were carried out at three
different concentrations of Zn(II) (0 µM, 1 µM, and 100 µM) for the
representative stereoisomers
metal chelation in solution. We observed that there is no obvious difference
between the inhibitory activities of and 12 to VIM-2 with or without excess
Zn(II) (Figure 2a), reflecting that similar to captopril and other thiol-based
MBL inhibitors [24,38], and 12 are not strong Zn(II) chelators in solution and
7 and 12, with aim to investigate the potential for
7
7
tolerated a high zinc concentration without loosing activity.
11

ACCEPTED MANUSCRIPT
Figure 2. (a) The dose-inhibition curves of compounds 7 and 12 with VIM-2 obtained
at three different zinc ion concentrations, revealing that they are not strong metal
chelators in solution. (b) The reversibility of the VIM-2 inhibition by L-captopril, 7,
and 12 examined by the jump-dilution method, revealing that comparing with
L-captopril, 7 and 12 dissociate more slowly.
We subsequently examined the reversibility of
7 and 12 inhibiting VIM-2 in
solution by using the rapid dilution approach as described previously [38,45].
The VIM-2 enzymes (the concentration equal to 100-fold used in the activity
assays, Figure 2b) were incubated with L-captopril, 7, or 12 (the concentration
of the inhibitor equal to 10-fold the IC50 value, Figure 2b) for 30 minutes at
room temperature. The samples were then rapidly diluted 100-fold and the
12

ACCEPTED MANUSCRIPT
enzyme activity was measured. Comparing with L-captopril, which has fully
reversible inhibition to VIM-2 (Figure 2b),
7 and particularly, 12 manifested in
slow dissociation kinetics (Figure 2b), reflecting that they form stable
complexes and have strong binding affinity with VIM-2.
2.3 Crystallographic Analyses.
In
order
to
investigate
how
the
most
potent
((
S)-3-mercapto-2-methylpropanamido)acetic acid derivatives bind to VIM-2,
we attempted to obtain X-ray crystal structures of complexes of VIM-2 with
them through co-crystallization experiments (Table S1). The VIM-2:11 (PDB
ID 5Y6D) and VIM-2:12 (PDB ID 5Y6E) complex structures were solved to
1.76 Å and 1.57 Å resolution (Table 2), respectively. The VIM-2:11 complex
crystallized in a
(ASU) and was refined to Rwork
from VIM-2:11, the VIM-2:12 complex structure contained two molecules of
VIM-2 in ASU, and was refined to Rwork Rfree value of 13.69/17.28% (Table 2).
In both structures, there was well-defined electron density (Fo Fc) in the VIM-2
C121 space groups with four molecules in the asymmetric unit
/
Rfree value of 16.32/21.88% (Table 2). Different
/
-
active site, into which 11 and 12 could be confidently modelled (Figure 3a-b).
Table 2. Data collection and refinement statistics.
Structure (PDB ID)
VIM-2:11 (5Y6D)
VIM-2:12 (5Y6E)
Data Collection
Space Group
C 1 2 1
C 1 2 1
Cell Dimensions: a, b, c (Å)
Cell Dimensions: α, β, γ (˚)
*Mol/ASU
189.03, 39.83, 133.84
90.00, 124.56, 90.00
4
101.36, 74.45, 64.46
90.00, 129.46, 90.00
2
Resolution Range (outer shell) (Å)
Number of Unique Reflections
46.60-1.76 (1.79-1.76)
90510
49.76-1.57 (1.60-1.57)
64049
13

ACCEPTED MANUSCRIPT
Completeness (%)
I/σ(I) (outer shell)
Wilson B Factor (Å2)
Refinement
92.2
98.4
15.4 (2.3)
18.86
16.9 (6.3)
13.25
Overall B Factor (Å2)
Protein B Factor (Å2)
Ligand B Factor (Å2) (occupancy)
Water B Factor (Å2)
‡RMSD from Ideal Bond Length (Å) 0.01
23.26
23.12
14.49
13.22
12.43 (1.0)
23.90
0.01
25.90 (1.0)
25.29
‡RMSD from Ideal Angles (˚)
work (%)
Rfree (%)
1.07
1.05
R
16.32
21.88
13.69
17.28
*Mol/ASU = molecules per asymmetric unit; ‡RMSD = root mean square deviation.
Figure 3. Crystallographic analyses reveal the binding modes of 11/12 to VIM-2.
Observed mFo-DFc electron density maps (OMIT maps) around (a) 11 and (b) 12
contoured at 3σ (blue mesh). View from crystal structures of VIM-2 in complex with
(c) 11 (PDB ID 5Y6D) and (d) 12 (PDB ID 5Y6E) reveal that the inhibitors binds to
chelate with active site zinc ions and form hydrophobic and electrostatic interactions
with residues on the L3 and L10 loops (e.g., cation–π interactions with Arg228, and
14

