A novel bifunctional phospholipase c that is regulated by Galpha 12 and stimulates the Ras/mitogen-activated protein kinase pathway.

Three families of phospholipase C (PI-PLCbeta, gamma, and delta) are known to catalyze the hydrolysis of polyphosphoinositides such as phosphatidylinositol 4,5-bisphosphate (PIP(2)) to generate the second messengers inositol 1,4,5 trisphosphate and diacylglycerol, leading to a cascade of intracellular responses that result in cell growth, cell differentiation, and gene expression. Here we describe the founding member of a novel, structurally distinct fourth family of PI-PLC. PLCepsilon not only contains conserved catalytic (X and Y) and regulatory domains (C2) common to other eukaryotic PLCs, but also contains two Ras-associating (RA) domains and a Ras guanine nucleotide exchange factor (RasGEF) motif. PLCepsilon hydrolyzes PIP(2), and this activity is stimulated selectively by a constitutively active form of the heterotrimeric G protein Galpha(12). PLCepsilon and a mutant (H1144L) incapable of hydrolyzing phosphoinositides promote formation of GTP-Ras. Thus PLCepsilon is a RasGEF. PLCepsilon, the mutant H1144L, and the isolated GEF domain activate the mitogen-activated protein kinase pathway in a manner dependent on Ras but independent of PIP(2) hydrolysis. Our findings demonstrate that PLCepsilon is a novel bifunctional enzyme that is regulated by the heterotrimeric G protein Galpha(12) and activates the small G protein Ras/mitogen-activated protein kinase signaling pathway.

There are three established families of PLC termed ␤ (ϳ150 kDa), ␥ (ϳ145 kDa), and ␦ (ϳ 85 kDa) (3). All three families of the PI-PLC family are able to recognize phosphatidylinositol (PI), phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphosphate (PIP 2 ) and to carry out the Ca 2ϩ -dependent hydrolysis of these inositol phospholipids. It is presumed that the primary substrate for hydrolysis is PIP 2 , which yields the second messengers inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol (DAG). IP 3 releases intracellular Ca 2ϩ from the endoplasmic reticulum via interaction with a specific receptor located on the surface of the endoplasmic reticulum. DAG, as well as increased intracellular Ca 2ϩ , activate protein kinase C leading to a cascade of intracellular events including regulation of cellular growth, smooth muscle contraction, and cardiac hypertrophy (2).
The mode of regulation differs considerably for members of the different isoform families. The ␤ isoforms are regulated by large heterotrimeric G proteins. After activation by agonists such as epinephrine, ␣ 1 adrenergic receptors are able to couple to G␣ subunits from the G q /G 11 family and stimulate hydrolysis of phosphatidyl inositol lipids via PLC ␤ isoforms. For some ␤ isoforms, G q alone is sufficient for activation, whereas for others the coordinated action of both G q and ␤␥ is necessary. The ␥ isoforms of PLC contain SH2 and SH3 domains; hence, they are activated by both receptor and nonreceptor tyrosine kinases. Until recently, the mode of regulation of the ␦ class was unknown. Work from our laboratory and that of others has determined that this class is regulated in vitro by lipid ligands and ionized free calcium (regulation by calcium via C2 domain is described by Lomasney et al.) 2 (4 -7).
Activation of G protein-coupled receptors modulates various aspects of cellular growth and proliferation, processes that are primarily controlled by small Ras-related G proteins and their downstream effector, the mitogen-activated protein (MAP) kinases (8,9). There appear to be multiple mechanisms involving heterotrimeric G␣ and ␤␥ subunits by which G protein-coupled receptors regulate small monomeric G protein function (10). G␤␥ subunits can induce phosphorylation of the Shc adapter protein leading to association with the Grb2 docking protein and eventual stimulation of Ras guanine nucleotide exchange activity (11). G␣ i and G␣ o subunits can lead to activation of MAP kinase via a protein kinase C-dependent pathway (12). Recently, a novel molecule, p115 RhoGEF, has been identified that serves as a direct link between the heterotrimeric G␣ subunit G␣ 13 and the small G protein Rho (13,14). G␣ 13 stimulates the nucleotide exchange activity of p115 RhoGEF for Rho. p115 RhoGEF also serves as a GTPase-activating protein for G␣ 13 and G␣ 12 . This is the first example of a protein that is able to directly link large and small G protein pathways. In this report we identify a novel fourth class of PI-PLC that we designate PLC⑀ and demonstrate that PLC⑀ interacts with large and small G proteins, although it is very different from p115 RhoGEF.

