A 5-nanosecond molecular dynamics trajectory for B-DNA: analysis of structure, motions, and solvation.

We report the results of four new molecular dynamics (MD) simulations on the DNA duplex of sequence d(CGCGAATTCGCG)2, including explicit consideration of solvent water, and a sufficient number of Na+ counterions to provide electroneutrality to the system. Our simulations are configured particularly to characterize the latest MD models of DNA, and to provide a basis for examining the sensitivity of MD results to the treatment of boundary conditions, electrostatics, initial placement of solvent, and run lengths. The trajectories employ the AMBER 4.1 force field. The simulations use particle mesh Ewald summation for boundary conditions, and range in length from 500 ps to 5.0 ns. Analysis of the results is carried out by means of time series for conformationalm, helicoidal parameters, newly developed indices of DNA axis bending, and groove widths. The results support a dynamically stable model of B-DNA for d(CGCGAATTCGCG)2 over the entire length of the trajectory. The MD results are compared with corresponding crystallographic and NMR studies on the d(CGCGAATTCGCG)2 duplex, and placed in the context of observed behavior of B-DNA by comparisons with the complete crystallographic data base of B-form structures. The calculated distributions of mobile solvent molecules, both water and counterions, are displayed. The calculated solvent structure of the primary solvation shell is compared with the location of ordered solvent positions in the corresponding crystal structure. The results indicate that ordered solvent positions in crystals are roughly twice as structured as bulk water. Detailed analysis of the solvent dynamics reveals evidence of the incorporation of ions in the primary solvation of the minor groove B-form DNA. The idea of localized complexation of otherwise mobile counterions in electronegative pockets in the grooves of DNA helices introduces an additional source of sequence-dependent effects on local conformational, helicoidal, and morphological structure, and may have important implications for understanding the functional energetics and specificity of the interactions of DNA and RNA with regulatory proteins, pharmaceutical agents, and other ligands.

Analysis of local helix bending in crystal structures of DNA oligonucleotides and DNA-protein complexes.

Sequence-dependent bending of the helical axes in 112 oligonucleotide duplex crystal structures resident in the Nucleic Acid Database have been analyzed and compared with the use of bending dials, a computer graphics tool. Our analysis includes structures of both A and B forms of DNA and considers both uncomplexed forms of the double helix as well as those bound to drugs and proteins. The patterns in bending preferences in the crystal structures are analyzed by base pair steps, and emerging trends are noted. Analysis of the 66 B-form structures in the Nucleic Acid Database indicates that uniform trends within all pyrimidine-purine and purine-pyrimidine steps are not necessarily observed but are found particularly at CG and GC steps of dodecamers. The results support the idea that AA steps are relatively straight and that larger roll bends occur at or near the junctions of these A-tracts with their flanking sequences. The data on 16 available crystal structures of protein-DNA complexes indicate that the majority of the DNA bends induced via protein binding are sharp localized kinks. The analysis of the 30 available A-form DNA structures indicates that these structures are also bent and show a definitive preference for bending into the deep major groove over the shallow minor groove.

Molecular dynamics studies of the human CD4 protein.

A dynamical model for an N-terminal fragment of the human CD4 protein has been determined by computer simulation. The protein has been studied both in vacuo and in solution. Data from both simulations agree moderately well with each other and with the crystal structure. All elements of secondary structure were retained during simulation. Point mutation and sequence replacement studies have shown that a loop in CD4, residues 40-52 is involved in binding with gp120, the human immunodeficiency virus surface glycoprotein. Our results show that the gp120-binding loop and a few regions which bind to monoclonal antibodies and class II MHC molecules are the most highly motile areas of the protein. These results are consistent with the suggestion that CD4 binds to target molecules by using induced-fit contacts.

Differential stability of beta-sheets and alpha-helices in beta-lactamase: a high temperature molecular dynamics study of unfolding intermediates.

beta-Lactamase, which catalyzes beta-lactam antibiotics, is prototypical of large alpha/beta proteins with a scaffolding formed by strong noncovalent interactions. Experimentally, the enzyme is well characterized, and intermediates that are slightly less compact and having nearly the same content of secondary structure have been identified in the folding pathway. In the present study, high temperature molecular dynamics simulations have been carried out on the native enzyme in solution. Analysis of these results in terms of root mean square fluctuations in cartesian and [phi, psi] space, backbone dihedral angles and secondary structural hydrogen bonds forms the basis for an investigation of the topology of partially unfolded states of beta-lactamase. A differential stability has been observed for alpha-helices and beta-sheets upon thermal denaturation to putative unfolding intermediates. These observations contribute to an understanding of the folding/unfolding processes of beta-lactamases in particular, and other alpha/beta proteins in general.

Molecular dynamics simulations of poly(dA).poly(dT): comparisons between implicit and explicit solvent representations.

