Backbone Atom Positions Research Articles

An end-to-end deep learning method for protein side-chain packing and inverse folding

Protein side-chain packing (PSCP), the task of determining amino acid side-chain conformations given only backbone atom positions, has important applications to protein structure prediction, refinement, and design. Many methods have been proposed to tackle this problem, but their speed or accuracy is still unsatisfactory. To address this, we present AttnPacker, a deep learning (DL) method for directly predicting protein side-chain coordinates. Unlike existing methods, AttnPacker directly incorporates backbone 3D geometry to simultaneously compute all side-chain coordinates without delegating to a discrete rotamer library or performing expensive conformational search and sampling steps. This enables a significant increase in computational efficiency, decreasing inference time by over 100× compared to the DL-based method DLPacker and physics-based RosettaPacker. Tested on the CASP13 and CASP14 native and nonnative protein backbones, AttnPacker computes physically realistic side-chain conformations, reducing steric clashes and improving both rmsd and dihedral accuracy compared to state-of-the-art methods SCWRL4, FASPR, RosettaPacker, and DLPacker. Different from traditional PSCP approaches, AttnPacker can also codesign sequences and side chains, producing designs with subnative Rosetta energy and high in silico consistency.

RNA Bricks—a database of RNA 3D motifs and their interactions

The RNA Bricks database (http://iimcb.genesilico.pl/rnabricks), stores information about recurrent RNA 3D motifs and their interactions, found in experimentally determined RNA structures and in RNA–protein complexes. In contrast to other similar tools (RNA 3D Motif Atlas, RNA Frabase, Rloom) RNA motifs, i.e. ‘RNA bricks’ are presented in the molecular environment, in which they were determined, including RNA, protein, metal ions, water molecules and ligands. All nucleotide residues in RNA bricks are annotated with structural quality scores that describe real-space correlation coefficients with the electron density data (if available), backbone geometry and possible steric conflicts, which can be used to identify poorly modeled residues. The database is also equipped with an algorithm for 3D motif search and comparison. The algorithm compares spatial positions of backbone atoms of the user-provided query structure and of stored RNA motifs, without relying on sequence or secondary structure information. This enables the identification of local structural similarities among evolutionarily related and unrelated RNA molecules. Besides, the search utility enables searching ‘RNA bricks’ according to sequence similarity, and makes it possible to identify motifs with modified ribonucleotide residues at specific positions.

Dynamics of the Streptavidin−Biotin Complex in Solution and in Its Crystal Lattice: Distinct Behavior Revealed by Molecular Simulations

We present a 250 ns simulation of the wild-type, biotin-liganded streptavidin tetramer in the solution phase and compare the trajectory to two previously published simulations of the protein in its crystal lattice. By performing both types of simulations, we are able to interpret the protein's behavior in solution in the context of its X-ray structure. We find that the rate of conformational sampling is increased in solution over the lattice environment, although the relevant conformational space in solution is also much larger, as indicated by overall fluctuations in the positions of backbone atoms. We also compare the distributions of chi1 angles sampled by side chains exposed to solvent in the lattice and in the solution phase, obtaining overall good agreement between the distributions obtained in our most rigorous lattice simulation and the crystallographic chi1 angles. We observe changes in the chi1 distributions in the solution phase, and note an apparent progression of the distributions as the environment changes from a tightly packed lattice filled with crystallization media to a bath of pure water. Finally, we examine the interaction of biotin and streptavidin in each simulation, uncovering a possible alternate conformation of the biotin carboxylate tail. We also note that a hydrogen bond observed to break transiently in previous solution-phase simulations is predominantly broken in this much longer solution-phase trajectory; in the lattice simulations, the lattice environment appears to help maintain the hydrogen bond, but more sampling will be needed to confirm whether the simulation model truly gives good agreement with the X-ray data in the lattice simulations. We expect that pairing solution-phase biomolecular simulations with crystal lattice simulations will help to validate simulation models and improve the interpretation of experimentally determined structures.

Automated Protein Structure Determination from NMR Spectra

Fully automated structure determination of proteins in solution (FLYA) yields, without human intervention, three-dimensional protein structures starting from a set of multidimensional NMR spectra. Integrating existing and new software, automated peak picking over all spectra is followed by peak list filtering, the generation of an ensemble of initial chemical shift assignments, the determination of consensus chemical shift assignments for all (1)H, (13)C, and (15)N nuclei, the assignment of NOESY cross-peaks, the generation of distance restraints, and the calculation of the three-dimensional structure by torsion angle dynamics. The resulting, preliminary structure serves as additional input to the second stage of the procedure, in which a new ensemble of chemical shift assignments and a refined structure are calculated. The three-dimensional structures of three 12-16 kDa proteins computed with the FLYA algorithm coincided closely with the conventionally determined structures. Deviations were below 0.95 A for the backbone atom positions, excluding the flexible chain termini. 96-97% of all backbone and side-chain chemical shifts in the structured regions were assigned to the correct residues. The purely computational FLYA method is suitable for substituting all manual spectra analysis and thus overcomes a main efficiency limitation of the NMR method for protein structure determination.

