Discussion
It is well established that the main function of RhoGDIs is to maintain Rho GTPases in inactive soluble complexes. In many canonical models of the Rho GTPase cycle, GDIs extract GTPases from the membrane and sequester them as inactive cytosolic complexes. RhoGDIs are therefore predominantly thought to act as negative regulators; however, they inhibit both activation [7] and inactivation [8,16,17] of GTPases. Little is known about how the opposing functions of GDIs influence the Rho GTPase cycle.
In the present study, we constructed a model of the Rho GTPase cycle, designated as the GDI-integrated model, in which GDIs inhibit the activities of GEFs and GAPs by interacting with them in addition to sequestering the Rho GTPases. This model indicated that GDIs sustain the activation of Rho GTPase by interacting with GAPs. Furthermore, as expected from the positive regulatory role of GDIs, (in other words, the inhibition of GAP activity by GDIs), an increase in the intracellular concentration of free GDIs enhanced the prolongation of Rho activation despite the overall decrease in the Rho activation level.
It was previously reported that the molar amount of RhoGDIα is roughly equal to the molar total of the RhoA, Rac1, and Cdc42 GTPases in several types of cultured cells [28]. RhoGDIβ is strongly expressed in hematopoietic cell lineages [39,40] and in other cell types [41-43], though it is not as ubiquitous as RhoGDIα. Specificities of RhoGDIs for Rho GTPases are largely unknown, but both RhoGDIα and RhoGDIβ can associate with RhoA, Cdc42, and Rac1 in some cell types [3]. Therefore, in RhoGDIβ-expressing cells, the total amount of RhoGDIα and RhoGDIβ may exceed the total amount of these Rho GTPases. In such cases, RhoGDIs may exist in a non-GTPase-complexed form in the cell and may function to sustain Rho activation for long periods.
RhoGDIβ is implicated in cancer progression, however, reports have presented contradictory evidence as to the nature of the correlation between cancer progression and RhoGDIβ expression level [44]. We have also reported that RhoGDIβ plays a positive [41,45,46] and negative [47] role in cancer progression. Several explanations for this contradictory behavior of RhoGDIβ have been proposed [44]. Our present study suggests that RhoGDI can act both as a positive and negative regulator of GTPases, and which role RhoGDI plays may depend on its expression level. This presents at least a partial explanation for the inconsistent correlation of RhoGDIβ with cancer progression.
It has been proposed that intracellular signals are transmitted through the dynamic activities of signaling molecules (defined as the temporal change in activity of a molecule) [48]. For example, in the case of ERK (extracellular signal-regulated kinases), transient and sustained activation states have been shown to result in different cellular responses [49]. It is well established that GEFs and GAPs function as positive and negative regulators of Rho GTPase cycles, respectively. We have shown that the functions of GEFs and GAPs are modulated by their interactions with GDIs, and that the interconversion between transient and sustained Rho activation occurs mainly through changes in the affinities of GDIs to GAPs and the concentrations of GAPs. The properties of GDIs and GAPs are regulated by posttranscriptional modifications [29-33,36-38] and the affinity between GDIs and GAPs may be altered by such modifications. Therefore, RhoGDIs and GAPs might participate in the switching between transient and sustained signals of the Rho GTPases. Although this mode seems not to be common in the regulation of Rho GTPases, certain sets of GTPases, GEFs, and GAPs may use this mode of regulation.
In the present study, we proposed a simplified model for positive regulation of Rho GTPases by GDIs. However, the model does not take into account GTPase cycling between membrane and cytosol. Cells contain membranous and cytoplasmic compartments, and typically, Rho GTPases function within the membranous compartments. It has been shown that the efficient cycling between inactive and active states of GTPases can occur entirely within protein complexes assembled on membrane surfaces [22]. RhoGDIs mediate the membrane-cytoplasmic shuttling of GTPases, and likely can alter the concentrations of GTPases and their RhoGDI-associated regulators at target sites in cells. Therefore, it is necessary to take into account the shuttling processes in developing a truly comprehensive model. Membrane-cytoplasmic shuttling has been considered in a simulation of the distribution of activated Cdc42 during the early phase of yeast bud formation [23]. Additionally, a modeling framework describing Rac cycling between membrane and cytosol has been reported [21]. Because our model for the Rho GTPase switch can be regarded as a basal signaling module, these studies that have taken into account the Rho GTPase shuttling processes should be incorporated into our model of the Rho GTPase switch for a more detailed and biologically-relevant model.