Pyk2 (proline-rich tyrosine kinase 2) and FAK (focal adhesion kinase) are highly related tyrosine kinases. One distinguishing feature is the differential regulation of the two enzymes in response to elevation of cytoplasmic calcium. In the latest issue of the Biochemical Journal, Sasaki and co-workers have provided insight into the calcium-dependent regulation of Pyk2. The findings suggest that calmodulin may bind the FERM (4.1/ezrin/radixin/moesin) domain to promote Pyk2 activation in response to calcium signals triggered by vasopressin. While the molecular details of the protein–protein interaction and mechanism of activation remain to be firmly established, this study is the first to provide mechanistic insight into the regulation of Pyk2 by calcium.
The FAK (focal adhesion kinase) family of proteins consists of two members, FAK and Pyk2 (proline-rich tyrosine kinase 2). They share the same domain organization of an N-terminal FERM (4.1/ezrin/radixin/moesin) domain, a central catalytic domain and a C-terminal domain with its docking sites for other signalling proteins, and in the case of FAK sequences essential for subcellular localization to sites of adhesion. FAK is broadly expressed and is essential for embryonic development, e.g. angiogenesis, heart development and axon guidance. In contrast, Pyk2 is more restricted in its expression and is dispensable for normal development. However, Pyk2 plays important physiological roles in macrophages and osteoclasts [3,8].
FAK is a well-established regulator of the cytoskeleton and cell migration. Pyk2 also regulates the cytoskeleton and cell migration in macrophages and osteoclasts [3,8]. Interestingly, there are several examples where the cellular activity of Pyk2 and FAK is apparently cell-context-dependent. The role of FAK in promoting fibroblast motility is well known, but Pyk2 has little effect on fibroblast migration . In several glioblastoma cell lines, Pyk2 expression is associated with increased motility and FAK expression is associated with decreased motility, but increased proliferation . These examples illustrate that biological responses of cells expressing both Pyk2 and FAK may be highly dependent upon the activation levels of the two kinases in response to different stimuli, and thus defining the mechanistic details of differential regulation is highly significant.
Pyk2 and FAK are regulated differently. FAK strongly co-localizes with integrins, which are receptors for proteins of the extracellular matrix, and integrin-dependent cell adhesion is a major activation stimulus for FAK. Pyk2 generally exhibits a diffuse cellular localization, and integrin-dependent cell adhesion acts as a minor stimulus for activation of Pyk2 . Soluble ligands, such as growth factors and neuropeptides, can stimulate both FAK and Pyk2, but these ligands are usually strong activators of Pyk2 and weak activators of FAK . This result could be attributed to the response of Pyk2 to elevated levels of cytoplasmic calcium. As FAK is refractory to signalling through changes in cytoplasmic calcium, this is an important signalling mechanism that distinguishes these two kinases. The mechanism of regulation of Pyk2 by calcium is the major focus of the article by Sasaki and co-workers , which appears in this issue of the Biochemical Journal and provides important new insights into the differential regulation of Pyk2 and FAK.
There are several molecular events that occur during activation of these kinases. A number of studies have clearly established that the N-terminal FERM domain of FAK is autoinhibitory and operates by complexing with the catalytic domain to block access to the active site of the enzyme . Based upon the conservation of sequence of the FERM/catalytic domain interface between FAK and Pyk2, it seems likely that Pyk2 can assemble into a similar autoinhibited conformation. An important step in kinase activation is the release of this autoinhibitory interaction and adoption of an active conformation. It should also be noted that the FERM domains of FAK and Pyk2 have additional functions. For example, the FAK FERM domain is required to promote growth-factor-induced motility of fibroblasts, and the Pyk2 FERM domain is important for the regulation of migration in glioblastoma cells [7,9]. Other key steps in the activation of Pyk2 and FAK include the phosphorylation of a key tyrosine residue between the FERM and catalytic domains: Tyr402 in Pyk2 and Tyr397 in FAK. Phosphorylation of these sites is critical for function, as they recruit other signalling molecules into the complex, including Src family kinases. This event is important for the transmission of downstream biochemical signals and most biological responses controlled by these kinases. Additional phosphorylation of tyrosine residues within the activation loop, Tyr579/580 in Pyk2 and Tyr576/577 in FAK, is important for maximal catalytic activity. It is possible that differential regulation of FAK and Pyk2 could be explained by differential modulation at any of these steps of activation.
