Sequence-Based Prediction of Cysteine Reactivity Using Machine Learning

Predicting cysteine reactivity changes upon phosphorylation using XGBoost

Cysteine reactivity serves as a significant indicator of protein function and can be affected by phosphorylation events. Experimental approaches have been developed to investigate this effect, but the scale is still relatively limited. Machine-learning approaches promise to accelerate the investigation of these phenomena. In this study, protein sequence information, distances to the closest phosphorylation sites, and the membership score of the intrinsically disordered region were used to represent the cysteine. Following the feature selection using an elastic net model, two groups of binary classifiers based on XGBoost were built to predict the occurrence and the direction of the reactivity change as a response to phosphorylation events, respectively. In addition, function enrichment analysis was performed on proteins/genes predicted to have reactivity changes. XGBoost performed the best in the independent test with AUC of 0.8192 and 0.9203 for the prediction of the change's occurrence and direction, respectively. The use of two binary classifiers successively resulted in an accuracy of 0.7568 in predicting whether reactivity would be unchanged, increased, or decreased. The enrichment analysis revealed the association of proteins carrying reactivity-changed cysteine residues with various disease-related pathways, particularly cancer, autosomal dominant diseases, and viral infections. Changes in cysteine reactivity influenced by phosphorylation are site-specific and can be predicted by XGBoost algorithms. Our model provides an efficient alternative way to explore the cysteine reactivity upon phosphorylation at the proteome-wide level, facilitating the investigation of protein functions and their clinical insights. Our code is available on GitHub (https://github.com/DarinaOsamu/predictors-of-cysteine-reactivity-changes).

Proteome-wide structural analysis identifies warhead- and coverage-specific biases in cysteine-focused chemoproteomics

Covalent drug discovery has undergone a resurgence over the past two decades and reactive cysteine profiling has emerged in parallel as a platform for ligand discovery through on- and off-target profiling; however, the scope of this approach has not been fully explored at the whole-proteome level. We combined AlphaFold2-predicted side-chain accessibilities for >95% of the human proteome with a meta-analysis of eighteen public cysteine profiling datasets, totaling 44,187 unique cysteine residues, revealing accessibility biases in sampled cysteines primarily dictated by warhead chemistry. Analysis of >3.5 million cysteine-fragment interactions further showed that hit elaboration and optimization drives increased bias against buried cysteine residues. Based on these data, we suggest that current profiling approaches cover a small proportion of potential ligandable cysteine residues and propose future directions for increasing coverage, focusing on high-priority residues and depth. All analysis and produced resources are freely available and extendable to other reactive amino acids.

Systematic conformation-to-phenotype mapping via limited deep sequencing of proteins

Non-native conformations drive protein-misfolding diseases, complicate bioengineering efforts, and fuel molecular evolution. No current experimental technique is well suited for elucidating them and their phenotypic effects. Especially intractable are the transient conformations populated by intrinsically disordered proteins. We describe an approach to systematically discover, stabilize, and purify native and non-native conformations, generated in vitro or in vivo, and directly link conformations to molecular, organismal, or evolutionary phenotypes. This approach involves high-throughput disulfide scanning (HTDS) of the entire protein. To reveal which disulfides trap which chromatographically resolvable conformers, we devised a deep-sequencing method for double-Cys variant libraries of proteins that precisely and simultaneously locates both Cys residues within each polypeptide. HTDS of the abundant E. coli periplasmic chaperone HdeA revealed distinct classes of disordered hydrophobic conformers with variable cytotoxicity depending on where the backbone was cross-linked. HTDS can bridge conformational and phenotypic landscapes for many proteins that function in disulfide-permissive environments.

