Biochemical Society Transactions

Rab GTPases and Their Interacting Proteins in Health and Disease

Rho family GTPases

Alan Hall

Abstract

Rho GTPases comprise a family of molecular switches that control signal transduction pathways in eukaryotic cells. A conformational change induced upon binding GTP promotes an interaction with target (effector) proteins to generate a cellular response. A highly conserved function of Rho GTPases from yeast to humans is to control the actin cytoskeleton, although, in addition, they promote a wide range of other cellular activities. Changes in the actin cytoskeleton drive many dynamic aspects of cell behaviour, including morphogenesis, migration, phagocytosis and cytokinesis, and the dysregulation of Rho GTPases is associated with numerous human diseases and disorders.

  • actin cytoskeleton
  • Cdc42
  • GTPase
  • Rac
  • Rho

Introduction

Rho GTPases constitute one of the five distinct families of the Ras superfamily [1]. Rho itself was discovered serendipitously in 1985 as a Ras-related protein (Ras homology) in Aplysia, and subsequently 18 genes encoding Rho-like proteins were identified in the human genome (Table 1), with Rho, Rac and Cdc42 being the far best characterized [2,3]. Rho-like proteins are found in all eukaryotic species so far examined and are highly conserved, even to the extent that the mammalian Cdc42 gene is capable of at least partially rescuing deletions in Saccharomyces cerevisiae Cdc42 [4]. Like Ras, the majority (although not all, see Table 1) of Rho family members act as molecular switches to regulate signal transduction pathways, by interconverting between inactive GDP-bound and active GTP-bound conformational states. Two breakthroughs in understanding the regulation of the switch came in 1991. First, a biochemical purification approach showed that the product of the bcr (breakpoint cluster region) gene acted as a GAP (GTPase-activating protein) to stimulate the intrinsic GTP hydrolysis rate of Rac [5]. Secondly, the oncoprotein Dbl was found to act as a GEF (guanine-nucleotide-exchange factor) and catalyse the exchange of GTP for GDP on Cdc42 [6]. Subsequent analysis of the human genome has revealed 67 Rho family GAPs harbouring a BH (bcr homology) domain and 71 Rho family GEFs with a DH (Dbl homology) domain (Figure 1). Later, the characterization of a protein DOCK (downstream of Crk-180 homologue) and its Drosophila and Caenorhabditis elegans orthologues led to the identification of a second sequence-unrelated family of GEFs, with 11 members in the human genome [7]. The substrates of DOCK family proteins appear to be Rac and Cdc42 only. Finally, a protein capable of inhibiting the dissociation of GDP from Rho [GDI (guanine-nucleotide-dissociation inhibitor)] was purified and three Rho family GDIs are encoded in the human genome [8]. The sizes of the GEF and GAP families point to highly complex pathways regulating Rho family GTPases, but the true significance of this is still unclear.

View this table:
Table 1 Rho GTPase family members

Eighteen Rho family members are found in mammals, although four (Rnd1, Rnd2, Rnd3/RhoE and RhoH/TTF, indicated by an asterisk) have no detectable intrinsic GTPase activity and are therefore unlikely to act as molecular switches. Drosophila and S. cerevisiae each have six Rho GTPases. Typically, Rho GTPases are approximately 190 amino acids in length and have a C-terminal CAAX box that is post-translationally prenylated. Four additional proteins are sometimes included in the mammalian Rho family (not shown): RhoBTB1 and RhoBTB2 have a stretch of relatively weak homology with Rho embedded within a larger protein (approximately 700 amino acids), and have no CAAX box. Two mitochondria-localized proteins, Miro1 and Miro2, have even less homology with Rho and are no longer considered part of the Rho family. Arabidopsis has 11 Rho family members, Rop 1–11 (not shown) [1].

