Spatial regulation of the cAMP-dependent protein kinase during chemotactic cell migration

Abstract

Historically, the cAMP-dependent protein kinase (PKA) has a paradoxical role in cell motility, having been shown to both facilitate and inhibit actin cytoskeletal dynamics and cell migration. In an effort to understand this dichotomy, we show here that PKA is regulated in subcellular space during cell migration. Immunofluorescence microscopy and biochemical enrichment of pseudopodia showed that type II regulatory subunits of PKA and PKA activity are enriched in protrusive cellular structures formed during chemotaxis. This enrichment correlates with increased phosphorylation of key cytoskeletal substrates for PKA, including the vasodilator-stimulated phosphoprotein (VASP) and the protein tyrosine phosphatase containing a PEST motif. Importantly, inhibition of PKA activity or its ability to interact with A kinase anchoring proteins inhibited the activity of the Rac GTPase within pseudopodia. This effect correlated with both decreased guanine nucleotide exchange factor activity and increased GTPase activating protein activity. Finally, inhibition of PKA anchoring, like inhibition of total PKA activity, inhibited pseudopod formation and chemotactic cell migration. These data demonstrate that spatial regulation of PKA via anchoring is an important facet of normal chemotactic cell movement.

Defining the biochemical mechanisms that control cell migration directly contributes to our understanding of physiologic events in which migration is normally required (e.g., embryonic development, wound healing, and angiogenesis) and those in which it is subverted for pathogenesis (e.g., cancer metastasis) (1). In vivo, the regulation of cell migration is achieved through a complex network of signals arising from cell surface receptors for extracellular matrix, other cells, and soluble factors such as peptide growth factors and bioactive lipids (2). Gradients of these extracellular ligands establish a gradient of engaged receptors on the cell surface that directs localized polymerization of actin and formation of protrusive leading edge structures, such as lamellipodia and filopodia, in the direction of eventual cell migration. Recent evidence indicates that establishing and maintaining cell polarity during migration depends on not only the bulk activity or abundance of cellular components that preside over cytoskeletal dynamics, but also the strict regulation of their activity or abundance in subcellular space (3, 4).

The cAMP-dependent protein kinase (PKA) is extremely promiscuous with its activity (5), with targets in the plasma membrane, cytoplasm, mitochondria, nucleus, and nearly every family of cytoskeletal network, including microtubules, intermediate filaments, and actin microfilaments (5, 6). Historically, PKA plays a dichotomous role in actin cytoskeletal organization and cell migration, exerting both negative (i.e., inhibitory) and positive (i.e., required or enhancing) effects (reviewed extensively in ref. 6). PKA can be both positively and negatively regulated by cell adhesion and during cell spreading (7-10). Furthermore, although some hallmarks of cell migration and cytoskeletal assembly require PKA activity [e.g., activation of Rac (9, 11), Cdc42 (12), and microfilament assembly (10)], others are inhibited by it [e.g., activation of Rho (13) and p21-activated kinase (7), interaction between vasodilator-stimulated phosphoprotein (VASP), and the c-Abl tyrosine kinase (8), actin polymerization (14)]. Finally, cell migration and invasion have been shown to either require PKA activity (9, 15-17) or be inhibited by it (15, 16, 18, 19). A prototypical example of the dichotomous nature of PKA regulation of cell migration is the data showing that αvβ3 integrin-dependent endothelial cell migration is both positively (11) and negatively (18) regulated by PKA. These ostensibly disparate observations indicate that PKA can both inhibit and facilitate cell migration and suggest a significant degree of complexity for regulation of PKA activity during migration.

These positive and negative effects and the number and variety of putative targets through which they might be elicited (6) suggest that there must be a mechanism in place for specifying or focusing PKA activity during cell migration to properly and accurately assign its functions. One hypothesis is that PKA activity might be spatially regulated during cell migration, and, thus, arrant activation or inhibition of PKA would disrupt the spatial specificity of its activity and hamper overall migration. The case for spatial regulation of PKA is strong, because a vast and growing literature indicates that recruitment or enrichment of PKA to specific subcellular regions or structures, through interaction with A kinase anchoring proteins (AKAPs), facilitates PKA-dependent signaling in those regions and is likely required for determining the specificity of the PKA signaling (20). Because of the abundant literature describing AKAP-mediated anchoring of PKA to membranous and cytoskeletal regions (20, 21), we investigated the hypothesis that PKA may be spatially regulated during cell migration. Here, we show that a subset of PKA holoenzyme, as well as PKA activity and phosphorylated substrates, are enriched in protrusive, leading-edge structures formed during chemotactic cell movement. Moreover, inhibition of PKA activity or AKAP-mediated localization prevents directed cell migration.