EXPERIMENTAL PROCEDURES
Cloning of hPLC⑀ cDNA-A computer search of the human Gen-Bank TM expressed sequence tag (EST) data base was conducted using three relatively short amino acid sequences from the conserved X and Y domains of the mammalian PLCs. An EST clone, zb59f12.s1, showed a high degree of homology and contained a putative open reading frame. Screening of a human placental cDNA library with the EST cDNA probe yielded a larger (3.0 kilobases) but still incomplete cDNA fragment. The full-length PLC⑀ cDNA was generated by 5Ј rapid amplification of cDNA ends utilizing the Marathon cDNA amplification kit (CLON-TECH). The cDNA was reverse transcribed from poly(A) ϩ mRNA obtained from human heart (CLONTECH), using the cDNA synthesis primer provided with the kit. The PCR amplification was carried out using adaptor primer 1 (5Ј-CCATCCTAATACGACTCACTATAGGGC-3Ј) and a gene specific reverse primer, 5Ј-TCCACCGTCTGCCACCAA-ACAACTCCACA-3Ј. The first round PCR was performed according to the recommendations of the manufacturer of 94°C for 1 min, five cycles at 94°C for 5 s and 72°C for 4 min, another five cycles at 94°C for 5 s and 70°C for 4 min, followed by 25 cycles at 94°C for 5 s and 68°C for 4 min. The second round of PCR was performed under the same conditions using a 1:200 dilution of the first round PCR amplification mixture as the template, adaptor primer 2 (5Ј-ACTCACTATAGGGCT-CGAGCGGC-3Ј) and a nested gene-specific primer (5Ј-AGCAGCGGG-CAGAGAGGTGTGTGTCC-3Ј). Southern blot analysis using an end-labeled internal oligonucleotide (5Ј-TGTCAACAGCATCTTTCAG-GTCATCC-3Ј) 5Ј from the PCR primers confirmed that the PCR product contained the expected sequence. The PCR product was subcloned into the PCR2.1 TA cloning vector (InVitrogen) according to the recommendations of the manufacturer. Positive clones were identified by hybridizing filter lifts with the same labeled oligonucleotide that was described for the Southern blots. Clones containing the expected size DNA fragment were then sequenced along both strands using the Taq dye terminator method at the University of Georgia Molecular Genetics Facility, Department of Genetics, using Applied Biosystems 373 and 377 automatic sequencers. The entire hPLC⑀ cDNA was obtained from heart cDNA by PCR using the following primers: 5Ј-ATGGTTTCAGA-AGGAAGTGCAGCAGGAA-3Ј (sense) and 5Ј-TCACTGTCGGTAATCC-ATTGTGTCACTGG-3Ј (antisense). The PLC⑀ cDNA was subcloned inframe into pBluescript KSϩ. The sequence of the inserted DNA was determined as previously mentioned.
Creation of the Phosphodiesterase-deficient Mutant hPLC⑀ H1144L-The phosphodiesterase-deficient mutant of hPLC⑀, PLC⑀ H1144L, was generated by changing the histidine residue at position 1144 to leucine using the QuickChange Kit (Stratagene). Briefly, the fragment of PLC⑀ containing the mutation was amplified with Pfu Turbo Polymerase (Stratagene) using sense (5Ј-GCTTGACGGCGCCTCCGG-3Ј) and antisense (5Ј-CTACCCTACGGGTAGTAAATAGAACCTGTATGCGACTG-TTGGTTC-3Ј) primers containing the mutation changing the histidine (CAT) to a leucine (CTT). The PCR was performed according to the recommendations of the manufacturer of 95°C for 30 s, 12 cycles at 95°C for 30 s, 1 min at 55°C and 4 min at 68°C. A second fragment overlapping the first was amplified using the sense primer, 5Ј-GATG-GGATGCCCATCATTTATCTTGGACATACGCTGACAACCAAG-3Ј and an antisense primer (5Ј-AACGGGGAGGGGGCACGG-3Ј) corresponding to pcDNA3 vector, 3Ј from the multiple cloning site of the vector. The PCR was also performed according to the recommendations of the manufacturer of 95°C for 30 s, 12 cycles at 95°C for 30 s, 1 min at 55°C, and 6 min at 68°C. The overlapping products were then used to amplify the entire PLC⑀ cDNA from nucleotide1278 -6507 using the sense primer 5Ј-GCTTGACGGCGCCTCCGG-3Ј and the antisense primer corresponding to the vector just 3Ј of the multiple cloning site. The PCR conditions were 95°C for 30 s, 12 cycles at 95°C for 30 s, 55°C for 1 min, and 10 min at 68°C. After ligation and plasmid preparation, the mutant insert was digested with SacII and XbaI. The mutated fragment was gel purified using the QIAquick gel extraction kit (Qiagen) and ligated into pcDNA3 in place of the wild type fragment. The sequence was then verified for the presence of the H1144L mutation and the absence of any additional mutations.
