Rapid identification of Keap1-Nrf2 small-molecule inhibitors through structure-based virtual screening and hit-based substructure search

Article abstract of DOI:10.1021/jm4017174

In this study, rapid structure-based virtual screening and hit-based substructure search were utilized to identify small molecules that disrupt the interaction of Keap1-Nrf2. Special emphasis was placed toward maximizing the exploration of chemical divers

Full text of DOI:10.1021/jm4017174

Brief Article
pubs.acs.org/jmc
Rapid Identification of Keap1−Nrf2 Small-Molecule Inhibitors
through Structure-Based Virtual Screening and Hit-Based
Substructure Search
Chunlin Zhuang, Sreekanth Narayanapillai, Wannian Zhang, Yuk Yin Sham,* and Chengguo Xing*
†
,‡
†
‡
,§
,†
†
Department of Medicinal Chemistry, University of Minnesota, 2231 Sixth Street SE, Minneapolis, Minnesota 55455, United States
‡
Department of Medicinal Chemistry, Second Military Medical University, 325 Guohe Road, Shanghai 200433, People’s Republic of
China
§
Center for Drug Design, University of Minnesota, 516 Delaware Street SE, Minneapolis, Minnesota 55455, United States
S Supporting Information
*
ABSTRACT: In this study, rapid structure-based virtual
screening and hit-based substructure search were utilized to
identify small molecules that disrupt the interaction of Keap1−
Nrf2. Special emphasis was placed toward maximizing the
exploration of chemical diversity of the initial hits while
economically establishing informative structure−activity rela-
tionship (SAR) of novel scaffolds. Our most potent non-
covalent inhibitor exhibits three times improved cellular
activation in Nrf2 activation than the most active noncovalent
Keap1 inhibitor known to date.
6
INTRODUCTION
Human Keap1 is a 70 kDa protein containing five domains:
■
(
a) N-terminal region (NTR), (b) BTB (broad complex,
tramtrack, and Bric a brac), an evolutionarily conserved
Oxidative stress can lead to chronic inflammation, which is
crucial to the development of many diseases, including cancer,
cardiovascular, and neurodegenerative diseases. Antioxidant
defense system is one of the major mechanisms to protect cells
from such stress. A number of enzymes, such as NADPH:qui-
none oxidoreductase 1 (NQO-1), heme oxygenase-1 (HO-1),
1
protein−protein interaction motif that dimerizes with Cullin
3
(Cul3)-based ubiquitin E3 ligase complex for Nrf2
ubiquitination; (c) intervening region (IVR), a cysteine-rich
region; (d) double glycine region (DGR), a domain that
comprises six-bladed Kelch motifs and binds to the Neh2
domain of Nrf2; and (e) C-terminal region (CTR). Human
Nrf2 has 605 amino acid residues that form six conserved
domains: Neh1−Neh6. Each domain has its own biological
function, particularly the N-terminus Neh2 domain, which
superoxide dismutase (SOD), and glutathione S-transferase
7
,8
2
(
GST), are the key components of this system. They are
mainly regulated by three cellular components: Kelch-like
ECH-associated protein 1 (Keap1), nuclear factor erythroid 2-
related factor 2 (Nrf2) and antioxidant response elements
8
3
contains two regions that bind to Keap1 Kelch domain. One
(
ARE). Under unstressed conditions, Nrf2 remains at low
region (amino acids 17−32) contains a DLG motif with a low
cellular concentration and is negatively regulated by Keap1
mediated proteasomal degradation. Upon exposure to
7
3
binding affinity (KD ≈ 1000 nM). Another region (amino
acids 77−82) contains an ETGE motif with a high binding
oxidative stress, Keap1 is deactivated. Nrf2 escapes Keap1-
mediated degradation, resulting in its nuclear translocation and
transcriptional activation of the ARE dependent genes,
including the above-mentioned entities. Thus, targeting the
Keap1−Nrf2−ARE signaling pathway is one logical strategy to
discover therapeutic agents for diseases and conditions
9
affinity (K ≈ 5.3 nM). On the basis of these two motifs,
D
several peptides disrupting the Keap1−Nrf2 interaction have
been developed. Unfortunately, their efficacy has not been
proven at the cellular level, probably due to poor cell
10
penetration.
