Synthesis and photophysical properties of a fluorescent cyanoquinoline probe for profiling ERBB2 kinase inhibitor response

Article abstract of DOI:10.1016/j.bmc.2017.09.034

A fluorescent probe targeting the ERBB2 receptor tyrosine was designed, synthesized and evaluated as reporter of ERBB2 dynamics in overexpressing BT474, i.e. Her2(+), cells. Two cyanoquinazoline (CQ) probes modeled after type-I (CQ1) or active state and t

Full text of DOI:10.1016/j.bmc.2017.09.034

Accepted Manuscript
Synthesis and photophysical properties of a fluorescent cyanoquinoline probe
for profiling ERBB2 kinase inhibitor response
Heajin Lee, Ralf Landgraf, James N. Wilson
PII:
S0968-0896(17)30930-6
https://doi.org/10.1016/j.bmc.2017.09.034
BMC 13994
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
20 July 2017
21 September 2017
22 September 2017
Please cite this article as: Lee, H., Landgraf, R., Wilson, J.N., Synthesis and photophysical properties of a fluorescent
cyanoquinoline probe for profiling ERBB2 kinase inhibitor response, Bioorganic & Medicinal Chemistry (2017),
doi: https://doi.org/10.1016/j.bmc.2017.09.034
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
Synthesis and photophysical properties of a
fluorescent cyanoquinoline probe for
profiling ERBB2 kinase inhibitor response
Leave this area blank for abstract info.
Heajin Lee, Ralf Landgraf, and James N. Wilson
Department of Chemistry, University of Miami, Coral Gables, Florida 33124, United States.
Department of Biochemistry and Molecular Biology, University of Miami, Miami, Florida 33101, United States.
Sylvester Comprehensive Cancer Center, University of Miami, Miami, Florida 33101, United States.
Cl
F
N
N
HN
CN
live cell imaging of
ERBB2 internalization
N
donor-acceptor π system
with red-shifted fluorescence

Bioorganic & Medicinal Chemistry
journal homepage: www.els evier.com
Synthesis and photophysical properties of a fluorescent cyanoquinoline probe for
profiling ERBB2 kinase inhibitor response
Heajin Lee,a Ralf Landgraf,b,c* and James N. Wilsona,c*
aDepartment of Chemistry, University of Miami, Coral Gables, Florida 33124, United States.
bDepartment of Biochemistry and Molecular Biology, University of Miami, Miami, Florida 33101, United States.
cSylvester Comprehensive Cancer Center, University of Miami, Miami, Florida 33101, United States.
ARTICLE INFO
ABSTRACT
Article history:
Received
A fluorescent probe targeting the ERBB2 receptor tyrosine was designed, synthesized and
evaluated as reporter of ERBB2 dynamics in overexpressing BT474, i.e. Her2(+), cells. Two
cyanoquinazoline (CQ) probes modeled after type-I (CQ1) or active state and type-II (CQ2) or
inactive state inhibitors were designed and synthesized with extended π systems that impart
binding-induced, turn-on fluorescence. Solution spectroscopy revealed that CQ1 exhibited
attractive photophysical properties and displayed turn-on emission in the presence of purified,
soluble ERBB2 kinase domain, while CQ2 was found to be non-emissive, likely due to
quenching via a photoinduced electron transfer mechanism. Live cell imaging with CQ1
revealed that this probe targeted an intracellular population of ERBB2, which increased
following treatment with type-I inhibitors, gefinitib and canertinib, but showed no response to
type-II inhibitors. CQ1 thus provides a novel means of imaging the dynamic response of
ERBB2(+) cells to kinase inhibitors.
Received in revised form
Accepted
Available online
Keywords:
receptor tyrosine kinase
kinase inhibitor
fluorescent probe
EGFR
2009 Elsevier Ltd. All rights reserved.
ERBB
1. Introduction
Cl
Cl
C
The advent of modern chemotherapies, including monoclonal
antibodies and small molecule kinase inhibitors, has
revolutionized the treatment of many cancers.1 Yet, rapidly
identifying the most effective treatment regime remains a
challenge that is complicated by tumor heterogeneity, evolving
mutations and compensatory shifts in signaling pathways that
minimize the impact of targeted therapies.2,3 Genomic analysis
can provide great predictive power for the efficacy of certain
drugs, but functional phenotypic assays are also a useful tool. To
this end, we recently reported fluorescent kinase probes, kProbes,
directed at the EGFR/ERBB family of receptor tyrosine kinases.4-
a)
b)
c)
F
N
O
HN
N
HN
N
N
H
O
H
N
H
N
CN
N
N
O
O
EWG
Thr
O
O
EKB-569, pelitinib
HKI-272, neratinib
Cl
Cl
F
O
F
N
N
N
N
HN
N
HN
N
H
O
O
H
N
N
6
Thr
Modification of an N-phenyl quinazoline core, an established
QA1
QA2
ERBB-targeted pharmacophore common to erlotinib, gefitinib
and lapatinib,7,8 introduced environmentally responsive
fluorescence switching, such that the probes are non-emissive, or
OFF in solution and become photoemissive upon binding, i.e.
emission is switched ON.4 This OFF/ON, or turn-on emission,
response allows the membrane-permeable probes to be utilized
on live cells, without the need for rinsing or removal of the
unbound probe. First generation probes, QA1 and QA2 (Figure
1) were shown to selectively stain and differentiate ERBB2-
overexpressing, i.e. Her2(+), cells, from low level expressing
cells classified clinically as Her2(-).6 While receptor activation
dynamics could be reported in real time at the single cell level,6
spectral overlap with the moderate inherent emission of small
Cl
Cl
F
O
F
N
N
N
N
HN
HN
N
CN
CN
N
CQ1
CQ2
Figure 1. A) Examples of type-I, pelitinib, and type-II, neratinib,
cyanoquinolines under investigation as EGFR/ERBB inhibitors; the
cyano moiety interacts with a threonine residue in the ATP binding
pocket, replacing a water bridge found for quinazoline inhibitors.12,13 B)
Fluorescent quinazoline probes developed in our lab. C) Cyanoquinoline
probes with modified donor-π-acceptor investigated in this study.

molecule inhibitors,9 such as lapatinib10 or gefitinib,11 prevented
the use of QA1 and QA2 for assaying the effect of these drugs.
