Monoacidic Inhibitors of the Kelch-like ECH-Associated Protein 1: Nuclear Factor Erythroid 2-Related Factor 2 (KEAP1:NRF2) Protein-Protein Interaction with High Cell Potency Identified by Fragment-Based Discovery

Article abstract of DOI:10.1021/acs.jmedchem.6b00228

KEAP1 is the key regulator of the NRF2-mediated cytoprotective response, and increasingly recognized as a target for diseases involving oxidative stress. Pharmacological intervention has focused on molecules that decrease NRF2-ubiquitination through covalent modification of KEAP1 cysteine residues, but such electrophilic compounds lack selectivity and may be associated with off-target toxicity. We report here the first use of a fragment-based approach to directly target the KEAP1 Kelch-NRF2 interaction. X-ray crystallographic screening identified three distinct "hot-spots" for fragment binding within the NRF2 binding pocket of KEAP1, allowing progression of a weak fragment hit to molecules with nanomolar affinity for KEAP1 while maintaining drug-like properties. This work resulted in a promising lead compound which exhibits tight and selective binding to KEAP1, and activates the NRF2 antioxidant response in cellular and in vivo models, thereby providing a high quality chemical probe to explore the therapeutic potential of disrupting the KEAP1-NRF2 interaction.

Full text of DOI:10.1021/acs.jmedchem.6b00228

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Mono-acidic inhibitors of the Kelch-like ECH-associated protein 1 :
Nuclear factor erythroid 2–related factor 2 (KEAP1:NRF2) protein-protein
interaction with high cell potency identified by Fragment-based Discovery
Thomas G. Davies, William E Wixted, Joseph E. Coyle, Charlotte Griffiths-Jones, Keisha Hearn,
Rachel L. McMenamin, David Norton, Sharna J. Rich, Caroline Richardson, Gordon Saxty, Henriette
Maria Gerarda Willems, Alison J.-A. Woolford, Joshua E. Cottom, Jen-Pyng Kou, John G. Yonchuk,
Heidi G. Feldser, Yolanda Sanchez, Joseph P Foley, Brian J. Bolognese, Gregory A Logan,
Patricia L. Podolin, Hongxing Yan, James F Callahan, Tom Daniel Heightman, and Jeffrey K Kerns
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00228 • Publication Date (Web): 31 Mar 2016
Downloaded from http://pubs.acs.org on March 31, 2016
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 67
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
Mono-acidic inhibitors of the Kelch-like ECH-associated protein 1 : Nuclear
factor erythroid 2–related factor 2 (KEAP1:NRF2) protein-protein
interaction with high cell potency identified by Fragment-based Discovery
0
1
2
3
4
5
6
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
†
2†
1
1
Thomas G. Davies , William E. Wixted , Joseph E. Coyle , Charlotte GriffithsꢀJones , Keisha
1
1
1
1
1
Hearn , Rachel McMenamin , David Norton , Sharna J. Rich , Caroline Richardson , Gordon
1
1
1
3
2
Saxty , Henriëtte M. G. Willems , Alison J.ꢀA. Woolford , Joshua E. Cottom , JenꢀPyng Kou ,
2
2
2
2
2
John G. Yonchuk , Heidi G. Feldser , Yolanda Sanchez , Joseph P. Foley , Brian J. Bolognese ,
2
2
2
2
Gregory Logan , Patricia L. Podolin , Hongxing Yan , James F. Callahan , Tom D.
1
2
Heightman *, Jeffrey K. Kerns *
1
Astex Pharmaceuticals, 436 Cambridge Science Park, Cambridge, CB4 0QA, UK
2
GlaxoSmithKline Pharmaceuticals, 709 Swedeland Road, King of Prussia, PA 19406, USA
3
GlaxoSmithKline Pharmaceuticals, 1250 South Collegeville Road, Collegeville, PA 19426,
USA
*
Corresponding authors
†
Coꢀfirst authors
1
ACS Paragon Plus Environment

Page 3 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
KEAP1 is the key regulator of the NRF2ꢀmediated cytoprotective response, and increasingly
recognized as a target for diseases involving oxidative stress. Pharmacological intervention has
focussed on molecules that decrease NRF2ꢀubiquitination through covalent modification of
KEAP1 cysteine residues, but such electrophilic compounds lack selectivity and may be
associated with offꢀtarget toxicity. We report here the first use of a fragmentꢀbased approach to
directly target the KEAP1 Kelch–NRF2 interaction. Xꢀray crystallographic screening identified
three distinct “hotꢀspots” for fragment binding within the NRF2 binding pocket of KEAP1,
allowing progression of a weak fragment hit to molecules with nanomolar affinity for KEAP1
whilst maintaining drugꢀlike properties. This work resulted in a promising lead compound which
exhibits tight and selective binding to KEAP1, and activates the NRF2 antioxidant response in
cellular and in vivo models, thereby providing a high quality chemical probe to explore the
therapeutic potential of disrupting the KEAP1ꢀNRF2 interaction.
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
0
1
2
3
4
5
6
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
Living organisms are routinely exposed to reactive oxidants and electrophiles, both from internal
metabolism and environmental toxins. Reactive oxygen and nitrogen species can elicit harmful
oxidative stress, yet their controlled production is an essential component of cell signalling,
necessitating a complex antioxidant regulatory network. The Nuclear factor erythroid 2–related
factor 2 (NRF2) has emerged as a master regulator of cellular resistance to oxidants, mediating
the upꢀregulation of multiple phase 2 and cytoprotective enzymes and proteins (e.g.,
NAD(P)H:quinone acceptor oxidoreductase 1 (NQO1) in response to electrophilic and oxidative
1ꢀ3
assault.
The protein Kelchꢀlike ECHꢀassociated protein 1 (KEAP1) plays a key role in the regulation of
4
NRF2. It consists of three domains: an Nꢀterminal Broadꢀcomplex, Tramtrack and BricꢀaꢀBrac
(BTB) domain, serving as an adaptor for the E3 ubiquitin ligase Cul3/Rbx1; an Intervening
Region (IVR) or BACK domain; and a Cꢀterminal Kelch repeat domain, which is a proteinꢀ
5
ꢀ8
recognition module for NRF2 with a βꢀpropeller fold (Figure 1a). Under basal conditions,
NRF2 is sequestered by the dimeric KEAP1/Cul3 complex, becoming ubiquitinated and
degraded by the proteasome. An increase in levels of electrophilic species results in covalent
modification of cysteine residues in the BTB and IVR domains, and leads to dissociation of Cul3
2
,4,9ꢀ12
and other conformational changes.
This perturbs the ubiquitination of NRF2, allowing
translocation to the nucleus where it activates antioxidant response elements (AREs) to increase
1
3,14
expression of cytoprotective proteins.
Inadequate NRF2 signalling in the face of chronic
oxidative stress has been proposed as a pathological mechanism in inflammatory diseases; upꢀ
3
ACS Paragon Plus Environment

