Blm10 Protein Promotes Proteasomal Substrate Turnover by an Active Gating Mechanism

Background: Association of the proteasome core with activators regulates proteasome activity. Results: Blm10 association increases proteasome activity toward peptides and the unstructured proteasome substrate tau-441. This process is mediated by the C terminus of Blm10. Conclusion: C-terminal docking-mediated proteasome activation by Blm10 facilitates the turnover of peptide and protein substrates. Significance: Blm10 contributes to the regulation of proteasome activity. We investigated the impact of the activator Blm10 on proteasomal peptide and protein degradation. We found enhanced degradation of peptide substrates in the of Blm10 and that Blm10 has the capacity to accel-erate proteasomal turnover of the unstructured protein tau-441 Mechanistically, proteasome requires the opening of a closed gate, which allows of unfolded proteins into the catalytic Our indicate that gate by Blm10 is achieved via engagement of its C-terminal segment with the CP. Crucial for this activity is a C-terminal YY X with the penultimate tyrosine a role. Blm10 a gate opening strategy to the proteasomal ATPases HbY X -dependent

For optimal proteolytic function, the central core of the proteasome (core particle (CP) or 20S) has to associate with activators. We investigated the impact of the yeast activator Blm10 on proteasomal peptide and protein degradation. We found enhanced degradation of peptide substrates in the presence of Blm10 and demonstrated that Blm10 has the capacity to accelerate proteasomal turnover of the unstructured protein tau-441 in vitro. Mechanistically, proteasome activation requires the opening of a closed gate, which allows passage of unfolded proteins into the catalytic chamber. Our data indicate that gate opening by Blm10 is achieved via engagement of its C-terminal segment with the CP. Crucial for this activity is a conserved C-terminal YYX motif, with the penultimate tyrosine playing a preeminent role. Thus, Blm10 utilizes a gate opening strategy analogous to the proteasomal ATPases HbYX-dependent mechanism. Because gating incompetent Blm10 C-terminal point mutants confers a loss of function phenotype, we propose that the cellular function of Blm10 is based on CP association and activation to promote the degradation of proteasome substrates.
Regulated protein degradation by the ubiquitin-proteasome system is essential for the maintenance of eukaryotic protein homeostasis (1). The central component of this system is the 26S proteasome, a 2.5-MDa protease present in the cytoplasm and nuclei of eukaryotic cells (2). The proteasome is crucial for the removal of misfolded or damaged proteins, for providing building blocks for protein synthesis during nutrient limitation, and for the timed degradation of most short-lived proteins (3). The latter function is essential for the correct execution of many cellular processes including cell cycle control, transcription, translation, DNA repair, apoptosis, and antigen presentation (4). To execute its function as a universal protein degradation machinery, the proteolytic capacity of the proteasome is broad and for the most part indiscriminate. However, many control mechanisms regulate and limit proteasome function to prevent nonspecific protein degradation (5).
The 26S proteasome consists of two modules: a proteolytic core cylinder and a regulatory/activating module that binds to the core (2). The core (termed core particle (CP) 6 or 20S proteasome) is a barrel-shaped cylinder formed by four stacked rings (6). The two inner rings consist of seven homologous ␤-subunits each. Three of the ␤-subunits contain active sites with different substrate side chain preferences (7). The active sites face the interior of the core particle, thus creating a sequestered proteolytic chamber. The two outer rings are composed of seven homologous ␣-subunits, which provide docking sites for proteasome activators. To enter the internal degradation chamber, unfolded substrates must pass through the center of the ␣-ring. This pore is gated by the N termini of the ␣ subunits (8), which must be displaced by proteasome activators (9).
