Cortactin Branches Out: Roles in Regulating Protrusive Actin Dynamics

Mechanisms of actin assembly

The cortical actin cytoskeleton is a complex and dynamic structure that participates in key aspects of cellular homeostasis. In response to extracellular stimuli, specific intracellular signaling cascades produce various actin-based protrusions of the plasma membrane, including lamellipodia, filopodia, dorsal waves, and invadopodia/podosomes. Lamellipodia formation can occur in response to treatment with EGF, leading to the activation of several kinase families, including Src, Erk, and PI3K (Yamaguchi and Condeelis 2007). PI3K in turn activates Rac, a member of the Rho family GTPases, which regulates the signaling pathway involved in barbed end generation and dendritic actin nucleation (Yamaguchi and Condeelis 2007). In addition to phosphorylating cortactin, Src mediates the phosphorylation of several downstream effector molecules, facilitating the activation of Arp2/3 complex and actin nucleation (Yamaguchi and Condeelis 2007). Stimulation with PDGF activates PI3K and Src, as well as PAK1, PKA, and Abl kinases (Buccione et al. 2004). The concerted action of these signaling proteins serves to aid in the formation of circular dorsal ruffles (CDRs), invadopodia, and lamellipodia. Regardless of the signaling pathway, initiation of actin polymerization remains the consistent factor responsible for initiating the formation of membrane protrusions required for motility and invasion, allowing the cell to react and adapt to a variety of environmental cues.

A critical step in dynamic remodeling of the cortical actin cytoskeleton is activation of the actin-related protein (Arp)2/3 complex. The Arp2/3 complex consists of seven subunits (Arp2, Arp3, and ARPC1-5) that function together to nucleate and initiate new actin filaments. Arp2/3 binds to the side of a preexisting F-actin (mother) filament and nucleates production of a new (daughter) filament at a 70° angle (Machesky et al. 1999; Higgs and Pollard 2001). Arp2/3-nucleated branched actin networks in vitro are very similar in appearance to the cortical actin network present in lamellipodia (Svitkina and Borisy 1999), and subsequent work has demonstrated a requirement of Arp2/3 activity in lamellipodia formation (Mullins et al. 1998; Bailly et al. 2001). Newly branched actin filaments elongate by addition of actin monomers to their barbed ends, and are terminated by the binding of capping protein (CP) to the barbed end. The addition of capping protein ensures that the newly produced filaments are short and rigid, with polymerization concentrated near the leading edge to produce protrusive force (Condeelis 2001). Activation of Arp2/3 results in significant conformational change of several complex members, allowing all seven subunits of the complex to interact with the mother filament (Blanchoin et al. 2000a). Spatially, Arp2/3 is positioned at the pointed end of the newly forming filament (Pollard 2007). The exposed surfaces of the Arp2 and Arp3 subunits assume a conformation similar to that of an F-actin barbed end, and these proteins function to mimic the first two G-actin subunits in daughter filaments (Beltzner and Pollard 2004)(reviewed in (Pollard 2007)). Subsequent network disassembly and actin filament depolymerization involves the inhibition of PAK and LIMK. Kinase inhibition results in ADF/cofilin dephosphorylation by the phosphatase slingshot, promoting ADF/cofilin-mediated severing and disassembly of actin filaments at the network rear, replenishing the actin monomer pool for subsequent rounds of polymerization (Lappalainen and Drubin 1997; Theriot 1997; Bamburg 1999; Carlier et al. 1999; Chen et al. 2000; Niwa et al. 2002). In addition to the Arp2/3 complex, several other proteins have been shown to initiate actin nucleation, including formins, spire and Cordon blue (reviewed in (Watanabe and Higashida 2004; Baum and Kunda 2005; Pollard 2007; Winckler and Schafer 2007)). The action of these proteins differs from Arp2/3 in that they nucleate straight actin filaments.

