Inhibition of Proteasome Activity Induces Formation of Alternative Proteasome Complexes

The proteasome is an intracellular protease complex consisting of the 20S catalytic core and its associated regulators, including the 19S complex , PA28αβ, PA28γ, PA200 and PI31. Inhibition of the proteasome induces autoregulatory de novo formation of 20S and 26S proteasome complexes. Formation of alternative proteasome complexes, however, has not been investigated so far. We here show that catalytic proteasome inhibition results in fast recruitment of PA28γ and PA200 to 20S and 26S proteasomes within 2-6 h. Rapid formation of alternative proteasome complexes did not involve transcriptional activation of PA28γ and PA200 but rather recruitment of preexisting activators to 20S and 26S proteasome complexes. Recruitment of proteasomal activators depended on the extent of active site inhibition of the proteasome with inhibition of β 5 active sites being


INTRODUCTION
The proteasome, one of the main proteolytic systems of the cell, is essential for maintenance of protein homeostasis and thus of cellular functions. The proteolytic activity of this multicatalytic protease resides inside the 20S core particle, built of four stacked rings of seven α and β subunits with α 7 β 7 β 7 α 7 organization. The internal cavity of the complex, defined by the two β rings, encloses three pairs of distinct active sites named according to their differential cleavage site specificity: chymotrypsin-like (CT-L), trypsin-like (T-L) and caspase-like (C-L) catalytic sites (1). Due to the closed conformation of the native 20S core complex, binding of different proteasome activators is required to induce opening of the 20S entry pores and allow injection of substrates into the catalytic chamber for degradation (2,3). Three different families of activators, highly conserved during evolution, have been identified (4): the 19S regulatory particle recognizes polyubiquitylated proteins and targets them for degradation in an ATP-dependent manner (5). Binding of either one or two 19S regulators to the 20S core gives rise to the 26S and 30S proteasome, respectively. The proteasome activator 28 (PA28) family, which comprises the heptameric PA28α/β and PA28γ complexes, and the proteasome activator 200 (PA200), however, function in an ubiquitin-and ATP-independent manner unless they assemble into so-called hybrid proteasome complexes that also contain one 19S regulator attached to one side of the 20S core (3,(6)(7)(8)(9). In addition to these three activators, another regulator (PI31) of the 20S complex has been identified but its functions for proteolysis remain obscure (10). Formation of PA28α/β heteroheptamers is induced by interferon-γ and LPS stimulation of cells. This activator is cytoplasmic and preferentially binds to the inducible 20S immunoproteasome (11)(12)(13)(14). Both PA28γ and PA200 are constitutively expressed in most cell types (7,15). PA28γ is exclusively found in the nucleus and associates as a homoheptamer with the 20S complex (16,17). PA28γ-containing proteasomes have been implicated in degradation of specific nuclear substrates involved in regulation of cell cycle progression and intranuclear dynamics (6,(18)(19)(20)(21). The monomeric PA200 is mainly present in the nucleus (7). Its function is not well understood, but involvement in genomic stability, maintenance of glutamine homeostasis, and spermatogenesis has been reported (22)(23)(24)(25)(26). Recruitment of proteasomal activators to the 20S catalytic core gives rise to a variety of different alternative proteasome complexes consisting of singly or doubly capped 20S proteasomes as well as hybrid complexes of different activators attached to 26S proteasomes (23,(27)(28)(29). Recent evidence indicates that the 20S and its activators may function as building blocks that assemble rapidly into different proteasome complexes with diverse functions and substrate specificities, allowing fast and dynamic adaptation of proteasome activity to cellular needs (30-32). Adaptation of 20S and 26S proteasome function has been particularly well studied in response to proteasome inhibition. Inactivation of the 20S catalytic sites by small molecule inhibitors results in concerted transcriptional upregulation of proteasomal genes and augmented formation of 20S and 26S/30S proteasome complexes (33). Moreover, inhibition of 26S proteasomes by silencing of 19S regulatory subunits caused enhanced transcription of proteasomal genes and increased formation of 20S proteasomes in Drosophila (34). Recent publications suggested a protective effect of 19S reduction in response to catalytic proteasome inhibition, but involvement of the proteasome activators PA28γ and PA200 in this response has not been investigated so far (35,36). Recruitment of PA28αβ in response to proteasome inhibition has been shown in reticulocytes lysates in vitro (29). Here, we show a so far unknown dynamic recruitment of PA28γ and PA200 to 20S and 26S proteasomes in cells in response to either proteolytic inhibition or siRNA-mediated 26S/30S proteasome knockdown.

