Consequences of COP9 signalosome and 26S proteasome interaction Xiaohua Huang 1 , Bettina K. J. Hetfeld 1 , Ulrike Seifert 2 , Thilo Ka ¨ hne 3 , Peter-Michael Kloetzel 2 , Michael Naumann 3 , Dawadschargal Bech-Otschir 4 and Wolfgang Dubiel 1 1 Division of Molecular Biology, Department of Surgery, Charite ´ , Universita ¨ tsmedizin Berlin, Germany 2 Institute of Biochemistry, Charite ´ , Universita ¨ tsmedizin Berlin, Germany 3 Institut fu ¨ r Experimentelle Innere Medizin, Universita ¨ t Magdeburg, Germany 4 MRC Human Genetics Unit, Western General Hospital, Edinburgh, UK The COP9 signalosome (CSN) has been discovered in plant cells as a negative regulator of photomorphogen- esis [1]. It occurs in all eukaryotic cells and consists of eight core subunits, CSN1–CSN8 [2]. Six of the CSN subunits contain PCI (proteasome, COP9 signalosome, initiation factor 3) domains and two contain MPN (Mpr-Pad1-N-terminal) domains [3]. These two charac- teristic domains have been found in three protein com- plexes: the CSN, the 26S proteasome lid complex (lid) and the eukaryotic translation initiation factor 3 (eIF3) complex. The two domains are composed of about 150–200 amino acids at the N- or C-terminus of the CSN subunits. The PCI domain has been demon- strated to be important for interactions between CSN subunits. Thus, it might have scaffolding function [4,5]. The MPN + or JAMM domain of CSN5 is responsible for an intrinsic metalloprotease activity of the complex [6]. The function of the MPN domain of CSN6 is unknown. The CSN is associated with a large number of pro- teins [7], most of which are substrates or regulators of the ubiquitin (Ub) system. Analysis of associated Keywords COP9 signalosome; lid; p53; PCI domain; 26S proteasome Correspondence Division of Molecular Biology, Department of Surgery, Charite ´ , Universita ¨ tsmedizin Berlin, Monbijoustr. 2, 10117 Berlin, Germany Fax: +49 30 450522928 Tel: +49 30 450522305 e-mail: wolfgang.dubiel@charite.de (Received 9 May 2005, accepted 6 June 2005) doi:10.1111/j.1742-4658.2005.04807.x The COP9 signalosome (CSN) occurs in all eukaryotic cells. It is a regula- tory particle of the ubiquitin (Ub)⁄ 26S proteasome system. The eight sub- units of the CSN possess sequence homologies with the polypeptides of the 26S proteasome lid complex and just like the lid, the CSN consists of six subunits with PCI (proteasome, COP9 signalosome, initiation factor 3) domains and two components with MPN (Mpr-Pad1-N-terminal) domains. Here we show that the CSN directly interacts with the 26S proteasome and competes with the lid, which has consequences for the peptidase activity of the 26S proteasome in vitro. Flag-CSN2 was permanently expressed in mouse B8 fibroblasts and Flag pull-down experiments revealed the forma- tion of an intact Flag-CSN complex, which is associated with the 26S proteasome. In addition, the Flag pull-downs also precipitated cullins indi- cating the existence of super-complexes consisting of the CSN, the 26S pro- teasome and cullin-based Ub ligases. Permanent expression of a chimerical subunit (Flag-CSN2-Rpn6) consisting of the N-terminal 343 amino acids of CSN2 and of the PCI domain of S9 ⁄ Rpn6, the paralog of CSN2 in the lid complex, did not lead to the assembly of an intact complex showing that the PCI domain of CSN2 is important for complex formation. The consequence of permanent Flag-CSN2 overexpression was de-novo assem- bly of the CSN complex connected with an accelerated degradation of p53 and stabilization of c-Jun in B8 cells. The possible role of super-complexes composed of the CSN, the 26S proteasome and of Ub ligases in the regula- tion of protein stability is discussed. Abbreviations CSN, COP9 signalosome; PCI, proteasome-COP9 signalosome-initiation factor 3; MPN, Mpr-Pad1-N-terminal; Ub, ubiquitin. FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 3909 enzymatic activities implies that the CSN is a compo- nent of the Ub pathway. Originally CSN associated kinase activity has been described [8]. Recently a num- ber of kinases associated with the complex have been identified [9,10]. CSN associated kinases phosphorylate important cellular proteins such as p53 and regulate the stability of the tumour suppressor towards the Ub system [11]. It has been demonstrated that subunits of the eIF3 complex interact with components of the CSN [12,13]. The exact function of this interaction is unknown. Recently many groups found that the CSN is associated with Ub ligases in particular with cullin- based ubiquitinating enzyme complexes. Cullins 1–7 are scaffolding proteins forming a family of highly diverse Ub ligase complexes, which are responsible for the ubiq- uitination of important cell cycle regulators and tran- scription factors. So far it has been shown that the CSN interacts with cullin 1 to cullin 4 [14–18]. An important reason for the relationship between the CSN and the cullin-based Ub ligase complexes seems to