Direct binding assay for the detection of type IV allosteric inhibitors of Abl

Article abstract of DOI:10.1021/ja303858w

Abelson (Abl) tyrosine kinase is an important cellular enzyme that is rendered constitutively active in the breakpoint cluster region (BCR)-Abl fusion protein, contributing to several forms of leukemia. Although inhibiting BCR-Abl activity with imatinib shows great clinical success, many patients acquire secondary mutations that result in resistance to imatinib. Second-generation inhibitors such as dasatinib and nilotinib can overcome the majority of these mutations but fail to treat patients with an especially prevalent T315I mutation at the gatekeeper position of the kinase domain. However, a combination of nilotinib with an allosteric type IV inhibitor was recently shown to overcome this clinically relevant point mutation. In this study, we present the development of a direct binding assay that enables the straightforward detection of allosteric inhibitors which bind within the myristate pocket of Abl. The assay is amenable to high-throughput screening and exclusively detects the binding of ligands to this unique allosteric site.

Full text of DOI:10.1021/ja303858w

Communication
pubs.acs.org/JACS
Direct Binding Assay for the Detection of Type IV Allosteric Inhibitors
of Abl
Ralf Schneider,† Christian Becker,‡ Jeffrey R. Simard,†,∥ Matthaus Getlik,†,⊥ Nina Bohlke,†,#
̈
Petra Janning,§ and Daniel Rauh*,†,‡
†Chemical Genomics Centre of the Max-Planck-Society, Otto-Hahn-Strasse 15, 44227 Dortmund, Germany
‡Department of Chemical Biology, Technical University Dortmund, Otto-Hahn-Strasse 6, 44227 Dortmund, Germany
§Max-Planck-Institute of Molecular Physiology, Otto-Hahn-Strasse 11, 44227 Dortmund, Germany
S
* Supporting Information
ABSTRACT: Abelson (Abl) tyrosine kinase is an
important cellular enzyme that is rendered constitutively
active in the breakpoint cluster region (BCR)-Abl fusion
protein, contributing to several forms of leukemia.
Although inhibiting BCR-Abl activity with imatinib
shows great clinical success, many patients acquire
secondary mutations that result in resistance to imatinib.
Second-generation inhibitors such as dasatinib and
nilotinib can overcome the majority of these mutations
but fail to treat patients with an especially prevalent T315I
mutation at the gatekeeper position of the kinase domain.
However, a combination of nilotinib with an allosteric type
IV inhibitor was recently shown to overcome this clinically
relevant point mutation. In this study, we present the
development of a direct binding assay that enables the
straightforward detection of allosteric inhibitors which
bind within the myristate pocket of Abl. The assay is
amenable to high-throughput screening and exclusively
detects the binding of ligands to this unique allosteric site.
Figure 1. Schematic representation of Abl. (a) The kinase domain of
Abl is shown with inhibitors addressing the active site (blue) and the
belson (Abl) tyrosine kinase is involved in many cellular
pathways and is associated with several human dis-
A
myristate pocket (yellow). Labeling positions analyzed in this study are
indicated (sticks). (b) Helix I is bent in the inactive conformation of
Abl (left, PDB code 3k5v), while being straight in the active one (right,
PDB code 2z60). This conformational change is reported by the
attached fluorophore (spheres).
