A White-Box Machine Learning Approach for Revealing Antibiotic Mechanisms of Action.

Current machine learning techniques enable robust association of biological signals with measured phenotypes, but these approaches are incapable of identifying causal relationships. Here, we develop an integrated "white-box" biochemical screening, network modeling, and machine learning approach for revealing causal mechanisms and apply this approach to understanding antibiotic efficacy. We counter-screen diverse metabolites against bactericidal antibiotics in Escherichia coli and simulate their corresponding metabolic states using a genome-scale metabolic network model. Regression of the measured screening data on model simulations reveals that purine biosynthesis participates in antibiotic lethality, which we validate experimentally. We show that antibiotic-induced adenine limitation increases ATP demand, which elevates central carbon metabolism activity and oxygen consumption, enhancing the killing effects of antibiotics. This work demonstrates how prospective network modeling can couple with machine learning to identify complex causal mechanisms underlying drug efficacy.

Integrative omics indicate FMRP sequesters mRNA from translation and deadenylation in human neuronal cells.

How fragile X syndrome protein (FMRP) binds mRNAs and regulates mRNA metabolism remains unclear. Our previous work using human neuronal cells focused on mRNAs targeted for nonsense-mediated mRNA decay (NMD), which we showed are generally bound by FMRP and destabilized upon FMRP loss. Here, we identify >400 high-confidence FMRP-bound mRNAs, only ∼35% of which are NMD targets. Integrative transcriptomics together with SILAC-LC-MS/MS reveal that FMRP loss generally results in mRNA destabilization and more protein produced per FMRP target. We use our established RIP-seq technology to show that FMRP footprints are independent of protein-coding potential, target GC-rich and structured sequences, and are densest in 5' UTRs. Regardless of where within an mRNA FMRP binds, we find that FMRP protects mRNAs from deadenylation and directly binds the cytoplasmic poly(A)-binding protein. Our results reveal how FMRP sequesters polyadenylated mRNAs into stabilized and translationally repressed complexes, whose regulation is critical for neurogenesis and synaptic plasticity.

Complementary environmental analysis and functional characterization of lower glycolysis-gluconeogenesis in the diatom plastid.

Organic carbon fixed in chloroplasts through the Calvin-Benson-Bassham Cycle can be diverted towards different metabolic fates, including cyoplasmic and mitochondrial respiration, gluconeogenesis, and synthesis of diverse plastid metabolites via the pyruvate hub. In plants, pyruvate is principally produced via cytoplasmic glycolysis, although a plastid-targeted lower glycolytic pathway is known to exist in non-photosynthetic tissue. Here, we characterized a lower plastid glycolysis-gluconeogenesis pathway enabling the direct interconversion of glyceraldehyde-3-phosphate and phospho-enol-pyruvate in diatoms, ecologically important marine algae distantly related to plants. We show that two reversible enzymes required to complete diatom plastid glycolysis-gluconeogenesis, Enolase and bis-phospho-glycerate mutase (PGAM), originated through duplications of mitochondria-targeted respiratory isoforms. Through CRISPR-Cas9 mutagenesis, integrative 'omic analyses, and measured kinetics of expressed enzymes in the diatom Phaeodactylum tricornutum, we present evidence that this pathway diverts plastid glyceraldehyde-3-phosphate into the pyruvate hub, and may also function in the gluconeogenic direction. Considering experimental data, we show that this pathway has different roles dependent in particular on day length and environmental temperature, and show that the cpEnolase and cpPGAM genes are expressed at elevated levels in high latitude oceans where diatoms are abundant. Our data provide evolutionary, meta-genomic and functional insights into a poorly understood yet evolutionarily recurrent plastid metabolic pathway.

Multi-omics analysis reveals the potential pathogenesis and therapeutic targets of diabetic kidney disease.

