Structure-based design of novel bicyclic nonpeptide inhibitors for the Src SH2 domain [1]

Full text of DOI:10.1021/jm0003337

© Copyright 2000 by the American Chemical Society
Volume 43, Number 21
October 19, 2000
Communications to the Editor
Str u ctu r e-Ba sed Design of Novel Bicyclic
Non p ep tid e In h ibitor s for th e Sr c SH2
Dom a in
William C. Shakespeare,* Regine S. Bohacek,
Mihai D. Azimioara, Karina J . Macek,
George P. Luke, David C. Dalgarno,
Marcos H. Hatada, Xiaode Lu, Shelia M. Violette,
Catherine Bartlett, and Tomi K. Sawyer
ARIAD Pharmaceuticals, Inc., 26 Landsdowne Street,
Cambridge, Massachusetts 02139-4234
Received August 2, 2000
In tr od u ction . The enzymes responsible for the selec-
tive phosphorylation of tyrosine residues, protein ty-
rosine kinases (PTKs),1 and the corresponding dephos-
phorylation, protein tyrosine phosphatases (PTPs),2
have been implicated in mediating a wide array of
intracellular events including cell proliferation, migra-
tion, differentiation, metabolism, and immune response.
These proteins selectively bind their targets and initiate
a cascade of signaling events through a series of
cytosolic modular domains that control protein-protein
interactions.3 One such domain, the Src homology 2
(SH2) domain, has been determined to play a critical
role in many signaling cascades by recognizing phos-
photyrosine (pTyr) sequences of cognate proteins.4 The
SH2 domain of the nonreceptor PTK Src, for example,
has been shown to interact with FAK, p130cas, p85, PI3-
F igu r e 1. Chemical structures of 1-3.
we were compelled to advance our efforts on the
structure-based design of inhibitors of the Src SH2
domain.
SH2 domains are relatively small protein modules of
approximately 100 amino acids that recognize pTyr-
containing sequences, thereby facilitating phosphoryl-
ation-dependent, protein-protein interactions. The sec-
ondary structure of the SH2 domain features a large
central antiparallel â-sheet and two flanking R-helices.
The pTyr-binding cleft is formed by three strands of the
â-sheet (âB, âC, and âD), the loop between the âB and
âC sheets, and Arg RA2 of the A helix.12-14 The first
structure of a ligated SH2 domain was solved using a
low-affinity phosphopeptide.15 Subsequently, a higher-
affinity ligated structure was determined for Src SH2
bound to an 11-residue phosphopeptide12,16 containing
the Src family-binding motif pTyr-Glu-Glu-Ile (1, Figure
1). These four residues make extensive contacts with
the protein. In fact, the majority of interactions with 1
involve the phosphotyrosine (in the basic pTyr pocket)
and pY+3 residue Ile. The pY+1 Glu(NH) forms a
hydrogen bond with His180(CO), and the Câ and Cø lie
on the protein surface in van der Waals contact with
K, and p68sam 5-8
Small molecules designed to inhibit
.
SH2-mediated protein-protein interactions provide po-
tential molecular probes to study cellular signaling
pathways and ultimately act as therapeutic agents
modulating such pathways demonstrated to be involved
in the pathogenesis of certain diseases. In the case of
Src, elevated levels of kinase activity have been associ-
ated with several different cancers including breast9 and
colon10 cancer. Additionally, genetic knockout studies
have implicated Src in bone resorption processes.11 Thus
* Corresponding author. Tel: 617-494-0400, ext 238. Fax: 617-494-
8144. E-mail: william.shakespeare@ariad.com.
10.1021/jm0003337 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/28/2000

3816 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21
Communications to the Editor
F igu r e 2. X-ray crystal structure of pTyr-Glu-Glu-Ile (1)
bound to the SH2 domain of Lck illustrating the solvent-
accessible surface detailed in terms of both hydrophobic and
hydrogen-bonding interactions. Yellow colors refer to hydro-
phobic surfaces, blue colors refer to hydrogen-bond-accepting
surfaces, and red colors refer to hydrogen-bond-donating
surfaces.
F igu r e 3. Model of 2 complexed with Lck SH2.
Tyr181 and Val178, respectively (Figure 2). No ad-
ditional hydrogen bonds are directly formed between the
backbone of pTyr-Glu-Glu-Ile and the protein. In fact,
between the pY+1 carbonyl and the pY+3 residue, the
backbone lies above the solvent-accessible surface. The
only interaction the pY+2 residue makes is an electro-
static one between the side-chain acid and Arg184.
