Comparative Analysis of Lysis Buffers for Enhanced Proteomic and Glycoproteomic Profiling

TL;DR

This study compares lysis buffers for proteomic and glycoproteomic profiling, finding that SDS provides the best coverage and reproducibility.

Contribution

Demonstrates that SDS lysis buffer outperforms others in proteome and N-glycoproteome depth and reproducibility.

Findings

SDS achieved the deepest proteome and N-glycoproteome coverage in HeLa and HEK293T cells.

SDS enabled comprehensive extraction from multiple subcellular compartments and glycosylation types.

SDS showed superior reproducibility with 85% of quantified proteins having coefficients of variation below 5%.

Abstract

Efficient and reproducible protein extraction is a critical prerequisite for high-quality proteomic and glycoproteomic analyses. In this study, four commonly used lysis buffers, sodium dodecyl sulfate (SDS), guanidine hydrochloride (GuHCl), urea (UA), and mammalian protein extraction reagent (MPER), were systematically evaluated within an integrated proteomic and N-glycoproteomic workflow. Using HeLa and HEK293T cells as model systems, we assessed buffer performance in terms of protein and intact N-glycopeptide identification depth, quantitative reproducibility, subcellular coverage, and glycan type distribution. Across both cell lines, SDS consistently achieved the deepest proteome and N-glycoproteome coverage, yielding the highest numbers of identified proteins, N-glycopeptides, glycoproteins, and glycosylation sites. Quantitative analysis demonstrated that SDS provided superior reproducibility, with approximately 85% of quantified proteins exhibiting coefficients of variation below 5%. Subcellular localization analysis at the global proteome level showed that SDS enabled more comprehensive extraction of proteins from multiple cellular compartments, including the nucleus, cytoplasm, mitochondria, and plasma membrane, indicating reduced extraction bias toward specific subcellular regions. Consistently, subcellular localization analysis of identified glycoproteins revealed enhanced coverage of membrane-associated compartments, particularly the plasma membrane, endoplasmic reticulum, Golgi apparatus, and lysosome. In addition, the analysis of glycan type classification for intact N-glycopeptides revealed that the SDS lysis buffer demonstrated the most comprehensive identification capability for glycopeptides with multiple glycosylation modifications in both cell lines. MPER and UA showed a highly consistent distribution across various glycosylation types, whereas the guanidine hydrochloride method was comparatively least effective. Overall, these results establish SDS as a robust lysis buffer for comprehensive, reproducible, and quantitatively stable proteomic and N-glycoproteomic analyses, providing practical guidance for buffer selection in quantitative glycosylation-focused studies.

Figures

Full text

1. Introduction

Proteomics has become a cornerstone of modern life sciences, offering powerful tools to characterize the dynamic landscape of protein expression, modification, and interaction networks in diverse biological systems [1]. With the continuous advancement of mass spectrometry (MS) platforms and sample preparation strategies, researchers are now able to profile proteomes with remarkable depth and precision [2]. Among the many layers of proteomic investigation, glycoproteomics has gained particular importance because N-linked glycosylation plays critical functional roles [3,4]. This post-translational modification regulates processes ranging from cell adhesion and signal transduction to immune surveillance and tumor progression [4,5]. Accurate and comprehensive characterization of proteins and their glycosylation status is therefore indispensable for advancing both basic biology and translational medicine [6,7].

Despite these technological advances, one persistent bottleneck in proteomic workflows remains the initial step of protein extraction. The lysis process directly determines the efficiency and stability of protein recovery, which in turn affects downstream peptide yield, identification depth, and quantification accuracy [8]. These challenges are particularly evident for membrane proteins, hydrophobic proteins, and N-glycoproteins, which are often poorly solubilized and consequently underrepresented in datasets. An inadequate lysis strategy can introduce substantial biases into proteome coverage, compromise reproducibility, and obscure biologically meaningful pathways. As a result, optimizing lysis buffer selection is a crucial prerequisite for reliable proteomics and glycoproteomics research [9].

