A novel group IIA phospholipase A2 interacts with v-Src oncoprotein from RSV-transformed hamster cells.

We have isolated a novel isoform of phospholipase A(2). This enzyme was designated srPLA(2) because it was discovered while analyzing the proteins interacting with different forms of the v-Src oncoproteins isolated from Rous sarcoma virus-transformed hamster cells. It contains all the functional regions of the PLA(2) group IIA proteins but differs at its C-terminal end where there is an additional stretch of 8 amino acids. The SrPLA(2) isoform was detected as a 17-kDa precursor in cells and as a mature 14-kDa form secreted in culture medium. A direct interaction of the 17-kDa precursor with the Src protein was observed in lysates of transformed cells. Both the 17- and 14-kDa forms were found to be phosphorylated on tyrosine. To our knowledge, this is the first report of a PLA(2) group II protein that is tyrosine phosphorylated. We surmise that srPLA(2) interacts with the Src protein at the cell membrane during the process of its maturation.

Previous work in our laboratories has focused on several transformed hamster cell lines named HET-SR, HET-SR-1, and HET-SR-8. Those lines were isolated as a result of independent infection of primary hamster fibroblasts with different stocks of Rous sarcoma virus (RSV). 1 All lines displayed high EMA (experimental metastatic activity) when injected intravenously in adult hamsters, but differed remarkably in SMA (spontaneous metastatic activity) when injected subcutaneously. HET-SR-1 and HET-SR-8 cells were highly metastatic (HM) and HET-SR cells were low metastatic (LM) (1). Each line has one integrated RSV provirus and express the Src-specific tyrosine kinase (2,3). Those lines secrete prostaglandin E (PgE), which inactivates the cytotoxic activity of NK cells (1). The production of PgE is in strict correlation with v-src gene expression, with tumor growth, and with the EMA potential of the transformed cells (3,4).
src genes from both cell lines were cloned and sequenced. It was found that v-src genes in both HM and LM cells had significant structural changes that were not previously observed in other members of the src family. The most interesting feature was that v-SrcHM and v-SrcLM differed from each other by several amino acids. Cells transformed with retroviral vectors carrying srcHM and srcLM display differences in morphology and in vivo metastatic activities, and we showed that structural changes in the C-terminal region of the v-Src proteins account for these differences in the metastatic properties of the corresponding transformed cells (5).
To assess the molecular mechanisms underlying the v-srcHMdependent high metastatic activity (or lack of such an activity) in v-srcLM-transformed cells, we tried to identify the products of cellular genes associated with the unique C-terminal regions of both v-Src variants, using a two-hybrid approach and a cDNA library made from low metastatic parental hamster fibroblasts. Several proteins interacting with v-Src via its Cterminal fragments were identified that were not characterized previously as v-Src protein partners.
We describe here one of these proteins related to the IIA group of phospholipase A 2 enzymes, srPLA 2 (Src-associated phospholipase A 2 ). This protein contains all the conserved functional residues typical for IIA PLA 2 enzymes and, in addition, 8 amino acids at its C terminus. In vitro, srPLA 2 protein interacts with v-SrcHM as well as with v-SrcLM oncoproteins. The 17-kDa precursor version of srPLA2 was identified by co-immunoprecipitation in lysates of RSV-transformed cells in a complex with the v-Src protein; it was found to be phosphorylated on a tyrosine.

EXPERIMENTAL PROCEDURES
Cell Cultures-Spontaneously transformed Syrian hamster fibroblasts and Syrian hamster embryo fibroblasts transformed independently in vitro by different stocks of RSV SR-D (from the Russian Cancer Research Center viral collection) were used (1). RSV-transformed lines had a typical transformed phenotype, were highly tumorigenic in syngenic hamsters, but differed in their SMA levels: HET-SR cells, low metastatic cell line; HET-SR-1 and HET-SR-8, high metastatic cell lines; HET-SR-2SC, high metastatic line selected from metastatic nodule in vivo. Transfectants HET-SR-N-Ras cl.6 and HET-SR N-Ras cl.34 lines were obtained by supertransfection of HET-SR and HET-SR-2SC, respectively with an activated N-ras gene mutated in the 12th codon (Gly3 Asp). Transfection of the cells by the N-ras oncogene was shown to induce the suppression of v-Src activity, coinciding with the decrease of prostaglandin secretion (4,6). NIH3T3-c-Src cells expressing the c-src gene were obtained after NIH3T3 cell transfection with a vector containing avian c-src (7). Cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 g/ml gentamycin, and 2 mM glutamine in the presence of 5% CO 2 at 37°C.
