發(fā)信人: newsci (樂(lè)科學(xué)), 信區(qū): BioTrends 標(biāo) 題: Nature:蛋白質(zhì)降解機(jī)理:偶聯(lián)去泛素化和降解的酶 發(fā)信站: 生命玄機(jī)站 (Thu Sep 26 05:29:09 2002) , 轉(zhuǎn)信 26 September 2002 Nature 419, 403 - 407 (2002); doi:10.1038/nature01071 Nature AOP, published online 1 September 2002 A cryptic protease couples deubiquitination and degradation by the proteasome TINGTING YAO AND ROBERT E. COHEN Department of Biochemistry, University of Iowa, 51 Newton Road, Iowa City, Iowa 52242, USA Correspondence and requests for materials should be addressed to R.E.C. (e-mail: robert-cohen@uiowa.edu). The 26S proteasome is responsible for most intracellular proteolysis in eukaryot es1, 2. Efficient substrate recognition relies on conjugation of substrates with multiple ubiquitin molecules and recognition of the polyubiquitin moiety by the 19S regulatory complex—a multisubunit assembly that is bound to either end of the cylindrical 20S proteasome core. Only unfolded proteins can pass through nar row axial channels into the central proteolytic chamber of the 20S core, so the attached polyubiquitin chain must be released to allow full translocation of the substrate polypeptide. Whereas unfolding is rate-limiting for the degradation o f some substrates and appears to involve chaperone-like activities associated wi th the proteasome3-5, the importance and mechanism of degradation-associated deu biquitination has remained unclear. Here we report that the POH1 (also known as Rpn11 in yeast) subunit of the 19S complex is responsible for substrate deubiqui tination during proteasomal degradation. The inability to remove ubiquitin can b e rate-limiting for degradation in vitro and is lethal to yeast. Unlike all othe r known deubiquitinating enzymes (DUBs) that are cysteine proteases6, 7, POH1 ap pears to be a Zn2+-dependent protease. To study the steps of proteasomal degradation that occur after unfolding, we dev eloped a substrate containing an amino-terminal ubiquitin fused to the ovomucoid first domain8. We used this substrate because it is amenable to unfolding and f luorescent labelling. The ovomucoid moiety (OM) was unfolded irreversibly by dis ulphide bond reduction and alkylation of six cysteines with a fluorophore, lucif er yellow (Fig. 1a). The ubiquitin–ovomucoid (UbOM) substrate also included an N-terminal six-His tag and a haemagglutinin (HA) epitope at the ubiquitin?Covomucoid junction. Figure 1 Proteasomal degradation of UbOM is targeted by means of the unfolded o vomucoid moiety. Full legend High resolution image and legend (38k) We mutated G76 to V76 in ubiquitin, as this mutation (G76V) can prevent release of ubiquitin by DUBs9. Ub(G76)OM is degraded rapidly by the 26S proteasome in an ATP-dependent manner (Fig. 1b). When we compared the degradation kinetics of Ub OM with ovomucoid alone (Fig. 1c), we found that both are degraded with the same Michaelis constant (Km), indicating that these substrates are targeted through the unfolded ovomucoid moiety and supporting previous conclusions that monoubiqu itin does not bind well to proteasomes. However, on mutation of G76 to V76 in ub iquitin (Ub(V76)OM), the catalytic rate constant (kcat) was reduced tenfold, whe reas Km was unchanged. Thus, the ubiquitin moiety appeared to have a role in a p ost-targeting step of proteasomal processing. The different degradation rates of the UbOM substrates suggested that Ub(G76)OM was deubiquitinated during degradation whereas with Ub(V76)OM the ubiquitin moie ty was degraded together with the unfolded OM extension. Figure 2a shows that th is was indeed the case. Degradation of Ub(G76)OM, but not Ub(V76)OM, generated a product that was recognized by anti-ubiquitin