Analytical Biochemistry 407 (2010) 253–260

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/locate /yabio

Surface dilution kinetics using substrate analog enantiomers as diluents:Enzymatic lipolysis by bee venom phospholipase A2

Jasmeet Singh a, Radha Ranganathan a,⇑, Joseph Hajdu b

a Department of Physics and Center for Supramolecular Studies, California State University, Northridge, CA 91330, USAb Department of Chemistry and Biochemistry and Center for Supramolecular Studies, California State University, Northridge, CA 91330, USA

a r t i c l e i n f o

Article history:Received 8 July 2010Received in revised form 9 August 2010Accepted 10 August 2010Available online 19 August 2010

Keywords:PhospholipaseLipid hydrolysisInterfaceKineticsSurface dilution

0003-2697/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.ab.2010.08.015

⇑ Corresponding author. Fax: +1 818 677 3234.E-mail address: radha.ranganathan@csun.edu (R. R

1 Abbreviations used: PLA2, phospholipase A2; NaTC, soa-phosphatidylcholine, dipalmitoyl; D-DPPC, D-a-phosBSA, bovine serum albumin; DMBP, 3,4-dimethyl benzofluorescence quenching; 5DSE, 5-doxyl stearic acid meresonance; DPS, 3-dodecyldimethylammoniopdimyristoylglycerophosphocholine.

a b s t r a c t

A novel assay employing D-enantiomers of phospholipids as diluents for characterizing surface kinetics oflipid hydrolysis by phospholipases is introduced. The rationales of the method are (i) D-enantiomers resisthydrolysis because of the stereoselectivity of the enzymes toward L-enantiomers and (ii) mixtures of L+D-lipids at various L/D ratios but constant L+D-lipid concentrations yield a surface dilution series of variableL-lipid concentration with constant medium properties. Kinetic characterization of bee venom phospho-lipase A2 activity at bile salt + phospholipid aggregate–water interfaces was performed using the mixedL+D-lipid surface dilution assay, and interface kinetic parameters were obtained. The assay applies to bio-membrane models as well. Activity was measured by pH-stat methods. Aggregation numbers and inter-face hydration/microviscosity measured by time-resolved fluorescence quenching and electron spinresonance, respectively, confirmed that interface properties were indeed invariant in a surface dilutionseries, supporting rationale (ii), and were used to calculate substrate concentrations. Activity data showexcellent agreement with a kinetic model derived with D-enantiomers as diluents and also that D-phos-pholipids bind to the enzyme but resist hydrolysis; underscoring rationale (i). The assay is significant forenabling determination of interface-specific kinetic parameters for the first time and thereby character-ization of interface specificity of lipolytic enzymes.

� 2010 Elsevier Inc. All rights reserved.

Phospholipase A2 (PLA2)1 catalyzes hydrolysis of the fatty acid– sensitive to the lipid/detergent ratio as well as to the total lipid +

ester linkage at the sn-2 position of phospholipids [1]. Enzymaticactivity is most significant when lipid substrates are in aggregatedforms [2–5]. The activity occurs at the aggregate–water interface [6].Kinetic investigation is a useful approach to elucidate the mechanisticrole of the interface in the rate of membrane lipid hydrolysis that is asyet not fully understood. The basic requirement for testing any kineticmodel is that the medium properties not change as the substrate con-centration is varied. Otherwise, the concentration effects and mediumeffects become inextricably linked and it is not possible to test kineticschemes. Designing a surface dilution series with invariable mediumproperties to test models of interfacial enzymatic kinetics has been along-standing problem in interfacial enzymology. The problem inthe case of lipid substrates solubilized in detergent micelles is thatthe interface properties change as the lipid concentration or lipid/detergent ratio is varied because aggregation numbers, micellar shape,surface charge, interface hydration, and microviscosity are quite

ll rights reserved.

anganathan).dium taurocholate; L-DPPC, L-phatidylcholine, dipalmitoyl;

phenone; TRFQ, time-resolvedthyl ester; ESR, electron spinropanesulfonate; DMPC,

detergent concentration. In the case of membrane models such asvesicle bilayers, the surface concentration cannot even be variedbecause the only component in the medium is the substrate lipid.

A new approach to designing a variable concentration interface,without drastically affecting interface properties, for characterizingkinetics of surface active enzymes is much needed. We propose themixed L+D-lipid assay as a novel solution to this long-standing prob-lem in surface enzymology. In this assay, the total L+D-lipid concen-tration is kept constant while the L/D ratio is varied. The enzyme isstereoselective for the L-lipid substrate [7,8]. Therefore, the sub-strate is the L-lipid and the D-lipid is the diluent together with anyother diluents that may be used such as detergents in the case ofmicellar aggregates. In traditional methods, surface dilution of sub-strate L-lipid concentration was achieved by the addition of deter-gents as diluents [3,4,9–12]. In contrast, substrate concentration inthe current system is diluted within the lipid portion of the kineticmedium with a neutral diluent in the form of the respective D-phospholipid enantiomer that has been found to resist hydrolysisand also to preserve the integrity of the medium when substitutedfor the L-phospholipid in the act of dilution. A surface dilution seriesis achieved by decreasing the L/D ratio. The advantage of such a seriesis that medium properties remain constant as substrate concentra-tion is varied because the total lipid concentration and detergentconcentration (in the case of micellar interfaces) remain constant.

254 Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260

This allows direct determination of the concentration dependence ofthe enzyme activity with respect to a specific target substrate with-out needing to adjust for medium property variations. In existingmethods, the presence of medium property variations and their im-pact on kinetic parameters have always been recognized to be asource of problems in data analysis and interpretation [3,12–14].Therefore, the current method should offer significantly bettergrounds for investigating kinetic schemes and deriving interface–enzyme and lipid–enzyme interaction parameters.