Str u ctu r e-Ba sed Design . The SH2 binding site
consists of two binding pockets: the pTyr and the pY+3
pocket. These pockets are separated by a solvent-
accessible â-strand (âD), with no significant indenta-
tions. However, when this region was carefully visual-
ized using an accessible surface colored to display both
hydrophobic and hydrogen-bonding sites,17 a large
hydrophobic region with a slight “dimple” in the middle
became apparent (Figure 2). This convex hydrophobic
surface is created by the phenyl ring of Tyr181.
The stacking of aromatic rings, either parallel or
perpendicular, is a well-known phenomenon observed
in protein structures18,19 and between ligands and
proteins in complexes. Therefore, we sought to capitalize
on the presence of the Tyr181 phenyl ring by designing
molecules which would position appropriately an aro-
matic ring to form energetically favorable stacking
interactions. For the initial design strategy, we chose
not to replace the phosphate group of pTyr, although,
ultimately, a suitable bioisosteric replacement would be
desired.
Compound 2 (Figure 1) was the first compound
advanced using an interactive structure-based design
strategy.20 To determine the likely binding of this
molecule to Src SH2, we docked the molecule into a
model of the binding site of Lck SH2 based on a high-
resolution (1.0 A) crystal structure of Lck SH2 com-
plexed with pTyr-Glu-Glu-Ile.21,22 The FLO9723 molec-
ular modeling program was used to simulate the
interactions compound 2 may form with binding site
atoms. Our docking studies predicted that compound 2
would position its phenyl ring above Tyr181; however,
the model showed that the cyclohexane ring was not
ideal in that it did not penetrate deeply into the pY+3
pocket (Figure 3). Also, other opportunities for hydro-
F igu r e 4. Model of 3a (S,S) complexed with Lck SH2
indicating the hydrophobic-accessible surface.
phobic and hydrogen-bonding interactions around the
Tyr181 were not utilized. Nonetheless, this compound
represented a useful starting point. Compound 2 was
synthesized and found to have an IC50 of 300 µM as a
mixture of diastereomers.
After this initial investigation, a novel series of de
novo designed nonpeptide ligands was reported, as
exemplified by the construction of compound 3 (3a , IC50
) 2 µM in our assay; Figure 1).24,25 Docking studies of
3a in our model showed numerous favorable protein-
ligand interactions (Figure 4): (1) the phenyl ring of 3a
stacks perpendicular to the phenyl ring of Tyr181; (2)
the primary amide of 3a forms two hydrogen bonds with
Lys182 (NH and CO) which displaces two water mol-
ecules and pulls the cyclohexane ring closer toward the
protein; and (3) the cyclohexane ring of 3a fits well into
the pY+3 pocket making five hydrophobic contacts.
However, analysis of the hydrophobic-accessible surface
adjacent to the central benzamide ring of 3a indicated
that the phenyl ring did not extend beyond Tyr181,
thereby not fully exploiting the complete hydrophobic
surface of the Src SH2 domain. We, therefore, focused
our search for ways to design compounds that would
more fully complement the complete convex architecture
created by Tyr181 while maintaining the other inter-
molecular contacts inherent in ligand 3a . Our search
resulted in compounds 4 and 5 (Figure 5), both which

Communications to the Editor
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21 3817
Sch em e 1. Chemical Synthesis of 4
F igu r e 5. Chemical structures of 4 and 5.
Formation of the carbobicylic amine 15 began from
the bicyclic phenol 11 (Scheme 2).27 Alkylation of phenol
Sch em e 2. Chemical Synthesis of 5
F igu r e 6. Model of 4a (white) and 5a (green) complexed with
the Lck SH2.
position a seven-membered ring over the complete
hydrophobic surface of Tyr181 (Figure 6). In addition
to the predicted enthalpic contributions, we also believed
that the increased rigidity of 4 and 5 should be entropi-
cally more favorable. In conclusion, our modeling studies
predicted that compounds 4a and 5a (the two correct
diastereomers) should bind to the SH2 domain of Src
with higher affinity than does 3a . Moreover, because
compound 4a positions an oxygen adjacent to the
hydrophobic surface, we expected it would exhibit
slightly weaker affinity than the related carbocycle 5a .
Ch em istr y. Preparation of the requisite amine 10 for
the construction of 4 began with commercially available
2,4-dihydroxybenzamide (6) (Scheme 1). Sequential
alkylation at the 4- and 2-positions resulted in the
formation of ester 7. Hydrolysis of 7, followed by a PPA
cyclization, afforded ketone 8. Reduction of 8 with
NaBH4 gave the corresponding alcohol, which was
smoothly transformed to the azide 9 with diphenyl
phosphorazidate.26 Reduction of 9 over Pd/C afforded
the desired amine 10. Amine 10 was then coupled with
an appropriately protected pTyr derivative; the result-
ing diastereomers were separated by RP-HPLC and the
benzyl groups removed with TFA to yield 4a and 4b.
