Discovery of Novel Peptidomimetic Boronate ClpP Inhibitors with Noncanonical Enzyme Mechanism as Potent Virulence Blockers in Vitro and in Vivo

Article abstract of DOI:10.1021/acs.jmedchem.9b01746

Caseinolytic protease P (ClpP) is considered as a promising target for the treatment of Staphylococcus aureus infections. In an unbiased screen of 2632 molecules, a peptidomimetic boronate, MLN9708, was found to be a potent suppressor of SaClpP function. A time-saving and cost-efficient strategy integrating in silico position scanning, multistep miniaturized synthesis, and bioactivity testing was deployed for optimization of this hit compound and led to fast exploration of structure-activity relationships. Five of 150 compounds from the miniaturized synthesis exhibited improved inhibitory activity. Compound 43Hf was the most active inhibitor and showed reversible covalent binding to SaClpP while did not destabilize the tetradecameric structure of SaClpP. The crystal structure of 43Hf-SaClpP complex provided mechanistic insight into the covalent binding mode of peptidomimetic boronate and SaClpP. Furthermore, 43Hf could bind endogenous ClpP in S. aureus cells and exhibited significant efficacy in attenuating S. aureus virulence in vitro and in vivo.

Full text of DOI:10.1021/acs.jmedchem.9b01746

Subscriber access provided by UNIV OF CAMBRIDGE
Article
Discovery of novel peptidomimetic boronate ClpP
inhibitors with noncanonical enzyme mechanism
as potent virulence blockers in vitro and in vivo
Yuan Ju, Lihui He, Yuanzheng Zhou, Tao Yang, Ke Sun, Rao
Song, Yang Yang, Chengwei Li, Zitai Sang, rui bao, and Youfu Luo
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01746 • Publication Date (Web): 07 Feb 2020
Downloaded from pubs.acs.org on February 8, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of novel peptidomimetic boronate
ClpP inhibitors with noncanonical enzyme
mechanism as potent virulence blockers in
vitro and in vivo
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
#,a
#,a
#,a
#,a,b
a
a
Yuan Ju, Lihui He, Yuanzheng Zhou, Tao Yang
Ke Sun, Rao Song, Yang
a
a,c
d
,a
,a
Yang, Chengwei Li, Zitai Sang, Rui Bao* and Youfu Luo*
a
State Key Laboratory of Biotherapy and Cancer/Collaborative Innovation Center for
Biotherapy, West China Hospital, Sichuan University, Chengdu 610041, China
b
Laboratory of Human Disease and Immunotherapies, West China Hospital, Sichuan
University, Chengdu 610041, China
c
Suzhou Zelgen Biopharmaceuticals Co., Ltd., Kunshan, Jiangsu 215301, China
d
Institute of Life Science, Luoyang Normal University, Luoyang, Henan 471934,
China
*
Corresponding authors. Email: luo_youfu@scu.edu.cn, baorui@scu.edu.cn
#
These authors contributed equally to this work.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
Caseinolytic protease P (ClpP) is considered as a promising target for the treatment of
Staphylococcus aureus infections. In an unbiased screen of 2632 molecules, a
peptidomimetic boronate, MLN9708, was found to be a potent suppressor of SaClpP
function. A time-saving and cost-efficient strategy integrating in silico position
scanning, multistep miniaturized synthesis, and bioactivity testing was deployed for
optimization of this hit compound and led to fast exploration of structure-activity
relationships. Five of 150 compounds from the miniaturized synthesis exhibited
improved inhibitory activity. Compound 43Hf was the most active inhibitor and
showed reversible covalent binding to SaClpP while did not destabilize the
tetradecameric structure of SaClpP. The crystal structure of 43Hf-SaClpP complex
provided mechanistic insight into the covalent binding mode of peptidomimetic
boronate and SaClpP. Furthermore, 43Hf could bind endogenous ClpP in S. aureus
cells, and exhibited significant efficacy in attenuating S. aureus virulence in vitro and
in vivo.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 3 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
The caseinolytic protease proteolytic subunit (ClpP) is a highly conserved serine
protease that plays an essential role in bacterial cell homeostasis by degrading
damaged and short-lived regulatory proteins. Alteration of ClpP activity is associated
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
with diverse biological functions in many pathogenic bacteria. In Actinomycetales
such as Mycobacterium tuberculosis and Corynebacterium glutamicum, ClpP was
2
3
found to be essential for bacterial viability. In Staphylococcus aureus, Listeria
4
5
6
monocytogenes, Streptococcus pneumoniae, and Pseudomonas aeruginosa, even
though ClpP is not essential for bacterial survival, it is vital for their pathogenicity.
Thus, as widespread bacterial resistance to antibiotics making the treatment of
infectious diseases more challenging, chemically targeting ClpP has been recognized
as a promising anti-virulence strategy. Such a strategy would combat pathogenesis
with less evolutionary pressure towards drug resistance, and limit the detrimental
7
impact of antibiotics on the host microbiome.
S. aureus is a major human pathogen that causes a variety of infections including
8
scalded skin syndrome, sepsis and meningitis. S. aureus ClpP (SaClpP) was found to
have a profound influence on the secretion of α-hemolysin (Hla) by regulating factors
9
encoded by the accessory gene regulator (agr) locus, such as RNAIII transcript.
Genetic clpP knockout in S. aureus attenuated virulence in a murine skin abscess
3
model, and a similar phenotype was observed upon inhibition of SaClpP with
1
0
β-lactones, the first-generation ClpP inhibitor. β-Lactones can covalently acylate the
catalytic serine residue of ClpP; a derivative compound, D3, lead to obvious
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
0
0,11
elimination of the extracellular virulence of S. aureus at nanomolar concentration.
1
However, the β-lactones are unstable and quickly hydrolyzed in human plasma. In
addition, compound AV170, a phenol ester, can covalently modify the active site of
SaClpP and depolymerize the tetradecameric proteolytic enzyme complex into
inactive heptamers. However, AV170 is not suitable for further in vivo investigation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
due to the instability of phenol ester bond. The first non-covalent oxazole inhibitor
of SaClpP, AV145, can bind to the handle region close to the active site; however, the
1
3
chaperone ClpX revoked the inhibitory effect.
Novel inhibitors of
ClpP from other species are also under development. Boron-based peptidomimetics,
including bortezomib (BTZ) and substrate-like peptide boronate inhibitors, have been
shown to be potent covalent inhibitors of MtClpP1P2 from M. tuberculosis and
1
4-16
hClpXP from human.
In addition, pyrimidines have been shown to inhibit PfClpP
1
7
from Plasmodium falciparum. The chloromethyl ketone Z-LY-CMK and α-amino
diphenyl phosphonates have been identified as inhibitors of EcClpP from Escherichia
coli, but some of these inhibitors lacked the efficacy in the nitric oxide stress
1
8,19
assay.
Nevertheless, the chemical structure types of SaClpP inhibitors are rather
quite limited. To expand the arsenal of ClpP inhibitors and further improve their
pharmacological application, there is a pressing need to develop novel compounds
with stable scaffolds and promising anti-virulence activity both in vitro and in vivo.
The traditional design–synthesis–purification–assay workflow is still widely used
in drug candidate discovery. However, the hazardous and time-consuming process,
which takes 1–8 weeks for every cycle and involves relatively high consumption of
ACS Paragon Plus Environment