We use this assay to characterize the kinetics of enzymatic activ-ity of bee venom PLA2 at the bile salt sodium taurocholate(NaTC) + (L+D-DPPC) aggregate–water interface (where DPPC isphosphatidylcholine, dipalmitoyl). The power of the assay is demon-strated most effectively with these aggregates but applies to bio-membrane models as well. Activity was measured by pH-statmethods. Biophysical characterization of the mixed bilesalt + phospholipid aggregates was performed to obtain the inter-face substrate concentration; which also confirmed that propertiesof aggregation number, interface hydration (polarity), and microvis-cosity do not change with changes in L/D-lipid ratio so long as the to-tal lipid (L+D-lipid) and total bile salt concentrations are each keptconstant. Thus, the series satisfies basic requirements for testing ki-netic models, namely that medium properties remain invariablewhile the substrate concentration is varied. This would not be possi-ble with the use of just the L-lipid alone because varying the lipid/bilesalt ratio to vary substrate concentrations significantly affects aggre-gate properties; that is, the parameters defining the kinetic mediumare not constant. Using the mixed lipid surface dilution series as thenew assay, interface kinetic parameters are obtained for the activityof bee venom PLA2. Excellent agreement of the measured enzymaticactivity with an interface kinetic model that includes the presence ofthe D-enantiomers shows that both the L- and D-lipids bind to the en-zyme but that the D-enantiomer is not hydrolyzed. The current assayallows mapping of the functional dependence of the kinetic param-eters on interface microstructure and, thus, a paradigm that the roleof the interface is expressed through the kinetic parameters.

Materials and methods

Materials

The lipids L-a-phosphatidylcholine, dipalmitoyl (L-DPPC, 100%,Avanti Polar Lipids) and D-a-phosphatidylcholine, dipalmitoyl (D-DPPC, >99%, Sigma) were obtained as lyophilized powders. The bilesalt NaTC (>95%) was obtained from Sigma and used as received.

PLA2 from honey bee venom was obtained from Sigma as a lyoph-ilized powder. The enzyme was dialyzed against 0.05 M sodiumphosphate buffer at pH 8.0 for 3 days, changing the buffer every8 h. Protein concentration was determined by the method of Lowryand coworkers [15] using bovine serum albumin (BSA) as standardand also by the extinction coefficient method [16,17]. Both of thesemethods gave results consistently within ±5%. The dialyzed enzymewas stored at pH 8.0 in 0.05 M sodium phosphate buffer at 4 �C.

Fluorescence probe pyrene (optical grade, 99%, Aldrich) andquencher 3,4-dimethyl benzophenone (DMBP, 99%, Aldrich) wereused for biophysical characterization of the aggregate by time-resolved fluorescence quenching (TRFQ). 5-Doxyl stearic acidmethyl ester (5-DSE, Sigma) was used as spin probe in the electronspin resonance (ESR) experiments. Double distilled water was usedto prepare all of the solutions.

Kinetic model

The kinetic scheme comprises three key sequential steps: (i) en-zyme–interface binding, with association and dissociation rate

constants k1 and k–1, respectively, and equilibrium dissociationconstant denoted by KS (= k–1/k1), to form the bound enzyme E*;(ii) enzyme–substrate lipid binding at the active site to form theinterfacial complex E*L with association and dissociation rate con-stants k2s and k–2s, respectively; and (iii) substrate lipid hydrolysiswith rate constant k3 [3]. When another entity that also binds tothe enzyme active site but resists hydrolysis is included in theinterface, the scheme gives for initial reaction velocity per moleof enzyme referred to as activity, A [18]:

A ¼ K3n½micelles�½l�sKSKMS þ KMSn½micelles� þ n½micelles�½l�s þ n½micelles�½d�s

KMSKDs

;

ð1Þ

where [micelles] is the concentration of bile salt + lipid aggregates insolution and n is the number of binding sites per aggregate. The sub-script ‘‘s” on the concentrations refers to interface or surface concen-trations; thus, [L]s and [D]s are the interface concentrations ofsubstrate and the hydrolysis-resistant D-enantiomer, respectively,KDs is the interfacial equilibrium D-enantiomer enzyme dissociationconstant, and

KMS ¼k�2s þ k3

k2sð2Þ

is the interface Michaelis–Menten parameter defined by the interfacereaction rate constants k3, k2s, and k–2s. When vesicles are used asmembrane models, [micelles] is to be replaced with [vesicles]. Eq.(1) is the generalized result that includes the special cases: (i) D-enan-tiomers or the other binding entity is absent, [D]s = 0, and we get backthe classic equation of Deems and coworkers [3]; (ii) the second en-tity is hydrolysis resistant and much stronger binding than substrate(like an inhibitor), and Eq. (1) is the same as the one derived by Bergand coworkers in the presence of inhibitors (also Eq. (1) in the refer-ence cited) [18]; and (iii) the D-enantiomer does not bind, then it wouldbe a neutral diluent just like the detergent. The activity would then begiven by Eq. (1) with KDs ?1 and expected to exhibit the well-knownsaturation behavior with increases in [L]s. The assay method does notrequire that the D-enantiomers bind to the enzyme. The function ofthe D-enantiomers is to facilitate variation of the substrate concentra-tion without compromising the underlying structure.

If the total surface lipid concentration, [L]s + [D]s = [Total Lipid]s,is constant, Eq. (1) may be rewritten as

A ¼ ½l�sa1 � a2½l�s

; ð3Þ

where

a1 ¼KSKMS

k3n½micelles� þ ½Total Lipids�KMS

k3KDsþ KMS

k3ð4Þ

and

a2 ¼1k3

KMS

KDs� 1

� �: ð5Þ

The consequence of the presence of a competitively binding buthydrolysis-resistant component is the negative sign in the denom-inator in Eq. (3), which leads to a continuous nonsaturating increasein activity with [L]s.

