• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of mbcLink to Publisher's site
Mol Biol Cell. Feb 2002; 13(2): 493–502.

Deletion Mutants in COP9/Signalosome Subunits in Fission Yeast Schizosaccharomyces pombe Display Distinct Phenotypes

John Pringle, Monitoring Editor


The COP9/signalosome complex is highly conserved in evolution and possesses significant structural similarity to the 19S regulatory lid complex of the proteasome. It also shares limited similarity to the translation initiation factor eIF3. The signalosome interacts with multiple cullins in mammalian cells. In the fission yeast Schizosaccharomyces pombe, the Csn1 subunit is required for the removal of covalently attached Nedd8 from Pcu1, one of three S. pombe cullins. It remains unclear whether this activity is required for all the functions ascribed to the signalosome. We previously identified Csn1 and Csn2 as signalosome subunits in S. pombe. csn1 and csn2 null mutants are DNA damage sensitive and exhibit slow DNA replication. Two further putative subunits, Csn4 and Csn5, were identified from the S. pombe genome database. Herein, we characterize null mutations of csn4 and csn5 and demonstrate that both genes are required for removal of Nedd8 from the S. pombe cullin Pcu1 and that their protein products associate with Csn1 and Csn2. However, neither csn4 nor csn5 null mutants share the csn1 and csn2 mutant phenotypes. Our data suggest that the subunits of the signalosome cannot be considered as a distinct functional unit and imply that different subunits of the signalosome mediate distinct functions.


Elements of the COP9/signalsome complex were first identified genetically in Arabidopsis through the isolation of COP mutants that result in constitutive photomorphogenesis (Chamovitz et al., 1996 blue right-pointing triangle; Wei and Deng, 1999 blue right-pointing triangle). The majority of the COP loci was subsequently found to correspond to subunits of a multiprotein complex (the COP9 complex) that is required for the nuclear accumulation of a transcriptional regulator, COP1. COP1 is a ring finger protein that accumulates in the nucleus when plants are germinated in the dark, but which is cytoplasmic when germination occurs in the light. Within the nucleus, COP1 binds to the HY5 transcription factor and promotes its degradation, thus negatively regulating a program of gene expression required for photomorphogenesis (Osterlund et al., 2000 blue right-pointing triangle). In Drosophila melanogaster, the homologous complex is essential for embryo development (Freilich et al., 1999 blue right-pointing triangle), and this has lead to the suggestion that the COP9/signalosome has a role in integrating different signaling pathways during differentiation.

An analogous human complex was identified biochemically during attempts to purify the regulatory lid component of the proteasome. The human complex copurifies with a kinase activity that is capable of phosphorylating c-Jun and IkB (Seeger et al., 1998 blue right-pointing triangle) and may be encoded by inositol 1,3,4-trisphosphate 5/6-kinase (Wilson et al., 2001 blue right-pointing triangle). Overexpression of some, but not all, subunits in human cells leads to differential effects on the stability of c-Jun (Naumann et al., 1999 blue right-pointing triangle). These observations have led to a model in which the signalosome is involved in the integration and/or modulation of multiple signal transduction pathways, and may be directly involved in regulating the transcriptional output of these pathways. A unified nomenclature for the individual subunits in different species has been agreed: the subunits are all named Csn (for COP9/signalosome) and numbered from 1 to 8 (Deng et al., 2000 blue right-pointing triangle).

The mammalian Csn5 subunit had previously been identified as c-Jun–interacting protein, Jun-activation domain binding protein JAB1 (Claret et al., 1996 blue right-pointing triangle). Csn5 also interacts with a number of nuclear hormone receptors (Chauchereau et al., 2000 blue right-pointing triangle; Li et al., 2000 blue right-pointing triangle) and with the cytoplasmic domain of the integrin LFA-1. Integrin adhesion receptors transduce signals that regulate of gene expression and activation of LFA-1 leads to an increase in the nuclear pool of Csn1 that is concomitant with an increase in transcription from an AP1-dependent promoter (Bianchi et al., 2000 blue right-pointing triangle). Csn5 is also found to interact with the CDK inhibitor protein p27. Overexpression of Csn5 decreases the nuclear abundance of p27, but the mechanism for this is unclear (Tomoda et al., 1999 blue right-pointing triangle). In Arabidopsis, two isoforms of Csn5 exist (also known as Ajh1 and Ajh2 with, respectively, 62 and 63% identity to human Csn5). Both Arabidopsis proteins exist in monomeric and complexed form, with complex formation dependent on Csn1. Although complexed Csn5 isoforms localizes predominantly to the nucleus in Arabidopsis, the monomeric forms are mainly cytoplasmic (Kwok et al., 1998 blue right-pointing triangle). In addition, Csn5 is the only subunit of the signalosome complex with an apparent homolog in Saccharomyces cerevisiae. This is further evidence that there might be a signalosome-independent function for Csn5.

