Protein representations: Encoding biological information for machine learning in biocatalysis.

Enzymes offer a more environmentally friendly and low-impact solution to conventional chemistry, but they often require additional engineering for their application in industrial settings, an endeavour that is challenging and laborious. To address this issue, the power of machine learning can be harnessed to produce predictive models that enable the in silico study and engineering of improved enzymatic properties. Such machine learning models, however, require the conversion of the complex biological information to a numerical input, also called protein representations. These inputs demand special attention to ensure the training of accurate and precise models, and, in this review, we therefore examine the critical step of encoding protein information to numeric representations for use in machine learning. We selected the most important approaches for encoding the three distinct biological protein representations - primary sequence, 3D structure, and dynamics - to explore their requirements for employment and inductive biases. Combined representations of proteins and substrates are also introduced as emergent tools in biocatalysis. We propose the division of fixed representations, a collection of rule-based encoding strategies, and learned representations extracted from the latent spaces of large neural networks. To select the most suitable protein representation, we propose two main factors to consider. The first one is the model setup, which is influenced by the size of the training dataset and the choice of architecture. The second factor is the model objectives such as consideration about the assayed property, the difference between wild-type models and mutant predictors, and requirements for explainability. This review is aimed at serving as a source of information and guidance for properly representing enzymes in future machine learning models for biocatalysis.

Macromolecular crowding effects on protein dynamics.

Macromolecular crowding experiments bridge the gap between in-vivo and in-vitro studies by mimicking some of the cellular complexities like high viscosity and limited space, while still manageable for experiments and analysis. Macromolecular crowding impacts all biological processes and is a focus of contemporary research. Recent reviews have highlighted the effect of crowding on various protein properties. One of the essential characteristics of protein is its dynamic nature; however, how protein dynamics get modulated in the crowded milieu has been largely ignored. This article discusses how protein translational, rotational, conformational, and solvation dynamics change under crowded conditions, summarizing key observations in the literature. We emphasize our research on microsecond conformational and water dynamics in crowded milieus and their impact on enzymatic activity and stability. Lastly, we provided our outlook on how this field might move forward in the future.

A lytic transglycosylase connects bacterial focal adhesion complexes to the peptidoglycan cell wall.

The Gram-negative bacterium Myxococcus xanthus glides on solid surfaces. Dynamic bacterial focal adhesion complexes (bFACs) convert proton motive force from the inner membrane into mechanical propulsion on the cell surface. It is unclear how the mechanical force transmits across the rigid peptidoglycan (PG) cell wall. Here, we show that AgmT, a highly abundant lytic PG transglycosylase homologous to Escherichia coli MltG, couples bFACs to PG. Coprecipitation assay and single-particle microscopy reveal that the gliding motors fail to connect to PG and thus are unable to assemble into bFACs in the absence of an active AgmT. Heterologous expression of E. coli MltG restores the connection between PG and bFACs and thus rescues gliding motility in the M. xanthus cells that lack AgmT. Our results indicate that bFACs anchor to AgmT-modified PG to transmit mechanical force across the PG cell wall.

Utilizing Molecular Dynamics Simulations, Machine Learning, Cryo-EM, and NMR Spectroscopy to Predict and Validate Protein Dynamics.

Protein dynamics play a crucial role in biological function, encompassing motions ranging from atomic vibrations to large-scale conformational changes. Recent advancements in experimental techniques, computational methods, and artificial intelligence have revolutionized our understanding of protein dynamics. Nuclear magnetic resonance spectroscopy provides atomic-resolution insights, while molecular dynamics simulations offer detailed trajectories of protein motions. Computational methods applied to X-ray crystallography and cryo-electron microscopy (cryo-EM) have enabled the exploration of protein dynamics, capturing conformational ensembles that were previously unattainable. The integration of machine learning, exemplified by AlphaFold2, has accelerated structure prediction and dynamics analysis. These approaches have revealed the importance of protein dynamics in allosteric regulation, enzyme catalysis, and intrinsically disordered proteins. The shift towards ensemble representations of protein structures and the application of single-molecule techniques have further enhanced our ability to capture the dynamic nature of proteins. Understanding protein dynamics is essential for elucidating biological mechanisms, designing drugs, and developing novel biocatalysts, marking a significant paradigm shift in structural biology and drug discovery.

