Gene UBP1 of Saccharomyces Cereuisiae”

Cloning and Functional Analysisof the Ubiquitin-specific Protease Gene UBP1 of Saccharomyces cereuisiae”

(Received for publication, February 5, 1991)

John W. Tobias and Alexander Varshavsky From the Departmentof Biology, Massachusetts Instituteof Technology, Cambridge, Massachusetts02139

In eukaryotes, both natural and engineered fusions erone function,in that its transient(cotranslational) covalent of ubiquitin to itself or other proteins are cleaved by association with specific ribosomal proteins promotes the processing proteases after the last (G~Y‘~)residue of assembly of ribosomal subunits (Finley et al., 1989). In vivo ubiquitin. Using the method ofsib selection,and taking acceptors of ubiquitin include chromosomal histones (Bonner advantage of the fact that bacteria such as Escherichia et al., 1988), integral membrane proteins such as the lympho- coli lack ubiquitin-specific enzymes, we have cloned a cyte homing receptor and the growth hormone receptor (Sie- gene, named UBPl, of the yeast Saccharomyces cere- gelman et al., 1986; Leung et al., 1987), actin (Ball et al., visiae that encodes a ubiquitin-specific processing pro- 1987), the intracellular neurofibrillary tangles in a variety of tease. With the exception of polyubiquitin, the UBPl neurodegenerative diseases (Perry et al., 1987; Lowe et al., protease cleaves at the carboxyl terminus of the ubiqui- 1988), the plant regulatory protein phytochrome (Jabben et tin moiety in natural and engineered fusions irrespective of their size or the presence of an amino-terminal al., 1989), the short-livedMATa2 repressor of the yeast ubiquitin extension. These properties of UBPl distin- Saccharomyces cereuisiae (Hochstrasser et al., 1991), and cy- guish it from the previously cloned yeast protease clins, a family of short-lived cell cycle regulators (Glotzer et YUH1, which deubiquitinates relatively short ubiqui- al., 1991). tin fusions but is virtually inactive withlonger fusions Unlike branched Ub-protein conjugates, which are formed such as ubiquitin-&galactosidase. The amino acid se- posttranslationally, linear Ub-protein adducts are formed as quence of the 809-residue UBPl lacks significantsim- the translationalproducts of either natural or engineered ilarities to other known proteins, including the 236- ubiquitin gene fusions. Thus, in S. cereuisiae for example, residue YUHl protease. Null ubpl mutants are viable, ubiquitin is generated exclusively by proteolytic processing of and retain the ability to deubiquitinate ubiquitin-0- precursors in which ubiquitin is joined either to itself, as in galactosidase, indicating that the family of ubiquitin- the linear polyubiquitin protein UBI4, or to unrelated amino specific proteases in yeast is not limited to UBPl and acid sequences, as in the hybrid proteins UBI1-UBI3 (Ozkay- YUH1. nak et al., 1987). In growing yeast cells, ubiquitin is generated from the UBI1-UBI3 precursors whose “tails”are specific ribosomal proteins (Finley et al., 1989). The polyubiquitin (UBI4) gene is dispensable in growing cells but becomes Ubiquitin (Ub),’ a highly conserved 76-residue protein, is essential (as the main supplier of ubiquitin) during stress present in eukaryotic cells either free or covalently joined to (Finley et al., 1987, 1988). The lack of genes encoding mature a great variety of proteins. The posttranslational coupling of ubiquitin and the fusion structure of ubiquitin precursors in ubiquitin to other proteins is catalyzed by a family of Ub- yeast are characteristic of other eukaryotes as well (Schles- conjugating enzymes (also called E2 enzymes (Pickart and inger and Bond, 1987). Thus, there must exist Ub-specific Rose, 1985)) and involves formation of an isopeptide bond processing proteases that produce ubiquitin, a protein essen- between the carboxyl-terminal glycine residue of ubiquitin tial for cell viability (Finley et al., 1984,1987, 1988), by and the E-amino group of a lysine residue in an acceptor cleaving Gly-X peptide bonds at thejunctions between ubiqui- protein (Pickart, 1988; Jentsch et al., 1990). A major function tin and itscarboxyl-terminal extensions. A priori, these pro- of ubiquitin is to mark proteins destined for selective degradation (reviewed by Hershko, 1988; Finley et al., 1988; Cie- teases may also have an isopeptidase activity, i.e. the ability chanover and Schwartz, 1989; Olson and Dice, 1989; Monia to deubiquitinate posttranslationally formed, branched Ub- et al., 1990). Ubiquitin has also been shown to have a chap- protein conjugates. A multiubiquitin chain is one such conjugate, in which ubiquitin itself serves as an acceptor, with * This work was supported by Grants AGO8991 and DK39520 (to several ubiquitin moieties attached sequentially to the initial A. V.) from the National Institutes of Health. The costs of publication acceptor protein to form a chain of branched Ub-Ub adducts of this article were defrayed in part by the payment of page charges. (Chau et al., 1989). Formation of a multiubiquitin chain on a This article must therefore be hereby marked “aduertisement” in targeted protein is required for the protein’s subsequent deg- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. radation (Chau et al., 1989; Gregori et al., 1990). The nucleotide sequence(s) reportedin thispaper has been submitted tothe GenBankTM/EMBL Data Bunk with accessionnumberls) Ub-specific, ATP-independent proteases capable of cleav- M63484. ing ubiquitin from its linearor branched conjugates have been ’ The abbreviations used are: Ub, ubiquitin; DHFR, dihydrofolate detected in all eukaryotes examined (Matsui et al., 1982; Rose, reductase; UB-X-DHFR, ubiquitin-X-mouse dihydrofolate reductase, 1988; Mayer and Wilkinson, 1989; Jonnalagadda et al., 1989) where X is an amino acid residue at theUb-DHFR junction; Ub-X- but not in bacteria such asEscherichia coli, which lack eukar- Pgal, ubiquitin-X-E. coli (%galactosidase;MTX, methothrexate; SDS- PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; yotic ubiquitin and Ub-specific enzymes (Finley et al., 1988). DTT, dithiothreitol; LB, Luria broth; HEPES, 4-(2-hydroxyethyl)-l- Recently, Miller et al. (1989) have cloned a S. cereuisiae gene, piperazineethanesulfonic acid; kb, kilobase(s); bp, base pair(s); ORF, named YUHl, for a Ub-specific protease that cleaves ubiqui- open reading frame. tin from its carboxyl-terminal fusions to relatively short 12021 12022 Ubiquitin-specific Protease amino acid sequences but is virtually inactive with larger centrifuge tube containing0.975 mlof 65 mM Na-EDTA, 50 mM Tris- fusions such as Ub-@-galactosidase(Ub-@gal). Wilkinson et HCl (pH 8.0), and protease inhibitors antipain, chymostatin, leupep- al. (1989) havealso cloned a cDNA for a mammalian homolog tin, pepstatin, and aprotinin, each at 25 pg/ml. 10 p1 of 10% (v/v) of the yeast YUHl protease. Triton X-100 was then added and dispersed by pipetting. The lysate was centrifuged at 40,000 X g for 30 min. The supernatant, containing We now report the cloning and functional analysis of an S. the 35S-labeledUb-M-DHFR, was frozen in liquid nitrogen and stored cereuisiue gene, named UBPl, which encodes a Ub-specific at -85 "C. To purify Ub-M-DHFR, a methothrexate (MTX) affinity processing protease whose amino acid sequence is dissimilar matrix was prepared as described by Kaufman (1974), using amino- to those of the YUHl protease and other known proteins. hexyl-Sepharose (Sigma). The MTX-Sepharose column (0.5-ml bed The UBPl protease cleaves at the carboxyl terminus of the volume) was washed with 10 ml of MTX buffer (0.2 M NaCl, 1 mM Na-EDTA, 0.2 mM dithiothreitol (DTT), 20mM Na-HEPES, pH ubiquitin moiety in linear fusions irrespective of their size or 7.5). The Ub-M-DHFR-containing 35Ssupernatant was thawed, clar- the presence of an amino-terminal ubiquitin extension. We ified by centrifugation at 12,000 X g for 1 min, and applied to the also show that null ubpl mutants are viable and have prop- column, which was then washed with 50 ml