The proteasome core interacts with three activator families: (a) complexes composed of or containing AAAϩ ATPases (the eukaryotic regulatory particle (RP/19S/PA700) or the archaeal proteasome activating nucleotidase PAN); (b) the PA28/11S protein family, with PA28␣␤ hetero-heptamers and the homoheptameric PA28␥ present in higher eukaryotes, as well as PA26 complexes found in Trypanosoma brucei (10,11); and (c) the conserved Blm10/PA200 proteins found in all eukaryotes (9). Yeast Blm10 and its human ortholog PA200 (for proteasome activator with a molecular mass of 200 kDa) are monomeric proteasome activators that have been identified in hybrid complexes, where Blm10/PA200 occupies one end and the RP/19S the other end of the core cylinder (12)(13)(14). The same proteasome topology has been reported for PA28 family proteasome activators (15)(16)(17)(18). Blm10/PA200 proteins are composed of HEAT repeats and display an elongated, curved tertiary structure (13,19,20). Similar to RP or PA28 binding, Blm10 or PA200 association with the CP enhances proteasomal peptidase activity (12,13,21), suggesting that proteasome activation is a conserved function for this class of proteins.
The 26 S proteasome and PA28 -20S, Blm10, and PA200 proteasomes exhibit an open axial channel into the CP as demonstrated by cryo-electron microscopic studies (21,22). These findings are supported by a recent x-ray structure of doubly capped reconstituted Blm10-CP complexes, which features a partially open, disordered CP gate (20). The structure also revealed that Blm10 binds to the CP via its C terminus, which occupies the same binding sites as PA26 and the proteasomal ATPases. From the position of the Blm10 C terminus within the binding pocket and its effect on the conformation of the outer surface of the ␣-subunit ring, it was proposed that gate opening by Blm10 might be achieved in a similar mechanism as suggested for the proteasomal ATPases (20).
Blm10 has been implicated in the final phase of CP maturation. It was identified in late stage CP assembly intermediates, yet deletion of BLM10 does not result in a pronounced CP maturation defect (23). Loss of BLM10, however, in a mutant, which abrogates the association of the RP with CP (sen3-1) produced a marked proteasome maturation defect. These data suggest that the function of Blm10 during the final steps of proteasome maturation can be compensated by the RP and that activator binding to the CP in general is required for the final activation of the CP (23). After completion of proteasome biogenesis, Blm10 remains stably bound to the CP and affects the turnover of proteasome substrates (24).
We demonstrated recently that Blm10 is required for proteasome-mediated turnover of the transcriptional activator Sfp1 (24). Although Sfp1 protein levels increase upon loss of BLM10 in vivo, the turnover of Ubc6, a ubiquitin-dependent proteasome substrate, is not affected. These findings argue for a model in which Blm10 mediates the degradation of a specific subset of proteasome targets.
Sfp1 is a transcription factor required for the correct adaptation of ribosome biogenesis to the metabolic status of the cell, and the absence of Blm10 results in dysregulated ribosome biosynthesis. BLM10 is strongly induced during oxidative growth or in the presence of rapamycin (24), which specifically inhibits TORC1 signaling, the major nutrient-sensitive signaling pathway. Deletion of BLM10 furthermore causes partial mitochondrial dysfunction (20). These observations suggest that Blm10 proteasomes are involved in the regulation of metabolism.
Mechanistically, it is unclear whether these cellular functions are related to the physical interaction of Blm10 with the CP to activate proteasomal degradation of Blm10-specific substrates. Here we explore whether the described cellular functions of Blm10 could be tied to regulation of proteasome gating by studying the molecular details of proteasome activation by Blm10.
As described for PA26 and proteasomal ATPase binding, Blm10 association with the CP is mediated by its C terminus (20,24). The C termini of Blm10 proteins contain a conserved motif, YYX or sometimes YFX, which is analogous to the C-terminal HbYX motif (where Hb indicates hydrophobic, and X is variable) in PAN and three (Rpt2, Rpt3, and Rpt5) of the six eukaryotic proteasomal ATPases (see Fig. 1 and Refs. 25 and 26). Biochemical and structural studies indicate that this motif is essential for ATPase-induced gate opening (25, 26, 29 -31).
We show here that mutagenesis in that region of Blm10 produces a loss of function phenotype. Interestingly, a peptide derived from the last seven residues of Blm10 is sufficient to activate the CP peptidase activity, indicating that Blm10 opens the CP gate via a C-terminal docking mechanism, analogous to proteasomal ATPases. A peptide where the penultimate tyrosine was exchanged against an alanine, however, failed to activate the CP. Thus, this residue is crucial for Blm10-mediated proteasome activation.
