Toward a systems-level view of dynamic phosphorylation networks

A Review of Machine Learning and Algorithmic Methods for Protein Phosphorylation Site Prediction

Post-translational modifications (PTMs) have key roles in extending the functional diversity of proteins and, as a result, regulating diverse cellular processes in prokaryotic and eukaryotic organisms. Phosphorylation modification is a vital PTM that occurs in most proteins and plays a significant role in many biological processes. Disorders in the phosphorylation process lead to multiple diseases, including neurological disorders and cancers. The purpose of this review is to organize this body of knowledge associated with phosphorylation site (p-site) prediction to facilitate future research in this field. At first, we comprehensively review all related databases and introduce all steps regarding dataset creation, data preprocessing, and method evaluation in p-site prediction. Next, we investigate p-site prediction methods, which are divided into two computational groups: algorithmic and machine learning (ML). Additionally, it is shown that there are basically two main approaches for p-site prediction by ML: conventional and end-to-end deep learning methods, both of which are given an overview. Moreover, this review introduces the most important feature extraction techniques, which have mostly been used in p-site prediction. Finally, we create three test sets from new proteins related to the released version of the database of protein post-translational modifications (dbPTM) in 2022 based on general and human species. Evaluating online p-site prediction tools on newly added proteins introduced in the dbPTM 2022 release, distinct from those in the dbPTM 2019 release, reveals their limitations. In other words, the actual performance of these online p-site prediction tools on unseen proteins is notably lower than the results reported in their respective research papers.

Hydrogen peroxide-dependent oxidation of ERK2 within its D-recruitment site alters its substrate selection

Extracellular signal-regulated kinases 1 and 2 (ERK1/2) are dysregulated in many pervasive diseases. Recently, we discovered that ERK1/2 is oxidized by signal-generated hydrogen peroxide in various cell types. Since the putative sites of oxidation lie within or near ERK1/2's ligand-binding surfaces, we investigated how oxidation of ERK2 regulates interactions with the model substrates Sub-D and Sub-F. These studies revealed that ERK2 undergoes sulfenylation at C159 on its D-recruitment site surface and that this modification modulates ERK2 activity differentially between substrates. Integrated biochemical, computational, and mutational analyses suggest a plausible mechanism for peroxide-dependent changes in ERK2-substrate interactions. Interestingly, oxidation decreased ERK2's affinity for some D-site ligands while increasing its affinity for others. Finally, oxidation by signal-generated peroxide enhanced ERK1/2's ability to phosphorylate ribosomal S6 kinase A1 (RSK1) in HeLa cells. Together, these studies lay the foundation for examining crosstalk between redox- and phosphorylation-dependent signaling at the level of kinase-substrate selection.

The alphavirus nonstructural protein 2 NTPase induces a host translational shut-off through phosphorylation of eEF2 via cAMP-PKA-eEF2K signaling

Chikungunya virus (CHIKV) is a reemerging alphavirus. Since 2005, it has infected millions of people during outbreaks in Africa, Asia, and South/Central America. CHIKV replication depends on host cell factors at many levels and is expected to have a profound effect on cellular physiology. To obtain more insight into host responses to infection, stable isotope labeling with amino acids in cell culture and liquid chromatography-tandem mass spectrometry were used to assess temporal changes in the cellular phosphoproteome during CHIKV infection. Among the ~3,000 unique phosphorylation sites analyzed, the largest change in phosphorylation status was measured on residue T56 of eukaryotic elongation factor 2 (eEF2), which showed a >50-fold increase at 8 and 12 h p.i. Infection with other alphaviruses (Semliki Forest, Sindbis and Venezuelan equine encephalitis virus (VEEV)) triggered a similarly strong eEF2 phosphorylation. Expression of a truncated form of CHIKV or VEEV nsP2, containing only the N-terminal and NTPase/helicase domains (nsP2-NTD-Hel), sufficed to induce eEF2 phosphorylation, which could be prevented by mutating key residues in the Walker A and B motifs of the NTPase domain. Alphavirus infection or expression of nsP2-NTD-Hel resulted in decreased cellular ATP levels and increased cAMP levels. This did not occur when catalytically inactive NTPase mutants were expressed. The wild-type nsP2-NTD-Hel inhibited cellular translation independent of the C-terminal nsP2 domain, which was previously implicated in directing the virus-induced host shut-off for Old World alphaviruses. We hypothesize that the alphavirus NTPase activates a cellular adenylyl cyclase resulting in increased cAMP levels, thus activating PKA and subsequently eukaryotic elongation factor 2 kinase. This in turn triggers eEF2 phosphorylation and translational inhibition. We conclude that the nsP2-driven increase of cAMP levels contributes to the alphavirus-induced shut-off of cellular protein synthesis that is shared between Old and New World alphaviruses. MS Data are available via ProteomeXchange with identifier PXD009381.

