Discoipyrroles A-D: Isolation, structure determination, and synthesis of potent migration inhibitors from Bacillus hunanensis

Article abstract of DOI:10.1021/ja403412y

Discoidin domain receptor 2 (DDR2) is a receptor tyrosine kinase involved in a variety of cellular response pathways, including regulation of cell growth, proliferation, and motility. Using a newly developed platform to identify the signaling pathway/molecular target of natural products, we identified a family of alkaloid natural products, discoipyrroles A-D (1-4), from Bacillus hunanensis that inhibit the DDR2 signaling pathway. The structure of 1-4, determined by detailed two-dimensional (2D) NMR methods and confirmed by X-ray crystallographic analysis has an unusual 3H-benzo[d]pyrrolo][1,3]oxazine-3,5- dione core. Discoipyrroles A-D potently inhibit DDR2 dependent migration of BR5 fibroblasts and show selective cytotoxicity to DDR2 mutant lung cancer cell lines (IC₅₀ 120-400 nM). Examination of the biosynthesis has led to the conclusion that the discoipyrroles are formed through a nonenzymatic process, leading to a one-pot total synthesis of 1.

Full text of DOI:10.1021/ja403412y

Article
pubs.acs.org/JACS
Discoipyrroles A−D: Isolation, Structure Determination, and
Synthesis of Potent Migration Inhibitors from Bacillus hunanensis
Youcai Hu,†,§ Malia B. Potts,‡,§ Dominic Colosimo,† Mireya L. Herrera-Herrera,‡ Aaron G. Legako,†
Muhammed Yousufuddin,∥ Michael A. White,‡ and John B. MacMillan*,†
‡
§
†Department of Biochemistry, Department of Cell Biology, Simmons Comprehensive Cancer Center, University of Texas
Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, Texas 75390, United States
∥Department of Chemistry, Center for Nanostructured Materials, University of Texas at Arlington, Arlington, Texas 76019, United
States
S
* Supporting Information
ABSTRACT: Discoidin domain receptor 2 (DDR2) is a
receptor tyrosine kinase involved in a variety of cellular
response pathways, including regulation of cell growth,
proliferation, and motility. Using a newly developed platform
to identify the signaling pathway/molecular target of natural
products, we identified a family of alkaloid natural products,
discoipyrroles A−D (1−4), from Bacillus hunanensis that
inhibit the DDR2 signaling pathway. The structure of 1−4,
determined by detailed two-dimensional (2D) NMR methods
and confirmed by X-ray crystallographic analysis has an unusual 3H-benzo[d]pyrrolo][1,3]oxazine-3,5-dione core. Discoipyrroles
A−D potently inhibit DDR2 dependent migration of BR5 fibroblasts and show selective cytotoxicity to DDR2 mutant lung
cancer cell lines (IC50 120−400 nM). Examination of the biosynthesis has led to the conclusion that the discoipyrroles are
formed through a nonenzymatic process, leading to a one-pot total synthesis of 1.
