Expression, purification, and kinetic characterization of a recombinant 80-kDa intracellular calcium-independent phospholipase A2.

A CHO cell-derived 80-kDa recombinant polypeptide (GenBank number I15470[GenBank]) putatively encoding a calcium-independent phospholipase A2 was expressed in S. frugiperda cells resulting in over a 15-fold increase in a calcium-independent phospholipase A1/A2 activity which was entirely inhibitable by (E)-6-(bromomethylene)-3-(1-naphthalenyl)-2H-tetrahydropyran-2-one. The recombinant polypeptide was purified from cytosol by sequential tandem affinity chromatographies employing ATP-agarose and calmodulin-Sepharose stationary phases. This strategy resulted in the rapid purification (36 h) of recombinant phospholipase A2 activity in 56% overall yield to a single intense 80-kDa protein band on SDS-polyacrylamide gel electrophoresis after silver staining. The purified protein possessed phospholipase A1, phospholipase A2, and lysophospholipase activities. Microbore anion exchange chromatography demonstrated that the 80-kDa protein band was comprised of multiple distinct isoforms including an anionic isoform which possessed over a 5-fold higher specific activity (5 μmol/mg·min) than earlier eluting isoforms. Collectively, these results unambiguously demonstrate that: 1) the 80-kDa polypeptide catalyzes phospholipase A1/A2 and lysophospholipase activities with distinct kinetic parameters; 2) calmodulin and ATP both interact with the catalytic polypeptide independent of regulatory proteins; and 3) distinct isoforms of this polypeptide exist which possess markedly different specific activities.

The intracellular phospholipases A 2 represent a rapidly expanding class of enzymes which have been categorized based on their calcium dependence into calcium-dependent and calciumindependent subtypes (1). Prior work has unambiguously identified a sequence encoding an 85-kDa intracellular calcium-dependent phospholipase A 2 which translocates to the membrane interface in the presence of calcium ion (2)(3)(4)(5)(6). Recently, a protein putatively catalyzing an intracellular calcium-independent phospholipase A 2 activity from CHO 1 cells has been described in abstract form (7), and its sequence determined (GenBank number I15470). Transient expression of this protein in the same cell type from which it was purified (i.e. CHO cells) results in enhanced calcium-independent phospholipase A 2 activity. Of course, the assignment of catalytic function to a recombinant protein expressed in the same context from which it was originally isolated requires purification of the protein to homogeneity to unambiguously demonstrate its role as a catalytic entity and not as an activator of an endogenous activity. However, due to intractable technical obstacles, the protein catalyzing this activity has never been purified to homogeneity to unambiguously discriminate between its potential role as an activator of an endogenous phospholipase activity versus a bona fide catalytic entity.
Since at least some members in the family of calcium-independent phospholipases A 2 exhibit specific high affinity interactions with ATP and/or calmodulin (8 -13), we sought to exploit potential interactions between these ligands and the expressed 80-kDa recombinant polypeptide to facilitate its purification to homogeneity to unambiguously identify its role in the catalytic process. We now report: 1) the expression of catalytically active recombinant 80-kDa calcium-independent phospholipase A 2 in a baculovirus expression system; 2) the rapid purification of the 80-kDa recombinant calcium-independent phospholipase A 2 to homogeneity through tandem sequential ATP and calmodulin affinity columns; and 3) the identification of multiple isoforms of this calcium-independent phospholipase A 2 which possess markedly different specific activities.
Cell Culture-S. frugiperda (Sf9) cells were cultured in 250-ml flasks equipped with a magnetic spinner containing supplemented 1 ϫ Grace's media (14). For expression of recombinant calcium-independent phospholipase A 2 , a 250-ml flask was prepared with 80 ml of 1 ϫ 10 6 cells/ml and incubated at 25°C for 24 h prior to infection with baculovirus. DUKX B1 CHO cells (ATTC #CRL 9010) were cultured in T-75 flasks containing minimal essential medium according to established methods (15).
