Interaction of the anaphase-promoting complex/cyclosome and proteasome protein complexes with multiubiquitin chain-binding proteins.

Fission yeast Rhp23 and Pus1 represent two families of multiubiquitin chain-binding proteins that associate with the proteasome. We show that both proteins bind to different regions of the proteasome subunit Mts4. The binding site for Pus1 was mapped to a cluster of repetitive sequences also found in the proteasome subunit SpRpn2 and the anaphase-promoting complex/cyclosome (APC/C) subunit Cut4. The putative role of Pus1 as a factor involved in allocation of ubiquitinylated substrates for the proteasome is discussed.


From the Medical Research Council Human Genetics Unit, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, United Kingdom
Fission yeast Rhp23 and Pus1 represent two families of multiubiquitin chain-binding proteins that associate with the proteasome. We show that both proteins bind to different regions of the proteasome subunit Mts4. The binding site for Pus1 was mapped to a cluster of repetitive sequences also found in the proteasome subunit SpRpn2 and the anaphase-promoting complex/cyclosome (APC/C) subunit Cut4. The putative role of Pus1 as a factor involved in allocation of ubiquitinylated substrates for the proteasome is discussed.
Ubiquitin-dependent protein degradation is a mechanism employed in eukaryotic cells not only to recycle damaged and misfolded proteins but also to control cellular processes by specific breakdown of regulatory proteins (1). Ubiquitinylation is accomplished in multiple steps (2). Initially ubiquitin is activated by a ubiquitin-activating enzyme (E1) in an ATP-dependent process whereby the E1 forms a thiol ester bond with the C terminus of ubiquitin. Subsequently the ubiquitin molecule is transferred to a ubiquitin-conjugating enzyme (E2). E2s carrying ubiquitin associate with substrate binding ubiquitin protein ligases (E3s) resulting in the covalent attachment of ubiquitin to the substrate protein. Several rounds of this conjugation process produce substrates carrying a chain of ubiquitin moieties. There is a range of different E2s that associate with various E3s, incorporating an element of substrate specificity to this process.
Ubiquitin protein ligases can be divided in two major groups, HECT domain E3s and RING finger domain E3s. The anaphase-promoting complex/cyclosome (APC/C) 1 belongs to a family of multimeric ubiquitin protein ligases of the RING finger type that also include the SCF and the VCB (2,3). It has a molecular mass of 700 kDa and consists of 11 different subunits. Mitotic events are controlled by the APC/C via multiubiquitinylation of cell cycle regulators like Cut2/Pds1p/securin and B-type cyclins. Once a ubiquitin chain is conjugated to these proteins they become targets for degradation by the 26 S proteasome enabling the continuation of downstream mitotic events.
The 2.5 MDa 26 S proteasome catalyzes the degradation of cellular proteins in an ATP-dependent manner (4). Its proteolytic component is the 20 S core complex, a cylindrical structure comprising four stacked rings each containing seven proteins. The inner rings enclose a central chamber harboring the catalytic sites. Access to the lumen is provided via the outer rings and regulated by the 19 S regulatory complex that is attached to one or both ends of the 20 S core (5).
The 19 S regulatory complex can be dissociated in two subcomplexes called the base and the lid (6,7). Six ATPases subunits are presumed to form a structure that associates with the 20 S core. Together with the two largest subunits of the proteasome SpRpn2/Rpn2/S1 and Mts4/Rpn1/S2 these ATPases form the base complex that was suggested to be participating in the unfolding and translocation of substrates. The lid complex is believed to contact the base via the subunits SpRpn2/Rpn2/S1, Mts4/Rpn1/S2, and Pus1/Rpn10/S5a. It is composed of a number of non-ATPase subunits whose function remains rather enigmatic. Proteasomes lacking the lid can cleave small peptides in an ATP-dependent manner, but they are unable to degrade ubiquitinylated proteins (6). Thus the lid is believed to be involved in the recognition and processing of those substrates.
