Cyclopentenone prostaglandins of the J series inhibit the ubiquitin isopeptidase activity of the proteasome pathway.

Electrophilic eicosanoids of the J series, with their distinctive cross-conjugated alpha,beta-unsaturated ketone, inactivate genetically wild type tumor suppressor p53 in a manner analogous to prostaglandins of the A series. Like the prostaglandins of the A series, prostaglandins of the J series have a structural determinant (endocyclic cyclopentenone) that confers the ability to impair the conformation, the phosphorylation, and the transcriptional activity of the p53 tumor suppressor with equivalent potency and efficacy. However, J series prostaglandins have a unique structural determinant (exocyclic alpha,beta-unsaturated ketone) that confers unique efficacy as an apoptotic agonist. In seeking to understand how J series prostaglandins cause apoptosis despite their inactivation of p53, we discovered that they inhibit the ubiquitin isopeptidase activity of the proteasome pathway. In this regard, J series prostaglandins were more efficacious inhibitors than representative members of the A, B, or E series prostaglandins. Disruption of the proteasome pathway with proteasome inhibitors can cause apoptosis independently of p53. Therefore, this finding helps reconcile the p53 transcriptional independence of apoptosis caused by Delta12-prostaglandin J(2). This discovery represents a novel mechanism for proteasome pathway inhibition in intact cells. Furthermore, it identifies isopeptidases as novel targets for the development of antineoplastic agents.

activity, as hypothesized (1,2,4,9,10), then two predictions should be valid. First, individual A series and J series PG should act rather uniformly on the cellular processes they affect. Second, their cellular effects should be self-consistent with established models of NFB and p53 function. However, not all experiments affirm these predictions. For example, the A and J series PG both repress NFB transcription and inhibit IB kinase (1,2); however, only ⌬12-PGJ 2 is anti-inflammatory (11). Likewise, the A and J series PG both repress p53 transcription; however, only the A series PG antagonize p53-dependent apoptosis (Ref. 4

and see below).
Herein we report the discovery of a molecular mechanism that clarifies the distinctive cellular effects of cyclopentenone PG. Namely, J series PG preferentially inhibit the ubiquitin isopeptidase activity (ubiquitin-specific protease) of the proteasome pathway. This pathway is the major nonlysosomal degradation pathway in cells (12,13). The degradation of target proteins via this pathway largely depends on their covalent modification with a ubiquitin polymer. This polymer consists of ubiquitin (8.5 kDa) subunits that are covalently attached via isopeptide bonds between the COOH terminus of one ubiquitin and the lysine ⑀-amino group of another ubiquitin. In turn, this polymer is covalently attached to a lysine ⑀-amino group on the target protein via another isopeptide bond. Ubiquitin isopeptidases are a large family of cysteine hydrolases that preferentially cleave these isopeptide bonds (14).
We demonstrate that inhibition of ubiquitin isopeptidase activity by PG depends on nuances of olefin-ketone conjugation. J series PG with a cross-conjugated ␣,␤-unsaturated dienone are more efficacious inhibitors compared with A series PG with a single ␣,␤-unsaturated ketone (Fig. 1). We further demonstrate that J series PG cause apoptosis, concurrently with inhibition of isopeptidase activity, and independently of p53-mediated gene transactivation. Our discovery has particular implications for the development of novel antineoplastic agents that are effective against cancers with dysfunctional p53 pathways.
Cell Culture-We maintained RKO, RKO-E6 (gift from Dr. Mark Meuth, Institute for Cancer Studies, University of Sheffield, Sheffield, United Kingdom), Ts20 cells (17) (gift from Dr. Harvey Ozer, Department of Microbiology and Molecular Genetics, UMDNJ-New Jersey Medical School, Newark, NJ) in Dulbecco's minimum essential medium in a humidified incubator with 5% CO 2 . We supplemented the media with 2 mM L-glutamine, 1 mM sodium pyruvate, 50 units/ml penicillin, and streptomycin, and 10% (v/v) fetal bovine serum. We maintained Ts20 cells at 32°C (permissive temperature) but placed them in a 40°C incubator (nonpermissive temperature) to inactivate the E1 ubiquitin-activating enzyme in experiments designed to observe disassembly of polyubiquitinated proteins (17). All other cells were maintained at 37°C.