Figure 1 Regulation of Rho family GTPases

Some 82 mammalian GEFs have been described: 71 in the Dbl (DH/PH domain) family and 11 in the DOCK (CZH1/CZH2) family. The DH and CZH2 domains encode catalytic activity respectively. The significance of having two distinct families is unknown. S. cerevisiae contains six DH/PH GEFs and one CZH2 domain-containing protein. Interestingly, Arabidopsis has no DH/PH GEFs and only one DOCK protein (SPIKE1). Instead, 14 Rop (Rho in plant) GEFs harbour a PRONE domain with catalytic GEF activity. The BH, Rho GAP domain is conserved in all eukaryotes. Some 67 GAPs are found in mammals, whereas Drosophila, S. cerevisiae and Arabidopsis have 20, five and eight respectively. Rho GDIs (~300 amino acids) inhibit spontaneous GDP dissociation and can solubilize GDP-bound Rho GTPases from membrane compartments. Mammals have three Rho GDIs, whereas Drosophila, S. cerevisiae and Arabidopsis each have one. A number of other mechanisms of regulating Rho GTPases have been described, although the generality and significance of these alternative pathways are not well understood and are likely to be cell-type- and/or context-dependent.

Downstream targets

The identification of effectors or targets of Rho family GTPases was not so straightforward, since the majority do not contain a recognizable conserved domain. However, through a combination of affinity chromatography, protein purification and yeast two-hybrid screening, over 100 targets have been reported for members of the Rho family [9]. They include approximately 30 kinases and a large number of scaffold/adaptor-like proteins. These data, which started to appear in the mid-1990s, revealed that Rho, Rac and Cdc42 are capable of interacting with perhaps 20 or 30 different proteins in a GTP-dependent manner, raising the likely prospect that they each regulate numerous distinct signal transduction pathways, not just one. The huge diversity of upstream regulators and downstream signals has made the analysis of Rho GTPase signalling pathways challenging.

Rho GTPases and the actin cytoskeleton

Two papers published in 1992 provided the first direct evidence that Rho and Rac regulate the assembly and organization of F-actin (filamentous actin) in response to extracellular cues [10,11]. In a fibroblast cell line, Rho was shown to promote the assembly of contractile actomyosin filaments in response to LPA (lysophosphatidic acid) addition, whereas Rac promoted the assembly of a peripheral actin meshwork, leading to membrane protrusions (lamellipodia/membrane ruffles) in response to PDGF (platelet-derived growth factor) or insulin. Furthermore, activated Ras strongly activated Rac, demonstrating cross-talk between the Ras and Rho families of small GTPases [11]. Later work revealed that, upon activation, Cdc42 induced peripheral actin-rich microspikes (filopodia) and also activated Rac (producing lamellipodia), demonstrating cross-talk within the Rho family [12,13]. Rho GTPases have since been shown to control the organized assembly of F-actin in all eukaryotic species examined so far, whereas the distinct effects of Rho, Rac and Cdc42 on actin are conserved in most, if not all, metazoan cell types [3].

The signal transduction pathways through which Rho GTPases mediate their effects on the actin cytoskeleton are now known in some detail. Rho promotes the polymerization of actin into linear filaments through a direct interaction with its target mDia [an FH (formin homology) domain-containing protein] [14]. In addition, activation of ROCK (Rho-associated kinase) leads to the phosphorylation and inactivation of myosin light chain phosphatase, thereby activating myosin II [15,16]. Although there are undoubtedly still many aspects of regulation to be uncovered, the activation of the two targets, mDia and ROCK, by Rho are thought to provide the trigger for the assembly of contractile actomyosin filaments. Activation of the heptameric actin-polymerizing complex Arp2/3 (actin-related protein 2/3), by the protein WAVE [WASP (Wiskott–Aldrich syndrome protein) verprolin homologous], leads to the meshwork of peripheral F-actin induced by Rac [17]. Rac is thought to release WAVE from an inhibited complex, and Arp2/3 elongates actin filaments from the sides of pre-existing filaments generating a meshwork [18]. Finally, Cdc42 also activates Arp2/3, in this case through a direct interaction with another Arp2/3 activator, N-WASP (neuronal WASP) [19]. The formation of filopodia is believed to require an Arp2/3-mediated initiation event followed by an mDia filament elongation event [20].

