Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Apr 1.
Published in final edited form as: Biochim Biophys Acta Mol Cell Biol Lipids. 2024 Feb 14;1869(3):159464. doi: 10.1016/j.bbalip.2024.159464

Role of Phosphatidic Acid Lipids on Plasma Membrane Association of the Ebola Virus Matrix Protein VP40

Michael D Cioffi 1, Monica L Husby 2, Bernard S Gerstman 1,3, Robert V Stahelin 2,4, Prem P Chapagain 1,3
PMCID: PMC11687635  NIHMSID: NIHMS1970258  PMID: 38360201

Abstract

The Ebola virus matrix protein VP40 is responsible for the formation of the viral matrix by localizing at the inner leaflet of the human plasma membrane (PM). Various lipid types, including PI(4,5)P2 (i.e. PIP2) and phosphatidylserine (PS), play active roles in this process. Specifically, the negatively charged headgroups of both PIP2 and PS interact with the basic residues of VP40 and stabilize it at the membrane surface, allowing for eventual egress. Phosphatidic acid (PA), resulting from the enzyme phospholipase D (PLD), is also known to play an active role in viral development. In this work, we performed a biophysical and computational analysis to investigate the effects of the presence of PA on the membrane localization and association of VP40. We used coarse-grained molecular dynamics simulations to quantify VP40 hexamer interactions with the inner leaflet of the PM. Analysis of the local distribution of lipids shows enhanced lipid clustering when PA is abundant in the membrane. We observed that PA lipids have a similar role to that of PS lipids in VP40 association due to the geometry and charge. Complementary experiments performed in cell culture demonstrate competition between VP40 and a canonical PA-binding protein for the PM. Also, inhibition of PA synthesis reduced the detectable budding of virus-like particles. These computational and experimental results provide new insights into the early stages of Ebola virus budding and the role that PA lipids have on the VP40-PM association.

1. Introduction

Filoviruses are a highly dangerous family of viruses that can cause severe hemorrhagic fever in humans leading to relatively high case fatality rates compared to other common viruses (1). The Ebola virus is a member of the Filovirus family and is mainly spread by human-to-human contact through various types of bodily fluids (i.e., blood, feces, sweat, vomit). The 2014–2016 outbreak of the Ebola virus in West Africa was the largest Ebola outbreak ever recorded (2). This led to the death of thousands with an average fatality rate of around 50% (3). It is essential that we understand the atomic and molecular level details of the virus’ life cycle so that we can continue developing more efficient vaccines and therapeutics.

The Ebola virus is a lipid-enveloped virus that will, during the budding process, obtain its lipid coating from the host cell plasma membrane (PM) (4). It does this by entering the host cell, unpacking its genetic content, replicating, and then localizing at the inner leaflet of the host cell’s PM. After membrane association, the virus will bud from the host cell and create new virions that will spread throughout the body and infect other cells (5). The Ebola virus matrix protein VP40 (VP40) is the major viral protein responsible for membrane association and the budding process (6). VP40 is the only Ebola virus protein necessary to form virus-like particles (VLPs) when expressed in mammalian cells (5, 7). Various lipids located within the human PM have been shown to contribute to this process, specifically, PI(4,5)P2 (PIP2) and phosphatidylserine (PS) lipids in the inner leaflet of the PM have been found to be major players in the binding and budding process of VP40 (8–10). The C-terminal domain (CTD) regions of VP40 have patches of amino acids containing cationic side chains that engage in electrostatic interactions with the anionic lipid headgroups which facilitate most of the initial VP40-PM interaction (11–14). Phosphatidic acid (PA) may also play an active role in VP40-PM interactions as it has been shown to be an important component of the PM inner leaflet (15, 16). When local concentrations of PA rise, it has been found to generate negative membrane curvature due to its small headgroup relative to other active phospholipids (16). The enzyme PLD has been found to synthesize PA in the inner leaflet of the PM by hydrolyzing phosphatidylcholine (PC) lipid molecules (16). This conversion of PC to PA lipids may be essential for efficient VP40 budding. The cellular role of PA, if any, in VP40 PM localization and virus-like particle (VLP) formation has not been explored.

To investigate the role of PA in VP40 PM localization and VLP formation, we employed a combination of computational and biophysical techniques. To investigate the role of PA on VP40-PM interactions at the molecular level, we performed coarse-grained molecular dynamics (CGMD) simulations with and without PA lipid molecules in the PM. The CGMD simulations were selected over all-atom simulations for this study due to the simulation lengths and computational cost considerations for the large size of the system (17). In the CG model, groups of heavy atoms (three to four on average) are mapped to a single interaction site (or a bead) (18). This enables us to investigate virus-membrane interactions and mechanisms at the tens of microseconds timescale. Complementary cellular experiments were performed to assess the impact of PA on VP40 PM localization and VLP formation. We find that PLD inhibition led to a significant reduction in VLP formation compared to vehicle control, highlighting the importance of PLD-derived PA on the VP40-dependent budding process. Our results, quantified through a variety of computational and biophysical tools, provide new insights into the influence PA lipids have in VP40-PM interactions.

2. Methods

2.1. Computational

2.1.1. VP40 hexamer model

The initial x-ray crystal structure of the VP40 dimer was obtained from the protein data bank (PDB) [PDB entry 7JZJ](19) which also supplies the structure of the CTD-CTD interface for the dimers to associate end-to-end. The missing residues were inserted using Modeller (20–22). For this, the left-CTD of the CTD-CTD complex was aligned/superimposed with the right-CTD of one dimer and the right-CTD of the CTD-CTD complex was aligned/superimposed with the left-CTD of a second dimer. This allowed us to create a dimer-dimer complex. This process was repeated to obtain a filament composed of three dimers, referred to as the hexamer in the rest of the paper.

2.1.2. Modeling of the VP40-membrane system

Two types of systems were examined: 1) protein-membrane systems with and without PA lipids and 2) membrane-only systems with and without PA lipids. For each system, an asymmetric lipid bilayer was used to represent the human PM composition which includes phosphatidylcholine (POPC, referred to as PC), phosphatidylinositol 4,5-bisphosphate (POP2, referred to as PIP2), phosphatidylethanolamine (POPE, referred to as PE), phosphatidylinositol (POPI, referred to as PI), sphingomyelin (POSM, referred to as SM), phosphatidylserine (POPS, referred to as PS), cholesterol (CHOL), and phosphatidic acid (POPA, referred to as PA) (23, 24).

