High-Throughput Simultaneous Quantification of Glycopeptides and Phosphopeptides Enabled by 12-plex DiLeu Isobaric Tags and Dual-Functional Titanium(IV)-IMAC Material

Abstract

Background:

Protein glycosylation and phosphorylation are critical post-translational modifications (PTMs) that regulate diverse physiological and pathological processes, yet their comprehensive characterization remains challenging due to low abundance and poor ionization efficiency. Recent advances using epoxy-ATP-Ti⁴⁺-IMAC materials have enabled simultaneous enrichment of N-glycopeptides, phosphopeptides, and mannose-6-phosphate glycopeptides. However, high-throughput, multiplexed quantification of these PTMs is still lacking. This work addresses the need for an efficient strategy capable of simultaneously enriching, identifying, and quantitatively comparing glycosylation and phosphorylation across multiple biological samples.

Results:

We developed a high-throughput workflow integrating epoxy-Ti⁴⁺-IMAC enrichment with custom N,N-dimethyl leucine (DiLeu) isobaric tags to achieve 12-plex quantitative analysis of N-glycosylation and phosphorylation for the first time. This streamlined one-tube sample preparation protocol enabled robust, simultaneous enrichment and quantification of PTMs from complex mouse brain samples. Application to APP/PS1 transgenic mice versus wild-type controls produced quantitative identification of 1,975 N-glycopeptides and 1,181 phosphopeptides. Comparative profiling revealed substantial PTM alterations associated with Alzheimer’s disease (AD)-related pathology. Differentially modified proteins mapped to key biological pathways, including synapse organization, synaptic membrane regulation, and cell adhesion. The abundance patterns highlighted broad disruptions in PTM-mediated signaling and provided molecular insights into synaptic dysfunction in the APP/PS1 model.

Significance and Novelty:

This integrated DiLeu isobaric labeling–epoxy-Ti⁴⁺-IMAC platform provides a powerful, high-throughput solution for the simultaneous quantification of glycosylation and phosphorylation, enabling detailed investigation of PTM interplay. By uncovering disease-associated modifications in AD mouse models, this method offers new opportunities to identify mechanistic biomarkers and therapeutic targets. Its versatility and scalability make it broadly applicable to PTM-centric studies across diverse biological systems, including biofluids, cell lysates and tissues.

Graphical Abstract

1. Introduction

Post-translational modifications (PTMs) are critical biochemical processes that occur in proteins either shortly after translation by ribosomes or once protein folding and localization are complete[1]. These modifications, typically catalyzed by specific enzymes, involve the covalent addition of various chemical groups, sugars, or even entire proteins to defined amino acid residues on target proteins. Among the many known PTMs, glycosylation and phosphorylation are two of the most extensively studied PTMs due to their pivotal roles in diverse biological processes. Phosphorylation is a dynamic and reversible PTM that involves the covalent attachment of a phosphate group to the hydroxyl group of serine, threonine, or tyrosine residues within a protein[2]. It is estimated that approximately one-third of proteins in mammalian cells are phosphorylated at any given time[3]. This process is catalyzed by kinases, which transfer the phosphate group from adenosine triphosphate (ATP) to target residues, while phosphatases remove phosphate groups to ensure tight regulation of protein function and signaling pathways[4]. On the other hand, glycosylation involves the enzymatic attachment of carbohydrate moieties (glycans) to specific amino acid residues on proteins[5]. Nearly half of all proteins undergo some form of glycosylation, which is essential for both physiological and pathological cellular functions [6,7]. Dysregulation of glycosylation and phosphorylation has been linked to the development and progression of numerous diseases, including Alzheimer's disease (AD) and various cancers[7–10]. Therefore, site-specific investigation of these PTMs is crucial for understanding their biological roles. Moreover, analyzing multiple PTMs concurrently not only deepens understanding of individual modifications but also enables exploration of crosstalk between different PTMs[11].

