Identifying the interacting positions of a protein using Boolean learning and support vector machines

It is known that in the three-dimensional structure of a protein, certain amino acids can interact with each other in order to provide structural integrity or aid in its catalytic function. If these positions are mutated the loss of this interaction usually leads to a non-functional protein. Directed evolution experiments, which probe the sequence space of a protein through mutations in search for an improved variant, frequently result in such inactive sequences. In this work, we address the use of machine learning algorithms, Boolean learning and support vector machines (SVMs), to find such pairs of amino acid positions. The recombination method of imparting mutations was simulated to create in silico sequences that were used as training data for the algorithms. The two algorithms were combined together to develop an approach that weighs the structural risk as well as the empirical risk to solve the problem. This strategy was adapted to a multi-round framework of experiments where the data generated in the present round is used to design experiments for the next round to improve the generated library, as well as the estimation of the interacting positions. It is observed that this strategy can greatly improve the number of functional variants that are generated as well as the average number of mutations that can be made in the library.     

Effective core potentials and the structures of metallocenes

The use of effective core potentials in the calculation of the geometrical parameters of the ferrocene molecule and its heavier analogs is reported. It is shown that a critical factor in these calculations is the efective core–core (ECC ) potential and, in the absence of ECC s for first-row atoms that are involved in short bonds, calculations of the geometrical parameters are not reliable. Good agreement with experimental geometries may be obtained by using the Los Alamos ECP s for atoms of the second and higher rows of the periodic table at the MP 2 level. DFT calculations have been performed and found to give numerical results comparable to MP 2 in the same basis. © 1995 John Wiley & Sons, Inc.     

Modulating protein structure with fluorous amino acids: increased stability and native-like structure conferred on a 4-helix bundle protein by hexafluoroleucine

There has recently been much interest in exploiting the unusual properties associated with fluorocarbons to modulate the physicochemical properties of proteins. Here we present a detailed investigation into the effect on structure and stability of systematically repacking the hydrophobic core of a model protein with the extensively fluorinated (fluorous) amino acid l-5,5,5,5',5',5'-hexafluoroleucine (hFLeu). The starting point was a 27-residue peptide, alpha(4)-H, that adopts an antiparallel 4-alpha-helix bundle structure, and in which the hydrophobic core comprises six layers of leucine residues introduced at the "a" and "d" positions of the canonical heptad repeat. A series of peptides were synthesized in which the central two (alpha(4)-F(2))(,) four (alpha(4)-F(4)), or all six layers (alpha(4)-F(6)) of the core were substituted hFLeu. The free energy of unfolding increases by 0.3 (kcal/mol)/hFLeu on repacking the central two layers and by an additional 0.12 (kcal/mol)/hFLeu on repacking additional layers, so that alpha(4)-F(6) is approximately 25% more stable than the nonfluorinated protein alpha(4)-H. One-dimensional proton, two-dimensional (1)H-(15)N HSQC, and (19)F NMR spectroscopies were used to examine the effect of fluorination on the conformational dynamics of the peptide. Unexpectedly, increasing the degree of fluorination also appears to result in peptides that possess a more structured backbone and less fluid hydrophobic core. The latter only occurs in alpha(4)-F(4) and alpha(4)-F(6), suggesting that crowding of the hFLeu residues may restrict the amplitude and/or time scales for rotation of the side chains.     

A novel strategy to determine protein structures using exclusively residual dipolar coupling

Recent advances in NMR techniques to measure anisotropic spin interactions such as residual dipolar coupling (RDC) have provided better insights into protein structure as well as dynamics. Exploitation of RDC, however, still remains challenging because its successful application requires a reasonable starting model. Using the singular value decomposition method, we have recently developed an RDC restraint potential to optimally extract orientational information from RDC without the prerequisite of any structural information. In the present study, its efficacy is further illustrated by folding a beta-hairpin and alpha-helix of protein G from extended conformations with RDC restraints alone by employing the replica exchange torsion angle molecular dynamics (REX-TAMD) technique. Subsequently, the entire structure of protein G has been determined accurately using the developed fragment superposition method (FRAGSUM). In FRAGSUM, each overlapping fragments (10 amino acids long) is first folded individually by REX-TAMD, and then the common amino acids are superimposed to determine the entire structure. Because FRAGSUM does not require any additional information besides RDC, it offers a new strategy for de novo structure determination using exclusively RDC.     

