Structure-Guided Design of Highly Selective and Potent Covalent Inhibitors of ERK1/2

Article abstract of DOI:10.1021/acs.jmedchem.5b00466

The RAS/RAF/MEK/ERK signaling pathway has been targeted with a number of small molecule inhibitors in oncology clinical development across multiple disease indications. Importantly, cell lines with acquired resistance to B-RAF and MEK inhibitors have been shown to maintain sensitivity to ERK1/2 inhibition by small molecule inhibitors. There are a number of selective, noncovalent ERK1/2 inhibitors reported along with the promiscuous hypothemycin (and related analogues) that act via a covalent mechanism of action. This article reports the identification of multiple series of highly selective covalent ERK1/2 inhibitors informed by structure-based drug design (SBDD). As a starting point for these covalent inhibitors, reported ERK1/2 inhibitors and a chemical series identified via high-throughput screening were exploited. These approaches resulted in the identification of selective covalent tool compounds for potential in vitro and in vivo studies to assess the risks and or benefits of targeting this pathway through such a mechanism of action.

Full text of DOI:10.1021/acs.jmedchem.5b00466

Subscriber access provided by UNIVERSITY OF LEEDS
Article
Structure-guided Design of Highly Selective
and Potent Covalent Inhibitors of ERK1/2
Richard A Ward, Nicola Colclough, Mairi Challinor, Judit Debreczeni, Key Eckersley,
Gary Fairley, Lyman Feron, Vikki Flemington, Mark Andrew Graham, Ryan Greenwood,
Philip Hopcroft, Tina D Howard, Michael James, Clifford D Jones, Christopher R Jones,
Jonathan Renshaw, Karen Roberts, Lindsay Snow, Michael Tonge, and Kay Yeung
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00466 • Publication Date (Web): 15 May 2015
Downloaded from http://pubs.acs.org on May 18, 2015
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 24
Journal of Medicinal Chemistry
Structure-guided Design of Highly Selective and Potent Covalent Inhibitors of ERK1/2
1
2
3
4
5
6
7
8
Richard A. Warda, Nicola Colclough, Mairi Challinor, Judit E. Debreczeni, Kay Eckersley, Gary Fairley, Lyman Feron, Vikki
Flemington, Mark A. Graham, Ryan Greenwood, Philip Hopcroft, Tina D. Howard, Michael James, Clifford D. Jones,
Christopher R. Jones, Jonathan Renshaw, Karen Roberts, Lindsay Snow, Michael Tonge and Kay Yeung.
9
10
11
Oncology & Discovery Sciences iMeds, AstraZeneca, Alderley Park, Macclesfield, SK10 4TG, UK.
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
a To whom correspondence should be addressed. Phone: +44 (0)1625 519045. Eꢀmail: richard.a.ward@astrazeneca.com.
Address: AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire, SK10 4TG, United Kingdom.
Abstract
The RAS/RAF/MEK/ERK signaling pathway has been targeted with a number of small molecule inhibitors in oncology clinical
development across multiple disease indications. Importantly, cell lines with acquired resistance to BꢀRAF and MEK inhibitors
have been shown to maintain sensitivity to ERK1/2 inhibition by small molecule inhibitors. There are a number of selective, nonꢀ
covalent, ERK1/2 inhibitors reported along with the promiscuous hypothemycin (and related analogues) which act via a covalent
mechanism of action. This manuscript reports the identification of multiple series of highly selective covalent ERK1/2 inhibitors
informed by structureꢀbased drug design (SBDD). As a starting point for these covalent inhibitors, both reported ERK1/2
inhibitors and chemical series identified via highꢀthroughput screening, were exploited. These approaches resulted in the
identification of selective covalent tool compounds for potential in vitro and in vivo studies to assess the risks and or benefits of
targeting this pathway through such a mechanism of action.
Abbreviations
ABT (1ꢀaminobenzotriazole), GSH (glutathione), HTS (high throughput screening), PD (pharmacodynamic), PK
(pharmacokinetics), LLE (ligand lipophilicity efficiency), LE (ligand efficiency), LCꢀMS (liquid chromatography–mass
spectrometry), MOA (mechanism of action), SAR (structureꢀactivity relationships), THP (tetrahydropyran).
Introduction
The RAS/RAF/MEK/ERK pathway plays a central role in controlling mammalian cell cycle progression, differentiation, protein
synthesis, metabolism, survival, migration and senescence1. This pathway is regulated by a wide range of extracellular stimuli
which result in activation of the RAS family of GTPases (KRAS, NRAS and HRAS), subsequently turning on the downstream
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 24
RAF protein kinases (BꢀRAF, cꢀRAF, and aꢀRAF). The main substrates of the RAF kinases are the MAPK/ERK kinases, MEK1,
and MEK2, which then phosphorylate their only substrates, extracellular signal–regulated kinase ERK1 and ERK22. ERK1 and
ERK2 control the output of the RAS/RAF/MEK/ERK pathway by activating their many substrates, both nuclear and
1
2
3
4
5
6
7
cytoplasmic3 4, which include nuclear components, transcription factors, membrane proteins and protein kinases. Activation of
,
8
9
the RAS/RAF/MEK/ERK signaling pathway results in extensive changes in gene expression mediated by phosphorylation of
transcription factors which promote cell cycle progression and also regulate negative feedback mechanisms, for example by
controlling the expression of dualꢀspecificity phosphatases (DUSPs) can dephosphorylate and inactivate ERK1/25.
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
Constitutive activation of the RAS/RAF/MEK/ERK pathway through genetic alterations is associated with multiple human
tumours. RAS genes are mutated in colorectal (50%), melanoma (20%) and pancreatic (90%) tumours, whereas mutations in bꢀ
RAF have been identified in melanoma (60%), thyroid (~50%) and colorectal (10%) tumours6. Due to the association of
activation of the RAS/RAF/MEK/ERK pathway and the ongogenic phenotype, several small molecule inhibitors targeting bꢀRAF
and MEK have been developed and entered clinical trials in multiple disease indications7,8. Data from clinical trials with the RAF
inhibitors such as vemurafenib and dabrafenib demonstrate dramatic regression of melanomas carrying the BRAF V600E
mutation9, whereas BRAF wildꢀtype tumours are refractory. However, acquired resistance to RAF inhibitors occurs in a
significant subset of patients, primarily from genetic alteration resulting in RAS/RAF/MEK/ERK pathway reactivation10ꢀ13
.
Additionally, several selective and potent nonꢀATPꢀcompetitive allosteric MEK1/2 inhibitors have been developed and assessed
in multiple clinical trials. Only limited efficacy as single agent therapies has been reported with MEK inhibitors, therefore
combinations of MEK1/2 inhibitors and cytotoxic chemotherapy, and/or other targeted agents are being studied to expand the
efficacy of this class of agents.
Direct targeting of ERK1/2, the key signaling node of the RAS/RAF/MEK/ERK pathway, provides another therapeutic option in
tumours with mutations in RAS or bꢀRAF genes. ERK1/2 inhibition may also have clinical utility in overcoming acquired
resistance to RAF and MEK inhibitors where RAS/RAF/MEK/ERK pathway reactivation has occurred. ERK1 and ERK2 are
85% identical by sequence (significantly higher in the ATP binding site) and share some, if not all, of their functions with reported
inhibitors apparently active against both. A range of reversible ERK1/2 small molecule inhibitors have been disclosed (figure 1),
including multiple examples in clinical development. These include MK8353/SCH900353 (likely related to SCH77298414),
BVDꢀ52315 and GDCꢀ099416, in addition to a selection of reported preꢀclinical inhibitors17 20. Hypothemycin and analogues such
ꢀ
as FR14808321 have also been reported as irreversible covalent inhibitors of ERK1/2, targeting the residue Cysꢀ166, which is
located in the ATP pocket. Sequence analysis across the kinome suggests that 48 kinases share a cysteine in this position22. These
reported inhibitors appear to be active across all of these kinases and therefore have limited utility in terms of assessing the
benefits and risks associated with selectively targeting ERK1/2 via a covalent mechanism of action.
ACS Paragon Plus Environment

