Regulation of a Novel Human Phospholipase C, PLC e , through Membrane Targeting by Ras*

Phosphoinositide-specific phospholipase C (PI-PLC) plays a pivotal role in regulation of intracellular signal transduction from various receptor molecules. More than 10 members of human PI-PLC isoforms have been identified and classified into three classes b , g , and d , which are regulated by distinct mechanisms. Here we report identification of a novel class of human PI-PLC, named PLC e , which is characterized by the presence of a Ras-associating domain at its C terminus and a CDC25-like domain at its N terminus. The Ras-associating domain of PLC e specifically binds to the GTP-bound forms of Ha-Ras and Rap1A. The dissociation constant for Ha-Ras is estimated to be approximately 40 n M , comparable with those of other Ras effectors. Co-expression of an activated Ha-Ras mutant with PLC e induces its translocation from the cytosol to the plasma membrane. Upon stimulation with epidermal growth factor, similar translocation of ectopically expressed PLC e is observed, which is inhibited by co-expression of dominant-nega-tive Ha-Ras. Furthermore, using a liposome-based reconstitution assay, it is shown that the phosphatidylinositol

The hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP 2 ) 1 by phosphoinositide-specific phospholipase C (PI-PLC) is a key event initiating intracellular signal transduction from various receptor molecules at the plasma membrane (1). This reaction yields two intracellular second messengers, diacylglycerol and inositol 1,4,5-trisphosphate, which induce activation of protein kinase C and mobilization of Ca 2ϩ from intracellular stores, respectively. Concurrently, the reduction in the PIP 2 concentration is likely an important signal because activities of various actin-binding proteins and pleckstrin homology (PH) domain-containing proteins are modulated through interaction with PIP 2 (2). PIP 2 concentration at the membrane has also been shown to be important in maintenance of the cell shape through regulation of the actin cytoskeleton (3). More than 10 members of human PI-PLC isoforms have been identified and classified into three classes, ␤, ␥, and ␦, exemplified by 150-kDa PLC␤1, 145-kDa PLC␥1, and 85-kDa PLC␦1, respectively (1). All the PI-PLC isoforms contain two regions of high sequence homology, designated X and Y, which constitute the PLC catalytic domain. In addition, other accessory modules, a PH domain, an EF-hand domain, and a C2 domain, are also shared by all the mammalian PLC isoforms reported so far (see Fig. 1A). The three classes of PI-PLCs are linked to receptors by distinct mechanisms (1): PLC␤ isozymes are activated by the ␣ subunits of the G q subfamily of heterotrimeric G proteins as well as by the G␤␥ dimer, whereas PLC␥ isozymes are activated by both receptor type and nonreceptor type protein tyrosine kinases through tyrosine phosphorylation.
Ras proto-oncoproteins are small GTP-binding proteins that function as molecular switches by cycling between the active GTP-bound state and the inactive GDP-bound state (4). They regulate a variety of biological responses including proliferation and differentiation of mammalian cells. In mammalian cells, membrane-anchored GTP-bound Ras interacts directly with the serine/threonine kinase Raf-1, leading to activation of the mitogen-activated protein kinase cascade (5). In addition, recent searches have identified a number of candidate Ras effectors other than Raf-1 and its isoforms B-Raf and A-Raf in mammalian cells (6). All these effector proteins exhibit a GTPdependent interaction with Ras, for which the intact effector region (residues 32-40 of mammalian Ras) is required (4 -6). This interaction, when made with post-translationally modified Ras, induces translocation of the effectors from the cytosol to the plasma membrane, where their substrates or activators are localized. The Ras-dependent membrane translocation has been shown to be prerequisite to activation of the Ras effectors, Raf-1 (7), phosphoinositide 3-kinase (8), and Ral guanine nucleotide dissociation stimulator (RalGDS) (9). Some of the candidate effectors of Ras, including RalGDS and AF-6/Afadin (10,11), have been shown to possess homologous motifs of about 100 amino acids in their Ras-associating (RA) regions (12). Recently, the RA domain of RalGDS was reported to share a * This work was supported by Ministry of Education, Science, Sports and Culture of Japan Grants 11470034, 12215098, 12670116, and 12670136 and by funds from the Sankyo Foundation of Life Science, Hyogo Science and Technology Association, and Kobe University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF190642.