Northern Blot Analysis-Human multiple-tissue Northern blot I and blot IV membranes (CLONTECH) containing approximately 2 mg of poly(A) ϩ RNA/lane were probed with full-length cDNA of hPLC⑀ or ␤-actin according to the manufacturer's instructions. Briefly, 30 ng of hPLC⑀ or ␤-actin were random-primed with [␣-32 P]dCTP (Ͻ6000 Ci/ mmol; PerkinElmer Life Sciences) using the Prime-It II random primer labeling kit (Stratagene). Unincorporated radionucleotide was removed from the probe by using Bio-Spin 30 chromatography columns (Bio-Rad). The radiolabeled probes were then denatured by boiling for 5 min and placed on ice. Membranes were prehybridized in 15 ml of Express hyb solution (CLONTECH) with continuous shaking at 68°C for 1 h. The prehybridization buffer was replaced with 10 ml of fresh Ex-pressHyb solution containing the radiolabeled probe. After a 1-h incubation at 68°C with continuous shaking, the blots were quickly rinsed several times with 2ϫ SSC, 0.05% SDS at room temperature and then washed twice for 10 min at room temperature in the same buffer. This was followed by two 20-min washes with fresh 0.1ϫ SSC, 0.1% SDS at 50°C and continuous shaking. The blots were then wrapped in plastic wrap and exposed to HyperFilm MP x-ray film (Amersham Pharmacia Biotech) at Ϫ80°C with two intensifying screens for 4 -24 h before developing the film. Blots exposed to multiple probes were stripped of the first probe as suggested by the manufacturer's protocols.
Transient Transfection of TSA201 Cells-The hPLC⑀ cDNA was subcloned in-frame into the unique HindIII and NotI sites of the mammalian expression vector pcDNA3 (InVitrogen) downstream of the cytomegalovirus promotor. A Kozak consensus sequence (TAAT) and a Myc tag (5Ј-ATGGAGCAGAAGCTGATCAGCGAGGAGGACCTG-3Ј) were incorporated in frame before the start codon. Plasmids were purified for transfection using Qiagen kits. TSA201 cells were cultured in complete Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing glutamine, high glucose, 10% fetal bovine serum (Life Technologies, Inc.), and 50 mg/ml gentamicin (Life Technologies, Inc.) at 37°C in a humidified 5% CO 2 incubator. The cells were transfected with Myc-hPLC⑀ or pcDNA3 empty vector using LipofectAMINE (Life Technologies, Inc.) according to manufacturer's instructions with slight modifications. Cells were harvested 48 h after transfection.
Immunoblotting-Samples were subjected to 6% or 15% SDS-polyacrylamide gel electrophoresis and were electrophoretically transferred to nitrocellulose as described previously (15). hPLC⑀ was detected by Western blot analysis using a primary anti-Myc tag antibody (InVitrogen). G␣* subunits were detected using specific anti-G␣ antibodies. Alkaline phosphatase or horseradish peroxidase-conjugated secondary antibody (IgG) was used for detection.
PLC Activity in TSA201 Cells-TSA201 cells were transfected with either pcDNA3 vector (control), hPLC⑀, hPLC⑀ H1144L, G␣ s *, G␣ i *, G␣ 12 *, G␣ 13 * cDNA, or Ras (2 mg/35-mm plate) alone or in combination using LipofectAMINE (Life Technologies, Inc). To maintain uniform amount of transfected DNA, empty vector was added to the transfection mixture when necessary. The cells were labeled with [ 3 H]myo-inositol (PerkinElmer Life Sciences) for 24 h and harvested 48 h after transfection. The amount of inositol phosphates (inositol 1-phosphate, inositol 1,4-bisphosphate, and IP 3 ) was determined using anion exchange chromatography (16). The percentage of PI hydrolyzed is expressed as the total inositol phosphates formed relative to the amount of [ 3 H]myoinositol incorporated into the phospholipid pool.
