An evaluation of implicit and explicit solvent model systems for the molecular dynamics simulation of bacteriophage T4 lysozyme

Machine learning/molecular dynamic protein structure prediction approach to investigate the protein conformational ensemble

Proteins exist in several different conformations. These structural changes are often associated with fluctuations at the residue level. Recent findings show that co-evolutionary analysis coupled with machine-learning techniques improves the precision by providing quantitative distance predictions between pairs of residues. The predicted statistical distance distribution from Multi Sequence Analysis reveals the presence of different local maxima suggesting the flexibility of key residue pairs. Here we investigate the ability of the residue-residue distance prediction to provide insights into the protein conformational ensemble. We combine deep learning approaches with mechanistic modeling to a set of proteins that experimentally showed conformational changes. The predicted protein models were filtered based on energy scores, RMSD clustering, and the centroids selected as the lowest energy structure per cluster. These models were compared to the experimental-Molecular Dynamics (MD) relaxed structure by analyzing the backbone residue torsional distribution and the sidechain orientations. Our pipeline allows to retrieve the experimental structural dynamics experimentally represented by different X-ray conformations for the same sequence as well the conformational space observed with the MD simulations. We show the potential correlation between the experimental structure dynamics and the predicted model ensemble demonstrating the susceptibility of the current state-of-the-art methods in protein folding and dynamics prediction and pointing out the areas of improvement.

Reproducing the Ensemble Average Polar Solvation Energy of a Protein from a Single Structure: Gaussian-Based Smooth Dielectric Function for Macromolecular Modeling

Typically, the ensemble average polar component of solvation energy (ΔG_polar^solv) of a macromolecule is computed using molecular dynamics (MD) or Monte Carlo (MC) simulations to generate conformational ensemble and then single/rigid conformation solvation energy calculation is performed on each snapshot. The primary objective of this work is to demonstrate that Poisson-Boltzmann (PB)-based approach using a Gaussian-based smooth dielectric function for macromolecular modeling previously developed by us (Li et al. J. Chem. Theory Comput. 2013, 9 (4), 2126-2136) can reproduce that ensemble average (ΔG_polar^solv) of a protein from a single structure. We show that the Gaussian-based dielectric model reproduces the ensemble average ΔG_polar^solv(⟨ΔG_polar^solv⟩) from an energy-minimized structure of a protein regardless of the minimization environment (structure minimized in vacuo, implicit or explicit waters, or crystal structure); the best case, however, is when it is paired with an in vacuo-minimized structure. In other minimization environments (implicit or explicit waters or crystal structure), the traditional two-dielectric model can still be selected with which the model produces correct solvation energies. Our observations from this work reflect how the ability to appropriately mimic the motion of residues, especially the salt bridge residues, influences a dielectric model's ability to reproduce the ensemble average value of polar solvation free energy from a single in vacuo-minimized structure.

The "Hot-Solvent/Cold-Solute" Problem Revisited

The temperature steers the equilibrium and nonequilibrium conformational dynamics of macromolecules in solution. Therefore, corresponding molecular dynamics simulations require a strategy for temperature control which should guarantee that the experimental statistical ensemble is also sampled in silico. Several algorithms for temperature control have been proposed. All these thermostats interfere with the macromolecule's "natural" dynamics as given by the Newtonian mechanics. Furthermore, using a single thermostat for an inhomogeneous solute-solvent system can lead to stationary temperature gradients. To avoid this "hot solvent/cold solute" problem, two separate thermostats are frequently applied, one to the solute and one to the solvent. However, such a separate temperature control will perturb the dynamics of the macromolecule much more strongly than a global one and, therefore, can introduce large artifacts into its conformational dynamics. Based on the concept that an explicit solvent environment represents an ideal thermostat concerning the magnitude and time correlation of temperature fluctuations of the solute, we propose a temperature control strategy that, on the one hand, provides a homogeneous temperature distribution throughout the system together with the correct statistical ensemble for the solute molecule while, on the other hand, minimally perturbing its dynamics.

Probing the folded state and mechanical unfolding pathways of T4 lysozyme using all-atom and coarse-grained molecular simulation

The Bacteriophage T4 Lysozyme (T4L) is a prototype modular protein comprised of an N-terminal and a C-domain domain, which was extensively studied to understand the folding/unfolding mechanism of modular proteins. To offer detailed structural and dynamic insights to the folded-state stability and the mechanical unfolding behaviors of T4L, we have performed extensive equilibrium and steered molecular dynamics simulations of both the wild-type (WT) and a circular permutation (CP) variant of T4L using all-atom and coarse-grained force fields. Our all-atom and coarse-grained simulations of the folded state have consistently found greater stability of the C-domain than the N-domain in isolation, which is in agreement with past thermostatic studies of T4L. While the all-atom simulation cannot fully explain the mechanical unfolding behaviors of the WT and the CP variant observed in an optical tweezers study, the coarse-grained simulations based on the Go model or a modified elastic network model (mENM) are in qualitative agreement with the experimental finding of greater unfolding cooperativity in the WT than the CP variant. Interestingly, the two coarse-grained models predict different structural mechanisms for the observed change in cooperativity between the WT and the CP variant--while the Go model predicts minor modification of the unfolding pathways by circular permutation (i.e., preserving the general order that the N-domain unfolds before the C-domain), the mENM predicts a dramatic change in unfolding pathways (e.g., different order of N/C-domain unfolding in the WT and the CP variant). Based on our simulations, we have analyzed the limitations of and the key differences between these models and offered testable predictions for future experiments to resolve the structural mechanism for cooperative folding/unfolding of T4L.

