Intramolecular Control of Protein Stability, Subnuclear Compartmentalization, and Coactivator Function of Peroxisome Proliferator-activated Receptor γ Coactivator 1α*

Peroxisome proliferator-activated receptor γ coactivator (PGC)-1 is a critical transcriptional regulator of energy metabolism. Here we found that PGC-1α is a short lived and aggregation-prone protein. PGC-1α localized throughout the nucleoplasm and was rapidly destroyed via the ubiquitin-proteasome pathway. Upon proteasome inhibition, PGC-1α formed insoluble polyubiquitinated aggregates. Ubiquitination of PGC-1α depended on the integrity of the C terminus-containing arginine-serine-rich domains and an RNA recognition motif. Interestingly, ectopically expressed C-terminal fragment of PGC-1α was autonomously ubiquitinated and aggregated with promyelocytic leukemia protein. Cooperation of the N-terminal region containing two PEST-like motifs was required for prevention of aggregation and targeting of the polyubiquitinated PGC-1α for degradation. This region thereby negatively controlled the aggregation properties of the C-terminal region to regulate protein turnover and intranuclear compartmentalization of PGC-1α. Exogenous expression of the PGC-1α C-terminal fragment interfered with degradation of full-length PGC-1α and enhanced its coactivation properties. We concluded that PGC-1α function is critically regulated at multiple steps via intramolecular cooperation among several distinct structural domains of the protein.

The primary structure of PGC-1␣ suggests a separation into two distinct regions. The N-terminal region contains a transcriptional activation domain, through which PGC-1␣ recruits histone acetyltransferase protein complexes, such as CBP/p300 and SRC-1. It also contains binding sites for nuclear receptors and other DNA-binding transcription factors. Notably, the N-terminal half of PGC-1␣ is sufficient for activating the PGC-1␣-responsive promoter to the same degree as full-length PGC-1␣. In contrast, the C-terminal region of PGC-1␣ contains a unique domain alignment not apparent in other nuclear hormone-receptor coactivators that includes serine-arginine-rich (RS) domains and an RNA recognition motif (RRM) (8). These structural features are characteristic of SR proteins, which facilitate spliceosome formation and orchestrate splice site selection. It is therefore likely that the C-terminal region of PGC-1␣ is involved in mRNA processing. This hypothesis is supported by the demonstrated physical interaction of the C-terminal half of PGC-1␣ with the hyperphosphorylated form of RNA polymerase II, positive transcription elongation factor-b (Cdk9 (cyclin-dependent kinase 9)-cyclin T), and RNA processing factors (9).
Both the protein expression and activity of PGC-1 must be tightly regulated to maintain the temporal and tissue-specific control of mitochondrial function in response to diverse metabolic demands, resulting in a necessarily short half-life. The signaling for induction of PGC-1␣ at the mRNA levels has been well studied (10 -15); however, it remains unknown how PGC-1␣ is inactivated to maintain adequate protein levels. PGC-1␣ is a highly potent inducer of mitochondrial biogenesis and function. Prolonged activation of PGC-1␣ may therefore not be permissive and possibly even pathogenic. Constitutive cardiac specific PGC-1␣ overexpression resulted in dilated cardiomyopathy with disruption of sarcomeric structures via the uncontrolled proliferation of mitochondria (16). It is generally thought that the levels of potent transcriptional activator proteins must be limited due to their cytotoxicity, and inverse correlations between intracellular levels of proteins and their potency as transcriptional activator have been demonstrated (17).
This study showed that the expression levels of PGC-1␣ are controlled by protein degradation through the ubiquitin-proteasome pathway and that PGC-1␣ is inherently prone to aggregation. Both proteasome inhibition and truncation of N-terminal regulatory motifs facilitated intranuclear aggregation of this protein. In addition, intramolecular interactions between the N-terminal and C-terminal regions regulated PGC-1␣ coactivator function by controlling protein stability and intranuclear localization.
