PROTEIN

EXPRESSION

AND PURIFICATION

3,27-35

(19%)

Purification of Recombinant Human Secretory Phospholipase A, (Group II) Produced in Long-Term Immobilized Cell Culture Wayne Levin, * Reynold F. Daniel,* Cheryl R. Stoner,? Timothy J. Staller,? Judith A. Wardwell-Swanson,* Yale M. Angelillo,$ Philip C. Familletti,$ and Robert M. Crowlt *Departments of Protein Biochemistry, tMolecular Genetics, and $Bioprocess Development, Hoffmann-La koche Inc., Nutley, New Jersey 07110 Received

November

7,1991,

and in revised

form

December

lo,1991

Recombinant human secretory phospbolipase A, (Group II) was expressed in long-term culture of immobilized Chinese hamster ovary cells utilizing a continuous-perfusion airlift bioreactor. The bioreactor was continuously perfused with cell-culture medium supplemented with 5% fetal calf serum at an average flow rate of 5 liters/day for 30 days. Recombinant phospholipase A,, at concentrations ranging from 100 to 500 pglliter, was purified to apparent homogeneity by an efficient two-step procedure involving a silica-based cation-exchange resin and hydrophobic interaction chromatography (~65% recovery of phospholipase A,). The purified recombinant protein has an apparent molecular weight of 16 kDa, identical to that of purified human placental or synovial fluid phospholipase Az, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Application of the puritled protein onto several different gel filtration columns resulted in elution of the protein at molecular weights corresponding to 3.14.7 kDa, suggesting an interaction of the protein with the column resins. However, analytical ultracentrifugation experiments revealed that the protein behaves as a monomer (13.8-14.2 kDa) over a protein concentration range of -10 pglml to 5 mg/ml. With autoclaved Escherichia coli membranes as substrate, the recombinant protein has catalytic properties (pH optimum, effects of bovine serum albumin, sodium chloride concentration, and requirement for calcium) similar to those of the protein purified from human placenta. 0 1992 Academic

Press, Inc.

Phospholipases A, (PLA,‘s,’ EC 3.1.1.4) are a family of lipolytic enzymes that catalyze the hydrolysis of the ’ Abbreviations nant PLA,; PBS,

used: PL4, phosphate-buffered

1046-5928192 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

phospholipase saline;

A,; rPLA,, SDS-PAGE,

recombisodium do-

2-acyl ester bond in cellular phospholipids thereby generating lysophospholipids and free fatty acids (1). When the released product is arachidonic acid, this fatty acid can be metabolized to one of several proinflammatory lipid mediators such as prostaglandins and leukotrienes. Lysophospholipids are also cytotoxic and are precursors of a potent chemical mediator, platelet activating factor. Most of these mediators are elevated in a variety of inflammatory fluids, suggesting an involvement of PLA, in the development and/or progression of inflammatory disorders (2,3). Much of the early research on the enzymology and protein structure of PLA, came from the secreted snake venom and pancreatic PLA,‘s. Characterization of the proteins from different sources has revealed a high degree of conservation among the proteins with regard to specificity for hydrolysis at the sn-2 position ofphospholipids, requirement for Ca’+, similar molecular weights (- 14-16 kDa), the presence of a large number of intrachain disulfide bonds, a His in the active site, and sequence homology (1). These proteins have been classified as group I or group II PLA,‘s based on structural features of the enzymes including the location of disulfide bonds in the molecules. Group I PLA,‘s include the mammalian pancreatic enzymes and those from the venoms of elapid and hydrophid snakes, while group II PLA,‘s include the enzymes found in the venoms of crotalid and viperid snakes as well as a mammalian protein found in a number of cell types and inflammatory exudates including human synovial fluid (4,5). Human group II PLA, has been purified from synovial fluid (69), platelets (8), placenta (9), and spleen (lo), but the low levels of protein present in these sources has pre-

decyl sulfate-polyacrylamide albumin; CHO, Chinese

hamster

gel electrophoresis; ovary; CMV,

BSA, bovine cytomegalovirus.

serum

27 Inc. reserved.

28

LEVIN

eluded its purification in quantities sufficient for investigating more thoroughly the structural and functional aspects of the protein. We report here a simple purification procedure for obtaining milligram quantities of human rPLA, (group II) from serum-containing medium of CHO cells transfected with the cDNA for human placental PLA,, MATERIALS

AND METHODS

Materials

The silica-based resins, NuGel P-SP and NuGel P-AF polyhydroxy, were obtained from Separation Industries, Inc. (Metuchen, NJ). Butyl Toyopearl TSK 650s was purchased from TosoHaas (Philadelphia, PA). Sephadex G-50, Superose 12, and Superdex 75 HR lo/ 30 columns were from Pharmacia LKB Biotechnology (Piscataway, NJ). Zorbax GF-250 and TSK-250 HPLC columns were from DuPont Co. (Wilmington, DE) and Bio-Rad Laboratories (Richmond, CA), respectively. Endoglycosidase F/N-glycosidase F was from BoehringerMannheim (Indianapolis, IN) and fatty acid-free bovine serum albumin and tunicamycin were obtained from Sigma Chemical Co. (St. Louis, MO). Iscove’s modified Dulbecco’s medium and heat-inactivated fetal calf serum were obtained from JRH Scientific, Inc. (Lenexa, KS). [35S]Cysteine was from Amersham Corp. (Arlington Heights, IL). Construction

