Detecting protein higher-order structural changes using kinase as a phospho-labeler Authors: Asato Maeda, Kosuke Ogata, Yasushi Ishihama Published: 2026 ABSTRACT SummaryWe developed an approach to detect protein structural changes on a proteome-wide scale by phosphate labeling through in vitro kinase reactions. This structural proteomics approach combines residue-specific labeling based on kinase substrate recognition with quantitative phosphoproteomics using phosphopeptide enrichment, enabling site-resolved profiling of structural alterations. Using myoglobin with and without heat denaturation and HEK293T cell extracts with and without protease treatment, we demonstrated that phosphorylation efficiency reflects differences in substrate protein structure. Moreover, by comparing phosphorylation efficiencies before and after RNA digestion, we successfully identified proteome-wide structural changes in non-denatured cell extracts. This approach enables a residue-specific readout of structural dynamics within the intracellular proteome.Keywords: structural proteomics, protein structure, mass spectrometry, in vitro kinase reaction, phosphorylation, phosphopeptide enrichment, phosphoproteomics Introduction Protein structural dynamics are closely linked to function.1,2 Within cells, proteins continuously undergo structural changes through interactions with other biomolecules, post-translational modifications, and degradation processes. These dynamics often regulate diverse cellular processes by transmitting signals to downstream proteins.3,4 However, such intermolecular communications are absent in purified protein systems commonly used to analyze protein structures using conventional biophysical methods. To fully elucidate the molecular mechanisms underlying physiological and pathological phenotypes, a comprehensive analysis of the higher-order structural dynamics of intracellular proteins—that is, structural proteomics—is essential.5 Mass spectrometry (MS)-based structural proteomics techniques are increasingly widespread6,7; these methods can be broadly classified as probing free and probing based. Probing-free approaches, such as thermal proteome profiling (TPP), detect changes in protein physicochemical properties.8,9,10 While they can comprehensively identify structurally altered proteins, they cannot elucidate structural changes in individual domains within the same molecule. In contrast, probing-based approaches, including hydrogen-deuterium exchange mass spectrometry (HDX-MS), hydroxyl radical footprinting (HRF), or limited proteolysis (LiP)-mass spectrometry (MS), modify exposed amino acid residues and identify the probed sites, enabling site-specific detection of structural changes and thus providing high-resolution structural information.11,12,13,14,15,16 However, the heterogeneity of probed peptides generated by amino-acid-nonspecific probing or nonspecific cleavage complicates MS detection.17,18 While a one-step digestion approach achieved high sensitivity for detecting structural changes,19 its structural resolution was limited to the peptide level. Alternative strategies use amino acid-specific probing reagents to enable proteome-wide monitoring of structural changes with residue-level resolution,20,21,22 but comprehensive probing of specific residues in complex biological samples remains difficult. Enrichment-tagged probing reagents have also been widely used to improve detection of low-abundance probed peptides.23,24 Unfortunately, biotin, the most widely used enrichment handle, binds strongly to streptavidin, making elution difficult,25,26,27 and the complex, poorly understood fragmentation patterns of biotinylated peptides hinder site-specific identification.28,29 Overall, residue-specific probing and the sensitive detection of probed peptides continue to be major bottlenecks in MS-based structural proteomics. Phosphorylation is a post-translational modification that is regulated by protein kinases. Kinases are usually categorized into serine/threonine or tyrosine kinases, which phosphorylate these residues with high specificity through substrate recognition mechanisms within milliseconds.30,31 Currently, phosphopeptides can be readily and efficiently enriched using well-established metal affinity chromatographic methods.32 Moreover, active kinases are commercially available and have been widely utilized in in vitro kinase assays with MS to investigate kinase-substrate interactions. These assays have been employed to profile kinase sequence specificity,33,34 identify direct substrates and consensus phosphorylation motifs,35,36 estimate phosphorylation stoichiometry,37 and enhance detection of endogenous phosphorylation sites.38,39 Here, we developed an approach called “phospho-probing,” which utilizes kinase as a phospho-labeler to profile protein structural changes on a proteome-wide scale. In this study, we used the tyrosine kinase SRC, which has been shown to phosphorylate a broad range of tyrosine residues in vitro,35,38 to perform phospho-probing of substrate proteins in distinct structural states. Our results demonstrate that in vitro phospho-probing can detect structural changes in heat-denatured myoglobin and can differentiate protein structural states in cell extracts. Furthermore, we applied phospho-probing to RNA-digested cell extracts and identified protein higher-order structural changes upon RNA dissociation. Results Phospho-probing workflow We present a strategy, termed “phospho-probing,” to analyze protein higher- order structures using an in vitro kinase assay. In the first step, proteins and protein complexes extracted under non-denaturing conditions undergo phospho-probing (Figure 1). Phosphorylation is carried out through a brief in vitro kinase reaction using a recombinant kinase, which adds phosphate groups specifically to phospho-acceptor residues located within sequences recognized by the kinase. The proteins are then digested, and the resulting phospho-probed peptides are selectively enriched from the peptide mixture via phosphopeptide enrichment. The enriched samples are analyzed by nanoscale liquid chromatography-tandem mass spectrometry (nanoLC-MS/MS) to identify and quantify the phosphorylated peptides. A control sample incubated without recombinant kinase is used to account for any remaining endogenous kinase activity or endogenous phosphorylation. Comparison with this control allows normalization of phosphorylation introduced by endogenous kinases and enables definition of the phosphorylation level specifically attributed to the recombinant kinase (phospho-probing efficiency). Phospho-probing workflow In the first step, the native (non-denatured) protein is phosphorylated through a brief in vitro kinase reaction using a recombinant kinase. Following trypsin digestion, the resulting phosphopeptides are enriched. The samples are then quantitatively analyzed by LC-MS/MS to detect residue-specific sites of protein structural change, based on the signal intensity ratios of the identified phosphorylation sites. In this study, tyrosine residues were selected as targets for phospho-probing, and recombinant SRC kinase, which exhibits high catalytic activity among tyrosine kinases,35 was used to perform phospho-probing reactions on substrates. Tyrosine residues are distributed across both hydrophilic and hydrophobic regions of proteins.40 Therefore, their labeling efficiency is expected to vary significantly depending on the protein’s structural state at each site. In addition, tyrosine kinases generally exhibit lower substrate sequence specificity than