Direct observation of T4 lysozyme hinge-bending motion by fluorescence correlation spectroscopy

Bacteriophage T4 Lysozyme (T4L) catalyzes the hydrolysis of the peptidoglycan layer of the bacterial cell wall late in the infection cycle. It has long been postulated that equilibrium dynamics enable substrate access to the active site located at the interface between the N- and C-terminal domains. Crystal structures of WT-T4L and point mutants captured a range of conformations that differ by the hinge-bending angle between the two domains. Evidence of equilibrium between open and closed conformations in solution was gleaned from distance measurements between the two domains but the nature of the equilibrium and the timescale of the underlying motion have not been investigated. Here, we used fluorescence fluctuation spectroscopy to directly detect T4L equilibrium conformational fluctuations in solution. For this purpose, Tetramethylrhodamine probes were introduced at pairs of cysteines in regions of the molecule that undergo relative displacement upon transition from open to closed conformations. Correlation analysis of Tetramethylrhodamine intensity fluctuations reveals hinge-bending motion that changes the relative distance and orientation of the N- and C-terminal domains with ≅ 15 μs relaxation time. That this motion involves interconversion between open and closed conformations was further confirmed by the dampening of its amplitude upon covalent substrate trapping. In contrast to the prevalent two-state model of T4L equilibrium, molecular brightness and number of particles obtained from cumulant analysis suggest that T4L populates multiple intermediate states, consistent with the wide range of hinge-bending angles trapped in the crystal structure of T4L mutants.

SRLS analysis of 15N relaxation from bacteriophage T4 lysozyme: a tensorial perspective that features domain motion

Bacteriophage T4L lysozyme (T4L) comprises two domains connected by a helical linker. Several methods detected ns domain motion associated with the binding of the peptidoglycan substrate. An ESR study of nitroxide-labeled T4L, based on the slowly relaxing local structure (SRLS) approach, detected ns local motion involving the nitroxide and the helix housing it. (15)N−H spin relaxation data from T4L acquired at magnetic fields of 11.7 and 18.8 T, and 298 K, were analyzed previously with the model-free (MF) method. The results did not detect domain motion. SRLS is the generalization of MF. Here, we apply it to the same data analyzed previously with MF. The restricted local N−H motion is described in terms of tilted axial local ordering (S) and local diffusion (D(2)) tensors; dynamical coupling to the global tumbling is accounted for. We find that D(2,⊥) is 1.62 × 10(7) (1.56 × 10(7)) s(−1) for the N-terminal (C-terminal) domain. This dynamic mode represents domain motion. For the linker D(2,⊥) is the same as the rate of global tumbling, given by (1.46 ± 0.04) × 10(7) s(−1). D(2,∥) is 1.3 × 10(9), 1.8 × 10(9) and 5.3 × 10(9) s(−1) for the N-terminal domain, the C-terminal domain, and the linker, respectively. This dynamic mode represents N−H bond vector fluctuations. The principal axis of D(2) is virtually parallel to the N−H bond. The order parameter, S(0)(2), is 0.910 ± 0.046 for most N−H bonds. The principal axis of S is tilted from the C(i−1)(α) −C(i)(α) axis by −2° to 6° for the N-, and C-terminal domains, and by 2.5° for the linker. The tensorial-perspective-based and mode-coupling-based SRLS picture provides new insights into the structural dynamics of bacteriophage T4 lysozyme.

A smooth and differentiable bulk-solvent model for macromolecular diffraction

Inclusion of low-resolution data in macromolecular crystallography requires a model for the bulk solvent. Previous methods have used a binary mask to accomplish this, which has proven to be very effective, but the mask is discontinuous at the solute-solvent boundary (i.e. the mask value jumps from zero to one) and is not differentiable with respect to atomic parameters. Here, two algorithms are introduced for computing bulk-solvent models using either a polynomial switch or a smoothly thresholded product of Gaussians, and both models are shown to be efficient and differentiable with respect to atomic coordinates. These alternative bulk-solvent models offer algorithmic improvements, while showing similar agreement of the model with the observed amplitudes relative to the binary model as monitored using R, R(free) and differences between experimental and model phases. As with the standard solvent models, the alternative models improve the agreement primarily with lower resolution (>6 A) data versus no bulk solvent. The models are easily implemented into crystallographic software packages and can be used as a general method for bulk-solvent correction in macromolecular crystallography.