4
Subsequently, the discovery of non-peptide Keap1−Nrf2
inhibitors has emerged. One recent study reported the
discovery of a potent small-molecule inhibitor 1 (K = 1.0 μM,
Figure 1) via a high throughput screening of NIH MLPCN
library of 337116 compounds (PubChem BioAssay ID: 504523,
504540) and hit optimization. This compound showed decent
involving oxidative stress. To date, most Nrf2 activators,
11
including natural products (e.g., sulforaphane and curcumin)
and synthetic compounds (e.g., oltipraz and bardoxolone
methyl), target this pathway through covalent modification
and deactivation of Keap1 protein via its reactive cysteine
i
4
,5
residues. Direct inhibition of the Keap1−Nrf2 protein−
protein interactions via a non-covalent mechanism is an
alternative for the discovery of small-molecule Nrf2 activators
Received: November 7, 2013
Published: January 13, 2014
4
with potential advantages such as lower toxicity.
©
2014 American Chemical Society
1121
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126

Journal of Medicinal Chemistry
Brief Article
6
5 compounds, representing 61 novel structural clusters with
Tanimoto coefficient <0.8, were purchased from the
commercial source (details in Supporting Information) for
biological evaluation.
Disrupting the Binding of Keap1−Nrf2. The inhibitory
activity (K ) of the 65 selected compounds from SBVS were
D2
18
measured by an established fluorescence anisotropy assay. On
the basis of its low micromolar inhibition activity, availability of
X-ray cocrystallographic structure, and ease of synthesis,
compound 3 was chosen as our reference compound. All
compounds were first evaluated at 100 μM, and 24 showed
detectable or decent Keap1−Nrf2 inhibition (Supporting
Information Table S1). These compounds were then evaluated
Figure 1. Recently published small molecule inhibitors of Keap1−Nrf2
interaction.
ARE-inducing activity with an EC50 of 18 μM in a cell-based
to determine the K . Nine possessed good inhibitory activity,
D2
12
functional assay. Another high throughput screening study
with KD2 ranging from 2.9 to 75.48 μM (Figure 2).
reported two novel scaffolds (2 and 3). Compound 3 (PDB
entry: 4IQK) showed a good binding inhibition (IC50 = 2.7
μM), while compound 2 (PDB entry: 4IN4) was rather weak
(
IC50 = 118 μM). Compound 3 was demonstrated to up-
regulate Nrf2 target gene NQO1 and stabilize Nrf2 protein
using an ARE-driven luciferase reporter system in DLD1 cells.
A small molecule 4 has been recently released in PDB bank
(
PDB entry: 3VNG, 3VNH). Comparison of these four crystal
structures showed that all ligands bind to the same protein−
protein interfacial pocket with minimal allosteric distortion of
the local side chains (Supporting Information Figure S3). This
provides a strong structural basis for in silico screening studies
by using a single X-ray crystallographic structure as a starting
point. Inhibitors with novel scaffolds are expected to comple-
ment current leads for future translational development.
Structure-based virtual screening (SBVS) and hit-based
substructure search (HBSS) are powerful tools for drug
1
3−15
discovery.
In this study, we successfully identified novel
Keap1−Nrf2 inhibitors with informative SARs by employing an
integrated SBVS and HBSS approach.
Figure 2. Chemical structures and Keap1−Nrf2 inhibitory activities of
RESULTS AND DISCUSSION
■
the hits from structure-based virtual screening.
Design Rationale. In this study, in-house synthesis was
replaced by taking the advantage that compounds in many
commercial chemical databases are inherently enriched with
analogues defined by the nature of their synthetic routes. Thus,
an SAR of the hits identified through SBVS can easily be
established through subsequent extraction of analogues of the
same chemotype within the same database. This approach is
economical and highly efficient to achieve a preliminary SAR
for large numbers of chemically diverse hits prior to focused
medicinal chemistry exploration. It is a standard best practice
for efficiently exploring commercial compound databases.
