Structure-based design and synthesis of phosphinate isosteres of phosphotyrosine for incorporation in Grb2-SH2 domain inhibitors. Part 1

Article abstract of DOI:10.1016/S0960-894X(00)00475-3

Based on X-ray crystal structure information, mono charged phosphinate isosteres of phosphotyrosine have been designed and incorporated in a short inhibitory peptide sequence of the Grb2-SH2 domain. The resulting compounds, by exploiting additional interactions, inhibit binding to the Grb2-SH2 domain as potently as the corresponding doubly charged (phosphonomethyl)phenylalanine analogue. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0960-894X(00)00475-3

Bioorganic & Medicinal Chemistry Letters 10 (2000) 2337±2341
Structure-Based Design and Synthesis of Phosphinate Isosteres
of Phosphotyrosine for Incorporation in Grb2-SH2
Domain Inhibitors. Part 1
Pascal Furet,a,* Giorgio Caravatti,a Alastair A. Denholm,b Alex Faessler,b
Heinz Fretz,a Carlos Garcõa-Echeverrõa,a Brigitte Gay,a Ed Irving,b
Neil J. Press,b Joseph Rahuel,a Joseph Schoepfera and Clive V. Walkerb
aNovartis Pharmaceuticals, Inc., Therapeutic Area Oncology, CH-4002 Basel, Switzerland
bNovartis Horsham Research Centre, Wimblehurst Road, Horsham, West Sussex, RH12 5AB, UK
Received 15 March 2000; revised 17 July 2000; accepted 8 August 2000
AbstractÐBased on X-ray crystal structure information, mono charged phosphinate isosteres of phosphotyrosine have been
designed and incorporated in a short inhibitory peptide sequence of the Grb2-SH2 domain. The resulting compounds, by exploiting
additional interactions, inhibit binding to the Grb2-SH2 domain as potently as the corresponding doubly charged (phosphono-
methyl)phenylalanine analogue. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
SH2 domains critically depends on strong polar inter-
actions involving the three highly charged terminal
oxygen atoms of the phosphate group of pTyr.8 For
example, replacement of pTyr, in a given phosphotyr-
osyl peptide sequence, by monocarboxylic acid mim-
etics, invariably reduces potency by several orders of
magnitude.9,10
Inhibiting the binding interactions between the tyrosine
kinase growth factor receptors and the Src homology 2
(SH2) domain of the adapter protein Grb2 constitutes a
possible strategy in the search for new anticancer drugs.1
Our e€orts in this direction have resulted in the identi®-
cation of small phosphotyrosine (pTyr)-containing
molecules that can eciently disrupt these interactions
in in vitro binding tests.2 5 However, in the perspective
of more advanced pharmacological studies, the phos-
phate group present in these inhibitors is a liability. The
two negative charges of pTyr at physiological pH can in
principle impair cell membrane penetration. In addition,
the possible hydrolysis of the phosphate ester bond by
phosphatases render pTyr-containing compounds unsuit-
able for cell and in vivo studies. Replacing pTyr in our
compounds by a less charged and phosphatase-resistant
mimeticÐand this without losing high binding anity
for the Grb2-SH2 domainÐis therefore desirable.
Obtaining stability towards phosphatases without major
loss of SH2 binding anity can be achieved by using a
(phosphonomethyl)phenylalanine (Pmp) mimetic.6,7
However, reducing the number of charges without
compromising binding anity is a far more challenging
task since recognition of phosphotyrosyl ligands by
As reported in this letter, we have addressed these issues
by designing mono charged phosphinate mimetics of
pTyr whose synthesis is described in Part 2 of this work.
The idea underlying this e€ort was to attenuate the
reduction of anity due to the removal of one charge by
exploiting the additional carbon position a€orded by a
phosphinate group to create new favourable interactions
with the SH2 domain. The designed phosphinate iso-
steres were incorporated in a short prototype inhibitory
peptide sequence of the Grb2-SH2 domain to assess
their value.
Design/Modelling
We based the design of phosphinate isosteres on the X-
ray crystal structure of the Grb2-SH2 domain in com-
plex with a peptide ligand containing Pmp as a replace-
ment for pTyr.11 Figure 1(a) shows the environment of
the phosphonate group of the Pmp moiety in this
structure.