An enzymatic characterization of proteasomal peptidase stimulation by Blm10 suggests that similar to its mammalian ortholog PA200 (12,14), Blm10 activates the caspase-like and tryptic-like activity of the proteasome more strongly than the chymotrypsin-like activity. Lastly, we provide evidence that Blm10 proteasomes are competent to contribute to the turnover of cellular proteasome targets by demonstrating that Blm10 enhances proteasome-mediated degradation in vitro utilizing hypophosphorylated unstructured tau-441 as a substrate.

EXPERIMENTAL PROCEDURES
Strains-The strains used in this study are listed in Table 1. Genomically integrated C-terminal mutants were constructed by homologous recombination using standard techniques (27,28) and were verified by sequencing. Primer sequences are available upon request.
Assay of Proteasomal Activity by C-terminal Peptides-Gate opening in yeast CP/20 S proteasome (0.2 g/100-l reaction) by C-terminal peptides was assayed in 50 mM Tris-HCl, 1 mm DTT, and 5% glycerol using the indicated fluorogenic substrates (100 M Suc-LLVY-AMC, 50 M Ac-nLPnLD-AMC, and 10 M Suc-LRR-AMC) at 32°C. The peptides derived from the C terminus of Blm10 (NH 2 -VLWRSYYA-COOH), from the C terminus of PA26 (NH 2 -GTDHMVS-COOH), and from the C terminus of PAN were synthesized by EZ Biolabs and were HPLC-purified to 98% purity. The peptides were dissolved in Me 2 SO prior to addition to the reaction. The rate of substrate hydrolysis (AMC liberation) was monitored in a Spectramax M5 spectrophotometric plate reader by exciting at 380 nm and measuring emission at 460 nm.
Phenotypic Analysis-Strains were grown overnight in YPD and diluted in 96-well plates to a density of 6 ϫ 10 6 cells/well followed by 5-fold serial dilutions. They were spotted onto YPD plates in the absence or presence of cycloheximide or on different carbon sources and incubated at the temperature indicated.

Analysis of Proteasome Complex Composition in Unfraction-
ated Lysates-Proteasomal complexes in lysates were detected as described previously (13), with the exception that the lysis was performed by cryogenic disruption using a Retsch MM301 grinder mill. Equal total protein (200 g) was loaded in each well.
Blm10 Proteasome Complex Purification-For the purification of WT and mutant Blm10-CP complexes, cells from yMS122 or yMS476 were collected and lysed by cryogenic disruption with a Retsch MM301 grinder mill. Purification was performed as described previously (13). Briefly, the cleared lysate was batch-incubated with IgG affinity gel (Cappel), and the beads were collected and washed. Proteasome complexes were eluted after cleavage with tobacco etch viral protease (Invitrogen). Subsequently, the purified complexes were first resolved on 3.5% acrylamide native gels (13), followed by an in-gel activity assay with the fluorogenic proteasome substrate Suc-LLVY-AMC as described previously (13) and SDS-PAGE followed by Sypro Ruby staining to assess complex composition and purity of the sample.
Proteasome Activity Assay-Three different fluorogenic peptide substrates were used to analyze proteasomal activity: Suc-LLVY-AMC (12.5-200 M), Ac-RLR-AMC (12.5-350 M), and Ac-nLPnLD-AMC (12.5-350 M). All of the substrates were purchased from BACHEM and prepared as 40ϫ stock solutions in Me 2 SO. Activity was assessed in 96-well assay plates (Costar 3631) by monitoring the kinetics of fluorescence increase of generated free AMC in a plate reader (Spectramax M5; excitation, 380 nm; emission, 460 nm). The assay buffer contained 50 mM Tris, pH 7.5, 5 mM MgCl 2 , 0.5 mM EDTA. Purified CP and Blm10-CP were used at a final concentration of 1 nM. The assay mixture was preincubated for 15 min at 30°C. Fluorescence increase was recorded at an interval of 30 s for 30 min with 5 s of mixing prior to each cycle. The data were analyzed with GraphPad Prism 5 to calculate the corresponding values of K half , V max , and R 2 . The equation was adapted from R. A. Copeland as provided on the GraphPad webpage.
