Introduction

Ubiquitin and its relatives are small proteins that are covalently attached to other proteins and macromolecules via their C-termini. Post-translational modification by ubiquitin and ubiquitin-like proteins (ubls) has emerged as a predominant cellular regulatory mechanism (reviewed in Hochstrasser, 2000b; Pickart, 2001), with important roles in cell division, the immune response, development, endocytic trafficking, biosynthetic pathways, cancer and many other biological processes (for examples and reviews, see Scheffner et al., 1990; Glotzer et al., 1991; Rock et al., 1994; Pagano et al., 1995; Aberle et al., 1997; Honda et al., 1997; Kishino et al., 1997; Matsuura et al., 1997; Salceda and Caro, 1997; Ghosh et al., 1998; Hicke, 1999, 2001a; Joazeiro et al., 1999; Kamura et al., 1999; Koepp et al., 1999; Lorick et al., 1999; Sidow et al., 1999; York et al., 1999; Yewdell and Bennink, 2001; Kwon et al., 2002). Modification by ubls rapidly and reversibly changes the function of the target. The different ubl modifications can affect different properties of the target (Hoege et al., 2002), including the protein's half-life, enzymatic activity, subcellular localization and protein–protein interactions (reviewed in Hochstrasser, 1996, 2000a, 2000b; Jentsch and Pyrowolakis, 2000; Yeh et al., 2000; Pickart, 2001). The best understood consequence of these modifications is ubiquitin-mediated proteolysis, in which polymeric chains containing four or more ubiquitins, isopeptide-linked between Lys48 on the surface of one ubiquitin and the C-terminus of the next in the chain, direct proteins for degradation by the proteasome (reviewed in Rechsteiner, 1998). Many other types of ubiquitin modification also play fundamental cellular regulatory roles. For example, multiubiquitin chains with linkages via Lys63, rather than Lys48, activate the IB kinase (Deng et al., 2000), and monoubiquitylation plays a role in processes ranging from endocytosis to transcriptional activation (reviewed in Hicke, 2001b). Other ubls that structurally resemble ubiquitin and are conjugated to macromolecules in vivo are being discovered at a rapid rate (reviewed in Hochstrasser, 1998, 2000a, 2000b; Jentsch and Pyrowolakis, 2000; Melchior, 2000; Yeh et al., 2000; Hay, 2001; Muller et al., 2001; Ohsumi, 2001; Schwartz and Hochstrasser, 2003). For example, the ubl NEDD8 (Rub1p in Saccharomyces cerevisiae) activates SCF ubiquitin ligases and is involved in cell cycle control, signaling and embryogenesis (Lammer et al., 1998; Liakopoulos et al., 1998; Pozo et al., 1998; Jones and Candido, 2000; Tateishi et al., 2001; Kurz et al., 2002). ISG15 is involved in the antiviral interferon response (Haas et al., 1987; Loeb and Haas, 1992; Narasimhan et al., 1996; Nicholl et al., 2000; Yuan and Krug, 2001; MacQuillan et al., 2003). Apg8p (also called Aut7p and Cvt5p) modifies the lipid phosphatidylethanolamine to modulate membrane dynamics (Mizushima et al., 1998a; Ichimura et al., 2000). Hub1p modifies cell polarity factors and plays a role in cell polarization (Dittmar et al., 2002). SUMO family members, including Smt3p in S. cerevisiae, modify a number of different proteins involved in cell division, nuclear transport, the stress response and signal transduction (reviewed in Hay, 2001; Johnson and Gupta, 2001; Muller et al., 2001; Takahashi et al., 2001; Tatham et al., 2001; Pichler et al., 2002; Schmidt and Muller, 2002). There are many other ubls, including FAT10, Urm1p and Apg12p whose important biological functions are just beginning to be discovered (Liu et al., 1999; Furukawa et al., 2000; Raasi et al., 2001).

The best understood conjugation pathway is that for ubiquitin. Following proteolytic cleavage at the C-terminus to end with the sequence Gly–Gly (reviewed in Hochstrasser, 1996; Wilkinson, 1997; Wilkinson and Hochstrasser, 1998), ubiquitin is conjugated to its targets via an enzymatic cascasde that involves three activities, carried out by proteins or protein complexes in classes known as E1, E2 and E3 (Figure 1) (Hershko et al., 1983; also reviewed recently in Hershko and Ciechanover, 1998; Hershko et al., 2000; Pickart, 2001; Weissman, 2001). For ubiquitin there is one E1, tens of E2s, and hundreds of E3s. The E1, or ubiquitin-activating enzyme, activates ubiquitin by C-terminal adenylation, and subsequently forms a highly reactive thioester bond between its catalytic cysteine and ubiquitin's C-terminus. The E1 also associates with the next enzyme in the cascade, the E2 ubiquitin conjugating enzyme, and promotes ubiquitin transfer to the E2's catalytic cysteine (Pickart and Rose, 1985; Haas et al., 1988; also reviewed in Hochstrasser, 1998, 2000b; Pickart, 2001). Genetic studies in S. cerevisiae indicate that E2s have evolved to orchestrate modification of proteins in a common pathway. For example, the ubiquitin E2s Rad6p and Cdc34p are involved in the DNA damage response and the G1–S transition of the cell cycle, respectively (Jentsch et al., 1987; Goebl et al., 1988). This functional specificity of E2s probably results from the ability of a particular E2 to associate with the appropriate E3s, which are also called ubiquitin protein ligases (Hershko et al., 1983). The E3 binds selectively to both the E2 and to the target. Substrates for modification are selected by a vast repertoire of E3s. Database searches suggest that there are over 100 E3s that fall into two classes, RING (really interesting new gene) and HECT (homologous to E6AP C-terminus), with RING being the major class (Scheffner et al., 1993; Huibregtse et al., 1995; Joazeiro et al., 1999; Venter et al., 2001). E3s have at least two domains. One is a protein–protein interaction domain, such as SH2, WW, WD40, PTB or other protein binding domain that recruits substrate (reviewed in Freemont, 2000; Jackson et al., 2000; Joazeiro and Weissman, 2000; Pawson and Nash, 2000; Tyers and Jorgensen, 2000; Pickart, 2001; Weissman, 2001; VanDemark and Hill, 2002; Yaffe, 2002; Pawson and Nash, 2003). The other domain, RING or HECT, recruits the E2 (Kumar et al., 1997; Huang et al., 1999; Joazeiro and Weissman, 2000; Zheng et al., 2000). RING domains contain two zinc ions coordinated by cysteines and histidines in a cross-brace structure (Bellon et al., 1997). Recently, variations of the RING motif have also been found in E3s. One of these is the U-box, in which the cysteines and histidines found in the RING motif are replaced with charged and polar residues that form salt-bridges and hydrogen bonds to stabilize a similar structure, but without coordinating zinc (Aravind and Koonin, 2000; Hatakeyama et al., 2001; Meacham et al., 2001; Pringa et al., 2001; Ohi et al., 2003). Like RING E3s and their variants, HECT E3s also have two domains, with a protein–protein interaction domain recruiting the substrate and the HECT domain recruiting the E2. However, there are significant mechanistic differences between RING and HECT E3s. No catalytic residues have been found in RING E3s. Instead, it is thought that the E3 functions as a scaffold to bring together the target with the catalytic E2. Unlike the RING E3s, the HECT E3s are catalytic. The E2 transfers ubiquitin to the HECT active site cysteine, which itself transfers ubiquitin to the substrate (Huibregtse et al., 1995).

Figure 1.

Ubiquitin (Ub) conjugation cascade. E1 – ubiquitin-activating enzyme. E2 – ubiquitin-conjugating enzyme. E3 – ubiquitin protein ligase. Distinct, parallel conjugation cascades exist for other ubls

Full figure and legend (19K)

Other ubls are conjugated to their targets by enzymes that are relatives of the E1, E2 and E3 proteins found in the ubiquitin pathway, with the greatest sequence homology among the E1s and the E2s for different ubls. The most well characterized ubl conjugation cascades are for SUMO and NEDD8. One E1, one E2 and several E3s have been found for the SUMO pathway (Johnson and Blobel, 1997; Johnson et al., 1997; Schwarz et al., 1998; Desterro et al., 1999; Gong et al., 1999; Johnson and Gupta, 2001; Kahyo et al., 2001; Sachdev et al., 2001; Takahashi et al., 2001; Pichler et al., 2002; Schmidt and Muller, 2002; Kagey et al., 2003). The PIAS-family of SUMO E3s contain a motif known as the SPRING finger, whose sequence resembles the RING finger found in ubiquitin E3s. The NEDD8 pathway also involves one dedicated E1 and one E2 (Lammer et al., 1998; Liakopoulos et al., 1998; Osaka et al., 1998; Pozo et al., 1998; Gong and Yeh, 1999). Interestingly, the target for the NEDD8 pathway is part of a ubiquitin RING-class E3, which also appears to serve also as an E3 for the NEDD8 pathway, except in this case promoting automodification, rather than transfer to a different target (Lammer et al., 1998; Osaka et al., 1998; del Pozo and Estelle, 1999; Freed et al., 1999; Liakopoulos et al., 1999; Wada et al., 1999a, 1999b; Read et al., 2000).

The conjugation machinery for the autophagy pathway is the only known example in which one E1 activates two different ubls (Mizushima et al., 1998a, 1998b; Shintani et al., 1999; Tanida et al., 1999; Ichimura et al., 2000; Tanida et al., 2001). The ubls Apg12p and Apg8p have the same activating enzyme, Apg7p (also called Cvt2p). Both Apg12p and Apg8p are transferred from Apg7p to the catalytic cysteine in downstream conjugation components, Apg10p and Apg3p (also called Aut1p), respectively, which are thought to be functional homologs of E2s, despite an absence of sequence homology (reviewed in Ohsumi, 2001). The only other ubls for which any component of the conjugation machinery has been identified are Urm1p and ISG15, whose activating enzymes have been identified (Furukawa et al., 2000; Yuan and Krug, 2001). Interestingly, Ube1L, the E1 for ISG15 is a candidate tumor suppressor for small-cell lung carcinoma (Kok et al., 1993; Pitterle et al., 1998; McLaughlin et al., 2000). While other ubls are currently orphans, it is likely that their conjugation machineries will be identified in the near future, either based on sequence similarities, or via biochemical purifications.

