Discovery of Novel Pyrazolo-pyridone DCN1 Inhibitors Controlling Cullin Neddylation

Article abstract of DOI:10.1021/acs.jmedchem.9b00410

Chemical control of cullin neddylation is attracting increased attention based largely on the successes of the NEDD8-activating enzyme (E1) inhibitor pevonedistat. Recently reported chemical probes enable selective and time-dependent inhibition of downstream members of the neddylation trienzymatic cascade including the co-E3, DCN1. In this work, we report the optimization of a novel class of small molecule inhibitors of the DCN1-UBE2M interaction. Rational X-ray co-structure enabled optimization afforded a 25-fold improvement in potency relative to the initial screening hit. The potency gains are largely attributed to additional hydrophobic interactions mimicking the N-terminal acetyl group that drives binding of UBE2M to DCN1. The compounds inhibit the protein-protein interaction, block NEDD8 transfer in biochemical assays, engage DCN1 in cells, and selectively reduce the steady-state neddylation of Cul1 and Cul3 in two squamous carcinoma cell lines harboring DCN1 amplification.

Full text of DOI:10.1021/acs.jmedchem.9b00410

Subscriber access provided by Macquarie University
Article
Discovery of Novel Pyrazolo-Pyridone DCN1
Inhibitors Controlling Cullin Neddylation
Ho Shin Kim, Jared T. Hammill, Daniel C. Scott, Yizhe Chen, Jaeki Min,
Jonah Rector, Bhuvanesh Singh, Brenda A. Schulman, and R. Kiplin Guy
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00410 • Publication Date (Web): 29 Aug 2019
Downloaded from pubs.acs.org on August 30, 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 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of Novel Pyrazolo-Pyridone DCN1 Inhibitors Controlling Cullin Neddylation
1
1
2
1
3
1
Ho Shin Kim ; Jared T. Hammill ; Daniel C. Scott ; Yizhe Chen ; Jaeki Min ; Jonah Rector ;
4
2,5
1
Bhuvanesh Singh ; Brenda A. Schulman ; R. Kiplin Guy *
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
1Department of Pharmaceutical Sciences, University of Kentucky, Lexington, Kentucky 40508,
United States
2Department of Structural Biology, St. Jude Children’s Research Hospital, Memphis,
Tennessee 38105, United States
3Department of Chemical Biology and Therapeutics, St. Jude Children's Research Hospital,
Memphis, Tennessee 38105, United States
4Department of Surgery, Laboratory of Epithelial Cancer Biology, Memorial Sloan Kettering
Cancer Center, New York, New York 10065, United States
5Department of Molecular Machines and Signaling, Max Planck Institute of Biochemistry,
Martinsried, Germany
*Correspondence: kip.guy@uky.edu
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 44
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
Chemical control of cullin neddylation is attracting increased attention based largely on the
successes of the NEDD8-activating enzyme (E1) inhibitor Pevonedistat. Recently reported
chemical probes enable selective and time-dependent inhibition of downstream members of
the neddylation tri-enzymatic cascade including the co-E3, DCN1. In this work, we report the
optimization of a novel class of small molecule inhibitors of the DCN1-UBE2M interaction.
Rational X-ray co-structure enabled optimization afforded a 25-fold improvement in potency
relative to the initial screening hit. The potency gains are largely attributed to additional
hydrophobic interactions mimicking the N-terminal acetyl group that drives binding of UBE2M
to DCN1. The compounds inhibit the protein-protein interaction, block NEDD8 transfer in
biochemical assays, engage DCN1 in cells, and selectively reduce the steady-state
neddylation of Cul1 and Cul3 in two squamous carcinoma cell lines harboring DCN1
amplification.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N-Acetyl Pocket
O O
NH
O O
NH
HN
N
Optimization
N
N
N
N
25 times
potency improvement
F
New scaffold HIT compound (1)
TR-FRET IC50 = 5.1 µM
New optimized compound (27)
TR-FRET IC50 = 0.2 µM
Introduction
NEDD8 (neural precursor cell expressed developmentally down-regulated protein 8) is a
ubiquitin-like protein (UBL) that is post-translationally appended to eukaryotic proteins in a
process termed neddylation. A three-step enzymatic cascade carries out neddylation. The first
step is the ATP-dependent formation of an E1-NEDD8 thioester intermediate. Next, the
NEDD8 is transferred by a transthioesterification reaction from the E1 to a Cys on the E2. E3s
then join the E2-NEDD8 complex and recruit specific target proteins. The E3 complex then
catalyzes the formation of an isopeptide bond connecting NEDD8’s C-terminus to the ɣ-amino
group of a lysine side-chain on the target protein (Figure 1).
2
ACS Paragon Plus Environment

