Cell Cycle-dependent Regulation of the Cyclin B1 Promoter*

Cyclin B1 mRNA expression varies through the cell cycle with its peak in G 2 /M. In cycling mammalian cells, its lowest level is in G 1 with a steady increase in S until a level 50-fold greater than that in G 1 is reached. In order to characterize the transcriptional component to this variation in expression, we cloned the upstream region 872 base pairs upstream from the start site of the cyclin B1 gene and have demonstrated that it confers cell cycle-dependent regulation onto two reporter genes, both chloramphenicol acetyltransferase and luciferase. Its activity was 25-fold greater in G 2 /M than in G 1 in HeLa cells with intermediate activity in S. This cyclical activity could be seen with sequences encom-passing only 90 base pairs upstream from the start site. Protein binding to this region was demonstrated using electrophoretic mobility shift assays, and the binding profiles appeared to vary depending upon the phase of the cycle in which the extracts are made. Thus, tran- scriptional control plays an important role in determin-ing cyclin B1 mRNA levels, and cell cycle-dependent activity is regulated through interactions with the region 90 bases upstream from the start site. binding sites. The cyclin D1-CAT and DNA polymerase (cid:97) -CAT plasmids were obtained from Bruno Calabretta (Thomas Jefferson University) (14). The hTK-CAT and histone H3.2-CAT plasmids were obtained from Amy S. Lee (University of Southern California) (15). Cell Culture, DNA Transfection, and Cell Synchronization— HeLa cells were grown in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal calf serum, penicillin, and streptomycin at 37 °C in 5%

Cyclin B1 mRNA expression varies through the cell cycle with its peak in G 2 /M. In cycling mammalian cells, its lowest level is in G 1 with a steady increase in S until a level 50-fold greater than that in G 1 is reached. In order to characterize the transcriptional component to this variation in expression, we cloned the upstream region 872 base pairs upstream from the start site of the cyclin B1 gene and have demonstrated that it confers cell cycle-dependent regulation onto two reporter genes, both chloramphenicol acetyltransferase and luciferase. Its activity was 25-fold greater in G 2 /M than in G 1 in HeLa cells with intermediate activity in S. This cyclical activity could be seen with sequences encompassing only 90 base pairs upstream from the start site. Protein binding to this region was demonstrated using electrophoretic mobility shift assays, and the binding profiles appeared to vary depending upon the phase of the cycle in which the extracts are made. Thus, transcriptional control plays an important role in determining cyclin B1 mRNA levels, and cell cycle-dependent activity is regulated through interactions with the region 90 bases upstream from the start site.
Cyclin B1 expression is required for the activation of mitosis or maturation promoting factor. This activation, which triggers the progression into mitosis is dependent both upon the binding of Cdk-1 (synonymous with p34 cdc2 ) to cyclin B1 as well as a series of phosphorylation and dephosphorylation events (1)(2)(3)(4)(5). Cyclin B1 was first identified as a protein whose level oscillated during the sea urchin oocyte cell cycle (6). Subsequently, it was shown that cyclin B1 was expressed ubiquitously in all eukaryotes examined and that the level of cyclin B1 was minimal in G 1 , began to rise in S phase, and peaked at the G 2 /M transition (5,7). The increase in cyclin B1 protein in somatic cycling mammalian cells is mirrored by oscillations in the expression of the cyclin B1 mRNA which increases by over 50-fold as cells progress through G 1 and S into G 2 /M (7)(8)(9). Thus, the increase in cyclin B1 protein that is required to activate maturation promoting factor is likely to depend upon its mRNA level.