EXPERIMENTAL PROCEDURES
Primary cell culture Primary human lung fibroblasts (phLF) were isolated from organ donor lungs and cultured as previously described (37-39). 24 h prior to treatment phLF were synchronized by incubation in starvation medium containing 1% FBS. Cells were treated with bortezomib (Millennium, Takeda), oprozomib (Onyx Pharmaceuticals) and epoxomicin (APExBIO) at the indicated concentrations and times.  (39).

Western blot analysis
Cells were lysed in TSDG or RIPA buffer as previously reported (40

Co-immunoprecipitation
Cells were lysed in TSDG buffer as described above. 3 µL of antibody specific for 20S subunit α4 (BML-PW8120, Enzo Life Sciences) were incubated with Protein G magnetic beads (Dynabeads, Thermo Fisher Scientific) for 15 min at 1200 rpm. For immunoprecipitation, 100 µg of protein were incubated with antibody-coupled beads at a total volume of 250 µL for 2 h on an overhead shaker at 4°C. 10 % of the total volume was removed as input control. After four times washing of the beads in 400 µL TSDG buffer containing 0.2% NP40 co-immunoprecipitated proteins were eluted by incubation of beads in Laemmli buffer at 95°C for 10 min. Subsequently, eluted samples and input control were analyzed by SDS-PAGE and Western blotting.
Quantitative real-time RT-PCR RNA extraction from phLF using Roti®-Quick-Kit (Carl Roth), reverse transcription and quantitative PCR was performed as previously described (40). The 60S ribosomal protein L19 (Rpl19) served as a housekeeping gene for relative expression analysis.

MTT assay
Metabolic activity of phLF in response to PA28γ silencing and bortezomib treatment was analyzed by 2,5-diphenyltetrazolium bromide (MTT) assay. 20 000 cells per well were cultured in a 24-well plate with transfection mixture for PA28γ silencing. After 24 h medium was exchanged to starvation medium containing 1% FBS. 48 h after silencing cells were treated with 10 nM bortezomib for 6 h. Cells were washed with fresh starvation medium and incubated for 24 h to allow recovery of the proteasome activity. For the MTT assay, the cells were incubated with a solution of thiazolyl blue tetrazolium bromide (5 mg/mL in PBS, Sigma) for 1 h at 37°C. Afterwards medium was aspirated and the blue crystals were dissolved in isopropanol + 0.1% Triton X-100. Absorbance was measured at 570 nM in a Sunrise TM plate reader (TECAN).

Cell count
For analysis of cell growth via cell count the same protocol was used as described for the MTT assay with the only difference that cells were cultured in 6-well-plates. Cells were trypsinized, mixed with trypan blue (Sigma-Aldrich) and cell number was determined using a Neubauer counting chamber.

Production and purification of PA28γ
The PA28γ (human) cDNA was subcloned into the pET-14B plasmid. After expression in E. coli, the protein complex was purified essentially as previously described for PA28αβ (43).