Cortactin structure and subcellular localization

Cortactin is a 63–65kDa protein initially characterized as a tyrosine phosphorylated substrate in v-Src-transformed chick embryo fibroblasts (Kanner et al. 1990; Wu et al. 1991). Cortactin was also independently identified as downstream target of FGFR-mediated c-Src activation (Zhan et al. 1993) and as a gene amplified and overexpressed in tumor cells with chromosome 11q13 amplification (Schuuring et al. 1993). When analyzed by SDS-PAGE, cortactin migrates as an 80/85kDa doublet that cannot be attributed to post-translational modification of the protein (Wu and Parsons 1993). The reason for the incongruity between the calculated MW and the observed Mr in SDS-PAGE experiments is currently unclear, but may involve an altered ability to bind SDS inherent in the primary sequence and/or spontaneous reacquisition of secondary structural elements.

Equilibrium sedimentation and deep etch electron microscopy has determined that cortactin is a monomeric rod-shaped protein ~220Å in length, thinner than the diameter of an actin filament (Weaver et al. 2002). This may represent the “open” form of the protein, since recent work indicates that cortactin can assume an alternative confirmation that is partially globular (Cowieson et al. 2008) (see below). Based on primary sequence analysis cortactin is composed of several distinct domains that either bind other proteins or are sites of post-translation modification (Fig. 1) (Wu et al. 1991). In the murine form, these regions include an amino terminal acidic domain (NTA) (amino acids 1–84). The NTA domain contains a conserved DDW region (amino acids 20–22) that is responsible for interaction with Arp3 and subsequent activation of the Arp2/3 complex (Weed et al. 2000; Higgs and Pollard 2001). Carboxyl terminal to the NTA domain is a series of 6 complete and one partial tandemly repeating segments (amino acids 85–326) termed cortactin repeats that comprise the actin binding region (ABR). The cortactin repeats region directly binds F-actin, with maximal binding activity centered on the fourth repeat. Following the repeats region is a α-helical domain (site of calpain cleavage (Huang et al. 1997b; Perrin et al. 2006)) and a proline-rich region (PRR; amino acids 327–494) enriched in sites of tyrosine and serine phosphorylation. The extreme C-terminus contains an Src homology (SH)3 domain (amino acids 495–542) that interacts with proline-rich binding sequences in several cortactin interacting partners. These include WIP (Kinley et al. 2003), N-WASP (Weaver et al. 2002), MLCK (Dudek et al. 2002), and dynamin 2 (McNiven et al. 2000) (reviewed in (Cosen-Binker and Kapus 2006)).

In serum-starved or unstimulated cells, cortactin is normally present in the cytoplasm as a non-phosphorylated protein (Head et al. 2003). Upon growth factor stimulation (EGF, FGF, or PDGF), activation of Rac1 leads to cortactin relocalization from internal cytoplasmic regions to the cortical actin network (Zhan et al. 1993; Zhan et al. 1994; Head et al. 2003). In addition to Src, cortactin is also phosphorylated by the cytoplasmic tyrosine kinases Fyn, Fer, cMet, c-Abl, Syk ((Gallet et al. 1999; Kapus et al. 2000; Crostella et al. 2001; Huang et al. 2003) Boyle et al. 2007). The serine/threonine kinases PAK1 and MAPK also directly phosphorylate cortactin (Campbell et al. 1999; Webb et al. 2006). The actions of these kinases enable cortactin to function in enhancing actin polymerization and branched actin network formation through its interaction with the actin nucleating complex Arp2/3, as well as the recruitment and positioning of other signaling proteins through SH2 and SH3-mediated interactions as described below.

Cortactin interactions with Arp2/3 and N-WASp

The formation of motility-associated plasma membrane protrusive structures requires the choreographed actions of specific actin effector proteins responsible for the proper spatio-temporal regulation of actin polymerization. Stabilization of branch points in the actin cytoskeletal network is essential for efficient membrane protrusion and directed cell motility. Arp2/3 branch points are inherently metastable, debranching spontaneously in vitro over time (Blanchoin et al. 2000b). Binding of cortactin to Arp2/3 at F-actin branch points stabilizes the branch juncture, inhibiting filament debranching and network breakdown (Weaver et al. 2001).

