Designed Nanomolar Small-Molecule Inhibitors of Ena/VASP EVH1 Interaction Impair Invasion and Extravasation of Breast Cancer

Designed nanomolar small-molecule inhibitors of Ena/VASP EVH1 interaction impair invasion and extravasation of breast cancer cells Matthias Baronea , Matthias Muller¨ a, Slim Chihab, Jiang Renc, Dominik Albatb, Arne Soickeb, Stephan Dohmenb, Marco Kleinb, Judith Brunsb, Maarten van Dintherc, Robert Opitza, Peter Lindemanna, Monika Beerbauma, Kathrin Motznya, Yvette Rosked, Peter Schmiedera , Rudolf Volkmera, Marc Nazare´ a , Udo Heinemannd,e , Hartmut Oschkinata,e, Peter ten Dijkec,1, Hans-Gunther¨ Schmalzb,1 , and Ronald Kuhne¨ a,1

aDepartment of Structural Biology, Leibniz-Forschungsinstitut fur¨ Molekulare Pharmakologie, 13125 Berlin, Germany; bDepartment of Chemistry, Universitat¨ zu Koln,¨ 50939 Koln,¨ Germany; cOncode Institute, Department of Cell and Chemical Biology, Leiden University Medical Center, 2333 Leiden, The Netherlands; dDepartment of Macromolecular Structure and Interaction, Max-Delbruck-Centrum¨ fur¨ Molekulare Medizin, 13125 Berlin, Germany; and eInstitute of Chemistry and Biochemistry, Freie Universitat¨ Berlin, 14195 Berlin, Germany

Edited by Samuel H. Gellman, University of Wisconsin–Madison, Madison, WI, and approved September 14, 2020 (received for review April 16, 2020)

Battling metastasis through inhibition of cell motility is con- The three vertebrate Ena/VASP family members, enabled sidered a promising approach to support cancer therapies. In homolog (ENAH), VASP, and Ena-VASP-like (EVL), share a this context, Ena/VASP-depending signaling pathways, in partic- tripartite structural organization in which two Ena/VASP homol- ular interactions with their EVH1 domains, are promising targets ogy domains (EVH1 and EVH2) are separated by a more for pharmaceutical intervention. However, protein–protein inter- divergent proline-rich central part. Interactions of the EVH2 actions involving proline-rich segments are notoriously difficult domain are involved in the elongation and protection of barbed- to address by small molecules. Hence, structure-based design end actin filaments from capping proteins and tetramerization efforts in combination with the chemical synthesis of additional (31, 32). EVH1 folds into a structured globular domain that molecular entities are required. Building on a previously devel- interacts with proteins at focal adhesions (33), the leading edge oped nonpeptidic micromolar inhibitor, we determined 22 crystal (34, 35), and invadopodia (36, 37) by recognizing the motif structures of ENAH EVH1 in complex with inhibitors and ratio- [F/W/L/Y]PxφP (35, 38) (φ hydrophobic, x any; SI Appendix, nally extended our library of conformationally defined proline- Fig. S3) in poly-L-proline type II helix (PPII) conformation. derived modules (ProMs) to succeed in developing a nanomolar In the course of our research into small molecules as poten- inhibitor (Kd = 120 nM, MW = 734 Da). In contrast to the previ- tial inhibitors of protein–protein interactions (39) we recently ous inhibitor, the optimized compounds reduced extravasation of in silico designed and stereo-selectively synthesized scaffolds, invasive breast cancer cells in a zebrafish model. This study rep- coined ProMs, which mimic pairs of prolines in PPII conforma- resents an example of successful, structure-guided development tion (40). The modular combination of different ProMs thereby of low molecular weight inhibitors specifically and selectively allowed us to generate nonpeptidic secondary-structure mimetics addressing a proline-rich sequence-recognizing domain that is that fulfill the steric requirements of the addressed proline-rich characterized by a shallow epitope lacking defined binding pock- motif-recognizing domain (41–47). For the EVH1 domain, our ets. The evolved high-affinity inhibitor may now serve as a proof-of-concept study yielded a canonically binding, nontoxic, tool in validating the basic therapeutic concept, i.e., the sup- pression of cancer metastasis by inhibiting a crucial protein– Significance protein interaction involved in actin filament processing and cell migration. Protein–protein interactions mediated by proline-rich motifs are involved in regulation of many important signaling cas- protein–protein interactions | metastasis | small molecules | cades. These motifs belong to the most abundant recognition peptide mimetics | proline-rich motif motifs in the eukaryotic genome and preferentially adopt a left- handed polyproline helix II, a secondary structure element that etastasis is a complex multistep process (1, 2) employing, has been notoriously difficult to mimic with small molecules. Mamong others, mechanisms governing actin cytoskeleton Here, we present a structure-guided design effort yielding a dynamics involving integrin signaling and actin regulatory pro- toolkit of chemical entities that enables rational construction of teins (3–5). So far, all approved antimetastatic drugs antagonize selective small molecule inhibitors for these protein domains. integrins (6) or inhibit downstream kinases (7, 8) (SI Appendix, We succeeded in developing an inhibitor for the Ena/VASP pro- Fig. S1). In the metastatic setting however, these drugs appear to tein family that is active in vivo and reduces extravasation of have only limited success (9–13) and 5-y survival is not increasing invasive breast cancer cells in a zebrafish model. satisfactorily (14, 15), making new approaches in antimetastatic drug development essential to meet this urgent medical need. Author contributions: M. Barone, P.t.D., H.-G.S., and R.K. designed research; M. Barone, M.M., S.C., J.R., D.A., A.S., S.D., M.K., J.B., M.v.D., R.O., P.L., M. Beerbaum, K.M., Y.R., P.S., The enabled/vasodilator stimulated phosphoprotein protein R.V., M.N., U.H., and H.O. performed research; M. Barone and R.O. analyzed data; and family (Ena/VASP) acts as a crucial hub in cell migration by M. Barone, R.O., H.O., P.t.D., H.-G.S., and R.K. wrote the paper.y linking actin filaments to invadopodia and focal adhesions (16– The authors declare no competing interest.y 22). Due to their role in the transformation of benign lesions This article is a PNAS Direct Submission.y into invasive and metastatic cancer, Ena/VASP proteins are dis- This open access article is distributed under Creative Commons Attribution-NonCommercial- cussed as part of the invasive signature and as a marker of breast NoDerivatives License 4.0 (CC BY-NC-ND).y carcinogenesis (23–25). At the advanced tumor stage, the pro- 1 To whom correspondence may be addressed. Email: [email protected], p.ten tein family is overexpressed (26–28), which has been shown to [email protected], or [email protected] increase migration speed in vivo (29) and to potentiate invasive- This article contains supporting information online at https://www.pnas.org/lookup/suppl/ ness (30). Yet, no sufficiently potent probes to interfere with doi:10.1073/pnas.2007213117/-/DCSupplemental.y Ena/VASP in vivo have been reported. First published November 12, 2020.

