Quantitative proteome-wide O-glycoproteomics analysis with FragPipe.

Identification of O-glycopeptides from tandem mass spectrometry data is complicated by the near complete dissociation of O-glycans from the peptide during collisional activation and by the combinatorial explosion of possible glycoforms when glycans are retained intact in electron-based activation. The recent O-Pair search method provides an elegant solution to these problems, using a collisional activation scan to identify the peptide sequence and total glycan mass, and a follow-up electron-based activation scan to localize the glycosite(s) using a graph-based algorithm in a reduced search space. Our previous O-glycoproteomics methods with MSFragger-Glyco allowed for extremely fast and sensitive identification of O-glycopeptides from collisional activation data but had limited support for site localization of glycans and quantification of glycopeptides. Here, we report an improved pipeline for O-glycoproteomics analysis that provides proteome-wide, site-specific, quantitative results by incorporating the O-Pair method as a module within FragPipe. In addition to improved search speed and sensitivity, we add flexible options for oxonium ion-based filtering of glycans and support for a variety of MS acquisition methods and provide a comparison between all software tools currently capable of O-glycosite localization in proteome-wide searches.

Comprehensive proteogenomic characterization of rare kidney tumors.

Non-clear cell renal cell carcinomas (non-ccRCCs) encompass diverse malignant and benign tumors. Refinement of differential diagnosis biomarkers, markers for early prognosis of aggressive disease, and therapeutic targets to complement immunotherapy are current clinical needs. Multi-omics analyses of 48 non-ccRCCs compared with 103 ccRCCs reveal proteogenomic, phosphorylation, glycosylation, and metabolic aberrations in RCC subtypes. RCCs with high genome instability display overexpression of IGF2BP3 and PYCR1. Integration of single-cell and bulk transcriptome data predicts diverse cell-of-origin and clarifies RCC subtype-specific proteogenomic signatures. Expression of biomarkers MAPRE3, ADGRF5, and GPNMB differentiates renal oncocytoma from chromophobe RCC, and PIGR and SOSTDC1 distinguish papillary RCC from MTSCC. This study expands our knowledge of proteogenomic signatures, biomarkers, and potential therapeutic targets in non-ccRCC.

Efficient Analysis of Proteome-Wide FPOP Data by FragPipe.

Monitoring protein structure before and after environmental alterations (e.g., different cell states) can give insights into the role and function of proteins. Fast photochemical oxidation of proteins (FPOP) coupled with mass spectrometry (MS) allows for monitoring of structural rearrangements by exposing proteins to OH radicals that oxidize solvent-accessible residues, indicating protein regions undergoing movement. Some of the benefits of FPOP include high throughput and a lack of scrambling due to label irreversibility. However, the challenges of processing FPOP data have thus far limited its proteome-scale uses. Here, we present a computational workflow for fast and sensitive analysis of FPOP data sets. Our workflow, implemented as part of the FragPipe computational platform, combines the speed of the MSFragger search with a unique hybrid search method to restrict the large search space of FPOP modifications. Together, these features enable more than 10-fold faster FPOP searches that identify 150% more modified peptide spectra than previous methods. We hope this new workflow will increase the accessibility of FPOP to enable more protein structure and function relationships to be explored.

Solid-Phase Compatible Silane-Based Cleavable Linker Enables Custom Isobaric Quantitative Chemoproteomics.

Mass spectrometry-based chemoproteomics has emerged as an enabling technology for functional biology and drug discovery. To address limitations of established chemoproteomics workflows, including cumbersome reagent synthesis and low throughput sample preparation, here, we established the silane-based cleavable isotopically labeled proteomics (sCIP) method. The sCIP method is enabled by a high yielding and scalable route to dialkoxydiphenylsilane fluorenylmethyloxycarbonyl (DADPS-Fmoc)-protected amino acid building blocks, which enable the facile synthesis of customizable, isotopically labeled, and chemically cleavable biotin capture reagents. sCIP is compatible with both MS1- and MS2-based quantitation, and the sCIP-MS2 method is distinguished by its click-assembled isobaric tags in which the reporter group is encoded in the sCIP capture reagent and balancer in the pan cysteine-reactive probe. The sCIP-MS2 workflow streamlines sample preparation with early stage isobaric labeling and sample pooling, allowing for high coverage and increased sample throughput via customized low cost six-plex sample multiplexing. When paired with a custom FragPipe data analysis workflow and applied to cysteine-reactive fragment screens, sCIP proteomics revealed established and unprecedented cysteine-ligand pairs, including the discovery that mitochondrial uncoupling agent FCCP acts as a covalent-reversible cysteine-reactive electrophile.