Human signalosome subunits interact with multiple cullins (Lyapina et al., 2001 blue right-pointing triangle). Cullins are components of the SCF ubiquitin ligases that contain core subunits of cullin, Skp1, and a ring finger protein, and which associate with multiple F-box proteins and ubiquitin-conjugating enzymes. In fission yeast, one of the three cullins (Pcu1) requires covalent attachment of Nedd8 (a small ubiquitin-like protein) to Lys 713 for function. In csn1 null mutants, Pcu1 is present exclusively in the Nedd8-modified form, whereas csn+ cells have ~50% of Pcu1 Nedd8 modified (Lyapina et al., 2001 blue right-pointing triangle). This observation has led to the suggestion that Nedd8 modification of Pcu1, and probably other cullins, may underlie the csn1-d and csn2-d phenotypes. We set out to determine whether S. pombe Csn4 and Csn5 were capable of interacting with Csn1 and with Csn2, to characterize the corresponding deletion mutants (csn4-d and csn5-d), and ascertain whether hypermodification of cullins correlated with phenotypes associated with csn1-d and csn2-d.


Cell Biology

Protocols for checkpoint measurements, cell scoring, and irradiation are previously described (Edwards and Carr, 1997 blue right-pointing triangle). γ-Irradiation was performed on exponentially growing cells in liquid culture, by using a Gammacell 1000 (Nordion, Ontario, Canada) 137Cs source (12.5 Gy/min). A known number of cells was plated after different doses of irradiation and colonies deriving from surviving cells were counted after 3 to 4 d, expressing survival as a percentage of colonies formed by mock-treated cells on equivalent plates. For UV survival analysis a known number of cells was plated and irradiated with different doses by using a germicidal UVC lamp (0.2 J/m2s). Survival was scored as for γ-irradiation. Survival in hydroxyurea was performed on asynchronous cultures. Aliquots were removed at hourly intervals after addition of 10 mM hydroxyurea and plated after dilution. Survival was assessed by counting the number of colonies formed compared with the untreated control (t0).


Proteins were identified in crude total extracts prepared by mechanical lysis in trichloroacetic acid, which prevents proteolysis. This is described in Caspari et al. (2000) blue right-pointing triangle.

Gel Filtration Analysis

Gel filtration analysis was performed as previously described (Caspari et al., 2000 blue right-pointing triangle). Briefly, extracts were prepared by dripping the cells resuspended in buffer B (50 mM NaH2PO4/Na2HPO4 pH 7, 150 mM NaCl, 10% glycerol, 0.5 mM dithiothreitol, 5 mM EGTA, 60 mM β-glycerolphosphate, 0.1 mM NaF, 10 μg/ml apoprotinin, 10 μg/ml leupeptin, 25 mM p-aminobenzamidin, 10 mM phenylmethylsulfonyl fluoride) into a mortar filled with liquid nitrogen and grinding them with a pestle. Frozen cell homogenate was allowed to thaw then was clarified (2 min 2000 rpm; Eppendorf), and the resulting supernatant was spun for 30 min at 30,000 rpm in a Sorvall AH650 at 4°C. The extract (~1 mg of total protein) was then passed through an equilibrated Superdex 200HR 10/30 column (Pharmacia) and 0.5-ml fractions were collected. The column was calibrated using the low- and high-molecular weight gel filtration calibration kit (Pharmacia).


For coimmunoprecipitation experiments, soluble, whole cell extracts were prepared in buffer B by grinding cells in liquid nitrogen as described above. Extract (400 μl, concentration 5 mg/ml) was precleared by incubation with prewashed protein G beads rotating for 30 min at 4°C. The supernatant was transferred to a fresh tube and 4 μl of primary rabbit anti-MYC antibody (Santa Cruz Biotechnology, Santa Cruz, CA), diluted 1:2, or 4 μl of mouse anti-hemagglutinin (HA) antibody (BabCO, Richmond, CA), diluted 1:5, was added for 1 h, rotating at 4°C. The extract was centrifuged for 5 min at 14,000 rpm in a microfuge and transferred to a fresh tube containing 15 μl of prewashed protein G beads. The immunocomplexes were left to absorb to the protein G beads by rotating incubation at 4°C for 30 min and subsequently washed once in buffer B and twice in buffer B containing 250 mM NaCl. SDS sample-buffer (2×, 100 μl) was added to the washed pelleted beads, boiled for 5 min, and separated by 8% SDS-PAGE. Standard protocols were used for SDS-PAGE and Western blot analysis. Detection was by chemiluminescence.

Immunofluorescence Microscopy of Csn1-MYC, Csn2-MYC, Csn4-MYC, and Csn5-MYC

For immunofluorescence staining cells were processed essentially as previously described (Hagan and Hyams, 1988 blue right-pointing triangle). Cells were fixed in mid-exponential growth by adding the culture to freshly prepared 30% paraformaldehyde solution to a final concentration of 3.75% and shaking for 10 min at 30°C. Cells were subsequently washed two times in PEM buffer [100 mM piperazine-N,N′-bis(2-ethanesulfonic acid pH 6.9; 1 mM EGTA, 1 mM MgSO4] and once in PEMS buffer (PEM + 1.2 M sorbitol). The cell wall was digested with 1.25 mg/ml zymolase (ICN Biomedicals, Cleveland, OH) in PEMS for 70 min at 37°C and membranes subsequently permeabilized with 1% Triton X-100 in PEMS. After two washes in PEM the cells were blocked in PEMBAL (PEM + 1% bovine serum albumin, 0.1% NaN3, 100 mM lysine hydrochloride pH 6.9) for >30 min. Cells were incubated overnight with primary antibody 12CA5 anti-HA antibody (BabCO) at a dilution of 1:250 or primary rabbit anti-MYC antibody (RDI) at a dilution of 1:150. They were washed three times in PEM and incubated with secondary Cy3-conjugated anti-mouse antibody (Jackson Immunoresearch Laboratories, West Grove, PA) at 1:250 or fluorescein isothiocyanate-conjugated anti-rabbit antibody (DAKO, Bucks, United Kingdom) at 1:150 for 2–6 h. After washing in PEM, phosphate-buffered saline (PBS), and PBS with 4,6-diamidino-2-phenylindole (DAPI) (1 μg/ml), the cells were resuspended and stored in PBS + 0.1% NaN3 at 4°C.

