Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. Jun 2009; 29(12): 3307–3318.
Published online Apr 13, 2009. doi:  10.1128/MCB.00240-09
PMCID: PMC2698738

Polyubiquitination by HECT E3s and the Determinants of Chain Type Specificity[down-pointing small open triangle]

Abstract

Polyubiquitination can mediate several different biochemical functions, determined in part by which lysine of ubiquitin is used to link the polyubiquitin chain. Among the HECT domain ubiquitin ligases, some, such as human E6AP, preferentially catalyze the formation of K48-linked polyubiquitin chains, while others, including Saccharomyces cerevisiae Rsp5 and human Itch, preferentially catalyze the formation of K63-linked chains. The features of HECT E3s that determine their chain type specificities have not been identified. We show here that chain type specificity is a function solely of the Rsp5 HECT domain, that the identity of the cooperating E2 protein does not influence the chain type specificity, that single chains produced by Rsp5 contain between 12 and 30 ubiquitin moieties, and that the determinants of chain type specificity are located within the last 60 amino acids of the C lobe of the HECT domain. Our results are also consistent with a simple sequential-addition mechanism for polyubiquitination by Rsp5, rather than a mechanism involving the formation of either E2- or E3-linked polyubiquitin chain transfers.

Ubiquitin can be covalently conjugated to proteins in several ways (20). Ubiquitin is sometimes conjugated via an isopeptide bond to a single lysine residue of a target protein and in other cases to multiple lysines. Less commonly, it is conjugated to the terminal amino group of a target (11) or even to cysteine side chains via a thioester bond (6). In all of these cases, a single ubiquitin might be conjugated at a given site (monoubiquitination), or multiple ubiquitins can be linked via one of the seven lysine residues of ubiquitin to form shorter oligoubiquitin chains (2- to 4-ubiquitin moieties) or longer polyubiquitin chains (>4-ubiquitin moieties). The chains might also be branched or linear, and if linear, either homogeneous or heterogeneous with respect to linkages (25). There are only a few cases for which we have a mechanistic understanding of how the enzymes of the ubiquitin system direct the generation of these distinct ubiquitin modifications. This is an important problem, because the different modes of ubiquitin conjugation have the potential to signal different biochemical fates. For example, lysine 29 (K29)- and K48-linked polyubiquitin chains are associated with proteasomal degradation, while K63-linked polyubiquitin chains have nonproteasomal functions in various signaling and trafficking pathways (4). All seven internal lysines of ubiquitin have been shown to be used for chain formation in vivo (34). The specific functions of some linkage types are uncharacterized, and there is the potential that some of these might mediate yet-to-be-discovered functions of polyubiquitin.

For polyubiquitination reactions that involve RING and RING-like U-box ubiquitin ligases, the type of polyubiquitin chains formed appears to be directed primarily by the cooperating E2 enzyme. This is presumably because the E3 is functioning primarily as a docking protein, with the chemistry of ubiquitination occurring between the E2 and the substrate protein. The best-characterized examples of this are reactions that involve Ubc13 and Mms2. Mms2 is a catalytically inactive E2-like enzyme that binds ubiquitin noncovalently and dimerizes with ubiquitin-charged Ubc13. The orientation of the Mms2-bound ubiquitin molecule allows only K63 to approach the Ubc13 ubiquitin-thioester bond, so that this combination of E2s promotes only K63-linked polyubiquitin chains (49). The Ubc13/Mms2-competent RING E3s include Rad5 (46, 47), Shprh (48), Chfr (5), and TRAF6 (12). Similar steric considerations, at the level of the E2~ubiquitin complex (the thioester bond is represented by ~), are likely to explain the chain type specificities of other E2/RING E3 combinations. The chemical environment surrounding specific lysine residues of ubiquitin might also contribute to chain type specificity by, for example, promoting deprotonation of the amino group of specific lysine side chains.

In contrast to RING E3s, HECT domain E3s directly catalyze protein ubiquitination. Ubiquitin-charged E2 enzymes transfer ubiquitin to the active-site cysteine within the HECT domain in a transthiolation reaction, preserving the high-energy ubiquitin thioester bond (40). Substrate ubiquitination then occurs by nucleophilic attack of the E3-ubiquitin thioester bond by a lysine side chain of the target protein, although the mechanism of polyubiquitination, as discussed below, has been unclear. The HECT domain is approximately 350 amino acids in length and is always found at the carboxyl-terminal end of HECT E3 proteins. Structural information is available for the HECT domains of human E6AP/Ube3A (19), WWP1 (50), Smurf2 (33), and Nedd4L1 (PDB 2ONI). All have an amino-terminal lobe (N lobe) of about 250 amino acids that contains the E2 binding site, and the structure of the E6AP/UbcH7 complex and detailed mutagenesis revealed the key determinants of E2/E3 interaction (13, 14, 19). All E2s that are known to function with HECT E3s belong to the group most similar to human UbcH7 and yeast Ubc4 and Ubc5. These E2s have a conserved phenylalanine residue in the loop between β3 and β4 strands (F63 in UbcH7) that is critical for the E2-HECT domain interaction. The carboxyl-terminal lobe (C lobe) of the HECT domain, consisting of about 100 amino acids, contains the active-site cysteine. The N and C lobes are connected by a short unstructured linker of 4 residues, and the differing orientations of the N and C lobes in the four HECT structures, together with mutagenesis of the WWP1 linker (50), strongly suggest that flexibility around the linker sequence is critical for protein ubiquitination. The flexibility is likely to be required to juxtapose the E2 and E3 active-site cysteines, which were separated by approximately 40 Å in the E6AP-UbcH7 structure, for the transthiolation reaction (19). The flexibility around the linker might also be required for the ability of the E3~ubiquitin complex to search three-dimensional space for accessible lysine residues on the target protein or on a polyubiquitin chain.

The simplest model for polyubiquitination by HECT E3s is that it is a processive reaction in which single ubiquitin molecules are added sequentially to the distal end of a substrate-linked chain during the course of a single round of substrate binding. Alternative models have been proposed (17), most notably the “indexation” model, in which a thioester-linked polyubiquitin chain is built on the E3 active-site cysteine and then the chain is transferred in bulk to the substrate (50). While indirect results supporting this model have been reported (53), the most obvious prediction of the model is that a thioester-linked chain on the E3 should be detectable, and this has not been reported or, to our knowledge, observed. An alternative “seesaw” model has also been suggested, in which multiple chain transfer events occur between the E2 and E3 to build a thioester-tethered chain, which would be transferred in bulk to the substrate. Each of these models faces significant conceptual problems, and all have been difficult to prove or disprove. The results presented here provide new support for the simple sequential-addition model.