Experimental Methods

Antibodies and Other Reagents. Detailed descriptions of antibodies and sundry materials can be found in the Supporting Text, which is published as supporting information on the PNAS web site, or can be requested from the corresponding author.

Immunofluorescence. Serum-starved cells were detached and replated (8) on coverslips coated with 10 μg/ml fibronectin. After incubation and treatment as described in the figure legends, cells were fixed in PBS/3.7% formaldehyde for 10 min, permeabilized in PBS/0.5% Triton X-100 for 5 min, blocked in PBS/2% BSA, stained with primary antibodies, phalloidin, Alexa-conjugated secondary antibodies, and/or DAPI (all diluted in block), then mounted on slides by using PermaFluor (Thermo Shandon, Pittsburgh). Cells were visualized through ×40 or ×60 PlanApo objectives on Olympus IX70 microscopes equipped for epifluorescence (with an F-View II charge-coupled device camera controlled by analysis software (Soft Imaging Systems) or confocal microscopy (FluoView 300 system and software).

Pseudopod Preparation and Quantification. These techniques were performed as described in ref. 22. Briefly, cells were replated for 2 h on fibronectin-coated, 3 μm-pore polycarbonate membranes in Costar Transwell inserts. Where indicated, pharmacological agents at the indicated concentrations were added to both chambers for 20 min, then chemoattractant was added to the lower chamber for 1 h. For pseudopod quantification or harvest, inserts were washed in PBS, cell bodies (CBs) were removed from the upper surface, and the pseudopodia (Pd) on the underside were scraped into lysis buffer. Alternatively, Pd were removed, and CB on the upper surface were harvested. Lysates were clarified by centrifugation and protein content quantified by bicinchoninic acid assay (Pierce).

Cell Migration Studies. Cells were cultured as for pseudopod preparations but by using 8-μm-pore Transwell inserts. Cultures were incubated at 37°C for 8 h, then nonmigrating cells were removed from the top chamber with cotton swabs and the remaining cells were fixed for 10 min in 3.7% formaldehyde/0.5% Triton X-100 in PBS, stained with DAPI, mounted, and counted by fluorescence microscopy.

Western Blotting, Immunoprecipitation, Transfections, and PKA Activity Assays. These techniques were performed as described in refs. 7 and 8. VASP tagged with an epitope from a vesicular stomatitis virus protein was expressed under control of the cytomegalovirus promoter as described in ref. 8.

Phosphatase, Rac, GTPase Activating Protein (GAP), and Guanine Nucleotide Exchange Factor (GEF) Activity Assays. The activity of immunoprecipitated protein tyrosine phosphatase containing a PEST motif (PTP-PEST) was assessed by using an in vitro phosphatase assay according to the manufacturer's instructions (Promega). Activation of Rac was determined with a pulldown assay by using a GST fusion with the p21-binding domain of p21-activated kinase as described in refs. 22 and 23. GAP and GEF assays were performed essentially as described in ref. 24. GST-Rac1 (1 μg) was incubated with 20 μCi (1 Ci = 37 GBq) of γ-³²P-GTP (for GAP assays) or α-³²P-GTP (for GEF assays) in nucleotide loading buffer (25 mM Tris, pH 7.5/50 mM NaCl/5 mM EDTA/1 mg/ml BSA/0.1 mM DTT) for 20 min at 25°C. MgCl₂ was added to 25 mM, and the mixture was kept on ice until use. Pd were harvested in lysis buffer (see ref. 24 and Supporting Text), cleared by centrifugation, and 250 μg was mixed with nucleotide-loaded Rac in the presence of 1 mM GTP and incubated at 20°C. Fifty-microliter aliquots were removed at the indicated times and filtered through nitrocellulose. Filters were washed extensively with wash buffer (loading buffer with 20 mM MgCl₂ instead of EDTA) and bound radioactivity quantified by scintillation counting.

