Discovery of 7-azaindole based anaplastic lymphoma kinase (ALK) inhibitors: Wild type and mutant (L1196M) active compounds with unique binding mode

Article abstract of DOI:10.1016/j.bmcl.2013.06.071

We have identified a novel 7-azaindole series of anaplastic lymphoma kinase (ALK) inhibitors. Compounds 7b, 7m and 7n demonstrate excellent potencies in biochemical and cellular assays. X-ray crystal structure of one of the compounds (7k) revealed a unique binding mode with the benzyl group occupying the back pocket, explaining its potency towards ALK and selectivity over tested kinases particularly Aurora-A. This binding mode is in contrast to that of known ALK inhibitors such as Crizotinib and NVP-TAE684 which occupy the ribose binding pocket, close to DFG motif.

Full text of DOI:10.1016/j.bmcl.2013.06.071

Accepted Manuscript
Discovery of 7-azaindole based anaplastic lymphoma kinase (ALK) inhibitors:
Wild type and mutant (L1196M) active compounds with unique binding mode
Venkateshwar Rao Gummadi, Sujatha Rajagopalan, Looi Chung Yeng,
Mohammadjavad Paydar, Girish Aggunda Renukappa, Bharathi Raja Ainan,
Narasimha Rao Krishnamurthy, Sunil Kumar Panigrahi, Kumari Mahasweta,
Sangeetha Raghuramachandran, Manoj Rajappa, Anuradha Ramanathan,
Anirudha Lakshminarasimhan, Murali Ramachandra, Wong Pooi Fong,
Mohammad Rais Mustafa, Srinivas Nanduri, Subramanya Hosahalli
PII:
S0960-894X(13)00804-4
DOI:
http://dx.doi.org/10.1016/j.bmcl.2013.06.071
BMCL 20635
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
21 March 2013
3 June 2013
25 June 2013
Please cite this article as: Gummadi, V.R., Rajagopalan, S., Yeng, L.C., Paydar, M., Renukappa, G.A., Ainan, B.R.,
Krishnamurthy, N.R., Panigrahi, S.K., Mahasweta, K., Raghuramachandran, S., Rajappa, M., Ramanathan, A.,
Lakshminarasimhan, A., Ramachandra, M., Fong, W.P., Mustafa, M.R., Nanduri, S., Hosahalli, S., Discovery of 7-
azaindole based anaplastic lymphoma kinase (ALK) inhibitors: Wild type and mutant (L1196M) active compounds
with unique binding mode, Bioorganic & Medicinal Chemistry Letters (2013), doi: http://dx.doi.org/10.1016/j.bmcl.
2013.06.071
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 (for review)
Graphical Abstract
Discovery of 7-azaindole based anaplastic lymphoma kinase (ALK) inhibitors: Wild type and mutant (L1196M) active
compounds with unique binding mode
Venkateshwar Rao Gummadia, §, Sujatha Rajagopalana, §, Looi Chung Yengb, Mohammadjavad Paydarb, Girish Aggunda
Renukappaa, Bharathi Raja Ainana, Narasimha Rao Krishnamurthya, Sunil Kumar Panigrahia , Kumari Mahaswetaa,
Sangeetha Raghuramachandrana, Manoj Rajappaa, Anuradha Ramanathana, Anirudha Lakshminarasimhana, Murali
Ramachandraa, Wong Pooi Fongb, Mohammad Rais Mustafab, Srinivas Nanduria, Subramanya Hosahallia, *
A novel series of 7-Azaindole derivatives have been identified as potent ALK inhibitors. Their synthesis and
preliminary SAR are described.

