Cell migration requires the coordination of multiple signaling pathways involved in membrane dynamics and cytoskeletal rearrangement. The Arf-like small GTPase Arl4A has been shown to modulate actin cytoskeleton remodeling. However, evidence of the function of Arl4A in cell migration is insufficient. Here, we report that Arl4A acts with the serine/threonine protein kinase Pak1 to modulate cell migration through their cooperative recruitment to the plasma membrane. We first observed that Arl4A and its isoform Arl4D interact with Pak1 and Pak2 and showed that Arl4A recruits Pak1 and Pak2 to the plasma membrane. The fibronectin-induced Pak1 localization at the plasma membrane is reduced in Arl4A-depleted cells. Unexpectedly, we found that Pak1, but not Arl4A-binding-defective Pak1, can recruit a cytoplasmic myristoylation-deficient Arl4A-G2A mutant to the plasma membrane. Furthermore, we found that the Arl4A-Pak1 interaction, which is independent of Rac1 binding to Pak1, is required for Arl4A-induced cell migration. Thus, we infer that there is feedback regulation between Arl4A and Pak1, in which they mutually recruit each other to the plasma membrane for Pak1 activation, thereby modulating cell migration through direct interaction.
ADP-ribosylation factor (Arf) and Arf-like (Arl) proteins belong to the Ras superfamily of small GTPases and are involved in membrane transport, organelle integrity maintenance, membrane lipid modification, and cytoskeletal dynamics via cyclic regulation between their GTP-bound active form and their GDP-bound inactive form (D'Souza-Schorey and Chavrier, 2006; Gillingham and Munro, 2007). The Arl4 small GTPases, which include Arl4A, Arl4C and Arl4D, are unique among the Arls in that their expression is tissue specific and tightly regulated during fetal development, whereas other Arf proteins appear to be expressed ubiquitously (Lin et al., 2000, 2002; Schurmann et al., 2002). Recent studies have shown that Arl4 subfamily members are involved in vesicle trafficking (Engel et al., 2004; Wei et al., 2009), organelle structure (Li et al., 2012; Lin et al., 2011) and cytoskeleton organization (Chiang et al., 2017), resembling the main function of most Arfs (Donaldson and Jackson, 2011; Gillingham and Munro, 2007). For example, Arl4A directly interacts with the golgin GCC185 (GCC2) through the Rab-binding domain and participates in the function of GCC185 to maintain Golgi structure and endosome-to-Golgi transport (Lin et al., 2011). Arl4A also complexes with ELMO/DOCK180 to regulate membrane protrusion and the actin cytoskeleton (Chiang et al., 2019). In addition to localizing to the cytosol and plasma membrane (PM), Arl4A also localizes to endosomes, Golgi and the nucleus (Lin et al., 2000, 2011). However, how Arl4 achieves differential regulation of signaling in different cellular contexts is poorly understood.
Cell migration occurs in physiological processes, including embryonic development, tissue repair and immune response, and pathological processes such as cancer (Parsons et al., 2010). The molecular mechanisms of cell migration have been extensively studied but not fully elucidated. Cell migration requires the coordination of multiple signaling pathways involved in membrane dynamics and cytoskeletal rearrangement. Previous studies have shown that Arl4a mRNA is specifically upregulated in hepatocyte growth factor-stimulated polarized MDCK cells (Balkovetz et al., 2004), SW620 metastatic colorectal cancer cells, and in the process of monocyte differentiation (Kubosaki et al., 2009). Additionally, our recent study showed that Arl4A participates in Slit2/Robo1 signaling to modulate cell motility (Chiang et al., 2019); however, evidence of the function of Arl4A in cell migration is insufficient.
p21-activated kinase 1 (Pak1) is a member of an evolutionarily conserved family of serine/threonine kinases that are involved in the regulation of cell motility, cytoskeletal dynamics, cell shape and adhesion (Delorme-Walker et al., 2011; Ma et al., 2012; Zhao and Manser, 2012). Paks are composed of an N-terminal regulatory domain and highly conserved C-terminal catalytic domain. The N-terminal region of Pak1 contains several sequence motifs responsible for interacting with partner proteins, such as Grb2, Nck1 (hereafter referred to as Nck) and the Cdc42/Rac exchange factor Arhgef7 (hereafter referred to as PIX) (Semenova and Chernoff, 2017; Zhao and Manser, 2012). For instance, the binding of activated Cdc42 or Rac1 to the p21-binding domain (PBD) releases Pak1 from the autoinhibitory domain (AID) and activates the kinase. Once activated, Pak1 undergoes autophosphorylation at specific sites and then phosphorylates a variety of substrate effector proteins (Bokoch, 2003). As previously shown, PIX and GIT1/2 (Arf GTPase-activating proteins) can activate Pak (Daniels et al., 1999; Loo et al., 2004), indicating that the GIT/PIX/Pak complex can function independently of binding of Cdc42/Rac GTPases. Notably, the efficient activation of Pak1 requires its targeting to the PM (del Pozo et al., 2000; Lu et al., 1997). The SH3-containing proteins Nck and Grb2 are believed to mediate membrane localization of Pak1 (Bokoch et al., 1996; Daniels et al., 1998; Lu et al., 1997), where it can be activated by signaling molecules, such as Pdk1 kinase (King et al., 2000), sphingosine (Bokoch et al., 1998; Zenke et al., 1999) and phosphatidylinositol (4,5)-bisphosphate (PIP2) (Malecka et al., 2013), indicating that Nck and Grb2 activate Pak1 in a GTPase-independent manner.
Although Rho GTPases are well-characterized activators of Pak1, the roles of other factors in Pak1 activation are less clear. Here, we identified Pak1 as a novel effector of Arl4A and found that Arl4A recruits Pak1 to the PM to promote Pak1 activity in a Rac1-independent manner. We also showed that Pak1 can facilitate the association of Arl4A with the PM and that the Arl4A-Pak1 interaction is crucial for the promotion of cell migration. Thus, our data provide a model for the targeting of Pak1 from the cytosol to the PM through an Arl4A-Pak1 interaction-dependent feedback loop, which further enhances Pak1 activation and cell migration.
