Bioisosteric investigation of ebselen: Synthesis and in vitro characterization of 1,2-benzisothiazol-3(2H)-one derivatives as potent New Delhi metallo-β-lactamase inhibitors

Article abstract of DOI:10.1016/j.bioorg.2020.103873

Carbapenem-resistant Enterobacteriaceae (CRE) producing New Delhi metallo-β-lactamase (NDM-1) cause untreatable bacterial infections, posing a significant threat to human health. In the present study, by employing the concept of bioisosteric replacement of the selenium moiety of ebselen, we have designed, synthesized and characterized a small compound library of 2-substituted 1,2-benzisothiazol-3(2H)-one derivatives and related compounds for evaluating their cytotoxicity and synergistic activity in combination with meropenem against the E. coli Tg1 (NDM-1) strain. The most promising compound 3a demonstrated potent synergistic activity against a panel of clinically isolated NDM-1 positive CRE strains with FICI as low as 0.09. Moreover, its IC₅₀ value and inhibition mechanism were also confirmed by using the enzyme inhibition assay and the ESI-MS analysis respectively. Importantly, compound 3a has acceptable toxicity and is not a PAINS. Because of its structural simplicity and potent synergistic activity in combination with meropenem, we propose that compound 3a may be a promising meropenem adjuvant and a new series of such compounds may worth further investigations.

Full text of DOI:10.1016/j.bioorg.2020.103873

Journal Pre-proofs
Bioisosteric investigation of ebselen: synthesis and in vitro characterization of
1
,2-benzisothiazol-3(2H)-one derivatives as potent New Delhi metallo-β-lac-
tamase inhibitors
Wen Bin Jin, Chen Xu, Qipeng Cheung, Wei Gao, Ping Zeng, Jun Liu,
Edward W.C. Chan, Yun-Chung Leung, Tak Hang Chan, Kwok-Yin Wong,
Sheng Chen, Kin-Fai Chan
PII:
S0045-2068(20)30194-2
DOI:
https://doi.org/10.1016/j.bioorg.2020.103873
YBIOO 103873
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
25 January 2020
7 April 2020
21 April 2020
Please cite this article as: W. Bin Jin, C. Xu, Q. Cheung, W. Gao, P. Zeng, J. Liu, E.W.C. Chan, Y-C. Leung, T.
Hang Chan, K-Y. Wong, S. Chen, K-F. Chan, Bioisosteric investigation of ebselen: synthesis and in vitro
characterization of 1,2-benzisothiazol-3(2H)-one derivatives as potent New Delhi metallo-β-lactamase inhibitors,
Bioorganic Chemistry (2020), doi: https://doi.org/10.1016/j.bioorg.2020.103873
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
©
2020 Elsevier Inc. All rights reserved.

Bioisosteric investigation of ebselen: synthesis and
in vitro characterization of 1,2-benzisothiazol-
3
(2H)-one derivatives as potent New Delhi metallo-
β-lactamase inhibitors
Wen Bin Jin, ‡,1,2 Chen Xu,‡,1,3 Qipeng Cheung, 1,3 Wei Gao, Ping Zeng, Jun Liu, Edward W. C.
1
1
1
1
1
1,4
1
,5
Chan, Yun-Chung Leung, Tak Hang Chan, Kwok-Yin Wong, Sheng Chen* and Kin-Fai
Chan*,1
1
State Key Laboratory of Chemical Biology and Drug Discovery and Department of Applied
Biology and Chemical Technology, The Hong Kong Polytechnic University, Hung Hom,
Kowloon, Hong Kong SAR, China
2
Faculty of Chinese Materia Medica, Yunnan University of Chinese Medicine, Kunming,
Yunnan, China
3
Shenzhen Key Laboratory for Food Biological Safety Control, Food Safety and Technology
Research Centre, The Hong Kong PolyU Shenzhen Research Institute, Shenzhen, China
4
Department of Chemistry, McGill University, Montreal, Quebec, H3A 2K6, Canada
1

5
Department of Infectious Diseases and Public Health, Jockey Club College of Veterinary
Medicine and Life Sciences, City University of Hong Kong, Kowloon, Hong Kong SAR, China
‡These authors contributed equally.
Corresponding authors: Kin-Fai Chan and Sheng Chen
*
For K.-F. C., email: kf.chan@polyu.edu.hk
For S. C., email: shechen@cityu.edu.hk
2

ABSTRACT
Carbapenem-resistant Enterobacteriaceae (CRE) producing New Delhi metallo-β-lactamase
(NDM-1) cause untreatable bacterial infections, posing a significant threat to human health. In
the present study, by employing the concept of bioisosteric replacement of the selenium moiety
of ebselen, we have designed, synthesized and characterized a small compound library of 2-
substituted 1,2-benzisothiazol-3(2H)-one derivatives and related compounds for evaluating their
cytotoxicity and synergistic activity in combination with meropenem against the E. coli Tg1
(NDM-1) strain. The most promising compound 3a demonstrated potent synergistic activity
against a panel of clinically isolated NDM-1 positive CRE strains with FICI as low as 0.09.
Moreover, its IC value and inhibition mechanism were also confirmed by using the enzyme
5
0
inhibition assay and the ESI-MS analysis respectively. Importantly, compound 3a has acceptable
toxicity and is not a PAINS. Because of its structural simplicity and potent synergistic activity in
combination with meropenem, we propose that compound 3a may be a promising meropenem
adjuvant and a new series of such compounds may worth further investigations.
Keywords: Carbapenem-resistant Enterobacteriaceae; New Delhi metallo-β-lactamase
inhibitors; 1,2-benzisothiazol-3(2H)-one; bioisosteric replacement; ebselen.
3