Cells were viewed after mounting in 90%glycerol containing 1 mg/ml para-phenylene diamine on a Leitz microscope and pictures were acquired using the MetaMorph imaging system (Universal Imaging, Downingtown, PA).

Genomic Tagging of Csn2, Csn4, and Csn5 and Gene Deletion

We genomically tagged the C terminus of csn2, csn4, and csn5 with 13-MYC epitopes by using the one-step polymerase chain reaction-tagging technology previously described by Bahler et al. (1998) blue right-pointing triangle. Replacements of the complete open reading frames of csn4 and csn5 with ura4+ were performed as described in Bahler et al. (1998) blue right-pointing triangle. Recombination at the correct locus was checked by Southern blot analysis and back-crossing of tagged alleles with the respective deletion mutants.

Strains Used and Constructed in This Analysis

The full list of strains is in Table Table1.1. All csn-d mutants have the open reading frame replaced by the ura4+ gene. All csn-MYC and csn-HA strains include a transcription termination site and a kanamycin resistance cassette after the stop codon. The diploid strain was created by mating the csn2-HA ade6.M210 and csn2-MYC ade6.M216. These two ade6 alleles exhibit intra-allelic complementation, and a diploid can be obtained by selecting for growth on adenine after zygote formation but before meiosis. The pREP41-HA-Nedd8 plasmid was introduced into relevant strains by standard techniques and cells maintained in minimal media without thiamine.

Table 1
Localization of Csn subunits


Csn1, Csn2, Csn4, and Csn5 Interact In Vivo

Our previous work demonstrated an interaction between Csn1 and Csn2 and suggested that these proteins formed a soluble protein complex of >500 kDa, consistent with the existence of a signalosome-like particle in fission yeast. The sequence similarity of Csn4 and Csn5 to signalosome subunits in higher eukaryotes predicts that these are subunits of a signalosome-like complex in fission yeast. To verify this, we added a C-terminal tag to both the csn4 and csn5 genes, which encodes the MYC epitope, and tested coimmunoprecipitation of Csn4 and Csn5 with Csn2-HA (Figure (Figure1A).1A). We were able to ascertain that these tagged proteins retained function (Figure (Figure4B).4B). With the anti-MYC antibody, immunoprecipitated MYC-tagged Csn1, Csn4, and Csn5 proteins were all found to be associated with HA-tagged Csn2 (Figure (Figure4B,4B, lanes 5–7). Csn2-HA was not precipitated in the absence of MYC-tagged Csn subunits (Figure (Figure4B,4B, lanes 2–4). Vice versa, coimmunoprecipitation of MYC-tagged Csn1, Csn4, or Csn5 protein with the anti-HA antibody (Figure (Figure4B,4B, lanes 5–7) required the presence of Csn2-HA. In addition, we also Flag-tagged Csn5 and used the anti-Flag antibody to coprecipitate Csn1 and Csn4 and demonstrated that Flag-tagged Csn5 could be immunoprecipitated with Csn1-MYC and Csn4-MYC (our unpublished data). Together, these data demonstrate that Csn1, Csn2, Csn4, and Csn5 interact in vivo. We have attempted to quantify the extent of coimmunoprecipitation (Figure (Figure1B).1B). Western blots of similar amounts of crude extract and depleted extract were compared after coprecipitation of Csn2-HA with either Csn1-MYC, Csn4-MYC, or Csn5-MYC and vice versa. The results are reasonably reproducible and suggest that, for example, although all of Csn2-HA is precipitated by anti-HA antibody (our unpublished data) the MYC-tagged partner is depleted by >50%. Such experiments do not allow us to conclude with certainty the proportions of protein in different complexes, but they are broadly consistent with size exclusion data discussed below.

Figure 1
Characterization of protein interactions. (A) Coimmunoprecipitation of Csn2-HA with Csn1-MYC, Csn4-MYC, or Csn5-MYC: either anti-MYC (left) or anti-HA (middle) were used to immunoprecipitate Csn subunits from total protein extracts (25 μg of which ...
Figure 4
Pcu1 modification by Nedd8. (A) Extracts from cells containing a Rep41-HA-Nedd8–expressing plasmid. The left two lanes contain extracts from pcu1+ cells. The remaining four lanes contain extracts from cells with a MYC-tagged pcu1 locus. ...

