Protein structure determination in solution by NMR spectroscopy

Challenge for Deep Learning: Protein Structure Prediction of Ligand-Induced Conformational Changes at Allosteric and Orthosteric Sites

In the realm of biomedical research, understanding the intricate structure of proteins is crucial, as these structures determine how proteins function within our bodies and interact with potential drugs. Traditionally, methods like X-ray crystallography and cryo-electron microscopy have been used to unravel these structures, but they are often challenging, time-consuming and costly. Recently, a breakthrough in computational biology has emerged with the development of deep learning algorithms capable of predicting protein structures based on their amino acid sequences (Jumper, J., et al. Nature 2021, 596, 583. Lane, T. J. Nature Methods 2023, 20, 170. Kryshtafovych, A., et al. Proteins: Structure, Function and Bioinformatics 2021, 89, 1607). This study focuses on predicting the dynamic changes that proteins undergo upon ligand binding, specifically when they bind to allosteric sites, i.e. a pocket different from the active site. Allosteric modulators are particularly important for drug discovery, as they open new avenues for designing drugs that can target proteins more effectively and with fewer side effects (Nussinov, R.; Tsai, C. J. Cell 2013, 153, 293). To study this, we curated a data set of 578 X-ray structures comprised of proteins displaying orthosteric and allosteric binding as well as a general framework to evaluate deep learning-based structure prediction methods. Our findings demonstrate the potential and current limitations of deep learning methods, such as AlphaFold2 (Jumper, J., et al. Nature 2021, 596, 583), NeuralPLexer (Qiao, Z., et al. Nat Mach Intell 2024, 6, 195), and RoseTTAFold All-Atom (Krishna, R., et al. Science 2024, 384, eadl2528) to predict not just static protein structures but also the dynamic conformational changes. Herein we show that predicting the allosteric induce-fit conformation still poses a challenge to deep learning methods as they more accurately predict the orthosteric bound conformation compared to the allosteric induce fit conformation. For AlphaFold2, we observed that conformational diversity, and sampling between the apo and holo state could be increased by modifying the MSA depth, but this did not enhance the ability to generate conformations close to the allosteric induced-fit conformation. To further support advancements in protein structure prediction field, the curated data set and evaluation framework are made publicly available.

Correlated rotational alignment spectroscopy: a new tool for high-resolution spectroscopy and the analysis of heterogeneous samples

Correlated rotational alignment spectroscopy correlates observables of ultrafast gas-phase spectroscopy with high-resolution, broad-band rotational Raman spectra. This article reviews the measurement principle of CRASY, existing implementations for mass-correlated measurements, and the potential for future developments. New spectroscopic capabilities are discussed in detail: signals for individual sample components can be separated even in highly heterogeneous samples. Isotopologue rotational spectra can be observed at natural isotope abundance. Fragmentation channels are readily assigned in molecular and cluster mass spectra. And finally, rotational Raman spectra can be measured with sub-MHz resolution, an improvement of several orders-of-magnitude as compared to preceding experiments.

Anti-Staphy Peptides Rationally Designed from Cry10Aa Bacterial Protein

Bacterial infections pose a significant threat to human health, constituting a major challenge for healthcare systems. Antibiotic resistance is particularly concerning in the context of treating staphylococcal infections. In addressing this challenge, antimicrobial peptides (AMPs), characterized by their hydrophobic and cationic properties, unique mechanism of action, and remarkable bactericidal and immunomodulatory capabilities, emerge as promising alternatives to conventional antibiotics for tackling bacterial multidrug resistance. This study focuses on the Cry10Aa protein as a template for generating AMPs due to its membrane-penetrating ability. Leveraging the Joker algorithm, six peptide variants were derived from α-helix 3 of Cry10Aa, known for its interaction with lipid bilayers. In vitro, antimicrobial assays determined the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) required for inhibiting the growth of Staphylococcus aureus, Escherichia coli, Acinetobacter baummanii, Enterobacter cloacae, Enterococcus facallis, Klebsiella pneumonia, and Pseudomonas aeruginosa. Time-kill kinetics were performed using the parental peptide AMPCry10Aa, as well as AMPCry10Aa_1 and AMPCry10Aa_5, against E. coli ATCC, S. aureus 111 and S. aureus ATCC strains showing that AMPCry10Aa_1 and AMPCry10Aa_5 peptides can completely reduce the initial bacterial load with less than 2 h of incubation. AMPCry10Aa_1 and AMPCry 10Aa_5 present stability in human serum and activity maintenance up to 37 °C. Cytotoxicity assays, conducted using the MTT method, revealed that all of the tested peptides exhibited cell viability >50% (IC50). The study also encompassed evaluations of the structure and physical-chemical properties. The three-dimensional structures of AMPCry10Aa and AMPCry10Aa_5 were determined through nuclear magnetic resonance (NMR) spectroscopy, indicating the adoption of α-helical segments. Electron paramagnetic resonance (EPR) spectroscopy elucidated the mechanism of action, demonstrating that AMPCry10Aa_5 enters the outer membranes of E. coli and S. aureus, causing substantial increases in lipid fluidity, while AMPCry10Aa slightly increases lipid fluidity in E. coli. In conclusion, the results obtained underscore the potential of Cry10Aa as a source for developing antimicrobial peptides as alternatives to conventional antibiotics, offering a promising avenue in the battle against antibiotic resistance.

Cellular and epigenetic perspective of protein stability and its implications in the biological system

Protein stability is a fundamental prerequisite in both experimental and therapeutic applications. Current advancements in high throughput experimental techniques and functional ontology approaches have elucidated that impairment in the structure and stability of proteins is intricately associated with the cause and cure of several diseases. Therefore, it is paramount to deeply understand the physical and molecular confounding factors governing the stability of proteins. In this review article, we comprehensively investigated the evolution of protein stability, examining its emergence over time, its relationship with organizational aspects and the experimental methods used to understand it. Furthermore, we have also emphasized the role of Epigenetics and its interplay with post-translational modifications (PTMs) in regulating the stability of proteins.

Structural Elucidation of Ubiquitin via Gas-Phase Ion/Ion Cross-Linking Reactions Using Sodium-Cationized Reagents Coupled with Infrared Multiphoton Dissociation

Accurate structural determination of proteins is critical to understanding their biological functions and the impact of structural disruption on disease progression. Gas-phase cross-linking mass spectrometry (XL-MS) via ion/ion reactions between multiply charged protein cations and singly charged cross-linker anions has previously been developed to obtain low-resolution structural information on proteins. This method significantly shortens experimental time relative to conventional solution-phase XL-MS but has several technical limitations: (1) the singly deprotonated N-hydroxysulfosuccinimide (sulfo-NHS)-based cross-linker anions are restricted to attachment at neutral amine groups of basic amino acid residues and (2) analyzing terminal cross-linked fragment ions is insufficient to unambiguously localize sites of linker attachment. Herein, we demonstrate enhanced structural information for alcohol-denatured A-state ubiquitin obtained from an alternative gas-phase XL-MS approach. Briefly, singly sodiated ethylene glycol bis(sulfosuccinimidyl succinate) (sulfo-EGS) cross-linker anions enable covalent cross-linking at both ammonium and amine groups. Additionally, covalently modified internal fragment ions, along with terminal b-/y-type counterparts, improve the determination of linker attachment sites. Molecular dynamics simulations validate experimentally obtained gas-phase conformations of denatured ubiquitin. This method has identified four cross-linking sites across 8+ ubiquitin, including two new sites in the N-terminal region of the protein that were originally inaccessible in prior gas-phase XL approaches. The two N-terminal cross-linking sites suggest that the N-terminal half of ubiquitin is more compact in gas-phase conformations. By comparison, the two C-terminal linker sites indicate the signature transformation of this region of the protein from a native to a denatured conformation. Overall, the results suggest that the solution-phase secondary structures of the A-state ubiquitin are conserved in the gas phase. This method also provides sufficient sensitivity to differentiate between two gas-phase conformers of the same charge state with subtle structural variations.

