The Intersectin Family
The Intersectin family refers to a group of proteins that share N-terminal EH domains, a central coiled-coil domain, and multiple SH3 domains. In mammals, there are Intersectin1 and 2 genes (Itsn1 and Itsn2, respectively), whereas in invertebrates, only one gene appears to be present.
Drosophila Dap160 and vertebrate Itsn1 are very highly expressed in the nervous system (Roos and Kelly 1998; Hussain et al. 1999; Sengar et al. 1999). A second, widely expressed Itsn gene has also been identified. This gene, originally designated Ese2, is now known as Itsn2 (Sengar et al. 1999; Pucharcos et al. 2000). There are many speculated functions for Itsn proteins in the literature. Based on protein domain structure and binding partners, as well as a series of overexpression experiments in cultured cells, it was accepted early on that Itsn proteins play a role in clathrin-mediated endocytic trafficking and in regulation of the actin cytoskeleton (Hussain et al. 1999, 2001; Sengar et al. 1999). Tissue culture experiments have been used to assign biological context to the endocytic and signaling functions. Indeed, the number of processes where Itsn proteins are reported to function has continued to grow over the past decade, mostly based on in vitro or ex vivo experiments. The first genetic studies on Itsn were performed in the fruit fly, where Dap160 was found to control synaptic structure and vesicle recycling at the neuromuscular synapse (see below) (Koh et al. 2004; Marie et al. 2004). Subsequent studies in nematodes have confirmed a role in synaptic vesicle recycling (Rose et al. 2007; Wang et al. 2008). Itsn1 mutant mice have also been reported (Yu et al. 2008). While these mice were born at the expected Mendelian frequency, some failed to thrive and died at a young age with evidence of enlarged endosomes in brain sections. These abnormalities were not seen in the majority of mutant animals, although exocytic and endocytic defects were observed in ex vivo cultures from the mice.
A number of studies in yeast have shown that endocytic proteins can regulate actin, and conversely, that actin regulatory proteins frequently control endocytosis. More recently, this phenomena has been demonstrated in higher organisms. Indeed, Wasp proteins, which regulate actin, also control internalization and trafficking of cell surface receptors. Consistent with this, mutants in Drosophila Wasp or its activator small GTPase, Cdc42, also enhance synaptic growth in a BMPR-dependent manner (Rodal et al. 2008; Nahm et al. 2010). Interestingly, Wasp and Dynamin are thought to compete for overlapping site(s) on Dap160, suggesting that Dap160-Wasp and Dap160-Dynamin complexes may perform distinct functions, perhaps during different steps of internalization and/or endocytic trafficking of BMPR (Rodal et al. 2008). Thus, at the periactive zone that surrounds each neuronal synapse, a region thought to perform functions in common with recycling endosomes of non-neuronal cells, Dap160/Itsn forms a complex with Wasp and Nwk to downregulate BMPR signaling (Fig. 2) (Rodal et al. 2008). In addition, Dap160/Itsn also forms complexes with Dynamin and Nwk, perhaps following dissolution of the Wasp complex (O’Connor-Giles et al. 2008). One or both of these complexes is likely associated with suppression of BMPR signaling, or recycling through the Rab11 compartment for return to the plasma membrane (Rodal et al. 2008). The Itsn partner Eps15 is also involved in these processes as Eps15 mutant neuromuscular junctions show the same overgrowth phenotype (Koh et al. 2007). In mammals, Itsn-1 also binds to TUC-4b, an orthologue of C. elegans UNC-33, which is thought to control neurite extension and branching (Quinn et al. 2003). Finally, in the differentiated synapse, Intersectins function within the periactive zone to coordinate events involved in efficient synaptic vesicle recycling. This function is independent of Nwk, and therefore involves distinct Itsn-protein complexes (see next section below).