Hamster cDNA Fusion Library-mRNA was isolated from Syrian hamster fibroblasts HET-SR cell line using an RNA extraction kit * This work was supported by INTAS Grant 96-1292 and by Russian National Foundation for Basic Research Grant 98-04-48504. 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 (Amersham Pharmacia Biotech). The cDNA was synthesized using TimeSaver cDNA synthesis kit (Amersham Pharmacia Biotech) and cloned into a pGAD3S2X vector, a derivative of pGAD-GH with a modified polylinker. The complexity of the library was 1.3 ϫ 10 6 with 75% of the clones containing an insert (8).
Sequence Analysis-cDNA inserts from selected yeast clones were sequenced using custom primers. Sequences were analyzed using the Sequenase sequencing kit (Amersham Pharmacia Biotech) or on an automatic sequencer (Applied Biosystems Inc., model 373A). Sequences were analyzed in GenBank TM /EBI using the Fasta 3.0t76 search program.
Northern Blot Hybridization-Total RNA was extracted from different hamster fibroblast cell lines by the guanidinium thiocyanate/phenol method (14), separated by electrophoresis in 1% agarose/formamide gel in 40 mM MOPS, 10 mM sodium acetate, and 10 mM EDTA, pH 7.0, and transferred to Hybond N ϩ membrane (Amersham Pharmacia Biotech). Hybridization was performed using ␣-32 P-labeled DNA of srPLA 2 or v-src as described previously (15). Following autoradiography, blots were stripped by heating them at 90°C for 2 h in 20 mM Tris-HCl, pH 8.0 containing 0.2 mM EDTA and 1.0% SDS and rehybridized for control with glyceraldehyde-3-phosphate dehydrogenase end-labeled with [␥-32 P]ATP.
Immunoprecipitation-The hamster cell lysates (1-2 mg of total protein) were obtained as mentioned above. To accumulate proteins in their tyrosine phosphorylated form (for immunoprecipitation with antiphosphotyrosine antibodies), the cells were incubated before lysis in Dulbecco's modified Eagle's medium with 100 M Na 3 VO 4 . Samples were precleared with normal rabbit antiserum and incubated with either anti-Src polyclonal Ab (Santa Cruz Biotechnology) or srPLA 2 antiserum or anti-phosphotyrosine mAb PY20 (Signal Transduction Biotechnology) at 4°C overnight. They were further incubated with protein A-Sepharose or protein G-Sepharose beads for 1 h. Bound complexes were collected by centrifugation, washed five times in lysis buffer, and eluted by boiling in SDS-sample buffer for 5 min. Samples were separated by SDS-PAGE in 15% or 17% polyacrylamide gels and immunoblotted with srPLA 2 antiserum or with anti-phosphotyrosine mAb PY20, To detect srPLA 2 secretion, 5 ml of culture medium from hamster fibroblasts (1 day of culture, ϳ80% of confluence) were clarified by centrifugation and incubated in rotating vials either with srPLA 2 antiserum or with anti-phosphotyrosine mAb PY20, or with normal rabbit serum as control, previously bound to protein A-Sepharose beads. After washing in lysis buffer, the material bound to beads was separated in SDS-PAGE and immunoblotted as described above.
Preparation of GST Fusion Proteins in Bacteria and in Vitro Interactions-The C-terminal fragment of srPLA 2 cDNA, corresponding to amino acids 50 -154 (the fragment obtained as a result of two-hybrid screening of hamster cDNA library) and the full-length srPLA 2 including the missing 5Ј-region of the cDNA (93 nucleotides of mature proteincoding region) were subcloned into pGEX4T2 (Amersham Pharmacia Biotech). GST (glutathione S-transferase) and GST⅐srPLA 2 proteins were produced in E. coli strain BL21 according to the manufacturer's instructions. The expressed proteins were extracted from bacteria by sonication in modified lysis buffer (20 mM Tris, pH 7.5, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 M KCl) and purified on glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) according to manufacturer's procedure. Purity and integrity of GST fusion proteins was assessed by SDS-PAGE and Coomassie Blue R staining of proteins. For in vitro binding studies, 2 mg of protein extracts, obtained as described above, were incubated with glutathione-Sepharose 4B-bound proteins GST⅐srPLA 2 or GST in binding buffer (50 mM Tris, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.3 mM Na 3 VO 4 , 0.02% NaN 3 , 1 mg/ml aprotinin, 1% ovalbumin (18). Following four washes in lysis buffer and boiling in SDS sample buffer, eluted proteins were separated by SDS-PAGE.