Page 3 of 24
Journal of Medicinal Chemistry
1
2
N
N
3
4
5
6
7
8
F
N
Cl
OH
Cl
N
NH
OH
O
HO
HO
N
O
O
N
9
H
N
O
O
10
11
12
13
14
15
16
17
HN
O
Cl
N
O
N
O
OH
N
N
N
N
H
HN
N
N
N
H
Figure 1 – Selection of published ERK1/2 inhibitors (Left to right, SCH772984, GDCꢀ0994, BVDꢀ523 and FR148083).
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
Covalent (irreversible and or reversible) inhibitors have a number of potential advantages over nonꢀcovalent small molecules
which have been well documented and can include improved selectivity and increased activity (or prolonged duration of activity)
against the primary target23. However, covalent inhibitors do have potential risks such as idiosyncratic toxicity, if they are
susceptible to binding covalently to a range of targets. In addition, if an inhibitor acting with a short duration of action is
preferred then an irreversible covalent inhibitor might not be suitable as the pharmacodynamic (PD) effect often continues after
the compound has been metabolised and cleared from the body. With specific reference to ERK, a covalent inhibitor may have
the advantage of maximum pathway knockdown along with a differential effect on feedback mechanisms to nonꢀcovalent
inhibitors; however there is the potential for toxicity through total inhibition of this key signalling pathway which is also utilised
by nonꢀtumorigenic cells. Therefore the identification of potent and selective covalent inhibitors of ERK1/2, with suitable
properties for in vitro and in vivo experiments, will enable the risks and benefits of such a profile to be further evaluated.
Results and discussion
Series Identification. Two hit identification approaches were taken to design and develop covalent inhibitors of ERK1/2. The
first approach consisted of the direct modification of published ERK1/2 inhibitors, through addition of a covalent ‘warꢀhead’ in a
suitable position on the molecule informed by knowledge of the likely binding mode (along with location of Cysꢀ166) through
molecular modelling. As an alternative approach, we identified a set of nonꢀcovalent ERK1/2 inhibitor template from high
throughput screening (HTS) to similarly convert to covalent inhibitors through molecular modelling and structurally informed
addition of a covalent functionality.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 24
Assessment of the binding mode of published ERK1/2 inhibitors via molecular modelling, suggested tetrahydroꢀpyridoꢀ
pyrimidine18, indazole24 and tetrahydroꢀpyrazoloꢀpyridine25 scaffolds might be suitable starting points for this approach.
Inhibitors containing these scaffolds were computationally docked into ERK2 crystal structures using Glide to model potential
binding modes (PDB Code 2OJI20). The Cysꢀ166 residue appears to be in a relatively challenging region of the ATP pocket to
target. It is potentially incompatible in terms of the vectors associated with some hinge scaffolds and established (irreversible)
covalent warheads. The impact of this is that there is a requirement for a specific geometry of suitable linker groups for the
covalent warhead to be correctly positioned. A range of target compounds for possible synthesis were evaluated using covalent
docking, where a variety of reactive groups were positioned on different regions of the described scaffolds. Molecular modelling
studies indicated that a sulphonamide group might facilitate a suitable geometry to target Cysꢀ166 from the highlighted scaffolds
which were predicted to bind to the hinge region. Three target compounds were synthesised (1, 2 and 3), for examples 1 and 2
additional compounds were made (4 and 5) without the double bond of the reactive group reduced as reversible control
compounds (table 1).
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
O
O
S
S
NH
NH
O
S
N
O
O
S
N
O
O
S
O
O
O
N
N
N
N
N
N
O
O
N
N
N
N
N
N
N
N
H
N
N
H
H
H
H
1
2
3
4
5
BVD-523
SCH772984
FR148083
1 (4)
2 (5)
3
ERK2
Biochemical, KM
ATP (µM)
<0.0003
<0.0003
5.1
0.37 (9.0)
0.22 (>10)
0.0075 (0.28)
0.017 (1.8)
0.074
0.11
0.38/1
1.4
ERK2
Biochemical,
High ATP (µM)
<0.0003
4.1/0.14
0.18
<0.0003
0.004/0.02
0.07
3.7
1.6/5.7
4.1
Cell MOA A375
pERK/pRSK
(µM)
0.24/1.1
(10/>30)
0.12/0.47
(10/>30)
Cell Proliferation
A375
ND (ND)
>10
1.3 (>30)
>10
(µM)
MEK1
>10
>10
3.4
4.9
Biochemical, KM
ATP (µM)
a
Log D7.4
>4
3.2
1.6
47
1
2.2
7
1.6
Aqueous
8.5
<2.1
760
300
Solubility (µM)b
Kinase Selectivity
ND
ND
ND
2 (CꢀRAF,
8 (AURB,
5 (CꢀRAF,
ACS Paragon Plus Environment

Page 5 of 24
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(N>75% out of
125 Kinase panel)
MEK1)
AURC, cꢀ
RAF, FLT1,
FLT3, KDR,
MKK7b,
FLT1, KDR,
MEK1,
PRAK)
TRKA)
Activity data into parenthesis indicates data on nonꢀcovalent matched pairs 4 and 5. All IC50 data are reported as micromolar and are the mean of
at least n=2 independent measurements. Kinase selectivity data measured at 1 µM compound concentration. Kinases in bold have a cysteine in
the same position as Cysꢀ166 in ERK2. ND is not determined. aDistribution coefficient between 1ꢀoctanol and aqueous phosphate buffer at pH
7.4. bSolubility of compounds in aqueous phosphate buffer at pH 7.4 after 24 hours at 25 °C (ꢁM).
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
Table 1 – Covalent inhibitors exploiting published ERK1/2 inhibitor templates.
All three of the designed compounds (1, 2 and 3) showed promising ERK2 biochemical activity at KM ATP (60 µM) and broadly
equivalent activity using higher concentrations of ATP (1 mM). For these compounds we also observed promising cellular
activity in a mechanism of action (MOA) assay measured by the levels of phosphorylated ERK2 (pERK) and of the
phosphorylation of the downstream kinase RSK (pRSK) in an A375 melanoma cell line, containing a bꢀRAFV600E mutation. In
contrast, in our assays FR148083 showed only weak biochemical activity and weak cellular activity. Additionally, the enzymeꢀ
cell dropꢀoff for disclosed inhibitors BVDꢀ523 and SCH772984 is significantly greater (likely driven by competition with cellular
ATP) compared to the covalent inhibitors, as expected with a covalent mechanism of action. The impact on the level of pErk
suggests that these compounds are also preventing the activation of ERK2 kinase (in addition to the inhibition of catalysis); this
shows a profile broadly similar to SCH772984 as opposed to BVDꢀ523 which has a profile consistent with the inhibition of
catalysis only. Antiꢀproliferative effects in this A375 melanoma cell line were also detected for these novel compounds, albeit
weakly which may be due to high intrinsic compound reactivity, discussed later. In both cases where we synthesised comparative
compounds without the reactive functionality (examples 4 and 5), the compounds were significantly less active across these assays
highlighting the importance of the covalent warꢀhead to activity. The covalent binding mechanism of 1 and 2 was confirmed by
Mass Spec (Supplementary Figures S1 and S2), there was not enough sample available to confirm this mechanism for 3. The
binding mode of 1 was resolved using coꢀcrystallisation (Figure 2). As predicted by the modelling work, the inhibitor makes key
hydrogen bonding interactions at the hinge region with residues Leuꢀ107 and Metꢀ108, both with hydrogen bond distances of
1.9Å. The sulphonamide group also makes an interaction with a bound water molecule, which forms an intricate network of
hydrogen bonds to Glnꢀ105, the backbone of Aspꢀ167 and Lysꢀ54. The covalent bond formed with Cysꢀ166 was clearly observed
in the electron density. Interesting, when compound 1 was soaked into ERK2 crystals, the inhibitor was unable to form a covalent
bond with Cysꢀ166 (data not shown). This may suggest that a degree of protein reorganisation is required for the covalent
reaction to occur which was not possible when the compound was soaked into the crystal system. The docking calculation largely
reproduced the crystallographic binding mode; the discussed bound water molecule however was not included in the simulation as
all crystallographic water molecules were removed. However, the predicted direct interaction of sulphonamide group with Lysꢀ54
appears to stabilise this interaction in the docking studies resulting in the crystallographic binding mode maintaining a favourable
GlideScore.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 24
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
Figure 2 ꢀ Binding mode of compound 1 elucidated in human ERK2 crystal structure via coꢀcrystallisation (PDB Code 4ZZM).
Top figure highlights the interaction of the pyrimidine group with the hinge region. The middle figure shows the hydrogen bond
network around the sulphonamide group and bound crystallographic water molecule. The lower figure shows the crystal structure
in blue, compared to the docked pose in brown using the covalent docking protocol in Glide (Using PDB Code 2OJI).
The lipophilicity of these molecules (measured by LogD at pH 7.4) is relatively low which, along with their relatively small
molecular size, suggests scope for further optimisation and the physicochemical properties such as solubility were also
encouraging. Broader kinase selectivity was assessed through testing in a Eurofins (Millipore)26 subꢀpanel of 125 kinases.
Compound 1 in particular showed very good kinome selectivity which is perhaps surprising for compounds of such low
ACS Paragon Plus Environment