Page 3 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. General Neddylation Cascade
The cullin family of ubiquitination E3s are the most well-characterized substrates of
neddylation. Upon neddylation, the cullins constellate the formation of the cullin-RING E3 UB
ligases (CRLs), a family of ligases that has approximately 300 members.1, 2 The CRLs regulate
diverse biological processes including cell cycle, signal transduction, DNA replication, and
viral modulation.3-6 CRL dysfunction is implicated in a number of human diseases, including
cancer.7-11 Drug discovery efforts targeting the CRLs and the associated proteasomal protein
degradation machinery have been extensive and continue to grow.9, 12-14 The neddylation
pathway has been successfully targeted by MLN4924 (Pevonedistat), an inhibitor of NEDD8’s
E1 enzyme, that completely blocks NEDD8 ligation to substrates. MLN4924 is currently being
15
tested in oncology clinical trials. An inhibitor of the COP9 signalosome, responsible for de-
16
neddylation of the CRLs, also displays anti-tumor activity. We have reported the discovery
of
inhibitors
of
DCN1
mediated
neddylation.17-19
Peptidomimetics
and
triazolo[1,5‑a]pyrimidine-based inhibitors of the DCN1-UBE2M interaction were also recently
reported by others.20-22 All the previously reported inhibitors potently and selectively inhibit the
DCN1/2-UBE2M protein interaction in biochemical assays, bind and thermally stabilize DCN1
in cells, and reduce the steady-state levels of cullin neddylation in a variety of cell lines
including HCC95, CAL33, KYSE70, H2170, SK-MES-1, and MGC-803 cells.17-22
DCN1 (Defective in Cullin Neddylation 1) is also known as DCUN1D1, DCNL1, or SCCRO
(Squamous Cell Carcinoma-related Oncogene); we use “DCN1” hereafter. DCN1 acts as a
co-E3 together with RBX1 to stimulate the transfer of NEDD8 from its E2 (UBE2M) to the cullin
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 44
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
23
proteins. Humans express five isoforms of DCNs. In vitro, the five DCNs can cooperate with
RBX1 to promote NEDD8 ligation from UBE2M or UBE2F to CULs 1-5, or with RBX2 to
24
promote NEDD8 ligation from UBE2F to CULs 1-5. In mammals, the genetic redundancy is
currently unclear although DCN1 is clearly not essential. Only DCN1 is broadly expressed
across all tissues. DCN1 and DCN2 share >80% sequence homology (100% conservation in
UBE2M binding pocket) and are the only isoforms expressed in both the cytosol and nucleus.
DCN3, 4, and 5 share less than 35% sequence homology with DCN1 and appear to have
more distinct tissue and cellular distributions.24-28 The high sequence homology, overlapping
expression patterns, and similar subcellular distributions suggest that DCN1 and 2 may be
redundant, while DCNs 3, 4, and 5 likely possess unique roles.
0
1
2
3
4
5
6
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
DCN1 is the most well characterized isoform due to its common amplification as part of a
large 3q26.3 amplicon in squamous cell carcinomas (SCC) and other tumors.29 DCN1
amplification in SCC negatively correlates with cause-specific survival suggesting that
targeting DCN1 may be of clinical utility.28-35 Although other important cancer gene(s) may be
present within the amplicon, several studies have demonstrated that DCN1 plays a critical role
in tumor progression and metastasis, and drives selection for 3q amplification in SCC.32, 36-39
Recently, DCN1 has been reported to have a driving role in prostate cancer and depletion of
DCN1 in LNCaP cells significantly reduced their proliferation, migration, and invasion
40
41
capability. Similar results have been reported in cervical cancers. DCN1’s emerging roles
in these diseases suggest it may present a new oncology drug discovery target.42, 43
A druggable pocket in DCN1 was initially revealed by crystallographic and biochemical
studies showing how DCN1’s activation of the NEDD8 pathway requires binding to UBE2M.
This interaction is driven by binding of UBE2M’s N-terminal acetyl-methionine and immediately
proximal amino acid side chains to a ≈350 Å hydrophobic pocket on DCN1.19, 20, 23, 27 This
interaction depends on the N-terminal acetylation of UBE2M, with the equilibrium dissociation
constant for the N-terminally acetylated protein being >100-fold more potent relative to the
non-acetylated protein.
3
4
ACS Paragon Plus Environment

Page 5 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We recently reported the optimization of a series of N-benzyl piperidines that occupy the
UBE2M-binding pocket on DCN1.17, 18 While this series of compounds, represented by NAcM-
OPT (Figure 1), were the first reported small molecule inhibitors of DCN1, they suffered
several limitations: 1) they do not access the N-acetyl pocket that controls the binding of the
native substrate;27 2) they possess only one stereocenter, thus lacking three-dimensional
character affording access the available sub-pockets within the binding pocket, and 3) they
have a moderate murine half-life that requires relatively high and frequent dosing to maintain
relevant concentrations in mouse models.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A common challenge associated with developing new classes of chemical probes is
determination of whether elicited cellular phenotypes are caused by engagement of the
desired target or by non-specific interaction(s) driven by the inhibitor’s core structure. Although
biological tools (CRISPR, siRNA, etc.) offer some insights, reduction of target protein levels
and acute pharmacologic inhibition of specific protein interactions may have different
consequences. Therefore, we sought a structurally unrelated class of small molecule inhibitors
that might overcome the inherent liabilities of the piperidine class. This work focuses on a
second class of inhibitors that contain a pyrazolo-pyridone core. The original lead was
discovered during our initial high-throughout campaign. This class has two chiral centers thus
imparting a greater degree of three-dimensional structure and affording opportunities for
increased binding potency, target selectivity, and improved solubility. Our overall strategy was
to first identify the minimum pharmacophore for this class, next define the structural drivers of
potency for binding to DCN1, and then use data gathered about optimal substituents from the
first-generation inhibitors to optimize binding. Here, new compounds were designed based on
hypotheses generated from combining empirically derived SAR and examination of X-ray co-
structures. New analogs were tested for potency using our previously reported TR-FRET
assay (vide infra). Key compounds were evaluated for cellular target engagement and
neddylation activity by immunoblotting for steady-state levels of cullin neddylation using two
cancer cell lines with amplified DCN1 expression (HCC95 and CAL33). Ultimately, these
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 44
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
studies identified compound 27, which is 25-fold more potent in our biochemical TR-FRET
assay than the original hit, thermally stabilizes DCN1 in cells, and selectively reduces steady-
state levels of CUL1 and CUL3 neddylation in HCC95 and CAL33 cells.
Results
0
1
2
3
4
5
6
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
19
We reported the results of our high-throughput screen of roughly 600,000 compounds
using a binding assay based on the TR-FRET signal between a biotinylated DCN1 protein,
recognized by Terbium-linked streptavidin, and the helical stapled peptide derived from N-
terminally acetylated UBE2M harboring a C-terminal AlexaFluor 488. This work identified
several structurally divergent inhibitors of the targeted binding event with potency (IC )
50
ranging from 1 to 30 micromolar.19 Prior manuscripts discussed the optimization of the hit
series leading to NAcM-OPT. This manuscript focuses on optimization of a series of pyrazolo-
pyridones, represented by compound 1, discovered during our screen.
Resynthesis of several analogs validated the potency of the pyrazolo-pyridone screening
samples (Figure 2, compound 1). Compound 1 also inhibited the transfer of NEDD8 from the
activated E2 to its cullin substrate in our pulse-chase assay23, 27 (Supplementary Figure C).
An X-ray co-structure of the compound 1 with DCN1 showed that it bound to the targeted
UBE2M binding pocket (Figure 2).
Cl
Cl
O
O
N
HN
HN
NH
N
N
N
O
Compound 1
NAcM-OPT
Gln114
Figure 2. Left: Overlay of 1(orange):DCN1(gray) and NAcM-OPT(blue):DCN1 X-ray crystal
co-structures (PDB 6P5W and 5V86), highlighting that 1 more efficiently occupies the N-acetyl
pocket (yellow highlight). Hashed orange line represents a potential hydrogen-bonding
6
ACS Paragon Plus Environment

