INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In mast cells and basophils, cross-linking the high affinity cell surface receptors for IgE (FcRI)¹ activates the tyrosine kinases Lyn and Syk (Eiseman and Bolen, 1992; Hutchcroft et al., 1992) and initiates a signaling cascade that leads to the secretion of histamine and other granule constituents, to changes in adhesive properties, cell shape, and surface topography and to the de novo synthesis of lipid mediators and cytokines (reviewed in Beaven & Metzger, 1993; Oliver et al., 1997). Tyrosine kinase activation by the FcRI and related members of the multisubunit immunoreceptor family, which includes the T cell antigen receptor, the B cell antigen receptor, and several Fc receptors, depends on immunoreceptor tyrosine-based activation motifs (ITAMs) located in the cytoplasmic domains of individual receptor subunits (Cambier, 1995). The heterotrimeric (₂) FcRI of RBL-2H3 mast cells contains ITAM motifs in both the  and  subunit cytoplasmic domains (Metzger, 1992). Recent studies indicate that cross-linking FcRI activates Lyn, leading to ITAM phosphorylation and Syk activation by its association with the  subunit phospho-ITAM. Syk activation, resulting in the phosphorylation of multiple protein substrates, in turn initiates the signaling cascade (Li et al., 1992; Oliver et al., 1994; Wilson et al., 1995; Rivera and Brugge, 1995).

Pharmacological studies have established that Syk-dependent tyrosine phosphorylation is required for the antigen-stimulated synthesis of inositol (1,4,5) trisphosphate (Ins(1,4,5)P₃), presumably mediated by phospholipase C (PLC) (Deanin et al., 1991; Oliver et al., 1994). It has been shown that RBL-2H3 cells contain two PLC isoforms, PLC1 and PLC2, with PLC2 being the more abundant species (S.G. Rhee, personal communication; also, below). PLC1 is phosphorylated on tyrosine and serine (Park et al., 1991; Li et al., 1992) and translocated to the membrane fraction (Atkinson et al., 1992) after FcRI cross-linking. Antigen-stimulated phosphorylation of PLC2 has not been reported previously, but Atkinson et al. (1993) reported its translocation to a particulate fraction after IgE receptor cross-linking.

The relative contributions of phosphorylation and redistribution to receptor-mediated PLC activation are not known in RBL-2H3 cells, and are incompletely understood in other systems. PLC1 and PLC2 are both monomeric enzymes that contain two pleckstrin homology (PH) domains, two SH2 domains and one SH3 domain (reviewed in Lee and Rhee, 1995). Nishibe et al. (1990) reported that PLC1 can be activated in vitro by incubation with epidermal growth factor (EGF) receptor preparations and ATP under conditions that also stimulate its tyrosine phosphorylation. However, PLC1 activated in vivo was only ~50% deactivated when PLC1 immunoprecipitates were incubated with tyrosine phosphatase, suggesting that several mechanisms contribute to maximal stimulation. Kim et al. (1991) showed that substituting phenylalanine for tyrosine 783 of PLC1 yielded a protein that could associate with tyrosine-phosphorylated sites in the platelet-derived growth factor (PDGF) receptor cytoplasmic domain and could become phosphorylated on other tyrosine sites, but was no longer activated by this interaction. These investigators also showed that mutations at Tyr1254 partially inhibited, and at Tyr771 enhanced, the PDGF-induced activation of PLC1. Several catalytically active PDGF and fibroblast growth factor receptors with cytoplasmic domain phenylalanine to tyrosine mutations that prevented their stable association with PLC also failed to stimulate tyrosine phosphorylation of PLC and Ins(1,4,5)P₃ production (Valilus et al., 1993; Mohammadi et al., 1992; Peters et al., 1992). Additionally, a C-terminal mutant of the EGF receptor, which lacked a PLC binding site at Tyr 992, could phosphorylate PLC1 without stimulating PtdIns(4,5)P₂ hydrolysis (Vega et al., 1992). Together, these studies suggest that a combination of the SH2 domain-mediated association of PLC1 with phosphotyrosine counterstructures and the tyrosine phosphorylation of PLC itself may be required for growth factor-mediated activation. Although the PH and SH3 domains of PLC isoforms have the potential to interact with membrane lipids and proteins and with cytoskeletal proteins (Pawson, 1995; Cohen et al., 1995), the roles of these domains in PLC translocation and activation have not been addressed.

Recently, we (Barker et al., 1995) confirmed the results of Yano et al. (1993) that FcRI cross-linking activates a form of phosphatidylinositol 3-kinase (PI 3-kinase) that is composed of a p85 adaptor subunit and a 110-kDa catalytic subunit. This enzyme phosphorylates phosphatidylinositols in the 3 position of the inositol moiety (reviewed in Stephens et al., 1993; Kapeller and Cantley, 1994). There is recent evidence that PI 3-kinase may regulate cellular activities via an additional role as a protein serine kinase (Dhand et al., 1994; Lam et al., 1994; Barker et al., 1995). We demonstrated that wortmannin, at nM concentrations thought to specifically inhibit PI 3-kinase, blocks secretion, macropinocytosis, and ruffling in antigen-stimulated RBL-2H3 cells (Barker et al., 1995). Despite the selective block of a subset of responses in antigen-stimulated RBL cells, wortmannin had no effect on the activation of Lyn, Syk or MAP kinases. From these results, we speculated that PI 3-kinase is located at a branch point in the FcRI signaling cascade. Unexpectedly, we also demonstrated that wortmannin treatment results in a significant loss of antigen-stimulated Ins(1,4,5)P₃ production (Barker et al., 1995). This novel finding prompted us to explore the potential role of PI 3-kinase in the FcRI-signaling cascade leading to PLC activation. In this report, we directly measure enzymatic activity of PLC1 and PLC2 isoforms in immune complex phospholipase assays, examine their intracellular localization, and determine the effects of wortmannin on the activation and phosphorylation of both PLC isoforms.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Material Sources

