Progress in Biophysics and Molecular Biology · 2017. 1. 23. · How to awaken your nanomachines: Site-specific activation of focal adhesion kinases through ligand interactions Katarzyna

lable at ScienceDirect

Progress in Biophysics and Molecular Biology 119 (2015) 60e71

Contents lists avai

Progress in Biophysics and Molecular Biology

journal homepage: www.elsevier .com/locate/pbiomolbio

How to awaken your nanomachines: Site-specific activation of focaladhesion kinases through ligand interactions

Katarzyna W. Walkiewicz a, Jean-Antoine Girault b, c, d, Stefan T. Arold a, *

a King Abdullah University of Science and Technology (KAUST), Division of Biological and Environmental Sciences and Engineering, ComputationalBioscience Research Center (CBRC), Thuwal, Saudi Arabiab Inserm, UMR-S 839, F-75005 Paris, Francec Universit�e Pierre & Marie Curie (UPMC), Sorbonne Universit�es, F-75005 Paris, Franced Institut du Fer �a Moulin, F-75005 Paris, France

a r t i c l e i n f o

Article history:Received 2 January 2015Received in revised form7 June 2015Accepted 14 June 2015Available online 17 June 2015

Keywords:Multidomain proteinAdhesomeX-ray crystallographyMolecular dynamicsHybrid methodsCalcium signalling

Abbreviations: FAs, focal adhesions; PR, proline ricPPII, protein-protein interaction inhibitor.* Corresponding author.

E-mail address: [email protected] (S.T. Aro

http://dx.doi.org/10.1016/j.pbiomolbio.2015.06.0010079-6107/© 2015 The Authors. Published by Elsevier

a b s t r a c t

The focal adhesion kinase (FAK) and the related protein-tyrosine kinase 2-beta (Pyk2) are highly versatilemultidomain scaffolds central to cell adhesion, migration, and survival. Due to their key role in cancermetastasis, understanding and inhibiting their functions are important for the development of targetedtherapy. Because FAK and Pyk2 are involved in many different cellular functions, designing drugs withpartial and function-specific inhibitory effects would be desirable. Here, we summarise recent progressin understanding the structural mechanism of how the tug-of-war between intramolecular and inter-molecular interactions allows these protein ‘nanomachines’ to become activated in a site-specificmanner.© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Contents

1. Multidomain proteins are versatile nanomachines and promising drug targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612. Multiple cellular functions of FAK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613. New insights into the structure and regulation of individual FAK domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614. Structural insight into full-length FAK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635. Functional implications of FAK dimerisation through the FERM domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 646. FERM and FAT synergise to stabilise FAK dimerisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647. The multifunctional basic motif of the FERM domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 648. pH dependent activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659. Site-specific functions of FAK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

9.1. FAK in the cytoplasm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659.2. FAK at FAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669.3. FAK in the nucleus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

10. Pyk2 idiosyncrasies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6610.1. Conserved mechanism despite differences? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6610.2. Nuclear Pyk2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

h region; SH, Src homology; FERM, band 4.1, ezrin, radixin, merlin; FAT, focal adhesion targeting; FAH, FAT homology;

ld).

Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 61

11. Targeted inhibition of FAK and Pyk2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6812. Conclusions & remaining challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

1. Multidomain proteins are versatile nanomachines andpromising drug targets

Flexible multidomain proteins are key players in eukaryotic sig-nalling networks. In response to the presence of ligands, post-translational modifications or subcellular conditions, these mole-cules adopt differently ‘assembled’ or ‘open’ conformations, whichproduce different functions (Vogel et al., 2004). Multidomain pro-teins are thus versatile nanomachines and their intra- and inter-molecular interactions are promising targets for inhibitors ofprotein:protein interactions (Falchi et al., 2014; Zhang et al., 2014).Given the correlation between structure and function, determiningthe 3D structures of multidomain proteins and their ligand com-plexes is important for understanding their different biological roles,and for rationally designing protein-protein interaction inhibitors(PPIIs) againstparticular subsets of functions.However,multidomainproteins are often too flexible for X-ray crystallography, too large fornuclear magnetic resonance (NMR) and too small for electron mi-croscopy (EM). Obtaining 3D structural information on these mole-cules consequently requires hybrid approaches, where partial high-resolution information (from X-ray crystallography or NMR) iscombinedwith low-resolution information [for example fromsmall-angle X-ray scattering (SAXS), EM or atomic force microscopy),computational methods (molecular modelling, molecular dynamicssimulations) and distance constraints coming from experimentalmethods [such as F€orster resonance energy transfer (FRET), chemicalcross-linking, mutation or binding studies]. For recent reviews, see(Adams et al., 2013; Graewert and Svergun, 2013; Kovacs et al., 2015;Lasker et al., 2012;RamboandTainer, 2013;Rozycki andBoura, 2014).

Herein we summarise recent advances in our understanding ofthe structure-function relationship of the non-receptor tyrosinekinase FAK and its close homologue Pyk2. Following the editor'srecommendation, this review is based on an oral presentationgiven at the 2014 EMBO workshop, ‘Advances in proteineproteinanalysis and modulation’, and thereby concentrates on how ligandsactivate kinase-dependent functions of FAK at focal adhesions(FAs). This concept introduces a certain bias, which we have tried tominimise where possible. We also briefly discuss how this mech-anistic framework helps us to understand kinase-independent FAKfunctions in other environments and how it may support calcium-sensing by Pyk2.

2. Multiple cellular functions of FAK

FAK is essential in embryonic development and wound healing(see (Arold, 2011;Hall et al., 2011; Schaller, 2010)). In adult tissue, FAKisexpressedonlyat lowlevels.However, FAK isoverexpressed inmostcancers, and it endows cancer cells with functions that normal adulttissue cells do not have, such as the capacity to survive followingdetachment from the supporting structure. In so doing, FAK becomesa key player in cancer cell metastasis and tissue invasion (Fu et al.,2012; Sulzmaier et al., 2014; Zhang and Hochwald, 2014).

As indicated by its name, FAK is a central player for the assemblyand disassembly of FAs (Arold, 2011; Hall et al., 2011; Mitra et al.,2005; Schaller, 2010). FAs are large and dynamic macromolecularassemblies through which the actin cytoskeleton is connected to

the extracellular matrix (ECM). In addition to anchoring the cell,FAs encode the state of the ECM into intracellular biochemicalpathways that control cell morphology, migration, differentiation,proliferation, and survival (Gumbiner, 1996; Ridley et al., 2003;Wehrle-Haller and Imhof, 2002). FAK is recruited to FAs inresponse to integrin-mediated cell adhesion. However, FAK is alsoinvolved in signalling of other cell surface receptors (including Gprotein-coupled receptors, the Tcell receptor, the deleted-in-colon-cancer netrin receptor and transmembrane tyrosine kinases) and isacting in many cellular environments, such as lamellipodia, mi-crotubules and the nucleus (Chapman and Houtman, 2014;Schaller, 2010; Zhao and Guan, 2009). FAK can promote differenteffects in the same subcellular localisation (such as assembly anddisassembly of FAs); it can also produce convergent effects atdifferent subcellular localisations (such as cancer cell invasion andmetastasis in lamellipodia or the nucleus) (Arold, 2011; Cance andGolubovskaya, 2008; Hall et al., 2011; Schaller, 2010). This func-tional versatility raises the question of how FAK achieves site-specific and cell stateespecific functions.

FAK consists of a central kinase domain, flanked by two non-catalytic domains, the band 4.1, ezrin, radixin, moesin (FERM)domain and the focal adhesion targeting (FAT) domain. These do-mains are separated by long linkers of about 50 (FERM-kinase) and220 (kinase-FAT) residues (Fig. 1). More than 50 ligands have beenreported for FAK, and each domain and linker region has its own setof ligands. Some ligands bind to FAK in the nucleus, and others inthe cytoplasm (Fig. 1; red and black, respectively). A subset ofcytoplasmic ligands (green) is incompatible with or counteractsautophosphorylation of FAK tyrosine 397.

The major role of the FAK kinase domain appears to be theautophosphorylation of FAK Y397 (Ciccimaro et al., 2006). Y397 issituated in the FERM-kinase linker, adjacent to a proline-rich motif(PR1), and autophosphorylation of Y397 has to proceed in trans inthe full-length standard form of FAK (Brami-Cherrier et al., 2014;Toutant et al., 2002). Once phosphorylated, this region constitutesa bi-dentate binding site for the Src-homology 2 (SH2) and SH3domains of Src family kinases Fyn and Src (which bind to pY397 andPR1, respectively) (Arold et al., 2001; Schaller et al., 1994; Thomaset al., 1998) (Fig. 1). The interaction with FAK activates these Src-family kinases, which then provide most kinase activity associ-ated with FAK, including phosphorylation of tyrosines within FAK(including Y576/Y577 in the kinase activation loop, Y861 in thekinase-FAT linker, and Y925 in the FAT domain) (Schaller et al.,1999, 1994; Schlaepfer et al., 1994; Xie et al., 2008). Thus, auto-phosphorylation of Y397 triggers the kinase-dependent functionsof the Src:FAK complex. In the absence of Y397 autophosphor-ylation, FAK binds to different ligands and performs differentfunctions (Corsi et al., 2009). Hence FAK can be seen as a scaffoldingprotein with a Y397-phosphorylation switch between differentconformational and functional states (Fig. 1).

3. New insights into the structure and regulation ofindividual FAK domains

The atomic-resolution 3D structures of the individual domainsof FAK have been previously described [(Fig. 2); for a detailed

Fig. 1. FAK is a multidomain scaffolding protein with many cellular ligands. Proteins associating with FAK in the nucleus are shown in red. Cytoplasmic ligands incompatible with, orcounteracting phosphorylation of Y397 are in green and other ligands in black. For references please see (Alam et al., 2014; Arold, 2011; Bongiorno-Borbone et al., 2005; Chen et al.,2001; Fujita et al., 1998; Garces et al., 2006; Hall et al., 2011; Ho et al., 2012; Nagano et al., 2010; Oh et al., 2009; Plaza-Menacho et al., 2011; Tanaka et al., 2010; Wang et al., 2014).PR: proline-rich region; SH: Src homology; FERM: band 4.1, ezrin, radixin, merlin; FAT: focal adhesion targeting. The colour code for the FERM, kinase and FAT domain is used for allillustrations.

Fig. 2. The structure and functional motifs of individual FAK domains. Dashed lines indicate loops not included in the crystallographic structural model. Key regions and residuesdiscussed in the text are highlighted. (A) Crystal structure of FERM-kinase domain fragment (31-686) in the ‘closed’ autoinhibited conformation (PDB accession number 2J0J). (B)Isolated FAT domain (purple) predominantly forms a four-helix bundle structure (PDB ID 3S9O). The S910PPP motif binds to site 1/4 located between helices H1 and H4. (C) FAT bindstwo paxillin LD motifs (yellow helices) on two sites, site 1/4 (located between helices 1 and 4) and site 2/3 (located between H2 and H3); PDB ID 1OW7. (D) FAT domain showingopening of H1 (PDB ID 1K04).