We transfected RKO cells with a p53-luciferase reporter construct to measure p53-dependent transactivation (4). We used expression of an independently transfected ␤-galactosidase gene to normalize for transfection efficiency and cell death.
Immunochemical Detection of Proteins-After treating cells with the indicated compounds for the indicated times, we removed the medium and lysed cells in 50 mM Tris, pH 7.4, 100 mM NaCl, 2 mM EDTA with 0.1% SDS, 0.1% deoxycholate, 1ϫ complete protease inhibitor mixture. We measured protein concentration by the Bradford method. We fractionated equal portions of the total cell lysate from each sample (12.5 g of protein) by SDS-PAGE. We transferred proteins to polyvinylidene difluoride blocked with 5% (w/v) nonfat dry milk in Tris-buffered saline T (20 mM Tris⅐HCl, pH 7.5, 100 mM sodium chloride, 0.5% (v/v) Tween 20). We measured proteins immunochemically by using primary antibodies directed against p53 (1:4,000), p21 WAF1/CIP1 (1:5,000), and ubiquitin (1:1000), followed by horseradish peroxidase-conjugated secondary antibodies (1:4000). We detected antigen-antibody complexes with enhanced chemiluminescence reagents. We scanned gels and quantified intensities using Kodak 1D Image Analysis Software.
Immunoprecipitation of p53-We treated RKO and RKO-E6 cells with the indicated compounds for the indicated times. We lysed cells in 250 mM sucrose, 50 mM Tris, pH 7.4, 25 mM KCl, 5 mM MgCl 2 , 1 mM EDTA, 1ϫ complete protease inhibitor, 1 mM NaF, and 1 mM sodium orthovanadate. We sonicated the lysate twice for 5 s at 4°C. After centrifugation at 13,000 ϫ g, we incubated samples containing 200 g of total protein, in 1 ml of lysis buffer, with 1 g of primary antibody, and 20 l of protein A/G PLUS-Agarose overnight at 4°C on a rocker. PAb 1620 precipitates p53 occurring in a wild type conformation; PAb 240 precipitates p53 in a mutant conformation (19 -21). DO-1 precipitates p53 occurring in either conformation or as a p53-ubiquitin conjugate. We centrifuged the samples at 2500 ϫ g to isolate the antigenantibody-bead complexes. We washed the immunoprecipitate four times with 1 ml of phosphate-buffered saline, 0.4% Tween 20. We then measured p53 or ubiquitin by immunochemical analysis with antibodies for p53 (FL-393) and ubiquitin (Ubi-1), respectively. Note that PAb 240 and PAb 1620 are monoclonal antibodies, despite their annotation.
Apoptosis Assays-We incubated 5 ϫ 10 4 RKO cells, plated on a Lab-Tek® II Chamber Slide TM (Nalge Nunc International Corp.), with the indicated compounds for the indicated times and then washed the cells twice with phosphate-buffered saline. We added 100 l of annexinbinding buffer (10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl 2 , pH 7.4) and 5 l of Alexa Fluor® 488-labeled annexin V and incubated samples for 15 min at room temperature. We washed the cells twice with 400 l of annexin-binding buffer and resuspended them in 15 l of 50:50 annexin-binding buffer:water. We stained the cells with propidium iodide (1 M) and Hoechst 33342 (1 M) and examined them by fluorescence microscopy. We determined caspase activity and DNA fragmentation as described previously (22). p21 WAF1/CIP1 Gene Expression-We determined p21 WAF1/CIP1 mRNA expression from cDNA microarray experiments (4). 20 S Proteasome Activity Assays-We determined 20 S proteasome catalytic activity by monitoring the cleavage of substrates for the post-glutamyl peptidase (z-LLE-MCA) and chymotrypsin-like (z-LLL-MCA and LLVY-MCA) proteolytic activities of the 20 S proteasome. Briefly, we incubated RKO cells with the indicated compounds for the indicated times. We lysed cells in 250 l of 25 mM HEPES, 5 mM EDTA, 0.1% CHAPS, 5 mM ATP, pH 7.5 with 2 mM dithiothreitol added just prior to lysis. We sonicated lysates and dispensed 45-l aliquots into four separate wells of a 96-well plate. We added 5 l of each proteasome substrate (500 M in lysis buffer) to separate 45-l samples (final substrate concentration: 50 M). We incubated the assay mixtures at 37°C for 15 min and quantified fluorescence (substrate cleavage) by exciting the samples at ϭ 360 nm and monitoring the emission at ϭ 460 nm.