There have been many excellent recent reviews on ubls, their conjugation machineries and their functions. This review serves to focus on recent structural and biochemical studies of the early steps of ubl conjugation cascades carried out by E1s and their relatives.

Biochemical mechanism of E1s, and their role in selecting a ubl and coordinating it with the cognate conjugation pathway

Ubls have their own dedicated E1s (Hatfield et al., 1990; Zacksenhaus and Sheinin, 1990; McGrath et al., 1991; Johnson et al., 1997; Lammer et al., 1998; Liakopoulos et al., 1998; Mizushima et al., 1998a; Osaka et al., 1998; Pozo et al., 1998; Desterro et al., 1999; Gong and Yeh, 1999; Gong et al., 1999; Okuma et al., 1999; Yuan and Krug, 2001), which are essential for all further conjugation (Ciechanover et al., 1984; Finley et al., 1984; McGrath et al., 1991; Swanson and Hochstrasser, 2000). E1s play at least three critical functions in initiating ubl conjugation cascades. First, the E1 selects the correct ubl for the pathway, associating noncovalently with the ubl. Second, the E1 activates the ubl C-terminus, allowing for further chemistry. Third, the E1 coordinates the ubl with the correct pathway, by transferring the ubl to its cognate E2. The enzymology of the E1 for ubiquitin has been studied extensively. Ubiquitin's E1 carries out these three functions through four enzymatic steps (Figure 2) (Ciechanover et al., 1981, 1982; Hershko et al., 1981; Haas and Rose, 1982; Haas et al., 1982, 1983, 1988; Hershko et al., 1983; Pickart and Rose, 1985; Haas and Bright, 1988; Wilkinson et al., 1990; Pickart et al., 1994) (also reviewed in Pickart, 2001).

Figure 2.

Reaction cascade for the E1 for ubiquitin. The E1 first adenylates ubiquitin's C-terminus and forms a noncovalent complex with ubiquitinadenylate. Second, the E1's catalytic cysteine attacks the ubiquitinadenylate, forming a covalent E1ubiquitin thioester complex. Next, the E1 adenylates a second molecule of ubiquitin, associating with the first covalently as a thioester and the second noncovalently as ubiquitinadenylate. Finally, E1 transfers the thioester-bound ubiquitin to E2's catalytic cysteine. The reactions of E1 serve at least three functions: (1) selection of the correct ubl for a pathway, (2) activation of the ubl's C-terminus for further chemistry and (3) coordination of the ubl with downstream components of the correct pathway

Full figure and legend (45K)

The first reaction catalysed by E1 is adenylation of ubiquitin's C-terminus (Ciechanover et al., 1981, 1982; Hershko et al., 1981). In addition to putting a good leaving group on the ubl's C-terminus, this step has also been suggested to be important for selection of the correct ubl. The adenylated ubiquitin forms a tight noncovalent complex with the E1 (Ciechanover et al., 1981; Haas and Rose, 1982; Haas et al., 1983). If the downstream steps of the conjugation cascade are blocked, either with a nonhydrolyzable analog of ubiquitinadenylate (Wilkinson et al., 1990) or by chemically blocking E1's catalytic cysteine with iodoacetamide (Haas et al., 1982), the adenylated ubiquitin remains stably associated with the E1. Studies of E1s for SUMO and NEDD8/Rub family members demonstrate that the first step of their reaction cycle also results in formation of an ubladenylate intermediate (Johnson et al., 1997; del Pozo and Estelle, 1999; Bohnsack and Haas, 2003).

The adenylation reaction has been most extensively studied for ubiquitin. For ubiquitin, the reaction proceeds by an absolutely ordered mechanism, with ATP binding preceding ubiquitin binding, prior to catalysis of ubiquitinadenylate formation (Haas and Rose, 1982; Haas et al., 1982; Haas et al., 1983; Hershko et al., 1983). Recent studies of the E1 for NEDD8 (Bohnsack and Haas, 2003), and of mutant versions of ubiquitin (Burch and Haas, 1994) demonstrate that it is also possible for E1s to function with random addition of ATP and the ubl. Thus, ubiquitin E1's requirement for ordered substrate addition is not a structural requirement for the catalytic competence of E1s, but reflects differences in the E1's affinities for ATP and ubiquitin as leading and trailing substrates (Bohnsack and Haas, 2003) (also, Z Tokgöz, R Bohnsack and A Haas, personal communication).

Insights into the interactions between E1 and ubiquitin have been provided from studies of mutant versions of ubiquitin. Ubiquitin's C-terminal glycine, as well as arginines 42 and 72 are involved in the interaction between ubiquitinadenylate and its E1 (Burch and Haas, 1994; Pickart et al., 1994). Residue 72 in ubiquitin plays a particularly critical role in the interaction with E1, and is the only known determinant of specificity in ubls for their E1s. Residue 72 is arginine in ubiquitin, and an alanine in ubiquitin's closest relative, NEDD8 (Larsen and Wang, 2002), which shares nearly 60% sequence identity with ubiquitin. Despite this high degree of sequence similarity, the E1 for ubiquitin is specific for ubiquitin and the E1 for NEDD8 is specific for NEDD8 (Osaka et al., 1998; Whitby et al., 1998). Much of this specificity can be attributed to residue 72. Mutation of NEDD8's alanine 72 to arginine is sufficient to allow NEDD8 to be activated by ubiquitin's E1 (Whitby et al., 1998). A mutation of ubiquitin's basic arginine 72 to the hydrophobic residue leucine, which is similar in nature to alanine, is sufficient to allow ubiquitin to be recognized by NEDD8's E1, resulting in a sixfold increase in catalytic specificity (Bohnsack and Haas, 2003). Kinetic studies of the E1 for NEDD8 with NEDD8 and the mutant version of ubiquitin with arginine 72 mutated to leucine suggest that the specificity of conjugation cascades for their particular ubl is related to affinity for a particular ubl at the adenylation step of the reaction (Bohnsack and Haas, 2003).

Following adenylation of ubiquitin's C-terminus, the E1's catalytic cysteine attacks the adenylate and forms a thioester with ubiquitin's C-terminus (Ciechanover et al., 1981, 1982; Haas and Rose, 1982; Haas et al., 1982). Formation of the E1ubiquitin thioester is very rapid, with a turnover rate of two per second (Haas and Rose, 1982). Next, the E1 adenylates a second molecule of ubiquitin, so the E1 is loaded with two molecules of ubiquitin – the first bound covalently as a thioester, and the second one bound noncovalently as an adenylate (Haas and Rose, 1982; Haas et al., 1982). A recent kinetic study demonstrated that the E1 for NEDD8 also proceeds through an analogous, double-loaded intermediate (Bohnsack and Haas, 2003), suggesting that the mechanism of all E1s is similar. Formation of the second molecule of ubladenylate prior to transfer of the ubl to E2 may serve several functions. First, it primes E1 to carry out another round of catalysis immediately upon transfer of a molecule of ubl to E2 (Haas and Rose, 1982). Second, studies of the E1 for ubiquitin show that binding of the second molecule of ubiquitin, in the form of ubiquitinadenylate promotes transfer of the thioester-linked molecule of ubiquitin to E2 (Pickart et al., 1994).

The final step catalysed by E1 is the transfer of ubiquitin to E2, the downstream component of the reaction cascade. This involves noncovalent association of the E1 with the E2, and a transthiolation reaction in which the ubl is transferred from E1's catalytic cysteine to E2's (Hershko et al., 1983; Haas et al., 1988).

It has been suggested that the E1 plays a major role in bringing the ubl together with the correct E2 (Hochstrasser, 2000b; Pickart, 2001). Support for this hypothesis comes from the findings that E1s associate noncovalently with ubiquitinadenylate (Haas et al., 1982), and that E1s also associate noncovalently with the E2 during the reaction (Haas et al., 1988). Consistent with this theory, while NEDD8 normally cannot be transferred to an E2 for ubiquitin, a NEDD8 mutant that is activated by the E1 for ubiquitin can be transferred to a ubiquitin E2 (Whitby et al., 1998). In addition, mutation of residues on either E1 or E2 that are involved in the noncovalent E1–E2 protein–protein interaction diminishes E2ubl thioester bond formation (Bencsath et al., 2002; Walden et al., 2003) (also Arthur Haas, personal communication).

Evolutionary origins of ubl adenylation from bacterial biosynthetic enzymes

Ubls and their activating enzymes appear to have evolved from primitive biosynthetic pathways, which exist in bacteria as well as eucaryotes. MoaD and ThiS are structural homologs of ubiquitin that play critical roles in the Escherichia coli molybdopterin and thiamin biosynthesis pathways, respectively (Lake et al., 2001; Rudolph et al., 2001; Wang et al., 2001). Like ubiquitin and other ubls, MoaD and ThiS are adenylated at their C-termini prior to downstream steps in the pathways (Taylor et al., 1998; Leimkuhler et al., 2001). Unlike ubiquitin and ubls such as NEDD8, SUMO, Apg8 and ISG15, however, the ultimate functions of MoaD and ThiS are not to be transferred to other proteins or macromolecules. Instead, MoaD and ThiS are temporarily modified, themselves, carrying sulfur atoms at their C-termini as thiocarboxylates, and are involved in sulfur transfer in the molybdopterin and thiamine biosynthetic pathways (Pitterle et al., 1993; Pitterle and Rajagopalan, 1993; Taylor et al., 1998). Thus, a common feature of the ubiquitin, ubl, MoaD and ThiS pathways is the involvement of the ubl's C-terminus in several chemical reactions. Interestingly, MoeB and ThiF, the enzymes that catalyse adenylation of the C-termini of the ubiquitin-like MoaD and ThiS proteins, respectively, share significant sequence homology to part of the E1s for ubiquitin and ubls (McGrath et al., 1991; Johnson et al., 1997). This sequence homology includes a G-X-G-X-X-G nucleotide binding motif (Walker et al., 1982), suggesting that the C-terminal adenylation reaction will be similar for ubiquitin, ubl, MoaD and ThiS pathways.

Modular nature of the primary sequences of ubl-activating enzymes

Analysis of primary sequences of ubl-activating enzymes, in conjunction with recently determined crystal structures described below (Lake et al., 2001; Walden et al., 2003), suggest that ubl-activating enzymes are modular proteins that have evolved multiple distinct domains to carry out their multiple specific functions (Figure 3). The common feature of all ubl-activating enzymes is a region of sequence homology to MoeB and ThiF, which suggests that the common function associated with E1s is C-terminal adenylation of a ubl. Additional sequences at the N- or C-termini, and in the middle of the MoeB/ThiF homology domain are likely to correspond to new domains that have evolved for specific functions of a particular ubl-activating enzyme.