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e7162

review, see (Alam et al., 2014; Arold, 2011; Hall et al., 2011)]. TheFAK FERM domain is composed of three lobes (F1eF3) and isstructurally similar to the FERM domains of those proteins afterwhich it was named (band 4.1, ezrin, radixin and moesin)(Ceccarelli et al., 2006; Girault et al., 1999). However, the FAK FERM

domain has adapted this scaffold to engage different interactions(such as binding to transmembrane receptor tails, or intra-molecular interactions), or to engage in the same type of in-teractions differently (for example binding to phospholipid headgroups) (Arold, 2011; Ceccarelli et al., 2006; Hall et al., 2011). The

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 63

crystal structure of a FAK fragment comprising the FERM and kinasedomains (residues 31-686) revealed that the FERM domain candock onto the kinase domain (Fig. 2) (Lietha et al., 2007). Thisassembled FERM-kinase conformation is stabilised by binding ofATP to the kinase domain, yet is incompatiblewith phosphorylationof Y576/Y577 in the kinase activation loop (Goni et al., 2014; Liethaet al., 2007; Zhou et al., 2015). The FAK kinase domain cannotefficiently autophosphorylate its activation loop tyrosines Y576/Y577, but requires Src kinases for this modification (Ciccimaro et al.,2006). Binding of the Src kinase SH2 domain to pY397 requiresdissociation of the contacts between pY397 and the FERM F1 lobe(Fig. 2A and data not shown), suggesting that this event weakensthe FERM:kinase interaction. Hence Y397 autophosphorylation,which leads to phosphorylation of Y576/Y577 by Src kinases, ulti-mately promotes an open, disassembled FAK conformation, asexperimentally observed using FRET sensors (Cai et al., 2008; Goniet al., 2014; Lietha et al., 2007). Thus, the presence of ATP (which isexpected to be constitutive, given the high ATP concentrations inthe cell) and phosphorylation of the kinase activation loop stabilisestructurally and functionally distinct structural states.

Despite forming a simple four-helix bundle structure, the ~140residue FAT domain also possesses several layers of regulation,linked to conformational plasticity (Fig. 2B) (Alam et al., 2014; Aroldet al., 2002; Dixon et al., 2004; Hayashi et al., 2002; Kadare et al.,2015; Prutzman et al., 2004). The interaction between FAT andpaxillin LD motifs results in FAK being recruited to FAs (Fig. 2C)(Brown et al., 1996). FAT binds two helical LD motifs, one betweenhelices H1 and H4 (site 1/4), the other one between helices H2 andH3 (site 2/3) (Bertolucci et al., 2005; Gao et al., 2004; Hoellereret al., 2003), whereas FAT-homology domains (FAHs) of other FA-localising proteins (CCM3, GIT1/2 and vinculin) bind only one LDmotif (Alam et al., 2014; Brown et al., 1996; Li et al., 2011;Schmalzigaug et al., 2007; Turner et al., 1990; Zhang et al., 2008).FAT binds also two CD4 endocytosis motifs, in a structurally similarway to LD motifs, allowing CD4 to recruit FAK for T cell receptorsignalling (Garron et al., 2008). Conversely, the FAT helix 4 is suf-ficient to bind to the talin FERM F3 lobe, and the FAT four-helixbundle structure is not required for this interaction (Hayashiet al., 2002; Lawson et al., 2012).

Phosphorylation and subsequent interaction of Y925 [located onhelix 1 (H1) of FAT] with the Grb2 SH2 domain requires opening ofH1 (Arold et al., 2002; Schlaepfer et al., 1994). H1-opening is pro-moted by the P944XPP motif between H1 and H2 (Fig. 2B and D),which functions as a molecular spring (Arold et al., 2002; Kadare

Fig. 3. Similarities in the molecular architecture of FAK and Pyk2 homodimeric structures. (crystallography (Brami-Cherrier et al., 2014). The dimeric FERM-kinase fragment was takenFAT domains, were placed using SAXS (ab inito SAXS model is shown as grey surface) (Bpositioning of FAT and linker residues (not shown). W266 (blue) is required for dimer forconformation (Brami-Cherrier et al., 2014). (B) The Pyk2 FERM dimer found in the crystal lawhere Pyk2 W273 (blue) at the dimer interface corresponds to FAK W266.

et al., 2015). Despite being only a low-probability event in vitro(Arold et al., 2002; Zhou et al., 2006), H1-opening is required incells for a subset of FAK functions, including phosphorylation ofY925 and FA turnover (Kadare et al., 2015). However, H1-opening isincompatible with FAT binding to paxillin LD motifs (Kadare et al.,2015). In crystal structures of apo-FAT, the S910PPPmotif, situated inan N-terminal extension of FAT, binds to site 1/4 (Arold et al., 2002;Kadare et al., 2015) (Fig. 2B). This interaction is incompatible withboth FAT H1 opening and binding of LD motifs to site 1/4 (Hoellereret al., 2003; Kadare et al., 2015). The S910PPP motif, when phos-phorylated by Erk 1/2 on serine 910, becomes a substrate for thepeptidyl-prolyl cis/trans isomerase PIN1. PIN1 recruits the proteintyrosine phosphatase PTP-PEST, which dephosphorylates FAK Y397,leading to increased FA turnover and tumour cell invasiveness(Zheng et al., 2009, 2011). The structural and mechanistic effects ofS910 phosphorylation and cis/trans isomerisation remain unknown.However, given the proximity of S910 to K1032 (located on FAT H4),we speculate that phosphorylation reinforces the interaction of thepS910PPP motif with the FAT domain core (Fig. 2B). Possible out-comes of S910 phosphorylationmight include, but not be limited to,competition with the LD motif bound to FAT helices 1/4, or stabi-lisation of the 4-helix bundle structure of FAT.

Recent structural insights now provide a framework to under-stand how different regulatory mechanisms combine to allow site-specific conformations and functions.

4. Structural insight into full-length FAK

Since full-length FAK does not form crystals suitable for X-rayanalysis, we have used SAXS to obtain structural insights (Brami-Cherrier et al., 2014). The ab initio low-resolution shape that weobtained suggested that FAK forms dimers under our experimentalconditions. Fitting of this shape by the known atomic-resolutionfragments was enabled by our observation that all crystal struc-tures that contained the FERM domain [including human (Brami-Cherrier et al., 2014) and avian FERM (Ceccarelli et al., 2006) andavian FERM-kinase fragments (Lietha et al., 2007)] contained thesame FERM:FERM dimer in the crystal lattice. These dimeric FERM-kinase structures fitted the full-length SAXS ab initio shape verywell and were used in conjunction with SAXS data on full-lengthFAK to investigate the position of the missing residues located C-terminal to the kinase domain. Although the low resolution pre-cluded a precise positioning of the FAT domain, SAXS data sup-ported that FATattached to the F2 lobe of the FERM domain (Brami-

A) The model of dimeric full length FAK was established by combining SAXS and X-rayfrom the crystal lattice of PDB 2J0J. The residues outside of this fragment, including therami-Cherrier et al., 2014). Note that the low resolution of SAXS precludes a precisemation in vitro and in cells. FAT:FERM interactions additionally stabilise the dimericttice of the Pyk2 FERM crystal structure (4EKU) corresponds to the FAK FERM dimers,

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e7164

Cherrier et al., 2014).

5. Functional implications of FAK dimerisation through theFERM domain

This full-length model suggested that FAK can be dimeric(Fig. 3A), and that this dimer conformation requires a FERM:FERMinteraction, centred on an interface formed by W266 (Brami-Cherrier et al., 2014). FAK dimerisation was not anticipated, sincenone of the FAK fragments (FERM, kinase, FERM-kinase or FAT)formed stable dimers in solution (with the exception of the arm-exchanged FAT dimers, which occur in vitro under high FAT con-centrations, and are unlikely to play a major role in cells) (Aroldet al., 2002; Kadare et al., 2015; Lietha et al., 2007). However,FAK self-association is apparent in vitro and in cells and requiresintact W266 (Brami-Cherrier et al., 2014). Moreover FAK transientdimerisation is needed for autophosphorylation of Y397 in vitroand in cells (Brami-Cherrier et al., 2014; Toutant et al., 2002). Inthis regard, full-length FAK behaves differently from the FERM-kinase fragment, which can autophosphorylate Y397 in cis(Lietha et al., 2007). Given the length of the FERM-kinase linker,and position of Y397 within this linker, transient dissociation ofthe FERM:kinase interaction is required for phosphorylation ofY397 in trans across the W266-mediated dimer. The apparentin vitro dimerisation dissociation constant (Kd) for the isolatedFERM domain is only ~30 mM, which explains why this interactionhas gone unnoticed in previous experiments and why it is onlyapparent in the highly concentrated FERM domains within crystalstructures (Brami-Cherrier et al., 2014). From our data we esti-mated the apparent dimerisation Kd of full-length FAK to be150 nM, whereas the FAK concentration in cells is only about10 nM (Brami-Cherrier et al., 2014). These findings suggested thattrans-autophosphorylation of FAK Y397, and hence activation ofkinase-dependent functions, requires local enrichment and clus-tering of FAK. Indeed, using acceptor photobleaching and confocalmicroscopy, we showed that FAK self-associates and autophos-phorylates specifically at FAs, suggesting that FAK reaches therequired local concentration specifically and exclusively at thesestructures. Moreover, our data also suggested that the weakFERM:FERM dimer needs to be stabilised by additional interactionsfrom the region that is C-terminal to the kinase domain (Brami-Cherrier et al., 2014).

6. FERM and FAT synergise to stabilise FAK dimerisation

The full-length structural data also suggested that the FERMdomain binds to the FAT domain. We confirmed this interactionexperimentally, establishing a Kd of 600 nM for recombinant FATand FERM domains. The observation that the presence of a 10-foldexcess of free FAT reduced substantially the amount of dimeric FAKsuggested that the FERM:FAT interaction stabilises FAK dimers byoccurring in trans (Brami-Cherrier et al., 2014), which is compatiblewith the length of the kinase-FAT linker. Counterintuitively, thepresence of paxillin LD4 peptides enhances the association be-tween FAT and FERM. Deletion of the N-terminal extension of FAT(residues 895 to 915) decreased the binding of FAT to FERM, but thepresence of LD4 peptides restored the interaction of the N-termi-nally deleted FAT (Brami-Cherrier et al., 2014). The exact region ofFAT that binds to FERM is unknown, as is the mechanism by whichLD4 increases this interaction.

