Ubiquitin-dependent Degradation of Cyclin B Is Accelerated in Polyploid Megakaryocytes*

During the endomitotic cell cycle of megakaryocytic cell lines, the levels of cyclin B1 and the activity of cyclin B1-dependent Cdc2 kinase, although detectable, are reduced as compared with megakaryocytes undergoing a mitotic cell cycle. The levels of cyclin A, however, are comparable during both cell cycles. The expression of cyclin B1 mRNA is also equivalent in proliferating and polyploidizing cells. In the current study, we found that the rate of cyclin B1 protein degradation is enhanced in polyploidizing megakaryocytes. This finding has led us to further investigate whether the ubiquitin-proteosome pathway responsible for cyclin B degradation is accelerated in these cells. Our data indicate that polyploidizing megakaryocytic cell lines and primary bone marrow cells treated with the megakaryocyte proliferation- and ploidy-promoting factor, the c-Mpl ligand, display increased activities of the ubiquitin-proteosome pathway, which degrades cyclin B, as compared with proliferating megakaryocytic cell lines or diploid bone marrow cells, respectively. This degradation has all the hallmarks of a ubiquitin pathway, including the dependence on ATP, the appearance of high molecular weight conjugated forms of cyclin B, and inhibition of the proteolytic process by a mutated form of the ubiquitin-conjugating enzyme Ubc4. Our studies also indicate that the ability to degrade cyclin A is equivalent in both the mitotic and endomitotic cell cycles. The increased potential of polyploid megakaryocytes to degrade cyclin B may be part of the cellular programming that leads to aborted mitosis.

During the mitotic cell cycle, cyclin B steadily accumulates at interphase and is rapidly degraded before the cells exit mitosis (reviewed in Ref. 1). Cyclin B degradation is essential for progression through to the next cell cycle. The specificity of cyclin B degradation is determined by the 9-amino acid motif conserved among the N termini of B-type cyclins (cyclin Box) (reviewed in Refs. 1 and 2). Mutation of the conserved arginine in the cyclin box inhibits ubiquitination of cyclins in Xenopus extracts, indicating that the cyclin box serves as a signal for ubiquitination (3,4). Cyclin A and B2, but not B1, require binding to p34Cdc2 for their proteolysis (5,6). While some components involved in cyclin B ubiquitination and degrada-tion have been identified, such as the ubiquitin-conjugating enzymes Ubc4 (4) and Ubc10 (7) and enzymes Cdc16p, Cdc26p, Cdc27p, and Cdc23p in yeast (reviewed in Ref. 8), much has yet to be explored.
Megakaryocytes undergo a unique cell cycle resulting in DNA replication in the absence of complete mitosis and cytokinesis. We and others found that during this cell cycle, DNA synthesis is not continuous, but rather consists of an S phase (of approximately 9 h) and a gap (of approximately 1-2 h) (9 -11). Studies aiming to investigate the nature of the gap phase during endomitosis led to the conclusion that mitosis is initiated but aborted at entry to anaphase (13). It was found that polyploid megakaryocytes derived from bone marrow cells treated with the ploidy-promoting factor, the c-Mpl ligand (reviewed in Ref. 12), can condense their chromosomes when treated with nocodazole (13). In this study, polyploid megakaryocytes were also found to display early features of prometaphase, such as spindle formation but not any characteristics of anaphase. Because of the rarity of megakaryocytes in the bone marrow and the inability to obtain a pure population of diploid megakaryocytes to be used for synchronization and for following biochemical parameters during the development of high ploidy cells, we and others have resorted to megakaryocytic cell lines. It was found that during the endomitotic cell cycle in MegT and HEL cells the activity of the mitotic kinase, cyclin B1-dependent Cdc2 kinase, is reduced as compared with the activity during the mitotic cell cycle in these cells (10,14). Under these conditions, however, the level of cyclin B1 protein is reduced, while the level of cyclin B1 mRNA is comparable in proliferating and polyploidizing megakaryocytes (10). Carrow et al. (15) detected cyclin B1 and a cyclin B1and Cdc2-dependent kinase activities in nonsynchronized, high ploidy bone marrow megakaryocytes treated with the c-Mpl ligand, although it has not yet been