|Home | About | Journals | Submit | Contact Us | Français|
Phospholipases constitute a ubiquitous class of membrane-active enzymes that play a key role in cellular signaling, proliferation, and membrane trafficking. Aberrant phospholipase activity is implicated in a range of diseases including cancer, inflammation, and myocardial disease. Characterization of these enzymes is therefore important, both for improving the understanding of phospholipase catalysis, and for accelerating pharmaceutical and biotechnological applications. This paper describes a novel approach to monitor, in-situ and in real-time, the activity of phospholipase D (PLD) and phospholipase C (PLC) on planar lipid bilayers. This method is based on enzyme-induced changes in the electrical charge of lipid bilayers and on the concomitant change in ion concentration near lipid membranes. The approach reports these changes in local ion concentration by a measurable change in the ion conductance through pores of the ion channel-forming peptide gramicidin A. This enzyme assay hence takes advantage of the amplification characteristics of gramicidin pores to sense the activity of picomolar to nanomolar concentrations of membrane-active enzymes without requiring labeling of substrates or products. The resulting method proceeds on lipid bilayers without the need for detergents, quantifies enzyme activity on native lipid substrates within minutes, and provides unique access to both leaflets of well-defined lipid bilayers; this method also makes it possible to generate planar lipid bilayers with transverse lipid asymmetry.
Phospholipases are membrane-active enzymes that catalyze the hydrolysis of specific ester bonds in phospholipids. These enzymes play a critical role in cell signaling, proliferation, and vesicle trafficking.1–3 As a result, phospholipases are implicated in a range of diseases including cancer, inflammation, and myocardial disease.1–3 To improve the understanding of these enzymes, characterization methods that monitor the activity of phospholipases in situ, in real time, and in a label-free fashion, would be useful. Due to the heterogeneous nature of catalytic reactions on lipid membranes it is, however, often challenging to monitor the activity of membrane-active enzymes.4,5 Established methods to characterize phospholipases measure the absorbance, fluorescence, or radioactivity of their enzymatic products.3,5,6 While radioactive and fluorescent assays offer high sensitivity, these assays are often not performed in situ, and the need for labeling of the substrate can limit the application or affect the results of these assays. Other assays employ coupled enzyme reactions1,7,8 in which the optimum experimental condition may be limited by the second enzyme. More recently, liquid crystals were employed for label-free detection of phospholipase activity, however, this approach is not quantitative.9
Here, we describe a novel, label-free, rapid, and quantitative method to monitor, in situ, the activity of phospholipases D and C (PLD, PLC) on planar lipid bilayers. This approach offers high sensitivity by ion channel amplification10,11 and employs planar lipid bilayers instead of micelles or liposomes. The assay hence provides access to both leaflets of a bilayer—a unique characteristic that makes this assay platform attractive for generating asymmetric lipid bilayers by adding different enzymes to each side of a membrane.
PLD cleaves the phosphodiester bond on the polar side of phospholipid headgroups (Fig. 1a) while PLC cleaves the bond on the glycerol side (Fig. 1b).1,3,5 These phospholipases, therefore, modify the headgroup of phospholipids and, for many lipid substrates, this modification results in a change in the net charge of the lipid molecule. In solutions of low ionic strength (i.e. close to, or below the physiologic range) this enzyme-induced change in the electric charge on the lipids in a membrane results in a change of the local concentration of counter ions near the membrane surface.12 This change in ion concentration can be detected by the channel-forming peptide gramicidin A (gA).11,13,14 Hence, the enzyme assays introduced here, take advantage of the change in the single-channel ion conductance through gA pores, γ, in response to enzyme-catalyzed modifications of the charge of lipids.11,14 Fig. 2 illustrates this concept for PLD.
In order to demonstrate the ability of an assay based on gA to monitor the activity of PLD, we formed planar lipid bilayers containing phosphatidylcholine (PC) lipids as substrate and monitored the changes in single channel conductance of gA, γ, upon addition of PLD from cabbage (EC 188.8.131.52) to both sides of the bilayer.17 After addition of the enzyme to the bilayer, we recorded short (≤ 60 s) current traces at different applied voltages and at distinct points of time after enzyme addition (Fig. 3). To analyze these current versus time traces, we first obtained a histogram that represented the distribution of currents for each current trace (Fig. 3). The difference in current between adjacent peaks in these histograms represented the current amplitude of single gA channel openings. These current amplitudes, therefore, reflect the single channel conductance of gA, γ, at a given time after addition of PLD. By averaging all current amplitudes within a histogram, we obtained the mean current amplitude of gA events and a standard deviation (SD) for these mean current amplitudes for each trace that corresponded to a certain time after addition of the enzyme. We plotted these mean current amplitudes with their SDs as a function of the applied voltage for each time point. From these I–V plots, we obtained a mean value for single channel conductance of gA, γ, at a given time point from the slope in the linear range ≤ |±100 mV| of the plot according to the equation I = R−1 · V = γ · V.
