Voltage-gated Kv2.1 Channel Enhances Cell Motility Through Increasing FAK Phosphorylation

FAK is a 125-kDa non-receptor protein tyrosine kinase in the plasma membrane;it colocalizes with integrins at focal adhesion sites and binds to several signaling proteins to form multicomponent complexes. FAK becomes phosphorylated and activated in response to integrin-mediated cell adhesion signal from the extracellular matrix. FAK activation is critical to focal adhesion dynamics, cell spreading, migration, tumor invasion, and cancer cell proliferation. Voltage-gated K⁺ channels are a family of membrane proteins that mediate K⁺ efflux and determine a neuron's intrinsic electrical excitability. Delayed rectifier Kv channels, specifically the Kv2.1 channel, are the major component of somatodendritic K⁺ channels in the brain of hippocampal and cortical pyramidal neurons. Because of their own important physiological and pathological roles, FAK and K⁺ channels have been enthusiastic but separate research areas;there has been no evidence establishing a direct structural or functional link between FAK and Kv.2.1 channel.Firstly, we checked the Kv2.1 expression level in two cancer cell lines. The data showed that there were some Kv2.1 protein in Hela and SK-Hep-1 cell lines. Then we checked whether the migration of cancer cells needed potassium channel activity using Hela cells. The result showed that both TEA and 25 mM K⁺ impaired the cell migration.In searching for regulatory factors of Kv2.1 channels, our immunoprecipitation and mass spectrometry experiment showed a close association between Kv2.1 and motility-associated proteins including vimentin and FAK. This observation led us to speculate that this ion channel might play a role in cell motility. To test this possibility, we first examined localization of Kv2.1 and FAK in several cell types in vivo and in vitro. Immunoprecipitation experiments using mouse cortex proteins revealed a complex containing Kv2.1, FAK and Src. Immunofluorescent staining in cortical neurons of 14 days in vitro (DIV) showed overlapped distribution of Kv2.1 and FAK on the cell soma. Colocalization of Kv2.1 and phospho-FAK^576/577 at the contact regions between cells was also observed in glial cultures of the mouse cortex. To further confirm these observations, we introduced EGFP-tagged Kv2.1gene into Chinese hamster ovary (CHO) cells that inherently lack endogenous delayed rectifier channels. Confocal imaging with the fluorescent antibody against phosphorylated pFAK397 illustrated a clustered distribution on the cell soma;FAK distribution was distinctly overlapped with transfected Kv2.1. A similar co-distribution pattern was also seen with phosphorylated pSrc416 and Kv2.1.To determine the functional domain of KV2.1 protein responsible for binding to FAK, a flag epitope tagged Kv2.1 and multiple deletion mutations (1-712, 1-416, 184-416,184-857, 416-712, and 416-857) were constructed for transfection and immunoprecipitation. Our data showed that only mutantes retained the Tl domain and P loop (1-857 of full length, 1-712, and 1-416) were able to form complex with FAK. To identify the region of FAK responsible for binding to KV2.1, we generated several FAK mutants tagged with myc epitope. Mutants containing the FAT domain (800-1055) did not interact with Kv2.1. However, the mutant containing only the N-terminus FERM domain showed similar binding activity as full length FAK, thus confirming the FERM domain was both required and sufficient for the interaction with Kv2.1. Focusing on the N-terminal domain of KV2.1, we noticed an amino acid sequence homologous to the LD motif (45-56) of paxillin that might bind FAK. Site-directed mutagenesis and immunoprecipitation experiments were performed to explore whether this LD-like motif was required for KV2.1 binding to FAK. Within the motif, when leucine amino acids (L45 and L48) were replaced with serines, FAK-KV2.1 binding activity was markedly decreased. On the other hand, FAK-KV2.1 binding was unaffected when R52 was mutated to leucine. To further confirm our observation, we carried out confocal imaging on CHO cells transfected with EGFP-Kv2.1 or a mutated Kv2.1 lacking the N-terminal 1-50 segment (Kv2.1Al-50). Clustered distribution of Kv2.1 and colocalization with FAK were observed only in CHO cells transfected with full length Kv2.1. Conversely, transfected Kv2.1Al-50 was evenly distributed throughout the cell and showed no overlap with FAK. Thus our data reveal the importance of the LD-like motif of Kv2.1 protein in binding to FAK. To elucidate whether the formation of K+ channel-FAK complex was channel subtype specific, EGFP-tagged KV1.5 protein was introduced into CHO