Page 7 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
interaction between the amide of 1 and backbone amide of Gln114. Right: 2D chemical
structures of compound 1 and NAcM-OPT.
Design and synthesis
Based on validation of the screening hit, we embarked on a hit-to-lead campaign aimed at
delivering an early lead that overcame the challenges associated with our first-generation
inhibitors embodied by NAcM-OPT. Our initial compound design was based on insights
derived from overlaying the X-ray co-structures of 1:DCN1 (PDB 6P5W), UBE2MNAc:DCN1
(PDB 5V83), and NAcM-OPT:DCN1 (PDB 5V86) (Figure 2) and is summarized in Figure 3.
Compound 1 aligns well with NAcM-OPT in addressing three binding sub-pockets: the Ile, Leu,
and hinge pockets. Therefore, initial substitutions in these pockets were informed by the
structure-activity relationships (SAR) derived during the optimization of first-generation lead
NAcM-OPT.17, 18 NAcM-OPT does not directly occupy the N-acetyl pocket (yellow highlight,
Figure 2). However, its piperidine ring is engaged in a putative hydrogen bonding network
with a bound water molecule within the N-acetyl pocket that strongly suggests fulfilling this
pocket could strengthen binding. This hypothesis is further supported by characterization of
analogs of the native binding peptide, which demonstrated that occupancy by even a small
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
27
hydrophobic group (acetyl vs. formyl) can enhance affinity 10-fold. The X-ray co-structure of
1:DCN1 shows that compound 1 positions the central pyrazolo-pyridone ring closer to the N-
acetyl pocket (Figure 2). Thus, we designed analogs to test whether substitution on the
pyridone’s nitrogen could induce tighter binding by direct hydrophobic packing rather than
indirect water-mediated binding (Figure 3). In addition, we explored the base pharmacophore
for the pyrazolo-pyridone series by utilizing reverse amides in the hinge pocket to determine
the significance of a predicted hydrogen bond between compound 1’s amide and Gln114 of
DCN1 (Figures 2, 3). Compound 1 has four aromatic rings that could negatively affect its
physiochemical properties and reduce its utility as a cellular probe to study the cullin
neddylation.44 To address this issue, we tested whether smaller, more flexible substituents
could replace these aryl rings. We also tried a variety of substituent modifications on the aryl
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 44
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
ring to introduce additional electronic or hydrophobic interactions (Figure 3). The overall goal
was to increase or maintain potency by targeted substitution of one ring while reducing total
molecular weight by removing a less critical ring or substituent.
[
Design]
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
[N-Acetyl Pocket]
•
Mimic N-Acetyl
N-Acetyl
Ile
Hinge
O O
HN
NH
N
N
Leu
[
Ile & Leu Pocket]
[Hinge Pocket]
•
•
Reduce # of Aryl rings
Refine Hydrophobic interactions
•Explore induced pocket;
multiple substitiution,
swap amide bond
•Add electronic interactions
Figure 3. Design strategy and X-ray co-structure overlay of compound 1:DCN1 (PDB 6P5W)
and UBE2M:DCN1 (PDB 5V83), highlighting the key regions targeted for optimization.
Compounds were synthesized by a short and efficient three-step procedure consisting of
45
46
preparation of the oxazolone intermediate , pyrazolo-pyridone ring formation , and
substitution using alkylation or acylation (Scheme 1). Formation of the core pyrazolo-pyridone
ring afforded a separable mixture of cis and trans diastereomers. Typically, the cis product
appeared as the less polar spot by TLC (hexane/ethyl acetate). The cis and trans isomers
1
were distinguished by H NMR and were assigned based on application of the Karplus
equation to the 3JH-H vicinal proton-proton coupling for the C4 or C5 protons of pyrazolo-
pyridone ring (Cis = 7-8 Hz, Trans = 9-11 Hz). The inactive trans-isomer could be converted
to the active cis product by stirring with a catalytic amount of Lewis acid (SnCl ) in refluxing
2
chlorobenzene. However, use of a large excess of Lewis acid (10+ equivalents) favors
formation of dehydrated side products rather than the pyrazolo-pyridone ring.
Alkyl substitution of the pyrazolo-pyridone amide proceeds under relatively mild conditions
presumably due to the presence of adjacent electron-withdrawing groups that increase amide
acidity. Most alkylations afforded a 5:1 mixture of N- vs. O-alkylation as determined by 2D
HMBC NMR analysis (Supplementary Figure A). Finally, a series of reverse amides were
synthesized from dimethyl malonate through an alternate route: condensation, hydrolysis, and
8
ACS Paragon Plus Environment

Page 9 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
amide coupling (Supplementary Scheme A).
O
NH2
O
H
R3
O
a or b
N
N
c
HO
N
R1
R2
O
N
R1
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
R2
H
O
R4
O O
O O
O
O
R4
R1
R3
R3
HN
N
N
R3
NH
R3
NH
R3
NH

N
N
C5
N
N

d or e
C4
N
N
R2
R2
R2
Scheme 1. General synthesis of pyrazolo-pyridones. Reagents and conditions: (a) NaOAc,
o
Ac O, 85 - 90 C, 1 - 2 h; (b) i) EDCI·HCl, DCM, r.t., 16 h; ii) Al O , 4Å MS, DCM, r.t., 16 h; (c)
2
2
3
SnCl , chlorobenzene, reflux, 16 h; (d) Cs CO , R-X (X= Br, I), DMF, r.t., 1 - 16 h; (e) AcCl,
2
2
3
pyridine, DCM, r.t., 1h.
The key annulation reaction (Scheme 1, step c) was previously reported using a mixture
4
6
of ethylene glycol and acetic acid as the solvent. In our hands, these conditions afforded low
yields and complex mixtures, particularly formation of undesired ethylene glycol adducts.
Switching from a nucleophilic polar protic solvent to a non-polar or polar aprotic solvent such
as chlorobenzene, DMF, or NMP improved yields and suppressed formation of side products.
As discussed below, the initial SAR showed that the cis-diastereomers were active and the
trans-diastereomers inactive. Therefore, we optimized the reaction to favor formation of the
cis-diastereomer. Addition of a catalytic amount of tin (II) chloride followed by refluxing in
chlorobenzene overnight proved the most selective, affording a 3:1 cis to trans ratio, and
moderate isolated yields of the pure cis product (30 - 40%).
Structure−Activity Relationships.
The main goal of the first phase of the work was to define and optimize the core
pharmacophore by testing requirements for key rings and functional groups. The secondary
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 44
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
goal was to define the best range of pendant substituents occupying the key sub-pockets. Our
overall optimization strategy uses a standard recursive process of sequential hypothesis-
based design of analog sets and testing of those sets in models of increasing complexity to
define possible structure-activity relationships (SAR) that drive further analog design. All novel
analogs were prepared and tested as racemates. Potency was assessed using our TR-FRET
binding assay. We targeted four regions of the compounds, named according to the peptide
ligand feature that they replace within the ligand-binding pocket, to systematically delineate
the SAR: the Ile, N-acetyl, Leu, and hinge pockets (Figure 3). During this study, we prepared
and tested over 140 analogs. However, to improve clarity and readability of the manuscript,
only key compounds, which clearly delineate SAR, are reported in the main text. A complete
list of all compounds prepared and evaluated during this study and supporting noted SAR
trends can be found in the Supplementary Tables.
0
1
2
3
4
5
6
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
First, we prepared roughly 30 compounds to test the inhibitory effects of the cis- and trans-
diastereomers about the core pyrazolo-pyridone ring. In general, only the cis diastereomers
inhibited DCN1-UBE2M binding at or below the maximum tested concentration of 15 µM. All
the trans isomers were inactive at the maximum tested concentration of 15 µM
(Supplementary Table A). These results demonstrate the three-dimensional orientation of
two aromatic rings is critical to activity. Therefore, subsequent work utilized only the purified
cis-diastereomers and all data reported below represent that set of compounds.
1
0
ACS Paragon Plus Environment

