Discovery, Optimization, and Target Identification of Novel Potent Broad-Spectrum Antiviral Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.9b00091

Viral infections are increasing and probably long-lasting global risks. In this study, a chemical library was exploited by phenotypic screening to discover new antiviral inhibitors. After optimizations from hit to lead, a novel potent small molecule (RYL-634) was identified, showing excellent broad-spectrum inhibition activity against various pathogenic viruses, including hepatitis C virus, dengue virus, Zika virus, chikungunya virus, enterovirus 71, human immunodeficiency virus, respiratory syncytial virus, and others. The mechanism of action and potential targets of RYL-634 were further explored by the combination of activity-based protein profiling and other techniques. Finally, human dihydroorotate dehydrogenase was validated as the major target of RYL-634. We did not observe any mutant resistance under our pressure selections with RYL-634, and it had a strong synergistic effect with some Food and Drug Administration-approved drugs. Hence, there is great potential for developing new broad-spectrum antivirals based on RYL-634.

Full text of DOI:10.1021/acs.jmedchem.9b00091

Subscriber access provided by IDAHO STATE UNIV
Article
Discovery, Optimization and Target Identification of
Novel Potent Broad-Spectrum Antiviral Inhibitors
Yiqing Yang, Lin Cao, Hongying Gao, Yue Wu, Yaxin
Wang, Fang Fang, Tianlong Lan, Zhiyong Lou, and Yu Rao
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00091 • Publication Date (Web): 02 Apr 2019
Downloaded from http://pubs.acs.org on April 3, 2019
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, Optimization and Target Identification of
Novel Potent Broad-Spectrum Antiviral Inhibitors
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
Yiqing Yang†, §, #, Lin Cao , Hongying Gao
‡, #
†, §, #
†
⊥
†
, Yue Wu , Yaxin Wang , Fang Fang , Tianlong
†
&,
†,
Lan , Zhiyong Lou *, Yu Rao *
†MOE Key Laboratory of Protein Sciences, School of Pharmaceutical Sciences, MOE Key
Laboratory of Bioorganic Phosphorus Chemistry & Chemical Biology, Tsinghua University,
Beijing 100084, P.R. China.
§Tsinghua University-Peking University Joint Center for Life Sciences, Beijing 100084, P.R.
China.
‡College of Life Sciences, Nankai University, Tianjin 300071, P.R. China.
⊥College of Life Sciences, Hebei Normal University, Shijiazhuang, Hebei 050024, P.R. China.
&School of Medicine and Collaborative Innovation Center of Biotherapy, Tsinghua University,
Beijing 100084, P.R. China.
ABSTRACT: Viral Infections are increasing and probably long-lasting global risks. In this study,
a chemical library was exploited by phenotypic screening to discover new antiviral inhibitors.
After optimizations from hit to lead, a novel potent small molecule (RYL-634) was identified,
showing excellent broad-spectrum inhibition activity against various pathogenic viruses, including
ACS Paragon Plus Environment
1

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
hepatitis c virus (HCV), dengue virus (DENV), zika virus (ZIKV), Chikungunya virus (CHIKV),
enterovirus 71 (EV71), human immunodeficiency virus (HIV), respiratory syncytial virus (RSV)
and others. The mechanism of action and potential targets of RYL-634 were further explored by
the combination of activity-based protein profiling (ABPP) and other techniques. Finally, human
dihydroorotate dehydrogenase (HsDHODH) was validated as the major target of RYL-634. We
did not observe any mutant resistance under our pressure selections with RYL-634, and it had
strong synergistic effect with some FDA-approved drugs. Hence, there is great potential for
developing new broad-spectrum antivirals based on RYL-634.
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
INTRODUCTION
Virus infection seriously threatens global health. For example, HCV, which can be cured now, still
1
annually infects over 185 million people and causes up to 500,000 deaths ; the infection numbers
of two other flaviviruses, DENV and ZIKV, are approximately 390 million around the world and
2
-3
4
million in Brazil alone, respectively . Since neither specific antiviral drugs nor vaccines are
available for newly emerging viruses, including DENV and ZIKV, related diseases result in heavy
4
-5
social burdens . Although the research community has made great efforts to discover therapeutic
6-8
small molecules for coping with such emergencies , novel potent molecules have been rarely
reported. The majority of reported effective agents in the literature are nucleotides or known
antivirals. To fight against future viral epidemics, which may emerge from known, currently
unknown or neglected viruses, broad-spectrum antiviral agents are particularly needed.
According to the mechanisms of action, antiviral reagents can be divided into virus-targeting
antivirals (VTAs) and host-targeting antivirals (HTAs)9-11. Although VTAs are more virus-
specific, they can relatively easily generate resistant mutants12-14. The allure of HTAs is their
ACS Paragon Plus Environment
2

Page 3 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
broad-spectrum antiviral activity. This activity is particularly attractive for some viral
emergencies, such as the severe acute respiratory syndrome coronavirus (SARS-CoV), the
influenza virus, the Ebola virus and the recent ZIKV outbreaks because a broad-spectrum agent
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
allows for immediate treatment without a lead time being required to develop a specific therapy15.
HTAs could also provide therapeutic opportunities for niche indication in cases when the patient
population is not large enough to provide an economic incentive for the development of virus-
1
5
specific drugs but when the medical need is acute . Although the discovery of HTAs has attracted
much attention13, 15-16, ribavirin is the only approved host-targeting antiviral small molecule drug,
and new HTAs are much needed13, 17
.
In the process of drug discovery, compounds discovered from phenotypic assays on the cellular
level may be more likely to be efficacious than target-based approaches in the following pre-
1
8
clinical or clinical studies . There has been a resurgence of phenotypic drug discovery in recent
years, which evades the incompletely understood complexity of diseases and does not rely on
1
9
knowledge of a specific target or its role in a disease . However, the mechanisms behind
biologically active molecules, which are usually first-in-class drugs from phenotypic screens,
urgently need to be revealed. Although various methods for targets identification, such as ABPP20-
2
1
22
23
24
25-
, reverse docking , genome-wide RNAi or CRISPR-Cas9 , proteome microarray and so on
26,
have been developed to solve this problem, target identification is still a challenge.
Here, we report the discovery of RYL-634 (compound 1, Figure 1) via a phenotypic screen, which
is a novel potent broad-spectrum and host-targeting antiviral small molecule against HCV, DENV,
ZIKV, CHIKV, EV71, HIV, Middle East respiratory syndrome coronavirus (MERS-CoV), RSV,
severe fever with thrombocytopenia syndrome virus (SFTSV) and the influenza virus. Its
mechanism of action and a target study are also described here.
ACS Paragon Plus Environment
3