As noted, the PLD enzyme can convert PC lipids to PA lipids. Therefore, to keep the systems with PA as similar as possible to the non-PA systems, we converted some of the PC lipids in the lower leaflet of the membrane to PA lipids and left the other lipid content unchanged. In total, five different systems were built and prepared using the Charmm-GUI (25) webserver with the MARTINI Bilayer Maker plugin (26): VP40-PM (no PA), VP40-PM (2% PA), VP40-PM (7% PA), PM only (no PA), and PM only (7% PA).

The composition of lipids in the upper (outer) leaflet for all five systems was in the ratio 40:4:8:4:20:4:20:0 (PC:PIP2:PE:PI:SM:PS:CHOL:PA). The ratio of lipids in the lower (inner) leaflet for the two systems without PA was (11:3:35:6:8:16:21:0). The inner leaflet of the three systems with PA contained either 2% or 7% of PA lipids, with a lipid ratio of 9:3:35:6:8:16:21:2 and 5:3:34:6:8:16:21:7, respectively. A higher concentration of PA (7%) was considered because the amount of PA lipids available to interact with the proteins was limited if only 2% PA is included in a small membrane system. There are 115 PA lipids in the 7% PA system, compared to only 32 in the 2% PA system, which may not be enough to represent the number of PA lipids available in the PM to interact with VP40. A significantly larger membrane system may allow enough PA lipids to interact with VP40 even at much lower concentrations. Unless otherwise specified as 2% PA, the analysis in the main text refers to the 7% PA system.

2.1.3. Coarse-grained Molecular Dynamics Simulations

We used the MARTINI 2.2 force field with the elastic network for amino acids combined with the MARTINI 2.0 force field for lipids and non-polarizable water (18, 27). Each system was solvated in water and neutralized with 0.15 M NaCl. The solvated VP40-PM systems contained approximately 142,000 to 158,000 CG beads. The PM-only systems each contained ~70,000 CG beads. All four CGMD simulations were completed using GROMACS (28, 29) and collected trajectories for a total of 20 μs with 20 fs timesteps at a temperature of 303.15 K.

2.2. Cellular Experiments

2.2.1. Confocal live cell imaging

HEK293 cells were used for confocal live cell imaging experiments to assess EGFP-VP40 localization (14, 30) in the absence or presence of the PA-binding reporter, RFP-PASS (31) (a kind gift from Dr. G. Dui, UT Health Houston). HEK293 cells were maintained as previously described (32) and transfected with EGFP-VP40, RFP-PASS, mRFP-LactC2 (from Dr. S. Grinstein, University of Toronto, Addgene #74061), myr-RFP (from Dr. A. Hadjantonakis, Addgene #32604) or mock (plasmid without protein expression insert). Cells were transfected with 0.4 μg EGFP-VP40 plasmid per well or 0.15 μg (low) or 0.4 μg (high) for RFP-PASS, myr-RFP, or mCherry-LactC2 protein expression.

For all imaging experiments, cells were stained for 10 minutes with WGA-Alexa647 (Thermo Fisher Scientific (Waltham, MA USA) for PM detection and Hoechst 33342 dye (Thermo Fisher Scientific) for nucleus visualization. Images were acquired on a Zeiss LSM880 microscope using a 40x water objective. Thresholding in ImageJ was used to determine the PM area. Percent PM localization of EGFP-VP40 or RFP-PASS was calculated by measuring the (488 nm or 561 nm excitation, respectively) fluorescence intensity at the PM compared to the intracellular fluorescence intensity.

2.2.2. Transmission Electron Microscopy

Transmission electron microscopy (TEM) was performed as previously described (33, 34) using HEK293 cells expressing Ebola virus VP40, GP and NP. HEK293 cells were treated with respective PLD inhibitors or vehicle (DMSO) 8 hours post-transfection. All treatments were performed at 1 μM and included either a PLD-1 inhibitor (VUO359595 (35)) from Avanti Polar Lipids, Inc. (Alabaster, AL USA), a PLD-2 inhibitor (VUO285655–1 (36)) from Avanti Polar Lipids, Inc. or a broad PLD isoform inhibitor (FIPI (37)) from Cayman Chemical Company (Ann Arbor, MI USA). Cells were then fixed at 24 hours post-transfection for TEM processing and imaging on a FEI Tecnai T12 electron microscope (Field Electron and Ion Company, Hillsboro, OR, USA).

3. Results

As shown recently, VP40 dimers assemble end-to-end at the PM to form matrix filaments (19). We modeled a VP40 hexamer by connecting three dimers end-to-end using the experimentally determined CTD-CTD interface. The resulting all-atom structure is shown in Fig. 1a. The hexamer model of VP40 is especially beneficial for this work as it allows us to examine the physical properties of VP40-PM interactions that may be missed with smaller systems. To simulate the lipid interactions with VP40, we set up a coarse-grained VP40-membrane system by placing the VP40 hexamer ~15Å below the inner leaflet of the PM (measured from amino acids to lipid tails), as shown in Fig. 1b. The membrane is composed of various lipids as described in the Methods section, and the lipids in CG representation are shown in Fig. 1c. In the Martini 2 CG model, PIP2 has six headgroup beads, PS and PC have two each, and PA has one. The headgroup beads for PIP2, PS, and PA accumulate net charges of −5, −1, and −2 respectively, while PC is neutral. At a physiological pH of 7.0, 50% of PA have −1 and 50% have −2 but the behavior shifts from having −1 to −2 charge when interacting with basic residues of the protein (38). Therefore, the CG model of PA with a single bead containing −2 charge is appropriate for our system.

Figure 1.

Figure 1.

Open in a new tab

a) All-atom model of the VP40 hexamer. b) The initial configuration of the coarse-grained (CG) model of the VP40 hexamer placed at the lower leaflet of the PM, composed of various lipids. c) Various lipids used in the simulations, represented as CG beads from the Martini model (18) (cholesterol not shown). Different bead types are colored differently and the total charges for the lipid headgroups are shown. For all lipid types, the phosphate group beads located just below the tail structures are named PO4 beads in Martini. The additional phosphate group beads at the lowest part of the PIP2 (furthest away from the tail) are named P1 and P2.