Mapping electron paramagnetic resonance spin label conformations by the simulated scaling method

In order to efficiently simulate spin label behavior when attached to the protein backbone we developed a novel approach that enhances local conformational sampling. The simulated scaling (SS) approach (Li, H., et al. J. Chem. Phys. 2007, 126, 24106) couples the random walk of a potential scaling parameter and molecular dynamics in the framework of hybrid Monte Carlo. This approach allows efficient barrier crossings between conformations. The method retains the thermodynamic detailed balance allowing for determination of relative free energies between various conformations. The accuracy of our method was validated by comparison with the recently resolved X-ray crystal structure of a spin labeled T4 lysozyme in which the spin label was in the interior of the protein. Consistent potentials of mean force (PMF) are obtained for the spin label torsion angles to illustrate their behavior in various protein environments: surface, semiburied, and buried. These PMFs reflect the experimentally observed trends and provide the rationale for the spin label dynamics. We have used this method to compare an implicit and explicit solvent model in spin label modeling. The implicit model, which is computationally faster, was found to be in excellent agreement with the explicit solvent treatment. Based on this collection of results, we believe that the presented approach has great potential in the general strategy of describing the behavior of the spin label using molecular modeling and using this information in the interpretation of EPR measurements in terms of protein conformation and dynamics.

De novo design of beta-helical polypeptides

Many proteins, including several growth factor receptors such as the IGF-1R and EGFR family, contain variants of the beta-helix fold. Inspection of the irregular protein beta-helices suggested that different families of regular beta-helical polypeptides can be designed using a series of hinged vectors and the constraints imposed by the geometry of a peptide backbone. We have conceived beta-helices with five and six beta-strands per turn and designed, in detail, a series of regular beta-helices with rhomboidal or triangular cross-sections. Each beta-helix was modeled by threading C(alpha) atoms to follow the vectorial beta-helix and then creating the H-bonded polypeptide backbone and appropriate side-chain orientations. The conformational stability of these regular beta-helices was assessed using molecular dynamics simulations. Several potential repeat amino acid sequences were identified for different geometries of beta-helix. Regular beta-helices offer new possibilities for the study of protein folding, the production of nanofibers, catalysts, inhibitors of growth factor receptors and drug carriers.

Solvent Interactions and Protein Dynamics in Spin-labeled T4 Lysozyme

Aspects of T4 lysozyme dynamics and solvent interaction are investigated using atomically detailed Molecular Dynamics (MD) simulations. Two spin-labeled mutants of T4 lysozyme are analyzed (T4L-N40C and T4L-K48C), which have been found from electronic paramagnetic resonance (EPR) experiments to exhibit different mobilities at the site of spin probe attachment (N- and C-terminus of helix B, respectively). Similarities and differences in solvent distribution and diffusion around the spin label, as well as around exposed and buried residues within the protein, are discussed. The purpose is to capture possible strong interactions between the spin label (ring) and solvent molecules, which may affect EPR lineshapes. The effect of backbone motions on the water density profiles is also investigated. The focus is on the domain closure associated with the T4 lysozyme hinge-bending motion, which is analyzed by Essential Dynamics (ED). The N-terminus of helix B is found to be a "hinge" residue, which explains the high degree of flexibility and motional freedom at this site.

Comparison of various implicit solvent models in molecular dynamics simulations of immunoglobulin G light chain dimer

The present study tests performance of different solvation models applied to molecular dynamics simulation of a large, dimeric protein molecule. Analytical Continuum Electrostatics (ACE) with two different parameter sets, older V98 and new V01, and Effective Energy Function (EEF) are employed in molecular dynamics simulation of immunoglobulin G (IgG) light chain dimer and variable domain of IgG light chain. Results are compared with explicit solvent and distance dependent dielectric constant (DDE) calculations. The overall analysis shows that the EEF method yields results comparable to explicit solvent simulations; however, the stability of simulations is lower. On the other hand, the ACE_V98 model does not seem to achieve the accuracy or stability expected in nanosecond timescale MD simulation for the studied systems. The ACE_V01 model greatly improves stability of the calculation; nonetheless, changes in radius of gyration and solvent accessible surface of the studied systems may indicate that the parameter set still needs to be improved if the method is supposed to be used for simulations of large, polymeric proteins. Additionally, electrostatic contribution to the solvation free energy calculated in the ACE model is compared with a numerical treatment of the dielectric continuum model. Wall clock time of all simulations is compared. It shows that EEF calculation is six times faster than corresponding ACE and 50 times faster than explicit solvent simulations.

Molecular dynamics simulations of peptides and proteins with a continuum electrostatic model based on screened Coulomb potentials

