Modelling of peptide and protein structures

Summary The modelling of protein structures (whether isolated, in solution, or involved in recognition processes) is reviewed, free of any mathematical apparatus, to provide an overview of the concepts as well as leading references. A general feeling for this field of work is first established by a sampling of some impressions on its difficulties and chances of success. Then, the main body of this work examines the information available (databases and parameters), presents the theoretical foundations for the modelling procedures (with emphasis on the potential energy functions), surveys the existing simulation techniques and prediction methods, and discusses the problems still to be faced. For completeness, a representative list of existing software packages is presented in the Appendix.Presented at the Third International Congress on Amino Acids, Vienna, August 23–27, 1993.     

Prediction and evaluation of side-chain conformations for protein backbone structures

A common approach to protein modeling is to propose a backbone structure based on homology or threading and then to attempt to build side chains onto this backbone. A fast algorithm using the simple criteria of atomic overlap and overall rotamer probability is proposed for this purpose. The method was first tested in the context of exhaustive searches of side chain configuration space in protein cores and was then applied to all side chains in 49 proteins of known structure, using simulated annealing to sample space. The latter procedure obtains the correct rotamer for 57% and the correct χ₁ value for 74% of the 6751 residues in the sample. When low-temperature Monte-Carlo simulations are initiated from the results of the simulated-annealing processes, consensus configurations are obtained which exhibit slightly more accurate predictions. The Monte-Carlo procedure also allows converged side chain entropies to be calculated for all residues. These prove to be accurate indicators of prediction reliability. For example, the correct rotamer is obtained for 79% and the correct χ₁ value is obtained for 84% of the half of the sample residues exhibiting the lowest entropies. Side chain entropy and predictability are nearly completely uncorrelated with solvent-accessible area. Some precedents for and implications of this observation are discussed. © 1996 Wiley-Liss, Inc.     

Identification of near-native structures by clustering protein docking conformations

Most state-of-the-art protein-protein docking algorithms use the Fast Fourier Transform (FFT) technique to sample the six-dimensional translational and rotational space. Scoring functions including shape complementarity, electrostatics, and desolvation are usually exploited in ranking the docking conformations. While these rigid-body docking methods provide good performance in bound docking, using unbound structures as input frequently leads to a high number of false positive hits. For the purpose of better selecting correct docking conformations, we structurally cluster the docking decoys generated by four widely-used FFT-based protein-protein docking methods. In all cases, the selection based on cluster size outperforms the ranking based on the inherent scoring function. If we cluster decoys from different servers together, only marginal improvement is obtained in comparison with clustering decoys from the best individual server. A collection of multiple decoy sets of comparable quality will be the key to improve the clustering result from meta-docking servers.     

Crystal structures of a Rab protein in its inactive and active conformations

We have determined crystal structures of Sec4, a member of the Rab family in the G protein superfamily, in two states: bound to GDP, and to a non-hydrolyzable GTP analog, guanosine-5'-(beta, gamma)-imidotriphosphate (GppNHp). This represents the first structure of a Rab protein bound to GDP. Sec4 in both states grossly resembles other G proteins bound to GDP and GppNHp. In Sec4-GppNHp, structural features common to active Rab proteins are observed. In Sec4-GDP, the switch I region is highly disordered and displaced relative to the switch I region of Ras-GDP. In two of the four molecules of Sec4-GDP in the asymmetric unit of the Sec4-GDP crystals, the switch II region adopts a conformation similar to that seen in the structure of the small G protein Ran bound to GDP. This allows residues threonine 76, glutamate 80, and arginine 81 of Sec4 to make contacts with other conserved residues and water molecules important for nucleotide binding. In the other two molecules in the asymmetric unit, these interactions do not take place. This structural variability in both the switch I and switch II regions of GDP-bound Sec4 provides a possible explanation for the high off-rate of GDP bound to Sec4, and suggests a mechanism for regulation of the GTPase cycle of Rab proteins by GDI proteins.     