3.1. VP40-membrane interactions and the role of PA

3.1.1. VP40-lipid contacts

To investigate the role of PA on VP40-PM interactions at the molecular level, we performed 20 μs coarse-grained molecular dynamics (CGMD) simulations for the VP40 hexamer-membrane systems with and without PA lipids in the PM. At the beginning of the simulation, the hexamer is free to move as it makes very few interactions with the PM. As the protein starts to drift toward the PM, it quickly starts to interact with the lipid headgroups, within the first 1 μs of the simulation. After the interactions start to occur, the hexamer movement slows, and the lateral diffusion of certain lipid types decreases significantly (Movie S1). By the first 10 μs, the protein is fully associated with the PM. A representative snapshot of the VP40-membrane interactions is shown in Fig. 2a, with the PIP2 (red) and PA (orange) lipids highlighted. We examined the lipid interactions with VP40 by calculating the normalized number of contacts for PIP2, PA, PC, and PS lipids from the trajectories of the entire 20 μs CGMD simulations for all three systems (non-PA, 2% PA, 7% PA). The PM lipids were determined to be in contact with VP40 if their phosphate group (PO4) beads were within 7Å of any amino acid side chain (SC2) bead of VP40, as was done in refs.(39, 40), and this distance encompasses the interaction pairs in the first coordination shell in CGMD simulations(41–43). The results were then normalized by dividing the total number of contacts (for any given lipid type) by the total number of that lipid type in the lower leaflet of the PM. In Figure 2, contacts over time are displayed for PIP2 (Fig. 2b), PC and PA (Fig. 2d) and PS (Fig. 2e). The normalized contacts for all other interacting lipids (PI, PE, and SM) are displayed in Supplementary Fig. S1 and are found to have relatively similar normalized contacts for PA vs. non-PA systems.

Figure 2.

Figure 2.

Open in a new tab

(a) A snapshot of the VP40-membrane interactions (taken at the end of 20 μs of the PA system) showing PIP2 (in red) and PA (in orange) lipids in contact with the VP40 filament. (b) Number of normalized contacts over time in which a PIP2 phosphate group CG bead is within 7Å of any VP40 side chain (SC2) bead. (c) Histogram showing the frequency distribution of normalized contacts for the last 5 μs of the simulations in (b) for all three systems. (d, e) Number of normalized contacts over time in which a PC, PA, or PS phosphate group CG bead is within 7Å of any VP40 side chain (SC2) bead.

To determine the distribution of contacts for each lipid type in Fig. 2(b, d, e), we calculated their average number of contacts from the 15 to 20 μs window, after the VP40 dimers were sufficiently associated with the PM. We find that the average normalized number of contacts for PA lipids were 0.050 ± 0.047 (for 2% PA) and 0.035 ± 0.013 (for 7% PA), whereas the contacts for PC lipids are much smaller in PA and non-PA systems (0.009 ± 0.008 (for 2% PA), 0.009 ± 0.011 (for 7% PA), and 0.007 ± 0.006 (for non-PA)). Standard deviations are relatively large since the contacts with these lipids can form and break often. PC contacts do not occur on their own, therefore are highly dependent on filament binding orientation. As other interacting lipids cause the filament to bind, more PC lipids than the average may come in contact with the filament due to random motion. While contacts for PC lipids do not change much between the systems, there is a large increase in overall protein-membrane contacts through the added PA lipid interactions. These new interactions from PA lipids seem to enhance overall membrane binding. The normalized contacts for PS lipids in all three systems are also quite similar: 0.016 ± 0.007 (for no PA), 0.022 ± 0.008 (for 2% PA), and 0.019 ± 0.009 (for 7% PA). PS-VP40 interactions don’t seem to be affected by the addition of PA. The normalized contacts of PIP2 lipids are high in all systems. However, we see 22.41 ± 0.04 contacts for the non-PA system and an increase in contacts of 25.47 ± 0.05 (for 2% PA) and 29.72 ± 0.04 (for 7% PA) for the PA systems. The additional interactions from PA lipids and VP40 seem to enhance the number of PIP2 contacts (by ~13.60% (for 2% PA) and ~32.62% (for 7% PA), respectively) which is a crucial lipid for the successful binding of VP40 and PM association (30).

To determine which lipid types and amino acid residues contributed to the lipid-protein interactions, we counted the number of frames (amount of time) various lipid types were within 7Å of any VP40 side chain bead SC2 for both the 7% PA and non-PA systems. We then determined what percentage of frames that lipid type is in contact with the VP40 residue and displayed the results per amino acid. This allowed us to assess which VP40 residues were interacting with which lipids the most and the relative strength of those interactions. The percentage of contacting frames between VP40 and PIP2, PA, PC, or PS lipids are displayed in Fig. 3 for both 7% PA and non-PA systems.

Figure 3.

Figure 3.

Open in a new tab

Bar chart showing the percentage of frames any phosphate group (PO4) CG bead is within 7Å of any cationic VP40 side chain CG bead (SC2) for (a) PIP2, (b) PS, or (c) PA/PC. The percent contacts in the y-axis signify the percentage of frames in which the respective lipid establishes contact with that residue listed in the x-axis. For clarity, lipid contacts below 20% are not shown. Notably, PC contacts for the PA system are not shown in (c) because none of the residues in this system make more than 20% contact with PC lipids. (d) The top view of a VP40 dimer (membrane interacting interface) showing the locations of the VP40 residues in contact with PA, shaded according to the percent contacts from (c). The locations of the three VP40 residues with PA contacts exceeding 80% are labeled.

In both systems and for all four lipid types, cationic lysine residues with a +1 charge dominate the interactions (Fig. 3). Most of the VP40 residues that interact with the headgroup PO4 beads more than 20% of the time are the same (K221, K224, K225, K236, H269, K270, K274, K275, K279, and K326); the differences lie in the percentages. As shown in Fig. 3a, for the PA-system, there is a noticeable increase in PIP2 percent contacts with K221 (70% increase) and H269 (50% increase) and the addition of novel K279 interactions. PIP2 percent contacts decrease slightly (~25%) with K236 but completely lose contacts with K326. PS is also in contact with most of the same residues as PIP2 (Fig. 3b). For the PA system, the residues with notable increases for PS contacts compared to non-PA system include H269 (~25%) and K270 (~20%), whereas K274 shows a decrease in PS contacts (~15%). PS in the PA system interacts with two additional residues, K279 and K236, not observed in the non-PA system. If we compare PA and PS as a whole, we find the percent contacts (or residency time) to be much greater for PA (even though it is significantly less abundant in the PM). In the system with PA, PA interacts with eight VP40 residues more than 20% of the time, whereas PC does not interact with any VP40 residues more than 20% of the time (Fig. 3c). In contrast, PC in the non-PA system interacts with only four VP40 residues more than 20% of the time. These four PC interactions in the non-PA system are all significantly weaker (20% - 40%) than PA interactions (30% - 85%). Based on the percent contact calculations, it’s clear that PA (and PS) lipids have, on average, more residency time (number of interacting frames) with VP40 residues than PC lipids. The difference in the percentage of contacting frames between PC and PA is substantial and suggests that PA contributes much more to the initial VP40-PM association. However, the residency time of PIP2 is much higher than the rest. Since the percentage of PIP2 lipids in the lower leaflet of the PM is comparably lower than the other types, the shorter residency time for PS and PA seems to be due to their lesser negative charge and/or their geometry.