Introduction
The global emergence and rapid spread of multidrug-resistant Gram-negative bacteria over the
past decades have posed a significant threat to our human health, particularly carbapenem-
resistant Enterobacteriaceae (CRE).1, 2 Among all the mechanisms of antibiotic resistance, such
as enzymatic degradation of drug, mutation of drug target, overexpression of membrane efflux
pumps, and reduced permeability of bacterial cell wall,3-6 the vast array of production of
hydrolytic enzymes, namely β-lactamases, in these pathogenic bacteria ranks first position in the
7
clinical settings. According to their amino acid sequences and functional mechanisms, β-
8
lactamases were systematically divided into four classes (A, B, C, and D). β-Lactamases of class
A, C, and D are termed serine β-lactamases (SBLs) as they generally perform a nucleophilic
attack on the carbonyl group of the β-lactam ring via the hydroxyl group of active-site serine.
These SBLs can be inhibited covalently by FDA-approved drugs, such as clavulanic acid,
sulbactam, tazobactam, avibactam, and vaborbactam.9-12 Unlike many SBLs, β-lactamases of
class B contain metal ions in the active site and are, therefore, termed metallo-β-lactamases
(MBLs). They attack the carbonyl group of the β-lactam ring through a zinc-bound hydroxyl ion
in a non-covalent manner.13 MBLs are even capable to inactivate the third-generation
1
1
cephalosporin antibiotics and the most powerful carbapenem antibiotics. Depending on the
amino acid sequences and zinc content, MBLs can be further classified into three subclasses B1,
1
4
B2, and B3. Particularly, New Delhi metallo-β-lactamase (NDM-1), a member of subclass B1
β-lactamases that was first isolated from Klebsiella pneumoniae in 2008, has aroused a
worldwide public health concern over the rapid dissemination of blaNDM-1 gene.15-18 More
importantly, the Center for Disease Control and Prevention has declared that the golden age of
antibiotics has come to an end since doctors have run out the treatment options for these so-
4

called “superbugs” carrying the blaNDM-1 gene, except for polymyxin, tigecycline but with
1
9
obvious toxicity. Crystallographic analysis of NDM-1 revealed that it contains one hydroxide
ion stabilized by two adjacent zinc ions in the active site, in which both zinc ions are associated
with the 3H site (His116, His118, His196) and the DCH site (Asp120, Cys221, His263) in a
2
0
tetrahedral geometry (Figure 1A). Zinc ion in the 3H site is tightly bound while zinc ion in the
DHC site is weakly coordinated. More interestingly, NDM-1 can be secreted outside the outer
membrane of bacteria via outer membrane vesicles, which could also help to confer drug
resistance to nearby pathogens.21
Figure 1. (A) Crystal structure of NDM-1 (PDB ID: 3Q6X) showing the active site of the
dinuclear zinc center, the 3H site (His116, His118, His196) and the DCH site (Asp120, Cys221,
His263). (B) Chemical structures of NDM-1 inhibitors reported in the literature. The functional
groups highlighted in blue are potential zinc ion chelating moieties.
Despite the tremendous effort to discover potential NDM-1 inhibitors, FDA-approved drugs
that target specifically NDM-1 are still unavailable.22, 23 Only bicyclic boronate taniborbactam
5

(VNRX-5133), combined with cefepime to treat complex urinary tract infection caused by
resistant bacteria such as CRE and carbapenem-resistant Pseudomonas aeruginosa carrying SBLs
2
4
and MBLs, has entered phase III clinical study. Yang has recently reported that both zinc ions
in the NDM-1 play essential roles during the hydrolysis of β-lactam antibiotics upon
investigating the effects of hetero-bimetallic analogs including CoCo- and ZnCo- styles of
2
5
NDM-1. As shown in Figure 1B, Klingler has systemically screened the FDA-approved thiol-
containing drugs including captopril, thiorphan, dimercaprol, and penicillamine as NDM-1
2
6
inhibitors. The pharmacophore of captopril, a subunit of 3-mercapto-2-methylpropanoic acid,
2
7
demonstrated moderate inhibitory activity against NDM-1. The bisthiazolidine compound L-
CS319 with a substrate-mimicking scaffold of thiol and carboxylate moieties is able to inhibit
NDM-1 activity.28 Dipicolinic acid derivatives with carboxylate moieties at C-2 and C-6
positions also represent a novel class of non-toxic and potent NDM-1 inhibitors.29, 30 Several
sulfur-containing heterocycles, such as 1-hydroxypyridine-2-thione, 1,2,4-triazole-3-thione and
azolylthioacetamides were also found to be potent inhibitors of MBLs.31-34 Of the greatest
concern is the discovery of the selective and potent NDM-1 inhibitor, aspergillomarasmine A
3
5-38
(
AMA), with a half maximal inhibitory concentration (IC ) value of 4.0 ± 1.0 μM.
These
5
0
results implied that small molecules possessing the property of bivalent zinc-chelating moieties
may exhibit potent inhibition against NDM-1. The antimicrobial peptide, thanatin, capable of
extracting the zinc ions in the NDM-1 also provides an alternative strategy for the NDM-1
inhibitor design.39
Our previous study revealed that an organoselenium compound ebselen (Ebs) exhibited potent
inhibition of NDM-1 by covalently binding to the Cys221 residue in the active site of the NDM-
6

1
enzyme.40 Further study of the structure-activity relationship (SAR) of Ebs led to the
identification of another promising Ebs derivative, namely compound 11_a38, which exhibited
more potent synergistic activity by restoring the antibacterial activity of meropenem (Mem)
against clinical NDM-1 positive CRE strains in a Galleria mellonella larvae infection model.41
More importantly, a phase 2 clinical trial of Ebs indicated that treatment with Ebs at a daily
dosage of 400 mg twice in humans was safe, suggesting that this type of organoselenium
compound is a promising lead without toxicity concern.42, 43 On the other hand, whether the five-
member ring of selenazol-3(2H)-one of Ebs derivatives replaced with other structurally related
analogs still active towards clinical NDM-1 positive CRE strains remains unclear. In the present
study, by employing the concept of bioisosteric replacement of the selenium moiety of Ebs, we
have designed and synthesized a small compound library of 2-substituted 1,2-benzisothiazol-
3
(2H)-one derivatives and related compounds for comparing their cytotoxicity and synergistic
activity in combination with Mem against the E. coli Tg1 (NDM-1) strain with Ebs and
compound 11_a38. The most promising compound 3a has been further tested towards a panel of
clinically isolated NDM-1 positive CRE strains. Moreover, its IC50 value and inhibition
mechanism were also confirmed by using the enzyme inhibition assay and the ESI-MS analysis,
respectively. Because of its structural simplicity and potent synergistic activity in combination
with Mem, a new series of such compounds may worth further investigations.
Results and Discussions
1
. Design and chemical synthesis
Notably, the New Delhi metallo-β-lactamase (NDM-1) is a powerful β-lactamase that
composes largely of hydrophobic amino acid residues and highly flexible loops in the active site,
7