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com
Discovery of 7-azaindole based anaplastic lymphoma kinase (ALK) inhibitors:
Wild type and mutant (L1196M) active compounds with unique binding mode
Venkateshwar Rao Gummadia, §, Sujatha Rajagopalana, §, Looi Chung Yengb, Mohammadjavad Paydarb,
Girish Aggunda Renukappaa, Bharathi Raja Ainana, Narasimha Rao Krishnamurthya, Sunil Kumar
Panigrahia , Kumari Mahaswetaa, Sangeetha Raghuramachandrana, Manoj Rajappaa, Anuradha
Ramanathana, Anirudha Lakshminarasimhana, Murali Ramachandraa, Wong Pooi Fongb, Mohammad Rais
Mustafab, Srinivas Nanduria, Subramanya Hosahallia, *
aAurigene Discovery Technologies Ltd, #39/40, KIADB Industrial Area, Hosur road, Electronic city Phase-II, Bangalore, India, 560100
bDepartment of Pharmacology, Faculty of Medicine, University of Malaya, 50603, Kuala Lumpur, Malaysia
 Corresponding author. Tel.: +91 80 71025316; fax: +91 80 28526285; e-mail: hosahalli_s@aurigene.com
§
These authors have contributed equally to this work
ARTICLE INFO
ABSTRACT
We have identified a novel 7-azaindole series of Anaplastic Lymphoma Kinase (ALK)
inhibitors. Compounds 7b, 7m-n demonstrate excellent potencies in biochemical and cellular
assays. X-ray crystal structure of one of the compounds (7k) revealed a unique binding mode
with the benzyl group occupying the back pocket, explaining its potency towards ALK and
selectivity over tested kinases particularly Aurora-A. This binding mode is in contrast to that of
known ALK inhibitors such as crizotinib and NVP-TAE684 which occupy the ribose binding
pocket, close to DFG motif.
Article history:
Received
Revised
Accepted
Available online
Keywords:
Anaplastic Lymphoma Kinase (ALK)
7-azaindole
2013 Elsevier Ltd. All rights reserved.
Anaplastic Large Cell Lymphoma (ALCL)
Cancer
Karpass 299
Anaplastic Lymphoma Kinase (ALK) is a transmembrane
receptor tyrosine kinase that belongs to the insulin receptor
superfamily. Physiologically ALK is involved in the early
development and function of nervous system. However its
expression is restricted to the central nervous system in normal
adults1. Several fusion/mutant proteins of ALK have been
implicated in oncogenesis. Nucleophosmin (NPM)-ALK was the
first fusion protein to be identified, with the N-terminus of NPM
fused to the intracellular domain of ALK. NPM-ALK is present
in 30-50% of advanced stage Anaplastic Large Cell Lymphoma
(ALCL) patients. Echinoderm microtubule associated protein like
4 (EML4)-ALK, which is identified in 3-7% of lung tumors, has
EML4 fused to the intracellular domain of ALK2. ALK fusion
protein has also been implicated in 30-60% of inflammatory
myofibroblastic tumors (IMT)3. These fusion proteins mediate
Raf/Mek/Erk and the PLCγ pathways7, resulting in increased cell
proliferation and resistance to apoptosis. Knock-out studies and
the use of selective ALK inhibitors in pre-clinical models have
validated this kinase as a potent oncogenic driver in cancers
dependant on ALK. Recently Crizotinib a c-Met/ALK dual
inhibitor has been given accelerated approval for the treatment of
ALK-dependant cancers, further validating the target and the
potential therapeutic application of inhibiting this oncogenic
kinase. Hence there is compelling evidence for targeting ALK-
dependant hematological and solid tumors. In patients, initial
response to Crizotinib has been exceptional. However resistance
has been known to develop within a year of starting therapy.