Inferring signaling pathways with probabilistic programming

Motivation

Cells regulate themselves via dizzyingly complex biochemical processes called signaling pathways. These are usually depicted as a network, where nodes represent proteins and edges indicate their influence on each other. In order to understand diseases and therapies at the cellular level, it is crucial to have an accurate understanding of the signaling pathways at work. Since signaling pathways can be modified by disease, the ability to infer signaling pathways from condition- or patient-specific data is highly valuable. A variety of techniques exist for inferring signaling pathways. We build on past works that formulate signaling pathway inference as a Dynamic Bayesian Network structure estimation problem on phosphoproteomic time course data. We take a Bayesian approach, using Markov Chain Monte Carlo to estimate a posterior distribution over possible Dynamic Bayesian Network structures. Our primary contributions are (i) a novel proposal distribution that efficiently samples sparse graphs and (ii) the relaxation of common restrictive modeling assumptions.

Results

We implement our method, named Sparse Signaling Pathway Sampling, in Julia using the Gen probabilistic programming language. Probabilistic programming is a powerful methodology for building statistical models. The resulting code is modular, extensible and legible. The Gen language, in particular, allows us to customize our inference procedure for biological graphs and ensure efficient sampling. We evaluate our algorithm on simulated data and the HPN-DREAM pathway reconstruction challenge, comparing our performance against a variety of baseline methods. Our results demonstrate the vast potential for probabilistic programming, and Gen specifically, for biological network inference.

Availability and implementation

Find the full codebase at https://github.com/gitter-lab/ssps.

Supplementary information

Supplementary data are available at Bioinformatics online.

Phosphoproteomics Meets Chemical Genetics: Approaches for Global Mapping and Deciphering the Phosphoproteome

Protein kinases are important enzymes involved in the regulation of various cellular processes. To function properly, each protein kinase phosphorylates only a limited number of proteins among the thousands present in the cell. This provides a rapid and dynamic regulatory mechanism that controls biological functions of the proteins. Despite the importance of protein kinases, most of their substrates remain unknown. Recently, the advances in the fields of protein engineering, chemical genetics, and mass spectrometry have boosted studies on identification of bona fide substrates of protein kinases. Among the various methods in protein kinase specific substrate identification, genetically engineered protein kinases and quantitative phosphoproteomics have become promising tools. Herein, we review the current advances in the field of chemical genetics in analog-sensitive protein kinase mutants and highlight selected strategies for identifying protein kinase substrates and studying the dynamic nature of protein phosphorylation.

Illuminating the kinome: Visualizing real-time kinase activity in biological systems using genetically encoded fluorescent protein-based biosensors

Genetically encoded fluorescent protein-based kinase biosensors are a central tool for illumination of the kinome. The adaptability and versatility of biosensors have allowed for spatiotemporal observation of real-time kinase activity in living cells and organisms. In this review, we highlight various types of kinase biosensors, along with their burgeoning applications in complex biological systems. Specifically, we focus on kinase activity reporters used in neuronal systems and whole animal settings. Genetically encoded kinase biosensors are key for elucidation of the spatiotemporal regulation of protein kinases, with broader applications beyond the Petri dish.

In Silico Tools and Phosphoproteomic Software Exclusives

Proteomics and phosphoproteomics have been emerging as new dimensions of omics. Phosphorylation has a profound impact on the biological functions and applications of proteins. It influences everything from intrinsic activity and extrinsic executions to cellular localization. This post-translational modification has been subjected to detailed study and has been an object of analytical curiosity with the advent of faster instrumentation. The major strength of phosphoproteomic research lies in the fact that it gives an overall picture of the workforce of the cell. Phosphoproteomics gives deeper insights into understanding the mechanism behind development and progression of a disease. This review for the first time consolidates the list of existing bioinformatics tools developed for phosphoproteomics. The gap between development of bioinformatics tools and their implementation in clinical research is highlighted. The challenge facing progress is ideally believed to be the interdisciplinary arena this field of research is associated with. For meaningful solutions and deliverables, these tools need to be implemented in clinical studies for obtaining answers to pharmacodynamic questions, saving time, costs and energy. This review hopes to invoke some thought in this direction.

