The HEAT repeat protein Blm10 regulates the yeast proteasome by capping the core particle

© 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES The HEAT repeat protein Blm10 regulates the yeast proteasome by capping the core particle Marion Schmidt1, Wilhelm Haas1, Bernat Crosas1, Patricia G Santamaria1,2, Steven P Gygi1, Thomas Walz1 & Daniel Finley1 Proteasome activity is fine-tuned by associating the proteolytic core particle (CP) with stimulatory and inhibitory complexes. Although several mammalian regulatory complexes are known, knowledge of yeast proteasome regulators is limited to the 19-subunit regulatory particle (RP), which confers ubiquitin-dependence on proteasomes. Here we describe an alternative proteasome activator from Saccharomyces cerevisiae, Blm10. Synthetic interactions between blm10⌬ and other mutations that impair proteasome function show that Blm10 functions together with proteasomes in vivo. This large, internally repetitive protein is found predominantly within hybrid Blm10–CP–RP complexes, representing a distinct pool of mature proteasomes. EM studies show that Blm10 has a highly elongated, curved structure. The near-circular profile of Blm10 adapts it to the end of the CP cylinder, where it is properly positioned to activate the CP by opening the axial channel into its proteolytic chamber. Protein breakdown underlies the control of a wide variety of cellular functions. In parallel with signal transduction pathways and the control of gene transcription, the regulated degradation of specific proteins is crucial to both the induction and termination of many cellular responses. Eukaryotic cells contain two major proteolytic systems: the lysosomal proteases and the 26S proteasome. Whereas the lysosome is a membrane-enclosed organelle containing many monomeric proteases, the proteasome is a complex oligomeric structure that is present in both the cytoplasm and nucleus of the cell. The proteasome is an ATP-dependent, 2.5-MDa protease composed of the 20S CP, which contains the proteolytic active sites, and the 19S RP, which is required for substrate recognition1. The crystal structure of the CP resembles a barrel-shaped cylinder composed of four stacked heptameric rings. The two outer rings are each formed by seven α subunits, and the two inner rings by seven β subunits, three of which harbor relatively nonspecific proteolytic active sites. To prevent indiscriminate protein degradation, the active sites are buried within the central cavity of the CP cylinder. Substrate entry into this cavity is restricted by a narrow channel that exhibits closed and open conformations. In the closed state the cavity is sealed by N-terminal tails of the α subunits2,3. The RP consists of six ATPases and 13 other subunits of mostly unknown function1. The ATPases have chaperonelike activity4,5 and are thought to form a ring structure that physically interacts with the CP6,7. The chaperone-like activity seems to be required to unfold substrate proteins8,9, a prerequisite for passage through the narrow pore of the CP. Specific subunits of the RP participate in the binding of polyubiquitinated proteins, the major class of substrates for the proteasome10–12, and mediate their deubiquitination prior to unfolding and degradation13–15. A key mechanistic feature of the proteasome is the regulated activation of the CP for substrate degradation. Through physical interaction of activating complexes with the CP, the gate into the central proteolytic cavity is opened, allowing substrates to be degraded. In mammalian cells four different proteasome activators have been identified: PA700 (RP or 19S), PA28αβ, PA28γ and PA200 (refs. 6,16,17). The crystal structure of a heterologous complex between Trypanosoma brucei PA26 (homologous to mammalian PA28αβ) and CP isolated from S. cerevisiae exemplifies the mechanism of activation18. Loop extensions from the PA26 subunits dock into surface crevices of the CP α subunits, thereby inducing a conformational change within the N-terminal tails of the α subunits. Upon binding of PA26, these segments migrate toward the activator and away from the long axis of the CP, thus opening a pore19. Mechanisms of this type are probably common to all CP-activating complexes. The ability of the CP to interact with various activating complexes adds an additional layer of proteasome regulation, which in mammalian cells is at least to some extent specific at the levels of tissues, subcellular compartments and physiological stimuli. For example, interferon-γ induces the expression of PA28αβ20. In S. cerevisiae the only known proteasome activator is the RP; no homolog of PA28 proteins has been identified. Although an apparent PA200 ortholog in yeast exists (Blm10, originally termed Blm3 (ref. 21)), the protein has been reported to interact with immature CP precursor complexes rather than with mature active proteasomes and has been proposed to function specifically in proteasome maturation22. In this study we used S. cerevisiae as a model system to explore the regulation of proteasome activity. We found that Blm10, the S. cerevisiae ortholog of mammalian PA200, associates with mature, active proteasome holocomplexes to form a previously unidentified 1Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, Massachusetts 02115, USA. 2Present address: Department of Pathology, MSB#599, NYU School of Medicine, 550 First Avenue, New York, New York 10016, USA. Correspondence should be addressed to D.F. (daniel_finley@hms.harvard.edu). Published online 20 March 2005; doi:10.1038/nsmb914 294 VOLUME 12 NUMBER 4 APRIL 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES Figure 1 Identification of a novel S. cerevisiae proteasome subcomplex containing core particle subunits and Blm10. (a) Identification of distinct proteasomal subcomplexes. Left, proteins from conventionally purified proteasomes were separated on a 10% (w/v) polyacrylamide gel and stained with Coomassie blue. Arrow indicates a previously unidentified proteasome component, subsequently identified as Blm10. The band below Blm10 is Ecm29 (ref. 14). Right, the same sample was subjected to native gel analysis followed by an in-gel activity assay with the fluorogenic peptide substrate suc-LLVY-AMC. Arrow indicates the previously uncharacterized proteasome subcomplex in S. cerevisiae. (b) Yeast proteasome core particles associate with Blm10. Proteasome complexes from an ecm29∆ strain bearing a protein A–tagged Rpn11 (yMS15) were purified by IgG-affinity chromatography and further fractionated on a size-exclusion column. Fractions were analyzed for peptidase activity after native gel separation (top) with the fluorogenic substrate suc-LLVY-AMC, followed by staining of the gels with Coomassie blue (middle). Active fractions were subjected to SDS-PAGE (bottom). CPx designates the novel proteasome species. The identity of the high-molecular-mass doublet band in lane 4 (arrow) was determined by tandem mass spectrometry. (c) Proteasome core particles isolated from a blm10∆ strain lack CPx. Core particles carrying a protein A–tagged Pre1 were isolated by IgG-affinity purification from wild-type (WT) (yMS31; lane 1) or blm10∆ cells (yMS94; lane 2). The purified complexes were subjected to SDS-PAGE separation (top) and to native gel electrophoresis followed by an in-gel activity assay (bottom). SDS-PAGE separation of CPx (purified either via a protein A tag at the RP subunit Rpn11 as in b or via a protein A tag at the CP subunit Pre1 as in c) revealed the presence of additional proteins, designated by asterisks. The presence of these proteins varied depending on the purification method used, suggesting that they are nonspecific copurifying proteins or degradation products. proteasome subpopulation in the cell. Blm10 binds to the axial pore of the CP cylinder and strongly activates the peptidase activity of the CP. Although a distinct subpopulation of proteasomes is associated with Blm10, deletion of the gene is not