Discovery and Structure-Based Optimization of Benzimidazole-Derived Activators of SOS1-Mediated Nucleotide Exchange on RAS

Article abstract of DOI:10.1021/acs.jmedchem.8b01108

Son of sevenless homologue 1 (SOS1) is a guanine nucleotide exchange factor that catalyzes the exchange of GDP for GTP on RAS. In its active form, GTP-bound RAS is responsible for numerous critical cellular processes. Aberrant RAS activity is involved in ～30% of all human cancers; hence, SOS1 is an attractive therapeutic target for its role in modulating RAS activation. Here, we describe a new series of benzimidazole-derived SOS1 agonists. Using structure-guided design, we discovered small molecules that increase nucleotide exchange on RAS in vitro at submicromolar concentrations, bind to SOS1 with low double-digit nanomolar affinity, rapidly enhance cellular RAS-GTP levels, and invoke biphasic signaling changes in phosphorylation of ERK 1/2. These compounds represent the most potent series of SOS1 agonists reported to date.

Full text of DOI:10.1021/acs.jmedchem.8b01108

Subscriber access provided by University of South Dakota
Article
Discovery and Structure-Based Optimization of Benzimidazole-
Derived Activators of SOS1-Mediated Nucleotide Exchange on RAS
Timothy Robert Hodges, Jason R. Abbott, Andrew J. Little, Dhruba Sarkar, James M. Salovich,
Jennifer Howes, Denis T. Akan, Jiqing Sai, Allison L. Arnold, Carrie Browning, Michael C. Burns,
Tammy Sobolik, Qi Sun, Yugandhar Beesetty, Jesse Coker, Dirk Scharn, Heinz Stadtmueller,
Olivia W. Rossanese, Jason Phan, Alex G. Waterson, Darryl B. McConnell, and Stephen W. Fesik
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01108 • Publication Date (Web): 11 Sep 2018
Downloaded from http://pubs.acs.org on September 11, 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 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biochemistry
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discovery and Structure-Based Optimization of Benzimidazole-Derived Activators of SOS1-
Mediated Nucleotide Exchange on RAS
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Timothy R. Hodges,◊,† Jason R. Abbott,◊,† Andrew J. Little,◊,†, ₣ Dhruba Sarkar, James M. Salovich, Jennifer
◊
◊
◊
◊
◊
◊
◊
◊,₵
E. Howes, Denis T. Akan, Jiqing Sai, Allison L. Arnold, Carrie Browning, Michael C. Burns, Tammy
Sobolik,◊,Ω Qi Sun,◊, Ⱡ Yugandhar Beesetty, Jesse A. Coker, Dirk Scharn, Heinz Stadtmueller, Olivia W.
◊
◊, ♯
‡
‡
Rossanese,◊,₽ Jason Phan, Alex G. Waterson, Darryl B. McConnell, and Stephen W. Fesik*
◊
§,¥
‡
,◊,§,¥
◊
§
Department of Biochemistry, Department of Pharmacology, Vanderbilt University School of Medicine,
Nashville, Tennessee 37232-0146, USA
¥Department of Chemistry, Vanderbilt University, Nashville, Tennessee 37232-0146, USA
‡Boehringer Ingelheim RCV GmbH & Co KG, Doktor-Boehringer-Gasse 5-11, 1120 Vienna, Austria
KEYWORDS: RAS, SOS1, Agonist, Cancer.
ABSTRACT: Son of sevenless homologue 1 (SOS1) is a guanine nucleotide exchange factor that catalyzes
the exchange of GDP for GTP on RAS. In its active form, GTP-bound RAS is responsible for numerous
critical cellular processes. Aberrant RAS activity is involved in ~30% of all human cancers; hence, SOS1 is
an attractive therapeutic target for its role in modulating RAS activation. Here, we describe a new series of
benzimidazole-derived SOS1 agonists. Using structure-guided design, we discovered small molecules that
increase nucleotide exchange on RAS in vitro at sub-micromolar concentrations, bind to SOS1 with low
double digit nanomolar affinity, rapidly enhance cellular RAS-GTP levels, and invoke biphasic signaling
changes in phosphorylation of ERK 1/2. These compounds represent the most potent series of SOS1
agonists reported to date.
1
ACS Paragon Plus Environment

Page 3 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
■
INTRODUCTION
RAS proteins act as molecular switches, cycling between active (GTP-bound) and inactive (GDP-
bound) states in order to regulate multiple intracellular signaling pathways that control cell proliferation,
growth, differentiation, and survival.1,2 Deregulation of these pathways via oncogenic RAS mutations can
result in cells displaying many of the hallmarks of cancer and is implicated in the oncogenesis and
progression of ~30% of all malignant tumors.3–7 Therefore, therapeutic targeting of oncogenic RAS has
been an important and long-standing goal in modern drug discovery. Unfortunately, only limited success
has been realized when directly targeting RAS due to the lack of chemically tractable pockets on the protein
surface.8–10 The recalcitrant nature of these proteins toward traditional drug discovery strategies has
inspired the pursuit of alternative approaches to inhibit RAS signaling.11–19 One strategy is to indirectly target
RAS through therapeutic modulation of the regulatory proteins that control its activity. The conversion of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
RAS-GDP to RAS-GTP is the rate-limiting step in the activation of RAS signaling and is controlled by
guanine nucleotide exchange factors (GEFs) such as son of sevenless homologue 1 (SOS1).7,20,21 Due to
its importance in RAS signaling, SOS1 has emerged as an attractive therapeutic target for the treatment of
RAS-driven cancers.22,23
Our group recently reported two unrelated chemical series that bind to the same hydrophobic
pocket on SOS1 as part of a RAS:SOS1:RAS complex.24–26 We have previously shown that these
compounds activate nucleotide exchange on RAS, resulting in increased levels of cellular RAS-GTP and,
at low concentrations, enhance phosphorylation of extracellular regulared kinase (ERK) 1/2 via the RAS-
MAPK signaling pathway.27 At higher doses of compound, phosphorylated ERK (pERK) phosphorylates
SOS1 leading to disruption of the SOS1-GRB2 complex and dissociation of RAS from SOS1, thereby
achieving a compound-mediated negative feedback loop that results in reduced ERK phosphorylation.
Although seemingly counterintuitive, a growing body of evidence suggest that activation of RAS may be a
viable therapeutic strategy.28–30 The “Goldilocks principle” suggests that cells can only tolerate changes in
signaling that stay within certain parameters, which do not go either above, or below a certain cellular
threshold.31 For example, enhancement of RAS-GTP levels in the cell by ectopic overexpression of mutant
RAS can lead to pronounced anti-tumoral responses.32 Furthermore, naturally occurring concomitant
expression of mutant RAS with either mutant BRAF or mutant EGFR is very rare, and overexpression of
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RAS is synthetically lethal in the context of an EGFR or BRAF mutation.33–35 Thus, the RAS activation that
we observe with SOS1 agonists may represent a promising approach for targeting cells with underlying
oncogenic RAS, BRAF, or EGFR mutations. Here, we describe the structure-guided design, synthesis, and
structure–activity relationships (SAR) of a new benzimidazole-derived series of SOS1 ligands. These efforts
led to the discovery of molecules that robustly activate SOS1-mediated nucleotide exchange in vitro, bind
with high affinity to SOS1, rapidly increase cellular RAS-GTP levels, and modulate ERK phosphorylation in
cells, resulting in the most potent SOS1 agonists yet reported.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
■
RESULTS AND DISCUSSION
Discovery of Benzimidazole-Derived SOS1 Compounds.
In order to identify compounds that bind to SOS1 and affect SOS1-mediated nucleotide exchange
in vitro, we conducted a high throughput screen (HTS) of the Vanderbilt Compound Collection.36 This screen
revealed multiple molecules in different chemical series that elicited robust and reproducible SOS1-
mediated activation of nucleotide exchange on RAS. Similar to our previously reported work,24–26,36
compound efficacy is described by relative percent activation (Act.), which is the rate of nucleotide
exchange at the maximum compound dose of 100 µM, relative to the previously reported indole-derived
control compound at the same concentration in the same run of the assay.25 The corresponding potency is
described as the half maximal effective concentration (EC50) determined from the dose response curve (0–
100 µM of compound) generated from the nucleotide exchange assay. From the HTS hits, several
benzimidazole-containing molecules emerged as potentially useful starting points. For example,
compounds 1 and 2 were found to activate nucleotide exchange with EC50 values of 63.0 µM and 23.2 µM,
respectively (Figure 1A). An early follow-up analog, 3, was synthesized that removed the ring constraints
from the 2-position nitrogen atom and introduced a substitution pattern on the benzyl group similar to that
of an aniline ring system in a previously reported anilinoquinazoline series.26 These structural modifications
afforded a 3-fold improvement in nucleotide exchange activation EC50 when comparing derivative 3 to 2.
Furthermore, an X-ray co-crystal structure obtained of 3 bound to the RAS:SOS1:RAS ternary complex
confirmed that it binds in the same hydrophobic pocket of the CDC25 domain of SOS1, adjacent to the
Switch II region of RAS, that is utilized by previously reported series of SOS1 ligands (Figure 1B).23–26,36
3
ACS Paragon Plus Environment

