Proteasome Activator 200: The HEAT is on…

Abstract

Proteasomes play a key regulatory role in all eukaryotic cells by removing proteins in a timely manner. There are two predominant forms: The 20S core particle (CP) can hydrolyze peptides and certain unstructured proteins, and the 26S holoenzyme is able to proteolyse most proteins conjugated to ubiquitin. The 26S complex consists of a CP barrel with a 19S regulatory particle (RP; a.k.a PA700) attached to its outer surface. Several studies purified another proteasome activator with a MW of 200 kDa (PA200) that attaches to the same outer ring of the CP. A role for PA200 has been demonstrated in spermatogenesis, in response to DNA repair and in maintenance of mitochondrial inheritance. Enhanced levels of PA200-CP complexes are observed under conditions in which either activated or disrupted CP prevail, suggesting it participates in regulating overall proteolytic activity. PA200, or its yeast ortholog Blm10, may also incorporate into 26S proteasomes yielding PA200-CP-RP hybrids. A three-dimensional molecular structure determined by x-ray crystallography of Blm10-CP provides a model for activation. The carboxy terminus of Blm10 inserts into a dedicated pocket in the outer ring of the CP surface, whereas multiple HEAT-like repeats fold into an asymmetric solenoid wrapping around the central pore to stabilize a partially open conformation. The resulting hollow domelike structure caps the entire CP surface. This asymmetric structure may provide insight as to how the 19S RP, with two HEAT repeatlike subunits (Rpn1, Rpn2) alongside six ATPases (Rpt1–6), attaches to the same surface of the CP ring, and likewise, induces pore opening.

The proteasome is the regulatory protease for removal of cellular proteins selected by the ubiquitin (Ub)¹ system (1). At the heart of the proteasome structure lies the barrel-shaped 20S core particle (CP) into which substrates must enter if they are to be proteolysed (2). Substrates access the central catalytic chamber through narrow pores defined by the outer α-rings of the 20S CP (3). In the absence of activators, these channels are closed by the N termini of the seven different α subunits that forge these rings, and proteolysis rates are repressed (4, 5). Disordering of the blocking residues or opening the channel accelerates entry of peptides, however in order to degrade globular or ubiquitinated substrates, attachment of proteasome activators (PAs) to the outer surface of the 20S CP is required (6–9). The Ub-binding, ATPase-containing, 19S regulatory particle (RP) is the best studied such activator (a.k.a “proteasome activator 700”; PA700) and is required for regulated proteolysis (Fig. 1). Attachment of the 19S RP to the outer surface of the α-ring induces an open pore, thereby enhancing the basal peptidase activity of the proteasome (2, 5, 10). Ubiquitin-binding subunits in the 19S RP anchor the tagged substrate whereas six ATPases (Rpt1–6) unfold and translocate it through the channel into the 20S CP for proteolysis. This RP-CP holoenzyme, termed the 26S proteasome, is thought to be the most abundant species in rapidly growing cells, although it is found in equilibrium with free 20S CP (11).

Fig. 1.

PA200/Blm10 association with different proteasome species, and related roles. Because most evidence for proteasome complexes with PA200 was obtained in yeast, we use the yeast nomenclature, Blm10, in this figure: A, Biological functions of Blm10. Blm10 is involved in different cellular functions such as enhancing hydrolysis of small peptides (21–23), maintenance of mitochondrial functions (24), response to DNA repair (21, 34), and in proteasome maturation and assembly (39–43). Whether it has an inhibitory or activating role in these processes is still a matter of debate. B, Blm10 and proteasome assembly. PA200/Blm10 was detected in inactive proteasome assembly precursors (39, 40), as well as in complexes with immature or nascent 20S CP (30). Following dimerization of half proteasomes (1/2 CP; a.k.a 15S proteasomes) the β-subunits are processed to yield the active mature form of the 20S CP. These active, yet latent species can be activated upon binding of proteasome activators such as the 19S RP (PA700) or by attachment of non-ATPase activators such as PA200/Blm10. Complexes of PA200 with activated 20S CP were found as symmetric or asymmetric versions (Blm10-CP and (Blm10)₂-CP) (22, 35, 41). These species were labeled CPx in (22)). PA200 is also found as an opposing appendage in hybrid complexes in the form of Blm10-CP-19S (22, 41).

