Paxillin is a focal adhesion-associated, phosphotyrosine-containing protein that may play a role in several signaling pathways. Paxillin contains a number of motifs that mediate protein–protein interactions, including LD motifs, LIM domains, an SH3 domain-binding site and SH2 domain-binding sites. These motifs serve as docking sites for cytoskeletal proteins, tyrosine kinases, serine/threonine kinases, GTPase activating proteins and other adaptor proteins that recruit additional enzymes into complex with paxillin. Thus paxillin itself serves as a docking protein to recruit signaling molecules to a specific cellular compartment, the focal adhesions, and/or to recruit specific combinations of signaling molecules into a complex to coordinate downstream signaling. The biological function of paxillin coordinated signaling is likely to regulate cell spreading and motility.
Paxillin was first identified as a 68–70 kDa phosphotyrosine-containing protein in cells transformed by the src oncogene (Glenney Jr and Zokas, 1989). Further characterization of the protein revealed that paxillin localized to discrete structures in the cell called focal adhesions, which are sites of close cellular contact with the underlying extracellular matrix (Turner et al., 1990). Tyrosine phosphorylation of paxillin was observed following integrin-dependent cell adhesion to extracellular matrix proteins, thus implicating paxillin in integrin-mediated signaling (Burridge et al., 1992). These early findings generated a great deal of interest in paxillin providing the impetus for extensive analysis of the protein. These studies have provided us with our current vision of paxillin, a focal adhesion-associated signaling molecule that functions as an adaptor protein to recruit diverse cytoskeleton and signaling proteins into a complex, presumably to coordinate the transmission of downstream signals.
The molecular cloning of paxillin revealed a number of motifs that are now known to function in mediating protein–protein interactions (see Figure 1) (Turner and Miller, 1994; Salgia et al., 1995a). The N-terminal half of paxillin contains a proline-rich region that could serve as an SH3 domain-binding site. Several tyrosine residues conforming to SH2 domain binding sites were also noted. In addition, the N-terminal domain of paxillin contains five copies of a peptide sequence, called the LD motif, which are now known to function as binding sites for other proteins (see Table 1) (Brown et al., 1998a). The C-terminal half of paxillin is comprised of four LIM domains, which are zinc-binding structures resembling a double zinc finger domain. These domains were first identified in three transcription factors, lin-11, isl-1, mec-3. LIM domains can also function to mediate protein-protein interactions (Dawid et al., 1998). The focal adhesion targeting sequence of paxillin is located in the C-terminal half of the protein. The 3rd LIM domain plays a major role in targeting paxillin to focal adhesions, whereas LIM domain 2 plays a minor role (Brown et al., 1996). Presumably these LIM domains engage binding partners to direct and/or tether paxillin in focal adhesions, but these binding partners have thus far eluded identification. Since paxillin does not exhibit enzyme activity, but contains an array of docking sites for other proteins, it is believed to function as an adaptor protein that facilitates the assembly of multi-protein complexes to coordinate signaling.
Paxillin: alternative splicing and related proteins
The first paxillin isoform isolated, referred to as paxillin α, is broadly expressed. Two alternatively spliced variants of paxillin, paxillin β and paxillin γ, have been described (Mazaki et al., 1997). Both of these isoforms contain a short exon inserted just downstream of LD motif 4, between lysine residue 277 and phenylalanine residue 278 of paxillin α. Paxillin β contains a 34 amino acid insertion and paxillin γ contains a 48 residue insertion that is unrelated in sequence to the paxillin β specific exon. The expression pattern of paxillin β and paxillin γ is more restricted than the expression of paxillin α (Mazaki et al., 1997). In fact, paxillin γ may be a human specific isoform since the murine paxillin gene lacks the exon encoding the γ-specific isoform sequences (Mazaki et al., 1998). Interestingly, the insertions appear to alter the ability of these paxillin isoforms to interact with other proteins (Mazaki et al., 1997). Compared to paxillin α, paxillin β exhibits reduced binding to vinculin in vitro, but equivalent binding to FAK. In contrast, paxillin γ is defective for FAK binding in vitro but retains vinculin-binding activity. It is not clear why these isoforms exhibit altered binding to FAK and vinculin since the insertions do not appear to disrupt the LD4 motif, which is a FAK/vinculin binding site (see below). Further, both FAK and vinculin can bind to other LD motifs in the N-terminus of paxillin (see below), which are distant from the isoform specific insertions. Perhaps the insertions disrupt interactions with other binding partners that stabilize the association of FAK or vinculin with paxillin. Alternatively, these insertions may alter the conformation of the protein masking recognition of both the LD2 and LD4 motifs by FAK or vinculin.
Paxillin is the founding member of a family of related proteins that now contains three members, paxillin, hic-5 and leupaxin (see Figure 2). Hic-5 was originally isolated from a screen for TGF-β and hydrogen peroxide inducible genes (Shibanuma et al., 1994). Hic-5 is 57% identical to paxillin and contains an organization of N-terminal LD motifs and C-terminal LIM domains similar to paxillin. Hic-5 localizes to focal adhesions and shares a number of binding partners with paxillin including FAK (Hagmann et al., 1998; Thomas et al., 1999a; Ishino et al., 2000a). The leupaxin cDNA was originally isolated from a macrophage library and its expression in lymphoid tissue documented (Lipsky et al., 1998). Leupaxin exhibits 37% identity with paxillin and contains the same domain organization as paxillin with N-terminal LD motifs and four C-terminal LIM domains (Lipsky et al., 1998). As expected from the sequence conservation, leupaxin can associate with a number of the same proteins that can bind paxillin including the FAK-related protein, CAKβ/Pyk2/CadTK/RAFTK (Lipsky et al., 1998). It is anticipated that some of the functions of paxillin, hic-5 and leupaxin overlap given the high degree of sequence identity within the LD motifs and LIM domains. However, within the N-terminal domains of these proteins are regions of sequence divergence, suggesting that each protein might also perform unique functions (see Figure 2).
The N-terminus of paxillin: LD motif binding partners
Paxillin-associated, actin-binding proteins
Vinculin, a focal adhesion-associated protein that functions in linking the actin cytoskeleton to focal adhesions was the first paxillin binding partner to be identified (Turner et al., 1990). Vinculin binds to the N-terminal domain of paxillin and appears to associate with three different sites, LD1, LD2 and LD4 (Brown et al., 1996; Turner et al., 1999). The binding site on vinculin resides between residues 978–1000 (Wood et al., 1994). Comparison of the sequence of this region of vinculin (actually residues 951–1000) with several other paxillin binding partners that also associate via the LD motifs has revealed sequence similarity (Tachibana et al., 1995; Nikolopoulos and Turner, 2000; Nikolopoulos and Turner, 2001). This sequence has been called the paxillin binding sequence (PBS) and where tested, mutations within the PBS disrupt the association of these proteins with paxillin (see Table 2). The PBS in vinculin is in the same region of the protein that may function in promoting the localization of vinculin to focal adhesions (Wood et al., 1994). Although this observation suggests that paxillin may play a role in targeting vinculin to focal adhesions, vinculin also contains a second focal adhesion targeting sequence in its N-terminus. Vinculin also contains talin- and actin-binding sites and the interaction of vinculin with these proteins is regulated by phosphatidylinositol-bisphosphate (Gilmore and Burridge, 1996). Vinculin may function to connect the actin cytoskeleton to paxillin.