Several examples of resistance to tyrosine kinase inhibitors
(TKIs) have been reported earlier 8,9. One of the most common
mutations has been with the gatekeeper residue observed in TKs
such as EGFR (T790M mutation) and BCR-ABL (T315I
mutation). Among the first identified resistant mutations in
patients treated with Crizotinib has been the critical gatekeeper
mutation, leucine to methionine (L1196M) 10. Hence our focus
has been the development of an ALK inhibitor that is potent
against both WT and the clinically relevant L1196M mutant form
of this kinase11.
ligand-independant
oligomerization
leading
to
autophosphorylation and hence constitutive activation of ALK. In
addition, overexpression or activating mutations of the protein
have been identified in 5-15% of neuroblastoma4,5. Recent
reports have also identified ALK amplification in 75% of
Inflammatory Breast Cancer (IBC) tumor samples that respond to
ALK inhibition in pre-clinical models6. Oncogenic ALK
mediates its effects through activation of PI3K/Akt, Jak/Stat,
Figure 1.

Screening of an in-house kinase specific library consisting of
3000 compounds resulted in the identification of 7-azaindole
compound 1 (Figure 1) as a potent inhibitor of ALK. Compound
1 showed an enzymatic potency of 90 and 141nM against wild
type and L1196M mutant ALK12. It also showed a cellular
potency of 3000nM in the ALK-driven Karpas299 proliferation
assay13. In addition to encouraging enzymatic and cellular
potency, the low molecular weight (354 Da) of compound 1
offered a potential for optimization. Profiling against an in-house
panel of kinases showed that compound 1 inhibited Aurora A
potently (84% inhibition at 100nM concentration) along with
ALK. We initiated the optimization of compound 1 with the
objectives of improving biochemical potency against ALK (WT
and L1196M), improving cellular activity and selectivity against
other kinases particularly Aurora A. Compound 1 was docked
(Figure 2) in the published ALK crystal structure (PDB ID:
2XP2) to design new compounds. The model reveals two
anchoring hydrogen bond interactions in the hinge region-pyrrole
NH with main chain carbonyl of Glu-1197 (G+1) residue and
pyridine nitrogen of azaindole scaffold with NH of Met-1199
residue (G+3). Substitution at C5- N-methyl pyrazole and C3-
pyrazole of compound 1 occupies solvent exposed region and
ribose pocket similar to that of crizotinib, respectively.
for this substitution. The reduced cellular potency observed with
compound 7g is presumably because of lack of permeability due
to carboxylic acid.
Table 2.
Based on the docking model (Figure 2), the benzyl ring of
compound 1 and the tri-substituted phenyl ring of crizotinib,
occupy a similar space in the ribose pocket. Although these two
moieties are separated by ~ 2 Å, sufficient space appears to be
available for substitution on benzyl ring. Hence we started
exploring multiple substitutions on benzyl ring of compound 7b.
We synthesized compounds with disubstitutions on the benzyl
group as shown in Table 2. Compound 7m with 3, 5-difluoro
substitution showed biochemical and cellular potencies
comparable to 7b.
Although these compounds showed anti-proliferative activity in
the functional assay (XTT), we sought to understand their
potency against the target in a cellular system. Hence we
measured the inhibition of the target pALK. Since Stat3 is one of
the primary mediators of ALK activity, we also measured pStat3
levels in cells following compound treatment. Dose dependent
inhibition of pALK was determined by in-cell western assay16,
while pStat3 inhibition (for selected compounds) was determined
by Western Blot as shown in Figure 3. The corresponding IC50
for pALK and pStat3 inhibition is shown in the accompanying
Table 3. Inhibition of pALK and pStat3 correlate well with the
anti-proliferative activity in the ALK-dependant Karpas 299
cells. Compound 7b with good biochemical and cellular potency
showed the best inhibition of pALK and pStat3. In contrast
compound 7c which showed weaker inhibition of pALK and
pStat3 was also weaker in the biochemical and anti-proliferation
assays.
Figure 2.
Since the docking mode14 of compound 1 indicated that the 5th
position substitution on azaindole occupies the solvent exposed
region, efforts were focused on expanding upon the SAR at the
C3 position. A general synthetic scheme of these compounds is
shown in Schemes 1 and 2. Compound 2 (5-bromo-3-iodo-1-
tosyl-1H-pyrrolo [2, 3-b] pyridine) was coupled with various
benzyl substituted pyrazole boronate esters to generate compound
415. Further coupling of compound 4 with the N-methyl pyrazole
boronate ester resulted in compound 6, which on basic hydrolysis
gave the desired product 7.
