INTRODUCTION

The PKC¹ family of enzymes is involved in the transduction of external signals from many growth factors, cytokines, and hormones (reviewed in Refs. 1-3). Many of the details of the receptor-mediated signaling pathways that lead to PKC activation have been elucidated (2, 3). A common feature of these pathways is receptor-mediated activation of lipid-metabolizing enzymes that generate DAG. DAG in turn activates PKC family members. The best characterized of these pathways involves the activation of phosphatidylinositol-specific phospholipase C (PI-PLC) activity (4). Two major classes of cellular receptor utilize the PI-PLC/PKC activation pathway. Growth factor receptors containing intrinsic tyrosine kinase activity activate PI-PLC  isoforms through direct tyrosine phosphorylation and activation of the enzyme (4). Many G-protein-coupled receptors activate PI-PLC isoforms through direct interaction with heterotrimeric G-proteins of the G_q class (4). Activation of PI-PLC enzymes leads to generation of inositol trisphosphate and DAG, two metabolites of phosphatidylinositol 4,5-bisphosphate that stimulate intracellular calcium release and activate PKC isozymes, respectively. Extracellular ligands can also stimulate activation of phosphatidylcholine-specific phospholipase C (PC-PLC) and/or phospholipase D (PLD) activities (5). These phospholipases can give rise to increased cellular DAG levels and PKC activation, including the novel, calcium-independent isozymes (3). More recently, it has been demonstrated that phosphatidylinositol 3,4,5-trisphosphate, a product of growth factor receptor-activated PI3 kinase, can directly and selectively activate the atypical PKC isozymes (6-10). Elucidation of these pathways has enhanced our understanding of how lipid metabolism and PKC activation are coupled to acute growth factor and hormone actions, including mitogenesis.

In addition to its well established role in acute mitogenic signaling, PKC has also been implicated in intrinsic signaling pathways, including those involved in cell cycle control (1). We recently demonstrated that activation of the _II PKC isoform at the nucleus is both cell cycle-regulated and necessary for entry of cells into mitosis (11-16). At the nucleus, _II PKC directly phosphorylates the nuclear envelope polypeptide lamin B at sites involved in mitotic nuclear lamina disassembly (12, 16). Inhibition of nuclear PKC activity leads to cell cycle arrest in the G₂ phase, indicating the importance of nuclear PKC in mitotic events (16). Although it is clear that nuclear PKC activity is cell cycle-regulated, the basis for this regulation is not clear. In the present study, we report that nuclear DAG levels fluctuate during cell cycle and that changes in nuclear DAG levels correlate with cell cycle progression through the G₂/M phase. Furthermore, a nuclear PI-PLC activity has been identified that is active during the G₂ phase. PI-PLC inhibitors lead to decreased nuclear PI-PLC activity and cell cycle blockade in the G₂ phase, suggesting a role for PI-PLC activity in nuclear events associated with entry into mitosis.

EXPERIMENTAL PROCEDURES

The human promyelocytic (HL60) leukemia cell line was grown and maintained in Iscove's medium supplemented with 10% calf serum as described previously (17). For cell cycle synchronization, cells were treated with aphidicolin (2 µg/ml) for 18 h to arrest cells in the G₁/S phase as described previously (16). Cells were released from the G₁ phase by washing four times with Iscove's medium and resuspending the cells in Iscove's medium containing 10% calf serum. At the indicated times, cells were either fixed and prepared for flow cytometric analysis of cell cycle progression as described previously (16) or harvested for nuclear DAG determination as described below. In some cases the phospholipase inhibitors 1-O-octadecyl-2-O-methyl-sn-glycero-3-phosphocholine (ET-18-OCH₃) or D609 or the cell cycle-blocking agents staurosporine or nocodazole were added to the cultures at the times and concentrations indicated in the figure and table legends. For determination of mitotic index, cells were stained with propidium iodide and viewed under phase fluorescence microscopy using a Nikon Axophot microscope equipped with appropriate filters as described previously (16).

Highly purified nuclei were isolated from cells in the indicated phases of cell cycle as described previously (17). Isolated HL60 nuclei are largely devoid of cytoplasmic and plasma membrane contamination as revealed by phase and electron microscopic examination and the absence of immunoreactive transferrin receptor, a prominent plasma membrane constituent of HL60 cells (17). Purified nuclei were assayed for DAG levels using the DAG kinase method of Priess et al. (18). Nuclear phospholipase activity was assayed in reaction buffer containing 20 mM Tris·HCl, pH 7.4, 100 mM NaCl, 10 µM CaCl₂, 1 mM dithiothreitol. Nuclei from 1 × 10⁷ cells were resuspended in reaction buffer and incubated at 37 °C for the indicated times and assayed for nuclear DAG levels as described above. The effect of the phospholipase C inhibitors ET-18-OCH₃, neomycin sulfate, and D609 on nuclear phospholipase activity was determined at the concentrations indicated in the figure legends. The possible involvement of phospholipase D activity in nuclear DAG generation was assessed by addition of either 1.5% ethanol or 100 µM propranolol to the reaction buffer.

