Alice Glättli

Derivation of an improved simple point charge model for liquid water: SPC/A and SPC/L

Different approaches to improve the simple point charge model for liquid water (SPC) were investigated. This led to a whole series of new water models with additional van der Waals interaction sites at the hydrogen atoms, modified partial charges and modified geometries. The properties of these models are analyzed and discussed. Particular emphasis has been put on the study of the dependence and sensitivity of water properties on the model parameters. We found that a simultaneous improvement of the dielectric permittivity and the diffusion coefficient is difficult to attain for a rigid, nonpolarizable three interaction site model. Nevertheless, two of the models presented here, SPC/A and SPC/L, show good agreement with experimental data on water and have been characterized in more detail. We conclude that SPC/L represents the overall properties of water better than SPC. Especially, it shows excellent dielectric properties, an improved shear viscosity and a slightly lower diffusion coefficient.

Unfolded state of peptides.

Publisher Summary This chapter illustrates the current knowledge on the unfolded state of peptides. Currently, the characterization of the populated microscopic states of a peptide is possible only by simulation methods. The focus is on molecular dynamics simulations of spontaneous (i.e., lacking biasing potentials or directed-sampling algorithms) and reversible folding of peptides in solution (i.e., with explicit solvent molecules), since the results from this type of studies are the least dependent on the methodology. A full characterization of the unfolded state becomes as essential as the determination of the folded conformation in two scenarios. The first is in the study of the physical, chemical, and biological properties of peptides. The folded conformation of a peptide is, in general, only marginally more stable than the lowest-free-energy unfolded conformation. As a result, any macroscopic observable of a peptide is weighted with both the folded and the unfolded states. Interpreting such observables in terms of the folded conformation only is therefore not correct. The second scenario is related to the study of peptide and protein folding. The description of equilibrium requires knowledge about each of the states involved. Therefore, it is fundamental to describe not only the folded state but also the unfolded state accurately in order to draw conclusions on the nature and mechanisms of peptide and protein folding.

A molecular dynamics study of the bee venom melittin in aqueous solution, in methanol, and inserted in a phospholipid bilayer

The structural properties of melittin, a small amphipathic peptide found in the bee venom, are investigated in three different environments by molecular dynamics simulation. Long simulations have been performed for monomeric melittin solvated in water, in methanol, and shorter ones for melittin inserted in a dimyristoylphosphatidylcholine bilayer. The resulting trajectories were analysed in terms of structural properties of the peptide and compared to the available NMR data. While in water and methanol solution melittin is observed to partly unfold, the peptide retains its structure when embedded in a lipid bilayer. The latter simulation shows good agreement with the experimentally derived 3J-coupling constants. Generally, it appears that higher the stability of the helical conformation of melittin, lower is the dielectric permittivity of the environment. In addition, peptide-lipid interactions were investigated showing that the C-terminus of the peptide provides an anchor to the lipid bilayer by forming hydrogen bonds with the lipid head groups.

An improved dimethyl sulfoxide force field for molecular dynamics simulations

A united-atom molecular simulation force field for liquid dimethyl sulfoxide has been found to produce unacceptably inaccurate densities when used with a reaction-field or Ewald treatment of the electrostatic interactions. The force field is mildly reparameterized leading to a smaller dipole moment and slightly larger methyl groups. In addition to being compatible with the more sophisticated treatment of Coulombic interactions, the new force field also results in a significantly better description of the diffusion coefficient, the shear viscosity and the dielectric constant. Other liquid properties remain at the satisfactory quality of the Liu et al. model.

A novel approach for designing simple point charge models for liquid water with three interaction sites

A simultaneous improvement of the diffusion and dielectric properties of the simple point charge (SPC) model for liquid water appears to be very difficult with conventional reparametrization of the commonly used Lennard–Jones and Coulomb interaction functions and without including a self‐energy correction in the effective pair‐potential as is done in the SPC/E model. Here, a different approach to circumvent this problem is presented. A short‐range interaction term, which corrects the oxygen–oxygen energy at small distances by small amounts of energy, was introduced in the nonbonded interaction function. This additional force‐field term allows to derive new parameter sets for SPC‐like water models that yield better agreement with experimental data on liquid water. Based on previous investigations of the force‐field parameter dependence of the water properties of SPC‐like models, the necessary parameter changes to obtain a lower diffusion coefficient and a larger dielectric permittivity were specified and accordingly six new models were developed. They all represent an improvement over SPC in terms of structural and diffusional properties, four of them show better dielectric properties also. One model, SPC/S, has been characterized in more detail, and represents most properties of liquid water better than SPC while avoiding the larger discrepancies with experimental values regarding density, thermal compressibility, energy, and free energy of the SPC/E model. We conclude that the use of a simple, short‐ranged additional oxygen–oxygen interaction term makes a simultaneous improvement of the diffusion coefficient and the dielectric properties of water feasible.