forming a relatively long and shallow drug-binding pocket that can accommodate and hydrolyze
a wide range of substrates. Our compound design strategy was to replace the core structure of
selenazol-3(2H)-one of Ebs with other substituted five-member rings of higher hydrophobicity
and smaller polar surface area to strengthen the interactions of van der Waals contact and
aromatic π-π stacking between the ligand and the DNM-1 enzyme.
Based on a previous report,44 a concise synthetic route for the preparation of various 2-
substituted 1,2-benzisothiazol-3(2H)-one analogs 3a-3x is depicted in Scheme 1. Starting from
the commercially available 2, 2’-dithiodibenzoyl acid 1a, conversion to acid chloride 2a by using
oxalyl chloride in the presence of a catalytic amount of dimethylformamide (DMF) followed by
subsequent amidation with a wide range of primary aryl and alkyl amines and intramolecular
cyclization of diamide under molecular catalysis of bromine using dichloromethane (DCM) as a
solvent, we successfully obtained the target molecules in three steps with reasonably good yields.
Such efficient syntheses of 2-substituted 1,2-benzisothiazol-3(2H)-one analogs allow the rapid
construction of the target compound library for subsequent biological studies. Twenty-four 1,2-
benzisothiazol-3(2H)-one derivatives were synthesized and their chemical structures are shown
in Table 1 respectively. Similarly, we also designed and synthesized compounds 5a-5d to
investigate the bioisosteric replacement of the sulfur atom with an oxygen atom, methylene
group, carbonyl group, and sulfinyl group respectively. Moreover, compound 5e with a pyridine
ring and compound 5f without a phenyl ring were synthesized to explore the importance of the
phenyl ring of 1,2-benzisothiazol-3(2H)-one analogs. Benzo[d]isoxazol-3(2H)-one 5a were
prepared in a moderate yield by nucleophilic alkylation of 3-hydroxybenzisoxazole (4a) with 4-
(trifluoromethyl)benzyl bromide in the presence of sodium hydride using DMF as the solvent.
An undesired side product 5g was also obtained as the consequence of the alkylation of O-
8

nucleophile B. Interestingly, alkylation of 1-isoindolinone 4b with 4-chlorobenzyl bromide under
the same condition gave only isoindolin-1-one 5b without the isolation of O-alkylated side
product. Coupling of potassium phthalimide 4c with 1-(3-(bromomethyl)phenyl)piperidine gave
the isoindoline-1,3-dione 5c in a good yield. Oxidation of 1,2-benzisothiazol-3(2H)-one 3a with
3
-chloroperoxybenzoic acid (m-CPBA) in an ice bath gave the 1,2-benzisothiazol-3(2H)-one-1-
oxide 5d in a low yield. Substitution of 2-chloronicotinonitrile 6 with thiourea, followed by
oxidative cyclization in concentrated sulfuric acid and subsequent alkylation with benzyl
bromide in the presence of potassium carbonate produced isothiazolo[5,4-b]pyridin-3(2H)-one
5
e with a moderate yield. Similarly, treatment of 3,3'-disulfanediyldipropanoic acid 1b with
thionyl chloride in the presence of a catalytic amount of DMF afforded the acid chloride 2b,
which was further reacted with excess 4-chloroaniline gave the diamide, followed by subsequent
chlorine-assisted cyclization of the diamide with sulfuryl chloride using DCM as a solvent in an
ice bath afforded the desired isothiazol-3(2H)-one 5f in overall good yield.45, 46
Scheme 1. The chemical synthesis of 1,2-benzisothiazol-3(2H)-one derivatives 3a-3x and related
bioisosteric analogs 5a-f.a
9

O
O
S
OH
S
Cl
S
O
a
b
S
N R
HO
Cl
S
O
O
1a
2a
3a-3x
O
O
O
O
O
c
NH
N
N
+
N
N
O
O
O
O
O
CF3
4
a
b
5a
5g
A
B
O
O
CF3
d
NH
N
4
5b
O
O
Cl
e
NK
N
O
c
O
N
4
5c
O
O
f
N
Cl
N
Cl
S
S
O
3a
5d
O
CN
Cl
CN
SH
g
h
N
N
N
N
S
6
7
5e
O
O
O
i
j
HO
S
Cl
S
N
Cl
S
OH
S
Cl
S
O
O
1b
5f
2
b
aReagents and reaction condition: (a) oxalyl chloride, DMF (cat.), DCM, rt, 3h. (b) (1) various
amine (excess), triethylamine, DCM, 0°C to rt, overnight; (2) Br , DCM, 0°C to rt, overnight,
2
3
(
8%-75%. The chemical structures of 3a-x were shown in Table 1. (c) NaH, DMF, 4-
trifluoromethyl)benzyl bromide, rt, overnight, 33% (5a); (d) NaH, DMF, 4-chlorobenzyl
bromide, rt, overnight, 60%; (e) DMF, 1-(3-(bromomethyl)phenyl)-piperidine, rt, overnight,
2%; (f) m-CPBA, DCM, 0°C, 25%; (g) thiourea, n-butanol, 4 h, reflux; (h) (1) conc. H SO ,
6
2
4
reflux, 4 h, (2) K CO , benzyl bromide, Et N, KI, DCM, rt, overnight, 55% over 3 steps. (i)
2
3
3
SOCl , DMF (cat.), DCM, 0°C; (j) (1) excess 4-chloroaniline, DCM; (2) SO Cl , DCM, 0°C, 3 h,
2
2
2
4
7% over 3 steps.
1
0