Results

Type-II PKA Regulatory Subunit Localizes to Protrusive Structures During Migration. Both inhibition and hyperactivation of PKA appear to have deleterious effects on chemotactic cell migration (reviewed extensively in ref. 6 and demonstrated in Fig. 6, which is published as supporting information on the PNAS web site). These disparate observations suggest that PKA activity is indeed required for successful chemotaxis, but in a tightly regulated manner, and that arrant activation or inhibition perturbs this regulation and impedes migration. To investigate whether the role of PKA in cell migration might involve spatial regulation of the kinase, we examined the localization of PKA subunits in migrating cells by immunofluorescence microscopy. We focused our efforts on the α isoform of the PKA type II regulatory subunit (RII), as AKAP-mediated localization of PKA occurs most frequently through type II, rather than type I, subunits (20). PKA RII was enriched in protrusive cellular structures formed in response to platelet-derived growth factor (PDGF) or lysophosphatidic acid (LPA) (Fig. 1A). Although previous work demonstrated localization of RIIα to actin-rich dorsal ruffles induced by PDGF (25), the current data demonstrate a polarized enrichment of RIIα within putative leading edge structures. This enrichment suggests that the function of PKA during cell migration might be modulated, in part, by localization of PKA RIIα to protrusive, lamellipodial structures formed during chemotaxis.

Fig. 1.

PKA subunits and activity are enriched in Pd. (A) REF52 or WI38 fibroblasts were plated onto fibronectin-coated coverslips for 90 min, then stimulated for 1 h with 10 ng/ml PDGF (REF52) (Top and Middle) or 100 ng/ml LPA (WI38) (Bottom). Cells were fixed and processed for immunofluorescence by using antibodies against the indicated proteins and fluorescent phalloidin to stain F-actin as indicated, then examined by confocal microscopy. (Middle) Enlargements are shown of the area indicated by the square in Upper Left. (Scale bar: 10 μm.) (B) REF52 cells were cultured for pseudopod formation as described in Experimental Methods, then fixed, stained with Alexa 488-phalloidin, and imaged by confocal microscopy. (Left) The diagram shows the preparation and orients the reader to Center, which shows a 3D reconstruction of a single cell prepared and imaged as described. The top of the cell appears flat because the image stack stops approximately halfway through the cell nucleus. (Right) The same cell is shown, rotated upward ≈30° about the horizontal axis to illustrate the Pd's fine structure. Movie 1, which is published as supporting information on the PNAS web site, shows a 180° rotation of this image. (C and D) Twenty micrograms of protein prepared from CB from unstimulated cells (Un), or from CB and Pd formed in response to PDGF, EGF, and PDGF (10 ng/ml each; E+P), or LPA, were immunoblotted with the indicated antibodies. (E) PKA activity was determined from equal amounts of protein prepared from unstimulated (Ctrl) REF52 cells or CB (open bars) and Pd (filled bars) formed in response to LPA or PDGF. Data (as cpm per μg of lysate) represent means ± SD for three independent Pd preparations processed simultaneously for kinase activity.

PKA Subunits and Activity are Enriched in Pd. To further explore localization of PKA during chemotaxis, we adapted a recently described, elegant subcellular fractionation technique that enriches for the growing Pd formed by migrating cells (22). In this technique, cells plated on one side of a thin, porous polymer filter coated with fibronectin are stimulated to migrate toward a chemoattractant on the opposite side and, thus, send one or more Pd through the filter pores (Fig. 1B). The filter provides a physical divider between the CB and Pd and, thus, allows separation and biochemical analysis of each fraction.

Equal amounts of CB and Pd protein were analyzed by Western blotting with antibodies against various PKA subunits. In agreement with immunofluorescence data, the relative level of PKA RIIα was greater in the Pd than in the CB (Fig. 1 C and D). The localization of the PKA regulatory subunit type Iα was more variable and cell type-specific, showing slight enrichment in the CB of REF52 (Fig. 1C) and WI38 fibroblasts (data not shown), but a slight (≈50%) enrichment in Pd of NIH 3T3 fibroblasts (Fig. 1C) and a greater enrichment in A7r5 smooth muscle cells and some epithelial cell lines (data not shown). Interestingly, although there was no obvious enrichment of PKA catalytic subunit (PKA-C) in either subcellular compartment in REF52 cells (Fig. 1D) or in NIH 3T3 cells (data not shown), actual PKA activity in Pd was significantly higher than in CB (Fig. 1E). Thus, the enrichment of PKA RII in lamellipodia/Pd observed by immunofluorescence and immunoblotting correlates with increased PKA activity in these protrusive structures.

