The Kinetics of the Phospholipase Az-catalyzed ... - Semantic Scholar

THEJOURNALOF BIOLOGICAL CHEMISTRY VoI. 256. No. 12, Issue of June 25. pp. 62744281. 1981 Printed m ( I S.A

The Kinetics of the Phospholipase Az-catalyzed Hydrolysis of Egg Phosphatidylcholine in Unilamellar Vesicles PRODUCT INHIBITION AND ITS RELIEF BY SERUM ALBUMIN* (Received for publication, October 6, 1980, and in revised form, February 11, 1981)

Josh P. Kupferbergf, Shinji Yokoyama,and Ferenc J. KezdyE From the Department of Biochemistry, The University of Chicago, Chicago, Illinois 60637

Only the lecithin in the outer leaflet (representing 70% of the total) of egg lecithin unilamellar vesicles is hydrolyzed by Crotalus atrm phospholipase A2.Hydrolyzed vesicles remain intact and impermeable to ionic solutes. The fatty acids produced in the hydrolysis remain on the vesicle and are only partially ionized at neutral pH due to electrostatic repulsions. About 40% of the lysolecithin product is desorbed from the vesicle. In the presence of a large excess of bovine serum albumin, the reaction is first order with respect to both the enzyme and the substrate. At 21 “CypH 7.2, I = 0.16 M, and [Ca”] = 7 mM, the second order rate constant is kelr(*) = 1.5 X lo6 M“ s-’. In the absence of albumin, the reaction is inhibited competitively by both the monomeric (K;” = 4.5 X IO-’ M) and micellar (nKIa= 3.7 X M) forms of lysolecithin ((critical micelle concentration] = 4.3 X M). Bovine serum albumin complexes two molecules o f lysolecithin with a dissociation constant, Kb = 5 x IO-” M. With substoichiometric albumin, the reaction is biphasic, and, when the albumin is saturated with lysolecithin, the kinetics become similar to those observed in the absence of albumin. The action of phospholipase A2 shows that in unilamellar vesicles there is only one major lecithin conformation in the outer leaflet, or that all conformations are rapidly interconvertible.

studied with shorter chain lecithin monomers (20,21), micelles (20) and insoluble monolayers (22,23), and with phospholipids in human serum lipoproteins (20, 21). Phospholipase AS hydrolysis of multiIamellar liposomes and unilamellar vesicles has also been investigated (24-26) and it was found that only the phospholipids in the outerleaflet of the vesicle bilayer are hydrolyzed (27, 28). With all but the monolayer substrates, the enzyme is subject to strong product inhibition which is relieved by serum albumin (29-31). Snake venom phospholipases AB have also been used t,o probe the structureof biological membranes and lipoproteins (29, 30, 32, 33), but the full time courseof the actionof these enzymes on membranes and membrane-like particles has not yet been analyzed (24, 27, 28).

For phospholipase AY to be a useful membrane structural probe, the full time course of the reactionof the enzyme with membrane-like particles and the effect of serum albumin must be understood. To this end,we have undertaken the studyof the hydrolysis of egg phosphatidylcholine in unilamellar vesicles by C.atrox phospholipase At and the following paper is the account of these results. MATERIALS AND METHODS

Chemicals and solvents purchased from Aldrich, Baker, Fisher, Mallinckrodt, or Sigma were reagent gradeunless otherwise indicated. Solvents used in high performance liquid chromatography were Fisher HPLC’ grade, Aldrich Spectroquality,orMallinckrodt Pesticide Grade; the latter two were filtered through 5-pm pore-sintered glass Many phospholipids can form unilamellar vesicles which discs. ThinLayerChromatography-Forphosphatidesand lysophosare closed bilayer particles of rather homogenous size; diam65:25:4 (v/v/v) was eters of 220-260 b, have been commonly reported (1).Unila- phatides, the solvent system CH~CL:CH.IOH:H~O; lysolecithin were employed. T h e R, values for fatty acid, lecithin, and mellar vesicles were first preparedby sonicating aqueouslipid 0.95, 0.50, and 0.25, respectively. For fatty acids, the solvent system dispersions (1-8). Batzri and Korn (9) showed that the same usedwas petroleumether:ethylether:acetic acid; 80:20:1 (v/v/v). vesicles may also be prepared by the injection of ethanolic With Silica Gel G plates (Analtech, Newark, DE) lipid spots were phospholipid solutions into aqueous media. Egg yolk phos- visualized by the iodine vapor method of Marinetti (35). With prophatidylcholine forms vesicles which are electrically neutral prietary hard layer silica gel plates with inorganic binder (Analtech, and are more stable and less permeable to ions than vesicles Newark, DE), lipid spots were charred at 170 “C for 20 min after immersion of the plate in 50% H&04 (v/v) (36). Phosphatides were formed from any other phospholipid commonly used for this also stained by the method of Dittmer and Lester (37). For quantipurpose (2, 9-16). The unilamellar vesicles have been widely tative TLC, Silica Gel 60 plates were used (E. Merck and Darmstadt, used as models of cell membranes and lipoprotein surfaces Germany). The dried plates were immersed first in the Dittmer and Lester stain and then in 0.1 M NaCI.’ The chromatograms were air(17-19). The kinetics of the action of Crotalus atrox and Crotalus dried until opaque and subjected to densitometry a t 600 nm using a adumunteus phospholipase AS (EC 3.1.1.4) has already been Schoeffel SD 300 spectrodensitometer and averaging the valuesof at least four determinat,ions.Accuracy of the densitometry was checked using standard mixtures of egg lecithin and egg lysolecithin. which ’Support for this research was provided by United States Public were spotted on the same plate as the unknowns. Health Service Grants HL-18577 (Program Project) and HL-15062 High Performance Liquid Chromatography-For product analy(Specialized Center of Research). The costs of publication of this sis, reaction mixtures which had been quenched by adding EDTA and article were defrayed in part by the payment of page charges. This adjusting the p H to 3.0 with glacial acetic acid were either extracted article must therefore be hereby marked “nduertisernent” in accord- __ __-~_ ance with 18 U.S.C. Section 1734 solely to indicate this fact. ’ Abbreviations used in this paper were: HPLC, high performance Supported by United States Public Health Service Medical Sciliquid chromatography; TLC, thin layer chromatography; CMC, critentist Traineeship 5TGM-07281. 9 To whom correspondence should be addressed at Box 10, 920 ical micelle concentration: BSA, bovine serum albumin. ’ R. A. Kleps and J. P. Kupferberg, unpublished observations. East 58th Street. Chicago. Illinois 60637.