ET

AL.

maintained in Iscove’s modified Dulbecco’s medium containing 10% heat-inactivated fetal calf serum and HT (10e4 M hypoxanthine and 10e5 M thymidine) at 37°C in a humidified atmosphere of 5% CO,. Prior to transfection, pBClB/CMV/PLA, (5 pg) was linearized by digestion with PvuI and mixed with 250 ng of similarly linearized pSV2-DHFR plasmid DNA (14) and 5 pg of high-molecular-weight carrier DNA. CHO/ DHFR- cells (5 X lo5 per loo-mm dish) were transfected by the calcium phosphate co-precipitation method (15) using reagents obtained from Stratagene (La Jolla, CA). The transfected cells were maintained in medium containing HT for 2 days and then subcultured and plated in medium without HT. After 14-16 days, colonies of surviving transfected cells (DHFR+) were observed. The DHFR+ cells were subcultured further in medium containing 1 pM amethopterin. Culture supernatants were analyzed by EIA (16) to quantitiate PLA, levels. After it was confirmed that cells resistant to 1 pM amethopterin were expressing human PLA, at levels higher than the starting population, single cell-derived cultures were generated by seeding 96-well plates at approximately one cell per well. The highest producing clone isolated from a screen on 11 96-well plates was designated CH0/4E9. A second round of cell cloning indicated that 4E9 was indeed a clonal and relatively stable cell line. One isolate from the second cloning procedure, 4E9.2H4, was selected for large-scale production.

of the PLA, Expression Plasmid

The cDNA clone (hpPLA, 9-l) used in these studies for expression of human secretory (group II) PLAz, which was isolated from a human placental cDNA library and subcloned into pGEM-7Zf(f), has been described previously (11). The mammalian expression vector pBC12/CMV/IL-2 has also been described (12). The entire coding region of human secretory PLA,, including the signal peptide coding sequence, was excised from pGEM/hpPLA, as a 887-bp DNA fragment by digestion with the restriction endonucleases HindIII and XhoI. The PLA, cDNA was inserted between the Hind111 andBamH1 sites of pBC12/CMV/IL-2 using T4 DNA ligase, replacing the IL-2 coding region in the original expression plasmid. The XhoI end of the PLA, gene fragment and the BamHI end of the expression vector were made blunt-ended by incorporation of deoxyribonucleotides with DNA polymerase I (Klenow fragment) in order to be ligated to each other, regenerating the XhoI site but not the BamHI site. The resulting expression plasmid was designated pBClB/CMVIPLA,. DNA Transfection Producing PLA,

and Isolation

Long-Term

Production

of PLA,

Clone 4E9.2H4 was grown and expanded in Iscove’s modified Dulbecco’s medium supplemented with 10% (v/v) heat-inactivated fetal calf serum in g-liter spinner flasks. A 3-liter continuous-perfusion airlift bioreactor (Bellco Biotechnology, Inc., Vineland, NJ), equipped with an immobilization matrix designed to increase the cell density of the culture, was used for production of PLA, by techniques previously described (17). The immobilization matrix used for production of PLA, was composed of glass cylinders with a diameter of 0.7 cm and a length of 2.0 cm and provided a cell-growth surface area of 16,500 cm’. The bioreactor was initially seeded with 2 X lo9 2H4 cells and was continuously perfused with IMDM supplemented with 5% (v/v) fetal calf serum (10% serum was used in the initial bioreactor) at an average rate of 5 liters/day for 30 days. After an initial growth period, a concentration of 100 to 500 pg PLA,/liter was obtained from the bioreactor. PLA, levels obtained from the bioreactor were monitored daily by an immunoassay described previously (16).

of Cell Clones

A Chinese hamster ovary cell line (13) which lacks dihydrofolate reductase activity (CHO/DHFR-) was

Purification

of Recombinant

Human

Placental PLA,

All chromatographic steps were performed at 4°C. CHO cell-culture medium (66 liters), adjusted to pH 7.5