Wortmannin was purchased from Sigma (St. Louis, MO). Immunoprecipitating anti-PLC1 antibodies were obtained from Calbiochem, La Jolla, CA (immunizing antigen amino acids 113-132, near the amino terminus of bovine PLC1) or Santa Cruz Biotechnology, Santa Cruz, CA (immunizing antigen amino acids 1249-1262, near the carboxyl terminus of bovine PLC1). The Santa Cruz anti-PLC1 antibody was also used for immunofluorescence and immunoelectron microscopy. For immunoblotting, anti-denatured-PLC1 antibodies (immunizing antigen amino acids 82-100 from bovine PLC1) were purchased from Transduction Labs, Lexington, KY. The anti-PLC2 monoclonal antibody used for immunofluorescence and immunoelectron microscopy was a generous gift from Dr. S.G. Rhee (NIH). Immunoprecipitating anti-PLC2 antibodies (immunizing antigen amino acids 1213-1232 from human PLC2) were also obtained from Santa Cruz. Phosphotyrosine was detected on Western blots with RC20-HRP antibody from Transduction. FITC-, HRP- and rhodamine-lissamine-conjugated secondary antibodies and the rabbit anti-mouse bridging antibody were from Jackson ImmunoResearch (West Grove, PA). Colloidal gold-labeled reagents were from Amersham (Arlington Heights, IL).

Cell Culture and Activation

RBL-2H3 cells were cultured on tissue culture flasks in minimal essential medium (MEM; Life Technologies, Gaithersburg, MD) supplemented with 15% fetal calf serum, penicillin-streptomycin, and L-glutamine. In some experiments, IgE receptors were primed by the addition of anti-DNP-IgE (1 µg/ml) for 12-20 h. Cells were then washed to remove excess IgE, incubated without or with 10 nM wortmannin for 15 min at 37°C, and activated by the addition of 1 µg/ml of the polyvalent antigen, DNP-BSA, at 37°C.

Western Blotting and Immune Complex Phospholipase Assays

Adherent, IgE-primed RBL-2H3 cells (1 × 10⁷ cells for PLC2 experiments and 2 × 10 ⁷ cells for PLC1) were activated for indicated times with DNP-BSA at 37°C. Culture dishes were transferred to a tray of ice, washed immediately with ice-cold PBS, and lysed with Buffer A (20 mM HEPES, pH 7.5, 150 mM NaCl, 1% Triton X-100, 5 mM -glycerophosphate, 0.2 mM sodium orthovanadate, 1 mM EGTA, 1 µg/ml aprotinin and leupeptin). Insoluble material was discarded after microcentrifugation (4 min at 13,000 × g, 4°C) and the supernatant rocked for 1 h with 30 µl Protein A/G Sepharose (Oncogene, Cambridge, MA) prebound to 1 µg anti-PLC1- or PLC2-specific antibodies.

For Western blotting, immune complexes were separated by SDS-PAGE, transferred to nitrocellulose and probed with anti-phosphotyrosine (anti-pY), or anti-PLC1 or 2 antibodies, followed by HRP-labeled second antibodies. Blots were developed with SuperSignal ULTRA (Pierce, Rockford, IL) and detected by autoradiography. The relative amounts of the two isoforms were analyzed with a Molecular Dynamics PhosphorImager with ImageQuant software.

For phospholipase assays, the beads were washed once with reaction buffer (35 mM NaH₂PO₄, pH 6.8, 70 mM KCl, 0.8 mM EGTA, 0.8 mM CaCl₂) and assayed for phospholipase activity with a procedure adapted from Wahl et al. (1992). The substrate was prepared by drying a 100-µl aliquot of PtdIns(4,5)-P₂ (1 mg/ml; Boerhinger-Mannheim-USA, Indianapolis, IN) together with 30 µl of Ptd[³H]Ins(4,5)P₂ (0.3 µCi; Dupont-New England Nuclear, Boston, MA), under a stream of nitrogen. The dried phospholipid was solubilized in 50 µl of 50 mM sodium phosphate, pH 6.8, 100 mM KCl with sonication, followed by adding 50 µl of 5% (80 mM) Triton X-100 and sonication to facilitate incorporation into Triton X-100 micelles. Excess wash buffer was removed from the immune complexes and 10 µl each of 5× reaction buffer and substrate solution added. The beads were incubated at 35°C for 20 min and reactions stopped by transfer to an ice bath with the addition of 100 µl of 1% (wt/vol) bovine serum albumin and 250 µl of 10% (wt/vol) TCA. Samples were centrifuged for 3 min in a swinging bucket microcentrifuge and release of [³H]Ins(1,4,5)P₃ into the supernatant quantified by liquid scintillation counting.