To maximize the utility of these fluorescent reporters in novel
drug screens or mechanistic dissection of signaling pathways
using established inhibitors, the optical properties of fluorescent
kProbes needs to be tuned to minimize spectral overlap with
typically UV to blue emissive pharmacophores.
a)
b)
N
N
HN
N
HN
N
CN
N
QA
CQ
In this contribution, we report the design, synthesis and optical
properties of cyanoquinoline kProbes, which exhibit significantly
red-shifted fluorescence compared to their quinazoline
predecessors. The optical properties of CQ1, in particular, make
it an attractive candidate for both detecting ERBB2-expressing
cells, as well as monitoring drug response. CQ1 exhibits a large
Stokes shift, with emission centered at 520 nm, and a high turn-
on, i.e. ON/OFF, ratio between the ERBB-bound/unbound forms.
With these improved photophysical parameters, CQ1 is optically
compatible with many ERBB2-targeted kinase inhibitors, and can
be utilized to explore cellular responses to drug exposure. CQ1
stains the intracellular pool of ERBB2, which is not typically
accessible to antibody labeling in a live cell setting. Finally, we
demonstrate the use of CQ1 to image the dynamics of Gefitinib-
induced internalization of ERBB2.
ESP
c)
d)
LUMO
HOMO
Figure 2. a) Chemical structures, b) electrostatic potential map and c, d)
frontier molecular orbitals of model compounds used to compare the
electronic properties of kinase probes buiilt on the quinazoline (QA) and
cyanoquinoline (CQ) heterocyclic cores. While the HOMOs map with nearly
identical distributions, the LUMO shows slightly more polarization on CQ, a
result of the electron-withdrawing nature of the cyano group, which also
appears electronegative in the ESP surface.
2. Results and Discussion
2.1. Design and Synthesis
R1
R2
Several quinazoline-based inhibitors, such as erlotinib,
gefitinib and lapatinib, target the ATP binding fold of the
EGFR/ERBB family.7,8 In some cases, the inherent fluorescence
of these inhibitors allows for tracking of their subcellular
distribution,10,11 yet their properties as fluorescent probes are less
than ideal, requiring UV excitation and exhibiting low quantum
yields. However, at clinically relevant concentrations (1-10 µM)
they are sufficiently emissive to contribute to moderate
background emission that overlaps with our recently reported
ERBB2-targeted kProbes, QA1 and QA2. Thus, we set out to
develop next generation probes with significantly red shifted
emission spectra required to image the dynamics of this receptor
tyrosine kinase family in response to these quinazoline-based
inhibitors. 3-Cyanoquinolines, including pelitinib and
neratinib,12,13 have also been demonstrated as inhibitors of these
receptor tyrosine kinases. Their binding mode differs slightly
with the cyano group replacing a bridging water molecule, which
interacts with a threonine residue via a hydrogen bond as shown
in Figure 1 with neratinib versus QA2. Our investigations into
the optical properties of several donor-acceptor quinazoline π-
systems demonstrated that strong electron-withdrawing groups
effectively lowered the energy of the S1 state producing a
concomitant red shift in both the absorption and excitation
spectra of model quinazoline kProbes.5
R2
R1
HN
N
Cl
N
R3
B(OH)2
CQ
CQ1
CQ2
H2N
CN
Br
CN
Br
a
b
1
2-4
2, CQ: R1 = R2 = H
3, CQ1: R1 = Cl, R2 = F
4, CQ2: R1 = Cl, R2 = OCH2(3-fluorophenyl)
CQ: R3 = NMe2
CQ1, CQ2: R3 = N
NMe
Figure 3. Synthesis of fluorescent cyanoquinoline probes. Conditions: a) 2-
propanol, 80° C, 24 h; b) Pd(OAc)2, PPh3, Et3N, DMF, 80° C, 24 h.
a)
absorption
PhCH3
1
25000
CHCl3
DMSO
emission
PhCH3
CHCl3
DMSO
0
0
300
400
500
600
λ / nm
b)
c)
QA1
CQ1
Based on quantum chemical calculations, we anticipated that
switching from the moderately electron-withdrawing nitrogen
atom at the 3-position of QA1 or QA2 to the stronger electron
withdrawing nitrile group of the cyanoquinoline probes, CQ1
and CQ2, could generate a probe with red-shifted excitation and
emission spectra. DFT calculations on truncated model
compounds QA and CQ (Figure 2) predicted an S1 state that was
only lowered by 0.3 eV, corresponding to a ~25 nm shift in the
absorption spectra. However, mapping of the frontier molecular
orbitals, as well as the dipoles of the molecules, 6.6 D for QA
versus 8.9 D for CQ, suggested that the excited state of the
cyanoquinoline probes should be more polarized than that of of
the quinazoline probes. This can be seen in a small, but non-
trivial shift in orbital density in the frontier molecular orbitals.
While the HOMOs of QA and CQ are virtually identical, the
ET(30) / kcal/mol
Figure 4. a) Absorption and emission spectra of CQ1 in several solvents
demonstrating the sensitivity of the excited state to solvent polarity. b)
Comparison of the emission maxima of CQ1 and QA1 in various solvents
reveals that the cyanoquinoline core of CQ1 is slightly more responsive to
solvent polarity. c) A visual comparison of QA1 and CQ1 in chloroform
shows the red-shifted emission of CQ1.

LUMO of CQ showed density shifted away from the
dimethylamino group and towards the cyano group, which could
contribute to thermal relaxation in the excited state and an
increased Stokes shift.
excitation
gefitinib
lapatinib
QA1
1
CQ1
The synthesis (Figure 3) of CQ1 and CQ2 proceeded from 6-
bromo-4-chloro-3-quinolinecarbonitrile (1) which was generated
emission
gefitinib
lapatinib
QA1
1
according to reference 14 and confirmed by H and 13C NMR
spectroscopy (see Supplementary Material) as well as HRMS.
The two pharmacophore arms, or phenyl moiety in the case of the
model compound, CQ, were installed via nucleophilic aromatic
substitution of the chlorine at the 4-position, followed by the
electron-donating anilino arm, which coupled under Suzuki
conditions to the 6-position. The compounds were obtained as
bright yellow solids following column chromatography and
crystallization. Ki values of 390 nM and 590 nM were
determined for CQ1 and CQ2, respectively, in receptor
stimulation assays (see Experimental section, below).