Detecting diagnostic features in MS/MS spectra of post-translationally modified peptides.

Post-translational modifications are an area of great interest in mass spectrometry-based proteomics, with a surge in methods to detect them in recent years. However, post-translational modifications can introduce complexity into proteomics searches by fragmenting in unexpected ways, ultimately hindering the detection of modified peptides. To address these deficiencies, we present a fully automated method to find diagnostic spectral features for any modification. The features can be incorporated into proteomics search engines to improve modified peptide recovery and localization. We show the utility of this approach by interrogating fragmentation patterns for a cysteine-reactive chemoproteomic probe, RNA-crosslinked peptides, sialic acid-containing glycopeptides, and ADP-ribosylated peptides. We also analyze the interactions between a diagnostic ion's intensity and its statistical properties. This method has been incorporated into the open-search annotation tool PTM-Shepherd and the FragPipe computational platform.

Efficient Analysis of Proteome-wide FPOP Data by FragPipe.

Monitoring protein structure before and after perturbations can give insights into the role and function of proteins. Fast photochemical oxidation of proteins (FPOP) coupled with mass spectrometry (MS) allows monitoring of structural rearrangements by exposing proteins to OH radicals that oxidize solvent accessible residues, indicating protein regions undergoing movement. Some of the benefits of FPOP include high throughput and lack of scrambling due to label irreversibility. However, the challenges of processing FPOP data have thus far limited its proteome-scale uses. Here, we present a computational workflow for fast and sensitive analysis of FPOP datasets. Our workflow combines the speed of MSFragger search with a unique hybrid search method to restrict the large search space of FPOP modifications. Together, these features enable more than 10-fold faster FPOP searches that identify 50% more modified peptide spectra than previous methods. We hope this new workflow will increase the accessibility of FPOP to enable more protein structure and function relationships to be explored.

Unraveling the glycosylated immunopeptidome with HLA-Glyco.

Recent interest in targeted therapies has been sparked by the study of MHC-associated peptides (MAPs) that undergo post-translational modifications (PTMs), particularly glycosylation. In this study, we introduce a fast computational workflow that merges the MSFragger-Glyco search algorithm with a false discovery rate control for glycopeptide analysis from mass spectrometry-based immunopeptidome data. By analyzing eight large-scale publicly available studies, we find that glycosylated MAPs are predominantly presented by MHC class II. Here, we present HLA-Glyco, a comprehensive resource containing over 3,400 human leukocyte antigen (HLA) class II N-glycopeptides from 1,049 distinct protein glycosylation sites. This resource provides valuable insights, including high levels of truncated glycans, conserved HLA-binding cores, and differences in glycosylation positional specificity between HLA allele groups. We integrate the workflow within the FragPipe computational platform and provide HLA-Glyco as a free web resource. Overall, our work provides a valuable tool and resource to aid the nascent field of glyco-immunopeptidomics.

MSFragger-Labile: A Flexible Method to Improve Labile PTM Analysis in Proteomics.

Posttranslational modifications of proteins play essential roles in defining and regulating the functions of the proteins they decorate, making identification of these modifications critical to understanding biology and disease. Methods for enriching and analyzing a wide variety of biological and chemical modifications of proteins have been developed using mass spectrometry-based proteomics, largely relying on traditional database search methods to identify the resulting mass spectra of modified peptides. These database search methods treat modifications as static attachments of a mass to particular position in the peptide sequence, but many modifications undergo fragmentation in tandem mass spectrometry experiments alongside, or instead of, the peptide backbone. While this fragmentation can confound traditional search methods, it also offers unique opportunities for improved searches that incorporate modification-specific fragment ions. Here, we present a new labile mode in the MSFragger search engine that provides the flexibility to tailor modification-centric searches to the fragmentation observed. We show that labile mode can dramatically improve spectrum identification rates of phosphopeptides, RNA-crosslinked peptides, and ADP-ribosylated peptides. Each of these modifications presents distinct fragmentation characteristics, showcasing the flexibility of MSFragger labile mode to improve search for a wide variety of biological and chemical modifications.