MAP Kinase Assay-TSA201 cells were co-transfected (12 mg DNA/ 100 mm plate) with vector, hPLC⑀, hPLC⑀ H1144L, or RasGRF2 in combination with HA-tagged MAP kinase (2 mg). 24 h after transfection the cells were serum starved over night. Cells were lysed 48 h after transfection as described previously with slight modifications (18).
Briefly, cells were washed twice with ice-cold phosphate-buffered saline (Life Technologies, Inc.) and lysed by addition of one volume 10ϫ MAP kinase lysis buffer (5.5% Triton X-100, 0.2 mM phenylmethanesulphonyl fluoride, 0.7 mg/ml pepstatin A, 10 mg/ml leupeptin, 2 mg/ml aprotinin, 20 mM sodium orthovanadate, and 20 mM sodium pyrophosphate) to 10 volumes phosphate-buffered saline. The cells were incubated for 10 min at 4°C and centrifuged at 4°C for 15 min at 10,000 ϫ g. Activated MAP kinase was immunoprecipitated by incubating the cells for 2 h with anti-HA antibodies (Upstate Biotechnology Inc., 0.4 mg/ml). At the end of the incubation period precleared protein A beads (50% slurry) were added, and the samples were rotated 2 h at 4°C. The ability of MAP kinase to phosphorylate myelin basic protein was measured following the manufacturer's protocol (Upstate Biotechnology Inc.). Protein samples were separated on 6% (PLC⑀ and H1144L) and 15% SDS-polyacrylamide gel and subsequently transferred to nitrocellulose at room temperature for 1 or 2 h at 5 or 15 V, respectively, using a semidry transfer cell (Bio-Rad). The HA monoclonal antibody (clone 12CA5; Roche Molecular Biochemicals) was used to detect HA-MAP kinase, anti-FLAG M2 antibody (Sigma) was used to detect FLAGtagged RasGRF2, and anti-Myc antibody (InVitrogen) was used to detect hPLC⑀ and the H1144L mutant. The horseradish peroxidase-coupled goat anti-mouse antibody (1:2,000; Sigma) was used as the secondary antibody. The blots were developed by ECL (Amersham Pharmacia Biotech).
Ras Pull-down Assays-TSA201 cells were transiently co-transfected as previously mentioned with either control plasmid (pcDNA3), hPLC⑀, hPLC⑀ H1144L, or RasGRF2 and Ha-Ras. The cells were serum-starved for 24 h. 48 h after transfection the cells from a 10-cm dish were lysed and scraped in 1 ml of RIPA buffer containing 50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% deoxycholate, 1% Nonidet P-40, 0.1% SDS, 0.1 mM aprotinin, 1 mM leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Lysates were centrifuged at 14,000 rpm for 8 min at 4°C to remove nuclei. The desired amount of bacterial lysate containing the expressed GST-Ras-binding domain (GST-RBD) of Raf1 (prepared as described by de Rooij and Bos (19)) was thawed on ice and incubated with glutathione-agarose beads at 4°C for 1 h. The beads were isolated by centrifugation and washed three times with RIPA buffer containing 50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% deoxycholate, 1% Nonidet P-40, 0.1% SDS, 0.1 mM aprotinin, 1 mM leupeptin, 1 mM pepstatin, and 1 mM phenylmethylsulfonyl fluoride at 4°C. Cell lysates were added to GST-RBD precoupled to glutathione-agarose beads and incubated at 4°C for 1 h. Beads were collected by centrifugation, washed three times with RIPA buffer, and resuspended in SDS sample buffer. GTP bound Ras was identified by precipitation with GST-Raf RBD followed by immunoblotting as described previously by de Rooij and Bos (19). The protein samples were separated on 6% (PLC⑀ and H1144L) and 15% SDSpolyacrylamide gel and subsequently transferred to nitrocellulose at room temperature for 1 h at 5 V using a semidry transfer cell (Bio-Rad). The monoclonal antibody Y13-259 (1:500, Transduction Labs) was used to detect Ras and anti-Myc tag antibody was used to detect hPLC⑀ and the H1144L mutant. The horseradish peroxidase-coupled goat antimouse antibody (1:2,000, Sigma) was used as the secondary antibody. The blot was developed by ECL (Amersham Pharmacia Biotech).