We have developed a natural products fraction library based
on a collection of marine-derived bacteria.8 The library of 6500
fractions has been utilized in a number of target-specific and
phenotypic HTS screens in the field of oncology. Screening for
cytotoxicity against relevant human cancer cell lines continues
to be a successful screening strategy but is limited by a lack of
mechanistic information.9 We have developed an assay that
augments phenotypic screening, producing a FUnctional
SIgnature-based ONtology (FUSION) map that links bioactive
compounds to the molecular target or regulatory network they
interact with in a mammalian cell (Figure 1a).10
This methodology utilizes pattern matching of a gene-
expression profile generated by exposing HCT-116 cells to a
single siRNA, miRNA, or natural product fraction. As the
molecular target of an individual siRNA is known, the
mechanism of action of natural product fractions can be
deduced by similar gene expression profiles (Figure S1, assay
protocol in SI). Using the FUSION map approach we
identified a natural product fraction from marine-derived
Bacillus hunanensis strain SNA-048 that provided a functional
signature identical to a siRNA knockout of DDR2 in HCT-116
cells (Figure 1b).10 Based on the functional signature, we
showed fraction SNA-048-7 from B. hunanensis inhibits DDR2
dependent migration of BR5 fibroblasts through a 2D-Matrigel
matrix at 1 μg/mL (Figure 1c).2a,11
INTRODUCTION
■
Discoidin domain receptors (DDR1 and DDR2) are cell-
surface receptors, which are activated by binding collagen,
which results in downstream signaling responses involved in
regulation of cell growth, differentiation, and motility.1
Prolonged stimulation of DDR2 has been associated with the
up-regulation of matrix metalloproteinase-1 expression. DDR2
also plays an important role in mediating fibroblast migration
and proliferation by matrix metalloproteinase-2 dependent
mechanisms.2 DDR2 has been implicated in obstructive
diseases of blood vessels by mediation of collagen turnover in
vascular smooth muscle cells.3
Recent reports have implicated DDR2 to play a role in a
variety of cancers.4 Sequencing efforts by Meyerson and co-
workers identified mutations in the DDR2 kinase gene in
∼4.0% of human squamous cell lung cancer (SCC) and went
on to establish that DDR2 mutant SCC cell lines were
selectively killed by knockdown of DDR2 by RNAi or by
treatment with the pan-kinase inhibitor dasatinib.5 This has
been further validated by successful treatment of tumors
established with DDR2 derived cell lines with dasatanib. In
breast cancer, increased collagen deposits in the stroma leads to
enhanced metastases.6 In that context, DDR2 has been
demonstrated to promotes breast cancer cell invasion and
migration in vitro and metastasis in vivo.7 With the emerging
importance of this pathway, there are still a number of
unknowns in the DDR2 signaling pathway and very few tools
available to interrogate the system.
Received: April 5, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
RESULTS AND DISCUSSION
■
Isolation and Structural Elucidation. B. hunanensis strain
SNA-048 was isolated from a sediment sample collected at
Galveston Bay, TX and selected using a humic acid based
media. Fermentation and extraction were carried out using
standard procedures.8 Purification of metabolites by sequential
flash C18, Sephadex LH-20, and reversed phase C18 HPLC led
to 1 (2.0 mg), 2 (0.8 mg), 3 (1.0 mg), and 4 (0.4 mg).
Discoipyrrole A (1) was obtained as a yellow solid and was
determined to have a molecular formula of C27H23NO5, based
on high resolution electrospray ionization mass spectrometry
(HR-ESIMS) [M + H]+ m/z 442.1652 and interpretation of
NMR data. The UV spectrum of 1 exhibited absorption bands
at 396, 324, 263, and 214 nm, indicative of a highly conjugated
Figure 1. (a) Overview of the FUSION assay to link natural products
molecular target or pathway. HCT-116 cells were individually
subjected to chemical and genetic perturbations and analyzed for
their gene expression of a set of select genes. (b) Hierarchical
clustering of natural product fractions with discoidin domain receptor
2 (DDR2) based on gene expression of six genes. (c) 2D-Matrigel cell
migration assay with BR5 fibroblasts, platelet derived growth factor,
and a natural product fraction from Bacillus hunanensis SNA-048
(fraction 7) at 1 μg/mL.
1
system. The H NMR spectrum (Table S1, dimethyl sulfoxide
DMSO-d6) revealed signals indicative of disubstituted aromatic
rings and a small ketide fragment, while the 13C NMR spectrum
contained two downfield signals indicative of carbonyls.
Because of the 11 quaternary carbons the structure elucidation
of 1 proved a challenge. Detailed analysis of 1D and 2D NMR
data allowed assignment of the simple substructures I−IV
(Figure 3a). The full details of the structural elucidation can be
found in the Supporting Information.
Here we report the BR5 2D-Matrigel migration assay-guided
isolation of discoipyrroles A−D (1−4) (Figure 2), a family of
Figure 3. Structure elucidation details of 1−4. (a) Substructures of 1
(I−IV) with COSY, HMBC correlations. (b) ORTEP drawing of
crystal structure of (−)-1a. (c) Key NMR correlations for 3.
Figure 2. Structures of 1−4 and biosynthetic precursors 5 and 6.