Rap1A, another member of Ras family small GTP-binding proteins, possesses an identical effector region with that of Ras and associates with almost all effector molecules of Ras (15). Although its physiological functions remain to be investigated, the findings that Rap1A can activate the Ras effector B-Raf and may cooperate with Ras in regulation of B-Raf-mediated responses in some cell types (16,17) suggest that Rap1A may regulate the activity of other Ras effectors as well. Recently, Rap1 has been implicated in integrin-mediated cell adhesion signaling following stimulation of cell surface molecules, such as the granulocyte colony-stimulating factor (18), CD31 (19), and the T cell receptor (20).
In this report, we have identified a novel class of human PI-PLC, named PLC⑀, which is characterized by the presence of an RA domain, as a human homolog of Caenorhabditis elegans PLC210 (21). We demonstrate that PLC⑀ is regulated through direct association with Ras, adding another member to the growing list of Ras effectors.

EXPERIMENTAL PROCEDURES
Cloning of PLC⑀ cDNA-The BLAST search (22) of GenBank TM entries identified a human EST clone zb59f12, which had a striking homology to the C-terminal segment of C. elegans PLC210 (see Fig. 1A). To isolate cDNAs coding for the upstream sequences, we performed 5Ј-RACE (23) by using a cDNA library, synthesized from human fetal brain mRNA (Invitrogen, San Diego, CA) by Marathon cDNA amplification procedure, as a template according to the manufacturer's instruction (CLONTECH, Palo Alto, CA). The nucleotide sequences of the 5Ј-RACE-amplified cDNAs were confirmed by isolating and sequencing multiple clones.
In Vitro Binding and Two-hybrid Assays-The post-translationally modified forms of human Ha-Ras and Rap1A were purified as described before (24 -26). Residues 2094 -2303 of PLC⑀, encompassing the RA domain, were expressed in Escherichia coli as an maltose-binding protein (MBP) fusion by using the pMAL-c plasmid. The resulting MBP-PLC⑀ (2094 -2303) was examined for in vitro association with Ha-Ras and Rap1A, which had been loaded with GTP␥S or GDP␤S, as described before (25,26). Interaction of PLC⑀ (2094 -2303) with various effector region mutants of Ha-Ras was examined by the yeast two-hybrid assay employing pGAD-PLC⑀ (2094 -2303) and pGBT-Ha-Ras G12V carrying the mutations as described (27). The ␤-galactosidase activity was measured by a filter assay as described (28) Adenylyl Cyclase Inhibition Assay-Measurements of Saccharomyces cerevisiae adenylyl cyclase activity dependent on GTP␥S-loaded Ha-Ras and its inhibition by purified MBP-PLC⑀ (2094 -2303) were carried out as described before (29).
Assay of PI-PLC Activity-Full-length PLC⑀ was expressed with a FLAG epitope tag in Spodoptera frugiperda Sf9 cells by using a baculovirus vector pVL1393 (PharMingen, San Diego, CA). FLAG-PLC⑀ was affinity purified with resin conjugated with anti-FLAG monoclonal antibody M2 (Sigma). The PI-PLC activities were measured essentially as described (21). Briefly, the fusion proteins were incubated in 50-l reaction mixtures containing 50 mM Mes, pH 6.8, 10 M Ca 2ϩ /EGTA, 100 mM NaCl, 0.2 mg/ml bovine serum albumin, 0.1 mM dithiothreitol, 90 M [ 3 H]PIP 2 (20,000 cpm), and 80 M phosphatidylethanolamine at 30°C for 30 min. [ 3 H]Inositol 1,4,5-trisphosphate produced was extracted and quantitated by liquid scintillation counting. Various EGTA/ Ca 2ϩ buffers giving different concentrations of free Ca 2ϩ were prepared as described (30). pGEX-PLC␦1 (provided by Dr. Tadaomi Takenawa, University of Tokyo, Tokyo, Japan) was used for expression of the full-length rat PLC␦1 as a glutathione S-transferase fusion in E. coli. For a liposome-based reconstitution assay, liposomes were made similarly as described (8) by sonication of dried lipids in buffer containing 20 mM Tris-HCl, pH 7.5, 1 mg/ml fatty acid-free bovine serum albumin, 7 mM EDTA, and 5 mM MgCl 2 . The final lipid concentration was 640 M phosphatidylethanolamine, 600 M phosphatidylserine, 280 M phosphatidylcholine, 60 M sphingomyelin and 160 nM [ 3 H]PIP 2 (5,000 cpm/pmol). Post-translationally modified Ha-Ras preloaded with either GTP␥S or GDP␤S was added to the liposomes at a final concentration of 35 nM and mixed at 4°C for 90 min. Liposomes were collected by centrifugation at 100,000 ϫ g for 2 h and resuspended in the same volume of the buffer. The PIP 2 -hydrolyzing activity of PLC⑀ was measured as described above by incubating 0.1 pmol of purified FLAG-PLC⑀ with the liposomes in a total volume of 50 l.
