Small-Molecule Activator of UNC-51-Like Kinase 1 (ULK1) That Induces Cytoprotective Autophagy for Parkinson's Disease Treatment

Article abstract of DOI:10.1021/acs.jmedchem.7b01575

UNC-51-like kinase 1 (ULK1), the yeast Atg1 ortholog, is the sole serine-threonine kinase and initiating enzyme in autophagy, which may be regarded as a target in Parkinson's disease (PD). Herein, we discovered a small molecule 33i (BL-918) as a potent activator of ULK1 by structure-based drug design. Subsequently, some key amino acid residues (Arg18, Lys50, Asn86, and Tyr89) were found to be crucial to the binding pocket between ULK1 and 33i by site-directed mutagenesis. Moreover, we found that 33i induced autophagy via the ULK complex in SH-SY5Y cells. Intriguingly, this activator displayed a cytoprotective effect on MPP⁺-treated SH-SY5Y cells, as well as protected against MPTP-induced motor dysfunction and loss of dopaminergic neurons by targeting ULK1-modulated autophagy in mouse models of PD. Together, these results demonstrate the therapeutic potential to target ULK1, and 33i, the novel activator of ULK1, may serve as a candidate drug for future PD treatment.

Full text of DOI:10.1021/acs.jmedchem.7b01575

Subscriber access provided by Queen Mary, University of London
A small-molecule activator of UNC-51-like kinase 1 (ULK1) that
induces cytoprotective autophagy for Parkinson's disease treatment
Liang Ouyang, Lan Zhang, Shouyue Zhang, Dahong Yao,
Yuqian Zhao, Guan Wang, Leilei Fu, Peng Lei, and Bo Liu
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01575 • Publication Date (Web): 21 Mar 2018
Downloaded from http://pubs.acs.org on March 21, 2018
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 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A small-molecule activator of UNC-51-like kinase 1 (ULK1)
that induces cytoprotective autophagy for Parkinson's
disease treatment
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
#
#
Liang Ouyang , Lan Zhang , Shouyue Zhang, Dahong Yao, Yuqian Zhao, Guan Wang,
Leilei Fu, Peng Lei, Bo Liu*
State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan
University, and Collaborative Innovation Center for Biotherapy, Chengdu 610041, China
ABSTRACT
UNCꢀ51ꢀlike kinase 1 (ULK1), the yeast Atg1 ortholog, is the sole serineꢀthreonine
kinase and initiating enzyme in autophagy, which may be regarded as a target in
Parkinson's disease (PD). Herein, we discovered a small molecule 33i (BLꢀ918) as a
potent activator of ULK1 by structureꢀbased drug design. Subsequently, some key amino
acid residues (Arg18, Lys50, Asn86 and Tyr89) were found to be crucial to the binding
pocket between ULK1 and 33i by siteꢀdirected mutagenesis. Moreover, we found that 33i
induced autophagy via the ULK complex in SHꢀSY5Y cells. Intriguingly, this activator
+
displayed a cytoprotective effect on MPP ꢀtreated SHꢀSY5Y cells, as well as protected
against MPTPꢀinduced motor dysfunction and loss of dopaminergic neurons by targeting
ULK1ꢀmodulated autophagy in mouse models of PD. Together, these results demonstrate
the therapeutic potential to target ULK1, and 33i, the novel activator of ULK1 may serve
as a candidate drug for future PD treatment.
Keywords: UNCꢀ51ꢀlike kinase 1; Autophagy; Parkinson's disease; ULK1 activator; The
ULK complex
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 50
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
INTRODUCTION
Autophagy is an evolutionarily conserved and multiꢀstep lysosomal degradation
1
,2
process that may degrade some longꢀlived proteins or damaged organelles in cells. It is
wellꢀknown that autophagy can be modulated by a number of autophagyꢀrelated (Atg)
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
genes, especially UNCꢀ51ꢀlike kinase 1 (ULK1) and its complex. As the homolog of Atg1
in mammals, ULK1 has similar functions with Atg1 which was found as the first
5
autophagyꢀrelated gene in yeast. And, the ULK complex formed by ULK1, mAtg13,
6,7
FIP200 and Atg101, is required to initiate the autophagic process.
Interestingly,
AMPꢀActivated Protein Kinase (AMPK) can activate ULK1 via directly phosphorylating
ULK1 or relieving the mammalian target of rapamycin complex 1 (mTORC1) negative
8
,9
regulation of ULK1 activity. Subsequently, the downstream signaling pathways can be
1
0,11
regulated by the activation of ULK1 and its complex.
As autophagy is a one of key mechanisms to maintain the nutrient and energy
homeostasis of cells, its dysregulation may impede the clearance of abnormal proteins
and damaged organelles, which is identified as one of pathogenesis of neurodegenerative
12,13
diseases, such as Parkinson’s disease (PD).
PD is often characterized by the
accumulation of αꢀsynuclein that can be visible as Lewy's body inclusions and by loss of
14
nigrostriatal dopaminergic neurons. Autophagy may promote the removal of such
1
5ꢀ17
aggregated proteins for protecting neuron cells against the toxicity,
which would be
1
8,19
regarded as an attractive approach for the treatment of PD.
Further, applications of
autophagy enhancers have been found to alleviate dopaminergic neurodegeneration of
PD models in vitro and in vivo, indicating that autophagic modulators have potential
2
0ꢀ22
therapeutic effects on PD.
More recently, ULK1 has been reported to trigger
starvationꢀinduced cytoprotective autophagy in SHꢀSY5Y cells, thus serving as a potential
23
target for PD therapy. Therefore, we hypothesize that discovery of a new activator of
ULK1 to regulate cytoprotective autophagy would be a promising avenue to treat PD.
Thus, in this study, we discovered a novel ULK1 activator 33i (BLꢀ918) that potently
activated ULK1. 33i could induce cytoprotective autophagy via the ULK1 complex in
SHꢀSY5Y cells, and also exerted its neuroprotective effects by targeting ULK1ꢀmodulated
2
ACS Paragon Plus Environment

