Project topics and materials
Get Complete Project Material File(s) Now! »
Tyrosine kinase signaling and the FAK family
In animals, phosphotyrosine (pTyr) signaling is an essential system that regulates hormone, growth factor, immune, and adhesion-based signaling, therefore allowing cell-cell communication (Hunter and Cooper, 1981; Shattil and Brugge, 1991; Myers et al., 1994; Weiss and Littman, 1994).
This signaling relies on a simple mechanism: a TK phosphorylates certain tyrosine residues on the substrate protein (or itself). Then, an effector protein recognize the pTyr by its Src homology 2 (SH2) domain or other pTyr-binding domains, and activates downstream signaling. The pathway can be stopped by a protein tyrosine phosphatase (PTP) which dephosphorylates the pTyr (Figure 1).
Figure 1. Mechanism of TK signaling. Tyrosines are phosphorylated by TKs and dephosphorylated by PTPs. The phosphorylated tyrosine is then recognized by an SH2-containing effector.
The origin of TKs and FAK family
The eukaryotic protein kinase (ePK) superfamily is divided into two kinase families according to their substrate specificity: the TKs and the serine/threonine kinases (STKs) families. TKs originally evolved from STKs. Although some STKs can phosphorylate tyrosine residues and are referred to as dual specificity protein kinases (Lindberg et al., 1992), TKs are discriminated from STKs by their overall sequence similarity and also notably by their characteristic catalytic loop motif (Hanks and Hunter, 1995; Manning et al., 2008).
The evolutionary history of TK family has been specifically studied and discussed for many years (Miller, 2012; Tong et al., 2017). It was initially believed that tyrosine kinases were specific to metazoans. In contrast with the high diversity of TKs in many metazoan phyla, they were initially not found in plants, fungi and other analyzed eukaryotes (Hanks and Hunter, 1995; Manning et al., 2002a). Although some bacteria evolved unique tyrosine kinases known as bacterial tyrosine kinases (BY-kinases), metazoan-like TKs were not found in bacteria (Grangeasse et al., 2012). Accordingly, TK were hypothesized to be a metazoan-specific evolutionary innovation that permitted cell-cell communication and thus contributed to the origin of metazoan multicellularity.
However, in 2001, King and Carroll discovered the first TK (a receptor tyrosine kinase designated MBRTK1) outside of the metazoan taxon (King and Carroll, 2001). This TK was found in Monosiga brevicollis, a unicellular member of choanoflagellates that are the closest known living relatives of metazoans (Richter and King, 2013). Successive papers then confirmed the existence of multiple active tyrosine kinases in this organism (King et al., 2003) and in several other choanoflagellate species (Segawa et al., 2006; Suga et al., 2008).
Genomic analyses then revealed that choanoflagellates contain a rich and complex repertoire of TKs that is comparable to those observed in the most complex multicellular animals (King et al., 2008; Manning et al., 2008; Pincus et al., 2008). Many choanoflagellates can form colonies, suggesting that their common ancestor with the metazoan could be a transitional form between unicellular and multicellular organisms and that these unicellular tyrosine kinases may have facilitated the evolution of multicellular animals (King, 2004).
Afterward, new genome/transcriptome sequencing has uncovered TKs in other eukaryotes thus forcing to reconsider every time the apparition date of TKs during evolution. In this way, in addition to the metazoans and the choanoflagellates, tyrosine kinases were discovered in three other sister lineages: the filastereans, the ichthyosporeans and the corallochytreans (forming, with the metazoans and the choanoflagellates, the holozoans) (Shalchian-Tabrizi et al., 2008; Suga et al., 2012, 2014; Fairclough et al., 2013; Sebé-Pedrós et al., 2016). TKs were also found in two amoebozoans (Clarke et al., 2013; Schaap et al., 2015) and in the apusozoan Thecamonas trahens (Suga et al., 2012). Moreover, TKs were discovered in some bikonts such as the green alga Chlamydomonas reinhardtii or the higher plants Arabidopsis thaliana and Oryza sativa (Shiu and Li, 2004; Miranda-Saavedra and Barton, 2007; Kerk et al., 2008; Wheeler et al., 2008) but also the oomycete Phytophthora infestans (Shiu and Li, 2004; Judelson and Ah-Fong, 2010). Nevertheless and so far, TKs were not found in the non-holozoan opistokonts (the fungi (Shiu and Li, 2004; Miranda-Saavedra and Barton, 2007; Suga et al., 2012, 2014), and the fungi relatives cristidiscoideans (Suga et al., 2014)), and in many other species (Shiu and Li, 2004; Miranda-Saavedra and Barton, 2007; Suga et al., 2008; Liu et al., 2011; Clarke et al., 2013) (Figure 2).