7. The multifunctional basic motif of the FERM domain

The binding of FERM to FAT requires the K216AKTRLK basicmotif located in the FERM F2 lobe (Fig. 3A), but the exact site of

interaction remains to be determined (Brami-Cherrier et al.,2014). Biosensor studies in live cells and cell-free assays haveshown that the same basic motif is required for binding ofphosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] and that thisinteraction correlates with disassembly of the inhibitory FERM:-kinase interactions and triggers autophosphorylation of Tyr397(Cai et al., 2008; Papusheva et al., 2009). Although the exact modeof PI(4,5)P2 binding is unknown, all-atom molecular dynamicssimulations suggested an allosteric link between binding ofPI(4,5)P2 to FERM F2 and the FERM F1 lobe (Zhou et al., 2015).Work by Goni et al. shows that both the charged head group andthe fatty acid chain of PI(4,5)P2 are needed to induce clusteringand partial structural opening of the inhibitory FERM:kinase as-sociation, which then stimulates Y397 autophosphorylation (Goniet al., 2014). Based on computational simulations, fluorescencemeasurements and crystallographic analysis of the FERMK216AKTRLK->AAATALK mutant, the authors proposed a model inwhich PI(4,5)P2 binding renders the FERM F2 lobe, and hence theassociation of the F2 lobe with the C-terminal lobe of the kinasedomain, more flexible (Goni et al., 2014). Rather than increasingthe propensity of the kinase domain to dissociate from FERM totrans-autophoshorylate Y397 (as required within the W266-mediated FAK dimer), the authors suggested that this increasedflexibility leads to dissociation of the link between the FERM F1lobe and the N-terminal kinase lobe, while maintaining theconnection between the kinase and the F2 lobe. To explain how aFERM-bound kinase can nonetheless phosphorylate Y397 in trans,the authors proposed that within PI(4,5)P2-induced clusters, FAKmolecules arrange themselves to allow their kinase domains tophosphorylate Y397 of the FERM-kinase linker from their neigh-bouring FAK molecule. Because we found that W266-mediatedFERM:FERM dimers are required for Y397 autophosphorylationin cells and in cell-free assays in the absence of PI(4,5)P2 (Brami-Cherrier et al., 2014), the lateral Y397 trans-phosphorylationwithout FERM:kinase dissociation proposed by Go~ni et al. mightbe an alternative and possibly synergistic route in clustered FAKpopulations. In both models, the FERM:kinase interaction isweakened by Src binding to pY397, and disrupted by subsequentSrc phosphorylation of the FAK kinase activation loop tyrosinesY576/Y577.

Indeed, K216AKTRLK-dependent binding of FAK appears to be amore general mechanism of ligand-induced coactivation. Thismode of action has been reported for two receptor tyrosine kinasesthat can activate FAK, namely the phosphorylated hepatocytegrowth factor receptor (pc-Met) (Chen and Chen, 2006; Chen et al.,2011) and rearranged during transfection (RET) (Plaza-Menachoet al., 2011). It is possible that the doubly phosphorylated pc-Metpromotes FAK autophosphorylation using a charge-based mecha-nism reminiscent of the one used by doubly phosphorylated lipidhead groups (Goni et al., 2014). The mechanism used by RET ap-pears to be different, because the RET:FERM interaction (whichleads to phosphorylation of FAK Y576/577 by RET) is disrupted byFAK-mediated RET phosphorylation, rather than promoted throughphosphorylation as observed for c-Met and lipid head groups(Plaza-Menacho et al., 2011). How all these, and perhaps other, co-activators compete or synergise with each other and with theFERM-bound FAT domain is currently unclear and awaits furtherstructural and functional analysis.

Intriguingly, in addition to enabling activation of FAK kinase-dependent functions at the plasma membrane, this basicK216AKTRLK motif is also part of a nuclear localisation signal (NLS),and mutation of K216 and K218 into alanine blocked nuclearaccumulation (Lim et al., 2008). Consequently, the interaction ofFAT with the FERM K216AKTRLK motif is expected to have animportant role in localisation and activation of FAK.

Fig. 4. Schematic model illustrating site-specific functions of FAK. FERM, kinase and FAT colours are as in Fig. 1. Flexible linker regions are in black. Ligands are not drawn to scale.Phosphorylated residues are indicated by a P on a coloured circle, where red denotes Y397 and orange is used for other tyrosine residues. Serine phosphorylation events are notincluded. The FERM:FAT interaction in dimeric FAK is shown in trans (see Section 6), however mechanisms in which a FERM:FAT interaction in cis stabilises FAK dimers cannot beruled out. At FAs, functionally important clustering of FAK molecules occurs. These clusters were not drawn for reasons of clarity. Trans-autophosphorylation of Y397 at FAs is shownto imply transient FERM:kinase dissociation across the W266-meditated dimer. Please note that autophosphorylation following PIP2-induced clustering was suggested to proceedwithout FERM:kinase dissociation (Goni et al., 2014). Phosphorylation of Y925 and subsequent recruitment of Grb2 is incompatible with the interaction of FAT with paxillin. Inmature FAs, other mechanisms can still attach FAK to focal adhesions, such as an interaction with talin, which does not need an assembled 4-helix bundle (Chen et al., 1995; Hayashiet al., 2002; Lawson et al., 2012). However interaction with Grb2 correlates with mechanisms leading to FA dissociation (Ezratty et al., 2005). Please refer to Sections 6 and 9.1e9.3.for a detailed description.

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 65

8. pH dependent activation

The FERM domain also harbours another trigger for auto-activation. In a recent study, Ritt et al. presented direct evidencelinking FAK activation to changes in physiological pH in living cells(Ritt et al., 2013). Cancer cells are known to have a higher intra-cellular pH than normal cells, which is linked to cell migration andcell proliferation, all of which are processes affected by FAK acti-vation. Decreasing intracellular pH enhances the FERM-kinaseinteraction and hence results in decreases in both Y397 phos-phorylation and kinase activity as evident from FRET sensorebasedassays (Ritt et al., 2013). As a result, in normal cells at slightly acidicpH FAK is in an autoinhibited state whereas in alkaline cancerouscells, FAK activity and phosphorylation of Tyr397 increase leadingto cell proliferation and metastasis. The ability of FAK to respond tochanges in pH provides an additional mechanism of control of FAKactivity in normal versus cancerous cells. Ritt et al. proposed thatthe mechanism behind the pH sensing ability of FAK relies onprotonation of H58 on the FERM's F1 lobe (Fig. 2A) (Ritt et al., 2013).Protonated H58 likely forms a network of electrostatic interactionswith E466 on the kinase domain and therefore enhances theautoinhibited conformation of FAK. In an independent study, basedon evidence from molecular dynamic studies, Choi et al. proposedthat deprotonation of His58 at higher pH, although not throughdirect interaction, causes conformational changes in the linker re-gion that expose the Y397 phosphorylation site (Choi et al., 2013).

Conversely, a slightly acidic pH appears to increase phosphor-ylation of the FAT domain. Using a combination of biochemical and

NMR structural analysis, Cable et al. showed that phosphorylationof Y925 and Y1007 by Src is enhanced at pH 6e6.5, as comparedwith pH 7.5. This effect might be linked to pH affecting the localdynamics of the FAT structure. How an increase of Src-dependentFAT phosphorylation under acidic pH conditions combines withthe opposing effect of decreased Y397 autophosphorylation and Srcrecruitment to allow pH-dependent functions remains to bedetermined (Cable et al., 2012).

9. Site-specific functions of FAK

Collectively, these data provide mechanistic explanations forhow intramolecular and intermolecular interactions determine thesite-specific functions of FAK. In the following we briefly outlinehow this frameworkmay explain FAK's different actions in differentsubcellular environments. Sections 9.1e9.3 are illustrated in Fig. 4.

9.1. FAK in the cytoplasm

In the cytoplasm, without mechanisms for local enrichment, theFAK concentration is too low for stable FAK dimerisation/clustering,and FAK is expected to remain monomeric (Brami-Cherrier et al.,2014). The FERM:FAT interaction in cis, although probably lessstable than in trans, may conceal the NLS on FERM and thuscounteract nuclear accumulation (Brami-Cherrier et al., 2014; Limet al., 2008; Ossovskaya et al., 2008). Because monomeric full-length FAK cannot autophosphorylate Y397 in cis (Brami-Cherrieret al., 2014), activation of kinase-dependent functions is inhibited

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e7166

in the cytoplasm. Binding of the FAK familyeinteracting protein of200 kDa (FIP200) to the FERM and kinase domains of FAK furtherinhibits residual kinase activity (Abbi et al., 2002). The exactbinding sites and mechanism for the inhibitory action of FIP200 areunknown.

9.2. FAK at FAs

Clustering of integrins at sites of FAs induces clustering of paxillinin the intracellular membrane-proximal space. Clustered paxillinrecruits FAK through the interaction between the LD motifs and theFAT domain [other, paxillin-independent ways to recruit FAK alsoexist (Cooley et al., 2000)]. This recruitment induces FAK dimers intwo ways: firstly, it locally enriches FAK, bringing the local FAKconcentration closer to the dimerisation Kd. Secondly, paxillinbinding to FAT reinforces the FERM:FAT interaction that stabilises thedimers (Brami-Cherrier et al., 2014). FAK enrichment at FAs bypaxillin likely synergises with FAK enrichment through PI(4,5)P2,which is locally synthesised at FAs (Goni et al., 2014). It is unclearhow PI(4,5)P2 binding to the KAKTLRK motif affects the FERM:FATinteraction. Further stimulated by the presence of other co-factors(such as cell surface receptor tails or local pH), this environmentpromotes loosening and/or disruption of the inhibitory FERM:kinaseinteraction, leading to Y397 phosphorylation within FAK clusters intrans. pY397-FAK recruits and activates Src kinases, stimulatingphosphorylation of FAK and hence activation of kinase-dependentfunctions of the Src-FAK complex. Within this activated Src:FAKcomplex, FAK adopts a fully open conformation [(Cai et al., 2008;Goni et al., 2014) and our unpublished data], maximally func-tioning as a scaffold for assembly, phosphorylation and co-localisation of ligands. It is unclear what triggers the conforma-tional opening of FAT required for Y925 phosphorylation and Grb2binding. It is possible that FAT opening is a probabilistic event,gaining biological significance in a clustered Src-FAK population(Kadare et al., 2015). FAT opening and Grb2 binding disrupt theinteractionwith paxillin, whereas a fully deployed and disassembledFAKmay be a better target for proteolytic cleavage and phosphatases.These and other events may trigger FA disassembly and turnover(discussed in greater detail in (Arold, 2011; Hall et al., 2011)).

The available mechanistic data may also explain why FAKfunctions as a kinase-independent adaptor in early spreading ad-hesions. In these structures paxillin binds to another molecule,Nudel, and hence does not cluster and dimerise FAK (Shan et al.,2009). The absence of FAK clusters and autophosphorylation mayexplain why FAK functions as a kinase-independent adaptor inearly spreading adhesions, in conjunction with n-Wasp and Arp2/3to control assembly and dynamics of the actin cytoskeleton (Serrelset al., 2007).