A continuum electrostatics approach for molecular dynamics (MD) simulations of macromolecules is presented and analyzed for its performance on a peptide and a globular protein. The approach incorporates the screened Coulomb potential (SCP) continuum model of electrostatics, which was reported earlier. The model was validated in a broad set of tests some of which were based on Monte Carlo simulations that included single amino acids, peptides, and proteins. The implementation for large-scale MD simulations presented in this article is based on a pairwise potential that makes the electrostatic model suitable for fast analytical calculation of forces. To assess the suitability of the approach, a preliminary validation is conducted, which consists of (i) a 3-ns MD simulation of the immunoglobulin-binding domain of streptococcal protein G, a 56-residue globular protein and (ii) a 3-ns simulation of Dynorphin, a biological peptide of 17 amino acids. In both cases, the results are compared with those obtained from MD simulations using explicit water (EW) molecules in an all-atom representation. The initial structure of Dynorphin was assumed to be an alpha-helix between residues 1 and 9 as suggested from NMR measurements in micelles. The results obtained in the MD simulations show that the helical structure collapses early in the simulation, a behavior observed in the EW simulation and consistent with spectroscopic data that suggest that the peptide may adopt mainly an extended conformation in water. The dynamics of protein G calculated with the SCP implicit solvent model (SCP-ISM) reveals a stable structure that conserves all the elements of secondary structure throughout the entire simulation time. The average structures calculated from the trajectories with the implicit and explicit solvent models had a cRMSD of 1.1 A, whereas each average structure had a cRMSD of about 0.8A with respect to the X-ray structure. The main conformational differences of the average structures with respect to the crystal structure occur in the loop involving residues 8-14. Despite the overall similarity of the simulated dynamics with EW and SCP models, fluctuations of side-chains are larger when the implicit solvent is used, especially in solvent exposed side-chains. The MD simulation of Dynorphin was extended to 40 ns to study its behavior in an aqueous environment. This long simulation showed that the peptide has a tendency to form an alpha-helical structure in water, but the stabilization free energy is too weak, resulting in frequent interconversions between random and helical conformations during the simulation time. The results reported here suggest that the SCP implicit solvent model is adequate to describe electrostatic effects in MD simulation of both peptides and proteins using the same set of parameters. It is suggested that the present approach could form the basis for the development of a reliable and general continuum approach for use in molecular biology, and directions are outlined for attaining this long-term goal.

MASKER: improved solvent-excluded molecular surface area estimations using Boolean masks

A fast algorithm for computing the solvent-accessible molecular surface area (SAS) using Boolean masks [Le Grand,S.M. and Merz,K.M.J. (1993). J. Comput. Chem., 14, 349-352) has been modified to estimate the solvent-excluded molecular surface area (SES), including contact, toroidal and re-entrant surface components. Numerical estimates of arc lengths of intersecting atomic SAS are used to estimate the toroidal surface and intersections between those arcs are used to estimate the re-entrant surface area. The new method is compared with an exact analytical method. Boolean molecular surface areas are continuous and pairwise differentiable and should be useful for molecular dynamics simulations, especially as the basis for an implicit solvent model.

Role of mg2+ in chromomycin a3 - DNA interaction: a molecular modeling study

Chromomycin A(3) (CHR) is an antitumor antibiotic that inhibits macromolecular biosynthesis by reversibly binding to double stranded DNA via the minor groove, with GC-base specificity. At and above physiological pH when CHR is anionic, interaction of CHR with DNA requires the presence of divalent metal ions like Mg(2+). However, at acidic pHthe molecule is neutral and it binds DNA even in absence of Mg(2+). Molecular dynamics simulation studies at 300K of neutral CHR and 1:1 CHR:Mg(2+) complexes formed at pH 5.2 and 8.0 show that hydrophobicity of CHR:Mg(2+) complex formed with the neutral drug is greater than that of the two other species. Interactions of CHR with DNA in presence and absence of Mg(2+) have been studied by simulated annealing to understand the role of Mg(2+) in the DNA binding potential of CHR. This shows that the antibiotic has the structural potential to bind to DNA even in the absence of metal ion. Evaluation of the direct interaction energy between the ligand and DNA does not explain the observed GC-base specificity of the antibiotic. When energy contributions from structural alteration of the interacting ligand and DNA as a sequel to complex formation are taken into account, atrue picture of the theoretical binding propensity emerges. This implies that DNA and/or the ligand undergo significant structural alterations during the process of association, particularly in presence of Mg(2+). Accessible surface area calculations give idea about the entropy contribution to the binding free energy which is found to be different depending upon the presence and absence of Mg(2+).

Effect of neighboring bases on base-pair stacking orientation: a molecular dynamics study

It is generally believed that base-pair stacking interaction in DNA double helix is one of the strongest interactions that governs sequence directed structural variability. However, X-ray crystal structures of some base-paired doublet sequences have been seen to adopt different structures when flanked by different base-pairs. DNA crystal database, however, is still too small to make good statistical inference about effect of such flanking residues. Influence of neighboring residue on the local helical geometry of a base-paired doublet in B-DNA has been investigated here using molecular dynamics simulation. We have generated ensembles of structures for d(CA).d(TG) and d(AA).d(TT) base-paired doublets located at the centers of d(CGCGCAAAGCG).d(CGCTTTGCGCG) and d(CGCGAAAACGCG).d(CGCGTTTTCGCG) sequences along with their analogs by varying the bases either at 5'- or 3'- position to the central doublet. Comparison of base paired doublet parameters for the ensembles of structures show that stacking geometry of d(CA).d(TG) doublet depends on some of the flanking base-pairs. On the other hand d(AA).d(TT) doublet remains nearly unperturbed when the flanking residues are altered.