Page 5 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
0
chemical materials, limits the use of these techniques on a larger scale.
For
bioassays that can couple with miniaturized chemical synthesis approaches, including
2
1
22
one-bead one-compound libraries,
DNA-encoding chemical libraries,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
mass-encoding technology, dynamic combinatorial chemistry, and nanoscale
24
synthesis with affinity ranking, the integration of synthesis and assay could greatly
reduce the material requirements and enable rapid analysis. Most such approaches
depend on delicate purification during the process of library synthesis, and integration
of miniaturized synthesis and testing without complicated workup has been
predominantly applied in one-step reactions. It remains challenging to use this
strategy for compounds requiring multistep synthesis.
In this study, considering that in silico structure-based activity predication can
2
5
significantly accelerate the hit to lead process, we used an integrated strategy
cascading in silico position scanning, multistep synthesis, and peptidase inhibition
activity testing to improve the inhibitory effect against SaClpP of our initial hit
compound MLN9708 (randomly screened from 2632 compounds). The integrated
strategy rationally and rapidly generated syntheses with high diversity at each position.
Notably, each reaction was performed at 0.1 M concentration in separated microplate
wells with minimum material consumption, the direct bioassay of crude products also
avoids the lengthy workup steps. Next, we performed scaled-up synthesis of the
promising products and tested their effect on virulence of S. aureus both in vitro and
in vivo. The most active compound, 43Hf, was co-crystallized with SaClpP, and the
detailed binding mode was elucidated. Compound 43Hf showed reversible covalent
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
binding with the active site serine residue of SaClpP while the protein maintained its
tetradecameric assembly.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Results
Identification of the peptidomimetic boronate MLN 9708 as an SaClpP inhibitor
To identify hit compounds targeting SaClpP, we performed a random screen using an
in-house library of 2632 bioactive chemicals (Fig. 1a). In the primary screen,
compounds were assayed at a single-point concentration of 10 µM and eight initial
hits (hit rate 0.3 %) were selected based on their >50% inhibitory effects on SaClpP
cleavage activity towards substrate Suc-LY-AMC (Figs. 1a, 2a and 2b). In the
secondary screen, initial hits were evaluated in a concentration-dependent manner and
five hits with low IC50 values were selected for further studies (Figs. 1a and 2c).
Differential scanning calorimetry analysis showed that, among the five hits,
compound F significantly shifted melting peak of SaClpP (from 50.8 °C to 77.5 °C)
and the EC50 value was 19.90 ± 1.17 µM, indicating direct binding (Fig. 2d).
Moreover, compound F also stood out in the substrate competition assay (Fig. 2e), in
which the increased substrate concentration had minimal influence on its inhibitory
activity.
ACS Paragon Plus Environment