In vitro reconstitution assay
Purified recombinant PA28γ was mixed with 1 µg of human 20S proteasome (purified from HeLa cells (43) and treated with 25 µM epoxomicin or solvent) in activity buffer (Tris-HCl 20 mM, NaCl 50 mM, DTT 1 mM, glycerol 10%) (final volume 20 µL). After incubation at 37°C for 5 min, the samples were loaded onto a native gel as described (44). After electrophoresis, the gel was incubated with 100 µM Suc-LLVY-AMC in activity buffer at 37°C for 20 min to visualize peptidase activity (44). The gel was then transferred on a PVDF FL membrane (40 V, overnight) and the 20S proteasome and PA28γ were immunoprobed simultaneously with mouse anti-α7 (BML-PW8110-0100, Enzo Life Sciences) and rabbit anti-PA28γ (BML-PW8190-0100, Enzo Life Sciences), respectively. The membrane was scanned using an Odyssey scanner (Licor) and the fluorescence was quantified using the ImageJ software.

Catalytic proteasome inhibition induces formation of alternative proteasome complexes.
Primary human lung fibroblasts (phLF) were treated with a single dose of the proteasome inhibitor bortezomib (BZ, 10 nM) for 24 h and formation of proteasome complexes was analyzed via native gel electrophoresis. In-gel activity assays with a specific substrate for the chymotrypsin-like (CT-L) activity demonstrated effective inhibition of 20S and 26S/30S proteasome complexes ( Figure 1). Subsequent immunoblotting for α1-7 20S subunits and the 19S subunit Rpt5 unambiguously identified the distinct proteasomes as 20S and 26S/30S complexes, respectively. Blotting of 20S α-subunits also revealed the formation of additional 20S-containing proteasome complexes of higher molecular weight suggesting recruitment of proteasomal activators. Indeed, immunostaining of native gels for PA28α, PA28γ and PA200 proteasome activators demonstrated formation of alternative proteasome complexes, composed of 20S and 26S proteasomes bound to these activators, in response to proteasome inhibitor treatment.

Proteasome activators PA28γ and PA200 are rapidly recruited to 20S and 26S complexes upon proteasome inhibition.
To determine the kinetics of activator recruitment, formation of alternative proteasome complexes was analyzed upon treatment with 10 nM BZ for 2, 6, 16 and 24 h, focusing on the activators PA28γ and PA200. Immunoblotting for the 20S subunit β5 indicated formation of additional 20S containing proteasome complexes at early time points of 2 to 6 h after addition of the proteasome inhibitor ( Figure 2A). Indeed, recruitment of PA28γ and PA200 to 20S and 26S proteasomes was observed already after 2 h whereas nontreated controls exhibited low levels of alternative proteasome complexes. Assembly of alternative proteasome complexes increased further over time peaking after 16 h of BZ treatment. Measuring the three different catalytic activities of the proteasome with synthetic fluorescent substrates revealed an almost complete inhibition of the CT-L activity after only 2 h of treatment ( Figure 2B). The caspase-like (C-L) active sites were inhibited to a much lesser extent (by 10-40%) than the CT-L activity and with a delayed kinetic. In contrast, the trypsin-like (T-L) activity was significantly elevated after 6 h and peaked at 16 h of inhibitor treatment. Taken together, inhibition of mainly the chymotrypsin-like activity induced rapid and pronounced formation of alternative proteasome complexes.

Proteasomal activators interact with 20S and 26S forming alternative proteasome complexes.
We confirmed our results on the proteasome inhibitor-induced formation of alternative proteasome complexes by two other methods, namely gel filtration and co-immunoprecipitation. Increased binding of PA28γ to 20S and 26S proteasomes was observed in gel filtration assays of BZ-treated HeLa cell extracts and thus confirmed our native gel results in a different cell type ( Figure 3A). In solvent treated control cells PA28γ mainly formed free multimers, which were recruited to 20S and 26S complexes in response to proteasome inhibition. Moreover, 20S pull down experiments with an antibody directed against the 20S subunit α4 revealed increased binding of proteasomal activators to the 20S proteasome in phLF upon treatment with 10 nM BZ for 6 h ( Figure 3B). Immunoblotting indicated enrichment of α4 via immunoprecipitation and comparable pulldown efficiencies. Altogether, these data confirm the increased interaction of the activators to 20S proteasomes complexes upon catalytic inhibition of the proteolytic sites in different cell types.

