Calculation of Lipid-Bilayer Permeabilities Using an Average Force

Cryptotanshinone-Induced Permeabilization of Model Phospholipid Membranes: A Biophysical Study

The Danshen terpenoid cryptotanshinone (CPT) is gaining enormous interest in light of its various outstanding biological activities. Among those, CPT has been shown to interact with cell membranes and, for instance, to have antibacterial activity. Several works have shown that CPT alone, or in combination with other drugs, can effectively act as an antibiotic against various infectious bacteria. Some authors have related the mechanism underlying this action to CPT-membrane interaction. This work shows that CPT readily partitions into phosphatidylcholine membranes, but there is a limiting capacity of accommodation of ca. 1 mol CPT to 3 mol phospholipid. The addition of CPT to unilamellar liposomes composed of 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) causes membrane permeabilization, as shown by fluorescent probe leakage. This process has been kinetically studied, as well as its modulation by incorporation of phosphatidylethanolamine or phosphatidylglycerol, as a model for pathogenic cell membranes. The thermotropic behavior of 1,2-dimyristoylphosphatidylcholine (DMPC) model membranes is weakly affected by CPT, but the terpenoid causes significant dehydration of the polar region of the bilayer and weak disordering of the acyl chain palisade, as observed in Fourier-transform infrared spectroscopy (FTIR) results. Small-angle X-ray scattering (SAXS) shows that CPT increases DMPC bilayer thickness, which could be due to localization near the phospholipid/water interface. Molecular dynamics (MD) simulations show that the lateral diffusion coefficient of the phospholipid increases with the presence of CPT. CPT extends from the polar head region to the center of the bilayer, being centered between the carbonyl groups and the unsaturated region of the POPC, where there is greater overlap. Interestingly, the free energy profiles of a water molecule crossing the lipid membrane show that the POPC membrane becomes more permeable in the presence of CPT. In summary, our results show that CPT perturbs the physicochemical properties of the phospholipid membrane and compromises its barrier function, which could be of relevance to explain part of its antimicrobial or anticancer activities.

How is Membrane Permeation of Small Ionizable Molecules Affected by Protonation Kinetics?

According to the pH-partition hypothesis, the aqueous solution adjacent to a membrane is a mixture of the ionization states of the permeating molecule at fixed Henderson-Hasselbalch concentrations, such that each state passes through the membrane in parallel with its own specific permeability. An alternative view, based on the assumption that the rate of switching ionization states is instantaneous, represents the permeation of ionizable molecules via an effective Boltzmann-weighted average potential (BWAP). Such an assumption is used in constant-pH molecular dynamics simulations. The inhomogeneous solubility-diffusion framework can be used to compute the pH-dependent membrane permeability for each of these two limiting treatments. With biased WTM-eABF molecular dynamics simulations, we computed the potential of mean force and diffusivity of each ionization state of two weakly basic small molecules: nicotine, an addictive drug, and varenicline, a therapeutic for treating nicotine addiction. At pH = 7, the BWAP effective permeability is greater than that determined by pH-partitioning by a factor of 2.5 for nicotine and 5 for varenicline. To assess the importance of ionization kinetics, we present a Smoluchowski master equation that includes explicitly the protonation and deprotonation processes coupled with the diffusive motion across the membrane. At pH = 7, the increase in permeability due to the explicit ionization kinetics is negligible for both nicotine and varenicline. This finding is reaffirmed by combined Brownian dynamics and Markov state model simulations for estimating the permeability of nicotine while allowing changes in its ionization state. We conclude that for these molecules the pH-partition hypothesis correctly captures the physics of the permeation process. The small free energy barriers for the permeation of nicotine and varenicline in their deprotonated neutral forms play a crucial role in establishing the validity of the pH-partitioning mechanism. Essentially, BWAP fails because ionization kinetics are too slow on the time scale of membrane crossing to affect the permeation of small ionizable molecules such as nicotine and varenicline. For the singly protonated state of nicotine, the computational results agree well with experimental measurements (P1 = 1.29 × 10-7 cm/s), but the agreement for neutral (P0 = 6.12 cm/s) and doubly protonated nicotine (P2 = 3.70 × 10-13 cm/s) is slightly worse, likely due to factors associated with the aqueous boundary layer (neutral form) or leaks through paracellular pathways (doubly protonated form).

Combining SILCS and Artificial Intelligence for High-Throughput Prediction of the Passive Permeability of Drug Molecules

Membrane permeability of drug molecules plays a significant role in the development of new therapeutic agents. Accordingly, methods to predict the passive permeability of drug candidates during a medicinal chemistry campaign offer the potential to accelerate the drug design process. In this work, we combine the physics-based site identification by ligand competitive saturation (SILCS) method and data-driven artificial intelligence (AI) to create a high-throughput predictive model for the passive permeability of druglike molecules. In this study, we present a comparative analysis of four regression models to predict membrane permeabilities of small druglike molecules; of the tested models, Random Forest was the most predictive yielding an R2 of 0.81 for the independent data set. The input feature vector used to train the developed prediction model includes absolute free energy profiles of ligands through a POPC-cholesterol bilayer based on ligand grid free energy (LGFE) profiles obtained from the SILCS approach. The use of the membrane free energy profiles from SILCS offers information on the physical forces contributing to ligand permeability, while the use of AI yields a more predictive model trained on experimental PAMPA permeability data for a collection of 229 molecules. This combination allows for rapid estimations of ligand permeability at a level of accuracy beyond currently available predictive models while offering insights into the contributions of the functional groups in the ligands to the permeability barrier, thereby offering quantitative information to facilitate rational ligand design.

Predictions from First-Principles of Membrane Permeability to Small Molecules: How Useful Are They in Practice?

Predicting from first-principles the rate of passive permeation of small molecules across the biological membrane represents a promising strategy for screening lead compounds upstream in the drug-discovery and development pipeline. One popular avenue for the estimation of permeation rates rests on computer simulations in conjunction with the inhomogeneous solubility-diffusion model, which requires the determination of the free-energy change and position-dependent diffusivity of the substrate along the translocation pathway through the lipid bilayer. In this Perspective, we will clarify the physical meaning of the membrane permeability inferred from such computer simulations, and how theoretical predictions actually relate to what is commonly measured experimentally. We will also examine why these calculations remain both technically challenging and overly computationally expensive, which has hitherto precluded their routine use in nonacademic settings. We finally synopsize possible research directions to meet these challenges, increase the predictive power of physics-based rates of passive permeation, and, by ricochet, improve their practical usefulness.

