Pore Formation by Amyloid-like Peptides: Effects of the Nonpolar-Polar Sequence Pattern.

One of the mechanisms accounting for the toxicity of amyloid peptides in diseases like Alzheimer's and Parkinson's is the formation of pores on the plasma membrane of neurons. Here, we perform unbiased all-atom simulations of the full membrane damaging pathway, which includes adsorption, aggregation, and perforation of the lipid bilayer accounting for pore-like structures. Simulations are performed using four peptides made with the same amino acids. Differences in the nonpolar-polar sequence pattern of these peptides prompt them to adsorb into the membrane with the extended conformations oriented either parallel [peptide labeled F1, Ac-(FKFE)2-NH2], perpendicular (F4, Ac-FFFFKKEE-NH2), or with an intermediate orientation (F2, Ac-FFKKFFEE-NH2, and F3, Ac-FFFKFEKE-NH2) in regard to the membrane surface. At the water-lipid interface, only F1 fully self-assembles into ß-sheets, and F2 peptides partially fold into an α-helical structure. The ß-sheets of F1 emerge as electrostatic interactions attract neighboring peptides to intermediate distances where nonpolar side chains can interact within the dry core of the bilayer. This complex interplay between electrostatic and nonpolar interactions is not observed for the other peptides. Although ß-sheets of F1 peptides are mostly parallel to the membrane, some of their edges penetrate deep inside the bilayer, dragging water molecules with them. This precedes pore formation, which starts with the flow of two water layers through the membrane that expand into a stable cylindrical pore delimited by polar faces of ß-sheets spanning both leaflets of the bilayer.

Peptide Self-Assembly into Amyloid Fibrils: Unbiased All-Atom Simulations.

Protein self-assembly plays an important role in biological systems, accounting for the formation of mesoscopic structures that can be highly symmetric as in the capsid of viruses or disordered as in molecular condensates or exhibit a one-dimensional fibrillar morphology as in amyloid fibrils. Deposits of the latter in tissues of individuals with degenerative diseases like Alzheimer's and Parkinson's has motivated extensive efforts to understand the sequence of molecular events accounting for their formation. These studies aim to identify on-pathway intermediates that may be the targets for therapeutic intervention. This detailed knowledge of fibril formation remains obscure, in part due to challenges with experimental analyses of these processes. However, important progress is being achieved for short amyloid peptides due to advances in our ability to perform completely unbiased all-atom simulations of the self-assembly process. This perspective discusses recent developments, their implications, and the hurdles that still need to be overcome to further advance the field.

Nucleation and Growth of Amyloid Fibrils.

The formation of amyloid fibrils is a complex phenomenon that remains poorly understood at the atomic scale. Herein, we perform extended unbiased all-atom simulations in explicit solvent of a short amphipathic peptide to shed light on the three mechanisms accounting for fibril formation, namely, nucleation via primary and secondary mechanisms, and fibril growth. We find that primary nucleation takes place via the formation of an intermediate state made of two laminated ß-sheets oriented perpendicular to each other. The amyloid fibril spine subsequently emerges from the rotation of these ß-sheets to account for peptides that are parallel to each other and perpendicular to the main axis of the fibril. Growth of this spine, in turn, takes place via a dock-and-lock mechanism. We find that peptides dock onto the fibril tip either from bulk solution or after diffusing on the fibril surface. The latter docking pathway contributes significantly to populate the fibril tip with peptides. We also find that side chain interactions drive the motion of peptides in the lock phase during growth, enabling them to adopt the structure imposed by the fibril tip with atomic fidelity. Conversely, the docked peptide becomes trapped in a local free energy minimum when docked-conformations are sampled randomly. Our simulations also highlight the role played by nonpolar fibril surface patches in catalyzing and orienting the formation of small cross-ß structures. More broadly, our simulations provide important new insights into the pathways and interactions accounting for primary and secondary nucleation as well as the growth of amyloid fibrils.

Peptide-Membrane Binding: Effects of the Amino Acid Sequence.