Page 7 of 24
Journal of Medicinal Chemistry
complexity, which make a relatively small number of interactions in the binding site. It was reassuring to confirm that these
compounds did not appear to be active against all kinases with a cysteine in this region, as was reported with analogues of
hypothemycin. It is important to note that the Eurofins (Millipore) assays for RAF and MEK1 are coupled assays, thereby
compounds active in ERK1/2 may also appear active against RAF and MEK1. The MEK1 assay contains unactive ERK2, with
the cꢀRAF assay containing unactive MEK1 and ERK2. Compounds 1 and 2 were subsequently confirmed as inactive (IC50 >10
µM) against an inꢀhouse MEK1 biochemical assay, with 3 showing weak MEK1 activity of 4.9 µM. Identifying which additional
kinases these compounds may act upon via a covalent mechanism of action was also of interest (i.e. kinases with a cysteine in the
same position as Cysꢀ166). Importantly, compound 1 did not appear to have significant kinase inhibition against this subset of
kinases. For compounds 2 and 3 activity was observed only against a limited number of kinases that included KDR, FLT1 and
FLT4. Although these compounds gave us confidence that Cysꢀ166 could be selectively targeted with small molecules, these
inhibitors proved difficult to further optimise as the sulphonamide warꢀhead has high intrinsic reactivity to thiols which we were
not able to substantially temper through modification whilst also maintaining target activity. In an assay measuring compound
reactivity versus glutathione as a model for the active site cysteine residue, compound 1 demonstrated a t1/2 of 15.4 minutes,
compared to our recent disclosure of our EGFR mutantꢀselective inhibitor AZD9291, which had t1/2 of 121 minutes27 (Note in pH
7.4 buffer alone compound 1 is stable to hydrolysis showing no reaction over 11 hrs at 37°C). This difference in reactivity is
driven by the alkenyl sulphonamide versus acrylamide warhead change. However, based on modelling, the alkenyl sulphonamide
warhead was judged to have the most suitable vector to target Cysꢀ166. High relative reactivity could lead to issues in the
optimisation of pharmacokinetic (PK) properties and may also be more likely to result in nonꢀspecific covalent binding to a wider
range of targets and instability when tested across in vitro assays.
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
From in-house screening using an ERK2 biochemical assay, a reversible inhibitor (6), first synthesised within our internal
FAK drug discovery program, was identified28. This compound showed moderate potency (0.044 µM and 0.34 µM in the Km and
high ATP ERK2 biochemical assays) and we believed it was a suitable lead compound to enable further optimisation through the
addition of reversible and irreversible interactions. ERK1 activity was also confirmed for this compound (0.095 µM) using the
Eurofins (Millipore) kinase panel. In a bid to improve ERK1/2 potency and broader kinase selectivity, the tetrahydropyran (THP)
group, which was predicted to bind in the solvent channel of disclosed ERK inhibitors18, was hybridised onto our identified hinge
binding group. We also decided to assess the impact of replacement of the triꢀfluoro methyl group with chlorine. The resulting
compounds 7 and 8 showed improvements in general kinase selectivity (4 other kinases with >75% inhibition compared to 10)
whilst potency remained largely unchanged (table 2).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 24
1
2
NH
NH
NH
3
4
O
O
O
5
6
7
8
HN
F
F
HN
F
HN
F
F
F
Cl
F
O
O
N
F
N
N
N
9
N
N
N
N
H
H
H
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
6
7
8
6
7
8
ERK2 Biochemical, KM
ATP (µM)
0.044
0.34
0.26
ND
0.04
0.38
2.1/2.4
3.4
ERK2 Biochemical, High
ATP (µM)
Cell MOA A375
8.7/5.1
1.4
7.5/10
ND
pERK/pRSK (µM)
Cell Proliferation A375
(µM)
MEK1 Biochemical,
KM ATP (µM)
>10
>10
>10
a
Log D7.4
3.8
15
2.8
290
ND
3.4
8.8
Aqueous Solubility (µM)b
Kinase Selectivity (N>75%
out of 125 Kinase panel)
10 (AURB, AURC,
4 (CDK7, cꢀRAF,
MET, FAK)
AXL, CDK1, CDK7,
cꢀRAF, IGFꢀ1R, IR,
MET, FAK, TRKA)
Kinase selectivity data measured at 1 µM compound concentration. All IC50 data are reported as micromolar and are the mean of at least n=2
independent measurements. ND is not determined. aDistribution coefficient between 1ꢀoctanol and aqueous phosphate buffer at pH 7.4.
bSolubility of compounds in aqueous phosphate buffer at pH 7.4 after 24 hours at 25 °C (ꢁM).
Table 2 – Reversible ERK1/2 inhibitors identified from inꢀhouse screening.
ACS Paragon Plus Environment

Page 9 of 24
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
Figure 3 ꢀ Binding mode of 7 (blue) and 10 (brown) elucidated in human ERK2 crystal structure. Top figures show compound 7
(PDB Code 4ZZN), focusing on the hinge region interactions (left) and bound water molecules interactions (right). Lower figures
show compound 10 (PDB Code 4ZZO), focusing on the hinge region interactions (left) and bound water molecules interactions
(right).
To guide further development of this series, a crystal structure using compound soaking was solved for compound 7 (Figure 3).
The crystal structure highlighted the important interaction of the pyridine group with the Metꢀ108 located on the hinge region (two
hydrogen bonds with lengths of 1.9Å). Also of particular interest were two identified crystallographic water molecules, one of
which interacts directly with the amide of compound 7, forming a network of hydrogen bonds with Gluꢀ71, Glnꢀ105 and Aspꢀ167.
The water molecule closest to the inhibitor was displaced by an oxygen of the sulphonamide of compound 1. The positioning of
this anilino group confirmed our hypothesis that substituting the aniline on the 2ꢀposition would be suitable to target Cysꢀ166. A
single compound was initially synthesised (9) which maintained the pyridine hinge binding group. Where the acrylamide
functionality was located at the 2ꢀposition of the aniline, the amide was reversed to enable this chemical modification. This
change appeared to be tolerated from the molecular modelling work as the key interactions appeared to be broadly maintained.
The activity data on this first example was highly encouraging as good biochemical (0.042 µM Km and 0.06 µM High ATP) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 24
cellular activity (0.12 µM pRSK and 0.4 µM pERK) were observed. Indeed, the potency increased when compared to the earlier
covalent examples (Table 3). These compounds appeared to show effects on both pERK and pRSK, consistent with the activity
profile of our earlier inhibitors and SCH772984. Subsequently, the matched pair with a pyrimidine hinge binding group was
synthesised (10) to assess impact of this change on activity, the pyrimidine hinge region binding group having been observed in
other ERK inhibitors such as GDCꢀ0994. This demonstrated approximately 10ꢀfold greater potency across our in vitro compound
activity assays. It was also reassuring that these compounds showed reduced intrinsic reactivity towards glutathione compared to
the earlier alkenyl sulphonamide series, with t1/2 152 minutes for 10, suggesting a greater contribution of reversible interactions to
overall compound potency for these compounds. The reversible matched pair of 10 (compound 11) was also synthesised as a tool
compound to assess the specific impact of the covalent bond, again showing significantly weaker activity across the assays. Mass
Spectrometry was used to confirm covalent binding of 10 to ERK2, which as expected was not observed with 11 (Supplementary
Figures S3 and S4). The binding mode of the covalent inhibitor 10 (brown) was solved using coꢀcrystallisation and was
remarkably similar to that of the reversible inhibitor 7 (Figure 3). It is interesting to note that even though the amide group has
been reversed in compound 10 (compared to that of compound 7), this apparently key interaction with the bound water molecule
is largely maintained. This series of covalent inhibitors appears to be able to exploit this complex of water molecules to produce
exquisite kinome selectivity; however compound 1 demonstrates that this selectivity can largely be maintained if these specific
interactions can be replaced by the inhibitor. Both compounds 7 and 10 appear to have the binding mode stabilised by an internal
hydrogen bond between the anilino NH and amide carbonyl, with a distance of 2Å in both cases. Compound 10 has consistent
hydrogen bonding interactions with the hinge region residue Metꢀ108 as 7 (distances of 1.9 Å and 2Å). One notable difference
was the rotation of the Cysꢀ166 sideꢀchain which enables the formation of the covalent bond, highlighting some flexibility in the
conformation of this residue.
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
O
O
O
O
HN
HN
HN
HN
HN
HN
HN
HN
N
Cl
Cl
Cl
Cl
O
O
O
N
N
O
N
N
N
N
N
H
N
N
N
N
H
H
H
9
10
11
12
9
10(11)
12
ERK2 Biochemical, KM
ATP (µM)
0.042
0.060
0.0085 (0.48)
0.0054 (4.0)
0.0031
0.003
ERK2 Biochemical, High
ATP (µM)
Cell MOA A375
0.12/0.4
0.019/0.069 (>30/>30)
0.0067/0.028
ACS Paragon Plus Environment

