Phosphatidylglycerol is a physiologic activator of nuclear protein kinase C.

A major mechanism by which protein kinase C (PKC) function is regulated is through the selective targeting and activation of individual PKC isotypes at distinct subcellular locations. PKC betaII is selectively activated at the nucleus during G2 phase of cell cycle where it is required for entry into mitosis. Selective nuclear activation of PKC betaII is conferred by molecular determinants within the carboxyl-terminal catalytic domain of the kinase (Walker, S. D., Murray, N. R., Burns, D. J., and Fields, A. P. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 9156-9160). We previously described a lipid-like PKC activator in nuclear membranes, termed nuclear membrane activation factor (NMAF), that potently stimulates PKC betaII activity through interactions involving this domain (Murray, N. R., Burns, D. J., and Fields, A. P. (1994) J. Biol. Chem. 269, 21385-21390). We have now identified NMAF as phosphatidylglycerol (PG), based on several lines of evidence. First, NMAF cofractionates with PG as a single peak of activity through multiple chromatographic separations and exhibits phospholipase sensitivity identical to that of PG. Second, purified PG, but not other phospholipids, exhibits dose-dependent NMAF activity. Third, defined molecular species of PG exhibit different abilities to stimulate PKC betaII activity. 1,2-Dioleoyl-PG possesses significantly higher activity than other PG species, suggesting that both fatty acid side chain composition and the glycerol head group are important determinants for activity. Fourth, in vitro binding studies demonstrate that PG binds to the carboxyl-terminal region of PKC betaII, the same region we previously implicated in NMAF-mediated activation of PKC betaII. Taken together, our results indicate that specific molecular species of nuclear PG function to physiologically regulate PKC betaII activity at the nucleus.

sical, or calcium-dependent PKC isotypes require calcium, diacylglycerol (DAG), and phosphatidylserine (PS) for activation. The novel PKCs require DAG and PS, but not calcium, whereas the atypical PKCs do not require DAG or calcium, but appear to require PS for activation. PKC isotypes exhibit tissue-and cell type-specific patterns of expression, suggesting specialization of function. Indeed, accumulating evidence indicates that PKC isotypes serve distinct, nonoverlapping functions in cellular physiology (reviewed in Refs. 1, 3, and 5).
An important mechanism by which PKC function is regulated is through the targeting of PKC isozymes to distinct subcellular locations (1,5). In human leukemia cells, which express PKC ␣, ␤ II , and , we have found that PKC ␤ II is selectively activated at the nucleus during the G 2 phase of cell cycle (6,7). At the nucleus, PKC ␤ II directly phosphorylates the nuclear envelope polypeptide lamin B at sites involved in mitotic nuclear lamina disassembly (8 -10). Inhibition of nuclear PKC ␤ II activity leads to cell cycle arrest in G 2 phase, demonstrating the importance of nuclear PKC ␤ II activation in the entry of cells into mitosis (7). In contrast, PKC ␣ and PKC are not observed at the nucleus, and we have demonstrated that they play key roles in leukemia cell differentiation and survival/apoptosis, respectively (11,12).
Given its involvement in cell cycle progression, we investigated the mechanisms underlying the selective nuclear translocation and activation of PKC ␤ II . Using chimeric PKC molecules, produced by exchanging the regulatory and catalytic domains of PKC ␣ and ␤ II , we determined that the catalytic domain of PKC ␤ II contains molecular determinants that are important for selective nuclear targeting of the enzyme (13). In related biochemical studies, we examined the mechanism by which PKC ␤ II is activated at the nucleus (14). We found that component(s) within the nuclear membrane selectively stimulate PKC ␤ II activity 3-6-fold above the level achieved in the presence of optimal concentrations of calcium, DAG, and PS (14). This nuclear membrane activation factor (NMAF) was shown to be soluble in nonionic detergents and organic solvents, and to be insensitive to protease treatment, suggesting that it is a lipid (14). In the present study, we identify NMAF as phosphatidylglycerol (PG) based on the fractionation profile of NMAF and the ability of purified PG to activate PKC ␤ II . Interestingly, individual purified PG species vary in their ability to activate PKC ␤ II , suggesting that the nuclear membrane contains specific PG species that serve to potently stimulate PKC ␤ II activity. Finally, we demonstrate that PG binds to the carboxyl-terminal region of PKC ␤ II , consistent with the role of this region in nuclear activation of PKC ␤ II (14). Our data indicate that nuclear membrane PG is an important physiologic regulator of PKC activity at the nucleus.

