Gene name - Rac1
Cytological map position - 61F5
Function - GTPase - signaling protein
Keyword(s) - cytoskeleton
Symbol - Rac1
FlyBase ID: FBgn0010333
Genetic map position - 3-
Classification - GTP-binding protein (rho-subfamily).
Cellular location - cytoplasmic
Rac, Rho (see Drosophila Rho1) and Cdc42 are members of the Rho family of small guanosine triphosphatases (GTPases) are all implicated in the temporal and spatial control of actin filament organization in the cytoskeleton. In mammalian fibroblasts actin filaments exist principally in three types of structures: the cortical actin network, actin stress fibers, and cell surface protursions including membrane ruffles and microspikes. Stress fibers emanate from distinct areas of the plasma membrane known as focal adhesions, where clusters of integrin receptors bind to extracellular matrix proteins such as fibronectin and collagen. A number of proteins are found associated with focal adhesions at the intracellular face of the plasma membrane, including vinculin, talin, tensin and alpha-actinin (Ridley, 1992a and references). Ruffling, in contrast to stress fibers which emanate throughout the cell, is a phenomenon of the actin reorganization at the membrane. It is thought that ruffling is a prelude to the accumulation of multiple large intracellular vesicles by pinocytosis. It is likely that these two processes are linked, as there is a close association between ruffling and pinocytosis (Ridley 1992b).
Although structurally related, the three proteins serve different functions, to judge by their diverse roles in cultured cells. Rho controls the assembly of actin stress fibers and focal adhesion complexes (Ridley, 1992a). Rac proteins, approximately 60% identical to Rho, regulate actin filament accumulation at the plasma membrane to produce lamellipodia (cellular protrusions involved in cell motility) and membrane ruffles. Rac also induces stress fiber formation but this function is dependent on the presence of Rho (Ridley, 1992b). Cdc42 stimulates the formation of filopodia (finger like projections that form around actin bundles whose barbed ends are oriented in the direction of outgrowth, involved in motility and axon growth cone spreading) (Nobes, 1995a).
Rac stimulates the assembly of multimolecular focal complexes at the plasma membrane. These complexes, which are associated with lamellipodia and filopodia, contain vinculin, paxillin, and focal adhesion kinase, but are formed independently of rho-induced focal adhesions. These Rac induced focal complexes at the leading edge of motile cells are distinct from focal adhesions in three ways: (1) they are much smaller, (2) they do not have the characteristic elongated, arrowhead shape of Rho-regulated focal adhesions and (3) they are not arranged within the cell in the same fashion as focal adhesions. Activation of Cdc42 in cultured cells leads to the sequential activation of Rac and then Rho, suggesting a molecular model for the coordinated control of cell motility by members of the Rho family of GTPases (Nobes, 1995a).
In Drosophila, Rac and Rho are both involved in dorsal closure. This involvement requires the cytoskeleton and downstream gene function, in particular the newly discovered serine/threonine kinase PAK, that binds and is activated by Rac and CDC42 (Harden, 1996). Hemipterous and Jun N-terminal kinase are downstream targets of Rac and CDC42 in the PAK initiated phosphorylation cascade. The dorsal closure phenotype is discussed more fully in the Hemipterous and JNK sites.
Rac and Cdc42 are GTP binding proteins that serve a function similar to that of Ras; all three are molecular switches that can activate protein kinases. A kinase cascade has been characterized in vertebrates that is homologous to that involving Hemipterous in Drosophila. Another direct target of Rac characterized in vertebrates is phosphatidylinositol (PI) 3-kinase (see Drosophila Pi3K92E). This interaction depends on Rac being in a GTP-bound state and requires an intact Rac effector domain (Bokoch, 1996). The function of PI3K is required for the generation of phospholipids, which leads to formation of diacylglycerol and the activation of protein kinase C.
