The SET Protein Regulates G 2 /M Transition by Modulating Cyclin B-Cyclin-dependent Kinase 1 Activity*

The SET protein and the cell cycle inhibitor p21 Cip1 interact in vivo and in vitro . We identified here the domain 157 LIF 159 of p21 Cip1 as essential for the binding of SET. We also found that SET contains at least two domains of interaction with p21 Cip1 , one located in the fragment amino acids 81–180 and the other one in the fragment including amino acids 181–277. SET and p21 Cip1 co-localize in the cell nucleus in a temporal man -ner. Overexpression of SET blocks the cell cycle at the G 2 /M transition in COS and HCT116 cells. Moreover, SET inhibits cyclin B-CDK1 activity both in vivo and in vitro in both cell types. This effect is specific for these complexes since SET did not inhibit either cyclin A-CDK2 or cyclin E-CDK2 complexes. SET and p21 Cip1 cooperate in the inhibition of cyclin B-CDK1 activity. The inhibitory effect of SET resides in its acidic C terminus, as demonstrated by the ability of this domain to inhibit cyclin B-CDK1 activity and by the lack of blocking G 2 /M tran- sition when a mutated form of SET lacking this C terminus domain was overexpressed in COS cells. These results

SET was first identified as a gene that was fused to the CAN gene in a patient with acute undifferentiated leukemia, apparently as a result of a translocation (1). SET is a 39-kDa phosphoprotein widely expressed in human and mouse tissues and found predominantly in the cell nuclei, although it has been recently found associated to the endoplasmic reticulum (1)(2)(3)(4). SET belongs to a family of proteins that also includes nucleosome assembly protein 1, testis-specific protein Y-encoded, the suppressor of presenilin 2 protein, the brain protein MB20, the differentially expressed nucleolar transforming growth factor ␤1 target, and the cell division autoantigen 1 (5)(6)(7)(8).
Not much is known about the role of SET in the regulation of cell activities, although several possible functions of this protein have been identified. SET has been found to be identical to the template-activating factor I, a host protein necessary for DNA replication of the adenovirus genome (9). A long acidic domain in the C-terminal region of SET is essential for template-activating factor I activity (9). SET has been also identified as a potent inhibitor of the protein phosphatase 2A (10), a phosphatase involved in the regulation of cell cycle progression (11). Moreover, SET has been shown to bind to nucleosomal histones and in that way protects histones from acetylation by a variety of histone acetyl transferases (11). Because histone acetylation has been involved in the regulation of chromatin condensation and transcription (12), SET might be involved in the regulation of these processes (13). Recently, it has been described that SET binds to the cyclin-dependent kinase (CDK) 1 inhibitor p21 Cip1 in vivo and in vitro and that SET might revert the inhibition produced by p21 Cip1 on cyclin E-CDK2 complexes but not on cyclin A-CDK2 (14). SET also binds to p35 nck5a (15), an activator of the neuronal CDK5 kinase functioning as a cyclin in neurons to support CDK5 activity (16). SET interacts specifically with B-type cyclins, although the functional significance of this interaction has not been elucidated (17). In general, the known functions of SET clearly relate it with the control of cell cycle progression. However, the cell cycle functional steps regulated by SET still remain unclear.
In the work reported here, we analyzed in detail the domains of p21 Cip1 and SET involved in their association, and we have identified that SET might play a role in mitosis by regulating cyclin B-CDK1 activity.
Green Fluorescent Protein (GFP) and DsRed Constructs-The GFP and DsRed expression vectors were gifts from Dr. H. P. Rahn (Munich). All SET constructs were cloned into BamHI-HindIII sites, and p21 Cip1 was cloned into EcoRI-HindIII sites.
Protein Expression and Purification-Glutathione S-transferase (GST) fusion proteins were expressed in Escherichia coli and subsequently purified by absorption to glutathione-Sepharose beads (Amersham Biosciences) as previously described (18). In several cases GST was separated from the GST fusion proteins by digestion with thrombin protease according to the manufacturer (Sigma).
Cell Cultures and Transfections-The colon carcinoma HCT116 cell line and the monkey COS-7 cell line were cultured in Dulbecco's modified Eagle's medium (Biological Industries) containing 10% heat-inactivated fetal calf serum. Both cell types were maintained at 37°C in a humidified atmosphere containing 5% CO 2 . Transfections of COS and HCT166 cells were performed with Effectene (Qiagen) as directed by the manufacturer. Cells were incubated overnight at 37°C, and fresh medium was added after 12-18 h. Then cells were harvested at different times after transfection. To measure the effect of GFP-SET overexpression on CDK1 activity, transfected cultures were subjected to cell sorting to separate GFP-or GFP-SET-expressing cells. Cell sorting gave a 90% of purity. Then separated cells were subjected to immunoprecipitation and kinase assay as described below.