Arl4A directly interacts with Pak1 through the Pak-PBD domain
Arl4A functions in the regulation of cytoskeleton rearrangement and membrane dynamics either by interacting with ARNO (Cyth2)/Arf6 signaling (Li et al., 2007; Matsumoto et al., 2014) or the ELMO/DOCK180 complex (Patel et al., 2011), which is required for the activation of Rac1 and cell proliferation mediated by integrin (Katoh and Negishi, 2003). Therefore, we hypothesized that Arl4A promotes cell migration through the activation of Rac1. We performed an activity pull-down assay of Rac1 by utilizing the Pak-p21 binding domain (PBD, Pak1 amino acids 70-127), which has a high affinity for the GTP-bound form of Cdc42/Rac1 (Manser et al., 1994). As shown previously (del Pozo et al., 2000), we found that Rac1 activity increased when HeLa cells were treated with fibronectin for 30 min (Fig. S1A). To examine whether Arl4A enhances Rac1 activation during cell proliferation, we compared Rac1 activity in fibronectin-treated cells transfected with or without Arl4A. As shown in Fig. 1A, Rac1 activity increased in cells after fibronectin treatment. Arl4A overexpression promoted Rac1 activity but did not further enhance Rac1 activity upon fibronectin stimulation. Notably, this pull-down assay revealed that Arl4A was strongly associated with Pak1-PBD (Fig. 1A). We also observed that the level of Arl4A associated with Pak1-PBD did not correlate with the amount of Rac1 pull-down, suggesting a possible interaction between Arl4A and Pak1. We next used immunoprecipitation to verify the in vivo results. We found that constitutively active Arl4A-Q79L (nucleotide-bound form mimic), but not inactive Arl4A-T34N (nucleotide-free form mimic), co-precipitated with wild-type (WT) Pak1 (Fig. 1B). Similarly, an in vitro glutathione-S-transferase (GST) pull-down assay demonstrated that Arl4A-Q79L, but not Arl4A-T34N, expressed in HeLa cells, had a stronger interaction with GST-fused Pak1-PBD (Fig. S1B). To examine the potential direct interaction between Arl4A and Pak1-PBD, we performed in vitro binding assays using purified recombinant proteins prepared from Escherischia coli. For comparison, His-tagged, constitutively active Rac-G12V had higher affinity binding to GST-Pak1-PBD than either WT or inactive (TN) Rac1 (Fig. S1C). Interestingly, we found that both Arl4A-WT and Arl4A-Q79L interacted more strongly with Pak1-PBD than did Arl4A-T34N (Fig. 1C). These results support the idea that Arl4A directly interacts with the Pak1-PBD domain in a GTP-dependent manner. Because Arl4A, Arl4C and Arl4D isoforms have high amino acid sequence identity, we also examined whether Pak1-PBD interacts with these Arl4 isoforms. We found that Arl4D, like Arl4A, interacted with Pak1-PBD in vitro, whereas only a weak interaction between Pak1-PBD and Arl4C was observed (Fig. S1D,E).
To identify the specific Pak1 regions that interact with Arl4A, we generated four GST-fusion Pak1 constructs: amino acids (a.a.) 1-70, 70-127 (PBD domain), 127-251 and 251-545 (kinase-containing domain) for the GST pull-down assay (Fig. 1D). Arl4A was overexpressed in HeLa cells and lysates were subjected to a GST pull-down assay. We found that exogenous Arl4A was specifically pulled down by the Pak1 70-127 fragment, but not by the other Pak1 fragments (Fig. 1E). As a positive control, we demonstrated that this PBD-domain fragment pulled down endogenous Rac1. It is well-documented that the Cdc42/Rac1 interactive-binding (CRIB) domain (a.a. 70-95) partially overlaps the AID (a.a. 83-149) in the N-terminal region of Pak1 (Lei et al., 2000). Thus, using additional Pak1 truncations we refined the Arl4A and Rac1 interaction within the PBD domain (Fig. 1F). The results showed that Arl4A, but not Rac1, interacted with Pak1 83-149 and suggest that Arl4A binds to the C terminus of the Pak1-PDB, which differs from the core Cdc42/Rac1-binding motif.
Arl4A induces the plasma membrane localization of Pak1
Several studies have shown that Pak1 translocates to the PM for potentiation of Pak1 activation upon stimulation (Bokoch et al., 1998; Lu et al., 1997; Lu and Mayer, 1999). As previously shown (Patel et al., 2011), active Arl4A colocalized with actin filaments at membrane protrusion sites (Fig. S2A). To examine whether the Arl4A-Pak1 interaction corresponds with the localization, we analyzed the subcellular localization of exogenous Pak1 in HeLa cells with or without Arl4A co-expression. As previously reported, exogenous Pak1 expression was predominantly cytosolic (Manser et al., 1997; Price et al., 1998) (Fig. 2A), whereas Arl4A was enriched at the PM and Golgi (Lin et al., 2011). When Pak1 and Arl4A were co-expressed, Pak1 was highly colocalized with Arl4A in membrane ruffles (Fig. 2A). Image quantification and subcellular fractionation assays further demonstrated subcellular redistribution of Pak1 from the cytoplasm to the PM in cells expressing Arl4A (Fig. 2A, lower panel, and 2B). We also analyzed endogenous Pak1 and found that it was also recruited to the PM when Arl4A was overexpressed (Fig. 2C). Because Pak1 and Pak2 exhibit 97% amino acid identity in the PBD domain, we also examined possible interaction between Arl4A and Pak2. Immunoprecipitation experiments demonstrated that Arl4A associates with Pak2 in HeLa cells (Fig. S1F). We also observed that Arl4A overexpression can induce Pak2 localization to the PM (Fig. S2B). Furthermore, it has been reported that β-PIX-GIT1 complexes with Pak1 and recruits Pak1 to focal adhesion sites (Nayal et al., 2006; Stofega et al., 2004). We thus examined whether Arl4A recruits Pak1 by translocating β-PIX and GIT1 to the PM. We observed that Arl4A co-expression translocated cytosolic β-PIX and GIT1 to Arl4A-enriched membrane protrusion sites (Fig. S2C,D). We further used total internal reflection fluorescence (TIRF) microscopy to show that β-PIX and GIT1 localization exhibited cytosolic distribution patterns with minor focal adhesion-like structures near the cell periphery (Fig. S2E,F). These results suggest that Arl4A overexpression can recruit Pak1 to the PM by inducing subcellular translocation of the β-PIX/GIT1 complex to membrane protrusion sites.
To identify cell lines that express higher levels of endogenous Arl4A, we examined 21 different human cancer cell lines and found that cervical cancer C33A cells exhibited the highest level of Arl4A protein (Fig. S3A). Protein levels of Arl4A, Arl4C and Pak1 were also higher in C33A cells than in HeLa cells (Fig. S3B). Further, Arl4A knockdown in HeLa or C33A cells resulted in a significant reduction of fibronectin-induced Pak1 localization signals at the PM (Fig. 2D,E). Subcellular fractionation analyses also demonstrated a decrease of fibronectin-induced membrane-associated Pak1 in Arl4A-knockdown C33A cells (Fig. 2F). These results collectively support a role of Arl4A in modulation of Pak1 localization at the PM.
Pak1 recruits cytoplasmic myristoylation-deficient Arl4A to the plasma membrane
To test whether Pak1 also affects the PM localization of Arl4A, we co-expressed Pak1 and different forms of Arl4A in HeLa cells and examined their location by immunofluorescence. Consistent with our previous findings (Lin et al., 2011), Arl4A-WT and -Q79L were localized to the PM, whereas Arl4A-T34N and N-terminal myristoylation-defective Arl4A-G2A mainly remained in the cytosol (Fig. 3A). Additionally, Arl4A-WT and -Q79L, but not nucleotide-binding defective Arl4A-T34N, could enrich Pak1 in the PM (Fig. 3B). Interestingly, cytoplasmic Arl4A-G2A also colocalized with Pak1 on the membrane ruffle structure (Fig. 3B). Furthermore, the fluorescence intensity of Arl4A-WT, -Q79L and -G2A on the PM was significantly increased in Pak1-expressing cells (Fig. 3C), suggesting that Pak1 can recruit Arl4A to the PM. We further examined whether Arl4A-G2A can interact with Pak1-PBD using a GST pull-down assay and found that the amount of Arl4A-G2A associated with Pak1-PBD was comparable to that of the GTP-bound forms of Arl4A (Fig. 3D). Pak1 is known to be translocated to the PM and activated in cells upon fibronectin treatment (Symons, 2000) (Fig. S4A). To examine whether fibronectin-induced active Pak1 could recruit Arl4A to the PM, we observed that more Arl4A was localized on the PM in HeLa cells after fibronectin treatment (Fig. S4B). Next, we evaluated Arl4A localization in Pak1-knockdown HeLa cells upon fibronectin treatment and observed that the PM localization of Arl4A decreased in Pak1-depleted cells (Fig. 3E). The efficacy of siRNA-mediated Pak1 knockdown was verified by immunoblot (Fig. 3F). These results indicate that active Pak1 could recruit Arl4A to the PM, suggesting that Arl4A and Pak1 may cooperatively recruit each other to activate Pak1 on the PM.