PURIFICATION

OF

RECOMBINANT

with dilute NaOH, was loaded onto a 25ml NuGel P-AI? (4.5 X 1.6 cm) precolumn that was connected directly to a 10 ml (2.2 X 2.6 cm) NuGel P-SP cation exchanger at a flow rate2 of 8-10 ml/min. Both columns were equilibrated with 100 mM Tris-HCl buffer (pH 7.5 at 4°C) prior to loading the supernatant. The columns were subsequently washed with 300 ml of 100 mM Tris-HCl (pH 7.5) containing 150 mM NaCl. After removal of the precolumn, the NuGel P-SP column was washed with 25 column volumes of the same Tris buffer containing 300 IIXM NaCl followed by elution of PLA, with Tris buffer containing 800 mM NaCl. The entire 800 mM NaCl fraction (100 ml) was diluted with Hz0 to bring the NaCl to 500 mM. Solid ammonium sulfate was added to 30% saturation, and the sample was applied (2.5 ml/min) to a 12-ml column of TSK Butyl Toyopearl65OS (2.2 X 3.2 cm) previously equilibrated with 50 mM Tris-HCl (pH 7.5) containing 500 mM NaCl and 30% ammonium sulfate. The column was subsequently washed stepwise with the Tris-NaCl buffer containing 30, 20, and 10% ammonium sulfate. The PLA, eluting in the 10% ammonium sulfate fraction was concentrated in an Amicon ultrafiltration cell with a YM-5 membrane. The purified PLA, was dialyzed against 50 mM Tris-HCl (pH 7.5) and stored at 4°C (usually in the presence of 0.02% sodium azide) . Enzyme Assays PLA, activity was measured as described by Rothhut et al. (18) with modification (7). Autoclaved Escherichia coli membranes, which had been labeled with [3H]oleic acid (19), were used as substrate. The assay mixture (50 ~1) contained 100 mM Tris-HCl (pH 8 at 22”C), 150 mM NaCl, 2 mM CaCl,, 250 pg fatty acid-free BSA, substrate (250,000 dpm), and enzyme. After a 4-min incubation at room temperature, 25 ~1 of 1 N HCl was added to stop the reaction followed by an additional 1.5 mg fatty acidfree BSA (in 15 ~1). Samples were centrifuged at 14,OOOg for 5 min, and the enzymatically released [3H]oleic acid in the supernatant was quantified. Enzyme extracts were diluted to ensure that no more than 15% of the labeled substrate was hydrolyzed during the incubations. Gel Permeation

Chromatography

The molecular weight of purified PLA, was determined by gel filtration on Zorbax GF-250 HPLC (two 9.4 mm X 25 cm columns in tandem), TSK-250 (7.5 mm ’ In theory, much higher flow rates could be used with this However, initial experiments established that rPLA, activity pletely stable in culture medium for at least 30 days when 4°C in the presence of 0.02% sodium azide. We decided to columns during the production runs at flow rates that would with termination of the bioreactor.

column, was comstored at load the coincide

PHOSPHOLIPASE

A2

29

x 30 cm) HPLC, Superose 12, Superdex 75, and Sephadex G-50 (2.5 X 35 cm) columns. Chromatography was performed at room temperature. Analytical

Ultracentrifugation

Purified PLA, (3 mg/ml) in either 50 mM Tris-HCl (pH 7.5 at 4°C) or 100 mM Tris-HCl (pH 8 at 4°C) containing 150 mM NaCl and 2 mM CaCl, was subjected to sedimentation equilibrium analysis at the Analytical Ultracentrifugation Facility, University of Connecticut (Storrs, CT). The samples of PLA, were evaluated at different protein concentrations and centrifugation conditions (32-44K rpm) such that data was obtained over a concentration range of approximately IO-30 pg/ ml to 5 mg/ml. Photographs were taken using an automated photosystem every 2 h beginning 16 h after starting the run in order to test for equilibrium. Blanks were run before and after experimental runs in order to correct for optical distortions. Data analysis was performed using a nonlinear least-squares program assuming a simple model (a single species which is either ideal or non-ideal) or more complex models (monomerdimer, monomer-trimer, etc.). Other Assay Methods Protein was determined by the method of Lowry et al. (20) with crystalline BSA as standard. SDS-PAGE was performed in 12.5% polyacrylamide gels (0.75 mm X 15 cm X 14 cm) by the method of Laemmli (21). The gels were stained with Coomassie blue R-250 and destained as described (22). NH,-terminal sequence analysis was performed using an Applied Biosystems gas phase sequencer (Model 470A) (23). Phenylthiohydantoin amino acid derivatives were identified on-line with an Applied Biosystems Model 120A PTH amino acid analyzer as described (24). Circular dichroism spectra were recorded at 22°C on a Jasco J-500A spectropolarimeter. RESULTS

AND

DISCUSSION

Expression of rPLA, in Cultured Mammalian Cells A cDNA clone coding for human placental PLA, (11) was expressed in mammalian cells under the transcriptional control of the CMV promoter using an expression vector used previously for expression of recombinant interleukin-2 (12). Preliminary experiments demonstrated that active PLA2 was synthesized and secreted in COS cells transiently transfected with pBC12ICMVI PLA, 9-l (data not shown). Immunoprecipitation analysis of radiolabeled proteins made in transfected cells confirmed that a protein with the expected molecular weight of human PLA, was synthesized at a substantial rate (11). In order to generate a cell line for long-term production of human PLA,, we employed gene-linked co-amplification techniques using DHFR-deficient

30

LEVIN

ET

TABLE Purification

of Human Total PLA, activity* (units X 10e9)

Fraction CHO culture supernatant NuGel P-SP Butyl Toyopearl Concentration/dialysis

806 762 605 540

rPLA,

(Group

Protein bd

AL. 1

II)

from

CHO

Cell

Medium”