0
CQ1
300
400
500
600
λ / nm
Figure 5. Comparison of excitation and emission spectra, in the presence
of purified ERBB2 kinase domain, of two quinazoline inhibitors,
gefitinib and lapatinib, with quinazoline probe, QA1 and cyanoquinoline
probe, CQ1. While all four compounds share similar excitation spectra,
the emission of CQ1 is significantly red-shifted, allowing for selective
imaging of ERBB2-binding in cells treated with ERBB2-targeted
inhibitors. Inset shows a comparison of QA1 and CQ1 emission in
CHCl3.
2.2. Optical Properties
We first examined the optical properties of CQ1 and CQ2 in
organic solvents to evaluate their sensitivity to their chemical
environment and polarity-dependent fluorescence (Figure 4).
Additionally, moderately polar organic solvents, such as CHCl3,
closely match the apparent polarity of the ATP binding pocket of
ERBB2 and emission in these solvents informs our imaging
parameters for fluorescence microscopy.5 UV-vis absorption
spectroscopy showed little variation between solvents, with peak
absorption at 330 nm and ε varying slightly from 2.5 x 104 cm-1
M-1to 2.9 x 104 cm-1 M-1 for both compounds, however,
fluorescence spectroscopy showed marked differences. CQ1 was
emissive in solvents of low polarity, such as toluene (Φem = 0.38)
with an ET30 value of 33.9 kcal/mol on Reichardt's scale, and
chloroform (Φem = 0.21) with an ET30 value of 39.1 kcal/mol, but
essentially nonemissive in solvents more polar that DMSO (ET30
= 45.0 kcal/mol).15 Emission was completely quenched in water,
octanol, methanol and acetonitrile, which have ET30 values
ranging from 46.0 to 63.1 kcal/mol. A linear correlation between
solvent polarity and emission wavelength was found by
examining several other solvents (Figure 4b) and the shift in
emission wavelength was found to be higher for CQ1 compared
to QA1, demonstrating that the more polar excited state was
better stabilized in polar solvents. The marked red-shift in the
emission spectra of CQ1, relative to QA1, is also apparent by
eye when comparing the probes in identical solvents, as shown in
Figure 4c. This again is testament to the presencet of the
electron-withdrawing nitrile, which lowers the energy of the S1
state. While CQ1 was found to be emissive, CQ2 showed no
emission in any solvent. By comparison, the quinazoline analog,
QA2, was weakly emissive in relation to QA1. The addition of
the type 2 pharmacophore arm appears to have a quenching effect
on both the 3-cyanoquinoline and quinazoline fluorophore cores,
which is exacerbated by the presence of the electron withdrawing
cyano group. We attribute this quenching to an excited state
electron transfer process, from the electron-rich N-phenyl-
benzyloxy ether.16 Thus, while CQ2 is cannot be utilized as a
fluorescent turn-on probe, CQ1 is ideally suited as a red-
emissive, type 1 inhibitor.
b) anti-ERBB2
c) overlay
a) CQ1
10 µm
Figure 6. CQ1 colocalizes with anti-ERBB2 staining: a) the false-
colored green channel shows CQ1 distribution, λex = 405 nm, λem = 500 -
600 nm; b) the false-colored red channel shows antibody staining of
ERBB2; c) overlay of the two channels shows a high degree of overlap.
which sufficiently segregates the emission of this probe from the
kinase inhibitors. The emission maximum of the bound form of
CQ1 is 520 nm, compared to 440 nm for ERBB2-bound lapatinib
and gefitinib, providing a unique optical window that allows for
selective monitoring of CQ1 emission in the presence of
ERBB2-directed inhibitors. Having established the optical
compatibility of CQ1 with quinazoline-based inhibitors, we next
examined the live cell imaging capabilities of this probe.
2.3. Confocal Microscopy
We next evaluated CQ1 as an ERBB2-targeted probe in a live
cell setting using BT474 cells, which are classified clinically as
Her2(+) and express ~106 copies of ERBB2. Cells were first
stained with ERBB2-directed antibody, then treated with 2 µM
solutions of CQ1. Figure 6 shows the two-channel image of the
CQ1 (green) and anti-ERBB2 (red) staining, which show a high
degree of overlap in the merged image. Minimal cell surface
staining was observed for CQ1, indicative of binding to
internalized receptors, and longer antibody staining times (12-16
h) were required to access this internalized pool of ERBB2. By
contrast, the small molecule, membrane-permeable probe
required only 5 min of staining to saturate the same population.
Having established that CQ1 targets ERBB2 in live cells, our
final goal was to demonstrate the compatibility of probe with
clinically-relevant inhibitors. Gefitinib provides an interesting
case-study for ERBB2 dynamics as it is a low-affinity inhibitor
of ERBB2, with IC50 values varying widely from 0.24 µM17 to
in excess of 5 µM18-20 and linked to factors such as expression
level,21,22 tumor type18 and binding pocket mutations.20,23 In terms
of receptor dynamics, Gefitinib has been shown to induce
heterodimerization24 and internalization of ERBB225 providing an
excellent opportunity to track dynamic ERBB2 pools with CQ1
as a fluorescent reporter. This is evident in the time-dependent
We next investigated the binding-induced, turn-on emission of
CQ1 and its optical compatibility with several commercially
available inhibitors in the presence of purified ERBB2 kinase
domain. As shown in Figure 5, the excitation spectra of ERBB2-
bound CQ1 is similar to ERBB2-bound QA1 and overlaps
significantly with quinazoline-based ERBB inhibitors such as
gefitinib and lapatinib. However, CQ1 is distinguishable from
QA1, gefitnib and lapatinib through a large, 200 nm Stokes shift,

second wave of emission increase, which stabilized after 5-6 min.
The gefitinib-induced response was found to be concentration
dependent, as shown in Figure 7b, with an EC50 of 4.4 µM.