11, followed by a regioselective oxidation with copper-
(II) sulfate and potassium peroxydisulfate,28 afforded
ketone 12 in 72% yield. This method of oxidation was
remarkably clean relative to the more conventional
reagents such as DDQ and PDC, both which gave lower
yields and more complex mixtures. Reduction of 12 with
NaBH4, followed by installation of the azido functional-
ity and reduction over palladium in the presence of
Boc2O, furnished the Boc-protected amine 13. Regiose-
lective bromination,29 followed by lithiation and car-
boxylation, yielded the carboxylic acid 14. Amide for-
mation and subsequent deprotection with TFA afforded
the requisite amine 15. Coupling as before, followed by
deprotection, furnished the desired compounds 5a and
5b.

3818 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21
Communications to the Editor
Ta ble 1. Comparative Src SH2 Binding Affinities (IC50’s) for
Compounds 3-5 Using a Competitive Fluorescence Polarization
Assay
F igu r e 7. Superposition of the model of 4a (white) and the
X-ray crystal structure of the 4a Lck SH2 (S164C) complex
(green). The water-mediated hydrogen bonds (determined
crystallographically) between the carboxamide of 4a and the
protein were not predicted as no water molecules were
included in our model.
knowledge, 4 and 5 represent some of the tightest
binding inhibitors known for this target SH2, and such
results demonstrate the effectiveness of exploiting ac-
cessible surface contacts using bicyclic templates. Fur-
ther modification of this series by non-phosphate-
containing pTyr isosteres holds the promise for future
inhibitors with increased cell penetration and in vivo
activity.31,32
*pTyr-Glu-Glu-IIe had an IC50 ) 5.6 µM in this assay.
Resu lts. Src SH2 binding data for compounds 3-5
were determined using a fluoresence polarization as-
say30 and are shown in Table 1. Incorporation of the
seven-membered ring systems resulted in a 10-20-fold
increased (relative to 3a ) affinity for compounds 4a and
5a . The carbocyclic analogue 5a is a 2-fold better binder
than the oxepin 4a . This is consistent with modeling
which predicted a slightly higher contact energy (hy-
drophobic interactions) for 5a (-34 vs -32.3 for 4a ) and
our own observations that when these ring systems are
attached to pTyr mimics, the carbocyclic-containing
compounds were uniformly more active than those
which contained oxygen (data not shown). Consistent
with the previously reported binding data for diastere-
omer 3b,24,25 both 4b and 5b were less active.
To confirm our hypothesis regarding the mode of
binding, the structure of 4a complexed with Lck SH2
(S164C) was determined by X-ray crystallography at a
resolution of 1.95 Å (Figure 7). As predicted, the seven-
membered ring of 4a adopts a concave conformation
which complements exquisitely the convex nature of the
hydrophobic surface created by Tyr181. The benzamide
carbonyl forms a hydrogen bond with the backbone NH
of Lys182 (for Src, Lys206) displacing one of the two
water molecules observed in the phosphopeptide com-
plex. The second water molecule is not displaced and is
hydrogen-bonded to the benzamide NH group of 4a and
the backbone carbonyl of Ile193 (for Src, Ile217). The
cyclohexyl group, although predicted to extend more
deeply into the hydrophobic pY+3 pocket, was not fully
extended and was observed in a higher-energy confor-
mation.
Ack n ow led gm en t. The authors thank Bonnie Mar-
mor and Ruth Yuan for analytical data, Andy Tyler
(Harvard University) for high-resolution MS, and the
financial support provided by Hoechst Marion Roussel
(Aventis).
Refer en ces
(1) Levitzki, A. Protein tyrosine kinase inhibitors as novel thera-
peutic agents. Pharmacol. Ther. 1999, 82, 231-239.
(2) Tonks, N. K. Protein tyrosine phosphatases and the control of
cellular signaling responses. Adv. Pharmacol. 1996, 36, 91-119.
(3) Cohen, G. B.; Ren, R.; Baltimore, D. Modular binding domains
in signal transduction proteins. Cell 1995, 80, 237-248.
(4) Thomas, S. M.; Brugge, J . S. Cellular functions regulated by Src
family kinases. Annu. Rev. Cell Dev. Biol. 1997, 13, 513-609.