Recent advances in computational algorithms and software for large-scale glycoproteomics.

Glycoproteomics, or characterizing glycosylation events at a proteome scale, has seen rapid advances in methods for analyzing glycopeptides by tandem mass spectrometry in recent years. These advances have enabled acquisition of far more comprehensive and large-scale datasets, precipitating an urgent need for improved informatics methods to analyze the resulting data. A new generation of glycoproteomics search methods has recently emerged, using glycan fragmentation to split the identification of a glycopeptide into peptide and glycan components and solve each component separately. In this review, we discuss these new methods and their implications for large-scale glycoproteomics, as well as several outstanding challenges in glycoproteomics data analysis, including validation of glycan assignments and quantitation. Finally, we provide an outlook on the future of glycoproteomics from an informatics perspective, noting the key challenges to achieving widespread and reproducible glycopeptide annotation and quantitation.

Ion Mobility-Mass Spectrometry Coupled to Droplet Microfluidics for Rapid Protein Structure Analysis and Drug Discovery.

Native mass spectrometry coupled to ion mobility (IM-MS) has become an important tool for the investigation of protein structure and dynamics upon ligand binding. Additionally, collisional activation or collision induced unfolding (CIU) can further probe conformational changes induced by ligand binding; however, larger scale screens have not been implemented due to limitations associated with throughput and sample introduction. In this work we explore the high-throughput capabilities of CIU fingerprinting. Fingerprint collection times were reduced 10-fold over traditional data collections through the use of improved smoothing and interpolation algorithms. Fast-CIU was then coupled to a droplet sample introduction approach using 40 nL droplet sample volumes and 2 s dwell times at each collision voltage. This workflow, which increased throughput by ∼16-fold over conventional nanospray CIU methods, was applied to a 96-compound screen against Sirtuin-5, a protein target of clinical interest. Over 20 novel Sirtuin-5 binders were identified, and it was found that Sirtuin-5 inhibitors will stabilize specific Sirtuin-5 gas-phase conformations. This work demonstrates that droplet-CIU can be implemented as a high-throughput biophysical characterization approach. Future work will focus on improving the throughput of this workflow and on automating data acquisition and analysis.

Enhancing Cysteine Chemoproteomic Coverage through Systematic Assessment of Click Chemistry Product Fragmentation.

Mass spectrometry-based chemoproteomics has enabled functional analysis and small molecule screening at thousands of cysteine residues in parallel. Widely adopted chemoproteomic sample preparation workflows rely on the use of pan cysteine-reactive probes such as iodoacetamide alkyne combined with biotinylation via copper-catalyzed azide-alkyne cycloaddition (CuAAC) or "click chemistry" for cysteine capture. Despite considerable advances in both sample preparation and analytical platforms, current techniques only sample a small fraction of all cysteines encoded in the human proteome. Extending the recently introduced labile mode of the MSFragger search engine, here we report an in-depth analysis of cysteine biotinylation via click chemistry (CBCC) reagent gas-phase fragmentation during MS/MS analysis. We find that CBCC conjugates produce both known and novel diagnostic fragments and peptide remainder ions. Among these species, we identified a candidate signature ion for CBCC peptides, the cyclic oxonium-biotin fragment ion that is generated upon fragmentation of the N(triazole)-C(alkyl) bond. Guided by our empirical comparison of fragmentation patterns of six CBCC reagent combinations, we achieved enhanced coverage of cysteine-labeled peptides. Implementation of labile searches afforded unique PSMs and provides a roadmap for the utility of such searches in enhancing chemoproteomic peptide coverage.

Multiattribute Glycan Identification and FDR Control for Glycoproteomics.