KINNTREX: a neural network to unveil protein mechanisms from time-resolved X-ray crystallography

Here, a machine-learning method based on a kinetically informed neural network (NN) is introduced. The proposed method is designed to analyze a time series of difference electron-density maps from a time-resolved X-ray crystallographic experiment. The method is named KINNTREX (kinetics-informed NN for time-resolved X-ray crystallography). To validate KINNTREX, multiple realistic scenarios were simulated with increasing levels of complexity. For the simulations, time-resolved X-ray data were generated that mimic data collected from the photocycle of the photoactive yellow protein. KINNTREX only requires the number of intermediates and approximate relaxation times (both obtained from a singular valued decomposition) and does not require an assumption of a candidate mechanism. It successfully predicts a consistent chemical kinetic mechanism, together with difference electron-density maps of the intermediates that appear during the reaction. These features make KINNTREX attractive for tackling a wide range of biomolecular questions. In addition, the versatility of KINNTREX can inspire more NN-based applications to time-resolved data from biological macromolecules obtained by other methods.

Exploring Kinase Asp-Phe-Gly (DFG) Loop Conformational Stability with AlphaFold2-RAVE

Kinases compose one of the largest fractions of the human proteome, and their misfunction is implicated in many diseases, in particular, cancers. The ubiquitousness and structural similarities of kinases make specific and effective drug design difficult. In particular, conformational variability due to the evolutionarily conserved Asp-Phe-Gly (DFG) motif adopting in and out conformations and the relative stabilities thereof are key in structure-based drug design for ATP competitive drugs. These relative conformational stabilities are extremely sensitive to small changes in sequence and provide an important problem for sampling method development. Since the invention of AlphaFold2, the world of structure-based drug design has noticeably changed. In spite of it being limited to crystal-like structure prediction, several methods have also leveraged its underlying architecture to improve dynamics and enhanced sampling of conformational ensembles, including AlphaFold2-RAVE. Here, we extend AlphaFold2-RAVE and apply it to a set of kinases: the wild type DDR1 sequence and three mutants with single point mutations that are known to behave drastically differently. We show that AlphaFold2-RAVE is able to efficiently recover the changes in relative stability using transferable learned order parameters and potentials, thereby supplementing AlphaFold2 as a tool for exploration of Boltzmann-weighted protein conformations (Meller, A.; Bhakat, S.; Solieva, S.; Bowman, G. R. Accelerating Cryptic Pocket Discovery Using AlphaFold. J. Chem. Theory Comput. 2023, 19, 4355-4363).

High-throughput prediction of protein conformational distributions with subsampled AlphaFold2

This paper presents an innovative approach for predicting the relative populations of protein conformations using AlphaFold 2, an AI-powered method that has revolutionized biology by enabling the accurate prediction of protein structures. While AlphaFold 2 has shown exceptional accuracy and speed, it is designed to predict proteins' ground state conformations and is limited in its ability to predict conformational landscapes. Here, we demonstrate how AlphaFold 2 can directly predict the relative populations of different protein conformations by subsampling multiple sequence alignments. We tested our method against nuclear magnetic resonance experiments on two proteins with drastically different amounts of available sequence data, Abl1 kinase and the granulocyte-macrophage colony-stimulating factor, and predicted changes in their relative state populations with more than 80% accuracy. Our subsampling approach worked best when used to qualitatively predict the effects of mutations or evolution on the conformational landscape and well-populated states of proteins. It thus offers a fast and cost-effective way to predict the relative populations of protein conformations at even single-point mutation resolution, making it a useful tool for pharmacology, analysis of experimental results, and predicting evolution.

Highly Accurate Prediction of NMR Chemical Shifts from Low-Level Quantum Mechanics Calculations Using Machine Learning

Theoretical predictions of NMR chemical shifts from first-principles can greatly facilitate experimental interpretation and structure identification of molecules in gas, solution, and solid-state phases. However, accurate prediction of chemical shifts using the gold-standard coupled cluster with singles, doubles, and perturbative triple excitations [CCSD(T)] method with a complete basis set (CBS) can be prohibitively expensive. By contrast, machine learning (ML) methods offer inexpensive alternatives for chemical shift predictions but are hampered by generalization to molecules outside the original training set. Here, we propose several new ideas in ML of the chemical shift prediction for H, C, N, and O that first introduce a novel feature representation, based on the atomic chemical shielding tensors within a molecular environment using an inexpensive quantum mechanics (QM) method, and train it to predict NMR chemical shieldings of a high-level composite theory that approaches the accuracy of CCSD(T)/CBS. In addition, we train the ML model through a new progressive active learning workflow that reduces the total number of expensive high-level composite calculations required while allowing the model to continuously improve on unseen data. Furthermore, the algorithm provides an error estimation, signaling potential unreliability in predictions if the error is large. Finally, we introduce a novel approach to keep the rotational invariance of the features using tensor environment vectors (TEVs) that yields a ML model with the highest accuracy compared to a similar model using data augmentation. We illustrate the predictive capacity of the resulting inexpensive shift machine learning (iShiftML) models across several benchmarks, including unseen molecules in the NS372 data set, gas-phase experimental chemical shifts for small organic molecules, and much larger and more complex natural products in which we can accurately differentiate between subtle diastereomers based on chemical shift assignments.

Deciphering the structure and mechanism of action of computer-designed mastoparan peptides

Mastoparans are cationic peptides with multifunctional pharmacological properties. Mastoparan-R1 and mastoparan-R4 were computationally designed based on native mastoparan-L from wasps and have improved therapeutic potential for the control of bacterial infections. Here, we evaluated whether these peptides maintain their activity against Escherichia coli strains under a range of salt concentrations. We found that mastoparan-R1 and mastoparan-R4 preserved their activity under the conditions tested, including having antibacterial activities at physiological salt concentrations. The overall structure of the peptides was investigated using circular dichroism spectroscopy in a range of solvents. No significant changes in secondary structure were observed (random coil in aqueous solutions and α-helix in hydrophobic and anionic environments). The three-dimensional structures of mastoparan-R1 and mastoparan-R4 were elucidated through nuclear magnetic resonance spectroscopy, revealing amphipathic α-helical segments for Leu3-Ile13 (mastoparan-R1) and Leu3-Ile14 (mastoparan-R4). Possible membrane-association mechanisms for mastoparan-R1 and mastoparan-R4 were investigated through surface plasmon resonance and leakage studies with synthetic POPC and POPC/POPG (4:1) lipid bilayers. Mastoparan-L had the highest affinity for both membrane systems, whereas the two analogs had weaker association, but improved selectivity for lysing anionic membranes. This finding was also supported by molecular dynamics simulations, in which mastoparan-R1 and mastoparan-R4 were found to have greater interactions with bacteria-like membranes compared with model mammalian membranes. Despite having a few differences in their functional and structural profiles, the mastoparan-R1 analog stood out with the highest activity, greater bacteriostatic potential, and selectivity for lysing anionic membranes. This study reinforces the potential of mastoparan-R1 as a drug candidate.

