Localization to the Proteasome Is Sufficient for Degradation

Abstract

The majority of unstable proteins in eukaryotic cells are targeted for degradation through the ubiquitin-proteasome pathway. Substrates for degradation are recognized by the E1, E2, and E3 ubiquitin conjugation machinery and tagged with polyubiquitin chains, which are thought to promote the proteolytic process through their binding with the proteasome. We describe a method to bypass the ubiquitination step artificially both in vivo and in a purified in vitro system. Seven proteasome subunits were tagged with Fpr1, and fusion reporter constructs were created with the Fpr1-rapamycin binding domain of Tor1. Reporter proteins were localized to the proteasome by the addition of rapamycin, a drug that heterodimerizes Fpr1 and Tor1. Degradation of reporter proteins was observed with proteasomes that had either Rpn10 or Pre10 subunits tagged with Fpr1. Our experiments resolved a simple but central problem concerning the design of the ubiquitin-proteasome pathway. We conclude that localization to the proteasome is sufficient for degradation and, therefore, any added functions polyubiquitin chains possess beyond tethering substrates to the proteasome are not strictly necessary for proteolysis. The majority of unstable proteins in eukaryotic cells are targeted for degradation through the ubiquitin-proteasome pathway. Substrates for degradation are recognized by the E1, E2, and E3 ubiquitin conjugation machinery and tagged with polyubiquitin chains, which are thought to promote the proteolytic process through their binding with the proteasome. We describe a method to bypass the ubiquitination step artificially both in vivo and in a purified in vitro system. Seven proteasome subunits were tagged with Fpr1, and fusion reporter constructs were created with the Fpr1-rapamycin binding domain of Tor1. Reporter proteins were localized to the proteasome by the addition of rapamycin, a drug that heterodimerizes Fpr1 and Tor1. Degradation of reporter proteins was observed with proteasomes that had either Rpn10 or Pre10 subunits tagged with Fpr1. Our experiments resolved a simple but central problem concerning the design of the ubiquitin-proteasome pathway. We conclude that localization to the proteasome is sufficient for degradation and, therefore, any added functions polyubiquitin chains possess beyond tethering substrates to the proteasome are not strictly necessary for proteolysis. ATP-dependent protease complexes degrade the majority of unstable cellular proteins, a process that is conserved across all three kingdoms of life. These molecular machines function both generally in protein turnover and specifically in the regulation of processes such as transcription, apoptosis, antigen presentation, and cell cycle progression (1Jesenberger V. Jentsch S. Nat. Rev. Mol. Cell. Biol. 2002; 3: 112-121Crossref PubMed Scopus (307) Google Scholar). A high degree of conservation is evident among them; the archaebacterial and eukaryotic 20S proteolytic core particles share both sequence and structural homology (2Bochtler M. Ditzel L. Groll M. Hartmann C. Huber R. Annu. Rev. Biophys. Biomol. Struct. 1999; 28: 295-317Crossref PubMed Scopus (419) Google Scholar), whereas eubacteria have functionally related complexes: ClpYQ, ClpXP, and ClpAP (3Bochtler M. Hartmann C. Song H.K. Bourenkov G.P. Bartunik H.D. Huber R. Nature. 2000; 403: 800-805Crossref PubMed Scopus (375) Google Scholar, 4Bochtler M. Ditzel L. Groll M. Huber R. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 6070-6074Crossref PubMed Scopus (166) Google Scholar, 5Groll M. Ditzel L. Lowe J. Stock D. Bochtler M. Bartunik H.D. Huber R. Nature. 1997; 386: 463-471Crossref PubMed Scopus (1923) Google Scholar). The 20S core particle is composed of four stacked heptameric rings structured in an α-β-β-α configuration. Access to the proteolytic central chamber is obstructed at both ends of the cylindrical assembly by N-terminal projections of the α-subunits, thus preventing uncontrolled proteolytic degradation (5Groll M. Ditzel L. Lowe J. Stock D. Bochtler M. Bartunik H.D. Huber R. Nature. 1997; 386: 463-471Crossref PubMed Scopus (1923) Google Scholar, 6Groll M. Bajorek M. Kohler A. Moroder L. Rubin D.M. Huber R. Glickman M.H. Finley D. Nat. Struct. Biol. 2000; 7: 1062-1067Crossref PubMed Scopus (641) Google Scholar). In eukaryotes, docking with the 19S regulatory particle (RP) 1The abbreviations used are: RP, regulatory particle; Ub, ubiquitin; ODC, ornithine decarboxylase; AZ1, antizyme; HA, hemagglutinin; Tor, (Tor11883-2078); DTL, degradation through localization. 