eases.1−3 It is composed of a kinase domain, two Src homology
domains (SH2, SH3), a C-terminal domain which defines its
intracellular localization, and a myristoylated N-terminal cap
region which is crucial for enzyme regulation.4 In normal cells,
Abl is predominantly found in an autoinhibited inactive state
which is stabilized by the N-terminal myristoylation site.5,6 In
the oncogenic breakpoint cluster region (BCR)-Abl fusion
protein, the N-terminus of Abl and its self-inhibiting
mechanism are lost, resulting in a constitutively active form
of the kinase which has been shown to be an important driver
in chronic myelogenous leukemia (CML) and acute lympho-
blastic leukemia (ALL).7
specific point mutations in Abl, particularly the T315I (Abl1a
numbering) gatekeeper mutation.10 None of the FDA-
approved drugs are able to overcome this mutation. Therefore,
research efforts have intensified to find novel inhibitors for the
treatment of imatinib-resistant patients carrying the T315I
mutation.11
A class of pyrimidines have shown antiproliferative activity
against BCR-Abl-expressing cells12 and received great attention
from oncologists worldwide, as they are also capable of
inhibiting BCR-Abl (T315I) when applied in combination with
ATP-competitive inhibitors such as imatinib and nilotinib.13
These pyrimidines, which include GNF-2 (1), were discovered
All inhibitors approved by the FDA so far are ATP-
competitive and bind completely (type I) or partially (type II)
within the ATP binding site of the kinase domain (Figure 1a).8
Innovative drug development efforts led to the discovery of
imatinib as a successful type II BCR-Abl inhibitor for the
treatment of CML.9 However, after a positive initial response to
treatment with imatinib, a significant number of patients
relapse. This event is mainly attributed to the acquisition of
Received: April 22, 2012
Published: May 21, 2012
© 2012 American Chemical Society
9138
dx.doi.org/10.1021/ja303858w | J. Am. Chem. Soc. 2012, 134, 9138−9141

Journal of the American Chemical Society
Communication
to act via a unique, truly allosteric (type IV) mechanism and
bind to a remote site outside of the ATP binding cleft, thus
providing an alternative mechanism of inhibition (Figure 1).
The binding of the myristoylated N-terminus of Abl within the
myristate binding pocket (MP) stabilizes the autoinhibited
inactive state of Abl.5 Similarly, the binding of ligands such as
GNF-2 can mimic this interaction by promoting a ∼90° bend
in helix I of Abl (Figure 1b), thereby allowing the adjacent SH2
domain of Abl to interact with the kinase domain in an
autoinhibitory manner.14,15 In contrast, helix I extends away
from the kinase domain in active Abl and disrupts the
interactions between the SH2 domain and the kinase domain.
Here we report the development of a direct binding assay
which is based on ligand-induced conformational changes of
helix I. We show that this system is a strong and reliable tool for
the identification of ligands which bind to the MP and can be
readily applied to high-throughput screening formats. This
assay technology (FLiK, “fluorescent labels in kinases”) was
originally developed for high-throughput screening of type II
and III kinase inhibitors by labeling either the highly flexible
activation loop16,17 or the glycine-rich loop18 of the target
kinase with acrylodan, a thiol-reactive fluorophore.19 Changes
in protein conformation lead to differences in the micro-
environment of acrylodan and alter its characteristic dual
emission spectrum.
(C305V/C330S/V338C), only the desired C338 was found
to be modified by acrylodan (Figure S5).
Prior to using this mutated construct for the characterization
of type IV Abl inhibitors, we first assessed the impact of these
substitutions on enzyme activity of wild-type and mutated Abl
(labeled and unlabeled) using standard activity-based assays.
We found that all constructs displayed similar ATP Km values
(as compared to the corresponding wild-type construct) and all
proteins were inhibited by dasatinib (3) and nilotinib (4) with
similar IC50 values, suggesting that neither the introduced
mutations nor the subsequent acrylodan labeling had a major
effect on enzyme activity or its affinity for known inhibitor
molecules (Figure S6).
Having biochemically characterized the acrylodan-labeled
Abl, we then analyzed its behavior in the FLiK assay. In the apo
form (no ligand bound), the labeled protein exhibited an
emission spectrum with a major maximum at 474 nm and a
secondary maximum (shoulder) at 510 nm (Figure 2a).
Specific labeling of the protein with acrylodan is accom-
plished by using site-directed mutagenesis to introduce a
cysteine at a site on the protein that will be perturbed by ligand-
induced conformation changes. Since two recent studies
provided evidence that bending of helix I does not require
the presence of SH2 and SH3 domains,13,20 we generated
recombinant Abl constructs consisting of only the catalytic
domain (as described elsewhere15). A comparison of published
crystal structures with either a straight or bent conformation of
helix I suggested three amino acids (V338, S503, or E505) as
potential labeling sites for the fluorophore, since a significant
change in the microenvironment of acrylodan would be
expected at these positions during the transition from the
active to the inactive state (Figure 1). Initial experiments
suggested that covalent labeling of V338C with acrylodan was
the most sensitive to conformation changes and therefore the
most suitable for the FLiK assay. This residue is close but not
part of the MP in either Abl conformation, suggesting that
labeling of V338C with the fluorophore should not disturb
ligand binding to the MP.