Fluorescence Microscopy-Full-length PLC⑀ was expressed in COS-7 cells as an enhanced green fluorescent protein (EGFP) fusion. COS-7 cells were cultured in a slide chamber in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and transfected with pcDNA-EGFP-PLC⑀ in combination with a pEFBOS-HA vector carrying Ha-Ras or its mutants (0.5 g each) by using superfect transfection reagent (Qiagen). Cells were transferred to serum-free medium 8 h post-transfection and further cultured for 18 h. After fixation with 3.7% formaldehyde and permeabilization with 0.2% Triton X-100, cells were stained with mouse anti-influenza virus HA monoclonal antibody (Roche Molecular Biochemicals) followed by rhodamine-conjugated secondary antibody. Localizations of PLC⑀ and Ha-Ras were analyzed under a confocal laser microscope (MRC-1024; Bio-Rad).
Immunoprecipitation and Fractionation-Full-length PLC⑀ was expressed with a FLAG epitope tag using pFLAG-CMV (Sigma). COS-7 cells transfected with pFLAG-CMV-PLC⑀ in combination with pEF-BOS-HA-Ha-Ras G12V or pEFBOS-HA-Ha-Ras G12V,Y32F were subjected to serum starvation for 18 h and lysed in buffer A (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM MgCl 2 , 1% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, and 1 M leupeptin) and centrifuged at 100,000 ϫ g for 60 min. HA-Ha-Ras was immunoprecipitated from the supernatant with anti-HA monoclonal antibody. Ha-Ras and PLC⑀ in the immunoprecipitates were detected by immunoblotting with anti-HA antibody and affinity purified rabbit polyclonal anti-PLC⑀ antibody raised against PLC⑀ (2094 -2303), respectively. For examination of subcellular localization, COS-7 cells expressing the above proteins were lysed in buffer A without 1% Triton X-100 by sonication. The lysates were separated into the membrane and cytosolic fractions by ultracentrifugation at 100,000 ϫ g for 60 min. FLAG-PLC⑀ in the membrane and cytosolic fractions was detected with anti-PLC⑀ antibody.