Because Csn1 and Csn2 are associated with a complex of >500 kDa (as judged by gel filtration analysis) our coimmunoprecipitation analysis suggests that Csn4 and Csn5 should be associated with the same complex. Indeed, gel filtration analysis of Csn4 and Csn5 demonstrates that these proteins are associated with a protein complex of >500 kDa, consistent with their being associated with a Csn1- and Csn2-containing complex (Figure (Figure1C).1C). It should be noted, however, that the majority of both Csn4 and Csn5 is present in fractions representing lower molecular weights. This may indicate that not all of the Csn4 and Csn5 is associated with a core signalosome complex. Arabidopis Csn5 has been found in both a nuclear (complexed) form and a cytoplasmic (noncomplexed) form. We therefore used immunolocalization by indirect fluorescence to establish the localization of Csn4 and Csn5 in S. pombe. As we previously reported for Csn1 and Csn2, both Csn4 and Csn5 are predominantly localized to the nucleus (Figure (Figure1D).1D). From the data described above we conclude that Csn4 and Csn5 are bona fide subunits of the Csn1- and Csn2-containing signalosome complex in fission yeast.

Deletion of csn4 and csn5

To characterize the genetic interactions of csn4 and csn5, we deleted both genes from the genome by using standard techniques. We have previously reported that both csn1-d and csn2-d are sensitive to ionizing irradiation and UV light and are delayed in progression through S phase during logarithmic growth (Mundt et al., 1999 blue right-pointing triangle). Unexpectedly, csn4-d and csn5-d cells display wild-type profiles of DNA damage sensitivity (Figure (Figure2A)2A) and the cell cycle parameters appear normal (our unpublished data). In addition, deletion of csn5 in a csn1-d background did not further exacerbate the radiation sensitivity of this mutant (Figure (Figure2A).2A).

Figure 2
Analysis of csn4-d and csn5-d. (A) Survival analysis. Top, percentage survival of csn+ (wt), csn1-d, csn2-d, csn4-d, csn5-d, and rad3-d cells after increasing doses of ionizing radiation. Middle, percentage survival of csn+ (wt), csn1-d ...

Loss of either csn1 or csn2 leads to activation of the DNA structure checkpoint pathway as judged by assaying phosphorylation of the DNA damage-specific checkpoint kinase Chk1 and the activation of the S-phase–specific checkpoint kinase Cds1 (Mundt et al., 1999 blue right-pointing triangle). Consistent with this, we also reported that both csn1-d and csn2-d mutant cells require the presence of a number of checkpoint genes for viability, including rad3, which encodes the central PI3-like checkpoint kinase (Mundt et al., 1999 blue right-pointing triangle). In contrast, we find that neither csn4-d nor csn5-d requires rad3 for survival. As shown in Figure Figure2B,2B, although csn1-d rad3ts and csn2-d rad3ts die at the restrictive temperature, both csn4-d rad3ts and csn5-d rad3ts are viable. This is consistent with the lack of a cell elongation phenotype for either csn4-d or csn5-d, which have previously been documented for csn1-d and csn2-d mutants. Taken together, these observations demonstrate that the csn4-d and csn5-d mutants do not suffer from the same intrinsic DNA damage as csn1-d and csn2-d and are not defective in S-phase progression.

Complex Formation and Nuclear Localization in Deletion Mutants of COP9/Signalosome Subunits

Intrigued by the distinct phenotypes resulting from the deletion of different COP9/signalosome subunits, we went on to analyze the effect of loss of csn1, csn2, csn4, and csn5 on complex formation (Figure (Figure3A)3A) and subunit localization (Figure (Figure3B3B and Table Table2).2). The size of the Csn1-containing complex is modestly affected by the deletion of csn2, but either not at all, or to an undetectable extent, by deletion of csn4 or csn5. It is particularly noticeable that the elution profile of Csn1-MYC spreads to lower molecular weight fractions in the csn2-d and csn5-d backgrounds. The profile of Csn2-MYC is shifted to a smaller size in csn1-d and, to a lesser extent in csn4-d, but not significantly in csn5-d mutants. Analysis of Csn4-MYC in csn1-d, csn2-d, and csn5-d reveals the most consistent result, with Csn4 not forming the higher molecular weight complex in either csn1-d, csn2-d, or csn5-d cells. Similarly, the amount of Csn5 in the higher molecular weight fractions is reduced in csn2-d and csn4-d and may be modestly affected in csn1-d cells. These data suggest that signalosome subunits Csn1 and Csn2 exist together in a multisubunit complex with Csn4 and Csn5, whereas Csn4 and Csn5 also exist in other, lower molecular weight forms.

Figure 3
Subunit localization and size exclusion chromatography in csn mutant backgrounds. (A) Size exclusion chromatography of MYC-tagged Csn1, Csn2, Csn4, and Csn5 proteins in csn mutant backgrounds. Extracts were prepared from the strains indicated and fractionated ...
Table 2
Strains used in this study

Surprisingly, examination of the localization of the signalosome subunits by indirect immunofluorescence did not correlate with the observed differences in phenotype between csn1 and csn2 mutants and csn4 and csn5 mutants. In strains deleted for any one of the four subunits, csn1-d, csn2-d, csn4-d, and csn5-d, the remaining epitope-tagged subunits were no longer observed to localize to the nucleus and the overall staining intensity was similar to that observed in control (nontagged) wild-type cells. The MYC-tagged Csn subunits are stable under these conditions (Figure (Figure5C)5C) so these data suggest that the proteins are either dispersed throughout the cell, or the tag is no longer accessible to the antibody under these conditions. Tubulin staining in the same cells excluded the possibility that a difference in cell wall permeability had led to this observation (our unpublished data).

