Antagonistic Roles for Phospholipase D Activities in B Cell Signaling: While the Antigen Receptors Transduce Mitogenic Signals Via a Novel Phospholipase D Activity, Phosphatidylcholine-Phospholipase D Mediates Antiproliferative Signals¹

Blymphocytes respond to Ag via clonotypic receptors (sIg).³ While ligation of these Ag receptors on mature B cells can lead to proliferation and differentiation into Ab-secreting cells, the majority of immature B cells respond by becoming anergic or undergoing apoptosis. The earliest signaling events detected following mitogenic stimulation of the B cell receptor include the activation of protein tyrosine kinases (PTKs) and the phospholipase C (PLC)-catalyzed hydrolysis of phosphatidylinositol 4,5-bisphosphate (PtdInsP₂) which generates the intracellular second messengers, inositol 1,4,5-trisphosphate and diacylglycerol (DAG) (reviewed in 1 . However, these pathways appear to be activated in both mature and immature B lymphocytes, suggesting that the Ag receptors on mature immunocompetent B cells may be coupled to additional signaling pathways to transduce differential responses. One candidate pathway for a central role in the transduction of proliferative responses is the phospholipase D (PLD)-mediated hydrolysis of phospholipids to generate phosphatidic acid (PtdOH) and a free polar head group. PtdOH has been identified as a putative lipid second messenger and has been suggested to modulate the activity of a number of signal transducing elements, including certain protein kinase C (PKC) isoforms and Ras-like G proteins, particularly Rho (2, 3, 4). In addition, PtdOH via its lysoderivative, lysoPtdOH, has been shown to be a potent mitogen for several cell systems, mediating its effects via pertussis toxin-insensitive and sensitive G protein receptors coupled to the mobilization of calcium and the downstream activation of the Ras-MAP kinase cascade (5). While the two mammalian PLD genes that have now been cloned are both PtdCho specific, distinct PLD subtypes have been identified in mammalian cells that differ in their coupling to cell surface receptors, subcellular localization, and phospholipid substrate specificities (2, 3, 4, 5, 6, 7). Differential activation of such PLD subtypes may generate a variety of PtdOH species in distinct subcellular compartments, with different species possibly eliciting unique downstream effects (2, 3, 4).

We now show that cross-linking of the Ag receptors on B cells induces DNA synthesis and proliferation in a PLD-dependent manner. Although phosphatidylcholine-PLD has been shown to play a central role in the transduction of proliferative responses for a wide variety of calcium-mobilizing receptors, we show that the Ag receptors are not coupled to this phospholipase, but rather to a novel putative phosphatidylinositol-specific PLD. The sIg-generated PtdOH, unlike phosphatidylcholine-derived phosphatidic acid, is converted to DAG. These data provide the first evidence that while the novel sIg-coupled PLD and resultant DAG generation may play a role in B cell proliferation, phosphatidylcholine-PLD may transduce, via phosphatidic acid, negative immunomodulatory signals in mature B lymphocytes.

The human B cell lines, Daudi, EDR, and Ramos, were maintained in RPMI 1640 medium supplemented with penicillin/streptomycin, glutamine, and 10% FCS at 37°C in a 5% CO₂ atmosphere. The following Abs were used in this study. Rabbit anti-mouse Ig (F(ab′)₂) and anti-human CD40 were purchased from Jackson ImmunoResearch (Stratech, U.K.) and Serotec (Oxford, U.K.), respectively. Hybridomas b.7.6. (anti-IgM), AMS28 (anti-IgD), Bet-2 (nonmitogenic anti-sIgM), and 34.5-3 (anti-IA^d) were grown, and the appropriate mAbs purified from culture supernatants as described previously (8, 9). FGK-45 (anti-murine CD40) was donated by Drs. J. Andersson and T. Rolink (Basel Institute for Immunology, Basel, Switzerland) and was purified by S-Sepharose chromatography (Pharmacia, St. Albans, U.K.). Purified NIMR-1 (anti-Thy-1) was donated by Dr. R. M. E. Parkhouse (IAH, Pirbright, U.K.).

PMA, LPS, Ficoll-Hypaque, and Percoll were purchased from Sigma (Poole, U.K.). NHS-biotin and avidin were purchased from Pierce & Warriner (Chester, U.K.). Guinea pig complement was purchased from Sera-Lab (Crawley Down, Sussex, U.K.). Tissue culture media, including, RPMI 1640, Ham’s F-10, FCS, and HEPES, were purchased from Life Technologies (Paisley, U.K.). [9,10-N-³H]palmitic acid (spec. act., 39 Ci/mmol) was purchased from New England Nuclear-DuPont (Boston, MA). [5,6,8,9,11,12,14,15-³H]arachidonic acid (spec. act., 213 Ci/mmol), [³H]myo-inositol (spec. act., 80–120 Ci/mmol), [9,10-N-³H]myristic acid (spec. act., 49 Ci/mmol), [9,10-N-³H]oleic acid (spec. act., 10 Ci/mmol), and [1-¹⁴C]stearate (60–62 mCi/mmol) were purchased from Amersham (Aylesbury, U.K.). [³H]butanol was a gift from Dr. Neil Thompson, Wellcome Laboratories (Beckenham, U.K.). Recombinant murine IL-4 was donated by Immunex (Seattle, WA).