Molecular Dynamics Simulation Study of the Protonation State Dependence of Glutamic Acid Transport through a Cyclic Peptide Nanotube

The effect of the protonation state of glutamic acid on its translocation through cyclic peptide nanotubes (CPNs) was assessed by using molecular dynamics (MD) simulations. Anionic (GLU-), neutral zwitterionic (GLU0), and cationic (GLU+) forms of glutamic acid were selected as three different protonation states for an analysis of energetics and diffusivity for acid transport across a cyclic decapeptide nanotube. Based on the solubility-diffusion model, permeability coefficients for the three protonation states of the acid were calculated and compared with experimental results for CPN-mediated glutamate transport through CPNs. Potential of mean force (PMF) calculations reveal that, due to the cation-selective nature of the lumen of CPNs, GLU-, so-called glutamate, shows significantly high free energy barriers, while GLU+ displays deep energy wells and GLU0 has mild free energy barriers and wells inside the CPN. The considerable energy barriers for GLU- inside CPNs are mainly attributed to unfavorable interactions with DMPC bilayers and CPNs and are reduced by favorable interactions with channel water molecules through attractive electrostatic interactions and hydrogen bonding. Unlike the distinct PMF curves, position-dependent diffusion coefficient profiles exhibit comparable frictional behaviors regardless of the charge status of three protonation states due to similar confined environments imposed by the lumen of the CPN. The calculated permeability coefficients for the three protonation states clearly demonstrate that glutamic acid has a strong protonation state dependence for its transport through CPNs, as determined by the energetics rather than the diffusivity of the protonation state. In addition, the permeability coefficients also imply that GLU- is unlikely to pass through a CPN due to the high energy barriers inside the CPN, which is in disagreement with experimental measurements, where a considerable amount of glutamate permeating through the CPN was detected. To resolve the discrepancy between this work and the experimental observations, several possibilities are proposed, including a large concentration gradient of glutamate between the inside and outside of lipid vesicles and bilayers in the experiments, the glutamate activity difference between our MD simulations and experiments, an overestimation of energy barriers due to the artifacts imposed in MD simulations, and/or finally a transformation of the protonation state from GLU- to GLU0 to reduce the energy barriers. Overall, our study demonstrates that the protonation state of glutamic acid has a strong effect on the transport of the acid and suggests a possible protonation state change for glutamate permeating through CPNs.

A Coarse-Grained Molecular Dynamics Description of Docetaxel-Conjugate Release from PLGA Matrices

Despite the extensive use of poly-lactic-glycolic-acid (PLGA) in biomedical applications, computational research on the mesoscopic characterization of PLGA-based delivery systems is limited. In this study, a computational model for PLGA is proposed, developed, and validated for the reproducibility of transport properties that can influence drug release, the rate of which remains difficult to control. For computational efficiency, coarse-grained (CG) models of the molecular components under consideration were built using the MARTINI force field version 2.2. The translocation free energy barrier ΔG_t^* across the PLGA matrix in the aqueous phase of docetaxel and derivatives of varying sizes and solubilities was predicted via molecular dynamics (MD) simulations and compared with experimental release data. The thermodynamic quantity ΔG_t^* anticipates and can help explain the release kinetics of hydrophobic compounds from the PLGA matrix, albeit within the limit of a drug concentration below a critical aggregation concentration. The proposed computational framework would allow one to predict the pharmacological behavior of polymeric implants loaded with a variety of payloads under different conditions, limiting the experimental workload and associated costs.

Sampling efficiency of the counting method for permeability calculations estimated with the inhomogeneous solubility-diffusion model

Permeability is a key property in various fields such as membrane technology for chemical separation and transport of substances through cell membranes. At the molecular scale, the counting method uses the number of membrane crossings in a conventional unbiased molecular dynamics simulation to predict the permeability. This contribution investigates under which conditions the counting method has insufficient statistics. An equation is derived for a compartmental model based on the inhomogeneous solubility-diffusion (Smoluchowski) model, giving insight into how the flux correlates with the solubility of permeants. This equation shows that a membrane crossing is a rare event not only when the membrane forms a large free energy barrier but also when the membrane forms a deep free energy well that traps permeants. Such a permeant trap has a high permeability; yet, the counting method suffers from poor statistics. To illustrate this, coarse-grained MD was run for 16 systems of dipalmitoylphosphatidylcholine bilayer membranes with different permeant types. The composition rule for permeability is shown to also hold for fluxes, and it is highlighted that the considered thickness of the membrane causes uncertainty in the permeability calculation of highly permeable membranes. In conclusion, a high permeability in itself is not an effective indicator of the sampling efficiency of the counting method, and caution should be taken for permeants whose solubility varies greatly over the simulation box. A practical consequence relevant in, e.g., drug design is that a drug with high membrane permeability might get trapped by membranes thus reducing its efficacy.

Molecular dynamics simulations of ethanol permeation through single and double-lipid bilayers

Permeation of small molecules through membranes is a fundamental biological process, and molecular dynamics simulations have proven to be a promising tool for studying the permeability of membranes by providing a precise characterization of the free energy and diffusivity. In this study, permeation of ethanol through three different membranes of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine (POPS), PO-phosphatidylethanolamine (POPE), and PO-phosphatidylcholine (POPC) is studied. Permeabilities are calculated and compared with two different approaches based on Fick's first law and the inhomogeneous solubility-diffusion model. Microsecond simulation of double bilayers of these membranes provided a direct measurement of permeability by a flux-based counting method. These simulations show that a membrane of POPC has the highest permeability, followed by POPE and POPS. Due to the membrane-modulating properties of ethanol, the permeability increases as functions of concentration and saturation of the inner leaflet in a double bilayer setting, as opposed to the customary definition as a proportionality constant. This concentration dependence is confirmed by single bilayer simulations at different ethanol concentrations ranging from 1% to 18%, where permeability estimates are available from transition-based counting and the inhomogeneous solubility-diffusion model. We show that the free energy and diffusion profiles for ethanol lack accuracy at higher permeant concentrations due to non-Markovian kinetics caused by collective behavior. In contrast, the counting method provides unbiased estimates. Finally, the permeabilities obtained from single bilayer simulations are combined to represent natural gradients felt by a cellular membrane, which accurately models the non-equilibrium effects on ethanol permeability from single bilayer simulations in equilibrium.

