Purification of a novel phospholipase A2 from bovine seminal plasma.

Phospholipases A2 are enzymes believed to play important roles in numerous physiological systems including sperm cell maturation. Relatively little work has, however, been devoted to study these enzymes in seminal plasma. We therefore undertook the purification and characterization of this enzyme from bovine seminal plasma. After a 330-fold purification, an activity corresponding to a protein of 100 kDa was identified by gel filtration. SDS-polyacrylamide gel electrophoresis analysis of the purified fraction revealed the presence of a 60-kDa band that comigrated with the activity during ion-exchange and gel filtration chromatography as well as polyacrylamide gel electrophoresis. The enzyme possessed a pH optimum around pH 6.5 and was calcium-dependent. Using isoelectric focusing, its isoelectric point was determined to be 5.6 +/- 0.07. The enzymatic activity was resistant to p-bromophenacyl bromide, but was sensitive to gossypol and dithiothreitol. The enzyme was 2 orders of magnitude more active toward micelles formed with deoxycholate than with Triton X-100. Slight differences in the specificity toward head groups and/or sn-2-side chains were found in both assay systems. The enzyme was acid-labile and did not display affinity for heparin. It would therefore appear that the phospholipase A2 form isolated from bovine seminal plasma is of a novel type.

Phospholipases A 2 are enzymes believed to play important roles in numerous physiological systems including sperm cell maturation. Relatively little work has, however, been devoted to study these enzymes in seminal plasma. We therefore undertook the purification and characterization of this enzyme from bovine seminal plasma. After a 330-fold purification, an activity corresponding to a protein of 100 kDa was identified by gel filtration. SDS-polyacrylamide gel electrophoresis analysis of the purified fraction revealed the presence of a 60-kDa band that comigrated with the activity during ion-exchange and gel filtration chromatography as well as polyacrylamide gel electrophoresis. The enzyme possessed a pH optimum around pH 6.5 and was calciumdependent. Using isoelectric focusing, its isoelectric point was determined to be 5.6 ؎ 0.07. The enzymatic activity was resistant to p-bromophenacyl bromide, but was sensitive to gossypol and dithiothreitol. The enzyme was 2 orders of magnitude more active toward micelles formed with deoxycholate than with Triton X-100. Slight differences in the specificity toward head groups and/or sn-2-side chains were found in both assay systems. The enzyme was acid-labile and did not display affinity for heparin. It would therefore appear that the phospholipase A 2 form isolated from bovine seminal plasma is of a novel type.
Phospholipases A 2 (PLA 2 ) 1 are ubiquitous enzymes capable of hydrolyzing the sn-2-position of phospholipids. Most PLA 2 characterized to date belong to either one of two main groups: high and low molecular mass PLA 2 (1,2). High molecular mass PLA 2 , also called cytoplasmic PLA 2 (cPLA 2 ), are 85-kDa proteins found in the cytoplasm of several cell types (3)(4)(5)(6). They are specific for arachidonic acid (6) and possess limited lysophospholipase (7,8) and phospholipase A 1 (9) activities. Low molecular mass PLA 2 (sPLA 2 ) form a family of homologous enzymes with molecular masses ranging from 14 to 20 kDa that are found in several secretory fluids as well as in the cytoplasm of various cell types (1,2,10). PLA 2 are believed to be important regulatory enzymes in numerous physiological systems such as inflammation, membrane remodeling, and cell signalization (11). Several PLA 2 that do not belong to either category have also been identified in various tissues and organisms (12)(13)(14)(15)(16)(17)(18)(19).
In the reproductive system, PLA 2 are widely accepted to play a major role in the late maturational events of spermatozoa, particularly in the acrosomal reaction (20 -23). The acrosomal reaction is a multifusion process that permits the release of hydrolytic enzymes, which are required for spermatozoa to penetrate the acellular layers surrounding the oocyte (24).
Although several studies have been undertaken to characterize the PLA 2 present in the spermatozoa and seminal plasma of various species (25)(26)(27)(28)(29)(30), only the enzyme from human seminal plasma has been purified to homogeneity and sequenced (31) so as to conclusively assign it to a particular PLA 2 group. The enzyme was found to be a 14-kDa protein, identical to the synovial enzyme (32), suggesting the same might be true of other mammalian species.
In bovine seminal secretions, the enzyme was partially purified, but was not characterized enough to assign it to a particular PLA 2 group (30). To determine the exact type(s) of PLA 2 present in bovine seminal plasma and to assess the generality of the occurrence of sPLA 2 in mammalian seminal plasma, we purified and characterized the major PLA 2 activity from bovine seminal plasma.