In vitro, many unfolded or loosely structured proteins are substrates for proteolysis by the isolated 20S CP, however it is unclear whether in cells, free 20S CP is biologically active as a regulatory protease on its own or whether it serves primarily as a reserve of core particles for rapid assembly with activators as needed. Nevertheless, there are increasing reports of 20S CP capable of proteolysis without requirement for ATP (12–16). Several non-ATPase activators have been isolated in complex with the 20S CP and may promote ATP-independent turn-over of certain loosely folded substrates. These include the heptameric 11S regulators (11S-Regs) that attach to the outer α-ring of the 20S barrel and activate proteolysis by enforcing channel gating (5, 7, 17). 11S-Regs form a hollow cone-shaped ring of seven 28-kDa subunits. The family includes the heteromeric interferon-γ induced PA28α/β involved in immune response and antigen presentation, the nuclear PA28γ, and a distantly related complex unique to trypanosomes termed PA26 (18–20). Another ATP-independent proteasome activator is the 250-kDa monomeric PA200 (Blm10 in yeast) that competes for the same binding sites on the 20S α-ring surface (21–24). Non-ATPase activators probably do not promote general proteolysis of intact globular proteins (9, 25). However, degradation of specific, loosely folded substrates aided by PA28/11S-Reg has been documented and may reflect a more general feature of non-ATPase activators (26–28). Attachment of a single copy of PA200 (or Blm10) to 20S CP to generate the asymmetric PA200-CP (Fig. 1) is correlated with enhanced hydrolysis of small peptides, hence its name “proteasome activator 200” (PA200) (21–23).

The Biology of PA200

Homologs of PA200 are found in many eukaryotes, though in some (fruit flies and fission yeast for example) it may have diverged beyond recognition or replaced by another non-ATPase activator. A large portion of our knowledge comes from the budding yeast model, where the gene encoding for the PA200 ortholog is termed BLM10 (YFL007w; previously known as BLM3). In this review we will use the term PA200 in reference to general properties of this protein regardless of origin, and revert to Blm10 to denote observations unique to yeast. PA200 is poorly conserved; yeast and mammalian orthologs maintain significantly less than 20% identity, and multiple sequence alignments are limited to key residues and structural motifs (23, 24, 29). In budding yeast, Blm10 was initially identified as a nuclear protein associated with nascent and de novo synthesized 20S CP and is considered to be the functional ortholog of PA200 (30). A later study in yeast suggested a role for Blm10 in the maintenance of mitochondrial functions (24). PA200 was detected in most mammalian organs by immuno-blotting, and found particularly abundant in testis (21). Knockdown cells (PA200ΔΔ) exhibited a decrease in the level of mature spermatozoa and multiple defects in spermatogenesis. However, PA200 deficiency did not lead to significant defects in either embryonic development or viability of adult mice, as shown with PA200 null mice (31). Even in Arabidopsis, PA200 null mutants are phenotypically indistinguishable from wild type under a variety of growth conditions, which include darkness, and short and long photoperiods (32). These observations suggest that PA200 is not strictly essential in most model organisms.

One condition that seems to require PA200 is response to radiation damage and in DNA repair. Following γ-irradiation, the predominantly nuclear PA200 relocalizes into subnuclear foci and associates with chromatin, suggesting a role in response to DNA repair (21). Relocalization is mediated by DNA-PK, which is a PI3-related kinase that responds to infra-red (IR)-induced damage, but not on ATM or p53 (33, 34). Accumulation of PA200 on chromatin correlates with formation of hybrid PA200-CP-RP proteasomes and an increase in general and postacidic proteolytic activity. However, in mammalian cells PA200 is not essential for the repair of DNA double-strand breaks generated upon treatment with ionizing radiation or bleomycin (31). Likewise, Δblm10 yeast mutants and null PA200 mutants in Arabidopsis survive stress conditions that accumulate damaged proteins or DNA damage such as induced by bleomycin or IR exposure (35). In contrast, deletion of a C-terminal domain of Blm10/PA200 prevents its nuclear import and compromises survival of cells exposed to bleomycin, a property attributed by the authors to dominant-negative cytoplasmic sequestration (22), though defects in nuclear assembly of proteasomes cannot be ruled out (30). Taken together, the studies above suggest that although not absolutely essential for survival, PA200 may play a supporting role in DNA repair in parallel with other factors (34).

