Discovery of Quinazolines That Activate SOS1-Mediated Nucleotide Exchange on RAS

Article abstract of DOI:10.1021/acsmedchemlett.8b00296

Proteins in the RAS family are important regulators of cellular signaling and, when mutated, can drive cancer pathogenesis. Despite considerable effort over the last 30 years, RAS proteins have proven to be recalcitrant therapeutic targets. One approach for modulating RAS signaling is to target proteins that interact with RAS, such as the guanine nucleotide exchange factor (GEF) son of sevenless homologue 1 (SOS1). Here, we report hit-to-lead studies on quinazoline-containing compounds that bind to SOS1 and activate nucleotide exchange on RAS. Using structure-based design, we refined the substituents attached to the quinazoline nucleus and built in additional interactions not present in the initial HTS hit. Optimized compounds activate nucleotide exchange at single-digit micromolar concentrations in vitro. In HeLa cells, these quinazolines increase the levels of RAS-GTP and cause signaling changes in the mitogen-activated protein kinase/extracellular regulated kinase (MAPK/ERK) pathway.

Full text of DOI:10.1021/acsmedchemlett.8b00296

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Letter
Discovery of Quinazolines That Activate
SOS1-Mediated Nucleotide Exchange on RAS
Jason R. Abbott, Pratiq A. Patel, Jennifer E. Howes, Denis T. Akan, J. Phillip Kennedy, Michael C. Burns,
Carrie F. Browning, Qi Sun, Olivia W. Rossanese, Jason Phan, Alex G. Waterson, and Stephen W. Fesik
ACS Med. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acsmedchemlett.8b00296 • Publication Date (Web): 08 Aug 2018
Downloaded from http://pubs.acs.org on August 8, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 7
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Discovery of Quinazolines That Activate SOS1ꢀMediated Nucleotide
Exchange on RAS
Jason R. Abbott,◊ Pratiq A. Patel,◊,† Jennifer E. Howes,◊ Denis T. Akan,◊ J. Phillip Kennedy,◊,‡ Michael
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
C. Burns,◊,⊥ Carrie F. Browning,◊,# Qi Sun,◊,¶ Olivia W. Rossanese,◊,£ Jason Phan,◊ Alex G. Waterson,§,¥
and Stephen W. Fesik◊,§,¥,*
◊Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tennessee 37232ꢀ0146, USA
§Department of Pharmacology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232ꢀ0146, USA
¥Department of Chemistry, Vanderbilt University, Nashville, Tennessee 37232ꢀ0146, USA
KEYWORDS RAS, SOS1, nucleotide exchange, structure-based design, quinazolines
ABSTRACT: Proteins in the RAS family are important regulators of cellular signaling and, when mutated, can drive cancer pathoꢀ
genesis. Despite considerable effort over the last thirty years, RAS proteins have proven to be recalcitrant therapeutic targets. One
approach for modulating RAS signaling is to target proteins that interact with RAS, such as the guanine nucleotide exchange factor
(GEF) son of sevenless homologue 1 (SOS1). Here, we report hitꢀtoꢀlead studies on quinazolineꢀcontaining compounds that bind to
SOS1 and activate nucleotide exchange on RAS. Using structureꢀbased design, we refined the substituents attached to the quinazoꢀ
line nucleus and built in additional interactions not present in the initial HTS hit. Optimized compounds activate nucleotide exꢀ
change at singleꢀdigit micromolar concentrations in vitro. In HeLa cells, these quinazolines increase the levels of RASꢀGTP and
cause signaling changes in the mitogenꢀactivated protein kinase/extracellular regulated kinase (MAPK/ERK) pathway.
The three RAS genes—HRAS, KRAS, and NRAS—constitute
the most frequently mutated gene family in human cancers.1
With RAS mutations found in approximately one third of all
human tumors, these oncogenes represent important targets in
the field of drug discovery.2ꢀ3 RAS genes encode small GTPꢀ
binding proteins that function as molecular switches, transmitꢀ
ting extracellular stimuli to intracellular signaling pathways.