Page 3 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
autophagy in a MPTPꢀinduced PD mouse model. Together, these results demonstrate the
therapeutic potential to target ULK1 in PD, and 33i, as the activator of ULK1 may serve as
a candidate drug in PD therapy.
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
RESULTS
Structure-based ligand design for ULK1.
The ULK complex which consists of ULK1, mAtg13, FIP200 and Atg101, is able to
1
1
initiate the autophagic process. ULK1 increases the phosphorylation of mAtg13 and
interacts with FIP200 and Atg101 to form the ULK complex; thereby eventually triggering
24
autophagy. Given that the ULK complex is required to trigger autophagy by an activator
of ULK1, we designed a small molecule that could activate ULK1 and the rest of ULK
complex (Fig. 1A). A recent study has reported a kinase activatorꢀAMPK complex
25
structure (PDB code: 4ZHX ). Of note, ULK1 and AMPK are the members of the
serine/threonine kinase family with similar molecular structures; therefore, they could be
speculated to have similar kinase binding modes. Compared ULK1 kinase domain
26
(
Gly7ꢀAla280) (PDB code: 4WNP ) with the kinase activatorꢀAMPK crystal structure, we
found that they share a similar structure with a low RMSD value of 0.985 in kinase domain,
indicating that ULK1 may possess a putative activation site in the corresponding region
(Glu9ꢀArg18, Ile48ꢀLeu53 and Gln82ꢀTyr89) which is far away from the known inhibitor
26,27
binding site (ATPꢀbinding site) (Fig. S1 and Fig. 1B).
In addition, we applied solvent
accessible surface (SAS) calculation to further identify the hot spots (Gly7ꢀLys55,
Asp80ꢀLeu90) that were druggable contact surfaces covering the putative activator
28
binding site (Fig. 1B). Based upon highꢀthroughput screening from ZINC database, we
achieved the top 50 potential smallꢀmolecule compounds according to their scores and
energies. Subsequently, considering the diversity of their chemical structures, we
reselected the top 20 hits (docked compounds). (Fig. 1C and Fig. S2). Combined with
kinase assay of ULK1, we found 15 of the topꢀranked 20 hits enhanced the ULK1 kinase
activity at different levels. And, we selected compound 3 as the best leading compound
(
Fig. 1D). According to the interactions between ULK1 and 3, four key amino acid
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 50
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
residues (Asn86, Met84, Lys50 and Tyr89) were identified, which may provide a clue to
optimize 3 for a better ULK1 activator. (Fig. 1E).
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. 1 Structure-based ligand design for ULK1. (A) The schematic model of autophagy initiation
process regulated by the ULK complex activation. (B) The kinase domain structure of ULK1 (PDB code:
4WNP) was colored in grey. Hot spots in activator binding sites were colored in blue and potential ULK1
activator binding sites were colored in red. (C) Docking the topꢀranked 20 candidate compounds. (D) The
topꢀranked 20 candidate compounds (1 ꢁM) were screened for kinase activity by ULK1 kinase assay.
Each point showed luminescence normalized to basal level at 10 ng ULK1. (E) A detailed view showed
the binding conformation between ULK1 and 3. Four key amino acid residues (Asn86, Met84, Lys50 and
Tyr89) were surrounded an activator binding pocket with 3.
Structural optimization and discovery of the ULK1 activator
Next, we drove structural optimization of the lead compound rationally (Fig. 2A). The
bridging oxygen atom was replaced by ester group to form an additional hydrogen bond
with Tyr89, and the imidazole ring was transformed into piperazine ring to yield
compounds 24aꢀs (Fig. 2B and Scheme 1). Based upon ULK1 kinase activity, 24c bear
the best maximum efficacy (Emax) with half maximal effective concentration (EC ) of
50
5
4.84 nM and, resulting in a 4.8ꢀfold (EC ) and 2.1ꢀfold (Emax) improvement compared to
50
3
(Table 1 and 4). According to the interactions between 24c and ULK1, an additional
4
ACS Paragon Plus Environment

Page 5 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PiꢀCation interaction and two hydrogen bonds were generated between 24c and Arg18,
Lys50 and Tyr89. Subsequently, the Dꢀ(ꢀ)ꢀ2ꢀPhenylglycine portion of the scaffold offered a
straightforward entry for the generation from phenylamine to obtain compounds 29a-t (Fig.
2
B and Scheme 2). Compared with 24c, 29n possessed a slight improvement of EC50 and
max in ULK1 kinase activity, but it showed 2ꢀfold improvement in autophagy activity
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
E
measured by flow cytometry analysis of monodansylcadaverine (MDC) staining (Fig. S3
and S4, Table 2 and 4). The residues located on the edge of active pocket were not fully
utilized, such as Asn86 and Tyr89. Thus, we synthesized some compounds 32a-i, 33a-i,
34a-i, 35a-i and 36a-i by introducing a urea group as the linker (Fig. 2B and Scheme 3).
The most potent compound 33i had 1.7ꢀfold (EC50) and 1.35ꢀfold (Emax) improvement in
ULK1 kinase activity, as well as a 1.5ꢀfold improvement in autophagy activity over 29n,
respectively (Fig. S3, 4 and 5, Table 3 and 4). Additionally, AMPK and eEF2K could not
be activated by 33i in the kinase assays (Fig. S6), indicating 33i has a priority in activating
ULK1. Moreover, 33i maintained all the critical interactions observed in compounds 24c
and 29n. Three additional halogen bonds were observed by the trifluoromethyl moiety
with Ala85 and Asn86, which reinforced this binding conformation, and the carbonyl of
scaffold and phenolic hydroxyl of Tyr89 formed a hydrogen bond as expected (Fig. 3A).
As mentioned above, the crucial residues Arg18 and Lys50 are required for the high
affinity of 33i. In addition, the residues Ala85, Asn86 and Tyr89, can improve its potency
and selectivity.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 50
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
6
ACS Paragon Plus Environment