Rac1 and CDC42 control actin-dependent processes in the fly's wing imaginal disc epithelium, and plays a role in the formation of the polarized outgrowth of wing hairs. At approximately 35 hours after puparium formation each wing epithelial cell forms a hair by extending a single process from its apical membrane. The emergence of a wing hair is presaged by the accumulation of actin on the distal side of the cell. Outgrowth initiates from this site and is oriented distally. At this stage, the outgrowth is termed a prehair. Subsequently, the prehair elongates and tilts up out of the plain of the epithelium so that its base comes to lie in the center of the apical membrane. Tissue polarity mutants such as frizzled, dishevelled, and prickle interfere with the choice of the site at which actin begins to accumulate before wing hair formation. Because these genes act early in the polarization pathway, they are likely candidates to act as molecules that transmit the polarization signal. Frizzled is unlikely to be involved in structural aspects of hair formation; rather it must polarize some feature of the cell that can later be used to determine either the site of hair outgrowth or cellular orientation. The tissue polarity genes inturned, fuzzy, and multiple wing hair act downstream of frizzled and are required only for hair polarity. These genes may help translate the cellular polarity generated by the Frizzled pathway into polarized hair outgrowth (Wong, 1993 and Eaton, 1996).
Rac and the other Ras family GTPases can be mutated to dominant negative forms that interfere with the functioning of wild type GTPases. When a dominant negative Rac1 is expressed in the wing, a polarity phenotype similar to that of inturned and fuzzy is produced. Cells that express a dominant negative Cdc42 often make no hair at all. Expression of dominant negative Cdc42 interferes with actin polymerization in wing hairs. In wild-type cells prior to hair formation, actin is distributed around the periphery in the apical junction region. Actin fibers are observed running across the cells in many directions, often with a radial arrangement. The first sign of hair formation is the accumulatin of actin on the distal side of the cell. At the level of the intercellular junction, actin filaments are often observed extending from the distal vertex into the center of the cell. As the hairs extend, actin disappears from the junctional region. Dominant negative Cdc42 causes dramatic defects in the actin organization that correlate with prehair outgrowth. By the time their wild-type neighbors have extended actin-filled prehairs, cells expressing dominant negative Cdc42 have not polarized the distribution of actin filaments distally, and no sign of outgrowth is evident. These data suggest that Cdc42 is specifically required for actin polymerization in developing wing hairs and that actin polymerization is required for outgrowth (Eaton, 1996).
While Cdc42 is critical for the outgrowth but not the placement of wing hairs, Rac1 is involved in the placement but not outgrowth. Dominant negative Rac1 causes duplication or triplication of wing hairs, but no defects in the appearance of the hairs themselves. Normally there is a continuous band of junctional actin around the perimeter of the apical aspect of wing epithelia. Upon expression of dominant negative Rac1 the normal continuous band of junctional actin is reduced in amount and fraught with gaps. The process by which duplicate hairs are filled with normal amounts of actin is delayed. During formation of hairs the majority of microtubules (see ß1 tubulin for related information) span the cell at the level of apical junctions, while 30 hours later, most microtubules comprise a longitudinal array that runs from the apex to the base of each cell. This suggests that microtubule organization is developmentally regulated during wing formation. Dominant negative Rac1 expression disorganizes the apical microtubule web. It is concluded that Rac1 is needed to maintain the proper organization of the apical microtubule web before prehair formation, at the time planar polarization is thought to occur. It is likely that Rac1 acts at the sites of cell-cell contact where junctions are located, and that dominant negative Rac1 disrupts planar polarization because of the disorganization of junctional actin and microtubules (Eaton, 1996).
Besides being involved in dorsal closure and hair polarity, the Rho family of GTPases are also involved in oogenesis. Rac is involved specifically in border cell migration. Analysis of heat shock regulated Rac1 indicates a requirement both for the initiation and continuation of migration. For more information about border cell migration see Slow border cells (Slbo). Expression of constitutively active Rac is unable to rescue the border cell migration defect in mutant slbo, suggesting that the two function in different pathways. Rac, Rho and Cdc42 are also required in the germ line for proper nurse cell cytoplamic transport. At stage 11 of oogenesis, a network of actin filaments polymerizes to form a cage around nurse cell nuclei, apparently preventing their movement during the transfer of nurse cell cytoplasm to the oocyte. This transfer process, known as "dumping," appears to result from the myosin-based contraction of subcortical actin. Limiting the amounts of each of the Rho family members results in the absence of the actin cage that normally surrounds stage 11 nurse cell nuclei. This results in the obstruction of ring canals and the failure of complete transfer, results similar to those occurring in other mutants that affect the cytoskeleton, such as chickadee, singed and quail (Murphy, 1996 and references).