Predicting internal protein dynamics from structures using coupled networks of hindered rotators

Internal motions in proteins, such as oscillations of internuclear vectors u(N(i)H(i) (N)) of amide bonds about their equilibrium position, can be characterized by a local order parameter. This dynamic parameter can be determined experimentally by measuring the longitudinal and transverse relaxation rates of (15)N(i) nuclei by suitable NMR methods. In this paper, it is shown that local variations of order parameters S(ii) (2) can be predicted from the knowledge of the structure. To this effect, the diffusive motion of the internuclear vector u(N(i)H(i) (N)) is described in a potential that takes into account the deviations of the angles theta(ij) between u(N(i)H(i) (N)) and neighboring vectors u(N(j)H(j) (N)) from their average value and similarly of deviations of the angles subtended between u(N(i)H(i) (N)) and u(X(j)Y(j)), where X(j) and Y(j) are heavy atoms in the vicinity of the u(N(i)H(i) (N)) vector under investigation. It is shown how the concept of vicinity can be defined by a simple cutoff threshold, i.e., by neglecting vectors u(X(j)Y(j)) with distances d(N(i),X(j))>7.5 A. The local order parameters S(ii) (2) can be predicted from the structure using a limited set of coordinates of heavy atoms. The inclusion of a larger number of heavy atoms does not improve the predictions. Applications to calmodulin, calbindin, and interleukin 4 illustrate the success and limitations of the predictions.     

Identifying i-motif formation using capillary electrophoresis

Over the past few years, intercalated motifs (i-motifs) have attracted attention due to the direct visualization of their existence in the nuclei of human cells. Traditionally, i-motifs have been studied using expensive and complicated NMR, and/or relatively inexpensive but less common circular dichroism spectrometry. The aim of this study was to investigate the feasibility of using less expensive, less complicated, and more widely available CE as an alternative for i-motif related research. The mobilities of two DNA and RNA i-motifs in CE were determined under different pH conditions. Our results demonstrate that CE is able to identify and differentiate mostly folded, partially folded, and mostly unfolded DNA and RNA i-motifs through changes in peak shape and migration time, thus providing a new method to study both i-motif conformation and the interactions between i-motifs and their ligands.     

Similarity searching in files of three-dimensional chemical structures: Representation and searching of molecular electrostatic potentials using field-graphs

This paper reports a method for the identification of those molecules in a database of rigid 3D structures with molecular electrostatic potential (MEP) grids that are most similar to that of a user-defined target molecule. The most important features of an MEP grid are encoded in field-graphs, and a target molecule is matched against a database molecule by a comparison of the corresponding field-graphs. The matching is effected using a maximal common subgraph isomorphism algorithm, which provides an alignment of the target molecule's field- graph with those of each of the database molecules in turn. These alignments are used in the second stage of the search algorithm to calculate the intermolecular MEP similarities. Several different ways of generating field-graphs are evaluated, in terms of the effectiveness of the resulting similarity measures and of the associated computational costs. The most appropriate procedure has been implemented in an operational system that searches a corporate database, containing ca. 173,000 3D structures.     

Three-dimensional visualization of protein expression in mouse brain structures using imaging mass spectrometry

We have developed a method to visualize matrix-assisted laser desorption ionization imaging mass spectrometry (MALDI IMS) data aligned with optically determinable tissue structures in three dimensions. Details of the methodology are exemplified using the 3-D reconstruction of myelin basic protein (MBP) in the corpus callosum of a mouse brain. In this procedure, optical images obtained from serial coronal sections are first aligned to each other to reconstruct a surface of the corpus callosum from segmented contours of the aligned images. The MALDI IMS data are then coregistered to the optical images and superimposed into the surface to create the final 3-D visualization. Correlating proteomic data with anatomical structures provides a more comprehensive understanding of healthy and pathological brain functions, and holds promise to be utilized in more complex anatomical arrangements.     

Identifying G-quadruplex-binding ligands using DNA-functionalized gold nanoparticles

The G-rich overhang of human telomere tends to form a G-quadruplex structure, and G-quadruplex formation can effectively inhibit telomerase activity in most cancer cells. Therefore, it is important to identify the formation and properties of the G-quadruplex, with the particular aim of selecting G-quadruplex-binding ligands that could potentially lead to the development of anticancer therapeutic agents. With this goal in mind, we report a fluorescence resonance energy transfer (FRET) assay system for the identification of G-quadruplex ligands using DNA-functionalized gold nanoparticles (DNA-GNPs) as the fluorescence quencher and a carboxyfluorescein (FAM)-tagged human telomeric sequence (F-GDNA) as the recognition probe. A thiolated complementary strand of human telomeric DNA (cDNA), which first adheres to the surface of the GNPs and then hybridizes with F-GDNA, results in the fluorescence quenching of F-GDNA by the GNPs. However, fluorescence is restored when single-stranded F-GDNA folds into a G-quadruplex structure upon the binding of quadruplex ligands, leading to the release of F-GDNA from the surface of the GNPs. Combined data from fluorescence measurements and CD spectroscopy indicated that ligands selected by this FRET method could induce GDNA to form a G-quadruplex. Therefore, this FRET G-quadruplex assay is a simple and effective approach to identify quadruplex-binding ligands, and, as such, it promises to provide a solid foundation for the development of novel anticancer therapeutic agents.     