In Vivo Phosphorylation of PLC1 and PLC2

In vivo phosphorylation of PLC species was measured in PLC immunoprecipitates from [³²P]orthophosphate-labeled, resting and antigen-activated RBL-2H3 cells. [³²P]-labeling was performed as described in Li et al. (1992). After 2 or 10 min activation with DNP-BSA, cells (4 × 10⁷) were placed on a tray of ice, washed with ice cold phosphate-free MEM and lysed with 1 ml Lysis Buffer B (50 mM Tris-HCl, pH 7.2, 150 mM NaCl, 1% Triton X-100, 1% Brij-96, 1 mM sodium orthovanadate, 1 µg/ml leupeptin and antipain). Lysates were centrifuged and enzymes isolated from the supernatants by immunoprecipitation with anti-PLC1 or anti-PLC2 antibodies and Protein A/G Sepharose. Beads were washed once with lysis buffer containing 0.1% Brij-96 and three times with lysis buffer without detergent. Laemmli buffer (40 µl) was added after the final wash, samples were boiled for 5 min, and [³²P] incorporation into PLC isoforms analyzed by SDS-PAGE as follows. In some experiments, gels were fixed and stained for protein with Coomassie Blue. Dried gels were used for autoradiography, and phosphoproteins were identified as PLC1 or PLC2 based on electrophoretic mobility relative to Life Technologies "ladder" protein standards, as determined by immunoblotting similar PLC immunoprecipitates prepared from nonradioactive cell lysates. Phosphate incorporation was quantified by PhosphorImager analysis. In other experiments, wet gels were placed in a Millipore Semidry blotter, proteins were transferred to PVDF and dried membranes put to film. PLC bands were excised, digested to constituent amino acids with 6N HCL, and analyzed for phosphoamino acid content with a Hunter Thin Layer Peptide Mapping System (CBS Scientific Co, Del Mar, CA).

Immunolocalization of PLC Isoforms

For fluorescence microscopy, monolayers of RBL-2H3 cells on glass coverslips were activated for 10 min with DNP-BSA. To visualize F-actin, cells were labeled by 30 min incubation in 2% paraformaldehyde, 0.02% saponin and 4 U/ml of rhodamine phalloidin (Molecular Probes, Eugene, OR). For PLC localization, cells were fixed for 10 min with 2% paraformaldehyde, followed by 10 min permeabilization with 0.05% Triton X-100 in PBS. The coverslips were washed in PBS, and incubated sequentially with primary antibodies (1 µg/ml rabbit anti-PLC1 or 1 µg/ml monoclonal anti-PLC2), followed by FITC-conjugated secondary antibodies. Coverslips were mounted on slides and photographed with a Zeiss Photomicroscope III equipped for epifluorescence microscopy. For immunoelectron microscopy, cell suspensions were activated for 2 or 10 min with DNP-BSA. Reactions were stopped by 10 min incubation at room temperature with fixative (10% paraformaldehyde, 0.075% glutaraldehyde, 0.2% picric acid). Cells were collected by centrifugation, washed twice in PBS and held serially in 50% ETOH (ethyl alcohol) (10 min), 70% ETOH (10 min), 2% uranyl acetate in 70% ETOH (60 min), 75% ETOH (10 min), 2:1 LR White in 75% ETOH (10 min), 100% LR white, (4 × 20 min). Cell pellets were embedded in gelatin capsules in 10 ml LR White containing 20 µl accelerator, held on ice for 30 min, and allowed to harden for 2 days at room temperature. Thin sections were mounted on 150 mesh nickel, formvar, and carbon-coated grids. Grids with sections were held in distilled H₂0 for 5 min at room temperature, and nonspecific protein-binding sites blocked for 15 min with 5% bovine calf serum, 0.5% BSA in TBS (20 mM Tris, 155 mM NaCl₂, 20 mM NaN₃, pH 7.6.). Samples were incubated overnight at room temperature with primary antibodies (anti-PLC1 at 0.1 µg/ml; anti-PLC2 at 1 µg/ml) in TBS supplemented with 1% serum, then washed through a series of 10 droplets of TBS, and incubated with 15 nm colloidal gold-conjugated Protein A or 30 nm colloidal gold-conjugated goat anti-mouse IgG in TBS, pH 8.2 (1:25; Amersham). They were again rinsed 10 times in TBS, pH 8.2, by the droplet method. Sections were postfixed with 2% glutaraldehyde, stained with uranyl acetate and lead citrate, and examined with an Hitachi 600 transmission electron microscope.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

FcRI Cross-linking Activates Both PLC1 and PLC2

FcRI cross-linking induces Ins(1,4,5)IP₃ synthesis that reaches a maximum at around 2 min and persists over at least 10 min (inset, Figure 1). This time course should predict the time course of PLC activation and decay in RBL-2H3 cells. To test this, we directly measured the activity of the two PLC isoforms in immunoprecipitates from antigen-stimulated rat tumor mast cells, using methods modified from Wahl et al. (1992). As shown in Figure 1, FcRI cross-linking of control cells (solid bars) causes a substantial increase in both PLC1 (A) and PLC2 (B) activities. We interpret the greater activity of PLC2 as a reflection of the greater abundance of PLC2 in RBL-2H3 cells. The increase in immune complex phospholipase activity is striking after 2 min of stimulation, when cytoplasmic Ins(1,4,5)IP₃ levels are highest (Inset). Consistent with the lower cytoplasmic Ins(1,4,5)IP₃ levels in cells treated with antigen for 10 min (Inset), the activities of both PLC isoforms are reduced, although PLC1 activity in particular is still well above basal levels in immune complexes prepared from cells that were exposed to antigen for 10 min.