9.3. FAK in the nucleus

Detachment of cells from the extracellular matrix can induceanoikis, mediated by the tumour suppressor p53. This detachment-triggered apoptosis is subverted in cancer cells, allowing metastatictumour cells to invade other tissues. It was proposed that FAK in-hibits p53 following cell detachment (Fu et al., 2012; Golubovskayaet al., 2008a; Lim et al., 2008). If focal adhesion contacts are lost,FAK translocates into the nucleus, where it sequesters and in-activates p53 and other pro-apoptotic effectors in a kinase-independent manner, thereby promoting cell survival. To achieveinactivation of p53, the FAK FERM domain functions as an adaptorbetween p53 and the ubiquitin E3 ligase murine double minute-2(Mdm2), to promote p53 ubiquitination and degradation(Golubovskaya et al., 2005; Lim et al., 2008). In the nucleus, FAKalso influences DNA methylation through direct interaction with

the methyl-CpG-binding domain protein 2 (MBD2) (Luo et al.,2009). MBD2 binds cytosine-methylated CpG islands, recruits his-tone deacetylases and DNA methyltransferases, and thus acts as atranscriptional repressor (Berger and Bird, 2005). The interactionwith FAK blocks MBD2 from recognising methylated CpG. It wasshown that this association enhances FAK's nuclear localisation andallows it to regulate heterochromatin remodelling and myogeninexpression during muscle differentiation (Luo et al., 2009). Thus,the controlled nuclear localisation allows FAK to logically linkmembrane-proximal stimuli with gene expression and hence toestablish a link between adhesion, motility, survival and epigeneticchanges (Cance and Golubovskaya, 2008; Fu et al., 2012).

Nuclear localization is promoted by an NLS that includes theK216AKTLRK basic cluster on the FERM domain (Lim et al., 2008)and that is also required for the interaction between FERM and FAT.Through yeast two-hybrid screening and cell lysate pull-down as-says, it was shown that MBD2 binds directly to the FAT domain ofboth FAK and Pyk2 (Luo et al., 2009). The MBD domain of MBD2 issufficient for this interaction, but additional residues from theglycineearginine rich (GR) region increase the affinity. Based on theavailable structural framework of FAK, we speculate that uponbinding to FAT, MBD2 disrupts the FAT:FERM interaction and re-veals the NLS, promoting nuclear import (Brami-Cherrier et al.,2014). In the absence of specific local enrichment, nuclear FAK isexpected to remain monomeric, a hypothesis compatible with itsfunction as a kinase-independent scaffold inhibiting pro-apoptoticfactors such as p53.

10. Pyk2 idiosyncrasies

In contrast to FAK, which is ubiquitous and highly expressedduring embryogenesis, Pyk2 appears at later developmental stagesand is abundant only in specific cell types, especially neurons(Menegon et al., 1999). Nonetheless, Pyk2 is of high physiologicaland biomedical relevance. In cell types where both FAK and Pyk2are present, Pyk2 can synergise with FAK, or compensate for a lossof FAK (resulting from knockout or specific inhibitors, for example),and thus promote cancer cell survival and metastasis (Sulzmaieret al., 2014; Weis et al., 2008). Indeed, Pyk2 expression isincreased in gliomas, hepatocellular carcinoma, lung, breast andprostate cancers, and has been proposed as a therapeutic target forinvasive cancers (Allen et al., 2009; Lipinski and Loftus, 2010; Loftuset al., 2009; Stanzione et al., 2001; Sun et al., 2008, 2011). Dereg-ulation of Pyk2 in cell types where it is abundant, such as neurons,osteoclasts and macrophages, has been linked to diseases associ-ated with these cell types, such as neurodegeneration, osteoporosisand inflammation (Gil-Henn et al., 2007; Lambert et al., 2013;Okigaki et al., 2003). Importantly, the Pyk2 gene (PTK2B) wasrecently discovered to be one of the genetic risk factors for Alz-heimer's disease (Lambert et al., 2013).

10.1. Conserved mechanism despite differences?

FAK and Pyk2 share the same domain organization (Giraultet al., 1999; Lev et al., 1995b) (Fig. 5). Given the high sequenceidentity of the FERM, kinase and FAT domains of these two proteins,it is expected that fundamental mechanistic aspects of proteinactivation are conserved. Indeed, Pyk2 can transiently self-associate through interactions involving both its N-terminal andC-terminal domains (Park et al., 2004). Self-association is requiredfor trans-autophosphorylation on Y402 (corresponding to FAKY397) (Bartos et al., 2010; Park et al., 2004). pY402 binds and ac-tivates Src family kinases, which then phosphorylate other residuesin Pyk2 and associated proteins (Dikic et al., 1996; Park et al., 2004).The Pyk2 FAT domain also binds to paxillin LD motifs (Lulo et al.,

Fig. 5. Pyk2 has adapted the domain organisation of FAK for calcium-sensing. (A) Sequence identities between Pyk2 and FAK are shown for individual regions. Reported Ca2þ/CaMbinding motifs of Pyk2 are shown in yellow (Kohno et al., 2008; Xie et al., 2008). (B) The structures of the isolated Pyk2 FERM (green; PDB 4EKU, lobes F1eF3 are indicated) andkinase (cyan; PDB 3FZO) domains are positioned as in the FAK FERM-kinase fragment (PDB 2J0J). Although this FERM:kinase association has not yet been observed experimentally,we speculate that Pyk2 also adopts such a FAK-like conformation, because the FAK FERM-kinase interface is conserved in Pyk2. The Pyk2 key residue F599 (blue), corresponding toFAK Y596, is highlighted. Reported Ca2þ/CaM binding motifs are mapped onto the FERM and kinase domains (orange). (C) Complex formed between the Pyk2 FAT domain (magenta)and paxillin LD4 motifs (PDB 3GM1) (Lulo et al., 2009).

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 67

2009). However, unlike FAK, Pyk2 is not strongly localized to focaladhesions in most cell types. Moreover, Pyk2 has the unique ca-pacity to sense calcium ions (Ca2þ). This capacity enables Pyk2 toact as a central transducer of the Ca2þ signal in FA turnover(Hashido et al., 2006) and allows it to execute specific functionsoutside of FAs (Andreev et al., 1999). For example, Pyk2 activation inneurons by Ca2þ promotes neurite outgrowth and is involved insynaptic plasticity (Girault et al., 1999).

Pyk2 has evolved through FAK gene duplication in vertebratesand subsequent specialisation (Corsi et al., 2006). How Pyk2 hasadapted the FAK scaffold to sense Ca2þ is unknown. Several path-ways may contribute to Ca2þ-dependent trans-autophosphor-ylation of Pyk2. Two groups reported that Pyk2 binds Ca2þ-boundcalmodulin (Ca2þ/CaM); however the proposed mechanisms differ(Kohno et al., 2008; Xie et al., 2008). Kohno et al. reported thatCa2þ/CaM binds to, and dimerises, the FERM domain, which isnecessary for Pyk2 trans-autophosphorylation (Kohno et al., 2008).Conversely, Xie et al. located a Ca2þ/CaM binding site within thePyk2 kinase domain (Xie et al., 2008). Both groups used sequencehomology to known CaM binding motifs to identify, and muta-tionally disrupt, their putative CaM binding site. However, in bothcases, the suggested binding site is mostly buried inside the FERMor kinase domain structure, inaccessible to canonical Ca2þ/CaMinteractions (Fig. 5B).

More indirect and often cell-type specific mechanisms alsoexist. In neurons Ca2þ/CaM has been reported to bind to the post-synaptic density protein 95 (PSD-95), releasing an autoinhibitoryinteraction and allowing PSD-95 to cluster and activate Pyk2(Bartos et al., 2010). The PSD-95 SH3 domain binds to a proline-richregion situated between the kinase and FAT domain of Pyk2

(Seabold et al., 2003). PSD-95 forms oligomers, and Pyk2 dimer-isation or clustering through Ca2þ/CaM:PSD-95 triggers Y402autophosphorylation in trans. Increased Ca2þ levels also triggerPyk2 autophosphorylation through activation of protein kinase C(PKC) (Lev et al., 1995a) or calmodulin-dependent kinase II (CaMKII)(Ginnan and Singer, 2002; Guo et al., 2004; Zwick et al., 1999). InPC12 cells, Ca2þ entry triggers both Pyk2 activation and cytonuclearaccumulation, and the Ca2þ-dependent Ser/Thr phosphatase cal-cineurin is required for these effects (Faure et al., 2007). The role ofprotein kinase A (PKA) and protein phosphatase 1(PP1) has alsobeen suggested (Battistone et al., 2014; Park et al., 2000; Rotfeld etal., 2014) {Park, 2000 #806; Battistone, 2014 #808; Rotfeld, 2014#807}. Pyk2 contains several putative (de)phosphorylation sites forthese kinases and phosphatases, mainly located in the kinase-FATlinker (Oppermann et al., 2009) {Oppermann, 2009 #606}. Thus,Pyk2 appears to be able to detect multiple signals triggered by Ca2þ.However, the molecular basis for how Pyk2 translates changes incytoplasmic calcium levels into biological responses remains to beestablished. Interestingly, the crystal structure of the isolated Pyk2FERM domain (PDB accession nr. 4EKU) shows identical FERM:-FERM dimers as observed and as demonstrated to be biologicallyrelevant to FAK (Brami-Cherrier et al., 2014) (Fig. 3B). We speculatethat Pyk2 has adapted the ligand-induced dimer-activation mech-anism of FAK so that Ca2þ-dependent mechanisms stabilise a weakFERM:FERM interaction for productive trans-autophosphorylation.

10.2. Nuclear Pyk2

Like FAK, Pyk2 is able to translocate to the nucleus (Faure et al.,2007). Pyk2 accumulates in the nucleus following Ca2þ influx

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e7168

(Faure et al., 2007), suggesting a correlation between Ca2þ/CaM-binding and nuclear localisation. This translocation involvesdephosphorylation of S778 by calcineurin, a Ca2þ/CaM activatedphosphatase, which appears to inactivate a nuclear exportsequence (Faure et al., 2013).

Like FAK, Pyk2 associates with MBD2 (Luo et al., 2009), and thisassociation correlates with a nuclear localisation of both proteins.The physiological effects of the Pyk2:MBD2 association, especiallyin neurons and other cell types where Pyk2 plays a central role,remain unknown, although it has been suggested that thePyk2:MBD2 complex might regulate gene expression (Mei andXiong, 2010). Formation of such a complex may reduce MBD2binding to methylated DNA and indirectly enhance gene expres-sion. The atomic basis for the interaction between Pyk2 and MBD2is unknown. Available data strongly suggest that MBD2 associatesin the same way with the FAT domains of either FAK or Pyk2 (Luoet al., 2009). Indeed, the FAK and Pyk2 FAT domains share 57%sequence identity, form the same 3D structure and bind paxillin LDmotifs in the same way (Fig. 5C) (Alam et al., 2014; Hoellerer et al.,2003; Lulo et al., 2009; Vanarotti et al., 2014).