Figure 3.
Table 3.
We co-crystallized one of the compounds (7k) with ALK kinase
domain17. The crystallographically determined binding mode of
compound 7k is shown in Figure 4.
Scheme 1.
Figure 4.
Alternatively these compounds can also be prepared as shown in
Scheme 2. Compound 8 (5-bromo pyrrolo [2, 3-b] pyridine) was
coupled with N-methyl pyrazole boronate ester and the resulting
compound 9 was iodinated and protected with tosyl to afford key
intermediate 11. Further coupling of 11 with various benzyl
substituted pyrazole boronate esters afforded compound 6, which
on basic hydrolysis gave the desired product 7.
Compound 7k occupies the ATP binding pocket of ALK.
Hydrogen bonding interactions are seen with pyrrole NH to main
chain carbonyl of Glu-1197 (G+1) and pyridine nitrogen with
main chain amine of Met-1199 (G+3) residue. N-methyl pyrazole
moiety of the compound is directed towards the solvent exposed
region. This is in agreement with the earlier modeling study
(Figure 2) and is similar to binding modes of reported ALK
inhibitors e.g. NVP-TAE684 and Crizotinib. Comparison of the
X-ray structures of crizotinib and compound 7k reveal no
conformational changes in secondary structure of the protein.
Significant movement is observed in some of the side chains of
residues including Glu-1167, Lys-1150, and Phe-1165 to
accommodate the benzyl ring. Among these, side chain of Glu–
1167 shows a maximum of 3.4Å movement. Surprisingly the 2,
5-difluoro substituted phenyl ring occupies a unique binding
pocket not predicted by modeling studies. There is a significant
difference in the binding mode of compounds in the back pocket
region. NVP-TAE684 and Crizotinib occupy ribose binding
pocket (close to DFG loop), while compounds reported in 4DCE,
4FNY, 4FNZ structures extend deep into the back pocket
adjacent to gate keeper residue18. In contrast, the 2, 5-difluoro
substituted phenyl ring of compound 7k occupies a unique
binding pocket adjacent to the catalytic Lys. This unique binding
mode of compound 7k is not observed with any of the
compounds reported in published ALK co-crystal structures.
Scheme 2.
Docking of compound 1 (Figure2) suggested that the benzyl
group occupies a hydrophobic pocket. Hence we synthesized a
set of compounds with fluoro substitution on benzyl group
resulting in compounds 7a-c. Table 1 shows biochemical potency
of compounds 7a-c against ALK and Aurora A kinases as well as
potency in Karpass-299 cells. Among these three compounds, 7b
(3-F) appeared to be most potent in enzymatic and cellular
assays. Compound 7b showed significantly enhanced
biochemical and cellular potency compared to compound 1
although there was no change in selectivity against Aurora A.
Table 1.
Since compound 7b with 3-F substitution showed an
improvement in potency, we explored various other substitutions
at this position of the benzyl group (Table 2). Although several of
these compounds were reasonably tolerated it did not show
improvement in either biochemical or cellular or selectivity
against Aurora A compared to compound 7b. Compound 7f
showed a significantly reduced potency indicating that bulkier
groups are not tolerated at this position although the docking
mode of compound suggested there is sufficient space available

Figure 5.
Table 4.
In summary screening of a library of compounds resulted in
identification of a novel series of azaindole compounds as potent
ALK inhibitors. Co-crystal structure of one of the compounds
revealed a unique binding mode for azaindole compounds
compared to Crizotinib and NVP-TAE684. A combination of
SAR and structure-guided approach was employed to optimize
initial hit compounds. These compounds showed excellent
potency against ALK WT as well as the clinically relevant
L1196M mutant enzyme. Some of the compounds showed good
cell based potency in both anti-proliferative and mechanistic
assays. Dimethyl substitution on right hand side pyrazole ring led
to enhanced selectivity against Aurora A kinase as indicated by
the crystal structure and docking studies. Further efforts to
optimize cellular and ADME properties of these molecules are in
progress.