Page 11 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O O
R
HN
O O
NH
N
N
HN
NH
N
N
R
F
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
IC50 (M)
TR-FRET)
IC50 (M)
(TR-FRET)
No
1
R
No
10
R
(
5.1 ± 0.6
4.0 ± 0.5
1.4 ± 0.2
13 ± 4
2
3
3
1.4 ± 0.2
3.5 ± 0.5
11
CF3
F
4
5
7.1 ± 2.6
3.5 ± 0.2
12
13
7.1 ± 0.9
>15
F
I
Cl
N
N
Cl
F
6
7
>15
14
15
>15
>15
3.0 ± 0.3
N
8
9
3.1 ± 0.2
1.3 ± 0.1
16
17
>15
>15
Table 1. SAR of Leu and hinge pocket. IC50 values were generated using our TR-FRET
binding assay and are represented as the mean of three replicates with errors reported as the
standard deviation.
Second, we examined the steric and electronic requirements of the Leu pocket (Table 1).
Introduction of a p-fluorine (3) increased potency relative to the unsubstituted benzene ring
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 44
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). A survey of other p-halogen substitutions (Cl or Br) or the p-methyl substitution (1)
revealed that they were less potent than the p-fluorine (3) analog (Supplementary Table B-
). Introduction of either the p-cyano (42) or p-methoxy (43) group completely abolished
activity (Supplementary Table B-1). Shifting the fluorine to the m-position of the phenyl ring
4) reduced potency by 5-fold. However, the m-, p-difluoro analog (38) retained the activity of
. Together, these results suggest that the p-fluorine substitution is critical for binding and
1
0
1
2
3
4
5
6
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
other small electron-withdrawing substituents, such as cyano, are not viable replacements.
This trend was reinforced by the inactive isosteric pyridine analog (6), whose activity could be
partially rescued by addition of a p-fluoro substituent (5).
We also investigated the effects of more flexible substituents. Replacement of the phenyl
ring with a benzyl group (7) retained the potency of the unsubstituted phenyl analog (2).
Replacement of the aryl ring with a variety of linear (propyl or butyl) or cyclic (cyclohexyl and
cyclopentyl) alkyl groups was well tolerated (8, 9, and 44-46, Supplementary Table B-1).
However, introduction of the shorter, isopropyl substituent (47) removed activity. The data
suggest that occupation of the hydrophobic Leu pocket is a critical driver of potency but that
the pocket is relatively forgiving, within the constraints of a fixed steric bulk. Thus, the Leu
pocket affords a site that could be further manipulated to modulate the physiochemical
properties of the inhibitors. Due to its superior potency and inert chemical nature, the
remainder of our SAR study fixed the substituent targeting the Leu pocket as the p-F–phenyl.
Surveying possible phenyl ring substituents for the hinge pocket (Table 1, Supplementary
Table B-2) revealed that m-substitution of the phenyl ring was critical for potency. Removal of
the m-methyl from the phenyl ring (10) reduced potency by 10-fold. In addition, changing the
position of methyl to either the p- (14) or o- (15) position completely abrogated activity. The
introduction of bulky m-substituents, such as tert-butyl (17), was not favorable. Replacement
of m-CH with the less metabolically labile m-CF group (11) resulted in a two-fold reduction
3
3
in potency. Replacement with smaller halogens (F (53), Cl (54)) was not tolerated but larger,
roughly isosteric, halogens (Br (55) and I (12)) retained some activity (Supplementary Table
1
2
ACS Paragon Plus Environment

Page 13 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
B-2). This sharp SAR about the hinge pocket is consistent with the SAR studies of NAcM-
OPT.17, 18 Similar to the binding of NAcM-OPT, the X-ray co-structure of 1:DCN1 revealed that
the hinge pocket undergoes a structural rearrangement to produce a deep narrow hydrophobic
cavity composed of alkyl and aromatic residues (Ile86, Phe89, Val102, Ile105, Ala106, Phe109,
Ala111, Phe117, Phe122, Phe164) that tightly interacts with the phenyl ring.
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
Rather unexpectedly, the 3,4-dichloro phenyl analog (13), which mimics NAcM-OPT, was
completely devoid of activity. This result suggests that the pyrazolo-pyridone analog does not
present the aryl ring to the induced hinge pocket in the same way that NAcM-OPT does.
Examination of the X-ray co-structures revealed that both series have a key H-bond with the
amide backbone of Gln114 on DCN1, which lock the compounds in a specific orientation.
However, while NAcM-OPT participates as an H-bond donor through its urea, compound 1
interacts as an H-bond acceptor through the oxygen of its amide. The highly restrictive steric
and electronic requirements of this pocket are further illustrated by the complete loss of
potency upon introduction of either a m-F-phenyl (57) or a pyridine replacement (16).
Converting the phenyl group to an aliphatic butyl chain (60) also results in loss of activity
(Supplementary Table B-2). Taken together, the SAR reveals that the hinge pocket requires
a phenyl ring with a small hydrophobic substituent, most efficiently met by a m-methyl
substituted phenyl ring.
Finally, we tested whether the reverse amide could maintain the key hydrogen bonding
interaction with Gln114 or switch compound 1 from an H-bond acceptor to an H-bond donor,
as observed with NAcM-OPT. Investigating five analogs, including the m-methyl (128) and m-,
p-dichloro phenyl (131) derivatives, we found that the reverse amide was not tolerated
(Supplementary figure D-1). Further optimization studies fixed the substituent targeting the
Leu pocket as the p-F–phenyl and the substituent targeting the hinge pocket as m-tolyl.
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 44
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
O O
R2
O O
HN
N
R
NH
R1
NH
N
N
N
N
0
1
2
3
4
5
6
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
F
IC50 (M)
TR-FRET)
IC50 (M)
(TR-FRET)
No
3
R
No
27
R1
R2
(
1.4 ± 0.2
>15
0.2 ± 0.1
4.1 ± 0.8
18
28
29
N
O
19
>15
7.1 ± 1.7
O
20
21
H
>15
30
5.6 ± 0.8
0.2 ± 0.1
1.2 ± 0.2
31
0
.7 ± 0.1
TLC-up)
.6 ± 0.1
1.2 ± 0.1
(TLC-up)
0.5 ± 0.2
(TLC-down)
(
0
22-a, b
32-a, b
(TLC-down)
23
24
25
0.5 ± 0.1
33
34
35
2.3 ± 0.8
0.8 ± 0.1
0.2 ± 0.1
0.3 ± 0.1
2.0 ± 0.1
Table 2. SAR of Ile and N-acetyl pocket. IC values were generated using our TR-FRET
5
0
binding assay and are represented as the mean of three replicates with errors reported as the
standard deviation.
Next, we explored the SAR of substituents reaching into the Ile pocket (Table 2,
Supplementary Table B-3). Introduction of a methylene between the phenyl and pyrazole
rings (18) reduced activity 10-fold. Thus, there is a limitation on the size of the substituent
1
4
ACS Paragon Plus Environment

