Nikolaos Georgakopoulos1,2, Sandeep Talapatra1, Dina Dikovskaya3, Sharadha Dayalan Naidu3, Maureen Higgins3, Jemma Gatliff1,2, Aysel Ayhan1, Roxani Nikoloudaki1,2, Marjolein Schaap1, Klara Valko1,4, Farideh Javid5, Albena T Dinkova-Kostova3,6, Frank Kozielski1, Geoffrey Wells1. 1. UCL School of Pharmacy, University College London, 29/39 Brunswick Square, London WC1N 1AX, U.K. 2. Stevenage Bioscience Catalyst, Keregen Therapeutics Ltd., Gunnels Wood Rd, Stevenage SG1 2FX, U.K. 3. Jacqui Wood Cancer Centre, Division of Cellular Medicine, School of Medicine, University of Dundee, Dundee DD1 9SY, Scotland, U.K. 4. Bio-Mimetic Chromatography Consultancy, 17 Cabot Close, Stevenage SG2 0ES, U.K. 5. Department of Pharmacy, University of Huddersfield, Queensgate, Huddersfield HD1 3DH, U.K. 6. Department of Pharmacology and Molecular Sciences and Department of Medicine, School of Medicine, Johns Hopkins University, Baltimore, Maryland 21205, United States.
Abstract
Inhibitors of Kelch-like ECH-associated protein 1 (Keap1) increase the activity of the transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) by stalling its ubiquitination and degradation. This enhances the expression of genes encoding proteins involved in drug detoxification, redox homeostasis, and mitochondrial function. Nrf2 activation offers a potential therapeutic approach for conditions including Alzheimer's and Parkinson's diseases, vascular inflammation, and chronic obstructive airway disease. Non-electrophilic Keap1-Nrf2 protein-protein interaction (PPI) inhibitors may have improved toxicity profiles and different pharmacological properties to cysteine-reactive electrophilic inhibitors. Here, we describe and characterize a series of phenyl bis-sulfonamide PPI inhibitors that bind to Keap1 at submicromolar concentrations. Structural studies reveal that the compounds bind to Keap1 in a distinct "peptidomimetic" conformation that resembles the Keap1-Nrf2 ETGE peptide complex. This is different to other small molecule Keap1-Nrf2 PPI inhibitors, including bicyclic aryl bis-sulfonamides, offering a starting point for new design approaches to Keap1 inhibitors.
Inhibitors of Kelch-like ECH-associated protein 1 (Keap1) increase the activity of the transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) by stalling its ubiquitination and degradation. This enhances the expression of genes encoding proteins involved in drug detoxification, redox homeostasis, and mitochondrial function. Nrf2 activation offers a potential therapeutic approach for conditions including Alzheimer's and Parkinson's diseases, vascular inflammation, and chronic obstructive airway disease. Non-electrophilic Keap1-Nrf2 protein-protein interaction (PPI) inhibitors may have improved toxicity profiles and different pharmacological properties to cysteine-reactive electrophilic inhibitors. Here, we describe and characterize a series of phenyl bis-sulfonamide PPI inhibitors that bind to Keap1 at submicromolar concentrations. Structural studies reveal that the compounds bind to Keap1 in a distinct "peptidomimetic" conformation that resembles the Keap1-Nrf2 ETGE peptide complex. This is different to other small molecule Keap1-Nrf2 PPI inhibitors, including bicyclic aryl bis-sulfonamides, offering a starting point for new design approaches to Keap1 inhibitors.
The transcription factor
Nrf2 (nuclear factor-erythroid 2-related
factor 2) is a key regulator protein involved in adaptive responses
to internal and external stress stimuli, particularly oxidative damage.
Nrf2 augments the expression of a wide range of cyto-protective proteins
by binding with high affinity to the antioxidant response elements
that are located in their promoter gene regions. Gene products regulated
by Nrf2 include proteins that maintain a reducing environment within
cells, phase I and II metabolic enzymes, anti-inflammatory mediators,
and proteins associated with general and selective types of autophagy.[1]The transcriptional activity of Nrf2 is
controlled by several regulatory
mechanisms, among which the Keap1 (Kelch-like ECH associating protein
1)/Cullin3 system is the most extensively studied. Keap1 is a homodimeric
protein that physically interacts through its Kelch domain with Nrf2
in a 2:1 stoichiometry.[2] Mutagenesis and
deletion analysis revealed that this interaction is mediated by two
discrete regions within the Neh2 (Nrf2–ECH homology 2) domain
of Nrf2 that consists of a low affinity 29DLG31 and a high affinity 79ETGE82 binding motif,
respectively.[3,4] This two-site recognition mechanism
allows Keap1 to suppress the activity of Nrf2 by facilitating its
ubiquitination and subsequent proteosomal degradation via a pendant
Cul3-based Rbx1 (ring-box 1) E3 ubiquitin ligase complex. However,
upon oxidative or electrophilic stress the repressor activity of Keap1
is impaired, allowing Nrf2 to evade degradation.[4] The stabilization of Nrf2 under such conditions stems from
the redox or covalent modification of Keap1 cysteine residues, which
induces conformational changes in the structure of Keap1 that diminish
its ubiquitin-facilitating activity.Given its pivotal role
in adjusting cellular defences under stress
conditions and during hormesis, Nrf2 has been proposed to hold great
promise as a drug target for the prevention or treatment of a wide
range of pathological and chronic conditions, including but not limited
to inflammatory conditions, neurodegenerative and autoimmune disorders
and cardiovascular diseases.[1,5−7] Compounds that inhibit Keap1 have been shown to enhance the transcriptional
activity of Nrf2 and have therefore been sought extensively as potential
therapeutic agents.Most Nrf2 inducers are electrophiles that
exert their effects by
covalently modifying sensor Keap1 cysteines and subsequently, halting
the ubiquitination and proteasomal targeting of the transcription
factor. Given their highly reactive nature however, such agents generally
demonstrate unspecific biological effects and interference with other
redox sensitive signaling pathways. Direct inhibition of the Keap1-Nrf2
protein–protein interaction (PPI) is an alternative approach
to inhibit the degradation of Nrf2 in a reversible and non-covalent
manner. We and others have described the development of peptides based
on the ETGE and DLG binding motifs of Nrf2 that compete with the transcription
factor for binding to the Kelch domain of Keap1.[8−11] More recently, a series of small
molecule modulators of the Keap1-Nrf2 PPI have been described,[12−22] and Keap1 was the focus of an ultralarge virtual screen of more
than 1 billion compounds.[23] Although in
the latter case promising levels of both in vitro and cellular potency
have been achieved, there is a relatively limited degree of structural
diversity among the different classes of PPI inhibitors reported.
A number of PPI inhibitors described to date share a common privileged
scaffold comprising a bicyclic ring system, usually a naphthalene
ring flanked by one or in most cases two benzenesulfonamide moieties
(Figure ). According
to the available crystallographic and molecular modeling data, the
bicyclic ring occupies the P3 sub-pocket of the Keap1 Kelch binding
site and enables the projection of the benzenesulfonamide group(s)
into the P4 and/or P5 sub-pockets. Previous attempts to replace the
naphthalene group with a single ring system have resulted in a reduced
Keap1 binding affinity, highlighting the favorable Keap1 binding properties
of this scaffold.[14,24] These can be attributed to the
formation of a network of π–cation and π–π
stacking interactions of the naphthalene ring with the guanidino group
of Arg415 and the benzenesulfonamide moieties with the aromatic side
chains of Tyr334, Tyr572 and Tyr525. Recent studies have revealed
that, in contrast to electrophiles, which impair the substrate adaptor
activity of Keap1, leading to the accumulation of newly synthesized
Nrf2, such non-electrophilic PPI inhibitors disrupt the binding of
Keap1 to the DLG motif of Nrf2 in preference to Keap1-ETGE binding.[25,26]
Figure 1
(A)
Structure of the Keap1 Kelch domain (PDB entry 4IQK) shown as a surface
representation, the sub-pockets of the Nrf2 binding site are shown:
P1 (maroon residues 415, 461, 462, 478, 483, and 508), P2 (green residues,
363, 380, 381, and 414), P3 (blue residues, 364, 509, 556, 571, 602,
and 603), P4 (yellow residues, 334, 572, and 577), and P5 (purple
residues, 525, 530, and 555); (B) the structures of published Keap1-Nrf2
PPI inhibitors containing a naphthalene central ring system. FP values
for compounds 1–8 are IC50s determined using FP; SA values are Kds from a secondary assay; (a) Kd determined
using SPR assay; (b) Kd determined using
ITC; (c) fluorescence anisotropy assay; and (d) not determined (DSF
thermal shift = −5 °C).[13,14,17,18,24,27]
(A)
Structure of the Keap1 Kelch domain (PDB entry 4IQK) shown as a surface
representation, the sub-pockets of the Nrf2 binding site are shown:
P1 (maroon residues 415, 461, 462, 478, 483, and 508), P2 (green residues,
363, 380, 381, and 414), P3 (blue residues, 364, 509, 556, 571, 602,
and 603), P4 (yellow residues, 334, 572, and 577), and P5 (purple
residues, 525, 530, and 555); (B) the structures of published Keap1-Nrf2
PPI inhibitors containing a naphthalene central ring system. FP values
for compounds 1–8 are IC50s determined using FP; SA values are Kds from a secondary assay; (a) Kd determined
using SPR assay; (b) Kd determined using
ITC; (c) fluorescence anisotropy assay; and (d) not determined (DSF
thermal shift = −5 °C).[13,14,17,18,24,27]Here, we report the development and biological characterization
of a novel Keap1-interactive scaffold identified through a structure-based
ligand design approach. Although structurally similar to previously
described phenyl bis-sulfonamides such as 9, several
of the new analogues exhibit sub-micromolar affinity for the Kelch
domain of Keap1 as determined by fluorescence polarization (FP) and
isothermal titration calorimetry (ITC) analyses.[14] Despite the presence of two carboxylate groups, selected
compounds interact with Keap1 and induce the transcriptional activity
of Nrf2 in cells at micromolar concentrations. The improved binding
affinity and inducer activity may be explained by a high-resolution
co-crystal structure of analogue 11 bound to the Keap1
Kelch domain that reveals a new and unexpected binding mode that closely
resembles the Keap1-Nrf2 ETGE peptide complex and is distinct from
the binding mode of any other known aryl bis-sulfonamides, which preferentially
disrupts the binding of Keap1 to Nrf2-DLG.[25] Thus, the new series of Keap1-Nrf2 PPI provide an alternative starting
point for future optimization that exploits this new binding mode.