One bead per residue can describe all-atom protein structures

Atomistic resolution is the standard for high-resolution biomolecular structures, but experimental structural data are often at lower resolution. Coarse-grained models are also used extensively in computational studies to reach biologically relevant spatial and temporal scales. This study explores the use of advanced machine learning networks for reconstructing atomistic models from reduced representations. The main finding is that a single bead per amino acid residue allows construction of accurate and stereochemically realistic all-atom structures with minimal loss of information. This suggests that lower resolution representations of proteins may be sufficient for many applications when combined with a machine learning framework that encodes knowledge from known structures. Practical applications include the rapid addition of atomistic detail to low-resolution structures from experiment or computational coarse-grained models. The application of rapid, deterministic all-atom reconstruction within multi-scale frameworks is further demonstrated with a rapid protocol for the generation of accurate models from cryo-EM densities close to experimental structures.

Structure-based, deep-learning models for protein-ligand binding affinity prediction

The launch of AlphaFold series has brought deep-learning techniques into the molecular structural science. As another crucial problem, structure-based prediction of protein-ligand binding affinity urgently calls for advanced computational techniques. Is deep learning ready to decode this problem? Here we review mainstream structure-based, deep-learning approaches for this problem, focusing on molecular representations, learning architectures and model interpretability. A model taxonomy has been generated. To compensate for the lack of valid comparisons among those models, we realized and evaluated representatives from a uniform basis, with the advantages and shortcomings discussed. This review will potentially benefit structure-based drug discovery and related areas.

Insights into protein structure using cryogenic light microscopy

Fluorescence microscopy has witnessed many clever innovations in the last two decades, leading to new methods such as structured illumination and super-resolution microscopies. The attainable resolution in biological samples is, however, ultimately limited by residual motion within the sample or in the microscope setup. Thus, such experiments are typically performed on chemically fixed samples. Cryogenic light microscopy (Cryo-LM) has been investigated as an alternative, drawing on various preservation techniques developed for cryogenic electron microscopy (Cryo-EM). Moreover, this approach offers a powerful platform for correlative microscopy. Another key advantage of Cryo-LM is the strong reduction in photobleaching at low temperatures, facilitating the collection of orders of magnitude more photons from a single fluorophore. This results in much higher localization precision, leading to Angstrom resolution. In this review, we discuss the general development and progress of Cryo-LM with an emphasis on its application in harnessing structural information on proteins and protein complexes.

Predicting Relative Populations of Protein Conformations without a Physics Engine Using AlphaFold 2

This paper presents a novel approach for predicting the relative populations of protein conformations using AlphaFold 2, an AI-powered method that has revolutionized biology by enabling the accurate prediction of protein structures. While AlphaFold 2 has shown exceptional accuracy and speed, it is designed to predict proteins' ground state conformations and is limited in its ability to predict conformational landscapes. Here, we demonstrate how AlphaFold 2 can directly predict the relative populations of different protein conformations by subsampling multiple sequence alignments. We tested our method against NMR experiments on two proteins with drastically different amounts of available sequence data, Abl1 kinase and the granulocyte-macrophage colony-stimulating factor, and predicted changes in their relative state populations with more than 80% accuracy. Our subsampling approach worked best when used to qualitatively predict the effects of mutations or evolution on the conformational landscape and well-populated states of proteins. It thus offers a fast and cost-effective way to predict the relative populations of protein conformations at even single-point mutation resolution, making it a useful tool for pharmacology, NMR analysis, and evolution.

Optimal control pulses for the 1.2-GHz (28.2-T) NMR spectrometers

The ability to measure nuclear magnetic resonance (NMR) spectra with a large sample volume is crucial for concentration-limited biological samples to attain adequate signal-to-noise (S/N) ratio. The possibility to measure with a 5-mm cryoprobe is currently absent at the 1.2-GHz NMR instruments due to the exceedingly high radio frequency power demands, which is four times compared to 600-MHz instruments. Here, we overcome the high-power demands by designing optimal control (OC) pulses with up to 20 times lower power requirements than currently necessary at a 1.2-GHz spectrometer. We show that multidimensional biomolecular NMR experiments constructed using these OC pulses can bestow improvement in the S/N ratio of up to 26%. With the expected power limitations of a 5-mm cryoprobe, we observe an enhancement in the S/N ratio of more than 240% using our OC sequences. This motivates the development of a cryoprobe with a larger volume than the current 3-mm cryoprobes.

Biomolecular NMR spectroscopy in the era of artificial intelligence

Biomolecular nuclear magnetic resonance (NMR) spectroscopy and artificial intelligence (AI) have a burgeoning synergy. Deep learning-based structural predictors have forever changed structural biology, yet these tools currently face limitations in accurately characterizing protein dynamics, allostery, and conformational heterogeneity. We begin by highlighting the unique abilities of biomolecular NMR spectroscopy to complement AI-based structural predictions toward addressing these knowledge gaps. We then highlight the direct integration of deep learning approaches into biomolecular NMR methods. AI-based tools can dramatically improve the acquisition and analysis of NMR spectra, enhancing the accuracy and reliability of NMR measurements, thus streamlining experimental processes. Additionally, deep learning enables the development of novel types of NMR experiments that were previously unattainable, expanding the scope and potential of biomolecular NMR spectroscopy. Ultimately, a combination of AI and NMR promises to further revolutionize structural biology on several levels, advance our understanding of complex biomolecular systems, and accelerate drug discovery efforts.

A practical guide to machine-learning scoring for structure-based virtual screening

Structure-based virtual screening (SBVS) via docking has been used to discover active molecules for a range of therapeutic targets. Chemical and protein data sets that contain integrated bioactivity information have increased both in number and in size. Artificial intelligence and, more concretely, its machine-learning (ML) branch, including deep learning, have effectively exploited these data sets to build scoring functions (SFs) for SBVS against targets with an atomic-resolution 3D model (e.g., generated by X-ray crystallography or predicted by AlphaFold2). Often outperforming their generic and non-ML counterparts, target-specific ML-based SFs represent the state of the art for SBVS. Here, we present a comprehensive and user-friendly protocol to build and rigorously evaluate these new SFs for SBVS. This protocol is organized into four sections: (i) using a public benchmark of a given target to evaluate an existing generic SF; (ii) preparing experimental data for a target from public repositories; (iii) partitioning data into a training set and a test set for subsequent target-specific ML modeling; and (iv) generating and evaluating target-specific ML SFs by using the prepared training-test partitions. All necessary code and input/output data related to three example targets (acetylcholinesterase, HMG-CoA reductase, and peroxisome proliferator-activated receptor-α) are available at https://github.com/vktrannguyen/MLSF-protocol , can be run by using a single computer within 1 week and make use of easily accessible software/programs (e.g., Smina, CNN-Score, RF-Score-VS and DeepCoy) and web resources. Our aim is to provide practical guidance on how to augment training data to enhance SBVS performance, how to identify the most suitable supervised learning algorithm for a data set, and how to build an SF with the highest likelihood of discovering target-active molecules within a given compound library.