Actopaxin is a 42-kDa protein that binds LD1 and LD4 of paxillin (Nikolopoulos and Turner, 2000). Actopaxin has two calponin homology domains, which usually function to bind actin, and actopaxin exhibits F-actin binding activity. There is a PBS in actopaxin and this sequence is part of the paxillin-docking site on actopaxin. In cells, actopaxin localizes to focal adhesions and correct localization is dependent upon paxillin binding. Initial experiments using a dominant negative strategy to perturb actopaxin function suggests that this protein functions in spreading and adhesion (Nikolopoulos and Turner, 2000).
Why does paxillin associate with several actin-binding proteins? Two functions can be proposed for these complexes. First, this could simply be an additional mechanism for anchoring the actin cytoskeleton in focal adhesions (see Figure 3). A second, more intriguing scenario is that actopaxin and vinculin tether paxillin (and associated proteins) to the actin cytoskeleton so that actin-based contractility causes clustering of paxillin and associated proteins (see Figure 3). This may be an important mechanism for activation of paxillin associated signaling molecules. In this regard it is noteworthy that stimuli that promote contractility induce tyrosine phosphorylation of paxillin and the activation of FAK, whereas inhibition of actin polymerization or contractility frequently inhibits signaling events in focal adhesions (Burridge and Chrzanowska-Wodnicka, 1996).
Paxillin associated enzymes
GTPase activating proteins (GAPs) for the ARF family of GTP-binding proteins
Recently, a number of proteins with ARF-GAP activity have been shown to bind to paxillin. PKL (paxillin-kinase linker) is a 95 kDa protein that can bind to the LD4 motif of paxillin and hic-5 (Turner et al., 1999). Avian PKL is highly related to murine Git2 and contains an N-terminal ARF-GAP domain. Two PBSs have been identified in PKL, one near the N-terminal ARF-GAP domain and the second near the C-terminus. The binding site for paxillin lies within the C-terminal 309 residues suggesting that the second PBS mediates binding to paxillin. Git1 is another ARF-GAP that is highly related in sequence to PKL (Turner et al., 1999). Like PKL, Git1 also associates with paxillin and the C-terminal PBS of Git1 is likely to mediate this interaction (Zhao et al., 2000). Both PKL and Git1 associate with PIX, a guanine nucleotide exchange factor for Rho family GTP-binding proteins, and paxillin can indirectly associate with PIX via interactions with PKL or Git1 (Turner et al., 1999; Zhao et al., 2000). Thus, in addition to their ARF-GAP catalytic function, PKL and Git1 also function as adaptor proteins recruiting an additional enzyme into complex with paxillin. Further, PIX associates with the PAK serine/threonine kinase, and paxillin has been shown to associate indirectly with PAK via PKL and PIX (Turner et al., 1999). Git2s is an alternatively spliced isoform of Git2, which is missing the C-terminal sequences of Git2 including the C-terminal PBS. Nevertheless, Git2s also associates with paxillin, albeit less strongly than full length Git2 (Mazaki et al., 2001). The N-terminal region of Git2s, which contains the N-terminal PBS, mediates its interaction with paxillin. In addition, another ARF-GAP called PAG3 (paxillin-binding proteins bearing ARF-GAP motifs) also binds to paxillin, although the details of the interaction have not been fully elucidated (Kondo et al., 2000). Interestingly, PAG3 was also identified as a CAKβ/Pyk2-associated ARF-GAP called PAPα (Andreev et al., 1999).
The adaptor function of PKL/Git1 may be a mechanism to recruit PIX and PAK into complex with paxillin, which may be very important for coordinating the activation and transmission of signals downstream of Rho proteins to regulate the dynamics of the actin cytoskeleton. However, the observation that four different ARF-GAP proteins can complex with paxillin, suggests that regulation of ARF proteins is an important aspect of paxillin signaling. The paxillin-associated ARF-GAPs may regulate subcellular localization of paxillin (see below). There is also some evidence that these ARF-GAP proteins can regulate cell motility (Zhao et al., 2000; Kondo et al., 2000; Mazaki et al., 2001). The mechanisms of regulation of cell motility by paxillin and these ARF-GAPs remain to be elucidated.
ILK is a serine/threonine protein kinase that was isolated as a β1 integrin binding partner in a yeast 2-hybrid assay (Dedhar, 2000). In cells, ILK co-localizes with integrins in focal adhesions. ILK can bind specifically to the LD1 motif of paxillin both in vitro and in vivo (Nikolopoulos and Turner, 2001). Within the C-terminal domain of ILK is a sequence related to the PBS found in vinculin and mutations within the PBS abrogate binding to paxillin in vitro. The C-terminal domain of ILK also contains a binding site for β1 and β3 integrins and interactions with the cytoplasmic domains of these integrin subunits may be important for localization (Dedhar, 2000). In addition, an interaction between the N-terminal domain of ILK and the LIM domain protein PINCH has also been implicated in targeting ILK to focal adhesions (Dedhar, 2000). It has now been shown that mutations in the PBS of ILK that disrupt paxillin binding also dramatically impair the ability of ILK to localize to focal adhesions (Nikolopoulos and Turner, 2001). Targeting of ILK to focal adhesions is clearly complex and it appears that association with multiple binding partners, one of which is paxillin, may be required for ILK localization to focal adhesions.
Although PAK can be indirectly recruited to paxillin via PKL and PIX interactions (Turner et al., 1999), there is also evidence that PAK3 can directly associate with paxillin. The PIX binding site on PAK3 resides between residues 184 and 205 (Hashimoto et al., 2001). In vitro, the N-terminal proline-rich fragment of PAK3 spanning residues 1 to 62 can bind paxillin (Hashimoto et al., 2001). Thus, at least in vitro, PAK3 can bind to paxillin independently of PIX. The binding site for the N-terminus of PAK3 on paxillin has been localized to LD4 (Hashimoto et al., 2001). Further, in coexpression experiments PIX and PAK3 can compete with each other for paxillin binding. These results suggest that there may be two mechanisms utilized by paxillin to recruit PAK3 into a signaling complex.
FAK and CAKβ/Pyk2/CadTK/RAFTK
The focal adhesion kinase, FAK, is another paxillin binding partner (Turner and Miller, 1994; Hildebrand et al., 1995). The paxillin binding site on FAK is within the C-terminal focal adhesion targeting sequence of FAK (Hildebrand et al., 1995). FAK binds to two sequences in paxillin, LD2 and LD4 (Brown et al., 1996). By comparison of the sequence of the regions of FAK and vinculin that associate with paxillin, a region of homology was defined, i.e. the PBS (Tachibana et al., 1995). While the PBS is a contiguous sequence within vinculin, the sequence is separated into two noncontiguous sequences in FAK, called PBS1 and PBS2. Mutations within PBS1 and PBS2 disrupt paxillin binding and focal adhesion targeting of FAK (Tachibana et al., 1995). This finding suggested that FAK localizes to focal adhesions by binding paxillin. However, additional mutations outside of this region of homology with vinculin also abrogate the paxillin binding activity of FAK (Cooley et al., 2000). Further, additional mutants of FAK that are defective for binding paxillin still retain the ability to target to focal adhesions (Schaller et al., 1993; Hildebrand et al., 1995; Cooley et al., 2000). Thus association with paxillin is dispensable for the correct localization of FAK. However, every mutant of FAK that fails to target to focal adhesions is also defective for paxillin binding (Hildebrand et al., 1995; Tachibana et al., 1995; Cooley et al., 2000). The most likely explanation for these paradoxical observations is that interactions with two distinct binding partners can promote focal adhesion targeting of FAK and that interactions with both of these binding partners, one of which is paxillin, must be disrupted to abolish the localization of FAK.