The importance of the fluctuations of cyclin B1 mRNA in controlling the cell cycle raises the question of how cyclin B1 mRNA level is regulated. Understanding cyclin B1 mRNA reg-ulation could be important in understanding the mechanisms leading to cell cycle control in G 2 /M. Both transcriptional and post-transcriptional controls are likely to be involved. Pines and Hunter (7), using nuclear run-on assays, suggested that the transcription of cyclin B1 was increased in cells synchronized by thymidine-aphidicolin treatment in G 2 /M as compared to cells in G 1 . Maity et al. (9) showed that the stability of cyclin B1 mRNA also fluctuated through the cell cycle with a half-life of 1-2 h in G 1 , 8 h in S, and 13 h in G 2 /M. While changes in the cyclin B1 mRNA stability are clearly significant, they alone are not of sufficient magnitude to account for the variation in level of 50-fold that is seen in many cycling cells. To further investigate the role of transcription in the cell cycle regulation of the human cyclin B1 gene, we have cloned its 5Ј upstream regulatory region and in this report describe its cell cycle-regulated promoter activity. We have also defined a short region of the promoter that can confer cell cycle-regulated transcription and that shows variation in protein binding through the cell cycle as evaluated by electrophoretic mobility shift assays.

MATERIALS AND METHODS
Library Screening, Polymerase Chain Reaction Amplification, and DNA Sequence Analysis-A human placental genomic library cloned in the Lambda FIX II vector (Stratagene) was screened using a full-length (1.4-kilobase) human cyclin B1 cDNA as a radiolabeled probe. We obtained a positive insert of approximately 20 kilobases. From this clone, we polymerase chain reaction-amplified a 949-bp 1 clone using 5Ј upstream and coding sequences of the human cyclin B1 gene as primers. This 949-bp clone was sequenced using the dideoxynucleotide chain termination method (10). It includes 872 bp upstream from the transcriptional start site identified by Pines and Hunter (7). Piaggio et al. (11) and Cogswell et al. (12) have also identified start sites at slightly different positions from that of Pines and Hunter (7). We have used the designation of Pines and Hunter (7) to position our clones numerically. The 949-bp fragment was subcloned into pCRII (Invitrogen) using TA cloning (13).
Plasmids-The 949-bp clone was subcloned into pCAT-BASIC (Promega) upstream of the chloramphenicol acetyltransferase (CAT) gene at the 5Ј HindIII and 3Ј XbaI sites. To delete part of the 5Ј region, the 949-bp clone was digested with PstI to generate a fragment including 207 bp of upstream sequence and part of the upstream untranslated region of the cyclin B1 gene. This fragment was subcloned into pCAT-BASIC at the PstI site. Two shorter constructs of 140 bp and 90 bp were generated with polymerase chain reaction using sequences at Ϫ140 and Ϫ90 as upstream primers and a sequence at ϩ72 as the downstream primer. The amplified fragments were then subcloned into pCAT-BASIC at the 5Ј HindIII and 3Ј PstI sites. The 949-bp clone was also subcloned into pGL2-BASIC (Promega) upstream of the luciferase gene at the 5Ј SacI and 3Ј KpnI sites. The cyclin D1-CAT and DNA polymerase ␣-CAT plasmids were obtained from Bruno Calabretta (Thomas Jefferson University) (14). The hTK-CAT and histone H3.2-CAT plasmids were obtained from Amy S. Lee (University of Southern California) (15).
Cell Culture, DNA Transfection, and Cell Synchronization-HeLa cells were grown in Dulbecco's Modified Eagle Medium supplemented with 10% fetal calf serum, penicillin, and streptomycin at 37°C in 5% * This work was supported by Grant GM 43759 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  CO 2 . Cells were grown to 50 -70% confluence and then transfected using the calcium phosphate method (10). Briefly, 30 g (for 100-mm dishes) or 18 g (for 60-mm dishes) of DNA in TE (10 mM Tris-HCl, pH 8.0, 1 mM EDTA) and 250 mM CaCl 2 was added dropwise to 2 ϫ HEPES-buffered saline (280 mM NaCl, 10 mM KCl, 1.5 mM Na 2 HPO 4 , 12 mM dextrose, 50 mM HEPES). The calcium phosphate-DNA coprecipitate was allowed to form for 30 min before being added dropwise to cells. The following day, the calcium phosphate-DNA coprecipitate was removed, and the cells were washed twice with serum-free media followed by treatment with one of the following drugs: nocodazole, 0.04 g/ml for 24 h, mimosine, 400 M for 24 h; thymidine, 2 mM for 16 h; or aphidicolin, 1 g/ml for 16 h. Cells were then harvested using Reporter Lysis Buffer (Promega).
