Interaction of Elongation Factor-1α and Pleckstrin Homology Domain of Phospholipase C-γ1 with Activating Its Activity

From the Department of Life Science, College of Natural Science, Daejin University, Kyeonggido 487-711, Korea, Department of Immunology, College of Medicine, Keimyung University, Taegu 700-712, Korea, Department of Physiology, College of Medicine, The Catholic University of Korea, Seoul 137-701, Korea, Department of Biochemistry, College of Medicine, Yeungnam University, Taegu 705-717, Korea, Department of Biological Science, College of Natural Science, Myongji University, Kyeonggido 449-728, Korea, Mitsubishi Kasei Institute of Life Sciences, Machida-shi, Tokyo 194, Japan, Institute of Biological Science, University of Tsukuba, Ibaraki 305-8572, Japan, In2Gen, Cancer Research Institute, Seoul National University, College of Medicine, Seoul 110-799, Korea, and Department of Life Science, Pohang University of Science and Technology, Kyungbuk 790-784, Korea

Many extracellular signals stimulate the hydrolysis of PIP 2 by the activation of PLC-␥1, which produces inositol 1,4,5trisphosphate (IP 3 ) and diacylglycerol. Both second messengers regulate the release of Ca 2ϩ from intracellular stores and activate protein kinase C, respectively (1,2). On the roles of PLC-␥1 in cell growth and differentiation, recent findings demonstrate that overexpression of PLC-␥1 induces malignant transformation in nude mice (3), and targeted deletion of PLC-␥1 results in embryonic lethality in mice (4).
The PH domain is a 120-amino acid residue stretch that has been identified in over 100 proteins (9 -12). The PH domain binds with high specificity and affinity to phosphoinositides including PIP, PIP 2 , and IP 3 (13)(14)(15). The PH domain of signaling molecules is involved in targeted translocation of molecules to cell membranes (13,16,17). Also, the PH domain mediates protein-protein interaction as well as protein-lipid interaction including the ␤␥-subunit of heteromeric G-protein (18,19), protein kinase C (20), actin (21), and BAP-135 (22). By analyzing the tertiary structure, the PH domain is an antiparallel ␤-sheet consisting of seven strands (23,24).
PLC-␥1 has two putative PH domains: one is located in the amino-terminal 150-amino acid residue, and the other is split by the SH2-SH2-SH3 domain (see Fig. 1A). Upon growth factor stimulation, the NH 2 -terminal PH domain of PLC-␥1 is targeted to the plasma membrane and binds to phosphatidylinositol 3,4,5-trisphosphate (PIP 3 ) but not to PIP 2 (25). In an effort to identify the PH domain ligands and to understand the phosphoinositide regulation mechanism of PLC-␥1, we used the GST⅐PH fusion protein system. We found that a split half of the PH domain of PLC-␥1 directly binds to EF-1␣, which is known for PI-4 kinase activating protein in plants (26,27).
Since the first step of PIP 2 biosynthesis is the phosphorylation of PI by PI-4 kinase to produce PI-4-P, which is phosphorylated by PI-4-P-5 kinase (PI-4-P kinase) to produce PIP 2 , it is probably true that EF-1␣, as a PI-4 kinase activator, has a pivotal role in regulating phospholipid metabolism. There is, however, no report on the roles of EF-1␣ as a PI-4 kinase activator in mammalian cells, so our present data are the first demonstration of the roles of eukaryotic EF-1␣ in mammalian phosphoinositide metabolism. In addition to the involvement of protein translation (28,29), EF-1␣ is involved in cytoskeletal rearrangement (30). Furthermore, overexpression of EF-1␣ correlates with metastasis (31) and leads to increased susceptibility to oncogenic transformation (32). Here we describe that EF-1␣ directly binds to the PH domain to activate PLC-␥1.