Clinicians have long been interested in understanding the molecular basis of diabetic kidney disease (DKD)and its potential treatment targets. Its pathophysiology involves protein phosphorylation, one of the most recognizable post-transcriptional modifications, that can take part in many cellular functions and control different metabolic processes. In order to recognize the molecular and protein changes of DKD kidney, this study applied Tandem liquid chromatography-mass spectrometry (LC-MS/MS) and Next-Generation Sequencing, along with Tandem Mass Tags (TMT) labeling techniques to evaluate the mRNA, protein and modified phosphorylation sites between DKD mice and model ones. Based on Gene Ontology (GO) and KEGG pathway analyses of transcriptome and proteome, The molecular changes of DKD include accumulation of extracellular matrix, abnormally activated inflammatory microenvironment, oxidative stress and lipid metabolism disorders, leading to glomerulosclerosis and tubulointerstitial fibrosis. Oxidative stress has been emphasized as an important factor in DKD and progression to ESKD, which is directly related to podocyte injury, albuminuria and renal tubulointerstitial fibrosis. A histological study of phosphorylation further revealed that kinases were crucial. Three groups of studies have found that RAS signaling pathway, RAP1 signaling pathway, AMPK signaling pathway, PPAR signaling pathway and HIF-1 signaling pathway were crucial for the pathogenesis of DKD. Through this approach, it was discovered that targeting specific molecules, proteins, kinases and critical pathways could be a promising approach for treating DKD.

CLIPPER 2.0: Peptide-Level Annotation and Data Analysis for Positional Proteomics.

Positional proteomics methodologies have transformed protease research, and have brought mass spectrometry (MS)-based degradomics studies to the forefront of protease characterization and system-wide interrogation of protease signaling. Considerable advancements in both sensitivity and throughput of liquid chromatography (LC)-MS/MS instrumentation enable the generation of enormous positional proteomics datasets of natural and protein termini and neo-termini of cleaved protease substrates. However, concomitant progress has not been observed to the same extent in data analysis and post-processing steps, arguably constituting the largest bottleneck in positional proteomics workflows. Here, we present a computational tool, CLIPPER 2.0, that builds on prior algorithms developed for MS-based protein termini analysis, facilitating peptide-level annotation and data analysis. CLIPPER 2.0 can be used with several sample preparation workflows and proteomics search algorithms and enables fast and automated database information retrieval, statistical and network analysis, as well as visualization of terminomic datasets. We demonstrate the applicability of our tool by analyzing GluC and MMP9 cleavages in HeLa lysates. CLIPPER 2.0 is available at https://github.com/UadKLab/CLIPPER-2.0.

An Automated and Fast Sample Preparation Workflow for Laser Microdissection Guided Ultrasensitive Proteomics.

Spatial tissue proteomics integrating whole-slide imaging, laser microdissection, and ultrasensitive mass spectrometry is a powerful approach to link cellular phenotypes to functional proteome states in (patho)physiology. To be applicable to large patient cohorts and low sample input amounts, including single-cell applications, loss-minimized and streamlined end-to-end workflows are key. We here introduce an automated sample preparation protocol for laser microdissected samples utilizing the cellenONE robotic system, which has the capacity to process 192 samples in 3 h. Following laser microdissection collection directly into the proteoCHIP LF 48 or EVO 96 chip, our optimized protocol facilitates lysis, formalin de-crosslinking, and tryptic digest of low-input archival tissue samples. The seamless integration with the Evosep ONE LC system by centrifugation allows 'on-the-fly' sample clean-up, particularly pertinent for laser microdissection workflows. We validate our method in human tonsil archival tissue, where we profile proteomes of spatially-defined B-cell, T-cell, and epithelial microregions of 4000 µm2 to a depth of ∼2000 proteins and with high cell type specificity. We finally provide detailed equipment templates and experimental guidelines for broad accessibility.

Galectin-3-Binding Protein Inhibits Extracellular Heparan 6-O-Endosulfatase Sulf-2.