Early recruitment of proteasomal activators is transcriptionally independent.
To investigate the underlying mechanism of recruitment, we analyzed the expression of PA28γ and PA200 in response to proteasome inhibition. Of note, both PA28γ and PA200 protein levels did not change until 16 h of proteasome inhibition as determined by Western blot analysis of phLF ( Figure 4A). Only after longer time points, i.e. after 16 and 24 h, levels of the activators were significantly increased compared to time matched controls. For PA200, this corresponded to delayed induction of RNA transcripts after 16 h while PA28γ mRNA expression was not upregulated suggesting differential transcriptional regulation in response to proteasome inhibition ( Figure 4B). Divergent transcriptional regulation of PA200 and PA28γ after proteasome inhibition is supported by in silico promoter analysis of both genes for putative binding sites of NRF1 (nuclear factor (erythroid-derived 2)-like 1) ( Figure 4C). NRF1 has been described as the critical transcription factor for regulation of proteasomal gene expression in response to proteasome inhibition (45,46). A highly conserved NRF1 binding site was identified in close proximity to the transcription start of the PSME4 promoter (gene name of PA200). In contrast, only one NRF1 binding site which is only conserved in primates was predicted far upstream of the transcriptional start site of the PSME3 promoter (gene name of PA28γ). These data thus support an early, transcriptionally independent recruitment of preexisting PA28γ and PA200 activators to 20S and 26S proteasome complexes as part of a rapid cellular response to catalytic proteasome inhibition. At later time points, expression of both PA28γ and PA200 is upregulated on the protein level. Only PA200, however, is regulated on the transcriptional level, in a manner possibly involving NRF1.

Inhibition of 26S/30S-proteasome activity induces formation of only PA200 alternative proteasomes.
To further assess whether formation of alternative proteasome complexes and induction of PA28γ and PA200 expression at later time points involves ubiquitin-mediated protein degradation, we selectively inhibited 26S/30S ubiquitin-mediated protein degradation by partial knockdown of the 19S subunit Rpn6 (39,47,48). After 72 h of Rpn6 silencing, we noted marked upregulation of PA200 on mRNA as well as on protein levels whereas expression of PA28γ was not significantly altered ( Figures 5A & B). Accumulation of K48-linked polyubiquitylated proteins proved efficient inhibition of ubiquitin-mediated protein degradation by partial knockdown of Rpn6 ( Figure 5B). In-gel activity and immunodetection of 20S and 19S subunits confirmed impaired assembly of 26S and 30S complexes ( Figure 5C). At the same time, the amount of active 20S was increased and alternative 20S proteasome complexes were formed. Of note, these alternative complexes contained only PA200 but not PA28γ as demonstrated by Western blotting of the native gels. These data further support the notion that delayed transcriptional regulation and alternative proteasome formation is specific for PA200 and involves factors that are sensitive to impairment of ubiquitin-mediated protein degradation such as NRF1, while PA28γ is regulated by mechanisms independent of 26S/30S proteasome activity.