rSNPBase 3.0: an updated database of SNP-related regulatory elements, element-gene pairs and SNP-based gene regulatory networks

Here, we present the updated rSNPBase 3.0 database (http://rsnp3.psych.ac.cn), which provides human SNP-related regulatory elements, element-gene pairs and SNP-based regulatory networks. This database is the updated version of the SNP regulatory annotation database rSNPBase and rVarBase. In comparison to the last two versions, there are both structural and data adjustments in rSNPBase 3.0: (i) The most significant new feature is the expansion of analysis scope from SNP-related regulatory elements to include regulatory element–target gene pairs (E–G pairs), therefore it can provide SNP-based gene regulatory networks. (ii) Web function was modified according to data content and a new network search module is provided in the rSNPBase 3.0 in addition to the previous regulatory SNP (rSNP) search module. The two search modules support data query for detailed information (related-elements, element-gene pairs, and other extended annotations) on specific SNPs and SNP-related graphic networks constructed by interacting transcription factors (TFs), miRNAs and genes. (3) The type of regulatory elements was modified and enriched. To our best knowledge, the updated rSNPBase 3.0 is the first data tool supports SNP functional analysis from a regulatory network prospective, it will provide both a comprehensive understanding and concrete guidance for SNP-related regulatory studies.

Channelrhodopsin-1 Phosphorylation Changes with Phototactic Behavior and Responds to Physiological Stimuli in Chlamydomonas

The unicellular alga Chlamydomonas (Chlamydomonas reinhardtii) exhibits oriented movement responses (phototaxis) to light over more than three log units of intensity. Phototaxis thus depends on the cell's ability to adjust the sensitivity of its photoreceptors to ambient light conditions. In Chlamydomonas, the photoreceptors for phototaxis are the channelrhodopsins (ChR)1 and ChR2; these light-gated cation channels are located in the plasma membrane. Although ChRs are widely used in optogenetic studies, little is known about ChR signaling in algae. We characterized the in vivo phosphorylation of ChR1. Its reversible phosphorylation occurred within seconds as a graded response to changes in the light intensity and ionic composition of the medium and depended on an elevated cytosolic Ca2+ concentration. Changes in the phototactic sign were accompanied by alterations in the phosphorylation status of ChR1. Furthermore, compared with the wild type, a permanently negative phototactic mutant required higher light intensities to evoke ChR1 phosphorylation. C-terminal truncation of ChR1 disturbed its reversible phosphorylation, whereas it was normal in ChR2-knockout and eyespot-assembly mutants. The identification of phosphosites in regions important for ChR1 function points to their potential regulatory role(s). We propose that multiple ChR1 phosphorylation, regulated via a Ca2+-based feedback loop, is an important component in the adaptation of phototactic sensitivity in Chlamydomonas.

Genetically Encoded Fluorescent Biosensors Illuminate the Spatiotemporal Regulation of Signaling Networks

Cellular signaling networks are the foundation which determines the fate and function of cells as they respond to various cues and stimuli. The discovery of fluorescent proteins over 25 years ago enabled the development of a diverse array of genetically encodable fluorescent biosensors that are capable of measuring the spatiotemporal dynamics of signal transduction pathways in live cells. In an effort to encapsulate the breadth over which fluorescent biosensors have expanded, we endeavored to assemble a comprehensive list of published engineered biosensors, and we discuss many of the molecular designs utilized in their development. Then, we review how the high temporal and spatial resolution afforded by fluorescent biosensors has aided our understanding of the spatiotemporal regulation of signaling networks at the cellular and subcellular level. Finally, we highlight some emerging areas of research in both biosensor design and applications that are on the forefront of biosensor development.

Synthesizing Signaling Pathways from Temporal Phosphoproteomic Data

We present a method for automatically discovering signaling pathways from time-resolved phosphoproteomic data. The Temporal Pathway Synthesizer (TPS) algorithm uses constraint-solving techniques first developed in the context of formal verification to explore paths in an interaction network. It systematically eliminates all candidate structures for a signaling pathway where a protein is activated or inactivated before its upstream regulators. The algorithm can model more than one hundred thousand dynamic phosphosites and can discover pathway members that are not differentially phosphorylated. By analyzing temporal data, TPS defines signaling cascades without needing to experimentally perturb individual proteins. It recovers known pathways and proposes pathway connections when applied to the human epidermal growth factor and yeast osmotic stress responses. Independent kinase mutant studies validate predicted substrates in the TPS osmotic stress pathway.