26 S Proteasome Activity Assay-We determined the 26 S proteasome regulatory activity by monitoring the cleavage of Ub-125 I-lysozyme, prepared as described previously (23). We incubated RKO cells with the indicated compounds for the indicated times. To monitor 26 S proteasome regulatory activity, we incubated 22.5 l of lysate, 2.5 l of an ATP regenerating system (0.5 mM ATP, 60 g/ml creatine phosphokinase, 6.6 mM phosphocreatine, 10 mM Tris-HCl, 0.5 mM MgCl 2 , 1 mM KCl, 0.05 mM dithiothreitol), and 25 l of 125 I-lysozyme-ubiquitin conjugates (136 cpm/l). After 30 min at 37°C, we removed a 20-l sample and added 400 l of bovine serum albumin solution (1% bovine serum albumin in 10 mM Tris, pH 7.5, 0.02% azide). We then isolated the acid soluble fraction by trichloroacetic acid precipitation (50 l 100% w/v trichloroacetic acid, on ice, 15 min) and centrifugation for 15 min at 16,000 ϫ g. We determined the amount of 125 I in the supernatant by ␥ spectrometry and calculated the percentage of soluble 125 I (supernatant 125 I/total 125 I) to determine the extent of substrate proteolysis. Isopeptidase Activity Assay-We determined isopeptidase activity by monitoring cleavage of a peptide substrate (z-LRGG-MCA) that mimics the carboxyl terminus of ubiquitin, which is involved in isopeptide bond formation. We observed that this substrate was subject to degradation by the proteasome; therefore, we incubated RKO cell lysates (0.5 g/ml in 25 mM HEPES, 5 mM EDTA, 0.1% CHAPS, 5 mM ATP, pH 7.5) with 30 M MG115 to inhibit proteasome activity (Ͼ90% inhibition was achieved). These conditions were used to eliminate background proteolytic activity and to permit selective determination of isopeptidase activity. We next added various concentrations of prostaglandins to identify isopeptidase inhibitory activity. Prostaglandins were incubated for 0.5 h on ice prior to substrate addition. We incubated the assay mixtures at 37°C for 3 h and quantified fluorescence as described above. Lysates were also monitored for proteasome activity to ensure that the isopeptidase inhibitory effect seen with the prostaglandins was not due to additive proteasome inhibition.