Despite their importance, direct mass spectrometry (MS) analysis of phosphopeptides and glycopeptides remains challenging due to their low abundance and poor ionization efficiency[12]. To address these issues, robust enrichment methods are essential prior to MS analysis. Phosphopeptides are commonly enriched using metal oxide affinity chromatography (MOAC) or immobilized metal affinity chromatography (IMAC)[13,14]. For glycopeptides, common enrichment methods include hydrazide chemistry, lectin affinity chromatography, boronic acid enrichment, and hydrophilic interaction chromatography (HILIC)[15,16]. However, a major limitation of these strategies is that they typically focus on a single PTM, even though PTMs frequently occur simultaneously on proteins and can influence one another through intricate crosstalk [17]. This interplay includes competitive or noncompetitive occupancy of proximal modification sites, glycosylation of kinases, and phosphorylation of glycosylation-related enzymes[18]. To investigate such interactions, it is essential to examine PTMs in a site-specific context on the same proteins. As a result, several studies have adopted sequential enrichment methods[19–21]. Previously, our lab explored the hydrophilic properties of commercially available centrifuge-assisted-extraction Ti⁴⁺-IMAC (CAE-Ti-IMAC) material for its potential to simultaneously enrich phosphopeptides, glycopeptides, and M6P glycopeptides through dual-mode enrichment[22,23]. This approach leverages a HILIC-like mechanism via hydroxyl groups on the material’s surface, while phosphopeptides are eluted through electrostatic interactions. However, insufficient hydrophilicity limited the material’s glycopeptide enrichment efficiency. To overcome this challenge, our group recently developed a hydrophilicity-enhanced bifunctional Ti-IMAC material grafted with ATP (epoxy-ATP-Ti⁴⁺), which is also easier to prepare. This new material demonstrates excellent performance in the simultaneous enrichment of both glycosylated and phosphorylated peptides, enabling large-scale profiling of both PTMs via liquid chromatography-tandem MS (LC-MS/MS)[24].

A major bottleneck, however, remains quantitative analysis. Dual-functional enrichment strategies often produce more than six fractions per sample, and when multiple samples must be analyzed, this significantly increases run-to-run variability as well as sample preparation and instrument time[25]. To achieve high-throughput analysis in a single MS/MS run, stable isotope labeling techniques, particularly isobaric labeling methods such as commercially available tandem mass tags (TMTs) and our custom-developed DiLeu isobaric tags, have become increasingly popular for comparative proteomic and PTM studies of different biological states[26–32]. These strategies have proven effective in generating accurate quantitative results with high multiplexing capability, reducing run-to-run variation and increasing analytical throughput in numerous proteomics studies[33–36]. Compared to commercial tags, DiLeu reagents are significantly more cost-effective and can be synthesized in-house with high yield [27]. When combined with epoxy-Ti⁴⁺-IMAC enrichment, DiLeu labeling enables high-throughput PTM quantitation compared to label-free single-shot analysis, which is essential for studying PTM crosstalk.

In this study, we developed a workflow combining 12-plex DiLeu isobaric labeling with epoxy-Ti⁴⁺-IMAC enrichment for comprehensive quantitative N-glycoproteomic and phosphoproteomic analyses (Figure 1). To enhance peptide recovery from precious or limited biological samples, we implemented a one-tube concept strategy using an acid-cleavable detergent, eliminating the need for a separate desalting step. The method was first validated using PANC-1 cell lysates and then applied to characterize N-glycosylation and phosphorylation in APP/PS1 AD model mouse brains. This approach proved highly effective for simultaneous quantification of glycopeptides and phosphopeptides with minimal inter-fraction interference and holds strong potential for future applications in site-specific quantitative analysis of both phosphorylation and N-glycosylation and their crosstalk.

Figure 1.

Workflow for the simultaneous quantification of N-glycopeptides and phosphopeptides using epoxy-ATP-Ti⁴⁺ IMAC material and 12-plex DiLeu isobaric labeling.

2. Experimental section

2.1. Chemicals

Dithiothreitol (DTT) and sequencing grade trypsin were from Promega (Madison, WI). Optima LC/MS grade solvents, formic acid (FA), urea, sodium chloride, phosphoric acid, ammonium hydroxide, sodium carbonate and 2,5-Dihydroxybenzoic acid (DHB) were from Fisher Scientific (Pittsburgh, PA). Trifluoroacetic acid (TFA), iodoacetamide (IAA), triethylammonium bicarbonate (TEAB), N,N-dimethylformamide (DMF), 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium tetrafluoroborate (DMTMM) were purchased from Sigma-Aldrich (St Louis, MO). Titanium sulfate was purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Oasis HLB 1 cc (10 mg) extraction cartridges, C18 SepPak cartridges and RapiGest SF (RapiGest) were purchased from Waters Corporation (Milford, MA). Empty 200 μL TopTips were from Glygen Corp (Columbia, MD). Strong cation exchange (SCX) spin tips and epoxy-functionalized silica bulk material (12 μm, 300 Å) were obtained were from PolyLC (Columbia, MD). Protease inhibitor cocktail tablets and phosphatase inhibitor cocktail tablets were from Roche (Mannheim, Germany). Centrifuge-assisted extraction Ti-IMAC (CAE-Ti-IMAC) microspheres were obtained from J&K Scientific Ltd. (Beijing, China). All other chemicals and LC-MS grade solvents were purchased from Fisher Scientific (Pittsburgh. PA).