CAKβ/Pyk2/CadTK/RAFTK is a FAK-related tyrosine kinase that is highly homologous to FAK within several regions, including the C-terminal focal adhesion targeting sequence. Not surprisingly, CAKβ also associates with paxillin (Salgia et al., 1996; Schaller and Sasaki, 1997; Li and Earp, 1997; Hiregowdara et al., 1997). Although CAKβ contains a functional focal adhesion targeting sequence (Schaller and Sasaki, 1997), CAKβ is not strongly localized to focal adhesions in many cell types, but rather is distributed in the cytoplasm (Schaller and Sasaki, 1997; Sieg et al., 1998; Zheng et al., 1998; Litvak et al., 2000). Under certain conditions CAKβ can be stimulated to translocate to focal adhesions and under these conditions the amount of paxillin associated with CAKβ is dramatically increased (Litvak et al., 2000). Although there are several interpretations for this observation, it is consistent with the theory that association with paxillin is important for the relocalization of CAKβ to focal adhesions.
The association of paxillin with FAK and CAKβ could also play an important role in directing phosphorylation of paxillin by these tyrosine kinases. A paxillin mutant that is unable to bind FAK exhibits reduced steady state levels of phosphotyrosine, reduced tyrosine phosphorylation following cell adhesion to fibronectin and reduced tyrosine phosphorylation in Src transformed cells (Thomas et al., 1999b). CAKβ can induce tyrosine phosphorylation of hic-5 and the induction of tyrosine phosphorylation depends upon the binding of hic-5 to the C-terminus of CAKβ (Ishino et al., 2000b). These data support the contention that association of paxillin with FAK and CAKβ is a mechanism to promote tyrosine phosphorylation of paxillin.
The translation product of a viral oncogene
The E6 oncoproteins from human papillomavirus 16 and bovine papillomavirus have several cellular binding partners whose function may be co-opted in the process of oncogenic transformation. Using different strategies, two groups have identified paxillin as a binding partner for E6 (Tong and Howley, 1997; Vande Pol et al., 1998). E6 interacts with several LD motifs of paxillin in vitro, although the most important in vivo binding site appears to be LD1 (Tong et al., 1997; Vande Pol et al., 1998). While it is clear that paxillin is not the sole E6 target within the cells, analysis of E6 mutants has provided a correlation between paxillin binding and transformation (Tong and Howley, 1997; Das et al., 2000). Thus paxillin may be one of several cellular proteins whose function is altered upon transformation by the papillomaviruses. The mechanism of action of the oncoprotein is not yet clear but likely involves disruption of normal paxillin-containing protein complexes. In vitro, recombinant E6 protein can associate with recombinant paxillin and block subsequent interactions between paxillin and FAK or vinculin (Tong et al., 1997). It is also interesting to observe that expression of E6 disrupts actin stress fibers. Since the actin cytoskeleton is potentially linked to paxillin via actopaxin and vinculin, this phenotype may be due to E6 binding to paxillin and disrupting interactions with actopaxin and vinculin (Tong et al., 1997; Tong and Howley, 1997; Nikolopoulos and Turner, 2000).
LD motif mediated interactions: specificity, affinity and regulation
Clearly the LD motifs within the N-terminal domain of paxillin play very important roles in the assembly of actin-binding proteins and signaling proteins into complexes (see Figures 3 and 4). The finding that many different proteins associate with LD motifs of paxillin raises some very important questions about these interactions. How do these LD motifs interact with their binding partners? What dictates specificity of the interactions? Why do some proteins bind to multiple LD motifs? How do different paxillin binding partners compete for binding? How are these interactions regulated?
Within every LD motif the core consensus sequence is conserved. However, comparison of the sequence of individual LD motifs between species or between family members reveals that sequences in addition to the core consensus motif are conserved. For example, comparison of the LD4 motif of human paxillin with the LD4 motif of frog paxillin and the LD4 motif of hic-5 reveals a high degree of sequence conservation extending beyond the core consensus. These sequences are likely to provide the specificity for interaction with individual paxillin-binding partners. While this theory remains to be rigorously tested, there is one reported point mutation within an LD motif that disrupts vinculin binding but does not affect FAK binding (Brown et al., 1996). Comparison of the sequences of the LD motifs might provide additional clues to the specificity of interaction. For example, FAK binds to LD2 and LD4 (Brown et al., 1996; Thomas et al., 1999b). Comparison of the sequences of LD2 and LD4 motifs among the paxillin family members and between species reveals three additional highly conserved residues upstream of the core LD motif consensus (see Table 3). These residues may provide the specificity in the LD2/LD4 interaction with FAK.
In addition to the problem of how binding partners distinguish different LD motifs, there can be multiple proteins that are capable of binding to a single LD motif. In some instances, e.g. PIX and PAK3 binding to LD4 and E6, FAK and vinculin binding to recombinant paxillin, these proteins were shown to compete for binding to the LD motif (Tong et al., 1997; Hashimoto et al., 2001). Clearly the nature of the signaling complex will be different depending upon which binding partner successfully competes to bind paxillin. The affinity of each of the paxillin binding partners for the LD motif might be an important factor in determining the nature of the assembled complex. The affinity of interaction between FAK and the paxillin LD2 motif and LD4 motif has been estimated by surface plasmon resonance (Thomas et al., 1999b). The dissociation constants for the interaction of the C-terminal domain of FAK with the LD2 and LD4 motifs are 3–4 μM. The dissociation constant for the interaction of the C-terminal domain of FAK with a fragment of paxillin containing both LD2 and LD4 is less than 1 μM. Although the mechanism hasn't been established, this result demonstrates that the presence of two FAK binding sites on paxillin dramatically increases the affinity for FAK. This may explain why individual paxillin binding partners have evolved to associate with more than one LD motif. Future studies exploring the affinity of paxillin for its binding partners, elucidating the components found in individual paxillin containing complexes and understanding how these interactions are regulated are essential for determining how these complexes function.
Association of paxillin with integrins
Paxillin can bind in vitro to synthetic peptides mimicking the cytoplasmic domain of the β1 and β3 integrins (Schaller et al., 1995). Further under some circumstances paxillin can coimmunoprecipitate with integrins (Chen et al., 2000). Beyond these initial observations, little is known about the interaction between paxillin and the cytoplasmic tails of the β integrin subunits. Paxillin also binds to the cytoplasmic domain of the α4 integrin and the interaction appears significantly stronger than the reported interaction between paxillin and β integrin subunits (Liu et al., 1999). The interaction of paxillin with α4 has been demonstrated in vitro using synthetic peptides and in vivo by coimmunoprecipitation. Two α4 residues that are critical for paxillin binding have been identified (Liu and Ginsberg, 2000). The site of interaction of paxillin has not been defined, but the α4 cytoplasmic domain can bind both hic-5 and leupaxin in addition to paxillin. Thus the site of interaction is likely in sequences conserved between the three, most likely the LD motifs or LIM domains. The function of paxillin binding to α4 is apparently to regulate cell spreading and motility (Liu et al., 1999).
Phosphorylation of paxillin
Multiple stimuli induce tyrosine phosphorylation of paxillin
The first stimulus shown to induce tyrosine phosphorylation of paxillin was integrin-dependent cell adhesion to extracellular matrix proteins. Stimulation with growth factors, neuropeptides, ligands for G-protein coupled receptors, antigen and physical stress have all been reported to stimulate tyrosine phosphorylation of paxillin. Thus a large number of very diverse stimuli converge to stimulate paxillin phosphorylation. A selected list of stimuli shown to induce tyrosine phosphorylation of paxillin is presented in Table 4.