An understanding of how the amino acid sequence affects the interaction of peptides with lipid membranes remains mostly unknown. This type of knowledge is required to rationalize membrane-induced toxicity of amyloid peptides and to design peptides that can interact with lipid bilayers. Here, we perform a systematic study of how variations in the sequence of the amphipathic Ac-(FKFE)2-NH2 peptide affect its interaction with zwitterionic lipid bilayers using extensive all-atom molecular dynamics simulations in explicit solvent. Our results show that peptides with a net positive charge bind more frequently to the lipid bilayer than neutral or negatively charged sequences. Moreover, neutral amphipathic peptides made with the same numbers of phenylalanine (F), lysine (K), and glutamic (E) amino acids at different positions in the sequence differ significantly in their frequency of binding to the membrane. We find that peptides bind with a higher frequency to the membrane if their positive lysine side chains are more exposed to the solvent, which occurs if they are located at the extremity (as opposed to the middle) of the sequence. Non-polar residues play an important role in accounting for the adsorption of peptides onto the membrane. In particular, peptides made with less hydrophobic non-polar residues (e.g., valine and alanine) are significantly less adsorbed to the membrane compared to peptides made with phenylalanine. We also find that sequences where phenylalanine residues are located at the extremities of the peptide have a higher tendency to be adsorbed.

Atomic Insights into Amyloid-Induced Membrane Damage.

Amphipathic peptides can cause biological membranes to leak either by dissolving their lipid content via a detergent-like mechanism or by forming pores on the membrane surface. These modes of membrane damage have been related to the toxicity of amyloid peptides and to the activity of antimicrobial peptides. Here, we perform the first all-atom simulations in which membrane-bound amphipathic peptides self-assemble into ß-sheets that subsequently either form stable pores inside the bilayer or drag lipids out of the membrane surface. An analysis of these simulations shows that the acyl tail of lipids interact strongly with non-polar side chains of peptides deposited on the membrane. These strong interactions enable lipids to be dragged out of the bilayer by oligomeric structures accounting for detergent-like damage. They also disturb the orientation of lipid tails in the vicinity of peptides. These distortions are minimized around pore structures. We also show that membrane-bound ß-sheets become twisted with one of their extremities partially penetrating the lipid bilayer. This allows peptides on opposite leaflets to interact and form a long transmembrane ß-sheet, which initiates poration. In simulations, where peptides are deposited on a single leaflet, the twist in ß-sheets allows them to penetrate the membrane and form pores. In addition, our simulations show that fibril-like structures produce little damage to lipid membranes, as non-polar side chains in these structures are unavailable to interact with the acyl tail of lipids.

Binding Mechanisms of Amyloid-like Peptides to Lipid Bilayers and Effects of Divalent Cations.

In several neurodegenerative diseases, cell toxicity can emerge from damage produced by amyloid aggregates to lipid membranes. The details accounting for this damage are poorly understood including how individual amyloid peptides interact with phospholipid membranes before aggregation. Here, we use all-atom molecular dynamics simulations to investigate the molecular mechanisms accounting for amyloid-membrane interactions and the role played by calcium ions in this interaction. Model peptides known to self-assemble into amyloid fibrils and bilayer made from zwitterionic and anionic lipids are used in this study. We find that both electrostatic and hydrophobic interactions contribute to peptide-bilayer binding. In particular, the attraction of peptides to lipid bilayers is dominated by electrostatic interactions between positive residues and negative phosphate moieties of lipid head groups. This attraction is stronger for anionic bilayers than for zwitterionic ones. Hydrophobicity drives the burial of nonpolar residues into the interior of the bilayer producing strong binding in our simulations. Moreover, we observe that the attraction of peptides to the bilayer is significantly reduced in the presence of calcium ions. This is due to the binding of calcium ions to negative phosphate moieties of lipid head groups, which leaves phospholipid bilayers with a net positive charge. Strong binding of the peptide to the membrane occurs less frequently in the presence of calcium ions and involves the formation of a "Ca2+ bridge".

Methane Clathrate Formation is Catalyzed and Kinetically Inhibited by the Same Molecule: Two Facets of Methanol.

Here, we perform molecular dynamics simulations to provide atomic-level insights into the dual roles of methanol in enhancing and delaying the rate of methane clathrate hydrate nucleation. Consistent with experiments, we find that methanol slows clathrate hydrate nucleation above 250 K but promotes clathrate formation at temperatures below 250 K. We show that this behavior can be rationalized by the unusual temperature dependence of the methane-methanol interaction in an aqueous solution, which emerges due to the hydrophobic effect. In addition to its antifreeze properties at temperatures above 250 K, methanol competes with water to interact with methane prior to the formation of clathrate nuclei. Below 250 K, methanol encourages water to occupy the space between methane molecules favoring clathrate formation and it may additionally promote water mobility.

