Dock180 was originally identified as a binding protein for the SH3 domain of the proto-oncogene product c-Crk through its C-terminal PxxP region [7
] (Box 1
). Later on, biochemical studies in mammalian cells demonstrated that Dock180 was positioned upstream of the Rho family member Rac [8
]. Subsequent genetic screens in Caenorhabditis elegans
and Drosophila melanogaster
suggested that the Dock180 orthologs in these organisms were also functioning upstream of Rac in a range of biological events, including phagocytosis of apoptotic cells, migration of gonad cells and myoblast fusion [9
]. Studies in C. elegans
have especially paved the way to a more detailed understanding of the regulation of the Dock180 signaling pathway.
Ced-12 (ortholog of mammalian Elmo; ‘Ced’ standing for ‘cell death abnormal’) was identified as a component of a genetic signaling cascade in C. elegans
that also contains Ced-2 (ortholog of mammalian c-Crk) and Ced-5 (ortholog of mammalian Dock180). This signaling cascade controls the activity of Ced-10 (ortholog of mammalian Rac) during cell migration and phagocytosis of apoptotic cells in C. elegans
). Ced-12 is an evolutionarily conserved protein, and three family members, Elmo1, -2 and -3, have been identified in mammals.
Box 2. Overview of the roles of Ced-5 and MBC, the orthologs of mammalian Dock180, in worms and flies, respectively
Genetic screens in C. elegans
were central to the identification of genes and gene products involved in the regulation of programmed cell death, or apoptosis (). Pioneering work led to the discovery of several Ced
genes and their positioning in genetic signaling cascades. Among these, two distinct sets of genes were shown to regulate the engulfment of dying cells, which is the last step in apoptosis, during embryogenesis: Ced-1
genes [orthologs of mammalian MEGF-10 (multiple EGF-like domains-10)] Gulp and ABC transporter, respectively) form one signaling pathway, and Ced-2
genes (orthologs of mammalian c-Crk, Elmo, Dock180 and Rac, respectively) form the other [46
]. Because Ced-5 and Ced-12 bind directly to a PtdSer receptor in phagocytes, the Ced-5–Ced-12 complex has been proposed to integrate the activation of the Ced-2–Ced-5–Ced-12–Ced-10 signaling cascade during the early stage of recognition of apoptotic cells, exposing the ‘eat-me’ signal PtdSer [23
] (). Interestingly, mutations in the Ced-2
genes result not only in defects in the engulfment process, but also in a failure of distal tip cells of the gonads to migrate correctly [26
], and also in defects in the outgrowth of D-type motor neurons and in the migration of P-cells during brain development [47
]. Presently, the molecular mechanisms downstream of Ced-10 that lead to cell motility and phagocytosis remain to be elucidated.
Figure 3 Mechanisms to target Dock180 to the membrane through Elmo. (a) Elmo contains a RhoG-binding domain (in yellow), which specifically recognizes the GTP-bound form of the RhoG GTPase. Following cell adhesion to fibronectin, GTP-bound RhoG recruits the Elmo–Dock180 (more ...)