29684–29690 | PNAS | November 24, 2020 | vol. 117 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.2007213117 Downloaded by guest on September 26, 2021 Downloaded by guest on September 26, 2021 aoee al. et Barone -emnlPo- nt(re)b rM9 rM1,Po-2 rM1,or ProM-15, ProM-12, ProM-13, the 1). ProM-9, replacing (Table by by ProM-17 achieved (green) was unit (pink) ProM-1 variation C-terminal Structural inhibitors study. (modu- preorganized this General in conformationally used (B) nonpeptidic, of (40). terminus C architecture permeable the lar) cell-membrane of inhibitors Esterification (red). the scaffold renders ProM-2 central a to attached the share compositions All 1. Fig. of terminus C the to adjacent from ActA monocytogenes of Listeria TEDEL with epitope binding complex C-terminal in well-resolved EVH1 successful (SI ENAH inhibitors report of different we structures crystal Here inhibitor resolution experiments. of vivo optimization was structure-based in improvement 2-chloro-(L)- for further a and inhibitor required domains, EVH1 synthetic scaffolds Ena/VASP the toward ProM While different (40). 1 (2-Cl-Phe) two phenylalanine of posed inhibitor 706-Da cell-membrane-permeable, stesatn on o iadotmzto.I ersnsthe represents It motif optimization. recognition ligand core for point pentapeptide starting the as Considerations. Initial 1. Table in of means K in by described assessed as was titration EVH1 binding fluorescence ENAH (47) The toward published. ProM-17 (45), inhibitors separately the and as ProM-2 of (47), synthesized 43), were ProM-15 1B) (41, (46), (Fig. ProM-1 ProM-13 (46), scaffolds (40). ProM-9 assembly ProM protocols modular coupling required established by using The synthesized blocks Inhibitors. building were the study of of this in Characterization employed and Synthesis Results extravasation cell cancer inhibiting 1 by studies. only preclinical i.e., at vivo, zebrafish in in act also shown were to Ena/VASP structural against compounds conserved Potent cell-membrane we permeability. and surface nontoxicity, weight, protein molecular flat low simplicity, rather affin- a the increasing against drastically ity While 1A). the (Fig. of scaffold modifications ProM-1 synthesized stereo-selectively and designed d ersnstecmon ihtehgetrpre affinity reported highest the with compound the represents r given are (∆G) binding upon energy free of change the and tutr ftefis-eeainEaVS V1inhibitor EVH1 Ena/VASP first-generation the of Structure (A) smnindaoe inhibitor above, mentioned As N ,teeypvn h a o future for way the paving thereby µM, 4) el dnie neato sites interaction identified Newly (48). alsS1–S6 Tables Appendix, aeyae -hoopeyaaieui (blue) unit 2-chloro-phenylalanine -acetylated 1 eeadesdb nsilico in by addressed were h obtained The Appendix. SI wt1 1 ftehigh-affinity the of ae n2 high- 22 on based 1 h compounds The ,icuigthe including ), com- 1A) (Fig. 1 served 1. .Tecvrg rao h attoprolines two last the of area coverage (SI The solvent 4 S8). from Fig. entirely site Appendix, sidechains interaction ligand hydrophobic The the C). S4 shield Fig. acid 2 Appendix , amino (SI (Fig. hydrophobic Met14 long contact a ing to TEDEL, of needed pitch is helix Phe77 the where to groove, binding contacts hydrophobic the of toward site extends interaction polar 8 terminal The by separated S4C). is Fig. Appendix, mutating (SI Consequently, W23NH. over group 9 of carboxylate length the suggested miss- sidechain conformation loop inhibitors the helical of that The structure 2B). water (Fig. complex conserved TEDEL every ing highly in a found displacing molecule group thereby