Phosphorylation of paxillin by tyrosine kinases
Although a large number of stimuli induce tyrosine phosphorylation of paxillin, only a few tyrosine kinases have been implicated in regulating tyrosine phosphorylation of paxillin. The focal adhesion kinase, FAK, and the FAK-related protein, CAKβ/Pyk2/CadTK/RAFTK, are major players in controlling tyrosine phosphorylation of paxillin. These proteins can colocalize with paxillin in vivo, physically associate with paxillin, and tyrosine phosphorylation/activation of these tyrosine kinases is frequently correlated with tyrosine phosphorylation of paxillin (Schlaepfer et al., 1999; Avraham et al., 2000). Overexpression of FAK and CAKβ promotes tyrosine phosphorylation of paxillin in vivo (Schaller and Parsons, 1995; Frisch et al., 1996; Schaller and Sasaki, 1997). Although fak−/− fibroblasts exhibit normal tyrosine phosphorylation of paxillin, this may be the result of compensatory expression of CAKβ in these cells (Ilic et al., 1995; Sieg et al., 1998; Owen et al., 1999). Inhibition of FAK signaling using a dominant negative strategy inhibits tyrosine phosphorylation of paxillin in fibroblasts (Richardson and Parsons, 1996). These lines of investigation provide compelling evidence that FAK/CAKβ regulate tyrosine phosphorylation of paxillin.
Members of the Src family of tyrosine kinases are also candidate kinases for the phosphorylation of paxillin. Paxillin was originally identified as a tyrosine phosphorylated protein in src-transformed fibroblasts (Glenney Jr and Zokas, 1989). Genetic evidence has also implicated Src family kinases in the regulation of tyrosine phosphorylation of paxillin under physiological conditions. Fibroblasts derived from csk−/− mice (Csk, the C-terminal Src kinase, is the major regulatory kinase of Src that functions to negatively regulate the catalytic activity of Src family kinases) exhibit enhanced activity of Src family kinases and increased tyrosine phosphorylation of their substrates including paxillin (Thomas et al., 1995). Further genetic evidence from the analysis of cells deficient for expression of Src family kinases also supports a role for these kinases in regulating tyrosine phosphorylation of paxillin. Fibroblasts lacking Src, Fyn and Yes exhibit dramatic reductions in adhesion dependent tyrosine phosphorylation of paxillin (Klinghoffer et al., 1999). Peritoneal macrophages from hck−/−/fgr−/− mice also exhibit defective tyrosine phosphorylation of paxillin upon adhesion to extracellular matrix proteins suggesting that these Src family kinases function in tyrosine phosphorylation of paxillin in macrophages (Suen et al., 1999). These results strongly implicate the Src family kinases in regulating paxillin phosphorylation.
The mechanism by which FAK/CAKβ and Src kinases induce tyrosine phosphorylation of paxillin may be complex. Activation of FAK/CAKβ results in the recruitment of Src kinases into complex with FAK and CAKβ. One substrate for the Src kinases is FAK/CAKβ and one consequence of Src-dependent phosphorylation is an elevation in the catalytic activity of FAK/CAKβ. Thus Src kinases could directly phosphorylate paxillin or indirectly promote paxillin phosphorylation by activation of FAK/CAKβ. In turn, FAK/CAKβ could directly phosphorylate paxillin or indirectly promote paxillin phosphorylation by recruiting Src kinases into complex, which then phosphorylate paxillin.
Tyrosine phosphorylation of paxillin is elevated in Bcr-Abl transformed 32D cells, suggesting that this oncogenic fusion protein might target paxillin for tyrosine phosphorylation (Salgia et al., 1995a). In normal fibroblasts, c-Abl may also contribute to tyrosine phosphorylation of paxillin. Upon adhesion, c-Abl transiently translocates from the nucleus to focal adhesions and associates with paxillin (Lewis and Schwartz, 1998). Further, c-Abl can phosphorylate paxillin in vitro (Lewis and Schwartz, 1998). Thus c-Abl potentially contributes to tyrosine phosphorylation of paxillin in response to cell adhesion.
Csk has also been proposed to phosphorylate paxillin since overexpression of Csk can result in enhanced paxillin phosphorylation (Sabe et al., 1994; Bergman et al., 1995; Takayama et al., 1999). However, the physiological significance of Csk-mediated tyrosine phosphorylation of paxillin is not clear.
Dephosphorylation of paxillin by tyrosine phosphatases
The regulation of dephosphorylation by tyrosine phosphatases may also be a very important mechanism of controlling paxillin signaling. Several PTPs have recently been implicated in regulating paxillin phosphorylation. PTP-PEST is a 125 kDa protein with an N-terminal catalytic domain and a long C-terminal domain containing proline-rich sequences and PEST sequences (Yang et al., 1993; Charest et al., 1995). Paxillin and hic-5 bind to one of the proline-rich regions of PTP-PEST (Shen et al., 1998; Cote et al., 1999; Nishiya et al., 1999; Shen et al., 2000). The PTP-PEST binding site on paxillin and hic-5 resides in the two C-terminal LIM domains (Cote et al., 1999; Nishiya et al., 1999; Shen et al., 2000). Overexpression of PTP-PEST results in dephosphorylation of paxillin and the association of paxillin with PTP-PEST via the proline-rich binding site is necessary for dephosphorylation (Shen et al., 2000). Thus the C-terminal domain of PTP-PEST functions as a docking site targeting paxillin for dephosphorylation. Further, PTP-PEST−/− fibroblasts exhibit enhanced tyrosine phosphorylation of paxillin (Angers-Loustau et al., 1999). It is noteworthy that PTP-PEST binds to the region of paxillin that functions in focal adhesion targeting. PTP-PEST is a cytoplasmic PTP that does not localize to focal adhesions (Charest et al., 1995). Thus, PTP-PEST may only associate with the population of paxillin that is not found in focal adhesions. This might be a mechanism to ensure dephosphorylation of paxillin as it exits from focal adhesions and maintain dephosphorylation of cytoplasmic pools of paxillin. Alternatively PTP-PEST might target non-focal adhesion-associated pools of paxillin that may function in other signaling pathways.
PTPΦ is also a candidate PTP for paxillin. Overexpression of PTPΦ in the BAC1.2F5 macrophage cell line reduces tyrosine phosphorylation of paxillin whereas overexpression of a catalytically inactive mutant of PTPΦ, which may act as a dominant negative mutant, enhances tyrosine phosphorylation of paxillin (Pixley et al., 2001). In addition, a substrate trapping mutant of PTPΦ can weakly associate with paxillin, which is usually indicative of an enzyme/substrate relationship (Pixley et al., 2001). Thus PTPΦ may also target paxillin for dephosphorylation.
Consequences of tyrosine phosphorylation of paxillin
Four sites of tyrosine phosphorylation of paxillin have been identified and a fifth potential site of phosphorylation suggested, all of which reside in the N-terminal domain of paxillin (see Figure 1). The major sites of tyrosine phosphorylation are tyrosines 31 and 118, whereas tyrosines 40 and 88 are other minor sites of phosphorylation (Schaller and Parsons, 1995; Bellis et al., 1995; Nakamura et al., 2000). Tyrosine 181 was also proposed as a site of phosphorylation, but has not definitely shown to become phosphorylated (Salgia et al., 1995b; Nakamura et al., 2000). The only known function of tyrosine phosphorylation of paxillin is to create binding sites for SH2 domain containing signaling proteins. Tyrosine residues 31, 118 and 181 are embedded in high affinity binding sites for the SH2 domain of the Crk adaptor protein and phosphorylation of these sites can result in the recruitment of Crk and CrkL into complex with paxillin (Birge et al., 1993; Schaller and Parsons, 1995; Salgia et al., 1995b). Crk and CrkL associate via their SH3 domains with other signaling molecules, e.g. the C3G guanine nucleotide exchange factor and Dock 180 (Feller et al., 1998). Therefore Crk and CrkL function in paxillin signaling to recruit additional signaling molecules into complex with paxillin. Csk and Chk, a Csk-related kinase, can also bind to paxillin via its SH2 domain (Sabe et al., 1994; Schaller and Parsons, 1995; Bergman et al., 1995; Grgurevich et al., 1999) (see Figure 5).