MBC, the Drosophila
ortholog of Dock180, was discovered in a screen for genes controlling the fusion of myoblasts into multi-nucleate muscle fibers [9
]. MBC was proposed to function directly upstream of dRac (Drosophila
Rac protein) during myoblast fusion [48
]. Interestingly, MBC mutant myoblasts can still align with one another and make initial cell–cell contact, but they are unable to remodel their membranes for fusion. MBC is also involved in, but is not absolutely required for, at least three different types of cell migration events in Drosophila
. First, the migration of border cells toward the oocyte, in which the migratory attractant Platelet-derived growth factor-vascular endothelial growth factor (PVF) is highly expressed, is regulated by the platelet-derived growth factor–vascular endothelial growth factor receptor (PVR) and the Rac GTPase. Clonal analyses revealed that among several candidate genes that regulate Rac activation, only MBC mutations significantly delayed the migration of border cells [49
]. Second, halfway through embryogenesis, MBC is expressed in the epidermis and seems to have an important role in the movement of the ventral and lateral epidermis to surround the embryo. This process, termed ‘dorsal closure’, involves a collective movement of epithelial cells, whereby the leading edge cells guide the movement of the sheets. In MBC mutant embryos, similarly to Rac mutants, F-actin is significantly less abundant in the leading edge cells, suggesting that the MBC–Rac signaling pathway might directly regulate the actin cytoskeleton; however, this remains to be demonstrated [34
]. Third, during metamorphosis, in a process termed ‘thorax closure’, the dorsal regions of the wing imaginal discs migrate towards each other, eventually to fuse at the midline to form the notum. This process is similar to dorsal closure, and many of the same genes are required for both events. A kinase cascade through MBC, Rac and c-Jun N-terminal kinase (Jnk) seems to be essential, downstream of the PVR, for thorax closure to occur normally [50
The Elmo proteins seem to be scaffold proteins, with no obvious catalytic activity. They share conserved domain features, including armadillo repeats at the N-terminus, an atypical PH domain, and a complex proline-rich region at the extreme C-terminus (). As depicted in , Elmo1 and Elmo2 proteins have been shown to interact physically with four mammalian Dock180 proteins that contain an SH3 domain, namely Dock180 (also known as Dock1), Dock2, Dock3 and Dock4 [11
] (Dock5 also contains an SH3 domain but its interaction with the Elmo proteins has not been experimentally demonstrated to date). The mechanism for interaction between Elmo and Dock180 seems to be complex and remains to be fully elucidated. Although the SH3 domain of Dock180 binds to the proline-rich region in Elmo, this binding is dispensable in coimmunoprecipitation and pull-down experiments [13
]. Clearly, the crucial domains required for the Elmo–Dock180 interaction remain to be mapped [4
]. What are the consequences of the Elmo–Dock180 interaction? Two independent (but not mutually exclusive) roles for Elmo in Dock180–Rac signaling that are now being uncovered are discussed here.
Figure 1 Schematic representation of the Elmo proteins and their interaction with the Dock180-related proteins. A large portion of the C-terminal domain of Elmo, including part of the PH domain and the proline-rich region (black bar), is required for Elmo to interact (more ...)
The Dock180–Elmo complex and the bipartite GEF model
Some studies suggest that Dock180 is catalytically active toward Rac only when in complex with Elmo [4
]. In support of this hypothesis, deletion mutants of Dock180 that fail to bind to Elmo do not significantly activate Rac when over-expressed in cells [11
]. One possible explanation put forth for this ‘synergy model’ is that the binding of Elmo to Dock180 increases the affinity of this protein complex towards nucleotide-free Rac, thus favoring the formation of a key intermediate during the catalysis of GDP for GTP exchange [4
]. Mechanistically, the atypical PH domain of Elmo was shown to be the key determinant in increasing the catalytic activity of Dock180 towards Rac. The PH domain of Elmo does not interact directly with either Dock180 or Rac. Instead, it can bind ‘in trans
’ to the Dock180–nucleotide-free Rac complex to stabilize a trimolecular complex between the three proteins [13
] (). This action of the Elmo PH domain was shown modestly to increase the catalytic activity of Dock180 toward Rac, by about twofold in vitro.
Nevertheless, it was found that this effect could be significant in vivo
because the expression of a form of Ced-12 with mutations in the PH domain failed to rescue the migration defects in Ced-12-null worms [13
]. However, many groups have demonstrated that overexpression of Dock180 alone is sufficient to activate Rac, and that an isolated DHR-2 domain of Dock180 (in the absence of Elmo) efficiently exchanges on Rac both in vitro
and in mammalian cells [14
] (Box 1
). Clearly, further studies are needed to reveal the exact contribution of Elmo in the Dock180-mediated activation of Rac. Elmo is clearly crucial for efficient Rac signaling by Dock180 during cell migration and phagocytosis. This could, in part, also be due to the ability of Elmo to localize the complex in defined cellular compartments, which is instrumental for the subsequent activation of biological signaling events downstream of Rac. This possibility is discussed later.
Figure 2 The bipartite GEF model. In this model, the GEF activity of Dock180 toward the Rac GTPase becomes detectable, or is greatly enhanced, when Dock180 is bound to Elmo. Mechanistically, it is thought that the PH domain of Elmo interacts in trans with the (more ...)
Elmo might aid targeting of Dock180 to the plasma membrane
Four different mechanisms for recruiting the Elmo–Dock180 complex to specific binding partners at the plasma membrane have recently been uncovered (). These partners include RhoG, Arf6, the phosphatidylserine (PtdSer) receptor and the IpgB-1 protein from Shigella. The potential importance of these interactions is now discussed.