of MTX buffer, 50 ml of erties indicating that the family of Ub-specific proteases in MTX buffer containing 2M urea, and thenagain with 50 ml of MTX yeast is not limited to UBPl andYUH1. buffer. Ub-M-DHFR was then eluted from the column with folic acid Since Ub-specific proteases cleave ubiquitin from its car- buffer (10 mM folic acid, 1 M KCI, 1 mM DTT, 1 mM Na-EDTA, 0.2 M potassium borate, pH 9.0). The folic acid buffer was applied in 1- boxyl-terminal extensions irrespective of the identity of the ml aliquots, and 1-ml fractions were collected. The fractions were extension's residue abutting the cleavage site (Bachmair et assayed for 35S;pooled fractions containing the peak of 35Swere al., 1986; Gonda et al., 1989), UBPl and its functional coun- dialyzed for 20 h at 4 "C against two changes of storage buffer (50% terparts make possible the in vivo or in vitro generation of glycerol, 1 mM MgCI2, 0.1 mM Na-EDTA, 40 mM Na-HEPES, pH proteins (including short polypeptides) bearing predetermined 7.5), and stored at -20 "C. The purified [35S]Ub-M-DHFRwas as- residues at their amino termini, a method with applications sayed by SDS-PAGE and fluorography and found to be greater than 95% pure. Other test proteins such as DHFR-Ub-M-pgal (Fig. L4, in both basic research (Baker and Varshavsky, 1991; Bartel VI), and the natural yeast ubiquitin fusions UBI2 (Fig. L4, I), UBI3 et al., 1990; Johnson et al., 1990; Townsend et al., 1988) and (Fig. lA, II), and UBI4 (polyubiquitin; Fig. lA, III), were expressed biotechnology (Lu et al., 1990; Butt et al., 1989; Sabin et al., in E. coli and assayed for deubiquitination with UBPl in mixed E. 1989; Ecker et al., 1989; Yo0 et al., 1989; Mak et al., 1989). coli extracts as described below. Sib Selection Cloning of the UBPl Gene-E.coli HBlOl cells EXPERIMENTALPROCEDURES carrying the S. cereuisiue genomic DNA library RB237 (see above) were spread on LB plates containing 100 pg/ml ampicillin (LB/amp) Strains, Media, and GeneticTechniques-E. coli strains HBlOl to yield -40 colonies per plate. After -16 h at 36"C, the colonies (pro leu thi lacy hsdS~20endA secA rpsL20 ara-14 galK2 xyl-5 mtl-1 were replica-plated onto LB/amp plates; the original plates were supE44 merBB; Boyerand Roulland-Dussoix, 1969) and JMlOl (supE stored at 4 "C, while their replicas were incubated at 36 "C for 24 h. thi A(lac-proAB)/F' traA36 proA+ proB+ lacIq lacZAM15; Yanisch- Cells on each of the replica plates were then eluted with 1 ml of LB/ Perron et al., 1985) were propagated in Luria broth (LB) andused as amp by repeated washing over the agar surface until all of the colonies hosts for plasmids and phage M13 derivatives. The S. cereuisk were dispersed into the liquid. The eluates were each diluted with 4 strains used were DF5 (MATaIMATa trpl-lltrpl-1 ura3-51/urd- ml of LB/amp and incubated overnight with shaking at 36 "C. 1.7 ml 51 hL3-A2000/his3-A2000 leu2-3,112/leu2-3,112 lys2-801/lys2-801 of each overnight culture was chilled on ice and centrifuged at 12,000 gallgal; Finley et al., 1987), BBY45(MATa trpl-I urd-52his3-A2000 X g for 1 min. The pellet was resuspended by vortexing in 50 pl of leu2-3,112 lys2-801 gal; Bartel et al., 1990), and BWGl-7a (MATa 25% (w/v) sucrose, 0.25 M Tris-HC1 (pH 8.0). 10 pl of freshly prepared hL4 ura3 adel leu2; Guarente et al., 1982). Rich (YPD) and synthetic egg-white lysozyme solution (10 mg/ml; Sigma) in 0.25 M Tris-HC1 yeast media were prepared according to Sherman et al. (1986). Trans- (pH 8.0) was added, mixed by light vortexing, and incubated at 0 "C formations of S. cereukiue and E. coli were carried out, respectively, for 5 min. 0.15 ml of 75 mM Na-EDTA, 0.33 M Tris-HC1 (pH 8.0) by the lithium acetate method (Ito et al., 1983) and by the calcium was then added, followed by incubation at 0°C for 5 min, with chloride method (Ausubel et al., 1989). The DNA library RB237 used occasional stirring. 1 p1 of 10% (v/v) Triton X-100 was then added for the isolation of UBPl was obtained by cloning a partial Sau3AI and mixed by pipetting. Eachof the resulting lysates was centrifuged digest of S. cereuisiaegenomic DNA into the BamHI site of the at 12,000 X g for 15 min at 4 "C, and the supernatantswere tested for URAS-CEN4-based vector YCp50 (Rose et al., 1987). Ub-specific protease activity as described below, using Ub-M-DHFR Ubiquitin-containing Test Proteins-Ub-M-pgal (ubiquitin-Met-& asthe test substrate. Upon finding a positive extract, overnight galactosidase) (see Fig. lA, V) and theotherwise identical Ub-P-pgal cultures were made as described above from each of the -40 individual containing proline (instead of methionine) at the Ub-pgal junction colonies on the plate that yielded the positive extract. AYCp50-based were purified by affinity chromatography of extracts from [35S]me- plasmid, named pJT55, was then isolated from a transformant that thionine-labeled E. coli (carryingpKKUb-M-pgal or pKKUb-P-pgal)) yielded a positive extract upon the second round of screening. The on aminophenylthiopyranogalactoside-Sepharose as described by yeast DNA insert of pJT55 was analyzed as described below. Gonda et al. (1989). Ub-M-DHFR (ubiquitin-Met-mouse dihydrofo- In Vitro Assaysfor Ub-specific Protease Actiuity-An E. coli extract late reductase) (Fig. lA, ZV) was prepared as follows: an overnight (9 pl) prepared as described above was added to 1 pl of an 35S-labeled culture (1 ml) of E. coli JMlOl carrying the plasmid pub-M-DHFR test protein(Ub-M-DHFR, Ub-M-pgal, or Ub-P-@gal) in storage (Bachmair and Varshavsky, 1989) was diluted into 50ml of LB buffer (see above), and incubated at 36°C for 3 h. 5 p1 of 3-fold supplemented with ampicillin at 40 pg/ml, and the cells were grown concentrated SDS-PAGE sample buffer (6% SDS, 33% glycerol, 30 with shaking at 37 "C to anA600 of -0.9. The culture was chilled on mM Na-EDTA, 0.3 M DTT, 0.6% bromphenol blue, 0.3 M Tris-HC1, ice for 15 min, centrifuged at 3000 X g for 5 min, and washed twice pH 6.8) was then added, followed by heating of the sample at 100 "C with M9 buffer (Ausubel et al., 1989) at 4°C. The cells were resus- for 3 min, electrophoresis in a 6% (for @-galactosidase)or a 12% (for pended in M9 buffer supplemented with glucose (0.2%), thiamine (2 DHFR) polyacrylamide-SDS gel, and fluorography. pg/ml), ampicillin (40 pg/ml), 1 mM isopropylthiogalactoside, and Similar tests using the unlabeled natural ubiquitin precursors methionine assay medium (0.0625%) (Difco). After incubation with UBI2, UBI3, and UBI4 (see introduction) were carried out as follows. shaking for 1 h at 37 "C, 1 mCi of Tran~~~S-label(ICN; at least 70% 5 pl of an E.coli extract to be tested for Ub-specific protease activity of 35Sis in L-methionine, with the rest in L-cysteine and related was mixed with 5 pl of an E. coli extract prepared as described above compounds) was added, and shaking was continued for 5 min. Unla- from E. coli carrying a plasmid that expressed UBI2, UBI3, or UBI4 beled L-methionine was then added to 1 mM, and shaking was (Ozkaynak et al., 1987; Finley et al., 1987).The mixture was incubated continued for another 10 min. Cells were harvested, washed twice for 1 h at 36"C, followed by SDS-PAGE as described above, and with M9 buffer, and resuspended in 0.5 ml of 25% (w/v) sucrose, 50 immunoblot analysis using a rabbit polyclonal anti-Ub antibody (a mM Tris-HC1, pH 8.0. 0.1 ml of freshly prepared egg-white lysozyme gift from Dr. Vincent Chau, Wayne State University, Detroit MI). solution (10 mg/ml; Sigma) in 0.25 M Tris-HC1, pH 8.0, was then Specifically, after SDS-PAGE, the gel and PVDF membrane (Milli- added, and the mixture was incubated for 5 min at 0 "C, followed by pore) were equilibrated with transfer buffer (20% methanol, 0.2 M the addition of0.1 ml of0.5 M Na-EDTA, pH 8.0, andfurther glycine, 25 mM Tris), followed by electroblotting of proteins from the incubation at 0 "C for 5 min. The suspension was transferred into a gel using the Genie apparatus (Idea Scientific) at 20 V for 25 min. Ubiquitin-specific Protease 12023 A B I UBI12: m-1

