Utilization of a nitrobenzoxadiazole (NBD) fluorophore in the design of a Grb2 SH2 domain-binding peptide mimetic

Article abstract of DOI:10.1016/j.bmcl.2005.01.017

Fluorescence labeling has become a general technique for studying the intracellular accumulation and localization of exogenously administered materials. Reported herein is a low nanomolar affinity Grb2 SH2 domain-binding antagonist that utilizes the environmentally-sensitive nitrobenzoxadiazole (NBD) fluorophore as a naphthyl replacement. This novel agent should serve as a useful tool to visualize the actions of this class of Grb2 SH2 domain-binding antagonists in whole cell systems.

Full text of DOI:10.1016/j.bmcl.2005.01.017

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1385–1388
Utilization of a nitrobenzoxadiazole (NBD) fluorophore in
the design of a Grb2 SH2 domain-binding peptide mimetic
Zhen-Dan Shi,a Rajeshri G. Karki,a Shinya Oishi,a Karen M. Worthy,b
Lakshman K. Bindu,b Pathirage G. Dharmawardana,c Marc C. Nicklaus,a
Donald P. Bottaro,c Robert J. Fisherb and Terrence R. Burke, Jr.a,*
aLaboratory of Medicinal Chemistry, CCR, NCI, NIH, Frederick, MD 21702, USA
bProtein Chemistry Laboratory, SAIC-Frederick, Frederick, MD 21702, USA
cUrology Oncologic Branch, CCR, NCI, NIH, Bethesda, MD 20892, USA
Received 28 October 2004; revised 4 January 2005; accepted 6 January 2005
Abstract—Fluorescence labeling has become a general technique for studying the intracellular accumulation and localization of
exogenously administered materials. Reported herein is a low nanomolar affinity Grb2 SH2 domain-binding antagonist that utilizes
the environmentally-sensitive nitrobenzoxadiazole (NBD) fluorophore as a naphthyl replacement. This novel agent should serve as a
useful tool to visualize the actions of this class of Grb2 SH2 domain-binding antagonists in whole cell systems.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Phosphotyrosyl (pTyr)-dependent protein–protein inter-
actions mediated by Src homology 2 (SH2) domains
provide important connectivity in a variety of patho-
genic signaling pathways. In particular, the Grb2 SH2
domain has been implicated in the etiology of several
cancers, including erbB-2-dependent breast cancers
and c-Met-dependent renal cancers.1 As one approach
toward a new class of anticancer therapeutic, pTyr mi-
metic-containing Grb2 SH2 domain binding antagonists
have been prepared typified by naphthylpropylamide-
containing analogues such as 1 and 2 (Fig. 1).2 Although
many of these compounds have exhibited high binding
affinity in extracellular assays, in physiological contexts
where Grb2 proteins reside in intracellular compart-
ments, persistent concerns exist regarding both the abil-
ity of these agents to transit cell membranes and their
binding selectivity once inside the cell. Fluorescence
labeling has become a general technique for studying
the intracellular accumulation and localization of exo-
genously administered materials.3 Ideally, use of this ap-
proach to study intracellular interactions of Grb2 SH2
domain-binding antagonists of type 2 would require
the preparation of highly fluorescent analogues that re-
tain characteristics of the parent structures. The current
report details the design and synthesis of nitrobenzox-
adiazole (NBD)-containing 3 as a variant of 2 poten-
tially suitable for whole cell Grb2 SH2 domain
binding studies.
1. Design and synthesis of target 3
The purpose in preparing a fluorescently-labeled Grb2
SH2 domain-binding inhibitor is to facilitate representa-
tive cellular distribution studies of the class of inhibitors
typified by parent compound 2. Therefore, structural
similarity of the fluorescent analogue with 2 was desired.