Immunofluorescence-Transfected cells were seeded in culture dishes containing glass coverslips and allowed to grow for at least 24 h. Cells were fixed with ethanol:acetic acid (95:5) and incubated with primary antibodies for 1 h at 37°C. After washing three times (5 min each) in phosphate-buffered saline, cells were incubated with secondary antibodies conjugated with cyanine 3 (Jackson) for 45 min at 37°C. After washing, coverslips were mounted on slides with Mowiol (Calbiochem). Immunofluorescence was recorded using a confocal laser fluorescence microscope.
To analyze the intracellular localization of DsRed-SET and GFP-p21 Cip1 , cells transfected with these plasmids were fixed with 4% paraformaldehyde for 10 min at room temperature. After washing, coverslips were mounted on slides with Mowiol, and immunofluorescence was recorded using a confocal laser fluorescence microscope.
Immunoprecipitation and CDK1 Kinase Assays-Cells were lysed in buffer A (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 5 mM EDTA, 50 mM NaF, 0.1% Triton X-100, 0.5 g/l aprotinin, 10 g/l leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 0.1 mM Na 3 VO 4 ) for 30 min on ice. Lysates (0.1-0.5 mg of protein) were incubated with 2 g of anti-cyclin B1 antibodies overnight at 4°C. Then, protein G-agarose beads (Sigma) were added, and samples were incubated for 1 h at 4°C. After 3 washes in buffer A and 2 in kinase buffer (50 mM Tri-HCl, pH 7.6, 20 mM MgCl 2 ), immunoprecipitates were resuspended in a final volume of 30 l of kinase buffer containing 4 g of histone H1 (Roche Molecular Biochemicals) and 2 mM dithiothreitol. Then they were incubated with different amounts of recombinant SET or/and GST-p21 Cip1 for 10 min. A kinase assay reaction was initiated by the addition of 30 M ATP including 10 Ci of [␥-32 P]ATP (Amersham Biosciences). After incubation for 30 min at 30°C, reactions were stopped by the addition of Laemmli sample buffer. Histone H1 was separated on 12% SDS-polyacrylamide gels, which were then stained with Coomassie Blue and dried. Phosphorylation was detected by autoradiography.
Pull-downs with Purified Proteins-Pull-down experiments were performed as previously described (19). SET protein, the different fragments of SET, and the different peptides and fragments of p21 Cip1 were coupled to BrCN-activated Sepharose 4B, as indicated by the manufacturer. Then, samples (1-4 g of protein) were incubated with 2 g of immobilized protein or peptide in the binding buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM dithiothreitol, 1% Triton X-100) for 2 h at room temperature. The beads were then extensively washed with binding buffer. The unbound fraction was the supernatant of the first wash. The bound fraction was extracted from the beads with Laemmli sample buffer. Unbound and bound proteins were analyzed by SDS-polyacrylamide gel electrophoresis followed by Coomassie Blue staining.
Laser-scanning Confocal Analysis-Cell cycle analysis was studied by laser-scanning cytometry to determine the cellular DNA content in transfected cells. After transfection, cells were fixed with 70% methanol for 2 h at room temperature and washed with phosphate-buffered saline. Then, they were incubated with 50 g/ml propidium iodide (Sigma) and 200 g/ml RNase for 10 min. The coverslips were seeded onto a slide using a mounting medium containing 25% propidium iodide (100 g/ml in phosphate-buffered saline) and 75% glycerol. Slides were scanned using the 20ϫ objective and 5 mW of Argon laser power. A red fluorescence threshold and a minimum cell size of 100 pixels identified cells for measurement. At least 10,000 fluorescent-positive cells were analyzed.
Binding Experiments-To analyze the binding of cyclin B to SET or to the mutated form SET1 ϩ 2, 15 mg of Molt 4 extracts (lysed in buffer A) were loaded onto a SET-Sepharose 4B or SET1 ϩ 2-Sepharose 4B columns (3 mg, each column). After washing in 50 volumes of the same buffer, proteins bound to the column were eluted with the same buffer containing 1 M NaCl. The eluted proteins were subjected to Western blotting using anti-cyclin B1 antibodies.