Cell Fractionation-To prepare nuclear and cytosolic extracts, cells were washed with phosphate-buffered saline, collected by centrifugation, and resuspended in hypotonic buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 M dithiothreitol, and a protease inhibitor mix). The swollen cells were homogenized and centrifuged at 10,000 rpm for 1 min. The supernatant was collected as the cytosolic fraction. The pellet was extracted in a high salt buffer (20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 M dithiothreitol, and a protease inhibitor mix), and the soluble fraction was collected as nuclear extracts following another centrifugation. The remaining insoluble pellet was resuspended in SDS lysis buffer (Upstate Biotechnology, Inc.) and solubilized by sonication on ice.
Western Blotting and Immunoprecipitation-Rabbit polyclonal antibodies against FLAG, HA, PGC-1␣, and MAT1 as well as mouse monoclonal antibodies against PML and Gal4 were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Goat polyclonal antibody against lamin B was also purchased from Santa Cruz Biotechnology, and a mouse monoclonal antibody against NuMA was purchased from BD Transduction Laboratories. For coimmunoprecipitation, nuclear proteins were resuspended in immunoprecipitation buffer (50 mM Tris, pH 8.0, 75 mM NaCl, 0.1% Nonidet P-40, 1 M dithiothreitol, and a protease inhibitor mix) and incubated with FLAG-agarose beads (Sigma). The bound proteins were resolved by SDS-PAGE. Protein expression was visualized using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (Amersham Biosciences) and detected using a LAS-3000 luminescence image analyzer (FUJIFILM).
Luciferase Reporter Assays-Cells were cotransfected in 6-well plates with expression vectors, firefly luciferase reporter genes (0.1 g), and a constitutive Renilla luciferase control (pRL-CMV; 50 ng; Promega) using Trans IT-LT1 (Mirus). Luciferase activity was determined 48 h after transfection, using the Dual-Luciferase Reporter Assay System (Promega) and a Monolight 2010 luminometer. All transfections comprised three or more independent experiments, each in duplicate.
Immunofluorescence-COS-7 cells or cardiomyocytes were plated onto glass coverslips in a 6-well plate. Cells were washed with phosphate-buffered saline and fixed in 4% paraformaldehyde for 10 min. Cells were then washed three times with cold phosphate-buffered saline and permeabilized for 5 min with phosphate-buffered saline containing 0.2% Triton X-100. Permeabilized cells were then blocked for 30 min with blocking buffer (3% bovine serum albumin and 0.1% Triton-X in Trisbuffered saline). Cells were stained with primary antibodies against FLAG (1:10,000; Sigma), anti-␣-actinin (1:800; Sigma), or PGC-1␣ (1:200; Santa Cruz Biotechnology) for 1 h at room temperature or overnight at 4°C. Secondary antibodies conjugated with either AlexaFluor 488 or TRITC (1:200; Dako Cytomation) were applied for 1 h at room temperature. Nuclei were stained with ToPro3 (Molecular Probes, Inc., Eugene, OR) in mounting medium. The stained cells were observed by confocal laser-scanning microscopy (LSM510; Carl Zeiss, Jena, Germany), with appropriate emission filters.
Statistical Analyses-Data are reported as mean Ϯ S.E. Comparisons were analyzed by analysis of variance, using a significance level of p Ͻ 0.05.