Specific activity (units X 10m9/mg)

352,000 18 13 11

CHO cells for transfection. The PLA, expression plasmid and a plasmid encoding functional DHFR were introduced into CHO cells, and the resulting cell population was subjected to selection for resistance to amethopterin in order to obtain cell isolates in which the number of copies of both DHFR and PLA, are amplified. A clonal cell line, CH0/4E9, was generated which, after 4 days in culture as a confluent monolayer, produced approximately 2 mg PLA,/liter. A culture of CH0/4E9 was expanded and inoculated into a bioreactor for long-term production of rPLA,. After an initial growth period, the culture produced between 100 and 500 pg PLA,/liter for several weeks. of rPLA,

We developed a two-step chromatographic procedure for the purification of human rPLA, from serum-containing tissue culture medium with an overall yield of 67% (Table 1). Despite the fact that expression levels of the protein (-0.25 pg/ml) are relatively low, we were able to take advantage of the high isoelectric point of human PLA, (PI > 9.5) to purify the protein on a cation-exchange resin at pH 7.5. The use of a precolumn was necessary to prevent clogging of the cation-exchange resin. When flow rates started to decrease because of clogging of the precolumn, it was replaced with fresh resin. Filtration of supernatant was not performed prior to loading the columns because we observed significant selective loss of rPLA, on cellulose acetate or polyvinylidene difluoride membranes, despite the fact that high levels of serum proteins (-5 mg/ml) were present in the medium. Essentially quantitative recovery of rPLA, was obtained from the NuGel P-SP column (Table 1) with a purification factor of >20,000-fold. A small additional amount of rPLA, was also recovered (3-5%) in later fractions of the 0.3 M NaCl eluate, after elution of contaminating proteins. This fraction of rPLA, could also

Purification (fold)

0.002 42 47 49

O1Sixty-six liters of culture medium containing 10% fetal calf serum was used for purification present in the culture medium is fetal calf serum proteins. *One unit of PLA, activity is arbitrarily defined as the amount of enzyme that releases membranes under conditions in which activity is directly proportional to the amount of sample analyzed such that activity was 515% of added radiolabeled substrate.

Purification

Culture

Recovery (%) 1

100 94 75 67

21,000 23,500 24,500 of rPLA,.

Greater

than

1 cpm of [aH]oleic acid analyzed. Serial dilutions

80%

of the protein

from labeled of all fractions

E. coli were

be further purified on the Butyl Toyopearl column (see below). The selection of hydrophobic column chromatography for the final purification step of rPLA, was based on the use of this resin in the purification of human synovial fluid PLA, (25) with the added advantage that rPLA, eluting from the NuGel P-SP column could be directly applied to the Butyl Toyopearl column. This second column successfully removed the trace contaminating proteins present in the rPLA, fraction from the cation-exchange resin. The final sample, after concentration and dialysis, was homogeneous as determined by SDS-PAGE (Fig. 1A). Results of silver-stained gels verified the high degree of purity of the protein (data not shown). The purified rPLA, comigrated with the same M, (16 kDa) as the native protein purified from human placenta. The predicted M, of each protein from the amino acid sequence is 13.9 kDa. The small amount of rPLA, purified from the 0.3 M NaCl eluate of the NuGel P-SP column contained an additional protein-staining band at 21 kDa (Fig. 1A). This protein was determined to be a modified form of rPLA by the results of the following experiments. First, on immunoblots both the 16- and 21-kDa proteins are recognized by monoclonal antibodies (26) prepared against purified human placental PLA, as well as a polyclonal antibody prepared against the purified 16-kDa rPLA, (data not shown). Second, the 21-kDa protein was eluted from an SDS gel with sodium acetate, pH 8.5, containing 0.1% SDS at 37°C overnight (27), and subsequent NH,-terminal sequencing (see below) verified that the eluted protein was rPLA,. Third, since human PLA, has a potential N-glycosylation site, the purified protein fraction was treated under denaturing conditions with Endoglycosidase F/N-glycosidase F (28) and subjected to SDS-PAGE. This treatment led to the virtual disappearance of the 21-kDa species without the appearance of a new protein-staining band, presumably because the protein migrated with the nonglycosylated 16-kDa rPLA, (data not shown). Fourth, upon

PURIFICATION

OF

RECOMBINANT

PHOSPHOLIPASE

31

A,

97 kDa 66 kDa 43

kDa

31 kDa

22 kDa rPLA2 14 kDa

-0l-t L

-NV -

rPLA2 O.SM NaCl

03

rPLA2 NaCl

PLA2

h-

PLA2

t

FIG. 1.