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
RESULTS AND DISCUSSION
Discovery and optimization of RYL-634 as a novel antiviral agent. In our previous project of
developing antimalarial PfNDH2 inhibitors, we constructed a mini focused library consisting of
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
27
~200 biaryl-substituted quinolones . Quinolone is a privileged structure in medicinal chemistry
that behaves many biological activities, but this kind of biaryl-substituted quinolone was never
reported as antiviral. Considering the current antiviral status mentioned above, we initiated an
antiviral drug discovery study via screening the library in our anti-HCV assays. First, the
compounds were screened by a high-content fluorescence microscope with a green fluorescent
protein (GFP)-fused HCV to rapidly find hit compounds (Figure 1A). Then, the hits were
quantitatively validated by RT-PCR or the luciferase reporter method. Once the EC values were
5
0
precisely calculated, the structure-activity relationship (SAR) was summarized and instructed us
to design and synthesize new molecules for hit optimization.
Serendipitously, some active compounds were found from this mini library which exhibited the
advantage of privileged structure in drug discovery. Compound RYL-552S (compound 2, Figure
1
B and Figure S1) stood out in the screening (EC = 0.39 μM) (Table 1) and was used as the hit
50
compound for the next optimization process. To synthesize the target compounds, various 1-biaryl-
propan-1-ones were constructed via Suzuki cross-coupling. These intermediates were then
condensed and cyclized with compound 64 under acidic condition to generate the target
compounds (Scheme 1, Table 1). The EC value dropped to 1.56 μM for the similar analogue
50
RYL-552 (compound 3) when the sulfur atom was merely replaced by the oxygen atom. It
indicated that the -SCF group may play an important role in the biological activity. Therefore,
3
some alternatives to -SCF were introduced into the molecules. (Figure 1B, Table 1). We first used
3
-OCH CF to probe the possible modifications (compound 5, Table 1). The insertion of methylene
2
3
ACS Paragon Plus Environment
4

Page 5 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
between the oxygen and -CF made it more potent than compound 3. Further methylation and
3
hydroxylation on the methylene improved the potency (compound 6), but it decreased largely
when only hydroxy group was kept on the methylene (compound 7). It seemed that some
hydrophobic groups like methyl groups would be good for the activity (compare compound 3 with
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
5
, 6 and 7). To verify the above hypothesis, we oxidized the sulfur atom of RYL-552S (3) to
generate more polar molecules like sulfoxide, sulfone and sulfoximines (compounds 8-11). All
these modifications interfered the anti-HCV activity (compare compound 2 with 8-11) which
indicated that the better potency of RYL-552s compared with RYL-552 resulted from the
hydrophobic property of the sulfur atom. Thus, compound 4 (Figure 1B) with substitution by a
hydrophobic t-butyl group was synthesized and was found to inhibit viral infection with similar
potency. Obviously, the -CF group was not essential for keeping the antiviral activity and it could
3
be replaced by hydrophobic groups. We also tried to simplify the substitution with -OCHF , -SF ,
2
5
-Cl and -OMe (compounds 12-15) as -SCF alternatives. These smaller functional groups didn’t
3
give good results possibly due to the unsuitable distance or orientation to the residues in the target.
Similarly, the introduction of constrained 6-member rings morpholine and piperidine gave bad
results too (compounds 16 and 17). When the -F and -O- groups was used to replace the -H on
benzene ring and the methylene between the di-aryl respectively, we found that compounds 3 and
1
8 (Table 2) behaved nearly same potency. With the -F and -O- groups remained, these 1-biaryl-
propan-1-one intermediates can be synthesized more easily via metal-free nucleophilic aromatic
substitution reaction (Scheme 1, A). Then some hydrophobic alkyl groups were introduced on the
nitrogen atom of the aniline, which led to the discovery of the most potent molecule RYL-634
(Figure 1B, Table 2, compound 1). In our test, the significant antiviral activity of RYL-634 could
be observed by fluorescence microscope in a dose-dependent manner (Figure 1C). It had EC50
ACS Paragon Plus Environment
5

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
values of approximately 5 nM from both RT-PCR and luciferase method with PSI-7977
(
sofosbuvir, EC = 0.17 μM) as the positive control (Figure 1D and 1E). The cell viability was
50
also monitored at the same time (Table 1-2), and the CC value of RYL-634 was higher than 2.5
5
0
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
μM. The high selectivity index (SI = CC /EC ) suggested that the excellent potency of RYL-634
5
0
50
was not due to cytotoxicity. During this round optimizing process, we found again that constrained
rings were worse than ring-opened substitutions (compare compound 1 with 19, 20 and 25 in Table
2
). The size of substitutions was also important so that modifications with di-ethyl gave the best
inhibition (compound 1). If the ethyl group was replaced by smaller methyl or larger propyl
(compounds 22-24 and 21), the inhibition significantly became weak.
ACS Paragon Plus Environment
6

Page 7 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 1. Discovery and optimization of novel antiviral small molecules. (A) The workflow for novel antiviral
drug discovery in this project. Each compound was screened at 10 μM with the anti-HCV assay, the intensity of
the fused-GFP was monitored quickly under a fluorescence microscope. Only compounds with over 50%
fluorescence decay were diluted to different concentrations and re-evaluated by qRT-PCR or luciferase reporter
method. The calculated IC50 values then guided the structure-activity relationship (SAR) study. (B) RLY-634
was identified after hit optimization. (C) The antiviral activity of RYL-634 was observed by fluorescence
microscope. The diagrams in the first line were GFP signals. The diagrams in the second line were stained
nucleus. The scale bar is 100 m. (D) The antiviral activity of RYL-634 was confirmed by RT-PCR and
luciferase reporter. (E) The antiviral activity of RYL-634 was compared with PSI-7977 (sofosbuvir) by
luciferase reporter. Each experiment was repeated three times. Error bars stand for SD.
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
ACS Paragon Plus Environment
7

