Introduction

Proteins that are metabolically relatively stable, with half-lives often exceeding the generation time coexist in the same cell with short-lived proteins that are rapidly degraded. The turnover rates of proteins depend on the physiological state of the cell, and the stability of individual proteins is subject to differential regulation. Metabolic instability is characteristic of damaged or otherwise abnormal proteins, and of many regulatory proteins. A short half-life allows for rapid adjustment of a protein's intracellular concentration through changes in its rate of synthesis or degradation. One important aspect of regulating a protein's activity through degradation is that it is irreversible. Only de novo synthesis, which can be regulated at the level of either transcription or translation, will result in a reappearance or increase of the respective activity. Another advantage of using degradation as a regulatory device is that it ascertains complete removal of a protein, thereby terminating its interactions with various partners via multiple domains, a task difficult to accomplish by reversible protein modifications such as phosphorylation.

In eukaryotes, most short-lived proteins are degraded by the ubiquitin/proteasome pathway. In this process, protein substrates are tagged by the attachment of polyubiquitin chains, which target them for degradation by the 26S proteasome. Ubiquitin itself is recycled upon substrate degradation. The signals within proteolytic substrates that determine their recognition by the ubiquitination/targeting machinery have been investigated for a variety of substrates. However, the often asked question as to whether a protein with a given sequence is likely to be metabolically stable or not, with a few exceptions, cannot be answered reliably at this point in time. The reason is that only a few motifs or structures that identify proteins as proteolytic substrates have been defined precisely, and in many cases their regulation is complex. The purpose of this review is to give an overview on the available information on these destruction signals and their recognition by the proteolytic machinery, with an emphasis on substrates whose degradation is ubiquitin-dependent and mediated by the proteasome. Proteases and their targets in bacteria have been reviewed recently1 and will not be covered in this chapter.

Primary and Secondary Destruction Signals

Features of proteins that render them metabolically unstable are called destruction or degradation signals, or degrons2. Destruction signals can be very simple, e.g. just the N-terminal amino acid residue (N-end rule, see below), and in other cases the signal can be a complex structural feature. Destruction signals may be conditional or masked such that recognition requires that they first be activated, for example by subunit separation, local unfolding or posttranslational modification.3,4,5,6 Another possibility is that the recognition of a destruction signal only occurs in a certain compartment. In that case, a protein's destruction can be induced by its translocation from one compartment to another, e.g. from the nucleus to the cytoplasm as shown recently for cyclin D, p53 and p27Kip1.7,8,9

Cells of the different branches of the kingdoms of life have developed more or less complex systems for the specific recognition and degradation of proteins with a great variety of destruction signals. The specificity of these systems ascertains that only proteins with destruction signals are rapidly degraded while proteins devoid of such signals are not. In more simple systems, the specificity is provided by the proteases or their cofactors which directly recognize the 'primary destruction signals' (for a review see reference 1). In a more complex system, the ubiquitin/proteasome pathway of eukaryotic cells, the specificity is provided by the substrate recognition components of the ubiquitin pathway, called ubiquitin-protein ligases (also called E3s or recognins), which interact with primary destruction signals and mediate the attachment of a polyubiquitin chain to the substrate. This modification serves as a 'secondary destruction signal', which actually targets the substrate to the 26S proteasome.10,11

Ubiquitin-protein ligases (E3s) are a class of enzymes with the common feature to bind both to the primary destruction signals of substrates and to ubiquitin-conjugating enzymes (also called E2 or Ubc).12,13,14 A variety of different types of ubiquitin-protein ligases have been described which do have little if any sequence similarity. Some are monomers like Ubr1, the recognition component of the N-end rule pathway in budding yeast (Fig. 1A), or E6-AP and related proteins like Rsp5 (described in later sections). The latter proteins are characterized by a C-terminal 'hect domain', a conserved ~350 amino acid sequence element that is required for ubiquitin ligation.15 Other E3 enzymes are complex multisubunit assemblies. Examples are the 'Anaphase Promoting Complex' (APC), and 'SCF complexes' composed of Skp1, cullin and F-box proteins (Fig. 1C and D). The large diversity of E3s enables the cell to funnel a variety of substrates whose degradation is independently regulated to the same protease.

In a polyubiquitin chain, Ubiquitin moieties are linked by isopeptide bonds between the e-amino group of a specific lysine (Lys-48) of one ubiquitin and the C-terminal glycine (Gly-76) of the following one11. Lys-48 linked polyubiquitin chains have structural motifs, hydrophobic patches, that are absent from free ubiquitin, from chains containing less than four ubiquitin moieties, and probably also from chains that are linked via lysine residues other than Lys-4816,17,18. These hydrophobic patches bind to specific subunits of the 19S regulatory subcomplex of the 26S proteasome and thus are thought to be an important feature of the polyubiquitin chain secondary destruction signal (for a review see reference 19). Recent data suggest that polyubiquitin chains with other linkages involving Lys-6, Lys-11 or Lys-29 also appear to be able to target proteins to the proteasome20,21,22. The structural properties of such assemblies and how they are recognized by the proteasome are not well studied.

The ability to recognize a very specific secondary destruction signal, namely certain types of polyubiquitin chains, makes the 26S proteasome a highly selective protease, which is nonetheless responsible for the degradation of substrates with a variety of primary destruction signals. The diverse cellular functions of the different substrates require a fine-tuning of the timing and speed of their destruction. The control is implemented at the targeting step whose spatial and temporal separation from the actual destruction allows the utilization of a highly selective, very complex and therefore costly protease (to which this book is dedicated) for a multitude of differentially regulated proteolytic processes.

Much of the proteolysis in eukaryotic cells that is not proteasome-dependent involves the membrane-enveloped proteolytic organelle of eukaryotic cells, the lysosome or vacuole (in yeast). This organelle contains a variety of proteolytic activities that degrade proteins down to reusable components and is thought to participate both in relatively non-specific turnover of cellular proteins and other organelles, and in the selective degradation of certain substrates. The low specificity turnover involves autophagocytic uptake of cytosolic material, which is particularly important under conditions of nutrient starvation when a higher flux of recycling is required23 (for a review see reference 24). The selective degradation of certain substrates in the vacuole is a result of their targeted transport into this organelle. Recently, it has been shown that a number of plasma membrane proteins undergo endocytotic targeting to the lysosome as the result of the addition of ubiquitin attachment (for a review see reference 25). Selective transport into the lysosome of cytosolic proteins to be degraded has also been described. This process is thought to be mediated by membrane transporters that recognize specific primary targeting signals within the protein substrates (for a review see reference 26).

In this review, the term 'primary destruction signal' is used to describe the structural features of proteins, including posttranslational modifications, that make them targets for destruction either directly or via ubiquitination, i.e. the addition of a secondary destruction signal.

Targeting by the amino terminal amino acid residue (N-end rule)

It was discovered by Hershko et al.27 that ubiquitin-mediated degradation of certain test substrates in their reticulocyte-derived in vitro degradation system was dependent on a free N-terminal a-NH2 group. The reason for this observation became clear a little later when Bachmair, Finley and Varshavsky constructed test proteins that were distinguished only by their N-terminal amino acid residues.28 Expression of all 20 such proteins in the yeast Saccharomyces cerevisiae revealed the surprising result that these proteins had extremely divergent half-lives, ranging from a few minutes to more than 20 hours. The relationship between the nature of the N-terminal amino acid residue of a protein and its in vivo half-life was termed 'N-end rule'.28

Detailed studies on the degradation of proteins with 'destabilizing' N-terminal residues ('N-degron') have since provided important insights into the mechanisms involved in ubiquitin-mediated proteolysis. These substrates are targeted for degradation by the Ubr1 recognin and the Ubc2 ubiquitin-conjugating enzyme (also known as Rad6), which form a complex that mediates the attachment of a Lys-48 polyubiquitin chain linked to a single lysine residue of the substrate (Fig. 1A).11,12,13,29,30

The N-end rule applies not only to yeast but to organisms ranging from bacteria (E. coli) to mammalian cells31,32. The sets of destabilizing amino acid residues differ but overlap between E. coli, yeast and mammals with the overall number of such residues increasing from the prokaryote to the higher eukaryote (Table 1).33,34 The detailed analysis of the 'rule books' for the different organisms unraveled a puzzling coincidence. Basically all proteins, with a few exceptions, are first synthesized with N-terminal methionine (Met). Depending on the nature of the second residue, Met is subsequently removed by methionine amino peptidase (MAP). After removal of Met, some N-terminal residues such as alanine, serine or threonine are acetylated. Interestingly, a look at the specificity of MAP reveals that amino acid residues that are destabilizing according to the the N-end rule are the same as those that, when in the second position, either prevent Met processing or become acetylated after it.35,36,37,38,39 In other words, according to the biochemical analyses, MAP and N-terminal acetylase appear to be designed to avoid the generation of destabilizing N-termini.