3.1.2. VP40-PIP2 interactions with different phosphate groups

Lipid charge is well known to play a key role in VP40 recruitment to the PM (44). Various types of phospholipids can interact with cationic protein complexes due to their negatively charged headgroups. As discussed earlier, the headgroup beads for PIP2, PS, and PA accumulate net charges of −5, −1, and −2 respectively, while PC is neutral. We see in Fig. 3c that the individual lipid-protein interactions between PA and VP40 are significantly stronger than those of PC. The actual interaction that is occurring between PC lipids and VP40 is quite weak (all four residues less than 40% residency time) compared to the interactions that are occurring between PA and VP40 (six out of seven residues greater than 50% with four higher than 75%). Since lipid geometry is quite similar between PA and PC, the increase in residency time seems to be associated with the stronger negatively charged headgroup (−2), compared to the neutral headgroup of PC. The same goes for PS; the stronger negative charge of PA seems to allow for stronger binding than PS. Noticeably, the residency time of any given interaction is closely correlated with the corresponding lipid’s headgroup charge. That is, PIP2 has the most residency time with VP40 residues out of the four, PA more than PS but lower than PIP2, and PC much lower than all.

As shown in Fig. 2, PIP2 has two additional phosphate groups compared to the other lipid types with only the PO4 group. They are located further down the headgroup (furthest away from the lipid tails) and their beads are named P1 and P2, each with charges of −2. To quantify the interactions with these two bead types, we calculated the percent contacts with P1 and P2 for both systems from the last 10 μs (10–20 μs) of the simulation and displayed the results in Fig. 4a,b.

Figure 4.

Figure 4.

Open in a new tab

Bar chart showing the percentage of frames PIP2’s P1 and P2 beads are interacting with VP40 residues for both PA and non-PA systems. a) Percentage contact with PIP2’s CG bead P1. b) Percentage contact with PIP2’s CG bead P2. Contacts below 20% are not shown for clarity. c) PIP2 (red) and PA (orange) lipids interacting with lysine residues at different VP40 depths at ~14 μs (top) and at ~16 μs (bottom). Lysine residues K104 (lower) and K275 (higher) are displayed in blue. Only PIP2 and PA lipids within 8Å of the VP40 hexamer are shown. The distance of 8Å was chosen here so that we can visualize both before and after the 7Å protein-lipid interaction occurs. The negatively charged headgroup beads (PO4) for PA and (PO4, P1, P2) for PIP2 interact with the positively charged lysine beads (SC2). After PA interacts with K275 and pulls the residue upwards, more PIP2 lipids are able to access the K104 residue located deeper.

As shown in Fig. 4a and 4b, most of the interacting residues are the same between the two systems. We find three novel contacts in the PA system (Y106, R151, and K212) and one in the non-PA system (K326). In addition to these novel contacts in the PA system, residues K127 and W191 also show large increases in percent contacts with the P1 beads (differences of ~25% and ~50% respectively) and K127 shows a large increase in the percent contacts with the P2 bead (difference of ~45%). In contrast, in the non-PA system, only one residue R134 has a notable increase in the percent contacts (~40%). Many of the VP40 residues (K104, Y106, F125, R134, R151, W191, K127, K212, and K221) that interact with PIP2 through P1 and P2 beads do not interact with PS or PA (which only have PO4 beads). Notably, these VP40 residues are all located further away from the inner leaflet of the PM (i.e., deeper within VP40) than the other residues. So, the interactions with these residues seem to be due to overall lipid geometry rather than lipid charge. PIP2 is longer than PS, PC, and PA (which can be seen in Fig. 1). Due to their relatively long headgroups, PIP2 is able to access deeper cationic VP40 residues whereas PS and PA cannot. Interactions at the surface with PS and PA seem to allow for PIP2 to access the deeper VP40 residues, enhancing membrane association (hence the increase in PIP2 interactions in the presence of PA). An example of the lipids interacting at different ‘depths’ can be seen in Fig. 4c. Here, PIP2 is able to access K104 as well as K275 whereas PA can only interact with K275. The PO4 beads in both PA and PIP2 interact with the SC2 beads of the surface-level lysine residues. However, the PIP2 (P1 and P2 beads) can interact with the deeper SC2 beads. The long length and strong negative charge of PIP2 allow for efficient interactions with many VP40 residues in various locations. PA and PS also interact with the same VP40 residues as PIP2 due to their negative charges, but these interactions are limited to the residues that are within reach (i.e., closer to the inner leaflet of the PM).

3.1.3. Radial distribution functions and PIP2 clustering

It was previously shown that lipid clustering directly affects membrane curvature (45). Specifically, PIP2 lipids have been shown to cluster(39) which plays an essential role in oligomerization as well as budding (9). Examination of our simulation trajectories also showed lipid clustering in both of our VP40-PM systems. We quantified the clustering behavior of various types of lipids in the PM, and the effects on this due to the presence of PA, by calculating radial distribution functions (RDFs or g(r)). The RDF can quantify the relative abundance of one molecule around another as it describes how density varies as a function of distance from a reference particle. To determine if the presence of PA in the PM had any effect on PIP2 clustering, we calculated and compared the RDF of various lipids relative to PIP2 (for both PA and non-PA systems) and relative to PA (for PA systems) from 10–20 μs (Figure 5).

Figure 5.

Figure 5.