carboxylate (W23NH), the Trp23 that shows surface of protein the and loop 2A S4–S8). (Fig. Figs. motif recognition Appendix, core SI the to vicinity terminal close the in surface positioning and solvent the to TEDEL that revealed an unambiguously adopts structures The 1.0 to S4A). up Fig. limits resolution with structures crystal substituting replacing However, failed. EVH1 Ac– 10-mer wild-type TEDEL-Containing cated with Complex in EVH1 Chimeras. ENAH of inhibitor Structures Crystal parent the the for TEDEL precondition of the a analogs affinity. as first of of surface we details protein design structural Therefore, the the featureless. with of interaction elucidating nearly side are C-terminal on is studies the focused groove docking on binding surface and protein the (50), the failed C-terminally since of unreliable had mechanism peptides binding gain to the elongated attempts into epi- earlier insights binding Noteworthy, (49). structural additional EVH1 an amino VASP represent of terminal tope TEDEL the of that Glu-Leu to presumption acids previous respect a with confirmed ing loss C-terminal affinity the that observed indicated selec- residues with TEDEL−NH flanking studies C-terminal Binding or peptides kJ/mol. to –8 truncated contributes about tively significantly with residues binding flanking the the of peptide truncation chimeric the and -emnlPo- ntt loadesteadtoa,s far so additional, the sites. address interaction nonpolar also and polar to uncovered, unit analogs ProM-1 improved TEDEL- of C-terminal inhibitor design parent of the task the in next knowledge of structures the this EVH1, exploit crystal ENAH to with was complex high-resolution in ligands of containing means by groove binding in 2D). the (Fig. to unaltered canonically bound TEDEL- still remains affinity-driving and inhibitors S4B) elongated independent conformation Fig. (SI Appendix , two (SI TEDEL energy chimeras be ProM-containing binding the to ProM- of as Interestingly, appear kJ/mol S11). elements TEDEL –3.5 Fig. and and of S8 1 gain and a S7 Tables to Appendix, leads which 1 ActA-derived r and Pro to Glu lsraayi fteitrcinbtenteTEDEL the between interaction the of analysis closer A aigaaye h otiuino h terminal the of contribution the analyzed Having 9 10 l om yrgnbn otebcbn md of amide backbone the to bond hydrogen a forms Glu e to Leu 9 5 10 s erae h idn fnt yafco f3 of factor a by affinity binding the decreased Asp .Due S3). Fig. Appendix , (SI specificity high with Pro 5 nta tepst orsalz h -emnlytrun- N-terminally the cocrystallize to attempts Initial α r a eelre yrpaigte ihProM- with them replacing by enlarged be can Pro e.Tepthi anydfie yMt4and Met14 by defined mainly is patch The Leu. hlcllo ofrainepsn cdcresidues acidic exposing conformation loop -helical 4 10 Pb rM1icesdteafiiyadyielded and affinity the increased ProM-1 by PP 2 PNAS l eut nnal he ie ekrbinding weaker times three nearly in results Ala wt2 oto a lotsll epnil o the for responsible solely almost was portion rmannoa ac htcnat the contacts that patch nonpolar a from A ˚ Tbe1.Cmaigteafiiisof affinities the Comparing 1). (Table | oebr2,2020 24, November pcfial,w ogtt oiythe modify to sought we Specifically, 1. 2 2∆N 1 oadEA V1rvasthat reveals EVH1 ENAH toward FPPPPTEDEL−NH 9 l snee opoel place properly to needed is Glu or 1 h yTpo -lPeand/or 2-Cl-Phe or Trp by Phe 2∆C .Cneunl,mutat- Consequently, C). | 2 o.117 vol. akn ihrteN- the either lacking Tbe1.Tefind- The 1). (Table 9 Loe h protein the over EL 1 IAppendix , (SI A 5 ˚ | ihimproved with r and Pro 2 o 47 no. 9 ihENAH with Lresidues EL | 10 9 29685 Leu Glu and 1