The benzyl group occupies the back pocket surrounded by
hydrophobic residues Ileu-1194 and Leu-1196. In addition polar
residues like Lys-1150, Glu-1167 and Arg-1275 are in close
proximity to this pocket. In this binding mode the 5-F is
positioned in close proximity to side chain of Arg-1275 residue
while 2-F is positioned near the gate keeper Leu-1196 residue.
This unique conformation and the binding mode is afforded by
the relative rigidity of the pyrazole moiety at 3rd position and the
potential interactions between the substituted phenyl ring and the
back pocket residues. In addition, the presence of catalytic Lys-
1150 in the close proximity of the pyrazole moiety can also have
a stabilizing hydrogen bond interaction. The 3-F substitution on
the benzyl group occupies a small hydrophobic pocket created by
Ile-1194 (3.2 Å) and Ile-1171 (3.6 Å) residues at the back pocket
and appear to be the optimal substitution as shown with
compound 7b. However bulkier substitution like 3-CF3 (7f) at
this position is not tolerated because of steric hindrance. The
enhanced biochemical potency of compound 7k with respect to
compound 1 appears to be due to the space filling effect of 2,5-
difluoro substituted phenyl group.
Acknowledgments
This research was supported by HIR grant (H-20001-
E00002) from the University of Malaya. The authors are thankful
to analytical, IPM department and Sivapriya Marappan of
Aurigene Discovery Technologies Ltd for providing support
during the research work.
Most of the synthesized compounds showed very high inhibition
of Aurora A kinase activity. ALK kinase activity appears to be
the main driver in ALK dependant tumors. Addition of Aurora A
has not been shown to have any synergistic/additive effect in
these tumors. In addition inhibition of Aurora kinases have been
References and notes:
1. Morris, S. W.; Kirstein, M. N.; Valentine, M. B.; Dittmer, K. G.;
Shapiro, D. N.; Saltman, D. L.; Look AT. Science 1994, 263
(5151), 1281.
2. Young Lim Choi.; Kengo Takeuchi.; Manabu Soda.; Kentaro
Inamura.; Yuki Togashi.; Satoko Hatano.; Munehiro Enomoto.;
Toru Hamada.; Hidenori Haruta.; Hideki Watanabe.; Kentaro
Kurashina.; Hisashi Hatanaka.; Toshihide Ueno.; Shuji Takada.;
Yoshihiro Yamashita.; Yukihiko Sugiyama.; Yuichi Ishikawa.;
Hiroyuki Mano. Cancer Res. 2008, 68(13), 4971.
3. Gleason, B. C.; Hornick, J. L. J. Clin. Pathology. 2008, 61, 428.
4. Helena Car´en.; Frida Abel.; Per Kogner.; Tommy Martinsson.
Biochem. J. 2008, 416, 153.
5. Rani E. George.; Takaomi Sanda.; Megan Hanna.; Stefan
Fröhling.; William LutherII.; Jianming Zhang.; Yebin Ahn.;
Wenjun Zhou.; Wendy B. London.; Patrick McGrady.; Liquan
Xue.; Sergey Zozulya.; Vlad Gregor.; Thomas R. Webb.;
Nathanael S. Gray.; D. Gary Gilliland.; Lisa Diller.; Heidi
Greulich.; Stephan W. Morris.; Matthew Meyerson.; Look, A. T.
Nature 2008, 455(7215), 975.
implicated
in
cardiotoxicity19.