View larger version (28K): [in this window] [in a new window]   Figure 1.   FcRI cross-linking activates both isoforms of PLC and PLC1 activation is selectively blocked by wortmannin. PLC activity was measured in anti-PLC immune complexes prepared from resting and antigen-stimulated RBL-2H3 cells from the hydrolysis of [3H]-PtdIns(4,5)P2 in the Triton-based mixed micelle assay described in MATERIALS AND METHODS. Results show activation of PLC1(A) and PLC2 (B) at 2 and 10 min after addition of cross-linking reagent (1 µg/ml DNP-BSA) to control cells (open bars). PLC1 activation, but not that of PLC2, is inhibited in cells pretreated for 15 min with 10 nM wortmannin (hatched bars). Results are the average of duplicate samples in one of at least two similar experiments. Inset: The time course of antigen-induced Ins(1,4,5)IP3 production in control cells, measured with the radioreceptor assay described in Deanin et al. (1991).

Antigen-induced Activation of PLC1, but not PLC2, Is Inhibited by Wortmannin

We showed previously that 10 nM wortmannin inhibits the production of Ins(1,4,5)IP₃ by 50-70% in antigen-stimulated cells (Barker et al., 1995). Therefore, we tested the effects of wortmannin on the activity of PLC isoforms in antigen-stimulated RBL-2H3 cells. As shown in Figure 1A (hatched bars), pretreatment of cells with 10 nM wortmannin, that irreversibly inhibits PI 3-kinase (Thelen et al., 1994), effectively blocks antigen-induced activation of PLC1, as measured in the immune complex phospholipase assay. In contrast, the activation of PLC2 in response to FcRI cross-linking is unaffected by wortmannin (Figure 1B). When wortmannin was added to PLC immune complexes together with substrate, there was no inhibition of phospholipase activity of either isotype (our unpublished results). Thus wortmannin inhibits antigen-stimulated Ins(1,4,5)P₃ production by selectively blocking a step upstream of PLC1 activation in the FcRI signaling cascade.

Phosphorylation of PLC1, but not PLC2, Is Inhibited by Wortmannin

The phosphorylation states of PLC1 and PLC2 were determined in immunoprecipitates prepared from [³²P]-orthophosphate-labeled RBL-2H3 cells. Immunoprecipitates were separated by SDS-PAGE and transferred to PVDF before autoradiography. The results of these experiments are shown in Figure 2A. Both PLC isoforms have very low levels of phosphorylation in resting cells (lanes 1, 4). After 2 min cross-linking of anti-DNP IgE-primed receptors with DNP-BSA, both PLC1 (lane 2) and PLC2 (lane 5) are phosphorylated. The antigen-stimulated phosphorylation of PLC1 is barely detectable in wortmannin-treated cells (lane 3). In contrast, wortmannin does not affect the antigen-induced phosphorylation of PLC2 (lane 6). Similar results were obtained from analyses of immune complexes generated from cells exposed to antigen for 10 min (unpublished observations). Results shown here were obtained with anti-N terminal antibodies to precipitate PLC1; similar results were seen with anti-C terminal antibodies to PLC1. The absence of phosphate-labeled bands in resting and wortmannin-treated cells is not a result of alterations in binding of antibodies to PLC1, as equal amounts of PLC1 are detected by immunoblotting methods in immunoprecipitates isolated under all three conditions (Figure 3B).

View larger version (27K): [in this window] [in a new window]   Figure 2.   A). Wortmannin inhibits the antigen-stimulated phosphorylation of PLC1 but not PLC2. IgE-primed, [32P]orthophosphate-labeled cells (4 × 107) were incubated for 15 min with or without 10 nM wortmannin, followed by 0 or 2 min with 1 µg/ml DNP-BSA. In lanes 1-3, PLC1 was immunoprecipitated from lysis supernatants, separated by SDS-PAGE and detected by autoradiography. PLC1 phosphorylation is not detected in resting cells (lane 1), although longer exposures to film can show low levels of phosphate incorporation (unpublished observation). FcRI cross-linking causes the phosphorylation of PLC1 (lane 2). Wortmannin treatment blocks the phosphorylation of PLC1 (lane 3). Lanes 4 to 6 show that phosphorylation of PLC2 is weakly detectable in resting cells (lane 4), and increases in response to antigen (lane 5). Lane 6 shows that the increase in PLC2 phosphorylation is unaffected by the presence of wortmannin. Similar results were obtained in three experiments.