While genetic studies in flies have shown a very important role for Dap160 and its orthologues in control of pre-synaptic growth, other work has revealed post-synaptic functions for Itsn. For example, the N-terminal EH domain region of Itsn1L binds to the kinase domain of EphB2 receptors in cultured hippocampal neurons (Irie and Yamaguchi 2002). Activation of EphB2 receptor signaling thus leads to activation of the DBL/PH domain Cdc42 GEF activity coded within the C-terminus of Itsn1L. N-Wasp and the endocytic protein Numb are bound to distinct SH3 domains of Itsn1L in this complex (Nishimura et al. 2006). This helps alleviate intrinsic SH3-mediated repression of the Itsn GEF domain and thereby enhances GTP-loading of Cdc42. In cultured neurons, Itsn1L colocalizes with F-actin at dendritic spines (Thomas et al. 2009). In turn, Cdc42-GTP and Itsn1L SH3 domains stimulate N-Wasp-mediated actin polymerization and dendritic spine formation, whereas Itsn1 siRNA knockdown disrupts spine maturation (Hussain et al. 2001; Thomas et al. 2009).
Synaptic Vesicle Recycling
A complex and specialized form of endocytosis is required in neurons to recover synaptic vesicle proteins and membranes in order to facilitate sustained synaptic transmission. This process appears to include fast and slow endocytic retrieval pathways that differ in their dependence on specific proteins, and that may also differ depending on whether they occur at small or large synapses, or in different synapses within the CNS. Based primarily on work in flies, a role for Itsn proteins in synaptic vesicle recycling has been definitively established (Pechstein et al. 2010b). Indeed, loss of Dap160 impaired membrane retrieval and caused altered localization and expression of key endocytic proteins Dynamin, Synaptojanin, Endophilin, LAP (Like-AP180) as well as the reserve pool marker Synapsin (Koh et al. 2004; Marie et al. 2004). Likewise, mutants of LAP in Drosophila show mislocalization of the presynaptic Ca++-sensor and cargo protein Synaptotagmin as well as Dap160. Proteomic analysis of Synaptotagmin complexes from rat brain synaptosomes has identified a complex comprised of Synaptotagmin, Stonin2, AP180, Itsn1, Epsin, Clathrin, Dynamin, and AP-2 (Khanna et al. 2006). These findings in conjunction with in vitro identification of binding partners strongly support the notion that Itsn proteins act as scaffolds to regulate activity and localization of the endocytic machinery. For example, in giant reticulospinal synapses, Itsn1 is localized to synaptic vesicle clusters where it sequesters Dynamin prior to membrane depolarization. Following K+ stimulation, however, Itsn1 redistributes to Clathrin-coated pits in the periactive zone where it recruits AP-2 to the budding vesicle and targets Dynamin to sites of vesicle fission (Evergren et al. 2007; Pechstein et al. 2010a). In the chicken calyx presynaptic terminal, an endocytic complex that includes Itsn1S and AP180 is tightly bound to CaV2.2 N-type calcium channels that cluster at transmitter release-sites. This complex also associates with a Dynamin, Clathrin, and Itsn1L containing subcomplex to facilitate rapid vesicle recovery (Khanna et al. 2007). Consistent with this conserved function, hippocampal neurons cultured from Itsn1 knockout mice show impaired synaptic vesicle recycling (Yu et al. 2008).
While most efforts have focused on dissecting the role of Itsn in endocytosis, there is some evidence to suggest that Itsn may function in exocytosis as well. Indeed, Itsn1 can bind directly to the SNARE protein, Snap25 (Okamoto et al. 1999), while cultured chromaffin cells from Itsn1 knockout mice and siRNA knockdown in the same cell type showed reduced secretion and lower Cdc42-dependent exocytic activity (Malacombe et al. 2006; Yu et al. 2008). Once again, details of how Itsn1 coordinates exo-endocytic trafficking remain to be determined.