ACCEPTED MANUSCRIPT
hydrogen-bonding interactions with Asn233; see Figure S3-4 for further details of
binding interactions).
The crystal structures revealed that 11 and 12 have very similar binding
modes. 11 and 12 are positioned to chelate to and bridge the active site zinc
ions via their thiol group (2.2~2.3 Å, Figure 3c-d and Figure S3-4); their
carboxylate moiety makes hydrogen-bonding interactions with Arg228 (3.0 Å)
and Asn233 (2.9 Å) on L10 loop; their phenyl moiety makes cation-π
interactions with Arg228 (3.4 Å, i.e. from the cation carbon atom of Arg228 to
the center of the phenyl moiety of 11 and 12); their phenyl and methyl moieties
form hydrophobic interactions with Phe61 and Tyr67 on L3 loop (Figure 3c-d
and Figure S3-4). Notably, the carboxylate group of 11 and 12 is observed to
have hydrogen-bonding interactions with the structural water molecule W2 (2.7
Å) (Figure 3c-d and Figure S3-4), which is positioned to form hydrogen bonds
with His196 (2.9 Å), Tyr224 (2.8 Å), and Asn233 (2.7 Å). Comparing with 11
,
12 (the phenolic hydroxyl group) has additional hydrogen-bonding interactions
with the water molecule W3 (2.9 Å), bridging to interact with Glu225 (2.9 Å)
(Figure 3d); this may, in part, explain why 12 is slightly more potent than 11. In
addition, the VIM-2:12 crystal structure clearly reveal the key interaction
features of 12 as a high ligand-efficiency inhibitor (LE = 0.61, Table 1), e.g. the
thiol chelating with the active site zinc ions, the amide forming hydrogen bonds
with Asn233, the carboxylate interacting with Asn233 and Arg228, and the
phenolic hydroxyl interacting with Arg228, Glu255, and Tyr67. The molecular
15

ACCEPTED MANUSCRIPT
docking analyses of 12 binding with NDM-1 revealed that 12 appears
positioned to chelate with active site zinc ions and interact with the catalytically
important residues, e.g. via forming hydrogen bonds with Lys224 and Asn233
(Figure S5a), similar with that observed in the NDM-1:Hydrolyzed
Benzylpenicillin (PDB ID 4EYF) [46] complex structure (Figure S5b).
Comparison of the complex crystal structures of VIM-2:11 (PDB ID 5Y6D),
VIM-2:12 (PDB ID 5Y6E), VIM-2:1 (PDB ID 5N4T) [38], VIM-2:2 (PDB ID
5N4S) [38], VIM-2:L-captopril (PDB ID 4C1D) [23], and VIM-2:D-captopril
(PDB ID 4C1E) [23], revealed the common pharmacophore features of the
((S)-3-mercapto-2-methylpropanamido)acetic acid derivatives, i.e., by chelating
with active site zinc ions and forming electrostatic and hydrogen-bonding
interactions with the catalytically important residues Arg228 and Asn233
(Figure S6). 11 and 12 have a very similar binding mode with 2; in addition to
the common pharmacophore features, their aromatic rings are positioned to
form hydrophobic/aromatic interactions with Phe61 and Tyr67, and cation-π
interactions with Arg228 (Figure 4a). Interestingly, 12 was observed to have
water-bridging interactions with Glu225 (Figure 4a), while no such interactions
were observed for
2 (Figure 4a), which may in part explain why 12 has more
potent VIM-2 inhibition than
2
.
16

ACCEPTED MANUSCRIPT
Figure 4. (a) Comparison of the complex crystal structures of VIM-2:12 (PDB ID
5Y6E) and VIM-2:2 (PDB ID 5N4S) [38], reveals that 12 and 2 have a very similar
binding mode to VIM-2. By superimposing structures of VIM-2:12 (PDB code: 5Y6E)
with (b) NDM-1:hydrolyzed benzylpenicillin (PDB code: 4EYF) [46] and (c)
17