Page 7 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Schematic illustration of chemical screening and structural optimization
process. (a) Identification of novel SaClpP inhibitor hits in an in-house library
containing 2632 bioactive small-molecule compounds. (b) Cartoon depicting the
integrated strategy to the optimize hit compound MLN9708.
The hit compound F is ixazomib citrate (MLN9708), which can be immediately
2
6
hydrolyzed to its biologically active form ixazomib (MLN2238) (Fig. 1a). The
dissociation constant (K ) for MLN9708 binding to SaClpP is about 7.3 ± 1.1 µM
d
according to the isothermal titration calorimetry (ITC) assay (Fig. S1). The active
form MLN2238 has a similar structure to that of the MtClpP1 inhibitor BTZ, which
also showed efficient inhibition of SaClpP (IC50 value of 5.3 ± 0.2 µM) (Fig. 2b;
1
4
Table 1), indicating that they may share a common mechanism of ClpP inhibition.
Boron-based peptidomimetic molecules have been recognized to be potent inhibitors
27
of proteases; because the boronate warhead is a strong electrophile and behaves as a
Lewis acid, it is capable of forming dative bonds with nucleophilic residues like
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
8,29
serine or threonine, resulting in an anionic tetragonal structure.
In SaClpP, the
active-site residue Ser98 is the most probable site for boronate binding, while
the inhibitory potency can be improved by optimizing the other moieties of the
boronate inhibitor.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Discovery of a novel SaClpP inhibitor MLN9708 from the chemical library
screen. (a) A random screen of 2362 compounds identified eight primary hits
with >50% inhibitory effect at 10 µM. Each circle represents the mean percentage
inhibition for a compound tested at 10 µM in triplicate. (b) Structures of the eight
initial hits and bortezomib (BTZ). (c) IC50 values of compounds A–H toward SaClpP.
(
d) Thermal shift assays of the binding between five hits (at 100 µM) and SaClpP. (e)
Inhibition of SaClpP peptidase activity by 3 µM compounds B, D, F, G and H in the
presence of 200 µM or 1 mM Suc-LY-AMC substrate. The error bars of panels (c)
and (e) represent the SD in three independent replicate determinations.
ACS Paragon Plus Environment

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Optimization of MLN9708/BTZ through the integrated strategy
The integrated strategy—including computational ranking, multistep microscale
synthesis, and cascaded biological testing—is illustrated in Figure 1b. The in silico
step involved virtual focused library construction and position scanning. A virtual
focused library containing the boronate warhead was automatically constructed by
substituent random combinations. To conduct the in silico position scanning, 17
libraries (libraries A–Q), which had different amino acids for II moiety (Figs. 3a and
S2), were generated. Each library contained 258 individual chemicals, including 43
carboxylic acids as building blocks for moiety I, and 6 boronate building blocks for
moiety III (Fig. S2). Docking of the ligands into the active site of SaClpP was
performed with the Schrödinger suite and the best pose of each molecule was ranked
according to the GlideScore. A heat map of the docking scores showed significant
differentiation at each position (Fig. 3a). Molecules with a score lower than that of
MLN2238 (−6.24 kcal/mol) were selected and the frequency of occurrence of each
group was counted: only seven fragments occurred <40 times in the I position, while
nine II fragments and all the III fragments occurred >100 times (Fig. S3). Eventually,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0 commercially available building blocks for the I substituent, 5 for the II substituent,
and 1 for the III substituent were selected for the next multistep miniaturized
synthesis experiments (Figs. 4 and S2).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure. 3 Ranking of the virtual focused library and activity of product mixtures
derived from the micro-scale synthesis. (a) Heat map of the GlideScores of small
molecules in virtual libraries A–Q. (b) Radial graph showing inhibitory effects of
crude products obtained in micro-scale synthesis. The outer edge of the wheel
represents 100% inhibition of SaClpP peptidase activity, while the center of the circle
represents 0% inhibition. The numbers around the ring represent the group on moiety
I, while the colors of the symbols represent the groups on moiety II. Each point
represents the mean percentage inhibition for the mixture tested at Σ[Pn]=10 µM in
triplicates.
To enable rapid and cost-efficient preparation of an extensive series of boronate
warhead compounds, we used a micro-scale and multi-step synthetic system that
could perform up to 96 reactions in parallel. In a proof of concept experiment, BTZ
was used as the synthesis template to verify the efficiency of the miniaturized
synthesis route (Fig. S4a). The multistep synthesis procedures were developed in
microplates, followed by straightforward solvent extraction and evaporation. In this
experiment, the ester product S1 (Fig. S4a) was obtained by coupling of
ACS Paragon Plus Environment