Page 5 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
regulation of the pathway might be beneficial in a range of therapeutic areas including
cardiovascular, neurodegeneration, and respiratory (e.g., chronic obstructive pulmonary disease)
15ꢀ19
conditions.
Several electrophilic small molecules that covalently modify the KEAP1 cysteine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
20
residues responsible for sensing oxidative stress show clinical promise. These include
2
1
derivatives of the triterpenoid 2ꢀcyanoꢀ3,12ꢀdioxooleanaꢀ1,9ꢀdienꢀ28ꢀoicꢀacid (CDDO), the
2
2
naturally occurring isothiocyanate sulforaphane, and dimethyl fumarate which was recently
2
3
approved by the FDA for the treatment of multiple sclerosis.
The direct, nonꢀcovalent antagonism of the KelchꢀNRF2 interaction (Figure 1b) may offer
advantages including a larger therapeutic index. NRF2 binds two Kelch subunits in the KEAP1
24ꢀ28
dimer through high affinity (“ETGE”) and low affinity (“DLG”)
motifs which interact with a
bowlꢀshaped pocket at the centre of the βꢀpropeller fold; a cellꢀpenetrant peptide containing the
ETGE” motif was recently reported to upꢀregulate protein levels associated with the antiꢀ
“
2
9
oxidant response. A number of small molecules binding at the KelchꢀNRF2 site have also been
3
0ꢀ37
33
identified,
and although encouraging levels of affinity have been achieved they show either
weak activity in cells or are poorly drugꢀlike. This is likely due to the presence of multiple
arginine residues in the NRF2ꢀbinding site (Arg 380, Arg 415 and Arg 483), favouring the
binding of poorly cellꢀpermeable compounds containing multiple carboxylic acids that mimic the
ETGE motif and its interactions with the Kelch domain (Figure 1c).
We report here the development of a small molecule inhibitor of the KEAP1 Kelch/NRF2
proteinꢀprotein interaction using a fragmentꢀbased drug discovery (FBDD) approach. The screen
identified three distinct “hotꢀspots” for fragment binding within the NRF2 binding pocket of
KEAP1, providing information on preferred interactions in these subsites which was used to
progress a weak fragment hit (K >1 mM) to molecules with nanomolar affinity for KEAP1. The
d
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 67
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
lead compound KIꢀ696 (compound 7, Table 1) exhibits tight and selective binding to the Kelch
domain, high potency in cellꢀbased assays, and activation of the NRF2 pathway in animal
models. It therefore provides a well validated in vivo active chemical probe with which to
explore the therapeutic potential of disrupting the Kelch/NRF2 interface.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
RESULTS
Identification of a potent and selective inhibitor of the KEAP1 Kelch/NRF2 interaction
It is well known that identifying proteinꢀprotein interaction inhibitors with favourable drugꢀlike
properties is challenging, and fragmentꢀbased approaches are increasingly recognised as a means
3
8,39
of generating potent and selective inhibitors with good ligand efficiency.
To enable Xꢀray crystallography as a primary screen, we used a mouse KEAP1 Kelch crystal
system which possesses an NRF2 binding site unoccluded by crystal contacts, and suitable for
highꢀthroughput soaking experiments. A crystallographic screen of approximately 330 fragments
led to the observation of binding in three discrete hotꢀspots within the NRF2 site, as exemplified
by compounds 1, 2 and 3, which form interactions in the vicinity of Arg 483, Tyr 525 and Ser
602 respectively (Figures 1d and 1e). These bound too weakly for reliable affinity measurement
by bioassay, as observed previously for fragment hits against “challenging” targets, which often
4
0
require sensitive biophysical techniques such as Xꢀray crystallography for their detection .
Nevertheless, the potency of such hits can often be rapidly optimized by structureꢀbased design
to bring them “onꢀscale”, and the phenyl acetic acid fragment 1 in particular was identified as a
promising “anchor fragment” for hit elaboration as discussed below.
5
ACS Paragon Plus Environment

Page 7 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Mimicking Glu 79 of the NRF2 ETGE motif, the carboxylic acid of 1 is engaged in a key
electrostatic interaction with Arg 483 on the KEAP1 Kelch domain (Figure 1f). It was known
from mutagenesis of NRF2 that electrostatic interactions with positively charged residues in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
27
Arg 483 region are important for binding , and an acidic moiety as exemplified in fragment 1
was anticipated to be required for potent activity in a lead molecule. Despite its low potency
(
IC >1 mM; Table 1), inspection of the complex structure suggested that 1 also had exit vectors
50
which would be well suited for fragment growing. For example, growth from the benzylic carbon
would allow interaction with the hotꢀspot occupied by fragment 2, whilst growth from the phenyl
ring, ortho to the chlorine, would allow access to the subsite identified by the sulphonamide
fragment 3. In addition, fragment binding elicited a conformational change in the side chain of
2
4
Arg 415 relative to apo and peptideꢀbound Kelch structures, creating an enclosed and
potentially more drugable interaction surface within the NRF2 pocket (Figure 1g).
This hit was subsequently modified in a stepꢀwise manner to fill the remainder of the NRF2 site
in a structureꢀdriven approach (Figure 2a), with potency assessed using both a fluorescence
polarization (FP) assay, and isothermal scanning calorimetry (ITC) (Table 1). We sought where
possible to exploit information provided by the ensemble of fragment hits, which provided a
direct insight into the minimal pharmacophore required for potent binding to the pocket.
The benzotriazole moiety was chosen for attachment directly to the benzylic carbon of the
phenyl acetic acid, supported by docking studies which suggested it could form a stacking
interaction with the sideꢀchain Tyr 525 whilst also satisfying the hydrogen bonding requirements
of both Gln 530 and Ser 555 with a bidentate acceptor motif. This initial modification gave a
sufficient increase in affinity to allow detection using the FP assay and ITC (IC = 61 µM; ITC
50
K = 59 µM). The coꢀstructure of compound 4 with KEAP1 shows that the benzotriazole
d
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 67
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
interacts with KEAP1 as predicted, and recapitulates the molecular recognition displayed by the
“aromaticꢀacceptor” fragment 2 (Figure 2a). Similarly, comparison of the KEAP1ꢀ4 structure
with that for sulfonamide fragment 3 suggested that growth from the 3ꢀposition of the
chlorophenyl ring of 4 with a sulfonamide group could reproduce the fragment’s hydrogen
bonding interaction with Ser 602. This was observed to be the case (compound 5, Figure 2b), and
resulted in a further 18ꢀfold increase in affinity (IC = 3.4 µM). Introduction of the sulfonamide
0
1
2
3
4
5
6
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
50
at this position was also observed to provide a useful exit vector with which to introduce
aromatic functionality to exploit the lipophilic edge of the pocket by πꢀstacking with Tyr 334
(
Figure 2c), as exemplified by compound 6 which exhibited subꢀmicromolar potency (IC = 0.27
50
µM). Throughout this optimization process there were only small perturbations of the binding
mode relative to the initial fragment hit, and good overlay of the iteratively introduced
pharmacophore elements such as the benzotriazole, suggesting well optimized interactions with
the NRF2ꢀbinding pocket (Figure 2c). The protein structure also displayed only modest
adjustments during the fragment elaboration, with the exception of Arg 415 which was observed
to adopt an alternative rotamer when bound to the metaꢀsubstituted compounds 5 and 6.
A number of small, structureꢀdriven modifications to compound 6 were then made to afford
further gains in potency. Replacement of the paraꢀchloroꢀphenyl with methyl was anticipated to
relieve potentially unfavourable intramolecular contacts with the sulfonamide centre, whilst an
electron donating methoxy group was introduced at the benzotriazole 7ꢀposition to increase the
hydrogen bond accepting potential of the ring nitrogens and add additional surface contacts.
Cyclization of the phenyl sulphonamide to form the fused 7ꢀmembered benzoxathiazepine was
introduced to rigidify the molecule in its bound conformation, and provided additional vectors to
fill the space between the sulphonamide and benzotriazole moieties. Although compounds 4 – 6
7
ACS Paragon Plus Environment

