40632_CH03_hochstrasser.qxd 3/26/07 11:45 AM Page 1 3 Intracellular degradation Mark Hochstrasser Yale University, New Haven, CT

CHAPTER OUTLINE

3.1 Introduction 3.8 Integral membrane that function at the cell 3.2 Overview of the ubiquitin- system surface are degraded within 3.3 Ubiquitin attachment to substrates requires multiple 3.9 Under stressful conditions cytoplasmic proteins can be engulfed into autophagosomes and degraded in the 3.4 Substrate recognition in the ubiquitin-ligation system 3.10 What’s next? 3.5 Degradation of proteins by the proteasome 3.11 Summary 3.6 Membrane proteins are degraded by several mechanisms 3.12 Supplement 1: Failures of the ubiquitin-proteasome 3.7 Retrotranslocation from the endoplasmic reticulum back system are a cause of human diseases into the allows degradation of short-lived ER proteins 3.13 Supplement 2: Ubiquitin has many functions beyond the proteasome

1 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 2

3.1 rapidly even without any change in the rate at Introduction which it is synthesized. For example, progres- Key concepts sion of the cell through the cell cycle requires • Different proteins in a cell have vastly different that certain inhibitors of mitosis be inactivated rates of degradation. The cell can regulate those once the chromosomes are aligned and are ready rates. to be separated to opposite poles. At that point • Regulatory proteins such as transcription factors the sensitivity of the inhibitors to proteolysis and cell cycle regulators often have high rates of suddenly increases and they quickly disappear, degradation to allow rapid changes in their con- centration. allowing the cell cycle to advance. • Misfolded or damaged proteins, which are poten- This example illustrates one of the advan- tially toxic, are also targets for efficient elimina- tages of protein degradation as a form of con- tion. trol: its irreversibility. Elimination of a protein • Most intracellular protein degradation is energy- removes any chance of its being reactivated in- dependent, highly specific, and processive. appropriately, which in the case of the cell cy- cle could be lethal to the cell. It is probably for Cells are densely packed with many different this reason that selective protein degradation kinds of proteins, and these proteins are in a is a common component of regulatory mecha- highly dynamic state, moving from place to nisms that determine the relative timing of place within the cell and associating and disso- events. For example, in the cell cycle, mitosis ciating with other proteins and ligands. In view must follow duplication of the chromosomes, of the energy that a cell expends synthesizing so certain factors required for DNA synthesis and organizing its proteins, it may be surpris- are degraded at the end of S phase and are only ing to learn that it is also constantly degrading synthesized again once mitosis has been com- many of them while they are still perfectly func- pleted. tional. Each has its own characteristic rate of In addition to degrading regulatory pro- degradation, or half-life. Most are normally teins, cells also degrade other types of proteins degraded slowly, with half-lives of many hours, while they are still functional. The purpose is days, or longer. Others, in contrast, have half- usually to remove proteins that the cell no longer lives that are much shorter than the doubling needs. This is often the case with proteins that time of the cell, in some cases being as short as help a cell adapt to a particular set of conditions; a few minutes. once the conditions cease to exist the proteins What purpose does proteolysis within a are degraded. Human muscles, for example, are cell serve, and why are the half-lives of some remodeled depending on how they are used proteins so short? One answer to both these and what energy source is available. Remodeling questions is that proteolysis is a very effective involves proteolysis of one set of structural and form of regulation. The rate at which a protein metabolic proteins and its replacement by an- is degraded has a profound effect on how quickly other specialized for a different purpose. its concentration can be changed following a Another function of intracellular proteol- change in its rate of synthesis. A protein that is ysis is to rid the cell of damaged or potentially degraded slowly will persist in the cell even if harmful proteins. Proteins can misfold or de- the cell stops making it. A rapidly degraded pro- nature, they may fail to assemble properly into tein, however, will quickly disappear after its complexes, or they can be altered by some ab- synthesis is stopped. Thus, rapid proteolysis of normal posttranslational modification. Such a protein allows rapid readjustment of its level aberrant proteins are potentially toxic and need within the cell. For this reason, proteins that to be eliminated. The eukaryotic cell has a re- may need to be activated or inactivated quickly markable ability to distinguish normal from ab- often have high rates of degradation, resulting normal proteins and selectively degrade the in short half-lives. This is true of many cellular latter. When this capacity is compromised, dis- regulatory proteins, such as transcription factors, ease often results. enzymes controlling rate-limiting steps in Several important properties characterize biosynthetic pathways, and cell cycle regula- the degradation of most intracellular proteins. tors. Perhaps most obvious is that it must be exquis- Regulation can also be achieved by abruptly itely specific. Only certain proteins should be changing the rate at which a protein is degraded. proteolyzed rapidly, and those often only under If a protein’s rate of degradation is suddenly in- very specific conditions. It is easy to imagine creased its concentration within the cell will fall the devastation that would result if a protease

2 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 3

that lacked such specificity were let loose within the trafficking of proteins into the lysosome. In a living cell. Proteases of the digestive system, this regard, the central question of how only for example, need to be able to cleave proteins particular proteins are degraded in the lyso- with high efficiency but low specificity, and it some is closely linked to the question of how spe- would be catastrophic if such a protease were cific proteins sort between different cellular released in the cytoplasm. A second feature of compartments. intracellular proteolysis is its dependence on the input of energy. Despite the fact that pep- tide bond cleavage is energetically favorable, 3.2 Overview of the ATP hydrolysis is necessary for efficient degra- dation of most proteins within the cell. This en- ubiquitin-proteasome ergy requirement derives largely from the need system for very high substrate specificity and can be understood in light of what we now know about Key concepts the molecular mechanisms of most cellular pro- • Cellular proteins to be degraded are first cova- teolytic pathways. These mechanisms will be lently linked to a polymer of ubiquitin polypep- the main topic of this chapter. tides; this polyubiquitin chain allows the substrate A final property that distinguishes most in- to bind to the proteasome, where the substrate protein is processively degraded to small peptides. tracellular proteolysis is that it is processive: • The ubiquitin-proteasome system performs most once a protein is selected for proteolysis, it is intracellular proteolysis. completely degraded to small peptides without larger intermediate forms being released. Processive degradation prevents the release Proteolysis of specific proteins in a cell is usu- of large protein fragments that might interfere ally performed by the ubiquitin-proteasome with cell function. For example, many tran- system. The enzymes comprising this system scription factors have their DNA-binding and can, when necessary, select and degrade almost transcriptional activation activities in separate any protein in the cell. Such proteolysis serves protein domains. If the DNA- a number of purposes. It orchestrates various cel- were proteolytically separated from the activat- lular events—such as critical cell cycle transi- ing domain, it would still be able to bind DNA tions—by eliminating regulatory proteins; it and might compete with the normal factor for removes proteins that the cell no longer needs; regulatory sites in the . and it provides an important quality control This chapter will focus on the major mech- function by continuously scanning the cell’s anisms of protein degradation in the eukary- collection of proteins and ridding it of those that otic cell. The majority of regulated degradation are damaged or abnormal. In all cases these en- occurs through the ubiquitin-proteasome sys- zymes act with great specificity, not only iden- tem, which functions in both the cytoplasm and tifying individual proteins from among the the nucleus. In order for a protein to be de- thousands in a cell, but often distinguishing dif- graded by this mechanism, it is first covalently ferent forms of the same protein depending on modified by the attachment of multiple copies factors such as subcellular localization, post- of a small polypeptide called ubiquitin. The translational modification, or inclusion in par- polyubiquitin-modified protein is recognized ticular multiprotein complexes. by an abundant, multisubunit protease called the The system is organized around ubiquitin, proteasome, which degrades the substrate into a small protein that is used as a label to desig- short peptides. The next section introduces the nate proteins for destruction. Ubiquitin is added ubiquitin-proteasome system in greater detail. to proteins in a tightly linked series of reactions Some proteolysis of cellular proteins also (FIGURE 3.1). A protein to be degraded is first rec- occurs in the lysosome, a membrane-enclosed ognized by a specific ubiquitin-protein lig- organelle. Protein degradation in the lysosome ase, also called an E3. The E3 protein directs the occurs by a completely different mechanism addition of ubiquitin onto the substrate. The cell than in the ubiquitin-proteasome system. The has many different E3s, each of which recog- interior of the lysosome contains an assortment nizes a specific structural feature of its targets, of hydrolytic enzymes, including many non- much as different antibodies recognize specific specific proteases. Proteins that are to be de- epitopes on their target proteins. As the factors graded are moved into the lysosome, where that first recognize a substrate, the E3s are re- they are exposed to the resident proteases. sponsible for much of the specificity of proteol- Substrate selectivity is achieved by controlling ysis by the ubiquitin-proteasome system.

3.2 Overview of the ubiquitin-proteasome system 3 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 4

The ubiquitin-proteasome pathway to proteins. Ubiquitin is a small protein of 76 RECOGNITION AND UBIQUITINATION UNFOLDING AND DEGRADATION amino acids that folds into a single, compact domain (FIGURE 3.2). It is one of the most highly Ubiquitin conserved proteins known, its sequence being 26S proteasome identical in species as distantly related as hu- ATP AMP + mans, frogs, and fruit flies, and differing only slightly in other species. Its name is derived from Ubiquitinating this ubiquity. enzymes Target ADP Degraded protein ATP protein Among the features of ubiquitin’s sequence are two glycine residues at its C-terminus. These FIGURE 3.1 Proteins destined for degradation are first covalently modified two amino acids do not fold as part of the pro- with one or more polymers of ubiquitin. Once polyubiquitinated, a target pro- tein can bind to the 26S proteasome, which unfolds the protein and translo- tein’s globular structure, instead forming a short, cates it into a central chamber containing protease active sites, where it is flexible tail extending from its surface (Figure degraded into short peptides. Ubiquitin is not degraded and is released for 3.2). This tail is the position at which ubiquitin reuse. is attached to substrate proteins. Linkage oc- curs between the carboxyl group of the C-ter- Ubiquitin molecules are often added to one minal glycine residue of ubiquitin and the another as well as to the substrate, resulting in -amino group of a lysine residue within the chains of ubiquitin extending from a protein tar- substrate, forming a type of amide bond often geted for degradation. Assembly of a polymeric referred to as an isopeptide bond (FIGURE 3.3). chain of at least four ubiquitins on the substrate Several lysine residues within a protein can be is needed for it to bind to the 26S proteasome, modified in this way at the same time, although a large complex of proteins in the cytoplasm that modification at multiple sites is generally not is a ubiquitin-dependent protease. Ubiquitinated necessary for subsequent proteolytic targeting. proteins often arrive at the proteasome still folded Ubiquitin is also often linked to a lysine residue and must be unfolded before they can be pro- in another ubiquitin molecule that is already teolyzed. Unfolding is performed by one of the attached to a protein (FIGURE 3.4). By the time two subcomplexes of the proteasome, the 19S it is finished being modified in this way, a sub- regulatory complex. In addition to unfolding strate protein may have a chain of as many as proteins, this complex removes ubiquitin from 10–30 ubiquitin molecules extending from its them before delivering them to the second sub- surface. Such polyubiquitination is required for complex, the 20S proteasome, where proces- a protein to be degraded. sive proteolysis all the way to small peptides In all known cases, ubiquitin is synthesized occurs. Processive proteolysis is possible because in a precursor form. It is translated as part of a of the proteasome’s unique architecture. Only in rare instances is a substrate protein only par- tially proteolyzed rather than being completely degraded.

