Fig. 1: Overexpression of p120 in NIH3T3 fibroblasts induces a dramatic “branching” phenotype. p120 is shown in green, filamentous actin in red.
p120, actin and Rho GTPases
Overexpression of p120 typically has a variety of morphological effects, depending on the cell type and the level of p120 overexpression. In fibroblasts, high-level overexpression causes a striking "branching" phenotype characterized by extreme arborization of cellular processes (see Anastasiadis et al., Nature Cell Biol, 2000). The term "branching" is used here to describe the cellular morphology and not the process by which arborization occurs. Less pronounced effects, including increased formation of lamellipodia, are observed in most epithelial cell types. The branching phenotype is also observed when p120 is overexpressed in cadherin-deficient cells. When overexpressed in normal cells, p120 rapidly saturates available cadherin-binding sites, and then accumulates in the cytoplasm. Thus, the branching morphology appears to be caused by cytoplasmic p120 and is potentially independent of cadherin presence.
Using the branching phenotype as an assay, we showed that p120 can potently inhibit RhoA (Anastasiadis et al., Nature Cell Biol, 2000). This conclusion is consistent with the observation that p120 overexpression perturbs actin stress fibers, structures which are dependent on RhoA activity. In addition, other manipulations that inhibit RhoA, including overexpression of p190 RhoGAP (Rho GTPase activating protein) or C3 exotransferase, induced similar branching phenotypes. Finally, we argued that in the context of the branching assay, the contribution of the related GTPases Rac and Cdc42 was relatively minor. However, in addition to inhibiting RhoA, p120 can also activate Rac and Cdc42. p120 overexpression was shown to block stress fiber-mediated contractility and maturation of focal contacts, effects that could be reversed by expression of a dominant active RhoA mutant. These effects of ectopically-expressed p120 are blocked by expression of various cadherin constructs that sequester the excess p120, suggesting that p120 can affect Rho GTPases only in the E-cadherin unbound state. Perhaps an important aspect of p120 function is the ability to shuttle between E-cadherin-bound and cytoplasmic pools. Therefore, the low affinity of p120 for cadherin binding (relative to β-catenin) may be a key feature of its mechanism of action. Another key feature relevant to cancer, is that upon E-cadherin loss during tumor progression, p120 mislocalizes to the cytoplasm, unlike β-catenin, which is rapidly degraded (See Thoreson et al., J. Cell Biol, 2000).
Our recent data indicate that depletion of endogenous p120 with a retrovirally transduced shRNA causes increased RhoA activity, which can be rescued by ectopic expression of full-length p120. Similarly, p120 depletion decreases basal or HGF-induced Rac activation. Combined with the p120 overexpression data, these results indicate that p120 is an important physiological modulator of Rho GTPases. In agreement with this conclusion, p120 overexpression or depletion in Xenopus embryos causes gastrulation defects that can be rescued by constitutively active, or dominant negative RhoA mutants, respectively (See Fang et al., J. Cell Biol, 2004). Furthermore, in mice, conditional knockout of p120 in the skin causes Rho-dependant, NFkB-mediated inflammation, confirming again the importance of endogenous p120 in the regulation of Rho activity.
Despite general overall agreement, recent studies differ in several details that provide substance for the next generation of experiments. A major issue is whether p120 interacts directly with the GTPases (RhoA, Rac, etc.), or indirectly, via proteins that regulate GTPase function. GTP exchange assays in vitro suggested that p120 inhibits the intrinsic GDP/GTP exchange activity of RhoA in a manner comparable to that of the well-characterized Rho inhibitor, Guanine Nucleotide Dissociation Inhibitor (GDI). p120-mediated inhibition of nucleotide exchange by RhoA has been demonstrated using highly purified proteins, implying a direct interaction between p120 and RhoA. This observation suggests a relatively simple mechanism of action whereby p120 sequesters the GTPase in an inactive state. However, the interaction between p120 and RhoA is quite weak in co-immunoprecipitation experiments, leaving open the possibility of more complicated mechanisms.