BIOPHYSICS AND COMPUTATIONAL BIOLOGY Table 1. Afﬁnities of various inhibitors toward ENAH EVH1 determined by ﬂuorescence titration

Ligand composition Kd, µM ∆G, kJ/mol wt1 Ac–FPPPP–OEt 153 (8) –21.9 (0.1)

wt2 Ac–SFEFPPPPTEDEL−NH2 13.0 (0.6) –27.9 (0.1) 1 Ac–[2-Cl-Phe][ProM-2][ProM-1]–OEt 4.1 (0.3) –30.8 (0.2) 1a Ac–[2-Cl-Phe][ProM-2][ProM-1]–OMe 4.4 (0.7) –30.6 (0.4) 1b Ac–[2-Cl-Phe][ProM-2][ProM-1]–OH 2.3 (0.2) –32.2 (0.2)

2 Ac–SFE[2-Cl-Phe][ProM-2][ProM-1]TEDEL−NH2 0.15 (0.02) –38.9 (0.2) 2∆N Ac–[2-Cl-Phe][ProM-2][ProM-1]TEDEL−NH2 0.33 (0.04) –37.0 (0.3) 2∆C Ac–SFE[2-Cl-Phe][ProM-2][ProM-1]–OEt 2.9 (0.2) –31.6 (0.1) 3a Ac–[2-Cl-Phe][ProM-2][ProM-12]–OMe 15 (1) –27.6 (0.2) 3b Ac–[2-Cl-Phe][ProM-2][ProM-12]–OH 13.5 (0.4) –27.8 (0.1) 4a Ac–[2-Cl-Phe][ProM-2][ProM-15]–OMe 0.47 (0.03) –36.1 (0.1) 4b Ac–[2-Cl-Phe][ProM-2][ProM-15]–OH 0.32 (0.04) –37.0 (0.3) 5a Ac–[2-Cl-Phe][ProM-2][ProM-17]–OMe 0.53 (0.09) –35.8 (0.4) 6 Ac–[2-Cl-Phe][ProM-2][ProM-9]–OEt 0.38 (0.05) –36.6 (0.3) 6b Ac–[2-Cl-Phe][ProM-2][ProM-9]–OH 0.12 (0.02) –39.5 (0.3) 6c Ac–[2-Cl-Phe]PP[ProM-9]–OH 0.23 (0.05) –37.9 (0.5) 7 Ac–[2-Cl-Phe][ProM-2][ProM-13]–OEt 0.12 (0.01) –39.5 (0.3)

Kd and ∆G values are given with the standard errors in parentheses.

Inhibitors 3, 4, and 5 Access the Polar Hot Spot. To mimic the polar Inhibitors 6 and 7 Cover the Hydrophobic Patch. The results interaction of 9Glu with W23NH, docking studies suggested obtained with the ProM-12–, ProM-15–, and ProM-17–derived that the configurational inversion of the C-terminal carboxyl inhibitors emphasize the particular importance to properly cover substituent could improve binding through directly contacting the hydrophobic binding site spanned by Phe77 and Met14 rather the H2O bound to W23NH (Fig. 2B). We therefore synthe- than accessing the H2O bound to W23NH (Fig. 3D). As indi- sized ProM-12 (epi-ProM-1) and the corresponding inhibitors cated by docking studies we envisioned that this hydrophobic 3a and 3b (methyl ester or free acid; Table 1) with the termi- hot spot could be specifically accessed by equipping 1 with an nal carboxylic group now pointing toward the protein surface. Even though these inhibitors established a 2.7-A-long˚ H-bond to the bound water molecules as intended (Fig. 3 A and B), both inhibitors 3a and 3b lost significant affinity relative to their parent congeners 1a and 1b (Table 1). This decreased affinity is reflected in the complex structures of 3a and 3b with resolution limits up to 1.42 A˚ by a detachment of the modified C-terminal part of the inhibitors from the hydrophobic protein surface locally defined by the Trp23 and Phe77 sidechains (Fig. 3B). To address the polar hot spot while keeping the hydrophobic contact in the C-terminal area undisturbed, docking studies suggested to formally open the C-terminal pyrrolidine ring of inhibitor 1 or 3. Such a modification opened the possibility to probe the underlying hydrophobic binding groove with different moieties (ProM-15 and ProM-17; Fig. 1B) (47). The result- ing less rigid inhibitors 4a and 5a showed a greatly improved affinity toward ENAH EVH1 by –5.5 kJ/mol compared to 1a, thereby nearly restoring the binding energy of the TEDEL- containing chimera 2∆N (Table 1). The complex structures of 4a and 4b with resolution limits up to 1.1 A˚ confirmed that the “ethyl finger” of ProM-15 indeed established the contact with the hydrophobic binding groove while the ester function maintained a 2.7-A-long˚ H-bond contact with the water bound to W23NH (Fig. 3C and SI Appendix, Fig. S9). Even though inhibitor 5a was expected to further enhance affinity by cov- Fig. 2. Crystal structure of chimera 2∆N in complex with ENAH EVH1 ering the hydrophobic area with an isobutyl group, it bound reveals the TEDEL loop binding close to the main binding groove through just as well as 4a. The structure of 5a in complex with ENAH polar and nonpolar interactions. (A) Interaction sites of terminal EL. 2∆N EVH1 at a resolution limit of 0.78 A˚ revealed the sterically more is displayed as ribbon and transparent sticks, color coded by a secondary demanding ProM-17 scaffold being a bit too bulky, causing the structure element: PPII (blue) and α-helix (yellow). (B) The backbone amide nitrogen of W23 is contacted by either 9Glu or a conserved water molecule inhibitor to be N-terminally shifted away from the ideal posi- (superposition of 19 H2O from 12 asymmetric units). (C) The hydrophobic tion (Fig. 3 C and D). Consequently, the H-bond contact to site provided by M14 and F77 is reachable only by a long aliphatic amino the W23NH-bound water widened to 3.0 A.˚ Due to the lack of acid such as 10Leu. (D) Superposition of inhibitor 1b (white) and 2∆N (blue- affinity gain, inhibitor 5a was not investigated further in cellular yellow). View along the main binding groove reveals only minor ligand assays. rearrangements.