Page 5 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. (A) Early hit benzimidazole-derived activators of SOS1-mediated nucleotide exchange. (B) X-ray co-crystal
structure of 3 (yellow; PDB ID code 6D5V) bound to SOS1 in the RAS:SOS1:RAS ternary complex. The RAS and
SOS1 protein surfaces are colored yellow and grey, respectively. Areas identified for SAR development are highlighted,
including: (i) a water-mediated interaction between the 2-amine functionality with Asp887 (blue oval), (ii) occupation of
a hydrophobic subpocket beneath Phe890 (red arrows), (iii) occupation of a hydrophobic region along the wall of the
CDC25 domain (pink rectangle), (iv) occupation of a hydrophobic subpocket behind Glu902 and adjacent to Phe890
(green circle), and (v) interaction with proximal residues of SOS1 from the 4-position of the benzimidazole scaffold
(white arrows).
C-2 Substituent SAR.
Analysis of the X-ray co-crystal structure of compound 3 revealed several opportunities to improve
binding to SOS1 (Figure 1B). We initiated our SAR studies by focusing on the 2-position of the
benzimidazole core with the goal of deploying a Lewis base closer to Asp887 in an attempt to establish a
direct, rather than water-mediated, interaction (Table 1). Notably, improved potency and efficacy were
observed with several 2-amine variants. This was highlighted by 2-piperazino 10, which demonstrated a
~4-fold and ~3-fold improvement in nucleotide exchange EC50 and relative percent activation, respectively.
Conversely, negligible nucleotide exchange activation was observed when the Lewis base was removed,
as exemplified by 2H-benzimidazole 4. The X-ray co-crystal structure of compound 10 (Figure 2A) revealed
that the piperazine moiety occluded the water molecule that facilitated the hydrogen bonding interaction
between Asp887 and 3. As a result, the piperazine nitrogen atom of 10 achieved a direct charge–charge
interaction with the carboxylate side chain of Asp887. Based on its combination of potent and efficacious
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
nucleotide exchange activity and limited number of hydrogen bond donors,37 piperazino-analog 10 was
selected as a template for continued structural modification.
Table 1. 2-Amino Substituents
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
a
Compd
R
EC50 (µM) Act. (%)
3
4
5
6
7
8
9
NH
2
8.1 ± 6.83
79 ± 8.0
--
b
H
--
37.1 ± 12.57 129 ± 1.8
26.0 ± 0.80 143 ± 18.9
8.5 ± 2.06 176 ± 10.1
6.6 ± 0.05 114 ± 6.3
5.6 ± 0.11 128 ± 0.6
2.2 ± 0.52 211 ± 20.5
1.9 ± 0.62 114 ± 12.2
1
1
0
1
Measurements are reported as the mean ± SD of two or more independent experiments, each conducted in duplicate.
aActivation values represent percentage relative to control.25 b “--“ denotes an EC50 value of >100 μM, an EC50 value
that was not calculated due to low efficacy, or an activation value that was not determined due to low efficacy.
Benzyl Substituent SAR.
With improvements in potency and efficacy achieved via the inclusion of the 2-piperazine moiety,
we turned our focus to investigating the chemical space around the benzyl group (Table 2). While ortho-
substitutions were generally poorly tolerated (not shown), the contribution of the aromatic halogen atoms
in other positions was quite evident, as unsubstituted benzyl analog 12 showed an EC50 that was ~4-fold
worse than 10. Interestingly, of the 4’-monosubstituted derivatives, only the 4’-fluoro compound 13
demonstrated improved nucleotide exchange potency relative to 12, and the activity dropped off sharply
with larger 4’-moieties. The benefit of meta-substitution also appeared to be limited by size as evidenced
5
ACS Paragon Plus Environment

Page 7 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
by the drop in potency observed for compounds featuring a 3’-trifluoromethyl (19) or a 3’-cyclopropyl (20)
group relative to 10. Together, these data suggested that limited steric bulk at the para- and meta-positions
provided improved potency. We noted that compound potency (EC50) was observed to correlate poorly with
compound efficacy (Act.). Empirically, it had been previously determined that SOS1 agonists achieving Act.
values greater than ~50% elicited more robust increases in cellular RAS–GTP levels.36 Beyond this
perceived minimum threshold, the origin and consequences of varying degrees of percent relative activation
is not well understood and is currently under further investigation.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Benzyl Substituents
3
’
4’
5’
a
Compd
10
R
R
R
EC50 (µM) Act. (%)
Cl
H
H
H
H
H
F
F
H
2.2 ± 0.52 211 ± 20.5
8.6 ± 1.64 122 ± 17.8
5.9 ± 4.88 130 ± 28.1
12.4 ± 2.70 164 ± 15.0
36.6 ± 17.26 93 ± 13.0
19.6 ± 5.01 148 ± 30.7
5.9 ± 1.33 183 ± 25.3
4.1 ± 2.00 190 ± 20.6
8.1 ± 0.52 263 ± 52.5
11.7 ± 2.74 151 ± 5.0
3.7 ± 0.29 270 ± 73.8
12
13
14
15
16
H
F
H
H
Cl
Me
H
H
CF
3
H
17
18
F
H
Me
F
H
19
20
21
CF
3
F
H
F
H
Me
F
Me
Measurements are reported as the mean ± SD of two or more independent experiments, each conducted in duplicate.
aActivation values represent percentage relative to control.25
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Arene 5,6-Substitution SAR.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. (A) X-ray co-crystal structure of 10 (magenta; PDB ID code 6D5M) bound to SOS1 in the RAS:SOS1:RAS
ternary complex. The curved blue line highlights the subpocket near His905. The RAS and SOS1 protein surfaces are
colored yellow and grey, respectively. (B) 28 (green; PDB ID code 6D5L) overlaid with 10. The green sphere represents
space filling of the 6-chlorine atom in 28, whereas the magenta and white spheres represent space filling of the 6-
methyl group in 10.
Having learned much about the tolerance of substitution around the benzyl group, we transitioned
to examining various hydrophobic substitutions at the 5- and 6-positions of the benzimidazole core to
optimize the occupancy of the small hydrophobic subpocket near His905 (Figure 2A). For this, the 2-
piperazino and 3’-chloro-4’-fluoro benzyl moieties were held constant to isolate the effects of the 5- and 6-
positions of the benzimidazole nucleus (Table 3). Removal of the 5-position substituent had a relatively
minor effect on compound potency. Compare, for example, the 5H-6-methyl derivative, 24, and the 5H-6-
chloro derivative, 28, to their respective 5-methyl analogs 10 and 22. It is interesting to note however, that
5H-substitution resulted in a strong decrease in relative percent activation. The X-ray co-crystal structure
of 28 overlaid with that of 10 (Figure 2B) showed that both ligands rest in the hydrophobic subpocket near
His905 and effected minimal differences in the positioning of the SOS1 binding pocket residues. This
suggested that the differences in observed Act. levels between analogs 10 and 28 are not due to a
significant discrepancy in compound binding mode. From these same co-crystal structures, it is apparent
that there is little room in the hydrophobic subpocket near His905 for 6-position moieties much larger than
a methyl group. This observation is supported by the substantial loss in potency observed for the 6-
cyclopropyl compound 26. Notably, removal of both 5- and 6-substitution, exemplified by compound 23,
7
ACS Paragon Plus Environment