I1 UBI3:

III UBI4:

VI FIG. 1. Test ubiquitin fusions and UBPl-containing plasmid pJT’7O. In A: I, the yeast UBI1 and UBI2 gene products (their amino acid sequences are identical) arethe natural fusion of ubiquitin to a specific ribosomal protein (Finley et al., 1989); 11, the yeast UBI3 protein is the naturalubiquitin fusion to another ribosomal protein; Ill, the yeast UBI4 protein is the natural head-to-tail, spacerless fusion of five ubiquitin moieties (Ozkaynak et al., 1987); IV, Ub-M-DHFR is an engineered fusion of yeast ubiquitin to mouse dihydrofolate reductase (DHFR), the junctional residue being methionine (M); V, same as IV buta fusion to E. coli P-galactosidase; VI, DHFR-Ub- M-pgal, same as V but with the ubiquitin moiety bearing at itsamino terminus the mouse DHFR moiety (connected to ubiquitin via a Ser-Gly-Gly-Gly linker). Black rectangles indicate the 45-residue sequence from an internal region of E. coli lac repressor that is present at the amino termini of DHFR and pgal in these engineered test proteins (Bachmair and Varshavsky, 1989). With the exception of pgal, the lengths of rectangles are proportional to thesizes of protein moieties. (See “Experimental Procedures” for details.) In B: pJT70, the plasmid that carries the -2.8-kb, UBPl-containing yeast genomic DNA insert. (See “Experimental Procedures” and text for details.)

The membrane was washed in ANT buffer (0.15 M NaC1,0.02% and to produce a reading frame that encoded a 4-residue linker NaN,, 50 mM Tris-HC1, pH 8.0), then incubated (at room tempera- sequence (Ser-Gly-Gly-Gly)between the carboxyl terminus of DHFR ture) with rocking in ANT buffer plus 20% fetal calf serum, followed andthe aminoterminus of ubiquitin, yielding the plasmid by a 2-h incubation at room temperature with the anti-Ub antibody pDHFRUbMpga1. in ANT/fetal calf serum buffer. The membrane was then washed E. coli JMlOl carrying pDHFRUbMPga1, alone or together with several times inANT buffer plus 0.05% Tween 20.An alkaline the compatible plasmid pJT184 (see below) that expressed UBP1, phosphatase-linked anti-rabbit second antibody (Promega Biotech) wasgrown in LB (in the presence of ampicillin at 40 pg/ml and in the ANT/Tween buffer was added, followed by a 30-min incubation chloramphenicol at 35 pg/ml to maintain the plasmids) to anAGW of at room temperature, washing of the membrane in ANT/Tween -0.3. 1 ml of the culture was centrifuged at 12,000 X g for 1 min. The buffer, and detection of the bound second antibody using a chromo- cells were resuspended in 50 p1 of M9 buffer containing glucose (0.2%) genic phosphatase substrate (Sigma). and thiamine (2 pg/ml) and incubated at 36 “C for 10 min, followed A positive control for a Ub-specific processing protease activity by the addition of 20 pCi of Tran~~~S-label(ICN) and a further 2- was an extract prepared from S. cerevisiae (strain BWG1-7a): cells min incubation. Unlabeled L-methionine was then added to a final from 1.5 ml of a stationary culture grown in YPD were pelleted by concentration of 1 mM, the cells were incubated for 10 min at 36 “C, centrifugation, washed twice with HzO, and resuspended in 0.15 ml then centrifuged at 12,000 X g for 30 s, washed once in cold M9 of 1% (v/v) Triton X-100, 1 mM Na-EDTA, 50 mM Na-HEPES (pH buffer, and resuspended in 50 p1 of lysis buffer (4% SDS, 125 mM 7.5), containingantipain, chymostatin, leupeptin, pepstatin, and Tris-HCI, pH 6.8), followed immediately by heating at 100°C for 4 aprotinin, each at 25 pg/ml, and (freshly added) 1 mM phenylmeth- min. The sample was then added to 1 ml of immunoprecipitation ylsulfonyl fluoride. 0.3 g of 0.5-mm glass beads was then added, and buffer (1%Triton X-100, 0.5% Na-deoxycholate, 0.15 M NaCl, 5 mM the cells were disrupted by vortexing at 4 “C for 1 min three times, Na-EDTA, 50 mM Na-HEPES, pH 7.5, plus freshly added 1 mM with 1-min pauses on ice. The suspension was centrifuged at 12,000 phenylmethylsulfonyl fluoride), centrifuged at 12,000 X g for 10 min x g for 5 min, and the supernatantwas used as a positive control in at 4 “C,and theupper 0.9 ml of the supernatantwas added to 6 pl of the in vitro deubiquitination assays. a concentrated hybridoma supernatant containing a monoclonal an- In Vivo Assay for Ub-specific Protease Activity-The UBPl pro- tibody to Pgal (Bachmair et al., 1986). The sample was incubated in tease activity with one of the test proteins, DHFR-Ub-M-pgal, in a microcentrifuge tube for 1 h at 0°C. 10 p1 of protein A-agarose which ubiquitin bore an amino-terminal and acarboxyl-terminal (Repligen) was then added; the suspension was incubated with rock- extension, was assayed using E. coli that expressed both UBPl and ing for 30 min at 4”C, and centrifuged for 15 s at 12,000 X g. The DHFR-Ub-M-pgal. A plasmid (p DHFRUbMPgal) that expressed the pellet was washed three times with immunoprecipitation buffer con- latter protein in E. coli from the yeast GALlO promoter was containing 0.1% SDS, dissolved in the SDS-PAGE, sample buffer, and structed by inserting the -0.67-kb BarnHIIHindIII fragment from electrophoresed in a 6% polyacrylamide-SDS gel, followed by fluoro- the plasmid pMet-1, which contained the complete mouse DHFR graphy. coding sequence (Bachmair and Varshavsky, 1989), into the KpnI Plasmid Manipulations and DNA Sequencing-Plasmids were con- site-containing version of the pUB23 plasmid (Gonda et al., 1989). structed using standard procedures (Ausubel et al., 1989). DNA frag- Site-directed mutagenesis using the Muta-Gene kit (Bio-Rad) was ments were isolated from agarose gels using Geneclean (BiolOl, La employed to position the start codon of the DHFR reading frame Jolla, CA). The YCp50-based plasmid (named pJT55) thatconferred downstream of the GALlO promoter of pUB23 (Bachmair etal., 1986) deubiquitinating activity onto E. coli (see above and “Results”) had a 12024 Protease Ubiquitin-specific