investigated how the level of activity of these kinases compares with that in proliferating diploid megakaryocytes or other bone marrow cells. In this context, it should be mentioned that while cyclin B levels peak prior to the early stages of mitosis, some level of cyclin B and cyclin B-dependent Cdc2 kinase are also detected during S phase and are essential for progression through this phase of the yeast cell cycle (16). Early stages of mitosis, such as chromosome condensation, could be initiated by the never-in-mitosis (NIMA)-type kinases in the absence of cyclin B-dependent Cdc2 kinase activity (17,18). Abrogation of mitosis during polyploidization could result from a variety of reasons, including a premature degradation of cyclin B. This latter possibility could lead to the detection of lower levels of this cyclin during endomitosis. In the current study, we sought to investigate whether the activity of the ubiquitin-proteosome pathway responsible for cyclin B degradation is altered in polyploidizing megakaryocytes. Our data indicate that polyploidization is associated with increased activity of the ubiquitin pathway re-sponsible for the recognition and proteolysis of cyclin B. This feature may be a part of cellular programming that leads megakaryocytes to abort mitosis during polyploidization.

MATERIALS AND METHODS
Culture Conditions-MegT cells (clone 37C1) were grown in a liquid culture, as described before (19). To induce ploidy, 1 ϫ 10 6 cells were seeded into a 75-cm 2 culture flask and incubated in 5% CO 2 at 39.5°C for 4 or 5 days (10,19). Cells were counted by hemocytometer, and cell viability was followed by staining with trypan blue. The attached cells as well as the detached cells were separately collected and subjected to various analyses. A clone (Y10/L8057) of the mouse megakaryocytic cell line L8057 (20), selected by limited dilution based on its ability to express acetylcholinesterase (a unique marker for rodent megakaryocytes) (21) and to respond to 25 ng/ml PEG-rHuMGDF 1 (generous gift of Amgen, Inc. Thousands Oaks, CA) by polyploidizing. Acetylcholinesterase activity was assayed as described before (9). Culture conditions for this cell line were as described elsewhere (20). Rat bone marrow cells were isolated from the femurs also as described before (22). Cells were cultured in 5% CO 2 at 37°C in a liquid culture under conditions that were previously shown to support maturation and ploidy of primary megakaryocytes (22,23). The cells were cultured in the presence of Iscove's modification of Dulbecco's medium (IMDM) and 20% horse serum supplemented with penicillin (2000 units/ml), streptomycin (200 g/ml), L-glutamine (0.592 mg/ml), and when indicated with 50 ng/ml PEG-rHuMGDF for 3 days prior to collection of large mature megakaryocytes.
Purification of Megakaryocytes-Polyploidizing megakaryocytes of varying sizes (20-m diameter and higher) were isolated from bone marrow cells derived from rat femurs (22) and cultured for 3 days in the presence of 50 ng/ml of PEG-rHuMGDF, to promote megakaryopoiesis, and in the presence of a mixture of growth factors including 1 unit/ml erythropoietin (Amgen), 10 units/ml of granulocyte-macrophage colonystimulating factor, and 5 units/ml of stem cell factor (Genzyme Corp., Boston, MA) to promote proliferation and survival of different bone marrow lineages, as we described before (22). The isolation procedure involved mesh filtration (nylon mesh screens from Spectramesh, Spectrum Medical, Inc., Los Angeles, CA). To this end, clumps were initially removed by filtration through a 200-m filter followed by filtration with a 17-m filter, which allowed diploid cells to pass through the mesh. Cells retained on the filter were retrieved by washing the inverted filter with phosphate-buffered saline (136 mM NaCl, 8 mM Na 2 HPO 4 , 2.6 mM KCl, 1.4 mM KH 2 PO 4 , pH 7.4). This procedure was repeated twice. We confirmed by counting cells under the light microscope and by staining for the unique marker for megakaryocytes, acetylcholinesterase (21), that megakaryocytes were enriched to account for 60 -70% of bone marrow cells (see Fig. 8A), a level comparable with the purification method involving elutriation (24). Because of the size and content of megakaryocytes, in comparison with the small diameter bone marrow cells, the enrichment of megakaryocytic material (e.g. protein) would be substantially higher.