Page 11 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2-pyrazinecarboxylic acid and methyl L-phenylalaninate, then, the ester was
hydrolyzed in basic conditions to furnish carboxylic acids product S2 (Fig. S4a).
Product S3 (N-((S)-1-(((R)-3-methyl-1-((3aS,4S,6S,7aR)-3a,5,5-trimethylhexahydro
-4,6-methanobenzo[d][1,3,2]dioxaborol-2-yl)butyl)amino)-1-oxo-3-phenylpropan-2-
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-
yl)pyrazine-2- carboxamide) was achieved by coupling (R)-BoroLeu-(+)-Pinanediol
trifluoroacetate with S2, and was subsequently deprotected to provide the target
boronate compound S4 (Fig. S4a). High-performance liquid chromatography (HPLC)
analysis indicated that BTZ was present in the crude product and the final yield was
about 28% (Fig. S4b). To be noted, none of the intermediate products demonstrated
inhibitory activity towards SaClpP (Fig. S5). Taken together, we concluded that it is
feasible to synthesize boronate warhead inhibitors in a micro-scale and multistep
system, and the resultant library members can be assayed directly without the need for
delicate purification procedures.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
I
II R2
III
O
R2
OH
B
O
NH2
O
R2
H
H
O
O
N
H2N
OH
OH
R1
OH
R3
B
R1
N
H
R1
N
O
H
O
O
R3
O
I
F
O
O
O
O
OH
O
O
O
O
OH
OH
O
OH
OH
OH
O2N
N
OH
Br
OH
OH
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
O
OH
O
O2N
1
O
2O
4
5
6
10
11
16
O
O
O
O
O
OH
N
H
N
O
OH HN
OH
OH
OH
OH
Cl
O
O
O
O
F
1
7
18
19
20
21
22
23
24
34
O
O
O
O
O
O
O
O
O2N
OH
OH
OH
OH
OH
OH
OH
OH
OH
N
O
Cl
N
Cl
O
N
Cl
NO2
2
5
26
27
28
29
32
33
O
Cl
OH
O
O
O
O
O
O
OH
N
N
S
OH
OH Br
OH
OH
O
N
Cl
3
5
36
37
40
42
43
III
II
OH
NH
NH2
B
H
NH
H
N
O
O
NH
2
O
O
O
O
NH
2
H N
O
O
H
2
N
NH2
2
O
O
O
O
H
f
A
Q
B
E
Figure 4. Synthetic strategy and building blocks for moieties I, II and III for the
combinatorial synthesis of a 150-member library of the boronate peptidomimetics.
To efficiently and rapidly explore the preliminary structure–activity relationship
(
SAR) of the boronate peptidomimetics toward SaClpP, we performed the micro-scale
combinatorial synthesis and bioassay in 96-well plates. In parallel, MLN2238 (43Af)
and BTZ (42Qf) were synthesized as controls. The multistep and micro-scale
synthesis in individual wells allowed us to trace the products structure clearly (Figs.
1
b and 4). In the first step, combinations of building blocks for moieties I and II were
employed to give 150 (30 × 5) intermediates. Next, by coupling the intermediates
with (R)-aminoboronate ester and deprotection, we totally obtained 150 target
ACS Paragon Plus Environment

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
boronate products (Fig. 4). Biological testing of the crude products as inhibitors of
SaClpP was performed at a single-point concentration of 10 µM. Certain compounds
that were better inhibitors than MLN2238 (43Af) or BTZ (42Qf) were selected to
further determine their IC50 values (Fig. 3b; Table S1). Note that the building blocks
were all inactive towards SaClpP (Fig. S6). Five products with low IC50 values were
qualitatively analyzed by ultra-performance liquid chromatography–mass
spectrometry (UPLC–MS) to confirm the formation of the correct chemicals. The
contents of the expected molecules, which proved to be the main products in the crude
products by UPLC–MS, was subsequently measured by HPLC and determined by
area normalization method (Fig. S7). Then, scaled-up synthesis of these five
inhibitors was performed in glassware vessels and they were purified by column
chromatography (Scheme 1). The dose-dependent inhibition of SaClpP was
determined (Table 1). The results validated the improved inhibitory activity of the
products optimized by the miniaturized system (IC50 values ranging from 0.9 µM to
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
.9 µM) compared with 5.7 µM for MLN9708 and 5.3 µM for BTZ.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 1. Synthesis of compounds 5Bf, 5Hf, 43Hf, 11Qf and 24Qf.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Reagents and conditions: (i) HATU, TEA, DCM, r.t., overnight; (ii) LiOH,
2
H O/THF/MeOH, 0 °C, overnight; (iii) isobutaneboronic acid, aqueous HCl,
EA/MeOH, r.t., overnight, yields: 15.1%–23.1%.
Thus, some preliminary SAR can be inferred from the micro-scale combinatorial
screening. For discussion purposes, we divided these 150 crude products into five
series (series A, B, E, H and Q) according to the different groups at the II position.
Compounds of series A and E with hydrogen or 1H-imidazole-4-methyl group at the
II position exhibited weakened inhibition potency compared with MLN2238 or BTZ;
none of these 60 compounds (30 in series A and 30 in series E) exhibited better
inhibitory effects than MLN2238 at 10 µM (Fig. 3b). These data suggest that Gly and
His may not be preferential amino acids as the building blocks for moiety II.
Compounds of series B, H and Q yielded several SaClpP inhibitors that were more
potent than MLN2238 (8 compounds in series B, 10 compounds in series H and 12
compounds in series Q). Among them, five compounds from series Q showed >70%
inhibition at 10 µM (Fig. 3b), suggesting that groups with phenyl substituent may
ACS Paragon Plus Environment