The increased activity of PKA in Pd suggests that the level of phosphorylation of PKA substrates might be higher in Pd than in CB. Indeed, exploratory analyses of CB and Pd fractions, separated by one- and two-dimensional electrophoresis, by staining with a phosphoprotein-specific stain (ProQ Diamond, Invitrogen) or by immunoblotting with a phospho-PKA-substrate antibody, showed a large number of reactive bands or spots present in Pd and not in CB (data not shown). To explore this increased phosphorylation in a more specific manner, we examined the phosphorylation state of two recently described PKA targets with particular relevance to cytoskeletal regulation and cell migration.

VASP Phosphorylation Is Enriched and VASP-Abl Interaction Is Disrupted in Pd. VASP is the founding member of a family of proteins with increasingly prominent and complex roles in actin cytoskeletal dynamics and cell migration (26). All mammalian VASP proteins are substrates for PKA (27), and phosphorylation has been shown to be crucial for regulating their function (26), particularly their role in regulating cell migration (28, 29). We therefore analyzed the relative phosphorylation of VASP in Pd and CB. Phosphospecific antibody reactivity and electrophoretic mobility shift both showed that VASP phosphorylation was significantly higher in Pd than in CB (Fig. 2 A and B), consistent with the increased level of PKA activity in Pd.

Fig. 2.

Localized PKA activity correlates with enrichment of PKA-phosphorylated VASP. (A) REF52 CB and Pd, formed in response to PDGF, EGF, or LPA, were blotted for PKA-phosphorylated VASP (pVASP; p157-VASP) or the retinoblastoma protein (Rb), which resides in the nucleus and is therefore present only in CB. (B) CB from unstimulated NIH 3T3 cells (Un), or CB and Pd formed in response to PDGF or EGF, were blotted with an antibody against VASP. (C) COS7 cells transfected with epitope-tagged VASP (VSV-VASP) were either treated with 25 μM Fsk for 20 min or cultured for EGF-induced pseudopod formation. Lysates from Fsk-treated cells (Fsk) and from CB and Pd were immunoprecipitated with antivesicular stomatitis virus antibody. Precipitates and whole cell extract (wce) from transfected cells were separated by SDS/PAGE and blotted with the indicated antibodies. A low-percentage gel was used to collapse the phosphorylation-sensitive electrophoretic profile of VASP (evident in B) to a single band for easier confirmation of equal loading. Immunoblotting unprecipitated CB and Pd lysates with anti-Abl antibodies confirmed the presence of c-Abl in both fractions (Lower).

An immediate consequence of PKA-mediated VASP phosphorylation is inhibition of VASP interaction with the c-Abl tyrosine kinase (8). We therefore investigated whether partitioning of PKA activity and VASP phosphorylation correlated with a spatial regulation of VASP-Abl interaction in migrating cells. To compensate for the small amount of protein harvested from Pd (22) and the requirement for a relatively large amount of material to observe VASP-Abl interaction (8), we used COS-7 cells transfected with an epitope-tagged version of wild-type VASP and prepared anti-epitope immunoprecipitates from Pd and CB extracts. As expected, coprecipitation of c-Abl with VASP was much lower in Pd than in CB (Fig. 2C), consistent with increased level of phosphorylated VASP in this fraction (Fig. 2 A and B).

PKA Phosphorylation of PTP-PEST is Enriched in Pd. PTP-PEST, a protein tyrosine phosphatase and an increasingly important regulator of cell migration (30, 31), can be directly phosphorylated and inhibited by PKA (32). To assess whether localization of PKA activity correlated with spatial regulation of PTP-PEST phosphorylation, we immunoprecipitated PTP-PEST from CB and Pd extracts and immunoblotted the precipitates with an antibody specific for PKA-phosphorylated substrates. Although phosphorylation of PTP-PEST was barely detectable in stably adherent but nonmotile cells on tissue culture plates, induction of chemotaxis resulted in PKA-mediated phosphorylation PTP-PEST specifically in Pd (Fig. 3A).