Page 9 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
resulted in a ~5-fold reduction in potency relative to 28. These data suggested that an appropriately sized
6-substituent was the primary contributor to potency in this region of the binding pocket.
Table 3. 5,6-Substituents
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
6
a
Compd
R
R
EC50 (µM) Act. (%)
1
2
2
2
2
0
2
3
4
5
Me
Me
H
Me
Cl
2.2 ± 0.52 211 ± 20.5
3.0 ± 0.56 200 ± 13.3
5.4 ± 1.22 109 ± 4.2
2.4 ± 0.33 106 ± 13.1
4.4 ± 0.54 78 ± 2.2
24.3 ± 5.33 65 ± 2.2
H
H
Me
H
CF
3
26
27
H
H
Br
Cl
1.3 ± 0.3
115 ± 5.4
2
8
H
1.0 ± 0.16 109 ± 1.0
Measurements are reported as the mean ± SD of two or more independent experiments, each conducted in duplicate.
aActivation values represent percentage relative to control.25
The SAR developed in Tables 2 and 3 prompted the synthesis of several matched pairs of analogs
to investigate the interplay between substituents on the aromatic ring of the benzyl moiety and the
substituents at the 5- and 6-positions of the benzimidazole nucleus. From this effort, crossover analog 29—
containing the 3’,5’-dimethyl-4’-fluorobenzyl moiety and a 6-position chlorine atom—was found to exhibit
an EC50 value below 1 µM in the nucleotide exchange assay (Figure 3A). As compounds continued to
improve, we observed a loss of SAR texture for derivatives with EC50 values in this range. To effectively
differentiate more potent compounds, a fluorescence polarization anisotropy (FPA) assay, based on the
displacement of a labeled probe from SOS1 (Supplementary Figure S2), was developed to provide a
measurement of binding affinity (Κ
d
) to SOS1.38,39 Several earlier compounds were re-evaluated using this
FPA assay, showing a general agreement between Κ
d
and nucleotide exchange EC50 for previously
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
investigated SAR trends (Supplementary Figure S3 and Table S1). This reaffirmed our established SAR
and prompted a switch to the FPA assay as the primary tool for evaluating analogs. Moving forward with a
new biochemical assay to drive optimization, compound 29 was selected as a template upon which to
conduct further work.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. (A) Nucleotide exchange and binding affinity data for compound 29. (B) X-ray co-crystal structure of
compound 29 (teal; PDB ID code 6D5J) bound to SOS1 in the RAS:SOS1:RAS ternary complex. The red dotted
spheres represent molecules of water and the space they occupy. The RAS and SOS1 protein surfaces are colored
yellow and grey, respectively.
C-4 Substituent SAR.
The binding mode of compound 29 (Figure 3B) highlighted several opportunities for continued
optimization. First, a largely hydrophobic space proximal to the 4-position of the benzimidazole remained
unoccupied, as demonstrated by the two molecules of water occupying the back of the binding pocket.
Additionally, several residues at the lip of the binding pocket, primarily Glu902, His905, and Tyr884, seemed
accessible for interaction. As these regions are largely solvent-exposed, a considerable tolerance for
groups that may offer the ability to tune the overall molecular properties was expected. To exploit these
opportunities, we explored a wide range of chemical diversity at the 4-position. A selection of these results
is shown in Table 4.
In an effort to fill the neighboring hydrophobic subpocket, several biaryl derivatives with lipophilic
substitutions were synthesized and evaluated by FPA (Table 4). Multiple 4-phenyl analogs demonstrated
improved binding affinity relative to the 4H-parent 29. In some instances, ortho-substitution of the 4-phenyl
9
ACS Paragon Plus Environment

Page 11 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ring offered further enhancement—best exemplified by the ortho-trifluoromethyl analog 35 and the ortho-
chloro compound 38—relative to the unsubstituted phenyl derivative 30. Increasing the size of the ortho-
alkyl groups resulted in a decline in binding affinity from 30, suggesting that the nearby pocket has a limited
size tolerance. Notably, biaryl analogs, such as 36 and 37, bearing meta- or para-substituents exhibited
diminished affinity when compared to their ortho-substituted or unsubstituted comparators.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Analysis of the X-ray co-crystal structure of 38 revealed that the aryl group is rotated out of plane
from the benzimidazole scaffold (Figure 4A). Further, when the co-crystal structure of 38 was overlaid with
that of 29 (Figure 4D), it was evident that 38 leaned over into the SOS1 binding pocket to drive the ortho-
chlorine atom down into the available lipophilic space, thus displacing the previously bound waters from
this subpocket. The presence of the hydrophobic aryl group in 38 disrupted the positioning of Glu902,
rotating the carboxylic acid side chain upward toward solvent compared to the downward pose observed in
the co-crystal structure of 29. The observed gains in binding affinity from the addition of phenyl groups at
the 4-position corresponded with an increase in compound lipophilicity as evidenced by a rise in ALogP
and correspondingly poor lipophilic ligand efficiency (LipE)40 when compared with the 4H-compound, 29.
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
Table 4. C-4 SAR
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
4
a
4
b
4
b
Compd
R
Κ d (µM)
AlogP
LipE Compd
R
Κ d (µM)
AlogP
LipE Compd
R
Κ d (µM)
AlogP
LipE
29
30
31
H
0.46 ± 0.078
0.27 ± 0.005
5.10
1.23
-0.06
-0.44
39
40
41
0.30 ± 0.078
0.20 ± 0.003
0.13 ± 0.030
7.59
-1.06
0.51
1.95
49
50
51
0.14 ± 0.009
0.24 ± 0.056
6.02
0.85
6.62
7.11
6.18
4.95
5.47
5.47
1.16
1.01
b
a
0.21 ± 0.010
0.33 ± 0.027
32
33
34
0.21 ± 0.029
7.56
7.82
7.90
-0.88
-1.41
-1.54
42
43
44
0.14 ± 0.013
5.22
5.14
4.41
1.65
1.46
2.21
52
53
54
0.45 ± 0.104
5.43
4.75
4.84
0.92
1.66
1.61
a
0.39 ± 0.093
0.44 ± 0.139
0.25 ± 0.057
0.24 ± 0.083
0.39 ± 0.064
0.35 ± 0.001
a
35
0.15 ± 0.048
0.67 ± 0.007
7.56
7.56
-0.74
-1.39
45
46
0.31 ± 0.083
0.38 ± 0.025
5.22
5.22
1.30
1.20
55
56
0.11 ± 0.018
0.10 ± 0.011
5.12
6.25
1.82
0.74
36
37
38
1.05 ± 0.056
0.12 ± 0.034
7.56
7.29
-1.59
-0.43
47
48
0.14 ± 0.009
0.24 ± 0.036
4.93
4.93
1.92
1.69
57
58
0.18 ± 0.081
<0.10
5.06
5.49
1.68
c
--
a
Measurements are reported as the mean ± SD of two or more independent experiments, each conducted in duplicate. Evaluated through displacement of probe S1
(Supplementary Figure S21). Evaluated through displacement of probe S2 (Supplementary Figure S2). LipE was not determined.
b
c
11
ACS Paragon Plus Environment