Currently, several types of lysis buffers are widely employed in proteomic studies, including sodium dodecyl sulfate (SDS), guanidine hydrochloride (GuHCl), urea (UA), and Mammalian Protein Extraction Reagent (MPER). Each of these reagents exhibits distinct physicochemical properties and therefore offers specific advantages and limitations [9]. SDS, a strong ionic detergent, is well known for its ability to efficiently disrupt membranes and solubilize hydrophobic proteins, but it requires careful downstream cleanup to avoid ion suppression in MS analysis [10,11]. GuHCl and UA act primarily as denaturants, dissolving proteins by disrupting intramolecular interactions, yet they may provide incomplete recovery of complex protein mixtures [12]. MPER, in contrast, is a mild non-denaturing detergent designed to preserve native protein conformations, making it suitable for functional studies such as immunoprecipitation, though often at the expense of extraction depth [13].

Previous extensive studies have systematically compared the performance of different lysis reagents across various biological samples and proteomic workflows, revealing the profound impact of lysis buffer selection on the completeness of protein extraction, class-specific biases, and downstream quantification accuracy. For example, a comparative study of different lysis buffers, including UA, GnHCl, sodium deoxycholate (SDC), and trifluoroacetic acid (TFA), for protein extraction from HeLa cells showed that the SDC-based buffer performed best in terms of the number of identified proteins and peptides. When combined with acetone precipitation, it further improved recovery rates. In contrast, the GnHCl system exhibited lower digestion efficiency and reduced proteome coverage [10]. Furthermore, an evaluation covering suspension cells, adherent cells, primary cells, and liver tissue demonstrated that SDS provided the most stable and reproducible lysis efficiency across different sample types. That study also proposed a two-step protein extraction strategy using urea followed by SDS, which significantly increased the recovery of hard-to-lyse components such as membrane and nuclear proteins [14]. In complex samples like plasma, combining SDS or GuHCl with single-pot solid-phase-enhanced sample preparation (SP3) technology markedly increased the number of identified proteins and effectively enriched membrane proteins [11]. For bacterial samples, a composite lysis buffer containing 5% SDS, 2 M UA, and 50 mM Tris-HCl maximized the number of protein identifications (an average of 2924), notably improving protein extraction from Gram-positive bacteria as well as from low-abundance phyla such as Proteobacteria [9].

Previous studies have typically assessed these buffers under specific experimental contexts, focusing on either solubilization efficiency, peptide yield, or compatibility with downstream workflows. However, there has been a lack of comprehensive evaluation that simultaneously considers identification depth, reproducibility across replicates, quantitative stability, and functional characteristics of uniquely identified proteins and glycopeptides. This gap hinders the rational selection of lysis reagents for high-throughput proteomic and glycoproteomic applications, particularly when robust quantitative performance is required. A systematic side-by-side comparison of these commonly used buffers within a standardized workflow would therefore provide essential guidance for both experimental design and data interpretation.

To benchmark the performance of our lysis buffer across diverse proteomic contexts, we selected two well-characterized human cell lines: HeLa and HEK293T. This choice was strategic, as both lines are widely used and supported by extensive public proteomic datasets, ensuring reproducibility and direct comparability with existing studies [15,16,17]. More importantly, their distinct biological origins provide divergent proteomic backgrounds. Evaluating our method in these two models allows us to rigorously assess buffer efficacy while controlling for biological variability, thereby addressing a key methodological gap in standardized sample preparation for comparative proteomics. Based on this, we conducted a comprehensive evaluation of SDS, GuHCl, UA, and MPER within an integrated proteomic and glycoproteomic pipeline. Using HeLa cells and HEK293T cells as model systems, proteins were digested by the filter-aided sample preparation (FASP) method [18], glycopeptides were enriched by hydrophilic interaction liquid chromatography (HILIC) [19,20], and the resulting datasets were analyzed by LC-MS/MS. Comparative analyses focused on protein identification depth, reproducibility across replicates, coefficient of variation (CV) distributions, and functional enrichment of uniquely identified proteins and glycopeptides. By directly contrasting the performance of the four buffers under identical conditions, we aimed to establish a systematic benchmark, providing a practical and evidence-based reference for lysis buffer selection in proteomic and glycoproteomic studies.

