Mark Hochstrasser

Ubiquitin, proteasomes, and the regulation of intracellular protein degradation.

Rapid degradation of specific proteins by ubiquitin/proteaseome-dependent pathways is a component of many cellular regulatory mechanisms. Recent work has shown that protein ubiquitination and deubiquitination are both mediated by large families of enzymes and that proteolysis can be modulated by alterations of the proteasome itself. The complexity of the ubiquitin system is reflected in the broad range of processes it regulates; these include key steps in cell cycle progression, processing of foreign proteins for presentation by class I MHC molecules, and the control of cell proliferation.

Substrate Targeting in the Ubiquitin System

As we have seen from the examples discussed above, the recognition mechanisms of many ubiquitin system substrates may be keyed to a fairly limited set of substrate features. Proteins whose degradation rates must be tightly coupled to environmental status, developmental state, or cell cycle stage are often phosphorylated via specific signal transduction cascades. These phosphorylated substrates are targets for particular variants of a large family of multisubunit ubiquitin–protein ligases, the SCF complexes. Interestingly, some of these substrates contain a rather short, closely related phosphopeptide motif.A very different type of degradation determinant, exemplified by the yeast α2 repressor, is a solvent–exposed hydrophobic protein surface. Such exposure usually only occurs in nonnative proteins or in the absence of a protein partner that contacts and buries the surface. Modulation of protein degradation rates by regulated folding or protein–protein interaction will in general not have the temporal precision of a transiently activated phosphorylation cascade, so the recognition of hydrophobic determinants may be utilized primarily for protein quality control and for bringing about changes in proteolytic rate that occur on somewhat longer time scales. It also seems particularly well-suited for ER-associated proteolysis, for the reasons discussed earlier.Notwithstanding these common themes, other types of protein motifs are also clearly recognized by the ubiquitin system. The first described E3 recognition determinant is a surprisingly simple one: the amino acid at the protein N terminus (Varshavsky 1997xVarshavsky, A. Genes Cells. 1997; 2: 13–28Crossref | PubMedSee all ReferencesVarshavsky 1997). In certain proteins, specific N-terminal residues stimulate polyubiquitination and rapid degradation in vivo. A different ubiquitination determinant has been found in the large subunit of yeast RNA polymerase II (Rpb1). Rpb1 is recognized through its conserved carboxy-terminal domain (CTD) by the Rsp5 ubiquitin ligase; this domain is both necessary and sufficient for Rsp5 binding and ubiquitination. The CTD contains the heptapeptide repeat sequence SPTSPSY, which is likely to be recognized by the WW domains of Rsp5. These WW domains, named for their two absolutely conserved tryptophan residues, are protein–protein interaction modules that bind directly to proline-rich sequences or phosphoserine- and phosphothreonine-containing elements in their targets (13xLu, P.J., Zhou, X.Z., Shen, M., and Lu, K.P. Science. 1999; 283: 1325–1328Crossref | PubMed | Scopus (471)See all References, 22xWang, G., Yang, J., and Huibregtse, J.M. Mol. Cell Biol. 1999; 19: 342–352Crossref | PubMedSee all References and references therein). Interestingly, a similar interaction occurs between the human epithelial sodium channel (ENaC) and Nedd4, a mammalian homolog of Rsp5. ENaC, whose impaired degradation in the heritable disorder Liddle’s syndrome is believed to cause hypertension, bears several proline-rich motifs that bind directly to the WW domains of Nedd4 (Goulet et al. 1998xGoulet, C.C., Volk, K.A., Adams, C.M., Prince, L.S., Stokes, J.B., and Snyder, P.M. J. Biol. Chem. 1998; 273: 30012–30017Crossref | PubMed | Scopus (141)See all ReferencesGoulet et al. 1998). These examples reinforce the notion that substrate discrimination can rely on very short sequence motifs, like those found in IκBα and β-catenin.Thus, a few general principles of substrate recognition in the ubiquitin system are at last beginning to emerge, and these should be useful guides for the challenges that lie ahead. These include questions about the details of the E3-substrate interaction and how a substrate can then become polyubiquitinated. Other central issues are how the activity of an E3 is modulated and what controls E2–E3 interactions. Finally, it is known that ubiquitin chains of different lengths and distinct ubiquitin–ubiquitin linkages can form on substrates, so investigation of what regulates the length and topology of these chains will certainly be of interest. There is little doubt that answers to many of these questions will not be long in coming, given the current pace of work in the field.*To whom correspondence should be addressed (e-mail: hoc1@midway.uchicago.edu).