sufficient to produce a global defect in proteasome activity. However, deletion of both BLM10 and RPN4, the general transcription factor controlling proteasome levels, leads to a reduced growth rate and to increased canavanine sensitivity, suggesting that if the cell is unable to adjust proteasome levels properly, Blm10 function becomes essential. Structural analysis of Blm10–CP complexes reveals that Blm10 docks onto the ends of the CP cylinder, thus enveloping the CP channel and fulfilling the topological requirements for a proteasome activator. We also present a structural analysis of isolated Blm10. This unusually large protein is exclusively composed of HEAT repeats, forming an α-solenoid structure. Its elongated, curved conformation exposes a large, solvent-accessible surface area, thus allowing Blm10 to interact with the heptameric ring of α subunits that form the axial surface of the CP. RESULTS Blm10 associates with proteasome core particles The only known CP-activating complex in S. cerevisiae is the RP23. However, native gel separation of conventionally purified proteasomes, followed by visualization of active proteasomes with fluorogenic peptides, revealed an uncharacterized proteasome species in S. cerevisiae (Fig. 1a). Purification of proteasomes in the absence of proteasome component Ecm29 or ATP results in dissociation of RP and CP14. By using an ecm29∆ strain, we enriched for the uncharacterized proteasome species. Proteasomes bearing a protein A tag on an RP subunit were bound to IgG resin. CP was eluted by high-salt treatment and subjected to size-exclusion chromatography, separating the novel proteasome species (CPx) from fractions containing free CP (Fig. 1b). The appearance of CPx in the native gel coincided with the presence of a high-molecular-mass protein in SDS-PAGE analysis (Fig.1b). This protein was identified by mass spectrometry as Blm10 (Supplementary Fig. 1 and Supplementary Methods online), the yeast homolog of PA200 (ref. 17). To confirm that Blm10 is indeed a component of CPx, we purified CP from a blm10∆ strain. Blm10 and the new CP species were missing in these preparations (Fig. 1c), whereas the CP itself seemed unaffected. Thus, Blm10 is specifically required to form CPx. Blm10 is a proteasome activator The mammalian homolog of Blm10 enhances the peptidase activity of CP17. To determine whether this property is conserved among Blm10 proteins, the Blm10–CP complex was purified from a strain expressing protein A–tagged CP and overexpressing HA3-tagged Blm10. CP samples were affinity-purified and then fractionated by size-exclusion chromatography. Blm10-containing fractions were identified by SDSPAGE (Fig. 2a). Each fraction was subjected to native gel electrophoresis followed by an in-gel activity assay and silver staining. A parallel gel was immunoblotted with an anti-HA probe to visualize Blm10, and with anti-α7 to visualize CP. Overexpression of Blm10 resulted in an apparent Blm102–CP complex. Thus, three CP-containing bands were detectable, two containing Blm10. The highest activity was detected in the Blm10-CP band (Fig. 2b, top, lane 1), although the highest protein amount was present in the band containing free CP, indicating that Blm10 association with the CP leads to CP activation. In a parallel purification from a blm10∆ mutant, no activated CP was detectable (Supplementary Fig. 2 online). To estimate the extent of CP activation by Blm10, peptide hydrolysis was assayed fluorometrically (Fig. 2c). A roughly four-fold activation was observed in the fraction containing the highest proportion of Blm10–CP complexes, compared with a fraction devoid of Blm10. Because the most abundant CP species in this sample (Fig. 2b, lane 1) did not contain Blm10, it must activate the CP by considerably more than four-fold. Thus, the effectiveness of CP activation by Blm10 may be comparable to that of the RP, which is often reported as eight- to ten-fold23. NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 4 APRIL 2005 295 © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES Figure 2 Blm10 interaction with the CP results in enhanced peptidase activity. (a) Coomassie blue stain of three consecutive Blm10containing Superose 6 fractions after SDS-PAGE (top, lanes 1–3). Blm10 degradation products were identified when the bands marked by an asterisk were subjected to mass spectrometry. (b) The fractions shown in a (3 µg) were subjected to native gel electrophoresis, followed by an in-gel activity assay using the CP substrate Suc-LLVYAMC (top) and silver nitrate staining (second panel). A parallel gel was immunoblotted and stained with an anti-HA probe (third panel), followed by stripping of the membrane and reprobing with an antiα7 (bottom). (c) The extent of CP activation by Blm10 was assessed by measuring the CP peptidase activity in a liquid-based assay using 0.5 µg of the fractions shown in a. The slopes of the individual reactions are presented in the bottom panel. (d) Reconstitution of Blm10–CP complexes is accompanied by CP activation. Left, purified input Blm10; middle, CP purified from blm10∆. Formation of complex was monitored by an in-gel peptidase activity assay in the presence of Suc-LLVY-AMC after native gel separation (top right) and Coomassie blue staining of the native gel (bottom right). The molar ratio of Blm10/CP was 2.7:1. To further verify the capacity of Blm10 to activate CP, we used a reconstitution strategy. Purified Blm10 (Fig. 2d) was reassociated with CP isolated from blm10∆ strains. The reconstituted Blm10–CP complex was found to be activated because, despite the strong signal observed in native gels via activity staining, the intensity of the same band stained with Coomassie blue was low. However, reconstitution was inefficient, possibly owing to instability of the large protein in the absence of CP. Structural characterization of Blm10–CP complexes The assignment of Blm10 as a CP activator predicts that Blm10 should dock directly onto the axial end of the CP cylinder, allowing Blm10 to regulate the state of the CP channel. To test this possibility, the structure of freshly purified Blm10–CP complexes was analyzed by negative-stain EM. Multireference alignment and classification of extracted particle images were used to generate projection averages of the different particle species (Fig. 3a, right). Indeed, Blm10 physically associates with the outer rims of the CP cylinder, forming a caplike structure that covers the channel into the proteolytic chamber. Blm10–CP has a high degree of structural similarity to cocomplexes of the mammalian CP and PA28 (refs. 20,24,25), a well-characterized mammalian proteasome activator (Fig. 3b). Notably, whereas all known CP activating particles, including PA28, are multisubunit complexes with a ring structure, the large, single polypeptide chain of Blm10 has apparently evolved to accommodate binding to a circular surface with seven-fold pseudo symmetry. To our knowledge a similar interaction between a single protein and a cylindrical oligomeric complex has not been described. Figure 3 Structural analysis of Blm10–CP complexes. (a) Freshly purified Blm10–CP complexes from yMS31 were imaged by EM. To obtain a more refined structure of the individual proteasome species (top panel, right half), 803 CP top-view particles, 777 CP side-view particles, 182 CP-Blm10 particles and 15 Blm102-CP particles were collected and averaged. The average and ten representative raw images are shown for each species. The size of Blm10 in association with the CP, as inferred from its electron density, is consistent with binding of a single polypeptide chain to the CP. The side length of the individual panels of averaged particles is 54 nm. (b) Structural comparison of Blm10–CP and PA28–CP complexes. Bottom panels were reproduced from ref. 25 with permission. 