A Liposome Membrane Permeability Assay for Investigating the Effects of Phosphatidylinositol Phosphate Groups on Membranotropic Action of Venom PLA2

Summary

We present a cost-effective method for assessing liposomal membrane permeability with superior sensitivity and resolution over conventional techniques. Using a novel encapsulated ligand substitution reaction, this technique detects phospholipase A₂-induced changes in model membranes. Easily integrated with standard analyses, this accessible approach enhances resolution for studying membrane biophysics and toxin interactions.

Abstract

We developed a novel, cost-effective, and highly sensitive assay for assessing liposomal membrane permeability based on ligand substitution reactions involving ligands encapsulated in the inner volume of liposomes, which we combined with a hydrolytic activity assay (using coenzyme A acylation) and molecular modeling simulation to investigate whether phosphates of the inositol ring in phosphatidylinositol protect membranes from the damaging effects of aberrantly expressed enzymes. Aberrant expression of phospholipases A₂ (PLA₂) leads to demyelination of axons that triggers the development of multiple sclerosis and Alzheimer's disease. To mimic the effects of aberrantly expressed PLA₂, we used the basic subunit of viper venom PLA₂, HDP-2P. The HDP-2P was tested on phosphatidylcholine (PC) liposomes enriched with phosphatidylinositol, phosphatidylinositol-4-phosphate (PI-4-P), or phosphatidylinositol-4,5-diphosphate (PI-4,5-P₂). While HDP-2P showed minimal hydrolytic activity and only slightly increased permeability in PC liposomes, enrichment of PC liposomes with PI enhanced both hydrolysis and permeability. Remarkably, PI-4-P and PI-4,5-P₂ dramatically inhibited HDP-2P's hydrolytic activity and completely reduced permeability. Molecular modeling simulations revealed that phosphates of the inositol ring in phosphatidylinositol sterically block phospholipid access to HDP-2P's active site, preventing hydrolysis. Both in vitro and in silico data suggest that phosphorylated phosphatidylinositols are associated with reduced PLA₂ activity, indicating a potential mechanism for mitigating excessive PLA₂ activity -- presumably as a means of protecting cell membranes from degeneration.

Introduction

Phosphatidylinositol (PI) is an anionic phospholipid consisting of a phosphatidic acid backbone linked to inositol, a hexahydroxy six-carbon ring, via a phosphate group. Inositol can be reversibly converted into phosphoinositides by phosphorylation at the 2, 3, 4, 5, and 6 carbon positions¹. PI and phosphoinositides are found in plasma membranes and organelles of various cells across animals, plants, and microorganisms. Seventy years after their discovery by Mable and Lowell Hokin^2,3,4, PI and phosphoinositides remain the subject of extensive research aimed at understanding their molecular mechanisms and their relevance to human health^5,6,7.

Apart from brain tissue, phosphoinositides are typically present at low concentrations, accounting for about 0.5 to 1% of the total lipids in the inner leaflet of the plasma membrane. The published data indicate that PI is unlikely to serve a major structural role under physiological conditions¹. However, during disease states, the levels of PI and phosphoinositides increase, and they flip-flop to the outer leaflet of the plasma membrane, indicating that these phospholipids may be linked to disease resistance¹. Although it has long been thought that PI plays a less significant physiological role compared to phosphoinositides, emerging evidence highlights the importance of PI-4-P and PI-4,5-P₂ in various physiological processes¹. These processes include intracellular signaling, intercellular signal transduction^8,9,10, membrane trafficking^11,12,13, modulation of gene expression¹, as well as cell proliferation¹⁴ and survival¹. Despite this, the molecular mechanisms by which PI and phosphoinositides drive these processes remain poorly understood.

PI and phosphoinositides are abundant in brain tissue cells, where they constitute approximately 10% of the total phospholipids, though their levels may transiently increase in the early stages of brain diseases suggesting that phosphoinositides may counteract the development of neurological pathologies¹ such as multiple sclerosis and Alzheimer's disease triggered by axonal demyelination caused by aberrant PLA₂s activity^15,16,17. These enzymes belong to the various classes and groups including group IIA secretory phospholipase A₂ (sPLA₂)¹⁵. Elevated levels of this group of enzymes are found in the cerebrospinal fluid of patients with multiple sclerosis and Alzheimer's, where they contribute to the breakdown of myelin membranes, leading to demyelination¹⁵. It has also been reported that sPLA₂s from snake venoms inhibit transmission at neuromuscular synapses^18,19, causing central nervous system disorders and adversely affecting the peripheral neuromuscular system²⁰.

In this study, we investigated whether PI and phosphoinositides can protect model phospholipid membranes from the effects of aberrantly expressed sPLA₂ enzymes. Model membranes were made of phosphatidylcholine (PC), the most abundant phospholipid in all mammalian cell membranes²¹, enriched with a 20% molar ratio of either PI, PI-4-P, or PI-4,5-P₂. To mimic the effects of aberrantly expressed sPLA₂ enzymes on model membranes, we used the basic subunit HDP-2P of the group IIA secretory viper venom neurotoxic PLA₂ enzyme²², whose hydrolytic activity has been previously studied on model myelin membranes²³.

For this work, we developed a novel approach to investigate changes in the permeability of liposomal membranes affected by HDP-2P. The novel method is based on a ligand substitution reaction in CuSO₄ solution outside the liposomes, involving an NH₃ ligand trapped inside the liposomes. If HDP-2P-induced membrane permeability allows NH₃ leakage from liposomes, the released ammonia will quantitatively substitute four water molecules in hydrated copper ions via the reaction: [Cu(H₂O)₆]²⁺ (aq) + 4NH₃ (aq) [Cu(H₂O)₂(NH₃)₄]²⁺ (aq) + 4H₂O (l). This ligand exchange produces a characteristic increase in solution optical density measurable by spectrophotometry (λ = 600 nm), serving as our quantitative readout for membrane permeability changes. Our ligand substitution-based approach demonstrates superior sensitivity for assessing membrane permeability compared to conventional techniques like ¹H-NMR^{23,24,25,26,27,28} and ion transport assays²⁹, while offering the additional advantage of not requiring expensive instrumentation. This simple and cost-effective method, applied in conjunction with HDP-2P hydrolytic activity measurements and the in silico simulation of PI, PI-4-P, and PI-4,5-P₂ binding to the molecular surface of HDP-2P, allowed us to conclude that the phosphates of the inositol ring in PI abolish the disruptive effects of HDP-2P on PC membranes enriched with PI-4-P or PI-4,5-P₂. Should this finding be confirmed in further studies, the ability of the inositol ring phosphates in PI to protect against the harmful effects of abnormally acting cellular proteins, potentially contributing to membrane stability in disease conditions involving aberrant protein activity, may pave the way for the development of novel pharmaceuticals for the treatment of neurological disorders such as Alzheimer's disease and multiple sclerosis.