Figure 3.

Comparison of the activating enzymes for Ub (ubiquitin) and ubls. The activating enzymes for E. coli MoaD and ThiS, MoeB and ThiF are 27 kDa, and the crystal structure of MoeB revealed a homodimer (Lake et al., 2001). The sequences corresponding to the N- and C-terminal subdomains in the MoeB structure are referred to as 'N' and 'C', respectively. The activating enzymes, or E1s, for ubiquitin, ISG15, NEDD8 and SUMO and their orthologs are 110–120 kDa proteins or heterodimeric complexes that contain a twofold repeat that corresponds to the sequence and structure of MoeB. The crystal structure of APPBP1-Uba3 revealed that extensions and insertions relative to the MoeB sequence play a role in E1-specific functions, such as transfer of the ubl to E2 (Walden et al., 2003). The sequences of other ubl-activating enzymes, such as Uba4p and Apg7p also contain extensions that may carry out functions specific to their particular conjugation cascades

Full figure and legend (139K)

The four ubls whose E1s are most closely related are ubiquitin, SUMO, NEDD8 and ISG15 and their family members. The E1s for these are 110–120 kDa, and are either a single polypeptide or a heterodimeric complex. In the heterodimeric E1s, one subunit corresponds to roughly the N-terminal half of the single-chain E1s, and the other corresponds to roughly the C-terminal half. The E1s for ubiquitin and ISG15 are single polypeptides, whereas the E1s for NEDD8 and SUMO family members are 110 kDa heterodimers, with one protein (APPBP1, NAE1, AXR1 or Ula1p for NEDD8 and its orthologs and Aos1p or SAE1 for SUMO) homologous to the N-terminal half of the E1s for ubiquitin and Isg15, and the other protein (Uba3, NAE2 or ECR1 for NEDD8 and its orthologs and Uba2 or SAE2 for SUMO) homologous to the C-terminal half of the E1 for ubiquitin (Figure 3) (Johnson et al., 1997; Lammer et al., 1998; Liakopoulos et al., 1998; Osaka et al., 1998; Pozo et al., 1998; Desterro et al., 1999; Gong et al., 1999; Gong and Yeh, 1999; Okuma et al., 1999; Yuan and Krug, 2001). The N- and C-terminal halves of the E1s for ubiquitin, SUMO, NEDD8 and ISG15 are partially homologous to each other, and the region of sequence homology between the two halves is also the region of sequence homology to MoeB and ThiF. The crystal structures of MoeB and of the APPBP1-Uba3 complex that is the E1 for NEDD8 revealed that this sequence homology region adopts a Rossmann fold, a structure frequently found in nucleotide binding proteins (Rossmann et al., 1974; Lake et al., 2001; Walden et al., 2003). Indeed, the G-X-G-X-X-G ATP binding motif (Walker et al., 1982) resides in the homology region in the C-terminal half of E1s. A subset of E1s also have retained this ATP-binding sequence in their N-terminal halves, although the function there is not known. In addition to playing a role in nucleotide binding, this sequence homology region is also involved in dimerization. It is involved in the homodimer interface in the perfectly symmetric MoeB–MoeB crystal structure (Lake et al., 2001), and in the protein–protein interface in the APPBP1–Uba3 complex (Walden et al., 2003).

The recently determined crystal structures of MoeB (Lake et al., 2001) and of the APPBP1–Uba3 complex (Walden et al., 2003) revealed that the structural homology between MoeB and both APPBP1 and Uba3 extends beyond the region of significant sequence homology, and also includes the C-terminal part of the MoeB sequence (Figure 3). This C-terminal portion of the MoeB sequence is involved in binding to the ubl MoaD, raising the possibility that even distally related ubl-activating enzymes will bind to their cognate ubls via a similar structure.

While in the MoeB sequence the two structural homology regions are separated by only an eight-residue linker, in both APPBP1 and Uba3 there are large insertions between the two MoeB structural homology regions. The insertions found in APPBP1 and in Uba3 are moderately conserved among the E1s for ubiquitin, SUMO, NEDD8 and ISG15. These E1s also contain a region of homology at the C-terminus. The sequences of these E1s suggest that the additional homology regions have evolved for E1-specific functions such as transfer of the ubl to the catalytic cysteine of one of a family of closely related E2s. Indeed, the insertion in the C-terminal half of E1 sequences contains the catalytic cysteine. This same region in MoeB is an eight-residue mobile loop, which also contains a cysteine.

The sequences of the activating enzymes for Urm1p, Apg12p and Apg8p/Aut7p are significantly divergent. Both of these more directly resemble the sequences of MoeB and ThiF, with only one copy of the structural homology region, and without an insertion between the two parts of the structural homology region. Based on structural constraints described below, it seems likely that the activating enzymes for Urm1p, Apg12p and Apg8p/Aut7p either function as homodimers, or else they have partner proteins resembling the N-terminal half of E1s that have yet to be reported. Indeed, Apg7p, the activating enzyme for Apg12p and Apg8p/Aut7p, has been reported to homodimerize (Komatsu et al., 2001). The sequences both of Apg7p and of Uba4p, the activating enzyme for Urm1p, contain extensions. Apg7p and its orthologs have a roughly 300-residue N-terminal extension that is unique to the Apg7p family members and is not homologous to other proteins. Uba4p has a roughly 120-residue C-terminal extension that is a rhodanese homology domain (Hofmann et al., 1998). Rhodaneses are widespread enzymes that catalyse sulfur transfer reactions, suggesting that Urm1p might function in a biosynthetic pathway in a manner similar to ThiS and MoaD. Consistent with this notion, some orthologs of MoeB (Appleyard et al., 1998), and another protein in the thiamin biosynthetic pathway, ThiI (Palenchar et al., 2000), also contain rhodanese homology domains. Intriguingly, an additional poorly understood link between rhodanese homology domains and the ubiquitin pathway comes from the prediction that some deubiquitinating enzymes contain rhodanese homology domains (Hofmann et al., 1998).

The different extensions and insertions in the MoeB/ThiF-like sequences of ubl-activating enzymes likely reflect the addition of different functionalities required for the different ubl-conjugation cascades.

Insights into ubl recognition and the adenylation reaction from the structure of a bacterial MoeB–MoaD complex

Three crystal structures of MoeB–MoaD complexes revealed the structural basis for ubl activation (Lake et al., 2001). Two of the structures are enzyme–substrate complexes: MoeB–MoaD alone and the MoeB–MoaD in complex with ATP. One structure is of the enzyme–product complex: MoeB–MoaDadenylate. Together, the three structures provide insight into ubl interactions, nucleotide interactions, and residues in an activating enzyme involved in the activation reaction.

The structure of MoeB has two subdomains, corresponding to the homology regions of the sequence (Figure 4). The N-terminal subdomain contains a variant Rossmann nucleotide binding fold, with a four-stranded parallel -sheet surrounded by helices. This subdomain contains the Gly-X-Gly-X-X-Gly nucleotide binding motif, and most of the catalytic residues. The C-terminal subdomain contains a four-stranded antiparallel -sheet and adopts a fold distantly related to sugar binding proteins. This second subdomain forms the bulk of the MoaD-interaction site. The sequence of this second subdomain is poorly conserved between MoeB and E1s for ubiquitin, SUMO, NEDD8 and ISG15, although the structure of the APPBP1–Uba3 complex showed that the structure of this region is conserved (Walden et al., 2003).

Figure 4.

Structure of the MoeB–MoaDadenylate complex (Lake et al., 2001). MoeB is shown in yellow, MoaD in orange and AMP in gray. The N-terminal subdomain of the MoeB structure adopts a variation of the Rossmann fold and is involved in ATP binding and catalysing the adenylation reaction. The C-terminal subdomain of the MoeB structure contains an antiparallel, four-stranded -sheet involved in binding to the globular domain of MoaD

Full figure and legend (181K)

Both the ATP binding site and the side-chains involved in adenylation are among the most highly conserved sequences between MoeB and E1s (Figure 5). A hydrophobic patch, involving MoeB's Val37, Phe63, Leu109 and Val134, interacts with the adenine ring. Notably, Val37, Phe63 and Val134 are conserved as hydrophobic residues in E1s. MoeB's Asn131 side-chain, which is conserved as a polar residue in E1s, forms a hydrogen bond with the N7 position of the adenine ring. The degree of conservation increases closer to the active site. MoeB's Asp62 and Gly40 contact the ribose ring and are identical among E1s. The backbone of Gly41, conserved as glycine or alanine in E1s, contacts ATP's -phosphate. The absolutely conserved Arg73 and Lys86 contact the -phosphate, and Ser69, conserved as a polar residue among E1s, and the absolutely conserved Asn70 interact with the -phosphate. The ATP in the MoeB–MoaD–ATP structure is in a strained conformation, with a kink at the -phosphate. Although a magnesium ion is not present in any of the structures, MoeB's Asp130 is in the correct position to ligate the magnesium ion.

Figure 5.

Close-up view of ATP bound near the C-terminus of MoaD (Lake et al., 2001). The ATP binding site consists of residues from both MoeB molecules in the crystallographic MoeB homodimer. One molecule of MoeB is shown in yellow, the other molecule of MoeB in the homodimer is shown in green, MoaD is shown in orange and ATP is shown in gray. Side-chains from many of the residues involved in ATP binding are indicated, and salt-bridges and hydrogen bonds to ATP are shown in black

Full figure and legend (247K)

MoeB interacts with a predominantly hydrophobic surface in MoaD, which is partially conserved in ubiquitin. At least one hydrogen bond, between MoaD's Arg11 and MoeB's Asp227 is also probably conserved between ubiquitin and its E1. However, MoaD–MoeB and ubiquitin–E1 interactions must be different, because MoaD lacks arginines corresponding to ubiquitin's arginines 42 and 72, known to be involved in interaction with E1. Small amino acids interact with the C-terminal Gly–-Gly motif in MoaD, conserved in ubiquitin and many other ubls. The C-terminus of MoaD is oriented for the adenylation reaction by interactions with the backbone of MoeB's Leu42 and by an interaction with MoeB's Arg135, which is absolutely conserved among E1s.

