A selective irreversible inhibitor targeting a PDZ protein interaction domain

Article abstract of DOI:10.1021/ja035540l

Irreversible inhibitors of proteases have proven themselves useful tools for determining which proteases are active under given conditions in tissues or cells and for studying the functional role that a protease plays in physiological processes. The application of such techniques to the study of the activity and function of protein-protein interactions has been hindered by the lack of guiding principles for the mechanistic design of irreversible inhibitors targeting the "active site" of a protein interaction. We report herein the first example of a mechanism-based irreversible inhibitor of a protein interaction that has been specifically targeted to one member of the PDZ family of protein interaction domains: the second PDZ domain of the membrane-associated guanylate kinase MAGI3. This inhibitor was designed using rationally directed computational evaluation to take advantage of a conserved histidine in the PDZ domain by introducing an ionizable group that will be held in close proximity to that nucleophile during binding. The novel compound exhibits all of the characteristics of an irreversible inhibitor of the interaction of the tumor suppressor PTEN with MAGI3 in in vitro models. In cells, the inhibitor can be shown to release PTEN from sequestration by MAGI3 and consequently upregulate the PKB signaling pathway. Copyright

Full text of DOI:10.1021/ja035540l

Published on Web 09/11/2003
A Selective Irreversible Inhibitor Targeting a PDZ Protein Interaction Domain
Naoaki Fujii,† Jose J. Haresco,§ Kathleen A. P. Novak,† David Stokoe,| Irwin D. Kuntz,§ and
R. Kiplin Guy*,†,‡,|
Department of Pharmaceutical Chemistry, Department of Cellular and Molecular Pharmacology, Laboratory for
Molecular Dynamics and Design, and Cancer Research Institute, UniVersity of California at San Francisco,
Genentech Hall, Mission Bay, 600 16th Street 2280, San Francisco, California 94143-2280
Received April 9, 2003; E-mail: rguy@cgl.ucsf.edu
The MAGIs are a small family of adaptors that are widely
expressed in the human body and function to maintain signaling
that regulates tissue organization and differentiation. MAGI3 binds
to the tumor suppressor PTEN, a lipid/protein phosphatase,1,2 and
colocalizes PTEN with the oncoprotein Akt/PKB, a protein kinase.3
In this context, PTEN prevents activation of PKB and suppresses
PKB signaling, whereas release of PTEN from MAGI-3 increases
Akt/PKB signaling. Normally, Akt/PKB signaling ensures cell
survival during response to cellular insults by suppressing apop-
tosis.4 However, PTEN mutants that cause constitutive Akt/PKB
Figure 1. The design of 1. Panel A shows an overlay of the peptide ligand
for PSD95-PDZ3 (gray) and the designed small molecule inhibitor of ligand
binding 1 (green). Panel B shows the structure of 1 and the rationalization
signaling have been associated with human cancers.5 The exact role
of the interaction of PTEN and MAGI in oncogenesis has been
controversial. Chemical disruption of this interaction would be a
unique and temporally controlled way to investigate the role of
Akt/PKB signaling in transformation and cancer. We report here
the structure-based design, synthesis, and evaluation of a potent
irreversible inhibitor of the interaction of PTEN and MAGI3 -
the first specifically targeted irreversible inhibitor of a protein
interaction.
of the placement of the ionizable hydroxyl group to capture the conserved
histidine in the binding site.
Proteins in the MAGI family contain six PDZ protein interaction
domains.6-11 MAGI3 binds to PTEN using its second PDZ domain
(MAGI3-PDZ2). Class I PDZ domains, such as MAGI3-PDZ2,
recognize the carboxy terminus of their binding partner (ligand)
through the consensus sequence X-S/T-X-V and include a conserved
histidine in the PDZ domain that interacts with the ligand’s -2
residue.12 The -1 and -3 residues of the ligand help determine
specificity for individual PDZ domains within the class. The high
structural homology between PDZ domains, with or without bound
ligand, suggested the feasibility of designing compounds targeted
to MAGI3 PDZ2 domain based on the crystal structure of liganded
PSD-95 PDZ313 using a combination of visual inspection and
DOCK.
The distances and angles between the R carbons of residues (0)
and (-1) of the PSD95-PDZ3 ligand CRIPT (amino acid sequence
Y-K-Q-T-S-V), bound to PSD95, were measured using Weblab and
matched by inspection to those of simple rigid cores. Potential
scaffolds were used to generate a virtual library with diversity
positions designed to mimic the crucial side chains, and the overall
predicted fit of these virtual compounds to the PDZ domain was
evaluated using DOCK.14,15 Initial studies indicated that positioning
an ionizable group near the conserved histidine in the PDZ domain
might produce an irreversible inhibitor. DOCKing experiments with
3-hydroxymethylindole 1 indicated that it could both place critical
functional groups for binding in proximity to their orientation in
the bound ligand and properly orient the hydroxymethyl group to
trap the histidine (Figure 1).