Despite the detailed knowledge on the nature of the N-end rule destruction signal and the mechanisms involved in targeting of proteins bearing destabilizing N-terminal residues in different organisms, the physiological role of this targeting pathway remains unclear. The reason is that only few physiological substrates have been identified. The first one to be discovered was the Sindbis virus RNA replicase which, as a result of posttranslational processing, bears N-terminal tyrosine, a destabilizing residue according to the N-end rule. Degradation of replicase is thought to reduce its cellular concentration relative to the structural components synthesized from the same template RNA.40

In yeast, two proteins that are targeted for degradation by the Ubr1 N-recognin have been identified recently. Gpa1, the Ga subunit of a heterotrimeric G protein involved in pheromone signaling was found to be ubiquitinated and degraded rapidly upon overexpression of Ubr1 and Ubc2.41 The Cup9 transcription repressor is an unstable protein that is targeted for degradation by Ubr1 and Ubc2 (or Ubc4).42 Interestingly, Cup9 controls the expression of a dipeptide transporter gene (PTR2). It was shown previously that certain dipeptides are potent competitive inhibitors of N-degron recognition43 suggesting that Ubr1 is involved in an autoregulatory circuit controlling dipeptide uptake. Degradation of Gpa1 and Cup9, however, does not depend on their N-termini suggesting that Ubr1 recognizes destruction signals other than the N-degron.41,42 The nature of these signals have not been defined as of yet.

None of the natural substrates discussed above have provided a clue about a general function of the N-end rule pathway. Neither has the genetic analysis of strains lacking the Ubr1 N-recognin, which do not display severe phenotypes, with the exception of their inability to take up dipeptides (see above). In an early hypothesis, it was speculated that this pathway removes proteins from the cytosol that have leaked out from compartments. Indeed, many secretory proteins have N-termini generated by signal peptide cleavage that would target them for degradation in the cytosol.28 The recent discovery that certain ER proteins are exported into the cytosol prior to degradation (see chapter by D.H. Wolf) is consistent with the above-mentioned hypothesis.

Another hypothesis suggested that permanently occurring damage caused by free radicals leads to fragmentation of proteins and thereby the generation of new N-termini, in particular with Asp and Glu residues.44 In this model, the N-end rule pathway would provide a means to destroy protein fragments with destabilizing N-termini. A related explanation would be that this pathway is involved in the destruction of intermediates generated by other targeting/proteolysis pathways.45 Neither hypothesis has received support from direct evidence thus far.

Goldberg and co-workers discovered recently, using specific inhibitors, that the N-end rule pathway is responsible for the degradation of most soluble proteins in extracts of skeletal muscle cells, which is in contrast to extracts from HeLa cells. Moreover, the increased protein turnover in atrophying muscles was found to be largely due to activation of the N-end rule pathway.46,47 These results indicate that this pathway plays an important role in the cell type-specific control of protein turnover in mammalian cells. However, they do leave open the question of whether the proteins' N-termini are the destruction signals operating in these cells.

PEST sequences

Searching for sequence motifs that may represent destruction signals, Roberts and Rechsteiner analyzed the sequence of short-lived proteins.48 They found that these proteins frequently contain regions with a high content of the amino acid residues proline (P), glutamate (E), serine (S), threonine (T) and to a lesser extent aspartic acid. This observation prompted these investigators to put forward the hypothesis that sequences rich in the above-mentioned amino acid residues, so called "PEST sequences", serve as destruction signals.48 They developed an algorithm that allows to calculate a probability (PEST-SCORE) for a given motif to confer metabolic instability (available at http://www.biu.icnet/uk/ projects/pest/ or at http://www.at. embnet/tools/bio/PESTFIND). There does not seem to be a consensus sequence, but it is rather the presence of hydrophilic stretches of more than 12 residues rich in PEST amino acid residues, flanked by lysine, arginine or histidine residues. Positively charged residues are disallowed within the PEST sequence. Since Xray crystallographic studies on certain PEST-positive proteins could not resolve the PEST regions, and because of the hydrophilic nature of these sequences, Rogers and Rechsteiner speculated that they might form conformationally flexible solvent-exposed loops or extensions.49

Since the original proposal of the PEST hypothesis, the relevance of these sequences for Ub-mediated degradation of a variety of proteolysis substrates has been demonstrated (reviewed in reference 49). There are, however, a number of examples where PEST sequences within proteins do not serve as destruction signals, in other cases they were found to be required but not sufficient for degradation.50,51,52,53,54 It is therefore presently not possible to predict reliably if a given PEST motif within a protein will target it for degradation. However, if a short-lived protein contains PEST sequences, it is a good guess that these motifs are involved in its targeting for degradation. In a number of cases in which PEST sequences have been implicated in the instability of proteolytic substrates, their ubiquitination was shown to be preceded by phosphorylation of serine or threonine residues within these regions.5,6,55,56,57,58 In these examples, therefore the phosphorylated PEST motifs represent primary destruction signals, the occurrence of which can be regulated through the activities of protein kinases and phosphatases. Recent work has demonstrated that several proteins whose degradation depends on phosphorylation and PEST sequences are targeted for degradation by Ubc/ligase complexes consisting of Ubc3/Cdc34, Skp1, cullin/Cdc53 and substrate selecting F-box proteins (Fig. 1D; discussed in more detail in a later section). Some of these F-box proteins (Cdc4, Met30) are characterized by the presence of WD40 repeats whose general property might be the binding of phosphorylated proteins.59,60,61 Another F-box protein, Grr1, that has been implicated in selecting phosphorylated G1 cyclins Cln1 and Cln2, however, lacks WD40 repeats.62,63 It remains to be seen whether recognition by F-box proteins and targeting by 'SCF complexes' (Skp1, cullin and F-box protein) will be a general mechanism for proteolysis of substrates containing PEST destruction signals.

Targeting of proteins with 'cyclin destruction boxes'

Cyclins were discovered in sea urchin eggs as proteins that accumulated in interphase and were rapidly degraded in mitosis.64 Cyclins are unstable regulatory subunits of MPF (Maturation Promoting Factor), a cyclin-Cdk complex controlling progression into mitosis (for review see references 65-68). In budding and fission yeast, respectively, the Cdc28 and Cdc2 Cdks associate with a variety of cyclins and control the cell cycle at several steps.68

Glotzer et al.69 showed that destruction of sea urchin cyclin B at the end of mitosis is mediated by the ubiquitin pathway. By deletion analysis they demonstrated that residues 13-66 were sufficient to promote degradation of a fusion protein in mitotic extracts of Xenopus eggs. They defined a 9 amino acid residue sequence extending from residue 42 to 50 as the 'destruction box' required for degradation. Several positions within the destruction box are conserved between B- and A-type cyclins (see table 1). Sequences outside of the destruction box are also required for degradation as a deletion of residues 54-66 resulted in stabilization of a test protein. This region bears Lys residues that are thought to serve as ubiquitination sites.70,71 The destruction boxes are essential for degradation of cyclins A, B1 and B2, but they are not interchangeable.70,72 The destruction box of sea urchin B1 cyclin can replace that of cyclin A but not vice versa.70 The difference in position nine, which is invariably Asn in B-type cyclins but is poorly conserved among A-type cyclins, is mainly responsible for the lack of function of the A-type destruction box in the context of cyclin B sequence. In contrast to cyclin B1, degradation of A-type cyclins require their interaction with Cdk.73 The difference in the exact nature of the destruction box and the requirement for additional sequence motifs are likely to account for the different timing of these cyclins' degradation. Cyclin A is degraded at metaphase; B-type cyclins are degraded at the end of anaphase.74,75,76 Ubiquitination of these mitotic cyclins is mediated by a complex ubiquitin-protein ligase, termed Anaphase-Promoting Complex (APC) or cyclosome,77,78,79 whose activity is regulated through phosphorylation by MPF (Fig. 1C).67,80