Toward the appropriate interpretation of Alphafold2

In life science, protein is an essential building block for life forms and a crucial catalyst for metabolic reactions in organisms. The structures of protein depend on an infinity of amino acid residues' complex combinations determined by gene expression. Predicting protein folding structures has been a tedious problem in the past seven decades but, due to robust development of artificial intelligence, astonishing progress has been made. Alphafold2, whose key component is Evoformer, is a typical and successful example of such progress. This article attempts to not only isolate and dissect every detail of Evoformer, but also raise some ideas for potential improvement.

Protein structure prediction from the complementary science perspective

A comparative analysis between two problems-apparently unrelated-which are solved in a period of ~400 years, viz., the accurate prediction of both the planetary orbits and the protein structures, leads to inferred conjectures that go far beyond the existence of a common path in their resolution, i.e., observation → pattern recognition → modeling. The preliminary results from this analysis indicate that complementary science, together with a new perspective on protein folding, may help us discover common features that could contribute to a more in-depth understanding of still-unsolved problems such as protein folding.

Fragment-based screening by protein-detected NMR spectroscopy

Fragment-based drug discovery (FBDD) identifies low molecular weight compounds that can be developed into ligands with high affinity and selectivity for therapeutic targets. Screening fragment libraries (<10,000 molecules) with biophysical techniques against macromolecules provides information about novel chemical spaces that bind the macromolecule and scaffolds that can be modified to increase potency. A fragment-screening pipeline requires a standardized protocol for target selection, library assembly and maintenance, library screening, and hit validation to ensure hit integrity. Herein, the fundamental aspects of a fragment screening pipeline-focusing on protein-detected NMR data collection and analysis-are discussed in detail for researchers to use as a resource in their FBDD projects. Selected screening targets must undergo rigorous stability and buffer testing by NMR spectroscopy to ensure the protein structure is stable for the entire screen. Biophysical instrumentation that rapidly measures protein thermostability is helpful in buffer screening. Molecules in fragment libraries are analyzed computationally and physically, stored at appropriate temperatures, and multiplexed in well plates for library conservation. The screening protocol is streamlined using liquid handling robotics for sample preparation and customized Python scripts for protein-detected NMR data analysis. Molecules identified from the screen are titrated to determine their binding site(s) and Kd values and confirmed with an orthogonal biophysical assay. This detailed FBDD screening pipeline developed by the Program in Chemical Biology at the Medical College of Wisconsin has successfully screened many unrelated target proteins to identified novel molecules that selectively bind to these target proteins.

Structure-based pathogenicity relationship identifier for predicting effects of single missense variants and discovery of higher-order cancer susceptibility clusters of mutations

We report the structure-based pathogenicity relationship identifier (SPRI), a novel computational tool for accurate evaluation of pathological effects of missense single mutations and prediction of higher-order spatially organized units of mutational clusters. SPRI can effectively extract properties determining pathogenicity encoded in protein structures, and can identify deleterious missense mutations of germ line origin associated with Mendelian diseases, as well as mutations of somatic origin associated with cancer drivers. It compares favorably to other methods in predicting deleterious mutations. Furthermore, SPRI can discover spatially organized pathogenic higher-order spatial clusters (patHOS) of deleterious mutations, including those of low recurrence, and can be used for discovery of candidate cancer driver genes and driver mutations. We further demonstrate that SPRI can take advantage of AlphaFold2 predicted structures and can be deployed for saturation mutation analysis of the whole human proteome.

Disruption of Spatial Proximities among Charged Groups in Equilibrium-Denatured States of Proteins Tracked Using Protein Charge Transfer Spectra

The absorption and luminescence originating from protein charge transfer spectra (ProCharTS) depend on the proximity between multiple charged groups in a protein. This makes ProCharTS absorbance/luminescence intensity a sensitive probe for detecting changes in the protein structure, which alter the proximity among charged groups in the protein. In this work, ProCharTS absorbance of charge-rich proteins like human serum albumin (HSA), α₃C, and α₃W was used to monitor structural changes upon chemical denaturant-induced protein unfolding under equilibrium conditions. The denaturation midpoints were estimated using nonlinear regression analysis. For HSA, absorbance at 325 and 340 nm estimated the GdnHCl-induced denaturation midpoints to be 0.80 and 0.61 M, respectively. A similar analysis of α₃C and α₃W ProCharTS absorbance yielded denaturation midpoints of 0.88 and 0.86 M at 325 nm and 0.96 and 0.66 M at 340 nm, respectively. A previously reported molten globule-like state in the GdnHCl-induced HSA unfolding pathway was detected by the increase in HSA ProCharTS absorbance at 0.5 M GdnHCl. To validate the above results, protein unfolding was additionally monitored using conventional methods like circular dichroism (CD), Trp, and dansyl fluorescence. Our results suggest that disruption of charged amino acid sidechain contacts as revealed by ProCharTS occurs at lower denaturant concentrations compared to the loss of secondary/folded structure monitored by CD and fluorescence. Further, HSA ProCharTS absorbance at 315-340 nm revealed that tertiary contacts among charged residues were disrupted at lower GdnHCl concentrations compared to sequence adjacent contacts. Our data underscore the utility of ProCharTS as a novel label-free tool to track unfolding in charge-rich proteins.

Rethinking the protein folding problem from a new perspective

One of the main concerns of Anfinsen was to reveal the connection between the amino-acid sequence and their biologically active conformation. This search gave rise to two crucial questions in structural biology, namely, why the proteins fold and how a sequence encodes its folding. As to the why, he proposes a plausible answer, namely, the thermodynamic hypothesis. As to the how, this remains an unsolved challenge. Consequently, the protein folding problem is examined here from a new perspective, namely, as an 'analytic whole'. Conceiving the protein folding in this way enabled us to (i) examine in detail why the force-field-based approaches have failed, among other purposes, in their ability to predict the three-dimensional structure of a protein accurately; (ii) propose how to redefine them to prevent these shortcomings, and (iii) conjecture on the origin of the state-of-the-art numerical-methods success to predict the tridimensional structure of proteins accurately.

Visualizing Heterogeneous Protein Conformations with Multi-Tilt Nanoparticle-Aided Cryo-Electron Microscopy Sampling

Obtaining the heterogeneous conformation of small proteins is important for understanding their biological role, but it is still challenging. Here, we developed a multi-tilt nanoparticle-aided cryo-electron microscopy sampling (MT-NACS) technique that enables the observation of heterogeneous conformations of small proteins and applied it to calmodulin. By imaging the proteins labeled by two gold nanoparticles at multiple tilt angles and analyzing the projected positions of the nanoparticles, the distributions of 3D interparticle distances were obtained. From the measured distance distributions, the conformational changes associated with Ca²⁺ binding and salt concentration were determined. MT-NACS was also used to track the structural change accompanied by the interaction between amyloid-beta and calmodulin, which has never been observed experimentally. This work offers an alternative platform for studying the functional flexibility of small proteins.