Western Blotting-Y10/L8057 cells grown in suspension were collected by centrifugation (380 ϫ g, 5 min) in growth medium, and MegT cells adhering to the culture dish were stripped from the plate with a cell scraper, while nonadhering cells were collected separately by centrifugation. Cells were washed twice with cold phosphate-buffered saline and then lysed in lysis buffer (0.5% Nonidet P-40, 50 mM Tris (pH 7.4), 250 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml n-tosyl-L-phenylalanine chloromethyl ketone, 10 g/ml soybean trypsin inhibitor) followed by centrifugation at 15,000 ϫ g for 5 min. For Western blotting, 10 g of lysed proteins were separated by 7.5 or 12% SDS-polyacrylamide gel electrophoresis (25) and electrophoretically transferred from the gel onto a nitrocellulose membrane (Schleicher & Schull) in a buffer containing 25 mM Tris, 192 mM glycine, and 20% methanol. The membrane was stained with 0.1% Ponceau S (Sigma) in 5% acetic acid to visualize protein bands. The membrane was destained with water, washed in TBS (10 mM Tris (pH 8.0), 150 mM NaCl) and blocked for 1 h with TBST (Tris-buffered saline with 0.1% Tween 20) containing 5% dry milk. The blot was incubated for 1 h at room temperature in the presence of 10 ml of TBST supplemented with a monoclonal antibody to cyclin B1 (GNS-1 diluted to 0.2 g/ml), gen-erated against the N-terminal domain of cyclin B between amino acids 1 and 21 (Pharmingen, San Diego, CA). This antibody cross-reacts with the corresponding murine protein, as tested by the manufacturers and by ourselves (10). When indicated, a polyclonal antibody to cyclin B1, which does not block exclusively the N-terminal region of cyclin B, was used (26). A limited amount of this antibody was available to us, and thus most of the studies were performed with the commercial monoclonal antibody to cyclin B1. The blot was washed four times, each time for 10 min, and incubated for 1 h with appropriate horseradish peroxidase-labeled secondary antibody (Amersham Corp.) at a 1:1500 to 1:3000 dilution in TBST. The blot was washed four times (each for 10 min) with TBST, and the Enhanced Chemiluminescence system (Amersham) was used for detection of proteins, as instructed by the manufacturer.
Preparation of Cytosolic Extract and Degradation Assay-Cells were washed twice with phosphate-buffered saline by centrifugation (380 ϫ g, 5 min). The volume of the pelleted cells was estimated, and the cells were washed again by centrifugation in the presence of 5 volumes of hypotonic buffer (10 mM Tris (pH 7.6), 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, 0.5 mM phenylmethanesulfolyl fluoride). Cells were resuspended in 2 volumes of hypotonic buffer and left on ice for 10 min prior to breaking by strokes in a glass homogenizer. Cells were spun down for 10 min in a microcentrifuge. Protein assay was performed on the supernatant as described in Ref. 10. Since adequate degradation was obtained with this preparation (27), we did not need to further clean the cytoplasmic extract by high speed centrifugation (100,000 ϫ g). Degradation of recombinant cyclin B was monitored as follows. Cytosolic extract (25 g of protein) was added to a reaction mixture containing the following components in a final volume of 25 l: 12 mM Tris (pH 7.5), 60 mM KCl, 3.5 mM MgCl 2 , 1 mM ATP (or ATP␥S), 20 mM creatine phosphate, 50 -200 ng of recombinant cyclin B1 or cyclin A (Pharmingen), purified from salt by buffer exchange with an Amicon 30-kDa Filtration Unit (Amicon Beverly, MA). The reaction was carried out at 30°C for the indicated time. When indicated, the cellular extract was pretreated with 0.15 units of apyrase (Sigma) for 20 min at 30°C, after which the reaction was processed in the absence of ATP or creatine phosphate. Ubc4 and Ubc2 were obtained as described elsewhere (28). Point mutations in Ubc4 and Ubc2 were generated by oligonucleotidedirected mutagenesis as described in Ref. 28.