This value correlates very well with the average IC50 value of
3.9 µM reported across several tumor types.18 While gefitnib and
CQ1 target the same binding fold, CQ1, with a Ki of 390 nM
and used at 2 µM can effectively compete with gefitinib to access
the binding pocket and report ERBB2 internalization. To verify
this behavior, we also evaluated the effect of canertinib and
compared it to two type-II, i.e. inactive state inhibitors, lapatinib
and neratinib (blue and purple traces, Figure 7a). Canertinib
(green trace, Figure 7a) induces a more moderate increase in
CQ1 emission that can be rationalized on the basis of higher
affinity and ability to bind covalently, blocking competition from
CQ1.26 By contrast, neither inactive state inhibitor elicited an
increase in CQ1, supporting the notion and structural features
that CQ1 binding is to the active state.27 The possible interactions
for ERBB2 with type-I and type-II probes, as well as CQ1, are
summarized Figure 8.27-29 CQ1 thus provides a route not only to
image the internal pool of ERBB2 and dynamics of
internalization, but also to differentiate drug binding modes
which is linked to their effects on receptor conformation,
dimerization and internalization.
CQ1 + Gef
CQ1 + Can
CQ1 + Lap
CQ1 + Ner
CQ1 + DMSO
a)
b)
20
15
10
5
40
30
20
10
0
+ inhibitor/DMSO
0
0
2
4
6
8
0
2
4
time / min
6
8
10
[Gef] / µM
c)
5 min
7 min
9 min
11 min
20 µm
Figure 7. Imaging of geftinib-induced, active conformation ERBB2-
internalization: a) following CQ1 saturation of ERBB2, the addition of
gefitinib, and to a lesser extent canertinib, both active conformer
inhibitors, produces an increase in CQ1 emission intensity, while the
addition of inactive state inhibitors, lapatinib or neratinib, does not; b)
the gefitinib-induced emission increase shows a concentration
dependence, with and EC50 of 4.4 µM that closes matches the reported
IC50 values of 3.7 to 3.9 µM;18,19 c) images of time dependent increase in
emission in BT474 cells;
3. Conclusions
In summary, we have developed an ERBB2-targeted kinase
probe, CQ1, utilizing a 3-cyanoquinoline core, studied the
optical as well as biochemical properties, and demonstrated its
utility in imaging dynamic shifts in receptor populations in
response to various ERBB2-targeted inhibitors. The optical
properties of CQ1 were improved over previous fluorescent
reporters, owing to the presence of the nitrile group serves to
stabilize the charge transfer excited state and red-shift emission
of the probe beyond that observed for early generation probes
and clinically relevant EGFR/ERBB inhibitors. CQ1 was found
to have moderate affinity for ERBB2 and targets the intracellular
pool, allowing this probe to function as a reporter of the rapid
dynamics of kinase internalization, following inhibitor treatment.
Finally, we were able to differentiate the response of inhibitors
targeting the active and inactive conformation. These results
serve as a demonstration of the utility of a small molecule
fluorescent reporter that is membrane permeable and can target
an intracellular binding pocket. Future development of kinase-
specific probes, adapted from existing and clinically available
inhibitors, can complement established antibody-labeling
strategies, providing information about dynamic shifts in kinase
populations in response to specific drug treatments.
ensemble of stabilized
inactive states
(majority of receptors) active conformation
conformation-induced
dimerization, ubiquitination
and internalization
T1 inhibitor-induced
D
C
B
A
T1
T2
Ubq
T2
T1
T1
HSP
Ubq
-
T1
CQ1
CQ1
H
internalization-induced
changes in structure and
association states
unbound, quenched
probes
CQ1
recycling
binding-induced
turn-on emission
Ubq
Ubq
CQ1
CQ1
+
CQ1
HSP
HSP
+
lysosome
degradation
G
E
F
various endosome stages
ATP pocket binders
ERBB2-associated proteins
Ubq
HSP
T1
T2
Ubiquitin
type 2 inhibitor
HSP90
type 1 inhibitor
CQ1
CQ1 (quenched)
other adaptors
4. Experimental section
Figure 8. Stages in ERBB receptor signaling and model for potential
interplay of CQ1 and inhibitor binding: The active and inactive kinase
conformations on the cell surface (A-D) preferentially associate with, and
are stabilized by, distinct interaction partners. Binding of type-I and type-
II inhibitors (T1, T2) shifts the equilibrium of association states and is
known to drive both dimerization and internalization.24-29 Internalization
results in a change in association states and pH-induced changes in
interactions and conformations of the extracellular domains (E-G).30-32 At
present, little is known about the relative affinity of established inhibitors
and competition with kinase directed probes in these distinct states.
Current data on probe CQ1 (shown in on- and off-state) suggest that
probe binding is favored in the internalized state.
4.1. Chemistry
4.1.1. Synthesis of 6-bromo-N-(phenyl)-quinoline-3-
carbnitrile-4-amine (2).
2.5 grams of 1 (9.4 mmol), 1.5 g (16.1 mmol) of aniline, and
o
2-propanol were added, vigorously stirred and heated at 80 C,
overnight with a condenser. After 24 hours, it was cooled to
room temperature, concentrated under reduced pressure, 100 mL
of 1M KOH solution was added and stirred for 20 min, 300 mL
of EtOAc was added, then organic solution was extracted through
separation funnel and concentrated. It was purified over silica
(100% DCM to 100% EtOAc) followed by crystallization from
methanol to afford yellow powder of 2 (1.363g, 42.1 %); mp:
emission intensity of cells first treated with CQ1, followed by
gefitinib treatment (red trace, Figure 7a). BT474 cells were
exposed to 2 µM solutions of CQ1 and after 5 min, the signal had
plateaued, indicating saturation of the internal ERBB2
population. Subsequent addition of gefitinib (5 µM), induced a
216-218 C
; IR νmax(cm-1): 3279.7, 3202.7, 3123.3, 3053.9,
2226.2, 1865.0, 1739.3, 1608.8, 1582.8, 1557.1, 1495.6, 1447.6,
1353.1, 1318.5, 1274.8, 1232.3, 1117.0, 1077.7, 971.2, 826.9,
̊

769.3, 696.5, 610.5; 1H NMR(500 MHz, DMSO-d6), δ: 7.26-7.32
(m, 3H), 7.41 (t, 2H, J = 8.0 Hz), 7.86 (d, 1H, J = 8.5 Hz), 7.97
(dd, 1H, J = 8.5, 1.5 Hz), 8.60 (s, 1H), 8.78 (s, 1H), 9.92 (s, 1H);
13C NMR (125 MHz, DMSO-d6 ), δ: 88.82, 116.91, 120.01,
121.16, 125.00, 125.86, 126.54, 129.50, 132.03, 135.30, 139.48,
147.88, 150.69, 154.02; HR-ESI (Q-TOF) m/z: calculated for
C16H11BrN3+ (M+H+): 324.0136, found: 324.0135.
reaction mixture was degassed for a further 30 min under a slow
stream of nitrogen, at which point 80 mg of Pd(OAc)2 was added.
