Explore the vast range of titles available.

The analysis of nuclear magnetic resonance (NMR) spectra for the comprehensive and unambiguous identification and characterization of peaks is a difficult, but critically important step in all NMR analyses of complex biological molecular systems. Here, we introduce DEEP Picker, a deep neural network (DNN)-based approach for peak picking and spectral deconvolution which semi-automates the analysis of two-dimensional NMR spectra. DEEP Picker includes 8 hidden convolutional layers and was trained on a large number of synthetic spectra of known composition with variable degrees of crowdedness. We show that our method is able to correctly identify overlapping peaks, including ones that are challenging for expert spectroscopists and existing computational methods alike. We demonstrate the utility of DEEP Picker on NMR spectra of folded and intrinsically disordered proteins as well as a complex metabolomics mixture, and show how it provides access to valuable NMR information. DEEP Picker should facilitate the semi-automation and standardization of protocols for better consistency and sharing of results within the scientific community. The analysis of NMR spectra of complex biochemical samples with respect to individual resonances is challenging but critically important. Here, the authors present a deep learning-based method that accelerates this process also for crowded NMR data that are non-trivial to analyze, even by expert NMR spectroscopists.

Rapid progress in machine learning offers new opportunities for the automated analysis of multidimensional NMR spectra ranging from protein NMR to metabolomics applications. Most recently, it has been demonstrated how deep neural networks (DNN) designed for spectral peak picking are capable of deconvoluting highly crowded NMR spectra rivaling the facilities of human experts. Superior DNN-based peak picking is one of a series of critical steps during NMR spectral processing, analysis, and interpretation where machine learning is expected to have a major impact. In this perspective, we lay out some of the unique strengths as well as challenges of machine learning approaches in this new era of automated NMR spectral analysis. Such a discussion seems timely and should help define common goals for the NMR community, the sharing of software tools, standardization of protocols, and calibrate expectations. It will also help prepare for an NMR future where machine learning and artificial intelligence tools will be common place.

We mined the most recent editions of the BioMagResDataBank and the protein data bank to parametrize a new empirical knowledge-based chemical shift predictor of protein backbone atoms using either a linear or an artificial neural network model. The resulting chemical shift predictor PPM_One accepts a single static 3D structure as input and emulates the effect of local protein dynamics via interatomic steric contacts. Furthermore, the chemical shift prediction was extended to most side-chain protons and it is found that the prediction accuracy is at a level allowing an independent assessment of stereospecific assignments. For a previously established set of test proteins some overall improvement was achieved over current top-performing chemical shift prediction programs.

Oral cancer is a public health problem with an incidence of almost 50,000 and a mortality of 10,000 each year in the USA alone. Black raspberries (BRBs) have been shown to inhibit oral carcinogenesis in several preclinical models, but our understanding of how BRB phytochemicals affect the metabolic pathways during oral carcinogenesis remains incomplete. We used a well-established rat oral cancer model to determine potential metabolic pathways impacted by BRBs during oral carcinogenesis. F344 rats were exposed to the oral carcinogen 4-nitroquinoline-1-oxide in drinking water for 14 weeks, then regular drinking water for six weeks. Carcinogen exposed rats were fed a 5% or 10% BRB supplemented diet or control diet for six weeks after carcinogen exposure. RNA-Seq transcriptome analysis on rat tongue, and mass spectrometry and NMR metabolomics analysis on rat urine were performed. We tentatively identified 57 differentially or uniquely expressed metabolites and over 662 modulated genes in rats being fed with BRB. Glycolysis and AMPK pathways were modulated during BRB-mediated oral cancer chemoprevention. Glycolytic enzymes Aldoa, Hk2, Tpi1, Pgam2, Pfkl, and Pkm2 as well as the PKA-AMPK pathway genes Prkaa2, Pde4a, Pde10a, Ywhag, and Crebbp were downregulated by BRBs during oral cancer chemoprevention. Furthermore, the glycolysis metabolite glucose-6-phosphate decreased in BRB-administered rats. Our data reveal the novel metabolic pathways modulated by BRB phytochemicals that can be targeted during the chemoprevention of oral cancer.

Elucidation of the mechanism of biomacromolecular recognition events has been a topic of intense interest over the past century. The inherent dynamic nature of both protein and ligand molecules along with the continuous reshaping of the energy landscape during the binding process renders it difficult to characterize this process at atomic detail. Here, we investigate the recognition dynamics of ubiquitin via microsecond all-atom molecular dynamics simulation providing both thermodynamic and kinetic information. The high-level of consistency found with respect to experimental NMR data lends support to the accuracy of the in silico representation of the conformational substates and their interconversions of free ubiquitin. Using an energy-based reweighting approach, the statistical distribution of conformational states of ubiquitin is monitored as a function of the distance between ubiquitin and its binding partner Hrs-UIM. It is found that extensive and dense sampling of conformational space afforded by the µs MD trajectory is essential for the elucidation of the binding mechanism as is Boltzmann sampling, overcoming inherent limitations of sparsely sampled empirical ensembles. The results reveal a population redistribution mechanism that takes effect when the ligand is at intermediate range of 1-2 nm from ubiquitin. This mechanism, which may be depicted as a superposition of the conformational selection and induced fit mechanisms, also applies to other binding partners of ubiquitin, such as the GGA3 GAT domain.