Previous studies had delineated the importance of C-
terminal 3-(arylbicyclo)propylamido functionality for
interaction with a hydrophobic patch formed by resi-
dues LeubD84 (Leu111) and PhebE3 (Phe119) of the
Grb2 SH2 domain protein.2a,5 The NBD ring system
was chosen as a fluorophore due to its bicyclic nature
and its ability to exhibit higher fluorescence when bound
to protein surfaces than when free in aqueous media.
Incorporation of the NBD unit into the inhibitor struc-
ture was achieved using a 1,2-diaminoethane linker,
which presented in target 3 the same number of linking
atoms from the (i + 2) Asn residue as provided by the
propylamino linker in parent 2.
*
Corresponding author. Tel.: +1 301 846 5906; fax: +1 301 846
6033; e-mail: tburke@helix.nih.gov
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.01.017

1386
Z.-D. Shi et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1385–1388
O
N
H
N
O
O
N
NH2
NH2
N
H
N
H
O
O
O
O
O2N
HN
HN
O
O
N
H
O
N
H
O
(HO)2P
(HO)2P
R
O
X
3
1
2
X = O ; R = NHCOCH3
X = CH2; R = CH2CO2H
OH
Figure 1. Structures of parent compounds 1 and 2 and the target fluorescent ligand 3.
Accordingly, 4 was obtained by reaction of commer-
cially-available 4-chloro-7-nitrobenzofurazane with
ethylenediamine as previously reported (Scheme 1).6
Coupling of 4 with Na-Boc-Asn-OH in the presence of
1-hydroxybenzotriazole (HOBt) and diisopropylcarbo-
diimide (DIPCDI) provided 5 (72% yield),7 which upon
Boc-deprotection using TFA gave 6 (84% yield).8 HOBt
mixed anhydride coupling of 6 with commercially avail-
able Na-Fmoc 1-aminocyclohexanecarboxylic acid (Na-
Fmoc Ac6c) gave 7 (78% yield),9 which was treated with
piperidine to yield the Na-deprotected 8 (79% yield).10
The next step required coupling of 8 with the known
pTyr mimetic 9.2c This proved troublesome, possibly
due to steric hindrance between the Ac6c ring and the
purification provided target 3 as an orange amorphous
powder.12 For binding analysis, 3 was converted to its
tri-sodium salt 3s by treatment with NaHCO3.
2. Determination of Grb2 SH2 domain-binding affinity
Binding affinity of 3s as determined using an ELISA-
based competition assay13 provided an IC50 value of
167 nM 44 nM. Complimentary analysis using surface
plasmon resonance (SPR) that measured the direct bind-
ing of 3s to chip-bound Grb2 SH2 domain protein was
conducted using a Biacore S51 instrument with biotinyl-
ated protein attached to neutravidin chips (Fig. 2).14 In
reasonable agreement with the ELISA results, a KD
value of 119 nM 0.6 nM was obtained, with kinetic
parameters being kon = 1.18 · 106 4.6 · 103 Mꢀ1 sꢀ1
and koff = 1.40 · 10ꢀ1 4.5 · 10ꢀ4 sꢀ1. This can be
contrasted to SPR data recently reported for the parent
analogue 2, where a KD value of 23.8 nM 1.8 nM
t
bulky a-(CH2CO2 Bu) group. However, application of
active ester coupling using 1-hydroxy-7-azabenzotri-
azole (HOAt) and 1-ethyl-3-(30-dimethylamino)pro-
pyl)carbodiimide (EDCI)) at elevated temperature
(50 ꢁC) produced 10 in 37% yield.11 Global removal of
But-ester protecting groups by treatment with TFA in
the presence of 1,2-ethanedithiol followed by HPLC
was obtained, with kinetic parameters being kon
=
O
O
N
H
N
N
H
N
O
α
N
-Boc Asn-OH
i
N
N
NH2
NH2
N
H
O
O2N
O2N
HNR
4
R = Boc
R = H
5
ii
6
O
OH
O
O
α
iii
N
-Fmoc Ac6c
O
t
(Bu O)2P
O
N
H
N
N
H
N
O
O
N
N
NH2
NH2
t
Bu O
N
N
9
H
O
O
H
O
O2N
O2N
HN
HN
O
v
O
N
H
R
N
H
O
(RO)2P
O
OR
7
8
R = Fmoc
R = H
iv
10
3
t
R = Bu
R = H
vi
3s
vii
Scheme 1. Reagents and conditions: (i) DIPCDI, HOBt, DMF, rt, overnight (72% yield); (ii) TFA, CH2Cl2 1 h (84% yield); (iii) HOBt, DIPCDI,
DMF, rt, overnight (78% yield); (iv) piperidine, CH3CN, rt, 1 h (79% yield); (v) EDCIÆHCl, HOAt, DMF, 50 ꢁC, 24 h (37% yield); (vi) TFA,
HS(CH2)2SH, H2O, rt, 1 h (42% yield); (vii) NaHCO3(aq) (quantitative).