The MoeB–MoaD complexes form a perfectly symmetric heterotetramer, involving extensive hydrophobic interactions between the two MoeB molecules. The MoeB homodimerization interface serves as a hydrophobic core, suggesting that a predominant function of dimerization is likely to be stabilization of the MoeB structure. In addition, there is also crosstalk between the two MoeB monomers in the ATP binding site. Arg14 from one monomer plays a major role in the adenylation of MoaD bound to the opposite monomer. Specifically, prior to the adenylation reaction, Arg14 contacts two of the oxygens on ATP's -phosphate. The Arg14 side-chain undergoes a major conformational change, allowing it to stabilize the negative charge that develops on the -phosphate during the reaction. Interestingly, the structure of the APPBP1–Uba3 complex, described in detail below, reveals that APPBP1 contributes an analogous arginine to the adenylation active site composed primarily of residues from Uba3. The absolute conservation of this arginine in the sequences of ubl-activating enzymes is consistent with both a key role for this arginine in the reaction, and also with the notion that other ubl-activating enzymes will adopt a similar structure, involving two Rossmann-fold containing domains.

The proposed adenylation reaction is as follows. First, the magnesium ion alleviates electrostatic repulsion between the MoaD C-terminus and ATP's -phosphate. A carboxylate oxygen of the C-terminal glycine attacks the -phosphate, resulting in a transient, penta-coordinated intermediate, prior to formation of the MoaDadenylate. The developing negative charge on the -phosphate is stabilized by Arg73, and also by Arg14 from the opposite MoeB monomer. Finally, the MoaDadenylate is formed.

Overall structure of E1s from the structure of APPBP1–Uba3, NEDD8's E1

Insights into the structural organization of E1s has come from the recent crystal structure of the E1 for NEDD8, the heterodimeric APPBP1–Uba3 complex (Figure 6). The structure reveals three domains, each specifying one of E1's three activities. One domain, corresponding to the common sequences among ubl-activating enzymes, resembles the structure of the MoeB homodimer and is involved in adenylation. A second domain arises from the E1-specific insertions in the middle of the two-part homology regions in both the N- and C-terminal halves of E1s. This domain is organized around the catalytic cysteine and is involved in thioester bond formation and the transthiolation reaction with E2. The third domain corresponds to the C-terminus of E1s. This domain adopts a -grasp fold, resembling the structure of ubiquitin and ubls, and is involved in E2 recognition.

Figure 6.

Overall structure of the APPBP1–Uba3 complex (Walden et al., 2003). Three views of the structure, separated by 60° rotations in y, with surface representations in the top panels, and secondary structural elements shown below. APPBP1 is colored cyan and Uba3 is red. The catalytic cysteine, Uba3's Cys 216 is shown in yellow. Cleft 1, Cleft 2, the nucleotide binding pocket, the second MoeB-like subdomain, APPBP1's four helix-bundle (4HB), Uba3's C-terminal domain and crossover loop are indicated

Full figure and legend (328K)

The three domains pack together to generate binding sites for ATP, the ubl NEDD8 and the E2 Ubc12. The overall structure resembles a canyon, with a broad deep groove in the middle. The adenylation domain resembling the MoeB homodimer serves as the base of the canyon. The E1-specific catalytic cysteine-containing and C-terminal domains serve as the canyon walls. An eight-residue 'crossover loop' (VanDemark and Hill, 2003) leads from the adenylation domain to the catalytic cysteine, and divides the canyon into two clefts. The two clefts are continuous with each other below and above the crossover loop. Each cleft is large enough to accommodate either two molecules of NEDD8 or a domain of E2.

The left cleft in Figure 6, Cleft 1, has the nucleotide binding pocket as its base. Its walls are formed by part of the Uba3 portion of the catalytic cysteine domain, and the C-terminal ubiquitin-like domain. The base of Cleft 2, on the right in Figure 6, is a four-stranded -sheet from Uba3 corresponding to the second MoeB-like subdomain. In the MoeB structure, this second subdomain is involved in MoaD binding. The catalytic cysteine domain portion of APPBP1 is the predominant wall for Cleft 2. Uba3's catalytic cysteine faces this second cleft.

Model for adenylation by E1

The adenylation domain of the E1 for NEDD8 contains the repeat sequence found in the N- and C-terminal halves of E1, which is also the region of sequence homology found in MoeB. In this domain, APPBP1 and Uba3 are structurally homologous to each other and to MoeB (Figure 7a). This structure consists of a mixed eight-stranded -sheet surrounded by eight -helices, and includes regions where sequence homology is weak. Both APPBP1 and Uba3 contain the two subdomains found in MoeB, with the N-terminal half of each adopting a variation of Rossmann nucleotide binding topology. In this domain, APPBP1 and Uba3 pack together in the same way that the two MoeB molecules pack in the MoeB homodimer (Figure 7b).

Figure 7.

Structural homology between APPBP1, Uba3 and MoeB (Lake et al., 2001; Walden et al., 2003). APPBP1 is shown in cyan, Uba3 in red, one MoeB molecule in the crystallographic MoeB homodimer is shown in yellow, and the other MoeB molecule in the crystallographic MoeB homodimer is shown in green. (a) In the adenylation domain of the E1 structure, APPBP1 and Uba3 are structurally homologous to each other. (b) APPBP1 and Uba3 are both structurally homologous to MoeB, and the APPBP1–Uba3 heterodimerization interface resembles the MoeB–MoeB homodimerization interface

Full figure and legend (415K)

The MoeB homodimer contains a perfect twofold symmetry, so each molecule of MoeB binds to a molecule of MoaD, forming a heterotetrameric complex. Thus, two molecules of MoaD can be adenylated simultaneously by the MoeB homodimeric complex. By contrast, the E1s appear to have evolved directionality in the reaction by restricting the activity to a single adenylation active site in Cleft 1, composed primarily of the C-terminal half of the E1, the Uba3 component in NEDD8's E1. Uba3 contains the glycine- rich nucleotide binding motif corresponding to the ATP inding site in MoeB (Figure 8a). By contrast, APPBP1 lacks G-X-G-X-X-G ATP binding residues in its MoeB-homology region. Also, the second subdomain of Uba3's adenylation domain is exposed in Cleft 2 to interact with NEDD8, much like MoeB's is exposed to interact with MoaD. By contrast, the second subdomain of APPBP1 is buried by a four-helix bundle, whose sequence is conserved in the N-terminal portions of other E1s, which would preclude interaction with a second molecule of NEDD8 (Figure 8b).

Figure 8.

Structural evidence for a single adenylation active site in the APPBP1–Uba3 complex. Uba3 in red, one MoeB molecule in the crystallographic MoeB homodimer is shown in yellow with its associated MoaD in orange, and the other MoeB molecule in the crystallographic MoeB homodimer is shown in green with its associated MoaD in purple. Structural alignment of Uba3 (a) or APPBP1 (b) with the MoeB–MoaD complex. The nucleotide binding site in Uba3 aligns with the nucleotide binding site in MoeB, and there is room in Cleft 2 to accommodate a ubl molecule. By contract, APPBP1 lacks a nucleotide binding motif, and a conserved four helix bundle in APPBP1 precludes ubl binding

Full figure and legend (213K)

Although there is limited sequence homology between MoeB and E1s in Cleft 2, the ubl-binding region, there is sufficient structural homology to place NEDD8 roughly in Cleft 2. The structure of the APPBP1–Uba3 adenylation domain was aligned with the MoeB–MoaDadenylate complex, allowing modeling of the structures of NEDD8 and AMP in the complex (Figure 9). Many features of the model have been confirmed by mutational analyses (Walden et al., 2003) (also A Haas, personal communication). The model suggests that both clefts in the E1 canyon are involved in formation of the ubladenylate, with residues involved in catalysing the adenylation reaction in Cleft 1, and extensive hydrophobic contacts with ubl made in Cleft 2. Residues important for the activation reaction as shown by the MoeB–MoaDadenylate structure are conserved in the structure of NEDD8's E1. These include a hydrophobic patch involving Uba3 Ile127, Gly55 and Ile54 that would contact the adenine, Uba3's Asp79 and Asp81 contacting the ribose ring, and Uba3's Asp146 coordinating magnesium. Just as one monomer of MoeB contributes a key arginine side-chain to coordinate the ATP bound to the opposite monomer, APPBP1's Arg15 is in the same position relative to the nucleotide binding site in Uba3.

Figure 9.

Model of an APPBP1–Uba3–NEDD8adenylate complex (Walden et al., 2003). The structure of NEDD8 (Whitby et al., 1998) was modeled into the site occupied by MoaD (Figure 8). APPBP1 is shown in cyan, Uba3 in red and NEDD8 and AMP in gray. Possible interacting residues from APPBP1, Uba3 and NEDD8 are labeled

Full figure and legend (437K)

The bulk of the NEDD8 binding site is predicted to involve Cleft 2. The model predicts extensive hydrophobic contacts between a hydrophobic patch on the surface of Uba3, involving Tyr 321, Tyr 331, Tyr 333 and Phe 335, which are conserved as tyrosine, proline, phenylalanine and phenylalanine in the E1 for ubiquitin, and a hydrophobic patch that is absolutely conserved between NEDD8 and ubiquitin including Leu 8, Ile 44 and Val 70. These residues in ubiquitin are essential for viability in yeast (Sloper-Mould et al., 2001) and mutation of Leu 8, or Leu 8 in combination with Val 70, reduces ubiquitin conjugate formation by more than 50% (Beal et al., 1996).

E1s display exquisite selectivity for their particular ubl. The model of the APPBP1–Uba3–NEDD8adenylate complex suggests an explanation for part of the preference of E1s for their cognate ubls (Walden et al., 2003). Residue 72 is the only known determinant of specificity in ubls for their E1s. Residue 72 is an arginine in ubiquitin and an alanine in NEDD8. In SUMO family members, the corresponding residue is either a glutamate or a glutamine. Evidence that residue 72 is a specificity determinant comes from the findings that mutation of ubiquitin Arg 72 to leucine reduces binding of the ubiquitinadenylate by 1000-fold (Burch and Haas, 1994), that an Ala 72 to arginine mutation in Nedd8 allows it to be activated at a similar rate as ubiquitin by the E1 for ubiquitin (Whitby et al., 1998), and that an Arg 72 to leucine mutation in ubiquitin allows it to be activated by the E1 for NEDD8 (Bohnsack and Haas, 2003). In the model, Uba3's hydrophobic residues Leu206 and Tyr207 are in a position to interact with Ala72 of Nedd8. There is a correlation between the nature of the amino-acid side-chains corresponding to Leu206 and Tyr207 in other E1s and the nature of the side-chain at position 72 in other ubls. In the E1 for ubiquitin, Leu206 and Tyr207 are replaced by an aspartate, and they are replaced by a lysine in the E1 for SUMO family members. There must be additional determinants of specificity, however, because mutation of Uba3's Leu206 and Tyr207 to aspartate is not sufficient to promote adenylation of ubiquitin (Walden et al., 2003).