Page 11 of 24
Journal of Medicinal Chemistry
1
2
pERK/pRSK (µM)
3
4
5
6
7
8
9
Cell Proliferation A375
(µM)
0.68
>10
2.5
0.057 (6.6)
>10
0.041
>10
2.3
MEK1 Biochemical,
KM ATP (µM)
a
Log D7.4
2.7
10
11
12
13
14
15
16
17
18
Aqueous Solubility (µM)b
40
46
150
Inꢀvitro Clearance HLM,
(Rat Heps)c
24 (270)
67 (200)
13 (>250)
Kinase Selectivity (N>75%
out of 125 Kinase panel)
2 (CꢀRAF, MEK1)
4 (AURB, AURC, cꢀ
2 (cꢀRAF, MEK1)
RAF, FLT1)
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
Activity data into parenthesis indicates data on nonꢀcovalent matched pair 11. All IC50 data are reported as micromolar and are the mean of at
least n=2 independent measurements. Kinase selectivity data measured at 1 µM compound concentration. Kinases in bold have a cysteine in the
same position as Cysꢀ166 in ERK2. aDistribution coefficient between 1ꢀoctanol and aqueous phosphate buffer at pH 7.4. bSolubility of
compounds in aqueous phosphate buffer at pH 7.4 after 24 hours at 25 °C (ꢁM). cHuman microsome and rat hepatocytes metabolism intrinsic
clearance (ꢁL min−1 mg−1).
Table 3 – Covalent ERK1/2 inhibitors developed from scaffold identified from inꢀhouse screening.
An additional compound (12) was synthesised in an effort to improve ERK1/2 potency through increasing the reactivity towards
Cysꢀ166, our previous studies indicated such a change is likely to increase the reactivity of the acrylamide towards glutathione29.
This was the most potent compound synthesised within this early work and showed exquisite panꢀkinase selectivity. As expected,
the addition of the nitrogen increased the inherent reactivity of the acrylamide, with a t1/2 42 minutes in the glutathione assay
(Both compounds 10 and 12 are stable to hydrolysis in pH 7.4 buffer alone at 37°C). Note as discussed previously, the
compounds in this table were tested against our MEK1 biochemical assay, where they were inactive (IC50>10µM) so it is likely
compounds such as 12 are not genuinely active against any kinases in the Eurofins (Millipore) panel. Compound 12 was
subsequently used as a highly selectivity and potent inꢀvitro tool for cellular washꢀout experiments where the compound inhibited
direct ERK substrate phosphorylation (p90RSK pThr359/Ser363) for up to 48 hours after the removal of compound (Figure 4a).
This is broadly consistent with recent literature data reporting a half life of 54 hours for ERK2 protein in mouse fibroblasts, ERK1
protein with a halfꢀlife of 120 hours30. In addition, the cellular potency of compound 12 was measured in the pRSK cellular assay
as a timeꢀcourse, showing maximal potency of the compound between 20ꢀ30 hours (Figure 4b), which also appears consistent
with these data.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 24
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
4a
4b
Figure 4 – Cellular washꢀout experiment with covalent ERK1/2 inhibitor 12 (4a) and pRSK cellular assay timeꢀcourse study (4b).
Further development of the series was focused towards understanding the wider SAR around the template, with a particular target
of identifying compounds which might be suitable for inꢀvivo testing. Our initial focus was to modify the 5ꢀposition of the
pyrimidine along with the solvent channel group to vary the specific structural motifs and broader lipophilicity, with a view to
impacting on compound plasma clearance. Due to the increased glutathione reactivity of compound 12, we chose to revert back to
the scaffold of 10, where the acrylamide was attached to the phenyl group, to maximise our chances of improving
pharmacokinetic properties.
Modification to the 5ꢀposition of the pyrimidine was tolerated with a number of groups. The CF3 analogue was a particularly
potent example (13), with methyl also showing activity (14, Table 4). We also observed that the THP group could be replaced by
a pyrazole (15), or certain basicꢀcontaining motifs such as compound 16. These changes were also combined in example 17. The
emerging data suggested that lowering LogD7.4 of compounds in this series broadly resulted in decreased inꢀvitro metabolic
clearance, correlations show an R2 of 0.69 when log human microsome clearance is plotted against LogD7.4 (Supplementary
Figure S4). We were however concerned with compounds where LogD7.4 has been significantly lowered as this might negatively
impact on passive compound permeability into cells. Therefore our aim was to target an area of liopophilicity which gave a
ACS Paragon Plus Environment