Ploidy Analysis-Bone marrow was collected from femurs of rats as described before (22,23). Marrows were suspended in 0.5 ml of CATCH buffer and incubated for 60 min at 4°C with 4A5 monoclonal antibody ascites (29) (generous gift of Sam Burstein, University of Oklahoma Health Sciences Center). The cells were then incubated with fluorescein-conjugated goat anti-rat IgG(FabЈ) 2 (Biosource International, Camarillo, CA) for 30 min at 4°C. Staining with propidium iodide was followed by determination of DNA content of 4A5 positive cells using a statistical package on a FACScan flow cytometer (Becton-Dickinson) as we described previously (10). Y10/L8057 cells were stained with propidium iodide and subjected to ploidy analysis also as described before (10).

Levels of Cyclin B1 and Cdc2 and the Half-life of Cyclin B1 in the Proliferating and Polyploidizing Megakaryocytic Cells Y10/
L8057-In the current study, we assayed for cyclin B1 and Cdc2 levels during polyploidization of a megakaryocytic cell line, Y10/L8057, which we subcloned from L8057 cells (20) (see "Materials and Methods"), based on its ability to polyploidize in response to the megakaryocyte ploidy-promoting factor PEG-rHuMGDF (Fig. 1). In the presence of PEG-rHuMGDF or of phorbol 12-myristate 13-acetate (PMA), a moderate or large fraction, respectively, of the cells underwent polyploidization, while the rest remained in the 2N proliferative mode (Fig. 1). Western blot analyses indicated that during polyploidization induced by these agents, the levels of cyclin B1, but not of Cdc2, were reduced (Fig. 2). We next explored the possibility that reduced levels of cyclin B reflected an accelerated degradation of this protein in polyploidizing cells. As shown in Fig. 3, the half-life of cyclin B1 in polyploid Y10/L8057/L8057 cells was substantially shorter than in proliferating cells. The enhanced degradation and reduced levels of cyclin B1 in polyploidizing megakaryocytes were not a result of cell cycle arrest, since cells induced to polyploidize displayed high rates of DNA synthesis (Table I).
Accelerated Degradation of Recombinant Cyclin B by a Ubiquitin-Proteosome Extract Derived from Polyploidizing Y10/ L8057 Cells-A shortened half-life of cyclin B could result from an enhanced activity of the ubiquitin-proteosome pathway in polyploidizing cells. To test this possibility, cytosolic extracts were prepared from proliferating megakaryocytes or cells undergoing endomitosis. These extracts were incubated with recombinant cyclin B1, and cyclin B1 degradation was followed. As shown in Fig. 4, a significant degradation of recombinant cyclin B was displayed by extracts from highly polyploidizing Y10/L8057 cells (treated with PMA), less from cells displaying moderate polyploidization (treated with PEG-rHuMGDF), but not in proliferating cells. A high level of cyclin B1 degradation (in PMA-treated cells) was dependent on the addition of ATP to the reaction, as typical of a ubiquitin-dependent degradation (30) (Fig. 4). Cellular ATP was sufficient, however, to promote a low level of cyclin B degradation in the PEG-rHuMGDFtreated cells.