Cooperative recruitment of Arl4A and Pak1 to the plasma membrane is independent of the Pak1-Rac1 interaction and Pak1 kinase activity
Previous studies reported that both Rac1 activation and Pak1 kinase activity are essential for the recruitment of Pak1 to the PM (Kimura et al., 2006; Manser et al., 1997; Sells et al., 2000; Symons, 2000). To test whether the formation of the Arl4A-Pak1 cooperative recruitment loop requires membrane Rac1 stimulation and Pak1 kinase activity (Sells et al., 1997), we co-expressed Arl4A-WT with HA-tagged constructs, including Pak1-H83L/H86L (Rac1-binding-defective mutant) or Pak1-K299R (Pak1 kinase-dead mutant) in HeLa cells. Both Pak1-H83L/H86L and Pak1-K299R were localized to the cytoplasm when transfected alone (Fig. 4A); however, both were recruited to the PM upon Arl4A expression (Fig. 4B, left panel). We also found that Pak1-H83L/H86L and Pak1-K299R could increase the PM localization of Arl4A-G2A (Fig. 4B, right panel). Pak1-CAAX, a constitutively active form with a prenylation motif (Manser et al., 1997), increased recruitment of Arl4A-WT and Arl4A-G2A to the PM (Fig. 4B). These results suggest that cooperative recruitment of Arl4A and Pak1 to the PM does not require Pak1-Rac1 interaction or Pak1 kinase activity.
Arl4A-Pak1 interaction is required for Pak1-mediated Arl4A membrane targeting
Previous data showed that the key region for Pak1-Arl4A binding is located in the C terminus of the Pak1-PDB (83-127 a.a.) (Fig. 1D,F). To define further the amino acids responsible for the Arl4A-Pak1 interaction, we performed alanine scanning between amino acids 102-127 within the Pak1-PBD (Fig. 5A). We found that alanine substitutions between amino acids 102-106 of Pak1 (Pak1-PBD-m5) diminished the interaction with Arl4A in both in vitro binding and pull-down assays (Fig. 5B,C). To ensure that the Pak1-PBD-m5 mutant did not lose intrinsic binding to other well-identified interacting proteins, constitutively active Rac1 and Cdc42 were transfected into HeLa cells and then lysates were incubated with Pak1-WT and Pak1-m5 (Fig. 5D). Lower amounts of Rac-G12V bound to Pak1-m5 than to Pak1-WT; however, Cdc42-Q61L bound equally to Pak1-WT and Pak1-m5, similar to both PIX and GIT (Fig. S4C). Overall, the interaction defect of Pak1-PBD-m5 was more specific to Arl4A (Fig. 5D). We next confirmed that Pak1-PBD-m5 decreased complex formation with Arl4A using a co-precipitation assay (Fig. 5E). To confirm whether Arl4A Pak1 recruitment is interaction dependent, Pak1-m5 and different forms of Arl4A were co-expressed in HeLa cells. Surprisingly, the results showed that Pak1-m5 could still be recruited to the PM by WT Arl4A (Fig. 5F), indicating that recruitment of Pak1 to the PM by Arl4A is interaction independent. In contrast, myristoylation-defective Arl4A-G2A was not recruited to the PM by Arl4A binding-deficient Pak1-m5 at the membrane ruffle, suggesting that the direct interaction is required for Pak1 to recruit Arl4A to the PM.
Arl4A recruitment of Pak1 to the plasma membrane is independent of β-PIX and PIP3 binding
Our results showed that recruitment of Pak1 to the PM by Arl4A is interaction independent (Fig. 5), which supports the possibility that Arl4A regulates the factors or provides the niche required for Pak1 to translocate to the PM. Because the Pak1-PIX interaction is known to be important for Pak1 PM localization (Chan et al., 2008; Manser et al., 1994), we further examined whether Arl4A recruits Pak1 to the PM through PIX. We observed that Arl4A still colocalized with Pak1 on the membrane ruffle in PIX-knockdown cells (Fig. 6A). siRNA-mediated knockdown of PIX was verified by immunoblot (Fig. 6B). These data indicate that Rac1 and PIX are not involved in the Arl4A-Pak1 feedback loop. In other words, Arl4A-induced Pak1 membrane targeting differs from the fibronectin-Rac1-Pak1 pathway.
A previous study showed that Pak1 membrane localization is mediated by binding of the Nck adaptor protein, which is sufficient to stimulate its kinase activity rather than requiring Cdc42/Rac1 GTPases (Bokoch et al., 1998). The N-terminal 82 amino acids of Pak1 were shown to be important for binding Nck1, Cdc42 and Rac1 (Bokoch et al., 1996). In this study, we observed that Nck-binding-defective Pak1 mutants (HA-Pak1-P13A and HA-Pak1-d82N) were still recruited by Arl4A to the PM, although to a lesser extent than Pak1-WT (Fig. 6C). This result suggests that Nck is partially involved in Arl4A-induced Pak1 membrane localization. In addition, it has been suggested that the phosphoinositide phosphatidylinositol (3,4,5)-trisphosphate (PIP3) is essential for both Pak1 activation and PM recruitment in response to extracellular signals (Burbelo et al., 1995; Chan et al., 2008). We used the PI3K inhibitor wortmannin to block PIP3 generation and observed that EGF-stimulated plasma translocation of Akt-PH-GFP was inhibited in wortmannin-treated cells (Fig. 6D, upper panel). However, wortmannin did not affect Arl4A-induced Pak1 PM localization (Fig. 6D, lower panels). A previous study showed that Nck bound to the Robo1 receptor recruits Pak to specific sites at the growth cone membrane (Fan et al., 2003). To test whether Arl4A-induced Pak1 membrane localization occurs through the Robo1-Nck-Pak1 interaction, we used shRNA to knockdown Robo1 (Chiang et al., 2019), but did not observe impaired Arl4A-induced Pak1 PM localization (Fig. S4D). Thus, we concluded that Arl4A recruits Pak1 to the PM independently of PIX and PIP3 binding, but is partially affected by Nck.