Autocatalytic Subunit Processing Couples Active Site Formation in the 20S Proteasome to Completion of Assembly

The eukaryotic 20S proteasome is responsible for the degradation of many cellular proteins, but how it is assembled and how its distinct active sites are formed are not understood. Like other proteasome subunits, the yeast Doa3 protein is synthesized in precursor form. We show that the N-terminal propeptide is required for Doa3 incorporation into the proteasome and, remarkably, that the propeptide functions in trans, suggesting it serves a chaperone-like function in proteasome biogenesis. Propeptide processing is not required for proteasome assembly but is needed for maturation of a specific subset of active sites. The likely nucleophile for these sites is provided by the N-terminal threonine of mature Doa3. Additional data indicate that precursor processing is autocatalytic and requires association of the two halves of the proteasome particle, thereby preventing formation of proteolytic sites until the central hydrolytic chamber has been sealed off from the rest of the cell.

Multiple ubiquitin-conjugating enzymes participate in the in vivo degradation of the yeast MATα2 repressor

Attachment of ubiquitin to proteins is catalyzed by a family of ubiquitin-conjugating (UBC) enzymes. Although these enzymes are essential for many cellular processes; their molecular functions remain unclear because no physiological target has been identified for any of them. Here we show that four UBC proteins (UBC4, UBC5, UBC6, and UBC7) target the yeast MAT alpha 2 transcriptional regulator for intracellular degradation by two distinct ubiquitination pathways. UBC6 and UBC7 define one of the pathways and can physically associate. The UBC6/UBC7-containing complex targets the Deg1 degradation signal of alpha 2, a conclusion underscored by the finding that UBC6 is encoded by DOA2, a gene previously implicated in Deg1-mediated degradation. These data reveal an unexpected overlap in substrate specificity among diverse UBC enzymes and suggest a combinatorial mechanism of substrate selection in which UBC enzymes partition into multiple ubiquitination complexes.

The Yeast ULP2 (SMT4) Gene Encodes a Novel Protease Specific for the Ubiquitin-Like Smt3 Protein

ABSTRACT Yeast Smt3 and its vertebrate homolog SUMO-1 are ubiquitin-like proteins (Ubls) that are reversibly ligated to other proteins. LikeSMT3, SMT4 was first isolated as a high-copy-number suppressor of a defective centromere-binding protein. We show here that SMT4 encodes an Smt3-deconjugating enzyme, Ulp2. In cells lacking Ulp2, specific Smt3-protein conjugates accumulate, and the conjugate pattern is distinct from that observed in a ulp1ts strain, which is defective for a distantly related Smt3-specific protease, Ulp1. The ulp2Δ mutant exhibits a pleiotropic phenotype that includes temperature-sensitive growth, abnormal cell morphology, decreased plasmid and chromosome stability, and a severe sporulation defect. The mutant is also hypersensitive to DNA-damaging agents, hydroxyurea, and benomyl. Although cell cycle checkpoint arrest in response to DNA damage, replication inhibition, or spindle defects occurs with normal kinetics, recovery from arrest is impaired. Surprisingly, either introduction of aulp1ts mutation or overproduction of catalytically inactive Ulp1 can substantially overcome theulp2Δ defects. Inactivation of Ulp2 also suppresses several ulp1ts defects, and the double mutant accumulates far fewer Smt3-protein conjugates than either single mutant. Our data suggest the existence of a feedback mechanism that limits Smt3-protein ligation when Smt3 deconjugation by both Ulp1 and Ulp2 is compromised, allowing a partial recovery of cell function.