Canonically, RAS proteins exist in two states: the guanosine
diphosphate (GDP)ꢀbound “off” state and the guanosine triꢀ
phosphate (GTP)ꢀbound “on” state.4ꢀ5 RAS activity is tightly
governed by GTPaseꢀactivating proteins (GAPs) and guanine
nucleotide exchange factors (GEFs).6 Mutations in RAS genes
can render the proteins that they encode constitutively active,
and cells harboring mutated RAS proteins often display the
hallmarks of cancer.7ꢀ9 To date, multiple strategies for targetꢀ
ing RAS have been developed, but each has met with limited
success.3,10ꢀ13
Our group has reported an approach for modulating RAS
signaling that entails binding to and activating the protein son
of sevenless homologue 1 (SOS1), a GEF that interacts directꢀ
ly with RAS and catalyzes the exchange of GDP for GTP.14ꢀ16
In these studies, we described compounds from an indole seꢀ
ries that bind to the CDC25 domain of SOS1 in the
RAS:SOS1:RAS complex, activate nucleotide exchange at
subꢀmicromolar concentrations in vitro, increase the levels of
RASꢀGTP in cancer cells, and cause biphasic signaling changꢀ
es in the mitogenꢀactivated protein kinase/extracellular reguꢀ
lated kinase (MAPK/ERK) pathway. Here, we describe the
discovery of a second series of compounds, based on a
quinazoline scaffold, that bind to the same pocket on SOS1,
but in a different manner, which is characterized by the rotaꢀ
tion of a Phe residue inside the binding pocket. Importantly,
compounds exhibiting this alternative binding mode elicit
similar effects in vitro and in cells as the previously reported
indoles.
From a high throughput nucleotide exchange assay used to
screen the Vanderbilt library of >160,000 compounds, 2,880
small molecules were identified that increased the rate of
SOS1ꢀmediated nucleotide exchange on RAS (1.8% initial hit
rate).17 Among the hit molecules in the screen, the 2,4ꢀ
diaminoquinazoline, 1, was identified for its ability to activate
nucleotide exchange with similar efficacy as the positive conꢀ
trol (Rel. Act. = 96%) and with good potency (EC50 = 9.7 ꢁM)
(Figure 1). An advantage of the quinazoline scaffold was the
potential for rapid diversification. Indeed, the amine substituꢀ
ents at positions 2 and 4 could each be readily exchanged, and
the carbon atoms of the quinazoline arene ring offered addiꢀ
tional vectors into distinct areas of chemical space.
To guide the first round of compound synthesis, the binding
mode of the HTS hit, 1, was identified by Xꢀray crystallogꢀ
raphy. As shown in the Xꢀray coꢀcrystal structure depicted in
Figure 1, compound 1 binds to a hydrophobic pocket on SOS1
in the RAS:SOS1:RAS complex. While this is the same bindꢀ
ing site occupied by compounds in the indole series, the aroꢀ
matic ring of Phe890 is flipped approximately 90 degrees relaꢀ
tive to its position in the coꢀcrystal structures obtained with the
indole compounds.14,16 A similar rotation of this residue was
also observed during a fragment screening campaign conductꢀ
ed by a group from AstraZeneca.18 This rotation unveiled a
new hydrophobic shelf inside of the binding pocket that is
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 2 of 7
aEach value represents the mean ± SD of at least two separate
experiments.
partially occupied by the 3′ꢀchloroaniline substituent on comꢀ
1
2
pound 1. This region of chemical space, which was unexꢀ
plored by the compounds in the indole series, provided an
excellent starting point for structureꢀbased design.
b
“—” denotes an EC50 value of >100 ꢁM, or an EC50 value that
was not calculated due to low efficacy in vitro.
3
4
5
6
7
8
9
cActivation values represent the percentage activation for each
compound relative to the activation of a positive control comꢀ
pound.14,16ꢀ17
Hydrophobic
Having explored the 3′ꢀposition of the aniline ring, we next
investigated alternative substitution patterns. Moving the subꢀ
stituent from the 3′ꢀposition to the 4′ꢀposition compromised
both potency and efficacy in vitro—compare, for example, the
4′ꢀmethyl analogue, 7, and the 4′ꢀbromo analogue, 8, with their
respective 3′ꢀregioisomers 3 and 4—suggesting that the 3′ꢀ
substituent was important for compound activity. With this
consideration in mind, several quinazolines featuring disubstiꢀ
tuted aniline rings were then prepared. Appending a second
substituent to either Cꢀ4′ or Cꢀ5′ of the aniline ring was detriꢀ
mental to activity when this substituent was either methyl (10
and 11) or chloro (12 and 13). However, smaller atoms such as
fluorine were well tolerated, as illustrated by compound 14,
which was of similar potency and efficacy to 1. Being among
the most potent and efficacious analogues in Table 1, comꢀ
pound 14 was identified as the scaffold upon which to build
further SAR.
shelf
Phe890
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 1. Xꢀray coꢀcrystal structure of compound 1 (beige; PDB
ID code 6CUO) bound to SOS1 in the RAS:SOS1:RAS ternary
complex. SOS1 protein surface is shown in gray with key binding
site residues displayed in pink. The alternative orientation of
Phe890 is shown in cyan.