*Corresponding author. Tel.: +41-61-67-990; fax: +41-61-696-69-29;
e-mail: pascal.furet@pharma.novartis.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00475-3

2338
P. Furet et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2337±2341
Figure 1. (a) Interactions between the phosphonate group of Pmp and
the pTyr binding pocket of Grb2-SH2 observed in a X-ray structure.
The hydrogen bonds are represented as dashed lines. (b) Chemical
structures of designed isosteres.
Two of the terminal phosphonate oxygen atoms, O1 and
O2, are involved in multiple electrostatic interactions
with residues of the phosphotyrosine binding pocket. In
particular, as previously observed in other SH2 domain
structures, salt bridges are formed with Arg bB5 and
Arg aA2.12 In contrast, the third terminal oxygen atom,
O3, does not make any direct interaction with the SH2
domain. It only forms a water mediated hydrogen bond
with the side chain of Ser BC2. This led to the idea of
removing one charge of the phosphonate group by
replacing O3 by a carbon atom. We reasoned that the
phosphinate analogue thus obtained would still have the
capacity to form the salt bridge interactions with the
arginines. In addition, the new carbon position would
allow the introduction of substituents to try to com-
pensate, by direct interaction with the SH2 domain, the
loss of the water mediated hydrogen bond with Ser BC2
and the overall weakening of electrostatic interactions
resulting from the removal of one negative charge.
Compounds 1 and 2 (Fig. 1(b)) were the ®rst phos-
phinate isosteres designed along this line. With these
compounds, we sought to create a direct hydrogen bond
with Ser BC2 by attaching a hydroxy group in the term-
inal position of the chain on the phosphorus atom.
Modelling13 indicated that spacers of one or two methyl-
ene units between the phosphorus atom and the terminal
hydroxy group would be adequate to position the
hydroxy oxygen atom within hydrogen bonding dis-
tance of the side chain of Ser BC2. However, as illus-
trated in Figure 2(a), the two carbon spacer was
considered to be more attractive because it positioned
the hydroxy group at the exact location of the water
molecule mediating the hydrogen bond between the
Figure 2. Models of phosphinate isosteres 2 (a) 3 (b) and 4 (c) inter-
acting with the phosphotyrosine pocket residues. Hydrogen bonds to
Ser BC2 are indicated as dashed lines. In 4, the absolute stereo-
chemistry of the carbon bearing the hydroxy substituent must be R to
form the hydrogen bond with Ser BC2.
phosphonate O3 atom and Ser BC2 in the X-ray struc-
ture. We speculated that this modi®cation would lead to
a displacement of the water molecule, with the asso-
ciated entropic advantage, without disrupting a hydrogen
bonding network involving other crystallographically
determined water molecules located at the bottom of the
phosphotyrosine pocket.
Keeping cell permeability considerations in mind, we
wondered whether it would be possible to use a purely
hydrophobic side chain to create new favourable con-
tacts with the protein. In previous work, we had shown

P. Furet et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2337±2341
2339
that the binding anity of phosphotyrosyl Grb2-SH2
ligands can be dramatically improved by attaching to
the N-terminus of pTyr an aromatic moiety able to
form a stacking interaction with the guanidinium group
of residue Arg aA2.5 A modelling experiment showed
that attachment of a benzyl group to the phosphorus
atom, at the position corresponding to O3 in the X-ray
structure, allowed the formation of the same stacking
interaction. As illustrated in Figure 2(b), this idea was
the basis of the design of 3, a more lipophilic phos-
phinate isostere compared to 1 or 2.
methylmorpholine in N,N-dimethylformamide. After
removal of the tert-butyloxycarbonyl protecting group
with 4 N HCl in dioxane, the dipeptide was coupled to
the required Na-Fmoc amino acids (1±5) using O-(1,2-
dihydro-2-oxo-1-pyridyl)-N,N,N0,N0-tetramethyl-uronium
tetra¯uoroborate18 and diisopropylethylamine in N-
methylpyrrolidinone. Treatment of the Fmoc-protected
compounds with a 20% solution of piperidine in N,N-
dimethylacetamide and subsequent acetylation of the free
N-terminal amino group with acetic anhydride a€orded
the target compounds 6±10.19 Compound 11 was synthe-
sised manually on a Rink amide MBHA resin,20 employ-
ing the Fmoc/tert-butyl strategy21 and coupling protocols
previously reported from our group.2,19,22
Finally, 4 was designed as a combination of both ideas:
targeting Ser BC2 for hydrogen bonding and Arg aA2
for stacking. We also envisaged the synthesis of 5 as a
reference to assess the success of our strategy aiming at
establishing new favourable interactions with the SH2
domain.