␤-Galactosidase, CAT, and Luciferase Assays-For ␤-galactosidase assays, equal amounts of cell lysate and assay buffer (120 mM Na 2 HPO 4 , 80 mM NaH 2 PO 4 , 2 mM MgCl 2 , 100 mM ␤-mercaptoethanol, 1.33 mg/ml o-nitrophenyl ␤-D-galactopyranoside) were mixed and incubated for 30 min before absorbance readings were taken at 420 nm. CAT assays were performed using a liquid scintillation counting method (16). Cell lysates were incubated with [ 3 H]chloramphenicol (50 mCi/ mmol) and n-butyryl coenzyme A (5 mg/ml) for 2-3 h. The reactions were terminated with mixed xylenes, and the organic phase was backextracted twice and then counted in a scintillation counter. For luciferase assays, cell lysates were mixed with luciferase assay reagent (Promega), and light intensity was measured for 10 s on a Berthold LB9501 luminometer.
Flow Cytometry and Cell Cycle Analysis-Cells were collected and resuspended in 40 mM citrate buffer and 5% dimethyl sulfoxide and frozen at Ϫ80°C until analysis. Prior to analysis, cells were thawed and stained with propidium iodide as described previously (17) and analyzed on the Becton-Dickinson FACScan flow cytometer. Data were interpreted using the CellFIT cell cycle analysis program.
Electrophoretic Mobility Shift Assays-Nuclear extracts were prepared as described by Schreiber et al. (18) from HeLa cells that were treated with thymidine-aphidicolin, mimosine, or nocodazole. Oligonucleotides corresponding to the minimal 90-bp cyclin B1 promoter were annealed and 32 P-end-labeled with T4 kinase. DNA-protein complexes were formed in a 15-l reaction mixture containing 20 mM HEPES pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 20% (v/v) glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 1 g/ml pepstatin A, 1 g/ml aprotinin, 2 ng (40 fmol) of the labeled oligonucleotide, 3 g of extract, and 300 ng of salmon sperm DNA (nonspecific competitor) for 20 min at room temperature. The DNAprotein complexes were run on 5% nondenaturing polyacrylamide gels in TGE buffer (25 mM Tris, 200 mM glycine, 1 mM EDTA). The gels were then dried, and autoradiography was performed with an intensifying screen.

Cloning of the 5Ј Upstream Region of the Cyclin B1 Gene-To
isolate the 5Ј upstream sequences of the human cyclin B1 gene, we screened a human placental genomic library with the human cyclin B1 cDNA as a probe. From this clone we subcloned a 949-bp fragment that comprised 872 bases upstream of the 5Ј start site and extended into the transcribed portion of the cyclin B1 gene. Fig. 1 shows the sequence of this clone. Several consensus sequences characteristic of eukaryotic promoter elements are present. At Ϫ25 upstream from the transcriptional start site published by Pines and Hunter (7), there is a "TATAlike" element and at Ϫ60 there is a CAAT box. Further upstream at Ϫ135 there is an Sp1 consensus element.
Promoter Activity of the 5Ј Upstream Region of Cyclin B1-In order to test the upstream sequences for promoter activity, we placed the 5Ј upstream region of the cyclin B1 gene into a chloramphenicol acetyltransferase (CAT) reporter vector, pCAT-BASIC, a plasmid that has no promoter or enhancer sequences. The plasmid containing the 5Ј upstream sequences, pCAT-upcB, was then transfected into HeLa cells using the calcium phosphate method. The transfected cells were treated with various drugs to arrest them in different phases of the cell cycle: nocodazole to block in G 2 /M phase, mimosine to block in late G 1 phase, or thymidine to block in S phase. Fig. 2 shows representative flow cytometry profiles of cell cycle position as determined by DNA content after treatment with these drugs.