serine/threonine kinases,41 and many tyrosine residues can be phosphorylated in vitro.35,38 Moreover, the overall level of endogenous tyrosine phosphorylation in mammalian cells is extremely low compared to serine/threonine phosphorylation,42 which minimizes the influence of endogenous phosphorylation on the analysis of phospho-probed sites. Phospho-probing detects higher-order structural changes in heat-denatured myoglobin We first investigated the feasibility of detecting protein structural changes by phospho-probing using denatured purified proteins. We selected horse apomyoglobin, a commonly used model for studying protein structural changes upon thermal denaturation.43,44 Since myoglobin contains two tyrosine residues, we first evaluated whether SRC kinase phosphorylates them, and we found that Y147 was phosphorylated by SRC, whereas Y104 was not (Figure S1A). We then examined whether the phospho-probing efficiency of Y147 reflects the higher-order structure of the protein. Myoglobin was heat treated at temperatures ranging from 37°C to 72°C. Circular dichroism (CD) spectra showed a decrease in α-helical content, confirming that myoglobin was denatured (Figure S1B). The melting temperature of apomyoglobin was consistent with values found in previous studies.45 Importantly, the phospho-probing efficiency of Y147 in heat-treated myoglobin increased as the α-helical content of myoglobin decreased (Figures 2B and S1C). These results indicate that phospho-probing can detect changes in the higher-order structure of the protein. Phospho-probing detects structural changes in heat-denatured apomyoglobin (A) Workflow of an experiment demonstrating the phospho-probing strategy with apomyoglobin as a model protein (see STAR Methods). Secondary structure analysis by circular dichroism (CD) spectroscopy and in vitro kinase reaction were performed at 37°C. The red dotted arrow indicates residue Y147. The structural image was generated based on PDB: 1AZI. (B) Comparison of the secondary structure profile of heat-denatured apomyoglobin and its phospho-probing efficiency. The α-helical content of apomyoglobin was assessed by monitoring ellipticity at 222 nm. Phospho-probing efficiency was calculated as the ratio of the peak area of phosphorylated Y147-containing peptides in each heat-treated sample to that of the control sample (heat treated at 37°C). The values represent the average of two or three technical replicates. Error bars indicate the standard deviation. Phospho-probing detects the disruption of proteome high-order structure by protease digestion Next, proteins extracted from HEK293T cells under non-denaturing conditions were used as a substrate pool for in vitro kinase reactions with SRC. To evaluate phospho-probing of substrates with disrupted higher-order structures, in vitro kinase reactions were also performed following trypsin digestion of the extracted proteins (Figure 3A). We identified 440 tyrosine phosphorylation sites in the non-denatured samples and 678 sites in the trypsin-digested samples, in at least two out of three technical replicates (Figures 3B and S2A; Table S2). Only 247 sites (56%) were common to both conditions, whereas 193 and 431 sites were uniquely identified in the non-denatured and trypsin-digested samples, respectively. These results suggest that phospho-probing profiles differ substantially depending on the higher-order structure of the substrate proteins. Phospho-probing of cell lysates in the non-denatured state (A) Experimental design for phospho-probing of non-denatured cell lysates (see STAR Methods). Phospho-probing reactions were performed in triplicate. (B) Venn diagram showing the overlap of tyrosine phosphorylation sites (pY sites) identified in non-denatured and trypsin-digested samples. (C) Proportion of tyrosine phospho-probed sites located in intrinsically disordered regions (IDRs), as predicted by StructureMap,46 among all identified tyrosine phospho-probed sites. Tyrosine residues from proteins not available in the AlphaFold database were excluded from the analysis. (D) Boxplot showing the comparison of phospho-probing efficiency of non-denatured lysate according to predicted disorder, solvent accessibility (pPSE at radius = 12 Å and angle = 70°), and predicted secondary structure. The box itself spans the interquartile range. The whiskers represent 5% and 95% quantiles. The thick horizontal line in each box indicates the median. Asterisks indicate results of Mann-Whitney U test (ns: p > 0.05, ∗p < 0.05, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001). Tyrosine residues of proteins for which no predicted structure was available in the AlphaFold database were excluded from the analysis. We next investigated the relationship between phospho-probing sites and protein higher-order structure on a proteome-wide scale and mapped the identified phospho-probing sites to predicted protein structures obtained from AlphaFold2.47,48 This enabled the extraction of comprehensive structural information for the labeled residues. Note that a lack of phospho-probing does not necessarily equate to structural hindrance, since kinases have sequence specificity, but we minimized this effect by comparing two structural states of identical sequence. Specifically, the StructureMap Python package was used to retrieve residue-level structural predictions from the AlphaFold Protein Structure Database, and intrinsically disordered regions (IDRs) were predicted for each residue.46 Among the 440 tyrosine phosphorylation sites identified in the non-denatured samples, 50% were in IDRs. In contrast, among the 431 sites uniquely identified in the trypsin-digested samples, more than 95% of the tyrosine residues were located in structured regions (Figures 3C and S2B), indicating that phospho-probing profiles reflect the increased accessibility of tyrosine residues upon disruption of protein higher-order structures. Interestingly, we also observed 193 tyrosine residues that were phospho-probed in the non-denatured sample but not in the trypsin-digested sample (Figure 3B). We observed that the distribution of lysine and arginine residues around tyrosines that were no longer phosphorylated in the tryptic digest differed significantly from those that were phosphorylated in the tryptic digest (p = 0.024, Fligner-Killeen test) (Figure S2C), suggesting that some phospho-probed sites in the non-denatured samples were affected by trypsin digestion. In addition, of the 193 tyrosine residues, 96 were detected as phospho-probed sites following prolonged incubation of the trypsin-digested sample (Figure S2D), suggesting that these sites were phosphorylated at a slower rate due to disruption of the substrate’s higher-order structure. Moreover, when analyzing phospho-probed sites where the primary sequence surrounding the tyrosine residues was preserved, their phospho-probing profiles also reflected the substrate’s higher-order structure (Figure S2E). We also examined the overlap between phospho-probed sites and those reported in vivo in PhosphoSitePlus49 and found that phospho-probing of substrate proteins in the non-denatured state selectively targets phosphorylation sites that are also modified in vivo (p < 10−5, Fisher’s exact test) (Figure S2F). In addition, tyrosine residues phosphorylated in the non-denatured state tended to be in regions with low predicted local distance difference test (pLDDT) scores in predicted structures (Figures S2G and S2H). This finding implies that phospho-probing of non-denatured substrates may provide access to regions for which current structure