29686 | www.pnas.org/cgi/doi/10.1073/pnas.2007213117 Barone et al. Downloaded by guest on September 26, 2021 Downloaded by guest on September 26, 2021 rtisadcnre htalihbtr eercgie ythe by recognized were inhibitors off-target all to that mode confirmed and binding proteins allowed the turn into in insights This 5A). structural (Fig. additional WW1 Yap1 and Profilin, SH3, of Fyn affinities binding tra- the perturbation assess to peak jectories assigned and experiments NMR (HSQC) We proteins. sensible motif-recognizing EVH1 used proline-rich Ena/VASP other toward to specificity compared Natural retained with inhibitors Compete mized Ena/VASP. and of Selective Partners Are Binding Inhibitors Optimized The ethyl additional the by 4 (Fig. Phe77 ProM-13 and of substituent Met14, Trp23, boost covering erly affinity observed the and from Thus, bias displacements. conformational marginal little only showed scaffolds optimized both inhibitors 4 with (Fig. Superposition hot intended hydrophobic the as accessed spot indeed scaffolds modified the that inhibitors with plex (51) efficiency ligand the improved S9). Table and Appendix, 4% (SI than less by weight Table Appendix, (SI S9 domains binding EVH1 inhibitor Ena/VASP 734-Da acid all potent to amino a canonically with flanking up eight coming in of and energy residues binding in succeeded entire we the substituent restoring ethyl single a of chimera introduction untruncated culated the as strong equally bound to used and (46) synthesized were inhibitors desired 1B), ProM- the i.e., (Fig. prepare ProM-1, ProM-13 of ring. derivatives and pyrrolidine alkylated 9 the C-terminal reason the this at For substituent alkyl additional spheres). (red ( molecules contact ProM-15 water maintaining bound of while respective groove their capability hydrophobic to the the address illustrating to (5a) view ProM-17 Detailed (D) green). unfavor- to comparison an of in to sidechain Superposition W23 associated the of spheres) from substituent detachment (red able molecules carboxyl water the bound of respective contact disturbed for (blue-white-red the hydrophobicity by inhibitors coded 12–containing of color Superposition is sur- (A) EVH1 hydrophilic–hydrophobic). solvent-accessible ENAH The of inhibitors. face ProM-17–containing and ProM-15–, 3. Fig. aoee al. et Barone outperformed inhibitor inhibitors Noteworthy, the (7). both lowered that and substantially showed then surements ihrslto rsa tutrso NHEH ncom- in EVH1 ENAH of structures crystal High-resolution .Tefra lyaino inhibitor of alkylation formal The ). 1 (K rsa tutrso NHEH ncmlxwt ProM-12–, with complex in EVH1 ENAH of structures Crystal d 4.1 1 2∆N )to µM) H- 15 wie ihinhibitors with (white) eeoula igeqatmcoherence quantum single heteronuclear N 6b 3a 7 and (K and A 7 d 3b and–. Jmlrltv to relative kJ/mol –8.7 gained and 0.12 7 efis etdwehrteopti- the whether tested first We C ealdve highlighting view Detailed (B) (green). 2∆N, 6 K t1.02 at and d B and )ceryrslsfo prop- from results clearly µM) n 2 nM 120 and (6) nM 380 to and or 1b, D). 7 1 4a Tbe1.Afiiymea- Affinity 1). (Table eouinconfirmed resolution A ˚ i.S10 ). Fig. Appendix, SI 2∆N nrae t molecular its increased 1, lgtgen and green) (light 4a 4a wie n h ProM- the and (white) and , 3a ihihe that highlighted hs ycal- by Thus, 2. 