View larger version (40K): [in this window] [in a new window]   Figure 3.   Phosphoamino acid analysis of PLC isoforms. Phosphate-labeled bands corresponding to PLC1 and PLC2 were excised from PVDF membranes after transfer from SDS-PAGE. Each band was digested to constituent amino acids and the entire sample was used for one-dimensional TLC analysis as described in MATERIALS AND METHODS. Mobilities of phosphoserine and phosphotyrosine were determined on the basis of authentic standards; neither isoform had detectable levels of phosphothreonine. The numbers under each lane represent the relative PhosphorImager units corresponding to phosphoserine and phosphotyrosine, adjusted for background. A) Phosphoamino acid content of PLC1 in resting cells (lane 1), after 2 min (lane 2) of cross-linking FcRI with antigen and after 10 nM wortmannin treatment followed by 2 min with antigen (lane 3). B+D). Western blot analysis of PLC1(B) or PLC2(D) immunoprecipitated from resting and 2 min antigen-activated cells (with and without wortmannin pretreatment) probed with anti-pY (B+D; lanes 1-3), stripped and reprobed with anti-PLC1 antibodies (B; lanes 4-6) or anti-PLC2 (D; lanes 4-6). C) Phosphoamino acid content of PLC2 in resting cells (lane 4), in control cells after 2 min of antigen (lane 5) or in 10 nM wortmannin-treated cells after 2 min of antigen (lane 6).

Because PLC1 was shown to be phosphorylated on both serine and tyrosine after FcRI cross-linking (Li et al., 1992), the PLC bands were excised, acid hydrolysed, and analyzed by TLC to determine their phosphoamino acid content. The results, shown in Figure 3A, establish that wortmannin markedly reduces the incorporation of phosphate into both serine and tyrosine residues of PLC1 in antigen-stimulated cells. Similar results were obtained in two additional experiments and in cells activated for 10 min as well as for 2 min. To confirm these findings, we also used Western blotting methods to probe anti-PLC1 immunoprecipitates with anti-phosphotyrosine antibodies. Results in Figure 3B show neglible levels of phosphotyrosine in PLC1 isolated from resting cells (lane 1) and significant phosphotyrosine immunoreactivity in PLC1 after 2 min of antigen stimulation (lane 2). Wortmannin pretreatment substantially reduces, but does not completely abolish, tyrosine phosphorylation (Figure 3B, lane 3). The Western blots were stripped and reprobed with anti-PLC1 antibodies (Figure 3B, lanes 4-6), showing that equivalent amounts of enzyme were immunoprecipitated from resting or antigen-stimulated cells.

In contrast, wortmannin had little or no effect on phosphorylation of PLC2. We conclude this based on no detectable differences when PLC2 immunoprecipitates were probed with anti-phosphotyrosine antibodies on Western blots (Figure 3D, lanes 1-3) and only slightly lower amounts (20-25%) of phosphoserine and phosphotyrosine in two separate phosphoamino acid analyses of PLC2 from antigen-stimulated cells after wortmannin treatment (Figure 3C). As was the case for PLC1, the amount of immunoprecipitable PLC2 is the same in lysates of resting; antigen-activated; and wortmannin-treated, antigen-activated cells (Figure 3D, lanes 4-6).

Different Distribution of PLC1 and PLC2 in Antigen-stimulated RBL-2H3 Cells: Immunofluorescence Microscopy

Cross-linking the FcRI on RBL-2H3 cells leads to cytoskeletal rearrangements, membrane ruffling, and increased cell adhesion and spreading (Pfeiffer et al., 1984; Pfeiffer and Oliver, 1994). These changes in cell morphology are illustrated in cells stained with rhodamine phalloidin to visualize filamentous actin. Resting cells (Figure 4A) have rounded cell bodies with one or more processes and a microvillous surface. After antigen stimulation, the cells spread and have prominent membrane ruffles (Figure 4B). Antigen-stimulated membrane ruffling, but not spreading, is markedly inhibited in wortmannin-treated cells (Figure 4C).

View larger version (113K): [in this window] [in a new window]   Figure 4.   PLC1 translocates to actin-rich membrane ruffles in antigen-stimulated cells. Filamentous actin was labeled with rhodamine phalloidin in fixed and permeablized cells before and after FcRI cross-linking. The actin stain shows the transformation from microvillous surface architecture in resting cells (A) to ruffles at 10 min of exposure to antigen (B); the majority of wortmannin-treated cells fail to make surface ruffles in response to antigen (C). PLC1 was localized by immunofluorescence microscopy in unstimulated RBL-2H3 cells (D) and antigen-stimulated (1 µg/ml DNP-BSA, 10 min) cells without (E) and with (F) wortmannin treatment. The anti-PLC1 stain shows a redistribution of enzyme from a generally diffuse distribution in resting cells to an association with membrane ruffles in antigen-stimulated cells. The presence of wortmannin inhibits membrane labeling of antigen-stimulated cells. The arrowheads in D and F point to occasional ruffles that form in resting or wortmannin-treated cells and label with PLC1 antibodies. Magnification 750×.

Using isoform-specific antibodies and immunofluoresence microscopy, we found that the majority of PLC1 has a diffuse cytosolic distribution in resting cells, with some labeling of the plasma membrane that is most obvious at the tips of membrane processes (arrowhead, Figure 4D). The membrane association is specific, since it is not seen when antibody is pretreated with immunizing peptide (amino acids 1249-1262 near the carboxyl terminus of bovine PLC1; unpublished observation). In contrast, PLC1 in antigen-activated cells is strongly associated with membrane ruffles (Figure 4E). Membrane association of PLC1 is not apparent in antigen-stimulated cells treated with the PI 3-kinase inhibitor, wortmannin, with the exception of the rare cells that display an incomplete ruffling response (arrowhead, Figure 4F).