The exact protein interactions required for Rac1 involvement in hair cell placement, follicle cell migration, and formation of the nurse cell actin cage are not yet known. Does Rac signal through Hemipterous in these instances as it does in dorsal closure, or are the effects of Rac confined to cytoskeletal organization independent of nuclear signaling? The involvement of Rac in two signaling cascades, one involving the serine/threonine kinase PAK and another potentially involving protein kinase C, leaves no shortage of pathways to be examined.
Rac GTPases regulate the actin cytoskeleton to control changes in cell shape. To date, the analysis of Rac function during development has relied heavily on the use of dominant mutant isoforms. Here, loss-of-function mutations have been used to show that the three Drosophila Rac genes, Rac1, Rac2 and Mtl, have overlapping functions in the control of epithelial morphogenesis, myoblast fusion, and axon growth and guidance. They are not required for the establishment of planar cell polarity, as had been suggested on the basis of studies using dominant mutant isoforms. The guanine nucleotide exchange factor, Trio, is essential for Rac function in axon growth and guidance, but not for epithelial morphogenesis or myoblast fusion. Different Rac activators thus act in different developmental processes. The specific cellular response to Rac activation may be determined more by the upstream activator than the specific Rac protein involved (Hakeda-Suzuki, 2002).
In Drosophila, studies using constitutively active and dominant negative mutants have implicated Rac1 in closure of the dorsal epidermis, myoblast fusion, the establishment of planar cell polarity, and the control of axon growth and guidance. Each of these processes requires dynamic remodelling of the actin cytoskeleton, although the extracellular signals and the cellular responses involved seem to be different in each case. Given the ability of dominant mutant Rac proteins to interfere with cytoskeletal dynamics, it is not surprising to find that they perturb each of these processes. But are endogenous Rac proteins actually required for these processes, and if so, which proteins are involved, and how are they regulated? These questions cannot be answered using dominant mutant proteins. They require the phenotypic analysis of loss-of-function mutations in each of the endogenous Rac genes (Hakeda-Suzuki, 2002).
The Drosophila genome contains two highly similar Rac genes, Rac1 and Rac2. A third gene, Mtl, encodes a closely related GTPase that is structurally similar to both Rac and Cdc42 GTPases, but functionally behaves like Rac1 and Rac2. Rac1, Rac2 and Mtl are therefore referred to collectively as the Drosophila Rac genes. All three genes are ubiquitously expressed during development. A loss-of-function mutation in the Mtl gene was generated by imprecise excision of a P-element inserted in the first non-coding exon. A 2,068-base pair (bp) deletion was recovered that removes the entire Mtl open reading frame, but no part of any other predicted gene. Animals homozygous for this deletion, MtlDelta, as well as both Rac1 and Rac2 single mutants, are viable and fertile. The Rac2 Mtl double mutant is also viable and fertile. All other combinations are homozygous lethal. These loss-of-function mutations have been used to assess the contribution of each Rac protein to a set of distinct cell-shape changes that occur during Drosophila development. Embryos were examined lacking both the maternal and zygotic contributions of one or more Rac gene, and also pupae and adults that were homozygous mutant either entirely or in large clones of cells. For pupae and adults, both the strong hypomorph Rac1J10 and the null allele Rac1J11 were used, together with the null deletion alleles for Rac2 and Mtl (Rac2Delta and MtlDelta). Analyses in the embryo were restricted to the use of the Rac1J10 allele, since triple mutant embryos could not be recovered using the null allele Rac1J11. Evidently, Rac proteins also have important but still uncharacterized functions during oogenesis and early embryogenesis (Hakeda-Suzuki, 2002).