Murine B cells were prepared as described previously (8, 10) Briefly, spleens from male BALB/c mice (12–20 wk old) were excised, and cells were dispersed by being pressed through stainless steel gauze. Depletion of erythrocytes was achieved by centrifugation through a Ficoll-Hypaque cushion (400 × g for 15 min). Lymphocytes were then depleted of T cells by treatment with anti-Thy-1 mAb (NIMR-1) and guinea pig complement. B cells were further purified on a discontinuous Percoll gradient of 85–65-50% (1200 × g for 15 min at 4°C). The higher density cells recovered from the 85–65% interface were termed resting cells, and the lower density cells recovered from the 65–50% interface were termed activated cells. Immature splenic B were prepared from the spleens of 4-wk-old mice and were classified as immature because they do not induce DNA synthesis in response to anti-Ig and are thus not functionally identical with B cells from adult mice (11).

Murine B cells were cultured (5 × 10⁵ cells/200 μl) in flat-bottom microtiter wells in triplicate in RPMI 1640 medium supplemented with glutamine (2 mM), sodium pyruvate (1 mM), nonessential amino acids, β2-ME (50 μM), penicillin (100 U/ml), streptomycin (100 μg/ml), and 5% FCS. [³H]thymidine at 0.5 μCi/well was added at 68 h, and cultures were harvested 4 h later using an automated cell harvester (Skatron, Sterling, VA), and label incorporated into DNA was estimated by liquid scintillation counting (10).

Murine PHA-blasts were prepared essentially as described previously (12). Briefly, erythrocyte-depleted splenocytes (see above) were resuspended (10⁷/ml) in RPMI medium supplemented with penicillin, streptomycin, glutamine, and 10% FCS and cultured overnight in the presence of PHA (50 μg/ml) at 37°C. Following incubation, the cells were separated on a discontinuous Percoll gradient of 85/65/50%. Cells were recovered from the 65–50% interface, fixed with paraformaldehyde, washed, and resuspended in RPMI medium containing 0.1% BSA and 20 mM HEPES, pH 7.4.

Cells to be examined for annexin V expression were washed in PBS and incubated with propidium iodide and/or annexin V-FITC conjugate in defined calcium and magnesium concentrations according to the instructions provided by Boehringer Mannheim (Lewes, East Sussex, U.K.). Cells were acquired (10,000 events) and immediately analyzed with a Becton Dickinson FACScan (Mountain View, CA) using LYSIS II software for analysis.

B cells were resuspended in RPMI 1640 medium containing penicillin, streptomycin, glutamine, and FCS; plated out (0.5 × 10⁷ cells in a final volume of 300 μl) into 24-well plates; and incubated with [³H]palmitate, [³H]arachidonate, [³H]myristate, or [³H]oleate (3 μCi/well) at 37°C for either 4 or 18 h. Methanol (500 μl) was then added to each well, and the disrupted cell suspension was transferred to glass vials. Lipid extracts were prepared by a Bligh-Dyer extraction as described previously (11) and were resolved by TLC on silica gel 60 plates (20 × 20 cm, 250 μm thick; Merck, Darmstadt, Germany) that were activated by prerunning in 1 mM EDTA, air-drying, and heating at 120°C for 1 h before use. Plates were developed in a paper-lined tank, pre-equilibrated with the solvent chloroform/acetone/methanol/glacial acetic acid/water (80/30/26/24/14, v/v/v/v/v). Phospholipids were detected by iodine vapor, and the associated radioactivities corresponding to [³H]PtdIns, -PtdCho, -PtdEth, -PtdSer, and -PtdOH standards were determined by scraping and liquid scintillation counting (11).

PLD activity was measured by transphosphatidylation as described previously (13). To identify PtdCho-specific PLD activity, cells were prelabeled with [³H]palmitate to preferentially label the PtdCho phospholipid pool (11). The specificity of the assay in B cells was defined by showing that PMA stimulated generation of [³H]PtdBut in cells cultured with butan-1-ol (maximal at 0.1–0.3% butan-1-ol) but not butan-2-ol (results not shown). Cells labeled for 4 h were resuspended (5 × 10⁷/ml) and cultured in RPMI 1640 medium, supplemented with 10% dialyzed FCS and the appropriate label (1 μCi/10⁶ cells). Cells labeled for 18 h were resuspended (10⁷/ml) and cultured in RPMI medium, supplemented with penicillin, streptomycin, glutamine, 10% dialyzed FCS, and the appropriate label (1 μCi/10⁶ cells) at 37°C in a 5% CO₂ atmosphere. Labeled cells were then washed and resuspended (3 × 10⁷/ml) in RPMI 1640 containing 20 mM HEPES, 0.1% BSA, and 0.3% butan-1-ol and equilibrated at 37°C for 15 min. To measure the total phospholipid-PLD activity, unlabeled cells were resuspended (3 × 10⁷/ml) in RPMI 1640 medium containing 20 mM HEPES (pH 7.4), 0.1% BSA, and [³H]butan-1-ol (20 μCi/ml) and then incubated at 37°C for 15 min before stimulation. Reactions (in glass vials) were initiated by the addition of cells (3 × 10⁶) to the appropriate stimulus prepared in butanol-containing medium (final volume, 150 μl). Where stock reagents were prepared in solvents such as DMSO, appropriate vehicle controls were added to the unstimulated control cells. After the indicated time, incubations were terminated, and lipid samples were prepared by a Bligh-Dyer extraction (14) for resolution by TLC in the organic phase of a solvent comprising 2,2,4-trimethylpentane/ethyl acetate/acetic acid/water (5/11/2/10, v/v/v/v) using an unlined, unequilibrated chromatography tank. [³H]PtdBut bands, identified by unlabeled or [¹⁴C]PtdBut standards, were scraped and quantitated by liquid scintillation counting (14).

