INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Secreted phospholipases A₂ (sPLA₂) are a group of soluble enzymes that catalyze the hydrolysis of the 2-acyl group in phospholipids, producing fatty acid and lysophospholipid as reaction products. As a class, sPLA₂ enzymes have a high degree of structural similarity and are believed to have a common catalytic mechanism (Heinrikson et al., 1977; Welches et al., 1993). Because of their stability and the simplicity of the purification protocols, sPLA₂s have been used as models to understand the more complicated cellular sPLA₂ groups.

The lateral organization of membrane lipids has been known to be an important parameter in the enzymatic hydrolysis of membrane phospholipids by sPLA₂. It was recognized early that the phospholipase enzymes preferred an organized lipid substrate near the lipid's phase transition and were particularly active against micellar lipids. To explain the activity of sPLA₂ against lipid bilayers, many researchers postulated the existence of transient bilayer "surface defects." These defects were imagined as membrane gaps that served to increase the affinity of the enzyme for the interface and facilitate enzyme access to the sn-2 position fatty acid through the disruption of the tightly organized lipid lattice. (Bell et al., 1996; Burack et al., 1997a, 1993; Grainger et al., 1989, 1990; Grandbois et al., 1998; Kensil and Dennis 1979; Op den Kamp et al., 1975). The exact molecular details of how the lipid packing influences sPLA₂ activity are still being investigated today.

Physical evidence of membrane defects and their correlation with sPLA₂ function is rare, but a number of studies have addressed this issue. One excellent example is the detailed work on mixed cholesterol-phospholipid in which a strong correlation between the formation of lipid superlattices and the precipitous decrease in sPLA₂ activity was clearly shown (Liu and Chong, 1999). Also, in phospholipid monolayer studies, physical spaces between phospholipid domains were identified as the starting points for sPLA₂ hydrolysis (Grainger et al., 1989). Along similar lines, products accumulated during the enzymatic lag period have been shown to lead to a characteristic burst in sPLA₂ hydrolysis activity, an effect which has been attributed to phase separation between substrate and products (Burack et al., 1997a, 1993; Dahmen-Levison et al., 1998; Grainger et al., 1989; Maloney and Grainger, 1993). Transient membrane irregularities created by lipid phase coexistence have been of interest to the phospholipase research community since the discovery that vesicles equilibrated at temperatures within their phase-coexistence region show the highest susceptibility to sPLA₂ attack. These types of defects are of likely biological significance, as natural membranes contain a variety of lipids, and the potential for forming packing arrangements containing mismatched borders is high.

Studies on the interaction of sPLA₂ and organized lipid interfaces have been conducted using a variety of systems, including small and large unilamellar vesicles, (SUVs and LUVs), multilamellar vesicles (MLVs), as well as monolayers at the air-water interface. Because of their particular characteristics (size and lamellarity) these model membrane systems are not necessarily accurate descriptions of cell membranes. Giant unilamellar vesicles (GUVs) with a mean diameter of 30 µm have a minimum curvature and mimic cell membranes in this respect. GUVs are ideal for studying lipid/lipid and lipid/protein interactions using microscopy techniques (Holopainen et al., 2000; Longo et al., 1998; Mengar and Keiper, 1998; Wick et al., 1996).

In this work, we explore two questions about the interaction of sPLA₂ with organized lipid interfaces: 1) does the interaction of the enzyme with organized lipids such as GUVs produce morphological change in the substrate vesicles, and 2) when the enzyme interacts with vesicles showing lipid domain separations and "defects," does the enzyme preferentially bind to a particular area, namely a particular domain or to the border between different domains?

In our experimental approach, we use two-photon microscopy, Crotalus atrox sPLA₂ and two classes of GUVs; namely, single lipid GUVs above their transition temperatures and mixed-lipid GUVs showing domain separation. In both mixed and homogeneous GUVs, we directly visualize the interaction of the vesicles with the enzyme by two-photon microscopy. To measure the binding of the enzyme to the vesicles, we used fluorescein-labeled sPLA₂ in the presence of Ba²⁺ (to inhibit activity but retain enzyme association with the interface). Activity measurements were done by addition of active enzyme (sPLA₂ plus Ca²⁺) to the GUVs. Fluorescent dyes were used to visualize the GUVs and to study the packing changes in the membrane (6-lauroyl-2-(dimethylamino) napthalene (LAURDAN)), and to visualize and identify liquid crystal and gel domains (Lissamine Rhodamine B 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (N-Rh-DPPE) and 6-propionyl-2-(dimethylamino) naphthalene (PRODAN)).