Figure 2. Synthesis of 1: (a) MeI (2 equiv), K2CO3 (3 equiv), DMF, 40
°C, 1 h. (b) H2, 10% Pd(C), MeOH, room temperature, 2 h, two steps quant.
(c) ICl (1.6 equiv), CaCO3 (3 equiv), MeOH, H2O, room temperature, 1 h,
34%. (d) 1-heptyne (5 equiv), PdCl2(PPh3)2 (0.15 equiv), CuI (0.3 equiv),
Et2NH (large excess), DMF, room temperature, 2 h. (e) PdCl2(PhCN)2 (0.2
equiv), DMF, 80 °C, 40 min, two steps overall 87%. (f) POCl3 (1.3 equiv),
DMF, 5 °C, quant. (g) BrCH2CH2Ph (10 equiv), Cs2CO3 (5 equiv), DMF,
room temperature, 19 h, 44%. (h) NaBH4 (excess), MeOH, 5 °C. (i) NaOH,
H2O, MeOH, 65 °C, 12 h.
The synthesis of 1 proceeded smoothly (Figure 2). Sequential
esterification of 2-methyl-5-nitrobenzoic acid, reduction of the nitro
group, and ortho iodination gave methyl 5-amino-4-iodo-2-methyl
benzoate. Crosscoupling with 1-heptyne gave alkynylaniline 3,
which afforded indole 4 after heating with palladium. Formylation
of 4 gave 5. Alkylation of 5 with phenylethyl bromide gave 6.
Sequential reduction of the aldehyde and hydrolysis of the ester
then produced 1. While compound 1 is unstable under acidic
conditions, it is stable in a neutral environment. Treatment of 1
with a large excess of imidazole (pH 7, PBS) afforded an imidazole
adduct, as expected (data not shown).
Compound 1 irreversibly blocks ligand binding to the MAGI3-
PDZ2 domain (Figure 3). Pretreatment of MAGI3-PDZ2 with 1
blocked ligand binding in a dose-dependent manner (Figure 3A).
While efficacy of peptide ligand binding dropped with increasing
concentrations of 1, ligand affinity did not change - consistent
† Department of Pharmaceutical Chemistry.
‡ Department of Cellular and Molecular Pharmacology.
§ Laboratory for Molecular Dynamics and Design.
| Cancer Research Institute.
9
12074
J. AM. CHEM. SOC. 2003, 125, 12074-12075
10.1021/ja035540l CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
allow full activation of PKB. Treatment of HCT116 cells expressing
endogenous PTEN, MAGI, and PKB with 1 results in activation
of PKB activity by roughly 3-fold, consistent with in vivo changes
(Figure 3C). Normal activation of PKB involves phosphorylation
of serine 473.16 Increased levels of phospho-Ser473-PKB were
detected in HCT116 cells treated with 1 by immunoblot using an
anti-phospho-Ser473 antibody (Figure 3D), thus indicating that the
effects on PKB signaling seem likely to be induced by blockade
of PTEN activity. At the concentrations that induce these effects,
there were no observable effects upon cell viability or morphology
not associated with PKB activity. These studies suggest that
inhibitor 1 is capable of physiologically inducing PKB signaling
through the normal cellular mechanism and that the reagent should
prove useful in studies of PKB mediated transformation.
The widespread occurrence of PDZ domains as organizers of
signaling pathways makes them an important subject for biological
studies. Compound 1 is the first cell permeable specific inhibitor
of PDZ domain function. The route used to produce 1 allows the
production of highly diverse libraries of analogues targeted to
discovering class- and domain-selective inhibitors. These materials
should have use both as functional probes of the activity of PDZ
domain containing proteins and as proteomics reagents17 for
isolation of active PDZ domains in cells. Studies toward these ends
are currently underway and will be reported in due course.
Acknowledgment. We acknowledge The Stewart Trust (R.K.G.,
K.A.P.N., N.F.), NIH (I.D.K., J.J.H., GM31497& GM56531), and
Fujisawa Pharmaceuticals (N.F.) for funding.