Resting B cells were labeled with [³H]myo-inositol (2.5 μCi/10⁶ cells) for 4 h at 37°C and stimulated with the appropriate ligand, and the reactions were quenched by the Bligh-Dyer two-phase separation. Aliquots of the organic phase were dried down in vacuo and applied to oxalic acid-pretreated silica gel 60 plates. Phospholipids were then separated by TLC, and [³H]PtdIns, [³H]PtdInsP, and [³H]PtdInsP₂ were located by visualization of unlabeled standards. Radioactivity was determined by scraping and liquid scintillation counting (15).

Resting B cells were stimulated with the appropriate ligand, and the reactions were quenched by Bligh-Dyer two-phase separation. Aliquots of the lower organic phase were dried in vacuo, the lipids were solubilized in a Triton X-100/phosphatidylserine mixture, and DAG mass was determined using the DAG kinase assay as previously described (16). The generated PtdOH was separated on silica TLC plates (Merck 5714; 5 × 20 cm 60F₂₅₄), developed in chloroform/methanol/acetic acid (38/9/4.5. v/v/v) to within 5 mm of the top of the plate. The PtdOH bands (relative to standards) were located by autoradiography or phosphorimaging and were scraped into scintillation vials, scintillant was added, and the associated radioactivity was determined by liquid scintillation counting.

Cells were stimulated as described above, and lipids were extracted into chloroform/methanol as previously described (17). Internal standard (1 μg 1, 2–12:0/12:0 DAG) and marker (1 μg 1, 2 + 1, 3–20:0/20:0 DAG) were added to each sample before lipid extraction. Following derivatization with 3,5-dinitrobenzoyl-chloride (50 mg/ml in dry pyridine, 60°C, 15 min) and solid phase extraction, the diradylglycerol derivatives were separated by HPLC, firstly into separate classes on a Kromasil silica column (5 μm, 2.1 × 250 mm) using a solvent gradient of 100% hexane/cyclohexane/diethyl ether/propan-2-ol (49/49/2/0.1, v/v/v/v), changing to 100% cyclohexane/diethyl ether/propan-2-ol (85/15/0.1, v/v/v) over 45 min at 0.5 ml/min and then into the individual molecular species on a Spherisorb S5ODS2 (Sigma-Aldrich, Poole, Dorset, U.K.) reverse phase column (5 μm, 3.2 × 250 mm), using a gradient of acetonitrile/propan-2-ol (9/1, v/v), changing to acetonitrile/propan-2-ol (1/1, v/v) over 45 min at 0.5 ml/min with detection at 254 nm. Identification was through the use of authentic standards and by reference to previously determined relative retention time data (17).

To investigate a role for PLD activities in Ag receptor-mediated B cell proliferation, we have taken advantage of the underlying principle of the definitive PLD transphosphatidylation assay. In this, the primary alcohol, butan-1-ol acts as a preferential nucleophilic acceptor for the phosphatidyl-moiety resulting from PLD-mediated cleavage of phospholipids to generate the inactive, nonmetabolizable product, PtdBut. Butan-1-ol abrogated anti-Ig-stimulated [³H]thymidine incorporation in murine resting splenic B cells (Fig. 1,A). In contrast, butan-1-ol did not substantially block the [³H]thymidine uptake resulting from stimulation of murine splenic B cells with LPS (50 μg/ml; Fig. 1,B). Nonspecific effects of butanol were ruled out by use of the stereoisoform, butan-2-ol, which cannot act as acceptor for the phosphatidyl moiety (Fig. 1).

Butan-1-ol appeared to decrease the levels of DNA synthesis in unstimulated as well as anti-Ig-treated cells; thus, to address whether the effects of butan-1-ol on B cell DNA synthesis reflected the induction of apoptosis and/or cell death, we stained B cells for cell surface expression of phosphatidylserine with annexin V (commitment to apoptosis) and/or uptake of propidium iodide (indicator of cell death). We have found that butanol-1-ol, but not butan-2-ol, leads to the significant induction of apoptosis in unstimulated and anti-Ig-stimulated B cells within 4 h (annexin V single positive cells; Fig. 2,A). Furthermore, cell death resulting from apoptosis (annexin V/propidium iodide double-positive cells) of B cells stimulated with anti-Ig in the presence of butan-1-ol, but not butan-2-ol, was almost complete (range, 75–95%) within 24 h (Fig. 2,B). Evidence that such butan-1-ol-mediated apoptosis is not simply the result of a cell stress-induced event but, rather, reflects specific targeting of a B cell survival/proliferative signal was provided by an additional series of experiments, which showed that culture with IL-4 (Fig. 2 C) could almost completely prevent the induction of cell death by apoptosis. Culture with anti-CD40 could also prevent such butan-1-ol-induced apoptotic cell death, but to a lesser degree (results not shown). Thus, taken together, these data suggest that one or more PLD activities play an important role in the sIg-mediated induction of survival and proliferation of B cells.