Thermodynamics and Mechanism of the Membrane Permeation of Hv1 Channel Blockers

The voltage-gated proton channel Hv1 mediates efflux of protons from the cell. Hv1 integrally contributes to various physiological processes including pH homeostasis and the respiratory burst of phagocytes. Inhibition of Hv1 may provide therapeutic avenues for the treatment of inflammatory diseases, breast cancer, and ischemic brain damage. In this work, we investigate two prototypical Hv1 inhibitors, 2-guanidinobenzimidazole (2GBI), and 5-chloro-2-guanidinobenzimidazole (GBIC), from an experimentally screened class of guanidine derivatives. Both compounds block proton conduction by binding the same site located on the intracellular side of the channel. However, when added to the extracellular medium, the compounds strongly differ in their ability to inhibit proton conduction, suggesting substantial differences in membrane permeability. Here, we compute the potential of mean force for each compound to permeate through the membrane using atomistic molecular dynamics simulations with the adaptive biasing force method. Our results rationalize the putative distinction between these two blockers with respect to their abilities to permeate the cellular membrane.

Unassisted N-acetyl-phenylalanine-amide transport across membrane with varying lipid size and composition: kinetic measurements and atomistic molecular dynamics simulation

Biological membranes are essential to preserve structural integrity and regulate functional properties through the permeability of nutrients, pharmaceutical drugs, and neurotransmitters of a living cell. The movement of acetylated and amidated phenylalanine (NAFA) across 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) membrane bilayers is investigated to probe physical transport. The rate of transport is measured experimentally applying parallel artificial membrane permeation assay (PAMPA). At the physiological temperature, 310 K, the measured time constants in the neutral pH were ∼6 h in DOPC and ∼3 h in POPC, while in a more acidic condition, at a pH 4.8, the time constants were ∼8 h in both lipids. Computationally, we have expanded our transport study of three aromatic dipeptides across a bilayer composed of DOPC¹⁸. In this study, we have examined the effects of lipid composition and bilayer size on the passive transport of NAFA by simulating the dipeptide in three different bilayers, with 50 DOPC lipids, 50 POPC lipids, and 40 POPC molecules. Specifically, atomistic molecular dynamics simulations with umbrella sampling were used to calculate the potential of mean force for the passive permeation of NAFA across the bilayers. Diffusion constants were then calculated by numerically solving the Smoluchowski equation. Permeability coefficients and mean first passage times were then calculated. Structural properties - Ramachandran plots, sidechain torsions, peptide insertion angles, radial distribution functions, and proximal peptide water molecules - were also examined to determine the effect of system size and lipid type. In terms of systems size, we observed a small decrease in the highest barrier of the potential of mean force and fewer sampled sidechain dihedral angle conformations with 40 versus 50 POPC lipids due to weaker membrane deformations within a smaller lipid bilayer. In terms of lipid type, DOPC contains two monounsaturated acyl chains compared to only one such acyl chain in POPC; therefore, DOPC bilayers are less ordered and more easily deformed, as seen by a much broader potential of mean force profile. The NAFA in DOPC lipid also transitioned to an internally hydrogen-bonded backbone conformation at lower membrane depths than in POPC. Similarly, as for other aromatic dipeptides, NAFA tends to insert into the membrane sidechain-first, remains mostly desolvated in the membrane center, and exhibits slow reorientations within the bilayer in both DOPC and POPC. With a joint experimental and computational study we have gained a new insight into the rate of transport and the underlying microscopic mechanism in different lipid bilayer conditions of the simplest hydrophobic aromatic dipeptide.Communicated by Ramaswamy H. Sarma.

Membrane permeability of small molecules from unbiased molecular dynamics simulations

Permeation of many small molecules through lipid bilayers can be directly observed in molecular dynamics simulations on the nano- and microsecond timescale. While unbiased simulations provide an unobstructed view of the permeation process, their feasibility for computing permeability coefficients depends on various factors that differ for each permeant. The present work studies three small molecules for which unbiased simulations of permeation are feasible within less than a microsecond, one hydrophobic (oxygen), one hydrophilic (water), and one amphiphilic (ethanol). Permeabilities are computed using two approaches: counting methods and a maximum-likelihood estimation for the inhomogeneous solubility diffusion (ISD) model. Counting methods yield nearly model-free estimates of the permeability for all three permeants. While the ISD-based approach is reasonable for oxygen, it lacks precision for water due to insufficient sampling and results in misleading estimates for ethanol due to invalid model assumptions. It is also demonstrated that simulations using a Langevin thermostat with collision frequencies of 1/ps and 5/ps yield oxygen permeabilities and diffusion constants that are lower than those using Nosé-Hoover by statistically significant margins. In contrast, permeabilities from trajectories generated with Nosé-Hoover and the microcanonical ensemble do not show statistically significant differences. As molecular simulations become more affordable and accurate, calculation of permeability for an expanding range of molecules will be feasible using unbiased simulations. The present work summarizes theoretical underpinnings, identifies pitfalls, and develops best practices for such simulations.

Interaction of Camptothecin with Model Cellular Membranes

Accurate and efficient prediction of drug partitioning in model membranes is of significant interest to the pharmaceutical industry. Herein, we utilize advanced sampling methods, specifically, the adaptive biasing force methodology to calculate the potential of mean force for a model hydrophobic anticancer drug, camptothecin (CPT), across three model interfaces. We consider an octanol bilayer, a thick octanol/water interface, and a model 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC)/water interface. We characterize the enthalpic and entropic contributions of the drug to the potential of mean force. We show that the rotational entropy of the drug is inversely related to the probability of hydrogen bond formation of the drug with the POPC membrane. In addition, in long-time microsecond simulations of a high concentration of CPT above the POPC membrane, we show that strong drug-drug aromatic interactions shift the spatial orientation of the drug with the membrane. Stacks of hydrophobic drugs form, allowing penetration of the drug just under the POPC head groups. These results imply that inhomogeneous membrane models need to take into account the effect of drug aggregation on the membrane environment.