Page 13 of 24
Journal of Medicinal Chemistry
suitable compromise between these properties. It was also observed that whilst compounds with the THP and basic functionality
in the solvent channel tested as inactive in the MEK1 biochemical assay, examples with pyrazole did show increased activity
(0.12µM for 15 and 0.26 uM for 17). These data might suggest that it is possible to find a relatively selective, dual MEK1/ERK,
covalent inhibitor through the approach described here.
1
2
3
4
5
6
7
8
9
O
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
HN
HN
R
R1
1
N
R2
N
N
H
13
14
15
16
17
R1
R2
CF3
Me
Cl
Cl
Me
O
O
N
N
N
N
N
ERK2 Biochemical,
KM ATP (µM)
0.0069
0.0085
0.077
0.088
0.0099
0.013
0.085
0.090
0.039
0.047
ERK2 Biochemical,
High ATP (µM)
Cell MOA A375
pERK/pRSK (µM)
0.022/0.065
0.053
0.17/0.74
0.47
0.048/0.12
0.091
0.072/0.2
0.12
0.066/0.28
0.24
Cell Proliferation
A375
(µM)
MEK1 Biochemical,
KM ATP (µM)
>10
>10
0.12
>10
0.26
a
Log D7.4
2.9
59
2
2.4
0.9
2.1
Aqueous Solubility
(µM)b
170
520
770
990
Inꢀvitro Clearance
HLM, (Rat Heps)
38 (270)
13 (59)
42 (120)
<3 (25)
19 (33)
ND
Kinase Selectivity
(N>75% out of 125
Kinase panel)
6 (AURB,
AURC, CDK7, cꢀ
RAF, MEK1,
SRC)
2 (cꢀRAF,
7 (AURB, AURC, cꢀ 4 (cꢀRAF, DYRK2,
RAF, FLT1, KDR,
MEK1, FAK)
MEK1)
FLT1, MEK1)
Kinase selectivity data measured at 1 µM compound concentration. All IC50 data are reported as micromolar and are the mean of at least n=2
independent measurements. Kinases in bold have a cysteine in the same position as Cysꢀ166 in ERK2. aDistribution coefficient between 1ꢀ
octanol and aqueous phosphate buffer at pH 7.4. bSolubility of compounds in aqueous phosphate buffer at pH 7.4 after 24 hours at 25 °C (ꢁM).
Table 4 – Expansion of SAR around the pyrimidine template
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 24
1
2
3
4
5
6
7
8
9
From this set of compounds, considering compound potency, broader kinase selectivity and physicochemical properties we
identified compound 13 as a potential inꢀvivo tool compound for mouse studies. Although in vivo plasma clearance in mouse
appeared high (note in vitro clearance in mouse hepatocytes was high, 143 ꢀL/min/106 cells), when compound 13 was coꢀdosed
with the irreversible inhibitor of cytochrome P450’s, 1ꢀaminobenzotriazole (ABT), we were able to estimate that the inhibitor
could achieve 9 and 7ꢀfold free plasma cover above cellular pRSK IC50 at the 0.5 and 2 hr time points (Mouse oral AUC0ꢀt (ꢀM.h)
of 21) respectively which for an irreversible agent could be adequate to explore the in vivo pharmadynamic (PD) activity using
established models in the mouse. In addition, the timecourse experiment shown in figure 4b suggests that the cellular potency of
these inhibitors may increase over time, potentially giving increased levels or duration of cover than this calculation may suggest.
As mouse or human plasma protein binding could not be measured due to poor recovery of compound, the foldꢀfree cover was
estimated using the protein binding value of the saturated analogue of compound 11 (Supplementary Figure S5).
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
Conclusion
We have described an approach for the identification of potent and selective irreversible covalent ERK1/2 inhibitors using the
exploitation of published ERK inhibitors along with inꢀhouse biochemical screening. To our knowledge these are the first
disclosed selective covalent ERK1/2 inhibitors fully characterised by biochemical and cellular assays along with elucidation of
binding mode by xꢀray crystallography31. Although our covalent inhibitors identified from published scaffold start points were
active and showed good selectivity, a more potent and potentially developable series of compounds was identified through the
optimisation of compounds identified from inꢀhouse screening. These compounds have excellent biochemical and cellular ERK
potency and highlighted examples showed exquisite selectivity in broad kinase panels. The cellular activity of these compounds
is equivalent to published inhibitors BVDꢀ523 and SCH772984. However, they have reduced molecular weight and lower
lipophilicity resulting in improved ligand efficiency (LE) and ligand lipophilicity efficiency (LLE)32 along with an alternative
mechanism of inhibition. We have also highlighted a compound which was dosed in vivo in mouse (with ABT) to give likely
target coverage of at least 2 hours, thereby providing a potential in vivo tool compound for the pathway where the brief inhibition
through a covalent mechanism of action may be adequate, or preferred. In summary, we have disclosed a series of highly potent
and selective ERK1/2 inhibitors that will enable improved understanding of how to optimally inhibit the RAS/RAF/MEK/ERK
signaling pathway.
Supporting Information Available: Additional data analyses, experimental information and compound characterisation can be
found in the supporting information. This material is available free of charge via the Internet at http://pubs.acs.org.
ACS Paragon Plus Environment

Page 15 of 24
Journal of Medicinal Chemistry
Acknowledgements
1
2
3
4
5
We would like to thank all colleagues at AstraZeneca who were involved in the work reported in this manuscript.
6
7
8
9
Experimental
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
Chemistry
General Analytical Information
Nuclear Magnetic Resonance spectra were recorded on a Bruker 300 MHz, 400 MHz or 500 MHz instrument at room
temperature. All 1H NMR spectra were measured in part per million (ppm) relative to the signals for residual dimethyl sulfoxide
(DMSO) in deuterated DMSO (2.52 ppm), or the signals for tetramethylsilane (TMS) added into the deuterated chloroform (0
ppm). Data for 1H NMR were reported as: chemical shift, integration, multiplicity (s = singlet, d = doublet, t= triplet, q = quartet),
coupling constants measures in Hz. Mass spectra were obtained on a Waters LCMS Micromass instrument with a XSELECT
CSH™ C18, 5 ꢁM, 2.1 x 50 mm column. All compounds were found to be of 95% purity or greater by LCMS.
Full compound synthesis experimental details and characterisation are detailed in the supplementary material with a selection of
key examples shown below.
Representative procedure for the synthesis of vinyl sulfonamides 1, 2 and 3.
Scheme 1 - Synthesis of 4 – 3ꢀ(2ꢀmethylpyridinꢀ4ꢀyl)ꢀ5ꢀ(vinylsulfonyl)ꢀ4,5,6,7ꢀtetrahydroꢀ1 Hꢀpyrazolo[4,3ꢀc]pyridine 3.
H
H
N
N
N
N
N
O
N
Cl
S
S
O
O
O
Cl
N
N
3
3a
3b
2ꢀChloroethanesulfonyl chloride (0.029 mL, 0.27 mmol) 3a was added to 3ꢀ(2ꢀmethylpyridinꢀ4ꢀyl)ꢀ4,5,6,7ꢀtetrahydroꢀ1 Hꢀ
pyrazolo[4,3ꢀc]pyridine 3b (58.8 mg, 0.27 mmol) and NꢀethylꢀNꢀisopropylpropanꢀ2ꢀamine (0.146 mL, 0.82 mmol) in DMF (2
mL). The resulting solution was stirred at room temperature for 24 hours. The solution was absorbed onto silica and purified by
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 24
flash silica chromatography, elution gradient 0 to 30% actetone in EtOAc. Pure fractions were evaporated to dryness to afford 3ꢀ
(2ꢀmethylpyridinꢀ4ꢀyl)ꢀ5ꢀ(vinylsulfonyl)ꢀ4,5,6,7ꢀtetrahydroꢀ1 Hꢀpyrazolo[4,3ꢀc]pyridine 3 (33.0 mg, 39.5 %) as a colourless gum.
1H NMR (400 MHz, DMSO) δ 2.57 (s, 3H), 2.88 (t, J = 5.7 Hz, 2H), 3.54 (t, J = 5.8 Hz, 2H), 4.52 (s, 2H), 6.16 (d, J = 10.0 Hz,
1H), 6.24 (d, J = 16.5 Hz, 1H), 6.95 (dd, J = 10.0, 16.5 Hz, 1H), 7.40 (d, J = 5.1 Hz, 1H), 7.49 (s, 1H), 8.54 (d, J = 5.2 Hz, 1H),
13.08 (s, 1H). HRMS: ESI+ m/z calculated for C14H16N4O2S [M+H]+ 305.1072; m/z observed 305.1076.
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
Scheme 2 - Experimental procedures for the synthesis of Nꢀ(2ꢀ((5ꢀchloroꢀ2ꢀ((tetrahydroꢀ2Hꢀpyranꢀ4ꢀyl)amino)pyrimidinꢀ4ꢀ
yl)amino)phenyl)acrylamide 10.
Cl
Cl
N
DIPEA
NH2
10d
O
Cl
Cl
n-BuOH
N
N
N
H
HN
O
N
H2N
110 °C
68%
DIPEA
Cl
N
N
H
NH2
NH2
10b
Cl
N
DMA
150 °C
84%
NH2
10e
10c
10a
O
Cl
DIPEA
THF
19%
10f
Cl
N
HN
O
N
N
H
HN
O
10
NꢀEthylꢀNꢀisopropylpropanꢀ2ꢀamine (1.90 mL, 10.9 mmol) was added to 2,4,5ꢀtrichloropyrimidine 10a (1.00 g, 5.45 mmol) and
benzeneꢀ1,2ꢀdiamine 10b (0.590 g, 5.45 mmol) in nꢀbutanol (25 mL) under nitrogen. The resulting solution was stirred at 110 °C
for 3 hours. The reaction mixture was evaporated and the residue was stirred with aqueous HCl (0.1M, 10 mL) for 30 minutes.
The solid was collected by filtration and dried under vacuum. The crude solid was triturated with DCM to give a solid which was
collected by filtration and dried under vacuum to afford N1ꢀ(2,5ꢀdichloropyrimidinꢀ4ꢀyl)benzeneꢀ1,2ꢀdiamine 10c (0.942 g, 68%)
as a white solid. Tetrahydroꢀ2 Hꢀpyranꢀ4ꢀamine 10d (0.791 mL, 7.64 mmol), N1ꢀ(2,5ꢀdichloropyrimidinꢀ4ꢀyl)benzeneꢀ1,2ꢀ
diamine 10c (650 mg, 2.55 mmol) and NꢀethylꢀNꢀisopropylpropanꢀ2ꢀamine (1.331 mL, 7.64 mmol) were dissolved in DMA (10
mL) and sealed into a microwave tube. The reaction was heated to 150 °C for 1 hour in the microwave reactor and cooled to 22
°C then evaporated to dryness. The residue was diluted with EtOAc (50 mL), and washed sequentially with water (50 mL) and
saturated brine (50 mL). The organic layer was dried over MgSO4, filtered and evaporated to afford crude product. The crude
product was purified by flash silica chromatography, elution gradient 0 to 5% MeOH in DCM. Pure fractions were evaporated to
dryness to afford N4ꢀ(2ꢀaminophenyl)ꢀ5ꢀchloroꢀN2ꢀ(tetrahydroꢀ2 Hꢀpyranꢀ4ꢀyl)pyrimidineꢀ2,4ꢀdiamine 10e (684 mg, 84%) as a
white solid. A solution of acryloyl chloride 10f (0.055 g, 0.611 mmol) in THF (3 mL) was added to a solution of N4ꢀ(2ꢀ
ACS Paragon Plus Environment