Protocol

1. In silico simulation of HDP-2P interaction with PI and phosphoinositides using molecular modeling software

1. Software installation and preparation
   1. Install the referenced chemical drawing editor for constructing ligand molecules as shown in Supplemental File 1-Supplemental Figure S1A.
   2. Install the molecular modeling software for molecular docking of ligand to receptor.
   3. Install the Automated Docking Tools for ligand and receptor preparation.
   4. As the molecular modeling simulation software depends on the Python environment, install Python.
2. Ligand preparation
   1. First, use the chemical drawing editor to build the atomic coordinates for molecular structures of PI, PI-4-P, PI-4,5-P₂, cardiolipin, and phosphatidylserine.
      ​NOTE: The example protocol in the following steps of the ligand preparation describes the steps only for PI.
   2. Build the molecular structure of PI using the chemical drawing editor as illustrated in Supplemental File 1-Supplemental Figure S1B and save the file in PDB format as shown in Supplemental File 1-Supplemental Figure S2.
   3. Open the PDB file in the automated docking tool software as shown in Supplemental File 1-Supplemental Figure S3A.
   4. Add hydrogen atoms as shown in Supplemental File 1-Supplemental Figure S3B and select either polar or all hydrogen atoms as shown in Supplemental File 1-Supplemental Figure S4A.
      NOTE: After completion of PI docking with HDP-2P, select only polar hydrogen atoms to simplify visualization of polar or hydrogen bonds.
   5. Set the PI molecule as the ligand for the automated docking tool software as shown in Supplemental File 1-Supplemental Figure S4B.
   6. Save the ligand as a PDBQT file as shown in Supplemental File 1-Supplemental Figure S5A. Make sure to make all bonds rotatable.
3. Receptor preparation
   1. Download the PDB file of phospholipase A₂ HDP-2 (PDB code 2I0U), consisting of two monomers, one is acidic, enzymatically inactive, A, and the other is basic, enzymatically active, E. To delete enzymatically inactive monomer A, import the 2I0U file to the molecular modeling software (Supplemental File 1-Supplemental Figure S5B), select all atoms of monomer A (Supplemental File 1-Supplemental Figure S6A), then go to Edit and delete all atoms of monomer A (Supplemental File 1-Supplemental Figure S6B). Keep enzymatically active monomer E, which is now designated as the receptor.
   2. Open the HDP-2P receptor PDB file in the automated docking tool software as shown in Supplemental File 1-Supplemental Figure S7.
   3. Remove all water molecules as shown in Supplemental File 1-Supplemental Figure S8A.
   4. Add hydrogen atoms as shown in Supplemental File 1-Supplemental Figure S8B. Please note that after completion of the HDP-2P docking with PI, only polar hydrogens have to be selected to simplify visualization of polar and hydrogen bonds.
   5. Set the HDP-2P as the receptor, as shown in Supplemental File 1-Supplemental Figure S9A, and save it as a PDBQT file. Ensure the receptor is set without rotatable bonds.
4. Docking parameter setup
   1. Open both the ligand and the receptor PDBQT files.
   2. Enter the grid box setup interface, as shown in Supplemental File 1-Supplemental Figure S9B, and define the grid box size (size x = 44.644; size y = 74.939; size z = 43.050) and center (center x = 25.626; center y = 9.582; center z = 41.825) coordinates, which fully enclose the entire surface of the receptor molecule as shown in Supplemental File 1-Supplemental Figure S10A.
   3. Close the grid box setup interface with the current settings as shown in Supplemental File 1-Supplemental Figure S10B and save the docking parameters in a config.txt file as shown in Supplemental File 1-Supplemental Figure S11A.
5. Running the molecular docking software
   1. Launch the molecular docking software by using the config.txt file as shown in Supplemental File 1-Supplemental Figure S11B.
   2. Execute the docking calculation (~30-90 min; Supplemental File 1-Supplemental Figure S12).
   3. Check the docking_log.txt file to review the docking scores (binding affinity).
   4. Select the docking result with the lowest energy conformation.
6. Analysis of docking data
   1. Use the automated docking tool software to open the docking result PDBQT file.
   2. Analyze the ligand-receptor binding sites, focusing on the receptor's active center and intermolecular interactions, including ionic, ion-polar, polar, and hydrogen bonding interactions. Make a table for PI and phosphoinositide, which interacts with the receptor's active site, to describe binding site affinities in kcal/mol, charged and polar moieties in PI/phosphoinositide polar head, and HDP-2P residues involved in intermolecular bonding, and types of bonds as shown in Table 1 and Table 2. Repeat docking for each receptor-ligand pair 3x to make sure that there is no deviation in binding sites, binding affinities, charged and polar moieties of PI/phosphoinositide and HDP-2P residues, and the type of bonds.

2. Hydrolytic activity of HDP-2P in PC liposomes enriched with PI, PI-4-P, PI-4,5-P₂ , PS, or CL monitored by using the coenzyme A acylated with free fatty acids

NOTE: Chloroform, methanol, NH₃, and CuSO₄ are hazardous. Always use nitrile gloves, goggles, and lab coat. Handle all protocols withthese reagents under the fume hood. Store chloroform in an amber glass bottle. Store methanol in flammable cabinets away from oxidizers. Keep methanol away from flames and sparks. Label all waste containers for these reagents and dispose according to the local hazardous waste disposal regulations.