EXPERIMENTAL PROCEDURES
Cell Culture and Isolation of Nuclear Membrane Extracts-Human promyelocytic (HL60) leukemia cells were maintained in suspension culture in Iscove's medium supplemented with 10% calf serum as described previously (7,14). Nuclear envelopes were prepared as described previously (7) and subjected to two-phase extraction in chloroform/methanol/water (8:4:3) to separate lipids from nonlipid constituents (15). The organic (lower) phase was isolated and dried under N 2 gas for subsequent analysis.
Thin Layer and Silica Column Chromatography of Nuclear Membrane Extracts-Preparative TLC was used to fractionate nuclear membrane extracts into major lipid classes based on the migration of purified lipid standards (Avanti Polar Lipids). Nuclear membrane extracts, isolated as described above, were resuspended in chloroform, spotted on Diamond K6F Silica Gel 60 TLC plates (Whatman) and developed in chloroform/methanol/water (65: 25:4). Developed plates were air-dried, and the migration of phospholipid standards was determined by staining with 0.1% 8-anilino-1-naphthalene sulfonic acid ammonium salt (Sigma) in water, pH 7.0, and visualized under short wave UV light (16). The TLC plates were partitioned based on the migration of purified standards, and the silica from each section was scraped from the plate and recovered. Phospholipids from each section were eluted into chloroform/methanol (3:1), and the resulting extracts were dried under N 2 gas and stored at Ϫ20°C until analyzed for activity as described below.
Nuclear membrane extracts were further fractionated on Bakerbond SPE 228 silica extraction columns (VWR). Samples were loaded onto the columns in chloroform and fractionated by stepwise elution into solvents containing increasing proportions of methanol in chloroform. Column fractions were dried under N 2 gas and stored at Ϫ20°C prior to analysis. The elution profile of standard lipids was assessed by TLC analysis of column fractions using the chromatographic system described above.
In Vitro PKC Activity Assay-The ability of nuclear membrane extracts, column fractions, and purified phospholipids to stimulate PKC ␤ II activity was assessed in a standard in vitro PKC assay (14) using purified recombinant baculovirus-expressed human PKC ␤ II as the source of enzyme (8). Briefly, PKC activity was measured by following incorporation of [ 32 P]phosphate from [␥-32 P]ATP into histone H1 (Sigma). Reactions were performed under optimal activating conditions (100 M CaCl 2 , 20 M dioleoylglycerol, 40 g/ml PS, and 10 g of histone H1) in the absence and presence of nuclear membrane extracts, column fractions, or purified phospholipids. All assays were carried out for 15 min at room temperature, conditions under which [ 32 P]phosphate incorporation is linear (14). Unless otherwise stated, nuclear membrane extract or column fractions from 10 7 cell equivalents were assayed in each reaction. The kinase reactions were terminated by addition of SDS sample buffer and the samples boiled for 5 min prior to SDS-polyacrylamide gel electrophoresis analysis. Histone phosphorylation was quantitated by phosphorimaging analysis of the gels as described previously (14).