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
Scheme 1. The synthetic routes for target molecules.
F
N
F
O
O
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
O
4NH2
6
R
N
H
R
TfOH, n-BuOH
49-63 and 67-80
1, 4-25 and 29-34
SCF3
O
O
O
HO
SCF3
O
H
N
O
NH2
EDCI, DMAP, HOBt
Et3N, DMF
76
48
SCF3
SCF3
O
77
1
1
.1 NaN3, DMF
.2 PPh3, THF
HO
K CO , DMF
2
3
O
O
O
R
O
B
B
R
R
O
O
O
Br
B
.
Pd(dppf)2Cl2 DCM
KOAc, DMF
.
O
Br
Pd(dppf)2Cl2 DCM, K3PO4, Toluene
R=:
F
OH
CF3
53
35-47
O
CF3
O
SF5
O
When R=-SCF3
49
5
0
51 F
52
.
Pd(dppf)2Cl2 DCM,
OH
O
O
O
O
NH
Br
K3PO4, Toluene
S
S
S
CF3
CF3
O
CF3
CF3
5
4
55
56
57
O
O
N
Cl
OMe
S
N
N
CF3
63
58
59
60
61
62
SCF3
NH2
O
O
O
O
R
NH2
F
F
F
Cl
Alkylation reagent
K2CO3, MeCN
HO
K2CO3, DMF
O
O
AlCl3
F
F
F
65
66
R=:
SCF3
H
N
HO
N
N
N
K2CO3, DMF
68
69
70
71
O
SCF3
NH
N
N
N
N
O
F
72
73
74
75
67
O
O
R=:
Alkylation reagent
K2CO3, Acetone
N
N
N
R
O
78
O
79
80
OH
ACS Paragon Plus Environment
8

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. The anti-HCV activity of compounds with -SCF alternatives.
3
F
O
R
N
H
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
Compound
R
Activity (EC , μM)
Toxicity (CC , μM)
SI
5
0
50
2
3
OCF3
SCF3
1.56±0.15
0.36±0.02
>10
>10
>6
>27
4
5
6
0.29±0.02
0.84±0.04
0.11±0.01
>5
>5
>17
>6
O
CF3
OH
>2.5
>23
CF3
OH
7
3.49±0.21
>10
>3
CF3
O
8
9
0.76±0.11
1.76±0.29
1.42±0.31
>5
>7
>2
>7
S
CF3
O
O
>2.5
>10
S
S
CF3
NH
O
O
1
1
0
1
CF3
N
5.39±0.36
1.01±0.11
>10
>5
>2
>5
S
CF3
F
O
12
F
1
3
SF5
0.73±0.07
4.40±0.42
2.17±0.37
>5
>7
>2
>5
14
15
Cl
>10
>10
OMe
O
1
1
6
7
3.65±0.11
1.92±0.07
>10
>10
>3
>5
N
N
ACS Paragon Plus Environment
9

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
Table 2. The anti-HCV activity of compounds with hydrophobic alternatives.
F
O
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
R
N
H
O
F
Compound
R
Activity (EC , μM)
Toxicity (CC , μM)
SI
50
50
SCF3
1
8
0.36±0.01
>10
>27
>416
1
0.006±0.007
>2.5
N
19
0.10±0.01
>10
>100
N
N
2
2
2
0
1
2
0.035±0.01
0.066±0.01
0.24±0.05
1.05±0.12
0.17±0.02
>10
>10
>10
>10
>10
>285
>151
>42
N
N
23
>10
N
H
N
2
4
>58
NH
25
2.57±0.15
>10
>4
N
The antiviral profile of RYL-634. With RYL-634 in hand, its antiviral potential was further
exploited. In current antiviral therapies, drug combinations or cocktails have been utilized more
commonly to avoid or lower the chance of developing drug-resistant mutations28-29. The
combination effects of RYL-634 with some FDA-approved antiviral drugs were tested on HCV
(Figure 2A). We treated HCVcc-infected cells with various concentrations and with RYL-634
either alone or in combination. The results revealed a concentration-dependent inhibition with all
ACS Paragon Plus Environment
10

Page 11 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of them alone and with any two in combination (Figure 2A). The data were further analyzed using
a mathematic model, MacSynergy II (Figure S2), and revealed that RYL-634-boceprevir, RYL-
634-cyclosporin A, RYL-634-ribavirin and RYL-634-sofosbuvir pairs have strong synergy (CI >
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
5
0), whereas the synergistic antiviral effect of RYL-634 with IFN-α, daclatasvir and telaprevir is
under moderate (50 > CI >10), according to the suggested criteria. Next, we investigated the
working step of RYL-634 in the HCV lifecycle. First, we observed that both HCVcc and HCV
subgenomic replicon (HCVrep) are potently inhibited by RYL-634 treatment; however, HCV
pseudo-typed virus (HCVpp) activity was not affected in the presence of RYL-634 at a high
concentration (Figure 2B). This meant that RYL-634 inhibited HCV at the step of viral replication.
Next, we assessed the step that RYL-634 worked on via time-of-addition analysis (Figure 2C).
Huh7.5.1 cells were infected with HCVcc at 4 °C for 2 h (T = -2 h). After removing unbound
viruses, cells were incubated at 37 °C (T = 0 h), and compounds were added to the infected cells
at different time points. We selected the anti-CD81 antibody as a control to represent typical
working HCV entry inhibitors. HCV inhibiting effects of the anti-CD81 antibody decreased at the
time of the temperature shift. The growth curve of HCV in the presence of RYL-634 indicated that
RYL-634 acts on the replication. Further study revealed that RYL-634 is a potent compound
against other RNA viruses (Figure 2D). It showed excellent potency against DENV (EC = 7 nM),
5
0
ZIKV (EC = 20 nM), EV71 (EC = 4 nM) and HIV (EC = 13 nM), and obvious inhibition
5
0
50
50
against CHIKV, RSV, SFTSV, MERS-CoV and influenza virus at sub-micromolar concentrations.
Thus far, there have been no effective inhibitors discovered for many of these viruses. Although
the recent ZIKV outbreak has resulted in some projects for anti-ZIKV and anti-DENV drug
discovery, RYL-634 inhibited these viruses as well as, if not better than, the best molecules from
those projects7, 30-31. The broad-spectrum antiviral activity indicated that RYL-634 may be a host-
ACS Paragon Plus Environment
11