During Drosophila embryogenesis, opposing lateral epidermal sheets move towards one another, meeting and fusing seamlessly at the dorsal midline. This process of dorsal closure resembles ventral enclosure in Caenorhabditis elegans; and wound healing in vertebrates. It is believed to be driven, at least in part, by an actomyosin contractile ring that assembles at the leading edge, with lamellipodial and filopodial protrusions facilitating adhesion and alignment as the epidermis is sealed. Expression of dominant negative Rac1 in epidermal cells prevents formation of the acto-myosin cable and completion of dorsal closure, suggesting that at least one endogenous Rac protein might be involved. This study has determined that all three Rac proteins contribute to dorsal closure. Triple mutant Rac embryos fail to complete dorsal closure. Little or no actin accumulation is seen at the leading epidermal edge, and both lamellipodia and filopodia are lacking. The underlying amnioserosa cells appear normal (Hakeda-Suzuki, 2002).
Weaker and less frequent defects are also seen in Rac1;Rac2 and Rac1;Mtl double mutant embryos. All remaining single and double mutant embryos successfully complete dorsal closure. Dorsal closure thus relies more on Rac1 than either Rac2 or Mtl, although any one of the three is largely sufficient. Quite different cell-shape changes occur during cell fusion, a striking example of which is the fusion of myoblasts to form multinucleate muscle fibers. The role of the actin cytoskeleton in myoblast fusion remains unclear. Most likely, it is involved in the formation of a vesicular prefusion complex that assembles at the apposed plasma membranes. Expression of either dominant negative or dominant active Rac1 in Drosophila myoblasts blocks their fusion, but here too the precise roles and contributions of individual Rac genes are unknown. Little or no myoblast fusion occurs in either Rac1 Rac2 double mutant or Rac1;Rac2;Mtl triple mutant embryos. In contrast, myoblast fusion appears to be complete in Rac1 and Mtl single and double mutants, whereas only a few isolated myoblasts fail to fuse in Rac2 single mutants and Rac2;Mtl double mutants. Myoblast fusion thus requires either Rac1 or Rac2, but not Mtl (Hakeda-Suzuki, 2002).
Actin rearrangements also underlie the establishment of planar cell polarity (PCP) within an epithelium. In Drosophila, PCP has been studied most extenstively in the context of eye and wing development. Photoreceptors in the eye are arranged in a trapezoidal fashion, giving each ommatidium a specific chirality and orientation. In the wing, each epithelial cell forms a single distally oriented hair, preceded by the assembly of an actin-based 'pre-hair' at the distal vertex of the cell. The involvement of Rac proteins in PCP has been suggested by the finding that both dominant negative and constitutively active forms of Rac1 disrupt ommatidial orientation in the eye, and that dominant negative Rac1 also induces the formation of multiple hairs per cell in the wing. To test more critically a requirement for Rac proteins in PCP, clones of cells in the eye and the wing were examined that were triply mutant for null alleles of Rac1, Rac2 and Mtl. No PCP defects could be detected within these clones in either tissue. Presumably, the PCP defects previously reported are due to cross-inhibition or cross-activation of other pathways by the dominant mutant Rac proteins used (Hakeda-Suzuki, 2002).