With adequate SAR established at the 2ꢀposition of the
quinazoline scaffold, our focus shifted to filling the unoccuꢀ
pied space near the (aminomethyl)furan of the HTS hit, 1. The
exocyclic nitrogen atom at the 4ꢀposition of the quinazoline
appeared to offer a vector into this area of the binding pocket;
however, in the Xꢀray coꢀcrystal structure depicted in Figure
1, this nitrogen atom is solvent exposed, and engaged in a
waterꢀmediated interaction with Glu902. It was not clear, a
priori, whether substituents on this nitrogen atom would be
tolerated. Thus, the Nꢀethyl analogue, 15, and the Nꢀacetyl
analogue, 16, were prepared (Table 2). Unfortunately, both of
these analogues displayed low efficacy in the nucleotide exꢀ
change assay, which suggested that substituting this nitrogen
atom would not be a viable design strategy.
In the Xꢀray coꢀcrystal structure depicted in Figure 1, the
space beneath Phe890—around the aniline ring of compound
1—appeared to be mainly hydrophobic in nature. Thus, lipoꢀ
philic groups seemed to be well suited for this area of the
binding pocket. For reference, the unsubstituted phenyl derivaꢀ
tive, 2, was prepared and proved to be inactive (Table 1). Reꢀ
introduction of a lipophilic substituent at the 3′ꢀposition of the
aniline ring restored activity, as in the 3′ꢀmethyl derivative, 3,
and the 3′ꢀbromo derivative, 4. On the other hand, less lipoꢀ
philic groups, such as the 3′ꢀmethoxy group in compound 5
and the 3′ꢀ(1ꢀhydroxyethyl) group in compound 6 were not
well tolerated.
Table 1. SAR of the Aniline Ringa
Table 2. SAR of the Quinazoline 4ꢀPositiona
Compd
R′
3′ꢀCl
EC50 (µM)b Rel. Act. (%)c
Compd
15
R
EC50 (µM)b
Rel. Act. (%)c
1
2
9.7 ± 1.16
96 ± 13.9
47 ± 13.6
96 ± 2.1
87 ± 12.5
38 ± 8.5
10 ± 4.8
8 ± 1.4
—
25 ± 4.4
H
—
3
3′ꢀMe
16.4 ± 0.35
16
17
—
16 ± 2.3
67 ± 1.1
4
3′ꢀBr
7.1 ± 1.93
5
3′ꢀOMe
3′ꢀ(1ꢀhydroxyethyl)
4′ꢀMe
—
21.9 ± 1.33
6
—
7
—
18
19
20
21
—
55 ± 3.1
111 ± 25.8
69 ± 4.9
8
4′ꢀBr
—
16 ± 4.0
19 ± 0.9
37 ± 11.3
44 ± 20.4
22 ± 5.6
41 ± 26.5
96 ± 10.6
16.3 ± 3.50
44.2 ± 19.35
—
9
4′ꢀOMe
3′,4′ꢀdiMe
3′,5′ꢀdiMe
3′,4′ꢀdiCl
3′,5′ꢀdiCl
3′ꢀClꢀ4′ꢀF
—
10
11
12
13
14
—
—
—
58 ± 4.1
—
22
6.4 ± 0.89
122 ± 77.7
13.1 ± 1.20
2
ACS Paragon Plus Environment

Page 3 of 7
ACS Medicinal Chemistry Letters
exemplified by the cyclobutyl analogues, 28 and 29, both of
which were inactive.
23
24
—
67 ± 5.9
1
2
3
6.2 ± 2.49
144 ± 7.0
4
Leu901
5
6
7
8
25
26
—
48 ± 0.1
Asp887
2.5 ± 0.79
74 ± 10.1
Glu902
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
27
—
—
28
29
—
—
43 ± 7.5
41 ± 3.0
Tyr884
His905
Figure 2. Xꢀray coꢀcrystal structure of compound 22 (green; PDB
ID code 6CUP) bound to SOS1 in the RAS:SOS1:RAS ternary
complex. SOS1 protein surface is shown in gray with key binding
site residues displayed in pink. Yellow circle indicates hydrophoꢀ
bic space near Leu901.
aEach value represents the mean ± SD of at least two separate
experiments.
b
“—” denotes an EC50 value of >100 ꢁM, or an EC50 value that
was not calculated due to low efficacy in vitro.