Structure-Based Virtual Screening. A schematic of the
overall virtual screening procedure in this study is presented in
Supporting Information Figure S1. Specs database (>300000
Interestingly, the phenylsulfonic amide turned out to be a
common scaffold (S47, S48, S49, S52, and S57), which was
different from the earlier reported inhibitors. Among the
identified hits, compounds S47 and S49 displayed excellent
Keap1−Nrf2 inhibitory activity, with K values of 2.9 and 3.6
D2
μM respectively, comparable to that of the reference compound
3 (KD2 = 3.3 ± 0.5 μM, Supporting Information Figure S12) in
our assay. Compound S6 and S40 showed moderate inhibitory
activity, with KD2 values of 15.2 and 10.4 μM. To obtain further
insight that would guide the hit-based substructure search, the
docking modes of these compounds with the Keap1 Kelch
domain were re-examined.
compounds) was processed, resulting in a screening library of
From the observed binding poses of our docking studies
(Figure 3 and Supporting Information S6), three recurring
features were observed that may play an important role in
selective binding of noncovalent Keap1 inhibitors, namely (a)
π−cation interaction of its aromatic scaffold with Arg415, (b)
hydrogen bonds with Ser363, Ser508, or Ser555, and (c) salt
bridge interaction between our negatively charged triazole or
acetate groups with surface Arg380, Arg415, or Arg483.
Interestingly, the most active and structurally similar com-
pounds, S47 and 3, involved a very different mode of binding
with deep insertion of their aromatic scaffold (cumenyl of S47
and naphthyl of compound 3) into the cationic Arg415 pocket.
6
1
53611 compounds with >10 conformations. Virtual screening
was carried out using Schrodinger’s Glide in three steps, each
with increasing level of computational sampling and precision,
resulting in 90 compounds with Glidescore less than that of
compound 3. While Glidescore performs poorly in the
estimation of binding affinity, it is commonly used as one
of the matrices for selecting poses in SBVS. To ensure
maximum chemical diversity and to limit redundancy in our
initial testing, these compounds were subjected to structure
similarity comparison and clustering based on their calculated
16
17
Tanimoto coefficient using the 2D fingerprint. The selected
1
122
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126

Journal of Medicinal Chemistry
Brief Article
Table 1. Inhibitory Activity of Compounds HS1−HS9 for
Keap1−Nrf2 Binding
Figure 3. Potential binding modes of the lead compounds (A) S6, (B)
S40, (C) S47, and (D) 3 with the Keap1 Kelch domain.
Table 2. Inhibitory Activity of Compounds HS10−HS21 for
Keap1−Nrf2 Binding
The differences are likely due to the fact that compound 3 is
structurally symmetric and eletroneutral while S47 is not. The
strong salt bridge interaction of S47 is also accompanied by
compensated replacement of the naphthyl group by the
cumenyl group inside Arg415 pocket (Figure 3C). The
structurally symmetric compound 3, on the other hand, binds
in a more symmetric way in relation to the 6-bladed fold
symmetry of the protein, involving two sulfonamide groups and
a 4-methoxyphenyl group, which formed additional hydrogen
bonds with Ser363 and Ser508 and π−π interactions with
Tyr334 to compensate for its potency (Figure 3D).
Hit-Based Substructure Search and Preliminary SARs.
Three scaffolds (S6, S40, and S47) representing novel
chemotypes for Keap1−Nrf2 inhibitors were selected for hit-
based substructure search in Specs database. Nine analogues of
S6, 13 analogues of S40, and 23 analogues of S47 were selected
and purchased for biological evaluation (25 of them have been
confirmed in-house for their identity and purity by NMR, MS,
and HPLC with details in Supporting Information).
The fluorescence anisotropy assay revealed that four
compounds (HS1, HS5, HS8, and HS9) within S6 series
(
Table 1) possessed moderate potency. SAR analysis revealed
that decreased inhibitory activity was observed for non-halogen
substituted derivative HS1 (KD2 = 42.9 ± 2.6 μM) compared to
the halogen-substituted one, S6 (KD2 = 15.2 ± 0.04 μM),
highlighting the importance of substitutions on R group. No
2
substitution on R (HS3) was totally inactive. Introducing
amide (HS7) or sulphonamide (HS2, HS4, and HS6) reduced
the inhibitory activity. Disubstitution on the same benzene ring
1
decreased activity compared to the monosubstituted derivative
S40. Removing either of the two benzene rings (HS20 and
HS21) showed decreased activity.
