(Received for publication, January 17, 1997, and in revised form, March 31, 1997)
From the Institute for Cancer Studies and * Department of Immunology, The University of Birmingham, Birmingham B15 2TH, United Kingdom
Stimulation of cells with certain agonists often activates both phospholipases C and D. These generate diacylglycerol and phosphatidate, respectively, although the two lipids are also apparently interconvertable through the actions of phosphatidate phosphohydrolase and diacylglycerol kinase. Diacylglycerol activates protein kinase C while one role for phosphatidate is the activation of actin stress fiber formation. Therefore, if the two lipids are interconvertable, it is theoretically possible that an uncontrolled signaling loop could arise. To address this issue structural analysis of diacylglycerol, phosphatidate, and phosphatidylbutanol (formed in the presence of butan-1-ol) from both Swiss 3T3 and porcine aortic endothelial cells was performed. This demonstrated that phospholipase C activation generates primarily polyunsaturated species while phospholipase D activation generates saturated/monounsaturated species. In the endothelial cells, where phospholipase D was activated by lysophosphatidic acid independently of phospholipase C, there was no activation of protein kinase C. Thus we propose that only polyunsaturated diacylglycerols and saturated/monounsaturated phosphatidates function as intracellular messengers and that their interconversion products are inactive.
Stimulation of cells by particular agonists which occupy either heterotrimeric G-protein-coupled receptors or those with an intrinsic tyrosine kinase activity induce an increase in the mass of diradylglycerols (collectively diacylglycerol, alkyl, acylglycerol and alkenyl, acylglycerol; DRG),1 in particular sn-1,2-diacylglycerol (DAG), the physiological activator of protein kinase C (PKC) (1). DAG is produced, together with inositol 1,4,5-trisphosphate which stimulates the elevation of intracellular free calcium concentration, by phospholipase C (PLC)-catalyzed phosphatidylinositol 4,5-bisphosphate hydrolysis. Agonist stimulation of this pathway is rapidly desensitized, DAG generation has been demonstrated to be rapid, but transient, declining toward basal levels within 1-2 min (2, 3). However, there is frequently a second sustained phase of DAG generation. This phase has been associated with an increase in the activation of phospholipase D (PLD)-catalyzed phosphatidylcholine (PC) hydrolysis, producing phosphatidate (PA) which can be converted to DAG by the action of phosphatidate phosphohydrolase. It has also been proposed that DAG can be derived from other pathways, e.g. through a PC-PLC pathway, although the evidence for stimulation of this pathway in mammalian cells remains mostly circumstantial (4, 5).
Cells contain multiple species of DAG, however, a limited subset of these change following stimulation. Comparison of the acyl chain DAG structures with those of the cellular phospholipids indicated that the initial phase of DAG increase was predominantly from inositol phospholipids, while the sustained phase, which was accompanied by an increase in choline release, was probably produced from PC (6-9). The initial phase of DAG generation was made up of specific polyunsaturated DAG species, in particular 18:0/20:3n-9, 18:0/20:4n-6, and 18:0/20:5n-3, while the second phase was predominantly represented by more saturated species (7).
The role of the PLD pathway remains incompletely defined. We have
recently demonstrated that PA, generated by the activation of PLD, can
stimulate rho-mediated actin stress fiber formation in porcine aortic
endothelial (PAE) cells (10). This would suggest that while PLD
activation clearly results in an increase in DAG, it is not certain
that this lipid plays a signaling role. Rather, it may be that the
initial product, PA, functions as a messenger and that it is converted
to DAG to attenuate the signal. In vitro studies have
demonstrated that essentially all DAG species can activate PKC,
however, the in vivo evidence for activation by PLD-derived
DAG remains mixed. One report has demonstrated no activation of PKC in
IIC9 fibroblasts stimulated with a concentration of thrombin which did
not activate inositol lipid hydrolysis but did activate PLD (11), while
another report found activation of PKC under the same conditions
(12). A number of signaling roles have been proposed for PA including
activation of a kinase (13), membrane fusion (14), and actin stress
fiber formation (10). Since DAG is converted to PA in cells by DAG
kinase activity and PA is metabolized to DAG by phosphatidate
phosphohydrolase activity, if both PA and DAG are second messengers an
uncontrolled signaling cycle could result. Therefore we have
re-examined the role of the PLD pathway in regulating PKC and have
analyzed the acyl chain structure of DAG and PA in stimulated cells. We
demonstrate here that the DAG and thus PA derived from PLC activation
is predominantly polyunsaturated while the lipids generated by PLD
activation are saturated or monounsaturated. In addition PLD-derived
DAG does not appear to regulate PKC activity in stimulated PAE
cells.