Page 15 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
targeting this pocket. Introduction of a fluoro substitution at either the o- (61) or p- (62)
positions had little effect on potency (Supplementary Table B-3). Since one of our design
goals was to reduce the aryl ring count, we investigated replacements of the phenyl ring.
Replacements of aryl ring with aliphatic chain were successful in many cases (22 – 25). Linear
alkyl chains three (21) or two (65) carbons in length proved roughly equipotent to the phenyl
derivative (3). Introduction of a linear four-carbon chain (63) reduced potency three-fold. The
unsubstituted (20) and methylated (66) pyrazoles were inactive (Supplementary Table B-3).
Introduction of substituted alkyl chains (22, 23, and 24), designed to mimic the Ile residue
displayed two - five times better potency than the phenyl derivative (3). Introduction of a methyl
group alpha to the pyrazole afforded an additional chiral center and the resulting
diastereomers were separable by silica gel chromatography. In the case of the 2-butyl
derivative, the diastereomers proved roughly equally potent (22a, b). However, in the cases
of the 1-cyclopropyl-2-propane (70a, b) and 1,1,1-trifluoro-2-propane (71a, b) derivatives, one
isomer was two to three times more potent. The cyclopropyl compound (25) was roughly
equipotent to the phenyl (3). However, the cyclopentyl (24) and cyclohexyl (72) derivatives
were two- and three-fold more potent respectively. The replacement of the phenyl ring with a
polar pyridine (19) or addition of a polar atom to the linear chain (26) significantly decreased
potency. Thus, demonstrating a strong preference for hydrophobic substituents in the Ile
pocket. We also examined the effects of substituting the other pyrazole nitrogen. All such
analogs were inactive at the maximum tested concentration of 15 µM (Supplementary figure
D-1). Therefore, the alignment of pyrazolo-pyridone scaffold’s amide bond and pyrazole’s
substituent are both crucial for binding.
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
To understand requirements for binding to the N-acetyl pocket, we investigated whether
substitution of the pyridone ring nitrogen (Table 2, Supplementary Table B-4) could mimic
the N-terminal acetyl of the native UBE2M substrate. Alkyl substitution significantly increased
potency with a relative order of no-substituent < propyl < methyl < ethyl (Supplementary
Table B-4). In addition to efficiently occupying the N-acetyl pocket, alkyl substitution of the
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 44
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
pyridone improves peripheral hydrophobic interactions by pushing the p-F-phenyl substituent
deeper into the Leu pocket (Figure 7). The five- to ten-fold observed improvement in potency
afforded by substitution of the pyridone ring is consistent with previous binding studies of the
native substrate. Careful characterization of the UBE2M-DCN1 interaction showed that N-
terminal acetylation of UBE2M enhances its binding affinity for DCN1 ten-times relative to N-
0
1
2
3
4
5
6
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
terminal formylation. Thus, occupancy of the N-acetyl binding pocket on DCN1 by even a
small hydrophobic group can significantly enhance binding. Interestingly, introduction of acyl
groups (29, 30), directly mimicking UBE2M’s N-terminal acetyl, gave compounds with poor
potency.
The compound reorientation, induced by substituting the pyridone, subtly effected the SAR
about the Ile pocket. Linear alkyl and aryl Ile substituents remained roughly equally potent.
However, sterically hindered alkyl Ile substituents (32 - 34) showed decreased potency relative
to the unmodified pyridone. These results imply a steric interaction between the two adjacent
substituents at the Ile and the N-acetyl pocket that induce an unfavorable change in their
orientation. The cyclopropyl group (35) was an exception to this trend and increased potency
about ten-fold relative to its parent (25). O-alkylated isomers were also prepared but proved
inactive at the maximum tested concentration of 15 µM (Supplementary Table C).
Biochemical and cellular profiling of compound 27
As discussed above, DCN1 is one of the five human paralogues (DCN1-5), which display
high structural similarity. In biochemical assays, all five isoforms can stimulate cullin
neddylation.24, 42 The selectivity of 27 was assessed using a pulse-chase assay to determine
the effects on NEDD8 transfer stimulated independently by each of the five DCN isoforms. At
the concentration employed for our cellular studies (10 µM), compound 27 was highly selective,
inhibiting DCN1 and DCN2, which are 100% identical in the N-Ac-Met binding pocket. DCN3,
DCN4, or DCN5 stimulated cullin neddylation were not inhibited at 10 µM (Figure 4). The
observed selectivity can be explained by subtle differences between the binding pockets on
1
6
ACS Paragon Plus Environment

Page 17 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
19
DCN1/2 and DCNs 3-5. The high specificity of 27 amongst the highly homologous family of
enzymes implies it is likely to be selective against other less-related binding pockets, as has
been demonstrated for NAcM-OPT.19
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. The DCN isoform selectivity of compound 27 and the known DCN1/2 selective
inhibitor NAcM-OPT (positive control). Pulse-chase assay monitoring transfer of NEDD8 from
UBE2M to CUL2 in the presence of 10 µM inhibitor or DMSO as tested against all five of the
DCN family members. Note different reaction times, based on differences in stimulation by the
different DCN family members. “N8” refers to NEDD8, and “~” refers to covalent complex.
A subset of inhibitors was tested for kinetic aqueous solubility (pH = 7.4 buffer). The
solubility of the compounds ranged from poor (≤ 1 µM) to good (> 60 µM) (Table 3). Notably,
switching the aryl substituent targeting the Ile pocket to an aliphatic group significantly
improved solubility (3 vs. 65, Table 3). Next, the compounds were evaluated for oxidative
metabolic stability using murine microsomes. For microsomal modeling, we chose a drug
concentration of 0.8 µM, which is routinely achieved in plasma by small molecule inhibitors
and is roughly comparable to the biochemical EC95 of our more potent compounds. Analogs
with alkyl substitution of the pyrazole ring were more rapidly metabolized than their phenyl
substituted counterparts (27 vs. 31, Table 3). Compound 3 and compound 27 showed roughly
comparable (within 2-fold) murine microsomal stability to that previously reported for NAcM-
OPT. Based on the combination of potency, solubility, and microsomal stability, compound 27
was selected as the top compound for further cellular characterization.
Given the goal of this work was to generate a novel class of cell active chemical probes,
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 44
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
we investigated the top compounds’ activity in cell-based studies. First, we tested compound
effects on monolayer proliferation using a non-transformed fibroblast (BJ) cell line lacking
DCN1 amplification. At the maximum tested concentration of 25 µM, none of the compounds
inhibited cell growth (Table 3). Next, we assessed growth inhibition using a squamous cell
carcinoma cell line (HCC95) known to express high levels of the DCN1 protein. Again, the
compounds did not inhibit cell growth at 25 µM (Table 3).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
ACS Paragon Plus Environment