Structural insights into the E1ubl thioester intermediate

A second domain comprises E1-specific insertions in the sequences of the adenylation domain portions of both APPBP1 and Uba3 (Figure 6). The catalytic cysteine, Cys216 of Uba3, which forms a thioester with NEDD8 and promotes transfer of NEDD8 to its E2 (Liakopoulos et al., 1998; Osaka et al., 1998), is the focus of this second domain. The elongated catalytic cysteine domain contains 11 helices from APPBP1 and four from Uba3 that radiate either toward or away from the active site Cys216. Although the insertions in the N-terminal portions of some other E1s are shorter than the insertion in APPBP1, and the insertions in the C-terminal portions of some other E1s are correspondingly longer than the insertion in Uba3, they are expected to be entirely helical by secondary structure prediction algorithms (Rost and Sander, 1994). Thus, the overall structure of the catalytic cysteine-containing domain will likely be similar in other E1s, but with variations in the relative order of the helices in the sequence.

Formation of the thioester complex between E1s and ubl likely proceeds through a nucleophilic attack on the ubladenylate by the E1 active site cysteine. This would require the active site cysteine be in close proximity with the ubl C-terminus. In the crystal structure of NEDD8's E1, the thiol of Cys216 is 30 Å away from the adenylation active site, suggesting that a conformational change in the complex would be required for the juxtaposition of the ubl C-terminus and the active site cysteine thiol. The gap might be reduced to only 10 Å, because the C-terminal four residues of ubls are flexible (Vijay-Kumar et al., 1985; Vijay-Kumar et al., 1987; Bayer et al., 1998; Rao-Naik et al., 1998; Whitby et al., 1998; Paz et al., 2000; Sheng and Liao, 2002; Ramelot et al., 2003), and could, in principle, extend directly toward the active site Cys while remaining noncovalently associated via hydrophobic interactions. The remainder of the gap could be closed by a change in the relative orientations of the adenylation and catalytic cysteine-containing domains. Indeed, the adenylation and catalytic cysteine-containing domains are linked by four extended tethers, APPBP1's -strands 7 and 8, and Uba3's loops 7 and 11, which could serve as hinges that allow the domains to rotate with respect to one another. Alternatively, a local conformational change might occur around the catalytic cysteine.

The NEDD8 E1 structure reveals that the Cys216 side-chain is exposed in the active site except for the hydroxyl of Thr217. Mutation of Uba3's Thr217 to alanine diminishes Uba3–NEDD8 thioester formation and the subsequent transfer of NEDD8 to its cognate E2, Ubc12, without affecting NEDD8adenylate formation (Walden et al., 2003). Thr217 is strictly conserved in the E1s for ubiquitin, SUMO, NEDD8, ISG15, Apg8p, Apg12p and their family members, all of which are known to form activating enzymeubl thioester intermediates. However, the residues corresponding to Uba3's Thr217 vary in MoeB, ThiF and Uba4p, and this deviation may reflect mechanistic differences at their catalytic cysteines. The roles of the cysteine residues corresponding to the catalytic cysteines of E1s are less clear for MoeB, ThiF and Uba4p. Although MoeB, ThiF and Uba4p all contain cysteines corresponding to the catalytic cysteines in the sequences of E1s, their close relative, HesA, which is involved in iron/sulfur protein formation during Anabena heterocyst biogenesis, does not (Borthakur et al., 1990). The region of MoeB that corresponds to the E1 catalytic cysteine domain insertions is an eight-residue mobile loop that is not visible in any of the MoeB–MoaD crystal complex structures. The role of MoeB's Cys187, corresponding to the catalytic cysteine in E1s, is unclear, because mutation of C187 to alanine does not significantly impair MoeB's ability to drive molybdopterin synthesis in vitro (Leimkuhler et al., 2001). By contrast, ThiF's Cys 184, corresponding to the catalytic cysteine of E1s, does play a role in the reaction cascade with ThiF: ThiF forms an acyl-disulfide complex with the C-terminus of ThiS (Xi et al., 2001). Uba4p's Cys 225 is also involved in forming a covalent complex with Urm1p, although the downstream reactions catalysed by Uba4p and the targets of Urm1p remain to be elucidated (Furukawa et al., 2000). Interestingly, the amino acid immediately following the potential catalytic cysteine is a valine in MoeB, an arginine in ThiF and a glutamate in Uba4p. The differences in residues corresponding to Uba3's Thr217 may result in differences in the reactions catalysed by the different activating enzymes.

Structural and mechanistic insights into ubl transfer from E1 to E2

The final function of E1 is to coordinate the ubl with its correct E2. This involves E1 interacting noncovalently with E2, and a subsequent transthiolation reaction in which the ubl is transferred from E1's catalytic cysteine to the E2's. For ubiquitin, a single E1 sits at the top of the cascade, transferring ubiquitin to many different E2s, one at a time. The fully loaded E1 has a strong affinity for E2s, with Kds in the subnanomolar to nanomolar range depending on the E2 (Haas et al., 1988). However, ubiquitin's E1 appears to have a low affinity for E2s in the absence of ubiquitin, as E1 readily separates from the E2s during purification (Hershko et al., 1983). The low affinity of free E1 for E2s might facilitate rapid cycling of E1, so that it can charge ubiquitin's many different E2s (Hershko et al., 1983). The situation may differ for the ubls SUMO and NEDD8 and their family members, because there is only one known E2 for each of these cascades. The Km for NEDD8's E2, Ubc12, during the reaction is similar to those for ubiquitin E2s during their reactions (Bohnsack and Haas, 2003). However, both Ubc12 and the E2 for SUMO family members, Ubc9, have been shown to associate with their E1s in the free state (Bencsath et al., 2002; Walden et al., 2003).

Although at present there are no crystal structures for E1–E2 protein complexes, the region of an E2 that is involved in the noncovalent interaction with an E1 has been mapped by mutational analysis. A mutational analysis of Ubc9p from Saccharomyces cerevisiae, the E2 for the SUMO family member Smt3p, mapped the E1 binding site to involve the N-terminal helix and the loop between the first and second -strands (Bencsath et al., 2002). The location of the E1 binding site is probably conserved among E2s for ubiquitin and other ubls, because two studies of E2s for ubiquitin revealed that mutations in the N-terminal helix diminished E2ubiquitin thioester bond formation (Sullivan and Vierstra, 1991; Pitluk et al., 1995). In addition, for another ubiquitin E2, Ubc13, a protein–protein interaction that blocks access to this surface impairs E2ubiquitin thioester bond formation (McKenna et al., 2001; Moraes et al., 2001; VanDemark et al., 2001). Interestingly, the structure of the loop between the first and second -strands is one of the most prominent differences between the structure of UBC9, the E2 for SUMO and the structures of E2s for ubiquitin (Tong et al., 1997; Giraud et al., 1998; Bernier-Villamor et al., 2002). These structural differences could serve as part of the basis for specificity in E1 binding.

It is appealing to hypothesize that the E2 would bind to E1's Cleft 1 (Figure 6) for several reasons. First, in the structural model of the APPBP1–Uba3–NEDD8adenylate complex, NEDD8 occupies Cleft 2, so there is space remaining in Cleft 1 to accommodate the E2 protein. Second, deletion of Uba3's C-terminal domain, which borders Cleft 1, is defective at associating with the E2 Ubc12, and at transferring NEDD8 to Ubc12 (Walden et al., 2003). Finally, the nucleotide binding site resides in Cleft 1, and for the E1 for ubiquitin, occupation of the nucleotide binding site has been shown to promote the transfer of E1's thioester-bound ubiquitin to E2 (Pickart et al., 1994).

Following association of E1 with E2, the active site cysteine in E1 is involved in the transfer of the ubl to the E2 catalytic cysteine. This transthiolation reaction likely involves deprotonation of E2's active site cysteine by a general base catalyst. Histidine often serves as a general base. Although a general base for this reaction has not yet been identified, APPBP1's His211 and Uba3's His227, which have the potential to deprotonate the catalytic cysteines from either E1 or E2 are located 6.8 and 9.8 Å away from the Cys216 thiol. Further structural and biochemical analyses will be required to understand the details of the transthiolation reaction.

Implications of the MoeB–MoaD and APPBP1–Uba3 structures for other ubl-activating enzymes

The sequences and structures of ubl-activating enzymes suggest that they are modular enzymes with a common domain involved in ubl adenylation and specific domains to carry out other particular functions. The four ubls whose activating enzymes are most closely related are ubiquitin, SUMO, NEDD8 and ISG15. The activating enzymes for these four ubls have evolved conserved insertions when compared with the sequences of MoeB and other ubl-activating enzymes. The structure of APPBP1–Uba3 revealed that these E1-specific insertions have evolved to play a role in E1-specific functions such as transfer of the ubl from E1 to E2. Although an E2 has not yet been identified for ISG15, the strong similarity between the sequences of Ube1L, the E1 for ISG15, and the E1s for ubiquitin, SUMO and NEDD8 family members suggests that the ISG15 pathway will involve an E2 that resembles E2s for ubiquitin, SUMO and NEDD8.

The MoeB–MoaD and APPBP1–Uba3 crystal structures suggest how E1s might orchestrate their multiple activities. Although different domains perform adenylation, thioester bond formation, and recruitment of E2, together they form a large canyon divided into two clefts that coordinate the E1's catalytic activities with its multiple intermolecular interactions. Both clefts are likely to be involved in the adenylation reaction, with one recognizing the bulk of the ubl and the other one catalysing adenylation, binding to nucleotide and the C-terminus of the ubl. The subsequent thioester bond formation is likely to be facilitated by the catalytic cysteine facing the cleft occupied by the bulk of the ubl. The finding that the opposite cleft contains the nucleotide binding pocket, and that it might also contain the E2 binding site, provides a structural explanation for how nucleotide binding might promote transfer of the ubl from E1 to E2. The bound nucleotide may cause subtle structural rearrangements that affect E2 binding, or it may actually occupy part of the E2 binding site. The MoeB–MoaD and APPBP1–Uba3 structures provide a framework for understanding how E1s drive the initial steps of ubl transfer cascades in an assembly-line fashion.