PA200 is composed almost entirely of α-turn-α modules identified as HEAT (Huntingtin - elongation factor 3 -PR65/A subunit of PP2A - lipid kinase TOR) repeats (29, 36, 37). Because of their concave nature, 32 such HEAT repeats in PA200 wrap around into a two-layered snail-like dome situated over the 20S surface, covering the translocation channel and well positioned to regulate traffic into the 20S (24, 29) (see Fig. 2). It is still unclear whether PA200 functions solely as a proteasome regulator or might harbor independent roles as well. A significant increase in cellular PA200 levels, followed by accumulation of PA200-CP, was obtained upon exposure of Arabidopsis seedlings to the proteasome chemical inhibitor MG132 or in a mutant that dampens RP activity, suggesting response of PA200 to cellular proteolytic demand (32). However, a Δblm10 mutant showed no synthetic sensitivity to accumulation of damaged or miss folded proteins, indicating that Blm10 might not be involved in general protein degradation, but rather have a more specific role. One example whereby PA200 may participate in degradation of a select subset of substrates is during the maintenance of mitochondrial functions (24). A double mutant deleted for BLM10 and the nonessential stress-induced proteasome transcriptional activator RPN4 showed synthetic sensitivity to accumulation of damaged or miss folded proteins that must be removed by the proteasome, indicating that Blm10 might have a positive role as a proteasome regulator when regulation of the proteasome is defective (22).

Fig. 2.

Structure of Blm10–20S complexes as revealed by x-ray crystallography (24). A, Side view of the symmetric Blm10-CP-Blm10 complex. The four-ringed 20S CP (each α or β ring is color coded separately) capped on both ends by Blm10 in white. Blm10 forms a dome-like structure, with a lateral opening. B, Top view of Blm10-CP complex. Blm10 (in red) contacts the α-ring surface of the 20S CP and opens the axial gates of each of the seven α-subunits of the 20S (in white). The cartoon presentation of Blm10 highlights the repetitive α-helical repeats that interact with all seven α-subunits below encircling the central pore. C, Docking of Blm10 onto 20S CP α-ring. The three C-terminal amino acids of Blm10 (red tail jutting downward from dome) insert into a pocket between proteasome subunits α5 and α6 (the “lys66” pocket). Binding of Blm10 reorients Pro17 of α5, leading to rearrangement of α5 N-terminal residues away from central pore and toward Blm10 cavity (white tail pointing upwards into Blm10 cavity). The resulting partially open gate (see also panel B) may alleviate some hindrance to substrate traffic. D, Solenoid fold of blm10, Top view. Blm10 forms a solenoid structure formed by the 32 HEAT repeats, each consisting of a α-turn-α of variable length (29). The helixes are color coded from Red at the C terminus (docking into the 20S; panel C) to Blue at the N terminus (peak of dome) to highlight the super helical nature of Blm10.

In addition (or instead) of being an activator of entry rates into the 20S CP chamber, PA200 may be a flusher, aiding exit of peptide products through a widened orifice as has been suggested for 11S-Regs (5, 7). Such a role would be a particularly appealing option for PA200–20S-19S hybrids (because the 19S RP would activate uptake of protein substrates and PA200 would accelerate excretion of peptide products). But PA200 may also be a bona fide activator of proteolysis, facilitating entry of substrates into PA200–20S complexes. The question is what for? Proteolysis of what kind of substrates would be enhanced by a non-ATPase activator? It has been shown that opening of the channel alone can permit 20S CP to proteolyse unstructured proteins, thus gating and substrate translocation through the open channel are not mandated to be ATP-dependent (11). Through its ability to open the channel, could PA200 be involved in proteolysis of stress-induced damaged or miss folded proteins, or possibly of defective unfolded polypeptides leaching off ribosomes (defective ribosomal intermediate products) (38)? If so, one could imagine that PA200 may be involved in Ub-independent proteolysis. So far, there is no evidence for this. Alternatively (and this is utter speculation), PA200 may yet be involved in more complex (and Ub-dependent) reactions by serving as a docking site for enzymes that feed into the Ub system, possibly forming an ersatz RP with faux lidlike complexes. Finally, Blm10/PA200 seems to be involved in proteasome assembly, although its specific role still remains an open question (39–43).