Ubiquitin 3.3 Ubiquitin attachment to N substrates requires multiple enzymes

Key concepts • Ubiquitin is a small polypeptide that is covalently attached to proteins in order to designate them for destruction. • Energy is required to activate ubiquitin before it can be used to modify proteins. • Attaching ubiquitin to a substrate requires that Glycine several enzymes act in sequence. The E1 tail activates ubiquitin and transfers it to an E2; an E3 C then transfers ubiquitin from the E2 to the sub- strate. FIGURE 3.2 The structure of ubiquitin. Extending from a single small, compact domain is a flexible tail of two glycine Central to the process of selective proteolysis is residues where the molecule is attached the ubiquitin molecule and how it is attached to target proteins.

4 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 5

An isopeptide bond A polyubiquitin chain

NH N- 2 TERMINUS NH2 Ubiquitin

Isopeptide COOH Ubiquitin bond

Ubiquitin Isopeptide NH bond Glycine CH Glycine 2 (C-TERMINUS) NH Lysine

COOH CH2

Carboxyl group CO

ε-amino group NH NH 2 CH2 CH 2 CH2 Lysine CH2 CH (SIDE CHAIN) 2 CH SUBSTRATE 2 CH2 N- C- CH2 NH CH CO TERMINUS TERMINUS FIGURE 3.4 Ubiquitin molecules are often attached NH COOH 2 ... NH CH CO ... SUBSTRATE to one another via isopeptide bonds, forming a SUBSTRATE chain of ubiquitins extending from the surface of a target protein. FIGURE 3.3 Ubiquitin molecules are attached to substrate proteins via isopep- tide bonds. Such bonds form between the carboxyl group at the C-terminus of the ubiquitin molecule and the amino group at the end of a lysine residue within the sequence of the target protein.

fusion protein with one of several ribosomal proteins, or as part of a long protein containing Ubiquitin activation FIGURE 3.5 Ubiquitin is activated in multiple ubiquitin units linked in tandem. In a series of reactions in which it is trans- both cases processing is required to form func- Free ubiquitin ferred from one enzyme to another. The enzyme E1 first forms a covalent tional ubiquitin, a task performed by enzymes bond between itself and ubiquitin in that can also remove ubiquitin from the lysine E1 binds E1 COO- ubiquitin a reaction driven by the energy of ATP side chains of proteins. Why ubiquitin is always hydrolysis. The ubiquitin is then trans- synthesized as a precursor is unclear. SH ATP ferred to a cysteine side chain of an Ubiquitin must be activated before it can AMP + E2, forming a second thioester. The energy of the ATP hydrolysis is pre- be attached to a substrate. As with most biolog- served in the thioester bonds, creat- ical reactions in which two units are joined, ing an activated form of ubiquitin. such as the polymerization of amino acids or Transfer of E2-associated ubiquitin serves a pool nucleotides, the input of energy is required to CO E2 ubiquitin from of ubiquitin for transfer to proteins. S E1 to E2 make the reaction favorable. Activation of ubiq- HS uitin occurs at its C-terminus in an ATP- dependent reaction performed by ubiquitin- activating enzyme (usually referred to as E1). The first part of the reaction, which is analo- The activated ubiquitin is gous to the activation of amino acids for their CO ready for use conjugation to tRNA by aminoacyl-tRNA syn- S thetases, results in the C-terminal carboxyl group of ubiquitin being linked to AMP. After this initial activation, the ubiquitin is transferred to a cysteine side chain within the E1 enzyme itself, forming a covalent thioester bond with This covalent complex of an E1 enzyme the enzyme and preserving the energy of acti- and ubiquitin serves as a pool of activated ubiq- vation (FIGURE 3.5). uitin within the cell. It is not, however, used

3.3 Ubiquitin attachment to substrates requires multiple enzymes 5 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 6

directly in the modification of substrates. subunits within a complex. The RING finger Instead, the ubiquitin is transferred a second binds directly to an E2, while a separate region time, this time onto a specific cysteine of one of the E3 binds the target protein; an example of a number of ubiquitin-conjugating en- is shown in FIGURE 3.8. The E3 positions the E2 zymes or E2s (Figure 3.5). Because ubiquitin and the substrate so that the ubiquitin can be is linked to the E1 and E2 proteins by the same transferred directly between them. Thus, one type of bond, no additional energy is required important function of this type of E3 is to estab- for this reaction. Thioester bonds are high-en- lish a geometry that permits the two compo- ergy bonds and are highly susceptible to nu- nents to interact productively. cleophilic attack, making reaction between As is clear from the structure of the E3 them and an amino group of a substrate pro- shown in Figure 3.8, this type of E3 can be com- tein favorable. This tightly choreographed se- posed of multiple subunits. In addition to the ries of ubiquitin transfers is common to all the RING finger, one larger subunit often provides well-characterized pathways by which ubiqui- the basic framework to align the E2 and the tin is added onto substrates. substrate-binding site, and another subunit con- The attachment of ubiquitin to a substrate tains the site itself. Different substrate-binding protein is catalyzed, finally, by a ubiquitin- subunits can be used with the same framework protein ligase or E3. These proteins act to- subunit, producing E3s with similar architec- gether with E2s, the two types of protein often tures but different substrate specificities. FIGURE forming a complex. Many different types of E3 3.9 shows the same framework subunit, RING are present in a cell, but most can be divided finger subunit, and E2 associated with three dif- into two major classes that differ in how they ferent substrate-binding subunits. These variable collaborate with an E2 to cause the transfer of subunits not only provide sites for binding dif- its bound ubiquitin onto a protein. More nu- ferent substrates, but also produce gaps of dif- merous are E3 proteins that cause ubiquitin to ferent sizes between the site and the E2. This be transferred directly from the E2 to the tar- presumably helps the E3 adapt to substrates of get protein (FIGURE 3.6). These E3 proteins con- different dimensions. tain a structural motif called a RING finger Proteins in the other major class of E3s have (FIGURE 3.7) that constitutes either a domain a single subunit with a segment called the HECT within a larger polypeptide or one of several domain. Within it is an absolutely conserved

FIGURE 3.6 The E3 is the protein most Attaching ubiquitin to a target protein The RING finger motif directly responsible for target protein recognition in ubiquitin-protein attach- E3 Target Three-dimensional Cross-brace ment. The E3 has distinct binding sites protein An E3 identifies a structure arrangement of zinc- target protein in coordinating residues for the target protein and for the ubiq- need of C uitin-charged E2. In this way, the E3 degradation 7 8 functions as an adaptor that brings the NH2 charged E2 and the target protein into 3 close proximity, thereby stimulating the 4 transfer of ubiquitin between the two. E2 Binding of the E2 to the E3 positions 5 Pre-formed ubiquitin chains may also be Ub activated ubiquitin 6 transferred from the E2 to the target pro- near target protein tein. NH2 C N 1 2 N Ubiquitin is HISTIDINE CYSTEINE transferred from ZINC the E2 to the target protein FIGURE 3.7 On the left is an example of a RING finger do- main showing how the zinc atoms (red) are essential to its structure. The zinc-coordinating residues and a conserved Repeated rounds are shown as stick figure representations. On the of ubiquitin right is a schematic representation of the cross-brace arrange- transfer produce a ment of the residues that contact the zinc atoms. Reprinted ubiquitin chain from Aravind and Koonin, “The U box is a modified RING fin- ger—A common domain in ubiquitination,” Curr. Biol. 10, pp. 132–134, © 2000; used with permission from Elsevier.

6 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/26/07 11:43 AM Page 7

Structure of an E3 Substrate-binding subunits FIGURE 3.9 E3s with different specificities can be cre- ated by exchanging substrate-binding subunits. Here three different substrate-binding subunits are shown E3 bound to the same larger framework, forming E3s that target different proteins. In the middle panel, a yel- E3 low peptide with the sequence recognized by the E3 is shown at the substrate-binding site. Reprinted from Substrate- Wu et al., “Structure of a -TrCP1-Skp1--catenin com- binding plex: Destruction motif binding and lysine specificity subunit Substrate- E2 Cys of the SCF-TrCP1 ubiquitin ligase,” Mol. Cell 11, pp. RING finger binding domain subunit 1445–1456, © 2003; used with permission from Elsevier. RING finger Cys domain E2

E3 FIGURE 3.8 The structure of a multisubunit E3. A RING finger subunit (red) and a substrate bind- ing subunit (green) are juxtaposed by a third large, extended subunit. The RING finger posi- tions the E2 and its bound ubiquitin so that it is immediately adjacent to where the substrate will E2 Cys bind. The cysteine residue where ubiquitin is at- tached to the E2 is circled. Reprinted from Wu et al., “Structure of a -TrCP1-Skp1--catenin com- plex: Destruction motif binding and lysine speci- E3 ficity of the SCF-TrCP1 ubiquitin ligase,” Mol. Cell 11, pp. 1445–1456, © 2003; used with permis- sion from Elsevier.