Results
and Discussion
Molecular Modeling and Ligand Design
Although several
studies describing the structure–activity relationships (SARs)
around the phenylsulfonamide[14,17] and acetate[28] portions of 4 have been published,
there is less information about the structural requirements for the
central ring system of this scaffold. Replacement of the naphthalene
with related bicyclic heterocycles such as isoquinolines, for example, 6 gives ligands of similar binding affinity and improved properties,
but other heterocycles had reduced binding affinity.[24] Recent reports showed that replacement of the naphthalene
core of 4 with a phenyl group reduces binding affinity;
however, 9 retains only low micromolar affinity for Keap1
that can be moderately improved by introducing a methoxy substitution
at the 2-position of the phenyl ring (Figure A, compound 10).[14,24] We carried out a molecular docking study using the Keap1 Kelch domain
structure derived from the co-crystal structure with 1 (PDB entry 4IQK, Figure B,C). We
confirmed that redocking 1 into the vacant Kelch binding
pocket gave a comparable bound conformation and that 4 docked in the expected manner (Figure D,E, respectively) to make additional polar
contacts with the arginine triad, residues 380, 415, and 483. Next,
we docked the ligands 11 and 12 in an analogous
manner (Figure F,G).
These new ligands have a central phenyl ring and either a 1,4- or
1,3-substitution pattern for the sulfonamide moieties. The docked
conformations suggested that the bis-sulfonamide was positioned in
a similar manner to those of 1 and 4 and
that the respective benzyloxy and carboxamide substituents of 11 and 12 project into the central channel through
the β-propeller structure, indicating that they may occupy the
P3 sub-pocket of Keap1 more effectively. These encouraging preliminary
data prompted the synthesis of a series of analogues based on the
selected scaffolds with the aim of determining the structural requirements
for interaction with the Keap1 Kelch domain central pore.
Figure 2
(A) Structures
of compounds 9–12; (B,C) co-crystallized
structure of 1 bound to the
Keap1 Kelch domain (PDB entry 4IQK); (D–G) docked conformations of
compounds 1, 4, 11, and 12, respectively, bound to the corresponding 4IQK Keap1 Kelch domain.
(A) Structures
of compounds 9–12; (B,C) co-crystallized
structure of 1 bound to the
Keap1 Kelch domain (PDB entry 4IQK); (D–G) docked conformations of
compounds 1, 4, 11, and 12, respectively, bound to the corresponding 4IQK Keap1 Kelch domain.
Chemical Synthesis
The synthesis
of compounds 31–36 was performed
following the general
procedure described in Scheme . The nitro groups of 13–18 were reduced with either Fe/AcOH or SnCl2 and the resulting
intermediates were reacted with an excess of 4-methoxybenzenesulfonyl
chloride and pyridine to afford the bis-sulfonamides 19–24. N-alkylation with ethyl bromoacetate and
K2CO3 followed by saponification of the resulting
ethyl esters 25–30 gave the bis-acids 10–11 and 31–34 in reasonable overall yields.
Scheme 1
Synthetic Route to
Compounds 10–11 and 31–34
Reagents and conditions: (a)
Fe, AcOH, EtOH/H2O, 65 °C, 2 h or SnCl2·2H2O, EtOH, 60 °C, 2 h; (b) 4-methoxybenzenesulfonyl
chloride, pyridine, DMAP, DCM, rt, o/n; (c) ethyl bromoacetate, K2CO3, DMF, rt, 6 h; and (d) 1 N NaOH, THF//MeOH
(1:1), rt, 2 h.
Synthetic Route to
Compounds 10–11 and 31–34
Reagents and conditions: (a)
Fe, AcOH, EtOH/H2O, 65 °C, 2 h or SnCl2·2H2O, EtOH, 60 °C, 2 h; (b) 4-methoxybenzenesulfonyl
chloride, pyridine, DMAP, DCM, rt, o/n; (c) ethyl bromoacetate, K2CO3, DMF, rt, 6 h; and (d) 1 N NaOH, THF//MeOH
(1:1), rt, 2 h.To further supplement our
compound library, we prepared compounds 36–39 from a common precursor (compound 28) following
the alternative and shorter synthetic sequence
outlined in Scheme . The benzyl ether of 28 was removed by catalytic transfer
hydrogenation, affording the phenol 35 in 61% yield. 35 was subsequently subjected to saponification to furnish 39, while the remaining analogues 36–39 were prepared by the alkylation of the phenol 35 with appropriate halides, followed by hydrolysis of the ethyl ester
protecting groups.
Scheme 2
Synthetic Route to Compounds 35–39
Reagents and conditions: (a)
triethylsilane, 5% Pd/C, EtOAc/EtOH (9:1 v/v), rt, 2 h, 61%; (b) R–X,
K2CO3, acetone, rt, o/n; and (c) 1 N NaOH, THF//MeOH
(1:1), rt, 2 h.
Synthetic Route to Compounds 35–39
Reagents and conditions: (a)
triethylsilane, 5% Pd/C, EtOAc/EtOH (9:1 v/v), rt, 2 h, 61%; (b) R–X,
K2CO3, acetone, rt, o/n; and (c) 1 N NaOH, THF//MeOH
(1:1), rt, 2 h.Compounds bearing carboxylate
bioisosteric replacements were prepared
according to the synthetic procedures shown in Scheme . The bis-amide 40 was synthesized
from 11 by treatment with Boc2O and ammonium
carbonate in DMF/pyridine. N-alkylation of 22 with bromoacetonitrile
gave 41 and then cycloaddition with NaN3 afforded
the bis-1H-tetrazole 42. The remaining
acid/1H-tetrazole analogues 51 and 52 were prepared from 16 via sequential assembly
of the tertiary sulfonamide moiety on each side of the central phenyl
ring, followed by [2 + 3] cycloaddition with sodium azide and subsequent
ethyl ester hydrolysis.
Scheme 3
Synthetic Route to Compounds 40–42, 51, and 52
Reagents and conditions: (a)
Boc2O, (NH4)2CO3, pyridine,
DMF, Ar, 0 °C to rt, 18 h, 74%; (b) bromoacetonitrile, K2CO3, DMF, rt, 8 h, 59%; (c) NaN3, NH4Cl, DMF, 100 °C, o/n; (d) 4-methoxybenzenesulfonyl chloride,
DCM, pyridine, rt, o/n; (e) ethyl bromoacetate, K2CO3, DMF, rt, 8 h; (f) SnCl2·2H2O,
EtOH, 65 °C, 1 h; and (g) 1 N NaOH, THF/MeOH (1:1 v/v), 2 h.
Synthetic Route to Compounds 40–42, 51, and 52
Reagents and conditions: (a)
Boc2O, (NH4)2CO3, pyridine,
DMF, Ar, 0 °C to rt, 18 h, 74%; (b) bromoacetonitrile, K2CO3, DMF, rt, 8 h, 59%; (c) NaN3, NH4Cl, DMF, 100 °C, o/n; (d) 4-methoxybenzenesulfonyl chloride,
DCM, pyridine, rt, o/n; (e) ethyl bromoacetate, K2CO3, DMF, rt, 8 h; (f) SnCl2·2H2O,
EtOH, 65 °C, 1 h; and (g) 1 N NaOH, THF/MeOH (1:1 v/v), 2 h.The related 3,5-bis-sulfonamide-N-substituted-benzamides 12 and 58–61 were prepared
from the readily available 3,5-diaminobenzoic acid following a similar
synthetic procedure (Scheme ). Sulfonylation of 53 using our standard conditions,
followed by amidation with EDCI·HCl and DMAP afforded intermediates 54 and 55–57, respectively.