⌬12-PGJ 2 Impairs the p53 Tumor Suppressor while Inducing
Apoptosis Concurrently-PGA 1 and A 2 can repress transcription by p53, thereby antagonizing p53-dependent apoptosis (4). If their electrophilic cyclopentenone ring confers this activity, as we hypothesized, then ⌬12-PGJ 2 should act analogously to PGA 1 and PGA 2 . Consistent with our hypothesis, we observed that RKO cells exposed to ⌬12-PGJ 2 acquired p53 that was transcriptionally inactive, as determined by a variety of assays. ⌬12-PGJ 2 transforms p53 protein in RKO cells from its wild type, native conformation into a mutant conformation ( Fig. 2A,  lanes 1-3). This mutant p53 conformation is a less efficient substrate for serine 15 phosphorylation. Phosphorylation of this residue on p53 is an early event in p53 activation. Etoposide increased total and phosphoserine 15 p53 in RKO cells relative to the control (Fig. 2B, lanes 1 and 2). However, coincubation with 20 -60 M ⌬12-PGJ 2 reduced serine 15 phosphorylation (Fig. 2B, lanes 3 and 4). The consequences of inhibiting this activating step are reflected in the transactivating ability of p53. Etoposide-stimulated transcription of a p53-dependent reporter gene was potently inhibited by 6 -60 M ⌬12-PGJ 2 (Fig. 2C). It became apparent why p53 failed to transcribe genes when we examined its ability to bind an oligonucleotide containing a p53 consensus binding sequence. In this assay, etoposide (50 M) promoted the formation of p53-32 P-labeled oligonucleotide complexes (Fig. 2D, lane 6) that could be super-shifted with anti-p53 antibody (Fig. 2D, lane 3), and competitively inhibited by excess "cold" oligonucleotide (Fig. 2D, lane 5). Excess cold mutant oligonucleotide inhibited only the nonspecific formation of p53-32 P-labeled oligonucleotide complexes (Fig. 2D, lane 4). However, 60 M ⌬12-PGJ 2 , alone (Fig. 2D,  lane 7) or combined with etoposide (Fig. 2D, lane 8) inhibited the formation of p53-32 P-labeled oligonucleotide complexes. These data suggest that, in terms of their pharmacological effects on p53, J series PG are quantitatively and qualitatively indistinguishable from A series PG. Cyclopentenone PG have similar effects on wild type p53 in cell lines other than RKO cells, including HCT 116, A549, and MCF-7 cells (data not shown).
In contrast to their uniform effects on p53 function, A and J series PG had divergent effects on apoptosis. As expected for agents that impair p53 function, PGA 1 and PGA 2 do not cause apoptosis (Fig. 3). However, ⌬12 PGJ 2 causes apoptosis, characterized by nucleosomal fragmentation, activation of caspase-3, and cellular association of annexin V (Fig. 3, A-C). The induction of apoptosis by ⌬12-PGJ 2 , concurrent with its impairment of p53 function in RKO cells is difficult to reconcile with the conventional model of p53 tumor suppression (6, 7). p21 WAF1/CIP1 Protein Levels Rise, Despite the Decline in Its Transcription-As expected, transcription of endogenous genes regulated by p53, typified by p21 WAF1/CIP1 , declined in RKO cells incubated with ⌬12-PGJ 2 (Fig. 4A). However, we noticed that these cells continued to accumulate p21 WAF1/CIP1 protein (Fig. 4B, lane 4). This implies that ⌬12-PGJ 2 had slowed the rate of p21 WAF1/CIP1 protein degradation. This effect is distinctive for ⌬12-PGJ 2 ; mRNA and protein levels of p21 WAF1/CIP1 fall, in parallel, in RKO cells treated with PGA 1 or PGA 2 (4) ( Fig. 4A and 4B, lanes 2 and 3). Proteolysis via the ubiquitindependent proteasome pathway modulates the activity of several transcription factors, tumor suppressors, and regulatory proteins, including p21 WAF1/CIP1 and p53 (24,25). Inhibitors of the 20 S proteasome catalytic activity can also cause apoptosis (26). Therefore, we hypothesized that ⌬12-PGJ 2 inhibited the ubiquitin-dependent 26 S proteasome pathway.
We used RKO and RKO-E6 cells as a model to test this hypothesis. RKO-E6 cells are derived from isogenic RKO cells stably transfected with the E6 oncoprotein. The E6 oncoprotein is an ubiquitin-conjugating enzyme that hastens proteasomemediated degradation of the p53 tumor suppressor protein (27). One predicts that RKO-E6 cells will accumulate p53 to a lesser extent than will RKO cells under various conditions ranging from basal to genomic stress. In contrast, one predicts that both cell lines will accumulate p53 to a similar extent if ⌬12-PGJ 2 inhibits the ubiquitin-dependent proteasome pathway.