Enhanced sampling of peptide and protein conformations using replica exchange simulations with a peptide backbone biasing-potential

During replica exchange molecular dynamics (RexMD) simulations, several replicas of a system are simulated at different temperatures in parallel allowing for exchange between replicas at frequent intervals. This technique allows significantly improved sampling of conformational space and is increasingly being used for structure prediction of peptides and proteins. A drawback of the standard temperature RexMD is the rapid increase of the replica number with increasing system size to cover a desired temperature range. In an effort to limit the number of replicas, a new Hamiltonian-RexMD method has been developed that is specifically designed to enhance the sampling of peptide and protein conformations by applying various levels of a backbone biasing potential for each replica run. The biasing potential lowers the barrier for backbone dihedral transitions and promotes enhanced peptide backbone transitions along the replica coordinate. The application on several peptide cases including in all cases explicit solvent indicates significantly improved conformational sampling when compared with standard MD simulations. This was achieved with a very modest number of 5-7 replicas for each simulation system making it ideally suited for peptide and protein folding simulations as well as refinement of protein model structures in the presence of explicit solvent.     

Occurrence and role of cis peptide bonds in protein structures

It has been widely assumed that the occurrence of cis peptide bonds in proteins is quite rare due to unfavorable contacts between adjacent amino acid residues in this isomeric form. To investigate this assumption, the Brookhaven Protein Data Bank was examined for the occurrences of cis peptide bonds. Out of 31,005 amide bonds, only 17, or 0.05%, are cis, while 99 of the 1534 imide bonds (X-Pro), or 6.5%, are cis. These figures are considerably less than the distribution predicted on the basis of the potential energy difference between the cis and trans isomeric forms and experimental data on small peptides. It is not known whether the lower than expected occurrence of cis peptide bonds arises from constraints imposed by the protein environment, or from assumptions made in the solution of the X-ray crystal structures. However, when the occurrence of cis bonds in the data base is examined relative to the resolution of the structures, the number of cis bonds increases with increasing resolution. The distributions seen for these peptide omega bonds in the data base are not the same shape as the distributions predicted from simple potential energy barriers. They are sharper in the main, but they are also broader at the base with significant numbers of nonplanar peptide bonds. Cis peptide bonds are found primarily in bends and turns and, in the case of cis imide bonds (X-PRO), this correlation is so high that it suggests a specific role for cis imide groups in such structures.     

Structural refinement of protein A mimetic peptide

A novel dendrimeric peptide ligand dubbed D-PAM-Φ was designed to achieve a high capacity for human IgG through the decoration of the D-PAM scaffold. The design criteria based on the introduction of small hydrophobic groups on the D-PAM structure were supported by the recently published solid-state structure of D-PAM complexed to the Fc fragment of a recombinant human IgG1 and by molecular dynamic simulations that provided information on the mode of binding of phenylacetyl-D-PAM (D-PAM-Φ). D-PAM-Φ was immobilised on an activated solid support and compared with the parent D-PAM affinity matrix. The newly obtained affinity sorbent was evaluated for its capacity to selectively capture polyclonal human IgG; the binding capacity was approximately 10 mg/ml, an almost 10-fold enhancement with respect to the D-PAM-functionalised matrices without the specificity of binding being reduced. The new ligand was also effective in the capturing of recombinant humanised IgG1 from a clarified cell culture supernatant. Under a typical laboratory-scale affinity chromatography assembly and preliminarily optimised binding conditions, the affinity purification of humanised IgG1 from culture supernatants rendered the desired product, with purity higher than 90%. The results suggest that the application of the computational approach on the structure of the D-PAM-Fc complex may be very valuable in the development of novel lead molecules for the downstream processing of human or humanised antibodies used in therapy.     

Hidden symmetry of peptide and protein primary structures]

The internal symmetry of peptide chains was considered. To identify symmetrically located equivalent amino acids, the signatures method and the code of amino acid codon roots were applied. There was revealed the hidden symmetry of amino acid sequences of peptides and proteins as well as of their active centres. Amino acids having common codon roots in primary (and supposedly in the spatial "biologically active") molecular structures, are located symmetrically. Definition of local symmetry of peptide chains was proposed to use as one of the elements of complex analysis to determine location of molecular active centres.     

Characterization of non-planar peptide groups in protein crystal structures

Peptide groups are generally assumed to be planar in protein structure, due to 'rigid' partial double bond character of peptide bonds, thus the value of peptide torsion angle omega should be restricted to 180 degrees for the usual trans form of peptide unit. However, on analyzing the ultra-high resolution protein crystal database, we find that in some cases, omega deviates >10 degrees from its usual value of 180 degrees, indicating significant non-planarity of peptide groups. Moreover, the non-planarity for most of the amino acids is found to be 'biased' towards values of omega smaller than 180 degrees. Similar trend for to is confirmed by the neutron diffraction data for proteins. The neutron diffraction database also reveals that non-planar peptide groups are generally correlated to 'pyramidal' structure of the peptide-nitrogen bonds. Consequently, the hydrogen atom of peptide group deviates from its planar position, as measured by the 'improper' torsion angle theta. Thus, we find that both the angles omega and theta point towards a significant amount of non-planarity of peptide groups, which cannot be ignored. The role of peptide nonplanarity in protein function is, however, not yet clear.     