How eluents define proteomic fingerprinting of protein corona on nanoparticles

Nanoparticles (NPs) have broad application prospects in the field of biomedicine due to their excellent physicochemical properties. When entering biological fluids, NPs inevitably encountered proteins and were subsequently surrounded by them, forming the termed protein corona (PC). As PC has been evidenced to have critical roles in deciding the biological fates of NPs, how to precisely characterize PC is vital to promote the clinical translation of nanomedicine by understanding and harnessing NPs' behaviors. During the centrifugation-based separation techniques for the PC preparation, direct elution has been most widely used to strip proteins from NPs due to its simpleness and robustness, but the roles of multifarious eluents have never been systematically declared. Herein, seven eluents composed of three denaturants, sodium dodecyl sulfate (SDS), dithiothreitol (DTT), and urea (Urea), were applied to detach PC from gold nanoparticles (AuNPs) and silica nanoparticles (SiNPs), and eluted proteins in PC have been carefully characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and chromatography coupled tandem mass spectrometry (LC-MS/MS). Our results showed that SDS and DTT were the main contributors to the efficient desorption of PC on SiNPs and AuNPs, respectively. The molecular reactions between NPs and proteins were explored and verified by SDS-PAGE analysis of PC formed in the serums pretreated with protein denaturing or alkylating agents. The proteomic fingerprinting analysis indicated the difference of the eluted proteins brought by the seven eluents was the abundance rather than the species. The enrichment of some opsonins and dysopsonins in a special elution reminds us that the possibility of biased judgments on predicting NPs' biological behaviors under different elution conditions. The synergistic effects or antagonistic effects among denaturants for eluting PC were manifested in a nanoparticle-type dependent way by integrating the properties of the eluted proteins. Collectively, this study not only underlines the urgent need of choosing the appropriate eluents for identifying PC robustly and unbiasedly, but also provides an insight into the understanding of molecular interactions during PC formation.

HyperCys: A Structure- and Sequence-Based Predictor of Hyper-Reactive Druggable Cysteines

The cysteine side chain has a free thiol group, making it the amino acid residue most often covalently modified by small molecules possessing weakly electrophilic warheads, thereby prolonging on-target residence time and reducing the risk of idiosyncratic drug toxicity. However, not all cysteines are equally reactive or accessible. Hence, to identify targetable cysteines, we propose a novel ensemble stacked machine learning (ML) model to predict hyper-reactive druggable cysteines, named HyperCys. First, the pocket, conservation, structural and energy profiles, and physicochemical properties of (non)covalently bound cysteines were collected from both protein sequences and 3D structures of protein-ligand complexes. Then, we established the HyperCys ensemble stacked model by integrating six different ML models, including K-nearest neighbors, support vector machine, light gradient boost machine, multi-layer perceptron classifier, random forest, and the meta-classifier model logistic regression. Finally, based on the hyper-reactive cysteines' classification accuracy and other metrics, the results for different feature group combinations were compared. The results show that the accuracy, F1 score, recall score, and ROC AUC values of HyperCys are 0.784, 0.754, 0.742, and 0.824, respectively, after performing 10-fold CV with the best window size. Compared to traditional ML models with only sequenced-based features or only 3D structural features, HyperCys is more accurate at predicting hyper-reactive druggable cysteines. It is anticipated that HyperCys will be an effective tool for discovering new potential reactive cysteines in a wide range of nucleophilic proteins and will provide an important contribution to the design of targeted covalent inhibitors with high potency and selectivity.

The Role of the Nucleotides in the Insertion of the bis-Molybdopterin Guanine Dinucleotide Cofactor into apo-Molybdoenzymes

The role of the GMP nucleotides of the bis-molybdopterin guanine dinucleotide (bis-MGD) cofactor of the DMSO reductase family has long been a subject of discussion. The recent characterization of the bis-molybdopterin (bis-Mo-MPT) cofactor present in the E. coli YdhV protein, which differs from bis-MGD solely by the absence of the nucleotides, now enables studying the role of the nucleotides of bis-MGD and bis-MPT cofactors in Moco insertion and the activity of molybdoenzymes in direct comparison. Using the well-known E. coli TMAO reductase TorA as a model enzyme for cofactor insertion, we were able to show that the GMP nucleotides of bis-MGD are crucial for the insertion of the bis-MGD cofactor into apo-TorA.