Open in a new tab

Radial distribution functions for all lipid types in the lower leaflet of the PM a) relative to PIP2 in the non-PA system, b) relative to PIP2 in the 2% PA system, c) relative to PIP2 in the 7% PA system, d) relative to PA in the 2% PA system, and e) relative to PA in the 7% PA system. All systems include VP40 hexamer. The x-axis represents the distance between the PO4 beads of PIP2 and the PO4 beads of all other lipid types or ROH beads of cholesterol. The RDFs were calculated with the GROMACS built-in RDF function.

The g(r) for PIP2 in Fig. 5 shows that the clustering of all other lipid types relative to PIP2 is similar in both the PA vs. non-PA systems. The clustering of PIP2 around itself (PIP2 -PIP2 clustering) showed an increase in both the first and second solvation layers when 7% PA was present in the PM (Fig. 5c) compared to when it was not (Fig. 5a). The increase in PIP2 clustering suggests that PA interactions with the PM facilitate tighter PIP2 interactions with VP40 residues. This can likely be attributed to PA increasing the residency time between the PM and VP40 due to its relatively high net anionic charge compared to that of PC only. While the PA interactions don’t last for as long and aren’t as abundant as PIP2 interactions, they appear to provide more contact times between VP40 and the PM than just having PC in the PM, thereby enhancing the clustering of PIP2 around VP40. When PA is only at 2% abundance (only 32 lipids available), PIP2 clustering is not significantly affected. This is most likely due to the limited number of PA lipids available in our membrane relative to the large three-dimer filament in the system.

We also calculated the RDF for various lipid types relative to PA lipids (Fig. 5d-e). We found the PA lipids cluster with themselves but unlike the clustering of PIP2 in Fig. 5b, PA lipid clusters with itself only in the second solvation layer (around 0.90 nm). Seemingly, PA will fill gaps between interacting lipids, similar to that of cholesterol (39, 46). Also, even in the second solvation layer, the PA-PA clustering is weaker than the PIP2 -PIP2 clustering. However, increasing the abundance of PA from 2% to 7% slightly increases PA clustering, which can be attributed to more direct interactions with the PM. Figure 6 shows the clustering of PIP2 around VP40 in the presence and absence of PA at various times during the simulation. The enhanced clustering in the presence of PA appears to be due to the interactions with VP40 as there was no notable difference in PIP2 or PA clustering in the absence of VP40. PS-PS clustering is much weaker compared to PIP2 -PIP2 or PA-PA and the behavior is little changed in the presence of PA (Fig. S2).

Figure 6.

Figure 6.

Open in a new tab

PIP2 (red) clustering in the presence of the VP40 hexamer (lime). The system without PA is indicated as the two top rows and the system with 7% PA is indicated as the bottom two rows. Each top and bottom image for the respective system is the exact same frame visualized without VP40 (upper row) and with the VP40 hexamer overlayed (bottom row) to see the location of clustering. Images are shown for 0 μs, 10 μs, and 20 μs.

3.1.4. Lipid Diffusion

We investigated how lipid clustering in the presence of VP40 affects overall lipid movement. We quantified the lipid movement by calculating the lateral diffusion coefficients of all lipid types used in the simulations. The diffusion coefficient is a valuable quantity that can give us useful information on the PM’s ability to recruit proteins (47, 48). The diffusion was determined using GROMACS, which calculates the mean squared displacement of a specified lipid type and performs a least squares linear fit to give us the diffusion coefficient. We calculated the lateral diffusion coefficients from 10–20 μs for all four systems which can be seen in Table 1. As expected, the diffusion decreases for all lipid types in the VP40-PM system compared to the PM-only system.

Table 1.

Diffusion constants (×10−7 cm2/s) for all lipids in the lower leaflet of the PM in the presence and absence of the VP40 hexamer. Errors are given as the difference in diffusion coefficients obtained from fits over two halves of the time interval.

Diffusion Constants (×10−7 cm2/s)
Lipid Type PM with VP40 PM only
No PA PA No PA PA
PC 2.14 ± 0.02 2.06 ± 0.30 2.50 ± 0.03 2.88 ± 0.21
PIP2 1.17 ± 0.10 0.70 ± 0.10 2.34 ± 0.11 1.87 ± 0.29
PE 1.92 ± 0.09 2.13 ± 0.23 2.81 ± 0.04 2.88 ± 0.24
PI 1.98 ± 0.06 1.89 ± 0.09 2.60 ± 0.29 3.36 ± 0.34
SM 2.05 ± 0.24 2.19 ± 0.20 3.24 ± 0.10 3.30 ± 0.07
PS 2.15 ± 0.19 2.55 ± 0.11 3.21 ± 0.06 2.94 ± 0.24
CHOL 2.80 ± 0.08 3.24 ± 0.19 3.41 ± 0.19 3.94 ± 0.07
PA N/A 1.84 ± 0.23 N/A 2.96 ± 0.28
Open in a new tab

When comparing the two VP40-PM systems, we find that PIP2 diffusion significantly decreases when PA is present in the PM. As noted above, the clustering is also enhanced for PIP2 when PA is present in the membrane. The decrease in PIP2 diffusion can be attributed to this clustering effect which is due to the increased number of electrostatic interactions when PA is present in the membrane. The diffusion of PC (neutral) lipids for both systems is within the error range, suggesting that PA presence doesn’t have much of an effect on the diffusion of PC. PS lipids have a marginal increase in diffusion. This is likely due to PA lipids being of similar geometry and slightly stronger charge, therefore ‘taking the spot’ of some PS lipids that would otherwise be involved in the same electrostatic interactions. When comparing the two PM-only systems, we find that the presence of PA does not have any clear effect on the lateral diffusion of lipids in the absence of VP40 as all the values are either within or near the error range. This suggests that changes in diffusion are only noticeable when the VP40 is present and in the process of associating with the membrane.