RESULTS
Cloning of Human PLC⑀-Nucleotide sequence determination of the human EST clone zb59f12 revealed a striking homology to the C-terminal region of C. elegans PLC210 including the Y, C2, and RA domains (Fig. 1, A and B), suggesting that this may represent a human homolog of PLC210. We isolated cDNAs coding for the upstream sequences by the 5Ј-RACE procedure using a human fetal brain cDNA library as a template. A putative initiator ATG was identified in the 5Ј-RACEamplified cDNAs as that matched the Kozak consensus sequence (31) and preceded by in-frame stop codons. A composite cDNA was reconstructed by joining the 5Ј-RACE-amplified clones and the EST clone zb59f12 and shown to encode the full-length protein with the size of 2,303 amino acid residues. Comparison of the deduced amino acid sequence with Gen-Bank TM entries indicated that this protein possessed the X, Y, and C2 domains, which were sandwiched between an N-terminal CDC25-like domain and a C-terminal RA domain (Fig. 1, A  and B). The CDC25-like domain is homologous to a family of guanine nucleotide exchange proteins for Ras, represented by S. cerevisiae CDC25. All of these structural elements are conserved with PLC210 except for a difference in the number of the RA domains, confirming that we identified a human homolog of PLC210. As observed with PLC210, this protein lacks a PH domain and an EF-hand domain, both of which are invariably present in all the known human PLC isoforms (Fig. 1A). The observations led us to propose that this protein defines a novel class of human PI-PLC, designated PLC⑀. The PLC⑀ mRNA was 7.6 kb in size and detected in various fetal and adult human tissues, such as brain, muscle, and lung (data not shown). The PI-PLC activity was confirmed by using recombinant full-length PLC⑀ expressed in Sf9 insect cells. Purified PLC⑀ specifically hydrolyzed PIP 2 with a specific activity of 1.35 mol/min/mg protein, which was comparable with that of PLC␦1, 304 nmol/min/mg protein, obtained under the same assay condition. Like most of other PI-PLC isoforms, the PIP 2hydrolyzing activity of PLC⑀ exhibited a dependence on the Ca 2ϩ concentration, and the maximal activity was obtained around 10 M Ca 2ϩ (Fig. 1C).
Specific Binding of PLC⑀ to Ha-Ras and Rap1A-MBP-PLC⑀ (2094 -2303), encompassing the RA domain, was examined for association with Ha-Ras and Rap1A, which had been loaded with GTP␥S or GDP␤S. A dose-dependent association was observed with the GTP-bound forms of Ha-Ras and Rap1A but not with their GDP-bound forms (Fig. 2A). The association with Ras was further analyzed quantitatively by the yeast adenylyl cyclase inhibition assay (29). A dose-dependent inhibition of Ha-Ras-dependent adenylyl cyclase activity was observed by purified PLC⑀ (2094 -2303) added into the reaction mixture (Fig. 2B). MBP-PLC⑀ (2094 -2303) had no effect on the Mn 2ϩdependent adenylyl cyclase activity, indicating that this protein exerted its effect by interacting with Ras but not with adenylyl cyclase (data not shown). We carried out this experiment in the presence of various concentrations of Ha-Ras and MBP-PLC⑀ (2094 -2303) to prove the competitive nature of the inhibition as described (29). At each point of Ha-Ras concentration in the presence of MBP-PLC⑀ (2094 -2303), we obtained free Ha-Ras concentration available for adenylyl cyclase activation as that required for giving the same adenylyl cyclase activity in the absence of MBP-PLC⑀ (2094 -2303). A difference between the original and the free concentrations of Ha-Ras was regarded as that bound to MBP-PLC⑀ (2094 -2303), and a reciprocal of this value was plotted against a reciprocal of the free Ha-Ras concentration (Fig. 2C). This gave a series of straight lines for each value of MBP-PLC⑀ (2094 -2303), which converged on the horizontal axis. The data indicated that MBP-PLC⑀ (2094 -2303) bound directly to Ha-Ras and competitively sequestered it from adenylyl cyclase. The K d value for Ha-Ras was calculated from the point of intersection with the horizontal axis and determined to be approximately 40 nM, which is comparable with those of other Ras effectors (27). The data also indicated that the maximal Ha-Ras binding capacity of PLC⑀ (2094 -2303), calculated from the point of intersection with the vertical axis (29), was about 0.5 pmol of Ha-Ras bound to 1 pmol of MBP-PLC⑀ (2094 -2303).
The interaction of the PLC⑀ RA domain with Ha-Ras was also confirmed by the yeast two-hybrid assay ( Table I). The Comparison with our past results obtained with other Ras effectors (21,27) indicated that the binding specificity of PLC⑀ was indistinguishable from those of PLC210 and Schizosaccharomyces pombe Byr2 but different from those of human Raf-1, RalGDS, and AF-6/Afadin. PLC⑀ (2094 -2303) exhibited a positive interaction with Rap1A but not with other small GTPbinding proteins such as R-Ras, RalA, RhoA, Rac1, and Cdc42 by the two-hybrid assay (data not shown). Furthermore, PLC⑀ was co-immunoprecipitated with activated Ha-Ras, Ha-Ras G12V , but not with Ha-Ras G12V,Y32F , from the total cellular extract of COS-7 cells co-expressing them (Fig. 2D), demonstrating their in vivo association.