Tris-HCl, 0.02% NaN 3 , pH 7.4) containing 10 mM sodium deoxycholate. The substrate solution was vortexed and mixed for 20 min. Ten l of substrate solution was added to each assay tube. A typical reaction mixture (final volume of 100 l) consisted of 1 mM CaCl 2 and 1 mM sodium deoxycholate in buffer A. After 30 min at 37°C, the reaction was stopped by adding 200 l of chloroform/methanol (2:1) containing 2 g/ml fatty acid tracer and 50 l of 4 M KCl. The assay tubes were then centrifuged, and the lower phase was applied onto a silica TLC plate, which was then developed in petroleum ether/ether/acetic acid (85:15: 1). The fatty acids were visualized with iodine, and the stained spots were cut into scintillation vials. The scintillation fluid was then added, and the radioactivity was determined in a liquid scintillation counter.

Purification Methods
Seminal Plasma Preparation-Pools of bovine ejaculates were centrifuged at low speed (300 ϫ g) to remove spermatozoa. The supernatant was then preserved at Ϫ20°C and used for purification within 2 weeks.
Butyl-Sepharose Chromatography-Ten ml of frozen seminal plasma was thawed, adjusted to 0.1 M choline chloride, and centrifuged at 10,000 ϫ g, and the supernatant was loaded (2 ml/min) on a 2.5 ϫ 10-cm butyl-Sepharose column equilibrated in buffer A containing 0.1 M choline chloride. The column was then washed at 7 ml/min with 700 ml of equilibration buffer followed by 350 ml of 5 M urea in buffer A (Fraction I).
Sephacryl S-300 Chromatography-Fraction I was concentrated by ultrafiltration (pore size of 10,000; Amicon, Inc.) and applied to a 1.5 ϫ 110-cm Sephacryl S-300 column (4°C) equilibrated in buffer A containing 0.15 M choline and 0.15 M NaCl. Fractions (5.8 ml) were collected at a flow rate of 0.3 ml/min. The fractions under the activity peak were pooled and concentrated (Fraction II). Calibration of the column was performed under the same conditions by passing RNase A, ovalbumin, and bovine serum albumin.
Q-Sepharose Chromatography-Fraction II was applied to a Q-Sepharose column (1 ϫ 1 cm) coupled to a fast protein liquid chromatography system and equilibrated in buffer A (without NaN 3 ) containing 0.2 M NaCl. The active fractions were eluted with a 0.2-1 M NaCl gradient in buffer A. Fractions (1 ml) were collected at a flow rate of 1 ml/min.

Characterization
A partially purified (190-fold) fraction, obtained by an alternative lower yield approach, was preserved at Ϫ20°C in 25% glycerol and used for all characterization studies unless otherwise specified. The substrate used was arachidonyl-PE unless specified otherwise.
Isoelectric Focusing-Isoelectric focusing was performed at 22°C for 7000 V-h on a post Sephacryl S-300 aliquot adjusted to 5 M urea and 2% ampholytes. The gel rods (0.3 ϫ 13 cm) consisted of 4% acrylamide, 2% ampholytes, pH 3-10, 2% Triton X-100, and 5 M urea. After completion of the electrophoresis, the gel rod was cut into 24 pieces, and proteins were eluted in 500 l of H 2 O/piece at 4°C for 16 h on an orbital shaker.
Inhibition Studies-For inhibition studies, PLA 2 -containing fractions were preincubated with the indicated concentrations of inhibitor dissolved in dimethyl sulfoxide (pBPB and gossypol) or H 2 O (dithiothreitol (DTT)) for 3 h (pBPB) or for 30 min (gossypol or DTT) at 37°C in buffer A. The sample was diluted 10 times prior to the enzymatic assay so that the final concentration of dimethyl sulfoxide in the assay tube was 1%.
Protein Estimation-During purification, protein concentration in each fraction was estimated by monitoring the absorbance at 280 nm. Protein content in pooled fractions was determined according to Bradford (33).

SDS-PAGE and Related
Techniques-SDS-polyacrylamide gel electrophoresis (PAGE) was performed essentially as described by Laemmli (34). PAGE was performed on a 6% gel according to Kramer et al. (35). The apparent molecular mass of the various protein bands was determined with the low molecular mass calibration kit from Pharmacia Biotech. Proteins were visualized using Coomassie Brilliant Blue R-250 (36).