In Situ Analysis of Blm10-CP Complex Activity-5 g of purified CP and Blm10-CP were separated on a 3.5% polyacryl-amide native gel. For the in-gel activity assay, the gel was incubated in 50 mM Tris, pH 7.4, 5 mM MgCl 2 , 1 mM ATP, 100 M Suc-LLVY-AMC (Bachem) for 10 min at 30°C. Activity was assessed by visualizing AMC fluorescence after proteasomal cleavage of LLVY-AMC on an ultraviolet screen. Subsequently, the total protein concentration was assessed using SYPRO Ruby Protein stain (Invitrogen) following the manufacturer's instructions. The same samples were analyzed by immunoblotting using anti-Blm10 (this study) and anti-proteasome 20 S core subunits (BioMol) antibodies. Images were taken on an EL Logic 100 imaging system (Kodak). For signal quantification, analysis was performed using the software ImageJ 1.41o. CP activity was corrected for the protein abundance in each individual band.
In Vitro Degradation of Tau-Reaction mixtures contained 0.5 mM EDTA, 5 mM MgCl 2 , 50 mM Tris-HCl, pH 7.4, and 600 nM recombinant tau-441 (recombinant human tau variant 2N4R; Enzo Life Sciences) in a final volume of 15 l. The reactions were supplemented with 1 nM of CP or Blm10-CP in the absence or presence of 40 M MG132 (CalBiochem) as indicated. The reactions were incubated at room temperature, and samples were taken at the indicated time points and boiled for 5 min with ELPHO sample buffer to terminate the reactions, followed by immunoblot analysis with anti-tau antibody (BIO-SOURCE). The signals were recorded using an ImageQuant LAS 4000 and analyzed by densitometry with the Image Quant software.

RESULTS
The Role of the C-terminal Residues of Blm10 in CP Binding and Activation-Sequence identity among Blm10 orthologs is as low as 17% (human versus yeast) (12). The physicochemical features of the C-terminal residues on the other hand are highly conserved (Fig. 1A). The last residue is always hydrophobic (Ala, Val, and Ile), and the penultimate is always aromatic (Phe or Tyr), whereas the preceding residue is always a Tyr. This motif thus resembles the HbYX motif in proteasomal ATPases and in the archaeal PAN (Fig. 1A) (29), which has been demonstrated to trigger gate opening. We have shown previously that the last three residues of Blm10 are required for CP binding and its function in vivo (24). Because CP gate opening can be induced by peptides derived from the eukaryotic ATPases Rpt2 and Rpt5 or from the archeal ATPase PAN (29 -31), we speculated that the C-terminal peptide of Blm10 (ct-Blm10) might be sufficient to induce gate opening and thus to stimulate CP peptidase activity. We tested this hypothesis with a WT Blm10 C-terminal peptide encompassing the last seven residues (ct-Blm10) and a con-trol peptide where the penultimate tyrosine was substituted with alanine (ct-Blm10Y-A). This conserved residue is critical for RP-and PAN-induced gate opening (29). We found that the C-terminal peptide markedly stimulated the hydrolysis of fluorogenic peptide substrates of the proteasome's trypsin-like (LRR-AMC) and the caspase-like (nLpnLD-AMC) activity (Fig.  1B). Because these substrates are cleaved by different sites and because their hydrolysis is limited by their entry rate, these findings suggest that Blm10 activates gate opening. Accordingly, the activity of a mutant CP with a constitutively open gate caused by N-terminal truncations at two of the ␣-subunits involved in the gating process (8) was not stimulated by ct-Blm10 (Fig. 1B). These data strongly suggest that the increased peptide hydrolysis by the CP, observed in the presence of ct-Blm10, is caused by gate opening and not direct stimulation of CP active sites. Furthermore, an exchange of the penultimate tyrosine in the peptide (ct-Blm10 Y-A) abrogated CP activation, confirming a crucial function either in CP binding or gate opening for this residue. The overall capacity of ct-Blm10 to stimulate yeast CP activity was comparable with that of the C-terminal peptide of PAN (ct-PAN), whereas the gating-incompetent C terminus of PA26 (ct-PA26) was ineffective in stimulating yeast CP (Fig. 1C), in accordance with results obtained for mammalian CP (29). In contrast to the trypsin-like or caspaselike activity, we did not observe stimulation of the chymotrypsin-like activity with ct-Blm10, which is consistent with data reported for the C termini of the proteasomal ATPases (29).