Ras Induces Membrane Translocation of PLC⑀ and Stimulates Its PIP 2 -hydrolyzing Activity-To examine whether the association with Ras has a function to recruit PLC⑀ to the plasma membrane, we co-expressed FLAG-tagged PLC⑀ with Ha-Ras G12V in serum-starved COS-7 cells and examined its distribution in the membrane and cytosolic fractions. PLC⑀ was predominantly localized in the cytosolic fraction when expressed alone (Fig. 3A). However, co-expression with Ha-Ras G12V caused a gross increment of the amount of PLC⑀ in the membrane fraction. In contrast, co-expression with Ha-Ras G12V,Y32F failed to affect the distribution of PLC⑀. Next, we took advantage of the GFP fusion and examined the intracellular localization of PLC⑀ under a confocal laser microscope. GFP-PLC⑀ was evenly distributed in the cytosol when expressed alone in the serum-starved COS-7 cells (Fig. 3B, top  panels). However, GFP-PLC⑀ became enriched at the plasma membrane when co-expressed with Ha-Ras G12V (Fig. 3B, middle panels). Again, Ha-Ras G12V,Y32F was found to be incompetent (Fig. 3B, bottom panels). Similar results were obtained by using NIH3T3 cells (data not shown). These data taken to-   gether indicate that Ras associates with PLC⑀ and induces its translocation from the cytosol to the plasma membrane.
We further investigated whether the PIP 2 -hydrolyzing activity of PLC⑀ is regulated through interaction with Ras. Because simple mixing of GTP-bound Ha-Ras with purified PLC⑀ had no effect on its activity (data not shown), we examined the effect of the Ras-induced membrane translocation, which was in principle expected to cause an increase of the PLC⑀ activity by bringing it to the proximity of its substrate PIP 2 . To this end, we employed an in vitro reconstitution assay, in which posttranslationally modified Ha-Ras and [ 3 H]PIP 2 were incorporated into liposomes and incubated with purified PLC⑀. As shown in Fig. 3C, GTP␥S-loaded Ha-Ras stimulated the PIP 2hydrolyzing activity of PLC⑀ by approximately 2.4-fold. In contrast, no stimulation was observed with the GDP␤S-loaded form. This result suggests that the Ras-dependent membrane translocation stimulates the PLC⑀ activity.
Epidermal Growth Factor Stimulates Translocation of PLC⑀ through Ras and Rap1A-To further examine whether the Ras-dependent translocation of PLC⑀ could be induced by a physiological stimulus, COS-7 cells expressing PLC⑀ were stimulated with EGF and analyzed for the distribution of PLC⑀ by immunoblotting. EGF induced transient membrane translocation of PLC⑀, starting from 5 min after the stimulation and terminating at 40 min (Fig. 4A, left panels). The EGF-stimulated membrane translocation was efficiently blocked by coexpression of the dominant-negative Ha-Ras mutant Ha-Ras S17N (Fig. 4A, right panels). Further, translocation of GFP-PLC⑀ to the plasma membrane upon EGF stimulation was observed (Fig. 4B). In addition, an enrichment of GFP-PLC⑀ in the perinuclear region was also noticed (Fig. 4B). We speculated that this perinuclear enrichment was mediated by Rap1A because Rap1A was reported to be activated upon EGF stimulation (32) and because Rap1A was found to be associated with PLC⑀ ( Fig. 2A). In fact, we observed co-localization of GFP-PLC⑀ with Rap1A G12V in this region (Fig. 4C). Moreover, following EGF stimulation, GFP-PLC⑀ was translocated to the plasma membrane in cells overexpressing wild type Ha-Ras, whereas in wild type Rap1A-overexpressing cells, GFP-PLC⑀ was translocated to the perinuclear region (Fig. 4D). Taken together, EGF may direct translocation of PLC⑀ to different subcellular regions through Ras or Rap1A. DISCUSSION We have discovered yet another class of PI-PLC, PLC⑀, which is characterized by the presence of the CDC25-like domain and the RA domain as well as by the absence of PH and EF-hand domains. As predicted from the close homology at the X, Y, and C2 domains with