Z.-D. Shi et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1385–1388
1387
Figure 2. Using a Biacore S51 instrument, a series of 5, 10, 20, 30, 50,
75, 100, 250, 500, 750, and 1000 nM injections of 3s were passed at a
flow rate of 30 lLminꢀ1 over a neutravidin chip containing 8024 RUꢀs
of biotinylated Grb2 SH2 domain protein. Data represents the results
of two duplicate injections, with mathematical fitting to a one
component binding model indicated in red.
Figure 3. Hypothetical Grb2 SH2 domain binding modes of 2 and 3.
Coloring of atoms is: carbon—yellow (for 2) or dark green (for 3);
nitrogen—blue; oxygen—red; hydrogen—white. Hydrophilic and
hydrophobic surface maps are colored in turquoise and orange,
respectively, with certain key residues closest to the surface maps being
displayed in wire representation.
4.18 · 106 2.88 · 103 Mꢀ1 sꢀ1 and koff = 9.91 · 10ꢀ2
2.94 · 10ꢀ3 sꢀ1
.
The 5-fold loss of binding affinity in-
14
curred by replacement of the 1-naphthyl ring system
by the NBD ring system was due primarily to a 3.5-fold
reduction in on-rate, with less effect arising from the off-
rate (1.4-fold increase).
module in the Insight II program.19 The default Delphi
formal charges were applied for the protein, while for 2
and 3 charges were obtained using ab initio calculations
with Gaussian 0320 (results not shown). It was observed
that the electrostatic potential of the 2-naphthyl ring
system of 2 was complementary to the protein surface
potential proximal to it, thereby promoting binding
interactions. In contrast, the electrostatic potentials on
the NBD moiety of 3 and on the hydrophilic protein
surface proximal to it were both electronegative, thus
promoting less favorable interactions. The noncomple-
mentary electrostatic potentials observed between the
NBD group of 3 and Grb2 SH2 domain protein may
contribute to its reduced binding affinity relative to 2.
3. Modeling studies
To compare the hypothetical binding modes of 2 and 3,
molecular dynamics simulations were performed on pre-
minimized protein–ligand complexes using Macromodel
8.6.15 Chain A from the crystal structure of mAZ-pY-
(aMe)pY-N-NH2 bound to the Grb2 SH2 domain
(1JYQ.pdb)16 was used as the starting geometry for
the modeling study. All minimizations and simulations
were performed using the Merck Molecular Force Field
(MMFFs)17 and a continuum solvation model for H2O
(GB/SA).18 Both compounds exhibited all binding inter-
actions that have previously been shown by X-ray crys-
tallography to be important for this class of compound.
However, two major differences were observed between
2 and 3. First, the a-CH2COOH of 3 was H-bonded to
GlnbD3 (Gln106), while the same group did not exhibit
any H-bonding interactions in 2. Although this addi-
tional H-bonding interaction should have favored the
binding of 3 as compared to 2, interactions of the C-ter-
minal NBD moiety of 3 were not as favorable as for the
naphthyl group of 2. While the C-terminal naphthylpro-
pyl portion of 2 made van der Waals interactions with a
hydrophobic region defined by residues LeubD8 and
PhebE3, the NMD group of 3 was found to occupy a
site closer to a hydrophilic region on the surface of the
Grb2 SH2 domain, and even here significant binding
interactions were not noted (Fig. 3).