Isolation of the cDNA Encoding Human PLC⑀, Determination of Expression in Human Tissues, and Identification of Structural
Features-X and Y domains are regions of ϳ170 and ϳ260 amino acids, respectively, which share 60% to 40% amino acid identity among all PLC isoforms. The X and Y domains are necessary for phosphodiesterase activity of PLC. We selected three relatively short amino acid sequences from the conserved X and Y domains to use in a BLAST search. The Basic Local Alignment Search Tool (BLAST) is an extremely powerful tool that allows one to simultaneously search multiple nucleotide sequence data bases. Using the BLAST server at the National Center for Biotechnology Information, we were able to identify an EST clone that partially encoded for a novel PLC isoform. Using the EST cDNA as a probe, a human placental cDNA library was screened yielding a larger but still incomplete cDNA. A full-length cDNA clone was ultimately generated with 5Ј rapid amplification of cDNA ends PCR using human heart mRNA as template that we termed PLC⑀. The full-length cDNA (Fig. 1A) possesses an open reading frame of 6.05 kilo-bases encoding a 1994-amino acid protein with a calculated molecular mass of 230,000, making this the largest PLC isolated to date. The next largest would be a member of the ␤ family at ϳ1300 amino acids and ϳ150 kDa. Hybridization of human multiple tissue Northern blots with PLC⑀ cDNA revealed a ϳ7.5-kilobase message corresponding to PLC⑀ expressed in a wide variety of tissues including brain, lung, kidney, testis, and colon with highest expression detected in the heart (Fig. 2). An additional transcript of larger size (ϳ9.5 kilobases) could be observed in most tissues, suggesting the possibility of an alternatively spliced form of PLC⑀ or differential polyadenylation. Results from Southern blotting indicate that isoforms likely do not exist (data not shown).
Structural analysis reveals that PLC⑀ contains the conserved catalytic X and Y domains, thus identifying this isoform conclusively as a Pl-PLC (Fig. 1B). Like members of the three other PLC families, PLC⑀ also contains the regulatory C2 domain. Unlike other eukaryotic PLCs, PLC⑀ appears not to contain a pleckstrin homology domain. A phylogenetic comparison of all known mammalian PLC isoforms (Fig. 1C) demonstrates that PLC⑀ shares little homology with other PLC families (␤, ␥, and ␦) and therefore constitutes a distinct family of PLCs. PLC⑀ is most similar to PLC210, a largely uncharacterized isoform from the nematode Caenorhabditis elegans (20). PLC210 was first identified by open reading frame prediction of genomic sequences generated from the C. elegans genome sequencing project. Although there are significant similarities, human PLC⑀ also differs considerably from PLC210. PLC⑀ is considerably larger than PLC210 (ϳ200 amino acids) and differs extensively in the primary structure of the C terminus and portions of the N terminus. We predict that the C terminus of PLC⑀ will mediate protein-protein interactions that are completely different from those of PLC210. Interestingly, both PLC⑀ and PLC210 share structural domains that are not present in any other PLC, further suggesting the existence of a novel PLC family. Both PLCs contain a Ras binding motif denoted as the RA domain (21). Although the functional role of RA domains is presently unknown, a recent report has demonstrated that PLC210 binds to Ha-Ras via this domain, suggesting a possible role of this isoform in Ras signaling (20). A number of other RA domain-containing proteins such as Ral-GDS and RGL are also known to bind to Ras (21). PLC⑀ has two RA domains at the C terminus (amino acids 1688 -1792 and 1813-1916), suggesting that it might interact with the effector region of Ras (Fig. 1B). PLC⑀ also contains domains in the N terminus, which suggests that it may interact with small G proteins of the Ras superfamily. Very significant homology (p Ͻ 0.00001) is found with aimless RasGEF from Dictyostelium, CDC25 RasGEF from yeast (both Candida albicans and Saccharomyces cerevisiae), human RasGEF homolog Sos1, human RasGEF H-GRF55, and son-of-sevenless RasGEF from mouse. The area of homology is encompassed by the RasGEF catalytic domain signature (G/A/P)CVP(F/Y)X 4 (L/I/M/F/Y)X(D/N)(L/I/ V/M) PROSITE121 (22). No other eukaryotic PLC contains these Ras binding motifs. This strongly suggests that PLC⑀ activates Ras signal transduction pathways.

Expression of PLC⑀ in TSA201 Cells and Characterization of Polyphosphoinositide Hydrolysis-Western blot analysis of cell lysates from TSA201 cells (a clone of human embryonic kidney 293 cells stably expressing simian virus 40 large T antigen)
revealed the presence of a single protein of approximately 230 kDa from cells transfected with PLC⑀ cDNA inserted into the mammalian expression vector pcDNA3 but not from control cells (Fig. 3A) (23). The observed apparent molecular mass of 230 kDa is consistent with the predicted molecular mass of PLC⑀. Extensive PLC⑀ immunoreactivity resides in the partic-ulate fraction of transfected TSA201 cells (data not shown) similar to eukaryotic PLC␤ isoforms but unlike PLC␥ and PLC␦ isozymes that are primarily localized in cytosolic fractions (24,25).
