Understanding the free-energy landscape of phase separation in lipid bilayers using molecular dynamics.

Liquid-liquid phase separation inside the cell often results in biological condensates that can critically affect cell homeostasis. Such phase separation events occur in multiple parts of cells, including the cell membranes, where the "lipid raft" hypothesis posits the formation of ordered domains floating in a sea of disordered lipids. The resulting lipid domains often have functional roles. However, the thermodynamics of lipid phase separation and their resulting mechanistic effects on cell function and dysfunction are poorly understood. Understanding such complex phenomena in cell membranes, with their diverse lipid compositions, is exceptionally difficult. For these reasons, simple model systems that can recapitulate similar behavior are widely used to study this phenomenon. Despite these simplifications, the timescale and length scales of domain formation pose a challenge for molecular dynamics (MD) simulations. Thus, most MD studies focus on spontaneous lipid phase separation-essentially measuring the sign (but not the amplitude) of the free-energy change upon separation-rather than directly interrogating the thermodynamics. Here, we propose a proof-of-concept pipeline that can directly measure this free energy by combining coarse-grained MD with enhanced sampling protocols using a novel collective variable. This approach will be a useful tool to help connect the thermodynamics of phase separation with the mechanistic insights already available from MD simulations.

Ionizable Amino Lipids Distribution and Effects on DSPC/Cholesterol Membranes: Implications for Lipid Nanoparticle Structure.

Lipid nanoparticles (LNPs) containing ionizable aminolipids are among the leading platforms for the successful delivery of nucleic-acid-based therapeutics, including messenger RNA (mRNA). The two recently FDA-approved COVID-19 vaccines developed by Moderna and Pfizer/BioNTech belong to this category. Ionizable aminolipids, cholesterol, and DSPC lipids are among the key components of such formulations, crucially modulating physicochemical properties of these formulations and, consequently, the potency of these therapeutics. Despite the importance of these components, the distribution of these molecules in LNPs containing mRNA is not clear. In this study, we used all-atom molecular dynamics (MD) simulations to investigate the distribution and effects of the Lipid-5 (apparent pKa of the lipid nanoparticle = 6.56), a rationally designed and previously reported ionizable aminolipid by Moderna, on lipid bilayers [Mol. Ther. 2018, 26, 1509-1519]. The simulations were conducted with half of the aminolipids charged and half neutral approximately to the expected ionization in the microenvironment of the LNP surface. In all five simulated systems in this work, the cholesterol content was kept constant, whereas the DSPC and Lipid-5 concentrations were changed systematically. We found that at higher concentrations of the ionizable aminolipids, the neutral aminolipids form a disordered aggregate in the membrane interior that preferentially includes cholesterol. The rules underlying the lipid redistribution could be used to rationally choose lipids to optimize the LNP function.

Understanding the Free Energy Landscape of Phase Separation in Lipid Bilayers using Molecular Dynamics.

Liquid-liquid phase separation (LLPS) inside the cell often results in biological condensates that can critically impact cell homeostasis. Such phase separation events occur in multiple parts of cells, including the cell membranes, where the so-called "lipid raft" hypothesis posits the formation of ordered domains floating in a sea of disordered lipids. The resulting lipid domains often have functional roles. However, the thermodynamics of lipid phase separation and their resulting mechanistic effects on cell function and dysfunction are poorly understood. Understanding such complex phenomena in cell membranes, with their diverse lipid compositions, is exceptionally difficult. For this reasons, simple model systems that can recapitulate similar behavior are widely used to study this phenomenon. Despite these simplifications, the timescale and and length scales of domain formation pose a challenge for molecular dynamics (MD) simulations. Thus, most MD studies focus on spontaneous lipid phase separation - essentially measuring the sign (but not the amplitude) of the free energy change upon separation - rather than directly interrogating the thermodynamics. Here, we propose a proof-of-concept pipeline that can directly measure this free energy by combining coarse-grained MD with enhanced sampling protocols using a novel collective variable. This approach will be a useful tool to help connect the thermodynamics of phase separation with the mechanistic insights already available from molecular dynamics simulations. SIGNIFICANCE: Standard molecular dynamics simulations can determine the sign the free energy change upon phase separation, but not the amplitude. We present a new method to determine the phase separation free energy for lipid membranes, based on a enhanced sampling using the weighted ensemble method combined with a novel collective variable, validated using coarse-grained simulations applied to several simple systems. The new method will be valuable as a way to develop models that connect molecular-level structural features to the thermodynamics of phase separation.