2.2. Mouse Brain Tissue Collection

The brain tissues were prepared per previous work[37]. Briefly, APP695/swe/PS1-dE9 (APP/PS1) double transgenic male mice (Stock No. 34832-JAX, Jackson Laboratory) were genotyped at weaning by Transnetyx (Cordova, TN) and studied at nine months of age (n = 6), alongside age-matched WT littermates as controls (n = 6). All male mice were housed in standard cages with 1–5 mice per cage under controlled conditions (12-hour light-dark cycle, 50% humidity, 24 °C) and provided standard chow and water ad libitum. Sex and gender-based analysis was not conducted. Animal experiments complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin-Madison (protocol #M005120). For tissue collection, mice were euthanized using a CO₂ chamber, and brains were rinsed with ice-cold PBS before dissection to isolate the cortex. The tissue was snap-frozen in liquid nitrogen and stored at −80 °C for further analysis.

2.3. Sample Preparation

The protocol of sample preparation followed previous publications[36]. The details of cell culture, protein extraction, and protein digestion are provided in the Supplementary Information.

DiLeu Labeling The synthesis and labeling of DiLeu tags were conducted, as previously reported[27]. DiLeu tags were activated in anhydrous DMF with DMTMM and NMM at 0.6 × molar ratios to tags. The reaction mixture was vortexed at room temperature for 1 h. Subsequently, the supernatant was added to each sample for labeling. After an additional 2 hours of vortexing at room temperature, the reaction was quenched by adding NH₂OH to a final concentration of 0.25%. Labeled peptides from each batch were pooled and dried in vacuo. For proteomics analysis, aliquots of the dried peptides were then subjected to cleanup using SCX SpinTips, following the manufacturer’s protocols, and subsequently desalted with C18 SepPak cartridges. Finally, the desalted peptides were dried in vacuo.

2.4. Simultaneous Enrichment of N-glycopeptides and Phosphopeptides

Synthesis of epoxy-ATP-Ti⁴⁺ IMAC material was prepared per previous work[24]. The material was packed into an empty TopTip, with 3 mg of cotton placed at the bottom. The TopTip was then positioned atop a 2 mL microcentrifuge tube using an adapter. For simultaneous glycopeptide and phosphopeptide enrichment, 500 μg peptide from PANC-1 cell digest or mouse brain tissue digest was loaded onto the epoxy-Ti⁴⁺-IMAC-cotton tip in a loading buffer composed of 80% ACN and 3% TFA. The flow-through was collected and reloaded five additional times. The epoxy-Ti⁴⁺-IMAC -cotton tip was subsequently washed with 300 μL loading buffer and centrifuged at 200 g for 2 min, a process repeated six times to remove non-modified peptides. Different numbers of elution fractions were collected at 200 g centrifugation speed depending on the sample type. For PANC-1 cell digests and mouse brain sample, six fractions were collected for analysis. Fraction 1 was eluted using 300 μL of 60% ACN/0.1% FA, Fraction 2 with 300 μL of 40% ACN/0.1% FA, and Fraction 3 with 300 μL of 20% ACN/0.1% FA followed by 300 μL of 0.1% FA. Fraction 4 comprised pooled eluates from 300 μL of 40% ACN/3% TFA, 300 μL of 50% ACN/6% TFA/200 mM NaCl, and 300 μL of 30% ACN/0.1% TFA, and this fraction was desalted using an HLB cartridge. Fraction 5 was eluted by 300 μL 60% ACN/10% NH₄OH and 300 μL 40% ACN/10% NH₄OH and fraction 6 consisted of eluates from 300 μL 20% ACN/10% NH₄OH, 300 μL 10% ACN/10% NH₄OH and 300 μL 10% NH₄OH. Between fractions 4 and 5, the IMAC-cotton tip was reconditioned by washing with 300 μL of 90% ACN/2.5% NH4OH three times. Samples were dried down in vacuo.