Multiresolution molecular dynamics simulations reveal the interplay between conformational variability and functional interactions in membrane-bound cytochrome P450 2B4.

Cytochrome P450 2B4 (CYP 2B4) is one of the best-characterized CYPs and serves as a key model system for understanding the mechanisms of microsomal class II CYPs, which metabolize most known drugs. The highly flexible nature of CYP 2B4 is apparent from crystal structures that show the active site with either a wide open or a closed heme binding cavity. Here, we investigated the conformational ensemble of the full-length CYP 2B4 in a phospholipid bilayer, using multiresolution molecular dynamics (MD) simulations. Coarse-grained MD simulations revealed two predominant orientations of CYP 2B4's globular domain with respect to the bilayer. Their refinement by atomistic resolution MD showed adaptation of the enzyme's interaction with the lipid bilayer, leading to open configurations that facilitate ligand access to the heme binding cavity. CAVER analysis of enzyme tunnels, AquaDuct analysis of water routes, and Random Acceleration Molecular Dynamics simulations of ligand dissociation support the conformation-dependent passage of molecules between the active site and the protein surroundings. Furthermore, simulation of the re-entry of the inhibitor bifonazole into the open conformation of CYP 2B4 resulted in binding at a transient hydrophobic pocket within the active site cavity that may play a role in substrate binding or allosteric regulation. Together, these results show how the open conformation of CYP 2B4 facilitates the binding of substrates from and release of products to the membrane, whereas the closed conformation prolongs the residence time of substrates or inhibitors and selectively allows the passage of smaller reactants via the solvent and water channels.

Nanoscale dynamics of the cadherin-catenin complex bound to vinculin revealed by neutron spin echo spectroscopy.

We report a neutron spin echo (NSE) study of the nanoscale dynamics of the cell-cell adhesion cadherin-catenin complex bound to vinculin. Our measurements and theoretical physics analyses of the NSE data reveal that the dynamics of full-length α-catenin, ß-catenin, and vinculin residing in the cadherin-catenin-vinculin complex become activated, involving nanoscale motions in this complex. The cadherin-catenin complex is the central component of the cell-cell adherens junction (AJ) and is fundamental to embryogenesis, tissue wound healing, neuronal plasticity, cancer metastasis, and cardiovascular health and disease. A highly dynamic cadherin-catenin-vinculin complex provides the molecular dynamics basis for the flexibility and elasticity that are necessary for the AJs to function as force transducers. Our theoretical physics analysis provides a way to elucidate these driving nanoscale motions within the complex without requiring large-scale numerical simulations, providing insights not accessible by other techniques. We propose a three-way "motorman" entropic spring model for the dynamic cadherin-catenin-vinculin complex, which allows the complex to function as a flexible and elastic force transducer.

Dissecting SOX9 dynamics reveals its differential regulation in osteoarthritis.

The transcription factor SOX9 is integral to tissue homeostasis and is implicated in skeletal malformation, campomelic dysplasia, and osteoarthritis (OA). Despite extensive research, the complete regulatory landscape of SOX9 transcriptional activity, interconnected with signaling pathways (TGFß, WNT, BMP, IHH, NFκB, and HIF), remains challenging to decipher. This study focuses on elucidating SOX9 signaling in OA pathology using Fluorescence Recovery After Photobleaching (FRAP) to assess SOX9 activity directly in live human primary chondrocytes (hPCs). Single cell FRAP data revealed two distinct subpopulations with differential SOX9 dynamics, showing varied distribution between healthy and OA hPCs. Moreover, inherently elevated SOX9-DNA binding was observed in healthy hPCs compared to preserved and OA counterparts. Anabolic factors (BMP7 and GREM1) and catabolic inhibitors (DKK1 and FRZb) were found to modulate SOX9 transcriptional activity in OA-hPCs. These findings provide valuable insights into the intricate regulation of SOX9 signaling in OA, suggesting potential therapeutic avenues for modulating SOX9 activity in diseased states.