To date, one subunit of the proteasome has been found to possess the ability to recognize multiubiquitinylated substrates, namely Pus1/Rpn10/S5a (8). However, given the fact that in yeast the corresponding RPN10/pus1 ϩ genes are not essential for cell viability, other mechanisms must clearly exist for this crucial step in the degradation process (9). Recently, it has been shown that the UBA/UBL domain proteins Rhp23/ Rad23p and Dph1/Dsk2p can interact with both the proteasome and with multiubiquitin suggesting that these factors play a role in the recognition and delivery of substrates for proteolysis (10 -13). The Pus1/Rpn10p protein contains a stretch of 20 amino acids recently defined as ubiquitin-interacting motif (UIM) that is involved in multiubiquitin chain binding, whereas in Dph1/Dsk2p and Rhp23/Rad23p the task is accomplished by a C-terminal ubiquitin pathway associated (UBA) domain (9,10,14,15,16). Both Dph1/Dsk2p and Rhp23/ Rad23p bind to the proteasome with their N-terminal UBL domain (10,17). As Pus1/Rpn10p does not have a UBL domain it must use different structures to associate with other proteasome subunits. Lambertson et al. (18) were able to precipitate the proteasome from cells deleted for RPN10 and RAD23 using tagged Rad23p or Rpn10p, respectively, demonstrating that these proteins bind to the proteasome independently.
So far little is known about how ubiquitinylated substrates arrive at the proteasome. Some ubiquitinylating enzymes associate directly with the proteasome as shown for the E2s Ubc1, Ubc2, Ubc4, and Ubc5 as well as for the E3s Ubr1p, Ufd4p, and KIAA10 (19 -23). It is also conceivable that multiubiquitinbinding proteins provide a link between components of the ubiquitinylation machinery and the proteasome as implied by the finding that hPLIC proteins, the human homologues of Dph1/Dsk2p, associate with E3 proteins (24). However, previously neither Rhp23/Rad23p nor Pus1/Rpn10p have been reported to interact with ubiquitinylating enzymes.
To improve our understanding of how the proteasome recruits ubiquitin substrates we started to investigate the association of these proteins with the 19 S regulatory complex. We characterize regions in proteasome subunits that bind Pus1 and Rhp23 and describe the interaction of these multiubiquitin-binding proteins with the APC/C.

EXPERIMENTAL PROCEDURES
Bioinformatics-Sequence similarity searches were carried out using Psi-BLAST (version 2.2.3) (25) using the BLOSUM62 substitution matrix for the first iteration. Sequences identified with a BLAST E-value of Ͻ1e Ϫ5 at convergence were filtered to remove overly similar sequences (80% identity in alignment) and aligned using ClustalW (1.74). The HMMer 2.0 package (version 2.2g;) was used to generate and optimize Hidden Markov models (HMMs) from the multiple sequence alignment. HMMsearch was used with the HMM to identify further examples of the UBL binding region in the SPTR data base.
Standard genetic methods and media were used as described, and S. pombe transformations were performed by the lithium acetate procedure (26).
Yeast Two-hybrid Assay-The full-length rhp23 ϩ gene and the UBL encoding domain of rhp23 ϩ were subcloned into the MCS of the pAS2 vector (Clontech). The yeast strain Y190 was transformed using this construct as bait. The obtained transformants were then transformed again, this time with a series of constructs encoding proteasomal subunits cloned into the pACT2 vector (Clontech) as prey. The obtained double transformants were tested for growth on plates containing 25 mM 3-aminotriazole (Sigma) and for their ability to activate the lacZ reporter by filter lifts assays. Only the Mts4 pACT2 prey construct scored positive in both of these tests.