Typically, assays of column eluents were conducted with 25 l of enzyme in a final volume of 210 l. After incubation at 37°C for 5 min, the reactions were quenched by addition of 1-butanol (100 l), vortexed, and the phases were separated by centrifugation. The reaction products from column assays were resolved by thin layer chromatography (Whatman Silica Gel 60A plates) utilizing a mobile phase consisting of 80:20:1 (petroleum ether:diethyl ether:acetic acid (v/v)). Radiolabeled fatty acid was identified by staining of a standard fatty acid sample with iodine vapor, scraped, and quantified by scintillation spectrometry. For kinetic assays, calcium-independent phospholipase A 1 /A 2 activities were assessed by quantifying the release of radiolabeled lysophospholipid or radiolabeled arachidonic acid, respectively, from L-␣1-palmitoyl-2-[1- 14 C]arachidonyl phosphatidylcholine. The release of radiolabeled fatty acid and lysophospholipid was quantified by incubating 120 ng of purified recombinant 80-kDa protein with the indicated concentrations of substrate in 100 mM Tris-HCl (pH 7.0) and 4 mM EGTA in a final volume of 100 l. After an incubation at 37°C for 30 s, the reactions were quenched by the addition of 100 l of 1-butanol and radiolabeled lysophospholipids and free fatty acids were quantitated by thin layer chromatography (Whatman Silica Gel 60A plates) utilizing a mobile phase consisting of 65:25:5 (chloroform:methanol:ammonium hydroxide (v/v)) or 80:20:1 (petroleum ether:diethyl ether:acetic acid (v/v)), respectively. Radiolabeled fatty acid and lysophospholipid were identified by staining of a standard fatty acid or lysophospholipid sample with iodine vapor, scraped, and quantified by scintillation spectrometry. Assays of lysophospholipase activity were conducted as described above except with L-1-[1-14 C]palmitoyl lysophosphatidylcholine as substrate. For studies employing BEL, enzyme samples were preincubated with 10 M BEL or ethanol vehicle alone for 5 min on ice prior to quantitating calcium-independent phospholipase A 2 activity.
RT-PCR Amplification of the 80-kDa Calcium-independent Phospholipase A 2 from CHO Cells-DUKX B1 CHO cells were grown to confluence in minimal essential media, washed in PBS, and total cellular RNA was prepared by guanidiline thiocyanate:phenol:chloroform extraction and isopropyl alcohol precipitation (16). cDNA was prepared from total RNA by an oligo(dT) 16 primed reverse transcriptase reaction (Perkin-Elmer) and the 80-kDa calcium-independent phospholipase A 2 DNA sequence (GenBank accession I15470) was amplified by PCR employing a 35-cycle reaction with steps at 62°C for 30 s, 72°C for 2 min, and 94°C for 30 s per cycle and utilizing oligos which flanked the 5Ј and 3Ј coding region, 5Ј-GAATTCATGCAGTTCTTCGGACGCC-3Ј and 5Ј-GAATTCTCAGGGCGACAGCAGCATTTG-3Ј, respectively. The amplified DNA product encoding the calcium-independent phospholipase A 2 (2.2 kb) was subcloned into a pCR-II vector (Invitrogen), sequenced, and subcloned onto either a pET-21a vector for expression in E. coli (17,18) or a pFasBac vector (Life Technologies, Inc.) for generation of baculovirus harboring the coding sequence for calcium-independent phospholipase A 2 (see below) (19). Standard molecular biology procedures were performed according to established methods (20).
Expression of Recombinant 80-kDa Calcium-independent Phospholipase A 2 in E. coli-The 2.2-kb DNA product encoding the 80-kDa polypeptide was subcloned into pET-21a vector, sequenced, and transformed into competent BL21(DE3) or BL21(DE3)pLysS E. coli for expression of the recombinant calcium-independent phospholipase A 2 . E. coli containing the recombinant 80-kDa polypeptide were grown to an OD 600 of 0.6 at 37°C in Luria-Bertani broth and the expression of recombinant protein was induced by the addition of isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM in the culture broth (18). Samples of the E. coli were removed at 30 min, 1 h, 2 h, and 3 h, pelleted, washed, and resuspended in 250 mM sucrose, 25 mM imidazole (pH 7.8), 5 mM KCl, and 1 mM EGTA. BL21(DE3) E. coli were lysed by sonication by 10 ϫ 1-s bursts from a Vibra Cell sonicator while BL21(DE3)pLysS E. coli were frozen at a temperature of Ϫ70°C and thawed on ice prior to sonication. Expression of recombinant 80-kDa calcium-independent phospholipase A 2 activity from E. coli lysates was measured as described above and the production of recombinant protein was analyzed by SDS-PAGE after staining with silver.