Compounds 12 and 58–61 were synthesized by nucleophilic substitutions using the appropriate
alkyl halides, followed by the removal of the protecting ethyl ester
groups for compounds 12 and 58–59.
Scheme 4
Synthetic Route to Compounds 12 and 58–61
Reagents and conditions:
(a)
4-R1-PhSO2Cl, pyridine, DMAP, DMF, 100 °C,
o/n; (b) amine, EDCI·HCl, DMAP, DMF, o/n; (c) bromoacetonitrile
or bromoacetamide, K2CO3, DMF, rt, o/n; (d)
ethyl bromoacetate, K2CO3, DMF, rt, o/n; and
(e) 1 N NaOH, THF/MeOH (1:1 v/v), o/n.
Synthetic Route to Compounds 12 and 58–61
Reagents and conditions:
(a)
4-R1-PhSO2Cl, pyridine, DMAP, DMF, 100 °C,
o/n; (b) amine, EDCI·HCl, DMAP, DMF, o/n; (c) bromoacetonitrile
or bromoacetamide, K2CO3, DMF, rt, o/n; (d)
ethyl bromoacetate, K2CO3, DMF, rt, o/n; and
(e) 1 N NaOH, THF/MeOH (1:1 v/v), o/n.
Structure–Activity
Relationships
The compounds
were screened for their ability to disrupt the interaction of the
Keap1 Kelch domain with a fluorescent peptide derived from the high
affinity ETGE binding motif of Nrf2 (FITC-β-DEETGEF-OH) in an
FP competition assay.[8] Generally, a variety
of substitutions were tolerated at the 2-position of the central phenyl
ring (Table ). The
phenol 39 demonstrated reasonable inhibition of the Keap1-Nrf2
PPI, while capping the phenolic OH with a methyl group (compound 10) did not affect the binding affinity. Replacing the 2-methoxy
substitution with an ethoxy (compound 31) led to improved
activity; however, increasing the size of the 2-alkoxy substituent
from ethoxy to n-propyloxy (compound 36) was not beneficial for activity. A complete loss of binding affinity
was recorded for branched isopropyloxy 37, presumably
due to the introduction of an unfavorable steric clash with the Keap1
binding pocket. Interestingly, a compromise in activity was achieved
by further extending the length of the 2-alkoxy moiety (n-butyl, compound 32), suggesting that introduction of
bulkier substituents at this position could lead to an improved binding
profile. In support of this notion, the benzyloxy analogue 11 demonstrated submicromolar activity in the FP assay with an IC50 value of 0.575 μM. Removing the benzylic methylene
bridge of 11 (compound 33) decreased the
activity by threefold, while further simplification of the structure
by replacing the ether substituent with a phenyl ring (compound 34) further diminished binding affinity. On the other hand,
functionalization of the benzyloxy group by introducing a meta-methoxy substitution was tolerated (compound 38) and resulted in a modest loss of binding activity.
Table 1
SARs and Physicochemical Properties
for Compounds 10–11, 31–34, 36–42, 51, and 52
compd
R1
R2
R3
FP IC50 (μM)
DSF (ΔTi °C)a
NQO1-fold induction
at 10 μMb
10
OMe
CO2H
CO2H
1.84 ± 0.22
2.4 ± 0.3
0.94 ± 0.13
31
OEt
CO2H
CO2H
0.815 ± 0.060
2.4 ± 0.3
1.65 ± 0.22
32
OnBu
CO2H
CO2H
0.677 ± 0.059
3.1 ± 0.2
1.66 ± 0.08
11
OBn
CO2H
CO2H
0.575 ± 0.072
3.3 ± 0.4
1.71 ± 0.09
33
OPh
CO2H
CO2H
1.50 ± 0.22
3.1 ± 0.4
0.92 ± 0.11
34
Ph
CO2H
CO2H
51%c
0.2 ± 0.3
1.08 ± 0.04
36
OnPr
CO2H
CO2H
1.19 ± 0.23
2.9 ± 0.2
1.29 ± 0.22
37
OiPr
CO2H
CO2H
<25%c
0.1 ± 0.3
1.09 ± 0.07
38
OBn(3-OMe)
CO2H
CO2H
0.805 ± 0.021
3.4 ± 0.3
1.84 ± 0.05
39
OH
CO2H
CO2H
1.11 ± 0.32
1.2 ± 0.3
1.00 ± 0.06
40
OBn
CONH2
CONH2
<25%c
NDd
1.19 ± 0.07
41
OBn
CN
CN
<25%c
0.1 ± 0.2
1.02 ± 0.05
42
OBn
TET
TET
37%c
NDd
1.06 ± 0.10
51
OBn
CO2H
TET
2.93 ± 0.33
NDd
1.06 ± 0.04
52
OBn
TET
CO2H
56%c
NDd
1.10 ± 0.04
Inflection temperature
(Ti) for Keap1 protein + vehicle = 61.3
±
0.1 °C, ΔTi for cpd 4 = 20.0 ± 0.3 °C.
Fold-induction of NQO1 enzymatic
activity relative to the DMSO control after 24 h inhibitor treatment,
cpd 4 = 2.85 ± 0.14 fold.
Percentage inhibition at 10 μM
concentration of inhibitor.
Not determined.
Inflection temperature
(Ti) for Keap1 protein + vehicle = 61.3
±
0.1 °C, ΔTi for cpd 4 = 20.0 ± 0.3 °C.Fold-induction of NQO1 enzymatic
activity relative to the DMSO control after 24 h inhibitor treatment,
cpd 4 = 2.85 ± 0.14 fold.Percentage inhibition at 10 μM
concentration of inhibitor.Not determined.To complement
our SAR model, we then explored the effect of carboxylate
isosteric transformations on binding activity. For this purpose, we
selected 11 as a model scaffold due to its superior potency
compared to other compounds of this series. Replacing both carboxylate
groups of 11 with carboxamides (compound 40) or nitriles (compound 41) abolished binding affinity,
possibly due to the inability of these groups to form the necessary
salt bridging interactions with the triad of positively charged arginine
residues that are located near to the P1 and P2 sub-pockets of the
Keap1 Kelch binding site. Similar results were obtained for the bis-1H-tetrazole 42, although some marginal inhibitory
activity was detected at the highest concentration screened (10 μM).
As expected, substituting only one carboxylic acid of 11 for a 1H-tetrazole (compounds 51 and 52) was better tolerated; however, activity was reduced by
at least sixfold in both cases.The 1,3-bis-sulfonamide benzamides 12 and 58–61 showed a consistently
weaker binding profile
in the FP assay compared to the previous examples (Table S1). The N-phenyl 58 and N-benzyl 59 analogues demonstrated modest activity
at 10 μM, causing approximately 50% inhibition of the interaction
between the Keap1 Kelch protein and the fluorescent peptide. On the
other hand, the N-ethyl 12 was unable
to markedly inhibit the FP signal at the same concentration. Analogues
having both carboxylates substituted for carboxamides (compound 60) or nitriles (compound 61) did not show activity
at 10 μM, results that are consistent with the SAR obtained
from the 2-benzyloxy derivatives 40 and 41.The binding affinity of compounds was confirmed in a secondary
orthogonal differential scanning fluorimetry (DSF) assay at a fixed
concentration of 10 μM using a label-less adaptation of a published
procedure in a Nanotemper Tycho instrument that measures changes in
the ratio of protein tyrosine and tryptophan intrinsic fluorescence
with temperature.[20] The fluorescence ratio
inflection temperature (Ti) for the Keap1
Kelch protein was estimated to be 61.3 ± 0.1 °C using this
method. Compound 4 was used as the positive control and
gave a Ti of 81.3 ± 0.3 °C, corresponding to an induced
shift in the Ti (ΔTi) of 20 °C. The Keap1 Kelch melting curves were
shifted to a smaller extent (ΔTi of ∼2.4–3.4 °C) relative to the positive control
in the presence of compounds 10, 11, 31–33, 36, and 38, confirming their direct interaction with the protein, while 34, 37, 39, and 41 gave
smaller ΔTi values (ΔTi ≈ 0.1–1.2 °C) consistent
with their lower activity in the FP assays.To further characterize
the binding profile of the series, we carried
out ITC experiments. Compound 11 had an estimated Kd of 0.500 ± 0.057 μM and compounds 31, 32, 36, and 38 had Kd values in the range 0.429–0.638 μM
(Figure S1), which were comparable with
the IC50 values recorded in the FP assay (Table ). The reference compound 4 had a Kd of 0.039 ± 0.010
μM, consistent with previous observations (Figure ).