After a 24-h treatment with nocodazole, 85-90% of the cells were arrested at the G 2 /M boundary. Mimosine, a plant amino acid that has been shown to arrest cells in late G 1 (19,20), blocked 65-75% of the cells in G 1 after 24 h of treatment. Thymidine blocked 55-65% of the cells in S phase as determined by flow cytometry after a 16-h treatment. Those cells blocked in G 1 at the G 1 /S interface will appear as G 1 cells by this type of analysis. Following cell cycle arrest for 16 or 24 h, the cells were harvested and assayed for CAT activity. A representative experiment is shown in Fig. 3. The transfected cells arrested in G 2 /M by nocodazole had 11-26-fold greater CAT activity than cells arrested in late G 1 by mimosine. Cells arrested in S phase by thymidine had an intermediate level.
pCAT-BASIC had essentially no activity in the transfected cells in any phase. A CAT vector with the SV40 early promoter and enhancer sequences (pCAT-CONTROL) was similarly transfected into HeLa cells which were then blocked in different phases of the cell cycle and showed no variation in activity with the cell cycle. The activity of pCAT-upcB in nocodazole-arrested cells was essentially equivalent to that of pCAT-CONTROL indicating that cyclin B1 promoter activity at that phase of the cycle was strong.
To be sure that the cell cycle differences we observed were specific for the cyclin B1 upstream sequences and not an artifact of the experimental assay, we tested several other promoters in the same system. We transfected plasmids containing the human DNA polymerase ␣, the human cyclin D1, the human thymidine kinase, or the hamster histone H3.2 promoters (14, 15) upstream of the CAT reporter gene into HeLa cells and then treated the cells with either nocodazole, mimosine, or thymidine. Each of these promoters has been characterized as being specific for S phase (14,15). The human DNA polymerase ␣ and the human cyclin D1 promoters have both been characterized as being dependent upon E2F activation (14,(21)(22)(23)(24). Since HeLa cells are effectively deficient for the retinoblastoma protein (pRb) due to the binding of pRb to the E7 protein of human papilloma virus (25), and the regulation of E2F transactivation is affected by its interaction with pRb, these cells may not regulate E2F-activated genes in the same way as cells having normal pRb activity (26). Cells deficient in pRb, for example, bypass the requirement for cyclin D1 in G 1 (26). In evaluating these experiments, it should also be noted that the CAT protein has a half-life of about 50 h (27). The long half-life can affect the interpretation of experiments because once a cell has synthesized the CAT protein, CAT activity would be detectable for the duration of the experiment. Since the majority of the cells in an exponentially growing population of HeLa cells are in either G 1 or S phase, the cells in S would lead to considerable synthesis of the CAT protein by an S phase-specific promoter. The CAT protein because of its long half-life would remain detectable as cells progressed into G 1 . In contrast, few cells would be in G 2 /M and spend little time at that point since it is the shortest component of the HeLa cell cycle. Thus, this method should clearly distinguish transcription in G 2 /M from that in G 1 or S, but would not be expected to clearly separate transcription in G 1 from S phase.
After transient transfection and cell cycle blockade, the S phase-specific promoters showed 1.3-3.4 times more CAT activity in S phase than in G 1 and slightly more CAT activity in S than in G 2 /M, as shown in Fig. 4. In contrast, however, the cyclin B1 upstream region vector had markedly higher CAT activity in G 2 /M compared to other phases of the cell cycle. Thus, it is unlikely that the cell cycle-regulated activity of the cyclin B1 promoter is an artifact of our experimental assay or of the drugs used to induce blockage in the cell cycle. These experiments indicate that the upstream region of the human cyclin B1 gene has promoter activity that is low in G 1 , higher in S, and maximal at G 2 /M.