An improved solution structure for psi-conotoxin PiiiE.

A revised, high-resolution structure of psi-conotoxin Piiie (psi-Piiie), a noncompetitive inhibitor of the nicotinic acetylcholine receptor (nAChR), produced through the use of NMR and molecular modeling calculations is presented. The original structures of psi-Piiie had a relatively high degree of disorder, particularly in the conformation of the disulfide bridges. Our studies utilized (13)C-labeling of selected cysteine residues allowing the resolution of all problems of resonance overlap for the cysteine residues. The improved data were used to produce a new set of structures by a molecular modeling process incorporating relaxation matrix methods for the determination of interproton distance restraints and a combination of distance geometry and simulated annealing for structure generation. The structures produced are very well converged with the RMSD of backbone atom positions of the main body of the peptide improving from 0.73 to 0.13 A. Other indicators of correlation with the experimental data and quality of covalent geometry showed significant improvement in the new structures. The overall conformation of the peptide backbone is similar between the two determinations with the exception of the N-terminus. This difference leads to a significant effect on the predicted distribution of positive charge within psi-Piiie, a property likely to influence interpretation of future mutational studies.

Structure-function relationships in Anabaena ferredoxin: correlations between X-ray crystal structures, reduction potentials, and rate constants of electron transfer to ferredoxin:NADP+ reductase for site-specific ferredoxin mutants.

A combination of structural, thermodynamic, and transient kinetic data on wild-type and mutant Anabaena vegetative cell ferredoxins has been used to investigate the nature of the protein-protein interactions leading to electron transfer from reduced ferredoxin to oxidized ferredoxin:NADP+ reductase (FNR). We have determined the reduction potentials of wild-type vegetative ferredoxin, heterocyst ferredoxin, and 12 site-specific mutants at seven surface residues of vegetative ferredoxin, as well as the one- and two-electron reduction potentials of FNR, both alone and in complexes with wild-type and three mutant ferredoxins. X-ray crystallographic structure determinations have been carried out for six of the ferredoxin mutants. None of the mutants showed significant structural changes in the immediate vicinity of the [2Fe-2S] cluster, despite large decreases in electron-transfer reactivity (for E94K and S47A) and sizable increases in reduction potential (80 mV for E94K and 47 mV for S47A). Furthermore, the relatively small changes in Calpha backbone atom positions which were observed in these mutants do not correlate with the kinetic and thermodynamic properties. In sharp contrast to the S47A mutant, S47T retains electron-transfer activity, and its reduction potential is 100 mV more negative than that of the S47A mutant, implicating the importance of the hydrogen bond which exists between the side chain hydroxyl group of S47 and the side chain carboxyl oxygen of E94. Other ferredoxin mutations that alter both reduction potential and electron-transfer reactivity are E94Q, F65A, and F65I, whereas D62K, D68K, Q70K, E94D, and F65Y have reduction potentials and electron-transfer reactivity that are similar to those of wild-type ferredoxin. In electrostatic complexes with recombinant FNR, three of the kinetically impaired ferredoxin mutants, as did wild-type ferredoxin, induced large (approximately 40 mV) positive shifts in the reduction potential of the flavoprotein, thereby making electron transfer thermodynamically feasible. On the basis of these observations, we conclude that nonconservative mutations of three critical residues (S47, F65, and E94) on the surface of ferredoxin have large parallel effects on both the reduction potential and the electron-transfer reactivity of the [2Fe-2S] cluster and that the reduction potential changes are not the principal factor governing electron-transfer reactivity. Rather, the kinetic properties are most likely controlled by the specific orientations of the proteins within the transient electron-transfer complex.

High Resolution Crystal Structures of Amphibian Red-Cell L Ferritin: Potential Roles for Structural Plasticity and Solvation in Function

Ferritin is a highly conserved multisubunit protein in animals, plants and microbes which assembles with cubic symmetry and transports hydrated iron ions and protons to and from a mineralized core in the protein interior. We report here the high resolution structures of recombinant amphibian red-cell L ferritin and two mutants solved under two sets of conditions. In one mutant, Glu56, 57, 58 and 60 were replaced with Ala, producing a lag phase in the kinetics of iron uptake. In the second mutant, His25 was replaced with Tyr with, at most, subtle effects on function. A molecule of betaine, used in the purification, is bound in all structures at the 2-fold axis near the recently identified heme binding site of bacterioferritin and horse spleen L ferritin. Comparisons of the five amphibian structures identify two regions of the molecule in which conformational flexibility may be related to function. The positions and interactions of a set of 10 to 18 side-chains, most of which are on the inner surface of the protein, are sensitive both to solution conditions and to the Glu→Ala mutation. A subset of these side-chains and a chain of ordered solvent molecules extends from the vicinity of Glu56 to 58 and Glu60 to the 3-fold channel in the wild type protein and may be involved in the transport of either iron or protons. The “spine of hydration” is disrupted in the Glu→Ala mutant. In contrast, H25Y mutation shifts the positions of backbone atoms between the site of the mutation and the 4-fold axis and side-chain positions throughout the structure; the largest changes in the position of backbone atoms are in the DE loop and E Helix, approximately 10 Å from the mutation site. In combination, these results indicate that solvation, structural plasticity and cooperative structural changes may play a role in ferritin function. Analogies with the structure and function of ion channel proteins such as annexins are noted.