Measurement of Phospholipase A 2 Activity-PLA 2 activity was assayed directly in the beads prepared as indicated above. Briefly, the beads were resuspended in the reaction mixture containing 140 mM NaCl, 30 mM KCl, 20 mM ATP, and 20 mM HEPES pH 7.5. Lipid kinase assays were initiated by addition of the mixture of 10 M phosphatidylcholine and 0.5 Ci/ml [ 14 C]phosphatidylcholine (Amersham Pharmacia Biotech). Reaction mixtures were incubated at room temperature for 40 min, then lipids were extracted by a chloroform/methanol mixture (1:1), and subsequently dotted onto silica gel plates. [ 14 C]arachidonic acid was separated from [ 14 C]phosphatidylcholine by thin-layer chromatography in chloroform, methanol, 11% ammonium hydroxide, water (65:35:6:3). The solvent was run up to 2 cm from the top of the plate. Plates were air dried, fixed, and autoradiographed to detect labeled lipids. Lipids were located by autoradiography or visualized with iodine vapor.
Quantitative Evaluation of the Data-Band intensities were quantified by scanning several autoradiograms obtained in independent experiments. Autographs scanned in an optical wavelength area using the three-dimensional scanner Extel, and grayscale images were obtained in TIFF files. These images were evaluated using Image Quant and TotalLab programs. This analysis produced quantitative data, in graphic format, for each band of the blot. Integration of each peak represented the relative quantities of material in the corresponding bands.

RESULTS
Two-hybrid Screening-A two-hybrid system was used to search for novel cellular proteins associated with the v-SrcHM and v-SrcLM (19). Full-length v-SrcHM and v-SrcLM and their C-terminal parts (v-SrcHMC and v-SrcLMC) were used as "baits" to screen the hamster cDNA library (8). The majority of the specific clones contained cDNA-encoding proteins not known as v-Src partners. Here we present detailed analysis of one gene identified as a result of screening with v-SrcHMC.
The ability of this clone to interact with two different regions of Src proteins, v-SrcHMN (v-Src without the HMC region), v-SrcHMC and v-SrcLMC was assessed by a two-hybrid test. The specific plasmid was introduced into yeast cells in the presence of pLexA-v-SrcHMN, pLexA-v-SrcHMC, pLex-v-Sr-cLMC, or pLexA-lamin (negative control), and ␤-galactosidase activity was assessed by filter and liquid assays (20). The interaction was found with v-SrcHMC and v-SrcLMC but not with v-SrcHMN (Table I). We concluded, therefore, that the presence of the C-terminal part of Src was essential for its association with this novel protein.
Sequence Analysis of the srPLA 2 Gene and Protein-The cloned cDNA was sequenced. A search in the GenBank TM data base revealed a homology with the 3Ј parts of different members of the phospholipase A 2 group II genes. The missing 5Ј section of the gene (150 nucleotides of coding region) was obtained as a result of polymerase chain reaction, using cDNA library and primers corresponding to the pGAD3S2X polylinker (sense) and to the isolated clone (antisense). The presence of full-length srPLA 2 RNA in hamster cells was confirmed by reverse transcriptase-polymerase chain reaction and subsequent analysis of the amplified DNA product.
The complete nucleotide and predicted amino acid sequences of srPLA 2 are shown in Fig. 1A. The highest level of homology (about 85%) was observed with the first 394 coding nucleotides of the rat PLA 2 gene from the IIA group of phospholipases A 2 (21). The homology observed between these two proteins was less obvious than the homology found at the DNA level (Fig.  1B).
Comparison of the srPLA 2 protein structure with amino acid sequences of other known members of II group phospholipase A 2, including IIA (21), IIB (22), IIC (23), IID (24), IIE (25), and IIF (25) confirms that this protein is related to the group IIA enzymes.