1. Prepare 100 mL of buffer solution for the HDP-2P-catalyzed phospholipid hydrolysis by dissolving 5.55 mg of CaCl₂ in 0.10 M Tris-HCl buffer (pH 7.4).
2. Prepare 50 mL of 1.0 M EDTA stock solution by dissolving 7.31 g of EDTA in deionized distilled water (ddH₂0).
3. Prepare 0.54 mM HDP-2P stock solution in 0.10 M Tris-HCl buffer (pH 7.4). Molar mass of HDP-2P is 13.827 kDa.
4. Prepare the following stock solutions of phospholipids at concentration 10.0 M in chloroform/methanol mixture (2 to 1 by volume): egg yolk L-α-phosphatidylcholine (PC), bovine heart cardiolipin (CL), bovine brain L-α-phosphatidyl-L-serine (PS), PI, PI-4-P, and PI-4,5-P₂. Molar masses of PC, CL, PS, PI, PI-4-P, and PI-4,5-P₂ are 768 Da, 1,466 Da, 792 Da, 867 Da, 962 Da, and 1,057 Da, respectively, as further described below.
5. Prepare five samples of PC liposomes. To prepare a sample of PC liposomes, add 75 µL of chloroform/methanol stock solution of PC to a glass test tube and dry the lipids at 25 °C under 10 Pa vacuum for 1 h until a lipid film is formed. Hydrate the lipid film for 1 h with 15 mL of buffer solution for phospholipid hydrolysis at 25 °C in a thermostat. Then, insert the titanium pistol of a 100-150 W ultrasonic dispenser into the lipid suspension in the glass test tube placed in an ice bath and sonicate the suspension at 22 kHz frequency for 10 min, not exceeding 50% of the maximum power. To remove titanium 'dust' particles, centrifuge liposomes at 200 × g (1,270 rpm with a 10 cm rotor) for 10 min at room temperature.
6. Prepare samples of PC liposomes enriched with either CL, PS, PI, PI-4-P, or PI-4,5-P₂ at the PC and other lipid molar ratio 4 to 1 by mixing chloroform/methanolphospholipid stock solutions-60 µL of PC with 15 µL of either CL, PS, PI, PI-4-P, or PI-4,5-P₂-in a glass tube. Then, dry the lipids in a 10 Pa vacuum or for 3 h at room temperature, hydrate the lipid film in buffer solution for phospholipid hydrolysis, and sonicate with ultrasonic waves as described above.
   NOTE: For each type of liposomes, prepare five samples.
7. Place 1.0 mL from five samples from each type of liposomes described above to separate tubes and add 10 µL of colorless coenzyme-A solution from the NEFA kit to each tube and then add 0.00, 0.93, 1.86, 2.79, or 3.72 µL of 0.54 mM HDP-2P stock solution to separate tubes to obtain in liposome samples the HDP-2P/phospholipid molar ratios of 0.000, 0.001, 0.002, 0.003, and 0.004, respectively.
8. Immediately after adding HDP-2P to the liposomes, incubate liposome samples for 15 min at 37 °C in a thermostat and then, stop phospholipid hydrolysis induced by HDP-2P by adding 100 µL of stock EDTA solution. After that, vortex the samples for 1 min.
9. Quantify the hydrolytic activity of HDP-2P by measuring the optical density of liposomal samples at 550 nm using a spectrophotometer, which would measure the extent that colorless coenzyme-A solution turns purple when it gets acylated with free fatty acids.
10. Repeat the above procedure two more times so that the average optical density values are derived from three independent trials.
    NOTE: Each data point obtained in this experimental trial is the average value of the optical density. Ideally, the standard deviation of experimental data points should lie within ±5% of the average.

3. Effects of HDP-2P on permeability of PC liposomal membranes enriched with PI, PI-4-P, PI-4,5-P₂ , PS, or CL measured by a ligand substitution reaction in CuSO​₄ solution involving a NH₃ ligand trapped inside the liposomes

NOTE: Chloroform, methanol, NH₃, and CuSO₄ are hazardous. Always use nitrile gloves, goggles, and lab coat. Handle all protocols with these reagents under the fume hood. Store chloroform in an amber glass bottle. Store methanol in flammable cabinets away from oxidizers. Keep methanol away from flames or sparks. Label all waste containers for these reagents and dispose according to the local hazardous waste disposal regulations.

1. Prepare 20 mL of 0.03 M stock solutions for each of six phospholipids - PC, CL, PS, PI, PI-4-P, PI-4,5-P₂- in chloroform/methanol (2 to 1 volume). Molecular masses of PC, CL, PS, PI, PI-4-P, PI-4,5-P₂ are 768, 1466, 792, 867, 962, and 1057 Da, respectively.
2. Prepare 100 mL of 0.10 M Tris-HCl (pH 7.4) buffer with at a final concentration of 0.05 M CuSO₄.
3. Prepare 4 mL of 1.5 × 10^-3 M HDP-2P stock in 0.05 M CuSO₄, 0.10 M Tris-HCl (pH 7.4) buffer. Molar mass of HDP-2P is 13.827 kDa.
4. To prepare 100 mL of 0.5 M NH₃ solution, use 25% NH₃ solution (density 905 g/L³⁰, equal to 13.30 M [905 g/L × 0.25 M ÷ 17.00 g/L = 13.30 M]). Thus, to prepare 100 mL of 0.5 M NH₃, use 3.76 mL of 25% NH₃ solution: (0.5 M × 1,000 mL ÷ 13.30 M) ÷ 10 = 3.76 mL of 25% NH₃ solution. Place 3.76 mL of 25% NH₃ solution in a 100 mL volumetric flask and bring the volume to 100 mL with 0.10 M Tris-HCl (pH 7.4) buffer.
5. To prepare liposomes with NH₃ in the inner volume of liposomes, place 12.5 mL of 0.03 M stock phospholipid solution in chloroform/methanol in a tube and dry phospholipids in 10 Pa vacuum or for 3 h at room temperature. Hydrate the phospholipid film for 1 h with 12.5 mL of 0.5 M NH₃, 0.10 M Tris-HCl (pH 7.4) at 25 °C in a thermostat; then sonicate the lipid suspension at 22 kHz frequency as described in step 2.5 to produce sonicated liposomes.
   NOTE: In the solution with sonicated liposomes, NH₃ is located inside and outside the liposomes.
6. To eliminate NH₃ outside the liposomes, dialyze the liposomes using a dialysis membrane (molecular cut-off 1.0 kDa) for 5 hat room temperature against 0.10 M Tris-HCl (pH 7.4) using a magnetic stirrer.
   NOTE: Dialysis for 5 h was sufficient to completely remove NH₃, as the addition of 0.5 mL of dialyzed liposomes to 0.5 mL of 0.10 M Tris-HCl, 0.05 M CuSO₄ buffer yielded the same optical density at 600 nm as that of liposomes prepared without NH₃ and mixed with the same buffer.
7. After dialysis, centrifuge the liposomes at 200 × g for 90 min as described in step 2.5 to pellet the liposomes. Discard the supernatant and resuspend the liposomes with 5.0 mL of 0.10 M Tris-HCl, 0.05 M CuSO₄ (pH 7.4) buffer, and divide the resuspended liposomes into five samples of 1.0 mL each. Add to these five 1.0 mL samples separately 80, 60, 40, 20, and 0 µL of 0.10 M Tris-HCl, 0.05 M CuSO₄ (pH 7.4) buffer and then 0, 20, 40, 60, 80 µL of HDP-2P stock buffer to adjust in the liposome samples the HDP-2P/phospholipid molar ratios 0.000, 0.001, 0.002, 0.003, and 0.004, respectively.
8. Incubate the liposome samples for 30 min in a thermostat at 37 °C; then, measure the optical density of the liposome samples at 600 nm using a spectrophotometer. Follow the steps described in section 3 for all six liposome samples and repeat 3x to ensure triplicate readings of optical density for each data point.
   NOTE: The standard deviation of experimental data points should always be within ±5% of the average.

Results

Molecular modeling simulation of HDP-2P Interaction with PI and phosphoinositides
The molecular modeling program calculates the best nine binding sites on the molecular surface of a receptor (larger molecule), which is targeted by a ligand (smaller molecule). These sites are determined by the release of the nine highest enthalpies, which designate the sites with the best binding affinities. The binding affinities are expressed as negative values, indicating that the intermolecular binding is an exothermic process. In the case of HDP-2P interacting with PI, the simulation predicted four binding sites (#3, #6, #7, and #8) for PI on the molecular surface of HDP-2P within the active center (Table 1). Additionally, the five binding sites for PI were found outside the active center of HDP-2P. The notations of the charged and polar moieties in the polar head of PI are illustrated in Figure 1.

The active center of HDP-2P consists of five key amino acid residues: Gly30, Gly32, His48, Asp49, and Lys69, which facilitate the proper alignment of phospholipid substrate in the active center to ensure catalytic hydrolysis. Some representative docked structures show the binding of PI to some amino acid residues of the active center, whereas in binding site #6, PI engages with all five key amino acid residues of the active center through ionic, ion-polar, and hydrogen bonds (Table 1). Figure 2 illustrates how PI interacts with all five key amino acid residues in the active center of HDP-2P within binding site #6.

The docking simulation of the interaction between HDP-2P and PI-4-P, which contains a phosphate group at position 4 of the inositol ring, predicted only binding site #1 as the location where PI-4-P binds to the active center of HDP-2P. Specifically, in this site, PI-4-P interacts with three key amino acid residues of the active center -- Gly30, Gly32, and Lys69 -- via ionic, ion-polar, and hydrogen bonds (Table 2). Figure 3 illustrates the interaction of PI-4-P with the active center of HDP-2P. In the remaining eight binding sites on the molecular surface of HDP-2P, PI-4-P binds to locations outside the active center.

The molecular docking simulations of HDP-2P interactions with PI-4,5-P₂ (containing phosphate groups at positions 4 and 5 of the inositol ring) revealed no binding to the enzyme's active center across all nine predicted binding sites. Instead, PI-4,5-P₂ consistently binds to a groove-like cavity adjacent to the active center (Figure 4), forming ion-polar and hydrogen bonds with the three HDP-2P residues, Trp31, Gly33, and Gln34).

While the binding of PI-4,5-P₂ to a groove-like cavity near the active center of cellular or venom PLA₂s has not been previously documented, structurally analogous features are well-established in pleckstrin homology³¹ (PH) and epsin N-terminal homology³² (ENTH) domains. These signaling domains utilize similar hydrophobic-polar-cationic groove architectures for specific PI-4,5-P₂ recognition^31,32, despite their non-enzymatic nature. The presence of a comparable structural motif in HDP-2P raises intriguing possibilities of either evolutionary mimicry or convergent evolution to facilitate membrane domain recognition by HDP-2P. If further verified by other methods, this study may represent the first reported instance of a venom PLA₂ utilizing such a mechanism for phosphoinositide recognition.

Interestingly, molecular docking simulations of CL and PS interacting with HDP-2P predicted the binding of both phospholipids to the active center of HDP-2P in eight out of nine binding sites, while the only binding site of CL and PS exterior to the active center was not located in the groove-like cavity (data not shown). It is also worth noting that, like PI, PC binds to the active center of HDP-2P in four out of nine binding sites, and both PI and PC, like CL and PS, do not bind to the groove-like cavity (data not shown). These findings suggest that the phosphate groups on the inositol ring of PI have a high binding affinity to the groove-like cavity located adjacent to the active center. Structural analyses have identified analogous groove-like cavities with hydrophobic-polar-cationic surface patches near the catalytic center in both cellular³³ and venom³⁴ PLA₂ enzymes. The binding of phosphoinositides to these cavities may induce physiologically significant conformational changes, potentially inhibiting enzymatic activity. Notably, among phosphoinositides, PI-4,5-P₂ demonstrates particularly poor substrate characteristics-human PLA₂ isoforms (cytosolic, calcium-independent, and secreted types) show minimal to undetectable hydrolytic activity toward PI-4,5-P₂³⁵. This suggests that phosphoinositide binding to these conserved structural features may serve a distinct physiological function (e.g., allosteric inhibition) rather than catalytic functions across the PLA₂ superfamily, which requires future experimental verification.

Hydrolytic activity of HDP-2P in PC liposomes enriched with PI, PI-4-P, PI-4,5-P₂, PS, or CL monitored using coenzyme A acylated with free fatty acids
The hydrolytic activity of HDP-2P on liposomal samples was assessed by using coenzyme A solution. The coenzyme A solution is colorless but turns purple upon acylation by free fatty acids released during phospholipid hydrolysis catalyzed by HDP-2P. The resulting increase in optical density, corresponding to the extent of coenzyme A acylation, was measured spectrophotometrically at 550 nm. Therefore, the increase in optical density in liposomal samples treated with coenzyme A and HDP-2P is directly proportional to the hydrolytic activity of HDP-2P.

Figure 5 presents the hydrolytic activity curves as a function of the HDP-2P/phospholipid molar ratio in various liposomal samples. The hydrolytic activity of HDP-2P is expressed in arbitrary units (a.u.) of optical density (OD) measured at 550 nm. We did not construct a calibration curve to relate a.u. values to the molar concentrations of free fatty acids, as our experimental trials showed that fatty acids in organic solvents, when added to aqueous solutions containing coenzyme A and acyl-coenzyme A synthetase, produced no detectable OD change at 550 nm. This indicates that such fatty acids form micelles rather than binding to coenzyme A. In contrast, only membrane-generated fatty acids (produced by PLA₂ activity) bind to membrane-associated coenzyme A, resulting in a measurable OD change. The amphipathic nature of coenzyme A -- having both hydrophobic and hydrophilic regions on its molecular surface -- preferentially localizes it at the membrane-water interface, which facilitates the binding of fatty acids produced by PLA₂ activity to coenzyme A. However, under these assay conditions, determining the molar concentration of fatty acids is not possible. For this reason, researchers using the coenzyme A assay express the hydrolytic activity of PLA₂ in arbitrary units of OD measured at 550 nm²⁰.

The highest increase in optical density, corresponding to the greatest hydrolytic activity of HDP-2P, was observed in PC liposomes enriched with CL or PS. This finding aligns well with the highest number of binding sites for CL and PS at the active center of HDP-2P. The hydrolytic activity of HDP-2P in PC liposomes enriched with PI was 3.0- and 2.5-fold lower than in liposomes enriched with CL or PS, respectively, but higher than in pure PC liposomes. This suggests that the binding affinity to the active center of HDP-2P of PI is higher than that of PC, despite both lipids having the same number of binding sites at the active center.

This apparent discrepancy may be explained by structural differences in the polar head groups of PI and PC. While PC is a neutral molecule, its cationic trimethylammonium group extends outward into the solution, whereas its anionic phosphate group is positioned closer to the hydrophobic area of the membrane. This charge distribution in the PC headgroup may create an electrostatic repulsion against the basic HDP-2P, reducing its binding affinity. In contrast, the acidic nature of PI makes it a more attractive lipid species for interaction with the basic HDP-2P, which could explain its higher hydrolytic activity compared to pure PC liposomes.

The hydrolytic activity of HDP-2P in PC liposomes enriched with PI-4-P, as shown in Figure 5, suggests minimal catalytic hydrolysis of phospholipids. Remarkably, the optical density curve obtained from PC liposomes enriched with PI-4,5-P indicates a complete abolition of hydrolytic activity by HDP-2P in this liposome sample. These findings align well with the docking simulations, which predict one binding site for PI-4-P at the active center of HDP-2P, while PI-4,5-P₂ shows no binding affinity to the active center at all. This suggests that the phosphate groups bound to the inositol ring hinder the phospholipid's binding to the active center of HDP-2P. We propose that the phosphate groups on the inositol ring extend outward into the solution, and by binding to HDP-2P via ionic and ion-polar intermolecular interactions, they keep HDP-2P at a distance from the membrane surface, thereby protecting the membrane from the hydrolytic activity of HDP-2P. This is consistent with recent findings revealing that human cytosolic, calcium-independent, and secreted PLA₂ enzymes showed barely detectable or no hydrolytic activity toward the PI-4-P and PI-4,5-P₂ substrates³⁵.

HDP-2P-induced permeability changes in PC liposomal membranes enriched with PI, PI-4-P, PI-4,5-P₂, PS, or CL: Assessment via ligand substitution reaction in CuSO₄ solution with encapsulated NH₃

The impact of HDP-2P hydrolytic activity on the membrane permeability of PC liposomes enriched with PI, PI-4-P, PI-4,5-P₂, PS, or CL was investigated using a ligand substitution reaction in a complex ion system, where the substituting ligand was encapsulated within the inner volume of the liposomes. This simple and reliable method is based on the displacement of four water molecules acting as ligands in the complex ion [Cu(H₂O)₆]²⁺ by NH₃

[Cu(H₂O)₆]²⁺ (aq) + 4NH₃ (aq) [Cu(H₂O)₂(NH₃)₄]²⁺ (aq) + 4H₂O (l)

During the process of NH₃ substituting water molecules in the complex ion [Cu(H₂O)₆]²⁺, the pale blue color of [Cu(H₂O)₆]²⁺ transitions to the dark blue color of [Cu(H₂O)₂(NH₃)₄]²⁺. In a system where NH₃ is encapsulated within the inner volume of liposomes and [Cu(H₂O)₆]²⁺ is present in the external solution, this color change -- monitored spectrophotometrically -- indicates an increase in liposomal membrane permeability to NH₃.

To establish such a system, we sonicated a phospholipid dispersion in Tris-HCl buffer containing NH₃, forming sonicated liposomes with NH₃ inside and outside the liposomes. External NH₃ was removed by dialysis, after which CuSO₄ solution was added to the liposome sample containing NH₃ in its inner volume. We then treated the liposomes with HDP-2P and monitored changes in optical density spectrophotometrically.