In Vitro Binding Assays-GST and GST fusion proteins were expressed in Escherichia coli BL21 (DE3) pLysS, bound to glutathione-Sepharose 4B beads (Amersham Biosciences) as described by the manufacturer and protein/beads ratio was adjusted to about 1 mg/ml. For assays using His-tagged constructs, proteins were expressed in E. coli M15 cells (Qiagen) and lysed in buffer A (50 mM Tris, pH 7.5, 100 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and Complete TM protease inhibitors (Roche Applied Science). 0.5 ml of the cleared extract were mixed with 15 l of beads bound to GST or GST fusion protein and incubated for 2 h at 4°C. Protein concentration of the extracts varied between 10 and 50 mg/ml depending on the expression level of the His-tagged protein. The beads were washed four times in buffer A and resuspended in 50 l of SDS sample buffer. 10 l of the samples were separated on 10 or 12% SDS gels and subjected to Western blot analysis. For binding assays using total yeast extracts, cells were grown to an OD 595 0.8, lysed in buffer B (1 mM dithiothreitol, 2 mM ATP and 10 mM MgCl 2 in buffer A) using glass beads (Braun) and cleared by centrifugation. The extracts were adjusted to a protein concentration of 10 mg/ml, and the binding assay was performed as described for the bacterial extracts except that Buffer B was used for washing the beads. Antisera used in Western blots were polyclonal anti-Mts2, monoclonal anti-HA12CA5 (Roche Diagnostics), and monoclonal anti-His 6 (a gift from V. Van Heiningen).
Co-precipitation-pREP1GST and pREP1GST-Pus1 were transformed into cut9HA cells and total extracts of S. pombe transformants expressing gst ϩ and cut9HA ϩ or gst-pus1 ϩ and cut9HA ϩ were prepared as described under "In vitro Binding Assays." 1 ml of the extract containing 5 mg of total protein were incubated with 15 l of glutathione-Sepharose (Amersham Biosciences) for 1 h at 4°C. The beads were washed four times in buffer B, resuspended in 50 l of SDS sample buffer, and 10 l were loaded onto a 10% SDS gel. GST and GST-Pus1 were visualized by Coomassie staining and Cut9HA by Western blotting with an anti-HA antibody (12CA5, Roche Diagnostics).

RESULTS
The UBL Domain of Rhp23 Binds to the Proteasome Subunit Mts4 -As a first step to characterize the interactions of Rhp23 with the proteasome we employed the yeast two-hybrid system. We tested the ability of Rhp23 to bind to the proteasomal subunits Mts1 (Rpn9), Mts2 (Rpt2p), Mts3 (Rpn12p), SpRpn2 (Rpn2), Mts4 (Rpn1), Pad1 (Rpn11), Pus1 (Rpn10), SpRpt1 (Rpt1), SpRpt6 (Rpt6), and to the proteasome-associated ubiquitin hydrolase Uch2. The two-hybrid assay detected an interaction of Rhp23 with Mts4, a subunit of the proteasomal base complex (data not shown). To test whether there is also a genetic interaction between rhp23 ϩ and mts4 ϩ we used the fission yeast mts4-1 strain that carries a mutation in the gene encoding Mts4 (27). By transforming mts4-1 cells with the plasmid pREP1Rhp23, in which Rhp23 expression is driven by the thiamine inducible nmt1 promoter, we were able to partially suppress the temperature sensitive (ts) phenotype of the mutant (Fig. 1A). As a control we also introduced the plasmid into mts2-1 cells, which have a mutation in the gene for the proteasome subunit Mts2 and display a temperature sensitive phenotype similar to mts4-1 (29). However, transformation of mts2-1 with pREP1Rhp23 could not rescue the mutant phenotype indicating that the interaction between rhp23 ϩ and mts4 ϩ is specific. In addition, the observed genetic interaction was dependent on the Rhp23 UBL domain as when this domain was deleted from the rhp23 ϩ gene suppression of the mts4-1 mutant no longer occurred. 2 In order to determine whether this genetic interaction reflected a direct physical interaction between Rhp23 and Mts4, we performed an in vitro binding assay. A GST-Rhp23 fusion protein coupled to glutathione-Sepharose was incubated with His-tagged full-length Mts4 as well as N-and C-terminal truncated versions of the protein expressed in Escherichia coli. As shown, GST-Rhp23 precipitated Mts4 from the bacterial extracts (Fig. 1B) demonstrating a direct interaction between these proteins. The use of truncated versions of Mts4 revealed that Rhp23 binds within a region stretching from amino acid position 181 to 407.