Surveying nonvisual arrestins reveals allosteric interactions between functional sites.

Arrestins are important scaffolding proteins that are expressed in all vertebrate animals. They regulate cell-signaling events upon binding to active G-protein coupled receptors (GPCR) and trigger endocytosis of active GPCRs. While many of the functional sites on arrestins have been characterized, the question of how these sites interact is unanswered. We used anisotropic network modeling (ANM) together with our covariance compliment techniques to survey all the available structures of the nonvisual arrestins to map how structural changes and protein-binding affect their structural dynamics. We found that activation and clathrin binding have a marked effect on arrestin dynamics, and that these dynamics changes are localized to a small number of distant functional sites. These sites include α-helix 1, the lariat loop, nuclear localization domain, and the C-domain ß-sheets on the C-loop side. Our techniques suggest that clathrin binding and/or GPCR activation of arrestin perturb the dynamics of these sites independent of structural changes.

S-nitroso-L-cysteine stereoselectively blunts the adverse effects of morphine on breathing and arterial blood gas chemistry while promoting analgesia.

S-nitrosothiols exert multiple effects on neural processes in the central and peripheral nervous system. This study shows that intravenous infusion of S-nitroso-L-cysteine (SNO-L-CYS, 1 µmol/kg/min) in anesthetized male Sprague Dawley rats elicits (a) sustained increases in minute ventilation, via increases in frequency of breathing and tidal volume, (b) a decrease in Alveolar-arterial (A-a) gradient, thus improving alveolar gas-exchange, (c) concomitant changes in arterial blood-gas chemistry, such as an increase in pO2 and a decrease in pCO2, (d) a decrease in mean arterial blood pressure (MAP), and (e) an increase in tail-flick (TF) latency (antinociception). Infusion of S-nitroso-D-cysteine (SNO-D-CYS, 1 µmol/kg/min, IV), did not elicit similar responses, except for a sustained decrease in MAP equivalent to that elicited by SNO-L-CYS. A bolus injection of morphine (2 mg/kg, IV) in rats receiving an infusion of vehicle elicited (a) sustained decreases in frequency of breathing tidal volume, and therefore minute ventilation, (b) a sustained decrease in MAP, (c) sustained decreases in pH, pO2 and maximal sO2 with sustained increases in pCO2 and A-a gradient, and (d) a sustained increase in TF latency. In rats receiving SNO-L-CYS infusion, morphine elicited markedly smaller changes in minute ventilation, arterial blood gas chemistry, A-a gradient and MAP. In contrast, the antinociceptive effects of morphine were enhanced in rats receiving the infusion of SNO-L-CYS. The morphine-induced responses in rats receiving SNO-D-CYS infusion were similar to vehicle-infused rats. These data are the first to demonstrate that infusion of an S-nitrosothiol, such as SNO-L-CYS, can stereoselectively ameliorate the adverse effects of morphine on breathing and alveolar gas exchange while promoting antinociception.

S-Nitroso-L-Cysteine Stereoselectively Blunts the Deleterious Effects of Fentanyl on Breathing While Augmenting Antinociception in Freely-Moving Rats.