Fig. 3.

Localized regulation of PTP-PEST by PKA. (A and B) PTP-PEST was immunoprecipitated from extracts of REF52 cells stably adherent to tissue culture plastic (TC), CB, and Pd formed in response to PDGF, or Pd treated with mPKI or StHt31 (as described in Fig. 2), then separated by SDS/PAGE and blotted with anti-PTP-PEST and anti-phospho-PKA substrate (p-PKA sub) antibodies. (C) PTP-PEST was immunoprecipitated from PDGF-stimulated Pd (Ctrl), or Pd treated with mPKI or StHt31 as described above, and subjected to an in vitro phosphatase assay (see Experimental Methods) in which absorbance at 630 nm is proportional to released phosphate and, thus, phosphatase activity. Data represent means ± SD for three independent Pd preparations immunoprecipitated and processed simultaneously for phosphatase activity.

The enrichment of PKA activity within Pd suggested that blocking PKA activity or anchoring within these structures should affect substrate phosphorylation. To test this hypothesis, cell-permeable peptides from the PKA inhibitor protein (myristoylated PKI peptide; mPKI) or the AKAP Ht31 (stearated Ht31 peptide; StHt31) were used to inhibit PKA activity or anchoring, respectively. Specifically, Pd were allowed to form as described, then PKA activity or anchoring was locally inhibited within these structures by the addition of inhibitors to the underside of the filter. Surprisingly, inhibition of neither PKA activity nor anchoring had any significant effect on the phosphorylation of VASP within Pd (Fig. 7, which is published as supporting information on the PNAS web site), but both interventions significantly inhibited the phosphorylation of PTP-PEST (Fig. 4B). As phosphorylation by PKA inhibits PTP-PEST, decreased phosphorylation should correlate with increased phosphatase activity. To test this hypothesis, phosphatase assays were performed on PTP-PEST immunoprecipitated from PDGF-stimulated Pd that were left untreated, or treated with mPKI or StHt31. As expected, inhibition of PKA activity or anchoring within Pd resulted in a significant increase in PTP-PEST activity (Fig. 4C). The importance of this effect is underscored by the observation that inhibition of PKA activity or anchoring dramatically inhibited the migration-induced tyrosine phosphorylation of p130Cas, a key PTP-PEST substrate (33) and important regulator of cell migration (34) (Fig. 8, which is published as supporting information on the PNAS web site). These data show that localized PKA activity regulates PTP-PEST activity during Pd formation.

Fig. 4.

PKA activity controls Rac by regulating Rac GEF and Rac GAP activities within Pd. (A and B) NIH 3T3 cells were cultured as in Fig. 2 D and E and CB and Pd extracts were subject to a pulldown assay by using a GST-p21-binding domain fusion protein to isolate the active form of Rac. A portion of the extracts were collected before pulldown and immunoblotted directly to determine total Rac levels. The bar graphs depict the average ratios of active to total Rac, ± SD, determined from three separate experiments by densitometry of the immunoblotted bands. (C and D) Control- or inhibitor-treated Pd from PDGF-stimulated NIH 3T3 cells were harvested and incubated with purified, recombinant GST-Rac1 loaded with α-³²P-GTP (C) or γ-³²P-GTP (D) for the indicated times to measure Rac GEF or GAP activity, respectively. The data are presented as the percent of radioactivity remaining bound to Rac1 in the absence of extract and represent means ± SD for four independent Pd preparations processed simultaneously for GEF or GAP activity. Note that the y axis in C does not go to zero.

An important potential downstream target for PKA-mediated regulation of PTP-PEST is the Rac GTPase, a sine qua non regulator of lamellipodia formation and cell migration whose activity has been shown to localize to the leading edge (35) and Pd (22), require PKA (9), and be downstream of both PTP-PEST and p130Cas (22, 31). To investigate the contribution of PKA activity and anchoring to localized regulation of Rac, Rac activity was assayed from CB and Pd extracts in the absence or presence of mPKI or StHt31. As reported in ref. 22, activation of Rac was almost entirely relegated to the Pd fraction (Fig. 4 A and B). Importantly, Rac activity within Pd was significantly inhibited by inhibition of either PKA activity or anchoring (Fig. 4 A and B).