It has been shown that Rhp23 binds the proteasome via its UBL domain (10). Therefore the binding of Rhp23 to Mts4 should also be dependent upon the UBL domain. Truncated versions of Rhp23 fused to GST were used in an in vitro binding assay similar to the one described above. We found that only the truncation of Rhp23 containing the UBL domain could precipitate Mts4 indicating that the UBL domain mediates the binding between them (Fig. 1C).
To demonstrate that binding to the Mts4/Rpn1 subunit of the 26 S proteasome is a general property of proteins that contain a UBL domain we assayed the fission yeast Udp7 protein (SwissProt accession number Q94580). The Udp7 protein, like the Rhp23 protein, contains a UBL domain in its N terminus.
When assayed in vitro GST-Udp7 was able to precipitate Mts4 protein from the bacterial extracts demonstrating an interaction (Fig. 1D). This result is consistent with the hypothesis that the UBL domain is a general Mts4/Rpn1 interaction domain.
Pus1 Binds Mts4 in a Region Containing Repetitive Sequences-It has been demonstrated in budding yeast that Rpn10p (Pus1) interacts with Rpn1 (Mts4) in an in vitro binding assay (20). Thus we investigated interactions between these two subunits. Previously, a genetic interaction was detected between the genes encoding these two proteins. The mts4-1 allele was found to be synthetically lethal with a null allele of pus1 (9). The Pus1 binding site within Mts4 was determined in an in vitro binding assay using the same Mts4 constructs employed for mapping the Rhp23 binding site. An N-terminal GST fusion of Pus1 immobilized on glutathione-Sepharose was used to precipitate full-length Mts4 as well as truncated versions of the protein. We could narrow down the binding site for Pus1 on Mts4 to a stretch of amino acids ranging from position 408 to 582 (Fig. 2, A and B). This particular region contains a cluster of repeats that can also be found in SpRpn2, the fission yeast homologue of the budding yeast Rpn2p subunit of the proteasome, and in Cut4 a subunit of the APC/C or anaphase-promoting complex (30). Although the repeats are rather weakly conserved they appear to be present only in orthologues of the proteins mentioned. In the Prosite data base (www.expasy.ch/ prosite/) this structure is referred to as APC_SEN3_REPEAT (accession number PS50248), whereas the Pfam data base (www.sanger.ac.uk/Pfam/) lists it as PC_rep for proteasome/ cyclosome repeat (accession number PF01851). We will therefore call it PC repeat.
Pus1 Also Binds to PC Repeat Modules of SpRpn2 and Cut4 -Having shown that Pus1 binds to one of the PC repeat modules found in Mts4 we asked whether Pus1 might also interact with the PC repeat modules present in SpRpn2 and Cut4. Moreover we wanted to narrow down the region of Pus1 that is responsible for binding the PC repeat structures. We expressed truncated versions of Pus1 with an N-terminal GST fusion. These constructs were tested in a binding assay using His-tagged PC repeat modules of Mts4, SpRpn2 and Cut4. Both the full-length Pus1 and PusC⌬1 were capable of binding to all three PC repeat modules tested, whereas no binding was observed using Pus1N⌬1 and Pus1N⌬2 (Fig. 2B). From these data we conclude that the N-terminal region of Pus1 up to amino acid position 84 appears to contain a region necessary for binding PC repeat modules.