Figure 2. FLiK assay. (a) Stepwise addition of GNF-2 to acrylodan-
labeled Abl (C305V/C330S/V338C) results in a significant and dose-
dependent change in the fluorescence. (b) The Kd can be determined
from plotting the logarithmic concentration of GNF-2 against the ratio
of the two acrylodan emission peaks. No significant change in the ratio
is observed for Me-GNF-2 and other control inhibitors. Due to
intrinsic compound fluorescence, titration of 3−5 did not exceed 10
μM. (c) The binding of GNF-2 is fast and reversible. (d) Addition of
vehicle alone had no effect.
The crystal structures also revealed that three native cysteines
(C305, C330, and C464) are exposed on the surface of the
protein (Figure S1). As these cysteines may be susceptible to
nonspecific labeling by acrylodan, they were to be substituted
by site-directed mutagenesis. While C330 was conservatively
mutated to serine, C305 was found to be conserved as a valine
among several Src family members, including cSrc, HCK, Lck,
Lyn, BLK, YES, FIN, FGR, and Abl (Figure S2), and was
therefore substituted by a valine. C464 is located on a loop C-
terminal to helix H inside the MP (Figure S1)15 and is
conserved in sequence (Figure S2) and structure (Figure S3)
among several Src family members, suggesting an important but
unknown function of this residue within the MP. To avoid any
unwanted alterations of the native binding site, we decided to
leave this cysteine in the recombinant construct. Mass
spectrometric and FLiK-based analyses suggested that this
cysteine is not prone to labeling (Figure S4). In MS/MS
measurements of tryptic digests of acrylodan-labeled Abl
Stepwise increase of the GNF-2 concentration resulted in a
significant change in the fluorescence spectrum of acrylodan
such that the first emission peak (474 nm) decreased
dramatically over the course of the titration, while the second
peak was less affected, shifted from 510 to 524 nm, and became
the dominant peak in the GNF-2-saturated form of Abl.
With this assay, it is possible to directly determine the
dissociation constant of inhibitors as described previously.16,18
This is done by monitoring the ratio of both acrylodan
emission peaks (R = I524 nm/I474 nm) in the presence of
increasing GNF-2 concentrations. Plotting these ratiometric
values against the inhibitor concentration on a logarithmic scale
produced a sigmoidal binding curve for GNF-2, with an
inflection point corresponding to Kd = 2.1 0.4 μM (Figure
2b). The affinity of GNF-2 for the MP of Abl was consistent (n
= 7) across two independent preparations (i.e., expression,
purification, and labeling) of the labeled Abl (C305V/C330S/
9139
dx.doi.org/10.1021/ja303858w | J. Am. Chem. Soc. 2012, 134, 9138−9141

Journal of the American Chemical Society
Communication
V338C) mutant and is in a similar range as compared to an
NMR-based study in which a ligand-induced change in the
chemical environment of V506 was measured by 15N-HSQC.20
Addition of the N-methylated version of GNF-2 (2, Me-GNF-
2) as a known negative control12 gave no response in our FLiK
assay, confirming that the change in fluorescence of acrylodan is
due to binding of GNF-2 to the MP and not due to an
unspecific interaction between compound and fluorophore.
We also tested highly potent inhibitors that bind in the ATP
binding site (DFG-in and DFG-out conformations) of Abl to
assess whether conformational changes in Abl distinct from the
MP would be detected by our new FLiK assay. The type I
inhibitor dasatinib (3) stabilizes the active “DFG-in” form of
the kinase, while the type II inhibitor nilotinib (4) stabilizes the
inactive “DFG-out” form.21,22 Both Abl-inhibitor complexes
differ by a very significant conformational change in the
activation loop of the kinase. Addition of both inhibitors to
acrylodan-labeled Abl did not induce a response in the FLiK
assay (Figure 2b), highlighting that this assay is insensitive to
other common conformational changes within the kinase
domain aside from those undergone by helix I, which occur
specifically upon binding of type IV allosteric inhibitors.
The small molecule 5 was reported in a NMR-based study to
bind to the MP with moderate affinity (Kd = 6 μM) but was
unable to induce bending of helix I due to a steric clash with
I502.20 Although known to bind to this site, titration of 5 up to
10 μM in the FLiK assay did not result in a significant change in
the ratio of the two emission peaks of acrylodan (Figure 2b),
thereby confirming that detected compounds must in fact
induce the required bend in helix I and that this event is
required to alter the acrylodan emission spectrum.