-15-kb yeast genomic DNA insert (data not shown). SphI digestion et dl., 1988). Specifically, thesib selection strategy(Mc- of pJT55 yielded a -10-kb fragment which, upon subcloning into the Cormick, 1987; Lederberg, 1989) was used to isolate a gene SphI-cut vector pUC19 (Yanisch-Perron et al., 1985), conferred the same deubiquitinating activity (plasmid pJT57). The -10-kb yeast from a S. cereuisiae genomic DNA library that could confer DNA insert of pJT57 was cut with SphI and XhoI, yielding a -5.5- Ub-specific protease activity onto E. coli. To be clonable by kb fragment, which was subclonedinto pUC19. The resulting plasmid, this strategy, a yeast gene must have a promoter that func- pJT60, alsoconferred deubiquitinatingactivity onto E. coli (see tions in E. coli (a minority of yeast promoters are able to do “Results”). so), must lack introns incoding region (most yeastgenes lack Since further subcloningof the pJT6O insert, using itssingle KpnI introns),and must encode a proteinthat functions as a site, did not yield a deubiquitinating activity-conferring fragment, the two fragments of the KpnI-cut, -5.5-kb insert of pJT6O were sub- monomer or a homooligomer and that is active in E. coli cloned into M13mp18 and M13mp19 (Yanisch-Perron et al., 1985), extracts. and sets of nested deletions were generated using the3’ exonuclease E. coli transformedwith a S. cereuisiue genomic DNA activity of T4 DNA polymerase (Dale et al., 1985). Single-stranded library (see “Experimental Procedures”)were spread on plates M13 DNA was sequenced using the Sequenase kit(U. S. Biochemical to yield -40 individual colonies per plate. Overnight cultures Corp.) underconditions recommended by themanufacturer. The were then seeded by pooled inoculums from a replica of each internal KpnI site in pJT6O was found to reside within a 2427-bp open reading frame (ORF),which was sequenced on both strands. plate. Extracts from these cultureswere tested for their ability The sequence information was then used to subclone further the to deubiquitinate UB-M-DHFR(Fig. LA IV), and one of the pJT6O insert. Following digestion with AccI, the 5’ overhangs of the extracts showed the activity (Fig. 2). Overnight cultures were resultant -2.8-kb fragment were filled in with Klenow PolI, followed then made asabove from each of the -40 individual colonies by ligation of SalI linkers. The fragment was then digested with SalI on the plate thatyielded the positive extract. Extracts from and RamHI, purified by agarose gel electrophoresis, and subcloned these cultures were tested for the deubiquitinating activity into SalIIBamHI-cutpUC19. The resultantplasmid, JT70, conferred the deubiquitinating activity ontoE. coli as efficiently as the initial with Ub-Met-DHFR, and a single positive extract was ob- pJT57 plasmid (data not shown), indicating that the corresponding tained. Ub-specific protease, named UBPl, was encoded by the 2427-bp ORF A plasmid(named pJT55) was then isolatedfrom the within the -2.8-kb insert of pJT70 (Fig. 1B). individual E. coli transformant that yielded the positive ex- For experiments that involved the simultaneous presence of two tract. Its -15-kb yeast DNA insert was subcloned to yield a distinct plasmids inE. coli, the UBPl-containing insertof pJT70 was -5.5-kb fragment (the plasmidpJT6O) that was indistinguish- excised with SalI and BamHI, and subcloned into the SalIIBamHI- cut vector pACYC184 (NewEngland Biolabs),yielding pJT184. able from the original insert inconferring the deubiquitinating pACYC184 bears the P15Aorigin of replication, which makes it activity onto E. coli (see “Experimental Procedures”). Since compatible with pMBl(ColE1)-based E. coli vectors such as pUCl9 further subcloning, using the KpnI site in themiddle of the and pBR322 (Sambrook et al., 1989). 5.5-kb insert of pJT60, did not produce activity-conferring The nucleotide and (predicted) amino acidsequences of the UBPl fragments, the insert was sequenced in both directions from gene and protein were compared to sequences in the GenBank (ver- the KpnI site, whichwas found to reside within an open sion 66.0), EMBL (version 23.0), and GENPEPT (a translated version of GenBank) databases, using the FastA algorithm (Pearson andreading frame(ORF). This sequence information was used to Lipman, 1988). subclone further the insert in pJT60, yielding pJT70 (Fig. Construction of a Null ubpI Mutant-A ubpl null allele was made lB), whose -2.8-kb, activity-conferring insert contained the in uitro by replacing 81% of the UBPl coding region with a fragment 2427-bp ORF (Fig. 3) encoding an acidic (calculated PI of carrying the yeast URA3 gene. First, the entire -15-kb yeast DNA 4.63), 809-residue (-93 kDa) protein,which wasnamed UBPl insert of pJT57 was subcloned into pUC9 to yield pJT9, which was (for ubiquitin-specific protease). This designation conforms then digested with ClaI and KpnI. The resultant -2-kb fragment, to the existing convention for naming genes encoding com- encompassing 81% of the UBPl ORF, was replaced with the URA3- containing -1.1-kb fragment of pRBU1, yielding pJT9K. The plasmid ponents of the ubiquitin system in S. cereuisiae ( UBIl-UBI4 pRBU1, kindlyprovided by Dr. R. T. Baker, was constructed by for the ubiquitingenes (Ozkaynak et al., 1984, 1987; Finley et isolating the -1.1-kb HindIIIISmaI fragment from YEp24 (Botstein al., 1987, 1988); UBC1-UBC7 for the genesencoding Ub- et al., 1979), fillingin theHind111 site usingKlenow PolI, and ligating conjugating enzymes (Jentsch et al., 1987, 1990); UBRl for the blunt-ended fragment into SmaI-cutpUC19. the gene encoding the recognition component of the N-end The -7.7-kb linear DNA from SalI-cut pJT9K that included the deletion/disruption ubpl-AI allele was usedto transform bothhaploid rule pathway, one Ub-dependentproteolytic pathway (Bartel (BBY45) and diploid (DF5) ura3 yeast strains. Southern hybridiza- et al., 1990)). tion analysis of haploid and diploid Ura’ transformants (with the The position of the start (ATG) codon in UBPl was in- latter expected to be heterozygous at the UBPl locus) confirmed the ferred so as toyield the longest ORF. No ATGs occur in any predicted structure of the ubpl-A allele by showing the presence of of the threeforward reading frames upstreamof this putative restriction endonuclease sites diagnostic of the transplacement (data not shown) (Rothstein,1983).