Ag receptor coupling to PtdCho-PLD activity was determined by measurement of stimulated [³H]PtdBut formation in [³H]fatty acid-labeled B cells. Labelling studies demonstrated that after 4-h incubation, [³H]palmitate, [³H]myristate, [³H]oleate, and [³H]arachidonate were all preferentially incorporated into PtdCho, indicating that the [³H]PtdBut species generated from such [³H]fatty acid-labeled B cells were derived from PtdCho (11, 14). Cross-linking of the Ag receptors on murine splenic B cells (sIg), using a mitogenic concentration (50 μg/ml) of anti-Ig Abs (F(ab′)₂), did not stimulate the formation of [³H]PtdBut over a 15-s to 60-min period from cells prelabeled with any of the above [³H]fatty acids (Fig. 3,A and results not shown). Stimulation of [³H]palmitate-labeled murine resting splenic B cells with a range of anti-Ig concentrations (0.05–500 μg/ml) or with anti-μ or anti-δ mAbs (50 μg/ml) did not stimulate PtdCho-PLD activity (Fig. 3, B and C, and results not shown). Moreover, stimulation of murine B cells with other polyclonal activators, such as LPS, also failed to stimulate PtdCho-PLD (Fig. 3 B). Similar results were obtained following cross-linking of the Ag receptors on in vivo activated murine splenic B cells and the human lymphoblastoid cell lines, Daudi and Ramos, under conditions in which anti-Ig induced release of [³H]InsPs from [³H]inositol-labeled cells, indicating that these cells were still responsive to stimulation via sIg (results not shown).

Anti-Ig also failed to stimulate the release of [³H]choline (polar head-group product of PtdCho-PLD) or [³H]phosphocholine from [³H]choline-labeled PtdCho; the hydrolysis of [³H]choline-, [³H]glycerol-, [³H]palmitate- or [³²P]-labeled PtdCho in murine B cells and the human B cell line, EDR; or the formation of [³H]PtdBut in [³H]lyso-PAF-labeled B cells (results not shown).

In contrast, PMA, which activates PtdCho-PLD in most cell types, strongly stimulated [³H]PtdBut formation from [³H]palmitate- and [³H]arachidonate-labeled B cells in a linear manner for up to 60 min (Fig. 3), indicating the presence of stimulatable PtdCho-PLD activity in murine B cells. These findings therefore suggest that it is unlikely that 1) the Ag receptors on B cells are coupled to PtdCho-PLD in a cell cycle-dependent manner or 2) PtdCho-PLD plays a role in the maintenance of Ag-driven B cell proliferation.

Stimulation of B cells with anti-Ig Abs probably represents polyclonal activation by type 2 T cell-independent Ags that typically are large molecules of repeating epitopes capable of effectively cross-linking sIg (1). However, the vast majority of soluble Ags are T cell dependent, and hence, both cytokine-directed (e.g., IL-4) and B-T cell contact-mediated (e.g., MHC class II, CD40) signals are required, in addition to those via the Ag receptors, for full B cell activation (1, 18). We therefore determined whether sIg or any of these additionally recruited receptor-mediated signals were coupled to PtdCho-PLD activity during T cell-dependent B cell activation.

Stimulation of murine B cells with IL-4 (1–1000 U/ml), either alone or in combination with anti-Ig (5–50 μg/ml), failed to induce PtdCho-PLD activation over a 60-min period (results not shown) despite costimulation with anti-Ig and IL-4 being mitogenic for B cells (19) (results not shown). Moreover, priming (1–16 h) of B cells with anti-Ig (50 μg/ml), IL-4 (100 U/ml), anti-Ig (5 μg/ml) plus IL-4 (100 U/ml), or LPS (50 μg/ml) failed to induce coupling of sIg to PtdCho-PLD (results not shown).

Secondly, we addressed whether mimicking B-T cell cooperation by ligation of MHC class II molecules by anti-Ia^d Abs would result in PtdCho-PLD activation. Cambier and co-workers have previously shown that class II molecules on murine B cells were differentially coupled to adenylyl cyclase or phospholipase C activation depending on the activation status of the B cells, i.e., resting B cells or cells primed by pretreatment (16 h) with IL-4 plus anti-Ig (18). We therefore investigated the coupling of class II molecules to PtdCho-PLD activity, either alone or in combination with anti-Ig (50 μg/ml), IL-4 (100 U/ml), or anti-Ig (5 μg/ml) plus IL-4 (100 U/ml), on freshly prepared or IL-4/anti-Ig-primed cells. In these experiments, the class II molecules on [³H]palmitate-labeled B cells were stimulated by either cross-linking of biotinylated anti-Ia^d Abs (anti-Ia^d mAbs 39-10-8 and 34.5-3; final concentration, 100 μg/ml) by avidin (25 μg/ml) or incubation on plates precoated with anti-Ia^d Abs. No activation of PtdCho-PLD could be detected under any of these conditions (results not shown).

Finally, to determine whether ligation of B cell surface Ags such as CD40 by T cell counterstructures during B-T cooperation induced PtdCho-PLD activation, [³H]palmitate-labeled resting B cells and B cells preincubated with anti-Ig plus IL-4 were incubated with fixed PHA-splenic T cell blasts in the presence and the absence of IL-4 (100 U/ml). No PtdCho-PLD activity was detected (results not shown). Furthermore, stimulation of resting murine B cells or the human B cell line, EDR, with appropriate anti-CD40 Abs did not induce PtdCho-PLD activity in these cells (results not shown).