Our results show, for C. atrox sPLA₂, Naja naja naja, and Agkistrodon piscivorus piscivorus, that the interaction of the enzyme with GUV induces dramatic morphological changes in the vesicles. The vesicles shrink until the liquid phase disappears and, in the case of GUVs presenting liquid/gel-phase separation (for C. atrox sPLA₂), only the solid domains remain at the end of the shrinkage processes. Binding experiments show homogeneous binding of the C. atrox sPLA₂ enzyme to the liquid crystal phospholipid domains with no measurable preference for the borders between domains. The hydrolysis kinetic results are consistent with this conclusion.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals

Phospholipids were obtained from Avanti Polar Lipids (Alabaster, AL) and used without further purification: phospholipids used were: 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC); 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC); 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC); 1,2-diarachidoyl-sn-glycero-3-phosphocholine (DAPC); 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine (DMPE); 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (DPPE); 1-palmitoyl-2-oleoyl-phosphatidylcholine (POPC); PRODAN; LAURDAN; and N-Rh-DPPE. These were purchased from Molecular Probes (Eugene, OR). EDTA and Tris (hydroxymethyl) aminomethane hydrochloride (Trizma HCl) were obtained from Sigma Biochemicals (St. Louis, MO).

PLA₂ and fluorescent conjugates

The dimeric sPLA₂ from C. atrox venom (Miami Serpentarium, Punta Gorda, FL) was purified in our laboratory by using published procedures (Hachimori et al., 1971). The monomeric sPLA₂ from N. n. naja venom was a gift from Dr. Edward Dennis (Dept. Chemistry, University of California, La Jolla, CA). The monomeric sPLA₂ from A. p. piscivorus venom was a gift from Dr. John Bell (Brigham Young University, Provo, UT). sPLA₂ from N. n. naja venom was obtained from Sigma, without further purification. The concentration of C. atrox, A. p. piscivorus, and N. n. naja sPLA₂ was determine by their extinction coefficients of ₂₈₀ = 25,000 M¹ cm¹ (Brunie et al., 1985), ₂₈₀ = 30,800 M¹ cm¹ (Bell et al., 1996) and ₂₈₀ = 29,300 M¹ cm¹(Darke et al., 1980), respectively.

The fluorescein conjugates of the C. atrox sPLA₂ were prepared by labeling the enzyme with fluorescein succinimidyl ester (Molecular Probes) as previously described (Sanchez et al., 2001). Conjugates were routinely labeled with an average of 1 fluorescein (₄₉₉ = 70,000 M¹ cm¹ (Jablonski et al., 1983) per sPLA₂ monomer. The sPLA₂ specific activity was measured using a pH-stat, and mixed micelle assay at pH = 8.0 (Dennis, 1973). No significant differences in enzyme activity between the fluorescein-labeled and unlabeled sPLA₂ species were found.

Vesicle preparation

Phospholipid stock solutions were prepared in chloroform at concentrations of 0.2 mg/ml. In our experiments we formed GUVs with POPC, DMPC, and DPPC and the binary mixtures DMPE/DMPC (7:3 molar ratio) and DLPC/DAPC (1:1 molar ratio).