be the intrin- sic metalloprotease activity of the CSN, which removes the Ub-like protein Nedd8 from cullins [6]. Cycles of neddylation and deneddylation of cullins seem to regu- late the ubiquitinating activity of the cullin-based Ub ligases [19]. The metalloprotease activity of CSN5 is also able to deubiquitinate proteins [15]. In addition, the CSN is associated with a deubiquitinating enzyme called Ubp12 in the fission yeast Schizosaccharomyces pombe [20], which counteracts autocatalytic degradation of components of cullin-based Ub ligases [20,21]. All described CSN interactions clearly indicate that the complex is a component of the Ub system. More- over, there are data on a direct interaction of the CSN with the proteolytic machinery of the Ub system, the 26S proteasome. Several years ago it was shown that the CSN cofractionates with the 26S proteasome from human cells [8]. The yeast two-hybrid screen revealed that the C-terminal domain of the Arabidopsis at CSN1 subunit interacts with at Rpn6 of the 26S proteasome lid [22]. Recently gel filtration size-fractionation of material from cauliflower in the presence of ATP and phosphatase inhibitors indicated that the CSN1 and CSN6 subunits coelute in the same fractions as subunits of the 26S regulatory complex. Moreover, coimmuno- precipitations revealed the existence of super-complexes consisting of the CSN, the 26S proteasome and cullin- based Ub ligase complexes [23]. Although these results indicate an association of the CSN and the 26S pro- teasome, the exact mode of this interaction, the role of PCI domains, and its consequences for the stability of cellular proteins is not known. Here we show for the first time that the CSN directly interacts with the 26S proteasome and that the purified human CSN has impact on 26S proteasome activity. The CSN seems to compete with the 26S pro- teasome lid. In cells permanently overexpressing CSN2 the amount of the CSN complex increases, which has consequences for the stability of p53 and c-Jun. Results Flag pull-downs with lysates from B8 fibroblasts permanently expressing Flag-CSN2 contain subunits of the 26S proteasome To study the interaction between the CSN and the 26S proteasome the human CSN2 cDNA was cloned into an eukaryotic expression vector coding for an N-ter- minal Flag-tag. The construct was permanently expressed in mouse B8 fibroblasts. Human and mouse CSN2 are identical on the amino acid level and there- fore we expected the integration into the mouse CSN. Interestingly, permanent expression of the Flag-CSN2 construct in HeLa cells was not successful, because cells died (data not shown). First it was tested whether the Flag-CSN2 was integ- rated into large protein complexes. Glycerol gradient centrifugation and subsequent western blots revealed that the Flag-CSN2 sediments into the same fractions as the CSN. In Fig. 1A, middle panel, two bands are seen with the anti-CSN2 Ig. The upper band corresponds to Flag-CSN2 and the lower one is endogenous CSN2, which occurred in a ratio of approximately 1 : 1. Both Flag-CSN2 and endogenous CSN2 were efficiently integrated into complexes. Due to de novo assembly the total amount of the CSN complex increased in B8 cells permanently expressing Flag-CSN2 (see below). To study CSN associated proteins, lysate of B8 cells expressing Flag-CSN2 was incubated with Flag-beads. After washing, bound proteins were specifically eluted with the Flag-peptide. The SDS ⁄ PAGE and subse- quent Coomassie stain of the Flag pull-down is shown in Fig. 1B. Selected bands were cut out and analyzed by mass spectrometry revealing the presence of all core CSN subunits. The eluted proteins were analyzed under nondenaturing conditions. In a nondenaturing gel followed by western blotting the complex migrates exactly to the position of the CSN. It can be detected by the anti-Flag as well as by the anti-CSN3 Ig. A smear was detected in the region of the 20S ⁄ 26S pro- teasome with the anti-Flag Ig (Fig. 1C). Previously it has been shown that the purified CSN phosphorylates c-Jun and p53 by CSN associated kin- ases [10,24]. To test whether eluted proteins were able to phosphorylate c-Jun and p53, kinase assays were performed. As shown in Fig. 1D, the two proteins as COP9 signalosome and 26S proteasome interaction X. Huang et al. 3910 FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS well as recombinant CSN2, another substrate of CSN associated kinases [25], were phosphorylated in a curcumin-sensitive manner; curcumin being a typical inhibitor of CSN associated kinases [10,26]. To study association of the CSN with the 26S pro- teasome western blots were performed. The data are summarized in Fig. 1E. First proteins of the Flag pull- downs were probed with antibodies against subunits of the CSN (Fig. 1E, left panel). All subunits of the CSN tested were detected supporting our data obtained by mass spectrometry (Fig. 1B). Figure 1E (right panel) shows the