All solvents were of AnalaR or HPLC grade from Rathburn Chemicals Ltd., Walkerburn, Scotland, United Kingdom. Lipid standards were purchased from Avanti Polar Lipids Inc., Alabaster, AL. Other chemicals were from Sigma-Aldrich Co. Ltd., Poole, Dorset, United Kingdom. Seectide and the anti-pan-PKC antibody were from Calbiochem, the other anti-PKC antibodies were obtained from Affiniti.
Swiss 3T3 fibroblasts were grown in Dulbecco's modified Eagle's medium with Glutamax + 10% newborn calf serum and PAE cells in Ham's F-12 nutrient mixture with Glutamax + 10% fetal calf serum and used upon reaching confluency and quiescence.
Cell StimulationCells were washed twice with phosphate-buffered saline, preincubated for 10 min in Dulbecco's modified Eagle's medium or Ham's F-12 containing 20 mM Hepes pH 7.4 and 0.1% bovine serum albumin (fraction V) at 37 °C followed by a further 10 min in fresh medium ± butanol before stimulation with 100 nM bombesin (Swiss 3T3 cells) or 10 µM sn-1-18:1n-9,2-lysophosphatidic acid (LPA) (PAE cells). Incubations were terminated with ice-cold methanol, 1 µg of 1,2-12:0/12:0 DAG or 1,2-17:0/17:0-PA added as internal standard and the lipids extracted.
DRG AnalysisThe lipid extract was derivatized with 3,5-dinitrobenzoyl-chloride, the DRG classes separated on a Kromasil HPLC column (5 µm, 2.1 mm x 250 mm; Hichrom Ltd., Reading, United Kingdom) and the DRG species separated on a Spherisorb S5ODS2 Excel column (5 µm, 4.6 × 250 mm; Hichrom Ltd.) as described previously (7, 8). Total DAG mass was determined upon lipid extracts using the DAG kinase assay method as described (15).
PA/PBut AnalysisThese lipids were isolated by separation on Silica Gel 60 TLC plates (Merck) using a first development with chloroform/methanol/acetic acid (65:15:7.5, by volume) to two-thirds of the distance up the plate and then with chloroform/methanol/acetic acid/water (75:45:3:1, by volume) to the top of the plate (RF values of 0.6 and 0.8 for PA and PBut, respectively). When PA alone was being analyzed, the lipid extract was spotted in the middle of a plastic-backed Silica Gel 60 TLC plate, developed to the top with chloroform/methanol/concentrated ammonia (65:30:5, by volume), cut 1 cm above the origin, the lower part turned through 180° and developed in the opposite direction with chloroform/methanol/acetic acid (80:20:10, by volume). The lipids were detected by brief exposure to iodine vapor, marked, the iodine removed by exposure to water vapor, the lipids eluted from silica with chloroform/methanol (2:1), dried, and methylated with 400 µl of 3% H2SO4 in dry methanol/diethyl ether (3:1, v/v) at 70 °C for 2 h. Following addition of 1 ml of 5% NaCl, the fatty acid methyl esters and dimethylacetals were extracted with 2 × 500 µl of redistilled hexane, dried, and analyzed by gas chromatography-mass spectrometry (5890GC/5972MSD; Hewlett Packard) on a polar DB-23 capillary column (0.25 µ film, 0.25 mm x 30 m; J & W Scientific) using splitless injection at 220 °C (12 p.s.i. head pressure, purge after 1 min), a temperature program of 55 °C for 2 min then to 140 °C at 70 °C/min and finally to 210 °C at 1 °C/min with interface heating at 270 °C. Identification was by reference to authentic standards. Total PA and PBut masses were quantified on TLC relative to co-chromatographed standards by Coomassie staining and densitometric scanning using an LKB 2400 UltraScan XL as described previously (15).
Analysis of Protein Kinase C IsozymesPAE cells were
stimulated for 5 min with vehicle, LPA (10 µM), or PMA
(100 nM), washed 3 times in ice-cold phosphate-buffered saline and harvested in 20 mM Hepes pH 7.4 containing 0.2 mM phenylmethylsulfonyl fluoride and 20 µg/ml leupeptin.
The samples were frozen overnight at 70 °C, thawed, sonicated for
20 s in a bath sonicator and membranes and cytosol separated by
centrifugation at 300,000 × g for 20 min. The
membranes were washed once and proteins separated by SDS-polyacrylamide
gel electrophoresis. Following transfer to polyvinylidene difluoride
membranes the PKC isozymes were probed using specific monoclonal
antibodies and detected using enhanced chemiluminescence.