The reaction mixture was heated at 80 oC for 24 h, cooled,
filtered through filter paper and eluted solution was added into 50
mL of H2O directly, and ethyl acetate was poured onto funnel to
wash the residues. Water and ethyl acetate solution mixture went
through extraction with three times with 200 mL of ethyl acetate
each. The organic layer was dried with MgSO4, filtered, and then
concentrated under reduced pressure. The solid was crystalized in
methanol, ethylacetate multiple times to afford light yellow
4.1.2. Synthesis of 6-bromo-N-(3-chloro-4-fluoro-
phenyl)-quinoline-3-carbnitrile-4-amine (3).
1.0 grams of 1 (3.8 mmol), 0.8 g (5.5 mmol) of 3-chloro-4-
fluoro-aniline, and 2-propanol were added, vigorously stirred and
powder of CQ. (165.0 mg, 18.1 %); mp: 275-276 C
̊
(dec); IR
max(cm-1): 3197.2, 3159.9, 3082.3, 3060.4, 2989.3, 2888.7,
2814.2, 2220.7, 1944.4, 1851.4, 1733.8, 1603.5, 1577.8, 1547.0,
ν
o
heated at 80 C, overnight with a condenser. After 24 hours, it
was cooled to room temperature, concentrated under reduced
pressure, 50 mL of 1M KOH solution was added and stirred for
20 min, 100 mL of EtOAc was added, then organic solution was
extracted three times through separation funnel and concentrated.
It was crystallization from methanol, ethyl acetate to afford
1494.6, 1445.7, 1363.5, 1308.5, 1297.3, 1263.6, 1063.1, 965.9,
898.4, 808.6, 762.3, 691.6, 622.3; H NMR(500 MHz, DMSO-
1
d6), δ: 2.97 (s, 6H), 6.84 (d, 2H, J = 9.0 Hz), 7.25 (t, 1H, , J = 7.5
Hz), 7.32 (d, 2H, J = 8.0 Hz), 7.42 (t, 2H, J = 7.5 Hz), 7.75 (d,
2H, J = 9.0 Hz), 7.93 (d, 1H, J = 9.0 Hz), 8.13 (dd, 1H, J = 9.0,
1.5 Hz), 8.51 (s, 1H), 8.67 (d, 1H, J = 1.5 Hz), 9.90(s, 1H); 13C
NMR (125 MHz, DMSO-d6), δ: 40.44, 88.51, 112.96, 117.44,
118.50, 120.09, 124.99, 126.26, 126.48, 128.09, 129.52, 130.33,
130.46, 138.79, 140.04, 147.76, 150.68, 151.38, 152.49; HR-ESI
yellow powder of 3 (553 mg, 38.8 %); mp: 210-212 C
ν
̊
; IR
max(cm-1): 3528.2, 3328.5, 3223.7, 3178.0, 3090.5, 3038.5,
2917.7, 2849.4, 2214.1, 1900.6, 1733.8, 1607.0, 1581.6, 1560.0,
1486.9, 1415.3, 1353.2, 1257.2, 1220.2, 1117.6, 1075.0, 978.2,
1
+
(Q-TOF) m/z: calculated for C24H21N4 (M+H+): 365.1766,
821.5, 771.3, 673.2, 650.9; H NMR(500 MHz, DMSO-d6), δ:
7.37 (s, 1H), 7.46 (t, 1H, J = 8.5 Hz), 7.62 (s, 1H), 7.89 (s, 1H),
7.98 (d, 1H, J = 8.0 Hz), 8.63 (s, 1H), 8.74 (s, 1H), 9.97 (s, 1H);
13C NMR (125 MHz, DMSO-d6 ), δ: 88.60, 117.01, 117.49-
117.66 (d, JC,F = 21.25 Hz), 120.05-120.10 (d, JC,F= 6.25 Hz),
120.20, 121.11, 125.81, 126.04-126.08 (d, JC,F = 5.0 Hz), 127.27,
131.76, 135.38, 137.06, 147.52, 150.77, 153.71, 154.83-156.78
(d, JC,F = 243.75 Hz); HR-ESI (Q-TOF) m/z: calculated for
C16H8BrClFN3+ (M+H+): 377.9632, found: 377.9632.
found:365.1763.
4.1.5. 6-(4-(4-Methylpiperazin-1-yl)phenyl)-N-(3-
chloro-4-fluoro-phenyl)-quinoline-3-carbonitrile-4-
amine (CQ1).
500 milligrams (1.3 mmol) of 6-bromo-N-(3-chloro-4-fluoro-
phenyl)-quinoline-3-carbnitrile-4-amine, 450 mg (2.0 mmol) of
4-(4-Methylpiperazin-1-yl)phenylboronic acid , 50 mg of PPh3,
10 mL of Et3N, and 10 mL of DMF were placed in a 50 mL
Schlenk flask with a stirrer bar under nitrogen purge. The
reaction mixture was degassed for a further 30 min under a slow
stream of nitrogen, at which point 50 mg of Pd(OAc)2 was added.
The reaction mixture was heated at 80 oC for 24 h, cooled,
filtered through filter paper and eluted solution was added into 50
mL of H2O directly, and ethyl acetate was poured onto funnel to
wash the residues. Water and ethyl acetate solution mixture went
through extraction with three times with 200 mL of ethyl acetate
each. The organic layer was dried with MgSO4, filtered, and then
concentrated under reduced pressure. It was purified over silica
(100 % EtOAc to 100% MeOH), followed by crystallization in
methanol, ethyl acetate multiple times to afford light yellow
4.1.3. Synthesis of 6-bromo-N-(3-chloro-4-((3-
fluorophenyl)methoxy)phenyl)-quinoline-3-
carbnitrile-4-amine (4).