Briefly, partial structures I (C-4 through C-8) and II (C-3, C-
9 through C-12) were assigned as 1,4-disubstituted phenyl rings
with an additional sp2 quaternary carbon at benzylic positions
(C-4, C-3). The nature of the quaternary carbon substituent at
the benzylic positions was not readily apparent from initial
investigation of the NMR data. However, the difference in
chemical shifts between H-6 and H-7 of ∼0.3 ppm, provided
evidence that C-4 was part of an olefin. This is predicated on
considerable literature NMR data for p-hydroxycinnamates;
where there is heteroatom substitution in the α or β position,
the Δδ value is always ∼0.3 ppm (Figure S2, S3). A similar Δδ
of ∼0.3 between H-10 and H-11 supports the structure of II.11
Partial structure III (C-17 through C-22) was readily
determined to be an anthranilate moiety, while the aliphatic
partial structure IV (C-13 through C-16) was established by
COSY correlations and key HMBC correlations.
polycyclic alkaloids comprised of a 3H-benzo[d]pyrrolo][1,3]-
oxazine-3,5-dione core (for 1, 2, and 4) and adorned with aryl
groups. The discoipyrroles are the first examples of a natural
product containing this highly functionalized alkaloid core. In
addition to 1−4, we have isolated hydroxysattabacin (5)12 and
the related enol (6) (Figure 2), which provide insight into the
biosynthetic origin of 1−4. The biosynthetic understanding
ultimately led to a one-step total synthesis of 1. Furthermore,
we report inhibition of DDR2 dependent migration of BR5
fibroblasts by discoipyrroles and further demonstrate their
dose-dependent effect on post-translational modifications of
DDR2. Evaluation of 1−4 against nonsmall cell lung cancer
(NSCLC) cell-lines showed selective cytotoxicity toward
DDR2 mutant cell-lines with IC50 values ∼200 nM.
B
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Based on 13C chemical shifts we could assign C-2 through C-
4 as an α,β-unsaturated ketone, which, in combination with I−
IV, account for all of the carbon and hydrogens and 15 of the
17 double bond equivalents, suggesting 1 contained two
additional rings. The distinct chemical shift of the C-1
quaternary carbon (δC 90.3) confirmed the presence of a
hemiaminal functionality, which allowed us to propose the
structure for 1 as depicted, containing a 3H-benzo[d]pyrrolo]-
[1,3]oxazine-3,5-dione core (Figure S4). A zero [α]D was
measured for 1, suggesting that the C1 quaternary stereocenter
is racemic, as will be addressed later.
To verify the structure, a number of attempts to crystallize
( )-1 were made with no success. Ultimately, ( )-1 was
separated via chiral-phase HPLC to give the individual
enantiomers (Figure S5), and (−)-1 converted to the bis-p-
bromobenzoate derivative (−)-1a (p-bromobenzoyl chloride,
Et3N, CH2Cl2). Crystallization of (−)-1a from 10:1
MeOH:CHCl3 resulted in yellow prisms, which gave an X-ray
crystal structure (Figure 3b) with 0.84 Å resolution and a Flack
parameter of 0.001 (6). Thus, we were able to confirm the
NMR assignment of 1 and establish that (−)-1a has the 1R
configuration.
The structures of the three remaining discoipyrroles 2−4
were assigned by NMR analysis and comparison to 1 (Figure
S6). Full details on the structure assignment can be found in
Tables S2−S4 and the Supporting Information. Of note,
discoipyrrole C (3), lacks the C-17 to C-23 anthranilate moiety,
instead is replaced by NH3. Acquisition of the 1H NMR of 3 in
DMSO-d6 showed the exchangeable −NH (δH 7.96) of the C-1
hemiaminal, which gave HMBC correlations to C-1, C-2, C-3,
C-4, C-5, and C-13, confirming the pyrrol-3-one core of 3
(Figure 3c). This provided further NMR evidence for the 3H-
benzo[d]pyrrolo][1,3]oxazine-3,5-dione core of 1.