ACCEPTED MANUSCRIPT
NDM-1:hydrolyzed cephalosporin (PDB code: 4RL0) [47], reveals that 12 likely
mimics the binding mode of ring-opened penicillin and cephalosporin intermediates to
NDM-1.
By superimposing structures of VIM-2:12 (PDB code: 5Y6E) with
NDM-1:hydrolyzed benzylpenicillin (PDB code: 4EYF) [46] and
NDM-1:hydrolyzed cephalosporin (PDB code: 4RL0) [47], we observed that 12
binds via a similar binding mode to that observed in complexes of NDM-1 with
ring-opened penicillin and cephalosporin intermediates (Figure 4b-c); the
carboxylate group of 12 forms electrostatic interactions with Arg228 in a
similar manner to that observed for the binding of the cephalosporin C-4
carboxylate to Lys224 (a structurally equivalent residue to Arg228) with
NDM-1 (Figure 4c). These results suggest that the inhibitor design via
mimicking the ring-open substrate intermediates would be an effective strategy
to obtain highly potent MBL inhibitors.
2.4 Cell Susceptibility and Preliminary Toxicity Assessment.
To assess the effects of the VIM-2 inhibitors on a cellular level, we
established the E. coli DH5α strain containing the VIM-2-encoding plasmid (E.
coli-VIM-2) as described in Experimental Section. Meropenem showed a
minimum inhibitory concentrations (MIC) value of 16 µg/mL against the E.
coli-VIM-2 strain, larger than that against the control E. coli strain (MIC < 0.25
µg/mL, Table 3), indicating that the E. coli-VIM-2 strain is resistant to
meropenem, and it can be used as a model to assess the VIM-2 inhibitors. The
18

ACCEPTED MANUSCRIPT
MIC data of meropenem in the absence or presence of
7 and 12 revealed that
both compounds resulted in 2-4 fold MIC reduction of meropenem to the E.
coli-VIM-2 cells (Table 3), indicating that they increase susceptibility of E.
coli-VIM-2 cells to meropenem. Although these inhibitors showed moderate
synergistic activity with meropenem to the established E. coli-VIM-2 cells, they
may be tested against clinically isolated VIM-2 producing strains like other
thiol-based inhibitors [23,38].
Table 3. MICs of meropenem (MEM) the absence or presence of 7 and 12 against
VIM-2-producing E. coli strain.
MIC (µg/mL)
E. coli-VIM-2
Assay
No.
1
Inhibitor
Concentration (µM)
7 (100 µM)
7 (200 µM)
7 (500 µM)
12 (100 µM)
12 (200 µM)
12 (500 µM)
7 (500 µM)
12 (500 µM)
-
Meropenem
E. coli
<0.25
<0.25
<0.25
<0.25
<0.25
<0.25
>128
+
+
+
+
+
+
-
16
16
8
8
8
2
3
4
5
6
7
8
9
4
>128
>128
16
-
+
>128
<0.25
We next evaluated the preliminary toxicity of
and zebrafish embryo development. The cultured HEK293T cells were treated
with 50 µM or 12 for 72 hours; for , no visible toxicity was observed, but for
7 and 12 on HEK293T cells
7
7
12 slight inhibition activity (~ 8%) was observed (Figure 5a). The in vivo live
fluorescent transgenic zebrafish (FLK-1:EGFP) assays were then used to
examine the effects of
7 and 12 at 20 µM on embryo development (see
Experimental Section). For all the tests, no deaths (i.e., no heartbeats) and no
19

ACCEPTED MANUSCRIPT
limited mobility were observed. In addition,
7
and 12 have no visible effects on
the development of zebrafish embryos, e.g. morphology and angiogenesis
(Figure 5b). These results reflected that
7 and 12 are not cytotoxic/embryotoxic
(zebrafish) compounds under the assay conditions.
Figure 5. Testing the toxic effects of 7 and 12 on HEK293T cells and zebrafish
embryos. (a) No obvious toxicity was observed for 7, but for 12 slight inhibition
activity to HEK293T cells was observed; (b) 7 and 12 have no visible effects on
livability, morphology and angiogenesis of zebrafish embryos.
3. Conclusion
In summary, structure-activity relationship studies with the synthesized
((S)-3-mercapto-2-methylpropanamido)acetic acid derivatives revealed that some of
them have very potent inhibition to VIM-2 and NDM-1. Kinetic studies revealed that
20