The most complex changes in cell shape that occur during development take place in the nervous system, as differentiating neurons extend axons and dendrites towards their specific target cells. Dominant mutant Rac proteins are crude tools with which to address the complexities and subtleties of neuronal differentiation, and it is perhaps not surprising that their use has sometimes led to conflicting results. To begin to tease apart the diverse functions of Rac GTPases in nervous system development, the embryonic central nervous system (CNS) and peripheral nervous system (PNS), and the adult visual system, were examined in single, double, and triple mutants for Rac1, Rac2 and Mtl. Embryonic CNS axon pathways were examined using anti-Fasciclin II (FasII) monoclonal antibody 1D4. FasII labels axons in three longitudinal fascicles on each side of the midline, and is thus a sensitive marker to detect any misrouting of longitudinal axons across the midline. Such midline guidance errors occur in 33% of segments in Mtl mutant embryos, but only at very low frequency (less than 2%) in embryos lacking one or both of Rac1 and Rac2. Mutations in Rac1, and to a lesser extent Rac2, enhance the frequency of the midline guidance errors in Mtl mutant embryos (to 75% and 42% of segments, respectively. Axon guidance defects also occur in the visual system of whole-eye Rac mosaics generated using eyFLP. In wild-type adults and control mosaics, photoreceptor axons establish precise topographic connections in the lamina and medulla of the optic lobe. These projection patterns are largely normal in each of the single mutants, and only mild defects occur in Rac1 Rac2 and Rac2 Mtl double mutants. Projection defects are more pronounced in Rac1 Mtl double mutants, and are severe in the triple mutant. These defects include local disruptions in topographic mapping, and a frequent misrouting of photoreceptor axons around and beyond the medulla. Specification of photoreceptor cell fate appears to be normal, even in triple mutant clones. The projection defects observed in the triple mutant can be rescued by reintroducing either Rac1 or Mtl specifically in the eye using a GMR transgene (Hakeda-Suzuki, 2002).
Together, these data establish a critical role for endogenous Rac proteins in axon guidance. The three Rac proteins have overlapping functions in axon guidance in both the CNS and visual system, just as they do in the mushroom bodies. Nevertheless, some degree of specialization can be discerned. For example, axon guidance at the CNS midline depends more on Mtl than Rac1, whereas Rac1 is more important than Mtl in the mushroom bodies (Hakeda-Suzuki, 2002).
Whereas single and double mutant embryos reveal a role for Rac proteins in axon guidance, triple mutant embryos demonstrate the essential function of Rac proteins in axon growth. Severe growth defects occur in Rac1;Rac2;Mtl homozygous mutant embryos. In the CNS, FasII-positive axons rarely extend from one segment into the next, and very few sensory axons from the PNS reach the CNS. Specification of neuronal and glial cell fate appears relatively normal, as does dendritic growth and morphology. These axon growth defects were quantified by determining the frequency with which axons from the dorsal cluster of sensory neurons reach the lateral cluster on their path towards the CNS. This analysis of the PNS confirmed the general impression gained from the CNS: axon stalling is severe in Rac1;Rac2;Mtl triple mutants, occurs occasionally in Rac1;Mtl double mutants, and is rare in all other combinations (Hakeda-Suzuki, 2002).
These data demonstrate that axon growth also requires Rac activity, but only at a very low level. This activity can be provided by any one of the three Rac proteins alone, although Rac2 is less effective than either Rac1 or Mtl. Axon growth is thus maintained at levels of Rac activity that are insufficient for accurate guidance, consistent with the idea that a low level of Rac activity is essential to drive the growth cone forward, while spatially restricted bursts of high activity may be needed to turn it (Hakeda-Suzuki, 2002).
Endogenous Rac GTPases thus function in morphogenesis of the epidermis, mesoderm, and nervous system. Are they regulated by the same or different upstream activators in each of these tissues? The guanine nucleotide exchange factor Trio activates Rac1, Rac2 and Mtl in vitro, and loss of trio function in the visual system results in projection errors of photoreceptor axons similar to those observed in Rac triple mutants. Axon guidance errors and occasional stalling defects also occur in embryos lacking zygotic trio function. Axon stalling becomes severe in both the CNS and PNS if the maternal trio function is also eliminated. As with the Rac proteins, low levels of Trio activity are sufficient but essential for axon growth. This critical requirement for Trio in axon growth is particularly striking, given that the Drosophila genome encodes at least 22 other Rho family GTPase exchange factors, several of which are also expressed in the developing nervous system (Hakeda-Suzuki, 2002).