3.2. Cellular Experiments

Previous lipid binding analysis of VP40 demonstrated that VP40 could associate with PA-containing vesicles albeit with a lower affinity than that of PS-containing vesicles (14). The cellular role of PA, if any, in VP40 PM localization and virus-like particle (VLP) formation has not been explored. To first assess any role for PA in VP40 PM localization, we employed a previously successful competition assay for VP40 PM localization (8, 30). In these experiments, GFP-VP40 was co-expressed for 24 hours with a PA-, PS-, or control lipid binding protein (RFP-PASS (PA), myr-RFP (PM anchored fluorescent protein), and mRFP-LactC2 (PS)) at low or high levels (Fig. 7). The PM marker WGA-647 was used to demarcate the PM to quantify PM localization of VP40 under different conditions. This work complements a more comprehensive co-expression analysis of VP40 and several lipid/membrane-binding proteins years ago (9, 30). In this earlier work, expression of PI(4,5)P2 and PS binding reporters at low or high levels greatly reduced VP40 PM localization in contrast to PI(3)P and PI(4)P binding reporters at low levels (Fig. 8). Notably, expression of the PA sensor RFP-PASS significantly reduced PM localization at low or high expression levels at either time point. The PM lipid-anchored RFP did not appreciably alter VP40 PM localization at low or high expression levels at either time point. As previously shown (8, 9), the PS sensor LactC2 reduced VP40 PM localization at the high expression level at either time point. This underscores the selective nature of reducing VP40 PM localization in a PA-dependent manner with comparable decreases in VP40 PM localization with co-expression of the PS- or PA-binding protein at high levels.

Figure 7.

Figure 7.

Open in a new tab

VP40 PM localization is decreased via competition with the PA-binding reporter, RFP-PASS. a) HEK293 cells expressing different fluorescently tagged proteins (GFP-eVP40, RFP-PASS, or mRFP-LactC2 at low (0.15 μg per well DNA) or high (0.4 μg per well DNA) levels were stained with WGA-Alexa647 and Hoechst 33342 dye at 24-hours post-transfection and imaged using confocal microscopy. Three independent replicates were performed for each condition.

Figure 8.

Figure 8.

Open in a new tab

Percent PM localization of VP40 was calculated by measuring the GFP-eVP40 fluorescence intensity at the PM compared to the intracellular fluorescence intensity. The plasma membrane marker WGA-Alexa647 was used to mark the plasma membrane to determine VP40 PM localization (N = 3 independent replicates performed). A two-way ANOVA was performed with multiple comparisons. *p<0.04, **p<0.006. Values are reported as mean ± standard deviation. Scale bars = 10 μm.

To assess if inhibition of PLD altered VP40-dependent VLP formation, we performed TEM to assess VLP formation as previously described (49, 50). Expression of the three Ebola virus proteins (VP40, GP, and NP) led to a marked increase in VLP formation from the PM in comparison with mock-transfected cells (Fig. 9). HEK293 cells producing VLPs under WT conditions were then compared to cells treated with vehicle (DMSO), a PLD-1 inhibitor (VUO359595) (51), a PLD-2 inhibitor (VUO285655–1) (36), and a broad PLD isoform inhibitor (FIPI) (37). Inhibition of PLD-1 or −2 led to a marked reduction in VLP formation compared to vehicle control (Fig. 9). This suggests that PLD-derived PA is an important part of the VP40-dependent budding process from the PM inner leaflet. However, PLD and PA can be involved in intracellular transport, so changes in VP40 transport to the PM cannot be ruled out as a contributing factor.

Figure 9.

Figure 9.

Open in a new tab

TEM reveals the decreased formation of VLPs post-inhibition of PLD. HEK293 cells expressing Ebola virus GP, NP, and VP40 were treated with control vehicle (DMSO) or 1 μM of respective PLD inhibitors (VUO359595 (PLD-1), VUO285655–1 (PLD-2), or FIPI (PLD-1&−2 inhibitor) at 8 hours post-transfections. Cells were then fixed for TEM processing at 24 hours post-transfection. The scale bar length is shown below each image.

4. Conclusion

We performed CGMD simulations to investigate the role lipid composition plays on the association of the Ebola Virus matrix protein VP40 with the human PM. We found that when PA lipids are in abundance within the PM (after PC conversion to PA), the ability for VP40 to associate at the inner leaflet of the PM is enhanced and facilitated largely by the increase in the number of electrostatic interactions. Cellular experiments demonstrated that a PA-binding protein competed with VP40 for PM localization akin to that of a PS-binding protein. We also find that PLD inhibition led to a marked reduction in VLP formation compared to vehicle control. These observations suggest that PLD-derived PA is an important part of the VP40-dependent budding process from the PM inner leaflet. Decreasing PA synthesis through PLD inhibition leads to a reduction in visible VLPs at the cell surface.

The charge difference between PC and PA, the relative length of PA, and the presence of other anionic lipid molecules (specifically PIP2 and PS) in the inner leaflet of the PM are responsible for successful VP40 association. PC conversion to PA seemingly has noticeable effects within the PM as PA acts similarly to PS by interacting with the same VP40 residues but for a notably longer time. PS has been shown to ‘pull’ VP40 and create a negative curvature in the membrane. While the curvature isn’t obvious in the simulations due to the coarse-grained elastic network, PA most likely has a similar pulling effect as PS since it is of a similar length but with a stronger negative charge. Both PA and PS pull on, or interact with, the nearest VP40 residues through electrostatic interactions to facilitate initial association, while the much longer PIP2 lipids are able to access deeper cationic residues of VP40 to facilitate a strong weave-like, anchored connection. Decreasing PC abundance while introducing PA lipids seems to increase overall amino acid and lipid residency time.

The effects of PA were also noticed in the RDFs. PIP2 clustered with itself more in the first and second solvation layers when PA was present in the membrane. This clustering effect is required for viral budding and has an influence on lipid movement which was seen in the decreased diffusion of PIP2. Further in-vitro experiments should be done to determine if these effects due to PA lipid presence occur on even larger timescales and if they do in fact create a more efficient budding process. Multi-microsecond timescale all-atom simulations may provide valuable insight into specific atomic-level details of PA and VP40 residue interactions. More detailed biophysical experiments are needed in vitro and in cell models to understand the mechanism by which PA alters the lipid diffusion of PIP2 and promotes VP40-dependent budding. Targeting specific interactions, like PC conversion to PA, which lead to favorable local lipid composition for VP40-PM association by increasing electrostatic interactions, may be a viable site for therapeutics.

Supplementary Material

1
2
Download video file (143.7MB, mov)

Highlights.

  • Phosphatidic acid (PA), resulting from the enzyme phospholipase D (PLD), plays an active role in viral development.

  • We used computational and biophysical analysis to investigate and demonstrate the importance of PA on the PM localization of Ebola virus matrix protein VP40 and virus-like particle (VLP) formation.

  • Our results show that the presence of PA enhances the lipid binding of VP40 and VLP formation.