2
. SAR study of 1,2-benzisothiazol-3(2H)-one analogs
In this study, carbapenem-resistant E. coli Tg1 (NDM-1) strain generated by integrating the
full-length blaNDM-1 gene isolated from a clinical Klebsiella pneumoniae strain into a wild type E.
coli Tg1 strain was employed for preliminary MICs screening. This E. coli Tg1 (NDM-1) strain
(Entry 1, Table 1) was highly resistant to Mem than its wild type E. coli Tg1 strain (MIC = 0.5
µg/mL) by 128-fold. The fractional inhibitory concentration index (FICI), used to quantify the
4
7
synergy between Mem and adjuvant, was calculated as previously described. The FICI ≤ 0.5 is
considered as a synergistic interaction. A total of thirty 1,2-benzisothiazol-3(2H)-one derivatives
were simultaneously subjected to assessment of their antimicrobial activity and competence to
restore the antimicrobial activity of Mem against E. coli Tg1 (NDM-1) strain by using standard
MICs screening. These results were summarized as shown in Table 1. Ebs and 11a_38 with FICI
values of 0.50 and 0.05 respectively were employed as positive controls (Entry 2-3, Table 1). In
general, the results of FICI showed that all tested 1,2-benzisothiazol-3(2H)-one analogs
exhibited a weaker synergistic activity than that of 11a_38 and only twelve 1,2-benzisothiazol-
3
(2H)-one analogs (Entry 4-15, Table 1) exhibited stronger synergistic activity than that of Ebs.
Moreover, the results of the compound MICs revealed that most of the analogs exhibit no
antimicrobial activity with MICs ≥ 128 µg/mL, implying that they are potential Mem adjuvants
of relatively non-cytotoxic and well-tolerated (Entry 4-25, Table 1). However, eight analogs
showed some extent of antimicrobial activity with MICs ranged from 2 µg/mL to 64 µg/mL,
suggesting that they may have unexpected interactions with protein targets other than NDM-1,
thus causing undesired off-target effects (Entry 26-33, Table 1). Therefore, these compounds are
not worthy to pursue any further.
1
1

Table 1. cLogP, tPSA and MIC screening of compounds 3a-3x, 5a-5f, 11_a38, Ebs and their
combination with Mem against E. coli Tg1 (NDM-1) strain.a
O
N
R
S
3a-3x
MIC (µg/mL)
Cpd Cpd + Mem
Cpd
No.
Entry
R
cLogP tPSA FICI
1
2
Mem
Ebs
N.A.
See Figure 1
See Figure 1
Cl
64
64
N.A.
16
2
N.A.
3.70
3.04
4.07
N.A. N.A.
20.31 0.50
58.64 0.05
20.31 0.09
3
4
5
6
11_a38
3a
≥128
≥128
4
Br
3b
3c
≥128
≥128
8
8
4.22
4.24
20.31 0.19
20.31 0.19
CF3
Cl
7
3d
≥128
8
4.66
20.31 0.19
Cl
Cl
8
9
3e
3h
3t
≥128
≥128
≥128
8
8
8
4.07
4.29
4.79
20.31 0.19
20.31 0.19
20.31 0.19
Cl
1
0
OH
1
1
1
1
1
2
3
4
3g
3i
3j
3l
≥128
≥128
≥128
≥128
16
16
16
16
2.32
3.35
3.27
4.78
40.54 0.38
20.31 0.38
29.54 0.38
20.31 0.38
OMe
1
1
5
6
3n
3k
≥128
≥128
16
32
4.38
3.90
20.31 0.38
29.54 0.75
Cl
OMe
NH
O
O
1
1
7
8
3o
3q
≥128
≥128
32
32
2.67
1.54
58.64 0.75
32.78 0.75
N
O
1
2

O
S
F
1
2
9
0
3r
≥128
≥128
32
32
2.71
2.92
54.54 0.75
58.64 0.75
O
O
O
3w
HN
O
Cl
2
2
1
2
5b
5d
≥128
≥128
32
32
3.40
2.18
20.31 0.75
37.88 0.75
N
O
N
Cl
S
O
F
2
2
3
4
3f
≥128
≥128
64
64
3.50
3.92
20.31 1.50
29.54 1.50
CF3
O
5a
N
O
O
N
N
2
5
5c
≥128
≥128
4.12
40.64 3.00
O
2
2
2
2
6
7
8
9
3s
3m
3v
64
64
32
32
8
32
8
3.52
2.82
1.58
0.61
20.31 0.25
29.54 1.00
20.31 0.38
29.10 0.38
OMe
3x
H
N
8
O
3
0
5e
32
8
1.41
32.67 0.38
N
S
OH
3
3
1
2
3u
3p
32
16
16
8
0.79
1.74
40.54 0.75
23.55 0.63
N
O
3
3
5f
2
1
2.47
20.31 0.52
N
Cl
S
a
cLogP and tPSA were calculated by ChemDraw Ultra (Version 12.0). FICI was defined as FIC
cpd) + FIC (Mem), where FIC (cpd) was the ratio of MICcombination to MICcpd and FIC (Mem)
(
was the ratio of MICcombination to MICMem. The drug combination is considered synergistic if the
value of FICI is ≤ 0.5. MIC values were determined using the double dilution method in
accordance with the CLSI guidelines and previous reports. All MIC experiments were performed
at least triplicate. N.A.: Not Applicable.
1
3

It is worthy to notice that 1,2-benzisothiazol-3(2H)-one derivatives with hydrophobic N-phenyl
rings of relatively larger cLogP values (4.07-4.79) exhibited more potent synergistic activity than
that of Ebs (Entry 4-10, Table 1). Interestingly, these phenyl rings were all substituted with
hydrophobic electron-withdrawing groups, such as 4-chloro group (3a), 4-bromo group (3b), 4-
trifluoromethyl group (3c), 3,4-dichloro group (3d), 3-chloro group (3e), and 4-phenyl group
(3t). In addition, replacing the phenyl ring with hydrogen, compound 3x displayed weaker
synergy with a FICI of 0.38 (Entry 29, Table 1). These results are likely to imply that strong π-π
interactions via the N-phenyl rings may exist between these ligands and the NDM-1 enzyme.
Unexpectedly, compound 3f with a 2-fluorophenyl group demonstrated no synergistic activity
with a FICI of 1.5. We reasoned that it may due to the steric hindrance of the 2-fluorophenyl
group, which results in the sulfur atom of 3f less accessible to the nucleophilic attack of the thiol
group of Cys221 of the NDM-1 enzyme. On the other hand, the installation of electron-donating
groups on the phenyl ring at the same position, such as 4-hydroxymethyl (3g), 4-methoxyl (3j),
4
-isopropyl (3l), and 3-ethyl (3n) resulted in no significant improvement on the synergistic
activity with a FICI of 0.38 (Entry 11-15, Table 1). Compound 3k with a 3-chloro-4-methoxyl
group even displayed no synergy with a FICI of 0.75 (Entry 16, Table 1). Replacing the phenyl
ring with various linear flexible alkyl side-chains generated compounds 3o, 3q, and 3w showing
no synergistic activity (Entry 17-18, 20, Table 1). These results also are likely to imply that
flexible alkyl side-chains may not provide favorable interactions between the ligands and the
NDM-1 enzyme.
3
. Comparison of FICI between 1,2-benzisoselenazol-3(2H)-ones (Se analogs) and 1,2-
benzisothiazol-3(2H)-ones (S analogs)
1
4