Page 19 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure
Proliferation EC50 (M)
(Normal Fibroblasts - BJ) (HCC 95)
t(1/2) /hr
(mouse)
CLint' (ml/min/kg) Avg. Sol. (µM)
No
R1
R2
(mouse)
(Kinetic)
R2
O O
N
R1
NH
N
N
F
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
H
> 25
> 25
> 25
> 25
1.1 ± 0.1
1.2 ± 0.1
50.3 ± 5.0
44.9 ± 3.7
5.5 ± 1.0
27
16.6 ± 0.6
6
5
1
H
H
> 25
> 25
> 25
> 25
0.6 ± 0.1
<0.1
89.7 ± 4.7
-
92.8 ± 2.7
10.9 ± 4.3
3
2
4
5
> 25
> 25
> 25
> 25
0.5 ± 0.1
<0.1
116.4 ± 10.7
-
17.1 ± 1.4
46.1 ± 3.6
3
N
O
N
N
O
123
> 25
> 25
ND
ND
ND
N
H
F
O O
HN
NH
N
N
1
8
> 25
> 25
0.3 ± 0.1
160.4 ± 5.0
1.1 ± 1.0
O O
HN
NH
N
N
7
> 25
> 25
ND
ND
ND
F
N
F
O
N
N
O
141
N
H
> 25
> 25
ND
ND
ND
Table 3. Kinetic solubility, microsomal stability, and cell proliferation effects in normal
BJ fibroblasts and malignant HCC95 cells. Values for the proliferation, half-life, clearance
and solubility are represented as means plus or minus standard deviation calculated from at
least one independent experiment, run in triplicate. ND indicates values were not determined.
A key feature of any chemical probe is that it must engage its intended target in cells. We
used the cellular thermal shift assay (CETSA)47 to demonstrate that 27 engages DCN1 in
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 44
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
HCC95 cells. Figure 5 demonstrates that the DCN1 protein is unstable at 52 °C in HCC95
cells treated with DMSO. However, in the presence of 10 µM of 27 or the known DCN1/2
inhibitor NAcM-OPT, the thermal stability of the DCN1 protein is enhanced. Compound 27
appears to more strongly enhance the thermal stability of DCN1 at 52 °C relative to NAcM-
OPT (Figure 5). This suggests that compound 27 effectively engages DCN1 in HCC95 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
Figure 5. Enhancement of DCN1 thermal stability by compounds 27 and NAcM-OPT (positive
control) but not by DMSO (negative control). HCC95 cells were treated with either DMSO or
10 μM of the indicated compound for 1 h, heated at the indicated temperature for 3 minutes,
lysed, and blotted with the indicated antibodies.
To confirm the cellular effects on neddylation, compound 27 and an inactive trans-isomer
(87) were tested for inhibition of steady-state cullin neddylation as detected by immunoblot
analysis. HCC95 and CAL33, two squamous cancer cell lines highly expressing DCN1 and
reported to be sensitive to DCN1 inhibition,19 were chosen for analysis (Figure 6). As
previously observed for other DCN1/2 selective inhibitors,17-21 compound 27 and NAcM-OPT
but not the negative control (87) nor DMSO selectively inhibited steady-state cullin neddylation.
In both cell lines, the most pronounced effects were observed on the neddylation of CUL1 and
CUL3 but more subtle effects were observed for CUL4A. In concordance with our previous
2
0
ACS Paragon Plus Environment

Page 21 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
studies,19 while complete blockade of neddylation by MLN4924 elicits similar cellular
responses across cell lines, acute pharmacologic inhibition of DCN1-mediated neddylation
show differential effects. In CAL33 cells, both NAcM-OPT and 27 display more pronounced,
albeit weak, inhibition effects on CUL2 neddylation compared to HCC95. Conversely, the
marginal effects on CUL5 neddylation are more robust for both compounds in HCC95 relative
to CAL33. Unexpectedly, there appears to be subtle differences in the potency of NAcM-OPT
and 27 in the two cell lines. In HCC95, 27 appears to more potently inhibit cullin neddylation
than NAcM-OPT. In direct contrast, in CAL33, NAcM-OPT appears more potent than 27.
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
Consistent with all known small molecule inhibitors of the DCN1-UBE2M interaction,17-21
acute pharmacologic inhibition of DCN1-mediated neddylation results in a smaller reduction
in neddylation compared to inhibition of the neddylation activating enzyme (E1) by MLN4924
or other NAE inhibitors. The difference in activity results from MLN4924’s irreversible inhibition
of the sole E1 enzyme in the neddylation cascade. Thus, no NEDD8 transfer to the E2 and
subsequently to the target protein (i.e. cullins) can occur. Alternatively, the residual
neddylation activity observed for the DCN1 inhibitors is caused by DCN1-independent
neddylation and likely represents the basal activity of RBX1, the other co-E3 enzyme that
19
works with DCN1 to stimulate neddylation. The most stark phenotypic consequence of these
different mechanisms of action, is the effect on gross substrate stabilization. Therefore, with
short exposures of western blots that enable direct comparison to MLN4924, small molecule
inhibition of the DCN1-UBE2M interaction does not elicit gross substrate stabilization (S.E. in
Figure 6). Longer exposures, which result in saturation of the signal for substrates with
MLN4924 treatment, permit the visualization of modest steady-state stabilization of some
canonical CUL1/CUL3 substrates. HCC95 cells appear more responsive and subtle
stabilization of several known CUL1/CUL3 substrates (p27, p21, cyclin E, and NRF2) can be
detected at longer exposures (L.E. in Figure 6). Alternatively, in CAL33, even at longer
exposures, only p27 and NRF2 stabilization is observed.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 44
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
The observed stabilization of NRF2 could also be explained in a DCN1 independent
manner if the tested compounds induced reactive oxygen species (ROS). Thus, ROS
induction was measured for each compound in dose-response using HCC95 cells and the cell
permeant reagent 2’,7’–dichlorofluorescin diacetate (DCFDA, also known as H2DCFDA and
as DCFH-DA), a fluorogenic dye that measures hydroxyl, peroxyl, and other ROS activity.
Even at 30 µM, neither 27, 87, nor NAcM-OPT have any measurable effects on reactive
oxygen species suggesting the observed effects on NRF2 is not induced by ROS
0
1
2
3
4
5
6
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
(Supplementary Figure D).
2
2
ACS Paragon Plus Environment