Results and Discussion
Binding anity was assessed by an ELISA-type assay
that measures the ability of a compound to inhibit the
binding of the phosphorylated C-terminal intracellular
domain of the epidermal growth factor receptor (EGFR)
to the Grb2-SH2 domain.23 IC50 values of compounds
6±10 in this assay are reported in Table 1. Also included
in the table is the IC50 value of the reference phospho-
nate analogue 11.
Chemistry
The synthesis of the Na-Fmoc derivatives of phos-
phinate isosteres 1±5 is described in the accompanying
paper (Part 2). We report here their incorporation in the
short prototype sequence Ac-Xxx-Ac6c-(1S,2R)-Achc-
NH2 (Xxx=1±5). This sequence stems from previous
work2,3 where we identi®ed suitable non-proteinogenic
amino acid replacements for residues pTyr+1 and
pTyr+214 of the minimal peptide motif recognized by
the Grb2-SH2 domain.
Our strategy to compensate the loss of one negative
charge by creating new interactions using a neutral side-
chain was successful. The potencies of phosphinate
derivatives 7, 8 and 9 are in the same range as that of
the doubly charged phosphonate analogue 11. Morever,
the IC50 values follow the trend that was expected from
our modelling studies. Replacement of one of the
hydroxy groups of the phosphonate moiety of 11 by a
simple methyl results in a signi®cant loss of anity
(compound 10), a consequence of weaker electrostatic
The (1S,2R)-2-amino-cyclohexanecarboxylic acid amide
b-amino acid (12, Achc) was prepared as depicted in
Scheme 1. rac,cis-2-Amino-cyclohexanecarboxylic acid
(13) was synthesised by known procedures15 using
sodium sul®te for the reduction of the N-chlorosulfonyl
b-lactam16 and concentrated hydrogen chloride for the
hydrolysis of the b-lactam. The cis orientation assigned
to the racemic 13 is that expected for a dipolar addition
of chlorosulfonyl isocyanate with ole®nic substrates.16
Resolution of the racemic mixture was performed by
liquid chromatography using a chiral support.17
Compounds 6±10 were synthesised in solution using
standard conditions. The enantiomerically pure cyclic b-
amino acid 12 was coupled to Na-Boc-Ac6c-OH (Ac6c,
1-aminocyclohexane carboxylic acid) using N-ethyl-N0-
(3-dimethylamino-propyl)-carbodiimide hydrochloride
and 1-hydroxybenzotriazole in the presence of N-
Figure 3. Chemical structures of target compounds.
Scheme 1. Conditions: a, (i) ClSO2NCO, DCM; (ii) Na2SO3; (iii) HClconcd, 53% (over three steps); b, (i) Cbz-Cl, Na2CO3, dioxane, 79%; chroma-
tographic chiral separation, 66% recovery; c, isobutylchloroformate, TEA, THF, 10 ꢀC then gaseous NH3, 78%; d, H2, Pd(C), MeOH, quant.

2340
P. Furet et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2337±2341
Table 1. Inhibitory activity of phosphinate derivatives Ac-Phe(4-
CH2-PO2H-R)-Ac6c-(1S,2R)-Achc-NH2 and reference phosphono-
peptide Ac-Pmp-Ac6c-(1S,2R)-Achc-NH2
R.; Panek, R. L.; Lu, G. H.; Dahring, T. K.; Rose, D. R. J.
Med. Chem. 1998, 41, 4329.
8. For a review on SH2 domain crystallographic or NMR
structures see Kuriyan, J.; Cowburn, D. Annu. Rev. Biophys.
Biomol. Struct. 1997, 26, 259.
Compounds
R
IC50, mMa
9. Burke, T. R., Jr.; Luo, J.; Yao, Z.-J.; Gao, Y.; Zhao, H.;
Milne, G. W. A.; Guo, R.; Voigt, J. H.; King, C. R.; Yang, D.
Bioorg. Med. Chem. Lett. 1999, 9, 347.