Figure 2. Distribution of tyrosine kinases and FAK family within eukaryotic evolution. The branches in the phylogenetic tree are not proportional to the divergence time. Presence or absence of tyrosine kinases or FAK orthologs in a clade is marked by a green tick or a red cross respectively. One can thus hypothesize that the appearance of TKs dates back up to the beginning of eukaryotic evolution and that some species then lost these enzymes (Shiu and Li, 2004; Miranda-Saavedra and Barton, 2007; Kerk et al., 2008; Wheeler et al., 2008; Schaap et al., 2015). However, one cannot exclude absolutely the possibility of multiple independent appearance of TKs or eventually some horizontal gene transfers. In any case, the prime hypothesis of a phosphotyrosine signaling at the origin of metazoan multicellularity is invalidated by these recent findings. Nevertheless, it is worth noting that the number of TKs underwent a great expansion in the holozoans (Manning et al., 2008; Suga et al., 2012, 2014; Fairclough et al., 2013; Sebé-Pedrós et al., 2016) suggesting a more prominent role of phosphotyrosine signaling in this monophyletic group. This expansion is the result of several gene duplications and domain shuffling (Shiu and Li, 2004; Suga et al., 2008, 2012, 2014; Liu et al., 2011; Jin and Pawson, 2012; Liu and Nash, 2012) and, as reviewed in Tong et al., it was certainly allowed by the fact that holozoan pTyr signaling had little cross-interference with pre-existing signaling systems such as pSer/Thr signaling and had therefore liberty to evolve novel functions (Tong et al., 2017).
Within holozoan lineage, a central set of TK families rapidly arose and remained conserved throughout the evolution. Among them, orthologs of the focal adhesion kinase (FAK) family have been discovered in filastereans and choanoflagellates but not in more evolutionary-distant clades (Sebe-Pedros et al., 2010; Fairclough et al., 2013; Suga et al., 2014) (Figure 2). This suggests an urholozoan origin of the FAK family.
However, the appearance of paralogs within FAK family seems to have occurred early in the vertebrate lineage. FAK family is composed of FAK and the proline rich kinase 2 (Pyk2) sharing together around 45% amino acid identity. Sequence alignments from different species suggests that gene duplication leading to the appearance of FAK and Pyk2 occurred after the urochordate branch and is consequently specific to vertebrates (Corsi et al., 2006) (Figure 3). Besides, the aforementioned alignments showed that FAK genes are more closely related to the common unique ancestor than Pyk2 genes. This gives rise to the idea that Pyk2 probably underwent less evolutionary pressures and had thus more possibilities to evolve.
Figure 3. Evolution of FAK family within eumetazoan clade. The branches in the phylogenetic tree are not proportional to the divergence time. The dendrogram summarizes the results obtained in Corsi et al., 2006.
FAK family
Identification of FAK family kinases
FAK was first identified in the beginning of the 90•s as a 120-kDa protein that is phosphorylated on tyrosine in cells attached to fibronectin-coated surfaces (Guan et al., 1991). It was then cloned independently by two laboratories from chicken embryo cells infected with v-Src (Schaller et al., 1992) and by sequence homology in mouse (Hanks et al., 1992). As its name implies, FAK is located to focal adhesions which are interaction sites mediated by integrins, between extracellular matrix and actin cytoskeleton through the plasma membrane (Carragher and Frame, 2004). FAK was then cloned in multiples vertebrate and invertebrate species including human (Whitney et al., 1993), xenope (Hens and DeSimone, 1995; Zhang et al., 1995), zebrafisch (Henry et al., 2001; Crawford et al., 2003), sea urchin (GarcÖ٨a et al., 2004) and drosophila (Fox et al., 1999; Fujimoto et al., 1999; Palmer et al., 1999).
Pyk2 (Lev et al., 1995), also known as cell adhesion kinase-Ƣ (CAKƢ) (Sasaki et al., 1995), calcium dependent-protein tyrosine kinase (CADTK) (Yu et al., 1996) or related adhesion focal tyrosine kinase (RAFTK) (Avraham et al., 2000) was initially described in PC12 cells as an nRTK activated by cytosolic calcium increase and stimulation of protein kinase C (PKC) (Lev et al., 1995). These two signaling pathways can be, in some case, independently activated by the same stimulus but can have an additive effect on Pyk2 activation (Brinson et al., 1998). As mentioned above, Pyk2 is a vertebrates-specific protein.