Figure 3. The activity of 1. Panel A shows that pretreating a GST‚MAGI3-
PDZ2 fusion protein with increasing concentrations of 1 prior to allowing
binding of a fluorescently labeled PTEN carboxy terminal peptide lowers
the amount of peptide ligand binding as monitored by fluorescence
polarization. Panel B shows that treating a GST‚MAGI3-PDZ2 fusion
protein with the biotin conjugate of 1 gives a covalently attached biotin
conjugate of the protein, as demonstrated by SDS-PAGE purification of
the adduct and avidin‚HRP blotting, and that this conjugation is inhibited
by pretreatment with 1. The GST control immunoblot shows there is no
change in overall protein present in each sample. Panel C shows that
treatment of HCT116 cells with 1 causes a 3-fold increase in endogenous
PKB activity, as measured by substrate turnover after immunopurification
of the enzyme from the cells. Panel D shows that the same treatment also
leads to phosphorylation of the endogenous PKB, as measured by blotting
of the cellular lysate with an antibody specific to phospho-PKB.
Supporting Information Available: Synthetic procedures, bio-
chemical protocols, and cellular protocols and activity data (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Wu, X.; Hepner, K.; Castelino-Prabhu, S.; Do, D.; Kaye, M. B.; Yuan,
X. J.; Wood, J.; Ross, C.; Sawyers, C. L.; Whang, Y. E. Proc. Natl. Acad.
Sci. U.S.A. 2000, 97, 4233-4238.
(2) Wu, Y.; Dowbenko, D.; Spencer, S.; Laura, R.; Lee, J.; Gu, Q.; Lasky,
L. A. J. Biol. Chem. 2000, 275, 21477-21485.
(3) Mayo, L. D.; Donner, D. B. Trends Biochem. Sci. 2002, 27, 462-467.
(4) Brazil, D. P.; Hemmings, B. A. Trends Biochem. Sci. 2001, 26, 657-
664.
(5) Maehama, T.; Dixon, J. E. Trends Cell Biol. 1999, 9, 125-128.
(6) Mino, A.; Ohtsuka, T.; Inoue, E.; Takai, Y. Genes Cells 2000, 5, 1009-
1016.
(7) Strochlic, L.; Cartaud, A.; Labas, V.; Hoch, W.; Rossier, J.; Cartaud, J.
J. Cell Biol. 2001, 153, 1127-1132.
(8) Dobrosotskaya, I.; Guy, R. K.; James, G. L. J. Biol. Chem. 1997, 272,
31589-31597.
(9) Aarts, M.; Liu, Y.; Liu, L.; Besshoh, S.; Arundine, M.; Gurd, J. W.; Wang,
Y. T.; Salter, M. W.; Tymianski, M. Science 2002, 298, 846-850.
(10) Sheng, M.; Sala, C. Annu. ReV. Neurosci. 2001, 24, 1-29.
(11) Fanning, A. S.; Anderson, J. M. J. Clin. InVest. 1999, 103, 767-772.
(12) Harris, B. Z.; Hillier, B. J.; Lim, W. A. Biochemistry 2001, 40, 5921-
5930.
with an irreversible inhibitor. Additionally, treatment of MAGI3-
PDZ2 with a biotin conjugate of 1 afforded a covalent adduct whose
formation was disrupted by pretreatment of MAGI3-PDZ2 with 1
(Figure 3B). This result is consistent with covalent modification
of a specific site on the protein by 1. Mass spectrometry of MAGI3-
PDZ2 treated with 1 revealed a single modified species with a mass
consistent with the addition of a single equivalent of 1 (see
Supporting Information). Mutation of the targeted conserved
histidine in the PDZ domain to alanine gave a stable protein that
was unreactive with reagent 1, indicating that the single site of
modification is most probably the targeted histidine (see Supporting
Information). As is consistent with the weak ionization potential
of the indole alcohol, complete reaction between MAGI3-PDZ2
and 1 required either high excess of 1 (as shown) or extended
reaction times (see Supporting Information).
(13) Doyle, D. A.; Lee, A.; Lewis, J.; Kim, E.; Sheng, M.; MacKinnon, R.
Cell 1996, 85, 1067-1076.
(14) Brooijmans, N.; Kuntz, I. D. Annu. ReV. Biophys. Biomol. Stuct. 2003,
32, 335-373.
(15) Zou, X.; Sun, Y.; Kuntz, I. D. J. Am. Chem. Soc. 1999, 121, 8033-8043.
(16) Stokoe, D.; Stephens, L. R.; Copeland, T.; Gaffney, P. R.; Reese, C. B.;
Painter, G. F.; Holmes, A. B.; McCormick, F.; Hawkins, P. T. Science
1997, 277, 567-570.
(17) Greenbaum, D.; Baruch, A.; Hayrapetian, L.; Darula, Z.; Burlingame, A.;
Medzihradszky, K. F.; Bogyo, M. Mol. Cell Proteomics 2002, 1, 60-68.
Functional inhibitors of the PTEN‚MAGI3-PDZ2 interaction
should prevent PTEN recruitment to the plasma membrane and
JA035540L
9
J. AM. CHEM. SOC. VOL. 125, NO. 40, 2003 12075