⌬12-PGJ 2 Inhibits the Ubiquitin-dependent Proteasome Pathway-Consistent with our prediction, the ratio of p53 protein in RKO-E6 cells/RKO cells was lower in every case tested, except cells incubated with ⌬12-PGJ 2 (Fig. 5A, lane 5). Note that RKO-E6 cells incubated with etoposide accumulate 3-fold less p53 compared with isogenic RKO cells (Fig. 5A, lane 2), confirming that the E6 oncoprotein hastens the degradation of p53. Note also that RKO-E6 cells incubated with PGA 1 and PGA 2 accumulated negligible amounts of p53 compared with the vehicle control (Fig. 5A, lanes 3 and 4 versus lane 1), indicating that these cyclopentenone PG do not inhibit proteasome-mediated degradation of p53 as efficaciously as ⌬12-PGJ 2 . Both RKO and RKO-E6 cells incubated with ⌬12-PGJ 2 accumulated appreciable p53 protein (ϳ2-3-fold compared with vehicle control (Fig. 5A, lane 5 versus lane 1). A separate control experiment showed that the 20 S proteasome inhibitor, MG115, affected RKO and RKO-E6 cells in a manner analogous to ⌬12-PGJ 2 (Fig. 5A, lane 6 versus lane 5). In panels A-C, RKO cells were incubated with ⌬12-PGJ 2 (0 -60 M, 6 h, 37°C). A, ⌬12-PGJ 2 causes p53 to obtain a mutant conformation. The p53 protein in cell lysates was immunoprecipitated with conformationally sensitive antibodies that distinguish between the native (Pab1620) and mutant (Pab240) conformers of p53. Positive controls were RKO cells transfected with plasmids that express either a conformational mutant of p53 (p53 (V143A)) or wild type p53 (p53 (wt)). B, ⌬12-PGJ 2 inhibits a key step in p53 activation, serine 15 phosphorylation. Total p53 protein and p53 phosphorylated at serine 15 (PO 4 -Ser15 p53) were determined by Western blot analysis. Densitometry was used to obtain the PO 4 -Ser 15 p53/total p53 ratios. This ratio in lane 2 (the positive control, etoposide) is designated as 1 for comparison purposes. C, ⌬12-PGJ 2 inhibits the transcriptional activity of p53. ⌬12-PGJ 2 inhibited the basal transactivation (E) and etoposide-induced transactivation (OE) of a p53-luciferase reporter plasmid in RKO cells. Data are the mean Ϯ S.D., n ϭ 3 experiments. D, ⌬12-PGJ 2 inhibits DNA binding by p53. RKO cells were incubated for 6 h with Me 2 SO vehicle, 50 M etoposide, 60 M ⌬12-PGJ 2 , alone, or combined with 50 M etoposide. Nuclear extracts from these cells were incubated with 32 P-labeled oligonucleotide containing a consensus binding site for p53, and as indicated, with PAB 421, an antibody that binds p53, excess cold oligonucleotide (competitor) identical to the 32 P-labeled oligonucleotide, or excess cold oligonucleotide (mutant) differing in sequence from 32 P-labeled oligonucleotide. Arrows indicate p53-32 P-labeled oligonucleotide complexes. NS ϭ nonspecific labeling; FP ϭ free 32 Plabeled oligonucleotide. Cellular p53 proteins are targeted for 26 S proteasomemediated degradation by conjugation with polyubiquitin (28). Consistent with our hypothesis, ⌬12-PGJ 2 caused RKO and RKO-E6 cells to accumulate high molecular weight p53 epitopes, recognized by anti-p53 antibody (Fig. 5B, lanes 5 and 10). Immunoprecipitation with anti-p53 antibody, followed by SDS-PAGE and immunochemical analysis with anti-ubiquitin antibody, confirmed that these high molecular weight p53 epitopes were polyubiquitinated isoforms of p53 (Fig. 5C, lanes 5 and 6). The accumulation of p53-polyubiquitin isoforms in Fig. 5C