Regardless of the mechanism, it is clear that different HECT E3s have specificities for the types of polyubiquitin chains that they synthesize. E6AP/Ube3A is a human HECT E3 that is hijacked by the human papillomavirus (HPV) E6 oncoprotein to target p53 for ubiquitin-mediated degradation, and disruption of expression in E6AP/Ube3A in brain neurons is the cause of Angelman syndrome (41). E6AP is highly specific for catalysis of K48-linked polyubiquitination (25, 53), consistent with the fact that p53 is targeted for proteasomal degradation in HPV E6-expressing cells. Human KIAA10 HECT E3 preferentially catalyzes both K48 and K29 linkages (52). In contrast, Saccharomyces cerevisiae Rsp5 preferentially synthesizes K63 chains in vitro and in vivo (23, 24, 38). Rsp5 contains an N-terminal C2 domain and three WW domains in the central region of the protein, and there are nine human HECT E3s that share this domain organization (21). Two of these, Nedd4-1 and Itch/AIP4, have also been shown to preferentially synthesize K63 chains in vitro (25, 44), although Itch/AIP4 was also reported to form K29 chains in vivo (7). The chain type specificity of human Huwe1/ArfBP1, whose reported substrates include p53, ARF, Mcl1, C-Myc, and N-Myc, is still unclear. It was shown to synthesize K48 chains on N-Myc (55), which was contrary to a previous report showing that Huwe1/ArfBP1 preferentially assembles K63 chains on C-Myc (1). Importantly, the chain type specificities of most HECT E3s have not been determined, and based on the few characterized examples, it is not possible to identify the sequence or structural determinants that distinguish or predict, for example, K48-specific and K63-specific E3s. The goals of the current study were to characterize the polyubiquitination reaction catalyzed by HECT E3s and to identify the determinants of HECT E3s that confer specificity for the synthesis of polyubiquitin chain types. Our results indicate that the C-terminal lobe of the HECT domain contains the critical features that determine chain type specificity.

MATERIALS AND METHODS

Plasmids.

Plasmids for the expression of p53, Wbp2, Ubc1, Ubc4, Ubc5, UbcH7, Rsp5, and NEDD4-1 were described previously (3, 23, 51). Clone KIAA0312 (HUGE protein database) was used as the template in PCR for the construction of chimeric proteins containing the C lobe of Huwe1. The genes encoding Sna3, Ubc3, Ubc6, Ubc8, Ubc11, and Ubc13 were amplified by PCR from genomic DNA of S. cerevisiae. Deletion of introns of Ubc8 and Ubc13; truncations of Wbp2, Rsp5, and Ubc1; and the generation of chimeric E3 genes were performed by standard PCR methods. In order to express 32P-labeled p53, ΔWW1,2,3-Rsp5, Wbp2, and Sna3, the genes encoding each protein were designed to encode the cyclic-AMP-dependent kinase recognition site (RRASV) at the 5′ end. All of the genes described above were cloned into pGEX6p-1 (GE Healthcare) for bacterial expression, and the sequences were verified by DNA sequencing.

Protein expression and purification.

pGEX6p-1 expression plasmids were transformed into Escherichia coli BL21. Bacterial cells were collected after a 5-h induction of protein expression in the presence of 0.1 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) and resuspended in phosphate-buffered saline containing 1% Triton X-100 and 100 μM phenylmethylsulfonyl fluoride. After lysis of the cells by sonication, glutathione S-transferase (GST) fusion proteins were affinity-purified on GST-Bind resin (Novagen). PreScission protease (GE Healthcare) was used to remove the GST portion of the proteins. Protein concentrations were normalized by Coomassie blue staining of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and infrared imaging (Li-Cor Biosciences). For labeling of proteins with 32P, GST-p53, GST-Wbp2 proteins, GST-Sna3 proteins, and GST-ΔWW1,2,3-Rsp5 proteins were incubated with adenosine 5′-[γ-32P]triphosphate (Perkin-Elmer Life Sciences) and cyclic-AMP-dependent protein kinase (Promega) on glutathione beads for 40 min at room temperature prior to GST cleavage by PreScission protease. E6AP and HPV type 33 (HPV33) E6 proteins were expressed using recombinant baculoviruses, as described previously (39). Briefly, both proteins were expressed as GST fusion proteins in High Five insect cells, affinity purified, and cleaved to remove the GST.

In vitro ubiquitination assays.

In vitro ubiquitination assays were carried out in the presence of 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM MgCl2, 0.1 mM dithiothreitol, 50 μg/ml ubiquitin (Boston Biochem), 1.25 μg/ml of E1 (Boston Biochem), 1.25 μg/ml of E2, and 5 μg/ml of E3 (40-μl total volume). Wild-type ubiquitin and all mutant forms of ubiquitin were from Boston Biochem. For the ubiquitination by Rsp5 or Rsp5-based chimeric E3s, Uba1 and Ubc4 were used as the E1 and the E2, respectively, while human E1 and UbcH7 were used for ubiquitination by E6AP, Itch, and Itch-based chimeric E3s. Purified HPV33 E6 (5 μl/ml) was added for the ubiquitination of p53 by E6AP. All ubiquitination reactions were initiated by the addition of ubiquitin and stopped by the addition of 4× SDS-PAGE loading buffer containing 0.4 M dithiothreitol. Unless otherwise indicated, reaction mixtures were incubated for 30 min at room temperature. For kinetic analysis of chain formation, the amount of ubiquitinated or unmodified substrate was quantitated and analyzed with a Bio-Rad phosphorimager and Quantity One software.

RESULTS

Chain type specificities of Rsp5 and E6AP and generation of simplified Rsp5 substrates.

To establish the chain type specificities of E6AP and Rsp5 in a purified in vitro system, ubiquitination reactions were performed with purified substrates and E1, E2, and E3 enzymes in the presence of wild-type or variant forms of ubiquitin. The substrate for E6AP was p53, which requires the additional presence of the E6 protein of a cancer-associated HPV type, such as HPV33 E6, for recruiting p53 to E6AP (2). Wild-type p53 was expressed as a GST fusion protein in bacteria, and after purification, the protein was radiolabeled with 32P at a kinase recognition site immediately downstream of the protease cleavage site separating the GST and p53 sequences. The cleaved 32P-labeled p53 protein was used as the ubiquitination substrate. Ubiquitin variants contained either a single lysine-to-arginine mutation (e.g., K63R) or six lysine-to-arginine mutations, so that only a single lysine remained [(e.g., K63(1) ubiquitin], or no lysines [K(0) ubiquitin]. As shown in Fig. Fig.1A1A (left), p53 was efficiently ubiquitinated by E6AP in the presence of E6 and wild-type ubiquitin and a cognate E2, UbcH7 (19), leading to the formation of very high-molecular-weight conjugates after 10 minutes of incubation. The products appeared identical in the presence of each of the ubiquitin mutants containing a single K-to-R mutation, with the exception of K48R ubiquitin, in which case the products were distributed around a much lower molecular weight. The products seen in the presence of K(0) ubiquitin (Fig. (Fig.1A,1A, right) appeared similar to those with K48R, suggesting that the products generated with K48R ubiquitin might represent p53 with multiple monoubiquitin modifications at several lysine residues. Consistent with this, the products observed with each of the ubiquitin mutants containing six K-to-R mutations, with the exception of K48(1), also appeared similar to those with K(0). Very high-molecular-weight conjugates were formed in the presence of K48(1) ubiquitin, similar to those seen with wild-type ubiquitin. These results indicate that E6AP has a strong preference for catalysis of K48-linked polyubiquitination of p53 in vitro.

FIG. 1.
Chain type specificities of E6AP and Rsp5. (A) In vitro ubiquitination of 32P-labeled p53 in the presence of human E1, UbcH7, E6AP, and HPV33 E6 proteins, with wild-type protein (WT) or the indicated ubiquitin (Ub) mutant proteins. Following a 10-min ...