cells. Immunostaining showed no overlapping between FAK and Kvl.5.Then we stably transfected CHO cells with full length wild type KV2.1 and dominant negative KV2.1, which has been shown resulted in depressed currents. The high expression level cell lines were selected using immunoblotting. Transfection of Kv2.1 gene into wildtype CHO cells resulted in sizable outward currents sensitive to block by the K+ channel blocker TEA. Wild type CHO cells and Kv2.1 transfected CHO cells (KV2.1-CH0) were then tested for adhesion to fibronectin-coated Petri dishes. It was apparent that Kv2.1 expression drastically increased the adhesion of CHO cells to fibronectin. TEA (5 mM) and elevated extracellular K+ (25 mM), two maneuvers to attenuate K+ efflux, reduced the cell adhesion activity. In confocal imaging study of CHO cells transfected with EGFP-tagged Kv2.1, we consistently observed that without the fibronectin coating Kv2.1 showed a lack of colocalization with FAK while in fibronectin-coated dishes Kv2.1 mostly colocalized with FAK. In 25 mM K+ medium, this colocalization was significantly reduced. Furthermore, immunoprecipitation with Kv2.1 antibody showed an increased formation of the Kv2.1-FAK complex in Kv2.1-CH0 cells after plating to fibronectin-coated dishes, TEA and elevated K+ attenuated the complex formation. The same results were obtained in fibroblasts overexpressing FAK (FAK+/+ cells). To further verify whether K+ efflux was critical for Kv2.1-FAK interaction, FAK+/+ cells were plated on dishes coated with fibronectin and the selective K+ ionophore valinomycin (1 nM) was added to stimulate K+ efflux. As expected, the formation of FAK-KV2.1 complex was increased dramatically. The binding of Src to the complex was also increased by valinomycin.Previous investigations have revealed that integrin clustering promotes FAK autophosphorylation at pFAK397, creating a binding site for the Src-homology 2 (SH2) domain. Src-mediated phosphorylation of FAK at pFAK576/577 promotes maximal FAK catalytic activity and facilitates downstream events including Rac activation, lamellipodia formation and cell migration. Therefore, the interaction between FAK and Src is important for FAK-mediated effects on cell motility. Immunoprecipitaion examination in KV2.1-CH0 and FAK+/+cells showed that the FAK-Src complex was largely inhibited when K+ channel activity or K+ efflux was reduced. Similar results were obtained with the specific Src family kinases inhibitor, PP2 (10 uM). We next tested the hypothesis that Kv2.1 binding to FAK could increase phosphorylation of FAK. In fibronectin-coated dishes, pFAK397 autophsophorylation and pFAK576/5r7 phosphorylation were much higher in KV2.1-CHO cells than that in wild type CHO cells. The pFAK576/577 phosphorylation is believed caused by Src kinase and occurs after pFAK397 autophosphorylation. Consistently, PP2 (10 uM) showed no effect on pFAK397 autophosphorylation but substantially blocked pFAK576/577 phosphorylation. Similarly, TEA (5 mM) and K+ (25 mM) had little effect on pFAK397autophosphorylation, but attenuated pFAK576'577 phosphorylation. Interestingly, TEA and elevated K+ showed no inhibitory effect on increased activity/phosphorylation of pSrc416, which is a prerequisite for pFAK576/577 phosphorylation. The differential effects of TEA, high K+ and PP2 on pFAK397, pFAK576/577 and pSrc416 were similarly observed in FAK+/+ cells. Stimulating K+ efflux by valinomycin selectively augmented phosphorylation ofpFAK576/577 an(j pSrc416 jn pAK+/+ ^.^Finally, we demonstrated that the interaction between Kv2.1 and FAK could regulate cell migration. In wound healing tests, Kv2.1-CH0 cells exhibited significantly enhanced cell motility compared to wild type CHO cells or CHO cells expressing dominant negative KV2.1. In line with their inhibitory effects on Kv2.I-FAK-Src complex formation, TEA and elevated extracellular K+ restrained cell migration. Likewise, PP2 but not its inactive analog PP3 substantially reduced cell migration. Expression of Kv2.1 promoted the morphological change of cell polarization and formation of lamellipodium. The fluorescent EGFP labeled Kv2.1 channel in CHO cells showed characteristic assembling of the channel at the cell bottom of adhesion regions. In fibroblast cells Kv2.1 and FAK were found concentrated at the rear part of the cell soma and at the focal adhesion sites of the leading edge of lamellipodium.Conclusion: The present investigation provided the novel evidence that LD-like motif localized in N-terminal of Kv2.1 could interact with FAK directly to regulate formation of FAK-Src complex. Furthermore, we also found that Kv2.1 could enhance cell motility through increasing FAK phosphorylattion.