signifies that ⌬12-PGJ 2 inhibits the ubiquitin-dependent proteasome pathway in a concentration-dependent manner. This effect is not restricted to p53 or to RKO cells; ⌬12-PGJ 2 also caused a corresponding accumulation of polyubiquitin conjugates in HL-60 cells that lack p53 alleles (data not shown). Polyubiquitin conjugate accumulation was most pronounced in cells treated with J series PG (Fig. 5D, lanes 2-4) compared with PGA 1 , PGA 2 , PGB 1, 15-keto-PGE 2 , or PGE 2 (Fig. 5D, lanes 5-9). J series PG share a cross-conjugated unsaturated ketone, while the other PG do not have this chemical feature. Note that while PGJ 2 does not possess cross-conjugation, it is rapidly converted to ⌬12-PGJ 2 in situ (29). It is unlikely that ⌬12-PGJ 2 inactivated the E6 oncoprotein, or other endogenous ubiquitin ligases, ubiquitin-activating enzymes, or ubiquitin-conjugating enzymes, because cells incubated with ⌬12-PGJ 2 accumulated p53-polyubiquitin conjugates. This implies that these enzymes are still functional in their roles of catalyzing ubiquitin conjugation (Fig. 5B). ⌬12-PGJ 2 retained its potency and efficacy as an apoptotic agonist in RKO-E6 cells (data not shown). This is consistent with our conclusion that PGJ 2 causes apoptosis via a process that is independent of p53 transcription.
⌬12-PGJ 2 Inhibits the Proteasome Pathway at a Site Distinct from the 26 S Proteasome-The 26 S proteasome consists of multiple subunits, including two 19 S regulatory subunits that recognize and bind polyubiquitinated protein substrates, and a 20 S proteolytic subunit having broad substrate specificity (24). MG115, which acts at the 20 S catalytic site, inhibited hydrolysis of several proteasome substrates in whole cell lysates. We examined hydrolysis of substrates for the chymotrypsin-like (z-LLL-MCA and LLVY-MCA) and post-glutamyl peptidase (z-LLE-MCA) proteolytic activities, as well as a 19 S regulatory subunit-dependent substrate, Ub-125 I-lysozyme (Fig. 6A). At a concentration exceeding that which causes polyubiquitin conjugate accumulation in intact cells, ⌬12-PGJ 2 (60 M) did not significantly inhibit hydrolysis of any of these proteasome substrates (Fig. 6A). In other words, MG115 (20 M) is significantly more efficacious (Ն90%) than ⌬12-PGJ 2 (Յ20%) at inhibiting the catalytic activity of the 26 S proteasome. Lactacystin, another agent that acts at the 20 S catalytic site, behaved similarly (data not shown). We noted that the differences in efficacy between MG115 and ⌬12-PGJ 2 as proteasome inhibitors (Fig.  6A) did not correlate with the extent or rate of polyubiquitin accumulation in cells. A dose-response experiment showed that, despite being a weaker proteasome inhibitor, ⌬12-PGJ 2 (6 -60 M) (Fig. 6B, lanes 5-7) caused intracellular polyubiquitin accumulation to a greater extent than MG115 (2-20 M) (Fig. 6B, lanes 2-4). Furthermore, we observed that the molecular weight distribution of polyubiquitin differed between RKO cells treated with MG115 and ⌬12-PGJ 2 . When incubated with 20 M MG115, RKO cells accumulated polyubiquitin distributed around a maximum of ϳ160 kDa. When incubated with 60 M ⌬12-PGJ 2 , RKO cells accumulated polyubiquitin distributed around a maximum of ϳ250 kDa (Fig. 6C). Collectively, these data (Figs. 5 and 6) suggest that ⌬12-PGJ 2 inhibits the terminal step in the proteasome pathway: isopeptidase-mediated disassembly of polyubiquitin.