一种位点注释的蛋白质数据库用于磷酸化肽段的鉴定

磷酸化修饰的分析一直是蛋白质组学研究的热点之一.在鸟枪法的蛋白质组学研究中,通过在数据库检索中设定磷酸化为可变修饰可以直接鉴定磷酸化修饰的位点.但是翻译后修饰的引入会增加数据检索空间,造成鉴定灵敏度的降低.为了解决这一问题,我们构建了一种位点注释的数据库,这种数据库包含蛋白质的磷酸化位点信息,并开发了一种新的数据库检索策略用于磷酸化肽段的可靠鉴定.用不同类型的数据作为分析对象,通过Mascot检索软件对这种新的数据库检索策略进行了考察,证明了这种方法在保证鉴定结果可靠性的前提下提高了磷酸化肽段鉴定的灵敏度.     

Identifying and localizing intracellular nanoparticles using Raman spectroscopy

Raman microscopy is employed to spectroscopically image biological cells previously exposed to fluorescently labelled polystyrene nanoparticles and, in combination with K-means clustering and principal component analysis (PCA), is demonstrated to be capable of localising the nanoparticles and identifying the subcellular environment based on the molecular spectroscopic signatures. The neutral nanoparticles of 50 nm or 100 nm, as characterised by dynamic light scattering, are shown to be non-toxic to a human lung adenocarcinoma cell-line (A549), according to a range of cytotoxicity assays including Neutral Red, Alamar Blue, Coomassie Blue and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). Confocal fluorescence microscopy identifies intracellular fluorescence due to the nanoparticle exposure, but the fluorescence distribution is spatially diffuse, potentially due to detachment of the dye from the nanoparticles, and the technique fails to unambiguously identify the distribution of the nanoparticles within the cells. Raman spectroscopic mapping of the cells in combination with K-means cluster analysis is used to clearly identify and localise the polystyrene nanoparticles in exposed cells, based on their characteristic spectroscopic signatures. PCA identifies the local environment as rich in lipidic signatures which are associated with localisation of the nanoparticles in the endoplasmic reticulum. The importance of optimised cell growth conditions and fixation processes is highlighted. The preliminary study demonstrates the potential of the technique to unambiguously identify and locate nonfluorescent nanoparticles in cells and to probe not only the local environment but also changes in the cell metabolism which may be associated with cytotoxic responses.     

Identifying reactive trajectories using a moving transition state

A time-dependent no-recrossing dividing surface is shown to lead to a new criterion for identifying reactive trajectories well before they are evolved to infinite time. Numerical dynamics simulations of a dissipative anharmonic two-dimensional system confirm the efficiency of this approach. The results are compared to the standard fixed transition state dividing surface that is well-known to suffer from recrossings and therefore requires trajectories to be evolved over a long time interval before they can reliably be classified as reactive or nonreactive. The moving dividing surface can be used to identify reactive trajectories in harmonic or moderately anharmonic systems with considerably lower numerical effort or even without any simulation at all.     

Impact of different potentials on the structures and energies of clusters

Cluster studies have attracted much interest in the past decades because of their extraordinary properties. To describe the interaction between atoms or molecules and predict the energies and structures, potential functions are developed. However, different potentials generally produce different structures and energies for a cluster. To study the effect of potentials on the structure of a cluster, He clusters in the size range of 13-140 are investigated by Lennard-Jones (LJ), Pirani, and Hartree-Fock-dispersion individual damping (HFD-ID) potential with dynamic lattice searching (DLS) method. Potential function curves, cluster structures, bonds, and energies of the global minima are compared. The results show that cluster energies decrease with the values of the potential functions, the differences between structures depend upon the disagreements of the potentials, and the preferable motif of a cluster changes from icosahedron to decahedron with the increase of the derivative of the short-range part of the potentials.     

Molecular dynamics study of perovskite structures with modified interatomic interaction potentials

The structure of compounds with the perovskite structure ABX₃ (A and B are cations, X are anions O^2—, F^—, Cl^—, Br^—, and I^—), which are widely used in engineering due to unique electrical, optical, and photovoltaic properties, has been considered. Hybrid organic—inorganic halide perovskites important for photovoltaics of a new generation are worth mentioning; they contain cations of organic nitrogen bases as monovalent cations. A molecular dynamics (MD) study of the CaTiO₃ base structure (Ca²⁺, Ti⁴⁺, and O^2—) has been performed in order to develop the methodology of computer simulation and optimization of the shape and parameters of atomic potentials for perovskite systems.     