Purification of Recombinant 80-kDa Calcium-independent Phospholipase A 2 from Sf9 Cells-Typically, an 80-ml spinner flask of Sf9 cells in supplemented 1 ϫ Grace's media was infected with baculovirus harboring recombinant 80-kDa calcium-independent phospholipase A 2 at an multiplicity of infection of 1.0 for 48 h. The Sf9 cells were pelleted at 500 ϫ g for 10 min, washed once with buffer A (250 mM sucrose, 25 mM imidazole (pH 8.0), 1 mM EGTA), and resuspended in 15 ml of buffer A. The cells were lysed by sonication (10 ϫ 1-s bursts) from a Vibra Cell sonicator and cytosol was prepared by subsequent sequential centrifugations at 10,000 ϫ g for 10 min and 100,000 ϫ g for 60 min. The 100,000 ϫ g supernatant (cytosol) was immediately loaded onto a DEAE-Sephacel column (2.6 ϫ 5 mm), previously equilibrated with buffer A and washed with buffer B (25 mM imidazole (pH 8.0), 1 mM EGTA). Calcium-independent phospholipase A 2 activity was eluted by the application of a linear gradient of 0 -1M NaCl in buffer B. The fractions which possessed calcium-independent phospholipase A 2 activity were pooled, and loaded directly onto a 1-ml ATP-agarose column previously equilibrated with buffer B. The column was sequentially washed with 10 bed volumes of buffer B, 10 mM AMP in buffer B, and buffer B, and phospholipase activity was eluted by application of 1 mM ATP in buffer B.
The ATP-agarose fractions which possessed calcium-independent phospholipase A 2 activity were pooled, adjusted to a final concentration of 5 mM CaCl 2 by addition of a 100 mM CaCl 2 buffer, and loaded onto a 0.5-ml column of calmodulin-Sepharose. Next, the column was washed with 10 column volumes of buffer containing 25 mM imidazole (pH 8.0), 500 M CaCl 2 prior to elution of calcium-independent phospholipase A 2 activity by the application of buffer containing 25 mM imidazole (pH 8.0) and 4 mM EGTA. For studies involving calmodulin-Sepharose chromatography of Sf9 cell cytosol containing recombinant calcium-independent phospholipase A 2 , freshly prepared cytosol was adjusted to 5 mM calcium and chromatographed as described above.
The calmodulin-Sepharose column fractions which possessed recombinant calcium-independent phospholipase A 2 activity were pooled and loaded onto a PC 1.6/5 Mono-Q column previously equilibrated with buffer B. Recombinant calcium-independent phospholipase A 2 was eluted by the application of a linear gradient of NaCl to 1 M in buffer B. Mono-Q column fractions which possessed calcium-independent phospholipase A 2 activity (150 l/fraction) were concentrated to dryness in the presence of 10% SDS by lyophilization, resuspended in 50 l of buffer containing 25 mM Tris-HCl (pH 6.8), 5% glycerol, 0.5% ␤-mercaptoethanol and the proteins were resolved by SDS-PAGE on 10% polyacrylamide gels and stained with silver.
Resolution of Multiple Isoforms of Recombinant Calcium-independent Phospholipase A 2 -Highly purified recombinant calcium-independent phospholipase A 2 isoforms from Mono-Q chromatography (peaks I and II) were individually pooled, diluted 3-fold in buffer B to reduce the NaCl concentration, and resubjected to Mono-Q chromatography as described above.

RESULTS
To investigate the biochemical characteristics of a recently described putative calcium-independent phospholipase A 2 (Ref. 7, GenBank accession I15470), RT-PCR was performed to amplify the cDNA encoding this protein from CHO cell total RNA utilizing oligonucleotide primers corresponding to its 5Ј and 3Ј coding sequence. The 2.2-kb DNA product from the RT-PCR amplification was subcloned into a pCR-II vector and sequenced to authenticate the fidelity of the amplified product. The 2.2-kb DNA product was subcloned into the E. coli expression vector, pET-21a, and competent E. coli were transformed. Although robust amounts of recombinant 80-kDa polypeptide were produced, induced E. coli did not demonstrate additional calcium-independent phospholipase A 2 activity in comparison to control cells. We hypothesized that expression of this polypeptide in the context of a mammalian system was necessary for proper post-translational modifications and/or protein folding for expression of catalytic activity. Accordingly, the 2.2-kb product was subcloned into the baculoviral vector, pFas-Bac, and transformed into competent DH10Bac E. coli cells for helper plasmid-mediated transposition of the recombinant sequence into a bMON14272 bacmid (19). The bMON14272 bacmid was purified from the DH10Bac E. coli and used to infect Sf9 cells for subsequent formation of recombinant baculovirus and expression of recombinant protein as described in detail under "Experimental Procedures." After initial amplification of the recombinant baculovirus harboring the calcium-independent phospholipase A 2 (ϳ1 ϫ 10 8 plaque-forming units/ml), Sf9 cells were infected for 48 h at a multiplicity of infection of 1.0, harvested, lysed by sonication, and calcium-independent phospholipase A 2 activity was quantified. Sf9 cells infected with recombinant 80-kDa polypeptide reproducibly demonstrated 15-fold increases in calcium-independent phospholipase A 2 activity (Fig. 1A) with greater than 80% of the recombinant activity partitioning into the cytosolic compartment (Fig. 1B). Furthermore, treatment of Sf9 cell cytosol which contained recombinant calcium-independent phospholipase A 2 with the mechanism-based inhibitor (E)-6-(bromomethylene)-3-(1naphthalenyl)-2H-tetrahydropyran-2-one (BEL) (21-23) completely ablated the expressed recombinant calcium-independent phospholipase A 2 activity (Fig. 1C).
To demonstrate that the 80-kDa recombinant polypeptide was responsible for catalyzing calcium-independent phospholipase A 2 activity (i.e. the polypeptide was not an activator or cofactor of an endogenous catalytic entity), a strategy was developed to chromatographically purify the recombinant polypeptide to homogeneity. Since the majority of the recombinant phospholipase A 2 activity partitioned into the cytosolic fraction, Sf9 cell cytosol was applied to a DEAE anion exchange column and recombinant calcium-independent phospholipase A 2 activity was eluted with a linear salt gradient. Recombinant phospholipase A 2 activity eluted at ϳ200 mM NaCl (Fig. 2).
Since at least some calcium-independent phospholipases A 2 interact in a highly selective fashion with ATP (8 -12) and calmodulin (13), we attempted to exploit the power inherent in double sequential affinity chromatographies to effect the rapid and facile purification of this recombinant phospholipase A 2 activity to homogeneity. First, fractions containing calciumindependent phospholipase A 2 activity from the DEAE eluent were pooled and immediately loaded onto an ATP-agarose affinity column. After loading, the column was sequentially washed with buffer, buffer containing 10 mM AMP, and finally the recombinant phospholipase A 2 activity was eluted by the application of buffer containing only 1 mM ATP (Fig. 3A). Over 95% of the calcium-independent phospholipase A 2 activity bound to the ATP-agarose column and was quantitatively eluted by the application of 1 mM ATP. Resolution of the proteins from the ATP-agarose column fractions by SDS-PAGE and visualization after silver-staining demonstrated the specificity of the interaction between the 80-kDa recombinant phospholipase A 2 and the ATP-agarose (Fig. 3B). Calcium-independent phospholipase A 2 activity was quantified as described under "Experimental Procedures" and is expressed as [ 14 C]arachidonic acid released from sn-2 radiolabeled phospholipid in nanomole/mg⅐min. B, the lysate from Sf9 cells which was infected with recombinant baculovirus harboring the sequencing encoding calcium-independent phospholipase A 2 was subjected to sequential differential centrifugation at 10,000 ϫ g for 20 min and 100,000 ϫ g for 60 min. Calcium-independent phospholipase A 2 activity in 25 l of the lysate (sample 1), the 10,000 ϫ g supernatant (sample 2), 100,000 ϫ g supernatant (sample 3), and the 100,000 ϫ g pellet (sample 4) fractions were assessed as described above and are expressed as disintegrations/min of Since at least one calcium-independent phospholipase A 2 interacts with calmodulin in a calcium sensitive fashion, we attempted to exploit the power of ternary complex affinity chromatography employing a calmodulin-Sepharose stationary phase (13). The fractions containing recombinant calcium-independent phospholipase A 2 activity from the ATP-agarose column were pooled, adjusted to 5 mM calcium, and directly loaded onto a calmodulin-Sepharose column as described under "Experimental Procedures." Recombinant calcium-independent phospholipase A 2 activity quantitatively bound to calmodulin-Sepharose in the presence of calcium ion and was quantitatively eluted by dispersal of the ternary complex with application of buffer containing EGTA (Fig. 4A). SDS-PAGE of fractions from calmodulin-Sepharose chromatography displayed only a single intense band after silver staining demon-strating that recombinant calcium-independent phospholipase A 2 activity copurified with the 80-kDa protein (Fig. 4B). Moreover, the resolving power of ternary complex affinity chromatography employing the calmodulin-Sepharose stationary phase was demonstrated by additional experiments in which crude Sf9 cytosol was loaded directly onto a calmodulin-Sepharose affinity column in the presence of calcium. Phospholipase A 2 activity was quantitatively bound to calmodulin-Sepharose in the presence of calcium ion and was quantitatively eluted by washing with EGTA resulting in a 10-fold purification step (Fig. 4, panels C and D).