Figure 5
Csn protein levels in csn mutants. (A) Coimmunoprecipitation of Csn2-HA with Csn2-MYC. Either anti-HA (left) or anti-MYC (middle) was used to immunoprecipitate Csn2 subunits from total protein extracts (25 μg of which is probed with both anti-HA ...

The size exclusion chromatography results suggest that Csn4 requires Csn5 to associate with the Csn1/Csn2-containing complex. To test this further, we established whether the interaction between Csn2 and Csn4 required the presence of the csn5 gene. By comparing coimmunoprecipitation from extracts prepared from csn2-HA csn4-MYC csn5+, and csn2-HA csn4-MYC csn5-d cells we can clearly show that the Csn2-Csn4 immunoprecipitation requires the integrity of csn5 (Figure (Figure33C).

Cullin Modification by Nedd8 in csn Mutants

Approximately half of the S. pombe Pcu1 cullin protein is modified by Nedd8 at lysine 713 in wild-type cells (Osaka et al., 2000 blue right-pointing triangle). This modification is required for Pcu1 function. Furthermore, purification of mammalian cullin-associated proteins identified association with Csn2 (Lyapina et al., 2001 blue right-pointing triangle). Interestingly, in csn1-d cells, all of the Pcu1 is modified by Nedd8, suggesting that the csn1 and by extension the signalosome, is required to remove Nedd8 from cullins (Lyapina et al., 2001 blue right-pointing triangle). Using an HA-Nedd8 expression vector, we have established an assay for Pcu1 cullin modification (Figure (Figure4A).4A). We find that loss of either csn1, csn2, csn4, or csn5 results in hyper-Nedd8 modification of Pcu1 (Figure (Figure4B).4B). Because loss of csn4 or csn5 does not result in the radiation sensitivity and slow S-phase phenotypes seen for loss of csn1 and csn2, this establishes that the inability to remove Nedd8 from Pcu1 (and presumably other cullins) is not the cause of the these phenotypes.

The reported interaction between Pcu1 and signalosome subunits raises the possibility that the signalosome is part of a stable complex with Pcu1. We have used the size exclusion chromatography assay to establish whether there is evidence that this is the case. Under the same conditions that clearly demonstrated the existence of an S. pombe signalosome complex (Figure (Figure3A),3A), we did not find that the elution profile of Pcu1-MYC was dependent on the integrity of the csn1, csn2, or csn4 genes (Figure (Figure44C).

Lack of Evidence for Csn2 Dimerization

Yeast offers a particularly useful system to study dimerization of proteins because diploid strains can be constructed containing two distinct tagged copies of a single gene locus. It has previously been suggested that Arabidopsis Csn2 may dimerize to form the core of the signalosome complex. We thus investigated the potential dimerization of Csn2 in fission yeast by constructing a diploid strain carrying both csn2-MYC and csn2-HA alleles and by preparing extracts from this strain. Soluble, whole cell extract contained the expected proportions of Csn2-HA and Csn2-MYC. Despite efficient immunoprecipitation of either protein by using the appropriate MYC or HA antibody, we did not detect coimmunoprecipitation of reciprocally tagged Csn2 protein (Figure (Figure5A).5A). Thus, we conclude that in fission yeast, at least, Csn2 is not present in multiple copies in a single protein complex. Although both Csn2-MYC and Csn2-HA are functional, judged by the lack of phenotype associated with the respective alleles, we cannot formally exclude the possibility of steric hindrance between the HA-MYC tags preventing heterodimerization. If this was the case, such steric hindrance presumably does not occur between HA-HA or MYC-MYC. If such HA-HA and MYC-MYC steric hindrance does occur, dimerization cannot be important for the signalosome functions we assay.

Loss of csn1 Leads to Instability of HA- but not MYC-tagged Csn2

Protein stability, and thus the steady-state levels, of subunits of complexes are sometimes affected by the presence or absence of the other subunits. Indeed, it has been reported that loss of Csn1 in Arabidopsis significantly reduces levels of other components of the signalosome. During our analysis, we observed that Csn2-HA levels in a csn1-d background were greatly reduced compared with those seen in csn1+ cells (Figure (Figure5B).5B). This suggests that Csn2-HA becomes unstable in the absence of csn1 (Figure (Figure5B,5B, lanes 2 and 3). After a long exposure (our unpublished data), low levels of HA-tagged protein can still be detected compared with a nontagged wild-type strain (Figure (Figure5B,5B, lane 6). As a further control, we crossed two independently isolated Csn2-HA csn1-d strains (Figure (Figure5B,5B, lanes 2 and 3) back into a csn1+ background. This rendered the protein stable and detectable (Figure (Figure5B,5B, lanes 4 and 5). The deletion of csn4 and csn5 had no obvious effect on the stability of Csn2-HA (Figure (Figure5B,5B, lanes 7 and 8). To verify and extend this result, we tested all four 13-MYC-epitope–tagged subunits under circumstances where one of each of the remaining subunits was deleted. Contrary to the result described above, we found that Csn2-MYC was more abundant, and thus apparently stabilized in the csn1-d background (Figure (Figure3B).3B). This result could be interpreted in two ways: either Csn2 is protected from degradation by the presence of the C-terminal MYC tags, or Csn2 is promoted for degradation (but only in the absence of its partner subunits) by the presence of C-terminal HA tags. This phenomenon of increased abundance in specific null mutant backgrounds was also apparent with Csn4-MYC in both the csn1-d and csn2-d backgrounds. In contrast, deletion of csn1, csn2, or csn4 did not affect the protein stability of Csn5, suggesting that Csn5 protein stability is not dependent on the Csn1/Csn2-containing signalosome complex (Figure (Figure3B).3B). It is intriguing that although all four subunits were tagged with 13-MYC epitopes, their apparent abundance in the csn+ background (Figure (Figure3B,3B, lanes 1, 5, 9, and 13) differs greatly and Csn1-MYC and Csn5-MYC appear to exist in excess over Csn2-MYC and Csn4-MYC.