The two variables for testing the model for activity in Eq. (1) orEq. (3) are [L]s and [micelles]. The micelle concentration is given by

½micelles� ¼ ½NaTC� þ ½l-DPPC� þ ½d-DPPC� � ½free NaTC�Nagg

; ð6Þ

where [NaTC], [L-DPPC], and [D-DPPC] are the solution concentra-tions of NaTC, L-DPPC, and D-DPPC, respectively, Nagg is the numberof molecules forming the aggregate, and [free NaTC] is the unasso-ciated free NaTC monomer concentration in the aqueous phase. The

Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260 255

concentrations of unassociated monomeric lipids can quite reason-ably be taken to be negligible.

Problems with using only the L-enantiomers

If only the L-enantiomers are present as in the usual assays,varying the concentration of the lipid while keeping the detergentconcentration constant, or varying the lipid/detergent ratio whilekeeping the total lipid + detergent constant, causes the aggregationnumber, micelle concentration, and interface properties of hydra-tion (polarity) and microviscosity to change quite significantly[10]. The parameters KS, KM, and k3 that define the kinetic mediumare sensitive to interface properties and are not constant anymore.Properties of ionic as well as nonionic micelles are also sensitive tothe total detergent + lipid concentration even if the lipid/detergentratio is held constant [10,19–21]. So, the idea of varying [micelles]while keeping the lipid/detergent ratio constant does not achieve areliable kinetic medium either. Furthermore, [L]s and [micelles]cannot be varied independently of each other.

Mixed L+D-lipid assembly

The ability to vary [L]s without varying [micelles] or any of theinterface properties is possible if the [detergent] and [L+D-lipids]are each constant and the L/D-lipid ratio is varied. This is the surfacedilution series, and Eq. (3) applies. The aggregation number, andhence [micelles], hydration, and microviscosity are constant forthis series, as shown in this work (Fig. A1 in the Appendix). There-fore, it is reasonable to expect that the kinetic parameters would beinvariant. Thus, the mixed lipid assay is close to being an ideal ki-netic medium to investigate kinetic models. [L]s and [micelles] canbe varied independently of each other.

Mixed micelle solution preparation

The ratio of [NaTC] to [Total Lipid] (i.e., [L-DPPC] + [D-DPPC]) waskept constant at 6 in all of the samples. The composition of the sam-ples was expressed in molar fraction of L-DPPC (XL-DPPC), defined by

Xl-DPPC ¼½l -DPPC�

½NaTC� þ ½l -DPPC� þ ½d-DPPC� : ð7Þ

For enzyme activity measurements, five different series of sam-ples were prepared, each with [Total Lipid] = 2, 5, 10, 15, and20 mM. In each of the surface dilution series, the fraction of L-DPPCin the lipid portion ranged from 0.1 to 1.0. For each series, twostock solutions each of NaTC + L-DPPC and NaTC + D-DPPC wereprepared by mixing an appropriate amount of lipid powder witha few drops of ethanol. The resulting mixture was vortexed thor-oughly to produce a clear solution, which was then dried underdry N2 flux to produce a film of lipid. Thereafter, the requiredamounts of the NaTC and water were added to the dry film toachieve the final concentrations. The solution was stirred overnightto ensure the complete solubilization of phospholipid. Appropriatevolumes of each of the two stock solutions were then mixed toachieve the final XL-DPPC. Finally, CaCl2 was added to a concentra-tion of 10 mM to all of the samples because calcium ion is a re-quired cofactor for hydrolysis by PLA2.

In the samples for TRFQ characterizations, pyrene as the probeand DMBP as the quencher were included. The concentration ofpyrene was kept at approximately one-hundredth the concentra-tion of micelles to ensure that the fraction of micelles with twoor more pyrenes is negligible. The quencher (DMBP) concentrationwas kept at approximately 1.3 [micelles]. The samples for ESR weresimilarly prepared with 5DSE as the spin probe at a concentrationof 0.003 [total surfactant].

Enzyme activity

The activity of enzyme was measured as the micromoles of fattyacid released per milligram of enzyme per unit time. The enzymeactivity measurements were conducted employing standard pH-statmethods [9]. Deprotonation of the fatty acid released by the chemi-cal reaction causes the pH to drop, and the amount of NaOH neededto bring the pH back to a preset value is a measure of enzyme activity.PLA2 activity in NaTC/DPPC aggregates was measured by the addi-tion of 30 ll of enzyme (�7 lg) into 5 ml of aggregate solutionsand monitoring the amount of 0.01 M NaOH required per minuteto maintain a constant pH of 8.0 using a Radiometer pH-stat assem-bly consisting of a titrator, an autoburette, and a pH meter, modelTIM 854 electrode, interfaced to a computer for recording data.The reaction was followed for approximately 6 min. The initial rateof activity was determined from the first 1 to 4 min of data. The yieldincreases linearly with time, and the slope of the line divided by themass of the enzyme in milligrams gives the activity in units of lmol/min/mg. The error in the fitted slopes was less than 1%. All activitymeasurements were conducted at 37 �C.