Proposed Biosynthesis and Precursor-Driven Biosyn-
thesis of 1 and Analogues. A proposed biosynthetic
pathway (Figure 4) accounts for the racemic nature of the
discoipyrroles as well as the ability to incorporate different
amines into the core structure. In the proposed pathway
hydroxysattabacin (5), isolated from the nonpolar partition of
the crude extract, is oxidized to give enol 6, which was also
isolated from B. hunanensis (Table S5). This is followed by an
aldol reaction with p-hydroxybenzaldehyde and subsequent
oxidation to give the triketone 7. Following keto−enol
tautomerization of 7, imine formation with anthranilic acid13
(or NH3 for 3) would give diketoenamine 8, which is followed
by a nonenzymatic hemiaminal formation to give 9 as a
racemate. We have isolated a small amount of 9 under stringent
isolation conditions, utilizing only LH20 size exclusion
chromatography, eliminating any trace acid from the
purification. However, upon C18 chromatography 9 undergoes
rapid dehydration-lactonization to give 1.
Figure 4. Proposed biosynthetic pathway to 1.
dehyde, p-bromobenzaldehyde, and p-methoxybenzaldehyde.
Separate additions of each precursor (1g·L−1) to A1 media and
shake fermentation resulted only in oxidations to the
corresponding benzoic acids; however, addition of p-hydrox-
ybenzaldehyde-1-13C (1 g·L−1) to solid agar A1 media resulted
in production of 4-[13C]-discoipyrrole A (10), as analyzed by
liquid chromatography/mass spectrometry (LC-MS) and NMR
(Scheme 1).
Not surprisingly, this verified that C-4 through C-8 originates
from p-hydroxybenzaldehyde. Similarly, we converted 5 to p-
[13CH3]-methoxysattabacin (11) and added to A1 liquid media
(0.250 g·L−1) and after 7 days of fermentation, observed
production of a compound with a mass corresponding to 12-O-
Scheme 1. Precursor-Directed Biosynthesis of 13C Labeled
Discoipyrrole Analogues 10 and 12
Discoipyrrole B originates from incorporation of p-
hydroxybenzaldehyde with sattabacin (lacking the phenoxyl
group), while discoipyrrole D may arise from an elaborated
anthranilate group or from nucleophilic attack of C-24 onto a
suitably substituted indole analogue. It should be noted that
another reasonable biosynthetic pathway can be proposed that
changes the order of events in a Mannich-type reaction,
whereby anthranilic acid and p-hydroxybenzaldehyde first form
an imine prior to nucleophilic attack by 6.14
Brief investigations of the biosynthetic hypotheses were
carried out using a precursor-directed approach with various
substituted benzaldehyde precursors, including p-nitrobenzal-
C
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
[13CH3]-methyl discoipyrrole A (12), which was verified by
ESIMS and NMR (Table S6).
dehydration, and two oxidations to generate 1. From our
biosynthetic proposal it is clear that the last two steps,
hemiaminal formation and dehydration, can be readily
rationalized. The oxidation steps on the other hand require
assistance from the media. The “spent” media does contain a
number of metal species, including Fe, Co, Mn, and Ni salts
and are capable of promoting the transition-metal mediated
benzylic oxidation.15 A second possibility is that the C-4
benzylic position is highly prone to air oxidation.
We further explored the possibility of carrying out a one-pot
transformation of the three precursors to 1 in a variety of
organic solvents, such as MeOH, MeOH/H2O, and DMSO and
at various pH ranges (2−11). Optimized conditions that lead to
a 20% conversion to 1 were 1% trifluoroacetic acid in DMSO at
45 °C for 30 h, but also produced a number of side-products,
including 13 (Scheme 2). Similarly, utilizing p-hydroxybenzal-
The foregoing precursor-directed experiments serve not only
to validate our biosynthetic proposal, but also as a means to
generate structural diversity for optimization of biological
activity. Therefore we decided to carry the precursor directed
biosynthetic studies one step further based on the observation
that the only enzymatic steps involved in our proposed
biosynthesis of 1 involve oxidations of an α-hydroxy ketone and
a benzylic alcohol. We hypothesized it would be possible to
generate the entire discoipyrrole framework by providing the
three substrates (5, p-hydroxybenzaldehyde, anthranilic acid) to
cell-free “spent” A1 media (Figure 5a).