Csn4 and Csn5 Are Bona Fide Subunits of COP9/Signalosome in Fission Yeast

The Csn1 subunit of the signalosome in fission yeast was initially identified through a genetic screen for suppression of Chk1-induced G2 arrest coupled with radiation sensitivity. Csn2 was identified as a two-hybrid interacting protein with Csn1. Csn4 and Csn5 were identified as putative subunits based on, respectively, 23 and 41% identity with their human counterparts. Herein, we have demonstrated that the Csn4 and Csn5 proteins interact with Csn2 in vivo. Consistent with our previous data, we also demonstrate that endogenous Csn1-MYC interacts with Csn2. Using a diploid strain, carrying both an HA- and an MYC-tagged allele of Csn2, we have further investigated the possibility that Csn2 dimerizes. Under identical conditions used to coimmmunoprecipitate the other subunits, we could not find any association of Csn2-HA with Csn2-MYC and vice versa. Csn2-HA and Csn2-MYC are functional in all the assays we have available. This strongly suggests that the Csn2 subunit is not present in more than one copy in the same protein complex. However, we should bear in mind the caveat that this specific combination of HA and MYC tags could prevent heterodimerization of Csn2.

Extensive size exclusion chromatography analysis indicated that, although Csn4 and Csn5 elute in similar size fractions to Csn1 and Csn2, significant quantities are also found in smaller molecular weight fractions. This could be indicative of the existence of either a smaller size subcomplex of the signalosome or of distinct additional complex(es) containing these proteins. Although this is consistent with data reported in Arabidopsis, we cannot exclude the possibility that the epitope tags affects the stability of the signalosome complex, although they clearly retain function as judged by the ability to promote the deneddulation of Pcu1. By analyzing size distribution of different subunits in various null mutant backgrounds, we are able to clearly show that the fraction of Csn4, which exists in a complex of >500 kDa, requires the presence of Csn1, Csn2, and Csn5. Indeed, we are also able to show that Csn5 is required for the coimmunoprecipitation of Csn4 with Csn2. The distribution of Csn2 in these experiments is clearly affected when Csn1 and Csn4 are absent, although a less dramatic change is evident when Csn5 is absent. Conversely, only subtle changes in molecular weight distribution are evident for Csn1 and Csn5 when other subunits are deleted.

Genetic Analysis Distinguishes csn4-d and csn5-d from csn1-d and csn2-d

Despite the association of Csn4 and Csn5 with Csn1 and Csn2 proteins, the deletion of csn4 and csn5 from the genome did not phenocopy csn1 and csn2 null mutants. csn4-d and csn5-d are not sensitive to UV or ionizing irradiation. Furthermore, csn5-d did not exacerbate the sensitivity of the csn1-d mutant. Unlike csn1-d and csn2-d, neither csn4-d nor csn5-d is synthetic lethal with rad3 mutants or with other checkpoint mutants (our unpublished data), suggesting that the intrinsic DNA damage generated in csn1 or csn2 null cells does not occur in csn4 or csn5 mutants. This is consistent with the lack of the cell elongation phenotype characteristic of csn1-d and csn2-d mutants, which correlates with S-phase delay and the constitutive activation of DNA-integrity checkpoint kinases. Despite trying a number of assays, we have yet to detect a clear phenotype associated with loss of either Csn4 or Csn5. This indicates two things. First, the different subunits of the signalosome do not act as a single functional unit in fission yeast, but presumably mediate a number of processes, some of which are dependent on specific subunits and may reflect distinct subcomplexes. Second, rapidly growing cells under laboratory conditions do not necessarily represent all the cellular requirements demanded of yeast cells in nature.

Nuclear Localization of Signalosome Subunits

Using indirect immunofluorescence, we find that Csn1, Csn2, Csn4, and Csn5 localize predominantly to the nucleus. It has been speculated that the signalosome may interact with the core 20S subunit of the proteasome, because it has structural similarity to the 19S regulatory lid complex. The proteasome in fission yeast has been shown to localize to the nuclear periphery (Wilkinson et al., 1998 blue right-pointing triangle). We do not detect any accumulation of COP9/signalosome subunits in this region. We cannot exclude the colocalization of a subpopulation of the signalosome with the proteasome.