A different mechanism was suggested by the observation that p120 interacts directly with Vav2, a RhoGEF (Rho GTPase exchange factor), which could account for the ability of p120 to activate Rac and Cdc42. Because Rac and Cdc42 can inhibit RhoA in some cells, it is possible that inhibition of RhoA by p120 occurs indirectly via activation of Rac and Cdc42. In support of this idea, a Vav2 deletion mutant can repress p120-induced branching. However, overexpression of wt Vav2, or dominant active forms of Rac or Cdc42, do not mimic the effects of p120 overexpression, making it less likely that this mechanism accounts for p120-induced branching and inhibition of RhoA activity.
p120 and microtubules
While there is general agreement that p120 increases cell motility by acting on Rho GTPases, some important mechanistic questions remain unanswered. One of these questions relates to the ability of ectopically expressed p120 to promote directional movement, while the other relates to the subcellular location of p120's action. Directional movement requires cell polarization, whereby a cell aquires a leading edge and a trailing end. Rho GTPases are believed to be important in this process. Rac and Cdc42 activities are thought to mediate protrusive events at the leading edge, while RhoA promotes motility by increasing contractility over the main cell body. Therefore, to promote motility, p120 needs to affect Rho GTPase activities locally and in a coordinated manner.
In addition, Rho GTPases are thought to be sequestered by RhoGDIs (GDP dissociation inhibitors) and kept at an inactive state in the cytoplasm. To be activated, Rho GTPases need to interact with exchange factors (RhoGEFs). However, RhoGEFs have lower affinity for Rho GTPases than GDIs, implying the existence of an intermediate step for Rho GTPase nucleotide exchange and activation. That step may involve targeting of the Rho GTPase/GDI complex to appropriate cellular sites, regulating the affinity of the complex and assuring activation by local GEFs. This seems to be the case in the activation of Rho GTPases at the plasma membrane via the CD44 transmembrane receptor. Therefore, to activate Rac and Cdc42, p120 may need to act at particular subcellular sites.
Fig. 2: Ectopically-expressed p120 (green) colocalizes with endogenous tubulin (red). The microtubule-disrupting agent nocodazole blocks co-localization.
Most migrating cells require an intact microtubule network to initiate and maintain directional movement. Integrin-mediated formation of focal contacts at the cell's leading edge, maturation of focal contacts to focal adhesions in the cell body, and dissolution of mature adhesions at the cell's rear are important for cell motility. Microtubules are thought to be important in promoting the polarized phenotype of motile cells, and may play a supporting role in promoting Rac activation and formation of a protruding lamellipodium. Microtubule filaments are polarized themselves, with a minus end attached to the centrosome and a plus end towards the cell periphery.
At the cell body, microtubules are thought to promote the establishment of mature focal adhesions, while at the trailing end, microtubules affect focal adhesion turnover, through a mechanism that is not well understood. Recent data argue that these effects of microtubules on the cell's body, as well as on leading and trailing ends, are mediated by differential regulation of Rho GTPases and require the function of a conventional kinesin. Kinesin is the prototypic member of a family of motor proteins that function to transport cargo on microtubule filaments. The mechanism by which microtubules affect Rho GTPases is still unclear. Nonetheless, several RhoGEFs associate with microtubules, and kinesin was suggested to promote the delivery of these GEFs to their site of action. Therefore, microtubules and kinesin may be essential for either the sequestration or the delivery of Rho regulators, thus modulating Rho GTPase activities and cell motility.
p120 slides on microtubules towards their plus ends. GFP-p120 (green) and Cy3-tubulin (red). (Yanagisawa et al., J Biol Chem, 2004).
Our recent data show that ectopically expressed p120 co-localizes with the microtubule network and associates with kinesin (movie 1 shows p120-GFP colocalization with tubulin, while movie 2 shows p120-GFP gliding along a microtubule). The kinesin interaction has important functional implications, as it could allow p120 to act in a localized and regulated manner towards Rho GTPases, in order to promote cell motility. The potential functional significance of this interaction is further underscored by preliminary data suggesting that kinesin binding blocks the ability of p120 to induce a branching morphology and loss of stress fibers in fibroblasts (Yanagisawa et al., J Biol Chem, 2004). Current projects focus on the possible role of the p120-kinesin interaction in Rho GTPase activation and the promotion of cell motility. We also plan to expand on our initial observations that the p120-kinesin interaction regulates the nuclear trafficking of p120, possibly regulating its nuclear functions.