Identification of the p21 Cip1 Domains That Interact with
SET-We previously reported that SET binds to a peptide of p21 Cip1 containing aa 145-164 (14). However, the binding analysis performed in this previous study did not cover all the p21 Cip1 sequence. To define more precisely the SET binding domains of p21 Cip1 , this protein was fractionated in two parts we named NT domain (aa 1-90) and CT domain (aa 91-164). Then, the binding of SET to these fragments was analyzed by pull-down experiments. Results showed that SET only bound to the CT domain of p21 Cip1 (data not shown).
We subsequently analyzed the binding of SET to four synthetic peptides from this CT domain of p21 Cip1 , 1 (aa 84 -98), 2 (aa 99 -121), 3 (aa 122-139), and 4 (aa 140 -164). Peptides were coupled to Sepharose 4B and pull down experiments were performed. Results showed that SET only significantly bound to peptide 4 (Fig. 1A). These results together with those previously reported (14) indicate that the p21 Cip1 region, including aa 140 -164, contains the SET binding domain of p21 Cip1 . To analyze in detail the putative SET-binding sites in this p21 Cip1 region, we studied the binding of SET to 4 overlapping peptides from this domain named 4A (aa 140 -150), 4B (aa 145-155), 4C (aa 150 -160), and 4D (aa 155-164). Pull-down experiments showed that SET bound to peptides 4A, 4C, and 4D but not to 4B (Fig. 1B), indicating that p21 Cip1 binds to SET by two separated domains (aa 140 -144 and aa 156 -160). To further confirm the relevance of the region including aa 155-160 in the binding to SET, we generated mutated p21 Cip1 fragments by changing aa 157 (Leu) and aa 159 (Phe) to Asp. Specifically, we generated the mutated peptides 4C-DID (aa 150 -160 with the mutations), 4-DID (aa 140 -164 with the mutations), and also the fragment CT-DID (aa 84 -164 with the mutations). The binding analysis clearly indicated that all these fragments containing the mutation did not associate with SET (Fig. 2). Thus, the 157 LIF 159 domain of the C terminus of p21 Cip1 is essential for the SET-p21 Cip1 interaction.
Nuclear Co-localization of SET and p21 Cip1 -COS cells were transfected with GFP-p21 Cip1 constructs alone or together with DsRed-SET, and the cellular localization of both proteins was subsequently analyzed by confocal microscopy. In some cells a clear co-localization between both proteins was observed (Fig.  4A). In many cases a clear co-localization was observed in perinucleolar areas (Fig. 4A). In contrast, in other cells p21 Cip1 and SET were separated (Fig. 4, B and C). These results indicate a temporal better than a constitutive association of SET with p21 Cip1 in the cells. Immunocytochemical studies to detect endogenous proteins using anti-SET and anti-p21 Cip1 antibodies also showed the co-localization of both proteins in a significant number of cells (Fig. 4D). Whether this temporal association is related to specific phases of the cell cycle is under study in our laboratory.
Overexpression of SET Blocks Cell Cycle at G 2 /M-To analyze the possible role of SET in cell cycle regulation we studied the effect of SET overexpression on cell cycle progression. By using laser-scanning confocal analysis, we determined the amount of DNA in COS and HCT116 cells transfected with GFP-SET or GFP. As shown in Fig. 5, SET overexpression induced a clear increase in the number of cells at G 2 /M and a decrease in the number of cells in G 1 in both cell types. These results suggest that overexpression of SET blocks cell cycle at G 2 /M. To further confirm this possibility, experiments analyzing the effect of SET overexpression on histone H3 phosphorylation (this phosphorylation only occurs in mitotic cells) were performed on COS and HCT116 cells. Results revealed that in both cell types transfected with SET, no phosphorylated histone H3 was observed (Fig. 6, A and B). In contrast, cells transfected only with GFP showed this phosphorylation (Fig. 6,  A and B). Overexpression of SET did not block the cell cycle at S phase since transfected cells showed bromodeoxyuridine incorporation (data not shown). These results support the hy-pothesis that SET might be involved in the regulation of G 2 /M transition.