6274

Kinetics of Lecithin Vesicle Hydrolysis withCHCI:I:CHaOH, 2:l(v/v)(38)and dried undernitrogen,or lyophilized withoutextraction. Lipidswereredissolved inHPLCgrade isopropanol prior toinjection. Columns used were pBondapack 5 p~ CN (0.32 x 30 cm) (Waters Associates, Waltham,MA)and Spherosorb CN (0.32 X 30 cm) (Phase Separations, packed by Altex Scientific, Berkeley, CA).The eluentwas either isocratic acetonitrile: water; 1:l (v/v) or an acetonitrile:water gradientfrom 3040% acetonitrile (v/v); the gradient was programmed for 10 min a t a flow rate of 80 ml/h. A Perkin Elmer LC-55 spectrophotometer set a t 205 nm was used to detect peaks. Compounds separated by chromatography were identified by rechromatography with authentic standards or by TLC. Egg phosphatidylcholine was isolated from fresh hen’s eggs by a modification of the method of Wells and Hanahan (39). Mallinckrodt CC-7-special grade silicic acid was used and columns were eluted stepwise with chloroform, successive ch1oroform:methanol mixtures in ratios; 3:1, 1:1, 1:3 (v/v), and finally with methanol. The fractions collected were assayed by TLC. The final product was shown to be pure lecithin by TLC and HPLC. T h e ultraviolet spect.rum of the purified egg lecithin showed no evidence of oxidation: A?IO/A,-,,~ 2 15. After lyophilization from dry redistilled benzene, the egg lecithin was stored a t -20 “C under dry nitrogen. Alternatively, egg lecithin was purchased, in drylyophilized form from Avanti Biochemicals (Birmingham, AL) and found to be as pure as the one prepared in our laboratory. Unilamellar vesicles prepared from both lecithin sources appeared identical by all criteria. Unilamellar vesicles were prepared by the method of Batzri and Korn (9).Yields of pure vesicles from starting materialwere typically >70%. Vesicles were stable by criteria of gel permeation chromatography and M : j , m (9). They remained impermeable to ionic solutes showing 99%) oleic acid (Sigma) (>998) or linoleic acid (Hormel Institute, Austin, MN). All fatty acids used were free of impurities detectable by high performance liquid chromatography and thinlayer

e,,,

’

by Phospholipase A2

6275

chromatography. Bovine serumalbumin lysolecithincomplexwas prepared by mixing smallamounts of concentratedethanolic egg lysolecithin solutions with lipid-free albumin solutions. Ovalbumin (>99%) was purchased from Sigma. Zon Trapping during Single Bilayer Vesicle Formation-To demonstrate ion trapping, single bilayer vesicle preparations were made in 0.16 M KC1 alone or with additional 50 mM nickel(I1) chloride or cadmium(I1) chloride. Cadmium-free KC1 solutions were prepared by adjusting the solutions to pH11 with NH:I, passingthem over Chelex 100 (Bio-Rad),and neutralizing with HCI. Cadmium-and nickelcontaining vesicles were freed of untrapped divalent cations in the course of the SepharoseCL-4B chromatography. Unilamellar vesicles for chloride trapping experiments were repeatedly concentrated by ultrafiltration through a n XMIOOA membrane (Amicon) and diluted with 0.16 M NaNO:]. Trapped ions were released by the addition of 20% (v) methanol. Chloride was measured by AgN0.r nephelometry. One-half ml of 1 M AgNO.1was added to 3 ml of solution and turbidity was measured at 400 nm using a Cary 15 spectrophotometer (Varian Instruments). Cadmium was measured using a Keithly model 616 electrometer with a Ruzicka solid-state electrode (Radiometer, Copenhagen) (46). Standardresponse curves were prepared using pCd” buffers (47). Nickel was measured by the dimethyl-glyoximate method of Nielsch and Giefer (48); the absorbance of the CHC1.1-extractable complex was measured at 455 nm. Critical micelle concentrations were determined from the concentration dependence of the surface tension of solutions (49). Surface tension was measured using a Cahn model KG electrobalance (Ventron, Cerritos, CA) equipped with a du Nouy ring (0.9 cm in diameter) fashioned from 0.02-mm Nichrome wire. Sample solutions were in round troughs (3.0 cm) machined fromsolidTeflon blocks. Glassdistilled deionized water was used in the preparation of all solutions. Enzyme Kinetics-Phospholipase A, reactions were monitored by measuringthe evolution of protonsproduced when lecithin was hydrolyzed. Reaction mixtures (3.5 ml) contained 0.16 M KC1 and 7 mN CaC12. Proton evolutionwas measured spectrophotometrically with 1.5 X 10“ M p-nitrophenolindicatorand 3.0 X 10.‘ M 3-(Nmorpho1ino)propanesulfonic acid or Tris-HCI buffer, pH 7.20, in the reaction mixture, monitoring the absorbance at 410 nm. The magnitude of protonevolution,measured by the decrease in A4,(,,was calibrated by back-titration of reaction mixtures withknown amounts of standardizedKOH.Alternatively,protonrelease was followed using a Radiometer (Copenhagen,Denmark) automatic titratormodel TTT 2, PHA 942, REA 300, and A B 0 11, in the pH stat mode. The “dead band” of the pH-stat had been reduced to less than 0.01 pH to accommodate concentrated protein solutions and other solutions of similarly high buffering capacity (29, 30). Otherproducts of thereaction were quantitated by TLCand HPLC. The earliest phase of the reaction, up to 0.58 reaction cornpleted, was followed using radioactively labeled egg phosphatidylcholine vesicles. Aliquots of the reaction mixture (0.02 M 3-(N-morpholino)propanesulfonic acid, pH 7.20, 0.16 M KCI, 4.4 X 10 M labeled phosphatidylcholine in vesicle form and 8.6 X 10 ” M dimer C. atrox phospholipase AS) were removed and extracted by a procedure modified from that of Nishijima (50). The 230-p1 aliquots were quenched with 200 pI of cold 20% trichloroacetic acid and mixed rapidly with 100 pl of 16R (v/v) Triton X-100 and 1 ml of hexane, containing 100 pg of oleic acid ascarrier. After phaseseparation, facilitated by centrifugation (3000 rpm X 10 min), 500plof the hexane layer was transferred to a glass scintillation vial containing 5 ml of scintillation cocktail (4 g/liter Omnifluor; New England Nuclear) (33%,v/v: Triton X-100, in toluene) and counted using a Nuclear Chicago Isocap 300 liquid scintillationcounter.Alternatively,aliquots were extracted with chloroform:methanol, 2:1 (v/v) and subjected to TLC. Spots and suitable blank areas of the chromatogramswere scraped and counted. The binding of C . atrox phospholipaseA, to egg lysophosphatidylcholine micelles was demonstrated by gel permeation chromatography using a Sephadex G-100 (Pharmacia, Uppsala, Sweden) column (0.9 x 29 cm). The eluent was 0.05 M 3-(N-morpholino)propanesulfonic acid, pH 7.2,0.16 M KCI, 7 mM CaClz, and 1 X 10” M dimeric C. atrox phospholipase AL. Egg lysolecithin, in micellar form, was introduced in the presence of 1 X 10”’ M enzyme.