In order to understand selectivity of these compounds, 7b was
docked in the X-ray structure of a representative Aurora A kinase
domain (PDB ID: 2XNG). Docking mode of compound 7b
indicated an interaction of pyrazole nitrogen with catalytic Lys-
162, while the benzylic group is placed in the ribose pocket. A
similar proximity of the pyrazole nitrogen to catalytic Lys-1150
is also observed in the docking mode of compound 7b with ALK.
However, orientation of pyrazole moiety appears to be slightly
different in ALK as compared to Aurora-A. Consequently,
pyrazole nitrogen is placed closer to the catalytic Lys-162 in case
of Aurora A (3 Å) compared to ALK (4.4 Å). It was
hypothesized that disruption of this interaction may result in
significant reduction of Aurora A kinase activity, while its impact
on ALK activity could be much lesser. Hence we synthesized few
substituted pyrazole analogs as shown in Table 4. Interestingly
some of these modifications improved selectivity against Aurora
A kinase.
6. Rabiya S Tuma. J. Natl. Cancer Inst. 2012, 104 (2), 87.
7. Shaw, A.; Solomon, B. Clin. Cancer. Res. 2011, 17(8), 2081
8. Ruriko Tanaka.; Shinya Kimura. Exp. Rev. of Anticancer
Therapy. 2008, 8(9), 1387.
9. Cai-Hong Yun.; Kristen E. Mengwasser.; Angela V. Toms.;
Michele S. Woo.; Heidi Greulich.; Kwok-Kin Wong.; Matthew
Meyerson.; Michael J. Eck. PNAS. 2008, 105 (6), 2070.
Compared to unsubstituted pyrazole, mono and dimethyl
pyrazole appear to give better selectivity against Aurora A (7n,
7o & 7q). Substitution on the benzyl group also appears to have
some influence on selectivity. 3,5-difluoro substitution on the
benzyl group along with dimethyl substitution on the pyrazole
(7q) appear to give best selectivity against Aurora A. Docking
mode of compound 7q with Aurora-A is shown in Figure 6. As
hypothesized, dimethyl substitution on pyrazole appears to
change the orientation of pyrazole moiety, resulting in loss of
interaction with catalytic Lys-162. This is also evident from the
biochemical activities of compounds 7q and 7b. Dimethyl
substitution at pyrazole in compound 7q results in significant loss
in Aurora A potency (64 folds) and marginal loss of ALK WT (4
folds) in comparison with compound 7b.
10. Young Lim Choi.; Manabu Soda.; Yoshihiro Yamashita.;
Toshihide Ueno.; Junpei Takashima.;Takahiro Nakajima.; Yasushi
Yatabe.; Kengo Takeuchi.; Toru Hamada.; Hidenori Haruta.;
Yuichi Ishikawa.; Hideki Kimura.; Tetsuya Mitsudomi.; Yoshiro
Tanio.; Hiroyuki Mano. N. Eng. J. Med. 2010, 363(18), 1734.
11. Refer Supplementary data, section C; part-I
12. In vitro enzyme assays: The ALK WT activity was determined, in
a 384-well TR-FRET format using recombinant human ALK
enzyme (Cat# 08-518, Carna Biosciences) and Ultra Light Poly
GT (Cat# TRF 0100D, Perkin Elmer) as a substrate. The final
assay conditions were 50 mM HEPES pH 7.1, 10 mM MgCl2, 2
mM MnCl2, 0.01% BSA, 2.5 mM DTT, 0.1 mM Na3 VO4, 40
nM Ultra Light Poly GT, 2.5 ng ALK WT enzyme, 1µM ATP and
125 nM Lance Eu-W1024 labeled anti phospho tyrosine antibody
(Cat# AD0203, Perkin Elmer) in 384 well format. The assay
reaction time was 30 minutes after which the antibody detection
mix is added. The L1196M mutant ALK assay was done in a
similar format. Mutant enzyme was from Carna Biosciences (Cat#
08-529) and 0.5µM ATP was used in the assay buffer. The
inhibitory activity against Aurora A was assessed by examining
the amount of ATP formed using Kinase Glo system (Cat #
Figure 6.