11. Targeted inhibition of FAK and Pyk2

Owing to their central role in linking adhesion, migration andsurvival, FAK and Pyk2 constitute potential therapeutic targets.Targeted small-molecule inhibitors for FAK produce encouragingresults in mousemodels, where they prevented growth, metastasis,vascular permeability and angiogenesis of tumours (reviewed in(Sulzmaier et al., 2014)). Of particular interest, these inhibitors alsosensitise cancer cells to chemotherapy, thus providing dual bene-fits. Currentlymost FAK inhibitors, and all those in clinical phase I/IItrials, are ATP-competitive kinase inhibitors (Sulzmaier et al., 2014).Some compounds achieve enhanced specificity by stabilising aparticular helical conformation of the FAK kinase activation loop‘DFG’ motif (Lietha and Eck, 2008; Roberts et al., 2008). Howeverdeveloping highly specific ATP-competitive agents for FAK remainschallenging, and PPIIs would provide an attractive novel tool forcancer therapy (Morelli et al., 2011). Several PPIIs were reported forFAK that showed anti-tumour activity in xenograft mouse models,and sensitised cancer cells to chemotherapy (Golubovskaya et al.,2012, 2013, 2008b; Hochwald et al., 2009; Kurenova et al., 2014,2009). The proposed targets of those compounds are Y397, thereported p53 binding site on FERM, and the reported VEGFR-3binding site on the FAT domain. These compounds, which wereidentified through in silico docking studies, are currently only weakinhibitors; they require high micromolar dosages and their selec-tivity and precise mode of action remain to be completelyestablished.

Currently, the only Pyk2 inhibitors in preclinical trials are ATP-competitive kinase inhibitors (Sulzmaier et al., 2014). Forfunction-specific anti-Pyk2 PPIIs, especially in diseases other thancancer, a better evaluation of the biological role of Pyk2 is requiredto assess the potential benefits or risks of inhibition. For example itis not known whether it is an increase or a decrease in Pyk2function that contributes to an increased risk of Alzheimer's disease(Kaufman et al., 2015; Xu et al., 2014).

12. Conclusions & remaining challenges

Despite recent advances in our mechanistic understanding ofFAK and Pyk2, several aspects of their complex structure-functionrelationship remain elusive. For example, it is unclear how allthose ligands that require the KAKTLRK basic motif on FERM syn-ergise or compete for this motif. In fact, the possibility that muta-tions in the motif increase conformational dynamics of the F2 lobe,

and thus allosterically block binding of ligands to different orremote sites on F2, cannot be excluded. Substantial F2 destabili-sation is expected to disrupt the F2-mediated autoinhibition be-tween the FERM and kinase domain. Hence, controlled F2destabilisation might also explain how FAK activation results fromc-Metepromoted phosphorylation of Y195 (Chen et al., 2011), orPIAS-1emediated sumoylation of K152 (Kadare et al., 2003),because both Y195 and K152 are buried inside the F2 fold.

It also remains to be determined how the 220 residueelongkinase-FAT linker blocks phosphorylation in cis of Y397, althoughthis residue is located between the FERM and kinase domains. Thislinker region presents the lowest sequence similarity between FAKand Pyk2 (Fig. 5), and it remains to be determined how thesedivergent sequences support the functional idiosyncrasies of bothhomologues.

Additional structural data would be required to better under-stand FAK's function in the nucleus, especially how FAK links p53andMdm2, and how FAK and Pyk2 interact with MBD2, resulting inchanges in DNAmethylation patterns. Further mechanistic insights,on an atomic level, would be needed to understand the role of thesemolecules in mechanotransduction (Bae et al., 2014; Berk et al.,1995). Finally, understanding the function of FAK and Pyk2 alsorequires investigating the importance of noise, synergy and struc-tural/functional heterogeneity within clustered FAK or Pyk2 pop-ulations embedded in adhesomes or other complex structures(Ladbury and Arold, 2012). Understanding such complex systemson a molecular and holistic level might now be possible throughhybrid methods combining experimental and computationalmethods with molecular imaging in cells or in cell-free systems(Kanchanawong et al., 2010; Wilhelm et al., 2014; Zaidel-Bar et al.,2007). A detailed knowledge of the complex structure-functionrelationship of FAK and Pyk2 would enable precise targeting ofproteineprotein interactions by inhibitors, and might provide aninspiration for novel biotechnological applications and tools.

Acknowledgements

We thank A. M. Jama and A. A. Momin for critical reading andcomments. This study was supported by the King Abdullah Uni-versity of Science and Technology (KAUST).Work in J.A. Girault's labis supported by Inserm, ANR, ERC, and FRM.

References

Abbi, S., Ueda, H., Zheng, C., Cooper, L.A., Zhao, J., Christopher, R., Guan, J.L., 2002.Regulation of focal adhesion kinase by a novel protein inhibitor FIP200. Mol.Biol. Cell 13, 3178e3191.

Adams, P.D., Baker, D., Brunger, A.T., Das, R., DiMaio, F., Read, R.J., Richardson, D.C.,Richardson, J.S., Terwilliger, T.C., 2013. Advances, interactions, and future de-velopments in the CNS, Phenix, and Rosetta structural biology software sys-tems. Annu. Rev. Biophys. 42, 265e287.

Alam, T., Alazmi, M., Gao, X., Arold, S.T., 2014. How to find a leucine in a haystack?structure, ligand recognition and regulation of leucine-aspartic acid (LD) motifs.Biochem. J. 460, 317e329.

Allen, J.G., Lee, M.R., Han, C.Y., Scherrer, J., Flynn, S., Boucher, C., Zhao, H.,O'Connor, A.B., Roveto, P., Bauer, D., Graceffa, R., Richards, W.G., Babij, P., 2009.Identification of small molecule inhibitors of proline-rich tyrosine kinase 2(Pyk2) with osteogenic activity in osteoblast cells. Bioorg Med. Chem. Lett. 19,4924e4928.

Andreev, J., Simon, J.P., Sabatini, D.D., Kam, J., Plowman, G., Randazzo, P.A.,Schlessinger, J., 1999. Identification of a new Pyk2 target protein with Arf-GAPactivity. Mol. Cell Biol. 19, 2338e2350.

Arold, S.T., 2011. How focal adhesion kinase achieves regulation by linking ligandbinding, localization and action. Curr. Opin. Struct. Biol. 21, 808e813.

Arold, S.T., Hoellerer, M.K., Noble, M.E., 2002. The structural basis of localization andsignaling by the focal adhesion targeting domain. Structure 10, 319e327.

Arold, S.T., Ulmer, T.S., Mulhern, T.D., Werner, J.M., Ladbury, J.E., Campbell, I.D.,Noble, M.E., 2001. The role of the Src homology 3-Src homology 2 interface inthe regulation of Src kinases. J. Biol. Chem. 276, 17199e17205.

Bae, Y.H., Mui, K.L., Hsu, B.Y., Liu, S.L., Cretu, A., Razinia, Z., Xu, T., Pure, E.,Assoian, R.K., 2014. A FAK-Cas-Rac-lamellipodin signaling module transduces

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 69

extracellular matrix stiffness into mechanosensitive cell cycling. Sci. Signal. 7,ra57.

Bartos, J.A., Ulrich, J.D., Li, H., Beazely, M.A., Chen, Y., Macdonald, J.F., Hell, J.W., 2010.Postsynaptic clustering and activation of Pyk2 by PSD-95. J. Neurosci. 30,449e463.

Battistone, M.A., Alvau, A., Salicioni, A.M., Visconti, P.E., Da Ros, V.G., Cuasnicu, P.S.,2014. Evidence for the involvement of proline-rich tyrosine kinase 2 in tyrosinephosphorylation downstream of protein kinase A activation during humansperm capacitation. Mol. Hum. Reprod. 20, 1054e1066.

Berger, J., Bird, A., 2005. Role of MBD2 in gene regulation and tumorigenesis. Bio-chem. Soc. Trans. 33, 1537e1540.

Berk, B.C., Corson, M.A., Peterson, T.E., Tseng, H., 1995. Protein kinases as mediatorsof fluid shear stress stimulated signal transduction in endothelial cells: a hy-pothesis for calcium-dependent and calcium-independent events activated byflow. J. Biomech. 28, 1439e1450.

Bertolucci, C.M., Guibao, C.D., Zheng, J., 2005. Structural features of the focaladhesion kinase-paxillin complex give insight into the dynamics of focaladhesion assembly. Protein Sci. 14, 644e652.

Bongiorno-Borbone, L., Kadare, G., Benfenati, F., Girault, J.A., 2005. FAK and PYK2interact with SAP90/PSD-95-associated protein-3. Biochem. Biophys. Res.Commun. 337, 641e646.

Brami-Cherrier, K., Gervasi, N., Arsenieva, D., Walkiewicz, K., Boutterin, M.C.,Ortega, A., Leonard, P.G., Seantier, B., Gasmi, L., Bouceba, T., Kadare, G.,Girault, J.A., Arold, S.T., 2014. FAK dimerization controls its kinase-dependentfunctions at focal adhesions. EMBO J. 33, 356e370.

Brown, M.C., Perrotta, J.A., Turner, C.E., 1996. Identification of LIM3 as the principaldeterminant of paxillin focal adhesion localization and characterization of anovel motif on paxillin directing vinculin and focal adhesion kinase binding.J. Cell Biol. 135, 1109e1123.

Cable, J., Prutzman, K., Gunawardena, H.P., Schaller, M.D., Chen, X., Campbell, S.L.,2012. In vitro phosphorylation of the focal adhesion targeting domain of focaladhesion kinase by Src kinase. Biochemistry 51, 2213e2223.

Cai, X., Lietha, D., Ceccarelli, D.F., Karginov, A.V., Rajfur, Z., Jacobson, K., Hahn, K.M.,Eck, M.J., Schaller, M.D., 2008. Spatial and temporal regulation of focal adhesionkinase activity in living cells. Mol. Cell. Biol. 28, 201e214.

Cance, W.G., Golubovskaya, V.M., 2008. Focal adhesion kinase versus p53: apoptosisor survival? Sci. Signal. 1, pe22.

Ceccarelli, D.F., Song, H.K., Poy, F., Schaller, M.D., Eck, M.J., 2006. Crystal structure ofthe FERM domain of focal adhesion kinase. J. Biol. Chem. 281, 252e259.

Chapman, N.M., Houtman, J.C., 2014. Functions of the FAK family kinases in T cells:beyond actin cytoskeletal rearrangement. Immunol. Res. 59, 23e34.

Chen, H.C., Appeddu, P.A., Parsons, J.T., Hildebrand, J.D., Schaller, M.D., Guan, J.L.,1995. Interaction of focal adhesion kinase with cytoskeletal protein talin. J. Biol.Chem. 270, 16995e16999.

Chen, R., Kim, O., Li, M., Xiong, X., Guan, J.L., Kung, H.J., Chen, H., Shimizu, Y., Qiu, Y.,2001. Regulation of the PH-domain-containing tyrosine kinase Etk by focaladhesion kinase through the FERM domain. Nat. Cell Biol. 3, 439e444.

Chen, S.Y., Chen, H.C., 2006. Direct interaction of focal adhesion kinase (FAK) withMet is required for FAK to promote hepatocyte growth factor-induced cell in-vasion. Mol. Cell. Biol. 26, 5155e5167.