PA200 in Proteasome Assembly

The role of Blm10/PA200 in proteasome maturation is a matter of debate. In an early study, the transition from the nascent to matured 20S CP was actually accelerated in yeast upon deletion of BLM10 (YFL007w; the gene coding for Blm10) pointing to a complicated relationship between Blm10 and proteasomes (30). In other studies, Blm10 was found to promote proteasome maturation. For example, Blm10 was detected in the inactive 15S-proteasome precursor (Fig. 1), which contained the following subunits: all α subunits, precursor forms of β2, β3, β4, and the assembly factors Ump1 and Pba1 and Pba2 (39). Blm10 was also found in the next assembly intermediate containing all 20S subunits except for the β7/Pre4 subunit, which is believed to be the last subunit to incorporate prior to dimerization of the so-called half proteasome (two α₇β₇ rings coming together) (40). However, deletion of the BLM10 gene alone had little effect on 20S CP formation from these half proteasomes, although a combined deletion of BLM10 with a C-terminal truncation in β7/Pre4 led to accumulation of half proteasomes, suggesting that both the C-terminal extension of β7/Pre4 and Blm10 have additive roles in stabilizing 20S CPs. Similarly, a combined deletion of BLM10 with a mutation in the SEN3/RPN2 gene led to accumulation of proteasome precursors. It was suggested that 19S RP can substitute for Blm10 in stabilizing nascent 20S CPs and promoting their maturation. Indeed, attachment of 19S subunits to these proteasome precursors was found in Δblm10 extracts, but not in wild-type extracts, indicating that 19S RPs bind to proteasome precursor complexes when Blm10 is absent. It was suggested that Rpn2 (independently shown to bind the 20S α-ring surface (44–46)) can partially substitute for Blm10 in stabilizing the nascent 20S CP (40). Likewise, it has also been claimed that the 19S RP substitutes Blm10 in binding the CP surface (41). 20S CP of yeast mutants with disordered or open CP gates (α3ΔN, α7ΔN, α3ΔN/α7ΔN, α4ΔN, Δpac3Δpac4) were found primarily as Blm10-CP-Blm10, Blm10-CP-RP, or RP-CP-RP. Yet upon deletion of Blm10, in α4ΔN/Δblm10 or α3ΔNα7ΔN/Δblm10 cells for example, most CPs were occupied with RPs, suggesting that sufficient RPs are available to substitute for Blm10 during CP maturation (41).

An up-regulation of Blm10 expression can be found in Δump1 cells, which are defective in CP maturation, resulting in increased levels of Blm10-CP-RP and especially Blm10-CP-Blm10 compared with wild type (WT) (Fig. 1). Because the symmetric complex containing two Blm10 units per CP was found primarily in Δump1 mutants, the lack of Ump1 might induce disordering in gate residues spawning activated CPs, and indeed, CPs with constitutively open gates (α3ΔN, α7ΔN, and α3ΔN/αΔN) were found primarily in complex with Blm10. One possible explanation may have been incomplete maturation of β precursors in this complex; however this was thought not to be the case (41). Apparently, interaction of Blm10 with the α-ring is not correlated with the presence or activation state of β subunits, as it can be found both with inactive or nascent, 20S CP species, as well as with mature or activated species. These observations are summarized schematically in Fig. 1. In fact, association of Blm10 with 20S CP may be correlated with ordering in the α-ring rather than with the status of β subunits, preferring open gate conformations of activated CP. This may explain why higher levels of Blm10 were found in proteasomes from open channel mutants (11, 41).