E2 Cys

Some E3s attach to ubiquitin FIGURE 3.10 E3s of the HECT class act as di- rect carriers of the activated ubiquitin. For the E3 HECT E3s, the activated ubiquitin is transferred E2 Ubiquitin is first from the E2 to a conserved cysteine side chain transferred from in the HECT domain. This bucket brigade-like Ub SH E2 onto E3 series of ubiquitin transfers terminates with NH cysteine residue. Remarkably, this cysteine func- 2 the attachment of ubiquitin to a lysine side tions in yet another ubiquitin transfer reaction, chain of the target protein. this time from the E2 to the E3 (FIGURE 3.10). Ubiquitin is then Only after the transfer has occurred can the transferred onto the target protein ubiquitin finally be attached to the substrate Target protein, which is bound by a distinct region of NH2 protein the E3. In addition to all the enzymes they have The process for adding ubiquitin to proteins, cells also con- repeats tain enzymes that remove it. Ubiquitin is some- SH times removed from a target protein very soon after it has been added, with the result that the protein is not degraded. Most eukaryotic Many rounds of ubiquitin transfer encode several dozen such deubiqui- SH produce a tinating enzymes. This suggests that, in some ubiquitin chain cases, ubiquitin may cycle on and off proteins constantly, and that the rate at which they are degraded is determined by how active a partic- ular deubiquitinating enzyme is.

3.3 Ubiquitin attachment to substrates requires multiple enzymes 7 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 8

How are proteins that need to be ubiquitinated 3.4 Substrate recognition in and degraded specifically selected from among all the ubiquitin-ligation the other proteins present in a cell? As suggested by the fact that they bind to the substrate, E3 pro- system teins and E3-E2 complexes are the major arbiters Key concepts of substrate specificity in most cases. Consideration • E3 enzymes determine the specificity of proteoly- of the sizes of the E1, E2, and E3 protein families sis. Most E3s bind to only a small number of sub- supports this conclusion. Most eukaryotes have strate proteins. only a single E1. This lone E1 transfers its activated • An E3 recognizes a degradation signal, or degron, ubiquitin to any of between one and two dozen dif- within a substrate protein. The degron contains all ferent E2s, depending on the organism. Why there the information necessary for a protein to be de- graded. are so many E2s is not completely understood. The • The degron is a structural feature of the substrate. simplest idea is that they can alter the specificity of In some cases it is a short, exact sequence of E3s as part of E3-E2 complexes. Different combi- amino acids; in others it is a less easily definable nations of E2s and E3s would thereby allow a wider property of the protein such as exposure of hy- range of substrates to be recognized than if the E3 drophobic amino acid residues on the protein’s alone was responsible for substrate degradation. surface. By far the greatest degree of molecular com- • Many degrons can be activated or inactivated by posttranslational modifications such as phosphory- plexity in the ubiquitin-ligation system is among lation. the E3s. There are hundreds of proteins con- taining RING-finger domains encoded in the

FIGURE 3.11 The crystal structure of Recognition of a degron by an E3 genomes of higher eukaryotes, and tens or hun- the substrate recognition domain of an dreds (depending on the species) of substrate- E3 bound to a short peptide. The pep- recognition subunits that form components of tide has the same sequence as the de- gron on proteins targeted by the E3. Substrate multisubunit RING-finger E3s. Assuming that Courtesy of Bing Hao and Nikola P. recognition each of these proteins exists in order to recog- domain of Pavletich. E3 nize a specific set of substrates, their number emphasizes how many different cellular events proteolysis is used to control. Degradation What is it about a substrate that is recog- signal of the protein that nized? Over the past ~15 years, a variety of E3 recognizes degradation signals or degrons have been described for both natural and engineered sub- strates of the ubiquitin-proteasome system. These FIGURE 3.12 Some target proteins Some degrons require phosphorylation are only polyubiquitinated after signals are structural features or sequences within they have been altered by another the substrate that are sufficient to cause its rapid Target DEGRON posttranslational modification. protein Signaling causes proteolysis. A degron, or a part of a degron, is rec- D S phosphorylation Phosphorylation is the most com- NH2 G ognized and bound directly by a specific E3 or L mon degron modification that trig- D of degron on ATP S target protein E3-E2 complex. Here we will discuss several dif- gers substrate binding to an E3. ADP ferent kinds of degrons in order to illustrate the different types of logic behind their use. In some cases, a single short sequence within D E3 with E2 binds S a substrate (as few as seven amino acids) is NH2 G phosphorylated L D S degron largely responsible for the binding of the E3. E3 FIGURE 3.11 shows the recognition domain of an E2 E3 bound to a short peptide with the same se- quence as the degron in the protein that the E3 Ubiquitin is repeatedly recognizes. The snugness of the fit between the transferred to two makes it clear how the specificity of recog- Ub target protein NH2 nition is achieved. In many proteins such seg- ments are recognized without being modified. In others they are only recognized once they Polyubiquitinated have undergone a specific posttranslational protein is then degraded modification—for example by a phosphate group or hydroxylation of a specific proline (FIGURE 3.12). This allows the rate of degradation

8 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 9

of the substrate to be regulated by signal trans- Degrons can be hidden by complex formation duction pathways or by changes in the cellular MONOMER DIMER environment, such as the changes in the con- Exposed centration of oxygen in the cytoplasm that lead hydrophobic to the hydroxylation of some proteins. surface One situation in which control of protein E3 degradation by signaling is important is devel- opment. Phosphorylation-dependent degrons are found in many developmentally regulated proteins, and the properly timed degradation of such proteins is essential. For example, a protein E3 called catenin functions as a central transcrip- tional regulator during Drosophila embryogene- sis. Up until a certain point in development it is constitutively phosphorylated, creating a phos- phate-dependent degron. The protein is con- stantly ubiquitinated and degraded as a result. Activation of the Wingless signaling pathway— An exposed hydrophobic When the protein is part surface on a protein of a complex, the one of several signaling pathways that control monomer serves as a hydrophobic surface is development—causes a major developmental degron and leads to shielded protein degradation transition by preventing catenin phosphoryla- tion, allowing catenin to avoid degradation and FIGURE 3.13 Hydrophobic protein regions that are buried in protein interiors or protein-protein interfaces are recognized activate needed for the next stage of de- as degrons when exposed at the protein surface. Such expo- velopment. This developmental mechanism is sure often signals the lack of proper or a fail- highly conserved and also functions in humans. ure to assemble into a functional . This gives Deregulation of the Wingless/Wnt pathway has the cell a quality control mechanism to ensure that only cor- been implicated in several cancers, one of a grow- rectly folded and assembled proteins perdure. ing number of examples where regulation by ubiquitin-mediated protein degradation is in- quality control—the elimination of damaged or volved in medically relevant events. (See Section misfolded proteins—may often utilize exposed 3.12, Supplement 1: Failures of the ubiquitin-protea- hydrophobic elements within proteins as de- some system are a cause of human diseases.) grons. Because the quality control system must In general, recognition of a specific protein be able to degrade any protein if it is misfolded, structure by an E3 or an E3-E2 complex follows these determinants are unlikely to be specific se- the same general principles found for other pro- quences of amino acids. Instead, it seems likely tein-protein interactions. For example, many well- that the system is designed to recognize the gen- known protein-interaction motifs, such as SH2 eral property of hydrophobicity rather than a domains and coiled coils, can function in the con- particular sequence of hydrophobic amino acids. text of E3s as substrate-binding sites. If an E3 has Some degradation signals are remarkably an SH2 domain, for example, it will recognize and complex. The reason is often to ensure that a ubiquitinate a substrate only if a specific tyrosine protein is degraded only under very specific in the substrate becomes phosphorylated. conditions. One example is a phosphorylation- A very different type of degradation deter- dependent degron found in a yeast inhibitor of minant is a hydrophobic surface exposed on the the onset of DNA replication. At least six of nine exterior of a protein. This feature is an effective potential sites in this inhibitor must be phos- degron because hydrophobic amino acids are phorylated by a cyclin-dependent kinase be- normally buried within the interiors of most fore the inhibitor is recognized by an E3 ligase properly folded proteins, or occur at the inter- and ubiquitinated. The requirement for phos- face between the subunits of a protein complex phorylation at multiple sites prevents inadver- and are covered when they have assembled cor- tent degradation of the inhibitor by small rectly (FIGURE 3.13). An exposed hydrophobic fluctuations in kinase activity, and helps make surface is therefore a likely indicator of an im- degradation an abrupt, all-or-nothing response properly folded or assembled protein. Such pro- that can only occur when a sufficiently strong teins have the potential to form toxic aggregates signal is received. This mechanism, based upon within the cell and must be eliminated before the complex nature of the degron, helps ensure they can do so. This logic suggests that protein a sharp and irreversible cell cycle transition.