1.0 grams of 1 (3.8 mmol), 1.38 g (5.5 mmol) of 3-chloro-4-
[(3-fluorophenyl)methoxy]-benzenamine, and 2-propanol were
o
added, vigorously stirred and heated at 80 C, overnight with a
condenser. After 24 hours, it was cooled to room temperature,
concentrated under reduced pressure, 50 mL of 1M KOH
solution was added and stirred for 20 min, 100 mL of EtOAc was
added, then organic solution was extracted three times through
separation funnel and concentrated. It was crystallization from
methanol, ethyl acetate to afford yellow powder of 4 (647 mg,
35.4 %); mp: 208-210 C
; IR νmax(cm-1):3366.4, 3074.0, 2923.7,
̊
powder of CQ1 (145.2 mg, 23.7 %); mp: 220-222 C
̊
(dec); IR
max(cm-1): 3571.1, 3331.2, 3170.1, 2965.5, 2937.8, 2888.0,
2836.3, 2811.5, 2211.6, 1867.8, 1766.6, 1663.0, 1602.4, 1587.4,
2205.9, 1870.6, 1739.3, 1606.5, 1582.9, 1559.9, 1500.7, 1447.1,
1353.5, 1311.4, 1288.5, 1268.8, 1221.9, 1117.7, 1059.4, 1027.2,
979.4, 894.3, 805.7, 745.8, 682.6, 610.5; 1H NMR(500 MHz,
DMSO-d6), δ: 5.28 (s, 2H), 7.17 (t, 1H, J = 8.0 Hz), 7.26-7.34
(m, 4H), 7.45-7.50 (m, 2H), 7.85 (d, 1H, J = 9.0 Hz), 7.97 (d, 1H,
J = 7.5 Hz), 8.57 (s, 1H), 8.76 (s, 1H), 9.87 (s, 1H); 13C NMR
ν
1533.9, 1494.3, 1414.9, 1361.1, 1254.5, 1128.5, 1077.6, 1009.2,
1
973.1, 844.4, 773.8, 655.2; H NMR(500 MHz, DMSO-d6), δ:
2.22 (s, 3H), 2.46 (broad s, 4H), 3.22 (broad s, 4H), 7.05 (d, 2H,
J = 7.5 Hz), 7.38 (s, 1H), 7.47 (t, 1H, , J = 9.0 Hz), 7.61 (s, 1H),
7.76 (d, 2H, J = 7.5 Hz), 7.93 (d, 1H, J = 8.5 Hz), 8.13 (d, 1H, J
= 8.0 Hz), 8.55 (s, 1H), 8.67 (s, 1H), 10.00 (s, 1H); 13C NMR
(125 MHz, DMSO-d6 + TFA), δ: 42.56, 45.65, 52.59, 88.10,
116.50, 117.15, 117.55-117.72 (d, JC,F= 21.25 Hz), 119.38,
119.85, 120.10-120.25 (d, JC,F= 18.75 Hz), 126.36-126.42 (d,
JC,F= 7.5 Hz), 127.62, 128.26, 129.41, 130.17, 131.08, 137.15,
(125 MHz, DMSO-d6 ), δ: 69.86, 87.72, 114.42-114.60 (d, JC,F
=
22.5 Hz), 114.75, 115.10-115.27 (d, JC,F= 21.25 Hz), 117.09,
119.94, 120.71, 122.09, 123.79-123.81 (d, JC,F= 2.5 Hz), 125.70,
126.22, 127.89, 130.98-131.04 (d, JC,F= 7.5 Hz), 132.01, 132.99,
135.27, 139.82-139.88 (d, JC,F= 7.5 Hz), 147.72, 151.12, 152.47,
154.09, 161.69-163.63 (d, JC,F= 242.5 Hz); HR-ESI (Q-TOF)
m/z: calculated for C23H15BrClFN3O+ (M+H+): 484.0051, found:
484.0051.
138.48, 146.65, 149.87, 152.02, 152.34, 154.94-156.89 (d, JC,F
243.75 Hz); HR-ESI (Q-TOF) m/z: calculated for C27H24ClFN5
(M+H+): 472.1704, found: 472.1673.
=
+
4.1.4. Synthesis of 6-(4-(Dimethyl-amino)phenyl)-
N-(phenyl)-quinoline-3-carbonitrile-4-amine (CQ).
4.1.6. 6-(4-(4-Methylpiperazin-1-yl)phenyl)-N-(3-
chloro-4-((3-fluorophenyl)methoxy)phenyl)-
quinoline-3-carbonitrile-4-amine (CQ2).
0.8
carbnitrile-4-amine,
g
(2.5 mmol) of 6-bromo-N-(phenyl)-quinoline-3-
650 mg (3.2 mmol) of 4-
(dimethylamino)phenylboronic acid hydrochloride, 80 mg of
PPh3, 10 mL of Et3N, and 10 mL of DMF were placed in a 50 mL
Schlenk flask with a stirrer bar under nitrogen purge. The
500 milligrams (1.0 mmol) of 4, 450 mg (2.0 mmol) of 4-(4-
Methylpiperazin-1-yl)phenylboronic acid, 50 mg of PPh3, 10 mL
of Et3N, and 10 mL of DMF were placed in a 50 mL Schlenk

flask with a stirrer bar under nitrogen purge. The reaction mixture
was degassed for a further 30 min under a slow stream of
nitrogen, at which point 50 mg of Pd(OAc)2 was added. The
false-colored green channel with the following parameters:
excitation was via the 488 nm laser line and emission was
captured between 500 and 550 nm. Alexfluor 546 was imaged in
a false-colored red channel with the following parameters:
excitation was via the 561 nm laser line and emission was
captured between 575 and 600 nm. Images were processed in
Fiji.33
o
reaction mixture was heated at 80 C for 24 h, cooled, filtered
through filter paper and eluted solution was added into 50 mL of
H2O directly, and ethyl acetate was poured onto funnel to wash
the residues. Water and ethyl acetate solution mixture went
through extraction with three times with 200 mL of ethyl acetate
each. The organic layer was dried with MgSO4, filtered, and then
concentrated under reduced pressure. It was purified over silica
(100 % EtOAc to 100% MeOH), followed by crystallization from
methanol, ethyl acetate multiple times to afford yellow powder of
For time course measurements, the focal plane was first
established using the phase contrast image and checked against
the autofluorescence of the cells under illumination from the 405
nm laser line. Laser intensity and photodetector gain were set by
imaging BT474 cells pretreated with either CQ1 (2 µM, 10
minutes). All data time course data was collected on the same
day with the same imaging settings. If z-drift was observed
during the course of a timed experiment, the experiment was
halted, the sample brought quickly (< 10 sec) into focus and the
experiment was resumed. If z-drift persisted or longer refocusing
times were required, the experiment was aborted and data
collection began anew on a separate sample.
