Multiple Splice Variants of the Human Calcium-independent Phospholipase A2 and Their Effect on Enzyme Activity*

Recently, the cloning of a novel Ca2+-independent phospholipase A2(iPLA2) from Chinese hamster ovary cells as well as from mouse and rat sources containing a C-terminal lipase motif and eight N-terminal ankyrin repeats has been described. In this report we describe the cloning of the human iPLA2 cDNA and its expression in B-cells and show that the iPLA2 gene undergoes extensive alternative splicing generating multiple isoforms that contribute to a novel mechanism to control iPLA2activity. The full-length cDNA clone encodes a 806-amino acid protein with a calculated molecular mass of 88 kDa. The protein contains a lipase motif, GXSXG, and ankyrin repeats, as described for the hamster and rodent forms of the enzyme but has an additional 54-amino acid proline-rich insertion in the last of the eight ankyrin repeats (residues 395–449). Furthermore, at least three additional isoforms most likely due to alternative splicing were identified. One that is present as a partial cDNA in the expressed sequence tag data base is similar to iPLA2 but terminates just after the lipase active site, and two other isoforms contain only the iPLA2 ankyrin repeat sequence (ankyrin-iPLA2-1 and -2). Ankyrin repeats are involved in protein-protein interactions and because the purified iPLA2enzyme exists as a multimeric complex of 270–350 kDa, the expression of just the ankyrin-iPLA2 sequence suggested that these may also interact with the iPLA2 oligomeric complexes and perhaps modulate PLA2 activity. Transfection of the human iPLA2 cDNA into COS cells resulted in a substantial increase in calcium-independent PLA2 activity in cell lysate. No activity above background was observed following ankyrin-iPLA2-1 cDNA transfection. However, co-transfection of the ankyrin-iPLA2-1 and the iPLA2 cDNAs resulted in a 2-fold reduction in activity compared with iPLA2 alone. A similar co-transfection of ankyrin-iPLA2-1 cDNA with the cPLA2cDNA had no effect on PLA2 activity. These results suggest that the ankyrin-iPLA2 sequence can function as a negative regulator of iPLA2 activity and that the alternative splicing of the iPLA2 gene can have a direct effect on the attenuation of enzyme activity.

Recently, the cloning of a novel Ca 2؉ -independent phospholipase A 2 (iPLA 2 ) from Chinese hamster ovary cells as well as from mouse and rat sources containing a C-terminal lipase motif and eight N-terminal ankyrin repeats has been described. In this report we describe the cloning of the human iPLA 2 cDNA and its expression in B-cells and show that the iPLA 2 gene undergoes extensive alternative splicing generating multiple isoforms that contribute to a novel mechanism to control iPLA 2 activity. The full-length cDNA clone encodes a 806-amino acid protein with a calculated molecular mass of 88 kDa. The protein contains a lipase motif, GXSXG, and ankyrin repeats, as described for the hamster and rodent forms of the enzyme but has an additional 54-amino acid proline-rich insertion in the last of the eight ankyrin repeats (residues 395-449). Furthermore, at least three additional isoforms most likely due to alternative splicing were identified. One that is present as a partial cDNA in the expressed sequence tag data base is similar to iPLA 2 but terminates just after the lipase active site, and two other isoforms contain only the iPLA 2 ankyrin repeat sequence (ankyrin-iPLA 2 -1 and -2). Ankyrin repeats are involved in protein-protein interactions and because the purified iPLA 2 enzyme exists as a multimeric complex of 270 -350 kDa, the expression of just the ankyrin-iPLA 2 sequence suggested that these may also interact with the iPLA 2 oligomeric complexes and perhaps modulate PLA 2 activity. Transfection of the human iPLA 2 cDNA into COS cells resulted in a substantial increase in calcium-independent PLA 2 activity in cell lysate. No activity above background was observed following ankyrin-iPLA 2 -1 cDNA transfection. However, co-transfection of the ankyrin-iPLA 2 -1 and the iPLA 2 cDNAs resulted in a 2-fold reduction in activity compared with iPLA 2 alone. A similar co-transfection of ankyrin-iPLA 2 -1 cDNA with the cPLA 2 cDNA had no effect on PLA 2 activity. These results suggest that the ankyrin-iPLA 2 sequence can function as a negative regulator of iPLA 2 activity and that the alternative splicing of the iPLA 2 gene can have a direct effect on the attenuation of enzyme activity.
Phospholipases A 2 (PLA 2 ) 1 are a rapidly growing family of diverse enzymes that hydrolyze fatty acids at the sn-2 position of phospholipids (1,2). PLA 2 enzymes can be subdivided into two classes, extracellular or intracellular, depending on the enzymes localization during catalysis (2). The intracellular PLA 2 s can be further categorized into calcium-dependent, best exemplified by the cytosolic phospholipase A 2 (cPLA 2 ) (3), and calcium-independent forms (iPLA 2 ), which tend to be quite diverse and have until recently been less characterized at the molecular level. The calcium-independent PLA 2 s have a wide tissue distribution (4) and have been purified from human myocardium (5), bovine brain (6), P388D 1 murine macrophages (7), and rabbit kidney (8). They all have distinct molecular masses, indicating the diversity of iPLA 2 s. Recently, an 85-kDa iPLA 2 was purified and cloned from CHO cells (9), and its sequence was found to be analogous to the 85-kDa iPLA 2 from P388D 1 cells (10) as well as the sequence for iPLA 2 from rat pancreatic islet (11). The amino acid sequence indicated the presence of eight ankyrin repeats and the GXSXG conserved catalytic sequence, as found in other lipases. Although there were apparent differences in ATP sensitivity among the enzymes, the biochemical, immunological, and sequence data indicate that these three enzymes are likely to be species variants of the same protein. In both P388D 1 cells and in rat pancreatic islets it is thought that iPLA 2 has a function in membrane phospholipid remodeling (12). It has been postulated that the rat islet iPLA 2 may be involved in arachidonic acid release leading to activation of ␤-cell ion channels (11).
Arachidonic acid is also the main precursor for important biological mediators such as leukotrienes (LT) (13). The oxygenation of arachidonic acid catalyzed by 5-lipoxygenase is the first step in the biosynthesis of leukotrienes and leads to the formation of LTA 4 , which can be further metabolized to LTB 4 and LTC 4 . In our ongoing studies of leukotriene synthesis and phospholipase activity in human B lymphocytes, we have demonstrated conversion of arachidonic acid to LTB 4 and expression of 5-lipoxygenase in human B lymphocytes (14,15). Although cellular homogenates of B lymphocytes can release arachidonic acid from phospholipids in vitro, exogenous arachidonic acid is a prerequisite for leukotriene synthesis in intact cells (16). To elucidate the expression of PLA 2 (s) in human B lymphocytes, we have examined in this report the expression of the different PLA 2 s at the transcriptional level and describe the cDNA cloning of the human 85-kDa iPLA 2 and its various isoforms and their effect on enzyme activity.  A comparison of the two EST sequences with the CHO iPLA 2 cDNA indicated that these ESTs contained the 3Ј end of the human iPLA 2 cDNA sequence. Each contained about a third of the full-length iPLA 2 sequence as judged from the size of the CHO sequence. However, at the 5Ј end of these clones, EST 46450 aligned more closely to the CHO sequence than did EST 30643. Identical sequence is boxed, and the termination codon for the CHO cDNA is double underlined. The human iPLA 2 PCR primers 1 and 2, which were based on the EST 46450 sequence, are underlined, and the polyadenylation signal is shown with asterisks. The dashes indicates the presence of gaps in the sequence. fied atmosphere with 5% CO 2 (16). The cultures were seeded at a cell density of 0.2 ϫ 10 6 cells/ml and harvested at approximately 1 ϫ 10 6 cells/ml.
Cloning of Human iPLA A2 -The CHO cell-derived iPLA 2 amino acid sequence (9) was used to perform a TBLASTN data base search of GenBank. The sequences of two human expressed sequence tag (EST) clones, 46450 (accession number H10676) and 30643 (accession number R18691), were found to show considerable identity to the CHO cell sequence. Bacteria containing the EST clones were obtained from Research Genetics, plasmid DNA was prepared, and the cDNA insert was sequenced on both strands using an ABI 373A automated DNA sequencer and m13 forward and reverse, as well as gene-specific primers. The DNA sequence identified the ESTs as human homologs of the CHO cell iPLA 2 . To obtain a full-length human iPLA 2 cDNA clone, 5Ј-rapid amplification of cDNA ends (RACE) was used to amplify the sequence from various cDNA sources (see Fig. 4). Amplification was carried out using a human iPLA 2 -specific 3Ј primer (iPLA 2 -2), a 5Ј anchor primer (CLONTECH), and Marathon Ready cDNAs (CLONTECH) as template. The Expand High Fidelity PCR system (Boehringer Mannheim) was used for the 5Ј-RACE using the following conditions: an initial denaturation step at 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 2.5 min, and a final extension at 72°C for 10 min. The amplified products were analyzed on 0.8% agarose gels, and DNA fragments were recovered (Qiagen). Taq polymerase was used to add a 3Ј A overhang to the fragment, which was then cloned into the pCR2.1 TA cloning vector (Invitrogen). Sequencing of both strands of the cloned 5Ј-RACE fragment was performed as described above.
A full-length human iPLA 2 clone was obtained by combining the 3Ј end of the iPLA 2 sequence in clone EST 46450 with the 5Ј iPLA 2 sequence obtained by the RACE reaction. Briefly, the 5Ј-RACE fragment was released from pCR2.1 by a HindIII/HpaI double digest. The released fragment was gel purified, extracted, and sticky/blunt end ligated into HindIII/HpaI-digested Lafmid BA/EST 46450. Sequencing of the full-length iPLA 2 was performed to verify the integrity of the sequence.
Transient Expression in COS-7 Cells-Plasmid DNA (5 g) containing iPLA 2 , ankyrin-iPLA 2 -1, or cPLA 2 cDNAs cloned into the eukaryotic expression vector pcDNA 3.1ϩ (Invitrogen) was transfected into COS-7 cells using LipofectAMINE (Life Technologies, Inc.). Transfections consisted of various combinations of the PLA 2 cDNAs and pcDNA 3.1 (see the legend to Fig. 6) to a total of 10 g of DNA transfection. Briefly, both DNA and LipofectAMINE (60 l) were each mixed with 800 l of medium (Opti-MEM) and then combined and incubated for 45 min at 20°C to allow formation of DNA-liposome complexes. Subsequently, 6.4 ml of medium was added to each tube, and the transfection mixture was transferred to washed COS-7 cells. Transfection was allowed to proceed for 5 h at 37°C in 80-mm dishes and terminated by replacement of the transfection mixture with Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
PLA 2 Enzyme Activity-Enzyme activities were determined using a vesicular based assay containing equal concentrations (5 M) of 1-palmitoyl 2-[1-14 C]arachidonyl phosphatidylcholine (55 mCi/mmol) and 1-palmitoyl 2-[1-14 C]arachidonyl phosphatidylethanolamine (55 mCi/mmol, NEN Life Science Products). The radioactive phospholipids were dried under nitrogen and resuspended in either 180 l of calciumfree assay buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, and 1 mg/ml albumin) or calcium-containing assay buffer (80 mM glycine, pH 9.0, 5 mM CaCl 2 , 2 mM dithiothreitol, and 1 mg/ml albumin). Cells were collected 48 h after transfection by scraping, washed twice, and resuspended in homogenization buffer (50 mM Tris-HCl, pH 7.5, 2 mM EDTA, 0.5 mM dithiothreitol, and 20% glycerol). Cells were then lysed by sonication (2 ϫ 5 s), the resulting homogenate was centrifuged at 10,000 ϫ g for 20 min at 4°C, and the supernatant was collected. Calcium-dependent and -independent PLA 2 activity in the supernatant was determined by the addition of 20 l of supernatant to 180 l of one of the above assay buffers, and the reaction was allowed to proceed for 20 min at 37°C before termination with the addition of 400 l of methanol containing 0.5% acetic acid and 10 M arachidonic acid. The sample was then applied to a Sep-Pak Vac (Waters) C18 cartridge, and fatty acids were eluted in 500 l of methanol. The [1-14 C]arachidonate content of the eluate was analyzed by high pressure liquid chromatography using a Nova-Pak C18 column (3.9 ϫ 150 mm) at a flow rate of 1 ml/min (the mobile phase was methanol/H 2 O/trichloroacetic acid (85: 15:0.01 by volume), and radioactivity was detected using a Beckman 171 radioisotope detector coupled on-line to a Waters 996 diode array spectrophotometer. Peak area integration was performed using Millenium software (Waters).

PLA 2 Expression in B-cell
Lines-Human monoclonal B-cells express both FLAP and 5-lipoxygenase and have the ability to make LTs only after the addition of exogenous arachidonic acid and stimulation with calcium ionophore and a reducing agent (14,15). The reason for the lack of cellular leukotriene biosynthesis in B-cells without these added factors is unclear. Furthermore, stimulation with various agents known to promote arachidonic acid release fails to induce a similar release in B-cells. However, B-cells possess phospholipase activity and sonicates of these cells release arachidonic acid (16). Therefore, to investigate the presence of PLA 2 in B-cells, we have examined the expression of the various PLA 2 enzymes at the tran- RT-PCR analysis of total RNA from BL-41 E95A and Raji cell lines using iPLA 2 primers 1 and 2 was performed as described under "Materials and Methods." A 10-l aliquot of the PCR reaction was analyzed on a 1.5% agarose gel containing ethidium bromide. Reactions containing reverse transcriptase are indicated with a plus sign (ϩ) and negative controls without reverse transcriptase with a minus sign (Ϫ). A glyceraldehyde 3-phosphate dehydrogenase (G3PDH) control amplification was also performed using the same reaction conditions. The DNA size markers are in bp. scriptional level using RT-PCR. Total RNA from the human monoclonal B-cell lines Raji and BL-41 E95A was used as template for the RT-PCR reaction, and primers were designed for the amplification of a fragment of the following PLA 2 cDNAs: cPLA 2 (3), iPLA 2 , sPLA 2 types I, II, and V (2,18), and lipoprotein-associated PLA 2 (19). The sequence used for the design of the iPLA 2 primers was obtained by a TBLASTN data base search using the CHO cell-derived amino acid sequence. Two human EST clones, 46450 and 30643, revealed 84 and 65% sequence identity, respectively, to the CHO cell-derived sequence, strongly suggesting that they represented the human form of the enzyme. Sequencing of the two clones revealed that EST 46450 contained a partial cDNA for the human homolog of iPLA 2 , whereas EST 30643, which also contained part of the human iPLA 2 cDNA sequence, differed from EST 46450 at the 5Ј end (Fig. 1). The difference in sequence at the 5Ј end of these clones was not due to a misspliced intron because the 3Ј intron junction consensus sequence was not present. Therefore, it could be either a cDNA library artifact or the iPLA 2 sequence is subject to alternative splicing.
The iPLA 2 primers 1 and 2 ( Fig. 1) along with the primers for the various PLA 2 s described above were used to amplify B-cell cDNA and revealed that the iPLA 2 sequence was easily amplified from these cells. A weak signal for the cPLA 2 sequence was detectable, but none of the other PLA 2 sequences were detected (data not shown). The weak signal obtained for cPLA 2 is in agreement with previous results showing that the cPLA 2 mRNA (20, 21) or protein (22) was either not present or detectable at low levels in B-cells. The results from one typical RT-PCR using primers iPLA 2 -1 and -2 is shown in Fig. 2. At least three distinct amplified products were observed. The most abundant product was the expected 217-bp fragment, whose identity was confirmed by sequencing, as well as two additional fragments with the apparent sizes of 330 and 380 bp, respectively. The 380-bp fragment was sequenced and found to contain the 5Ј end of EST clone 30643 (Fig. 3A). This iPLA 2 EST contains a 168-bp insert between iPLA 2 primers 1 and 2, indicating that there are splice variants of iPLA 2 . The insertion introduces a premature termination codon that would result in the expression of a truncated iPLA 2 (Fig. 3A). The nature of the 330-bp fragment is currently under investigation, although it is related to iPLA 2 because it hybridizes to an iPLA 2 oligonucleotide probe (data not shown).
Cloning of the Human iPLA 2 cDNA-The results from the RT-PCR indicated that iPLA 2 is at least one of the more significant PLA 2 enzymes in B-cells, and therefore we decided to characterize it more extensively. To obtain a full-length iPLA 2 cDNA clone, 5Ј-RACE was performed on various cDNAs to amplify the remainder of the sequence. Fig. 4 shows the results from one such reaction and the substantial amplification of a 2.2-kilobase DNA fragment from testis cDNA. This fragment was subcloned, and sequence data were obtained for five different clones. Three clones had identical sequence and contained an open reading frame encoding the human iPLA 2 cDNA sequence (Fig. 5). One of the clones had an insertion just before the iPLA 2 catalytic domain that produced a frameshift, thus leading to a truncated iPLA 2 without the catalytic domain (Fig. 3B). This clone was identified as ankyrin-iPLA 2 -1 to indicate that it coded for only the ankyrin repeats. Similarly, the other clone, ankyrin-iPLA 2 -2, had the identical 53-bp insertion as the previous isoform but also had an additional 52-bp insertion 80 bp 5Ј to the ankyrin-iPLA 2 -1 insertion, thus generating a different C terminus but again without the catalytic domain (Fig. 3B). In addition, this clone contained a 216-bp in-frame deletion in the 5Ј region of the cDNA resulting in the removal of amino acids 71-143 (Figs. 3C and 5). To prove that the observed different forms of iPLA 2 were not due to artifacts in the 5Ј-RACE reaction, testis cDNA as well as two different testis cDNA libraries and B-cell cDNA were subjected to PCR using primers iPLA 2 -33 and -34 (Fig. 3B), which spanned the region of interest. Two detectable amplified products were obtained, one major product corresponding to iPLA 2 and a minor one corresponding to ankyrin-iPLA 2 -1, whereas ankyrin-iPLA 2 -2 was not observed (data not shown). The absence of isoform 2 is not too surprising because it could be at levels that would require more rounds of amplification or the use of ankyrin-iPLA 2 -2-specific primers.
A full-length iPLA 2 cDNA sequence was obtained by combining the 3Ј end of EST 46450 with the appropriate 5Ј-RACE fragment (see "Materials and Methods"). The inferred human iPLA 2 amino acid sequence was compared with the recently reported sequences and found to have an overall identity of 90% to hamster, rat, and mouse iPLA 2 sequences (Fig. 5A). The major difference between the human sequence and the other species is the insertion of an additional 54 amino acids that would interrupt the last putative ankyrin repeat as defined in the hamster and rat iPLA 2 sequences (9, 11) (Fig. 5). This sequence is also present in the other human iPLA 2 isoforms. A search of GenBank™ with this additional 54-amino acid sequence did not reveal any significant homologies.