Crystal Structure of the
Keap1–Compound 11 Complex
The overall
structure of the Keap1 Kelch domain
in complex with compound 11 was solved at a resolution
of 1.75 Å (PDB entry 6HWS). The Keap1 Kelch domain adopts a six-bladed β-propeller
structure with a central hollow channel.[30] Compound 11 binds to the Nrf2 binding region which
is located over the central pore through the protein domain. The central
phenyl ring of compound 11 occupies the P3 sub-pocket
of the Keap1 Kelch binding site, forming a π–cation interaction
with the positively charged guanidino group of Arg415 (Figure A,B). Similarly to the naphthalene-based
inhibitors reported previously,[17,19] the flanking 4-methoxybenzenesulfonamide
moieties form π–π stacking interactions with the
aromatic side chains of Tyr334, Tyr572, Phe577, and Tyr525 of the
P4 and P5 sub-pockets, whereas one of the sulfonamide oxygen atoms
forms a hydrogen bond with the carboxamide side chain group of Gln530
(Figure C). The two
acetate side chains of 11 are projected toward the P1
and P2 sub-pockets with the respective carboxylate groups participating
in a network of hydrogen bonding and charge–charge interactions
with Arg415, Ser508, Asn414, and Arg380. Interestingly, the benzyloxy
phenyl group stabilizes the bound conformation of 11 by
forming a face-to-face π–π stacking interaction
with the 4-methoxybenzene ring of the adjacent sulfonamide side chain,
while it also participates in an edge-to-face cation−π
stacking with Arg483 (Figure B,C).
Figure 3
(A–C) Bound conformation of compound 11 (green
in A and B, purple in C) co-crystallized with the Keap1 Kelch domain;
(D) overlay of the bound conformation of 11 and the ETGE
peptide bound to Keap1 (PDB entry 2FLU);[29] (E) comparison
of the Keap1 Kelch domain binding pocket in the compound 11 co-crystal structure (cyan) and the 2FLU ETGE peptide co-crystal (blue); (F) comparison
of the Keap1 Kelch domain binding pocket in the compound 11 co-crystal structure (cyan) and the compound 1 co-crystal
structure (green).[19]
(A–C) Bound conformation of compound 11 (green
in A and B, purple in C) co-crystallized with the Keap1 Kelch domain;
(D) overlay of the bound conformation of 11 and the ETGE
peptide bound to Keap1 (PDB entry 2FLU);[29] (E) comparison
of the Keap1 Kelch domain binding pocket in the compound 11 co-crystal structure (cyan) and the 2FLU ETGE peptide co-crystal (blue); (F) comparison
of the Keap1 Kelch domain binding pocket in the compound 11 co-crystal structure (cyan) and the compound 1 co-crystal
structure (green).[19]The bound conformation of 11 contrasts with the predicted
docked conformation (Figure F) in which the benzyloxy substituent occupies the central
channel through the protein and instead has a greater similarity to
that of ETGE peptide-bound forms of the Keap1 Kelch domain (e.g.,
PBD entry 2FLU)[29] in which the binding pocket is adapted
to fit the peptide. With respect to the ligand, one of the carboxylic
acids of 11 and the ETGE peptide glutamate residue equivalent
to Glu79 overlay (Figure D, lower right) and the peptide backbone overlays with the
adjacent sulfonamide motif. The Keap1 binding pocket residues are
also in a very similar orientation in the 11-bound and
peptide-bound structure (Figure E), particularly with regard to the Arg415 residue,
which is usually rotated in small molecule-bound structures to expose
the central channel through the β-propeller structure (e.g.,
the complex with 1, PDB entry 4IQK, Figure F).
Cell-Based Experiments
The ability
of compounds of
this series to induce the transcriptional activity of Nrf2 was evaluated
in Hepa1c1c7 mouse hepatoma cells using the colorimetric NAD(P)H quinone
oxidoreductase 1 (NQO1) enzyme activity assay, an assay that is widely
used to characterize the biological activity of Nrf2 inducers.[12,31,32] We initially screened the compounds
at a fixed dose of 10 μM (Table ). This preliminary screening gave encouraging results
as it revealed a good correlation between the in vitro and cell-based
activities of the series. Compounds 31, 32, 11, and 38 retained a good level of potency
in cells, inducing the enzymatic activity of NQO1 by more than 1.5-fold
relative to the DMSO control at 10 μM. On the other hand, compounds
with a less tight binding to Keap1 failed to cause a marked induction
of NQO1 at the same concentration. Motivated by these results, we
carried out concentration–response studies with selected candidates
in order to gain a more detailed understanding of the cellular activity
profile of the series. As shown in Figure , the NQO1 activity was generally induced
in a dose-dependent manner after a 24 h treatment by all compounds
screened. The methoxy-containing compound 10 was the
least potent among the analogues tested, displaying only marginal
cellular activity (1.4-fold induction) at 100 μM, results that
are consistent with its weak binding affinity for Keap1. On the other
hand, 38 showed the most promising cellular activity
profile, causing a reasonable increase in NQO1 (1.5-fold to 2-fold)
at low micromolar concentrations that reaches a threefold induction
at the highest concentration tested (100 μM). The maximal NQO1
activation induced by 38 is similar to that observed
for the positive controls bardoxolone methyl (CDDO-Me) and dimethyl
fumarate (DMF); however, the latter compounds are more active at lower
concentrations. In contrast to CDDO-Me that was cytotoxic at concentrations
higher than 1 μM, none of the phenyl ethers showed signs of
cytotoxicity in this assay up to a final concentration of 100 μM.
This observation was supported by cytotoxicity assays carried out
in ARPE19 retinal pigment epithelial cells.[33] We demonstrated that 11 was not cytotoxic at concentrations
up to 100 μM. Under the same conditions, the corresponding naphthalene
bis-sulfonamide 4 showed a small reduction in cell viability
(70% cell viability @ 100 μM) and sulforaphane had an IC50 of ∼10 μM, consistent with its behavior in
other cell lines (Table S2).[12] Based on these data, compound 11 was evaluated using real-time quantitative PCR to confirm the induction
of NQO1 that was observed in the enzymatic assays in mouse Hepa1c1c7
cells. After 16 h of exposure, an increase in the expression of NQO1
was observed (Figure A; consistent with the NQO1 enzymatic assay results). The increase
in mRNA levels of other Nrf2 target genes including heme oxygenase
1 (Hmox1), glutathione S-transferase P (GstP), and
glutamate-cysteine ligase catalytic (Gclc) and modulatory (Gclm) subunits
(Figure B–E)
indicated that the effect was not limited to NQO1.
Figure 4
Dose-dependent induction
of NQO1 [NAD(P)H-dependent quinone oxidoreductase-1]
by selected (A) Keap1-Nrf2 PPI inhibitors: 10 (◀), 11 (▼), 31 (▲), 32 (●), 36 (⧫), and 38 (■)
and (B) electrophilic Nrf2 activators: DMF (★) and CDDO-Me
(▶).
Figure 5
Induction of the Nrf2-transcriptional targets
Nqo1 (A), Hmox1 (B),
Gstp (C), Gclc (D), and Gclc (E) in Hepa1c1c7 cells (n = 3) by compound 11 (50 μM). The data are represented
graphically as mean values ± 1 standard deviation (SD). Student’s t-test was used to test for statistical significance, and
in all cases, the difference in gene expression between vehicle- and
compound 11-treated cells was statistically significant
(p < 0.05).
Dose-dependent induction
of NQO1 [NAD(P)H-dependent quinone oxidoreductase-1]
by selected (A) Keap1-Nrf2 PPI inhibitors: 10 (◀), 11 (▼), 31 (▲), 32 (●), 36 (⧫), and 38 (■)
and (B) electrophilic Nrf2 activators: DMF (★) and CDDO-Me
(▶).Induction of the Nrf2-transcriptional targets
Nqo1 (A), Hmox1 (B),
Gstp (C), Gclc (D), and Gclc (E) in Hepa1c1c7 cells (n = 3) by compound 11 (50 μM). The data are represented
graphically as mean values ± 1 standard deviation (SD). Student’s t-test was used to test for statistical significance, and
in all cases, the difference in gene expression between vehicle- and
compound 11-treated cells was statistically significant
(p < 0.05).In a proof of principle study, target engagement by compound 11 was further supported by use of the cellular thermal shift
assay (CETSA).[34] Lysates of human promyelocytic
leukemia HL-60 cells were incubated with compound 11 for
1 h at 37 °C and then subjected to a temperature gradient. Immunoblotting
analysis of the protein levels of Keap1 in the soluble fractions after
the removal of heat-induced aggregates showed that the thermal stability
of Keap1 was increased by the treatment with compound 11 (Figure S2), indicating that the compound
binds to endogenous human Keap1 in the cellular environment.The activities that we describe are consistent with inhibition
of the PPI between Keap1 and Nrf2 and the subsequent upregulation
of Nrf2 target genes. However, as with other studies of Keap1 inhibitors,
we cannot rule out the possibility that the compounds have other biological
activities that we have not determined in our assays.