Additional Functions for Itsn
Intersectins are also expressed in non-neuronal tissues and are therefore expected to have non-neuronal functions (Sengar et al. 1999). In support of this, Itsn1S has been implicated in endocytosis of the renal K+ channel ROMK1 and Itsn2L in T-cell antigen receptor endocytosis (McGavin et al. 2001; He et al. 2007). Itsn1 and 2 have also been reported to regulate caveolar-based endocytosis. In endothelial cells, Itsn1S associates with Dynamin and Snap23 patches to help facilitate caveolae-mediated internalization (Predescu et al. 2003, Klein et al. 2009). In contrast, knockdown of Itsn2L by siRNA increased endocytosis through caveolae, whereas overexpression of the Cdc42-activating GEF domain of Itsn2L blocked it, suggesting that Itsn2L-dependent actin polymerization may inhibit caveolae-mediated endocytosis (Klein et al. 2009).
Itsn1 is also thought to control cell survival, polarity and mitosis. For example, the Itsn1 mouse knockout has decreased levels of NGF in the septal region of the brain, suggesting a role for Itsn1 in NGFR endocytosis or trafficking (Yu et al. 2008). Indeed, Itsn1 regulates survival signaling (Loeb et al. 2006; Das et al. 2007; Predescu et al. 2007). siRNA knockdown of Itsn1 in differentiated neurons causes cell death, which may be associated with loss of an Itsn1/PI3K-C2β/Akt survival signal (Das et al. 2007). siRNA knockdown of Itsn1 in endothelial cells also causes cell death (Predescu et al. 2007). Itsn1 can activate Epidermal Growth Factor Receptor (EGFR) internalization and signaling to the mitogenic transcription factor, Elk-1. Regulation of this pathway may involve Sos, Ras, and Jnk signaling but not Erk1/2 (Mohney et al. 2003). Expression of a dominant active form of Itsn1L in NIH3T3 cells was strongly transforming as a result of Cdc42 and Ras activation (Wang et al. 2005). In Drosophila, Dap160 directly binds and activates aPKC in an endocytosis-independent pathway to modulate neuroblast polarity and cell cycle progression (Chabu and Doe 2008). In mammalian cells, Itsn2 is thought to control activation of Cdc42 near centrosomes in order to control spindle orientation and lumen formation. In this context, Itsn2L SH3 domains bound to p150Glued, a subunit of the Dynactin complex (Rodriguez-Fraticelli et al. 2010). Interestingly, a Dynactin complex protein was also identified in association with C. elegans ITSN1 (Wang et al. 2008).
Thus, a series of elegant studies in organisms ranging from flies to worms to mice have revealed a conserved role for Itsn proteins in coordinating endocytosis and signaling in neurons. For example, Itsn functions in the periactive zone of neuromuscular junctions to regulate BMP-mediated synaptic growth (O’Connor-Giles et al. 2008). It also functions in distinct complexes to coordinate synaptic vesicle recycling at this location. Many non-neuronal Itsn functions have been identified on the basis of in vitro or ex vivo experiments. Genetic studies in Drosophila, C. elegans and mice have yet to support such roles (Koh et al. 2004; Marie et al. 2004; Rose et al. 2007; Wang et al. 2008; Yu et al. 2008). This may be due to redundancy between Itsn and other scaffolding or signaling proteins, and also to an early focus of genetic studies on synaptic transmission. Since an Itsn2 knockout mouse has yet to be described, it is difficult to interpret the rather limited effect of deleting Itsn1 in mice (Yu et al. 2008). Indeed, Itsn2 deletion may cause major problems in endocytosis and/or trafficking of channel proteins, antigen receptors, caveolae-based internalization, or lumen formation. Of course, Itsn1 and Itsn2 may function redundantly in a number of processes. Double mutant mice will be required to test this possibility. Beyond expected redundancies based on the homology of Itsn1 and 2 with each other, or homologies between Itsn proteins and other proteins with EH domains, SH3 domains, or DBL/PH domains that hint at shared functions, a small screen for synthetic lethal or synthetic sick interactions in C. elegans has revealed that Itsn proteins can function redundantly with unrelated proteins like Disabled, Dab-1 (Wang et al. 2008). This redundancy may be based on parallel Itsn- and Dab-mediated pathways for endocytosis and/or signal transduction downstream of common cargo proteins. Future genetic studies are likely to yield insights into novel functions for Itsn proteins in neurons and many other cell types.