Nuclear envelopes were isolated from G₁ phase cells as described previously (11). Purified nuclear envelopes were incubated with purified recombinant human _II PKC in the absence or presence of exogenous dioleoylglycerol (Avanti Polar Lipids) in PKC reaction buffer as described previously (11). _II PKC lamin kinase activity was assessed by monitoring _II PKC-mediated phosphorylation of the nuclear envelope component lamin B as described previously (11).

RESULTS AND DISCUSSION

We previously demonstrated that protein kinase C is activated at the nucleus of human leukemia cells in response to a number of proliferative stimuli (11-15). Similar observations have been made in many other cell systems, indicating that nuclear PKC activation is an important physiologic response involved in proliferative signaling to a wide variety of stimuli (reviewed in Ref. 1). In human leukemia cells, we have found that nuclear PKC corresponds to the _II PKC isoform and that regions within the catalytic domain of _II PKC mediate selective translocation of this isoform to the nucleus (19, 20). At the nucleus, _II PKC directly phosphorylates the nuclear envelope polypeptide lamin B at mitosis-specific sites involved in mitotic nuclear lamina disassembly (11, 13, 15). Nuclear translocation of _II PKC is cell cycle-regulated, occurring in the G₂ phase prior to mitosis (11, 16). Inhibition of nuclear _II PKC leads to cell cycle blockade in the G₂ phase, demonstrating the importance of nuclear PKC activity in entry into mitosis (16). However, the mechanism by which nuclear PKC activation is coupled to cell cycle progression remains unknown.

We therefore assessed whether the nuclear levels of the major physiologic PKC activator DAG are responsive to cell cycle phase. Specifically, we wished to determine whether changes in nuclear DAG levels could account for the observed cell cycle-regulated activation of PKC at the nucleus during the G₂ phase. We first compared nuclear DAG levels in cells in the G₁ and the G₂ phase. For this purpose, cells were synchronized in the G₁ phase with aphidicolin and released into medium to allow synchronous progression through S phase, G₂ phase, mitosis, and the subsequent G₁ phase (16). Nuclei were isolated from cells in the G₁ phase or the G₂ phase (8 h after release from aphidicolin) and assayed for DAG levels (Fig. 1A). G₁ phase nuclei contain 10.0 ± 1.4 pmol of DAG/10⁶ nuclei, whereas cells in the G₂ phase contain 26.2 ± 2.2 pmol of DAG/10⁶ nuclei. The nuclear DAG level in G₁ phase cells is comparable with that in unsynchronized cells, indicating that the synchronization procedure itself does not affect nuclear DAG levels (data not shown). These results demonstrate that nuclear DAG levels fluctuate during cell cycle and are ~2.5-3-fold higher in G₂ phase cells than in G₁ phase cells.

In contrast to the 2.5-3-fold change in nuclear DAG levels, total cellular DAG levels are only slightly elevated during the G₂ phase when compared with the levels in G₁ phase cells (Fig. 1B). G₁ phase cells contain 170 ± 9.9 pmol of DAG/10⁶ cells, whereas G₂ phase cells contain 193 ± 4.1 pmol of DAG/10⁶ cells, an increase of about 14%. Given the relative levels of total cellular and nuclear DAG levels, the 2.5-3-fold increase in nuclear DAG is sufficient to account for most of the observed increase in total cellular DAG in G₂ phase cells. These data demonstrate that the observed cell cycle changes in DAG levels occur predominantly at the nucleus and cannot be accounted for by contamination of the isolated nuclei with DAG from other cellular sources. These results are consistent with those obtained by others in regenerating rat liver (21) and mitogen-stimulated Swiss 3T3 cells (22) where specific increases in nuclear DAG levels of 2-3-fold are observed in association with proliferation in the absence of similar changes in total cellular DAG levels.

We next determined whether nuclear DAG levels fluctuate during cell cycle progression (Fig. 2A). For this purpose, cells were synchronized in the G₁ phase and allowed to progress synchronously through cell cycle. Nuclei were isolated from cells at different times after release from the G₁ phase and assessed for cell cycle distribution and nuclear DAG levels. As cells progress through cell cycle, nuclear DAG levels rise to a peak at 8-9 h after release from the G₁ phase. Flow cytometric and mitotic index analyses indicate that this peak of DAG coincides with the G₂/M phase transition. The time course of nuclear DAG generation corresponds well with that of nuclear _II PKC activation, which also occurs at the time of the G₂/M phase transition (16). Nuclear DAG levels subsequently drop to basal levels as the cells reenter the G₁ phase. These results are consistent with the findings in regenerating rat liver hepatocytes and Swiss 3T3 cells, where a specific increase in nuclear DAG levels are seen without significant changes in total cellular DAG (21, 22).