Modulation of membrane permeability by carbon dioxide

Promoting drug delivery across the biological membrane is a common strategy to improve bioavailability. Inspired by the observation that carbonated alcoholic beverages can increase the absorption rate of ethanol, we speculate that carbon dioxide (CO₂ ) molecules could also enhance membrane permeability to drugs. In the present work, we have investigated the effect of CO₂ on the permeability of a model membrane formed by 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine lipids to three drug-like molecules, namely, ethanol, 2',3'-dideoxyadenosine, and trimethoprim. The free-energy and fractional-diffusivity profiles underlying membrane translocation were obtained from μs-timescale simulations and combined in the framework of the fractional solubility-diffusion model. We find that addition of CO₂ in the lipid environment results in an increase of the membrane permeability to the three substrates. Further analysis of the permeation events reveals that CO₂ expands and loosens the membrane, which, in turn, facilitates permeation of the drug-like molecules. © 2019 Wiley Periodicals, Inc.

Dynamic Protonation Dramatically Affects the Membrane Permeability of Drug-like Molecules

Permeability (P_m) across biological membranes is of fundamental importance and a key factor in drug absorption, distribution, and development. Although the majority of drugs will be charged at some point during oral delivery, our understanding of membrane permeation by charged species is limited. The canonical model assumes that only neutral molecules partition into and passively permeate across membranes, but there is mounting evidence that these processes are also facile for certain charged species. However, it is unknown whether such ionizable permeants dynamically neutralize at the membrane surface or permeate in their charged form. To probe protonation-coupled permeation in atomic detail, we herein apply continuous constant-pH molecular dynamics along with free energy sampling to study the permeation of a weak base propranolol (PPL), and evaluate the impact of including dynamic protonation on P_m. The simulations reveal that PPL dynamically neutralizes at the lipid-tail interface, which dramatically influences the permeation free energy landscape and explains why the conventional model overestimates the assigned intrinsic permeability. We demonstrate how fixed-charge-state simulations can account for this effect, and propose a revised model that better describes pH-coupled partitioning and permeation. Our results demonstrate how dynamic changes in protonation state may play a critical role in the permeation of ionizable molecules, including pharmaceuticals and drug-like molecules, thus requiring a revision of the standard picture.

Multiscale Simulations of Biological Membranes: The Challenge To Understand Biological Phenomena in a Living Substance

Biological membranes are tricky to investigate. They are complex in terms of molecular composition and structure, functional over a wide range of time scales, and characterized by nonequilibrium conditions. Because of all of these features, simulations are a great technique to study biomembrane behavior. A significant part of the functional processes in biological membranes takes place at the molecular level; thus computer simulations are the method of choice to explore how their properties emerge from specific molecular features and how the interplay among the numerous molecules gives rise to function over spatial and time scales larger than the molecular ones. In this review, we focus on this broad theme. We discuss the current state-of-the-art of biomembrane simulations that, until now, have largely focused on a rather narrow picture of the complexity of the membranes. Given this, we also discuss the challenges that we should unravel in the foreseeable future. Numerous features such as the actin-cytoskeleton network, the glycocalyx network, and nonequilibrium transport under ATP-driven conditions have so far received very little attention; however, the potential of simulations to solve them would be exceptionally high. A major milestone for this research would be that one day we could say that computer simulations genuinely research biological membranes, not just lipid bilayers.

Prediction of Membrane Permeation of Drug Molecules by Combining an Implicit Membrane Model with Machine Learning

Lipid membrane permeation of drug molecules was investigated with Heterogeneous Dielectric Generalized Born (HDGB)-based models using solubility-diffusion theory and machine learning. Free energy profiles were obtained for neutral molecules by the standard HDGB and Dynamic HDGB (DHDGB) to account for the membrane deformation upon insertion of drugs. We also obtained hybrid free energy profiles where the neutralization of charged molecules was taken into account upon membrane insertion. The evaluation of the predictions was done against experimental permeability coefficients from Parallel Artificial Membrane Permeability Assays (PAMPA), and effects of partial charge sets, CGenFF, AM1-BCC, and OPLS, on the performance of the predictions were discussed. (D)HDGB-based models improved the predictions over the two-state implicit membrane models, and partial charge sets seemed to have a strong impact on the predictions. Machine learning increased the accuracy of the predictions, although it could not outperform the physics-based approach in terms of correlations.

Characterizing the Molecular Mechanisms for Flipping Charged Peptide Flanking Loops across a Lipid Bilayer

The cell membrane largely prevents the passive diffusion of charged molecules due to the large free energy barrier associated with translocating charged groups across the hydrophobic lipid bilayer core. Despite this barrier, some peptides can interconvert between transmembrane and surface-adsorbed states by "flipping" charged flanking loops across the bilayer on a surprisingly rapid second-minute time scale. The transmembrane helices of some multispanning membrane proteins undergo similar reorientation processes, suggesting that loop-flipping may be a mechanism for regulating membrane protein topology; however, the molecular mechanisms underlying this behavior remain unknown. In this work, we study the loop-flipping behavior exhibited by a peptide with a hydrophobic transmembrane helix, charged flanking loops, and a central, membrane-exposed aspartate residue of varying protonation state. We utilize all-atom temperature accelerated molecular dynamics simulations to predict the likelihood of loop-flipping without predefining specific loop-flipping pathways. We demonstrate that this approach can identify multiple possible flipping pathways, with the prevalence of each pathway depending on the protonation state of the central residue. In particular, we find that a charged central residue facilitates loop-flipping by stabilizing membrane water defects, enabling the "self-catalysis" of charge translocation. These findings provide detailed molecular-level insights into charged loop-flipping pathways that may generalize to other charge translocation processes, such as lipid flip-flop or the large-scale conformational rearrangements of multispanning membrane proteins.

Effects of lipid composition on membrane permeation

Passive permeation through lipid membranes is an essential process in biology. In vivo membranes typically consist of mixtures of lamellar and nonlamellar lipids. Lamellar lipids are characterized by their tendency to form lamellar sheet-like structures, which are predominant in nature. Nonlamellar lipids, when isolated, instead form more geometrically complex nonlamellar phases. While mixed lamellar/nonlamellar lipid membranes tend to adopt the ubiquitous lamellar bilayer structure, the presence of nonlamellar lipids is known to have profound effects on key membrane properties, such as internal distributions of stress and elastic properties, which in turn may alter related biological processes. This work focuses on one such process, i.e., permeation, by utilising atomistic molecular dynamics simulations in order to obtain transfer free energy profiles, diffusion profiles and permeation coefficients for a series of thirteen small molecules and drugs. Each permeant is tested on two bilayer membranes of different lipid composition, i.e., purely lamellar and mixed lamellar/nonlamellar. Our results indicate that the presence of nonlamellar lipids reduces permeation for smaller molecules (molecular weight < 100) but facilitates it for the largest ones (molecular weight > 100). This work represents an advancement towards the development of more realistic in silico permeability assays, which may have a substantial future impact in the area of rational drug design.