4
1
We have constructed a small library of Ebs analogs for SAR study in the previous study. This
allows us to systemically compare the synergistic activity between both analogs of 1,2-
benzisoselenazol-3(2H)-one (Se analogs) and 1,2-benzisothiazol-3(2H)-one (S analogs). The
summarized results were shown in Figure 2. Among all the Se or S analogs that contain
hydrophobic N-phenyl rings (R group), the 4-chlorophenyl ring of Se or S analogs exhibited the
most potent synergistic activity. Other Se or S analogs, such as 4-hydroxymethyl, 4-
methoxyphenyl, and 4-isopropyl phenyl exhibited the same synergistic activity with a FICI value
of 0.38. Surprisingly, the Se analog with a 4-methoxy benzyl group or a CH CH NHBoc group
2
2
demonstrated more potent synergistic activity by 5-fold and 15-fold respectively than the S
analog, suggesting that Se analog may be a better Mem adjuvant than the S analog. We reasoned
that it is likely due to the fact that the N−Se bond has longer bond length than that of the N-S
bond, resulting in the benzisoselenazol-3(2H)-one are less stable and are more susceptible to the
ring-opening than the 1,2-benzisothiazol-3(2H)-one. To confirm this hypothesis, bioisosteres of
CH group (5b), oxygen atom (5a), and carbonyl group (5c) were tested and their bond lengths
2
of N−X bond were calculated by using Chem3D Pro. As illustrated in Table 2, shortening the
bond length of the N−X bond results in the increase of the FICI value. Compound 5a, 5b, and 5c
displayed no synergy in combination with Mem. Evidently, it seems to be in correlation with the
hypothetical trend that the shorter the bond length the higher the bond dissociation energy.
1
5

Figure 2. FICI comparison between the 1,2-benzisoselenazol-3(2H)-one analogs and the 1,2-
benzisothiazol-3(2H)-one analogs. The drug combination is considered synergistic if the value of
FICI is ≤ 0.5.
Table 2. Comparison of FICI and bond lengths of different structurally related analogs.a
O
N R
X
Cpd no
X =
Se
S
N−X
N−Se
N−S
N−C
N−O
N−C
Bond length of N−X FICI
1
3
5
5
5
1_a38
1.843 Å
1.705 Å
1.451 Å
1.320 Å
1.381 Å
0.05
0.19
0.75
1.50
3.00
a
b
a
c
CH
O
2
C=O
4
. In vitro studies of compound 3a
1
6

Next, to confirm whether the interaction between compound 3a and NDM-1 enzyme is the
primary cause of the observed synergy, the NDM-1 enzyme was expressed and purified for the
4
1
colorimetric nitrocefin assay as previously described. The IC value of compound 3a was
5
0
determined to be 0.44 ± 0.02 μM, which is smaller than that of AMA (IC = 4.0 ± 1.0 μM)
5
0
3
8
(
Figure 3A). The mode of inhibition of compound 3a could also be evaluated by the steady-
state kinetic studies. By plotting the velocity (reaction rate) against the nitrocefin concentration
substrate), Vmax (maximum rate) and K (Michaelis constant) were calculated according to
(
m
Michaelis equations. When the concentration of compound 3a increased, Km remained
unchanged and Vmax decreased, revealing that the NDM-1 inhibition of compound 3a is non-
competitive (Figure 3B). We discovered for the first time that compound 3a was an active site-
binding inhibitor of NDM-1, which exhibited a non-competitive inhibition. Additionally, the
reversibility of inhibition can be determined by measuring the recovery of enzymatic activity
after a rapid and large dilution assay. After a sudden 100-fold dilution of a 20 minutes pre-
incubation mixture of the enzyme at a concentration 100-fold of that used in the normal assay
and 10-fold of IC of compound 3a, approximately 46% residue hydrolytic activity of NDM-1
5
0
was observed (Figure 3C). This result implied that interaction between compound 3a and the
NDM-1 was slowly reversible, unlike the positive control compound Ebs which was irreversible
interaction. One of the possible explanations for the reversibility of compound 3a may be due to
the presence of hydroxyl ions in water after the rapid and large dilution. The covalent adduct of
compound 3a and enzyme may undergo disproportionation and give rise to the formation of
benzenesulfinate, mercatobenzamide, and the active enzyme respectively.
1
7

Figure 3. Biochemical and cytotoxic studies of compound 3a. (A) Inhibition of NDM-1 by
compound 3a. (B) The plot of the velocity of NDM-1 against nitrocefin concentration with
different concentrations of compound 3a. (C) Residual activity after fast dilution of the enzyme-
inhibitor complex. (D) Cytotoxicity of compound 3a against HEK-293T cell lines.
Given that the toxicity issue has been the greatest obstacle to the development of NDM-1
inhibitors, a human embryonic kidney cell line HEK-293T was selected to verify the safety
profile of compound 3a. As shown in Figure 3D, compound 3a displayed acceptable toxicity
against HEK-293T cell lines with 75% cell viability at a concentration of 10 mg/mL, which is
much higher than the effective concentration in combination study. Moreover, microscopic
analysis of cell morphology indicated that no obvious morphological changes of HEK-293T cells
1
8

were observed after prolonged incubation with compound 3a at a concentration of 10 mg/mL,
exhibiting negligible toxicity.
5
. ESI-MS analysis
The mechanism of the interaction between compound 3a and the NDM-1 enzyme indicated that
Ebs, Ebsulfur combined with rhodamine B and 11_a38 also interacted in a similar manner, in
which compound 3a is likely to bind covalently to the thiol group of Cys221 of the NDM-1.40, 41,
4
8
The results of ESI-MS analysis shown in Figure 4 further confirmed the hypothesis. The
molecular weight of the denatured NDM-1 enzyme was found to be 25867 Da. Incubating the
enzyme with compound 3a at a molar ratio of 1:2 led to an obvious mass shift of the complex to
a peak of molecular weight of 26130 Da, which is the sum of the molecular weight of NDM-1
enzyme and compound 3a. Furthermore, the addition of the reducing agent dithiothreitol (DTT)
into the complex and incubating overnight at room temperature resulted in the recovery of the
molecular weight of NDM-1, confirming the formation of the disulfide linkage between the
compound 3a and the NDM-1 enzyme.
1
9