Page 9 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were prepared and soaked as the racemates, only the (S) enantiomers were observed to bind by
crystallography. The combination of the modifications to the initial fragment described above,
and the observation of this stereoselective binding led to the identification of 7 (Table 1; Figures
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
d and 2e). This compound fulfils the threeꢀpoint pharmacophore identified from the fragment
screen (acid, aromaticꢀacceptor and sulfonamide) and exhibits very high affinity for the KEAP1
Kelch domain (ITC K = 1.3 nM, Figure 2f).
d
To assess its selectivity towards the Kelch domain of KEAP1, compound 7 was screened against
4
1
GSK’s Enhanced Crossꢀscreen Panel (eXP), which consists of 49 in vitro functional assays for
targets which may represent toxicity liabilities. With the exception of the organic anion
transporting polypeptide 1B1 (OATP1B1) (IC = 2.5 µM), the bile salt export pump BSEP (IC
50
5
0
= 4.0 µM), and the phosphodiesterase PDE3A (IC = 10 µM), no significant crossꢀreactivity
5
0
was observed (see supplemental Table S1). No cytotoxicity was observed towards BEASꢀ2B
cells with compound 7 at concentrations up to 10 µM (cell titer glo assay). We therefore believe
that compound 7 possesses a suitable selectivity profile as a tool compound for cellꢀbased and in
vivo studies.
Comparison of NRF2 and compound 7 binding to the KEAP1 Kelch domain
To compare the interactions between compound 7 and KEAP1 with those formed by the tight
binding ETGE motif of NRF2, we overlaid Kelchꢀbound structures of the lead compound with
that for a fragment of NRF2 (Leu 76–Leu 84; sequence LDEETGEFL) containing the ETGE
24
sequence embedded within a 9ꢀmer βꢀturn (PDB 1x2r ; Figure 3a). A key observation from this
is that compound 7 is not a peptidomimetic, with little overlap between its binding footprint and
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 67
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
that of NRF2. Nevertheless a careful analysis does reveal a conservation of key pharmacophore
points, as described below.
As evident from the fragment hit, the main point of similarity with the natural ligand is the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
carboxylic acid of compound 7, which occupies a similar region of space to Glu 79 of NRF2
(Figure 3b), both of which form electrostatic interactions with the basic sideꢀchain of Arg 483.
As mentioned earlier, mutational analysis of NRF2 had previously demonstrated the importance
27
of Glu 79 for tight binding , and our data confirm the presence of this key hotꢀspot in the Kelchꢀ
NRF2 interface. In addition, stacking interactions of the planar peptide backbone between Glu 78
and Glu 79 are partially mimicked by the benzotriazole ring, and there is a good overlay of the
peptide carbonyl oxygen and the N2 benzotriazole nitrogen, both of which accept a hydrogen
bond from the sideꢀchain of Gln 530 (Figure 3c). The potential to build potency by interaction
with this region was evident from the fragment screen with a number of small aromatic
heterocycles (eg 2) observed to bind here. Away from these regions, the interactions formed by
the lead and ETGE peptide diverge. Although the phenylꢀsulfonamide of compound 7 occupies a
region of the pocket broadly similar to the NRF2 Glu 82 sideꢀchain, and both exploit a
hydrophobic stack with Tyr 334, the electrostatic interaction formed between Arg 380 and NRF2
Glu 82 is not present for our lead. This result, which is consistent with the detection of the
uncharged sulphonamide fragment 3, highlights that a second acidic moiety at this position is not
required for potency if other interactions are efficiently optimized. Similar to the chloroꢀphenyl
in the fragment hit, the tolyl group of the lead occupies a region of the pocket at the entrance to
the central tunnel which is unexploited by the ETGE peptide, and leads to a change in rotamer of
Arg 415.
9
ACS Paragon Plus Environment

Page 11 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Much of the remainder of the ETGEꢀcontaining turn makes little contact with the protein, and
presumably provides a scaffold to preꢀform a globular architecture complementary to the bowl of
the KEAP1 pocket whilst rigidly orientating the glutamate residues in an optimal position. By
contrast, compound 7 fills the pocket more efficiently, with a much larger proportion forming
direct interactions with the NRF2 site (Figure 3d). The result is that 7 (MW = 551) exhibits
higher affinity and efficiency in binding compared with much larger and more hydrophillic
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
26
NRF2ꢀderived peptides (e.g., Leu 76ꢀLeu 84 (Figure 3a); K = 182 nM; MW = 1052), resulting
d
in a compound with physicochemical properties that are suitable for progression to cellꢀbased
and in vivo studies. Of particular note is that compound 7 contains only a single carboxylic acid
functionality, compared with five in the Leu 76ꢀLeu 84 peptide.
Assessment of compound 7 in cellular and in vivo models
Chronic obstructive pulmonary disease (COPD) is a disease associated with excessive levels of
electrophilic stress and inflammation, and activation of the KEAP1/NRF2 pathway by covalent
modification of KEAP1 cysteine residues has been shown to reduce cigarette smokeꢀinduced
21,42,43
emphysema (a key feature of COPD) in an NRF2ꢀdependent fashion in mice.
An
assessment in cellular and in vivo models of relevance to this disease area therefore provides an
opportunity to assess the pharmacological effects and therapeutic relevance of disrupting the
KelchꢀNRF2 interaction, and allows comparison with covalent activators of the NRF2 pathway.
Multiple sources of human lung epithelial cells were treated with compound 7 to gain an
understanding of pathway activation in vitro, followed by in vivo studies with Han Wistar rats.
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 67
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
Compound 7 increases NRF2 Nuclear Translocation in Normal Human Bronchial Epithelial
cells
To assess whether disruption of the KelchꢀNRF2 interaction by compound 7 can increase levels
of NRF2 in a cellular context, we firstly measured the amount of nuclear NRF2 protein
accumulation in Normal Human Bronchial Epithelial (NHBE) cells. Cells were treated with
compounds or DMSO vehicle for 6 hours, after which cytoplasmic and nuclear fractions were
generated for assessment of NRF2 protein levels. Compound 7 increased the accumulation of
nuclear NRF2 protein in a concentrationꢀdependent manner (supplemental Figure S1). The
commercially available NRF2 activators tertꢀbutylhydroquinone (tBHQ) and CDDOꢀMe were
included as positive controls.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 7 upꢀregulates NRF2ꢀdependent gene expression
We next assessed the ability of compound 7 to upregulate NRF2ꢀdependent gene expression in
human primary lung epithelial cells. NHBE cells were transfected with 25 nM nonꢀtargeting or
NRF2 siRNA for 48 hours followed by 24 hours treatment with DMSO, compound 7, or CDDOꢀ
Me, which activates NRF2 through covalent modification of KEAP1 BTB. Compound 7
increased mRNA expression of the NRF2ꢀdependent genes NQO1 and GCLM in NHBE cells
transfected with the nonꢀtargeting siRNA, while NRF2 gene silencing significantly decreased
compound activity (Figure 4a). Similar results were seen in cells treated with CDDOꢀMe.
Confirmation that NRF2 siRNA treatment effectively decreases NRF2 mRNA expression is
shown in supplemental Figure S2.
11
ACS Paragon Plus Environment