Pus1 and Rhp23 Interact with APC/C-We have demonstrated that Pus1 can bind to the PC repeat module present in the APC/C subunit Cut4 in vitro. If Pus1 binds to Cut4 in vivo, it should be possible to use the GST-Pus1 fusion protein to precipitate the APC/C from fission yeast extracts. Therefore we incubated GST-Pus1 and GST-Rhp23 with extracts from cut9HA cells that carry a genomically HA-tagged version of the APC/C subunit Cut9 and tested whether these GST fusions could precipitate the APC/C and/or the proteasome. Because Rhp23 and Pus1 bind to multiubiquitin chains it had to be considered that they might interact with multiubiquitinylated substrates attached to the APC/C. Taking that into account, we also used Pus1N5 and Rhp23PP, which have mutations in their ubiquitin binding domains and do not bind multiubiquitin (9,10). Furthermore, we tested truncated versions of Pus1, which lack the C-terminal multiubiquitin binding motif (Pus1C⌬1) as well as versions of Rhp23 lacking either its UBL domain (Rhp23N⌬1) or both UBA domains (Rhp23C⌬1). We were able to precipitate the APC/C and the proteasome using Pus1, Pus1N5, or Pus1C⌬1 fusion proteins but not Pus1N⌬1 or Pus1N⌬2 (Fig. 3A). These findings indicate that the interaction of Pus1 with the APC/C and the proteasome is independent of its ability to bind ubiquitin conjugates but can be assigned to the Pus1 N terminus up to amino acid position 84, which is consistent with the results of the mapping studies involving the Cut4 PC repeat region.
Rhp23, Rhp23PP, and Rhp23 but not Rhp23N⌬1 were found to precipitate the proteasome, confirming that the UBA domains of Rhp23 are not involved in proteasome binding (10). However, the APC/C binding activity of Rhp23 appears to be mediated by its multiubiquitin binding UBA domains as we could detect APC/C when beads coated with Rhp23 or Rhp23N⌬1 but not with Rhp23PP or Rhp23C⌬1 were used. Thus it is likely that Rhp23 binds via multiubiquitinylated substrates associated with the APC/C.
Given that Pus1 and Cut4 interact directly, we investigated whether any genetic interactions exist using existing mutant alleles of these genes. We crossed the⌬pus1 strain to the temperature sensitive mutant cut4-533 (31). Spores carrying both the cut4-533 and the ⌬pus1 allele were not viable. As a control ⌬ pus1 was crossed to cut9-665 another temperature sensitive APC/C mutant (31). Synthetic lethality was not observed between cut9-665 and ⌬pus1 indicating a specific genetic interaction between pus1 ϩ and cut4 ؉ .
In order to elucidate whether the genetic data mirror an in vivo interaction between Pus1 and the APC/C we performed co-precipitation studies. Extracts from cut9HA cells expressing either GST or GST-Pus1 were incubated with glutathione beads and analyzed by SDS-PAGE and Western blotting. As presented in Fig. 3B we detected the APC/C subunit Cut9 in precipitates from cut9HA/GST-Pus1 but not from cut9HA/GST indicating an in vivo interaction between Pus1 and the APC/C. DISCUSSION We have characterized the interactions of Pus1 and Rhp23, representing two families of multiubiquitin-binding proteins, with the proteasome and the APC/C. Our studies demonstrate that Rhp23 binds to Mts4, a base component of the proteasomal 19 S regulatory complex. The UBL domain of Rhp23, which is responsible for the interaction with the proteasome, appears to bind the Mts4 protein between amino acid positions 181 and 407. Using this sequence to search the data base demonstrates that it is only found in other Mts4/Rpn1 orthologues from other species. Interestingly, this sequence comprises part of a pfamB domain, pfam B_4211. A multiple sequence alignment of the conserved pfamB domain is shown in Fig. 4. A homology search of the recently sequenced S. pombe genome with this pfamB domain demonstrated that it was only present in the Mts4 protein.