The 2 isoform of PLC has a distinctly different distribution from PLC1 in RBL-2H3 cells. Immunofluorescence microscopy with a monoclonal antibody to PLC2 showed strong reactivity in the Golgi region and a patchy distribution along the plasma membrane (Figure 5A) of resting RBL-2H3 cells. Antigen stimulation induces cell spreading, which most likely accounts for the more dispersed appearance of the patches of membrane-associated anti-PLC2 reactivity (Figure 5B). However, the Golgi region still contains the highest concentration of PLC2. There is no detectable association of PLC2 with membrane ruffles at the dorsal surface of the cells. No difference in PLC2 labeling between antigen-stimulated control and wortmannin-treated cells was apparent at the level of immunofluorescence microscopy; in addition, there was no labeling above background levels in cells stained with FITC-conjugated anti-mouse IgG secondary alone (these, and other negative control illustrations below, omitted for space consideration).

View larger version (47K): [in this window] [in a new window]   Figure 5.   Immunoelectron microscopy of PLC1. Resting (A) and antigen-activated (B-D) (1 µg/ml DNP-BSA, 10 min) RBL-2H3 cells were embedded in LR White and thin sections labeled sequentially with anti-PLC1 and 15 nm Protein A-gold particles. Membrane-associated gold particles marking the location of PLC1 are circled in all panels to emphasize the concentration of membrane-associated PLC1 in membrane projections. Panels A and B show the presence of additional gold particles in the cell interior, including the nucleus. Micrographs are representative of results in three different experiments. Bar = 0.5 µm.

Different Distributions of PLC1 and PLC2: Immunoelectron Microscopy

Because ruffles represent folds in membranes, there is potential to misinterpret brightly stained structures as targeted sites of membrane translocation. We therefore localized PLC1 at the ultrastructural level by anti-PLC1 immunogold-labeling of thin sections of LR-White-embedded RBL-2H3 cells and transmission electron microscopy. Typical micrographs of PLC1 distribution in resting and antigen-stimulated cells are shown in Figure 6. Gold labeling was essentially absent from identical samples treated with Protein A-15 nm gold alone. The majority of gold particles labeling PLC1 in resting cells are found in the cell interior. However, a subpopulation of gold particles bound to resting cells are membrane-associated, and these are predictably located on microvilli, rather than on smooth membrane regions (Figure 6A). Gold particles are also present in the cytoplasm and nucleus and at the membrane of antigen-activated cells. The micrographs in Figure 6B-D show that most of the membrane-associated particles in activated cells are associated with lamellae. The preferential labeling of PLC1 in surface projections (microvilli, ruffles) suggests that it targets to regions of high membrane curvature, which are rich in actin.

View larger version (119K): [in this window] [in a new window]   Figure 6.   PLC2 distribution is not markedly altered after antigen stimulation. PLC2 was localized by immunofluorescence microscopy in unstimulated (A) and antigen-stimulated (B) (1 µg/ml DNP-BSA, 10 min) RBL-2H3 cell monolayers. PLC2 is distributed in small clusters associated with the Golgi region (G) and plasma membrane under both conditions. Magnification 750×.

To verify that FcRI cross-linking results in PLC1 recruitment to the plasma membrane, a series of micrographs from replicate experiments were blind-coded and scored for gold particles located within 60 nm of the plasma membrane, within the nucleus, and over the remainder of the cell. Because of the limited contrast of cytoplasmic organelles in LR White-embedded samples, no attempt was made to assign gold particles to intracellular organelles other than the nucleus.

In resting cells, approximately 6% of total gold particles identifying PLC1 were found within 60 nm of the plasma membrane (Figure 8A). The remaining gold particles were distributed between the nucleus (12%) and the cytoplasm (82%). Our observation of intranuclear PLC1 is consistent with earlier reports that detected PLC in nuclear fractions by immunoblotting methods (Marmiroli et al., 1994; Martelli et al., 1994).

FcRI cross-linking of control cells for 2 min increased the proportion of membrane-associated gold particles to approximately 10% (Figure 8A). Almost 15% of gold particles were membrane-associated after 10 min exposure to antigen. We showed previously that the antigen-induced transition of surface topography from a microvillous to a lamellar architecture is visible at 30 seconds, advanced after 2 min (when fringed lamellae are often visible), and complete by 5 to 10 min (Oliver et al., 1997). It thus appears that PLC1 recruitment may accompany the ruffling response. In these experiments, we also observed that the proportion of gold particles in the nucleus did not change after antigen stimulation. Thus, PLC1 is recruited from the cytoplasmic pool to the plasma membrane in response to FcRI cross-linking.

Results in Figure 8B show that the proportion of PLC1 at the plasma membrane of wortmannin-treated cells was slightly elevated at 2 min of antigen stimulation and had returned to basal levels by 10 min with antigen. These data implicate PI 3-kinase directly or indirectly in the process of antigen-induced PLC1 recruitment to the plasma membrane of RBL-2H3 cells.