prediction models are less reliable. We further quantified the phospho-probing efficiency of individual tyrosine residues in the substrate proteins and analyzed their correlation with predicted structural features (see STAR Methods). Phospho-probing efficiencies correlated with predicted disorder, solvent accessibility, and secondary structure of the corresponding regions (Figure 3D). These results indicate that quantitative phospho-probing reflects the higher-order structural state of the substrate proteins. Effects of the higher-order structure of substrate proteins on the optimum phospho-probing reaction time were also evaluated (Figure S3A). We observed that the difference in phospho-probing efficiency between tyrosine sites located in disordered regions and those in structured regions decreased as incubation time increased (Figure S3B). This result suggests that brief kinase incubation times enhance the sensitivity of phospho-probing to differences in protein higher-order structure. Based on these findings, we adopted a 15-min incubation for all in vitro kinase reactions to maximize the responsiveness of phospho-probing profiles to substrate structural states. Overall, these results demonstrate that phospho-probing enables the profiling of protein structural differences in complex biological samples. Phospho-probing enables the residue-specific profiling of protein structural changes induced by RNA digestion in complex biological samples Next, we further examined the applicability of phospho-probing to detect structural changes in protein subpopulations. Since SRC substrates are enriched in RNA-binding proteins (Figure S4A), we used phospho-probing to profile protein higher-order structural changes induced by the loss of protein-RNA interactions. The phospho-probing reaction with SRC was performed using non-denatured HEK293T cell lysates (RNA-preserved) or lysates digested with an RNase cocktail (RNA-digested) (Figures 4A and S4B). To improve the detectability of phosphotyrosine-containing peptides, we included an isobarically labeled carrier sample enriched in tyrosine-phosphorylated peptides generated by a prolonged in vitro kinase reaction.38 As a result, we successfully identified 3,126 tyrosine phosphorylation sites across 1,881 proteins (Table S3). Among these, 307 phospho-probed sites were identified in RNA-preserved samples and 689 in RNA-digested samples (Figure S4; see STAR Methods). Comparison of phospho-probing efficiency between the two conditions showed that 301 sites on 246 proteins exhibited a significantly increased probing efficiency following RNA digestion (Figures 4B, S4D, and S4E). Notably, more than 40% of the phospho-probed tyrosine residues showed increased probing upon RNA digestion, suggesting that tyrosine residues involved in protein-RNA interactions are widely distributed across the proteome. Application of phospho-probing to detect protein structural changes induced by RNA digestion and comparison of RNase-sensitive proteins detected by phospho-probing and solubility assay (A) Experimental workflow for phospho-probing using RNA-preserved [RNase(−)] and RNA-digested [RNase(+)] HEK293T cell lysates (see STAR Methods). In vitro kinase reactions were performed in triplicate. (B) Volcano plot showing differences in tyrosine phospho-probing efficiency between RNA-preserved and RNA-digested samples. 711 tyrosine residues that were phospho-probed in at least one of the two conditions were analyzed (see STAR Methods). The x axis represents the log2-transformed fold change in phospho-probing efficiency following RNase treatment. The y axis indicates the p value from a two-tailed Welch’s t test. 301 sites that showed a significant increase (fold change ≥2; p < 0.05) in RNA-digested sample are highlighted in red. (C) Gene Ontology (GO) molecular function enrichment analysis of proteins containing 246 RNase-sensitive tyrosine residues. A total of 587 proteins with phospho-probed tyrosines were used as the background set. (D) Mapping of RNase-sensitive tyrosine residues onto the RPL5 and 5S rRNA complex structure (PDB: 4UG0). Tyrosine residues with increased phospho-probing efficiency in RNA-digested samples (Y183, Y207, Y240, and Y253) are indicated in red. The minimum Euclidean distance (in angstroms) between each tyrosine and RNA is indicated. (E) Visualization of RNase-sensitive (red) and RNase-insensitive (blue) tyrosine residues on representative RNA-binding proteins (HNRNPA0, SYNCRIP, and HNRNPK). Schematic of protein domains are based on UniProt annotations. RRM, RNA recognition motif; KH, K homology domain. (F) Overlap of all identified proteins between phospho-probing and solubility assay datasets. (G) Overlap of RNase-sensitive proteins identified by each method among the 480 phospho-probed proteins within the 1,507 proteins commonly analyzed by both approaches. (H) Pie chart showing the proportion of the 166 RNase-sensitive proteins uniquely identified by phospho-probing that are annotated as RNA-binding proteins (RBPs) in the RBP2GO database. We then investigated the relationship between protein-RNA interactions and tyrosine residues with increased phospho-probing efficiency upon RNA digestion. Gene Ontology (GO) enrichment analysis using the Database for Annotation, Visualization, and Integrated Discovery (DAVID)50,51 showed strong enrichment of RNA-binding proteins, indicating that phospho-probing can detect higher-order structural changes in RNA-binding proteins (Figure 4C). We then focused on 60S ribosomal protein L5 (RPL5), in which Y253 showed the largest increase in phospho-probing efficiency following RNA digestion. In total, four tyrosine residues with increased phospho-probing efficiency were identified in RPL5. Mapping these sites onto the crystal structure of the protein-RNA complex showed that all four were located within 8 Å of the RNA, indicating close proximity in the native complex52,53 (Figure 4D). These results demonstrate that phospho-probing can detect sites of structural change in proteins associated with the loss of protein-RNA interactions. We further examined the relationship between domain localization of phospho-probed tyrosine residues and RNase sensitivity. Among these, 125 were located within UniProt-annotated domains, of which 58 (46%) were RNase sensitive. Notably, 30 were located within RNA recognition motifs (RRMs), of which 21 (70%) were RNase sensitive, indicating that most tyrosines within RNA-binding domains are RNase sensitive. For example, in HNRNPA0, SYNCRIP, and HNRNPK, tyrosines located on the RRM or K homology (KH) domain were RNase sensitive, whereas tyrosines located outside these domains were not (Figure 4E). These examples suggest that phospho-probing can detect protein structural changes in a site-specific manner. To gain insight into the in vivo function of RNase-sensitive tyrosine residues, we analyzed a comprehensive phosphorylation site dataset from the literature to examine phosphorylation within RNA-binding domains.54 In this dataset, SNRNP70, a subunit of the small nuclear ribonucleoprotein (snRNP) complex, showed enhanced phospho-probing of Y126 upon RNA digestion (Figure S4F). Structural mapping of this tyrosine residue onto the crystal structure of the snRNP complex revealed that Y126 is located at the interface between SNRNP70 and RPB2, a subunit of RNA polymerase II. These findings highlight the potential of phospho-probing to detect structural alterations at functionally relevant phosphorylation sites. Performance comparison of phospho-probing and solubility assay Sridharan et al. previously