ACCEPTED MANUSCRIPT
the potent inhibitors (at least 7 and 12) are not strong zinc chelators in solution, and
they manifest in slow dissociation kinetics. Crystallographic analyses revealed that
the inhibitors bind to chelate the active site zinc ions and interact with catalytically
important residues, similar to the binding modes of hydrolyzed substrates (e.g.,
penicillin and cephalosporin), suggesting that the inhibitor design via mimicking the
ring-open substrate intermediates would be an effective strategy to obtain highly
potent MBL inhibitors. Moreover, the tested inhibitors have no visible toxicity to the
viability of HEK293T cells and zebrafish embryo development. The overall results
revealed the potential of ((S)-3-mercapto-2-methylpropanamido)acetic acid
derivatives as lead compounds to develop clinically useful MBL inhibitors, and the
complex crystal structures will provide clues and insights into inhibitor optimization.
4. Experimental Section
4.1 Chemistry
Unless otherwise indicated, all starting materials, reagents, and solvents
were purchased from market and were used as supplied without further
purification. Reactions were monitored by thin layer chromatography (TLC) on
Qingdao Haiyang silica gel F-254 thin layer plates. Column chromatography,
performed on Qingdao Haiyang silicagel 60 (300–400 mesh), was used to
1
purify the intermediates and target compounds. H NMR spectra were recorded
13
at 400 or 600 MHz, and C NMR spectra were recorded at 100 or 150 MHz.
Chemical shifts were expressed in parts per million (ppm, δ) relative to
tetramethylsilane (TMS) as an internal standard. Proton coupling patterns are
21

ACCEPTED MANUSCRIPT
described as singlet (s), doublet (d), triplet (t) and multiplet (m). Low-resolution
and high-resolution mass spectral (MS) data were determined on an Agilent
1100 series LC−MS with UV detection at 254 nm in low-resonance
electrospray ionization (ESI) mode. HPLC analysis was carried out a Waters
2695 HPLC system equipped with a Kromasil C18 column (4.6 mm × 250 mm,
5 µm).
4.1.1 General Procedures for the Target Compounds (Compound 3 as An
Example)
The target compounds were synthesized through three steps.
First step: The methyl-esterification reaction of L-phenylalanine. To a suspension
of L-phenylalanine (1 g, 6.05 mmol, 1.0 equiv) in methanol (25 mL) cooled to 0 °C
was added thionyl chloride (475 µL, 6.66 mmol, 1.5 equiv) slowly. After 10 min, the
reaction mixture was heated to 75 oC to reflux for 2 h. After completion (monitored by
TLC), the solvent was removed in vacuo, and at this time, a white solid was formed
methyl L-phenylalaninate hydrochloride (1.23g, quantitative yield).
Second
step:
The
condensation
acid. The
reaction
product
with
(S)-3-(acetylthio)-2-methylpropanoic
crude
methyl
L-phenylalaninate hydrochloride (200 mg, 0.99 mmol, 1.0 equiv) was dissolved in
DMF (4 mL), then 1-hydroxybenzotriazole (HOBT, 273 mg, 1.98 mmol, 2.0 equiv),
1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride (EDCI, 378 mg,
1.98 mmol, 2.0 equiv), N,N-diisopropylethylamine (DIEA, 516 µL, 2.97 mmol, 3.0
equiv, ) and (S)-3-(acetylthio)-2-methylpropanoic acid (161 mg, 0.99 mmol, 1.0 equiv)
22

ACCEPTED MANUSCRIPT
were added. The reaction was stirred at ambient temperature for 12 h and the
conversion of the reaction was monitored by TLC analysis. The reaction mixture was
extracted with EtOAc (3 × 15 mL) after adding H2O (15 mL), then the combined
organic layers were dried over Na2SO4, filtered, and concentrated in vacuo. The crude
product was purified by column chromatography (PE/EtOAc, 15:1) to yield methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-L-phenylalaninate as a white solid (225 mg,
yield 70.1%).
Third
step:
Hydrolyse-Reaction.
To
a
solvation
of
methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-L-phenylalaninate (40 mg, 1.0 equiv) in
degassed methanol (3 mL), 1 M NaOH (3 mL) was added and stirred at ambient
temperature for 2 h. Upon completion of the reaction, the methanol was evaporated in
vacuo, and the residue was acidified with 1M HCl solution (pH 1-2) and extracted
several times with ethyl acetate, dried by Na2SO4, filtered and concentrated. The
crude product was loaded onto silica and purified by column chromatography
(DCM/MeOH, 5:1) to yield the desired product compound 3 as colorless oil (24 mg,
yield 72.7%).
4.1.2 Structural Characterization of the Target Compounds
(S)-2-Benzyl-((S)-3-mercapto-2-methylpropanamido)acetic acid (3). The title
compound was obtained by the general procedures as described above using
L-phenylalanine (1 g, 6.05 mmol), 3 was isolated as colorless oil (24 mg, yield
72.7%), after column chromatography (DCM/MeOH 5 : 1), 96.5% HPLC purity. mp: >
o
300 C. 1H NMR (400 MHz, CDCl3) δ 8.14 (s, 1H), 7.24 – 7.1 (m, 5H), 4.65 (s, 1H),
23