Acknowledgments

These studies were funded by the NIH (AI142651 and OD027043) to R.V.S, (AI158220) to R.V.S. and P.C., and (GM075762) to M.L.H.

Footnotes

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Feldmann H, Geisbert TW. Ebola haemorrhagic fever. The Lancet. 2011;377(9768):849–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Control CfD, Prevention. 2014–2016 Ebola outbreak distribution in West Africa. Centers for Disease Control and Prevention (CDC), Atlanta, GE, USA; 2017. [Google Scholar]
  • 3.Kuhn J Filoviruses: a compendium of 40 years of epidemiological, clinical, and laboratory studies 2008. [PubMed] [Google Scholar]
  • 4.Madara JJ, Han Z, Ruthel G, Freedman BD, Harty RN. The multifunctional Ebola virus VP40 matrix protein is a promising therapeutic target. Future virology. 2015;10(5):537–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Jasenosky LD, Neumann G, Lukashevich I, Kawaoka Y. Ebola virus VP40-induced particle formation and association with the lipid bilayer. Journal of virology. 2001;75(11):5205–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Harty RN. No exit: targeting the budding process to inhibit filovirus replication. Antiviral research. 2009;81(3):189–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Noda T, Sagara H, Suzuki E, Takada A, Kida H, Kawaoka Y. Ebola virus VP40 drives the formation of virus-like filamentous particles along with GP. Journal of virology. 2002;76(10):4855–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Adu-Gyamfi E, Johnson KA, Fraser ME, Scott JL, Soni SP, Jones KR, Digman MA, Gratton E, Tessier CR, Stahelin RV. Host cell plasma membrane phosphatidylserine regulates the assembly and budding of Ebola virus. Journal of virology. 2015;89(18):9440–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Johnson KA, Taghon GJ, Scott JL, Stahelin RV. The Ebola Virus matrix protein, VP40, requires phosphatidylinositol 4, 5-bisphosphate (PI (4, 5) P2) for extensive oligomerization at the plasma membrane and viral egress. Scientific reports. 2016;6(1):1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Soni SP, Stahelin RV. The Ebola virus matrix protein VP40 selectively induces vesiculation from phosphatidylserine-enriched membranes. Journal of Biological Chemistry. 2014;289(48):33590–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Bornholdt ZA, Noda T, Abelson DM, Halfmann P, Wood MR, Kawaoka Y, Saphire EO. Structural rearrangement of ebola virus VP40 begets multiple functions in the virus life cycle. Cell. 2013;154(4):763–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Adu-Gyamfi E, Soni SP, Xue Y, Digman MA, Gratton E, Stahelin RV. The Ebola virus matrix protein penetrates into the plasma membrane. Journal of Biological Chemistry. 2013;288(8):5779–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Gc JB, Gerstman BS, Chapagain PP. Membrane association and localization dynamics of the Ebola virus matrix protein VP40. Biochim Biophys Acta Biomembr. 2017;1859(10):2012–20. doi: 10.1016/j.bbamem.2017.07.007. [DOI] [PubMed] [Google Scholar]
  • 14.Del Vecchio K, Frick CT, Gc JB, Oda S-I, Gerstman BS, Saphire EO, Chapagain PP, Stahelin RV. A cationic, C-terminal patch and structural rearrangements in Ebola virus matrix VP40 protein control its interactions with phosphatidylserine. The Journal of biological chemistry. 2018;293(9):3335–49. doi: 10.1074/jbc.m117.816280. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Kassas N, Tanguy E, Thahouly T, Fouillen L, Heintz D, Chasserot-Golaz S, Bader M-F, Grant NJ, Vitale N. Comparative characterization of phosphatidic acid sensors and their localization during frustrated phagocytosis. Journal of Biological Chemistry. 2017;292(10):4266–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Frohman MA. The phospholipase D superfamily as therapeutic targets. Trends in pharmacological sciences. 2015;36(3):137–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Marrink SJ, Tieleman DP. Perspective on the Martini model. Chemical Society Reviews. 2013;42(16):6801–22. [DOI] [PubMed] [Google Scholar]
  • 18.Marrink SJ, Risselada HJ, Yefimov S, Tieleman DP, De Vries AH. The MARTINI force field: coarse grained model for biomolecular simulations. The journal of physical chemistry B. 2007;111(27):7812–24. [DOI] [PubMed] [Google Scholar]
  • 19.Wan W, Clarke M, Norris MJ, Kolesnikova L, Koehler A, Bornholdt ZA, Becker S, Saphire EO, Briggs JA. Ebola and Marburg virus matrix layers are locally ordered assemblies of VP40 dimers. Elife. 2020;9:e59225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Fiser A, Do RKG, Šali A. Modeling of loops in protein structures. Protein science. 2000;9(9):1753–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Eswar N, Eramian D, Webb B, Shen M-Y, Sali A. Protein structure modeling with MODELLER. Structural proteomics: high-throughput methods. 2008:145–59. [DOI] [PubMed] [Google Scholar]
  • 22.Webb B, Sali A. Comparative protein structure modeling using MODELLER. Current protocols in bioinformatics. 2016;54(1):5.6.1–5.6.37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zachowski A Phospholipids in animal eukaryotic membranes: transverse asymmetry and movement. Biochemical Journal. 1993;294(Pt 1):1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Van Meer G, Voelker DR, Feigenson GW. Membrane lipids: where they are and how they behave. Nature reviews Molecular cell biology. 2008;9(2):112–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Jo S, Kim T, Iyer VG, Im W. CHARMM-GUI: a web-based graphical user interface for CHARMM. Journal of computational chemistry. 2008;29(11):1859–65. [DOI] [PubMed] [Google Scholar]
  • 26.Qi Y, Ingólfsson HI, Cheng X, Lee J, Marrink SJ, Im W. CHARMM-GUI martini maker for coarse-grained simulations with the martini force field. Journal of chemical theory and computation. 2015;11(9):4486–94. [DOI] [PubMed] [Google Scholar]
  • 27.Monticelli L, Kandasamy SK, Periole X, Larson RG, Tieleman DP, Marrink S-J. The MARTINI coarse-grained force field: extension to proteins. Journal of chemical theory and computation. 2008;4(5):819–34. [DOI] [PubMed] [Google Scholar]
  • 28.Van Der Spoel D, Lindahl E, Hess B, Groenhof G, Mark AE, Berendsen HJ. GROMACS: fast, flexible, and free. Journal of computational chemistry. 2005;26(16):1701–18. [DOI] [PubMed] [Google Scholar]