The electroformation method, developed by Angelova and Dimitrov (Angelova and Dimitrov, 1986; Angelova et al., 1992; Dimitrov and Angelova, 1987) was used to prepare the vesicles. GUVs were formed in a temperature-controlled chamber that allows a working temperature range from 9°C to 80°C (Bagatolli and Gratton, 1999, 2000a). GUVs were prepared using the following steps: ~2 µl of the lipid stock solution was spread on each of the two sample chamber platinum wires under a stream of dry N_2. The chamber was then lyophilized for ~1 h to remove any remaining trace of organic solvent. The chamber and the buffer (Tris 0.5 mM, pH = 8) were separately equilibrated to temperatures above the lipid mixture phase transition(s) (~10°C over the corresponding transition temperature) and then 4 ml of buffer was added to cover the wires. Immediately after buffer addition, the platinum wires were connected to a function generator (Hewlett-Packard, Santa Clara, CA) and a low-frequency alternating field (sinusoidal wave function with a frequency of 10 Hz and an amplitude of 2V) was applied for 90 min. The AC field was turned off after GUVs were formed and the temperature was adjusted to the desired temperature. In the binary lipid mixtures the temperature was decreased to the gel/fluid phase-coexistence temperature regime (44°C for DMPE/DMPC and 46.6°C for DLPC/DAPC). The temperature was measured inside the chamber at the platinum wires, using a digital thermometer (model 400B, Omega, Stamford, CT) with a precision of 0.1°C. The experiments were carried out in the same chamber after the vesicle formation with an inverted fluorescence microscope (Axiovert 35, Zeiss, Thornwood, NY). The vesicles remained attached to the platinum wires, which allowed us to perform the particular experiments on single GUVs without vesicle drifting. A charge-coupled device video camera (CCD-Iris, Sony, Tokyo, Japan) was used to follow GUV formation and to select the target vesicle. For experiments involving LAURDAN and PRODAN, fluorophore (in dimethyl sulfoxide (DMSO)) was added to the sample chamber after the vesicles were formed. The final bilayer phospholipid:fluorophore ratio was kept greater than 100:1. The vesicles were homogeneously labeled within 15 min of fluorophore addition. The final concentration of DMSO in the sample was <0.1%. In the case of N-Rh-DPPE, the fluorescent phospholipid was premixed in chloroform with the primary lipid(s) and spread onto the sample chamber wires before GUVs were formed. The concentration of N-Rh-DPPE in the samples was 0.5 mol%.The mean diameter of the GUVs was ~30 µm, as previously reported (Bagatolli and Gratton 2000a, 2000b).

Two-photon intensity images

Images were collected on a scanning two-photon fluorescence microscope designed in our laboratory (So et al., 1995, 1996). A LD-Achroplan 20X long working distance air objective with a N.A. of 0.4 (Zeiss, Holmdale, NJ) was used. A mode-locked titanium-sapphire laser (Mira 900, Coherent, Palo Alto, CA) pumped by a frequency-doubled Nd:vanadate laser (Verdi, Coherent) set to 780 nm, was used as the two-photon excitation light source. A galvanometer-driven x-y scanner was positioned in the excitation path (Cambridge Technology, Watertown, MA) to achieve beam scanning in both the x and y directions. The samples received from 5 to 9 mW of 780 nm excitation light, and a frame rate of 9 s/frame was used to acquire the 256 × 256 pixel images. A quarter wave-plate (CVI Laser Corporation, Albuquerque, NM) was aligned and placed before the light entered the microscope to minimize the polarization effects of the excitation light. The fluorescence emission was observed through a broad band-pass filter from 350 nm to 600 nm (BG39 filter, Chroma Technology, Brattleboro, VT). A miniature photomultiplier (R5600-P, Hamamatsu, Bridgewater, NJ) was used for light detection in the photon counting mode. A commercialized version (ISS card) of a card designed in our lab (Eid et al., 2000) was used to acquired the counts. A two-channel detection system was attached for generalized polarization function (GP) image collection (see below for details).