V1931, Promega) in the presence of 2µM ATP and Kemptide as
substrate.
13. In vitro cell-based assays: Karpas 299 cells were treated with
various concentrations of compound for 72h to determine cell
growth. Inhibition of proliferation was determined using XTT.
14. Schrodinger suite of software (Schrodinger suite 2012 release)
was used for all modeling study. ALK kinase domain (PDB ID:
2XP2 & 2NXG) were prepared using protein preparation wizard.
This process takes care of hydrogen atom addition, correction of
missing side chains, finding accurate tautomeric state of residues
like Histidine at specified pH (pH adjusted to 7.4) and geometric
optimization of hydrogen atoms. Similarly 3-D structure of all
synthesized compounds were prepared using ligprep module
(Version 2.5) prior to docking. Docking was carried out using
Glide. Top ranked pose was shortlisted for analysis. All figures
corresponding to docking and x-ray structure were generated using
Pymol (version 1.5) visualization software from Schrodinger.
15. Semghee Hong; Jinhee Kim; Ju Hyeon Seo; Kyung Hee Jung;
Soon-Sun Hong; Sungwoo Hong. J. Med. Chem. 2012, 55(11),
5337.
16. Briefly Karpas 299 cells were treated with various concentrations
of compound for 1h following which they were fixed,
permeabilized and inhibition of pALK was determined by in-cell
western. For western blotting: Briefly, 4 × 105 Karpas cells were
treated with test compound or positive control Crizotinib at the
indicated doses for 6 hours. Next, cells were harvested and lysed
in RIPA buffer (Santa Cruz). 30 µg of total cell lysates were
loaded onto 10% polyacrylamide gel and transferred to
microporous polyvinylidene difluoride (PVDF) membrane
(Milipore). Immunoblotting was performed with anti-Stat3 and
anti-phospho Stat3 (Tyr705) (Cell Signaling Technology,
Danvers, MA) antibodies. Membranes were detected using ECL
Plus Chemiluminescence Reagent (Amersham, Chalfont, UK).
17. ALK protein was expressed and purified as reported in Biochem.
J. 2010, 430 (3), 425. Freshly purified protein was incubated over
night at 4oC with AUAK-39 (in-house compound, structure not
disclosed) at 1: 2 ratio. Protein was concentrated to 7mg/ml and
crystallization was set up in and around the reported conditions.
Crystals were obtained in most of the drops. Crystals were soaked
with compound 7k at 1mM concentration for 3 nights before data
collection. The crystals were flash frozen at 100K using 20%
glycerol as cryo-protectant. The diffraction data were collected
using in-house Rigaku RU300 X-ray generator with R-AXIS IV
++detector to
a maximum resolution of 2.5 Å. Indexing,
integration and scaling were performed using DENZO and
SCALEPACK. The structure was solved by molecular
replacement (MR) method using the reported ALK structure (PDB
Code; 2XP2). Model building and ligand fitting were done using
COOT and refinement was carried out using REFMAC 5.2.0001
software. The final model is complete except for some residues
due to poor electron density. These missing residues are Gly1123-
All1130 (P-loop region), Ser1136-Leu1145, Pro1153-Ala1164,
Ala-1214-Leu1221, Tyr1278-Cys1288 and Ala1300-Gly1304.
18. Refer Supplementary data, section C; part-II
19. Thomas Force.; Kyle L. Kolaja. Nat. Rev. Drug Disc. 2011, 10,
111.
Supplementary data
Supporting information is attached as a separate file.

