COP9 signalosome interacts ATP-dependently with p97/valosin-containing protein (VCP) and controls the ubiquitination status of proteins bound to p97/VCP

Abstract

Ubiquitinated proteins can alternatively be delivered directly to the proteasome or via p97/VCP (valosin-containing protein). Whereas the proteasome degrades ubiquitinated proteins, the homohexameric ATPase p97/VCP seems to control the ubiquitination status of recruited substrates. The COP9 signalosome (CSN) is also involved in the ubiquitin/proteasome system (UPS) as exemplified by regulating the neddylation of ubiquitin E3 ligases. Here, we show that p97/VCP colocalizes and directly interacts with subunit 5 of the CSN (CSN5) in vivo and is associated with the entire CSN complex in an ATP-dependent manner. Furthermore, we provide evidence that the CSN and in particular the isopeptidase activity of its subunit CSN5 as well as the associated deubiquitinase USP15 are required for proper processing of polyubiquitinated substrates bound to p97/VCP. Moreover, we show that in addition to NEDD8, CSN5 binds to oligoubiquitin chains in vitro. Therefore, CSN and p97/VCP could form an ATP-dependent complex that resembles the 19 S proteasome regulatory particle and serves as a key mediator between ubiquitination and degradation pathways.

FIGURE 1.

CSN5 interacts with p97/VCP in vivo and in vitro. A, lysates from NIH 3T3 and HEK 293T cells (Input, lane 1) were used for IP with anti-CSN5 antibody (lane 2) and isotype control antibody (Ctr., lane 3) and analyzed by immunoblotting (IB). B, reciprocal IP was performed using anti-p97/VCP antibody (lane 2). Immunoprecipitates were immunoblotted for the detection of p97/VCP and CSN5. C, CSN5 expression was analyzed in NIH 3T3 cells transfected with empty vector (lane 1) or FLAG-p97/VCP (lane 2). In the supernatant of lysed cells, FLAG-p97/VCP (lane 4) was pulled down using anti-FLAG antibody. Coprecipitated CSN5 was detected by immunoblotting. D, GST-CSN5 (1 μg) immobilized on glutathione-Sepharose 4B beads was incubated with increasing amounts of purified His-p97/VCP (lanes 1–4). Binding of His-p97/VCP was detected by immunoblotting (lanes 1–4, top panel). As control, unloaded beads were incubated with 2 μg of His-p97/VCP (lane 5). Input of His-p97/VCP: 0.25 μg (lane 1), 0.5 μg (lane 2), 1 μg (lane 3), and 2 μg (lanes 4 and 5).

FIGURE 2.

CSN5 and p97/VCP colocalize in vivo. FRET-CLSM and double-labeling indirect immunofluorescence were used to detect a close in vivo association of CSN5 and p97/VCP in NIH 3T3 cells ectopically expressing both proteins (9). In acceptor bleaching experiments, the fluorescence of the donor (p97/VCP labeled with Cy3-conjugated secondary antibody; A and B) and of the acceptor (CSN5 labeled with Cy5-conjugated secondary antibody; C and D) was visualized and quantified in a defined ROI. Scale bar, 20 μm. After photobleaching of the acceptor fluorophore (compare ROI 1 in C and D), a significant increase in donor fluorescence in Cy3 was detected only in the ROI that was chosen for acceptor photobleaching (compare ROI 1 in A and B). ROI 2–5 are control regions outside the bleached area. 18–20 cells in each experiment (n = 76 for the experimental group and n = 82 for the control group (see below) obtained from four different experiments) were analyzed. A positive FRET signal in the bleached ROI, calculated as increase in fluorescence (ΔIF), was robustly detected in every experiment with a median increase of 7.9 (E). ***, p ≤ 0.001, Mann-Whitney U test. Boxplots: percentiles 0, 25, median, 75, 100. Open circles denote extreme values of the data set. A false-positive FRET signal that can be caused by cross-reactivity of secondary antibodies was excluded by performing control experiments in which both secondary antibodies were applied together with Cy5-labeled anti-CSN5 antibody (acceptor) only (median ΔIF = 2.8). Significant FRET occurred only in the cytoplasm, indicating that p97/VCP and CSN5 must be closer than 10 nm in this compartment, whereas no indication of colocalization in the nucleus was found (data not shown). CSN5 is distributed evenly over the nucleus and cytoplasm, whereas p97/VCP was mainly located in the cytoplasm.

FIGURE 3.