To further characterize the role of the C-terminal residues, we constructed genomically integrated point mutations where the last five residues were substituted with the residues corresponding to the same positions in PA26. Because the PA26 C terminus binds to the same pockets as Blm10, this strategy is expected to minimize structural disturbances during CP docking. We have shown previously that Blm10 is required for correct repression of ribosome biogenesis and that BLM10 deleted cells exhibit cycloheximide resistance at low concentrations of this translation inhibitor (24). We therefore tested whether the C-terminal mutants produce a loss of function phenotype and found that an exchange of the penultimate tyrosine (Y2142V) indeed mimicked BLM10 deletion, showing increased viability with cycloheximide present (Fig. 2A). A similar increase in viability was observed for S2140H, whereas a partial loss of function phenotype is observed for R2139D ( Fig. 2A). Mutation of the last residue (A2143S) or Y2141M had no effect on viability in the presence of cycloheximide. To explore whether the phenotypes were caused by disrupted Blm10-CP interaction, we studied the relative distribution of proteasomal complexes in unfractionated lysates after native gel separation followed by an in-gel activity assay. Y2142V and S2140H mutations abrogated binding, which was preserved in R2139D (Fig. 2B). We speculate that the partial loss of function phenotype of the latter mutant is caused by an impaired gating mechanism. These findings are in agreement with previously published data, which revealed that the deletion of the last residue of Blm10 or the substitution of the last seven residues with the gating incompetent PA26 C-terminal peptides allowed CP binding yet produced a loss of function phenotype (24). Right panel, sequence alignments of the C-terminal segments of Blm10 orthologs. S.c., Saccharomyces cerevisiae; C.a., Candida albicans; C.e., Caenorhabditis elegans; H.s., Homo sapiens; A.t., Arapidopsis thaliana. B, heptapeptides derived from the wild-type Blm10 C terminus (ct-Blm10), as well as a peptide where the penultimate tyrosine has been replaced by alanine (ct-Blm10 Y-A) were tested for their ability to stimulate CP peptidase activity against three fluorogenic peptides assaying the chymotryptic (Suc-GGL-AMC), the tryptic (Suc-LRR-AMC), and the caspase-like activity (Ac-nLPnLD-AMC) of the proteasome. The assays were performed with 0.2 g of purified wild-type CP and an open gate mutant, where the N termini of ␣3 and ␣7 were deleted. The open gate mutant was purified from either WT cells or cells deleted for BLM10. The data are normalized for the activity of wild-type CP with the respective substrate. C, to compare ct-Blm10-mediated stimulation with ct-PAN and ct-PA26, the same assay as in A was performed with Suc-LRR-AMC, the proteasome substrate that showed the highest stimulation for the C-terminal peptide of Blm10. DMSO, dimethyl sulfoxide.
The Impact of Blm10 on Enzymatic Activities of the Proteasome-The C-terminal Blm10 peptide impacted the different proteasomal peptidase activities of the CP differentially. Although the chymotrypsin-like activity remained largely unaffected, the trypsin-like and the caspase-like activities were elevated (Fig. 3). Differential stimulation has also been observed for purified PA200 -20S complexes (12). To test whether the effects obtained with ct-Blm10 are also observed with the endogenous Blm10-CP complex, they were purified and subjected to a detailed enzymatic analysis. We obtained essentially the same results as for the ct-Blm10 stimulated CP peptidase activities: stimulation of trypsin-like and caspase-like activities, but unaffected chymotrypsin-like activity (Fig. 3). Prior studies have in fact shown that chymotrypsin-like fluorogenic peptide substrates can induce gate opening (32), and because of their very rapid cleavage by the chymotrypsin-like-site, their hydrolysis is less dependent on gate opening than the basic and acidic substrates, whose entry is rate-limiting. Although we were unable to determine the half-maximal binding constant or a proper V max for the trypsin-like activity because saturation could not be reached, we did not observe a change in binding affinity for the caspase or the chymotrypsin substrate, yet observed changes in V max for the caspase substrate, arguing for a model in which Blm10-mediated stimulation of peptide hydrolysis does not originate from an allosteric effect on the active sites. These results are consistent with an active gateopening mechanism induced by Blm10 and are corroborated by the observation that ct-Blm10 could not activate peptide hydrolysis in open-gated CP mutants (Fig. 1B).