ACCEPTED MANUSCRIPT
4.40 - 4.36 (m, 1H), 3.07-3.04 (m, 1H), 2.91-2.87 (m, 1H), 2.77-2.73 (m, 1H),
2.65-2.62 (m, 1H), 1.03 (d, J = 6.7 Hz, 3H); 13C NMR (100 MHz, DMSO-d6) δ 174.2,
173.5, 138.1, 129.6, 128.6, 126.8, 53.6, 43.5, 37.1, 27.8, 17.3; HRMS (ESI) Calcd. for
C13H17NO3S [M - H]- 266.0856, found 266.0854.
(S)-4-Hydroxybenzyl-((S)-3-mercapto-2-methylpropanamido)acetic acid (4). The title
compound was obtained by the general procedures as described above. Starting from
L-tyrosine (1 g, 0.006 mol), after first step, white solid was formed Methyl
L-tyrosinate hydrochloride (1.28g, quantitative yield). Next, Methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-L-tyrosinate was obtained as white solid (258
mg, 88.4% yield) according to the general second step with methyl L-tyrosinate
hydrochloride (200 mg, 0.86 mmol).
4
was synthesized from Methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-L-tyrosinate (40 mg, 0.118 mmol) according
to the general third step, isolated as colorless oil (25 mg, 75.1%) after column
o
1
chromatography (DCM/MeOH 5 : 1), 96.5% HPLC purity. mp: > 300 C. H NMR
(400 MHz, CDCl3) δ 12.49 (s, 1H), 9.18 (s, 1H), 8.18-8.13 (m, 1H), 7.18 (d, J = 8.2
Hz, 2H), 6.74 (d, J = 8.2 Hz, 2H), 5.22-5.16 (m, 1H), 4.14-3.99(m, 1H), 2.74 – 2.57
13
(m, 2H), 2.41-2.31 (m, 2H), 2.18 (t, J = 8.1 Hz, 1H), 1.01 (d, J = 6.4 Hz, 3H); C
NMR (100 MHz, DMSO-d6) δ 179.0, 178.2, 161.0, 135.2, 132.8 (d, J = 12 Hz),
120.1, 58.7, 48.2, 41.2, 32.5, 22.4; HRMS (ESI) Calcd. for C13H17NO4S [M + Na]+
306.0770, found 305.0793.
(S)-2-Phenyl-2-((S)-3-mercapto-2-methylpropanamido)acetic acid (5). The title
compound was obtained by the general procedures as described above. Starting from
24