Page 7 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Boc
Boc
R2
HN
26
HN
H
NH2
O
HN
H
(
a)
N
(b)
N
COOH
R1
R1 (c)
X
NH
H
O
N
R1
27a,b;30aꢀc
2
8a,b;31aꢀc
O
3
2aꢀi;33aꢀi;34aꢀi;35aꢀi;36aꢀi
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
2a: X=O, R =(S)ꢀ1ꢀphenylethly, R =4ꢀmethoxyphenyl
32b: X=O, R =(S)ꢀ1ꢀphenylethly, R =2ꢀchlorophenyl
2c: X=O, R =(S)ꢀ1ꢀphenylethly, R =3ꢀchlorophenyl
2d: X=O, R =(S)ꢀ1ꢀphenylethly, R =4ꢀmethylphenyl
2e: X=O, R =(S)ꢀ1ꢀphenylethly, R =2,6ꢀdimethylphenyl
32f: X=O, R =(S)ꢀ1ꢀphenylethly, R =4ꢀtrifluoromethylphenyl
2g: X=O, R =(S)ꢀ1ꢀphenylethly, R =1ꢀnaphthyl
2h: X=S, R =(S)ꢀ1ꢀphenylethly, R =phenyl
32i: X=S, R =(S)ꢀ1ꢀphenylethly, R =3,5ꢀditrifluoromethylphenyl 35e: X=O, R =3ꢀbromophenyl, R =2,6ꢀdimethylphenyl
3a: X=O, R =2,4ꢀdifluorophenyl, R =4ꢀmethoxyphenyl
3b: X=O, R =2,4ꢀdifluorophenyl, R =2ꢀchlorophenyl
3c: X=O, R =2,4ꢀdifluorophenyl, R =3ꢀchlorophenyl
33d: X=O, R =2,4ꢀdifluorophenyl, R =4ꢀmethylphenyl
3e: X=O, R =2,4ꢀdifluorophenyl, R =2,6ꢀdimethylphenyl
3f: X=O, R =2,4ꢀdifluorophenyl, R =4ꢀtrifluoromethylphenyl
3g: X=O, R =2,4ꢀdifluorophenyl, R =1ꢀnaphthyl
33h: X=S, R =2,4ꢀdifluorophenyl, R =phenyl
3i: X=S, R =2,4ꢀdifluorophenyl, R =3,5ꢀditrifluoromethylphenyl 36e: X=O, R
4a: X=O, R =2ꢀmethylphenyl, R =4ꢀmethoxyphenyl
4b:X=O, R =2ꢀmethylphenyl, R =2ꢀchlorophenyl
34c: X=O, R =2ꢀmethylphenyl, R =3ꢀchlorophenyl
4d: X=O, R =2ꢀmethylphenyl, R =4ꢀmethylphenyl
4e: X=O, R =2ꢀmethylphenyl, R =2,6ꢀdimethylphenyl
34f: X=O, R =2ꢀmethylphenyl, R =4ꢀtrifluoromethylphenyl
1 2
1
2
34g: X=O, R
34h: X=S, R
34i: X=S, R
35a: X=O, R
35b: X=O, R
35c: X=O, R
35d: X=O, R
=2ꢀmethylphenyl, R
=2ꢀmethylphenyl, R
=2ꢀmethylphenyl, R
=3ꢀbromophenyl, R
=3ꢀbromophenyl, R
=3ꢀbromophenyl, R
2
=3ꢀbromophenyl, R =4ꢀmethylphenyl
=1ꢀnaphthyl
=phenyl
=3,5ꢀditrifluoromethylphenyl
2
=4ꢀmethoxyphenyl
1
2
1
2
3
3
3
1
2
1
2
1
2
1
2
1
2
1
=2ꢀchlorophenyl
2
=3ꢀchlorophenyl
1
2
1
2
3
3
1
1
2
1
2
1
1
2
1
2
3
3
3
35f: X=O, R =3ꢀbromophenyl, R =4ꢀtrifluoromethylphenyl
1 2
1
2
35g: X=O, R
35h: X=S, R
35i: X=S, R
36a: X=O, R
36b: X=O, R
=3ꢀbromophenyl, R
=3ꢀbromophenyl, R
=3ꢀbromophenyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=3,4ꢀdimethyloxyphenylethyl, R
=1ꢀnaphthyl
=phenyl
=3,5ꢀditrifluoromethylphenyl
=4ꢀmethoxyphenyl
1
2
1
2
1
2
1
2
1
2
1
2
3
3
3
1
2
1
2
=2ꢀchlorophenyl
2
=3ꢀchlorophenyl
1
2
1
2
36c: X=O, R
36d: X=O, R
1
2
1
=4ꢀmethylphenyl
1
2
1
2
3
1
2
=2,6ꢀdimethylphenyl
1
2
3
3
36f: X=O, R
36g: X=O, R
36h: X=S, R
36i: X=S, R
1
2
=4ꢀtrifluoromethylphenyl
2
=1ꢀnaphthyl
1
2
1
2
1
=phenyl
1
2
1
2
3
3
1
2
=3,5ꢀditrifluoromethylphenyl
1
2
1
2
Scheme 3 Reagents and conditions: (a) NꢀMethylmorpholine, isoꢀbutylchloroformate, THF, ꢀ20
℃
1 2
, R NH ; (b)
HCl/MeOH, r.t.; (c) CH Cl , TEA, R NCO or R NCS, r.t.
2
2
2
2
Fig. 2 Structural optimization and discovery of the ULK1 activator. (A) Biological evaluation guided
optimization towards 33i. (B) ULK1ꢀ33i interactions included hydrogen bonds (green dash), PiꢀSulfur
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 50
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
(
(
yellow dash), Halogen interaction (blue dash), PiꢀCation (orange dash) and Carbon Hydrogen bond
light blue). Insets showed selected data from the structureꢀactivity relationship.
Table 1. Kinase activities of compounds 24a-s against recombinant human ULK1 and
relevant MDC positive ratio in SHꢀSY5Y cells
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
Kinase
MDC positive
Compound
R1
R2
activity%(100
b
ratio%(1 ꢁM)
a
nM)
2
4a
4b
155.76±4.09
148.68±4.28
9.59±0.91
6.56±1.25
2
2
4c
4d
197.06±6.55
180.73±3.31
14.72±0.80
12.49±0.59
2
2
4e
150.60±7.49
141.88±3.86
10.04±0.44
7.41±0.44
2
4f
2
2
4g
4h
138.59±4.07
143.06±4.56
6.22±0.57
9.99±0.39
8
ACS Paragon Plus Environment