Page 15 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
perform better at position II. Considering moiety I, compounds with
1
-(4-isobutylphenyl)ethan-1-yl or m-methylphenyl substituent from series B,
compounds with 1-(4-isobutylphenyl)ethan-1-yl or 4-fluorophenyl substituent of
series H, compounds with phenethyl, 2-methyl-4-methoxyphenyl, indole-3-methyl,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
,5-dimethoxyphenyl or 4-methoxyphenyl substituents from series Q showed >70%
inhibition at 10 µM (Fig. 3b), suggesting that aromatic rings at position I may
contribute to SaClpP inhibition. IC50 values of compounds that showed greater
inhibition than the hit controls were determined (Table S1). Compounds with
2,5-dichlorophenyl, 1-(4-isobutylphenyl)ethan-1-yl, tert-butylphenyl, m-methylphenyl,
3
,5-dinitrophenyl, 2,4-dimetoxyphenyl, 2,5-dimetoxyphenyl, p-methoxyphenyl,
o-chlorophenyl, pyridine-2-yl, 2-chloropyridine-4-yl, 2-methyl-4-metoxyphenyl or
indole-3-methyl groups at position I exhibited greater inhibition potency (IC50 < 10
µM) against SaClpP than those with other substituents at this position. Taken together,
we conclude that aromatic groups with phenyl group at position II and aromatic rings
with electron-donating groups at position I were more beneficial for inhibition of
SaClpP.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Peptidase inhibitory activity, anti-virulence activity, hemolytic activity and
cytotoxicity of the purified inhibitors.
e
a
b
c
d
GI50 (µM)
IC50
k
inact/K
2
i
EC50
HC50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Entry
(
µM)
(10 /sM)
(µM)
(µM)
HeLa
HepG2
>200
MLN
BTZ
5.7 ± 0.01 2.4 ± 0.07
3.6 ± 0.2
408 181.8 ± 2.1
5.3 ± 0.2
1.4 ± 0.1
1.9 ± 0.3
1.5 ± 0.2
1.9 ± 0.2
2.2 ±0.1
7.2 ± 0.3
8.0 ± 1.0
9.1 ± 0.6
6.7 ± 0.7
23.5 ± 0.6 >500 168.1 ± 1.7 100.6 ± 2.5
5
Bf
9.6 ± 0.5
2.6 ± 0.4
>500 83.0 ± 2.1 116.8 ± 3.4
5
Hf
>500 71.2 ± 2.2
54.8 ± 3.3
>200
11Qf
14.8 ± 0.9 >500
>200
2
4Qf
3.3 ± 0.2
>500 195.8 ± 1.9
>200
43Hf
0.9 ± 0.05 16.7 ± 0.2 1.4 ± 0.02 >500 108.0 ± 3.1 83.4 ± 4.6
a
b
Dose-dependent inhibition activity against SaClpP of the purified product. Values determined
c
from time-dependent IC50 curves. Determination of EC50 values for hemolysis after treatment of
d
e
MRSA with the inhibitors. Hemolytic activity of inhibitors. Values that inhibited the growth of
the cells by 50%. The data represent the mean ± SD in three independent replicate determinations.
Biochemical and biophysical characterization of the interaction between SaClpP
and the optimized compounds
Because of the time-dependent inhibition properties of covalent inhibitors, it is
ACS Paragon Plus Environment

Page 17 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
i
suitable to further evaluate their activities by comparing the inhibition constant (K )
3
0
and enzyme inactivation rate (kinact).
i
K and kinact values were calculated directly
3
1
from the time-dependent IC50 values as previously described. Compared with the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
k
inact/K
i
values of MLN9708 (2.4 × 10 /sM) and BTZ (2.2 × 10 /sM), all of the
optimized compounds showed approximately 3-fold to 7-fold higher kinact/K values,
i
2
2
ranging from 6.7 × 10 /sM to 16.7 × 10 /sM (Table 1). This result confirmed that the
optimization strategy had successfully improved the potency of the primary hits.
Compound 43Hf was verified as the most potent inhibitor.
To examine the direct interaction between the optimized inhibitors and SaClpP,
thermal analysis was conducted by using differential scanning fluorimetry. All the
optimized compounds exhibited enhanced abilities to progressively shift the transition
midpoint of the heat denaturation curve by 29.5–42.2 °C comparing to the DMSO
m
treated SaClpP (45.5 °C) (Fig. 5a). 43Hf exerted the most significant effect (ΔT =
+42.2 °C), with a half-maximal effective concentration of 1.27 µM (Figs. 5a and 5b),
which confirmed that the affinity of 43Hf for SaClpP was on the micromolar level. It
worth noting that the reported covalent inhibitor AV170 barely impacted the thermal
stability of SaClpP (Fig. 5a), indicating that the boronate peptidomimetics have a
different inhibitory mechanism and distinct interaction pattern from that of AV170.
This deduction was further supported by the size exclusion chromatography column
and native-PAGE assays, in which the oligomeric states of SaClpP with or without
boronate inhibitors were determined (Figs. 5c, 5d and S8). Distinct from the
mechanism of compound AV170 which disrupted the ClpP tetradecamer into inactive
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
2
heptamers, the boronates warhead inhibitors did not alter the tetradecameric
assembly of SaClpP.
Next, we performed a dialysis experiment and measured the recovery rate of
peptidase activity to test the reversibility of binding. The minimal changes in SaClpP
inhibition by AV170 even after 72-h dialysis confirmed the permanent binding of this
inhibitor to the protease active site (Fig. 5e). In contrast, for the boronate inhibitors,
SaClpP activities were restored to nearly 100% after different periods of dialysis (Fig.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
e). The rates of recovery of activity were slower for the optimized compounds than
for MLN9708 and BTZ. In particular, compounds with an indolyl group at the II
position (5Hf and 43Hf) showed the slowest recovery rates. These results suggest that
the boronate inhibitors bound to SaClpP in a reversible, covalent fashion, and the
indolyl group at position II might strengthen the binding. The reversible covalent
inhibitory nature of the boronate compounds toward SaClpP would prevent the
formation of permanent covalent adducts with off-target proteins, and it may be
possible to optimize the inhibitor residence time at the active-site serine of SaClpP,
3
3,34
which could facilitate the development of inhibitors with therapeutic applications.
In S. aureus cells, SaClpP is in complex with SaClpX or SaClpC to execute the
1
ATPase-dependent protein degradation function. The SaClpXP activity can be
determined in vitro by monitoring the degradation of fluorescein
35
isothiocyanate-labeled casein (FITC-casein) substrate. Using this approach, the
optimized inhibitors were assayed for their inhibition of SaClpXP activity at
concentrations of 100, 10 and 1 µM. In contrast to the initial hits BTZ and MLN9708,
ACS Paragon Plus Environment