Pak1 kinase activity is crucial for Arl4A-induced cell migration
Our previous study suggests that Arl4A is a putative regulator of cell migration (Chiang et al., 2019). To examine whether Pak1 plays a role in Arl4A-induced cell migration, we confirmed that Arl4A is involved in cell migration using a Boyden chamber transwell assay. We found that exogenous expression of Arl4A promoted HeLa cell migration upon fibronectin stimulation (Fig. S5A). We then used two individual siRNAs to knockdown endogenous Arl4A in C33A cells and evaluated migration using fibronectin as an attractant. Arl4A depletion in C33A cells decreased cell migration (Fig. S5B), which could be rescued by transfection with RNAi-resistant Arl4A (Fig. S5C). Thus, our results demonstrated the positive role of exogenous and endogenous levels of Arl4A in cell migration.
To determine whether Pak1 is involved in Arl4A-mediated cell migration, migration was analyzed in Pak1-knockdown HeLa cells. We observed that Arl4A-induced cell migration was impaired in Pak1-knockdown cells, which could be restored by expression of wild-type Pak1 (Pak1-WT-Res) but not a kinase-dead mutant (Pak1-K299R-Res) (Fig. 7A). These data indicate that Pak1 kinase activity contributes to Arl4A-induced cell migration.
Cooperative recruitment of Arl4A and Pak1 to the PM plays an important role in Arl4A-induced cell migration
To determine whether cooperative recruitment of Arl4A and Pak1 to the PM is important for cell migration, we transiently transfected HeLa cells with Arl4A (WT or G2A) and Pak1 (WT or m5). We found that compared to Pak1-WT, Pak1-m5 had a significantly lower ability to induce cell migration (Fig. 7B). Co-expression of Arl4A-WT and Pak1-WT triggered stronger migration than the expression of either construct alone. However, unlike Pak1-WT, Pak1-m5 failed to enhance cell migration upon co-expression with Arl4A. These results indicate that the Pak1-Arl4A interaction is important for Arl4A-induced cell migration. As expected, the membrane-targeting mutant Arl4A-G2A by itself could not induce cell migration. However, when co-expressed with Pak1-WT, but not Pak1-m5, Arl4A-G2A could induce cell migration similar to Arl4A-WT. This finding indicates that Pak1 can recruit Arl4A-G2A to the PM and confer Arl4A translocation to activate downstream signaling for cell migration. Consistently, expression of Pak1-WT, but not Pak1-m5, restored the cell migration defect in Pak1-knockdown C33A cells (Fig. S5D). This further supports a crucial role of Arl4A-Pak1 cooperativity in Arl4A-induced migration.
We showed that Rac1 is not involved in the Arl4A-Pak1 cooperative recruitment feedback loop, thus we investigated whether Arl4A-Pak1 promotes migration in a Rac1-dependent manner. Both Pak1-WT and Pak1-H83L/H86L enhanced HeLa cell migration and their co-expression with Arl4A-WT evoked even more migration of HeLa cells (Fig. 7C). As we have confirmed in vitro that Arl4A and Rac1-binding-deficient Pak1-H83L/H86L can interact (Fig. 1F), these results suggest that Arl4A-Pak1-induced cell migration is independent of Rac1-Pak1 interaction. Together, our data support the idea that the Arl4A-Pak1 interaction and cooperative recruitment to the PM play important roles in cell migration.
Arl4A-Pak1-induced cell migration differs from Rac1-Pak1 signaling
Previous studies showed that Pak1 phosphorylation is related to its kinase activity or localization (Sells et al., 2000). In particular, Rac1-activated Pak1 is highly phosphorylated on several specific sites (Parrini et al., 2002; Pirruccello et al., 2006). Serine 144 is a known marker of activated Pak1, and its phosphorylation protects its conformation from autoinhibition (Buchwald et al., 2001; Chong et al., 2001). Phosphorylation on Thr423 during the activation process is crucial for substrate accessibility and catalytic activity (Chong et al., 2001; Zenke et al., 1999). We examined whether Arl4A, similar to Rac1, can trigger Pak1 translocation to the PM and facilitate Pak1 phosphorylation. We co-expressed Pak1 with Arl4A or Rac-G12V in HeLa cells and examined phosphorylation at Thr423, Ser199/204 and Ser144 using phospho-specific antibodies. Pak1-CAAX served as a positive control for highly phosphorylated Pak1. We found that both Rac-G12V expression and fibronectin treatment increased Pak1 phosphorylation at these sites. However, Arl4A overexpression did not appear to affect phosphorylation at any of these sites (Fig. 8A), suggesting that the action of Arl4A on Pak1 differs from fibronectin or Rac1 signaling. We further examined whether Arl4A depletion affects fibronectin-induced phosphorylation of Pak1. By quantifying the amount of fibronectin-induced Pak1 phosphorylation, we found that Arl4A depletion decreased Pak1 phosphorylation levels in HeLa and C33A cells (Fig. S6A,B). This result suggests that Arl4A is involved in fibronectin-induced Pak1 activation. Considering that Pak1 kinase activity is required for Arl4A-induced migration (Fig. 7A), we performed in vitro kinase assays using a GST-Paktide substrate to confirm the impact of Arl4A on Pak1 activation. We found that, unlike recombinant GTPγs-loaded Rac1, recombinant Arl4A-QL could not activate Pak1 to phosphorylate the GST-Paktide substrate (Fig. 8B). However, we detected activation of Pak1 kinase activity in vitro using HA-tagged Pak1 isolated from cells co-expressing Arl4A-WT and Pak1 (Fig. 8C). Cells co-transfected with Rac1-G12V and Pak1-WT were used as a positive control, as reported previously (Strochlic et al., 2010). These results suggest that Arl4A indirectly activates Pak1 through an unidentified kinase associated with Arl4A.
Pak is known to associate with activated Rac1 and Cdc42 and mediate actin organization, motility, adhesion and gene transcription through multiple regulatory targets (Chiang and Jin, 2014). Here, we report that the small GTPase Arl4A can mediate membrane targeting and activation of Pak1 through a cooperative recruitment loop. We showed that Arl4A directly interacts with Pak1 via the Pak-PBD domain and induces Pak1 membrane targeting in a Cdc42/Rac1-independent manner. Interestingly, Pak1 also binds to and recruits Arl4A to the PM. We further showed that the Arl4A-Pak1 interaction is crucial for Pak1-dependent cell migration. Thus, we conclude that Arl4A and Pak1 recruit one another to the PM in an interaction-dependent manner, thereby modulating cell motility.
A previous study of high-resolution structures of Pak1 showed that the N-terminal regulatory region of Pak1 contains a CRIB domain (a.a. 70-95) that partially overlaps with the AID (a.a. 83-149) (Lei et al., 2000). Binding of GTP-loaded Cdc42 and Rac1 to the CRIB motif causes an AID conformational change resulting in disinhibition of the catalytic domain, dissociation of the dimer, and phosphorylation of the regulatory region and activation loop. We showed that Arl4A, but not Rac1, directly interacts with the AID (C-terminal end of Pak1-PDB) (Fig. 1F). Arl4A also interacts with Cdc42/Rac1-binding-deficient Pak1 (H83L/H86L). These results indicate that the Arl4A-binding site on Pak1 is unique from the core Cdc42/Rac1-binding motif and suggest that Cdc42/Rac1 binding is not required for the Arl4A-Pak1 interaction. Furthermore, our results showed that the Pak1-m5 mutant site (a.a.102-106), within a helical stretch in the linker between the CRIB and autoinhibition segments, is crucial for Arl4A binding. We speculate that Arl4A binding to the AID may release Pak1 from autoinhibition and induce kinase activity.