For determination of PKC activity, lysates were prepared in 20 mM Tris/HCl, pH 7.6, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM sodium orthovanadate, 0.5% Nonidet P-40, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 80 µg/ml leupeptin, 20 µg/ml aprotinin, precleared with protein A-Sepharose and immunoprecipitated with an anti-pan-PKC antibody (Calbiochem) coupled to protein A-Sepharose. Activity was determined as described by Monks et al. (16) in the presence or absence of 100 ng/ml PMA. Assays were performed in duplicate.
Bombesin stimulated a time-dependent increase in total DAG mass (Table I) and changes in the species profile in Swiss 3T3 cells. As we have previously reported the DAG mass changes observed after 25 s were in species considered characteristic of the inositol phospholipids (7), in particular 18:0/20:5n-3, 18:0/20:4n-6, and 18:0/20:3n-9, although smaller changes were also seen in several other polyunsaturated species. The increases in DAG mass at 5 and 30 min were largely due to elevation of more saturated species, particularly 18:1n-9/18:1n-9, 16:0/18:1n-9, 16:0/16:0, 18:0/18:1n-9, and 16:0/18:0.
|
LPA stimulated a small increase in total DAG mass in PAE cells, but
only by a significant amount after 5 min (Table I, Fig. 1). While there was little overall change in DAG profile
(Fig. 2a), there was an increase in the mass
of 18:1n-9/18:1n-9, 16:0/16:1, and 14:0/18:1n-9 (Fig. 2b).
The basal DAG in this cell was very different to the Swiss 3T3, being
enriched in polyunsaturated species, particularly 18:0/20:4n-6 and
18:0/20:3n-9 (Fig. 2).
Role of PLD Activation in DAG Generation
Cells were stimulated in the presence of 0.3% butanol. This led to the formation of PBut, which as a poor substrate for phosphatidate phosphohydrolase, accumulated and prevented the formation of DAG as a consequence of PLD activation. In Swiss 3T3 cells the alcohol abolished the elevation of total DAG mass at 30 min (Table I), but not at the shorter time points. This result confirms that the early rise in DAG is derived from PLC rather than PLD activation. The presence of butanol did not significantly alter the DAG species profiles at any of the times (data not shown). In PAE cells LPA stimulates PLD activity in the absence of an effect upon PLC (10). Therefore, as expected, butan-1-ol, but not butan-2-ol (which does not function in the transphosphatidylation reaction), prevented the LPA-stimulated increase in DAG mass in the PAE cells (Fig. 1). Similarly, the PMA-stimulated increase in DAG mass was also completely abolished by pretreatment of the cells with butan-1-ol but not butan-2-ol (Fig. 1).
Changes in PA and PButIn bombesin-stimulated Swiss 3T3 cells
PA mass (determined both in the presence and absence of butanol)
reached a maximum between 1 and 5 min before declining, while PBut
levels continued rising for at least 30 min (Fig. 3).
Alkyl,acyl species represented less than 5% of the total PA and PBut
diradyl forms while alkenyl,acyl species represented approximately
0.5% of the total for PBut and less than 0.2% for PA. This contrasts
with Madin-Darby canine kidney cells where PLD was apparently selective
for alkyl,acyl-PC (17), indicating cell and/or species differences for
this enzyme.
LPA stimulated an increase in PA mass in PAE cells from 3.9 nmol/107 cells to 5.0 nmol/107 cells after 20 s and 5.8 nmol/107 cells after 1 min. The mass then declined to 3.5 nmol/107 cells after 10 min. In addition to the effects upon stimulated increases in DAG mass (Fig. 1), we have previously reported that the LPA-stimulated increase in PA mass in PAE cells was inhibited by butan-1-ol but not by butan-2-ol (10).