Page 13 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compound 7 increased NQO1 Activity in an NRF2ꢀDependent Manner
To confirm that compound 7 activated NRF2 and led to the induction of the NRF2ꢀdependent
downstream signalling cascade, we measured the ability of 7 to increase NQO1 enzyme activity.
NHBE cells were transfected with NRF2 or nonꢀtargeting siRNA for 48 hours before treatment
with DMSO, 100 nM CDDOꢀMe, or 1 µM compound 7 for another 48 hours. NQO1 specific
enzyme activity was then assayed in crude cell lysates. Figure 4b shows that compound 7
induced NQO1 specific enzyme activity by 6ꢀfold over the DMSO control and NRF2 siRNA
abolished approximately 88% of the activity above vehicleꢀtreated, nonꢀtargeting siRNA levels.
Similar NRF2ꢀdependent activity was seen in cells treated with CDDOꢀMe. The ability of 7 to
increase NQO1 activity was also confirmed in a human lung epithelial cell line, BEASꢀ2B, with
an EC of 12 nM (Fig 4c).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
Compound 7 prevents tBHPꢀinduced glutathione depletion
The in vitro functional activity and cytoprotective effects of compound 7 were next assessed in
response to oxidative stress. NHBE cells were preꢀtreated with compounds or DMSO for
approximately 18 hours before addition of 300 µM tertꢀbutyl hydroperoxide (tBHP), an oxidative
stress agent that causes rapid oxidation of reduced glutathione (GSH) and other events that
4
4
decrease cell viability. As shown in Figure 4d, treatment with 300 µM tBHP decreased GSH
levels by approximately 63% compared to the media control. The γꢀglutamylcysteine synthetase
inhibitor buthionine sulfoxide (BSO) and glutathione precursor Nꢀacetyl cysteine (NAC) were
used as internal controls. The application of 1 mM NAC increased total glutathione levels
approximately 4ꢀfold over media alone, whereas 3 mM BSO decreased glutathione content by
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 67
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
approximately 37%. Preꢀtreatment with NAC or the NRF2ꢀactivating compound CDDOꢀMe
prevented tBHPꢀinduced glutathione depletion. The Kelch targeting compound 7 increased
intracellular glutathione levels in a concentrationꢀdependent manner. In addition, morphological
changes were also examined in NHBE cells with DMSO, tBHP, or compound 7 + tBHP
treatment. Treatment with tBHP clearly had a detrimental effect on cell health and appearance
while preꢀtreatment of cells with 1 µM compound 7 before the exposure to tBHP maintained cell
morphology consistent with the DMSO control (supplemental Figure S3).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 7 Induces the Expression of NRF2ꢀRegulated Genes in COPD patientꢀderived
bronchial epithelial cells
We next evaluated the ability of compound 7 to upꢀregulate NRF2ꢀdependent gene expression in
human diseaseꢀderived cells. Bronchial epithelial cells from human COPD patient lung brushings
were grown and differentiated into airꢀliquid interface (ALI) culture conditions. ALI cells were
treated for 6 hrs with compound 7 and expression of NRF2ꢀdependent genes, including NQO1,
TXNRD1, GCLM and HOꢀ1 was measured using realꢀtime PCR. As shown in Figure 4e,
compound 7 efficiently upꢀregulated expression of each of these genes in a concentrationꢀ
dependent manner. The EC obtained for NQO1 mRNA expression in the lung brushing cells
5
0
correlates closely with the EC observed with the NQO1 activity assay in BEASꢀ2B cells (22
50
nM and 12 nM, respectively). These data demonstrate the ability of compound 7 to upregulate
NRF2 signalling not only in healthy cells as previously shown, but also in a translational model
of human disease.
13
ACS Paragon Plus Environment

Page 15 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compound 7 Induces the Pulmonary Expression of NRF2ꢀRegulated Genes in vivo
In order to study the in vivo effects of compound 7, IV and oral pharmacokinetic parameters
were determined (Figure S4). A relatively high clearance and short halfꢀlife were observed after
IV dosing, with low oral bioavailability (F = 7%). Hence, IV infusions were used in order to
achieve sustained target engagement when studying the in vivo pharmacology of compound 7.
To confirm engagement of the NRF2 pathway by compound 7 in vivo, Han Wistar rats were
administered the compound by IV infusion over 6 hours, and lung tissue was analyzed for
expression of the NRF2ꢀregulated genes Nqo1, Hoꢀ1, Txnrd1, Srxn1, Gsta3 and Gclc. As shown
in Figure 5a, compound 7 induced the expression of each of the 6 genes in a doseꢀdependent
manner, with maximum increases over vehicle controls of 37ꢀ (Nqo1), 17ꢀ (Hoꢀ1), 9ꢀ(Txnrd1),
28ꢀ (Srxn1), 15ꢀ (Gsta3) and 13ꢀfold (Gclc) occurring at the 50 µmol/kg dose. EC values were
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
4
4.0, 25.7, 42.6, 33.8, 28.4, and 44.1 µmol/kg, respectively, giving an average EC value of 36.4
50
+
3.4 µmol/kg.
Compound 7 Attenuates OzoneꢀInduced Pulmonary Inflammation in vivo
To examine the effect of compound 7 under conditions of oxidative stress, rats were
administered the compound at 35 µmol/kg (the approximate average EC50 value for gene
expression) by IV infusion over 6 hours, and after 24 hours were exposed to ozone (1 ppm for 3
hours). Fifteen minutes following the termination of ozone exposure, the numbers of total cells,
neutrophils and mononuclear cells in the bronchoalveolar lavage (BAL) fluid were significantly
increased in vehicleꢀtreated ozoneꢀexposed rats compared to vehicleꢀtreated airꢀexposed rats
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 67
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
(
Figure 5b). The ozoneꢀinduced increase in each cell subset was significantly reduced as a result
of compound 7 treatment.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 7 Restores OzoneꢀInduced Depletion of Lung GSH Levels in vivo
In a subsequent study, compound 7 was administered to rats at 10, 35 and 50 ꢀmol/kg by IV
infusion, resulting in steady state compound concentrations in the blood of 407 + 44 nM, 946 +
5
0 nM and 1437 + 186 nM, respectively, over the 6 hour infusion period. Exposure to ozone 24
hours postꢀdose produced a significant depletion in lung levels of the antiꢀoxidant molecule,
GSH, which was restored by compound 7 in a doseꢀdependent manner (Figure 5c).
Synthesis
Fragments 1 ꢀ 3 were purchased from commercial sources. The propionic acid 4 was synthesized
according to Scheme 1. Addition of 4ꢀchlorophenylmagnesium bromide to the benzotriazolyl
aldehyde 10 gave the carbinol 11, which was oxidized to the ketone 12 using DessꢀMartin
periodinane. The carboxy side chain was then introduced by a HornerꢀWadsworthꢀEmmons
condensation with ketone 12 followed by reduction with nickel borohydride and in situ
saponification to give acid 4 in four steps overall.
The methanesulfonamideꢀsubstituted aryl propionic acid 5 was prepared according to Scheme 2.
First, the Nꢀmethyl methanesulfonamide group was introduced by displacement of the benzylic
bromide 14. Next, HornerꢀWadsworthꢀEmmons condensation gave the cinnamate ester 16,
15
ACS Paragon Plus Environment