Scheme 2. Synthesis of 1 in Organic Solvent
dehyde-1-13C, 11 and anthranilic acid with 1% trifluoroacetic
acid (TFA) in DMSO at 45 °C for 30 h, we observed the
production of 4-[13C]-12-O-[13CH3]-methyl discoipyrrole A
(14), as characterized by MS and NMR (Table S7). It is likely
that DMSO and O2 are participating in the two oxidation steps
of the one-pot transformation, as when the reaction is carried
out with deoxygenated DMSO under a N2 atmosphere, no
transformation occurs.
We monitored the oxidation of 5 to 6 by NMR and LC-MS
in DMSO-d6 and 1% deuterated TFA and could observe a 3:1
ratio of 5 to 6 after 6 h (Figure S7). In addition to 6, we
observed conversion of 5 to p-hydroxybenzaldehyde (∼5%),
which presumably derives from tautomerization of 6 to give the
quinone, followed by nucleophilic addition of H2O and a
subsequent retro-aldol (Figure S8).
The combination of the “spent” media induced formation
and the DMSO and TFA promoted formation of 1 offer
complementary routes for generation of discoipyrrole ana-
logues for further evaluation.
Inhibition of DDR2 Signaling by 1−4. Based on our
initial prediction, we evaluated the ability of 1−4 to inhibit BR5
fibroblast migration in a 2D-Matrigel matrix at 1 μM, with 1
displaying the most potent inhibition at 95% followed by 2 and
4 with 85% inhibition and only modest inhibition by 3 at 50%
(Figure 6a, Figure S9). All four compounds exhibited no
cytotoxicity against BR5 fibroblasts up to 10 μM, ruling out the
possibility that migration inhibition was caused by reduced cell
viability (Figure 6a). As a positive control in these assays, we
measured the migration inhibition ability of the pan-kinase
inhibitor dasatinib, which has been shown to be an inhibitor of
the kinase domain of DDR2. As expected, dasatinib inhibited
cell migration at 0.1 μM, however, was modestly cytotoxic to
the BR5 fibroblasts at this concentration.16
Figure 5. (a) Production of 1 from three biosynthetic precursors in
“spent” media. (b) SIM chromatograms, monitoring at m/z z 440 [M
− H]− for purified 1 from B. hunanensis strain SNA-048; B. aquamaris
strain SNE-038; spent media from B. aquamaris strain SNE-038 with 5,
p-hydroxybenzaldehyde, and anthranilic acid after 5 days at 37 °C.
We wanted to use “spent” media because it contains a variety
of metal ions and cofactors that are produced in the course of
bacterial growth. Discoipyrrole producing strain B. hunanensis
was avoided to preclude background production of 1. Instead,
we chose media from cultures of B. aquamaris strain SNE-038
which was devoid of discoipyrroles or 5 (LC-MS analysis,
Figure 5b) and displayed, after a 7 day fermentation, a pH of
9.0, the same as spent media from B. hunanensis grown under
the same conditions. After removal of the cells by
centrifugation, the supernatant was heated to 90 °C for five
minutes and filtered through a 5KD filter to denature and/or
remove enzymes. The resulting spent media was dosed with of
5 (10 mg·L−1), benzaldehyde (8 mg·L−1), and anthranilic acid
(8 mg·L−1) at 37 °C and allowed to shake for 5 days.
Gratifyingly, LC-MS analysis of the ethyl acetate extract
revealed production of significant levels of 1 (Figure 5). This
remarkable one-pot transformation involves the formation of
one C−C bond, imine formation, hemiaminal formation,
Closer investigations of the direct effect of 1 on DDR2 was
carried out with human BR-5 fibroblast cells in culture with
indicated concentrations of 1 (DMSO control) for 18 h. Whole
D
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
second class of natural products that inhibit the DDR2
signaling pathway.
In addition to the structural novelty and biological activity,
the nonenzymatic production of the discoipyrrole family of
compounds is an example of a growing number of natural
products that involve nonenzymatic steps in their production.