  • 29.Abraham MJ, Murtola T, Schulz R, Páll S, Smith JC, Hess B, Lindahl E. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015;1:19–25. [Google Scholar]
  • 30.Johnson KA, Taghon GJ, Scott JL, Stahelin RV. The Ebola Virus matrix protein, VP40, requires phosphatidylinositol 4, 5-bisphosphate (PI (4, 5) P2) for extensive oligomerization at the plasma membrane and viral egress. Scientific reports. 2016;6(1):19125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Zhang F, Wang Z, Lu M, Yonekubo Y, Liang X, Zhang Y, Wu P, Zhou Y, Grinstein S, Hancock JF. Temporal production of the signaling lipid phosphatidic acid by phospholipase D2 determines the output of extracellular signal-regulated kinase signaling in cancer cells. Molecular and cellular biology. 2014;34(1):84–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Nguyen CT, Dhakal D, Pham VTT, Nguyen HT, Sohng J-K. Recent advances in strategies for activation and discovery/characterization of cryptic biosynthetic gene clusters in Streptomyces. Microorganisms. 2020;8(4):616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wijesinghe KJ, McVeigh L, Husby ML, Bhattarai N, Ma J, Gerstman BS, Chapagain PP, Stahelin RV. Mutation of hydrophobic residues in the C-terminal domain of the Marburg virus matrix protein VP40 disrupts trafficking to the plasma membrane. Viruses. 2020;12(4):482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Amiar S, Husby ML, Wijesinghe KJ, Angel S, Bhattarai N, Gerstman BS, Chapagain PP, Li S, Stahelin RV. Lipid-specific oligomerization of the Marburg virus matrix protein VP40 is regulated by two distinct interfaces for virion assembly. Journal of Biological Chemistry. 2021;296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.O’Reilly MC, Scott SA, Brown HA, Lindsley CW. Further evaluation of novel structural modifications to scaffolds that engender PLD isoform selective inhibition. Bioorganic & medicinal chemistry letters. 2014;24(24):5553–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lavieri R, Scott SA, Lewis JA, Selvy PE, Armstrong MD, Brown HA, Lindsley CW. Design and synthesis of isoform-selective phospholipase D (PLD) inhibitors. Part II. Identification of the 1, 3, 8-triazaspiro [4, 5] decan-4-one privileged structure that engenders PLD2 selectivity. Bioorganic & medicinal chemistry letters. 2009;19(8):2240–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Su W, Yeku O, Olepu S, Genna A, Park J-S, Ren H, Du G, Gelb MH, Morris AJ, Frohman MA. 5-Fluoro-2-indolyl des-chlorohalopemide (FIPI), a phospholipase D pharmacological inhibitor that alters cell spreading and inhibits chemotaxis. Molecular pharmacology. 2009;75(3):437–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Shin JJ, Loewen CJ. Putting the pH into phosphatidic acid signaling. BMC Biol. 2011;9:85. doi: 10.1186/1741-7007-9-85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Gc JB, Gerstman BS, Stahelin RV, Chapagain PP. The Ebola virus protein VP40 hexamer enhances the clustering of PI(4,5)P(2) lipids in the plasma membrane. Phys Chem Chem Phys. 2016;18(41):28409–17. doi: 10.1039/c6cp03776c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.E C, Dai L, Tian J, Da LT, Yu J. Structure-Based Simulation and Sampling of Transcription Factor Protein Movements along DNA from Atomic-Scale Stepping to Coarse-Grained Diffusion. J Vis Exp. 2022(181). doi: 10.3791/63406. [DOI] [PubMed] [Google Scholar]
  • 41.Alessandri R, Souza PCT, Thallmair S, Melo MN, de Vries AH, Marrink SJ. Pitfalls of the Martini Model. J Chem Theory Comput. 2019;15(10):5448–60. doi: 10.1021/acs.jctc.9b00473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Atilgan AR, Durell SR, Jernigan RL, Demirel MC, Keskin O, Bahar I. Anisotropy of fluctuation dynamics of proteins with an elastic network model. Biophys J. 2001;80(1):505–15. doi: 10.1016/S0006-3495(01)76033-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Abdizadeh H, Jalalypour F, Atilgan AR, Atilgan C. A Coarse-Grained Methodology Identifies Intrinsic Mechanisms That Dissociate Interacting Protein Pairs. Front Mol Biosci. 2020;7:210. doi: 10.3389/fmolb.2020.00210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Stahelin RV. Membrane binding and bending in Ebola VP40 assembly and egress. Frontiers in microbiology. 2014;5:300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Koldsø H, Shorthouse D, Hélie J, Sansom MS. Lipid clustering correlates with membrane curvature as revealed by molecular simulations of complex lipid bilayers. PLoS computational biology. 2014;10(10):e1003911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Risselada HJ, Marrink SJ. The molecular face of lipid rafts in model membranes. Proceedings of the National Academy of Sciences. 2008;105(45):17367–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Goose JE, Sansom MS. Reduced lateral mobility of lipids and proteins in crowded membranes. PLoS computational biology. 2013;9(4):e1003033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Gerstman BS, Stahelin RV, Chapagain PP. The Ebola virus protein VP40 hexamer enhances the clustering of PI (4, 5) P 2 lipids in the plasma membrane. Physical Chemistry Chemical Physics. 2016;18(41):28409–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Amiar S, Stahelin RV. The Ebola virus matrix protein VP40 hijacks the host plasma membrane to form virus envelope. J Lipid Res. 2020;61(7):971. doi: 10.1194/jlr.ILR120000753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Wijesinghe KJ, McVeigh L, Husby ML, Bhattarai N, Ma J, Gerstman BS, Chapagain PP, Stahelin RV. Mutation of Hydrophobic Residues in the C-Terminal Domain of the Marburg Virus Matrix Protein VP40 Disrupts Trafficking to the Plasma Membrane. Viruses. 2020;12(4). doi: 10.3390/v12040482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.O’Reilly MC, Scott SA, Brown HA, Lindsley CW. Further evaluation of novel structural modifications to scaffolds that engender PLD isoform selective inhibition. Bioorg Med Chem Lett. 2014;24(24):5553–7. doi: 10.1016/j.bmcl.2014.11.017. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1970258-supplement-1.docx (1.2MB, docx)
2
Download video file (143.7MB, mov)

RESOURCES