Since stimulation of resting immunocompetent B cells under mitogenic conditions did not stimulate PtdCho-PLD activity despite the potential to couple to one or more such activities, we determined whether the Ag receptors were coupled to PtdCho-PLD activity under conditions that induced B cell anergy and/or apoptosis. However, ligation of the Ag receptors on [³H]palmitate-labeled immature splenic B cells or the WEHI 231 immature B lymphoma cell line under conditions that induce anergy in immature splenic B cells (11) (results not shown) and growth arrest and apoptosis in WEHI 231 cells (11, 20) did not induce stimulation of PtdCho-PLD activity. In addition, no stimulation of PtdCho-PLD activity could be detected following stimulation of WEHI 231 cells with LPS (results not shown).

The DNA synthesis studies performed in the presence and the absence of butan-1-ol (Fig. 1) suggested that the Ag receptors were coupled to one or more phospholipase D activities under conditions of mitogenic stimulation. That the Ag receptors on B cells were indeed coupled to such a PLD activity was shown by experiments in which B cells were mitogenically stimulated in the presence of [³H]butan-1-ol (20 μCi/ml) to trap [³H]PtdBut products of all PLD activities regardless of phospholipid substrate specificity (Fig. 4). Stimulation of B cells with anti-Ig (50 μg/ml) under these conditions induced a rapid accumulation of [³H]PtdBut (Fig. 4,A), which reached a maximum over a period of approximately 10 min. As expected, PMA stimulation of B cells also resulted in the accumulation of [³H]PtdBut (Fig. 4,A). Thus, the Ag receptors on murine splenic B cells are coupled to one or more PLD activities. In contrast, stimulation of sIg on WEHI-231 B cells under conditions that lead to growth arrest and apoptosis did not induce the generation of [³H]PtdBut when these B cells were cultured in the presence of [³H]butanol (Fig. 4 B).

The sIg-coupled PLD activity is not a PtdEth-PLD activity, as anti-Ig does not 1) induce detectable hydrolysis of [³H]palmitate-, [³H]arachidonate-, or [³H]ethanolamine-labeled PtdEth (the level of [³H]ethanolamine-labeled-PtdEth detected in anti-Ig treated cells was 112 ± 5% of that observed in control cells following 2-min incubation with stimulus, a time point of optimal [³H]PtdBut generation), or 2) stimulate the release of free [³H]ethanolamine (alternative product of PtdEth-PLD) from [³H]PtdEth in [³H]ethanolamine-labeled B cells. For example, following stimulation of B cells with anti-Ig for 2.5 min (conditions that induce formation of [³H]PtdBut; Fig. 4), the level of free [³H]ethanol amine detected was 7560 ± 440 dpm relative to a control value of 7880 ± 400 dpm. Likewise, the values for [³H]ethanolamine phosphate formation were 6961 ± 307 and 7076 ± 500 dpm, respectively. In contrast, [³H]inositol labeling studies demonstrated that anti-Ig Abs induced a substantial PtdIns hydrolysis, with kinetics consistent with the formation of [³H]PtdBut, in mature (Fig. 4,C), but not immature (Fig. 4 D), murine splenic B cells, suggesting that PtdIns is probably the substrate for the sIg-coupled PLD activity found in mature B cells.

ATP has been widely reported to exhibit immunomodulatory properties on human B cell activation and lymphocyte proliferation (21, 22, 23, 24). ATP exerts its effects on human B cells via P₂-purinoceptors and is coupled to PtdInsP₂-PLC and calcium mobilization (21). In addition, this class of receptors has been shown to be coupled to PtdCho-PLD activation in a number of cell types (reviewed in 25 ; we therefore determined whether P₂-purinoceptors could modulate murine B cell responses and are coupled to PtdCho-PLD activation in murine B cells. We found that ATP inhibits anti-Ig-mediated DNA synthesis in murine B cells in a dose-dependent manner (Fig. 5,A). Stimulation of [³H]palmitate-labeled murine resting splenic B cells with ATP at concentrations that inhibited sIg-mediated DNA synthesis induced generation of [³H]PtdBut, with maximal activation occurring at approximately 200 μM ATP (Fig. 5 B).

P₂-purinoceptors are G protein-coupled receptors (25); we therefore investigated the effects of pertussis toxin, which ADP-ribosylates and hence covalently modifies one or more G_i-like G proteins in B cells (results not shown) (26), on ATP-coupled PtdCho-PLD activation in B cells. Treatment of murine B cells (Fig. 6,A) or Daudi and Ramos B cells (results not shown) with pertussis toxin led to an increased basal level of [³H]PtdBut generation, suggesting that one or more PtdCho-PLD activities could be under the negative control of a G_i-coupled receptor on B cells. Despite the elevated basal level of [³H]PtdBut generation resulting from pertussis toxin treatment, ATP further stimulated [³H]PtdBut formation (Fig. 6,A). In addition, pretreatment of murine B cells with selective tyrosine kinase inhibitors, such as genistein and tyrphostin, under conditions that inhibit sIg-stimulated tyrosine phosphorylation (results not shown) led to an enhanced ATP-mediated PtdCho-PLD signal (Fig. 6 B), suggesting that the ATP-coupled PtdCho-PLD pathway may be under the negative control of a tyrosine kinase(s).

