Computer simulation structures , experiments

P. H0YRUP, K. Jorgensen, and O.G. Mouritsen. Nano-scale structure in membranes in relation to enzyme action - computer simulation vs. experiment. Comp. Phys. Commun., 2002, 147, 313— 320. 

While the long-standing picture of the solvent as a structureless medium has been very useful for offering a qualitative understanding of the solvent s effect on structure and dynamics, computer simulations and experiments clearly suggest that a microscopic molecular description of the solvent is necessary. This is particularly so for interfacial phenomena, because the interfacial region itself is only a few molecular diameters thick. 

We carry out computer simulations in the hope of understanding bulk, macroscopic properties in temis of the microscopic details of molecular structure and interactions. This serves as a complement to conventional experiments, enabling us to leam something new something that cannot be found out in other ways. 

To enable an atomic interpretation of the AFM experiments, we have developed a molecular dynamics technique to simulate these experiments [49], Prom such force simulations rupture models at atomic resolution were derived and checked by comparisons of the computed rupture forces with the experimental ones. In order to facilitate such checks, the simulations have been set up to resemble the AFM experiment in as many details as possible (Fig. 4, bottom) the protein-ligand complex was simulated in atomic detail starting from the crystal structure, water solvent was included within the simulation system to account for solvation effects, the protein was held in place by keeping its center of mass fixed (so that internal motions were not hindered), the cantilever was simulated by use of a harmonic spring potential and, finally, the simulated cantilever was connected to the particular atom of the ligand, to which in the AFM experiment the linker molecule was connected. 

The comparison with experiment can be made at several levels. The first, and most common, is in the comparison of derived quantities that are not directly measurable, for example, a set of average crystal coordinates or a diffusion constant. A comparison at this level is convenient in that the quantities involved describe directly the structure and dynamics of the system. However, the obtainment of these quantities, from experiment and/or simulation, may require approximation and model-dependent data analysis. For example, to obtain experimentally a set of average crystallographic coordinates, a physical model to interpret an electron density map must be imposed. To avoid these problems the comparison can be made at the level of the measured quantities themselves, such as diffraction intensities or dynamic structure factors. A comparison at this level still involves some approximation. For example, background corrections have to made in the experimental data reduction. However, fewer approximations are necessary for the structure and dynamics of the sample itself, and comparison with experiment is normally more direct. This approach requires a little more work on the part of the computer simulation team, because methods for calculating experimental intensities from simulation configurations must be developed. The comparisons made here are of experimentally measurable quantities. 

Colloidal crystals . At the end of Section 2.1.4, there is a brief account of regular, crystal-like structures formed spontaneously by two differently sized populations of hard (polymeric) spheres, typically near 0.5 nm in diameter, depositing out of a colloidal solution. Binary superlattices of composition AB2 and ABn are found. Experiment has allowed phase diagrams to be constructed, showing the crystal structures formed for a fixed radius ratio of the two populations but for variable volume fractions in solution of the two populations, and a computer simulation (Eldridge et al. 1995) has been used to examine how nearly theory and experiment match up. The agreement is not bad, but there are some unexpected differences from which lessons were learned. 

To illustrate the relationship between the microscopic structure and experimentally accessible information, we compute pseudo-experimental solvation-force curves F h)/R [see Eq. (22)] as they would be determined in SEA experiments from computer-simulation data for T z [see Eqs. (93), (94), (97)]. Numerical values indicated by an asterisk are given in the customary dimensionless (i.e., reduced) units (see [33,75,78] for definitions in various model systems). Results are correlated with the microscopic structure of a thin film confined between plane parallel substrates separated by a distance = h. Here the focus is specifically on a simple fluid in which the interaction between a pair of film molecules is governed by the Lennard-Jones (12,6) potential [33,58,59,77,79-84]. A confined simple fluid serves as a suitable model for approximately spherical OMCTS molecules confined... 

In this situation computer simulation is useful, since the conditions of the simulation can be chosen such that full equihbrium is established, and one can test the theoretical concepts more stringently than by experiment. Also, it is possible to deal with ideal and perfectly flat surfaces, very suitable for testing the general mechanisms alluded to above, and to disregard in a first step all the complications that real substrate surfaces have (corrugation on the atomistic scale, roughness on the mesoscopic scale, surface steps, adsorbed impurities, etc.). Of course, it may be desirable to add such complications at a later stage, but this will not be considered here. In fact, computer simulations, i.e., molecular dynamics (MD) and Monte Carlo (MC) calculations, have been extensively used to study both static and dynamic properties [11] in particular, structural properties at interfaces have been considered in detail [12]. 

Besides crystalline order and structure, the chain conformation and segment orientation of polymer molecules in the vicinity of the surface are also expected to be modified due to the specific interaction and boundary condition at the surface between polymers and air (Fig. 1 a). According to detailed computer simulations [127, 128], the chain conformation at the free polymer surface is disturbed over a distance corresponding approximately to the radius of gyration of one chain. The chain segments in the outermost layers are expected to be oriented parallel to the surface and chain ends will be enriched at the surface. Experiments on the chain conformation in this region are not available, but might be feasible with evanescent wave techniques described previously. Surface structure on a micrometer scale is observed with IR-ATR techniques [129],... 