⌬12-PGJ 2 Inhibits the Isopeptidase Activity of the Proteasome Pathway-To test our hypothesis, we sought an experimental system that would allow us to examine polyubiquitin disassembly in the absence of de novo ubiquitin conjugation. The murine cell line, ts20, harbors a temperature-sensitive E1 ubiquitin-activating enzyme that is catalytically impaired at the nonpermissive temperature, 40°C (17). We first uniformly accumulated polyubiquitinated proteins at the permissive temperature (32°C) by incubating ts20 cells with MG115 (1 M, 12 h). We then added fresh media containing either MG115 (6 M) or ⌬12-PGJ 2 (20 M) and immediately incubated the samples at the nonpermissive temperature and monitored polyubiquitin disassembly by ubiquitin isopeptidases (Fig. 7A). Under these conditions, ⌬12-PGJ 2 slowed the rate of polyubiquitin disassembly significantly (Fig. 7B, lanes 8 -10) compared with Me 2 SO (Fig. 7B, lanes 2-4) or MG115 (Fig. 7B, lanes 5-7). Densitometric analysis of the polyubiquitinated proteins in Fig. 7A showed that, in the absence of de novo ubiquitin conjugation, polyubiquitin disappeared by first-order kinetics with a half-life (t1 ⁄2 ) Х 6 h in ts20 cells treated with 20 M ⌬12-PGJ 2 (Fig. 7C). This rate is 3-fold slower than the rate of polyubiquitin disappearance, t1 ⁄2 Х 2 h, in ts20 cells treated with MG115, consistent with ⌬12-PGJ 2 and MG115 having different sites of action. We note that this effect could not be reproduced with a cell-penetrable, peptide inhibitor of cathepsin B and calpains (MDL 28170 (30)), indicating that polyubiquitin disassembly does not reflect activity of other major cellular proteolytic pathways (data not shown). We also note that ts20 cells respond to PG treatment in a manner similar to RKO cells; ⌬12-PGJ 2 causes a greater accumulation of polyubiquitin conjugates in ts20 cells than an equimolar concentration of MG115 (data not shown).
Isopeptidases comprise a large family of cysteine proteases. Currently, little is known about the substrate specificity of the individual members of this family. Therefore, we measured inhibition of isopeptidase activity by quantifying the degradation of a general isopeptidase substrate, z-LRGG-MCA, that mimics the ubiquitin COOH-terminal isopeptide linkage (Fig.  8). We observed a concentration-dependent inhibition of isopeptidase substrate degradation with ⌬12-PGJ 2 , while PGA 1 , PGB 1 , PGE 2 , and 15-keto-PGE 2 demonstrated no inhibitory  6 h). A, ⌬12-PGJ 2 inhibits HPV-E6-mediated degradation of p53. Cellular levels of p53 protein were determined by immunochemical analysis. MG115 (20 M, 4 h) was used as a positive control for inhibiting proteasome-mediated degradation of p53. Densitometric analysis was used to obtain relative p53 abundance ratios. B, ⌬12-PGJ 2 causes the accumulation of high molecular weight p53. Prolonged exposure of the Western blot in A was used to identify high molecular weight epitopes recognized by anti-p53. C, ⌬12-PGJ 2 causes the accumulation of polyubiquitinated p53. p53 was immunoprecipitated with anti-p53, fractionated by SDS-PAGE, and analyzed immunochemically with an antibody to ubiquitin. D, only J series PG cause the accumulation of polyubiquitin conjugates. Whole cell lysates were collected from RKO cells that were treated as indicated. Ubiquitin was then analyzed immunochemically. activity in this assay. We note that MDL 28170 (180 M) did not impair the degradation of z-LRGG-MCA under these conditions (data not shown). Furthermore ⌬12-PGJ 2 (Յ180 M) did not inhibit the degradation of a proteasome substrate, z-LLE-MCA, or a caspase-3 substrate, Ac-DEVD-MCA, indicating selectivity for the inhibition of isopeptidase activity (data not shown).