3.4 Substrate recognition in the ubiquitin-ligation system 9 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 10

tory complex or cap, sits atop each end of the 3.5 Degradation of proteins 20S core particle, where it controls entry and by the proteasome exit into the central hydrolytic chamber. The regulatory complex includes polyubiquitin-bind- Key concepts ing sites and a ring of ATPase subunits that use • Ubiquitinated proteins are degraded by the protea- the energy of ATP hydrolysis to unfold substrates some. and translocate them into the interior of the • The proteasome is composed of three protein sub- core particle. complexes: a central, cylindrical complex, the 20S proteasome, is capped at either end by a copy of The architecture of the 26S proteasome re- the 19S ATPase complex. flects several fundamental constraints. First, it • An internal chamber within the 20S proteasome must degrade only specifically targeted pro- contains the protease active sites. Access to the teins. Second, it must be able to cleave almost chamber is through a narrow channel at each end any targeted protein down to short peptides of the cylinder. without releasing any of the intermediate pro- • The 19S complex acts as a gate to the channel and teolytic products. The first requirement is met uses the energy of ATP hydrolysis to unfold sub- strates and move them into the proteolytic cham- because substrates can only enter the interior ber. of the 20S subcomplex via the narrow chan- • The polyubiquitin chain is released from the sub- nels at its ends (FIGURE 3.15). By enclosing its strate protein prior to or during protein degrada- proteolytic sites within a central chamber that tion. can only be reached via small openings, the proteasome prevents random proteolysis: the Ubiquitinated proteins are degraded by a so- dimensions of most folded and assembled pro- phisticated and highly specialized protease called teins are simply too large to allow them to en- the proteasome. The 26S proteasome is a very ter. Moreover, in the absence of the 19S cap large (~2.5 MDa) structure constructed from at complex, the structure of the 20S proteasome least 32 different polypeptides and designed to changes slightly so that its entry ports are closed, function as an efficient but highly selective pro- preventing unregulated entry into the prote- teolytic machine (FIGURE 3.14). This machine is olytic chamber. composed of two major subcomplexes. The cen- The ability of the proteasome to degrade al- tral 20S proteasome or core particle is the pro- most any polypeptide without allowing large teolytically active component. It is a hollow intermediates to escape is also reflected in its cylindrical structure formed from a stack of four architecture. The proteolytic active sites are in rings, each composed of seven subunits. Narrow the 14 subunits that comprise the central pair channels run through its ends to a large cen- of rings in the 20S proteasome and face inward tral chamber that houses the protease active so that they line the walls of the central cham- sites. The second subcomplex, the 19S regula- ber. In eukaryotes, six of the 14 subunits bear active sites. The active subunits are all related FIGURE 3.14 The structure of the 26S protea- The proteasome to one another but differ sufficiently within their some, based upon a combination of electron microscopy and X-ray crystallography. It is com- STRUCTURE active sites to give them distinct peptide cleav- posed of three large complexes, two of which 19S cap 20S cylinder 19S cap age preferences. Each individual site also has (19S, green) cap the ends of a central cylindri- Base the ability to cleave a range of sequences, a prop- cal complex (20S, blue). The 20S complex is erty that is enhanced because of the high effec- composed of four stacked rings that enclose a Lid tive concentration of the substrate when it is central cavity that is only accessible through the ends of the cylinder. The different parts of trapped within the proteolytic chamber. This al- the overall structure play different roles in the lows even weak binding interactions between recognition and proteolysis of substrates. The Central the substrate and the active site pockets to be suf- cavity upper panel is reprinted from Larsen and Finley, 50 Å ficient for peptide-bond cleavage. Together, these “Protein translocation channels in the protea- properties allow the proteasome to cleave al- some and other proteases,” Cell 91, pp. FUNCTIONALITY 431–434, © 1997; used with permission from Polyubiquitin Substrate most any substrate protein into short fragments. Elsevier. removal unfolding The narrow and protected entry channels that prevent nonubiquitinated proteins from entering the proteasome serve the additional purpose of making it difficult for polypeptides to escape once they are within the central cham- Processive proteolysis ber. Proteolytic products cannot easily diffuse out until they have been cleaved to fairly short

10 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 11

lengths, and this diffusion barrier limits the re- Entry points into the proteolytic chamber lease of peptides until they have reached a length

of ~8 to 12 amino acids. This is what makes the 20S 19S proteasome a processive protease. Interestingly, Channel is Open the lengths of the peptides produced by the pro- Substrate can enter 20S via narrow teasome are similar to those of peptides pre- opening when 19S is sented on the surface of the cell by class I MHC present molecules. The proteasome is the primary source of such antigenic peptides, which must be trans- ported from the cytosol across the membrane 20S of the endoplasmic reticulum (ER) in order to bind to class I MHC molecules. Channel is Closed Substrate cannot For the 20S proteasome to degrade a folded enter due to substrate protein, the 19S regulatory cap must structural changes in 20S when 19S is first bind the polyubiquitinated protein, remove absent its polyubiquitin tag, unfold the protein, and translocate it into the 20S proteasome (FIGURE 3.16). A variety of subunits within the cap co- FIGURE 3.15 Cut-away structures of the 20S proteasome when operate to perform these functions. Some of its part of a complete 26S proteasome (top)—with 19S complexes at each end—or when alone (bottom). When part of the com- components are specific “ubiquitin receptors,” plete proteasome, the 20S complex has small openings in its ends, able to recognize and bind directly to ubiqui- allowing substrate proteins prepared by the 19S complexes to tin. Binding by one of these proteins to the gain access to the central chamber where the proteolytic sites polyubiquitin attached to a substrate is likely are located. These openings are absent from the free 20S com- to be the means by which many substrates first plex, preventing random proteins in the cytoplasm from enter- ing. Four of the proteolytic sites are shown as small blue dots interact with the proteasome. In other cases, in the upper panel. substrates can interact initially with soluble “adaptor” proteins containing both ubiquitin- and proteasome-binding sites. Path of a protein within the proteasome Once ubiquitin receptors have captured a substrate, other components within the 19S Proteins in the lid complex unfold it. The 19S cap includes a half- of the 19S cap dozen related but distinct ATPase subunits be- bind polyubiquitin chain, linking longing to the so-called AAA+ family of ATPases, substrate to the proteasome. or AAA+ ATPases. These subunits are believed Lid Base to form a hexameric ring around the end of the 19S CAP pore leading into the 20S proteasome, and drive both the unfolding and the translocation of sub- ATPases in the strate proteins into its interior. Substrates are base unfold and push the thought to bind to several of the subunits si- substrate into multaneously, with ATP binding or hydrolysis the central pore of the 20S then causing the individual subunits in the ring ADP to move relative to one another, pulling apart ATP the folded parts of the protein and propelling The substrate is them through the pore and into the proteolytic degraded within cavity. At some point in the process, the polyu- the proteolytic cavity and at biquitin chain must be released so that it will not some point the impede the translocation of the substrate. polyubiquitin chain is detached Proteasome-associated deubiquitinating en- Proteolytic from the zymes, including several that are integral sub- cavity substrate units of the 19S cap, have been implicated in this FIGURE 3.16 The 19S regulatory complexes encounter a substrate process. Whether the polyubiquitin chain is re- first. They are responsible for recognizing the polyubiquitin chain, leased prior to or during the initiation of sub- removing it from the substrate, and unfolding and translocating strate unfolding is not yet known. Once it is the substrate into the 20S proteasome core. ATP hydrolysis by the 19S complex is required to unfold the protein, which may occur released, the chain is cleaved into individual at the same time it is being translocated into the 20S complex. ubiquitin monomers and they are again used Proteolytic sites within the 20S proteasome cleave the substrate to mark proteins for proteolysis. into short peptides, which are then released.

3.5 Degradation of proteins by the proteasome 11 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 12

of the lysosome. Getting them there requires 3.6 Membrane proteins are several steps (FIGURE 3.18). Proteins from the degraded by several Golgi apparatus or the plasma membrane that need to be degraded are first collected in vesi- mechanisms cles and sent to the late endosome. There, sec- Key concepts tions of the membrane that contain the protein • Transmembrane proteins require special mecha- invaginate and vesiculate, creating an internal nisms for their full degradation. vesicle within the endosome. Fusion of the en- • Proteins are extracted from the membrane of the tire endosome with the lysosome releases the ER and degraded in the cytoplasm by the protea- vesicle into the interior of the lysosome, where some. the multiple nonspecific proteases within it can • Membrane proteins from the plasma membrane and degrade the protein. The catch here is that dif- some other membranes are moved in vesicles to the interior of the lysosome, where both the lipids ferent parts of the protein are still separated by and the proteins are degraded. the membrane of the internalized vesicle. This mechanism of proteolysis therefore requires not Degradation of transmembrane proteins poses only proteases, but also lipases that can destroy a special problem because different parts of the intervening membrane. Lysosomes (called such proteins are on opposite sides of a mem- vacuoles in yeast) contain both lipases and pro- brane. Eukaryotic cells have devised several teases and can fully degrade both the lipids and solutions to this problem, all of which permit the transmembrane proteins in a vesicle. degradation of transmembrane proteins with- Remarkably, these enzymes spare the mem- out the accumulation of large proteolytic frag- brane of the lysosome itself as well as the many ments. In the first, complete degradation of a resident proteins of this organelle. In Section 3.8, is enabled by remov- Integral membrane proteins that function at the cell ing it from the bilayer so that it ends up entirely in a single compartment, where it can be de- Membrane protein degradation in lysosomes graded in the same way as a soluble protein (FIGURE 3.17). This mechanism is employed in Protein to be degraded the ER as well as in mitochondria. As will be discussed in more detail in the next section, proteins are often extracted from the mem- Vesicle brane of the ER and degraded in the cytoplasm formation by the proteasome. and fusion to endosome Proteins in some of the cell’s other mem- branes are degraded by a much different mech- CYTOPLASM anism that involves sending them to the interior

Degradation of membrane proteins by retrotranslocation ENDOSOME

ENDOPLASMIC RETICULUM Channel Formation of Integral multivesicular membrane body protein Membrane fusion releases the Ubiquitin Proteasome vesicles into LATE lysosome CYTOPLASM ENDOSOME Degraded protein

The transmembrane The protein is The protein is LYSOSOME protein is transferred polyubiquitinated on degraded by the to the channel the cytoplasmic side proteasome FIGURE 3.18 Transmembrane proteins on the cell’s surface of the membrane can be collected by vesicle formation. The vesicles are sent FIGURE 3.17 Proteins originally translocated into the lumen of the en- to endosomes, where membrane fusion between the vesicle doplasmic reticulum or inserted in the membrane can be retrotranslocated and the endosome incorporates the proteins into the endo- back into the cytoplasm if the proteins need to be degraded. Channels some’s membrane. Membrane invagination at the late endo- for retrotranslocation must exist, but their identity is not yet certain. some forms internal vesicles and creates a multivesicular body Polyubiquitin attachment might occur before, during, or even after the (MVB). Each MVB fuses with a lysosome, releasing the vesi- retrotranslocation process. For some proteins, modification by polyubiq- cles into its interior. There, lipases and proteases degrade uitin is necessary for their full ejection back into the cytoplasm. both the protein and lipid components of the vesicles.