1The abbreviations used are: RP, regulatory particle; Ub, ubiquitin; ODC, ornithine decarboxylase; AZ1, antizyme; HA, hemagglutinin; Tor, (Tor11883-2078); DTL, degradation through localization. to form the complete 26S proteasome is sufficient to relieve this block, opening a channel into the core (6Groll M. Bajorek M. Kohler A. Moroder L. Rubin D.M. Huber R. Glickman M.H. Finley D. Nat. Struct. Biol. 2000; 7: 1062-1067Crossref PubMed Scopus (641) Google Scholar, 7Kohler A. Cascio P. Leggett D.S. Woo K.M. Goldberg A.L. Finley D. Mol. Cell. 2001; 7: 1143-1152Abstract Full Text Full Text PDF PubMed Scopus (339) Google Scholar). Eukaryotes have evolved an elaborate system that operates in conjunction with the proteasome to facilitate the temporal and specific regulation of intracellular proteolysis. Most proteins are targeted for degradation through ubiquitination, mediated by the E1, E2, E3 ubiquitin (Ub) conjugation machinery. These three consecutively acting enzymes are necessary for target recognition, transfer of a ubiquitin moiety to the substrate, and subsequent elongation of the ubiquitin branched chain (8Hershko A. Ciechanover A. Annu. Rev. Biochem. 1998; 67: 425-479Crossref PubMed Scopus (6736) Google Scholar). Modularity and the large number of E2 Ub-conjugating enzymes and E3 Ub ligases allow for greater specificity and flexibility in recognizing a diverse range of substrates. Once a protein is polyubiquitinated, it is targeted to and degraded by the 26S proteasome. The polyubiquitin chain is thought to play two possible roles. The first is to target the protein to the proteasome; the second is to initiate the process of degradation. The targeting hypothesis is supported by the identification of several proteasome subunits that either bind or crosslink to ubiquitin chains (9Deveraux Q. Ustrell V. Pickart C. Rechsteiner M. J. Biol. Chem. 1994; 269: 7059-7061Abstract Full Text PDF PubMed Google Scholar, 10Lam Y.A. Lawson T.G. Velayutham M. Zweier J.L. Pickart C.M. Nature. 2002; 416: 763-767Crossref PubMed Scopus (365) Google Scholar). Hypotheses for how ubiquitin-dependent initiation of degradation might occur include allosteric regulation, channel opening, and assistance in the unfolding of the target (11Groll M. Huber R. Int. J. Biochem. Cell Biol. 2003; 35: 606-616Crossref PubMed Scopus (156) Google Scholar). However, little data have been reported to support these ideas. The elucidation of the mechanism for proteolysis of ornithine decarboxylase (ODC) established that polyubiquitination is not necessary for proteasome-mediated degradation (12Coffino P. Nat. Rev. Mol. Cell. Biol. 2001; 2: 188-194Crossref PubMed Scopus (295) Google Scholar). ODC is an enzyme whose degradation is mediated by its binding to the cofactor antizyme 1 (AZ1). Once bound, ODC-AZ1 can be recognized by the proteasome, and ODC is degraded in an ubiquitin-independent manner. However, it is unclear whether the means by which AZ1 promotes degradation differs fundamentally from that of polyubiquitin chains. A recent study demonstrated that ODC-AZ1 competes with substrate-linked and free polyubiquitin chains for the occupancy of the same binding site on the proteasome (13Zhang M. Pickart C.M. Coffino P. EMBO J. 2003; 22: 1488-1496Crossref PubMed Scopus (177) Google Scholar). Thus, the mechanism for the degradation of ODC seems to represent a specialized evolutionary adaptation that closely mimics ubiquitination. The binding of ubiquitin-conjugated substrates (or ODCAZ1) to the proteasome may itself serve as the activation step in proteolytic degradation. We hypothesize that such an activation step is not necessary, that localization to the proteasome can be sufficient for degradation. In this study, we support this hypothesis by demonstrating proteasome-mediated degradation both in vivo and in a purified in vitro system, while artificially bypassing the need for ubiquitin-dependence. Construction of Parental Strain DY001—All experiments were performed in derivatives of strain DY001 to ensure that the components of the heterodimerization system would minimally interact with endogenous proteins, thus preventing cell cycle arrest and mislocalization of the reporter upon the addition of rapamycin (14Heitman J. Movva N.R. Hall M.N. Science. 1991; 253: 905-909Crossref PubMed Scopus (1506) Google Scholar). The Fpr1-rapamycin binding domain (nucleotides 5656-6243) of the dominant allele TOR1-2 was amplified from strain JHY17-9C (15Lorenz M.C. Heitman J. J. Biol. Chem. 1995; 270: 27531-27537Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar) and subcloned into the integrating plasmid pRS306 (16Sikorski R.S. Hieter P. Genetics. 