The sPLA₂, either native or labeled with fluorescein, was prepared in the buffer used to prepare the vesicles (0.5 mM Tris pH = 8). Ca²⁺ or Ba²⁺ was added to the buffer as needed for the specific protocol. The desired amount of PLA₂ was then deposited with a microinjector needle or with a pipette into the chamber containing the target GUV. The addition of buffer and salts in the absence of the sPLA₂ was used as the control. For the binding experiments, the GUV images were collected following the addition of the fluorescein-labeled sPLA₂ (in presence of Ba²⁺). Imaging the fluorescein-labeled sPLA₂ bound on the GUV surface required integration of ~10-80 frames (9 s/frame) because of the low fluorescence intensity from the bound fluorescein-sPLA₂. The lipid phase-state of the enzyme-bound region was determined through the addition of PRODAN, which shows a preferential partition into the lipid fluid phase (Bagatolli and Gratton, 2000b; Krasnowska et al., 1998). For the protein activity experiments, successive images of LAURDAN and N-Rh-DPPE labeled GUV were taken immediately after the addition of the active sPLA₂ (in presence of Ca²⁺). Single lipid vesicles, LAURDAN-doped and N-Rh-DPPE-labeled, were sufficiently bright to image with one frame.

Lipid domain area calculations were carried out using procedures written within Igor (Wave Metrics, Lake Oswego, OR). Pixels were defined as "gel" or "liquid" by their GP values and summed to determine the corresponding areas (see below for GP description).

LAURDAN GP measurements

LAURDAN was used as a membrane probe because of its ability to report the extent of water penetration into the bilayer surface. Water penetration has been strongly correlated with lipid packing and membrane fluidity (Parasassi et al., 1990, 1998; Parasassi and Gratton, 1995). The emission spectrum of LAURDAN in a single phospholipid bilayer is centered at 490 nm when the membrane is in the gel phase and at 440 nm when in the liquid crystalline phase. The GP gives a mathematically convenient and quantitative way to measure the emission shift. The function is given by:
where I_B and I_R are the blue (440 nm) and red (490 nm) emission intensities, respectively. A full discussion on the use and mathematical significance of GP can be founded in the literature (Bagatolli and Gratton, 1999; Parasassi et al., 1991). In our two-photon, dual-channel instrument the excitation wavelength used for the two-photon images was 780 nm. The fluorescence was split into red and blue channels by using a Chroma Technology 470DCXR-BS dichroic beam splitter in the emission path. Interference filters, either an Ealing 490 or an Ealing 440, were placed in the appropriate emission paths to further isolate the red and blue parts of the emission spectrum. Separate detectors were used for each channel to simultaneously collect the 490 nm and 440 nm emission. Corrections for the wavelength dependence of the emission detection system was accomplished through the comparison of known solutions (LAURDAN in DMSO or LAURDAN in DMPC vesicles at 20°C (Bagatolli and Gratton, 1999)), taken on an ISS Inc (Champaign, Urbana, IL) model PC1 steady-state fluorometer. The separate red and blue images were simultaneously collected and then recombined to form the GP image of the sample.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Interaction of C. atrox sPLA₂ with single lipid GUVs

GUVs were visualized with the membrane fluorophore LAURDAN during experiments on sPLA₂ hydrolysis. LAURDAN integrates into the membrane and emits with sufficient intensity to follow the changes produced on the GUVs after the addition of the sPLA₂. LAURDAN also allows us to quantify the fluidity changes in the membrane by measuring GP (Parasassi et al., 1991). Fig. 1 a shows the time sequence of fluorescence intensity images obtained before and after the addition of dimeric C. atrox sPLA₂ to DPPC GUVs at 53°C in the liquid phase. Calcium cofactor, essential for enzymatic activity, was added together with the enzyme. After the addition of the protein, the vesicle shrinks until there is virtually nothing of the vesicle remaining. Time-dependent changes in the estimated surface area of the DPPC GUV as hydrolysis proceeds are plotted in Fig. 1 b. Shrinking responses were also observed with DMPC (Fig. 2) and POPC (data not shown) vesicles.

View larger version (39K): [in this window] [in a new window]   FIGURE 1   PLA2-dependent shrinking of a DPPC GUV at 53°C is shown. Fluorescence intensity images of LAURDAN-containing DPPC GUV were collected before and after the addition of C. atrox sPLA2. The point of enzyme addition (10 µl of 8.5× 106 M sPLA2 in 100 µM Ca2+) is noted by an arrow. The images are shown using false color representation according to the figure scales, and the numbers indicate the time in seconds. The vesicle radii, in pixels, were determined directly from the vesicle cross-sections. The surface areas were then calculated assuming a spherical shape, and a plot of the surface area as a function of time is shown in (b).