the known PLC isoforms, PLC⑀ exhibited a Ca 2ϩ -dependent PI-PLC activity with the maximal activity observed at 10 M Ca 2ϩ . PLC⑀ possessed the RA domain, which exhibited a specific binding activity to GTP-bound Ha-Ras with an affinity comparable with those of other Raseffector proteins (27). This binding was abolished by specific mutations in the effector region, which constitutes a major interface for interaction with Ras effectors (4). Further, it was shown that Ras could induce translocation of PLC⑀ to the plasma membrane and stimulate its PIP 2 -hydrolyzing activity in the liposome reconstitution assay. These observations indicate that PLC⑀ functions as a direct downstream effector of Ras. Considering that full-length PLC⑀ expressed in Sf9 cells by itself possessed a higher specific activity than PLC␦ and could not be activated in vitro by mixing with Ras, the association with Ras is likely to induce activation of PLC⑀ through its recruitment to the plasma membrane, where its substrate PIP 2 exists. This mechanism is similar to those observed for other Ras effectors, phosphoinositide 3-kinase (8) and RalGDS (9).
However, we cannot exclude the possibility that Ras has another role of directly activating PLC⑀ as observed for Raf-1 (7).
The physiological function of PLC⑀ remains to be elucidated. The observed membrane recruitment of PLC⑀ by EGF treatment suggests that PLC⑀ may be involved in signal transduction from growth factor receptors. In mammalian cells, an increase in PI-PLC activity upon treatment with growth factors has been largely attributed to the activation of PLC␥ (1). However, the observation that the rate of phosphoinositide turnover in Ras-transformed NIH3T3 cells was three times that in untransformed cells implied a persistent stimulation of PI-PLC in these cells (33). Further, injection of anti-PI-PLC antibody has been shown to inhibit Ras-induced mitogenesis (34). These observations are consistent with the possibility that an as yet unidentified species of PI-PLC regulated by Ras may play a role in mammalian cell proliferation, although they might be explained by accessory events accompanying the Ras-induced FIG. 4. EGF-induced translocation of PLC⑀. A, COS-7 cells were transfected with pcDNA-EGFP-PLC⑀ alone or in combination with pEFBOS-HA-Ha-Ras S17N (0.5 g each) by using superfect transfection reagent. The cells were transferred to serum-free medium 8 h posttransfection and further cultured for 18 h before stimulation with EGF (100 ng/ml) for the indicated periods. PLC⑀ in the membrane fraction was detected with anti-PLC⑀ antibody. B, subcellular localization of GFP-PLC⑀ was examined after stimulation with EGF (100 ng/ml) for the indicated periods under the confocal laser microscope. The arrowhead indicates the plasma membrane enrichment of GFP-PLC⑀, and the arrow indicates the perinuclear enrichment. C, HA-Rap1A G12V (left) and GFP-PLC⑀ (middle) co-expressed in COS-7 cells were examined for their subcellular localization as described in Fig. 3B. D, COS-7 cells expressing GFP-PLC⑀ in combination with either wild type Ha-Ras (upper panels) or wild type Rap1A (lower panels) were left unstimulated (left panels) or stimulated with EGF (100 ng/ml) for 5 min (right panels). transformation such as autocrine stimulation of receptors coupled to the known PI-PLCs. Our present findings suggest that they might be accounted for by PLC⑀.
PLC⑀ binds not only to Ras but also to Rap1A as shown in Fig. 2A. Furthermore, Rap1A may mediate translocation of PLC⑀ to the perinuclear region in response to EGF (Fig. 4). Considering that Rap1A (35,36) and protein kinase C (37) were reported to be localized at the Golgi apparatus, PLC⑀, downstream of Rap1A, may have some role in the regulation of Golgi functions mediated by protein kinase C. Further analysis of the function of PLC⑀ may reveal its unique role in some cellular phenomena.