Rapidly improving methods for glycoproteomics have enabled increasingly large-scale analyses of complex glycopeptide samples, but annotating the resulting mass spectrometry data with high confidence remains a major bottleneck. We recently introduced a fast and sensitive glycoproteomics search method in our MSFragger search engine, which reports glycopeptides as a combination of a peptide sequence and the mass of the attached glycan. In samples with complex glycosylation patterns, converting this mass to a specific glycan composition is not straightforward; however, as many glycans have similar or identical masses. Here, we have developed a new method for determining the glycan composition of N-linked glycopeptides fragmented by collisional or hybrid activation that uses multiple sources of information from the spectrum, including observed glycan B-type (oxonium) and Y-type ions and mass and precursor monoisotopic selection errors to discriminate between possible glycan candidates. Combined with false discovery rate estimation for the glycan assignment, we show that this method is capable of specifically and sensitively identifying glycans in complex glycopeptide analyses and effectively controls the rate of false glycan assignments. The new method has been incorporated into the PTM-Shepherd modification analysis tool to work directly with the MSFragger glyco search in the FragPipe graphical user interface, providing a complete computational pipeline for annotation of N-glycopeptide spectra with false discovery rate control of both peptide and glycan components that is both sensitive and robust against false identifications.

Collision-Induced Unfolding Reveals Stability Differences in Infliximab Therapeutics under Native and Heat Stress Conditions.

Ion mobility-mass spectrometry (IM-MS) and collision-induced unfolding (CIU) assays of monoclonal antibody (mAb)-based biotherapeutics have proven sensitive to disulfide bridge structures, glycosylation patterns, and small molecule conjugation levels. Despite promising prior reports detailing the capabilities of IM-MS and CIU to differentiate biosimilars, generic mAb therapeutics, there remain questions surrounding the sensitivity of CIU to mAb structure changes that occur upon stress, the reproducibility of such measurements across IM-MS platforms, and the correlation between CIU and differential scanning calorimetry (DSC) datasets. In this report, we describe a comprehensive IM-MS and CIU dataset acquired for three Infliximabs: Remicade, Inflectra, and Renflexis. We subject each infliximab sample to forced degradation through heat stress and observe broadly similar yet subtly different stability patterns for these three biotherapeutics. We find that CIU is capable of tracking differences in mAb higher-order structure (HOS) imparted during forced heat stress degradation and that DSC is less sensitive to these alterations in comparison. Furthermore, we collected our comprehensive IM-MS and CIU data across two instrument platforms (Waters G2 and Agilent 6560), with both producing similar abilities to differentiate mAbs while also revealing minor differences between the results obtained on the two instruments. Finally, we demonstrate that CIU-based heatmaps and classification allow for rapid assessment of the most differentiating charge states for the analysis of infliximab, and using multiplexed classification, we conservatively estimate a 30-fold improvement in the time required to perform mAb stability and HOS measurements over standard DSC tools.

Fast Deisotoping Algorithm and Its Implementation in the MSFragger Search Engine.

Deisotoping, or the process of removing peaks in a mass spectrum resulting from the incorporation of naturally occurring heavy isotopes, has long been used to reduce complexity and improve the effectiveness of spectral annotation methods in proteomics. We have previously described MSFragger, an ultrafast search engine for proteomics, that did not utilize deisotoping in processing input spectra. Here, we present a new, high-speed parallelized deisotoping algorithm, based on elements of several existing methods, that we have incorporated into the MSFragger search engine. Applying deisotoping with MSFragger reveals substantial improvements to database search speed and performance, particularly for complex methods like open or nonspecific searches. Finally, we evaluate our deisotoping method on data from several instrument types and vendors, revealing a wide range in performance and offering an updated perspective on deisotoping in the modern proteomics environment.

Fast and comprehensive N- and O-glycoproteomics analysis with MSFragger-Glyco.