2.5. NanoLC-MS/MS analysis

Samples were analyzed using LC–MS/MS with an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, San Jose, CA) connected to a Dionex Ultimate 3000 UHPLC system (Thermo Fisher Scientific, San Jose, CA). Glycopeptide fractions eluted under acidic conditions were reconstituted in 0.1% FA and phosphopeptide fractions collected under basic conditions were reconstituted in 20 mM citric acid/1% FA solution before MS analysis[38]. Peptide separation was conducted on a 15 cm length, 75 μm i.d. in-house packed BEH C18 (1.7 μm, 130 Å, Waters) capillary column with an 80-min gradient from 3 to 30% ACN (0.1% FA) at a flow rate of 0.3 μL/min. Data acquisition was conducted in top speed mode with a cycle time of 3 seconds. For N-glycopeptides analysis, peptide precursor scans were performed from m/z 400 to 2000 at a resolution of 60,000, with an automatic gain control (AGC) target of 4E5 and a maximum injection time (IT) of 50 ms. The selected precursors underwent higher-energy C-trap dissociation (HCD) with a normalized collision energy (NCE) of 30 and a ±3% stepped HCD collision energy. Tandem MS acquisition was also performed at a resolution of 60,000, with a lower mass limit of m/z 110, and a dynamic exclusion time of 12 seconds with a 10 ppm mass tolerance. For phosphopeptide analysis, the MS scan range covered m/z 300 to 1800, also at a resolution of 60,000, with an AGC target of 2E5 and a maximum IT of 50 ms. The MS/MS method utilized a top 20 data-dependent acquisition (DDA) mode, with all MS/MS dissociations conducted using an NCE of 30 and ±8% stepped HCD collision energy. The MS/MS parameters included a resolution of 60,000, an AGC target of 5E4, and a maximum IT of 118 ms. Each sample was acquired in technical triplicates.

2.6. Data Analysis

Byonic software (version 2.9.38, Protein Metrics Inc, San Carlos, CA) embedded with the Proteome Discover 2.5 (PD 2.5, Thermo Fisher Scientific) was used to process intact N-glycopeptide, phosphopeptide and proteomics data. Raw files were searched against UniProt Homo sapiens reviewed database (December 2023) or UniProt Mus musculus reviewed database (April 2024). Trypsin was selected as the enzyme and two maximum missed cleavages were allowed. Searches were performed with a precursor mass tolerance of 10 ppm and a fragment mass tolerance of 0.02 Da. Fixed modifications were specified as carbamidomethylation (+57.02146 Da) on cysteine residues and 12-plex DiLeu (+145.12801 Da) on peptide N-terminus and lysine residues. Dynamic modifications included oxidation of methionine (+15.99492 Da), phosphorylation on serine, threonine and tyrosine (+79.96633 Da) and N-glycosylation. Glycan modifications were searched against a human or mammalian glycan database expanded from Byonic embedded N-glycan database to include N-linked M6P glycans consisting of HexNAc (2–4) Hex (3–9) Phospho (1–2) modifications[23]. Peptide identification results were filtered at Byonic score > 150, PEP 2D < 0.05, |Log Prob| > 1. MS/MS spectra of M6P glycopeptide PSMs were manually inspected to ascertain the existence of phosphorylated hexose diagnostic ions. Proteomics data were searched the same as above except without dynamic modifications of phosphorylation or N-glycosylation. A two-sample Student’s t test with a two-detailed distribution for binary comparison were conducted using Perseus (version 2.0.11.0) and the p value was subjected to permutation-based FDR for multiple testing corrections. The MS data has been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD061371[39].