296 VOLUME 12 NUMBER 4 APRIL 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES Figure 4 Structural characterization of Blm10. Immediately after sizeexclusion chromatography of Blm10 that was dissociated from the CP (see Fig. 2d), Blm10-containing Superose 6 fractions were analyzed by EM. (a) Averages of multiple Blm10 configurations. Between 100 and 200 particles for each configuration were collected and averaged. Averages of CP and Blm10–CP complexes (Fig. 3a) are presented for comparison with isolated Blm10 in the lower panels. The side length of the individual panels is 32 nm. (b) Size comparison of averaged CP top views (Fig. 3a) with an averaged circular Blm10 configuration. (c) An averaged image of the intact Blm10–CP complex as shown in Figure 3a (left) is compared with a composite figure of averaged free CP and a compact free Blm10 configuration. be found after purification as a component of mature, highly active CP complexes. We therefore considered whether the presence of Blm3 in purified mature CP (Fig. 1a) could be misleading. Because Blm10 might redistribute onto mature CP artifactually during the course of prolonged purification procedures, we characterized the complexes formed by Blm10 in the absence of chromatographic manipulations. Fresh, unfractionated yeast extracts were applied directly to native gels, and after electrophoresis proteasomal complexes were visualized using the in-gel activity assay. We expected a defect in proteasome maturation to be manifested either in inherent instability of mature proteasome holocomplexes or in a lower abundance of mature proteasomes. Such effects were not detected in blm10∆ lysates under the experimental conditions used (Fig. 5a). Unexpectedly, we could not detect free Blm10–CP complexes by this method (Fig. 5a), although we initially identified Blm10 in a complex with CP (Fig. 1). Instead only RP-capped proteasome complexes were detected, suggesting that free CP complexes are nonabundant in yeast cells, and that Blm10 in the cell might reside primarily in RP–CP complexes. To test for the presence of Blm10 in RP–CP complexes, bands exhibiting proteasome activity were excised from native gels that had been loaded with fresh lysate (Fig. 5b) and subjected to tandem mass spectrometry (Supplementary Methods online). The Blm10 peptides obtained are shown in Supplementary Figure 3 online. In summary, this analysis indicated that Blm10 in unperturbed lysates is predominantly bound to RP–CP complexes. Consistent with our finding that Blm10 binds the cylindrical ends of CP, Blm10 was found in association with RP-CP but not RP2-CP species (Fig. 5b). Notably, extended electrophoresis led to a broadening of the RP-CP band, suggesting that it might represent two distinct species (Fig. 5b). By subjecting lysates to protracted electrophoresis, the RP-CP band was resolved into a doublet, with the upper band absent in blm10∆ lysates (Fig. 5c). Thus, Blm10 forms a novel Blm10-CP-RP hybrid proteasome. From comparing the relative abundance of proteasomal subcomplexes in freshly lysed cells to that after mild affinity purification (Fig. 5), we further conclude that Blm10–CP–RP as well as RP–CP complexes are dynamic and labile, dissociating rapidly during the course of either conventional or affinity-based purification. Blm10-CP-RP proteasomes are especially under-represented after purification, in comparison to RP-CP. Structural analysis of isolated Blm10 Isolated Blm10 was subjected to EM immediately after purification. Individual particles were selected from the micrographs, classified and used to calculate averages (Fig. 4a). We identified several conformations. Circular and spiral Blm10 configurations are assumed to represent Blm10 top views. Their diameters are well matched to the upper surface of CP cylinders (Fig. 4b). These species exhibit central stain accumulation, suggesting a pore-like entry for proteasome substrates, a feature not clearly visible in electron microscopic images of intact Blm10–CP complexes, because only side views could be obtained. A second set of images, designated ‘compact views’ (Fig. 4a), is likely to represent Blm10 side views, because a composite of a representative of this group of Blm10 averages (indicated by a black box) with free CP side views matches closely the structure of the averages we obtained for intact Blm10–CP complexes (Fig. 4c). A bioinformatic approach predicted that Blm10 and its orthologs are composed exclusively of HEAT repeats26. Based on known crystal structures27–30 of HEAT repeat proteins, Blm10 would thus be expected to adopt an extended Figure 5 Blm10 forms a distinct proteasomal subcomplex composed of Blm10–CP–RP in fresh α-solenoid, curved tertiary structure. The struc- unfractionated cell lysates. (a) Freshly prepared yeast lysates from wild type (WT) (BY4741) or a ture of Blm10, as visualized by EM, supports this blm10∆ strain (yMS63) (50 µg) were separated by native gel electrophoresis for 90 min followed by hypothesis and provides general insights into the visualization of proteasomal complexes by an in-gel activity assay with suc-LLVY-AMC. Lane 1 contains affinity purified proteasome holocomplexes as a marker. (b) Identification of the composition of cellular interaction of Blm10 with the CP. Identification of hybrid RP–CP–Blm10 complexes A recent report suggested that Blm10 is involved specifically in CP maturation22. In contrast, our findings indicate that Blm10 can proteasome subcomplexes by tandem mass spectrometry. Affinity-purified proteasome holocomplexes (left lane) and 100 µg of freshly prepared yeast lysates from WT cells (BY4741) (lane 2) were subjected to native gel electrophoresis for 3 h, followed by an in-gel activity assay. Boxed regions were excised and analyzed by tandem mass spectrometry. Proteasomal proteins identified from these regions are shown on the right. (c) Fresh lysate (300 µg) from WT (BY4741) or a blm10∆ strain (yMS63) were subjected to native gel electrophoresis for 4 h. Active proteasome complexes were visualized by an in-gel activity assay in the presence of suc-LLVY-AMC. NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 4 APRIL 2005 297 © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES Figure 6 Effects of BLM10 deletion on cells lacking RPN4 and ECM29. (a) Blm10 and Ecm29 cooperate in maintaining proper proteasome function. Five-fold serial dilutions of cultures of WT (BY4741), blm10∆ (7275), ecm29∆ (yMS4) or the double mutant blm10∆ ecm29∆ (yMS10) were spotted onto YPD plates and incubated either at 30 °C (top), at 37 °C (middle) or onto synthetic medium plates lacking arginine and supplemented with 1.5 µg ml–1 canavanine. In contrast to the results reported here, BLM10 (referred to therein as BLM3) deletion has been reported previously to result in a temperature-sensitive growth defect22. This discrepancy might reflect genetic background differences, because the strain background used in an earlier study22, WCG4, has been described previously to enhance the severity of various proteasome mutants51. (b) Total lysate from WT (BY4741) or a blm10∆ strain (yMS63) (5 µg) was resolved by 10% (w/v) SDS-PAGE and probed for ubiquitin conjugate levels with anti-ubiquitin. (c) Growth of WT (BY4741), rpn4∆ (3716), blm10∆ (yMS63), or rpn4∆ blm10∆ cultures (yMS151) were spotted at 30 °C in YPD was determined by measuring the A660 at the times indicated. (d) Five-fold serial dilutions of exponentially growing WT (BY4741), rpn4∆ (3716) or rpn4∆ blm10∆ cultures (yMS151) were spotted onto YPD or synthetic medium lacking arginine and supplemented with 1.5 µg ml–1 canavanine. (e) To test for proteasome levels, total lysate (5 µg) of WT (BY4741, lane 1), rpn4∆ (lane 2), blm10∆ (lane 3), or rpn4∆ blm10∆ (lane 4) was subjected to SDS-PAGE followed by immunodetection with an anti-Rpn8 (top) or anti-Arc15 as a loading control (bottom). Synthetic interaction of Blm10 with Rpn4 and with Ecm29 The presence of Blm10 in RP–CP complexes suggested a potential function of Blm10 in the degradation of ubiquitinated proteins by the proteasome. However, we did not observe increased