Molecular dynamics simulation of lipid bilayers with GROMOS96: Application of surface tension

The GROMOS96 force fields 45A3 and 53A5, when applied to dipalmitoylphosphatidylcholine (DPPC) membranes, have a tendency to result in a reduced area per lipid in constant pressure simulations. The application of surface tension is effective in increasing the area per lipid, a measure of the phase of the membrane, but only if the area is already close to the experimental range. Therefore the surface tension cannot compensate for strong inadequacies in the force-field parameters. The behaviour of the 45A3 force field from long NP n γT simulations of tens of nanoseconds is analysed over a range of different surface tensions. Comparisons are made with the corresponding NP n AT simulations.

Membrane protein dynamics in different environments: simulation study of the outer membrane protein X in a lipid bilayer and in a micelle

The bacterial outer membrane protein OmpX from Escherichia coli has been investigated by molecular dynamics simulations when embedded in a phospholipid bilayer and as a protein-micelle aggregate. The resulting simulation trajectories were analysed in terms of structural and dynamic properties of the membrane protein. In agreement with experimental observations, highest relative stability was found for the β-barrel region that is embedded in the lipophilic phase, whereas an extracellular protruding β-sheet, which is a unique structural feature of OmpX that supposedly plays an important role in cell adhesion and invasion, shows larger structure fluctuations. Additionally, we investigated water permeation into the core of the β-barrel protein, which contains a tight salt-bridge and hydrogen-bond network, so that extensive water flux is unlikely. Differences between the bilayer and the micellar system were observed in the length of the barrel and its position inside the lipid environment, and in the protein interactions with the hydrophilic part of the lipids near the lipid/water interface. Those variations suggest that micelles and other detergent environments might not offer a wholly membrane-like milieu to promote adoption of the physiological conformational state by OmpX.

Isotopically Labelled and Unlabelled β‐Peptides with Geminal Dimethyl Substitution in 2‐Position of Each Residue: Synthesis and NMR Investigation in Solution and in the Solid State

The preparation of (S)-β2,2,3-amino acids with two Me groups in the α-position and the side chains of Ala, Val, and Leu in the β-position (double methylation of Boc-β-HAla-OMe, Boc-β-Val-OMe, and Boc-β-Leu-OMe, Scheme 2) is described. These β-amino acids and unlabelled as well as specifically 13C- and 15N-labelled 2,2-dimethyl-3-amino acid (β2,2-HAib) derivatives have been coupled in solution (Schemes 1, 3 and 4) to give protected (N-Boc, C-OMe), partially protected (N-Boc/C-OH, N-H/C-OMe), and unprotected β2,2- and β2,2,3-hexapeptides, and β2,2- and β2,2,3-heptapeptides 1–7. NMR Analyses in solution (Tables 1 and 2, and Figs. 2–4) and in the solid state (2D-MAS NMR measurements of the fully labelled Boc-(β2,2-HAib)6-OMe ([13C30, 15N6]-1e; Fig. 5), and TEDOR/REDOR NMR investigations of mixtures (Fig. 6) of the unlabelled Ac-(β2,2-HAib)7-OMe (4) and of a labelled derivative ([13C4,15N2]-5; Figs. 7–11, and 19), a molecular-modeling study (Figs. 13–15), and a search in the Cambridge Crystallographic Data Base (Fig. 16) allow the following conclusions: i) there is no evidence for folding (helix or turn) or for aggregation to sheets of the geminally dimethyl substituted peptide chains in solution; ii) there are distinct conformational preferences of the individual β2,2- and β2,2,3-amino acid residues: close to eclipsing around the C(O)C(Me2(CHR)) bond (τ1,2), almost perfect staggering around the C(2)C(3) ethane bond (τ2,3), and antiperiplanar arrangement of H(C3) and H(N) (τ3,N; Fig. 12) in the solid state; iii) the β2,2-peptides may be part of a turn structure with a ten-membered H-bonded ring; iv) the main structure present in the solid state of F3CCO(β2,2-HAib)7-OMe is a nonfolded chain (>30 A between the termini and >20 A between the N-terminus and the CH2 group of residue 5) with all CO bonds in a parallel alignment (±10°). With these structural parameters, a simple modelling was performed producing three (maybe four) possible chain geometries: one fully extended, two with parallel peptide planes (with zick-zack and crankshaft-type arrangement of the peptide bonds), and (possibly) a fourth with meander-like winding (D–G in Figs. 17 and 18).

Molecular dynamics simulations and free energy calculations on the enzyme 4-hydroxyphenylpyruvate dioxygenase

4‐Hydroxyphenylpyruvate dioxygenase is a relevant target in both pharmaceutical and agricultural research. We report on molecular dynamics simulations and free energy calculations on this enzyme, in complex with 12 inhibitors for which experimental affinities were determined. We applied the thermodynamic integration approach and the more efficient one‐step perturbation. Even though simulations seem well converged and both methods show excellent agreement between them, the correlation with the experimental values remains poor. We investigate the effect of slight modifications on the charge distribution of these highly conjugated systems and find that accurate models can be obtained when using improved force field parameters. This study gives insight into the applicability of free energy methods and current limitations in force field parameterization.