2. Materials and Methods

2.1. Cell Culture and Collection

HeLa and HEK293T cell lines were obtained from Procell Life Science & Technology Co., Ltd. (Wuhan, China) and cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin at 37 °C in a humidified incubator with 5% CO_2_. When cells reached ~90% confluence, they were washed with phosphate-buffered saline (PBS), detached using trypsin, and pelleted by centrifugation (1500 rpm, 3 min). Cell pellets were washed three times with PBS and stored at −80 °C until use. All the cell culture reagents used in this study were obtained from Procell Life Science & Technology Co., Ltd. (Wuhan, China).

2.2. Protein Extraction and Quantification

Cell pellets were lysed on ice using one of four buffers: 8 M urea (Sigma-Aldrich, St. Louis, MO, USA), MPER (Thermo Fisher Scientific, Rockford, IL, USA), 6 M guanidine hydrochloride (Sigma-Aldrich, St. Louis, MO, USA), or 2% SDS (Sigma-Aldrich, St. Louis, MO, USA), each supplemented with protease inhibitors. Lysis was performed by sonication (except for MPER) under the following conditions: 25 W power, 2 s on/2 s off pulses, for 5 min. Lysates were centrifuged (16,000× g, 20 min, 4 °C), and supernatants were collected. Protein concentrations were determined using the Bradford assay [21].

2.3. Protein Digestion

Protein digestion was performed using the filter-aided sample preparation (FASP) protocol [18]. Briefly, 200 μg of protein extract was transferred into a 30 kDa ultrafiltration device (Millipore) (Merck Millipore Ltd., Tullagreen, Carrigtwohill, Co. Cork, Ireland) and centrifuged at 14,000× g for 15 min to remove residual buffer. Proteins were washed twice with 200 μL of 8 M urea (UA) solution prepared in 100 mM Tris-HCl (pH 8.0) (Sigma-Aldrich, St. Louis, MO, USA). Reduction was carried out by adding 25 mM dithiothreitol (DTT) (Sigma-Aldrich, St. Louis, MO, USA) and incubating at 37 °C for 4 h. After centrifugation, alkylation was performed with 50 mM iodoacetamide (IAA) (Sigma-Aldrich, St. Louis, MO, USA) in the dark at room temperature for 30 min, followed by centrifugation and one additional wash with 8 M UA. To quench excess IAA, 25 mM DTT was added and incubated for 15 min at room temperature in the dark, followed by centrifugation.

Subsequently, the buffer was exchanged three times with 8 M UA and then three times with 50 mM ammonium bicarbonate (NH_4_HCO_3_, pH 8.0) (Sigma-Aldrich, St. Louis, MO, USA) to remove residual detergents and denaturants. Proteins were digested on the filter with trypsin (Enzyme & Spectrum, Beijing, China) at an enzyme-to-protein ratio of 1:10 (w/w) for 16 h at 37 °C. Peptides were collected by centrifugation, followed by two additional washes with 50 mM NH_4_HCO_3_ and distilled water to maximize recovery. The combined peptide solution was dried in a SpeedVac concentrator (Eppendorf, Hamburg, Germany) at 45 °C (V-AQ mode) and stored at −80 °C until LC–MS/MS analysis. Unless otherwise noted, all centrifugation steps were performed at 14,000× g for 15 min at room temperature.