References

1. Aberle H, Bauer A, Stappert J, Kispert A and Kemler R. (1997). EMBO J., 16, 3797–3804. | Article | PubMed | ISI | ChemPort |
2. Appleyard MV, Sloan J, Kana'n GJ, Heck IS, Kinghorn JR and Unkles SE. (1998). J. Biol. Chem., 273, 14869–14876. | Article | PubMed | ISI | ChemPort |
3. Aravind L and Koonin EV. (2000). Curr. Biol., 10, R132–R134. | Article | PubMed | ISI | ChemPort |
4. Bayer P, Arndt A, Metzger S, Mahajan R, Melchior F, Jaenicke R and Becker J. (1998). J. Mol. Biol., 280, 275–286. | Article | PubMed | ISI | ChemPort |
5. Beal R, Deveraux Q, Xia G, Rechsteiner M and Pickart C. (1996). Proc. Natl. Acad. Sci. USA, 93, 861–866. | Article | PubMed | ChemPort |
6. Bellon SF, Rodgers KK, Schatz DG, Coleman JE and Steitz TA. (1997). Nat. Struct. Biol., 4, 586–591. | Article | PubMed | ISI | ChemPort |
7. Bencsath KP, Podgorski MS, Pagala VR, Slaughter CA and Schulman BA. (2002). J. Biol. Chem., 277, 47938–47945. | Article | PubMed | ISI | ChemPort |
8. Bernier-Villamor V, Sampson DA, Matunis MJ and Lima CD. (2002). Cell, 108, 345–356. | Article | PubMed | ISI | ChemPort |
9. Bohnsack RN and Haas AL. (2003). J. Biol. Chem., 278, 26823–26830. | Article | PubMed | ISI | ChemPort |
10. Borthakur D, Basche M, Buikema WJ, Borthakur PB and Haselkorn R. (1990). Mol. Gen. Genet., 221, 227–234. | Article | PubMed | ISI | ChemPort |
11. Burch TJ and Haas AL. (1994). Biochemistry, 33, 7300–7308. | Article | PubMed | ISI | ChemPort |
12. Ciechanover A, Elias S, Heller H and Hershko A. (1982). J. Biol. Chem., 257, 2537–2542. | PubMed | ISI | ChemPort |
13. Ciechanover A, Finley D and Varshavsky A. (1984). Cell, 37, 57–66. | Article | PubMed | ISI | ChemPort |
14. Ciechanover A, Heller H, Katz-Etzion R and Hershko A. (1981). Proc. Natl. Acad. Sci. USA, 78, 761–765. | Article | PubMed | ChemPort |
15. del Pozo JC and Estelle M. (1999). Proc. Natl. Acad. Sci. USA, 96, 15342–15347. | Article | PubMed | ChemPort |
16. Deng L, Wang C, Spencer E, Yang L, Braun A, You J, Slaughter C, Pickart C and Chen ZJ. (2000). Cell, 103, 351–361. | Article | PubMed | ISI | ChemPort |
17. Desterro JM, Rodriguez MS, Kemp GD and Hay RT. (1999). J. Biol. Chem., 274, 10618–10624. | Article | PubMed | ISI | ChemPort |
18. Dittmar GA, Wilkinson CR, Jedrzejewski PT and Finley D. (2002). Science, 295, 2442–2446. | Article | PubMed | ISI | ChemPort |
19. Finley D, Ciechanover A and Varshavsky A. (1984). Cell, 37, 43–55. | Article | PubMed | ISI | ChemPort |
20. Freed E, Lacey KR, Huie P, Lyapina SA, Deshaies RJ, Stearns T and Jackson PK. (1999). Genes Dev., 13, 2242–2257. | Article | PubMed | ISI | ChemPort |
21. Freemont PS. (2000). Curr. Biol., 10, R84–R87. | Article | PubMed | ISI | ChemPort |
22. Furukawa K, Mizushima N, Noda T and Ohsumi Y. (2000). J. Biol. Chem., 275, 7462–7465. | Article | PubMed | ISI | ChemPort |
23. Ghosh S, May MJ and Kopp EB. (1998). Annu. Rev. Immunol., 16, 225–260. | Article | PubMed | ISI | ChemPort |
24. Giraud MF, Desterro JM and Naismith JH. (1998). Acta Crystallogr. D Biol. Crystallogr., 54, 891–898. | Article | PubMed | ISI |
25. Glotzer M, Murray AW and Kirschner MW. (1991). Nature, 349, 132–138. | Article | PubMed | ISI | ChemPort |
26. Goebl MG, Yochem J, Jentsch S, McGrath JP, Varshavsky A and Byers B. (1988). Science, 241, 1331–1335. | Article | PubMed | ISI | ChemPort |
27. Gong L, Li B, Millas S and Yeh ET. (1999). FEBS Lett., 448, 185–189. | Article | PubMed | ISI | ChemPort |
28. Gong L and Yeh ET. (1999). J. Biol. Chem., 274, 12036–12042. | Article | PubMed | ISI | ChemPort |
29. Haas AL, Ahrens P, Bright PM and Ankel H. (1987). J. Biol. Chem., 262, 11315–11323. | PubMed | ISI | ChemPort |
30. Haas AL and Bright PM. (1988). J. Biol. Chem., 263, 13258–13267. | PubMed | ISI | ChemPort |
31. Haas AL, Bright PM and Jackson VE. (1988). J. Biol. Chem., 263, 13268–13275. | PubMed | ISI | ChemPort |
32. Haas AL and Rose IA. (1982). J. Biol. Chem., 257, 10329–10337. | PubMed | ChemPort |
33. Haas AL, Warms JV, Hershko A and Rose IA. (1982). J. Biol. Chem., 257, 2543–2548. | PubMed | ISI | ChemPort |
34. Haas AL, Warms JV and Rose IA. (1983). Biochemistry, 22, 4388–4394. | Article | PubMed | ISI | ChemPort |
35. Hatakeyama S, Yada M, Matsumoto M, Ishida N and Nakayama KI. (2001). J. Biol. Chem., 276, 33111–33120. | Article | PubMed | ISI | ChemPort |
36. Hatfield PM, Callis J and Vierstra RD. (1990). J. Biol. Chem., 265, 15813–15817. | PubMed | ISI | ChemPort |
37. Hay RT. (2001). Trends Biochem. Sci., 26, 332–333. | Article | PubMed | ISI | ChemPort |
38. Hershko A and Ciechanover A. (1998). Annu. Rev. Biochem., 67, 425–479. | Article | PubMed | ISI | ChemPort |
39. Hershko A, Ciechanover A and Rose IA. (1981). J. Biol. Chem., 256, 1525–1528. | PubMed | ISI | ChemPort |
40. Hershko A, Ciechanover A and Varshavsky A. (2000). Nat. Med., 6, 1073–1081. | Article | PubMed | ISI | ChemPort |
41. Hershko A, Heller H, Elias S and Ciechanover A. (1983). J. Biol. Chem., 258, 8206–8214. | PubMed | ISI | ChemPort |
42. Hicke L. (1999). Trends Cell Biol., 9, 107–112. | Article | PubMed | ISI | ChemPort |
43. Hicke L. (2001a). Cell, 106, 527–530. | Article | PubMed | ISI | ChemPort |
44. Hicke L. (2001b). Nat. Rev. Mol. Cell. Biol., 2, 195–201. | Article | PubMed | ISI | ChemPort |
45. Hochstrasser M. (1996). Annu. Rev. Genet., 30, 405–439. | Article | PubMed | ISI | ChemPort |
46. Hochstrasser M. (1998). Genes Dev., 12, 901–907. | PubMed | ISI | ChemPort |
47. Hochstrasser M. (2000a). Science, 289, 563–564. | Article | PubMed | ISI | ChemPort |
48. Hochstrasser M. (2000b). Nat. Cell Biol., 2, E153–E157. | Article | PubMed | ISI | ChemPort |
49. Hoege C, Pfander B, Moldovan GL, Pyrowolakis G and Jentsch S. (2002). Nature, 419, 135–141. | Article | PubMed | ISI | ChemPort |
50. Hofmann K, Bucher P and Kajava AV. (1998). J. Mol. Biol., 282, 195–208. | Article | PubMed | ISI | ChemPort |
51. Honda R, Tanaka H and Yasuda H. (1997). FEBS Lett., 420, 25–27. | Article | PubMed | ISI | ChemPort |
52. Huang L, Kinnucan E, Wang G, Beaudenon S, Howley PM, Huibregtse JM and Pavletich NP. (1999). Science, 286, 1321–1326. | Article | PubMed | ISI | ChemPort |
53. Huibregtse JM, Scheffner M, Beaudenon S and Howley PM. (1995). Proc. Natl. Acad. Sci. USA, 92, 2563–2567. | Article | PubMed | ChemPort |
54. Ichimura Y, Kirisako T, Takao T, Satomi Y, Shimonishi Y, Ishihara N, Mizushima N, Tanida I, Kominami E, Ohsumi M, Noda T and Ohsumi Y. (2000). Nature, 408, 488–492. | Article | PubMed | ISI | ChemPort |
55. Jackson PK, Eldridge AG, Freed E, Furstenthal L, Hsu JY, Kaiser BK and Reimann JDR. (2000). Trends Cell Biol., 10, 429–439. | Article | PubMed | ISI | ChemPort |
56. Jentsch S, McGrath JP and Varshavsky A. (1987). Nature, 329, 131–134. | Article | PubMed | ISI | ChemPort |
57. Jentsch S and Pyrowolakis G. (2000). Trends Cell Biol., 10, 335–342. | Article | PubMed | ISI | ChemPort |
58. Joazeiro CA and Weissman AM. (2000). Cell, 102, 549–552. | Article | PubMed | ISI | ChemPort |
59. Joazeiro CA, Wing SS, Huang H, Leverson JD, Hunter T and Liu YC. (1999). Science, 286, 309–312. | Article | PubMed | ISI | ChemPort |
60. Johnson ES and Blobel G. (1997). J. Biol. Chem., 272, 26799–26802. | Article | PubMed | ISI | ChemPort |
61. Johnson ES and Gupta AA. (2001). Cell, 106, 735–744. | Article | PubMed | ISI | ChemPort |