Chen, T.H., Chan, P.C., Chen, C.L., Chen, H.C., 2011. Phosphorylation of focal adhesionkinase on tyrosine 194 by Met leads to its activation through relief of auto-inhibition. Oncogene 30, 153e166.

Choi, C.H., Webb, B.A., Chimenti, M.S., Jacobson, M.P., Barber, D.L., 2013. pH sensingby FAK-His58 regulates focal adhesion remodeling. J. Cell Biol. 202, 849e859.

Ciccimaro, E., Hevko, J., Blair, I.A., 2006. Analysis of phosphorylation sites on focaladhesion kinase using nanospray liquid chromatography/multiple reactionmonitoring mass spectrometry. Rapid Commun. Mass Spectrom. 20,3681e3692.

Cooley, M.A., Broome, J.M., Ohngemach, C., Romer, L.H., Schaller, M.D., 2000. Paxillinbinding is not the sole determinant of focal adhesion localization or dominant-negative activity of focal adhesion kinase/focal adhesion kinase-related non-kinase. Mol. Biol. Cell 11, 3247e3263.

Corsi, J.M., Houbron, C., Billuart, P., Brunet, I., Bouvree, K., Eichmann, A., Girault, J.A.,Enslen, H., 2009. Autophosphorylation-independent and -dependent functionsof focal adhesion kinase during development. J. Biol. Chem. 284, 34769e34776.

Corsi, J.M., Rouer, E., Girault, J.A., Enslen, H., 2006. Organization and post-transcriptional processing of focal adhesion kinase gene. BMC Genomics 7, 198.

Dikic, I., Tokiwa, G., Lev, S., Courtneidge, S.A., Schlessinger, J., 1996. A role for Pyk2and Src in linking G-protein-coupled receptors with MAP kinase activation.Nature 383, 547e550.

Dixon, R.D., Chen, Y., Ding, F., Khare, S.D., Prutzman, K.C., Schaller, M.D.,Campbell, S.L., Dokholyan, N.V., 2004. New insights into FAK signaling andlocalization based on detection of a FAT domain folding intermediate. Structure12, 2161e2171.

Ezratty, E.J., Partridge, M.A., Gundersen, G.G., 2005. Microtubule-induced focaladhesion disassembly is mediated by dynamin and focal adhesion kinase. Nat.Cell Biol. 7, 581e590.

Falchi, F., Caporuscio, F., Recanatini, M., 2014. Structure-based design of small-molecule protein-protein interaction modulators: the story so far. FutureMed. Chem. 6, 343e357.

Faure, C., Corvol, J.C., Toutant, M., Valjent, E., Hvalby, O., Jensen, V., El Messari, S.,Corsi, J.M., Kadare, G., Girault, J.A., 2007. Calcineurin is essential fordepolarization-induced nuclear translocation and tyrosine phosphorylation ofPYK2 in neurons. J. Cell Sci. 120, 3034e3044.

Faure, C., Ramos, M., Girault, J.A., 2013. Pyk2 cytonuclear localization: mechanismsand regulation by serine dephosphorylation. Cell. Mol. Life Sci. 70, 137e152.

Fu, W., Hall, J.E., Schaller, M.D., 2012. Focal adhesion kinase-regulated signalingevents in human cancer. Biomol. Concepts 3, 225e240.

Fujita, H., Kamiguchi, K., Cho, D., Shibanuma, M., Morimoto, C., Tachibana, K., 1998.Interaction of Hic-5, A senescence-related protein, with focal adhesion kinase.J. Biol. Chem. 273, 26516e26521.

Gao, G., Prutzman, K.C., King, M.L., Scheswohl, D.M., DeRose, E.F., London, R.E.,Schaller, M.D., Campbell, S.L., 2004. NMR solution structure of the focal adhe-sion targeting domain of focal adhesion kinase in complex with a paxillin LDpeptide: evidence for a two-site binding model. J. Biol. Chem. 279, 8441e8451.

Garces, C.A., Kurenova, E.V., Golubovskaya, V.M., Cance, W.G., 2006. Vascularendothelial growth factor receptor-3 and focal adhesion kinase bind and sup-press apoptosis in breast cancer cells. Cancer Res. 66, 1446e1454.

Garron, M.L., Arthos, J., Guichou, J.F., McNally, J., Cicala, C., Arold, S.T., 2008. Struc-tural basis for the interaction between focal adhesion kinase and CD4. J. Mol.Biol. 375, 1320e1328.

Gil-Henn, H., Destaing, O., Sims, N.A., Aoki, K., Alles, N., Neff, L., Sanjay, A.,Bruzzaniti, A., De Camilli, P., Baron, R., Schlessinger, J., 2007. Defectivemicrotubule-dependent podosome organization in osteoclasts leads toincreased bone density in Pyk2(-/-) mice. J. Cell Biol. 178, 1053e1064.

Ginnan, R., Singer, H.A., 2002. CaM kinase II-dependent activation of tyrosine ki-nases and ERK1/2 in vascular smooth muscle. Am. J. Physiol. Cell Physiol. 282,C754eC761.

Girault, J.A., Labesse, G., Mornon, J.P., Callebaut, I., 1999. The N-termini of FAK andJAKs contain divergent band 4.1 domains. Trends Biochem. Sci. 24, 54e57.

Golubovskaya, V.M., Figel, S., Ho, B.T., Johnson, C.P., Yemma, M., Huang, G.,Zheng, M., Nyberg, C., Magis, A., Ostrov, D.A., Gelman, I.H., Cance, W.G., 2012.A small molecule focal adhesion kinase (FAK) inhibitor, targeting Y397 site: 1-(2-hydroxyethyl)-3, 5, 7-triaza-1-azoniatricyclo [3.3.1.1(3,7)]decane; bromideeffectively inhibits FAK autophosphorylation activity and decreases cancer cellviability, clonogenicity and tumor growth in vivo. Carcinogenesis 33,1004e1013.

Golubovskaya, V.M., Finch, R., Cance, W.G., 2005. Direct interaction of the N-ter-minal domain of focal adhesion kinase with the N-terminal transactivationdomain of p53. J. Biol. Chem. 280, 25008e25021.

Golubovskaya, V.M., Finch, R., Zheng, M., Kurenova, E.V., Cance, W.G., 2008a. The 7-amino-acid site in the proline-rich region of the N-terminal domain of p53 isinvolved in the interaction with FAK and is critical for p53 functioning. Bio-chem. J. 411, 151e160.

Golubovskaya, V.M., Ho, B., Zheng, M., Magis, A., Ostrov, D., Morrison, C.,Cance, W.G., 2013. Disruption of focal adhesion kinase and p53 interaction withsmall molecule compound R2 reactivated p53 and blocked tumor growth. BMCCancer 13, 342.

Golubovskaya, V.M., Nyberg, C., Zheng, M., Kweh, F., Magis, A., Ostrov, D.,Cance, W.G., 2008b. A small molecule inhibitor, 1,2,4,5-benzenetetraaminetetrahydrochloride, targeting the y397 site of focal adhesion kinase decreasestumor growth. J. Med. Chem. 51, 7405e7416.

Goni, G.M., Epifano, C., Boskovic, J., Camacho-Artacho, M., Zhou, J., Bronowska, A.,Martin, M.T., Eck, M.J., Kremer, L., Grater, F., Gervasio, F.L., Perez-Moreno, M.,Lietha, D., 2014. Phosphatidylinositol 4,5-bisphosphate triggers activation offocal adhesion kinase by inducing clustering and conformational changes. Proc.Natl. Acad. Sci. U. S. A. 111, E3177eE3186.

Graewert, M.A., Svergun, D.I., 2013. Impact and progress in small and wide angle X-ray scattering (SAXS and WAXS). Curr. Opin. Struct. Biol. 23, 748e754.

Gumbiner, B.M., 1996. Cell adhesion: the molecular basis of tissue architecture andmorphogenesis. Cell 84, 345e357.

Guo, J., Meng, F., Fu, X., Song, B., Yan, X., Zhang, G., 2004. N-methyl-D-aspartatereceptor and L-type voltage-gated Ca2þ channel activation mediate proline-richtyrosine kinase 2 phosphorylation during cerebral ischemia in rats. Neurosci.Lett. 355, 177e180.

Hall, J.E., Fu, W., Schaller, M.D., 2011. Focal adhesion kinase: exploring Fak structureto gain insight into function. Int. Rev. Cell Mol. Biol. 288, 185e225.

Hashido, M., Hayashi, K., Hirose, K., Iino, M., 2006. Ca2þ lightning conveys cell-cellcontact information inside the cells. EMBO Rep. 7, 1117e1123.

Hayashi, I., Vuori, K., Liddington, R.C., 2002. The focal adhesion targeting (FAT) re-gion of focal adhesion kinase is a four-helix bundle that binds paxillin. Nat.Struct. Biol. 9, 101e106.

Ho, B., Olson, G., Figel, S., Gelman, I., Cance, W.G., Golubovskaya, V.M., 2012. Nanogincreases focal adhesion kinase (FAK) promoter activity and expression anddirectly binds to FAK protein to be phosphorylated. J. Biol. Chem. 287,18656e18673.

Hochwald, S.N., Nyberg, C., Zheng, M., Zheng, D., Wood, C., Massoll, N.A., Magis, A.,Ostrov, D., Cance, W.G., Golubovskaya, V.M., 2009. A novel small molecule in-hibitor of FAK decreases growth of human pancreatic cancer. Cell Cycle 8,2435e2443.

Hoellerer, M.K., Noble, M.E., Labesse, G., Campbell, I.D., Werner, J.M., Arold, S.T.,2003. Molecular recognition of paxillin LD motifs by the focal adhesion tar-geting domain. Structure 11, 1207e1217.

Kadare, G., Gervasi, N., Brami-Cherrier, K., Blockus, H., El Messari, S., Arold, S.T.,Girault, J.A., 2015. Conformational dynamics of the focal adhesion targetingdomain control specific functions of focal adhesion kinase in cells. J. Biol. Chem.290, 478e491.

Kadare, G., Toutant, M., Formstecher, E., Corvol, J.C., Carnaud, M., Boutterin, M.C.,Girault, J.A., 2003. PIAS1-mediated sumoylation of focal adhesion kinase

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e7170

activates its autophosphorylation. J. Biol. Chem. 278, 47434e47440.Kanchanawong, P., Shtengel, G., Pasapera, A.M., Ramko, E.B., Davidson, M.W.,

Hess, H.F., Waterman, C.M., 2010. Nanoscale architecture of integrin-based celladhesions. Nature 468, 580e584.

Kaufman, A.C., Salazar, S.V., Haas, L.T., Yang, J., Kostylev, M.A., Jeng, A.T.,Robinson, S.A., Gunther, E.C., van Dyck, C.H., Nygaard, H.B., Strittmatter, S.M.,2015. Fyn inhibition rescues established memory and synapse loss in Alzheimermice. Ann. Neurol. 77, 953e971.