The localization of Pak1 to the PM is indispensable for its kinase activation (del Pozo et al., 2000). Although Pak1 mainly resides in the cytosol in resting cells, upon stimulation by growth factors, Pak1 translocates to membrane ruffles and lamellipodia. Membrane targeting leads to a conformational change in Pak1 (Parrini et al., 2009), and membrane attachment sequentially leads to the activation of Pak kinase activity (Bokoch et al., 1998; Lu et al., 1997) through Cdc42/Rac GTPases and/or lipid stimulation (Chan et al., 2008). Consistent with these reports, we found that the majority of the Pak1 protein appeared to be located in the cytosol in HeLa cells and that Arl4A can induce Pak1 localization to the PM. Interestingly, Arl4A can recruit Arl4A-binding-defective Pak1 and Cdc42/Rac1-binding-deficient Pak1 mutants to the PM. These data indicate that the Arl4A-induced translocation of Pak1 to the PM is independent of the Arl4A-Pak1 interaction and Cdc42/Rac binding. Notably, we also demonstrated that Arl4A recruitment of Pak1 to the PM is independent of PIX and PIP3 binding. Thus, our data are consistent with previous studies showing that Pak1 membrane localization is sufficient to activate Pak1, and that subcellular localization, but not the availability of GTP-loaded Cdc42/Rac GTPases, can limit Pak1 activation. Whether Arl4A-mediated factor(s) in the PM play roles in recruiting Pak1 remains unclear and needs to be investigated.
Our data showed that Pak1, but not Arl4A-binding-defective Pak1, recruits cytoplasmic myristoylation-deficient Arl4A-G2A to the PM, indicating that the Arl4A-Pak1 interaction is required for Pak1-mediated recruitment of Arl4A to the PM. Pak1 is known to be translocated to the PM and activated in cells upon fibronectin treatment (Symons, 2000). We also found that Arl4A PM localization is decreased in Pak1-depleted cells upon fibronectin treatment. Although overexpression of Pak1 can recruit cytosolic Arl4A-G2A to the PM, we found that the level of activated endogenous Pak1 in fibronectin-treated HeLa cells is not sufficient to recruit cytosolic Arl4A-G2A to the PM. We speculated that overexpressing Pak1 could initiate a low level of Pak1 activation on the membrane, which triggers the Pak1-Arl4A interaction-dependent feedback recruitment loop. Another interpretation is that most of the activated endogenous Pak1 under fibronectin-treated conditions may form a complex with other proteins, resulting in a decreased amount of active Pak1 interacting with Arl4A-G2A. The effects of cooperative recruitment of Arl4A and Pak1 are clearly crucial for the intrinsically transient properties of cellular control circuits, such as excitability and oscillations. Together, our results demonstrate that Arl4A and Pak1 form a cooperative recruitment loop for the translocation of Pak1 to the PM.
Filamin A, an actin-binding protein, is a substrate for Pak1 at Ser152 (Vadlamudi et al., 2002). Filamin A interacts with Pak1 through residues 52-132 and stimulates the activity and function of Pak1 during cytoskeletal remodeling. Merlin (Nf2), a FERM-containing protein, is also a substrate of Pak1. A direct interaction between Merlin and Pak1 has been identified at residues 70-143 (Kissil et al., 2003). Merlin reduces the interaction of Rac1 and paxillin from Pak1 and inhibits Pak1's kinase activity. These effects imply that proteins that interact with Pak1 through the PBD domain could generate positive or negative feedback to affect the regulation of Pak1 activity and function. Consistent with this notion, our data showed that fibronectin-induced activation of Pak1 enhances Arl4A PM localization, which is decreased in Pak1-depleted cells. It will be interesting to explore whether Pak1 can regulate Arl4A activation through its kinase activity.
Pak1 is known to be autophosphorylated at several sites, including Ser144, Ser199, Ser204 and Thr423, upon activation by Cdc42 or Rac. These sites are essential for releasing Pak1 from the AID and for the maintenance of kinase activity (Chong et al., 2001; Zenke et al., 1999). In particular, phosphorylation at Thr423 strongly correlates with Pak1 activation and the substitution of the acidic residue glutamic acid at this site yields a constitutively active enzyme (Manser et al., 1997; Sells et al., 1997). Although overexpression of Arl4A did not increase the activation loop phosphorylation of Pak1 at Thr423, our in vitro kinase assay showed that Arl4A could promote Pak1 kinase activity. This is surprising and leaves unresolved the question of whether or not Pak1 kinase activity is increased by Arl4A-mediated recruitment. Further work in our laboratory is directed at determining whether Arl4A may mediate non-canonical phosphorylation site(s) of Pak1 that may contribute to Pak1 activity.
Cell migration is a highly regulated event that is initiated by the protrusion of the cell membrane (Le Clainche and Carlier, 2008). Among the members of the Rho GTPase family, Cdc42 reportedly plays a major role in regulating cell polarity, cell migration and actin reorganization (Etienne-Manneville, 2004; Raftopoulou and Hall, 2004; Ridley, 2015). Our previous study showed that Arl4A plays a role in actin cytoskeleton rearrangement via a pathway that stimulates ELMO/DOCK180-induced Rac signaling (Patel et al., 2011). Arl4A can recruit cytohesin-2 to the PM, implicating it in Rac1 activation via signaling through cytohesin-2/Arf6, which functions in actin reorganization and efficient cell migration. Recently, we reported that Arl4A interacts with the transmembrane receptor roundabout (Robo1) to modulate cell migration by promoting Cdc42 activation. In this study, we found that Arl4A interacts with Cdc42/Rac1-binding-deficient Pak1-H83L/H86L to enhance cell migration further. This finding is consistent with previous studies showing that Cdc42/Rac1-binding-deficient Pak1 can induce neurite outgrowth (Daniels et al., 1998; Kiosses et al., 1999; Vadlamudi et al., 2000). An Arl4A-binding-defective Pak1 mutant could not increase Arl4A-induced cell migration, even though Arl4A recruited it to the PM. Thus, the Arl4A-Pak1 interaction is pivotal not only for their cooperative recruitment to the PM but also for promoting Pak1 activity and Arl4A-induced cell migration.
Based on our findings, we propose a model of Arl4A-mediated Pak1 activation (Fig. 8D). In resting cells, Pak1 mainly resides in the cytosol. When Arl4A is in its GTP-bound active form, it anchors to the membrane and triggers a membrane composition change that leads to the recruitment of Pak1 to the PM in a Cdc42/Rac1-binding-independent manner. Next, the membrane-localized active Pak1 recruits more Arl4A via direct interaction, thus forming a cooperative recruitment loop. Although active Arl4A can recruit Arl4A-binding-defective Pak1-m5 to the PM, Pak1-m5 lacks the ability to recruit Arl4A, resulting in a disruption of the cooperative recruitment loop. Through this feedback mechanism, Arl4A facilitates Pak1 membrane targeting and activation, thereby promoting cell migration.