Cooperativity in human glucokinase (GCK), the body’s primary glucose sensor and a major determinant of glucose homeostatic diseases, is fundamentally different from textbook models of allostery because GCK is monomeric and contains only one glucosebinding site. Prior work has demonstrated that millisecond timescale order-disorder transitions within the enzyme’s small domain govern cooperativity. Here, using limited proteolysis, we map the site of disorder in unliganded GCK to a 30-residue active-site loop that closes upon glucose binding. Positional randomization of the loop, coupled with genetic selection in a glucokinase-deficient bacterium, uncovers a hyperactive GCK variant with substantially reduced cooperativity. Biochemical and structural analysis of this loop variant and GCK variants associated with hyperinsulinemic hypoglycemia reveal two distinct mechanisms of enzyme activation. In α-type activation, glucose affinity is increased, the proteolytic susceptibility of the active site loop is suppressed and the ¹H-13C heteronuclear multiple quantum coherence (HMQC) spectrum of13C-Ile–labeled enzyme resembles the glucose-bound state. In β-type activation, glucose affinity is largely unchanged, proteolytic susceptibility of the loop is enhanced, and the ¹H-13C HMQC spectrum reveals no perturbation in ensemble structure. Leveraging both activation mechanisms, we engineer a fully noncooperative GCK variant, whose functional properties are indistinguishable from other hexokinase isozymes, and which displays a 100-fold increase in catalytic efficiency over wild-type GCK. This work elucidates specific structural features responsible for generating allostery in a monomeric enzyme and suggests a general strategy for engineering cooperativity into proteins that lack the structural framework typical of traditional allosteric systems.

There is a critical need to accurately diagnose, prevent, and treat biofilms in humans. The biofilm forming P. aeruginosa bacteria can cause acute and chronic infections, which are difficult to treat due to their ability to evade host defenses along with an inherent antibiotic-tolerance. Using an untargeted NMR-based metabolomics approach, we identified statistically significant differences in 52 metabolites between P. aeruginosa grown in the planktonic and lawn biofilm states. Among them, the metabolites of the cadaverine branch of the lysine degradation pathway were systematically decreased in biofilm. Exogenous supplementation of cadaverine caused significantly increased planktonic growth, decreased biofilm accumulation by 49% and led to altered biofilm morphology, converting to a pellicle biofilm at the air-liquid interface. Our findings show how metabolic pathway differences directly affect the growth mode in P. aeruginosa and could support interventional strategies to control biofilm formation.

Intrinsically disordered proteins (IDPs) are highly dynamic systems that play an important role in cell signaling processes and their misfunction often causes human disease. Proper understanding of IDP function not only requires the realistic characterization of their three-dimensional conformational ensembles at atomic-level resolution but also of the time scales of interconversion between their conformational substates. Large sets of experimental data are often used in combination with molecular modeling to restrain or bias models to improve agreement with experiment. It is shown here for the N-terminal transactivation domain of p53 (p53TAD) and Pup, which are two IDPs that fold upon binding to their targets, how the latest advancements in molecular dynamics (MD) simulations methodology produces native conformational ensembles by combining replica exchange with series of microsecond MD simulations. They closely reproduce experimental data at the global conformational ensemble level, in terms of the distribution properties of the radius of gyration tensor, and at the local level, in terms of NMR properties including 15 N spin relaxation, without the need for reweighting. Further inspection revealed that 10–20% of the individual MD trajectories display the formation of secondary structures not observed in the experimental NMR data. The IDP ensembles were analyzed by graph theory to identify dominant inter-residue contact clusters and characteristic amino-acid contact propensities. These findings indicate that modern MD force fields with residue-specific backbone potentials can produce highly realistic IDP ensembles sampling a hierarchy of nano- and picosecond time scales providing new insights into their biological function.

Despite their importance for biological activity, slower molecular motions beyond the nanosecond range remain poorly understood. We have assembled an unprecedented set of experimental NMR data, comprising up to 27 residual dipolar couplings per amino acid, to define the nature and amplitude of backbone motion in protein G using the Gaussian axial fluctuation model in three dimensions. Slower motions occur in the loops, and in the β-sheet, and are absent in other regions of the molecule, including the α-helix. In the β-sheet an alternating pattern of dynamics along the peptide sequence is found to form a long-range network of slow motion in the form of a standing wave extending across the β-sheet, resulting in maximal conformational sampling at the interaction site. The alternating nodes along the sequence match the alternation of strongly hydrophobic side chains buried in the protein core. Confirmation of the motion is provided through extensive cross-validation and by independent hydrogen-bond scalar coupling analysis that shows this motion to be correlated. These observations strongly suggest that dynamical information can be transmitted across hydrogen bonds and have important implications for understanding collective motions and long-range information transfer in proteins.

Despite the prominent role of the K-Ras protein in many different types of human cancer, major gaps in atomic-level information severely limit our understanding of its functions in health and disease. Here, we report the quantitative backbone structural dynamics of K-Ras by solution nuclear magnetic resonance spectroscopy of the active state of wild-type K-Ras bound to guanosine triphosphate (GTP) nucleotide and two of its oncogenic P-loop mutants, G12D and G12C, using a new nanoparticle-assisted spin relaxation method, relaxation dispersion and chemical exchange saturation transfer experiments covering the entire range of timescales from picoseconds to milliseconds. Our combined experiments allow detection and analysis of the functionally critical Switch I and Switch II regions, which have previously remained largely unobservable by X-ray crystallography and nuclear magnetic resonance spectroscopy. Our data reveal cooperative transitions of K-Ras·GTP to a highly dynamic excited state that closely resembles the partially disordered K-Ras·GDP state. These results advance our understanding of differential GTPase activities and signaling properties of the wild type versus mutants and may thus guide new strategies for the development of therapeutics. Here the authors describe the backbone structural dynamics of K-Ras in its active state by solution NMR. Comparing wild-type K-Ras to oncogenic mutants unveils cooperative transitions to a highly dynamic excited state that advances our understanding of the GTPase activities of K-Ras.