western blots with antibodies against sub- units of the 26S proteasome. In all Flag pull-downs the components of the 26S proteasome base S1 ⁄ Rpn2, S4 ⁄ Rpt2, S6b ⁄ Rpt3 and S6a ⁄ Rpt5 and the 20S protea- some were clearly identified. Under our conditions the subunits of the lid S10a ⁄ Rpn7 and S12 ⁄ Rpn8, but not S13 ⁄ Rpn11, were detected. There were no unspecific proteins eluted from the Flag-beads as demonstrated by control pull-downs with lysate from B8 cells. The CSN binds directly to the 26S proteasome and most likely competes with the lid It has been shown that the CSN forms super-com- plexes with the 26S proteasome and with cullin-based AB CD E Fig. 1. Flag-CSN2 permanently expressed in B8 cells is integrated into an intact CSN complex, which interacts with the 26S pro- teasome. (A) Glycerol gradient centrifugation was performed with lysates of B8 cells expressing Flag-CSN2. Subsequent western blotting with glycerol gradient fractions using antibodies against Flag, CSN2 and the 20S proteasome (20S) revealed sedimenta- tion of the Flag-CSN2 into the same frac- tions as the CSN complex. The asterisk indicates that the anti-CSN2 Ig interacts with the Flag-CSN2 (upper band) as well as with endogenous CSN2 (lower band). (B) Mass spectrometry of selected bands from SDS ⁄ PAGE of Flag-CSN2 pull-downs. (C) Flag-CSN2 pull downs were analyzed by nondenaturing gel electrophoresis. Subse- quently proteins were blotted to nitrocellu- lose and tested with anti-Flag and anti-CSN3 Igs. Positions of the CSN, the 20S and the 26S proteasome were determined with spe- cific antibodies. (D) Flag-CSN2 pull-downs were used as a source of kinase activity in kinase assays. Recombinant CSN2, p53 and c-Jun were used as substrates. The reaction was inhibited by the kinase inhibitor curcu- min. (E) Western blot analyses with antibod- ies against CSN subunits (left panel) and antibodies against 26S proteasome subunits (right panel) were performed with lysate from B8 cells (controls) and lysate from B8 cells permanently expressing Flag-CSN2. X. Huang et al. COP9 signalosome and 26S proteasome interaction FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 3911 Ub ligases [23]. However, direct interaction between the CSN and the 26S enzyme had not been demonstra- ted so far. Therefore, we performed in vitro immuno- precipitations with isolated human CSN and 26S proteasome. The two purified particles were preincu- bated with a molar ration of 1 : 1 for 30 min in the presence of ATP. Then immunoprecipitation was per- formed with the anti-CSN7 Ig or with preimmune serum as a control. The data are shown in Fig. 2A. The western blot of the precipitate revealed the coim- munoprecipitation of the CSN and S1 ⁄ Rpn2 of the 26S proteasome indicating a direct interaction of the two complexes. To test whether the CSN competes with the 26S pro- teasome lid complex purified CSN and 26S proteasome were incubated as in Fig. 2A using different molar rations of the two complexes. After incubation immuno- precipitations with a monoclonal antibody against the 20S proteasome subunit a6 ⁄ C2 were performed. The data in Fig. 2B demonstrate that CSN1 and CSN5 can be well detected in immunoprecipitates after incuba- tion with a 20-fold molar excess of the CSN. Of note, after long-term exposure CSN1 and CSN5 were also seen in samples with 1 : 1 ratios (data not shown). In contrast, the lid subunit S10a ⁄ Rpn7 was well seen in the absence of the CSN, but was not detectable at a molar ratio of 1 : 20. The base subunit S4 ⁄ Rpt2 and the 20S proteasome did not change depending on the molar ratios of 26S ⁄ CSN (Fig. 2B). Based on these data we speculated that the CSN might replace the lid. Possible competition between the CSN and the lid complexes might be also reflected by changed proteasome activity. To see whether the direct interaction of the CSN with the 26S proteasome has an effect on proteasome peptidase activity, assays with purified CSN and 26S proteasome and with succinyl- Leu-Leu-Val-Tyr-AMC as substrate were carried out. As shown in Fig. 2C, measured fluorescence revealed that 26S proteasome peptidase activity in the presence of ATP is slightly inhibited by a molar excess of the CSN. There was no effect detected without ATP. Ubiquitinated proteins are the physiological substrates of the 26S proteasome. Unfortunately, because of the high deubiquitinating activity associated with the CSN [20], the impact of the CSN on the degradation of model ubiquitinated substrates by the 26S proteasome was difficult to estimate under in vitro condition. Sub- strates were quickly deubiquitinated before the 26S proteasome had a chance to degrade them (data not shown). The PCI domain of CSN2 is essential for CSN complex assembly and its interaction with the 26S proteasome The paralog subunit of CSN2 in the lid complex is Rpn6. We were interested to see whether substitution of the PCI domain of CSN2 by the PCI domain of Rpn6 has