’

RESULTSANDDISCUSSION

The hydrolysis of egg phosphatidylcholine in unilamellar vesicles by C. atron phospholipase A2 is inhibited by that portion of lysolecithin produced which desorbs from the vesicle surface. Both monomeric and micellar lysolecithin form

6276

Kinetics of Lecithin Vesicle Hydrolysis

complexes with the enzyme, inhibiting its activity toward vesicles. Bovine serum albumin, when present in the reaction mixture, binds desorbed lysolecithin, mitigating its inhibitory activity as long as thealbumin is in excess.Until the albumin is saturated, V / S ([reaction velocity]/[substrate concentration]) is linearly related to the product concentration. The two equivalent and independent high affinity lysolecithin binding sites of bovine serum albumin are distinct from the high affinity fatty acid binding sites. The time course of the hydrolysis, in the presence of an excess of albumin, is consistent with a simple second order reaction,exhibiting no enzyme processivity or denaturationon the surface. The results, upon which these conclusions are based, follow. Interaction of Lysolecithin with Proteins-In agreement with previously published results (27,28),we have found that C. atrox phospholipase APhydrolyzed phosphatidylcholine in unilamellar vesicles. A considerable decrease in the rate is observed, however, after the initial phase of the reaction, suggesting unusually strong product inhibition. For this reason, the interaction of fatty acid and lysolecithin with the enzyme had to be studied before the analysis of the kinetics could be undertaken. The CMC of egg lysolecithin was determined to be4.3 X M underourexperimental conditions (Fig. 1).Theapparent CMC increases to 9.9 X M when thesubphase contains 2.0 X M bovine serum albumin. The non-zero surface pressure belowCMC and the lower maximum value seen with serum albumin are due to the intrinsic surface activity of albumin (51). The apparent increase of CMC in the presence of albumin can be accounted for by postulating the complexation of lysolecithin by the protein. The surfacepressure of 1.5 X M lysolecithin over a protein-free surface is 2.5 dynes/cm. When the subphase contains 2.0 X M bovine serum albumin, this surface pressure is not exceeded until a lysolecithin concentration of 4.3 X M is reached. This is consistent with a binding stoichiometry of two lysolecithins per albumin molecule. If at this point the concentration of unoccupied binding sites on the albumin is approximately equal to the free lysolecithin concentration, one may estimate that the dissociation constant of the complex is KB 2 5 X lo-* M of the complex. The slope of the surface pressure (a)uersus In [lysolecithin] curves (Fig. 1) is not changed by the presence of albumin in the subphase. This implies that the micellar aggregation number is independent of albumin (52) and that the formation of lysolecithin-albumin-mixed micelles is unlikely. Gel permeation chromatography demonstratedthe binding of C. atrox phospholipase As to egg lysolecithin micelles