The protein encoded by the newly identified gene contains essential structural characteristics typical of the PLA 2 group II family of proteins: a signal peptide with a potential transmembrane helix (positions 1-21); the PLA 2 histidine (63-71), and aspartic acid (109 -119) in the active enzyme site; the calcium binding loop (45-52) and all conserved 14 cysteine residues typical for this group of proteins (Fig. 1, A and B). The amino acid differences are located in almost all regions of the protein except in the functional zones situated in the middle of the protein (Fig. 1B). The main difference between srPLA 2 and the majority of the other known PLA 2 proteins (with the exception of group IIF) is the existence of an additional 8-amino acid fragment at the C-terminal end of the hamster protein (Fig.  1B). This peptide appears as a result of changes in the last exon of srPLA 2 encoding DNA in comparison with PLA 2 IIA from phylogenetically related rodent cells. In the hamster cDNA, the stop codon present in the rat and mouse PLA 2 genes is missing, and the gene is terminated by a TAG codon localized further downstream. These structural changes were identified in all tested Syrian hamster DNAs isolated from different cell lines and animals (data not shown). Despite differences in the structure, our data strongly suggest that this novel srPLA 2 is a member of the so-called "secreted" PLA 2 gene family and, most probably, belongs to the IIA group of phospholipases A 2 .
In Vitro Protein Binding-To further demonstrate the interaction existing between srPLA 2 and v-Src, binding ability was tested in vitro. A fragment of srPLA 2 (srPLA 2 -(50 -154)) iso-TABLE I Yeast two-hybrid analysis of srPLA 2 -src interaction N-and C-terminal parts of v-SrcHM fusion to LexA and srPLA 2 fusion to pGAD were co-expressed in yeast strain L40. ␤-galactosidase activity reflecting proteins interaction was measured by a quantitative ␤-galactosidase solution assay. Yeast co-transformed with pLexA-lamin and pGAD-srPLA 2 were used as negative controls; yeast co-transformed with pLexA-Ras and pGAD-Raf were used as positive controls, and the data were normalized according to this interaction arbitrarily considered as a standard (value 1). A, coding nucleotides and corresponding amino acid sequences (GenBank TM /EBI nucleotide sequence data base accession number AJ251361). The signal peptide and PLA 2 active sites are overlined. signal peptide, 1-21 amino acids; calcium binding loop, 45-52 amino acids; histidine active site, 63-71 amino acids; aspartic acid active site, 109 -119 amino acids. The hypothetical phosphorylation site (tyrosine at position 91) is indicated by an asterisk. B, comparison of amino acid sequences of srPLA 2 with PLA 2 proteins from different members of group II. Amino acid residues coinciding with the srPLA 2 protein sequence are shown. Yellow indicates seven identical amino acids at a particular position, green indicates six identical amino acids, blue indicates five identical amino acids, gray indicates four identical amino acids, pink indicates three identical amino acids, purple indicates two identical amino acids, and red indicates amino acids specific only for srPLA 2 . Cysteine residues, whose positions are conserved in all primary structures of group II PLA 2 , are indicated by asterisks. The peptide selected for the preparation of srPLA 2 -specific serum is overlined. lated by two-hybrid screening was expressed as a GST fusion protein and prepared from E. coli. Protein extracts from low metastatic HET-SR and high metastatic HET-SR-8 RSV-transformed cell lines were incubated with purified GST⅐srPLA 2 or GST proteins bound to the glutathione-Sepharose 4B beads, and retained proteins were analyzed by SDS-PAGE; mAbs against c-Src/v-Src were used for immunodetection. Previously, it was found that the proteins of both the HM and LM v-Src isoforms had a molecular mass of 62 kDa. This unusual molecular size (62 kDa instead of 60 kDa) is due to an insertion of 20 additional amino acids in the unique domain of the oncoproteins (5). Fig. 2A shows approximately the same level of v-Src and c-Src proteins in RSV-transformed cells, independent of their metastatic activity (lanes 1 and 2). In both cell lines, GST⅐srPLA 2 binds with the v-Src protein ( Fig. 2A, lanes 5 and  6), whereas GST does not ( Fig. 2A, lanes 3 and 4).
The truncated srPLA 2 , isolated by two-hybrid screening, lacked the first amino acids that constitute the N-terminal ␣-helix and the calcium-binding loop, both regions important structurally and functionally. To test the ability of full-length srPLA 2 to interact with the Src product, GST-full-length sr-PLA 2 fusion protein was used in protein binding assay in vitro. Fig. 2B shows that the fusion protein carrying full-length sr-PLA 2 protein also effectively binds v-Src protein from transformed cells.
The phospholipase activity of bacterially expressed GST-⅐srPLA 2 protein was measured. Fig. 3 demonstrates that the purified bacterial fusion protein carrying full-length srPLA 2 stimulates the formation of arachidonic acid after incubation with phosphatidylcholine substrate: i.e. this fusion protein preserves its enzymatic function. According to quantitative evaluation, 1 mg of total bacterial protein contains approximately 20 units of PLA 2 activity.
In RSV-transformed hamster cells expressing v-Src, no interaction between the c-Src protein and srPLA 2 was found ( Fig.  2A, lanes 5 and 6). The interaction of the fusion GST⅐srPLA 2 with c-Src proteins from cells not expressing v-Src, was analyzed (Fig. 2C). NIH3T3 cells overproducing avian c-Src protein and RSV-transformed hamster cells with depressed v-src gene transcription (HET-SR-N-ras cl.34) were compared (Fig. 2C,  lanes 1 and 2). In cell lines overproducing exogenous avian c-Src, a minimal amount of p60 c-src was found to be associated with srPLA 2 (Fig. 2C, lane 5). The binding of p60 c-src with GST⅐srPLA 2 from HET-SR-N-ras cl.34 cells producing only the c-Src protein was not observed (Fig. 2C, lane 6). The interaction between GST⅐srPLA 2 and c-Src from spontaneously transformed hamster fibroblasts (STHE) producing high level of Src kinase activity was also observed (data not shown). The protein binding data suggests that both v-Src and c-Src proteins can potentially interact with the srPLA 2 protein, but the affinity of the latter is significantly higher for the viral oncoprotein. srPLA 2 Expression in RSV-transformed Cells-To detect the srPLA 2 protein and its interaction with Src proteins in cells, rabbit anti-srPLA 2 serum elicited against a peptide in the C-terminal part of the protein (amino acids 131-145) was prepared. The chosen peptide includes the residues from the amino acid extension at the C terminus and shares less homology with the same region of other known phospholipases (Fig.  1B).
This family of enzymes is referred as "secreted PLA 2 " (26). We compared the srPLA 2 protein in HET-SR culture medium and in cell lysates. The molecular mass of srPLA2 in cell lysates corresponds to the precursor forms of the PLA 2 II group of proteins (about 17 kDa) (Fig. 4A, lane 5). The extracellular protein has a molecular mass of 14 kDa corresponding to the mature form of PLA 2 proteins without the N-terminal signal peptide (Fig. 4A, lane 4). Total cell lysate and culture medium (M) immunoprecipitated with normal rabbit antiserum were used as negative controls (Fig. 4A, lanes 1 and 2). The immunodetection by anti-srPLA 2 antibodies was specific because incubation of the serum with the antigenic peptide precluded detection of the 17-kDa band (Fig. 4B). The absence of detectable 14-kDa srPLA 2 in the total cell lysate blot was probably due to low amounts of this protein in cytosolic fractions (Fig.  4A, lane 5).
A direct association between the v-Src protein and the precursor form of srPLA 2 in transformed cells was detected. After immunoprecipitation of cell lysates with anti-Src antibodies, a protein (17 kDa) recognized by anti-srPLA 2 antibodies was observed (Fig. 4A, lane 3).
The v-Src protein contains tyrosine kinase activity. We surmised that the interaction between srPLA 2 and Src could bring about phosphorylation of srPLA 2 on tyrosine. To examine this hypothesis, the blots reacting with srPLA 2 anti-serum were stripped and reprobed with the anti-phosphotyrosine monoclonal antibody PY20. Fig. 4C shows that both 17- (Fig. 4C, lane  3) and 14-kDa forms (Fig. 4C, lane 4) are phosphorylated on tyrosine. The phosphorylation of srPLA 2 protein was confirmed using the reciprocal procedures: the total phosphotyrosine-containing proteins were precipitated from culture medium and cell lysates by anti-phosphotyrosine antibodies and tested by anti-srPLA 2 serum. Among the tyrosine-phosphorylated proteins, the 17-and 14-kDa bands were identified in culture medium and cell lysates, correspondingly (Fig. 4D, lanes 1 and  2). A complex formed between the v-Src protein and the tyrosine-phosphorylated 17-kDa form of srPLA 2 was also detected in lysates of high metastatic cell lines (data not shown).