Degradation of Cyclin A Is Not Accelerated in Polyploidizing Y10/L8057 Cells-Cyclin A protein is highly expressed in the megakaryocytic cell line, MegT, in HEL cells (10,31) and in Y10/L8057 cells undergoing polyploidization (Fig. 5A). Since cyclin A is also subjected to degradation via the ubiquitin pathway (reviewed in Ref. 32), we examined whether the ability of polyploid megakaryocytes to degrade this cyclin is accelerated as was the case for cyclin B1. Interestingly, and in accordance with the detection of high levels of cyclin A in polyploid megakaryocytes, the ability of this latter cell type to degrade recombinant cyclin A was comparable with the rate displayed by proliferating cells (Fig. 5B).

Accelerated Degradation of Recombinant Cyclin B by a Ubiquitin-Proteosome Extract Derived from Polyploidizing MegT
Cells-MegT cells are megakaryocytes that have been immortalized with the temperature-sensitive large T antigen (19). At the permissive temperature (34°C), the cells proliferate. At the nonpermissive temperature (39.5°C), a fraction of the cells remain adhering to the dish and undergo a mitotic cell cycle, while the other fraction (in which T antigen is destroyed) detaches from the plate and undergoes an endomitotic cell cycle (10). During this endomitotic process, the level of cyclin B protein, but not of cyclin A protein nor of cyclin B mRNA, is reduced, as compared with proliferating cells (10). Cytosolic extracts were prepared from these proliferating megakaryocytes or from cells undergoing endomitosis. These extracts were incubated with recombinant cyclin B1, and cyclin B1 degradation was followed. As shown in Fig. 6A, a significant degradation of recombinant cyclin B was displayed by extracts prepared from polyploidizing cells. Cyclin B degradation was prevented in the presence of apyrase, which degrades ATP, or in the presence of the nonhydrolyzable (by the proteosomes) ATP analog, ATP␥S. An additional hallmark of a ubiquitin-dependent degradation is the appearance of high molecular weight conjugates of cyclin B and ubiquitin. To detect these high molecular weight complexes, we also used a polyclonal antibody to cyclin B protein. This antibody, in contrast to the monoclonal antibody used in Fig. 6A, is not directed against the N-terminal domain of cyclin B and thus should recognize ubiquitin-bound cyclin B. Indeed, as shown in Fig. 6B, high molecular weight conjugates of recombinant cyclin B appeared only in polyploid cells, depending on the availability of ATP.

PEG-rHuMGDF-treated Primary Bone Marrow Cells Display an Increased Ability to Degrade Recombinant Cyclin B-To test
that the phenotype observed in Y10/L8057 and MegT cells is not exclusive to these cell lines, we prepared extracts from primary rat bone marrow cells cultured for 3 days in the absence or presence of PEG-rHuMGDF (reviewed in Ref. 12). Under our liquid culture conditions, PEG-rHuMGDF did indeed increase the frequency and ploidy of megakaryocytes (Table II). A cellular extract prepared from PEG-rHuMGDFtreated bone marrow cells degraded recombinant cyclin B more efficiently than extract from diploid bone marrow cells (Fig. 7). This was accompanied by a large shift of cyclin B to high molecular weight forms. It is technically impossible at this stage to purify significant quantities of proliferating, small megakaryocytes from primary bone marrow, thus preventing us from comparing the extent of cyclin B degradation in diploid versus polyploid megakaryocytes. We were, however, able to obtain a preparation of highly enriched polyploid megakaryocytes from bone marrow cells cultured with PEG-rHuMGDF (Fig. 8A). As shown in Fig. 8B, the rate of degradation of recombinant cyclin B1 by an extract derived from large megakaryocytes is significantly higher than the rate of degradation by diploid bone marrow cells. We realize that these results, although indicative, do not demonstrate conclusively that polyploidizing primary megakaryocytes display enhanced activity of the ubiquitin-proteosome pathway as compared with diploid megakaryocytes.