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
targeting molecule. If RYL-634 was a viral-targeting molecule, viral strains with resistance against
RYL-634 might be generated under the sustained pressure of high concentrations. However, any
mutant resistant strain was never observed after 8 rounds of screening in our pressure selections
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
with boceprevir as the control32.
Figure 2. The antiviral profile of RYL-634. (A) The combination of RYL-634 with FDA-approved drugs (IFN-
α, ribavirin, cyclosporin A, telaprevir, boceprevir, daclatasvir and sofosbuvir). The combination indexes were
shown in Figure S2. (B) Inhibitory activities of RYL-634 on HCVcc, HCVpp and HCVrep. (C) The antiviral
time course of RYL-634. (D) The broad-spectrum antiviral activity of RYL-634. SI, selectivity index. Each
experiment was repeated three times. Error bars stand for SD.
Target identification by ABPP. To explore the potential targets of RLY-634 and understand the
mechanism of action at the molecular level, we next focused on target identification. ABPP has
ACS Paragon Plus Environment
12

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
20
become a powerful method for target identification . In a typical ABPP procedure (Figure 3A), a
3
3
probe with a crosslinker and alkyne tag is first designed and synthesized . Then, the probe is
incubated with live cells in vivo or in cell lysate in vitro to bind with target proteins. Once the
covalent binding is triggered by UV, the probe-labeled sample reacts with azide-tagged rhodamine
and biotin (Figure 3B) via click chemistry. The results are monitored by in-gel fluorescence and
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
4
LC-MS/MS . ABPP may be quite suitable for the target identification of RYL-634 because our
SAR study has indicated that some hydrophobic functional groups are essential for its biological
activity (Figure 1B, Table 1-2). As illustrated in Scheme 2, the terminal alkyne-containing
diazirine photocrosslinker could be used as a hydrophobic functional group to replace the alkyl
chain in RYL-634. Compound 87 was used as the positive probe (EC = 8 nM on HCV). When a
5
0
methylation reaction occurred on the oxygen atom of compound 87, the quinolone ring became a
quinoline. This change made compound 88 lose the most potency and serve as an ideal negative
control (EC > 5 μM on HCV), which also indicated that the quinolone scaffold is vital for the
5
0
antiviral activity. In the following ABPP studies, we firstly observed the labeling position via in-
gel fluorescence scanning for different probe concentrations (Figure 3C and S3). Three
experimental groups were set up for each concentration in which the positive probe, negative probe
and competitor (RYL-634) were studied. After optimizations, the labeling process can be
efficiently conducted with a probe concentration between 1.0 and 10.0 μM. In the first pull-down
experiment, 833 proteins were identified that bound with the positive probe more than the negative
or DMSO control (Figure 3D and Table S1). If a threshold (positive/negative ratio of >3 and
positive/DMSO of >5) was set to sort these proteins, the amount could be narrowed down to 290
proteins over it (Figure 3D and Table S1). Interestingly, only human proteins were found in the
proteomic data, despite the presence of viral components in the pull-down experiment. Both this
ACS Paragon Plus Environment
13

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
point and the broad-spectrum antiviral property described above (Figure 2A) strongly suggested
that RYL-634 is a host-targeting antiviral molecule. Five proteins (DOCK1, SIAE, DHODH,
UTRO and MLC1) occupied the highest region of the P/N ratio in the first experiment (Figure 3D
and 3E). However, these candidates were treated with caution because the protein scores in the
hazardous group were very low, indicating that their signal may be caused by the stochastic nature
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
5
of data-dependent acquisition by LC-MS/MS . In contrast, the abundance of proteins in the
conservative group is always high; however, the selectivity (P/N ratio) may not be good. Proteins
in that group may be originally abundant in the biological system, and their non-specific binding
with the probes may be strong. Then, we further conducted several replicates of the experiment to
eliminate random signals and to further enrich the target proteins. In this case, 78 proteins appeared
in all three replicates, and another 302 proteins appeared in two of the three replicates (Figure 3F).
For the sake of insurance, all the proteins that appeared in at least two replicates were considered
in the following study.
ACS Paragon Plus Environment
14

Page 15 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 3. Target identification by activity-based protein profiling (ABPP). (A) The workflow for ABPP. (B)
The structures of rhodamine and biotin with azide tag. (C) The in-gel fluorescence of ABPP. Left, in-gel
fluorescence scanning. Right, the corresponding Commassie brilliant blue staining of the same gel which
indicated the same sample loading. Neg, negative probe. Pos, positive probe. Com, competitive (positive and
RYL-634). For the enlarged version, see Figure S3. (D) The detected proteins by LC-MS/MS in the first
experiment. (E) Protein distribution in a coordinate system of P/N (positive probe to negative probe) ratio and
protein scores. (F). The Venn diagram of three experiments.
Bioinformatic analysis of the potential targets. We performed a bioinformatic analysis next.
First, the subcellular distributions of the candidates were concluded from the gene ontology
analysis (Figure 4A and Table S1). Most of them were distributed in cytosol (29%), in the nucleus
ACS Paragon Plus Environment
15

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
(
29%) and in the mitochondrion (15%). Some others also dispersed in the extracellular matrix
7%), the endoplasmic reticulum (ER, 6%), the Golgi apparatus (4%), on the cell surface (3%), the
(
proteasome (3%), the ribosome (2%) and the endosome (2%). In the following analysis of the
biological process, we found that these candidates were involved in many biological processes,
and the representative ones are summarized in Figure 4B, especially the small molecule metabolic
process (Table S1). To our delight, some of the enriched proteins participated in the viral process,
the viral life cycle and viral transcription, which might be related with the broad-spectrum antiviral
activity of RYL-634 (Figure 4B). They correlated with the entire viral life cycle, including viral
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, molecular metabolism, transcription and translation. (Table S2, Figure 4C and 4D)36-37
.
More importantly, some of them, such as HSP90, HSPA5, IMPDH2, DHODH, lipid-related
enzymes and PDI, have been purposed to be or validated as broad-spectrum antiviral targets37-38
.
RYL-634 had a high chance to act on them.
Scheme 2. Design and synthesis of clickable and photoreactive probes.
O
O
H
NH2
EtI, K2CO3, MeCN
N
O
O
O
N
F
66
F
I
67
2 3
K CO , MeCN
OH
o
O
1. NaBH , MeOH, -30 C, 2h
OH
O
4
2
. I2, imidazole, PPh3, DCM
F
84
HO
81
83
F
F
O
F
N
F
F
O
O
Dess-Martin, DCM
MeI, K2CO3. DMF
N
N
N
N
H
NH2
N
H
64
O
OH
O
86
TfOH, n-BuOH
O
F
F
F
F
85
O
O
N
1.1 NH3 in MeOH
.2 NH2OSO3H
.3 Et3N, I2
1
1
N
N
H
N
N
N N
O
O
87
88
Postive probe, EC50=8 nM on HCV
Negative probe, EC50>5 M on HCV
ACS Paragon Plus Environment
16