Page 17 of 24
Journal of Medicinal Chemistry
aminophenyl)ꢀ5ꢀchloroꢀN2ꢀ(tetrahydroꢀ2 Hꢀpyranꢀ4ꢀyl)pyrimidineꢀ2,4ꢀdiamine 10e (0.195 g, 0.611 mmol) and NꢀethylꢀNꢀ
isopropylpropanꢀ2ꢀamine (0.156 mL, 0.910 mmol) in THF (10 mL) at 0 °C. The resulting solution was stirred at 0 °C under a
nitrogen atmosphere for 35 minutes, then evaporated to dryness to afford a brown solid. The crude product was purified by silica
gel chromatography, eluting with 25% to 75% ethyl acetate in heptane. Fractions containing pure product were combined and
evaporated to dryness to afford Nꢀ(2ꢀ((5ꢀchloroꢀ2ꢀ((tetrahydroꢀ2 Hꢀpyranꢀ4ꢀyl)amino)pyrimidinꢀ4ꢀyl)amino)phenyl)acrylamide 10
(0.042 g, 19%). 1H NMR (400 MHz, DMSO) 1.39 – 1.54 (m, 2H), 1.71 – 1.82 (m, 2H), 3.23 – 3.40 (m, 3H), 3.86 (d, J = 10.3
Hz, 2H), 5.85 (dd, J = 1.9, 10.1 Hz, 1H), 6.36 (dd, J = 1.8, 17.0 Hz, 1H), 6.55 (dd, J = 10.1, 17.0 Hz, 1H), 6.94 (s, 1H), 7.27 (dtd,
J = 1.5, 7.5, 26.4 Hz, 2H), 7.43 (dd, J = 1.4, 7.8 Hz, 1H), 7.84 (d, J = 7.7 Hz, 1H), 7.99 (s, 1H), 8.33 (s, 1H), 10.20 (s, 1H).
HRMS: ESI+ m/z calculated for C18H20ClN5O2 [M+H]+ 374. 1384; m/z observed 374.1398.
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
The nonꢀcovalent matched pair examples 4, 5 and 11 were synthesised using the same methodology with the variation at the final
step. The nonꢀcovalent inhibitors 7 and 8 were prepared in analogous way to 628, starting from the appropriate commercially
available 2ꢀchloropyridine.
Biochemical and Cellular Assays
A full description of compound handling and experimental methods for all biochemical and cellular assays referred to can be
found in the supporting information section.
Molecular Modelling
The published ERK2 structure (pdb code 2OJI20) and inꢀhouse 4ZZN structures were used for the modelling of potential inhibitor
binding modes from the fragment and HTSꢀderived work respectively. All crystallographic water molecules were deleted for the
calculations as it was not deemed possible to confidently predict important water molecules upfront across the different series
modelled. However, to validate this protocol, a series of docking calculations were performed to show the crystallographic
inhibitors could be reliably reꢀdocked into the protein structures with and without the water molecules highlighted in the
manuscript. Crystal structures were prepared for docking using the protein preparation wizard in Maestro (Version 9.1,
Schrodinger), which optimises hydrogen placements. The active site was defined by the bound ligand. Glide was used for the
proteinꢀligand docking, in XP protocol. Multiple tautomers and protonation states were enumerated for ligands using our inꢀ
house software Leatherface33. A 3D structure of each ligand as input to the docking calculation was initially generated by Corina
and optimised with the MMFF force field (Sheffield solvation model34). Docked binding modes were ranked using the docking
score (GlideScore35) and visually inspected for the retention of key interactions to the hinge region (Leuꢀ107) along with other
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 24
close contacts. Binding modes were deprioritised if the conformation of the docked molecule was considered unsatisfactory. The
covalent docking algorithm in Glide was also used to assist the prioritisation of structures to covalently target Cysꢀ166.
1
2
3
4
5
6
7
Mass Spectrometry and protein structures
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
Protein was incubated with compound present at 3 times the molar concentration of the ERK2 protein. After incubation at room
temp for 2 hours, a small amount was diluted 10 fold for loading onto mass spectrometer immediately. Denaturing mass spec was
performed by loading samples into Waters Premier QTOF via Agilent LC in solvents containing 0.05% formic acid. Each run
lasted 5 minutes and was a single reverse phase binding step, which was followed by elution into the mass spectrometer in 90%
acetonitrile to desalt sample. The remaining sample was used to set up crystallisation drops. Protein was setꢀup 1:1 with 30%
polyethylene glycol monomethyl ether 2000, 200mM ammonium sulphate and 100 mM Hepes pH 7.6. Crystals appeared within a
week. They were flash frozen in the presence of 15% v/v 2,3ꢀbutandiol and xꢀray data collected at Dimond Light Source, (Didcot,
UK).
6His-TEV-HsERK2(1-360)-LE
1
MHHHHHHGGG ENLYFQGSHM AAAAAAGAGP EMVRGQVFDV GPRYTNLSYI
GEGAYGMVCS AYDNVNKVRV AIKKISPFEH QTYCQRTLRE IKILLRFRHE
NIIGINDIIR APTIEQMKDV YIVQDLMETD LYKLLKTQHL SNDHICYFLY
QILRGLKYIH SANVLHRDLK PSNLLLNTTC DLKICDFGLA RVADPDHDHT
GFLTEYVATR WYRAPEIMLN SKGYTKSIDI WSVGCILAEM LSNRPIFPGK
HYLDQLNHIL GILGSPSQED LNCIINLKAR NYLLSLPHKN KVPWNRLFPN
ADSKALDLLD KMLTFNPHKR IEVEQALAHP YLEQYYDPSD EPIAEAPFKF
DMELDDLPKE KLKELIFEET ARFQPGYRSL E
51
101
151
201
251
301
351
Expected mw: 43829.8
Data collections were carried out at the Diamond Light Source synchrotron beamlines at cryogenic temperatures, using ethylene
glycol as cryoprotectant. The pipedream and autoBUSTER software packages were used to solve and refine the structures, the
program Coot was used for manual building of the models. Data collection and refinement statistics together with PDB accession
codes of the final models are listed below. The structures have been deposited in the pdb as 4ZZM, 4ZZN and 4ZZO (Table 5).
Compound
1
7
10
PDB accession code
4ZZM
4ZZN
4ZZO
Data collection statistics
Space group and cell par:
P21
P21
P21
ACS Paragon Plus Environment