sensitivity of blm10∆ mutants toward the arginine analog canavanine (Fig. 6a). Upon incorporation into newly synthesized proteins, canavanine promotes increased protein misfolding and imposes stress on the ubiquitin-proteasome system, as does increased growth temperature. blm10∆ strains did not exhibit an increase in the abundance of ubiquitinated proteins (Fig. 6b), which would be expected if Blm10 contributed substantially to the degradation of ubiquitinated proteasome substrates. These data might argue against a general requirement for Blm10 in the turnover of proteasome substrates and suggest instead a specialized function for Blm10–CP–RP complexes. Alternatively, the cell might compensate for the loss of Blm10 by upregulation of proteasome levels or other means. Eukaryotic cells upregulate proteasome subunit expression if either assembly or activity of the proteasome is impaired31,32. The transcription factor responsible for regulating proteasome subunit expression in yeast is Rpn4, which, as part of a feedback circuit, is itself a proteasome substrate33. Consequently, deletion of RPN4 enhances the phenotypes of proteasome-defective mutants32. A strain with deletions in both RPN4 and BLM10 was viable but grew slowly at 30 °C (Fig. 6c), and showed enhanced canavanine sensitivity in comparison to rpn4∆ mutants (Fig. 6d). The rpn4∆ blm10∆ synthetic phenotype could potentially be explained by Rpn4-mediated upregulation of proteasome synthesis in the blm10∆ mutant. However, deletion of 298 BLM10 is itself not sufficient to upregulate proteasome levels (Fig. 6e). Global analysis of mRNA levels in the blm10∆ mutant also indicated that proteasome genes are not induced (R. Gali, Bauer Center for Genomic Research, Harvard University, and M.S., unpublished data). Although the absence of Blm10 is not deleterious enough to trigger upregulation of the overall amount of proteasomes, the reduced growth rate of strains carrying deletions both in RPN4 and BLM10 and their enhanced sensitivity toward canavanine suggest that Blm10 function becomes crucial if proteasome levels drop below a critical threshold. The synthetic phenotype observed for rpn4∆ blm10∆ suggests that the functions of Blm10 are mediated at least in part through its interaction with the proteasome. To further verify this hypothesis we tested for potential synthetic effects between BLM10 and ECM29, which encodes a protein that is required for maintaining the structural integrity of yeast proteasomal complexes14. The ecm29∆ blm10∆ double mutant showed sensitivity to both high temperature and canavanine, whereas the single-deletion strains were unaffected, suggesting that Blm10 and Ecm29 function to some extent redundantly to promote proteasome activity (Fig. 6a). The C-terminal domain of Blm10 We searched the primary sequence of Blm10 orthologs for conserved regions as potential targets for mutational analysis. The tertiary structure of HEAT repeats allows for a high degree of residue variation34, and the overall sequence conservation in the Blm10 protein family in particular is low, with the human and S. cerevisiae counterparts having only 17% identity. The only region with a higher degree of identity among Blm10 proteins encompasses the 339 C-terminal residues (Supplementary Fig. 4 online). In this region, the identity between human and budding yeast orthologs rises to 24.8% with 36.2% similarity. To characterize this region functionally, we constructed strains carrying a C-terminally deleted Blm10. The truncated protein, expressed from the endogenous BLM10 promoter, was fused to GFP to study its localization (Fig. 7a). Like the proteasome, Blm10 is predominantly localized to the nucleus of yeast cells (Fig. 7b and ref. 22). Notably, deletion of the conserved C terminus led to a redistribution of the protein to the cytoplasm, suggesting that this region contains the nuclear localization signal (NLS) required VOLUME 12 NUMBER 4 APRIL 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY ARTICLES © 2005 Nature Publishing Group http://www.nature.com/nsmb Figure 7 The conserved C terminus of Blm10 is required for nuclear import. (a) Schematic representation of the C-terminal Blm10 truncation generated for this study. (b) Cellular localization of Blm10-GFP (top), Blm10∆C-GFP (middle) and Pre1-GFP (bottom) determined by live-cell fluorescence of strains yMS114, yMS7 and yMS153, respectively. Bar, 5 µm. (c) Deletion of the conserved Blm10 C terminus results in bleomycin sensitivity. Fivefold serial dilutions of exponentially growing WT cells (BY4741), blm10∆ (yMS63) and a C-terminally truncated Blm10 mutant (blm10∆C) (5622) were spotted onto plates. These contained either YPD, YPD plus 5 µg ml–1 bleomycin A2, or synthetic medium lacking arginine and supplemented with 1.5 µg ml–1 canavanine. Plates were incubated at 30 °C for 2–4 d. for nuclear import. Although we could not identify a canonical NLS sequence motif in the S. cerevisiae protein in that region, a strong NLS motif (Supplementary Fig. 4 online) was found within this segment in all Blm10 orthologs. It is also apparent (Fig. 7b) that blm10∆C cells are swollen and rounded—phenotypes not observed for blm10∆ cells (data not shown). To further characterize the truncation, we analyzed the growth of wild-type, blm10∆ and blm10∆C cultures on plates containing either canavanine or bleomycin. As observed for complete deletion, growth on canavanine was not impaired in blm10∆C mutants (Fig. 7c, middle). BLM10 was initially identified as an extragenic suppressor of blm3-1 (ref. 21). blm3-1 was obtained in a genetic screen to detect yeast genes controlling sensitivity to the chemotherapeutic drug bleomycin, which induces DNA double-strand breaks (DSBs)35. Thus, a potential role of Blm10 might be to link proteasomal function to DNA repair mechanisms. In fact, proteasomes are recruited to the sites of DSBs in yeast36 and PA200 colocalizes with nuclear foci that were generated upon γ-irradiation of mammalian cells17. From the latter results, the authors concluded that PA200 is recruited to the sites of DNA repair and involved in repair mechanisms. However, it remains unclear whether proteasome recruitment to DSBs is dependent on Blm10, because the protein might passively localize to DSBs by virtue of its association with the proteasome. Thus, current evidence is only weakly suggestive that Blm10containing proteasomes play a role in DNA repair that is not shared by the general pool of proteasomes. A specific role for Blm10 in DNA repair would be indicated by a hypersensitivity to DNA-damaging agents upon deletion of BLM10, which has not been tested. We find that blm10∆ mutants are sensitive to bleomycin, but to a modest extent (Fig. 7c), and as compared with other proteasome mutants36 blm10∆ cells are not exceptional in the severity of their bleomycin sensitivity. Truncation of the conserved Blm10 C terminus, on the other hand, led to a gain-offunction phenotype of marked bleomycin sensitivity (Fig. 7c, right). As described above, Blm10 nuclear import is impaired upon deletion of its C terminus. A possible scenario explaining the apparent gain of function of a cytoplasmically localized Blm10 truncation on bleomycin sensitivity is that the truncated protein still physically interacts with either the proteasome or other factors involved in DNA repair, thus trapping them in the cytoplasm and attenuating their recruitment to the sites of DSBs. DISCUSSION Activating complexes play a crucial role in the regulation of proteasome activity by inducing the CP gate to open, thus allowing translocation of substrate proteins into the proteolytic chamber. In S. cerevisiae the only previously known proteasome activator is the RP. Here we demonstrate that, similarly to mammalian CP, yeast CP can interact with different activators. We provide structural and biochemical evidence that Blm10, a 240-kDa protein, can associate with mature purified CP and RP–CP complexes, resulting in the