Enzyme reactions, like all chemical events, are dynamic. Information coming to us from experiments is not dynamic even though the intervals of time separating observations may be quite small. In addition, much information is denied to us because of technological limitations in the detection of chemical changes. Our models would be improved if we could observe and record all concentrations at very small intervals of time. One approach to this information lies in the creation of a model in which we know all of the concentrations at any time and know something of the structural attributes of each ingredient. A class of models based on computer simulations, such as molecular dynamics, Monte Carlo simulations, and cellular automata, offer such a possibility. 

Levy (Chapter 6) has also explored the use of supercomputers to study detailed properties of biological macromolecule that are only Indirectly accessible to experiment, with particular emphasis on solvent effects and on the Interplay between computer simulations and experimental techniques such as NMR, X-ray structures, and vltratlonal spectra. The chapter by Jorgensen (Chapter 12) summarizes recent work on the kinetics of simple reactions In solutions. This kind of calculation provides examples of how simulations can address questions that are hard to address experimentally. For example Jorgensen s simulations predicted the existence of an Intermediate for the reaction of chloride Ion with methyl chloride In DMF which had not been anticipated experimentally, and they Indicate that the weaker solvation of the transition state as compared to reactants for this reaction In aqueous solution Is not due to a decrease In the number of hydrogen bonds, but rather due to a weakening of the hydrogen bonds. 

The use of computer simulations to study internal motions and thermodynamic properties is receiving increased attention. One important use of the method is to provide a more fundamental understanding of the molecular information contained in various kinds of experiments on these complex systems. In the first part of this paper we review recent work in our laboratory concerned with the use of computer simulations for the interpretation of experimental probes of molecular structure and dynamics of proteins and nucleic acids. The interplay between computer simulations and three experimental techniques is emphasized (1) nuclear magnetic resonance relaxation spectroscopy, (2) refinement of macro-molecular x-ray structures, and (3) vibrational spectroscopy. The treatment of solvent effects in biopolymer simulations is a difficult problem. It is not possible to study systematically the effect of solvent conditions, e.g. added salt concentration, on biopolymer properties by means of simulations alone. In the last part of the paper we review a more analytical approach we have developed to study polyelectrolyte properties of solvated biopolymers. The results are compared with computer simulations. 

Additional insights into the dynamics and structure of bimodal elastomers have been obtained by dynamic light-scattering experiments [129], neutron scattering experiments [130] and calculations [131], dual cross-linking system experiments [132], non-affine swelling [133], and the computer simulations already mentioned. 

Fig. 15. The KATAPAK-S structure and its computational grid. Reprinted from Chemical Engineering Science, Vol. 56, van Baten et al., Radial and Axial Dispersion of the Liquid Phase within a KATAPAK-S Structure Experiments vs. CFD Simulations, pp. 813-821, Copyright (2001), with permission from Elsevier.

Models that are too complicated for the analytical methods of statistical mechanics can often be explored by computer simulations. In a certain sense these are a theoretician s experiment One can devise a model for a certain system, and then investigate with the aid of the computer its consequences. By varying the system parameters, or modifying features of the model, one can gain insight into the structure or dynamics of the system, which one could not obtain by other means. While computer simulations are not as intellectually satisfying as explicit calculations, they are often the only way to test a model. Sometimes they are also used to check the validity of approximations made in analytical calculations. 

An increase in theoretical activity. Any exact science requires a healthy balance between experiment and theory. The wealth of structural information that has become available requires models and theories for its interpretation, and has thus spurred increased activity both in proper theory and in computer simulations. 

The molecular models can be used in computational simulations of functional mechanisms to generate and/or probe specific mechanistic hypotheses for structural changes involved in the various states of the receptors, both wild-type and mutant constructs. The structural context makes these hypotheses testable in collaborative experiments designed to probe specific predictions and refine functional insights (for comprehensive review, see ref. [5]). 

The microscopic structure of water at the solution/metal interface has been the focus of a large body of literature, and excellent reviews have been published summarizing the extensive knowledge gained from experiments, statistical mechanical theories of varied sophistication, and Monte Carlo and molecular dynamics computer simulations. To keep this chapter to a reasonable size, we limit ourselves to a brief summary of the main results and to a sample of the type of information that can be gained from computer simulations. 

EARN distributions the yield along the azimuth 4>= —30° was preferentially reduced with respect to = + 30°. In agreement with intuition, the calculations confirm that the oxygen atom resides in the C-site. Thus from the cooperation of EARN experiments and computer simulations the coverage and nature of the adsorption site of 0/Rh(l 11) has been determined. It will be of interest to see if other surface structure techniques can be used to confirm these s Kcific surface structures. 

In order to learn more about the Rouse model and its limits a detailed quantitative comparison was recently performed of molecular dynamics (MD) computer simulations on a 100 C-atom PE chain with NSE experiments on PE chains of similar molecular weight [52]. Both the experiment and the simulation were carried out at T=509 K. Simulations were imdertaken,both for an explicit (EA) as well as for an united (l/A) atom model. In the latter the H-atoms are not explicitly taken into account but reinserted when calculating the dynamic structure factor. The potential parameters for the MD-simulation were either based on quantum chemical calculations or taken from literature. No adjusting... 

---