cActivation values represent the percentage activation for each
compound relative to the activation of a positive control comꢀ
pound.14,16ꢀ17
The Xꢀray coꢀcrystal structure of compound 22 presented an
additional avenue for further compound design. As shown in
Figure 2, the exocyclic nitrogen at the 2ꢀposition of the
quinazoline scaffold and the endocyclic Nꢀ1 ring nitrogen
engage the carboxylate group of Asp887 through a molecule
of water. During the AstraZeneca fragment screening camꢀ
paign (vide supra), several ligands were also shown to interact
directly with this Asp residue.18 Furthermore, whereas mutaꢀ
tion of Asp887 did not affect the activity of our previously
reported indole compounds,14 several of the HTS hits lost the
ability to activate nucleotide exchange when tested with the
Asp887Ala and Asp887His mutant forms of SOS1.22 Thus, we
speculated that this residue could be leveraged to improve the
activity of the compounds in this quinazoline series.
A more productive approach for increasing the occupancy
of the binding pocket involved branching from the carbon
atoms of the amine substituent. This strategy allowed us to
introduce more sp3 character into these quinazoline comꢀ
pounds,19 and provided an opportunity to move away from the
furan moiety, a known metabolic liability.20ꢀ21 As illustrated in
Table 2, promising results were obtained with αꢀbranched
amines, such as the pentanꢀ3ꢀamine featured in compound 17,
and smaller cycloalkyl rings, such as the cyclopropyl group
featured in compound 19. Further potency improvements were
realized when these two design elements were combined, as in
the (R)ꢀ1ꢀcyclopropylethanꢀ1ꢀamine featured in compound 22,
which activated the nucleotide exchange process at 122% relaꢀ
tive to control, with an EC50 value of 6.4 ꢁM. Remarkably, the
(S)ꢀenantiomer matched pair 23 demonstrated a large dropꢀoff
in exchange activity. To clarify the binding mode of the new
lead compound 22, and to further guide structureꢀbased deꢀ
sign, the Xꢀray coꢀcrystal structure of this compound bound to
the RAS:SOS1:RAS complex was obtained.
As shown in Figure 2, the cyclopropyl substituent of comꢀ
pound 22 sits behind His905 of SOS1, while the methyl
branch partially fills the hydrophobic space near Leu901 (Figꢀ
ure 2, yellow circle). This arrangement is enforced by the (R)ꢀ
configuration of the stereogenic carbon. Our analysis of this
coꢀcrystal structure further suggested that additional pocket
occupancy could be achieved by extending the alkyl branch
into the hydrophobic area near Leu901. Thus, several comꢀ
pounds were prepared to explore this region of chemical
space. The preference for the (R)ꢀenantiomer was confirmed
with the ethyl derivatives, 24 and 25, but no marked improveꢀ
ment in activity was realized. While the propyl derivative, 26,
was more potent than the methyl comparator, 22, a noticeable
decrease in efficacy was observed, and further increasing the
size of the alkyl branch, as with the isobutyl analogue, 27,
ablated nucleotide exchange activity altogether. Decreases in
activity were also observed as the size of the cycloalkyl subꢀ
stituent extended beyond cyclopropyl. This general trend is
To this end, several 8ꢀsubstituted quinazolines intended to
directly engage Asp887 via a charge–charge interaction or a
hydrogen bond were designed. For these studies, the (R)ꢀ1ꢀ
cyclopropylethanꢀ1ꢀamine was retained at the 4ꢀposition of the
quinazoline scaffold. The data for compounds 22 and 24 reꢀ
ported in Table 2, as well as data for other matched pairs (not
shown) suggested that the methyl branch (e.g. 22) and the
ethyl branch (e.g. 24) could be used to similar effect, and the
commercial availability of the (R)ꢀ1ꢀcyclopropylethanꢀ1ꢀ
amine made it a more attractive choice for subsequent rounds
of compound synthesis.
Table 3. SAR of the Quinazoline 8ꢀPositiona
Compd
30
R
EC50 (µM)b
Rel. Act. (%)c
2.5 ± 0.16
54 ± 8.8
31
32
3.0 ± 0.01
2.7 ± 0.18
160 ± 4.5
203 ± 1.9
33
34
2.0 ± 0.58
1.8 ± 0.05
148 ± 31.5
201 ± 14.7
3
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 4 of 7
phorylation, where increased levels of phosphorylated ERK1/2
(pERK1/2T202/Y204) are observed at lower compound concentraꢀ
tions and decreased levels of pERK1/2T202/Y204 are observed at
higher concentrations. Mechanistically, inhibition of ERK1/2
phosphorylation is achieved at higher treatment concentrations
35
36
1.1 ± 0.37
4.2 ± 0.29
132 ± 31.5
204 ± 7.4
1
2
3
4
5
6
7
8
through negative feedback on SOS1 by pERK1/2T202/Y204 15ꢀ17
.