Physicochemical
Properties
Given the promising results
obtained from our SAR studies, we sought to further characterize this
novel series of Keap1-Nrf2 PPI inhibitors by determining their aqueous
solubility and transmembrane permeability. Solubility measurements
were carried out at pH 7.4 using the Multiscreen HTS-PCF filter plates
(Merck Millipore, MSSLBPC10), while transmembrane permeability was
evaluated using the parallel artificial membrane permeation assay
(PAMPA).[35] Generally, the solubility of
the series was good and within the range of 250–350 μg/mL
(Table S3), which is comparable to the
reported solubility values for the naphthalene-containing compound 4. The highest-affinity analogue 11 had an aqueous
solubility of 348 μg/mL, while modifications on the ether portion
of the scaffold did not cause significant changes. On the other hand,
bioisosteric replacements of both carboxylate groups of 11 with nitriles, amides, or 1H-tetrazoles reduced
solubility by more than twofold. As expected, compounds of this chemical
series were found to have a relatively low permeability at pH 7.4
due to the presence of the two ionizable carboxylate groups (Table S4). Modifications to the ether moiety
of the scaffold had minimal effects on the permeability of the series.
Similarly, replacing either carboxylic acid group with a 1H-tetrazole did not cause major changes in the log Pe and led to a similar degree of passive diffusion. Overall,
compound 38 demonstrated the most promising combination
of physicochemical properties among the compounds screened, which
could explain its improved cellular activity profile.These
data prompted us to further investigate the potential of 38 as a lead molecule for further optimization by evaluating a series
of its physicochemical and pharmaceutical properties using high-performance
liquid chromatography (HPLC) methods.[36] Chromatographic approaches have been widely applied to drug discovery
for the determination of lipophilicity and biomimetic properties of
preclinical drug candidates. For example, the chromatographic hydrophobicity
index (CHI) provides an alternative approach to determine the lipophilicity
of compounds that is based on their retention time in reversed-phase
HPLC columns. CHI values can be converted into the log D scale (CHI log D) to enable comparisons with data
obtained using the octanol–water partition coefficient method.
Compound 38 demonstrated good lipophilicity in our HPLC
measurements that was consistently higher compared to the naphthalene 4 in different pH environments (Table ).
Table 2
Comparison of Physicochemical
and
Pharmaceutical Properties of Compounds 38 and 4
compd
CHI log D 2.0
CHI log D 7.4
CHI log D 10.5
IAM binding log K(IAM)
Vd (L/kg)
DEmax (%)
HSA binding (%)
AGP binding (%)
4
2.75
0.57
0.61
1.64
0.01
0.94
>99.9
20.3
38
3.15
0.85
0.88
1.83
0.09
4.55
99.1
47.7
The improved hydrophobic character of 38 translated
into an increased immobilized artificial membrane (IAM) binding, which
represents the affinity of ligands for phospholipids and thus relates
to permeability through the main physiological barriers. These results
are consistent with the improved in vitro binding affinity to cell-based
activity ratio of compound 38 relative to 4 (Table , Figure . We also carried
out HPLC biomimetic experiments using stationary phases containing
immobilized human serum albumin (HSA) and α-glycoprotein (AGP)
to predict the binding of the compounds to the respective plasma proteins.
Our results suggest that the two molecules have a relatively different
plasma protein binding profile. In particular, naphthalene 4 demonstrated a very tight binding to HSA, as indicated by its apparent
infinite retention inside the HSA-coated HPLC column, while 38 bound to HSA more weakly. The binding profiles of 38 and 4 to the AGP were also different, with
the two compounds showing 47.7 and 20.3% binding, respectively. Compound 38 demonstrated a more than fourfold improvement in maximum
drug efficiency percentage (DEmax)[37] relative to the naphthalene 4, which predicts an improvement
in free plasma concentration. Similarly, the volume of distribution
(Vd)[38] of 38 is predicted to be larger than that of 4.
Optimizing the physicochemical properties of the compounds is relevant
to their cellular permeability, which could be improved and also to
the potential utility of the compounds in interventions in neurological
conditions where the role of Nrf2 induction has a number of promising
applications.[1,7,39,40]
Conclusions
In
this article we present the design and synthesis of a series
of phenyl bis-sulfonamide inhibitors of Keap1. Compound 11 binds to a Keap1 conformation that is comparable to its peptide
bound form which distinguishes it from all other Keap1 PPI-inhibitory
small molecules to date. The SARs for the compound series (Figure ) demonstrate the
effects of substitution at the newly introduced ether substituent,
the carboxylate moieties and the positioning of the sulfonamide moieties.
Together with the structural data from compound 11, this
suggests that the core structure could be further modified, for example,
by changes to the linker length for the acid motifs, exploration of
additional carboxylate isosteres, replacement of the benzyloxy substituent
with substituted analogues or heterocycles and modifications to the
sulfonamide substituents. The current compounds bind to Keap1 with
mid-nanomolar binding affinity, interact with Keap1 in cells, and
increase the expression of Nrf2 target genes. Compound 11 lacked cytotoxic activity in ARPE19 cells. The structural characterization
of the unique bound conformation of the ligands and the assessment
of their physicochemical properties and liabilities provide a basis
for further optimization of this new subtype of the Keap1 inhibitor.
Figure 6
SARs for
compounds 10–12, 31–42, 51–52, and 58–61.
SARs for
compounds 10–12, 31–42, 51–52, and 58–61.
Experimental Section
Chemistry Methods
General
All anhydrous solvents and reagents were purchased
from commercial suppliers and used without further purification.1H NMR spectra were recorded at ambient temperature in
deuterated solvents (CDCl3 or DMSO-d6) on a Bruker ADVANCE 400 spectrometer at 400.13 MHz or Bruker
ADVANCE 500 spectrometer at 500 MHz. Chemical shifts are reported
in parts per million (ppm) downfield from the tetramethylsilane reference
(δ = 0) using the residual protonated solvent as an internal
standard [1H: δ (CDCl3) = 7.26 ppm, δ
(DMSO-d6) = 2.50 ppm]. Data for 1H NMR are given as follows: chemical shift {multiplicity, coupling
constants [J, given in hertz (Hz)], integration and
assignment}. Multiplicities in the 1H NMR spectra are quoted
as follows: s = singlet, d = doublet, t = triplet, q = quartet, m
= multiplet, dd = double doublet, and ddd = double double doublet.
Splitting patterns that could not be interpreted or easily visualized
were recorded as multiplets (m) or broad peaks (br).13C NMR spectra were recorded at an ambient temperature
in deuterated solvents (CDCl3 or DMSO-d6) on a Bruker ADVANCE 400 spectrometer at 100.61 MHz
or Bruker ADVANCE 500 spectrometer at 125 MHz. Chemical shifts were
measured in ppm relative to tetramethylsilane (δ = 0) using
the following internal references: δ (CDCl3) = 77.0
ppm, δ (DMSO-d6) = 39.4 ppm.
High-Resolution
Mass Spectrometry
High-resolution mass
spectrometry (HRMS) spectra were recorded on a Micromass Q-TOF Premier
Tandem Mass Spectrometer coupled to an HPLC instrument using electrospray
ionization (ESI) mass spectrometry. Calibration was performed with
an internal standard—positive mode: reserpine which gives m/z [M + H]+ = 609.2812, negative
mode: taurocholic acid which gives m/z [M – H]− = 514.2839.
Analytical
TLC
Analytical TLC was performed on pre-coated
Merck glass backed silica gel plates (Silica gel 60 F254) and visualized by exposure to ultraviolet light (254 nm) and/or
staining with an appropriate reagent followed by heating if required.
Flash Column Chromatography
Flash column chromatography
was carried out on Merck Kieselgel 60 (40–63 μm) under
a positive pressure of N2 gas.
Liquid Chromatography–Mass
Spectrometry
Liquid
chromatography–mass spectrometry (LC–MS) spectra were
recorded using a Shimadzu LCMS-2020 equipped with an XTerra MS C18
column (4.6 × 50 mm, 2.5 μm) and a flow rate of 0.6 mL/min.
The eluent system consisted of eluent A (H2O with 0.1%
formic acid, HPLC grade) and eluent B (MeCN with 0.1% formic acid,
HPLC grade) with the following conditions: 0.0–2.5 min 90%
A/10% B, then a linear gradient from 2.5 to 5.5 min to a final composition
of 5% A/95% B that was maintained for a further 2.5 min, and then
adjusted to 10% A/90% B over 10.5 min and held for 1.5 min. Total
run time = 12 min.
Analytical Reverse-phase HPLC
Analytical
reverse-phase
HPLC was carried out on an XSELECT CSH C18 column 50 × 6 mm (particle
size: 2.5 μm) at a flow rate of 1.0 mL/min. The eluent system
consisted of eluent A (H2O with 0.1% TFA, HPLC grade) and
eluent B (MeCN with 0.1% TFA, HPLC grade) with the following gradient
conditions: initial fixed composition 5% B to 50% B over 20 min, then
increased to 95% B over 2 min, held for 2 min at 95% B, and then returned
to 5% B in 1 min. Total duration of the gradient run was 25 min.All compounds that underwent biological testing were >95% pure
by
HPLC analysis.