The substrates used for Rsp5 were Wbp2 and Sna3, both of which contain “PY” motifs (PPXY sequences) that are recognized by the WW domains of Rsp5. Wbp2 is a human protein identified in a screen for WW domain binding proteins (8), and it contains three PY motifs and 15 lysine residues. Sna3 is a physiologic target of Rsp5 that contains a single PY motif and four lysines, with K125 being the in vivo modification site (45). Because the modification of several lysine residues can lead to ambiguities in the analysis of reaction products (i.e., whether a multiubiquitinated substrate is monoubiquitinated at several sites or polyubiquitinated at one site), we sought to generate Rsp5 substrates containing a single lysine residue. This was possible for Rsp5 substrates because of the small and well-defined modules that are the basis for enzyme substrate recognition (WW domain-PY motifs), but it was not possible for E6AP, which recognizes the DNA binding domain of p53 (42), which has several lysine residues that are critical for folding or function (9). In preliminary experiments, we determined that a C-terminal fragment of Wbp2 containing the three PY motifs and only a single lysine residue (Wbp2-C-K222) (Fig. (Fig.1B)1B) was ubiquitinated as efficiently as full-length wild-type Wbp2 (not shown). Wbp2-C-K222 was purified and radiolabeled as described above and used in ubiquitination reactions with Rsp5 and ubiquitin variants, using yeast Ubc4 as the E2. Wbp2-C-K222 was very efficiently ubiquitinated in the presence of wild-type ubiquitin, and of the single K-to-R ubiquitin mutants (Fig. (Fig.1C,1C, left), K63R had the greatest effect on the size distribution of the ubiquitinated products. Interestingly, Wbp2-C-K222, containing only a single lysine, appeared to be diubiquitinated in the presence of K(0) ubiquitin (Fig. (Fig.1C,1C, right), suggesting that Rsp5 might ubiquitinate the terminal amino group of the protein (addressed below). Ubiquitination in the presence of K6(1), K27(1), and K29(1) ubiquitins was similar to that with K(0) ubiquitin, indicating that Rsp5 cannot extend ubiquitin chains off of K6, K27, or K29. Higher-molecular-weight products were observed with K11(1), K33(1), K48(1), and K63(1) ubiquitins, although the products were consistently of much higher molecular weight with K63(1). Together, these results indicate that Rsp5 preferentially catalyzes K63-linked polyubiquitination of Wbp2 but that it will synthesize shorter K11, K33, or K48 chains in the absence of K63.

The same type of analysis was performed with a single-lysine derivative of Sna3 (Sna3-K125) (Fig. (Fig.1B).1B). The results were very similar to those seen with the Wbp2 substrate, with an even more pronounced effect on the size distribution of products seen with K63R ubiquitin compared to wild-type ubiquitin and the other K-to-R single mutants (Fig. (Fig.1D,1D, left). Again, the chains formed with K63(1) ubiquitin were longer than chains formed with the other single-lysine derivatives of ubiquitin, but shorter chains were formed with K11(1), K33(1), and K48(1) ubiquitins (Fig. (Fig.1D,1D, right). Like the single-lysine mutant of Wbp2, Sna3-K125 also appeared to be modified with two ubiquitin molecules in the presence of K(0) ubiquitin. Together, the results shown in Fig. Fig.11 indicate that different HECT E3s can have strikingly different characteristics with respect to the synthesis of polyubiquitin chains. Furthermore, the chain type specificity was not dependent on the identity of the substrate, as shown here for Rsp5 and as suggested earlier for both Rsp5 and E6AP (23, 24).

The Rsp5 reactions described above were performed at a single incubation time (10 min). Figure Figure22 shows a time course of Wbp2-C-K222 ubiquitination by Rsp5 in the presence of wild-type ubiquitin and K63R, K48R, K63(1), and K48(1) ubiquitins (Fig. 2A to E, respectively). The reactions were initiated by the addition of ubiquitin after a 10-min preincubation of all other components (E1, Ubc4, Rsp5, 32P-labeled substrate protein, and ATP). The overall rates of disappearance of the unmodified substrate were similar in each set of reactions, although there was a slight delay in the disappearance of substrate with K63(1) and K48(1) ubiquitins (Fig. 2D and E), which was likely related to a delay in charging of one or more of the enzymes with a ubiquitin molecule that contains six surface residue alterations. There was no significant difference between the rates of appearance of high-molecular-weight conjugates in the presence of wild-type ubiquitin and K48R ubiquitin (Fig. 2A and B), as predicted from the results shown in Fig. Fig.1.1. In contrast, the rate of generation of long chains was significantly lower in the presence of K63R ubiquitin (Fig. (Fig.2C).2C). K63(1) ubiquitin resulted in kinetics of chain formation similar to those with wild-type ubiquitin, although the chains were slightly shorter at each time point, and even at the latest time point, the conjugates did not reach the very tight distribution of high-molecular-weight products seen with wild-type ubiquitin (Fig. (Fig.2D).2D). While the basis of the latter effect is not known, the six K-to-R surface mutations of K63(1) might affect the conformation of the elongating chain in a manner that prevents formation of the very longest chains. K48(1) ubiquitin showed delayed chain formation, and very high-molecular-weight products were not observed (Fig. (Fig.2E).2E). These results were quantitated as the fraction of the total ubiquitinated Wbp2 protein that contained chains of more than four ubiquitin molecules, and this was plotted as a function of the reaction time (Fig. (Fig.2F).2F). This demonstrates that the rates of formation of long chains were very similar so long as K63 was intact (wild-type, K63(1), and K48R ubiquitins). Chain formation was significantly slower in the presence K63R, and given the results presented in Fig. Fig.1,1, these chains were likely to contain K11, K33, or K48 linkages. The rate of polyubiquitination was lower still in the presence of K48(1) ubiquitin. These results further substantiate the preferential formation of K63-linked polyubiquitin chains by Rsp5.

FIG. 2.
Time course of ubiquitination by Rsp5, with 32P-labeled Wbp2-C-K222 as a substrate, in the presence of wild-type (A), K48R (B), K63R (C), K63R(1) (D), and K48(1) (E) ubiquitin (Ub). Mono-, di-, tri-, and tetraubiquitinated substrates are designated Ub1, ...

N-terminal substrate modification and the length of Rsp5-catalyzed chains.

We had planned to utilize a single-lysine Rsp5 substrate (Wbp2-C-K222 or Sna3-K125), which we expected to be modified at a single site, to determine the lengths of the K63 chains that Rsp5 can synthesize. However, the results presented above suggested that Rsp5 also modified the terminal amino group of both proteins. To confirm this, a lysineless derivative of Wbp2-C was generated. Figure Figure3A3A shows the ubiquitination of Wbp2-C, Wbp2-C-K222, and Wbp2-C-K(0) (containing three, one, or no lysines, respectively). Consistent with N-terminal modification, high-molecular-weight ubiquitinated products were generated with Wbp2-C-K(0), although a significant amount of unmodified substrate remained compared to Wbp2-C and Wbp2-C-K222. In addition, the migration of the products generated in the presence of K(0) ubiquitin indicated that Rsp5 ubiquitinated four sites on Wbp2-C, two sites on Wbp2-C-K222, and one site on Wbp2-C-K(0). To rule out an acceptor site on Wbp2-C-K(0) other than the terminal amino group, GST-Wbp2-C-K222 and GST-Wbp2-C-K(0), with the GST moiety still attached, were ubiquitinated by Rsp5 in the presence of K(0) ubiquitin (Fig. (Fig.3B).3B). Multiubiquitin conjugates were formed on both proteins, indicative of internal lysines of GST serving as ubiquitin acceptors. In a parallel set of reactions, the GST tags were removed by proteolysis after the ubiquitination reactions were terminated, with the prediction being that the GST moiety would protect from modification the normally exposed terminal amino group of Wbp2-C. Consistent with this, the postcleavage Wbp2-C-K(0) protein (which retained the 32P label) was not ubiquitinated, while Wbp2-C-K(222) was monoubiquitinated. Together, these results indicate that Wbp2-C-K(0) is indeed modified at the terminal amino group by Rsp5. Similar results showing that Rsp5 can modify terminal amino groups were recently reported (38). Figure Figure3C3C shows a time course of modification of Wbp2-C-K(0) and Wbp2-C-K222, performed in the presence of K(0) ubiquitin. Wbp2-C-K222 was modified in a biphasic manner, with the second phase of modification corresponding to the rate of monoubiquitination of Wbp2-C-K(0), indicating that Rsp5 modifies the internal lysine more efficiently than the terminal amino group.

