Structural Complexity in Ubiquitin Recognition
2006; Cell Press; Volume: 124; Issue: 6 Linguagem: Inglês
10.1016/j.cell.2006.03.009
ISSN1097-4172
AutoresJ. Wade Harper, Brenda A. Schulman,
Tópico(s)Autophagy in Disease and Therapy
ResumoUbiquitinated proteins are sorted into distinct pathways via association with several classes of ubiquitin binding domain-containing proteins. A virtual explosion in the field of ubiquitin binding proteins has revealed several new classes and interactions with distinct surfaces on ubiquitin, providing a clearer understanding of how sorting of ubiquitinated proteins is achieved. Ubiquitinated proteins are sorted into distinct pathways via association with several classes of ubiquitin binding domain-containing proteins. A virtual explosion in the field of ubiquitin binding proteins has revealed several new classes and interactions with distinct surfaces on ubiquitin, providing a clearer understanding of how sorting of ubiquitinated proteins is achieved. Protein modification by ubiquitin and ubiquitin-like (Ubl) proteins has emerged as a central mechanism through which cellular pathways are regulated (Pickart, 2004Pickart C.M. Cell. 2004; 116: 181-190Abstract Full Text Full Text PDF PubMed Scopus (567) Google Scholar). In this process, the C-terminal glycine of ubiquitin becomes linked to primarily lysine residues in target proteins via an E1-E2-E3 cascade. As ubiquitin itself contains seven lysine residues, multiple molecules of ubiquitin can also become linked to each other to form polyubiquitin chains. Thus, ubiquitination can take many forms which can differentially control the fate of the target protein. The best understood function of ubiquitination is proteolysis, whereby lysine-48 (K48)-linked polyubiquitin chains allow recognition by the 26S proteasome. However, proteins can also be monoubiquitinated or polyubiquitinated through alternative (e.g., K63) linkages, and such modifications are thought to control protein activity or localization (Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar). The ubiquitination event can be viewed as a molecular zip code, which is used to sort different ubiquitination products to different destinations. Errors in delivery of ubiquitinated proteins to the proteasome or other destinations could be disastrous for cells. Ubiquitin zip codes are read by several families of ubiquitin binding domain (UBD)-containing proteins that recognize a conserved “hydrophobic pocket” on ubiquitin centered at isoleucine-44 (I44; Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar). Now, new work—including two papers in this issue of Cell—reveal additional molecular mail carriers that sort and process ubiquitinated proteins and provide evidence that additional surfaces of ubiquitin are recognized by different classes of UBDs (Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar Hirano et al., 2006Hirano S. Kawasaki M. Ura H. Kato R. Raiborg C. Stenmark H. Wakatsuki S. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 272-277https://doi.org/10.1038/nsmb1051Crossref PubMed Scopus (143) Google Scholar, Hoeller et al., 2006Hoeller D. Crosetto N. Blagoev B. Raiborg C. Tikkanen R. Wagner S. Kowanetz K. Breitling R. Mann M. Stenmark H. Dikic I. Nat. Cell Biol. 2006; 8 (Published online January 22, 2006): 163-169https://doi.org/10.1038/ncb1354Crossref PubMed Scopus (251) Google Scholar, Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Mattera et al., 2006Mattera R. Tsai Y.C. Weissman A.M. Bonifacino J.S. J. Biol. Chem. 2006; 281: 6874-6883Crossref PubMed Scopus (93) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar, Reyes-Turcu et al., 2006Reyes-Turcu F.E. Horton J.R. Mullally J.E. Heroux A. Cheng X. Wilkinson K.D. Cell. 2006; (this issue)PubMed Google Scholar). These studies further our understanding of how specificity is achieved. One important question is whether the code for ubiquitin recognition will be as simple as one UBD = one mode of ubiquitin recognition. Studies of other types of protein-protein interaction motifs have revealed diversity in the ways a given class of protein interaction module can interact with partners, and the fact that distinct ubiquitin binding CUE domains bind ubiquitin in substantially different ways (Figure 1A; Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar) indicates that such complexity may exist with UBDs. Indeed, recent work on by Hirano et al., 2006Hirano S. Kawasaki M. Ura H. Kato R. Raiborg C. Stenmark H. Wakatsuki S. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 272-277https://doi.org/10.1038/nsmb1051Crossref PubMed Scopus (143) Google Scholar on Hrs, a ubiquitin-interaction motif (UIM)-containing protein involved in receptor endocytosis (Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar), suggests that this is the case. The structure of Hrs-UIM in complex with ubiquitin revealed a new subclass of UIMs. In the structure, the Hrs-UIM binds two ubiquitins, in stark contrast with previously characterized UIM domains which bind one (Figures 1A and 1B). In the Hrs-UIM complex, both ubiquitin molecules (Ub1 and Ub2) use their I44 surface to interact with UIM, but they bind on opposite sides of the helix. Importantly, the interaction surface for Ub1 and Ub2 are very similar, reflecting two conserved ubiquitin-interacting sequences within the single UIM. The two ubiquitin-interacting sequences are displaced by two residues. Can one-sided and two-sided UIMs be distinguished on the basis of sequence? Hirano et al., 2006Hirano S. Kawasaki M. Ura H. Kato R. Raiborg C. Stenmark H. Wakatsuki S. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 272-277https://doi.org/10.1038/nsmb1051Crossref PubMed Scopus (143) Google Scholar aligned UIM motifs from several proteins and found that the two classes of UIM can be identified by the presence or absence of an internal repeat sequence e-x-e-x-Φ-x-Φ-A-Φ-A-z-S-z-A/S-e (where e is a negatively charged residue, x is any residue, z is a large hydrophobic or polar residue; underlined residues indicate binding site for Ub2; Φ, hydrophobic residue). All UIMs which maintain alanine at position 10 and either serine or alanine at position 14 are expected to form two-sided UIMs. This idea was substantiated via biochemical analysis of three two-sided UIMs (Hrs-UIM, Hsj1-UIM2, and Eps15-UIM2). Importantly, examination of EGF endocytosis in cells expressing mutations in either or both of the UBDs reveals that interaction of both sides of the UIM are required for full Hrs function. What is the purpose of two-sided UIMs? One possibility is that having a two-sided UIM provides additional avidity in recognition of targets, particularly those targets that contain either a polyubiquitin chain or contain multiple monoubiquitins. In principle, this could be achieved by having two UIM domains in the same polypeptide. Indeed, many UIM-containing proteins contain tandem UIM domains. However, it is also the case that some proteins that contain two-sided UIMs also have tandem UIMs, sometimes of the one-sided class and sometimes of the two-sided class (Hirano et al., 2006Hirano S. Kawasaki M. Ura H. Kato R. Raiborg C. Stenmark H. Wakatsuki S. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 272-277https://doi.org/10.1038/nsmb1051Crossref PubMed Scopus (143) Google Scholar). An alternative possibility is that two-sided and tandem UIMs have evolved to bind ubiquitinated proteins which present multiple ubiquitin molecules in a particular stereotyped architecture with respect to each other. Further studies are required to distinguish these possibilities and to establish any functional differences between one- and two-sided UIMs. Two other new UBDs have been identified from studies of the Rabex-5 GTPase exchange factor, the mammalian ortholog of yeast VPS9 (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Mattera et al., 2006Mattera R. Tsai Y.C. Weissman A.M. Bonifacino J.S. J. Biol. Chem. 2006; 281: 6874-6883Crossref PubMed Scopus (93) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). A C-terminal CUE-UBD (Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar) allows recruitment of VPS9 to ubiquitinated cell