by Phospholipase A2

(enzyme assayed by the radioactive substrate method). Micelles, 3 X M in lysolecithin, eluted from Sephadex G-100 column at K,, = 0.1, corresponding approximately to a M , = 100,000 and a micellar aggregation number, n, of 200. When M phosphothe sample applied to the column had 1.1 X lipase Az, nearly all the enzyme coeluted with the micelles, whereas the elution volume of the dimeric C. atrox phospholipase Az in the absence of lipid was at K,, = 0.5. From these results, the estimated upper limit of the dissociation constant was E7 lo-‘ M for the enzyme-lysolecithin complex. We also observed that C. atrox phospholipase AP does not bind to lysolecithin micelles in the absence of Ca2+,indicating that the binding occurs at theactive site of the enzyme. Reaction Products-The products of the hydrolysis of egg phosphatidylcholine in unilamellar vesicles by C. atrox phospholipase AP, identified by TLC and HPLC, were fatty acid and lysolecithin. When vesicles containing radioactive phosphatidylcholine were hydrolyzed, the only radioactive product was fatty acid. Since the fatty acid moiety in the 2-position of the glycerol is labeled, the positional specificity of C. atrox phospholipase Al for phosphatidylcholine in vesicles is the same as for that in the wet ether system (37). No more than 70% (+2%) of the phosphatidylcholine in vesicles was ever hydrolyzed, while in ether-water or by addition of chloroform to the70% hydrolyzed vesicles, the samephosphatidylcholine was quantitatively hydrolyzed by the enzyme. Vesicles were still intact up to 6 h afterexhaustive enzymatic hydrolysis as they coeluated with undigested vesicles on Sepharose CL-4B. They were impermeable to previously trapped Cd2’ (2% limit of detection) and theirASwdid not change, again implying an absence of fusion. More than 99% of the fatty acid produced in the enzymatic hydrolysis coeluted with the intact vesicles during gel permeation chromatography and was retained in ultrafiltration througha cellulose acetate membrane with 100-A pores (Amicon XM-100 A). In contrast, between 25-50% of the lysolecithin was found to be associated with low molecular weight fractions from both the above separations. This lysolecithin, dissociated from vesicles, was either bound to bovine serum albumin (when it was present) or free in either monomeric or micellar form. That only the phosphatidylcholine in the outer leaflet of the vesicle bilayer is accessible to phospholipase A2 has been previously reported (27,28).Assuming a vesicle radius r = 115 A (3,9,12,19,53),a bilayer thickness of 44 A (53),an excluded molecular area of66 A’ (54) for the phosphatidylcholine headgroup, and maximum density of head group packing in the inner leaflet, one may calculate that there should be 960 I I I I I phosphatidylcholine molecules on an inner leaflet surface. Since 30% of the phosphatidylcholine in unilamellar vesicles is on the inner leaflet, there should be 3200 phosphatidylcholine molecules/vesicle. The 2240 phosphatidylcholine molecules that occupy the 1.66 X lo”A’ surface of the outer leaflet have an average area of74 A’ available to each headgroup. By varying within reasonable limits the radius (*IO A), the bilayer thickness (+5 A), and the excluded molecular area (+-IOA*) the number of phospholipid molecules per vesicle changes no more than 25%. Total number of phosphatidylcholine molecules per vesicle estimated by our method is slightly a t variance with the numbers proposed previously (53, log1 l y s o l e c ~ t h ~ n ] 55, 56). We feel, however, that the partitioning between the FIG. 1. Egg lysolecithin critical micelle concentration and inner and outerleaflets is one of the most important structural bindingto bovine serumalbumin at 22 “C. Surfacepressure aspects to be taken into consideration, and for this reason we (dynes em”) versus log [lysolecithin] plot (O),subphase; 0.1 M NaC1, will use the numbers derived above for our calculations. 0.05 3-(N-morpholino)propanesulfonicacid, pH 7.2, 1%ethanol (v%). Product Ionization-Comparison of the proton yield measubphase as before but with 2.0 X IO-” M bovine serum albumin. Egg lysolecithin concentration at upper discontinuity corresponds to sured acidimetrically with the fatty acid yield measured by HPLC or TLC showed that the fattyacid produced ionized to the CMC.

.,

Kinetics of Lecithin Vesicle Hydrolysis

by Phospholipase A2

6277

a lesser and lesser extent as the reaction progressed. Only 40% of the fatty acid was ionized in phosphatidylcholine vesicles whose outer leaflet had been quantitatively hydrolyzed by phospholipase Aa at pH 7.20. Similar results have been observed inthe hydrolysis of phospholipid in serum lipoproteins (29, 30), and have been attributed to electrostatic repulsion. Such phenomena obey the following equation (30, 57): '0

where, w is a constant, N is the number of fatty acids/per vesicle, 2 is the average charge per vesicle, and Kin, is the proton dissociation constant when 2 = 0. The value of Kin, was estimated to be 4.9 by titration of the fatty acids on unilamellar vesicles which were 4 % hydrolyzed. Using the value n = 3200 and the fact that 70% of the lipid is in the outer leaflet, product yield was converted to N, and proton Substituting these values into Equation 1 yielded yield to w = 3.10 X One can calculate apriori the value of w with the help of the following equation (57):

X

z.

I

W=L(LL) 2DkT r 1 + Kr

I

200

I

400 800

I

I

I

600

Tlme k e c )

where E is the charge of the electron in electrostatic units, D the dielectric constant, k is the Boltzman constant, T is the temperature, and K is defined as K=

[ (s) K C DKT

lo00

(CsZS)

Here, N is Avagadro's number, C, is the concentration of ions in the buffer, and Z, is their charge. For our reaction conditions, K = 1.31 X 10' and, for a radius r = 115 8, w = 3.10 X 10". The excellent agreement of the experimental and theoretical values of w further justifies the assumption that electrostatic effects are solely responsible for the suppression of ionization of the vesicles. Accordingly, in our kinetic studies the proton yield measured experimentally was converted to product yield with the help of Equation 1. Since the electrostatic suppression of ionization is modest for all but the last phase of the reaction, the 25% uncertainty in the value of N does not alter significantly the product yield. Activation of theEnzymaticHydrolysis by AlbuminQuantitative hydrolysis of the lipid in the outer leaflet was not attained unless at least 1 mol of serum albumin/4 mol of available substrate was included in the reaction mixture. At smaller albumin-to-lipid ratios, the reaction was biphasic with the magnitude of product evolved in thefast phase, P,, increasing as thealbumin increased (Fig. 2). The dependence of P , on the concentration of bovine serum albumin allowed us to estimateR , the fraction of the lysolecithin product which desorbed from the surface of the unilamellar vesicles. Fig. 3 shows a plot of P , uersus albumin concentration. The linearity of the plot indicates that a constant fraction of the lysolecithin is desorbed as the hydrolysis proceeds. The stoichiometry, derived from the slope, was 4.8 egg lysolecithins produced per bovine serum albumin. Since the serum albumin binds 2 mol of lysolecithin per mole of protein (see above), R = 0.42 f 0.04. When vesicles containing "-labeled phosphatidylcholine were hydrolyzed by the enzyme, the initial rate was the same with and without bovine serum albumin, thoughsignificant differences in the rate of hydrolysis were already apparent by the time the reaction was 0.5% complete ([So]= M). This enhancement of phospholipase A2 activity by bovine or human serum albumin has been reported previously for other enzymes and otherforms of substrates (29,31).We also observed that ovalbumin is ineffective in that respect. When we separated bovine serum albumin from hydrolyzed phos-

FIG. 2. Time course of phospholipase A2 hydrolysis of unilamellar egg phosphatidylcholine vesicles in the presence of substoichiometric bovine serum albumin. pH-stat assay system: 2.43 X IO-' M (dimer) phospholipase A.L,2.81 X IO",' M egg phosphatidylcholine ([SO] = 1.97 X M ) , 22 "C. Bovine serum albumin concentrationsare: curue a, 2.6 X M; curve b, 4.3 X M; curve c, 5.5 X M;curve d,6.4 X M; curve e, 9.2 X M. Theoretical curves were calculated from numerical integration of Equation 11 for the fast phase and Equation 12 for the slow phase.