1989; 122: 19-27Crossref PubMed Google Scholar). This vector was then digested with HindIII to cut once within TOR1-2 and transformed into the strain BY4742 Δfpr1::kanr (Research Genetics). Integration and subsequent loop-out was selected for on the appropriate plates. The correct strain was verified by PCR and sequencing. Genomic Tagging of Proteasome Subunits—Tagging of proteasome subunits was performed both by homologous recombination of linear fragments containing 40 bp of flanking homology to the target site and by two-step integration with a non-replicating plasmid. FPR1 was amplified by PCR from the strain FY4 (17Winston F. Dollard C. Ricupero-Hovasse S.L. Yeast. 1995; 11: 53-55Crossref PubMed Scopus (769) Google Scholar) and subcloned into the plasmid pUG-spHIS5 (18Wach A. Brachat A. Alberti-Segui C. Rebischung C. Philippsen P. Yeast. 1997; 13: 1065-1075Crossref PubMed Scopus (506) Google Scholar) with a C-terminal hemagglutinin (HA)-tag, forming FPR1-HA-pUG-spHIS5. Integration primer pairs were designed for tagging each of four proteasome subunits (PRE10, RPN2, RPN6, RPN11). For each pair, one primer contained 40 bp of genomic homology to the 3′ end of the proteasome subunit, excluding the stop codon, and 20 bp of homology to the 5′ end of FPR1 on pUG-spHIS5, excluding ATG. The second primer contained 40 bp of genomic homology ∼50 bp downstream of the proteasome subunit gene stop codon and 20 bp of homology to pUG-spHIS5 immediately downstream of the spHIS5 marker flanked by loxP sites. Two confirmatory primers were also designed that flanked the integration site of each proteasome subunit. Strain DY001 was transformed with pSH47, a plasmid with a galactose-inducible cre gene and a URA3 selection marker (19Guldener U. Heck S. Fielder T. Beinhauer J. Hegemann J.H. Nucleic Acids Res. 1996; 24: 2519-2524Crossref PubMed Scopus (1335) Google Scholar). A 2-kb linear fragment from FPR1-HA-pUG-spHIS5 was amplified using each integration primer pair to generate linear 2-kb fragments suitable for genomic integration. 15 μg of each fragment was transformed into DY001 carrying pSH47, and selection was performed on SC-URA-HIS. Colonies were then picked and streaked onto SC-URA GAL to induce cre and to select for the loopout of the spHIS5 marker. Colonies were finally streaked onto a 5-fluoroorotic acid-containing plate to remove pSH47. All tagged subunits were verified by sequencing. Approximately 400-500 bp of the carboxyl-terminal end (without the stop codon) and 3′ untranslated region of proteasome subunits RPT2, RPT5, and RPN10 were amplified from strain FY4. Each pair was subcloned into the integration plasmid pRS306 (16Sikorski R.S. Hieter P. Genetics. 1989; 122: 19-27Crossref PubMed Google Scholar) along with FPR1 so that the final structure at the cloning site was 5′-proteasome subunit C-term-FPR1-proteasome subunit UTR. Each derivative of pRS306 was cut at a unique site within the carboxy terminus of the proteasome subunit and transformed into DY001. Selection for integration was done on SC-URA; loopout of the marker was on 5-fluoroorotic acid-containing plates. All tagged subunits were verified by sequencing. Δpdr5 strains were generated by recombination of a URA3 marker flanked by 40 bp homologous to sequence immediately 5′ and 3′ to genomic PDR5. Preparation of Reporter Plasmids—All versions of the reporters were derivatives of the vector pRS415 (16Sikorski R.S. Hieter P. Genetics. 1989; 122: 19-27Crossref PubMed Google Scholar). TOR1(S1972R) is an allele of TOR1 that has a severely impaired binding affinity to FPR1-rapamycin (14Heitman J. Movva N.R. Hall M.N. Science. 1991; 253: 905-909Crossref PubMed Scopus (1506) Google Scholar). The sequence corresponding to amino acids 1883-2078 for both Tor1 and Tor1(S1972R) were amplified and inserted into the vector pRS415 along with HIS3 amplified from strain FY4. HA-tagged versions of the reporters had the hemagglutinin epitope-fused C-terminal to His3. Screen and Western Assays—In both assays, rapamycin was added to a final concentration of 1 μm, cycloheximide was added to a final concentration of 30 ng/ml, and PS-341 was added to a final concentration of 200 μm. For liquid cultures, samples were extracted at fixed time intervals (0, 30, 60, 90 min) and whole-cell extracts were made. 25 μgof total protein was used from each sample, resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Membranes were blocked with 10% powdered nonfat milk in phosphate-buffered saline plus Tween overnight at 4 °C and incubated with and then was done using Western system. of the proteasomes were from strains and Rpn10 FPR1 using the D.S. J. A. M. T. Finley D. Mol. Cell. 2002; Full Text Full Text PDF PubMed Scopus Google Scholar). were in in 1 and by were at for 25 and incubated with for 1 at 4 was in a with of 1 and with 1 1 was performed by the in of containing