2∆N and 7 loagainst also 3b wie.(C (white). 5a otheir to and 4a) 1 (dark and 1 ) etsc steJra ellst.Inhibitors lysate. cell environ- Jurkat complex the more specificity a as in such The held ment EVH1. also experiments Ena/VASP HSQC in for seen also selectivity but boosts the affinity desirable increased highly yielded inhibitors only optimized not mization the for prominent decreased 4a more was drastically and EVH1, decrease with ENAH to this compared bound proteins At off-target inhibitors the ). S21 to all and affinity sec- time, S13 PPII Figs. same mimic Appendix, the to (SI scaffolds WW1, elements ProM Yap1 structure and ondary of Profilin, ability SH3, the Fyn corroborating of sites interaction canonic y H,wietesnhtcppieAc–RPLPPLP−NH for peptide measured (52) a synthetic with ADAP the partner while binding SH3, natural Fyn the with competing not h oeaeo 7 n W23. and F77 of coverage inhibitors the of Superposition (D) displacements. marginal by pyrrolidine addressed inhibitors C-terminal spot of the hot sition apolar at the alkylation hydrophilic– accesses additional ring the for how (blue-white-red shows ENAH hydrophobicity (white) by of (A coded surface hydrophobic). color solvent-accessible is The EVH1 inhibitors. ProM-13–containing 4. Fig. atesmgtafc h naieeso acrcls Indeed, cells. cancer of binding inhibitors invasiveness we with natural MDA–MB–231 the (56), incubating migra- with affect matrix interaction cell might extracellular their partners during of inhibiting degradation structures that and expected protrusive 18) of (16, tion formation the in Invasion. Cell Can- cer Breast Ena/VASP-Mediated Inhibit Potently Inhibitors Optimized 0.67 concen- IC50 inhibitory of half-maximal (IC50) (55) the the tration that of Zyxin revealed Analysis and 5B). blot (34), (Fig. Western manner RIAM and for concentration-dependent (16), compete (1) a RAPH1 in to initial with able the EVH1 were that ENAH inhibitor find (7) to structure-optimized lysate the experiments cell pulldown MDA–MB–231 vitro in with performing by EVH1 Ena/VASP assessed of which we first efficiency At (MDA–MB–231), the (54). challenge therapeutic cells major a cancer represent on breast and effect receptor-, their HER2-negative) progesterone investigated receptor-, we (estrogen activity, triple-negative biological expected the pulldown; vitro (in S22 concentration Fig. same Appendix, the at SH3 GST-Fyn hmtxsaogaftlbvn eu FS rdeti a in gradient (FBS) serum bovine fetal a along chemotaxis opoewehrtedvlpdcmonsas exhibit also compounds developed the whether probe To and rsa tutrso NHEH ncmlxwt rM9 and ProM-9– with complex in EVH1 ENAH of structures Crystal 7 al S10 Table Appendix, (SI s 12 vs. µM PNAS and K 7 7 sEaVS rtisaecnrlyinvolved centrally are proteins Ena/VASP As d 7 and ). ocmeewt nw idn atesfor partners binding known with compete to f8.2 of uepsto finhibitors of Superposition B) | a 8tmssalrta htof that than smaller times 18 was µM, 1b oebr2,2020 24, November wie eel itecnomtoa isand bias conformational little reveals (white) P ipae DPfo bead-bound from ADAP displaced µM < 0.0002). ,sgetn hti iioopti- silico in that suggesting ), | 10 o.117 vol. 1, e in Leu 4a 1, 6b , 4a 7 and 6, (C 2∆N. | and and , o 47 no. and 6, 4a 7 7 Superpo- ) with highlights 1 reduced | 7 (mean 2 were 29687 (53) 2∆N SI

