Protein structure prediction and potential energy landscape analysis using continuous global minimization

Protein-Peptide Complex Prediction through Fragment Interaction Patterns

A well-behaved physics-based all-atom scoring function for protein structure prediction is analyzed with several widely used all-atom decoy sets. The scoring function, termed AMBER/Poisson-Boltzmann (PB), is based on a refined AMBER force field for intramolecular interactions and an efficient PB model for solvation interactions. Testing on the chosen decoy sets shows that the scoring function, which is designed to consider detailed chemical environments, is able to consistently discriminate all 62 native crystal structures after considering the heteroatom groups, disulfide bonds, and crystal packing effects that are not included in the decoy structures. When NMR structures are considered in the testing, the scoring function is able to discriminate 8 out of 10 targets. In the more challenging test of selecting near-native structures, the scoring function also performs very well: for the majority of the targets studied, the scoring function is able to select decoys that are close to the corresponding native structures as evaluated by ranking numbers and backbone Calpha root mean square deviations. Various important components of the scoring function are also studied to understand their discriminative contributions toward the rankings of native and near-native structures. It is found that neither the nonpolar solvation energy as modeled by the surface area model nor a higher protein dielectric constant improves its discriminative power. The terms remaining to be improved are related to 1-4 interactions. The most troublesome term is found to be the large and highly fluctuating 1-4 electrostatics term, not the dihedral-angle term. These data support ongoing efforts in the community to develop protein structure prediction methods with physics-based potentials that are competitive with knowledge-based potentials.

The transcriptional antiterminator protein LicT regulates the expression of Bacillus subtilis operons involved in beta-glucoside metabolism. It consists of an N-terminal RNA-binding domain (co-antiterminator (CAT)) and two phosphorylatable phosphotransferase system regulation domains (PRD1 and PRD2). In the activated state, each PRD forms a dimeric unit with the phosphorylation sites totally buried at the dimer interface. Here we present the 1.95 A resolution structure of the inactive LicT PRDs as well as the molecular solution structure of the full-length protein deduced from small angle x-ray scattering. Comparison of native (inactive) and mutant (constitutively active) PRD crystal structures shows massive tertiary and quaternary rearrangements of the entire regulatory domain. In the inactive state, a wide swing movement of PRD2 results in dimer opening and brings the phosphorylation sites to the protein surface. This movement is accompanied by additional structural rearrangements of both the PRD1-PRD1 ' interface and the CAT-PRD1 linker. Small angle x-ray scattering experiments indicate that the amplitude of the PRD2 swing might even be wider in solution than in the crystals. Our results suggest that PRD2 is highly mobile in the native protein, whereas it is locked upon activation by phosphorylation.

ABSTRACTWe present a knowledge‐based function to score protein decoys based on their similarity to native structure. A set of features is constructed to describe the structure and sequence of the entire protein chain. Furthermore, a qualitative relationship is established between the calculated features and the underlying electromagnetic interaction that dominates this scale. The features we use are associated with residue–residue distances, residue–solvent distances, pairwise knowledge‐based potentials and a four‐body potential. In addition, we introduce a new target to be predicted, the fitness score, which measures the similarity of a model to the native structure. This new approach enables us to obtain information both from decoys and from native structures. It is also devoid of previous problems associated with knowledge‐based potentials. These features were obtained for a large set of native and decoy structures and a back‐propagating neural network was trained to predict the fitness score. Overall this new scoring potential proved to be superior to the knowledge‐based scoring functions used as its inputs. In particular, in the latest CASP (CASP10) experiment our method was ranked third for all targets, and second for freely modeled hard targets among about 200 groups for top model prediction. Ours was the only method ranked in the top three for all targets and for hard targets. This shows that initial results from the novel approach are able to capture details that were missed by a broad spectrum of protein structure prediction approaches. Source codes and executable from this work are freely available at http://mathmed.org/#Software and http://mamiris.com/. Proteins 2014; 82:752–759. © 2013 Wiley Periodicals, Inc.