of of at 30 °C for 1 was and for each The of the 26S complexes and the of and subunits was by SDS-PAGE, and In and were cut of their pRS415 inserted into and transformed into cells A overnight was used to 1 of which was for 4 at The fusion protein was with of and the was for 4 at 30 of purified protein was done to the using a of of the was 1 and to The were performed in 1 containing 4 μg of proteasome and μg of or in an of at 30 was with 1 rapamycin, and proteasome was with were by of from the in Full and all primer and plasmid sequence can be on the at of are a of that facilitate the of any two have been used in a number of such as localization of proteins to of and of gene D.M. S.L. Science. PubMed Scopus Google Scholar, D.M. S.L. Nature. 1996; PubMed Scopus Google Scholar, Nature. 1996; PubMed Scopus Google Scholar). the hypothesis that localization to the proteasome is sufficient for we and such a system in of the pair of is to the proteasome, whereas the is to the reporter to be for degradation. of the in an system the two thus the reporter to the site of the proteasome. degradation as a of the it can be with the appropriate In S. the rapamycin has been to bind with high affinity to the protein Fpr1, and this in to the domain of Tor1 or (15Lorenz M.C. Heitman J. J. Biol. Chem. 1995; 270: 27531-27537Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar). Seven proteasome subunits that range in from the 20S proteolytic core were with Fpr1 and subunits not be of their not DY001 strains with and Rpn10 subunits to as strain were and contained as by and as by Western not A Screen for was used to the strains containing proteasome subunits for The marker HIS3 a protein in and is necessary for on All of the strains used in this have their of HIS3 and of for Two reporter constructs were designed for in the The reporter is an fusion of the with The reporter with a that has a affinity for the (14Heitman J. Movva N.R. Hall M.N. Science. 1991; 253: 905-909Crossref PubMed Scopus (1506) Google Scholar). of was on the of strains on with or strains that the reporter were not to any in the two the fusion protein would not bind rapamycin and, therefore, not be to the proteasome. rapamycin and, therefore, have an with the proteasome in the of the this was sufficient for containing strains would be to a on the with The DY001 and the four strains were transformed with the and reporter All strains on containing not Each was then as a on two of either containing or rapamycin Strain the Strain had a that was and and not with degradation of the reporter for these two strains be Western was a for strains to the However, on its the not that localization was sufficient for degradation function have been in a but manner. this degradation of the reporters was and were transformed with that versions of and The were in liquid to cycloheximide was added to protein Each was then into two and rapamycin was added to one were at and whole-cell protein extracts were and used for Western with both and an in the of degradation of the reporter rapamycin was added to the cultures, with of in the of degradation was with strains the that degradation of the reporter was mediated by the proteasome and not through the experiments were with the addition of the proteasome PS-341 and A. J. S. D. T. L. R. J. Proc. Natl. Acad. Sci. U. S. A. 1998; PubMed Scopus Google Scholar). addition of PS-341 degradation of in the of These experiments that the on was to the degradation of the and that the 26S proteasome is necessary for this to for and that the 26S proteasome is not necessary but also sufficient for DTL, in vitro experiments were performed in which the components were purified proteasomes and reporter 26S proteasomes have been for in D.S. J. A. M. T. Finley D. Mol. Cell. 2002; Full Text Full Text PDF PubMed Scopus Google Scholar). The same was used to proteasomes from strain versions of and were purified from a system and then with the with or were extracted at and reporter degradation was by Western The in vitro was with the in vivo All of the experiments using purified degradation. Degradation of was observed in the of rapamycin and proteasomes from strain The of the reporter was on the of 15 with was observed in Once proteolysis of be upon the addition of the proteasome These in with the in vivo that localization to the proteasome is sufficient for the initiation of degradation. is not polyubiquitin chains serve beyond targeting substrates to the proteasome. In this study, we demonstrated that once a is localized to the proteasome, are not for degradation to we designed a system using of to artificially reporters with proteasome We then established a to strain for hypothesis was we to