Quantitative Measurement of PG Mass-The amount of PG in nuclear membrane extracts and column fractions was assessed using the specific, enzymatic PG assay described by Jones and Ashwood (20). Briefly, membrane extracts or column fractions were incubated with phospholipase D, glycerokinase, and glycerol-3-phosphate oxidase to generate hydrogen peroxide from PG. Hydrogen peroxide formed in this step was quantitated after addition of 4-aminoantipyrine, 3,5-dichloro-2-hydroxybenzenesulfonic acid, and peroxidase, by following generation of red chromogen at its absorption maximum of 510 nm. A blank was generated for each unknown using the same reaction conditions in the absence of phospholipase D. Absorbance is directly proportional to the quantity of PG in the sample. The total mass of PG in each sample was determined using a standard curve generated with purified PG (Avanti Polar Lipids). Specificity of the assay was confirmed using other purified phospholipid classes, none of which gave absorbance above background at 100 g/ml phospholipid.
Lipid Vesicle Binding Assay-A fusion protein between glutathione S-transferase (GST) and the carboxyl terminus of PKC ␤ II (amino acids 576 -673; GST-PKC ␤ II CT) was generated by polymerase chain reaction of the carboxyl-terminal fragment of the human PKC ␤ II cDNA using the following primers: forward 5Ј-CGGGATCCCACTGATGAC-CAAACACC-3Ј and reverse 5Ј-CCCTCGAGGATTAGCTCTTGACT-TCG-3Ј. These primers allowed introduction of a 5Ј-BamHI and a 3Ј-XhoI restriction site to facilitate directional cloning of the polymerase chain reaction product into the pGEX-5X-3 expression vector (Amersham Pharmacia Biotech). GST fusion protein was expressed in Escherichia coli and purified on glutathione affinity columns according to the manufacturer's protocol (GST gene fusion system, Amersham Pharmacia Biotech). Purified GST or GST-PKC ␤ II CT was mixed with sonicated lipid vesicles (250 g of lipid) or buffer alone (50 mM Tris, pH 7.4, 100 mM NaCl, 10 mM MgSO 4 , and 0.5 mM CaCl 2 ) for 15 s and then filtered through 100-kDa molecular mass cutoff filters (Amicon) according to Rebecchi et al. (21). Lipid vesicles contained dioleoyl-PG, PC, or both as indicated in the figure legend. In some instances, PS and DAG (20 and 8 mol %, respectively) were included in place of PC as indicated in the text. Lipid vesicles were quantitatively retained by the filter as determined by TLC analysis. Filtrate (unbound) and bound fractions were resolved by SDS-polyacrylamide gel electrophoresis and the presence of GST fusion protein determined by immunoblot analysis with anti-PKC ␤ II (Santa Cruz) and anti-GST antibody (PharMingen) and chemiluminescence. Quantification was by densitometric analysis of the developed films.

Initial Characterization of Nuclear Membrane Activation
Factor-Based on our initial observation that NMAF activity is solubilized from nuclear envelopes by either nonionic detergent or organic solvent extraction, we hypothesized that NMAF may be a lipid (14). Further characterization demonstrated that NMAF activity fractionated completely into the organic phase of a two-phase (Folch) extraction and that NMAF activity is largely resistant to heat inactivation by boiling for 5 min, supporting the hypothesis that NMAF is lipid and not protein.
In addition, NMAF exhibits the same stimulatory effect on PKC ␤ II activity in both our standard vesicle assay and a mixed micelle assay (22), indicating that NMAF activity is not the result of an artifact inherent to the assay system (data not shown). These data, along with our previous observation that NMAF activity is resistant to exhaustive protease treatment (14), provide convincing evidence that NMAF is lipid and not protein.