In sharp contrast to the distribution of PLC1, the majority of gold particles localizing PLC2 were observed in close proximity to the Golgi stacks of both resting and activated cells (Figure 7C,D). Furthermore, the membrane-associated fraction of PLC2 failed to show a preferential localization to membrane ruffles in either resting (Figure 7A) or activated (Figure 7B) cells. Instead, PLC2 labeling was frequently just interior to the cortical actin network (arrow, Figure 7A). It was also noted in association with invaginations at the plasma membrane of activated cells (Figure 7E,F). Although clathrin coats are not visible in the LR White sections, our previous experience analyzing the membrane architecture of RBL-2H3 cells (Mao et al., 1993) allows us to identify these structures as coated pits. Again, gold labeling was essentially absent from samples treated with Protein A 15 nm gold alone in combination with the rabbit anti-mouse bridge or with 30 nm goat anti-mouse colloidal gold.

View larger version (124K): [in this window] [in a new window]   Figure 7.   Immunoelectron microscopy of PLC2. anti-PLC2 was localized in thin sections of LR White-embedded resting (A) and antigen-activated (B-F) RBL-2H3 cells by labeling with anti-PLC2 followed by 15 nm Protein A-gold with a rabbit anti-mouse bridge (B,C,E,F) or with 30 nm anti-mouse IgG-colloidal gold particles (A,D). 15 nm gold particles in B,C are circled for emphasis. Gold particles marking PLC2 associate with the plasma membrane and also prominently label vesicles in the Golgi region (G) in panels C,D. Occasional gold particles label Golgi stacks (arrows, D). Gold labeling was also found in association with coated pits (E,F). Micrographs are representative of results in 3 different experiments. Bar = 0.5 µm.

The results of morphometric analyses showed little or no recruitment of PLC2 to the plasma membrane of activated cells (Figure 8C).

View larger version (33K): [in this window] [in a new window]   Figure 8.   Morphometric measurements of gold-labeled PLC isoforms show wortmannin-sensitive translocation of PLC1, but not PLC2, to the plasma membrane. A) Percent of anti-PLC1 gold label within 60 nm of the plasma membrane of resting, control RBL-2H3 cells, followed by 2 or 10 min of antigen stimulation. B) Percent of anti-PLC1 gold label at the plasma membrane after 15 min wortmannin treatment, followed by 2 or 10 min of antigen stimulation. C) Percent of anti-PLC2 gold label at the plasma membrane of resting, control RBL-2H3 cells, followed by 2 or 10 min of antigen. Micrographs were blind-coded before counting; each experimental point represents the percent of total gold counts from replicate thin sections cut from blocks generated in 2 separate experiments. At least 1,000 gold particles were counted for each experimental condition.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Antigen-induced PLC activation, leading to Ins(1,4,5)IP₃ production, is an early and important event in the FcRI signaling cascade. The present study explores the relative contributions of PLC1 and PLC2 to total PLC activity in antigen-stimulated RBL-2H3 cells and begins to address mechanisms involved in activating these enzymes. The results of assays for phospholipase activity in isozyme-specific immunoprecipitates establish that FcRI cross-linking activates both isoforms of PLC. Activation of both isoforms is associated with their increased phosphorylation on serine and tyrosine. The results of immunofluorescence and immunoelectron microscopic localization studies demonstrate that a small percentage of PLC1 associates with the plasma membrane of resting cells and, where present at the membrane, is primarily associated with membrane projections such as the leading edges of lamellae. After antigen stimulation, additional PLC1 is translocated to the plasma membrane, where it associates with membrane ruffles. In contrast, the majority of PLC2 is concentrated in the Golgi region of both resting and activated cells, and the plasma membrane-associated portion of PLC2 does not increase appreciably after FcRI cross-linking.

The topographical results add PLC1 to a growing list of cytoplasmic proteins that specifically target to specialized regions of the plasma membrane. Other examples include phospholipase A₂, Ras and Grb2 (Bar-Sagi et al., 1988; Bar-Sagi et al., 1993) that are found in microvilli and membrane ruffles of rat fibroblasts and the transmembrane protein, E-selectin, that is localized to the tips of microvilli in resting neutrophils (Erlandsen et al., 1995). Bar-Sagi et al. (1993) also found that a truncated form of PLC1, containing only the SH3 domain, localizes to the stress fibers in fibroblasts. Since antigen-stimulated RBL-2H3 cells do not make stress fibers, but rather concentrate actin in surface projections, it is possible that association with the cytoskeleton is a primary means to recruit this PLC isoform to the plasma membrane. We note however that the other principal antigen-induced actin structures in RBL-2H3 cells, phosphotyrosine-containing adhesive structures known as actin plaques (Pfeiffer and Oliver, 1994), fail to label with anti-PLC1 antibodies. There is also precedent for the patchy distribution of the plasma membrane-associated component of PLC2. Wilson et al. (1994) reported a similar distribution for the heterotrimeric G protein, Gi₂. Additionally, although FcRI molecules are randomly distributed on the plasma membrane before activation, the cross-linked FcRI redistributes away from membrane ruffles and projections and concentrates in clusters in the planar (flat) regions of the plasma membrane. These clusters are subsequently internalized through coated pits. Thus, the plasma membrane distributions of the PLC2 isoform and the cross-linked FcRI partially overlap.