Since there was a difference in the phospholipid substrate of the sIg- and P₂-coupled PLDs, we considered it possible that there would be a distinction between the PtdOH or DAG products generated in response to stimulation of murine B cells with anti-Ig or ATP. The sIg-mediated generation of DAG is biphasic and comprises an early phase believed to correspond to sIg-coupled PtdInsP₂ hydrolysis followed by a second phase that is maximal within 10 min and is sustained for at least 30 min (27) (results not shown). Although the results in this paper suggest that this late phase generation of DAG is not due to either PLD- or PLC-mediated hydrolysis of PtdCho or PtdEth, it was possible that at least some of the sIg-mediated DAG signal was derived, via PtdOH phosphohydrolase activity, from PtdOH generated by the novel sIg-coupled PLD activity. This was tested by determining the mass level of DAG generated over a 30-min period in response to anti-Ig in B cells preincubated with butanol (0.3%) to trap any potential PtdOH species formed. To our surprise, DAG generation at 1 min, which was previously thought to be due to hydrolysis of PtdInsP₂, was almost fully blocked by preincubation with butanol and was therefore derived from PtdOH (Fig. 7,A). Taken together with the results presented in Fig. 4,A, which show that sIg couples to one or more PLD activities within 30 s, these data suggest that a PLD is predominantly responsible for early sIg-mediated DAG signals in B cells. However, only about 50% of the DAG generated at 10 min appears to be derived from PtdOH (Fig. 7,B). Although IP₃ production appears to be desensitized within 2 min of stimulation, high levels (up to about 20% total [³H]inositol-labeled lipids) of total InsPs continue to be generated in a linear manner for at least 2 h following anti-Ig stimulation of B cells (27), and thus, the late and sustained phase DAG may be produced by both PLC- and PLD-dependent mechanisms of inositol lipid hydrolysis. In contrast, similar studies in WEHI-231 cells showed that sIg-stimulated DAG generation was not blocked by preincubation with butanol (Fig. 7 C and results not shown), suggesting that in these immature B cells the DAG generated under conditions leading to growth arrest is not derived via a PLD-derived mechanism.

To rule out the possibility that we were detecting DAG derived from PtdCho-PLD and/or PLC activities not detected by the radiolabel-dependent regimens outlined above, we investigated whether butanol differentially blocked DAG generation resulting from stimulation via sIg vs the PtdCho-PLD activators, ATP and PMA (Fig. 7). Despite ATP stimulating PtdCho-PLD in murine B cells and reports that it stimulates PtdInsP₂ hydrolysis in at least human B cells (21), ATP did not induce DAG generation at any of the time points tested (1, 15, or 30 min; Fig. 7,D and results not shown). Likewise, PMA did not induce any significant generation of DAG over this time course (Fig. 7,A and results not shown). However, following pretreatment with butanol, both PMA and ATP stimulated substantial levels of DAG production (Figs. 7, A and D), suggesting that in B cells, PtdCho-derived PtdOH is not converted to DAG and, indeed, appears to act to suppress DAG levels perhaps by inducing DAG kinase or lipase activities or even, in an analogous manner to PMA (28), by inhibiting PtdInsP₂-PLC.

The finding that sIg and purinoceptors are coupled to the generation of distinct species of PtdOH and DAG was confirmed by experiments analyzing the DAG species generated in B cells following stimulation with either anti-Ig or ATP. Following isolation of the diradylglycerols from control and stimulated cells, separation of the three diradylglycerol classes (DAG, alkylacylglycerol, and alkenylacylglycerol) showed that the cells contained very little alkylacylglycerol or alkenylacylglycerol and that the levels of these diradylglycerols did not vary in response to any stimulation. However, a complex pattern of changes was observed in the DAGs. As we have observed with other cell types (17, 29), the majority of DAG species obtained were saturated/mono- or diunsaturated: in particular, there were high levels of 14:0/16:1, 16:0/16:0, 14:0/18:1, 18:0/18:2, 16:0/18:1n-9, 16:0/18:0, and 18:0/18:0. Together these species made up approximately 52% of the total basal levels of DAG; however, there was also a significant resting level (9%) of an 18:0/20:4n-6 DAG species. Stimulation of the cells with anti-Ig caused an increase in DAG mass that was blocked by inclusion of butanol, as was also observed using the DAG kinase assay (Fig. 7, A and B). Analysis of the sIg-stimulated DAG species following inclusion of butanol revealed a dramatic reduction in the increase in the combined levels of 18:0/18:2 and 16:0/18:1n-9 species (control, 9% of the total DAG; control plus butan-1-ol, 6.8%; sIg, 20%; sIg plus butan-1-ol, 5.9%) and a small, but not statistically significant, reduction in the 18:0/20:4n-6 species. No increase in the relative levels of DAG species was observed in response to stimulation of murine B cells with ATP or PMA, results consistent with those obtained by the DAG mass assay (Fig. 7, A and C). The surprising increases in DAG mass observed in response to ATP or PMA in the presence of butanol in the DAG kinase assay (Fig. 7, A and C) were, however, confirmed by DAG species analysis: for example, in the control cells the level of 16:0/16:0 increased from 6.8 ± 0.1 to 9.0 ± 1.2% in the presence of butanol, while in the ATP-stimulated cells the combined levels of 18:0/18:2 and 16:0/18:1n-9 were increased from 4.2 ± 0.2 to 6.8 ± 1.7% by the alcohol, and in the PMA-stimulated cells butanol increased the level of 18:0/20:4n-6 from 3.7 ± 1.0 to 7.1 ± 1.9%.