Endogenous and exogenously administered S-nitrosothiols modulate the activities of central and peripheral systems that control breathing. We have unpublished data showing that the deleterious effects of morphine on arterial blood-gas chemistry (i.e., pH, pCO2, pO2, and sO2) and Alveolar-arterial gradient (i.e., index of gas exchange) were markedly diminished in anesthetized Sprague Dawley rats that received a continuous intravenous infusion of the endogenous S-nitrosothiol, S-nitroso-L-cysteine. The present study extends these findings by showing that unanesthetized adult male Sprague Dawley rats receiving an intravenous infusion of S-nitroso-L-cysteine (100 or 200 nmol/kg/min) markedly diminished the ability of intravenous injections of the potent synthetic opioid, fentanyl (10, 25, and 50 µg/kg), to depress the frequency of breathing, tidal volume, and minute ventilation. Our study also found that the ability of intravenously injected fentanyl (10, 25, and 50 µg/kg) to disturb eupneic breathing, which was measured as a marked increase of the non-eupneic breathing index, was substantially reduced in unanesthetized rats receiving intravenous infusions of S-nitroso-L-cysteine (100 or 200 nmol/kg/min). In contrast, the deleterious effects of fentanyl (10, 25, and 50 µg/kg) on frequency of breathing, tidal volume, minute ventilation and non-eupneic breathing index were fully expressed in rats receiving continuous infusions (200 nmol/kg/min) of the parent amino acid, L-cysteine, or the D-isomer, namely, S-nitroso-D-cysteine. In addition, the antinociceptive actions of the above doses of fentanyl as monitored by the tail-flick latency assay, were enhanced by S-nitroso-L-cysteine, but not L-cysteine or S-nitroso-D-cysteine. Taken together, these findings add to existing knowledge that S-nitroso-L-cysteine stereoselectively modulates the detrimental effects of opioids on breathing, and opens the door for mechanistic studies designed to establish whether the pharmacological actions of S-nitroso-L-cysteine involve signaling processes that include 1) the activation of plasma membrane ion channels and receptors, 2) selective intracellular entry of S-nitroso-L-cysteine, and/or 3) S-nitrosylation events. Whether alterations in the bioavailability and bioactivity of endogenous S-nitroso-L-cysteine is a key factor in determining the potency/efficacy of fentanyl on breathing is an intriguing question.

A reversible mitochondrial complex I thiol switch mediates hypoxic avoidance behavior in C. elegans.

C. elegans react to metabolic distress caused by mismatches in oxygen and energy status via distinct behavioral responses. At the molecular level, these responses are coordinated by under-characterized, redox-sensitive processes, thought to initiate in mitochondria. Complex I of the electron transport chain is a major site of reactive oxygen species (ROS) production and is canonically associated with oxidative damage following hypoxic exposure. Here, we use a combination of optogenetics and CRISPR/Cas9-mediated genome editing to exert spatiotemporal control over ROS production. We demonstrate a photo-locomotory remodeling of avoidance behavior by local ROS production due to the reversible oxidation of a single thiol on the complex I subunit NDUF-2.1. Reversible thiol oxidation at this site is necessary and sufficient for the behavioral response to hypoxia, does not respond to ROS produced at more distal sites, and protects against lethal hypoxic exposure. Molecular modeling suggests that oxidation at this thiol residue alters the ability for NDUF-2.1 to coordinate electron transfer to coenzyme Q by destabilizing the Q-binding pocket, causing decreased complex I activity. Overall, site-specific ROS production regulates behavioral responses and these findings provide a mechanistic target to suppress the detrimental effects of hypoxia.

D-cysteine ethyl ester and D-cystine dimethyl ester reverse the deleterious effects of morphine on arterial blood-gas chemistry and Alveolar-arterial gradient in anesthetized rats.

We determined whether intravenous injections of the membrane-permeable ventilatory stimulants, D-cysteine ethyl ester (ethyl (2 S)- 2-amino-3-sulfanylpropanoate) (D-CYSee) and D-cystine dimethyl ester (methyl (2 S)- 2-amino-3-[[(2 S)- 2-amino-3-methoxy-3-oxopropyl]disulfanyl] propanoate) (D-CYSdime), could overcome the deleterious actions of intravenous morphine on arterial blood pH, pCO2, pO2 and sO2, and Alveolar-arterial (A-a) gradient (i.e., the measure of exchange of gases in the lungs) in Sprague Dawley rats anesthetized with isoflurane. Injection of morphine (2 mg/kg, IV) caused pronounced reductions in pH, pO2 and sO2 accompanied by elevations in pCO2, all which are suggestive of diminished ventilation, and elevations in A-a gradient, which suggests a mismatch of ventilation-perfusion. Subsequent boluses of D-cysteine ethyl ester (2 ×100 µmol/kg, IV) or D-cystine dimethyl ester (2 ×50 µmol/kg, IV) rapidly reversed of the negative actions of morphine on pH, pCO2, pO2 and sO2, and A-a gradient. Similar injections of D-cysteine (2 ×100 µmol/kg, IV) were without effect, whereas injections of D-cystine (2 ×50 µmol/kg, IV) produced a modest reversal. Our data show that D-cysteine ethyl ester and D-cystine dimethyl ester readily overcome the deleterious effects of morphine on arterial blood gas (ABG) chemistry and A-a gradient by mechanisms that may depend upon their ability to rapidly enter cells. As a result of their known ability to enter the brain, lungs, muscles of the chest wall, and most likely the major peripheral chemoreceptors (i.e., carotid bodies), the effects of the thiolesters on changes in ABG chemistry and A-a gradient elicited by morphine likely involve central and peripheral mechanisms. We are employing target prediction methods to identify an array of in vitro and in vivo methods to test potential functional proteins by which D-CYSee and D-CYSdime modulate the effects of morphine on breathing.