Human extracellular 6-O-endosulfatases Sulf-1 and Sulf-2 are the only enzymes that post-synthetically alter the 6-O sulfation of heparan sulfate proteoglycans (HSPG), which regulates interactions of HSPG with many proteins. Oncogenicity of Sulf-2 in different cancers has been documented, and we have shown that Sulf-2 is associated with poor survival outcomes in head and neck squamous cell carcinoma (HNSCC). Despite its importance, limited information is available on direct protein-protein interactions of the Sulf-2 protein in the tumor microenvironment. In this study, we used monoclonal antibody (mAb) affinity purification and mass spectrometry to identify galectin-3-binding protein (LG3BP) as a highly specific binding partner of Sulf-2 in the conditioned media of HNSCC cell lines. We validated their direct interaction in vitro using recombinant proteins and have shown that the chondroitin sulfate (CS) covalently bound to the Sulf-2 influences the binding to LG3BP. We confirmed the importance of the CS chain for the interaction by generating a mutant Sulf-2 protein that lacks the CS. Importantly, we have shown that the LG3BP inhibits Sulf-2 activity in vitro in a concentration-dependent manner. As a consequence, the addition of LG3BP to a spheroid cell culture inhibited the invasion of the HNSCC cells into Matrigel. Thus, Sulf-2 interaction with LG3BP may regulate the physiological activity of the Sulf-2 enzyme as well as its activity in the tumor microenvironment.

Direct sequencing of total Saccharomyces cerevisiae tRNAs by LC-MS/MS.

Among RNAs, transfer RNAs (tRNAs) contain the widest variety of abundant posttranscriptional chemical modifications. These modifications are crucial for tRNAs to participate in protein synthesis, promoting proper tRNA structure and aminoacylation, facilitating anticodon:codon recognition, and ensuring the reading frame maintenance of the ribosome. While tRNA modifications were long thought to be stoichiometric, it is becoming increasingly apparent that these modifications can change dynamically in response to the cellular environment. The ability to broadly characterize the fluctuating tRNA modification landscape will be essential for establishing the molecular level contributions of individual sites of tRNA modification. The locations of modifications within individual tRNA sequences can be mapped using liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). In this approach, a single tRNA species is purified, treated with ribonucleases, and the resulting single-stranded RNA products are subject to LC-MS/MS analysis. The application of LC-MS/MS to study tRNAs is limited by the necessity of analyzing one tRNA at a time, because the digestion of total tRNA mixtures by commercially available ribonucleases produces many short digestion products unable to be uniquely mapped back to a single site within a tRNA. We overcame these limitations by taking advantage of the highly structured nature of tRNAs to prevent the full digestion by single-stranded RNA-specific ribonucleases. Folding total tRNA prior to digestion allowed us to sequence Saccharomyces cerevisiae tRNAs with up to 97% sequence coverage for individual tRNA species by LC-MS/MS. This method presents a robust avenue for directly detecting the distribution of modifications in total tRNAs.

Comprehensive investigation of pathway enrichment methods for functional interpretation of LC-MS global metabolomics data.

BACKGROUND: Global or untargeted metabolomics is widely used to comprehensively investigate metabolic profiles under various pathophysiological conditions such as inflammations, infections, responses to exposures or interactions with microbial communities. However, biological interpretation of global metabolomics data remains a daunting task. Recent years have seen growing applications of pathway enrichment analysis based on putative annotations of liquid chromatography coupled with mass spectrometry (LC-MS) peaks for functional interpretation of LC-MS-based global metabolomics data. However, due to intricate peak-metabolite and metabolite-pathway relationships, considerable variations are observed among results obtained using different approaches. There is an urgent need to benchmark these approaches to inform the best practices. RESULTS: We have conducted a benchmark study of common peak annotation approaches and pathway enrichment methods in current metabolomics studies. Representative approaches, including three peak annotation methods and four enrichment methods, were selected and benchmarked under different scenarios. Based on the results, we have provided a set of recommendations regarding peak annotation, ranking metrics and feature selection. The overall better performance was obtained for the mummichog approach. We have observed that a ~30% annotation rate is sufficient to achieve high recall (~90% based on mummichog), and using semi-annotated data improves functional interpretation. Based on the current platforms and enrichment methods, we further propose an identifiability index to indicate the possibility of a pathway being reliably identified. Finally, we evaluated all methods using 11 COVID-19 and 8 inflammatory bowel diseases (IBD) global metabolomics datasets.