FIG. 3.
N-terminal substrate modification and the lengths of Rsp5-catalyzed chains. (A) In vitro ubiquitination of 32P-labeled Wbp2-C, Wbp2-C-K222, and Wbp2-C-K(0) with K(0) or wild-type (WT) ubiquitin (Ub). The reaction mixtures were incubated for 10 min at ...

Having characterized the single site of modification of Wbp2-C-K(0), we used this substrate to estimate the length of the K63 polyubiquitin chain that Rsp5 is capable of generating. Wbp2-C-K(0) was ubiquitinated by Rsp5 in the presence of K63(1) ubiquitin or in the presence of a combination of K63(1) and K63(0) at a ratio of 16:1. The latter was done in order to create chain termination products that could be resolved more clearly by SDS-PAGE. As shown in Fig. Fig.3D,3D, a ladder of ubiquitinated products containing up to 12 ubiquitin moieties could be discerned; however, higher-molecular-weight products that could not be resolved into discrete bands clearly extended beyond this length. Assuming that the migration of higher-molecular-weight products continued to be proportional to chain length, we calculated that the longest chains contained approximately 30 ubiquitin moieties. We therefore conclude that Rsp5 can synthesize homogeneous K63 chains containing at least 12 and perhaps up to 30 ubiquitin molecules. While it is not known whether Rsp5 products ever reach these lengths in vivo, mechanistic models for ubiquitination by HECT E3s must be consistent with this result.

The identity of the cooperating E2 does not influence the chain type specificity of Rsp5.

Results with the APC (anaphase promoting complex) and other RING E3s indicate that their polyubiquitination products are highly dependent on the identity of the cooperating E2 enzyme(s) (10, 16, 25, 27, 37). In order to determine whether the type of polyubiquitin chain formed by Rsp5 was also dependent on the E2, 8 of the 11 yeast E2s (Ubc1, -3, -4, -5, -6, -8, -11, and -13) were expressed and purified for in vitro ubiquitination reactions, and each was shown to be active in ubiquitin-thioester assays (not shown). Based on previous genetic and biochemical interactions (18, 29, 32), we expected that only Ubc1, -4, and -5 would function with Rsp5. Ubc1 contains a C-terminal UBA ubiquitin binding domain that has been suggested to influence its polyubiquitination activities (30, 37), so a Ubc1-ΔUBA protein was also analyzed. As shown in Fig. Fig.4A,4A, Ubc1, -4, and -5 were indeed the only E2s that supported the ubiquitination of Wbp2-C-K222 by Rsp5, and Ubc1-ΔUBA functioned similarly to wild-type Ubc1. The polyubiquitin chain types formed by Rsp5 were examined in the presence of Ubc1, Ubc5, and Ubc1-ΔUBA. As shown in Fig. 4B to D, the Rsp5 chain preference was the same with Ubc1, Ubc5, and Ubc1-ΔUBA as it was with Ubc4 (used in the experiments in Fig. Fig.1).1). Furthermore, the Rsp5 chain preference was the same in the presence of human UbcH7 (Fig. (Fig.4E),4E), which, as shown in Fig. Fig.1,1, cooperated with E6AP in catalyzing K48 polyubiquitin chains. Therefore, while HECT E3s cooperate with only a subset of E2s (e.g., those most closely related to yeast Ubc4 and human UbcH7), the chain type specificity of HECT E3-catalyzed reactions is solely a function of the E3, and not of the E2. This result has important implications for the indexation model of polyubiquitination by HECT E3s (see Discussion).

FIG. 4.
The chain type specificity of Rsp5 is not influenced by the identity of E2. (A) 32P-labeled Wbp2-C-K222 was ubiquitinated in the presence of Uba1, Rsp5, ubiquitin, and the indicated yeast E2s. (B to E) The chain type specificity of Rsp5 was determined ...

The determinants of RSP5 chain type specificity are within the HECT domain.

We next sought to determine whether sequences outside of the HECT domain influence the chain type specificity of Rsp5 or whether the HECT domain contained all the determinants of chain type specificity. A series of truncated Rsp5 proteins were expressed and purified (Fig. (Fig.5A).5A). The N-terminal C2 domain and WW domains 1 and 2 are not required for binding to Wbp2 (not shown), and both the ΔC2 and ΔWW1,2 proteins ubiquitinated Wbp2-C-K222 with the same chain type preference as wild-type Rsp5 (Fig. 5B and C). The ΔWW1,2,3 mutant, lacking a substrate recruitment domain, did not bind or ubiquitinate Wbp2, as expected (not shown). We therefore used Rsp5 autoubiquitination to analyze the chain type preference of this enzyme. Rsp5-ΔWW1,2,3 preferentially conjugated K63 chains to itself, with shorter chains formed in the presence of K11(1), K33(1), or K48(1). These results indicate that the C2 and WW domains do not influence the chain type specificity of Rsp5 and that chain type specificity is an inherent property of the HECT domain itself.

FIG. 5.
Rsp5 chain type specificity is a function of the HECT domain. (A) Schematic of Rsp5 truncation mutants used for in vitro ubiquitination assays. (B and C) ΔC2 Rsp5 was used in ubiquitination assays with 32P-labeled Wbp2-C-K222 in the presence of ...

Chain type specificity determinants map to the HECT C lobe.

To more precisely map the determinants of chain type specificity within the Rsp5 HECT domain, an initial set of chimeric enzymes was created that exchanged the complete HECT domain, the N lobe, the C lobe, or the N-C lobe linker sequence of Rsp5 with the analogous region of a K48-specific HECT E3, E6AP (Fig. 6A and B). Chimeras A and B, containing the complete E6AP HECT domain or the E6AP N lobe, were not biochemically active, based on both ubiquitin thioester assays and substrate ubiquitination assays, and therefore could not be analyzed. Chimeras C, D, and E were active. Chimera C contained three amino acid substitutions so that the N-C lobe linker sequence was converted from GIAE to GSRN (residues 690 to 693 of Rsp5). This protein efficiently ubiquitinated Wbp2 and had the same specificity for synthesis of the various types of polyubiquitin chains as wild-type Rsp5 (Fig. (Fig.6C),6C), indicating that the N-lobe-C-lobe linker sequence does not represent a determinant of chain type specificity of Rsp5.

FIG. 6.
Chain type specificities of chimeric Rsp5-E6AP proteins. (A) Structure of the E6AP HECT domain. The N lobe (blue), C lobe (green), hinge (red), and active-site cysteine (yellow) are indicated. (B) Schematic diagram of wild-type Rsp5 (top) and chimeras ...