3. Results and discussion

3.1. Simultaneous Quantification of N-glycopeptides and Phosphopeptides Using 12-plex DiLeu Isobaric Tags and Epoxy-ATP-Ti⁴⁺-IMAC Material

PANC-1 cells were used in initial experiments to validate the strategy for simultaneous quantitative analysis of N-glycopeptides and phosphopeptides by combining DiLeu isobaric labeling with epoxy-Ti⁴⁺-IMAC enrichment. Given the inherently low-abundance of PTMs, a one-tube, desalting-free DiLeu labeling workflow was adopted to minimize sample loss and reduce experimental steps. As described in a previous report[24], the epoxy-ATP-Ti⁴⁺-IMAC material enables dual-mode enrichment, allowing a single sorbent to capture two classes of peptides through synergistic interactions and gradient elution. In this study, 500 μg of 12-plex DiLeu-labeled PANC-1 tryptic digests were loaded onto the epoxy-ATP-Ti⁴⁺-IMAC material under HILIC conditions, and six fractions were collected for subsequent analysis. The first three fractions, eluted under mildly acidic conditions with decreasing concentrations of ACN, were expected to release glycopeptides retained via hydrophilic interactions, predominantly neutral glycopeptides. The fourth fraction, eluted with varying ACN and salt concentrations under strongly acidic conditions, resembled the traditional loading and washing buffer used in the IMAC-based phosphopeptide enrichment. This fraction was expected to release the sialylglycopeptides, which were captured through a combination of hydrophilic and electrostatic interactions. The strong acid protonates sialic acid residues, disrupting electrostatic binding, while 200 mM NaCl reduces non-specific ionic adsorption of non-phosphorylated peptides. Remaining sialylglycopeptides are eluted in alkaline conditions by a high concentration of hydroxide ions, which competes with sialic acid residues for interaction with Ti⁴⁺ ions in the material. Phosphopeptides and M6P glycopeptides, however, remained bound to the IMAC material because their phosphate groups have lower pKa values than sialic acid. The final two elution steps with ammonium hydroxide were designed to release the remaining phosphopeptides and M6P glycopeptides[22,24]. Despite their low abundance and susceptibility to ion suppression, M6P glycopeptides were successfully separated from phosphopeptides under high-pH HILIC conditions and confidently detected by MS. Figures 2A and 2 B show the identification numbers of modified peptides enriched using the epoxy-ATP-Ti⁴⁺-IMAC material with both a 1:1:1:1:1:1:1:1:1:1:1:1 ratio and a 1:2:5:10:5:2:1:2:5:10:5:2 DiLeu mixing ratio, demonstrating accurate quantitative performance in both conditions. The increasing abundance of sialylglycopeptides across the fractions confirmed successful dual-mode separation for both the 1:1:1:1:1:1:1:1:1:1:1:1 equal-mix ratio (Figures S1A) and the 1:2:5:10:5:2:1:2:5:10:5:2 dynamic-range ratio (Figure S1B). Such separation is crucial for enhancing the detection sensitivity of sialylated and M6P glycopeptides, as their ionization is often suppressed by coeluting neutral glycopeptides due to their negative charge. Similarly, epoxy-ATP-Ti⁴⁺-IMAC material demonstrated comparable affinity and performance and for DiLeu-labeled phosphopeptides under both mixing conditions (Figures S1C and S1D).

Figure 2.

Simultaneous quantification of glycopeptides and phosphopeptides from PANC-1 cells using epoxy-ATP-Ti⁴⁺-IMAC materials. Identification numbers refer to peptides in each fraction enriched by epoxy-ATP-Ti⁴⁺ IMAC material in (A) 1:1 ratios and in (B) 1:2:5:10ratios. The mixed TFA/ACN fraction was pooled from three eluates: 40% ACN/3% TFA, 50% ACN/6% TFA/200 mM NaCl, and 30% ACN/0.1% TFA, and was desalted prior to LC-MS/MS analysis.

We further evaluated the quantitative accuracy and dynamic range of 12-plex DiLeu-labeled peptides enriched using epoxy-ATP-Ti⁴⁺-IMAC material for high-throughput quantification of glycopeptides and phosphopeptides. As illustrated in Figure 3, the median ratios were closely aligned with the expected values, with deviations within 5% for both the 1:1:1:1:1:1:1:1:1:1:1:1 equal-mix and the 1:2:5:10:5:2:1:2:5:10:5:2 dynamic-range ratio samples. These findings highlight the excellent reliability and robustness of this platform across a practical dynamic range, establishing it as a powerful tool for high-throughput PTM quantification. Compared with label-free approaches, this method offers substantial advantages. DiLeu isobaric labeling dramatically reduces instrument time and mitigates run-to-run variability, particularly when multiple fractions must be analyzed. The DiLeu strategy is also fully compatible with widely used software platforms, including MaxQuant, PEAKS, and MSFragger. Specifically, up to 12 biological samples can be analyzed in parallel across the fractionated workflow, yielding approximately 14 hours of LC–MS acquisition time. Additionally, by adapting the Nanogram TMT Processing in One Tube (NanoTPOT) strategy [40] to enable desalting-free isobaric labeling, the dual-PTM enrichment workflow eliminates the need for sample desalting, thereby minimizing sample loss and enhancing analytical throughput. Consequently, this approach is both efficient and scalable, making it highly suitable for large-cohort studies where multiplexed quantification is required.