To account for possible batch-to-batch or temporal variations inenzymatic activity, hydrolytic activity toward standard assay sys-tems was used as reference to obtain normalization factors. Two dif-ferent reference assay mixtures used in this study were (i) standardegg yolk emulsion as described by Nieuwenhuizen and coworkers[22] and (ii) an internal standard mixture of a zwitterionic surfactant3-dodecyldimethylammoniopropanesulfonate (DPS) and dim-yristoylglycerophosphocholine (DMPC) as explained in our previouswork [9]. The composition of the DPS + DMPC mixture used as stan-dard was [DMPC] = 4 mM and [DPS] = 8 mM. The observed activityin the DPS + DMPC assay mixture was 2200 (±10%) lmol/min/mg.The stock solution of PLA2 showed no loss of enzymatic activity overthe time period of study. All of the experiments were conducted withthe same batch of enzymes to avoid batch-to-batch variations inenzymatic activity. The precision error in the activity was deter-mined from several repeated measurements on a few samples, withthe same batch of enzyme under the same experimental conditions,over the course of the entire set of experiments and was found to be±5%. Care was taken to maintain identical experimental conditions(instrument warmup time, sample preparation, and equilibrationtimes). These cautionary measures are believed to have minimizedscatter in the data so that reliable conclusions could be drawn.

Results and discussion

The dependence of the activity of bee venom PLA2 on interfacialsubstrate concentration is shown in Fig. 1 for five surface dilutionseries differing in total [NaTC] + [Total Lipid] concentration butwith the same detergent/total lipid ratio. Therefore, each of thecurves refers to some constant micelle concentration that is differ-ent from the others. [L]s was determined from the measured aggre-gation number, interface hydration, and model calculations asdescribed in the Appendix.

The initial slow increase in activity with [L]s followed by a rapidrise with no saturation agrees with the prediction of Eq. (3) con-firming that the D-enantiomer binds but resists hydrolysis.

The constant a1 extracted by fitting of Eq. (3) to each of the curvesvaries with [micelles], as shown in Fig. 2. Examination of the behav-ior of a1 as a function of [micelles] must take into account the depen-dence of [Total Lipid]s on [micelles]. This is because aggregationnumber varies with NaTC + lipid concentration. Therefore, the lipidsurface concentration is not constant even though the lipid/deter-gent ratio in solution is constant. Fit of Eq. (4) to a1 versus [micelles]after including the empirical dependence of [Total Lipid]s on [mi-celles] given by Eq. (A6) in the Appendix yields

Fig.1. Enzymatic activity of bee venom PLA2 (±5%) on L-DPPC in mixed aggregatesof NaTC + L-DPPC + D-DPPC as a function of the interface substrate concentration,[L]s, for five surface dilution series. The concentration ratio of bile salt to total DPPCis constant at 6. The total DPPC concentration varies from 0.002 to 0.020 M. Withineach of the series, variation in [L]s is achieved by varying the L-DPPC/D-DPPC ratio.[D]s was calculated from measured values of aggregation number and interfacehydration according to the methods in the Appendix. The lines are fits to Eq. (3).

256 Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260

1:41� 10�11 KMS

k3KCþ KMS

k3¼ 2:78� 10�14 ð8Þ

KSKMS

k3þ 7:34� 10�16 KMS

k3KC¼ n 2:72� 10�18 ð9Þ

4:76� 10�10nKMS

k3KC¼ 2:62� 10�12: ð10Þ

The values of a2 from each of the fits in Fig. 1 have a larger precisionerror, and the average value is

a2 ¼ 0:0009� 0:00016: ð11ÞThe individual kinetic parameters were then calculated as solutionsto the fitting results in Eqs. (8)–(11), and the values are reported inTable 1.

Fig.2. Value of a1 from the fit of Eq. (3) to the data in Fig. 1 viewed as a function ofthe measured concentration of [micelles]. The solid and dashed lines are the fitsobtained from Eq. (4) with and without inclusion of the observed [Total Lipid]s

dependence on [micelles], respectively, according to Eq. (A6) and Fig. A4 in theAppendix.

The fits in Figs. 1 and 2 show the appropriateness of expressingthe lipid concentration as surface concentrations rather than molefractions. If the lipid surface concentrations were expressed asmole fractions, [Total Lipid]s in Eq. (4) would be replaced by a con-stant (= 1/7 in this work) and would appear not to contribute to thevariation of a1 with [micelles]. The dashed line is the fit obtainedwhen only the inverse variation of a1 with [micelles] through justthe first term alone in Eq. (4) is considered and the [Total Lipid]s

variation with [micelles] is ignored. However, the surface concen-tration does vary because of the variations of the aggregation num-ber, aggregate shape, and hydration with total lipid + detergentconcentration (see Fig. A4 in the Appendix).

Thus, biophysical characterization of aggregates allows deter-mination of surface concentrations and verification of not onlythe appropriateness of the interface medium for kinetic character-ization but also the effects on the substrate concentration due tochanges in size and shape.

Data presented in Fig. A1 confirm that aggregate properties donot vary with L/D-lipid ratio. Therefore, it is possible to realize atrue surface dilution series, that is, one with constant mediumproperties. On the other hand, a micelle dilution series is notdependable, with Figs. A2 and A3 clearly showing that aggregateproperties and substrate concentration vary with micelle concen-tration, although the lipid mole fraction is constant.

The three step kinetic scheme and the resulting model for activityin Eq. (1) (with [D]s = 0) was originally proposed and tested by Deemsand coworkers with cobra venom PLA2 acting on phospholipids inTriton X-100 micelles [3]. Activity versus total detergent + lipid con-centration was measured for each of several different constant lipid/detergent ratios. The data of Deems and coworkers and subsequentexperiments along the same lines, by other researchers, exhibit sat-uration kinetics and general viability of the surface dilution kineticmodel [4,5,11–14]. However, the scatter in the data and the lack ofagreement of the model with the observed activity dependence onlipid mole fraction have led the same authors to question the validityof their assumptions on lipid mole fraction effects on micellar struc-ture [3,12]. The micelle physicochemical characterization data in theAppendix indeed show that the aggregation number, interface polar-ity, and microviscosity vary not only with lipid mole fraction but alsowith total concentration (Figs. A2 and A4). Therefore, series of vary-ing L-lipid mole fraction or varying [micelles] using mixed micellesof detergent and only L-lipid are not ideal systems for testing kineticschemes and deriving kinetic parameters [13]. Nevertheless, mixedmicelle kinetic media continue to be in use and have yielded usefulinformation on kinetic parameters, schemes, and protein–lipidinteractions, notwithstanding problems due to noninvariable inter-face characteristics within a series that can lead to scatter in the data[11,12]. This type of information can be reliably obtained and sub-stantially improved with the platform of mixed lipid systems pre-sented here because the mixed lipid surface dilution series are aninnovative way to vary the substrate lipid concentration while keep-ing the total L+D-lipid mole fraction, and hence medium properties,constant.