Page 9 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
2
4i
4j
158.91±4.50
151.86±7.45
140.89±6.17
11.42±0.67
11.80±0.65
8.56±0.92
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
4k
2
4l
162.89±8.39
138.92±8.63
162.08±11.02
13.42±0.59
4.51±0.84
11.58±1.15
2
4m
CH3
2
2
4n
4o
170.86±7.79
167.93±4.95
14.17±0.28
11.05±0.57
24p
2
4q
150.37±6.81
154.81±6.58
147.00±5.78
11.76±0.52
11.21±0.40
9.20±0.40
2
4r
24s
a
Each compound was determined by three independent experiments (values are the mean ±
SEM).
b
Determined by flow cytometry analysis using MDC staining (values are the mean ± SEM).
Table 2. Kinase activities of compounds 29a-t against recombinant human ULK1 and
relevant MDC positive ratio in SHꢀSY5Y cells
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 50
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
Kinase
MDC positive
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
R1
R2
activity%(100
b
ratio%(1 ꢁM)
a
nM)
2
9a
137.05±5.74
142.75±6.47
10.96±0.81
13.52±0.88
2
9b
29c
146.10±5.67
139.01±5.39
15.30±0.58
13.80±0.60
2
9d
2
9e
142.11±5.59
139.88±8.68
147.13±8.17
16.61±1.01
16.21±0.61
18.13±0.78
29f
2
9g
10
ACS Paragon Plus Environment

Page 11 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
9h
144.09±5.26
158.24±6.19
163.77±8.51
152.80±6.20
16.75±0.40
21.04±0.83
24.47±0.56
20.31±1.08
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
9i
9j
2
2
9k
2
9l
161.03±9.18
148.19±4.23
23.44±1.24
20.55±0.59
2
9m
2
2
9n
9o
201.73±9.12
171.04±6.13
30.13±1.48
28.24±1.47
2
9p
149.18±4.44
20.25±1.23
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 50
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
9q
150.69±5.32
22.42±1.21
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
9r
198.31±7.32
167.90±6.56
174.10±12.15
32.10±1.72
26.18±1.60
31.36±1.48
29s
2
9t
a
Each compound was determined by three independent experiments (values are the mean ±
SEM).
b
Determined by flow cytometry analysis using MDC staining (values are the mean ± SEM).
Table 3. Kinase activities of compounds 32a-i, 33a-i, 34a-i, 35a-i, 36a-i against
recombinant human ULK1 and relevant MDC positive ratio in SHꢀSY5Y cells
Kinase
MDC positive
Compound
X
R
1
R
2
a
b
activity%(100 nM)
ratio%(1 ꢁM)
3
2a
2b
O
145.21±9.96
158.14±8.91
19.19±0.95
19.29±0.95
3
O
12
ACS Paragon Plus Environment

Page 13 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3
3
2c
2d
2e
O
O
O
O
O
S
169.07±9.96
22.70±1.10
16.56±1.17
16.24±0.82
38.34±1.61
19.61±0.77
32.25±2.37
32.92±2.07
CH3
148.05±11.74
144.21±11.2
209.33±10.17
149.21±8.35
169.03±4.54
177.85±7.24
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
2f
3
3
2g
2h
3
2i
S
3
3a
3b
O
O
152.79±6.18
143.91±2.75
30.95±0.39
19.65±0.60
3
3
3c
3d
O
O
159.01±2.61
152.79±2.84
24.72±0.93
24.02±0.86
CH3
3
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 50
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
3
3e
O
O
O
S
S
147.09±1.05
24.89±1.67
43.61±1.11
22.26±0.84
39.44±1.35
43.90±1.73
3
3f
226.95±2.17
151.17±5.10
196.87±4.77
243.21±4.45
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
3
3g
3h
33i
3
4a
O
O
157.78±3.62
30.35±0.93
30.35±0.93
36.75±1.03
34b
34c
34d
34e
O
O
O
O
O
156.17±5.41
174.42±5.64
149.37±4.52
182.68±3.04
145.81±5.19
27.35±0.93
34.88±0.86
24.25±0.80
38.38±0.59
24.40±1.13
CH3
34f
3
4g
14
ACS Paragon Plus Environment

Page 15 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
4h
S
S
179.30±4.30
36.21±0.94
43.09±1.17
3
4i
5a
5b
5c
204.95±5.78
157.00±5.34
160.81±6.31
151.86±5.58
147.06±4.86
154.80±3.50
179.01±2.35
150.33±3.40
175.71±3.59
217.27±5.26
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
3
3
O
O
O
O
O
O
O
S
28.61±1.01
26.34±0.86
24.11±1.10
25.43±0.67
27.12±0.63
33.57±2.74
24.63±0.54
32.93±1.39
39.63±1.03
CH3
35d
35e
3
5f
3
3
5g
5h
3
5i
S
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 50
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
3
3
3
3
3
6a
6b
6c
6d
6e
O
O
O
O
O
O
140.82±5.51
21.05±0.30
27.89±0.67
31.09±0.96
24.31±1.28
23.21±1.33
33.48±1.20
152.12±4.46
167.93±7.71
148.01±6.56
143.99±4.30
162.67±4.14
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
CH3
3
6f
3
3
6g
6h
O
S
S
139.48±4.13
170.26±1.82
184.33±7.33
20.95±0.75
36.66±1.11
38.95±0.81
3
6i
a
Each compound was determined by three independent experiments (values are the mean ±
SEM).
b
Determined by flow cytometry analysis using MDC staining (values are the mean ± SEM).
Table 4. Emax values and EC50 values for representative compounds
a
max
Compound
E
EC50 (nM)
263.1
3
0.324 ± 0.024
16
ACS Paragon Plus Environment