Protein Degradation or Regulation: Ub the Judge

The experimental setting for the work of Hicke and Riezman 1996xHicke, L. and Riezman, H. Cell. 1996; 84: 277–287Abstract | Full Text | Full Text PDF | PubMed | Scopus (594)See all ReferencesHicke and Riezman 1996 is very different from that of Chen et al. 1996xChen, Z.I., Parent, L., and Maniatis, T. Cell. 1996; 84See all ReferencesChen et al. 1996, but the inference of a nonproteolytic role for ubiquitin is equally compelling. Many cell surface proteins are rapidly endocytosed and, if not recycled to the plasma membrane, are routed to the lysosome/vacuole, where they are destroyed by vacuolar proteases. Endocytosis of the S. cerevisiae α-factor pheromone receptor, Ste2, takes place constitutively at a slow rate but is stimulated 5-10 fold by α-factor binding. Pheromone binding to the G protein-coupled Ste2 receptor activates the MAP kinase signal transduction pathway, resulting ultimately in alterations in transcription needed by the cell to mate with α-factor secreting cells. Internalized Ste2 follows the endocytic pathway to the vacuole, where it is degraded. For endocytosis, the cytoplasmic C-terminal tail of the polytopic Ste2 protein is critical; within this tail, a nine-residue sequence, SINNDAKSS, is sufficient for receptor internalization (seeHicke and Riezman 1996xHicke, L. and Riezman, H. Cell. 1996; 84: 277–287Abstract | Full Text | Full Text PDF | PubMed | Scopus (594)See all ReferencesHicke and Riezman 1996, for references).In their new work, Hicke and Riezman have examined more closely the requirements for receptor endocytosis. As shown previously, exposure of cells to α-factor led to rapid disappearance of the receptor. Degradation was strongly inhibited in a yeast end4 mutant, which is defective at an early step in receptor endocytosis. Based on SDS-polyacrylamide gel separations, however, Ste2 did not accumulate in its normal form in the mutant cells but rather, as a set of higher molecular mass species, the most prominent of which differed in apparent molecular mass by intervals of ∼9 kD. The higher molecular mass forms of Ste2 are also detectable as transient intermediates in pheromone-stimulated wild-type cells. These results immediately suggested that Ste2 might be modified by ubiquitin, a suspicion confirmed by immunoprecipitation with an antibody to ubiquitin. Furthermore, internalization of ligand-bound receptors was markedly inhibited in yeast cells lacking members of the Ubc1/Ubc4/Ubc5 subfamily of E2 enzymes. Ubiquitination of Ste2 is correspondingly reduced in ubc4 ubc5 mutants. Some of these mutated strains are quite sick, but the defect in Ste2 endocytosis could not be explained by a general defect in membrane trafficking.All this is well and good, but it does not prove ubiquitination of Ste2 is what triggers its endocytosis. For that matter, a very similar story has already been told about yeast Ste6, an ATP-binding cassette transporter protein involved in a-factor export (Kolling and Hollenberg 1994xKolling, R. and Hollenberg, C.P. EMBO J. 1994; 13: 3261–3271PubMedSee all ReferencesKolling and Hollenberg 1994). However, Hicke and Riezman went on to analyze the effect of mutations in the SINNDAKSS internalization signal on Ste2 ubiquitination, and the results were telling. Mutation of the lysine in this element to arginine (in a truncated derivative of Ste2) almost completely blocked endocytosis. The mutation also eliminated Ste2 ubiquitination. There is one other lysine in the cytoplasmic tail of this derivative of Ste2. Changing it to arginine was of no consequence for either receptor internalization or ubiquitination. Finally, mutation to alanine of all three serines in the endocytosis signal eliminated receptor hyperphosphorylation in response to α-factor and inhibited receptor internalization. Correspondingly, ubiquitination of Ste2 was no longer detectable. The simplest interpretation of these data is that endocytosis of Ste2 specifically requires ubiquitination of the lysine in the SINNDAKSS element and that this modification depends on phosphorylation of the flanking serines. Phosphorylation-dependent ubiquitination has been reported for a number of proteins in the last few years, one of the best characterized examples being IκBα, as discussed above.The Ste2 results raise some interesting questions. For one, how does ubiquitination of Ste2 lead to its endocytosis? The ubiquitin (chain?) may provide a binding site for a component of the endocytic machinery or may alter Ste2 structure in a way that facilitates such binding. As an alternative, the authors suggest ubiquitin may promote movement of Ste2 into regions of membrane that actively endocytose. Another question is why ubiquitinated Ste2 is not targeted by the proteasome, as demonstrated convincingly by Hicke and Riezman 1996xHicke, L. and Riezman, H. Cell. 1996; 84: 277–287Abstract | Full Text | Full Text PDF | PubMed | Scopus (594)See all