EXPERIMENTAL PROCEDURES
Reagents-Anti-EF-1␣ monoclonal antibody and horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody were purchased from Upstate Biotechnology Inc. (Lake Placid, NY). Anti-GST antibody and HRP-conjugated donkey anti-goat antibody were from Amersham Biosciences and Jackson ImmunoResearch Laboratories (West Grove, PA), respectively. All the phospholipids including phosphatidylethanolamine (PE), PIP, and PIP 2 were from Sigma. Lysylendopeptidase AP-1 was obtained from Wako Pure Chemical (Osaka, Japan).
Determination of the Partial Amino Acid Sequence-GST⅐PH-bound proteins were separated by 10% SDS-PAGE and stained with Coomassie Brilliant Blue R-250. The prominent band was excised and digested with lysylendopeptidase AP-1 for 14 h, and the resulting peptides were separated by reverse-phase high pressure liquid chromatography C 8 column chromatography as described previously (35). Amino acids from the NH 2 terminus of the peptides were analyzed by a pulse-liquid phase protein sequencer (PE-Biosystems, model 492 cLC).
PLC-␥1 Activity Assay-PLC-␥1 activity was measured as described previously (36). Briefly, substrate was prepared as sonicated vesicles of 75 mM [ 3 H]PIP 2 (9,000 -10,000 cpm/assay, PerkinElmer Life Sciences) and 750 mM PE in 50 mM HEPES buffer (pH 7.0) containing 2 mM CaCl 2 . Reactions were performed for 20 min at 30°C in a 60-l final volume and terminated by the addition of 1 ml of chloroform/methanol/ HCl (50:50:0.3) and 0.45 ml of 1 N HCl. The mixtures were vortexed and centrifuged for 10 min at 2,000 rpm. The aqueous phase containing [ 3 H]IP 3 was collected and subjected to a scintillation counter. The effect of EF-1␣ was examined by adding the indicated amount of EF-1␣ to the PLC-␥1 assay mixture. Tetrahymena pyriformis EF-1␣ was homogeneously purified by the method described before (37). Wild type PLC-␥1 and its mutant form (Y509A/F510A) were homogeneously prepared as described previously (38).
Far Western Blot Analysis-Purified EF-1␣ (0.2 g/lane) was resolved in 10% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane. Nonspecific binding to the membrane was blocked by adding 2% skim milk in Tris-buffered Tween 20 (TBT) for 1 h at room temperature. The membranes were then incubated with GST, GST⅐nPH 2 , or GST⅐nPH 2 ⅐Y509A/F510A mutant proteins (0.5 g/ml) in blocking buffer for 14 h at 4°C. After washes in TBT buffer, the membranes were incubated with anti-GST antibody for 2 h at room temperature. After washing the membrane with TBT buffer, bound proteins were detected by successive incubation with HRP-conjugated anti-goat antibody as a second antibody using the ECL detection system.
Dot-blot Analysis-The ability of the proteins to bind different phospholipids was examined using Dot-blot analysis (39). Briefly, chloroform-solubilized phospholipids (3 g of each) were spotted onto nitrocellulose membrane (PROTRAN, Schleicher & Schuell), and then the membrane was dried at room temperature for 1 h. The following steps are exactly the same method as for Far Western blotting. The membrane was blocked with 2% non-fat skim milk in TBT buffer for 1 h. The membranes were then incubated with purified EF-1␣, GST, GST⅐EF-1␣,

FIG. 1. Isolation of EF-1␣ as a split PH domain of PLC-␥1 binding protein.