Structure of FAK family kinases
FAK and Pyk2 are composed of three conserved domains (Figure 4): an amino-terminal four-point-one, ezrin, radixin, moesin (FERM) domain, a central kinase domain and carboxy-terminal focal adhesion targeting (FAT) domain (Girault et al., 1999b; Lipinski and Loftus, 2010; Hall et al., 2011; Walkiewicz et al., 2015). These three domains are connected by two linker regions which contain proline-rich (PR) motifs. Unlike many nRTKs, neither FAK nor Pyk2 encompass Src-homology domains 2 and 3 (SH2 and SH3). FERM and FAT domains both contribute to the regulation of the enzymatic activity of FAK and Pyk2 and allow their interaction with many proteins playing a key role in signal transduction.
Figure 4. Comparison of human FAK and Pyk2 structures. FERM, kinase, FAT and PR domains, as well the main phosphorylated tyrosines are indicated. Percentages of identity (aminoacids) of each domain are specified.
FERM domain
FERM domains are roughly 300 amino-acids domains commonly found in proteins that bind cytoplasmic regions of transmembrane proteins and often act as linker between the cytoskeleton and plasma membrane (Chishti et al., 1998; Girault et al., 1999b; Riggs et al., 2011). Besides, FERM domains can also mediate intramolecular interactions. For instance, the functional activity of the prototypical FERM domain proteins ezrin, radixin, and moesin is regulated by FERM domain mediated intramolecular associations (Pearson et al., 2000; Edwards and Keep, 2001).
The FERM domain has three lobes, namely: F1, F2 and F3, together forming a cloverleaf-shaped structure that mediates both protein-membrane targeting as well as protein-protein interactions (Figure 5). It is worth noting FAK FERM domain shares only 12²15% identity with the sequences of other FERM domains but they adopt a quite similar tertiary structure as shown by hydrophobic cluster analysis (Girault et al., 1999b).
FERM proteins are targeted to the membrane due to the interaction between basic residues in a cleft between subdomains F1 and F3 and PIP2 (Hirao et al., 1996; Hamada et al., 2000). This interaction further induces conformational changes of FERM proteins that would stimulate their interaction with the cytoplasmic tails of transmembrane proteins (Hamada et al., 2003).
The FERM domain also appears be a domain of interaction with various cytosolic and nuclear proteins. Among them, the three membrane-associated phosphatidylinositol transfer proteins (PITPNMs), a family of protein associated with metastasized cancers, interact with the FERM domain of Pyk2 but not FAK (Lev et al., 1999). In contrast, the transcription factor p53 was shown to interact with the FERM domain of both FAK (Golubovskaya and Cance, 2011) and Pyk2 (Lim et al., 2010). The mitogen-activated protein kinase kinase kinase kinase 4 (MAP4K4) was shown to interact with the FERM domain of Pyk2 but not of FAK (Loftus et al., 2013).
Another function of the FERM domain is a potential autoregulatory role due to interactions with the kinase domain. This autoinhibitory interaction has been described in FAK (Lietha et al., 2007), and also, more recently, in Pyk2 (Loving and Underbakke, 2019). Moreover, in both FAK and Pyk2, the FERM domain is supposed to be mandatory for the activation-induced homodimerization (Kohno et al., 2008; Riggs et al., 2011; Brami-Cherrier et al., 2014).
Finally, a nuclear localization sequence (NLS) and a nuclear export signal (NES) were found in the F2 and F1 subdomain respectively of both FAK and Pyk2 showing a particular standing of this region for FAK family kinases nucleocytoplasmic shuttling (Lim et al., 2008a; Ossovskaya et al., 2008).
Figure 5. Cartoon structure of FAK (left) and Pyk2 (right) FERM domain. FAK FERM structure was taken from Ceccarelli et al., 2006. Pyk2 FERM structure was taken from PDB ID: 4EKU Savarimuthu et al.
Linker 1
The FERM domain of FAK family kinases is followed by a linker of approximately 70 aminoacids which precedes the kinase domain. This sequence is poorly conserved with the exception of PR1 and the Tyr-397/Tyr-402 (FAK /Pyk2) whose phosphorylation plays a key role in the activation of FAK and Pyk2 and in the recruitment of Src-family kinases (SFKs) (Dikic et al., 1996; Hall et al., 2011). Indeed, this recruitment is based on a two-point interaction between: 1. the SH2 domain of the SFK and the phospho-tyrosine (pTyr-397 of FAK/pTyr-402 of Pyk2) 2. The SH3 domain of the SFK and the PR1 of FAK/Pyk2 (Schaller et al., 1994; Thomas et al., 1998; Arold et al., 2001).