Cortactin is able to directly influence actin polymerization by serving as a nucleation promoting factor (NPF) for Arp2/3 (Uruno et al. 2001; Weaver et al. 2001). NPFs are a class of proteins that can activate Arp2/3 through direct binding, resulting in a conformational change in the Arp2/3 complex that facilitates de novo actin nucleation and filament elongation as described above (Welch and Mullins 2002). The Wiskott-Aldrich syndrome protein (WASP) family of proteins, including WASP and N-WASp, and the related WAVE/Scar proteins also serve as NPFs for Arp2/3 (Pollard 2007). Cortactin functions as an NPF for Arp2/3 through its ability to bridge Arp3 with ARPC2, 4 and 5, stabilizing the active conformation of the Arp2/3 complex and facilitating the addition of actin monomers to the Arp2 and Arp3 subunits on the side of the mother filament (Weaver et al. 2002; Pollard 2007) (Fig. 2A). Interestingly, the NPF function of cortactin requires both the NTA and the actin binding region (ABR) of cortactin, demonstrating the importance of interaction with both F-actin and Arp2/3 for efficient actin polymerization (Weaver et al. 2001).

Similar to cortactin, N-WASp and WAVE/Scar proteins bind to F-actin, though this interaction occurs through a basic region and not an ABR (Suetsugu et al. 2001; Kelly et al. 2006). These proteins can bind G-actin monomers and the Arp2/3 complex by virtue of a Wasp Homology-2 (WH2)-central-acidic (WCA) region ((Marchand et al. 2001), reviewed in (Takenawa and Suetsugu 2007)). While F-actin binding is important for WASp-family protein function, the binding of G-actin to the WCA region results in dramatic enhancement of NPF activity by placing actin monomers in close proximity to activated Arp2/3 (Machesky et al. 1999).

Although cortactin-mediated Arp2/3 activation is weaker than that of N-WASp, cortactin synergizes with N-WASp in activating Arp2/3, enhancing N-WASp mediated actin nucleation (Weaver et al. 2001). How N-WASp and cortactin cooperate to initiate Arp2/3 activation is somewhat controversial. The interaction of N-WASp with the side of an actin filament enhances recruitment of and activates Arp2/3, through binding to Arp2, Arp3, and ARPC1 subunits (Suetsugu et al. 2001; Pollard 2007). The binding of the cortactin NTA domain displaces the N-WASp WCA domain from the Arp3 subunit, but not Arp2 or ARPC1 (Weaver et al. 2002). In this manner, N-WASp and cortactin may both bind to and activate Arp2/3 complex concomitantly, forming a ternary complex of Arp2/3/WCA/NTA that results in enhanced NPF activity. Alternately, evidence exists that the activation of Arp2/3 by N-WASP and cortactin occurs in a sequential manner. Cortactin has a higher binding affinity for activated Arp2/3 complex at branch points compared to non-activated, non-actin associated Arp2/3 (Uruno et al. 2003). After Arp2/3 is activated by N-WASp, cortactin effectively binds and displaces N-WASp from the Arp2/3, presumably stabilizing the Arp2/3/cortactin/F-actin complex (Uruno et al., 2003). Whether one or both of these mechanisms are employed at the cellular level to drive Arp2/3 activity remains to be determined.

In addition to directly interacting with Arp2/3 via its NTA, cortactin also promotes actin-nucleation activity indirectly through the SH3 domain. The SH3 domain of cortactin binds directly to a proline-rich region on N-WASp, liberating N-WASp from its auto-inhibited state, resulting in N-WASp-mediated Arp2/3 nucleation activity (Mizutani et al. 2002; Martinez-Quiles et al. 2004) (Fig. 2B). The interaction of cortactin and N-WASp promotes the localization of N-WASp to sites of actin polymerization within invadopodia/podosomes (Mizutani et al. 2002) and enhances cell movement (Kowalski et al. 2005). Collectively these studies suggest cortactin and N-WASp are intimately intertwined in regulating Arp2/3 activity responsible for cortical actin assembly.