BIOPHYSICS AND COMPUTATIONAL BIOLOGY Fig. 5. In vitro studies with the initial and optimized inhibitors. (A) The 1H- 15N HSQC titration experiments to assess the off-target specificity of 1, 4a, and 7. Each inhibitor was tested against a selection of domains recognizing proline-rich segments (Fyn SH3, Profilin, and Yap1 WW) and peak perturbations farther than 0.1 ppm were taken into account. Plotted are fitted Kd on a logarithmic scale, together with the according affinities for ENAH EVH1 taken from Table 1. (B) Pulldown experiments and fluorescence-based analysis of GST-tagged ENAH EVH1 from MDA–MB–231 cell lysates treated with 1 or 7. Natural binding partners RAPH1, RIAM, and Zyxin are displaced by the inhibitors from ENAH EVH1 in a concentration-dependent manner with different IC50 values (P = 1.6e-4).

dose-dependent manner (Fig. 6A and SI Appendix, Fig. S23). data indicate that the structure-based optimization of our EVH1 The calculated IC50 values confirmed that optimized inhibitors inhibitors (Fig. 1B) not only resulted in a gain of binding affin- exhibit a superior potency compared to 1 (Fig. 6 A, Right; mean ity but also opened options to fine-tune their pharmacological IC50 1, 88 µM; 4a, 53 µM; 6, 24 µM; and 7, 26 µM). The properties. increased potency of 7 was also confirmed in a Boyden chamber invasion assay (SI Appendix, Fig. S24). To exclude possible toxic Discussion effects, we tested inhibitors 1, 4a, 6, and 7 in MDA–MB–231 but Based on recent evidence that Ena/VASP is a critical factor in also Ena/VASP triple-knockout mouse fibroblastic MVD7 and metastatic cancer progression, we targeted its EVH1 adapter Ena/VASP triple-knockout mouse melanoma cell lines. None domain which recognizes proline-rich segments on a shallow, of the cell lines showed toxic effects at 180 µM (SI Appendix, rigidly shaped, and nearly featureless protein surface. The chal- Fig. S25). lenge to develop a small molecule inhibitor for this seemingly To determine whether the increased potency seen for “undruggable” protein–protein interaction was successfully met inhibitors 4a, 6, and 7 also holds in vivo, we investigated their by combining X-ray crystallography, in silico design, and chem- effect on the extravasation of MDA–MB–231 cells in trans- ical synthesis of modular inhibitors displaying a preorganized genic zebrafish embryos (fli:EGFP) (57) (Fig. 6A). Simulating PPII secondary structure. To extend our toolbox of conforma- the late stage of the metastatic cascade, we measured the ability tionally defined proline-derived synthetic scaffolds (ProMs) we of MDA–MB–231 cells to extravasate into the avascular collagen first identified two affinity-driving hot spots on the surface of matrix-rich tailfin (58). Counting the number of cells that have ENAH EVH1 by characterizing the binding mode of TEDEL- extravasated and invaded the tailfin tissue, we could indeed ver- elongated ligands. In a second step we then specifically modified ify that inhibition of the Ena/VASP-mediated processes caused our initial inhibitor 1 to exploit these interaction sites: The pronounced antiinvasive effects (Fig. 6B). In contrast to the ini- ProM-15– and ProM-17–derived inhibitors 4a and 5a, designed tial inhibitor 1, which at 1 µM concentration did not show any to contact the discovered polar hot spot, showed an impressive effect compared to zebrafish treated with phosphate-buffered gain of affinity (Kd of 470 and 530 nM, respectively). These saline (PBS) (P > 0.2), the structure-optimized inhibitors signifi- compounds were even surpassed by the ProM-9– and ProM- cantly reduced cancer cell extravasation (4a and 7, P < 0.0001; 6, 13–derived inhibitors 6 and 7, designed to cover a hydrophobic P < 0.005). Surprisingly, the weaker binding inhibitor 4a exhib- patch, the latter binding to ENAH EVH1 with a Kd of 120 nM. ited a similarly strong in vivo activity to that of 7 probably due to Thus, by calculated introduction of a single ethyl substituent different secondary properties such as cell permeability. These close to the C terminus of 1 we could gain as much binding