consequences for the complex integration of the chimerical CSN2-Rpn6 protein. The cDNA enco- ding the first 343 amino acids of human CSN2 and the PCI domain of human Rpn6 (Rpn6 amino acids 291– 422) were linked together and cloned into an eukaryotic expression vector possessing a N-terminal Flag-tag (Fig. 3A). The construct was permanently expressed in B8 cells. In glycerol gradients the Flag-chimerical pro- tein sedimented into similar fractions as the CSN, which is shown by western blotting in Fig. 3B. Again, AB C Fig. 2. Direct interaction between the CSN and the 26S protea- some in vitro. (A) Co-immunoprecipitation of the CSN and the 26S proteasome in vitro using the anti-CSN7 Ig or preimmune serum (control). Purified CSN and 26S proteasome were incubated at a molar ratio of 1 : 1 in the presence of ATP. The precipitate was analyzed by western blotting with the anti-S1 ⁄ Rpn2 Ig. (B) Co-im- munoprecipitations of the CSN and the 26S proteasome using the monoclonal anti-a6 ⁄ C2 Ig. Isolated 26S proteasome and different amounts of the purified CSN were incubated for 30 min in the pres- ence of ATP. After incubation immunoprecipitations were per- formed and precipitates were analyzed by western blotting using anti-CSN1, anti-CSN5, anti-S10a ⁄ Rpn7, anti-S4 ⁄ Rpt2 and anti-20S proteasome Igs. (C) Fluorescence was measured with isolated 26S proteasome (0.15 pmol per sample) in the presence of succinyl- Leu-Leu-Val-Tyr-MCA as substrate with or without ATP. Purified CSN was added in molar rations indicated. COP9 signalosome and 26S proteasome interaction X. Huang et al. 3912 FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS the anti-CSN2 Ig revealed two bands, the Flag-CSN2- Rpn6 (upper band) and the endogenous CSN2 (lower band), which were integrated into complexes. In this case the expression of the chimerical protein was signi- ficantly less than that of the endogenous CSN2. Flag pull-downs were performed as described above to determine: (a) if the chimerical CSN2-Rpn6 protein was integrated into an intact CSN or into the 26S proteasome lid complex; (b) if there was an interaction with the 26S proteasome? Proteins specifically eluted with the Flag-peptide were analyzed with antibodies against subunits of the CSN as well as the 26S protea- some. According to the data shown in Fig. 3C the Flag-CSN2-Rpn6 protein was not integrated into either an intact CSN or lid complex. The Flag pull- downs contained significant amounts of CSN1 protein, but only S1 ⁄ Rpn2 and traces of S4 ⁄ Rpt2 indicating that there is no interaction with the 26S proteasome complex. Are there CSN-26S proteasome super- complexes? We were interested to see whether the Flag pull-downs contain additional B8 cell proteins besides the CSN and the 26S proteasome. Therefore western blots were carried out after Flag pull-downs using antibodies against p53, c-Jun, cullin 1, cullin 3, Ub, associated kinase CK2a subunit and a subunit of the eIF3 com- plex, INT6. The large number of blots is not shown and the data are summarized in Table 1. Positive reac- tions are indicated. Again, controls with B8 cell lysate alone showed that no unspecific proteins were eluted from the Flag-beads. In pull-downs with the wild-type Flag-CSN2, p53 and c-Jun, two typical substrates of the CSN [24], were detected. In addition, cullin 1 and cullin 3 were found, suggesting an association with cullin-based Ub ligase complexes. These results confirm earlier observations on the existence of super-complexes consisting of the CSN, the 26S proteasome and cullin- based Ub ligases [23]. The anti-Ub Ig reacted with high-molecular weight material indicating the binding of Ub conjugates. The CSN interaction with CK2 [10] and with INT6 [12] has been published before. Table 1 shows that basically none of the tested proteins, except p53, interacted with the Flag-CSN2-Rpn6 chimerical A B C Fig. 3. The PCI domain of CSN2 is essential for CSN complex for- mation. (A) The Flag-CSN2-Rpn6 construct codes for the first 343 amino acids of CSN2 and for the PCI domain of its lid paralog Rpn6 (amino acids 291–422). (B) The Flag-CSN2-Rpn6 chimera was stably expressed in B8 cells and the cell lysate was analyzed by glycerol gradients and subsequent western blotting. The asterisk indicates that the anti-CSN2 Ig interacts with both the upper Flag-CSN2- Rpn6 protein and the lower endogenous CSN2. (C) Western blot analyses with antibodies against CSN subunits (left panel) and anti- bodies against 26S proteasome subunits (right panel) were per- formed with lysate from B8 cells (controls) and lysate from B8 cells permanently expressing the Flag-CSN2-Rpn6 chimera. Table 1. Proteins detected in Flag pull-downs by western blotting. Flag pull-downs were performed as described (Experimental proce- dures). Eluted proteins were separated by SDS ⁄ PAGE, blotted to nitrocellulose and probed with the antibodies indicated in the table. The + symbol indicates a positive antibody