37
38
6.6 ± 0.02
7.7 ± 0.18
140 ± 14.7
106 ± 8.5
9
Asp887
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
39
40
—
—
36 ± 1.9
Glu902
83 ± 17.6
41
42
—
—
87 ± 11.4
76 ± 11.8
aEach value represents the mean ± SD of at least two separate
experiments.
b
“—” denotes an EC50 value of >100 ꢁM, or an EC50 value that
was not calculated due to low efficacy in vitro.
cActivation values represent the percentage activation for each
compound relative to the activation of a positive control comꢀ
pound.14,16ꢀ17
Figure 3. Xꢀray coꢀcrystal structure of compound 34 (magenta;
PDB ID code 6CUR) bound to SOS1 in the RAS:SOS1:RAS
ternary complex. SOS1 protein surface is shown in gray with key
binding site residues displayed in pink.
As the data in Table 3 illustrate, many of the 8ꢀsubstituted
quinazolines demonstrated improved potency and efficacy
relative to the parent 8Hꢀquinazoline, 22. In particular, anaꢀ
logues containing a substituent bearing an amino group perꢀ
formed especially well in the assay. The length of the tether
connecting the amine to the quinazoline had a noticeable efꢀ
To test whether these quinazolines, which bind to SOS1
with Phe890 in a different conformation than was observed
previously,14,16 elicit the same signaling changes in cells as our
indole compounds, the effects of four exemplar compounds—
22, 30, 32 and 34—were assessed by western blotting. The
levels of endogenous cellular RASꢀGTP were measured in
response to compound treatment, along with the associated
levels of pERK1/2T202/Y204 and the total ERK1/2 protein levels.
As shown in Figure 4, an increase in the levels of RASꢀGTP
was observed after treatment with compounds 32 and 34.
Compound 30 elicited a small increase in RASꢀGTP levels
only at higher concentrations, consistent with its lower relative
percent activation. All three of these compounds elicited the
expected biphasic modulation of ERK1/2 phosphorylation.
The cellular activity of quinazolines 30, 32, and 34 is thus
consistent with that of the previously reported indoles,14ꢀ16 and
exemplifies the characteristic activity of our SOS1 agonist
compounds. Together, these data suggest that the compounds
from these two distinct chemical series, which demonstrate
different binding modes to SOS1, likely act via the same bioꢀ
logical mechanism. In contrast, compound 22 did not induce a
marked increase in the levels of RASꢀGTP at the concentraꢀ
tions tested. Despite this lack of RASꢀGTP induction, a subꢀ
stantial increase in the levels of pERK1/2T202/Y204 was noted.
The anticipated decrease in ERK1/2 phosphorylation however,
was not observed. The effects on pERK1/2T202/Y204 signaling
elicited by five less active compounds—13, 25, 27, 28, and
29—were also measured and resemble the effects elicited by
compound 22, namely, an increase in ERK1/2 phosphorylation
at higher compound concentrations (Figure S1).
fect
(aminomethyl)quinazoline 30 exhibited good potency, but was
less efficacious than the homologous 8ꢀ
on
compound
activity.
For
example,
8ꢀ
(aminoethyl)quinazoline 32. Cyclic amines were well tolerated
(e.g. 34–36) as were acyclic amines (e.g. 32, 37, and 38). On
the other hand, compounds featuring amino alcohols (e.g. 39)
and ethers (e.g. 40–42) were essentially inactive, establishing
a clear preference for amineꢀcontaining substituents that preꢀ
sumably engage in a charge–charge interaction with the carꢀ
boxylate side chain of Asp887. To test this hypothesis, the Xꢀ
ray coꢀcrystal structure of compound 34 bound to SOS1 in the
RAS:SOS1:RAS ternary complex was obtained (Figure 3). In
the crystallized binding pose, the exocyclic nitrogen atom at
the 4ꢀposition of the quinazoline nucleus interacts with Glu902
via a molecule of water. A second molecule of water bridges
the exocylic amine at the 2ꢀposition and the endocylic Nꢀ1
nitrogen atom to Asp887. As designed, the secondary amine of
the tetrahydropyridine moiety at the 8ꢀposition of the quinazoꢀ
line interacts directly with the carboxylate side chain of
Asp887 via a chargeꢀassisted hydrogen bond.23ꢀ25 This addiꢀ
tional interaction appears to improve both potency and efficaꢀ
cy in vitro, as evidenced by comparing the nucleotide exꢀ
change activity of compound 34 with that of the parent 8Hꢀ
quinazoline, 22, and the dihydropyran analogue, 41.