A, FLAG-p97/VCP interacts with the N-terminal domain of CSN5. Plasmids expressing FLAG-p97/VCP, Myc-CSN5 (wild type (wt) 1–334), or the deletion mutant Myc-1–191, Myc-1–110, or Myc-110–191 (47) were cotransfected into HEK 293T cells. Expression of FLAG-p97/VCP and Myc-tagged CSN5 proteins was detected by immunoblotting (IB) before (Input) and after IP of FLAG-p97/VCP. The asterisk denotes IgG heavy chains. B, interaction between CSN5 and p97/VCP is independent of a functional JAMM motif. Plasmids expressing FLAG-p97/VCP and a Myc-CSN5 mutant with point mutations in the JAMM motif (see “Experimental Procedures”) were cotransfected into HEK 293T cells. Expression of FLAG-p97/VCP and Myc-tagged CSN5 mutant was detected by immunoblotting before (Input) and after IP of Myc-CSN5 mutant (Myc-mCSN5, left panel) and FLAG-p97/VCP (right panel). C, all domains of p97/VCP contribute to the interaction interface with CSN5/CSN. Wild type p97/VCP or truncated proteins (see schematic drawing of constructs) were expressed as FLAG-tagged fusion proteins in HEK 293T cells together with CSN5. Expression of FLAG-tagged p97/VCP proteins and Myc-tagged CSN5 was detected by immunoblotting before (Input) and after IP of CSN5. V, empty vector. The asterisk denotes an unspecific band. The input blot (left) was reprobed, whereas duplicate gels were used for FLAG and CSN5 immunoblots after IP (right).

FIGURE 4.

p97/VCP interacts with the CSN complex. A, p97/VCP cofractionates together with CSN1 and CSN5 during Sephacryl S-200 gel chromatography. Proteins in 30% of each fraction were acetone-precipitated, separated by 12.5% SDS-PAGE, and immunoblotted (IB) using the antibodies indicated on the left. Fraction numbers are provided at the bottom, and molecular mass markers are at the top. Input, 2% of lysate. B, the remainder of fraction 5 was used for IP of CSN1 and detection of coprecipitated p97/VCP, CSN1, and CSN5 by immunoblotting. Ctr., IP with isotype control antibody. C, purified CSN (1.4 μg) and p97/VCP (6 μg; 1.2 μg in lane 4) were analyzed by immunoblotting of native 2–18% polyacrylamide-agarose composite gels either alone or after 2-h preincubation of both complexes in the absence or presence of 1 mm ATP. 4 m (lane 6) and 6 m urea (lane 7) were used to dissociate p97/VCP hexamers (p97hex). The asterisk denotes a p97/VCP homotrimer.

FIGURE 5.

CSN5 binds to oligoubiquitinated proteins in vitro and in vivo. A, oligoubiquitin chains Ub(2–7) were incubated with GST or GST-CSN5 immobilized on glutathione-Sepharose 4B or His-p97/VCP immobilized on Ni-NTA-agarose. Bound proteins were resolved by SDS-PAGE and identified by sequential immunoblotting. Protein samples of GST-CSN5 and His-p97/VCP were directly loaded in lanes 4 and 6, respectively. All lanes were on the same immunoblot; a few intervening lanes were omitted as indicated by vertical lines. B, HEK 293T cells transfected with empty vector (V) or Myc-CSN5 mutants were lysed and immunoprecipitated (IP) using a monoclonal anti-Myc antibody. Myc tag and ubiquitin were detected by immunoblotting (IB). Coprecipitated polyubiquitinated proteins were quantified with ImageJ (see numbers below the ubiquitin immunoblot (IB: Ub). Vector control was set to 0%, wild type (wt) Myc-CSN5 to 100%. The asterisk denotes IgG heavy chains.

FIGURE 6.

A functional CSN complex with associated USP15 deubiquitinase is required for deubiquitination of polyubiquitinated substrates bound to p97/VCP. A, HEK 293T cells were transfected with empty vector, CSN5, or the JAMM mutant of CSN5 (mCSN5) and treated with the proteasome inhibitor MG132. p97/VCP was immunoprecipitated from transfected cell lysates, and precipitates were immunoblotted (IB) for ubiquitin, p97/VCP, and CSN5. β-Actin was used as loading control. The fraction of coprecipitated polyubiquitinated proteins normalized against precipitated p97/VCP was quantified with ImageJ. Numbers below the ubiquitin immunoblots (IB: Ub) indicate the fold increase versus control, which was set to 1. B, cells were transiently transfected with control, CSN1, CSN5, and USP15 siRNAs. The intensities of protein bands detected in immunoblots were normalized for β-actin and quantified with ImageJ. CSN1 siRNA reduced CSN1 protein to 40%, CSN5 siRNA CSN5 protein to 18%, and USP15 siRNA USP15 protein to 24%. After treatment with the proteasome inhibitor MG132 for 1 h, cell lysates were analyzed by immunoblotting using specific antibodies as indicated on the left. After IP of p97/VCP, ubiquitin was detected in precipitates by immunoblotting. Asterisks denote IgG heavy chains, # indicates Myc-mCSN5. C, functional controls. Lysates of CSN5 knock-down cells were analyzed by immunoblotting for the presence of CSN5, cullin 1, IκBa, CSN1, and β-actin (upper panel). Lysates of mCSN5-transfected cells were analyzed by immunoblotting for the presence of CSN5, cullin 1, and β-actin (lower panel).