2.4. HILIC Enrichment of Glycopeptides

Intact glycopeptides were enriched using a previously established HILIC protocol [19]. Briefly, Venusil HILIC material (5 μm, 100 Å) (Agela Technologies, Tianjin, China) was activated with 0.1% TFA (three times) and equilibrated with 80% acetonitrile (ACN)/0.2% TFA (Thermo Fisher Scientific, Rockford, IL, USA) (three times). Dried peptides were dissolved in 80% ACN/0.2% TFA and incubated with pre-equilibrated HILIC material for 2 h at room temperature with gentle rotation. The mixture was transferred to a C8 membrane-packed Tip column (Jinteng Experimental Equipment Co., Ltd., Tianjin, China), and the flow-through was reapplied for a second binding step. Non-specifically bound peptides were removed by washing with 80% ACN/0.2% TFA, and glycopeptides were eluted with 0.1% aqueous TFA in three cycles (eluates pooled). The enriched glycopeptides were dried in a SpeedVac concentrator (Eppendorf) at 45 °C (V-AQ mode) and stored at −80 °C until LC-MS/MS analysis.

2.5. LC-MS/MS

All samples were analyzed using an Easy-nLC 1200 system coupled to an Orbitrap Fusion Lumos mass spectrometer both from Thermo Fisher Scientific (Rockford, IL, USA). Peptide separation was performed on an in-house packed C18 column (30 cm × 100 μm i.d., 1.9 μm ReproSil-Pur C18-AQ) (Polymicro Technologies, Phoenix, AZ, USA). Mobile phase A consisted of 0.1% formic acid (FA) in water, and mobile phase B consisted of 80% acetonitrile (ACN)/0.1% FA. Peptide concentrations were determined at 205 nm using a NanoDrop OneC spectrophotometer (Thermo Fisher Scientific, Rockford, IL, USA).

Proteomic DDA acquisition. Approximately 1 μg of peptides was loaded and separated with a 78 min linear gradient at 600 nL/min: 4–10% B (0–5 min), 10–22% B (5–48 min), 22–35% B (48–66 min), 35–90% B (66–76 min), and 90% B (76–78 min). Data were acquired in Orbitrap–Orbitrap mode. Full MS scans were acquired at 120,000 resolution across 350–1550 m/z (AGC target 4 × 10^5^; maximum injection time 50 ms). MS/MS spectra were acquired at 15,000 resolution with higher-energy collisional dissociation (HCD, 30% normalized collision energy), an isolation window of 1.6 m/z, dynamic exclusion of 18 s, and a 3 s cycle time.

Glycoproteomic DDA acquisition. For glycopeptides, 1 μg was loaded onto the same C18 column and separated with a 90 min gradient: 5–10% B (0–8 min), 10–22% B (8–68 min), 22–32% B (68–82 min), 32–90% B (82–88 min), and 90% B (88–90 min). MS data were collected in Orbitrap–Orbitrap mode. MS1 scans were acquired at 120,000 resolution over 800–2000 m/z (AGC target 4 × 10^5^; maximum injection time 100 ms). MS2 scans were acquired at 15,000 resolution (AGC target 5 × 10^4^; maximum injection time 250 ms). Precursor ions with charge states 2–6 were fragmented using stepped HCD (20%, 30%, 40%), with dynamic exclusion set to 15 s. Data acquisition was controlled by XCalibur (version 4.3).

2.6. Data Analysis

The Homo sapiens protein database was downloaded from UniProt (27 April 2025, 20,471 entries) [22] for database searching. Raw proteomic data were processed with MaxQuant (v2.0.3) [23] using the following parameters: trypsin as the protease with up to two missed cleavages; precursor mass tolerance of 20 ppm and fragment mass tolerance of 4.5 ppm; carbamidomethylation (C) as a fixed modification; methionine oxidation and N-terminal acetylation as variable modifications; and a false discovery rate (FDR) threshold of <1%.

Glycoproteomic raw files were analyzed using pGlyco3.1 [24] and Panda (v1.2.7.) [25]. For pGlyco3.1, the settings included trypsin digestion with up to three missed cleavages, carbamidomethylation (C) as a fixed modification, methionine oxidation and N-terminal acetylation as variable modifications, precursor tolerance of ±4 ppm, and FDR < 1%. Default parameters were applied for all other settings. Similarly, Panda analysis was performed with an FDR < 1%.

Quantitative and qualitative analyses of proteins, peptides, and N-glycosylation modifications were performed using Excel. Linear correlation analyses of protein identifications were conducted with Perseus (version 2.0.10.0) [26]. For the calculation of the coefficient of variation (CV) for each protein, the data were first log_2_-transformed in Perseus. Missing values were then imputed using the “Gaussian distribution-based stochastic sampling” method with the following parameters: width = 0.3 and down shift = 0.8. Subcellular localization analysis of identified proteins and glycoproteins from each group was performed using WoLF PSORT (https://wolfpsort.hgc.jp/ accessed on 23 November 2025) [27]. Data visualization was performed using GraphPad Prism (v10.1.2) and OriginPro 2024b [28].

3. Results

3.1. Experimental Design and Overall Workflow

The overall workflow for evaluating the effects of four lysis buffers (SDS, GuHCl, MPER, and UA) on proteomic and glycoproteomic profiling is illustrated in Figure 1. After thawing and culturing to 90% confluence, HeLa and HEK293T cells were evenly divided into four portions and seeded into four culture dishes, corresponding to the four lysis buffer treatment groups of SDS, UA, GuHCl, and MPER. Once the cells in each dish reached 90% confluence again, the cells from each dish were further divided into five equal portions and transferred to new culture dishes for continued growth. When all newly seeded cells again reached 90% confluence, they were collected. Thus, five biological replicates were obtained for each treatment group.

Extracted proteins were reduced, alkylated, and digested with trypsin following the FASP protocol to generate peptides. The peptide mixtures were then separated into two parallel analyses: one directly subjected to LC-MS/MS for global proteomic profiling, and the other enriched for intact glycopeptides using hydrophilic interaction liquid chromatography (HILIC) before LC-MS/MS analysis.

Mass spectrometry was performed in DDA mode with optimized scan ranges for proteomic (350–1500 m/z) and glycoproteomic (800–2000 m/z) acquisition. The proteomic datasets were searched against the UniProt human database using MaxQuant, while glycoproteomic data were analyzed with pGlyco3.1 and Panda. The results from both proteomic and glycoproteomic analyses were systematically compared to assess identification depth, reproducibility, quantitative stability, and functional enrichment, thereby enabling a comprehensive evaluation of buffer performance.

3.2. Systematic Evaluation of Lysis Buffer Performance in Proteomic and Glycoproteomic Profiling

Building on this workflow, we performed a systematic evaluation of SDS, GuHCl, MPER, and UA in terms of their effects on protein extraction efficiency, proteome identification depth, reproducibility across replicates, quantitative stability, and the functional relevance of uniquely identified proteins and glycopeptides.

3.2.1. Proteome Identification Depth

The protein identification depth of the four lysis buffers was evaluated. The results showed that SDS buffer identified the highest number of proteins in both cell lines, but the number of identified peptides varied significantly between the different cell lines. Specifically, in HeLa cells, the protein identification numbers ranked from highest to lowest as SDS, GuHCl, MPER, and UA (Figure 2A). In HEK293T cells, the order was SDS, MPER, UA, and GuHCl (Figure 2C). In terms of peptide identification, the performance of the different buffers varied by cell line. In HeLa cells, the number of peptides identified by GuHCl and SDS was similar, both higher than those identified by UA and MPER (Figure 2B). However, in HEK293T cells, SDS identified the fewest peptides, while MPER identified the most (Figure 2D).