Examples of non-cyclin proteins containing cyclin-type destruction boxes are budding yeast Pds1 and fission yeast Cut2. These anaphase inhibitors appear to be involved in controlling sister chromatids cohesion until the onset of anaphase.81,82 Pds1 and Cut2 degradation in anaphase also depends on APC.83,84 In budding yeast, several cohesins have been identified that are involved in holding sister chromatids together. One of them, Scc1/Mcd1, whose association with chromatin is regulated by Pds1 is also degraded by APC.85,86 Budding yeast Ase1 is another protein with a destruction box required for its APC-mediated degradation when cells exit from mitosis and enter G1. Ase1 is a microtubule-binding protein, its degradation appears to be required for disassembly of the mitotic spindle at the end of mitosis.87 Two positive regulators of APC activity have recently been shown to be APC substrates themselves.88,89 In budding yeast, the WD40 repeat protein Cdc20/fizzy activates degradation of the anaphase inhibitor Pds1 by APC.88,90 Clb2 cyclin degradation at the end of anaphase, in contrast, requires activation of APC by Hct1, another WD40 protein,91 and the Polo-like kinase Cdc5.88,89 Clb2 degradation is important for disassembly of the mitotic spindle, cytokinesis and rereplication of the genome.88 Cdc20 and Cdc5 accumulate during G2/M phase and disappear as a consequence of APC-mediated proteolysis at late stages of anaphase.88 Cdc5 has several cyclin destruction box-like motifs including two RXXL sequences in the N-terminal half. Mutation of the Arg and Lys residues in either of these elements had no effect on the protein's stability whereas changing them in both elements led to a partial stabilization of Cdc5.89 Deletion of the first 70 residues including both RXXL motifs resulted in a much stronger stabilization suggesting that the destruction signal is contained within this N-terminal region.88 Cdc20 also bears two sequences similar to the cyclin destruction box. Deletion of one of them resulted in reduced proteolysis whereas deletion of both led to strong stabilization of Cdc20.88

In all the examples described above, cyclin-type destruction boxes appear to be signals for cell cycle-regulated and APC-mediated proteolysis (Fig. 1C). The regulation is likely to occur at the level of APC, which appears to be controlled by phosphorylation67,80,92 and to depend on certain cofactors that control the degradation of subsets of substrates at specific stages of the cell cycle.88,90,91 The nature of the APC or APC-associated polypeptide(s) that interacts with destruction boxes are still unknown.71 Possible candidates are the WD40 proteins Hct1 and Cdc20.88,90,91

Recently, other Cdk complexes containing cyclin-like proteins that regulate transcription and other processes have been discovered in budding yeast (for a recent review see reference 93). The C-type cyclin Ume3 (Srb11 or Ssn8)94,95 is not subject to cell-cycle regulated proteolysis. Its degradation is induced during meiosis and heat shock96 and by various other stess conditions (see below). Ume3, together with the Cdk Ume5 (Srb10 or Ssn3), negatively regulates the expression of invertase95, of an HSP70 gene (SSA1) and of the meiotic gene SPO13.96 Ume3p contains a PEST sequence close the C terminus and a RXXL motif reminiscent of the cyclin destruction box close to the N-terminus. Mutational analysis revealed that both sequences are required for normal degradation of Ume3 upon heat shock in addition to a conserved 'cyclin box' that mediates interaction with Cdks. An interaction with Ume5 Cdk, however, was not required for normal degradation rates upon heat shock. Recent studies by Cooper and Strich demonstrated that, in addition to during heat shock stress and meiosis, Ume3 is destroyed during ethanol shock, oxidative stress as well as carbon starvation or growth on non-fermentable carbon sources. Analysis of the three domains whose mutation stabilized Ume3 during heat shock revealed that the cyclin box domain is required for the rapid turnover of Ume3 during ethanol shock, oxidative stress or carbon starvation, wheras the PEST region mediates down regulation of this cyclin during growth on non-fermentable carbon sources. The cyclin destruction box-like sequence is only required for heat-induced degradation of Ume3 (K.F. Cooper and R. Strich, personal communication). These data suggest that distinct regulatory pathways impinge on the different destruction signals of Ume3. Surprisingly, the degradation of Ume3 was not inhibited by mutations in genes for vacuolar proteases or the proteasome.96 The mechanics of targeting and degradation of this C-type cyclin remain to be elucidated.

An example of a non-cyclin protein whose destruction depends on the presence of a cyclin-type destruction box is Cdc25, the guanine nucleotide exchange factor (GEF) for Ras in budding yeast.97 Cdc25 is degraded constitutively, i.e. independent of the cell cycle. The other Ras GEF present in yeast, Sdc25, is also unstable and bears a cyclin-type destruction box.97 The targeting machinery involved in the degradation of these substrates is presently unknown. Its identification will resolve whether the recognition of the destruction box-like sequences in the latter substrates involves similar components as described above for mitotic cyclins.

Targeting of G1 cyclins and other proteins by SCF complexes

G1 cyclins like budding yeast Cln1, Cln2 and Cln3 do not contain a sequence similar to the destruction box of A- and B-type cyclins. Extensive deletion analyses aimed at the identification of destruction signals have been carried out for Cln2 and Cln3.5,53 These studies demonstrated that PEST sequences close to the C-termini of these cyclins are major determinants of rapid degradation. Furthermore, it was demonstrated that phosphorylation by the Cln/Cdc28 (cyclin/Cdk) complex is important for subsequent ubiquitination of Cln2 and Cln3.5,6 Recent evidence suggests that phosphorylated Cln1 and Cln2 are ubiquitinated by a Ubc/recognin complex formed by Ub-conjugating enzyme Ubc3/Cdc34 and (SCFGrr1) which is composed of Skp1, cullin (Cdc53) and an F-box protein (Grr1)62,63,66,98,99,100,101,102. A related complex, SCFCdc4, containing the F-box protein Cdc4 instead of Grr1 is required for degradation of budding yeast Clb/Cdk inhibitor Sic159,62 and of Cdc6.60 Degradation of Sic1 at the G1-S transition triggers initiation of DNA synthesis.103 Phosphorylation of Sic1 by Cln/Cdc28 is required for its binding to Cdc4 and its ubiquitination by SCFCdc4 and Cdc34.57 Cdc6 (as its fission yeast homolog Cdc18) is also required for initiation of DNA replication. The first 47 amino acids contain a high score PEST sequence essential for Cdc6 degradation via SCFCdc4. The same 47 amino acids showed two-hybrid interaction with Cdc4 but were unable to confer instability when fused to a b-gal reporter. These data suggested that additional sequences beside the PEST element-containing N-terminus are essential for degradation of Cdc6. This assumption was supported by the identification of a point mutation within the C-terminal half that resulted in a significant stabilization of Cdc6.60

Within the new class of SCF-type ligases, the ability to recognize primary destruction signals appears to reside in the F-box proteins (Grr1, Cdc4 and Met30).59,60,61,62. In the substrates for which this has been studied in detail (Cln1, Cln2, Sic1, Gcn4) the destruction signal recognized by these F-box proteins involve PEST sequences and phosphorylation.6,55,57,59,62

The role of phosphorylation in regulating destruction signals

As indicated in the previous section there are a number of substrates whose ubiquitination is preceded by phosphorylation. A few other important examples should be mentioned briefly.

Members of the Rel family of transcriptional activators as NF-kB and Dorsal are controlled by the Ub/proteasome pathway in several ways (for a review see reference 104). They are inhibited by ankyrin-repeat proteins such as IkBa, IkBb or Cactus that prevent their translocation into the nucleus. Signal-induced phosphorylation of these inhibitors, which bear C-terminal PEST elements triggers their ubiquitin-mediated degradation. As a result the liberated activators can enter the nucleus and mediate transcription of their target genes. Ubiquitination and therefore degradation of IkBa was recently shown to be inhibited by its conjugation to the small ubiquitin-like protein SUMO-1 (also called Sentrin, GMP1, SMT3, UBL1 or PIC1). SUMO-1 is attached to the same Lys residue in IkBa that is also used for ubiquitination thereby making it resistant to degradation.105 Generation of the p50 subunit of NF-kB from a p105 precursor is also mediated by the proteasome.106 Processing requires a glycine-rich region upstream of the processing site and has been proposed to occur cotranslationally suggesting that the proteasome is attracted by a destruction signal that is exposed in the nascent chain but obstructed in the completed polypeptide.107

The large subunit of RNA polymerase II has been shown to be phosphorylated on its C-terminal domain followed by ubiquitination and degradation after UV-radiation. This domain is formed by heptapeptide repeats (consensus: SPTSPSY) rich in PEST amino acids. In yeast, the ubiquitin ligase Rsp5 is involved in ubiquitination of this polymerase subunit.108,109

Phosphorylation also appears to be the trigger for ubiquitin-mediated endocytosis of plasma membrane proteins110,111 (R. Kölling, personal communication). Surprizingly, the ubiquitination of the uracil permease Fur4p requires the activity of the same ligase (Rsp5/Npi1) that mediates tagging of RNA polymerase II.110

There are also examples of an inhibition of proteolysis by phosphorylation. In Xenopus, the Mos protein kinase controls entry of immature oocytes into the meiotic cell cycle and is also involved in arresting mature oocytes at metaphase II. Upon fertilization or activation of the eggs, Mos is rapidly degraded allowing cells to enter a mitotic division cycle.112 An essential determinant of Mos degradation is an N-terminal Pro residue suggesting that the N-end rule pathway may be involved. This degradation signal is regulated by phosphorylation of a Ser residue immediately following the N-terminal Pro. Ser-2 phosphorylation in arrested mature oocytes inhibits ubiquitin-mediated proteolysis of Mos. Rapid dephosphorylation is the trigger for its degradation upon fertilization.112

Mos is also a proto-oncogene. Ectopic expression of Mos induces a mitogen-activated protein kinase cascade ultimately resulting in the phosphorylation of the transcriptional activator c-Jun.113 The phosphorylation on specific Ser residues within a PEST rich region results in a stabilization of c-Jun which appears to be responsible for the transforming effect of Mos induction.