Page 17 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
which allowed rhodiumꢀcatalysed conjugate addition of the boronate ester 18 to give, after
saponification, the acid 5 in five steps overall.
The benzenesulfonamide 6 was prepared using an analogous, but complementary route (Scheme
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
). HornerꢀWadsworthꢀEmmons condensation with the benzotriazolyl aldehyde 10 gave the
cinnamate ester 19, which was subjected to rhodiumꢀcatalysed conjugate addition of 4ꢀchloroꢀ3ꢀ
hydroxymethylphenyl pinacol boronate ester to give the intermediate bisaryl propionate 20,
containing a hydroxymethyl group as a handle for further derivatisation. Displacement of the
hydroxyl group with Nꢀmethyl benzenesulfonamide under Mitsunobu conditions followed by
chiral HPLC gave the (S)ꢀmethyl ester 21, which was saponified to give the acid 6 in five steps
overall.
The benzoxathiazepine 7 was prepared according to Scheme 4. This required the synthesis of the
5
ꢀbromoꢀ1ꢀmethylꢀ7ꢀmethoxybenzotriazole intermediate 7. Starting from 2ꢀaminoꢀ3ꢀnitrophenol
(22), selective Oꢀmethylation followed by bromination under acidic conditions gave anisole 24,
which was Nꢀmethylated and then reduced to give the diaminoanisole 26. Diazotization with
sodium nitrite then gave the benzotriazole 27. This was converted to the cinnamate ester 28 by
Heck coupling, allowing rhodiumꢀcatalysed conjugate addition of 4ꢀchloroꢀ3ꢀ
hydroxymethylphenyl pinacol boronate ester to give the intermediate bisaryl propionate 29. The
key benzoxathiazepine intermediate 32 was prepared by sulfonylation of (R)ꢀ1ꢀaminopropanol
(30) with 2ꢀfluorobenzenesulfonamide followed by cyclization under basic conditions. Coupling
of the two intermediates 29 and 32 under Mitsunobu conditions followed by saponification and
chiral HPLC gave the desired stereoisomer 7 in a total of 12 steps.
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 67
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
DISCUSSION
We have described here the successful identification of compound 7, a highly potent and
selective inhibitor of the KEAP1 KelchꢀNRF2 interaction with the ability to stimulate the NRF2
pathway in both human cellꢀbased assays, and in vivo. KEAP1 is increasingly recognized as a
key node for therapeutic intervention in a range of diseases involving chronic oxidative stress,
including respiratory and cardiovascular indications, diabetic nephropathy and
neurodegeneration. Indeed, the approved multiple sclerosis drug Tecfidera is a NRF2 activator,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
and is believed to operate at least in part through this mechanism of action. However,
compounds that have so far been progressed to clinical trials lack selectivity due to their
electrophilic mode of action, and this may lead to pleiotropic effects. For example, an analogue
of the widely studied NRF2 inducer CDDO has been shown to react with over 500 molecular
4
5
species, and the therapeutic outcome of such compounds is likely to result from activities
against multiple targets. The inherent lack of selectivity observed for this class of compound may
lead to a low therapeutic index, but also hinders attempts to clearly dissect and understand the
consequences of pharmacological activation of NRF2. By contrast, nonꢀcovalent disruption of
the KelchꢀNRF2 interaction is likely to afford the opportunity for more specific modulation of
this pathway.
Although a number of smallꢀmolecule inhibitors of the KEAP1 KelchꢀNRF2 interaction have
recently been identified, an important aspect of this work was the use of a fragmentꢀbased
approach which allowed the development of a novel, nonꢀpeptidomimetic inhibitor that
combines high potency with more drugꢀlike physicochemical properties. For example, the
fragment screen highlighted the ability of nonꢀcharged moieties to bind close to the region
occupied by the NRF2 Glu 82 sideꢀchain, and we have shown here that a single acidic
17
ACS Paragon Plus Environment

Page 19 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
functionality is sufficient to achieve highly potent binding in an elaborated compound. This
detailed understanding of the KEAP1 Kelch pharmacophore afforded by the fragment screen,
together with a parsimonious fragments to leads design approach, allowed the progression from a
hit with >1 mM potency to a nanomolar lead compound which was well suited for biological
evaluation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 7 showed benefit in both cell and animal models of oxidative stressꢀinduced
respiratory disease. Compound 7 induced the expression of NRF2ꢀregulated genes such as
NQO1, as well as the corresponding NQO1 enzyme activity, in normal human bronchial
epithelial cells. The inhibition of these inductions in cells treated with NRF2ꢀtargeting siRNA,
but not with nonꢀtargeting siRNA, confirms that compound 7 exerts these effects through NRF2.
Similar induction of NRF2ꢀregulated gene expression was observed in the lungs of compound 7ꢀ
treated rats, demonstrating compound activity in vivo. Importantly, compound 7ꢀinduced gene
expression was also observed in bronchial epithelial cells derived from COPD patients,
providing a translational link to the human disease state.
GSH is a molecule that plays a crucial role in protecting cells from oxidative damage, leading to
the pursuit of therapeutic strategies that modulate GSH content. In both our in vitro and in vivo
models, exposure to oxidants such as tBHP and ozone resulted in decreases in GSH content,
which were restored in a doseꢀdependent manner by compound 7. In vivo ozone exposure also
induces the influx of inflammatory cells into the lungs, allowing for examination of the antiꢀ
inflammatory effects of potential therapeutics. Here we demonstrate that the ozone induces
increases in BAL fluid total cells, neutrophils and mononuclear cells that can be significantly
inhibited by compound 7.
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 67
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
Collectively, these studies provide consistent evidence across human cellꢀbased and in vivo
models that compound 7 activates the NRF2 pathway, resulting in induction of target gene
expression and downstream antiꢀoxidant and antiꢀinflammatory activities. To the best of our
knowledge, compound 7 represents the most potent smallꢀmolecule KEAP1ꢀNRF2 inhibitor so
far reported, and the demonstration of its high selectivity and detailed mode of action studies
both in vitro and in vivo allow its observed phenotypic effects to be confidently related to onꢀ
target activity. In fulfilling these criteria, compound 7 provides an important new chemical probe
with which to further study the biology of the KEAP1/NRF2 axis, and represents a promising
lead compound for diseases where maladaptive NRF2 pathway activation is implicated.
0
1
2
3
4
5
6
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
ACS Paragon Plus Environment