Scaffolds as Structural Tools for Bone-Targeted Drug Delivery

Although bone has a high potential to regenerate itself after damage and injury, the efficacious repair of large bone defects resulting from resection, trauma or non-union fractures still requires the implantation of bone grafts. Materials science, in conjunction with biotechnology, can satisfy these needs by developing artificial bones, synthetic substitutes and organ implants. In particular, recent advances in materials science have provided several innovations, underlying the increasing importance of biomaterials in this field. To address the increasing need for improved bone substitutes, tissue engineering seeks to create synthetic, three-dimensional scaffolds made from organic or inorganic materials, incorporating drugs and growth factors, to induce new bone tissue formation. This review emphasizes recent progress in materials science that allows reliable scaffolds to be synthesized for targeted drug delivery in bone regeneration, also with respect to past directions no longer considered promising. A general overview concerning modeling approaches suitable for the discussed systems is also provided.

Computer simulations of the diffusion of Na+ and Cl- ions across POPC lipid bilayer membranes

We have carried out molecular dynamics simulations using NAMD to study the diffusivity of Na and Cl ions across a POPC lipid bilayer membrane. We show that an imbalance of positively and negatively charged ions on either side of the membrane leads to the diffusion of ions and water molecules. We considered the cases of both weak and very strong charge imbalance across the membrane. The diffusion coefficients of the ions have been determined from the mean square displacements of the particles as a function of time. We find that for strong electrochemical gradients, both the Na and Cl ions diffuse rapidly through pores in the membrane with diffusion coefficients up to ten times larger than in water. Rather surprisingly, we found that although the Na ions are the first to begin the permeation process due to the lower potential barrier that they experience compared to the Cl ions, the latter complete the permeation across the barrier more quickly due to their faster diffusion rates.

Link between Membrane Composition and Permeability to Drugs

Prediction of membrane permeability to small molecules represents an important aspect of drug discovery. First-principles calculations of this quantity require an accurate description of both the thermodynamics and kinetics that underlie translocation of the permeant across the lipid bilayer. In this contribution, the membrane permeability to three drugs, or drug-like molecules, namely, 9-anthroic acid (ANA), 2',3'-dideoxyadenosine (DDA), and hydrocortisone (HYL), are estimated in a pure 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and in a POPC:cholesterol (2:1) mixture. On the basis of independent 2-5-μs free-energy calculations combined with a time-fractional Smoluchowski determination of the diffusivity, the estimated membrane permeabilities to these chemically diverse permeants fall within an order of magnitude from the experimental values obtained in egg-lecithin bilayers, with the exception of HYL in pure POPC. This exception is particularly interesting because the calculated permeability of the sterol-rich bilayer to HYL, in close agreement with the experimental value, is about 600 times lower than that of the pure POPC bilayer to HYL. In contrast, the permeabilities to ANA and DDA differ by less than a factor of 10 between the pure POPC and POPC:cholesterol bilayers. The unusual behavior of HYL, a large, amphiphilic compound, may be linked with the longer range perturbation of the lipid bilayer it induces, compared to ANA and DDA, suggestive of a possibly different translocation mechanism. We find that the tendency of lower permeabilities of the POPC:cholesterol bilayer relative to those of the pure POPC one is a consequence of increased free-energy barriers. Beyond reporting accurate estimates of the membrane permeability, the present contribution also demonstrates that rigorous free-energy calculations and a fractional-diffusion model are key in revealing the molecular phenomena linking the composition of a membrane to its permeability to drugs.

Free Energy Calculations of Membrane Permeation: Challenges Due to Strong Headgroup-Solute Interactions

Understanding how different classes of molecules move across biological membranes is a prerequisite to predicting a solute's permeation rate, which is a critical factor in the fields of drug design and pharmacology. We use biased molecular dynamics computer simulations to calculate and compare the free energy profiles of translocation of several small molecules across 1,2-dioleoyl- sn-glycero-3-phosphocholine (DOPC) lipid bilayers as a first step toward determining the most efficient method for free energy calculations. We study the translocation of arginine, a sodium ion, alanine, and a single water molecule using the metadynamics, umbrella sampling, and replica exchange umbrella sampling techniques. Within the fixed lengths of our simulations, we find that all methods produce similar results for charge-neutral permeants, but not for polar or positively charged molecules. We identify the long relaxation time scale of electrostatic interactions between lipid headgroups and the solute to be the principal cause of this difference and show that this slow process can lead to an erroneous dependence of computed free energy profiles on the initial system configuration. We demonstrate the use of committor analysis to validate the proper sampling of the presumed transition state, which in our simulations is achieved only in replica exchange calculations. On the basis of these results we provide some useful guidance to perform and evaluate free energy calculations of membrane permeation.

Predicting Binding Free Energies: Frontiers and Benchmarks

Binding free energy calculations based on molecular simulations provide predicted affinities for biomolecular complexes. These calculations begin with a detailed description of a system, including its chemical composition and the interactions among its components. Simulations of the system are then used to compute thermodynamic information, such as binding affinities. Because of their promise for guiding molecular design, these calculations have recently begun to see widespread applications in early-stage drug discovery. However, many hurdles remain in making them a robust and reliable tool. In this review, we highlight key challenges of these calculations, discuss some examples of these challenges, and call for the designation of standard community benchmark test systems that will help the research community generate and evaluate progress. In our view, progress will require careful assessment and evaluation of new methods, force fields, and modeling innovations on well-characterized benchmark systems, and we lay out our vision for how this can be achieved.