In FAK specifically, this linker also contains the Tyr-407, which phosphorylation was shown to negatively regulate FAK activity (Lim et al., 2007).
Kinase domain
The central region is the kinase domain which is highly conserved between FAK and Pyk2 (Sasaki et al., 1995). This domain adopts a typical bi-lobal structure, very similar to that of other kinase domains, surrounding the catalytic site and the ATP binding pocket (Han et al., 2009) (Figure 6). The N-lobe contains a five-stranded beta-sheet and one alpha-helix whereas the C-lobe is mainly composed of alpha-helixes and the activation loop. Phosphorylation of Tyr-579 and Tyr-580 (Tyr-576 and Tyr-577 for FAK) within this activation loop is crucial to maximize Pyk2 kinase activity.
This kinase domain was also shown to interact directly with the retinoblastoma 1-inducible coiled-coil 1 (RB1CC1) which is supposed to be an inhibitor of Pyk2 kinase activity (Ueda et al., 2000). Moreover, a NES was found in the C-lobe of both FAK and Pyk2 suggesting an unexpected role of the kinase domain in nucleocytoplasmic shuttling of FAK family kinases (Ossovskaya et al., 2008).
Linker 2
The linker 2 is located between the kinase domain and the FAT domain. This sequence contains two PR sequences, named PR2 (residues 713 to 720 in Pyk2) and PR3 (residues 855 to 860, ibid.), that mediate the interaction of Pyk2 with a number of SH3 domain-containing proteins that also interact with FAK such as p130Cas (Astier et al., 1997; Xiong et al., 1998; Lakkakorpi et al., 1999), SAPAP3, and Graf (Ohba et al., 1998; Xiong et al., 1998). More specifically in Pyk2, these PR sequences were also shown to interact with PSD-95 (Seabold et al., 2003), synapse-associated protein 102 (SAP102) (Seabold et al., 2003), ArfGAP with SH3 domain, ankyrin repeat and PH domain 2 (ASAP2) (Andreev et al., 1999), Arg kinase binding protein 2 (ArgBP2) (Haglund et al., 2004), nephrocystin (Benzing et al., 2001) and proline-rich acidic protein (PRAP) (Takahashi et al., 2003).
Additionally, a NES and a nuclear targeting sequence (NTS) were found in the linker 2 of Pyk2, playing an important role in subcellular trafficking of Pyk2 (Faure et al., 2013).
FAT domain
The C-terminal FAT domain of FAK family kinases is highly conserved between Pyk2 and FAK. It exhibits an anti-parallel four-helix bundle structure in both FAK (Hayashi et al., 2002; Liu et al., 2002) and Pyk2 (Lulo et al., 2009) (Figure 7). In FAK, this domain is consubstantially linked to focal adhesion targeting (Hildebrand et al., 1993; Shen and Schaller, 1999). It was reported to interact with talin (Chen et al., 1995) and paxillin (Tachibana et al., 1995; Brown et al., 1996), two proteins highly enriched in focal adhesions. Conversely, Pyk2 FAT domain was shown to interact with paxillin (Lulo et al., 2009) but not talin (Zheng et al., 1998). Moreover, it was demonstrated that the binding mechanism between Pyk2 and FAK for paxillin was different and that paxillin thus formed a much more stable complex with the FAT domain of FAK than with the FAT domain of Pyk2 (Vanarotti et al., 2014). Instead of paxillin, Pyk2 FAT domain showed a preferential interaction with some proteins from the same paxillin superfamily such as leupaxin (Vanarotti et al., 2016) or the molecular scaffold and transcription co-regulator Hic-5 (Matsuya et al., 1998). These differences between FAK and Pyk2 perhaps explain why only a small proportion of Pyk2 is localized in focal contacts in most cell types.
Moreover, the FAT domain of FAK and Pyk2 associates with gelsolin, an actin binding protein, showing a common regulatory role of FAK and Pyk2 in actin cytoskeleton organization (Wang et al., 2003; Chan et al., 2009). Finally, within the FAT domain, Tyr-925 in FAK and Tyr-881 in Pyk2, when phosphorylated by Src, constitute a binding site for the adaptor Grb2 leading to the initiation of the MAP kinase signaling pathway (Schlaepfer and Hunter, 1996; Blaukat et al., 1999). In Pyk2, this phosphorylated tyrosine was also reported to be an anchoring site for the oncogenic TK c-Abl (Zrihan-Licht et al., 2004).