3.2.2. Stability of Proteome Identification

The reproducibility and stability of proteomic analysis were evaluated through correlation analysis, sequence coverage, detection frequency, and coefficient of variation distribution. The reproducibility across all experimental groups was high, with mean R^2^ values ranging from 0.88 to 0.92 (Figure 3A). Performance in sequence coverage varied across cell lines and methods. In HeLa cells, SDS yielded the highest number of proteins with ≥10% coverage, followed by GuHCl, MPER, and UA. Conversely, in HEK293T cells, MPER demonstrated superior coverage, outperforming the other three methods (Figure 3B). To assess the reproducibility of protein identification, the number of proteins detected in at least three replicates in each group (i.e., quantifiable proteins) was counted. The SDS method demonstrated the best reproducibility in both cell lines. It identified 3057 quantifiable proteins (77.98% of the total) in HeLa cells, exceeding the performance of all other buffers. This advantage was maintained in HEK293T cells, where it identified 3434 quantifiable proteins (Figure 3C). Coefficient of variation (CV) analysis further underscored the quantitative stability of SDS. In HeLa cells, SDS identified more proteins with CV < 5% (2795) than other buffers. Similarly, in HEK293T cells, SDS led with 3138 such proteins, matching the performance of MPER and UA but surpassing GuHCl (Figure 3D).

At the peptide identification level, we counted the number of peptides that were successfully identified in all five replicate experiments across each group. The results showed that the SDS group identified a higher number of peptides than the UA group, but fewer than the MPER and GuHCl groups. In HEK293T cells, the peptide count for the SDS group was the lowest (Table S1).

Although different methods vary in sequence coverage, the SDS method demonstrates robust superiority in key metrics such as sample correlation, reproducibility (number of quantifiable proteins), and quantitative stability (CV values), proving its outstanding overall performance.

3.2.3. Analysis of Subcellular Localization of Proteins Identified in Each Group

This study conducted subcellular localization analysis on the proteins identified in each group. As shown in Figure 4, these proteins were primarily distributed in the cytoplasm, nucleus, mitochondria, plasma membrane, and extracellular regions. However, the subcellular compartment with the highest number of identified proteins differed between the two cell lines.

Further analysis revealed that in both cell lines, SDS lysis buffer consistently demonstrated higher identification efficiency in extracting proteins from various subcellular compartments (Figure 4). In HeLa cells, the SDS group identified 1300 cytoplasmic proteins and 1086 nuclear proteins, numbers significantly higher than those obtained with the other three lysis buffers (Figure 4A, Table S2). This advantage was similarly evident in other localization categories such as mitochondria, plasma membrane, and extracellular proteins (Figure 4A, Table S2). In contrast, the protein counts across various subcellular localizations obtained with GuHCl, MPER, and UA in HeLa cells were highly consistent (Figure 4A, Table S2). In HEK293T cells, SDS maintained its superior extraction efficiency, with higher protein identification numbers in categories such as the nucleus (1110 proteins) and cytoplasm (1050 proteins) compared to the other buffers, a trend that also held true for mitochondria, extracellular proteins, and other compartments (Figure 4B, Table S3). Notably, in this cell line, MPER and UA still yielded highly similar protein extraction results, while GuHCl consistently showed the lowest overall identification counts across all subcellular categories (Figure 4B, Table S3).

In summary, under the experimental conditions employed, SDS lysis buffer proved more efficient in extracting and identifying a larger number of proteins from different subcellular compartments. The extraction performance of MPER and UA remained quite similar, whereas the effectiveness of GuHCl appeared to vary depending on the cell line used.

3.3. Systematic Evaluation of Four Lysis Buffers on N-Glycosylation Modifications

3.3.1. Statistics of N-Glycosylation Modification Identification Results

The impact of four lysis buffers (SDS, GuHCl, MPER, and UA) on N-glycosylation profiling was systematically assessed in terms of intact N-glycopeptides, glycoproteins, N-glycans, and glycosylation sites in HeLa and HEK293T cells. Among the groups, SDS consistently achieved the highest identification depth. In HeLa cells, SDS yielded a significantly higher number of intact N-glycopeptides compared with the other three buffers (Figure 5A), along with markedly greater numbers of identified glycoproteins and glycosylation sites (Figure 5B,D), as well as broader coverage of N-glycans (Figure 5C). Similarly, in HEK293T cells, SDS showed a clear advantage, identifying substantially more intact N-glycopeptides than the remaining buffers (Figure 5A), and maintaining superior performance in glycoprotein, N-glycan, and glycosylation site identification (Figure 5B–D). The performance ranking of the four lysis buffers was consistent between the two cell lines. These findings clearly establish the superior performance of SDS in glycoproteome coverage.