To determine the optimal wavelength for measuring optical density changes, we recorded the absorbance spectra of [Cu(H₂O)₂(NH₃)₄]²⁺ in a 0.025 M CuSO₄, 0.100 M NH₃ solution and [Cu(H₂O)₆]²⁺ in a 0.025 M CuSO₄ solution, using visible light in the range of 500 nm to 900 nm. As shown in Figure 6, irradiation at 600 nm yielded maximal absorbance for [Cu(H₂O)₂(NH₃)₄]²⁺ and minimal absorbance for [Cu(H₂O)₆]²⁺, leading us to select 600 nm as the optimal wavelength for our membrane permeability experiments.

Figure 7 shows that membrane permeability to NH₃ is highest in PC liposomes enriched with anionic CL and PS. This observation is consistent with our previously reported data showing that HDP-2P, a group IIA PLA₂, efficiently interacts with the anionic phospholipid palmitoyloleoyl-glycerophosphoglycerol, leading to substrate hydrolysis-an effect not observed with the neutral phospholipid palmitoyloleoylglycerophosphocholine³⁶. In general, most group IIA PLA₂ enzymes exhibit significantly higher hydrolytic activity toward anionic phospholipids^33,37,38. The permeability of PC+PI liposomes is lower than that of PC+CL and PC+PS liposomes but remains higher than that of pure PC liposomes. These findings align well with the results of docking simulations and HDP-2P hydrolytic activity assays. In other words, lyposomal samples that show low permeability are expected to be structurally stable due to the inability of HDP-2P to hydrolyze phospholipids and due to a lower affinity of HDP-2 to bind to PI and phosphoinositides, as predicted by the molecular docking data (Figure 2, Figure 3, and Figure 4). Therefore, enzyme hydrolytic activity by HDP-2P (Figure 5) and membrane permeability, as shown in Figure 7, should be positively correlated.

However, in PC+PI-4-P and PC+PI-4,5-P₂ liposomes, permeability did not increase upon treatment with HDP-2P. While the permeability results in PC+PI-4,5-P₂ liposomes correlate well with docking simulations and HDP-2P hydrolytic activity data for the same liposomes, the absence of increased permeability in PC+PI-4-P liposomes does not fully align with hydrolytic activity results, as a small degree of hydrolytic activity was recorded in these liposomes. These findings suggest that a slight localized increase in hydrolytic activity in PC+PI-4-P liposomes does not disturb bilayer packing of phospholipids to a degree that would change membrane permeability to NH₃.

Figure 1: Notations of charged and polar moieties of phosphatidylinositol's polar head. Please click here to view a larger version of this figure.

Figure 2: Interaction of PI with the active center of HDP-2P predicted for binding site #6 by molecular docking software. PI is given in stick representation and HDP-2P is given in line (diagram on the left) and molecular surface (diagram on the right) representations. For the stick representations, emerald represents carbon, orange is phosphorus, red is oxygen, and white is hydrogen. For the molecular surface and lines, green represents carbon, blue is nitrogen, red is oxygen, white is hydrogen, and yellow is sulfur. Yellow broken lines identify intermolecular bonds. Arrows point to amino acid residues in the active center. Abbreviations: PI = Phosphatidylinositol; HDP-2P = basic subunit of viper venom phospholipase. Please click here to view a larger version of this figure.

Figure 3: Interaction of PI-4-P with the active center of HDP-2P predicted for binding site #1 by molecular docking software. PI-4-P is given in stick representation and HDP-2P is given in lines (diagram on the left) and molecular surface (diagram on the right) representations. For the sticks, emerald represents carbon, orange is phosphorus, red is oxygen, and white is hydrogen. For molecular surface and lines, green represents carbon, blue is nitrogen, red is oxygen, white is hydrogen, and yellow is sulfur. Yellow broken lines identify intermolecular bonds. Arrows point to amino acid residues in the active center. Abbreviations: PI-4-P = Phosphatidylinositol-4-phosphate; HDP-2P = basic subunit of viper venom phospholipase. Please click here to view a larger version of this figure.

Figure 4: Interaction of PI-4,5-P₂ with the molecular surface of HDP-2P predicted for binding site #6 by molecular docking software. PI-4,5-P₂ binds to a groove-like cave located below the active center of HDP-2P. The five key amino acid residues in the active center are given in the three-letter notation. PI-4,5-P₂ is given in stick representation and HDP-2P is given in line (diagram on the left) and molecular surface (diagram on the right) representations. For the sticks, emerald represents carbon, orange is phosphorus, red is oxygen, and white is hydrogen. For molecular surface and lines, green represents carbon, blue is nitrogen, red is oxygen, white is hydrogen, and yellow is sulfur. Abbreviations: PI-4,5-P₂ = Phosphatidylinositol-4,5-diphosphate; HDP-2P = basic subunit of viper venom phospholipase. Please click here to view a larger version of this figure.

Figure 5: Hydrolytic activity curves from samples of liposomes at various HDP-2P/phospholipid molar ratios. Liposomes made of PC+CL (dark green), PC+PS (orange), PC+PI (dark blue), PC+PI-4-P (blue), PC+PI-4,5-P₂ (green), and PC (purple). The optical density values in arbitrary units represent the means of triplicate readings at 550 nm. Error bars indicate the SD of the triplicate readings. Statistical significance of the differences in optical density values among the liposome samples was assessed using ANOVA, with all p-values found to be well below 0.05. Abbreviations: PC = phosphatidylcholine; CL = bovine heart cardiolipin; PS = phosphatidylserine; PI = Phosphatidylinositol; PI-4-P = Phosphatidylinositol-4-phosphate; PI-4,5-P₂ = Phosphatidylinositol-4,5-diphosphate; a.u. = arbitrary units; HDP-2P = basic subunit of viper venom phospholipase. Please click here to view a larger version of this figure.