Native gels are a powerful tool to resolve and capture distinct PA200-containing proteasome configurations even straight from whole cell extracts (21, 22, 39, 47). These “alternative” configurations can then be excised and contents analyzed (for example by immunoblotting or by tandem MS (MS-MS)). Comparison of protein levels and peptidase activity within the gel after resolution verified that a 1:1 PA200–20S CP complex had higher specific peptidase activity than 20S CP, yet less than the activated 19S RP-20S CP specie that defines the “classic” 26S proteasome holoenzyme. Activity of Blm10-CP-Blm10 by contrast, was hardly detectable, suggesting that binding of two Blm10 molecules to CP represses peptide hydrolysis (41). The three-dimensional model obtained from the crystal structure (see below) lends support to repression of activity. Overall, six distinct populations of proteasomes were discernable by native gel, three of which contain Blm10 (41) (Fig. 1B). In one study, the Blm10-CP-19S RP hybrid complex may actually represent the predominant form that contains PA200/Blm10 in rapidly growing yeast (22). PA200–20S-19S RP hybrids were also detectable in cell extract (21–23) (Fig. 1).

A distantly related high molecular weight proteasome-interacting protein, Ecm29 (48, 49) is also predicted to accommodate multiple HEAT repeats and resemble a highly curved solenoid, but may differ significantly from PA200 as it contains at least 50% more repeats with some atypical properties (29). Ecm29 partakes in cellular recovery from oxidative stress, apparently by abetting oxidation-induced 26S proteasome disassembly (50). Ecm29 has also been shown to serve as an adaptor for coupling 26S proteasomes to specific cellular compartments (51). It has been proposed that both Blm10 and Ecm29 compensate for deficiencies in 20S CP-dedicated chaperons and assist in 20S CP maturation and RP-CP assembly, but at different molecular checkpoints. Thus, Ecm29-bound RP-CP-RP and RP-CP-Blm10 complexes from Δump1 cells—lacking α3 and with incomplete β5 processing—can be rescued upon addition of α3, resulting in dissociation of the 20S CP and maturation of β5 (52). As mentioned above, Blm10 serves as a scaffold protein for half CP proteasome assembly, facilitates pre-holo CP formation and remains bound to CP with unusually configured α-rings, whereas Ecm29 recognizes aberrations within the β-ring (52).

The issue of proteasome activation by Blm10 is somewhat of a mystery; as mentioned above, peptidase hydrolysis by Blm10-CP was enhanced relatively to latent 20S CP, whereas similar hydrolysis by Blm10-CP-Blm10 was reduced (41). Evidently, PA200 is found in a stable complex with mature and active 20S CP and not only with precursors or nascent species during assembly (Fig. 1), although mature 20S CP can still be subject to regulation by shifting between latent and activated forms. To address such questions, functional PA200-CP has been purified from extract using chromatography and biochemical fractionation for use in enzymology or structural studies (53). High-salt treatment combined with sizing column or affinity purification enriches samples with PA200-containing proteasome species, particularly PA200–20S complexes (22, 23). PA200-CP complexes can be purified from Arabidopsis crude seedlings extracts using affinity purification by tagged proteasome subunits (32). Once obtained, PA200 was efficient in activating 20S CP for peptidase activity suggesting that it is positioned to facilitate substrate entry by enforcing opening the entry pore (21, 22, 30, 35). Structural studies have been instrumental in deciphering the effect PA200 has on 20S CP channel opening and proteolytic activation.

Structure of PA200 in Complex with 20S CP

PA200 is composed almost entirely of HEAT repeats. HEAT repeats found within functionally diverse proteins are α-helical domains of around 50 residues that pack together to form an elongated superhelix or “solenoid” with a high degree of flexibility. Two adjoining helices linked by a sort turn form the basic unit of HEAT repeats in the resulting solenoid structure (36, 54). Thirty two such HEAT repeats have been delineated in PA200. Because of distinctive amino acid composition, one helix in each repeat is narrower, forming the concave surface; the other wider helix forms the convex surface. Depending on how compact these solenoids fold, HEAT repeat domains could vary from a straight sheet to a horseshoe, a circle or even a spiral wrapping around itself (29, 36, 55–58). The specific HEAT repeats in PA200 were predicted to form highly curved structures, possibly wrapping around into a two-layered dome that when situated over the distal surface of the 20S could participate in gating the translocation channel (29). A spiral, representing this solenoid structure, was visualized by negative staining EM of stand-alone Blm10 (22). Cryo-EM and three-dimensional reconstitution by x-ray crystallography of Blm10–20S complexes provided greater detail revealing a solenoid that contacts each of the seven α subunits of the 20S CP in a defined orientation (23, 24, 35) (Figs. 2A, 2B).