Page 13 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We hypothesized that directly engaging residues Glu902, His905, and Tyr884 with heteroatoms
might provide significant affinity enhancements. Furthermore, successful incorporation of polar functionality
at the 4-position was expected to result in improved solubility and lower lipophilicity relative to the biaryl
compounds. We thus explored the use of a protonatable nitrogen atom to target the carboxylate of Glu902
with a potential charge–charge interaction. As the data in Table 4 illustrate, amine substitutions at the 4-
position were well-tolerated and generally demonstrated lower lipophilicity. This is best exemplified by the
N-methylpiperazine analog, 42, which demonstrated a greater than 3-fold improvement over the 4H-
benzimidazole predecessor 29. Notably, 42 had a similar binding affinity to that of 38, but with much
improved AlogP and LipE values. The X-ray co-crystal structure of the tetrahydropyridine analog 43 (Figure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4B) revealed that Glu902 was rotated down toward the nitrogen atom of 43 compared to the position of the
same residue in the X-ray co-crystal structure of 38 (Figure 4A). In contrast to 29 and 38, compound 43
leaned toward the RAS:SOS1 interface to accommodate the positioning of Glu902 (Figure 4D), confirming
that the tetrahydropyridine moiety of 43 did indeed interact with the carboxylate of Glu902. However,
replacement of the basic nitrogen atom in the piperazine moiety with a methylene group (40) or oxygen
atom (41) proved to have little effect on binding affinity, suggesting that the charge–charge interaction with
Glu902 offered minimal benefit to binding.
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. X-ray co-crystal structures of (A) 38 (salmon; PDB ID code 6D5H), (B) 43 (orange; PDB ID code 6D5G), (C)
47 (green; PDB ID code 6D5E), and (D) 38, 43, and 47 overlaid with 29 (teal) bound to SOS1 in the RAS:SOS1:RAS
ternary complex. The RAS and SOS1 protein surfaces are colored yellow and grey, respectively.
Since engaging Glu902 from the 4-position of the benzimidazole scaffold improved lipophilicity, but
provided only limited binding affinity improvements over 4-aryl analogs, the focus turned toward creating a
vector for a heteroatom to engage His905 or Tyr884 via a hydrogen bond, as exemplified by compounds
45–49 (Table 4). These entities also proved to be widely tolerated, affording binding affinities similar to that
of other 4-substituents highlighted thus far. The X-ray co-crystal structure for 47 revealed that the extended
nitrogen atom did not engage His905, but was orientated toward Tyr884 instead (Figure 4C). Interaction
with the phenol of Tyr884 caused compound 47 to lean toward the RAS:SOS1 interface, in similar fashion
to 43 (Figure 4D). Notably, Glu902 was rotated upward toward solvent—as it was in the X-ray co-crystal
structure of 38—to accommodate the hydrocarbon portion of the pyrrolidine ring.
Importantly, the examples incorporating polar functionality at the 4-position of the benzimidazole
scaffold were widely tolerated, and resulted in compounds with greatly improved physical metrics relative
13
ACS Paragon Plus Environment

Page 15 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to the highly lipophilic biaryl analogs. It was hypothesized that heteroaromatic substituents at the 4-position
might merge the observed tolerance of polar functionality with occupation of the proximal hydrophobic
subpocket and generate compounds with a favorable combination of potency, physiochemical properties,
and a reduced number of hydrogen bond donors (Table 4). Unsubstituted heterocyclic compounds
generally showed a marginal decrease in binding affinity relative to unsubstituted phenyl congener 30, but
exhibited improved AlogP and LipE values. As was the case in the 4-phenyl series, ortho-substitution
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
provided a marked boost in ligand binding affinity, exemplified by the 3-fold decrease in Κ
d
from pyrimidine
54 to ortho-methyl pyrimidine 55 and ortho-isopropyl pyrimidine 56. Furthermore, disubstituted pyrazole 58
generated an affinity below the reliable detection limit (K
d
<0.10 µM) afforded by the assay using the
fluorescent probe S2 (Supplementary Figure S2).
Inspection of the X-ray co-crystal structure of 58 revealed that the disubstituted pyrazole
functionality was rotated out of plane from the benzimidazole scaffold (Figure 5A). This oriented the
lipophilic methyl group toward the back of the hydrophobic pocket in a comparable fashion to the ortho-
chloro derivative 38. This caused the ligand to lean away from the RAS:SOS1 interface (Figure 5D) and
overrode a potential interaction between the pyrazole nitrogen atoms and Glu902, as evidenced by the
rotation of Glu902 up toward solvent. The ortho-substituted heterocycle subset of compounds consistently
offered the best SOS1 affinities measured to this point. In addition, these compounds possessed improved
physical properties relative to their more lipophilic phenyl counterparts.
C-2 Diazaspiro[3.3]heptane Analogs.
After a thorough structure-guided SAR study, several compounds were discovered to bind to SOS1
with binding affinities near or below 0.10 µM (e.g. 58). We postulated that the numerous modifications to
the benzimidazole scaffold may have affected earlier investigated SAR. Thus, the 2-position was re-
evaluated using several amines in the context of our more advanced benzimidazole derivatives. From these
efforts, a multitude of C-2 diazaspiro[3.3]heptane derivatives were found to have binding affinities below
0
.10 µM,41–43 leading to the synthesis of a new fluorescent probe (S3, Supplementary Figure S2) bearing
this moiety. The FPA competition assay with this probe proved capable of reproducible measurement of
affinities well below 0.10 µM. Indeed, data for these analogs (Table 5) showed up to an order of magnitude
improvement in affinity from their 2-piperazino counterparts.
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5. C-2 Diazaspiro[3.3]heptane Derivatives
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
3'
5'
a
Compd
R
R , R
Κ d (µM)
ALogP LipE
59
60
61
Me
Me
H
0.012 ± 0.0020 6.92
1.02
3.05
2.98
0.016 ± 0.0037 4.76
b
0.18 ± 0.003
3.78
62
63
Me
H
0.012 ± 0.0008 4.53
0.036 ± 0.0055 3.56
3.38
3.88
64
65
Me
Me
0.044 ± 0.0067 4.85
0.009 ± 0.0017 5.12
2.50
2.91
Measurements are reported as the mean ± SD of two or more independent experiments, each conducted in duplicate.
a
b
Evaluated through displacement of fluorescent probe S3 (Supplementary Figure S2). Evaluated through displacement
of fluorescent probe S2 (Supplementary Figure S2).
Analysis of the X-ray co-crystal structure for pyrazole 65 showed a very similar binding pose to that
demonstrated by 58, with Glu902 rotated up and the ligand leaning in the pocket to place the methyl group
of 65 into the hydrophobic subpocket (Figure 5B, 5D). The X-ray co-crystal structure for 64 (Figure 5C)
showed a similar binding pose to 4-tetrahydropyridine 43, with Glu902 rotated down to establish a charge–
charge interaction with the piperazine nitrogen atom and causing the ligand to lean toward the RAS:SOS1
interface (Figure 5D). Since little difference in the binding mode for either 64 or 65 from comparable 2-
piperazine variants was evident, it was speculated that the more rigid C-2-diazaspiro[3.3]heptane moiety
benefited from a lower entropic penalty as a result of binding compared with the more flexible C-2 piperazine
functionality. Additionally, the diazaspiro[3.3]heptane functionality is more basic than its piperazine
counterpart,43 which might potentiate a stronger charge–charge interaction with Asp887.
15
ACS Paragon Plus Environment