Page 23 of 44
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 6. Western blot for inhibition of cellular neddylation at 48 hours by DMSO, MLN4924
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 44
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
(
positive control, 1 µM), 87 (negative control, 10 µM), 27 (10 µM), and NAcM-OPT (10 µM) as
indicated by steady-state levels of CUL1~NEDD8. “N8” refers to NEDD8, and “~” refers to
covalent complex.
0
1
2
3
4
5
6
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
Discussion and Conclusions
In conclusion, a series of pyrazolo-pyridone analogs were investigated as novel inhibitors
of the DCN1-UBE2M interaction and cullin neddylation. The efficient synthetic routes outlined
in Scheme 1 enabled rapid access to analogs to define the SAR and permitted gram scale
preparation of 27. Comparative analysis of multiple X-ray co-crystal structures enabled
rational drug design to access the N-acetyl pocket, providing a 10-fold boost in potency.
Empirical systematic investigations of the structure-activity relationships led to a refinement of
peripheral interactions. Together, these studies yielded compound 27, which is with 25-fold
more potent than the HTS hit compound (1). The SAR studies revealed four critical drivers for
potency (Figure 7): (1) the stereochemistry of the pyrazolo-pyridones, with the cis isomer
being much more active than the trans isomer; (2) filling the N-acetyl binding pocket, with the
ethyl substitution on pyrazolo-pyridone ring increases potency 5 – 10 fold; (3) replacement of
either aryl ring targeting the Ile or Leu pocket with saturated alkyl chains does not compromise
potency; and (4) the hinge pocket affords the sharpest SAR and even subtle electronic or
steric deviation from the meta-substituted aryl significantly compromise potency. A high-
resolution X-ray co-crystal structure of compound 27 bound to DCN1 confirmed the binding
mode and our SAR hypotheses that the ethyl group on pyrazolo-pyridone ring efficiently fills
the N-acetyl pocket and pushes the compound deeper into the Leu pocket (Figure 7).
2
4
ACS Paragon Plus Environment

Page 25 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
[
SAR summary]
[N-Acetyl Pocket]
5-10 times potency improvement
Best: Ethyl
O O
N
NH
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
N
N
Tight induced fit
Best: m-Tolyl
Tyr 181
Aryl  Alkyl >> H
Hydrophobic
F
Aryl  Alkyl
Best: p-Fluorophenyl
Gln114
Figure 7. SAR summary and X-ray crystal structure of compound 27(yellow):DCN1(gray)
PDB 6P5V). The ethyl substituent is represented as space filling spheres to highlight the key
(
driver of potency. Hashed orange line represents the 2.6 Å hydrogen-bonding interaction
between the compound’s hinge amide and backbone amide of DCN1 Gln114. DCN1 Tyr181
is shown as sticks to highlight the hydrophobic interaction and potential electrostatic
interaction.
A subset of compounds was profiled for kinetic aqueous solubility (pH = 7.4 buffer), murine
microsomal stability, and their effects on the cellular proliferation of both normal (BJ) and
transformed (HCC95) mammalian cell lines (Table 3). The lack of growth toxicity and range
of solubility and stability values observed for a variety of biochemically potent analogs suggest
that future rounds of optimization are likely to improve in vitro pharmacokinetic characteristics.
Together, these studies suggested compound 27 had the best combination of properties and
thus was selected for further profiling.
Compound 27 binds to the targeted pocket on DCN1 (Figure 7) and potently and
selectively inhibits DCN1/2 mediated cullin neddylation in biochemical assays (Table 2 and
Figure 4). Cell-based studies demonstrated that compound 27 engages cellular DCN1
(Figure 5) and selectively reduces the level of steady-state CUL1 and CUL3 neddylation in
HCC95 and CAL33 cells (Figure 6). As seen previously, acute pharmacologic inhibition of the
DCN1-UBE2M interaction has different consequences from complete blockade of the NEDD8
pathway by MLN4924 and does not grossly stabilize canonical CRL substrates.
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 44
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
In HCC95 cells, compound 27 appears to better enhance the thermal stability of DCN1
relative to NAcM-OPT (Figure 5), which translates to more potent inhibition of cellular
neddylation. Thus, supporting our original design hypothesis that fulfilling the N-acetyl pocket
could afford more potent inhibitors of cullin neddylation in HCC95 cells. Overall, 27 overcomes
two of the three noted limitations of NAcM-OPT by efficiently occupying the N-acetyl pocket
and incorporating additional three-dimensional character. The promising preliminary ADME-
Tox data suggests future optimization studies, with a focus on metabolic stability, may yield
compounds suitable for single daily oral dosing in murine models and overcome the final
liability of NAcM-OPT.
0
1
2
3
4
5
6
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
Perhaps most intriguingly, the increase in potency observed in HCC95 cells does not
translate to increased potency relative to NAcM-OPT in CAL33 cells. These results
demonstrate that two different chemical classes with similar biochemical potencies can exhibit
differential, cell line specific, effects on steady-state cullin neddylation. It also further supports
the hypothesis that having multiple structural classes of inhibitors blocking the DCN1-UBE2M
interaction is crucial to permit interrogation of the function of DCN1-mediated neddylation.
Currently, it is unknown when neddylation of a particular cullin would require DCN1/2
activity or why different chemical classes of inhibitors would elicit selective cellular effects. A
broad range of factors could be responsible for dictating this selectivity such as expression
levels and cellular sub-localization of DCN1 and the other components of the neddylation
pathway. In the long-term, developing multiple chemical scaffolds of DCN1 inhibitors may
prove clinically relevant for different sub-types of cancers. Alternatively, combinations of
various chemotypes of DCN1 inhibitors may afford simultaneous access to different sub-
cellular pools of DCN1, effecting a broader spectrum of activity across multiple cancers. The
emerging role of DCN1 in multiple tumor types and in vitro data for the DCN1 inhibitors
suggests that they may offer an orthogonal, more selective, means to inhibit the NEDD8
pathway relative to the clinically validated NAE inhibitor MLN4924. Future studies, outside the
scope of this work, will focus on elucidating the underlying mechanisms responsible for when
2
6
ACS Paragon Plus Environment