Mitochondria-related Phenotypes of Blm10 C-terminal Mutants and of Blm10 Overexpression-We reported recently that a chimeric Blm10 protein, where the last seven residues were exchanged against the corresponding residues in PA26 (Blm10PA26C), allowed complex formation with the CP, yet cells expressing this protein are characterized by a loss of function phenotype (24). Purification resulted in low yields of intact Blm10PA26C-CP complexes, which might be indicative of reduced binding affinity (24). To improve the purification yield, we substituted the endogenous promoter of the mutant with the overexpressing promoter of GAL1; however, the yield of intact complexes remained low (supplemental Fig. S1). Because a biochemical characterization was therefore impeded, we studied potential effects of this mutation in vivo by comparing growth-related phenotypes of cells overexpressing BLM10 or the BLM10PA26C chimera.
To investigate potential epistatic effects with proteasome hypomorphs, we generated the same strains deleted for the proteasome-related transcription factor Rpn4. rpn4⌬ cells are characterized by a reduced proteasome pool (34,35). If the overexpression of BLM10 or BLM10PA26C causes proteasome dysfunction, the phenotype should be exacerbated in cells deleted for RPN4.
Blm10 has recently been implicated in mitochondrial function and oxidative metabolism because loss of BLM10 results in a high frequency of cells with dysfunctional mitochondria (20). We observed decreased viability of cells overexpressing BLM10 during growth on nonfermentable carbon sources at higher temperature in the absence of RPN4 (Fig. 4A, lower panel). The data support a role for Blm10 in correct maintenance of mitochondrial activity and indicate that this function is linked to the proteasome. Loss of BLM10, on the other hand, did not result in significant growth defects under these conditions. The C terminus of PA26 is unable to induce gate opening. Thus, replacing the last seven residues of Blm10 with the corresponding residues in PA26 should abrogate the gate-opening function of Blm10. Consistent with this hypothesis. we observed a gain-of-function phenotype in cells overexpressing the Blm10PA26 chimera. After induction of the gene in the presence of galactose, the cells were not viable at higher temperature or on nonfermentable carbon sources (YPGly). Furthermore, co-deletion of RPN4 aggravated the phenotype, resulting in lethality under inducing conditions (YPGal) already at normal growth temperatures (Fig. 4A, upper panel). In summary, although cells tolerate Blm10 overexpression unless oxidative metabolism is induced at higher growth temperature, elevated expression of a gating incompetent Blm10 version is toxic at higher temperature or in cells with reduced proteasome capacity as in cells deleted for RPN4.
Although previous reports suggested that plasmid-derived overexpression of WT Blm10 might be lethal (33) or might result in a slow growth phenotype (21), overexpression from the genomic locus (GAL1pBLM10) did not produce a visible phenotype at normal growth temperature under the conditions chosen (Fig. 4A, upper panel). Because the lethal phenotype was previously speculated to be caused by blocking RP-binding sites in the presence of elevated Blm10 levels (33), we investigated proteasomal complex formation in cells overexpressing BLM10 via native gel electrophoresis in unfractionated lysates. Although we observed a partial reduction in RP 2 -CP complexes, the dominant change upon Blm10 overexpression was a complete shift of the RP-CP band to the hybrid Blm10-CP-RP complex (Fig. 4B). This finding argues for a model in which Blm10 preferentially interacts with the unoccupied surface of RP-CP complexes rather than competing for RP binding. A significant population of Blm10-CP species without co-associated RP was not observed (Fig. 4B).