In addition to using the assay to obtain valuable data about
affinities of allosteric Abl inhibitors, it is also possible to
measure the kinetic parameters for the association and
dissocation of MP ligands. Due to the greater change in
acrylodan emission at 474 nm with ligand binding (Figure 2a),
this wavelength was used to monitor the binding kinetics of
GNF-2 in real-time. As shown in Figure 2c, the assay can
reliably detect fast and reversible binding of GNF-2. Addition of
vehicle alone (Figure 2d) or the negative control Me-GNF-2
(Figure S7) did not induce the same type of response.
Finally, to facilitate the use of our assay in formats
compatible with high-throughput screening, we placed
acrylodan-labeled Abl in 384-well microtiter plates, measured
the emission spectra in the presence and absence of GNF-2
(Figure 3a), and found a striking similarity to the spectral
behavior observed in the cuvette format (Figure 2a). We then
assessed the quality of the assay by measuring the ratiometric
fluorescence in response to negative (DMSO) and positive (20
μM GNF-2) controls (Figure 3b) and determined Z′ = 0.84
0.05 (n = 5), suggesting a high robustness of the assay in HTS
formats. Traditionally, Z′ > 0.5 is considered suitable for a
quality high-throughput screening assay.23
In summary, we have established a novel direct binding assay
for the detection of type IV allosteric inhibitors addressing the
myristate pocket of Abl. We show that this assay is robust and
compatible with high-throughput screening in microtiter plates
and will serve as a powerful tool for exclusively detecting
conformational changes which occur in the vicinity of the MP.
These changes are associated directly with conformational
changes in helix I of Abl, which occur when type IV allosteric
inhibitors bind within the MP of Abl.
ASSOCIATED CONTENT
■
S
* Supporting Information
Details on compound synthesis, protein production (construct
design, expression and purification), the sequence and
structural alignment, mass spectroscopic analyses, and
representative plots of the activity-based assay. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
daniel.rauh@tu-dortmund.de
Present Addresses
∥Amgen Inc., Cambridge, MA 02142, United States
⊥Ontario Institute for Cancer Research, Toronto, ON M5G
0A3, Canada
#Department of Chemistry, Technical University Berlin, 10587
Berlin, Germany
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Matthias Rabiller, Guido Zaman, and Rogier Buijsman are
thanked for helpful discussions. The German Federal Ministry
for Education and Research supported this work through the
German National Genome Research Network-Plus
(NGFNPlus) (Grant No. BMBF 01GS08104). MSD, Bayer
HealthCare, Merck-Serono, and Bayer CropScience are
thanked for financial support.
REFERENCES
■
(1) Van Etten, R. A. Trends Cell. Biol. 1999, 9, 179−186.
(2) Ren, R. Nat. Rev. Cancer 2005, 5, 172−183.
(3) Lin, J.; Arlinghaus, R. Oncogene 2008, 27, 4385−4391.
(4) Hantschel, O.; Superti-Furga, G. Nat. Rev. Mol. Cell. Biol. 2004, 5,
33−44.
(5) Hantschel, O.; Nagar, B.; Guettler, S.; Kretzschmar, J.; Dorey, K.;
Kuriyan, J.; Superti-Furga, G. Cell 2003, 112, 845−857.
(6) Pluk, H.; Dorey, K.; Superti-Furga, G. Cell 2002, 108, 247−259.
(7) Pendergast, A. M.; Witte, O. N. Baillieres Clin. Haematol. 1987, 1,
1001−1020.
(8) An, X.; Tiwari, A. K.; Sun, Y.; Ding, P. R.; Ashby, C. R., Jr.; Chen,
Z. S. Leuk. Res. 2010, 34, 1255−1268.
Figure 3. FLiK assay in the microtiter plate format. (a) Spectra of
ligand-free and GNF-2-bound acrylodan-labeled Abl resemble the
spectra measured in cuvettes (see Figure 2a). (b) Ratio of the two
acrylodan emission peaks in the apo form (open squares) and with 20
μM GNF-2 (filled squares), used to calculate Z′ for the assay.
(9) Capdeville, R.; Buchdunger, E.; Zimmermann, J.; Matter, A. Nat.
Rev. Drug Discov. 2002, 1, 493−502.
(10) Jabbour, E.; Cortes, J.; Kantarjian, H. Cancer 2011, 117, 897−
906.