In the embryonic nervous system and adult visual system, loss of trio function thus results in defects remarkably similar to those observed upon loss of Rac function, consistent with the idea that Trio and Rac proteins act in a common pathway in vivo. An epistasis experiment was performed to test this. Overexpression of the Trio GEF1 domain using the eye-specific GMR promoter results in a severely disrupted eye morphology and highly aberrant photoreceptor axon projections. If Trio signals through Rac proteins in vivo, then these defects should be dependent on Rac function. This is indeed the case. Both the eye morphology and axon projection defects are almost completely suppressed in animals homozygous for loss-of-function mutations in either Rac1 or Rac2. Mtl alone does not suppress this trio gain-of-function phenotype. The Rac1;Rac2;Mtl triple mutant phenotype is completely epistatic to the trio gain-of-function phenotype. These data demonstrate that Trio GEF1 does indeed act through Rac proteins in vivo, and further suggest that Rac1 and Rac2 are its preferred substrates. The trio loss-of-function phenotype is however much more severe than the Rac1;Rac2 double mutant phenotype, suggesting that endogenous Trio may also activate Mtl, at least when Rac1 and Rac2 are lacking (Hakeda-Suzuki, 2002).
Having identified Trio as the primary activator of Rac proteins during axon growth, whether Trio is required for any of the other Rac functions was investigated. Dorsal closure occurs normally in embryos lacking both maternal and zygotic trio function. Myoblast fusion also appears complete in these embryos, but myotubes often fail to attach themselves correctly to the epidermis. Thus, although it is expressed in both the epidermis and mesoderm, Trio is not required for either dorsal closure or myoblast fusion (Hakeda-Suzuki, 2002).
Thus endogenous Rac proteins control cell-sheet spreading, cell fusion, and axon growth and guidance, and they also regulate axon branching. Each of these processes involves its own characteristic restructuring of the cytoskeleton, and hence is likely to be mediated by a different set of Rac effectors. What determines which of these effector pathways will be stimulated when Rac proteins are activated? One possibility would be that distinct Rac proteins have distinct effectors. This may well be the case for myoblast fusion, which can be mediated by Rac1 or Rac2, but not Mtl. However, in most cases Rac1, Rac2 and Mtl have largely overlapping functions, indicating that they also share a common set of effectors. A similar pattern of overlapping functions in diverse processes has also recently been reported for the three C. elegans Rac genes. In general, the cellular response is therefore unlikely to be dictated by the specific Rac protein involved. These results suggest an alternative possibility. Trio, despite its widespread expression, is required for only a limited subset of Rac functions. This suggests that the set of effectors a Rac protein engages, and hence the cellular response it induces, might also depend on how or where it has been activated. Trio, for example, might activate Rac proteins to a level, for a duration, or in a subcellular location, that allows it to stimulate only those effector pathways that control motility and guidance. Exploring the basis for specificity in Rac function is an important task for the future (Hakeda-Suzuki, 2002).
Bases in 5' UTR - 265
Bases in 3' UTR - 923
Rac1 is 92% identical to human Rac1 (Luo, 1994). Rac1 and Rac2 of Drosophila are are 93% identical (Harden, 1995).
Rac, a small GTPase in the Ras superfamily, regulates at least two biological processes in animal cells: (1) the polymerization of actin and the assembly of integrin complexes to produce lamellipodia and ruffles; and (2) the activity of an NADPH oxidase in phagocytic cells. NADPH oxidase activation is mediated through a Rac effector protein, p67phox. Using chimeras made between Rac and the closely related GTPase, Rho, two distinct effector sites have been identified in Rac, one N-terminal and one C-terminal, both of which are required for activation of p67phox. The same two effector sites are essential for Rac-induced actin polymerization in fibroblasts. p65PAK, a ubiquitous serine/threonine kinase, interacts with Rac at both the N- and C-terminal effector sites, but the GTPase-activating protein, bcr interacts with Rac in a different region. This makes p65PAK a candidate effector of Rac-induced lamellipodium formation (Diekmann, 1995).
date revised: 2 June 2002
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.