The observation that nuclear DAG levels are elevated during the G₂ phase of cell cycle, coupled with our previous results demonstrating that _II PKC is translocated and activated at the nucleus during the G₂ phase (11), led us to determine whether the observed cell cycle-dependent changes in nuclear DAG level are capable of stimulating nuclear PKC activity. We previously demonstrated that nuclei from G₂ phase cells contain detectable levels of _II PKC, whereas G₁ phase nuclei do not (16). Therefore, analysis of the ability of G₁ and G₂ phase nuclei to support PKC activity is complicated by the difference in endogenous PKC associated with these nuclei. To overcome this problem, we isolated nuclear envelopes from G₁ phase cells, which contain little or no endogenous PKC activity, and incubated them with equivalent amounts of recombinant _II PKC in the absence or presence of additional DAG (15 pmol/10⁶ nuclei) to the level found in G₂ phase nuclei (Fig. 2B). PKC activity was directly assessed by monitoring PKC-mediated phosphorylation of its major nuclear envelope substrate, lamin B, as described previously (11). This approach allowed us to directly assess the ability of increased DAG levels, in the amount observed in G₂ phase nuclei, to activate PKC without the complication of differences in endogenous nuclear PKC. As can be seen, G₁ phase nuclei (containing 10 pmol of DAG/10⁶ nuclei) are able to support modest _II PKC-mediated phosphorylation of lamin B (Fig. 2B, lane 1). However, addition of DAG to the level observed in the G₂ phase (25 pmol/10⁶ nuclei) leads to significant activation of _II PKC lamin kinase activity (Fig. 2B, lane 2). We conclude that the 2.5-fold increase in nuclear DAG levels observed in G₂ phase cells is sufficient to support significant activation of nuclear _II PKC. These data are consistent with our previous finding that _II PKC is translocated and activated at the nucleus during the G₂ phase (11).

Given the elevated levels of nuclear DAG in G₂ phase cells, we next determined whether G₂ phase nuclei contain an active phospholipase activity capable of generating DAG. Isolated nuclei from G₂ phase cells were incubated at 37 °C for up to 2.5 h, and nuclear DAG levels were determined throughout the incubation period (Fig. 3A). As can be seen, nuclear DAG levels increase in a linear fashion over a 2-h time course. DAG generation is enzymatic since nuclear DAG levels do not increase over the 2.5-h time period when nuclei are incubated at 4 °C rather than 37 °C (data not shown). Furthermore, aphidicolin (2 µM), the agent used in the synchronization of these cells, has no effect on nuclear DAG generation (data not shown). These results indicate that nuclear DAG comes from an intrinsic nuclear phospholipase activity and argue against the translocation of DAG, produced in other cellular membranes, to the nucleus.

We next determined the nature of the nuclear phospholipase responsible for generating nuclear DAG. For this purpose, we assessed the effect of various phospholipase inhibitors on nuclear DAG generation (Fig. 3B). Incubation of nuclei for 2 h at 37 °C leads to accumulation of DAG to levels 3-5-fold that of starting G₂ phase nuclei (Fig. 3B, columns 1 and 2). Inclusion of the PI-PLC-selective inhibitor ET-18-OCH₃ leads to dose-dependent inhibition of nuclear DAG generation (40% inhibition at 10 µM (Fig. 3B, column 3) and 85% inhibition at 100 µM (Fig. 3B, column 4)). Likewise, the nonselective phospholipase inhibitor neomycin sulfate leads to inhibition of nuclear phospholipase activity (Fig. 3B, columns 7 and 8). The data are in good agreement with the published IC₅₀ values of these compounds (IC₅₀ ET-18-OCH₃ = 10 µM (23), neomycin sulfate = 65 µM (24)). In contrast, addition of the PC-PLC inhibitor D609 at 10 µM and 30 µM (Fig. 3B, columns 5 and 6) had no effect on nuclear DAG generation.

To assess the possible contribution of PLD activity to nuclear DAG generation, we determined the effect of ethanol on nuclear DAG production. In the presence of 1.5% ethanol, PLD generates phosphatidylethanol rather than phosphatidic acid (25). Since phosphatidylethanol is not a substrate for phosphatidic acid phosphohydrolase, the enzyme responsible for conversion of phosphatidic acid to DAG, 1.5% ethanol effectively blocks PLD-mediated generation of DAG (25). Inclusion of 1.5% ethanol in the nuclear phospholipase assay had little effect on nuclear DAG generation (Fig. 3B, column 9). Likewise, 100 µM propranolol, a phosphatidic acid phosphohydrolase inhibitor (26), had no effect on nuclear DAG generation (Fig. 3B, column 10). Taken together, these data indicate that the major phospholipase activity responsible for generating nuclear DAG in G₂ phase nuclei is a PI-PLC activity. Although it is possible that other phospholipase activities may contribute to nuclear DAG levels, the inhibitor studies clearly demonstrate that the major nuclear phospholipase activity present in G₂ phase nuclei is PI-PLC.

