Current Biology, Vol. 14, R754–R756, September 21, 2004, ©2004 Elsevier Ltd. All rights reserved. DOI 10.1016/j.cub.2004.09.012
Protein Degradation: Recognition of Ubiquitinylated Substrates Rasmus Hartmann-Petersen1 and Colin Gordon2
A cell-free system has been developed in budding yeast that provides direct evidence that the Dsk2/Dph1, Rad23/Rhp23 and Rpn10/Pus1 multiubiquitin-binding proteins, long implicated in substrate recognition and presentation to the 26S proteasome, actually fulfil such a role.
Ubiquitin, a 76 amino acid protein highly conserved in all eukaryotes, is covalently linked to lysine residues in target proteins and regulates many intracellular processes such as endocytosis, vesicle sorting and virus budding [1]. Ubiquitin’s main function is to act as a tag, targeting proteins for degradation by the proteasome, the multiprotein protease responsible for protein turnover. Ubiquitin chains of at least four moieties in length are required to target proteins for destruction. Ubiquitin itself has seven lysine residues, all of which could potentially nucleate ubiquitin chains. Biochemical analysis has shown, however, that only Lys27- and Lys48-linked chains are able to direct proteins for degradation [2]. A key question is how are multi-ubiquitinylated proteins recognized by the proteasome? Recent studies [3,4] using a cell-free yeast system have led to a clearer understanding of substrate recognition and presentation to the proteasome. Initial biochemical studies showed that one subunit of the proteasome — S5a, Rpn10, Pus1 in humans, budding yeast and fission yeast, respectively — had the properties expected for a multi-ubiquitin receptor. Rpn10 was able to bind in vitro to chains of four moieties or more but had little affinity for mono-ubiquitin. Deletion analysis localized the binding to a carboxyterminal ubiquitin-interacting motif (UIM) [5]. Genetic experiments in yeast, however, showed that deletion of the gene resulted in little loss of viability [6,7]. Most other proteasome genes are essential so this surprising result raised the possibility that multi-ubiquitin recognition was degenerate and other proteins were able to compensate for the multi-ubiquitin recognition by the Rpn10/Pus1 protein. A number of laboratories set out to identify such factors. It was found that two proteins, Dsk2/Dph1 and Rad23/Rhp23, had the biochemical and genetic properties required for such a role — they were both able to interact with ubiquitin chains via their ubiquitin-associated (UBA) domains [8] and the proteasome via their ubiquitin-like (UBL) domains. Single mutants of DSK2/Dph1+ or RAD23/Rhp23+ were viable. When 1August
Krogh Institute, University of Copenhagen, Universitetsparken 13, DK-2100 Copenhagen, Denmark. 2MRC Human Genetics Unit, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK. E-mail:
[email protected]
Dispatch
the mutants were crossed with each other and the rpn10∆/pus1∆ strain to obtain double and triple mutants, however, these strains showed synthetic phenotypes, implying that they each had overlapping functions with the RPN10/pus1+ gene [9–15]. Further evidence that these proteins played a role in substrate recognition was revealed by the finding that in fission yeast the cyclin-dependent kinase (Cdk) inhibitor Rum1 (a known target of proteasome-mediated ∆rhp23∆ mutant degradation), accumulated in the pus1∆ strain [10]. This work led to a model in which these proteins act as substrate-binding proteins that transport ubiquitinylated substrates from the E3 ubiquitin ligases (which add the ubiquitin chains to proteasome substrates) on to the proteasome [16] (Figure 1). After these initial studies on the function of the shuttle proteins, it was realized that a cell-free system would be required to demonstrate that these multiubiquitin-binding proteins played a direct role in recognition and turnover of ubiquitinylated substrates. The