10. Beaulieu, P. L.; Cameron, D. R.; Ferland, J.-M.; Gauthier,
J.; Ghiro, E.; Gillard, J.; Gorys, V.; Poirier, M.; Rancourt, J.;
Wernic, D.; Llinas-Brunet, M.; Betageri, R.; Cardozo, M.;
Hickey, E. R.; Ingraham, R.; Jakes, S.; Kabcenell, A.; Kirrane,
T.; Lukas, S.; Patel, U.; Proudfoot, J.; Sharma, R.; Tong, L.;
Moss, N. J. Med. Chem. 1999, 42, 1757.
6
7
8
9
10
11
CH2OH
CH2CH2OH
CH2Ph
7.6
1.6
1.2
0.53
25.5
0.7
CH(OH)Phb
CH3
OH
aValues are means of three experiments.
bMixture of two diastereomers (R and S at the hydroxy substituted
benzylic carbon).
11. Rahuel, J. Unpublished results. The main di€erence between
this structure and that of the Grb2-SH2 phosphotyrosyl peptide
complex previously published (Rahuel, J.; Gay, B.; Erdmann,
D.; Strauss, A.; Garcõa-Echeverrõa, C.; Furet, P.; Caravatti, G.;
Fretz, H.; Schoepfer, J.; Grutter, M. Nature Struct. Biol. 1996, 3,
586) is a movement of the BC loop that positions Ser BC2 at a
longer distance from the ligand. As a consequence the direct
hydrogen bond existing between the phosphate group and Ser
BC2 in the latter structure becomes water mediated in the case of
the phosphonate analogue structure.
interactions with the SH2 domain. Introduction of an
alcohol chain targeting Ser BC2 for hydrogen bonding
enhances potency (6 and 7 compared to 10). The e€ect is
more pronounced with 7, in agreement with our hypoth-
esis that a water molecule is displaced. A substantial gain
in potency (20 times) is also obtained with the benzyl
substituent (8 compared to 10), which supports the
existence of the designed stacking interaction with Arg
aA2. As anticipated, combining the two favourable
structural modi®cations results in the most active deriva-
tive 9 of the series.
12. For the nomenclature of SH2 domain residues see: Lee, C.
H.; Kominos, D.; Jacques, S.; Margolis, B.; Schlessinger, J.;
Shoelson, S. E.; Kuriyan, J. Structure 1994, 2, 423.
13. The design and molecular modelling work was carried out
interactively in MacroModel using the AMBER force ®eld
(dielectric constant of 4 r) to energy minimise the protein±ligand
complexes. (a) MacroModel v.4.0: Mohamadi, F.; Richards, N.
G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.; Cau®eld, C.;
Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem.
1990, 11, 440. (b) Weiner, S. J.; Kollman, P. A.; Case, D. A.;
Singh, U. C.; Ghio, C.; Alagona, G.; Profeta, S., Jr.; Weiner,
P. J. Am. Chem. Soc. 1984, 106, 765.
Conclusion
This work demonstrates that it is possible to ®nd mono
charged isosteres of pTyr that give Grb2-SH2 domain
inhibitors as potent as that obtained with the classical
doubly charged Pmp isostere when incorporated in the
same peptide sequence. In addition to bearing only one
charge, the resulting compounds reported here do not
contain any proteinogenic amino acids. These molecules
pave the way for a new generation of Grb2-SH2 domain
blockers useful in cell and in vivo studies.
14. Ligand residues are numbered relative to the position of
the phosphotyrosine which is denoted pTyr0.
15. Malpass, J. R.; Tweddle, N. J. J. Chem. Soc., Perkin Trans. 1
1997, 874. The synthesis of an analogue of 13 (rac,cis-2-amino-
cyclohex-3-ene carboxylic acid) was previously reported by our
group in ref 2. Additional experimental SAR data on potent
Grb2-SH2 antagonists containing asparagine mimetics have
been reported in: Furet, P.; Gay, B.; Schoepfer, J.; Zeller, M.;
Rahuel, J.; Satoh, Y.; Garcõa-Echeverrõa, C. Peptides: Frontiers
of Peptide Science, Proc. Am. Pept. Symp. 2000, in press.