Recruitment of proteasomal activators depends on the extent of proteasome inhibition.
In a next step, we investigated whether acute formation of alternative proteasome complexes directly correlates with the degree of proteasome inhibition. Treatment of phLF with increasing doses of BZ for 6 h (1, 10, 50, 100 nM) led to a dose-dependent inhibition of the CT-L and C-L catalytic sites whereas the T-L like activity was not inhibited. We observed dose-dependent accumulation of alternative proteasomes as shown by native gel electrophoresis and immunoblotting for PA28γ and PA200 ( Figure 6A). The small molecule inhibitor oprozomib (OZ) was used to only inhibit the chymotrypsin-like activity of the proteasome ( Figure 6B) (49). Again, dose dependent inhibition of the CT-L catalytic site resulted in enhanced recruitment of PA28γ and PA200 to 20S and 26S proteasomes. Interference with total proteasome activity by treatment with proteasome inhibitor epoxomicin resulted in an even more pronounced formation of alternative proteasome complexes ( Figure 6C). Therefore, these results suggest first that inhibition of the chymotrypsin-like activity is sufficient for recruitment of PA200 and PA28γ and second that formation of alternative proteasome directly correlates with the degree of proteasomal inhibition.
PA28γ recruitment to purified 20S proteasomes does not depend on active site modifications by proteasome inhibitors. Previous data suggested an allosteric opening of 20S core particles upon active site occupancy by peptidic ligands and allosteric stabilization of the 20S-19S interaction in inhibitor-treated 26S proteasomes (50)(51)(52)(53). We thus investigated whether recruitment of PA28γ depended on activesite modification of the 20S proteasome in vitro.
In an in vitro reconstitution assay, purified 20S proteasomes were first inactivated by addition of epoxomicin, and binding of purified PA28γ was investigated. Native gel analysis revealed activation of the CT-L activity upon addition of PA28γ, which was completely blocked by epoxomicin treatment (Figure 7). Immunoblotting of the native gels confirmed formation of PA28γ-containing proteasomes. Importantly, binding of PA28γ to the purified 20S proteasome was not enhanced upon active site inhibition as indicated by quantification of PA28γ-bound 20S proteasomes ( Figure 7A). Similar results were obtained using bortezomib instead of epoxomicin, and using PA28γ either in limiting or excess molar ratios compared to the 20S proteasome (data not shown). These data suggest that recruitment of the proteasome activator PA28γ is not induced via allosteric reorganization of 20S core complex in vitro.
PA28γ recruitment allows cells to cope with proteasome inhibition. In order to further investigate the functional role of PA28γ-containing alternative proteasome complexes, we performed recovery experiments where cells were first treated with BZ for 6 h and then allowed to recover in fresh medium for 24 h.
In this setting, the chymotrypsin-and caspase-like activities, which had been considerably inhibited in response to 6 h BZ treatment, recovered ( Figure 8A). The C-L and T-L active sites even showed increased activity compared to controls. Native gel analysis revealed that the CT-L activity was fully restored in 26S/30S proteasome complexes while 20S complexes were still partially inhibited indicating that activity assays with total extracts are not fully reflecting the differential activity profiles of 26S and 20S proteasome complexes ( Figure 8B). Of note, PA28γ-containing 20S proteasome complexes persisted upon recovery of proteasome activity suggesting that these alternative complexes may play a functional role for rescuing cells from proteasome inhibition ( Figure 8B). We tested this hypothesis by silencing of PA28γ and analysis of the cellular response to 6 h pulse exposure of BZ and 24 h of recovery. While phLF transfected with control siRNAs recovered well from proteasome inhibition, silencing of PA28γ significantly decreased metabolic activity of bortezomibchallenged cells ( Figure 8C). Diminished metabolic activity was accompanied by reduced cell numbers in PA28γ-deficient cells in response to BZ treatment ( Figure 8D). Of note, the different treatments did not induce cell death but only significantly affected cell numbers of living cells as determined by trypan-blue exclusion assay (data not shown). These data indicate that loss of PA28γ results in cell cycle arrest. In accordance with these findings, protein levels of cyclin D1, whose timely degradation is required for cell cycle progression, and expression of the cell cycle inhibitor p21, a substrate of PA28γ, were significantly elevated upon loss of PA28γ and BZ treatment compared to its controls ( Figure 8E) (19,54). Although we cannot distinguish whether this increase is due to accumulation of p21 in the absence of functional PA28γ or a secondary effect upon cell cycle arrest, these data indicate that cells become more sensitive to proteasome inhibition upon loss of PA28γ and suggest that PA28γ-containing alternative proteasome complexes may play an important role in the cellular stress response to impaired proteasome function.