Page 17 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In addition to improved binding affinities, compounds featuring the diazaspiro[3.3]heptane moiety
demonstrated improved ALogP and LipE values compared to previously examined SOS1 ligands.
Compound lipophilicity was further reduced by appending an amine to the para-position of the pyrimidine,
as exemplified by 62. Removal of the 3’- and 5’-methyl substituents, as in 63, generated the best ALogP
and LipE metrics for this series of SOS1 agonists, albeit with a binding affinity ~3-fold worse than the parent
molecule 62, as might be expected from the decreased occupancy of the hydrophobic pocket beneath
Phe890. As a result of this extensive SAR campaign, these analogs represent the best binders to SOS1
reported to date, highlighted by pyrazole 65 with a binding affinity of 9 nM.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. X-ray co-crystal structure of (A) 58 (pink; PDB ID code 6D59), (B) 65 (yellow; PDB ID code 6D56), (C) 64
(white; PDB ID code 6D55), and (D) 58, 64, and 65 overlaid with 29 (teal) bound to SOS1 in the RAS:SOS1:RAS
ternary complex. The RAS and SOS1 protein surfaces are colored yellow and grey, respectively.
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Cellular Analysis.
With compounds that bind tightly to SOS1 in hand, we sought to assess whether this series of
benzimidazole-derived molecules also elicited the previously observed, characteristic biphasic ERK
modulation and RAS-GTP activation in cancer cells.25,26 To assess the effects on pERK signaling, an In-
Cell Western (ICW) assay was used to evaluate cellular pERK1/2T202/Y204 levels relative to total ERK1/2
levels in response to treatment with varying concentrations of compounds (Figure 6). Two different cell
lines—HeLa (WT KRAS) and NCI-H727 (G12V KRAS)—were treated with compounds at concentrations
ranging from 1.5 to 100 µM for 30 min. The levels of pERK1/2T202/Y204 were normalized to the levels of total
ERK1/2 protein and the DMSO vehicle control. Accordingly, the ICW EC50 values reported in Table 6
represent the half maximal concentration of compound required to elicit peak induction of pERK1/2T202/Y204
levels. Thus, a lower value is considered indicative of a more potent SOS1 agonist. Guided by our previous
mechanistic observations,27 we hypothesized that potent SOS1 agonists would typically demonstrate
biphasic modulation of pERK1/2T202/Y204. This is characterized by an increase in pERK1/2T202/Y204 levels at
lower doses of compound followed by a decrease below baseline in pERK1/2T202/Y204 levels at higher doses
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Figure 6A). Observing an increase in pERK1/2T202/Y204 levels with no corresponding decrease over the
concentration range of compound tested (Figure 6B) was considered to indicate an insufficiently potent
SOS1 agonist. Alongside EC50 values, Table 6 also indicates whether biphasic modulation of
pERK1/2T202/Y204 was elicited by each derivative. A selection of molecules featuring diverse chemical motifs
and binding affinities, spanning five orders of magnitude, were examined by ICW to evaluate whether any
inferences could be drawn between structural features, binding affinities, and phenotypic cellular
responses.
17
ACS Paragon Plus Environment

Page 19 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Representative ICW data for benzimidazole compounds that elicit (A) biphasic modulation of pERK1/2T202/Y204
protein levels, or (B) only an increase in pERK1/2T202/Y204 protein levels. The green channel of the LI-COR infrared
scanner was used to detect pERK1/2T202/Y204 protein levels. In the same well, the red channel of the LI-COR infrared
scanner was used to detect total ERK1/2 protein levels. The graphs represent the quantified immunofluorescence data,
where pERK1/2T202/Y204 levels were normalized to total ERK1/2 levels from the same well and to the vehicle control
treated well. The vertical black dotted line indicates how EC50 values were determined, and the horizontal black dotted
line indicates baseline pERK1/2T202/Y204 signal. Each ICW experiment was conducted three independent times, and in
duplicate on each plate.
Table 6. In-Cell Western Analysis of pERK1/2T202/Y204 Levels
Compd
47
64
42
62
56
58
55
38
60
59
63
65
29
23
10
3
61
HeLa ICW
3.2
5.9
6.0
7.8
8.8
9.0
10.2 11.0 11.4 14.0 14.0 14.1 14.5 21.0 21.3 32.2 33.2
EC50 (μM) ± 0.1 ± 0.4 ± 1.2 ± 1.9 ± 2.9 ± 2.4 ± 4.1 ± 5.0 ± 3.8 ± 1.5 ± 0.8 ± 1.3 ± 1.8 ± 7.6 ± 10.0 ± 11.0 ± 10.9
H727 ICW
7.3 10.0 7.3 13.4 13.9 15.6 14.5 36.8 15.7 19.7 27.5 21.1 21.2 51.4 29.1 18.7 42.5
EC50 (μM) ± 3.1 ± 3.2 ± 1.5 ± 2.6 ± 0.2 ± 5.5 ± 1.0 ± 4.7 ± 2.2 ± 6.4 ± 2.8 ± 6.1 ± 8.4 ± 16.8 ± 1.8 ± 5.8 ± 15.1
K
d
(μM)
0.14 0.049 0.14 0.012 0.10 <0.10 0.11 0.12 0.013 0.012 0.036 0.009 0.46 14.8 2.59 55.0 0.18
HeLa
biphasic
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
–
–
+
+
–
–
–
H727
biphasic
–
–
+
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ICW EC50 measurements are reported as the mean ± SD of three individual experiments, each conducted in duplicate.
K
d
measurements are reported as the mean of two or more independent experiments, each conducted in duplicate. A
“
+” is used to indicate biphasic modulation of pERK1/2T202/Y204, as characterized by an increase in pERK1/2T202/Y204
levels at lower concentrations and a decrease in pERK1/2T202/Y204 levels below baseline at higher concentrations. A “–
” sign indicates that compound treatment caused an increase in pERK1/2T202/Y204 levels, but no subsequent decrease
below baseline over the concentrations tested. All pERK1/2T202/Y204 data are normalized to the DMSO vehicle control
and to the total ERK1/2 protein levels from the same well.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The majority of compounds examined by ICW exhibited the characteristic biphasic modulation that
we have come to expect from potent activators of SOS1-mediated nucleotide exchange, based on prior
work.25–27 On the other hand, less potent compounds with higher ICW EC50 values were less likely to elicit
biphasic modulation of ERK1/2 phosphorylation within the concentration range tested. We hypothesized
that compounds with higher binding affinities would modulate pERK1/2T202/Y204 signaling with greater potency.
Indeed, analogs with weaker binding affinities, such as 3, 23, and 10, all modulated pERKT202/Y204 signaling
with low potency. However, compounds with higher binding affinities did not necessarily modulate
pERK1/2T202/Y204 signaling with the greatest potencies. This is best illustrated by comparing the data for 47
and 65. Although the K value for 47 is an order of magnitude higher than that of 65, the ICW EC50 values
d
for 47 were 3–4-fold better than those of 65 in both cell lines. These data suggest that other factors beyond
only SOS1 binding affinity can influence pERK1/2T202/Y204 modulation in cells. These factors may include the
physical properties of the compounds such as lipophilicity and solubility or issues with membrane
permeability and/or plasma protein binding.44,45 Indeed, 4-N-methylpiperazino 64 displayed superior
solubility at pH 6.8 (>110 μg/mL) relative to its direct comparator 4-pyrimidino 60 (3 μg/mL). Furthermore,
several compounds, while capable of showing adequate passive permeability, appear to suffer from
excessive efflux (Supplementary Table S2). In addition to these characteristics, cellular considerations such
as protein localization,46 post-translational modifications,27,47 and additional layers of allosteric regulation48
that are not accounted for in a biochemical system could play a role in a disconnect between binding affinity
to SOS1 and cellular activity.
Finally, to assess whether compounds from the benzimidazole series activate RAS signaling in
cells we measured RAS-GTP levels and accompanying pERKT202/Y204 levels in HeLa cells (Figure 7). We
19
ACS Paragon Plus Environment