1

I

I

I

I

4

FIG. 3. Capacity of bovine serum albumin to relieve product inhibition by desorbed egglysolecithin. Product liberated in the fast phase ([Px]) versus bovine serum albumin ([BSA]). The solid line represents the linear leastsquare fit. phatidylcholine vesicles by ultrafitration using an Amicon XM lOOA membrane, the albumin collected had lost its ability to enhance the rate or extent of the phospholipase reaction. In contrast, bovine serum albumin complexed with 2-mol equivalents of palmitic, oleic, or linoleic acid was just as efficient as thelipid-free albumin in enhancing the enzymatic hydrolysis, whereas albumin saturated with 2-mol equivalents

6278

Kinetics of Lecithin Vesicle

Hydrolysis by Phospholipase A2

of lysolecithin became totally ineffective for this purpose. Thus, the role of albumin in promoting the phospholipase reaction consists of sequestering free lysolecithin from the reaction mixture. On the other hand, since the addition of albumin restores the initial rate of the reaction while interacting only with the desorbed lysolecithin, we can conclude the surface-bound lysolecithin is not the inhibitor. Analysis of the Time Course-The apparent initial rates, as measured by the spectrophotometric dyeassay, were found to be directly proportional to both enzyme and substrate concentrations up to M phospholipid. The order was the same in the presence and absence of bovine serum albumin but the apparentsecond order rate constant was 3 X lo4 M" s-' without albumin and 1.5 X lo6 M" s" in the presence of 1.5 X M serum albumin. When reaction mixtures contained less than optimal amounts of bovine serum albumin, some product inhibitionwas already evident before the albumin had been saturated, in spite of the fact thatunder these conditions no miceller lysolecithin was present. Likewise, during the hydrolysis of the [3H]phosphatidylcholine vesicles without albumin, inhibition occurred when the product was still below the CMC. The formation of lysolecithin micelles in the reaction mixture resulted in a considerable additional decrease in the reaction rate, showing that both forms of the lysolecithin inhibited the reaction. If the inhibitions are competitive, then the time course of the phospholipase A:, hydrolysis of egg phosphatidylcholine in vesicles without bovine serum albumin should follow the rate law:

Tlme

kecl

I

I

I

1

5

10

15

20

8

[Product]

X

Jd fM)

FIG. 4. Time course of the phospholipase Az hydrolysis of where V is the reaction velocity, P is the product concentration, t is time, kex(zIthe second-order rate constant, Eo the e g g phosphatidylcholine in unilamellar vesicles without bototal enzyme concentration, So the initial concentration of vine serum albumin. A , Product uersus time plot. pH-stat assay system: 3.83 X IO-' M (dimer) phospholipase A?, 2.43 x IO"' M egg accessible phosphatidylcholine, P, the molarity of desorbed M ) , 22 "C. 0, experimental phosphatidylcholine ([So]= 1.70 X lysolecithin monomer and P, the molarity of lysolecithin in points. The theoretical time course curve was generated by using a the miceller form. Likewise KF is the dissociation constant of third order Runge-Kutta numerical integration of Equation 6 with the enzyme-lysolecithin micelle complex expressed as molarity the values I($ = 4.6 X lo-' M, nKjl = 3.7 x W 7 M, So = 1.7 x M. B, linearized plot, S / V uersus [prodof micelle, n is the miceller aggregation number, and K? is M, and CMC = 4.3 X the dissociation constant of the enzyme-lysolecithin monomer uct], according to Equation 6. complex. In the absence of albumin, the concentration of desorbed for the full time course is second order. It is first order with lysolecithin exceeds the CMC very early in the reaction. Thus, respect to both enzyme and substrate; thus the rate law is we may assume P,,, = CMC. Equation 4 may then be trans- described by the equation, ( d P / d t )= kexlS)Eo (So- P). This formed to equation accounts for the full time course. Fig. 5A shows an example of the timecourse of a typical reaction. Fig. 5B is the dP kexw Eo ( S o P ) v="= semilogarithmic first order plot for the same reaction. From CMC RP - CMC dt I+a linear least squares fit of the pseudo-fist order plot (Fig. KT nK? 5B), the slope, k e X c lwas , determined to be 4.6 X s-I, Equation 5 may be further transformed to the double recip- yielding = 1.46 X lo6 M" s-', in excellent agreement with the value obtained from the initial rates. The apparentsecondrocal: is independent of the enzyme conorder rateconstant centration except a t low enzyme concentrations (Eo < lo-' M) when a monomer-dimer equilibrium, with only the dimer active, is significant. From our data, we find the dissociation which is of the form: constant of the dimer, Kd =: 6 x 10"" M. This value is very S/V=A + B*P (7) close to the dissociation constants of C. adamanteus phosFig. 4A shows the time course of a reaction without albumin pholipase A2 dimer, Kd = 3.7 X lo-"' M determined elsewhere and Fig. 4B is a plot of S/V versus P. The latterplot is linear, (22), using the dioctanoyl lecithin monolayer assay system. The second order rate constant is the same when derived indicating that Equation 5 correctly describes the reactions. The values of the constants, KY and n E in equation were from the time course of the reaction with excess albumin or extracted from the slope and intercept, using the constants from the initial reactionrates, with and without albumin. The kexv2,= 1.5 X lo-' M" sa', R = 0.4, and CMC = 4.3 X M. surface of the vesicle retains most of the reaction products and albumin does not bind to the surface of the partially We found KT = 4.6 X lo-' M andnK7 = 3.7 X 10" M. When bovine serum albumin is present inexcess with hydrolyzed vesicle. Identity of the rate constants means that respect to thedesorbed lysolecithin, the overall reaction order on the surface of the vesicles, neither the lysolecithin nor fatty +