In conclusion, phosphorylation appears to be a widespread strategy for regulating destruction signals, either by activating or by inhibiting them.

Amphipathic and hydrophobic domains as destruction signals

The Mata2 repressor controls cell type-specific genes in budding yeast. This transcriptonal regulator contains at least two destruction signals, one in each of its two domains, that are recognized by different pathways. The N-terminal 67 amino acid residues (DEG1) are sufficient to mediate degradation of a fusion protein via a pathway involving the Ubc6 and Ubc7 enzymes.50 Both enzymes are associated with the ER membrane and therefore with the nuclear envelope, a property that was found to be required for their function in the DOA (degradation of alpha2) pathway.114,115 Mata2 represses a-specific genes in haploid a cells and, together with Mata1, haploid-specific genes in diploid cells.50 Recent studies revealed that both Mata1 and Mata2 are highly unstable in haploid cells but are stabilized in diploids due to the formation of Mata1/Mata2 complexes. A close inspection of the Mata1/Mata2 interacting domains, which in Mata2 overlap with DEG1, revealed that they are characterized by amphipathic helices116,117 whose hydrophobic face apparently represent destruction signals that are masked upon complex formation.118

The intriguing mechanism for cell-type specific control of repressor function via destruction, or its inhibition by complex formation could be viewed as specialized case of a common theme. Subunits of complexes missing their partners are usually rapidly degraded. Gottesman and Maurizi45 hypothesized that the amphipathic nature of the exposed surfaces in dissociated proteins might be recognized as destruction signals. Finley and co-workers117 who studied synthetic signals in budding yeast came to the same conclusion based on their observation that an amphipathic helix likely to be formed by one of their signals was responsible for targeting a test protein for ubiquitin-mediated destruction.

Misfolded or otherwise abnormal proteins probably expose similar determinants or hydrophobic patches usually buried in the correctly folded structure that result in their ubiquitin-mediated destruction.45,117,119,120

Selection of ubiquitination site

The presence of an appropriately positioned Lys residue to which ubiquitin can be attached appeared to be an essential determinant of the N-end rule degradation signal of certain substrates. Elimination of this determinant by mutation of critical Lys residues resulted in the stabilization of the substrates used in these studies.29 However, when triose phosphate isomerase from S. cerevisiae was modified to have a destabilizing residue at its N-terminus, it was found that elimination of single or multiple possible ubiquitin accepting Lys residues (mutated to Arg) did not result in a strong stabilization of such mutants (Dohmen and Varshavsky, unpublished data). Similar results were obtained with substrates of other targeting pathways suggesting that the ubiquitination complexes involved are quite promiscuous with respect to the selection of the ubiquitination site.55,70,121

At this point in time it can only be speculated whether it is the spatial proximity or the regional structural flexibility of a given Lys residue that makes it a suitable site for ubiquitination. The mechanism that allows the utilization of various ubiquitin-accepting Lys residues on certain substrates is also unclear. One possibility is that recognition/targeting complexes scan a protein directionally for an appropriate Lys residue, in which case the first one to be encountered would be used ('Scanning model').2 It is also conceivable that, in a substrate that is binding to a recognin via its destruction signal, different Lys residues have varying statistic probabilities, due to thermal fluctuations, to get into close proximity of the active site of the targeting complex ('Stochastic Capture model').2,70

Trans targeting

The fact that a b-galactosidase test substrate with a destabilizing N-terminal residue was stable after mutation of the two ubiquitination sites close to its N-terminus29 enabled the discovery of "trans targeting", an intriguing mechanism with important implications.

As indicated ealier, a primary destruction signal can be a complex structural feature that may even involve more than one polypeptide (quaternary structure). One potential example of such a situation is that the two determinants of a primary signal, namely a feature that is bound by a recognin and the ubiquitin acceptor site, are located on two different polypeptides of a complex. That this is indeed possible was demonstrated by Johnson et al.122 who could show that a polypeptide bearing the first determinant of a destruction signal, i.e. a destabilizing N-terminal residue, can target an interacting polypeptide for ubiquitination (trans targeting) that lacks the first determinant but contains an ubiquitin accepting internal Lys residue (Fig. 1B). If the subunit with the destabilizing N-terminus itself does not contain a suitable Lys residue, it will not be ubiquitinated and hence not be degraded. Such a trans targeting polypeptide may thus act catalytically. The subunit-selective nature of ubiquitin-mediated proteolysis underlying this process has already occurred in the context of proteolytic removal of cyclins from cyclin/Cdk complexes (see also reference 50).

Another fascinating example of trans targeting is that of p53 targeting by E6 protein of oncogenic human papilloma virus (for details see chapter by M. Scheffner). E6 binds to the p53 tumor suppressor protein and to E6-AP, the latter being a cellular ubiquitin-protein ligase. The viral E6 thereby brings E6-AP and p53 together. As a result p53 is ubiquitinated by an E6AP/Ubc complex and degraded by the proteasome.123 Another example of trans targeting is that of c-Fos by c-Jun. These proteins form a heterodimeric transcription factor involved in controlling cell proliferation. Degradation of c-Fos is triggered by phosphorylation of the interacting c-Jun.124

The possibility of utilizing trans targeting proteins that can recruit Ubc/ligase complexes to substrates lacking determinants that would directly mediate binding to these complexes provides an additional level of controlling substrate specificity of the targeting machinery. Trans targeting or 'ligase recruiting proteins' may provide a means to recruit a limited number of Ubc/ligase complexes for targeting of a greater variety of proteolysis substrates under a greater variety of physiological conditions. An example of such a modular system are the recently discovered SCF-type recognins (see above). In this view, Cdc34, Skp1 and Cdc53 would form a Ubc/ligase complex that is recruited for ubiquitination of different types of substrates by different F-box proteins, such as Grr1, Cdc4 or Met30 (Fig. 1D).61

Ubiquitin as a degradation signal

Work by Varshavsky and coworkers22,28,125 lead to the discovery that a fusion protein consisting of Ub and another protein is degraded by the Ub/proteasome pathway. Within such a fusion protein, Ub serves as a primary degradation signal to which additional Ub moieties are attached by enzymes of the UFD pathway ('Ubiquitin fusion degradation'). Physiological substrates of the UFD pathway are not known to date. One possibility is that this pathway is involved in the degradation of Ub-like proteins or substrates that are modified by the attachment of such proteins. In the yeast S. cerevisiae, in which the UFD pathway has been studied, there are two proteins containing Ub-like domains, Rad23 and Dsk2, that are involved in DNA repair and function of the spindle pole body, respectively.126,127 These proteins are not posttranslationally attached to other proteins since they lack the diglycine motif at the C-terminus that is essential for conjugation.128 Rad23 was shown to interact directly with the 26S proteasome suggesting a function of the proteasome in DNA repair.129 In addition, variants of Rad23 were reported to be ubiquitinated and degraded by the proteasome.129 Degradation, however, was not inhibited by mutations that are specific for the UFD pathway22 (K. Madura, personal communication). Several other Ub-like proteins, UCRP, Smt3 (also called Sentrin, SUMO, GMP or PIC1) and Rub1 (homolog of NEDD8), have recently been shown to be posttranslationally attached to other proteins (for review see references 130-132). To date it has not been reported for any of these Ub-like proteins or their substrates that they are modified by the attachment of Ub and thereby targeted for degradation. The physiological relevance of the UFD pathway and the role of Ub (or possibly Ub-like proteins) as primary destruction signals thus remain unclear.