Page 17 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 4. Bioinformatic analysis of the potential target proteins. (A) Cellular component analysis of the potential
target proteins. (B) Biological process analysis of the potential target proteins. (C) Replication cycle and
polyprotein organization of flaviviruses. (D) Some potential target proteins in different viral infection stages
were selected by searching literatures. See Table S1 and S2 for details.
Target enrichment by SAR-guided selection in reverse docking. The 3D structures for the most
enriched proteins have been provided (Figure S4), which inspired us to also probe the structure-
based methods. Reverse docking (Figure 5A) with its low-cost and high throughput is proving to
be a useful tool for drug repositioning and target identification39-40. Therefore, some programs or
4
1
databases, such as TarFishDock , have been developed. Different scoring functions (DOCK,
GOLD, AutoDock, Glide, etc.) have been utilized in reverse docking for protein ranking; however,
4
2
reverse docking for protein ranking can be strongly influenced by the different scoring functions .
ACS Paragon Plus Environment
17

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
Some specific factors should be introduced into the manual selection stage to filter more possible
targets from the candidate pool. Since the SAR for RYL-634 has been extensively studied as
described above (Figure 1B, Table 1-2) and because SAR represents a unique scaffold feature for
bioactive compounds, we envisioned that the consistency between SAR and reverse docking can
be used as a filter to specifically enrich the target candidates.
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
In this workflow (Figure 5A), the conformation of RYL-634 with the lowest energy was calculated
first; meanwhile, a protein library was constructed from PDTD. The top 120 binding models with
Glide scores below -8.0 were further sorted according to SAR. This meant that the binding model
should excellently explain the SAR but should not conflict with it. Finally, 20 proteins were
enriched by this SAR-guided selection (Figure 5B and Figure S5). We noticed that dihydroorotate
dehydrogenase (DHODH) appeared again in the enrichment. Furthermore, its binding model had
the best consistency with the SAR. In the binding model (Figure 5C and Figure 5D), the carbonyl
group and the fluorine atom on the quinolone ring formed a hydrogen bonding network with R136
and Q47 of DHODH. Accordingly, the potency abated if the fluorine atom was moved to other
positions on the quinolone ring (Table 3, compounds 3, 26-28) or if a methyl group was installed
on the oxygen atom, which would disrupt these hydrogen bonds (the positive probe 87 vs. negative
probe 88). The right benzene ring of RYL-634 was vised by F62 and Y38 via π-π stacking so that
the potency was mostly lost when the benzene ring was replaced by some saturated groups (Table
3
, compounds 29-31). At the end of RYL-634, the nitrogen atom and the diethyl group contributed
to a hydrogen bond with Y38 and hydrophobic interactions around the Leu residues, respectively.
This can explain why the diethyl-substituted amine group could be used to replace other
hydrophobic groups and give the best activity (Figure 1B, Table 1-2). Since the above described
π-π stacking, hydrogen bonding and hydrophobic interactions with DHODH in the model, keeping
ACS Paragon Plus Environment
18

Page 19 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a one-atom length, such as an oxygen atom, on the linker between di-aryl was reasonable (Figure
5
D). Regardless of the abstraction of an atom or insertion of more atoms, the tandem interactions
were disarranged, and the potency decreased (Table 3, compounds 32-34). A similar binding style
of the positive probe (Scheme 2, compound 87) with DHODH was also observed from the
molecular docking (Figure 5E) by which the low abundance of DHODH in LC-MS/MS may be
elucidated (Figure 3E). In the light of the high protein expression level of DHODH in cells (Figure
S6), the low abundance may be caused by the low reaction efficiency of the probe with the protein.
In detail, the additional terminal alkyne-containing diazirine photocrosslinker in the positive probe
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
(Scheme 2) brought about more hydrophobic interactions with several Leu residues compared with
RYL-634 (Figure 5E). The C-H bonds from the Leu residues had a lower reactivity with the
43
carbene intermediate than the polar protic O-H or N-H bonds ; however, the diazirine group was
fixed at an orientation toward the solvent exposure by these hydrophobic interactions. The
diazirine could rearrange into a diazo compound, and this binding position could make it form an
unstable intermediate, which rapidly reacted with nucleophiles including water44.
Thus, three clues to DHODH have been collected. First, DHODH was identified by LC-MS/MS
in the pull-down experiment. Second, the bioinformatic analysis showed that DHODH is closely
related with the viral life cycle (Figure 4D). DHODH supplies pyrimidine for viral replication, and
we observed that RYL-634 extensively inhibits the viral replication stage (Figure 2D). Finally, it
was enriched again by the SAR-guided selection in the reverse docking (Figure 5B). All these data
encouraged us to focus on DHODH in the following study.
ACS Paragon Plus Environment
19

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. Target enrichment by SAR-guided selection in reverse docking. (A) The workflow for reverse docking.
(B) The selected proteins after SAR-guided filtration. (C) The binding model of RYL-634 with HsDHODH. (D)
The binding model of RYL-634 with HsDHODH was consistent with its SAR very well. (E) The binding model
of positive probe 87 with HsDHODH.
ACS Paragon Plus Environment
20

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. The anti-HCV activity of compounds with other modifications.
Compound
Structure
Activity (EC , μM)
Toxicity (CC , μM)
SI
5
0
50
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
F
O
3
0.36±0.02
2.48±0.2
0.97±0.1
1.85±0.32
>10
>10
>5
>27
SCF3
N
H
O
F
2
2
2
6
7
8
>4
>5
>3
SCF3
N
H
O
SCF3
F
N
H
O
>5
SCF3
N
H
F
F
O
2
9
28.12±1.34
>100
>4
N
H
O
N
O
F
O
3
3
0
1
25.56±1.94
16.65±0.9
>100
>100
>4
>6
N
H
N
O
O
F
O
N
H
N
F
O
3
3
2
3
2.33±0.13
1.59±0.17
11.05±0.88
>10
>10
>50
>4
>6
>5
N
H
SCF3
F
O
N
H
O
SCF3
F
O
34
SCF3
N
H
H
N
O
ACS Paragon Plus Environment
21