Study of the electrostatics treatment in molecular dynamics simulations

This article considers the treatment of long-range interactions in molecular dynamics simulations. We investigate the effects of using different cutoff distances, constant versus distance-dependent dielectric, and different smoothing methods. In contrast to findings of earlier studies, we find that increasing the cutoff over 8 A does not significantly improve the accuracy (Arnold and Ornstein, Proteins 1994;18:19-33), and using a distance-dependent dielectric instead of a constant dielectric also does not improve accuracy (Guenot and Kollman, Protein Sci 1992;1:1185-1205). This might depend on differences in simulation protocols or force fields, or both, because we use the CHARMM22 force field with stochastic boundary conditions, whereas earlier studies used other protocols and energy functions. We also note that the stability of the simulations is highly dependent on the starting structure, showing that accurate molecular simulations not only depend on a realistic simulation protocol but also on correct initial conditions.

Molecular dynamics study of time-correlated protein domain motions and molecular flexibility: cytochrome P450BM-3

Time-correlated atomic motions were used to characterize protein domain boundaries from atomic coordinates generated by molecular dynamics simulations. A novel application of the dynamical cross-correlation matrix (DCCM) analysis tool was used to help identify putative protein domains. In implementing this new approach, several DCCM maps were calculated, each using a different coordinate reference frame from which protein domain boundaries and protein domain residue constituents could be identified. Cytochrome P450BM-3, from Bacillus megaterium, was used as the model protein in this study. The analyses indicated that the simulated protein comprises three distinct domain regions; in contrast, only two protein domains were identified in the original crystal structure report. Specifically, the DCCM analyses showed that the F-G helix region was a separate domain entity and not a part of the alpha domain, as previously designated. The simulations demonstrated that the domain motions of the F-G helix region effected both the size and shape of the enzyme active site, and that the dynamics of the F-G helix domain could possibly control access of substrate to the binding pocket.

Three-dimensional modelling of human cytochrome P450 1A2 and its interaction with caffeine and MeIQ

The three-dimensional modelling of proteins is a useful tool to fill the gap between the number of sequenced proteins and the number of experimentally known 3D structures. However, when the degree of homology between the protein and the available 3D templates is low, model building becomes a difficult task and the reliability of the results depends critically on the correctness of the sequence alignment. For this reason, we have undertaken the modelling of human cytochrome P450 1A2 starting by a careful analysis of several sequence alignment strategies (multiple sequence alignments and the TOPITS threading technique). The best results were obtained using TOPITS followed by a manual refinement to avoid unlikely gaps. Because TOPITS uses secondary structure predictions, several methods that are available for this purpose (Levin, Gibrat, DPM, NnPredict, PHD, SOPM and NNSP) have also been evaluated on cytochromes P450 with known 3D structures. More reliable predictions on alpha-helices have been obtained with PHD, which is the method implemented in TOPITS. Thus, a 3D model for human cytochrome P450 1A2 has been built using the known crystal coordinates of P450 BM3 as the template. The model was refined using molecular mechanics computations. The model obtained shows a consistent location of the substrate recognition segments previously postulated for the CYP2 family members. The interaction of caffeine and a carcinogenic aromatic amine (MeIQ), which are characteristic P450 1A2 substrates, has been investigated. The substrates were solvated taking into account their molecular electrostatic potential distributions. The docking of the solvated substrates in the active site of the model was explored with the AUTODOCK programme, followed by molecular mechanics optimisation of the most interesting complexes. Stable complexes were obtained that could explain the oxidation of the considered substrates by cytochrome P450 1A2 and could offer an insight into the role played by water molecules.

Protein hinge bending as seen in molecular dynamics simulations of native and M61 mutant T4 lysozymes

A dynamical model of interdomain "hinge bending" of T4 lysozyme in aqueous solution has been developed on the basis of molecular dynamics (MD) simulation. The MD model study provides a description of the conformational reorganization expected to occur for the protein in aqueous solution as compared to the crystalline environment. Three different 500 ps molecular dynamics simulations were calculated, each using a distinctly different crystal conformation of T4 lysozyme as the starting points of the MD simulations. Crystal structures of wild-type lysozyme and "open" and "closed" forms of M61 variant structures were analyzed in this study. Large-scale, molecular-conformational rearrangements were observed in all three simulations, and the largest structural change was found for the open form of the M61 allomorph. All three simulated proteins had closed relative to the wild-type crystal structure, and the closure of the "jaws" of the active site cleft occurred gradually over the time course of the trajectories. The time average MD structures, calculated over the final 50 ps of each trajectory, had all adapted to conformations more similar to each other than to their incipient crystal forms. Using a similar MD protocol on cytochrome P450BM-3 [M. D. Paulsen and R. L. Ornstein (1995) Proteins: Structure Function and Genetics, Vol. 27, pp. 237-243] we have found that the opposite type of motion relative to the starting crystal structure, that is, the open form of the crystal structure, had opened to a greater degree relative to the incipient crystal structure form. Therefore we do not believe that either result is merely a simulation artifact, but rather the protein dynamics are due to protein relaxation in the absence of crystal packing forces in the simulated solution environments.