Degradation of non-ubiquitinated proteins by the proteasome

An interesting substrate of proteasomal degradation is ornithine decarboxylase (ODC), an enzyme involved in polyamine biosynthesis. As a result of a variety of stimuli ODC becomes rapidly degraded (for review see reference 133). Degradation of ODC, normally a homodimeric protein, is induced by its interaction with a protein called antizyme that binds to ODC monomers.63 Synthesis of antizyme is controlled by a sophisticated regulatory mechanism involving a ribosomal frame shift mechanism that is modulated by changes in polyamine concentration.134 Binding of antizyme results in the destruction of ODC monomers by the 26S proteasome. According to in vivo and in vitro studies, this degradation does not involve ubiquitination.135,136,137 Apparently, the exposure of the C-terminal ODC destruction signal and its presentation by antizyme is sufficient to make ODC a substrate of the proteasome.138 In another in vitro study it was reported that ubiquitination is not an absolute requirement for degradation of the c-Jun oncoprotein by the 26S proteasome.139

These studies suggest that the 26S proteasome has the capacity to degrade certain non-ubiquitinated substrates, which could either be directly recognized by the 19S regulatory cap or be presented to the proteasome by co-factors like antizyme in the case of ODC or possibly by chaperones in other cases.

Applying destruction signals

A detailed knowledge of the nature of destruction signals and the intracellular mechanism of their recognition can be applied to experimental and possibly pharmaceutical purposes. Several approaches have taken advantage of the N-degron for the generation of conditional null mutants in budding yeast.4,140,141 Varshavsky suggested that destruction signals could be utilized as features of drugs.142 A peptide drug could be designed that is rapidly degraded and therefore non-toxic in normal cells, but is stabilized in abnormal cells by specific binding to one or more proteins that are unusually abundant in these cells. That such an approach is possible in principle has been demonstrated recently.143 A diphtheria toxin modified to have a destabilizing N-terminus was taken up by Vero cells and became a substrate of the N-end rule pathway in the cytosol of these cells. The degradation of this modified toxin in an in vitro system was inhibited by an antibody specifically binding to it.

Conclusion

Due to the increased awareness of selective protein degradation as an important regulatory mechanism, the number of identified substrates that are regulated by proteolysis is increasing nearly every day. Analyses bearing on the nature of the primary destruction signals of such substrates, however, are still at an early stage. What seems to be emerging is that there are a variety of oftentimes complex destruction signals specific for certain proteolysis substrates or substrate classes. For each type of destruction signal there appear to be specific recognition proteins that mediate attachment of ubiquitin chains as secondary destruction signals targeting the substrates for degradation by the proteasome. Some of the destruction signals, such as cyclin-type destruction boxes, PEST sequences, and phosphorylated regions appear to be characteristic for larger sets of proteolysis substrates. Some of these signals are subject to posttranslational regulation, in other cases the targeting machinery appears to be regulated, both mechanisms allowing for adjustment of proteins' stabilities to the physiological states of the cell. Precise definitions of primary destruction signals and the mode of their recognition are challenging topics for important studies to come. The results will not only help to explain vital regulatory mechanisms in cell biology but might in addition provide useful tools for the treatment of disease.

Acknowledgments

I am grateful to Katrina Cooper, Randy Strich, Alfred Goldberg, Mark Hochstrasser, Ralf Kölling and Kiran Madura for communicating results prior to publication, and to Erica Johnson, Ralf Kölling, Paula Ramos for comments on the manuscript.

Note added in proof:

While this manuscript was in the editing process, several studies have led to the identification of zinc-binding ring-H2 domains as a common feature of several otherwise diverse ubiquitin ligases such as SCF complexes, APC and Ubr1. The role of this domain has been reviewed recently (Deshaies, R.J. SCF and Cullin/Ring H2 based ubiquitin ligases. Annu Rev Cell Dev Biol 1999; 15:435-467)

References

1. Gottesman S. Proteases and their targets in Escherichia coli. Annu Rev Genet 1996; 30:465-506.

2. Varshavsky A. The N-end rule. Cell 1992; 69:725-735.

3. Campenaro MR, Flemington EK. Regulation of E2F through ubiquitin-proteasome-dependent degradation: stabilization by the pRB tumor suppressor protein. Proc Natl Acad Sci USA 1997; 94:2221-2226.

4. Dohmen RJ, Wu P-P, Varshavsky A. Heat-inducible degron: a method for constructing temperature-sensitive mutants. Science 1994; 263:1273-1276.

5. Yaglom J, Linskens MHK, Sadis S et al. P34Cdc28-mediated control of Cln3 degradation. Mol Cell Biol 1995; 15:731-741.

6. Lanker S, Valdivieso MH, Wittenberg C. Rapid degradation of the G1 cyclin Cln2 induced by CDK-dependent phosphorylation. Science 1996; 271:1597-1600.

7. Cheng M, Olivier P, Diehl JA et al. The p21(Cip1) and p27(Kip1) CDK 'inhibitors' are essential activators of cyclin D-dependent kinases in murine fibroblasts. Genes Dev 1998; 12:3499-3511.

8. Freedman DA, Levine AJ. Nuclear Export is required for degradation of endogenous p53 by MDM2 and human papillomavirus E6. Mol Cell Biol 1998; 18:7288-7293.

9. Tomoda K, Kubota Y, Kato J. Degradation of the cyclin-dependent-kinase inhibitor p27Kip1 is instigated by Jab1. Nature 1999; 398:160-165.

10. Hershko A, Heller H. Occurance of a polyubiquitin structure in ubiquitin-protein conjugates. Biochem Biophys Res Commun 1985; 128:1079-1086.

11. Chau V, Tobias JW, Bachmair A et al. A multiubiquitin chain is confined to a specific lysine in a targeted short-lived protein. Science 1989; 243:1576-1583.

12. Dohmen RJ, Madura K, Bartel B et al. The N-end rule is mediated by the UBC2(RAD6) ubiquitin-conjugating enzyme. Proc Natl Acad Sci USA 1991; 88:7351-7355.

13. Madura K, Dohmen RJ, Varshavsky A. N-recognin/Ubc2 interactions in the N-end rule pathway. J Biol Chem 1993; 268:12046-12054.

14. Jentsch S. The ubiquitin-conjugation system. Ann Rev Genet 1992; 26:179-207.

15. Huibregtse JM, Scheffner M, Beaudenon S et al. A family of proteins structurally and functionally related to the E6-AP ubiquitin-protein ligase. Proc Natl Acad Sci USA 1995; 92:2563-2567.

16. Beal R, Deveraux Q, Xia G et al. Surface hydrophobic residues of multubiquitin chains essential for proteolytic targeting. Proc Natl Acad Sci USA 1996; 93:861-866.

17. Piotrowski J, Beal R, Hoffman L et al. Inhibition of the 26S proteasome by polyubiquitin chains synthesized to have a defined length. J Biol Chem 1997; 272:23712-23721.

18. Cook WJ, Jeffrey LC, Kasperek E et al. Structure of tetraubiquitin shows how multiubiquitin chains can be formed. J Mol Biol 1994; 236:601-609.

19. Pickart C. Targeting of substrates to the 26S proteasome. FASEB J 1997; 11:1055-1066.

20. Baboshina OV, Haas AL. Novel multiubiquitin chain linkages catalyzed by the conjugating enzymes E2EPF and Rad6 are recognized by 26S proteasome subunit 5. J Biol Chem 1996; 271:2823-2831.

21. Arnason T, Ellison MJ. Stress resistance in Saccharomyces cerevisiae is strongly correlated with assembly of a novel type of multiubiquitin chain. Mol Cell Biol 1994; 14:7876-7883.

22. Johnson ES, Ma P, Ota IM et al. A proteolytic pathway that recognizes ubiquitin as a degradation signal. J Biol Chem 1995; 270:17442-17456.

23. Teichert U, Mechler B, Müller H et al. Lysosomal (vacuolar) proteinases of yeast are essential catalysts for protein degradation, differentiation, and cell survival. J Biol Chem 1989; 264:16037-16045.

24. van den Hazel HB, Kiellan-Brandt MC, Winther JR. Biosynthesis and function of yeast vacuolar proteases. Yeast 1996; 12:1-16.

25. Hicke L. Ubiquitin-dependent internalization and down-regulation of plasma membrane proteins. FASEB J 1997; 11:1215-1226.

26. Dice JF. Peptide sequences that target cytosolic proteins for lysosomal proteolysis. Trends Biochem Sci 1990; 15:305-309.

27. Hershko A, Heller H, Eytan E et al. Role of the a-amino group of protein in ubiquitin-mediated protein breakdown. Proc Natl Acad Sci USA 1984; 81:7021-7025.

28. Bachmair A, Finley D, Varshavsky A. In vivo half-life of a protein is a function of its amino-terminal residue. Science 1986; 234:179-186.