It was not possible to analyze PA and PBut molecular species in a
similar manner to that performed with DRG since quantitative removal of
phosphate (with alkaline phosphatase) or phosphobutanol (with
phospholipase C) was not achievable. Direct derivatization of PA with a
nitrophenyl group (to enable UV detection on HPLC) using
dicyclohexylcarbodiimide, diisopropyl carbodiimide, triisopropylbenzene sulfonyl chloride, or toluene sulfonyl chloride as a condensing agent
also proved unsuitable due to very low efficiencies,
cis-trans isomerization of the unsaturated fatty acids and
loss of polyunsaturated species. Therefore PA and PBut species changes
were analyzed by determining alterations in fatty acid content of the
purified lipid fractions. Both lipids showed no significant changes in the relative fatty acid compositions, which were predominantly 16:0,
18:0, and 18:1n-9 at all time points in both bombesin-stimulated Swiss
3T3 and LPA-stimulated PAE cells (Fig. 4). The PA
species profile in the absence of butan-1-ol was essentially identical to that obtained in its presence (data not shown), however, inclusion of the alcohol prevented any increase in PA fatty acid mass in PAE
cells, but not in Swiss 3T3 cells (Fig. 4). Only trace levels of
polyunsaturated fatty acids (e.g. 20:4n-6) were detected,
but these did not change in stimulated PAE cells and appeared to
decline slightly following stimulation of Swiss 3T3 cells, probably
reflecting agonist-stimulated inositol phospholipid biosynthesis.
Analysis of the fatty acids in the PBut fractions of both stimulated
cell types showed an increase in saturated (14:0, 16:0, and 18:0) and in monounsaturated (predominantly 18:1n-9) species, with no change in
the polyunsaturated species (Fig. 4).
Activation of Protein Kinase C
DAG stimulates PKC activity,
thus elevated DAG levels in cells should be accompanied by an increase
in PKC activation. We examined the activation of PKC initially by
determining the translocation of the enzyme from the cytosol to the
membrane. PAE cells were found to express PKC,
,
, and
.
Western blotting demonstrated that the
and
were found
exclusively in the membrane fraction in resting and stimulated cells
(Fig. 5). PKC
was found equally distributed between
the membrane and cytosolic fractions in control, PMA- and
LPA-stimulated cells (data not shown). PKC
was also found in both
membrane and cytosolic fractions. However, while LPA did not stimulate
PKC
translocation, PMA stimulated a complete translocation of this
isoform to the membrane (Fig. 5).
To confirm that the translocation reflected changes in PKC activity,
anti-PKC immunoprecipitates were prepared from control, LPA- and
PMA-stimulated PAE cells, and phosphorylation of the selectide PKC
substrate peptide determined. Fig. 6 shows that LPA did
not stimulate PKC activity, while incubation of the cells with PMA
increased activity approximately 3-fold. As a control inclusion of PMA
in the assay stimulated PKC to an equal extent in lysates prepared from
control, LPA- and PMA-stimulated cells.
Agonist-stimulated PLD activity has been proposed to provide the source of sustained DAG generation in cells leading to the sustained activation of protein kinase C. However, it is becoming increasingly apparent that the PA product of PLD activation itself functions as an intracellular messenger. The potential for interconversion of PA and DAG in cells thus raises the possibility of uncontrolled signaling in a normal cell and makes it difficult to reconcile the possibility that both lipids can indeed function as messenger molecules. The results presented in this report provide a molecular basis for separate signaling functions of DAG and PA.
We and others have previously reported that in stimulated cells the rapidly generated DAG is predominantly polyunsaturated and apparently derived from inositol phospholipids (6-9, 18). We now show that the acyl chain structure of the PA in stimulated Swiss 3T3 and PAE cells is predominantly saturated or monounsaturated, in particular 16:0, 18:0, and 18:1n-9. Conversion of this PA to DAG by the action of phosphatidate phosphohydrolase thus produces a saturated/monounsaturated rather than a polyunsaturated species. We have only been able to detect extremely small quantities of polyunsaturated acyl groups in PA, e.g. 20:4n-6 in the presence or absence of butan-1-ol. This presumably reflects the rapid utilization of these species by PA-cytidyl transferase, forming cytidyl monophosphate-PA for the resynthesis of inositol phospholipids, a possibility supported by the observed reduction in polyunsaturated PA mass in stimulated Swiss 3T3 cells. Attempts to trap cytidyl monophosphate-PA, by inositol depletion and LiCl treatment as described by other groups (e.g. Ref. 20) were largely unsuccessful, with no significant accumulation of this lipid (data not shown). Others have also found that this trapping technique did not to work with all cell types (19-21). Gas chromatography-mass spectrometry analysis of cytidyl monophosphate-PA showed predominantly 16:0, 18:0, and 18:1n-6 fatty acids with no detectable 20:3n-9, 20:4n-6, or 20:5n-3 at any time point, suggesting that molecular species containing these fatty acids are selectively metabolized more rapidly than other species.