Comparison of low oxidoreduction potential cytochrome c553 from Desulfovibrio vulgaris with the class I cytochrome c family

The cytochrome c553 from Desulfovibrio vulgaris (DvH c553) is of importance in the understanding of the relationship of structure and function of cytochrome c due to its lack of sequence homology with other cytochromes, and its abnormally low oxido-reduction potential. In evolutionary terms, this protein also represents an important reference point for the understanding of both bacterial and mitochondrial cytochromes c. Using the recently determined nuclear magnetic resonance (NMR) structure of the reduced protein we compare the structural, dynamic, and functional characteristics of DvH c553 with members of both the mitochondrial and bacterial cytochromes c to characterize the protein in the context of the cytochrome c family, and to understand better the control of oxide-reduction potential in electron transfer proteins. Despite the low sequence homology, striking structural similarities between this protein and representatives of both eukaryotic [cytochrome c from tuna (tuna c)] and prokaryotic [Pseudomonas aeruginosa c551 (Psa c551)] cytochromes c have been recognized. The previously observed helical core is also found in the DvH c553. The structural framework and hydrogen bonding network of the DvH c553 is most similar to that of the tuna c, with the exception of an insertion loop of 24 residues closing the heme pocket and protecting the propionates, which is absent in the DvH c553. In contrast, the Psa c551 protects the propionates from the solvent principally by extending the methionine ligand arm. The electrostatic distribution at the recognized encounter surface around the heme in the mitochondrial cytochrome is reproduced in the DvH c553, and corresponding hydrogen bonding networks, particularly in the vicinity of the heme cleft, exist in both molecules. Thus, although the cytochrome DvH c553 exhibits higher primary sequence homology to other bacterial cytochromes c, the structural and physical homology is significantly greater with respect to the mitochondrial cytochrome c. The major structural and functional difference is the absence of solvent protection for the heme, differentiating this cytochrome from both reference cytochromes, which have evolved different mechanisms to cover the propionates. This suggests that the abnormal redox potential of the DvH c553 is linked to the raised accessibility of the heme and supports the theory that redox potential in cytochromes is controlled by heme propionate solvent accessibility.

Rational approach to improving reductive catalysis by cytochrome P450cam

Although halogenated hydrocarbons are noted for low chemical reactivity, small amounts are toxic to humans. Cytochromes P450 have been implicated in transforming these compounds to more reactive species, under anaerobic conditions, through reduction at the heme. A significant amount of effort has been directed toward turning this catalytic ability to our advantage by engineering P450 variants than can efficiently remediate these compounds in situ, before they come in contact with the human population. We have taken a 'rational' approach to this problem, in which a combination of theory and molecular modeling is applied to identify which properties of the enzyme have the greatest influence over reductive dehalogenation. Recent progress in this area is briefly reviewed. Two novel mutants, incorporating tryptophan (positions 87 and 396) and histidine (position 96, neutral and protonated) amino acid substitutions in the active site, are proposed and evaluated using molecular dynamics simulations. The upper bound on rate enhancement relative to wild-type is estimated in each mutant using electron transfer theory. The most significant rate enhancement is predicted for the His 96 mutant in the protonated state; while some His residues of certain proteins exhibit a pKa high enough to support a large protonated population, such information is not presently available for this proposed mutant.

Molecular dynamics simulations indicate that F87W,T185F-cytochrome P450cam may reductively dehalogenate 1,1,1-trichloroethane

Cytochrome P450cam is capable of reductively dehalogenating several chlorinated alkanes at low, but measurable, rates. In previous investigations of structure-function relationships in this enzyme using molecular dynamics simulations, we noticed that 1,1,1-trichloroethane (TCA) exhibits a very high degree of mobility in the active site due to its smaller molecular volume relative to the native substrate, camphor(1,2). Several amino acid sidechains lining the active site also exhibit significant dynamic fluctuations, possibly as a result of poor steric complementary to TCA. Guided by these results, we modeled double (F87W, T185F) and triple (F87W, T185F, V295I) mutants of P450cam, which provide additional bulk in the active site and increase the frequency of heme-substrate collision. Molecular dynamics simulations (300 ps on each protein) indicate that these mutants do not significantly perturb the three-dimensional fold of the enzyme, or local structure in the region of the active site. Both mutants bind the substrate more stably near the heme than the wild-type. Interestingly, however, the bulkier triple mutant seems to actually inhibit heme-substrate interactions relative to the double mutant. Over the final 200 ps of simulation, TCA is within 1 A of nonbonded contact with the heme 25% more often in the double mutant versus the wild-type. The triple mutant, on the other hand, binds TCA within 1 A of the heme only 15% as often as the wild-type. These results indicate that the double mutant may reductively dehalogenate TCA, a property not observed for the native protein. Implications for other experimentally measurable parameters are discussed.