12 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 13

surface are degraded within lysosomes, this type of are collectively referred to as ER-associated degradation will be described in more detail; it degradation (ERAD). is how the levels of most plasma of proteins into the ER or inser- proteins are decreased or “downregulated” when tion of transmembrane proteins into its mem- necessary. brane requires a transporter called the Sec61 , which provides an aqueous chan- nel through which polypeptide segments move 3.7 Retrotranslocation from into the ER lumen or the . Retrotranslocation of proteins out of the ER ap- the endoplasmic pears to require a specific transport apparatus reticulum back into the as well, possibly the same Sec61 translocon used for import. Here the emphasis will be on how cytoplasm allows proteins are recognized by ubiquitin ligases (E3s) degradation of short- associated with the ER, and how polyubiquiti- nation of the substrate is thought to be involved lived ER proteins in the retrotranslocation process. Such sub- Key concepts strates include not only aberrant proteins that • A protein machine based on a hexamer of AAA+ failed quality control screening but also func- ATPases, Cdc48/p97, helps retrotranslocate pro- tional proteins that are degraded when they are teins from the ER to the cytosol. not needed. For example, the rate-limiting en- • A polyubiquitin chain on the targeted substrate zymes in both the and unsaturated fatty helps bind it to the Cdc48/p97 complex. acid biosynthetic pathways are located in the • Ubiquitination of a retrotranslocation substrate occurs at the cytoplasmic face of the ER mem- membrane of the ER, and their levels are reg- brane. ulated by degradation stimulated by the end • For some transmembrane substrates, the protea- products or intermediates of the pathways. some is also required for retrotranslocation. Ubiquitin-dependent degradation of these trans- membrane enzymes is crucial for reducing their As the site where a large fraction of a cell’s newly activity when levels of or unsaturated synthesized proteins are folded, modified, and fatty acids, respectively, become too high. assembled, the ER must also have the capacity The nature of the degrons of ERAD sub- to monitor these activities and eliminate proteins strates remains poorly understood in most cases. that fail to complete them successfully. All of a An exception is the target of a particular cytoso- cell’s secretory proteins and most of its trans- lic E3 that specifically recognizes a type of man- membrane proteins are processed by the ER, and nose-rich sugar chain that marks glycosylated they all enter it in an unfolded state, many be- proteins unable to pass the ER quality control ing directly inserted into the organelle as they system. Recognition by the E3 occurs on the are synthesized. Once such proteins are within cytoplasmic face of the ER membrane, but this the ER, they must fold, assemble into complexes is preceded by binding of the same sugar chain with other proteins, and undergo various post- to surveillance proteins on the interior of the ER. translational modifications. If these proteins fail Glycosylation is not necessary for targeting to fold or assemble properly, or lack the appro- many other ERAD substrates, however. Most priate disulfide bonds or sugar modifications, ERAD degrons are probably based on properties they are prevented from proceeding to the Golgi. of the sequence or structure of the polypeptide Through the action of ER chaperones and other itself and are likely to be similar to those that enzymes, further attempts at folding and assem- control the degradation of cytosolic proteins. As bly are made. However, if these attempts fail, the noted earlier in Section 3.4, Substrate recognition in system gives up and the protein is designated for the ubiquitin-ligation system, an exposed hydropho- destruction. Remarkably, it is degraded in the cy- bic protein surface is a degradation determinant toplasm: terminally misfolded or misassembled for a number of cytosolic proteins that lack their proteins in the ER are translocated back to the normal protein partners or are misfolded. cytosol (called retrotranslocation or disloca- Analogous determinants probably allow molec- tion), where they are polyubiquitinated and de- ular chaperones to retain proteins within the ER. graded by the 26S proteasome (Figure 3.17). The If retained in the ER long enough, misassembled mechanisms that monitor folding and assembly or misfolded proteins are returned to the translo- within the ER are called ER quality control, con and retrotranslocated into the cytosol for and the proteolytic pathways linked to the ER degradation. During retrotranslocation, hy-

3.7 Retrotranslocation from the endoplasmic reticulum back into the cytoplasm allows degradation of short-lived ER proteins 13 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 14

Ubiquitin-driven retrotranslocation polyubiquitin has been suggested to activate the ATPase, stimulating the extraction of the sub- Substrate ENDOPLASMIC RETICULUM strate from the membrane. For some transmembrane proteins, the pro- Channel teasome also contributes to retrotranslocation. As discussed earlier, the 19S regulatory com- plex of the proteasome also contains a hexam- eric array of AAA ATPases. These might extract a protein from the membrane in the same way Cdc48 Ubiquitin ATPase as does Cdc48/p97, but would feed it directly into complex chain the core of the proteasome. CYTOSOL Both the proteasome and the Cdc48/p97 complex also participate in another type of ER- Substrate is free Substrate's Binding of Substrate is to slide back and movement back polyubiquitinated pulled into the linked proteolysis, namely, the processing of forth through to ER is blocked chain activates cytosol to be membrane-tethered transcription factors. These membrane by Cdc48 the ATPase degraded complex portion of Cdc48 proteins consist of at least two domains. When complex separated from the rest of the protein, one can FIGURE 3.19 Before the beginning of this sequence the green protein was act as a transcription factor. It is, however, syn- identified as unnecessary or defective and brought to the channel from within thesized fused to a transmembrane domain. As either the lumen of the ER or its membrane. The Cdc48 complex binds the sub- a result, the proteins are inserted into the mem- strate while it is still within the channel, but the ATPase activity of the Cdc48 complex is not stimulated until the substrate has been ubiquitinated. The ubiq- brane of the ER, where they cannot influence uitin ligase is not shown in the figure. transcription. When transcription is needed, they are ubiquitinated at the surface of the ER, allow- drophobic elements within these aberrant pro- ing them to bind to Cdc48/p97 and the protea- teins are likely to be recognized a second time, some. Instead of being removed from the this time by ubiquitin ligases associated with the membrane and completely destroyed, however, cytosolic face of the ER membrane. This mode of a limited number of cuts separates the cytosolic recognition is suggested by the fact that several and membrane domains, releasing the cytoplas- of the E2 and E3 enzymes that act upon ER qual- mic domain from the membrane and allowing it ity control substrates are located in the mem- to enter the nucleus through nuclear pore com- brane of the ER and also appear to recognize plexes and activate transcription. What prevents exposed hydrophobic patches in cytoplasmic pro- complete degradation of these transcription fac- teins in need of degradation. This unexpected tors is a fascinating but unanswered question. observation lends credence to the idea that many ERAD substrates may use degrons similar to those in soluble cytoplasmic and nuclear proteins. 3.8 Integral membrane Ubiquitination of most ER proteins is needed not only for their actual degradation proteins that function at but also for their movement back into the cy- the cell surface are toplasm. This is clear because in many cases experimentally blocking the ubiquitination of degraded within lysosomes an ERAD substrate prevents it from ever leav- Key concepts ing the ER. Ubiquitination may be required • Unneeded plasma membrane proteins are collected for retrotranslocation because a major means by endocytosis and delivered via vesicles to the by which ER proteins are moved out of the or- membrane of the endosome. ganelle is by the activity of a multifunctional • Vesicles containing the proteins bud into the inte- rior of the endosome, creating a multivesicular ATPase of the AAA class called Cdc48 or p97 body (MBV) with many internal vesicles. (FIGURE 3.19). A hexameric complex of this en- • The outer membrane of MVBs fuses with the mem- zyme, together with several accessory proteins, brane of the lysosome, exposing the internal vesi- is located on the cytoplasmic surface of the cles to the hydrolytic enzymes of this organelle. membrane, and binds directly to polyubiqui- • Newly synthesized plasma membrane proteins still tin and polyubiquitinated proteins. Once bound within the secretory pathway can be rerouted from to a substrate protein, the Cdc48/p97 complex the TGN to the endosome, and from there to the lysosome for degradation. is believed to undergo ATP-dependent confor- • Ubiquitin plays a role in these events as a sorting mational changes that help pull or ratchet the signal. substrate through the ER channel. Binding to

14 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 15

Among the various membranes of a cell, the Ubiquitin directs membrane trafficking to the lysosome plasma membrane has a special need for the ENDOPLASMIC ability to degrade its proteins selectively. Unlike RECTICULUM LYSOSOME proteins that function within the membranes of CYTOPLASM the cell’s organelles, those in the plasma mem- GOLGI brane must often be changed in order to allow APPARATUS the cell to adapt to new environmental condi- Degradation tions. This requires the destruction of perfectly Protein Protein is LATE functional membrane proteins either after they is needed not needed ENDOSOME have already reached the plasma membrane or while they are on their way there. Ubiquitination Destruction of newly synthesized but un- needed plasma membrane proteins occurs by a EARLY selective rerouting of the secretory pathway. ENDOSOME Proteins destined for the plasma membrane move from the ER through the Golgi and on to the trans-Golgi network (TGN) before finally being Ubiquitination sent to the plasma membrane. When specific EXTRACELLULAR SPACE proteins need to be degraded, this pathway can FIGURE 3.20 Ubiquitination can serve as a sorting signal that directs un- be selectively modified for them at the TGN. needed membrane proteins to lysosomes for degradation. At both the trans- There, proteins to be degraded are polyubiqui- Golgi and the cell surface ubiquitination causes proteins to be diverted to tinated and, as a result, are sorted to the lysosome endosomes. There, the ubiquitin is removed and invagination of the mem- rather than being sent on to the plasma mem- brane collects the proteins in internal vesicles, which are degraded after fu- brane. This allows ubiquitination to determine sion of an endosome with a lysosome. what fraction of the newly synthesized protein trafficks to the plasma membrane under differ- proteins to the proteasome. Although endocy- ent environmental conditions. For example, tosed proteins that are not modified with ubiq- when yeast cells are grown on a good source of uitin can be recycled to the cell surface, those nitrogen, their demand for amino acids from that are ubiquitinated are generally routed to the their surroundings is reduced, and the perme- lysosome for degradation. ase that allows their uptake is destroyed more Many plasma membrane proteins, however, quickly. This is accomplished by elaboration of can be endocytosed without ubiquitination and polyubiquitin chains on the newly synthesized yet are still degraded in the lysosome when nec- protein, which causes it to be sent from the trans- essary. Degradation of these proteins often de- Golgi to the yeast’s vacuole (the equivalent of the pends on ubiquitination after they have arrived lysosome in larger eukaryotic cells) rather than in the endosome. Unmodified proteins are re- on to the plasma membrane (FIGURE 3.20). cycled to the plasma membrane, while those Although ubiquitin participates in this form of that have been ubiquitinated are sent on to the protein degradation, it does not act by directing lysosome. Ubiquitin attached to cargo proteins proteins to the proteasome. Instead, it serves as allows such sorting by directing them to sites of a signal that allows the proteins to be sorted to involution in the membrane of the endosome for an organelle where they are degraded. selective incorporation into internalized vesi- Membrane proteins that must be degraded cles. Ubiquitinated proteins congregate at these after they have reached the plasma membrane sites because of the assembly there of large com- are collected by endocytosis. This pathway is plexes of proteins that include ubiquitin-bind- sometimes triggered by the binding of a ligand ing proteins as well as others that orchestrate to its , and serves to remove the recep- the budding of vesicles into the lumen of the tor from the surface of the cell so that the cell endosome. Once the formation of vesicles has be- becomes less sensitive to the ligand. Here, too, gun, the role of ubiquitin on the target proteins ubiquitin often plays a crucial role because ubiq- is apparently complete because it is removed by uitination of the cytoplasmic portion of the tar- deubiquitinating enzymes before the vesicles get protein serves as an essential step in the pinch off into the interior of the organelle. As in- recruitment of the protein into -coated ternalized vesicles accumulate within it, the en- pits, the first step in endocytosis (Figure 3.20). dosome matures into a multivesicular body A single ubiquitin is often sufficient, unlike the (MVB) (FIGURE 3.21). Its subsequent fusion with chains of ubiquitin that are required to direct the lysosome releases the internal vesicles into