Other examples of de novo nonenzymatic natural products
synthesis include jadomycin (DNA alkylator),20 elansolid
(antibiotic),21 and the ammosamides (quinone reductase 2
inhibitors, myosin modifiers).22 In these examples, the
nonenzymatic reaction pathway has been exploited to generate
analogues with enhanced biological activity. Nonenzymatic
multistep synthesis of analogues of 1 provide us an opportunity
to explore and discover analogues with enhanced potency and
pharmacokinetic attributes.
Figure 6. (a) Quantification of migration inhibition in a BR5 2D-
Matrigel fibroblast migration assay with 1 μM of 1−4 + platelet
derived growth factor, DMSO as a negative control, and 0.1 μM
dasatinib (0.1 μM) as a positive control. To ensure activity was not
due to cell death, cell viability of BR5 fibroblasts was measured by Cell
Titer-glo at the same concentration. (b) Dose-dependent effect of 1 on
DDR2 in BR5 fibroblasts.
ASSOCIATED CONTENT
* Supporting Information
■
S
General procedures, bioassay protocols, purification proceures,
data tables, and NMR spectra. This material is available free of
charge via the Internet at http://pubs.acs.org
cell lysate was collected, separated by electrophoresis through
an SDS-PAGE gel, transferred to a PVDF membrane, and
detected by anti-DDR2 antibody. In DMSO-treated cells,
DDR2 protein ran as a two bands of approximately 100 kDa
apparent size. Treatment with the initial active fraction SNA-
048-7 as well as 1 and 2 resulted in appearance of a third
DDR2-reactive band of approximately 130 kDa apparent size,
indicating modification of the DDR2 protein (Figure 6b). The
nature of the modification is unknown at this time, but is
consistent with glycosylation, based on electrophoretic
mobility.17 Further work is necessary to elucidate the full
mechanistic details for inhibition and to identify the molecular
target.
As part of our UTSW panel of nonsmall cell lung cancer
(NSCLC) lines, we have identified lines HCC366 as harboring
a DDR2 mutation. Measurement of cytotoxicity of 1 against
HCC366 gave an IC50 = 120 nM, while the cytotoxicity against
the DDR2 wild-type NSCLC cell-line A549 had a decrease in
cytotoxicity with an IC50 = 10 μM. The IC50 values for 2 and 4
are similar to 1 (0.190 μM and 0.275 μM against HCC366,
respectively; 8.8 μM and 13.4 μM against A549, respectively),
while the IC50 values for 3 are significantly reduced (0.7 μM
against HCC366 and 19.6 μM against A549, also see Table S8).
As a positive control, dasatinib had IC50 values of 170 nM
against HCC366 and 9.8 μM against A549, consistent with
previous reports.5,18
AUTHOR INFORMATION
Corresponding Author
john.macmillan@utsouthwestern.edu
■
Author Contributions
§These authors contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank S. Wang, B. Posner, and members of the
UTSW High-Throughput Screening Core Facility. This work
was funded by grants from the US National Institutes of Health
(R01 CA149833, RC2 CA148225 and U01 CA176284) and the
Welch Foundation (I-1689). J.B.M. is the Chilton/Bell Scholar
in Biochemistry. M.B.P. was supported by a Komen for the
Cure Postdoctoral Fellowship.
REFERENCES
■
(1) (a) Vogel, W.; Gish, G. D.; Alves, F.; Pawson, T. Mol. Cell 1997,
1, 13−23. (b) Shrivastava, A.; Radziejewski, C.; Campbell, E.; Kovac,
L.; McGlynn, M.; Ryan, T. E. Mol. Cell 1997, 1, 25−34.
(2) (a) Olaso, E.; Labrador, J.-P.; Wang, L.; Ikeda, K.; Eng, F. J.;
Klein, R.; Lovett, D. H.; Lin, H. C.; Friedman, S. L. J. Biol. Chem. 2002,
277, 3606−3613. (b) Leitinger, B. J. Biol. Chem. 2003, 278, 16761−
16769.