but not anti-HA antibody. We isol ated this product on Ni2+ resin and determined by matrix-assisted laser desorpti on/ionization–time of flight (MALDI-TOF) mass spectrometry that its molecular mass was within 8 Da of six-His-tagged ubiquitin (data not shown). Thus, the ubiquitin moiety was processed precisely at G76. For two reasons, these results indicated the action of an unusual DUB. First, the degradation reactions contained 2 μM ubiquitin-aldehyde (Ubal), a potent DUB inhibitor10. Pre-incubation of purified proteasomes with Ubal eliminated deubiquitination of K48-linked Ub2 and ubiquitin Asp 77, whereas even fivefold more Ubal did not inhibit UbOM degradation (data not shown). Therefore, Ub(G76)OM is processed by a Ubal-insensitive DUB. Second, deubiquitination by this activity promotes degradation (Fig. 1c). When incubated without Ubal, Ub(G76)OM was rapidly deubiquitinated (probably by UCH37 (ref. 11)) and then degraded slowly (that is, at the same rate as ovomucoid alone; Fig. 1b). Because this deubiquitinating activity was observed only with a proteasomal degradation substrate, we named it the 'cryptic DUB'. Figure 2 A ubiquitin-aldehyde (Ubal)-insensitive DUB couples release of ubiquit in with substrate degradation (western blots). Full legend High resolution image and legend (66k) The properties of the cryptic DUB suggested that it could be responsible for pol yubiquitin release from conjugates during degradation. To investigate this possi bility, we prepared protein conjugates containing a pentaubiquitin chain (Ub5) b y conjugation of K48-linked tetraubiquitin (Ub4) to K48 of the ubiquitin in the fusion proteins3. As shown in Fig. 2b (lanes 1–3), Ub5 was released from Ub5(G76)OM incubated with 26S proteasomes in the presence of Ubal. We confirmed that the ovomucoid moiety was degraded by measuring the generation of trichloroacetic acid (TCA)-soluble fluorescence. Similar results were observed with a folded substrate, Ub5(G76)DHFR (Fig. 2b, lanes 4?C6). A related substrate, Ub5(V76)DHFR, is targeted efficiently to the proteasom e (Km = 35 nM), but its degradation is slow (kcat = 0.05 min-1) and is accompani ed by degradation of the proximal mutant ubiquitin Ub(V76)3. These properties ar e reminiscent of our findings with Ub(V76)OM. Indeed, comparison of Ub5(G76)DHFR and Ub5(V76)DHFR, in which DHFR contained a carboxy-terminal HA-tag9, showed th at the former substrate was degraded more rapidly (Fig. 2b, lanes 4–6, lower panel) and with release of Ub5 (lanes 4?C6, upper panel). In contrast, degradation of Ub5(V76)DHFR was slow (lanes 7–9) and Ub4 rather than Ub5 was produced, consistent with degradation of Ub(V76). With both substrates, the DHFR moiety was degraded (that is, free HA-tagged DHFR was not detected). Thermodynamically, ubiquitin is exceedingly stable12, and the need to unfold Ub(V76) before its degradation may contribute to the slow turnover of the Ub(V76)-containing substrates. Deubiquitination of UbOM by the cryptic DUB in 26S proteasomes is ATP-dependent (Fig. 2c, lanes 1, 6, 8 and 9), and ATPS could not substitute for ATP (not shown ). However, Ubal-insensitive deubiquitination by the isolated 19S complex is ATP -independent (lanes 1, 2, 4, 5). Similar results were obtained with the irrevers ible DUB inhibitor ubiquitin vinylsulphone (UbVS)13 (lanes 3, 7) or the Ub5(G76) OM substrate (data not shown). Thus, ATP hydrolysis is not required for deubiqui tination per se; rather, the ATP-dependence of the cryptic DUB must originate fr om other substrate processing step(s). The timing of deubiquitination during sub strate processing is critical: the polyubiquitin chain needs to fulfil its targe ting