Page 19 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
which showed nearly no inhibition of SaClpXP, two of the optimized compounds
were effective inhibitors; in particular, 43Hf had an IC50 value of 69.62 ± 2.94 µM
(Fig. 5f). These results indicated that 43Hf was also an efficient inhibitor of the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
SaClpXP complex.
Considering that the initial hit ixazomib is a marketed proteasome inhibitor and
SaClpP is a serine protease, the target compounds were tested for their inhibitory
3
6
16
effects on 20S proteasome, homologous PaClpP1 from P. aeruginosa, hClpP and
3
7
other serine proteases, including HtrA-like serine protease AlgW and trypsin-like
3
8
serine protease Factor Xa (FXa). The tested compounds MLN9708, 5Bf, 43Hf and
2
4Qf inhibited 20S proteasome with nanomolar IC50 values, and compound 43Hf
showed slightly lower activity than the others (Figure S9 and Table S2). All the tested
analogues had slight selectivity for SaClpP against hClpP and PaClpP1, with IC50
between 3 to 25 µM for hClpP and PaClpP1. Nevertheless, these boronate compounds
showed little inhibition to AlgW and FXa at concentration up to 100 µM.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Inhibition potency and binding affinity of the optimized compounds. (a)
Box plots representation of the direct interaction between the boronate compounds
(50 µM) and SaClpP using differential scanning fluorimetry determination. Data were
from three independent experiments, and the black lines represent the mean. (b)
Strong stabilization of SaClpP folding by 43Hf in a concentration-dependent manner.
The data represent the mean ± SD in three independent replicate determinations. (c,
d) Size-exclusion chromatograms and native-PAGE illustrating the retention of the
tetradecameric state of SaClpP on treatment with the boronate compounds. (e)
Peptidase activity recovery after dialysis of SaClpP pretreated with the inhibitor or
DMSO. (f) Inhibition of SaClpXP protease activity by the inhibitors. The error bars of
panels (e) and (f) represent the SD in three independent replicate determinations.
ACS Paragon Plus Environment

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structural insights into the boronate inhibitor-bound SaClpP
To determine the exact binding mode between the boronate compounds and SaClpP,
BTZ and 43Hf were respectively co-crystallized with SaClpP. The protein was
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
prepared using similar expression and purification methods to those described
12,39,40
previously.
The crystal structures were determined by molecular replacement
3
9
using the SaClpP Y63A mutant structure (PDB code: 5C90) as the search model .
The final structure of 43Hf-bound SaClpP was refined to 2.3Å resolution with Rwork
of 19% and Rfree of 20%, and that of BTZ-bound SaClpP was refined to 2.2 Å
resolution with Rwork of 24% and Rfree of 25% (Table S3). Like other reported ClpP
13,16,39,41,42
structures,
SaClpP assembled into two heptameric rings that form a
tetradecameric barrel. The inhibitors covalently bound to the active-site residue Ser98
in each subunit (Fig. 6). 43Hf-bound SaClpP and BTZ-bound SaClpP were nearly
identical in their overall structures, displaying a root mean square deviation (RMSD)
value of 0.39Å for 1400 Cα atoms in tetradecamer–tetradecamer comparisons.
Despite variations in occupancy, model building into the averaged 2Fo−Fc
electron-density map allowed the unambiguous positioning of 43Hf (or BTZ)
between helices α5 and α3 (Fig. 6b and 6c). Since the two boronates had similar
binding patterns, we mainly focused on the 43Hf-bound SaClpP structure in the
following analysis. 43Hf bound to the S1–S3 subsites of SaClpP and formed several
hydrogen bonds with conserved residues, including Gly69, Ser70, Met99, His123 and
Leu126 (Figs. 6b and 6d). The difference of hydrogen bond with Lue126 for 43Hf
comparing to BTZ interpreted the enhanced reversible covalent binding (Fig. 6d and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6e). A network of hydrophobic interaction contacts with Val71, Pro125, Gly127 and
Ile143 further stabilized the compound binding. The boron atom of 43Hf covalently
interacted with the nucleophilic oxygen lone pair of Ser98O, while Gly69N and
Met99N hydrogen-bridges to one of the acidic boronate hydroxyl groups. These
boron atom-mediated interactions strengthened the anchoring in the active site
oxyanion hole.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Co-crystal structure of SaClpP in complex with covalent inhibitor 43Hf or
BTZ. (a) Overall structures of 43Hf-bound SaClpP. (b, c) Structural details of (b)
4
3Hf and (c) BTZ binding pocket in SaClpP. The monomer SaClpP is shown as a blue
cartoon, the inhibitors are shown as sticks. 2Fo−Fc electron-density maps of 43Hf and
BTZ in SaClpP are also shown. (d, e) Binding mode of the 43Hf (d) and BTZ (e)
interactions with critical residues of SaClpP. Residues that form hydrogen bonds with
the ligand are shown as sticks and are labeled. (f) Comparison of the position of
ACS Paragon Plus Environment