Negative staining electron micrographs of purified PA200–20S complexes showed the activator as a domelike structure (23, 35), not unlike the roughly 200 kDa 11S-Reg complexes to which it was compared (5, 22). However PA200 is a monomer. Its detailed structure and mode of attachment to the surface of the 20S CP must be inherently different from symmetric heptameric 11S-Reg rings. Initially, an EM study at a 23Å resolution showed that PA200 attached to the α-ring surface in a unique and defined conformation, coming in contact with all subunits except for α7 (23). This is a radically different picture than the hollow cones formed by the 11S-Reg family (23). Nevertheless, as attachment of either PA28 or PA200 enhances peptidase activity to a similar extent, both undoubtedly facilitate traffic through the gated channel. A cross section of the three-dimensional reconstruction illuminated how this may occur. Upon binding of PA200, the central region of the α-ring rearranged to yield an apparent open channel conformation. Density maps suggested that extensive rearrangement extended beyond the central-pore region (23). Cryo-EM, followed by three-dimensional reconstitution of Blm10 in complex with 20S CP at a 18 Å resolution revealed a cave-like dome with lateral opening, just over α7 (35). Blm10 when in complex with the mature 20S yeast proteasome also opened the axial gates on the α-rings in a manner similar to its mammalian homologue PA200. Thus, both Blm10 and PA200 bind to the α-rings of the 20S core proteasome as hollow asymmetric caps. Yet, PA200 does not appear to contact the α7 subunit, whereas Blm10 appears to bind all α subunits.

A high-resolution image obtained from the crystal structure of 20S capped on both ends by Blm10 confirmed the solenoid nature of the domelike structure covering the α-ring surface with a lateral opening situated over the α7 subunit (24). Blm10 snaked around itself one and a half times in a superhelical solenoid that was largely hollow (Figs. 2B, 2D). The dimensions of the lateral opening through the Blm10 dome were 13 Å by 22 Å (Fig. 2A). This is consistent with the observation of Blm10/PA200 stimulating hydrolysis of small peptides but not of proteins (24). However, it is still unclear whether traffic of substrates (or products) through this dome and into the 20S CP occurs via this lateral window, or whether conformational changes in Blm10 alter between active and suppressive modes based on their ability to accommodate or translocate substrates. Blm10 makes numerous contacts with the α-ring surface (24), however it is insertion of the final three amino residues (including the key penultimate tyrosine residue Tyr2142) into the Lysine pocket at the interface of α5 and α6 (Fig. 2C) that seems to induce partial pore opening (Fig. 2B). The three-dimensional structure obtained from Blm10-CP crystals illuminates the multiple contacts made between Blm10 HEAT and the 20S surface. It is interesting to note that HEAT repeat karyopherins make contacts with subunits of the nuclear pore complex, facilitating traffic of macromolecules (10, 29, 46). In analogy, do HEAT-containing proteasome-caps (such as Blm10 or Ecm29) also partake in nuclear transport of proteasome species through the nuclear pore, or by nature of their flexibility and potential for myriad interactions, do they facilitate traffic of substrates or products through the internal proteolysis pore within the proteasome?