Figure 4. ESI-MS analysis and their cartoon representations of (A) denatured NDM-1 enzyme,
B) denatured NDM-1 enzyme treated with compound 3a and (C) compound 3a-treated
denatured NDM-1 enzyme further treated with DTT.
(
6
. Computational docking studies of compound 3a
To gain more insights into the potential interactions of compound 3a in the active site of the
NDM-1 enzyme, the X-ray crystal structure of NDM-1 was employed to perform the
computational docking studies. As shown in Figure 5, the highest docking score positioned
2
0

compound 3a into the substrate-binding site of NDM-1, which is in good agreement with the
results of the enzymatic assays. Compound 3a was predicted to be well-situated in the substrate-
binding site, forming potential hydrogen-bonding interactions with Asp120 and the hydroxide
ion respectively. Moreover, strong π-π stacking interaction was also predicted to form between
the 4-chlorophenyl ring of compound 3a and the imidazole ring of His263. This result is in good
agreement with the results of SAR studies. The amino acid residues of Lys211, Ile35, Val37,
Trp93, Cys221, and Gln123 were closely adjacent to compound 3a, implying that they may also
be involved in the van der Waals interaction with compound 3a.
Figure 5. Model of compound 3a (green sticks) docked into the substrate-binding site of NDM-1
using the X-ray crystal structure of NDM-1 (PDB ID: 3Q6X) with labeled important amino acid
residues. The blue dashed lines indicated the hydrogen-bonding interactions.
2
1

7
. Combination of compound 3a and Mem against clinically isolated CRE strains
Next, the most potent compound 3a was further screened against a panel of 23 clinically
isolated CRE strains and the results were illustrated in Table 3. These CRE strains are highly
resistant to Mem antibiotic with MICs ranged from 32 µg/mL to ≥ 128 µg/mL as they produced
not only NDM-1 but also other β-lactamases such as VIM-1, IMP-4, KPC-2, CTX-M-3, CTX-
M-14, CTX-M-15, TEM-1, and SHV-12 to confer drug resistance. Encouragingly, compound 3a
was able to restore the antimicrobial activity of Mem against nearly all of these strains with FICI
ranging from 0.09 to 0.38, except for the strains KP04 and CF05. Compared with the positive
control Ebs, compound 3a exhibited much stronger synergistic activity against NDM-1 positive
meropenem-resistant clinical isolates.
Table 3. MIC (µg/mL) screening of 3a, Ebs, Mem and their combination with Mem against
clinically isolated CRE strains carrying NDM-1.a
O
O
N
Cl
N
S
Se
3a
Ebs
MIC (µg/mL)
MIC (µg/mL)
CRE
Strains
Additional β-
lactamases
Entry
3a +
Mem
FICI
Ebs + FICI
Mem
3a
Ebs
Mem
1
2
3
4
5
6
7
8
9
KP04
KP14
EC06
EC33
EC34
EC45
KO03
KO16
CF05
CF17
CF35
b5, b1, b7, b8 ≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
32
16
16
16
16
16
16
16
64
4
0.50
0.25
0.25
0.38
0.38
0.25
0.25
0.25
1.00
0.09
0.25
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
64
1.00
b5, b3, b8
b4, b5, b8
b6
≥128
≥128
64
≥128 2.00
64
16
16
16
32
32
16
16
32
1.00
0.38
0.38
0.25
0.50
0.50
0.25
0.38
0.50
b6, b7, b8
none
64
≥128
≥128
≥128
≥128
64
b8
b8
b8
1
0
1
b8
1
b4, b3, b8
≥128
16
2
2

1
1
1
1
1
1
1
1
2
2
2
2
2
3
4
5
6
7
8
9
0
1
2
3
CF43
none
≥128
64
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
16
16
8
0.25
0.38
0.32
0.25
0.38
0.32
0.25
0.25
0.32
0.32
0.09
0.38
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
≥128
32
16
32
0.50
0.38
1.25
ECL07 b5, b7, b8
ECL08 b5, b7, b8
32
ECL09 b4, b5, b7, b8 ≥128
16
16
8
≥128 2.00
ECL10 b5, b7, b8
ECL18 b2, b7
64
64
64
1.50
2.50
32
ECL19 b5, b3
≥128
≥128
16
16
8
≥128 2.00
≥128 2.00
ECL22 b5, b7, b8
ECL27 b4, b5, b7, b8 32
64
64
64
64
2.50
2.50
1.50
1.50
ECL28 b7, b8
MM23 none
MM26 none
32
64
64
8
4
16
aEC, Escherichia coli; KO, Klebsiella oxytoca; CF, Citrobacter freundii; EL, Enterobacter
cloacae; KP, Klebsiella pneumoniae; MM, Morganella morganii. Additional β-lactamase
determinants: b1: VIM-1, b2: IMP-4, b3: KPC-2, b4: CTX-M-3, b5: CTX-M-14, b6: CTX-M-15,
b7: TEM-1, and b8: SHV-12. The synergistic effect is depicted by the FIC index, which is
calculated as FIC (cpd) + FIC (Mem), where FIC (cpd) is the (MIC of cpd in combination with
Mem)/(MIC of cpd alone) while FIC (Mem) is (MIC of cpd in combination with Mem)/(MIC of
Mem alone). The MIC value was determined using the double dilution method in accordance
with the CLSI guidelines and previous reports. The drug combination is considered synergistic if
the FICI ≤0.5. All experiments were performed in at least triplicates.
8
. Compound 3a is not a PAINS
It is widely accepted that pan-assay interference compounds (PAINS) cause false-positive
assay signals due to reactivity under assay conditions and more than 450 compound classes have
been designated as PAINS including 1,2-benzisothiazol-3(2H)-one. However, it has been
4
9
confirmed that not all 1,2-benzisothiazol-3(2H)-ones are PAINS. More importantly, the activity
of compound 3a was proved to be dependent on the ratio of compound and Mem. It was found
that increasing the Mem to compound 3a ratio from the 1:0.15 to 1:6 could further reduce the
MIC by 256 fold (Table S1), revealing the dose-dependent effect of 3a. As shown in Table 4,
we also demonstrated that compound 3a did not act as a colloidal aggregator which was a
2
3