CQ2 (74.0 mg, 12.8 %); mp: 254-256 C
(dec); IR νmax(cm-1):
̊
3353.9, 3071.4, 2970.2, 2937.4, 2847.0, 2804.1, 2199.9, 1870.6,
1605.7, 1583.1, 1514.2, 1443.3, 1359.9, 1299.3, 1264.8, 1215.1,
1161.9, 1060.5, 1030.4, 1009.4, 953.2, 889.5, 784.5, 748.6,
1
690.0, 657.3; H NMR(500 MHz, DMSO-d6), δ: 2.23 (s, 3H),
2.47 (broad s, 4H), 3.23 (broad s, 4H), 5.29 (s, 2H), 7.06 (d, 2H,
J = 9.0 Hz), 7.17 (t, 1H, , J = 8.0 Hz), 7.28-7.35 (m, 4H), 7.45 (q,
1H, J = 8.0 Hz), 7.52 (s, 1H), 7.77 (d, 2H, J = 9.0 Hz), 7.92 (d,
1H, J = 8.5 Hz), 8.14 (d, 1H, J = 7.5 Hz), 8.50 (s, 1H), 8.69 (s,
1H), 9.88 (s, 1H); 13C NMR (125 MHz, DMSO-d6), δ: 46.26,
48.07, 54.97, 69.86, 87.26, 114.42-114.60 (d, JC,F= 22.5 Hz),
114.80, 115.10-115.27 (d, JC,F= 21.25 Hz), 115.76, 117.57,
118.82, 119.56, 122.11, 123.79-123-81 (d, JC,F= 2.5 Hz), 126.24,
4.3. Quantum Chemical Calculations
Quantum chemical calculations were performed in Spartan.
The structures of QA and CQ were optimized in the ground state
(DFT 6-31G*) with the N-phenyl ring locked with dihedral
angles matching the crystal structure of the ERBB1-bound
inhibitor, gefitinib (PDB entry: 4WKQ).
127.94, 128.04, 129.01, 130.33, 130.50, 130.98-131.05 (d, JC,F=
8.75 Hz), 133.49, 138.39, 139.85-139.91 (d, JC,F= 7.5 Hz),
147.74, 151.25, 151.85, 152.32, 152.82, 161.70-163.64 (d, JC,F
=
242.5 Hz); HR-ESI (Q-TOF) m/z: calculated for C34H30ClFN5O+
(M+H+): 578.2123, found: 578.2117.
Acknowledgments
This work was supported by the National Cancer Institute
Innovative Molecular Analysis Technologies Program,
CA182341-03 (J. N. W. and R. L.). J. N. W. also acknowledges
seed funding from the American Cancer Society (98-277-07)
R.L. also acknowledges the support of National Cancer Institute,
CA98881-05 and the Braman Family Breast Cancer Institute.
4.2. Biological assays and imaging
4.2.1. Cell culture and reagents
MCF7 and BT474 cells were cultured as previously described6
in sterile T-35 or T-75 flasks. Cells were maintained in RPMI
containing 10% dialyzed FBS, penicillin (100 units/mL) and
streptomycin (0.01%) solution under a humidified 5% CO2
atmosphere.
References and notes
1. Chabner, B. A.; Roberts, T. G. Nat. Rev. Cancer 2005, 5, 65-72.
2. Bagnyukova, T.; Serebriiskii, I. G.; Zhou, Y.; Hopper-Borge, E.
A.; Golemis, E. A.; Astsaturov, I. Cancer Biol. Ther. 2010, 10,
839-853.
4.2.2. Ligand stimulation assays
MCF7 cells were seeded with equal quantity (200
thousand/well) in six-well plates. After 48 hours, cells were pre-
treated with the small molecule inhibitors of various
concentrations for 1 hour before induction by neuregulin
(NRGb1, 30 nM, 30 minutes). Cell lysates were generated
immediately by SDS lysis. Equal aliquots were subjected to SDS-
PAGE and Western blot analysis. ERBB2 phosphorylation was
evaluated for Tyr1139 located close to the extreme cytoplasmic
C-terminus of the receptor (validated by pan-TyrP detection
(4G10)). The signal obtained for pTyr1139 relative to the ERBB2
receptor levels was determined as the relative receptor
phosphorylation.
3. McGranahan, N.; Swanton, C. Cancer Cell 2015, 27, 15-26.
4. Sicard, R.; Dhuguru, J.; Liu, W.; Patel, N.; Landgraf, R.; Wilson,
J. N. Bioorg. Med. Chem. Lett. 2012, 22, 5532-5535.
5. Dhuguru, J.; Liu, W.; Gonzalez, W. G.; Babinchak, W. M.;
Miksovska, J.; Landgraf, R.; Wilson, J. N. J. Org. Chem. 2014,
79, 4940-4947.
6. Lee, H.; Liu, W.; Brown, A. S.; Landgraf, R.; Wilson, J. N. Anal.
Chem. 2016, 88, 11310-11313.
7. Bridges, A. J. Chem. Rev. 2001, 101, 2541-2571.
8. Rewcastle, R.W.; Denny, W. A.; Bridges, A. J.; Zhou, H.; Cody,
D. R.; McMichael, A.; Fry, D.W. J. Med. Chem. 1995, 38, 3482-
3487.
9. Levinson, N. M.; Boxer, S. G. PLoS One, 2012, 7, e29828.
10. Wilson, J. N.; Liu, W.; Brown, A. S.; Landgraf, R. Org. Biomol.
Chem. 2015, 13, 5006-5011.