WASP-interacting protein (WIP) also binds to the SH3 domain of cortactin and enhances cortactin-mediated Arp2/3 activation (Kinley et al. 2003) (Fig. 2C, D). Similar to other SH3 binding partners, localization of WIP to the cell cortex area is dependent upon the cortactin binding domain (Kinley et al. 2003). Maximal Arp2/3 activity is achieved when the cortactin-WIP complex is associated with actin filaments, suggesting a novel role for cortactin in linking WIP to pre-existing filaments. WIP also inhibits the depolymerization of actin filaments (Martinez-Quiles et al. 2001), resulting in the stabilization of actin branch points and inhibition of filament depolymerization.

Characterization of cortactin binding to F-actin

Cortactin was initially identified as an F-actin binding protein in cosedimentation assays using recombinant truncation mutants (Wu and Parsons 1993). Topical mapping analysis determined that the repeat region was responsible for F-actin binding, with a binding stoichiometry of one cortactin molecule per 14 actin monomer subunits. Deletion mapping of the cortactin repeats region indicates that the presence of the fourth repeat is required for optimal F-actin binding (Weed et al. 2000). This finding is supported by F-actin binding studies with different naturally-occurring cortactin splice variants isolated from human squamous cell carcinoma cell lines (van Rossum et al. 2003). Alternate splicing of the cortactin transcript results in the loss of the 6^th ABR (termed splice variant 1; SV1-cortactin) or the loss of the 5^th and 6^th ABR (SV2-cortactin). Both splice variants retain the fourth repeat and display similar F-actin binding activities as compared to wild-type cortactin. In addition, these splice variants retain the ability to activate Arp2/3-mediated actin nucleation, although the SV2-cortactin variant is not as effective as SV1-cortactin. Collectively these studies support the importance of the fourth cortactin repeat in maintaining the actin binding activity of the protein. While other studies have shown cortactin splice variants lacking the fourth repeat are capable of binding F-actin, these isoforms are not as abundant as wild type, SV-1 or SV-2 cortactin (Katsube et al. 2004).

To determine which residues within actin filaments interact with cortactin, Pant et al. (Pant et al. 2006) performed a 3D reconstruction analysis from electron micrographs of actin filaments decorated with a recombinant protein encompassing the cortactin repeats region. Their results indicate that cortactin likely binds to the actin filament near amino acids 338–348 in an individual actin subunit, a known binding region for the F-actin associated proteins gelsolin (McLaughlin et al. 1993), vitamin D binding protein (DBP) (Otterbein et al. 2001), and ADF/cofilin (reviewed in (Dominguez 2004)). Binding of the cortactin repeats region to this area deepens the cleft between adjacent actin monomers. The authors hypothesize that this leads to a transient instability within the helix, allowing for the binding of other proteins to the filament, (i.e.; Arp2/3 complex). In addition, cortactin preferentially binds to F-actin filaments containing ATP or ATP/ADP-Pi, demonstrating a higher affinity for newly polymerized actin filaments as opposed to older ADP-loaded filaments (Bryce et al. 2005). This is in agreement with the requirement for cortactin for maximal incorporation of fluorescently-labeled G-actin monomers within extending lamellipodia (Bryce et al. 2005).

Detailed conformational information on cortactin has not been available due in part to the lack of a published crystal structure. Recently, cortactin was analyzed by a combined approach utilizing chemical cross-linking, circular dichroism and small angle X-ray scattering (Cowieson et al. 2008). This study revealed that recombinant cortactin assumes a globular “lollipop” shape in solution, with the helical domain residing next to the actin-binding region. This conformation results in a folding back of the C-terminal region, placing the second actin repeat in close proximity to the fifth repeat, and SH3 domain near the helical domain and the second repeat. The conformation deduced from these studies has important ramifications regarding how cortactin can interact with F-actin, since binding to F-actin does not induce changes in cortactin secondary structure (Cowieson et al. 2008). Actin binding studies at low temperature reveal that cortactin can assemble F-actin into sheets ~ one filament width wide, thereby bundling F-actin into parallel arrays. Whether or not this activity takes place in a cellular context remains to be evaluated.