Identifying mechanisms of interfacial dynamics using single-molecule tracking

The "soft" (i.e., noncovalent) interactions between molecules and surfaces are complex and highly varied (e.g., hydrophobic, hydrogen bonding, and ionic), often leading to heterogeneous interfacial behavior. Heterogeneity can arise either from the spatial variation of the surface/interface itself or from molecular configurations (i.e., conformation, orientation, aggregation state, etc.). By observing the adsorption, diffusion, and desorption of individual fluorescent molecules, single-molecule tracking can characterize these types of heterogeneous interfacial behavior in ways that are inaccessible to traditional ensemble-averaged methods. Moreover, the fluorescence intensity or emission wavelength (in resonance energy transfer experiments) can be used to track the molecular configuration and simultaneously directly relate this to the resulting interfacial mobility or affinity. In this feature article, we review recent advances involving the use of single-molecule tracking to characterize heterogeneous molecule-surface interactions including multiple modes of diffusion and desorption associated with both internal and external molecular configuration, Arrhenius-activated interfacial transport, spatially dependent interactions, and many more.     

Label-free and direct protein detection on 3D plasmonic nanovoid structures using surface-enhanced Raman scattering

In this paper we use surface-enhanced Raman spectroscopy (SERS) on 3D metallic structures for label-free detection and characterization of proteins of interest at low concentrations. The substrates are prepared via nanopatterning with latex nano/microparticles and Cr and Ag sputtering, yielding stable, tunable, and mechanically flexible plasmonic structures. The nanovoids generate a SERS signal of the proteins of interest that is background free and independent of the protein charge. Concentrations as low as 0.05 μg mL⁻¹ could be detected for 4 different proteins. The proteins also exhibit significantly different SERS spectra on these substrates, which is an important feature for future label-free direct detection schemes.     

A benchmark of optimally folded protein structures using integer programming and the 3D-HP-SC model

The Protein Structure Prediction (PSP) problem comprises, among other issues, forecasting the three-dimensional native structure of proteins using only their primary structure information. Most computational studies in this area use synthetic data instead of real biological data. However, the closer to the real-world, the more the impact of results and their applicability. This work presents 17 real protein sequences extracted from the Protein Data Bank for a benchmark to the PSP problem using the tri-dimensional Hydrophobic-Polar with Side-Chains model (3D-HP-SC). The native structure of these proteins was found by maximizing the number of hydrophobic contacts between the side-chains of amino acids. The problem was treated as an optimization problem and solved by means of an Integer Programming approach. Although the method optimally solves the problem, the processing time has an exponential trend. Therefore, due to computational limitations, the method is a proof-of-concept and it is not applicable to large sequences. For unknown sequences, an upper bound of the number of hydrophobic contacts (using this model) can be found, due to a linear relationship with the number of hydrophobic residues. The comparison between the predicted and the biological structures showed that the highest similarity between them was found with distance thresholds around 5.2–8.2 Å. Both the dataset and the programs developed will be freely available to foster further research in the area.     

Mapping low-resolution three-dimensional protein structures using chemical cross-linking and Fourier transform ion-cyclotron resonance mass spectrometry

Techniques in mass spectrometry (MS) combined with chemical cross-linking have proven to be efficient tools for the rapid determination of low-resolution three-dimensional (3-D) structures of proteins. The general procedure involves chemical cross-linking of a protein followed by enzymatic digestion and MS analysis of the resulting peptide mixture. These experiments are generally fast and do not require large quantities of protein. However, the large number of peptide species created from the digestion of cross-linked proteins makes it difficult to identify relevant intermolecular cross-linked peptides from MS data. We present a method for mapping low-resolution 3-D protein structures by combining chemical cross-linking with high-resolution FTICR (Fourier transform ion-cyclotron resonance) mass spectrometry using cytochrome c and hen egg lysozyme as model proteins. We applied several homo-bifunctional, amine-reactive cross-linking reagents that bridge distances from 6 to 16 A. The non-digested cross-linking reaction mixtures were monitored by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS) to determine the extent of cross-linking. Enzymatically digested reaction mixtures were separated by nano-high-performance liquid chromatography (nano-HPLC) on reverse-phase columns applying water/acetonitrile gradients with flow rates of 200 nL/min. The nano-HPLC system was directly coupled to an FTICR mass spectrometer equipped with a nano-ESI (electrospray ionization) source. Cross-linking products were identified using a combination of the GPMAW software and ExPASy Proteomics tools. For correct assignment of the cross-linking products the key factor is to rely on a mass spectrometric method providing both high resolution and high mass accuracy, such as FTICRMS. By combining chemical cross-linking with FTICRMS we were able to rapidly define several intramolecular constraints for cytochrome c and lysozyme.