Although Arl4A, Arl4C and Arl4D have high amino acid sequence identities (∼60-70%), the subcellular localization, effectors, and post-translational modifications of the Arl4 isoforms differ. Besides cycling between a GTP-bound and GDP-bound state, Arl4A/C/D also undergo a membrane targeting cycle; however, the precise steps of activation and/or membrane trafficking of Arl4A/C/D remain unclear. Pak1 and Pak2 also have high amino acid sequence identities (76%), but a previous study showed functional differences between the Pak1 and Pak2 isoforms (Coniglio et al., 2008). Interestingly, we also found that Arl4A can associate with Pak2 in HeLa cells and showed that Arl4A can induce Pak2 translocation to the PM. Thus, we propose that Arl4 GTPases serve to organize microdomains by recruiting specific sets of effector proteins, like Pak1 or Pak2, to distinct regions of the PM. Our finding that Arl4A and Pak1 are cooperatively recruited to the PM suggests that Arl4–effector interactions are key to Arl4 isoform localization. Much work needs to be done to understand the determinants of each Arl4 isoform that are recognized by effectors and which effector interactions are most important for a given Arl4 protein.
Mouse Arl4a mRNA expression is developmentally regulated and highly expressed in neurons in the embryonic stage (Lin et al., 2000). Arl4A expression levels are increased in several cancer cells, such as colorectal SW620 cells, which exhibit a more aggressive metastatic phenotype. We also found that the cervical cancer cell line C33A expresses a high level of Arl4A and showed that depletion of Arl4A significantly decreases C33A cell migration. The function of Arl4A in cell migration highlights its potential role in neuronal migration in embryos and embryonic development. Our preliminary data showed that Arl4A activity is reduced when cells are grown in serum starvation conditions, suggesting that Arl4A activation requires certain extrinsic factors. Further exploration is required to identify possible promigratory or environmental cues to reveal the regulation of Arl4A activation and its effect on cell migration.
In conclusion, we discovered an unexpected regulation of Pak1 by Arl4A. Our results suggest that Arl4A and Pak1 cooperatively recruit each other to the PM and modulate cell motility through an interaction-dependent manner. Arl4A and Pak1 both function to regulate cytoskeletal dynamics via crosstalk involving the Rho/Cdc42/Rac1 and Arf GTPase signaling pathways. To perform such crucial roles, it is obvious that Arl4A and Pak1 activity must be highly regulated spatially and temporally. Although both Arl4A and Rac1 can induce membrane ruffling and cell migration, the scenarios underlying the common outcome could be derived from divergent developmental signaling.
MATERIALS AND METHODS
HeLa (CCL-2) and C33A (HTB-31) cells were purchased from the American Type Culture Collection (ATCC) and grown in a humidified incubator with 5% CO2 at 37°C. HeLa cells were maintained in high-glucose Dulbecco's modified Eagle's medium (Hyclone Laboratories) containing sodium bicarbonate supplemented with 10% fetal bovine serum (FBS; Hyclone Laboratories). C33A cells were maintained in RPMI (Gibco) with sodium bicarbonate supplemented with 10% FBS and non-essential amino acids (Gibco).
Transient transfection in mammalian cells
Transient transfections with the indicated siRNAs and plasmids were performed using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's protocol. Subconfluent cells were incubated with DNA:Lipofectamine 2000 complexes (1:2 ratio) in growth medium. All siRNAs were custom synthesized by Dharmacon, Inc. (GE Healthcare). HeLa (6×105) or C33A (1×106) cells were transfected with 20 nM siRNA transfection complexes. Cells were harvested 24 or 48 h after adding the transfection reagent for subsequent analyses. For Robo1 knockdown, HeLa cells were transfected with shRNA pSuper-GFP plasmids using Lipofectamine 2000 reagent. Arl4A knockdown by lentiviral shRNA silencing was performed by using packaging plasmids to produce shRNA lentivirus from pLKO.1-GFP-shRNA vectors (RNAi Core, Academia Sinica, Taiwan). After 24 h of infection, cells were selected with puromycin for 3 days. Green fluorescent protein (GFP) was used as reporter for virus infection and the GFP-positive cells were included to determine infection efficiency. The sequences of the siRNAs and shRNA used in this study are listed in Table S1.
The open reading frame (ORF) of human ARL4A was amplified from a brain cDNA library (Clontech) by polymerase chain reaction (PCR). For the generation of functional defect mutants, a two-step PCR technique was performed using codon-replacement paired-primers. For eukaryotic expression, the ORFs of ARL4A and its mutants were subcloned into pSG5 (Stratagene), pCMV-HA and pcDNA3.1-Myc (Invitrogen) vectors. For prokaryote expression, the ORFs of ARL4A and its mutants were subcloned into the pET32a vector (Novagen). The pSG5-Arl4A siRNA-resistant construct was generated by a two-step PCR technique using site-directed mutation primers. The primer sequences for PCR were as follows: F: 5′-CAAGATTTGAGAAACTCTTTGTCACTGTCAGAAATTGAG-3′; R: 5′-CTCAATTTCTGACAGTGACAAAGAGTTTCTCAAATCTTG-3′.
The ORF of human PAK1 in pCMV6M-HA and pCMV-FLAG were kindly provided by Professor Jau-Song Yu, Graduate Institute of Biomedical Sciences, Chang Gung University, Taoyuan, Taiwan. For the generation of different fragments of PAK1, the following set of primers were subjected to PCR using pCMV-HA-Pak1 as the template: Pak1-1-70 (F: 5′-GAATTCCGATGTCAAATAACGGCCTAGAC-3′; R: 5′-GTCGACTTATTTCTCTTTCTTTTTATTTG-3′); Pak1-70-127 (F: 5′-GGTACCTTAGTGATTGTTCTTTGTTGCCTC-3′; R: 5′-GTCGACCACATCCAGAACAGCCTGCG-3′); Pak1-127-251 (F: 5′-GAATTCATGGTGTTGGAGTTTTACAAC-3′; R: 5′-GTCGACTTACTCATCAGACATTTTAGGCTTC-3′); Pak1-251-545 (F: 5′-GAATTCGAGGAGATCTTGGAGAAATTAC-3′; R: 5′-GGTACCTTAGTGATTGTTCTTTGTTGCCTC-3′); and Pak1-83-149 (F: 5′-GAATTCCACACAATTCATGTCGGTTTTG-3′; R: 5′-GTCGACTTATGACTTATCTGTAAAGCTCATG-3′). The PAK1 fragments were cloned in the bacterial expression vector pGEX4T-1 (Clontech). All sequences were subjected to DNA sequencing for sequence verification. A plasmid expressing Akt-PH-GFP, which contains AKT1 (a.a. 1-164) with the PH domain (a.a. 6-107), was obtained from Dr Z. F. Chang (National Taiwan University, Taipei, Taiwan). For cloning of GST-Paktide, oligonucleotides encoding the peptide sequence GRRRRRSWYWDG were annealed and ligated into pGEX-4T-1 plasmid via EcoRI/XhoI restriction enzyme sites.