Detailed kinetic analysis and kinetic parameters are potentiallykeys to understanding the differences between different forms ofenzymes. A significant point about the current mixed lipid surfacedilution is that it is capable of addressing specificity of the kineticparameters to the type of interface, substrate lipid, and enzymeindividually precisely because the interface properties can be held

Table 1Numerical values for the kinetic parameters derived as solutions to Eqs. (8)–(11).

n KC/10–12 (Mcm–2) Ks/10–3 (M) Km/10–12 (Mcm–2) k3 (lmol/min/mg)

4 6.05 1.12 2.2 2658

Fig.A1. The aggregation number N (by TRFQ), hydration index H (by ESR), and

Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260 257

constant and interaction parameters of a lipid-specific enzyme at aspecific interface can be determined. In this context, for example,the activity of two forms of phosphatidylinositol 4-kinases at phos-phatidylinositol + Triton X-100 micellar interfaces were shown toexhibit significantly different interfacial enzymological character-istics and kinetic parameters [11,12]. These experiments clearlymake the case for the presence of specificity of kinetic parametersnot only to the lipid and enzyme but also to the interface. Kineticcharacterization with the current mixed lipid assay system wouldadvance the approach to better elucidate protein–lipid interactionsat interfaces through derivation and comparisons of kinetic param-eters of mutant enzymes and different isoforms [14,23].

The application presented here shows how kinetic parametersspecific to an interface, lipid, and enzyme can be derived. It is partof a broader approach toward interfacial enzymology because themixed lipid assay and proposed surface kinetic schemes apply toall types of lipid interfaces, including mixed micelles, and lipid bi-layer membrane models such as vesicles and liposomes [14]. In thecase of membrane models, one would use mixtures of L+D-lipids invarious proportions to prepare vesicles or liposomes. Our initialinvestigations using ESR of spin probes incorporated into mixedL-DPPC and D-DPPC vesicles to measure the temperature depen-dence of the micropolarity (hyperfine coupling constant) andmicroviscosity of the bilayer medium [24,25] showed that theseproperties not only are independent of the proportion in the mix-ture but also show a sudden change at 41 �C for all mixtures, cor-responding to the gel–liquid phase transition temperature of DPPC.Phase transition temperatures are quite sensitive to intermolecularinteractions. The observation that each of L-DPPC, D-DPPC, andtheir mixtures exhibits the same phase transition temperature isindicative that any complexation between L- and D-lipids either isabsent or does not affect physicochemical properties. The otherchiral molecule in the sample solution is the enzyme. Possibilityof complexation of the D-phospholipids with the enzyme is takeninto account through the inclusion of the interfacial enzyme–lipiddissociation constant KDs in the kinetic scheme (Eq. (1)).

microviscosity g (by ESR) of NaTC + L-DPPC + D-DPPC aggregates do not change withthe L-DPPC fraction in the lipid portion. Aggregate properties remain invariant withsubstrate concentration. The NaTC and total L+D-DPPC concentrations are eachconstant. The L-DPPC/D-DPPC ratio is varied. XL-DPPC = l-DPPC

½l-DPPC�þ½d-DPPC�. H and g arereferred to their values H0 (= 0.43) and g0 (= 0.17 cP) at XL-DPPC = 0 (XD-DPPC = 1).

Fig.A2. The aggregation number (by TRFQ) and hydration index (by ESR) ofNaTC + L-DPPC aggregates change with total NaTC + L-DPPC concentration evenwhen the NaTC/lipid ratio is constant.

Conclusions

The new surface dilution assay using the D-enantiomers of thephospholipid substrate as diluents represents a significant advancein interface enzymology because the surface substrate concentra-tion can be varied without affecting the underlying kinetic med-ium. The results show that the D-enantiomers bind to enzyme,resist hydrolysis, and do not change medium properties whensubstituted for the substrate L-enantiomers. In this sense, theyare neutral diluents. The medium effects can be disentangled fromthe substrate concentration effects, and kinetic parameters specificto the interface medium can be determined. The assay permits sep-aration and characterization of the individual steps of hydrolysis.

As an illustration, the assay was employed to determine the ki-netic parameters for the interfacial activity of bee venom PLA2 inNaTC + phospholipid aggregates for a particular lipid mole fractionof 0.14. Conducting the assay on surface dilution series at differentmole fractions would allow the investigation of the dependence ofthe kinetic parameters on interface structure.

The idea of using enantiomeric mixtures should be applicable tokinetic characterization of other stereoselective surface acting en-zymes. The assay is also applicable to membrane models andshould open the door to characterize the kinetics of membranelipolysis and the role of the membrane interface in it. There havebeen no studies of activity dependence on interface substrate con-centration in membranes, presumably due to lack of a method tovary this concentration. Mixed L+D-lipid bilayers can be a platform

for investigating enzymatic kinetics in membranes, and it is the fu-ture course of the current project.

Acknowledgment

The authors thank the National Institutes of Health for its sup-port of this project through grant S06 GM048680.