Breaking free from the crystal lattice: Structural biology in solution to study protein degraders

Structural biology offers a versatile arsenal of techniques and methods to investigate the structure and conformational dynamics of proteins and their assemblies. The growing field of targeted protein degradation centres on the premise of developing small molecules, termed degraders, to induce proximity between an E3 ligase and a protein of interest to be signalled for degradation. This new drug modality brings with it new opportunities and challenges to structural biologists. Here we discuss how several structural biology techniques, including nuclear magnetic resonance, cryo-electron microscopy, structural mass spectrometry and small angle scattering, have been explored to complement X-ray crystallography in studying degraders and their ternary complexes. Together the studies covered in this review make a case for the invaluable perspectives that integrative structural biology techniques in solution can bring to understanding ternary complexes and designing degraders.

Quantitatively Visualizing Bipartite Datasets

As experiments continue to increase in size and scope, a fundamental challenge of subsequent analyses is to recast the wealth of information into an intuitive and readily interpretable form. Often, each measurement conveys only the relationship between a pair of entries, and it is difficult to integrate these local interactions across a dataset to form a cohesive global picture. The classic localization problem tackles this question, transforming local measurements into a global map that reveals the underlying structure of a system. Here, we examine the more challenging bipartite localization problem, where pairwise distances are available only for bipartite data comprising two classes of entries (such as antibody-virus interactions, drug-cell potency, or user-rating profiles). We modify previous algorithms to solve bipartite localization and examine how each method behaves in the presence of noise, outliers, and partially observed data. As a proof of concept, we apply these algorithms to antibody-virus neutralization measurements to create a basis set of antibody behaviors, formalize how potently inhibiting some viruses necessitates weakly inhibiting other viruses, and quantify how often combinations of antibodies exhibit degenerate behavior.

AlphaFold2 and its applications in the fields of biology and medicine

AlphaFold2 (AF2) is an artificial intelligence (AI) system developed by DeepMind that can predict three-dimensional (3D) structures of proteins from amino acid sequences with atomic-level accuracy. Protein structure prediction is one of the most challenging problems in computational biology and chemistry, and has puzzled scientists for 50 years. The advent of AF2 presents an unprecedented progress in protein structure prediction and has attracted much attention. Subsequent release of structures of more than 200 million proteins predicted by AF2 further aroused great enthusiasm in the science community, especially in the fields of biology and medicine. AF2 is thought to have a significant impact on structural biology and research areas that need protein structure information, such as drug discovery, protein design, prediction of protein function, et al. Though the time is not long since AF2 was developed, there are already quite a few application studies of AF2 in the fields of biology and medicine, with many of them having preliminarily proved the potential of AF2. To better understand AF2 and promote its applications, we will in this article summarize the principle and system architecture of AF2 as well as the recipe of its success, and particularly focus on reviewing its applications in the fields of biology and medicine. Limitations of current AF2 prediction will also be discussed.

Full opening of helix bundle crossing does not lead to NaK channel activation

A critical part of ion channel function is the ability to open and close in response to stimuli and thus conduct ions in a regulated fashion. While x-ray diffraction studies of ion channels suggested a general steric gating mechanism located at the helix bundle crossing (HBC), recent functional studies on several channels indicate that the helix bundle crossing is wide-open even in functionally nonconductive channels. Two NaK channel variants were crystallized in very different open and closed conformations, which served as important models of the HBC gating hypothesis. However, neither of these NaK variants is conductive in liposomes unless phenylalanine 92 is mutated to alanine (F92A). Here, we use NMR to probe distances at near-atomic resolution of the two NaK variants in lipid bicelles. We demonstrate that in contrast to the crystal structures, both NaK variants are in a fully open conformation, akin to Ca2+-bound MthK channel structure where the HBC is widely open. While we were not able to determine what a conductive NaK structure is like, our further inquiry into the gating mechanism suggests that the selectivity filter and pore helix are coupled to the M2 helix below and undergo changes in the structure when F92 is mutated. Overall, our data show that NaK exhibits coupling between the selectivity filter and HBC, similar to K+ channels, and has a more complex gating mechanism than previously thought, where the full opening of HBC does not lead to channel activation.

Advances in antibody discovery from human BCR repertoires

Antibodies make up an important and growing class of compounds used for the diagnosis or treatment of disease. While traditional antibody discovery utilized immunization of animals to generate lead compounds, technological innovations have made it possible to search for antibodies targeting a given antigen within the repertoires of B cells in humans. Here we group these innovations into four broad categories: cell sorting allows the collection of cells enriched in specificity to one or more antigens; BCR sequencing can be performed on bulk mRNA, genomic DNA or on paired (heavy-light) mRNA; BCR repertoire analysis generally involves clustering BCRs into specificity groups or more in-depth modeling of antibody-antigen interactions, such as antibody-specific epitope predictions; validation of antibody-antigen interactions requires expression of antibodies, followed by antigen binding assays or epitope mapping. Together with innovations in Deep learning these technologies will contribute to the future discovery of diagnostic and therapeutic antibodies directly from humans.

Eraldo Antonini Lectures, 1983-2019

“Can order spring from Chaos?” is the title of an extensive Report on Italian science published by NATURE on 12 May 1983 and written by Robert Walgate, the Chief European Correspondent. It is a twenty pages complete paper touching all aspects of the struggle of Italian scientists to work in the “curious amalgam of ingenuity and muddle, a reflection of the political system”. (Nature, 1983; 303: 109–128). To read it after four decades is interesting but somewhat depressing since the main problems unfolded in the paper have not been solved, starting with the largely insufficient support of fundamental curiosity driven research. At page 114 you could find a item called: ITALY’s TOP SCIENTISTS: Four in the top one thousand. The Author refers to the data reported by the ISI (Institute of Scientific Information) that took two years to scan 3,000 major journals over the period 1965–78 and covered 5 millions articles and 67 millions references. The four top Italian scientists working in Italy were: Eraldo Antonini (3127 citations), Enrico Clementi (4001), Silvio Garattini (2833), and Giorgio Giacomelli (2483); 3 out of four were 52 years old, and one 55. Antonini did not see the Report since he passed away on March 18, 1983. However the information leaked before the publication of Nature because I remember the Messaggero of Rome reporting a whole page with the ranking of the four Italians, and even a picture of Eraldo. The students of the first year Medical course, his Class, welcomed the Professor with a standing ovation. After a short time the Board of the SIB (Società Italiana di Biochimica) casted a unanimous vote in favour of the motion of President Noris Siliprandi to begin the annual Congress with an Antonini Lecture, forever. As reported below, the tradition began immediately at the Congress in Saint-Vicent, Italy, and is continuing. In this paper I report an account of the Eraldo Antonini Lectures that I attended over the years and until September 2019, a few months before the pandemics lock down.