Ub-M-DHFR- RESULTS M-DHFR- The UBPl Gene-The previously cloned Ub-specific proc- > essing proteaseof S. cereuisiae, YUH1, cleavesubiquitin from relatively short carboxyl-terminal extensionsof ubiquitin but abcd is virtually inactive with larger ubiquitin fusions suchas Ub- FIG. 2. Cloning of the UBPl gene. A sib selection strategy was Pgal (Miller et al., 1989).2 The efficient deubiquitination of used to isolate a gene from a S. cereuisiae genomic DNA library that the latterfusion in yeastcells and cell-free extracts (Bachmair could conferUb-specific protease activity onto E. colicells that carried et al., 1986; Bachmair and Varshavsky, 1989; see also below), such a gene. Lane a, partial in uitro deubiquitination of the ”S-labeled as well as the properties of null yuhl mutants (Miller et al., Ub-M-DHFR by an extract from the (pooled) 40 individual E. coli tranoformanto producod uoing tho ycnot DNA lihary (stiti 1080) indicatcd that S. ~-cr.suiduecullLains a h1lily of Ub- IlD337 specific processingproteases. To clone thecorresponding “Experimental Procedures”). Lune b, same as lane a but an extract from apool that lacked the deubiquitinatingactivity. Lane c, purified, genes, we took advantage of the fact that bacteria suchas E. “S-labeled Ub-M-DHFR (control). Lane d, same as lane c but after coli lack eukaryoticubiquitin andUb-specific enzymes (Finley incubation with a whole cell yeast extract (see “Experimental Pro- cedures”). Arrowhead denotes an unrelated cleavage product of Ub- J. Tobias and A. Varshavsky, unpublished results. M-DHFR thatformed in the control (UBP1-lacking)E. coli extract. Ubiquitin-specific Protease 12025

-192 "T"ATAGWAllWAAGWTAGPAGXATETiXk" -104 TTTCCATTATAA~TCATTIT~C7AGlT"MTCT -12 ~A~A~TMCAGTITA~TlnTA~~T MDLFIESKINSLLQFLFGSRQDFLRN 26 79 ~~TM~~~~A~~M~T~A~T~T FKTWSNNNNNLSIYLLIFGIVVFFYKKPDH 56 169 Cl'AMCTACA-AGTGMAPTMTMTWP LNYIVESVSEMTTNFRNNNSLSRWLPRSIF 86 AczcxxT" 259 AczcxxT" "=-A- THLDEEILKRGGFIAGLVNDGNTCFMNSVL 116 349 CAA~TCATUXGAGRAlTPTMTGXATAAQZ?GC~T~TGMCACA&T QSLASSRELMEFLDNNVIRTYEEIEQNEHN 146 439 ~KT3XQEA~~~A~MWATAAG EEGNGQESAQDEATHKKNTRKGGKVYGKHK 176 529 MG"T" TAPAAGa2ATCTA KKLNRKSSSKEDEEKSQEPDITFSVALRDL 206 619 ~"ITA~T~A~T~A~~V LSALNAKYYRDKPYFKTNSLLKAMSKSPRK 236 709 AATA-?dXAWAAP T~GPAAGTAKGITAMXAllGAAT NILLGYDQEDAQEFFQNILAELESNVKSLN 266 799 hcpGAAAAAcTAGTACXA~~lT~~~~CITM~CT TEKLDTTPVAKSELPDDALVGQLNLGEVGT 296 889 OmA(aTTCCAACPGAACAGATTGATOCTAACTCTATACTR VYIPTEQIDPNSILHDKSIQNFTPFKLMTP 326 979 T~TMGATA~A~~~A LDGITAERIGCLQCGENGGIRYSVFSGLSL 356 1069 AATTT~TA~AWXITA~A~GT~TCA~m NLPNENIGSTLKLSQLLSDWSKPEIIEGVE 386 TOPAAaX;ITOPGCCCICACAGCACCGCACTCTULmATTT FIG. 3. Nucleotide sequence of the 1159 CNRCALTAAHSHLFGQLKEFEKKPEGSIPE 416 UBPl gene and deduced amino acid 1249 ACAGTTATTG?CGA"lTATCAA sequence of the UBPl protein. KPINAVKDRVHQIEEVLAKPVIDDEDYKKL 446 1339 CdTA~TA~A~~T~~~M~~~M~~~~~AlTA~~TA~~ HTANMVRKCSKSKQILISRPPPLLSIHINR 476 1429 WTTlGATCCAAGUC%TACdTGAlT~TM~ATXTTTAAGTCAAGGTXSA~~ SVFDPRTYMIRKNNSKVLFKSRLNLAPWCC 506 1519 GATATAATGAAATTCOXMCCUATGT~~TGhuacATT DINEINLDARLPMSKKEKAAQQDSSEDENI 536 1609 ~T~ATA~~ACdlUA~~~~~~~~TA~~~ GGEYYTKLHERFEQEFEDSEEEKEYDDAEG 566 1699 MCTATGCGTCTCATTACAATCATACCAAGGATATCAGTA NYASHYNHTKDISNYDPLNGEVDGVTSDDE 596 1789 U\TGAC.TACAITGAAWLAACCULTGCTPTACAAT~~T~Tm~~~T~~~~ DEYIEETDALGNTIKKRIIEHSDVENENVK 626 1879 G~~T~~A~~T~T~ DNEELQEIDNVSLDEPKINVEDQLETSSDE 656 1969 ULAGI\TOPTATACCAGCKYaCCTAmAATTATGCTAGGT EDVIPAPPINYARSFSTVPATPLTYSLRSV 686 2059 A~C7~A~TMlTATGGmATTACATTG IVHYGTHNYGHYIAFRKYRGCWWRISDETV 716 2149 TA~~~~A~~ATTIT~TAT~~~c~~GA~ YVVDEAEVLSTPGVFMLFYEYDFDEETGKM 746 2239 MGGATGAlnGGW&XATKKAGTMTMlUAGAAGATGA"A~~Q3AAGC KDDLEAIQSNNEEDDEKEQEQKGVQEPKES 776 2329 C~~GA~~~Cd~~CXAT~~TA~~ QEQGEGEEQEEGQEQMKFERTEDHRDISGK 806 2419 VGTGT~TAA~TM~ATM~T~~~TM~A~A~ D V N ter 809 2509 G~~~~TA~TAT~~A 2599 M~C7CC2653 initiator ATG until position -200 (Fig. 3). The codon adap- the purified 3sS-labeled testproteins) (Fig. 4, A-C) or by tation index (calculated according to Sharp and Li, 1987) for immunoblotting with anti-Ub antibodies (Fig. 40). UBPl is 0.19, characteristic of weakly expressed yeast genes. TheUBPl protease was ATP-independent(data not Computer-aided comparisons of the UBPl nucleotide se- shown), and cleaved ubiquitin from the tested carboxyl-ter- quence and of the predicted amino acid sequence of UBPl to minal extensions of ubiquitin, such as Ub-X-pgal or UB-X- sequence databases (see "Experimental Procedures") revealed DHFR, where X is any of the 20 amino acid residues except no significant similarities to sequenced genes or gene prod- proline (Figs. 1, 4A, and data not shown). When the residue ucts. In particular, UBPl lacks sequence similarities to pro- X was proline, the UBP1-mediated deubiquitination was teinsthat are either known or expected to interact with strongly inhibited (Fig. 4A). This pattern of substrate speci- ubiquitin, including several yeast UB-conjugating (UBC) en- ficity was indistinguishable from that observed with S. cere- zymes (Jentschet al., 1990), UBCs from other organisms uisiae using the same test proteins either in uiuo (Bachmair (Sullivan and Vierstra, 1989; Reynolds et al., 1990; Schneider et al., 1986; Bachmair and Varshavsky, 1989) or in whole cell et al., 1990), and, in addition, more surprisingly, the YUHl extracts (Fig. 1 and data notshown). Ub-specific processing protease (also called ubiquitin hydro- UBPl could also deubiquitinate natural ubiquitin fusions lase) (Miller et al., 1989) and itsmammalian homologs (Wilk- such as UBI1-UBI3, which are precursors of both ubiquitin inson et al., 1989). and specific ribosomal proteins (see Introduction). Deubiqui- Substrate Specificity of the UBPl Protease-Extracts from tination of UBI2 and UBI3by the UBPlprotease (the amino E. coli carrying either the pUC19 vector alone or the pUC19- acid sequence of UBI1 is identical to thatof UBI2 (Ozkaynak based pJT7O plasmid that expressed the yeast UBPl protease et al., 1987)) was demonstrated by mixing extracts of E. coli were added to a variety of Ub-containing test proteins (see that either contained or lacked UBPl with E. coli extracts "Experimental Procedures"). The reaction products were frac- containing UBI2 or UBI3. Deubiquitination of the test sub- tionated by SDS-PAGE, followed either by fluorography (for strates was monitored by SDS-PAGE and immunoblotting 12026 Proteme Ubiquitin-specific