One regulatory mechanism recently identified that governs cortactin binding to F-actin involves the histone deacetylase HDAC6 (Zhang et al. 2007). Cortactin is a substrate for the histone acetyltransferase PCAF, which acetylates eleven lysine residues within the protein. It is not clear if PCAF is the only histone aceytltransferase that acetylates cortactin, as there is no significant difference in the acetylation status of cortactin from PCAF^−/− fibroblasts as compared to controls (Zhang et al. 2007). Eight of the targeted lysines are located within the ABR of cortactin. These eight lysine residues are postulated to form two positively “charged patches” that facilitate the favorable interaction of cortactin with F-actin. Acetylation by PCAF neutralizes the lysine patches, inhibiting binding of the ABR to F-actin. Deacetylation by HDAC6 reverses this process and restores the ability to bind F-actin. The acetylation/deacetylation status of cortactin also has physiological consequences, as acetylated cortactin is unable to translocate to the cell periphery in a Rac-dependent manner, while deacetylated cortactin retains this ability. Cells containing acetylated cortactin also showed decreased cell migration, likely due to impaired cortactin binding to F-actin. The reversible acetylation status of cortactin provides a unique signaling mechanism that regulates F-actin binding activity.

The effects of cortactin phosphorylation on the actin cytoskeleton

The phosphorylation of cortactin and resulting functional consequences have been an intense area of study for many years, with several recent reports providing a better understanding with regard to actin dynamics and cellular behavior. In the case of tyrosine phosphorylation, initial mapping studies identified the major Src phosphorylation sites on murine cortactin as Y421, Y466 and Y482, with Y486 phosphorylated to a lesser extent (Huang et al. 1997a; Huang et al. 1998). Tyrosine 475 has also been identified as a phosphorylation site by mass spectroscopy, although the responsible kinase(s) have not been identified (Martin et al. 2006). Phosphorylation of Y421, Y466 and Y482 has known biochemical and cellular consequences with regards to actin dynamics. Tyrosine phosphorylation of cortactin at these sites is hypothesized to lead to a conformational change in the protein, resulting in more efficient cleavage by calpain proteases (Huang et al. 1997b; Perrin et al. 2006), possibly affecting the ability of cortactin to bind to and cross-link actin filaments. The majority of reports indicate that high levels of tyrosine phosphorylation correlate with elevated cell migration and cancer metastasis (Huang et al. 1998; Liu et al. 1999; Bourguignon et al. 2001; Li et al. 2001; Huang et al. 2003), although recent evidence indicates that cortactin tyrosine phosphorylation is inhibits motility in some tumor types (Jia et al. 2008). Cell type or context differences may play a role in explaining these discrepancies. Tyrosine phosphorylation can also regulate cortactin function by enhancing the binding of select SH3-ligands, including MLCK, CD2AP, and dynamin2 (Dudek et al. 2002; Lynch et al. 2003; Zhu et al. 2007). Cortactin was recently identified as a specific substrate for the tyrosine phosphatase PTP1B (Stuible et al. 2008). PTP1B binds to cortactin primarily at phosphorylated Y446, although surrounding residues may indirectly affect this interaction. EGFR activation leads to phosphorylation of Y446, as does hyperosmolarity, which activates Fyn and Fer (Kapus et al. 2000) resulting in phosphorylation of the Src-targeted sites Y421, Y470, Y486 in addition to Y446. Expression of PTP1B reduces Y421 and Y446 phosphorylation, and a Y-F mutation at codon 446 results in increased hyperosmotic apoptotic signaling, indicating that phosphorylation of Y446 is may be important for regulating apoptotic responses based on its phosphorylation status.