Having improved the biochemical activity of these quinazoꢀ
line compounds using structureꢀbased design, we sought to
assess compoundꢀmediated effects in HeLa cells. We have
previously suggested that rapid and robust activation of RAS
signaling to an intolerably high threshold may elicit antiꢀ
cancer effects.14ꢀ17 In our proposed biological mechanism,
SOS1 agonist compounds induce (1) an increase in the levels
of RASꢀGTP and (2) biphasic modulation of ERK1/2 phosꢀ
At present, the discrepancy between RASꢀGTP and
pERK1/2T202/Y204 induction after treatment with compound 22
is not fully understood, but may be indicative of either insuffiꢀ
cient cellular potency or offꢀtarget effects resulting in
pERK1/2T202/Y204 activation independent of RASꢀGTP inducꢀ
tion. It is notable that, although compound 22 was found to
weakly inhibit ERK1/2 kinase activity (Table S3), this comꢀ
pound elicits enhanced pERK1/2T202/Y204 levels in cells.
4
ACS Paragon Plus Environment

Page 5 of 7
ACS Medicinal Chemistry Letters
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 4. RASꢀGTP and corresponding pERK1/2T202/Y204 levels from HeLa cells that were treated for 30 min with up to 100 ꢁM of comꢀ
pound 22, 30, 32, or 34. EGF treatment (50 ng/mL for 5 min) was used as a positive control for pathway activation. Quantification values
for RASꢀGTP and pERK1/2T202/Y204 levels are displayed under the respective blots for each compound and are also provided in Table S1.
Data are representative of three independent experiments.
In this paper, an SAR study involving the structureꢀguided
General experimental considerations, detailed experimental proꢀ
cedures, and characterization data for all new compounds; bioꢀ
chemical and cellular assay conditions; quantification tables for
western blotting experiments; western blotting experiments inꢀ
volving compounds 13, 25, 27, 28, and 29; details regarding proꢀ
tein expression and purification; details regarding crystallization,
Xꢀray data collection, structure solution, and refinement; a table
containing Xꢀray data collection and refinement statistics; and a
Reaction Biology Corp. kinase profiling report for compound 22.
(PDF)
design and synthesis of a series of quinazolines is described.
These molecules bind to SOS1 in the same pocket as the preꢀ
viously reported indoles, but in a distinctive manner, which is
characterized by the rotation of Phe890. This rotation revealed
a new area of chemical space and provided a starting point for
the structureꢀbased design of improved compounds. By exꢀ
ploring different substituents on the aniline ring at the 2ꢀ
position of the quinazoline nucleus, optimizing the branched
alkyl amine at the 4ꢀposition, and tethering an amino group to
the 8ꢀposition, the overall nucleotide exchange activity of the
initial HTS hit, 1, was improved. Lead compounds identified
in the biochemical assay, 32 and 34, were tested in HeLa cells
where they elicited an increase in the levels of RASꢀGTP and
caused biphasic changes in ERK1/2 phosphorylation. These
new findings suggest that the preferred profile for a SOS1
agonist from this quinazoline series should include both a high
Rel. Act. (approximately ≥50%) and a low EC50 (approximateꢀ
ly ≤2.5 ꢁM) to elicit potent activation of RAS and biphasic
modulation of ERK1/2 phosphorylation in HeLa cells.
Accession Codes
Compound 1 (PDB ID code 6CUO)
Compound 22 (PDB ID code 6CUP)
Compound 34 (PDB ID code 6CUR)
Authors will release the atomic coordinates and experimental data
upon article publication.
SOS1 agonist compounds from two distinct chemical seꢀ
ries—indole and quinazoline—elicit the same biochemical and
cellular effects that we propose are characteristic of SOS1
activation. In particular, the biphasic modulation of ERK1/2
phosphorylation elicited by compounds from these two chemiꢀ
cally unrelated series further illustrates and supports our reꢀ
cently proposed biological mechanism.15ꢀ17 Importantly, these
studies show that compounds that bind to SOS1 with the
Phe890 residue in the alternative conformation described here
also activate nucleotide exchange. Based on our hypothesis
that rapid and robust activation of RAS signaling may elicit
antiꢀcancer effects, current efforts in this series are focused on
improving compound potency in the biochemical and cellular
assays, and on identifying potential sources of offꢀtarget activꢀ
ity. The goal of these future studies is to identify a tool comꢀ
pound that can be used to explore the in vivo consequences of
modulating RAS signaling through SOS1 in cancer cells.