Doubly Capped Blm10-CP Complexes Retain Proteasomal Peptidase Activity-Recently, a study performed on Blm10-CP in unfractionated lysates or with reconstituted complexes suggested that although the binding of one Blm10 promotes the peptidase activity of the CP, Blm10 binding to both ends of the cylinder inhibits CP activity (36). This model is difficult to reconcile with a proposed function of Blm10 as a proteasome activator and is not supported by an x-ray crystallographic study, which demonstrates that both gates are partially open in CP flanked on both ends with Blm10 (20). To assess this model, we purified endogenous Blm10-CP complexes under conditions that yielded a high portion of doubly capped Blm10-CP (Fig.  5A). The contribution of the different complexes toward CP peptidase stimulation was evaluated through an in situ analysis, where we determined the peptidase activity and the relative protein content after an in-gel activity assay (Fig. 5A, top two  panels). Because of its broad linear dynamic range, protein content was assessed with the fluorogenic dye SYPRO Ruby. To ascertain that all of the signals analyzed were in the linear range, we performed standard curves for fluorescence increase after cleavage of the fluorogenic peptide Suc-LLVY-AMC by the CP, as well as for protein stain using Sypro Ruby (supplemental Fig. S2). As shown in Fig. 5B, doubly capped Blm10 proteasomes are fully active. To demonstrate that the slowest migrating complex indeed contains doubly capped Blm10-CP, we analyzed the Blm10 content in the bands with an anti-Blm10 antibody (Fig. 5A, second to  last panel).
Blm10 Promotes Protein Degradation in Vitro-Cryo-EM (21) and x-ray crystallographic studies (20) revealed a fully or partially open gate upon Blm10 binding, respectively. Furthermore, the x-ray structure revealed a pore of 13 ϫ 22 Å within the Blm10 dome. The largest opening in the PAN complex from Methanocaldococcus jannaschii has a diameter of 13 Å (37), and this ATPase complex promotes the degradation of unfolded proteins by the proteasome (38). Thus, we speculated that the overall topology of the Blm10-CP complex might accommodate the passage and subsequent degradation of an unstructured protein. To test this, we established an in vitro assay to investigate whether Blm10-CP complexes are able to promote the degradation of recombinant unstructured tau-441 (supplemental Fig. S3), a substrate that has been demonstrated to be degraded by the proteasome in vitro in the absence of ATP or ubiquitination (39). For this assay, equal molar amounts of CP or Blm10-CP (Fig. 6, A and B) were incubated with tau in the presence or absence of the proteasome inhibitor MG132. While in the presence of the proteasome inhibitor MG132, tau turnover was blocked, both CP and Blm10-CP complexes promoted tau degradation. Blm10-CP complexes exhibited an ϳ2-fold accelerated turnover of the unstructured protein (Fig. 6, C and D) as compared with CP alone. We conclude that Blm10 binding activates the proteolytic activity of the CP, resulting in a more efficient turnover of unstructured proteins.

DISCUSSION
The cylindrical proteasome core particle is assembled in a latent form and has its active sites sequestered within the interior of the cylinder (6). Stimulation of the proteolytic capacity of the CP requires its association with a proteasome activator (9).   mini of the CP ␣-subunits. While they form an interlaced layer in the closed form, thus occluding substrate entry, upon binding of PA26, they rotate away from the body of the cylinder and adopt an ordered 7-fold symmetric conformation in the open form (40). The structural motifs in the N-terminal portion of the ␣-subunits, which are involved in the gating mechanism, are the rigid helix H0, a reverse turn, and the extreme N termini of each ␣-subunit. Two mechanistically and structurally different strategies for proteasome gating have been reported or suggested so far. Biochemical, cryo-EM, and x-ray diffraction studies indicate that the proteasomal ATPases induce gate opening through the docking of their C termini, but their final position in the CP binding pockets has not been resolved yet (25, 26, 29 -31). The crystal structure of PA26, a homolog of the mammalian PA28 proteasome activator family (11), in a complex with yeast CP revealed a different proteasome activation mechanism: although the heptameric PA26 ring still associates with the CP ␣-ring through their extreme C termini, activation is achieved via an internal structure, the activation loop (40,41). It interacts with a conserved Pro-13 within the reverse turn in the proteasome ␣-subunits. Displacement of Pro-13 appears to drive the formation of the open pore by dislocating the reverse turn by ϳ1 Å, a movement that is propagated to the N termini (40).