2.5 kb E E E 1 kI UBPl )1 H V - ROk I kb

I,+- I,+- .@? E E

3.7 kb B UBIJ- FIG.5. Wild-type and null alleles of the UBPl gene. Re- placement of the wild-type UBPI allele with the deletion/disruption ubpl-A1::URAd allele generates a new arrangement of EcoRI sites detectable by Southern hybridization analysis of Ura’ S. cereuisiae transformants using the indicated DNA probe. (See “Experimental Procedures” for details.)

proteases whose in vitro substrate specificity is differentfrom those of UBPl and YUH1. More recent data3 indicate that uhpl-AI ahcde Igh i UBP1, while inactive with polyubiquitin in vitro (in E. coli FIG. 4. Substrate specificity of the UBPl protease. A, puri- extracts), was ableto cleave itin uiuo, when UBPl and fied, :”S-labeled Ub-M-@gal (Fig. IA, V)and Ub-P-@gal (the latter polyubiquitin were expressed together in E. coli (see “Discus- containing proline instead of methionine at the Ub-@galjunction) sion’’). were incubated with UBPI-containing or UBP1-lacking E. coli extracts, orwith a S. cereuisiae extract, asindicated on topof the lanes, The UBPlprotease also did not cleave in uitro a branched followed by electrophoretic analysisof pgal. The bandsof Ub-X-pgal ubiquitin conjugate suchas the multiubiquitin chain (see and X-pgal are indicated on the left. B, kinetics of in oitro deubiqui- Introduction). Specifically, no cleavage of the invitro-formed, tination of the ‘”S-labeled Ub-M-@gal (Fig. lA, V)by extracts from P-galactosidase-attached multiubiquitin chain (Chau et al., S. cereoisiae carryingeither the wild-type ordisruption/deletion 1989) by the UBPl protease was observed when the UBP1- (ubpl-AI) alleles of theyeast UBP1 gene. After incubations for containing E. coli extract was added to %-labeled, multiubi- specified periods of time, p-galactosidasewas analyzed by SDS- PAGE. C, in oioo deubiquitination of DHFR-Ub-M-pgal (Fig. lA, VI) quitinated @-galactosidase(data not shown). in [:%]methionine-labeled E. coli expressing the yeast UBPI gene. TheUBPl proteasedeubiquitinated carboxyl-terminal Lane a, E. coli expressing both UBPl and Ub-M-Bgal (Fig. L4, V) ubiquitin fusionseven if the ubiquitinmoiety bore in addition were labeled with [“‘Slmethionine, followed by extraction, immuno- an amino-terminal extension. This was shown by expressing precipitation, and electrophoretic analysisof &galactosidase. Lane b, a fusion protein DHFR-Ub-M-pgal (Fig. lA, VI) in E. coli same as lane a but with E. coli lacking the yeast UBPI gene and either by itself or together with UBPl (see “Experimental expressing DHFR-Ub-M-Bgal. Lane c, same as laneb but withE. coli expressing both UBPl and DHFR-Ub-M-@gal. Arrowheads indicate Procedures”). The cells were labeled with [”S]methionine, the species which are either partial (apparently in oioo) degradation extracts were prepared, and immunoprecipitatedwith a mono- products of DHFR-Ub-M-@gal, withcleavages in its DHFR moiety, clonal antibodyto p-galactosidase (Fig. 4C). When E. coli did or the primary translation products due to initiations within the not carrya UBPl -expressing plasmid, severalprotein species, DHFR-coding region of the ORF for DHFR-Ub-M-@gal (allof these including the one whose apparent molecular mass equaled species were converted into M-cgal in the presence of the UBPl that of DHFR-Ub-M-@gal, were precipitated by the anti-pgal protease;lane c). D, in vitro deubiquitination assays with natural ubiquitin fusions. Lane a, an extract from E. coli expressing UBI2 antibody (Fig. 4C, lane b). However, in E. coli that expressed (Fig. lA, I)was mixed with an extract from E. coli lacking the yeast both DHFR-Ub-M-pgal and UBP1, all of these P-galactosid- UBPl protease, followed by an incubation, SDS-PAGE, and detection ase species were converted to the one thatcomigrated with of ubiquitin-containing species using immunoblot analysis with an M-pgal (Fig. 4C, lane c). Note that smaller p-galactosidase- anti-Ub antibody (see “Experimental Procedures”). Lane b, same as containing proteins (Fig. 4C, arrowheads) in E. coli lacking lane a but with the second extract being from yeast. Lane c, same as UBPl and expressing DHFR-Ub-M-@gal were due either to lane b but with the second extract beingfrom E. coli expressing UBPI. Lane d, same as lane a but with the first extract being from alternativetranslation start sites or proteolytic cleavages E. coli expressing UBI3 (Fig. lA,II). Lane e, same as laned but with within the DHFRmoiety of the fusion protein, inasmuch as the second extract being from yeast. Lane f, same as lanee but with all of these species were converted by UBPl intoM-pgal (Fig. the second extract being from E. coli expressing UBPl. Note that 4C, lune c). UBPl cleaves ubiquitin from both UBI3 and its partial degradation Null ubpl Mutants-Haploid S. cereuiside strains in which products seen in c. Lane g, same as lanea, but with the first extract 81% of the UBPl ORF was deleted and replaced with the being from E. coli expressing UBI4 (Fig. lA, III). Lane h, same as lane g but with the second extract being from yeast. Lane i, same as URA3 gene were constructed asdescribed under “Experimen- lane h but with the second extract beingfrom E. coli expressing tal Procedures” (Fig. 5). These ubpl-A1 strainsgrew normally UBPI. Partially deubiquitinated derivatives of UBI4 (which is de- and lacked obvious phenotypicabnormalities (data not noted as UB5) are indicated on the right. (See “Experimental Proce- shown). In addition, extracts from ubrl-A1 cells were indis- dures” for details.) tinguishable from wild-type extracts in their ability todeubi- quitinate ubiquitinfusions such as Ub-M-DHFR andUb-M- with anti-Ub antibodies(Fig. 40). pgal (Fig. 4B and data not shown), indicating that thefamily Polyubiquitin, a naturally occurring spacerless, head-to-tail of Ub-specific proteases in yeast is not limited to UBPl and fusion of several ubiquitin moieties (Ozkaynak et al., 1984, YUH1. 