Chimera D switched the C lobe and linker sequence of Rsp5 with those from E6AP, and this protein, like E6AP, catalyzed K48-linked ubiquitination nearly exclusively (Fig. (Fig.6D),6D), suggesting that the C lobe of the HECT domain contains the main determinants of chain type specificity. Chimera D also catalyzed K48 polyubiquitination of Sna3 (not shown). Finally, chimera E contained E6AP sequences from only the C-terminal portion of the C lobe, replacing the C-terminal 62 amino acids of Rsp5 with the analogous region of E6AP. This protein, like the complete C-lobe chimera (chimera D), catalyzed primarily K48 polyubiquitination (Fig. (Fig.6E),6E), with a slight increase in K11 and K33 chains relative to chimera D. These results suggest that the critical determinants of chain type specificity are located primarily within the C-terminal 60 amino acids of the C lobe.

Three additional Rsp5 chimeras that contained the C lobes of different HECT E3s were expressed: NEDD4-1, Itch/AIP4, and Huwe1/ArfBP1 (Fig. (Fig.7A).7A). NEDD4-1 and Itch/AIP4 are, like Rsp5, C2 and WW domain HECT E3s. Both have been previously reported to catalyze K63-linked chains (confirmed for full-length Itch) (Fig. (Fig.8).8). The chain type specificity of Huwe1/ArfBP1, which is not an Rsp5 homolog, is unclear, with one report of K63 polyubiquitination (1) and another of K48 polyubiquitination (55). Figure 7B and C shows that the Rsp5-NEDD4 and Rsp5-Itch C-lobe chimeras preferentially formed K63-linked polyubiquitin chains, with the secondary ability to form shorter K11, K33, and K48 chains, like wild-type Rsp5. The pattern of ubiquitination with the Rsp5-Huwe1 C-lobe chimera [Huwe1 (a)] was more complex but appeared to be a combination of Rsp5-like and E6AP-like activities: it retained K63 polyubiquitination activity at a level similar to that of the K11 and K33 chain formation activity while having the ability to synthesize slightly longer K48 chains (Fig. (Fig.7D).7D). An Rsp5-Huwe1 chimera that contained only the C-terminal half of the Huwe1 C lobe [Huwe1 (b)] showed a pattern of ubiquitination similar to that of the chimera that contained the complete Huwe1 C lobe (Fig. (Fig.7E),7E), consistent with the notion that the C-terminal portion of the C lobe contains the critical determinants for chain type synthesis.

FIG. 7.
Chain type specificities of chimeric Rsp5-Nedd4, -Itch, and -Huwe1 proteins. (A) Schematic diagram of the Rsp5-based chimeric E3s in which the C lobe of Rsp5 was replaced by the C lobe of NEDD4 (red), Itch (green), or Huwe1 (blue). The Huwe1 (a) chimera ...
FIG. 8.
Chain type specificities of chimeric Itch-E6AP and Itch-Huwe1 proteins. (A) Schematic diagram of Itch-based chimeras, with the C lobe of Itch replaced by that of E6AP (yellow) or Huwe1 (blue). (B to D) Wild-type Itch (B) or the Itch-E6AP (C) or Itch-Huwe1 ...

We next sought to determine whether the results with the Rsp5-based chimeras applied to other HECT E3s. Unfortunately, all of the E6AP-based chimeras (equivalent to Rsp5 chimeras A, B, D, and E) were inactive and could not be analyzed. A similar finding with E6AP chimeric E3s was reported previously (43). Chimeras were therefore created in the context of full-length Itch/AIP4 (Fig. (Fig.8A),8A), which also binds and ubiquitinates Wbp2-C-K222. As shown in Fig. Fig.8B,8B, wild-type Itch had a very pronounced specificity for the synthesis of K63 chains. The replacement of the C lobe of Itch/AIP4 with that of E6AP led to a complete switch to K48 chain type specificity (Fig. (Fig.8C).8C). The Huwe1 C lobe had an effect on Itch very similar to the effect it had on Rsp5, where K48 chains were the longest by a small margin, while K63 chain activity persisted at a level comparable to the K11 and K33 chain formation activities. Together, the results presented here indicate that the chain type specificity of HECT E3s is a function of the last ~60 amino acids of the C lobe of the HECT domain, encompassing the terminal three β-strands, the active-site loop, and the terminal α-helix.

DISCUSSION

The results presented here address several aspects of the mechanism of ubiquitination and polyubiquitination by HECT E3s. First, in the simplest model of polyubiquitination by HECT E3s, the sequential-addition model, a single ubiquitin is conjugated to a substrate protein, and the chain is elongated by conjugation of additional ubiquitin monomers to the growing (distal) end of the chain. A conceptual difficulty with this model is that it must account for the formation of potentially long chains, and we have shown here, using a substrate with a single site of modification, that these chains have the potential to contain between 12 and least 30 molecules. An alternative model, the “indexation” model, has been proposed that might account for chain formation by HECT E3s (17, 50). In this model, a thioester-tethered polyubiquitin chain is built on the E3 active-site cysteine by sequential reactions with E2~ubiquitin, with a lysine residue of the E3-bound ubiquitin attacking the E2~ubiquitin thioester in multiple rounds of reaction. The polyubiquitin chain could then theoretically be transferred in a one-step reaction to a substrate lysine. This model was initially proposed based on the differing C-lobe positions seen in the E6AP and WWP1 HECT domain structures (19, 50), which suggested that a ratcheting of the C lobe away from the E2 as a chain was built on the E3 active-site cysteine might account for the differing C-lobe positions seen in the two structures. Importantly, this model faces a conceptual problem similar to that of the sequential model with respect to forming long chains (i.e., finding and orienting the distal end of the chain), the only difference being that in the indexation model the growing chain is tethered to the E3 (via a thioester bond) rather than the substrate (through an isopeptide bond). Alternatively, a combination of the sequential and indexation models could be imagined, where a short (e.g., four-ubiquitin) chain might be built on the E3 and transferred to the substrate, with the E3 being recharged and transferring another short chain to the end of the first, and so on. In this case, the model predicts that polyubiquitination might occur in groups of x ubiquitins, where x is the length of the chain built on the E3. As clearly shown here by time course experiments, this is not the case for Rsp5 or, to our knowledge, for any other HECT E3s examined to date. A monoubiquitinated product is the major product at the earliest reaction time points. While it has been reported that E6AP can catalyze diubiquitination by a mechanism consistent with the indexation model (53), no direct evidence was found for the existence of the key predicted intermediate in this reaction (diubiquitin in a thioester linkage at the E3 active site). We have also been unsuccessful in attempting to observe HECT-ubiquitin thioesters that contain more than a single ubiquitin moiety (not shown).

A critical distinction between the simple sequential-addition and indexation models is that in the sequential model the last ubiquitin of the chain must be oriented toward the active site of the E3 for addition of the next ubiquitin molecule, while in the indexation model, the last ubiquitin of the E3-tethered chain must be oriented toward the active site of the E2. An important result in our study was that the identity of the E2 did not influence the type of ubiquitin chains formed by HECT E3s. That is, a single E2 (UbcH7) functioned with E6AP to catalyze K48-linked polyubiquitination and with Rsp5 to catalyze K63-linked polyubiquitination. It is difficult to imagine that the identity of the E2 would not influence the chain type in the indexation model, since the E2 is an active participant in the formation of the E3-tethered chain. By comparison, the formation of the E3-tethered chain would be equivalent to an E2-dependent RING E3 reaction, and the identity of the E2 has clearly been shown to determine the type of chains formed in several such cases (10, 25, 37). In contrast, the only essential function of the E2 in the sequential model is to transfer ubiquitin to the E3, and there would be no obvious reason to expect that the E2 would influence downstream steps, including the type of chain that would be formed by the E3. The simplest direct evidence against a requirement for an indexation mechanism is that monoubiquitination precedes polyubiquitination in substrate ubiquitination reactions and that conversion of a substrate to a ubiquitinated form occurs with similar kinetics in the presence of K(0) ubiquitin or wild-type ubiquitin. Finally, while yeast Ubc1, -4, and -5 were the only E2s found to function with Rsp5, as predicted from earlier work on HECT-E2 binding (14, 32), it is not yet known whether these E2s are unique in being able to cooperate catalytically with the HECT domain or whether other ubiquitin E2s might also support HECT E3 activity if they were able to physically engage the E2 binding site of the HECT domain (36).