Proteasome Inhibition Stabilizes PGC-1 Protein and Docks It in an Insoluble
Subnuclear Compartment-Initially, this study examined how PGC-1␣ protein stability might be regulated. A number of other rapidly degraded proteins contain stretches of amino acids enriched in proline, glutamic acid, serine, and threonine, termed PEST domains, and these regions have been implicated in promoting protein turnover (20). Analysis of the human PGC-1␣ protein sequence using a PEST-FIND program identified two putative PEST sequences with high scores (11.45 and 9.98, respectively), one located in the N-terminal transcriptional activation domain (amino acids 70 -146) and the other located between amino acids 255-270 (Fig. 1A). PGC-1 orthologs are found to be limited in higher vertebrates, such as mammals, birds, and fish. As shown in Fig. 1B, the two identified PEST-like motifs are highly preserved in PGC-1 orthologs of humans, pigs, rats, chickens, puffer fish, and zebrafish. This observation led us to further investigate whether degradation of PGC-1 is coupled to the ubiquitinproteasome pathway. We approached this by monitoring protein levels of PGC-1␣ after treatment with proteasome inhibitors. COS-7 cells were transiently transfected to express FLAG-tagged PGC-1␣ and cultured in the presence of a reversible proteasome inhibitor, MG132. Nuclear extracts were obtained from these cells at different time points after the addition of MG132 and then subjected to Western blotting. PGC-1␣ protein levels in the nuclear extracts were increased by at least 5-fold at 1 h after treating the cells with MG132 relative to untreated cells ( Fig. 2A, top). Sur-  were treated with MG132 (5 M) for the indicated periods of time. Protein fractions were prepared from the nuclear extracts (NE) and nuclear pellet (NP) and subjected to Western blotting. B, COS-7 cells were transiently transfected with FLAG-PGC-1␣ or -␤ vector, and the expressed protein was detected by anti-FLAG antibodies. Membranes were stripped and reprobed with anti-NuMA (nuclear matrix protein) and anti-MAT-1-specific antibodies. C, endogenous PGC-1␣ protein in neonatal rat cardiomyocytes was determined by anti-PGC-1␣ antibodies.
prisingly, longer incubations with MG132 decreased PGC-1␣ protein levels in a time-dependent manner. Given that protein degradation is blocked by proteasome inhibition, this paradoxical reduction of PGC-1␣ protein in the nuclear extracts raised the possibility that PGC-1␣ might translocate into specific subnuclear compartments, where it is tightly bound to a structure such as the nuclear matrix, thus rendering it insoluble and not subject to degradation. Based on this idea, we analyzed the pellet remaining after preparation of the nuclear extract, which contained insoluble nuclear proteins resistant to extraction in buffer containing high salt (400 mM NaCl) and nonionic detergent. The pellet was resuspended in SDS lysis buffer, solubilized by sonication (referred to hereafter as the nuclear pellet fraction), and then immunoblotted in parallel with the nuclear extracts. As shown in the bottom panel of Fig. 2A, the untreated nuclear pellet showed no PGC-1␣ activity. Following MG132 treatment, however, the protein levels of PGC-1␣ in the nuclear pellet gradually increased until they had almost recovered at 6 h after the addition of MG132. Similarly, treatment with lactacystin, another proteasome inhibitor, increased the levels of PGC-1␣ protein in the nuclear extracts initially, followed by a gradual decrease with a reciprocal increase in the nuclear pellet levels. 4 Transiently expressed PGC-1␤ in COS-7 cells exhibited similar behavior in response to proteasome inhibition (Fig. 2B). The fidelity of our cell fractionation procedure was verified by the absence of MAT-1 (Ménageà-trois-1) in the nuclear pellet, whereas NuMA (nuclear matrix marker nuclear mitotic apparatus protein) was absent in the nuclear extracts (21). Taken together, such dynamic behavior of PGC-1 following inhibition of proteasome activity was ascribed to a subnuclear compartment shift. To confirm that this shift occurred with endogenous PGC-1␣, the same biochemical fractionation was performed using cultured neonatal rat cardiomyocytes. As shown in Fig. 2C, significant levels of endogenous PGC-1␣ were detected in the nuclear pellet when the cardiomyoctes were preincubated with MG132 (Fig. 2C).