PLC⑀ has Pl-PLC activity, measured by its ability to hydrolyze exogenous PIP 2 , a selective substrate of PI-PLCs. As can be seen in Fig. 3B, 10 g of plasma membranes obtained from PLC⑀ transfected TSA201 cells had a 2-3-fold (61-114 pmol PIP 2 hydrolyzed) greater PLC activity over a 15-min incubation period than membranes of cells transfected with control vector (20 -46 pmol of PIP 2 hydrolyzed).
The Heterotrimeric G Protein G␣ 12 Selectively Stimulates PLC⑀-mediated Hydrolysis of Polyphosphinositides-As depicted in the phylogenetic tree (Fig. 1C), the closest mammalian homolog of PLC⑀ is PLC␤, an effector for heterotrimeric G protein G␣ and ␤␥ subunits. Currently, there are no identifia- ble motifs for proteins that bind to heterotrimeric G proteins. However, truncation of the terminal 112 amino acids of PLC␤1 (Gln 1030 -Leu 1142 ) has been shown to totally abolish regulation of this isoform by G␣ q , suggesting that the C terminus is necessary for G␣ subunit interaction (26). Like PLC␤, PLC⑀ has a long C terminus, suggesting the possibility for regulation by heterotrimeric G proteins. The ability of eight different constitutively active (GTPase-deficient) mutants of G protein ␣ subunits (G␣*) to stimulate PLC⑀ activity in TSA201 cells was determined. In the absence of any extracellular activators, co-transfection of G␣ 12 * with PLC⑀ augmented PLC activity nearly 3-fold greater than control (Fig. 4A). Cells transfected with G␣ 12 * alone did not have increased PLC activity. Cotransfection with G␣ 13 * also led to an increase in PLC activity; however, G␣ 13 alone increased endogenous PLC activity, suggesting that the effects obtained with G␣ 13 * co-transfected with PLC⑀ were in fact equal in magnitude to the sum of the individual effects of PLC⑀ and G␣ 13 * alone. Because co-expression of G␣ 13 * with PLC⑀ tended to decrease PLC⑀ expression somewhat, we cannot be sure that G␣ 13 * has no effect. In many systems G␣ 12 and G␣ 13 are interchangable, that is they both regulate the same effectors. A clear exception has been described for p115RhoGEF. Although both G␣ 12 and G␣ 13 can bind to this novel RhoGEF (p115RhoGEF serves as a GAP for both G␣ subunits), only G␣ 13 can activate the GEF activity toward Rho (14). G␣ s * and G␣ i * had no effect on PLC⑀ activity. The ability of other G␣ subunits: G␣ i2 *, G␣ z *, and G␣ o * to regulate PLC⑀ were also determined; however, none of these G␣ subunits increased PLC activity when co-transfected with PLC⑀ (data not shown). G␣ q * expressed in cells greatly in-creased basal PLC activity most likely by activation of endogenous PLC␤ found in these cells. Co-transfection of PLC⑀ with G␣ q * did not increase PLC activity above G␣ q * transfected cells, suggesting that this G␣ subunit does not activate PLC⑀. PLC⑀ activity was obtained under conditions where similar amounts of PLC⑀ and G␣ subunits were expressed (Fig. 4, B and C).
PLC⑀ is one of very few known effectors for G␣ 12 , because the G␣ 12 /G␣ 13 family of heterotrimeric G proteins is currently poorly characterized (27). G␣ 12 and G␣ 13 appear to play an important role in regulating cellular and cytoskeletal changes and may themselves be regulated by the receptors for thrombin and lysophosphatidic acid and other mitogenic agonists (9). G␣ 12 is in fact a highly oncogenic G␣ subunit. A GTPase deficient mutant G␣ 12 Q229L fully transforms NIH 3T3 cells. The transformed cells form foci, grow in semisolid medium, and form tumors in nude mice (28). G␣ 12 has also been found to regulate extracellular signal-regulated kinase and c-Jun kinase pathways (29,30). It appears that G␣ 12 stimulates c-Jun via activation of Ras; however, the mechanism by which G␣ 12 activates Ras is unknown (31). Perhaps PLC⑀ can act as a direct link between G␣ 12 and Ras.