The SH2 domain of Src can also associate with paxillin in vitro (Schaller and Parsons, 1995). The Src family kinase Lck can associate with paxillin in vivo in T cells (Ostergaard et al., 1998). This interaction is direct and is mediated by the SH2 domain of Lck. Although the binding site on paxillin has not been firmly established, one of the minor sites of tyrosine phosphorylation, tyrosine 40, is within a sequence resembling a high affinity Src SH2 domain binding site (Songyang et al., 1993; Schaller and Parsons, 1995). It is also of interest that this tyrosine residue is very close to a proline rich sequence in paxillin (46PPPVPPPP53) that has been shown to bind the SH3 domain of Src in vitro (Weng et al., 1993). One possibility is that both the SH2 and SH3 domains of Lck may engage paxillin in vivo.
Whereas the LD motifs and LIM domains are highly conserved between paxillin, hic-5 and leupaxin, other sequences in the N-terminal domains of these proteins, including the sites of tyrosine phosphorylation in paxillin, are not conserved. Thus functions specific to each protein might be mediated by these unique sequences. Although hic-5 does not contain the Crk-binding sites in its N-terminus, tyrosine 60 is a site of phosphorylation and serves as a binding site for Csk (Ishino et al., 2000b). Recruitment of Csk could lead to further tyrosine phosphorylation of paxillin or to phosphorylation of Src family kinases to inhibit their catalytic activity.
Serine/threonine phosphorylation of paxillin
Although tyrosine phosphorylation of paxillin has been extensively examined, serine phosphorylation is a far more prevalent posttranslational modification. Paxillin becomes increasingly phosphorylated on serine when fibroblasts attach to fibronectin or when macrophages adhere to vitronectin (De Nichilo and Yamada, 1996; Bellis et al., 1997). In addition to adhesion-dependent serine phosphorylation, paxillin also becomes phosphorylated on serine/threonine residues during mitosis and upon stimulation of MCF-7 cells with heregulin (Yamakita et al., 1999; Vadlamudi et al., 1999).
Several serine/threonine kinases have been suggested to phosphorylate paxillin. The phosphorylation of paxillin in adhering macrophages might be mediated by PKC since inhibitors of PKC block serine phosphorylation of paxillin upon adhesion (De Nichilo and Yamada, 1996). However, in some cell types activation of PKC can also lead to Erk activation. PMA stimulation of EL4 cells induces a mobility shift in paxillin that is likely due to serine/threonine phosphorylation and this shift can be impaired with MEK inhibitors (Ku and Meier, 2000). Further paxillin can be phosphorylated in vitro with Erk (Ku and Meier, 2000). In MCF-7 cells, heregulin induces the activation of p38 and this serine/threonine kinase may be responsible for phosphorylation of paxillin on serine in response to heregulin (Vadlamudi et al., 1999). PAK3, which associates directly and indirectly with paxillin can also phosphorylate paxillin on serine residues in vitro (Hashimoto et al., 2001).
In addition to these known kinases, serine/threonine kinase activities can associate with and target three different fragments of paxillin for phosphorylation in vitro. A fragment of the N-terminus of paxillin (residues 168–191) can interact with a serine/threonine kinase that can phosphorylate two serine residues on paxillin, residues 188 and 190 (Bellis et al., 1997). Phosphopeptide mapping suggests that these sites might become phosphorylated upon adhesion of fibroblasts to fibronectin (Bellis et al., 1997). LIM domain 2 can bind a kinase that can phosphorylate paxillin on threonine 403 in vitro (Brown et al., 1998b). LIM domain 3 can bind a kinase that appears to phosphorylate serine residues 457 and 481 in vitro (Brown et al., 1998b). It has not been established under what conditions the sites of serine/threonine phosphorylation within the LIM domains of paxillin are phosphorylated in vivo.
Three outstanding issues remain in the analysis of serine/threonine phosphorylation of paxillin. Several phosphorylation sites have been identified, but additional studies are required to determine if additional sites of phosphorylation exist. Although a number of candidate kinases for paxillin phosphorylation have been identified, further studies are required to define which kinases control paxillin phosphorylation in response to different stimuli. Finally, it is not yet clear what the consequences of serine/threonine phosphorylation of paxillin may be. There is some evidence that serine phosphorylation within LIM domain 3 might function in regulating localization of paxillin (Brown et al., 1998b). Further, overexpression of paxillin mutants that cannot be phosphorylated on LIM domain phosphorylation sites slightly retards adhesion to fibronectin, suggesting that paxillin phosphorylation might regulate cell adhesion (Brown et al., 1998b). While these results may provide the first clues that phosphorylation of these sites regulate localization and adhesion, additional studies are required to fully elucidate the function of phosphorylation of paxillin on serine/threonine residues.
How does paxillin localize to focal adhesions?
The major focal adhesion targeting sequence of paxillin is within LIM domains 2 and 3 (Brown et al., 1996). These LIM domains presumably function in mediating protein–protein interactions to target and or tether paxillin in focal adhesions, although the binding partner remains elusive. Although LIM domain binding partners for paxillin and hic-5 have been described, none of these binding partners are found in focal adhesions. As described above, the N-terminal half of paxillin contains binding sites for two focal adhesion-associated proteins, FAK and vinculin, but association with these proteins is not the major mechanism of focal adhesion targeting since they bind outside of the major targeting sequence of paxillin. However, interactions with these proteins, most likely vinculin, may play a minor role in localization of paxillin to focal adhesions, since the N-terminus of paxillin, lacking the LIM domains, weakly targets to focal adhesions (Thomas et al., 1999a).
One surprising finding related to focal adhesion targeting of paxillin came from the analysis of streptolysin-O permeabilized Swiss 3T3 cells. Treatment of permeabilized cells with GTP-γ-S promoted the translocation of paxillin from a perinuclear location to structures resembling focal adhesions (Norman et al., 1998). This effect was rho-independent, since activated rho did not promote the same robust change in the location of paxillin and the GTP-γ-S dependent translocation of paxillin was unaffected by C3 toxin. Introduction of ARF1 into permeabilized Swiss 3T3 cells or microinjection of ARF1 into intact cells induced the translocation of paxillin to focal adhesion like structures suggesting that ARF1 could control targeting of paxillin to focal adhesions (Norman et al., 1998). The ARF family of GTP-binding proteins, including ARF1, generally function in membrane/vesicle trafficking, although ARF6 may play a role in regulating the actin cytoskeleton (Chavrier and Goud, 1999; Donaldson and Jackson, 2000). A role for ARF proteins in regulating localization of paxillin to focal adhesions is further supported by studies where ARF-GAP proteins are overexpressed. Overexpression of PAG3, Git2s, Git1 or ASAP1 impair the localization of paxillin to focal adhesions (Zhao et al., 2000; Kondo et al., 2000; Randazzo et al., 2000; Mazaki et al., 2001). Although three of these ARF-GAPs are paxillin associated, and the fourth is found in focal adhesions, it is not clear whether overexpression of these proteins is mimicking the function of the endogenous protein or indiscriminately inhibiting ARF function. At present, the mechanism by which ARF proteins may regulate the translocation of paxillin to focal adhesions is completely speculative. It is intriguing that ARF-GAPs associate with paxillin and could further function in regulating localization of paxillin. It is also interesting to note that PKL and Git1 can localize in focal adhesions (Turner et al., 1999; Zhao et al., 2000), whereas Git2s and PAG3 do not (Kondo et al., 2000; Mazaki et al., 2001). Perhaps association with different ARF-GAPs serves to regulate paxillin localization. Binding to Git2s or PAG3 in a non-focal adhesion compartment might inhibit ARF-dependent translocation to focal adhesions, maintaining a cytoplasmic population of paxillin.