(A) SDS-PAGE of purified rPLA,. Gel electrophoresis was performed as described under Materials and Methods. Various amounts of protein purified from the 0.3 and 0.8 M NaCl fractions of the NuGel P-SP column are shown. Note the detectable protein-staining band at 21 kDa in the rPLA, preparation purified from the 0.3 M NaCl fraction. Purified PLA, from human placenta is also shown. (B) Immunoprecipitation and SDS-PAGE of radiolabeled PLA, expressed in transfected cells. COS-7 cells were transfected with the PLA, expression plasmid as described (11). Sixty-six hours later, cells were washed with PBS and incubated in 1 ml serum-free medium with or without 5 pg tunicamycin/ ml. After 1.5 h at 37°C the medium was removed, 0.5 ml of serum-free medium containing tunicamycin and [%]cysteine (0.5 mCi/ml) was added, and the cells were incubated for 4 h. The radiolabeled proteins were immunoprecipitated as described (ll), resolved by SDS-PAGE, and visualized by fluorography. The arrowhead denotes the location of the Pl-kDa form of PLA, compared to the nonglycosylated PLA, (16 kDa). The + and - denote samples processed in the presence or absence of tunicamycin.

reexamination of our previously published data (ll), we noted the presence of a 21-kDa protein which was immunoprecipitated from transfected COS cells with monoclonal antibodies against human group II PLAz. Figure 1B shows that the 21-kDa protein is not present when the cells are treated with tunicamycin prior to radiolabeling, while the level of the 16-kDa PLA, is unaffected. These data indicate that the 21-kDa protein is a glycosylated form of rPLA,. If we assume equal recoveries of the 21- and 16-kDa proteins through the purification procedures, then the 21-kDa protein represents ~1% of the total rPLA, secreted by CHO cells.

Structural

Characterization

of rPLAz

Expression of rPLA, in CHO cells was performed by transfecting the cells with human placental cDNA containing the entire protein coding region including the putative signal peptide. Purified rPLA,, as well as the glycosylated form of the enzyme (21 kDa), had the expected NH,-terminal sequence (6,7), NLVNFHRMIKL, indicating that the signal peptide had been correctly pro-

cessed in CHO cells during transit through the secretory pathway. The high degree of purity of rPLA, was confirmed by the lack of any detectable secondary sequence. Gel filtration of rPLA, on a Superose 12 column in 50 mM Tris (pH 7) containing 150 mM NaCl yielded a sharp, symmetrical peak eluting at M, 4.5 kDa (Fig. 2). Recovery of PLA, from the column was 94% as determined by enzymatic assays. When the NaCl concentration was increased to 500 mM with or without 2 mM CaCl,, identical elution profiles were obtained. No rPLA, was recovered from the column in the absence of NaCl. Since the molecular weight of rPLA, from the amino acid sequence is 13.9 kDa, the protein apparently was interacting with the column resin, resulting in a marked underestimate of its molecular weight. As a result of these observations, other types of sizing columns were used in an attempt to eliminate this problem (Table 2). Gel filtration on a Superdex 75 column, which is composed of dextran linked to agarose beads, yielded a A& 4.7 kDa for rPLA,. On a TSK-250 column, rPLA, eluted with a A4,3.1 kDa in 500 mM NaCl and did not

32

-

----.. LEVIN

lo

20

30

Elution FIG. rPLA, HCl min. (158 (1.3

40 time

50

60

T--

the purified protein as well as the protein secreted in culture medium are identical. In order to determine whether the purified enzyme exists as a monomer in solution, the protein was subjected to equilibrium analytical ultracentrifugation in an attempt to overcome the problems encountered in the gel filtration experiments. Data obtained from the ultracentrifugation experiments indicated that the enzyme had a calculated M, 13.8 f 0.1 kDa in 50 mM TrisHCl (pH 7.5 at 4’C) and A4,14.2 * 0.1 kDa in 100 mM Tris-HCl (pH 8.0 at 4”C), 150 mM NaCl, 2 mM CaCl,. Thus, rPLA, exists as a monomer in solution over a protein concentration range of lo-30 pg/ml to -5 mg/ml. Figure 3 shows the CD spectra of rPLA in the absence or presence of 2 mM CaCl,. At pH 7, rPLA, had two negative maxima at 211 and 222 nm with molecular ellipticity values of -1.56 X lo6 and -1.635 X 106, respectively. The spectrum of rPLA, in the region of 240-290 nm exhibited negative bands at 264, 271, 279, and 287 nm. There was no significant difference in the CD spectrum in the presence or absence of CaCl,. Addition of 5 mM EGTA to the sample containing CaCl, had no effect on the CD spectrum (data not shown). The a-helical content of rPLA was calculated (29,30) to be 37-38%.

70

(min.)

2.

Gel filtration of rPLA, on a Superose 12 HR lo/30 column. (125 gg protein) was loaded onto the column in 50 mM Tris(pH 7 at 22°C) containing 0.15 M NaCl. Flow rate was 0.3 ml/ Molecular weight standards were thyroglobulin (670 kDa), IgG kDa), ovalbumin (44 kDa), myoglobin (17 kDa), and vitamin B,, kDa).