dominant source of artifacts. The results of MIC assay in the presence or absence of 0.1% or
.01% (v/v) non-ionic detergent Triton X-100 suggested that the activity of compound 3a was
0
not attenuated by the addition of detergent, suggesting that compound 3a was not an aggregator.
The database of aggregator advisor further provide stronger support.50
Table 4. MIC values of different treatments against E.coli Tg1 (NDM-1) strain.
Treatment
MIC (µg/mL)
Triton X-100 (2%)
≥128
≥128
64
3
a
Mem
a + Mem
3
4
Triton X-100 (0.1%) + 3a
Triton X-100 (0.1%) + Mem
Triton X-100 (0.1%) + Mem + 3a
≥128
64
4
Triton X-100 (0.01%) + 3a
Triton X-100 (0.01%) + Mem
Triton X-100 (0.01%) + Mem + 3a
≥128
64
4
Conclusion
In summary, 1,2-benzisothiazol-3(2H)-one derivatives were identified as promising NDM-1
inhibitors that can restore the antimicrobial activity of Mem against E. coli Tg1 (NDM-1) and
clinical NDM-1 producing CRE isolates. Particularly, compound 3a displayed the most potent
synergistic activity with Mem and acceptable toxicity. SAR study and bioisosteric investigation
of the scaffold suggested that hydrophobic N-phenyl ring with halogen substituents would be
preferable and the synergistic activity was correlated with the bond length (N−X), which was
responsible for the bond dissociation energy. The most potent compound 3a, exhibiting IC50
2
4

value 0.44 ± 0.02 μM, inhibited NDM-1 via a non-competitive and covalent manner, providing
promising chemical scaffolds and opportunities for further development.
Experimental section
General procedure for the preparation of compounds 3a-3x
The starting material, 2,2’-dithiodibenzoyl chloride 2a, was synthesized from 2,2’-
5
1
dithiodibenzoic acid 1a as described previously. To a solution of the corresponding amine
(14.5 mmol) in dry DCM (25 mL) was added dropwise a solution of freshly prepared 2,2’-
dithiodibenzoyl chloride 2a (1.0 g, 2.9 mmol) in dry DCM (20 mL) at 0 °C with stirring. The
resulting mixture was allowed to slowly warm to room temperature and stirring was continued
for overnight. Removal of DCM under reduced pressure gave a reddish residue. An aqueous 3 M
HCl (20 mL) solution was then poured into the residue to remove the excess amine. After 15
mins, the precipitates were filtered, washed with DCM, and dried in vacuo to yield the desired
diamide intermediate as a solid with sufficient purity for the next step. To a well-stirred solution
of the corresponding diamide (1.0 mmol) in dry DCM (20 mL) was slowly added liquid bromine
(1.5 mmol in 10 mL DCM) at 0 °C over 10 mins. The reaction mixture was allowed to slowly
warm to room temperature and stirring was continued for overnight. Activated basic aluminum
oxide (0.30 g, Sigma-Aldrich, 150 mesh) was added to the mixture and the solvent was
evaporated under reduced pressure to give a crude mixture. Purification of the residue was
performed by using the flash column chromatography on silica gel to afford the desired
compound in moderate to good yield.
2
5

2
-(4-chlorophenyl)-1,2-benzisothiazol-3(2H)-one (3a)
This compound 3a (0.20 g, yield 75%) was prepared from 2,2'-disulfanediylbis(N-(4-
chlorophenyl)benzamide) (0.52 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
CDCl ) δ 8.10 (d, J = 7.82 Hz, 1H), 7.64 - 7.72 (m, 3H), 7.56 - 7.62 (m, 1H), 7.40 - 7.50 (m,
3
1
3
3
1
+
H), 1.24 - 1.33 (m, 1H), 0.90 (t, J = 6.85 Hz, 1H); C NMR (101 MHz, CDCl ) δ 164.1, 139.6,
3
35.9, 132.6, 132.4, 129.4, 127.2, 126.0, 125.5, 124.6, 120.2; HRMS m/z calcd for [C H NClOS
1
3
8
+
H] 262.0088, found 262.0086.
2
-(4-bromophenyl)-1,2-benzisothiazol-3(2H)-one (3b)
This compound (0.13 g, yield 43%) was prepared from 2,2'-disulfanediylbis(N-(4-
bromophenyl)benzamide) (0.61 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
1
3
CDCl ) δ 8.08 (d, J = 7.82 Hz, 1H), 7.50 - 7.75 (m, 6H), 7.35 - 7.50 (m, 1H); C NMR (101
3
MHz, CDCl ) δ 164.0, 139.6, 136.4, 132.6, 132.4, 127.2, 126.0, 125.7, 124.6, 120.2, 120.2;
3
+
HRMS m/z calcd for [C H BrNOS + H] 305.9583, found 305.9584.
1
3
8
2
-(4-(trifluoromethyl)phenyl)-1,2-benzisothiazol-3(2H)-one (3c)
This compound (0.12 g, yield 41%) was prepared from 2,2'-disulfanediylbis(N-(4-
(trifluoromethyl))benzamide) (0.59 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
2
6