Page 17 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
4c
9n
33i
0.692 ± 0.032
0.742 ± 0.028
1.000 ± 0.037
54.84
40.80
24.14
2
a
All Emax values were normalized to the maximal response of 33i.
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
Identification of the key amino acids between 33i and ULK1
To determine the key amino acids in activator binding site between 33i and ULK1, we
constructed several mutants of ULK1 using siteꢀdirected mutagenesis. The amino acid
K50A
R18A
residues of Lys50, Arg18, Asn86 and Tyr89 were mutated to alanine (ULK1
, ULK1
,
N86A
Y89A
K50A
N86A
ULK1
ULK1
ULK1
). ULK1 kinase activity assay indicated that ULK1
, ULK1
and
Y89A
WT
mutants induced substantial decrease of kinase activity compared to ULK1
R18A
after treatment with 33i, while ULK1
mutant only induced relatively little reduction of
kinase activity (Fig. 3B and Fig. S7). Subsequently, we applied in vitro kinase assay to
demonstrate that 33i could enhance the phosphorylation level of mAtg13 in HEKꢀ293T
WT
29,30
cells transfected with ULK1 , indicating 33i activates ULK1 in vitro.
And, we found
R18A
that ULK1
mutant did not result in apparent reduction on phosphorylation of mAtg13 in
the presence of 33i, but the other three mutants induced a significant decrease of mAtg13
N86A
Y89A
phosphorylation, especially for ULK1
and ULK1
mutants (Fig. 3C). Intriguingly, we
directly coꢀtransfected Flagꢀtagged ULK1 (WT or mutant ULK1) and GSTꢀtagged mAtg13
into HEKꢀ293T cells, and found the similar results with the in vitro kinase assay indicated
by the phosphorylation levels of pꢀmAtg13 (Fig. 3D). Moreover, in an analysis of Surface
Plasmon Resonance (SPR), we found that 33i bound to ULK1 with a high binding affinity
K50A
N86A
Y89A
(
K
D
= 0.719 ꢁM), but the ULK1
, ULK1
and ULK1
mutants lead to obvious
WT
decrease of binding affinity than ULK1
at different levels (Fig. S8). Notably, these
biochemical experiments suggested that Lys50, Asn86 and Tyr89 may be more crucial to
the activator binding site of ULK1. More importantly,they were in consistent with the
aforementioned key amino acid residues to the activator binding pocket with 3. Together,
these results suggest that Lys50, Asn86 and Tyr89 are the key amino acids between
ULK1 and its activator.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 50
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
Fig. 3 Identification of the key amino acids between ULK1 and 33i. (A) A docking pose based upon
the kinase domain structure of ULK1 (PDB code: 4WNP) showed the interaction between 33i and ULK1.
The key amino acid residues are marked red. (B) Lys50, Arg18, Asn86 and Tyr89 in the kinase domain of
ULK1 were mutated to alanine by siteꢀdirected mutagenesis. And, the recombinant proteins were purified
WT
from insect cells expressing ULK1 and ULK1 mutants. Then, ULK1 kinase activity was measured by
using kinase assay and purified ULK1 mutant proteins in the presence of different concentrations of 33i.
WT
The relative kinase activity of ULK1 mutants were normalized to the maximal response of ULK1 . (C)
WT
Flagꢀtagged ULK1 and ULK1 mutants were expressed in HEKꢀ293T cells and immunoprecipitated by
antiꢀFlag antibody, then incubated with GSTꢀtagged mAtg13 in a kinase reaction buffer in the presence
or absence of 33i. The reaction was stopped and analyzed by western blot with pꢀmAtg13 antibody. (D)
HEKꢀ293T cells was coꢀtransfected with Flagꢀtagged ULK1 (WT or mutant ULK1) and GSTꢀtagged
mAtg13, then treated with or without 33i. The phosphorylation of mAtg13 in total cell lysates were
18
ACS Paragon Plus Environment

Page 19 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
detected with pꢀmAtg13 antibody.
3
3i induces autophagy in neuron-like cells
To explore the mechanisms of 33iꢀinduced autophagy, we observed the cellular
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
ultrastructure by the electron microscopy. 33i treatment induced some vacuolar elements
that were most likely to be of autophagic origin in SHꢀSY5Y cells (Fig. 4A and 4B). And,
we found that 33i timeꢀdependently elevated the expression levels of LC3ꢀII (a key marker
of autophagy), Beclinꢀ1 and its phosphorylation status, whereas the level of the selective
autophagy substrate SQSTM1/p62 was reduced after treatment with 33i (Fig. 4C).
Moreover, LC3 and SQSTM1/p62 were significantly accumulated after coꢀtreatment with
3
3i and Bafilomycin A1, indicating that 33i treatment enhances the autophagic flux (Fig.
D). To further demonstrate that whether 33i could induce autophagic flux in other
4
neuronꢀlike cells, we treated highly differentiated PCꢀ12 cell with 33i and Bafilomycin A1.
Interestingly, 33i treatment induced autophagosome accumulation in PCꢀ12 cells, which
was indicated by increased expression levels of LC3ꢀII and SQSTM1/p62, as well as the
aggregated LC3 puncta in PCꢀ12 cells (Fig. 4E and 4F). These results show that 33i can
induce autophagy in both undifferentiated and differentiated neuronꢀlike cells. Next, we
examined whether 3ꢀmethyladenine (3ꢀMA), a class III PI3K autophagy inhibitor, could
block 33iꢀinduced autophagy. MDC staining analysis were conducted by 33iꢀtreated
SHꢀSY5Y cells within or without 3ꢀMA. We observed an increase in green dots with
fluorescence in 33iꢀtreated cells; these green dots were significantly reduced by the 3ꢀMA
treatment (Fig. 4G). Meanwhile, 33i treatment led to the increase of the GFPꢀLC3 puncta
in SHꢀSY5Y cells, which was markedly decreased under the treatment of 3ꢀMA (Fig. 4H).
In addition, 3ꢀMA was found to restrain the transformation of LC3ꢀI to LC3ꢀII, as well as the
degradation of SQSTM1/p62, indicating that 3ꢀMA can block 33iꢀinduced autophagy (Fig.
4
I).
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 50
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
Fig. 4 33i induces autophagy in neuron-like cells. (A) SHꢀSY5Y cells were treated with or without 5
µM 33i for 24 h. The ultrastructure was examined via transmission electron microscopy (TEM). Arrows
indicate autophagic bodies. (B) The numbers of cells with autophagosome and numbers of autophagic
***
vacuoles per cell were analyzed from at least 30 randomly chosen fields in the TEM analysis,
p<0.001,
compared to control. (C) SHꢀSY5Y cells were treated with 5 µM 33i for 0, 6, 12, 24 and 36 h. The
expression levels of LC3, pꢀBeclinꢀ1, Beclinꢀ1 and SQSTM1/p62 were examined by western blot. (D)
SHꢀSY5Y cells were treated with 5 µM 33i for 24 h with or without 10 nM Bafilomycin A1. The levels of
LC3 and SQSTM1/p62 were examined by western blot. (E) PCꢀ12 cells were treated with 5 µM 33i for 24
h with or without 10 nM Bafilomycin A1. The expression levels of LC3 and SQSTM1/p62 were examined
by western blot. (F) PCꢀ12 cells were transfected with GFPꢀmRFPꢀLC3 plasmid, followed by treatment
with 5 µM 33i for 24 h with or without 10 nM Bafilomycin A1. Then, the GFPꢀLC3 puncta were observed
by fluorescence microscopy. Scale bar = 20 ꢁm. (G) SHꢀSY5Y cells were treated with 5 µM 33i for 24 h
with or without 2 mM 3ꢀMA, then treated with MDC staining and observed using a fluorescence
microscopy. The MDC positive ratio is shown in the graphs. Scale bar, 20 ꢁm. (H) SHꢀSY5Y cells were
transfected with a GFPꢀLC3 plasmid, followed by treatment with 5 µM 33i for 24 h with or without 2 mM
3ꢀMA. GFPꢀLC3 puncta were observed using a fluorescence microscopy. The number of LC3 puncta
20
ACS Paragon Plus Environment