Enzymatic

TABLE Weight

Determinations

Characterization

of rPLA,

Purified rPLA, was characterized with [3H]oleic acidlabeled, autoclaved E. coli membranes as substrate. The pH optimum of the purified enzyme was -8 with E. coli membranes as substrate (data not shown). The optimum Ca2+ concentration for catalytic activity is shown in Fig. 4. In the absence of added Ca2+, rates of reaction were less than 10% of maximum, which were achieved with 0.5 mM Ca2+. A lo-fold higher Ca2+ concentration had no further effect on the reaction. The low but detectable reaction obtained in the absence of added Ca2+ could be inhibited >50% by 1 PM EGTA and completely

elute from the column in the absence of NaCl. On a Zorbax GF-250 column, which is a silica-based resin as is the TSK-250 column, rPLA, could not be eluted from the column in PBS or PBS containing 500 InM NaCl. However, rPLA, did elute from a Sephadex G-50 column with a M, 11.5 kDa, very close to the known monomeric molecular weight of the protein (13.9 kDa). A similar molecular weight was also obtained for rPLA, in CHO cell culture supernatant, indicating that the M, of

Molecular

. AL.

ti’l’

2

of rPLA,

on Gel

Filtration

Columns

Column Buffer PBS f0.5 M NaCl 50 mM Tris f0.15 M NaCl f0.5 M NaCl

Zorbax

GF-250 nd. nd. n.d. n.d.

TSK-250

Superose

n.d. 3.1

n.d. 4.5 4.5

12

Superdex

Note. Molecular weight of rPLA, (kDa) on various sizing columns with regard to known protein standards: kDa), ovalbumin (44 kDa), myoglobin (17 kDa), and vitamin B,, (1.3 kDa). The amount of purified rPLA, except for the Sephadex G-50 column (300 pg). The “-” indicates that the column was not used and “nd.” eluted from the column.

75

Sephadex

G-50

-

-

4.7

11.5

thyroglobulin (670 kDa), IgG (158 applied on the columns was 125 pg indicates that no detectable PLA,

PURIFICATION

OF

RECOMBINANT

PHOSPHOLIPASE

33

AZ

6000 T 4 > Fg

II

+ 2mM CaC12

4000

z 8 N 2

2000

I 100

1

I-

I

I

I

I

200

300

400

500

NaCl

(mM)

FIG. 5. The effect of various concentrations lytic activity of rPLA, with [3H]oleic acid-labeled as substrate.

1600

Ll

1700 ---Z--TO

Wavelength FIG. 3. Circular dichroism ence of 2 mM CaCl,. Protein Tris-HCl (pH 7 at 22°C).

(nm)

spectra of rPLA, in the absence or presconcentration was 300 fig/ml in 50 mM

inhibited by increasing the concentration of EGTA to 10 ELM (Fig. 4). Addition of 2 mM Ca2+ to samples containing 10 PM EGTA restored complete catalytic activity. A

10000

10000 .

z- 6000 .E fz-. 6000 5 0N 4000 2

x

.

8000 6000 4000

2000

2000

L

1J1

la34 EGTA

izl-.

0 ---_ I&

0.5mM

l.cenu CaCl 2

2.anM

5.OldA

EGTA

FIG. 4. The effects of Ca2* and EGTA on the catalytic activity of rPLA, using radiolabeled E. coli membranes as substrate. Assays were performed as described under Materials and Methods. For the experiment with various concentrations of CaCl,, -0.12 ng rPLA, was used. For evaluating the effects of EGTA on enzyme activity, a 20-fold higher amount of rPLA, (-2.4 ng) was used in order to increase the sensitivity of the assay. When rPLA, (-0.12 ng) was pretreated with 10 PM EGTA followed by addition of 2 mM CaCl,, the observed catalytic activity is indicated with an X.

of NaCl on the cataE. coli membranes

number of divalent cations (Cd”, Zn2+, Sr2+, Ba2+, Mn2+, and M$+) were tested for their ability to replace the Ca2’ requirement in the reaction. At final concentrations of 0.1, 1, and 10 InM, none of the divalent cations were capable of replacing Ca2+ in the reaction (data not shown). Sodium chloride had a stimulatory effect on the hydrolysis of oleic acid from autoclaved E. coli membranes (Fig. 5). An approximate three- to fourfold enhancement in metabolism was obtained with a NaCl concentration of loo-150 mM. High concentrations of NaCl (3500 mM) inhibited the reaction. Similar results were obtained by replacing NaCl with various concentrations of Tris-HCl buffer (pH 8.0), suggesting that increased ionic strength was the cause of the stimulation of rPLA, activity. The stimulatory effect of fatty acid-free BSA on the catalytic activity of rPLA, (with labeled E. coli membranes as substrate) was due to at least two effects of BSA. When concentrated enzyme was diluted in buffer (100 mM Tris, pH 8, 150 mM NaCl, 2 mM CaC1,) containing various concentrations of BSA (1 pg-5 mg/ml) and subsequently incubated in assay mix containing 5 mg BSA/ml, a >lO-fold increase in activity over that of enzyme diluted in buffer with no BSA present was observed. Maximal stimulation was obtained with 0.5 mg BSA/ml dilution buffer (Fig. 6). In a subsequent experiment, when enzyme was diluted in buffer containing 1 mg BSA/ml and then incubated with E. coli membranes such that the final BSA concentration in the assay varied between 30 pg and 5 mg BSA/ml, a 2.5-fold maximal enhancement of activity was obtained when the assay mix contained -1.0 mg BSA/ml. These effects were not observed when purified IgG was used to replace BSA in the dilution buffer or assay mix. Thus, BSA prevents

34

LEVIN

IOOOO

1

ET

AL.