DISCUSSION
Our data demonstrate a novel and rapid regulation of alternative proteasome complex formation upon proteasome inhibition. Earlier studies by our group already unraveled enhanced de novo assembly of 20S and 26S proteasomes after catalytic 20S inhibition but regulation of alternative proteasomes in cells has not been investigated so far (33). Here, we show for the first time that preexisting proteasome activators PA28γ and PA200 are rapidly recruited to 20S and 26S proteasomes in response to catalytic inhibition of the proteasome in different cell types. In contrast, only PA200 is recruited upon inhibition of 26S ubiquitin-mediated protein degradation. In this latter case, transcriptional activation of PA200 is observed at late stages. To our knowledge such rapid recruitment of proteasomal activators is a novel finding which has only been described for PA28αβ in rabbit reticulocyte lysates in vitro (29).
Two central questions arise from this observation: how are proteasomal activators recruited to 20S and 26S proteasomes and why are alternative proteasome complexes formed in the cell?
As proteasomal activators were rapidly recruited to the 20S and 26S proteasomes before any alteration in their expression levels, this suggests the presence of a reservoir of free PA28γ and PA200 in the cell. Indeed, in untreated cells, we observed low levels of both activators associated with 20S or 26S complexes and PA28γ was mainly found as free heptamers while 20S and 26S proteasomes were mostly present in a PA28γ-free form ( Figure 3A). Similarly, Fabre et al. observed that the majority of 20S proteasomes is uncapped and that less than 5% of total 20S is associated with PA28γ and PA200 (14). Moreover, proteasome subunits have been described as stable and long-lived proteins, supporting the presence of a backup pool of proteasomal subunits, which can quickly be regulated upon certain stimuli (55). Hence, this pool of available activators may serve as a reserve of building blocks enabling fast recruitment to 20S and 26S proteasomes upon cellular stimuli such as proteotoxic stress. Rapid recruitment may involve allosteric effects of inhibited 20S core particles, posttranslational modifications or other cofactors rather than induction of protein levels. In our study, recruitment of PA28γ and PA200 directly correlated with the extent of catalytic inhibition with the inhibition of the β5 active site being sufficient for recruitment. This suggests that the decrease in catalytic activity allosterically regulates the association of the 20S core particle with the activators. An allosteric mechanism was proposed for purified yeast proteasomes where inhibition of the β5 active site stabilized the interaction between 20S and 19S complexes (50). Similarly, active site modifications were proposed to induce conformational changes to PA28 binding-sites in T. acidophilum (53). Binding of peptidic proteasome inhibitors to the catalytic active sites has been shown to induce an allosteric conformational shift of the α-subunits that propagates opening of the 20S pore (51,52). However, we did not observe any effects of catalytic proteasome inhibition on the recruitment of purified PA28γ to 20S proteasomes using an in vitro reconstitution assay. In addition, PA200-20S alternative proteasomes were formed upon specific inhibition of 26S/30S proteasomes when 20S active sites are not modified by proteasome inhibitors. The underlying mechanism for recruitment remains to be determined and may involve accumulation of substrates, dynamic posttranslational modifications, competing binding partners and cofactors in a complex cellular environment. Cofactors might also be molecules that prevent recruitment of activators to the 20S core complex under normal conditions, but that dissociate upon proteasome inhibition. A similar mechanism has been shown for the ER resident chaperone immunoglobulin-binding protein (BiP), which binds to ER stress sensor proteins at baseline conditions but is recruited to misfolded proteins upon activation of the unfolded