(5) Fukui, Y.; Hanafusa, H. Requirement of phosphatidylinositol-3
kinase modification for its association with p60src. Mol. Cell. Biol.
1991, 11, 871-874.
(6) Schaller, M. D.; Hilderbrand, J . D.; Shannon, J . D.; Fox, J . W.;
Vines, R. R.; Parsons, J . T. Autophosphorylation of the focal
adhesion kinase, pp125FAK, directs SH2-dependent binding of
pp60src. Mol. Cell. Biol. 1994, 14, 1680-1688.
(7) Taylor, S. J .; Shalloway, D. An RNA-binding protein associated
with Src through its SH2 and SH3 domains in mitosis. Nature
1994, 368, 867-871.
(8) Petch, L. A.; Bockholt, S. M.; Bouton, A.; Parsons, J . T.; Burridge,
K. Adhesion-induced tyrosine phosphorylation of the p130src
substrate. J . Cell. Sci. 1995, 108, 1371-1379.
(9) Ottenhof-Kalff, A. E.; Rijksen, G.; Beurden, A. C. M. v.; Hen-
nipman, M. P.; Michels, A. A.; Staal, G. E. Characterization of
protein tyrosine kinases from human breast cancer: involvement
of the c-src oncogene product. J . Cancer Res. 1992, 49, 4773.
(10) Weber, T. K.; Steele, G.; Summerhayes, I. C. Different pp60c-
src activity in well and poorly differentiated human colon
carcinomas and cell lines. J . Clin. Invest. 1992, 90, 372.
(11) Soriano, P.; Montgomery, C.; Geske, R.; Bradley, A. Targeted
disruption of the c-src proto-oncogene leads to osteoporosis in
mice. Cell 1991, 64, 693-702.
(12) Waksman, G.; Shoelson, S. E.; Pant, N.; Cowburn, D.; Kuriyan,
J . Binding of a high affinity phosphotyrosyl peptide to the Src
SH2 domain: Crystal structures of the complexed and peptide-
free forms. Cell 1993, 72, 779-790.
Using X-ray data available from both peptide and
nonpeptide inhibitors as well as the FLO97 molecular
modeling program, we have designed a series of second-
generation inhibitors of the Src SH2 domain. To our

Communications to the Editor
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 21 3819
(13) Gilmer, T.; Rodriguez, M.; J ordan, S.; Crosby, R.; Alligood, K.;
Green, M.; Kimery, M.; Wagner, C.; Kinder, D.; Charifson, P.;
Hassell, A. M.; Willard, D.;Luther, M.; Rusnak, D.; Sternbach,
D. D.; Mehrotra, M.; Peel, M.; Shampine, L.; Davis, R.; Robbins,
J .; Patel, I. R.; Kassel, D.; Burkhart, W.; Moyer, M.; Bradshaw,
T.; Berman, J . Peptide inhibitors of Src SH3-SH2-phosphop-
rotein interactions. J . Biol. Chem. 1994, 269, 31711-31719.
(14) Eck, M. J .; Shoelson, S. E.; Harrison, S. C. Recognition of a high-
affinity phosphotyrosyl peptide by the Src homology-2 domain
of p561ck. Nature 1993, 362, 87-91.
(15) Waksman, G.; Kominos, D.; Robertson, S. C.; Pant, N.; Balti-
more, D.; Birge, R. B.; Cowbum, D.; Hanafusa, H.; Mayer, B. J .;
Overduin, M.; Resh, M. D.; Rios, C. B.; Silverman, L.; Kuriyan,
J . Crystal structure of the phosphotyrosine recognition domain
(SH2) of the v-src tyrosine kinase complexed with tyrosine
phosphorylated peptides. Nature 1992, 358, 646-653.
(16) Eck, M. J .; Shoelson, S. E.; Harrison, S. C. Recognition of a high
affinity phosphotyrosyl peptide by the Src homology 2 domain
of p561ck. Nature 1993, 362, 87-91.
(17) Bohacek, R. S.; McMartin, C. Definition and display of steric,
hydrophobic and hydrogen-bonding properties of ligand binding
sites in proteins using Lee and Richards accessible surface:
Validation of a high-resolution graphical tool for drug design.
J . Med. Chem. 1992, 35, 1671-1684.
(18) Burely, S. K.; Petsko, G. A. Aromatic-aromatic interaction: a
mechanism of protein structure stabilization. Science 1985, 229,
23.
determination (X-ray and NMR). Where relevant, to compare
Lck with Src, residue numbers for Src are given in parentheses.