A

3s h 24 zc 2

8000 T &

6000 -

.

B

r t 2 5

N 4000

-

2000

-

6000 -

4000

-

2000

-

2

N

2

1

, 0.0

I 0.5

I 1.0

'

I 1.5

r

I 2.0

c

BSA

I 2.5

'

I 3.0

z

II 3.5

II 4.0

I 4.5

r

/ .05

I 5.0

I .5

I 1

I 2

(mg/ml)

BSA

I 4

I 3

I 5

(mg/ml)

FIG. 6.

Effect of BSA on the catalytic activity of rPLA,. (A) Purified rPLA, was diluted in 100 mM Tris-HCl (pH 8 at 22°C) containing 2 mM CaCl, and 150 mu NaCl containing various concentration of BSA. The diluted enzyme was then incubated with [3H]oleic acid-labeled E. coli membranes in the same buffer mixture containing 5 mg BSA/ml. (B) Enzyme was diluted in 100 mM Tris-HCl (pH 8 at 22°C) containing 2 mM CaCl,, 150 mM NaCl, and 1 mgBSA/ml. The diluted enzyme was then assayed for PLA, activity such that the final concentration of BSA in the assay varied from 30 rg/ml to 5 mg/ml.

loss of catalytic activity during dilution of the enzyme, probably by preventing binding of the highly basic rPLA, to walls of containers, and stimulates activity with E. coli membranes as substrate, perhaps by binding inhibitory product(s) of the reaction as previously suggested (31). The pH optimum for the reaction as well as the effects of divalent cations, NaCl, and BSA on enzymatic activity with E. coli membranes as substrate were indistinguishable for purified rPLA, or human placental PLA, (data not shown). Thus, by all criteria examined, the recombinant protein is identical to the enzyme purified from human placenta. The purification procedure described here for rPLA, results in good recovery of purified protein and is amenable for processing large volumes of cell culture supernatant. As much as 300 liters of supernatant has been applied to a single lo-ml NuGel P-SP column with no detectable overloading of the column and no decrease in recovery or purity of rPLA, subsequently eluted from the column. The purified protein is presently being utilized for NMR analysis in the presence and absence of inhibitors of enzymatic activity (S. Narasimhan, R. Daniel, W. Levin, and D. Fry, unpublished observations). ACKNOWLEDGMENTS We Dharm

thank Bogda for assistance

Wegrxynski with the

for obtaining the CD gel filtration experiments,

spectra, Liz Yu-Ching

Pan for NH,-terminal sequence ration of the manuscript.

analyses,

and Lisa

Nieves

for prepa-

REFERENCES 1. Waite, M. Research” York. 2. Vadas, P., pholipase 3. Pruzanski, tor between

(1987) (Hanahan,

in “The

Phospholipases: Handbook of Lipid D. J., Ed.), Vol. 5, pp. l-332, Plenum, New

and Pruzanski, W. (1984) Role of extracellular phosA, in inflammation. Adv. Znfi’nmmation Res. 7, 51-59. W., and Vadas, P. (1991) Phospholipase Ax-A mediaproximal and distal effecters of inflammation. Zm-

munol. Today 12,143-146. 4. Heinrikson, R. L., Kreuger, E. T., and Keim, P. S. (1977) Amino acid sequence of phospholipase AZ-a. from the venom of Crotalus adamantew. A new classification of phospholipases A, based upon structural determinants. J. Biol. Chem. 252,4915-4921. 5. Davidson, F. F., and Dennis, E. A. (1990) Evolutionary relationships and implications for the regulation of phospholipase A, from snake venom to human secreted forms. Mol. Evol. 31,228238. 6. Hara, (1988) quence novial

S., Kudo, I., Matsuta, Amino acid composition of human phospholipase fluid. J. Biochem. 104,

K., Miyamoto, T., and Inoue, K. and NH,-terminal amino acid seA, purified from rheumatoid sy-

326-328.

Lai, C-Y., and Wada, K. (1988) Phospholipase A2 from human synovial fluid: Purification and structural homology to the placental enzyme. Biochem. Biophys. Res. Commun. l&57,488-493. Kramer, R. M., Hession, C., Johansen, B., Hayes, G., McGray, P., Chow, E. P., Tizard, R., and Pepinsky, R. B. (1989) Structure and properties of a human non-pancreatic phospholipase A,. J. Biol.

Chem. 264,5768-5775. Seilhamer,

J. J., Plant,

S., Pruzanski,

W., Schilling,

J., Stefanski,

PURIFICATION E., Vadas, pholipase 10.

P., and Johnson, L. K. (1989) A2 in arthritic synovial fluid.