In addition to the human iPLA 2 cDNA there are at least three other iPLA 2 isoforms. The EST 30643 sequence, if it were full-length, could encode an iPLA 2 that contained the ankyrin repeats and lipase active site but with a truncated C terminus (Fig. 5). The other two iPLA 2 isoforms ankyrin-iPLA 2 -1 and -2 contain only the ankyrin repeats because they terminate before the lipase active site (Fig. 5). A diagrammatic representation of the iPLA 2 isoforms is illustrated in Fig. 5B. The ankyrin-iPLA 2 -1 sequence is identical to the human iPLA 2 up to the last three C-terminal amino acids and terminates about 40 residues before the active site serine. In contrast, although the ankyrin-iPLA 2 -2 sequence has the same iPLA 2 ankyrin repeats, it also contains more structural and sequence variation than does ankyrin-iPLA 2 -1. First, there is the 72-amino acid deletion between residues 70 -143, and second, the additional insertion in the iPLA 2 -2 cDNA (Fig. 3C) alters the sequence of the last 50 C-terminal amino acids (Fig. 5, A and B). The presence of these iPLA 2 isoforms indicates that the iPLA 2 sequence is subjected to a significant amount of alternative splicing.
Effect of Ankyrin-iPLA 2 Sequence on iPLA 2 Activity-The ankyrin structural motif appears to have a function in the formation of various types of protein-protein interactions (23). Deletion of the ankyrin repeats of the CHO cell iPLA 2 results in the loss of lipase activity, suggesting that this structure is required for enzyme activity (9). A possible function for the ankyrin repeats in iPLA 2 is to participate in the formation of the large oligomeric structures (270 -350 kDa) found for iPLA 2 upon gel filtration (7,9). If this is the case then the ankyrin-iPLA 2 sequences (iPLA 2 -1 and -2) may also participate in the formation of the iPLA 2 oligomeric structures in the cell and have some effect on iPLA 2 activity. To test this, iPLA 2 cDNA was co-transfected into COS cells with either pcDNA 3.1 vector or ankyrin-iPLA 2 -1 cDNA, and PLA 2 activity was measured in the cell lysate in the presence and the absence of calcium (Fig.  6, A and B). A control transfection for PLA 2 activity using the cPLA 2 cDNA was also performed. Transfection of iPLA 2 and control vector resulted in a substantial increase in PLA 2 activity in cell lysate over control (vector alone) in the absence (Fig.  6A) or the presence of calcium (Fig. 6B), indicating that the iPLA 2 cDNA sequence codes for an active enzyme. Increased PLA 2 activity was also observed upon transfection of the cPLA 2 cDNA but only in the presence of calcium, as would be expected (Fig. 6B). Transfection of just the ankyrin-iPLA 2 cDNA sequence did not result in any PLA 2 activity over background (data not shown). However, replacement of the vector control in the iPLA 2 cDNA transfection with ankyrin-iPLA 2 -1 cDNA caused a 2-fold decrease in PLA 2 activity in the cell lysate in both assays (Fig. 6, A and B). Including the ankyrin-iPLA 2 -1 cDNA in the cPLA 2 cDNA transfection had no effect on PLA 2 activity (Fig. 6B). These results would suggest that the ankyrin-iPLA 2 sequences can specifically modulate iPLA 2 activity.