Optimization and validation of multi-state NMR protein structures using structural correlations

Recent advances in the field of protein structure determination using liquid-state NMR enable the elucidation of multi-state protein conformations that can provide insight into correlated and non-correlated protein dynamics at atomic resolution. So far, NMR-derived multi-state structures were typically evaluated by means of visual inspection of structure superpositions, target function values that quantify the violation of experimented restraints and root-mean-square deviations that quantify similarity between conformers. As an alternative or complementary approach, we present here the use of a recently introduced structural correlation measure, PDBcor, that quantifies the clustering of protein states as an additional measure for multi-state protein structure analysis. It can be used for various assays including the validation of experimental distance restraints, optimization of the number of protein states, estimation of protein state populations, identification of key distance restraints, NOE network analysis and semiquantitative analysis of the protein correlation network. We present applications for the final quality analysis stages of typical multi-state protein structure calculations.

Early aggregation mechanism of Aβ16-22 revealed by Markov state models

Aβ_16-22 is believed to have critical role in early aggregation of full length amyloids that are associated with the Alzheimer's disease and can aggregate to form amyloid fibrils. However, the early aggregation mechanism is still unsolved. Here, multiple long-term molecular dynamics simulations combining with Markov state model were used to probe the early oligomerization mechanism of Aβ_16-22 peptides. The identified dimeric form adopted either globular random-coil or extended β-strand like conformations. The observed dimers of these variants shared many overall conformational characteristics but differed in several aspects at detailed level. In all cases, the most common type of secondary structure was intermolecular antiparallel β-sheets. The inter-state transitions were very frequent ranges from few to hundred nanoseconds. More strikingly, those states which contain fraction of β secondary structure and significant amount of extended coiled structures, therefore exposed to the solvent, were majorly participated in aggregation. The assembly of low-energy dimers, in which the peptides form antiparallel β sheets, occurred by multiple pathways with the formation of an obligatory intermediates. We proposed that these states might facilitate the Aβ_16-22 aggregation through a significant component of the conformational selection mechanism, because they might increase the aggregates population by promoting the inter-chain hydrophobic and the hydrogen bond contacts. The formation of early stage antiparallel β sheet structures is critical for oligomerization, and at the same time provided a flat geometry to seed the ordered β-strand packing of the fibrils. Our findings hint at reorganization of this part of the molecule as a potentially critical step in Aβ aggregation and will insight into early oligomerization for large β amyloids.

Measurement of Secondary Structure Changes in Poly-L-lysine and Lysozyme during Acoustically Levitated Single Droplet Drying Experiments by In Situ Raman Spectroscopy

Drying processes such as spray drying, as commonly used in the pharmaceutical industry to convert protein-based drugs into their particulate form, can lead to an irreversible loss of protein activity caused by protein secondary structure changes. Due to the nature of these processes (high droplet number, short drying time), an in situ investigation of the structural changes occurring during a real drying process is hardly possible. Therefore, an approach for the in situ investigation of the expected secondary structural changes during single droplet protein drying in an acoustic levitator by time-resolved Raman spectroscopy was developed and is demonstrated in this paper. For that purpose, a self-developed NIR–Raman sensor generates and detects the Raman signal from the levitated solution droplet. A mathematical spectral reconstruction by multiple Voigt functions is used to quantify the relative secondary structure changes occurring during the drying process. With the developed setup, it was possible to detect and quantify the relative secondary structure changes occurring during single droplet drying experiments for the two chosen model substances: poly-L-lysine, a homopolypeptide widely used as a protein mimic, and lysozyme. Throughout drying, an increase in the β-sheet structure and a decrease in the other two structural elements, α-helix, and random coil, could be identified. In addition, it was observed that the degree of structural changes increased with increasing temperature.

Positional SHAP (PoSHAP) for Interpretation of machine learning models trained from biological sequences

Machine learning with multi-layered artificial neural networks, also known as "deep learning," is effective for making biological predictions. However, model interpretation is challenging, especially for sequential input data used with recurrent neural network architectures. Here, we introduce a framework called "Positional SHAP" (PoSHAP) to interpret models trained from biological sequences by utilizing SHapely Additive exPlanations (SHAP) to generate positional model interpretations. We demonstrate this using three long short-term memory (LSTM) regression models that predict peptide properties, including binding affinity to major histocompatibility complexes (MHC), and collisional cross section (CCS) measured by ion mobility spectrometry. Interpretation of these models with PoSHAP reproduced MHC class I (rhesus macaque Mamu-A1*001 and human A*11:01) peptide binding motifs, reflected known properties of peptide CCS, and provided new insights into interpositional dependencies of amino acid interactions. PoSHAP should have widespread utility for interpreting a variety of models trained from biological sequences.

Membrane Molecular Interactions and Induced Structures of CPPs

Cell penetrating peptides (CPPs) are generally defined as short positively charged peptides, containing 5-30 amino acids. Based on their physicochemical properties, they are classified as three main groups, namely hydrophobic, amphipathic, and hydrophilic. They are capable of interacting with the cell membrane without inducing serious toxicity, and they can carry cargo molecules across the membrane. Cargo molecules could be different therapeutics which makes CPPs valuable in the field of drug delivery into living cells. Nowadays, CPPs are considered as potential parts of therapeutics against several diseases.Despite similarities in their primary structure, the interactions of CPPs with a cell membrane may vary a lot. This is even more complicated when the CPP is bound to the cargo molecule. The mechanism(s) of their cellular uptake and endosomal escape have not been completely resolved. Understanding the mechanism of membrane interaction will help us designing a CPP with enhanced, selective cargo delivery, hopefully resulting in better disease treatments. So far energy independent direct membrane penetration and energy-dependent endocytosis have been suggested as two main mechanisms of cellular entry for CPPs, and both may be applicable for the same CPP-complex, depending on the conditions.In order to understand which mechanism is associated with a particular CPP 's cellular uptake in a particular cell (sometimes including endosomal escape), different biological and biophysical methods and strategies have been applied. In this chapter, we will address several biophysical methods, such as fluorescence spectroscopy, circular dichroism (CD) spectroscopy, dynamic light scattering, and NMR .We also review different membrane model systems which are suitable for the biophysical studies. These include large unilamellar phospholipid vesicles (LUVs ), which are the most commonly used in the lipid-peptide interaction studies. Detergent micelles and mixed micelles (bicelles) are also suitable membrane model systems, particularly in high-resolution NMR studies.

Automatic Assembly and Calibration of Models of Enzymatic Reactions Based on Ordinary Differential Equations

Enzymatic reactions have been studied for more than a 100 years. Indeed, isolation of enzymes from biological materials is no longer the main source of enzymes today, as they are now largely produced using recombinant technology, or can even be synthesized from scratch. Studies of the three-dimensional structures of enzymes can provide answers to many questions, but the kinetics of enzymatic reactions is the only method that can lead to better understanding of their function. The complexity of high-throughput analysis of progress curves of data obtained can only be achieved through numerical solutions of a suitable system of ordinary differential equations. We have developed the freely available server ENZO: a web tool for derivation and evaluation of kinetic models of enzyme-catalyzed reactions ( http://enzo.cmm.ki.si/ ). ENZO can be used for simultaneous analysis of a series of progress curves obtained under many different conditions. In this chapter, we exemplify the principles and possibilities of this type of high-throughput analysis.