Position-Dependent Diffusion Tensors in Anisotropic Media from Simulation: Oxygen Transport in and through Membranes

A Bayesian-based methodology is developed to estimate diffusion tensors from molecular dynamics simulations of permeants in anisotropic media, and is applied to oxygen in lipid bilayers. By a separation of variables in the Smoluchowski diffusion equation, the multidimensional diffusion is reduced to coupled one-dimensional diffusion problems that are treated by discretization. The resulting diffusivity profiles characterize the membrane transport dynamics as a function of the position across the membrane, discriminating between diffusion normal and parallel to the membrane. The methodology is first validated with neat water, neat hexadecane, and a hexadecane slab surrounded by water, the latter being a simple model for a lipid membrane. Next, a bilayer consisting of pure 1-palmitoyl 2-oleoylphosphatidylcholine (POPC), and a bilayer mimicking the lipid composition of the inner mitochondrial membrane, including cardiolipin, are investigated. We analyze the detailed time evolution of oxygen molecules, in terms of both normal diffusion through and radial diffusion inside the membrane. Diffusion is fast in the more loosely packed interleaflet region, and anisotropic, with oxygen spreading more rapidly in the membrane plane than normal to it. Visualization of the propagator shows that oxygen enters the membrane rapidly, reaching its thermodynamically favored center in about 1 ns, despite the free energy barrier at the headgroup region. Oxygen transport is quantified by computing the oxygen permeability of the membranes and the average radial diffusivity, which confirm the anisotropy of the diffusion. The position-dependent diffusion constants and free energies are used to construct compartmental models and test assumptions used in estimating permeability, including Overton's rule. In particular, a hexadecane slab surrounded by water is found to be a poor model of oxygen transport in membranes because the relevant energy barriers differ substantially.

Permeability of a Fluid Lipid Bilayer to Short-Chain Alcohols from First Principles

Computational prediction of membrane permeability to small molecules requires accurate description of both the thermodynamics and kinetics underlying translocation across the lipid bilayer. In this contribution, well-converged, microsecond-long free-energy calculations are combined with a recently developed subdiffusive kinetics framework to describe the membrane permeation of a homologous series of short-tail alcohols, from methanol to 1-butanol, with unprecedented fidelity to the underlying molecular models. While the free-energy profiles exhibit barriers for passage through the center of the bilayer in all cases, the height of these barriers decreases with the length of the aliphatic chain of the alcohol, in quantitative agreement with experimentally determined differential solvation free energies in water and oil. A unique aspect of the subdiffusive model employed herein, which was developed in a previous article, is the determination of a position-dependent fractional order which quantifies the degree to which the motion of the alcohol deviates from classical diffusion along the thickness of the membrane. In the aqueous medium far from the bilayer, this quantity approaches 1.0, the asymptotic limit for purely classical diffusion, whereas it dips below 0.75 near the center of the membrane irrespective of the permeant. Remarkably, the fractional diffusivity near the center of membrane, where its influence on the permeability is the greatest, is similar among the four permeants despite the large difference in molecular weight and lipophilicity between methanol and 1-butanol. The relative permeabilities, which are estimated from the free-energy and fractional diffusivity profiles, are therefore determined predominantly by differences in the former rather than the latter. The predicted relative permeabilities are highly correlated with existing experimental results, albeit they do not agree quantitatively with them. On the other hand, quite unexpectedly, the reported experimental values for the short-tail alcohols are nearly three orders of magnitude lower than the available experimental measurement for water. Plausible explanations for this apparent disagreement between theory and experiment are considered in detail.

The role of intramolecular nonbonded interaction and angle sampling in single-step free energy perturbation

Single-step free energy perturbation (sFEP) has often been proposed as an efficient tool for a quick free energy scan due to its straightforward protocol and the ability to recycle an existing molecular dynamics trajectory for free energy calculations. Although sFEP is expected to fail when the sampling of a system is inefficient, it is often expected to hold for an alchemical transformation between ligands with a moderate difference in their sizes, e.g., transforming a benzene into an ethylbenzene. Yet, exceptions were observed in calculations for anisole and methylaniline, which have similar physical sizes as ethylbenzene. In this study, we show that such exceptions arise from the sampling inefficiency on an unexpected rigid degree of freedom, namely, the bond angle θ. The distributions of θ differ dramatically between two end states of a sFEP calculation, i.e., the conformation of the ligand changes significantly during the alchemical transformation process. Our investigation also reveals the interrelation between the ligand conformation and the intramolecular nonbonded interactions. This knowledge suggests a best combination of the ghost ligand potential and the dual topology setting, which improves the accuracy in a single reference sFEP calculation by bringing down its error from around 5k_BT to k_BT.

Subdiffusion in Membrane Permeation of Small Molecules

Within the solubility-diffusion model of passive membrane permeation of small molecules, translocation of the permeant across the biological membrane is traditionally assumed to obey the Smoluchowski diffusion equation, which is germane for classical diffusion on an inhomogeneous free-energy and diffusivity landscape. This equation, however, cannot accommodate subdiffusive regimes, which have long been recognized in lipid bilayer dynamics, notably in the lateral diffusion of individual lipids. Through extensive biased and unbiased molecular dynamics simulations, we show that one-dimensional translocation of methanol across a pure lipid membrane remains subdiffusive on timescales approaching typical permeation times. Analysis of permeant motion within the lipid bilayer reveals that, in the absence of a net force, the mean squared displacement depends on time as t0.7, in stark contrast with the conventional model, which assumes a strictly linear dependence. We further show that an alternate model using a fractional-derivative generalization of the Smoluchowski equation provides a rigorous framework for describing the motion of the permeant molecule on the pico- to nanosecond timescale. The observed subdiffusive behavior appears to emerge from a crossover between small-scale rattling of the permeant around its present position in the membrane and larger-scale displacements precipitated by the formation of transient voids.

Generalized Langevin Methods for Calculating Transmembrane Diffusivity

The membrane permeability coefficient of a solute can be estimated using the solubility-diffusion model. This model requires the diffusivity profile (D(z)) of the solute as it moves along the transmembrane axis, z. The generalized Langevin equation provides one strategy for calculating position-dependent diffusivity from straightforward molecular dynamics simulations where the solute is restrained to a series of positions on the z-coordinate by a harmonic potential. The diffusivity of the solute is calculated from its correlation functions, which are related to the friction experienced by the solute. Roux and Hummer have derived expressions for the diffusion coefficient from the velocity autocorrelation function (VACF) and position autocorrelation function (PACF), respectively. In this work, these methods are validated by calculating the diffusivity of H₂O and O₂ in homogeneous liquids. These methods are then used to calculate transmembrane diffusivity profiles. The VACF method is less sensitive to thermostat forces and has incrementally lower errors but is more sensitive to the spring constant of the harmonic restraint. For the permeation of a solute through a lipid bilayer, the diffusion coefficients calculated using these methods provided significantly different results. Long-lived correlations of the restrained solute due to inhomogeneities in the bilayer can result in spuriously low diffusivity when using the PACF method. The method based on the VACF does not have this issue and predicts higher rates of diffusion inside the bilayer.