reported a structural proteomics method that utilizes solubility differences based on surfactant extraction and found that digesting RNA in cell lysates increases the solubility of proteins involved in RNA-protein interactions.55 However, proteins that are already highly soluble under RNA-preserved conditions, or those exhibiting only modest solubility changes upon RNA digestion, are likely to have their structural changes overlooked by the solubility assay. To evaluate the phospho-probing approach, we compared RNase-sensitive proteins identified by phospho-probing with those identified by the solubility assay. A total of 1,507 proteins were commonly present in both datasets (Figure 4F), including 480 proteins that were phospho-probed by SRC. Among these proteins, 166 were identified as RNase sensitive only by phospho-probing, 31 were uniquely identified by the solubility assay, and 43 were identified by both methods (Figure 4G). This result suggests that phospho-probing is highly sensitive for detecting protein structural changes. Most of the 166 RNase-sensitive proteins uniquely identified by phospho-probing are annotated in RBP2GO,56 a database of RNA-binding proteins, supporting that they are RNA-interacting proteins (Figures 4H, S4G, and S4H). This indicates that phospho-probing can capture RNA-dependent structural changes that are not detectable by solubility-based methods. This difference likely reflects the distinct resolution of the two approaches; whereas the solubility assay reports changes in the biophysical properties of the entire protein, phospho-probing enables site-specific detection of local structural alterations. Interestingly, five of the 166 RNase-sensitive proteins (AP1AR, FNBP1L, MIER3, NUCB2, and PNPO) were not annotated in RBP2GO. For instance, MIER3 (mesoderm induction early response 1, family member 3) is a transcriptional repressor that forms a complex with HDAC1/HDAC2 and Snail and binds to the E-cadherin promoter to repress its expression.57 Given this regulatory role, it would be of interest to assess the RNA-binding activity of MIER3. Our findings provide a valuable resource for exploring the function and mechanisms of unknown RNA-binding proteins. In contrast, 31 proteins were identified as RNase sensitive exclusively by the solubility assay. In these cases, phospho-probed sites appeared to differ from those involved in RNA binding, as fewer were located within RRMs (Figure S4I). This suggests that phospho-probing may fail to detect structural changes if labeling occurs in protein regions unrelated to those changes. For example, although SRSF1 forms a complex with RNA via its RRM domain, the phospho-probed tyrosine residue Y170 was located far from the RRM-RNA interaction interface (Figure S4J). This implies that while Y170 resides within a domain associated with RNA binding, the local structure surrounding this residue did not change significantly upon RNA dissociation. Therefore, phospho-probing could overlook such changes when labeling does not occur near the site of structural alteration. In summary, phospho-probing with kinases enabled the proteome-wide and site-specific detection of structural changes in substrate proteins that interact directly or indirectly with RNA. Discussion Our results demonstrate the potential of in vitro phospho-probing as a strategy for detecting protein higher-order structural changes. However, to observe changes in in vitro phospho-probing efficiency, two conditions are required: (1) kinase access to the phospho-acceptor and (2) compatibility of the surrounding sequence with the kinase’s sequence specificity. Accordingly, the magnitude of changes in phospho-probing efficiency is influenced by both factors. When the same recombinant kinase is used to compare different structural states of the identical sequence, its intrinsic sequence specificity and enzymatic activity remain constant under the same reaction conditions. Therefore, changes in phospho-probing efficiency can be interpreted as changes in protein higher-order structure. Although this study does not aim to directly compare phosphorylation levels between different sites, if the amino acid sequence preferences of kinases can be corrected quantitatively and with high precision, it may become possible to make relative comparisons of accessibility between sites based on phospho-probing efficiency, thereby potentially improving the resolution of phosphorylation probing as a structural readout. It remains unclear whether phospho-probing of cell lysates, as performed in this study, can capture structural changes occurring within cells; however, given that recent studies have successfully detected structural changes in intracellular proteins by introducing proteases into cells via electroporation,58 the development of in-cell phospho-probing is likely to expand the possibilities of structural proteomics. Theoretically, proteins with known phosphorylation sites are amenable to higher-order structural analysis using phospho-probing. According to PhosphoSitePlus,49 tyrosine phosphorylation sites have been reported in 57% of human proteins (11,743 of 20,601), and 84% (17,346 of 20,601) have at least one phosphorylation site when sites on serine and threonine are included. Any active kinase can be used in in vitro kinase reactions, thus leveraging multiple kinases could expand the analytical coverage of protein higher-order structural changes. Moreover, recent studies have reported kinases with exceptionally broad or low sequence specificity,59 suggesting that even proteins with few or poorly conserved phosphorylation motifs could be amenable to phospho-probing. Additionally, when specific substrates are of interest, phosphorylation site databases can be utilized to identify known phosphorylation motifs within those proteins.33,34 Such prior knowledge will support the rational selection of kinases for phospho-probing and will facilitate targeted analysis of structural changes. We used a substrate pool generated by trypsin digestion as a control to evaluate the effect of protein higher-order structure on phospho-probing. However, we also found that loss of the amino acid sequence surrounding phospho-acceptor sites due to trypsin digestion can affect labeling efficiency. Therefore, selection of the control substrate pool requires careful consideration. Although trypsin digests were chosen for their practicality, alternative proteolytic enzymes may generate peptide pools that enable phospho-probing analysis with reduced sequence bias. Limitations of the study The current phospho-probing method remains limited to detecting pronounced changes in protein higher-order structure. Detecting subtle conformational changes, such as those induced by protein-ligand interactions, post-translational modifications, or minor differences in protein complex states, remains future work. In addition, because the reaction requires an active kinase, it cannot be performed under conditions that denature or inactivate the enzyme. Developing kinases with engineered substrate specificity or enhanced tolerance to denaturing conditions through protein design technologies could expand the applicability of this method to a broader range of structural analysis. We could not distinguish whether changes in phospho-probing efficiency were due to changes in protein higher-order structure or secondary structural changes resulting from the probing process itself. Although we minimized the impact of secondary effects by adopting brief reaction times, we cannot completely remove their influence. Time-course profiling of phospho-probing could help differentiate primary changes from secondary effects. In the current workflow, phosphate groups introduced by phospho-probing cannot be distinguished from endogenous phosphate groups because of their identical mass, limiting accurate assessment of labeling efficiency. Incorporating non-natural ATP analogs, such as 18O-ATP, into the phospho-probing reaction may enable clear differentiation between introduced and endogenous phosphate groups. In this study, only singly phosphorylated peptides were analyzed, whereas multi-phosphorylated peptides were excluded to simplify interpretation. However, capturing higher-order structural changes at multi-phosphorylation sites remains of considerable interest. Future improvements in the analytical workflow would enable inclusion of multi-phosphorylated peptides, allowing for a more comprehensive understanding of structural dynamics at these complex regulatory sites. Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Yasushi Ishihama (yishiham@pharm.kyoto-u.ac.jp). Materials availability This study did not generate new, unique reagents. Data and code availability •The raw MS data and analysis files have been deposited with the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (https://jpostdb.org/) with the dataset identifier PXD063369. The raw MS data and analysis files have been deposited with the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (https://jpostdb.org/) with the dataset identifier PXD063369. •This paper does not report original code. This paper does not report original code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Acknowledgments We would like to thank the members of the Department of Molecular Systems BioAnalysis for the fruitful discussion. We also thank the Kyoto University Innovative Support Alliance for Life Science (iSAL) for the J-820 circular dichroism spectrometer, with support from Prof. Kenichi Kawano, and for the Orbitrap Exploris 480 mass spectrometer. This work was supported by JST SPRING (JPMJSP2110 to A.M.), JST ACT-X (JPMJAX2324 to K.O.), JSPS KAKENHI (23K18185 and 23H04924 to Y.I. and 25K18607 to K.O.), the FY 2024 Researcher Exchange Program between JSPS and ETH to K.O., the Nakatani Foundation to K.O., and the Research Foundation for Pharmaceutical Sciences to K.O. Author contributions Conceptual design of the project was done by K.O. and Y.I. All experiments and data analysis were carried out by A.M. The first draft of the manuscript was written by A.M., and the final manuscripts were prepared by A.M., K.O., and Y.I. Declaration of interests The authors declare no competing interests. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used GPT-4 in order to improve language and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERChemicals, peptides, and recombinant proteinsMyoglobin from equine heartSigma-AldrichCat# M1882DMEM(High Glucose) with L-Glutamine, Phenol Red and Sodium PyruvateFUJIFILM WakoCat# 043-30085Fetal Bovine Serum (FBS)Thermo Fisher ScientificCat# 10270106UltraPure™ Tris Buffer (powder format)Thermo Fisher ScientificCat# 15504-020Phosphatase Inhibitor Cocktail 2SigmaCat# P5726Phosphatase Inhibitor Cocktail 3SigmaCat# P0044100 x protease inhibitor cocktailMerckCat# P8340RNasin® Plus Ribonuclease InhibitorPromegaCat# N2611RNase Cocktail™ Enzyme MixThermo Fisher ScientificCat# AM2286SRCCarna BiosciencesCat# 08-173Adenosine 5′-Triphosphate (ATP) Disodium Salt TrihydrateFUJIFILM WakoCat# 014-16913Magnesium Chloride HexahydrateFUJIFILM WakoCat# 135-00165Sodium deoxycholate (SDC)FUJIFILM WakoCat# 194-08311Sodium N-lauroylsarcosinate (SLS)FUJIFILM WakoCat# 192-10382(+/−)-Dithiothreitol (DTT)FUJIFILM WakoCat# 045-08974Iodoacetamide (IAA)FUJIFILM WakoCat# 095-02151Ammonium BicarbonateFUJIFILM WakoCat# 017-02875Lysyl endopeptidase (LysC)FUJIFILM WakoCat# 129-02541Sequencing Grade Modified TrypsinPromegaCat# V5111Ethyl AcetateFUJIFILM WakoCat# 051-00356Trifluoroacetic Acid (TFA)FUJIFILM WakoCat# 208-02741Acetonitrile (ACN) (Guaranteed Reagent)FUJIFILM WakoCat# 014-00381Empore SDB-XC disksCDS AnalyticalCat# 13-110-020InertSep PLS-2GL SciencesCat# 5010-69100Titanium dioxide (10 μm)GL SciencesCat# 5020-75010Empore C8 disks3MCat# 2214Lactic AcidFUJIFILM WakoCat# 128-00056PiperidineFUJIFILM WakoCat# 166-02773Phosphoric AcidFUJIFILM WakoCat# 160-08636Heptafluorobutyric acid (HFBA)TCICat# Q0054TMTpro 16plex Label Reagent SetThermo Fisher ScientificCat# A44520TMTpro 18-plex, TMTpro-134C, and TMTpro-135N Label ReagentsThermo Fisher ScientificCat# A52046InertSep RP-C18 beadsGL SciencesCat# 5010-27133Disodium Hydrogenphosphate 12-WaterFUJIFILM WakoCat# 196-02835Sodium Dihydrogenphosphate DihydrateFUJIFILM WakoCat# 192-0281525% Ammonia SolutionFUJIFILM WakoCat# 010-03166MethanolFUJIFILM WakoCat# 137-01823Reprosil-Pur 120 C18-AQ 3.0 μmDr. Maisch GmbHCat# r13.aq.Reprosil-Pur 120 C18-AQ 1.9 μmDr. Maisch GmbHCat# r119.aq.Critical commercial assaysPierce BCA Protein Assay KitThermo Fisher ScientificCat# 23227Deposited dataProteomics datasetThis paperjPOST: PXD063369Phosphorylation_site_dataset (ver. 2023-8-22)PhosphoSitePlus49https://www.phosphosite.orgHigh-confidence, large-scale phosphorylation site datasetRamasamy et al.54pr2c00164_si_001.xlsxProtein solubility-assay datasetSridharan et al.5541589_2022_1062_MOESM2_ESM.xlsx; 41589_2022_1062_MOESM17_ESM.xlsxComprehensive dataset of RNA-binding proteins (downloaded in 2024-05, 6100 proteins)RBP2GO56https://rbp2go.dkfz.de/Experimental models: Cell linesHEK293TRIKEN BRCCat# RCB2202Software and algorithmsMascot (ver. 2.7.0)Matrix Sciencehttps://www.matrixscience.com/MaxQuant (ver. 1.6.17.0)Cox and Mann60https://maxquant.org/FAIMS MzXML GeneratorHebert et al.61https://github.com/coongroup/FAIMS-MzXML-GeneratorSkyline (ver. 21.2.0)MacLean et al.62http://skyline.msPython (ver. 3.12.3)Python Software Foundationhttps://www.python.org/StructureMapBludau et al.46https://github.com/MannLabs/structuremap_analysisDAVIDSherman et al.51https://davidbioinformatics.nih.gov/Perseus (ver. 2.0.1.1)Tyanova et al.63https://maxquant.net/perseus/PyMOL (ver. 3.1.5.1)Schrödingerhttps://pymol.org/Scipy (ver. 1.13.1)Virtanen et al.64https://github.com/scipy/scipyStatannotations (ver. 0.6.0)Charlier et al.65https://zenodo.org/records/8396665OtherAmicon Ultra-0.5 Centrifugal Filter, 3 kDa MWCOMilliporeCat# UFC500396VeritiPro Thermal Cycler, 96-wellThermo Fisher ScientificCat# A48141 Experimental model and study participant details Cell culture HEK293T cells (RIKEN BRC cell bank) were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum in 10 cm or 15 cm diameter dishes at 37°C under 5% CO2 in air until 80% confluency was reached. For harvesting, cells were washed twice with ice-cold phosphate-buffered saline (PBS), detached using a cell scraper, and pelleted by centrifugation. Method details Apomyoglobin preparation Apomyoglobin was prepared by removing heme from holomyoglobin (Sigma) as previously described.66 The resulting apomyoglobin solution was exchanged to reaction buffer [40 mM Tris-HCl (pH 7.5), 20 mM MgCl2] using a 3-kDa ultrafiltration device (Amicon Ultra, Millipore). The final protein concentration was adjusted to 1 mg/mL. The concentration of apomyoglobin was estimated based on the dry weight of the lyophilized holomyoglobin. Thermal denaturation of myoglobin 50 μL aliquots of apomyoglobin solution were transferred to a 0.2 mL polymerase chain reaction (PCR) tube and heated in parallel at 37°C, 42°C, 47°C, 52°C, 57°C, 62°C, 67°C, and 72°C for 15 min using the VeritiPro Thermal Cycler (Thermo Fisher Scientific). After heating, the samples