To further confirm the cell cycle-regulated activity of the cyclin B1 promoter, we subcloned the cyclin B1 promoter upstream of the luciferase reporter gene in the pGL2-BASIC vector (pGL2-upcB). While the CAT mRNA has a short halflife, its protein has a long half-life of 50 h. In contrast, both the luciferase mRNA and protein have short half-lives. The halflife of the luciferase protein is approximately 3 h (27). Thus, we could transfect a luciferase expression vector driven by the cyclin B1 promoter, follow the cells as they cycled, and assess the activity of the cyclin B1 promoter throughout the cell cycle by monitoring luciferase activity. These experiments did not depend upon pharmacological blockade of the cell cycle. HeLa HeLa cells were transfected with one of the following CAT plasmids: the SV40 promoter and enhancer sequences (pCAT-CONTROL), no promoter or enhancer sequences (pCAT-BASIC), or the 949-bp upstream region of cyclin B1 (pCAT-upcB) upstream from CAT. A ␤-galactosidase plasmid (pSV␤-gal) was included in the transfection as an internal control for transfection efficiency. 18 h after transfection, the cells were treated with either mimosine to block in G 1 , thymidine to block in S, or nocodazole to block in G 2 /M. Cells were then harvested and assayed for CAT and ␤-galactosidase activity. After normalization to ␤-galactosidase values, the ratios of CAT activity in different phases of the cell cycle were the same as the ratios when comparing absolute CAT activity levels. CAT activity shown here has not been normalized. cells were cotransfected with pGL2-upcB and pSV␤-gal (as an internal control for transfection efficiency) and treated with aphidicolin for 16 h. After the cells were released from the aphidicolin block, cells were harvested every 3 h for a 24-h period. Cell lysates were assayed for luciferase and ␤-galactosidase activity and by flow cytometry for DNA content. As shown in Fig. 5A, the cells were predominantly in G 1 at the start of the experiment. At this point, the level of luciferase activity was low. As the cells entered S phase, luciferase activity rose. Luciferase activity peaked (at 9 times greater than the level at the time of release) 10 h after release when the majority of cells were in G 2 /M. Once the cells exited mitosis and reentered G 1 , luciferase activity began to drop. At 21 h after release, only 11% of the cells remained in G 2 /M while 62% were in G 1 . At this point, the luciferase activity had fallen to a level only twice that of the level at the time of release from the aphidicolin block. This decline in luciferase activity was not a result simply of activity declining with time after transfection, but was a result of cell cycle progression. This can be seen from the experiment shown in Fig. 5B. Some of the transfected synchronized cells from the experiment shown in Fig. 5A were also treated with nocodazole. By blocking the cells in G 2 /M and following them with time, we could assess the effect that increasing time after transfection had on the luciferase activity. As shown in Fig. 5B, nocodazole blocked the cells in G 2 /M, yet luciferase activity remained high throughout the remainder of the experiment. Thus, the cyclin B1 promoter conferred cell cycle variation to the expression of the luciferase reporter gene. It resulted in the lowest level of luciferase activity in G 1 , an increasing level in S phase, and the maximal level in G 2 /M corresponding to a pattern paralleling the level of cyclin B1 mRNA in cycling cells.
Identification of a Minimal Promoter Sequence-To map the sequences responsible for conferring the cell cycle-regulated transcription of cyclin B1, we generated a series of plasmids containing smaller fragments of the cyclin B1 promoter. The full-length upcB construct was further truncated to generate fragments that are 207, 140, and 90 bp upstream of the transcriptional start site. All of these constructs included the putative TATA element and the CAAT box. The Ϫ207 and Ϫ140 constructs include the Sp1 consensus site, whereas the Ϫ90 construct does not. These constructs were then subcloned into pCAT-BASIC and used in transient transfection assays as described in Fig. 3. As shown in Fig. 6, the Ϫ207 construct was found to have 7-fold greater CAT activity in transfected cells blocked in G 2 /M compared to transfected cells blocked in late G 1 . The Ϫ140 and Ϫ90 constructs were found to have 16-fold greater CAT activity in transfected cells blocked in G 2 /M compared to transfected cells blocked in late G 1 . These results indicate that a fragment as small as 90 bp upstream of the transcriptional start site of the human cyclin B1 gene is suffi- FIG. 4. Promoter activity of S phase-specific promoters. HeLa cells were transfected with S phase-specific promoter-CAT plasmids (human DNA polymerase ␣, human cyclin D1, human thymidine kinase, or hamster histone H3.2) and treated with either mimosine to block in G 1 , thymidine to block in S, or nocodazole to block in G 2 /M. Cells were then harvested and assayed for CAT activity. Cotransfection with pSV␤-gal indicated equivalent transfection efficiency in all samples.