Recent advances in methods for enrichment and mass spectrometric analysis of intact glycopeptides have produced large-scale glycoproteomics datasets, but interpreting these data remains challenging. We present MSFragger-Glyco, a glycoproteomics mode of the MSFragger search engine, for fast and sensitive identification of N- and O-linked glycopeptides and open glycan searches. Reanalysis of recent N-glycoproteomics data resulted in annotation of 80% more glycopeptide spectrum matches (glycoPSMs) than previously reported. In published O-glycoproteomics data, our method more than doubled the number of glycoPSMs annotated when searching the same glycans as the original search, and yielded 4- to 6-fold increases when expanding searches to include additional glycan compositions and other modifications. Expanded searches also revealed many sulfated and complex glycans that remained hidden to the original search. With greatly improved spectral annotation, coupled with the speed of index-based scoring, MSFragger-Glyco makes it possible to comprehensively interrogate glycoproteomics data and illuminate the many roles of glycosylation.

Fast Quantitative Analysis of timsTOF PASEF Data with MSFragger and IonQuant.

Ion mobility brings an additional dimension of separation to LC-MS, improving identification of peptides and proteins in complex mixtures. A recently introduced timsTOF mass spectrometer (Bruker) couples trapped ion mobility separation to TOF mass analysis. With the parallel accumulation serial fragmentation (PASEF) method, the timsTOF platform achieves promising results, yet analysis of the data generated on this platform represents a major bottleneck. Currently, MaxQuant and PEAKS are most used to analyze these data. However, because of the high complexity of timsTOF PASEF data, both require substantial time to perform even standard tryptic searches. Advanced searches (e.g. with many variable modifications, semi- or non-enzymatic searches, or open searches for post-translational modification discovery) are practically impossible. We have extended our fast peptide identification tool MSFragger to support timsTOF PASEF data, and developed a label-free quantification tool, IonQuant, for fast and accurate 4-D feature extraction and quantification. Using a HeLa data set published by Meier et al. (2018), we demonstrate that MSFragger identifies significantly (∼30%) more unique peptides than MaxQuant (1.6.10.43), and performs comparably or better than PEAKS X+ (∼10% more peptides). IonQuant outperforms both in terms of number of quantified proteins while maintaining good quantification precision and accuracy. Runtime tests show that MSFragger and IonQuant can fully process a typical two-hour PASEF run in under 70 min on a typical desktop (6 CPU cores, 32 GB RAM), significantly faster than other tools. Finally, through semi-enzymatic searching, we significantly increase the number of identified peptides. Within these semi-tryptic identifications, we report evidence of gas-phase fragmentation before MS/MS analysis.

Pervasive Charge Solvation Permeates Native-like Protein Ions and Dramatically Influences Top-down Sequencing Data.

Post-translational modifications create a diverse mixture of proteoforms, leading to substantial challenges in linking proteomic information to disease. Top-down sequencing of intact proteins and multiprotein complexes offers significant advantages in proteoform analysis, but achieving complete fragmentation for such precursor ions remains challenging. Intact proteins that undergo slow-heating generally fragment via charge directed (i.e., mobile proton) or charge remote fragmentation pathways. Our efforts seek to alter this paradigm by labeling proteins with trimethyl pyrylium (TMP), which forms a stable, positively charged label at lysine residues. Fixing positive charges to the protein sequence reduces the availability of mobile protons, driving fragmentation to charge remote channels. Furthermore, we demonstrate that capping acidic side chains with carbodiimide chemistry obstructs this pathway, restoring charge-directed fragmentation and resulting in more even coverage of the protein sequence. With large amounts of fixed charge and few mobile protons, we demonstrate that it is also possible to direct fragmentation almost exclusively to lysine residues containing the charged label. Finally, our data indicate that when electrosprayed under native conditions, protein ions possess an immense capacity to accommodate excess positive charge. Molecular dynamics simulations of such ions bearing intrinsically charged labels reveal evidence of numerous charge solvation processes, including the preferential formation of helices that solvate charged labels through interactions with their macro-dipoles. Thus, these studies reveal the extent to which intact gas-phase protein ions are capable of solvating charge, and provide the most complete indication to date of the extensive physical forces opposing the goal of comprehensive top-down protein sequencing.

Collision Induced Unfolding Classifies Ligands Bound to the Integral Membrane Translocator Protein.