Advances in targeted liquid chromatography-tandem mass spectrometry methods for endocannabinoid and N-acylethanolamine quantification in biological matrices: A systematic review.

Liquid chromatography paired with tandem mass spectrometry (LC-MS/MS) is the gold standard in measurement of endocannabinoid concentrations in biomatrices. We conducted a systematic review of literature to identify advances in targeted LC-MS/MS methods in the period 2017-2024. We found that LC-MS/MS methods for endocannabinoid quantification are relatively consistent both across time and across biomatrices. Recent advances have primarily been in three areas: (1) sample preparation techniques, specific to the chosen biomatrix; (2) the range of biomatrices tested, recently favoring blood matrices; and (3) the breadth of endocannabinoid and endocannabinoid-like analytes incorporated into assays. This review provides a summary of the recent literature and a guide for researchers looking to establish the best methods for quantifying endocannabinoids in a range of biomatrices.

O-Glycomic and Proteomic Signatures of Spontaneous and Butyrate-Stimulated Colorectal Cancer Cell Line Differentiation.

Gut microbiota of the gastrointestinal tract provide health benefits to the human host via bacterial metabolites. Bacterial butyrate has beneficial effects on intestinal homeostasis and is the preferred energy source of intestinal epithelial cells, capable of inducing differentiation. It was previously observed that changes in the expression of specific proteins as well as protein glycosylation occur with differentiation. In this study, specific mucin O-glycans were identified that mark butyrate-induced epithelial differentiation of the intestinal cell line CaCo-2 (Cancer Coli-2), by applying porous graphitized carbon nano-liquid chromatography with electrospray ionization tandem mass spectrometry. Moreover, a quantitative proteomic approach was used to decipher changes in the cell proteome. It was found that the fully differentiated butyrate-stimulated cells are characterized by a higher expression of sialylated O-glycan structures, whereas fucosylation is downregulated with differentiation. By performing an integrative approach, we generated hypotheses about the origin of the observed O-glycome changes. These insights pave the way for future endeavors to study the dynamic O-glycosylation patterns in the gut, either produced via cellular biosynthesis or through the action of bacterial glycosidases as well as the functional role of these patterns in homeostasis and dysbiosis at the gut-microbiota interface.

HowDirty: An R package to evaluate molecular contaminants in LC-MS experiments.

Contaminants derived from consumables, reagents, and sample handling often negatively affect LC-MS data acquisition. In proteomics experiments, they can markedly reduce identification performance, reproducibility, and quantitative robustness. Here, we introduce a data analysis workflow combining MS1 feature extraction in Skyline with HowDirty, an R-markdown-based tool, that automatically generates an interactive report on the molecular contaminant level in LC-MS data sets. To facilitate the interpretation of the results, the HTML report is self-contained and self-explanatory, including plots that can be easily interpreted. The R package HowDirty is available from https://github.com/DavidGZ1/HowDirty. To demonstrate a showcase scenario for the application of HowDirty, we assessed the impact of ultrafiltration units from different providers on sample purity after filter-assisted sample preparation (FASP) digestion. This allowed us to select the filter units with the lowest contamination risk. Notably, the filter units with the lowest contaminant levels showed higher reproducibility regarding the number of peptides and proteins identified. Overall, HowDirty enables the efficient evaluation of sample quality covering a wide range of common contaminant groups that typically impair LC-MS analyses, facilitating corrective or preventive actions to minimize instrument downtime.

On the excessive use of coefficient of variation as a metric of quantitation quality in proteomics.

The coefficient of variation (CV) is often used in proteomics as a proxy to characterize the performance of a quantitation method and/or the related software. In this note, we question the excessive reliance on this metric in quantitative proteomics that may result in erroneous conclusions. We support this note using a ground-truth Human-Yeast-E. coli dataset demonstrating in a number of cases that erroneous data processing methods may lead to a low CV which has nothing to do with these methods' performances in quantitation.

Global profiling of protein lactylation in Caenorhabditis elegans.