Figure 6: Absorbance spectra of aqueous solutions of copper complexes. Copper complexes [Cu(H₂O)₆]²⁺ (orange line) and [Cu(H₂O)₂(NH₃)₄]²⁺ (blue line), irradiated by visible light in the range from 500 nm to 900 nm. Copper complex [Cu(H₂O)₆]²⁺ was made in a solution of 0.025 M CuSO₄ and copper complex [Cu(H₂O)₂(NH₃)₄]²⁺ was made in a solution of 0.025 M CuSO4 containing 0.100 M NH₃. Absorbance values in this figure for each data point are the means of the absorbance readings done in triplicate. Error bars represent the SD from the triplicate readings. Please click here to view a larger version of this figure.

Figure 7: Membrane permeability curves from samples of liposomes. Liposomes made of PC+CL (purple), PC+PS (blue), PC+PI (dark blue), PC+PI-4-P (dark green), PC+PI-4,5-P₂ (orange), and PC (green) at various HDP-2P/ phospholipid molar ratios. Membrane permeability values are expressed as optical density values in arbitrary units (a.u.) measured at 600 nm. Each data point represents the mean of triplicate readings. Error bars indicate the SD of the triplicate readings. Statistical significance of the differences in optical density values among the liposome samples was assessed using ANOVA, with all p-values found to be well below 0.05. Please click here to view a larger version of this figure.

Table 1: Predicted affinity binding energies for candidate molecular complexes of HDP-2P with PI. Molecular modeling simulations were used to predict the types of bonds and binding-site affinities, as well as the charged and polar moieties of both the phosphatidylinositol headgroup and the HDP2P amino acid residues that participate in their intermolecular interactions. The charged and polar groups of phosphatidylinositol are highlighted in Figure 1. Note that Pb in NH_pb^δ+ denotes a peptide bond. Abbreviations: PI = Phosphatidylinositol; HDP-2P = basic subunit of viper venom phospholipase. Please click here to download this Table.

Table 2: Predicted affinity binding energies for candidate molecular complexes of HDP-2P with PI-4-P. Molecular modeling simulations were used to predict the types of bonds and binding-site affinities, as well as the charged and polar moieties of both the phosphatidyl-inositol-4-phosphate (PI-4-P) headgroup and the HDP2P amino acid residues that participate in their intermolecular interactions. Note that Pb in NH_pb^δ+ denotes a peptide bond. Please click here to download this Table.

Supplemental File 1: Molecular docking workflow. Overview of the molecular docking procedure using AutoDock, including ligand construction, receptor preparation, grid box setup, and execution of docking simulations to predict PI-HDP-2P interactions. Please click here to download this File.

Discussion

In this study, we investigated whether phosphate groups covalently bound to the inositol ring in phosphatidylinositol (PI) can prevent the degradation of phospholipid membranes by phospholipase A₂ (PLA₂) enzyme. To assess the progression of phospholipid membrane degradation caused by the hydrolytic activity of PLA₂, we designed a simple yet highly sensitive and accurate method to measure changes in membrane permeability, which should increase as structural degradation progresses.

The method is based on a ligand substitution reaction involving the hexaaquacopper(II) complex, [Cu(H₂O)₆]²⁺, suspended in the aqueous environment outside phospholipid liposomes, while ammonia (NH₃) is encapsulated inside the liposomes. Ammonia has a higher affinity for Cu²⁺ than water molecules, but NH₃ must permeate through the liposomal membrane to access the external aqueous phase, where it substitutes H₂O molecules in [Cu(H₂O)₆]²⁺. If the structural integrity of the liposomal membrane is compromised due to PLA₂-catalyzed phospholipid hydrolysis, increased membrane permeability to NH₃ will facilitate ligand substitution, leading to a color change from the pale blue of [Cu(H₂O)₆]²⁺ to the dark blue of [Cu(H₂O)₂(NH₃)₄]²⁺. We have monitored this transformation quantitatively by employing spectrophotometry to record the change in optical density at 600 nm.

We have previously used ¹H-NMR with paramagnetic ion ferrocyanide, [Fe(CN)₆]^3-, to study changes in membrane permeability affected by HDP-2P²⁴ and by cobra venom cardiotoxins^{24,25,26,27,28,39}. We have also used a range of fluorescent and phosphorescent probes to study the effects of cobra venom cardiotoxins on the permeability of liposomal membranes^20,23,40. However, ferrocyanide and fluorescent/phosphorescent probes are large non-physiological molecular entities that penetrate lipid membranes when membrane integrity is significantly damaged. In contrast, our novel ligand substitution assay utilizes NH₃ -- a small, natural cellular metabolite capable of permeating lipid membranes at subtle lipid-packing perturbations. This enhanced sensitivity of our novel membrane permeability method is evidenced by its ability to detect changes in membrane permeability at an HDP-2P/lipid molar ratio of 0.001, compared to 0.004 required for ¹H-NMR/ferrocyanide detection²³.

In the same study, we used erythrosine phosphorescence quenching by ferrocene to examine the effects of HDP-2P on intermembrane lipid exchange, assessed by the change in the ratio of erythrosine lifetime in test samples (τ_n) to that in the control sample (τ₀)²³. At an HDP-2P/lipid molar ratio of 0.004, the τ_n/τ₀ value decreased to 0.4²³, indicating a state of membrane fluidity permeable to divalent cations²⁴. Thus, our NH₃-based ligand substitution assay detects membrane permeability at an HDP-2P/lipid molar ratio of 0.001, whereas conventional assays (¹H-NMR/ferrocyanide and phosphorescence probe quenching) detect membrane permeability only at a molar ratio of 0.004. This demonstrates that our NH₃-based assay is four times more sensitive than conventional assays.

The exceptional sensitivity of our NH₃-based assay yielded permeability data that closely correlated with hydrolytic activity measurements, with only one apparent discrepancy: PC+PI-4-P liposomes showed insignificant hydrolysis without permeability changes, suggesting that minimal local lysophospholipid/fatty acid production does not affect the global lipid packing, which does not influence membrane structural integrity and permeability. Thus, this study demonstrated that the ligand substitution reaction in a complex ion system, involving a ligand encapsulated within the inner volume of liposomes, produced results that excellently aligned with the hydrolytic activity data of HDP-2P. These findings, in turn, were in strong agreement with in-silico studies on the binding of PI and phosphoinositides to the molecular surface of HDP-2P.