Appendix A. Characterization of NaTC + DPPC aggregates

The methods used to characterize micellar aggregates and ob-tain the substrate concentration were described in some detail in

258 Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260

our previous publications [20,26–28]. Only a brief outline and dataare presented here.

Aggregation numbers were determined by TRFQ. DMBPquenched decay of the fluorescence from pyrene solubilized inthe aggregates was measured [29–31]. TRFQ measurements werecarried out on several NaTC/L+D-DPPC solutions of various compo-sitions. The aggregate concentrations and the aggregation numberswere determined from fits of the decay curves to the Infelta–Tachiya micellar quenching model [32].

The decay curves indicated small globular (spherical, ellipsoi-dal, or small cylinders with length/diameter ratios of �6 or less)polydisperse aggregates. The weight-averaged aggregation num-bers, Nagg, were determined from the quencher concentrationdependence of the decay fit values [10,33,34]. The unassociated

NaTCH2O DPPC

Lmic

Rmic

HO

CH3

CH3OH

H3C

OH

NaTC

HO

CH3

CH3OH

HNO

H3C

SO OO

OHHO

CH3

CH3OH

HNO

H3C

SO OO

OHH

OH H

OH

H

OH H

OH

H OHH

HO

C

HO

CH3

CH3OH

HN

H3C

O

OH

HO

CH3

CH3OH

HNO

H3C

SO OO

OH

OH

H

OH

H

OH H

HO

CH3

CH3OH

HNO

H3C

SO OO

OH

OH

H

OH

H

OH

H

14.6Ao

OH

H

OH H

HO

CH3

CH3OH

HNO

H3C

SO OO

OH

OH

H

OHH

OH

H

A

C

B

Fig.A3. (A) Components of the aggregate. (B) NaTC + DPPC cylindrical aggregate. (C) Strucaggregate. The dimensions shown are the local radii, estimated by ChemDraw 3D Ultra

NaTC monomer concentration is expected to be less than 10% ofthe [NaTC] + [Total Lipid] (>12 mM in this work) and is neglectedin the calculation of Nagg and [micelles] [10].

The aggregation number does not vary with the L-DPPC/D-DPPCratio for the surface dilution series where the NaTC and total L+D-DPPC are each kept constant as confirmed by Fig. A1. Howeverthe aggregation number does vary with the total concentration ofNaTC and lipid as shown in Fig. A2. The data in Fig. A2 are obtainedwith all L-DPPC in the lipid portion of the mixture because Nagg

does not depend on the proportions of the enantiomers (Fig. A1).Also shown in Figs. A1 and A2 are the hydration index H and

microviscosity g, measured by ESR using the spin probe 5-DSEincorporated into the aggregate [10,26,35]. The hydration indexis the polarity of the environment of the spin probe and is given

HNO

SO OO

OO O

O

O

O

N

POO

H

15. 6Ao

OH H

DPPC

HO

CH3

CH3OH

HNO

H3C

SO OO

OHO

H

OH

H

OH

H

OH H

HO

CH3

CH3OH

HNO

H3C

SO OO

OH

H3

CH3OH

HNO

H3C

SO OO

OH

O

S OO

OH H

OH H

OH

H

OH H

OH

H

OHH

tural formulas and arrangement of NaTC, DPPC, and water molecules in a cylindrical5.0, from which the average radius was derived as in Eq. (A3).

Fig.A4. The total lipid (all L-DPPC in this measurement) interface concentration inunits of Mcm–2 varies with micelle concentration due to the dependence of theaggregation number and hydration on the total NaTC + lipid concentration (asshown in Fig. A2) even when the lipid mole fraction is constant at 1/7. The solid lineis the empirical fit given in Eq. (A6).

Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260 259

by the volume fraction of OH dipoles (including those from wateras well as NaTC) in that environment [35,36]. The precision errorsin H and g are less than 1%.

Appendix B. Model calculations of [L]s

Taking the entire lipid population to be solubilized in the aggre-gates, the surface substrate concentration is

½l�s ¼½l -DPPC�

Amic½micelles� ; ðA1Þ

where Amic is the aggregate surface area. The following procedure isused to calculate Amic.

Mixed aggregates of bile salts and phospholipids are generallyknown to form cylindrical aggregates [10]. The aggregate and itscomponents are shown in Fig. A3. In the current calculation, theaggregate form is represented as a cylinder whose volume, Vmic

(with radius and length denoted by Rmic and Lmic, respectively), is

Vmic ¼ NDPPCVL þ NNaTCVTC—OH þ VOH ¼ pR2micLmic: ðA2Þ

Hydration index data (Figs. A1 and A2) show that the OH dipolesfrom NaTC and the associated water occupy 43% of the volume inthe spin probe environment. Therefore, water is an important mi-celle component. The volume VOH in Eq. (A2) includes the volumeof aggregate associated water and the volume of the OH portionof the NaTC. VTC–OH is the volume of the taurocholate excludingthe OH (ffi 496 Å3), and VL is the volume of a phospholipid molecule(ffi 797 Å3) [10,37]. NDPPC and NTC are the numbers of the phospho-lipid and taurocholate molecules in the aggregate, respectively. Thecounterion (sodium) volume is not included because it is negligiblein comparison with volumes of the phospholipid and taurocholateheadgroups. The arrangement of the molecules in the aggregate isvisualized in Fig. A3. Such an organization is attributed to the non-linear boat-like architecture of the bile salt molecule resulting fromthe differences in the planes of the headgroup and each of the ste-roid rings and to the flexibility of the lipid chains [10]. The surfaceof the cylinder is nonuniform because of the heterogeneity in thesizes and shapes of the molecules in the aggregate. The occurrenceof the largest radius of the aggregate is due to the (taurocholateheadgroup + steroid rings) (14.6 Å) + cylindrical radius of lipid chain(�1 Å). The smallest radius occurs in the bile salt-rich region and isapproximately 14.6 Å. Thus, the compositionally averaged radius is

Rmic ¼ 15:6XmicDPPC þ 14:6Xmic

TC ¼ 15:6XDPPC þ 14:6XTC ¼ 14:7 Å;

ðA3Þ

where XmicDPPC and Xmic

TC are the mole fractions of DPPC and NaTC in theaggregate, respectively, and are taken to be equal to their solutionconcentrations because the free monomer concentrations are insig-nificant compared with the total concentrations, particularly for thelipids. For the same reason, the NDPPC and NTC are also taken as theirmolar concentration fractions of the total aggregation number.