Does paxillin have a nuclear function?
LIM domains were originally discovered as a motif that was shared by three transcription factors, lin 11, isl-1 and mec-3. Given that these domains are double zinc fingers and that zinc fingers are important elements in DNA binding proteins like steroid receptors, the possibility that some of the LIM domain containing, focal adhesion-associated proteins might also perform a nuclear function was an attractive hypothesis. In fact, zyxin, another focal adhesion-associated LIM protein, was proposed to shuttle between focal adhesions and the nucleus (Nix and Beckerle, 1997). Several lines of evidence suggest that hic-5, and perhaps paxillin as well, might have a nuclear function. First, a small fraction of hic-5 and paxillin have been reported to localize to the nucleus and even associate with the nuclear matrix, and this localization was apparently mediated via the C-terminal, LIM domain-containing half of the protein (Thomas et al., 1999a; Yang et al., 2000). Second, the LIM domains of hic-5 have been reported to bind DNA, although no consensus binding motif for hic-5 has been defined (Nishiya et al., 1998). Third, two different two hybrid screens, one using the androgen receptor and the other using the glucocorticoid receptor as bait, have identified hic-5 as a steroid receptor binding protein. Importantly, coexpression of hic-5 with steroid receptors enhances expression of reporter genes in vivo, suggesting that hic-5 can potentiate the activation of gene expression by these transcription factors (Fujimoto et al., 1999a; Yang et al., 2000). While the hypothesis that hic-5 functions in both the focal adhesions and the nucleus is tantalizing, the importance of hic-5 in signaling to the nucleus is not firmly established and a similar role for paxillin is presently only speculative. Nevertheless, these findings seem sufficiently encouraging that further investigation is warranted.
Biological functions of paxilllin
Paxillin plays a critical role in embryonic development since paxillin−/− mice exhibit embryonic lethality by day 9.5 (S Thomas, personal communication). Analysis of paxillin−/− fibroblasts reveals defects in cell spreading and in cell motility (S Thomas, personal communication). Additional studies using other approaches to address the role of paxillin in the cell have produced similar conclusions, that paxillin controls spreading and motility. Expression of a dominant negative mutant of FAK impairs endogenous FAK signaling and delays cell spreading (Richardson and Parsons, 1996). The effects of this dominant negative mutant can be reversed by overexpression of FAK or Src (Richardson et al., 1997). There is a good correlation in this series of experiments between reduced tyrosine phosphorylation of paxillin and delayed spreading.
To address the role of tyrosine phosphorylation of paxillin in controlling cell motility, two different groups have utilized tyrosine to phenylalanine substitution mutants of paxillin as dominant negative mutants. Expression of a paxillin mutant with phenylalanine substitutions for both tyrosine 31 and 118 in the NBT-II rat bladder carcinoma cell line impaired random motility on collagen by approximately 50% (Petit et al., 2000). In NBT-II cells, adhesion to collagen promotes the association of Crk with paxillin and overexpression of Crk enhances random motility in NBT-II cells. Further, overexpression of Crk rescues the motility defect in NBT-II cells stably expressing the tyrosine to phenylalanine mutants of paxillin (Petit et al., 2000). These results suggest that tyrosine phosphorylation of paxillin promotes cell motility by recruiting Crk to sites of paxillin tyrosine phosphorylation. In contrast to these findings, overexpression of wild type paxillin in Cos7 cells or NMuMG cells impaired haptotaxis toward collagen (Yano et al., 2000). Further, overexpression of wild type paxillin inhibited the ability of LPA stimulated MM-1 rat hepatoma tumor cells to invade a monolayer of primary mesothelial cells (Yano et al., 2000). A paxillin mutant with substitutions for the four sites of tyrosine phosphorylation on paxillin had no effect or even enhanced motility/invasion in these assays (Yano et al., 2000). These results suggest that tyrosine phosphorylation of paxillin inhibits cell motility. It is perhaps noteworthy that Crk and CrkI, were shown to associate with p130cas in the NMuMG cells but not with paxillin. It is difficult to reconcile these two disparate findings. Differences in cell type or the experimental design to measure motility in the two studies may be the reason for the opposing results. Alternatively, tyrosine phosphorylation of paxillin might have opposing roles in random motility and haptotactic motility. Further investigation will be required to resolve this issue.
Paxillin has also been implicated in the control of cell spreading and motility by the α4 integrin subunit. The α4 integrin functions to inhibit cell spreading and promote motility when expressed exogenously in cells (Liu et al., 1999). Paxillin plays an essential role in the inhibition of cell spreading by α4 since a α4 mutant that is defective for paxillin binding fails to block spreading. Further, the α4 integrin does not block the spreading of paxillin−/− fibroblasts, but can block the spreading of paxillin−/− fibroblasts that have been engineered to re-express paxillin.
Since the original discovery of paxillin as a tyrosine phosphorylated Src substrate, a number of the details of the function of paxillin have been elucidated. It is now well established that paxillin is a focal adhesion associated adaptor protein that recruits signaling molecules into a complex. Some paxillin-binding partners associate via LD motif interactions, whereas the interactions of others are regulated by phosphotyrosine-dependent interactions. The functions of paxillin appear to include the regulation of cell spreading and cell motility, and these functions are presumably elicited by coordinating signaling events by paxillin associated proteins. Future challenges in the area of paxillin signaling are to define the molecular pathways by which individual paxillin binding partners can affect spreading and or motility, and to determine how the signals from multiple proteins associated with paxillin are integrated to produce a coordinated biological response.