The only other natural product that has been shown to
inhibit DDR2 activity is actinomycin D. Although actinomycin
D has been shown to possess potent DNA intercalating activity
(<10 nM), when cells are treated with actinomycin D for short
periods (<8 h) there is significant inhibition of DDR2
activation by collagen.19 The mechanistic details suggest
actinomycin D disrupts the ability of collagen to bind to
DDR2 but does not interrupt DDR2 localization in the cell
membrane.
(3) Xu, L.; Servais, J.; Polur, I.; Kim, D.; Lee, P. L.; Chung, K.; Li, Y.
Arth. Rheum. 2010, 62, 2736−2744.
(4) Ford, C. E.; Lau, S. K.; Zhu, C. Q.; Andersson, T.; Tsao, M. S.;
Vogel, W. F. Br. J. Cancer 2007, 96, 808−814.
(5) Hammerman, P. S.; et al. Cancer Disc. 2011, 1, 78−89.
(6) Kauppila, S.; Stenback, F.; Risteli, J.; Jukkola, A.; Risteli, L. J.
Pathol. 1998, 186, 262−268.
(7) Zhang, K.; Corsa, C. A.; Ponik, S. M.; Prior, J. L.; Piwnica-
Worms, D.; Eliceiri, K. W.; Keely, P. J.; Longmore, G. D. Nat. Cell Biol.
2013, 15, 677−687.
CONCLUSIONS
■
(8) (a) Hu, Y.; Legako, A. G.; Espindola, A.-P. D. M.; MacMillan, J.
B. J. Org. Chem. 2012, 77, 3401−3307. (b) Pan, E.; Jamison, M.;
Yousufuddin, M.; MacMillan, J. B. Org. Lett. 2012, 14, 2390−2393.
(c) Hu, Y.; MacMillan, J. B. Org. Lett. 2011, 13, 6580−6583. (d) Hu,
Y.; Espindola, A.-P. D. M.; Stewart, N. A.; Wei, S.; Posner, B. A.;
MacMillan, J. B. Bioorg. Med. Chem. 2011, 19, 5183−5189. (e) Hu, Y.;
Wang, K.; MacMillan, J. B. Org. Lett. 2013, 15, 390−393.
Through the use of an unbiased screening strategy using the
recently developed FUSION platform, we were able to identify
inhibitors of the DDR2 pathway. DDR2 has emerged as an
important target in cancer therapy, particularly in a significant
percentage of squamous cell carcinomas, such as nonsmall cell
lung cancers. The discovery of the discoipyrroles represents a
E
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(9) Garrett, M. D.; Walton, M. I.; McDonald, E.; Judon, I.;
Workman, P. Prog. Cell Cycle Res. 2003, 5, 145−158.
(10) (a) Potts, M. B.; Kim, H. Y.; Fisher, K. W.; Hu, Y.; Carrasco, Y.;
Bulut, G. B.; Ou, Y.-H.; Herrera-Herrera, M. L.; Cubillos, F.;
Mendiratta, S.; Xiao, G.; Hofree, M.; Ideker, T.; Xie, Y.; Huang, L.
J.-S.; Lewis, R. E.; MacMillan, J. B.; White, M. A. Science Signaling,
accepted. (b) Lamb, J.; Crawford, E. D.; Peck, D.; Modell, J. W.; Blat,
I. C.; Wrobel, M. J.; Lerner, J.; Brunet, J.-P.; Subramanian, A.; Ross, K.
N.; Reich, M.; Hieronymus, H.; Wei, G.; Armstrong, S. A.; Haggarty,
S. J.; Clemons, P. A.; Wei, R.; Carr, S. A.; Lander, E. S.; Golub, T. R.
Science 2006, 313, 1929−1935.
(11) (a) Menezes, G. C.; Miron-Mendoza, M.; Ho, C.-H.; Jiang, H.;
Grinnell, F. Exp. Cell Res. 2008, 314, 3081−3091. (b) Grinnell, F.;
Rocha, L. B.; Iucu, C.; Rhee, S.; Jiang, H. Exp. Cell Res. 2006, 312, 86−
94.
(12) Lampis, G.; Deidda, D.; Maullu, C.; Madeddu, M. A.; Pompei,
R.; Monachie, F.; Satta, G. J. Antibiot. 1995, 48, 967−972.