Improved Method for Determining Absolute Phosphorylation Stoichiometry Using Bayesian Statistics and Isobaric Labeling

Phosphorylation stoichiometry, or occupancy, is one element of phosphoproteomics that can add useful biological context (Gerber et al. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 6940-5). We previously developed a method to assess phosphorylation stoichiometry on a proteome-wide scale (Wu et al. Nat. Methods 2011, 8, 677-83). The stoichiometry calculation relies on identifying and measuring the levels of each nonphosphorylated counterpart peptide with and without phosphatase treatment. The method, however, is problematic in that low stoichiometry phosphopeptides can return negative stoichiometry values if measurement error is larger than the percent stoichiometry. Here, we have improved the stoichiometry method through the use of isobaric labeling with 10-plex TMT reagents. In this way, five phosphatase treated and five untreated samples are compared simultaneously so that each stoichiometry is represented by five ratio measurements with no missing values. We applied the method to determine basal stoichiometries of HCT116 cells growing in culture. With this method, we analyzed five biological replicates simultaneously with no need for phosphopeptide enrichment. Additionally, we developed a Bayesian model to estimate phosphorylation stoichiometry as a parameter confined to an interval between 0 and 1 implemented as an R/Stan script. Consequently, both point and interval estimates are consistent with the plausible range of values for stoichiometry. Finally, we report absolute stoichiometry measurements with credible intervals for 6772 phosphopeptides containing at least a single phosphorylation site.

Phosphoproteomics-Based Profiling of Kinase Activities in Cancer Cells

Cellular signaling, predominantly mediated by phosphorylation through protein kinases, is found to be deregulated in most cancers. Accordingly, protein kinases have been subject to intense investigations in cancer research, to understand their role in oncogenesis and to discover new therapeutic targets. Despite great advances, an understanding of kinase dysfunction in cancer is far from complete.A powerful tool to investigate phosphorylation is mass-spectrometry (MS)-based phosphoproteomics, which enables the identification of thousands of phosphorylated peptides in a single experiment. Since every phosphorylation event results from the activity of a protein kinase, high-coverage phosphoproteomics data should indirectly contain comprehensive information about the activity of protein kinases.In this chapter, we discuss the use of computational methods to predict kinase activity scores from MS-based phosphoproteomics data. We start with a short explanation of the fundamental features of the phosphoproteomics data acquisition process from the perspective of the computational analysis. Next, we briefly review the existing databases with experimentally verified kinase-substrate relationships and present a set of bioinformatic tools to discover novel kinase targets. We then introduce different methods to infer kinase activities from phosphoproteomics data and these kinase-substrate relationships. We illustrate their application with a detailed protocol of one of the methods, KSEA (Kinase Substrate Enrichment Analysis). This method is implemented in Python within the framework of the open-source Kinase Activity Toolbox (kinact), which is freely available at http://github.com/saezlab/kinact/ .

RF-Hydroxysite: a random forest based predictor for hydroxylation sites

Protein hydroxylation is an emerging posttranslational modification involved in both normal cellular processes and a growing number of pathological states, including several cancers. Protein hydroxylation is mediated by members of the hydroxylase family of enzymes, which catalyze the conversion of an alkyne group at select lysine or proline residues on their target substrates to a hydroxyl. Traditionally, hydroxylation has been identified using expensive and time-consuming experimental methods, such as tandem mass spectrometry. Therefore, to facilitate identification of putative hydroxylation sites and to complement existing experimental approaches, computational methods designed to predict the hydroxylation sites in protein sequences have recently been developed. Building on these efforts, we have developed a new method, termed RF-hydroxysite, that uses random forest to identify putative hydroxylysine and hydroxyproline residues in proteins using only the primary amino acid sequence as input. RF-Hydroxysite integrates features previously shown to contribute to hydroxylation site prediction with several new features that we found to augment the performance remarkably. These include features that capture physicochemical, structural, sequence-order and evolutionary information from the protein sequences. The features used in the final model were selected based on their contribution to the prediction. Physicochemical information was found to contribute the most to the model. The present study also sheds light on the contribution of evolutionary, sequence order, and protein disordered region information to hydroxylation site prediction. The web server for RF-hydroxysite is available online at .