Abedi H, Dawes KE, Zachary I . 1995 J. Biol. Chem. 270: 11367–11376
Andreev J, Simon JP, Sabatini DD, Kam J, Plowman G, Randazzo PA, Schlessinger J . 1999 Mol. Cell Biol. 19: 2338–2350
Angers-Loustau A, Cote JF, Charest A, Dowbenko D, Spencer S, Lasky LA, Tremblay ML . 1999 J. Cell Biol. 144: 1019–1031
Avraham H, Park SY, Schinkmann K, Avraham S . 2000 Cell Signal. 12: 123–133
Bellis SL, Miller JT, Turner CE . 1995 J. Biol. Chem. 270: 17437–17441
Bellis SL, Perrotta JA, Curtis MS, Turner CE . 1997 Biochem. J. 325: 375–381
Berg MM, Krafft GA, Klein WL . 1997 J. Neurosci. Res. 50: 979–989
Bergman M, Joukov V, Virtanen I, Alitalo K . 1995 Mol. Cell Biol. 15: 711–722
Birge RB, Fajardo JE, Reichman C, Shoelson SE, Songyang Z, Cantley LC, Hanafusa H . 1993 Mol. Cell Biol. 13: 4648–4656
Bocciardi R, Mograbi B, Pasini B, Borrello MG, Pierotti MA, Bourget I, Fischer S, Romeo G, Rossi B . 1997 Oncogene 15: 2257–2265
Brown MC, Curtis MS, Turner CE . 1998a Nat. Struct. Biol. 5: 677–678
Brown MC, Perrotta JA, Turner CE . 1996 J. Cell Biol. 135: 1109–1123
Brown MC, Perrotta JA, Turner CE . 1998b Mol. Biol. Cell 9: 1803–1816
Burridge K, Chrzanowska-Wodnicka M . 1996 Annu. Rev. Cell Dev. Biol. 12: 463–518
Burridge K, Turner CE, Romer LH . 1992 J. Cell Biol. 119: 893–903
Charest A, Wagner J, Shen SH, Tremblay ML . 1995 Biochem. J. 308: 425–432
Chavrier P, Goud B . 1999 Curr. Opin. Cell Biol. 11: 466–475
Chen LM, Bailey D, Fernandez-Valle C . 2000 J. Neurosci. 20: 3776–3784
Cooley MA, Broome JM, Ohngemach C, Romer LH, Schaller MD . 2000 Mol. Biol. Cell 11: 3247–3263
Cote JF, Turner CE, Tremblay ML . 1999 J. Biol. Chem. 274: 20550–20560
Das K, Bohl J, Vande Pol SB . 2000 J. Virol. 74: 812–816
Dawid IB, Breen JJ, Toyama R . 1998 Trends Genet. 14: 156–162
De Nichilo MO, Yamada KM . 1996 J. Biol. Chem. 271: 11016–11022
Dedhar S . 2000 Curr. Opin. Cell Biol. 12: 250–256
Donaldson JG, Jackson CL . 2000 Curr. Opin. Cell Biol. 12: 475–482
Feller SM, Posern G, Voss J, Kardinal C, Sakkab D, Zheng J, Knudsen BS . 1998 J. Cell Physiol 177: 535–552
Fernandez R, Boxer LA, Suchard SJ . 1997 J. Immunol. 159: 5568–5575
Flinn HM, Ridley AJ . 1996 J. Cell Sci. 109: 1133–1141
Fox GL, Rebay I, Hynes RO . 1999 Proc. Natl. Acad. Sci. USA 96: 14978–14983
Frisch SM, Vuori K, Ruoslahti E, Chan-Hui PY . 1996 J. Cell Biol. 134: 793–799
Fujimoto J, Sawamoto K, Okabe M, Takagi Y, Tezuka T, Yoshikawa S, Ryo H, Okano H, Yamamoto T . 1999b J. Biol. Chem. 274: 29196–29201
Fujimoto N, Yeh S, Kang HY, Inui S, Chang HC, Mizokami A, Chang C . 1999a J. Biol. Chem. 274: 8316–8321
Fuortes M, Jin WW, Nathan C . 1994 J. Cell Biol. 127: 1477–1483
Gilmore AP, Burridge K . 1996 Nature 381: 531–535
Glenney Jr JR, Zokas L . 1989 J. Cell Biol. 108: 2401–2408
Graham IL, Anderson DC, Holers VM, Brown EJ . 1994 J. Cell Biol. 127: 1139–1147
Grgurevich S, Mikhael A, McVicar DW . 1999 Biochem. Biophys. Res. Commun. 256: 668–675
Hagmann J, Grob M, Welman A, van Willigen G, Burger MM . 1998 J. Cell Sci. 111: 2181–2188
Hamawy MM, Swaim WD, Minoguchi K, de Feijter AW, Mergenhagen SE, Siraganian RP . 1994 J. Immunol. 153: 4655–4662
Hashimoto S, Tsubouchi A, Mazaki Y, Sabe H . 2001 J. Biol. Chem. 276: 6037–6045
Hildebrand JD, Schaller MD, Parsons JT . 1995 Mol. Biol. Cell 6: 637–647
Hiregowdara D, Avraham H, Fu Y, London R, Avraham S . 1997 J. Biol. Chem. 272: 10804–10810
Ilic D, Furuta Y, Kanazawa S, Takeda N, Sobue K, Nakatsuji N, Nomura S, Fujimoto J, Okada M, Yamamoto T . 1995 Nature 377: 539–544
Ishino K, Kaneyama, Shibanuma M, Nose K . 2000a J. Cell Biochem. 76: 411–419
Ishino M, Aoto H, Sasaski H, Suzuki R, Sasaki T . 2000b FEBS Lett. 474: 179–183
Khan MA, Okumura N, Okada M, Kobayashi S, Nakagawa H . 1995 FEBS Lett. 362: 201–204
Kintscher U, Wakino S, Kim S, Fleck E, Hsueh WA, Law RE . 2001 Hypertension 37: 587–593
Klinghoffer RA, Sachsenmaier C, Cooper JA, Soriano P . 1999 EMBO J. 18: 2459–2471
Kondo A, Hashimoto S, Yano H, Nagayama K, Mazaki Y, Sabe H . 2000 Mol. Biol. Cell. 11: 1315–1327
Ku H, Meier KE . 2000 J. Biol. Chem. 275: 11333–11340
Leduc I, Meloche S . 1995 J. Biol. Chem. 270: 4401–4404
Lewis JM, Schwartz MA . 1998 J. Biol. Chem. 273: 14225–14230
Li X, Earp HS . 1997 J. Biol. Chem. 272: 14341–14348
Lipsky BP, Beals CR, Staunton DE . 1998 J. Biol. Chem. 273: 11709–11713
Litvak V, Tian D, Shaul YD, Lev S . 2000 J. Biol. Chem. 275: 32736–32746
Liu S, Ginsberg MH . 2000 J. Biol. Chem. 275: 22736–22742
Liu S, Thomas SM, Woodside DG, Rose DM, Kiosses WB, Pfaff M, Ginsberg MH . 1999 Nature 402: 676–681
Mazaki Y, Hashimoto S, Okawa K, Tsubouchi A, Nakamura K, Yagi R, Yano H, Kondo A, Iwamatsu A, Mizoguchi A, Sabe H . 2001 Mol. Biol. Cell 12: 645–662
Mazaki Y, Hashimoto S, Sabe H . 1997 J. Biol. Chem. 272: 7437–7444
Mazaki Y, Uchida H, Hino O, Hashimoto S, Sabe H . 1998 J. Biol. Chem. 273: 22435–22441
Murakami H, Iwashita T, Asai N, Iwata Y, Narumiya S, Takahashi M . 1999 Oncogene 18: 1975–1982
Nakamura K, Yano H, Uchida H, Hashimoto S, Schaefer E, Sabe H . 2000 J. Biol. Chem. 275: 27155–27164
Needham LK, Rozengurt E . 1998 J. Biol. Chem. 273: 14626–14632
Nikolopoulos SN, Turner CE . 2000 J. Cell Biol. 151: 1435–1448
Nikolopoulos SN, Turner CE . 2001 J. Biol. Chem. 276: 23499–23505
Nishiya N, Iwabuchi Y, Shibanuma M, Cote JF, Tremblay ML, Nose K . 1999 J. Biol. Chem. 274: 9847–9853
Nishiya N, Sabe H, Nose K, Shibanuma M . 1998 Nucleic Acids Res. 26: 4267–4273
Nix DA, Beckerle MC . 1997 J. Cell Biol. 138: 1139–1147