reaction; ND, not deter- mined. Antibodies Flag-CSN2 Flag-CSN2-Rpn6 B8 Anti-p53 + + – Anti-c-Jun + – – Anti-cullin 1 + – – Anti-cullin 3 + – – Anti-ubiquitin + – – Anti-CK2a +ND – Anti-INT6 + ND – X. Huang et al. COP9 signalosome and 26S proteasome interaction FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 3913 protein demonstrating that the intact CSN is essential for most bindings including the interaction with the 26S proteasome and with cullin-based Ub ligases. Changes of the CSN and of CSN substrates in B8 cells permanently expressing Flag-CSN2 or Flag-CSN2-Rpn6 chimera To test whether the CSN was modified in fibroblast per- manently expressing Flag-CSN2 or Flag-CSN2-Rpn6 western blots with cell lysates were performed using antibodies against CSN subunits. As demonstrated in Fig. 4A, permanent expression of wild-type Flag-CSN2 led to elevated levels of CSN3 and CSN5 subunits in B8 cells suggesting a de novo assembly of the CSN complex. In contrast, expression of the Flag-CSN2-Rpn6 chimera did not change the amount of CSN subunits in B8 cells as compared with control cells. Typical CSN substrates p53, c-Jun, p27 and IjBa are phosphorylated by the associated kinases, which regulate the stability of the proteins [24]. Therefore the influence of Flag-CSN2 or Flag-CSN2-Rpn6 expres- sion on the stability of these proteins in B8 cells was studied by western blotting. As shown in Fig. 4B, sig- nificant changes of p53 and c-Jun levels were detected in B8 cells permanently expressing Flag-CSN2 as compared to control cells. While p53 in Flag-CSN2 B8 cells almost completely disappeared, c-Jun was clearly stabilized. The impact on p27 and IjBa steady state levels in B8 cells is less pronounced as compared to p53 and c-Jun. Discussion Here we show for the first time that the CSN directly interacts with the 26S proteasome. It can compete with the lid, which has consequences for 26S proteasome peptidase activity. In addition, we demonstrate the essential role of the PCI domain of CSN2 for complex formation and consequences of permanently overex- pressed CSN2 for the stability of p53 and c-Jun. The CSN interacts directly with the 26S proteasome and influences proteasome cleavage activity Although the exact mode of CSN ⁄ 26S proteasome inter- action is still obscure, it has been speculated that the CSN might be an alternative lid [27]. It has been known for many years that the CSN copurifies with subunits of the 26S proteasome [8] and analyses by mass spectrome- try revealed components of the 26S base in our CSN preparation from red blood cells (our unpublished data). To study CSN ⁄ 26S interaction, Flag-CSN2 was perma- nently expressed in mouse B8 cells. Data show that human CSN2 with a Flag-tag at its N-terminus was integrated into a complete mouse CSN complex. This is not surprising, as mouse and human CSN2 protein are 100% identical. No unspecific proteins were detected in our control pull downs with B8 cell lysate and using the Flag-peptide for specific elution. Western blot analysis of Flag-CSN2 pull-downs revealed the presence of 20S core particle and 26S proteasome base subunits. Under our conditions, it was difficult to detect components of the lid (Fig. 1E, right panel). The direct interaction between the CSN and the 26S proteasome is shown by in vitro coimmunoprecipitation (Fig. 2A). A possible competition between the CSN and the lid is demonstra- ted by immunoprecipitations after incubations of the 26S proteasome and the CSN at different molar ratios. The data shown in Fig. 2B demonstrate that a molar excess of the CSN most likely replaces the lid. Because of significant sequence homologies between the compo- nents of the CSN and the lid, it is likely that the two complexes can be substituted by and compete with each other. Competition is also indicated by measuring pepti- dase activity with a fluorogenic peptide in the presence of purified human 26S proteasome and purified human CSN. Increasing amounts of the CSN slightly sup- AB Fig. 4. Permanent expression of Flag-CSN2 causes de novo assem- bly of the CSN complex in B8 cells connected with degradation of endogenous p53 and stabilization of c-Jun. (A) Lysates of B8 cells (controls), B8 cells permanently expressing Flag-CSN2-Rpn6 and B8 cells permanently expressing Flag-CSN2 were tested by western blotting using antibodies against CSN3 and CSN5. (B) Lysates of B8 cells (controls), B8 cells permanently expressing Flag-CSN2- Rpn6 and B8 cells permanently expressing Flag-CSN2 were tested by western blotting using antibodies against Flag, p53, c-Jun, p27 and IjBa. The anti-actin Ig was used as an internal control demon- strating equal loading of proteins. COP9 signalosome and 26S proteasome interaction X. Huang et al. 3914 FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS pressed 26S proteasome peptidase activity indicating a direct interaction with the 26S complex. The effect of the CSN on 26S enzyme