Modes of 20S CP Activation

The 11S regulators (PA26 and PA28) are hollow cones that form a contiguous channel leading straight into the proteolytic chamber (5, 17). In the case of 11S-Regs, the symmetry match between the two heptameric rings (11S-Reg and 20S CP are both heptamers), imposes an ordered open conformation of the channel blocking residues at the center of the α-ring that facilitates substrate traffic (7). Upon binding of PA26 to the yeast 20S CP for instance, the seven α subunits pivot outward, opening the pore (7). This movement is most pronounced for α3 and α4. It was shown that indeed α3 plays a pivotal role in gating (6). Activation by the 11S regulators is bimodal: binding to the 20S CP surface is via their C termini that insert into pockets (often referred to as Lys66 pockets) in the α-ring, and gate opening relies on an activation loop in the activator complex to induce conformation changes in residues that block the central pore. In contrast, both the 19S RP and Blm10 seem to be able to simultaneously bind 20S CP and stabilize an open gate conformation using their C-terminal residues (24). In the case of the 19S RP it is the C termini of ATPase subunits (primarily Rpt2 and Rpt5) that insert into the lysine pockets on the 20S CP (59–61). Naturally, as a monomer, Blm10 has only one C terminus. The penultimate tyrosine residue of Blm10, Tyr2142 (Phenylalanine in some homologs) is invariably conserved among PA200 orthologs and seems to make the same interactions as seen for the penultimate Tyr/Phe of ATPase subunits of the archaeal PAN or of the 19S RP that were shown to be critical for gate opening. The last three amino acids of Blm10 insert into the Lysine pocket at the interface of α5 and α6 (Fig. 2C), positioning the penultimate tyrosine residue (Tyr2142) to contact Gly17 of α5, stabilize Pro17 of the same chain in a reverse turn, and reposition the N-terminal tail of α5 to induce partial pore opening (Fig. 2B). The Pro17 turns of neighboring subunits, α2, α3, and α4, lack direct contact with Blm10 and thus become disordered. In summary, PA200 may provide a lesson as to how other asymmetric regulators (such as the 19S RP) could recognize minor differences among α subunits to activate the pore.

That PA200 is correlated with activated 20S CP when found in the asymmetric PA200-CP complex, whereas the symmetric PA200-CP-PA200 seems to be repressed, raises the possibility that PA200 is not so much an activator per se, but rather a “cap” for previously activated proteasomes. Thus the enhanced activity of the asymmetric complex may reflect the unhindered open channel on the opposite “naked” side of the 20S CP barrel. Indeed a preponderance of Blm10 with 20S CP was found with activated proteasomes identified under some stress conditions or open channel mutants (41). That the symmetric PA200-CP-PA200 is not fully activated may be a consequence of a closed dome like structure that does not facilitate traffic of large polypeptides or structured proteins. The partial channel opening induced by PA200 may be insufficient on its own to fully activate latent 20S CP (24).

Closing Words

Traditionally, proteasomes come in two flavors: 20S proteasomes (consisting of the core particle) and 26S holoenzymes in which this core is attached to the 19S regulatory particle (a.k.a PA700). The identification of alternative proteasome activators in stable complexes with 20S CP can be seen as a turning point, challenging the strict definition of what is actually a proteasome (22, 23). Like a game of musical chairs, several large proteins or complexes compete for interactions with the same α-ring surface, at times binding to the two opposing surfaces (“hybrid proteasomes”). Some, such as PA200/Blm10, the 11S regulators (PA26, PA28 etc), and the ATPase-containing regulators 19S (eukaryotes) or PAN (archaea) activate various proteolytic properties. Others, like PI31, enforce inhibitory effects (62). The growing list of proteins that compete for space on the same α-ring reveals this interface as a hot spot of 26S proteasomes regulation. The most prominent property of PA200 is its repetitive α-helices that fold into a closed solenoid. It is noteworthy that most of the non-ATPase proteasome activators consist of multiple, repetitive domains, with high α-helix content, manifested in PA200 by the HEAT repeats. It would seem that the α-ring surface has a particular affinity for such domains. As mentioned above, another large HEAT containing protein, Ecm29, interacts with the proteasome, though precisely where it attaches and for what purpose has yet to be clarified. Likewise there is evidence that Rpn1 and Rpn2, the two largest 19S RP subunits, are made up of α-helical repeats akin to HEAT motifs (37) and may come in close proximity to the outer surface of the 20S CP α-ring (44–46, 57, 63, 64), though in these cases too, the precise arrangement of these peripheral proteins is unclear. Undoubtedly, still others are waiting to be unearthed and characterized. In pursuit of proteasome activators, the HEAT is on.