first such cell-free system was developed by Raasi and Pickart [17]. Surprisingly, they showed that the Rad23 protein did not appear to stimulate degradation by the proteasome, as predicted by the shuttle hypothesis, but actually inhibited degradation. In the new studies, Deshaies and co-workers [3] have now developed a different cell-free system that actually behaves in a manner predicted by the shuttle hypothesis. In this system, 26S proteasomes are affinity-purified from budding yeast using an epitopetagged version of the Pre1 20S subunit from wild-type, rpn10∆, dsk2∆ or rad23∆ strains. As a substrate they use a ubiquitinylated form of the Cdk inhibitor Sic1. Wild-type proteasomes degraded ubiquitinylated Sic1 efficiently while those prepared from the rpn10∆ and rad23∆ strains were defective in Sic1 turnover. By adding epoxomicin, an inhibitor of proteolysis by the 20S proteasome core particle, the authors could show that the rpn10∆ and rad23∆ 26S proteasomes were also defective in deubiquitination of the Sic1 substrate. Importantly, if recombinant Rpn10 protein was added to the rpn10∆ proteasomes, efficient degradation and deubiquitination of the Sic1 substrate was restored. Interestingly, this rescue was dosage dependent. The addition of low levels of Rpn10 protein was sufficient to rescue activity and had little effect on the activity of wild-type proteasomes. However, a twofold molar excess of Rpn10 protein was enough to observe an inhibition of Sic1 degradation by the wild-type proteasomes, and a fourfold molar excess inhibited all activity. Mutational analysis showed that the UIM domain, required for multi-ubiquitin binding in Rpn10, played a critical role in ubiquitinylated Sic1 degradation. This important result demonstrates for the first time that the UIM domain of Rpn10 plays a direct role in protein degradation. The authors went on to show that recombinant Rad23 could rescue the deubiquitinating activity of rad23∆ proteasomes. This rescue
Current Biology R755
Sic1, Far1
?
Sic1, Clb2, Gic2
Rpn10
Rad23
was also concentration dependent. At low concentrations Rad23 protein restored deubiquitinating activity, while it inhibited at higher concentrations. A threefold molar excess of Rad23 protein was used in the earlier experiments of Raasi and Pickart [17], so their observed inhibition of degradation is consistent with the findings of Verma et al. [3]. The genetic studies in yeast predicted that Rpn10/Pus1 and Rad23/Rhp23 had overlapping functions. It is therefore gratifying that in the in vitro system Rpn10 protein rescued the deubiquitinating activity when added to Rad23-deficient proteasomes, meaning that Rpn10 and Rad23 proteins acted redundantly. Surprisingly, although Rad23 protein could rescue the deubiquitinating activity of the Rad23-deficient proteasomes, it was not able to rescue the deubiquitinating of the Rpn10-deficient proteasomes. Rescue of the Rpn10-deficient proteasomes by Rad23 protein required the presence of at least the aminoterminal von Willebrand factor type A domain of Rpn10. A ‘facilitator’ function was thus proposed for Rpn10 to explain its role in stimulating rescue of Rpn10-deficient proteasomes by Rad23. Presumably this facilitator function is mediated by Rpn10 within the 26S proteasome complex and not by free Rpn10 acting as a shuttle protein. Additional experiments demonstrated that rescue of the deubiquitinating activity of Rad23- and Rpn10-deficient proteasomes required both the proteasomeinteracting UBL domain and the multi-ubiquitin-binding UBA domain of Rad23. Again this is consistent with studies in yeast which showed that both domains have to be present to rescue the defects in rad23∆rpn10∆ and rhp23∆pus1∆ yeast strains [10,13]. In addition these observations reveal another inconsistency with the earlier in vitro system in which the UBL domain of Rad23 did not appear to be required for activity [17]. Finally, further evidence that the budding yeast Rad23 and Rpn10 proteins acted as alternative multi-ubiquitin receptors for the 26S proteasome was provided by a recent study [4], which demonstrated that the proteasome could bind ubiquitinylated substrates directly via Rad23 and/or Rpn10. However, substrate turnover was not investigated in this system. As all the studies up to this point had been on the contribution of Rad23 and Rpn10 to the degradation of Sic1, it was important to assess if similar findings were also observed for other known ubiquitin–proteasome substrates. To test this, a number of different epitope-tagged physiological substrates of the ubiquitin–proteasome system were expressed in different mutant backgrounds. Interestingly, there appeared to be some specificity in that the substrate protein Cln2, a G1 cyclin, was not stabilized in the rpn10∆, dsk2∆, ddi1∆, ufd1-1, rad23∆, rad23∆dsk2∆, rpn10∆rad23∆ or rad23∆ddi1∆ mutant strains. In contrast, Far1, a G1 Cdk inhibitor, showed higher levels of accumulation in a rad23∆ than in a rpn10∆ strain, while the Cdc42 effector Gic2 and the mitotic cyclin Clb2 both behaved in a manner similar to the Sic1 protein. In addition, the endoplasmic reticulum associated degradation (ERAD) substrate CPY* was not stabilized in the rpn10 ∆ and rad23 ∆ strains but only in a
E3s Rpn1 Rpt5
20S core particle
Current Biology
Figure 1. Shuttle model for recognition and presentation of substrates to the 26S proteasome. Substrate proteins (green thread) which have become ubiquitinylated (black beads) by ubiquitin-protein ligases (E3s) are transferred to the proteasome by shuttle proteins such as Rad23 and Rpn10. These proteins contain UBA/UIM domains (blue) that bind ubiquitin and UBL/VWA domains (red) that mediate their interaction with the 26S proteasome subunit Rpn1 and, in the case of Rpn10, also Rpn2 (not shown). Some substrates, like Far1, Clb2 and Gic2 are specific for either Rad23 or Rpn10, as indicated, whereas others, like Sic1, are unspecific. As the ubiquitin-binding activity of the Rpt5 subunit is insufficient to target substrate degradation in vitro, it possibly serves downstream of Rad23 and Rpn10 to conduct ubiquitinylated substrates to the 20S core once they have been brought to the 26S proteasome. For simplicity other less well-characterized shuttle proteins like Dsk2 and Cdc48 have been excluded. Also the lid sub-complex of the 26S proteasome is not shown.
ufd1-1 strain. Ufd1, in complex with Npl4 and the ATPase Cdc48 had previously been implicated in the ERAD pathway and these results implied that neither Rad23 nor Rpn10 played any part in presentation of substrates from this pathway [18]. It was surprising therefore that Rad23 was identified to be a critical component of ERAD in a genome-wide screen carried out in S. cerevisiae [19]. Further work will be required to resolve this apparent paradox. From the results described above, however, the different multi-ubiquitin-binding proteins seem to have some specificity for different ubiquitin–proteasome substrates. One simple model that might be proposed to account for this specificity is that the shuttle proteins interact specifically with the different ubiquitin ligases that add the ubiquitin chain to the substrate proteins. This does not appear to be the case, however, as both Cln2 and Gic2 protein substrates are targeted by the SCFGrr1 ubiquitin ligase but are turned over differently in the different multi-ubiquitin receptor mutant strains. In conclusion, the development of a cell-free extract that behaves in a manner predicted from the previous genetic data represents a significant step forward in the study of recognition and presentation of ubiquitinylated substrates to the 26S proteasome. The combination of this cell-free system with microbial genetics should provide powerful tools with which to investigate this important topic in more detail.