A, PLC-␥1 has two putative PH domains (PH 1 and split PH 2 ) in addition to the SH2n, SH2c, SH3, and catalytic X and Y domains. A split PH domain consists of NH 2 -terminal portion (nPH 2 ) and COOH-terminal portion (cPH 2 ). B, three GST fusion proteins of GST⅐PH 1 , GST⅐nPH 2 , and GST⅐cPH 2 incubated with (ϩ) or without (Ϫ) NIH 3T3 cell lysates. The bound proteins were isolated by pull-down and subjected to 10% SDS-PAGE. A prominent protein with 48 kDa (indicated by an open arrowhead) was detected from GST⅐nPH 2 fusion protein followed by Coomassie Brilliant Blue staining. C, various GST fusion proteins incubated with NIH 3T3 cell lysates. The bound proteins were resolved on 10% SDS-PAGE followed by immunoblotting using anti-EF-1␣ monoclonal antibody. WCL indicates the whole cell lysates used for each pull-down experiment. GST⅐nPH 2 , GST⅐mutant (Y509A/F510A) nPH 2 proteins (0.5 g/ml), respectively, in blocking buffer for 14 h at 4°C. After washes with TBT buffer, the membranes were incubated with anti-EF-1␣ monoclonal antibody and anti-GST antibody for 2 h at room temperature, respectively. After washing the membrane again with TBT buffer extensively, bound proteins were detected by successive incubation with HRP-conjugated anti-mouse for EF-1␣ and anti-goat antibody as second antibody for GST fusion proteins using the ECL detection system, respectively.

RESULTS
A Split PH Domain of PLC-␥1 Directly Binds to EF-1␣-PLC-␥1 has two putative PH domains in the molecule. To search for proteins that specifically bind to the PH domains of PLC-␥1, we prepared three kinds of GST⅐PH domain fusion proteins in Escherichia coli (GST⅐PH 1 , GST⅐nPH 2 , and GST⅐cPH 2 ). These purified GST⅐PH fusion proteins were incubated with lysate of NIH 3T3 cells, respectively (Fig. 1, A and  B). Among them, GST⅐nPH 2 fusion proteins specifically pulled down a prominent protein with a molecular size of 48 kDa. To identify the protein, the band was cut from the gel and subjected to a protein sequencer after lysylendopeptidase digestion. We obtained two peptide sequences, P1 and P2. P1 is YYVTIIDAPGHRDFIK, and P2 is TGHLIYK. When these sequences were searched by the NCBI data base of SWISS-PLOT, they were found to match 58 species of EF-1␣ sequence reported. P1 and P2 correspond to the 105-120th and 24 -30th amino acids of human EF-1␣, respectively. Then we confirmed the band with 48 kDa as EF-1␣ with Western immunoblotting using anti-EF-1␣ monoclonal antibody (Fig. 1C). To further clarify the binding region of PLC-␥1 to EF-1␣, we examined binding capacity using several other GST fusion proteins including GST⅐PH 1 , GST⅐cPH 2 , GST⅐SH2, GST⅐SH3, GST⅐PH⅐GAP, and PH⅐GAP. As shown in Fig. 1C, only GST⅐nPH 2 associates with EF-1␣, as judged by Western immunoblotting. We next tested whether the binding is direct or not; Far Western blotting using purified protozoan T. pyriformis EF-1␣ (tEF-1␣) was used for this. Since EF-1␣ is highly conserved and has very similar biochemical properties among different species in eukaryotes (28), we used tEF-1␣ due to its success with purification steps with high purity (37). The result of Far Western blotting clearly showed a direct binding between GST⅐nPH 2 and tEF-1␣ (Fig. 2, A and B). Moreover, a double point mutation in GST⅐nPH 2 fusion protein (Y509A/F510A) lost its binding affinity to tEF-1␣ (Fig. 3, A and B). To confirm the interaction between the PH domain and EF-1␣ in vivo, immunoprecipitation was carried out to detect a PLC-␥1⅐EF-1␣ complex in COS-7 cells. The immunoprecipitates of EF-1␣ isolated by anti-EF-1␣ antibody included PLC-␥1, detected by Western immunoblot by anti-PLC-␥1 antibody or vice versa (Fig. 2C). Also, the yeast two-hybrid assay was introduced to show in vivo interaction between EF-1␣ and the nPH 2 domain of PLC-␥1 (Fig.  2D). The mutant nPH 2 domain did not bind to EF-1␣ in either yeast two-hybrid assay. These results clearly demonstrate that the nPH 2 domain of PLC-␥1 directly binds to EF-1␣.