Lecture 21

It is hard to construct theories for the folding of globular proteins because they are large and complicated molecules having enormous numbers of nonnative conformations and having native states that are complicated to describe. Statistical mechanical theories of protein folding are constructed around major simplifying assumptions about the energy as a function of conformation and/or simplifications of the representation of the polypeptide chain, such as one point per residue on a cubic lattice. It is not clear how the results of these theories are affected by their various simplifications. Here we take a very different simplification approach where the chain is accurately represented and the energy of each conformation is calculated by a not unreasonable empirical function. However, the set of amino acid sequences and allowed conformations is so restricted that it becomes computationally feasible to examine them all. Hence we are able to calculate melting curves for thermal denaturation as well as the detailed kinetic pathway of refolding. Such calculations are based on a novel representation of the conformations as points in an abstract 12-dimensional Euclidean conformation space. Fast folding sequences have relatively high melting temperatures, native structures with relatively low energies, small kinetic barriers between local minima, and relatively many conformations in the global energy minimum's watershed. In contrast to other folding theories, these models show no necessary relationship between fast folding and an overall funnel shape to the energy surface, or a large energy gap between the native and the lowest nonnative structure, or the depth of the native energy minimum compared to the roughness of the energy landscape.

Proteins play a key role in cells' function and metabolism. Their functions are directly related with the three-dimensional (3D) native structure. Different algorithms have been proposed to predict the 3D protein structure from the amino acids sequence by minimizing its free energy, nonetheless, this problem still a great challenge in structural biology. The space of possible conformations becomes very large for long proteins, making it difficult to search for the optimal minimum free energy with the available computational resources. Due to the complexity involved in the protein folding problem the researchers have developed some simplified protein models as the hydrophobic-polar (HP). In this work we propose a parallel algorithm to predict protein conformations in HP model by dividing tasks among processing nodes distributed in a server cluster. With this approach we circumvent the memory constraints observed by other approaches. We discuss the correctness of the algorithm based on the verification results achieved by model checking. The performance of our algorithm was evaluated but although benefits of parallel execution are noticeable for small proteins, we demonstrated that our algorithm has an exponential complexity.

This paper summarizes some fundamental statistical mechanical aspects of protein folding and the prediction of protein structure. In order to predict the native structure of a protein, it is necessary to understand the physical conditions that determine its unique and thermodynamically-stable native structure, and to surmount the numerous local energy minima to arrive at the native structure. A statistical mechanical approach has been used to address the problem of foldability of polypeptides, and global minimization techniques have been developed to solve the multiple-minima problem. Some recent progress in these areas, made in our laboratory, is described.

The spontaneous folding of protein has posed a fundamental problem popularly known as Levinthal paradox. Effort has been harnessed both in vivo and in vitro to obtain the native (functional) structure of proteins which is the most stable thermodynamically from the experimental and theoretical point of view. A coarse-grained simplified model has been a very useful tool for the study of protein structure by representing proteins as linear and self-avoiding chains which contain two types of monomers H (hydrophobic) and P (polar) on a lattice model. We present the results of numerical studies of the energy landscape on three sequences of amino acid and distinct funnels were generated with contact interactions to obtain the ground state conformation on a square lattice using a move-biased Monte Carlo simulation (MBMC). This method is very efficient and outperforms the conventional Monte Carlo method in the state-of-the-art results.

We present a new structurally derived pair-to-pair substitution matrix (P2PMAT). This matrix is constructed from a very large amount of integrated high quality multiple sequence alignments (Blocks) and protein structures. It evaluates the likelihoods of all 160,000 pair-to-pair substitutions. P2PMAT matrix implicitly accounts for evolutionary conservation, correlated mutations, and residue-residue contact potentials. The usefulness of the matrix for structural predictions is shown in this article. Predicting protein residue-residue contacts from sequence information alone, by our method (P2PConPred) is particularly accurate in the protein cores, where it performs better than other basic contact prediction methods (increasing accuracy by 25-60%). The method mean accuracy for protein cores is 24% for 59 diverse families and 34% for a subset of proteins shorter than 100 residues. This is above the level that was recently shown to be sufficient to significantly improve ab initio protein structure prediction. We also demonstrate the ability of our approach to identify native structures within large sets of (300-2000) protein decoys. On the basis of evolutionary information alone our method ranks the native structure in the top 0.3% of the decoys in 4/10 of the sets, and in 8/10 of sets the native structure is ranked in the top 10% of the decoys. The method can, thus, be used to assist filtering wrong models, complementing traditional scoring functions.