surface receptors at the plasma membrane or the endosome, facilitating Rab activation. Upon EGF stimulation, the EGF receptor (EGFR) gets ubiquitinated, which recruits Rabex-5 to the plasma membrane (Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). However, it was not clear how Rabex-5 is recruited to ubiquitin because Rabex-5 lacks strong sequence identity with VPS9 in its C terminus. Recent work revealed the presence of two new functionally distinct and independent UBDs in the N terminus of Rabex-5: (1) a linear peptide sequence that resembles a UIM (described below), and (2) the ZnF_A20 motif (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). The first motif has been named “MIU” (motif interacting with ubiquitin) or “IUIM” (inverted UIM). Both the MIU/IUIM and the ZnF_A20 sequences, which are capable of binding both monomeric ubiquitin and K48-linked chains, are conserved in Rabex-5 orthologs as well as in proteins with unrelated functions. The Rabex-5 MIU/IUIM interacts with ubiquitin's I44 hydrophobic patch (Figure 1C). However, the sequence orientation of the MIU/IUIM is inverted with respect to UIM, such that the ubiquitin C terminus is directed toward the C terminus of the MIU/IUIM domain (Figures 1A and 1C). Consistent with structural data, mutation of the MIU/IUIM's central alanine or ubiquitin's I44 dramatically reduces binding (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). The MIU/IUIM's N terminus contains an additional helical turn, providing additional contacts with ubiquitin that increase its affinity for ubiquitin by ∼10-fold relative to most UIMs. The ZnF_A20 motif represents another new UBD, distinct from two other known zinc binding UBDs, the NZF and ZnF-UBP motifs (Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar; Figure 1). The ZnF_A20 motif interacts with the D58-centered hydrophobic face of ubiquitin and represents the first UBD interacting with a surface of ubiquitin other than the I44 patch, a finding supported by mutational data (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). UBD-containing proteins often have multiple UBDs of different classes, raising the question of how these different UBDs function together or independently to recognize ubiquitinated targets. The Rabex-5 studies provide the first structural insight into combinatorial recognition of ubiquitin by multiple UBDs in a single protein. Although the structures presented contain one molecule of Rabex-5 per ubiquitin, this apparent 1:1 stoichiometry appears to be a result of crystal packing in which higher-order interactions are observed. Indeed, both studies present biophysical data indicating that both the MIU/IUIM and ZnF_A20 domains bind ubiquitin independently (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). However, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar provide evidence that two molecules of Rabex-5 can interact with a single ubiquitin, and vice-versa. Thus, multiple UBDs within a single protein may allow for higher-order complexes, with different stoichiometries that allow recognition of particular ubiquitinated targets (Figure 1C). What are the functional roles of Rabex-5 UBDs? Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar found that Rabex-5 binds ubiquitinated EGFR in EGF-stimulated cells in a manner that depends on both UBDs. Mutation of either domain alone had little impact on EGFR association, whereas mutation of both abolished the EGFR interaction. However, the effect of these mutations on the process of EGFR endocytosis and degradation has yet to be performed, so it is currently unclear whether mutants that still maintain the ability to bind EGFR will retain full biological activity. In principle, although binding of ubiquitinated EGFR through either the ZnF_A20 or MIU/IUIM motifs is possible, interaction with only one of the two domains may place EGFR in the proper configuration for appropriate downstream steps. Interestingly, the ZnF_A20 domain from Rabex-5 was found to have E3 ubiquitin ligase activity that is analogous to that seen with zinc binding