The program AMBER 3.0 has been used to generate molecular dynamics trajectories of a poly(dA).poly(dT) decamer. The simulations were performed using different methods to treat solvent effects. Results of a simulation including 18 counterions NH4+ and 4109 water molecules under (N, P, T) conditions were compared to simulation runs with implicit solvent representation in which solvent screening effects were represented by the use of a sigmoidal distance-dependent dielectric function. In the latter case, the system was simulated under microcanonical (N, V, E) and canonical (N, V, T) conditions. For the fully hydrated system simulation, a preequilibration protocol was developed since it was observed that long and progressive periods of heating and equilibration on the overall system were necessary in order to avoid energetic collisions between the solute and the solvent molecules, leading to severe irreversible deformation of the solute. A detailed analysis of DNA conformations, sugar puckers, and stability of the hydrogen bonds, Watson-Crick and three-center H bonds, is reported. The results show that DNA remains essentially in the B conformer with a tendency in the hydrated model to adopt a slightly distorted, unwound, and stretched conformation in comparison to standard B-DNA. Concerning sugar puckers, the mean pseudorotation phases of the adenine residues are systematically higher than those of the thymine residues, except in the case of the hydrated model for which a articular behavior is observed for the adenine strand. In this case, the terminal bases oscillate between C2'-endo and O4'-endo and the central ones stay in the C3'-endo domain. The mean lifetimes of the internal Watson-Crick H-bond (A) HN6...O4(T) are also dependent on the base pairs included in the calculation, excepted for the implicit solvent simulation at constant temperature. The three-center H bonds have very small mean lifetimes in all three cases of MD simulation. In the minor groove of the hydrated model, a spine of hydration is found as observed by x-ray crystallography and other theoretical simulations. On the basis of the rms deviations, it appears that the fully hydrated simulation has not reached a plateau at the end of the run, while the implicit simulation at constant energy seems to have converged. At constant temperature, very large oscillations in rms deviations are observed.

Persistence analysis of the static and dynamical helix deformations of DNA oligonucleotides: application to the crystal structure and molecular dynamics simulation of d(CGCGAATTCGCG)2.

A theory and graphical presentation for the analysis of helix structure and deformations in oligonucleotides is presented. The parameters "persistence" and "flexibility" as defined in the configurational statistics of polymers of infinite length are reformulated at the oligonucleotide level in an extension of J. A. Schellman's method [(1974) Biopolymers, Vol. 17, pp. 217-226], and used as a basis for a systematic "Persistence Analysis" of the helix deformation properties for all possible subsequences in the structure. The basis for the analysis is a set of link vectors referenced to individual base pairs, and is limited to sequences exhibiting only perturbed rod-like behavior, i.e., below the threshold for supercoiling. The present application of the method is concerned with a physical model for the angular component of bending, so the link vectors are defined as the unit components of a global helix axis obtained by the procedure "Curves" of R. Lavery and H. Sklenar [(1988) J. Biomol. Struct. Dynam., Vol. 6, pp. 63-91; (1989) ibid., Vol. 6, pp. 655-667]. A discussion of the relationship between global bending and relative orientation of base pairs is provided. Our approach is illustrated by analysis of some model oligonucleotide structures with intrinsic kinks, the crystal structure of the dodecamer d(CGCGAATTCGCG)2, and the results of two molecular dynamics simulations on this dodecamer using two variations of the GROMOS force field. The results indicate that essentially all aspects of curvature in short oligonucleotides can be determined, such as the position and orientation of each bend, the sharpness or smoothness, and the location and linearity of subsequences. In the case of molecular dynamics simulations, where a Boltzmann ensemble of structures is analyzed, the spatial extent of the deformations (flexibility) is also considered.

Conformational and helicoidal analysis of the molecular dynamics of proteins: "curves," dials and windows for a 50 psec dynamic trajectory of BPTI.

A new procedure for the graphic analysis of molecular dynamics (MD) simulations on proteins is introduced, in which comprehensive visualization of results and pattern recognition is greatly facilitated. The method involves determining the conformational and helicoidal parameters for each structure entering the analysis via the method "Curves," developed for proteins by Sklenar, Etchebest, and Lavery (Proteins: Structure, Function Genet. 6:46-60, 1989) followed by a novel computer graphic display of the results. The graphic display is organized systematically using conformation wheels ("dials") for each torsional parameter and "windows" on the range values assumed by the linear and angular helicoidal parameters, and is present in a form isomorphous with the primary structure per se. The complete time evolution of dynamic structure can then be depicted in a set of four composite figures. Dynamic aspects of secondary and tertiary structure are also provided. The procedure is illustrated with an analysis of a 50 psec in vacuo simulation on the 58 residue protein, bovine pancreatic trypsin inhibitor (BPTI), in the vicinity of the local minimum on the energy surface corresponding to a high resolution crystal structure. The time evolution of 272 conformational and 788 helicoidal parameters for BPTI is analyzed. A number of interesting features can be discerned in the analysis, including the dynamic range of conformational and helicoidal motions, the dynamic extent of 2 degrees structure motifs, and the calculated fluctuations in the helix axis. This approach is expected to be useful for a critical analysis of the effects of various assumptions about force field parameters, truncation of potentials, solvation, and electrostatic effects, and can thus contribute to the development of more reliable simulation protocols for proteins. Extensions of the analysis to present differential changes in conformational and helicoidal parameters is expected to be valuable in MD studies of protein complexes with substrates, inhibitors, and effectors and in determining the nature of structural changes in protein-protein interactions.