Anionic cardiolipin stabilizes the transmembrane region of hyaluronan synthase and promotes catalysis-relevant dynamics.

Hyaluronic acid (HA) is a glycosaminoglycan essential for cellular processes and finding increasingly applications in medicine, pharmaceuticals, and cosmetics. While membrane-integrated Class I hyaluronan synthase (HAS) catalyzes HA synthesis in most organisms, the molecular mechanisms by which HAS-lipid interactions impact HAS catalysis remain unclear. This study employed coarse-grained molecular dynamics simulation combined with dimensionality reduction to uncover the interplay between lipids and Streptococcus equisimilis HAS (SeHAS). A minimum of 67 % cardiolipin is necessary for HA synthesis, as determined through simulations using gradient-composed membranes. The anionic cardiolipin stabilizes the cationic transmembrane regions of SeHAS and thereby maintains its conformation. Moreover, the highly dynamic cardiolipin is required to modulate the catalysis-relevant motions in HAS and thus facilitate HA synthesis. These findings provide molecular insights essential not only for understanding the physiological functions of HAS, but also for the development of cell factories and enzyme catalysts for HA production.

Predicting Mutation-Induced Allosteric Changes in Structures and Conformational Ensembles of the ABL Kinase Using AlphaFold2 Adaptations with Alanine Sequence Scanning.

Despite the success of AlphaFold2 approaches in predicting single protein structures, these methods showed intrinsic limitations in predicting multiple functional conformations of allosteric proteins and have been challenged to accurately capture the effects of single point mutations that induced significant structural changes. We examined several implementations of AlphaFold2 methods to predict conformational ensembles for state-switching mutants of the ABL kinase. The results revealed that a combination of randomized alanine sequence masking with shallow multiple sequence alignment subsampling can significantly expand the conformational diversity of the predicted structural ensembles and capture shifts in populations of the active and inactive ABL states. Consistent with the NMR experiments, the predicted conformational ensembles for M309L/L320I and M309L/H415P ABL mutants that perturb the regulatory spine networks featured the increased population of the fully closed inactive state. The proposed adaptation of AlphaFold can reproduce the experimentally observed mutation-induced redistributions in the relative populations of the active and inactive ABL states and capture the effects of regulatory mutations on allosteric structural rearrangements of the kinase domain. The ensemble-based network analysis complemented AlphaFold predictions by revealing allosteric hotspots that correspond to state-switching mutational sites which may explain the global effect of regulatory mutations on structural changes between the ABL states. This study suggested that attention-based learning of long-range dependencies between sequence positions in homologous folds and deciphering patterns of allosteric interactions may further augment the predictive abilities of AlphaFold methods for modeling of alternative protein sates, conformational ensembles and mutation-induced structural transformations.

Innovative Strategies in X-ray Crystallography for Exploring Structural Dynamics and Reaction Mechanisms in Metabolic Disorders.

Enzymes are crucial in metabolic processes, and their dysfunction can lead to severe metabolic disorders. Structural biology, particularly X-ray crystallography, has advanced our understanding of these diseases by providing 3D structures of pathological enzymes. However, traditional X-ray crystallography faces limitations, such as difficulties in obtaining suitable protein crystals and studying protein dynamics. X-ray free-electron lasers (XFELs) have revolutionized this field with their bright and brief X-ray pulses, providing high-resolution structures of radiation-sensitive and hard-to-crystallize proteins. XFELs also enable the study of protein dynamics through room temperature structures and time-resolved serial femtosecond crystallography, offering comprehensive insights into the molecular mechanisms of metabolic diseases. Understanding these dynamics is vital for developing effective therapies. This review highlights the contributions of protein dynamics studies using XFELs and synchrotrons to metabolic disorder research and their application in designing better therapies. It also discusses G protein-coupled receptors (GPCRs), which, though not enzymes, play key roles in regulating physiological systems and are implicated in many metabolic disorders.

Exploring protein structural ensembles: Integration of sparse experimental data from electron paramagnetic resonance spectroscopy with molecular modeling methods.