Thin Layer Chromatographic Fractionation of NMAF-In order to characterize the lipid component(s) of nuclear membranes responsible for NMAF activity, nuclear membrane extracts were resolved into major lipid classes by TLC. TLC plates were spotted with nuclear membrane extract in chloroform and developed using a solvent system that allows resolution of the major phospholipid classes. The developed TLC plates were then divided into regions based on the migration of lipid standards, scraped, and eluted as described under "Experimental Procedures." Individual TLC fractions were then assayed for NMAF as measured by the ability to stimulate PKC ␤ II activity as described previously (14). NMAF was quantitatively recovered as a single peak of activity in two consecutive fractions (F5 and F6) from the TLC plate (Fig. 1). Comparison of the migration of NMAF activity with purified phospholipid standards reveals that NMAF comigrates with several major phospholipid classes. Specifically, PG, PC, and phosphatidylethanolamine (PE) all migrated within the region of the TLC containing NMAF activity. In contrast, NMAF activity was completely resolved from other major phospholipid classes, including phosphatidic acid, PS, phosphatidylinositol (PI), and lysophospholipids, and from the neutral lipid metabolite DAG. These data suggest that NMAF may correspond to PG, PC, or PE.
Ability of Individual Phospholipids to Activate PKC ␤ II -Our previous studies indicated that NMAF is not PS or DAG, since NMAF activates PKC ␤ II in the presence of maximal concentrations of these cofactors (14). Substantiating this finding, PS and DAG are clearly resolved from NMAF activity after TLC separation (Fig. 1). The migration of NMAF on TLC plates suggested that NMAF might correspond to one of the three phospholipids PG, PC, or PE. Therefore, we directly determined the ability of these phospholipids to activate PKC ␤ II (Fig. 2). As can be seen, PG was able to stimulate PKC ␤ II activity in a dose-dependent fashion above the level induced by the conventional PKC activators, DAG, PS, and calcium. In contrast, PC and PE showed little or no stimulatory activity at concentrations up to 250 g/ml, demonstrating that the NMAFlike activity exhibited by phospholipid addition is specific for PG.
If PG were NMAF, one would predict that NMAF and purified PG would exhibit the same sensitivity to phospholipases. Therefore, we assessed the effect of PI-PLC, PC-PLC, and PLA 2 treatment on the ability of NMAF and purified PG to stimulate PKC ␤ II activity (Fig. 3). Treatment of NMAF or purified PG with either PI-PLC or PC-PLC had little effect on their PKC stimulatory activity. In contrast, treatment with PLA 2 led to substantial inhibition of both NMAF and PG activity. These data demonstrate that NMAF and PG exhibit a similar pattern of sensitivity to phospholipases, consistent with the suggestion that NMAF is PG. These data also provide direct confirmation that NMAF is distinct from PC and PI and that it is not a lysophospholipid.
PG Is Present in Nuclear Membrane Extracts and Cofractionates with NMAF Activity-Given the ability of PG to stimulate PKC activity, we next assessed whether PG is present in nuclear membrane extracts and whether endogenous PG cofractionates with NMAF activity. For this purpose, we subjected nuclear membrane extracts to fractionation by silica column chromatography (Fig. 4A). Samples were loaded onto silica columns in chloroform, and lipid constituents were eluted into solvents of increasing polarity. Individual fractions were collected and assayed for the ability to stimulate PKC activity (Fig. 4A). NMAF activity was retained on the column and eluted as a single peak in 1:1 chloroform/methanol (CHCl 3 /CH 3 OH). Purified PG standard also eluted as a single peak in 1:1 CHCl 3 /CH 3 OH, further suggesting that PG corresponds to NMAF (data not shown). Therefore, we directly assessed the levels of PG in the silica column fractions from nuclear membrane extracts using a specific enzymatic assay for PG mass (20). Nuclear membrane PG elutes specifically in 1:1 CHCl 3 /CH 3 OH along with NMAF activity (Fig. 4B). From the quantitative PG assay, we calculate that nuclear membranes contain ϳ0.39 g of PG/ 10 7 cells. This level of PG corresponds to a concentration of ϳ10 g/ml in our standard assay. These data confirm the presence of PG in the nuclear membrane and demonstrate that nuclear PG comigrates with NMAF activity.