16. Durst, T.; O'Sullivan, M. J. J. Org. Chem. 1970, 35, 2043.
17. The resolution of the racemic mixture was carried out
using a Chiralpak AD column (5Â50 cm; 20 mm) eluted with
hexane:MeOH (24:1, v/v; 0.1% TFA) at a ¯ow rate of 150 mL/
min and detection at 210 nm (recovery: 66%). The purity of
the ®nal compounds was veri®ed by analytical HPLC on a
Chiralpak AD column (0.46Â25 cm) eluted with hexane:
MeOH (19:1, v/v; 0.1% TFA) at a ¯ow rate of 1 mL/min and
detection at 210 nm; single peak at tR=12.34 (1S,2R) min and
tR=15.35 min (1R,2S).
Acknowledgements
We gratefully acknowledge the technical assistance of Mrs.
D. Arz, Mrs. V. von Arx, Mr. R. Wille, Mrs. C. Stamm,
Mr. P. Haner and Mr. R. Vogelsanger.
References and Notes
1. Smithgall, T. E. J. Pharma. Toxico. Methods 1995, 34, 125.
2. Furet, P.; Garcõa-Echeverrõa, C.; Gay, B.; Schoepfer, J.;
Zeller, M.; Rahuel, J. J. Med. Chem. 1999, 42, 2358.
3. Furet, P.; Gay, B.; Caravatti, G.; Garcõa-Echeverrõa, C.;
Rahuel, J.; Schoepfer, J.; Fretz, H. J. Med. Chem. 1998, 41,
3442.
4. Garcõa-Echeverrõa, C.; Furet, P.; Gay, B.; Fretz, H.;
Rahuel, J.; Schoepfer, J.; Caravatti, G. J. Med. Chem. 1998,
41, 1741.
5. Furet, P.; Gay, B.; Garcõa-Echeverrõa, C.; Rahuel, J.; Fretz,
H.; Schoepfer, J.; Caravatti, G. J. Med. Chem. 1997, 40, 3551.
6. Yao, Z.-J.; King, C. R.; Cao, T.; Kelley, J.; Milne, G. W.
A.; Voigt, J. H.; Burke, T. R., Jr. J. Med. Chem. 1999, 42, 25.
7. Eaton, S. R.; Cody, W. L.; Doherty, A. M.; Holland, D.
18. Knorr, R.; Trzeciak, A.; Bannwarth, W.; Gillessen, D.
Tetrahedron Lett. 1989, 30, 1927.
19. The crude compounds 6±11 were puri®ed by medium-
pressure liquid chromatography on a C18-column using an
acetonitrile±water gradient. The purity of the ®nal compounds
was veri®ed by reversed-phase analytical HPLC on a Nucleosil
column (250Â4.0 mm; 5 mm, 100 A): linear gradient over 10min
of MeCN/0.09% TFA and H2O/0.1% TFA from 1:49 to 3:2,
¯ow rate 2.0 mL/min, detection at 215 nm; single peak at
tR=6.79 min (6), tR=6.90 min (7), tR=6.67 min (8), tR=
8.10 min (9), tR=7.19 min (10), and tR=6.65 min (11). Mass
spectral analyses (matrix-assisted laser-desorption ionisation
time-of-¯ight mass spectrometry, MALDI-TOF): 536.8 (calcd

P. Furet et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2337±2341
2341
563.6, C27H40N4O7P1; 6); 577.9 (calcd 577.6, C28H42N4O7P1;
7); 623.6 (calc. 623.7, C33H44N4O6P1; 8); 639.6 (calcd 639.7,
C33H44N4O7P1; 9); 549.7 (calcd 549.6, C27H42N4O6P1; 10);
and 549.6 (calcd 549.6, C26H38N4O7P1; 11).
22. Na-Fmoc-[4-(O-di-tert-butyl)-phosphonomethyl]-l-phenyl-
alanine was used in the solid-phase synthesis of compound 11.
Attempts to incorporate the Fmoc isosteres 1±5 on a solid
support had little success (Garcõa-Echeverrõa, C., unpublished
results).
20. Rink, H. Tetrahedron Lett. 1987, 28, 3787.
21. Atherton, E.; Sheppard, R. C. In Solid-Phase Peptide
SynthesisÐA Practical Approach; Richwood, D., Hames, B.
D., Eds.; IRL Press at Oxford University Press: Oxford, 1989.
23. Gay, B.; Furet, P.; Garcõa-Echeverrõa, C.; Rahuel, J.;
Chene, P.; Fretz, H.; Schoepfer, J.; Caravatti, G. Biochemistry
1997, 36, 5712.