Other activities of Rho GTPases

The first reported activity of a Rho family protein, in 1991, was for Rac acting as an allosteric regulator of the phagocytic NOX (NADPH oxidase) enzyme complex [21]. This enzyme is a major component of the killing machinery for micro-organisms phagocytosed by neutrophils and macrophages and mutations in any of its four subunits underlie CGD (chronic granulomatous disease), an immunodeficiency. Rac interacts with p67phox, a cytosolic component that is recruited to the membrane during NOX activation [22]. Since Rac is also required for the actin polymerization driving phagocytosis, it is tempting to speculate that this dual role of Rac (actin and NOX) facilitates co-ordination of these activities in space and time. Interestingly, although p67phox is not conserved in plants, the Rho family protein Rop2 regulates NOX activity in Arabidopsis [23]

The large number of target proteins points to additional activities of Rho family members and many have been reported over the years. Cdc42 is a key regulator of cell polarity in budding yeast, being required for polarized bud site establishment during cell division and for positioning the membrane extension formed during mating [24]. One of Cdc42′s targets in metazoans is the adaptor protein Par6, which is required for establishing anterior–posterior polarity in the C. elegans embryo, as well as apical–basal polarity in epithelial cells and axon–dendrite polarity in neurons [25,26]. A number of signal transduction pathways associated with the regulation of gene transcription are controlled by Rho GTPases. Rho, for example, activates the transcription factor SRF (serum-response factor), albeit indirectly through its effects on actin, whereas Rac and Cdc42 activate the JNK (c-Jun N-terminal kinase) and p38 MAPK (mitogen-activated protein kinase) pathways through targets such as the mixed lineage kinases [27,28]. These effects are likely to be cell-type- and context-dependent, since there are examples of JNK/p38 activation that do not seem to involve a Rho family protein.

Rho GTPases and cell biology

It has been known for some 40 years that the actin cytoskeleton drives many dynamic processes in cells, including migration, phagocytosis, endocytosis, morphogenesis and cytokinesis (Figure 2). Work over the last 15 years, using (i) dominant-negative versions of Rho, Rac and Cdc42, (ii) bacterial toxins such as C3 (Rho inhibitor) or toxin B (Rho, Rac and Cdc42 inhibitor), (iii) genetic analysis in model organisms, and (iv) RNAi (RNA interference), has revealed a central role for Rho GTPases in regulating each of these processes. However, in only a very few cases has the specific signal transduction pathway involved (e.g. GEF, GAP and target) been fully characterized.

Figure 2 Cell biological roles of the Rho family

Rho GTPases control the assembly and organization of the actin cytoskeleton in most, if not all, eukaryotic cells. It has been known for at least 40 years, through the use of actin polymerization poisons, notably cytochalasin, that actin drives many of the dynamic properties of cells, including the ones shown schematically. Many other cell biological processes involving Rho GTPases are not shown.

Cell migration requires the activities of multiple Rho family members. Classic mesenchymal cell migration in two-dimensional cultures involves actin-driven protrusions at the front co-ordinated with actomyosin contractility at the cell body and rear. This has led to the hypothesis that, together, active Rac at the front and active Rho at the back promote migration [29]. In addition, single-cell chemotaxis and scratch-induced migration in cell monolayer assays have indicated a role for Cdc42 in interpreting extracellular, directional cues, an activity that is in keeping with its general role as a regulator of cell polarity [30,31]. However, as work on cell migration has progressed and become more sophisticated over recent years, it is now clear that these simple ideas require significant modification. Thus cells can move in different ways depending on the environmental context. For example, in a three-dimensional matrix, some tumour cells move using a blebbing-type motion, which is driven by Rho and actomyosin, but is independent of Rac [32].

One take-home message from the work on cell migration is that the signal transduction pathways activated by Rho GTPases must be spatially restricted. This has led several groups to develop novel biosensors capable of visualizing the activation of Rho family members in time and space [33,34]. These reagents have led to new insights into the roles of and interrelationships between Rho, Rac and Cdc42. Probably the most dramatic observation in the migration field from the use of biosensors has been that Rho and Rac are both active at leading edge protrusions [35]. This contrasts starkly with the early hypothesis of active Rac at the front and active Rho at the back. The biochemical implications of the result are still unclear: one possibility is that Rho at the front is activating a different signalling pathway from Rho at the back. The spatiotemporal control of Rho GTPase activities is clearly a distinguishing feature of all the cell biological processes shown in Figure 2. Thus spatially restricted actomyosin contractility (Rho) can promote the specific changes in cell shape seen during morphogenesis or cytokinesis, whereas highly localized protrusive activity (Rac/Cdc42) is seen not only during migration, but also around particles as they are phagocytosed.