␤2-Sheet of nPH 2 Domain Is Critical for Binding to EF-1␣-Fine mapping of the EF-1␣ binding site was carried out within the nPH 2 domain of PLC-␥1. Since aromatic residue has a potential for protein-protein interaction via hydrophobic interaction, we substituted aromatic residues including tyrosine and phenylalanine for alanine within the ␤2and ␤3-sheet (Fig.  3A). Several mutants of the nPH 2 domain induced by sitedirected mutagenesis were expressed as GST fusion proteins, mixed with lysates of NIH 3T3 cells, pulled down, and sub-FIG. 2. nPH 2 domain of PLC-␥1 directly binds to EF-1␣. A, GST⅐nPH 2 fusion protein was incubated with purified Tetrahymena EF-1␣ (tEF-1␣) in Nonidet P-40 buffer containing 1% bovine serum albumin. The bound tEF-1␣ was subjected to immunoblotting with anti-EF-1␣ antibody. B, purified tEF-1␣ (0.2 g/lane) was subjected to 10% SDS-PAGE. The protein was then transferred to a nylon membrane and probed with either anti-EF-1␣ antibody (left, Western blot) or purified GST, GST⅐nPH 2 and GST⅐nPH 2 ⅐Y509A/F510A proteins (right, Far Western blot), respectively. Left panel, the filter was probed with HRP-conjugated goat anti-mouse antibody. Right panel, the filter was incubated with anti-GST antibody followed by HRP-conjugated donkey anti-goat antibody. C, immunoprecipitation (IP) analysis. COS-7 cell lysate was immunoprecipitated using anti-EF-1␣ antibody and anti-PLC-␥1 antibody (F-7), and the immunoprecipitates were subjected to immunoblotting with anti-PLC-␥1 antibody (upper) or anti-EF-1␣ antibody (lower). WCL, the whole cell lysates used for each pull-down experiment. D, yeast two-hybrid assay. cDNAs from GST⅐nPH 2 and GST⅐nPH 2 ⅐Y509A/F510A, respectively, were inserted into the EcoRI/ XhoI site of the pGilda LexA vector (CLONTECH). A full-length EF-1␣ cDNA was inserted into the EcoRI/XhoI site of the pB42AD (activation domain) vector (CLONTECH). The yeast two-hybrid assay was carried out according to the manufacturer's specifications. jected to immunoblotting with anti-EF-1␣ antibody. The results demonstrate that amino acid residues Tyr-509 and Phe-510 of the ␤2-sheet play critical roles in the interaction between EF-1␣ and the PH domain of PLC-␥1 (Fig. 3B). A double point mutant Y509A/F510A completely abolishes the interaction, and another double point mutant Y506A/P507A of the ␤2-sheet of PLC-␥1 also shows reduced binding affinity to EF-1␣. However, the truncated mutant GST⅐tnPH 2 (amino acids 495-547) that lacks a portion of the ␤1-sheet has an equal binding capacity to that of nPH 2 (amino acids 477-547) (data not shown). These results indicate that the association between EF-1␣ and the nPH 2 domain of PLC-␥1 is due to hydrophobic interaction via the ␤2-sheet of PLC-␥1. nPH 2 Domain of PLC-␥1 Specifically Binds to PIP and PIP 2 -To explore the binding region of PLC-␥1 to phosphoinositides, we used GST⅐nPH 2 , mutant GST⅐nPH 2 ⅐Y509A/ F510A, GST⅐EF-1␣, and GST proteins for dot-blotting (lipidprotein blotting). Different lipids (each 3 g) including PE, PIP, and PIP 2 were spotted onto nitrocellulose membrane, and the membrane was blotted as described under "Experimental Procedures." As shown in Fig. 4, a double point mutant of GST⅐nPH 2 ⅐Y509A/F510A without binding affinity for EF-1␣ binds to PIP and PIP 2 with a similar capacity as wild type GST⅐nPH 2 , whereas GST⅐EF-1␣ and GST as a control do not show any binding capacity to phospholipids. Also, purified tEF-1␣ did not show any binding affinity to phospholipids (data not shown). These results suggest that the nPH 2 domain of PLC-␥1 has different binding sites for EF-1␣ and phospholipid. It is noteworthy that the nPH 2 region serves as a substrate PIP 2 -binding site, whereas the NH 2 -terminal PH domain (PH 1 ) of PLC-␥1 has been reported to interact with PIP 3 for membrane-targeted translocation (25).