29. Bachmair A, Varshavsky A. The degradation signal in a short-lived protein. Cell 1989; 56:1019-1032.

30. Bartel B, Wünning I, Varshavsky A. The recognition compenent of the N-end rule pathway. EMBO J 1990; 9:3179-3189.

31. Tobias JW, Shrader TE, Rocap G et al. The N-end rule in bacteria. Science 1991; 254:1374-1377.

32. Lévy F, Johnsson N, Rümenapf T et al. Using ubiquitin to follow the metabolic fate of a protein. Proc Natl Acad Sci USA 1996; 93:4907-4912.

33. Varshavsky A. The N-end rule: functions, mysteries, uses. Proc Natl Acad Sci USA 1996; 93:12142-12149.

34. Varshavsky A. The N-end rule. Cold Spring Harbor Symp Quant Biol 1996; 60:461-478.

35. Huang S, Elliott RC, Liu P-S et al. Specificity of cotranslational amino-terminal processing of proteins in yeast. Biochem 1987; 26:8242-8246.

36. Ben-Bassat A, Bauer K, Chang S-Y et al. Processing of the initiation methionine from proteins: properties of the Escherichia coli methionine aminopeptidase and its gene structure. J Bact 1987; 169:751-757.

37. Boisseel J-P, Kaper TJ, Bunn HF. Cotranslational amino-terminal processing of cytosolic proteins. J Biol Chem 1988; 263:8443-8449.

38. Moerschell RP, Hosokawa Y, Tsunasawa S et al. The specificities of yeast methionine aminopeptidase and acetylation of amino-terminal methionine in vivo. J Biol Chem 1990; 265:19638-19643.

39. Chang Y-H, Teichert U, Smith JA. Purification and characterization of a methionine aminopeptidase from Saccharomyces cerevisiae. J Biol Chem 1990; 265:19892-19897.

40. deGroot RJ, Rümenapf T, Kuhn RJ et al. Sindbis virus RNA polymerase is degraded by the N-end rule pathway. Proc Natl Acad Sci USA 1991; 88:8967-8971.

41. Madura K, Varshavsky A. Degradation of Ga by the N-end rule pathway. Science 1994; 265:1454-1458.

42. Byrd C, Turner GC, Varshavsky A. The N-end rule pathway controls the import of peptides through degradation of a transcriptional repressor. EMBO J 1998; 17:269-277.

43. Baker RT, Varshavsky A. Inhibition of the N-end rule pathway in living cells. Proc Natl Acad Sci USA 1991; 87:2374-2378.

44. Dean RT. A mechanism for accelerated degradation of intracellular proteins after limited damage by free radicals. FEBS Lett 1987; 220:278-282.

45. Gottesman S, Maurizi MR. Regulation by proteolysis: energy-dependent proteases and their targets. Microbiol Rev 1992; 56:592-621.

46. Solomon V, Lecker SH, Goldberg AL. The N-end rule pathway catalyzes a major fraction of the protein degradation in skeletal muscle. J Biol Chem 1998; 273:25216-25222.

47. Solomon V, Baracos V, Sarraf P et al. Rates of ubiquitin conjugation increase when muscle atrophy, largely through activation of the N-end rule pathway. Proc Natl Acad Sci USA 1998; 95:12602-12607.

48. Rogers S, Wells R, Rechsteiner M. Amino acid sequences common to rapidly degraded proteins: The PEST hypothesis. Science 1986; 234:364-368.

49. Rechsteiner M, Rogers SW. PEST sequences and regulation of proteolysis. Trends Biochem Sci 1996; 21:267-271

50. Hochstrasser M, Varshavsky A. In vivo degradation of a transcriptional regulator: the yeast a2 repressor. Cell 1990; 61:697-708.

51. Pakdel F, Le Goff P, Katzenellenbogen BS. An assessment of the role of domain F and PEST sequences in estrogen receptor half-life and bioactivity. J Steroid Biochem Mol Biol 1993; 46:663-672.

52. Nixon PJ, Komenda J, Barber J et al. Deletion of the PEST-like region of photosystem two modifies the QB-binding pocket but does not prevent rapid turnover of D1. J Biol Chem 1995; 270:14919-14927.

53. Salama SR, Hendricks KB, Thorner J. G1 cyclin degradation: the PEST motif of yeast Cln2 is necessary, but not sufficient, for rapid protein turnover. Mol Cell Biol 1994; 14:7953-7966.

54. King RW, Deshaies RJ, Peters JM et al. How proteolysis drives the cell cycle. Science 1996; 274:1652-1659.

55. Kornitzer D, Raboy B, Kulka RG et al. Regulated degradation of the transcription factor Gcn4. EMBO J 1994; 13:6021-6030.

56. Lin R, Beauparlant P, Makris C et al. Phosphorylation of IkBa in the C-terminal PEST domain by casein kinase II affects intrinsic protein stability. Mol Cell Biol 1996; 16:1401-1409.

57. Verma R, Annan RS, Huddleston MJ et al. Phosphorylation of Sic1p by G1 cyclin/Cdk required for its degradation and entry into S phase. Science 1997; 278:455-459.

58. Liu ZP, Galindo RL, Wasserman SA. A role for CKII phosphorylation of the cactus PEST domain in dorsoventral patterning of the Drosophola embryo. Genes Dev 1997; 11:3413-3422.

59. Feldmann RMR, Correll CC, Kaplan KB et al. A complex of Cdc4p, Skp1p and Cdc53p/Cullin catalyzes ubiquitination of the phosphorylated inhibitor Sic1p. Cell 1997; 91:221-230.

60. Drury LS, Perkins G, Diffley JFX. The Cdc4/34/53 pathway targets Cdc6p for proteolysis in budding yeast. EMBO J 1997; 16:5966-5976.

61. Patton EE, Willems AR, Sa D et al. Cdc53 is a scaffold protein for multiple Cdc34/Skp1/F-box protein complexes that regulate cell division and methionine biosynthesis in yeast. Genes Dev 1998; 12:692-705.

62. Skowyra D, Craig KL, Tyers M et al. F-box proteins are receptors that recruit phosphorylated substrates to SCF ubiquitin-ligase complex. Cell 1997; 91:209-219.

63. Li FN, Johnston M. Grr1 of S. cerevisiae is connected to the ubiquitination machinery through Skp1: coupling glucose sensing to gene expression and the cell cycle. EMBO J 1997; 16:5629-5638.

64. Evans T, Rosenthal ET, Youngblom J et al. Cyclin: a protein specified by maternal mRNA in sea urchin eggs that is destroyed at each cleavage division. Cell 1983; 33:389-396.

65. Murray AW, Kirschner MW. Dominoes and Clocks: the union of two views of the cell cycle. Science 1998; 24:614-621.

66. Deshaies RJ, Chau V, Kirschner M. Ubiquitination of the G1 cyclin Cln2p by a Cdc34p-dependent pathway. EMBO J 1995; 14:303-312.

67. Hershko A. Roles of ubiquitin-mediated proteolysis in cell cycle control. Curr Opin Cell Biol 1997; 9:788-799.

68. Nasmyth K. Control of the yeast cell cycle by the Cdc28 protein kinase. Curr Opin Cell Biol 1993; 5:166-179.

69. Glotzer M, Murray AW, Kirschner MW. Cyclin is degraded by the ubiquitin pathway. Nature 1991; 349:132-138.

70. King RW, Glotzer M, Kirschner MW. Mutagenic analysis of the destruction signal of mitotic cyclins and structural characterization of ubiquitinated intermediates. Mol Biol Cell 1996; 7:1343-1357.

71. Yamano H, Tsurumi C, Gannon J et al. The role of the destruction box and its neighbouring lysine residues in cyclin B for anaphase ubiquitin-dependent proteolysis in fission yeast: defining the D-box receptor. EMBO J 1998; 17:5670-5678.

72. Klotzbücher A, Stewart A, Harrison D et al. The "destruction" box of cyclin A allows B-type cyclins to be ubiquitinated, but not efficiently destroyed. EMBO J 1996; 15:3053-3064.

73. Stewart E, Kobayashi H, Harrison D et al. Destruction of Xenopus cyclins A and B2, but not B1, requires binding to p34cdc2. EMBO J 1994; 13:584-594.

74. Minshull J, Golsteyn R, Hill CS et al. The A- and B-type cyclin-associated cdc2 kinases in Xenopus turn on and off at different times in the cell cycle. EMBO J 1990; 9:2865-2875.

75. Whitfield WG, GonzaleZ C, Maldonado CG et al. The A- and B-type cyclins of Drosophila are accumulated and destroyed in temporally distinct events that define separable phases of the G2-M transition. EMBO J 1990; 9:2563-2572.