Kohno, T., Matsuda, E., Sasaki, H., Sasaki, T., 2008. Protein-tyrosine kinase CAKbeta/PYK2 is activated by binding Ca2þ/calmodulin to FERM F2 alpha2 helix and thusforming its dimer. Biochem. J. 410, 513e523.

Kovacs, E., Zorn, J.A., Huang, Y., Barros, T., Kuriyan, J., 2015. A structural perspectiveon the regulation of the epidermal growth factor receptor. Annu. Rev. Biochem.84, 739e764. http://dx.doi.org/10.1146/annurev-biochem-060614-034402.

Kurenova, E., Ucar, D., Liao, J., Yemma, M., Gogate, P., Bshara, W., Sunar, U.,Seshadri, M., Hochwald, S.N., Cance, W.G., 2014. A FAK scaffold inhibitor dis-rupts FAK and VEGFR-3 signaling and blocks melanoma growth by targetingboth tumor and endothelial cells. Cell. Cycle 13, 2542e2553.

Kurenova, E.V., Hunt, D.L., He, D., Magis, A.T., Ostrov, D.A., Cance, W.G., 2009. Smallmolecule chloropyramine hydrochloride (C4) targets the binding site of focaladhesion kinase and vascular endothelial growth factor receptor 3 and sup-presses breast cancer growth in vivo. J. Med. Chem. 52, 4716e4724.

Ladbury, J.E., Arold, S.T., 2012. Noise in cellular signaling pathways: causes andeffects. Trends Biochem. Sci. 37, 173e178.

Lambert, J.C., Ibrahim-Verbaas, C.A., Harold, D., Naj, A.C., Sims, R., Bellenguez, C.,DeStafano, A.L., Bis, J.C., Beecham, G.W., Grenier-Boley, B., Russo, G., Thorton-Wells, T.A., Jones, N., Smith, A.V., Chouraki, V., Thomas, C., Ikram, M.A.,Zelenika, D., Vardarajan, B.N., Kamatani, Y., Lin, C.F., Gerrish, A., Schmidt, H.,Kunkle, B., Dunstan, M.L., Ruiz, A., Bihoreau, M.T., Choi, S.H., Reitz, C.,Pasquier, F., Cruchaga, C., Craig, D., Amin, N., Berr, C., Lopez, O.L., De Jager, P.L.,Deramecourt, V., Johnston, J.A., Evans, D., Lovestone, S., Letenneur, L., Moron, F.J.,Rubinsztein, D.C., Eiriksdottir, G., Sleegers, K., Goate, A.M., Fievet, N.,Huentelman, M.W., Gill, M., Brown, K., Kamboh, M.I., Keller, L., Barberger-Gateau, P., McGuiness, B., Larson, E.B., Green, R., Myers, A.J., Dufouil, C., Todd, S.,Wallon, D., Love, S., Rogaeva, E., Gallacher, J., St George-Hyslop, P., Clarimon, J.,Lleo, A., Bayer, A., Tsuang, D.W., Yu, L., Tsolaki, M., Bossu, P., Spalletta, G.,Proitsi, P., Collinge, J., Sorbi, S., Sanchez-Garcia, F., Fox, N.C., Hardy, J., DenizNaranjo, M.C., Bosco, P., Clarke, R., Brayne, C., Galimberti, D., Mancuso, M.,Matthews, F., European Alzheimer's Disease, I., Genetic, Environmental Risk inAlzheimer's, D., Alzheimer's Disease Genetic, C., Cohorts for, H., Aging Researchin Genomic, E., Moebus, S., Mecocci, P., Del Zompo, M., Maier, W., Hampel, H.,Pilotto, A., Bullido, M., Panza, F., Caffarra, P., et al., 2013. Meta-analysis of 74,046individuals identifies 11 new susceptibility loci for Alzheimer's disease. Nat.Genet. 45, 1452e1458.

Lasker, K., Velazquez-Muriel, J.A., Webb, B.M., Yang, Z., Ferrin, T.E., Sali, A., 2012.Macromolecular assembly structures by comparative modeling and electronmicroscopy. Methods Mol. Biol. 857, 331e350.

Lawson, C., Lim, S.T., Uryu, S., Chen, X.L., Calderwood, D.A., Schlaepfer, D.D., 2012.FAK promotes recruitment of talin to nascent adhesions to control cell motility.J. Cell Biol. 196, 223e232.

Lev, S., Moreno, H., Martinez, R., Canoll, P., Peles, E., Musacchio, J.M., Plowman, G.D.,Rudy, B., Schlessinger, J., 1995a. Protein tyrosine kinase PYK2 involved in Ca2þ-induced regulation of ion channel and MAP kinase functions. Nature 376,737e745.

Lev, S., Moreno, H., Martinez, R., Canoll, P., Peles, E., Musacchio, J.M., Plowman, G.D.,Rudy, B., Schlessinger, J., 1995b. Protein tyrosine kinase PYK2 involved inCa(2þ)-induced regulation of ion channel and MAP kinase functions. Nature376, 737e745.

Li, X., Ji, W., Zhang, R., Folta-Stogniew, E., Min, W., Boggon, T.J., 2011. Molecularrecognition of leucine-aspartate repeat (LD) motifs by the focal adhesion tar-geting homology domain of cerebral cavernous malformation 3 (CCM3). J. Biol.Chem. 286, 26138e26147.

Lietha, D., Cai, X., Ceccarelli, D.F., Li, Y., Schaller, M.D., Eck, M.J., 2007. Structural basisfor the autoinhibition of focal adhesion kinase. Cell 129, 1177e1187.

Lietha, D., Eck, M.J., 2008. Crystal structures of the FAK kinase in complex withTAE226 and related bis-anilino pyrimidine inhibitors reveal a helical DFGconformation. PLoS One 3, e3800.

Lim, S.T., Chen, X.L., Lim, Y., Hanson, D.A., Vo, T.T., Howerton, K., Larocque, N.,Fisher, S.J., Schlaepfer, D.D., Ilic, D., 2008. Nuclear FAK promotes cell prolifera-tion and survival through FERM-enhanced p53 degradation. Mol. Cell 29, 9e22.

Lipinski, C.A., Loftus, J.C., 2010. Targeting Pyk2 for therapeutic intervention. ExpertOpin. Ther. Targets 14, 95e108.

Loftus, J.C., Yang, Z., Tran, N.L., Kloss, J., Viso, C., Berens, M.E., Lipinski, C.A., 2009. ThePyk2 FERM domain as a target to inhibit glioma migration. Mol. Cancer Ther. 8,1505e1514.

Lulo, J., Yuzawa, S., Schlessinger, J., 2009. Crystal structures of free and ligand-boundfocal adhesion targeting domain of Pyk2. Biochem. Biophys. Res. Commun. 383,347e352.

Luo, S.W., Zhang, C., Zhang, B., Kim, C.H., Qiu, Y.Z., Du, Q.S., Mei, L., Xiong, W.C., 2009.Regulation of heterochromatin remodelling and myogenin expression duringmuscle differentiation by FAK interaction with MBD2. EMBO J. 28, 2568e2582.

Mei, L., Xiong, W.C., 2010. FAK interaction with MBD2: a link from cell adhesion tonuclear chromatin remodeling? Cell Adh. Migr. 4, 77e80.

Menegon, A., Burgaya, F., Baudot, P., Dunlap, D.D., Girault, J.A., Valtorta, F., 1999.FAKþ and PYK2/CAKbeta, two related tyrosine kinases highly expressed in the

central nervous system: similarities and differences in the expression pattern.Eur. J. Neurosci. 11, 3777e3788.

Mitra, S.K., Hanson, D.A., Schlaepfer, D.D., 2005. Focal adhesion kinase: in commandand control of cell motility. Nat. Rev. Mol. Cell Biol. 6, 56e68.

Morelli, X., Bourgeas, R., Roche, P., 2011. Chemical and structural lessons from recentsuccesses in protein-protein interaction inhibition (2P2I). Curr. Opin. Chem.Biol. 15, 475e481.

Nagano, M., Hoshino, D., Sakamoto, T., Kawasaki, N., Koshikawa, N., Seiki, M., 2010.ZF21 protein regulates cell adhesion and motility. J. Biol. Chem. 285,21013e21022.

Oh, M.A., Choi, S., Lee, M.J., Choi, M.C., Lee, S.A., Ko, W., Cance, W.G., Oh, E.S.,Buday, L., Kim, S.H., Lee, J.W., 2009. Specific tyrosine phosphorylation of focaladhesion kinase mediated by Fer tyrosine kinase in suspended hepatocytes.Biochim. Biophys. Acta 1793, 781e791.

Okigaki, M., Davis, C., Falasca, M., Harroch, S., Felsenfeld, D.P., Sheetz, M.P.,Schlessinger, J., 2003. Pyk2 regulates multiple signaling events crucial formacrophage morphology and migration. Proc. Natl. Acad. Sci. U. S. A. 100,10740e10745.

Oppermann, F.S., Gnad, F., Olsen, J.V., Hornberger, R., Greff, Z., Keri, G., Mann, M.,Daub, H., 2009. Large-scale proteomics analysis of the human kinome. Mol. Cell.Proteomics. 8, 1751e1764.

Ossovskaya, V., Lim, S.T., Ota, N., Schlaepfer, D.D., Ilic, D., 2008. FAK nuclear exportsignal sequences. FEBS Lett. 582, 2402e2406.

Papusheva, E., Mello de Queiroz, F., Dalous, J., Han, Y., Esposito, A., Jares-Erijmanxa, E.A., Jovin, T.M., Bunt, G., 2009. Dynamic conformational changes inthe FERM domain of FAK are involved in focal-adhesion behavior during cellspreading and motility. J. Cell Sci. 122, 656e666.

Park, J.H., Park, J.K., Bae, K.W., Park, H.T., 2000. Protein kinase A activity is requiredfor depolarization-induced proline-rich tyrosine kinase 2 and mitogen-acti-vated protein kinase activation in PC12 cells. Neurosci. Lett. 290, 25e28.

Park, S.Y., Avraham, H.K., Avraham, S., 2004. RAFTK/Pyk2 activation is mediated bytrans-acting autophosphorylation in a Src-independent manner. J. Biol. Chem.279, 33315e33322.

Plaza-Menacho, I., Morandi, A., Mologni, L., Boender, P., Gambacorti-Passerini, C.,Magee, A.I., Hofstra, R.M., Knowles, P., McDonald, N.Q., Isacke, C.M., 2011. Focaladhesion kinase (FAK) binds RET kinase via its FERM domain, priming a directand reciprocal RET-FAK transactivation mechanism. J. Biol. Chem. 286,17292e17302.

Prutzman, K.C., Gao, G., King, M.L., Iyer, V.V., Mueller, G.A., Schaller, M.D.,Campbell, S.L., 2004. The focal adhesion targeting domain of focal adhesionkinase contains a hinge region that modulates tyrosine 926 phosphorylation.Structure 12, 881e891.

Rambo, R.P., Tainer, J.A., 2013. Super-resolution in solution X-ray scattering and itsapplications to structural systems biology. Annu. Rev. Biophys. 42, 415e441.