SDS-PAGE and immunoblotting
For immunoblot analysis, proteins were separated by SDS-PAGE and transferred to a PVDF membrane (Immobilon P; Millipore). Following transfer, the membrane was blocked with 5% non-fat dry milk/TBST (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM KCl, 0.1% Tween-20) at room temperature for 30 min. The membrane was incubated with a primary antibody diluted in the blocking buffer at 4°C overnight. After extensive washing, the membrane was incubated with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody diluted in blocking buffer at room temperature for 1 h. Blots were washed, developed using an enhanced chemiluminescence system (Immobilon Crescendo, Millipore), and visualized with a digital imaging system (LAS4000; GE Healthcare).
Polyclonal Arl4A antibodies were prepared as previously described (Lin et al., 2000). The polyclonal antibodies obtained from Cell Signaling Technology were as follows: anti-Pak1 (1:1000, #2602S, Lot:7), anti-Pak1-P Thr423 (1:1000, #2601S, Lot:14), anti-Pak1-P Ser199/204 (1:1000, #2605S, Lot:5), anti-Pak1-P Ser144 (1:1000, #2606S, Lot:6) and β-PIX (1:1000, #4515S, Lot:1). The monoclonal antibodies were as follows: anti-Rac1 (1:1000, 05-389, clone 23A8, Lot: DAM 1604948; Millipore), anti-α-tubulin (1:5000, T5168; Sigma-Aldrich), anti-HA (1:1000, SC-7392; Santa Cruz Biotechnology), anti-Myc (1:1000, MMS-150R; Covance), anti-FLAG M2 (1:1000, F-3165; Sigma-Aldrich) and anti-phalloidin-488 (1:500, 8878; Cell Signaling Technology). DAPI (4′,6-diamidino-2-phenylindole) solution was purchased from Millipore (1:5000, S7113). For immunofluorescence, an anti-Pak1 antibody was obtained from Santa Cruz Biotechnology (1:50, SC-881 C-19). HRP-conjugated goat anti-rabbit and anti-mouse immunoglobulin antibodies were purchased from GE Healthcare (1:5000, NA934V and NA931V). Alexa Fluor 594 and 488 anti-rabbit and anti-mouse secondary antibodies were obtained from Invitrogen (1:500; A-11012 for Alexa Fluor 594 rabbit; A-11034 for Alexa Fluor 488 rabbit; A-11001 for Alexa Fluor 488 mouse; A-11032 for Alexa Fluor 594 mouse).
Generation and purification of recombinant proteins
For the production of His-tagged Arl4A and glutathione S-transferase (GST)-fusion Pak1 fragment proteins, expression vectors were transformed into BL21(DE3) cells. After growth on plates overnight, ampicillin-resistant E. coli were inoculated in 100 ml of Luria broth/ampicillin (50 mg/ml) at 37°C and grown to a density of A600=0.7∼0.8. Recombinant protein induction was initiated by addition of 0.25 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) followed by incubation for 1 h at 37°C. E. coli were harvested by centrifugation at 8000 g for 10 min, pellets were suspended in 4 ml of lysis buffer (20 mM Tris-HCl, pH 7.9, 500 mM NaCl, 5 mM imidazole, protease inhibitor cocktail and 100 mg/ml lysozyme), sonicated, and kept on ice for 30 min. After centrifugation for 14,000 rpm (18,000 g) for 20 min, 4 ml of supernatant (soluble form) was incubated with 0.5 ml of Ni2+-NTA agarose (Qiagen, 1018244) or glutathione-Sepharose resins (GE Healthcare) at 4°C for 2 h. Columns were washed three times with washing buffer (20 mM Tris-HCl, pH 7.9, 500 mM NaCl, 25 mM imidazole). The beads of GST-tagged fusion protein were re-suspended in PBS then subjected to GST pull-down assay. His-tagged proteins were eluted in elution buffer containing 200 mM imidazole. Fractions were collected and analyzed by SDS-PAGE and Coomassie Blue staining to identify and evaluate target protein purity and concentration.
In vitro binding assay
Two micrograms of purified GST-tagged proteins bound to glutathione-Sepharose beads were incubated with 2 μg of His-tagged small GTPases in 1 ml of binding buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM DTT, 10% glycerol, 1% Triton X-100, 1 mM NaN3 and 1× protease inhibitor cocktail) for at least 1 h at 4°C. The beads were washed three times with 1 ml of binding buffer. After removing the buffer, the beads were boiled in 20 µl of SDS-PAGE sample buffer for 10 min and then subjected to immunoblot analysis.
GST pull-down assay
Subconfluent cells were transiently transfected with the indicated expression plasmids using Lipofectamine 2000 reagent (Invitrogen). After 16 h, the cells were lysed in lysis buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM DTT, 250 mM sucrose, 10% glycerol, 0.1% Triton X-100, 1 mM NaN3 and protease inhibitor). Cell lysates were centrifuged at 13,000 rpm (15,000 g) for 10 min, and the supernatants were incubated with 20 µg of GST-fusion proteins bound to glutathione-Sepharose beads for 1 h at 4°C with end-over-end rotation. The beads were washed three times with lysis buffer and the pull-down complexes were subjected to immunoblot analysis.
Rac1 activity pull-down
Cells were transfected with the indicated plasmids for 24 h and then lysed in activity pull-down assay buffer (50 mM Tris-HCl, pH 7.5; 150 mM NaCl, 1 mM DTT, 1% NP-40, 5 mM MgCl2, 5% glycerol and 1× protease inhibitor). The samples were then incubated with Pak-PBD-GST beads for 1 h at 4°C. The beads were washed with assay buffer and the bound proteins were subjected to immunoblot analysis.
Subconfluent HeLa cells were co-transfected with pSG5-Arl4A-Q79L-HA and empty vector, pCMV-tag-2C-Pak1-WT, or pCMV-tag2C-Pak1-m5 plasmids using Lipofectamine 2000 reagent (Invitrogen). Twenty-four hours after transfection, cells were lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 10 mM MgCl2, 0.2% NP-40, 10% glycerol and protease inhibitor cocktail) and incubated on ice for 1 h. The lysates were harvested by centrifugation at 13,000 rpm (15,000 g) for 15 min to remove insoluble material. For purification of FLAG-tagged proteins, lysates were incubated with anti-FLAG M2 magnetic beads (Sigma-Aldrich) overnight at 4°C with end-over-end rotation. Beads were washed three times with 500 µl of immunoprecipitation washing buffer (lysis buffer containing 300 mM NaCl) followed by 500 µl of TBS buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl). Native protein:protein complexes were eluted by incubating the magnetic beads with 300 µg/ml FLAG Peptide (Sigma-Aldrich) for 15 min twice. Arl4A-WT-Myc immunoprecipitation was performed using Myc-Trap magnetic beads (Chromotek; Planegg-Martinsried, Germany) according to the manufacturer's protocol. HeLa cell lysates were harvested and then incubated with beads for 1 h at 4°C. The eluted protein complexes were then subjected to immunoblot analysis.
