Mimicking proline-rich motifs

A novel structure-guided approach in drug design to develop inhibitors for proline-rich mediated protein-protein interactions

Proline-rich sequences are amongst the most frequent motifs in eukaryotic cells and play a pivotal role in many signaling pathway. They mediate the assembly of molecular complexes by interacting with versatile recognition domains contained in various intracellular proteins. Signaling proline-rich motife recognizing domains (PRDs) include Src-homology 3 (SH3), WW, GYF, Profilin and Drosophila enabled (Ena)/vasodilator-stimulated phosphoprotein (VASP) homology 1 (EVH1) domains. Small-molecule competitors of yet so far undruggable PRDs are urgently needed for functional analysis of large-scale genomics and proteomics data.

Polyproline helix II display a trigonal profile as structural recognition motif. The threefold axial symmetry conserved two different recognition patterns PxxP or xPPx, depending on which helical face is bound by a central aromatic residue (Trp, orange). Left panel: Core recognition motif FPxxP bound by Mena EVH1 (PDB code 1EVH). Right panel: Core recognition motif PPxY bound by Nedd4 WW (PDB code 4N7H).
The initial inhibitor Ac-[2-Cl-F][ProM-2][ProM-1]-OH mimics the natural binding mode of Ena/VASP EVH1. Comparison of cocrystal structures of (a) Ac--FPPPPT-\NH and (b) inhibitor 1 bound to ENAH EVH1 mimics the binding mode of FPPPP. Both ligands are colored according to the pattern found in the SPOT array.
a) SPOT array substitution experiment (top middle) reveals that ENAH EVH1 allows replacement of the orange proline by aliphatic residues. b) The modular concept of ProM scaffolds allows adaptations of separate building blocks while keeping the rest of the ligand constant. c)-e) The orange building block was synthesized in three versions to explore the epitope underneath the proline.

We developed a modular strategy to obtain an extendable toolkit of chemical fragments (ProMs) designed to replace pairs of conserved prolines in recognition motifs. These rationally designed chemical fragments mimic dipeptide motifs (e.g. PP, xP, Px), adopting backbone angles typical of a left-handed polyproline II helix (PPII). Up to now more than 15 new chemical entities were synthezised using an innovative modular synthesis concept developed by  Prof. H.-G. Schmalz at the Universität zu Köln (Inst. Organische Synthese).  As a proof of concept we were able to develop, for the first time, low molecular weight inhibitors of Ena/VASP EVH1 domains. Our collaboration focused on the design of sp3-rich, non toxic scaffolds that are mimicking the challenging recognition of proline-rich motifs. We used an interdisciplinary approach in which scaffolds are in silico designed, new synthesis routes found, inhibitors cocrystallized and evaluated in biophysical and cell biological assays.

Structure-guided drug development relies on high-resolution structures for in-silico docking studies. Shown is the mF0-DFc difference density map contured at 3sigma revealing missing density of the inhibitor Ac-[2-Cl-F][ProM-2][ProM-3]-OH. Map calculated at 1.15A after refining the molecular replacement solution against a merged data set.
High-resolution crystal structures are obtained by micro seeding experiments. ENAH EVH1 thereby crystallizes within 3-50 days, depending on the conditions of the precipitation agent and salt additive.
Protein crystals are measured at the BESSY-II radiation light source (MX-beamlines in Adlershof, Berlin)
Protein crystals are measured at the BESSY-II radiation light source (MX-beamlines in Adlershof, Berlin)

Ena/VASP - a druggable antimetastatic target

Metastasis is a hallmark of cancer and the leading cause of mortality among cancer patients. Strikingly, the last 10 years of cancer research increased the survival rate for patients with diagnosed metastatic disease by less than 3%. The lack of metastasis-directed drugs and the limited progress of metastasis-directed drug development efforts make new approaches essential in drug design.
Cells of the primary tumor undergo a developmental regulatory program which leads to a specific signature of gene expression, which results in a migration state of cancer cells. Pivotal in this process are the Ena/VASP proteins involved in the regulation of the actin cytoskeleton.

Therapeutic inhibitors (red) are targeting upstream proteins involved in signalling pathways remodelling actin (green). Our new approach enables downstream-inhibition of Ena/VASP to interfere with the locomotion of cancer cells.
(A) Inhibitor 1 Ac-[2-Cl-F][ProM-2][ProM-1]-OH with ProM-2 in red and ProM-1 in green (right) and the corresponding peptid Ac-FPPPP-OH (left). (B) Overlay of Mena-EVH1 with Ac-FPPPPT-OH in green (PDB: 1EVH) and EnaH-EVH1 with Ligand 1 in blue (PDB: 4MY6). (C) Binding constants to VASP-, EnaH- and EVL-EVH1 domains.

In contrast to current drug targets, Ena/VASP acts most downstream of cancer signaling cascades where bypassing these routes is unlikely. Since localization of Ena/VASP depends on the EVH1 domain, the domain displays a favourable drug target. However, small molecules interfering with EVH1 remain elusive.
We succeeded in designing a 706 Da compound that bound with a Kd of 4 μM and displaced natural binding partners in pull-down experiments from a native lysate of invasive breast cancer cell line. Immunofluorescence confirmed that Ena/VASP was delocalized from cellular membrane protrusions. Furthermore, the cell-membrane-permeable compound inhibited the invasion in a boyden chamber assay by two-thirds. Based on these encouraging results, the group of Prof. van Dijke (University Leiden) contacted us. In cooperation we tested the inhibitor in an in vivo zebra fish study that verified the antimetastatic effect.
Our approach yielded in patented inhibitors that are unique and the only small molecules known to address Ena/VASP at the time.

Pulldown experiments wit GST- EnaH-EVH1 on Glutathion-sepharose beads with MDA-MB-231 lysate. The natural binding partners zyxin (ZYX) und RAPH1 are displaced by inhibitor 1 in a concentration dependent manner. Loading controls in coomassie stain. GST alone as a control.
MDA-MB-231 cells. Control (col. 1, 3) and 100 μM inhibitor 1 (col. 2, 4). Binding partners Zyxin (ZYX) and RAPH1 in top row; VASP, middle row; merge (ZYX, blue; VASP, red; colocal., purple; RAPH1, green; colocal., yellow), lower row. Treated cells show displacement of VASP from its unchanged binding partners. Scale bars: 5μm.
Invasion of breast cancer cells trough a Matrigel™ matrix along a gradient of chemoattractant in a Boyden chamber. Controls (light grey) with (left) and without (right) gradient. Tested compounds (dark grey): nsb, a non sense combination of ProMs, has no effect; Inhibitor 1b reduces invasion by 60%.

Leibniz-Forschungsinstitut für Molekulare Pharmakologie im Forschungsverbund Berlin e.V. (FMP)
Campus Berlin-Buch
Robert-Roessle-Str. 10
13125 Berlin, Germany
+4930 94793 - 100 
+4930 94793 - 109 (Fax)