Effects of cholesterol on the mechanism of fengycin, a biofungicide.

Fengycins are a class of antifungal lipopeptides synthesized by the bacteria Bacillus subtilis, commercially available as the primary component of the agricultural fungicide Serenade. They are toxic to fungi but far less to mammalian cells. One key difference between mammalian and fungal cell membranes is the presence of cholesterol only in the former; recent experimental work showed that the presence of cholesterol reduces fengycin-induced membrane leakage. Since our previous all-atom and coarse-grained simulations suggested that aggregation of membrane-bound fengycin is central to its ability to disrupt membranes, we hypothesized that cholesterol might reduce fengycin aggregation. Here, we test this hypothesis using coarse-grained molecular dynamics simulations, with sampling enhanced via the weighted ensemble method. The results indicate that cholesterol subtly alters the size distribution for fengycin aggregates, limits the lateral range of their membrane disordering, and reduces the ability of aggregates to bend the membrane. Taken together, these phenomena may account for cholesterol's effects on fengycin activity.

Nitrosyl factors play a vital role in the ventilatory depressant effects of fentanyl in unanesthetized rats.

There is an urgent need to understand the intracellular mechanisms by which synthetic opioids, such as fentanyl, depress breathing. We used L-NAME (NG-nitro-L-arginine methyl ester), a nitric oxide synthase (NOS) inhibitor, to provide evidence for a role of nitric oxide (NO) and nitrosyl factors, including S-nitrosothiols, in fentanyl-induced suppression of breathing in rats. We measured breathing parameters using unrestrained plethysmography to record the changes produced by bolus administration of fentanyl (25 µg/kg, IV) in male Sprague Dawley rats that were pretreated with vehicle (saline), L-NAME (50 µmol/kg, IV) or the inactive D-isomer, D-NAME (50 µmol/kg, IV), 15 min previously. L-NAME produced a series of ventilatory changes that included (i) sustained elevations in breathing frequency, due to the reductions in the durations of inspiration and expiration, (ii) sustained elevations in minute ventilation, accompanied by minimal changes in tidal volume, and (iii) increases in inspiratory drive and expiratory drive, and peak inspiratory flow and peak expiratory flow. Subsequent administration of fentanyl in rats pretreated with vehicle produced negative effects on breathing, including decreases in frequency, tidal volume and therefore minute ventilation. Fentanyl elicited markedly different responses in rats that were pretreated with L-NAME, and conclusively, the negative effects of fentanyl were augmented by the NOS inhibitor. D-NAME did not alter ventilatory parameters or modulate the effects of fentanyl on breathing. Our study fully characterized the effects of L-NAME on ventilation in rats and is the first to suggest a potential role of nitrosyl factors in the ventilatory responses to fentanyl. Our data shows that nitrosyl factors reduce the expression of fentanyl-induced changes in ventilation.

D-Cystine di(m)ethyl ester reverses the deleterious effects of morphine on ventilation and arterial blood gas chemistry while promoting antinociception.

We have identified thiolesters that reverse the negative effects of opioids on breathing without compromising antinociception. Here we report the effects of D-cystine diethyl ester (D-cystine diEE) or D-cystine dimethyl ester (D-cystine diME) on morphine-induced changes in ventilation, arterial-blood gas chemistry, A-a gradient (index of gas-exchange in the lungs) and antinociception in freely moving rats. Injection of morphine (10 mg/kg, IV) elicited negative effects on breathing (e.g., depression of tidal volume, minute ventilation, peak inspiratory flow, and inspiratory drive). Subsequent injection of D-cystine diEE (500 µmol/kg, IV) elicited an immediate and sustained reversal of these effects of morphine. Injection of morphine (10 mg/kg, IV) also elicited pronounced decreases in arterial blood pH, pO2 and sO2 accompanied by pronounced increases in pCO2 (all indicative of a decrease in ventilatory drive) and A-a gradient (mismatch in ventilation-perfusion in the lungs). These effects of morphine were reversed in an immediate and sustained fashion by D-cystine diME (500 µmol/kg, IV). Finally, the duration of morphine (5 and 10 mg/kg, IV) antinociception was augmented by D-cystine diEE. D-cystine diEE and D-cystine diME may be clinically useful agents that can effectively reverse the negative effects of morphine on breathing and gas-exchange in the lungs while promoting antinociception. Our study suggests that the D-cystine thiolesters are able to differentially modulate the intracellular signaling cascades that mediate morphine-induced ventilatory depression as opposed to those that mediate morphine-induced antinociception and sedation.