As demonstrated in Fig. 3, the current amplitude of gA opening and closing events increased over time after addition of PLD to the bilayer setup. To further illustrate the shift in the amplitude of single channel currents upon addition of PLD, we plotted the distribution of the step amplitudes as a consequence of each opening and closing event in the original current versus time traces in Fig. 4. These event histograms clearly illustrate the time-dependent increase in step amplitude of gA events after addition of PLD.
Given the reaction equation of PLD (Fig. 1a), we hypothesized that this increase in current amplitude of gA events was due to the formation of the negatively-charged PA lipids (the product of PLD-catalyzed hydrolysis of neutral PC lipids) in the bilayer. In order to test this hypothesis, we measured the current amplitude of gA events in membranes of various PA contents. Fig. 5a and b show the statistical distribution of the current amplitude of gA events in PC bilayers as a function of the mole fraction of PA in the bilayer. As expected, these results illustrated a significant increase in the mean current amplitude of gA events as a result of increasing amounts of PA lipids in the bilayer.
In order to relate, quantitatively, the changes in γ to the changes in amount of produced PA lipids in the membrane, we acquired a calibration curve (Fig. 5c) of γ (in units of pS = 10−12 Ω −1) as a function of the mole fraction of phosphatidic acid (XPA, unitless). Eq. 1, which resulted from the curve fit in Fig. 5c, describes this calibration curve:18
Fig. 6 demonstrates the time-dependent increase in γ after addition of different amounts of PLD to a PC bilayer. Interestingly, this increase progressed with an initial lag phase of slow change, followed by a rapid increase in γ. We considered two possible effects to explain this lag phase (which lasted 5–9 min depending on the enzyme concentration). First, Kuppe et al. have previously reported a similar lag phase of slow activity of PLD on PC bilayers.8 These authors demonstrated that Ca+-induced formation of PA domains in PC membranes19 is a key parameter that facilitates binding of PLD to the membrane surface and significantly increases the hydrolysis rate of PLD.8 Second, mixing of PLD in the electrolyte of the bilayer chamber and binding of PLD to PC membranes may take time.
To estimate the change in XPA in membranes upon addition of PLD, we used Eq. 1. Fig. 7a shows the linear, time-dependent changes in XPA after the lag phase for different PLD concentrations. We defined the lag time as the interval from addition of PLD until XPA reached a value of ~0.03 (see Supporting Information for details). The inset in Fig. 7a reveals that the rate of product formation (dXPA/dt) increased linearly with an increase in enzyme concentration.
To confirm a functional dependence of γ on the enzymatic activity of PLD, we performed a control experiment with heat-denatured PLD and found no increase in γ (Fig. 7b). Furthermore we monitored the changes in γ upon addition of PLD in the presence of two PLD inhibitors, resveratrol3,20 and cyclosporine A.21 Fig. 7b shows that both inhibitors reduced the reaction rate significantly. We also examined the effect of the soluble product of the reaction, choline, on γ and found no significant change in the presence of 20 μM choline (see Supporting Information).
In addition, we examined the effect of the lipid product of the reaction, PA, on the permeability of ions through the bilayer19 in the absence of gA pores and found no significant effect in the examined range of 0–30 mol% PA content (see Supporting information).
Together, these control experiments confirmed that the increase in γ upon addition of PLD was due to enzymatic activity and not due to artifacts.
To explore the potential of this gA-based assay platform to create asymmetric planar lipid bilayers, we exposed one side of a PC bilayer to active PLD and the other side to inactivated PLD (Fig. 8a). We hypothesized that asymmetric activity of the enzyme would create a planar bilayer with transverse asymmetry, in which one leaflet (exposed to active PLD) would contain the catalytic product, i.e. negatively-charged PA lipids, and the other leaflet (exposed to inactivated PLD) would retain unmodified substrate, i.e. neutral PC lipids. We have previously demonstrated that γ is affected only by the electric charges present near the entrance of gA channels and not by the charges near the exit of the pore.11 Since gA is permeable to cations, the entrance of gA pores is always located in the positively polarized compartment of the bilayer setup. Hence, the entrance of gA channels can be switched from one compartment to the other based on the polarity of applied voltage. We, therefore, hypothesized that if asymmetric activity of PLD on the PC bilayer produces a bilayer with transverse lipid asymmetry, gA pores would have a large conductance when the polarity of the voltage would place the entrance of gA on the side of the negatively-charged leaflet (i.e., the side with active PLD), whereas gA pores would have a low conductance when the polarity of the voltage would place the entrance on the side of the neutral leaflet of the bilayer (i.e., the side with inactive PLD).