Effects of Ions and Small Compounds on the Structure of Aß42 Monomers.

Aggregation of amyloid-ß (Aß) proteins in the brain is a hallmark of Alzheimer's disease. This phenomenon can be promoted or inhibited by adding small molecules to the solution where Aß is embedded. These molecules affect the ensemble of conformations sampled by Aß monomers even before aggregation starts. Here, we perform extensive all-atom replica exchange molecular dynamics (REMD) simulations to provide a comparative study of the ensemble of conformations sampled by Aß42 monomers in solutions that promote (i.e., aqueous solution containing NaCl) and inhibit (i.e., aqueous solutions containing scyllo-inositol or 4-aminophenol) aggregation. Simulations performed in pure water are used as our reference. We find that secondary-structure content is only affected in an antagonistic manner by promoters and inhibitors at the C-terminus and the central hydrophilic core. Moreover, the end of the C-terminus binds more favorably to the central hydrophobic core region of Aß42 in NaCl adopting a type of strand-loop-strand structure that is disfavored by inhibitors. Nonpolar residues that form the dry core of larger aggregates of Aß42 (e.g., PDB ID 2BEG) are found at close proximity in these strand-loop-strand structures, suggesting that their formation could play an important role in initiating nucleation. In the presence of inhibitors, the C-terminus binds the central hydrophilic core with a higher probability than in our reference simulation. This sensitivity of the C-terminus, which is affected in an antagonistic manner by inhibitors and promoters, provides evidence for its critical role in accounting for aggregation.

Role of Cholesterol on Binding of Amyloid Fibrils to Lipid Bilayers.

Molecular dynamics simulations are used to provide insights into the molecular mechanisms accounting for binding of amyloid fibrils to lipid bilayers and to study the effect of cholesterol in this process. We show that electrostatic interactions play an important role in fibril-bilayer binding and cholesterol modulates this interaction. In particular, the interaction between positive residues and lipid head groups becomes more favorable in the presence of cholesterol. Consistent with experiments, we find that cholesterol enhances fibril-membrane binding.

Thermodynamic Stability of Polar and Nonpolar Amyloid Fibrils.

Thermodynamic stabilities of amyloid fibrils remain mostly unknown due to experimental challenges. Here, we combine enhanced sampling methods to simulate all-atom models in explicit water in order to study the stability of nonpolar (Aß16-21) and polar (IAPP28-33) fibrils. We find that the nonpolar fibril becomes more stable with increasing temperature, and its stability is dominated by entropy. In contrast, the polar fibril becomes less stable with increasing temperature, while it is stabilized by enthalpy. Our results show that the nature of side chains in the dry core of amyloid fibrils plays a dominant role in accounting for their thermodynamic stability.

Effects of Trimethylamine- N-oxide (TMAO) on Hydrophobic and Charged Interactions.

Effects of trimethylamine- N-oxide (TMAO) on hydrophobic and charge-charge interactions are investigated using molecular dynamics simulations. Recently, these interactions in model peptides and in the Trp-Cage miniprotein have been reported to be strongly affected by TMAO. Neopentane dimers and Na+Cl- are used, here, as models for hydrophobic and charge-charge interactions, respectively. Distance-dependent interactions, i.e., potential of mean force, are computed using an umbrella sampling protocol at different temperatures which allows us to determine enthalpy and entropic energies. We find that the large favorable entropic energy and the unfavorable enthalpy, which are characteristic of hydrophobic interactions, become smaller when TMAO is added to water. These changes account for a negligible effect and a stabilizing effect on the strength of hydrophobic interactions for simulations performed with Kast and Netz models of TMAO, respectively. Effects of TMAO on the enthalpy are mainly due to changes in terms of the potential energy involving solvent-solvent molecules. At the molecular level, TMAO is incorporated in the solvation shell of neopentane which may explain its effect on the enthalpy and entropic energy. Charge-charge interactions become stronger when TMAO is added to water because this osmolyte decreases the enthalpic penalty of bringing Na+ and Cl- close together mainly by affecting ion-solvent interactions. TMAO is attracted to Na+, becoming part of its solvation shell, whereas it is excluded from the vicinity of Cl-. These results are more pronounced for simulation performed with the Netz model which is more hydrophobic and has a larger dipole moment compared to the Kast model of TMAO.

Effects of Trimethylamine-N-oxide on the Conformation of Peptides and its Implications for Proteins.

To provide insights into the stabilizing mechanisms of trimethylamine-N-oxide (TMAO) on protein structures, we perform all-atom molecular dynamics simulations of peptides and the Trp-cage miniprotein. The effects of TMAO on the backbone and charged residues of peptides are found to stabilize compact conformations, whereas effects of TMAO on nonpolar residues lead to peptide swelling. This suggests competing mechanisms of TMAO on proteins, which accounts for hydrophobic swelling, backbone collapse, and stabilization of charge-charge interactions. These mechanisms are observed in Trp cage.

Thermodynamic properties of amyloid fibrils in equilibrium.

In this manuscript we use a two-dimensional coarse-grained model to study how amyloid fibrils grow towards an equilibrium state where they coexist with proteins dissolved in a solution. Free-energies to dissociate proteins from fibrils are estimated from the residual concentration of dissolved proteins. Consistent with experiments, the concentration of proteins in solution affects the growth rate of fibrils but not their equilibrium state. Also, studies of the temperature dependence of the equilibrium state can be used to estimate thermodynamic quantities, e.g., heat capacity and entropy.

Thermodynamics of Aß16-21 dissociation from a fibril: Enthalpy, entropy, and volumetric properties.

Here, we provide insights into the thermodynamic properties of A ß16-21 dissociation from an amyloid fibril using all-atom molecular dynamics simulations in explicit water. An umbrella sampling protocol is used to compute potentials of mean force (PMF) as a function of the distance ξ between centers-of-mass of the A ß16-21 peptide and the preformed fibril at nine temperatures. Changes in the enthalpy and the entropic energy are determined from the temperature dependence of these PMF(s) and the average volume of the simulation box is computed as a function of ξ. We find that the PMF at 310 K is dominated by enthalpy while the entropic energy does not change significantly during dissociation. The volume of the system decreases during dissociation. Moreover, the magnitude of this volume change also decreases with increasing temperature. By defining dock and lock states using the solvent accessible surface area (SASA), we find that the behavior of the electrostatic energy is different in these two states. It increases (unfavorable) and decreases (favorable) during dissociation in lock and dock states, respectively, while the energy due to Lennard-Jones interactions increases continuously in these states. Our simulations also highlight the importance of hydrophobic interactions in accounting for the stability of A ß16-21.

Role of side-chain interactions on the formation of α-helices in model peptides.

The role played by side-chain interactions on the formation of α-helices is studied using extensive all-atom molecular dynamics simulations of polyalanine-like peptides in explicit TIP4P water. The peptide is described by the OPLS-AA force field except for the Lennard-Jones interaction between Cß-Cß atoms, which is modified systematically. We identify values of the Lennard-Jones parameter that promote α-helix formation. To rationalize these results, potentials of mean force (PMF) between methane-like molecules that mimic side chains in our polyalanine-like peptides are computed. These PMF exhibit a complex distance dependence where global and local minima are separated by an energy barrier. We show that α-helix propensity correlates with values of these PMF at distances corresponding to Cß-Cß of i-i+3 and other nearest neighbors in the α-helix. In particular, the set of Lennard-Jones parameters that promote α-helices is characterized by PMF that exhibit a global minimum at distances corresponding to i-i+3 neighbors in α-helices. Implications of these results are discussed.

Driving ß-strands into fibrils.

In this work we study contributions of mainchain and side chain atoms to fibrillization of polyalanine peptides using all-atom molecular dynamics simulations. We show that the total number of hydrogen bonds in the system does not change significantly during aggregation. This emerges from a compensatory mechanism where the formation of one interpeptide hydrogen bond requires rupture of two peptide-water bonds, leading to the formation of one extra water-water bond. Since hydrogen bonds are mostly electrostatic in nature, this mechanism implies that electrostatic energies related to these bonds are not minimized during fibril formation. Therefore, hydrogen bonds do not drive fibrillization in all-atom models. Nevertheless, they play an important role in this process since aggregation without the formation of interpeptide hydrogen bonds accounts for a prohibitively large electrostatic penalty (~9.4 kJ/mol). Our work also highlights the importance of using accurate models to describe chemical bonds since Lennard-Jones and electrostatic contributions of different chemical groups of the protein and solvent are 1 order of magnitude larger than the overall enthalpy of the system. Thus, small errors in modeling these interactions can produce large errors in the total enthalpy of the system.