How To Be a Good Member of a Scientific Software Community [Article v1.0].

Software is ubiquitous in modern science - almost any project, in almost any discipline, requires some code to work. However, many (or even most) scientists are not programmers, and must rely on programs written and maintained by others. A crucial but often neglected part of a scientist's training is learning how to use new tools, and how to exist as part of a community of users. This article will discuss key behaviors that can make the experience quicker, more efficient, and more pleasant for the user and developer alike.

Persistent Protein Motions in a Rugged Energy Landscape Revealed by Normal Mode Ensemble Analysis.

Proteins are allosteric machines that couple motions at distinct, often distant, sites to control biological function. Low-frequency structural vibrations are a mechanism of this long-distance connection and are often used computationally to predict correlations, but experimentally identifying the vibrations associated with specific motions has proved challenging. Spectroscopy is an ideal tool to explore these excitations, but measurements have been largely unable to identify important frequency bands. The result is at odds with some previous calculations and raises the question what methods could successfully characterize protein structural vibrations. Here we show the lack of spectral structure arises in part from the variations in protein structure as the protein samples the energy landscape. However, by averaging over the energy landscape as sampled using an aggregate 18.5 µs of all-atom molecular dynamics simulation of hen egg white lysozyme and normal-mode analyses, we find vibrations with large overlap with functional displacements are surprisingly concentrated in narrow frequency bands. These bands are not apparent in either the ensemble averaged vibrational density of states or isotropic absorption. However, in the case of the ensemble averaged anisotropic absorption, there is persistent spectral structure and overlap between this structure and the functional displacement frequency bands. We systematically lay out heuristics for calculating the spectra robustly, including the need for statistical sampling of the protein and inclusion of adequate water in the spectral calculation. The results show the congested spectrum of these complex molecules obscures important frequency bands associated with function and reveal a method to overcome this congestion by combining structurally sensitive spectroscopy with robust normal mode ensemble analysis.

Structures of SF3b1 reveal a dynamic Achilles heel of spliceosome assembly: Implications for cancer-associated abnormalities and drug discovery.

The pre-mRNA splicing factor SF3b1 exhibits recurrent mutations among hematologic malignancies and cancers, and consequently is a major therapeutic target of clinically-advanced spliceosome inhibitors. In this review, we highlight and rigorously analyze emerging views of SF3b1 conformational transitions, including the human SF3b particle either in isolation or bound to spliceosome inhibitors, and human or yeast spliceosome assemblies. Among spliceosome states characterized to date, an SF3b1 α-helical superhelix significantly closes to surround a U2 small nuclear RNA duplex with the pre-mRNA branch point sequence. The SF3b1 torus is locally unwound at an active site adenosine, whereas protein cofactors appear to stabilize overall closure in the spliceosome. Network analyses demonstrates that the natural SF3b1 dynamics mimic its conformational change in the spliceosome, raising the possibility of conformational selection underpinning spliceosome assembly. These dynamic SF3b1 conformations have consequences for gatekeeping of spliceosome assembly and therapeutic targeting of its cancer-associated dysfunction.

Best Practices for Quantification of Uncertainty and Sampling Quality in Molecular Simulations [Article v1.0].

The quantitative assessment of uncertainty and sampling quality is essential in molecular simulation. Many systems of interest are highly complex, often at the edge of current computational capabilities. Modelers must therefore analyze and communicate statistical uncertainties so that "consumers" of simulated data understand its significance and limitations. This article covers key analyses appropriate for trajectory data generated by conventional simulation methods such as molecular dynamics and (single Markov chain) Monte Carlo. It also provides guidance for analyzing some 'enhanced' sampling approaches. We do not discuss systematic errors arising, e.g., from inaccuracy in the chosen model or force field.

Chemically Accurate Relative Folding Stability of RNA Hairpins from Molecular Simulations.