Protein structure prediction from inaccurate and sparse NMR data using an enhanced genetic algorithm

Computer simulation of secondary structure folding of RNA chains demonstrates that these chains under some temperature and solvent conditions find their native conformations much faster than by exhaustive sorting of all the possible chain conformations. This is true both for the random and the edited chains. However, for a random chain the native structure is thermodynamically stable in one region of temperature and solvent quality, while fast folding (and even faster unfolding) of the random chain occurs in another region. Some editing of a chain is necessary to make its native structure thermodynamically stable at the optimal for a fast folding range of conditions. Only for the edited chains can the thermodynamically stable native secondary structures fold fast. The dependence of folding time on the chain length is not an exponential one but much less steep.

Membrane transporters including glucose transporters (GLUTs) are involved in cellular energy supplies, cell metabolism and other vital biological activities. They have also been implicated in cancer proliferation and metastasis, thus they represent an important target in combatting cancer. However, membrane transporters are very difficult to study due to their multispan transmembrane properties. The new computational tool, AlphaFold2, offers highly accurate predictions of three-dimensional protein structures. The glutamine, threonine and tyrosine (QTY) code provides a systematic method of rendering hydrophobic sequences into hydrophilic ones. Here, we present computational studies of native integral membrane GLUTs with 12 transmembrane helical segments determined by X-ray crystallography and CryoEM, comparing the AlphaFold2-predicted native structure to their water-soluble QTY variants predicted by AlphaFold2. In the native structures of the transmembrane helices, there are hydrophobic amino acids leucine (L), isoleucine (I), valine (V) and phenylalanine (F). Applying the QTY code, these hydrophobic amino acids are systematically replaced by hydrophilic amino acids, glutamine (Q), threonine (T) and tyrosine (Y) rendering them water-soluble. We present the superposed structures of native GLUTs and their water-soluble QTY variants. The superposed structures show remarkable similar residue mean square distance values between 0.47 and 3.6Å (most about 1-2Å) despite >44% transmembrane amino acid differences. We also show the differences of hydrophobicity patches between the native membrane transporters and their QTY variants. We explain the rationale why the membrane protein QTY variants become water-soluble. Our study provides insight into the differences between the hydrophobic helices and hydrophilic helices, and offers confirmation of the QTY method for studying multispan transmembrane proteins and other aggregated proteins through their water-soluble variants.

Membrane transporters including glucose transporters (GLUTs) are involved in cellular energy supplies, cell metabolism and other vital biological activities. They have also been implicated in cancer proliferation and metastasis, thus they represent an important target in combatting cancer. However, membrane transporters are very difficult to study due to their multispan transmembrane properties. The new computational tool, AlphaFold2, offers highly accurate predictions of three-dimensional protein structures. The glutamine, threonine and tyrosine (QTY) code provides a systematic method of rendering hydrophobic sequences into hydrophilic ones. Here, we present computational studies of native integral membrane GLUTs with 12 transmembrane helical segments determined by X-ray crystallography and CryoEM, comparing the AlphaFold2-predicted native structure to their water-soluble QTY variants predicted by AlphaFold2. In the native structures of the transmembrane helices, there are hydrophobic amino acids leucine (L), isoleucine (I), valine (V) and phenylalanine (F). Applying the QTY code, these hydrophobic amino acids are systematically replaced by hydrophilic amino acids, glutamine (Q), threonine (T) and tyrosine (Y) rendering them water-soluble. We present the superposed structures of native GLUTs and their water-soluble QTY variants. The superposed structures show remarkable similar residue mean square distance values between 0.47 and 3.6 Å (most about 1–2 Å) despite >44% transmembrane amino acid differences. We also show the differences of hydrophobicity patches between the native membrane transporters and their QTY variants. We explain the rationale why the membrane protein QTY variants become water-soluble. Our study provides insight into the differences between the hydrophobic helices and hydrophilic helices, and offers confirmation of the QTY method for studying multispan transmembrane proteins and other aggregated proteins through their water-soluble variants.

Protein Structure Prediction (PSP) is an NP-hard optimization problem that has been solved by many existing algorithms. Simply, it can be thought of a process of predicting the native 3D structure from its amino acid sequence. Chemical Reaction Optimization (CRO) is a recent metaheuristic algorithm that has been applied to many well-known problems and has shown better performance compared to the existing ones. So, we have applied CRO algorithm to solve the PSP problem. The four operators of CRO: on-wall ineffective collision, decomposition, inter-molecular ineffective collision, and synthesis have been designed to solve PSP problem. We have also designed a repair mechanism to get the correct structure. The experiment results show that CRO performs well in the case of PSP.