Lactylation, as a novel posttranslational modification, is essential for studying the functions and regulation of proteins in physiological and pathological processes, as well as for gaining in-depth knowledge on the occurrence and development of many diseases, including tumors. However, few studies have examined the protein lactylation of one whole organism. Thus, we studied the lactylation of global proteins in Caenorhabditis elegans to obtain an in vivo lactylome. Using an MS-based platform, we identified 1836 Class I (localization probabilities > 0.75) lactylated sites in 487 proteins. Bioinformatics analysis showed that lactylated proteins were mainly located in the cytoplasm and involved in the tricarboxylic acid cycle (TCA cycle) and other metabolic pathways. Then, we evaluated the conservation of lactylation in different organisms. In total, 41 C. elegans proteins were lactylated and homologous to lactylated proteins in humans and rats. Moreover, lactylation on H4K80 was conserved in three species. An additional 238 lactylated proteins were identified in C. elegans for the first time. This study establishes the first lactylome database in C. elegans and provides a basis for studying the role of lactylation.

Identification of a rare [Gγ(Aγδß)0] -thalassemia using tandem mass spectrometry.

Thalassemias are a group of inherited monogenic disorders characterized by defects in the synthesis of one or more of the globin chain subunits of the hemoglobin tetramer. Delta-beta (δß-) thalassemia has large deletions in the ß globin gene cluster involving δ- and ß-globin genes, leading to absent or reduced synthesis of both δ- and ß-globin chains. Here, we used direct globin-chain analysis using tandem mass spectrometry for the diagnosis of δß-thalassemia. Two cases from unrelated families were recruited for the study based on clinical and hematological evaluation. Peptides obtained after trypsin digestion of proteins extracted from red blood cell pellets from two affected individuals and their parents were analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Mass spectrometric analysis revealed a severe reduction in δ, ß, and Aγ globin proteins with increased Gγ globin protein in the affected individuals. The diagnosis of Gγ(Aγδß)0 -thalassemia in the homozygous state in the affected individuals and in the heterozygous state in the parents was made from our results. The diagnosis was confirmed at the genetic level using multiplex ligation-dependent probe amplification (MLPA). Our findings demonstrate the utility of direct globin protein quantitation using LC-MS/MS to quantify individual globin proteins reflecting changes in globin production. This approach can be utilized for accurate and timely diagnosis of hemoglobinopathies, including rare variants, where existing diagnostic methods provide inconclusive results.

Oktoberfest: Open-source spectral library generation and rescoring pipeline based on Prosit.

Machine learning (ML) and deep learning (DL) models for peptide property prediction such as Prosit have enabled the creation of high quality in silico reference libraries. These libraries are used in various applications, ranging from data-independent acquisition (DIA) data analysis to data-driven rescoring of search engine results. Here, we present Oktoberfest, an open source Python package of our spectral library generation and rescoring pipeline originally only available online via ProteomicsDB. Oktoberfest is largely search engine agnostic and provides access to online peptide property predictions, promoting the adoption of state-of-the-art ML/DL models in proteomics analysis pipelines. We demonstrate its ability to reproduce and even improve our results from previously published rescoring analyses on two distinct use cases. Oktoberfest is freely available on GitHub (https://github.com/wilhelm-lab/oktoberfest) and can easily be installed locally through the cross-platform PyPI Python package.

Identification of the native Torpedo californica nicotinic acetylcholine receptor's glycan composition after a multi-step sequential purification method using MALDI-ToF MS.

The Cys-loop pentameric ligand-gated ion channels comprise a dynamic group of proteins that have been extensively studied for decades, yielding a wealth of findings at both the structural and functional levels. The nicotinic acetylcholine receptor (nAChR) is no exception, as it is part of this large protein family involved in proper organismal function. Our efforts have successfully produced a highly pure nAChR in detergent complex (nAChR-DC), enabling more robust studies to be conducted on it, including beginning to experiment with high-throughput crystallization. Our homogeneous product has been identified and extensively characterized with 100% identity using Nano Lc MS/MS and MALDI ToF/ToF for each nAChR subunit. Additionally, the N-linked glycans in the Torpedo californica-nAChR (Tc-nAChR) subunits have been identified. To study this, the Tc-nAChR subunits were digested with PNGase F and the released glycans were analyzed by MALDI-ToF. The MS results showed the presence of high-mannose N-glycan in all native Tc-nAChR subunits. Specifically, the oligommanose population Man8-9GlcNac2 with peaks at m/z 1742 and 1904 ([M + Na]+ ions) were observed.