enhancement of CP peptidase activity. It has been similarly demonstrated that the PA200 protein can bind to the corresponding human proteasome complexes and enhances the peptidase activity of the CP17. In this and perhaps other respects, Blm10 shows functional conservation from yeast to mammals. The structural analysis in this study demonstrates binding of Blm10 to the outer-ring surface of the core particle cylinder, thus fulfilling the topological requirements for a proteasome activator. In contrast to the PA26 and PA28 activators, which form hepta- or hexameric rings, and also to the RP, which is thought to interact with the CP surface via the six-membered ring formed by the conserved ATPase subunits, Blm10 seems to bind proteasomes as a single polypeptide chain. To gain insight into the structural properties of this protein we dissociated Blm10 from the CP and obtained electron microscopic images of isolated Blm10. Blm10 has an elongated, flexible, curved tertiary structure. A variety of distinct conformations were observed that may reflect a limited flexibility of the molecule upon dissociation from the CP. The shape that Blm10 adopts upon interaction with CP is most likely a circular or spiral structure. This particular conformation exhibits a central stain accumulation, suggesting that Blm10 might form a channel that is continuous with the central pore leading into the proteolytic cavity of the CP, thus allowing substrate entry and exit. The electron microscopic analysis of Blm10 agrees well with a recent bioinformatic approach to predict the tertiary structure of Blm10 orthologs26. This study identified regions in PA200 with high homology to PR65/A26. The crystal structure of PR65/A has recently been solved27, revealing a tertiary structure built exclusively of 15 tandem HEAT repeats. Precise definition of the repeat boundaries poses a major challenge in the prediction of repeat structures, because these residues exhibit a high degree of variation and degeneration37. By using structure-based alignments, the bioinformatic study defined the individual repeat boundaries and identified 18 imperfect HEAT repeats in PA200, spanning the entire molecule26. This approach revealed that Blm10 and its orthologs are composed of a distinct subgroup of HEAT repeats, designated HEAT-PA200 (ref. 26). Few structures of proteins with multiple structural repeats have been solved so far, because their elongated, flexible tertiary structures and high molecular masses hamper crystallization and NMR studies38, thus favoring structural characterization by EM. The general architecture of HEAT repeats, as suggested by the PR65/A structure, is a consecutive repetition of a module containing two NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 4 APRIL 2005 299 © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES antiparallel α-helices connected by turns. Table 1 Strains used in this study Individual repeats are arranged along a Strain Genotype common axis to form an open, superheli- BY4741 MATa his3D1 leu2D0 met15D0 ura3D0 cal, coiled-coil conformation, also termed 933 MATa his3D1 leu2D0 met15D0 ura3D0 ecm29∆:Kan α-solenoid tertiary structure. This particular 7275 MATa his3D1 leu2D0 met15D0 ura3D0 blm10s∆:Kan open conformation gives rise to an extensive 5622 MATa his3D1 leu2D0 met15D0 ura3D0 blm10∆C:Kan solvent-exposed surface, enabling the bind3716 MATa his3D1 leu2D0 met15D0 ura3D0 rpn4∆:Kan ing of large interaction partners such as proyMS4 MATa his3D1 leu2D0 met15D0 ura3D0 ecm29∆:URA3 37 tein complexes and polynucleotides . With yMS6 MATα his3D1 leu2D0 lys2D0 ura3D0 RPN11 tevProA (HIS3) respect to Blm10, this structural feature MATa his3D1 leu2D0 met15D0 ura3D0 BLM10∆C-GFP:Kan most likely underlies its ability to interact yMS7 yMS10 MATa his3D1 leu2D0 met15D0 ura3D0 blm10s∆:Kan ecm29∆:URA3 with the CP cylinder. MATα his3D1 leu2D0 met15D0 ura3D0 blm10s∆:Kan The majority of proteasomal substrates in yMS14 the cell are ubiquitinated proteins. RP func- yMS15 MATa his3D1 leu2D0 met15D0 ura3D0 RPN11 tevProA (HIS3) ecm29∆:Kan tion is essential for recognition, binding and yMS31 MATa his3D1 leu2D0 met15D0 ura3D0 PRE1 tevProA (HIS3) unfolding of such substrates. The question of yMS32 MATα his3D1 leu2D0 ura3D0 PRE1 tevProA (HIS3) whether Blm10 functions in physical associa- yMS46 MATa his3D1 leu2D0 met15D0 ura3D0 ecm29∆:Kan PRE1 tevProA (HIS3) tion with the RP is important because hybrid yMS63 MATa his3D1 leu2D0 met15D0 ura3D0 blm10l∆:Nat complexes composed of Blm10–CP–RP yMS92 MATα his3D1 leu2D0 met15D0 ura3D0 blm10∆:Nat might link Blm10-induced CP activation to yMS94 MATa his3D1 leu2D0 met15D0 ura3D0 PRE1 tevProA (HIS3) blm10:Nat the degradation of the major group of proyMS114 MATa his3D1 leu2D0 met15D0 ura3D0 BLM10-GFP: Kan teasome substrates. Similar hybrid complexes yMS122 MATa his3D1 leu2D0 met15D0 ura3D0 PRE1tevProA (HIS3) GalpHA3Blm10(Kan) have been identified for PA28 (refs. 24,25). To yMS151 MATa his3D1 leu2D0 lys2D0 ura3D0 blm10∆:Nat rpn4∆:Kan identify Blm10-containing proteasome comyMS153 MATa his3D1 leu2D0 met15D0 ura3D0 PRE1-GFP: Kan plexes in the cell, we established a method for The strains BY4741, 933, 7275, 5622 and 3716 were obtained from the Research Genetics yeast gene deletion the detection of proteasomal complexes in collection45. All other strains have been generated for this study. yMS10 was derived from crossing 7275 and yMS4; freshly lysed, unfractionated cell extracts, fol- yMS15 from crossing yMS6 and 933; yMS46 from crossing 933 and yMS32; yMS94 from crossing yMS31 and lowed by the identification of complex com- yMS92, and yMS151 from crossing 3716 and yMS92. All other strains were generated by genomic integration. Strain , which has been demonstrated to be a partial deletion of BLM10 (known ponents by tandem mass spectrometry. We 7275 contained a deletion of ORF YFL007w as BLM3 at the time of the correction52), owing to a sequencing error in the yeast genome database. The corrected provide evidence that the bulk of the cellular ORF YFL007w contains both previous designated genes, YFL006w and YFL007w51. To distinguish between the fullpool of proteasome-bound Blm10 exists in a length deletion and the partial deletion, the latter is referred to in this study as blm10s∆. Strain 5622 is a deletion of hybrid Blm10–CP–RP complex, in which one YFL006W and as such is a partial C-terminal deletion of Blm10 and is therefore designated here as blm10∆C:Kan. end of the CP cylinder is capped with RP and the other with Blm10. Blm10–CP and free CP complexes are not readily detected in fresh yeast lysates and are generAlthough the presence of Blm10 in RP–CP complexes suggests ated through the disassembly of Blm10–CP–RP complexes even dur- a potential function for Blm10 in the degradation of ubiquitinated ing rapid and mild affinity purification. Although the importance of substrates, we did not detect any changes in the level of ubiquitinated loosely associated protein complexes in the cell is well recognized, few conjugates in BLM10-deleted cells, nor do these cells show increased experimental approaches are suited to define such labile complexes. sensitivity to canavanine. Several alternative explanations might account Native gel electrophoresis, as applied here, may be of general use in for these observations: Blm10–CP–RP complexes might mediate a very specific function, unrelated to the degradation of ubiquitinated proteins; clarifying the in vivo assembly states of such complexes. Blm10 (referred to as Blm3 in ref. 22) was previously identified in they might have activity toward only a specific subset of ubiquitinated association with late-stage CP assembly intermediates, but was not substrates; or the effects of the BLM10 deletion might be masked by detected in complexes with mature CP or RP-CP22. Late-stage CP compensating processes in the cell. In general, impaired proteasome assembly intermediates are fully assembled core particles, which still function is counteracted