Figure 3.

Quantitative performance. The 12-plex DiLeu-labeled glycopeptides from PANC-1 cell lysate were combined in (A) 1:1 ratios across all channels and in (B) 1:2:5:10ratios. The 12-plex DiLeu-labeled phosphopeptides were combined in (C) 1:1 ratios and in (D) 1:2:5:10:5:2:1:2:5:10:5:2 ratios.

3.2. 12-plex DiLeu Labeling Coupled with Epoxy-ATP-Ti⁴⁺-IMAC in Mouse Brain

AD is a progressive neurodegenerative disorder characterized by the accumulation of amyloid-beta plaques and neurofibrillary tangles[42]. While extensive research has explored glycosylation and phosphorylation in AD, a critical gap remains in understanding the molecular mechanisms underlying the crosstalk between these PTMs and their contributions to disease progression. Therefore, it is crucial to quantitatively and qualitatively investigate dysregulated glycosylation and phosphorylation in AD models to identify PTM signatures associated with disease progression. Here, we adapted our high-throughput platform to study dysregulated glycosylation and phosphorylation in AD mice.

First, we validated the feasibility of this strategy in mouse cortex. Figure 4A demonstrates effective separation of glycopeptides and phosphopeptides. Specifically, 447, 551, 653, and 300 glycopeptides were detected in the acidic fractions, while 543 and 629 phosphopeptides were identified in the basic fractions. Phosphopeptides were largely absent from the first four fractions, and only a few glycopeptides appeared in the last two fractions. Multisialylated glycopeptides, which have high affinity for Ti⁴⁺ ions, eluted only at high pH (Figure 4B). Similarly, Figure 4C shows that the proportion of phosphopeptides and multiphosphopeptides increased with decreasing ACN concentration, and all M6P glycopeptides were found in the last fraction. The dual-mode elution strategy at different pH levels enabled the identification of 1975 unique N-glycopeptides and 1181 unique phosphopeptides, including 188 sialylated glycopeptides, 8 M6P glycopeptides, and 236 multi-phosphorylated peptides (Figure 4D). These PTM-modified peptides were mapped to 353 glycoproteins and 493 phosphoproteins, with 44 proteins exhibiting both glycosylation and phosphorylation, and 7 M6P glycoproteins also showing phosphorylation (Figure 4E). These results demonstrate that this method facilitates quantitative enrichment and separation of diverse glycopeptides and phosphopeptides, highlighting its strong potential for analyzing PTM crosstalk.

Figure 4.

Simultaneous quantification of glycopeptides and phosphopeptides from mouse cortex tissue. (A) Number of peptides identified in each fraction; (B) Percentages of sialylated and multi-sialylated glycopeptides in six fractions; (C) Percentages of phosphopeptides and multi-phosphopeptides in six fractions; (D) Total number of identified modified peptides; (E) Venn diagram showing the coexistence of N-glycosylation and phosphorylation at the protein level. * The mixed TFA/ACN fraction was pooled from three eluates: 40% ACN/3% TFA, 50% ACN/6% TFA/200 mM NaCl, and 30% ACN/0.1% TFA, and was desalted prior to LC-MS/MS analysis.