RING finger E3s (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Mattera et al., 2006Mattera R. Tsai Y.C. Weissman A.M. Bonifacino J.S. J. Biol. Chem. 2006; 281: 6874-6883Crossref PubMed Scopus (93) Google Scholar). This ZnF_A20 domain binds ubiquitin-charged Ubc5 (an E2) in a manner that depends upon residues involved in ubiquitin binding. Precisely how this ZnF_A20 domain achieves specificity for particular charged E2s is unclear, as there does not appear to be direct interaction with the uncharged E2 and it does not bind charged Ubc7 (another E2 of the same family as Ubc5). Further work is required to understand how ZnF_A20 domains promote ubiquitin conjugation, whether this activity is important for coupled monoubiquitination of Rabex-5 seen in vivo, and whether other ZnF_A20 domains function similarly (Lee et al., 2006Lee S. Tsai Y.C. Mattera R. Smith W.J. Kostelansky M.S. Weissman A.M. Bonifacino J.S. Hurley J.H. Nat. Struct. Mol. Biol. 2006; 13 (Published online February 5, 2006): 264-271https://doi.org/10.1038/nsmb1064Crossref PubMed Scopus (165) Google Scholar, Mattera et al., 2006Mattera R. Tsai Y.C. Weissman A.M. Bonifacino J.S. J. Biol. Chem. 2006; 281: 6874-6883Crossref PubMed Scopus (93) Google Scholar, Penengo et al., 2006Penengo L. Mapelli M. Murachelli A.G. Confalonieri S. Magri L. Musacchio A. Paolo Di Fiore P. Polo S. Schneider T.R. Cell. 2006; (Published online February 16, 2006)https://doi.org/10.1016/j.cell.2006.02.020Abstract Full Text Full Text PDF PubMed Scopus (222) Google Scholar). The ZnF-UBP domain, which is also referred to as the polyubiquitin-associated zinc finger (PAZ) domain (Hicke et al., 2005Hicke L. Schubert H.L. Hill C.P. Nat. Rev. Mol. Cell Biol. 2005; 6: 610-621Crossref PubMed Scopus (609) Google Scholar), is found in diverse protein families, including deubiquitinating enzymes (DUBs) and RING E3s. Structural and biochemical studies from Wilkinson and colleagues now reveal a new mode of ubiquitin binding by the ZnF-UBP domain from the DUB IsoT/USP5 and demonstrate the function of this new ubiquitin recognition in the disassembly of free ubiquitin chains (Reyes-Turcu et al., 2006Reyes-Turcu F.E. Horton J.R. Mullally J.E. Heroux A. Cheng X. Wilkinson K.D. Cell. 2006; (this issue)PubMed Google Scholar). Previous studies demonstrated that although the IsoT/USP5 is capable of cleaving the artificial substrate ubiquitin-amino-methylcoumarin (AMC), the reaction is activated by free ubiquitin (Wilkinson et al., 1995Wilkinson K.D. Tashayev V.L. O'Connor L.B. Larsen C.N. Kasperek E. Pickart C.M. Biochemistry. 1995; 34: 14535-14546Crossref PubMed Scopus (253) Google Scholar). This stimulation required the C-terminal Gly-Gly sequence in ubiquitin. This suggested that the stimulatory function of free ubiquitin may mimic the action of the proximal ubiquitin in the polyubiquitin chain. However, the mechanism underlying this activation, as well as how IsoT/USP5 properly positions the polyubiquitin chain substrate for cleavage, has remained a mystery, in part because in addition to its ZnF-UBP domain, IsoT/USP5 also contains two ubiquitin-associated (UBA) domains (Figure 1). To identify the basis for IsoT's recognition of the proximal ubiquitin, Reyes-Turcu et al., 2006Reyes-Turcu F.E. Horton J.R. Mullally J.E. Heroux A. Cheng X. Wilkinson K.D. Cell. 2006; (this issue)PubMed Google Scholar looked for fragments of IsoT/USP5 that bind ubiquitin containing an intact Gly-Gly motif, but not ubiquitin in which the C-terminal Gly is mutated. Through this process, they identified the ZnF-UBP domain which binds a single ubiquitin molecule with a Kd of 3 μM, significantly tighter than most UBDs. The structure of ZnF-UBP and its complex with free ubiquitin reveals a new mode of recognition: ZnF-UBP binds the C-terminal tail of ubiquitin (Figure 1D). Thus, the ZnF-UBP-ubiquitin structure represents the second structure of a UBD that binds to ubiquitin independently of I44. The ZnF-UBP adopts a compact α/β fold, stabilized by the coordination of a zinc atom (Figure 1D). Evidence of a role for the ZnF-UBP domain