Page 21 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
chose to test matched sets of compounds that featured the same substituents at the 4-position of the
benzimidazole scaffold, but between these sets, differed in their ability to modulate pERK1/2T202/Y204 signaling
as measured by ICW. Regardless of the increase in binding affinity afforded by the diazaspiro[3.3]heptane
substituent at the 2-position, an N-methylpiperazine substituent at the 4-position, as in 42 and 64, resulted
in enhanced potency by pERK1/2T202/Y204 ICW, relative to the ortho-methylpyrimidine of 55, 60, and 61, which
modulated pERK1/2T202/Y204 signaling less potently.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. RAS-GTP levels and corresponding pERK1/2T202/Y204 levels from HeLa cells that were treated for 10 min with
up to 30 μM of compound 42, 64, 55, 60, or 61. EGF (50 ng/mL for 5 min) was used as a positive control for pathway
stimulation. Data are representative of two independent experiments.
Treatment with compounds 42 or 64 caused substantial increases in RAS-GTP levels at 10–30
μM, and induced small increases in RAS-GTP at concentrations as low as 0.3 μM. On the other hand, only
modest increases in RAS-GTP levels were observed with 55, 60, and 61 over the same range of dosage
concentrations. These findings are consistent with ICW pERK1/2T202/Y204 EC50 data, which identified
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compounds 42 and 64 as amongst the most potent compounds tested in cells. Our laboratory has typically
assessed RAS-GTP levels after 30 min of compound treatment with all previous chemical series of SOS1
agonists that we have discovered, and generally, treatment concentrations up to 100 μM were required to
observe notable changes in RAS-GTP levels.25,26 However, compounds 42 and 64 activated RAS-GTP and
caused biphasic modulation of pERK1/2T202/Y204 signaling with such high potencies that a lower maximum
concentration of 30 µM and only 10 min of treatment were required to sufficiently visualize their effects.
Furthermore, 64 elicited robust activation of RAS-GTP at lower concentrations than 42, which may be a
result of the higher SOS1 binding affinity of 64 compared to 42. Overall, compared to the other series of
SOS1 agonists discovered in our lab, advanced examples from this series of benzimidazole-derived
compounds modulate RAS-GTP and associated downstream signaling with the greatest potency that we
have observed, to date.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound Synthesis.
To access various benzimidazoles with 5,6-dimethyl moieties, commercially available bisaniline
derivative 66 was acylated with 1,1-carbonyldiimidazole (CDI) and then chlorinated in refluxing POCl to
3
afford 2-chloro benzimidazole 67 (Scheme 1). From there, alkylation of the benzimidazole nitrogen atom
with a variety of substituted benzyl bromides and subsequent displacement of the 2-chlorine atom with an
array of amines via an S Ar-type reaction furnished analogs 5–8, 10–19 and 21. The 5,6H-derivative 23
N
was constructed in a similar manner, beginning from commercially available benzene-1,2-diamine (68).
Compound 20 was accessed via a Suzuki coupling between 10 and cyclopropylboronic acid. 2-Amino and
2H-derivatives 3 and 4 were prepared by alkylation of commercially available compounds 69 and 70,
respectively, with 3-chloro-4-fluorobenzyl bromide. Lastly, 66 was condensed with piperidine-4-carboxylic
acid followed by Boc-protection of the piperidine nitrogen atom to afford intermediate 71. Subsequent
benzylation of the benzimidazole nitrogen atom and Boc-removal furnished analog 9.
21
ACS Paragon Plus Environment

Page 23 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of Various Benzimidazole Derivatives.a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aReaction conditions: (a) CDI, DMAP, CH
diamine, DMA, 165 °C in microwave; TFA, CH
acid, Pd(OAc) , PCy ·HBF , DME/H O, 90 °C. (f) urea, xylenes, 130 °C. (g) piperidine-4-carboxylic acid, 4 M HCl(aq),
05 °C. (h) Boc O, Et N, THF/DMF, 35 °C. (i) NaH, 3-chloro-4-fluorobenzyl bromide, DMF, rt. (j) TFA, CH Cl , rt.
Cl
, rt. (b) POCl
, 90 °C. (c) benzyl bromide derivative, K
2
2
3
2 3
CO , DMF, rt. (d)
2
Cl2, rt when Boc removal necessary. (e) K
3
PO , cyclopropylboronic
4
2
3
4
2
1
2
3
2
2
In order to synthesize derivatives with varying 5,6-substitution patterns, commercially available 1-
fluoro-2-nitrobenzenes, represented by 72, were treated with 3-chloro-4-fluorobenzylamine via an S Ar-
type reaction and then reduced using Zn/HCl to provide anilines 73 (Scheme 2). These intermediates were
acylated using CDI and then chlorinated by heating with POCl to form 2-chlorobenzimidazoles 74.
Substitution of the 2-chlorine atom with piperazine afforded derivatives 22, 24, 25, 27, and 28. Compound
N
3
26 was accessed via a Suzuki coupling between 27 and cyclopropylboronic acid.
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 2. Synthesis of 5,6-Substituted Benzimidazole Derivatives.a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aReaction conditions: (a) i-Pr
CDI, DMAP, CH Cl , rt. (d) POCl
cyclopropylboronic acid, dioxane, 100 °C.
Cl
, rt. (c)
2
Cl ,
2
NEt, 3-chloro-4-fluorobenzylamine, NMP, rt. (b) Zn°, 12 M HCl, EtOH/CH
2
2
2
2
3
, 100 °C. (e) piperazine, DMA, 165 °C in microwave. (f) K PO , PdCl
3
4
2
dppf·CH
2
Compounds with various 4-position substituents were constructed by treating 3-bromo-5-
chlorobenzene-1,2-diamine 75 with CDI and then POCl to furnish 2-chlorobenzimidazole intermediate 76
Scheme 3). Benzylation with 4-fluoro-3,5-dimethylbenzyl bromide or 4-fluorobenzyl bromide and
3
(
successive displacement of the 2-chlorine atom with piperazine, 1-Boc-piperazine or 1-Boc-2,6-
diazaspiro[3.3]heptane provided 4-bromo compounds 77, 78, 79, or 80. These intermediates were
functionalized at C-4 by various metal catalyzed C–C and C–N bond forming reactions, or reduced via
palladium catalyzed hydrogenolysis to afford analogs 29–51 and 54–65 after exposure to trifluoroacetic
acid to effect N-Boc deprotection.
23
ACS Paragon Plus Environment