were placed on ice until CD measurements or until kinase reaction was initiated. Circular dichroism (CD) spectroscopy Secondary structure analysis by circular dichroism (CD) spectroscopy was performed at 37°C. CD spectra were recorded using a J-820 spectropolarimeter (Jasco). The heated protein solution was diluted 5-fold with reaction buffer, and far-UV CD spectra were acquired at 37°C in the range of 200–250 nm. A quartz cell with a 1 mm path length was used for measurements. Protein extraction from cells and RNA digestion treatment Protein extraction was performed by suspending HEK293T cells in surfactant-free buffer under ice-cold conditions, followed by mechanical shearing. HEK293T cell pellets were resuspended in ice-cold lysis buffer [40 mM Tris-HCl (pH 7.5), protease inhibitor cocktail (Sigma), phosphatase inhibitor cocktail 2 and 3 (Sigma), 1 U/mL RNasin Plus Ribonuclease Inhibitor (Promega)]. The suspension was passed 10 times through a 22G needle and 3 times through a 26G needle, then centrifuged at 20000 × g for 20 min at 4°C using a benchtop centrifuge (Eppendorf). The protein concentration in the resulting supernatant was measured using the Pierce BCA protein assay (Thermo Fisher Scientific) and diluted to 1 mg/mL with lysis buffer. The extracted protein solution was divided into two aliquots. One aliquot was used as the RNA-preserved lysate, and the other one was used to prepare RNA-digested lysate. RNA digestion was performed based on the method described by Sridharan et al.55 For the digestion, 2 μL of RNase cocktail (Thermo Fisher Scientific) was added per 100 μL of protein solution, followed by incubation at 4°C for 30 min. In vitro kinase reaction of substrate proteins The in vitro kinase reaction was performed using recombinant tyrosine kinase SRC (human, Carna Biosciences) at an enzyme-to-substrate (E:S) ratio of 1/1000 (w/w) in the presence of 1 mM ATP and 20 mM MgCl2 at 37°C for 15 min. For evaluation of the SRC kinase activity toward myoglobin, the reaction was extended to 3 h. For evaluation of the optimal phospho-probing reaction time, the reaction was performed for 15, 30, 60 and 180 min. The kinase reaction was quenched by the addition of sodium deoxycholate (final concentration: 12 mM) and sodium N-lauroylsarcosine (final concentration: 12 mM), followed by heat treatment at 95°C for 5 min. The number of technical replicates for each in vitro kinase reaction is listed in Table S1. Protein digestion and desalting Protein digestion and desalting were performed based on the sample preparation method using phase transfer surfactant (PTS).67 Proteins were reduced with 10 mM dithiothreitol (DTT) for 30 min and alkylated with 50 mM iodoacetamide (IAA) for 30 min in the dark. The sample solution was diluted 5-fold with 50 mM ammonium bicarbonate, and Lys-C (Wako) and trypsin (Promega) were added at an E:S ratio of 1:100 (w/w) each, followed by overnight digestion at 37°C. After digestion, an equal volume of ethyl acetate was added and acidified with 0.5% trifluoroacetic acid (TFA). After shaking for 2 min, the sample was centrifuged at 15,800 × g for 2 min. The aqueous and organic phases were completely separated, and the aqueous phase was dried using a SpeedVac (Thermo Fisher Scientific), reconstituted in 0.1% TFA and 5% acetonitrile (ACN), and desalted using SDB-XC StageTips.68,69,70 In vitro kinase reaction of tryptic digests Desalted peptides were dried and then reconstituted in 40 mM Tris-HCl (pH 7.5) at 1 mg/mL. Note that the peptide amount corresponds to the protein amount measured by the BCA assay prior to digestion. The in vitro kinase reaction was performed using recombinant tyrosine kinase SRC at an E:S ratio of 1/1000 (w/w) in the presence of 1 mM ATP and 20 mM MgCl2 at 37°C for 15 min. For characterization of the phospho-probed sites uniquely identified in non-denatured samples, the reaction was performed for 60 min. For evaluation of the optimal phospho-probing reaction time, the reaction was performed for 15, 30, 60 and 180 min. For preparation of the carrier sample used for isobaric labeling signal amplification, the in vitro kinase reaction was performed at an E:S ratio of 1/500 (w/w) for 3 h. The kinase reaction was quenched by acidifying the solution to a final concentration of 0.5% TFA, and the peptides were desalted using SDB-XC StageTips. The number of technical replicates for each in vitro kinase reaction is listed in Table S1. Phosphopeptide enrichment Phosphopeptide enrichment was performed using hydroxy acid-modified metal oxide chromatography (HAMMOC).32 Briefly, TiO2 beads were packed into a StageTip equipped with an Empore C8 disk (3M) and coated with lactic acid. The digested peptides were then loaded onto the StageTip, allowing phosphorylated peptides to coordinate with the TiO2 beads. Non-phosphorylated peptides that interacted weakly with TiO2 were removed by washing. Phosphopeptides were subsequently eluted using 0.5% piperidine, acidified with phosphoric acid, and desalted using an SDB-XC StageTip. The desalted peptides were dried using a SpeedVac and reconstituted in the LC/MS/MS solvent containing 0.5% TFA and 4% ACN. For stable isotope labeling experiments, heptafluorobutyric acid (HFBA) was added instead of phosphoric acid before introducing the sample into the StageTip-based microreactor. The information on the peptide amount in each phosphopeptide enrichment for all experiments is shown in Table S1. TMT labeling Isobaric labeling of phosphopeptides was performed using TMTpro (Thermo Fisher Scientific) by a solid-phase reaction method.71 SDB-XC StageTips packed with 0.5 mg of InertSep RP-C18 beads (GL Sciences) were activated with 0.1% TFA and 80% ACN and equilibrated with 0.1% TFA and 5% ACN. Phosphopeptides enriched from 50 μg of protein were loaded onto the StageTips and washed with 50 mM phosphate buffer (pH 6.5). To avoid peptide elution, TMT reagents were dissolved in ACN and then diluted with phosphate buffer to reduce the ACN concentration to less than 5%. A volume of 10 μL of this TMT solution (10 μg/10 μL) was applied to each tip. The particles were kept moist, and samples were incubated at room temperature for 1 h. The StageTips were then washed with 0.1% TFA and 5% ACN, and peptides were eluted with 0.1% TFA and 80% ACN. The information on the peptide amount in each TMT channel is shown in Table S1. High-pH reverse-phase chromatography-based fractionation High-pH reverse-phase chromatography-based fractionation of TMT-labeled phosphopeptides was performed using StageTips as previously described72 with minor modifications. TMT-labeled phosphopeptides were pooled and dried in a SpeedVac, then reconstituted in 50 μL of 0.1% ammonia. SDB-XC StageTips were activated with methanol, 0.1% ammonia, and 80% ACN, followed by equilibration with 0.1% ammonia. The phosphopeptides were then loaded onto the StageTip, and the flow-through fraction was collected. Peptides were sequentially eluted with 0.1% ammonia containing 5%, 7.5%, 10%, 12.5%, 15%, 20%, and 80% ACN. The flow-through and the first fraction (5%) were pooled, as were the last two fractions (20% and 80%). The recovered peptides were dried using a SpeedVac and reconstituted in solvent consisting of 0.5% TFA and 4% ACN for nanoLC/MS/MS analysis. NanoLC/MS/MS analysis The peptide digests of myoglobin were analyzed using a nanoLC/MS/MS system consisting of a PAL HTC-xt autosampler (CTC-Analytics), an UltiMate 3000 RSLCnano pump (Thermo Fisher Scientific), and a Q Exactive mass spectrometer (Thermo Fisher Scientific). The mass spectrometric analyses were carried out in