In summary, title compound 3 represents a low nano-
molar affinity Grb2 SH2 domain-binding inhibitor that
utilizes the environmentally-sensitive NBD fluorophore.
Although modeling studies suggest that the level of
bonding interactions afforded by the NBD ring system
may not be as great as those provided by the naphthyl
ring, which it replaces, this novel agent should serve as
a useful tool to investigate the actions of this class of
Grb2 SH2 domain-binding antagonists in whole cell
systems.
Acknowledgements
Appreciation is expressed to Drs. Christopher Lai and
James Kelley of the LMC for mass spectral anaysis.
References and notes
In order to further understand the binding modes of 2
and 3, the electrostatic potential on the solvent-accessi-
ble surface of the Grb2 SH2 domain protein as well as
for compounds 2 and 3 were calculated using the Delphi
1. (a) Daly, R. J.; Binder, M. D.; Sutherland, R. L. Oncogene
1994, 9, 2723; (b) Garbay, C.; Liu, W. Q.; Vidal, M.;
Rogues, B. P. Biochem. Pharmacol. 2000, 60, 1165; (c)
Dankort, D.; Maslikowski, B.; Warner, N.; Kanno, N.;

1388
Z.-D. Shi et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1385–1388
Kim, H.; Wang, Z. X.; Moran, M. F.; Oshima, R. G.;
(s, 1H), 3.48 (m, 1H), 2.97 (d, 2H, J = 21.4 Hz), 3.03–2.92
(m, 3H), 2.67 (dd, 1H, J = 4.4 Hz & 15.5 Hz), 2.57–2.50
(m, 2H), 2.32 (dd, 1H, J = 2.7 Hz & 17.6 Hz), 1.97 (m,
1H), 1.85–1.15 (m, 10H), 1.40 (s, 9H), 1.39 (s, 9H), 1.37 (s,
9H). FABMS (+Ve) m/z 915 [MH+].
Cardiff, R. D.; Muller, W. J. Mol. Cell. Biol. 2001, 21,
1540; (d) Saucier, C.; Papavasiliou, V.; Palazzo, A.;
Naujokas, M. A.; Kremer, R.; Park, M. Oncogene 2002,
21, 1800; (e) Osada, S.; Saji, S. Curr. Med. Chem., Anti-
Cancer Agent 2003, 3, 119; (f) Feller, S. M.; Tuchscherer,
G.; Voss, J. Leukemia Lymphoma 2003, 44, 411.
12. Compound 3:
H NMR (d6-DMSO) d 9.38 (t, 1H,
J = 6.0 Hz), 8.51 (d, 1H, J = 9.0 Hz), 8.29 (s, 1H), 7.70–
7.66 (m, 2H), 7.35 (s, 1H), 7.16–7.09 (m, 4H), 6.89 (s, 1H),
6.44 (d, 1H, J = 9.0 Hz), 4.33 (m, 1H), 3.58–3.52 (m, 2H),
3.41–3.37 (m, 2H), 3.10 (m, 1H), 2.97 (dd, 1H, J = 5.0 Hz
& 13.3 Hz), 2.89 (d, 2H, J = 21.3 Hz), 2.62 (dd, 1H,
J = 6.6 Hz & 15.6 Hz), 2.50 (dd, 1H, J = 5.1 Hz &
15.5 Hz), 2.43 (dd, 1H, J = 10.4 Hz & 16.9 Hz), 2.32 (dd,
1H, J = 9.9 Hz & 13.1 Hz), 1.98 (dd, 1H, J = 4.1 Hz &
16.8 Hz), 1.84–1.07 (m, 10H). FABMS (+Ve) m/z 745
[MꢀH].