Multiple

OF forms

RECOMBINANT of phos-

J. Biochem. 106, 38-42.

PHOSPHOLIPASE 21. Laemmli, assembly

4

35

U. K. (1970) Cleavage of structural of the head of bacteriophage T4.

proteins

during

the

Nature 227,680-6&X.

Kanda, A., Ono, T., Yoshida, N., Tojo, H., and Okamoto, M. (1989) The primary structure of a membrane-associated phospholipase A, from human spleen. Biochem. Biophys. Res. Com-

22. Ryan, somal

mun. 163,42-48.

23. Hewick, R. M., Hunkapiller, M. W., Hood, L. E., and Dreyer, W. J. (1981) A gas-liquid solid phase peptide and protein sequenator. J. Biol. Chem. 256,7990-7997. 24. Hunkapiller, M. W., Granlund-Moyer, K., and Whiteley, N. W. (1986) Analysis of phenylthiohydantoin amino acids by HPLC, in “Methods of Protein Microcharacterization” (Shively, J. E., Ed.), pp. 315-327, Humana Press, Clifton, NJ. 25. Hara, S., Kudo, I., Chang, H. W., Matsuta, K., Miyamoto, T., and Inoue, K. (1989) Purification and characterization of extracellular phospholipase A, from human synovial fluid in rheumatoid arthritis. J. Biochem. 105, 395-399. 26. Stoner, C. R., Reik, L. M., Donohue, M., Levin, W., and Crowl, R. M. (1991) Human group II phospholipase A,: Characterization of monoclonal antibodies and immunochemical quantitation of the protein in synovial fluid. J. Immunol. Methods 145,127-136.

11. Crowl, R., Stoner, C., Stoller, T., Pan, Y.-C., and Conroy, (1990) Isolation and characterization of cDNA clones from man placenta coding for phospholipase A,. Adu. Exp. Med.

R. hu-

Biol.

279,173-184. 12. Cullen, ciency

B. R. (1986) Trans-activation of human immunodefivirus occurs via a bimodal mechanism. Cell 46,973-982.

13. Urlaub, G., and Chasin, cell mutants deficient

L. A. (1980) in dihydrofolate

Isolation of Chinese hamster reductase activity. Proc.

Natl. Acad. Sci. USA 77,4216-4220. 14. Subramani, S., Mulligan, the mouse dihydrofolate cleic acid in simian virus 15. Graham, the assay

F. L., and of infectivity

Van

R., and Berg, P. (1981) Expression of reductase complementary deoxyribonu40 vectors. Mol. Cell. Biol. 1, 854-864. Der Eb, A. J. (1973) New technique of human adenovirus 5 DNA. Virology

for

52,

456-467. 16. Crowl, R. M., Stoller, (1991) Induction hepatoma cells

17. Familletti,

P. C., and Fredericks, J. E. (1988) Techniques for mammalian cell immobilization. BiolTechrwlogy 6,41-44. 18. Rothhut, B., Russo-Marie, F., Wood, J., DiRosa, M., and Flower, R. J. (1983) Further characterization of the glucocorticoid-induced antiphospholipase protein “renocortin.” B&hem. Biophys.

Res. Commun. 117,878-884. 19. Rock, C. O., and Jackowski, S. (1985) Pathways for the incorporation of exogenous fatty acids into phosphatidylethanolamine

Chem. 255,7941-7955.

J., and Stoffel, W. (1990) A facile method for the isolation and preparation of proteins and peptides for sequence analysis in the picomolar range. Biol. Chem. Hoppe-Seyler 371.675-685.

28. Freeze, H. H., and Varki, A. (1986) Endo-glycosidase F and peptide N-glycosidase F release the great majority of total cellular N-linked oligosaccharides: Use in demonstrating that sulfated Nlinked oligosaccarides are frequently found in cultured cells. Bio-

them. Biophys. Res. Commun. 140,967-973. 29. Jibson, M. D., and dichroism of synthetic in methanol solutions. 30.

in

Escherichia coli. J. Biol. Chem. 260, 12,720-12,724. (1951)

0. H., Rosebrough, Protein measurement

Biol. Chem. 193,265-275.

microJ. Biol.

27. Kurth,

T. J., Conroy, R. R., and Stoner, C. R. of phospholipase A, gene expression in human by mediators of the acute phase response. J. Biol.

Chem. 266,2647-2651.

20. Lowry,

D. E., Thomas, P. E., and Levin, W. (1980) Hepatic cytochrome P-450 from rats treated with isosafrole.

N. J., Farr, with the

A. L., and Randall, R. J. Folin phenol reagent. J.

31.

Li, C. human

Znt. J.

H. (1981) /3-Endorphin. Circular analogs with various chain lengths Peptide Protein Res. 18, 297-301.

Bewley, T. A., Brovetto-Cruz, J., and Li, C. H. (1969) Human pituitary growth hormone. Physicochemical investigations of the native and reduced-alkylated protein. Biochemistry 8,4701-4708. Conricode, K. M., and Ochs, R. S. (1989) Mechanism for the inhibitory and stimulatory actions of proteins on the activity of phospholipase A,. B&him. Biophys. Acta 1003,36-43.