the data-dependent acquisition (DDA) mode. The LC mobile phases consisted of solvent A (0.5% acetic acid) and solvent B (0.5% acetic acid and 80% ACN). Peptides were separated using an in-house-packed needle column73 (150 mm, 100 μm i.d.) containing Reprosil-Pur 120 C18-AQ 3.0 μm (Dr. Maisch). The flow rate was set to 500 nL/min. The gradient conditions for mobile phase B were as follows: 5–10% B in 0.5 min, 10–40% B in 20 min, 40–99% B in 1 min, followed by 99% B for 3.5 min. The electrospray voltage was set to 2.4 kV in the positive mode. The mass range of the survey scan was from 350 to 1500 m/z with a resolution of 70,000, an automatic gain control (AGC) target of 3 × 106, and a maximum injection time (MIT) of 100 ms. The first mass of the MS/MS scan was set automatically with a resolution of 17,500, an AGC target of 1 × 105, and a MIT of 100 ms. The fragmentation was performed by higher energy collisional dissociation with a normalized collision energy of 27%. The dynamic exclusion time was set to 30 s. For HEK293T proteome samples, peptide digests were analyzed using a nanoLC/MS/MS system consisting of a WPS-3000PL RS autosampler (Thermo Fisher Scientific), an UltiMate 3000 RSLCnano pump, a FAIMS Pro interface (Thermo Fisher Scientific), and an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific). The mass spectrometric analyses were carried out in the DDA mode. The LC mobile phases consisted of solvent A (0.5% acetic acid) and solvent B (0.5% acetic acid and 80% ACN). Peptides were separated using an in-house-packed needle column (250 mm, 100 μm i.d.) containing Reprosil-Pur 120 C18-AQ 1.9 μm resin (Dr. Maisch) at 50°C in a PRSO-V2 column oven (Sonation GmbH). The flow rate was set to 400 nL/min. The gradient conditions were as follows: 5% B in 5 min, 5–19% B in 55.3 min, 19–29% B in 21 min, 29–40% B in 8.7 min, 40–99% B in 0.1 min, followed by 99% B for 4.9 min. The electrospray voltage was set to 2.4 kV in the positive mode. The FAIMS mode was set to standard resolution, and the total carrier gas flow was 4.0 L/min. The CV was set to −40 and −60, and the cycle time of each CV experiment was set to 1 s. The mass range of the survey scan was from 375 to 1500 m/z with a resolution of 60,000, 300% normalized AGC target, and auto MIT. The first mass of the MS/MS scan was set to 120 m/z with a resolution of 15,000, standard AGC target, and auto MIT. The fragmentation was performed by higher energy collisional dissociation with a normalized collision energy of 30%. The dynamic exclusion time was set to 20 s. For TMT-labeled proteome samples, analysis conditions were essentially the same as above, with the following changes: the gradient conditions were as follows: 5% B in 5 min, 5–12.5% B in 5 min, 12.5–19% B in 55.3 min, 19–29% B in 31 min, 29–40% B in 13.7 min, 40–99% B in 0.1 min, followed by 99% B for 4.9 min. The cycle time of each CV experiment was set to 1.5 s. The mass range of the survey scan was from 375 to 1575 m/z with a resolution of 60,000, 300% normalized AGC target, and a MIT of 50 ms. The first mass of the MS/MS scan was set to 110 m/z with a resolution of 45,000, 200% normalized AGC target, and a MIT of 120 ms. The fragmentation was performed by higher energy collisional dissociation with a normalized collision energy of 32%. The dynamic exclusion time was set to 30 s. Quantification and statistical analysis MS data processing For myoglobin datasets, peptides were identified by automated database searches with Mascot ver. 2.7.0 (Matrix Science) against the horse myoglobin FASTA database (ver. 2020-05), including the MSFragger contaminants fasta database. The peptide mass tolerance was set to 5 ppm, and the fragment ion mass tolerance was set to 20 ppm. Trypsin was specified as the digestion enzyme, allowing up to two missed cleavage sites. Carbamidomethylation (C) was set as a fixed modification, while oxidation (M) and phosphorylation (STY) were set as variable modifications. Peptides were accepted if the Mascot score was over the 95% confidence limit (p < 0.05) based on the identity score of each peptide. For HEK293T proteome datasets, the raw data files were analyzed and processed using MaxQuant (ver. 1.6.17.0),60 and peptides and proteins were identified by automated database searches with Andromeda74 against the UniProt human fasta file (ver. 2022-03, 20,601 human protein entries). Raw data files collected from FAIMS experiments were split into a set of MaxQuant compliant MzXML files using FAIMS MzXML Generator (https://github.com/coongroup/FAIMS-MzXML-Generator).61 The precursor mass tolerance was set to 4.5 ppm, and the fragment ion mass tolerance was set to 20 ppm. Trypsin and LysC were specified as the digestion enzymes, allowing up to two missed cleavage sites. For label-free quantification experiments, carbamidomethylation (C) was set as a fixed modification, and oxidation (M) and phosphorylation (STY) were set as variable modifications. For TMT quantification experiments, TMTpro (peptide N-term and K) and carbamidomethylation (C) were set as fixed modifications. The false discovery rate (FDR) cut-off of 1% was applied for both peptide-spectrum match (PSM) and protein identifications. The minimum Andromeda score cut-off of 40 was applied for modified peptide identifications. For label-free quantification, the match-between-run function was used with default settings (0.7 min match window and 20 min alignment time). Quantification of phospho-probing efficiency of myoglobin Identified myoglobin peptides were quantified by processing MS1 extracted ion chromatograms using Skyline (ver. 21.2.0).62,75 The phospho-probing efficiency of myoglobin was calculated based on the average quantification values of phosphorylated peptides from heat-denatured myoglobin. Quantification of phospho-probing efficiency in cell lysate Phospho-probing efficiency in cell lysates was calculated using the evidence.txt output file from MaxQuant. The Phospho(STY)Sites.txt file was not used because quantitative values for individual phosphorylation sites may be derived from multiple distinct peptides. Similarly, entries in the modificationSpecificPeptides.txt file may contain phosphorylation sites at different positions within the same peptide sequence. The calculation of phospho-probing efficiency was performed using Python (version 3.12.3) based on the evidence.txt file exported from MaxQuant. The data were filtered to exclude entries corresponding to the reverse decoy database, potential contaminants, and peptides with phosphorylation site localization probabilities below 0.75. Only monophosphorylated peptides were used for further analysis. In TMT-based quantification, an additional filter was applied, as previously described,38 to minimize the ion sampling effect76,77; the ratio of the total TMT signal in the carrier channel to the total TMT signal in the sample channel was less than 100. To correct for differences in MS injection amounts across samples, the quantification values of tyrosine-phosphorylated peptides were normalized to the median intensity of identified serine-phosphorylated peptides. For label-free quantification experiments, the quantitative value of the phosphorylation site was based on the peptide species with the largest sum of the intensities of the phosphorylated peptides at the same position in the same peptide sequence. For analyzing the correlation between phospho-probing efficiency of individual tyrosine residues in the substrate proteins with predicted structural features, we defined the intensity ratio of phospho-probed tyrosine-containing peptides