76. Hunt T, Luca FC, Ruderman JV. The requirements for protein synthesis and degradation and the control of destruction of cyclins A and B in the meiotic and mitotic cycles of the clam embryo. J Cell Biol 1992; 116:707-724.

77. Sudakin V, Ganoth D, Dahan A et al. The cyclosome, a large complex containing cyclin-selective ubiquitin ligase activity, targets cyclins for destruction at the end of mitosis. Mol Biol Cell 1995; 6:185-198.

78. Irniger S, Piatti S, Michaelis C et al. Genes involved in sister chromatid separation are needed for B-type cyclin proteolysis in budding yeast. Cell 1995; 81:269-278.

79. King RW, Peters JM, Tugendreich S et al. A 20S complex containing CDC27 and CDC16 catalyzes the mitosis-specific conjugation of ubiquitin to cyclin B. Cell 1995; 81:279-288.

80. Zachariae W, Schwab M, Nasmyth K et al. Control of cyclin ubiquitination by CDK-regulated binding of Hct1 to the anaphase promoting complex. Science 1998; 282:1721-1724.

81. Yamamoto A, Guacci V, Koshland D. Pds1p is required for faithful execution of anaphase in the yeast Saccharomyces cerevisiae. J Cell Biol 1996; 133:85-97.

82. Yamamoto A, Guacci V, Koshland D. Pds1p, an inhibitor of anaphase in budding yeast, plays a critical role in the APC and checkpoint pathway(s). J Cell Biol 1996; 133:99-110.

83. Cohen-Fix O, Peters J-M, Kirschner MW et al. Anaphase initiation in Saccharomyces cerevisiae is controlled by the APC-dependent degradation of the anaphase inhibitor Pds1p. Genes Dev 1997; 10:3081-3093.

84. Funabiki H, Yamano H, Kumada K et al. Cut2 proteolysis required for sister-chromatid separation in fission yeast. Nature 1996; 381:438-441.

85. Michaelis C, Ciosk R, Nasmyth K. Cohesins: chromosomal proteins that prevent premature separation of sister chromatids. Cell 1997; 91:35-45.

86. Guacci V, Koshland D, Strunnikov A. A direct link between sister chromatid cohesion and chromosome condensation revealed through the analysis of MCD1 in S. cerevisiae. Cell 1997; 91:47-57.

87. Juang Y-L, Huang J, Peters J-M et al. APC-mediated proteolysis of Ase1 and the morphogenesis of the mitotic spindle. Science 1997; 275:1311-1314.

88. Shirayama M, Zachariae W, Ciosk R et al. The Polo-like kinase Cdc5p and the WD-repeat protein Cdc20p/fizzy are regulators and substrates of the anaphase promoting complex in Saccharomyces cerevisiae. EMBO J 1998; 17:1336-1349.

89. Charles JF, Jaspersen SL, Tinker-Kulberg RL et al. The Polo-related kinase Cdc5 activates and is destroyed by the mitotic cyclin destruction machinery in S. cerevisiae. Curr Biol 1998; 8:497-507.

90. Visintin R, Prinz S, Amon A. CDC20 and CDH1: a family of substrate-specific activators of APC-dependent proteolysis. Science 1997; 278:460-463.

91. Schwab M, Lutum AS, Seufert W. Yeast Hct1 is a regulator of Clb2 cyclin proteolysis. Cell 1997; 90:683-693.

92. Lahav-Baratz S, Sudakin V, Ruderman JV et al. Reversible phosphorylation controls the activity of cyclosome-associated cyclin-ubiqutin ligase. Proc Natl Acad Sci USA 1995; 92:9303-9307.

93. Andrews B, Measday V. The cyclin family of budding yeast: abundant use of a good idea. Trends Genet 1998; 14:66-72.

94. Liao S-M, Zhang J, Jeffery DA et al. A kinase-cyclin pair in the RNA polymerase II holoenzyme. Nature 1995; 374:193-196.

95. Kuchin S, Yeghiayan P, Carlson M. Cyclin-dependent protein kinase and cyclin homologs SSN3 and SSN8 contribute to transcriptional control in yeast. Proc Natl Acad Sci USA 1995; 92:4006-4010.

96. Cooper KF, Mallory MJ, Smith JB et al. Stress and developmental regulation of the yeast C-type cyclin Ume3p (Srb11p/Ssn8p). EMBO J 1997; 16:4665-4675.

97. Kaplon T, Jacquet M. The cellular content of Cdc25p, the Ras exchange factor in Saccharomyces cerevisiae, is regulated by destabilization through a cyclin destruction box. J Biol Chem 1995; 270:20742-20747.

98. Bai C, Sen P, Hofmann K et al. SKP1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box. Cell 1996; 86:263-274.

99. Barral Y, Jentsch S, Mann C. G(1) cyclin turnover and nutrient-uptake are controlled by a common pathway in yeast. Genes Dev 1995; 9:275-285.

100. Willems AR, Lanker S, Patton E et al. Cdc53 targets phosphorylated G1 cyclins for degradation by the ubiquitin proteolytic pathway. Cell 1996; 86:453-463.

101. Deshaies RJ. Phosphorylation and proteolysis: partners in the regulation of cell division in budding yeast. Curr Opin Genet Dev 1997; 7:7-16.

102. Mathias N, Steussy CN, Goebl MG. An essential domain within Cdc34p is required for binding to a complex containing Cdc4p and Cdc53p in Saccharomyces cerevisiae. J Biol Chem 1998; 273:4040-4045.

103. Schwob E, Böhm T, Mendenhall MD et al. The B-type cyclin inhibitor p40SIC1 controls the G1 to S transition in S. cerevisiae. Cell 1994; 79:233-244.

104. Thanos D, Maniatis T. NF-kB: a lesson in family values. Cell 1995; 80:529-532.

105. Desterro JMP, Rodriguez MS, Hay RT. SUMO-1 modification of IkBa inhibits NF-kB activation. Mol Cell 1998; 2:233-239.

106. Palombella VJ, Rando OJ, Goldberg AL et al. The ubiquitin-proteasome pathway is required for processing the NF-kB1 precursor protein and the activation of NF-kB. Cell 1994; 78:773-785.

107. Lin L, DeMartino GN, Greene WC. Cotranslational biogenesis of NF-kB p50 by the proteasome. Cell 1998; 92:819-828.

108. Huibregtse JM, Yang JC, Beaudenon SL. The large subunit of RNA polymerase II is a substrate of the Rsp5 ubiquitin-protein ligase. Proc Natl Acad Sci USA 1997; 94:3656-3661.

109. Ratner JN, Balasubramanian B, Corden J et al. Ultraviolet radiation-induced ubiquitination and proteasomal degradation of the large subunit of RNA polymerase II. Implications for transcription-coupled DNA repair. J Biol Chem 1998; 273:5184-5189.

110. Marchal C, Haguenauer-Tsapis R, Urban-Grimal D. A PEST-like Sequence mediates phosphorylation and efficient ubiquitination of yeast uracil permease. Mol Cell Biol 1998; 18:314-321.

111. Hicke L, Zanolari B, Riezman H. Cytoplasmic tail phosphorylation of the a-factor receptor is required for its ubiquitination and internalization. J Cell Biol 1998; 141:349-358.

112. Nishizawa M, Okazaki K, Furuno N et al. Degradation of Mos by the N-terminal (Pro2)-dependent ubiquitin pathway on fertilization of Xenopus eggs: possible significance of natural selection for Pro2 in Mos. EMBO J 1993; 11:2433-2446.

113. Okazaki K, Sagata N. The Mos/MAP kinase pathway stabilizes c-Fos by phosphorylation and augments its transforming activity in NIH 3T3 cells. EMBO J 1995; 14:5048-5059.

114. Chen P, Johnson P, Sommer T et al. Multiple ubiquitin-conjugating enzymes participitate in the in vivo degradation of yeast MATa2 repressor. Cell 1993; 74:357-369.

115. Biederer T, Volkwein C, Sommer T. Role of Cue1p in ubiquitination and degradation at the ER surface. Science 1997; 278:1806-1809.

116. Ho C-H, Adamson JG, Hodges RS et al. Heterodimerization of the yeast MATa1 and MATa2 proteins is mediated by two leucine zipper-like coiled-coil motifs. EMBO J 1994; 13:1403-1413.

117. Sadis S, Atienza C, Finley D. Synthetic signals for ubiquitin-dependent proteolysis. Mol Cell Biol 1995; 15:4086-4094.

118. Johnson P, Swanson R, Rahhilina L et al. Degradation signal masking by heterodimerization of MATa2 and MATa1 blocks their mutual destruction by the ubiquitin-proteasome pathway. Cell 1998; 94:217-227.