1987; Dworlrin Root1 ct al., 1084;Schlasillgar a~dDud, 1987), is the only linear ubiquitin fusion that was not cleaved in a DISCUSSION UBP1-containing E. coli extract; however, the same yeast We have cloned a gene for UBP1, a new member of the polyubiquitin (encoded by UBZ4 (Ozkaynak et al., 1987)) was family of Ub-specific processing proteases in S. cereuisiae. converted into mature ubiquitinby a whole-cell yeast extract (Fig. 40), indicating that S. cereukiue contains Ub-specific ’R. Baker and A. Varshavsky, manuscript in preparation. Ubiquitin-specific Protease 12027 The UBPl protease is an 809-residue protein that has no eukaryotes was inferred from the existence of polyubiquitin significant sequence similarities to known proteins,including genes (Ozkaynak et al., 1984,1987; Dworkin-Rastl eta1.,1984), the only other previously cloned yeast Ub-specific processing and from the demonstration that engineered ubiquitinfusions protease, the 236-residue YUHl (Miller et al., 1989). While such as Ub-X-pgalswere precisely deubiquitinated in vivo or the YUHlprotease is virtually inactive againstrelatively long in cell-free extracts (Bachmair etal., 1986; Gonda etal., 1989). ubiquitin fusions such as Ub-@gal (Miller et al., 1989),2 the Another finding, that Ub-X-@galswere deubiquitinated irre- UBPl protease of the present work cleaves at the carboxyl spective of the identityof a residue X at theUb-@gal junction, terminus of ubiquitin in natural and engineered fusion pro- led to the discovery of the N-end rule, which relates the in teins irrespective of their size or the presence of an amino- vivo half-life of a protein to the identityof its amino-terminal terminal ubiquitin extension(see “Results”). The latter prop- residue (Bachmair et a1.,1986). Distinct versionsof the N-end erty of UBPl makes it possible to express ubiquitin fusions rule havesince been shown tooperate in alleukaryotes to proteins of interest as “triple” fusions, whose amino-ter- examined, from yeast to mammals (Bachmair et al., 1986; minalubiquitin extension, e.g. thebiotin-binding protein Varshavsky et al., 1988; Reiss et al., 1988; Bachmairand streptavidin, could be used to purify the fusion by affinity Varshavsky, 1989; Gonda et al., 1989; Johnson et al., 1990; chromatography (from bacteria such as E. coli that lack Ub- Bartel et al., 1990). However, since ubiquitin and Ub-specific specific proteases) before the UBP1-mediated deubiquitina- enzymes are absent from bacteria (Finley et al., 1988), the tion step. ubiquitin fusion technique could not be used to determine Among the ubiquitin fusions tested in the present work, whether the N-end rule operates in prokaryotes aswell. The polyubiquitin is theonly linear fusion that is nota substrate isolation of UBPl has allowed us to bypass this difficulty, as forthe UBPl protease in vitro (see “Results”).Since the E. coli transformed witha UBPl -expressing plasmid acquires amino-terminal region of ubiquitin,and in particular its the ability to deubiquitinate Ub-X-pgals irrespective of the amino-terminal methionineresidue, is an integral partof the identity of the junctional residue X, proline being the single compact ubiquitin globule (Vijay-Kumar et al., 1987; Wilkin- exception.Measurements of thein vivo half-lives of the son, 1988), thecleavable Gly-Met bondbetween two adjacent resultant X-@gals inE. coli have shown that a distinct version ubiquitin moieties in polyubiquitin may be sterically shielded of the N-end rule operates in bacteria well.4as to a greater extent than areanalogous Gly-X bonds inubiqui- Availability of the UBPl gene andthe purified UBPl tin fusions that lack the Ub-Ub motif. It has recently been protease will increase the versatility of the ubiquitin fusion- found that UBP1,while unable to deubiquitinatepolyubiqui- basedtechnique for thein vivo or in vitro generation of tin in vitro, can do so in vivo, when coexpressed with polyu- proteins (including short polypeptides) that bear predeter- biquitinin E. Alikely explanationis that the UBP1- mined residues at their amino termini, a method with appli- mediated deubiquitinationof polyubiquitin invivo is cotrans- cations in both basic research (Baker and Varshavsky, 1991; lational or, more generally, occurs before the folding of the Bartel etal., 1990; Johnson et al., 1990; Townsend etal., 1988) newly formed, multidomain polyubiquitin protein complete.is and biotechnology (Lu et al., 1990; Butt et al., 1989; Sabin et UBPl also did not cleave in vitro a branched ubiquitin al., 1989; Ecker et al., 1989; Yo0 et al., 1989; Mak et al., 1989). conjugate such as the multiubiquitin chain (see Introduction Acknowledgments-We thank Vincent Chau for the anti-Ub anti- and “Results”). It remains to be determined whether UBPl body and Rohan Baker for the pRBUlplasmid. We also thank Erica fails tocleave the multiubiquitin chain in vivo as well. Johnson, Tom Shrader, and especially Rohan Baker for comments Although UBPl is able to cleave all of the tested natural on the manuscript and Barbara Doranfor secretarial assistance. and engineered ubiquitin fusions at least in vivo (see above), and although the processing of natural ubiquitin precursors REFERENCES is essential for cell viability (Finley et al., 1989), UBPl is a Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, nonessential gene: the null ubpl mutant is viable, grows at a J. G.,Smith, J. A., and Struhl, K. (eds)(1989) Current Protocols in wild-type rate, and is otherwise phenotypically normal (see Molecular Biology, John Wiley and Sons, New York Bachmair, A,, and Varshavsky,A. (1989) Cell 56, 1019-1032 “Results”). Moreover, the double ubpl yuhl mutant, which Bachmair, A,, Finley, D., and Varshavsky, A. (1986) Science 234, lacks both the UBPl and YUHl proteases, is also phenotyp- 179-186 ically normal, and retains the ability to deubiquitinate both Baker, R. T., and Varshavsky, A. (1991) Proc. Natl. Acad. Sci. U. S. linear and branched Ub-proteinconjugates.3 The multiplicity A. 88,1090-1094 of functionally overlapping Ub-specific processing proteases Ball, E., Karlik, C. C., Beall, C. J., Saville, D. L., Sparrow, J. C., Bullard, B., and Fyrberg, E. A. (1987) Cell 51, 221-228 in S. cerevisiae accounts, in hindsight, for the failure of our Bartel, B., Wunning, I., and Varshavsky,A. (1990) EMBOJ. 