The volume of the spin probe environment is taken to be the to-tal micelle volume, so that

H ¼ VOH

Vmic: ðA4Þ

In micelles of conventional linear surfactants, the ESR signal fromthe spin probe is due to the polar nitroxide moiety that points tothe polar interface of micelles, and in this case the spin probe envi-ronment is the interface [26]. On the other hand, in bile salt assem-blies containing lipids, the majority of the steroid ring volume ispart of the aggregate surface as a result of the boat shape of the bilesalt molecule (Fig. A3). The contribution of the NaTC headgroup tothe aggregate radius is much greater than that of the steroid rings.Therefore, there is no clear subdivision of the aggregate volume into

an interface volume and hydrophobic core volume as in micelles ofconventional linear surfactants. (For this reason, bile salt assembliesare often described as aggregates rather than as micelles.)

Applying Eq. (A4), VOH in Eq. (A2) is replaced byHVmic ¼ HpR2

micLmic. Eq. (A2) can then be solved for Lmic using thenumerical value of Rmic given by Eq. (A3). Amic is then calculatedusing

Amic ¼ 2pR2mic þ 2pRmicLmic: ðA5Þ

Substitution in Eq. (A1) yields [L]s.The importance of the calculation of the interface substrate con-

centration is illustrated in Fig. A4, which shows that it is not con-stant, although XDPPC is constant. The lipid surface concentrationdependence on [micelles] is an outcome of the aggregation numberdependence on the total solute concentrations of bile salts and lip-ids. The data are described empirically by the fit

½Total Lipid�s ¼ 1:41� 10�11 þ 7:34� 10�16

½micelles� � 4:76

� 10�10 ½micelles�: ðA6Þ

This dependence is of consequence in treating the data of a1 versus[micelles] in Fig. A2 because of its presence in Eq. (4).

References

[1] E.A. Dennis, The growing phospholipase A2 superfamily of signal transductionenzymes, Trends Biochem. Sci. 22 (1997) 1–2.

[2] W.A. Pieterson, J.C. Vidal, J.J. Volwerk, G.H. de Haas, Zymogen-catalyzedhydrolysis of monomeric substrates and the presence of a recognition site forlipid–water interfaces in phospholipase A2, Biochemistry 13 (1974) 1455–1460.

[3] R.A. Deems, B.R. Eaton, E.A. Dennis, Kinetic analysis of phospholipase A2activity toward mixed micelles and its implications for the study of lipolyticenzymes, J. Biol. Chem. 250 (1975) 9013–9020.

[4] W.J. Hoffman, M. Vahey, J. Hajdu, Pancreatic porcine phospholipase A2catalyzed hydrolysis of phosphatidycholine in lecithin–bile salt mixedmicelles: kinetic studies in a lecithin–sodium cholate system, Arch. Biochem.Biophys. 221 (1983) 361–370.

[5] M.K. Jain, O. Berg, The kinetics of interfacial catalysis by phospholipase A2 andregulation of interfacial activation: hopping versus scooting, Biochim. Biophys.Acta 1002 (1989) 127–156.

260 Enzymatic lipolysis by bee venom PLA2 / J. Singh et al. / Anal. Biochem. 407 (2010) 253–260

[6] O.G. Berg, M.H. Gelb, M.-D. Tsai, M.K. Jain, Interfacial enzymology: the secretedphospholipase A2 paradigm, Chem. Rev. 101 (2001) 2613–2653.

[7] F. Pattus, A.J. Slotboom, G.H. De Haas, Regulation of the interaction ofpancreatic phospholipase A2 with lipid–water interfaces by calcium2+ ions: amonolayer study, Biochemistry 18 (1979) 2698–2702.

[8] M. Gudmand, S. Rocha, N.S. Hatzakis, K. Peneva, K. Müllen, D. Stamou, H. Uji-I,J. Hofkens, T. Bjørnholm, T. Heimburg, Influence of lipid heterogeneity andphase behavior on phospholipase A2 action at the single molecule level,Biophys. J. 98 (2010) 1873–1882.

[9] J. Singh, R. Ranganathan, J. Hajdu, Kinetics of bacterial phospholipase C activityat micellar interfaces: effect of substrate aggregate microstructure and a modelfor the kinetic parameters, J. Phys. Chem. B 112 (2008) 16741–16751.

[10] J. Singh, Z. Unlu, R. Ranganathan, P.C. Griffiths, Aggregate properties of sodiumdeoxycholate and dimyristoylphosphatidylcholine mixed micelles, J. Phys.Chem. B 112 (2008) 3997–4008.

[11] J.T. Nickels, R.J. Buxeda, G.M. Carman, Purification, characterization, andkinetic analysis of a 55-kDa form of phosphatidylinositol 4-kinase fromSaccharomyces cerevisiae, J. Biol. Chem. 267 (1992) 16297–16304.

[12] R.J. Buxeda, J.T. Nickels, C.J. Belunis, G.M. Carman, Phosphatidylinositol 4-kinase from Saccharomyces cerevisiae: kinetic analysis using Triton X-100/phosphatidylinositol-mixed micelles, J. Biol. Chem. 266 (1991) 13859–13865.