Several recent studies have examined the downstream cellular effects of cortactin tyrosine phosphorylation. Tyrosine phosphorylation of cortactin in osteoclasts is essential for the formation and turnover of podosomes and invadopodia, structures associated with invasive potential (Luxenburg et al. 2006; Tehrani et al. 2006; Ayala et al. 2008). Cortactin tyrosine phosphorylation is also required for efficient metastasis of cancer cells, where breast cancer cell lines expressing a tyrosine phosphorylation-deficient mutant of cortactin demonstrate a marked reduction of osteolytic metastases compared to control cells (Li et al. 2001). The phosphorylation status of cortactin also plays a role in centrosome separation at the G2-M phase in the cell cycle though its ability to link the actin cytoskeleton to the centrosome (Wang et al. 2008). Suprastimlation of pancreatic acini, used to mimic the disease state of pancreatitis, leads to an increase in cortactin tyrosine phosphorylation and the reorganization of the actin cytoskeleton, resulting in the formation of cortactin-rich membrane blebs (Singh and McNiven 2008). This process is dependent on cortactin tyrosine phosphorylation since expression of a phosphorylation-null mutant decreases actin reorganization and diminishes membrane bleb formation. Future work will likely uncover new roles for cortactin tyrosine phosphorylation in normal and pathogenic cellular processes.

While cortactin phosphorylation plays a role in several discrete cellular functions, a link between tyrosine phosphorylation and Arp2/3 activation was initially dismissed since cortactin fragments lacking the α-helical, PRR and SH3 domains activate Arp2/3 in vitro nearly as effectively as the full-length protein (Weaver et al. 2001). However, additional work now points to a positive effect for Src-mediated cortactin phosphorylation on Arp2/3 mediated actin polymerization through Nck and N-WASp (Tehrani et al. 2007) (Fig. 2E–G). Nck1 binds to Src-phosphorylated cortactin through its SH2 domain, with the Nck1/phosphocortactin complex in turn interacting with either WIP (Fig. 2E) or N-WASp (Fig. 2F) through its SH3 domain. Trimeric cortactin/Nck1/N-WASp or /WIP complexes enhance Arp2/3 nucleation activity. Interestingly, when cortactin is the only NPF present (i.e., in conditions lacking N-WASp but retaining NCK and WIP), the Arp2/3-binding DDW motif within the cortactin NTA domain is essential for stimulating Arp2/3-mediated actin nucleation. Additionally, the SH3 domain of Src-phosphorylated cortactin is not required for Arp2/3 activation when Nck1 and N-WASP or WIP are present, suggesting that the Nck1 SH3 domain can perform in a surrogate manner to bring N-WASp or WIP to cortactin, resulting in Arp2/3 complex activation. The binding of Nck1 to phosphorylated cortactin therefore provides an indirect link to Arp2/3 complex regulation, as well as involving cortactin in an additional layer of complexity in regulating actin nucleation (Fig. 2G).

In addition to tyrosine phosphorylation, cortactin is also a substrate for several serine/ threonine kinases. Serine residues 405 and 418 were first identified to be phosphorylated downstream of MEK and can be directly phosphorylated by Erk1/2 (Campbell et al. 1999; Webb et al. 2006). Phosphorylation of S405 and 418 is associated with an 80 to 85kDa shift in Mr on SDS-PAGE, potentially due to the disruption of intramolecular interactions between the cortactin SH3 domain and amino-terminal residues (Campbell et al. 1999; Martinez-Quiles et al. 2004; Stuible et al. 2008). In support of this, phosphorylation of S405 and S418 corresponds with increased N-WASp binding and activation of Arp2/3 actin nucleation (Martinez-Quiles et al. 2004). Interestingly, Src-mediated tyrosine phosphorylation inhibits the binding between cortactin and N-WASp in this setting, resulting in decreased Arp2/3 nucleation (reviewed in (Lua and Low 2005)). This may be due to the absence of Nck in these assays. Phosphorylation of S405 and S418 is also required for efficient invadopodia formation and extra cellular matrix degradation (Ayala et al. 2008).