ACCEPTED MANUSCRIPT
L-phenylglycine (1 g, 0.007 mol), after first step, white solid was formed Methyl
(S)-2-amino-2-phenylacetate hydrochloride (1.30g, 97.7% yield). Next, Methyl
(S)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-phenylacetate was obtained as
white solid (247 mg, 80.4% yield) according to the general second step with Methyl
(S)-2-amino-2-phenylacetate hydrochloride (200 mg, 0.992 mmol). 5 was synthesized
from Methyl (S)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-phenylacetate (40 mg,
0.129 mmol) according to the general third step, isolated as colorless oil (22 mg,
67.2%) after column chromatography (DCM/MeOH 5 : 1), 97.6% HPLC purity. mp: >
o
1
300 C. H NMR (400 MHz, DMSO-d6) δ 12.74 (s, 1H), 8.67 (d, J = 7.2 Hz, 1H),
7.47 – 7.26 (m, 5H), 5.34 (d, J = 7.2 Hz, 1H), 2.75 – 2.57 (m, 2H), 2.48 – 2.37 (m,
1H), 2.08 (t, J = 7.7 Hz, 1H), 1.02 (d, J = 6.2 Hz, 3H); 13C NMR (100 MHz,
DMSO-d6) δ 174.3, 172.3, 137.9, 128.9, 128.0, 127.8, 56.6, 43.0, 27.9, 17.3; HRMS
(ESI) Calcd. for C12H15NO3S [M - H]- 252.0700, found 252.0686 .
(S)-2-(4-Fluorophenyl)-2-((S)-3-mercapto-2-methylpropanamido)acetic acid (6). The
title compound was obtained by the general procedures as described above. Starting
from L-4-fluorophenylglycine (1 g, 5.91 mmol), after first step, white solid was
formed Methyl (S)-2-amino-2-(4-fluorophenyl) acetate hydrochloride (1.29g,
quantitative
yield).
Next,
Methyl
(S)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-phenylacetate was obtained as
white solid (248 mg, 75.8% yield) according to the general second step with Methyl
(S)-2-amino-2-(4-fluorophenyl) acetate hydrochloride (200 mg, 1.09 mmol). 6 was
synthesized from Methyl (S)-2-((S)-3-(acetylthio)-2-methylpropanamido)
25

ACCEPTED MANUSCRIPT
-2-(4-fluorophenyl)acetate(30 mg, 0.092 mmol) according to the general third step,
isolated as colorless oil (17 mg, 60.7%) after column chromatography (DCM/MeOH
5 : 1), 97.2% HPLC purity. mp: > 300 oC. 1H NMR (400 MHz, DMSO-d6) δ 12.83 (s,
1H), 8.66 (d, J = 7.2 Hz, 1H), 7.47-7.44 (m, 2H), 7.21 (t, J = 8.4 Hz, 2H), 5.37 (d, J =
7.2 Hz, 1H), 2.64-2.62 (m, 2H), 2.47 – 2.37 (m, 1H), 2.08 (t, J = 7.4 Hz, 1H), 1.08 (d,
J = 5.3 Hz, 3H); 13C NMR (100 MHz, DMSO-d6) δ 174.6, 172.2, 162.2 (d, J = 161
Hz), 133.7 (d, J = 19 Hz), 130.2 (d, J = 5.0 Hz), 115.7 (d, J = 14 Hz), 55.9, 43.2, 27.7,
17.6; 19F NMR (376 MHz, CDCl3): δ = -114.4; HRMS (ESI) Calcd. for C12H14FNO3S
[M - H]- 270.0606, found 270.0598.
(S)-2-(4-Hydroxyphenyl)-2-((S)-3-Mercapto-2-methylpropanamido)acetic acid (7).
The title compound was obtained by the general procedures as described above.
Starting from L-4-hydroxyphenylglycine (1 g, 5.98 mmol), after first step, white solid
was formed Methyl (S)-2-amino-2-(4-hydroxyphenyl) acetate hydrochloride(1.31 g,
quantitative
yield).
Next,
Methyl
(S)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-(4-hydroxyphenyl) acetate was
obtained as white solid (249 mg, 76.6% yield) according to the general second step
with Methyl (S)-2-amino-2-(4-hydroxyphenyl) acetate hydrochloride (200 mg, 1.10
mmol).
7
was
synthesized
from
Methyl
(S)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-(4-hydroxyphenyl) acetate (30 mg,
0.092 mmol) according to the general third step, isolated as colorless oil (18 mg,
72.5%) after column chromatography (DCM/MeOH 5 : 1), 97.0% HPLC purity. mp: >
300 oC. 1H NMR (600 MHz, DMSO-d6) δ 12.58 (s, 1H), 9.49 (s, 1H), 8.47 (d, J = 7.1
26