[13] M.F. Roberts, Assays of phospholipases on short-chain phospholipids, MethodsEnzymol. 197 (1991) 95–113.

[14] G.M. Carman, R.A. Deems, E.A. Dennis, Lipid signaling enzymes and surfacedilution kinetics, J. Biol. Chem. 270 (1995) 18711–18714.

[15] O.H. Lowry, N.J. Rosebrough, A. Lewis Farr, R.J. Randall, Protein measurementwith the folin phenol reagent, J. Biol. Chem. 193 (1951) 265–275.

[16] R.K. Scopes, Measurement of protein by spectrophotometry at 205 nm, Anal.Biochem. 59 (1974) 277–282.

[17] C.M. Stoscheck, Quantitation of protein, Methods Enzymol. 182 (1990) 50–69.[18] O.G. Berg, B.Z. Yu, J. Rogers, M.K. Jain, Interfacial catalysis by phospholipase A2:

determination of the interfacial kinetic rate constants, Biochemistry 30 (1991)7283–7297.

[19] F.H. Quina, P.M. Nassar, J.B.S. Bonilha, B.L. Bales, Growth of sodium dodecylsulfate micelles with detergent concentration, J. Phys. Chem. 99 (1995) 17028–17031.

[20] R. Ranganathan, L. Tran, B.L. Bales, Surfactant and salt induced growth ofnormal alkyl sodium sulfate micelles well above their CMC, J. Phys. Chem. B104 (2000) 2260–2264.

[21] R. Ranganathan, L.T. Okano, C. Yihwa, F.H. Quina, Growth ofcetyltrimethylammonium chloride and acetate micelles with counterionconcentration, J. Colloid Interface Sci. 214 (1999) 238–242.

[22] W. Nieuwenhuizen, H. Kunze, G.H. De Haas, Phospholipase A2 (phosphatideacylhydrolase, EC 3.1.1.4) from porcine pancreas, Methods Enzymol. B 32(1974) 147–154.

[23] A. Aloulou, J.A. Rodriguez, S. Fernandez, D. van Oosterhout, D. Puccinelli, F.Carrière, Exploring the specific features of interfacial enzymology based onlipase studies, Biochim. Biophys. Acta 1761 (2006) 995–1013.

[24] M. Alves, M. Peric, An EPR study of the interfacial properties ofphosphatidylcholine vesicles with different lipid chain lengths, Biophys.Chem. 122 (2006) 66–73.

[25] M. Alves, B.L. Bales, M. Peric, Effect of lysophosphatidylcholine on the surfacehydration of phospholipid vesicles, Biochim. Biophys. Acta 1778 (2008) 414–422.

[26] J.S. Singh, J. Miller, R. Ranganathan, Physicochemical characterization ofphospholipid solubilized mixed micelles and a hydrodynamic model ofinterfacial fluorescence quenching, J. Phys. Chem. B 111 (2007) 9317–9324.

[27] B.L. Bales, R. Ranganathan, P.C. Griffiths, Characterization of mixed micelles ofSDS and a sugar based nonionic surfactant as a variable reaction medium, J.Phys. Chem. B 105 (2001) 7465–7473.

[28] R. Ranganathan, C. Giongo, M.S. Bakshi, B.L. Bales, J. Hajdu, Phospholipidcontaining mixed micelles: characterization of diheptanoylphosphatidylcholine(DHPC) and sodium dodecyl sulfate and DHPC and dodecyl trimethylammoniumbromide, Chem. Phys. Lipids 135 (2005) 93–104.

[29] M.H. Gehlen, F.C. De Schryver, Time-resolved fluorescence quenching inmicellar assemblies, Chem. Rev. 93 (1993) 199–221.

[30] M. Almgren, J.E. Löfroth, Determination of micelle aggregation numbers andmicelle fluidities by time-resolved fluorescence quenching studies, J. ColloidInterface Sci. 81 (1981) 486–499.

[31] M. Almgren, in: M. Grätzel, K. Kalyanasundaram (Eds.), Kinetics and Catalysisin Microheterogeneous Systems, Marcel Dekker, New York, 1991, pp. 63–113.

[32] M. Tachiya, Application of a generating function to reaction kinetics inmicelles: kinetics of quenching of luminescent probes in micelles, Chem. Phys.Lett. 33 (1975) 289–292.

[33] M. Almgren, S. Swarup, Size of sodium docecyl sulfate micelles in the presenceof additives: 2. Aromatic and saturated hydrocarbons, J. Phys. Chem. 86 (1982)4212–4216.

[34] S. Reekmans, D. Bernik, M. Gehlen, J. Van Stam, M. Van, F.C. der Auweraer, De.Schryver, Change in the micellar aggregation number or in the sizedistribution? A dynamic fluorescence quenching study of aqueouscetyltrimethylammonium chloride, Langmuir 9 (1993) 2289–2296.

[35] B.L. Bales, L. Messina, A. Vidal, M. Peric, O.R. Nascimento, Precision relativeaggregation number determinations of SDS micelles using a spin probe: amodel of micelle surface hydration, J. Phys. Chem. 102 (1998) 10347–10358.

[36] P. Mukerjee, C. Ramachandran, R.A. Pyter, Solvent effects on the visible spectraof nitroxides and relation to nitrogen hyperfine coupling constants:nonempirical polarity scales for aprotic and hydroxylic solvents, J. Phys.Chem. 86 (1982) 3189–3197.

[37] Y.H. Zhao, M.H. Abraham, A.M. Zissimos, Fast calculation of van der Waalsvolume as a sum of atomic and bond contributions and its application to drugcompounds, J. Org. Chem. 68 (2003) 7368–7373.