Under physiological conditions, proteins continuously undergo structural fluctuations on different timescales. Some conformations are only sparsely populated, but still play a key role in protein function. Thus, meaningful structure-function frameworks must include structural ensembles rather than only the most populated protein conformations. To detail protein plasticity, modern structural biology combines complementary experimental and computational approaches. In this review, we survey available computational approaches that integrate sparse experimental data from electron paramagnetic resonance spectroscopy with molecular modeling techniques to derive all-atom structural models of rare protein conformations. We also propose strategies to increase the reliability and improve efficiency using deep learning approaches, thus advancing the field of integrative structural biology.

The structural biology and dynamics of malate dehydrogenases.

Malate dehydrogenase (MDH) enzymes catalyze the reversible oxidoreduction of malate to oxaloacetate using NAD(P) as a cofactor. This reaction is vital for metabolism and the exchange of reducing equivalents between cellular compartments. There are more than 100 structures of MDH in the Protein Data Bank, representing species from archaea, bacteria, and eukaryotes. This conserved family of enzymes shares a common nucleotide-binding domain, substrate-binding domain, and subunits associate to form a dimeric or a tetrameric enzyme. Despite the variety of crystallization conditions and ligands in the experimental structures, the conformation and configuration of MDH are similar. The quaternary structure and active site dynamics account for most conformational differences in the experimental MDH structures. Oligomerization appears essential for activity despite each subunit having a structurally independent active site. There are two dynamic regions within the active site that influence substrate binding and possibly catalysis, with one of these regions adjoining the subunit interface. In this review, we introduce the reader to the general structural framework of MDH highlighting the conservation of certain features and pointing out unique differences that regulate MDH enzyme activity.

Differential conformational dynamics in two type-A RNA-binding domains drive the double-stranded RNA recognition and binding.

Trans-activation response (TAR) RNA-binding protein (TRBP) has emerged as a key player in the RNA interference pathway, wherein it binds to different pre-microRNAs (miRNAs) and small interfering RNAs (siRNAs), each varying in sequence and/or structure. We hypothesize that TRBP displays dynamic adaptability to accommodate heterogeneity in target RNA structures. Thus, it is crucial to ascertain the role of intrinsic and RNA-induced protein dynamics in RNA recognition and binding. We have previously elucidated the role of intrinsic and RNA-induced conformational exchange in the double-stranded RNA-binding domain 1 (dsRBD1) of TRBP in shape-dependent RNA recognition. The current study delves into the intrinsic and RNA-induced conformational dynamics of the TRBP-dsRBD2 and then compares it with the dsRBD1 study carried out previously. Remarkably, the two domains exhibit differential binding affinity to a 12-bp dsRNA owing to the presence of critical residues and structural plasticity. Furthermore, we report that dsRBD2 depicts constrained conformational plasticity when compared to dsRBD1. Although, in the presence of RNA, dsRBD2 undergoes induced conformational exchange within the designated RNA-binding regions and other residues, the amplitude of the motions remains modest when compared to those observed in dsRBD1. We propose a dynamics-driven model of the two tandem domains of TRBP, substantiating their contributions to the versatility of dsRNA recognition and binding.

Water determines the intramolecular dynamics of proteins. En example of bovine serum albumin.

In this work, the terahertz time-domain spectroscopy method analyzed solutions of bovine serum albumin (BSA) in two high concentrations (50 and 334 mg/mL) at three pH values (2.5, 6.5, 8.5) and the same solvents without protein, at 25°C. The spectra of dry BSA were also recorded. For the first time, a method for determining the complex dielectric permittivity of protein molecules in aqueous solutions, without the dielectric contribution of the aqueous phase, is proposed. It is shown that the dielectric permittivity of dissolved and dry BSA (lyophilized, in the native conformation) differ significantly in the terahertz frequency range. These differences are small near 70 cm-1, but they increase greatly with decreasing frequency. It was found that the dielectric losses of protein molecules in solution are close to the dielectric losses of the aqueous environment, which in this frequency range are determined by intermolecular relaxation processes of water. Since dielectric losses are directly related to molecular dynamics, this fact shows that the intramolecular dynamics of the protein completely adjusts to the intermolecular dynamics of the aqueous environment. It also indicates that the native conformation does not determine all the fundamental characteristics of a protein molecule, in particular, it does not determine the dynamics of the protein, which significantly depends on the water environment.