Page 19 of 24
Journal of Medicinal Chemistry
1
2
3
4
a, b,c [Å]
β [°]
48.88, 70.31, 60.41
109.12
48.76, 69.70, 59.80
48.75, 70.07, 60.10
109.25
108.71
1.50
5
Resolution [Å]
1.89
1.63
6
7
8
9
Unique reflections
Multiplicity
30202 (2182)
3.2 (3.2)
55281 (1580)
3.1 (2.2)
47270 (3474)
3.3 (3.2)
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
Completeness [%]
Rsym [%]
97.5 (94.1)
8.6 (57.3)
8.9 (1.9)
91.1 (53.1)
2.9 (9.3)
99.3 (98.8)
7.2 (56.7)
9.9 (2.1)
Mean(I)/sd
22.4 (6.3)
Refinement statistics
Number of reflections
(working /test)
R/Rfree [%]
30183 (1903)
21.04/25.44
61548 (2460)
17.47/20.23
47250 (2593)
19.13/23.73
Deviation from ideal
geometry
bond lengths [Å]
bond angles [◦]
Ramachandran plot [%]
preferred region
generously allowed region
disallowed region
0.010
0.990
0.012
1.060
0.010
1.010
95.6
3.5
96.71
2.4
96.1
2.9
0.8
0.9
0.9
ǂ Data in parentheses refer to the highest resolution shell.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 5 – Protein crystal structure statistics.
Immunoblotting
Cells were washed on ice with cold PBS before the addition lysis buffer containing 25mmol/L TrisꢀHCl, 3 mmol/L EDTA, 3
mmol/L EGTA, 50 mmol/L sodium fluoride, 2 mmol/L sodium orthovanadate, 10mmol/L βꢀglycerophosphate, 5mmol/L
pyrophosphate, 0.5% Triton Xꢀ100, 0.1% βꢀmercaptoethanol supplemented with protease inhibitor cocktail (Roche). After
scraping, lysates were transferred to microcentrifuge tubes and shaken at 4°C for 20 minutes before 10ꢀminute centrifugation and
protein quantification with Bradford Reagent (BioꢀRad). Equal protein amounts were loaded for SDSꢀPAGE using 4ꢀ20% Trisꢀ
HCl gels (BioꢀRad) followed by transfer to nitrocellulose membranes. After blocking in 5% BSAꢀTBST or milkꢀTBST,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 24
membranes were blotted with total p90RSK (BD Bioscience 610226), phosphoꢀp90RSK (Thr359/Ser363, Cell Signaling
1
2
3
Technology 9344) and vinculin (Sigma V9131) followed by horseradish peroxidase (HRP)–conjugated secondary antibodies (Cell
4
5
Signaling Technology; 7074 or 7076). Signals were detected with SuperSignal West Dura Chemiluminescent Substrate, Thermo
6
7
Scientific.
8
9
References
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. Schirmer, A.; Kennedy, J.; Murli, S.; Reid, R.; Santi, D. V. Targeted covalent inactivation of protein kinases by resorcylic acid
lactone polyketides. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 4234ꢀ4239.
2. Dhillon, A. S.; Hagan, S.; Rath, O.; Kolch, W. MAP kinase signalling pathways in cancer. Oncogene 2007, 26, 3279ꢀ3290.
3. Hanahan, D.; Weinberg, R. A. The hallmarks of cancer. Cell 2000, 100, 57ꢀ70.
4. Yoon, S.; Seger, R. The extracellular signalꢀregulated kinase: multiple substrates regulate diverse cellular functions. Growth
Factors 2006, 24, 21ꢀ44.
5. Owens, D. M.; Keyse, S. M. Differential regulation of MAP kinase signalling by dualꢀspecificity protein phosphatases.
Oncogene 2007, 26, 3203ꢀ3213.
6. Vakiani, E.; Solit, D. B. KRAS and BRAF: drug targets and predictive biomarkers. J. Pathol. 2011, 223, 219ꢀ229.
7. Solit, D. B.; Garraway, L. A.; Pratilas, C. A.; Sawai, A.; Getz, G.; Basso, A.; Ye, Q.; Lobo, J. M.; She, Y.; Osman, I.; Golub,
T. R.; SeboltꢀLeopold, J.; Sellers, W. R.; Rosen, N. BRAF mutation predicts sensitivity to MEK inhibition. Nature 2006, 439,
358ꢀ362.
8. Ji, Z.; Flaherty, K. T.; Tsao, H. Targeting the RAS pathway in melanoma. Trends Mol. Med. 2012, 18, 27ꢀ35.
9. Flaherty, K. T.; Infante, J. R.; Daud, A.; Gonzalez, R.; Kefford, R. F.; Sosman, J.; Hamid, O.; Schuchter, L.; Cebon, J.;
Ibrahim, N.; Kudchadkar, R.; Burris, H. A.,3rd; Falchook, G.; Algazi, A.; Lewis, K.; Long, G. V.; Puzanov, I.; Lebowitz, P.;
Singh, A.; Little, S.; Sun, P.; Allred, A.; Ouellet, D.; Kim, K. B.; Patel, K.; Weber, J. Combined BRAF and MEK inhibition in
melanoma with BRAF V600 mutations. N. Engl. J. Med. 2012, 367, 1694ꢀ1703.
10. Montagut, C.; Sharma, S. V.; Shioda, T.; McDermott, U.; Ulman, M.; Ulkus, L. E.; DiasꢀSantagata, D.; Stubbs, H.; Lee, D.
Y.; Singh, A.; Drew, L.; Haber, D. A.; Settleman, J. Elevated CRAF as a potential mechanism of acquired resistance to BRAF
inhibition in melanoma. Cancer Res. 2008, 68, 4853ꢀ4861.
ACS Paragon Plus Environment