3.8 Integral membrane proteins that function at the cell surface are degraded within lysosomes 15 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 16

FIGURE 3.21 The three-dimensional structure Multivesicular bodies In addition to being the normal site of degrada- of a multivesicular body as reconstructed from tion of many transmembrane proteins from the a series of electron micrographs. Electron mi- 3D reconstruction Electron plasma membrane and several other cellular crographs of one area of a cell were taken at micrograph different focal planes. One of the series is shown compartments, the lysosome or vacuole can at the upper right with various types of mem- also degrade cytosolic proteins. Cytosolic pro- branes—including several MVBs—traced in teins can be transported into the lysosome by color. The three-dimensional structure of the several different pathways, the best understood MVB corresponding region of the cytoplasm is shown of which is . Autophagy is a mech- on the upper left. A three-dimensional view of 250 nm 250 nm one of the MVBs is shown at the bottom. A anism for randomly enclosing organelles and large collection of vesicles is visible within it. large volumes of cytosol within membranes so Reprinted with permission from Murk et al., that their constituents can be nonspecifically Proc. Natl. Acad. Sci. USA 100, pp. degraded (FIGURE 3.22). This pathway is acti- 13332–13337, 2003. vated primarily when cells are starved, but is also important for certain kinds of programmed cell death. Autophagy begins with the forma- Front view 3/4 view tion of a cup-shaped preautophagosome that looks like a vesicle that has collapsed on itself the lysosomal interior, where the proteins and (FIGURE 3.23). The exact origin of this double lipids within them are degraded. membrane structure is still controversial, but it Each of these mechanisms of protein degra- is thought by many to derive from the smooth dation involves the use of ubiquitin as a sort- ER. The edges of the cup-shaped structure fold ing signal, diverting a target protein into a back over a volume of cytoplasm and fuse, form- pathway that leads eventually to the lysosome. ing the autophagosome, a transient organelle Ultimately, the targeted protein is degraded by surrounded by two lipid bilayers. lysosomal proteases, usually without the pro- Trafficking to the lysosome is achieved by teasome ever being involved. Although ubiqui- fusion of the outer membrane of the autophago- tin is best known for its role in targeting proteins some with the membrane of the lysosome. This to the proteasome, more and more roles for it releases a single membrane-delimited vesicle, an outside of that context—some outside the con- autophagic body, into the interior of the lyso- text of proteolysis completely—are being dis- some. From this point on, the breakdown of covered. (See Section 3.13, Supplement 2: Ubiquitin autophagic bodies is analogous to that of the has many functions beyond the proteasome.) vesicles formed within multivesicular bodies (Section 3.8, Integral membrane proteins that func- tion at the cell surface are degraded within lysosomes), 3.9 Under stressful conditions which also end up inside the lysosome. Lipases, proteases, and other hydrolytic enzymes break cytoplasmic proteins can down these structures to allow recycling of be engulfed into amino acids and lipids, which is particularly im- portant in nutrient-starved cells. autophagosomes and Until recently, none of the components of degraded in the lysosome the molecular machinery controlling autophagy was known. Genetic studies with yeast, how- Key concepts ever, finally provided a means of identifying • Under starvation conditions, autophagy allows the random degradation of proteins so that their con- them. Because autophagy is induced by stress, stituent amino acids can be reused for synthesis of particularly starvation, some of the components essential proteins. identified act as parts of autophagy-specific • Large membrane structures called autophagosomes mechanisms for nutrient sensing and signal engulf random portions of cytoplasm that can in- transduction. Three other sets of factors are cur- clude organelles. rently known to be part of the machinery for au- • After fusion of the outer membrane of an au- tophagosome formation. One is a membrane tophagosome with the lysosome, the inner mem- brane and the enclosed cytoplasm are degraded. protein complex that is thought to mark the site • Much of the molecular machinery that choreo- of preautophagosome membrane formation, graphs autophagy is also responsible for formation but the details of how this complex works are of smaller vesicles that engulf specific proteins in not yet known. Remarkably, the other two sets the cytosol for transport into the lysosome. of well-characterized factors necessary for the formation of double-membrane vesicles are cen-

16 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 17

Autophagy Autophagy in a yeast cell FIGURE 3.23 An electron micrograph of a yeast cell with several au- MEMBRANE MEMBRANE CUP tophagic bodies inside its vacuole. CUP An autophagosome that has not yet fused with the vacuole is at the lower CYTOPLASM right. The inset shows a magnified Protein VACUOLE view of an isolation membrane as it is closing around an area of cyto- AUTO- plasm. The two advancing edges will Autophagic soon meet and fuse with one another Advancing bodies Fusion with edges to complete the formation of an au- the lysosome tophagosome. Courtesy of Yoshinori Ohsumi. Auto- phagosome

AUTOPHAGIC BODY

LYSOSOME

Degraded protein FIGURE 3.22 Stress, especially nutrient lim- which functions in Atg8 ligation in much the itation, initiates the formation of an “iso- same way that the -amino group of lysine par- lation membrane,” which elongates and ticipates in the attachment of ubiquitin to pro- closes on itself to enclose a volume of cy- teins. A specific enzyme can cleave Atg8 from toplasm within two concentric lipid bilay- PE, and its activity is also required for autophagy. ers. While closing it resembles a cup. The outer bilayer of the resulting autophago- Here, too, the exact mechanistic role of this Ubl some then fuses with the membrane of the in autophagosome formation remains to be de- lysosome, releasing an autophagic body sur- termined. One proposal is that Atg8 acts as a rounded by a single bilayer into its interior. kind of covalent carrier, deposit- Lysosomal enzymes then degrade all the ing PE in the newly forming precursor mem- components of the autophagic body. brane and possibly contributing to its curvature. Although autophagy appears to be rela- tered on two small proteins, called Atg8 and tively nonspecific with regard to what cytoplas- Atg12 (for “autophagy”), which show many mic contents are trapped and delivered to the similarities to ubiquitin. Although neither has lysosome, a related process occurs in nonstarved any strong primary sequence similarity to ubiq- cells for bringing a subset of lysosomal enzymes uitin, they have sequence similarity to one an- into the lysosome from their site of synthesis other, and one (Atg8) has been shown to have and assembly in the cytosol. Strikingly, many of a three-dimensional structure very similar to the same genes important for autophagy are re- ubiquitin. Like ubiquitin, the C-termini of both quired for this highly specific enzyme delivery proteins are activated with ATP by an enzyme system. The requirements for each process are related to E1 before being covalently attached not identical, however, which presumably ac- to targets. Also like ubiquitin, both Atg8 and counts for the differences in specificity and vesi- Atg12 form transient thioesters with the E1- cle size that characterize each. like enzyme and are each transferred to an E2- like protein before their attachment to substrates. The detailed roles of these ubiquitin-like pro- 3.10 teins (Ubls) in autophagy have not yet been What’s next? worked out. It is clear, however, that they do not Our appreciation of the significance of intracel- direct the destruction of their targets by the pro- lular proteolysis and our understanding of in- teasome. tracellular protein degradation mechanisms has Atg12 must be attached to another au- undergone enormous changes over the past tophagy protein and is required for either the decade and a half. We now know that many initial extension of the walls of the autophago- regulatory events involve proteolysis and that some outward or the final fusion of membranes ubiquitin ligation is involved in the majority of that closes the organelle. Remarkably, the tar- regulated protein degradation in the cell through get of Atg8 ligation is not a protein but the com- both proteasomal and nonproteasomal routes. mon phospholipid phosphatidylethanolamine Ubiquitin-like proteins have also been found (PE). The head group of PE is a primary amine, to be regulators of certain ubiquitin ligation en-