Rho GTPases and disease

The founder member of the Ras superfamily was intensely studied in the 1980s, since it was found to be an oncogene in some 30% of all human cancers. Further analysis revealed a single amino acid substitution resulting in a protein that was insensitive to GAPs and therefore constitutively active. The enormous amount of sequencing information derived from human cancers has uncovered no similar mutations in Rho GTPases. Many of the Rho GEFs, including Dbl, the founder member, were originally identified as oncogenes in experimental tissue culture transformation assays. Despite this, there are just a few sporadic examples of genetic alterations in Rho GEFs in human cancer. The possibility that Rho family GAPs might act as tumour suppressors has also been explored. Indeed, there is strong evidence supporting such a role for DLC1 (deleted in liver cancer 1), a GAP active on Rho [36]. Heterozygous deletions are found in approximately 50% of human cancers, and loss of DLC1 can promote tumorigenesis in mouse cancer models. However, homozygous deletions of DLC1 are not found in human cancer, and, whereas expression from the remaining allele appears to be attenuated in some cancers, definitive proof for a tumour-suppressor role has yet to be obtained [37].

Genetic alterations in Rho family signalling pathways have, on the other hand, been reported in other diseases and disorders. Mutations in GAPs, GEFs and targets are found in a variety of mental retardation syndromes and it has been proposed that these lead to defects in spine morphogenesis, an alteration in the size and shape of dendritic spines associated with memory and learning [38]. Genetic alterations are also associated with immunological disorders [Rac2 in LAD (leucocyte adhesion deficiency) and the Cdc42 target WASP in Wiskott–Aldrich syndrome], and developmental disorders {the Cdc42 GEF FGD1 [FYVE, RhoGEF and PH (pleckstrin homology) domain-containing 1] in faciogenital dysplasia syndrome} [3941]. Finally, there is a great deal of interest in ROCK as a pharmaceutical target. Aberrant actomyosin contraction has been implicated in a variety of human diseases, including hypertension, atherosclerosis and ischaemia. A ROCK inhibitor is already in clinical use in Japan for cerebral ischaemia and Phase II trials are underway in the US for treatment of glaucoma [42].

Conclusions

The literature on Rho GTPases is now huge and it has only been possible to give the briefest of overviews in the present paper. Outside of specific details, there are, however, a number of general issues related to Rho GTPases that are still not understood. It is not clear, for example, how target specificity is achieved. One solution to this, for which there is evidence in yeast, is that the GEF determines the signal output of the GTPase. This can be achieved if the GEF is physically associated with one or more specific targets and it could partly explain why there is such a diversity of GEFs. A second question relates to spatial organization of signalling. This is probably best understood in S. cerevisiae, where localized accumulation of active Cdc42 at the membrane during bud formation involves: (i) a positive-feedback loop due to the recruitment of a target protein that is already associated with a GEF, and (ii) Cdc42 inactivation and dissociation from the membrane, through a GAP, a GDI and endocytosis. Mathematical modelling suggests that this cycle of activation and inactivation promotes the highly localized accumulation of an active patch of Cdc42 at the membrane. Whether similar mechanisms regulate Rho GTPases in mammalian cells remains to be seen.

Footnotes

  • Rab GTPases and Their Interacting Proteins in Health and Disease: A Biochemical Society Focused Meeting held at University College Cork, Cork, Ireland, 11–13 June 2012. Organized and Edited by Mary McCaffrey (University College Cork, Ireland).

Abbreviations: Arp2/3, actin-related protein 2/3; bcr, breakpoint cluster region; BH, bcr homology; DH, Dbl homology; DLC1, deleted in liver cancer 1; DOCK, downstream of Crk-180 homologue; F-actin, filamentous actin; GAP, GTPase-activating protein; GDI, guanine-nucleotide-dissociation inhibitor; GEF, guanine-nucleotide-exchange factor; JNK, c-Jun N-terminal kinase; NOX, NADPH oxidase; PH, pleckstrin homology; ROCK, Rho-associated kinase; WASP, Wiskott–Aldrich syndrome protein; WAVE, WASP verprolin homologous

References

View Abstract