2. (a) Furet, P.; Gay, B.; Caravatti, G.; Garcia-Echeverria,
C.; Rahuel, J.; Schoepfer, J.; Fretz, H. J. Med. Chem.
1998, 41, 3442; (b) Yao, Z. J.; King, C. R.; Cao, T.; Kelley,
J.; Milne, G. W. A.; Voigt, J. H.; Burke, T. R. J. Med.
Chem. 1999, 42, 25; (c) Wei, C.-Q.; Li, B.; Guo, R.; Yang,
D., Jr.; Burke, T. R., Jr. Bioorg. Med. Chem. Lett. 2002,
12, 2781.
3. Numerous reports have appeared detailing the use of
fluorescent ligands to study protein interactions within
cellular contexts. For recent examples, see: (a) Berque-
Bestel, I.; Soulier, J.-L.; Giner, M.; Rivail, L.; Langlois,
M.; Sicsic, S. J. Med. Chem. 2003, 46, 2606; (b) Ettmayer,
P.; Billich, A.; Baumruker, T.; Mechtcheriakova, D.;
Schmid, H.; Nussbaumer, P. Bioorg. Med. Chem. Lett.
2004, 14, 1555.
13. ELISA procedure: Streptavidin-coated 96 well plates were
first blocked with 0.2% I-Block (Tropix Applied Biosys-
tems, Bedford, MA). A recombinant chimeric protein
encompassing the human Grb2 SH2 domain sequence
(residues 60–151, accession number P29354) fused in
frame with the complete maltose binding protein sequence
and a carboxyl terminal 6-histidine tag was expressed in E.
coli, biotinylated, and purified by conventional methods.
Approximately 1 mg/mL of protein solution was diluted to
10 lg/mL in 0.1% BSA/TPBS before adding to appropri-
ate wells (100 lL/well). Plates were then incubated at RT
for 2 h followed by three washes with TPBS. To each well
was added human epidermal growth factor receptor-
derived Grb2 SH2 domain recognition sequence (NH2-
biotin-DDTFLPVPE-pYINQSVPK-COOH) conjugated
to horse radish peroxidase (HRP, 100 lL at 20 lg/mL)
in the presence or absence of varying concentrations of 3s,
then the plates were incubated for 1 h at rt. Following
incubation, plates were washed with TPBS, then to each
well was added TPBS (50 lL) and ELISA detection
reagent OPD (50 lL; Abbott Laboratories Diagnostic
Division, Abbott Park, IL; Cat # 0617230). Absorbance
was measured in an ELISA plate reader at 450 nm. A
minimum of four replicate wells was analyzed for each
antagonist concentration. Regression analysis, IC50 val-
ues, and statistically significant differences were deter-
mined using GraphPad InStat Software analysis of two or
more independent ELISA experiments.
4. Nomenclature as suggested by Eck et al. Nature 1994,
362, 87.
5. Schoepfer, J.; Fretz, H.; Gay, B.; Furet, P.; Garcia-
Echeverria, C.; End, N.; Caravatti, G. Bioorg. Med. Chem.
Lett. 1999, 9, 221.
6. Cotte, A.; Bader, B.; Kuhlmann, J.; Waldmann, H. Chem.
Eur. J. 1999, 5, 922.
7. Compound 5: H NMR (d6-DMSO) d 9.18 (m, 1H), 8.53 (d,
1H, J = 8.9 Hz), 7.73–7.70 (m, 2H), 7.25 (s, 1H), 6.86 (s,
1H), 6.45 (d, 1H, J = 9.1 Hz), 4.15 (m, 1H), 3.53–3.50 (m,
2H), 3.46–3.41 (m, 2H), 2.42 (dd, 1H, J = 5.5 Hz &
15.1 Hz), 2.34 (m, 1H), 1.38 (s, 9H). FABMS (+Ve) m/z
438 [MH+].