CDCl ) δ 8.13 (d, J = 7.82 Hz, 1H), 7.94 (d, J = 7.82 Hz, 2H), 7.67 - 7.80 (m, 3H), 7.57 - 7.65
3
1
3
(
m, 1H), 7.48 (t, J = 7.34 Hz, 1H); C NMR (101 MHz, CDCl ) δ 164.2, 139.5, 132.9, 127.4,
3
+
1
2
26.6, 126.5, 126.5, 126.5, 126.1, 124.7, 123.6, 120.2; HRMS m/z calcd for [C H F NOS + H]
14 8 3
96.0351, found 296.0352.
2
-(3,4-dichlorophenyl)-1,2-benzisothiazol-3(2H)-one (3d)
This compound (0.11 g, yield 39%) was prepared from 2,2'-disulfanediylbis(N-(3,4-
dichlorophenyl)benzamide) (0.59 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
CDCl ) δ 8.11 (d, J = 7.82 Hz, 1H), 7.93 (d, J = 1.96 Hz, 1H), 7.67 - 7.75 (m, 1H), 7.58 - 7.66
3
1
3
(
m, 2H), 7.52 - 7.57 (m, 1H), 7.49 (t, J = 7.82 Hz, 1H); C NMR (101 MHz, CDCl ) δ 164.1,
3
1
39.5, 133.3, 132.9, 130.9, 127.4, 126.2, 125.8, 123.2, 120.2; HRMS m/z calcd for
+
[
C H Cl NOS + H] 295.9698, found 295.9699.
1
3
7
2
2
-(3-chlorophenyl)-1,2-benzisothiazol-3(2H)-one (3e)
This compound (0.14 g, yield 52%) was prepared from 2,2'-disulfanediylbis(N-(3-
chlorophenyl)benzamide) (0.52 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
CDCl ) δ 8.11 (d, J = 7.82 Hz, 1H), 7.82 (s, 1H), 7.57 - 7.74 (m, 3H), 7.36 - 7.51 (m, 2H), 7.29
3
1
3
(
d, J = 7.83 Hz, 1H); C NMR (101 MHz, CDCl ) δ 164.1, 139.6, 138.5, 135.0, 132.7, 130.3,
3
2
7

+
1
2
27.3, 127.0, 126.0, 124.6, 124.3, 122.2, 120.1; HRMS m/z calcd for [C H ClNOS + H]
13 8
62.0088, found 262.0094.
2
-(2-fluorophenyl)-1,2-benzisothiazol-3(2H)-one (3f)
This compound (0.09 g, yield 38%) was prepared from 2,2'-disulfanediylbis(N-(2-
fluorophenyl)benzamide) (0.49 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
CDCl ) δ 8.14 (d, J = 7.83 Hz, 1H), 7.64 - 7.76 (m, 1H), 7.57 - 7.64 (m, 1H), 7.53 (t, J = 7.34
3
1
3
Hz, 1H), 7.39 - 7.50 (m, 2H), 7.22 - 7.32 (m, 2H); C NMR (101 MHz, CDCl ) δ 164.7, 158.3
3
(d, J = 254.56 Hz), 141.3, 132.5, 130.6, 130.6 (d, J = 8.1 Hz), 130.0, 127.3, 125.8, 124.8 (d, J =
+
3
.7 Hz), 123.3, 120.3, 117.0 (d, J = 19.8 Hz); HRMS m/z calcd for [C H FNOS + H] 246.0383,
1
3
8
found 246.0385.
2
-(4-(hydroxymethyl)phenyl)-1,2-benzisothiazol-3(2H)-one (3g)
This compound (0.13 g, yield 49%) was prepared from 2,2'-disulfanediylbis(N-(4-
(hydroxymethyl)phenyl)benzamide) (0.52 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24
1
g, 1.5 mmol) according to the general preparation procedure described above. H NMR (400
MHz, MeOH-d ) δ 8.00 (d, J = 7.82 Hz, 1H), 7.84 (d, J = 7.83 Hz, 1H), 7.73 (t, J = 7.82 Hz,
4
1
3
1
1
6
H), 7.61 - 7.68 (m, 2H), 7.46 - 7.54 (m, 3H), 4.67 (s, 2H); C NMR (101 MHz, MeOH-d ) δ
4
64.7, 141.3, 140.7, 135.7, 132.5, 128.4, 127.5, 126.0, 125.8, 125.8, 124.8, 124.7, 124.2, 120.7,
+
3.1; HRMS m/z calcd for [C H NO S + H] 258.0583, found 258.0588.
1
4
11
2
2
8

2
-(4-chlorophenethyl)-1,2-benzisothiazol-3(2H)-one (3h)
This compound (0.15 g, yield 53%) was prepared from 2,2'-disulfanediylbis(N-(4-
chlorophenethyl)benzamide) (0.58 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5
1
mmol) according to the general preparation procedure described above. H NMR (400 MHz,
CDCl ) δ 8.00 (d, J = 8.80 Hz, 1H), 7.51 - 7.62 (m, 1H), 7.45 - 7.51 (m, 1H), 7.35 (t, J = 7.83
3
Hz, 1H), 7.23 (d, J = 7.82 Hz, 2H), 7.08 - 7.19 (m, 2H), 4.08 (t, J = 7.34 Hz, 2H), 3.01 (t, J =
1
3
7
.34 Hz, 2H); C NMR (101 MHz, CDCl ) δ 165.2, 140.2, 136.2, 132.6, 131.8, 130.2, 128.7,
3
+
1
26.6, 125.5, 124.4, 120.4, 45.0, 34.9; HRMS m/z calcd for [C H ClNOS + H] 290.0401,
1
5
12
found 290.0402.
2
-phenyl-1,2-benzisothiazol-3(2H)-one (3i)
This compound (0.16 g, yield 72%) was prepared from 2,2'-disulfanediylbis(N-
phenylbenzamide) (0.46 g, 1.0 mmol), DCM (20 mL) and liquid bromine (0.24 g, 1.5 mmol)
1
according to the general preparation procedure described above. H NMR (400 MHz, CDCl ) δ
3
8
7
1
2
.11 (d, J = 7.82 Hz, 1H), 7.72 (d, J = 7.58 Hz, 2H), 7.62 - 7.69 (m, 1H), 7.55 - 7.62 (m, 1H),
1
3
.41 - 7.53 (m, 3H), 7.29 - 7.37 (m, 1H); C NMR (101 MHz, CDCl ) δ 164.1, 139.9, 137.3,
3
+
32.4, 129.4, 127.2, 127.0, 125.8, 124.9, 124.6, 120.1; HRMS m/z calcd for [C H NOS + Na]
1
3
9
50.0297, found 250.0308.
2
-(4-methoxyphenyl)-1,2-benzisothiazol-3(2H)-one (3j)
2
9