Page 25 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 3. Synthesis of 4-Substituted Benzimidazole Derivatives.a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aReaction conditions: (a) CDI, DMAP, CH
fluorobenzyl bromide, K CO , DMF, rt. (d) diamine, K
Cl , rt. (g) 10% Pd/C, H , AcOH, i-PrOH, rt. (h) NaBH
4
Cl
, rt. (b) POCl
, 110 °C. (c) 4-fluoro-3,5-dimethylbenzyl bromide or 4-
or i-Pr NEt, DMF or DMSO, 80–100 °C. (e) Pd cat. C–C or
, MeOH, rt; 4 M HCl in dioxane.
2
2
3
2
3
2
CO
3
2
C–N coupling. (f) TFA, CH
2
2
2
■ CONCLUSION
A set of early hit molecules based on a benzimidazole scaffold were found to activate SOS1-
mediated nucleotide exchange on RAS. The SAR of these compounds was systematically investigated
using two different biochemical assays and X-ray co-crystal structures of select SOS1 ligands bound to the
RAS:SOS1:RAS complex were used as guides for analog design. Through these studies, several novel
derivatives were found to exhibit low double-digit nanomolar affinity for SOS1, which represent the tightest
binders of SOS1 reported to date. These compounds also elicited increases of RAS-GTP levels in cells
with the highest potency observed from any SOS1 ligands yet tested, while inducing a robust biphasic
response in downstream pERK1/2T202/Y204 signaling that is characteristic of SOS1 agonists. These
observations are in agreement with our proposed mechanism that compound-mediated agonism of SOS1
can modulate RAS signaling via negative feedback on SOS1 by active pERK1/2T202/Y204. The
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
benzimidazole-derived compounds described here and future derivatives may serve as chemical tools for
studying modulation of RAS signaling via binding to SOS1 and could inform future discoveries regarding
small molecule therapeutics for the treatment of RAS-driven malignancies.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
■ EXPERIMENTAL
General Procedures. All chemical reagents and reaction solvents were purchased from commercial
1
suppliers and used as received. Proton nuclear magnetic resonance ( H NMR) spectra were recorded at
1
either 400 MHz or 600 MHz on a Bruker spectrometer. For H NMR spectra, chemical shifts are reported
in parts per million (ppm) and are reported relative to residual non-deuterated solvent signals. Coupling
constants are reported in hertz (Hz). The following abbreviations (or a combination, thereof) are used to
describe splitting patterns: s, singlet; d, doublet; t, triplet; q, quartet; quint, quintet; m, multiplet; comp,
overlapping multiplets of non-magnetically equivalent protons; br, broad. All final compounds were of 95%
purity or higher, unless otherwise noted, as measured by analytical reversed-phase HPLC. Analytical HPLC
was performed on an Agilent 1200 series system with UV detection at 214 and 254 nm, along with
evaporative light scattering detection (ELSD). Low-resolution mass spectra were obtained on an Agilent
6
140 mass spectrometer with electrospray ionization (ESI). LC-MS experiments were performed with the
following parameters: Phenomenex Kinetex 2.6 μm XB-C18 100 Å, LC column 50 x 2.1 mm; 2 min gradient,
%–95% MeCN in H O, and 0.1% TFA or 0.1% formic acid. Analytical thin layer chromatography (TLC)
5
2
was performed on Kieselgel 60 F254 glass plates precoated with a 0.25 mm thickness of silica gel. TLC
plates were visualized with UV light and iodine. Silica gel chromatography was performed using a Teledyne
®
ISCO Combiflash Rf system, eluting with varying concentrations of EtOAc in hexanes or MeOH in CH
2
Cl .
2
Preparative reversed-phase HPLC was performed on a Gilson instrument equipped with a Phenomenex
Kinetex C18 column, using varying concentrations of MeCN in H O, and 0.1% TFA. Compounds 1 and 2
2
were purchased from ChemDiv and used as received. Compounds 52 and 53 were purchased from Viva
Biotech (Shanghai, China) and used as received. Starting materials 66, 68, 69, 70, and 1-fluoro-2-
nitrobenzene derivatives, represented by 72, were purchased from commercial suppliers and used as
received. Compound 75 was prepared according to a previously reported protocol.49
25
ACS Paragon Plus Environment