The results presented here highlight the flexibility of Rsp5 in recognizing multiple modification sites of target proteins. Rsp5 ubiquitinated multiple internal lysines of Wbp2 and Sna3, as well as the terminal amino group of both proteins. While it is not known whether Rsp5 ubiquitinates any of its physiologic substrates at the terminal amino group in vivo, our in vitro results suggest that caution should be used in concluding that lysineless Rsp5-associated proteins are not subject to ubiquitination (54). Targeting of multiple sites on target proteins suggests that, in addition to the flexibility of domain architecture within the HECT domain, there may be a large degree of flexibility in the linkage between the HECT domain and the rest of the protein so that many different surfaces of the substrate are presented to the active site of the E3. Perhaps consistent with this has been the failure so far to obtain structures for any full-length HECT E3s. Interestingly, the protein used for determining the HECT domain structure of Smurf2 contained two WW domains; however, these were disordered in the structure (33). The flexibility between the N and C lobes of the HECT domain itself might also promote recognition of alternative modification sites by presenting the reactive ubiquitin-thioester bond in a wide variety of positions in three-dimensional space.

The analyses of chimeric and truncated HECT E3s presented here show that chain type specificity is a function of the C lobe of the HECT domain. More specifically, chain type specificity maps to a region encompassing the last ~60 amino acids of the C lobe. According to the known HECT domain structures, this region contains three β-strands (S8 to S10 in the E6AP HECT domain structure), followed by the terminal α-helix (H14) and four to six unstructured residues (Fig. 9A and B). The active-site cysteine is located in the short loop between S9 and S10. Substitution of the entire ~60-amino-acid region from E6AP or Huwe1 into Rsp5 altered the chain type specificity to those of E6AP and Huwe1, respectively. The most obvious difference between Rsp5 and E6AP within this region is the presence of an extra 7-amino-acid loop in Rsp5, preceding S8 (Fig. (Fig.9).9). This loop is present in all human WW-HECT E3s, as well as Huwe1, where it is consists of 9 residues. However, deletion of this loop from Rsp5 did not alter its chain type specificity, nor did insertion of the Rsp5 loop sequence into chimera D, which contained the E6AP C lobe (Fig. (Fig.6),6), alter its chain type specificity (not shown). Therefore, this most obvious distinguishing feature between the Rsp5 and E6AP HECT domain C lobes is not a determinant of chain type specificity. Additionally, Rsp5 proteins containing the E6AP S8-to-S10 region alone, or the H14 region alone, did not switch the chain type specificity of Rsp5 (not shown). Thus, determinants from both the β-strand region and the terminal α-helix appear to function together to control the chain type specificity of Rsp5.

FIG. 9.
(A) Structures of the HECT domain C lobes of E6AP (19) and WWP1 (50). The structural elements of the last ~60 amino acids of E6AP are colored, as are the corresponding regions of WWP1 (β-strands, blue; α-helix, red; active-site ...

We have shown previously that suppression of the temperature-sensitive phenotype of an rsp5-1 mutant is dependent on the ability to form K63 chains (24); however, none of the Rsp5-based chimeras, including those that catalyzed primarily K63 polyubiquitination in vitro, complemented the rsp5-1 temperature-sensitive mutant in vivo (not shown). Because all of the proteins analyzed were expressed at levels similar to that of wild-type Rsp5 (not shown), the simplest explanation for the inability of the Rsp5/Nedd4 and Rsp5/Itch chimeras to complement the rps5-1 phenotype is that they were not fully active in vivo. Alternatively, they might not have been able to productively interact in vivo with certain substrates, substrate adaptors (22, 26, 28, 31, 35), or regulators (23, 24) of Rsp5. These possibilities are currently being explored.

How does the C lobe influence the chain type specificity of Rsp5? As discussed above, in the sequential-addition model for polyubiquitination, the distal end of the polyubiquitin chain must be oriented toward the active site of the E3 for further reactivity. A noncovalent ubiquitin binding site on the HECT domain not only might facilitate the polyubiquitination reaction, but could also specify the chain type by determining which lysine of the distal ubiquitin is able to approach the active-site thioester of the E3. Such a binding site on the C lobe might differ between an Rsp5-like E3, which might orient K63 of the distal ubiquitin toward the active site, and an E6AP-like E3, which might orient K48 toward the active site. Alternatively the relative positioning of the C lobe might determine which lysine of a noncovalently bound ubiquitin molecule, perhaps bound to the N lobe, could access the active-site cysteine. In this model, the last 60 amino acids of the C lobe domain would be predicted to determine the positioning of the C lobe, and hence, the chain type specificity of the E3. We have so far not detected a noncovalent ubiquitin binding site on the C lobe of the HECT domain of either Rsp5 or E6AP, although we imagine that the affinity of this interaction would be quite low. Two other studies have identified ubiquitin binding sites either within the N lobe of the HECT domain of Rsp5 (15) or directly upstream of the HECT domain of KIAA10 (53). While both studies indicated that these sites facilitated ubiquitination, there was no evidence that the binding sites specifically affected the chain type. Finally, in a mechanism that is not dependent on the presumption of a binding site for the distal ubiquitin of the chain, the C lobe might influence the chain type specificity by determining how the thioester-bound ubiquitin molecule is oriented on the E3, so that only specific lysines of the distal ubiquitin molecule can access the active site (Fig. (Fig.9C).9C). According to this model, Rsp5-like E3s would orient the thioester-bound ubiquitin so that only K63 of the incoming ubiquitin molecule could approach the active site, while in E6AP-like E3s, the orientation would allow only K48 to approach the active site. This is our favored model, since is it consistent with our previous interpretation of results showing that the conserved phenylalanine, located most commonly 4 amino acids from the ends of almost all HECT E3s, is critical for substrate ubiquitination (Fig. (Fig.9B).9B). Alteration of the −4 F residue does not affect the transfer of ubiquitin from the E2 to the E3 but nearly completely blocks the transfer of ubiquitin from the E3 to the substrate. We proposed that the C-terminal portion of the HECT domain, and the −4 F residue in particular, was critical for orienting the thioester-bound ubiquitin molecule to provide access of the attacking amino group to the active-site cysteine (39). Therefore, both our current and previous results may point to a function of the C lobe in orienting the thioester-bound ubiquitin molecule.

In summary, we have shown here that the chain type specificity of HECT ubiquitin ligases is a function of the C lobe of the HECT domain, that the identity of the cooperating E2 does not influence the chain type specificity, and that Rsp5 can synthesize polyubiquitin chains containing at least 12 and perhaps up to 30 ubiquitin moieties. While aspects of the apparently highly dynamic process of polyubiquitination are still subject to speculation, the current available evidence, including that presented here, most strongly supports the simple sequential-addition model for polyubiquitination by HECT E3s. Structural studies may ultimately be required to determine how the terminal 60 amino acids of the C lobe influence the chain type specificity of HECT ubiquitin ligases.