The steady-state level of srPLA 2 -specific RNA was analyzed

FIG. 2. Comparative analysis of the Src protein in vitro binding with the GST⅐SrPLA2 fusion protein. A and C,
GST with the truncated srPLA 2 isolated after a two-hybrid screening; B, GST with the full-length srPLA 2 . Cell lysates were incubated either with purified GST or with GST⅐SrPLA2 proteins bound to glutathione-Sepharose beads, washed, run on SDS-PAGE in parallel with aliquots of the same cell lysates, and transferred to nitrocellulose. The blots were probed with mAb against v-Src/c-Src. A, lanes 1, 3, 5 in high and low metastatic cell lines. A single 0.8-kilobase RNA transcript was identified in cell cultures expressing the v-src gene. In cells with a depressed v-src gene expression, the amount of srPLA 2 -specific messenger was decreased (Fig. 5) In normal hamster embryo fibroblasts, this RNA was practically undetectable (data not shown).

DISCUSSION
By analyzing the protein partners of various isoforms of the v-Src oncoprotein (v-SrcLM and v-SrcHM), we identified a novel gene encoding a low molecular weight phospholipase A 2 . This gene (srPLA 2 ) has a high degree of homology with the group II PLA 2 genes from different species, including rodent, human, and viper.
The nearest evolutionary rat IIA group PLA 2 exhibits 70% amino acid sequence identity with srPLA 2 . srPLA 2 contains all the functional areas of the known group II PLA 2 including the signal peptide region, histidine-aspartate active sites, the calcium binding loop, and the 14 conserved cysteine residues involved in the disulfide bonds (Fig. 1, A and B). As in the other members of group II, srPLA 2 contains amino acids localized identically in PLA 2 from groups IB and V (27). No homology was observed with PLA 2 from different groups in the C-terminal part of the srPLA 2 protein (Fig. 1B). The most unusual structural alteration in srPLA 2 was the presence of an additional eight amino acid residues at the C-terminal end of the protein. The function exerted by these amino acids remains, at present, unclear.
We show here that there is a direct interaction of this novel srPLA 2 with the v-Src oncoprotein. This was shown by different methodological approaches in vitro as well as in vivo: (i) by using the yeast two hybrid system (ii) by the association of a GST⅐srPLA 2 fusion protein with the v-Src protein in lysates of transformed cells, and (iii) by using antibodies prepared against srPLA 2 : the complex between v-Src and srPLA 2 proteins could be immunoprecipitated from cell lysates. srPLA 2 was detected as a 17-kDa precursor in the cells and as a mature 14-kDa form secreted into the culture medium. According to the molecular weight of srPLA 2 in the complex, it is suggested that the target of the Src protein could be the cellular 17-kDa precursor form of srPLA 2 . We also showed that both the 17-and anti-peptide serum after immunological exhaustion with corresponding peptide. C, the membrane previously blotted with anti-srPLA 2 serum was stripped in the stripping buffer (100 mM 2-mercaptoethanol, 2% SDS 62, 5 mM Tris-HCl, pH 6.7) at 50°C for 30 min and washed, blocked, and reprobed with anti-phosphotyrosine Ab PY20. D, the cell lysate and the culture medium were immunoprecipitated with anti-phosphotyrosine Ab, and immunoblotted with anti-srPLA 2 serum. The positions of the precursor srPLA 2 (17 kDa) and mature form of srPLA 2 (14 kDa) are indicated by arrows. the 14-kDa proteins, identified by Ab srPLA 2, respectively in RSV-transformed cells lysates and in culture medium, were phosphorylated on tyrosine.
Although v-srcHM and v-srcLM present structural changes in their C-terminal region, we observed no difference in their interaction with srPLA 2 , which appears, therefore, to recognize both isoforms irrespective of their metastatic potential.