Recapitulation and design of protein binding peptide structures and sequences

An important objective of computational protein design is the generation of high affinity peptide inhibitors of protein-peptide interactions, both as a precursor to the development of therapeutics aimed at disrupting disease causing complexes, and as a tool to aid investigators in understanding the role of specific complexes in the cell. We have developed a computational approach to increase the affinity of a protein-peptide complex by designing N or C-terminal extensions which interact with the protein outside the canonical peptide binding pocket. In a first in silico test, we show that by simultaneously optimizing the sequence and structure of three to nine residue peptide extensions starting from short (1-6 residue) peptide stubs in the binding pocket of a peptide binding protein, the approach can recover both the conformations and the sequences of known binding peptides. Comparison with phage display and other experimental data suggests that the peptide extension approach recapitulates naturally occurring peptide binding specificity better than fixed backbone design, and that it should be useful for predicting peptide binding specificities from crystal structures. We then experimentally test the approach by designing extensions for p53 and dystroglycan-based peptides predicted to bind with increased affinity to the Mdm2 oncoprotein and to dystrophin, respectively. The measured increases in affinity are modest, revealing some limitations of the method. Based on these in silico and experimental results, we discuss future applications of the approach to the prediction and design of protein-peptide interactions.     

Patterns and conformations of commonly occurring supersecondary structures (basic motifs) in protein data bank

Short peptides connecting-helices and-strands have been analyzed in 240 proteins refined at resolutions of 0.25 nm or better. Connecting peptides of lengths between one and five residues have been classified as part of supersecondary motifs of four types:, , , and. Careful consideration has been given to the definition of secondary structures on the basis of hydrogen bonds and main-chain conformational angles. Using five classes of residue conformation—a, b, e, l, t—in the nonregular structure regions of, space, 34 classes of supersecondary motifs occurring at least five times have been identified. Among these 34 classes, 11 classes that occur more than 25 times are commonly occurring supersecondary structure motifs. The patterns and conformations of the 11 commonly occurring supersecondary structure motifs have been characterized, demonstrating that patterns and conformations adopted by supersecondary structure motifs are limited. The results have relevance to structure prediction, comparative modeling, and protein folding.     

Perturbation of peptide conformations induced in anisotropic environments

Reversed-phase high performance liquid chromatography (RP HPLC) has been found to be a convenient and powerful tool for the study of the secondary structure of peptides. Here, the ability of proline to perturb the secondary structures of peptides induced at aqueous-lipid interfaces and the induced conformation of polyproline peptides were investigated by means of RP HPLC. For these studies, four different complete sets of substitution analogues of model peptides expected to have specific induced conformations were used. In the first two studies, a single lysine was “walked” through two 18-residue polyproline sequences (one N-acetylated, the other not). In the remaining two studies, a proline was “walked” through two different sequences that had been found earlier to be induced into an α-helical conformation during RP HPLC (an 18-residue polyalanine sequence and the amphipathic 14-residue sequence Ac-LLKLLKKLLKKLKK-NH₂). Sixty-eight individual analogues were synthesized for this study and the effect of the respective substitutions on retention times was determined. The results are consistent with the concept that, upon interaction with the C-18 of the stationary phase during RP HPLC, polyproline is induced into a type II helical conformation, polyalanine into an α-helical conformation, and Ac-LLKLLKKLLKKLKK-NH₂ into an amphipathic α-helical array. In an extension of this study, the antimicrobial activities of Ac-LLKLLKKLLKKLKK-NH₂ and its 18 proline substitution analogues were found to be inversely correlated with their RP HPLC retention times.     