Role of SET on Cyclin B-CDK1 Activity-Evidence presented here showing that overexpression of SET blocked cell cycle progression at G 2 /M and the fact that SET binds to cyclin B (17) suggests that SET alone or together with p21 Cip1 might be involved in the regulation of cyclin B-CDK1 activity. Thus, we first studied the effect of SET overexpression on histone H1 phosphorylation (histone H1 phosphorylation depends on cyclin B-CDK1 activity). These experiments were performed on COS and HCT116 cells transfected with GFP-SET or GFP as a control. Results revealed that cells overexpressing SET did not show H1 phosphorylation, whereas those transfected only with GFP did (Fig. 6, C and D). We subsequently studied the effect of SET overexpression on cyclin B-CDK1 activity. At 48 h after transfection with GFP-SET or GFP alone, COS and HCT116 cells were subjected to immunoprecipitation using anti-cyclin B antibodies. Then immunoprecipitates were analyzed for CDK1 activity. Results indicated that after SET transfection cyclin B-CDK1 activity was clearly inhibited in both cell types (Fig.  7A). To analyze whether SET might directly regulate cyclin B-CDK1 activity, HCT116 cell extracts were immunoprecipitated with specific anti-cyclin B antibodies. Then CDK1 activity was determined in the immunoprecipitates in the presence or absence of exogenous SET. As shown in Fig. 7B, CDK1 activity was inhibited by the addition of 1-2 M purified SET. Inhibition was dose-dependent, and the effect was similar to that observed by recombinant GST-p21 Cip1 (Fig. 7B). Interestingly, this effect was specific for cyclin B-CDK1 because cyclin A-CDK2 and cyclin E-CDK2 activities were not affected by the addition of purified SET (Fig. 8A). SET and p21 Cip1 generate an additive inhibitory effect. The addition of 0.5 M SET and 0.5 M p21 Cip1 inhibited cyclin B-CDK1 activity, whereas at these doses SET and p21 Cip1 added alone only produced a limited inhibition (Fig. 8B). To determine which domain of SET was involved in the inhibition of cyclin B-CDK1 activity, we studied the effect of SET fragments (SET1, SET2, and SET3) on this activity. Results indicate that SET3 fragment contains the cyclin B-CDK1 inhibitory domain (Fig. 8C). To further confirm this possibility, we analyzed the binding of cyclin B to a mu- tated form of SET lacking the SET3 fragment (SET1 ϩ 2). Affinity columns of SET or SET1 ϩ 2 were prepared and loaded with cell extracts. After extensive washing, bound proteins were eluted, and the presence of cyclin B in the eluates was analyzed by Western blot. As shown in Fig. 9A, cyclin B bound to SET but not to SET1 ϩ 2, indicating that SET3 was responsible for cyclin B binding. Overexpression of SET1 ϩ 2 in COS cells showed an intracellular localization both nuclear and cytoplasmic, similar to the distribution of GFP but different to SET-GFP that was mainly nuclear (Fig. 9B). The overexpression of this mutated form of SET was not able to block cell cycle at G 2 /M (Fig. 9C). These results support the SET3 domain as responsible for the regulation of G 2 /M transition. DISCUSSION SET is a cyclin B-interacting protein that recently has been shown to associate with the CDK inhibitor p21 Cip1 (14,17). We report here that SET inhibits cyclin B-CDK1 activity both in vivo and in vitro and that this inhibitory capacity is additive to that of p21 Cip1 . Thus, SET might be involved in the regulation of G 2 /M transition.
Our results revealed that SET binds to two specific domains of p21 Cip1 located at the C terminus (aa 140 -144 and aa 156 -164). The later domain is essential for the binding of SET to p21 Cip1 since specific mutations in this region block the association between both proteins. Thus, domain 140 -144 might help the binding, but it is not enough for maintaining a stable association with SET when domain 156 -164 is mutated. The C terminus of p21 Cip1 , which includes the 156 -164 domain, also binds a number of proteins including proliferating cell nuclear antigen, the E7 oncoprotein of the human papilloma virus, Gadd45, c-Myc, and calmodulin among others (19 -23). This fact suggests an important role for this p21 Cip1 region in the control of cellular functions. Studies to analyze the ability of SET to compete with these other p21 Cip1 -binding proteins for the association with p21 Cip1 are still lacking. Thus, the possible involvement of SET on the modulation of the binding of p21 Cip1 to these important cell cycle regulators is still unknown. SET binds to p21 Cip1 also by two domains, one located at the central region of SET, whereas the other one is placed at the C-terminal region. These results suggest that SET might associate with two p21 Cip1 molecules at the same time or, alternatively, that the two SET-binding sites of p21 Cip1 (aa 140 -145 and 156 -164) might bind to two different domains of SET.
Experimental work to answer this question is under way in our laboratory.