The role of calcium in the regulation of Pyk2, which is also called calcium-dependent tyrosine kinase, has been known for many years. The mechanism of regulation is indirect, since Pyk2 does not bind calcium and the activity of the purified enzyme is not regulated by calcium in in vitro assays. Sasaki and co-workers now provide evidence that calmodulin regulates Pyk2 activation, since two structurally related inhibitors of calmodulin block vasopressin-induced tyrosine phosphorylation of Pyk2 in the WFB rat fibroblast cell line . This finding is substantiated since Pyk2, but not FAK, associates with calmodulin–agarose beads in a calcium-dependent manner, and calmodulin can be co-immunoprecipitated with Pyk2. Interestingly, the regulatory FERM domain of Pyk2 interacts with calmodulin, suggesting that calmodulin–FERM domain interactions might be important for calcium-dependent activation of Pyk2, although additional interactions with other domains cannot be excluded. Although these results are consistent with a direct interaction between calmodulin and the FERM domain of Pyk2, it is important to note that all of these studies were performed using cell lysates, and thus an indirect mechanism of binding cannot be excluded. One important area of future research is the elucidation of the precise molecular mechanism of interaction of these proteins. The results are likely to be highly significant, given the importance of the FERM domain in regulating this kinase family.
FERM domains contain three subdomains that structurally resemble ubiquitin (the F1 subdomain), the acyl-CoA-binding protein (the F2 subdomain) and a PH (pleckstrin homology) domain (the F3 subdomain). Substitution of the F2 subdomain of FAK for the F2 subdomain of the Pyk2 FERM domain abolishes binding to calmodulin, and this loss of function is consistent with the hypothesis that calmodulin binding occurs through this subdomain of the FERM domain . This could be highly significant, since the F2 subdomain provides the major contact with the catalytic domain in the autoinhibited conformation of FAK . In addition, another conserved sequence that is important for FAK activation in response to multiple stimuli is also contained within the F2 subdomain . Sasaki and co-workers  identify an α-helical sequence within the F2 subdomain of the Pyk2 FERM domain that resembles a reverse basic 1–8–14 motif that functions as a calmodulin-binding site in the HIV gp160 protein and Ca2+/calmodulin-dependent protein kinase α. The location of the motif is intriguing, since the loop following this α-helix is a critical component of the catalytic-domain-binding site within the FERM domain of FAK. However, structural studies suggest that this motif is unlikely to function in calmodulin binding. Calmodulin has N- and C-terminal globular domains, which contain Ca2+-binding EF-hand motifs. These two domains clamp down on peptide ligands to surround the α-helical motif. Although the structure of the FERM domain of Pyk2 has not been solved, structural studies on FAK suggest that this interaction could not occur, since the α-helix is buried in the core of the F2 subdomain. To verify the role of this peptide motif in calmodulin binding, Sasaki and co-workers  mutated two residues within the reverse basic 1–8–14 motif in Pyk2 (the LQ/AA mutant). This resulted in attenuation of calmodulin binding and tyrosine phosphorylation in vivo. These two residues are near the surface, but partially buried in the autoinhibited structure of FAK, and presumably in Pyk2. Thus these mutations might also alter the conformation of the FERM domain to obscure a distant binding site. Regardless, these results provide correlative evidence between calmodulin binding and tyrosine phosphorylation in vivo, and thus support the overall hypothesis. Further studies are required to fully elucidate the molecular mechanism of the interaction of calmodulin with the Pyk2 FERM domain.
The other intriguing observation described in the manuscript by Sasaki and co-workers  is the calcium-dependent assembly of Pyk2 into higher-order complexes containing Pyk2 dimers in complex with calmodulin. Again, this activity is mediated by the FERM domain, since the Pyk2 FERM domain can form a similar complex. The assembly of this complex may be significant, as it provides several additional hypothetical mechanisms of Pyk2 activation by calmodulin. By promoting dimerization, calmodulin could enhance autophosphorylation of Pyk2, which is important for recruiting other signalling molecules into the complex and transmission of a downstream signal. Through a second, but not mutually exclusive, mechanism, dimerization could promote activation loop phosphorylation and optimal Pyk2 kinase activity. Interestingly, inhibition of cytoplasmic calcium is reported to block AngII (angiotensin II)-induced phosphorylation of Pyk2 at Tyr580, but not at Tyr402 . This is consistent with two mechanisms of regulation: (1) disruption of the autoinhibited conformation (since the activation loop is sequestered between the catalytic and FERM domains in the autoinhibited conformation of FAK) and (2) increased activation loop phosphorylation due to dimerization. Additional studies are required to distinguish between these possibilities.
In summary, Sasaki and co-workers  have provided the first evidence of a role for calmodulin in the calcium-dependent regulation of Pyk2. The model of activation envisions calmodulin binding to the FERM domain of Pyk2 leading to altered conformation and/or dimerization, and consequently elevated Pyk2 signalling. While the mechanistic details of calmodulin binding and its precise role in regulating Pyk2 remain to be established, this provocative paper will certainly stimulate additional studies to fully elucidate the relationship between calmodulin and Pyk2 in calcium-dependent regulation.
I thank Derek Ceccarelli for valuable discussions during the preparation of this article. Research in the author's laboratory is supported by NIH grant HL45100.