Acknowledgments

We thank members of the Huibregtse laboratory for helpful discussions and comments.

This work was supported by Public Health Service grant CA072943 from the National Cancer Institute and by the Institute for Cellular and Molecular Biology at the University of Texas.

Footnotes

[down-pointing small open triangle]Published ahead of print on 13 April 2009.

REFERENCES

1. Adhikary, S., F. Marinoni, A. Hock, E. Hulleman, N. Popov, R. Beier, S. Bernard, M. Quarto, M. Capra, S. Goettig, U. Kogel, M. Scheffner, K. Helin, and M. Eilers. 2005. The ubiquitin ligase HectH9 regulates transcriptional activation by Myc and is essential for tumor cell proliferation. Cell 123409-421. [PubMed]
2. Beaudenon, S., and J. M. Huibregtse. 2008. HPV E6, E6AP and cervical cancer. BMC Biochem. 9(Suppl. 1)S4. [PMC free article] [PubMed]
3. Beaudenon, S. L., M. R. Huacani, G. Wang, D. P. McDonnell, and J. M. Huibregtse. 1999. Rsp5 ubiquitin-protein ligase mediates DNA damage-induced degradation of the large subunit of RNA polymerase II in Saccharomyces cerevisiae. Mol. Cell. Biol. 196972-6979. [PMC free article] [PubMed]
4. Bernassola, F., M. Karin, A. Ciechanover, and G. Melino. 2008. The HECT family of E3 ubiquitin ligases: multiple players in cancer development. Cancer Cell 1410-21. [PubMed]
5. Bothos, J., M. K. Summers, M. Venere, D. M. Scolnick, and T. D. Halazonetis. 2003. The Chfr mitotic checkpoint protein functions with Ubc13-Mms2 to form Lys63-linked polyubiquitin chains. Oncogene 227101-7107. [PubMed]
6. Cadwell, K., and L. Coscoy. 2005. Ubiquitination on nonlysine residues by a viral E3 ubiquitin ligase. Science 309127-130. [PubMed]
7. Chastagner, P., A. Israel, and C. Brou. 2006. Itch/AIP4 mediates Deltex degradation through the formation of K29-linked polyubiquitin chains. EMBO Rep. 71147-1153. [PMC free article] [PubMed]
8. Chen, H. I., A. Einbond, S. J. Kwak, H. Linn, E. Koepf, S. Peterson, J. W. Kelly, and M. Sudol. 1997. Characterization of the WW domain of human yes-associated protein and its polyproline-containing ligands. J. Biol. Chem. 27217070-17077. [PubMed]
9. Cho, Y., S. Gorina, P. D. Jeffrey, and N. P. Pavletich. 1994. Crystal structure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations. Science 265346-354. [PubMed]
10. Christensen, D. E., P. S. Brzovic, and R. E. Klevit. 2007. E2-BRCA1 RING interactions dictate synthesis of mono- or specific polyubiquitin chain linkages. Nat. Struct. Mol. Biol. 14941-948. [PubMed]
11. Ciechanover, A., and R. Ben-Saadon. 2004. N-terminal ubiquitination: more protein substrates join in. Trends Cell Biol. 14104-106. [PubMed]
12. Deng, L., C. Wang, E. Spencer, L. Yang, A. Braun, J. You, C. Slaughter, C. Pickart, and Z. J. Chen. 2000. Activation of the IκB kinase complex by TRAF6 requires a dimeric ubiquitin-conjugating enzyme complex and a unique polyubiquitin chain. Cell 103351-361. [PubMed]
13. Eletr, Z. M., D. T. Huang, D. M. Duda, B. A. Schulman, and B. Kuhlman. 2005. E2 conjugating enzymes must disengage from their E1 enzymes before E3-dependent ubiquitin and ubiquitin-like transfer. Nat. Struct. Mol. Biol. 12933-934. [PubMed]
14. Eletr, Z. M., and B. Kuhlman. 2007. Sequence determinants of E2-E6AP binding affinity and specificity. J. Mol. Biol. 369419-428. [PMC free article] [PubMed]
15. French, M. E., B. R. Kretzmann, and L. Hicke. 2009. Regulation of the RSP5 ubiquitin ligase by an intrinsic ubiquitin-binding site. J. Biol. Chem. doi:.10.1074/jbc.M901106200 [PMC free article] [PubMed] [Cross Ref]
16. Haas, A. L., and K. D. Wilkinson. 2008. DeTEKting ubiquitination of APC/C substrates. Cell 133570-572. [PubMed]
17. Hochstrasser, M. 2006. Lingering mysteries of ubiquitin-chain assembly. Cell 12427-34. [PubMed]
18. Horak, J., and D. H. Wolf. 2001. Glucose-induced monoubiquitination of the Saccharomyces cerevisiae galactose transporter is sufficient to signal its internalization. J. Bacteriol. 1833083-3088. [PMC free article] [PubMed]
19. Huang, L., E. Kinnucan, G. Wang, S. Beaudenon, P. M. Howley, J. M. Huibregtse, and N. P. Pavletich. 1999. Structure of an E6AP-UbcH7 complex: insights into ubiquitination by the E2-E3 enzyme cascade. Science 2861321-1326. [PubMed]
20. Ikeda, F., and I. Dikic. 2008. Atypical ubiquitin chains: new molecular signals. ‘Protein modifications: beyond the usual suspects’ review series. EMBO Rep. 9536-542. [PMC free article] [PubMed]
21. Ingham, R. J., G. Gish, and T. Pawson. 2004. The Nedd4 family of E3 ubiquitin ligases: functional diversity within a common modular architecture. Oncogene 231972-1984. [PubMed]
22. Kee, Y., and J. M. Huibregtse. 2007. Regulation of catalytic activities of HECT ubiquitin ligases. Biochem. Biophys. Res. Commun. 354329-333. [PMC free article] [PubMed]
23. Kee, Y., N. Lyon, and J. M. Huibregtse. 2005. The Rsp5 ubiquitin ligase is coupled to and antagonized by the Ubp2 deubiquitinating enzyme. EMBO J. 242414-2424. [PMC free article] [PubMed]
24. Kee, Y., W. Munoz, N. Lyon, and J. M. Huibregtse. 2006. The deubiquitinating enzyme Ubp2 modulates Rsp5-dependent K63-linked polyubiquitin conjugates in Saccharomyces cerevisiae. J. Biol. Chem. 28136724-36731. [PubMed]
25. Kim, H. T., K. P. Kim, F. Lledias, A. F. Kisselev, K. M. Scaglione, D. Skowyra, S. P. Gygi, and A. L. Goldberg. 2007. Certain pairs of ubiquitin-conjugating enzymes (E2s) and ubiquitin-protein ligases (E3s) synthesize nondegradable forked ubiquitin chains containing all possible isopeptide linkages. J. Biol. Chem. 28217375-17386. [PubMed]
26. Leon, S., Z. Erpapazoglou, and R. Haguenauer-Tsapis. 2008. Ear1p and Ssh4p are new adaptors of the ubiquitin ligase Rsp5p for cargo ubiquitylation and sorting at multivesicular bodies. Mol. Biol. Cell 192379-2388. [PMC free article] [PubMed]
27. Li, W., D. Tu, A. T. Brunger, and Y. Ye. 2007. A ubiquitin ligase transfers preformed polyubiquitin chains from a conjugating enzyme to a substrate. Nature 446333-337. [PubMed]