Individual Species of PG Differ in Ability to Activate PKC ␤ II -Having identified NMAF as PG, we wished to determine whether the fatty acid side chain composition of individual PG Nuclear membrane extracts were prepared from HL60 cell nuclear envelopes as described previously (14). Extracts in chloroform were spotted onto Silica Gel 60 plates and resolved by one-dimensional TLC in chloroform/methanol/water (65:25:4, v/v/v). The migration of purified phospholipid standards was determined as described under "Experimental Procedures," and the positions of standards are depicted in the schematic diagram. The TLC plate was divided into eight fractions as indicated, based on the migration of the lipid standards and the silica scraped from each fraction and recovered. Lipids in each fraction were eluted with chloroform/methanol (3:1), dried under N 2 gas, resuspended in aqueous buffer, and assayed for activation of PKC ␤ II in the standard histone kinase assay under conditions supporting maximal PKC activity (100 M Ca 2ϩ , 20 M diacylglycerol, 40 g/ml phosphatidylserine).
Results are plotted as -fold activation relative to control (no lipid extract added) for each fraction and compared with the -fold activation obtained with unfractionated NMAF (top bar). Results represent the mean of three independent determinations Ϯ SD. LPC, lysophosphatidylcholine; PA, phosphatidic acid; PS, phosphatidylserine; PI, phosphatidylinositol. species is important for PKC stimulatory activity. Therefore, we assessed the ability of individual purified molecular species of PG to activate PKC ␤ II (Fig. 5). The PG preparation used in the studies shown in Figs. 2 and 3 is a mixture of PG species with a complex distribution of fatty acid side chain constituents (Avanti Polar Lipids catalog). The two predominant side chain constituents of this PG mixture are C16:0 palmitic acid (34%) and C18:1 oleic acid (31%). Therefore, we compared the ability of defined PG species containing these two fatty acid side chain constituents to stimulate PKC ␤ II activity (Fig. 5A). 1,2-Dipalmitoyl-PG and the mixed fatty acid species 1-palmitoyl-2oleoyl PG exhibited activities comparable with or weaker than the original mixture. In contrast, 1,2-dioleoyl-PG was more potent, exhibiting about 1 log higher activity than the original mixture.
In order to determine whether the difference in activity between 1,2-dioleoyl-PG and other PG species was merely an effect of fatty acid side chain length, the activity of other C18 fatty acid-containing PG species was compared with that of 1,2-dioleoyl-PG (Fig. 5B). Neither distearoyl (C18:0) PG nor dielaidoyl (C18:1 ⌬9 trans)-PG exhibited enhanced activity compared with the original mixture and in fact were less potent than the mixture in stimulating PKC ␤ II activity. Interestingly, 1,2-dioleoyl (C18:1⌬9 cis) PG had high activity relative to the mixture, whereas its stereoisomer 1,2-dielaidoyl (C18:1 ⌬9 trans)-PG was less active. These data demonstrate that both the glycerol head group and the fatty acid side chain composition are important determinants for NMAF activity. Taken together, these data suggest that particular molecular species of PG present in the nucleus may be responsible for the potent stimulation of nuclear PKC ␤ II activity observed in the presence of nuclear membranes. We next wished to directly compare the ability of nuclear membrane-derived PG and purified PG to stimulate PKC ␤ II activity (Fig. 6). Nuclear PG was isolated by silica column chromatography as described above and assayed for activity along with purified 1,2-dioleoyl-PG and a mixture of PG species. Nuclear membrane-derived PG and purified 1,2-dioleoyl-PG exhibit comparable activity that is more potent than that exhibited by the PG mixture. From these data, an apparent FIG. 4. NMAF cofractionates with phosphatidylglycerol on silica column chromatography. A, assay for NMAF activity in silica column fractions. Nuclear membrane extracts were prepared as described under "Experimental Procedures." Dried extracts were solubilized in chloroform and loaded onto a silica column. Components of the extracts were eluted with increasing amounts of methanol in chloroform. The fractions were dried under N 2 gas, resuspended in aqueous buffer, and assayed for the ability to activate PKC ␤ II activity. Results represent the mean of three independent determinations Ϯ S.D. Fractions contain chloroform:methanol in the indicated ratios. ON, onput; FT, flow through; 100% MeOH, 100% methanol. B, quantitation of PG in silica column fractions. Nuclear membrane extracts were fractionated by silica column chromatography as described in A. Fractions are labeled as in A. The amount of PG in each fraction was measured using an enzymatic assay specific for PG (Ref. 20; see "Experimental Procedures"). Results are the mean of duplicate determinations.