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
His123 in tetradecameric wild-type SaClpP (PDB: 3V5E) [in pink], in heptameric
SaClpP (PDB: 3ST9) [in light yellow], and in SaClpP with 43Hf bound [in blue].
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Like the β-lactone D3 and phenyl ester-containing compound ML90, which also
covalently bound to the ClpP active site and resulted in retention of the tetradecameric
1
2,32,43
assembly,
the boronate compounds did not lead to SaClpP disassembly.
However, the insight into the mechanism of compounds triggering or disrupting the
oligomerization remains to be elucidated. As previously proposed, the conserved
active site histidine is critical in a hydrogen bonding network that is important for
maintaining the oligomerization state, and alerting the orientation of its side chain
4
4
causes deoligomerization. In the boronate-bound SaClpP structures, the hydroxyl of
boronic acid formed a hydrogen bond with the nitrogen atom of the His123 imidazole
ring and kept this residue in place (Fig. 6f).
Attenuation of S. aureus virulence by optimized compounds in vitro and in vivo
To identify a pharmacological candidate, we examined the activity of the optimized
compounds in cell-based studies and cytotoxicity tests. To determine the inhibitory
effects of boronate compounds on S. aureus growth, we measured the bacterial
growth rate in the presence of different inhibitors. Compounds 43Hf and 5Hf
demonstrated only minor inhibitory activity of bacterial growth of
methicillin-resistant S. aureus (MRSA) ATCC33591 at concentrations up to 64 µM
(Figs. 7a and S9). In cytotoxicity assays, 5Hf showed moderate inhibition of HepG2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cells with a GI50 value of 54.81 µM, but the other boronates displayed lower
cytotoxicity toward HeLa and HepG2 cell lines (Table 1; Fig. S11).
Since ClpP controls the expression of bacterial Hla, which can cause cystitis,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
5
pyelonephritis, and sepsis , we treated MRSA with various concentrations of the
boronate inhibitors and incubated the culture supernatants with erythrocytes to assess
the level of secreted Hla. The boronate inhibitors themselves (other than MLN9708)
showed no hemolytic activity even at 500 µM (Table 1). They all resulted in an
obvious reduction in hemolysis caused by pretreated MRSA supernatant, with IC50
values ranging from 1.4 to 23.5 µM (Fig. 7b; Table 1). 43Hf was the best compound,
approximately 30-fold more potent than BTZ. Western-blotting assays confirmed its
downregulatory effect on the levels of Hla in the bacteria (Fig. S12). 43Hf could
almost eliminate Hla at 16 µM after the supernatant was concentrated 25-fold (Fig.
7
c). To confirm the requirement of functional ClpP for 43Hf inhibiting the Hla, we
next investigated
a
ClpP-deletion (ΔclpP) strain and
a
complemented
+
ΔclpP/pYJ335::clpP (ΔclpP ) strain of S. aureus NCTC8325. We found that the
hemolytic activity of the ΔclpP strain was greatly reduced compared with that of the
wild-type (WT), and was not affected by the presence of 43Hf (Fig. 7d). The WT and
+
the complemented (ΔclpP ) strains showed nearly identical hemolytic activity, and
43Hf could suppress the production of Hla by these strains in
a
concentration-dependent manner (Fig. 7d). These results indicated that functional
ClpP was a requirement for attenuation of S. aureus virulence by 43Hf. Thus, we
conclude that 43Hf is a potent antivirulence candidate with a mild antibacterial effect
ACS Paragon Plus Environment

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(minimum inhibitory concentration = 64–128 µg/mL) and low cytotoxicity.
To gain more detailed insight into the virulence inhibition activity of 43Hf, we
next determined the expression of several important virulence-related genes in S.
aureus after 43Hf treatment. According to quantitative polymerase chain reaction
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(qPCR) results (Fig. 7e), RNAIII effector, which controls a large number of virulence
factors including Hla, was significantly repressed (>10-fold) by 16 µM 43Hf.
Expression of agrB, the agr quorum-sensing regulator, was reduced around 2-fold.
Compound 43Hf also demonstrated significant repression effects on expression of
4
5,46
genes encoding toxins (hla, psm, lukS) and protease (aur).
Cellular thermal shift assay (CETSA) was performed to confirm that 43Hf
engaged SaClpP in bacteria. CETSA monitors ligand-induced changes in thermal
stability of a target protein to determine the binding affinity of the inhibitor for the
4
7
target protein in whole cells. Figure 7f showed that the SaClpP protein was unstable
at 59 °C in MRSA treated with DMSO. The thermal stability of the endogenous
SaClpP was enhanced by 100 µM 43Hf; the protein remained stable even at 83 °C
(Fig. 7f). These results provided direct evidence that 43Hf is able to target SaClpP in
bacteria.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Compound 43Hf affected the virulence of S. aureus virulence in vitro. (a)
Growth curves of MRSA treated with various concentrations of 43Hf in MHB
medium. (b) The hemolytic activity of MRSA was reduced by 43Hf. The images
represent the released hemoglobin of in supernatant from mixtures of culture
supernatant and sheep red blood cells after incubation for 1 h. (c) Western blot
showing the production of Hla (25-fold concentrated) in 43Hf-treated MRSA. (d)
Comparison of the hemolytic activity of S. aureus WT, ΔclpP and complemented
+
ΔclpP treated with 43Hf at different concentrations. (e) Expressions level of the
virulence genes (hla, psm, luks, RNAIII, agrB and aur) in MRSA in response to 43Hf
treatment. (f) CETSA of SaClpP in MRSA in the absence or presence of 43Hf. The
MRSA culture was incubated with 100 µM 43Hf for 2 h before performing CETSA.
The error bars of panels (a), (d) and (e) represent SD in three independent replicate
ACS Paragon Plus Environment