Purification of Bovine Seminal PLA 2
Seminal plasma was first passed through a butyl-Sepharose resin (Fig. 1a). Extensive washing (14 column volumes) was required to remove all the weakly adsorbed proteins. The urea-  1. Purification scheme. a, butyl-Sepharose chromatographic pattern; b, Sephacryl S-300 chromatographic pattern; c, Q-Sepharose chromatographic pattern. Chromatography was performed as described under "Experimental Procedures." The fractions under the bar were pooled and assayed for purity as indicated in Table I. Where appropriate, the gradients used are indicated. The arrow in a indicates the point of addition of the 5 M urea buffer. In b, the approximate void volume of the Sephacryl S-300 column is indicated by the arrow. q, absorbance; E, PLA 2 activity. desorbed fractions (Fraction I) contained most of the recovered activity. Fraction I was concentrated and loaded onto a Sephacryl S-300 gel sieving column (Fig. 1b). A single active peak was obtained whose elution position corresponded to the behavior of a 100-kDa protein as determined by calibration of the column. The active peak was then concentrated and applied onto a Q-Sepharose ion exchanger. The activity was again eluted in one major activity peak, which well overlapped the protein pattern (Fraction III).
The purification results are summarized in Table I. This scheme resulted in a purification of 330-fold with a 45% recovery of the activity.

Characterization of Bovine Seminal PLA 2
The Purified Enzyme Behaves as a 60-kDa Protein on SDS-PAGE-The active fractions from the successive purification steps were analyzed by SDS-PAGE (Fig. 2)  The 60-kDa Band Is Responsible for the Activity-Fraction III was subjected to PAGE. Measurement of the activity eluted from the gel slices revealed that it was recovered at a position corresponding to the protein (Fig. 3).
Calcium Requirement and pH Optimum-In a manner similar to most phospholipases characterized thus far, the enzyme was calcium-dependent and was maximally active at ϳ2 mM calcium (Fig. 4a), while analysis of the pH dependence of the activity revealed a single activity maximum at pH 6.5 (Fig. 4b).
Sensitivity of Bovine Seminal PLA 2 to Known PLA 2 Inhibitors-Purified PLA 2 was resistant to pBPB, whereas the two positive controls, porcine pancreatic and C. atrox PLA 2 , were inhibited (Fig. 5a). Seminal PLA 2 was inhibited by gossypol at inhibitor concentrations higher than those required to inhibit crotal PLA 2 , but similar to those required to inhibit the porcine pancreatic enzyme (Fig. 5b). The porcine enzyme and seminal PLA 2 also shared similar sensitivities to the thiol reagent DTT (Fig. 5c); the sensitivity of the crotal enzyme toward DTT was not investigated in this study.
Determination of the Enzyme pI-To determine the pI of PLA 2 , isoelectric focusing of a partially purified enzyme was performed (Fig. 6). The gel rod was cut into 24 pieces, which were then left to elute in H 2 O. The supernatants were assayed for PLA 2 activity, and their pH was measured. Several (n ϭ 8) such experiments revealed a single activity peak at pH 5.6 Ϯ 0.07 (mean Ϯ S.E.). Typical activity recoveries on the order of 10 -20% were obtained. The true recovery is expected to be higher since the Triton X-100 concentration in the supernatants (ϳ0.001% final concentration) inhibited the activity of a partially purified fraction by ϳ50% (data not shown).
Substrate Specificity of Seminal PLA 2 -The substrate specificity was studied in the presence of phospholipid micelles consisting of either PC or PE and deoxycholate or Triton X-100. As summarized in Table II, PLA 2 was 2 orders of magnitude more active toward the deoxycholate-containing substrate than toward the Triton X-100-containing substrate or vesicular substrate (data not shown). In the presence of deoxycholate, the enzyme discriminated between the sn-2-fatty acid as it was less active toward PC carrying linoleoyl (1111 Ϯ 98) than arachi- donyl (1716 Ϯ 73). For a given sn-2-side chain, no selectivity was observed between PE/deoxycholate-or PC/deoxycholatecontaining micelles as both substrates were hydrolyzed at similar rates, suggesting that the enzyme shows little, if any, head group specificity in this assay system. When micelles comprising Triton X-100 were used, however, head group specificities were observed. The ethanolamine phospholipid was cleaved more efficiently than the corresponding choline phospholipid (56 Ϯ 5.8 versus 17 Ϯ 1.5), although the total amount hydrolyzed remained much lower than when deoxycholate was present. Interestingly, the side chain specificity observed with deoxycholate-containing micelles was reversed when Triton X-100 micelles were used, as linoleyl was then preferred over arachidonyl (48 Ϯ 1.7 versus 17 Ϯ 1.5).