Inhibition of cellular ubiquitin isopeptidase activity should have five predictable effects. First, polyubiquitin and/or polyubiquitinated proteins should accumulate (Figs. 5, B-D, and  6B). Second, the cellular pool of monomeric ubiquitin should diminish and its depletion should slow the rate of ubiquitinprotein conjugation, leading to accumulation of unmodified protein (e.g. p53) (Fig. 5A). Third, cells with active 26 S proteasome and inactive isopeptidases (i.e. ⌬12-PGJ 2 treatment) should retain polyubiquitin-protein conjugates differing in composition compared with cells with active isopeptidases and inactive 26 S proteasome (i.e. MG115 treatment) (Fig. 6C). Fourth, the rate of polyubiquitin disassembly by isopeptidases should slow, increasing the half-life of polyubiquitin conjugates (Fig. 7). Last, the cleavage of peptide substrates by isopeptidases should be inhibited (Fig. 8). Our experiments affirm each of these predictions.
The exact trigger for apoptosis due to proteasome pathway inhibition is unknown. However, disruption of proteasome pathway function could result in the buildup of detrimental proteins (i.e. pro-apoptotic or damaged proteins). Also, as others and we have shown, this disruption alters monoubiquitin/ polyubiquitin dynamics, which could affect essential processes that rely on ubiquitin modification for function. Thus, isopeptidase inhibition by ⌬12-PGJ 2 could cause apoptosis by these and/or other mechanisms.
⌬12-PGJ 2 is the first isopeptidase inhibitor reported to exert its effects in intact cells. Nonhydrolyzable ubiquitin dimer analogs (ϳ16 kDa) (38) and ubiquitin aldehyde (ϳ8.5 kDa) (39,40) inhibitors are suitable for investigation of isolated ubiquitin isopeptidases, but not for investigations with intact cells. We are presently investigating whether ⌬12-PGJ 2 acts directly, via Michael adduct formation at their active site sulfhydryl, or indirectly by modulating cellular sulfhydryl-disulfide status analogous to the mechanism described by Liu et al. (41). We are also investigating whether selective inhibition exists between the various isopeptidase family members.
As a novel modulator of the ubiquitin-dependent proteasome pathway, ⌬12-PGJ 2 differs chemically and mechanistically from lactacystin (42), eponemycin (43), nitric oxide (44), 4-hydroxynonenal (45), and peptide-aldehyde or boronate inhibitors (26), all of which slow the disassembly of ubiquitin-protein conjugates by irreversibly modifying the 20 S catalytic subunit of the 26 S proteasome. The fact that agents as chemically diverse as these are each apoptotic agonists reinforces the proposition that the ubiquitin-proteasome pathway contains molecular targets for antineoplastic drug development (26). Importantly, our results suggest that inhibition of the isopeptidase component of this pathway causes apoptosis independently of p53-mediated gene transactivation; ⌬12-PGJ 2 inactivated p53-mediated transcription, caused apoptosis, and inhibited cellular isopeptidase activity under identical conditions. Furthermore, it caused apoptosis in cells expressing dysfunctional allelic variants of p53 (46,47). Agents that cause apoptosis independently of p53 are of considerable interest, because the majority of drugs used in oncology have lower efficacy in cells with mutant p53 compared with cells with wild type p53 (48). The response to chemotherapy is complex, and investigations focused on a single genetic factor, like p53, may exaggerate its role. However, numerous investigations show that disruption of p53 impairs the efficacy of p53-dependent drugs, e.g. 5-fluorouracil, used for a range of oncological disorders (49 -53).
⌬12-PGJ 2 can originate from rapid dehydration of its parent compound, PGD 2 (29,54). We conclude that the cross-conjugated dienone (⌬9,⌬12-olefin flanking the C11 carbonyl) accounts for the activity we describe. Naturally occurring A series PG with a single endocyclic ␣,␤-unsaturated ketone are not as efficacious as apoptotic agonists and isopeptidase inhibitors, compared with ⌬12-PGJ 2 . It is noteworthy that endogenous formation of PGD 2 , the parent of ⌬12-PGJ 2 , correlates inversely with the metastatic potential of murine malignant melanoma cells (55,56). Although we originally attributed this beneficial effect to PGD 2 and its inhibition of platelet-tumor cell aggregates, it is plausible that ⌬12-PGJ 2 might also have contributed as an apoptotic agonist.