Pressure-Dependent Properties of Elementary Hydrophobic Interactions: Ramifications for Activation Properties of Protein Folding.

Hydration effects on a pair of methane molecules are investigated by extensive constant-pressure (NPT) sampling using the TIP4P model of water under 1, 1000, 2000, and 3000 atm. The volume distributions of pure water and of methanes plus water are determined directly as functions of methane-methane distance ξ. The corresponding excess isothermal and adiabatic compressibilities are estimated from the pressure-dependent methane excess volume. The dependence of excess volume on ξ is oscillatory for small ξ. The maxima of excess volume and compressibility are seen near the desolvation barrier (db) of the potential of mean force (PMF). These features may be understood by the development, near the db, of a void volume encased by a molecular (Connolly) surface defined using a water-sized probe. These db properties for two methanes are consistent with well-corroborated experimental observations of positive activation volumes for protein folding and some experiments suggesting a slightly higher compressibility for the folding transition state than the unfolded state. At high pressures, the volumes at the PMF solvent-separated minimum and the contact-minimum configurations are both smaller than the volume at large ξ. This trend provides a rationalization for the compactness of pressure-denatured states of proteins. Taking the packing densities of pure nonpolar phases into consideration, our simulation results suggest that whether the activation volume of unfolding is positive or negative hinges on the packing compactness of the protein core. Volume change can be but is not necessarily monotonic along the folding pathway.

Hydration of non-polar anti-parallel ß-sheets.

In this work we focus on anti-parallel ß-sheets to study hydration of side chains and polar groups of the backbone using all-atom molecular dynamics simulations. We show that: (i) water distribution around the backbone does not depend significantly on amino acid sequence, (ii) more water molecules are found around oxygen than nitrogen atoms of the backbone, and (iii) water molecules around nitrogen are highly localized in the planed formed by peptide backbones. To study hydration around side chains we note that anti-parallel ß-sheets exhibit two types of cross-strand pairing: Hydrogen-Bond (HB) and Non-Hydrogen-Bond (NHB) pairing. We show that distributions of water around alanine, leucine, and valine side chains are very different at HB compared to NHB faces. For alanine pairs, the space between side chains has a higher concentration of water if residues are located in the NHB face of the ß-sheet as opposed to the HB face. For leucine residues, the HB face is found to be dry while the space between side chains at the NHB face alternates between being occupied and non-occupied by water. Surprisingly, for valine residues the NHB face is dry, whereas the HB face is occupied by water. We postulate that these differences in water distribution are related to context dependent propensities observed for ß-sheets.

Exploring the free energy landscape of a model ß-hairpin peptide and its isoform.

Secondary structural transitions from α-helix to ß-sheet conformations are observed in several misfolding diseases including Alzheimer's and Parkinson's. Determining factors contributing favorably to the formation of each of these secondary structures is therefore essential to better understand these disease states. ß-hairpin peptides form basic components of anti-parallel ß-sheets and are suitable model systems for characterizing the fundamental forces stabilizing ß-sheets in fibrillar structures. In this study, we explore the free energy landscape of the model ß-hairpin peptide GB1 and its E2 isoform that preferentially adopts α-helical conformations at ambient conditions. Umbrella sampling simulations using all-atom models and explicit solvent are performed over a large range of end-to-end distances. Our results show the strong preference of GB1 and the E2 isoform for ß-hairpin and α-helical conformations, respectively, consistent with previous studies. We show that the unfolded states of GB1 are largely populated by misfolded ß-hairpin structures which differ from each other in the position of the ß-turn. We discuss the energetic factors contributing favorably to the formation of α-helix and ß-hairpin conformations in these peptides and highlight the energetic role of hydrogen bonds and non-bonded interactions.

Properties of the Lennard-Jones dimeric fluid in two dimensions: an integral equation study.

The thermodynamic and structural properties of the planar soft-sites dumbbell fluid are examined by Monte Carlo simulations and integral equation theory. The dimers are built of two Lennard-Jones segments. Site-site integral equation theory in two dimensions is used to calculate the site-site radial distribution functions for a range of elongations and densities and the results are compared with Monte Carlo simulations. The critical parameters for selected types of dimers were also estimated. We analyze the influence of the bond length on critical point as well as tested correctness of site-site integral equation theory with different closures. The integral equations can be used to predict the phase diagram of dimers whose molecular parameters are known.