in IsoT/USP5 function comes from the finding that mutations in zinc ligands reduce the rate of Ub-AMC hydrolysis and disrupt activation by free ubiquitin. ZnF-UBP also interacts with a surface centered on ubiquitin's Ile36 (Figure 1D), a residue which is required for ubiquitin function in vivo (Sloper-Mould et al., 2001Sloper-Mould K.E. Jemc J.C. Pickart C.M. Hicke L. J. Biol. Chem. 2001; 276: 30483-30489Crossref PubMed Scopus (194) Google Scholar). These data lead to a model in which the ZnF-UBP domain is critical for recognition of the proximal ubiquitin in polyubiquitin chains. In the case of DUBs, this presumably helps anchor the terminal Ub-Ub linkage in the proper orientation for cleavage. Further studies are required to understand how the catalytic domain recognizes ubiquitin and how IsoT's UBA domains help to organize the polyubiquitin chain to facilitate binding or cleavage. Reyes-Turcu et al., 2006Reyes-Turcu F.E. Horton J.R. Mullally J.E. Heroux A. Cheng X. Wilkinson K.D. Cell. 2006; (this issue)PubMed Google Scholar also found that the ZnF-UBP in the E3 IMP also binds free ubiquitin in a manner that depends on the presence of its C-terminal glycine. One speculative idea is that binding to free ubiquitin serves to stimulate the activities of enzymes which contain ZnF-UBP. Indications that surfaces of ubiquitin other than the I44 patch are functionally important (Sloper-Mould et al., 2001Sloper-Mould K.E. Jemc J.C. Pickart C.M. Hicke L. J. Biol. Chem. 2001; 276: 30483-30489Crossref PubMed Scopus (194) Google Scholar) caused Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar to search for proteins that can interact with both ubiquitin and its I44A mutant. This study identified the UBM and UBZ domains, both of which are found in Y family DNA polymerases which catalyze DNA synthesis across damaged templates (translesion synthesis; TLS). The 18 residue UBM sequence is found in tandem copies in Polι as well as in the Rev1 Y-polymerase. Isolated UBMs interact with ubiquitin with a Kd of ∼180 μM, and NMR studies reveal that the UBM interacts with a surface adjacent to the I44 patch (Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar). The UBZ domain is zinc-finger-like motif, distinct from other UBDs (Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar), which is conserved in the two remaining Y family polymerases (Polη and Polκ). What is the role of the UBDs, and therefore ubiquitin, in TLS? In response to UV irradiation, TLS polymerases interact with proliferating cell nuclear antigen (PCNA) in replication foci (Vidal et al., 2004Vidal A.E. Kannouche P. Podust V.N. Yang W. Lehmann A.R. Woodgate R. J. Biol. Chem. 2004; 279: 48360-48368Crossref PubMed Scopus (73) Google Scholar). Interestingly, Polι- and Polη-containing mutations in UBM or UBZ motifs are defective in recruitment to these UV-induced replication foci, suggesting a role for binding to ubiquitin in the recruitment process. TLS polymerases associate directly with PCNA, and this interaction becomes stronger upon PCNA ubiquitination (Vidal et al., 2004Vidal A.E. Kannouche P. Podust V.N. Yang W. Lehmann A.R. Woodgate R. J. Biol. Chem. 2004; 279: 48360-48368Crossref PubMed Scopus (73) Google Scholar), which occurs in response to DNA damage. Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar found that UBM and UBZ domains are critical for tight PCNA-TLS polymerase binding. Interestingly, the UBM-PCNAUb interaction was position independent because fusion of ubiquitin in frame with PCNA led to assembly with the UBM domain. Using an in vivo complementation system, Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar demonstrated that the UBZ domain of Polη is required for the response of cells to UV irradiation. How might the UBZ/UBM-PCNA interaction be regulated? Insight into this question comes from the finding that both Polι and Polη are monoubiquitinated in vivo, dependent upon an intact UBM or UBZ domain. Importantly, ubiquitinated TLS polymerases fail to associate with ubiquitinated PCNA, raising the possibility that monoubiquitination of TLS