We established previously that nM wortmannin concentrations inhibit antigen-stimulated Ins(1,4,5)P₃ synthesis by 50 to 70% (Barker et al., 1995). This result led us to hypothesize that PI 3-kinase may play a role in PLC activation. We report here that wortmannin blocks the activation of the PLC1 isoform, as measured in an in vitro phospholipase assay. This inhibition is associated with the inhibition of PLC1 phosphorylation and of PLC1 translocation to the plasma membrane. In contrast, wortmannin does not inhibit PLC2 activation and has little or no effect on PLC2 phosphorylation or distribution. These data suggest the possibility that, even though PLC2 is the more abundant enzyme, PLC1 may play a predominant role in mediating antigen-induced PtdIns(4,5)P₂ hydrolysis in RBL-2H3 cells. The results of immunolocalization studies, showing that a substantial proportion of PLC2 resides near the Golgi complex, may partially explain these results. Even though PLC2 is activated by FcRI cross-linking, its ability to contribute to Ins(1,4,5)IP₃ production may be severely limited by its poor access to substrate, presumed to be most abundant at the plasma membrane.

On the basis of evidence that nM concentrations of wortmannin specifically inhibit PI 3-kinase (Thelen et al., 1994; Wymann et al., 1996), we hypothesize that PI 3-kinase contributes to the pathway leading to activation of PLC1. One likely mechanism involves a role for PI 3-kinase lipid products in PLC1 recruitment to the membrane. Both PLC isoforms have binding motifs, such as src (SH2, SH3) and pleckstrin homology (PH) domains, that are implicated in the interaction of other proteins with specific inositol phospholipids (Rameh et al., 1995; Lemmon et al., 1995; Hemmings, 1997). We suppose that distinct features within the hypervariable regions of the two PLC1 isoforms further defines the preferential targeting of the two isoforms to their predominant intracellular localizations. Once recruitment has occurred, interaction with PI 3-kinase lipid products has the potential to directly alter enzymatic activity. For example, D-3 phosphoinositides are known to activate protein kinase C types ,, and  (Toker et al., 1994). Lu et al. (1996) found basal PLC activity was enhanced approximately twofold in lipid micelle assays that incorporated PtdIns(3,4,5)P₃ or PtdIns(3,4)P₂. Thus, it is possible that PLC1 is recruited to the membrane and activated as a result of direct interaction with 3-phosphorylated phosphoinositides. Alternatively, PLC1 recruitment could occur indirectly via a membrane-associated platform/adaptor complex whose assembly is controlled by PI 3-kinase and its metabolites. Once at the membrane, PLC1 would be in close proximity to PtdIns(4,5)P₂ and its other substrates, PtdIns and PtdIns(4)P, as well as to membrane-associated tyrosine kinases. These possibilities-i.e., that D-3 phosphoinositides both recruit PLC1 to the membrane for activation by phosphorylation and directly enhance PLC activity-are not mutally exclusive. There is precedence for multiple roles for PI 3-kinase products in the activation of the serine kinase, c-Akt (also known as protein kinase B or PKB). PI 3-kinase lipid products bind and directly activate c-Akt (Franke et al., 1997). In addition, PtdIns(3,4,5)P₃ activates PDK1, that phosphorylates c-Akt on threonine-308 and up-regulates Akt activity (Stokoe, et al., 1997).

A large fraction of ³²P incorporated into PLC1 after IgE receptor stimulation is on serine (Li et al., 1991; see also Figure 3). In earlier reports, Yamada et al. (1992) showed that several serine/threonine kinase inhibitors reduce the antigen-stimulated hydrolysis of total inositol phospholipids and tyrosine phosphorylation of PLC1 and we showed (Barker et al., 1995) that nM concentrations of wortmannin block both serine and lipid kinase activities of PI 3-kinase. Thus it is also possible that the serine phosphorylation of PLC1 by PI 3-kinase may contribute to its maximal stimulation after receptor cross-linking. Alternatively PI 3-kinase may be upstream of another serine kinase, such as Akt (Burgering and Coffer, 1995; Bos, 1995), that in turn phosphorylates PLC1. The possibility that wortmannin directly inhibits another serine kinase, even at the low nM concentrations used in this study, also cannot be completely excluded. Previous reports of serine phosphorylation of PLC1 in other cell types implicated protein kinase C (PKC) and cAMP-dependent protein kinase (PKA) in the negative regulation of PLC1 (reviewed in Rhee et al., 1993). It follows that, if serine phosphorylation of PLC1 in RBL-2H3 cells is obligatory for maximal activation, then the target serine must be distinct from the PKC or PKA phosphorylation site at serine 1248.

Finally, we note the strong correlation between the ability of wortmannin treatment to inhibit both membrane ruffling and PLC1 activation. It is possible that activation of PLC1 precedes, or is dependent on, its assembly into macromolecular signaling complexes. These signaling complexes are likely to be associated with actin and other cytoskeletal elements that also participate in the formation of plasma membrane ruffles. Our current efforts are focused at defining which of these possible mechanisms underlies our observation that the FcRI-mediated phosphorylation, translocation, and activation of PLC1 is blocked by wortmannin.