3.10 What’s next? 17 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 18

zymes and of specialized lysosomal targeting strates, assembly of a ubiquitin chain(s) on the and degradation pathways such as autophagy. protein is generally observed, which allows tight Much remains to be learned. We have a binding to the 26S proteasome. The proteasome very limited knowledge of the nature of natu- breaks down targeted substrates to short pep- rally occurring degrons, and particularly of the tides but recycles the ubiquitin molecules. degrons that define proteins as abnormal or Nonproteasomal mechanisms of protein misfolded. Structural analyses of ubiquitin lig- degradation regulated by ubiquitin or Ubl liga- ases complexed with the degrons they recog- tion are now also known to be widespread. In nize will be required for a full understanding particular, routing to the lysosome of either cy- of this problem. How ubiquitin ligase complexes tosolic proteins via autophagy or of membrane assemble the long polyubiquitin chains that proteins via endocytosis and other membrane serve as signals for proteasome binding is poorly trafficking pathways is controlled not only by understood as well. The mechanics of substrate ubiquitin ligation to the substrates themselves protein binding, unfolding, and translocation but by ubiquitin or Ubl ligation to the compo- within the 26S proteasome are just beginning nents of the molecular machinery that execute to be studied, as is the assembly of this complex or regulate these trafficking events. and fascinating 2500 kDa proteolytic machine. Responsibility for recognizing a substrate Similarly, in pathways such as endocytosis falls primarily to a ubiquitin-protein ligase or and autophagy, ubiquitin or ubiquitin-like pro- E3. Each cell contains hundreds of different ubiq- tein attachments function as sorting or regula- uitin-protein ligases, and they are responsible tory signals, but studies of the mechanisms of for much of the specificity of intracellular pro- action of these various signals are in their in- tein degradation. Each recognizes and binds to fancy. (See Section 3.13, Supplement 2: Ubiquitin a small region in its substrate called a degrada- has many functions beyond the proteasome.) Many tion signal or degron, and then transfers a se- cell biological issues remain to be addressed: ries of ubiquitin molecules onto the protein. How does protein localization modulate the Because ubiquitination usually leads to degra- ability of a protein to be ubiquitinated or mod- dation, the activities of ubiquitin-protein ligases ified by another Ubl? Conversely, how do such are often as tightly regulated as are the substrates modifications affect protein localization and themselves. Regulation can also occur through protein-protein interactions? Where are the exposure or modification of the degrons. many different enzymes of the ubiquitin sys- Degradation signals in many proteins are only tem located? Are there still more Ubls awaiting exposed when they are unfolded or improperly discovery? What additional physiological path- assembled, while the degrons in other sub- ways are regulated by the ubiquitin system and strates—particularly regulatory proteins—can how? Investigations of these and many other be activated or inhibited by posttranslational questions will keep the field vibrant for years to modifications such as phosphorylation or by al- come. tering their interactions with other proteins. Modification of the degradation signal allows the degradation of many proteins in response 3.11 Summary to the activation of signaling pathways. Protein degradation is a widespread means of controlling the activities of cellular regulatory proteins and is also essential as a protein sur- 3.12 Supplement 1: Failures of veillance or quality control mechanism that re- moves defective proteins from the cell. For most the ubiquitin-proteasome short-lived eukaryotic proteins, conjugation to system are a cause of the polypeptide ubiquitin is an obligatory step in their degradation. Ubiquitin is joined reversibly human diseases to proteins by covalent linkage between its car- Not surprisingly for a system that contributes boxyl-terminus and, most frequently, the - in so many ways to cell , links between amino group of a lysine within the acceptor defects in the ubiquitin-proteasome system proteins. Ubiquitinated proteins are in a dynamic and human disease are being made with ever- state, subject either to further rounds of ubiq- increasing frequency. Our understanding of uitin addition or to ubiquitin removal by the mechanistic basis of these connections and deubiquitinating enzymes. For proteolytic sub- our ability to use this knowledge to intervene

18 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 19

clinically are still very limited. However, this and the damage done to the cell. One idea, in area of ubiquitin biology is beginning to come the example of Parkinson’s disease, is that in under much closer scrutiny. certain neurons of afflicted individuals, specific There are now both sporadic (i.e., sponta- proteins begin to aggregate with time and in neously arising) and inherited forms of cancer this form can inhibit the proteasome, leading that have been linked to defects in components eventually to the death of the neuron. of the ubiquitination machinery. For instance, Finally, human pathogens often bring in ei- a component of a ubiquitin ligase has recently ther their own invidious pieces of ubiquitin sys- been identified as a tumor suppressor. This lig- tem machinery or factors that interfere with ase tags several key growth-stimulating pro- the host ubiquitin-conjugation apparatus, al- teins, including cyclin E and the c-Myc lowing evasion of immune defenses, more vir- oncoprotein, for rapid, proteasome-mediated ulent infections, and in some cases triggering proteolysis. When the ubiquitin ligase is mu- specific cancers. For example, in some herpes tated, these regulatory proteins accumulate to family viruses, the virus directs the synthesis of high levels and help drive cells into S phase of a ubiquitin ligase that inserts into the ER mem- the cell cycle under conditions in which they brane. There it triggers the ubiquitin-depend- should be quiescent. ent retrotranslocation and proteasomal In contrast, interfering with the ubiquitin- degradation of the major histocompatibility proteasome system may help keep some forms (MHC) class I heavy chain. This prevents the of cancer under control. An active-site inhibitor presentation of viral antigens to T lymphocytes of the proteasome is now being used to treat at the cell surface, so the infected cells are not multiple myeloma, an otherwise incurable can- destroyed by the cellular immune system. cer affecting B lymphocytes. The exact reasons why inhibition of something as widespread as the proteasome should selectively kill these ma- 3.13 Supplement 2: Ubiquitin lignant cells are not certain, but one idea is that they are uniquely sensitive because of the heavy has many functions demands B cells place on the protein quality beyond the proteasome control system of the ER. B cells are the cells responsible for synthesizing and secreting an- Although ubiquitin is best known as a tag that tibodies, and are therefore among the cell types marks proteins for destruction, it is now clear in the body with the greatest volume of protein that ubiquitin also plays many other roles in moving through the secretory system. the cell. In particular, protein ubiquitination Impairment of the proteasome in this context has emerged as a major signal for modulating would lead to the accumulation of a much protein binding and intracellular membrane greater quantity of damaged proteins than in trafficking, and it is used for these purposes in other cell types, and programmed cell death a wide variety of cellular contexts. might be the outcome. These two examples Ubiquitination is particularly common in make clear the delicate balance of growth con- membrane trafficking events. In this context, trols that can be tipped either way by impair- ubiquitin is attached to its target proteins differ- ment of the ubiquitin-proteasome system. In ently than when it is used as a signal for bind- some cases, this will lead to abnormal cell pro- ing to the proteasome. Targeting of proteins to liferation, but in other situations, such as with the proteasome (Section 3.5, Degradation of pro- myeloma cells, cancerous cells will be pushed teins by the proteasome) requires their attachment into the apoptotic pathway. to a polyubiquitin chain with a particular isopep- Degenerative disorders, particularly neu- tide linkage between the ubiquitin molecules rodegenerative diseases such as Parkinson’s dis- in the chain. In contrast, the pathways that sort ease and Alzheimer’s disease, have also been and move membrane proteins to different lo- associated with defects in the ubiquitin-protea- cations in the cell involve modification of cargo some pathway. In certain heritable forms of proteins with either a single ubiquitin or a polyu- Parkinson’s disease, for example, the neuronal biquitin chain with a linkage different from that defects have been traced to mutations in genes used for targeting proteins to the proteasome. encoding components of the ubiquitin system. These pathways use ubiquitination to direct the These diseases are complex, and thus there are transport of proteins at several different points likely to be multiple ways by which the ubiq- within the cell and sometimes use different types uitin system is involved in their development of ubiquitin modification to specify different

3.12 Supplement 2: Ubiquitin has many functions beyond the proteasome 19 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 20

destinations. For instance, some sorting deci- verse surfaces available for macromolecular in- sions at the trans-Golgi depend on whether the teractions. Ubiquitin and some Ubls can also transmembrane cargo protein is attached to a be assembled in topologically distinct chains. single ubiquitin or to a ubiquitin polymer. Thus, these protein modifiers can act as highly Sorting by ubiquitin modification also occurs flexible adaptor modules for altering protein at the plasma membrane, where ubiquitination conformation or protein-protein interactions. controls what proteins—such as cell surface re- Being proteins, Ubls are also subject to evolu- ceptors—are endocytosed and directed to the tionary divergence through duplication and di- lysosome. In all cases, ubiquitinated proteins versification of Ubl-encoding genes. At present, are recognized by various “ubiquitin receptors” there is evidence for about a dozen different with distinct ubiquitin-binding motifs. Ubl-ligation systems in eukaryotes. As discussed Yet another point of ubiquitin intervention earlier, several of the most divergent Ubl- in membrane protein sorting occurs by modi- systems are crucial for the biogenesis of au- fication of components of the sorting machin- tophagosomes in response to starvation. Others ery itself. Many endocytic and membrane sorting regulate the ubiquitin pathway itself by regu- factors can both bind ubiquitin and be modi- lating the activity of ubiquitin-protein ligases fied by it. This allows, in principle, a network of or by competing for the modification of a sub- dynamic protein-protein interactions. The gen- strate, sometimes targeting the same lysine. It eral model here is based on the idea that cova- would not be surprising if additional ubiqui- lent addition of a ubiquitin molecule to a protein tin-related modifiers were found. One can also enhances or reduces the ability of that protein predict with some confidence that some of these to interact with other factors, such as proteins Ubls will participate in the regulation of protein with ubiquitin-binding domains. Assembly of degradation, and that ubiquitin itself will be specific multiprotein complexes on membrane discovered to have additional mechanistic roles surfaces drives changes in membrane curva- in the panoply of cellular regulatory pathways. ture and can alter the local lipid and protein content of the membrane. Regulated ubiquitin References attachment and removal might drive the as- sembly of such complexes and control mem- 3.1 Introduction brane protein sorting events as a result. It is now also clear that a significant num- Reviews ber of small proteins related to but distinct from Glickman, M. H. and Ciechanover, A. (2002). The ubiquitin also function by covalent attachment ubiquitin-proteasome proteolytic pathway: to other macromolecules, primarily proteins. destruction for the sake of construction. Such ubiquitin-like proteins (Ubls) greatly ex- Physiol. Rev. 82, 373–428. pand the range of physiological processes af- Gottesman, S. and Maurizi, M. R. (1992). Regulation by proteolysis: energy-dependent fected by protein ligation mechanisms. For proteases and their targets. Microbiol. Rev. 56, example, SUMO (for small ubiquitin-related 592–621. modifier) is a Ubl that is ligated to tens if not Hochstrasser, M. (1996). Ubiquitin-dependent pro- hundreds of different proteins and directly reg- tein degradation. Annu. Rev. Genet. 30, ulates many aspects of transcription, chromatin 405–439. structure, nucleocytoplasmic trafficking, and Levine, B. and Klionsky, D. J. (2004). cell cycle regulation. As we have seen, the Apg8 Development by self-digestion: molecular and Apg12 pathways in yeast have specific func- mechanisms and biological functions of au- tions in autophagy. In higher eukaryotes, there tophagy. Dev. Cell 6, 463–477. are multiple Apg8-related proteins, and they Pickart, C. M. and Cohen, R. E. (2004). may have more diverse roles in membrane dy- and their kin: proteases in the machine age. Nat. Rev. Mol. Cell Biol. 5, namics than in yeast. For example, one of the 177–187. human Apg8 paralogs is an -specific protein and another is localized to the Golgi. 3.2 Overview of the ubiquitin-proteasome What features of ubiquitin and ubiquitin- system like proteins might make them better for reg- Reviews ulating proteins than other forms of protein Glickman, M. H. and Ciechanover, A. (2002). The modification? In comparison to modifiers such ubiquitin-proteasome proteolytic pathway: as phosphoryl or acetyl groups, ubiquitin and destruction for the sake of construction. Ubls are much larger and have much more di- Physiol. Rev. 82, 373–428.