PLC⑀ Activates Ras and a Downstream Serine/Threonine Kinase MAP Kinase-The presence of a RasGEF motif in the N terminus of PLC⑀ suggests that PLC⑀ can activate Ras by acting as an exchange factor by promoting the exchange of GTP for bound GDP. Ras mediates its effects on cellular growth and transformation mainly by activating a cascade of serine/threonine kinases including Raf, MEK, and MAP kinase (32). Because MAP kinase (ERK1/2) is a downstream effector of Ras signaling, the ability of PLC⑀ to activate Ras was determined by measuring phosphorylation of MAP kinase. To determine whether PLC⑀ could activate Ras and the MAP kinase pathway independent of PI hydrolysis, the ability of an X domain phosphodiesterase deficient mutant of PLC⑀ was examined. X and Y domains are regions of ϳ170 and ϳ260 amino acids, respectively, which share 60% to 40% amino acid identity among all eukaryotic PI-PLCs. The X and Y domains are necessary for phosphodiesterase activity and make up the catalytic core of the enzyme. Bacterial PLCs contain only the X domain. We have previously demonstrated through extensive site-directed mutagenesis of PLC ␦1 that the X domain is responsible for the catalytic hydrolysis of polyphosphoinositides, whereas the Y domain is responsible for substrate binding (33). Mutation of amino acid residues Arg 338 , Glu 341 , and His 356 in the X domain of PLC ␦1 lead to cleavage defective enzymes (33). These residues are absolutely conserved in all eukaryotic PLCs including PLC⑀. A cleavage-defective PLC⑀ was created by mutating the conserved histidine at position 1144 to leucine. This residue is analogous to the histidine at position 356 of PLC ␦1. PLC⑀ H1144L is expressed normally as determined by Western blot using anti-Myc antibodies (Fig. 5B) but does not hydrolyze substrate (data not shown). Thus, any effects on Ras mediated by PLC⑀ H1144L would be due to the Ras binding domains of PLC⑀ rather than due to indirect effects of PI hydrolysis and the production of the second messengers DAG and IP 3 .
To determine whether PLC⑀ and PLC⑀ H1144L could indeed activate the MAP kinase pathway, TSA201 cells were co-transfected with either pcDNA3 vector (control), PLC⑀, PLC⑀ H1144L, or RasGRF2 (as a control for activation of the Ras/ MAP kinase pathway) and HA-tagged MAP kinase. Immunoprecipitates obtained from cells expressing PLC⑀ show a 3-fold (454 pmol/min/mg total protein) increase in phosphorylation of MAP kinase relative to control immunoprecipitates (139 pmol/ min/mg) obtained from cells transfected with vector alone (Fig.  5A). Expression of PLC⑀ H1144L also stimulates phosphorylation of MAP kinase, indicating that PI hydrolysis is not necessary for activation of the Ras effector MAP kinase. In fact, PI hydrolysis seems to inhibit the activation of MAP kinase, because H1144L-stimulated MAP kinase approximately 2-fold greater than wild type PLC⑀ (907 and 454 pmol/min/mg, respectively). The expression of wild type PLC⑀ and H1144L was nearly identical (31.9 versus 29.3 relative units, respectively) in TSA201 cells, as was the level of HA-MAP kinase. Therefore, levels of expression did not account for the differences in activity (Fig. 5B). The products of PI hydrolysis IP 3 and DAG could act as indirect negative regulators of the RasGEF activity and thereby act as a negative feedback loop. For example, stimulation of protein kinase C by DAG might lead to phosphorylation of PLC⑀ and inhibition of the RasGEF activity. The known RasGEF, RasGRF2, gave the most robust stimulation (1231 pmol/min/mg) yet was comparable with the stimulation by H1144L and PLC⑀, demonstrating that PLC⑀ is a fairly robust activator of MAP kinase.