DISCUSSION
The seminal PLA 2 activity bound specifically to the butyl-Sepharose resin, thus permitting a 130-fold purification in a single step. Choline had to be included throughout this step to prevent the heparin-binding proteins, the main component of bovine seminal plasma (37), from strongly binding to the resin.
Rechromatography of the unadsobed fraction did not permit further binding of the activity, thus suggesting the presence of another form of PLA 2 , which was not further investigated in this study. Chromatography on both gel filtration and ionexchange resins (Fig. 1, b and c) resulted in activity and protein absorbance patterns that eluted closely together, indicating that the major protein (absorbance at 280 nm) was also responsible for the activity. When analyzed by SDS-PAGE and stained with Coomassie Brilliant Blue, a major 60-kDa band was visible in both chromatographic runs (Fig. 2). Further confirmation that the 60-kDa band was responsible for the activity was obtained by PAGE. Fractions that consisted of eluates of gel slices were assayed for PLA 2 activity, and again, the band intensity and the corresponding enzymatic activity variations matched closely (Fig. 3). Gel filtration revealed that the activity behaved as a 100-kDa protein (Fig. 1b), whereas SDS-PAGE analysis showed a 60-kDa band (Fig. 2). This discrepancy might be attributed to dimerization of the 60-kDa enzyme. This dimer appears stable since moderately stringent conditions (0.1% deoxycholate or 5 M urea) failed to shift the elution position of PLA 2 (data not shown). Since the omission of 2-mercaptoethanol did not change its behavior on SDS-PAGE (data not shown), it appears that the interaction is noncovalent. Consistent with the dimer hypothesis, the enzyme behaves on native PAGE as a much larger protein than bovine serum albumin despite a very similar pI (Fig. 3).
Binding to Q-Sepharose at pH 7.4 (Fig. 1c) as well as isoelectric focusing (Fig. 6) indicate that the enzyme is acidic. In comparison, most mammalian sPLA 2 are neutral to basic proteins, with one notable exception (10). cPLA 2 , on the other hand, possess pI values similar to those of the seminal enzyme ( Fig. 6) (8, 35). Besides this similarity, however, the seminal enzyme shares little in common with cPLA 2 . Using two different assay systems, the seminal enzyme did not show the char-  6. Isoelectric focusing of seminal PLA 2 . A partially purified aliquot was adjusted to 5 M urea and 2% ampholytes, applied onto an isoelectric focusing polyacrylamide gel rod, and subjected to isoelectric focusing. The gel rod was then cut into 24 equally sized pieces, which were incubated for 16 h in distilled water. The PLA 2 activity content of the supernatant was then determined. The pH in each fraction was measured using a glass electrode. q, pH; E, PLA 2 activity.
FIG. 5. Sensitivity of the seminal enzyme to PLA 2 inhibitors. Seminal (ϫ), crotal (E), and porcine pancreatic (q) PLA 2 were pretreated with the indicated PLA 2 inhibitors as described under "Experimental Procedures." The PLA 2 activity of an aliquot (10-fold diluted) that underwent the appropriate treatment was then assayed over a 30-min period. The data shown here represent the mean of three independent determinations. For clarity, the standard error is shown only in c; for a and b, it was typically below 10% and never above 15% of the corresponding mean.
acteristic specificity for arachidonylphospholipids found in high molecular mass PLA 2 . In the Triton X-100 assay system, the seminal plasma PLA 2 activity toward sn-2-arachidonyl was ϳ3-fold lower than the activity toward linoleyl, whereas cPLA 2 , in a similar assay system, displayed a 3-fold higher activity (5). Moreover, while cPLA 2 is inhibited by deoxycholate micelles relative to sonicated vesicles (4), the reverse is observed for the seminal enzyme (data not shown).
The resistance of the enzyme to pBPB supports the view that this enzyme is novel. pBPB inactivates sPLA 2 by alkylating a histidine residue located in the active site of the enzyme (38). It also inactivates cPLA 2 (39) by an unknown mechanism, which is likely to be quite different from sPLA 2 since cPLA 2 does not possess an active-site histidine (3). At the pBPB concentrations used, both enzyme types should be inactivated, and yet, the seminal enzyme is unaffected. As expected, the two PLA 2 controls, the type I porcine pancreatic and the type II C. atrox enzymes, were inactivated (Fig. 5a). The greater resilience of the crotal enzyme is most likely due to its tendency to shield its active site through dimerization (40,41). This raises the possibility that seminal PLA 2 possesses a histidine or some other susceptible residue in its active site, which would be completely shielded from the environment in the absence of substrate and/or Ca 2ϩ .