Identification of Iron-Sulfur (Fe-S) Cluster and Zinc (Zn) Binding Sites Within Proteomes Predicted by DeepMind's AlphaFold2 Program Dramatically Expands the Metalloproteome

DeepMind's AlphaFold2 software has ushered in a revolution in high quality, 3D protein structure prediction. In very recent work by the DeepMind team, structure predictions have been made for entire proteomes of twenty-one organisms, with >360,000 structures made available for download. Here we show that thousands of novel binding sites for iron-sulfur (Fe-S) clusters and zinc (Zn) ions can be identified within these predicted structures by exhaustive enumeration of all potential ligand-binding orientations. We demonstrate that AlphaFold2 routinely makes highly specific predictions of ligand binding sites: for example, binding sites that are comprised exclusively of four cysteine sidechains fall into three clusters, representing binding sites for 4Fe-4S clusters, 2Fe-2S clusters, or individual Zn ions. We show further: (a) that the majority of known Fe-S cluster and Zn binding sites documented in UniProt are recovered by the AlphaFold2 structures, (b) that there are occasional disputes between AlphaFold2 and UniProt with AlphaFold2 predicting highly plausible alternative binding sites, (c) that the Fe-S cluster binding sites that we identify in E. coli agree well with previous bioinformatics predictions, (d) that cysteines predicted here to be part of ligand binding sites show little overlap with those shown via chemoproteomics techniques to be highly reactive, and (e) that AlphaFold2 occasionally appears to build erroneous disulfide bonds between cysteines that should instead coordinate a ligand. These results suggest that AlphaFold2 could be an important tool for the functional annotation of proteomes, and the methodology presented here is likely to be useful for predicting other ligand-binding sites.

Human milk oligosaccharides and non-digestible carbohydrates reduce pathogen adhesion to intestinal epithelial cells by decoy effects or by attenuating bacterial virulence

This work investigated the effects of different chemical structures of human milk oligosaccharides (hMOs) and non-digestible carbohydrates (NDCs) on pathogen adhesion by serving as decoy receptors. Pre-exposure of pathogens to inulins and low degree of methylation (DM) pectin prevented binding to gut epithelial Caco2-cells, but effects were dependent on the molecules' chemistry, pathogen strain and growth phase. Pre-exposure to 3-fucosyllactose increased E. coli WA321 adhesion (28%, p < 0.05), and DM69 pectin increased E. coli ET8 (15 fold, p < 0.05) and E. coli WA321 (50%, p < 0.05) adhesion. Transcriptomics analysis revealed that DM69 pectin upregulated flagella and cell membrane associated genes. However, the top 10 downregulated genes were associated with lowering of bacteria virulence. DM69 pectin increased pathogen adhesion but bacterial virulence was attenuated illustrating different mechanisms may lower pathogen adhesion. Our study illustrates that both hMOs and NDCs can reduce adhesion or attenuate virulence of pathogens but that these effects are chemistry dependent.

Protein Modifications: From Chemoselective Probes to Novel Biocatalysts

Chemical reactions can be performed to covalently modify specific residues in proteins. When applied to native enzymes, these chemical modifications can greatly expand the available set of building blocks for the development of biocatalysts. Nucleophilic canonical amino acid sidechains are the most readily accessible targets for such endeavors. A rich history of attempts to design enhanced or novel enzymes, from various protein scaffolds, has paved the way for a rapidly developing field with growing scientific, industrial, and biomedical applications. A major challenge is to devise reactions that are compatible with native proteins and can selectively modify specific residues. Cysteine, lysine, N-terminus, and carboxylate residues comprise the most widespread naturally occurring targets for enzyme modifications. In this review, chemical methods for selective modification of enzymes will be discussed, alongside with examples of reported applications. We aim to highlight the potential of such strategies to enhance enzyme function and create novel semisynthetic biocatalysts, as well as provide a perspective in a fast-evolving topic.

Ebselen, a multi-target compound: its effects on biological processes and diseases

Ebselen is a well-known synthetic compound mimicking glutathione peroxidase (GPx), which catalyses some vital reactions that protect against oxidative damage. Based on a large number of in vivo and in vitro studies, various mechanisms have been proposed to explain its actions on multiple targets. It targets thiol-related compounds, including cysteine, glutathione, and thiol proteins (e.g., thioredoxin and thioredoxin reductase). Owing to this, ebselen is a unique multifunctional agent with important effects on inflammation, apoptosis, oxidative stress, cell differentiation, immune regulation and neurodegenerative disease, with anti-microbial, detoxifying and anti-tumour activity. This review summarises the current understanding of the multiple biological processes and molecules targeted by ebselen, and its pharmacological applications.