Inhibition of Rac activity may be due to decreased activity of Rac-specific GEFs or increased activity of Rac-specific GAPs. To investigate these possibilities, GEF and GAP activities in control-, mPKI-, and StHt31-treated Pd were measured by using recombinant Rac. Interestingly, inhibition of PKA activity or anchoring resulted in both decreased Rac-GEF and increased Rac-GAP activities (Fig. 4 C and D). These critically important data show that localized PKA activity can preside over Rac, a crucial regulator of leading edge events and cell migration, through modulation of GEF and GAP activities.

PKA Activity and Anchoring Are Required for Pseudopod Stability, Formation, and Chemotaxis. The dramatic effect of the mPKI and StHt31 peptides on Rac activity, along with the published observation that Rac activity is greatly diminished in retracting Pd (22), suggested that inhibition of PKA activity or anchoring might result in Pd destabilization and retraction. Indeed, anecdotal evidence from optimizing conditions for the experiments in Figs. 3 B and C and 4 suggested that prolonged incubation with inhibitors severely diminished the amount of Pd protein recovered. To formally test this hypothesis, Pd were induced then treated with mPKI or StHt31 for increasing periods of time before quantification. The amount of Pd material did indeed diminish over time in the presence of mPKI or StHt31 (Fig. 5A), indicating that prolonged inhibition of either PKA activity or anchoring leads to destabilization and/or retraction of Pd.

Fig. 5.

PKA activity and anchoring are required for pseudopod stability and formation and for chemotaxis. (A) REF52 cells were cultured for Pd formation toward PDGF for 1 h. PBS (Ctrl), 20 μM mPKI, or 50 μM StHt31 was added to the filter undersides and, at the indicated times, pseudopod formation was quantified by measuring the amount of pseudopod protein by using a bicinchoninic acid assay. Note that the y axis does not go to zero. (B and C) REF52 cells were cultured for Pd formation (B) or migration (C) by adding the indicated concentrations of StHt31, StHt31P, mPKI, cytochalasin D, or Fsk (in μM) to the cells 20 min before addition of growth factor. Pd formation was measured as above, whereas cell migration was measured as described in Experimental Methods. Similar results were seen by using NIH 3T3 cells (data not shown).

The data to this point suggested that both PKA activity and anchoring might be required for chemotaxis. To test this hypothesis, we first determined the effect of inhibition of PKA activity or anchoring on Pd formation. Treatment of cells with either StHt31 or mPKI before growth factor stimulation inhibited Pd formation in a dose-dependent fashion and, to an extent, comparable with total disruption of the actin cytoskeleton with cytochalasin D (Fig. 5B). Pd formation was unaffected by treatment with a control peptide incapable of disrupting PKA-AKAP interaction (StHt31P) but was ablated by arrant activation of PKA with forskolin (Fsk) (Fig. 5B). Moreover, specific inhibition of PKA anchoring, like inhibition of PKA activity, also inhibited chemotactic cell migration in a dose-dependent manner (Fig. 5C). These data show that successful chemotactic cell migration requires not only PKA activity but also localization of that activity through interaction with AKAPs.

Discussion

Chemotactic migration is a fundamentally important cellular behavior. PKA has long been shown to exert both negative and positive effects on cytoskeletal dynamics and cell migration. However, little work has been done to reconcile these disparate observations and carefully elucidate the contribution of this venerable and ubiquitous kinase to the regulation of cell migration. Because the consequences of PKA activity are many and because the targets for PKA activity are scattered far and wide about the cellular landscape, the need for focusing or specifying PKA activity is ostensibly greater, although perhaps less obvious, than for proteins dedicated to a single purpose. The current data reconcile prior observations by showing that PKA is spatially regulated during cell migration and demonstrate the requirement of PKA anchoring for mammalian somatic cell migration. It is important to note, however, that although the current data implicate that localization of PKA within leading edge structures is important for Pd formation and cell migration, they do no not formally obviate potential roles for PKA anchored elsewhere to preside over other processes related to cell migration (e.g., tail retraction).

Given the significant enrichment in PKA activity within Pd, the observed lack of enrichment of PKA-C in Pd may seem curious. However, the currency of PKA activity is carried by the amount of free (i.e., unbound to an R subunit) C subunit, rather than its total amount. Thus, the bulk distribution of catalytic subunit per se is not a faithful indicator of the distribution of PKA activity, and this stresses the importance for subcellular and/or spatial analysis of PKA function. Indeed, this idea is a logical extension of the concept that PKA signaling can be spatially regulated through interaction with AKAPs and is one of the central tenets of the current work. Compensatory enrichment of RI subunits in the CB in some cell types (this study) and the potential for excess R over C subunits (36) may also contribute to this disparity.