Fig. 8b shows that, as expected, γ changed asymmetrically depending on the polarity of the applied voltage: at +100 mV, γ increased strongly over time (with this polarity, the entrance of the gA pores was located on the side with active PLD), whereas γ did not increase significantly at −100 mV, when the polarity placed the gA entrance on the side with inactive PLD.
To describe the relationship between γ and the time after addition of PLD, we derived an equation to fit the data in Fig. 8b. As shown by Eq. 1, γ is a hyperbolic function of the mole fraction of PA in the membrane, XPA, while we used a simple model to describe the relationship between XPA and time, t, after addition of PLD by a rate equation for pseudo-first order kinetics of the form:22
where the fitting parameter k (min−1) is a pseudo-first order rate constant of the overall reaction at a given concentration of PLD. Combining Eq. 1 and Eq. 2, we obtained the functional dependence of γ on the time t after addition of PLD:
We introduced t0 (min) in Eq. 3 to account for the lag phase. The solid red curve in Fig. 8b illustrates the best fit of the six points shown in red after the lag phase to Eq. 3, with k = 0.02 min−1 and t0 = 3.16 min for the best fit.
The strong difference between γ as a function of the polarity of the applied voltage in Fig. 8b demonstrates the ability of the platform presented here to create asymmetric lipid bilayers by asymmetric addition of an enzyme to a symmetric planar bilayer. Although not demonstrated in this paper, we think that this capability is attractive since intracellular lipases process lipids in natural biomembranes with strong transverse asymmetry. The assay developed here, may hence make it possible to study lipase activity in bilayers that are better models of physiologic membranes than liposomes or micelles. In addition, the capability of generating asymmetric lipid bilayers by exploiting lipase activity as demonstrated here, may be useful for studying the effect of transverse asymmetry on the activity of other membrane proteins, such as ion channels and transport proteins. And finally, the results from asymmetric addition of enzymes shown in Fig. 8, provide additional evidence that the observed changes in γ were indeed due to the enzyme activity of phospholipase and not due to artifacts.
In order to extend this label-free, ion channel-based approach of quantifying enzyme activity to a second relevant enzyme, we monitored the activity of phosphatidylinositol-specific PLC (PI-PLC) from Bacillus cereus (EC 184.108.40.206). In this case, we recorded the changes in γ in PC bilayers that contained initially 10 % PI (i.e. XPI,0 = 0.1) upon addition of the PI-PLC enzyme.
In these experiments, as PI-PLC catalyzed the hydrolysis of negatively-charged PI lipids and produced neutral diacylglycerol (DAG) lipids15 (Fig. 1b), the concentration of cations close to the membrane surface gradually decreased, leading to a measurable reduction of γ. Fig. 9a shows a calibration curve of γ as a function of XPI. Eq. 4, which resulted from the curve fit in Fig. 9a, describes this calibration curve:
Based on the measured values of γ we used Eq. 4, to estimate XPI as a function of time after addition of PI-PLC. Fig. 9b shows the decrease in XPI after addition of ~40 and ~80 pM of PI-PLC. The initial slopes of the two resulting exponential fits in Fig. 9a, differed by a factor of 2.1 and were hence close to the expected value of 2.0.
The limit of detection (LOD) of these assays (based on LOD = 3 × standard deviation) was equal to a mole fraction of PA, XPA, of 0.003 for the PLD assay, and a mole fraction of PI, XPI, of 0.078 for the PLC assay (see Supporting Information for details). We obtained these limits of detection with an ionic strength of ~21.5 mM; the sensitivity of the assay decreases with increasing ionic strength because the effect of electrostatic attraction of cations near a negatively-charged membrane surface will contribute relatively less to γ than the increased concentration of cations in the bulk solution.11,14 Nevertheless, we and others showed previously that negatively-charged membranes still increase γ significantly at ionic strengths close to the physiologic range (e.g. at 100 mM KCl).11,14
In summary, this work presents a novel, label-free assay to monitor, in real time, the enzymatic activity of phospholipases D and C on planar lipid bilayers. This ion channel-based assay provides unique access to both leaflets of a bilayer, proceeds on membranes with well-defined lipid composition, and exploits amplification characteristics of an ion pore11 (see Supporting Information). In addition, this assay detects enzyme activity within minutes at picomolar to nanomolar concentrations of membrane-active enzymes on native lipid substrates without requiring detergents or labels. Finally, this assay platform makes it possible to generate planar lipid bilayers with transverse lipid asymmetry. The application of the present ion channel-based assay may be somewhat limited by the poor mechanical stability of planar lipid bilayers. Recently reported strategies for fabrication of mechanically-stable bilayers23–25 such as embedding lipid bilayers in hydrogels26 may, however, address this issue in future work.