29688 | www.pnas.org/cgi/doi/10.1073/pnas.2007213117 Barone et al. Downloaded by guest on September 26, 2021 Downloaded by guest on September 26, 2021 aoee al. et Barone eufrdb asn vraHLa 66 uedx7 ieexclusion size and 75 column Superdex exchange 16/60 ion HiLoad Sepharose a with SP over overnight column. an passing cleaved on was by GST rebuffered loaded GSH. and with glutathione eluted a thrombin on and loaded was matrix fraction Sepharose soluble the centrifugation and cation in proteins Purification. Protein Methods and Materials assays. inhibitors cellular EVH1 beyond also Ena/VASP cell activity synthetic biological cancer that exhibit proved inhibit 1 we at to way, model zebrafish compounds This a in optimized vivo in the extravasation of ability the the potency inhibitor increased initial as the with to cells such compared MDA–MB–231 environment of invasion complex inhibitors the a optimized with All in lysate. the and cell vitro Jurkat for in selectivity both pro- recognizing teins, increased proline-rich related to exhibit compared domain EVH1 and Ena/VASP partners known interaction with compete residues ligand inhibitors acid structure-optimized chimeric amino flanking the eight the of by contributed as energy zebrafish in extravasation cells cancer MDA–MB–231 (B) arrows). group. (white a tailfin 3 as avascular with plotted the curves are into IC50 and extravasate fitted assay inhibitors (red) the to MDA–MB–231 or of normalized (control) increase count, PBS signal object with object phase treated phase (representative linear manner during dose-dependent calculated a were in inhibited is Invasion 6. Fig. nuyeceoai sa ihihbtrtetdMAM–3.(Left MDA–MB–231. inhibitor-treated with assay chemotaxis IncuCyte (A) inhibitors. optimized and initial the with studies vivo in and Cellular .coli E. L1D3 rw n2Tmdu vrih.Atrsonifi- After overnight. medium 2YT in grown BL21(DE3) n/APEH oan eeepesda GST-fusion as expressed were domains EVH1 Ena/VASP utemr,w hwdta the that showed we Furthermore, 2. ial,w ol demonstrate could we Finally, 1. 1, 4a, or 6, 4a 7 t1 at , concentration. µM and 6, ocnrto.(C concentration. µM 7 interfere ersnaiefloecneiae fzbas mro ae ndy5. day on taken embryos zebrafish of images fluorescence Representative ) agtseicatbde a edotb ursec-ae secondary fluorescence-based scanner. a infrared an by on out antibody lysate. read 4 at was cell beads antibodies on Jurkat target-specific overnight inhibitors incubating or adding lysate by the achieved MDA–MB–231 to was partners using binding natural beads of Displacement 4B Sepharose tathione Assay. Pulldown Sparky. eercre naBue V0 t30Kwt 6sas aasz 1,024 size spectra data NMR scans, HSQC 16 TCEP. with sodium K mM 300 mM 2 at 40 AV600 and in Bruker ( NaCl, a K mM on 298 recorded 100 were at 7.3, calorimetry pH titration phosphate, isothermal or titration positions Studies. in Binding atomic rendered five were images the and using Q79CD and match Chimera. Y16CG, body F77CG, rigid W23N, with acces- W23NE, by the in K and superposed listed 300 Protein statistics were are the to refinement at codes and replace- 293 deposited Diffraction molecular sion and at databank. suite, by program EVH1 Base solved PHENIX Data ENAH were the by structures mg/mL refined All ment, 26 excess. to ligand different 8 and microseeding Crystallization. 1 H) × 5 ( 256 15 rsaswr bandb itn ao ifso using diffusion vapor sitting by obtained were Crystals ) n ekprubto rjcoiswr rce in tracked were trajectories perturbation peak and N), S-NHEH rGTFnS3wr moiie nglu- on immobilized were SH3 GST-Fyn or EVH1 GST-ENAH iscaincntnswr eemndvafluorescence via determined were constants Dissociation PNAS | oebr2,2020 24, November ope structures Complex S1–S6. Tables Appendix, SI | σ o.117 vol. ◦ ahdln) (Right line). dashed .TeWsenbo with blot Western The C. | o 47 no. | IC50s ) 29689 )

BIOPHYSICS AND COMPUTATIONAL BIOLOGY IncuCyte Assay. Migration of 2,000 cells per well was monitored by the 34 ◦C and the numbers of MDA–MB–231 cells that extravasated individually IncuCyte S3 live cell imaging system over a gradient of 0.25% FBS (insert) from circulation into the collagen fibers of the tail fin were counted 5 d to 1% FBS (lower chamber) in Dulbecco’s Modified Eagle Medium (DMEM) postinjection. high-glucose medium, with the inhibitors supplied in both chambers. The experiment was set up as a triplet for the inhibitor concentrations and Data and Code Availability. All study data are included in this article and SI quadruplets for the positive (PBS) and negative controls (20 µM PI3K Appendix. inhibitor LY294002). Raw data were averaged and processed with a script ACKNOWLEDGMENTS. We thank K. Franke, K. Piotukh, M. Neuenschwander, written in R language. J. O. Jost, T. Bock-Bierbaum, K. Le Cong, R. Metternich, M. Macias, and J. Faix for their valuable input and discussion. This work is supported by Zebrafish Assay. mCherry-labeled MDA–MB–231 were injected into the duct the German Federal Ministry of Education and Research (16GW0186K) and of curvier of transgenic zebrafish embryos (fli:EGFP) and administrated with the Cancer Genomics Center Netherlands. J.R. is supported by the China corresponding inhibitor treatment. Zebrafish embryos were maintained at Scholarship Council.

1. I. J. Fidler, S. Paget, The pathogenesis of cancer metastasis: The ‘seed and soil’ 31. C. Bachmann, L. Fischer, U. Walter, M. Reinhard, The evh2 domain of the vasodilator- hypothesis revisited. Nat. Rev. Cancer 3, 453–458 (2003). stimulated phosphoprotein mediates tetramerization, f-actin binding, and actin 2. P. Friedl, S. Alexander, Cancer invasion and the microenvironment: Plasticity and bundle formation. J. Biol. Chem. 274, 23549–23557 (1999). reciprocity. Cell 147, 992–1009 (2011). 32. J. E. Bear et al., Antagonism between Ena/VASP proteins and actin filament capping 3. L. Blanchoin, R. Boujemaa-Paterski, C. Sykes, J. Plastino, Actin dynamics, architecture, regulates fibroblast motility. Cell 109, 509–521 (2002). and mechanics in cell motility. Physiol. Rev. 94, 235–263 (2014). 33. J. E. Bear et al., Negative regulation of fibroblast motility by Ena/VASP proteins. Cell 4. M. Vicente-Manzanares, C. K. Choi, A. R. Horwitz, Integrins in cell migration–the actin 101, 717–728 (2000). connection. J. Cell Sci. 122, 199–206 (2009). 3