AUTHOR INFORMATION
Corresponding Author
* Eꢀmail: stephen.fesik@vanderbilt.edu.
Phone: +1 (615) 322ꢀ6303
Fax: +1 (615) 875ꢀ3236
Present Addresses
†P.A.P.: PharmAgra Labs, Brevard, NC
‡J.P.K.: Truetiva, Inc., Monterey, CA
⊥M.C.B.: Northwestern University Feinberg School of Medicine,
Chicago, IL
#C.F.B.: Wright Medical Group N.V., Franklin, TN
¶Q.S.: AbbVie, Inc., North Chicago, IL
£O.W.R.: The Institute of Cancer Research, London, UK
Author Contributions
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the manuꢀ
script.
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website.
Funding Sources
5
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 6 of 7
Funding for this work came from the following sources:
12. Ostrem, J. M. L.; Shokat, K. M. Direct SmallꢀMolecule
Inhibitors of KRAS: From Structural Insights to Mechanismꢀ
Based Design. Nat. Rev. Drug Discovery 2016, 15 (11), 771–785.
13. Wilson, C. Y.; Tolias, P. Recent Advances in Cancer Drug
Discovery Targeting RAS. Drug Discovery Today 2016, 21 (12),
1915–1919.
U. S. National Institutes of Health, NIH Director’s Pioneer Award
(DP1OD006933/DP1CA174419) to S. W. Fesik
Lustgarten Foundation Research Investigator Grant to S. W. Fesik
National Cancer Institute SPORE Grant in GI Cancer
(5P50A095103ꢀ09) to R. J. Coffey
1
2
3
4
5
6
7
8
14. Burns, M. C.; Sun, Q.; Daniels, R. N.; Camper, D.; Kennedy,
J. P.; Phan, J.; Olejniczak, E. T.; Lee, T.; Waterson, A. G.;
Rossanese, O. W.; Fesik, S. W. Approach for Targeting Ras With
Small Molecules That Activate SOSꢀMediated Nucleotide
Exchange. Proc. Natl. Acad. Sci. U.S.A. 2014, 111 (9), 3401–
3406.
15. Howes, J. E.; Akan, D. T.; Burns, M. C.; Rossanese, O. W.;
Waterson, A. G.; Fesik, S. W. Small MoleculeꢀMediated
Activation of RAS Elicits Biphasic Modulation of PhosphoꢀERK
Levels That Are Regulated Through Negative Feedback on SOS1.
Mol. Cancer Ther. 2018, 17 (5), 1051–1060.
16. Abbott, J. R.; Hodges, T. R.; Daniels, R. N.; Patel, P. A.;
Kennedy, J. P.; Howes, J. E.; Akan, D. T.; Burns, M. C.; Sai, J.;
Sobolik, T.; Beesetty, Y.; Lee, T.; Rossanese, O. W.; Phan, J.;
Waterson, A. G.; Fesik, S. W. Discovery of Aminopiperidine
Indoles That Activate the Guanine Nucleotide Exchange Factor
SOS1 and Modulate RAS Signaling. J. Med. Chem. 2018, 61 (14),
6002–6017.
17. Burns, M. C.; Howes, J. E.; Sun, Q.; Little, A. J.; Camper, D.
V.; Abbott, J. R.; Phan, J.; Lee, T.; Waterson, A. G.; Rossanese,
O. W.; Fesik, S. W. HighꢀThroughput Screening Identifies Small
Molecules That Bind to the RAS:SOS:RAS Complex and Perturb
RAS Signaling. Anal. Biochem. 2018, 548, 44–52.
18. Winter, J. J. G.; Anderson, M.; Blades, K.; Brassington, C.;
Breeze, A. L.; Chresta, C.; Embrey, K.; Fairley, G.; Faulder, P.;
Finlay, M. R. V.; Kettle, J. G.; Nowak, T.; Overman, R.; Patel, S.
J.; Perkins, P.; Spadola, L.; Tart, J.; Tucker, J. A.; Wrigley, G.
Small Molecule Binding Sites on the Ras:SOS Complex Can Be
Exploited for Inhibition of Ras Activation. J. Med. Chem. 2015,
58 (5), 2265–2274.
19. Lovering, F.; Bikker, J.; Humblet, C. Escape from Flatland:
Increasing Saturation as an Approach to Improving Clinical
Success. J. Med. Chem. 2009, 52 (21), 6752–6756.
20. Smith, G. F. Designing Drugs to Avoid Toxicity. In Prog.
Med. Chem.; Lawton, G.; Witty, D. R., Eds.; Elsevier: 2011; Vol.
50, pp 1–47.