DISCUSSION
The human iPLA 2 sequence presented here contains the ankyrin and the GXSXG lipase motifs described for the hamster and rodent enzymes (9,11). A major difference between the human enzyme and that of other species is the insertion of a 54-amino acid proline-rich sequence in the eighth ankyrin motif (residues 395-449). A recent report from Tang et al. (9), which describes a partial human iPLA 2 cDNA clone, found at the exact same position in their clone some sequence that was not present in the CHO iPLA 2 , which they attributed to potential unspliced intron sequences. The insertion described here for human iPLA 2 , which was detected in all sequenced iPLA 2 clones, does not contain any splice junction consensus sequence, nor does it disrupt the reading frame. Whether the human gene contains an additional exon or this form is also present in the other species but has not yet been cloned remains to be determined. The possibility that iPLA 2 expression could give rise to splice variants was suggested by finding two partial iPLA 2 cDNAs in the EST data base with different 5Ј ends. The difference was found to be due to a 168-bp insertion in EST 30643. Both EST 30643 and iPLA 2 cDNA sequences were easily detected by RT-PCR of B-cell RNA suggesting that both iPLA 2 isoforms are expressed. Presently, we are trying to obtain a full-length EST 30643 clone to confirm that this truncated isoform of iPLA 2 is catalytically active. Two additional iPLA 2 splice variants were also detected, ankyrin-iPLA 2 -1 and -2, which contain only the ankyrin repeats and no lipase catalytic site. The existence of these truncated forms of iPLA 2 is intriguing because the ankyrin motif has been shown to be involved in various types of protein-protein interactions (23)(24)(25)(26). The truncated forms could function as negative regulatory proteins by docking to iPLA 2 binding sites in the cell and thereby prevent docking of the catalytically active enzyme. Alternatively, they may interfere with the formation of the quaternary structure of iPLA 2 and in this way alter enzyme activity. An oligomeric form of the enzyme may indeed be the active state of the enzyme because removal of ankyrin repeats results in loss of enzyme activity (9). The fact that ankyrin-iPLA 2 -1 can alter iPLA 2 activity was shown by co-transfection of the iPLA 2 and ankyrin-iPLA 2 -1 cDNAs into COS cells. The co-transfection of both constructs results in a 2-fold decrease in PLA 2 enzyme activity compared with that observed for the co-transfection of iPLA 2 cDNA and vector DNA. Co-transfection of ankyrin-iPLA 2 -1 cDNA with the cPLA 2 cDNA had no effect on PLA 2 enzyme activity. Thus the interaction of ankyrin-iPLA 2 -1 with iPLA 2 results in a decrease in enzyme activity. The most likely explanation for this is a competition between ankyrin-iPLA 2 -1 and iPLA 2 monomers to form the oligomeric species. In cells transfected with the iPLA 2 cDNA alone, the iPLA 2 oligomer is most likely composed entirely of active iPLA 2 monomers. Cotransfection of both iPLA 2 and ankyrin-iPLA 2 cDNAs could result in iPLA 2 oligomers that contain various combinations of both iPLA 2 and ankyrin-iPLA 2 -1 monomers. What is not known is the stoichiometry of the active iPLA 2 enzyme complex. Does it contain only iPLA 2 monomers, or can it tolerate formations with ankyrin-iPLA 2 ? The size of the iPLA 2 complex, based on gel permeation chromatography, ranges from 270 to 350 kDa (7,9), suggesting that there are at most four iPLA 2 monomers/complex. If this is the case and we assume that there is equal expression of both subunits in the transfected cells and binding affinity is not changed between the different subunits, then iPLA 2 -(ankyrin-iPLA 2 -1) complexes should be enzymatically active, because enzyme activity was only decreased 2-fold in the co-transfected cells. However, a more thorough investigation using purified subunits will have to be done to confirm this. Although we can show that ankyrin-iPLA 2 can have an effect on iPLA 2 activity, its in vivo role remains to be determined. The level of ankyrin-iPLA 2 -1 expression is at least 10-fold lower than that of iPLA 2 , 2 suggesting that if it participates in the iPLA 2 enzyme complex, it must have a much more subtle effect on enzyme activity in the cell. Nevertheless, what we have shown is that alternative splicing of the iPLA 2 gene can have a direct effect on iPLA 2 activity.