62. Johnson ES, Schwienhorst I, Dohmen RJ and Blobel G. (1997). EMBO J., 16, 5509–5519. | Article | PubMed | ISI | ChemPort |
63. Jones D and Candido EP. (2000). Dev. Biol., 226, 152–165. | PubMed |
64. Kagey MH, Melhuish TA and Wotton D. (2003). Cell, 113, 127–137. | Article | PubMed | ISI | ChemPort |
65. Kahyo T, Nishida T and Yasuda H. (2001). Mol. Cell, 8, 713–718. | Article | PubMed | ISI | ChemPort |
66. Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin Jr WG, Elledge SJ, Conaway RC, Harper JW and Conaway JW. (1999). Science, 284, 657–661. | Article | PubMed | ISI | ChemPort |
67. Kishino T, Lalande M and Wagstaff J. (1997). Nat. Genet., 15, 70–73. | Article | PubMed | ISI | ChemPort |
68. Koepp DM, Harper JW and Elledge SJ. (1999). Cell, 97, 431–434. | Article | PubMed | ISI | ChemPort |
69. Kok K, Hofstra R, Pilz A, van den Berg A, Terpstra P, Buys CH and Carritt B. (1993). Proc. Natl. Acad. Sci. USA, 90, 6071–6075. | Article | PubMed | ChemPort |
70. Komatsu M, Tanida I, Ueno T, Ohsumi M, Ohsumi Y and Kominami E. (2001). J. Biol. Chem., 276, 9846–9854. | Article | PubMed | ISI | ChemPort |
71. Kumar S, Kao WH and Howley PM. (1997). J. Biol. Chem., 272, 13548–13554. | Article | PubMed | ISI | ChemPort |
72. Kurz T, Pintard L, Willis JH, Hamill DR, Gonczy P, Peter M and Bowerman B. (2002). Science, 295, 1294–1298. | Article | PubMed | ISI | ChemPort |
73. Kwon YT, Kashina AS, Davydov IV, Hu RG, An JY, Seo JW, Du F and Varshavsky A. (2002). Science, 297, 96–99. | Article | PubMed | ISI | ChemPort |
74. Lake MW, Wuebbens MM, Rajagopalan KV and Schindelin H. (2001). Nature, 414, 325–329. | Article | PubMed | ISI | ChemPort |
75. Lammer D, Mathias N, Laplaza JM, Jiang W, Liu Y, Callis J, Goebl M and Estelle M. (1998). Genes Dev., 12, 914–926. | PubMed | ISI | ChemPort |
76. Larsen CN and Wang H. (2002). J. Proteome Res., 1, 411–419. | Article | PubMed | ISI | ChemPort |
77. Leimkuhler S, Wuebbens MM and Rajagopalan KV. (2001). J. Biol. Chem., 276, 34695–34701. | Article | PubMed | ISI | ChemPort |
78. Liakopoulos D, Busgen T, Brychzy A, Jentsch S and Pause A. (1999). Proc. Natl. Acad. Sci. USA, 96, 5510–5515. | Article | PubMed | ChemPort |
79. Liakopoulos D, Doenges G, Matuschewski K and Jentsch S. (1998). EMBO J., 17, 2208–2214. | Article | PubMed | ChemPort |
80. Liu YC, Pan J, Zhang C, Fan W, Collinge M, Bender JR and Weissman SM. (1999). Proc. Natl. Acad. Sci. USA, 96, 4313–4318. | Article | PubMed | ChemPort |
81. Loeb KR and Haas AL. (1992). J. Biol. Chem., 267, 7806–7813. | PubMed | ISI | ChemPort |
82. Lorick KL, Jensen JP, Fang S, Ong AM, Hatakeyama S and Weissman AM. (1999). Proc. Natl. Acad. Sci. USA, 96, 11364–11369. | Article | PubMed | ChemPort |
83. MacQuillan GC, Mamotte C, Reed WD, Jeffrey GP and Allan JE. (2003). J. Med. Virol., 70, 219–227. | Article | PubMed | ISI | ChemPort |
84. Matsuura T, Sutcliffe JS, Fang P, Galjaard RJ, Jiang YH, Benton CS, Rommens JM and Beaudet AL. (1997). Nat. Genet., 15, 74–77. | Article | PubMed | ISI | ChemPort |
85. McGrath JP, Jentsch S and Varshavsky A. (1991). EMBO J., 10, 227–236. | PubMed | ISI | ChemPort |
86. McKenna S, Spyracopoulos L, Moraes T, Pastushok L, Ptak C, Xiao W and Ellison MJ. (2001). J. Biol. Chem., 276, 40120–40126. | Article | PubMed | ISI | ChemPort |
87. McLaughlin PM, Helfrich W, Kok K, Mulder M, Hu SW, Brinker MG, Ruiters MH, de Leij LF and Buys CH. (2000). Int. J. Cancer, 85, 871–876. | PubMed |
88. Meacham GC, Patterson C, Zhang W, Younger JM and Cyr DM. (2001). Nat. Cell Biol., 3, 100–105. | Article | PubMed | ISI | ChemPort |
89. Melchior F. (2000). Annu. Rev. Cell Dev. Biol., 16, 591–626. | Article | PubMed | ISI | ChemPort |
90. Mizushima N, Noda T, Yoshimori T, Tanaka Y, Ishii T, George MD, Klionsky DJ, Ohsumi M and Ohsumi Y. (1998a). Nature, 395, 395–398. | Article | PubMed | ISI | ChemPort |
91. Mizushima N, Sugita H, Yoshimori T and Ohsumi Y. (1998b). J. Biol. Chem., 273, 33889–33892. | Article | PubMed | ISI | ChemPort |
92. Moraes TF, Edwards RA, McKenna S, Pastushok L, Xiao W, Glover JN and Ellison MJ. (2001). Nat. Struct. Biol., 8, 669–673. | Article | PubMed | ISI | ChemPort |
93. Muller S, Hoege C, Pyrowolakis G and Jentsch S. (2001). Nat. Rev. Mol. Cell Biol., 2, 202–210. | Article | PubMed | ISI | ChemPort |
94. Narasimhan J, Potter JL and Haas AL. (1996). J. Biol. Chem., 271, 324–330. | Article | PubMed | ISI | ChemPort |
95. Nicholl MJ, Robinson LH and Preston CM. (2000). J. Gen. Virol., 81, 2215–2218. | PubMed |
96. Ohi MD, Vander Kooi CW, Rosenberg JA, Chazin WJ and Gould KL. (2003). Nat. Struct. Biol., 10, 250–255. | Article | PubMed | ISI | ChemPort |
97. Ohsumi Y. (2001). Nat. Rev. Mol. Cell Biol., 2, 211–216. | Article | PubMed | ISI | ChemPort |
98. Okuma T, Honda R, Ichikawa G, Tsumagari N and Yasuda H. (1999). Biochem. Biophys. Res. Commun., 254, 693–698. | Article | PubMed | ISI | ChemPort |
99. Osaka F, Kawasaki H, Aida N, Saeki M, Chiba T, Kawashima S, Tanaka K and Kato S. (1998). Genes Dev., 12, 2263–2268. | PubMed | ISI | ChemPort |
100. Pagano M, Tam SW, Theodoras AM, Beer-Romero P, Del Sal G, Chau V, Yew PR, Draetta GF and Rolfe M. (1995). Science, 269, 682–685. | Article | PubMed | ISI | ChemPort |
101. Palenchar PM, Buck CJ, Cheng H, Larson TJ and Mueller EG. (2000). J. Biol. Chem., 275, 8283–8286. | Article | PubMed | ISI | ChemPort |
102. Pawson T and Nash P. (2000). Genes Dev., 14, 1027–1047. | PubMed | ISI | ChemPort |
103. Pawson T and Nash P. (2003). Science, 300, 445–452. | Article | PubMed | ISI | ChemPort |
104. Paz Y, Elazar Z and Fass D. (2000). J. Biol. Chem., 275, 25445–25450. | Article | PubMed | ISI | ChemPort |
105. Pichler A, Gast A, Seeler JS, Dejean A and Melchior F. (2002). Cell, 108, 109–120. | Article | PubMed | ISI | ChemPort |
106. Pickart CM. (2001). Annu. Rev. Biochem., 70, 503–533. | Article | PubMed | ISI | ChemPort |
107. Pickart CM, Kasperek EM, Beal R and Kim A. (1994). J. Biol. Chem., 269, 7115–7123. | PubMed | ISI | ChemPort |
108. Pickart CM and Rose IA. (1985). J. Biol. Chem., 260, 1573–1581. | PubMed | ISI | ChemPort |
109. Pitluk ZW, McDonough M, Sangan P and Gonda DK. (1995). Mol. Cell. Biol., 15, 1210–1219. | PubMed |
110. Pitterle DM, Johnson JL and Rajagopalan KV. (1993). J. Biol. Chem., 268, 13506–13509. | PubMed | ISI | ChemPort |
111. Pitterle DM, Jolicoeur EM and Bepler G. (1998). In Vivo, 12, 643–658. | PubMed |
112. Pitterle DM and Rajagopalan KV. (1993). J. Biol. Chem., 268, 13499–13505. | PubMed | ISI | ChemPort |
113. Pozo JC, Timpte C, Tan S, Callis J and Estelle M. (1998). Science, 280, 1760–1763. | Article | PubMed | ChemPort |
114. Pringa E, Martinez-Noel G, Muller U and Harbers K. (2001). J. Biol. Chem., 276, 19617–19623. | Article | PubMed | ISI | ChemPort |
115. Raasi S, Schmidtke G and Groettrup M. (2001). J. Biol. Chem., 276, 35334–35343. | Article | PubMed | ISI | ChemPort |
116. Ramelot TA, Cort RJ, Yee AA, Semesi A, Edwards AM, Arrowsmith CH and Kennedy MA. (2003). J. Struct. Funct. Genom., 4, 25. | Article | ChemPort |
117. Rao-Naik C, delaCruz W, Laplaza JM, Tan S, Callis J and Fisher AJ. (1998). J. Biol. Chem., 273, 34976–34982. | Article | PubMed | ISI | ChemPort |
118. Read MA, Brownell JE, Gladysheva TB, Hottelet M, Parent LA, Coggins MB, Pierce JW, Podust VN, Luo RS, Chau V and Palombella VJ. (2000). Mol. Cell Biol., 20, 2326–2333. | Article | PubMed | ISI | ChemPort |