FIG. 5. Cyclical activity of the cyclin B1 promoter. A, HeLa cells were co-transfected with a plasmid containing the promoter of cyclin B1 upstream of the luciferase reporter gene (pGL2-upcB) and a ␤-galactosidase plasmid (pSV␤-gal) as an internal control for transfection efficiency. Cells were then synchronized with aphidicolin for 16 h, released, and harvested for luciferase and ␤-galactosidase activity and flow cytometry every 3 h for a 24-h period. To control for transfection efficiency, luciferase values were normalized to ␤-galactosidase values. B, some of the transfected synchronized cells were treated with nocodazole to keep them arrested in G 2 /M. Cells were then harvested for luciferase and ␤-galactosidase activity and flow cytometry. cient for cell cycle-regulated promoter activity.
Electrophoretic Mobility Shift Assays Using the Minimal Cyclin B1 Promoter-In order to search for proteins that may be important for the cell cycle-regulated transcription of the human cyclin B1 gene, we performed electrophoretic mobility shift assays using the cyclin B1 promoter as the DNA probe. The 90-bp region of the cyclin B1 promoter that was found to be sufficient for conferring cell cycle-regulated activity was 32 Pend-labeled and incubated with nuclear extracts prepared from HeLa cells that were synchronized and harvested when the majority of cells had exited from mitosis into G 1 (early G 1 ), cells that were blocked in G 1 with mimosine (late G 1 ), or blocked in G 2 /M with nocodazole. As shown in Fig. 7, four bands were seen when the 90-bp fragment of the cyclin B1 promoter was incubated with G 1 extracts. The arrow indicates the position of a fainter, but reproducible band that was seen when extracts were made from late G 1 cells, but not from extracts of early G 1 cells. When the assay was performed using extracts from nocodazole-blocked cells, only three faint bands were seen. This suggests that there may be inhibitory protein(s) present during G 1 that down-regulate transcription of cyclin B1. DISCUSSION In this study, we have described the isolation of the 5Ј upstream sequences of the human cyclin B1 gene, and, using two different experimental assay systems, we have shown that these sequences show promoter activity that is strongly cell cycle-regulated. Its activity was lowest in G 1 , rose to an intermediate level in S phase, and reached a peak in G 2 /M. This activity parallels the pattern of expression of the cyclin B1 mRNA. We found that the cyclin B1 promoter was 11-26 times more active in G 2 /M than in G 1 . Previous studies have shown that the level of cyclin B1 mRNA present during G 2 /M is up to 50-fold greater than the amount of mRNA present in G 1 . Data from our laboratory have shown that variations in cyclin B1 mRNA stability during the cell cycle contribute to the vast differences in the cyclin B1 mRNA level (9). The half-life of the cyclin B1 message is 13 h in G 2 and 2 h in G 1 . The cyclical activity of the cyclin B1 promoter, along with the differences in the cyclin B1 half-life, suggest that both changes in the rate of transcription and changes in mRNA stability are responsible for the cell cycle-regulated expression of this gene.
Piaggio et al. (11) have recently cloned the upstream region of the human cyclin B1 gene. They found that the upstream region had promoter activity that was minimal in quiescent or G 0 cells, but increased in activity gradually after serum stim-ulation. However, they were unable to show any cell cyclespecific regulation of activity. In their experiments, the activity of the cyclin B1 promoter did not fall following mitosis and entry into G 1 . This may have been due to their use of chloramphenicol acetyltransferase (CAT) as their reporter gene, since CAT has a half-life that is considerably longer than the cell cycle time of about 24 h in the cells they used. High levels of CAT protein present during mitosis would still be present once cells cycled back into G 1 .
Cogswell et al. (12) have also reported the isolation of a cyclin B1 promoter element using the upstream sequences from the cyclin B1 gene. They found that the activity of the cyclin B1 promoter was enhanced in vivo during G 2 /M as compared to S. Their experiments also use CAT as the reporter gene and hence do not show cyclical activity of the cyclin B1 promoter. In this study, Cogswell et al. (12) identified a region as being important for cell cycle regulation that is distal to the Ϫ90 region that we studied. This region contained a binding site for the upstream stimulatory factor and is required for in vitro transcription of the cyclin B1 gene. Using an electrophoretic mobility shift assay, they were able to identify upstream stimulatory factor binding to this region. This binding was greater in extracts from G 2 /M phase cells than in S phase cells, suggesting that upstream stimulatory factor stimulates transcription of cyclin B1 during G 2 /M. This is in contrast to our finding that suggests the presence of inhibitory protein(s) that repress cyclin B1 transcription during G 1 . Together these experiments suggest that regulation of the cyclical activity of the cyclin B1 promoter may be controlled by several regions and that their action may be redundant.