Membrane proteins represent most current therapeutic targets, yet remain understudied due to their insolubility in aqueous solvents and generally low yields during purification and expression. Ion mobility-mass spectrometry and collision induced unfolding experiments have recently garnered attention as methods capable of directly detecting and quantifying ligand binding within a wide range of membrane protein systems. Despite prior success, ionized surfactant often creates chemical noise patterns resulting in significant challenges surrounding the study of small membrane protein-ligand complexes. Here, we present a new data analysis workflow that overcomes such chemical noise and then utilize this approach to quantify and classify ligand binding associated with the 36 kDa dimer of translocator protein (TSPO). Following our denoising protocol, we detect separate gas-phase unfolding signatures for lipid and protoporphyrin TSPO binders, molecular classes that likely interact with separate regions of the protein surface. Further, a detailed classification analysis reveals that lipid alkyl chain saturation levels can be detected within our gas-phase protein unfolding data. We combine these data and classification schemes with mass spectra acquired directly from liquid-liquid extracts to propose an identity for a previously unknown endogenous TSPO ligand.

An Algorithm for Building Multi-State Classifiers Based on Collision-Induced Unfolding Data.

Collision-induced unfolding (CIU) has emerged as a valuable method for distinguishing iso-cross-sectional protein ions through their distinct gas-phase unfolding trajectories. CIU shows promise as a high-throughput, structure-sensitive screening technique with potential applications in drug discovery and biotherapeutic characterization. We recently developed a CIU classification workflow to support screening applications that utilized CIU data acquired from a single protein charge state to distinguish immunoglobulin (IgG) subtypes and membrane protein lipid binding. However, distinguishing highly similar protein structures, such as those associated with biotherapeutics, can be challenging. Here, we present an expansion of this classification method that includes CIU data from multiple charge states, or indeed any perturbation to protein structure that differentially affects CIU, into a combined classifier. Using this improved method, we are able to improve the accuracy of existing, single-state classifiers for IgG subtypes and develop an activation-state-sensitive classifier for selected Src kinase inhibitors when data from a single charge state was insufficient to do so. Finally, we employ the combination of multiple charge states and stress conditions to distinguish a highly similar innovator/biosimilar biotherapeutic pair, demonstrating the potential of CIU as a rapid screening tool for drug discovery and biotherapeutic analysis.

A Modified Drift Tube Ion Mobility-Mass Spectrometer for Charge-Multiplexed Collision-Induced Unfolding.

Collision-induced unfolding (CIU) of protein ions and their noncovalent complexes offers relatively rapid access to a rich portfolio of biophysical information, without the need to tag or purify proteins prior to analysis. Such assays have been characterized extensively for a range of therapeutic proteins, proving exquisitely sensitive to alterations in protein sequence, structure, and post-translational modification state. Despite advantages over traditional probes of protein stability, improving the throughput and information content of gas-phase protein unfolding assays remains a challenge for current instrument platforms. In this report, we describe modifications to an Agilent 6560 drift tube ion mobility-mass spectrometer in order to perform robust, simultaneous CIU across all precursor ions detected. This approach dramatically increases the speed associated with typical CIU assays, which typically involve mass selection of narrow m/ z regions prior to collisional activation, and thus their development requires a comprehensive assessment of charge-stripping reactions that can unintentionally pollute CIU data with chemical noise when more than one precursor ion is allowed to undergo simultaneous activation. By studying the unfolding and dissociation of intact antibody ions, a key analyte class associated with biotherapeutics, we reveal a predictive relationship between the precursor charge state, the amount of buffer components bound to the ions of interest, and the amount of charge stripping detected. We then utilize our knowledge of antibody charge stripping to rapidly capture CIU data for a range of antibody subclasses and subtypes across all charge states simultaneously, demonstrating a strong charge state dependence on the information content of CIU. Finally, we demonstrate that CIU data collection times can be further reduced by scanning fewer voltage steps, enabling us to optimize the throughput of our improved CIU methods and confidently differentiate antibody variant ions using ∼20% of the data typically collected during CIU. Taken together, our results characterize a new instrument platform for biotherapeutic stability measurements with dramatically improved throughput and information content.