Page 21 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL PROCEDURES
Synthesis of Compounds 4ꢀ7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
All solvents and commercially available reagents were used as received. All reactions were
followed by TLC analysis (TLC plates GF254, Merck) or LCꢀMS (liquid chromatography mass
spectrometry). LCꢀMS analysis was performed with an Agilent or Shimadzu LC system with
variable wavelength UV detection using reverse phase chromatography with a CH CN and water
3
gradient with a 0.02 or 0.1% TFA modifier (added to each solvent) and using a reverse phase
column, e.g., Thermo Hypersil Gold C18. MS was determined using either PE Sciex 150EX LCꢀ
MS, Waters ZQ LCꢀMS, or Agilent 6140 LCꢀMS Single Quadrupole instruments. Column
chromatography was performed on prepacked silica gel columns (3090 mesh, IST) using a
1
Biotage SP4 or similar. NMR spectra are referenced as follows: H (400 MHz), internal standard
TMS at δ = 0.00. Abbreviations for multiplicities observed in NMR spectra: s; singlet; br s,
broad singlet; d, doublet; t, triplet; q, quadruplet; p, pentuplet; spt, septuplet; m, multiplet. All
compounds reported are of at least 95% purity according to LCꢀMS unless stated otherwise.
3-(4-Chlorophenyl)-3-(1-methyl-1H-1,2,3-benzotriazol-5-yl) propanoic acid (4)
A stirred solution of tꢀBuOK (0.372 g, 3.31 mmol) in THF (22 mL) at 0°C under N was treated
2
with trimethyl phosphonoacetate (0.903 g, 4.86 mmol). After 10 mins, 5ꢀ(4ꢀchlorobenzoyl)ꢀ1ꢀ
methylꢀ1Hꢀ1,2,3ꢀbenzotriazole (0.600 g, 2.21 mmol) was added slowly. The mixture was
allowed to warm to ambient temperature and then stirred for 72 h. The mixture was quenched
with NH Cl (aq., sat.), diluted with water and then extracted with nꢀheptane (x3). The combined
4
organic layers were washed with water and brine, dried over MgSO , filtered and concentrated to
4
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 67
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
dryness to give a colourless solid, used without further purification (640 mg, 88%). LCꢀMS
+
[
M+H] 328/330 (Cl), rt 1.40/1.43 min (2 isomers). A stirred solution of this material (0.640 g,
1
.95 mmol) in MeOH (20 mL) under N was treated with NiCl .6H O (0.093 g, 0.39 mmol), then
2 2 2
0
1
2
3
4
5
6
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
cooled to 0°C and treated with NaBH (0.222 g, 5.86 mmol) slowly in portions. The mixture was
4
allowed to warm to ambient temperature and stirred for 4 h, then treated with NaOH (2M, 40
mL) and stirred overnight. The mixture was then quenched with NH Cl (sat., aq.) and extracted
4
with EtOAc. The organic layer was dried over MgSO , filtered and concentrated under reduced
4
pressure. The residue was purified by phenomenix silica cartridges by gradient elution (100%
DCM to 50 % EtOAc in DCM) to give the product as a buff solid (0.307 g, 50%). LCꢀMS
+
1
[
M+H] 316/318 (Cl), rt 1.24 min. H NMR (400 MHz, Meꢀd ꢀOD): 7.92 (s, 1H), 7.68 (d, 1H),
3
7.49 (dd, 1H), 7.38ꢀ7.26 (m, 4H), 4.73 (t, 1H), 4.31 (s, 3H), 3.23ꢀ3.10 (m, 2H).
3-{4-Chloro-3-[(N-methylmethanesulfonamido)methyl]phenyl}-3-(1-methyl-1H-1,2,3-
benzotriazol-5-yl)propanoic acid (5)
A stirred mixture of 5ꢀbromoꢀ1ꢀmethylbenzotriazole (prepared according to PCT Int. Appl.,
2012119046; 1.28 g, 6.04 mmol), bis(pinacolato)diboron (1.69 g, 6.04 mmol), KOAc (1.18 g,
12.08 mmol) and PdCl dppf (0.221 g, 0.30 mmol) in dioxane was degassed with N for 10 mins
2
2
and then heated at 90°C for 5 hours. After cooling the mixture was partitioned between water
and EtOAc. The organic phase was dried (MgSO ), filtered and concentrated to dryness. The
4
residue was purified by flash chromatography (EtOAc/petrol gradient 0ꢀ40%) to give 1ꢀmethylꢀ
5ꢀ(4,4,5,5ꢀtetramethylꢀ[1,3,2]dioxaborolanꢀ2ꢀyl)ꢀ1Hꢀbenzotriazole (1.21 g, 72%). LCꢀMS
+
[
M+H] 260, rt 1.30 min. A stirred mixture of this material (0.191 g, 0.74 mmol), (E)ꢀ3ꢀ{4ꢀ
21
ACS Paragon Plus Environment

Page 23 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
chloroꢀ3ꢀ[(methanesulfonylꢀmethylꢀamino)ꢀmethyl]ꢀphenyl}ꢀacrylic acid methyl ester (0.117 g,
0
.37 mmol), [RhCl(cod)] (0.009 g, 0.02 mmol), triethylamine (0.056 g, 0.55 mmol), 1,4ꢀdioxane
2
(0.18 mL) and water (1.23 mL) was heated in a reactivial at 95°C for 18 hours. LiOH (1M, 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
mL) was added and the mixture stirred at ambient temperature for 2 hours. The mixture was
treated with citric acid (5%, aq.) and extracted with CHCl :IPA (3:1, x2). The combined organic
3
layers were washed successively with water and brine, dried over MgSO , filtered and
4
concentrated to dryness. Purification by preparative HPLC (Agilent, formic acid method and
+
Agilent TFA) gave the product as a colourless oil (0.028 g, 17%). LCꢀMS [M+H] 437/439 (Cl),
1
rt 0.99 min. H NMR (400 MHz, Meꢀd ꢀOD): 7.92 (s, 1H), 7.69 (d, 1H), 7.56ꢀ7.46 (m, 2H), 7.37
3
(d, 1H), 7.29 (dd, 1H), 4.75 (t, 1H), 4.44 (s, 2H), 4.31 (s, 3H), 3.18 (d, 2H), 2.92 (s, 3H), 2.76 (s,
3H).
(3S)-3-{4-Chloro-3-[(N-methylbenzenesulfonamido)methyl]phenyl}-3-(1-methyl-1H-1,2,3-
benzotriazol-5-yl)propanoic acid (6)
A stirred solution of (S)ꢀmethyl 3ꢀ(4ꢀchloroꢀ3ꢀ((Nꢀmethylphenylsulfonamido) methyl)phenyl)ꢀ3ꢀ
(1ꢀmethylꢀ1Hꢀbenzo[d][1,2,3]triazolꢀ5ꢀyl)propanoate (0.180 g, 0.351 mmol) in methanol (15
mL) and THF (1 mL) at ambient temperature was treated with a 1M solution of lithium
hydroxide (16 mL, 16.00 mmol). The resulting suspension was diluted with methanol (10 mL).
After 1 hr 20 mins at ambient temperature the reaction mixture was acidified to pH 1ꢀ2 and the
solution diluted with ethyl acetate. The organic phase was washed with water (2x), dried over
magnesium sulfate, filtered, and concentrated to give 0.156 g of a light yellow solid. This
material was dissolved in chloroform and concentrated (3x) to give 0.141 g of the title compound
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 67
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
1
as a light yellow solid. LCꢀMS [M+H] 499/501 (Cl), rt 1.03 min. H NMR (400 MHz,
methanolꢀd ) δ 7.83 ꢀ 7.90 (m, 3H), 7.59 ꢀ 7.71 (m, 4H), 7.46 (d, J=8.8 Hz, 1H), 7.39 (s, 1H),
4
7
.30 ꢀ 7.34 (m, 1H), 7.24 ꢀ 7.28 (m, 1H), 4.70 (t, J=8.2 Hz, 1H), 4.28 ꢀ 4.33 (m, 5H), 3.12 (dd,
0
1
2
3
4
5
6
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
J=7.8, 3.8 Hz, 2H), 2.61 (s, 3H).
(
3S)-3-(7-Methoxy-1-methyl-1H-benzo[d][1,2,3]triazol-5-yl)-3-(4-methyl-3-(((R)-4-methyl-
1
,1-dioxido-3,4-dihydro-2H-benzo[b][1,4,5]oxathiazepin-2-yl)methyl)phenyl)propanoic
acid (7)
To a solution of ethyl 3ꢀ(3ꢀ(hydroxymethyl)ꢀ4ꢀmethylphenyl)ꢀ3ꢀ(7ꢀmethoxyꢀ1ꢀmethylꢀ1Hꢀ
benzo[d][1,2,3]triazolꢀ5ꢀyl)propanoate (980 mg, 2.56 mmol) in THF (30 mL) was added (R)ꢀ4ꢀ
methylꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[b][1,4,5]oxathiazepine 1,1ꢀdioxide (818 mg, 3.83 mmol), PSꢀPPh3
(
3195 mg, 5.11 mmol) and then was added diisopropyl azodicarboxylate (DIAD) (0.994 mL,
5
.11 mmol) in tetrahydrofuran (THF) (10 mL). The resulting reaction mixture was stirred at
room temperature for 30 min. The reaction mixture was filtered and the filtrate was concentrated
under reduced pressure, purified with flash chromatography over silica gel column to afford
desired crude intermediate ethyl 3ꢀ(7ꢀmethoxyꢀ1ꢀmethylꢀ1Hꢀbenzo[d][1,2,3]triazolꢀ5ꢀyl)ꢀ3ꢀ(4ꢀ
methylꢀ3ꢀ(((R)ꢀ4ꢀmethylꢀ1,1ꢀdioxidoꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[b][1,4,5]oxathiazepinꢀ2ꢀ
yl)methyl)phenyl)propanoate (1.30 g, 88%). This crude intermediate was dissolved in methanol
(30.0 mL) then was added NaOH (2 N, 6.39 mL, 12.8 mmol). The resulting reaction mixture was
stirred at 80ºC for 40 min. The reaction mixture was acidified with HCl (1 N) to pH ~3,
concentrated under reduced pressure, extracted with EtOAc (3 x 50 mL). The combined organic
layers were washed with brine (50 mL), dried over MgSO , filtered, and concentrated under
4
reduced
pressure
to
afford
the
desired
product
3ꢀ(7ꢀmethoxyꢀ1ꢀmethylꢀ1Hꢀ
23
ACS Paragon Plus Environment