Proteasome Inhibition Protects PGC-1␣ from Proteolysis but Changes the Intranuclear Distribution Pattern-Next, we examined the dynamic behavior of intranuclear PGC-1␣ by immunocytochemistry using laser confocal microscopy. COS-7 cells were transiently transfected with FLAG-PGC-1␣ expression plasmid and analyzed by immunofluorescence staining using an anti-FLAG antibody. In the absence of MG132, PGC-1␣ showed a diffuse nucleoplasmic distribution with concentrated staining in fine speckles (Fig. 3A, left). Treatment of the cells with MG132 for 16 h drastically altered the distribution of PGC-1␣ staining into multiple large speckles or aggregates (Fig. 3A, right). We next immunolocalized endogenous PGC-1␣ in neonatal rat cardiomyocytes following proteasome inhibition to exclude the possibility of the changed staining pattern being due to protein overexpression. In the absence of MG132, endogenous PGC-1␣ showed a fine speckled pattern throughout the nucleoplasm, whereas treatment with MG132 caused the formation of discrete nuclear foci stained for PGC-1␣ (Fig. 3B). Although these foci were fewer in number and smaller in size than those observed in the transfected COS-7 cells, these results supported the notion that PGC-1␣ dynamically shifts its subnuclear location upon proteasome inhibition. To examine the effect of protein synthesis on the steady-state levels of PGC-1␣ , COS-7 cells were treated with cycloheximide to inhibit de novo protein synthesis 24 h after transfection with FLAG-PGC-1␣ and observed at various time points up to 5 h. The intensity of PGC-1␣ immunofluorescent signals rapidly decreased to undetectable levels by 5 h after cycloheximide treatment (Fig. 3C, upper panels). Similar results were observed with endogenous PGC-1␣ in rat cardiomyocytes (Fig. 3D), strongly indicating the requirement of ongoing protein synthesis to maintain steady-state levels of PGC-1␣ protein. Proteasome inhibition with MG132 eventually changed the fine speckled diffuse nucleoplasmic distribution of PGC-1␣ to the distinctive concentration in large speckles, most possibly via protection of PGC-1␣ from proteolysis (Fig. 3C, lower panels). In contrast, neither the staining intensity nor pattern of an irrelevant protein, such as Cdk9 (22), was affected by treatment with cycloheximide and/or MG132 during the same time course. 4 We therefore concluded that the nuclear distribution of PGC-1␣ is dynamically regulated under certain circumstances, such as proteasome inhibition, such that it shifts from its nucleoplasm-predominant location to dense large speckles within the nucleus. Moreover, this type of intranuclear compartment shift might be coupled to the ubiquitin-proteasome pathway.
PGC-1␣ Is Degraded via the Ubiquitin Proteasome Pathway-We next asked which subnuclear compartment of PGC-1 is targeted for ubiquitination. Immunofluorescence analyses revealed significant overlap of ubiquitin staining with the fine speckled pattern of PGC-1␣ throughout the nucleoplasm (Fig.  4A). In the presence of MG132, the accumulated PGC-1␣ protein in the larger nuclear speckles also colocalized with ubiquitin staining (Fig. 4B). To examine the ubiquitination state of PGC-1␣ biochemically, COS-7 cells overexpressing FLAG-PGC-1␣ and hemagglutinin (HA)-tagged ubiquitin either in the presence or absence of MG132 were immunoprecipitated using anti-FLAG-agarose and then analyzed for the presence of ubiquitination using Western blotting with an anti-HA antibody. As shown in Fig. 4C, higher molecular weight bands representing polyubiquitinated PGC-1␣ were detected in the presence of MG132. Together, these results demonstrated that the PGC-1␣ localized in fine speckles throughout the nucleoplasm is targeted for ubiquitination. Inhibition of the proteasome triggers the relocation of polyubiquitinated PGC-1␣ into the large nuclear speckles, where the protein is tightly complexed and not easily extracted.
Promyelocytic leukemia (PML) nuclear bodies are tightly bound to the nuclear matrix and are enriched in multiple components of the proteasome and ubiquitinated proteins (23). In this study, we also colocalized polyubiquitinated PGC-1␣ with PML, confirming these compartments as nuclear bodies (Fig. 4D).

N-and C-terminal Parts of PGC-1␣ Cooperatively Regulate Proteolysis and Localization via the Ubiquitin-Proteasome
Pathway-The results thus far suggested that both degradation and subnuclear localization of PGC-1␣ are tightly coupled to the ubiquitin-proteasome system. To further delineate the domains of PGC-1␣ regulating these phenomena, we constructed a panel of C-and N-terminal deletion mutants of PGC-1␣, depicted in Figs. 5A and 6A. The N-terminal region of PGC-1␣ contains functional domains for transcription initia-tion, including transcriptional activation domains and the binding sites for transcription factors. The C terminus of PGC-1␣ contains two RS domains and an RRM, through which PGC-1␣ is proposed to act in mRNA processing of target genes during transcriptional elongation (9).