To a solution of 13 (1.63 g, 9.7
mmol) in EtOH/H2O (220 mL, 4:1 v/v) was added AcOH (9.0
mL) and the mixture was brought to reflux. Fe (2.7 g, 48.25 mmol)
was then added portion-wise and stirring was continued for 1 h. On
completion, the reaction was left to cool to rt, diluted with H2O (50 mL) and basified to pH 8–9 with sat. NaHCO3. The reaction mixture was filtered through Celite and the
filtrate was extracted with EtOAc (3 × 80 mL). The combined organic
layers were washed with H2O (3 × 150 mL), sat. brine
(3 × 150 mL), dried over anhydrous Na2SO4, and evaporated to dryness under reduced pressure. The crude product
was sequentially dissolved in DCM (20 mL) and a solution of 4-methoxybenzenesulfonyl
chloride (4.21 g, 20.3 mmol) in pyridine (2.4 mL, 29.10 mmol) was
added to the reaction mixture. The reaction was stirred at rt overnight
and following the addition of a small volume of hexane to aid compound
precipitation, the crude solid was collected by filtration and washed
with Et2O. Flash chromatography (EtOAc/hexane 7:3 v/v)
afforded 19 (962 mg, 35%) as a white powder. 1H NMR (500 MHz, DMSO): δ 10.09 (s, 1H), 9.16 (s, 1H), 7.66
(d, J = 8.9 Hz, 2H), 7.58–7.41 (m, 2H), 7.06
(d, J = 8.9 Hz, 2H), 7.01 (d, J =
8.5 Hz, 1H), 6.98 (d, J = 8.9 Hz, 2H), 6.63–6.50
(m, 2H), 3.81 (s, 6H). 13C NMR (126 MHz, DMSO): δ
162.4, 162.1, 153.0, 136.6, 132.1, 130.0, 128.9, 128.7, 126.4, 121.2,
114.2, 113.7, 111.5, 103.5, 55.6, 55.5, 55.2.
To a solution of 35 (400
mg, 0.63 mmol) and K2CO3 (287 mg, 2.08 mmol)
in acetone (10 mL) was added 1-iodopropane (100 μL, 1.04 mmol)
and the resulting suspension was stirred at reflux for 16 h. On completion,
the reaction mixture was cooled to rt, diluted with H2O
(80 mL), and extracted with EtOAc (3 × 100 mL). The combined
organic layers were washed with H2O (3 × 200 mL) and
sat. brine (3 × 200 mL), dried over anhydrous MgSO4, and evaporated to dryness under reduced pressure. The crude product
was dissolved in THF/MeOH (10 mL, 1:1 v/v) and a solution of 1 N NaOH
(6.3 mL) was added. The reaction was stirred at 40 °C for 2 h
and was then cooled to rt. The reaction mixture was acidified with
1 N HCl to pH 1–2 and the resulting precipitate was filtered,
washed with H2O, DCM, and EtOAc, and dried in vacuo to
afford 36 (160 mg, 42%) as a white solid. 1H NMR (500 MHz, DMSO-d6): δ (ppm)
13.15–12.66 (br, 2H), 7.61 (d, J = 8.9 Hz,
2H), 7.46 (d, J = 8.9 Hz, 2H), 7.35 (d, J = 8.5 Hz, 1H), 7.09 (d, J = 8.9 Hz, 2H), 7.04 (d, J = 8.9 Hz, 2H), 6.81 (dd, J = 8.5, 2.2
Hz, 1H), 6.57 (d, J = 2.2 Hz, 1H), 4.41 (s, 2H),
4.19 (s, 2H), 3.85 (s, 3H), 3.84 (s, 3H), 1.27–1.15 (m, 2H),
0.72 (t, J = 7.4 Hz, 3H). 13C NMR (125
MHz, DMSO-d6): δ (ppm) 170.3, 170.0,
162.8, 162.5, 154.5, 140.8, 133.3, 131.3, 129.8, 129.6, 129.0, 125.2,
119.0, 114.2, 113.9, 111.5, 69.0, 55.7, 55.6, 51.8, 50.4, 21.3, 10.0.
LC–MS m/z (ESI): 1242.00
[2M – H]−, tR = 5.96 min, purity: >95%. HRMS (ESI): calcd for C27H30N2O11S2 [M –
H]−, 621.1213; found, 621.1186. HPLC: tR = 9.01 min, purity: >95%.
To a solution of 16 (10.0 g, 29.04
mmol) in DCM (100 mL) was added 4-methoxybenzenesulfonyl chloride
(6.6 g, 31.94 mmol), DMAP (355 mg, 2.90 mmol), and pyridine (5.2 mL,
63.89 mmol), and the reaction was stirred for 24 h at rt. On completion,
the reaction was diluted with H2O (100 mL) and acidified
to pH 2–3 with 1 N HCl. The organic layer was isolated and
the aqueous layer was extracted with DCM (3 × 100 mL). The combined
organic portions were washed with 1 N HCl (3 × 200 mL), H2O (3 × 200 mL), and sat. brine (3 × 200 mL), dried
over anhydrous Na2SO4, and evaporated to dryness
under reduced pressure. The crude product was triturated with Et2O and used in the next step without further purification.To a solution of the crude product in DMF (10 mL) was added ethyl
bromoacetate (4.0 mL, 36.29 mmol) and K2CO3 (6.0
g, 43.55 mmol), and the reaction was stirred for 8 h at rt. On completion,
the reaction mixture was acidified with 1 N HCl to pH 4–5 and
the resulting precipitate was collected by filtration, washed with
H2O, Et2O, and MeOH, and dried in vacuo to afford
the ester 43 (7.6 g, 52%) as an off-white powder. 1H NMR (500 MHz, DMSO-d6): δ
(ppm) 7.96–7.86 (m, 2H), 7.71 (d, J = 8.6
Hz, 1H), 7.53 (d, J = 8.9 Hz, 2H), 7.37–7.30
(m, 3H), 7.16–7.07 (m, 2H), 6.90 (d, J = 8.9
Hz, 2H), 5.02 (s, 2H), 4.39 (s, 2H), 4.06 (q, J =
7.1 Hz, 2H), 3.75 (s, 3H), 1.15 (t, J = 7.1 Hz, 3H). 13C NMR (125 MHz, DMSO-d6): δ
(ppm) 168.6, 162.7, 154.9, 147.7, 135.5, 134.0, 132.9, 130.6, 129.1,
128.2, 127.8, 127.0, 115.5, 114.2, 108.2, 70.2, 60.8, 55.5, 50.2,
13.9.
To a solution of 3,5-diaminobenzoic
acid 53 (1.0 g; 6.60 mmol) in DMF (5 mL) were added 4-methoxybenzenesulfonyl
chloride (3.0 g, 14.52 mmol), pyridine (1.3 mL, 16.50 mmol), and DMAP
(80 mg, 0.66 mmol), and the reaction was stirred at 100 °C overnight.
On completion, the reaction was cooled to rt and diluted with EtOAc
(100 mL). The organic layer was washed with 1 N HCl (3 × 50 mL),
H2O (3 × 50 mL), and sat. brine (3 × 50 mL),
dried over anhydrous MgSO4, and evaporated to dryness under
reduced pressure to afford 54 (1.73 g, 53%) as a white
solid. LC–MS m/z (ESI): 490.95
[M – H]−, tR =
5.36 min, purity: >95%.
To a solution
of 54 (250 mg, 0.51
mmol) in DMF (5 mL) were added aniline (55 μL, 0.61 mmol), EDCI·HCl
(245 mg, 1.28 mmol), and DMAP (125 mg, 1.02 mmol), and the reaction
mixture was stirred at rt overnight. On completion, the solution was
diluted with EtOAc (10 mL), washed with 1 N HCl (3 × 10 mL),
sat. NaHCO3 (3 × 10 mL), and sat. brine (3 ×
10 mL). The organic layer was dried over anhydrous MgSO4 and evaporated to dryness under reduced pressure to afford 55 (179 mg, 63%) as a white solid. LC–MS m/z (ESI): 566.05 [M – H]−, tR = 6.19 min, purity: >95%.