Considering that there are a number of proteins carrying UBL domains two conclusions could be drawn. Either the structures that interact with the different UBL domains do not resemble each other, or all proteins containing a UBL domain can bind to Mts4/Rpn1. In this study we have shown that the UBL domain containing protein Udp7 is also able to interact directly with Mts4/Rpn1 indicating that the latter hypothesis seems more probable. Consistent with this hypothesis the Dph1/Dsk2, Bag1, and Ubp6/Usp14 UBL containing proteins In Pus1N5 the multiubiquitin binding LALAL motif is substituted by NNNNN and in Rhp23PP the conserved glycine residues 158 and 333 are changed to proline, causing a loss of the multiubiquitin binding ability for both proteins. After washing, beads were subjected to SDS-PAGE and Western blot analysis to detect precipitated proteasome or APC/C by using anti-Mts2 or anti-HA antisera respectively. B, extracts from cut9HA cells expressing either gst ϩ or gst-pus1 ϩ were incubated with glutathione beads, and precipitates were analyzed by SDS-PAGE and Western blotting. In the left panel a Coomassie stain of the precipitates is shown, visualizing GST and GST-Pus1 bands respectively. In the right panel 10% of the input cut9HA/pREP1GST and cut9HA/ pREP1GST-Pus1 extracts was loaded, and on the left panel 10 l of the precipitates were loaded and analyzed by anti-HA Western blotting.
have already been shown to bind the proteasome, presumably by interaction with the Mts4/Rpn1 subunit, although this has only been shown for the budding yeast Dsk2 protein (32-34).
Using protein cross-linking studies it has been reported that the budding yeast Dsk2 and Rad23 UBLs interact with both the Rpn1 and Rpn2 subunits (35). In contrast, in our in vitro binding experiments we found no interaction with the fission yeast Rpn2 orthologue and the fission yeast Rad23 orthologue, Rhp23. A recent study in budding yeast also indicated that the UBL domain of Rad23 specifically bound to the Rpn1 subunit and showed little affinity for the Rpn2 protein (32). However the UBL binding domain identified in Rpn1 by this group was different to that found by our binding studies of the Mts4 protein. However, in each case the UBL-binding region was defined by a series of nested deletions and the binding region left was internally consistent. Therefore, further experiments will have to be carried out to resolve this apparent paradox in the different UBL binding sites, which might reflect a difference in the budding and fission yeast proteins.
The structure of the UBL domain has recently been solved by NMR spectroscopy and shown to be related to ubiquitin (36). Therefore, we tested whether mono-ubiquitin or tetra-ubiquitin could bind to Mts4. In an in vitro binding assay we were not able to detect binding between those proteins (data not shown). In addition, in mammalian cells it has been shown that the hHR23 and S5a proteins, the human versions of the fission yeast Rhp23 and Pus1 proteins, can directly interact with each other (37). Using deletion analysis the S5a binding site was determined to be in the second UIM binding domain present in this protein. This domain is not present in the simpler versions of the fission yeast Pus1 and budding yeast Rpn10 proteins (9,38). Consistent with this we observed no direct interaction between the fission yeast Pus1 and Rhp23 proteins. 3 The proteasome subunits SpRpn2 and Mts4 appear to play a central role in the structural organization of the 19 S regulatory complex. They link the ATPases of the base and components of the lid and also harbor binding sites for proteins that recognize multiubiquitinated substrates (See Fig. 5). This provides a simple model for the delivery of ubiquitinated substrates between the lid subcomplex and the ATPase ring. Presumably upon release from the multiubiquitin-binding proteins the ubiquitnated subtrates are captured by the multiubiquitin binding Rpt5/S6Ј ATPase, before the substrate protein is unfolded by the action of the ATPase ring and the polypeptide translocated to the 20 S catalytic complex for degradation (39).