These results clearly show that anti-Ig and ATP/PMA induce differential effects on both the mass of DAG and the individual species of DAG generated. Moreover, they indicate that these stimuli use distinct lipid signaling pathways to generate their respective DAG species. Indeed, given that anti-Ig and ATP, but not PMA, similarly stimulate PtdInsP₂-PLC and that butanol blocks sIg-coupled, but not ATP/PMA-coupled, DAG generation in B cells, these results demonstrate that sIg and the purinoceptors couple to distinct PLD signaling pathways in murine B cells. These results also support our contention that the lipid products of stimulated PLD activity are predominantly saturated/monounsaturated or saturated/disunsaturated rather than polyunsaturated (29).

Activation of PLD isoforms, with the resultant generation of distinct PtdOH lipid second messengers, has been proposed to play a central role in the transduction of a variety of cellular responses, including membrane trafficking, the respiratory burst, exocytosis, and proliferation (2, 3, 4, 30). We therefore investigated the role of PLD activities in mitogenic signaling of B lymphocytes via the Ag receptors and now show that one or more PLD activities play an important role in the sIg-mediated survival and proliferation of mature B cells (Figs. 1, 2, and 4). However, we have found, in contrast to the situation with the Ag receptors on T cells (14), that the Ag receptors on mature B cells are not coupled to the classical PtdCho-PLD following mitogenic stimulation with anti-Ig Abs (Fig. 3). Moreover, stimulation of B cells under conditions designed to mimic T cell-dependent activation of B cells failed to induce PtdCho-PLD activation (results not shown). In contrast, ATP, which inhibits murine B cell activation, stimulates PtdCho-PLD activation (Figs. 5 and 6 and results not shown). However, PtdCho-PLD does not appear to transduce the early sIg-mediated signals leading to anergy or apoptosis of immature B cells (results not shown).

The Ag receptors on mature resting murine B cells, however, are coupled to a PtdIns-PLD following mitogenic stimulation with anti-Ig Abs (Fig. 4). The potential importance of this novel sIg-coupled PLD pathway to mitogenic signaling is underscored by the facts that blockage of this pathway results in B cell death by apoptosis (Fig. 2) and that sIg is not coupled to this pathway in immature B cells or the immature B cell line, WEHI-231 (Figs. 4 and 7), which undergoes growth arrest and apoptosis following ligation of sIg. Consistent with this, an undefined PLD activity was recently similarly reported to be coupled to mitogenic stimulation of the Ag receptors on EBV-transformed human B cells (31), and PtdOH has previously been shown to be rapidly generated (within 1 min) following ligation of sIg on human and murine B cells (32). Moreover, our labeling studies, which ruled out a role for a PtdEth-specific PLD and suggested PtdIns to be the mitogenic PLD substrate in murine B cells (Fig. 4), are consistent with studies on human B cells that suggested that formation of [³H]glycerol-labeled PtdOH was not derived from [³H]glycerol-labeled PtdCho or PtdEth but, rather, might have resulted from sequential activation of an acyltransferase and an undefined PLD (32). Evidence that sIg is coupled to a PtdIns-PLD rather than to the glycosylphosphatidylinositol (GPI)-PLD previously shown to play a role in cytokine signaling (33, 34) is provided by both our labeling and our species analysis studies. GPI-PLD activity is characterized by the formation of 14:0/14:0 dimyristoyl-DAG (33); however, there was no formation of [³H]myristate-labeled PtdBut or dimyristoyl-DAG, determined either by species analysis or by reference to standards in the DAG kinase assay. Moreover, in the TLC solvent system used to resolve phospholipids in this study, glycosylphosphatidylinositol migrates in the PtdInsP₂ region and is thus clearly separated from PtdIns (33). As we detect loss of label in the PtdIns region of the TLC plate, it is not a GPI-lipid that is being hydrolyzed (Fig. 4). Furthermore, a precedent for a role for PtdIns-PLD in cellular activation has been set by the report that bradykinin stimulates a rapid and transient PtdIns-PLD, rather than a PtdCho-PLD, response in Madin-Darby canine kidney cells (35).

We therefore propose that while Ag-driven survival and proliferative signals in mature B cells are transduced at least in part by a PtdIns-PLD activity, PtdCho-PLD plays an antagonistic role in transducing antiproliferative signals in mature, immunocompetent B cells. This is the first report, to our knowledge, that PtdCho-PLD pathways may play a role in the negative regulation of cellular proliferation.