The isolation of the cyclin B1 promoter and its further characterization will provide us with a better understanding of how the cyclin B1 gene is regulated and thus how transcription may be triggered in a cell cycle-specific fashion. The cdc25C promoter has been identified as being specific for G 2 /M transcription (28). Cdc25C dephosphorylates the inhibitory phospho-FIG. 6. Promoter activity of truncations of the cyclin B1 promoter. The full-length cyclin B1 promoter (Ϫ872) was truncated to generate fragments of 207, 140, and 90 bp upstream of the transcriptional start site and subcloned into a promoterless CAT vector (pCAT-BASIC). HeLa cells were transfected with these truncated cyclin B1 promoter-CAT plasmids and treated with nocodazole to block in G 2 /M or mimosine to block in G 1 . Cells were then harvested and assayed for CAT activity indicated as counts/min. FIG. 7. Electrophoretic mobility shift assays using the minimal cyclin B1 promoter. Electrophoretic mobility shift assays were performed using double-stranded oligonucleotides corresponding to the cyclin B1 sequence from Ϫ90 to ϩ1 as the radiolabeled probe. Nuclear extracts were prepared from HeLa cells that were blocked in late G 1 with mimosine or in G 2 /M with nocodazole. HeLa cells were synchronized in early G 1 by a double block of thymidine and aphidicolin (35,36). After release from the final block, the cells were cultured until they had progressed through the cell cycle into early G 1 (2-3 h after completion of mitosis) (9). These cells were used in the preparation of the early G 1 extracts. Extracts prepared from cells in these different phases of the cell cycle were incubated with the probe as described under "Materials and Methods." Arrows indicate a faint, but reproducible band seen in assays from late but not early G 1 cells. rylations on p34 cdc2 (cdk-1) and is required for activation of maturation promoting factor and for transition into M and G 1 (3,5). Its mRNA cycles with the highest levels in G 2 /M and the lowest in G 1 , yet in spite of the cyclical expression of the mRNA, the protein itself appears to be present in constant amounts through the cell cycle making the physiological significance of the cell cycle regulation of cdc25C mRNA difficult to understand (29,30). The cdc25C promoter lacks a canonical TATA element (28). An element termed the cycle-dependent element was identified to act as a cis-acting repressor element that is responsible for the low level of expression of cdc25C during G 0 and G 1 . This cycle-dependent element contains the sequence GGCGG that when mutated resulted in increased expression of the promoter in G 1 . The cyclin B1 gene contains in the 5Јuntranslated region of its mRNA a GGCGG sequence. However, the adjacent bases are not homologous to the cycle-dependent element of cdc25C. Furthermore, the GGCGG sequence that was identified as important in the cycle-dependent element is located downstream of the transcriptional start site of cyclin B1 making it less likely to play a role in the regulation of cyclin B1, but not excluding that possibility.
The finding of promoters that are actively transcribed during G 2 /M raises another interesting question with respect to cell cycle-dependent transcription. Generally, transcription is turned off during mitosis at which point the chromosomes are condensed (31,32). Since the cyclin B1 gene is maximally active during this phase of the cell cycle, there must be novel mechanisms that allow transcription of this gene to occur under conditions in which transcription generally is repressed. Further study of this promoter should reveal important insights into transcriptional control during chromosome condensation and perhaps will reveal regions protected from chromosome condensation.
Currently, the B-type cyclins are a well-characterized class of proteins, but the mechanisms underlying their transcriptional regulation are poorly understood. Although genes have been identified in yeast that regulate the transcription of cyclin B1, their mammalian homologs are unknown (33,34). A transacting factor that acts positively during the G 2 /M transition or negatively during G 1 has yet to be identified in mammalian cells. The existence of such proteins is likely given that the cyclin B1 promoter is maximally active during G 2 /M and given the differential patterns of gel retardation found through the cell cycle using portions of the cyclin B1 promoter.