Page 21 of 24
Journal of Medicinal Chemistry
11. Wagle, N.; Emery, C.; Berger, M. F.; Davis, M. J.; Sawyer, A.; Pochanard, P.; Kehoe, S. M.; Johannessen, C. M.; Macconaill,
L. E.; Hahn, W. C.; Meyerson, M.; Garraway, L. A. Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor
genomic profiling. J. Clin. Oncol. 2011, 29, 3085ꢀ3096.
1
2
3
4
5
6
7
8
9
12. Emery, C. M.; Vijayendran, K. G.; Zipser, M. C.; Sawyer, A. M.; Niu, L.; Kim, J. J.; Hatton, C.; Chopra, R.; Oberholzer, P.
A.; Karpova, M. B.; MacConaill, L. E.; Zhang, J.; Gray, N. S.; Sellers, W. R.; Dummer, R.; Garraway, L. A. MEK1 mutations
confer resistance to MEK and BꢀRAF inhibition. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 20411ꢀ20416.
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
13. Poulikakos, P. I.; Persaud, Y.; Janakiraman, M.; Kong, X.; Ng, C.; Moriceau, G.; Shi, H.; Atefi, M.; Titz, B.; Gabay, M. T.;
Salton, M.; Dahlman, K. B.; Tadi, M.; Wargo, J. A.; Flaherty, K. T.; Kelley, M. C.; Misteli, T.; Chapman, P. B.; Sosman, J. A.;
Graeber, T. G.; Ribas, A.; Lo, R. S.; Rosen, N.; Solit, D. B. RAF inhibitor resistance is mediated by dimerization of aberrantly
spliced BRAF(V600E). Nature 2011, 480, 387ꢀ390.
14. Morris, E. J.; Jha, S.; Restaino, C. R.; Dayananth, P.; Zhu, H.; Cooper, A.; Carr, D.; Deng, Y.; Jin, W.; Black, S.; Long, B.;
Liu, J.; Dinunzio, E.; Windsor, W.; Zhang, R.; Zhao, S.; Angagaw, M. H.; Pinheiro, E. M.; Desai, J.; Xiao, L.; Shipps, G.; Hruza,
A.; Wang, J.; Kelly, J.; Paliwal, S.; Gao, X.; Babu, B. S.; Zhu, L.; Daublain, P.; Zhang, L.; Lutterbach, B. A.; Pelletier, M. R.;
Philippar, U.; Siliphaivanh, P.; Witter, D.; Kirschmeier, P.; Bishop, W. R.; Hicklin, D.; Gilliland, D. G.; Jayaraman, L.; Zawel, L.;
Fawell, S.; Samatar, A. A. Discovery of a novel ERK inhibitor with activity in models of acquired resistance to BRAF and MEK
inhibitors. Cancer. Discov. 2013, 3, 742ꢀ750.
15. BVDꢀ523. BioMed Valley Discoveries. https://biomedꢀvalley.com/portfolio/bvdꢀ523, Accessed April 2015.
16. GDCꢀ0994. Proceedings: AACR Annual Meeting 2014; April 5ꢀ9, 2014; San Diego, CA. Robarge, K et al. Genentech, Inc.,
South San Francisco, CA; Array Biopharma, Boulder, CO.
17. Deng, Y.; Shipps, G. W.,Jr; Cooper, A.; English, J. M.; Annis, D. A.; Carr, D.; Nan, Y.; Wang, T.; Zhu, H. Y.; Chuang, C. C.;
Dayananth, P.; Hruza, A. W.; Xiao, L.; Jin, W.; Kirschmeier, P.; Windsor, W. T.; Samatar, A. A. Discovery of novel, dual
mechanism ERK inhibitors by affinity selection screening of an inactive kinase. J. Med. Chem. 2014, 57, 8817ꢀ8826.
18. Blake, J. F.; Gaudino, J. J.; De Meese, J.; Mohr, P.; Chicarelli, M.; Tian, H.; Garrey, R.; Thomas, A.; Siedem, C. S.; Welch,
M. B.; Kolakowski, G.; Kaus, R.; Burkard, M.; Martinson, M.; Chen, H.; Dean, B.; Dudley, D. A.; Gould, S. E.; Pacheco, P.;
ShahidiꢀLatham, S.; Wang, W.; West, K.; Yin, J.; Moffat, J.; Schwarz, J. B. Discovery of 5,6,7,8ꢀtetrahydropyrido[3,4ꢀ
d]pyrimidine inhibitors of Erk2. Bioorg. Med. Chem. Lett. 2014, 24, 2635ꢀ2639.
19. Aronov, A. M.; Tang, Q.; MartinezꢀBotella, G.; Bemis, G. W.; Cao, J.; Chen, G.; Ewing, N. P.; Ford, P. J.; Germann, U. A.;
Green, J.; Hale, M. R.; Jacobs, M.; Janetka, J. W.; Maltais, F.; Markland, W.; Namchuk, M. N.; Nanthakumar, S.; Poondru, S.;
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 24
Straub, J.; ter Haar, E.; Xie, X. Structureꢀguided design of potent and selective pyrimidylpyrrole inhibitors of extracellular signalꢀ
1
2
3
regulated kinase (ERK) using conformational control. J. Med. Chem. 2009, 52, 6362ꢀ6368.
4
5
6
7
8
20. Aronov, A. M.; Baker, C.; Bemis, G. W.; Cao, J.; Chen, G.; Ford, P. J.; Germann, U. A.; Green, J.; Hale, M. R.; Jacobs, M.;
Janetka, J. W.; Maltais, F.; MartinezꢀBotella, G.; Namchuk, M. N.; Straub, J.; Tang, Q.; Xie, X. Flipped out: structureꢀguided
design of selective pyrazolylpyrrole ERK inhibitors. J. Med. Chem. 2007, 50, 1280ꢀ1287.
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
21. Ohori, M.; Kinoshita, T.; Yoshimura, S.; Warizaya, M.; Nakajima, H.; Miyake, H. Role of a cysteine residue in the active site
of ERK and the MAPKK family. Biochem. Biophys. Res. Commun. 2007, 353, 633ꢀ637.
22. Leproult, E.; Barluenga, S.; Moras, D.; Wurtz, J. M.; Winssinger, N. Cysteine mapping in conformationally distinct kinase
nucleotide binding sites: application to the design of selective covalent inhibitors. J. Med. Chem. 2011, 54, 1347ꢀ1355.
23. Singh, J.; Petter, R. C.; Baillie, T. A.; Whitty, A. The resurgence of covalent drugs. Nat. Rev. Drug Discovery 2011, 10, 307ꢀ
317.
24. Cooper, A.; Deng, Y.; Shipps Jr., G. W; Shih, NꢀY.; Zhu, H.; Sun, R.; Kelly, .J; Doll, R.; Nan, Y.; Wang, T.; Desai, J.;
Wang, J.; Dong, Y.; Madison, V. S.; Li, X.; Hruza, A.; Siddiqui, M. A.; Samatar, A.; Paliwal, S.; Tsui, HꢀC.; Celebi, A. A.; Wu,
Y.; Boga, S. B. WO2007070398, Schering Corporation; 2000 Galloping Hill Road, Kenilworth, NJ 07033ꢀ0530, US.
25. Cooper, A.; Deng, Y.; Shipps Jr., G. W.; Sun, B.; Wang, J.; Zhu, L. WO2012036997, Schering Corporation; 2000
Galloping Hill Road, Kenilworth, NJ 07033-0530, US.
26. Eurofins (Millipore), http://www.eurofins.co.uk, Accessed April 2015.
27. Finlay, M. R.; Anderton, M.; Ashton, S.; Ballard, P.; Bethel, P. A.; Box, M. R.; Bradbury, R. H.; Brown, S. J.; Butterworth,
S.; Campbell, A.; Chorley, C.; Colclough, N.; Cross, D. A.; Currie, G. S.; Grist, M.; Hassall, L.; Hill, G. B.; James, D.; James, M.;
Kemmitt, P.; Klinowska, T.; Lamont, G.; Lamont, S. G.; Martin, N.; McFarland, H. L.; Mellor, M. J.; Orme, J. P.; Perkins, D.;
Perkins, P.; Richmond, G.; Smith, P.; Ward, R. A.; Waring, M. J.; Whittaker, D.; Wells, S.; Wrigley, G. L. Discovery of a potent
and selective EGFR inhibitor (AZD9291) of both sensitizing and T790M resistance mutations that spares the wild type form of
the receptor. J. Med. Chem. 2014, 57, 8249ꢀ8267.
28. Barlaam, B. C; Foote, K. M; Ple, P. WO2009153589, AstraZeneca UK Ltd; 15 Stanhope Gate London Greater London W1K
1LN, UK.
ACS Paragon Plus Environment

Page 23 of 24
Journal of Medicinal Chemistry
29. Ward, R. A.; Anderton, M. J.; Ashton, S.; Bethel, P. A.; Box, M.; Butterworth, S.; Colclough, N.; Chorley, C. G.; Chuaqui,
C.; Cross, D. A.; Dakin, L. A.; Debreczeni, J. E.; Eberlein, C.; Finlay, M. R.; Hill, G. B.; Grist, M.; Klinowska, T. C.; Lane, C.;
Martin, S.; Orme, J. P.; Smith, P.; Wang, F.; Waring, M. J. Structureꢀ and reactivityꢀbased development of covalent inhibitors of
the activating and gatekeeper mutant forms of the epidermal growth factor receptor (EGFR). J. Med. Chem. 2013, 56, 7025ꢀ7048.
1
2
3
4
5
6
7
8
9
30. Schwanhaeusser, B.; Busse, D.; Li, N.; Dittmar, G.; Schuchhardt, J.; Wolf, J.; Chen, W.; Selbach, M. Global quantification of
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
mammalian gene expression control [Erratum to document cited in CA155:060687]. Nature (London, U. K. ) 2013, 495, 126ꢀ127.
31. Compounds sharing the pyrimidine scaffold have recently been disclosed as covalent inhibitors of Erk in a patent from
Celgene Avilomics Research, Inc. Haq, N; Niu, D; Petter, R. C; Qiao, L; Singh, J; Zhu, Z. WO2014124230, Celgene Avilomics
Research, Inc.
32. Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful metric for lead selection. Drug Discov Today 2004, 9,
430ꢀ431.
33. Lyne, P. D.; Kenny, P. W.; Cosgrove, D. A.; Deng, C.; Zabludoff, S.; Wendoloski, J. J.; Ashwell, S. Identification of
Compounds with Nanomolar Binding Affinity for Checkpoint Kinaseꢀ1 Using KnowledgeꢀBased Virtual Screening. J. Med.
Chem. 2004, 47, 1962ꢀ1968.
34. Grant, J. A.; Pickup, B. T.; Sykes, M. J.; Kitchen, C. A.; Nicholls, A. A simple formula for dielectric polarisation energies:
The Sheffield Solvation Model. Chem. Phys. Lett. 2007, 441, 163ꢀ166.
35. Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.; Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D.
T. Extra Precision Glide: Docking and Scoring Incorporating a Model of Hydrophobic Enclosure for ProteinꢀLigand Complexes.
J. Med. Chem. 2006, 49, 6177ꢀ6196.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 24
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
Table of Contents graphic
ACS Paragon Plus Environment