28. Lin, C. H., J. A. MacGurn, T. Chu, C. J. Stefan, and S. D. Emr. 2008. Arrestin-related ubiquitin-ligase adaptors regulate endocytosis and protein turnover at the cell surface. Cell 135714-725. [PubMed]
29. Medintz, I., H. Jiang, and C. A. Michels. 1998. The role of ubiquitin conjugation in glucose-induced proteolysis of Saccharomyces maltose permease. J. Biol. Chem. 27334454-34462. [PubMed]
30. Merkley, N., K. R. Barber, and G. S. Shaw. 2005. Ubiquitin manipulation by an E2 conjugating enzyme using a novel covalent intermediate. J. Biol. Chem. 28031732-31738. [PubMed]
31. Nikko, E., J. A. Sullivan, and H. R. Pelham. 2008. Arrestin-like proteins mediate ubiquitination and endocytosis of the yeast metal transporter Smf1. EMBO Rep. 91216-1221. [PMC free article] [PubMed]
32. Nuber, U., and M. Scheffner. 1999. Identification of determinants in E2 ubiquitin-conjugating enzymes required for hect E3 ubiquitin-protein ligase interaction. J. Biol. Chem. 2747576-7582. [PubMed]
33. Ogunjimi, A. A., D. J. Briant, N. Pece-Barbara, C. Le Roy, G. M. Di Guglielmo, P. Kavsak, R. K. Rasmussen, B. T. Seet, F. Sicheri, and J. L. Wrana. 2005. Regulation of Smurf2 ubiquitin ligase activity by anchoring the E2 to the HECT domain. Mol. Cell 19297-308. [PubMed]
34. Peng, J., D. Schwartz, J. E. Elias, C. C. Thoreen, D. Cheng, G. Marsischky, J. Roelofs, D. Finley, and S. P. Gygi. 2003. A proteomics approach to understanding protein ubiquitination. Nat. Biotechnol. 21921-926. [PubMed]
35. Polo, S., and P. P. Di Fiore. 2008. Finding the right partner: science or ART? Cell 135590-592. [PubMed]
36. Ptak, C., C. Gwozd, J. T. Huzil, T. J. Gwozd, G. Garen, and M. J. Ellison. 2001. Creation of a pluripotent ubiquitin-conjugating enzyme. Mol. Cell. Biol. 216537-6548. [PMC free article] [PubMed]
37. Rodrigo-Brenni, M. C., and D. O. Morgan. 2007. Sequential E2s drive polyubiquitin chain assembly on APC targets. Cell 130127-139. [PubMed]
38. Saeki, Y., T. Kudo, T. Sone, Y. Kikuchi, H. Yokosawa, A. Toh-e, and K. Tanaka. 2009. Lysine 63-linked polyubiquitin chain may serve as a targeting signal for the 26S proteasome. EMBO J. 28359-371. [PMC free article] [PubMed]
39. Salvat, C., G. Wang, A. Dastur, N. Lyon, and J. M. Huibregtse. 2004. The −4 phenylalanine is required for substrate ubiquitination catalyzed by HECT ubiquitin ligases. J. Biol. Chem. 27918935-18943. [PubMed]
40. Scheffner, M., U. Nuber, and J. M. Huibregtse. 1995. Protein ubiquitination involving an E1-E2-E3 enzyme ubiquitin thioester cascade. Nature 37381-83. [PubMed]
41. Scheffner, M., and O. Staub. 2007. HECT E3s and human disease. BMC Biochem. 8(Suppl. 1)S6. [PMC free article] [PubMed]
42. Scheffner, M., T. Takahashi, J. M. Huibregtse, J. D. Minna, and P. M. Howley. 1992. Interaction of the human papillomavirus type 16 oncoprotein with wild-type and mutant p53 proteins. J. Virol. 665100-5105. [PMC free article] [PubMed]
43. Schwarz, S. E., J. L. Rosa, and M. Scheffner. 1998. Characterization of human hect domain family members and their interaction with UbcH5 and UbcH7. J. Biol. Chem. 27312148-12154. [PubMed]
44. Scialpi, F., M. Malatesta, A. Peschiaroli, M. Rossi, G. Melino, and F. Bernassola. 2008. Itch self-polyubiquitylation occurs through lysine-63 linkages. Biochem. Pharmacol. 761515-1521. [PubMed]
45. Stawiecka-Mirota, M., W. Pokrzywa, J. Morvan, T. Zoladek, R. Haguenauer-Tsapis, D. Urban-Grimal, and P. Morsomme. 2007. Targeting of Sna3p to the endosomal pathway depends on its interaction with Rsp5p and multivesicular body sorting on its ubiquitylation. Traffic 81280-1296. [PMC free article] [PubMed]
46. Tsui, C., A. Raguraj, and C. M. Pickart. 2005. Ubiquitin binding site of the ubiquitin E2 variant (UEV) protein Mms2 is required for DNA damage tolerance in the yeast RAD6 pathway. J. Biol. Chem. 28019829-19835. [PubMed]
47. Ulrich, H. D., and S. Jentsch. 2000. Two RING finger proteins mediate cooperation between ubiquitin-conjugating enzymes in DNA repair. EMBO J. 193388-3397. [PMC free article] [PubMed]
48. Unk, I., I. Hajdu, K. Fatyol, B. Szakal, A. Blastyak, V. Bermudez, J. Hurwitz, L. Prakash, S. Prakash, and L. Haracska. 2006. Human SHPRH is a ubiquitin ligase for Mms2-Ubc13-dependent polyubiquitylation of proliferating cell nuclear antigen. Proc. Natl. Acad. Sci. USA 10318107-18112. [PMC free article] [PubMed]
49. VanDemark, A. P., R. M. Hofmann, C. Tsui, C. M. Pickart, and C. Wolberger. 2001. Molecular insights into polyubiquitin chain assembly: crystal structure of the Mms2/Ubc13 heterodimer. Cell 105711-720. [PubMed]
50. Verdecia, M. A., C. A. Joazeiro, N. J. Wells, J. L. Ferrer, M. E. Bowman, T. Hunter, and J. P. Noel. 2003. Conformational flexibility underlies ubiquitin ligation mediated by the WWP1 HECT domain E3 ligase. Mol. Cell 11249-259. [PubMed]
51. Wang, G., J. Yang, and J. M. Huibregtse. 1999. Functional domains of the Rsp5 ubiquitin-protein ligase. Mol. Cell. Biol. 19342-352. [PMC free article] [PubMed]
52. Wang, M., D. Cheng, J. Peng, and C. M. Pickart. 2006. Molecular determinants of polyubiquitin linkage selection by an HECT ubiquitin ligase. EMBO J. 251710-1719. [PMC free article] [PubMed]
53. Wang, M., and C. M. Pickart. 2005. Different HECT domain ubiquitin ligases employ distinct mechanisms of polyubiquitin chain synthesis. EMBO J. 244324-4333. [PMC free article] [PubMed]
54. Watson, H., and J. S. Bonifacino. 2007. Direct binding to Rsp5p regulates ubiquitination-independent vacuolar transport of Sna3p. Mol. Biol. Cell 181781-1789. [PMC free article] [PubMed]
55. Zhao, X., J. I. Heng, D. Guardavaccaro, R. Jiang, M. Pagano, F. Guillemot, A. Iavarone, and A. Lasorella. 2008. The HECT-domain ubiquitin ligase Huwe1 controls neural differentiation and proliferation by destabilizing the N-Myc oncoprotein. Nat. Cell Biol. 10643-653. [PMC free article] [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...