Riemannian geometry for efficient analysis of protein dynamics data.

An increasingly common viewpoint is that protein dynamics datasets reside in a nonlinear subspace of low conformational energy. Ideal data analysis tools should therefore account for such nonlinear geometry. The Riemannian geometry setting can be suitable for a variety of reasons. First, it comes with a rich mathematical structure to account for a wide range of geometries that can be modeled after an energy landscape. Second, many standard data analysis tools developed for data in Euclidean space can be generalized to Riemannian manifolds. In the context of protein dynamics, a conceptual challenge comes from the lack of guidelines for constructing a smooth Riemannian structure based on an energy landscape. In addition, computational feasibility in computing geodesics and related mappings poses a major challenge. This work considers these challenges. The first part of the paper develops a local approximation technique for computing geodesics and related mappings on Riemannian manifolds in a computationally feasible manner. The second part constructs a smooth manifold and a Riemannian structure that is based on an energy landscape for protein conformations. The resulting Riemannian geometry is tested on several data analysis tasks relevant for protein dynamics data. In particular, the geodesics with given start- and end-points approximately recover corresponding molecular dynamics trajectories for proteins that undergo relatively ordered transitions with medium-sized deformations. The Riemannian protein geometry also gives physically realistic summary statistics and retrieves the underlying dimension even for large-sized deformations within seconds on a laptop.

Quantifying Protein Dynamics by Kymograph Analysis.

Time-lapse imaging of the subcellular localization and dynamic behavior of proteins is critical to understand their biological functions in cells. With the advent of various methodologies and computational tools, the precise tracking and quantification of protein spatiotemporal dynamics have become feasible. Kymograph analysis, in particular, has been extensively adopted for the quantitative assessment of proteins, vesicles, and organelle movements. However, conventional kymograph analysis, which is based on a single linear trajectory, may not comprehensively capture the complexity of proteins that alter their course during intracellular transport and activity. In this chapter, we introduced an advanced protocol for whole-cell kymograph analysis that allows for three-dimensional (3D) tracking of protein dynamics. This method was validated through the analysis of tip-focused endocytosis and exocytosis processes in growing tobacco pollen tubes by employing both the advanced whole-cell and classical kymograph methods. In addition, we enhanced this method by integrating pseudo-colored kymographs that enables the direct visualization of changes in protein fluorescence intensity with fluorescence recovery after photobleaching to advance our understanding of protein localization and dynamics. This comprehensive method offers a novel insight into the intricate dynamics of protein activity within the cellular context.

Allostery at a Protein-Protein Interface Harboring an Intermolecular Motional Network.

Motional properties of proteins govern recognition, catalysis, and regulation. The dynamics of tightly interacting residues can form intramolecular dynamic networks, dependencies fine-tuned by evolution to optimize a plethora of functional aspects. The constructive interaction of residues from different proteins to assemble intermolecular dynamic networks is a similarly likely case but has escaped thorough experimental assessment due to interfering association/dissociation dynamics. Here, we use fast-MAS solid-state 15N R1ρ NMR relaxation dispersion aided by molecular-dynamics simulations to mechanistically assess the hierarchy of individual µs timescale motions arising from a crystal-crystal contact, in the absence of translational motion. In contrast to the monomer, where particular mutations entail isolated perturbations, specific intermolecular interactions couple the motional properties between distant residues in the same protein. The mechanistic insights obtained from this conceptual work may improve our understanding on how intramolecular allostery can be tuned by intermolecular interactions via assembly of dynamic networks from previously isolated elements.

Pitfalls in measurements of R1 relaxation rates of protein backbone 15N nuclei.