Table(s)
Table 1. SAR of halo substituted benzyl group
Compound
R
ALK WT
IC50 a nM
90
ALK (L1196M)
IC50 a nM
141
Aurora A %
inhibition at 100nM
Karpass-299 cell
IC50 b nM
3000
1
H
84
64
84
74
7a
7b
7c
2-F
3-F
4-F
85
5.6
70
223
35
256.4
2088
466
1781
aIC50 values determined by TR-FRET; see ref-12
bIC50 values determined by XTT; see ref-13
Table 2. SAR of the substituted benzyl group
Compound
R
ALK WT
ALK
(L1196M)
IC50 a nM
141
35
143
150
465.4
811
ND
ND
200.5
124
379
47
Aurora A %
inhibition at
Karpass-299 cell
IC50 b nM
IC50 a nM
100nM
84
84
87
85
84
ND
ND
ND
93
1
7b
7d
7e
7f
7g
7h
7i
H
3-F
3-Cl
3-CN
3-CF3
3-COOH
3-NO2
3-NH2
90
5.6
30
32
247
98
70
28
41
29
84
6
3000
466
855
1609
1008
>10000
1852
3627
3363
1200
2261
663.8
7j
7k
7l
3-OH
2,5-difluoro
3,4- difluoro
3,5- difluoro
75
84
74
7m
aIC50 values determined by TR-FRET; see ref-12
bIC50 values determined by XTT; see ref-13, ND – Not determined

Table 3. Comparison of potency of the compounds in functional (XTT) and mechanistic (pStat3
inhibition) assays in Karpas 299 cells. IC50 for pStat3 was determined by densitometric evaluation of
western blots
Compound
ALK WT IC50 nM
pALK IC50 (nM)
pStat3 IC50 (nM)
XTT IC50 (nM)
7b
7c
7d
7k
7l
5.6
70
30
29
84
219.5
3914
3121
900
138.5
3011
985.6
1592
466
1781
855
1200
3300
985.6
2261
Table 4. SAR towards Aurora A selectivity
Compound
R
R1
ALK WT
ALK (L1196M)
AuroraA
IC50 nM
11.8
37.13
355.9
216.2
768.3
Fold
selectivity
IC50 a nM
IC50 a nM
35
7b
7n
7o
7p
7q
Shown in Table 1
3-F
3-F
5.6
6
31
63
26
2
6
11
3
H
56
274.9
428
CH3
CH3
CH3
2,5-difluoro
3,5- difluoro
310.2
30
aIC50 values determined by TR-FRET; see ref-12

Figure(s)
Figure 1. Some of the reported ALK Inhibitors and 7-Azaindole derivative
Figure 2. Docking model of compound 1 (yellow) in the X-ray structure of crizotinib bound ALK (PDB ID:
2XP2). Critical hinge hydrogen bond interactions are shown in dotted lines for compound 1 and
Crizotinib (magenta). Picture made with Pymol (version 1.5) visualization software.

Figure 3. Suppression of Stat3 phosphorylation by ALK inhibitors in Karpas 299. The cells were treated
with compounds 7b, 7c, 7d, 7k, 7l at the indicated concentrations or known inhibitor Crizotinib at 10 μM
(C) for 6 hours. Cell lysates were harvested and evaluated for levels of total and phospho-Stat3 by
Western blotting.
Figure 4. Co-crystal structure of compound 7k (yellow) in complex with ALK kinase domain (PDB ID:
4J0A). Hydrogen bond interactions are represented by dotted lines and the active site residues are
labeled. The 2Fo-Fc electron density map contoured at 1σ level for the bound 7k compound is shown as
mesh (blue).

Figure 5. Overlay of X-ray structures of ALK showing critical differences in the binding modes of
compounds. Compound 7k (yellow), Crizotinib (magenta) and NVP-TAE684 (green).
Figure 6. Docking model of 7b (magenta) and 7q (yellow) shown in Aurora A kinase domain. Critical
differences in binding model of compound 7b and 7q are highlighted.

Scheme 1. General synthesis of 3, 5-disubstituted 7-azaindoles
Scheme 2. General synthesis of 3, 5-disubstituted 7-azaindoles