(S6) exhibited higher inhibition than the mono ones (HS8).
For the S40 series (Table 2), the carboxyl group on R was
The inhibitory activity of S47 analogues is summarized in
Tables 3 and 4. Compounds S47, S49, HS26, HS27, HS28,
HS31, HS39, and HS41 showed activity lower than 10 μM.
Notably, compounds S47, S49, HS26, HS31, and HS39 had
KD2 values around 3 μM, comparable to that of the reference
compound 3 (KD2 = 3.3 ± 0.5 μM). The key observations of
the SARs from these analogues have been summarized as
follows: generally, the triazole derivatives were more potent
than the carboxylates (Table 3). Introducing substituents on
the NH group of triazole (HS33) or replacing the whole
triazole by benzene rings (HS34−HS35) or other aromatic
1
critical for the inhibitory activity, likely because it may form
hydrogen bonds with Arg380 and Arg415. The corresponding
hydroxyl (HS18) or methoxy groups (HS12 and HS19) were
completely inactive. Compounds with various substitutions on
R2 position also showed quite different inhibitory activity. The
methyl derivatives (S40, HS10, HS11, and HS13) were better
than the halogen derivatives (HS15 and HS16). When the
methyl group on position 2 (S40) was moved to position 3
(
HS10), decreased activity was observed. Similarly, disubstitu-
tion on the same benzene ring (HS11, HS13, and HS14) led to
1
123
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126

Journal of Medicinal Chemistry
Brief Article
Table 3. Inhibitory Activity of Compounds HS22−HS38 for
Keap1−Nrf2 Binding
Table 4. Inhibitory Activity of Compounds HS39−HS45 for
Keap1−Nrf2 Binding
potentially due to its insufficient occupation of the binding
site. Removing one aromatic ring of the naphthyl (S57 and
HS45) was unfavorable.
The most potent compounds (S6, S40, S47, and HS26) of
each class were then confirmed to bind with Keap1 protein
non-covalently in a reversible binding assay by reproducing
similar binding affinities of dialyzed Keap1 to its fluorescence
probe (Supporting Information Figure S13).
Differential Scanning Fluorimetry (DSF) Assay. The
affinity of selected compounds to Keap1 protein was then
confirmed by a qualitative orthogonal DSF assay at a single
concentration used in previous validation studies for Keap1−
rings (HS32 and HS36−HS38) dramatically reduced the
inhibitory activity, presumably due to the loss of the salt bridge
interaction with Arg483. A hydroxyl group was critical for the
inhibitory activity because of the hydrogen bond interaction
with the backbone of Ser555. Thus, acetylation of the hydroxyl
group (HS43) or cyclization with sulfur (HS44) rendered the
compounds completely inactive. As the compound with
isopropylphenyl group (S47) could be inserted deeply into
the central cavity of Keap1 protein, the derivative with no
substitution on the benzene ring (HS30) had a poor inhibition
1
9,20
Nrf2 inhibition.
One advantage of DSF is that it can be
done with higher throughput without requiring large amounts
of protein. As expected, compounds S47 and HS31 showed
significant decrease (ΔT = 5 and 4 °C respectively) in melting
m
temperatures of Keap1 protein, while the reference compound
3 was not very sensitive in this assay (Figure 4 and Supporting
Information Figure S14). The data also showed S6 and S40
with moderate potency only had 1 or 2 °C difference compared
to DMSO control, while the negative control compounds
HS32 and HS44 were inactive. These results overall confirmed
the binding interaction of the lead compounds with Keap1
protein
(
KD2 = 61.2 ± 5.4 μM). Introducing a methoxy group on
position 4 (HS28) led to a more than 10-fold increase in
activity. Ethoxyl derivative (HS31) also enhanced the activity
with a KD2 value of 3.1 ± 0.3 μM.
The compounds with aliphatic groups (S47, S49, S52, HS25,
HS27−HS28, and 31) on the benzene ring were more potent
than the halogen substituted ones (HS22−HS24, HS29). Tri-
or disubstituted analogues were more potent than the
Quantification of mRNA of Nrf2 Downstream Genes.
To determine whether the compounds activate Nrf2, qRT-PCR
was carried out to quantify the mRNA levels of HO-1 and
NQO-1 using PC12 cells (Figure 5). HO-1 mRNA was
significantly enhanced by compound S47 in dose−response
and time−response manners with no obvious effect by the
reference compound 3. Noticeably, the HO-1 mRNA was three
times of that of compound 3 at 4.5 and 6 h. For NQO-1
mRNA, it was elevated at both doses of S47 but no change at
monosubstituted ones, such as HS26 (2,4,6-trimethyl, KD2
=
4.2 ± 0.4 μM), HS25 (4-methyl, KD2 = 15.4 ± 0.9 μM), and
HS27 (4-isopropyl, KD2 = 6.4 ± 1.4 μM). Naphthyl (HS39)
and quinolyl (HS40) were tolerated with similar activity
compared with the phenyl ones (Table 4). The thienyl
compound (HS42) was 10-fold less potent, probably
1
124
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126

Journal of Medicinal Chemistry
Brief Article
Nrf2. Focused medicinal chemistry work is ongoing to expand
the SAR of novel inhibitors.
EXPERIMENTAL SECTION
Fluorescent Anisotropy Assay. Determination of equilibrium
■
dissociation constants K for each compound was performed using the
D
fluorescence anisotropy assay with FITC−βAla−DEETGEF−OH
10
(Supporting Information Figure S10) as the fluorescence probe.
The fluorescence anisotropy was measured on SpectraMax M5e
microplate reader with 485 nm excitation and 535 nm emission after a
6
0 min incubation at room temperature. The KD2 of each compound
tested was determined by fitting the displacement curves as described
in the literature (details can be found in Supporting Information).
1
8
Figure 4. Thermal shifts of Keap1 protein with selected compounds
(
100 μM in 0.4% DMSO) determined by DSF assay.
ASSOCIATED CONTENT
Supporting Information
■
*
S
Computational protocols, FITC−βAla−DEETGEF−OH pro-
tein sequence and characterization, dose−response curves for
inhibition of Keap1−Nrf2 binding of the hits, synthesis of the
reference compound, NMR, MS spectra, and HPLC purity of
selected compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
For C.X.: phone, 612-626-5675; fax, 612-624-0139; E-mail,
■
*
Figure 5. The mRNA levels of downstream target genes of Nrf2
measured by qRT-PCR in PC12 cells: (A) HO-1, (B) NQO-1.
xingx009@umn.edu.
For Y.Y.S.: phone, 612-625-8255; fax, 612-625-8154; E-mail,
shamx002@umn.edu.
low dose of compound 3. These results mechanistically
supported our hits disrupting the protein−protein interaction
of Keap1 and Nrf2.
Western Blotting Assay for Nrf2 Nuclear Trans-
location. To further validate the mechanism of action,
Western blotting was used to characterize the subcellular
distribution of Nrf2 protein in PC12 cells. Given that
downstream gene up-regulation was observed at 4.5 h (Figure
*
Author Contributions
Ideas and experiment design: C. Zhuang, Y. Y. Sham, C. Xing.
Computational development: C. Zhuang, Y. Y. Sham.
Chemistry and Biology: C. Zhuang, S. Narayanapillai, C.
Xing. Analysis and data interpretation: C. Zhuang, Y. Y. Sham,
C. Xing. Writing and review of the manuscript: C. Zhuang, S.
Narayanapillai, W. Zhang, Y. Y. Sham, C. Xing. Study
supervision: Y. Y. Sham, C. Xing.
5
), Nrf2 nuclear translocation should occur before that. Indeed,
S47 treatment caused a significant decrease in the level of
cytosolic Nrf2 protein and an increase in nuclear Nrf2 protein
Notes
1
.5 h after treatment (Figure 6). The level of Nrf2 protein was
The authors declare no competing financial interest.
reversed after 4.5 h treatment, likely because of the feedback
loop of Nrf2 and Keap1 regulation.
21
ACKNOWLEDGMENTS
■
We thank the Department of Medicinal Chemistry, Minnesota
Supercomputing Institute, and the Center for Drug Design for
providing the necessary resources. We thank Ran Dai and Dr.
Barry Finzel at the University of Minnesota for assistance of
DSF assay. We are also grateful to the China Scholarship
Council’s Ph.D. Abroad Training Plan.
Figure 6. Western blotting of S47 (4 μM) induced Nrf2 nuclear
translocation in PC12 cells.
ABBREVIATIONS USED
■
NQO-1, NADPH:quinone oxidoreductase 1; HO-1, heme
oxygenase-1; SOD, superoxide dismutase; GST, glutathione S-
transferase; GPx, glutathione peroxidase; Keap1, Kelch-like
ECH-associated protein 1; Nrf2, nuclear factor erythroid 2-
related factor 2; ARE, antioxidant response elements; SBVS,
structure-based virtual screening; HBSS, hit-based substructure
search; SAR, structure−activity relationship; DSF, differential
scanning fluorimetry
CONCLUSION
■
Three classes of novel inhibitors disrupting Keap1−Nrf2
protein−protein interaction were successfully identified by
structure-based virtual screening. Further hit-based substructure
search of the potent compounds concluded an informative
SAR, which was a high-efficiency strategy and best use of the
commercial compound library. Our most potent noncovalent
inhibitor exhibits three times improved cellular activation of
Nrf2 than the most active noncovalent Keap1 inhibitor known
to date. Further cell-based mechanism study supported our new
hits disrupting the protein−protein interaction of Keap1 and
REFERENCES
■
(
1) Uttara, B.; Singh, A. V.; Zamboni, P.; Mahajan, R. T. Oxidative
stress and neurodegenerative diseases: a review of upstream and
downstream antioxidant therapeutic options. Curr. Neuropharmacol.
2009, 7, 65−74.
1
125
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126

Journal of Medicinal Chemistry
Brief Article
(
2) Dinkova-Kostova, A. T.; Talalay, P. Direct and indirect
(19) Niesen, F. H.; Berglund, H.; Vedadi, M. The use of differential
scanning fluorimetry to detect ligand interactions that promote protein
stability. Nature Protoc. 2007, 2, 2212−2221.
antioxidant properties of inducers of cytoprotective proteins. Mol.
Nutr. Food Res. 2008, 52 (Suppl 1), S128−S138.
(
3) Itoh, K.; Wakabayashi, N.; Katoh, Y.; Ishii, T.; Igarashi, K.; Engel,
(20) Wang, L.; Lewis, T.; Zhang, Y. L.; Khodier, C.; Magesh, S.;
Chen, L.; Inoyama, D.; Chen, Y.; Zhen, J.; Hu, L.; Beamer, L. J.;
Faloon, P. W.; Dandapani, S.; Perez, J. R.; Munoz, B.; Palmer, M.;
Schreiber, S. The identification and characterization of non-reactive
inhibitor of Keap1−Nrf2 interaction through HTS using a
fluorescence polarization assay. In Probe Reports from the NIH
Molecular Libraries Program; National Institutes of Health: Bethesda,
MD, 2010.
J. D.; Yamamoto, M. Keap1 represses nuclear activation of antioxidant
responsive elements by Nrf2 through binding to the amino-terminal
Neh2 domain. Genes Dev. 1999, 13, 76−86.
(
4) Magesh, S.; Chen, Y.; Hu, L. Small molecule modulators of
Keap1−Nrf2−ARE pathway as potential preventive and therapeutic
agents. Med. Res. Rev. 2012, 32, 687−726.
(
and molecular regulation of enzymes that detoxify carcinogens. Proc.
Natl. Acad. Sci. U. S. A. 1993, 90, 2965−2969.
(
5) Prestera, T.; Holtzclaw, W. D.; Zhang, Y.; Talalay, P. Chemical
(21) Lee, O. H.; Jain, A. K.; Papusha, V.; Jaiswal, A. K. An auto-
regulatory loop between stress sensors INrf2 and Nrf2 controls their
cellular abundance. J. Biol. Chem. 2007, 282, 36412−36420.
6) Wakabayashi, N.; Dinkova-Kostova, A. T.; Holtzclaw, W. D.;
Kang, M. I.; Kobayashi, A.; Yamamoto, M.; Kensler, T. W.; Talalay, P.
Protection against electrophile and oxidant stress by induction of the
phase 2 response: fate of cysteines of the Keap1 sensor modified by
inducers. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 2040−2045.
(
7) McMahon, M.; Thomas, N.; Itoh, K.; Yamamoto, M.; Hayes, J. D.
Redox-regulated turnover of Nrf2 is determined by at least two
separate protein domains, the redox-sensitive Neh2 degron and the
redox-insensitive Neh6 degron. J. Biol. Chem. 2004, 279, 31556−
31567.
(
8) Lo, S. C.; Li, X.; Henzl, M. T.; Beamer, L. J.; Hannink, M.
Structure of the Keap1:Nrf2 interface provides mechanistic insight into
Nrf2 signaling. EMBO J. 2006, 25, 3605−3617.
(
9) Tong, K. I.; Katoh, Y.; Kusunoki, H.; Itoh, K.; Tanaka, T.;
Yamamoto, M. Keap1 recruits Neh2 through binding to ETGE and
DLG motifs: characterization of the two-site molecular recognition
model. Mol. Cell. Biol. 2006, 26, 2887−2900.
(
10) Inoyama, D.; Chen, Y.; Huang, X.; Beamer, L. J.; Kong, A. N.;
Hu, L. Optimization of fluorescently labeled Nrf2 peptide probes and
the development of a fluorescence polarization assay for the discovery
of inhibitors of Keap1−Nrf2 interaction. J. Biomol. Screening 2012, 17,
4
(
35−447.
11) Hu, L.; Magesh, S.; Chen, L.; Wang, L.; Lewis, T. A.; Chen, Y.;
Khodier, C.; Inoyama, D.; Beamer, L. J.; Emge, T. J.; Shen, J.;
Kerrigan, J. E.; Kong, A. N.; Dandapani, S.; Palmer, M.; Schreiber, S.
L.; Munoz, B. Discovery of a small-molecule inhibitor and cellular
probe of Keap1−Nrf2 protein−protein interaction. Bioorg. Med. Chem.
Lett. 2013, 23, 3039−3043.
(
12) Marcotte, D.; Zeng, W.; Hus, J. C.; McKenzie, A.; Hession, C.;
Jin, P.; Bergeron, C.; Lugovskoy, A.; Enyedy, I.; Cuervo, H.; Wang, D.;
Atmanene, C.; Roecklin, D.; Vecchi, M.; Vivat, V.; Kraemer, J.;
Winkler, D.; Hong, V.; Chao, J.; Lukashev, M.; Silvian, L. Small
molecules inhibit the interaction of Nrf2 and the Keap1 Kelch domain
through a non-covalent mechanism. Bioorg. Med. Chem. 2013, 21,
4
(
2
(
011−4019.
13) Shoichet, B. K. Virtual screening of chemical libraries. Nature
004, 432, 862−865.
14) Zhuang, C.; Miao, Z.; Zhu, L.; Dong, G.; Guo, Z.; Wang, S.;
Zhang, Y.; Wu, Y.; Yao, J.; Sheng, C.; Zhang, W. Discovery, synthesis,
and biological evaluation of orally active pyrrolidone derivatives as
novel inhibitors of p53−MDM2 protein−protein interaction. J. Med.
Chem. 2012, 55, 9630−9642.
(
15) Muegge, I. Synergies of virtual screening approaches. Mini-Rev.
Med. Chem. 2008, 8, 927−933.
16) Enyedy, I. J.; Egan, W. J. Can we use docking and scoring for
hit-to-lead optimization? J. Comput.-Aided Mol. Des. 2008, 22, 161−
68.
17) Maggiora, G.; Vogt, M.; Stumpfe, D.; Bajorath, J. Molecular
(
1
(
Similarity in Medicinal Chemistry. J. Med. Chem. 2013, DOI: 10.1021/
jm401411z.
(
18) Roehrl, M. H.; Wang, J. Y.; Wagner, G. A general framework for
development and data analysis of competitive high-throughput screens
for small-molecule inhibitors of protein−protein interactions by
fluorescence polarization. Biochemistry 2004, 43, 16056−16066.
1
126
dx.doi.org/10.1021/jm4017174 | J. Med. Chem. 2014, 57, 1121−1126