Membrane Structure of Aquaporin Observed with Combined Experimental and Theoretical Sum Frequency Generation Spectroscopy

High-resolution structural information on membrane proteins is essential for understanding cell biology and for the structure-based design of new medical drugs and drug delivery strategies. X-ray diffraction (XRD) can provide angstrom-level information about the structure of membrane proteins, yet for XRD experiments, proteins are removed from their native membrane environment, chemically stabilized, and crystallized, all of which can compromise the conformation. Here, we describe how a combination of surface-sensitive vibrational spectroscopy and molecular dynamics simulations can account for the native membrane environment. We observe the structure of a glycerol facilitator channel (GlpF), an aquaporin membrane channel finely tuned to selectively transport water and glycerol molecules across the membrane barrier. We find subtle but significant differences between the XRD structure and the inferred in situ structure of GlpF.

Probing Polymer Chain Folding in Solution Using Second Harmonic Light Scattering

Periodically grafted amphiphilic copolymers (PGACs) were earlier shown by us to adopt a zigzag folded conformation in the solid state, which enabled the backbone and pendant segments to segregate and occupy alternate layers in a lamellar structure. The conformational transition from a random coil to a zigzag folded chain in solution is an interesting problem, which is largely unexplored. To examine this, an orthogonally clickable parent polyester was sequentially clicked with two types of poly(ethylene glycol) (PEG) segments: one is a simple PEG and the other is a PEG that carries a dipolar chromophore. These two hydrophilic PEG segments, installed in a periodic and alternating fashion along the hydrocarbon-rich (HC) polyester backbone, ensure that the Janus folded chains are formed upon folding and carry chromophoric dipoles oriented along the same direction, thereby generating a large net dipole. The folding-induced alignment of chromophores in solution was followed using second harmonic light scattering (SHLS), wherein the intensity of the frequency-doubled scattered light (I_2ω) is measured. Folding was induced by adding a polar solvent, like methanol, to a chloroform solution of the polymer; methanol desolvates the HC backbone but solubilizes the pendant PEG segments, thus inducing folding. The second harmonic intensity (I_2ω) increased initially with methanol concentration and then saturated; in contrast, I_2ω remained invariant with the solvent composition in the case of an analogous model chromophore. Furthermore, in a model PGAC carrying chromophore-bearing PEG segments on every repeat unit, I_2ω decreased with increasing methanol composition, revealing the formation of a centrosymmetric folded chain, wherein the chromophoric dipoles on either side cancel each other. Thus, this study clearly reveals that the zigzag chain folding of PGACs can be induced by a segment-selective solvent, resulting in the rather elusive directional ordering of chromophoric dipoles in solution.

Evaluation of the discriminatory potential of antibodies created from synthetic peptides derived from wheat, barley, rye and oat gluten

Celiac disease (CD) is triggered by ingestion of gluten-containing cereals such as wheat, barley, rye and in some cases oat. The only way for affected individuals to avoid symptoms of this condition is to adopt a gluten-free diet. Thus, gluten-free foodstuffs need to be monitored in order to ensure their innocuity. For this purpose, commercial immunoassays based on recognition of defined linear gluten sequences are currently used. These immunoassays are designed to detect or quantify total gluten regardless of the cereal, and often result in over or underestimation of the exact gluten content. In addition, Canadian regulations require a declaration of the source of gluten on the label of prepackaged foods, which cannot be done due to the limitations of existing methods. In this study, the development of new antibodies targeting discrimination of gluten sources was conducted using synthetic peptides as immunization strategy. Fourteen synthetic peptides selected from unique linear amino acid sequences of gluten were bioconjugated to Concholepas concholepas hemocyanin (CCH) as protein carrier, to elicit antibodies in rabbit. The resulting polyclonal antibodies (pAbs) successfully discriminated wheat, barley and oat prolamins during indirect ELISA assessments. pAbs raised against rye synthetic peptides cross-reacted evenly with wheat and rye prolamins but could still be useful to successfully discriminate gluten sources in combination with the other pAbs. Discrimination of gluten sources can be further refined and enhanced by raising monoclonal antibodies using a similar immunization strategy. A methodology capable of discriminating gluten sources, such as the one proposed in this study, could facilitate compliance with Canadian regulations on this matter. This type of discrimination could also complement current immunoassays by settling the issue of over and underestimation of gluten content, thus improving the safety of food intended to CD and wheat-allergic patients.

Protein and solutes freeze-concentration in water/glycerol mixtures revealed by pulse EPR

Lyophilization can extend protein drugs stability and shelf life, but it also can lead to protein degradation in some cases. The development of safe freeze-drying approaches for sensitive proteins requires a better understanding of lyophilization on the molecular level. The evaluation of the freezing process and its impact on the protein environment in the nm scale is challenging because feasible experimental methods are scarce. In the present work we apply pulse EPR as a tool to study the local concentrations of the solute in the 20 nm range and of the solvent in the 1 nm range for a spin labeled 27 kDa monomeric green fluorescent protein, mEGFP, and the 172 Da TEMPOL spin probe, frozen in different water/glycerol-d5 mixtures. For average glycerol volume fractions, φ_gly-d5^avg, ≥ 0.4 we observed transparent glassy media; the local concentration and the 1 nm solvent shell of TEMPOL and the protein correspond to those of a uniform vitrified glass. At φ_gly-d5^avg ≤ 0.3 we observed partial ice crystallization, which led to ice exclusion of glycerol and TEMPOL with freeze-concentration up to the glycerol maximal-freeze local volume fraction, φ_gly-d5^loc, of 0.64. The protein concentration and its shell behavior was similar except for the lowest φ_gly-d5^avg (0.1), which showed a 4.7-fold freeze-concentration factor compared to sevenfold for TEMPOL, and also a smaller φ_gly-d5^loc. We explain this behavior with an increased probability for proteins to get stuck in the ice phase during fast freezing at higher freeze-concentration and the related large-scale mass transfer.

NMR hawk-eyed view of AlphaFold2 structures

The prediction of the three‐dimensional (3D) structure of proteins from the amino acid sequence made a stunning breakthrough reaching atomic accuracy. Using the neural network‐based method AlphaFold2, 3D structures of almost the entire human proteome have been predicted and made available (https://www.alphafold.ebi.ac.uk). To gain insight into how well AlphaFold2 structures represent the conformation of proteins in solution, I here compare the AlphaFold2 structures of selected small proteins with their 3D structures that were determined by nuclear magnetic resonance (NMR) spectroscopy. Proteins were selected for which the 3D solution structures were determined on the basis of a very large number of distance restraints and residual dipolar couplings and are thus some of the best‐resolved solution structures of proteins to date. The quality of the backbone conformation of the AlphaFold2 structures is assessed by fitting a large set of experimental residual dipolar couplings (RDCs). The analysis shows that experimental RDCs fit extremely well to the AlphaFold2 structures predicted for GB3, DinI, and ubiquitin. In the case of GB3, the accuracy of the AlphaFold2 structure even surpasses that of a 1.1 Å crystal structure. Fitting of experimental RDCs furthermore allows identification of AlphaFold2 structures that are best representative of the protein's conformation in solution as seen for the EF hands of the N‐terminal domain of Ca²⁺‐ligated calmodulin. Taken together, the analysis shows that structures predicted by AlphaFold2 can be highly representative of the solution conformation of proteins. The combination of AlphaFold2 structures with RDCs promises to be a powerful approach to study structural changes in proteins.

