|Home | About | Journals | Submit | Contact Us | Français|
Cell adhesion, migration, proliferation and differentiation are tightly linked and coordinated cellular processes. Cell adhesion dependent gene expression is believed to contribute to such coordination. Focal adhesion kinase (FAK) and its related protein, PYK2 (proline rich tyrosine kinase 2), are a major family of cell adhesion activated tyrosine kinases that play important roles in these cellular processes. Whereas FAK or PYK2 is known to be a scaffold protein, recruiting many cytoplasmic proteins into the focal adhesion complex and regulating focal adhesion turnover and cell migration, how FAK or PYK2 links to the nuclei and regulates gene expression remain largely unclear. We recently report a new signaling of FAK in regulating heterochromatin remodeling by its interaction with MBD2 (Methyl CpG binding domain protein 2), which may underlie FAK regulation of myogenin expression and muscle differentiation. Two insights have been obtained through the analysis of FAK-MBD2 interaction. The interaction appears to be sufficient, but not necessary, for FAK translocation into or maintaining in the nucleus. The nuclear FAK-MBD2 complexes cause altered heterochromatin organization and decreased MBD2 association with HDAC1 (histone deacetylase complex 1) and methyl CpG site in the myogenin promoter, thus, inducing myogenin expression. These results demonstrate a new mechanism underlying FAK regulation of gene expression, and suggest a potential link between cell adhesion and cell differentiation.
FAK and its related protein, PYK2, are a family of cytoplasmic tyrosine kinases that are activated upon cell-ECM (extracellular matrix) adhesion and mediate cell adhesion signaling in cell migration, survival and differentiation.1–5 FAK/PYK2 functions as a scaffold protein at adhesion site, where they interact with many cytoplasmic proteins and integrate signals of different environmental cues.1,2 Structurally, both contain a large N-terminal FERM domain, a central catalytic domain, and a large C-terminal domain, where two proline rich motifs and focal adhesion targeting (FAT) domain are identified.2 Many proteins, most of them are cytoplasmic proteins, have been found to interact with FAK/PYK2 via the C-terminal domain. For examples, the proline rich sequences in the C-terminal domain of FAK/PYK2 bind to SH3 domains of p130Cas (Crk associated substrate) family (e.g., Cas and HEF1),6 Pap family (PH domain and SH3 domain containing ArfGAP proteins, e.g., ASAP1 and Pap),7 and Graf/PSGAP family proteins (PH and SH3 domain containing RhoGAP proteins, e.g., Graf and PSGAP).8,9 The FAT domain of FAK/PYK2 is critical for its association with cytoskeletal proteins (talin, paxillin family proteins).3,10–12
It is note worthy that FAK and PYK2, although showing remarkable similarities as described above, are involved in many different signaling and functions.13 For examples, FAK, but not PYK2, interacts with DCC (deleted in colorectal cancer) family netrin receptors and such interaction is important for FAK regulation of neural axon pathfinding.14–16 PYK2, but not FAK, interacts with gelsolin, a F-actin severing protein, and this interaction appears to be crucial for PYK2 regulation of podosome turnover in osteoclasts and osteoclast activation.17 Moreover, FAK knock out mouse leads to early embryonic lethal18 and PYK2 knock out mice can survive to adult stage.19,20 Many phenotypes of FAK knock out are often not rescued by expression of PYK2.18
Although FAK or PYK2 is a resident at the focal adhesions or cytoplasm, they, under specific circumstances, translocate to the nuclei and play a role in the control of nuclear functions such as gene expression and cell cycle. Serine 722, but not tyrosine 397, phosphorylated FAK, is found to localized at the nuclei in hypertrophic myocardium.21 Via KLF8, a transcriptional factor, FAK regulates cyclin D1 and cellcycle progression.5 PYK2, a FAK related tyrosine kinase, can be translocated to the nuclei upon depolarization in cultured neurons/PC12 cells or in cultured hippocampal slices.22 Exactly how does FAK or PYK2 translocate to the nuclei and regulate gene expression or cell cycle remains unclear. Recently, it has been reported that FAK N-terminal FERM domain contains a NLS (nuclear localization signal) that is necessary for FAK nuclear translocation23 and FAK kinase domain contains a functional NES (nuclear export signal).24 In our studies, we found that FAK and PYK2, via their FAT domains, bind to MBD2.25 This interaction is sufficient for FAK/PYK2 translocation into the nuclei.25 We thus speculate that MBD2 binding to the FAK FAT domain may change the conformation of FAK, preventing FAK targeting to focal adhesions, exposing FAK N-terminal NLS, thus, increasing FAK nuclear translocation. In addition, MBD2 binding to FAK at the nuclei may also prevent FAK export from nucleus, maintaining FAK localization at the nuclei. Whereas this notion requires additional experimental evidence, these studies provide novel insights into FAK/PYK2 translocation to and function in the nuclei, implicating a novel link for coordination between cell adhesion and gene expression at the nuclei.