119. Pacifici RE, Kono, Y, Davies KJA. Hydrophobicity as the signal for selective degradation of hydroxyl radical-modified hemoglobin by the multicatalytic proteinase complex, proteasome. J Biol Chem 1993; 268:15405-15411.

120. Gilon T, Chomsky O, Kulka RG. Degradation signals for ubiquitin system proteolysis in Saccharomyces cerevisiae. EMBO J 1998; 17:2759-2766.

121. Treier M, Staszewski LM, Bohmann D. Ubiquitin-dependent c-Jun degradation in vivo is mediated by the delta domain. Cell 1984; 78:787-798.

122. Johnson ES, Gonda DK, Varshavsky A. Cis-trans recognition and subunit-specific degradation of short-lived proteins. Nature 1990; 346:287-291.

123. Scheffner M, Huibregtse JM, Vierstra RD et al. The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell 1993; 75:495-505.

124. Papavassiliou AG, Treier M, Chavrier C et al. Targeted degradation of c-Fos, but not v-Fos, by phosphorylation-dependent signal on c-Jun. Science 1992; 258:1941-1944.

125. Johnson ES, Bartel B, Seufert W et al. Ubiquitin as a degradation signal. EMBO J 1992; 11:497-505.

126. Watkins JF, Sung P, Prakash L et al. The Saccharomyces cerevisiae DNA repair gene RAD23 encodes a nuclear protein containing a ubiquitin-like domain required for its biological function. Mol Cell Biol 1993; 13:7757-7765.

127. Biggins S, Ivanovska I, M.D. R. Yeast ubiquitin-like genes are involved in duplication of the microtubule organizing center. J Cell Biol 1996; 133:1331-1346.

128. Johnson ES, Schwienhorst I, Dohmen RJ et al. The ubiquitin-like protein Smt3p is activated for conjugation to other proteins by an Aos1p/Uba2p heterodimer. EMBO J 1997; 16:5509-5519.

129. Schauber C, Chen L, Tongaonkar P et al. Rad23 links DNA repair to the ubiquitin/proteasome pathway. Nature 1998; 715:715-718.

130. Saitoh H, Pu RT, Dasso M. SUMO-1: wrestling with a new ubiquitin-related modifier. Trends Biochem Sci 1997; 22:374-376.

131. Johnson PR, Hochstrasser M. SUMO-1: ubiquitin gains weight. Trends Cell Biol 1997; 7:408-413.

132. Hochstrasser M. There's the Rub: a novel ubiquitin-like modification linked to the cell cycle regulation. Genes Dev 1998; 12:901-907.

133. Hayashi S, Murakiami Y, Matsufuji S. Ornithine decarcoxylase: a noval type of regulatory protein. Trends Biochem Sci 1996; 21:27-30.

134. Matsufuji M, Matsufuji T, Miyazaki Y et al. Autoregulatory frameshifting in decoding mammalian ornithine decarboxylase antizyme. Cell 1995; 80:51-60.

135. Rosenberg-Hasson Y, Bercovich Z, Ciechanover A et al. Degradation of ornithine decarboxylase in mammalian cells is ATP dependent but ubiquitin independent. Eur J Biochem 1989; 185:469-474.

136. Bercovich Z, Rosenberg-Hasson Y, Ciechanover A et al. Degradation of ornithine decarboxylase in reticulocyte lysate is ATP dependent but ubiquitin independent. J Biol Chem 1989; 264:15949-15952.

137. Murakami Y, Matsufuji S, Kameji T et al. Ornithine decarboxylase is degraded by the 26S proteasome without ubiquitination. Nature 1992; 360:597-599.

138. Ghoda L, Sidney D, Macrae M et al. Structural elements of ornithine decarboxylase required for intracellular degradation and polyamine-dependent regulation. Mol Cell Biol 1992; 12:2178-2185.

139. Jariel-Encontre I, Pariat M, Martin F et al. Ubiquitinylation is not an absolute requirement for degradation of c-Jun by the 26S proteasome. J Biol Chem 1995; 270:11623-11627.

140. Park E-U, Finley D, Szostak JW. A strategy for the generation of conditional mutations by protein destabilization. Proc Natl Acad Sci USA 1992; 89:1249-1252.

141. Moqtaderi Z, Bai Y, Poon D et al. TBP-associated factors are not generally required for transcriptional activation in yeast. Nature 1996; 383:188-191.

142. Varshavsky A. Codominance and toxins: a path to drugs of nearly unlimited selectivity. Proc Natl Acad Sci USA 1995; 92:3663-3667.

143. Falnes PØ, Olsnes S. Modulation of the intracellular stability and toxicity of diphteria toxin through degradation by the N-end rule pathway. EMBO J 1998; 17:615-625.

Abbreviations used in this chapter: APC, Anaphase Promoting complex; Cdk, cyclin-dependent kinase; DOA, degradation of alpha2; E2, Ub-conjugating enzyme (Ubc); E3, ubiquitin protein ligase; ER, endoplamic reticulum; GEF, guanine nucleotide exchange facor; MAP, methionine amino peptidase; ODC, ornithine decarboxylase; SCF, complex formed by Skp1p, Cdc53/cullin and F-box protein; Ub, ubiquitin; Ubp, Ub processing protease; UFD, Ub fusion degradation;

Table 1: Primary destruction signals.

species	destabilizing N-terminal residues
E. coli	F,L,W,Y,R,K
S. cerevisiae	F,L,W,Y,R,K,H,I,N,Q,D,E
Mammallian cells	F,L,W,Y,R,K,H,I,N,Q,D,E,C,A,T,(S),(P)

cyclin destrucion box	(Pos.) amino acid sequence
cyclin B (A. punctulata, sea urchin)	(42) RAALGNISN
cyclin B2 (X. laevis)	(30) RAALGEIGN
cyclin B1 (X. laevis)	(36) RTALGDIGN
cyclin B1 (H. sapiens)	(42) RTALGDIGN
cdc13 (S. pombe)	(59) RHALDDVSN
Clb2p (S. cerevisiae)	(25) RLALNNVTN
cyclin A1 (X. laevis)	(41) RTVLGVIGD
cyclin A2 (X. laevis)	(26) RTVLGVLQE
cyclin A (H. sapiens)	(47) RAALAVLKS
Cdc5p (S. cerevisiae)	(17) RSKLVHTPI (61) REKLSALCK
Cdc20p (S. cerevisiae)	(17) RSVLSIASP (60) RPSLQASAN
Pds1p (S. cerevisiae)	(85) RLPLAAKDN
Ase1p (S. cerevisiae)	(760) RQLFPIPLN
Cut2 (S. pombe)	(33) RAPLGSTKQ (52) RTVLGGKST
Ume3p (S. cerevisiae)	(25) RQKLWLLEC
Cdc25p (S. cerevisiae)	(148) RSSLNSLGN

Table 1: Primary destruction signals. Upper part, amino acid residues that are destabilizing in the N-end rule. Ser (S) was destabilizing in rabbit reticulocyte extracts but not in mouse L-cells.32 Pro (P) at the N-terminus was shown to be an essential destruction signal of cMos,112 but it was not destabilizing in the context of other test substrates.32 Lower part; amino acid sequences of cyclin destruction boxes (CDB) and CDB-like sequences required for degradation of the proteins listed. Residues that have been shown to be important for a function as destruction signal are in bold face. The positions of the first Arg (R) residues of CDBs are given relative to the N-termini of the respective proteins.

Fig.1: Targeting of proteolytic substrates.

Fig.1: Targeting of proteolytic substrates. A, the Ubr1 ligase is drawn with two binding sites, I for basic residues, II for bulky hydrophobic residues, and in association with N-terminal amidase (Nta1) and Arg-tRNA-protein transferase (Ate1p) which modify proteins with certain N-termini.34 Rad6/Ubc2 interacts with Ubr1 via an acidic extension at its C terminus. A question mark indicates that it is unclear whether Ub is transferred from Ubc to which it is linked via a thioester bond or whether transfer involves an intermediate thioester between Ub and Ubr1. B, same as in a A, but with the two determinants of the destruction signal, the N-terminal destabilizing residue (d) and the ubiquitin accepting Lys residue, on different polypeptides of a multimeric protein. C, targeting of proteins (examples listed in the lower right corner) with cyclin destruction boxes (CDB) by APC that is activated by Cdc5 and Hct1, or by Cdc20, and regulated by Cdk-mediated phosphorylation . Question marks indicate that the exact role of the latter proteins in activation of APC and the nature of the subunit recognizing the CDB are not clear to date. D, targeting of phosphorylated substrates containing PEST sequences (examples in the lower right corner) by a complex of Cdc34/Ubc3 and SCF (Skp1, cullin and F-box protein).