Recent studies have demonstrated the N-terminal portion of Elmo to be a binding partner for the constitutively active RhoG GTPase [15
]. It was shown that the recruitment of the Elmo–Dock180 complex to the active RhoG molecule, which is mainly located at the plasma membrane, is crucial for efficient Rac-dependent epithelial cell spreading on the matrix protein fibronectin [15
]. Thus, abrogation of the Elmo–Dock180 complex by dominant-negative proteins or by small interfering RNAs (siRNAs) against Elmo, RhoG or Dock180 significantly delayed cell spreading and reduced cell migration on fibronectin [15
]. In addition, in PC 12 rat pheochromocytoma cells, interfering with the RhoG–Elmo pathway blocked Rac-dependent neurite outgrowth induced by both neural growth factor and serum. Interestingly, RhoG has also been shown to mediate the recruitment of the Elmo–Dock180 complex to the plasma membrane during engulfment of latex beads that mimic apoptotic cells [18
]. In C. elegans
, UNC-73 (mammalian Trio), which is a GEF for Mig-2 (mammalian RhoG), functions as the direct upstream activator enabling coupling between Mig-2 and Ced-12, triggering the Ced-5-mediated activation of Ced-10 at the membrane during engulfment [18
]. These genetic results are not completely understood because a double mutation in the UNC-73 and Mig-2 genes does not induce engulfment defects. Nevertheless, the UNC-73 Mig-2
double mutant enhances the defects observed in Ced-2
gene mutants. The biological importance of this signaling cascade in mammalian cells remains to be fully determined because deletion of the RhoG
gene in mice does not result in any obvious phenotype [19
Similarly to RhoG, Arf6, another GTPase known to activate Rac, could also mediate the recruitment of the Elmo–Dock180 complex to the plasma membrane [20
]. Thus, Arno, which is a specific GEF for Arf6, has been shown to promote strong targeting of Elmo–Dock180 to the membrane, and dominant-negative forms of both Dock180 and Elmo block Arno-induced lamellipodia formation and cell migration in a Madin–Darby canine kidney cell model [20
]. It is not clear whether a direct interaction between Arf6 and Elmo (or Dock180) exists. Interestingly, in Drosophila
, Arf6 and its GEF, Loner, which is an Arno-related Sec7 domain-containing GEF, were shown to regulate Rac localization and activation during myoblast fusion [21
]. It will be interesting to investigate if Arf6 and Loner do so by recruiting the dElmo–Myoblast City (MBC; Drosophila
Dock180) complex to promote dRac activation during myoblast fusion.
During apoptosis, dying cells expose the lipid PtdSer on the outer side of the plasma membrane as an ‘eat-me’ signal [22
]. Phagocytes subsequently recognize this signal to engulf the apoptotic cells through a PtdSer receptor. Identification of the receptor for PtdSer in mammals preceded its discovery in C. elegans
. However, genetic studies in worms placed the PtdSer receptor in the same pathway as Ced-2, Ced-5, Ced-10 and Ced-12 during engulfment of cell corpses [23
]. Mechanistically, both Ced-5 and Ced-12 have been shown to interact directly with the PtdSer receptor in in vitro
binding assays [23
]. This suggests that the recruitment of the Ced-10-activating machinery at the membrane can be accomplished through specific interactions with the PtdSer receptor. It remains to be determined whether the mammalian PtdSer receptor(s) also directly recruit Elmo–Dock180 following ligand binding. It also raises the question of the role of Ced-2 in this pathway because this molecule was expected to have such a receptor-targeting role. Further studies are thus needed to understand the precise roles of these genes in the engulfment of apoptotic cells.
It has been shown that the evolutionarily conserved pathways for phagocytosis described earlier are used by some pathogens to infect mammalian cells. Thus, phagocytic-like events regulate the ability of certain types of bacteria, such as Shigella
, to enter intestinal epithelial cells. To promote membrane protrusions, Shigella
injects proteins into epithelial cells, and one of these bacterial proteins, IpgB1, was recently shown to bind directly to Elmo and recruit the Elmo–Dock180 complex to the membrane [24
]. Interfering with IpgB1–Elmo complex formation significantly decreased the ability of Shigella
to infect epithelial cells in culture [24
]. Inhibition of the Elmo–Dock180 pathway could thus provide a novel means to prevent infection by this pathogen.