(13) Dalgliesh, C. E.; Knox, W. E.; Neuberger, E. Nature 1951, 168,
20−22.
(14) Arend, M.; Westermann, B.; Risch, N. Angew. Chem., Int. Ed.
1998, 37, 1044−1070.
(15) (a) Moore, C. M.; Helmann, J. D. Curr. Opin. Microbiol. 2005, 8,
188−195. (b) Imlay, J. A. Annu. Rev. Microbiol. 2003, 57, 395−418.
(16) Shor, A. C.; Keschman, E. A.; Lee, F. Y.; Muro-Cacho, C.;
Letson, G. D.; Trent, J. C.; Pledger, W. J.; Jove, R. Cancer Res. 2007,
67, 2800−2808.
(17) (a) Konitsiotis, A. D.; Raynal, N.; Bihan, D.; Hohenester, E.;
Farndale, R. W.; Leitinger, B. J. Biol. Chem. 2008, 283, 6861−6868.
(b) Ali, B. R.; Xu, H.; Akawi, N. A.; John, A.; Karuvantevida, N. S.;
Langer, R.; Al-Gazali, L.; Leitinger, B. Hum. Mol. Genet. 2010, 19,
2239−2250.
(18) Johnson, F. M.; Saigal, B.; Talpaz, M.; Donato, N. J. Clin. Cancer
Res. 2005, 11, 6924−6932.
(19) Siddiqui, K.; Kim, G. W.; Lee, D. H.; Shin, H. R.; Yang, E. G.;
Lee, N. T.; Yang, B.-S. Biol. Pharm. Bull. 2009, 32, 136−141.
(20) (a) Rix, U.; Zheng, J.; Remsing Rix, L. L.; Greenwell, L.; Yang,
K.; Rohr, J. J. Am. Chem. Soc. 2004, 126, 4496−4497. (b) Cottreau, K.
M.; Spencer, C.; Wentzell, J. R.; Graham, C. L.; Borissow, C. N.;
Jakeman, D. L.; McFarland, S. A. Org. Lett. 2010, 12, 1172−1175.
(c) Dupuis, S. N.; Robertson, A. W.; Veinot, T.; Monro, S. M. A.;
Douglas, S. E.; Syvitski, R. T.; Goralski, K. B.; McFarland, S. A.;
Jakeman, D. L. Chem. Sci. 2012, 3, 1640−1644.
(21) (a) Steinmetz, H.; Gerth, K.; Jansen, R.; Schlager, N.; Dehn, R.;
̈
Reinecke, S.; Kirschning, A.; Muller, R. Angew. Chem., Int. Ed. 2011,
̈
50, 532−536. (b) Jansen, R.; Gerth, K.; Steinmetz, H.; Reinecke, S.;
Kessler, W.; Kirschning, A.; Muller, R. Chem.Eur. J. 2011, 17, 7739−
̈
7744. (c) Dehn, R.; Katsuyama, Y.; Weber, A.; Gerth, K.; Jansen, R.;
Steinmetz, H.; Hofle, G.; Muller, R.; Kirschning, A. Angew. Chem., Int.
̈
Ed. 2011, 50, 3882−3887. (d) Steinmetz, H.; Zander, W.; Shushni, M.
A. M.; Jansen, R.; Gerth, K.; Dehn, R.; Drager, G.; Kirschning, A.;
̈
Muller, R. ChemBioChem 2012, 13, 1813−1817.
̈
(22) (a) Hughes, C. C.; MacMillan, J. B.; Gaudencio, S. P.; Jensen, P.
R.; Fenical, W. Angew. Chem., Int. Ed. 2009, 48, 725−727. (b) Hughes,
C. C.; MacMillan, J. B.; Gaudencio, S. P.; Fenical, W.; LaClair, J. J.
Angew. Chem., Int. Ed. 2009, 48, 728−732. (c) Pan, E.; Oswald, N.;
Life, J. M.; Posner, B. A.; MacMillan, J. B. Chem. Sci. 2013, 4, 482−
488.
F
dx.doi.org/10.1021/ja403412y | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX