A novel human phosphatidylethanolamine-binding protein resists tumor necrosis factor alpha-induced apoptosis by inhibiting mitogen-activated protein kinase pathway activation and phosphatidylethanolamine externalization.

The phosphatidylethanolamine (PE)-binding proteins (PEBPs) are an evolutionarily conserved family of proteins with pivotal biological functions. Here we describe the cloning and functional characterization of a novel family member, human phosphatidylethanolamine-binding protein 4 (hPEBP4). hPEBP4 is expressed in most human tissues and highly expressed in tumor cells. Its expression in tumor cells is further enhanced upon tumor necrosis factor (TNF) alpha treatment, whereas hPEBP4 normally co-localizes with lysosomes, TNFalpha stimulation triggers its transfer to the cell membrane, where it binds to Raf-1 and MEK1. L929 cells overexpressing hPEBP4 are resistant to both TNFalpha-induced ERK1/2, MEK1, and JNK activation and TNFalpha-mediated apoptosis. Co-precipitation and in vitro protein binding assay demonstrated that hPEBP4 interacts with Raf-1 and MEK1. A truncated form of hPEBP4, lacking the PE-binding domain, maintains lysosomal co-localization but has no effect on cellular responses to TNFalpha. Given that MCF-7 breast cancer cells expressed hPEBP4 at a high level, small interfering RNA was used to silence the expression of hPEBP4. We demonstrated that down-regulation of hPEBP4 expression sensitizes MCF-7 breast cancer cells to TNFalpha-induced apoptosis. hPEBP4 appears to promote cellular resistance to TNF-induced apoptosis by inhibiting activation of the Raf-1/MEK/ERK pathway, JNK, and PE externalization, and the conserved region of PE-binding domain appears to play a vital role in this biological activity of hPEBP4.

The phosphatidylethanolamine-binding protein (PEBP) 1 family consists of a number of 21-23-kDa basic proteins, first identified in bovine brain, with preferential in vitro affinity for phosphatidylethanolamine, a component of the cell membrane. This family is an evolutionarily conserved group found in species of flowering plants (Antirrhinum (1)), parasites (Plasmodium falciparium (2)), nematodes (Toxocara canis (3)), insects (Drosophila melanogaster (4)), and mammals, including cattle, monkeys, and humans (5). A number of functions have been suggested for the mammalian PEBP proteins, including lipid binding and inhibition of serine proteases (6). These proteins can also act as precursors for a bioactive peptide HCNP (hippocampal cholinergic neurostimulating peptide), important in hippocampus development (5). Plant PEBP homologues are involved in the control of a morphogenic switch between shoot growth and flower structures (7). Yeast two-hybrid screen analysis has shown that human PEBP1 (hPEBP1, also called Raf kinase inhibitory protein or RKIP) acts as a suppressor of Raf-1 kinase activity and mitogen-activated protein kinase signaling in fibroblasts via its ability to sequester and inactivate Raf-1 and MEK1 (8,9). Both Raf-1 and MEK bind to the highly conserved phosphatidylethanolamine-binding domain of hPEBP; hPEBP induces dissociation of Raf-1⅐MEK complexes and behaves as a competitive inhibitor of MEK phosphorylation. Mapping of the binding domains has shown that MEK and Raf-1 bind to overlapping sites in hPEBP1, whereas MEK and hPEBP1 associate with different domains in Raf-1, and Raf-1 and hPEBP1 bind to different sites in MEK. In addition, mPEBP2, a recently characterized novel testis-specific member of the phosphatidylethanolamine-binding protein family, colocalizes with members of the mitogen-activated protein kinase pathway in late spermatocytes and spermatids and on the midpiece of epididymalsperm (10).
Most aminophospholipids, including phosphatidylserine (PS) and, to a lesser extent, phosphatidylethanolamine (PE), are located in the inner leaflet of the plasma membrane (11,12). One of the earliest events during apoptosis is the loss of membrane phospholipid asymmetry, leading to the exposure of PS and PE on the outer leaflet (13), a phenomenon called "phospholipid flip-flop". This mechanism facilitates the recognition of apoptotic cells; Ro, a 19-amino acid tetracyclic polypeptide with a molecular mass of 2041 Da isolated from Streptoverti-cillium griseoverticillatum, can be used to monitor the transbilayer movement of PE in biological membranes during cell division and apoptosis (14 -18). PE is usually exposed on the surface of apoptotic cells (15, 19 -21), and the fluorescencelabeled Ro peptide, FL-SA-Ro, has proven to be a useful tool for detecting apoptotic cells via specific PE binding.