3.3. Site-specific Quantitative Analysis of Glycopeptides in AD Mouse Brain versus WT

In brief, glycopeptides and phosphopeptides were enriched from pooled 12-plex DiLeu- labeled brain tissues of AD mice and WT controls using epoxy-Ti⁴⁺-IMAC material. To ensure reproducible and unbiased analysis, we collected six biological replicates for both the AD and control groups, and MS data were acquired in technical replicates as well. For glycopeptides, we performed unsupervised hierarchical clustering of significantly altered quantified glycoforms to explore their profiles across groups (Figure 5A). This heatmap illustrates column-wise clustering of biological replicates within either the AD or control group, indicating larger intergroup differences than intragroup variations and thus demonstrating the robustness of our method for simultaneous quantitative analysis of protein glycosylation. We then examined the differences between the AD and control groups in greater detail. We found that 47 glycoforms from 36 proteins were significantly altered between the two groups (two-sided t-test, p value < 0.05, permutation-based FDR correction)(Figure 5B; Table S1). Detail information on the significantly altered glycoforms is listed in Table S1. For example, one glycoform of ADAM10 (N279_HexNAc(2)Hex(7)) displayed a downward trend in the AD group. ADAM10 is a critical alpha-secretase involved in the non-amyloidogenic processing of amyloid precursor protein (APP). Reduced levels of ADAM10 could shift APP processing toward the amyloidogenic pathway, increasing amyloid-beta production and contributing to plaque formation[43]. Additionally, less-studied glycoproteins such as Neurofascin and GRM3 also exhibited downregulation, providing promising avenues for understanding synaptic and axonal dysfunctions in AD[44,45]. Their dysregulation highlights critical yet underexplored molecular mechanisms that warrant further investigation. The comprehensive intact glycoproteome coverage also enabled the global quantification of site-specific microheterogeneity in brain glycosylation, as visualized in Figure 5C. This glycoprotein-glycan network diagram maps which glycans (outer nodes, 28 total) modified which proteins (inner bar, 45 total). The glycoproteins were sorted by the number of glycosylation sites, and all nodes and edges were colored according to glycosylation type. Of note, several glycosylation patterns were observed, including the prevalence of high mannose structures. AD mouse brain displayed a roughly even distribution of high-mannose and fucosylated glycans, whereas complex/hybrid and sialylated species appeared less frequently across multiple sites within the same protein (Figure 5C). These findings highlight alterations in the specificity of individual glycosylation events influenced by pathological processes in the AD model.

Figure 5.

High-throughput investigation of glycosylation changes in AD mice and healthy controls. (A) Hierarchical clustering of DiLeu reporter ion intensities of significantly altered glycoforms in WT and AD groups. (B) Volcano plot comparing glycopeptide expression levels between AD and WT groups. Points above the horizontal dash line represent significantly altered peptides (two-sided t-test, p value < 0.05). (C) Glycoprotein-glycan network diagram mapping which glycans (outer nodes) modify which proteins (inner bars). Glycoproteins are sorted by the number of glycosites. Glycan nodes and their links to proteins are colored according to glycosylation type.

To functionally categorize the quantified glycopeptides, GO annotations of the glycoproteins identified by the DiLeu isobaric labeling-epoxy-Ti⁴⁺-IMAC workflow were performed across three categories: biological process (Figure S2A), cellular component (Figure S2B), and molecular function (Figure S2C). The top 10 most significant categories were plotted. Among biological processes, the most enriched terms were associated with synapse organization and assembly, representing key pathways in brain physiology. The top cellular component categories revealed that the identified glycoproteins were predominantly synapse-related, highlighting the potential critical role of glycoproteins in maintaining brain function. This is consistent with studies showing that aberrant glycosylation of synaptic proteins is linked to synaptic dysfunction and cognitive decline[46] ​. Molecular function analysis indicated that most of the glycoproteins were involved in binding activities, suggesting alterations in receptor-ligand interactions and enzymatic processes in the AD brain[47].

3.4. Investigating the phosphoproteome in AD Mice and its crosstalk with the glycoproteome

Another major advantage of our method is its ability to facilitate comprehensive quantification of the phosphoproteome from the same sample. In total, we identified 1181 unique phosphopeptides corresponding to 493 phosphoproteins (Figure 4D). Notably, in the AD mouse brain, we detected four significantly altered phosphopeptides (Figure S3; Table S2), all of which exhibited an upregulation trend.