Transfected cells seeded on cover slides were fixed with 4% paraformaldehyde/PBS for 15 min. After washing with PBS, cells were permeabilized with 0.01% Triton X-100/PBS at room temperature for 5 min, then blocked with 0.5% bovine serum albumin in PBS. Primary antibodies were diluted in blocking buffer at the following ratios: anti-Arl4A (1:200), anti-HA (1:200), anti-Myc (1:200), anti-FLAG M2 (1:200). Alexa Fluor 594 and 488 anti-rabbit and anti-mouse secondary antibodies were used at 1:1000. Images were captured using an Axioplan microscope (Carl Zeiss, Inc.) or confocal microscope (Nikon C1). For total internal reflection fluorescence (TIRF) imaging, HeLa cells were transduced to express β-Pix-Myc or FLAG-GIT1 in the presence or absence of Arl4A for 24 h and re-plated onto 35-mm diameter, 14-mm microwell glass-bottom plates (No. P35G-1.5-14-C uncoated coverslip, MatTek Corporation). The next day, cells were fixed and stained with anti-Myc, anti-FLAG and anti-Arl4A antibodies and then monitored by TIRF microscopy (Carl Zeiss Laser TIRF 3).
Transwell migration assay
Transwell migration assays were performed using Boyden transwell chambers with 82-mm polycarbonate membranes (Corning). The membrane was pre-coated on the underside with 10 mg/ml fibronectin in PBS for at least 1 h at 37°C. C33A cells (1×106) or HeLa cells (4×105) were transiently transfected with the indicated plasmids or siRNAs. The transfected cells were serum-starved for 16 h, harvested with trypsin-EDTA (Invitrogen), and suspended in serum-free medium. Cells were counted and 1-2×105 cells were added to the top of each well and allowed to migrate to the underside of the membrane for 13-16 h at 37°C in a 5% CO2 incubator. The lower chamber contained 10% FBS in culture medium. Each assay was performed in duplicate. Cells were fixed with 4% paraformaldehyde/PBS for 30 min at room temperature. Non-migrating cells on the upper membrane surface were removed with a cotton swab, and migratory cells attached to the lower surface were stained with Crystal Violet for 15 min. After washing, the membrane was air-dried and the migrated cells were imaged using a phase-contrast microscope (Eclipse TS-100, Nikon) with an imaging system (DS-5M, Nikon). The number of migrated cells in at least three fields of each well was counted using ImageJ software.
Quantification of membrane-targeting protein
To analyze the fluorescence ratio on the cell membrane, cell fluorescence intensity analysis was performed using efficient and unbiased image analysis approaches (Barry et al., 2015). ImageJ software was used to measure the fluorescence intensities for plasma membrane:cytosol (PM:C) ratio calculations. Using the ‘Make Binary’ command, the image was converted to black and white so that the region of cell body can clearly be distinguished. The ‘Wand’ (tracing) tool was used to mark the cell outline, which can be memorized by Regions of Interest (ROI) Manager. The ‘Erode’ command was used to remove pixels from the edges of the marked cell outline to further define the cytosol region, which can also be memorized by ROI Manager. The PM:C ratio was calculated using the following equation:
PM:C ratio=intensity of the cell body region – intensity of the cytosol region/intensity of the cytosol region.
Pak1 immunoprecipitation kinase assay
The Pak1 immunoprecipitation kinase assay was performed as described previously (Strochlic et al., 2010). Briefly, HeLa cells (10-cm plates) were transfected with HA-tagged Pak1-WT and Arl4A-WT or Rac-G12V for 24 h. Cells were lysed in 1 ml of lysis buffer (25 mM HEPES, pH 7.4, 0.3 M NaCl, 1.5 mM MgCl2, 0.5 mM EGTA, 0.5% Triton X-100, 5% glycerol containing protease inhibitor). Lysates were centrifuged at 15,000 g for 15 min and the resulting supernatants were used to isolate HA-Pak1 with anti-HA-agarose beads (Sigma-Aldrich, A2095) at 4°C for 2 h. HA-Pak1-beads were washed three times with kinase buffer (50 mM HEPES, pH 7.4, 5 mM NaCl, 10 mM MgCl2, 2 mM MnCl2, 1 mM dithiothreitol, 0.05% Triton X-100). Kinase reactions were started by addition of a mix containing 10 µg GST-Paktide and 30 µM cold ATP with 10 µCi of [32P]-γ-ATP at 30°C. To perform Pak kinase assays using recombinant small GTPases, His-tagged Arl4A-Q79L and Rac1-WT were generated and purified from E. coli. Recombinant Rac1-WT was activated by loading GTPγS in 5 mM MgCl2-containing buffer for 10 min at 30°C. His-Arl4A-Q79L and Rac1- GTPγS were pre-incubated with HA-Pak1-WT-beads for 10 min, then the kinase reactions were initiated by adding GST-Paktide and ATP/[32P]-γ-ATP. The reactions were stopped after 30 min by adding sample buffer and analyzed by SDS-PAGE and radiography. Quantification was performed by ImageJ software.
HeLa (1×107) and C33A (2×107) cells were harvested by trypsin and washed thoroughly with PBS. Cell pellets were fractionated using the CNM Compartmental Protein Extraction Kit (Biochain, K3012010). Cell membranes were disrupted to release the nuclei by passing the cells through a 30-gauge needle 15 times. In general, 90-95% of the nuclei were released from the cells as monitored under the microscope. The disrupted cells presented a non-transparent and non-circular cell shape. After continuous extraction, each fractionation was analyzed by immunoblotting using specific antibodies of various cellular compartment markers including cytoplasmic (α-tubulin, 1:5000, T5168, mouse, Sigma-Aldrich), nuclear (HP1α, 1:1000, 05689, mouse, Upstate Biotechnology), and plasma membrane (Na+/K+ ATPase, 1:1000, sc21712, mouse, Santa Cruz Biotechnology).
All data are expressed as the mean±s.d., and P-values were calculated by two-tailed Student's t-test or one-way analysis of variance (ANOVA) followed by Dunnett's post-hoc multiple comparison test. Statistical significance was determined by using Prism 6 software. Significant differences (*P≤0.05; **P≤0.01; ***P≤0.001) are indicated. For each independent in vitro experiment, three biological replicates were included.
We thank Drs Randy Haun, Ya-Wen Liu and Pei-Hsin Huang for their critical review of the manuscript. We also express our deep gratitude to Dr Jau-Song Yu for providing the Pak1 expression plasmids.
Conceptualization: K.-J.C., F.-J.S.L.; Methodology: K.-J.C., T.-C.C., C.-J.Y., F.-J.S.L.; Validation: K.-J.C.; Formal analysis: K.-J.C., T.-C.C.; Investigation: K.-J.C., T.-C.C.; Resources: K.-J.C., T.-C.C., C.-J.Y., F.-J.S.L.; Data curation: K.-J.C., T.-C.C., C.-J.Y.; Writing - original draft: K.-J.C., T.-C.C.; Writing - review & editing: K.-J.C., C.-J.Y., F.-J.S.L.; Visualization: K.-J.C., C.-J.Y., F.-J.S.L.; Supervision: F.-J.S.L.; Project administration: F.-J.S.L.; Funding acquisition: F.-J.S.L.
This work was supported by grants from the National Health Research Institutes (NHRI) in Taiwan (NHRI-EX106-10601B1) and the Center of Precision Medicine from the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan awarded to F.-J.S.L.
The authors declare no competing or financial interests.