Ridley, A.J., Schwartz, M.A., Burridge, K., Firtel, R.A., Ginsberg, M.H., Borisy, G.,Parsons, J.T., Horwitz, A.R., 2003. Cell migration: integrating signals from frontto back. Science 302, 1704e1709.

Ritt, M., Guan, J.L., Sivaramakrishnan, S., 2013. Visualizing and manipulating focaladhesion kinase regulation in live cells. J. Biol. Chem. 288, 8875e8886.

Roberts, W.G., Ung, E., Whalen, P., Cooper, B., Hulford, C., Autry, C., Richter, D.,Emerson, E., Lin, J., Kath, J., Coleman, K., Yao, L., Martinez-Alsina, L., Lorenzen, M.,Berliner, M., Luzzio, M., Patel, N., Schmitt, E., LaGreca, S., Jani, J., Wessel, M.,Marr, E., Griffor, M., Vajdos, F., 2008. Antitumor activity and pharmacology of aselective focal adhesion kinase inhibitor, PF-562,271. Cancer Res. 68,1935e1944.

Rotfeld, H., Hillman, P., Ickowicz, D., Breitbart, H., 2014. PKA and CaMKII mediatePI3K activation in bovine sperm by inhibition of the PKC/PP1 cascade. Repro-duction. 147, 347e356.

Rozycki, B., Boura, E., 2014. Large, dynamic, multi-protein complexes: a challengefor structural biology. J. Phys. Condens. Matter 26, 463103.

Schaller, M.D., 2010. Cellular functions of FAK kinases: insight into molecularmechanisms and novel functions. J. Cell Sci. 123, 1007e1013.

Schaller, M.D., Hildebrand, J.D., Parsons, J.T., 1999. Complex formation with focaladhesion kinase: a mechanism to regulate activity and subcellular localizationof Src kinases. Mol. Biol. Cell 10, 3489e3505.

Schaller, M.D., Hildebrand, J.D., Shannon, J.D., Fox, J.W., Vines, R.R., Parsons, J.T.,1994. Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent binding of pp60src. Mol. Cell. Biol. 14, 1680e1688.

Schlaepfer, D.D., Hanks, S.K., Hunter, T., van der Geer, P., 1994. Integrin-mediatedsignal transduction linked to Ras pathway by GRB2 binding to focal adhesionkinase. Nature 372, 786e791.

Schmalzigaug, R., Garron, M.L., Roseman, J.T., Xing, Y., Davidson, C.E., Arold, S.T.,Premont, R.T., 2007. GIT1 utilizes a focal adhesion targeting-homology domainto bind paxillin. Cell Signal 19, 1733e1744.

Seabold, G.K., Burette, A., Lim, I.A., Weinberg, R.J., Hell, J.W., 2003. Interaction of thetyrosine kinase Pyk2 with the N-methyl-D-aspartate receptor complex via theSrc homology 3 domains of PSD-95 and SAP102. J. Biol. Chem. 278,15040e15048.

Serrels, B., Serrels, A., Brunton, V.G., Holt, M., McLean, G.W., Gray, C.H., Jones, G.E.,Frame, M.C., 2007. Focal adhesion kinase controls actin assembly via a FERM-mediated interaction with the Arp2/3 complex. Nat. Cell Biol. 9, 1046e1056.

Shan, Y., Yu, L., Li, Y., Pan, Y., Zhang, Q., Wang, F., Chen, J., Zhu, X., 2009. Nudel andFAK as antagonizing strength modulators of nascent adhesions through paxillin.PLoS Biol. 7, e1000116.

Stanzione, R., Picascia, A., Chieffi, P., Imbimbo, C., Palmieri, A., Mirone, V.,Staibano, S., Franco, R., De Rosa, G., Schlessinger, J., Tramontano, D., 2001.

K.W. Walkiewicz et al. / Progress in Biophysics and Molecular Biology 119 (2015) 60e71 71

Variations of proline-rich kinase Pyk2 expression correlate with prostate cancerprogression. Lab. Invest. 81, 51e59.

Sulzmaier, F.J., Jean, C., Schlaepfer, D.D., 2014. FAK in cancer: mechanistic findingsand clinical applications. Nat. Rev. Cancer 14, 598e610.

Sun, C.K., Man, K., Ng, K.T., Ho, J.W., Lim, Z.X., Cheng, Q., Lo, C.M., Poon, R.T., Fan, S.T.,2008. Proline-rich tyrosine kinase 2 (Pyk2) promotes proliferation and inva-siveness of hepatocellular carcinoma cells through c-Src/ERK activation. Carci-nogenesis 29, 2096e2105.

Sun, C.K., Ng, K.T., Lim, Z.X., Cheng, Q., Lo, C.M., Poon, R.T., Man, K., Wong, N.,Fan, S.T., 2011. Proline-rich tyrosine kinase 2 (Pyk2) promotes cell motility ofhepatocellular carcinoma through induction of epithelial to mesenchymaltransition. PLoS One 6, e18878.

Tanaka, T., Moriwaki, K., Murata, S., Miyasaka, M., 2010. LIM domain-containingadaptor, leupaxin, localizes in focal adhesion and suppresses the integrin-induced tyrosine phosphorylation of paxillin. Cancer Sci. 101, 363e368.

Thomas, J.W., Ellis, B., Boerner, R.J., Knight, W.B., White II, G.C., Schaller, M.D., 1998.SH2- and SH3-mediated interactions between focal adhesion kinase and Src.J. Biol. Chem. 273, 577e583.

Toutant, M., Costa, A., Studler, J.M., Kadare, G., Carnaud, M., Girault, J.A., 2002.Alternative splicing controls the mechanisms of FAK autophosphorylation. Mol.Cell. Biol. 22, 7731e7743.

Turner, C.E., Glenney Jr., J.R., Burridge, K., 1990. Paxillin: a new vinculin-bindingprotein present in focal adhesions. J. Cell Biol. 111, 1059e1068.

Vanarotti, M.S., Miller, D.J., Guibao, C.D., Nourse, A., Zheng, J.J., 2014. Structural andmechanistic insights into the interaction between Pyk2 and paxillin LD motifs.J. Mol. Biol. 426, 3985e4001.

Vogel, C., Bashton, M., Kerrison, N.D., Chothia, C., Teichmann, S.A., 2004. Structure,function and evolution of multidomain proteins. Curr. Opin. Struct. Biol. 14,208e216.

Wang, Q., Wang, Y., Fritz, D., Rajshankar, D., Downey, G.P., McCulloch, C.A., 2014.Interactions of the protein-tyrosine phosphatase-alpha with the focal adhesiontargeting domain of focal adhesion kinase are involved in interleukin-1signaling in fibroblasts. J. Biol. Chem. 289, 18427e18441.

Wehrle-Haller, B., Imhof, B., 2002. The inner lives of focal adhesions. Trends CellBiol. 12, 382e389.

Weis, S.M., Lim, S.T., Lutu-Fuga, K.M., Barnes, L.A., Chen, X.L., Gothert, J.R., Shen, T.L.,Guan, J.L., Schlaepfer, D.D., Cheresh, D.A., 2008. Compensatory role for Pyk2during angiogenesis in adult mice lacking endothelial cell FAK. J. Cell Biol. 181,43e50.

Wilhelm, B.G., Mandad, S., Truckenbrodt, S., Krohnert, K., Schafer, C., Rammner, B.,Koo, S.J., Classen, G.A., Krauss, M., Haucke, V., Urlaub, H., Rizzoli, S.O., 2014.

Composition of isolated synaptic boutons reveals the amounts of vesicle traf-ficking proteins. Science 344, 1023e1028.

Xie, J., Allen, K.H., Marguet, A., Berghorn, K.A., Bliss, S.P., Navratil, A.M., Guan, J.L.,Roberson, M.S., 2008. Analysis of the calcium-dependent regulation of proline-rich tyrosine kinase 2 by gonadotropin-releasing hormone. Mol. Endocrinol. 22,2322e2335.

Xu, J., Chatterjee, M., Baguley, T.D., Brouillette, J., Kurup, P., Ghosh, D., Kanyo, J.,Zhang, Y., Seyb, K., Ononenyi, C., Foscue, E., Anderson, G.M., Gresack, J.,Cuny, G.D., Glicksman, M.A., Greengard, P., Lam, T.T., Tautz, L., Nairn, A.C.,Ellman, J.A., Lombroso, P.J., 2014. Inhibitor of the tyrosine phosphatase STEPreverses cognitive deficits in a mouse model of Alzheimer's disease. PLoS Biol.12, e1001923.

Zaidel-Bar, R., Itzkovitz, S., Ma'ayan, A., Iyengar, R., Geiger, B., 2007. Functional atlasof the integrin adhesome. Nat. Cell Biol. 9, 858e867.

Zhang, J., Hochwald, S.N., 2014. The role of FAK in tumor metabolism and therapy.Pharmacol. Ther. 142, 154e163.

Zhang, X., Betzi, S., Morelli, X., Roche, P., 2014. Focused chemical librariesedesignand enrichment: an example of protein-protein interaction chemical space.Future Med. Chem. 6, 1291e1307.

Zhang, Z.M., Simmerman, J.A., Guibao, C.D., Zheng, J.J., 2008. GIT1 paxillin-bindingdomain is a four-helix bundle, and it binds to both paxillin LD2 and LD4 mo-tifs. J. Biol. Chem. 283, 18685e18693.

Zhao, J., Guan, J.L., 2009. Signal transduction by focal adhesion kinase in cancer.Cancer Metastasis Rev. 28, 35e49.

Zheng, Y., Xia, Y., Hawke, D., Halle, M., Tremblay, M.L., Gao, X., Zhou, X.Z., Aldape, K.,Cobb, M.H., Xie, K., He, J., Lu, Z., 2009. FAK phosphorylation by ERK primes ras-induced tyrosine dephosphorylation of FAK mediated by PIN1 and PTP-PEST.Mol. Cell 35, 11e25.

Zheng, Y., Yang, W., Xia, Y., Hawke, D., Liu, D.X., Lu, Z., 2011. Ras-induced andextracellular signal-regulated kinase 1 and 2 phosphorylation-dependentisomerization of protein tyrosine phosphatase (PTP)-PEST by PIN1 promotesFAK dephosphorylation by PTP-PEST. Mol. Cell. Biol. 31, 4258e4269.

Zhou, J., Bronowska, A., Le Coq, J., Lietha, D., Grater, F., 2015. Allosteric regulation offocal adhesion kinase by PIP(2) and ATP. Biophys. J. 108, 698e705.

Zhou, Z., Feng, H., Bai, Y., 2006. Detection of a hidden folding intermediate in thefocal adhesion target domain: implications for its function and folding. Proteins65, 259e265.

Zwick, E., Wallasch, C., Daub, H., Ullrich, A., 1999. Distinct calcium-dependent pathways of epidermal growth factor receptor trans-activation and PYK2 tyrosine phosphorylation in PC12 cells. J. Biol. Chem.274, 20989e20996.