According to the NetPhos 2.0 protein phosphorylation prediction program (28), tyrosine 91 in the srPLA 2 protein is the potential site for phosphorylation. There is no direct evidence, as yet, that this protein is phosphorylated as a result of Src tyrosine kinase activity because none of the tyrosines in sr-PLA 2 (including Tyr-91) are in a known preferred environment for phosphorylation by the Src tyrosine kinase (29). Structure preferences for tyrosine kinases, however, are weakly pronounced (30). Considering that a 17-kDa srPLA 2 was observed in complexes with the v-Src product, we believe that Src could be involved in srPLA 2 phosphorylation.
According to our preliminary indirect immunofluorescence imaging, srPLA 2 is localized in the endoplasmic reticulum (ER) and is presented on the surface of transformed cells. We speculate that in the process of maturation, the srPLA 2 precursor interacts with the Src protein at the cell membrane; thereafter, the mature form of tyrosine phosphorylated srPLA 2 is secreted. How srPLA2, which is stored in either granules or ER, can interact with v-Src at the plasma membrane, and how it affects the enzyme secretion is questionable. It was shown that induction of growth factor independence and granulocyte-macrophage colony-stimulating factor secretion by the v-Src oncogene does not require membrane association of pp60v-src (31). Therefore, v-Src functions may not be dependent on membrane localization. Whether the interaction of srPLA 2 with Src protein is specific only for RSV-transformed v-Src-producing cells or plays a general role in the regulation of group IIA PLA 2 remains to be determined.
The initial aim of the present research was to use a unique cell system model to detect cellular genes that could possibly play a role in the acquisition (or loss) of the metastatic activity of transformed cells. The identification of srPLA 2 , related to the phospholipase A 2 family, interacting with Src as a potential target protein, was an inadvertent finding. Phospholipases A 2 (phosphatidylcholine 2-acylhydrolases; EC 3.1.1.4) catalyze the hydrolysis of the sn-2 ester bond in phospholipids, generating free fatty acids, such as arachidonic acid and lysophospholipids (32,33). Arachidonic acid is the key substrate for the synthesis of potent lipid mediators of inflammation (e.g. prostaglandin, leukotrienes, etc.). Group IIA PLA 2 has been proposed to play a role in many pathological conditions, such as rheumatoid arthritis (34), septic shock (35), pancreatitis (36), and psoriasis (37) that are mediated by arachidonic acid release.
Several reports demonstrate that PLA 2 are probably also involved in the processes of malignant progression of the transformed cell. (i) The basal PLA 2 activity was found to be increased in rodent fibroblasts transformed by membrane-associated oncogenes (38). (ii) The Pla2 g2a gene, encoding group IIA PLA 2 , has been implicated in colon cancer. The "murine intestinal neoplasia" or min1 gene is the murine homologue of the APC gene, mutated in human familial adenomatosis polyposis, a hereditary form of colon cancer. It is worth noting that, recently, it was shown that c-Src-activating mutations in the regulatory C-terminal end region of the protein are observed in 12% of advanced human colon cancer cases. This change in c-Src is activating, transforming, tumorigenic, and promotes metastasis (39). (iii) specific antisense oligonucleotide ASGII blocked the appearance of a heparin-releasable group IIA PLA 2 in the culture supernatants of macrophage-like P388D1 cells. The disappearance of this protein correlated with reduced prostaglandin E2 production by activated cells, indicating that an extracellular pool of group IIA PLA 2 is involved in prostaglandin production by P388D1 cells (40). In the treat- ment of NIH3T3, fibrosarcoma, and sarcoma, but not lymphoma or mastocytoma cells, with both porcine pancreatic and Naja naja phospholipases A 2 causes these cells to dramatically invade the extracellular matrix in vitro in a dose-dependent manner (41).
As it was mentioned above, RSV-transformed hamster cells are secreting PgE, which inactivates the cytotoxic activity of NK cells (1). Among normal cells, the same phenotype is expressed in activated macrophages. The significant decrease of v-src gene expression correlates, in turn, with the complete suppression of PgE secretion as well as with other properties related to resistance of cells to NK lymphocytes and macrophages. According to preliminary data, the production of sr-PLA 2 was lower in cells not expressing active Src and PgE.