Page 21 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
normalized to cell number is shown. Control: n = 98 cells, 33i: n = 93 cells, 33i+3ꢀMA: n = 96 cells. Scale
bar, 20 ꢁm. (I) SHꢀSY5Y cells were treated with 5 µM 33i for 24 h with or without 2 mM 3ꢀMA. The levels
of LC3 and SQSTM1/p62 were examined by western blot, βꢀactin was used as a loading control.
3
3i induces autophagy by triggering the ULK complex
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
As the ULK complex is highly correlated to autophagosome formation, we examined
the expression levels of ULK1, mAtg13, Atg101 and FIP200, but the total protein levels
have no change after 33i treatment. Notably, 33i treatment elevated ser317 and ser555
phosphorylation of ULK1, as well as decreased ser757 phosphorylation of ULK1. In
addition, the phosphorylation level of mAtg13 was also increased after activation of ULK1.
mTOR, known as an upstream negative regulator of the ULK complex, was
dephosphorylated following 33i treatment (Fig. 5A). Moreover, the result of
coꢀimmunoprecipitation showed that the binding activity of the ULK complex was
markedly increased after treatment with 33i (Fig. 5B), indicating the specificity of the
interactions amongst ULK1, mAtg13, Atg101 and FIP200. To further explore the
association between ULK1 and autophagic activation in 33iꢀtreated SHꢀSY5Y cells, we
knocked down ULK1 by ULK1ꢀsiRNA. We observed that MDC positive staining and
GFPꢀLC3 dots were markedly decreased in ULK1ꢀsiRNA treated cells (Fig. 5C and 5D).
To confirm whether the absence of ULK1 could cause autophagic inhibition in the ULK
complex, we found that the repression of ULK1 suppressed the formation of the ULK
complex, as determined by upregulation of FIP200, mAtg13, Atg101 and downregulation
of pꢀmAtg13 (Fig. 5E). Moreover, the expression levels of pꢀBeclinꢀ1, Beclinꢀ1 and LC3ꢀII
were decreased and that of SQSTM1/p62 was increased (Fig. 5E). Because ULK2 is
31
highly homologous to ULK1, we knocked down both ULK1 and ULK2 simultaneously to
examine whether the two proteins show a functional redundancy. Interestingly, we found
that knockdown of ULK1/2 also induced inhibition of phosphorylation of mAtg13 and
Beclinꢀ1, as well as SQSTM1/p62 degradation and transformation of LC3ꢀI to LC3ꢀII (Fig.
S9). Thus, it is established that 33i induces cytoprotective autophagy via the ULK
complex.
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 50
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
Fig. 5 33i induces autophagy via the ULK complex. (A) SHꢀSY5Y cells were treated with 5 µM 33i for
the indicated amount of time. The levels of ULK1, pꢀULK1, mAtg13, pꢀmAtg13, FIP200, Atg101, mTOR
and pꢀmTOR were examined by western blot. (B) SHꢀSY5Y cells were treated with 33i for 6h. Then, the
interactions among ULK1 and mAtg13, Atg101, FIP200 were determined by coꢀimmunoprecipitation. (C)
SHꢀSY5Y cells were transfected with siRNAꢀNC or siRNAꢀULK1, followed by treatment with 5 µM 33i for
24 h and then stained with MDC and observed using a fluorescence microscopy. The MDC positive ratio
is shown in the graphs. Scale bar, 20 ꢁm. (D) SHꢀSY5Y cells were coꢀtransfected with a GFPꢀLC3
plasmid and siRNAꢀNC or siRNAꢀULK1, followed by treatment with 5 µM 33i for 24 h. Then, the
GFPꢀLC3 puncta were observed using a fluorescence microscopy. The number of LC3 puncta
normalized to cell number is shown. Control: n = 101 cells, 33i: n = 97 cells, 33i+siꢀNC: n = 103 cells,
33i+siꢀULK1: n = 94 cells. Scale bar, 20 ꢁm. (E) SHꢀSY5Y cells were transfected with siRNAꢀNC or
siRNAꢀULK1, followed by treatment with or without 5 µM 33i for 24 h. The levels of ULK1, mAtg13,
pꢀmAtg13, LC3, pꢀBeclinꢀ1, Beclinꢀ1 and SQSTM1/p62 were examined by western blot, βꢀactin was used
as a loading control.
22
ACS Paragon Plus Environment