PIP 2 Potentiates the Binding Affinity of PH Domain to
EF-1␣-Since both substrate PIP 2 and EF-1␣ bind to the nPH 2 domain of PLC-␥1, we investigated whether they compete with each other for binding to the PH domain. However, we found that the association between the nPH 2 domain of PLC-␥1 and EF-1␣ significantly increased in the presence of PIP 2 but not in either PE or PIP up to its concentration of 100 g/ml phospholipid (Fig. 5). Complex formation of GST⅐nPH 2 ⅐EF-1␣ increased in a PIP 2 dose-dependent manner.
EF-1␣ Activates PLC-␥1 Activity-To examine whether the complex formation of both proteins affects PLC-␥1 enzymatic activity, we measured its catalytic activity. Since we confirmed that tEF-1␣ specifically associates with PLC-␥1 (Fig. 2, A and  B), we used tEF-1␣ for its effect on PLC-␥1 activity. After preincubation of the purified tEF-1␣ with either PLC-␥1 or mutant PLC-␥1 (Y509A/F510A) at 4°C for 1 h, [ 3 H]PIP 2 hydrolyzing activity was measured. As shown in Fig. 6, EF-1␣ activates wild type PLC-␥1 activity in a bell-shaped manner, whereas mutant PLC-␥1 (Y509A/F510A) shows basal level activity even in the presence of EF-1␣. The activity of wild type PLC-␥1 is accelerated about 3-fold under the condition of 1:2 molar ratio (PLC-␥1 to EF-1␣). DISCUSSION Many reports have described the activation mechanisms and functional roles of PLC-␥1 in cellular signaling. Generally, growth factor stimulation leads to the binding of the SH2 domains of PLC-␥1 to the autophosphorylated receptor, and then PLC-␥1 is subsequently activated by tyrosine phosphorylation of Tyr-783 followed by PIP 2 hydrolysis to IP 3 and diacylglycerol. However, the degree of tyrosine phosphorylation of PLC-␥1 does not correlate well with its enzyme activity. For example, some ligands strongly stimulate tyrosine phosphorylation of PLC-␥1 with low IP 3 production (6), and some ligands highly induce the production of IP 3 with weak tyrosine phosphorylation of PLC-␥1 (7). There might be alternative controlling mechanism(s) of PLC-␥1 activity in cellular signaling. To identify a responsible molecule(s) for regulation of PLC-␥1, we have searched for binding proteins to PLC-␥1 and found EF-1␣, a PI-4 kinase activator.
We showed that an NH 2 -terminal split PH domain of PLC-␥1 specifically binds to EF-1␣ and that PIP 2 , a substrate of PLC-␥1, increases its association with EF-1␣. It is noteworthy that the strict region of PLC-␥1 plays a role for protein-protein interaction other than the SH2 and SH3 domains in PLC-␥1.
Although extensive studies on the role of the PH domains were done in PLC-␤ (17), -␦ (41,42), and -␥1 (25) and PI-4 kinase (39), those of a split PH domain of PLC-␥1 had not been examined. By pull-down experiments with GST⅐nPH 2 using a detergent lysate of NIH 3T3 cells, EF-1␣ was identified by peptide sequence analysis. The association between the nPH 2 domain and EF-1␣ is highly specific. Since EF-1␣ has been reported to be an activating protein of PI-4 kinase (26,27), it is meaningful that the association might play a critical role for PLC-␥1 in cellular signaling.