In the present study, we report the molecular cloning and characterization of a new member of the PEBP family derived from human bone marrow stromal cells (BMSCs). This protein contains a typical phosphatidylethanolamine-binding domain and has been designated as human phosphatidylethanolaminebinding protein 4 (hPEBP4). The ability of hPEBP4 to function as an anti-apoptotic molecule and the mechanisms that underlie hPEBP4 effects were investigated.

MATERIALS AND METHODS
Reagents and Cell Culture-All of the cells were grown in RPMI 1640 or Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal calf serum, 4.5 g/liter D-glucose, nonessential amino acids (100 M each), 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine at 37°C in a 5% CO 2 atmosphere.
Isolation of hPEBP4 cDNA-Full-length hPEBP4 cDNA was directly isolated from a human BMSC cDNA library by random sequencing as described previously (22). A plasmid cDNA library of pCMV SPORT6.0 vector (Invitrogen) was constructed using the Superscript plasmid system (Invitrogen) for cDNA synthesis. The full-length cDNA of clone HNC7B1 appeared to encode a protein with a typical phosphatidylethanolamine-binding domain and was hence designated as hPEBP4. The full-length sequence is available in the GenBank TM data base under accession number AY037148.
Isolation of Human Blood Cells-Buffy coats from healthy donors were obtained from the transfusion center of Changhai Hospital (Shanghai, China). Human peripheral blood mononuclear cells were prepared from buffy coats by Ficoll-Paque 1.077 (Sigma) density gradient centrifugation. Peripheral monocytes, CD4 ϩ T cells, CD8 ϩ T cells, and B cells were directly isolated from human peripheral blood mononuclear cells using anti-CD14, anti-CD3, anti-CD4, anti-CD8, and anti-CD19 monoclonal antibody-conjugated magnetic microbeads (Miltenyi Biotech), respectively. Some monocyte samples were stimulated with lipopolysaccharide (1 g/ml for 24 h), and hPEBP4 expression examined in all groups by RT-PCR, as described below. The purity of each cell population was generally 90 -95%.
RT-PCR and Northern Blot Analysis of hPEBP4 mRNA Expression-RT-PCR and Northern blotting were performed as described previously (23). Primers specific for hPEBP4 were 5Ј-GGGTTGGACAATGAG-GCTG-3Ј (sense) and 5Ј-TGTGCTTGGGCTCGCTGGC-3Ј (antisense). Northern blot filters containing human poly(A) ϩ RNA (2 g/lane) were purchased from Clontech. Full-length hPEBP4 cDNA was used as a template for probe synthesis for Northern blot analysis. The filters were hybridized with the 32 P-labeled hPEBP4 cDNA probe in ExpressionHyb hybridization solution (Clontech).
Recombinant Expression of hPEBP4-Glutathione S-Transferase Fusion Protein and Generation of Anti-hPEBP4 Polyclonal Antibody-Analysis of the primary amino acid sequences of the varied PEBP family members revealed that the N terminus is poorly conserved in this family. The N-terminal 99-amino acid section of hPEBP4 was thus expressed, and rabbit antibodies that specifically recognize the N terminus of hPEBP4 were generated. cDNA encoding the 99 N-terminal amino acids of hPEBP4 was cloned into pGEX-2T (Amersham Biosciences). The soluble glutathione S-transferase fusion protein was obtained under isopropyl ␤-D-thiogalactopyranoside induction (0.2 mM) at 37°C for 4 h with Escherichia coli strain BL21 as a host and purified by glutathione-Sepharose 4B affinity chromatography (Pierce). The purified fusion protein was used to immunize rabbits according to a conventional procedure, as described (24). anti-hPEBP4 serum was purified using protein A affinity chromatography (Pierce), and titration analysis was performed by Western blot.
Cell Transfection-The expression vectors hPEBP4-B and p75PEBP4-B were transiently transfected into L929 cells using Lipo-fectAMINE reagent (Invitrogen) for 40 h with pcDNA3.1/Myc-His (Ϫ)B as a mock control. After serum starvation for 24 h, the cells were exposed to 20 ng/ml TNF␣ for 10 min or the indicated concentration of TNF␣ overnight and then analyzed by Western blot or apoptosis assay. In some experiments, 1 g of hPEBP4-B, p75PEBP4-B, or control vector were co-transfected together with the Ras-B expression vector (1 g) into L929 cells, and the cells were harvested 48 h later for subsequent analysis.
Cytolocalization of hPEBP4 by Fluorescence Confocal Microscopy-HEK 293 cells transiently transfected with hPEBP4-GFP, p75PEBP4-GFP, or GFP control vector, growing on glass cover slides, were placed in 6-well plates, treated with 20 ng/ml TNF␣ (Sigma) for 10 min, and then incubated with LysoTracker Red DND-99 (Molecular Probes) for 15 min at room temperature in the dark. At the same time, GFP, hPEBP4-GFP, or p75PEBP4-GFP vector were co-transfected into some cells with either pDsRed-mem (encoding a membrane-localized form of red fluorescent protein), Raf-1-RFP, or MEK1-RFP. The samples were washed briefly in phosphate-buffered saline and fixed in 4% polyformaldehyde prior to observation with a fluorescence confocal microscopy (LSM confocal microscope; Carl Zeiss).
Co-precipitation and in Vitro Protein Binding Assay-36 h after transfection, transiently transfected L929 cells were serum-starved for 24 h, stimulated with 20 ng/ml TNF␣ for 10 min, harvested, and then solublized in lysis buffer (Cell Signaling). The cell lysates were mixed with Ni-NTA beads (Qiagen), gently agitated overnight at 4°C, and washed four times in lysis buffer containing 10 mM amidazole and 0.1% Triton X-100, and Western blotting was performed. For in vitro protein binding assay, the extract of 293HEK cells transiently transfected with hPEBP4-B or p75PEBP4-B for 48 h were mixed with Ni-NTA beads as described above. After incubation with cell lysis supernatant of MEK-1-FLAG or Raf-1-FLAG-transfected cells for 2-3 h at 4°C, the samples were washed four times with phosphate-buffered saline and analyzed by Western blot.
Construction of FL-SA-Ro-Ro 09-0198, kindly provided by Roche Applied Science, was biotinylated using Biotin (Long Arm) NHS (Vector Laboratories) (14). FL-SA (Vector Laboratories) was mixed with biotinylated Ro, and the resulting FL-SA-Ro complexes were purified by gel filtration, as described (16).
Apoptosis Assay-The cells were washed, resuspended in staining buffer, and stained with PI and either annexin V (ApoAlert annexin V apoptosis kit, Becton Dickinson) or R123 (R-302; Molecular Probes), according to the manufacturer's instructions, or FL-SA-Ro, as described (17). Stained cells were analyzed by fluorescence-activated cell sorter (FACSCalibur, Becton Dickinson).
hPEBP4 siRNA Assay-21-nucleotide sequences of hPEBP4 siRNA were synthesized by Proligo: 5Ј-GGAAAAGUCAUCUCUCUCCTT (sense) and 5Ј-GGAGAGAGAUGACUUUUCCTT (antisense). hPEBP4 mutation control siRNA oligonucleotides were 5Ј-GGAAAAUCUACU-CUCUCCTT (sense) and 5Ј-GGAGAGAGUAGACUUUUCCTT (antisense). For annealing, 20 M single-stranded 21-nucleotide RNAs were incubated in annealing buffer (100 mM potassium acetate, 30 mM HEPES-KOH at pH 7.4, 2 mM magnesium acetate) for 1 min at 90°C and then for 1 h at 37°C. siRNA duplexes were transfected into MCF-7 breast cancer cells using Oligofectamine reagent (Invitrogen), as described (25). All of the data shown in this article are representative of at least three independent experiments. bers of PEBP family from human, bovine, mouse, rat, monkey, D. melanogaster, and others ( Fig. 1). At the primary amino acid level, it shared the highest homology with an unnamed murine PEBP-like protein, BAB24810 (45% identity and 57% similarity). As such, the two appeared to represent interspecies orthologues, whereas the other PEBPs represented paralogues, exhibiting significantly lower sequence homology. Based on its sequence similarity with other PEBP family members, the conservation of PEBP features, and its putative capacity to bind phosphatidylethanolamine, the novel molecule was designated hPEBP4. The human PEBP4 cDNA corresponded to Unigene cluster Hs.352388, located on human chromosome 8p21.2.

Identification and Sequence Analysis of hPEBP4 -
Based on sequence analysis and existing PEBP structures, a number of features characteristic of the PEBP family, with likely functional importance, were identified in hPEBP4. As with all PEBP family members, human PEBP4 was predicted to contain a PE-binding domain, in this case between amino acids 84 and 191 (Fig. 1). Residues previously demonstrated to be conserved between all members of the family (Pro 97 , Asp 98 , Pro 100 , His 112 , and Arg 149 ) were also found in hPEBP4. These residues are thought to be involved in determining the local structure of a biologically important ligand-binding site in PEBPs (26). The two main regions of high sequence conservation among hPEBP4 and other PEBPs occurred between residues 91-117 and residues 141-153 (based on hPEBP4 numbering). Residues 91-101 (including cis-Pro 100 ) lined one side of the putative PE-binding site, with residues 96 -98 and 100 forming an Asp-Pro-Asp-Xaa-Pro motif that is universally conserved in all family members. The first aspartate of this motif (Asp 96 ), along with His 112 , lay at the base of the binding pocket. Arg 149 is part of a second highly conserved region running between residues 146 and 149, with the motif Gly-Xaa-His-Arg. The 141-153 conserved region formed part of the second side of the binding pocket, as well as a loop adjacent to this site. These structural features suggest that hPEBP4 might possess PE binding properties similar to those seen in other PEBPs, whose structures have been determined by x-ray crystallography.
The effects of TNF␣ on hPEBP4 mRNA expression in A549 and LoVo cells, which do not normally express hPEBP4, were examined by quantitative real time PCR analysis. Following TNF␣ stimulation, hPEBP4 expression increased in the cells, reaching a maximum at 12 h and then decreasing (Fig. 2E). The inducible, time-dependent, hPEBP4 expression pattern suggests that hPEBP4 might play roles in cellular responses to extracellular stimuli and in self-protection from damage.
Cytolocalization of hPEBP4 -The cytolocalization of GFPfused hPEBP4 protein in HEK293 cells was examined by fluorescence confocal microscopy. 48 h after transfection with hPEBP4-GFP (full-length) or p75PEBP-GFP (75 N-terminal hPEBP4 Inhibits TNF␣-induced Apoptosis residues of hPEBP4, lacking the PE-binding domain), GFP fluorescence displayed a cytoplasmic distribution in transfected cells, with signal particularly strong in the perinuclear region. Further subcellular localization analysis was performed using specific probes for cell organelles including mitochondria, endoplasmic reticulum, Golgi apparatus, and lysosomes. The majority of hPEBP4-GFP and p75PEBP-GFP green staining correlated with the red lysosome signal (LysoTracker), as displayed by yellow fluorescence in overlaid images (Fig.  3A), but did not correlate with red staining associated with mitochondria, endoplasmic reticulum, or Golgi apparatus (data not shown). Cells transfected only with GFP showed fluorescence in both cytoplasm and nucleus and no evidence of colocalization with lysosomes. The results suggest that hPEBP4 localizes primarily to lysosomes, in a PE-binding domain-independent manner.
TNF␣-induced hPEBP4 Translocation from Lysosome to Cell Membrane-Because TNF␣ could induce hPEBP4 expression, we examined its influence on hPEBP4 cytolocalization. hPEBP4-GFP, p75PEBP-GFP, or GFP were co-transfected into HEK293 cells together with pDsRed-mem, which encoded a membrane-localized form of red fluorescent protein. When transfected cells were stimulated with TNF␣ (20 ng/ml) for 10 min, hPEBP4-GFP translocated from lysosomes to the cell membrane (Fig. 3B). In contrast, p75PEBP-GFP, which lacked the PE-binding domain, did not demonstrate translocation. Expression of GFP, GFP-fused p75PEBP4, GFP-fused hPEBP4 proteins in transfected cells was confirmed by Western blot (ϳ27 kDa for GFP, 36 kDa for GFP-p75PEBP4, and 53 kDa for GFP-hPEBP4 fusion protein; Fig. 3C). These results provide evidence that hPEBP4 might be induced to translocate from lysosomes to the plasma membrane following cellular exposure to extracellular stimuli such as TNF␣ and that this translocation requires the PE-binding domain.
hPEBP4 Attenuates TNF␣-induced Activation of ERK1/2 and MEK1-Expression vectors encoding Myc epitope-and His 6tagged full-length and truncated forms of hPEBP4 (hPEBP4-B and p75PEBP4-B) were transfected into TNF␣-sensitive L929 cells. Lysates of transfected cells were resolved by SDS-PAGE and immunoblotted to confirm expression (Fig. 3C). Proteins of ϳ30 and 13 kDa were detected specifically by anti-Myc antibody in cells transfected with hPEBP4-B and p75PEBP4-B, respectively, but not in cells transfected with control vector. Phosphorylation of Raf-1, MEK1, and ERK1/2 were examined following 10 min of stimulation with TNF␣. As shown in Fig.  4A, TNF␣-induced phosphorylation of ERK-1/2 and MEK1 in hPEBP4 Inhibits TNF␣-induced Apoptosis L929 cells was blocked by overexpression of hPEBP4 but not by overexpression of PE-binding domain-deficient p75PEBP4. Also, when hPEBP4-B was co-transfected with the Ras expression vector Ras-B, Ras-mediated activation of MEK1 and ERK1/2 was also inhibited (Fig. 4B). In contrast, Raf-1 phosphorylation triggered by TNF␣ stimulation or co-transfection with Ras-B was not influenced by either form of hPEBP4.
hPEBP4 Associates with Raf-1 and MEK1-To investigate whether hPEBP4 could interact with endogenous Raf-1 and MEK1, lysates of L929 cells transiently transfected with hPEBP4-B, p75PEBP4-B, or control vector and then treated with TNF␣ were precipitated using Ni-NTA beads. The precipitates were resolved by SDS-PAGE and probed for both Raf-1 and MEK1 but not ERK1/2, which is downstream of MEK1 and unlikely to directly interact with hPEBP. As shown in Fig. 4C, hPEBP4 co-precipitated with Raf-1 and MEK1 after TNF␣ stimulation, whereas p75PEBP4 did not, indicating that TNF␣ prompts hPEBP4 to associate with Raf-1 and MEK1 and that this association requires the PE-binding domain of hPEBP4.
An in vitro protein binding assay was carried out to further demonstrate this interaction. Cell lysates of hPEBP4-B-, p75PEBP4-B-, or control vector-transfected L929 cells were immobilized on Ni-NTA-beads and incubated with cell extracts of HEK293 cells transfected with Raf-1-FLAG or MEK1-FLAG, respectively. Fig. 4D showed Western blot of protein-protein interaction. hPEBP4 specially bound to Raf-1 and MEK1. No binding was detected of p75PEBP4 and control. These results were further confirmed using confocal microscopy (Fig. 4E). hPEBP4 translocated from lysosomes to the membrane and co-localized with Raf-1 and MEK1 upon TNF␣ treatment, whereas p75PEBP4 did not. Given that the Ras/Raf-1/MEK/ ERK signaling pathway activated by TNF␣ was inhibited in hPEBP4-overexpressing cells, it appears that extracellular stimuli such as TNF␣ cause hPEBP4 to depart from its usual lysosomal location, traveling to a membrane proximal position where it binds to Raf-1 or MEK-1, dissociating the Raf-1⅐MEK complex and thus behaving as a competitive inhibitor of MEK phosphorylation. Notably, p75PEBP4, a truncated form lacking a PE-binding domain, had no effect on any of the above described events, save for co-localization with lysosomes, indicating that it is the PE-binding domain that binds Raf-1 and MEK-1 and exerts a negative effect on the Raf-1/MEK1/ERK cascade.
hPEBP4 Inhibits TNF␣-induced Cellular Apoptosis-Phospholipids such as PE and PS are located on the inner leaflet of the plasma membrane, and loss of membrane phospholipid asymmetry is one of the earliest events during apoptosis. PE, as well as PS, can be detected on the cell surface during the   FIG. 4. Effects of hPEBP4 overexpression on Ras/Raf-1/MEK1/ ERK1/2 signaling pathway and interaction of hPEBP4 with Raf-1 and MEK1. A, hPEBP4-B, p75PEBP4-B, or mock control vector transfected L929 cells were serum-starved for 24 h and then treated with 10 ng/ml TNF␣ for 10 min. Equivalent protein loadings of each lysate were immunoblotted with antibodies recognizing the phosphorylated, active forms of pp42/pp44 ERK, MEK1, or Raf-1. B, Ras-B was co-transfected together with hPEBP4-B, p75PEBP4-B, or mock control vector into L929 cells. 48 h after transfection, the cells were lysed and subjected to Western blot analysis with the same antibody as described in A. C, hPEBP4 binding to Raf-1 and MEK1 in vivo. Lysates of Hislabeled hPEBP transfected L929 cells stimulated with TNF␣ for 10 min were incubated with Ni-NTA beads. The beads were collected by centrifugation, and samples were immunoblotted using Raf-1 or MEK1 or anti-hPEBP4 antibody. D, hPEBP4 binding with Raf-1 and MEK1 in vitro. Lysates of hPEBP4-B or p75PEBP4-B transfected cells were immobilized with Ni-NTA-beads. After incubation with cell extracts of Raf-1-FLAG or MEK1-FLAG transfectants for 2-3 h at 4°C, agarose pellets were washed and subjected to SDS-PAGE. E, hPEBP4 co-localization with Raf-1 and MEK1 upon TNF␣ treatment. HEK 293 cells were transiently transfected with hPEBP4-GFP, p75PEBP4-GFP, or control GFP vector, together with Raf-1-RFP or MEK1-RFP vectors and then stimulated with 20 ng/ml TNF␣ for 10 min. Co-localization of hPEBP4-GFP (green) and Raf-1-RFP or MEK1-RFP (red) is demonstrated by yellow fluorescence in overlay images.
hPEBP4 Inhibits TNF␣-induced Apoptosis early stages of apoptosis, indicating a loss of asymmetric distribution of aminophospholipids. Because hPEBP4 contained a conserved PE-binding domain and translocated from lysosomes to the membrane upon TNF␣ treatment, we wondered whether it played a role in the process of cellular apoptosis. To investigate the possibility, L929 cells were transiently transfected with hPEBP4-B, p75PEBP4-B, or control vector, stimulated with the indicated concentrations of TNF␣ for 20 h and apoptosis assessed. Loss of mitochondrial inner transmembrane potential is also often associated with the early stages of apoptosis and may be one of the central features of the process (27). The green fluorescent cationic dye Rhodamine 123 (R-302) binds selectively to the inner mitochondrial membrane and accumulates in the charged membrane compartments of living cells; apoptotic cells exhibit a loss of R123 binding. As shown in Fig. 5A, hPEBP4-B transfectants did not differ in their baseline level of apoptosis, when compared with p75PEBP4-B or mock transfectants but did exhibit lower levels of apoptosis following TNF␣ treatment. Fluorescein isothiocyanate-conjugated annexin V/PI staining was also performed to detect apoptotic cells. Cells acquire annexin V-binding sites during apoptosis, providing another method for detecting cells undergoing apoptosis. Apoptotic cells are annexin V-positive (including PI Ϫ or PI ϩ ), whereas necrotic cells are annexin V-negative and PI-positive. As shown in Fig. 5B, the percentage of hPEBP4overexpressing L929 cells that were annexin V positive (apoptotic) was much less than that of p75PEBP4-overexpressing cells, mock control, or untransfected cells following exposure to TNF␣. The same result was obtained in human lung carcinoma cells A549 (data not shown). Together these results suggest that hPEBP4 might be involved in cellular resistance to TNF␣induced apoptosis and that the conserved region of PE-binding domain appears to play a vital role in this biological activity of hPEBP4.
Overexpression of hPEBP4 Inhibits TNF␣-induced Activation of JNK and PE Externalization-TNF receptor-associated factor 2 (TRAF2) in TNF signaling activates a mitogen-activated protein kinase kinase kinase, which then activates ERK or JNK. The roles of JNK activation in apoptosis are highly controversial, with reports suggesting pro-apoptotic, anti-apoptotic, and neutral roles (28 -31). Because hPEBP4 plays a negative role in TNF-induced apoptosis, its effects on TNFinduced JNK activation were tested. L929 cells were transiently transfected with mock, hPEBP4-B, or p75PEBP4-B vectors. After serum starvation for 24 h and stimulation with TNF␣ for various lengths of time, the phosphorylation levels of JNK were examined. As shown in Fig. 6A, TNF␣-induced phosphorylation of JNK in L929 cells was inhibited by overexpression of hPEBP4 but not by overexpression of p75PEBP4. However , no binding between hPEBP4 with TNFR1, TNFR2,  TRAF1, TRAF2, TRAF3, TRAF4, TRAF5, TRAF6, TRADD, or FADD was detected, as assayed by immunoprecipitation with Ni-NTA beads from lysates of L929 cells stably transfected with hPEBP4 vector (data not shown). Meanwhile, TNF-mediated activation of caspase-3 (Fig. 6B), caspase-8, and p38 (data not shown) were not influenced by hPEBP4 overexpression.

hPEBP4 Inhibits TNF␣-induced Apoptosis
Ro 09-0198 (cinnamycin), a tetracyclic peptide antibiotic that recognizes the structure of phosphatidylethanolamine in a highly specific manner, forming a tight equimolar complex with PE on biological membranes, has been used to monitor the transbilayer movement of PE in biological membranes during cell division and apoptosis. Because hPEBP4 had a PE-binding domain and translocates from lysosomes to the cell membrane after TNF␣ stimulation, we wondered whether it could associate with PE and inhibit its exposure, thus rendering the cell insensitive to apoptotic stimuli. We conjugated biotinylated Ro to FL-SA to study the molecular movement of PE in L929 cells, prepared as described for the apoptosis assay. As shown in Fig.  6C, no difference was observed among the groups in the absence of TNF␣. However, when cells were subjected to incubation with 20 ng/ml TNF␣ for 20 h, PE exposure was suppressed only by overexpression of hPEBP4. This result demonstrates that the insensitivity of hPEBP4 transfectants to TNF␣-induced apoptosis is at least partly due to inhibition of PE externalization.
Silencing of hPEBP4 Expression Sensitizes MCF-7 Breast Cancer Cells to TNF␣-induced Apoptosis-Overexpression of hPEBP4 in L929 cells showed that this protein represents a potent survival molecule. Given its high expression in breast cancer cells, the possibility that silencing hPEBP4 expression could sensitize MCF-7 breast cancer cells to apoptosis induction was investigated. siRNA was used to silence the expression of hPEBP4 protein in MCF-7 breast cancer cells, with a mutant siRNA duplex and no siRNA as controls. Western blot-ting and RT-PCR confirmed the complete silencing of hPEBP4 expression in MCF-7 cells (Fig. 7A). Annexin V/PI staining was used to detect apoptotic cells after exposure to 20 ng/ml TNF␣ for 20 h. As shown in Fig. 7B, hPEBP4-silenced MCF-7 cells were more sensitive to TNF␣-induced apoptosis. The result further demonstrates that hPEBP4 promotes cellular resistance to TNF␣-induced apoptosis.

DISCUSSION
The PEBP family is a highly conserved group of proteins with homologues in a wide variety of organisms. We report here the molecular cloning and characterization of a new member of the PEBP family, hPEBP4, from human BMSCs. hPEBP4 appears to promote cellular resistance to TNF-induced apoptosis by inhibiting activation of the Raf-1/MEK/ERK pathway, JNK, and PE externalization.
Sequence analysis revealed that hPEBP4 shares significant homology to other PEBP family members but that it also represents a distinct subset within the family. According to O'Bryan and colleagues (10), previously identified PEBPs can be grouped into three subfamilies; the majority of differences between these subfamilies are in the N-terminal residues (those preceding the PE-binding domain), especially the first 10 amino acids. Members of the three subfamilies share similar overall features, including length (each being ϳ190 residues) and amino acid composition in functional domains. In contrast, hPEBP4, and its closest homologue, the unnamed mouse protein BAB24810, differ in these aspects. They are both more hPEBP4 Inhibits TNF␣-induced Apoptosis than 220 residues in length and contain at least two insertions and one deletion in their protein sequence, when compared with proteins belonging to the three existing subfamilies. Therefore, hPEBP4 belongs to the fourth PEBP subfamily; this is represented in its nomenclature, hPEBP4. Although the N-terminal sequence of hPEBP4 varied notably from those of known PEBPs, and the two insertions (between residues 55 and 56 and residues 102 and 103) and the deletion (residues 131-134 of hPEBP numbering) were all located in regions identified as PE-binding structures, it seems that the overall functions of hPEBP4 were not affected because it could associate with Raf-1 and MEK1 and inhibit TNF␣-induced activation of the Ras/Raf-1/MEK/ERK signaling pathway, all PEBP family member attributes (8 -10).