Mono-Q chromatography of the calmodulin-Sepharose eluent demonstrated an elution profile identifying the presence of multiple isoforms of the 80-kDa polypeptide which each cochromatographed with phospholipase A 2 enzymic activity (Fig. 5). The majority of recombinant phospholipase A 2 activity eluted at ϳ50 mM NaCl with a specific activity of 1 mol/mg⅐min while a second, more anionic, peak eluted at ϳ120 mM NaCl with a specific activity of 5 mol/mg⅐min. Each of three early eluting peaks as well as the late eluting peak contained calcium-independent phospholipase A 1 , phospholipase A 2 , and lysophospholipase activities in similar ratios. Peaks I and II were individually pooled, diluted, and subsequently rechromatographed on a re-equilibrated Mono-Q stationary phase demonstrating that each peak chromatographed according to its elution profile during the initial chromatography (i.e. rechromatography of peak I resulted in a single peak eluting at 50 mM NaCl while rechromatography of peak II resulted in a single peak eluting at 120 mM NaCl) (Fig. 6). These results demonstrate that the isoforms are long-lived entities and not the result of a rapidly equilibrating mixture.
To examine the kinetic characteristics of the phospholipase A 1 , phospholipase A 2 , and lysophospholipase activities catalyzed by the recombinant 80-kDa polypeptide, substrate-activity profiles were compared. First, phospholipase A 1 , phospholipase A 2 , and lysophospholipase activities were linear with respect to time over the incubation times utilized (30 s). Second, each of the activities displayed saturation kinetics (Fig. 7,  A and B). Third, phospholipase A 1 activity (as assessed by the production of sn-2 labeled lysophospholipid from the substrate L-␣1-palmitoyl-2- with an apparent K m of 3.3 M. Fatty acid production catalyzed by the recombinant 80-kDa polypeptide (which reflects both direct phospholipase A 2 catalyzed release of the sn-2 radiolabeled arachidonic acid moiety as well as fatty acid generated through sequential phospholipase A 1 and lysophospholipase activities) demonstrated an apparent maximum velocity of 0.9 mol/mg⅐min and an apparent K m of 1.6 M. We point out that a substantial amount of fatty acid release could result from sequential phospholipase A 1 and lysophospholipase activities. Incubation of enzyme with L-␣1-O-hexadecyl-2-[ 3 H]arachidonyl phosphatidylcholine demonstrated the expressed 80-kDa polypeptide possessed a phospholipase A 2 activity of 0.5 mol/ mg⅐min. Since the sn-1 alkyl ether linkage is not susceptible to hydrolysis by esterolytic processes, these results unambiguously demonstrate phospholipase A 2 activity as an inherent catalytic function of the expressed polypeptide. Fourth, kinetic analysis of lysophospholipase activities demonstrated that the enzyme rapidly catalyzes hydrolysis of monomeric lysophosphatidylcholine (critical micellular concentration of palmitoyl lysophosphatidylcholine ϭ 7 M (24)) with only modest in-creases of enzymic activity at supramicellar concentrations of substrate.

DISCUSSION
The present results demonstrate that the recombinant chromatographically pure 80-kDa polypeptide catalyzes phospholipase A 1 /A 2 and lysophospholipase activities when expressed in a baculoviral expression system. Moreover, individual isoforms of the 80-kDa polypeptide can be chromatographically resolved and calcium-independent phospholipase A 1 /A 2 and lysophospholipase activities comigrate with each peak of 80-kDa protein mass. Collectively, these results identify the 80-kDa polypeptide as a catalytic entity mediating phospholipolysis. Furthermore, they underscore the necessity of expression of this protein in the context of a mammalian cell since E. coli expressed robust quantities of recombinant protein which did FIG. 5. Mono-Q chromatography of recombinant calcium-independent phospholipase A 2 . A, recombinant calcium-independent phospholipase A 2 activity from the calmodulin-Sepharose column eluent (10 g of protein) was loaded onto a PC1.6/5 Mono-Q column previously equilibrated with buffer B. After washing, recombinant calcium-independent phospholipase A 2 activity was eluted by the application of a linear gradient of NaCl in buffer B. UV absorption at 280 nm (-) and NaCl concentration (---). B, fractions from the Mono-Q column elute were dried in the presence of 10% SDS, reconstituted in 50 l of 50 mM Tris-HCl (pH 6.8), 5% glycerol, and 0.5% ␤-mercaptoethanol and the proteins were resolved by SDS-PAGE on 10% polyacrylamide gels and stained with silver as described under "Experimental Procedures." The calciumindependent phospholipase A 2 activity in each fraction was measured as described under "Experimental Procedures" and is expressed in dintegrations/min of [ 14 C]arachidonic acid released from 5 M L-␣1-palmitoyl-2-[1-14 C]arachidonyl phosphatidylcholine and shown on the top of the gel.
FIG. 6. Rechromatography of previously resolved recombinant calcium-independent phospholipase A 2 isoforms on a Mono-Q stationary phase. A, purified recombinant calcium-independent phospholipase A 2 from the calmodulin-Sepharose eluent (5 g of protein) was purified on a Mono-Q column and resolved into earlier eluting (peak I) and later eluting (peak II) peaks as described under "Experimental Procedures." B, peak I from the Mono-Q chromatography in A was diluted 3-fold in buffer B to reduce the NaCl concentration and reapplied to Mono-Q resin and rechromatographed employing identical conditions as above. C, peak II from three iterative preparations similar to those represented in A was diluted 3-fold in buffer B to reduce the NaCl concentration and resubjected to Mono-Q chromatography as described above.

FIG. 7.
Purified recombinant 80-kDa polypeptide catalyzes calcium-independent phospholipase A 1 /A 2 and lysophospholipase activities. A, calcium-independent phospholipase A 1 /A 2 activities were measured from 120 ng of highly purified recombinant 80-kDa polypeptide (i.e. calmodulin-Sepharose eluent) by quantitating the release of radiolabeled arachidonic acid (ϩ) and sn-2 radiolabeled lysophospholipid (᭛) from the indicated concentrations of sn-2 radiolabeled L-␣1-palmitoyl-2-[1-14 C]arachidonyl phosphatidylcholine after incubation at 37°C for 30 s as described under "Experimental Procedures." The production of radiolabeled lysophospholipid from L-␣1-palmitoyl-2-[1-14 C]arachidonyl phosphatidylcholine represents phospholipase A 1 activity while the production of radiolabeled free fatty acid represents combined phospholipase A 2 activity as well as sequential phospholipase A 1 and lysophospholipase activities. B, calcium-independent lysophospholipase activity was measured from 120 ng of highly purified recombinant 80-kDa polypeptide (i.e. calmodulin-Sepharose eluent) by measuring the release of radiolabeled [ 14 C]palmitic acid from the indicated concentrations of sn-1 radiolabeled palmitoyl lysophosphatidylcholine after incubation at 37°C for 30 s as described under "Experimental Procedures." not possess either inherent or inducible (in our hands) catalytic activity.