View larger version (32K): [in this window] [in a new window]   FIGURE 2   Time-dependent changes in the LAURDAN GP of DMPC GUVs after the addition of C. atrox sPLA2. Numbers indicates the time in minutes, and the arrow indicates enzyme addition (of 7 × 107 M sPLA2 in 100 µM Ca2+). The temperature was kept constant at 26.1°C during the experiment. The GP images are obtained by processing the simultaneously collected red and blue images as described in the Materials and Methods section.

GUV membrane-phase state was evaluated during the hydrolysis process through the LAURDAN GP images. A series of GP images that follow the time course for sPLA₂ hydrolysis of a DMPC vesicle cluster is shown in Fig. 2. The GUV cluster was equilibrated at 26.1°C, several degrees above the main transition phase for this phospholipid. The GP images (Fig. 2) show an increase in the observed average membrane GP from 0.37 to 0.47 (Fig. 3, a and b) as the bilayer phospholipids are hydrolyzed. The GP histograms for these images have a clear GP center and distribution, which can be plotted as a function of time (Fig. 3 b). The width of the distribution remains, within experimental error, constant throughout the hydrolysis process.

View larger version (18K): [in this window] [in a new window]   FIGURE 3   The membrane GP histograms shown were calculated from the images in Fig. 2. In (a) the GP histograms sPLA2 before sPLA2 addition () (first image in Fig. 2) and during the hydrolysis but before vesicle collapse () (second to last image in Fig. 2). The Gaussian center and width of the GP histograms are shown as a function of time in (b).

The binding of the sPLA₂ to the single lipid GUVs was observed using a fluorescein-labeled C. atrox sPLA₂ conjugate. Because sPLA₂ activity on the GUV's interface would alter the membrane character, which would influence the sPLA₂ interfacial binding, Ba²⁺, a mimic of the Ca²⁺ cofactor, known to inhibit enzymatic activity but permit interfacial association (Yu et al., 1998, 1993), was used in the binding studies. The association of the fluorescein-sPLA₂ conjugate to single lipid vesicles in the liquid phase is shown in Fig. 4. The cross-section and surface images of POPC and DPPC vesicle were obtained by changing the position of the z-focus in the microscope. Under conditions in which the phospholipids were in the liquid crystalline state, the signal from the C. atrox fluorescein-labeled sPLA₂ was found to evenly cover the POPC and DPPC vesicle surfaces.

View larger version (96K): [in this window] [in a new window]   FIGURE 4   Two-photon excitation images of single lipid GUVs after addition of C. atrox fluorescein-sPLA2. (a) Sequence of events when a microinjector needle was used to spread the labeled protein on the surface of a POPC GUV; the needle is located close to the target vesicle (just the needle containing the fluorescent protein can be seen) (left panel), the labeled enzyme is spread on the GUV surface which can already be visualized as a diffuse bright region in the center panel, and the needle is removed and the homogeneous labeled GUV is observed after some of the excess labeled protein has diffused away (right panel). (b) The cross-section view of POPC vesicles after the addition of 63 × 103 fmol (added by microinjector) of C. atrox sPLA2 8× 106 M in buffer containing 150 mM Ba2+ on the GUV surface. Temperature was maintained at 51°C. (c) DPPC vesicles after the addition (added by pipette) of 20 µl of sPLA2 3.2× 106M in 100 µM Ba2+.

Interaction of C. atrox sPLA₂ with mixed lipid GUVs

The effect of enzyme activity toward mixed-lipid GUVs was examined using N-Rh-DPPE, which has been shown to partition into the liquid domains in these binary lipid GUVs (Bagatolli and Gratton, 2000b). Images were collected after the addition of sPLA₂ to single GUVs composed of DMPE/DMPC (Fig. 5 a) and DLPC/DAPC (Fig. 5 b) at 44°C and 46°C, respectively. For both lipid mixtures, and before the addition of the enzyme, the GUVs displayed leaf-shaped gel domains surrounded by a fluid phase. This phase separation is stable if the temperature is kept constant, in agreement with previous observation (Bagatolli and Gratton, 2000b). After the addition of the sPLA₂ (arrow in Fig. 5, a and b), the vesicles were observed to decrease in size to a small mass of lipid. For both binary lipid mixtures, as the vesicle size decreased, the solid domains became deformed and the fluorescent (liquid) domains disappeared. Fig. 6, a and b, shows the time-dependent changes of the area corresponding to the total, the gel and the fluid regions of the fluorescent images illustrated in Fig. 5. A faster decrease of the liquid crystalline phase with respect to the gel phase is observed in both mixtures. The time needed to eliminate the liquid crystalline part of the vesicle was ~300 s for both experiments.