MBD2 belongs to a family of MBD domain containing proteins, including MBD1, MBD2, MBD3, MBD4 and MeCP2. They associate with heterochromatin in the nuclei by their interaction with methylated DNA at CpG islands. They appear to “read” and “translate” the DNA methylation signal into transcriptional repression at least partially by recruiting silencing complexes and histone deacetylases, thereby stabilizing and consolidating the heterochromatic state.26,27 MBD2 is the methyl-binding component of the MeCP1 complex, which contains the histone deacetylase (HDAC) proteins HDAC1, HDAC2 and the RbAp46 and RbAp48 proteins (also known as RBBP7 and RBBP4), allowing MBD2 to target HDAC/chromatin remodeling activity to methylated templates.26,27 MBD2 can also act synergistically with other proteins; for example, the zinc finger protein MIZF interacts with MBD2 and acts as an HDAC-dependent transcriptional repressor.28 In light of our observations that FAK-MBD2 interaction at the nucleus reduces MBD2 binding to the CpG island of myogenin promoter, and enhances myogenin induction during muscle differentiation, we speculate that MBD2 may also act as a “docking site” of chromatin, recruiting “activators” to turn on gene expression during muscle differentiation. MBD2-HDAC or MeCP1 complex, a dominant complex for heterochromatin remodeling associated with transcriptional repression, may be regulated under specific circumstances. When muscles exposed to oxidative stress, FAK/PYK2 associates with MBD2 in the nuclei, such interaction may displace HDAC1/2’s binding with MBD2, attenuates MBD2’s suppressor activity (Fig. 1A). On the other hand, it is possible that FAK/PYK2-MBD2 interaction may also enhance or recruit additional molecular component to activate MBD2’s “demethylase” activity, thus, reducing methylation of myogenin promoter and increasing myogenin transcription. It is of interest to note that MBD2 has been reported to be an enzyme capable of actively demethylating DNA.29–32 However, the latter speculation needs additional experiments to support.
Our results have suggested a role for FAK-MBD2 interaction in muscle differentiation. Muscle cell differentiation involves the fusion of myoblasts into multi-nucleated myotubes, a process regulated by the activation of myogenic basic helix-loop-helix (bHLH) factors (e.g., MyoD, Myf5, myogenin and MRF4). Myogenin is believed to be crucial for the expression of genes associated with muscle terminal differentiation, while MyoD and Myf5 are required for commitment to or initiation of the myogenesis. C2C12 cells provide an excellent culture model system for studying muscle differentiation, as it is technically easy to be differentiated in culture and genetically manipulated. Using this model, our results have suggested that myogenin is induced under oxidative stress by the FAK-MBD2 pathway.25 Muscle differentiation requires both up and down-regulation of many genes, and this process appears to be associated/coupled with muscle heterochromatin remodeling.33 Formation of large clusters of heterochromatin, characterized by high levels of DNA methylation, specifically methylated forms of histone H3, and deacetylated histone H4, is observed during muscle differentiation.34 This event correlates with overexpression of the methyl CpG-binding domain (MBD) proteins, MeCP2 or MBD2, in muscle cells,34 corroborating with our observation that increased FAK-MBD2 complex leads to the formation of large clusters of heterochromatin.25 The nuclear FAK-MBD2 complex thus may regulate expression of multiple genes, in addition to myogenin, a question that remains to be further investigated.
There is increasing evidence that oxidative stress (e.g., H2O2) appears to mediate dual functions on myogenesis, depending on its dose and time of treatment. Not merely as a damaging agent, it also acts as a useful signaling molecule to regulate growth, differentiation, proliferation and apoptosis, at least within the physiological concentration. When higher dose of H2O2 (e.g., >100 µM) is applied prior to differentiation induction in C2C12 cells, but not after the differentiation program, it blocks myogenin expression and myogenic differentiation. However, exposure of low level of H2O2 can enhance muscle differentiation and function.35 Our paper suggests that FAK nuclear activity in myotubes is enhanced by H2O2 exposure. However, it remains unclear the physiological source of H2O2 in myotubes. We speculate that increased metabolic activity during muscle contraction may produce low levels of oxidants and that FAK/PYK2-MBD2 complex at the nuclei may be involved in oxidative stress-induced chromatin remodeling and multiple gene expression (e.g., anti-oxidant genes).
In addition to muscles, our working model may be extended to other cell types, such as neurons (Fig. 1B). Of interest to note is that neuronal activity induces expression of BDNF (brain derived nerve growth factor), a NGF related growth factor that plays an important role for neural development and neuronal synaptic function. This BDNF induction is associated with the DNA methylation related chromatin remodeling at its promoter region.36,37 The increased BDNF transcription involves dissociation of the MeCP2-histone deacetylase-mSin3A repression complex from its promoter in depolarized neurons.37 This event appears to be due to membrane depolarization induced calcium-dependent phosphorylation and release of MeCP2 from BDNF promoter III.36 This view is remarkably similar to what we have found in muscle cells25 (Fig. 1A). We thus speculate that in addition to MeCP2, MBD2 may also contribute to BDNF induction (Fig. 1B). This working model is in line with the observations that PYK2, but not FAK, is highly expressed in mature neurons and translocates to the nuclei of hippocampal neurons after membrane depolarization,22 that PYK2, but not FAK, is activated in response to increased calcium flux,38,39 and that PYK2-MBD2 binding can reduces MBD2 association with HDAC complex.25 This working hypothesis also raises many questions. Is PYK2 involved in BDNF induction by neuronal membrane depolarization Is PYK2-MBD2 interaction regulated by neuronal activity Does MBD2 bind to the BDNF promoter III These questions remain to be further investigated.
Together, our results suggest that FAK translocates into nuclei under specific circumstances (e.g., exposure to oxidative stress of muscle cells), where it forms a complex with MBD2, attenuates MBD2’s transcriptional repressor activity, activates gene (e.g., myogenin) expression, and promotes muscle terminal differentiation. These results suggest a previously unrecognized mechanism underlying FAK/PYK2 regulation of gene expression, and reveal a new mechanism of MBD2 regulation by FAK/PYK2 tyrosine kinases.
This work is supported by NIH (National Institutes of Health) (to L.M. and W.-C. Xiong).
Previously published online: www.landesbioscience.com/journals/celladhesion/article/10343