A previous report demonstrated by immunoprecipitation experiments that SET and p21 Cip1 might be associated in vivo (14). Here we demonstrate the in vivo co-localization of both proteins by ectopically expressing p21 Cip1 and SET followed by confocal microscopy analysis but also by immunofluorescence studies using specific antibodies. Results reported here indicate that p21 Cip1 and SET clearly co-localize in the nucleus; however, patterns showing that both proteins might also be separated inside the nucleus were also observed in a significant number of cells. These results suggest that both proteins might interact at specific moments of the cell cycle but not in others. A possibility is that both proteins might interact during midlate S phase. The perinucleolar pattern of co-localization ( Fig.  4A) suggests that both proteins might interact during the mid-S phase since a perinucleolar pattern of replication factories has been observed at this stage of S phase (24). In addition, preliminary results obtained in our laboratory indicate that SET co-localizes with bromodeoxyuridine in the replication factories. Studies to define the kinetics of p21 Cip1 -SET interactions during cell cycle are under way in our laboratory.
Overexpression of SET in two different cellular types, COS and HCT116, induced the inhibition of cell cycle progression at G 2 /M transition. This block is associated with the inhibition of cyclin B-CDK1 activity, indicating that SET might be involved in the control of mitosis entry by regulating the activity of cyclin B-CDK1 complexes. Results from in vitro experiments revealing that SET inhibits cyclin B-CDK1 activity at concentrations similar to those of p21 Cip1 strongly support this hypothesis. Interestingly, the inhibitory effect of SET and p21 Cip1 are additive, suggesting that they might cooperate in regulating cyclin B-CDK1 activity under specific circumstances, as for instance after DNA damage. This cooperation in the inhibition of cyclin B-CDK1 might be similar to that occurring between Gadd45 and p21 Cip1 after DNA damage (25,26). The specific domain of SET involved in cyclin B-CDK1 inhibition is located at the C-terminal region (SET3) that contains a long very acidic domain (aa 225-277). Recently, it has been shown that Gadd45 binds to CDK1 and inhibits the activity of cyclin B-CDK1 complexes (27,28). Interestingly, the Gadd45 domain responsible for the inhibition is an acidic patch including the sequence DEDDDR (28). Moreover, Ran, a small nuclear GTPase implicated in both cell cycle progression and nuclear export (29,30), also contains an acidic motif with similar amino acid composition in its C-terminal domain (DEDDDL) (31). It has been shown that overexpression of Ran induces G 2 /M arrest, whereas deletion of the DEDDDL motif abolishes this activity (29). Interestingly, the C-terminal region of SET contains an acidic sequence 260 DEDDDE 265 that is very similar to those of Gadd45 and Ran, involved in the regulation of cyclin B-CDK1 activity and G 2 /M transition. All these results suggest that this sequence included in the C terminus of SET is the one involved in the inhibition of cyclin B-CDK1 activity and, consequently, in G 2 /M block. However, this still remains to be proved.
Results reported here together with recent reports indicate that SET has a relevant role in the regulation of CDK activities. A previous report revealed that SET blocks the inhibitory effect of p21 Cip1 on cyclin E-CDK2, thus allowing a cyclin E-CDK2 activity in the presence of p21 Cip1 (14). In this case, SET plays a role as an "activator" of these complexes in G 1 . Recently, it has been shown that SET binds to p35 nck5a and p35 nck5ai (15). Both proteins are activators of the neuronal CDK5 (16). SET and p35 nck5a co-localize in the nucleus of cultured cortical neurons, and in vitro experiments indicated that SET enhanced the activity of p35 nck5a -CDK5. Interestingly, the acidic tail of SET is required for the stimulatory effect of p35 nck5a -CDK5 (15). Results reported here indicate that in addition to the positive regulation of CDK2 and CDK5 activities, SET might also play a role as a negative regulator of CDK1. This information reveals that SET plays a dual role in the cell cycle as a positive regulator of G 1 /S and as a negative regulator of mitosis entry. A possibility that has to be investigated is that SET might be involved in blocking cyclin B-CDK1 during S phase and G 2 , avoiding a premature activation of these complexes before cells would be ready to enter mitosis.
SET is also a potent inhibitor of protein phosphatase 2A, which is involved in different cellular functions, including cell proliferation (10). This fact suggest a concerted regulation by SET of phosphorylation/dephosphorylation of nuclear proteins during cell proliferation but also during other cellular processes requiring the activity of CDKs.
Results of this study are the first to demonstrate that SET is a negative regulator of cyclin B-CDK1 in vivo and in vitro and as a consequence of mitosis entry. The fact that SET is also involved in the regulation of chromatin structure by modulating histone acetylation (11) places this protein in a key position in the regulation of cell cycle. On the one hand, SET might be involved in the control the phosphorylation status of nuclear proteins by regulating CDK and protein phosphatase 2A activities. On the other hand, SET might control transcriptional activity of chromatin by regulating histone acetylation and chromatin condensation. The protein p21 Cip1 might participate in these processes, although its role still remains unclear.