In fact, alternative splicing of genes containing sequences that encode both an activity domain and some kind of protein binding domain appears to be a common mechanism to control activity. A similar situation to that of iPLA 2 is the IB␥ inhibition of NF-B activity. IB proteins contain six or seven ankyrin-like repeats, which have been shown to be essential for retaining NF-B in the cytoplasm and inhibit DNA binding by Rel/NF-B (25, 27). IB␥ is derived by alternative splicing of the murine p105 gene, which is the precursor for the p50 component of the NF-B, p50-p65 heterodimer (28,29). The N-terminal half of p105 contains p50, which is derived by proteolytic cleavage, whereas the C-terminal half has eight ankyrin repeats (29). In addition, there are multiple isoforms of IB␥, all derived by alternative splicing of the p105 gene and each with unique IB activities (30). The two ankyrin-iPLA 2 isoforms described here do have structural and sequence differences, but whether or not they have unique inhibitory activities remains to be determined. Additional examples of alternative splicing having positive and negative effects on activity are two genes involved in programmed cell death: ich-1 (Caspase-2) (31), which encodes a cysteine protease, and bcl-x, a bcl-2-related regulatory gene (32). Again it is analogous to the iPLA 2 situation described above in that both genes produce a long transcript that codes for a functional product that is inhibited by a truncated version encoded by a shorter alternatively spliced transcript (31)(32)(33).
In the two B-cell lines tested, the iPLA 2 cDNA sequence was easily amplified, whereas only a weak signal was obtained for the cPLA 2 cDNA. The sPLA 2 groups I, II, and V and the lipoprotein-associated PLA 2 sequences were not detected. The weak PCR signal obtained for cPLA 2 is consistent with previ- The seven ankyrin repeats in the human sequence are overlined (I-VII), and the eighth repeat, which is underlined in the CHO sequence, is interrupted by an insertion in the human sequence. The active site (GTSTG) is indicated in bold, and termination is indicated by an asterisk. The EST 30643 sequence is the partial sequence from the EST clone. Shading denotes identity, and the dots indicate the presence of gaps in the sequence. B, a diagrammatic representation of the iPLA 2 isoforms. The 54-amino acid insertion into the eighth ankyrin repeat of the human sequences and the 72-amino acid deletion in ankyrin-iPLA 2 -2 are illustrated. The active site serine is indicated for both the CHO and human sequence. The C-terminal hatched region in ankyrin-iPLA 2 -2 denotes nonidentical sequence, and the number of amino acids in each sequence is also shown. ous findings that showed that cPLA 2 expression in B-cells was either very low or undetectable (20 -22). Based on this it would suggest that iPLA 2 may be one of the more significant PLA 2 s in B-cells. In fact the partial human iPLA 2 cDNA sequence recently described (9) was cloned from a B-cell line cDNA library. However, no physiological stimuli is at present known to induce leukotriene synthesis in human B lymphocytes; hence they are dependent on exogenous arachidonic acid for leukotriene synthesis. Perhaps, what is required is the correct stimulus to activate the iPLA 2 enzyme. It has been demonstrated that the majority of the 5-lipoxygenase enzyme in nonstimulated B lymphocytes is located at the nucleus (16). Thus, an increase in the intracellular calcium concentration, which renders translocation of 5-lipoxygenase to the nucleus, might not be a prerequisite for leukotriene synthesis in B lymphocytes, because the 5-lipoxygenase already is located at the nucleus. Therefore, it is tempting to speculate that some stimulus that activates the calcium-independent iPLA 2 without increasing the intracellular calcium concentration may be sufficient for the induction of leukotriene synthesis in B lymphocytes. However, in a recent review, Balsinde and Dennis suggest that the major function of iPLA 2 is in membrane remodeling and not in arachidonic acid release, although involvement in the latter cannot be completely ruled out (34). Therefore, the function and the regulation of iPLA 2 in B lymphocytes and its role in leukotriene synthesis remain to be determined.
In conclusion, we describe in this report the cDNA sequence of the human iPLA 2 and its various splice variants. We furthermore present data indicating that a splice variant of the iPLA 2 containing only the ankyrin motifs and not the active site specifically modulates iPLA 2 activity when the proteins are co-expressed in COS-7 cells. These findings suggest that alternative splicing of the iPLA 2 pre-mRNA can result in the production of regulatory subunits that can modify iPLA 2 in vivo activity.