To a solution of 55 (180
mg, 0.32 mmol) in DMF (2 mL) was added K2CO3 (133 mg, 0.96 mmol) and ethyl bromoacetate (90 μL, 0.79 mmol)
and the reaction mixture was stirred at rt overnight. On completion,
the solution was diluted with H2O (30 mL), acidified to
pH 5 with 2 M HCl, and extracted with EtOAc (4 × 20 mL). The
combined organic layers were washed with H2O (3 ×
50 mL) and sat. brine (3 × 50 mL), dried over anhydrous MgSO4, and evaporated to dryness under reduced pressure to afford
the ester (220 mg, 96%) as a white solid that was used in the next
step without further purification.To a solution of the crude
product (180 mg, 0.24 mmol) in THF/MeOH (10 mL, 1:1 v/v) was added
a solution of NaOH (39 mg, 0.97 mmol) in H2O (5 mL), and
the reaction mixture stirred at rt overnight. On completion, the reaction
was acidified with 2 M HCl to pH 2, diluted with H2O (75
mL), and extracted with EtOAc (4 × 50 mL). The combined organic
layers were washed with H2O (3 × 50 mL), sat. brine
(3 × 50 mL), dried over anhydrous MgSO4 and evaporated
to dryness under reduced pressure to afford 58 (148 mg,
83%) as a white solid. 1H NMR (400 MHz, DMSO): δ
(ppm) 13.03–12.60 (br, 2H), 10.12 (s, 1H), 7.61 (d, J = 2.0 Hz, 2H), 7.57 (dd, J = 8.5, 1.0
Hz, 2H), 7.50–7.43 (m, 4H), 7.31–7.22 (m, 2H), 7.18
(t, J = 2.0 Hz, 1H), 7.06–6.99 (m, 1H), 6.98–6.90
(m, 4H), 4.27 (s, 4H), 3.73 (s, 6H). 13C NMR (101 MHz,
DMSO): δ (ppm) 169.6, 163.5, 162.8, 140.5, 138.6, 135.6, 129.6,
129.5, 128.6, 125.4, 124.1, 120.8, 114.3, 55.5, 51.7, 15.2. HRMS (ESI):
calcd for C31H29N3O11S2 [M – H]−, 682.1165; found, 682.1170.
LC–MS m/z (ESI): 682.00 [M
– H]−, tR = 6.04
min, purity: >95%.
Compound 57 (0.25 g, 0.43 mmol) was
dissolved in DMF (2 mL) in a round-bottomed flask. K2CO3 (0.178 g, 3 equiv) was added with stirring for 3 min, followed
by 2-bromoacetamide (0.148 g, 2.5 equiv). The reaction mixture was
stirred overnight at rt and monitored by TLC. The solution was diluted
in water (30 mL) and adjusted to pH 5 with 2 M HCl, then extracted
with EtOAc (4 × 20 mL). The combined organic layers were washed
with water (3 × 50 mL) then with brine (1 × 50 mL). The
organic layer was dried over MgSO4, and the solvent was
removed under vacuum. The residue was suspended in Et2O
(3 mL) and then the solvent evaporated under vacuum. The residue was
stirred with hexane (5 mL) overnight. The hexane was decanted carefully
and the resulting solid 60 (0.278 g, 93%) was dried under
vacuum. 1H NMR (400 MHz, DMSO): δ (ppm) 8.96 (t, J = 5.9 Hz, 1H), 7.68 (d, J = 2.0 Hz, 2H),
7.54–7.49 (m, 4H), 7.37–7.29 (m, 5H), 7.28–7.23
(m, 3H), 7.16 (s, 2H), 7.10–7.04 (m, 4H), 4.45 (d, J = 5.8 Hz, 2H), 4.11 (s, 4H), 3.83 (s, 6H); 13C NMR (101 MHz, DMSO): δ (ppm) 168.6, 164.4, 162.9, 140.5,
139.2, 135.0, 129.7, 128.9, 128.3, 127.2, 126.8, 125.6, 114.4, 55.7,
42.6. HRMS (ESI): calcd for C32H33N5O9S2 [M – H]−, 694.1641;
found, 694.1633. LC–MS m/z (ESI): 694.00 [M – H]−, tR = 5.57 min, purity: >95%.
The Keap1 Kelch domain protein coordinates
were obtained from the PDB (ref: 4IQK); the ligand, water molecules, and ions
were removed from the structure file and the protein structure was
parameterized and saved in pdbqt format using AutoDockTools. The small
molecules were prepared in 2D format using ChemDraw and then converted
into 3D coordinates using Chem3D (ChemOffice 19, PerkinElmer). The
ligands were energy minimized using the MM2, MMFF94, and Mopac (PM7)
protocols with default parameters and then saved as mol2 files. The
mol2 files were converted into pdbqt files using openbabel 2.3.2.[41] The compounds were docked to the Keap1 Kelch
domain using qvina 2.1[42] with default parameters
and an exhaustiveness setting of 10. The best docked conformations
were saved and analyzed in UCSF Chimera 1.14.[43] Images of the docked structures and protein conformers were prepared
in UCSF Chimera, LigPlot Plus,[44] and Pymol
(The PyMOL Molecular Graphics System, Version 2.0 Schrödinger,
LLC).
Sub-Cloning, Expression, and Purification of the Keap1 Kelch
Domain
The coding sequence for the Keap1 Kelch domain was
sub-cloned into a modified pEt15b vector with a His-SUMO tag and Ulp1
cleavage site. The plasmid was transformed into Escherichia
coli BL21 CodonPlus cells (Novagen) and grown in 6
L of the Terrific Broth (TB) medium at 37 °C, supplemented with
100 mg/L ampicillin, to an A600 of 1.0
and induced for 24 h at 20 °C with 0.5 mM isopropyl β-d-thiogalactoside (IPTG; Melford). Cells were harvested by spinning
them down in a Beckman centrifuge using a JLA 16.25 rotor at 6000
rpm for 20 min. Protein purification was carried out in three steps.
The cells were initially resuspended in lysis buffer (50 mM HEPES,
pH 7.4, 300 mM NaCl and 10 mM imidazole) and sonicated at 16 μm
amplitude for 10 cycles of 30 s ON and 45 s OFF. The lysate was then
cleared by centrifugation at 20,000 rpm at 4 °C using a JA 25.5
rotor. To perform the first affinity chromatography, the lysate was
loaded onto a 5 mL His-Trap column (GE Healthcare, UK) pre-equilibrated
with lysis buffer, washed for 50 column volumes of wash buffer (lysis
buffer supplemented with 50 mM imidazole), and eluted with elution
buffer (lysis buffer supplemented with 250 mM imidazole). The purified
protein was pooled and cleaved overnight with UlP1 protease supplemented
with 3 mM DTT. Simultaneous with the cleavage, the protein was also
dialyzed against buffer containing 50 mM HEPES, pH 7.4, and 300 mM
NaCl. Thereafter, the protein was passed through the His-Trap column
for the second time using the same buffers as before. The cleaved
protein without the His-SUMO tag was captured in the flow, which was
then pooled and passed through a 16/60 Superdex 200 pg size exclusion
column with a buffer containing 50 mM HEPES, pH 7.5, and 200 mM NaCl
to further purify Keap1. Lastly, the protein was then pooled and concentrated
to 10 mg/mL using an Amicon ultraconcentration device (Millipore),
aliquoted, flash-frozen in liquid nitrogen, and stored at −80
°C for subsequent use.
FP Assays
The FP assays were carried
out according
to previously described methods.[8,9,12,45] Briefly, varying concentrations
of the small molecule inhibitor dissolved in DMSO were plated into
untreated Corning black 96 well plates containing a solution of the
Keap1 Kelch domain (200 nM final concentration) and the fluorescent
peptide FITCβ-DEETGEF-OH (1 nM final concentration) in Dulbecco’s
phosphate-buffered saline (DPBS) at pH 7.4 (11% final DMSO concentration,
100 μL final volume). Following incubation under slow agitation
(30 min at RT in the absence of light), the plates were transferred
to a PHERAstar microplate reader and the FP was recorded. All measurements
were recorded in triplicate. The data were normalized to the control
and then fitted to a standard four-parameter logistic function using
Origin Pro software.
DSF Assays
Solutions of the Keap1
Kelch domain protein
(5 μM final concentration) and the small molecule inhibitor
(10 μM final concentration) were prepared in DPBS at pH 7.4
(10% final DMSO concentration, 30 μL final volume) in an Eppendorf
tube. The solution was centrifuged briefly (1500 rpm, 60 s) to remove
any particulates and mix the samples, followed by a 30 min incubation
at RT. The samples were loaded into Tycho NT.6 capillaries (Nanotemper)
in triplicate and positioned in the Tycho NT 1.6 instrument. The samples
were heated from 35 °C to 95 °C using the default settings
and the ratio of the intrinsic tyrosine and tryptophan fluorescence
intensities (350 and 330 nm) for each capillary tube was recorded.
The fluorescence ratio versus temperature was plotted and the inflection
temperature (Ti) of the profile was calculated
and recorded within the instrument software. Changes in inflection
(ΔTi) for the inhibitors were calculated
by subtracting the mean Ti of the Keap1
sample plus vehicle (n = 6) from the mean Keap1 plus
inhibitor value (n = 3).
ITC Experiments
The measurements were performed as
described previously.[45] Briefly, the protein
was dialyzed overnight against buffer containing 25 mM HEPES, pH 7.4,
and 200 mM NaCl and concentrated to 50 μM in the presence of
5% DMSO to match the ligand solution. The small molecule inhibitor
was diluted to a concentration of 500 μM in the dialysis buffer
(5% max final concentration of DMSO). ITC experiments were performed
with a MicroCal PEAQ-ITC instrument (Malvern Instruments, UK). Titrations
were carried out at 25 °C with a stirring speed of 750 rpm. The
inhibitor was titrated into the cell containing the protein solution
over 30 injections with the first injection of 0.3 μL, followed
by 29 injections of 1.3 μL and a gap of 120 s between each injection.