We purchased cesium acetate, resveratrol, and cyclosporin A from Sigma Aldrich; potassium chloride (KCl) from EMD Chemicals; cesium chloride (CsCl) from International Biotechnologies, Inc.; calcium chloride (CaCl2), pentane, and hexadecane from Fluka; and HEPES from Fisher Scientific. Gramicidin A (gA) was purchased as gramicidin D from Sigma Aldrich and purified by silica chromatography as described previously27 to afford a final purity of 97% of gA. We purchased the following phospholipids form Avanti Polar Lipids, Inc.: 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DiPhyPC), 1,2-diphytanoyl-sn-glycero-3-phosphate (sodium salt) (DiPhyPA), and 1,2-dioleoyl-sn-glycero-3-phosphoinositol (ammonium salt) (PI). Phospholipase D (PLD) from cabbage (EC 220.127.116.11) was obtained from Sigma Aldrich and phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus cereus (EC 18.104.22.168) was purchased from Invitrogen.
We received PLD as a lyophilized powder and immediately dissolved it in a buffer solution containing 10 mM CsCl, 0.5 mM CaCl2, and 10 mM cesium acetate with a pH of 5.5 (the same buffer was used for single channel recordings with this enzyme) to a final activity of 2,500 units mL−1. We aliquoted and stored this PLD solution at −80° C until usage. According to Sigma Aldrich, one unit of PLD liberates 1.0 μmol of choline from L-α-phosphatidylcholine (egg yolk) per hour at pH 5.6 at 30 °C. The specific activity of this enzyme, provided by Sigma Aldrich, was more than 1670 units per milligram of protein. Assuming a pure enzyme and considering a molecular weight of ~92,000 Da,28 a concentration of 1 unit mL−1 corresponds to a concentration of ~17.1 nM. We used this enzyme in a concentration range of ~15–40 nM in the bilayer chamber. For asymmetric bilayer experiments, we obtained inactivated PLD from a preparation that went through at least three freeze-thaw cycles, followed by extended storage at 4° C.
We received PI-PLC with an activity of 100 units mL−1 in a solution containing 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 0.01 % sodium azide, and 50% glycerol from Invitrogen, and stored it at −80° C. After thawing this enzyme preparation for the first time, we diluted it 10-fold in a buffer solution containing 20 mM KCl, 10 mM HEPES, pH 7.4 (this buffer was used for single channel recordings with this enzyme). We then aliquoted and stored the diluted enzyme solution at −80° C until usage. The unit definition for this enzyme, according to Invitrogen, is the amount of enzyme that converts 1 μmol of substrate to product per minute under the conditions of the assay. The specific activity of this enzyme, according to Invitrogen, was at least 1,000 units per milligram of protein. Assuming a pure enzyme and considering a molecular weight of ~35,000 Da, a concentration of 1 unit mL−1 corresponds to ~10.9 nM. We used this enzyme in a concentration range of ~43–86 pM in the bilayer setup. Therefore, with the standard bilayer setup used here, which had a volume of 3–4 mL, we were able to sense the activity of 172 femtomoles of PLC enzyme. It may be worth to point out that microfabricated bilayer setups24,26 have the capacity of measuring single ion channel conductance in volumes as small as 10 μL. Employing such miniaturized setups for the assays developed here, may make it possible to quantify the activity of 430 attomoles of PLC.
To minimize loss of enzyme activity due to storage of these enzymes in solution, we always thawed fresh aliquots of enzyme solutions shortly before addition to the bilayer setup.