Despite this resistance, some common structural features between pancreatic and seminal PLA 2 are suggested by the inhibition patterns of DTT and gossypol. The pancreatic enzyme is inhibited by gossypol at concentrations very close to those required to inhibit the porcine enzyme (Fig. 5b). Although the precise structural modifications induced by gossypol are unknown, the similar concentrations required to inhibit pancreatic PLA 2 and the seminal enzyme suggest some common structural elements. This resemblance appears to be quite specific as the inhibition pattern of the crotal enzyme, which shares strong structural homologies with the pancreatic enzyme (1,2), is completely different. The shared DTT sensitivities (Fig. 5c) further support the view that common features between mammalian sPLA 2 and seminal PLA 2 exist. Biochemical characterization revealed that seminal PLA 2 shows catalytic properties common to most sPLA 2 identified so far: the enzyme is Ca 2ϩ -dependent (Fig. 4a) and is optimally active in the neutral to alkaline pH range (Fig. 4b) (42).
The substrate selectivity profile of purified PLA 2 is also reminiscent of mammalian sPLA 2 (43,44). For instance, these enzymes are activated by the introduction of negative charges (as with deoxycholate versus Triton X-100) in the lipid substrate, most likely due to the accumulation of positive charges near the phospholipid-binding site (45). In the absence of deoxycholate, for a given acyl side chain, they are more active toward the anionic phospholipid PE than toward the zwitterionic phospholipid PC (43).
Beside these catalytic similarities, major structural differences appear to exist between these enzymes. For instance, mammalian sPLA 2 are low molecular mass (14 -20 kDa) and mostly basic proteins, whereas the seminal enzyme possesses a 60-kDa mass and an acidic pI. Pancreatic PLA 2 and the human seminal/synovial enzyme demonstrate affinity for heparin (46 -48), while bovine seminal PLA 2 does not (data not shown). Moreover, sPLA 2 are resistant to acidic conditions as relatively good recoveries are routinely obtained following chromatography performed under acidic conditions (49 -51), whereas the major PLA 2 activity found in seminal plasma is acid-labile (data not shown).
The seminal enzyme displays a specific activity (under suboptimal conditions) of ϳ0.01 mol/min/mg, which is rather low compared with that of low molecular mass PLA 2 (for instance, ϳ40 and 1500 mol/min/mg for bovine pancreatic and Naja naja venom PLA 2 , respectively) or with that of cPLA 2 (ϳ0.6 mol/min/mg) (4). The activity range of these well characterized PLA 2 thus covers 5 orders of magnitude. The resistance of seminal PLA 2 to pBPB (Fig. 5a) might indicate that it acts via a different, less efficient catalytic mechanism than the established enzymes. The lower catalytic efficiency of bovine seminal PLA 2 could be required for its proper function in seminal plasma. Alternatively, it could possess some yet undetermined advantages over other types of PLA 2 that would render it better suited to the particularity of the bovine reproductive physiology.
These results differ significantly from those reported previously concerning bovine (30) or human (28,29,31,47,52) seminal plasma PLA 2 . The major human seminal plasma PLA 2 has been found to be identical to the synovial enzyme (31,32). A minor form that was not recognized by the anti-synovial PLA 2 antibody was also reported (31). In the bovine species, the preliminary characterization of the enzyme published previously (30) did not permit any definitive conclusions to be drawn as to the nature of the seminal enzyme. Two different enzymatic activities were partially purified from seminal vesicle secretions. SDS-PAGE of the most purified fraction showed a doublet migrating as 14 -16-kDa proteins. This enzyme may represent a minor PLA 2 form. The human prostate enzyme has also been partially purified and characterized (53). Overall, its biochemical properties appear to be quite distinct from those of bovine seminal plasma PLA 2 .
The activities found in bovine, ram, and porcine seminal plasma amount to ϳ1, 10, and 0.03%, respectively, of the human seminal plasma PLA 2 activity (29), suggesting that qualitative differences might exist between the PLA 2 types found in these species. The structural characterization of the enzyme that is currently underway should reveal the reasons behind these differences.