Kinetics of Lecithin Vesicle Hydrolysis I

6279

tion of the complexed lysolecithin is, to a good approximation, equal to RP and theconcentration of the unoccupied binding sites on the albumin is equal to 2 [BSA] - RP, where [BSA] is the molarity of bovine serum albumin added to thereaction mixture. Thus, (2[BSA] - RP)P,,, Krc = (8) RP

I

I

by Phospholipase Aa

Since the end of the fastphase is determined by the saturation of the albumin by lysolecithin, then 2 [BSA] = RP,, Thus, from Equation 8, it follows that

Substituting Equation 9 into Equation 4, with Po = 0, and taking the reciprocal, one obtains:

b I

I

I

I

I

200

100 Tlme

kecl

By plotting (So- P ) / V versus P/(P=- P ) for the fast phase of the reaction, a straight line was obtained (not shown), and within experimental error, the slope equaled the intercept, indicating that K N z KF. For this reason, Equation 10 could be simplified to:

-

"1

0'2 if 2[BSA) [BSA]'

=

RP, and R

=

0.4

Fig. 6 shows plots of V/(Sn - P) versus P for the time courses shown in Fig. 2. The data yield straight lines with negative slopes, andthe slopes are inversely proportional to P,, whereas the intercepts are same the within experimental error. Thus, Equation 11 correctly describes the fast phase of the reaction. From the Y-intercepts we obtained an average of Tlme bec 1 hex,l,= 1.5 X IO-' M" s", in good agreement with the rate FIG. 5. Time course of phospholipase A2 hydrolysis of egg constant obtained in the presence of excess albumin. The phosphatidylcholine in unilamellar vesicles in the presence of bovine serum albumin. A , [product] uersus time.pH-stat assay theoretical curves in Fig.2 were generated using the integrated system: 2.47 X lo-' M phospholipase Az, 9.0 X 10 ~' M egg phosphati- form of Equation 11 and the parameters calculated from Fig. dylcholine ([S,]= 6.3 X IO-' M ) , 1.28 X 10 M bovine serum albumin 6. 22 "C. e,experimental points. Theoretical curve calculated as a first Once the albumin is saturated, the slower phase of the order reaction, using k e x ( z lEll= 4.6 X 10 ~ " s - 'B, . semilogarithmic plot hydrolysis follows the rate law: of data in Fig. 4.

acid inhibit the enzyme. This is not surprising in light of the fact that C. atrox phospholipase A2 is not saturated with egg phosphatidylcholine up to at least 10" M phospholipid when the substrate is in the form of unilamellar vesicles. The rate law is not consistent with a slow surface penetration of the enzyme proposed for the hydrolysis of phospholipid monolayers by pancreatic phospholipase A, (58-60). Only twopools of substrate in phosphatidylcholine vesicles are kinetically significant and they areidentifiable with the inner and outer leaflets of the bilayer (28). This implies either that all the phosphatidylcholine molecules in the outer leaflet have the same physical state, that all physical states arerapidly interconvertible relative to the rate of the enzymatic reaction, or that these differences are not discernible to the enzyme. The general rate law, Equation 4,leads to the prediction thatthe time course of reactions with substoichiometric amounts of bovine serum albumin is biphasic because the micellar inhibition occurs when the albumin is exhausted. This is indeed the case, as shown in Fig. 2. The concentration of lysolecithin desorbed from the vesicle is equal to RP and it will be almost quantitatively bound to the albumin, since the albuminconcentration is greater than K R during the fast phase of the reaction. Under these conditions, the concentra-

;Product]

FIG. 6. Phospholipase

A2

X

Id

(MI

hydrolysis of unilamellar egg

phosphatidylcholine vesicles. Plot according to Equation 11, showing inhibition by egg lysolecithin monomer during the fast phase of the reaction. Reaction conditions are identical to those in Fig. 5. The solid lines represent the least square fit for the fast phases.

Kinetics of Lecithin Vesicle Hydrolysis by Phospholipase A2

6280

which is derived from Equation 4 with P, = RP - 2 [BSA] - CMC. This is readily transformed to: (So - P ) / V 1

CMC

KT

2[BSA]

+ CMC +-,R

nK7

nKY

( 13)

PI

Plots of (So - P ) / V versus P again yielded straight lines (not shown). The values of the y-intercepts were negative, but the dissociation constant of lysolecithin monomer from phospholipase A, could not be reliably extracted from them. The values of nK7 estimated from the slopes agree with the value derived from the phospholipase A, hydrolysis without albumin. The latter portions of the time course inFig. 2 ( P > Pm),fit the theoretical curve derived from the numerical integration of Equation 12, showing the validity of the original assumptions. In the case where there is a high enough concentration of bovine serum albumintoensure that time course of the hydrolysis of the lecithin by phospholipase AS is monophasic but where all of the albumin becomes saturated during the course of the reaction, the time course is also adequately described by Equation 1. Likewise, Equation 10 described the fast phase of reactions in which the albumin had been pretitrated with varying amounts of exogenous egg lysolecithin. Fig. 7 summarizes the results of a series of such biphasic reactions. The values of P, and the apparentfirst order rate constant,k:!& = Vo/Sowere derived, as described above, from plots of VIS uersus P,. As Fig. 7 shows, the titration of bovine serum albumin here is again -2 mol of lysolecithin/mol of albumin. In conclusion, the rate behavior of the hydrolysis ofegg lecithin in unilamellar vesicles by C. atrox phospholipase A2 is that of a simple second order reactionwith the inhibition of the enzyme by one product. The kinetics of this reaction would be hopelessly complex were it not for circumstances which simplify the situation. The most important of these is the desorption of a constant fraction of the lysolecithin product, -40% throughout the reaction. The question may be asked why lysolecithin desorbs and why a constant fraction of it desorbs. A possible answer may be found in hydrolysis of phosphatidylcholine monolayers (61) at the air-water interface, where the enzymatic reaction is accompanied by an increase of the apparent molecular areas. Thus, in the hydrolysis of lecithin in vesicles, a constant surface area and surface pressure could only be preserved by the desorption of some of the surface components, especially the water-soluble