Page 25 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzo[d][1,2,3]triazolꢀ5ꢀyl)ꢀ3ꢀ(4ꢀmethylꢀ3ꢀ(((R)ꢀ4ꢀmethylꢀ1,1dioxidoꢀ3,4ꢀdihydroꢀ2Hꢀ
benzo[b][1,4,5]oxathiazepinꢀ2ꢀyl)methyl)phenyl)propanoic acid (1.41 g, 100%). The title
compound was obtained by SFC chiral purification to afford the desired enantiomerically pure
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
1
product (0.367 g, 26%). LCꢀMS [M+H] 551, rt 1.02 min. H NMR (400 MHz, DMSOꢀd ) δ
6
1.11 (d, J=6.27 Hz, 3 H), 2.24 (s, 3 H), 2.76 (d, J=15.06 Hz, 1 H), 3.06 ꢀ 3.16 (m, 2 H), 3.62 (dd,
J=15.31, 10.29 Hz, 1 H), 3.82 (d, J=14.05 Hz, 1 H), 3.94 (s, 3 H), 4.34 ꢀ 4.47 (m, 6 H), 4.51 (t,
J=7.78 Hz, 1 H), 6.93 (s, 1 H), 7.12 (d, J=7.78 Hz, 1 H), 7.28 ꢀ 7.32 (m, 2 H), 7.36 ꢀ 7.38 (m, 2
1
H), 7.45 (s, 1 H), 7.66 (t, J=7.65 Hz, 1 H), 7.78 (d, J=7.53 Hz, 1 H), 12.08 (br s, 1 H). The H
NMR spectrum for compound 7 is given in Supplemental Figure S5.
(4-Chlorophenyl)(1-methyl-1H-1,2,3-benzotriazol-5-yl)methanol (11)
A stirred solution of 1ꢀmethylꢀ1Hꢀ1,2,3ꢀbenzotriazoleꢀ5ꢀcarbaldehyde (0.500 g, 3.10 mmol) in
THF at 0°C under N was treated with a solution of 4ꢀchlorophenylmagnesium bromide in THF
2
(1M, 3.72 mL, 3.72 mmol). After 2 hours the mixture was quenched with NH Cl (sat., aq.) and
4
warmed to ambient temperature. The mixture was extracted with DCM (x3) and the combined
organic layers were washed with water followed by brine, dried over MgSO , filtered and
4
concentrated to dryness to give a yellow solid, used without further purification (0.800 mg,
+
1
9
4%). LCꢀMS [M+H] 274/276 (Cl), rt 1.23 min. H NMR (400 MHz, DMSOꢀd ): 8.03 (s, 1H),
6
7
.75 (d, 1H), 7.49 (dd, 1H), 7.43 (d, 2H), 7.36 (d, 2H), 6.15 (d, 1H), 5.91 (d, 1H), 4.27 (s, 3H).
5
-(4-Chlorobenzoyl)-1-methyl-1H-1,2,3-benzotriazole (12)
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 67
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
A solution of (4ꢀchlorophenyl)(1ꢀmethylꢀ1Hꢀ1,2,3ꢀbenzotriazolꢀ5ꢀyl)methanol (0.800 g, 2.92
mmol) in dichloromethane (29 mL) was treated with DessꢀMartin periodinane (1.49 g, 3.51
mmol) at 0°C, under N . After stirring for 1 h at ambient temperature, the reaction mixture was
2
0
1
2
3
4
5
6
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
diluted with sodium bicarbonate (sat., aq.) and extracted with CHCl (x3). The combined
3
organic layers were washed with water followed by brine, dried over MgSO , filtered and
4
concentrated to dryness to leave a yellow solid. This was purified by phenomenix silica cartridge
(
100% CHCl ) and the residue was washed with a small amount of MeOH to yield a colourless
3
+
1
solid (600 mg, 76%). LCꢀMS [M+H] 272/274 (Cl), rt 1.36 min. H NMR (400 MHz, DMSOꢀ
d6): 8.37 (s, 1H), 8.04 (s, 1H), 7.98 (dd, 1H), 7.82 (d, 2H), 7.67 (d, 2H), 4.39 (s, 3H).
(E)-3-{4-Chloro-3-[(methanesulfonyl-methyl-amino)-methyl]-phenyl}-acrylic acid methyl
ester (16)
A mixture of 4ꢀchloroꢀ3ꢀ(bromomethyl)benzaldehyde (0.25g, 1.07 mmol), Nꢀmethyl
methanesulfonamide (0.129g, 1.18 mmol) and K CO (0.163g, 1.18 mmol) in MeCN was stirred
2
3
at ambient temperature under N overnight. The mixture was diluted with water and extracted
2
with DCM (x3). The combined organic layers were washed with brine and dried over MgSO4.
Filtration and concentration to dryness afforded Nꢀ(2ꢀchloroꢀ5ꢀformylꢀbenzyl)ꢀNꢀmethylꢀ
methanesulfonamide as a yellow oil (0.247g, 88%), used without further purification. LCꢀMS
+
[
M+18] 279/281 (Cl isotopes), rt 1.15 min. A stirred solution of tꢀBuOK (0.212g, 1.89 mmol) in
THF (10 mL) at 0°C under N was treated with trimethyl phosphonoacetate (0.403g, 2.17 mmol).
2
After 30 mins, a solution of Nꢀ(2ꢀchloroꢀ5ꢀformylꢀbenzyl)ꢀNꢀmethylꢀmethanesulfonamide
(0.247g, 0.94 mmol) in THF (2 mL) was added. The mixture was stirred for 3 hours, then treated
25
ACS Paragon Plus Environment