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
DHODH was validated as a target of RYL-634. DHODH is a flavin-dependent mitochondrial
enzyme that catalyzes the oxidation of the intermediate, dihydroorotate, to orotate with coenzyme
4
5
Q as a cofactor in the de novo biosynthesis of pyrimidine (Figure 6A) . Both de novo and salvage
pathways provide precursors for further synthesis of RNA, DNA, glycoproteins and phospholipids.
The inhibition of DHODH has been validated as a promising therapeutic strategy for viral
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
infection, cancer, arthritis and immunosuppression . To verify if RYL-634 is an inhibitor of
DHODH, a chemical rescue experiment was designed. In the hypothesis, the de novo pathway was
repressed by RYL-634 by inhibiting DHODH, and the downstream supply of intermediates for
viral replication was blocked; then, the viral replication would be rescued if the intermediate is
compensated from the salvage pathway. The uridine, a metabolite in the pyrimidine salvage
pathway (Figure 6A), was supplemented to the viral infection system, and the antiviral activity of
RYL-634 was most significantly reversed (Figure 6B). In the presence of 0.5 mM uridine, the EC50
value of RYL-634 on HCV declined near 200-fold (Figure 6C). This phenomenon was as same as
previously reported cases for other DHODH inhibitors46-47. Then, the recombinant DHODH
enzyme was purified. In a DARTS study (Figure 6D), a protein might be less susceptible to
4
8
proteolysis when it is ligand-bound than when it is ligand-free . RYL-634 protected DHODH
from the degradation of pronase in a dose-dependent manner, while GAPDH was mostly degraded.
Obviously, RYL-634 can bind with DHODH very well. We also conducted enzymatic activity
assays to reveal the functional inhibition of RYL-634 to DHODH. The inhibition of RYL-634 to
DHODH was stronger 4-fold than teriflunomide, the only DHODH-targeted immunomodulatory
49
drug approved by the FDA for multiple sclerosis (Figure 6E) . Accordingly, the positive probe
(compound 87) had a better inhibitory effect than the negative probe (compound 88). Although
DHODH had been purposed as antiviral target, its inhibitors had been rarely well studied for the
ACS Paragon Plus Environment
22

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
broad-spectrum antiviral activity. Teriflunomide was purposed to be used as antiviral by targeting
5
0
DHODH very recently from a genome-wide siRNA screen , but its antiviral activity of
micromolar level was not good enough, while RYL-634 had better antiviral activity, corresponding
to its stronger inhibition to DHODH. DHODH was a conserved protein between different species
such as human, monkey, dog and mouse (Figure S7), that could explain why the antiviral activity
assays from host cells of different species were all good (Figure 2E). All the above results made
us believe that the enzymatic activity of RYL-634 can mostly explain its cellular antiviral activity
and DHODH should be a major antiviral target of RYL-634. Noteworthy, these biaryl-substituted
quinolone compounds have been mainly used as antimalarial PfNDH2 inhibitors in previous
studies27,51. Their inhibition on DHODH have never been reported. What’s more, there are no
NDH2 or homologous proteins in both human cells and in the virus. RYL-634, a HsDHODH
inhibitor, has a completely different mechanism of action with the previous compounds. Although
DHODH could be viewed as a major target of RYL-634, these results did not exclude the
possibility that besides inhibiting DHODH, it may also exert its antiviral activities through other
mechanisms. In parallel, we also attempted to validate other potential targets during this process;
however, we only confirmed that RYL-634 has no activity on purified enzymes, including HSP90,
PRPS1 and IMPDH2.
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
ACS Paragon Plus Environment
23

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
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. DHODH was validated as a target of RYL-634. (A) The de novo and salvage pathways for pyrimidine
biosynthesis. (B) The antiviral activity of RYL-634 was reversed by supplementing uridine. (C) The EC value
5
0
of RYL-634 significantly increased in the presence of uridine. (D) DHODH was protected by RYL-634 in the
study of DARTS. (E) The inhibition of DHODH activity by teriflunomide, RYL-634, positive probe and negative
probe. Each experiment was repeated three times. Error bars stand for SD.
CONCLUSIONS
In summary, RYL-634, a novel potent broad-spectrum antiviral small molecule, has been
identified by phenotypic screening and validated as a DHODH inhibitor. A detailed structure-
activity-relationship (SAR) of these compounds has been discussed. Although some viral
infections, such as HCV and HIV, can be controlled by medications, there are still no good choices
for many others, such as DENV, ZIKV, CHIKV, and SFTSV. Moreover, viruses usually generate
resistances to drugs. The known broad-spectrum antivirals including some DHODH inhibitors
usually are not as potent as RYL-634 in cellular assays, or else show serious cytotoxicity which
extremely limited their clinical utilization36, 52. For example, brequinar, a strong DHODH inhibitor,
ACS Paragon Plus Environment
24

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
had gained a lot attention in previous but failed in the clinical study due to its side effect and poor
solubility53-54. Moreover, the DENV resistance to brequinar via enhancement of polymerase
5
5
activity was reported . Most of the DHODH inhibitors including the only FDA-approved
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
teriflunomide behaved very weak antiviral activity that couldn’t be competent for a drug56-57. Some
of the DHODH inhibitors contained “soft groups” like labile esters that easily metabolized in vivo
and presumably led to poor pharmacokinetic stability46, 58-59. The potent broad-spectrum antiviral
activity of RYL-634 with only weak cytotoxicity that does not show detectable resistance in anti-
HCV assays and its combination with other drugs provide new insights on the development of
antivirals that are likely to avoid drug resistance. Considering the importance of pharmacokinetic
4
6
properties which might lead to failure of previous DHODH inhibitor in vivo , tremendous efforts
for lead optimization to give more drug-like analogues of RYL-634 in the following pre-clinical
study are essential.
Target identification is a time- and budget-consuming process, especially for some promiscuous
molecules. In ABPP, the sensitivity is limited for compounds that exhibit a low binding affinity
toward their target or for targets expressed at low levels. In these cases, the target protein can be
lost during washing steps, or its binding may be obscured by the presence of highly abundant and
6
0
non-specific binding proteins . In the reverse docking, which is limited by the volume of the PDB
database, protein preparation and scoring functions, the bias and false positives are always
observed. Therefore, combining multiple methods for target identification is helpful for
overcoming the corresponding limitations and making the process more direct. DHODH has been
validated as an antiviral target in previous studies13, 37, 61-64. In addition, it can be used as a
therapeutic target for cancer, rheumatoid arthritis, sclerosis, malarial and immunosuppression45, 65-
66. RYL-634, a new hit compound for human DHODH, could be used in the development of
ACS Paragon Plus Environment
25