View larger version (106K): [in this window] [in a new window]   FIGURE 5   Series of two-photon excitation images of DMPE:DMPC (a) and DLPC:DAPC (b) GUVs before and after the addition of a C. atrox sPLA2. The red and the dark areas in the vesicles correspond to the fluid and gel phases, respectively (false color representation is according to the scale on the bottom of the figure). The fluorescent images were taken at the vesicle surface. (a) The addition of 15 µl of sPLA2 2 mg/ml in 150 mM Ca2+ at 44°C. (b) The addition of 20 µl of sPLA2 2 mg/ml in 150 mM Ca2+ at 46°C. N-Rh-DPPE final concentration was 0.5%. Numbers indicate the time in seconds.

View larger version (18K): [in this window] [in a new window]   FIGURE 6   Time-dependent changes of the area corresponding to the total (), the gel () and the fluid (black Xs) regions of the N-Rh-DPPE-labeled GUVs fluorescent images from Fig. 5. (a) DMPE/DMPC 7:3 mol at 44°C. (b) DLPC/DAPC 1:1 mol at 46°C.

Direct observation of sPLA₂ binding to the GUVs was monitored using a fluorescein-labeled sPLA₂ (with Ba⁺²). Labeled enzyme was added to unlabeled GUVs composed of DMPE/DMPC (7:3 molar ratio) and DLPC/DAPC (1:1 molar ratio) mixtures at temperatures corresponding to the gel/fluid phase coexistence. After addition, the enzyme was found to decorate a defined region of the binary GUVs (Fig. 7, a and b, left panel). To determine the phase-state of the protein-labeled membrane region, PRODAN was added to the vesicle immediately after sPLA₂ binding was observed. In a bilayer membrane, PRODAN partitions into the lipid liquid crystalline phase 35-fold greater than the lipid gel phase (Bagatolli and Gratton, 2000b; Krasnowska et al., 1998). PRODAN was found to label the same GUV region that contained the fluorescein-sPLA₂, indicating that the sPLA₂ was binding to the liquid crystalline lipid domains (Fig. 7, a and b, right panel).

View larger version (111K): [in this window] [in a new window]   FIGURE 7   Two-photon excitation images of (a) DMPC/DMPE 7:3 mol GUV at 46.6°C and (b) DLPC/DAPC 1:1 mol GUV at 48.1°C after the addition of 10 µl of C. atrox sPLA2 fluorescein-labeled (12 × 106 M in buffer containing 150 mM Ba2+) (left panel) and after PRODAN addition (right panel). The fluorescent images were taken at the vesicle surface with the false color representation denoted in the figure scale.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

C. atrox sPLA₂ activity and binding on GUVs

Single lipid GUVs

Mixed lipid GUVs

Morphology changes in GUVs

Our primary observation of sPLA₂-dependent GUV morphology change has been the shrinking response to sPLA₂. Gradual loss of vesicle size because of sPLA₂ is surprising. Historically, the general view in PLA₂ research is that membranes, normally SUVs, LUVs, or erythrocytes, remain intact and unaffected as sPLA₂ hydrolysis of the outer leaflet phospholipids proceeds (Gul and Smith, 1974; Jain et al., 1991, 1980). In fact, studies on the hydrolysis kinetics of sPLA₂ toward vesicle phospholipids have often rested on the assumption that the vesicles will remain intact (Jain et al., 1991). Our results here would bring into question some of these long-standing beliefs.