products. As in the case of the serum high density lipoproteins (20), the lysolecithin of the vesicles seems to desorb more readily than the fatty acids. Our results show that it is the desorbed lysolecithin which inhibits the enzyme by acting as a competitive inhibitor in the monomeric, as well as in the micellar state. The reversible binding of lysolecithin to bovine serum albumin and the inhibition of the enzyme cause the rate of the reaction to be finely regulated by the degree of saturation of albumin. The affinity of the two lysolecithin binding sites on bovine serum albumin is comparable to thatof the two highest affinity binding sites for fatty acid (62-64). However, the fact that bovine serum albumin complexed with two fatty acids still activates phospholipase demonstrates thenonidentity of the lysolecithin and fattyacid sites. This role of bovine serum albumin in reducing the lysolecithin product inhibition of the phospholipase AS reaction may have physiological significance. While phospholipase A2 activity has been reported in human serum (34,65), no enzyme with pure phospholipase A, activity has been isolated from that source. The enzyme 1ecithin:cholesterol acyltransferase has been shown to act as a phospholipase A, (66). Both enzymes remove the acyl moiety from the 2-position of lecithin, leaving a lysolecithin product that inhibits their activity (67). Bovine serum albumin, in vitro, relieves the inhibition in both cases, presumably by binding lysolecithin. Serum albumin might, therefore, also be an importantregulator of 1ecithin:cholesterol acyltransferase inplasma, controling serum lipoproteinmetabolism by its degree of saturation by lysolecithin. Acknowledgments-We wish tothank AndrewRepasy for his expert technical assistance and Altex Scientific for the generous gift of high performance liquid chromatography columns. We further wish to acknowledge the valuable advice received from Dr. John H. Law. We thank Drs. Robert L. Heinrikson, Robert A. Kleps, Nikhil M. Pattnaik, and Betty W. Shen for their helpful suggestions. REFERENCES

1. Huang, C. H. (1969) Biochemistry 8,344-352 2. Johnson, S.M., Bangham, A. D., Hill, M. W., and Korn, E. D. (1971) Biochim. Biophys. Acta 233, 820-826 3. Chapman, D., Fluck,D. J., Penkett, S. A., and Shipley, G.G. (1968) Biochim. Biophys. Acta 163,255-261 4. Attwood, D., and Saunders, L. (1965) Biochim. Biophys. Acta 98, 344-350. 5. Bangham, A. D., and Horne, R. M. (1964) J . Mol. Biol. 8, 660668 6. Lucy, J . A., and Glauert, A. M. (1964) J . Mol. Biol. 8, 727-748 7. Saunders, L., Perrin, J., and Gammack, D. (1962) J . Pharm. Pharmacol. 14, 567-572 8. Papahadjopoulos, D., and Watkins, J. C. (1967) Biochim. Biophys. Acta 135,639-652 9. Batzri, S.,and Korn, E. D. (1973) Biochim. Biophys. Acta 298, 1015-1019 10. Blok, M. C., van der Neut-Kok, E. C. M., van Deenen, L. L. M., and de Gier, J. (1975) Biochim. Biophys. Acta 406, 187-196 11. Kantor,H. L., Mabrey, S., Prestegard, J . H., andSturtevant, J. M. (1977) Biochim. Biophys. Acta 466,402-410 12. Suurkuusk, J., Lentz, B. R., Barenholtz, Y., Biltonen, R. L., and Thompson, T. E. (1976) Biochemistry 15, 1393-1401 13. Mao, S. J . T., Sparrow, J. T., Gillian, E. B., Gotto, A. M., Jr., and Jackson, R. L. (1977) Biochemistry 16,4150-4156 14. Inoue, K. (1974) Biochim. Biophys. Acta 339,390-402 15. van Dijck, P. W. M., Ververgaert, P. H.J. Th., Verkleij, A. J., van I 2 Deenen, L. L. M., and deGier, J. (1975) Biochim. Biophys. Acta 406,465-478 [ELPC] / [ E A ] 16. Chapman, D., Peel, W. E., Kingston, B., and Lilley, T. H. (1977) Biochim. Biophys. Acta 464,260-275 FIG. 7. Titration of binding site on bovine serum albumin by 17. Papahajopoulos, D. (1973) Form a n d Function ofPhospholipids exogenous egg lysolecithin. 0, [Pl] (M); m, apparent (S-I) (Ansell, G. B., Hawthorne, J . N., and Dawson, R. M. C., eds) uersus egg lysolecithin [ELPC] added perbovine serum albumin. T h e 143-170, Elsevier, Amsterdam solid lines represent the least square fits. See text for experimental 18. Morrisett, J. D., Jackson, R. L., andGotto, A. M., Jr. (1977) details.