Page 27 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
General Procedure A: Synthesis of 5–8, 10–19, 21, and 23. Step 1: A mixture of 67 (1.0 equiv), an
appropriate benzyl bromide derivative (1.1 equiv), and K
2
CO (2.0 equiv) in DMF (0.25 M) was stirred at rt
3
for 16 h. The mixture was diluted with EtOAc, poured into a 50% sat. brine solution, and extracted with
EtOAc (3x). The combined organics were washed with a 50% sat. brine solution, a sat. brine solution, dried
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Na
2
SO ), and concentrated under reduced pressure. The crude residue was purified by silica gel
4
chromatography. Step 2: The material obtained from step 1 (1.0 equiv) was mixed with an appropriate
diamine (10.0 equiv) in N,N-dimethylacetamide (0.2 M) and heated to 165 °C in a microwave reactor for 45
min. The mixture was cooled to rt, diluted with EtOAc, poured into a 50% sat. brine solution, and extracted
with EtOAc (3x). The combined organics were washed with a 50% sat. brine solution, a sat. brine solution,
dried (Na
2
SO ), and concentrated under reduced pressure. The crude residue was purified by reversed-
4
phase preparative HPLC and the pure fractions were concentrated under a stream of warm air to afford the
target compound as a TFA salt. Alternatively, the free-base was obtained by neutralizing the pure HPLC
fractions with solid Na
reduced pressure.
2
CO
3
, extraction with CH
Cl
2 2
(3x), drying over Na
2
SO , and concentrating under
4
General Procedure B: Synthesis of 22 and 24–28. Step 1: A mixture of an appropriately substituted 1-
fluoro-2-nitrobenzene derivative (72, 1.0 equiv) and N,N-diisopropylethylamine (1.5 equiv) in N-methyl-2-
pyrrolidinone (0.5 M) was charged with 3-chloro-4-fluorobenzylamine (1.2 equiv) and stirred at rt for 16 h.
The reaction mixture was poured into ice cold brine and stirred for 5 min. The resultant yellow precipitate
was collected via filtration, rinsed with water, and dried under reduced pressure. Step 2: The crude material
obtained from step 1 was suspended in a mixture of EtOH/CH
dust (10.0 equiv) was added. After the reaction was complete, as indicated by LC-MS, the reaction mixture
was filtered. Water was added to the filtrate and the suspension was extracted with CH Cl (3x), dried
), and concentrated. Step 3: The crude residue obtained from step 2 was re-suspended in CH Cl
0.3 M), charged with DMAP (0.1 eq) and CDI (2.0 eq), and stirred at rt. After 2 h, 1 M HCl was added and
2
2
Cl /12 M HCl (3:1:1 v/v/v, 0.2 M) and Zn
2
2
(
(
Na
2
SO
4
2
2
the mixture was stirred rapidly. The resultant solid was collected via filtration, rinsed with water, rinsed with
CH Cl , and dried under reduced pressure. Step 4: The crude solid obtained from step 3 was suspended
in POCl (1.0 M) and heated to 100 °C. After the reaction was complete, as indicated by LC-MS, the mixture
was cooled to 0 °C and poured slowly into water. After all the POCl
2
2
3
3
was quenched, the mixture was slowly
26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
neutralized with solid NaOH. The resulting precipitate was collected via filtration, rinsed with water, and
dried under reduced pressure. Step 5: The crude solid obtained from step 4 and piperazine (10.0 equiv)
were dissolved in a mixture of N,N-dimethylacetamide (0.2 M) and heated to 165 °C in a microwave reactor
for 45 min. The mixture was cooled to rt, diluted with EtOAc, poured into a 50% sat. brine solution, and
extracted with EtOAc (3x). The combined organics were washed with a 50% sat. brine solution, a sat. brine
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
solution, dried (Na
2
SO ), and concentrated under reduced pressure. The crude residue was purified by
4
reversed-phase preparative HPLC and the pure fractions were concentrated under a stream of warm air to
afford the target compound as a TFA salt.
General Procedure C: Synthesis of 30–39, 43, 54–63, and 65. Step 1: A mixture of an appropriate
boronic acid or boronic acid pinacol ester (1.5 equiv), PdCl
7, 78, 79, or 80 (1.0 equiv) under Ar(g) was charged with degassed DMF/ethanol (4:1 v/v, 0.35 M) and
heated to a temperature of 90 °C. After 18 h, the reaction mixture was cooled to rt, diluted with sat. aqueous
ammonium chloride, and extracted with CH Cl (3x). The organic phases were combined, dried (Na SO ),
and concentrated under reduced pressure. The crude residue was purified by silica gel chromatography.
Step 2: The compound obtained from step 1 (1.0 equiv) was dissolved in CH Cl (0.05 M) and stirred at rt.
2
dppf·CH
Cl
2 2
2
(0.1 equiv), K CO
3
(3.0 equiv), and
7
2
2
2
4
2
2
TFA (0.2 M) was added and the mixture was stirred for 2 h before being concentrated under reduced
pressure. The residue was purified by reversed-phase preparative HPLC and the pure fractions were
concentrated under a stream of warm air to afford the target compound as a TFA salt. Alternatively, the
free-base was obtained by neutralizing the pure HPLC fractions with solid Na
3x), drying over Na SO , and concentrating under reduced pressure.
General procedure D: Synthesis of 40–42, 44–48, and 64. A mixture of amine (2.0 equiv), Pd(OAc)
equiv), BINAP (0.2 equiv), and K PO (3.0 equiv) under Ar(g) was charged with a solution of either 77 or 79
1.0 equiv) in degassed toluene (0.1 M) and heated to 100 °C with vigorous stirring. Once the starting
2
CO
3
, extraction with CH
2
Cl
2
(
2
4
2
(0.1
3
4
(
material was consumed, as indicated by LC-MS, the mixture was diluted with EtOAc, filtered over
diatomaceous earth, and concentrated under reduced pressure. The crude residue was dissolved in a
mixture of TFA and CH
2
Cl (1:2 v/v, 0.05 M) and stirred at rt for 1 h. The reaction mixture was concentrated
2
under reduced pressure and purified by reversed-phase preparative HPLC. The pure fractions were
concentrated under a stream of warm air to afford the target compound as a TFA salt. Alternatively, the
27
ACS Paragon Plus Environment

Page 29 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
free-base was obtained by neutralizing the pure HPLC fractions with solid Na
2
CO
3
, extraction with CH
2
Cl
2
(
3x), drying over Na
General procedure E: Synthesis of 50 and 51. Step 1: A mixture of 77 (1.0 equiv), an appropriate
potassium pyridine trifluoroborate (1.1 equiv), Pd(OAc) (0.1 equiv), RuPhos (0.2 equiv), and Na CO (2.0
2
SO
4
, and concentrating under reduced pressure.
2
2
3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
equiv) under Ar(g) was charged with degassed ethanol (0.2 M). The reaction mixture was heated to 80 °C
for 48 h. The reaction mixture was cooled to rt, diluted with sat. aqueous ammonium chloride, and extracted
with dichloromethane (3x). The organic phases were combined, dried (Na
reduced pressure. The residue was purified by silica gel chromatography. Step 2: The compound obtained
from step 1 (1.0 equiv) was dissolved in CH Cl (0.05 M) and TFA (0.2 M) was added. The mixture was
2
SO
4
), and concentrated under
2
2
stirred for 2 h before being concentrated under reduced pressure. The crude residue was purified by
reversed-phase preparative HPLC and the pure fractions were concentrated under a stream of warm air to
afford the target compound as a TFA salt.
1
-(3-Chloro-4-fluorobenzyl)-5,6-dimethyl-1H-benzo[d]imidazol-2-amine (3 as TFA salt). A mixture of
9 (50 mg, 0.31 mmol), 3-chloro-4-fluorobenzyl bromide (50 µL, 0.37 mmol), and NaH (60% dispersion in
6
mineral oil, 25 mg, 0.63 mmol) in EtOH (4 mL) was stirred at 60 °C until starting material was completely
consumed, as indicated by LC-MS. The mixture was cooled to rt, acidified with a 10% HCl solution,
extracted witch CH
Cl
2 2
(3x), dried (Na
2
SO ), and concentrated under reduced pressure. The crude residue
4
was purified by reversed-phase preparative HPLC and the pure fractions were concentrated under a stream
1
of warm air to afford 8 mg of the title compound as the corresponding TFA salt (6% yield). H NMR (400
MHz, CD
3
OD): δ 7.41 (dd, J = 6.8, 2.2 Hz, 1H), 7.25 (t, J = 8.8 Hz, 1H), 7.22–7.16 (comp, 2H), 7.14 (s, 1H),
5
.35 (s, 2H), 2.34 (s, 3H), 2.32 (s, 3H). MS (ESI) m/z = 304.0 [M+H]+.
1-(3-Chloro-4-fluorobenzyl)-5,6-dimethyl-1H-benzo[d]imidazole (4 as TFA salt). A mixture of 70 (50
mg, 0.34 mmol), 3-chloro-4-fluorobenzyl bromide (50 µL, 0.38 mmol), and K
2
CO (71 mg, 0.51 mmol) in
3
DMF (3 mL) was stirred at rt for 1 h. The mixture was acidified with a 10% HCl solution, diluted with EtOAc,
poured into a 50% sat. brine solution, and extracted with EtOAc (3x). The combined organics were washed
with a 50% sat. brine solution, a sat. brine solution, dried (Na
2
SO ), and concentrated under reduced
4
pressure. The crude residue was purified by reversed-phase preparative HPLC and the pure fractions were
concentrated under a stream of warm air to afford 15 mg of the title compound as the corresponding TFA
28
ACS Paragon Plus Environment