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
therapeutic agents for these conditions. Especially for malarial, this kind of compounds are likely
to inhibit PfDHODH too, because the HsDHODH and PfDHODH are near homologues59, 67-68
.
Since these compounds were originally used as antimalarial PfNDH2 inhibitors27, 51, there is great
chance to develop dual inhibitors of PfNDH2 and PfDHODH based on RYL-552 and RYL-634.
This work provided some important information for the further mechansim study and potential
clinical utilizations of RYL-634 in the future.
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
EXPERIMENTAL SECTION
Chemistry. All commercial chemical materials (Energy, Bide, Macklin, J&K Chemical Co. Ltd.)
1
13
were used without further purification. All solvents were analytical grade. The H NMR and C
NMR spectra were recorded on a Bruker AVANCE III 400 MHz spectrometer in CDCl , CD OD,
3
3
13
or DMSO-d using tetramethylsilane (TMS) or solvent peak as a standard. All C NMR spectra
6
were recorded with complete proton decoupling. Low resolution mass spectral analyses were
performed with a Waters AQUITY UPLCTM/MS. Analytical thin-layer chromatography (TLC)
was performed on silica gel F254 plates from Yantai Chemical Industry Research Institute, and
flash column chromatography was performed on silica gel (200 − 300 mesh) from Qingdao
Haiyang Chemical Co. Ltd. A BUCHI Rotavapor R-3 was used to remove solvents by evaporation.
The purities of all the final tested compounds was more than 95%, which were confirmed by NMR
or UPLC.
2
7
Compounds 2, 3 and 26 – 28 were prepared according to the previously reported procedures .
All the other final compounds (compounds 1, 4 – 25, 29 – 34 and 85) were synthesized according
to the following synthetic procedure for compound 1, when the corresponding intermediates were
replaced. The details for preparing the intermediates were shown in Supporting Information.
ACS Paragon Plus Environment
26

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
-(4-(4-(Diethylamino)phenoxy)-3-fluorophenyl)-5-fluoro-3-methylquinolin-4(1H)-one
(1).To a 10 ml round bottom flask was added 560 mg compound 68 (2 mmol, 1,.0 eq), 400 mg 64
o
(1.0 eq), 170 μL TfOH (1.0 eq) and 3 mL n-BuOH as solvent. The mixture was refluxed at 130 C
overnight. Compound 1 was obtained after the purification by column chromatography (EtOAc:
hexane = 1: 1.5, isolated yield = 25%). To a 10 ml round bottom flask was added 560 mg compound
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
6
8 (2 mmol, 1,.0 eq), 400 mg 64 (1.0 eq), 170 μL TfOH (1.0 eq) and 3 mL n-BuOH as solvent.
o
The mixture was refluxed at 130 C overnight. Compound 1 was obtained after the purification by
1
column chromatography (EtOAc: hexane = 1: 1.5, isolated yield = 25%). H-NMR (400 MHz,
CD OD: CDCl = 0.2: 0.3, ppm): 7.48 (dd, J = 12.92 Hz, J = 6.92 Hz, 1H), 7.34 (d, J = 8.40 Hz,
3
3
1H), 7.29 (d, J = 11.04 Hz, 1H), 6.97 – 6.87 (m, 4H), 6.80 (d, J = 8.28 Hz, 2H), 3.33 (q, J = 7.08
13
Hz, 4H), 1.98 (s, 3H), 1.13 (t, J = 7.08 Hz, 3H). C-NMR (100 MHz, CD OD: CDCl = 0.2: 0.3,
3
3
ppm): 178.64, 163.09, 160.50, 154.48, 152.00, 148.62, 148.51, 147.93, 146.34, 146.15, 142.39,
142.35, 132.65, 132.55, 129.52, 129.45, 125.91, 125.88, 118.89, 118.28, 118.01, 117.82, 114.51,
+
1
4
14.40, 114.31, 109.81, 109.60, 45.48, 12.68, 12.49. LC-MS: calcd for C H F N O [M+H] :
2
6
25
2
2
2
35.18, found 435.23.
1
2
-(4-(4-(Tert-butyl)benzyl)phenyl)-5-fluoro-3-methylquinolin-4(1H)-one (4). H-NMR (400
MHz, CDCl , ppm): 11.54 (s, 1H), 7.77 (d, J = 8.44 Hz, 1H), 7.48 – 7.42 (m, 1H), 7.26 (d, J =
3
8
.08 Hz, 2H), 7.03 (d, J = 8.16 Hz, 2H), 6.99 (d, J = 8.44 Hz, 2H), 6.80 (dd, J = 11.64Hz, J = 7.96
1
3
Hz, 1H), 3.83 (s, 2H), 1.82 (s, 2H), 1.29 (s, 9H). C-NMR (100 MHz, CDCl , ppm): 177.59,
3
162.56, 159.97, 149.19, 148.14, 142.60, 142.08, 142.04, 137.29, 134.82, 132.60, 131.41, 131.31,
1
3
28.85, 128.80, 128.71, 125.51, 117.55, 114.71, 114.01, 113.92, 108.88, 108.67, 41.16, 34.49,
+
1.51, 12.43. LC-MS: calcd for C H FNO [M+H] : 400.20, found 400.23.
2
7
27
ACS Paragon Plus Environment
27

