A streamlined technique for in-depth ubiquitomics analysis of formalin-fixed paraffin-embedded tissues

Abstract

Formalin-fixed paraffin-embedded (FFPE) tissues offer advantages over fresh frozen (FrFr) tissues in terms of preserving tissue integrity, reducing long-term storage costs, and increasing sample availability, thus providing an invaluable resource for biomarker discovery and exploration of disease mechanisms. As a reversible and highly versatile post-translational modification (PTM), ubiquitination is involved in protein homeostasis and the regulation of almost all physiological and pathological processes, but ubiquitomics studies of FFPE tissues are lacking due to technical challenges. In this study, we proposed an optimized, streamlined technique for extracting GlyGly proteins from FFPE tissues, including improvements in the components of lysis buffer, the duration of heating for tissue sections and the strategy for protein digestion. The application of this strategy to clinical samples revealed that 7000–10 000 GlyGly sites could be identified from several FFPE tissue sections of hepatocellular carcinoma (HCC). Finally, we showed that the storage duration had a minor effect on the stability of ubiquitomics information within 7 years, while phosphoproteomics information remained stable for more than 10 years. Our strategy proves that FFPE tissues can serve as a valuable alternative to FrFr tissues for comprehensive ubiquitomics investigations.

---

Introduction

Ubiquitination regulates protein homeostasis and is linked with most pathophysiological processes.^1–7 A system-wide investigation of the ubiquitomics of clinical tissue samples facilitates a deeper understanding of the regulatory mechanisms underlying a multitude of diseases and assists in the identification of potential biomarkers and drug targets.

Formalin-fixed, paraffin-embedded (FFPE) tissues are the gold standard for preserving clinical samples for retrospective pathological diagnosis.^8,9 A substantial number of FFPE biospecimens have been collected in hospital repositories worldwide. These samples, along with high-quality patient follow-up data, offer invaluable resources for comprehensive studies based on clinical samples.^10,11 In recent years, several efforts have focused on the proteomics and PTMs of FFPE tissues.^12–18 The primary challenge in analyzing the proteomics of FFPE tissues is the crosslinking reactions that occur among biological molecules due to the presence of formalin.^19–21 These reactions hinder the effective and reproducible extraction of proteins and their subsequent identification.²² Different approaches, including heat treatment, the use of extraction solvents with highly denaturing performance, or high concentrations of primary amine-containing buffers, are typically employed to increase protein solubility and prevent protein crosslinking.^19,23–29 While significant insights have been gained from explorations of protein glycosylation, phosphorylation and other modifications in FFPE samples, ubiquitomics analysis of FFPE tissues has never been attempted. Additionally, the stability of ubiquitinated proteins during the storage of FFPE tissues remains poorly understood. To simplify the following text, ubiquitinated proteins, peptides and sites are referred to as GlyGly proteins, peptides and sites, respectively.

All seven lysine (K) residues of ubiquitin can react with formalin, which may lead to the formation of more crosslinks on GlyGly proteins than on non-GlyGly proteins and increase the difficulty of extracting GlyGly proteins. Furthermore, the additional modifications in proteins from FFPE tissues prevent trypsin or other endopeptidases from reaching their active cleavage sites;^21,27 thus, the K-ε-GG antibody recognition sites cannot be exposed, and the efficacy of GlyGly peptide enrichment may be diminished, which may also hinder ubiquitomics analyses. Collectively, these challenges hinder comprehensive ubiquitomics analyses of FFPE tissues. There is an unmet need to develop effective sample pretreatment approaches to reverse protein crosslinking and ensure highly efficient extraction and subsequent analysis of GlyGly proteins.³⁰

The objective of this study was to establish a method to extract, digest and profile ubiquitomics effectively from FFPE tissues. To this end, seven protocols using different lysis buffers and digestion methods were compared for FFPE tissue sample treatment. Furthermore, the impacts of other variables, such as the incorporation of the reducing agent dithiothreitol (DTT) in the lysis buffers and the duration of heat treatment, on the depth of coverage of the ubiquitomics were assessed. Since FFPE samples are generally stored for long periods at room temperature, which may affect protein stability, we further evaluated the impact of storage duration on the stability of ubiquitomics and phosphoproteomics of FFPE samples.^31–33