protein response thereby activating the sensors (56). The cellular response to proteasome inhibition involves transcriptional upregulation of 20S proteasome subunits and certain proteasome activators (57). NRF1 was identified as the critical transcription factor for this type of regulation (45,46). Accordingly, in silico analysis of the PA200 promoter indicated one potential, highly evolutionary conserved NRF1 binding site close to the transcription start site, whereas only one poorly conserved binding site was detected in the extended promoter region of PA28γ. This more reliable prediction of a NRF1 binding site in the promoter sequence of PA200 compared to PA28γ is well in line with our observed specific transcriptional upregulation of PA200 after 16 h of BZ treatment and after 72 h of Rpn6 silencing, while PA28γ was not regulated on the mRNA level at all. In the cell, a vast pool of free PA28γ heptamers exists ( Figure 3A). This may function as a backup reservoir and therefore transcriptional induction might not be required for formation of new alternative proteasome complexes.
What is the function of alternative proteasome complexes in the cell that are formed upon proteasome inhibition? We reason that recruitment of proteasome activators to the 20S and 26S proteasomes is part of a protective cellular stress response to imbalanced protein homeostasis upon proteasome inhibition (30, [58][59][60]. This notion is supported by the observed increased sensitivity of cells to proteasome inhibition upon silencing of PA28γ. Proteasome activators may recruit distinct substrates for proteasomal degradation thereby facilitating cellular recovery after protein stress. In line with this concept, both, PA28γ and PA200, have been shown to target specific proteins for degradation via the 20S core particle: steroid receptor coactivator-3 (SRC-3), cyclin-dependent kinase inhibitors p16, p19 and p21 as well as the deacteylase Sirt1 have been identified as substrates for PA28γ-containing alternative complexes (6,19,20,54). Moreover, PA28γ was shown to be involved in ubiquitin-dependent proteasomal degradation of p53 (61,62). PA200 has been speculated to contribute to the degradation of peptides and damaged or misfolded proteins arising from cellular stress, since its binding to the 20S core particle induces opening of its channel and facilitates the entry of substrates (7,27,63). So far acetylated core histones are the only specific substrates described for PA200 (26). This raises the exciting possibility that rapid formation of PA28γ-and PA200-containing alternative proteasomes directs proteasomal protein degradation to specific sets of substrates in order to allow the cell to cope with conditions of proteotoxic stress upon proteasome inhibition. It is tempting to speculate that this is part of a conserved stress response to misfolded and damaged proteins. While proteasome inhibitor treatment of cells may not represent a physiologic condition, it is important to understand the therapeutic effects of proteasome inhibitors as bortezomib (Velcade) and carfilzomib (Kyprolis) are approved drugs for the treatment of multiple myeloma (64,65). The reduction of 26S proteasome function is a physiologically relevant situation. Impairment of 26S proteasome activity has been reported for aging, (66, 67), neurodegeneration (68), and oxidative stress (69). In these studies, however, formation of alternative proteasome complexes as part of the cellular response to impaired 26S proteasome function was not investigated.
In conclusion, our findings demonstrate rapid remodeling of proteasome complexes and support the building block concept in which proteasome activators rapidly assemble into distinct proteasome complexes according to cellular needs (30). Identification of novel compounds that specifically interfere with PA28γ-or PA200mediated activation of 20S proteasomes should help to unravel the exact cellular roles of these alternative complexes and may also prove to be suitable to manipulate proteasome subcomplexes in a more specific fashion.