strains that were for on and such Western performed on whole-cell extracts from these strains that of the reporter proteins was to their proteolysis. we that the proteasome was necessary for this degradation by the process with the addition of the in a in vitro system we demonstrated that purified 26S proteasome was sufficient to a in which polyubiquitin chains or that any functions beyond tethering substrates to the proteasome are not strictly necessary for proteolysis to Seven proteasome subunits were selected for as it was not would a with The tagged subunit that the was one of the proteasome subunits that has been to bind to polyubiquitin chains (9Deveraux Q. Ustrell V. Pickart C. Rechsteiner M. J. Biol. Chem. 1994; 269: 7059-7061Abstract Full Text PDF PubMed Google Scholar). little is an protein that of the proteolytic core particle L. C. L. S. S. A. A. M. R. R. Brachat S. S. M. A. P. F. M. D. U. Hegemann J.H. S. S.L. P. D. L. C. M. R. P. S.L. J.L. M. P. S. M. J. M. Philippsen P. M. Nature. 2002; PubMed Scopus Google Scholar) and which a but in and on containing or Western of strains and established that both fusion proteins at their molecular and, therefore, were is subunit that has been reported to ubiquitin chains as by experiments Y.A. Lawson T.G. Velayutham M. Zweier J.L. Pickart C.M. Nature. 2002; 416: 763-767Crossref PubMed Scopus (365) Google Scholar). The that not occur for these strains that either of the was or that localization to the proteasome, degradation was not The of an subunit to promote is upon a number of for the of the and of a target to any site such as the We are whether of the in strains and these strains be used to to the proteasome their proteolysis. In addition to used as a the also as an are of of ubiquitin-independent degradation of substrates in vitro Woo K.M. Goldberg A.L. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar, C. J. 1994; Google Scholar). is also that the proteasome degrade but unstable substrates in a ubiquitin-independent R. L. M. J. 2002; PubMed Scopus Google Scholar, M. F. S. C. M. J. Biol. Chem. 1995; 270: Full Text Full Text PDF PubMed Scopus Google Scholar, S. C.M. Goldberg A.L. J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar). In the substrates were The reporters were as demonstrated by the and also by the of to bind to we that binding of rapamycin itself not the a was not observed in the DY001 strain free Fpr1 in addition to the reporters not or proteins need to be to the channel into the 20S proteolytic core (5Groll M. Ditzel L. Lowe J. Stock D. Bochtler M. Bartunik H.D. Huber R. Nature. 1997; 386: 463-471Crossref PubMed Scopus (1923) Google Scholar, Goldberg A.L. Nat. Cell. 2000; 2: PubMed Scopus Google Scholar). This process is thought to be by the which both and and form of the for the 19S regulatory particle Goldberg A.L. Nat. Cell. 2000; 2: PubMed Scopus Google Scholar, Glickman M. R. Finley D. M. Nat. Cell. 1999; PubMed Scopus Google Scholar, J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar). The with these data is that substrates at the proteasome are an through their with the and that degradation once this is C. S. M. A. Mol. Cell. 2001; 7: Full Text Full Text PDF PubMed Scopus Google Scholar, A. Struct. Biol. 2003; 13: PubMed Scopus Google Scholar). The of study support this as to elaborate that for proteolysis to We observed a range of in that was upon the strain degradation was observed at all that of the with to the proteasome is not strictly necessary for proteolysis. However, the observed in study that is an for substrates to the for This is the by proteins to the proteasome through polyubiquitin chains. The of the reporter protein in the strain was which is and with that of substrates. Our a to protein The in is to of processes that are in However, the of that both specifically bind to and the function of a protein target is The of K.M. A. F. C.M. Proc. Natl. Acad. Sci. U. S. A. 2001; PubMed Scopus Google Scholar, K.M. R. A. C.M. Mol. Cell 2003; 2: Full Text Full Text PDF PubMed Scopus Google Scholar) to the of but not by using to artificially target proteins for ubiquitination. We to bypass the ubiquitination step and generate that promote the degradation of proteins through heterodimerization with the proteasome. In we are the of the system used in this study into a method for the of protein the to endogenous proteins of in an strain of degradation. such a system would have such as A. S. Nature. 1998; PubMed Scopus Google Scholar) and U. 2000; PubMed Scopus Google Scholar) is that it would target proteins for whereas the two and therefore, upon the of for We Heitman for the strain for the pUG-spHIS5 and pSH47, for strain for the pRS306 and and and for of the