Blm10 orthologs contain a universally conserved motif at their C termini, YYX or YFX, that is analogous to the HbYX motif found in proteasomal ATPases involved in CP gating. The data presented in this study demonstrate that this motif is actively involved in the CP gate-opening mechanism by Blm10 proteins. Mutagenesis in that region affects the function of Blm10. Two different phenotype classes were observed. Manipulating the penultimate tyrosine (Tyr-2142) either through deletion of the last two residues (24) or through exchange against a valine as shown in Fig. 2 abrogates proteasome binding and produces a loss of function phenotype. The same observations were made for an exchange of Ser-2140 against histidine. Exchange of Arg-2139 against an aspartate, on the other hand, promotes binding. The cells, however, exhibit a partial loss of function phenotype, which we speculate to be caused by inefficient CP gating. Similar characteristics were observed for a Blm10 chimera, where the C terminus of Blm10 was exchanged against the seven last residues of the gating-incompetent PA26 C terminus (24).
We furthermore report that a heptameric peptide derived from the Blm10 C terminus is sufficient to stimulate the peptidase activity of WT CP but not of a CP mutant with a constitutively open gate. A Blm10 C-terminal peptide where the penultimate tyrosine has been replaced by an alanine loses the ability to stimulate. These data are in agreement with a model in which docking of the C terminus of Blm10 occurs in a similar fashion as the proteasomal ATPases C termini and induces structural changes in the ␣-ring surface that result in gate opening. Thus, our data demonstrate that the HbYX motif is crucial for the Blm10 proteasome interaction as to mediate gate opening. It furthermore supports information obtained from a recent crystal structure demonstrating that the HbYX motif in Blm10 repositions crucial residues of the reverse turn within its binding pocket and induces a similar conformation as observed in the open conformation of the PA26-CP complex (9,20).
Blm10 is an unusual proteasome activator because it is a monomeric protein of 246 kDa in contrast to PA28 activators or the proteasomal ATPases, which form oligomeric rings (42)(43)(44). Blm10 binding to the proteasome breaks the 7-fold pseudosymmetry of the core cylinder, because it engages only with a single binding pocket. However, breaking the CP symmetry during activation is not without precedent. Biochemical analyses of the activation capacity of the RP ATPase C termini revealed that only three of the six ATPases effectively stimulate CP activity (29), and it is likely that only one HbYX C terminus may interact with the CP at any given time (45). This hypothesis is supported by the finding that although the C terminus of Rpt3 does contain a HbYX motif, its function appears to be restricted to CP docking rather than gating (46). These reports argue for a mechanism in which an asymmetric engagement of the activator binding pockets within the ␣-ring surface is sufficient for gate opening, although in the case of the RP, the binding of ubiquitin conjugates somehow further enhances gate opening and substrate entry (47).
In summary, whereas all proteasome activators bind via their C termini to the same binding pockets located at the CP surface, two different gate opening mechanisms can be used: C-terminal docking-mediated activation by the proteasomal ATPases and Blm10 and activation loop-mediated gate opening by the PA26/PA28 family of proteasome activators. Mechanistically, proteasome activators can also be classified according to whether they possess ATPase activity or not. Even an open CP gate allows only restricted access to the internal degradation chamber of the proteasome, and substrate proteins must be unfolded before they can enter (38,43,48). The proteasomal ATPases within the RP and PAN are able to perform this activity, whereas PA28 and Blm10 activators do not.
Very few targets for PA28 have been described (49), and there is little information on which substrates are specifically degraded by Blm10 proteasomes. The only protein so far that has been reported to require Blm10 for correct proteasomemediated turnover is the ribosome-related transcription factor Sfp1 (24). Many transcription factors or transcriptional co-regulators are predicted to contain large intrinsically unstructured regions (50). The graphic webserver FoldIndex© (51) provides an algorithm to predict unstructured regions in proteins based on their primary sequence. Such an analysis of the Sfp1 sequence predicts a high portion of intrinsically unstructured regions (supplemental Fig. S3A). A preference for degrading unstructured proteins might explain the lack of ATPase activity in Blm10 activators. This hypothesis is corroborated by our findings that Blm10 is able to stimulate CP-mediated degradation of the unstructured proteasome substrate tau-441 (supplemental Fig. S3B) in vitro as shown in this report. We infer from these data that the structural change of the CP gate upon Blm10 binding is sufficient for passage of both peptides and unfolded protein substrates. We propose that the ability of Blm10 proteasomes to promote the degradation of unfolded proteasome substrates might represent a critical and perhaps principal function of Blm10 in vivo.