Page 23 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
3i has a therapeutic potential on PD models in vitro and in vivo
To evaluate whether 33iꢀinduced autophagy could have a cytoprotective efficacy, we
+
added 1 mM MPP to SHꢀSY5Y cells within or without 33i, and the inhibitory ratio 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
+
measured. 33i could partially reverse MPP ꢀinduced cell death, which was determined by
enhancing cell viability (0.5 ꢁM: +42%, p=0.015; 5 ꢁM: +67%, p=0.004, Fig. 6A); however,
the restoration of 33i on cell viability was remarkably decreased after 3ꢀMA treatment (0.5
ꢁM: ꢀ24%, p=0.038; 5 ꢁM: ꢀ32%, p=0.009, Fig. 6A). Additionally, knockdown of ULK1
could reverse the cytoprotective effect of 33i (ꢀ30%, p=0.008, Fig. 6B). Thus, these results
indicate that 33i can induce autophagy, which has a cytoprotective effect on SHꢀSY5Y
+
cells. Since 33i induced cytoprotective autophagy in MPP ꢀtreated SHꢀSY5Y cells, we
next examined the therapeutic efficacy of 33i in a MPTPꢀinduced PD mouse model. To
assess the protective effect of 33i on MPTPꢀinduced motor dysfunction, we performed
some behavior tests on the MPTPꢀtreated mice, including pole test and swimming test.
And, the time to turn and time to finish for the MPTPꢀtreated mice were longer than that for
the vehicleꢀtreated mice (p<0.001), which are significantly restored in the medianꢀ
(40mg/kg: p<0.001,) and highꢀdose (80mg/kg: p<0.001) 33iꢀtreated mice, but barely
changed for the lowꢀdose (20mg/kg) 33iꢀtreated mice (Fig. 6C). Additionally, the
swimming tests (MPTP: p<0.001, compared to the control; 20mg/kg: p=0.039; 40mg/kg:
p<0.001; 80mg/kg: p<0.001; compared to MPTP) were similar with those of the pole test
(
(
Fig. 6C). We found that the levels of dopamine (DA), 3,4ꢀdihydroxyphenylacetic acid
DOPAC) and homovanillic acid (HVA) (DA: ꢀ82.2%, p=0.0093; DOPAC: −87.9%,
p=0.0046; HVA: −69.2%, p=0.0011) were obviously decreased in the Striatum (ST) of
MPTPꢀtreated mice, whereas 33i treatment attenuated the loss of DA and its metabolites
(DA: +499%, p=0.0183; DOPAC: +559%, p=0.0378; HVA: +211%, p=0.0334; Fig. 6D). To
determine whether 33i restrains the death of dopaminergic neuron cells in the Substantia
nigra (SN), we evaluated the expression of tyrosine hydroxylase (TH) that was visualized
by immunofluorescence. In addition, we demonstrated that MPTP could remarkably
lessen THꢀpositive neuron cells in SN, compared to control group (ꢀ63%, p<0.001, Fig. 6E
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 50
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
and 6F). By contrast, 33i treatment could reduce MPTPꢀinduced loss of THꢀpositive
neuron cells, compared to MPTP group (20mg/kg: +49%, p=0.003; 40mg/kg: +111%,
p<0.001; 80mg/kg: +142%, p<0.001; Fig. 6E and 6F). The expression of TH in ST was
also decreased after MPTP treatment, and restored by 33i (Fig. 6G). Moreover, we found
that the body weights of mice were not affected by during MPTP and 33i treatments (Fig.
S10A). And, no apparent toxicity in blood (Fig. S10B) or normal tissues (heart, thymus,
spleen, kidney, colon and ileum) were observed following 33i treatment (Fig. S10C).
Taken together, these results demonstrate that 33i has a good therapeutic potential on PD
models 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
+
Fig. 6 33i has a therapeutic potential on PD models in vitro and in vivo. (A) MPP (1 mM) was added
to SHꢀSY5Y cells with 0.5, 5, 50 µM 33i with or without 2 mM 3ꢀMA. Then, the cell viability was
24
ACS Paragon Plus Environment

Page 25 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
*
*
*
+
determined by an MTT assay, p<0.01, p<0.05, compared to MPP . (B) SHꢀSY5Y cells were transfected
with siRNAꢀNC or siRNAꢀULK1, followed by treatment with or without 5 µM 33i for 24 h in the presence of
+
**
MPP (1 mM). Then, the cell viability was determined by an MTT assay, p<0.01, compared to siꢀNC. (C)
The pole test and swimming test were performed to assess mouse motor dysfunction. The data were
#
##
***
*
expressed as the mean ± SEM (n=8).
p<0.001, compared to the control group; p<0.001, p<0.05,
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
compared to the MPTPꢀtreated group. (D) The levels of DA, DOPAC and HVA in the Striatum were
measured by highꢀperformance liquid chromatography (HPLC). The data were expressed as the mean ±
##
*
SEM (n=3). p<0.01, compared to the control group; p<0.05, compared to the MPTPꢀtreated group. (E)
The expression of TH in the Substantia nigra was visualized using immunofluorescence. Scale bar, 200
ꢁ
m. (F) Quantification of THꢀpositive cells in different groups. The data were expressed as the mean ±
###
***
**
SEM (n=5).
p<0.001, compared to the control group;
p<0.001, p<0.01, compared to the
MPTPꢀtreated group. (G) The protein level of TH in the Striatum was detected by western blot, βꢀactin
was used as a loading control.
3
3i induces cytoprotective autophagy by targeting ULK1 in vivo
To demonstrate whether 33i induces cytoprotective autophagy by targeting ULK1, we
examined the expression levels of LC3, Beclinꢀ1, SQSTM/p62 and pꢀULK1 in ST and SN,
respectively. Notably, the levels of pꢀBeclinꢀ1, Beclinꢀ1 and LC3ꢀII, as well as the
degradation of SQSTM1/p62 were enhanced in the medianꢀ and highꢀdose groups, rather
than the lowꢀdose group. The phosphorylation of ULK1 was increased in all 33iꢀtreated
groups, especially in the medianꢀ and highꢀdose groups (Fig. 7A and 7B). In addition,
MPTP treatment induced apoptosis in both ST and SN, as determined by caspaseꢀ3
activation and the decreasing ratio of Bclꢀ2/Bax; however, apoptosis was reversed in the
medianꢀdose group following 33i treatment (Fig. 7A and 7B). And, the expression of
pꢀULK1 in SN was increased, as determined by immunofluorescence (Fig. 7C and 7D).
Moreover, we found that 3ꢀMA partially reversed the neuroprotective effects induced by
3
3
8
3i via the immunofluorescence analysis of TH (ꢀ64%, p<0.001, Fig. 8A and 8B). And,
ꢀMA reversed 33iꢀinduced LC3ꢀII upregulation and SQSTM1/p62 degradation in ST (Fig.
C). Altogether, these results indicate that 33i exerts its cytoprotective autophagic effect
to prevent MPTPꢀinduced apoptosis, by targeting ULK1 in vivo.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 50
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
Fig. 7 33i induces cytoprotective autophagy by targeting ULK1 in vivo. The expression levels of Bax,
Bclꢀ2, Caspaseꢀ3, LC3, pꢀBeclinꢀ1, Beclinꢀ1, SQSTM1/p62, ULK1 and pꢀULK1 in the Striatum (A) and
Substantia nigra (B) were detected by western blot, βꢀactin was used as a loading control. (C) The
expression of pꢀULK1 in the Substantia nigra was visualized using immunofluorescence. Scale bar, 200
ꢁ
m. (D) Quantification of pꢀULK1 positive cells in different groups. The data were expressed as means ±
###
***
**
SEM (n=5).
p<0.001, compared to the control group; p<0.001, p<0.01, compared to the
MPTPꢀtreated group.
26
ACS Paragon Plus Environment

