Crystal structure and versatile functional roles of the COP9 signalosome subunit 1

The constitutive photomorphogenesis 9 (COP9) signalosome (CSN) plays key roles in many biological processes, such as repression of photomorphogenesis in plants and protein subcellular localization, DNA-damage response, and NF-κB activation in mammals. It is an evolutionarily conserved eight-protein complex with subunits CSN1 to CSN8 named following the descending order of molecular weights. Here, we report the crystal structure of the largest CSN subunit, CSN1 from Arabidopsis thaliana (atCSN1), which belongs to the Proteasome, COP9 signalosome, Initiation factor 3 (PCI) domain containing CSN subunit family, at 2.7 Å resolution. In contrast to previous predictions and distinct from the PCI-containing 26S proteasome regulatory particle subunit Rpn6 structure, the atCSN1 structure reveals an overall globular fold, with four domains consisting of helical repeat-I, linker helix, helical repeat-II, and the C-terminal PCI domain. Our small-angle X-ray scattering envelope of the CSN1–CSN7 complex agrees with the EM structure of the CSN alone (apo-CSN) and suggests that the PCI end of each molecule may mediate the interaction. Fitting of the CSN1 structure into the CSN–Skp1-Cul1-Fbox (SCF) EM structure shows that the PCI domain of CSN1 situates at the hub of the CSN for interaction with several other subunits whereas the linker helix and helical repeat-II of CSN1 contacts SCF using a conserved surface patch. Furthermore, we show that, in human, the C-terminal tail of CSN1, a segment not included in our crystal structure, interacts with IκBα in the NF-κB pathway. Therefore, the CSN complex uses multiple mechanisms to hinder NF-κB activation, a principle likely to hold true for its regulation of many other targets and pathways.The constitutive photomorphogenesis 9 (COP9) signalosome (CSN) is a more than 300-kDa complex that was first identified as a negative regulator of Constitutive Photomorphogenesis (COP) in plants (1, 2). In the subsequent years, the highly conserved protein complex was also found in fungi (3, 4), Caenorhabditis elegans (5), Drosophila melanogaster (6), and mammals (7, 8). The most studied function of the CSN complex in eukaryotes is the regulation of protein degradation through two pathways, deneddylation (9–11) and deubiquitination (12, 13). In the deneddylation pathway, the CSN complex can influence the cullin-RING ligase activity by removing Nedd8, a ubiquitin-like protein, from a cullin (9, 14). On the other hand, the CSN complex can also suppress cullin activity through recruitment of the deubiquitination enzyme USP15 (12) or Ubp12p, the Schizosaccharomyces pombe ortholog of human USP15 (13). Other functions of the CSN complex identified in mammalian cells include regulating the phosphorylation of ubiquitin–proteasome pathway substrates through CSN-associated kinases (7, 15–18). Overall, the CSN complex appears to be a key player in protein subcellular localization (19, 20), DNA-damage response (21), NF-κB activation (22), development, and cell cycle control (23, 24). Thus, the functions of the CSN complex are beyond the regulation of light-dependent reaction in plants.The CSN complex in most of the species contains eight subunits named CSN1 to CSN8, in order of decreasing size. All eight subunits share homologous sequences with “lid” components of the 26S proteasome regulatory particle and the eukaryotic translation initiation factor 3 (eIF3) (7, 25). Among these eight subunits, CSN6 and catalytic CSN5 contain a conserved MPN-domain (MOV34, Pad1N-terminal) (26), and the rest of the CSN subunits bear a PCI-domain (Proteasome, COP9 signalosome, Initiation factor eIF3). The MPN-domain within CSN5 contains a metal chelating site and forms the catalytic region of the isopeptidase for deneddylation (27). Recently, the crystal structures of the CSN6–MPN domain and the CSN5 subunit have been revealed (28, 29). Interestingly, amino acids 97–131, a flexible segment within the CSN5–MPN domain, were proven to be essential in regulating the isopeptidase states of CSN5 (29). PCI is an ∼200-amino acid domain, beginning with an N-terminal helical bundle arrangement and ending with a globular winged-helix subdomain (30, 31). A number of interactions between PCI domains of CSN subunits have been identified by the yeast two-hybrid system (32, 33). Dessau et al. reported the crystallographic data of the PCI domain of Arabidopsis thaliana subunit 7, and their in vitro studies also suggested that the PCI domain mediates and stabilizes protein–protein interactions within the complex (34).Although many speculated on how the CSN subunits interact with each other and come into a functional unit, the architecture of the CSN complex was accessed by electron microscopy (EM) (35, 36) and native mass spectrometry approaches (37). These studies confirmed structural similarities among CSN, the proteasome lid, and eIF3. Furthermore, the CSN appears to contain two dominant subcomplexes, CSN1/2/3/8 and CSN 4/5/6/7 (37), which correspond to the large and the small segments, respectively, in an EM study of the CSN alone (apo-CSN) (36). An EM study of the CSN in complex with an Skp1-Cul1-Fbox (SCF) E3 ligase was also reported, showing reciprocal regulation between CSN and SCF (38). To date, unfortunately, there is no high-resolution mapping on these subunit interactions.To further define the CSN structure and to study its functional significance, we feel the need to obtain structures of CSN subunits at an atomic level. In our study, we used Arabidopsis thaliana CSN1 (atCSN1) as a guide to facilitate our understandings of the PCI-containing CSN subunits. The atCSN1, encoded in the chromosome 3, has 441 amino acids that are 45% identical in sequence to Homo sapiens CSN1. Among all of the subunits of the complex, CSN1 is known to be the longest and to play a crucial role in complex integrity (39–41). Here, we report the crystal structure of atCSN1 and describe its integration within the complex as well as its interaction with IκBα in the NF-κB signaling pathway.