polymerases results in intramolecular inhibition of their own UBD (Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar). The notion that monoubiquitination of UBD-containing proteins inhibits them from binding in trans to ubiquitinated partner proteins was demonstrated recently by Hoeller et al., 2006Hoeller D. Crosetto N. Blagoev B. Raiborg C. Tikkanen R. Wagner S. Kowanetz K. Breitling R. Mann M. Stenmark H. Dikic I. Nat. Cell Biol. 2006; 8 (Published online January 22, 2006): 163-169https://doi.org/10.1038/ncb1354Crossref PubMed Scopus (251) Google Scholar. The authors fused ubiquitin to the C terminus of the endocytic adaptor proteins Sts1, Sts2, Eps15, and Hrs, all of which contain UBDs that interact with monoubiquitinated partners via the I44 patch on ubiquitin. The ubiquitin-fusion proteins fail to associate with monoubiquitin, but this interaction was restored by an I44A mutation in the ubiquitin fusion. Hoeller et al., 2006Hoeller D. Crosetto N. Blagoev B. Raiborg C. Tikkanen R. Wagner S. Kowanetz K. Breitling R. Mann M. Stenmark H. Dikic I. Nat. Cell Biol. 2006; 8 (Published online January 22, 2006): 163-169https://doi.org/10.1038/ncb1354Crossref PubMed Scopus (251) Google Scholar further demonstrated a conformational change resulting from intramolecular UBD-ubiquitin interactions using a FRET assay. As noted by Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar, conjugation and removal of ubiquitin from TLS polymerases may provide a mechanism for shuttling these enzymes in and out of replication foci during and after DNA damage. Further studies are required to determine how polymerase ubiquitination is regulated. The new work reviewed here represents an expansion in the number of known UBDs and provides high-resolution insights into the structural details for ubiquitin's interactions with four new classes of UBDs. These UBDs represent the building blocks that are often found assembled in different combinations within ubiquitin recognition and processing machines. Is any UBD-ubiquitin interaction sufficient for function, or are specific UBD-interactions important within a particular ubiquitin recognition machinery? We now know of one case in which UBDs appear to have been swapped during evolution: although yeast VPS9 contains a CUE domain at its C terminus, the presumed mammalian VPS9 ortholog, Rabex-5, contains N-terminal ZnF_A20 and MIU/IUIM domains. On the flip side, in at least some cases, linear fusions to ubiquitin proteins appear to maintain at least some function of their counterparts monoubiquitinated on a specific lysine (Bienko et al., 2005Bienko M. Green C.M. Crosetto N. Rudolf F. Zapart G. Coull B. Kannouche P. Wider G. Peter M. Lehmann A.R. et al.Science. 2005; 310: 1821-1824Crossref PubMed Scopus (552) Google Scholar, Hoeller et al., 2006Hoeller D. Crosetto N. Blagoev B. Raiborg C. Tikkanen R. Wagner S. Kowanetz K. Breitling R. Mann M. Stenmark H. Dikic I. Nat. Cell Biol. 2006; 8 (Published online January 22, 2006): 163-169https://doi.org/10.1038/ncb1354Crossref PubMed Scopus (251) Google Scholar). Still lingering is also the question of how other ubiquitin-like proteins are recognized. Recently, peptide sequences have been identified that bind SUMO in a manner distinct from known UBD-ubiquitin interactions (Song et al., 2004Song J. Durrin L.K. Wilkinson T.A. Krontiris T.G. Chen Y. Proc. Natl. Acad. Sci. USA. 2004; 101: 14373-14378Crossref PubMed Scopus (439) Google Scholar, Song et al., 2005Song J. Zhang Z. Hu W. Chen Y. J. Biol. Chem. 2005; 280: 40122-40129Crossref PubMed Scopus (206) Google Scholar). Future studies will likely reveal whether other Ubl proteins are recognized by new or related classes of binding motifs. The increased identification of UBDs is matched by an expansion in our knowledge of the ways in which ubiquitin modifies its targets. A remaining challenge will be to understand how the different combinations of UBDs are assembled into a particular architecture to read out the different ubiquitin modifications and translate them into a particular cellular outcome.
Referência(s)