function, but not impede subsequent substrate translocation. The ATP-depend ence seen with intact proteasomes probably reflects coupling of deubiquitination to degradation, which may insure proper timing of polyubiquitin release by cont rolling the position of the proximal ubiquitin at the DUB active site. To date, ATP hydrolysis has been found to have a role in productive polyubiquitin chain b inding14, substrate unfolding4, and opening of the channels that lead into the 2 0S proteasome15; polypeptide translocation may also require ATP. It is unclear a t present to which step deubiquitination is coupled. Unlike other DUBs, the activity of the cryptic DUB is insensitive to Ubal or UbV S. The fact that similar rates of deubiquitination were observed with the 19S co mplex and the intact 26S proteasome (Fig. 2c) suggests that a subunit of the 19S complex is responsible. Three proteasome-associated DUBs are known. UCH37 is a subunit of the 19S complex that disassembles ubiquitin conjugates11. However, un like the cryptic DUB, its activity is restricted to the distal rather than proxi mal ubiquitin of a polyubiquitin chain, and it is completely inhibited by Ubal. Ubp6/USP14 is a DUB of unknown specificity that can associate with the proteasom e13, 16, but Ubp6 is largely inhibited by Ubal or UbVS (ref. 13, and our own unp ublished data). Moreover, although the bovine 19S and 26S complexes contained ab out 20-fold different levels of Ubp6 (0.005 and 0.12 molecules per complex, resp ectively; data not shown), they displayed similar Ubal-insensitive DUB activity (Fig. 2c). Ubp6 was excluded definitively as the cryptic DUB by comparison of 26 S proteasomes purified from wild-type and ubp6 yeast; both contained the Ubal-in sensitive and ATP-dependent deubiquitination activity (Supplementary Information Fig. S1). Doa4 is a yeast DUB originally thought to participate in substrate pr ocessing by the proteasome because its deficiency leads to accumulation of small peptides linked to mono- or polyubiquitin17. However, the ability of Doa4 to hy drolyse linkages between ubiquitin molecules17 and its substoichiometric associa tion with the 26S proteasome18 make it unlikely that Doa4 is the cryptic DUB. What other protein might account for the cryptic DUB activity? The 19S regulator y complex is composed of two subcomplexes: the base, which includes Rpn1, Rpn2 a nd six ATPases, and the lid, which contains at least eight non-ATPase subunits19 . Among the non-ATPases, Rpn11 (POH1 in humans) is the most conserved subunit, w ith 65% identity between the yeast and the human proteins (Fig. 3). This high co nservation suggested that a catalytic activity might be associated with Rpn11. E ukaryotes contain another multiprotein complex, the COP9 signalosome (CSN), whos e eight subunits are markedly similar to those of the proteasome lid19. The CSN recently was shown to contain a deconjugation activity specific for the ubiquiti n-like protein Nedd8 (ref. 20). Thus, we hypothesized that Rpn11/POH1 is the cry ptic DUB, and that yeast Rri1 (JAB1 in humans), the CSN orthologue, is responsib le for de-conjugation of Nedd8. Proteasome subunits Rpn11/POH1 and Rpn8/S12, CSN subunits Rri1/JAB1, and also eIF3 subunits p40 and p47 each contain an MPN doma in21. We identified nine amino acids within this domain (Fig. 3) as potential ca talytic site residues of a protease; importantly, eight of these are conserved b etween Rpn11 and JAB1, but not between Rpn11 and S12 or eIF3 p40. Figure 3 Alignment of the MPN domain of MPN family proteins. Full legend High resolution image and legend (45k) We predicted that loss of activity of the cryptic DUB would severely impede prot