Based on high-mannose (HM), fucosylated (Fuc), sialylated (Sia), sialylated-fucosylated complex types, and other glycoforms, the glycopeptides from each group were categorized into five types. Overall, glycopeptides with high-mannose modifications were identified in the largest numbers in both cell lines (Figure 5E,F). In HeLa cells, the SDS lysis buffer yielded higher numbers of identified glycopeptides than GuHCl, MPER, and UA in the four glycan modification types: HM, Sia, Fuc, and Sia+Fuc (Figure 5E, Table S4). This indicates that SDS comprehensively enhanced the identification capability for glycopeptides with various glycan modifications. Additionally, UA and MPER were highly consistent in both the total number of identified N-glycopeptides and the distribution of glycopeptide types based on glycan categories (Figure 5E, Table S4). The GuHCl group identified the lowest total number of glycopeptides among the four lysis methods, mainly due to a significantly lower count of HM-type glycan-modified glycopeptides compared to the other lysis buffers (Figure 5E). In HEK293T cells, SDS remained the leading method in the number of identified glycopeptides across all glycan types (Figure 5F, Table S5). Similarly to the observations in HeLa cells, MPER and UA continued to display similar glycan type distribution patterns (Figure 5F, Table S5). GuHCl yielded lower identification counts for glycopeptides with HM, Fuc, Sia, and Sia+Fuc modifications than the other lysis buffers (Figure 5F, Table S5). In summary, the SDS lysis buffer demonstrated the most comprehensive identification capability for glycopeptides with various glycan modifications in both cell lines. MPER and UA showed highly consistent distributions across different glycan modification types, while the GuHCl method was relatively the least effective.

3.3.2. Stability of Intact Glycopeptide Identification

The stability of glycopeptide identification was evaluated based on reproducibility across replicates and quantitative variation. To assess the reproducibility of intact N-glycopeptides identification, the number of intact N-glycopeptides detected in at least three replicates in each group (i.e., quantifiable intact N-glycopeptides) was counted. In HeLa cells, the SDS group demonstrated optimal performance in quantifying intact N-glycopeptides, showing both the highest total number of quantifiable intact N-glycopeptides and the highest proportion relative to all glycopeptides detected, significantly surpassing the GuHCl, MPER, and UA groups (Figure 6A).

Notably, even in terms of the most stringent metric of overlapping identification number (glycopeptides detected in all five experimental replicates), SDS outperformed the other three lysis buffers. Moreover, this advantage was consistently observed in both tested cell lines, indicating better reproducibility (Table S1).

With respect to quantification stability (CV < 5%), the SDS group again exhibited a clear advantage, with a substantially greater proportion of its quantifiable glycopeptides meeting this stability criterion compared to the other groups (Figure 6C). Similarly, in HEK293T cells, the SDS group also yielded the highest number of quantifiable intact N-glycopeptides, followed by the UA, MPER, and GuHCl groups (Figure 6B). Regarding quantitative stability (CV < 5%), the performance ranking of the four lysis buffers was consistent with that observed in HeLa cells (Figure 6D). Collectively, these results highlight SDS as the most robust buffer for stable and reproducible glycopeptide analysis, with UA performing moderately well and GuHCl and MPER showing relatively poor reproducibility.

3.3.3. Analysis of Subcellular Localization of Glycoproteins Identified in Each Group

Overall, the glycoproteins identified in both cell lines were predominantly enriched in the plasma membrane and extracellular space (Figure 7), which is highly consistent with the biological functions of glycoproteins, such as their primary localization on the cell surface and involvement in intercellular communication and secretion. Glycoproteins associated with the endoplasmic reticulum, the main site of glycosylation modification, were also consistently detected (Figure 7).