To benchmark RNA force fields, we compared the folding stabilities of three 12-nucleotide hairpin stem loops estimated by simulation to stabilities determined by experiment. We used umbrella sampling and a reaction coordinate of end-to-end (5' to 3' hydroxyl oxygen) distance to estimate the free energy change of the transition from the native conformation to a fully extended conformation with no hydrogen bonds between non-neighboring bases. Each simulation was performed four times using the AMBER FF99+bsc0+χOL3 force field, and each window, spaced at 1 Å intervals, was sampled for 1 µs, for a total of 552 µs of simulation. We compared differences in the simulated free energy changes to analogous differences in free energies from optical melting experiments using thermodynamic cycles where the free energy change between stretched and random coil sequences is assumed to be sequence-independent. The differences between experimental and simulated ΔΔ G° are, on average, 0.98 ± 0.66 kcal/mol, which is chemically accurate and suggests that analogous simulations could be used predictively. We also report a novel method to identify where replica free energies diverge along a reaction coordinate, thus indicating where additional sampling would most improve convergence. We conclude by discussing methods to more economically perform these simulations.

Lipids Alter Rhodopsin Function via Ligand-like and Solvent-like Interactions.

Rhodopsin, a prototypical G protein-coupled receptor, is a membrane protein that can sense dim light. This highly effective photoreceptor is known to be sensitive to the composition of its lipidic environment, but the molecular mechanisms underlying this fine-tuned modulation of the receptor's function and structural stability are not fully understood. There are two competing hypotheses to explain how this occurs: 1) lipid modulation occurs via solvent-like interactions, where lipid composition controls membrane properties like hydrophobic thickness, which in turn modulate the protein's conformational equilibrium; or 2) protein-lipid interactions are ligand-like, with specific hot spots and long-lived binding events. By analyzing an ensemble of all-atom molecular dynamics simulations of five different states of rhodopsin, we show that a local ordering effect takes place in the membrane upon receptor activation. Likewise, docosahexaenoic acid acyl tails and phosphatidylethanolamine headgroups behave like weak ligands, preferentially binding to the receptor in inactive-like conformations and inducing subtle but significant structural changes.

Phosphorylation-mediated structural changes within the SOAR domain of stromal interaction molecule 1 enable specific activation of distinct Orai channels.

Low-conductance, highly calcium-selective channels formed by the Orai proteins exist as store-operated CRAC channels and store-independent, arachidonic acid-activated ARC channels. Both are activated by stromal interaction molecule 1 (STIM1), but CRAC channels are activated by STIM1 located in the endoplasmic reticulum membrane, whereas ARC channels are activated by the minor plasma membrane-associated pool of STIM1. Critically, maximally activated CRAC channel and ARC channel currents are completely additive within the same cell, and their selective activation results in their ability to each induce distinct cellular responses. We have previously shown that specific ARC channel activation requires a PKA-mediated phosphorylation of a single threonine residue (Thr389) within the cytoplasmic region of STIM1. Here, examination of the molecular basis of this phosphorylation-dependent activation revealed that phosphorylation of the Thr389 residue induces a significant structural change in the STIM1-Orai-activating region (SOAR) that interacts with the Orai proteins, and it is this change that determines the selective activation of the store-independent ARC channels versus the store-operated CRAC channels. In conclusion, our findings reveal the structural changes underlying the selective activation of STIM1-induced CRAC or ARC channels that determine the specific stimulation of these two functionally distinct Ca2+ entry pathways.

Selectivity and Mechanism of Fengycin, an Antimicrobial Lipopeptide, from Molecular Dynamics.

Fengycin is a cyclic lipopeptide used as an agricultural fungicide. It is synthesized by Bacillus subtilis as an immune response against fungal infection and functions by damaging the target's cell membrane. Previous molecular dynamics simulations and experiments have led to the hypothesis that the aggregation of fengycins on the membrane surface plays a key role in cell disruption. Here, we used microsecond-scale all-atom molecular dynamics simulations to understand the specificity, selectivity, and structure of fengycin oligomers. Our simulations suggest that fengycin is more likely to form stable oligomers in model fungal membranes (phosphatidylcholine) compared to the model bacterial membranes (phosphatidylethanolamine:phosphatidylglycerol). Furthermore, we characterize the differences in the structure and kinetics of the membrane-bound aggregates and discuss their functional implications.