To assess potential cross-regulation between glycosylation and phosphorylation, we specifically analyzed proteins that carry both PTMs. Remarkably, 44 proteins were found to contain both glycosylation and phosphorylation sites (Figure 4E). A representative example is cell adhesion molecule 3, a member of the immunoglobulin superfamily that plays a pivotal role in mediating cell-cell adhesion, particularly within the nervous system[48]. In our study, cell adhesion molecule 3 was found to possess six distinct glycoforms and one phosphosite (Ser386), underscoring its dual role in PTM regulation (Figure 5). Importantly, these glycoforms represent novel findings for this protein, as none have been previously reported. Among these glycoforms, three are sialylated and five are fucosylated, highlighting the critical involvement of these glycosylation types in AD pathogenesis [49,50]. The identification of these previously uncharacterized glycoforms provides new insights into the complex regulatory mechanisms governing cell adhesion molecule 3, particularly regarding its role in synaptic function and its potential involvement in neurodegenerative disorders such as AD. Our simultaneous enrichment and quantification strategy establishes a robust framework for investigating the cross-regulation of glycosylation and phosphorylation in modulating the functions of cell adhesion molecule 3 and related proteins. Future studies focusing on this protein and other dually modified targets may uncover critical mechanisms of disease progression and contribute to the identification of novel therapeutic targets. Overall, this study not only deepens our understanding of the AD phosphoproteome and glycoproteome but also underscores the importance of integrated PTM-focused approaches in unraveling the molecular complexity of neurodegenerative diseases.

4. Conclusions

This study presents a robust and highly effective analytical method that enables simultaneous, high-throughput quantification of both glycosylation and phosphorylation. By integrating a streamlined, one-tube sample processing workflow with DiLeu isobaric labeling, we simplified the experimental procedures and minimized sample loss. As a proof-of-principle experiment, the DiLeu-epoxy-ATP-Ti⁴⁺-IMAC strategy was applied to enrich different PTMs from AD mouse brains and WT samples. Our approach identified 1975 N-glycopeptides corresponding to 353 glycoproteins, among which 47 glycopeptides showed significant dysregulation. GO analysis revealed a strong association between these glycopeptides and synaptic function, underscoring their potential roles in maintaining brain physiology. Similarly, 1181 phosphopeptides were identified from 493 phosphoproteins, with four showing significant alterations. These findings provide valuable insights into the roles of glycosylation and phosphorylation in disease pathogenesis, highlighting their intricate interplay in the molecular mechanisms underlying AD. Importantly, both the epoxy-ATP-Ti^4_IMAC material and the DiLeu isobaric tags are straightforward to prepare, making this workflow readily adoptable by other MS laboratories as a general analytical strategy for high-throughput quantitative analyses of protein glycosylation and phosphorylation. This approach can be further extended to investigate a wide range of diseases in which dysregulation of these PTMs is implicated, enabling high-throughput, multiplexed PTM analysis. Future efforts will focus on extending the strategy to quantitatively enrich additional glycosylation types, including O-glycosylation[51,52]. We anticipate that this newly developed method, together with the associated fractionation approach, will greatly facilitate related studies and advance our understanding of the N-glycoproteome, phosphoproteome, and their potential interactions in health and disease.

Supplementary Material

Supplementary Figures

Figure S1. Simultaneous quantification of glycopeptides and phosphopeptides from PANC-1 cells using epoxy-ATP-Ti⁴⁺-IMAC materials. Distribution of glycopeptides with different numbers of sialic acids enriched by epoxy-ATP-Ti⁴⁺-IMAC material in (A) 1:1 ratios and in (B) 1:2:5:10 ratios; percentages of phosphopeptides and multi-phosphorylated peptides in fractions enriched by epoxy-ATP-Ti⁴⁺-IMAC material in (C) 1:1 ratios and in (D) 1:2:5:10 ratios.

Figure S2. Gene ontology analysis of glycoproteins identified in the strategy, categorized by (A) biological process, (B) cellular component, and (C) molecular function. The top 10 most significant categories are shown.

Figure S3. Volcano plot of phosphopeptides from AD and WT mouse brains (two-sided t test, p value < 0.05).

This article following supporting information available free of charge.

Additional experimental methods, Figures S1–S3, Tables S1–S2.

Figure 6.

Glycan diversity and phosphorylation on cell adhesion molecule 3 (N: HexNAc, H: Hex, F: fucose, S: NeuAc).

Data availability

The MS data has been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD061371, with Token “y9qLgGU0wLn7”. Alternatively, reviewers can get access by logging in to the PRIDE website using the username “reviewer_pxd061371@ebi.ac.uk“ and password “CDUDAI2oGkNW”.