VASP and its related proteins are increasingly important regulators of actin dynamics during cell migration, and their phosphorylation has been shown to be critical for regulating their function in this regard (26-29). One consequence of VASP phosphorylation is regulation of its interaction with c-Abl, a nonreceptor tyrosine kinase closely linked to regulation of cytoskeletal dynamics and cell migration in several systems (37). Our data show that VASP-Abl interaction is specifically disrupted within protrusive structures formed during cell migration. It should be noted, however, that unlike Ena (the Drosophila ortholog of VASP) and N-Mena (its mammalian neuron-specific counterpart), VASP is not phosphorylated by Abl (26). Indeed, the biochemical consequences of VASP-Abl interaction for the function of either protein are currently unknown. Nonetheless, the importance of VASP and Abl proteins in cell migration, the dynamic regulation of their binding during cell spreading (8), and the current data all support continued investigation of the role of this interaction in cytoskeletal regulation.

The lack of effect of PKA inhibition on VASP phosphorylation within Pd is somewhat surprising. However, despite its proven importance, the details of regulation of VASP phosphorylation are still largely unknown. The existence of a VASP phosphatase(s) has been directly implicated by pharmacological studies (38) and may be inferred from the rapid dephosphorylation of VASP upon cell adhesion (8). Thus, one hypothesis is that under the current culture conditions, the phosphorylation and dephosphorylation of VASP may not be in rapid equilibrium. Thus, inhibition of PKA, in the absence of specific activation of a putative phosphatase(s), does not result in demonstrable dephosphorylation. Nonetheless, the importance of VASP proteins, and their phosphorylation, in cell migration suggest that regulation of VASP phosphorylation will still prove to be an important function of localized PKA activity.

Unlike phosphorylation of VASP, PTP-PEST phosphorylation and activity as well as the activity of Rac within Pd were significantly affected by localized disruption of PKA function. Indeed, our data make an important connection between previous reports demonstrating a requirement of PKA activity for activation of Rac (9, 11) and the localization of Rac activity to leading-edge structures (35). Our demonstration that regulation of Rac by PKA ostensibly occurs through modulation of both GEF and GAP function is made more intriguing by two recent reports. First, PKA has been shown to directly phosphorylate the Rac/Cdc42 GEF βPIX and, thus, regulate its translocation and activation (39). Second, a complex comprising the α4 integrin, paxillin, and Arf-GAP appears to be responsible for restricting Rac activation to the leading edge during cell migration (40). The potential importance of this pathway is underscored our observations that PKA-mediated phosphorylation of the α4 integrin subunit, which controls α4-paxillin interaction (41), is not only enriched in Pd during smooth muscle cell migration (42) but is also blocked by disruption of PKA anchoring (Fig. 9, which is published as supporting information on the PNAS web site).

Although it is clear that regulation of Rac represents a pathway of major importance for localized PKA activity during cell migration, the growing list of relevant substrates for PTP-PEST (see refs. 22 and 31) provides several other potential targets. Indeed, the complexity of cytoskeletal regulation and chemotaxis (1, 43) and the promiscuity of PKA as a kinase (5) make identification of a single, key target for PKA in chemotaxis unlikely. Even within the scope of the current study, several targets have been identified. It is most likely that localization of PKA activity during migration affects many targets, with the relevance of each depending on how important and how unique its contribution is to a given cell's migration. Clearly, more experimentation is needed to fully elucidate the mechanism of localization, identify additional relevant targets, and define cell-type specificities for the localization of PKA during cell migration. The ability of AKAPs to serve not only as location-specific anchors but as scaffolds for multiple signaling enzymes (20) portends an even greater level of spatial regulation and functional heterogeneity for PKA. The current data provide a foundation in this regard by demonstrating that it is not simply a matter of PKA signaling exerting a blanket negative or positive effect on the cytoskeleton or cell migration. Rather, it appears to be a balance of PKA activity, in extent, time, and subcellular space, that is crucial for successful cell movement.