26.
Qian   For clarity, the intensity of the signal has been normalized to the maximum value of each curve, and the total areas under each curve were made equal. B) Analysis of interaction of proteasomal activators with the 20S proteasome using co-immunoprecipitation. Lysates of primary human lung fibroblasts (phLF) were treated with BZ for 6 h or with solvent, lysed and subjected to immunoprecipitation using an antibody against the 20S subunit α4.
Co-immunoprecipitated proteins as well as total protein lysate (input, 10% of total volume) were separated by SDS-PAGE. Direct interaction of proteasomal activators PA28γ and PA200 with the 20S proteasome subunit α4 was visualized via immunoblotting. Representative results of experiments in cells from three different donors are shown.

Figure 4: Early recruitment of proteasomal activators is transcriptionally independent.
A) Western blot analysis showing expression of total PA28γ and PA200 in primary human lung fibroblasts (phLF) treated with 10   A) Primary human lung fibroblasts (phLF) were treated with Rpn6 or scrambled control siRNAs for 72 h and mRNA expression of Rpn6, PA28γ and PA200 upon Rpn6 silencing was assessed by quantitative RT-PCR. Bars indicate fold change of mRNA levels compared to control siRNA-transfected cells (One sample t-Test, n=3). B) Western blot analysis for K48-polyubiquitylated (UbiK48) proteins, the 19S regulatory subunit Rpn6 and proteasomal activators PA28γ and PA200 in response to Rpn6 knockdown. Densitometric analysis indicates β-actin normalized signals relative to the respective scrambled siRNA control (One sample t-Test, n=3). C) Native gel analysis of Rpn6 silencing. In-gel CT-L activity assays and immunoblotting for the 19S subunit Rpt5, the 20S subunits α1-7 and PA28γ and PA200 indicate the activity and composition of proteasome complexes. Experiments were performed in phLF from three different organ donors and representative results of one donor are shown. Primary human lung fibroblasts (phLF) were treated with different doses of proteasome inhibitors (bortezomib (BZ) in A); oprozomib (OZ) in B) for 6 h, lysed and chymotrypsin-like (CT-L), caspase-like (C-L) and trypsin-like (T-L) activities were measured using fluorescent substrates. Bars indicate % activity compared to the solvent treated control for three experiments using three different phLF lines (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=3). Dose dependent recruitment of proteasomal activators was detected via native gel electrophoresis including in-gel CT-L activity assay and immunoblotting for PA28γ and PA200 using the respective antibodies. C) Analysis of proteasome activity and proteasome complex formation in response to treatment of phLF with 20 µM epoxomicin for 4.5 h. Bars show % activity compared to the solvent treated control. All experiments were performed in phLF from three different organ donors and native gels show representative results of one donor (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=3).  A) Analysis of proteasome activity in primary human lung fibroblasts (phLF) after 6 h bortezomib (BZ) treatment and recovery in fresh medium for 24 h by using fluorescent substrates specific for the three catalytic sites of the proteasome. Bars indicate % activity compared to the solvent treated control for three experiments using three different phLF lines (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=3). B) Analysis of proteasome complex formation of lysates used in A) via native gel analysis including ingel CT-L activity assay and immunoblotting for 20S subunit β5 and PA28γ using the respective antibody. C) Metabolic activity of phLF after PA28γ silencing and co-treatment with BZ. 48 h after transient silencing of PA28γ, phLF were treated with 10 nM BZ for 6 h followed by 24 h recovery in fresh medium. Metabolic activity of BZ and siRNA co-treated cells was assessed using the MTT assay. Bars indicate % metabolic activity compared to control siRNA and solvent treated control (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=3). D) Determination of cell numbers of upon treatment as in C). The number of dead and living cells was determined by trypan blue exclusion. Bar diagram shows % cell count of living cells compared to control siRNA and solvent treated control. The number of dead cells was negligible in all conditions (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=4). E) Analysis of the accumulation of cyclin D1 and p21 in phLF upon PA28γ silencing and co-treatment with BZ. 48 h after transient silencing of PA28γ, phLF were treated with 10 nM BZ for 6 h followed by 24 h recovery in fresh medium. PA28γ, Cyclin D1 and p21 protein levels were analyzed by Western blotting using the respective antibodies. Bar diagram indicates densitometric analysis of β-actin normalized signals relative to the scrambled siRNA control (One-way ANOVA, Bonferroni's Multiple Comparison Test, n=4).