First, we examined the effect of C-terminal deletions on the subcellular localization, protein stability, and ubiquitination status of PGC-1␣. Although full-length PGC-1␣ protein was exclusively localized in the nucleoplasm, the RS/RRM deletion mutant PGC-1␣-(1-565) was present in both cytoplasm and nucleoplasm, and further deletion toward the N-terminal end (residues 1-292) completely abrogated the nuclear localization (Fig. 5B, upper panels). These findings together with the fact that a C-terminal fragment of PGC-1␣-(565-798) remained exclusively localized in the nucleus implied that at least one functional nuclear localization signal exists in the C-terminal region. Consistent with the immunofluorescent staining, Western blotting showed PGC-1␣-(1-565) distributed equally between the cytosol and nuclear extract fractions, whereas the majority of PGC-1␣-(1-292) protein was recovered in the cytosol fraction (Fig. 5C). The intensity of PGC-1␣-(1-565) immunofluorescence was unchanged 5 h after cycloheximide treatment, whereas that of full-length PGC-1␣ was rapidly decreased (Figs. 3C and 5E). Truncation of the C-terminal RS/RRM region thus stabilized PGC-1␣-(1-565) protein in the nucleus, and exposure of the cells to MG132 had no effect on the protein levels or subcellular distribution of PGC-1␣-(1-565). In support of this, nucleoplasmic PGC-1␣-(1-565) protein failed to form large speckles in the nucleus upon proteasome inhibition (Fig. 5B, lower panels). Furthermore, the ubiquitination of PGC-1␣ was blocked by deletion of the C-terminal RS/RRM region (Fig. 5D).

Exogenous Expression of the C-terminal PGC-1␣ Fragment Interferes with Degradation of Full-length PGC-1␣ and
Enhances Coactivator Function-Our results indicated that the C-terminal fragment of PGC-1␣-(565-798) predominantly localized in the nucleus and was sufficient to bind the ubiquitination machinery. Furthermore, PGC-1␣-(565-798) was resistant to degradation by the proteasome. We therefore postulated that PGC-1␣-(565-798) might act as a "decoy protein," competitively binding to and sequestering the ubiquitination machinery for PGC-1␣. To test this hypothesis, we examined the impact of PGC-1␣-(565-798) expression on full-length PGC-1␣ expression levels. In agreement with our hypothesis, PGC-1␣-(565-798) expression enhanced the protein levels of full-length PGC-1␣ in nuclear extracts in a dose-dependent manner (Fig. 7A) but had no effect on the biochemical or structural compartmentalization of PGC-1␣. In contrast, PGC-1␣-(565-798) did not affect the levels of Gal4-VP16, an irrelevant chimeric transcriptional activator whose expression is limited at very low levels by rapid degradation via the ubiquitin-proteasome pathway (24) (Fig. 7B). These results indicated that PGC-1␣-(565-798) interferes with the ubiquitination of full-length PGC-1␣, thereby protecting it from degradation.
The activity of transcriptional factors seems to be coupled to the ubiquitin-dependent degradation of these proteins (17). Based on this, we finally examined the effect of blocking PGC-1␣ proteolysis on its transcriptional activity. To do this, we exogenously expressed PGC-1␣-(565-798) and observed the impact on promoters of two mitochondrial respiratory chain complex components, cytochrome c and ATP synthetase ␤. These two promoters have been well characterized and contain ERR-responsive elements and nuclear respiratory factorbinding sites, both of which are known coactivation targets for PGC-1␣. Coexpression of PGC-1␣-(565-798) activated both promoters in a dose-dependent manner but had no effect on the control cytomegalovirus promoter (Fig. 7, C-E).