activity can be explained by a conformational change caused through the replacement of the lid by the CSN. At the moment it is unclear whe- ther this effect of the CSN on 26S proteasome peptidase activity has physiological relevance. The fraction of CSN particles associated with the 26S proteasome and vice versa is small. As seen in the nondenaturing gel in Fig. 1C, most of Flag-CSN2 integrated into the free CSN complex and only small amounts migrated into regions of the 26S proteasome. This is also reflected by the fact that immunoprecipita- tions in cell lysates using antibodies against CSN or 26S components sometimes failed to detect coimmuno- precipitation of the two complexes. With phosphatase inhibitors and ATP included in buffers a fraction of 5–10% of CSN was estimated to be associated with the 26S proteasome in plant cells [23]. Under our con- ditions without adding ATP or inhibitors of phospha- tases this fraction seems to be below 5%. The role of the PCI domain of CSN2 for complex formation It has been shown before that PCI domains are important for CSN complex formation [4,28]. How- ever, the exact role of PCI domains in the complex assembly process or in subunit targeting to the right complex is still obscure. Here we demonstrate that the PCI domain of CSN2 is essential for the formation of an intact CSN particle and subsequently for the forma- tion of super-complexes consisting of the CSN, the 26S proteasome and Ub ligases. The rationale for generating a chimerical protein consisting of the N-terminal part of CSN2 and the PCI domain of its paralog lid subunit, Rpn6, was to test, whether the PCI domain has the information for targeting it to the right complex. Our data revealed that the PCI domain of Rpn6 is not sufficient to serve as an address for the lid complex. The chimerical pro- tein did not show any interaction with lid subunits. In contrast, the CSN2-Rpn6 chimera interacted with CSN1 obviously independently of the CSN2-PCI domain. The formation of a CSN1–CSN2 subcomplex with specific function in cell cycle would explain the exclusive phenotypes obtained with csn1 and csn2 dele- tions in S. pombe. Knockouts of the two subunits, but not of other CSN subunits, cause cell cycle delay in S-phase [14,29]. In csn1 and csn2 deletion mutants the cell cycle inhibitor Spd1 accumulates causing Suc22-dependent suppression of ribonucleotide reduc- tase connected with S-phase delay and DNA damage sensitivity [14]. The existence of a CSN1–CSN2 sub- complex has to be verified in the future. Possible functions of CSN-based super-complexes The presented data confirm our earlier findings that overexpression of Flag-CSN2 leads to de novo assembly of the CSN complex followed by the stabilization of c-Jun transcription factor [30]. In addition, here we show that an increase of the CSN in B8 cells signifi- cantly accelerated the degradation of the tumour sup- pressor p53 (Fig. 4B). This is not surprising, as the CSN targets p53 to degradation by the Ub system [11] and increased amounts of the CSN accelerate the degrada- tion. Currently the question whether this process is mediated by super-complexes consisting of the CSN, the 26S proteasome and Ub ligases cannot be answered. At the moment two hypothesis on the function of the super-complexes can be distinguished. First, the super- complexes are proteolytic machineries for the degrada- tion of a certain set of substrates, which are channelled from substrate labelling by phosphorylation and ubiqui- tination to complete proteolytic cleavage. In this model the CSN would act as an alternative lid or a platform bringing together specific Ub ligases and the 26S protea- some. Second, the CSN is a platform that allows Ub ligase re-assembly. This hypothesis is based on an idea by Wolf and coworkers assuming that the CSN blocks cullin-based complex activity including auto-ubiquiti- nation and provides an environment necessary for the assembly of new cullin-based complexes [21,31]. Accord- ing to the second hypothesis one would expect that association of the 26S proteasome to the super-complex might also cause inhibition of the protease to protect cullins and other components from degradation. How- ever, our data indicate that the CSN does not efficiently inhibit the 26S proteasome activity in vitro. In addition, elevated CSN amounts in B8 cells did not cause a gen- eral inhibition of Ub-dependent proteolysis. Therefore we favour the first model in which super-complexes are large proteolytic machines that carry out specific proteo- lysis. Future work is necessary to fully understand the function of CSN ⁄ 26S proteasome interaction and of the super-complexes. Experimental procedures Materials The kinase inhibitor curcumin was obtained from Sigma. Antibodies against CSN5 (a gift from B. Christy), Ub (Dako, Glostrup, Denmark), Flag (Sigma, St Louis, Missouri, USA), 20S proteasome (Affiniti ⁄ Biomol, Hamburg, X. Huang et al. COP9 signalosome