(23) McMartin, C.; Bohacek, R. S. QXP: Powerful, rapid computer
algorithms for structure-based design. J . Comput.-Aided Mol.
Des. 1997, 11, 333-344.
(24) Lunney, E. A.; Para, K. S.; Rubin, J . R.; Humblet, C.; Fergus, J .
H.; Marks, J . S.;Sawyer, T. K. Structure-based design of a novel
series of nonpeptide ligands that bind to the pp60src SH2 domain.
J . Am. Chem. Soc. 1997, 119, 12471-12476.
(25) Lunney, E. A.; Para, K. S.; Plummer, M. S.; Prasad, J . V. N. V.;
Saltiel, A. R.; Sawyer, T.; Shahripour, A. Compounds, composi-
tions and methods for inhibiting the binding of proteins contain-
ing an SH2 domain to cognate phosphorylated proteins. WO 97/
12903, 1997.
(26) Thompson, A. S.; Humphrey, G. R.; DeMarco, A. M.; Mathre, D.
J .; Grabowski,E. J . J . Direct conversion of activated alcohols to
azides using diphenyl phosphorazidate. A practical alternative
to Mitsunobu conditions. J . Org. Chem. 1993, 58, 5886-5888.
(27) Prelog, v. V.; Ruzicka, L.; Metzler, O. Zur kenntnis des kohlen-
stoffringes. U¨ ber cis-3,4-polymethylen-cyclohexanone. Helv. Chim.
Acta 1947, 30, 1883-1895.
(28) Hauser, F. M.; Ellenberger, S. R. Regiospecific oxidation of
methyl groups in dimethylanisoles. Synthesis 1987, 723-724.
(29) Nilsson, J . L. G.; Selander, H.; Sievertsson, H.; Skanberg, I.;
Svensson, K.-G. Directive effects in the bromination of bicyclic
phenols. Acta Chem. Scand. 1971, 25, 94-100.
(30) Lynch, B. A.; Loiacono, K. A.; Tiong, C. L.; Adams, S. E.; Macneil,
I. A. A fluoresence polarization based Src-SH2 binding assay.
Anal. Biochem. 1997, 247, 77-82.
(19) Burely, S. K.; Petsko, G. A. Weakly polar interactions in proteins.
Adv. Protein Chem. 1988, 39, 125.
(20) Buchanan, J . L.; Bohacek, R. S.; Luke, G. P.; Hatada, M.; Lu,
X.; Dalgarno, D. C.; Narula; S. S.; Yuan, R.; Holt, D. A.
Structure-based design and synthesis of a novel class of Src SH2
inhibitors. Bioorg. Med. Chem. Lett. 1999, 9, 2353-2358.
(21) Tong, L.; Warren, T. C.; King, J .; Betageri, R.; Rose, J .; J akes,
S. Crystal structures of the human p561ck SH2 domain in
complex with two short phosphotyrosyl peptides at 1.0 ANG. and
1.8 ANG. resolution. J . Mol. Biol. 1996, 256, 601-610.
(22) When this work was initiated, no high-resolution crystal struc-
tures of Src SH2 were available for docking ligands which span
the entire binding site. Therefore, we used a high-resolution (1.0
Å) crystal structure of Lck SH2 complexed with the phospho-
peptide pYEEI. To the extent that the active site of Lck SH2 is
identical to that of Src SH2, we have successfully exploited this
model in the design of highly potent Src SH2 inhibitors and in
the accurate prediction of their binding mode prior to structural
(31) Violette, S. M.; Shakespeare, W. C.; Bartlett, C.; Guan, W.;
Smith, J . A.; Rickles,R. J .; Bohacek, R. S.; Holt, D. A.; Baron,
R.; Sawyer, T. K. A Src SH2 selective binding compound in-
hibits osteoclast-mediated resorption. Chem. Biol. 2000, 7, 225-
235.
(32) Shakespeare, W.; Yang, M.; Bohacek, R.; Cerasoli, F.; Stebbins,
K.; Sundaramoorthi, R.; Vu, C.; Pradeepan, S.; Metcalf, C.;
Haraldson, C.; Merry, T.; Dalgarno, D.; Narula, S.; Hatada, M.;
Lu, X.; Schravendijk, M. R. v.; Adams, S.; Violette, S.; Smith,
J .; Guan, W.; Catherine Bartlett; Herson, J .; Iuliucci, J .; Weigele,
M.; Sawyer, T. Structure-based design of a novel, osteoclast-
selective, nonpeptide Src SH2 inhibitor with in vivo anti-
resorptive activity. Proc. Natl. Acad. Sci. USA 2000, 97,
9373-9378.
J M0003337