Measuring Kinase Activity-A Global Challenge

The kinase enzymes within a cell, known collectively as the kinome, play crucial roles in many signaling pathways, including survival, motility, differentiation, stress response, and many more. Aberrant signaling through kinase pathways is often linked to cancer, among other diseases. A major area of scientific research involves understanding the relationships between kinases, their targets, and how the kinome adapts to perturbations of the cellular system. This review will discuss many of the current and developing methods for studying kinase activity, and evaluate their applications, advantages, and disadvantages. J. Cell. Biochem. 118: 3595-3606, 2017. © 2017 Wiley Periodicals, Inc.

Integrated Strategies to Gain a Systems-Level View of Dynamic Signaling Networks

In order to survive and function properly in the face of an ever changing environment, cells must be able to sense changes in their surroundings and respond accordingly. Cells process information about their environment through complex signaling networks composed of many discrete signaling molecules. Individual pathways within these networks are often tightly integrated and highly dynamic, allowing cells to respond to a given stimulus (or, as is typically the case under physiological conditions, a combination of stimuli) in a specific and appropriate manner. However, due to the size and complexity of many cellular signaling networks, it is often difficult to predict how cellular signaling networks will respond under a particular set of conditions. Indeed, crosstalk between individual signaling pathways may lead to responses that are nonintuitive (or even counterintuitive) based on examination of the individual pathways in isolation. Therefore, to gain a more comprehensive view of cell signaling processes, it is important to understand how signaling networks behave at the systems level. This requires integrated strategies that combine quantitative experimental data with computational models. In this chapter, we first examine some of the progress that has recently been made toward understanding the systems-level regulation of cellular signaling networks, with a particular emphasis on phosphorylation-dependent signaling networks. We then discuss how genetically targetable fluorescent biosensors are being used together with computational models to gain unique insights into the spatiotemporal regulation of signaling networks within single, living cells.

RF-Phos: A Novel General Phosphorylation Site Prediction Tool Based on Random Forest

Protein phosphorylation is one of the most widespread regulatory mechanisms in eukaryotes. Over the past decade, phosphorylation site prediction has emerged as an important problem in the field of bioinformatics. Here, we report a new method, termed Random Forest-based Phosphosite predictor 2.0 (RF-Phos 2.0), to predict phosphorylation sites given only the primary amino acid sequence of a protein as input. RF-Phos 2.0, which uses random forest with sequence and structural features, is able to identify putative sites of phosphorylation across many protein families. In side-by-side comparisons based on 10-fold cross validation and an independent dataset, RF-Phos 2.0 compares favorably to other popular mammalian phosphosite prediction methods, such as PhosphoSVM, GPS2.1, and Musite.

Predicting CK2 beta-dependent substrates using linear patterns

CK2 is a constitutively active Ser/Thr protein kinase deregulated in cancer and other pathologies, responsible for about the 20% of the human phosphoproteome. The holoenzyme is a complex composed of two catalytic (α or α´) and two regulatory (β) subunits, with individual subunits also coexisting in the cell. In the holoenzyme, CK2β is a substrate-dependent modulator of kinase activity. Therefore, a comprehensive characterization of CK2 cellular function should firstly address which substrates are phosphorylated exclusively when CK2β is present (class-III or beta-dependent substrates). However, current experimental constrains limit this classification to a few substrates. Here, we took advantage of motif-based prediction and designed four linear patterns for predicting class-III behavior in sets of experimentally determined CK2 substrates. Integrating high-throughput substrate prediction, functional classification and network analysis, our results suggest that beta-dependent phosphorylation might exert particular regulatory roles in viral infection and biological processes/pathways like apoptosis, DNA repair and RNA metabolism. It also pointed, that human beta-dependent substrates are mainly nuclear, a few of them shuttling between nuclear and cytoplasmic compartments.

•

The designed linear patterns assist CK2 beta-dependent substrates prediction.

•

A high-throughput prediction of CK2 beta-dependent substrates was performed in several organisms including human, mouse and rat.

•

The functional classification indicated a role of CK2 beta-dependent regulation in viral infection, apoptosis, DNA repair and RNA metabolism.

•

The functional classification indicated that human CK2 beta-dependent substrates are mainly nuclear with a number of them also found in cytoplasm.