and 26S proteasome interaction FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 3915 Germany), p53 (BD Biosciences, San Jose, CA, USA), c-Jun as well as CK2a (Calbiochem, Schwalbach, Germany), p27 as well as IjBa (Santa Cruz, CA, USA) and cullin 1 (Onco- gene, Schwalbach, Germany) were used in western blots. The anti-S1 ⁄ Rpn2 Ig was a gift from K. Hendil, August Krogh Institute, Copenhagen, Denmark and the anti-INT6 Ig was a gift from C. Norbury, University of Oxford, UK. Preparations, assays and immunoprecipitation Preparation procedure for the 26S proteasome as well as the CSN from human red blood cells and kinase assays were described before [8]. Peptidase assays with the purified proteasome and with succinyl-Leu-Leu-Val-Tyr-AMC as substrate were outlined earlier [32]. Mass spectrometry was performed as described [8]. In vitro immunoprecipitation was carried out with 2.3 pmol of the CSN as well as the 26S proteasome. First the two particles were incubated in the presence of 2 mm ATP at 37 °C for 30 min. The immuno- precipitation with the anti-CSN7 Ig or with the preimmune serum was carried out as before [10]. Cell culture, Flag pull-downs B8 mouse fibroblast cells were cultured using Iscove’s MEM (Biochrom, Berlin, Germany) with 125 lgÆmL )1 G418. Stable transfected B8 cells were established using cal- cium phosphate precipitation and selected with 1 lgÆmL )1 puromycin. Human CSN2 and human CSN2-Rpn6 chimera cDNAs were cloned into pcDNA3.1 vector (Invitrogen, Carlsbad, CA, USA) coding for an N-terminal Flag-tag. Expression of Flag-CSN2 or Flag-CSN2-Rpn6 protein was tested by western blots with an anti-Flag Ig. Flag pull-downs with B8 cells were performed as recom- mended by the manufacturer (Sigma). Briefly, stably trans- fected cells were rinsed twice with ice-cold 1· NaCl ⁄ P i and collected. Ice-cold lysis buffer (50 mm Tris ⁄ HCl pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% Triton X-100) with freshly added phenylmethylsulfonyl fluoride (1 mgÆmL )1 ) was added to the cells on ice. After centrifugation at 15 000 g for 10 min at 4 °C supernatants were loaded onto the pre- pared ANTI-FLAG M2 affinity column. After washing with 20 column volumes of 1· TBS (50 mm Tris ⁄ HCl pH 7.4, 150 mm NaCl, 1 mm EDTA), proteins were eluted by competition with the Flag peptide (100 lgÆmL )1 ). Eluted proteins were used for western blots, nondenaturing electrophoresis and kinase assay. Glycerol gradients, nondenaturing electrophoresis and western blots Glycerol gradient centrifugation was performed as outlined before [32]. For nondenaturing electrophoresis 2 lL of Flag pull-downs were separated on a 4–15% (w ⁄ v) Phast-gel (Pharmacia Biotech., Inc.) at 300 VÆh )1 . Proteins were blot- ted onto nitrocellulose and probed with an anti-Flag or anti-CSN3 Ig. All western blots were developed by ECL technique (Amersham, Buckinghamshire, UK). Acknowledgements This work was supported by a grant from the G.I.F., the German-Israeli Foundation for Scientific Research and Development, and grant DU 229 ⁄ 6–2 from the Deutsche Forschungsgemeinschaft to W. D. References 1 Wei N, Chamovitz DA & Deng XW (1994) Arabidopsis COP9 is a component of a novel signaling complex mediating light control of development. Cell 78, 117– 124. 2 Deng XW, Dubiel W, Wei N, Hofmann K, Mundt K, Colicelli J, Kato J, Naumann M, Segal D, Seeger M et al. (2000) Unified nomenclature for the COP9 sig- nalosome and its subunits: an essential regulator of development. Trends Genet 16, 202–203. 3 Hofmann K & Bucher P (1998) The PCI domain: a common theme in three multiprotein complexes. Trends Biochem Sci 23, 204–205. 4 Tsuge T, Matsui M & Wei N (2001) The subunit 1 of the COP9 signalosome suppresses gene expression through its N-terminal domain and incorporates into the complex through the PCI domain. J Mol Biol 305, 1–9. 5 Kim T, Hofmann K, von Arnim AG & Chamovitz DA (2001) PCI complexes: pretty complex interactions in diverse signaling pathways. Trends Plant Sci 6, 379–386. 6 Cope GA, Suh GS, Aravind L, Schwarz SE, Zipursky SL, Koonin EV & Deshaies RJ (2002) Role of predicted metalloprotease motif of Jab1 ⁄ Csn5 in cleavage of Nedd8 from Cul1. Science 298, 608–611. 7 Wei N & Deng XW (2003) The COP9 signalosome. Annu Rev Cell Dev Biol 19, 261–286. 8 Seeger M, Kraft R, Ferrell K, Bech-Otschir D, Dumdey R, Schade R, Gordon C, Naumann M & Dubiel W (1998) A novel protein complex involved in signal trans- duction possessing similarities to 26S proteasome subunits. FASEB J 12, 469–478. 9 Wilson MP, Sun Y, Cao L & Majerus PW (2001) Ino- sitol 1,3,4-trisphosphate 5 ⁄ 6-kinase is a protein kinase that phosphorylates the transcription factors c-Jun and ATF-2. J Biol Chem 276, 40998–41004. 