The dynamics of the backbone and side-chains of protein are routinely studied by interpreting experimentally determined 15N spin relaxation rates. R1(15N), the longitudinal relaxation rate, reports on fast motions and encodes, together with the transverse relaxation R2, structural information about the shape of the molecule and the orientation of the amide bond vectors in the internal diffusion frame. Determining error-free 15N longitudinal relaxation rates remains a challenge for small, disordered, and medium-sized proteins. Here, we show that mono-exponential fitting is sufficient, with no statistical preference for bi-exponential fitting up to 800 MHz. A detailed comparison of the TROSY and HSQC techniques at medium and high fields showed no statistically significant differences. The least error-prone DD/CSA interference removal technique is the selective inversion of amide signals while avoiding water resonance. The exchange of amide with solvent deuterons appears to affect the rate R1 of solvent-exposed amides in all fields tested and in each DD/CSA interference removal technique in a statistically significant manner. In summary, the most accurate R1(15N) rates in proteins are achieved by selective amide inversion, without the addition of D2O. Importantly, at high magnetic fields stronger than 800 MHz, when non-mono-exponential decay is involved, it is advisable to consider elimination of the shortest delays (typically up to 0.32 s) or bi-exponential fitting.

Molecular basis of the interactions between the disordered Neh4 and Neh5 domains of Nrf2 and CBP/p300 in oxidative stress response.

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a major transcription factor that functions in maintaining redox homeostasis in cells. It mediates the transcription of cytoprotective genes in response to environmental and endogenous stresses to prevent oxidative damage. Thus, Nrf2 plays a significant role in chemoprevention. However, aberrant activation of Nrf2 has been shown to protect cancer cells from apoptosis and contribute to their chemoresistance. The interaction between Nrf2 and CBP is critical for the gene transcription activation. CBP and its homologue p300 interact with two transactivation domains in Nrf2, Neh4, and Neh5 domains through their TAZ1 and TAZ2 domains. To date, the molecular basis of this crucial interaction is not known, hindering a more detailed understanding of the regulation of Nrf2. To close this knowledge gap, we have used a set of biophysical experiments to dissect the Nrf2-CBP/p300 interactions. Structural properties of Neh4 and Neh5 and their binding with the TAZ1 and TAZ2 domains of CBP/p300 were characterized. Our results show that the Neh4 and Neh5 domains of Nrf2 are intrinsically disordered, and they both can bind the TAZ1 and TAZ2 domains of CBP/p300 with micromolar affinities. The findings provide molecular insight into the regulation of Nrf2 by CBP/p300 through multi-domain interactions.

NMR Dynamic View of the Stabilization of the WW4 Domain by Neutral NaCl and Kosmotropic Na2SO4 and NaH2PO4.

The Hofmeister series categorizes ions based on their effects on protein stability, yet the microscopic mechanism remains a mystery. In this series, NaCl is neutral, Na2SO4 and Na2HPO4 are kosmotropic, while GdmCl and NaSCN are chaotropic. This study employs CD and NMR to investigate the effects of NaCl, Na2SO4, and Na2HPO4 on the conformation, stability, binding, and backbone dynamics (ps-ns and µs-ms time scales) of the WW4 domain with a high stability and accessible side chains at concentrations ≤ 200 mM. The results indicated that none of the three salts altered the conformation of WW4 or showed significant binding to the four aliphatic hydrophobic side chains. NaCl had no effect on its thermal stability, while Na2SO4 and Na2HPO4 enhanced the stability by ~5 °C. Interestingly, NaCl only weakly interacted with the Arg27 amide proton, whereas Na2SO4 bound to Arg27 and Phe31 amide protons with Kd of 32.7 and 41.6 mM, respectively. Na2HPO4, however, bound in a non-saturable manner to Trp9, His24, and Asn36 amide protons. While the three salts had negligible effects on ps-ns backbone dynamics, NaCl and Na2SO4 displayed no effect while Na2HPO4 significantly increased the µs-ms backbone dynamics. These findings, combined with our recent results with GdmCl and NaSCN, suggest a microscopic mechanism for the Hofmeister series. Additionally, the data revealed a lack of simple correlation between thermodynamic stability and backbone dynamics, most likely due to enthalpy-entropy compensation. Our study rationalizes the selection of chloride and phosphate as the primary anions in extracellular and intracellular spaces, as well as polyphosphate as a primitive chaperone in certain single-cell organisms.