ACCEPTED MANUSCRIPT
Hz, 1H), 7.19 (d, J = 8.0 Hz, 2H), 6.74 (d, J = 8.0 Hz, 2H), 5.19 (d, J = 7.1 Hz, 1H),
2.69 – 2.57 (m, 2H), 2.45 – 2.37 (m, 1H), 2.07 (t, J = 7.5 Hz, 1H), 1.07 (d, J = 5.9 Hz,
3H); 13C NMR (100 MHz, DMSO-d6) δ 179.1, 177.5, 162.3, 134.1, 132.3, 120.3,
60.9, 47.8, 32.47, 22.38; LC-MS: m/z 270.1 [M + H]+.
(R)-2-Benzyl-((S)-3-mercapto-2-methylpropanamido)acetic acid (8). The title
compound was obtained by the general procedures as described above. Starting from
D-phenylalanine (1 g, 6.05 mmol), after first step, white solid was formed Methyl
D-phenylalaninate hydrochloride (1.22 g, quantitative yield). Next, Methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-D-phenylalaninate was obtained as white
solid (278mg, 86.6% yield) according to the general second step with Methyl
D-phenylalaninate hydrochloride (200 mg, 0.99 mmol). 8 was synthesized from
Methyl ((S)-3-(acetylthio)-2-methylpropanoyl)-D-phenylalaninate (40 mg, 0.124
mmol) according to the general third step, isolated as colorless oil (27 mg, 81.8%)
after column chromatography (DCM/MeOH 5 : 1), 96.9% HPLC purity. mp: > 300 oC.
1H NMR (600 MHz, DMSO-d6) δ 12.66 (s, 1H), 8.23 (d, J = 7.2 Hz, 1H), 7.33 – 7.12
(m, 5H), 4.47-4.41 (m, 1H), 3.10-3.06 (m, 1H), 2.89–2.83 (m, 1H), 2.64-2.55 (m, 1H),
2.46 – 2.41 (m, 1H), 2.38 – 2.30 (m, 1H), 2.08 (t, J = 7.7 Hz, 1H), 0.88 (d, J = 6.4 Hz,
3H); 13C NMR (100 MHz, DMSO-d6) δ 174.2, 173.5, 138.0, 129.5, 128.4, 126.7,
53.5, 43.5, 37.1, 27.7, 17.4; HRMS (ESI) Calcd. for C13H17NO3S [M - H]- 266.0856,
found 266.0859 .
(R)-4-Hydroxybenzyl-((S)-3-mercapto-2-methylpropanamido)acetic acid (9). The title
compound was obtained by the general procedures as described above. Starting from
27

ACCEPTED MANUSCRIPT
D-tyrosine (500 mg, 2.76 mmol), after first step, white solid was formed Methyl
D-tyrosinate
hydrochloride(1.25g,
97.8%
yield).
Next,
Methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-D-tyrosinate was obtained as white solid (198
mg, 68.3% yield) according to the general second step with methyl D-tyrosinate
hydrochloride (200 mg, 0.86 mmol).
9
was synthesized from Methyl
((S)-3-(acetylthio)-2-methylpropanoyl)-D-tyrosinate (40 mg, 0.118 mmol) according
to the general third step, isolated as a colorless oil (21 mg, 63.6%) after column
o
1
chromatography (DCM/MeOH 5 : 1), 96.3% HPLC purity. mp: > 300 C. H NMR
(600 MHz, DMSO-d6) δ 12.60 (s, 1H), 9.18 (s, 1H), 8.15 (d, J = 8.1 Hz, 1H), 7.01 (d,
J = 7.7 Hz, 2H), 6.64 (d, J = 7.7 Hz, 2H), 4.34 (d, J = 3.6 Hz, 1H), 2.94 (dd, J = 13.7,
3.9 Hz, 1H), 2.73 (dd, J = 13.5, 9.7 Hz, 1H), 2.65 – 2.55 (m, 1H), 2.45 (dd, J = 13.5,
6.7 Hz, 1H), 2.41 – 2.30 (m, 1H), 2.08 (t, J = 8.1 Hz, 1H), 0.91 (d, J = 6.8 Hz, 3H);
13C NMR (100 MHz, DMSO-d6) δ 175.3, 173.8, 156.0, 130.5, 128.9, 115.0, 55.1,
43.6, 36.7, 27.9, 17.6; HRMS (ESI) Calcd. for C13H17NO4S [M + NH4]+ 301.1217,
found 301.1422.
(R)-2-Phenyl-2-((S)-3-mercapto-2-methylpropanamido)acetic acid (10). The title
compound was obtained by the general procedures as described above. Starting from
D-phenylglycine (1 g, 0.007 mol), after first step, white solid was formed Methyl
(R)-2-amino-2-phenylacetate hydrochloride (1.32g, 98.9% yield). Next, Methyl
(R)-2-((S)-3-(acetylthio)-2-methylpropanamido)-2-phenylacetate was obtained as
white solid (260 mg, 84.7% yield) according to the general second step with Methyl
(R)-2-amino-2-phenylacetate hydrochloride (200 mg, 0.992 mmol). 10 was
28