in cells by increased expression of proteasomal require partial autocatalytic processing of the active subunits39. In this genes31,32, a process regulated by Rpn4 in yeast. Such an upregulation of report a function for Blm10 in proteasome maturation was suggested. proteasome levels, however, was not detected in BLM10-deleted strains. The nature of a maturation factor, like Ump1, is generally to transiently Nonetheless, the rpn4 blm10 double-deletion strain showed a growth associate with a nascent form of a complex. In the cellular context, defect at 30 °C and strong canavanine sensitivity. We conclude from however, Blm10 is found in association with mature active RP–CP these results that, although the function of Blm10 is not essential for complexes. Furthermore, if an important assembly factor is missing, normal growth, it becomes critical if the proper regulation of proteaRpn4-dependent upregulation of proteasome synthesis is expected32, some levels in the cell is impaired. That Blm10 is a positive regulator which is not observed in blm10 mutants. It is unclear at the moment of proteasome function is also implied by the synthetic phenotype of why a previous study22 did not detect Blm10 in fractions containing double mutants between BLM10 and ECM29, which encodes a proteamature proteasomes. One possible reason might be the purification some-associated protein. In summary, a simple loss-of function mutation in BLM10 exhibits procedures used. As demonstrated here, Blm10–CP–RP complexes are labile and dissociate even under mild and rapid purification condi- canonical phenotypes of proteasome functional deficiency when in a tions as used in this study. Another possibility is that the antibody, genetic background conferring marginal impairment of proteasome raised against the N terminus of Blm10, and used to identify Blm10 in function, such as rpn4∆ or ecm29∆. In contrast, frameshift mutations near association with nascent CP, recognizes a segment of the protein that the 3′ end of BLM10, such as blm10∆C, result in a gain-of-function phenotype of bleomycin hypersensitivity accompanied by loss of the most is subject to cleavage. 300 VOLUME 12 NUMBER 4 APRIL 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES conserved sequences on the protein and gross mislocalization. Thus, there seem to be two distinct aspects to the Blm10 phenotype. Our data indicate that Blm10 functions in vivo as a component of the proteasome; whether the bleomycin sensitivity exhibited by BLM10 gain-of-function mutations is linked to its interaction with the proteasome remains to be resolved. Proteasome activators, such as PA28, and proteasome inhibitors, such as PI31, have been studied intensively for years, yet little of how they regulate proteasomes is understood40,41. Because the presence of PA28 increases the efficiency of cell-surface presentation of several peptides42,43, this activator is believed to be important for MHC class I antigen presentation. However the role of the PA28 family does not seem to be restricted to antigen presentation. For example, PA28 is found in some organisms, such as trypanosomes, that do not have a class I cell-surface presentation pathway18. Also, mice lacking PA28γ have the unrelated phenotype of small size44. Because Blm10 is so widely conserved in evolution, and found in one of the most experimentally accessible organisms, budding yeast, it presents a unique opportunity to dissect the function of a proteasome activator in detail. METHODS Strains, strain construction and culture. All strains and primers used in this work are listed in Table 1 and Supplementary Table 1 online, respectively. They were obtained by standard genetic techniques. All strains are isogenic to BY4741 and are S288C-derived. BY4741, BY4742, 933, 7275, 5622 and 3716 were obtained from Research Genetics45. The primers used for a marker exchange of strain 933 to create yMS4 were PTEF1 and TTEF1 (Supplementary Table 1 online). Primers and templates for the genomic integration of the ProA tag in Rpn11 and Pre1 were as described14 and were used to construct yMS6, yMS31 and yMS32 (see Supplementary Table 1 online). Complete BLM10 gene deletion for constructing yMS63 was achieved by gene replacement with a natMX3 cassette by homologous recombination46. The following primers were used: KOBclonf and KOBclnor. The template for marker integration was pAG25 (ref. 46). For the integration of a C-terminal GFP tag, yEGFP and a kanMX4 cassette were either fused in frame to the C terminus of the target open reading frame (ORF) replacing the stop codon by homologous recombination or fused after residue 1804 to delete the C-terminal 400 amino acids. To amplify the integration construct, template pYM12 (ref. 47) was used with the following primer pairs: Blm10fexCtag and Blm10rexCtag (Blm10-GFP) to construct strain yMS114; Blm10tagf and Blm10tagr (Blm10∆C-GFP) to construct strain yMS7; and Pre1CGFPtagF and Pre1CGFPtagR (for Pre1-GFP) to construct yMS153. yMS122 was obtained by genomic integration of a construct carrying a triple HA tag, replacing the start codon of Blm10 preceded by the GAL1 promoter in yMS31. The primers used for obtaining the integration construct were Blm3Ntagf and Blm3NHAr with pFA6a-3HA-kanMX6 (ref. 48) as a template. All strains used for proteasome purifications were grown at 30 °C in YPD to early stationary phase unless otherwise noted. Proteasome purification. Conventional proteasome purification was carried out as described23. The initial purification of Blm10–CP complexes was done by affinity purification via a ProA tag integrated at the C terminus of Rpn11 in an ecm29∆ background (yMS15). Cells (100 g) were harvested and lysed by using a French press in 200 ml lysis buffer A (50 mM Tris-HCl, pH 8, 5 mM MgCl2, 1 mM ATP, 0.5 mM EDTA). After the cell debris was removed by centrifugation the lysate was incubated with 4 ml IgG resin (ICN Kappel) for 1 h at 4 °C. The resin was washed with 100 bed volumes of lysis buffer plus 100 mM NaCl. Subsequently, to destabilize the RP-CP interaction, ATP was removed by washing the column with five bed volumes of buffer A (50 mM Tris-HCl, pH 7.6, 100 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA), followed by a 30-min incubation of the resin in five bed volumes of buffer B (50 mM Tris-HCl, pH 7.6, 300 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA) to elute the core particle. The effectiveness of the CP elution was tested by an LLVY activity assay as described49. The eluate was concentrated to 0.5 ml and subjected to size-exclusion chromatography using a Superose 6 column with 24-ml bed volumes (Pharmacia), pre-equilibrated with buffer A. Peak fractions were applied to a 10% (w/v) SDS-polyacrylamide gel, followed by Coomassie blue staining and native gel analysis (see below). Because Blm10 was found to interact with the core particle, all subsequent Blm10-CP purifications were done using a ProA-tagged Pre1 strain. To obtain CP from wild-type and blm10∆ strains, 100 g of cells from yMS31 and yMS94 were lysed in 45 ml of lysis buffer B (50 mM Tris-HCl, pH 8, 5 mM MgCl2, 0.5 mM EDTA) by French press cell disruption. The cell extracts were cleared by centrifugation and incubated with 4 ml of IgG resin for 1 h at 4 °C. The resin was washed with 400 ml of buffer A. Subsequently the resin was incubated with 50 ml of buffer B for 30 min to remove RP. Salt was removed from the column by washing with buffer C (50 mM Tris-HCl, pH 7.6, 5 mM MgCl2, 0.5 mM EDTA). To elute the core particle the resin was incubated with 100 U 6× HisTEV (Gibco) per ml resin for 2 h at RT and eluted with five column volumes of buffer C, concentrated, and subjected to SDS-PAGE and native gel analysis. To obtain a higher yield of Blm10–CP complexes, yMS122, a strain where the endogenous BLM10 promoter was replaced by a GAL1 promoter, was grown in YPGal to early stationary phase. Cells (100 g) were harvested and subjected to the same purification procedure as