Distinct PtdCho-PLD subtypes have been identified in mammalian cells that appear to differ in their coupling to cell surface receptors, subcellular localization, and regulation (2, 3, 4, 30). For example, the two major classes of PtdCho-PLD activity that have been well characterized differ in their cofactor requirements for low m.w. GTPases and PKC, while an oleate-dependent activity remains rather ill defined. Although we have no definitive information regarding the coupling of the ATP-stimulated PtdCho-PLD or, indeed, the novel sIg-coupled PLD in B cells, our data on the regulation of PtdCho-PLD activity by pharmacological agents have highlighted some novel and interesting features of PtdCho-PLD regulation in B cells. In contrast to studies reporting that pertussis toxin and tyrosine kinase inhibitors inhibited agonist-stimulated PLD in neutrophils, fibroblasts, and RBL cells (reviewed in 2 , in murine B cells, pretreatment with pertussis toxin (4 h) and the tyrosine kinase-selective inhibitors, genistein and tyrphostin (1 h), increased both basal and ATP (enhanced in an additive manner)-stimulated PtdCho-PLD activity (Fig. 6). A negative regulatory role for G_i-like G proteins is further supported by our finding that pertussis toxin pretreatment similarly stimulates basal PtdCho-PLD activity in Daudi and Ramos B cells (results not shown). These results with pertussis toxin and tyrosine kinase inhibitors therefore suggest that in B cells PtdCho-PLD activation may be antagonistic to cell growth and, hence, subject to negative control resulting from the cross-talk between G_i-like G proteins and PTKs produced by cross-linking of the B cell receptor and proposed to be involved in sIg coupling to PtdInsP₂-PLC and downstream activation events (36, 37). A role for PtdCho-PLD in the transduction of negative immunomodulatory signals could therefore contribute to the pertussis toxin-mediated inhibition of sIgM-coupled up-regulation of c-fos mRNA expression and DNA synthesis in primary B cells (37).

The finding that the Ag receptors on B cells are not coupled to PtdCho-PLD was initially rather surprising, since many of the receptors known to couple to PtdCho-PLD are, like the Ag receptors, calcium-mobilizing receptors and capable of transducing mitogenic signals (2, 3, 4, 30). Also in common with many studies on a wide range of cell types (2), PMA (presumably acting via PKC) (Fig. 3) stimulated one or more PtdCho-PLD activities in B cells. These results are perhaps at first sight inconsistent with sIg coupling to PKC, but not with PtdCho-PLD activation following mitogenic stimulation of B cells via the Ag receptors (38). However, in B cells, PMA may stimulate one or more PKC isoforms not recruited following ligation of sIg. Our data also suggest that PtdCho-PLD activity is under negative control resulting from the cross-talk between G_i-like G protein(s) and tyrosine kinase-mediated signals (Fig. 6) following ligation of the Ag receptors on B cells (36, 37, 38). Taken together with our findings that ATP and pertussis toxin, which inhibit sIg-mediated DNA synthesis, can also stimulate PtdCho-PLD activity (Fig. 6 and results not shown), our data suggest that PtdCho-PLD may be antagonistic for sIg-driven B cell activation and that cross-talk between Ag receptor-coupled G protein and tyrosine kinase-mediated signals may indeed act to suppress PtdCho-PLD activity. Interestingly, sIg is not coupled to PtdCho-PLD in immature B cells, and thus, PtdCho-PLD does not appear to transduce early sIg-directed anergic and/or apoptotic signals in immature B cells (results not shown).

The immunoregulatory role of P₂-purinoceptors on B cells is unclear, but it is well established that ATP suppresses lymphocyte cytotoxicity and proliferation (22, 23, 24). Previous studies have demonstrated that ATP-specific P₂Z receptors are expressed on human B cells and coupled to PtdInsP₂-PLC and mobilization of calcium (21). In addition, ATP-stimulation of P₂-purinoceptors has been shown to induce the expression of c-fos, c-myc, IL-2R, and transferrin receptors in human B cells (21). However, since sIg is coupled to the generation of similar signals in B cells these data suggested that the proliferative effects mediated via sIg and/or the antiproliferative effects of ATP were probably due to additional transduction events. P₂-purinoceptors have been shown to be coupled to PtdCho-PLD in a number of cell types (25). Thus, the coupling of the P₂-purinoceptors to PtdCho-PLD may at least in part provide a biochemical mechanism for the transduction of ATP-mediated antiproliferative signals in murine B cells.

Although the mechanisms underlying the contrasting biological responses to stimulation of differential PLD activities in B cells (and other cell types) are currently unknown, our results show that while at least some of the sIg-stimulated PtdOH species are metabolized to enhance the DAG signal, PtdOH species derived from PtdCho in response to ATP are not converted to DAG, and indeed appear to suppress DAG that would be expected to be generated from other signaling sources, such as PtdInsP₂. These findings not only support a central role for DAG in the transduction of proliferative signals in B cells but are also consistent with recent reports that PtdCho- and PtdIns-specific PLD activities generate distinct PtdOH species (differing in their fatty acid composition) in different cellular compartments and with potentially differential downstream effector mechanisms. The importance of this is demonstrated by the findings that saturated/monounsaturated, PLD-derived DAGs are unable to activate PKC in porcine aortic endothelial cells (29), and only the steroyl-arachidonyl-species of PtdOH are capable of stimulating the activity of GTPase-inhibiting protein and inhibiting the activity of GTPase-activating protein (2, 3, 4, 39). Moreover, the fate of such distinct PtdOH and DAG species and their downstream effectors may therefore provide a biochemical rationale for our observations that while the putative PtdIns-PLD plays a role in mitogenic signaling in B cells, PtdCho-specific PLD(s) may be involved in the transduction of antiproliferative signals via P₂-purinoceptors, and possibly other immunomodulatory receptors, in mature, immunocompetent B cells.