Characterization of the Pus1 association with Mts4 revealed a specific interaction between Pus1 and a cluster of PC repeats in Mts4. These repetitive sequences contain an alternating pattern of large aliphatic residues and glycine or alanine (30). It has been suggested that these repeats form a structure possessing a concave surface of a hydrophobic nature which might represent a binding site. According to the Prosite data base, Mts4 contains two PC repeat clusters with five and three repeats, respectively, which we call here PC-1 and PC-2. The mapping experiments show that PC-1 is sufficient to bind Pus1. It is conceivable that PC-1 and PC-2 together form one binding site. This would imply that deletion of PC-2 does not interfere 3  with the integrity of the binding site, but that PC-2 on its own is unable to form a functional binding structure itself. Mts4 binds not only Rhp23 and Pus1, its human homologue S2 was also shown to associate with the HECT E3 KIAA10 (22). SpRpn2 can bind Pus1 and for its budding yeast orthologue Rpn2p, an interaction with the RING finger E3 Ubr1p was demonstrated (20).
Data base searches revealed that PC repeats can be found only in the two largest subunits of the proteasome Mts4/Rpn1 and SpRpn2 and in the APC/C component Cut4/APC1. In binding studies we demonstrated that Pus1 could bind also to the PC repeat clusters of SpRpn2 and Cut4. This indicates that binding of Pus/Rpn10 represents a general and conserved function of these PC repeat structures.
The Pus1/Rpn10 protein represents a special case among regulatory subunits of the proteasome as it is the only subunit that also occurs in low molecular weight fractions of cell extracts (9,38,40). Furthermore it has been suggested that Rpn10p might have a role in linking the proteasome base and lid, based on the observation that the 19 S regulatory complex dissociates easily when purified from a null allele of RPN10 (6).
Our binding studies using truncated forms of Pus1 show that deletion of the N-terminal region up to amino acid position 84 is sufficient to abolish not only the ability of Pus1 to bind PC repeat modules but also its association with the proteasome and the APC/C. Glickman et al. (6) demonstrated that budding yeast Rpn10p deleted for its N-terminal region up to position 61 can still bind to the proteasomal base complex, but not to the lid. As the PC repeat-containing subunits of the proteasome are part of its base component the data suggests that in Pus1 the region between amino acids 61 and 84 is indispensable for the interaction with the PC repeat modules.
Interestingly the N terminus of Rpn10p up to position 50 has been shown to be critical for the degradation of Ub-Pro-␤-gal and resistance to amino acid analogues (14). Therefore the ability of Pus1/Rpn10p to bind the proteasome might be crucial for these phenotypic functions of the protein. However, considering that Pus1 binds multiubiquitin chains as well as associating with the APC/C and the proteasome, it is conceivable that it has a role in the delivery of multiubiquitinylated substrates to the proteasome. Pus1 and UBA/UBL proteins might represent separate but, in terms of substrate specificity, overlapping pathways that facilitate the transfer of substrates from ubiquitin protein ligases to the proteasome.
Alternatively one could speculate that Pus1 acts as a factor involved in delivery of multiubiquitinylated substrates, be it from the APC/C to UBA/UBL proteins and to the proteasome or directly from E3s to the proteasome. Depending on which pathway is used to deliver the substrate, Pus1 might associate with the appropriate PC repeat structure in order to facilitate the process. Therefore one would predict that Pus1 binds to Mts4 when substrates are acquired via Rhp23 or KIAA10 or it uses the SpRpn2 PC repeat for the Ubr1p pathway.
Recently it has been demonstrated that valosin-containing protein, a member of the AAA (ATPases associated with a variety of cellular activities) family, can bind ubiquitin chains and is essential for the ubiquitin-dependent degradation of certain proteasome substrates (41). Future work will shed light on the relationship between the various factors involved in the recruitment of multiubiquitinylated proteins to the proteasome and elucidate further the role of individual proteasomal subunits in this process.