FIG. 5. Different molecular species of PG differ in ability to
activate PKC ␤ II . The ability of various species of PG to activate PKC ␤ II was assessed. Kinase assays were conducted under the conditions described in the legend to Fig. 1 in the absence or presence of the indicated amount (2.5-250 g/ml) of purified PG species. Results are expressed as -fold activation of kinase activity in the presence of purified PG species. Results represent the mean of three independent determinations Ϯ S.E. A, the major PG species in the original PG mixture differ in activity. The activity of PG species containing the major fatty acid side chain constituents represented in the original mixture, oleic acid and palmitic acid, were compared for NMAF activity. E, PG mixture; q, 1,2-dioleoyl-PG; ϫ, 1,2-dipalmitoyl-PG; and f, 1-palmitoyl-2oleoyl-PG. B, activity is stereospecific for dioleoyl (C18:1 ⌬9 cis) PG. C18 PG species were compared for NMAF activity. E, original PG mixture; q 1,2-dioleoyl (C18:1 ⌬9 cis)-PG; f, 1,2-dielaidoyl (c18:1 ⌬9 trans)-PG; and ϫ, 1,2-distearoyl (C18:0)-PG.
EC 50 was estimated for each PG source. The apparent EC 50 for the purified PG mixture was ϳ90 g/ml. In contrast, both 1,2-dioleoyl-PG and nuclear membrane PG were more potent than the PG mixture, with apparent EC 50 values of ϳ10 g/ml. Therefore PG, possibly 1,2-dioleoyl-PG, exhibits sufficient activity to account for the activity we previously identified as NMAF.
The Carboxyl-terminal Region of PKC ␤ II Selectively Binds PG-Our previous results demonstrated that NMAF-mediated activation of PKC ␤ II is isozyme selective and suggested that the carboxyl-terminal region of PKC ␤ II is important for this activation (14). Therefore we devised a lipid vesicle binding assay to directly assess whether the carboxyl-terminal region of PKC ␤ II binds to PG. In this assay, PG vesicles were incubated with a GST fusion protein containing the carboxyl terminus of PKC ␤ II (GST-PKC ␤ II CT). Vesicle-bound protein was separated from unbound protein by centrifugation through a 100-kDa cutoff filter and the bound fraction analyzed for the presence of the GST fusion protein (Fig. 7). Under these conditions, GST-␤ II CT bound to PG-containing vesicles (Fig. 7A). Binding was selective for GST-PKC ␤ II CT, since little or no binding of GST to PG-containing vesicles was detected (Fig.  7B). Likewise, binding of GST-PKC ␤ II CT was selective for PG, since PC vesicles did not bind GST-PKC ␤ II CT (Fig. 7A, lane 3). Binding to PG vesicles was dependent on the PG content of mixed vesicles containing different proportions of PG and PC, indicating that binding is both concentration-dependent and saturable (Fig. 7C). The presence of PS and DAG did not influence binding to PG (data not shown), consistent with the fact that PS and DAG bind to the C2 region within the regulatory domain of PKC ␤ II , which is not present in the GST-PKC ␤ II CT fusion protein. In conclusion, our data demonstrate that PG binds selectively to the carboxyl-terminal region of PKC ␤ II . These data are consistent with our finding that this region of PKC ␤ II contains important determinants involved in the nuclear translocation and activation of PKC ␤ II (13,14). DISCUSSION Protein kinase C function is regulated by multiple mechanisms, including the tissue-and cell type-specific expression of individual PKC isotypes. In addition, PKC isotypes exhibit intrinsic differences in cofactor requirements and sensitivities to lipid second messengers. Appropriate intracellular targeting of PKC also