Protein conformational landscapes: energy minimization and clustering of a long molecular dynamics trajectory

Using energy minimization and cluster analysis, we have analyzed a 1020 ps molecular dynamics trajectory of solvated bovine pancreatic trypsin inhibitor. Elucidation of conformational substates in this way both illustrates the degree of conformational convergence in the simulation and reduces the structural data to a tractable subset. The relative movement of structures upon energy minimization was used to estimate the sizes of features on the protein potential energy surface. The structures were analyzed using their pairwise root-mean-square C alpha deviations, which gave a global measure of conformational changes that would not be apparent by monitoring single degrees of freedom. At time scales of 0.1 ps, energy minimization detected sharp transitions between energy minima separated by 0.1 A rms deviation. Larger conformational clusters containing these smaller minima and separated by 0.25 A were seen at 1 ps time scales. Both of these small features of the conformational landscape were characterized by movements in loop regions associated with small, correlated backbone dihedral angle shifts. On a nanosecond time scale, the main features of the protein energy landscape were clusters separated by over 0.7 A rms deviation, with only seven of these substates visited over the 1 ns trajectory. These substrates, discernible both before and after energy minimization, differ mainly in a monotonic pivot of the loop residues 11-18 over the course of the simulation. This loop contains lysine 17, which specifically binds to trypsin in the active site. The trajectory did not return to previously visited clusters, indicating that this trajectory has not been shown to have completely sampled the conformational substates available to it. Because the apparent convergence to a single region of conformation space depends on both the time scale of observation and the size of the conformational features examined, convergence must be operationally defined within the context of the simulation.

Structural features of isolated M2 helices of nicotinic receptors. Simulated annealing via molecular dynamics studies

The nicotinic acetylcholine receptor is an integral membrane protein and a ligand-gated cation channel. It has stoichiometry alpha 2 beta gamma delta, the subunits arranged symmetrically around an approximate five-fold axis. Five M2 helices, one from each subunit, form a parallel helix bundle surrounding a central pore. Simulated annealing via restrained molecular dynamics (SA/MD) has been employed to generate ensembles of isolated M2 transmembrane helices. Four ensembles of two different M2 helix sequences, M2 delta and M2 gamma, have been generated by SA/MD. The ensembles differed in their treatment of electrostatic interactions. Analysis of the simulated structures showed that intra-helical H-bonds were more strongly conserved in the C-terminal (and more hydrophobic) segment of M2 helices. Conformations of polar sidechains have been analyzed, placing particular emphasis on EK (and QK) pairs at the N-termini of M2 delta (and M2 gamma) helices. Conformations of EK sidechain pairs were obtained for the high resolution structures in the protein database in order to guide our analysis of simulated structures. Serine and threonine sidechain conformations in the M2 models also have been determined. Implications of studies of isolated M2 helices for models of the intact pore region of the nicotinic receptor are discussed.

De novo prediction of polypeptide conformations using dihedral probability grid Monte Carlo methodology

We tested the dihedral probability grid Monte Carlo (DPG-MC) methodology to determine optimal conformations of polypeptides by applying it to predict the low energy ensemble for two peptides whose solution NMR structures are known: integrin receptor peptide (YGRGDSP, Type II beta-turn) and S3 alpha-helical peptide (YMSEDEL KAAEAAFKRHGPT). DPG-MC involves importance sampling, local random stepping in the vicinity of a current local minima, and Metropolis sampling criteria for acceptance or rejection of new structures. Internal coordinate values are based on side-chain-specific dihedral angle probability distributions (from analysis of high-resolution protein crystal structures). Important features of DPG-MC are: (1) Each DPG-MC step selects the torsion angles (phi, psi, chi) from a discrete grid that are then applied directly to the structure. The torsion angle increments can be taken as S = 60, 30, 15, 10, or 5 degrees, depending on the application. (2) DPG-MC utilizes a temperature-dependent probability function (P) in conjunction with Metropolis sampling to accept or reject new structures. For each peptide, we found close agreement with the known structure for the low energy conformational ensemble located with DPG-MC. This suggests that DPG-MC will be useful for predicting conformations of other polypeptides.

Diffuse scattering in protein crystallography

The understanding of flexibility and deformability in proteins is one of the current major challenges of structural molecular biology. The knowledge of the average atomic positions of three-dimensional folding of proteins, which is obtained either by X-ray diffraction or n.m.r. spectroscopy, is generally not sufficient to explain their functional mechanisms. Very often it is necessary to consider the existence of other concerted atomic motions as, for example, in the well-known case of the CO molecule fixation at the active site of myoglobin which requires the concerted displacement of a large number of atoms in order to open a channel down to this site. This opening, which depends on the physico-chemical conditions, plays the role of a regulator in the biochemical reactions (Janin & Wodak, 1983; Tainer et al. 1984; Westhof et al. 1984; Ormos et al. 1988).

Substrate recognition by ribosome-inactivating protein studied by molecular modeling and molecular electrostatic potentials

A computer model of dianthin 30, a type 1 ribosome-inactivating protein (RIP), is constructed by homology modeling using two known X-ray structures; a type 1 RIP, pokeweed antiviral protein (PAP), and chain A of a type 2 RIP, ricin. The 3D structure is refined by molecular dynamics and its binding site compared with those of PAP and ricin using molecular electrostatic potential mapping. The differences in the maps obtained clearly show how, despite the similarity of the topology of the binding site, differences in electrostatic potential can account for the experimentally observed differences in substrate recognition and binding. This demonstrates the potential of these techniques for guiding further experimental analyses.