9140
dx.doi.org/10.1021/ja303858w | J. Am. Chem. Soc. 2012, 134, 9138−9141

Journal of the American Chemical Society
Communication
(11) O’Hare, T.; Deininger, M. W.; Eide, C. A.; Clackson, T.; Druker,
B. J. Clin. Cancer Res. 2011, 17, 212−221.
́
(12) Adrian, F. J.; Ding, Q.; Sim, T.; Velentza, A.; Sloan, C.; Liu, Y.;
Zhang, G.; Hur, W.; Ding, S.; Manley, P.; Mestan, J.; Fabbro, D.; Gray,
N. S. Nat. Chem. Biol. 2006, 2, 95−102.
(13) Zhang, J.; Adrian, F. J.; Jahnke, W.; Cowan-Jacob, S. W.; Li, A.
́
G.; Iacob, R. E.; Sim, T.; Powers, J.; Dierks, C.; Sun, F.; Guo, G. R.;
Ding, Q.; Okram, B.; Choi, Y.; Wojciechowski, A.; Deng, X.; Liu, G.;
Fendrich, G.; Strauss, A.; Vajpai, N.; Grzesiek, S.; Tuntland, T.; Liu, Y.;
Bursulaya, B.; Azam, M.; Manley, P. W.; Engen, J. R.; Daley, G. Q.;
Warmuth, M.; Gray, N. S. Nature 2010, 463, 501−506.
(14) Nagar, B.; Hantschel, O.; Seeliger, M.; Davies, J. M.; Weis, W. I.;
Superti-Furga, G.; Kuriyan, J. Mol. Cell 2006, 21, 787−798.
(15) Nagar, B.; Hantschel, O.; Young, M. A.; Scheffzek, K.; Veach,
D.; Bornmann, W.; Clarkson, B.; Superti-Furga, G.; Kuriyan, J. Cell
2003, 112, 859−871.
(16) Simard, J. R.; Getlik, M.; Grutter, C.; Pawar, V.; Wulfert, S.;
̈
Rabiller, M.; Rauh, D. J. Am. Chem. Soc. 2009, 131, 13286−13296.
(17) Simard, J. R.; Kluter, S.; Grutter, C.; Getlik, M.; Rabiller, M.;
̈
̈
Rode, H. B.; Rauh, D. Nat. Chem. Biol. 2009, 5, 394−396.
(18) Simard, J. R.; Getlik, M.; Grutter, C.; Schneider, R.; Wulfert, S.;
̈
Rauh, D. J. Am. Chem. Soc. 2010, 132, 4152−4160.
(19) Prendergast, F. G.; Meyer, M.; Carlson, G. L.; Iida, S.; Potter, J.
D. J. Biol. Chem. 1983, 258, 7541−7544.
́
(20) Jahnke, W.; Grotzfeld, R. M.; Pelle, X.; Strauss, A.; Fendrich, G.;
Cowan-Jacob, S. W.; Cotesta, S.; Fabbro, D.; Furet, P.; Mestan, J.;
Marzinzik, A. L. J. Am. Chem. Soc. 2010, 132, 7043−7048.
(21) Tokarski, J. S.; Newitt, J. A.; Chang, C. Y.; Cheng, J. D.;
Wittekind, M.; Kiefer, S. E.; Kish, K.; Lee, F. Y.; Borzillerri, R.;
Lombardo, L. J.; Xie, D.; Zhang, Y.; Klei, H. E. Cancer Res. 2006, 66,
5790−5797.
(22) Weisberg, E.; Manley, P. W.; Breitenstein, W.; Bruggen, J.;
̈
Cowan-Jacob, S. W.; Ray, A.; Huntly, B.; Fabbro, D.; Fendrich, G.;
Hall-Meyers, E.; Kung, A. L.; Mestan, J.; Daley, G. Q.; Callahan, L.;
Catley, L.; Cavazza, C.; Azam, M.; Neuberg, D.; Wright, R. D.;
Gilliland, D. G.; Griffin, J. D. Cancer Cell 2005, 7, 129−141.
(23) Inglese, J.; Johnson, R. L.; Simeonov, A.; Xia, M.; Zheng, W.;
Austin, C. P.; Auld, D. S. Nat. Chem. Biol. 2007, 3, 466−479.
9141
dx.doi.org/10.1021/ja303858w | J. Am. Chem. Soc. 2012, 134, 9138−9141