easomal degradation and be lethal to the cell. Therefore, we tested the abilitie s of rpn11 mutants to support growth of yeast (Fig. 4a). Mutant alleles of rpn11 tagged with protein A at the C terminus (rpn11–ProA) were cloned into a low-copy plasmid under the control of the native RPN11 promoter. These plasmids were transformed into yeast that had the chromosomal copy of RPN11 deleted but carried a wild-type RPN11 (tagged with HA at the C terminus) on a low-copy URA3 plasmid. When cured of the URA3 plasmid by transfer to plates containing 5-fluoro-orotic acid (5-FOA), these yeast cells were forced to rely on rpn11?CProA for growth. Figure 4a shows that among nine mutants, five of them support growth as well as the wild type. However, the H111Q, H109Q and D122E mutations a re lethal, and D142A only partially supports growth. Overexpression of the latte r four mutant alleles together with wild-type RPN11 leads to slowed growth and r educed cell density in stationary phase (Supplementary Information Fig. S2). Mor eover, we observed elevated levels of ubiquitin conjugates in these strains (Fig . 4b), an indication of defective proteasomes. To exclude the possibility that t he lethality of the mutations was due to impaired proteasome assembly, we isolat ed 26S proteasomes from yeast strains that harboured low-copy plasmids to expres s wild-type Rpn11-HA together with each of the mutant Rpn11–ProA proteins. Immunoblotting with anti-protein A and anti-HA antibodies revealed that both wild-type and mutant Rpn11 proteins were present (Fig. 4c). Note that wild-type Rpn11?Cprotein A was incorporated at a higher level than the mutated forms. This corre lated with higher levels of the wild-type fusion protein in the cell lysates (da ta not shown). It is possible that the uncomplexed mutant Rpn11 subunits were in trinsically unstable, or that dysfunctional proteasomes containing inactive Rpn1 1 were degraded selectively. Figure 4 Evidence that Rpn11/POH1 is a Zn2+-dependent DUB and characterization of rpn11 mutants. Full legend High resolution image and legend (48k) The above results suggest that Rpn11 is the cryptic DUB and that H109, H111 and D122 are active-site residues; D142 may also assist catalysis. Importantly, the viability of the C116S mutant excludes a cysteine protease mechanism; this is de spite similarity of residues 113–121 in Rpn11 with the catalytic Cys-box of the ubiquitin-specific protease (UBP) family of DUBs22. Rather, the identities of the critical residues suggest that Rpn11 is a Zn2+-dependent metalloprotease23. When bovine 19S complex was incubated with the Zn2+ chelator TPEN (N,N,N',N'-tetrakis(2-pyridylmethyl)ethylenediamine), deubiquitination of Ub(G76)OM was inhibited; this effect was prevented when TPEN was pre-incubated with excess Zn2+ (Fig. 4d). TPEN did not inhibit the cryptic DUB in intact bovine 26S proteasomes (data not shown), possibly because the 19S and 26S complexes differ with respect to stability of the lid subcomplex or access to the Rpn11 active site. An ATP-dependent, Ubal-insensitive DUB activity in 26S proteasomes that is inhibited by o-phenanthroline has been reported previously24. We think this activity probably corresponds to the cryptic DUB identified here. However, we found that inhibition by o-phenanthroline, which is not prevented by excess Zn2+, is apparently due to a general structural disruption of both the 19S and 26S complexes (data not shown). Thus far, our attempts to generate an active deubiquitinating enzyme from Rpn11 expressed in Escherichia coli or in vitro have been unsuccessful. Rpn11 may be u nable to fold as an isolated subunit into a native, active conformation. However , substrate interactions with neighbouring subunits in the 19S complex are proba bly required to position conjugates for cleavage by Rpn11. Normally, these inter actions would involve a portion of the substrate protein that is unfolded and ne ar the (poly)ubiquitin–protein attachment site. In the case of the Ub5(V76)DHFR substrate, the proximal ubiquitin was itself unfolded and translocated until the next ubiquitin, which possessed a G76 residue, was in position for cleavage and Ub4 was released. Except for residues 113–121 (see above), Rpn11/POH1 bears little resemblance to either ubiquitin C-terminal hydrolase (UCH)- or UBP-type DUB enzymes25. Rpn11/POH1 is also unique among known DUBs in that it is essential for growth, and its discovery helps to place deubiquitination both functionally and mechanistically in the scheme of proteasomal degradation. MPN domain proteins are represented in all three biological domains of life (Archaea, Bacteria and Eukaryota)21, 26. The properties of Rpn11/POH1 provide insight into the possible functions and mechanisms of those MPN proteins that contain the putative Zn2+-binding site. In particular, our results suggest that JAB1 in the COP9 signalosome is responsible for de-conjugation of Nedd8. Another Rpn11 homologue of interest is C6.1A, the aberrant expression of which has been implicated in cases of T-cell leukaemia27. Methods Strains and plasmids The UbOM plasmid was constructed by putting a six-His-tagge d ubiquitin28 between NdeI and HindIII sites of vector pRSet, then inserting HAO M (made by polymerase chain reaction (PCR) from pOMCHI1-X (ref. 8) and a primer encoding an HA epitope) at the HindIII site. The Ub(G76)DHFR plasmid was derived from Ub(V76)DHFR3 by site-directed mutagenesis. Wild-type RPN11 was cloned into pRS316 with 200 base pairs of upstream sequence and the addition of a C-termina l HA tag. RPN11–ProA was cloned from the yeast sDL73 (ref. 16) and put into pRS313 with a 200-bp upstream sequence of RPN11, or into pESC-LEU (Stratagene) without the upstream sequence. Mutations were introduced into those plasmids by site-directed mutagenesis. A RPN11/rpn11 diploid strain (Research genetics, strain number 25683) was transformed with pRS316 RPN11-HA, sporulated, and tetrads dissected to obtain haploid strain RCY519 that was rpn11[pRS316-RPN11-HA]. Proteasome isolation 26S proteasomes were purified from bovine erythrocytes29, a nd the bovine 19S complex (PA700) was a gift from G. DeMartino. Yeast 26S protea somes were affinity-purified using sDL73 as the wild type16 or a ubp6 derivative of sDL73 made by standard techniques. For rpn11 mutant proteasomes, strain RCY5 19 was transformed with various pRS313 rpn11–ProA mutants and grown to stationary phase in SC (- URA, -LEU) medium. Lysates were prepared in 50 mM HEPES, pH 8, containing 5 mM MgCl2, 2 mM ATP, 1 mM EDTA, 0.3 M sorbitol, 0.1 M NaCl, 40 μM chymostatin, and 1 protease inhibitor cocktail (Sigma, number P8215). After 1 h centrifugation at 100,000g, the supernatant was further centrifuged at 400,000g for 1 h and the clear pellets were dissolved in degradation assay buffer (see below); after separation by non-denaturing polyacrylamide gel electrophoresis (PAGE), proteasomes were detected by soaking the gel in Suc-LLVY-AMC19. Gel pieces containing the 26S species were excised, heated with SDS sample buffer, and analysed by SDS?CPAGE and immunoblotting. Substrate proteins Recombinant UbOM (G76 or V76) and UbDHFR (G76 or V76) were ex pressed in E. coli and purified with Ni2+-NTA agarose (Qiagen) followed by gel f iltration (Superdex 75). Ovomucoid and UbOM were unfolded by 6 M guanidine-HCl a nd 1 mM dithiothreitol (DTT) in 50 mM