In addition to chemical ligand stimulation, cortactin phosphorylation and translocation can also be affected by mechanical or osmotic stress. Endothelial cell monolayers exposed to shear stress demonstrate enhanced stress fiber formation and Rac-dependent cortactin translocation to the cortical actin network independent of cortactin tyrosine phosphorylation (Birukov et al. 2002). However, monolayer exposure to sphingosine 1-phosphate (S1P) increases barrier function in a cortactin-dependent fashion, as cortactin knock-down significantly inhibits S1P-mediated barrier enhancement (Dudek et al. 2004). Translocation and tyrosine phosphorylation of cortactin is also increased in response to S1P treatment, and is critical to regulating barrier function. In addition, the interaction between MLCK and cortactin also affects barrier function, as cortactin blocking peptides inhibit S1P-mediated effects on MLC phosphorylation and barrier enhancement (Dudek et al. 2004). In the case of osmotic stress, hyperosmolarity results in cell shrinkage, triggering the tyrosine phosphorylation of cortactin independent of changes in osmolarity or intracellular pH (Kapus et al. 1999). Cortactin phosphorylation under these conditions is mediated by the Src family member Fyn. Rac-dependent cortactin translocation in response to hypertonicity is independent of shrinkage-induced phosphorylation, localizing cortactin and Arp2/3 to areas of active cortical actin reorganization known to occur in response to osmotic stress (Di Ciano et al. 2002).

Separate from exogenous stimuli, cortactin phosphorylation can also be effected by the polymerization status of actin (Fan et al. 2004). Depolymerization of actin by latrunculin B leads to enhanced tyrosine phosphorylation of cortactin mediated by Fer. Alternately, filament stabilization by jasplakinolide blocks cortactin phosphorylation induced by either hypertonic stress or treatment with latrunculin B. Stabilization results in the redistribution of cortactin to F-actin rich patches. In this way, the polymerization status of the F-actin network may regulate the actions of cortactin by altering its phosphorylation status, resulting in a feedback loop alternating between network expansion and reorganization (Gunst 2004).

Recently, the Cell Migration Consortium reported a comprehensive mass spectroscopy analysis in an attempt to identify all cortactin phosphorylation sites (Martin, et al 2006). In addition to the sites discussed above, sixteen additional phosphorylation sites were identified, twelve occurring on serine and four on threonine. Of these sites, serine 113 was independently identified as a PAK phosphorylation site, serving to downregulate F-actin binding to cortactin (Webb, et al 2006). Another intreuging phosphorylation site identified from this analysis is threonine 24, which is directly adjacent to the DDW motif in the NTA region (Martin, et al 2006). This raises the possibility that phosphorylation of this site may play a role in the regulation of Arp2/3 binding to the cortactin NTA. Future work will likely address the role of T24 phosphorylation and associated kinase(s) in cortactin function, as well as how other cortactin phosphorylation sites regulate cortactin activity.

Additional cortactin binding proteins and their role in actin organization

Given its involvement in multiple actin-based cellular processes, cortactin is capable of exerting differential effects on cortical actin cytoskeletal organization and assembly. In addition to the proteins discussed above, the ability of cortactin to alternatively regulate cortical actin dynamics stems from its direct association with additional actin-associated proteins. These interactions provide further insights into how cortactin regulates actin polymerization and organization.

Cortactin function in the formation of actin-based protrusive structures

The localization of cortactin with the cortical actin cytoskeleton combined with its role in regulating actin polymerization positions it to play a governing role in a variety of cellular events involving actin-based protrusion of the cell membrane. Consequently, the contribution of cortactin in the formation and function of lamellipodia, invadopodia and CDRs has been the subject of much recent investigation, leading to a better understanding about the molecular events and signaling pathways that regulate these structures.