Overall, the AutoDock simulation in the present study revealed that PI binds to the active center at four out of nine identified binding sites, whereas PI-4-P binds to the active center at only one of the nine predicted sites. The hydrolytic activity of HDP-2P in PC+PI-4-P liposomes was approximately 5x lower than in PC+PI liposomes. Furthermore, in-silico studies predicted no binding of PI-4,5-P₂ to the active center of HDP-2P. Consistently, PLA₂ activity assays and ligand substitution reaction experiments showed that PI-4,5-P₂ completely inhibited both phospholipid hydrolysis and the increase in membrane permeability, while PI-4-P showed less, non-significant hydrolytic activity and entirely prevented membrane permeability changes. These observations are in good agreement with in vitro activity studies of human cytosolic, calcium-independent, and secreted PLA₂ enzymes toward PI, PI-4-P, and PI-4,5-P₂³⁵. These studies showed that only Group IVA cytosolic PLA₂ exhibited relatively high catalytic activity toward PI and PI-4-P, while its activity toward PI-4,5-P₂ was barely detectable. All other tested PLA₂ enzymes showed no activity toward PI-4,5-P₂³⁵. This lack of activity has been attributed to the highly hydrophilic phosphate groups on the inositol ring, which interfere with productive interactions between the membrane surface and PLA₂ enzymes³⁵. We suggest that the outward extension of phosphate groups on the inositol ring into the solution and their ability to form ion-polar and hydrogen bonds with residues of HDP-2P's groove-like cavity likely prevents HDP-2P from approaching the membrane surface closely enough to exert its hydrolytic activity, thereby stabilizing the membrane and reducing degradation of phospholipid membranes by HDP-2P which can be detrimental to cells by activating cell death pathways. This mechanism seems to align with the potential protective role of phosphoinositides in disease conditions involving aberrant enzyme activity.

The strong correlation between results obtained from membrane permeability assays, hydrolytic activity measurements, and in silico molecular docking studies highlights the reliability of the ligand substitution reaction method. This approach, utilizing a ligand encapsulated within liposomes, proves to be an accurate and effective tool for assessing membrane permeability changes caused by structural perturbations in the stable bilayer organization of phospholipids. While we acknowledge that molecular docking studies using AutoDock presented in this paper do not provide a thorough understanding of how phospholipids can interact with HDP-2P in a dynamic and longitudinal manner, molecular dynamics studies need to be performed in the future to identify additional molecular mechanisms of how HDP-2 can permeabilize phospholipid bilayers composed of various phospholipids. It is important to note that performing molecular dynamics studies requires high computational resources, such as high-performance computers and clusters, which are not available at many academic institutions that have limited resources and research funding. Finally, we acknowledge the use of a rigid receptor(HDP-2P) in the present techniques paper and the lack of a solvent/membrane environment. However, we want to emphasize that performing molecular dockings of HDP-2P with various phospholipids using a rigid receptor is the conventional approach to do initial docking studies to identify a first pass of docked candidate structures which can be followed up in the future with a flexible receptor once the critical amino acid residues within the catalytic side that bind to substrates are identified to focus on molecular regions that can be rendered flexible in follow-up molecular docking studies and in a solvated environment. Hence, leveraging collaborations with other institutions that have more resources can be a way to help corroborate some of the initial findings provided by AutoDock and identify additional molecular mechanisms of substrate-enzyme interactions.

In conclusion, the findings of this study demonstrate that the ligand substitution reaction method, utilizing a ligand encapsulated within liposomes, can be effectively applied to studies of membrane permeability in various membrane-mediated processes. These include ligand-receptor interactions, lipid-protein interactions, intermembrane exchange, membrane fusion and fission, and neuromuscular contractions, among others. Furthermore, this approach holds significant potential for investigating membranotropic pharmaceuticals. At the same time, we recognize that while novel methods in model membranes are invaluable for dissecting specific mechanisms of membrane interactions, their inherent simplifications limit their ability to fully replicate the structural, compositional, and functional complexity of biological membranes in vivo. It is therefore essential to interpret data from model systems with caution and to validate findings using cellular or organismal models wherever possible.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Basic subunit HDP-2P of V. nikolskii viper venom PLA2, HDP-2, 98.5% purity	Institute of Bioorganic Chemistry, Moscow		HDP-2 isolated according to potocol in PMID:18083205. HDP-2P purified from HDP-2 by DEAE Sephadex A-50
Cardiolipin (Heart, Bovine) (sodium salt),>99%	Macklin	C916138
Coenzyme A, FUJIFILM Wako’s NEFA kit	Lexington, MA, USA
Dialysis tubing cellulose membrane molecular weight cut-off 1.0 kDa	Sigma Aldrich Trading Co. Ltd., Shanghai, China
Digitally controlled water-tank thermostat	YoungGuangMing Medical Instruments Ltd. Co., Beijing, China
Electronic high precision analytical balance FA2004	Shanghai Maiyi Ltd. Co., China
JingHua 752 UV-Visible range spectrophotometer	Shanghai Jinghua Instruments Co., Ltd., Shanghai, China
L-α-phosphatidylcholine (Egg, Chicken),≥95%	Macklin	L799415
L-α-phosphatidylinositol (Liver, Bovine) (sodium salt),≥99%	Macklin	L780292
L-α-phosphatidylinositol-4,5-bisphosphate (Brain, Porcine) (ammonium salt),≥99%	Macklin	L996223
L-α-phosphatidylinositol-4-phosphate (Brain, Porcine) (ammonium salt), 99%	Macklin	L970250
L-α-phosphatidylserine (Brain, Porcine) (sodium salt),>99%	Macklin	L864755
Magnetic Stirrer JB-3	Shanghai Leici JB-3
pH Meter PHS-3C	Shanghai INESA Scientific Instruments Co., Ltd., China
TGL-16B microcentrifuge	Shanghai 3 Anting Co., Shanghai, China
Ultrasonic Dispenser Yt-JY96-IIN	Shanghai Yetuo technology Co., Ltd., China
Vacuum Pump TW-1M	Tingwei Co. Ltd., Wenling, China