Tris-HCl pH 7.4, 0.5 mM EDTA, then exchang ed into buffer containing 1 mM tris(2-carboxyethyl)phosphine-HCl instead of DTT. Lucifer yellow iodoacetamide (molecular Probes) was added in ninefold excess ov er cysteines. After 5 h in the dark at 20 °C, free dye was removed by desalting (Sephadex G25); 80% of the cysteines were modified. Ub4 chains prepared from Ub C48 and UbD77 (ref. 30) were transferred to monoubiquitin substrates3 and the Ub 5 protein products were purified on Ni2+-NTA agarose. Degradation assays and immunoblots Conditions for in vitro degradation or deubiq uitination were essentially as described3. Unless specified otherwise, reactions were at 37 °C and included 42 nM 26S proteasomes or 19S complex and 2 μM Ubal. Proteasomes were pre-incubated with Ubal for 5 min before addition of sub strates; 10 U ml-1 apyrase (Sigma) was used for ATP depletion. To quantify degra dation of lucifer yellow-labelled substrates, samples were precipitated with 20% TCA, centrifuged, and fluorescence in the supernatant was measured (426 nm exci tation, 531 nm emission) after neutralization with Tris base. Immunoblotting fol lowed standard protocols and detection was by means of chemiluminescence (Pierce ). The antibodies used were: rabbit polyclonal affinity-purified anti-ubiquitin, anti-protein A (Sigma), and monoclonal anti-HA (clone 16B12; Covance). Supplementary information accompanies this paper. Received 19 June 2002;accepted 16 August 2002 References 1. Hershko, A. & Ciechanover, A. The ubiquitin system. Annu. Rev. Bio chem. 67, 425-479 (1998) | Article | PubMed | 2. Zwickl, P., Baumeister, W. & Steven, A. Dis-assembly lines: the proteasome an d related ATPase-assisted proteases. Curr. Opin. Struct. Biol. 10, 242-250 (2000 ) | Article | PubMed | 3. Thrower, J. S., Hoffman, L., Rechsteiner, M. & Pickart, C. M. Recognition of the polyubiquitin proteolytic signal. EMBO J. 19, 94-102 (2000) | Article | PubM ed | 4. Braun, B. C. et al. The base of the proteasome regulatory particle exhibits c haperone-like activity. Nature Cell Biol. 1, 221-226 (1999) | Article | PubMed | 5. Liu, C. et al. Conformational remodeling of proteasomal substrates by PA700, the 19S regulatory complex of the 26S proteasome. J. Biol. Chem. 277, 26815-2682 0 (2002) | Article | PubMed | 6. D'Andrea, A. & Pellman, D. Deubiquitinating enzymes: a new class of biologica l regulators. Crit. Rev. Biochem. Mol. Biol. 33, 337-352 (1998) | PubMed | 7. Amerik, A. Y., Li, S. J. & Hochstrasser, M. Analysis of the deubiquitinating enzymes of the yeast Saccharomyces cerevisiae. Biol. Chem. 381, 981-992 (2000) | PubMed | 8. DeKoster, G. T. & Robertson, A. D. thermodynamics of unfolding for Kazal-type serine protease inhibitors: entropic stabilization of ovomucoid first domain by glycosylation. Biochemistry 36, 2323-2331 (1997) | Article | PubMed | 9. Johnson, E. S., Ma, P. C., Ota, I. M. & Varshavsky, A. A proteolytic pathway that recognizes ubiquitin as a degradation signal. J. Biol. Chem. 270, 17442-174 56 (1995) | Article | PubMed | 10. Hershko, A. & Rose, I. A. Ubiquitin-aldehyde: a general inhibitor of ubiquit in-recycling processes. Proc. Natl Acad. Sci. USA 84, 1829-1833 (1987) | PubMed | 11. Lam, Y. A., Xu, W., DeMartino, G. N. & Cohen, R. E. Editing of ubiquitin con jugates by an isopeptidase in the 26S proteasome. Nature 385, 737-740 (1997) | P ubMed | 12. Wintrode, P. L., Makhatadze, G. I. & Privalov, P. L. Thermodynamics of ubiqu itin unfolding. Proteins 18, 246-253 (1994) | PubMed | 13. Borodovsky, A. et al. A novel active site-directed probe specific for deubiq uitylating enzymes reveals proteasome association of USP14. EMBO J. 20, 5187-519 6 (2001) | Article | PubMed | 14. Lam, Y. A., Lawson, T. G., Velayutham, M., Zweier, J. L. & Pickart, C. M. A proteasomal ATPase subunit recognizes the polyubiquitin degradation signal. Natu re 416, 763-767 (2002) | Article | PubMed | 15. Rubin, D. M., Glickman, M. H., Larsen, C. N., Dhruvakumar, S. & Finley, D. A ctive site mutants in the six regulatory particle ATPases reveal multiple roles for ATP in the proteasome. EMBO J. 17, 4909-4919 (1998) | Article | PubMed | 16. Leggett, D. S. et al. Multiple associated proteins regulate proteasome struc ture and function. Mol. Cell (in the press) 17. Papa, F. R. & Hochstrasser, M. The yeast DOA4 gene encodes a deubiquitinatin g enzyme related to a product of the human tre-2 oncogene. Nature 366, 313-319 ( 1993) | PubMed | 18. Papa, F. R., Amerik, A. Y. & Hochstrasser, M. Interaction of the Doa4 deubiq uitinating enzyme with the yeast 26S proteasome. Mol. Biol. Cell 10, 741-756 (19 99) | PubMed | 19. Glickman, M. H. et al. A subcomplex of the proteasome regulatory particle re quired for ubiquitin-conjugate degradation and related to the COP9-signalosome a nd eIF3. Cell 94, 615-623 (1998) | PubMed | 20. Lyapina, S. et al. Promotion of NEDD-CUL1 conjugate cleavage by COP9 signalo some. Science 292, 1382-1385 (2001) | Article | PubMed | 21. Hofmann, K. & Bucher, P. The PCI domain: a common theme in three multiprotei n complexes. Trends Biochem. Sci. 23, 204-205 (1998) | Article | PubMed | 22. Glickman, M. H., Rubin, D. M., Fried, V. A. & Finley, D. The regulatory part icle of the Saccharomyces cerevisiae proteasome. Mol. Cell Biol. 18, 3149-3162 ( 1998) | PubMed | 23. Alberts, I. L., Nadassy, K. & Wodak, S. J. Analysis of zinc binding sites in protein crystal structures. Protein Sci. 7, 1700-1716 (1998) | PubMed | 24. Eytan, E., Armon, T., Heller, H., Beck, S. & Hershko, A. Ubiquitin C-termina l hydrolase activity associated with the 26 S protease complex. J. Biol. Chem. 2 68, 4668-4674 (1993) | PubMed | 25. Wilkinson, K. D. Regulation of ubiquitin-dependent processes by deubiquitina ting enzymes. FASEB J. 11, 1245-1256 (1997) | PubMed | 26. Ponting, C. P., Aravind, L., Schultz, J., Bork, P. & Koonin, E. V. Eukaryoti c signalling domain homologues in archaea and bacteria. Ancient ancestry and hor izontal gene transfer. J. Mol. Biol. 289, 729-745 (1999) | Article | PubMed | 27. Thick, J., Mak, Y. F., Metcalfe, J., Beatty, D. & Taylor, A. M. A gene on ch romosome Xq28 associated with T-cell prolymphocytic leukemia in two patients wit h ataxia telangiectasia. Leukemia 8, 564-573 (1994) | PubMed | 28. Beers, E. P. & Callis, J. Utility of polyhistidine-tagged ubiquitin in the p urification of ubiquitin-protein conjugates and as an affinity ligand for the pu rification of ubiquitin-specific hydrolases. J. Biol. Chem. 268, 21645-21649 (19 93) | PubMed | 29. Hoffman, L., Pratt, G. & Rechsteiner, M. Multiple forms of the 20S multicata lytic and the 26S ubiquitin/ATP- dependent proteases from rabbit reticulocyte ly sate. J. Biol. Chem. 267, 22362-22368 (1992) | PubMed | 30. Piotrowski, J. et al. Inhibition of the 26S proteasome by polyubiquitin chai ns synthesized to have defined lengths. J. Biol. Chem. 272, 23712-23721 (1997) | Article | PubMed | Acknowledgements. We thank D. Finley and D. Leggett for yeast strains, advice on proteasome purification, and a sample of yeast proteasomes; G. DeMartino for bo vine PA700; A. Robertson for ovomucoid protein; C. Pickart for the Ub(V76)DHFR p lasmid; K. Wilkinson, H. Ploegh and A. Borodovsky for UbVS, recombinant Ubp6 pro tein and anti-human Ubp6; C. Pickart, A. Lam and K. Wilkinson for discussions; a nd L. Weisman and C. Pickart for comments on the manuscript. This work was