Two Relations to Estimate Membrane Permeability Using Milestoning

Prediction of passive permeation rates of solutes across lipid bilayers is important to drug design, toxicology, and other biological processes such as signaling. The inhomogeneous solubility-diffusion (ISD) equation is traditionally used to relate the position-dependent potential of mean force and diffusivity to the permeability coefficient. The ISD equation is derived via the Smoluchowski equation and assumes overdamped system dynamics. It has been suggested that the complex membrane environment may exhibit more complicated damping conditions. Here we derive a variant of the inhomogeneous solubility diffusion equation as a function of the mean first passage time (MFPT) and show how milestoning, a method that can estimate kinetic quantities of interest, can be used to estimate the MFPT of membrane crossing and, by extension, the permeability coefficient. We further describe a second scheme, agnostic to the damping condition, to estimate the permeability coefficient from milestoning results or other methods that compute a probability of membrane crossing. The derived relationships are tested using a one-dimensional Langevin dynamics toy system confirming that the presented theoretical methods can be used to estimate permeabilities given simulation and milestoning results.

Simulation-Based Approaches for Determining Membrane Permeability of Small Compounds

Predicting the rate of nonfacilitated permeation of solutes across lipid bilayers is important to drug design, toxicology, and signaling. These rates can be estimated using molecular dynamics simulations combined with the inhomogeneous solubility-diffusion model, which requires calculation of the potential of mean force and position-dependent diffusivity of the solute along the transmembrane axis. In this paper, we assess the efficiency and accuracy of several methods for the calculation of the permeability of a model DMPC bilayer to urea, benzoic acid, and codeine. We compare umbrella sampling, replica exchange umbrella sampling, adaptive biasing force, and multiple-walker adaptive biasing force for the calculation of the transmembrane PMF. No definitive advantage for any of these methods in their ability to predict the membrane permeability coefficient Pm was found, provided that a sufficiently long equilibration is performed. For diffusivities, a Bayesian inference method was compared to a generalized Langevin method, both being sensitive to chosen parameters and the slow relaxation of membrane defects. Agreement within 1.5 log units of the computed Pm with experiment is found for all permeants and methods. Remaining discrepancies can likely be attributed to limitations of the force field as well as slowly relaxing collective movements within the lipid environment. Numerical calculations based on model profiles show that Pm can be reliably estimated from only a few data points, leading to recommendations for calculating Pm from simulations.

Permeability across lipid membranes

Molecular permeation through lipid membranes is a fundamental biological process that is important for small neutral molecules and drug molecules. Precise characterization of free energy surface and diffusion coefficients along the permeation pathway is required in order to predict molecular permeability and elucidate the molecular mechanisms of permeation. Several recent technical developments, including improved molecular models and efficient sampling schemes, are illustrated in this review. For larger penetrants, explicit consideration of multiple collective variables, including orientational, conformational degrees of freedom, are required to be considered in addition to the distance from the membrane center along the membrane normal. Although computationally demanding, this method can provide significant insights into the molecular mechanisms of permeation for molecules of medical and pharmaceutical importance. This article is part of a Special Issue entitled: Biosimulations edited by Ilpo Vattulainen and Tomasz Róg.

Comparative exploration of hydrogen sulfide and water transmembrane free energy surfaces via orthogonal space tempering free energy sampling

Hydrogen sulfide (H2 S), a commonly known toxic gas compound, possesses unique chemical features that allow this small solute molecule to quickly diffuse through cell membranes. Taking advantage of the recent orthogonal space tempering (OST) method, we comparatively mapped the transmembrane free energy landscapes of H2 S and its structural analogue, water (H2 O), seeking to decipher the molecular determinants that govern their drastically different permeabilities. As revealed by our OST sampling results, in contrast to the highly polar water solute, hydrogen sulfide is evidently amphipathic, and thus inside membrane is favorably localized at the interfacial region, that is, the interface between the polar head-group and nonpolar acyl chain regions. Because the membrane binding affinity of H2 S is mainly governed by its small hydrophobic moiety and the barrier height inbetween the interfacial region and the membrane center is largely determined by its moderate polarity, the transmembrane free energy barriers to encounter by this toxic molecule are very small. Moreover when H2 S diffuses from the bulk solution to the membrane center, the above two effects nearly cancel each other, so as to lead to a negligible free energy difference. This study not only explains why H2 S can quickly pass through cell membranes but also provides a practical illustration on how to use the OST free energy sampling method to conveniently analyze complex molecular processes. © 2015 Wiley Periodicals, Inc.

Sampling errors in free energy simulations of small molecules in lipid bilayers

Free energy simulations are a powerful tool for evaluating the interactions of molecular solutes with lipid bilayers as mimetics of cellular membranes. However, these simulations are frequently hindered by systematic sampling errors. This review highlights recent progress in computing free energy profiles for inserting molecular solutes into lipid bilayers. Particular emphasis is placed on a systematic analysis of the free energy profiles, identifying the sources of sampling errors that reduce computational efficiency, and highlighting methodological advances that may alleviate sampling deficiencies. This article is part of a Special Issue entitled: Biosimulations edited by Ilpo Vattulainen and Tomasz Róg.

Diffusive Models of Membrane Permeation with Explicit Orientational Freedom

Accurate calculation of permeabilities from first-principles has been a long-standing challenge for computer simulations, notably in the context of drug discovery, as a route to predict the propensity of small, organic molecules to spontaneously translocate biological membranes. Of equal importance is the understanding of the permeation process and the pathway followed by the permeant from the aqueous medium to the interior of the lipid bilayer, and back out again. A convenient framework for the computation of permeabilities is provided by the solubility-diffusion model, which requires knowledge of the underlying free-energy and diffusivity landscapes. Here, we develop a formalism that includes an explicit description of the orientational motion of the solute as it diffuses across the membrane. Toward this end, we have generalized a recently proposed method that reconciles thermodynamics and kinetics in importance-sampling simulations by means of a Bayesian-inference scheme to reverse-solve the underlying Smoluchowski equation. Performance of the proposed formalism is examined in the model cases of a water and an ethanol molecule crossing a fully hydrated lipid bilayer. Our analysis reveals a conspicuous dependence of the free-energy and rotational diffusivity on the orientation of ethanol when it lies within the headgroup region of the bilayer. Specifically, orientations for which the hydroxyl group lies among the polar lipid head groups, while the ethyl group recedes toward the hydrophobic interior are associated with free-energy minima ∼2kBT deep, as well as significantly slower orientational kinetics compared to the bulk solution or the core of the bilayer. The conspicuous orientational anisotropy of ethanol at the aqueous interface is suggestive of a complete rotation of the permeant as it crosses the hydrophobic interior of the membrane.