High-resolution Three-dimensional Structure of Ribonuclease A in Solution by Nuclear Magnetic Resonance Spectroscopy

High-resolution three-dimensional structures of bovine pancreatic ribonuclease A in aqueous solution have been determined by nuclear magnetic resonance (NMR) spectroscopy combined with restrained molecular dynamics calculations. The structures are based on: (1) 464 interproton distance constraints with accurate upper and lower limits, determined from build-up rates of nuclear Overhauser effects (NOE) by using the complete relaxation matrix; (2) 999 more approximate upper limits for interproton distances; and (3) 42 dihedral angle constraints (37 for φ and 5 for χ1). A total of 16 structures were calculated, which show a root-mean-square (r.m.s.) deviation of 0?66 Å for the backbone atoms and 1·68 Å for all heavy-atoms. The converged structures are highly similar to those found in the crystal state. r.m.s. deviation of backbone atom positions in the crystal as compared to those in the average solution structure is 0·92 Å. Observed differences are concentrated in loop regions and in the neighborhood of His119 and His48 side-chains. Dynamics aspects, such as H/D amide proton exchange and side-chain mobility have been examined.

Determination of a high-quality nuclear magnetic resonance solution structure of the bovine pancreatic trypsin inhibitor and comparison with three crystal structures

A high-quality three-dimensional structure of the bovine pancreatic trypsin inhibitor (BPTI) in aqueous solution was determined by 1H nuclear magnetic resonance (n.m.r.) spectroscopy and compared to the three available high-resolution X-ray crystal structures. A newly collected input of 642 distance constraints derived from nuclear Overhauser effects and 115 dihedral angle constraints was used for the structure calculations with the program DIANA, followed by restrained energy minimization with the program AMBER. The BPTI solution structure is represented by a group of 20 conformers with an average root-mean-square deviation (RMSD) relative to the mean solution structure of 0·43 Å for backbone atoms and 0·92 Å for all heavy atoms of residues 2 to 56. The pairwise RMSD values of the three crystal structures relative to the mean solution structure are 0·76 to 0·85 Å for the backbone atoms and 1·24 to 1·33 Å for all heavy atoms of residues 2 to 56. Small local differences in backbone atom positions between the solution structure and the X-ray structures near residues 9, 25 to 27, 46 to 48 and 52 to 58, and conformational differences for individual amino acid side-chains were analyzed for possible correlations with intermolecular protein-protein contacts in the crystal lattices, using the pairwise RMSD values among the three crystal structures as a reference.

The solution structure of motilin from NMR distance constraints, distance geometry, molecular dynamics, and an iterative full relaxation matrix refinement

A model of the structure of the 22 amino acid residue gastrointestinal peptide hormone motilin in 30% hexafluoro-2-propanol has been obtained by using distance constraints obtained from two-dimensional nuclear Overhauser enhancements. A set of initial structures have been generated by using the distance geometry program DIANA, and 10 of these structures have been refined by using restrained molecular dynamics (AMBER). The resulting structures are virtually indistinguishable in terms of constraint violations and energies and display less than 0.5-A root mean square deviations (RMSD) of the backbone atom positions from Tyr7 to Lys20. A comparison of back-calculated and experimental NOE intensities indicates that RMSD's are not the best indicators of the goodness of fit or of the precision with which the structure is defined. The structure was further refined by fitting the experimental NOE data using an iterative full relaxation matrix analysis. The mean error between the observed and calculated backbone NOE intensities for the final refined structure was 0.23 for the full length of the molecule, 0.18 for the region from Glu9 to Lys20, and 0.29 for the region from Phe1 to Gly8. R factors for the same regions were 0.27, 0.19, and 0.43, respectively. All of the NOE-determined structures consistently display an alpha-helix which extends from Glu9 to Lys20. Considerable lack of definition of structure exists at the amino and carboxyl ends of the molecule and also in the vicinity of Thr6-Tyr7-Gly8. A tendency to form a wide turn appears to exist over the sequence Pro3-Ile4-Phe5-Thr6, but the structure in this region is not well defined by the NOE data.