in non-denatured samples to those in trypsin-digested samples (Non-denatured/Trypsin-digested ratio) as the phospho-probing efficiency for each residue. A lower ratio indicates reduced phospho-probing efficiency in the non-denatured state, while a higher ratio indicates increased efficiency. For TMT experiments, the quantitative value of phosphorylation sites was based on the peptide species with the lowest posterior error probability (PEP) among those with valid reporter ion signals across all sample channels. Tyrosine phospho-probing efficiency was calculated as the ratio of the mean quantification value of tyrosine phosphorylation sites in the kinase-treated sample (SRC(+)) to that in the non-treated sample (SRC(−)). Phospho-probed sites were defined as those exhibiting a statistically significant (p < 0.05) and greater than 2-fold increase in phospho-probing efficiency. Volcano plot Volcano plots were generated based on the log2 fold change (FC) in the average quantitative values of tyrosine phosphorylation sites (x axis) and the –log10 p-values from Welch’s two-tailed t test (y axis). For differential analysis between kinase-treated and non-treated samples, the quantification values of tyrosine phosphorylation sites in kinase-treated and non-treated samples were used for statistical testing. Tyrosine sites with log2 FC > 1 and p < 0.05 were considered significantly phospho-probed in vitro. For differential analysis of phospho-probing efficiency between RNA-digested and RNA-preserved samples, the SRC(+)/SRC(−) ratios in each sample were used for statistical testing. Tyrosine phosphorylation sites with log2 FC > 1 and p < 0.05 were considered to exhibit a significant increase in phospho-probing efficiency upon RNA digestion. Annotation of structural information on phospho-probed sites Annotation of the structural information of the identified phospho-probed sites was performed using the StructureMap python package.46 Of the proteins in the UniProt human protein FASTA file (2022-03 release, 20,601 proteins), 20,220 predicted structures were downloaded from the AlphaFold database. Those that could not be downloaded were proteins for which AlphaFold did not provide predicted structures, such as proteins with sequence lengths longer than 2,700 and proteins containing non-canonical amino acids. Using the downloaded predicted structures, we obtained the predicted structure confidence level (pLDDT score) of the tyrosine residue and information about the protein secondary structure where the tyrosine residue is located, then the prediction-aware partially sphere exposure (pPSE) was calculated and the intrinsically disordered region (IDR) was predicted. The obtained structural information was matched to the identified phosphorylated receptor residues based on the Uniprot accession ID and residue number. Integration of other databases Known phosphorylation site information was annotated using PhosphoSitePlus (https://www.phosphosite.org/). Among the phosphorylation sites listed in the “Phosphorylation site dataset” (downloaded in 2023-08), sites reported in human proteins were defined as “Reported”. Gene Ontology (GO) annotation and GO enrichment analysis of the identified proteins were performed using DAVID (https://davidbioinformatics.nih.gov/). Protein domain information corresponding to the identified phospho-probed sites was annotated using the “Add sequence features” function in Perseus (ver. 2.0.1.1).63 Analysis of RNA-binding domains known to be phosphorylated in vivo was conducted using a high-confidence, large-scale phosphorylation site dataset filtered by Ramasamy et al. (pr2c00164_si_001.xlsx).54 Protein domain information for phosphorylation sites in this dataset was also annotated using Perseus (ver. 2.0.1.1), and the dataset was matched to the identified phospho-probed sites. A comprehensive dataset of RNA-binding proteins was obtained from RBP2GO (https://rbp2go.dkfz.de/).56 The human RNA-binding protein dataset (downloaded in 2024-05; 6,100 proteins) from RBP2GO was matched to the identified proteins. Comparison with solubility-assay data For comparison with the solubility-assay data reported by Sridharan et al.55 we downloaded the supplementary data containing protein solubility profiles in RNA-preserved and RNA-digested lysates from Sridharan et al. (41589_2022_1062_MOESM2_ESM.xlsx; 41589_2022_1062_MOESM17_ESM.xlsx). The first identifier of “gene_name” used in the supplemental data was mapped to UniProt accession IDs and matched to the proteins identified in this study. Based on the criteria described in the original publication, solubility-assay target proteins were classified as “RNase-sensitive,” “RNase-insensitive,” or “Predominantly soluble”. Mapping of phospho-probed sites on protein 3D structures Phospho-probed sites were mapped onto representative 3D protein structures obtained from either the PDB database (https://www.rcsb.org/) or the AlphaFold database (https://alphafold.ebi.ac.uk/). Protein structures were visualized using PyMOL (https://pymol.org/). The structures of horse myoglobin, RPL5, SNRNP70, and SRSF1 were retrieved from the PDB. For RPL5, the minimum Euclidean distance between tyrosine residues that exhibited increased phospho-probing efficiency upon RNA digestion and nearby RNA molecules was calculated using the “distance” function in PyMOL. Predicted structures for EIF3D and HNRNPL were obtained from the AlphaFold database. Statistical analysis Quantitative data are expressed in bar and line graphs as mean ± SD (error bar). Two-tailed Welch’s t test, the Fligner–Killeen test and two-sided Fisher’s exact test were performed using Scipy64 (version 1.13.1). Mann-Whitney U test and statistical annotations on boxplots were performed using statannotations65 (version 0.6.0). Published: May 27, 2026 Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.crmeth.2026.101466. Contributor Information Kosuke Ogata, Email: ogata.kosuke.7w@kyoto-u.ac.jp. Yasushi Ishihama, Email: yishiham@pharm.kyoto-u.ac.jp. Supplemental information Document S1. Figures S1–S4 Table S1. All experimental information of phospho-probing (substrate amount, reaction time, TMT channel, etc.), related to STAR Methods Table S2. Identified phosphorylation sites and quantification results in the protease-digested proteome, related to Figure 3 Table S3. Identified phosphorylation sites, quantification results, and statistical analyses in the RNA-digested proteome, related to Figure 4 Document S2. Article plus supplemental information References Associated Data Supplementary Materials Document S1. Figures S1–S4 Table S1. All experimental information of phospho-probing (substrate amount, reaction time, TMT channel, etc.), related to STAR Methods Table S2. Identified phosphorylation sites and quantification results in the protease-digested proteome, related to Figure 3 Table S3. Identified phosphorylation sites, quantification results, and statistical analyses in the RNA-digested proteome, related to Figure 4 Document S2. Article plus supplemental information Data Availability Statement •The raw MS data and analysis files have been deposited with the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (https://jpostdb.org/) with the dataset identifier PXD063369. The raw MS data and analysis files have been deposited with the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the jPOST partner repository (https://jpostdb.org/) with the dataset identifier PXD063369. •This paper does not report original code. This paper does not report original code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.