Breaking News

Ubiquitylation of ubiquitin-binding proteins

Experiments carried out by van Delft et al. showed that, upon stimulation of cells with EGF or transforming growth factor alpha, Eps15 is transiently phosphorylated on tyrosine residues and is modified by monoubiquitylation [175]. It was recently reported that two UIM motifs at the extreme C-terminus of Eps15 are essential for the monoubiquitylation of this molecule [105], and for the binding of Eps15 to ubiquitin-containing proteins, with a preference for polyubiquitin chains. Mutations in the first UIM motif abolish the monoubiquitylation of Eps15 but do not affect the ability of the protein to bind to ubiquitin-binding proteins, whereas mutations in the distal motif did both [105, 176]. Finally, Eps15 UIM domains are required for ubiquitylation but are not the sites of ubiquitylation [105, 176]. Eps15 ubiquitylation appears to be Nedd4-dependent [105]. Initial attemps at understanding the function of Eps15 ubiquitylation by construction of various mutant forms of Eps15 with deletions of the UIM-containing region, or point mutations that abolish ubiquitylation have not revealed clear effects, as the mutant proteins were correctly targeted to clathrin-coated pits, and internalization of certain receptors was not inhibited [176, 177].
Very similar results were obtained for mammalian epsins, known Eps15. Interestingly, epsin possesses two UIMs (Fig. 1). It was demonstrated that epsin is predominantly monoubiquitylated and that these UIM motifs are necessary for epsin ubiquitylation but that they are not the site of ubiquitylation [178]. Epsin is possibly ubiquitylated in its ENTH domain. This domain interacts with phosphoinositides. This interaction leads to conformational changes, suggesting that it may initiate membrane curvature (reviewed in [179]). Epsin ubiquitylation may thus regulate these events. The observation that mammalian epsin is ubiquitylated confirms earlier genetic reports on liquid facets, the Drosophila epsin. Mutations in liquid facets were found to increase endocytic defects associated with mutations in the gene encoding the de-ubiquitylating enzyme Fat facets [180]. Formal prove that Fam, the mammalian homolog of Fat facet is indeed the specific ubiquitin isopeptidase responsible for epsin deubiquitylation was recently obtained in experiments where si-RNA-mediated suppression of Fam inhibits in vivo epsin deubiquitylation triggered by Ca2+ influx into synaptosomes [181]. These overall data indicate that epsin deubiquitylation plays a key role in the endocytic process that remains to be determined.
Isolated UIMs from both mammalian epsins and Eps15 are sufficient to promote the ubiquitylation of a chimeric glutathione-S transferase (GST)-UIM fusion protein, suggesting that UIMs may serve as a general signal for ubiquitylation [178]. Thus, the same motif in several endocytic proteins may be responsible for ubiquitin recognition and monoubiquitylation, a notion that extends to UIM-containing proteins involved in MVB sorting [3], and to a subset of proteins containing other ubiquitin binding domains. In the case of UIM-containing proteins, a general rule seems to emerge. Ubiquitin attachment occurs outside the UIM domain, is limited to monoubiquitylation, and catalyzed by HECT E3, that can also catalyze polyubiquitylation of different types. One possible interpretation of these observations is that the E3/UIM recognition is mediated by the ubiquitin present in the thiol-ester intermediate of the E3. Once ubiquitin is transferred to a UIM-containg protein, the E3 would dissociate, yielding a monoubiquitylated substrate. 
Eps15 and epsins have yeast homologs that have been shown to be involved in endocytosis [38, 61], Deletion of the EDE1 gene, or of one of the two ENT genes, together with a heat-sensitive mutation in the second gene (the deletion of both ENT genes is lethal) leads to defective fluid-phase endocytosis and to the defective internalization of Fur4p, Ste6p, and Ste2p, respectively [38, 61, 182]. The yeast epsins, Ent1p and Ent2p, each display two UIMs, and Ede1p, a yeast homolog of the mammalian Eps15, has a C-terminal UBA domain (Fig. 1). The yeast epsins and Ede1p were shown to bind monoubiquitin in vitro, in a way dependent of several conserved UIM residues [182], and the two UIMs cooperate for the interaction [183]. The two UIMs of epsins may display functional redundancy in endocytosis with Ede1p, possibly due to the ubiquitin-binding properties of the UBA domain of Ede1p [182]. Data reported by Aguilar and Wendland slightly modify the view of Ede1p and Ent1/2 properties. These authors showed that Ede1p and Ent1p bind one another, like their mammalian counterpart. The UBA Domain of Ede1p binds yeast membranes in a ubiquitin-dependent way, and the ENTH domain of Ent1p cooperates with the UIMs for membrane recruitment [183]. These data emphasize that the properties of these ubiquitin-binding proteins are intimately linked to their interactions with lipids. Whereas these data are compatible with the attractive hypothesis of Ent/epsin acting as adaptors for ubiquitylated plasma membrane proteins, the formal proof of an interaction in situ with a precise ubiquitylated cargo is still lacking, as are informations relative to the affinity of UBA/UIM domains of these proteins for monoubiquitin versus ubiquitin chains (UbK63-based). It will also be important to determine whether Ent1/2p and Ede1p are ubiquitylated, like mammalian epsins and Eps15.

No comments