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
determinations.
In vivo efficacy of compound 43Hf was evaluated using a murine skin abscess mice
model generated by infection with MRSA ATCC33591. Different bacterial recovery
levels from skin were observed between the control groups and the treated mice
(p=0.024 for vehicle and 1.5 mg/kg 43Hf, and p=0.013 for vehicle and 3 mg/kg 43Hf)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Fig. 8a). The body weights of the mice were monitored daily, and 43Hf caused no
evident weight loss or other signs of adverse effects in the efficacy experiment (Fig.
8b). Serum biochemistry analyses were also performed to assess the potential toxicity.
As shown in Figure S13, the kidney indicators creatinine (CREA), uric acid (UA) and
blood urea nitrogen (BUN), and the heart indicator lactate dehydrogenase (LDH)
maintained at normal levels. 43Hf treatment increased the level of direct bilirubin
(DBIL) and aspartate transaminase (AST) on day 4 but these returned to the normal
level on day 14. All other hepatic indicators (albumin (ALB), total protein (TP),
alkaline phosphatase (ALP) and alanine transaminase (ALT)) remained in the normal
range. These findings demonstrated that 43Hf may affect the hepatic function in the
short term after administration, but without long-term impact.
Skin tissue and major organs (heart, liver, spleen, lung and kidney) were fixed in 4%
paraformaldehyde, embedded, thin-sectioned and stained with hematoxylin and eosin
(H&E). Pathological damages to skin tissue such as thickened dermis and
inflammatory cell infiltration of the dermis was observed in the MRSA-infected group
compared with the control (PBS treated groups) (Fig. 8c). Skin integrity
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was preserved in mice treated with 43Hf after the infection, and vancomycin
(Van)-treated infected skin showed minimal inflammation (Fig. 8c). The MRSA
infection caused severe alveolar destruction and interstitium thickening in the lung
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
tissue of the vehicle group as S. aureus usually spread to the lung through the blood
48
from skin, while treatment with 43Hf and Van obviously reduced the changes (Fig.
S14). H&E stained sections of heart, liver, kidney and spleen revealed no
histopathological changes (Fig. S14). All of these results demonstrated that 43Hf was
efficacious in attenuating the virulence of MRSA in vivo.
Figure 8. In vivo anti-virulence activity of compound 43Hf in a mouse model of
MRSA skin infection. (a) Bacterial counts in skin samples (n=5 mice/ group). The p
value vs the vehicle group determined with t-test. (b) Animal body weight change.
Data are represented as the mean ± SD (n=5 mice/ group). (c) H&E stained of
infected skin on day 4 after the first administration of treatment were analyzed by
microscopy. Scale bar 200 µm.
ACS Paragon Plus Environment

Page 29 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discussion and Conclusions
The traditional small-molecule drug discovery process is still costly and
time-consuming, involving iterative rounds of molecular design, chemical synthesis,
biological assay, and data analysis. The relatively large consumption of chemical
materials, hazardous workup, low signal-to-noise outputs, and cumbersome
equipment required for recent techniques make new drug development increasingly
difficult. In this study, we describe a low-cost and simple combinatorial approach that
consolidated computational ranking, multistep miniaturized synthesis in microplates,
and cascaded bioactivity testing for rapid exploration of preliminary SARs of
inhibitor compounds of SaClpP. The coupled miniaturized synthesis (<0.1 mg of
reactants for each reaction) enabled parallel synthesis of large collections of
compounds with minimal consumption. The cascaded SaClpP inhibition assay
directly screened out active compounds from the crude products without
time-intensive and laborious reaction purification.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Nevertheless, we cannot avoid false negative hits in this integration platform of
synthesis and testing because the reaction yield depends on the starting materials and
intermediates to some extent. Importantly, our combinatorial and focused compound
library exploration and enzymatic screening identified five compounds suitable for
scaled-up synthesis. Eventually, from these five active SaClpP inhibitors, compound
4
3Hf was proved to be a promising submicromolar inhibitor of SaClpP that is capable
of targeting endogenous ClpP in bacterial cells. Several specific ClpP inhibitors
have been developed to decrease the virulence of bacteria and be used in combination
ACS Paragon Plus Environment