Stoichiometric Thiol Redox Proteomics for Quantifying Cellular Responses to Perturbations

Post-translational modifications regulate the structure and function of proteins that can result in changes to the activity of different pathways. These include modifications altering the redox state of thiol groups on protein cysteine residues, which are sensitive to oxidative environments. While mass spectrometry has advanced the identification of protein thiol modifications and expanded our knowledge of redox-sensitive pathways, the quantitative aspect of this technique is critical for the field of redox proteomics. In this review, we describe how mass spectrometry-based redox proteomics has enabled researchers to accurately quantify the stoichiometry of reversible oxidative modifications on specific cysteine residues of proteins. We will describe advancements in the methodology that allow for the absolute quantitation of thiol modifications, as well as recent reports that have implemented this approach. We will also highlight the significance and application of such measurements and why they are informative for the field of redox biology.

Selective and Effective: Current Progress in Computational Structure-Based Drug Discovery of Targeted Covalent Inhibitors

Targeted covalent inhibitors are currently showing great promise for systems that are normally difficult to target with small molecule therapies. This renewed interest has spurred the refinement of existing computational methods as well as the designof new ones, expanding the toolbox for discovery and optimization of selectiveand effective covalent inhibitors. Commonly applied approaches are covalentdocking methods that predict the conformation of the covalent complex with known residues. More recently, a new predictive method, reactive docking, was developed, building on the growing corpus of data generated by large proteomics experiments. This method was successfully used in several 'inverse drug discovery' programs that use high-throughput techniques to isolate effective compounds based on screening of entire compound libraries based on desired phenotypes.

Predicting Reactive Cysteines with Implicit-Solvent-Based Continuous Constant pH Molecular Dynamics in Amber

Cysteines existing in the deprotonated thiolate form or having a tendency to become deprotonated are important players in enzymatic and cellular redox functions and frequently exploited in covalent drug design; however, most computational studies assume cysteines as protonated. Thus, developing an efficient tool that can make accurate and reliable predictions of cysteine protonation states is timely needed. We recently implemented a generalized Born (GB) based continuous constant pH molecular dynamics (CpHMD) method in Amber for protein pKa calculations on CPUs and GPUs. Here we benchmark the performance of GB-CpHMD for predictions of cysteine pKa's and reactivities using a data set of 24 proteins with both down- and upshifted cysteine pKa's. We found that 10 ns single-pH or 4 ns replica-exchange CpHMD titrations gave root-mean-square errors of 1.2-1.3 and correlation coefficients of 0.8-0.9 with respect to experiment. The accuracy of predicting thiolates or reactive cysteines at physiological pH with single-pH titrations is 86 or 81% with a precision of 100 or 90%, respectively. This performance well surpasses the traditional structure-based methods, particularly a widely used empirical pKa tool that gives an accuracy less than 50%. We discuss simulation convergence, dependence on starting structures, common determinants of the pKa downshifts and upshifts, and the origin of the discrepancies from the structure-based calculations. Our work suggests that CpHMD titrations can be performed on a desktop computer equipped with a single GPU card to predict cysteine protonation states for a variety of applications, from understanding biological functions to covalent drug design.

The biosynthesis of the molybdenum cofactors in Escherichia coli

The biosynthesis of the molybdenum cofactor (Moco) is highly conserved among all kingdoms of life. In all molybdoenzymes containing Moco, the molybdenum atom is coordinated to a dithiolene group present in the pterin-based 6-alkyl side chain of molybdopterin (MPT). In general, the biosynthesis of Moco can be divided into four steps in in bacteria: (i) the starting point is the formation of the cyclic pyranopterin monophosphate (cPMP) from 5'-GTP, (ii) in the second step the two sulfur atoms are inserted into cPMP leading to the formation of MPT, (iii) in the third step the molybdenum atom is inserted into MPT to form Moco and (iv) in the fourth step bis-Mo-MPT is formed and an additional modification of Moco is possible with the attachment of a nucleotide (CMP or GMP) to the phosphate group of MPT, forming the dinucleotide variants of Moco. This review presents an update on the well-characterized Moco biosynthesis in the model organism Escherichia coli including novel discoveries from the recent years.