We analyzed the effects of the deletion of the different signalosome subunits on nuclear localization. We had anticipated a correlation among nuclear localization, complex formation (as judged by size exclusion chromatography), and the pleiotrophic phenotypes associated with loss of csn1 and csn2. However, the emerging picture, summarized in Table Table1,1, was more complex and does not yet allow a straightforward interpretation. When any one subunit was deleted from the genome, we were unable to detect specific nuclear localization of any of the remaining MYC-tagged subunits. Western blot analysis indicated that the proteins were present in the cells. In particular, the loss of nuclear staining for Csn1-MYC and Csn2-MYC in csn4-d or csn5-d mutant backgrounds was surprising because it did not correlate with the radiation sensitivity associated with csn1 and csn2 deletion mutants. Although it is possible that nuclear localization of Csn1 and Csn2 is not important for their function, it is also possible that Csn1 and Csn2 proteins are dispersed throughout the cell, and that sufficient material is present in the nucleus to retain function.

Cullin Modification in Signalosome Mutants

The Nedd8 modification of Pcu1 lys 713 is required for its essential function. Intriguingly, in the absence of csn1, Pcu1 is hyper-Nedd8 modified. This suggested that the lack of regulation of cullin modification by Nedd8 may underlie the phenotypes associated with csn1 and csn2 deletion mutants. However, we have shown that Pcu1 is hyper-Nedd8 modified in csn4 and csn5 null mutants. Because neither of these mutants exhibit any obvious phenotype, this precludes a causative relationship between Pcu1 modification status and the csn1-d or csn2-d phenotypes. S. pombe contains two further cullin homologs, Pcu3 and Pcu4. Although we have not assessed their Nedd8 modification status in the signalosome mutants, it is likely that they are also hypermodified.

It remains unclear what the biochemical function(s) of the signalosome subunits is. The accumulating evidence suggests a close involvement in aspects of ubiquitin-dependent protein degradation. Csn1 and Csn2 are apparently involved in a specific pathway or pathways that are required for a normal DNA damage response and the proper execution of S phase. Csn4 and Csn5 are not required for this set of functions. We have been unable to identify any biological endpoints for the loss of Csn4 and Csn5 function and thus we are not able to assess whether their function(s) require Csn1 and Csn2. However, based on our size exclusion chromatography and immunolocalization data, we speculate that the different subunits, or subsets of subunits, participate in distinct functions. Clearly, in our model these subcomplexes must aggregate, presumably for a purpose that is, as yet, undetermined but that may affect the global regulation of SCF ubiquitin ligase complexes. A complete and satisfactory explanation will have to await the characterization of the other missing subunits of the fission yeast signalosome-like complex and the determination of its structure and function.

A Cautionary Tail of Tags

In Arabidopsis, mutations in csn1 (COP11) lead to instability of Csn8 (COP9) protein (Wei et al., 1994 blue right-pointing triangle). We observed that Csn2-HA becomes unstable in the csn1-d background. Unexpectedly, while attempting to verify this observation by using the csn2-myc cells we found that the effect of csn1-d on Csn2-MYC stability was opposite: Csn2-MYC was stabilized and not destabilized, as was observed with the Csn2-HA protein. This difference could possibly be due to protection of the MYC-tagged protein from degradation by the long C-terminal extension, which would subsequently lead to Csn2-MYC accumulation. However, it is also formally possible that the HA-tag promotes degradation of Csn2-HA protein when it is not in the requisite complex. We must await the generation of a suitably sensitive antibody to the endogenous protein before resolving this point, but it is noticeable how much of the literature relies on HA- or MYC-tagged proteins to support conclusions concerning protein abundance.


K.M. acknowledges Marie Curie Fellowship number ERBFMBICT-9272313. We thank members of the laboratory for support, Thomas Caspari for inspiration, and Takashi Toda for reagents.


Article published online ahead of print. Mol. Biol. Cell 10.1091/mbc.01–10-0521. Article and publication date are at www.molbiolcell.org/cgi/10.1091/mbc.01–10-0521.

Article published online ahead of print. Mol. Biol. Cell 10.1091/mbc.01–10-0521. Article and publication date are at .