ReferencesHicke and Riezman 1996 with proteasome mutants. It is not simply because Ste2 is a membrane protein. The CFTR transmembrane protein, particularly the mutant form commonly found in cystic fibrosis patients, is multiubiquitinated when in the ER and is apparently degraded by the proteasome (Ward et al. 1995xWard, C.L., Omura, S., and Kopito, R.R. Cell. 1995; 83: 121–127Abstract | Full Text PDF | PubMed | Scopus (976)See all ReferencesWard et al. 1995). Under some conditions, certain receptor tyrosine kinases can also be degraded by a ubiquitin/proteasome-dependent mechanism (Sepp-Lorenzino et al. 1995xSepp-Lorenzino, L., Ma, Z., Lebwohl, D.E., Vinitsky, A., and Rosen, N. J. Biol. Chem. 1995; 270: 16580–16587CrossRef | PubMed | Scopus (174)See all ReferencesSepp-Lorenzino et al. 1995). Structural features of the targeted protein may be important in determining its fate. It is also conceivable that ubiquitin chains on Ste2 rarely reach the lengths thought to be preferred by the 26S proteasome because either chain assembly is slow or chain disassembly by deubiquitinating enzymes is rapid. An intriguing possibility is that the type of ubiquitin chain formed on Ste2 determines its metabolic fate. Several groups have reported that ubiquitin chains with non-lysine 48 isopeptide linkages can form in yeast cells (1xArnason, T. and Ellison, M.J. Mol. Cell Biol. 1994; 14: 7876–7883PubMedSee all References, 13xSpence, J., Sadis, S., Haas, A.L., and Finley, D. Mol. Cell Biol. 1995; 15: 1265–1273CrossRef | PubMedSee all References, 9xJohnson, E.S., Ma, P.C.M., Ota, I.M., and Varshavsky, A. J. Biol. Chem. 1995; 270: 17442–17456CrossRef | PubMed | Scopus (550)See all References, 2xBaboshina, O.V. and Haas, A.L. J. Biol. Chem. 1996; 271PubMedSee all References). Interestingly, lysine 63-linked ubiquitin chains require Ubc4/Ubc5, the same E2 enzymes implicated in receptor endocytosis, and it had previously been argued that such chains serve nonproteolytic functions (Arnason and Ellison 1994xArnason, T. and Ellison, M.J. Mol. Cell Biol. 1994; 14: 7876–7883PubMedSee all ReferencesArnason and Ellison 1994). Such alternative chains could provide distinct binding surfaces that target the ubiquitin conjugate to different binding factors or cellular structures. These considerations are of course relevant not only for membrane protein endocytosis but also for other ubiquitin-dependent, proteasome-independent processes, such as the IκB kinase activation discussed above.A final question is how often ubiquitination is utilized as a signal for endocytosis and vacuolar targeting. Almost certainly, the answer will be that it is extremely common. As mentioned above, the yeast Ste6 protein is very likely to follow a degradation pathway very similar to that which works on Ste2. Other recent studies in yeast suggest the same is true for uracil permease, the general amino acid transporter Gap1, and the multidrug transporter Pdr5 (Egner and Kuchler 1996xEgner, R. and Kuchler, K. FEBS Lett. 1996; 378: 177–181Abstract | Full Text PDF | PubMed | Scopus (97)See all ReferencesEgner and Kuchler 1996references therein). Interestingly, uracil permease appears to use a ubiquitination signal related to the “destruction box” of mitotic cyclins. Internalization and vacuolar degradation of uracil permease and Gap1 require the Npi1/Rsp5 protein, which has characteristics of an E3 ubiquitin-protein ligase. These are not just peculiarities of yeast: a number of mammalian membrane proteins are also known to be multiubiquitinated at the cell surface and degraded in the lysosome (Ciechanover 1994xCiechanover, A. Cell. 1994; 79: 13–21Abstract | Full Text PDF | PubMed | Scopus (1316)See all ReferencesCiechanover 1994).At present, we have little more than a first adumbration of the complexities of the ubiquitin system. Many new physiological roles for this intricate metabolic system are still being discovered. At a very basic mechanistic level, it is still unclear how specific proteins are identified by the ubiquitin conjugation machinery or what the relative importance of the E2 and E3 enzymes is in substrate recognition. We are also just becoming aware of the remarkably large and diverse group of deubiquitinating enzymes that exist (yeast have at least 17 such enzymes); these enzymes are likely to be functionally diverse as well (Papa and Hochstrasser 1993xPapa, F. and Hochstrasser, M. Nature. 1993; 366: 313–319CrossRef | PubMed | Scopus (298)See all ReferencesPapa and Hochstrasser 1993). We now have two clear examples where ubiquitin operates in a way other than as a direct proteolytic signal. With the realization that ubiquitination does more than deliver proteins to the proteasome, questions about the molecular basis of substrate specificity and targeting are given new urgency. Combining mechanistic analysis of the ubiquitin system with the study of its varied biological functions should prove a productive enterprise for some years to come.