Data analyses were performed as previously described. For each experiment,
at least two titrations were performed. Titration data were analyzed
independently, and the obtained values were averaged.
Crystallization
and Structure Determination of the Keap1–Compound 11 Complex
Crystals for the Keap1 Kelch domain in
complex with 11 appeared after 2–3 days in 3.7
M sodium formate, pH 7.0 at 18 °C. The crystals were further
optimized by streak seeding to obtain single crystals for diffraction
measurements. The single crystals of the complex were cryo-protected
with 20% w/v ethylene glycol in 1.2-fold of the crystallization solution
and flash-frozen in liquid nitrogen.Diffraction data for individual
crystals were collected at beamlines I03 at Diamond Light Source.
Data were processed using either XDS[46] or
iMosflm[47] and scaled to resolutions as
mentioned in Table S4.[48] The structure of the complex was solved by molecular replacement
using the native Keap1 structure (PDB entry 1ZGK)[49] as a search model. The crystallographic statistics are
given in Table S4. Crystals of the binary
Keap1 Kelch domain in complex with 11 contain one molecule
in the asymmetric unit.
Solubility and PAMPA Assays
Solubility
measurements
were performed using Multiscreen HTS-PCF filter plates (Merck Millipore,
MSSLBPC10) at pH 7.4 according to the manufacturer’s instructions.
Membrane permeability was evaluated using the PAMPA using the hexadecane-method
PAMPA protocol with HDM-PAMPA multiscreen permeability filter plates
(Merck Millipore, MAIPNTR10) according to the manufacturer’s
instructions.
Biomimetic HPLC Methods
The chromatographic
measurements
of CHIlogD (CHI) were carried out using the compounds’ calibrated
gradient retention times obtained from an Agilent 1100 HPLC fitted
with a Gemini NX-C-18 column (Phenomenex Ltd Macclesfield, UK) with
dimensions of 50 × 3 mm and 5 μm particle size. The mobile
phase A was either 0.01 M formic acid (pH 2.6), a 50 mM ammonium acetate
buffer with an adjusted pH of 7.4, or a 50 mM ammonium acetate buffer
with an adjusted pH of 10.5. The mobile phase B was 100% acetonitrile.
The flow rate was 1.0 mL/min. An acetonitrile linear gradient was
used from 0 to 100%. The acetonitrile concentration reached 100% in
3.5 min. The 100% acetonitrile mobile phase was maintained for an
additional 1 min before it was returned to 0% at 4.7 min. The gradient
run cycle time was 6 min, with an additional equilibration time of
1 min before the next injection. The gradient retention times were
calibrated using the CHI values of reference compounds.The
phospholipid-binding was measured using an IAM PC.DD2 column with
dimensions of 100 × 4.6 mm (Regis Technologies Inc., Morton Grove,
IL, USA). The gradient retention times were measured using a 50 mM
ammonium acetate mobile phase with the pH adjusted to 7.4. The mobile
phase flow rate was 1.5 mL/min. The acetonitrile gradient was applied
to reach 90% in 4.75 min. The 90% acetonitrile concentration was maintained
for an additional 0.5 min (to 5.25 min) and then returned to 0% by
5.5 min. The cycle time was 6 min, plus an additional 1 min equilibration
time was applied while the injector prepared for the next injection.
The gradient retention times were calibrated with the acetophenone
homologues for which the CHI IAM values have been established using
isocratic measurements.The protein binding measurements were
carried out on Chiralpak
HSA and Chiralpak AGP columns with dimensions of 3 × 50 mm and
5 μm particle size (Chiral Technologies Europe, France). The
mobile phase was 50 mM ammonium acetate adjusted to pH 7.4, with a
1.2 mL/min flow rate. The standard isopropanol gradient reached 35%
in 3.5 min, which was maintained for 1 min, before returning to 0%
at 4.7 min. The cycle time was 6 min with an additional 1 min re-equilibration
time. The racemic warfarin showed separation of its enantiomers. The
retention times were calibrated using literature protein binding data
of nine marketed drugs.
Cytotoxicity Assay
Retinal pigment
epithelial ARPE19
cells (non-carcinoma cells) were grown and maintained in a DMEM F12
medium supplemented with 10% fetal bovine serum and 2 mM l-glutamine at 37 °C, 5% CO2. The cells were plated
in 96-well culture plates at a density of 10,000 cells per mL and
allowed to adhere at 37 °C for 24 h. The following day, various
doses of drugs or vehicle were added to the cells and further incubated
for 96 h. Then, the supernatant was removed and MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide] was added for 4 h. The ability of cells to form formazan
crystals by active mitochondrial respiration was determined by using
a Microplate reader after dissolving the crystals in DMSO. Cytotoxicity
was expressed as a relative percentage of the absorbance measured
at 540 nm in the control and drug-treated cells.
Real-Time
PCR
Hepa1c1c7 cells were seeded at 5 ×
105 cells/well of six-well plates. On the following day,
cells were treated with compound 11 (50 μM) or
vehicle (0.1% DMSO) for 16 h, in triplicates. Cells were then washed
twice in PBS, lysed, RNA was extracted, and cDNA was synthesized using
standard methods. The levels of mRNA for each gene were determined
by real-time quantitative PCR (TaqMan) and normalized for the levels
of 18S mRNA.
Cellular Thermal Shift Assay
HL-60
cells (3 ×
107) grown in RPMI 1640 (Invitrogen) supplemented with
10% (v/v) heat-inactivated fetal bovine serum (Invitrogen) were pelleted
by centrifugation at 300g for 4 min at room temperature
(RT), washed with phosphate-buffered saline (PBS) twice, resuspended
in PBS (3 mL) containing 1× protease inhibitor cocktail (Roche),
and snap frozen in liquid N2. Cell lysates were obtained
by four freeze–thaw cycles and cell debris were removed by
centrifugation at 17,000g for 15 min at 4 °C.
The supernatant (lysate) was split into two tubes, each containing
1.3 mL of lysate, and incubated with either 0.1% (v/v) DMSO or 50
μM of compound 11 at 37 °C for 1 h. Following
incubation, 100 μL of lysate was aliquoted into 12 PCR tubes
and subjected to various temperatures (38–60 °C) for 3
min using the VeriFlex blocks in the Veriti 96-well thermal cycler
(Thermo Fisher Scientific). After cooling for 3 min at 25 °C,
the samples were snap frozen. On the following day, the samples were
thawed out at 25 °C, and the insoluble fractions were removed
by centrifugation at 17,000g for 40 min at 4 °C.
The soluble fractions (60 μL) were transferred to fresh microcentrifuge
tubes containing 20 μL of 4× LDS buffer (Thermo Fisher
Scientific) and 8 μL of Sample Reducing Agent (Thermo Fisher
Scientific), mixed, and incubated at RT for 30 min. Proteins were
resolved by electrophoresis on 4–12% NuPAGE Bis-Tris gels (Thermo
Fisher Scientific) with MOPS running buffer and transferred onto a
0.45 μm nitrocellulose membrane. The membrane was blocked in
5% (w/v) non-fat milk in PBS-0.1% Tween 20 (Milk-PBST) for 1 h, incubated
overnight (16 h) at 4 °C with the primary rat monoclonal Keap1
antibody (1:2000 in Milk-PBST, MABS514, Millipore), washed thrice
in PBS-0.1% (v/v) Tween 20 (PBST) for 30 min, and incubated with goat
anti-rat 680/800 IRDye secondary antibody (1:20,000 in Milk-PBST)
for 1 h at RT protected from light. The blots were washed thrice in
PBST for 30 min before scanning using Odyssey (LI-COR) imager, where
the Keap1 band fluorescence intensity for each condition was quantified
and normalized to the 38 °C sample intensity.
Authors: V Calabrese; A Copani; D Testa; A Ravagna; F Spadaro; E Tendi; V G Nicoletti; A M Giuffrida Stella Journal: J Neurosci Res Date: 2000-06-01 Impact factor: 4.164
Authors: Phillip R Lazzara; Brian P David; Aparna Ankireddy; Benjamin G Richardson; Katherine Dye; Kiira M Ratia; Sekhar P Reddy; Terry W Moore Journal: J Med Chem Date: 2019-11-14 Impact factor: 7.446
Authors: Jong Yeon Hwang; Leggy A Arnold; Fangyi Zhu; Aaron Kosinski; Thomas J Mangano; Vincent Setola; Bryan L Roth; R Kiplin Guy Journal: J Med Chem Date: 2009-07-09 Impact factor: 7.446
Figure 1
Figure 2
Scheme 1
Scheme 2
Scheme 3
Scheme 4
Figure 3
Figure 4
Figure 5
Figure 6