Analysis of the acyl structure of the PBut formed in cells stimulated in the presence of 30 mM butan-1-ol demonstrated that the saturated/monounsaturated PA was produced by PLD activation, rather than by PLC-catalyzed phospholipid hydrolysis followed by DAG kinase-catalyzed phosphorylation of the generated DAG. The reduction in stimulated DAG generation in the presence of the alcohol defined the fraction generated by PLD activation. Thus the results in Fig. 1 and Table I demonstrate that PLD activation is responsible for sustained DAG generation in both Swiss 3T3 and PAE cells. It was considered possible that the "alcohol trap" of generated PA was incomplete, however, in the LPA-stimulated PAE cells, where PLD is the only agonist-stimulated phospholipase, butan-1-ol, but not butan-2-ol completely prevented DAG generation (Fig. 1).
It has been proposed that PLD-derived DAG can stimulate PKC activity in
chronically stimulated cells (see Ref. 5, for review). This implies
that the acyl chain structure of the DAG is not relevant to the ability
of a species to function as an activator. The PAE cell provided a
useful experimental model to test this hypothesis since LPA stimulation
only increased saturated/monounsaturated DAGs which were produced as a
result of PLD activation. In the PAE cells all the PKC and -
and
a significant proportion of the PKC
and -
were found in the
membrane fraction under basal conditions. This membrane localization
was probably a consequence of the high level of 18:0/20:4n-6 and
18:0/20:3n-9 DAGs (Fig. 2) found in the resting cells. This provides
support for a specific role for the polyunsaturated DAGs in activating
PKC. Fig. 5 clearly shows that the increase in the
saturated/monounsaturated DAG species in LPA-stimulated PAE cells was
unable to induce the translocation of PKC
to a membrane fraction.
PKC
is a calcium independent isoform, thus it would be expected to
be translocated by an increase in DAG mass; as a control inclusion of
PMA clearly induced complete translocation. This lack of translocation
reflected the inability of LPA to stimulate PKC activity in the PAE
cells (Fig. 6). Thus the DAG species produced as a result of PLD
activation do not appear to be regulators of PKC, at least in this cell
line. Therefore, while they can activate in vitro, we
suggest that saturated/monounsaturated DAG species do not regulate PKC
activity in an intact normal cell.
Previous reports (11) have suggested that PKC translocation is a
consequence of phosphatidylinositol 4,5-bisphosphate, rather than PC
hydrolysis since an increase in both DAG and [Ca2+] are
required. It was previously proposed that the DAG derived from PC
hydrolysis did not activate PKC (11). However, this study only examined
PKC
and a later study in the same cell line suggested that the DAG
derived from PC hydrolysis could stimulate PKC
translocation (12),
the results reported here differ from that report. A possible
explanation for the differences in results may be that in the work
reported here we have been able to clearly demonstrate that all of the
increased DAG in the stimulated PAE cells is indeed PLD derived.
An alternative explanation for the lack of stimulation of PKC by
PLD-derived DAG is that the phospholipase has been activated in a
compartment devoid of PKC. Subcellular fractionation studies have
provided evidence for PLD activity in plasma membranes, Golgi membranes, endoplasmic reticulum, and the nuclear membrane (22-26). Thus it is unlikely that PLD-derived DAG would be formed in a membrane
devoid of PKC, particularly since PKC isoenzymes appear to be able to
translocate to most membranes in the cell including those where PLD
activity has been detected. Additional support for our proposal that
PLD-derived DAG does not activate PKC is the observation that
incubation of HL-60 cells with Streptomyces chromofuscus PLD
had no effect upon PKC redistribution, while incubation with
Bacillus cereus phosphatidylinositol-specific phospholipase
C induced cytosol to membrane translocation (27). In keeping with this
result, we were only able to observe sustained activation of PKC and
-
in Swiss 3T3 cells where polyunsaturated DAG species were elevated
(28).
Thus we propose that the DAG derived from PLD hydrolysis is not involved in signaling, rather it is a metabolite utilized in the resynthesis of phospholipids. This also suggests that the monounsaturated/saturated PA species themselves play a signaling role in cells. Indeed, in our recent demonstration of PA-stimulated actin stress fiber formation in PAE cells, the stimulant was the dioleoyl structure (10). In order for PA to function as a signal a target molecule must exist and thus proof of PA's signaling function is dependent upon the identification of specific PA targets. It is, however, implicit in the proposed function for PA, that the polyunsaturated PA species do not bind to and activate such target proteins under physiological conditions and thus do not function as signaling molecules. In conclusion we suggest that, while both PLD and PLC pathways directly generate second messengers, it is only the polyunsaturated DAGs and the saturated/monounsaturated PAs which serve as signals under physiological conditions.