Page 27 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with NH Cl (aq., sat.), diluted with water and extracted with DCM (x3). The combined organic
4
layers were washed with water and brine and dried over MgSO . Filtration and concentration to
4
dryness followed by purification by silica gel chromatography (100% DCM) gave the product as
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
1
a colourless solid (0.277g, 92%). LCꢀMS [M+H] 335/337 (Cl), rt 1.30 min. H NMR (400 MHz,
Meꢀd ꢀOD): 7.75 (d, 1H), 7.71 (d, 1H), 7.59 (dd, 1H), 7.50 (d, 1H), 6.58 (d, 1H), 4.50 (s, 2H),
3
3.81 (s, 3H), 3.01 (s, 3H), 2.85 (s, 3H).
(E)-Methyl 3-(1-methyl-1H-benzo[d][1,2,3]triazol-5-yl)acrylate (19)
A stirred suspension of potassium tertꢀbutoxide (108 mg, 0.96 mmol) in THF (4 mL) at 0°C was
treated with methyl 2ꢀ(dimethoxyphosphoryl)acetate (0.18 mL, 1.12 mmol) in THF (4 mL). The
reaction mixture was stirred at 0°C for 1.5 h. Then 1ꢀmethylꢀ1Hꢀbenzo[d][1,2,3] triazoleꢀ5ꢀ
carbaldehyde (100 mg, 0.62 mmol) was added in one portion. The suspension was stirred at 0ºC
for 1.5 h. The reaction was quenched via addition of saturated ammonium chloride solution and
diluted with water. A precipitate formed which was filtered and dried to give the title compound
+
1
(
105 mg, 78%). LCꢀMS [M+H] 218.0, rt 0.68 min. H NMR (400 MHz, chloroformꢀd) δ 8.18
(s, 1H), 7.86 (d, J=16.1 Hz, 1H), 7.72 (dd, J=8.7, 1.4 Hz, 1H), 7.53 (d, J=8.8 Hz, 1H), 6.52 (d,
J=15.8 Hz, 1H), 4.32 (s, 3H), 3.84 (s, 3H).
Methyl 3-(4-chloro-3-(hydroxymethyl)phenyl)-3-(1-methyl-1H-benzo[d][1,2,3]triazol-5-
yl)propanoate (20)
A stirred suspension of (2ꢀchloroꢀ5ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)phenyl)
methanol (500 mg, 1.86 mmol), (E)ꢀmethyl 3ꢀ(1ꢀmethylꢀ1Hꢀbenzo[d][1,2,3]triazolꢀ5ꢀyl)acrylate
26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 67
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
(
640 mg, 2.95 mmol) and chloro(1,5ꢀcyclooctadiene)rhodium(I) dimer (92 mg, 0.19 mmol) in
water (8 mL) and 1,4ꢀdioxane (1.5 mL) at room temperature was treated with triethylamine (0.5
mL, 3.61 mmol). The reaction mixture was stirred at 95°C for 2 h and then at room temperature
for 17 h. The reaction was diluted with water and ethyl acetate. The organic phase was washed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
with water and brine, dried over MgSO , then filtered. The filtrate was adsorbed onto isolute and
4
+
purified by silica gel chromatography to give the title compound (98 mg, 15%). LCꢀMS [M+H]
3
60, rt 0.81 min. 1H NMR (400 MHz, methanolꢀd ) δ 7.89 (s, 1H), 7.66 (d, J=8.5 Hz, 1H), 7.44 ꢀ
4
7
.51 (m, 2H), 7.26 ꢀ 7.31 (m, 1H), 7.19 ꢀ 7.24 (m, 1H), 4.74 (t, J=8.0 Hz, 1H), 4.64 (s, 2H), 4.29
(s, 3H), 3.56 (s, 3H), 3.21 (dd, J=7.9, 3.9 Hz, 2H).
(S)-Methyl 3-(4-chloro-3-((N-methylphenylsulfonamido)methyl)phenyl)-3-(1-methyl-1H-
benzo[d][1,2,3]triazol-5-yl)propanoate (21)
stirred solution of methyl 3ꢀ(4ꢀchloroꢀ3ꢀ(hydroxymethyl)phenyl)ꢀ3ꢀ(1ꢀmethylꢀ1Hꢀ
benzo[d][1,2,3]triazolꢀ5ꢀyl)propanoate (0.485 g, 1.348 mmol), Nꢀmethylbenzenesulfonamide
0.243 g, 1.419 mmol), and (E)ꢀdiazeneꢀ1,2ꢀdiylbis(piperidinꢀ1ꢀylmethanone) (0.715 g, 2.83
mmol) in tetrahydrofuran (THF) (18 mL) at 0 °C was treated with tributylphosphine (0.75 mL,
.04 mmol). Upon addition the colour turned from deep orange to a lighter yellow. The iceꢀbath
A
(
3
was removed and the solution allowed stirred at ambient temperature for 15 h during which time
a precipitate formed. The supernatant was removed and the solution diluted with acetone and
ethyl acetate and washed with water (2x), brine (1x), dried over MgSO , filtered and
4
concentrated to give a yellow oil. This was adsorbed onto isolute and purified by silica gel
chromatography to give a white solid, which was subjected to chiral SFC HPLC resolution to
give (R)ꢀmethyl 3ꢀ(4ꢀchloroꢀ3ꢀ((Nꢀmethylphenylsulfonamido)methyl)phenyl)ꢀ3ꢀ(1ꢀmethylꢀ1Hꢀ
27
ACS Paragon Plus Environment

Page 29 of 67
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
,2,3ꢀbenzotriazolꢀ5ꢀyl)propanoate (0.106 g, 15%) and (S)ꢀmethyl 3ꢀ(4ꢀchloroꢀ3ꢀ((Nꢀ
methylphenylsulfonamido)
methyl)phenyl)ꢀ3ꢀ(1ꢀmethylꢀ1Hꢀbenzo[d][1,2,3]triazolꢀ5ꢀ
+
1
yl)propanoate (0.113 g, 16%). LCꢀMS [M+H] 513/515 (Cl), rt 1.08 min. H NMR (400 MHz,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
methanolꢀd ) δ 7.83 ꢀ 7.90 (m, 3H), 7.59 ꢀ 7.72 (m, 4H), 7.45 (d, J=8.8 Hz, 1H), 7.41 (s, 1H),
4
7.29 ꢀ 7.34 (m, 1H), 7.21 ꢀ 7.27 (m, 1H), 4.72 (t, J=7.9 Hz, 1H), 4.27 ꢀ 4.34 (m, 5H), 3.58 (s,
3H), 3.14 ꢀ 3.22 (m, 2H), 2.62 (s, 3H).
2-Methoxy-6-nitroaniline (23)
A stirred solution of 2ꢀaminoꢀ3ꢀnitrophenol (35 g, 227 mmol) in DMF (400 mL) was treated with
K CO (37.7 g, 273 mmol) and iodomethane (17.04 mL, 273 mmol). After 16 h the mixture was
2
3
poured into water. The resulting precipitate was collected by filtration and the solid was washed
+
with water and dried to give the title compound (35 g, 89%). LCꢀMS [M+H] 169, rt 1.71 min.
1
H NMR (400 MHz, DMSOꢀd ) δ 7.59 (m, 1H), 7.08 (m, 3H), 6.62 (m, 1H), 3.88 (s, 3H).
6
4-Bromo-2-methoxy-6-nitroaniline (24)
A solution of 2ꢀmethoxyꢀ6ꢀnitroaniline (35 g, 208 mmol) in acetic acid (500 mL) was treated
with sodium acetate (27.3 g, 333 mmol) and bromine (11.80 mL, 229 mmol). Then the reaction
mixture was stirred at room temperature for 20 minutes. The resulting precipitate was filtered
and washed with water and dried inꢀvacuum pump to give 50 g (95%) of the title compound.
+
1
LCꢀMS [M+H] 249, rt 1.78 min. H NMR (400 MHz, DMSOꢀd ) δ 7.75 (s, 1 H), 7.22 (m, 3 H),
6
3.92 (s, 3 H).
28
ACS Paragon Plus Environment