Sequence analysis and preliminary homology modeling based on the crystal structure of bovine and human PEBPs (26,32) suggest that hPEBP4 also contains a putative PE-binding pocket, which could potentially allow anchoring of this protein to the inner leaflet of membranes. In addition, hPEBP4 has high sequence and predicted topology homology with analogous Raf-1-and MEK-binding regions predicted in deletion studies (9). In this article we observed that hPEBP4 was involved in the regulation of mitogen-activated protein kinase signaling. Overexpression of hPEBP4 suppressed signaling through this pathway, whereas down-regulation of hPEBP4 by siRNA relieved suppression; hPEBP4 did not inhibit Raf-1 activation but specifically interfered with the phosphorylation of MEK by Raf-1, thus inhibiting downstream ERK activation. We demonstrated that hPEBP4 associates with both MEK1 and Raf-1 using co-immunoprecipitation and in vitro protein binding assays. In accordance with this result, the recruitment of hPEBP4 from lysosomes to cell membrane and co-localization with Raf-1 or MEK1 following TNF␣ treatment was observed by confocal microscopy. Because the highly conserved PE-binding domain of other PEBPs has been proven to bind to Raf-1 and MEK1 (8,9), we constructed vectors expressing a truncated form of hPEBP4, p75PEBP4, which lacked a PE-binding domain, and found that it could not inhibit the Raf-1/MEK/ ERK pathway. p75PEBP4 was also unable to associate with either Raf-1 or MEK. These data are in agreement with the hypothesis put forward by Yeung et al. (8,9), derived from their studies on known PEBPs, suggesting that both Raf-1 and MEK bind to phosphatidylethanolamine-binding domains (accession number PS01220) and that PEBPs are targeted to the cell membrane following mitogenic stimulation.
PE and PS, located on the cytoplasmic surface of the cell membrane, play important roles in maintaining plasma membrane asymmetry. Translocation of both PE and PS from the inner to the outer leaflet of the plasma membrane precedes apoptosis, suggesting that redistribution of PS and PE may be an early symptom of apoptotic cell death. Cells overexpressing hPEBP4 were resistant to apoptosis induced by TNF␣, whereas silencing of hPEBP4 using siRNA sensitized the cells to apoptotic stimuli, consistent with the previous observation that TNF␣-resistant cKDH-8/11 cells had increased levels of hPEBP compared with TNF␣-sensitive KDH-8/YK cells (33). Using FL-SA-Ro, which stringently recognizes the structure of PE, we found that TNF␣-induced PE exposure on the surface of cells was significantly increased during hPEBP4 mRNA silencing and decreased during hPEBP4 overexpression. The data suggest that hPEBP4 is targeted to the cell membrane and binds to PE situated on the inner leaflet of the plasma membrane, FIG. 7. hPEBP4 RNA interference in MCF-7 breast cancer cells enhances TNF␣-induced apoptosis. A, functional silencing of endogenous hPEBP4 expression by RNA interference in MCF-7 cells. MCF-7 cells were transfected with hPEBP-specific siRNA (hPEBP4 siRNA) or mutated hPEBP4 siRNA control using Oligofectamine reagent. 48 h after transfection, the cells were harvested, and hPEBP4 expression was detected by Western blotting with anti-hPEBP4 antibody (upper panels) and RT-PCR analysis (lower panels). B, MCF-7 cells transfected with specific hPEBP4 siRNA or hPEBP4 mutation siRNA control were stained with annexin V and PI following incubation with TNF␣ for 20 h. The percentages of apoptotic cells in each sample are indicated.
The TNFR1 signaling complex is composed of the trimerized receptor, among which the FADD recruits and activates procaspase 8, initiating the apoptotic pathway, and the TNF receptor-associated factors 2 and 5 (TRAF2 and TRAF5) and the receptor-interacting protein are involved in the activation of JNK (34,35). The contribution of JNK activation to apoptosis depends on both cell type and stimulus (35), and its precise role in TNF-induced apoptosis is unclear (36,37). JNK has recently been shown to positively regulate TNF-induced apoptosis (38). We found that up-regulation of hPEBP4 expression in L929 cells inhibited TNF␣-induced ERK and JNK activation, as well as apoptosis. However, hPEBP4 had no effect on p38, caspase-3, or caspase-8 activation, and did not bind to TNFR1, TNFR2, TRAF1, TRAF2, TRAF3, TRAF4, TRAF5, TRAF6, TRADD, or FADD (data not shown). hPEBP4-mediated suppression of ERK and JNK activation thus appears to be partially responsible for the observed inhibition of TNF␣-induced apoptosis; further investigation is required to identify other mechanisms by which hPEBP promotes apoptosis resistance.
In sum, we have cloned and characterized hPEBP4, a novel member of the PEBP family that potentially functions as a survival-enhancing molecule, inhibiting TNF␣-induced apoptosis by interfering with Ras/Raf/MEK/ERK signaling, JNK activation, and PE externalization via its phosphatidylethanolamine-binding domain.