Norman JC, Jones D, Barry ST, Holt MR, Cockcroft S, Critchley DR . 1998 J. Cell Biol. 143: 1981–1995
Ostergaard HL, Lou O, Arendt CW, Berg NN . 1998 J. Biol. Chem. 273: 5692–5696
Owen JD, Ruest PJ, Fry DW, Hanks SK . 1999 Mol. Cell Biol. 19: 4806–4818
Palmer RH, Fessler LI, Edeen PT, Madigan SJ, McKeown M, Hunter T . 1999 J. Biol. Chem. 274: 35621–35629
Petit V, Boyer B, Lentz D, Turner CE, Thiery JP, Valles AM . 2000 J. Cell Biol. 148: 957–970
Pixley FJ, Lee PS, Condeelis JS, Stanley ER . 2001 Mol. Cell Biol. 21: 1795–1809
Randazzo PA, Andrade J, Miura K, Brown MT, Long YQ, Stauffer S, Roller P, Cooper JA . 2000 Proc. Natl. Acad. Sci. USA 97: 4011–4016
Rankin S, Rozengurt E . 1994 J. Biol. Chem. 269: 704–710
Richardson A, Malik RK, Hildebrand JD, Parsons JT . 1997 Mol. Cell Biol. 17: 6906–6914
Richardson A, Parsons T . 1996 Nature 380: 538–540
Ryu H, Lee JH, Kim KS, Jeong SM, Kim PH, Chung HT . 2000 J. Immunol. 165: 2116–2123
Sabe H, Hata A, Okada M, Nakagawa H, Hanafusa H . 1994 Proc. Natl. Acad. Sci. USA 91: 3984–3988
Salgia R, Avraham S, Pisick E, Li JL, Raja S, Greenfield EA, Sattler M, Avraham H, Griffin JD . 1996 J. Biol. Chem. 271: 31222–31226
Salgia R, Li JL, Lo SH, Brunkhorst B, Kansas GS, Sobhany ES, Sun Y, Pisick E, Hallek M, Ernst T . 1995a J. Biol. Chem. 270: 5039–5047
Salgia R, Uemura N, Okuda K, Li JL, Pisick E, Sattler M, de Jong R, Druker B, Heisterkamp N, Chen LB . 1995b J. Biol. Chem. 270: 29145–29150
Schaller MD, Borgman CA, Parsons JT . 1993 Mol. Cell Biol. 13: 785–791
Schaller MD, Otey CA, Hildebrand JD, Parsons JT . 1995 J. Cell Biol. 130: 1181–1187
Schaller MD, Parsons JT . 1995 Mol. Cell Biol. 15: 2635–2645
Schaller MD, Sasaki T . 1997 J. Biol. Chem. 272: 25319–25325
Schlaepfer DD, Hauck CR, Sieg DJ . 1999 Prog. Biophys. Mol. Biol. 71: 435–478
Seko Y, Takahashi N, Sabe H, Tobe K, Kadowaki T, Nagai R . 1999 Biochem. Biophys. Res. Commun. 262: 290–296
Seufferlein T, Rozengurt E . 1994 J. Biol. Chem. 269: 27610–27617
Shen Y, Lyons P, Cooley M, Davidson D, Veillette A, Salgia R, Griffin JD, Schaller MD . 2000 J. Biol. Chem. 275: 1405–1413
Shen Y, Schneider G, Cloutier JF, Veillette A, Schaller MD . 1998 J. Biol. Chem. 273: 6474–6481
Shibanuma M, Mashimo J, Kuroki T, Nose K . 1994 J. Biol. Chem. 269: 26767–26774
Sieg DJ, Ilic D, Jones KC, Damsky CH, Hunter T, Schlaepfer DD . 1998 EMBO J. 17: 5933–5947
Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG, King F, Roberts T, Ratnofsky S, Lechleider RJ . 1993 Cell 72: 767–778
Suen PW, Ilic D, Caveggion E, Berton G, Damsky CH, Lowell CA . 1999 J. Cell Sci. 112: 4067–4078
Tachibana K, Sato T, D'Avirro N, Morimoto C . 1995 J. Exp. Med. 182: 1089–1099
Takayama Y, Tanaka S, Nagai K, Okada M . 1999 J. Biol. Chem. 274: 2291–2297
Tang D, Mehta D, Gunst SJ . 1999 Am. J. Physiol. 276: C250–C258
Thomas JW, Cooley MA, Broome JM, Salgia R, Griffin JD, Lombardo CR, Schaller MD . 1999b J. Biol. Chem. 274: 36684–36692
Thomas SM, Hagel M, Turner CE . 1999a J. Cell Sci. 112: 181–190
Thomas SM, Soriano P, Imamoto A . 1995 Nature 376: 267–271
Tong X, Howley PM . 1997 Proc. Natl. Acad. Sci. USA 94: 4412–4417
Tong X, Salgia R, Li JL, Griffin JD, Howley PM . 1997 J. Biol. Chem. 272: 33373–33376
Turner CE, Brown MC, Perrotta JA, Riedy MC, Nikolopoulos SN, McDonald AR, Bagrodia S, Thomas S, Leventhal PS . 1999 J. Cell Biol. 145: 851–863
Turner CE, Glenney Jr JR, Burridge K . 1990 J. Cell Biol. 111: 1059–1068
Turner CE, Miller JT . 1994 J. Cell Sci. 107: 1583–1591
Turner CE, Pietras KM, Taylor DS, Molloy CJ . 1995 J. Cell Sci. 108: 333–342
Vadlamudi R, Adam L, Talukder A, Mendelsohn J, Kumar R . 1999 Oncogene 18: 7253–7264
Vande Pol SB, Brown MC, Turner CE . 1998 Oncogene 16: 43–52
Wang F, Nohara K, Olivera A, Thompson EW, Spiegel S . 1999 Exp. Cell Res. 247: 17–28
Weng Z, Taylor JA, Turner CE, Brugge JS, Seidel-Dugan C . 1993 J. Biol. Chem. 268: 14956–14963
Wheeler GN, Hynes RO . 2001 Gene 262: 291–299
Williams LM, Ridley AJ . 2000 J. Immunol. 164: 2028–2036
Wood CK, Turner CE, Jackson P, Critchley DR . 1994 J. Cell Sci. 107: 709–717
Yamakita Y, Totsukawa G, Yamashiro S, Fry D, Zhang X, Hanks SK, Matsumura F . 1999 J. Cell Biol. 144: 315–324
Yang L, Guerrero J, Hong H, DeFranco DB, Stallcup MR . 2000 Mol. Biol. Cell 11: 2007–2018
Yang Q, Co D, Sommercorn J, Tonks NK . 1993 J. Biol. Chem. 268: 6622–6628
Yano H, Uchida H, Iwasaki T, Mukai M, Akedo H, Nakamura K, Hashimoto S, Sabe H . 2000 Proc. Natl. Acad. Sci. USA 97: 9076–9081
Zachary I, Sinnett-Smith J, Turner CE, Rozengurt E . 1993 J. Biol. Chem. 268: 22060–22065
Zhang Z, Hernandez-Lagunas L, Horne WC, Baron R . 1999 J. Biol. Chem. 274: 25093–25098
Zhao ZS, Manser E, Loo TH, Lim L . 2000 Mol. Cell Biol. 20: 6354–6363
Zheng C, Xing Z, Bian ZC, Guo C, Akbay A, Warner L, Guan JL . 1998 J. Biol. Chem. 273: 2384–2389
Zhu T, Goh EL, Lobie PE . 1998 J. Biol. Chem. 273: 10682–10689
Research in the author's laboratory is supported by grants from the American Cancer Society (RPG-96-021-04-CSM) and the National Institutes of Health (GM57943 and CA90901).
About this article
Journal of Cellular Physiology (2019)
Oleic acid induces migration through a FFAR1/4, EGFR and AKT-dependent pathway in breast cancer cells
Endocrine Connections (2019)
Pathology - Research and Practice (2019)
Journal of Cell Science (2018)
Preparation of Cytocompatible ITO Neuroelectrodes with Enhanced Electrochemical Characteristics Using a Facile Anodic Oxidation Process
Advanced Functional Materials (2018)