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
1
5
-Fluoro-3-methyl-2-(4-(4-(2,2,2-trifluoroethoxy)benzyl)phenyl)quinolin-4(1H)-one (5). H-
NMR (400 MHz, CDCl , ppm): 11.78 (s, 1H), 7.81 (d, J = 8.28 Hz, 1H), 7.46 – 7.43 (m, 1H), 7.22
3
(d, J = 7.56 Hz, 2H), 7.00 (d, J = 8.08 Hz, 2H), 6.92 (d, J = 7.48 Hz, 2H), 6.78 (d, J = 7.84 Hz,
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
13
2
1
1
6
H), 4.26 (q, J = 7.96 Hz, 2H), 3.79 (s, 2H), 1.76 (s, 3H). C-NMR (100 MHz, CDCl , ppm):
3
77.54, 162.46, 159.87, 156.02, 148.22, 142.52, 142.09, 142.05, 134.45, 132.62, 131.47, 131.36,
30.29, 128.91, 124.6, 122.10, 117.45, 115.11, 114.82, 113.94, 113.85, 108.88, 108.67, 66.53,
+
6.17, 65.82, 65.47, 40.69, 12.40. LC-MS: calcd for C H F NO [M+H] : 442.14, found 442.22.
2
5
20
4
2
5
4
-Fluoro-3-methyl-2-(4-(4-(1,1,1-trifluoro-2-hydroxypropan-2-yl)benzyl)phenyl)quinolin-
1
(1H)-on (6). H-NMR (400 MHz, CDCl : d -MeOD = 0.2:0.3, ppm): 7.51 (d, J = 7.84 Hz, 2H),
3
4
7.46 – 7.42 (m, 1H), 7.36 – 7.29 (m, 5H), 7.19 (d, J = 7.80 Hz, 2H), 6.89 – 6.83 (m, 1H), 4.00 (s,
13
2
1
H), 1.95 (s, 3H), 1.71 (s, 3H). C-NMR (100 MHz, CDCl : d -MeOD = 0.2:0.3, ppm): 177.41,
3
4
61.87, 159.28, 148.48, 142.50, 141.20, 141.16, 140.14, 137.34, 131.99, 131.35, 131.25, 129.81,
128.55, 128.34, 127.95, 126.97, 126.07, 124.13, 116.94, 113.40, 113.18, 113.09, 108.53, 108.32,
+
7
3.86, 73.57, 73.28, 73.00, 40.60, 22.25, 11.30. LC-MS: calcd for C H F NO [M+H] : 456.15,
2
6
22
4
2
found 456.22.
5
-Fluoro-3-methyl-2-(4-(4-(2,2,2-trifluoro-1-hydroxyethyl)benzyl)phenyl)quinolin-4(1H)-
1
one (7). H-NMR (400 MHz, CDCl : d -MeOD = 0.2:0.3, ppm): 7.48 – 7.33 (m, 6H), 7.30 (d, J =
3
4
1
3
7
.96 Hz, 2H), 7.20 (d, J = 7.88 Hz, 2H), 4.96 (q, J = 6.92 Hz, 1H), 4.01 (s, 2H), 1.94 (s, 3H). C-
NMR (100 MHz, CDCl : d -MeOD = 0.2:0.3, ppm): 179.25, 163.72, 161.14, 150.36, 144.36,
3
4
143.06, 143.02, 142.95, 134.91, 133.88, 133.23, 133.12, 130.40, 130.22, 130.07, 129.15, 127.66,
1
1
24.86, 118.76, 115.29, 115.24, 115.01, 114.92, 110.39, 110.17, 73.58, 73.27, 72.96, 72.64, 42.54,
+
3.07. LC-MS: calcd for C H F NO [M+H] : 441.14, found 441.21.
2
5
20
4
2
ACS Paragon Plus Environment
28

Page 29 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
-Fluoro-3-methyl-2-(4-(4-((trifluoromethyl)sulfinyl)benzyl)phenyl)quinolin-4(1H)-one (8).
1
H-NMR (400 MHz, CDCl , ppm): 11.46 (s, 1H), 7.69 (m, 3H), 7.52 – 7.40 (m 1H), 7.35 – 7.27
3
1
3
(m, 4H), 7.04 (d, J = 8.24 Hz, 2H), 6.80 (t, J = 9.60 Hz, 1H), 3.98 (s, 2H), 1.84 (s, 3H). C-NMR
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
(100 MHz, CDCl , ppm): 177.50, 162.58, 160.01, 147.72, 146.89, 141.98, 140.84, 133.49, 133.23,
3
1
1
31.61, 130.24, 129.23, 128.89, 126.41, 124.77, 123.09, 122.04, 117.48, 114.53, 113.93, 108.98,
+
08.77, 41.51, 12.44. LC-MS: calcd for C H F NO S [M+H] : 460.09, found 460.12.
24 18
4
2
5
-Fluoro-3-methyl-2-(4-(4-((trifluoromethyl)sulfonyl)benzyl)phenyl)quinolin-4(1H)-one (9).
1
H-NMR (400 MHz, d -DMSO, ppm): 11.55 (s, 1H), 8.10 (d, J = 8.24 Hz, 2H), 7.78 (d, J = 8.36
6
Hz, 2H), 7.54 – 7.50 (m, 5H), 7.36 (d, J = 8.44 Hz, 1H), 6.93 (dd, J = 11.96 Hz, J = 7.88 Hz, 1H),
1
3
4.27 (s, 2H), 1.82 (s, 3H). C-NMR (100 MHz, d -DMSO, ppm): 175.42, 161.73, 159.29, 152.00,
6
1
1
46.61, 141.80, 140.98, 132.92, 131.67, 131.56, 131.10, 129.32, 129.08, 127.20, 121.10, 117.88,
16.07, 114.09, 114.05, 113.01, 108.34, 108.14, 40.57, 12.02. LC-MS: calcd for C H F NO S
2
4
18
4
3
+
[M+H] : 476.09, found 476.19.
5
-Fluoro-3-methyl-2-(4-(4-(S-(trifluoromethyl)sulfonimidoyl)benzyl)phenyl)quinolin-4(1H)-
1
one (10). H-NMR (400 MHz, CDCl :d -MeOD = 0.2:0.3, ppm): 8.03 (d, J = 7.96 Hz, 2H), 7.52
3
4
–
6
7.47 (m, 3H), 7.43 (d, J = 7.80 Hz, 2H), 7.36 (d, J = 7.68 Hz, 2H), 7.30 (d, J = 8.40 Hz, 1H),
1
3
.89 (dd, J = 11.72 Hz, J = 8.16 Hz, 1H), 4.19 (s, 2H), 1.97 (s, 3H). C-NMR (100 MHz, CDCl :
3
d4-MeOD = 0.2:0.3, ppm): 179.16, 163.61, 161.03, 150.65, 149.77, 142.90, 142.40, 134.34,
1
33.09, 132.98, 131.95, 131.15, 131.05, 130.43, 130.38, 123.93, 120.62, 118.72, 115.03, 114.98,
+
114.84, 110.28, 110.07, 42.57, 13.02. LC-MS: calcd for C H F N O S [M+H] : 475.10, found
2
4
19
4
2
2
4
75.26.
ACS Paragon Plus Environment
29