appears to be critical for proper PKC isotype function in vivo. Intracellular targeting can be achieved through specific interactions of PKC with a growing family of docking proteins. For instance, members of the protein kinase A anchoring protein family appear to serve a role in PKCmediated signaling at the postsynaptic density (23). Protein kinase A anchoring proteins target PKC and other signaling molecules to the same intracellular compartment through simultaneous binding to distinct binding sites on the protein kinase A anchoring protein (23). Still other PKC-binding proteins serve as receptors for activated protein kinase C and play functional roles in PKC targeting and translocation events (24,25). It will be of interest to determine whether similar mechanisms aid in the targeting of PKC ␤ II to the nucleus of human leukemia cells. In this paper, we provide direct evidence that specific lipid components within the nuclear membrane play a key role in PKC signaling by stimulating PKC activity at the nucleus.
In previous studies, we demonstrated that PKC ␤ II is selectively translocated to the nucleus of human leukemia cells in response to proliferative stimuli (1,(7)(8)(9). Nuclear PKC ␤ II translocation and activation is cell cycle-regulated, occurring during the G 2 phase of cell cycle (6,7). At the nucleus, PKC ␤ II phosphorylates sites on the nuclear envelope polypeptide lamin B that are involved in the process of mitotic nuclear lamina disassembly (6,7,9). Inhibition of nuclear PKC activity leads to cell cycle arrest in G 2 phase prior to mitosis, demonstrating that nuclear PKC activity is required for cell cycle progression through the G 2 /M phase transition (7). Using PKC chimera, we demonstrated that nuclear translocation of PKC ␤ II is dependent upon the carboxyl-terminal catalytic domain of the enzyme (13). At the nucleus, PKC ␤ II activity is stimulated by a component within the nuclear membrane, termed NMAF, that serves to potently activate the enzyme (14). An active component of NMAF has now been identified.
Identification of NMAF as Phosphatidylglycerol-Based on several lines of evidence, we have shown that NMAF corresponds to PG. First, NMAF is a heat-stable, lipophilic activity that comigrates with PG on thin layer and silica column chromatographies. Second, purified PG, but not other phospholipids, exhibits NMAF-like activity. Third, NMAF and PG exhibit similar sensitivities to phospholipases. Fourth, PG is present in the nuclear membrane in sufficient quantities to stimulate PKC ␤ II activity. Fifth, specific PG species exhibit different activities indicating that the selectivity of NMAF activity lies not only in the glycerol head group, but also in the fatty acid side chain constituents. 1,2-Dioleoyl PG was found to be significantly more potent at stimulating PKC ␤ II activity than the other PG species tested. The selectivity for oleic acid is stereospecific since 1,2-dielaidoyl-PG, which is identical to 1,2dioleoyl PG except for the orientation of the ⌬9 double bond in the C18:1 fatty acid chain, exhibits activity that is about one log lower than 1,2-dioleoyl-PG. These results argue against the possibility that PG causes a nonspecific membrane effect, such as a change in membrane charge density, leading to PKC activation. Whereas it is possible that the nuclear membrane contains other lipid components that contribute to NMAF activity, we provide convincing evidence that PG has sufficient activity and is present in appropriate quantities to account for NMAF activity.