9,3179- earlier efforts to clone the genes for these proteases using 3189 yeast-based genetic screens.While the strategyof sib selection Bonner, W., Hatch, C. L., and Wu, R. S. (1988) in Ubiguitin (Re- and functional cloningin E. coli that yielded the UBPl gene chsteiner, M., ed) pp.157-172, Plenum Press, New York bypassed the problem of genemultiplicity, it was initially Botstein, D., Falco, S. C., Stewart, S. E., Brennan, M., Scherer, S., Stinchcomb,D. T., Struhl, K., and Davis, R. W. (1979) Gene viewed as a strategy unlikely to succeed for other reasons(see (Amst.) 8,17-24 “Results”). More recent work: using another E. coli-based Boyer, H. W., and Roulland-Dussoix,D. (1969) J. Mol. Bid. 41,459- genetic screen, has yielded two more S. cerevisiae genes for 472 Ub-specific processing proteases, named UBP2 and UBP3. Butt, T., Jonnalagadda, S., Monia, B. P., Sternberg, E. J., Marsh, J. Thus, the yeastUBP gene family contains at least fourgenes, A., Stadel, J. M., Ecker, D. J., and Crooke, S.T. (1989) Proc. Natl. including YUHI; a variety of evidence indicates that other Acad. Sci. U. S. A. 86,2540-2544 Chau, V., Tobias, J. W., Bachmair, A., Marriott, D., Ecker, D. J., eukaryotes have at least as many (Rose, 1988; Wilkinson et Gonda, D. K., and Varshavsky, A. (1989) Science 243, 1576-1583 al., 1989; Jonnalagadda et al., 1989). The continuing genetic Ciechanover, A., and Schwartz, A. L. (1989) TrendsBiochem. Sci. and biochemical analyses of the single and multiple null ubp 14, 483-488 mutants should allow more precise definitions of the appar- Dale, R. M. K., McClure, B. A., and Houchins, J. P. (1985) Plasmid ently overlapping functions of these proteases. 13, 31-40 Prior to the cloning of YUHl and UBPl, the presence of ‘ J. Tobias, T. Sbrader, G. Rocap, and A. Varshavsky, manuscript UB-specific processing proteases in S. cerevisiae and other in preparation. 12028 Ubiquitin-specificProtease Dworkin-Rastl, E., Shrutkowski, A., and Dworkin, M. B. (1984) Cell Monia, B. P., Ecker, D. J., and Crooke, S. P. (1990) BiolTechnology 39,321-325 8,209-215 Ecker, D. J., Stadel, J. M., Butt, R. T., Marsha, J. A., Monia, B. P., Olson, T. S., and Dice, J. F. (1989) Curr. Opin. CellBiol. 1, 1194- Powers, D. A., Gorman, J. A., Clark, P. E., Warren, F., Shatzman, 1200 A., and Crooke, S. T. (1989) J. Biol. Chem. 264, 7715-7719 Ozkaynak, E., Finley, D., and Varshavsky, A. (1984) Nature 312, Finley, D., Ciechanover, A., and Varshavsky, A. (1984) Cell 37, 43- 663-666 55 Ozkaynak, E., Finley, D., Solomon, M. J., and Varshavsky, A. (1987) Finley, D., Ozkaynak, E., and Varshavsky, A. (1987) Cell 48, 1035- EMBO J. 6,1429-1439 1046 Pearson, W. R., and Lipman, D. J. (1988) Proc. Natl. Acad. Sci. U. Finley, D., Ozkaynak, E., Jentsch, S., McGrath, J. P., Bartel, B., S. A. 85, 2444-2448 Pazin, M., Snapka, R. M., and Varshavsky, A. (1988) in Ubiquitin Perry, G., Friedman, R., Shaw, G., and Chau, V. (1987) Proc. Natl. (Rechsteiner, M., ed) pp. 39-75, Plenum Press, New York Acd. Sci. U. S. A. 84, 3033-3036 Finley, D., Bartel, B., and Varshavsky, A. (1989) Nature 338, 394- Pickart, C. M., and Rose, I. A. (1985) J. Biol. Chem. 260,1573-1581 401 Pickart, C. M. (1988) in Ubiquitin (Rechsteiner, M., ed) pp. 77-100, Glotzer, M., Murray, A., and Kirschner, M. (1991) Nature 349, 132- Plenum Press, New York 138 Reiss, Y., Kaim, D., and Hershko, A. (1988) J. Biol.Chem. 263, Gonda, D. K., Bachmair, A., Wunning, I., Tobias, J. W., Lane, W. S., 2693-2698 and Varshavsky, A. (1989) J. Biol. Chem. 264, 16700-16712 Reynolds, P., Koken, M. H. M., Hoeijmakers, J. H. J., Prakash, S., Gregori, L., Poosch, M. S., Cousins, G., and Chau, V. (1990) J. Biol. and Prakash, L. (1990) EMBO J. 9, 1423-1430 Chem. 265,8354-8357 Rose, I. A. (1988) in Ubiquitin (Rechsteiner, M., ed) pp. 135-155, Plenum Press, New York Guarente, L., Yocum, R. R., and Gillard, R. (1982) Proc. Natl. Acad. Rose, M. D.,Novick, P., Thomas, J. H., Botstein, D., and Fink, G. R. Sci. S. A. 79,7410-7414 U. (1987) Gene (Amst.) 60, 237-243 Hershko, A. (1988) J. Biol. Chem. 263, 15237-15240 Rothstein, R. (1983) Methods Enzymol. 101, 202-211 Hochstrasser, M., Ellison, M., Chau, V., and Varshavsky, A. (1991) Sabin, E. A., Lee-Ng, C. T., Shuster, J. R., and Barr, P. J. (1989) Proc. Natl. Acad. Sci. U.S. A. 88, 4606-4610 BiolTechnology 7, 705-709 Ito, H., Fukuda, Y., Murata, K., and Kimura, A. (1983) J. Bacteriol. Sambrook, J., Fritsch,E. F., and Maniatis, T. (1989) Molecular 153, 163-168 Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Jabben, M., Shanklin, J., and Vierstra, R. D. (1989) J. Biol. Chem. Press, Cold Spring Harbor, NY 264,4998-5005 Schlesinger, M., and Bond, U. (1987) Oxf. Sum. Eukaryotic Genes 4, Jentsch, S., McGrath, J. P., and Varshavsky, A. (1987) Nature 329, 77-91 131-134 Schneider. R.. Eckerskom. C.. Lottsaeich.- .. F.. and Schweiger.-. M. Jentsch, S., Seufert, W., Sommer, T., andReins, H.-A. (1990) Trends (1990) EMiO J. 9,143111437 Biochem. Sci. 15, 195-198 Sham. P. M.. and Li. W.-H. (1987) Nucleic Acids Res. 15.1281-1295 Johnson, E. S., Gonda, D. K., and Varshavsky, A. (1990) Nature Sherman, F.,' Fink, G. R., an'd Hicks, J. B. (1986) Methods in Yeast 346,287-291 Genetics, Cold Spring Harbor Laboratory Press, Cold Spring Har- Jonnalagadda, S., Butt, T. R., Monia, B. P., Mirabelli, C. K., Gotlib, bor, NY L., Ecker, D. J., and Crooke, S. T. (1989) J. Biol. Chem. 264, Siegelman, M., Bond, M. W., Gallatin, W. M., St. John, T., Smith, 10637-10642 H. T., Fried, V. A., and Weissman, I. L. (1986) Science 231, 823- Kaufman, B. T. (1974) Methods Enzymol. 34, 272-281 829 Lederberg, J. (1989) Genetics 121, 395-399 Sullivan, m. L., and Vierstra, R. D. (1989) Proc. Natl. Acad. Sci. U. Leung, D.W., Spencer, S. A., Cachianes, G., Hammonds, R.G., S. A. 86,9861-9865 Collins, C., Henzel, W. J., Barnard, R., Waters, M. J., and Wood, Townsend, A,, Bastin, J., Gould, K., Brownleee, G.,Andrew, M., Coupar, B., Boyle, D., Chan, S., and Smith, G. (1988) J. Exp. Med. W. I. (1987) Nature 330, 537-543 168,1211-1224 Lowe, J., Blanchard, A., Morrel, K., Lennox, G., Reynolds, L., Billett, Varshavsky, A,, Bachmair, A., Finley, D., Gonda, D., and Wunning, M., Landon, M., and Mayer, R. J. (1988) J. Pathol. 155,9-15 I. (1988) in Ubiouitin (Rechsteiner.. M..I ed) DD. 287-321, Plenum Lu, C., Yang, Y.-F., Ohashi, H., and Walfish, P. G. (1990) Biochem. Press, New York' Biophys. Res. Commun. 171, 138-142 Vijay-Kumar, S., Bugg, C. E., and Cook, W. J. (1987) J. Mol. Biol. Mak, P., Mcdonnell, D. P., Weigel, N. L., Schrader, W. T., and 194,531-544 O'Malley, B. W. (1989) J. Biol. Chem. 264, 21613-21618 Wilkinson, K. D. (1988) in Ubiquitin (Rechsteiner, M., ed) pp. 5-38, Matsui, S., Sandberg, A. A., Negoro, S., Seon, B. K., and Goldstein, Plenum Press, New York G. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 1535-1539 Wilkinson, K. D., Lee,K., Deshpande, S., Duerksen-Hughes, P., Mayer, A. N., and Wilkinson, K. D. (1989) Biochemistry 28, 166-172 Boss, J. M., and Pohl, J. (1989) Science 246, 670-673 McCormick, M. (1987) Methods Enzymol. 151,445-449 Yanisch-Perron, C., Vieira, J., and Messing, J. (1985) Gene (Amst.) Miller, H. I., Henzel, W. J., Ridgway, J. B., Kuang, W.-J., Chisholm, 33, 103-119Yo0, Y., Rote, K., and Rechsteiner, M. (1989) J. Biol. V., and Liu, C.-C. (1989) BiolTechnology 7,698-704 Chem. 264, 17078-17083