In vivo disassembly of free polyubiquitin chains by yeast Ubp14 modulates rates of protein degradation by the proteasome

Degradation of many eukaryotic proteins requires their prior ligation to polyubiquitin chains, which target substrates to the 26S proteasome, an abundant cellular protease. We describe a yeast deubiquitinating enzyme, Ubp14, that specifically disassembles unanchored (‘free’) ubiquitin chains in vitro, a specificity shared by mammalian isopeptidase T. Correspondingly, deletion of the UBP14 gene from yeast cells results in a striking accumulation of free ubiquitin chains, which correlates with defects in ubiquitin‐dependent proteolysis. Increasing the steady‐state levels of ubiquitin chains in wild‐type cells (by expressing a derivative of ubiquitin with an altered C‐terminus) inhibits protein degradation to a degree comparable with that observed in ubp14Δ cells. Inhibition of degradation is also seen when an active site mutant of Ubp14 is overproduced in vivo. Surprisingly, overproduction of wild‐type Ubp14 can inhibit degradation of some proteins as well. Finally, Ubp14 and human isopeptidase T are shown to be functional homologs by complementation analysis. We propose that Ubp14 and isopeptidase T facilitate proteolysis in vivo by preventing unanchored ubiquitin chains from competitively inhibiting polyubiquitin–substrate binding to the 26S proteasome.

Degradation Signal Masking by Heterodimerization of MATα2 and MATa1 Blocks Their Mutual Destruction by the Ubiquitin-Proteasome Pathway

Proteolysis by the ubiquitin-proteasome pathway is often regulated, but the mechanisms underlying such regulation remain ill-defined. In Saccharomyces cerevisiae, cell type is controlled by the MAT transcription factors. The alpha2 repressor is a known ubiquitin pathway substrate in alpha haploid cells. We show that a1 is rapidly degraded in a haploids. In a/alpha diploids, alpha2 and a1 are stabilized by heterodimerization. Association depends on N-terminal coiled-coil interactions between a1 and alpha2. Residues in alpha2 important for these interactions overlap a critical determinant of an alpha2 degradation signal, which we delimit by extensive mutagenesis. Our data provide a detailed description of a natural ubiquitin-dependent degradation signal and point to a molecular mechanism for regulated turnover in which proteolytic signals are differentially masked in alternative multiprotein complexes.

Eukaryotic 20S proteasome catalytic subunit propeptides prevent active site inactivation by N‐terminal acetylation and promote particle assembly

Proteins targeted for degradation by the ubiquitin‐proteasome system are degraded by the 26S proteasome. The core of this large protease is the 20S proteasome, a barrel‐shaped structure made of a stack of four heptameric rings. Of the 14 different subunits that make up the yeast 20S proteasome, three have proteolytic active sites: Doa3/β5, Pup1/β2 and Pre3/β1. Each of these subunits is synthesized with an N‐terminal propeptide that is autocatalytically cleaved during particle assembly. We show here that the propeptides have both common and distinct functions in proteasome biogenesis. Unlike the Doa3 propeptide, which is crucial for proteasome assembly, the Pre3 and Pup1 propeptides are dispensable for cell viability and proteasome formation. However, mutants lacking these propeptide‐encoding elements are defective for specific peptidase activities, are more sensitive to environmental stresses and have subtle defects in proteasome assembly. Unexpectedly, a critical function of the propeptide is the protection of the N‐terminal catalytic threonine residue against Nα‐acetylation. For all three propeptide‐deleted subunits, activity of the affected catalytic center is fully restored when the Nat1‐Ard1 Nα‐acetyltransferase is mutated. In addition to delineating a novel function for proteasome propeptides, these data provide the first biochemical evidence for the postulated participation of the α‐amino group in the proteasome catalytic mechanism.

Biogenesis, structure and function of the yeast 20S proteasome.

Intracellular degradation of many eukaryotic proteins requires their covalent ligation to ubiquitin. We previously identified a ubiquitin‐dependent degradation pathway in the yeast Saccharomyces cerevisiae, the DOA pathway. Independent work has suggested that a major mechanism of cellular proteolysis involves a large multisubunit protease(s) called the 20S proteasome. We demonstrate here that Doa3 and Doa5, two essential components of the DOA pathway, are subunits of the proteasome. Biochemical analyses of purified mutant proteasomes suggest functions for several conserved proteasome subunit residues. All detectable proteasome particles purified from doa3 or doa5 cells have altered physical properties; however, the mutant particles contain the same 14 different subunits as the wild‐type enzyme, indicating that most or all yeast 20S proteasomes comprise a uniform population of hetero‐oligomeric complexes rather than a mixture of particles of variable subunit composition. Unexpectedly, we found that the yeast Doa3 and Pre3 subunits are synthesized as precursors which are processed in a manner apparently identical to that of related mammalian proteasome subunits implicated in antigen presentation, suggesting that biogenesis of the proteasome particle is highly conserved between yeast and mammals.