Page 27 of 50
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
Fig. 8 3-MA reverses a neuroprotective effect induced by 33i via inhibition of autophagy. (A) The
expression of TH in the Substantia nigra was visualized using immunofluorescence. Scale bar, 200 ꢁm.
(B) Quantification of THꢀpositive cells in different groups. The data were expressed as the mean ± SEM
###
***
(
n=5). p<0.001, compared to the control group. p<0.001, compared to the MPTPꢀtreated group. (C)
The expression levels of LC3 and SQSTM1/p62 in the Striatum were detected by western blot, βꢀactin
was used as a loading control.
DISCUSSION
23
Hitherto, ULK1 has been reported to be a biomarker to modulate autophagy in PD.
Intriguingly, a recent study has revealed that ULK1 expression is partial downregulated in
3
2
PD patients compared with the controls, suggesting that ULK1 may be a potential target
of PD. In our study, we discovered a novel ULK1 activator 33i that potently activated ULK1
to bind to some key amino acids such as Arg18, Lys50, Asn86 and Tyr89. Interestingly, we
have recently reported an antiꢀtumor activator of ULK1 named compound 37 (LYNꢀ1604)
33
that induced autophagyꢀassociated cell death and apoptosis in tripleꢀnegative breast
cancer (TNBC). Different from 37, we put forward a designing strategy to discover a totally
new ULK1 activator 33i which could regulate the ULK complex and thus eventually
triggering cytoprotective autophagy. Moreover, 33i has a distinctive chemical structure
from 37, which bears more interaction surfaces with ULK1, as well as displays a negligible
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 50
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
toxicity and remarkable cytoprotective effect on PD models. Thus, 33i has distinctive drug
design strategies, chemical structures and autophagic functions, which would be utilized
in the different diseases.
Currently, levodopa, the most common used drug in PD treatment can attenuate the
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
34,35
symptom of PD, but long term use still lead to motor fluctuation and other side effects.
In addition, cholinergic inhibitors and 5HT1A receptor agonists have demonstrated some
therapeutic effects on PD symptoms, but also be associated with adverse effects in
36
cognition. In avoid of such side effects, some smallꢀmolecule agents that enhance
autophagic activity, have been discovered to have therapeutic potential on PD. For
instance, Latrepirdine, a neuroactive compound, can reduce αꢀsynuclein accumulation in
3
7
mouse neurons by stimulating autophagy. Isorhynchophylline was found to activate
38
autophagy and expedite the degradation of aggregated αꢀsynuclein in neuron cells.
Trehalose is also an autophagic regulator to accelerate the clearance of αꢀsynuclein in a
39
PD model. Although these compounds can exert neuroprotective effects by modulating
autophagy; however, they have not any identified target and intricate mechanism.
Distinctive from them, we found that 33i could induce cytoprotective autophagy via the
ULK complex in SHꢀSY5Y cells, whereas silencing of ULK1 or blocking autophagy
induction led to decreased level of cytoprotective autophagy. Also, 33i exerted its
neuroprotective effects by targeting ULK1ꢀmodulated autophagy in a MPTPꢀinduced PD
mouse model, which is characterized by significant reduction in loss of THꢀpositive neuron
cells and MPTPꢀinduced apoptosis. Importantly, due to its low molecular weight, 33i may
transport well across the bloodꢀbrain barrier and enter into brain, thus making it to be a
leading compound for PD drug development.
CONCLUSIONS
In summary, ULK1, which is known as the autophagic initiator, has been recently
reported to be a potential therapeutics target in many diseases, such as PD. Based on
structureꢀbased drug design and highꢀthroughput screening, we eventually discover a
novel activator 33i, which may bind to ULK1 in some key amino acid residues (Arg18,
28
ACS Paragon Plus Environment

Page 29 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Lys50, Asn86 and Tyr89). Moreover, this ULK1 activator induces cytoprotective
autophagy by modulating the ULK complex (ULK1ꢀmAtg13ꢀFIP200ꢀAtg101) in
+
MPP ꢀtreated SHꢀSY5Y cells. More importantly, we demonstarte that this compound can
alleviate MPTPꢀinduced motor dysfunction and loss of dopaminergic neurons by targeting
ULK1ꢀinitiating autophagy in mouse models of PD. Taken together, our findings
demonstrate that this activator of ULK1 would be further exploited as a smallꢀmolecule
candidate drug for the future PD therapeutics.
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
EXPERIMENTAL SECTION
Screening potential ULK1 activator. The Xꢀray crystal structure of ULK1 kinase domain
(PDB code: 4WNP) was downloaded from the Protein Databank (PDB). Based upon the
identified activator binding site, we performed virtual screening of small molecule
compounds from ZINC library (http://zinc.docking.org/) by Accelrys Discovery Studio
4
0
(
version 3.5; Accelrys, SanDiego, CA, USA) using LibDock protocol.
Energy
4
1
minimization of each compound was performed by the CHARMm force field. The top 50
small molecule compounds were reꢀranked by semiꢀflexible docking approach using
42
CDOCKER protocol. The other parameters were set as default values.
ADP-Glo kinase activity assays. ULK1 and AMPK kinase activity assays were
performed using ADPꢀGlo Kinase Assay + ULK1 or AMPK Kinase Enzyme System
33, 43
according to our previous studies.
And EEF2K kinase activity assay was performed as
previously reported. In brief, the compound, substrate, ATP and kinase enzyme were all
diluted in a kinase buffer consisting of 50 µM DTT, 20 mM MgCl , 40 mM Tris (pH 7.5) and
.1 mg/ml BSA. Then, 2 µl of ULK1 kinase enzyme or purified wildꢀtype and mutant ULK1
K50A, R18A, N86A, Y89A) (10 ng), 2 µl of MBP (0.1 µg/µl)/ATP (10 µM) mix, 1 µl of
44
2
0
(
compound (5% DMSO) were added to the 384ꢀwell plates. Subsequently, 5 µl of ADPꢀGlo
reagent was added to per well after incubation at room temperature for 60 min. The plates
were continuously incubated at room temperature for 40 min, then 10 µl of kinase
detection reagent were added into each well. After incubation at room temperature for 30
min, the luminescence were recorded by microplate reader. The EC50 values were
29
ACS Paragon Plus Environment