Kinetics of Lecithin Vesicle Hydrolysis BiochLm. Biophys. Acta 472,93-133 19. Kroon, D. J., Kupferberg, J. P., Kaiser, E. T., and Kezdy, F. J . (1978)J.Am. Chem. SOC.100,5976-5977 20. Wells, M. A. (1972) Biochemistry 11, 1030-1041 21. de Haas, G. H., Bonsen, P. P. M., Pieterson, W. A,, andvan Deenen, L. L. M. (1971) Biochim. Biophys. Acta 239,252-266 22. Shen, B. W., Tsao, F. H. C., Law, J. H., and Kezdy, F. J . (1975)J . Am. Chem. SOC.97, 1205-1208 23. Cohen, H., Shen, B. W., Snyder, W. R., Law, J. H., and Kezdy, F. J . (1976) J. Colloid Interface Sci. 56, 240-250 24. Wilschut, J . C., Regts, J., Westernberg, H., and Scherphof, G. (1978) Biochim. Biophys. Acta 508, 185-196 25. Op den Kamp, J . A. F., de Gier, J., and van Deenen, L. L. M. (1974) Biochim. Biophys. Acta 345, 253-256 26. Op den Kamp, J . A. F., Kanuerz, M. Th., and van Deenen, L. L. M. (1975) Biochim. Biophys. Acta406, 169-177 27. Kupferberg, J . P., and Kezdy, F. J . (1978) Fed. Proc. 37, 1834 28. Sundler, R., Alberts, A.W., and Vagelos, P. R. (1978) J. Biol. Chem. 253,5299-5304 29. Pattnaik, N.M., Kezdy, F. J., and Scanu, A. M. (1976) J . Biol. Chem. 251, 1984-1990 30. Aggerbeck, L.P., Kezdy, F. J., and Scanu, A. M. (1976) J. Biol. Chem. 251,3823-3830 31. Smith, A. D., Gul, S., and Thompson, R. H. S.(1972) Biochzm. Biophys. Acta 289, 147-157 32. Dernel, R. A,, Geurts van Resseim, W. S. M., Zwaal, R. F. A,, Roelofsen, B., and van Deenen, L. L. M. (1975) Biochim. Biophys. Acta 406,97-I07 33. van Alphen, L., Lugtenberg, B., van Boxtel, R., and Vernhoef, K. (1977) Biochim. Biophys. Acta466,257-268 34. Shen, B. W. and Law, J. H. (1979) The Biochemistry of Atherosclerosis (Scanu, A. M., ed) pp. 275-291, Marcel Dekker, New York 35. Marinetti, G. V. (1962) J. Lzpid Res. 3, 1-20 36. Rouser, G., Kritchevsky, G., and Yamamoto, A. (1967) in Lipid Chromatographic Analysis (Marinetti, G. V., ed) pp. 99-162, Marcel Dekker, New York 37. Dittmer, J. C., and Lester, R. L. (1964) J. Lipid Res. 5, 126-127 38. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochern. Physiol. 37,911-917 39. Wells, M. A., and Hanahan, D. J. (1969) Methods Enzymol. 14, 178-184 40. Hachimori, Y., Wells, M. A,, and Hanahan, D. J. (1971) Biochemistry 10,4084-4089 41. Ornstein, L. (1964) Ann. N . Y.Acad. Sci. 121, 321-349 42. Lands, W. E. M. (1960) J . Biol. Chem. 235, 2233-2237 43. Saito, K., and Hanahan, D. J. (1962) Biochemistry 1, 521-532

by Phospholipase A2

6281

44. White, D. A., Ansell, G. B., Hawthorne, J. N., and Dawson, R. M. C. (1973) Form and Function in Phospholipids, pp. 441-482, Elsevier, Amsterdam 45. Elsbach, P., and Pettis, P. (1973) Biochim. Biophys. Acta 296, 89-93 46. Ruzicka, J., Lamm, C. G., and Tjell, J . Chr. (1972) Anal. Chim. Acta 62, 17-28 47. Ruzicka, J., and Hansen, E. H. (1973) Anal. Chim. Acta 63, 115128 48. Nielsch, W., and Giefer, L. (1956) Mikrochim. Acta 1956, 522527 49. Lagocki, J. W., Emken, E. A., Law, J . H., and Kezdy, F. J. (1976) J. Biol. Chem. 251,6001-6006 50. Nishijima, M., Akamatsu, Y., and Nojima, S. (1974) J. Biol. Chem. 249,5658-5667 51. Evans, M. T. A., Mitchell, J., Mussellwhite, P. R., and Irons, L. (1970) in Surface Chemistry of Biological Systems. (Blank, M., ed) pp. 1-22, Plenum, New York 52. Tanford, C.(1973) The Hydrophobic Effect, pp. 60-70, John Wiley & Sons, New York 53. Lagner, P., Gotto, A. M., and Morrisett, J. D. (1979)Biochemistry 18, 164-171 54. Shen, B. W., Scanu, A.M., and Kezdy, F. J . (1977) Proc. Natl. Acad. Sci. U. S. A. 74,837-841 55. Huang, C. H. (1969) Biochemistry 8, 344-352 56. Mason, J. T., and Huang, C. (1978) Ann. N. Y . Acad. Sci. 308, 29-48 57. Tanford, C. (1961) Physical Chemistry of Macromolecules, pp. 548-572, John Wiley & Sons, New York 58. Pattus, F., Slotboom, A. J., and de Haas, G. H. (1979) Biochernistry 18, 2691-2697 59. Pattus, F., Slotboom, A. J., and de Haas, G. H. (1979) Biochemistry 18, 2698-2702 60. Pattus, F., Slotboom, A. J., and de Haas, G. H. (1979) Biochemistry 18, 2703-2707 61. Dawson, R. M. C. (1969) Methods Enzymol. 14, 633-648 62. Spector, A. A., John, K., and Fletcher, J. E. (1969) J . Lipid Res. 10, 56-67 63. Spector, A. A., Fletcher, J . E., and Ashbrook, J. I>. (1971) Biochemistry 10, 3229-3232 64. Switzer, S., and Eder, H. A. (1965) J. Lipid Res. 6, 506-511 65. Etienne, J., Paysant, M., Gruber, A., and Polonovski, J . (1969) Bull. Soc. Chim. Biol. 51, 709-716 66. Aron, L., Jones, S., and Fielding, C. J. (1978) J. Biol. Chem. 253, 7220-7226 67. Fielding, C. J., Shore, V. G., and Fielding, P. E. (1972) Biochim. Biophys. Acta 270, 513-518