20 CHAPTER 3 Intracellular protein degradation 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 21

Hochstrasser, M. (1996). Ubiquitin-dependent pro- masking by heterodimerization of MAT2 and tein degradation. Annu. Rev. Genet. 30, MATa1 blocks their mutual destruction by the 405–439. ubiquitin-proteasome pathway. Cell 94, Pickart, C. M. and Cohen, R. E. (2004). 217–227. Proteasomes and their kin: proteases in the Morin, P. J., Sparks, A. B., Korinek, V., Barker, N., machine age. Nat. Rev. Mol. Cell Biol. 5, Clevers, H., Vogelstein, B., and Kinzler, K. W. 177–187. (1997). Activation of -catenin-Tcf signaling References in colon cancer by mutations in -catenin or Hershko, A., Heller, H., Elias, S., and Ciechanover, APC. Science 275, 1787–1790. A. (1983). Components of ubiquitin-protein Nash, P., Tang, X., Orlicky, S., Chen, Q., Gertler, ligase system. Resolution, affinity purification, F. B., Mendenhall, M. D., Sicheri, F., Pawson, and role in protein breakdown. J. Biol. Chem. T., and Tyers, M. (2001). Multisite phosphory- 258, 8206–8214. lation of a CDK inhibitor sets a threshold for the onset of DNA replication. Nature 414, 3.3 Ubiquitin attachment to substrates 514–521. requires multiple enzymes Winston, J. T., Strack, P., Beer-Romero, P., Chu, C. Y., Elledge, S. J., and Harper, J. W. (1999). Reviews The SCFbeta-TrCP-ubiquitin ligase complex as- Glickman, M. H. and Ciechanover, A. (2002). The sociates specifically with phosphorylated de- ubiquitin-proteasome proteolytic pathway: struction motifs in IB and -catenin and destruction for the sake of construction. stimulates IB ubiquitination in vitro. Genes Physiol. Rev. 82, 373–428. Dev. 13, 270–283. Hochstrasser, M. (1996). Ubiquitin-dependent pro- Yaron, A., Gonen, H., Alkalay, I., Hatzubai, A., tein degradation. Annu. Rev. Genet. 30, Jung, S., Beyth, S., Mercurio, F., Manning, 405–439. A. M., Ciechanover, A., and Ben-Neriah, Y. (1997). Inhibition of NF--B cellular function References Hershko, A., Heller, H., Elias, S., and Ciechanover, via specific targeting of the I- -B-ubiquitin lig- A. (1983). Components of ubiquitin-protein ase. EMBO J. 16, 6486–6494. ligase system. Resolution, affinity purification, 3.5 Degradation of proteins by the and role in protein breakdown. J. Biol. Chem. proteasome 258, 8206–8214. Huang, L., Kinnucan, E., Wang, G., Beaudenon, S., Reviews Howley, P. M., Huibregtse, J. M., and Baumeister, W. et al. (1998). The proteasome: par- Pavletich, N. P. (1999). Structure of an E6AP- adigm of a self-compartmentalizing protease. UbcH7 complex: insights into ubiquitination Cell 92, 367–380. by the E2-E3 enzyme cascade. Science 286, Guterman, A. and Glickman, M. H. (2004). 1321–1326. Deubiquitinating enzymes are intrinsic to pro- Huibregtse, J. M., Scheffner, M., Beaudenon, S., teasome function. Curr. Protein Pept. Sci. 5, and Howley, P. M. (1995). A family of proteins 201–211. structurally and functionally related to the E6- AP ubiquitin-protein ligase. Proc. Natl. Acad. References Sci. USA 92, 2563–2567. Amerik, A. Y., Nowak, J., Swaminathan, S., and Lorick, K. L., Jensen, J. P., Fang, S., Ong, A. M., Hochstrasser, M. (2000). The Doa4 deubiquiti- Hatakeyama, S., and Weissman, A. M. (1999). nating enzyme is functionally linked to the RING fingers mediate ubiquitin-conjugating vacuolar protein-sorting and endocytic path- enzyme (E2)-dependent ubiquitination. Proc. ways. Mol. Biol. Cell 11, 3365–3380. Natl. Acad. Sci. USA 96, 11364–11369. Groll, M., Ditzel, L., Löwe, J., Stock, D., Bochtler, Zheng, N. et al. (2002). Structure of the Cul1- M., Bartunik, H. D., and Huber, R. (1997). Rbx1-Skp1-F boxSkp2 SCF ubiquitin ligase Structure of 20S proteasome from yeast at 2.4 complex. Nature 416, 703–709. Å resolution. Nature 386, 463–471. Hough, R., Pratt, G., and Rechsteiner, M. (1986). 3.4 Substrate recognition in the ubiquitin- Ubiquitin-lysozyme conjugates. Identification ligation system and characterization of an ATP-dependent protease from rabbit reticulocyte lysates. J. Reviews Biol. Chem. 261, 2400–2408. Laney, J. D. and Hochstrasser, M. (1999). Substrate Peters, J. M., Cejka, Z., Harris, J. R., Kleinschmidt, targeting in the ubiquitin system. Cell 97, J. A., and Baumeister, W. (1993). Structural 427–430. features of the 26 S proteasome complex. J. References Mol. Biol. 234, 932–937. Johnson, P. R., Swanson, R., Rakhilina, L., and 3.6 Membrane proteins are degraded by Hochstrasser, M. (1998). Degradation signal several mechanisms

References 21 40632_CH03_hochstrasser.qxd 3/23/07 4:52 PM Page 22

Reviews Reviews Hicke, L. and Dunn, R. (2003). Regulation of Hicke, L. and Dunn, R. (2003). Regulation of membrane protein transport by ubiquitin and membrane protein transport by ubiquitin and ubiquitin-binding proteins. Annu. Rev. Cell ubiquitin-binding proteins. Annu. Rev. Cell Dev. Biol. 19, 141–172. Dev. Biol. 19, 141–172. Isacke, C. M. and Yarwood, H. (2002). The Katzman, D.J., Odorizzi, G., and Emr, S. (2002). hyaluronan receptor, CD44. Int. J. Biochem. Nat. Rev. Mol. Cell Biol. 3, 893–905. Cell Biol. 34, 718–721. Katzman, D.J., Odorizzi, G., and Emr, S. (2002). References Nat. Rev. Mol. Cell Biol. 3, 893–905. Hicke, L. and Riezman, H. (1996). Ubiquitination of a yeast plasma membrane receptor signals 3.7 Retrotranslocation from the endoplasmic its ligand-stimulated endocytosis. Cell 84, reticulum back into the cytoplasm allows 277–287. degradation of short-lived ER proteins Kölling, R. and Hollenberg, C. P. (1994). The ABC- transporter Ste6 accumulates in the plasma Reviews membrane in a ubiquitinated form in endocy- Hampton, R. Y. (2002). ER-associated degradation tosis mutants. EMBO J. 13, 3261–3271. in protein quality control and cellular regula- tion. Curr. Opin. Cell Biol. 14, 476–482. 3.9 Under stressful conditions cytoplasmic Kostova, Z. and Wolf, D. H. (2003). For whom the proteins can be engulfed into bell tolls: protein quality control of the endo- autophagosomes and degraded in plasmic reticulum and the ubiquitin-protea- the lysosome some connection. EMBO J. 22, 2309–2317. Reviews References Levine, B. and Klionsky, D. J. (2004). Hiller, M. M., Finger, A., Schweiger, M., and Wolf, Development by self-digestion: molecular D. H. (1996). ER degradation of a misfolded mechanisms and biological functions of au- luminal protein by the cytosolic ubiquitin- tophagy. Dev. Cell 6, 463–477. proteasome pathway. Science 273, 1725–1728. Ohsumi, Y. (2001). Molecular dissection of au- Johnson, P. R., Swanson, R., Rakhilina, L., and tophagy: two ubiquitin-like systems. Nat. Rev. Hochstrasser, M. (1998). Degradation signal Mol. Cell Biol. 2, 211–216. masking by heterodimerization of MAT2 and MATa1 blocks their mutual destruction by the References ubiquitin-proteasome pathway. Cell 94, Baba, M., Takeshige, K., Baba, N., and Ohsumi, Y. 217–227. (1994). Ultrastructural analysis of the au- Wiertz, E. J., Jones, T. R., Sun, L., Bogyo, M., tophagic process in yeast: detection of au- Geuze, H. J., and Ploegh, H. L. (1996). The tophagosomes and their characterization. J. human cytomegalovirus US11 product Cell Biol. 124, 903–913. dislocates MHC class I heavy chains from the Ichimura, Y. et al. (2000). A ubiquitin-like system endoplasmic reticulum to the cytosol. Cell 84, mediates protein lipidation. Nature 408, 769–779. 488–492. Yoshida, Y. et al. (2002). E3 ubiquitin ligase that Marzella, L., Ahlberg, J., and Glaumann, H. recognizes sugar chains. Nature 418, (1982). Isolation of autophagic vacuoles from 438–442. rat liver: morphological and biochemical char- acterization. J. Cell Biol. 93, 144–154. 3.8 Integral membrane proteins that Mizushima, N., Noda, T., Yoshimori, T., Tanaka, Y., function at the cell surface are Ishii, T., George, M. D., Klionsky, D. J., degraded within lysosomes Ohsumi, M., and Ohsumi, Y. (1998). A pro- tein conjugation system essential for au- tophagy. Nature 395, 395–398.

22 CHAPTER 3 Intracellular protein degradation