described for CP purification utilizing a tagged Pre1. Because, however, affinity-purified complexes were to be further separated on a size-exclusion column, the stringent washing of the IgG resin with 300 mM NaCl was omitted. To assess CP activation by Blm10, 0.5 µg of the fractions were tested for peptidase activity with 100 µM of the fluorogenic substrate Suc-LLVY-AMC in 20 mM Na-phosphate, pH 7.4, 50 mM NaCl by monitoring the generation of free AMC at 25 °C in a Cary Eclipse fluorescence spectrophotometer. Blm10 purification and reconstitution of complex. Blm10 was purified from intact Blm10–CP complexes via the Pre1-ProA tag. The Blm10-CP interaction was destabilized by high salt followed by elution of the dissociated Blm10 with subsequent size-exclusion chromatography to obtain the pure protein. To minimize copurification of RP, an ecm29∆ strain (yMS46) was used. This strain (100 g) was harvested and lysed in 150 ml of lysis buffer B by French press cell disruption. After removal of cell debris by centrifugation the lysate was incubated with 4 ml of IgG resin for 2 h at 4 °C, followed by washing with 100 bed volumes of 50 mM Tris-HCl, pH 7.6, 50 mM NaCl. Dissociation of Blm10 was achieved by incubation of the resin in 50 mM Tris-HCl, pH 7.6, 500 mM NaCl, for 3 h at 22 °C. The elution of Blm10 was monitored by SDS-PAGE. The eluate was immediately diluted with 50 mM TrisHCl, pH 7.6, concentrated to 500 µl, and subjected to size-exclusion chromatography on a Superose 6 column equilibrated with 50 mM Tris, pH 7.6, 50 mM NaCl, 0.5 mM EDTA. Blm10-containing fractions were concentrated and quantified using a Bradford protein assay (Bio-Rad). CP still bound to the resin was eluted by TEV cleavage as described above, concentrated and subjected to sizeexclusion chromatography. For reconstitution of Blm10–CP complexes, 680 nM Blm10, isolated from yMS31, and 250 nM CP, isolated from a yMS94 as described above, were combined and incubated for 30 min at room temperature in buffer C in a final volume of 20 µl. Formation of complex was monitored by subjecting the sample to native gel separation followed by an in-gel activity assay. All experiments involving pure Blm10 were done in rapid succession, because the protein was found to be highly unstable upon dissociation from the proteasome. Electrophoresis and immunoblotting. SDS-PAGE and immunoblotting was carried out as described14. The anti-ubiquitin and the anti-α7 were obtained from Affiniti, the anti-HA probe (12CA5) was obtained from Roche, and the antibody against Arc15 was a gift from R. Li (Harvard Medical School). Purified proteasomal complexes were applied to native gels composed of 3.5% (w/v) acrylamide, 90 mM Tris-base, 90 mM boric acid, 1 mM ATP, 5 mM MgCl2, 0.5 mM EDTA, which were run in the same buffer omitting the acrylamide. The gels were run at 100 V for 2 h at 10 °C, incubated for 30 min at 30 °C in substrate buffer (50 mM Tris-HCl, pH 7.4, 5 mM MgCl2, 1 mM ATP, 0.05 µM Suc-LLVYAMC (Bachem)) for the in-gel activity assay unless otherwise indicated. Active bands were visualized using an ultraviolet screen at 340 nm. Subsequently the gels were stained with Coomassie blue or silver nitrate. To analyze proteasomal complexes in cell lysates, harvested cells were resuspended in lysis buffer B at a ratio of 1.5 ml g–1 wet cell mass. Cell disruption was carried out with a French press. After clarification of the lysate by centrifugation, protein content was assessed by a Bradford assay and the sample was subjected to native gel electrophoresis and a subsequent in-gel activity assay as described NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 4 APRIL 2005 301 ARTICLES © 2005 Nature Publishing Group http://www.nature.com/nsmb above. For mass spectrometry experiments the active bands were excised and the proteins were denatured in 40% (v/v) methanol, 10% (v/v) glacial acetic acid overnight at 22 °C. EM and image processing. Sample solution (5 µl) was applied to a glowdischarged carbon-coated grid and negatively stained with uranyl formate as described50. Samples were inspected with a FEI Philips T12 transmission EM operated at 120 kV. Images were taken at a nominal magnification of 52,000× and a defocus of –1.5 µm using low-dose procedures. Images were digitized with a Zeiss SCAI scanner using a step size of 7 µm. For images of the CP–Blm10 complex, 5 × 5 pixels were averaged, yielding a pixel size of 6.7 Å on the specimen level, and 3 × 3 pixels were averaged for images of isolated Blm10, yielding a pixel size of 4 Å. All image processing steps were done using the SPIDER suite49. Approximately 3,000 particles for CP–Blm10 complexes and 8,000 particles for isolated Blm10 were selected from the digitized electron micrographs and windowed into 80 × 80 pixel images. The images were subjected to ten rounds of alignment, multivariate statistical analysis, classification, and image averaging to generate the averaged projection structures shown in Figures 3 and 4. Fluorescence microscopy. Strains yMS7, yMS114 and yMS153 carrying C-terminally GFP-tagged Pre1, Blm10∆C or Blm10 were grown overnight in synthetic complete media at 30 °C. Cultures were then diluted back to an A660 of 0.1 and further incubated at 30 °C for 4 h to obtain exponentially growing cultures. Cells were harvested and applied to glass slides. GFP fluorescence images were collected using a Nikon Eclipse1000 upright automated microscope equipped with a Hamamatsu Orca II Dual Mode cooled CCD camera and a MetaMorph data acquisition workstation. Phenotypic analysis of gene deletion. To analyze the phenotypes of gene deletions, overnight cultures were diluted back to an A660 of 0.1 and further incubated for 4 h at 30 °C to obtain exponentially growing cultures. The A660 was measured and the cultures were diluted in 96-well plates with YPD to a density of 6 × 106 cells per well, followed by five-fold serial dilutions and spotted onto YPD plates in the absence or presence of 5 µg ml–1 bleomycin A2 (Calbiochem) or onto synthetic arginine drop out plates supplemented with 1.5 µg ml–1 canavanine. Note: Supplementary information is available on the Nature Structural & Molecular Biology website. ACKNOWLEDGMENTS This work was supported by a grant from the Deutsche Forschungsgemeinschaft to M.S. (SCHM 884/3-1), and US National Institutes of Health (NIH) grants to D.F. (GM65592) and S.G. (GM67945). We thank B. Petre for collecting the EM images. The molecular EM facility at Harvard Medical School was established by a generous donation from the Giovanni Armenise Harvard Center for Structural Biology and is maintained by funds from NIH grant GM62580 to T.W. We are also grateful to the Harvard Medical School Nikon Imaging Center for technical assistance and access to their instruments, to R. Li for providing us with an antibody against Arc15, and to S. Elsasser, J. Hanna and J. Roelofs for critical reading of the manuscript. We thank R. Gali, and the Bauer Center for Genomics Research, Harvard University, for help with preliminary results of microarray analysis on the Rosetta Resolver microarray analysis platform. We furthermore are grateful to C. Hill for communicating results prior to publication and to A.L. Goldberg for his permission to include PA28-20S electron micrographs in our study. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Received 14 October 2004; accepted 10 February 2005 Published online at http://www.nature.com/nsmb/ 1. Pickart, C.M. & Cohen, R.E. Proteasomes and their kin: proteases in the machine age. Nat. Rev. Mol. Cell Biol. 5, 177–187 (2004). 2. Groll, M. et al. Structure of 20S proteasome from yeast at 2.4 Å resolution. Nature 386, 463–471 (1997). 3. Groll, M. et al. A gated channel into the proteasome core particle. Nat. Struct. Biol. 7, 1062–1067 (2000). 4. Braun, B.C. et al. The base of the proteasome regulatory particle exhibits chaperone-like activity. Nat. Cell Biol. 1, 221–226 (1999). 