The addition of PIP 2 , a PLC-␥1 substrate, to the incubation mixture of the GST⅐nPH 2 domain fusion protein and cell extracts containing EF-1␣, the complex formation of GST⅐nPH 2 ⅐ EF-1␣, was dramatically increased in a PIP 2 concentration-dependent manner. On this point, it is interesting that the stable complex between PLC-␥1 and its substrate PIP 2 was detected by Dot-blot analysis (Fig. 4). Therefore, we can speculate that the PH domain of PLC-␥1 associates with PIP 2 first, and the PH domain/PIP 2 complex formation induces the conformational change to allow EF-1␣ to bind PLC-␥1. EF-1␣ binding to PLC-␥1 might facilitate the hydrolysis of PIP 2 by PLC-␥1. In this context, the role of PLC-␥1-bound EF-1␣ is a possible regulator for PIP 2 hydrolysis.
The activation of PLC-␥1 activity by EF-1␣ showed a bellshaped curve (Fig. 6). The maximum activity was at around a 1:2 molar ratio, whereas the activity decreased to basal level at higher than 1:8 molar ratios. There might be several reasons to explain the bell-shaped curve. One is that EF-1␣ has a very basic isoelectric point and is easily aggregated at high density (43,44). Another possibility is that component(s) such as Ca 2ϩ / CaM molecules are contaminated in EF-1␣ preparation. Although the preparation of tEF-1␣ is highly pure, the contamination of Ca 2ϩ /CaM molecules or other components could not be completely excluded. Generally, EF-1␣ preparation contains Ca 2ϩ /CaM molecules to some extent (40,45). In this regard, Ca 2ϩ /CaM might sequester Ca 2ϩ supplements for maximal PLC-␥1 activity at high doses of EF-1␣ addition.
EF-1␣ promotes the production of PIP and PIP 2 by the activation of PI-4 kinase, and eventually this newly produced PIP 2 hydrolysis is also accelerated by EF-1␣ via PLC-␥1 activation. EF-1␣ activates both PI-4 kinase (26, 27) and PLC-␥1, which FIG. 5. PIP 2 increases binding activity of GST⅐nPH 2 domain to EF-1␣. A, GST⅐nPH 2 (4 g)-coupled beads incubated with NIH 3T3 cell lysate in the presence of 0, 10, 50, or 100 g/ml sonicated phospholipid vesicles. Each resulting bead was divided into two equal portions for measuring the bound EF-1␣ (upper) and for normalization by amounts of GST⅐nPH 2 proteins (lower), respectively. Both proteins were accessed by immunoblotting with anti-EF-1␣ antibody or by anti-GST antibody followed by the second antibody. B, amounts of bound EF-1␣ expressed by relative image density (Quantity One, Bio-Rad), which was normalized by GST⅐nPH 2 . The experiments were carried out three times with similar results. can bind to PIP and PIP 2 . However, they regulate the level of PIP and PIP 2 in a different manner. The former regulates the level of PIP and PIP 2 by phosphorylation of PI at the D-4 position of the inositol ring, whereas the latter regulates the phospholipid level by hydrolyzing PIP 2 to IP 3 and diacylglycerol. Therefore, EF-1␣ has the potential to induce a rapid PI turnover in a cell.
In vivo, the complex of PLC-␥1⅐EF-1␣ was detected by immunoprecipitation not only from quiescent cells but also from epidermal growth factor-or platelet-derived growth factorstimulated cells, and so far no significant difference was observed between them. However, using green fluorescent protein fusion proteins, serum, and lysophosphatidic acid increased their complex formation around the cell membrane. 2 Although we need more detailed analysis for the PLC-␥1 activation mechanism, our results show a direct interaction between PLC-␥1 and EF-1␣ that elucidates the phospholipid metabolisms induced by PLC-␥1 in cellular signaling.