21. Peterson, L. A. Reactive Metabolites in the Biotransformation
of Molecules Containing a Furan Ring. Chem. Res. Toxicol. 2013,
26 (1), 6–25.
Notes
RAS activator compounds have been licensed to Boehringer
Ingelheim.
ACKNOWLEDGMENT
The authors would like to thank the following core facilities for
their contributions to this work:
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
Experiments were performed in the Vanderbilt High Throughput
Screening (HTS) Core Facility with assistance provided by C.
David Weaver, Paige Vinson, Chris Farmer, and Corbin Whitꢀ
well. The HTS Core receives support from the Vanderbilt Institute
of Chemical Biology and the Vanderbilt Ingram Cancer Center
(P30 CA68485).
The Vanderbilt University Biomolecular NMR Facility for inꢀ
strumentation. This facility receives support from an NIH SIG
Grant (1Sꢀ10RR025677ꢀ01) and Vanderbilt University matching
funds.
The U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences for use of the Advanced Photon Source
(Contract: DEꢀAC02ꢀ06CH11357).
ABBREVIATIONS
EGF, epidermal growth factor; ERK1/2, extracellular regulated
kinases 1 and 2; GAP, GTPaseꢀactivating protein; GEF, guanine
nucleotide exchange factor; EC50, half maximal effective concenꢀ
tration; HTS, high throughput screen/screening; MAPK, mitogenꢀ
activated protein kinase; pERK1/2, phosphorylated ERK1/2;
SOS1, son of sevenless homologue 1
REFERENCES
1. Papke, B.; Der, C. J. Drugging RAS: Know the Enemy.
Science 2017, 355 (6330), 1158–1163.
2. Prior, I. A.; Lewis, P. D.; Mattos, C. A Comprehensive Survey
of Ras Mutations in Cancer. Cancer Res. 2012, 72 (10), 2457–
2467.
3. Cox, A. D.; Fesik, S. W.; Kimmelman, A. C.; Luo, J.; Der, C.
J. Drugging the Undruggable RAS: Mission Possible? Nat. Rev.
Drug Discovery 2014, 13 (11), 828–851.
4. Pulgar, T. G. D.; Lacal, J. C. GTPase. In Encyclopedia of
Cancer; Schwab, M., Ed.; Springer Berlin Heidelberg: Berlin,
Heidelberg, 2011; pp 1609–1613.
5. Cox, A. D.; Der, C. J. Ras History. Small GTPases 2010, 1 (1),
2–27.
6. Vigil, D.; Cherfils, J.; Rossman, K. L.; Der, C. J. Ras
Superfamily GEFs and GAPs: Validated and Tractable Targets for
Cancer Therapy? Nat. Rev. Cancer 2010, 10 (12), 842–857.
7. Hanahan, D.; Weinberg, R. A. The Hallmarks of Cancer. Cell
2000, 100 (1), 57–70.
8. Hanahan, D.; Weinberg, R. A. Hallmarks of Cancer: The Next
Generation. Cell 2011, 144 (5), 646–674.
9. PylayevaꢀGupta, Y.; Grabocka, E.; BarꢀSagi, D. RAS
Oncogenes: Weaving a Tumorigenic Web. Nat. Rev. Cancer
2011, 11 (11), 761–774.
10. Sacco, E.; Spinelli, M.; Vanoni, M. Approaches to Ras
Signaling Modulation and Treatment of RasꢀDependent
Disorders: A Patent Review (2007–Present). Expert Opin. Ther.
Pat. 2012, 22 (11), 1263–1287.
11. Singh, H.; Longo, D. L.; Chabner, B. A. Improving Prospects
for Targeting RAS. J. Clin. Oncol. 2015, 33 (31), 3650–3659.
22. Burns, M. C. Small Molecule Modulation of the RasꢀSOS
Interaction in Cancer. Dissertation, Vanderbilt University,
Nashville, TN, 2014.
23. Gilli, G.; Gilli, P. Towards an Unified HydrogenꢀBond
Theory. J. Mol. Struct. 2000, 552 (1), 1–15.
24. MeotꢀNer, M. The Ionic Hydrogen Bond. Chem. Rev. 2005,
105 (1), 213–284.
25. Gilli, P.; Gilli, G. Hydrogen Bond Models and Theories: The
Dual Hydrogen Bond Model and Its Consequences. J. Mol. Struct.
2010, 972 (1), 2–10.
6
ACS Paragon Plus Environment

Page 7 of 7
ACS Medicinal Chemistry Letters
TABLE OF CONTENTS GRAPHIC
1
2
Structure-based design
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
7
ACS Paragon Plus Environment