Integrative Structural Biology in the Era of Accurate Structure Prediction

Characterizing the three-dimensional structure of macromolecules is central to understanding their function. Traditionally, structures of proteins and their complexes have been determined using experimental techniques such as X-ray crystallography, NMR, or cryo-electron microscopy-applied individually or in an integrative manner. Meanwhile, however, computational methods for protein structure prediction have been improving their accuracy, gradually, then suddenly, with the breakthrough advance by AlphaFold2, whose models of monomeric proteins are often as accurate as experimental structures. This breakthrough foreshadows a new era of computational methods that can build accurate models for most monomeric proteins. Here, we envision how such accurate modeling methods can combine with experimental structural biology techniques, enhancing integrative structural biology. We highlight the challenges that arise when considering multiple structural conformations, protein complexes, and polymorphic assemblies. These challenges will motivate further developments, both in modeling programs and in methods to solve experimental structures, towards better and quicker investigation of structure-function relationships.

Structural Changes in Insulin at a Soft Electrochemical Interface

Understanding the interaction of proteins at interfaces, which occurs at or within cell membranes and lipoprotein vesicles, is central to our understanding of protein function. Therefore, new experimental approaches to understand how protein structure is influenced by protein-interface interactions are important. Herein we build on our previous work exploring electrochemistry at the interface between two immiscible electrolyte solutions (ITIES) to investigate changes in protein secondary structure that are modulated by protein-interface interactions. The ITIES provides an experimental framework to drive protein adsorption at an interface, allowing subsequent spectroscopic analysis (e.g., Fourier transform infrared spectroscopy) to monitor changes in protein structure. Here, we reveal that the interaction between insulin and the interface destabilizes native insulin secondary structure, promoting formation of α helix secondary structures. These structural alterations result from protein-interface rather than protein-protein interactions at the interface. Although this is an emerging approach, our results provide a foundation highlighting the value of the ITIES as a tool to study protein structure and interactions at interfaces. Such knowledge may be useful to elucidate protein function within biological systems or to aid sensor development.

Contribution of macromolecules to brain 1 H MR spectra: Experts' consensus recommendations

Proton MR spectra of the brain, especially those measured at short and intermediate echo times, contain signals from mobile macromolecules (MM). A description of the main MM is provided in this consensus paper. These broad peaks of MM underlie the narrower peaks of metabolites and often complicate their quantification but they also may have potential importance as biomarkers in specific diseases. Thus, separation of broad MM signals from low molecular weight metabolites enables accurate determination of metabolite concentrations and is of primary interest in many studies. Other studies attempt to understand the origin of the MM spectrum, to decompose it into individual spectral regions or peaks and to use the components of the MM spectrum as markers of various physiological or pathological conditions in biomedical research or clinical practice. The aim of this consensus paper is to provide an overview and some recommendations on how to handle the MM signals in different types of studies together with a list of open issues in the field, which are all summarized at the end of the paper.

Studying protein folding in health and disease using biophysical approaches

Protein folding is crucial for normal physiology including development and healthy aging, and failure of this process is related to the pathology of diseases including neurodegeneration and cancer. Early thermodynamic and kinetic studies based on the unfolding and refolding equilibrium of individual proteins in the test tube have provided insight into the fundamental principles of protein folding, although the problem of predicting how any given protein will fold remains unsolved. Protein folding within cells is a more complex issue than folding of purified protein in isolation, due to the complex interactions within the cellular environment, including post-translational modifications of proteins, the presence of macromolecular crowding in cells, and variations in the cellular environment, for example in cancer versus normal cells. Development of biophysical approaches including fluorescence resonance energy transfer (FRET) and nuclear magnetic resonance (NMR) techniques and cellular manipulations including microinjection and insertion of noncanonical amino acids has allowed the study of protein folding in living cells. Furthermore, biophysical techniques such as single-molecule fluorescence spectroscopy and optical tweezers allows studies of simplified systems at the single molecular level. Combining in-cell techniques with the powerful detail that can be achieved from single-molecule studies allows the effects of different cellular components including molecular chaperones to be monitored, providing us with comprehensive understanding of the protein folding process. The application of biophysical techniques to the study of protein folding is arming us with knowledge that is fundamental to the battle against cancer and other diseases related to protein conformation or protein–protein interactions.

VoroCNN: Deep convolutional neural network built on 3D Voronoi tessellation of protein structures

MOTIVATION

Effective use of evolutionary information has recently led to tremendous progress in computational prediction of three-dimensional (3D) structures of proteins and their complexes. Despite the progress, the accuracy of predicted structures tends to vary considerably from case to case. Since the utility of computational models depends on their accuracy, reliable estimates of deviation between predicted and native structures are of utmost importance.

RESULTS

For the first time, we present a deep convolutional neural network (CNN) constructed on a Voronoi tessellation of 3D molecular structures. Despite the irregular data domain, our data representation allows us to efficiently introduce both convolution and pooling operations and train the network in an end-to-end fashion without precomputed descriptors. The resultant model, VoroCNN, predicts local qualities of 3D protein folds. The prediction results are competitive to state of the art and superior to the previous 3D CNN architectures built for the same task. We also discuss practical applications of VoroCNN, for example, in recognition of protein binding interfaces.

AVAILABILITY

The model, data, and evaluation tests are available at https://team.inria.fr/nano-d/software/vorocnn/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Membrane interactions of the anuran antimicrobial peptide HSP1-NH2: Different aspects of the association to anionic and zwitterionic biomimetic systems

Studies have suggested that antimicrobial peptides act by different mechanisms, such as micellisation, self-assembly of nanostructures and pore formation on the membrane surface. This work presents an extensive investigation of the membrane interactions of the 14 amino-acid antimicrobial peptide hylaseptin P1-NH₂ (HSP1-NH₂), derived from the tree-frog Hyla punctata, which has stronger antifungal than antibacterial potential. Biophysical and structural analyses were performed and the correlated results were used to describe in detail the interactions of HSP1-NH₂ with zwitterionic and anionic detergent micelles and phospholipid vesicles. HSP1-NH₂ presents similar well-defined helical conformations in both zwitterionic and anionic micelles, although NMR spectroscopy revealed important structural differences in the peptide N-terminus. ²H exchange experiments of HSP1-NH₂ indicated the insertion of the most N-terminal residues (1-3) in the DPC-d₃₈ micelles. A higher enthalpic contribution was verified for the interaction of the peptide with anionic vesicles in comparison with zwitterionic vesicles. The pore formation ability of HSP1-NH₂ (examined by dye release assays) and its effect on the size and surface charge as well as on the lipid acyl chain ordering (evaluated by Fourier-transform infrared spectroscopy) of anionic phospholipid vesicles showed membrane disruption even at low peptide-to-phospholipid ratios, and the effect increases proportionately to the peptide concentration. On the other hand, these biophysical investigations showed that a critical peptide-to-phospholipid ratio around 0.6 is essential for promoting disruption of zwitterionic membranes. In conclusion, this study demonstrates that the binding process of the antimicrobial HSP1-NH₂ peptide depends on the membrane composition and peptide concentration.