Conformational and helicoidal analysis of 30 PS of molecular dynamics on the d(CGCGAATTCGCG) double helix: "curves", dials and windows.

A new procedure for the analysis of the structure and molecular dynamics of duplex DNA is introduced, in which comprehensive visualization of results and pattern recognition is greatly facilitated. The method involves determining the values of the conformational and helicoidal parameters for each structure entering the analysis using the method "Curves" developed by Lavery and Sklenar, J. Biomol. Str. Dyn. 6, 63 (1988), followed by a novel computer graphic display of the results. The graphic display is organized systematically using conformation wheels, or "dials", for each IUPAC torsional parameter and "windows" on the range of values assumed by the linear and angular helicoidal parameters, and is presented in a form isomorphous with the structure per se. The complete time evolution of the conformational and helicoidal parameters of a DNA double helix can then be depicted in a set of six composite figures. Dynamical aspects of helix bending are also subsumed in this analysis. The procedure is illustrated with an analysis of the structures of canonical A and B forms of DNA and the 300 degrees K native dodecamer duplex d(CGCGAATTCGCG). The "dials and windows" are then used for a comprehensive analysis of 30 psec of molecular dynamics on the dodecamer in the vicinity of a canonical B-DNA energy minimum. This involves presentation of the time evolution of 206 conformational and 230 helicoidal parameters for the dodecamer. A number of interesting structural features can be recognized in the analysis, including crankshaft motions, BI - BII transitions, sugar repuckerings, and a description of spontaneous helix bending at what corresponds to the 1 degrees and 2 degrees "hinge points" indicated in the crystal structure. Our approach is expected to be directly useful for critical analysis of the effects of various assumptions about force field parameters, hydration and electrostatic effects and thus contribute to the development of reliable simulation protocols for nucleic acid systems. Extension of the method to present differential changes in conformational and helicoidal parameters is expected to be valuable for the analysis of structural and molecular dynamics studies of the reorganization and adaptation of DNA on complexation with various drugs and regulatory proteins.

Theoretical considerations on the "spine of hydration" in the minor groove of d(CGCGAATTCGCG).d(GCGCTTAAGCGC): Monte Carlo computer simulation.

A theoretical description of aqueous hydration in the minor groove of a B-form DNA is presented on the basis of a liquid-state Monte Carlo computer simulation on a system consisting of the oligonucleotide duplex d(CGCGAATTCGCG).d(GCGCTTAAGCGC) in a canonical B-form together with 1777 water molecules contained in a hexagonal prism cell and treated under periodic boundary conditions. The results are analyzed in terms of solvent density distributions. The calculated minor-groove solvent density shows considerable localization, indicative of discrete solvation sites and providing theoretical evidence for a well-defined ordered water structure. In the AATT sequence, this corresponds to the "spine of hydration" described by H. R. Drew and R. E. Dickerson [(1981) J. Mol. Biol. 151, 535-556] based on the x-ray crystal structure of the dodecamer hydrate. We find, however, that the calculated ordered water structure also extends into the CGCG flanking sequences, supported by the N2 hydrogen bond donors of the guanine residues and indicating that the spine of hydration could thus extend throughout the minor groove of a B-form DNA. This provides a possible explanation of the positive binding entropies observed by L. A. Marky and K. J. Breslauer [(1984) Proc. Natl. Acad. Sci. USA 84, 4359-4363] for both A.T and C.G sequences on the complexation of netropsin to the minor groove of DNAs. Implications of these results with regard to the thermodynamic stability of DNA in water and the sequence specificity of the minor groove hydration are discussed.

Aqueous hydration of nucleic acid constituents: Monte Carlo computer simulation studies.

Monte Carlo computer simulations were performed on dilute aqueous solutions of thymine, cytosine, uracil, adenine, guanine, the dimethyl phosphate anion in the gauche-gauche conformation and a ribose and deoxyribose derivative. The aqueous hydration of each molecule was analysed in terms of quasi-component distribution functions based on the Proximity Criterion, and partitioned into hydrophobic, hydrophilic and ionic contributions. Color stereo views of selected hydration complexes are also presented. A preliminary discussion of the transferability of functional group coordination numbers is given. The results enable to comment on two current problems related to the hydration of nucleic acids: a) the theory of Dickerson and coworkers on the role of water in the relative stability of the A and B form of DNA and b) the idea of water bridges and filaments emerging from the computer simulation results on the hydration of DNA fragments by Clementi.