To assess the potential physiologic role of nuclear PI-PLC, we determined the effect of PI-PLC inhibition on cell cycle progression through the G₂/M phase transition. For this purpose, cells were synchronized, released from the G₁ phase, and assessed for cell cycle progression by flow cytometric analysis as described previously (16). Using this protocol, HL60 cells progress synchronously through the G₂/M phase and return to the G₁ phase by 12 h after release (Fig. 4, top panel). However, when 10 µM ET-18-OCH₃ is included in the culture medium, cells progress through the S phase but exhibit a blockade in the G₂/M phase (Fig. 4, middle panel). This blockade is not seen with the PC-PLC inhibitor D609, indicating that it is selective for PI-PLC inhibition (Fig. 4, bottom panel). To assess the level of cell cycle inhibition, we compared the cell cycle distribution of ET-18-OCH₃-treated cells with those treated with other cell cycle inhibitors. The spindle poison nocodazole was used to arrest cells in M phase, and the nonselective protein kinase inhibitor staurosporine was used to arrest cells in the G₂ phase (16, 27) as positive controls for complete arrest in these cell cycle phases (Table I). From this analysis, we estimate that 10 µM ET-18-OCH₃ leads to 46% inhibition in the G₂ phase, consistent with the published IC₅₀ of 10 µM for inhibition of PI-PLC (21). Higher concentrations of ET-18-OCH₃ could not be used in these cells as they caused progressive induction of apoptosis, an effect of ET-18-OCH₃ that has been well documented (28-31).

Table I. Cell cycle distribution of HL60 cells after treatment with various cell cycle inhibitors HL60 cells were synchronized in the G1 phase as described under "Experimental Procedures." Synchronized cells were washed and resuspended in control medium (no addition) or medium containing nocodazole (40 ng/ml), staurosporine (150 nM), ET-18-OCH3 (10 µM), or D609 (30 µM). The cells were incubated at 37 °C for 12 h, at which time the cells were fixed and assessed for cell cycle distribution by flow cytometric and mitotic index analysis as described under "Experimental Procedures." Treatment G1 S G2 M % % % % No addition 65 22 2 11 Nocodazole 20 30 2 48 Staurosporine 14 26 48 12 ET-18-OCH3 40 28 23 9 D609 67 22 2 9

Concomitant with the G₂ phase cell cycle arrest, treatment with 10 µM ET-18-OCH₃ led to a 35% reduction of nuclear DAG levels when compared with untreated G₂ phase cells, in line with the level of cell cycle arrest induced by ET-18-OCH₃. In addition, ET-18-OCH₃ treatment leads to an 83% reduction in nuclear PI-PLC activity in vitro (Fig. 5). A similar G₂ phase cell cycle arrest has been reported in colony-stimulating factor 1-dependent cells after treatment with ET-18-OCH₃ (32). Our data demonstrate that ET-18-OCH₃ leads to inhibition of nuclear PI-PLC activity concomitant with cell cycle arrest in the G₂ phase, providing a plausible mechanism for the cell cycle effects of ET-18-OCH₃.

In conclusion, our data provide direct evidence that nuclear DAG levels are cell cycle-regulated and that they rise to a peak corresponding to the G₂/M phase transition. The major phospholipase responsible for the generation of nuclear DAG during the G₂ phase of cell cycle is a PI-PLC activity. Nuclear PI-PLC activity appears to be important for the G₂/M phase transition of cell cycle, since when this activity is inhibited, cell cycle arrest in the G₂ phase is observed. Recent studies have provided evidence implicating nuclear PI-PLC activities in the G₁ to S phase transition and in DNA replication (33-36). Our present data indicate an additional, novel role for nuclear PI-PLC during the G₂/M phase. We hypothesis that the DAG generated by nuclear PI-PLC activity is responsible for activating nuclear PKC during the late G₂ phase, which is required for entry of cells into mitosis (16). Here, we provide direct evidence that the elevated nuclear DAG levels observed during the G₂ phase of cell cycle are sufficient to stimulate nuclear _II PKC-mediated phosphorylation of its physiologic nuclear substrate lamin B. Future studies will focus on identifying which PI-PLC isozyme(s) correspond to nuclear PI-PLC in G₂ phase cells, determining the molecular mechanisms by which this nuclear PI-PLC activity is regulated during cell cycle and further assessing its role in the G₂/M phase transition.