Previous studies indicated that PG-mediated activation of PKC ␤ II might involve the carboxyl-terminal region of the catalytic domain of PKC ␤ II (14). Our lipid vesicle binding studies provide direct evidence in support of this conclusion. Specifically, we demonstrate that the carboxyl terminus of PKC ␤ II binds selectively to PG-containing vesicles. These results are interesting in light of our previous demonstration that the carboxyl terminus of PKC ␤ II is also important for the nuclear translocation and activation of PKC ␤ II (13). Taken together, these results suggest that nuclear PG functions to modulate nuclear PKC ␤ II translocation and activation. In a recent study, we determined that the cell cycle-dependent activation of PKC ␤ II in the nucleus during G 2 phase is coupled to the generation of nuclear DAG through the action of a nuclear PI-PLC activity (26). We have determined nuclear PG levels during cell cycle progression and find that they do not change appreciably during cell cycle. 2 It is possible therefore that nuclear PG functions primarily to facilitate or enhance the selective binding of PKC ␤ II to the nuclear membrane where it can be fully activated in the presence of elevated DAG generated during G 2 phase. Further studies will be aimed at elucidating the relative contribution of PG, PS, DAG, and calcium in cell cycle-regulated activation of nuclear PKC ␤ II .
Effects of Acidic Phospholipids on PKC Membrane Binding and Kinase Activity in Vitro-A number of studies have demonstrated that the phospholipid environment of the target membrane can influence PKC activity in vitro. Utilizing defined vesicle and/or mixed micelle assays, the requirement for calcium, DAG, and phospholipids in PKC membrane binding and activation has been investigated. In the presence of calcium and DAG, PKC exhibits an identical sigmoidal depend-ence on PS for membrane binding and activation (27). DAG increases the affinity of PKC for PS, but not for other acidic phospholipids (28). Though PE and PG can reduce the amount of PS required for maximal binding and activity, they cannot replace the requirement for PS (28). Based on these studies, it has been suggested that PKC exhibits a dual requirement for acidic phospholipid, a specific requirement for PS binding and nonspecific electrostatic interactions with other acidic phospholipids (28). Our data also indicate that PG, in addition to PS, can influence PKC activity. However, our data suggest that, like the specific requirement for PS, the interactions involving PG are specific. Furthermore, our binding studies demonstrate that PG binds to the carboxyl-terminal region of PKC ␤ II , a site distinct from the binding site for PS in the regulatory domain. In addition, our data indicate that specific PG species can differentially influence PKC activity, indicating that specific lipid-PKC interactions underlie the stimulatory effects of PG.
Phosphatidylglycerol Is a Physiologic Activator of Nuclear PKC ␤ II -The effects of phospholipid composition and membrane fluidity on PKC activity have been assessed by selective enrichment of rat liver membranes with various phospholipids (29). Addition of PG, PS, PE, or dioleoyl-PC to membrane can lead to enhanced activation of PKC, with PG being the most effective activator (29). Our data are consistent with these findings and indicate that nuclear membrane PG is a physiologically relevant regulator of PKC activity. However, the exact mechanism by which PG stimulates PKC ␤ II activity remains to be fully elucidated. One potential mechanism stems from the observation that PKC interacts well with acidic phospholipids (28,30). Bazzi and Nelsestuen (30) demonstrated that calciumdependent binding of PKC to phospholipid vesicles induces clustering of acidic phospholipids including PG. They suggested that PKC may induce certain acidic phospholipids to form microdomains within physiologic membranes. Therefore, the presence of specific phospholipids, such as certain PG species, may influence PKC activity by forming clustered subdomains that serve to enhance the membrane binding and catalytic activity of PKC. Likewise this clustering may influence substrate selection through enhanced interactions of PKC and substrate at the membrane surface. Future studies will focus on determining the mechanisms by which nuclear PG stimulates PKC ␤ II activity, the fatty acid side chain composition and PKC stimulatory activity of individual nuclear PG species, and the potential involvement of nuclear PG in the temporal and spatial regulation of nuclear PKC ␤ II activity.