How general are our observations? Can our observation on GUVs be extrapolated to standard bulk experiments on LUVs? Reduction in GUV dimensions indicates a loss of lipid mass and a dramatic change in vesicle volume. The GUVs in the absence of C. atrox sPLA₂ are stable for many hours and thus the mass loss clearly must be attributable to the formation of fatty acid and lysophospholipid from enzymatic digestion by sPLA₂. Lipid loss is likely to occur through the escape of hydrolysis products, alone or in mixtures with undigested phospholipids, in the form of monomers, micelles, or smaller vesicles. The lysophospholipids have a sensible solubility in the aqueous phase and have been shown to desorb from the membrane surface (Kupferberg et al., 1981; Speijer et al., 1996). Let us consider the DPPC GUVs and LUVs, as DPPC is a common lipid used in vesicle studies. Palmitic acid, which has a lower solubility in the aqueous phase than the 1-palmitate-lysophospholipid, would be expected to remain in the GUV at higher concentrations, but it too must eventually be removed for the vesicle to continue the shrinking process. It is our belief that the vesicle initially shrinks through loss of monomer and micellar lysophospholipid. The loss of lysolipid results in a decrease in the GUV internal volume stressing the GUV bilayer, which serves to further extrude mixtures of fatty acid, lysophospholipid, and phospholipid in the form of micelles or smaller vesicles below the resolution of our microscope. Indeed, the partition coefficient measured for palmitate and that estimated for 1-palmitate-lysophospholipid (Anel et al., 1993; Bent and Bell, 1995; Brown et al., 1993) would suggest that these species will significantly desorb from the membrane under the low concentrations of phospholipid used in our studies, ~0.3 uM. In fact, ~99% of the lysolipid and 70% of the fatty acid would be expected to leave the membrane under equilibrium conditions. In contrast, the phospholipid concentrations generally used for bulk vesicle studies are almost three orders of magnitude and the vesicles should retain most of the hydrolyzed product. In retaining the hydrolysis products, the membranes could very well stay intact, as indicated in a number of studies (Kupferberg et al., 1981). However, care should be exercised, as the partition coefficients are temperature dependent and may be affected by the ionic strength and pH of the buffer used. In the low lipid concentration case, a buildup of products will still occur but will be limited by the rate of hydrolysis, the product desorption rates, and the partition coefficients.

The hypothesis that there may be profound effects on vesicle structure because of PLA₂ is not new and has been discussed by Biltonen et al. (Burack et al., 1997a, 1995). Through the use of electron microscopy, ¹³PNMR, and light-scattering techniques, Biltonen collected data that clearly indicated fundamental changes in bulk LUVs attributable to A. p. piscivorus sPLA₂ addition. Their data are consistent with the observed shrinking we report here and suggest that LUVs and GUVs have similar properties with respect to sPLA₂ hydrolysis and vesicle morphology change.

Our results demonstrate that the addition of C. atrox sPLA₂ to GUVs of single-lipid (DPPC and DMPC) and mixed-lipid at their domain's coexisting temperature (DPPC/DPPE, DLPC/DAPC) shows gross morphological changes, both distortions in the vesicle membrane (data not shown) and shrinking of the vesicle size. These effects will likely disturb the local packing structure of the lipids, potentially affecting the binding and hydrolysis of the surface phospholipids by sPLA₂ and thus influence the kinetic profiles collected.

In conclusion, we would like to make two general remarks. First, C. atrox sPLA₂ is a dimeric enzyme unlike most of the sPLA₂s under active investigation and may have its own unique membrane association and interactions. Based on the common opinion that the secreted PLA₂s share a common mechanism of hydrolysis, we may speculate that the interaction observed for C. atrox sPLA₂ and the GUVs may reveal part of that mechanism which is common. However, particular differences depending on the enzyme's origin may change this scenario. We conclude from this study that the C. atrox sPLA₂ binds to the liquid region of zwitterionic vesicles and hydrolyzes the liquid crystalline phase lipids in a homogeneous fashion. Second, independent of the sPLA₂ issues, we show here that two-photon microscopy, as a technique, together with the GUVs as an ideal model for biological membranes, represent a powerful methodology to study the interaction of proteins with organized, unsupported bilayer membranes.