8. Compound 6: H NMR (d6-DMSO) d 8.84 (m, 1H), 8.54 (d,
1H, J = 8.5 Hz), 7.90 (m, 1H), 7.67 (m, 1H), 7.45 (m, 1H),
7.20 (s, 1H), 6.50 (d, 1H, J = 8.6 Hz), 4.01 (m, 1H), 3.64–
3.47 (m, 4H), 2.73 (dd, 1H, J = 5.0 Hz & 15.3 Hz), 2.63
(dd, 1H, J = 7.8 Hz & 15.4 Hz). FABMS (+Ve) m/z 338
[MH+].
9. Compound 7: H NMR (CDCl3) d 8.13 (d, 1H, J = 8.6 Hz),
7.98 (d, 1H, J = 8.6 Hz), 7.77 (m, 1H), 7.64 (dd, 2H,
J = 0.6 Hz & 7.6 Hz), 7.43 (dd, 2H, J = 1.0 Hz & 7.6 Hz),
7.33–7.29 (m, 3H), 7.22–7.18 (m, 3H), 5.84 (s, 1H), 5.79 (d,
1H, J = 8.4 Hz), 5.38 (s, 1H), 4.68 (m, 1H), 4.26 (dd, 1H,
J = 7.0 Hz & 10.5 Hz), 4.20 (dd, 1H, J = 7.2 Hz &
10.5 Hz), 4.08 (m, 1H), 3.56 (m, 1H), 3.42–3.31 (m, 2H),
3.24 (m, 1H), 3.08 (dd, 1H, J = 4.4 Hz & 15.7 Hz), 2.40
(dd, 1H, J = 4.5 Hz & 15.6 Hz), 2.09–1.18 (m, 10H).
FABMS (+Ve) m/z 685 [MH+].
14. Oishi, S.; Karki, R. G.; Kang, S.-U.; Wang, X.; Worthy,
K. M.; Bindu, L. K.; Nicklaus, M. C.; Fisher, R. J.; Burke,
T. R., Jr. J. Med. Chem., in press.
15. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.;
Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11,
440.
10. Compound 8:
H
NMR (CD3OD)
d
8.36 (d, 1H,
16. Nioche, P.; Liu, W. Q.; Broutin, I.; Charbonnier, F.;
Latreille, M. T.; Vidal, M.; Roques, B.; Garbay, C.;
Ducruix, A. J. Mol. Biol. 2002, 315, 1167.
17. Halgren, T. A. J. Comput. Chem. 1996, 17, 490.
18. Qiu, D.; Shenkin, P. S.; Hollinger, F. P.; Still, W. C.
J. Phys. Chem. A 1997, 101, 3005.
J = 8.8 Hz), 6.32 (d, 1H, J = 8.9 Hz), 4.67 (m, 1H), 3.68–
3.53 (m, 4H), 2.78 (dd, 1H, J = 6.7 Hz & 15.4 Hz), 2.72
(dd, 1H, J = 5.9 Hz & 15.5 Hz), 1.93–1.30 (m, 10H).
FABMS (+Ve) m/z 463 [MH+].
11. Compound 10:
H NMR (CDCl3) d 8.34 (d, 1H,
J = 8.6 Hz), 7.98 (s, 1H), 7.88 (s, 1H), 7.81 (d, 1H,
J = 7.9 Hz), 7.15 (dd, 2H, J = 2.4 Hz & 8.1 Hz), 7.03 (d,
2H, J = 7.9 Hz), 6.98 (s, 1H), 6.66 (s, 1H), 6.12 (d, 1H,
J = 8.2 Hz), 5.73 (s, 1H), 4.56 (m, 1H), 3.77 (m, 1H), 3.64
19. Computational results were obtained using the software
program Insight II by Accelrys Inc., San Diego, CA
(http://www.accelrys.com/).
20. Frisch, M. J.; et al. Gaussian 03, Revision B.04 2003.