Dependency between consecutive local conformations helps assemble protein structures from secondary structures using Go potential and greedy algorithm

Discretization of protein conformational space and fragment assembly methods simplify the search of native structures. These methods, mostly of Monte Carlo and genetic-type, do not exploit, however, the fact that short fragments describing consecutive parts of proteins are conformation-dependent. Yet, this information should be useful in improving ab initio and comparative protein structure modeling. In a preliminary study, we have assessed the possibility of using greedy algorithms for protein structure reconstruction based on the assembly of fragments of four-residue length. Greedy algorithms differ from Monte Carlo and genetic approaches in that they grow a polypeptide chain one fragment after another. Here, we move one step further in complexity, and provide strong evidence that the dependence between consecutive local conformations during assembly makes possible the reconstruction of protein structures from their secondary structures using a Go potential. Overall our procedure can reproduce 20 protein structures of 50-164 amino acids within 2.7 to 6.5 A RMSd and is able to identify native topologies for all proteins, although some targets are stabilized by very long-range interactions.     

重金属结合肽（蛋白）的结构分析

金属离子与金属结合肽（蛋白）的相互作用与应用研究,一直是生物无机化学的重点和热点,也是分子间相互作用研究领域的难点。本研究利用ClustalX、BLAST等生物信息技术与方法对大量已知的重金属结合肽进行分析与数据挖掘。确定筛选获得的重金属结合肽常富含His,无Cys,无金属结合肽模式序列,进化不保守;部分氨基酸序列结构（如六肽）可在蛋白数据库中找到相似序列。序列特征主要为Zn^2＋相关的转录因子。本研究为重金属结合蛋白-重金属离子的相互作用分析简化为重金属结合肽-重金属离子的结构模拟与分析提供了重要的理论基础和研究手段。     

Structural interrogation of phosphoproteome identified by mass spectrometry reveals allowed and disallowed regions of phosphoconformation

Background

Many biologically active compounds bind to plasma transport proteins, and this binding can be either advantageous or disadvantageous from a drug design perspective. Human serum albumin (HSA) is one of the most important transport proteins in the cardiovascular system due to its great binding capacity and high physiological concentration. HSA has a preference for accommodating neutral lipophilic and acidic drug-like ligands, but is also surprisingly able to bind positively charged peptides. Understanding of how short cationic antimicrobial peptides interact with human serum albumin is of importance for developing such compounds into the clinics.

Results

The binding of a selection of short synthetic cationic antimicrobial peptides (CAPs) to human albumin with binding affinities in the μM range is described. Competitive isothermal titration calorimetry (ITC) and NMR WaterLOGSY experiments mapped the binding site of the CAPs to the well-known drug site II within subdomain IIIA of HSA. Thermodynamic and structural analysis revealed that the binding is exclusively driven by interactions with the hydrophobic moieties of the peptides, and is independent of the cationic residues that are vital for antimicrobial activity. Both of the hydrophobic moieties comprising the peptides were detected to interact with drug site II by NMR saturation transfer difference (STD) group epitope mapping (GEM) and INPHARMA experiments. Molecular models of the complexes between the peptides and albumin were constructed using docking experiments, and support the binding hypothesis and confirm the overall binding affinities of the CAPs.

Conclusions

The biophysical and structural characterizations of albumin-peptide complexes reported here provide detailed insight into how albumin can bind short cationic peptides. The hydrophobic elements of the peptides studied here are responsible for the main interaction with HSA. We suggest that albumin binding should be taken into careful consideration in antimicrobial peptide studies, as the systemic distribution can be significantly affected by HSA interactions.     

Predicting inactive conformations of protein kinases using active structures: conformational selection of type-II inhibitors

Protein kinases have been found to possess two characteristic conformations in their activation-loops: the active DFG-in conformation and the inactive DFG-out conformation. Recently, it has been very interesting to develop type-II inhibitors which target the DFG-out conformation and are more specific than the type-I inhibitors binding to the active DFG-in conformation. However, solving crystal structures of kinases with the DFG-out conformation remains a challenge, and this seriously hampers the application of the structure-based approaches in development of novel type-II inhibitors. To overcome this limitation, here we present a computational approach for predicting the DFG-out inactive conformation using the DFG-in active structures, and develop related conformational selection protocols for the uses of the predicted DFG-out models in the binding pose prediction and virtual screening of type-II ligands. With the DFG-out models, we predicted the binding poses for known type-II inhibitors, and the results were found in good agreement with the X-ray crystal structures. We also tested the abilities of the DFG-out models to recognize their specific type-II inhibitors by screening a database of small molecules. The AUC (area under curve) results indicated that the predicted DFG-out models were selective toward their specific type-II inhibitors. Therefore, the computational approach and protocols presented in this study are very promising for the structure-based design and screening of novel type-II kinase inhibitors.