A Permeability Study of O2 and the Trace Amine p-Tyramine through Model Phosphatidylcholine Bilayers

We study here the permeability of the hydrophobic O2 molecule through a model DPPC bilayer at 323K and 350K, and of the trace amine p-tyramine through PC bilayers at 310K. The tyramine results are compared to previous experimental work at 298K. Nonequilibrium work methods were used in conjunction to simultaneously obtain both the potential of mean force (PMF) and the position dependent transmembrane diffusion coefficient, D(z), from the simulations. These in turn were used to calculate the permeability coefficient, P, through the inhomogeneous solubility-diffusion model. The results for O2 are consistent with previous simulations, and agree with experimentally measured P values for PC bilayers. A temperature dependence in the permeability of O2 through DPPC was obtained, with P decreasing at higher temperatures. Two relevant species of p-tyramine were simulated, from which the PMF and D(z) were calculated. The charged species had a large energetic barrier to crossing the bilayer of ~ 21 kcal/mol, while the uncharged, deprotonated species had a much lower barrier of ~ 7 kcal/mol. The effective in silico permeability for p-tyramine was calculated by applying three approximations, all of which gave nearly identical results (presented here as a function of the pKa). As the permeability value calculated from simulation was highly dependent on the pKa of the amine group, a further pKa study was performed that also varied the fraction of the uncharged and zwitterionic p-tyramine species. Using the experimental P value together with the simulated results, we were able to label the phenolic group as responsible for the pKa1 and the amine for the pKa2, that together represent all of the experimentally measured pKa values for p-tyramine. This agrees with older experimental results, in contrast to more recent work that has suggested there is a strong ambiguity in the pKa values.

Extracting the diffusion tensor from molecular dynamics simulation with Milestoning

We propose an algorithm to extract the diffusion tensor from Molecular Dynamics simulations with Milestoning. A Kramers-Moyal expansion of a discrete master equation, which is the Markovian limit of the Milestoning theory, determines the diffusion tensor. To test the algorithm, we analyze overdamped Langevin trajectories and recover a multidimensional Fokker-Planck equation. The recovery process determines the flux through a mesh and estimates local kinetic parameters. Rate coefficients are converted to the derivatives of the potential of mean force and to coordinate dependent diffusion tensor. We illustrate the computation on simple models and on an atomically detailed system-the diffusion along the backbone torsions of a solvated alanine dipeptide.

Hydrophobic aggregation and collective absorption of dioxin into lipid membranes: insights from atomistic simulations

Dioxins are a highly toxic class of chlorinated aromatic chemicals. They have been extensively studied, but several molecular-level details of their action are still missing. Here we present molecular dynamics simulations of their absorption and diffusion through cell membranes. We show that, due to their hydrophobic character, dioxins can quickly penetrate into a lipid membrane, both as single molecules and as aggregates. We find clear evidence for their ability to accumulate in cell membranes. Our free energy calculations indicate that subsequent transport into the cell is unlikely to be a simple diffusive process.

Microsecond molecular dynamics simulations of lipid mixing

Molecular dynamics (MD) simulations of membranes are often hindered by the slow lateral diffusion of lipids and the limited time scale of MD. In order to study the dynamics of mixing and characterize the lateral distribution of lipids in converged mixtures, we report microsecond-long all-atom MD simulations performed on the special-purpose machine Anton. Two types of mixed bilayers, POPE:POPG (3:1) and POPC:cholesterol (2:1), as well as a pure POPC bilayer, were each simulated for up to 2 μs. These simulations show that POPE:POPG and POPC:cholesterol are each fully miscible at the simulated conditions, with the final states of the mixed bilayers similar to a random mixture. By simulating three POPE:POPG bilayers at different NaCl concentrations (0, 0.15, and 1 M), we also examined the effect of salt concentration on lipid mixing. While an increase in NaCl concentration is shown to affect the area per lipid, tail order, and lipid lateral diffusion, the final states of mixing remain unaltered, which is explained by the largely uniform increase in Na(+) ions around POPE and POPG. Direct measurement of water permeation reveals that the POPE:POPG bilayer with 1 M NaCl has reduced water permeability compared with those at zero or low salt concentration. Our calculations provide a benchmark to estimate the convergence time scale of all-atom MD simulations of lipid mixing. Additionally, equilibrated structures of POPE:POPG and POPC:cholesterol, which are frequently used to mimic bacterial and mammalian membranes, respectively, can be used as starting points of simulations involving these membranes.

The adaptive biasing force method: everything you always wanted to know but were afraid to ask

In the host of numerical schemes devised to calculate free energy differences by way of geometric transformations, the adaptive biasing force algorithm has emerged as a promising route to map complex free-energy landscapes. It relies upon the simple concept that as a simulation progresses, a continuously updated biasing force is added to the equations of motion, such that in the long-time limit it yields a Hamiltonian devoid of an average force acting along the transition coordinate of interest. This means that sampling proceeds uniformly on a flat free-energy surface, thus providing reliable free-energy estimates. Much of the appeal of the algorithm to the practitioner is in its physically intuitive underlying ideas and the absence of any requirements for prior knowledge about free-energy landscapes. Since its inception in 2001, the adaptive biasing force scheme has been the subject of considerable attention, from in-depth mathematical analysis of convergence properties to novel developments and extensions. The method has also been successfully applied to many challenging problems in chemistry and biology. In this contribution, the method is presented in a comprehensive, self-contained fashion, discussing with a critical eye its properties, applicability, and inherent limitations, as well as introducing novel extensions. Through free-energy calculations of prototypical molecular systems, many methodological aspects are examined, from stratification strategies to overcoming the so-called hidden barriers in orthogonal space, relevant not only to the adaptive biasing force algorithm but also to other importance-sampling schemes. On the basis of the discussions in this paper, a number of good practices for improving the efficiency and reliability of the computed free-energy differences are proposed.