PGC-1 possesses an intrinsic activation function when directly tethered to DNA as a chimeric protein fused to the DBD derived from yeast transcription factor Gal4. This Gal4-DBD-PGC-1␣ fusion protein markedly activated the Gal4-responsive reporter activity, compared with negligible activity for Gal4-DBD alone. Cotransfection of an expression plasmid encoding PGC-1␣-(565-798) increased the transcriptional activity of Gal4-DBD-PGC-1␣ in a dose-dependent manner but had no effect on Gal4-DBD fused to the VP16 activation domain (Fig. 7, G and H). PGC-1␣ binds nuclear receptors through its leucine-rich LXXLL motifs, designated as L2 and L3. Furthermore, binding to an endogenous repressor, p160 Myb-binding protein, also requires L2 and L3 (26) (18). Mutation of these two leucine residues to alanine will thus allow a more detailed and specific examination of intrinsic transactivation function of Gal4-DBD-PGC-1␣, since the nuclear receptors and endogenous repressor no longer bind PGC-1␣. As expected from the disruption of p160 binding, the mutant L2/3A showed greater transcriptional activity than wild-type PGC-1␣, and coexpression of PGC-1␣-(565-798) enhanced transcription to the same degree as achieved with wild-type PGC-1␣ (Fig. 7G, lines 9 -12). Taken together, these findings indicated that PGC-1␣-(565-798) stabilizes full-length PGC-1␣, thereby enhancing the transcriptional activity of PGC-1␣.

DISCUSSION
This study demonstrated that PGC-1␣ is inherently prone to aggregate within the nucleus, and its activity is tightly coupled with the ubiquitin-proteasome system. The protein stability, intranuclear localization, and coactivator function of PGC-1␣ are also finely regulated, at least in part, via intramolecular interactions among distinct structural domains.
Deletion of the C-terminal region containing RS domains and RRM completely blocked ubiquitination and stabilized the PGC-1␣ protein. In contrast, a truncated protein comprising only the C-terminal region was autonomously polyubiquitinated. These results suggested that the ubiquitination of PGC-1␣ depends on the integrity of the C-terminal region. This notion was further strengthened by ectopic expression of the C-terminal fragment of PGC-1␣ disrupting the degradation of full-length PGC-1␣. Surprisingly, the polyubiquitinated C-terminal fragment of PGC-1␣ was resistant to proteasome-mediated degradation and prone to aggregate formation. The N-terminal region of PGC-1␣, containing PEST-like motifs, targeted the polyubiquitinated PGC-1␣ to the proteasome and inhibited aggregate formation. These findings prompted us to speculate that the N-terminal region regulates the aggregation properties of the C-terminal region by protein degradation, probably utilizing the PEST-like motifs. This would regulate protein turnover of PGC-1␣ and maintain this coactivator availability at transcriptionally active sites (27). In this context, it should be noted that p38 MAPK phosphorylates PGC-1␣ on three amino acid residues located within and adjacent to the PEST-like motifs (Thr 262 , Ser 265 , and Thr 298 ) and increases the half-life of PGC-1␣ protein 3-fold (28). Further research is needed to clarify how the N-terminal region of PGC-1␣ is regulated in terms of degradation or aggregation properties in response to, for example, environmental stresses.
Biochemical fractionation revealed that PGC-1 protein levels in the nuclear extracts were increased in the presence of proteasome inhibitors; however, prolonged treatment resulted in a time-dependent translocation of PGC-1 into the insoluble nuclear pellet with a reciprocal decrease of protein in the soluble nuclear extract fraction. Consistent with these results, immunofluorescent study revealed a drastic shift in the nucleoplasm-predominant distribution of PGC-1 into large, discrete intranuclear aggregates on proteasome inhibition. These findings suggested that PGC-1 is an aggregate-prone protein that forms intranuclear inclusions when the protein degradation machinery is impaired.