The recombinant protein expressed calcium-independent phospholipase A 1 /A 2 and lysophospholipase activities in similar amounts. Potentially, two mechanisms can be responsible for the sn-2 fatty acid release from choline glycerophospholipids including: 1) the sequential hydrolysis of the sn-1 acyl group followed by lysophospholipase activity; or 2) direct hydrolysis of the sn-2 acyl group and the concomitant generation of sn-1 acyl-lysophospholipids. Both mechanisms likely contribute to the production of radiolabeled free fatty acid from specifically radiolabeled sn-2 L-␣1-palmitoyl-2-[1-14 C]arachidonyl phosphatidylcholine. We point out the potential possibility that all of the observed release of free fatty acid could result from sequential phospholipase A 1 and lysophospholipase activities since the recently generated sn-2 labeled lysophospholipid is present at the active site of the enzyme and could serve as the preferred substrate for a second round of hydrolysis. In this paradigm, the relative fractional percentage of radiolabeled fatty acid to lysophospholipid generated reflects the relative rates of a second round of enzymatic cleavage in comparison to the rate of release of the radiolabeled lysophospholipid bound at the active site. However, the recombinant protein has substantive amounts of phospholipase A 2 activity since L-␣1-Ohexadecyl-2-[ 3 H]arachidonyl phosphatidylcholine (where the sn-2 acyl group is the only acyl group which can be hydrolyzed by this enzyme) is a good substrate. Finally, the rate of lysophospholipase activity was similar to that of phospholipase A 2 activity, suggesting the similar interactions of the carboxyl moiety destined for hydrolysis with critical amino acids at the active site are present (i.e. the activation energies are similar for the hydrolysis of both substrates).
There are several features of the purification strategy which merit consideration. First, the utilization of tandem sequential affinity columns facilitates the procurement of geometric increases in the high purification factors typically obtained through affinity chromatographic approaches. Second, ternary complex affinity chromatography with calmodulin-Sepharose resin facilitates the discrimination not only between entities which can recognize calmodulin but also between those proteins which possess an obligatory requirement for binding to the calcium-calmodulin complex. Third, the ATP-agarose affinity chromatography exploited the molecular recognition of ATP by the protein in the context of an anionic stationary phase. Since the proteins had been previously selected for binding to a cationic stationary phase (DEAE-Sephacel resin), the subsequent forced affinity binding to a negatively charged affinity column further amplified the power and selectivity of ATP affinity chromatography. The integration of these three key principles in the purification strategy allowed the complete purification of this polypeptide in 36 h in Ͼ56% overall yield (Table I) while prior attempts employing conventional chromatographic strategies have resulted in technically demanding approaches accompanied by poor yields of inhomogenous preparations.
The separation of the 80-kDa phospholipase A 2 into distinct chromatographically resolvable isoforms with differing specific activities was unanticipated. Rechromatography of constituent isoforms eluting with their original chromatographic profiles demonstrated that these isoforms are isolatable entities (on a laboratory time scale) and are not the result of a dynamic interchange of an equilibrium mixture. It is intriguing to speculate that the higher specific activity isoform of this polypeptide has a higher phosphorylation state giving rise to an increased retention time on an anionic exchange resin and a higher specific activity. Whatever covalent modifications are eventually determined to be responsible for these effects, the results suggest a potential biochemical mechanism for agonistinduced increases in calcium-independent phospholipase A 2 activity whereby a high specific activity isoform can be generated from a lower specific activity pool during cellular stimulation.
Recently, studies of the crystal structure of phospholipase C␦ identified the unanticipated finding of an EF-hand calcium binding motif (residues 133-281) which shares substantial homology with calmodulin (25). The present results demonstrate a direct interaction between calmodulin and the recombinant 80-kDa phospholipase A 2 . Accordingly, it is intriguing to speculate that the phospholipase C␦ structure recapitulates the bimolecular calmodulin-phospholipase A 2 complex demonstrated herein, suggesting an ancestral relationship between these two regiospecific phospholipases. Although comparisons of the primary sequences of phospholipase C␦ and calciumindependent phospholipase A 2 have not revealed sequence homology, we anticipate that comparisons of the three-dimensional structure of the recombinant phospholipase A 2calmodulin complex and phospholipase C␦ could show regions of homology and provide insights into the structure of an evolutionarily distant ancestral polypeptide from which both phospholipases are derived. Calcium-independent phospholipase A 2 activity in the Sf9 cell cytosol and eluents from the DEAE-Sephacel, ATP-agarose, and calmodulin-Sepharose columns was measured by quantitating the release of radiolabeled fatty acid from 5 M L-␣1-palmitoyl-2-[1- 14