Dramatic differences in the motions of the mouth of open and closed cytochrome P450BM-3 by molecular dynamics simulations

Molecular dynamics trajectories were calculated separately for each of the two molecules in the asymmetric unit of the crystal structure of the hemoprotein domain of cytochrome P450BM-3. Each simulation was 200 ps in length and included a 10 A layer of explicit solvent. The simulated time-average structure of each P450BM-3 molecule is closer to its crystal structure than the two molecular dynamics time-averaged structures are to each other. In the crystal structure, molecule 2 has a more accessible substrate binding pocket than molecule 1, and this difference is maintained throughout the simulations presented here. In particular, the substrate docking regions of molecule 1 and molecule 2 diverge in the solution state simulations. The mouth of the substrate binding pocket is significantly more mobile in the simulation of molecule 2 than in the simulation of molecule 1. For molecule 1, the width of the mouth is only slightly larger than its X-ray value of 8.7 A and undergoes fluctuations of about 1 A. However, in molecule 2, the mouth of the substrate binding pocket is dramatically more open in the time-average molecular dynamics structure (14.7 A) than in the X-ray structure (10.9 A). Furthermore, this region of the protein undergoes large amplitude motions during the trajectory that are not seen in the trajectory of molecule 1, repeatedly opening and closing up to 7 A. Presumably, the binding of different substrates will induce the mouth region to adopt different conformations from within the wide range of structures that are accessible.

An evaluation of molecular models of the cytochrome P450 Streptomyces griseolus enzymes P450SU1 and P450SU2

P450SU1 and P450SU2 are herbicide-inducible bacterial cytochrome P450 enzymes from Streptomyces griseolus. They have two of the highest sequence indentities to camphor hydroxylase (P450cam from Pseudomonas putida), the cytochrome P450 with the first known crystal structure. We have built several models of these two proteins to investigate the variability in the structures that can occur from using different modeling protocols. We looked at variability due to alignment methods, backbone loop conformations and refinement methods. We have constructed two models for each protein using two alignment algorithms, and then an additional model using an identical alignment but different loop conformations for both buried and surface loops. The alignments used to build the models were created using the Needleman-Wunsch method, adapted for multiple sequences, and a manual method that utilized both a dot-matrix search matrix and the Needleman-Wunsch method. After constructing the initial models, several energy minimization methods were used to explore the variability in the final models caused by the choice of minimization techniques. Features of cytochrome P450cam and the cytochrome P450 superfamily, such as the ferredoxin binding site, the heme binding site and the substrate binding site were used to evaluate the validity of the models. Although the final structures were very similar between the models with different alignments, active-site residues were found to be dependent on the conformations of buried loops and early stages of energy minimization. We show which regions of the active site are the most dependent on the particular methods used, and which parts of the structures seem to be independent of the methods.

Investigation of domain motions in bacteriophage T4 lysozyme

Hinge-bending in T4 lysozyme has been inferred from single amino acid mutant crystalline allomorphs by Matthews and coworkers. This raises an important question: are the different conformers in the unit cell artifacts of crystal packing forces, or do they represent different solution state structures? The objective of this theoretical study is to determine whether domain motions and hinge-bending could be simulated in T4 lysozyme using molecular dynamics. An analysis of a 400 ps molecular dynamics simulation of the 164 amino acid enzyme T4 lysozyme is presented. Molecular dynamics calculations were computed using the Discover software package (Biosym Technologies). All hydrogen atoms were modeled explicitly with the inclusion of all 152 crystallographic waters at a temperature of 300 K. The native T4 lysozyme molecular dynamics simulation demonstrated hinge-bending in the protein. Relative domain motions between the N-terminal and C-terminal domains were evident. The enzyme hinge bending sites resulted from small changes in backbone atom conformations over several residues rather than rotation about a single bound. Two hinge foci were found in the simulation. One locus comprises residues 8-14 near the C-terminal of the A helix; the other site, residues 77-83 near the C-terminal of the C helix. Comparison of several snapshot structures from the dynamics trajectory clearly illustrates domain motions between the two lysozyme lobes. Time correlated atomic motions in the protein were analyzed using a dynamical cross-correlation map. We found a high degree of correlated atomic motions in each of the domains and, to a lesser extent, anticorrelated motions between the two domains. We also found that the hairpin loop in the N-terminal lobe (residues 19-24) acted as a mobile 'flap' and exhibited highly correlated dynamic motions across the cleft of the active site, especially with residue 142.

Active-site mobility inhibits reductive dehalogenation of 1,1,1-trichloroethane by cytochrome P450cam

Recent studies by Wackett and co-workers have shown that cytochrome P450cam is capable of reductively dehalogenating hexachloroethane at a significant rate, but that no appreciable dehalogenation of 1,1,1-trichloroethane is observed. A growing body of evidence indicates that differences in intrinsic reactivity can not completely explain this observation. We therefore explored the possible role of differences in preferred binding orientation and in active-site mobility. A detailed analysis of molecular dynamics trajectories with each of these substrates bound at the active site of P450cam is presented. While the dynamics and overall time-average structure calculated for the protein are similar in the two trajectories, the two substrates behave quite differently. The smaller substrate, 1,1,1-trichloroethane, is significantly more mobile than hexachloroethane and has a preferred orientation in which the substituted carbon is generally far from the heme iron. In contrast, for hexachloroethane, one of the chlorine atoms is nearly always in van der Waals contact with the heme iron, which should favor the initial electron transfer step.