119. Rechsteiner M. (1998). Ubiquitin and the BIOLOGY of the Cell, Peters J-M, Harris JR, Finley D (eds) Plenum Press: New York, pp 147–189.
120. Rock KL, Gramm C, Rothstein L, Clark K, Stein R, Dick L, Hwang D and Goldberg AL. (1994). Cell, 78, 761–771. | Article | PubMed | ISI | ChemPort |
121. Rossmann MG, Moras D and Olsen KW. (1974). Nature, 250, 194–199. | Article | PubMed | ISI | ChemPort |
122. Rost B and Sander C. (1994). Proteins, 19, 55–72. | Article | PubMed | ISI | ChemPort |
123. Rudolph MJ, Wuebbens MM, Rajagopalan KV and Schindelin H. (2001). Nat. Struct. Biol., 8, 42–46. | Article | PubMed | ISI | ChemPort |
124. Sachdev S, Bruhn L, Sieber H, Pichler A, Melchior F and Grosschedl R. (2001). Genes Dev., 15, 3088–3103. | Article | PubMed | ISI | ChemPort |
125. Salceda S and Caro J. (1997). J. Biol. Chem., 272, 22642–22647. | Article | PubMed | ISI | ChemPort |
126. Scheffner M, Huibregtse JM, Vierstra RD and Howley PM. (1993). Cell, 75, 495–505. | Article | PubMed | ISI | ChemPort |
127. Scheffner M, Werness BA, Huibregtse JM, Levine AJ and Howley PM. (1990). Cell, 63, 1129–1136. | Article | PubMed | ISI | ChemPort |
128. Schmidt D and Muller S. (2002). Proc. Natl. Acad. Sci. USA, 99, 2872–2877. | Article | PubMed | ChemPort |
129. Schwartz DC and Hochstrasser M. (2003). Trends Biochem. Sci., 28, 321–328. | Article | PubMed | ISI | ChemPort |
130. Schwarz SE, Matuschewski K, Liakopoulos D, Scheffner M and Jentsch S. (1998). Proc. Natl. Acad. Sci. USA, 95, 560–564. | Article | PubMed | ChemPort |
131. Sheng W and Liao X. (2002). Protein Sci., 11, 1482–1491. | Article | PubMed | ISI | ChemPort |
132. Shintani T, Mizushima N, Ogawa Y, Matsuura A, Noda T and Ohsumi Y. (1999). EMBO J., 18, 5234–5241. | Article | PubMed | ISI | ChemPort |
133. Sidow A, Bulotsky MS, Kerrebrock AW, Birren BW, Altshuler D, Jaenisch R, Johnson KR and Lander ES. (1999). Nat. Genet., 23, 104–107. | Article | PubMed | ISI | ChemPort |
134. Sloper-Mould KE, Jemc JC, Pickart CM and Hicke L. (2001). J. Biol. Chem., 276, 30483–30489. | Article | PubMed | ISI | ChemPort |
135. Sullivan ML and Vierstra RD. (1991). J. Biol. Chem., 266, 23878–23885. | PubMed | ISI | ChemPort |
136. Swanson R and Hochstrasser M. (2000). FEBS Lett., 477, 193–198. | PubMed |
137. Takahashi Y, Kahyo T, Toh EA, Yasuda H and Kikuchi Y. (2001). J. Biol. Chem., 27, 27.
138. Tanida I, Mizushima N, Kiyooka M, Ohsumi M, Ueno T, Ohsumi Y and Kominami E. (1999). Mol. Biol. Cell, 10, 1367–1379. | PubMed | ISI | ChemPort |
139. Tanida I, Tanida-Miyake E, Ueno T and Kominami E. (2001). J. Biol. Chem., 276, 1701–1706. | PubMed | ISI | ChemPort |
140. Tateishi K, Omata M, Tanaka K and Chiba T. (2001). J. Cell Biol., 155, 571–579. | Article | PubMed | ISI | ChemPort |
141. Tatham MH, Jaffray E, Vaughan OA, Desterro JM, Botting CH, Naismith JH and Hay RT. (2001). J. Biol. Chem., 276, 35368–35374. | Article | PubMed | ISI | ChemPort |
142. Taylor SV, Kelleher NL, Kinsland C, Chiu H-J, Costello CA, Backstrom AD, McLafferty FW and Begley TP. (1998). J. Biol. Chem., 273, 16555–16560. | Article | PubMed | ISI | ChemPort |
143. Tong H, Hateboer G, Perrakis A, Bernards R and Sixma TK. (1997). J. Biol. Chem., 272, 21381–21387. | Article | PubMed | ISI | ChemPort |
144. Tyers M and Jorgensen P. (2000). Curr. Opin. Genet. Dev., 10, 54–64. | Article | PubMed | ISI | ChemPort |
145. VanDemark AP and Hill CP. (2002). Curr. Opin. Struct. Biol., 12, 822–830. | Article | PubMed | ISI | ChemPort |
146. VanDemark AP and Hill CP. (2003). Nat. Struct. Biol., 10, 244–246. | Article | PubMed | ISI | ChemPort |
147. VanDemark AP, Hofmann RM, Tsui C, Pickart CM and Wolberger C. (2001). Cell, 105, 711–720. | Article | PubMed | ISI | ChemPort |
148. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, Gocayne JD, Amanatides P, Ballew RM, Huson DH, Wortman JR, Zhang Q, Kodira CD, Zheng XH, Chen L, Skupski M, Subramanian G, Thomas PD, Zhang J, Gabor Miklos GL, Nelson C, Broder S, Clark AG, Nadeau J, McKusick VA, Zinder N, Levine AJ, Roberts RJ, Simon M, Slayman C, Hunkapiller M, Bolanos R, Delcher A, Dew I, Fasulo D, Flanigan M, Florea L, Halpern A, Hannenhalli S, Kravitz S, Levy S, Mobarry C, Reinert K, Remington K, Abu-Threideh J, Beasley E, Biddick K, Bonazzi V, Brandon R, Cargill M, Chandramouliswaran I, Charlab R, Chaturvedi K, Deng Z, Di Francesco V, Dunn P, Eilbeck K, Evangelista C, Gabrielian AE, Gan W, Ge W, Gong F, Gu Z, Guan P, Heiman TJ, Higgins ME, Ji RR, Ke Z, Ketchum KA, Lai Z, Lei Y, Li Z, Li J, Liang Y, Lin X, Lu F, Merkulov GV, Milshina N, Moore HM, Naik AK, Narayan VA, Neelam B, Nusskern D, Rusch DB, Salzberg S, Shao W, Shue B, Sun J, Wang Z, Wang A, Wang X, Wang J, Wei M, Wides R, Xiao C and Yan C et al. (2001). Science, 291, 1304–1351. | Article | PubMed | ISI | ChemPort |
149. Vijay-Kumar S, Bugg CE and Cook WJ. (1987). J. Mol. Biol., 194, 531–544. | Article | PubMed | ChemPort |
150. Vijay-Kumar S, Bugg CE, Wilkinson KD and Cook WJ. (1985). Proc. Natl. Acad. Sci. USA, 82, 3582–3585. | PubMed | ChemPort |
151. Wada H, Yeh ET and Kamitani T. (1999a). Biochem. Biophys. Res. Commun., 257, 100–105. | Article | PubMed | ISI | ChemPort |
152. Wada H, Yeh ET and Kamitani T. (1999b). J. Biol. Chem., 274, 36025–36029. | Article | PubMed | ISI | ChemPort |
153. Walden H, Podgorski MS and Schulman BA. (2003). Nature, 422, 330–334. | Article | PubMed | ISI | ChemPort |
154. Walker JE, Saraste M, Runswick MJ and Gay NJ. (1982). EMBO J., 1, 945–951. | PubMed | ISI | ChemPort |
155. Wang C, Xi J, Begley TP and Nicholson LK. (2001). Nat. Struct. Biol., 8, 47–51. | Article | PubMed | ISI | ChemPort |
156. Weissman AM. (2001). Nat. Rev. Mol. Cell Biol., 2, 169–178. | Article | PubMed | ISI | ChemPort |
157. Whitby FG, Xia G, Pickart CM and Hill CP. (1998). J. Biol. Chem., 273, 34983–34991. | Article | PubMed | ISI | ChemPort |
158. Wilkinson KD. (1997). FASEB J., 11, 1245–1256. | PubMed | ISI | ChemPort |
159. Wilkinson KD and Hochstrasser M. (1998). Ubiquitin and the Biology of the Cell, Peters J-M, Harris JR, Finley D (eds) Plenum Press: New York, pp. 99–125.
160. Wilkinson KD, Smith SE, O'Connor L, Sternberg E, Taggart JJ, Berges DA and Butt T. (1990). Biochemistry, 29, 7373–7380. | PubMed | ISI | ChemPort |
161. Xi J, Ge Y, Kinsland C, McLafferty FW and Begley TP. (2001). Proc. Natl. Acad. Sci. USA, 98, 8513–8518. | Article | PubMed | ChemPort |
162. Yaffe MB. (2002). Nat. Rev. Mol. Cell Biol., 3, 177–186. | Article | PubMed | ISI | ChemPort |
163. Yeh ET, Gong T and Kamitani T. (2000). Gene, 248, 1–14. | Article | PubMed | ISI | ChemPort |
164. Yewdell JW and Bennink JR. (2001). Curr. Opin. Immunol., 13, 13–18. | Article | PubMed | ISI | ChemPort |
165. York IA, Goldberg AL, Mo XY and Rock KL. (1999). Immunol. Rev., 172, 49–66. | Article | PubMed | ISI | ChemPort |
166. Yuan W and Krug RM. (2001). EMBO J., 20, 362–371. | Article | PubMed | ISI | ChemPort |
167. Zacksenhaus E and Sheinin R. (1990). EMBO J., 9, 2923–2929. | PubMed | ISI | ChemPort |
168. Zheng N, Wang P, Jeffrey PD and Pavletich NP. (2000). Cell, 102, 533–539. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We are grateful to Art Haas for helpful discussions and communication of results prior to publication. This work was supported by ALSAC, the NIH (P30CA21765 NCI Cancer Center Core grant to St. Jude, R01GM69530 to BAS) and a Pew Scholar Award in Biomedical Sciences to BAS.