Megakaryocyte-induced Degradation of Cyclin B Depends on the Ubiquitin-conjugating Enzyme Ubc4 -It has been reported that cyclin B degradation in other systems is mediated via the ubiquitin-conjugating enzyme Ubc4 (4). We thus used a mutated form of Ubc4 in which cysteine 85 was replaced by serine.
Such a protein was shown to act in vitro as a dominant negative mutant (28). Our data indicated that mutated Ubc4, but not mutated Ubc2 protein, slowed the rate of degradation of recombinant cyclin B by purified megakaryocytes (Fig. 9), indicating that Ubc4 mediates cyclin B degradation in this cell type. DISCUSSION In some systems investigated, the lack of cyclin B1 alone is sufficient to drive endoreduplication. Endoreduplication in some Drosophila cell types is indeed associated with the lack of cyclin B1 (33). Also, the metaphase II arrest in mouse oocytes is controlled through destruction of cyclin B1 (34). Certain treatments, such as with inhibitors of protein kinases in mammalian cells or high levels of the protein encoded by rum1, which inhibits the mitotic kinase in fission yeast, block M phase and induce repeats of S phase (35)(36)(37). We found previously that polyploidization in megakaryocytic cell lines is associated with reduced levels of cyclin B1 protein, but not of cyclin A or of cyclin B1 mRNA, as compared with the levels in diploid cells (9,10). In this context, it is important to note that although cyclin B is important for the G 2 /M transition, low levels of cyclin B may also be needed for the S phase of yeast cell cycles (16,38). In the current study, we found that cyclin B1 half-life in polyploidizing cells is decreased, as compared with that in proliferating cells. We  thus explored the possibility that a reduced level of cyclin B1 during endomitosis reflects an enhanced ability of these cells to degrade this cyclin, since this may result in abrogated mitosis (13). We found that polyploidization in the megakaryocytic cell lines MegT and Y10/L8057 or in primary megakaryocytes was associated with an enhanced ability of the cells to degrade recombinant cyclin B1 protein, as compared with the diploid proliferating megakaryocytic cell lines or nonmegakaryocytic bone marrow cells, respectively. Also, an extract of total bone marrow cells treated with the ploidypromoting factor, the c-Mpl ligand, exhibited an augmented ability to degrade cyclin B, as compared with nontreated bone marrow cells. This proteolytic process displayed all of the hallmarks of a ubiquitin-dependent degradation (39), such as formation of high molecular weight complexes of cyclin, dependence on ATP, and competition by the mutated form of the ubiquitin-conjugating enzyme Ubc4. Many components are involved in ubiquitination and degradation of cyclin B. Cyclin B ubiquitination is regulated via the ubiquitin-conjugating enzymes Ubc4 (4) and UbcX/Ubc10 (7,40), and on ubiquitin ligase (E3) activity, the latter associated with a complex of proteins termed the anaphase-promoting complex (APC) or cyclosome (2,41,42). Together with the ubiquitin-activating enzyme, these components promote polyubiquitination of cyclin B, which targets the protein to proteolysis by the 26 S proteosome. Besides Cdc16 and Cdc27, the identities of the components of APC in mammalian cells remain to be explored. The APC determines the specificity of degradation of different cyclins. King and colleagues (4) concluded that different cyclins are recognized by different components of the ubiquitination machinery. It was found that the cyclin A destruction box cannot functionally substitute for its B-type counterpart, indicating the existence of different components of the APC. Since cyclin A levels are elevated during the G 1 /S phase of the mitotic or endomitotic cell cycles in megakaryocytes (10) and since our current study demonstrates that polyploidizing megakaryocytes do not display an enhanced ability to degrade recombinant cyclin A, it is reasonable to assume that a cyclin B1-specific component of the APC may be up-regulated during polyploidization. Future identification of the APC components will allow a further examination of this process, and one may speculate that ploidy-promoting factors may affect the expression of such an APC-specific component.