However, differences were observed between the cell lines and among the lysis buffers. In HeLa cells, the glycoproteins extracted by the four lysis buffers exhibited highly similar distributions across subcellular locations, with only minor variations in protein counts among categories such as plasma membrane, extracellular space, and endoplasmic reticulum (Figure 7A, Table S6). This indicates that for HeLa cells, the four lysis methods performed comparably in terms of broad glycoprotein extraction, without showing a pronounced preference for glycoproteins from specific subcellular compartments. In HEK293T cells, the choice of lysis buffer significantly influenced the glycoprotein identification profile. The SDS lysis buffer demonstrated superior extraction efficiency for glycoproteins from the plasma membrane and extracellular compartments, identifying significantly higher numbers of plasma membrane glycoproteins (187) and extracellular glycoproteins (109) compared to the other three lysis buffers (Figure 7B, Table S7). This result suggests that SDS offers a distinct advantage in solubilizing membrane-associated glycoproteins in HEK293T cells.

4. Discussion

This study systematically evaluated the performance of four commonly used lysis buffers (SDS, GuHCl, MPER, UA) in HeLa and HEK293T cells using an integrated proteomic and glycoproteomic workflow. The results demonstrate that in FASP-based sample preparation, SDS buffer exhibits significant advantages across the majority of key performance metrics. Its excellent reproducibility and quantitative stability make it the preferred lysis buffer for combined proteomic and glycoproteomic studies.

At the proteome level, SDS demonstrated the highest protein identification depth in both cell lines, along with superior reproducibility and quantitative stability. Specifically, SDS consistently identified the largest number of reproducible proteins (detected in at least three experimental replicates) in both HeLa and HEK293T cells, and showed the highest count of proteins with low coefficients of variation (CV < 5%). This finding confirms the crucial role of strong denaturants in achieving comprehensive coverage of complex proteomes, particularly in the effective extraction of hydrophobic and membrane proteins [29]. Additionally, subcellular localization analysis indicated that SDS more efficiently extracted proteins from diverse compartments, including the cytoplasm, nucleus, mitochondria, and plasma membrane, underscoring its broad solubilization capacity.

At the glycoproteomic level, SDS demonstrated decisive advantages. It significantly outperformed other methods in the number of identified intact N-glycopeptides, glycoproteins, glycan structures, and glycosylation sites. Particularly noteworthy was its exceptional reproducibility and quantitative stability in glycopeptide analysis. Moreover, SDS demonstrated a clear advantage in enriching glycoproteins localized to the plasma membrane and extracellular space, which are key compartments for glycosylated molecules involved in cell communication and signaling. This aligns with its strong membrane-disrupting properties, allowing for more effective recovery of membrane-associated and secreted glycoproteins that are often underrepresented under milder lysis conditions.

In conclusion, our comparative analysis provides clear experimental evidence regarding the performance of these four lysis buffers in an integrated proteomic and glycoproteomic workflow. Under the conditions tested, SDS delivered the deepest coverage, highest reproducibility, and best quantitative precision for both the global proteome and the N-glycoproteome in HeLa and HEK293T cells.

5. Conclusions

In summary, this systematic comparative study provides a solid evidence base for the selection of lysis buffers. SDS is established as the most reliable lysis reagent for in-depth, quantitative proteomic and glycoproteomic analyses. Meanwhile, MPER retains its value in applications requiring the preservation of native protein activity. Ultimately, the choice of lysis strategy should align with specific research objectives. This work not only offers practical guidance for methodological optimization but also underscores the profound impact of the initial sample preparation step on final data quality and biological insight depth, providing important direction for advancing robust and reproducible omics research. Although HeLa and HEK293T cells are widely used in proteomics method development, they may not fully represent other biological contexts, including suspension cells, tissue environments, and bacterial samples. Therefore, we will expand the range of sample types for more in-depth research in the future.