Quantitative proteomics investigating the intrinsic adaptation mechanism of Aeromonas hydrophila to streptomycin.

Aeromonas hydrophila, a prevalent pathogen in the aquaculture industry, poses significant challenges due to its drug-resistant strains. Moreover, residues of antibiotics like streptomycin, extensively employed in aquaculture settings, drive selective bacterial evolution, leading to the progressive development of resistance to this agent. However, the underlying mechanism of its intrinsic adaptation to antibiotics remains elusive. Here, we employed a quantitative proteomics approach to investigate the differences in protein expression between A. hydrophila under streptomycin (SM) stress and nonstress conditions. Notably, bioinformatics analysis unveiled the potential involvement of metal pathways, including metal cluster binding, iron-sulfur cluster binding, and transition metal ion binding, in influencing A. hydrophila's resistance to SM. Furthermore, we evaluated the sensitivity of eight gene deletion strains related to streptomycin and observed the potential roles of petA and AHA_4705 in SM resistance. Collectively, our findings enhance the understanding of A. hydrophila's response behavior to streptomycin stress and shed light on its intrinsic adaptation mechanism.

Dysregulated proteome and N-glycoproteome in ALG1-deficient fibroblasts.

Asparagine-linked glycosylation 1 protein is a ß-1,4-mannosyltransferase, is encoded by the ALG1 gene, which catalyzes the first step of mannosylation in N-glycosylation. Pathogenic variants in ALG1 cause a rare autosomal recessive disorder termed as ALG1-CDG. We performed a quantitative proteomics and N-glycoproteomics study in fibroblasts derived from patients with one homozygous and two compound heterozygous pathogenic variants in ALG1. Several proteins that exhibited significant upregulation included insulin-like growth factor II and pleckstrin, whereas hyaluronan and proteoglycan link protein 1 was downregulated. These proteins are crucial for cell growth, survival and differentiation. Additionally, we observed a decrease in the expression of mitochondrial proteins and an increase in autophagy-related proteins, suggesting mitochondrial and cellular stress. N-glycoproteomics revealed the reduction in high-mannose and complex/hybrid glycopeptides derived from numerous proteins in patients explaining that defect in ALG1 has broad effects on glycosylation. Further, we detected an increase in several short oligosaccharides, including chitobiose (HexNAc2 ) trisaccharides (Hex-HexNAc2 ) and novel tetrasaccharides (NeuAc-Hex-HexNAc2 ) derived from essential proteins including LAMP1, CD44 and integrin. These changes in glycosylation were observed in all patients irrespective of their gene variants. Overall, our findings not only provide novel molecular insights into understanding ALG1-CDG but also offer short oligosaccharide-bearing peptides as potential biomarkers.

Lipidomics and mass spectrometry imaging unveil alterations in mice hippocampus lipid composition exposed to hypoxia.

Lipids are components of cytomembranes that are involved in various biochemical processes. High-altitude hypoxic environments not only affect the body's energy metabolism, but these environments can also cause abnormal lipid metabolism involved in the hypoxia-induced cognitive impairment. Thus, comprehensive lipidomic profiling of the brain tissue is an essential step toward understanding the mechanism of cognitive impairment induced by hypoxic exposure. In the present study, mice showed reduced new-object recognition and spatial memory when exposed to hypobaric hypoxia for 1 day. Histomorphological staining revealed significant morphological and structural damage to the hippocampal tissue, along with prolonged exposure to hypobaric hypoxia. Dynamic lipidomics of the mouse hippocampus showed a significant shift in both the type and distribution of phospholipids, as verified by spatial lipid mapping. Collectively, a diverse and dynamic lipid composition in mice hippocampus was uncovered, which deepens our understanding of biochemical changes during sustained hypoxic exposure and could provide new insights into the cognitive decline induced by high-altitude hypoxia exposure.