Dispatch R756
References 1. Hicke, L. (2001). Protein regulation by monoubiquitin. Nat. Rev. Mol. Cell Biol. 2, 195-201. 2. Glickman, M.H., and Ciechanover, A. (2002). The ubiquitin-proteasome proteolytic pathway: destruction for the sake of construction. Physiol. Rev. 82, 373-428. 3. Verma, R., Oania, R., Graumann, J., and Deshaies, R.J. (2004). Multiubiquitin chain receptors define a layer of substrate selectivity in the ubiquitin-proteasome system. Cell 118, 99-110. 4. Elsasser, S., Chandler-Militello, D., Muller, B., Hanna, J., and Finley, D. (2004). Rad23 and Rpn10 serve as alternative ubiquitin receptors for the proteasome. J. Biol. Chem. 279, 26817-26822. 5. Deveraux, Q., Ustrell, V., Pickart, C., and Rechsteiner, M. (1994). A 26 S protease subunit that binds ubiquitin conjugates. J. Biol. Chem. 269, 7059-7061. 6. van Nocker, S., Sadis, S., Rubin, D.M., Glickman, M., Fu, H., Coux, O., Wefes, I., Finley, D., and Vierstra, R.D. (1996). The multiubiquitin-chain-binding protein Mcb1 is a component of the 26S proteasome in Saccharomyces cerevisiae and plays a nonessential, substrate-specific role in protein turnover. Mol. Cell. Biol. 16, 60206028. 7. Wilkinson, C.R., Ferrell, K., Penney, M., Wallace, M., Dubiel, W., and Gordon, C. (2000). Analysis of a gene encoding Rpn10 of the fission yeast proteasome reveals that the polyubiquitin-binding site of this subunit is essential when Rpn12/Mts3 activity is compromised. J. Biol. Chem. 275, 15182-15192. 8. Hofmann, K., and Bucher, P. (1996). The UBA domain: a sequence motif present in multiple enzyme classes of the ubiquitination pathway. Trends Biochem. Sci. 21, 172-173. 9. Schauber, C., Chen, L., Tongaonkar, P., Vega, I., Lambertson, D., Potts, W., and Madura, K. (1998). Rad23 links DNA repair to the ubiquitin/proteasome pathway. Nature 391, 715-718. 10. Wilkinson, C.R., Seeger, M., Hartmann-Petersen, R., Stone, M., Wallace, M., Semple, C., and Gordon, C. (2001). Proteins containing the UBA domain are able to bind to multi-ubiquitin chains. Nat. Cell Biol. 3, 939-943. 11. Lambertson, D., Chen, L., and Madura, K. (1999). Pleiotropic defects caused by loss of the proteasome-interacting factors Rad23 and Rpn10 of Saccharomyces cerevisiae. Genetics 153, 6979. 12. Funakoshi, M., Sasaki, T., Nishimoto, T., and Kobayashi, H. (2002). Budding yeast Dsk2p is a polyubiquitin-binding protein that can interact with the proteasome. Proc. Natl. Acad. Sci. USA 99, 745750. 13. Rao, H., and Sastry, A. (2002). Recognition of specific ubiquitin conjugates is important for the proteolytic functions of the UBA domain proteins Dsk2 and Rad23. J. Biol. Chem. 277, 11691-11695. 14. Saeki, Y., Saitoh, A., Toh-e, A., and Yokosawa, H. (2002). Ubiquitinlike proteins and Rpn10 play cooperative roles in ubiquitin- dependent proteolysis. Biochem. Biophys. Res. Commun. 293, 986-992. 15. Chen, L., and Madura, K. (2002). Rad23 promotes the targeting of proteolytic substrates to the proteasome. Mol. Cell. Biol. 22, 49024913. 16. Hartmann-Petersen, R., Seeger, M., and Gordon, C. (2003). Transferring substrates to the 26S proteasome. Trends Biochem. Sci. 28, 26-31. 17. Raasi, S., and Pickart, C.M. (2003). Rad23 ubiquitin-associated domains (UBA) inhibit 26 S proteasome-catalyzed proteolysis by sequestering lysine 48-linked polyubiquitin chains. J. Biol. Chem. 278, 8951-8959. 18. Ye, Y., Meyer, H.H., and Rapoport, T.A. (2003). Function of the p97Ufd1-Npl4 complex in retrotranslocation from the ER to the cytosol: dual recognition of nonubiquitinated polypeptide segments and polyubiquitin chains. J. Cell Biol. 162, 71-84. 19. Medicherla, B., Kostova, Z., Schaefer, A., and Wolf, D.H. (2004). A genomic screen identifies Dsk2p and Rad23p as essential components of ER-associated degradation. EMBO Rep. 5, 692-697.