10 Uhle S, Medalia O, Waldron R, Dumdey R, Henklein P, Bech-Otschir D, Huang X, Berse M, Sperling J, Schade R et al. (2003) Protein kinase CK2 and protein kinase D are associated with the COP9 signalosome. EMBO J 22, 1302–1312. COP9 signalosome and 26S proteasome interaction X. Huang et al. 3916 FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 11 Bech-Otschir D, Kraft R, Huang X, Henklein P, Kape- lari B, Pollmann C & Dubiel W (2001) COP9 signalo- some-specific phosphorylation targets p53 to degradation by the ubiquitin system. EMBO J 20, 1630–1639. 12 Hoareau Alves K, Bochard V, Rety S & Jalinot P (2002) Association of the mammalian proto-oncoprotein Int-6 with the three protein complexes eIF3, COP9 sig- nalosome and 26S proteasome. FEBS Lett 527, 15–21. 13 Yahalom A, Kim TH, Winter E, Karniol B, von Arnim AG & Chamovitz DA (2001) Arabidopsis eIF3e (INT- 6) associates with both eIF3c and the COP9 signalo- some subunit CSN7. J Biol Chem 276, 334–340. 14 Liu C, Powell KA, Mundt K, Wu L, Carr AM & Caspari T (2003) Cop9 ⁄ signalosome subunits and Pcu4 regulate ribonucleotide reductase by both checkpoint- dependent and -independent mechanisms. Genes Dev 17, 1130–1140. 15 Groisman R, Polanowska J, Kuraoka I, Sawada J, Saijo M, Drapkin R, Kisselev AF, Tanaka K & Nakatani Y (2003) The ubiquitin ligase activity in the DDB2 and CSA complexes is differentially regulated by the COP9 signalosome in response to DNA damage. Cell 113, 357–367. 16 Lyapina S, Cope G, Shevchenko A, Serino G, Tsuge T, Zhou C, Wolf DA, Wei N & Deshaies RJ (2001) Pro- motion of NEDD-CUL1 conjugate cleavage by COP9 signalosome. Science 292, 1382–1385. 17 Schwechheimer C, Serino G, Callis J, Crosby WL, Lyapina S, Deshaies RJ, Gray WM, Estelle M & Deng XW (2001) Interactions of the COP9 signalosome with the E3 ubiquitin ligase SCFTIRI in mediating auxin response. Science 292, 1379–1382. 18 Geyer R, Wee S, Anderson S, Yates J & Wolf DA (2003) BTB ⁄ POZ domain proteins are putative substrate adaptors for cullin 3 ubiquitin ligases. Mol Cell 12 , 783– 790. 19 Cope GA & Deshaies RJ (2003) COP9 signalosome: a multifunctional regulator of SCF and other cullin-based ubiquitin ligases. Cell 114, 663–671. 20 Zhou C, Wee S, Rhee E, Naumann M, Dubiel W & Wolf DA (2003) Fission yeast COP9 ⁄ signalosome sup- presses cullin activity through recruitment of the deubiquitylating enzyme Ubp12p. Mol Cell 11, 927–938. 21 Wee S, Geyer RK, Toda T & Wolf DA (2005) CSN facilitates Cullin-RING ubiquitin ligase function by counteracting autocatalytic adapter instability. Nat Cell Biol 7, 387–391. 22 Kwok SF, Staub JM & Deng XW (1999) Characteriza- tion of two subunits of Arabidopsis 19S proteasome regulatory complex and its possible interaction with the COP9 complex. J Mol Biol 285, 85–95. 23 Peng Z, Shen Y, Feng S, Wang X, Chitteti BN, Vierstra RD & Deng XW (2003) Evidence for a physical associa- tion of the COP9 signalosome, the proteasome, and spe- cific SCF E3 ligases in vivo. Curr Biol 13, R504–R505. 24 Bech-Otschir D, Seeger M & Dubiel W (2002) The COP9 signalosome: at the interface between signal transduction and ubiquitin-dependent proteolysis. J Cell Sci 115, 467–473. 25 Kapelari B, Bech-Otschir D, Hegerl R, Schade R, Dumdey R & Dubiel W (2000) Electron microscopy and subunit–subunit interaction studies reveal a first architecture of COP9 signalosome. J Mol Biol 300, 1169–1178. 26 Henke W, Ferrell K, Bech-Otschir D, Seeger M, Schade R, Jungblut P, Naumann M & Dubiel W (1999) Com- parison of human COP9 signalsome and 26S protea- some lid. Mol Biol Rep 26, 29–34. 27 Li L & Deng XW (2003) The COP9 signalosome: an alternative lid for the 26S proteasome? Trends Cell Biol 13, 507–509. 28 Freilich S, Oron E, Kapp Y, Nevo-Caspi Y, Orgad S, Segal D & Chamovitz DA (1999) The COP9 signalo- some is essential for development of Drosophila melano- gaster. Curr Biol 9, 1187–1190. 29 Mundt KE, Porte J, Murray JM, Brikos C, Christensen PU, Caspari T, Hagan IM, Millar JB, Simanis V, Hof- mann K et al. (1999) The COP9 ⁄ signalosome complex is conserved in fission yeast and has a role in S phase. Curr Biol 9, 1427–1430. 30 Naumann M, Bech-Otschir D, Huang X, Ferrell K & Dubiel W (1999) COP9 signalosome-directed c-Jun activation ⁄ stabilization is independent of JNK. J Biol Chem 274, 35297–35300. 31 Wolf DA, Zhou C & Wee S (2003) The COP9 signalo- some: an assembly and maintenance platform for cullin ubiquitin ligases? Nat Cell Biol 5, 1029–1033. 32 Huang X, Seifert U, Salzmann U, Henklein P, Preissner R, Henke W, Sijts AJ, Kloetzel PM & Dubiel W (2002) The RTP site shared by the HIV-1 Tat protein and the 11S regulator subunit alpha is crucial for their effects on proteasome function including antigen processing. J Mol Biol 323, 771–782. X. Huang et al. COP9 signalosome and 26S proteasome interaction FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS 3917 . degradation of p53 and stabilization of c-Jun in B8 cells. The possible role of super-complexes composed of the CSN, the 26S proteasome and of Ub ligases. loading of proteins. COP9 signalosome and 26S proteasome interaction X. Huang et al. 3914 FEBS Journal 272 (2005) 3909–3917 ª 2005 FEBS pressed 26S proteasome