11. Trummer, B. J.; Iyer, V.; Balu-Iyer, S. V.; O’Connor, R.;
Straubinger, R. M. J. Pharm. Sci., 2012, 101, 2763-2776.
12. Wissner, A.; Berger, D. M.; Boschelli, D. H.; Floyd, M. B.;
Greenberger, L. M.; Gruber, B. C.; Johnson, B. D.; Mamuya, N.;
Nilakantan, R.; Reich, M. F.; Shen, R.; Tsou, H. -R.; Upeslacis,
E.; Wang, Y. F.; Biqi Wu, B.; Fei Ye, F.; Zhang, N. J. Med.
Chem. 2000, 43, 3244-3256.
4.2.3. Confocal imaging
For imaging, cells were seeded at a density of 105 cell/cm2 in
96 microwell plates or on glass coverslips. Cells maintained a
normal morphology during the course of the experiments
(maximum of 1 h) and remained adhered to the imaging plate or
coverslip. Single photon imaging was performed on a Leica SP5
confocal microscope housed within the UM Biology Imaging
Core Facility. For the imaging of CQ1 a channel was created
with excitation via a 405 nm laser line with emission between
500 and 600 nm captured and false-colored blue. Cells were
labeled an anti-ERBB2 antibody directed towards the
extracellular domain of ERBB2 followed by either Alexafluor
488 or 546 secondary antibody. Alexfluor 488 was imaged in a
13. Wissner, A.; Mansour, T. S. Arch. Pharm. Chem. Life Sci. 2008,
341, 465-477.
14. Pannala, M.; Kher, S.; Wilson, N.; Gaudette, J.; Sircar, I.; Zhang,
S. -H.; Bakhirev, A.; Yang, G.; Yuen, P.; Gorcsan, F.; Sakurai, N.;
Barbosa, M.; Cheng, J. -F. Bioorg. Med. Chem. Lett. 2007, 17,
5978-5982.

15. Reichardt, C. Angew. Chem. Int. Ed. Engl. 1979, 18, 98-110.
16. Zhao, C.; An, J.; Zhou, L.; Fei, Q.; Wang, F.; Tan, J.; Shi, B.;
Wang, R.; Guo, Z.; Zhu, W. -H. Chem. Commun. 2016, 52, 2075-
2078.
26. Citri, A.; Alroy, I.; Lavi, S.; Rubin, C.; Xu, W.; Grammatikakis,
N.; Patterson, C.; Neckers, L.; Fry, D. W.; Yarden, Y. EMBO J.
2002, 21, 2407-2417.
27. Sánchez-Martín, M.; Pandiella, A. Int. J. Cancer. 2012, 131, 244-
252.
17. Zhang, Y. –M.; Cockerill, S.; Guntrip, S. B.; Rusnak, D.; Smith,
K.; Vanderwall, D.; Wood, E.; Lackey, K. Bioorg. Med. Chem.
Lett., 2004, 14, 111-114.
28. Zhang, Y.; Zhang, J.; Liu, C.; Du, S.; Feng, L.; Luan, X.; Shi, Y.;
Wang, T.; Wu, Y.; Cheng, W.; Meng, S.; Li, M.; Liu, H. Cancer
Lett., 2016, 382, 176-185.
18. Knight, L. A.; Di Nicolantonio, F.; Whitehouse, P.; Mercer, S.;
Sharma, S.; Glaysher, S.; Johnson, P.; Cree, I. A. BMC Cancer,
2004, 4, 83-90.
29. Citri, A.; Gan, J.; Mosesson, Y.; Vereb, G.; Szollosi, J.; Yarden,
Y. EMBO Rep., 2004, 5, 1165-1170.
19. Jain, A.; Tindell, C. A.; Laux, I.; Hunter, J. B.; Curran, J.; Galkin,
A.; Afar, D. E.; Aronson, N.; Shak, S.; Natale, R. B.; Agus, D. B.
Proc. Nat. Acad. Sci. USA, 2005, 102, 11858-11863.
20. Piechocki, M. P.; Yoo, G. H.; Dibbley, S. K.; Lonardo, F. Cancer
Res. 2007, 67, 6825-6843.
30. Warren, C. M.; Landgraf, R. Cell. Signal. 2006, 18, 923-933.
31. Bertelsen, V.; Stang, E. Membranes 2014, 4, 424-446.
32. Sigismund, S.; Confalonieri, S.; Ciliberto, A.; Polo, S.; Scita, G.;
Di Fiore, P. P. Physiol. Rev. 2012, 92, 273-366.
33. Schindelin, J; Arganda-Carreras, I.; Frise, E.; Kaynig, V.; Longair,
M.; Pietzsch, T.; Preibisch, S.; Rueden, C.; Saalfeld, S.; Schmid,
B.; Tinevez, J. –Y.; White, D. J.; Hartenstein, V.; Eliceiri, K.;
Tomancak, P.; Cardona, A. Nature Meth. 2012, 9, 676.
21. Hendriks, B. S.; Opresko, L. K.; Wiley, H. S.; Lauffenburger, D.
J. Biol. Chem. 2003, 278, 23343-23351.
22. Hirata, A.; Hosoi, F.; Miyagawa, M.; Ueda, S.-i.; Naito, S.; Fujii,
T.; Kuwano, M.; Ono, M. Cancer Res. 2005, 65, 4253-4260.
23. Paez, J. G.; Jänne, P. A.; Lee, J. C.; Tracy, S.; Greulich, H.;
Gabriel, S.; Herman, P.; Kaye, F. J.; Lindeman, N.; Boggon, T. J.;
Naoki, K.; Sasaki, H.; Fujii, Y.; Eck, M. J.; Sellers, W. R.;
Johnson, B. E.; Meyerson, M. Science 2004, 304, 1497-1500.
24. Björkelund, H.; Gedda, L.; Barta, P.; Malmqvist, M.; Andersson,
K. PLoS One, 2011, 6, e24793.
Supplementary Material
1
Please see H and 13C NMR spectra for 2-4, CQ, CQ1 and
CQ2 as well as atomic coordinates from DFT calculations for
CQ and QA in accompanying file.
25. Björkelund, H.; Gedda, L.; Barta, P.; Malmqvist, M.; Andersson,
K. Mol. Clin. Oncol. 2013, 1, 343-352.