What is the molecular mechanism by which PGC-1 forms the observed intranuclear aggregates? Protein aggregation is based on a nucleated polymerization reaction (29). Self-aggregation of PGC-1 might occur when the protein levels of PGC-1 reach a critical level (30). Intranuclear inclusions of nuclear poly(A) binding protein PABPN1 are pathological hallmarks of inherited oculopharyngeal dystrophy, and this protein tends to aggregate and form intranuclear inclusions in a ubiquitinationindependent manner when exogenously expressed in cells (31). Interestingly, both an arginine-rich domain and RNA-binding domain are essential for the aggregation-prone properties of PABPN1. As mentioned above, the C-terminal region of PGC-1 contains an RS domain and an RRM. These domain structures might therefore be common to a certain class of aggregationprone proteins. Alternatively, PGC-1 proteins might facilitate self-aggregation when polyubiquitinated. Some subnuclear compartments, such as nuclear matrix, could assist the aggregation by scaffolding polyubiquitinated PGC-1 through ubiquitin-binding domains (32,33). Moreover, it remains unknown why PGC-1 aggregates are resistant to proteasome-mediated degradation. A possible explanation is that the proteasome precludes the clearance of oligomeric and aggregated protein, since substrates need to be unfolded before passing through the narrow pore of the proteasome barrel (34).
The nucleus is organized into defined territories specialized for distinct functions (27). Clastosome is a subset of PML bodies that harbors a privileged site of nuclear protein degradation and high concentrations of proteasome components, ubiquitin conjugates, and PML. Clastosome assembles when proteolysis is highly active in the nucleus but disappears when proteasome is inhibited (23). Although endogenous clastosome might prevent the accumulation of PGC-1 before aggregation is initiated, disruption of the balance between clastosome activity and PGC-1 protein levels by inappropriate degradation and/or production might lead to aggregate formation. Colocalization of PML with PGC-1 aggregates supports this notion, and we could speculate that PGC-1␣ aggresomes impair the normal operation of PML bodies, including proteolysis, by altering the organization of nuclear domains containing PML protein (35,36). Conversely, PML was shown to block the ubiquitin-proteasome-dependent degradation of particular proteins, including p73 (37).
Ubiquitination of transcriptional activators and subsequent proteasome-mediated degradation provide an inhibitory feedback mechanism and control of the level of protein expression (38). Moreover, functional overlap of sequences that activate transcription and signal ubiquitin-mediated degradation indicates that the ubiquitin-proteasome pathway is sometimes tightly coupled with the activities of transcriptional activators (39,40). Ligand-dependent and cyclic assemblies and disassemblies of diverse nuclear hormone receptor complexes on responsive promoters further supports the notion that degradation and transcriptional activity are mutually interdependent (41).
Our results suggested that protein degradation of PGC-1␣ followed by C terminus ubiquitination does not directly couple with the potent coactivation function of PGC-1␣. Deletion of the N-terminal transactivation domain (residues 1-185) had no effect on protein stability and intranuclear localization. In addition, the C-terminal fragment of PGC-1␣-(565-798) interferes with the ubiquitination and subsequent degradation of fulllength PGC-1␣, thereby activating transcription of PGC-1␣dependent promoters in a dose-dependent manner. This result also indicated a limit to the endogenous ubiquitination machinery available to act on the C-terminal part of PGC-1␣ in the nucleus and that the C-terminal fragment of PGC-1␣-(565-798) might competitively sequester PGC-1␣-specific E3 ligases. Isolation of such candidates would clearly facilitate our understanding of the ubiquitin-proteasome pathway in transcriptional regulation by PGC-1␣.
Reduced activity of PGC-1␣ has been associated with many human diseases in different tissues, including heart failure (5,42), and aging (43). However, the molecular mechanism by which PGC-1␣ activity is suppressed in certain pathological states remains elusive (5). It is plausible to speculate that altered intramolecular control of protein stability and localization of PGC-1␣ in response to particular stress signals might change the PGC-1␣ activity. The aggregation and sequestration of polyubiquitinated PGC-1␣ into a transcriptionally inactive compartment resulted in reduced PGC-1␣ activity. Furthermore, the formation of nuclear aggregates can impact adversely on nuclear architecture and function (36,44,45). In this regard, intranuclear aggregates of PGC-1␣ might be a pathognomonic feature of certain genetic or acquired human diseases, as seen in poly(Q) (glutamine) diseases (46).