Table of contents :
INTRODUCTION
Tyrosine kinase signaling and the FAK family
1. The origin of TKs and FAK family
2. FAK family
Identification of FAK family kinases
Structure of FAK family kinases
2.2.1. FERM domain
2.2.2. Linker 1
2.2.3. Kinase domain
2.2.4. Linker 2
2.2.5. FAT domain
Expression of FAK and Pyk2
Cellular localization of FAK family kinases
Isoforms of FAK family kinases
2.5.1. FAK isoforms
2.5.2. Pyk2 isoforms
Biological functions of FAK
2.6.1. Cellular functions of FAK
2.6.2. Physiological functions of FAK
Pyk2: a nRTK of FAK family
1. Regulation of Pyk2 activity in non-neuronal cells
Activation and phosphorylation of Pyk2
1.1.1. Canonical activation of Pyk2
1.1.2. Regulation of Pyk2 activation by Ca2+-activated kinases
1.1.2.1. PKC
1.1.2.2. CaMKII
Dephosphorylation of Pyk2
1.2.1. Tyrosine phosphatases
1.2.1.1. SHP-1
1.2.1.2. SHP-2
1.2.1.3. PTP-PEST
1.2.1.4. STEP
1.2.2. Ser/Thr phosphatases
SUMOylation of Pyk2
S-nitrosylation of Pyk2
2. Pyk2 functions in non-neuronal cells
Pyk2 cellular functions
2.1.1. Cell adhesion
2.1.2. Cell migration
2.1.3. Cell division
2.1.4. Cell survival
2.1.5. Cell differentiation
Physiological role of Pyk2
2.2.1. Bone physiology
2.2.2. Vascular system integrity
2.2.3. Immune system function
2.2.4. Kidney function
2.2.5. Sperm capacitation
2.2.6. Generation of Pyk2 knockout mice
3. Pathological role of Pyk2
Pyk2 and inflammatory diseases
Pyk2 and cancers
Pharmacological inhibitors of Pyk2
Roles of Pyk2 in the CNS
1. Specific regulation of Pyk2 in the CNS
Activation of Pyk2 in neurons
Regulation of Pyk2 localization
2. Pyk2 biological functions in the CNS
Ionic channels regulation
2.1.1. Kv1.2
2.1.2. BK channels
2.1.3. NMDA receptor (NMDAR)
Development
Synaptic plasticity
Neuronal survival
Pyk2 in glial cells
3. Pyk2 in CNS diseases
Alzheimer·s disease
Parkinson·s disease
Huntington·s disease
Neuroinflammation
Glioma and neuroblastoma
Cerebral ischemia
Psychiatric disorders
RESULTS
Pyk2 modulates hippocampal excitatory synapses and contributes to cognitive deficits in a Huntington’s disease model
1. Context and objectives
2. Contribution to the work
3. Article
4. Summary of the findings and conclusions
Pyk2 in the amygdala modulates chronic stress sequelae via PSD-95-related microstructural changes
1. Context and objectives
2. Contribution to the work
3. Article
4. Summary of the findings and conclusions
PTK2B/Pyk2 overexpression improves a mouse model of Alzheimer’s disease
1. Context and objectives
2. Contribution to the work
3. Article
4. Summary of the findings and conclusions
Conditional BDNF Delivery from Astrocytes Rescues Memory Deficits, Spine Density, and Synaptic Properties in the 5xFAD Mouse Model of Alzheimer Disease.
1. Context and objectives
2. Contribution to the work
3. Article
4. Summary of the findings and conclusions
Pyk2 in nucleus accumbens D1 receptor-expressing neurons is selectively involved in the acute locomotor response to cocaine
1. Context and objectives
2. Contribution to the work
3. Article
4. Summary of the findings and conclusions
Supplementary data: spine density and morphology in the NAc of Pyk2-/- mice
1. Materials and methods
2. Results
DISCUSSION
Role of Pyk2 in memory
Kinase-dependent and independent functions of Pyk2
Antagonistic effect of Pyk2 on spine density and morphology
BDNF and Pyk2 merging functions
Pyk2 and AD: risk or rescue factor?
Contrasted function of Pyk2 in the striatum
BIBLIOGRAPHY