The role of the Ras-associating or RA domains is presently unknown. Proteins such as RalGDS that contain both Ras-GEF and RA domains are assumed to serve as links between different Ras family members. Although RA domains are known to bind to the effector loop of activated members of the Ras superfamily, the RA domain seems to not be necessary for regulation of GDP/GTP exchange by some proteins. For example, activated Ras stimulates RGL (for RalGDS-like), which exchanges GDP for GTP on Ral in vivo. In vitro, however, the RA domain of RGL is not necessary for GDP/GTP exchange on Ral (34). It is presumed that the role of the RA domain in RGL activation of Ral is to mediate redistribution of RGL to the membranes where Ral is located. To differentiate the role of the RasGEF domain from that of the RA domain of PLC⑀ in stimulating MAP kinase, a Myc-tagged construct PLC⑀-GEF was made of the N-terminal 600 amino acids of PLC⑀. This construct contains the entire RasGEF domain but lacks the X and Y phosphodiesterase domains and both of the C-terminal RA domains. PLC⑀-GEF was able to stimulate MAP kinase to a level comparable with the wild type holo enzyme (Fig. 6), suggesting that the RasGEF domain is sufficient for activation of the Ras/MAP kinase pathway by PLC⑀ and that the C-terminal RA domains are not necessary. This result is consistent with the previously described findings for RGL (34).
It is difficult to predict the function of the RA domains in PLC⑀ from these studies. They do not seem to be involved with membrane association, because the PLC⑀-GEF construct is highly targeted to membranes (data not shown). The domains might be closely associated with the phosphodiesterase catalytic domains (X and Y) because when one or more of the RA domains are truncated, the resulting enzyme has very little phosphodiesterase activity, suggesting that these domains are necessary for proper folding of the enzyme (data not shown). This is similar to truncation of the ␦1 isoform of PLC, where deletion of even a few C-terminal residues leads to inactivation of phosphodiesterase activity. 3 Because of the potential interaction of the RA domains with the X and Y domains, the ability of activated Ras to regulate the phosphodiesterase activity of PLC⑀ was assessed by co-transfection. As can be seen from Fig.  7A, v-Ras does not stimulate but may inhibit PLC⑀ in vivo. Transfection of cells with PLC⑀ alone increased PI hydrolysis 3.5-fold, whereas co-transfection with PLC⑀ and v-Ras lead to an increase of only 2-fold. Expression levels of v-Ras and PLC⑀ did not vary substantially among different conditions. Actually, PLC⑀ expression was slightly greater when co-transfected with v-Ras. Overall, these results suggest a potential role for PLC⑀ in regulating Ras activation and thus activation of the MAP kinase pathway in a manner dependent upon the RasGEF domain and independent of PI hydrolysis.
PLC⑀ Activates Ras-More direct evidence of the ability of PLC⑀ to activate Ras comes from experiments in which GTP-Ras is trapped using the Ras effector Raf1. The minimal RBD of Raf1 (amino acids 51-131) binds very tightly and specifically to the GTP-bound form of Ras (K d 20 nM), whereas the affinity for RasGDP is 3 orders of magnitude lower (35). To determine whether PLC⑀ could indeed act as an exchange factor for Ras, TSA201 cells were transiently co-transfected with either pcDNA3 (control), PLC⑀, PLC⑀ H1144L, or RasGRF2 (as a GEF control for activation of Ras) and Ha-Ras. The cells were harvested after 48 h and RasGTP was identified by precipitation with GST-RBD and immunoblotting using anti-Ras antibody. Fig. 8A demonstrates that GST-RBD bound RasGTP was increased significantly in cells transfected with PLC⑀, PLC⑀ H1144L, and the known RasGEF, RasGRF2. The fold increase in RasGTP for PLC⑀, H1144L, and RasGRF was 4.2-, 9.3-, and 4.1-fold, respectively. Although levels of expression for PLC⑀ and the phosphodiesterase deficient mutant H1144L were very similar (Fig. 8B), H1144L was 2.2-fold more potent for activation of Ras. As discussed in the previous section, H1144L was also more potent in activating MAP kinase. The increase in H1144L activity suggests that the phosphodiesterase activity may serve as a negative feedback regulator of the RasGEF activity. Levels of expression for Ha-Ras were similar for each of the experimental conditions (Fig. 8B). There was no appreciable RasGTP detected in control cells. Basal levels of RasGTP were reduced significantly by preincubation of the cells in serum-free medium for 12 h. Taken together, these results provide more direct evidence that PLC⑀ indeed activates Ras by acting as a RasGEF.
This work demonstrates that PLC⑀ is a widely expressed unique PLC enzyme that constitutes a new PLC family possessing bifunctional activity, both PLC and RasGEF activity.
As such, PLC⑀ may mediate the effects of G protein-coupled receptors, especially those coupled with G␣ 12 /G␣ 13 through two divergent pathways involving phosphatidylinositol hydrolysis as well as direct activation of the Ras/MAP kinase pathway. Thus, this new member of the PLC family may play a vital role in transducing signals from the plasma membrane to the nucleus through multiple pathways to modulate cytoskeletal changes, cell growth, and mitogenesis.