Redox Systems Biology: Harnessing the Sentinels of the Cysteine Redoxome

Significance: Cellular redox processes are highly interconnected, yet not in equilibrium, and governed by a wide range of biochemical parameters. Technological advances continue refining how specific redox processes are regulated, but broad understanding of the dynamic interconnectivity between cellular redox modules remains limited. Systems biology investigates multiple components in complex environments and can provide integrative insights into the multifaceted cellular redox state. This review describes the state of the art in redox systems biology as well as provides an updated perspective and practical guide for harnessing thousands of cysteine sensors in the redoxome for multiparameter characterization of cellular redox networks. Recent Advances: Redox systems biology has been applied to genome-scale models and large public datasets, challenged common conceptions, and provided new insights that complement reductionist approaches. Advances in public knowledge and user-friendly tools for proteome-wide annotation of cysteine sentinels can now leverage cysteine redox proteomics datasets to provide spatial, functional, and protein structural information. Critical Issues: Careful consideration of available analytical approaches is needed to broadly characterize the systems-level properties of redox signaling networks and be experimentally feasible. The cysteine redoxome is an informative focal point since it integrates many aspects of redox biology. The mechanisms and redox modules governing cysteine redox regulation, cysteine oxidation assays, proteome-wide annotation of the biophysical and biochemical properties of individual cysteines, and their clinical application are discussed. Future Directions: Investigating the cysteine redoxome at a systems level will uncover new insights into the mechanisms of selectivity and context dependence of redox signaling networks.

Identification of YdhV as the First Molybdoenzyme Binding a Bis-Mo-MPT Cofactor in Escherichia coli

The oxidoreductase YdhV in Escherichia coli has been predicted to belong to the family of molybdenum/tungsten cofactor (Moco/Wco)-containing enzymes. In this study, we characterized the YdhV protein in detail, which shares amino acid sequence homology with a tungsten-containing benzoyl-CoA reductase binding the bis-W-MPT (for metal-binding pterin) cofactor. The cofactor was identified to be of a bis-Mo-MPT type with no guanine nucleotides present, which represents a form of Moco that has not been found previously in any molybdoenzyme. Our studies showed that YdhV has a preference for bis-Mo-MPT over bis-W-MPT to be inserted into the enzyme. In-depth characterization of YdhV by X-ray absorption and electron paramagnetic resonance spectroscopies revealed that the bis-Mo-MPT cofactor in YdhV is redox active. The bis-Mo-MPT and bis-W-MPT cofactors include metal centers that bind the four sulfurs from the two dithiolene groups in addition to a cysteine and likely a sulfido ligand. The unexpected presence of a bis-Mo-MPT cofactor opens an additional route for cofactor biosynthesis in E. coli and expands the canon of the structurally highly versatile molybdenum and tungsten cofactors.

Applications of Reactive Cysteine Profiling

Cysteine thiols are involved in a diverse set of biological transformations, including nucleophilic and redox catalysis, metal coordination and formation of both dynamic and structural disulfides. Often posttranslationally modified, cysteines are also frequently alkylated by electrophilic compounds, including electrophilic metabolites, drugs, and natural products, and are attractive sites for covalent probe and drug development. Quantitative proteomics combined with activity-based protein profiling has been applied to annotate cysteine reactivity, susceptibility to posttranslational modifications, and accessibility to chemical probes, uncovering thousands of functional and small-molecule targetable cysteines across a diverse set of proteins, proteome-wide in an unbiased manner. Reactive cysteines have been targeted by high-throughput screening and fragment-based ligand discovery efforts. New cysteine-reactive electrophiles and compound libraries have been synthesized to enable inhibitor discovery broadly and to minimize nonspecific toxicity and off-target activity of compounds. With the recent blockbuster success of several covalent inhibitors, and the development of new chemical proteomic strategies to broadly identify reactive, ligandable and posttranslationally modified cysteines, cysteine profiling is poised to enable the development of new potent and selective chemical probes and even, in some cases, new drugs.