5. Liu, C.W. et al. Conformational remodeling of proteasomal substrates by PA700, the 19 S regulatory complex of the 26 S proteasome. J. Biol. Chem. 277, 26815–26820 (2002). 302 6. Schmidt, M., Lupas, A.N. & Finley, D. Structure and mechanism of ATP-dependent proteases. Curr. Opin. Chem. Biol. 3, 584–591 (1999). 7. Hartmann-Petersen, R., Tanaka, K. & Hendil, K.B. Quaternary structure of the ATPase complex of human 26S proteasomes determined by chemical cross-linking. Arch. Biochem. Biophys. 386, 89–94 (2001). 8. Prakash, S., Tian, L., Ratliff, K.S., Lehotzky, R.E. & Matouschek, A. An unstructured initiation site is required for efficient proteasome-mediated degradation. Nat. Struct. Mol. Biol. 11, 830–837 (2004). 9. Navon, A. & Goldberg, A.L. Proteins are unfolded on the surface of the ATPase ring before transport into the proteasome. Mol. Cell 8, 1339–1349 (2001). 10. Lam, Y.A., Lawson, T.G., Velayutham, M., Zweier, J.L. & Pickart, C.M. A proteasomal ATPase subunit recognizes the polyubiquitin degradation signal. Nature 416, 763–767 (2002). 11. Elsasser, S., Chandler-Militello, D., Mueller, B., Hanna, J. & Finley, D. Rad23 and Rpn10 serve as alternative ubiquitin receptors for the proteasome. J. Biol. Chem. 279, 26817–26822 (2004). 12. Verma, R., Oania, R., Graumann, J. & Deshaies, R.J. Multiubiquitin chain receptors define a layer of substrate selectivity in the ubiquitin-proteasome system. Cell 118, 99–110 (2004). 13. Yao, T. & Cohen, R.E. A cryptic protease couples deubiquitination and degradation by the proteasome. Nature 419, 403–407 (2002). 14. Leggett, D.S. et al. Multiple associated proteins regulate proteasome structure and function. Mol. Cell 10, 495–507 (2002). 15. Guterman, A. & Glickman, M.H. Complementary roles for Rpn11 and Ubp6 in deubiquitination and proteolysis by the proteasome. J. Biol. Chem. 279, 1729–1738 (2004). 16. Wojcik, C., Tanaka, K., Paweletz, N., Naab, U. & Wilk, S. Proteasome activator (PA28) subunits, α, β and γ (Ki antigen) in NT2 neuronal precursor cells and HeLa S3 cells. Eur. J. Cell Biol. 77, 151–160 (1998). 17. Ustrell, V., Hoffman, L., Pratt, G. & Rechsteiner, M. PA200, a nuclear proteasome activator involved in DNA repair. EMBO J. 21, 3516–3525 (2002). 18. Whitby, F.G. et al. Structural basis for the activation of 20S proteasomes by 11S regulators. Nature 408, 115–120 (2000). 19. Forster, A., Whitby, F.G. & Hill, C.P. The pore of activated 20S proteasomes has an ordered 7-fold symmetric conformation. EMBO J. 22, 4356–4364 (2003). 20. Tanahashi, N. et al. Hybrid proteasomes. Induction by interferon-γ and contribution to ATP-dependent proteolysis. J. Biol. Chem. 275, 14336–14345 (2000). 21. Doherty, K., Pramanik, A., Pride, L., Lukose, J. & Wood Moore, C. Expression of the expanded YFL007w ORF and assignment of the gene name BLM10. Yeast 21, 1021 (2004). 22. Fehlker, M., Wendler, P., Lehmann, A. & Enenkel, C. Blm3 is part of nascent proteasomes and is involved in a late stage of nuclear proteasome assembly. EMBO Rep. 4, 959–963 (2003). 23. Glickman, M.H., Rubin, D.M., Fried, V.A. & Finley, D. The regulatory particle of the Saccharomyces cerevisiae proteasome. Mol. Cell. Biol. 18, 3149–3162 (1998). 24. Kopp, F., Dahlmann, B. & Kuehn, L. Reconstitution of hybrid proteasomes from purified PA700-20 S complexes and PA28αβ activator: ultrastructure and peptidase activities. J. Mol. Biol. 313, 465–471 (2001). 25. Cascio, P., Call, M., Petre, B.M., Walz, T. & Goldberg, A.L. Properties of the hybrid form of the 26S proteasome containing both 19S and PA28 complexes. EMBO J. 21, 2636–2645 (2002). 26. Kajava, A.V., Gorbea, C., Ortega, J., Rechsteiner, M. & Steven, A.C. New HEAT-like repeat motifs in proteins regulating proteasome structure and function. J. Struct. Biol. 146, 425–430 (2004). 27. Groves, M.R., Hanlon, N., Turowski, P., Hemmings, B.A. & Barford, D. The structure of the protein phosphatase 2A PR65/A subunit reveals the conformation of its 15 tandemly repeated HEAT motifs. Cell 96, 99–110 (1999). 28. Chook, Y.M. & Blobel, G. Structure of the nuclear transport complex karyopherin-β2Ran t • GppNHp. Nature 399, 230–237 (1999). 29. Vetter, I.R., Arndt, A., Kutay, U., Gorlich, D. & Wittinghofer, A. Structural view of the Ran-importin β interaction at 2.3 Å resolution. Cell 97, 635–646 (1999). 30. Goldenberg, S.J. et al. Structure of the Cand1-Cul1-Roc1 complex reveals regulatory mechanisms for the assembly of the multisubunit cullin-dependent ubiquitin ligases. Cell 119, 517–528 (2004). 31. Meiners, S. et al. Inhibition of proteasome activity induces concerted expression of proteasome genes and de novo formation of mammalian proteasomes. J. Biol. Chem. 278, 21517–21525 (2003). 32. London, M.K., Keck, B.I., Ramos, P.C. & Jurgen Dohmen, R. Regulatory mechanisms controlling biogenesis of ubiquitin and the proteasome. FEBS Lett. 567, 259–264 (2004). 33. Xie, Y. & Varshavsky, A. RPN4 is a ligand, substrate, and transcriptional regulator of the 26S proteasome: a negative feedback circuit. Proc. Natl. Acad. Sci. USA 98, 3056–3061 (2001). 34. Andrade, M.A., Petosa, C., O’Donoghue, S.I., Muller, C.W. & Bork, P. Comparison of ARM and HEAT protein repeats. J. Mol. Biol. 309, 1–18 (2001). 35. Moore, C.W. Further characterizations of bleomycin-sensitive (blm) mutants of Saccharomyces cerevisiae with implications for a radiomimetic model. J. Bacteriol. 173, 3605–3608 (1991). 36. Krogan, N.J. et al. Proteasome involvement in the repair of DNA double-strand breaks. Mol. Cell 16, 1027–1034 (2004). 37. Andrade, M.A., Perez-Iratxeta, C. & Ponting, C.P. Protein repeats: structures, functions, and evolution. J. Struct. Biol. 134, 117–131 (2001). 38. Kajava, A.V. Review: proteins with repeated sequence—structural prediction and modeling. J. Struct. Biol. 134, 132–144 (2001). VOLUME 12 NUMBER 4 APRIL 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY © 2005 Nature Publishing Group http://www.nature.com/nsmb ARTICLES 39. Chen, P. & Hochstrasser, M. Autocatalytic subunit processing couples active site formation in the 20S proteasome to completion of assembly. Cell 86, 961–972 (1996). 40. Forster, A. & Hill, C.P. Proteasome degradation: enter the substrate. Trends Cell Biol. 13, 550–553 (2003). 41. McCutchen-Maloney, S.L. et al. cDNA cloning, expression, and functional characterization of PI31, a proline-rich inhibitor of the proteasome. J. Biol. Chem. 275, 18557–18565 (2000). 42. Groettrup, M. et al. A role for the proteasome regulator PA28α in antigen presentation. Nature 381, 166–168 (1996). 43. Sijts, A. et al. The role of the proteasome activator PA28 in MHC class I antigen processing. Mol. Immunol. 39, 165–169 (2002). 44. Murata, S. et al. Growth retardation in mice lacking the proteasome activator PA28γ. J. Biol. Chem. 274, 38211–38215 (1999). 45. Brachmann, C.B. et al. Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 14, 115–132 (1998). 46. Goldstein, A.L. & McCusker, J.H. Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15, 1541–1553 (1999). 47. Knop, M. et al. Epitope tagging of yeast genes using a PCR-based strategy: more tags and improved practical routines. Yeast 15, 963–972 (1999). 48. Longtine, M.S. et al. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 14, 953–961 (1998). 49. Frank, J. et al. SPIDER and WEB: processing and visualization of images in 3D electron microscopy and related fields. J. Struct. Biol. 116, 190–199 (1996). 50. Ohi, M., Li, Y., Cheng, Y. & Walz, T. Negative staining and image classification—powerful tools in modern electron microscopy. Biol. Proceed. Online 6, 23–34 (2004). 51. Gerlinger, U.M., Guckel, R., Hoffmann, M., Wolf, D.H. & Hilt, W. Yeast cycloheximideresistant crl mutants are proteasome mutants defective in protein degradation. Mol. Biol. Cell 8, 2487–2499 (1997). 52. Robben, J., Hertveldt, K., & Volckaert, G. Revisiting the yeast chromosome VI DNA sequence reveals a correction merging YFL007w and YFL006w to a single ORF. Yeast 19, 699–702 (2002). NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 4 APRIL 2005 View publication stats 303