We formed planar lipid bilayers29 with the “folding technique”11,30 in a custom-made bilayer setup fabricated in Teflon with two compartments of 3 or 4 mL capacity. We separated these two compartments by a thin Teflon film that contained one aperture with a diameter of ~100 μm (Eastern Scientific Inc.). We used vacuum grease (Corning) to attach the Teflon film to the chamber and to create a water tight seal between the two compartments. To facilitate bilayer formation, we pretreated the area surrounding the aperture of the Teflon film on each side with 2 μL of 5% (v/v) hexadecane in pentane. After addition of 1 mL of appropriate electrolyte (see below) to each compartment, we spread 3–5 μL of a lipid solution in pentane (25 mg mL−1 for DiPhyPC or 5 mg mL−1 for mixtures containing 90 mol% DiphyPC and 10 mol% phosphatidylinositol) at the air-water interface of the electrolyte solution and raised the liquid level by adding another 2 or 3 mL (depending on the compartment size) of electrolyte. Raising the liquid level in these two compartments above the aperture in the Teflon film resulted in the formation of a lipid bilayer from apposition of two lipid monolayers as described originally by Montal and Mueller31. If raising the liquid level did not immediately result in formation a bilayer over the aperture, we lowered the liquid level in one or both compartments, followed by raising the electrolyte solution again. We repeated this cycle until we obtained a bilayer with a minimum capacitance of 70 pF.
The electrolyte solution for the experiments with PLD contained 10 mM CsCl, 0.5 mM CaCl2, and 10 mM cesium acetate at a pH of 5.5 (adjusted with HCl). For experiments with PLC, the electrolyte contained 20 mM KCl and 10 mM HEPES at a pH of 7.4 (adjusted with KOH). We chose electrolytes with low ionic strength for these single channel recordings because the charge-induced local increase in cation concentration close to a membrane affects the gA conductance more significantly under conditions of low ionic strength compared to conditions of high ionic strength.11,14
Once a stable lipid bilayer was obtained, we gradually added small volumes (0.1 μL) of a solution of 10 ng mL−1 gA in isopropanol (Acros Organics) to both compartments of the bilayer setup until one to six gA channels were inserted in the bilayer at the same time. We chose isopropanol as solvent for gA since tertiary alcohols are not substrates for PLD as opposed to primary and secondary alcohols. After each addition of gA, we mixed the bilayer chambers by stirring the solutions in both compartments for at least 3 min (we confirmed previously that a small droplet of food dye was completely mixed in the setup within 2 min of stirring). In these experiments the final concentration of gA in the bilayer chamber was in the range of 0.1–2.0 pM. This range of gA concentration corresponds to 0.3–8 femtomoles of gA in the chamber with a volume of 3–4 mL; the total amount of lipid in each compartment was ~150 nanomoles, leading to a lipid to gA ratio of ~108. To measure the single channel conductance of gA pores,23,32 we recorded current traces versus time while applying different voltages in the range of ±100 mV. We performed these single channel recordings in “voltage clamp mode” using Ag/AgCl pellet electrodes (Warner Instruments) in both compartments of the bilayer setup. Data acquisition and storage was carried out using a custom software written in Labview by Daniel J. Estes in combination with a Geneclamp 500 amplifier from Axon Instruments (set to a gain of 100 mV pA−1 and a filter cutoff frequency of 2 kHz). The data acquisition board (National Instruments) that was connected to the amplifier was set to a sampling frequency of 15 kHz. All recordings were carried out at a temperature of ~22° C.
We performed the analysis of the single channel current traces by computing histograms of the currents from the original current versus time traces with Clampfit 9.2 software (Axon Instruments). From these histograms, we extracted the mean current amplitude of gA channel opening and closing events. All conductance values were obtained from the slopes of current amplitudes versus voltage (I–V) curves (see the next section for more details on this analysis).
In order to reduce the noise in these single channel recordings, we mounted the bilayer setup on a low noise stir plate (Stir-2, Warner Instruments), which was placed on a bench top vibration isolation platform (50 BM-4, Minus K Technology) inside a Faraday cage from Warner Instruments. In addition, we found that we reached the minimum noise level by placing the aforementioned setup on top of a second vibration isolation table (63–500 series from TMC) with a Faraday cage.
In order to ensure recordings from high quality bilayers without contaminations from previous experiments, we cleaned the bilayer chamber and setup in the following way: We removed the Teflon film that was previously mounted by vacuum grease to separate two compartments of the bilayer setup, and we cleaned the film by several washes with a stream of chloroform from a Teflon squirt bottle, followed by air drying before storage. Second, we rinsed the chamber with deionized water, followed by ethanol, and finally by chloroform. After two chloroform rinses, we used a Kimwipe tissue to clean the inside walls of the chamber such that all the vacuum grease was removed, and we, again, rinsed the chamber with chloroform. We stored the chamber in chloroform between experiments. Note, all steps with solvents (in particular chloroform) were carried out in a chemical fume hood.
This work was supported by NIH (M.M., 1R01GM081705), a Barbour Scholarship (S.M.), the UCSD Academic Senate, a Hellman faculty fellowship for J.Y., and the UCSD Center for AIDS Research (NIAID 5 P30 AI36214).