  • Bahler J, Wu JQ, Longtine MS, Shah NG, McKenzie A, Steever AB, Wach A, Philippsen P, Pringle JR. Heterologous modules for efficient and versatile PCR-based gene targeting in Schizosaccharomyces pombe. Yeast. 1998;14:943–951. [PubMed]
  • Bianchi E, Denti S, Granata A, Bossi G, Geginat J, Villa A, Rogge L, Pardi R. Integrin LFA-1 interacts with the transcriptional co-activator JAB1 to modulate AP-1 activity. Nature. 2000;404:617–621. [PubMed]
  • Caspari T, Dahlen M, Kanter-Smoler G, Lindsay HD, Hofmann K, Papadimitriou K, Sunnerhagen P, Carr AM. Characterization of Schizosaccharomyces pombe Hus1: a PCNA-related protein that associates with Rad1 and Rad9. Mol Cell Biol. 2000;20:1254–1262. [PMC free article] [PubMed]
  • Chamovitz DA, Wei N, Osterlund MT, von Arnim AG, Staub JM, Matsui M, Deng XW. The COP9 complex, a novel multisubunit nuclear regulator involved in light control of a plant developmental switch. Cell. 1996;86:115–121. [PubMed]
  • Chauchereau A, Georgiakaki M, Perrin-Wolff M, Milgrom E, Loosfelt H. JAB1 interacts with both the progesterone receptor and SRC-1. J Biol Chem. 2000;275:8540–8548. [PubMed]
  • Claret FX, Hibi M, Dhut S, Toda T, Karin M. A new group of conserved coactivators that increase the specificity of AP-1 transcription factors. Nature. 1996;383:453–457. [PubMed]
  • Deng X-W, et al. Unified nomenclature for the COP9 signalosome and its subunits: an essential regulator of development. Trends Genet. 2000;16:202–203. [PubMed]
  • Edwards RJ, Carr AM. Analysis of radiation-sensitive mutants of fission yeast. Methods Enzymol. 1997;283:471–494. [PubMed]
  • Freilich S, Oron E, Kapp Y, Nevo-Caspi Y, Orgad S, Segal D. The COP9 signalosome is essential for development of Drosophila melanogaster. Curr Biol. 1999;9:1187–1190. [PubMed]
  • Hagan IM, Hyams JS. The use of cell division cycle mutants to investigate the role of microtuble distribution in the fission yeast Schizosaccharomyces pombe. J Cell Sci. 1988;89:343–357. [PubMed]
  • Kwok SF, Solano R, Tsuge T, Chamovitz DA, Ecker JR, Matsui M, Deng X-W. Arabidopsis homologs of a c-Jun coactivator are present both in monomeric form and in the COP9 complex, and their abundance is differentially affected by the pleiotropic cop/det/fus mutations. Plant Cell. 1998;10:1779–1790. [PMC free article] [PubMed]
  • Li S, Liu X, Ascoli M. p38JAB1 binds to the intracellular precursor of the lutropin/choriogonadotropin receptor, and promotes its degradation. J Biol Chem. 2000;275:13386–13393. [PubMed]
  • Lyapina S, Cope G, Shevchenko A, Serino G, Tsuge T, Zhou C, Wolf DA, Wei N, Shevchenko A, Deshaies RJ. Promotion of NEDD-CUL1 conjugate cleavage by COP9 signalosome. Science. 2001;292:1382–1385. [PubMed]
  • Martinho RG, Lindsay HD, Flaggs G, DeMaggio A, Hoekstra M, Carr AM, Bentley NJ. Analysis of Rad3 and Chk1 protein kinases defines different checkpoint responses. EMBO J. 1998;17:7239–7249. [PMC free article] [PubMed]
  • Mundt KE, et al. The COP9/signalosome is conserved in fission yeast and has a role in S-phase. Curr Biol. 1999;9:1427–1430. [PubMed]
  • Murray JM, Doe C, Schenk P, Carr AM, Lehmann AR, Watts FZ. Cloning and characterization of the S. pombe rad15 gene, a homologue to the S. cerevisiae RAD3 and human ERCC2 genes. Nucleic Acids Res. 1992;20:2673–2678. [PMC free article] [PubMed]
  • Naumann M, Bech-Otschir D, Huang X, Ferrell K, Dubiel W. COP9 signalosome-directed c-Jun activation/stabilization is independent of JNK. J Biol Chem. 1999;274:35297–35300. [PubMed]
  • Osaka F, Saeki M, Katayama S, Aida N, Toh-E A, Kominami K, Toda T, Suzuki T, Chiba T, Tanaka K, Kato S. Covalent modifier NEDD8 is essential for SCF ubiquitin-ligase in fission yeast. EMBO J. 2000;19:3475–3484. [PMC free article] [PubMed]
  • Osterlund MT, Hardtke CS, Wei N, Deng X-W. Targeted destabilization of HY5 during light-regulated development of Arabidopsis. Nature. 2000;405:662–466. [PubMed]
  • Seeger M, Kraft R, Ferrell K, Bech-Otschir D, Dumdey R, Schade R, Gordon C, Naumann M, Dubiel W. A novel protein complex involved in signal transduction possessing similarities to 26S proteosome subunits. FASEB J. 1998;12:469–478. [PubMed]
  • Tomoda K, Kubota Y, Kato J. Degradation of the cyclin-dependent-kinase inhibitor p27Kip1 is instigated by Jab1. Nature. 1999;398:160–165. [PubMed]
  • Wei N, Chamovitz DA, Deng XW. Arabidopsis COP9 is a component of a novel signaling complex mediating light control of development. Cell. 1994;78:117–124. [PubMed]
  • Wei N, Deng X-W. Making sense of the COP9 signalosome. A regulatory protein complex conserved from Arabidopsis to human. Trends Genet. 1999;15:98–103. [PubMed]
  • Wilkinson CR, Wallace M, Morphew M, Perry P, Allshire R, Javerzat JP, McIntosh JR, Gordon C. Localization of the 26S proteasome during mitosis and meiosis in fission yeast. EMBO J. 1998;17:6465–6476. [PMC free article] [PubMed]
  • Wilson MP, Sun Y, Cao L, Majerus PW. Inositol 1,3,4-trisphosphate 5/6-kinase is a protein kinase that phosphorylates the transcription factors c-jun and ATF-2. J Biol Chem. 2001;276:40998–41004. [PubMed]

Articles from Molecular Biology of the Cell are provided here courtesy of American Society for Cell Biology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • EST
    Published EST sequences
  • Gene
    Gene links
  • GEO Profiles
    GEO Profiles
    Related GEO records
  • HomoloGene
    HomoloGene links
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • Pathways + GO
    Pathways + GO
    Pathways, annotations and biological systems (BioSystems) that cite the current article.
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links
  • Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...