Page 27 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cellular neddylation of a particular cullin would involve DCN1 activity and why different
structural scaffolds of DCN1 inhibitors would elicit cell line specific effects.
Experimental Section
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
TR-FRET Assay.19
The TR-FRET assay was carried out in black 384-well microtiter plates at a final volume of 20
μM per well. To screen library compounds, the assay cocktail was prepared as a mixture of
50 nM biotin-DCN1, 20 nM Ac-UBE2M12-AlexaFluor488, 2.5 nM Tb streptavidin
(ThermoFisher) in assay buffer (25 mM HEPES, 100 mM NaCl, 0.1% Triton X-100, 0.5 mM
DTT, pH 7.5). The assay cocktail was then incubated for 1 h at room temperature and
distributed using a WellMate instrument (Matrix). Compounds to be screened were added to
assay plates from DMSO stock solutions by pin transfer using 50SS pins (V&P Scientific). The
assay mixture was incubated for 1 h at room temperature prior to measuring the TR-FRET
signal with a PHERAstar or Clariostar plate reader (BMG Labtech) equipped with excitation
modules at 337 nm and emissions at 490 and 520 nm. We set the integration start to 100 μs
and the integration time to 200 μs. The number of flashes was set to 100. The relative
fluorescence (Ex/Em = 520:490) was used for TR-FRET signal calculations. Assay end points
were normalized from 0% (DMSO only) to 100% inhibition (unlabeled competitor peptide) for
hit selection and curve fitting. All compounds were tested in triplicate or more.
General Cell Culture
BJ (ATCC®, CRL-2522™), a normal human foreskin fibroblast cell line was purchased from
the American Type Culture Collection (ATCC) and cultured in Eagle's Minimum Essential
Medium (EMEM) medium (ATCC, 30-2003) supplemented with 10% FBS (ATCC, 30-2020).
HCC95, a squamous cell lung carcinoma line, was cultured in RPMI-1640 medium with 2 mM
L-glutamine (ATCC, 30-2001) supplemented with 10% fetal bovine serum (ATCC, 30-2020).
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 44
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
CAL33 (DSMZ no. ACC 447), a tongue squamous cell carcinoma, was purchased from the
Leibniz-Institut DSMZ - Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH
and was cultured in Dulbecco's Modified Eagle's Medium (DMEM) (ATCC® 30-2002™)
supplemented with 10% fetal bovine serum (ATCC, 30-2020). Cells were routinely tested for
mycoplasma contamination using the MycoAlert Mycoplasma Detection Kit (Lonza, LT07-218).
0
1
2
3
4
5
6
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
Evaluation of Cell Proliferation Effects
Exponentially growing cells were plated in 384-well polystyrene, white, flat bottom, low flange,
tissue culture treated assay plates (Corning, 3570) using a Matrix Wellmate liquid dispenser
(Thermo Fisher), and incubated overnight at 37 ºC in a humidified 5% CO2 incubator. Each
cell line was plated to a previously validated density to ensure logarithmic growth. For BJ, the
cells were plated to 1000/cells per well in 30 microliters of complete media. For HCC95, the
cells were plated to 600/cells per well in 30 microliters of complete media. Test articles were
prepared as stock solutions in DMSO. Drug plates were prepared by serially diluting (ten, 3-
fold dilutions) stock solutions into a 384-well clear polypropylene plate (Corning, 3657). Test
articles were manually transferred (73 nL) from the drug plate to the assay plate by
hydrodynamic pin transfer using a pin tool (AFIX384FP1, V&P Scientific) adapted for manual
transfer (BGPK, VP 381D-N, V&P Scientific) and equipped with FP1S50 pins (V&P Scientific).
After drugging, the plates were incubated at 37 ºC in a humidified 5% CO2 incubator for 72
hours, and then cell viability was measured using Cell TiterGlo (Promega, G7573) according
to the manufacturer’s recommendation. Luminescence was measured on a Clariostar plate
reader (BMG Labtech). Assay endpoints were normalized from 0% (DMSO only) to 100%
inhibition (58 µM Idarubicin) and fit to a semi-log plot using at least n = 3 technical replicates
and n = 2 biological replicates. The curve fit was performed using the Collaborative Drug
Discovery software and the standard Hill equation.
Cellular Thermal Shift Assay.
2
8
ACS Paragon Plus Environment

Page 29 of 44
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The cellular thermal shift assay (CETSA) was performed according to the previously published
47
procedure. Briefly, 500,000 HCC95 cells were aliquoted in 100 μL of PBS and DMSO or the
indicated compound (10 μM) was added (0.1 % final DMSO). Mixtures were incubated on ice
for 1 h, and the cells were washed three times with PBS buffer. Cell pellets were resuspended
in 50 μL of PBS and heated for 3 min at the indicated temperature in a thermocycler. Cells
were lysed by three rounds of freezing in liquid nitrogen and thawing on ice. After pelleting at
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
o
20,000 rpm for 20 min at 4 C, equal amounts of supernatant were removed and blotted with
the indicated antibodies.
Western Blot Analysis
6
Exponentially growing HCC95 or CAL33 cells were plated in 6-well plates at 0.4 x 10 cells/well
in 2 ml of media and incubated overnight at 37 °C in a humidified 5% CO incubator. 24 and
2
48 hrs. after plating, the media was aspirated and replenished with 2 ml fresh media containing
either 4 μL of DMSO or a 500x compound DMSO stock solution. The cells were harvested
after 72 hr via trypsinization, thoroughly washed with PBS, pelleted, and stored at -80 °C. Cell
pellets were thawed on ice and lysed by resuspension in 30-40 μl of lysis buffer [50 mM Tris,
150 mM NaCl, 0.5% NP-40, 0.1% SDS, 6.5 M Urea, 2 mM 1,10-orthophenanthroline, 1X Halt
Protease and Phosphatase inhibitor cocktail (ThermoFisher), 0.25 kU Universal Nuclease
(ThermoFisher), pH 7.5]. Cell suspensions were incubated on ice for 25 minutes with
occasional mixing by pipetting up and down. Lysates were cleared by centrifugation at 13,000
rpm for 20 minutes and the supernatant collected. The protein concentration of total cell lysate
was determined by BCA assay (Pierce) using BSA as a control. Cell lysates were diluted into
2X SDS-PAGE sample buffer such that 25 μg of total protein was loaded per well. Samples
were heated at 95 °C for 2 minutes, briefly cleared by pulse centrifugation, separated on 4-12%
NuPAGE gels (Invitrogen), and transferred to PVDF membranes (BIO-RAD) at 100 V for 90
minutes at 4 °C. Membranes were blocked for 1 hour in blocking buffer consisting of 1x TBS,
0.1% Tween-20, and 5% blotting grade non-fat dry milk (BIO-RAD). Primary antibodies were
2
9
ACS Paragon Plus Environment