Experimental procedures

Preparation of mouse liver samples

After the liver tissue was removed from each mouse, half of the tissue were immediately frozen and stored at −80 °C. The remaining half of the liver tissue was immersed in 4% paraformaldehyde fixative, and then the tissue blocks were dehydrated with 75%, 80%, 90%, 95%, and 100% ethanol sequentially. The tissue blocks were then immersed in liquid paraffin wax for three hours, replaced with fresh liquid paraffin wax, and immersed for another hour to obtain the FFPE tissues of the mouse liver. The FFPE blocks were sectioned into 10 µm thick slices with a paraffin slicer (Leica, HistoCore AUTOCUT 230 V AC 50 Hz 10DVA), dried, and stored at room temperature for subsequent experiments. This study followed the 3R principles and was approved by the ethics committee of the Beijing Institute of Lifeomics (approval number: NCPSB-20250313-46MT).

Clinical sample collection

HCC FFPE tissues were collected at the National Cancer Centre, and all human FFPE tissues were obtained according to the approved protocol. FFPE tissues were sectioned into 10 µm sections with a paraffin slicer (Craftek, CR-601ST) and collected in tubes. This study was approved by the Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, approval number: NCC2019XC-007). Detailed information of all the patients is provided in supplemental Table S1.

Protein extraction from FFPE tissue sections

For the FFPE tissues fixed on glass slides, the sections were immersed in xylene for 30 min, followed by sequential immersion in 100%, 85%, 70%, and 50% ethanol for 5 min each. The tissues were then scraped into tubes filled with water and centrifuged (14 000×g, 10 min), and the supernatant was discarded to obtain dewaxed and rehydrated tissues. After adding 500 µL of lysis buffer (2% SDS/100 mM Tris–HCl, pH 8.0/50 mM DTT), the tissues were lysed by sonication (50 W, 1s on, 2s off, and total sonication time of 12 min) and then heated in a metal bath. The tissues were then sonicated again according to the above parameters. Finally, the protein concentration was determined using the Bradford assay (Thermo Scientific, USA).

For the FFPE tissues collected in the tubes, 1 mL of xylene was added to the tubes, which were subsequently incubated for 15 min at RT and 7 rpm. The mixture was subsequently centrifuged, the supernatant was discarded, and the procedure was repeated once. After 1 mL of 100%, 85%, 70%, or 50% ethanol or cold water was added, the supernatant was discarded to obtain dewaxed or rehydrated mouse liver tissues, and the remaining procedure was the same as above.

Protein extraction from frozen mouse liver tissue

The mouse liver tissues were weighed in 1.5 mL tubes to an approximate wet weight of 50 mg and then placed on ice. To wash the tissue, 1 mL of PBS was added to the tube, and the tissue was cut into small cubes. After centrifugation at 4 °C, the supernatant was discarded. Five hundred microliters of lysis buffer (1% SDC/100 mM Tris–HCl, pH 8.0) was added, and the tissue was lysed by sonication. Finally, the protein concentration was quantified via a BCA assay (Thermo Scientific, USA).

Protein digestion: FASP digestion

Two hundred to 500 µg of protein was diluted with wash buffer (8 M urea (UA)/100 mM Tris–HCl, pH 8.0) on a 30 kDa filter (Merck Millipore) and replaced with 200 µL of wash buffer twice. Protein alkylation was achieved by incubation for 4 h in alkylation buffer (20 mM DTT/8 M UA/100 mM Tris–HCl, pH 8.0) and then incubation for 30 min in 50 mM iodoacetamide (IAA)/8 M UA/100 mM Tris–HCl. The samples were purified using wash buffer and further purified via 50 mM ABC three times. The proteins were subsequently digested overnight at 37 °C via trypsin (1 : 50 w/w; Promega). The peptides were eluted via centrifugation, and 200 µL of 50 mM ABC and 100 µL of water were used for further washing.

GlyGly peptide enrichment

The lyophilized peptide samples were redissolved in immunoaffinity purification (IAP) buffer (50 mM MOPS pH 7.2/10 mM Na₂HPO₄/50 mM NaCl), and the peptide concentration was estimated using a Nanodrop. The enrichment of GlyGly peptides was performed following a previously reported protocol.³⁴ In brief, K-ε-GG antibody beads (PTM Scan® Ubiquitin Remnant Motif (K-ε-GG) Kit (Cell Signaling Technology, USA)) were washed with 1 mL of PBS four times. Peptide solution (200–300 µg) was added to the beads, the total volume was adjusted to 1 mL with IAP, and the mixture was incubated for 2 hours at 4 °C and 7 rpm. The beads were washed twice with cold IAP and three times with cold water. The supernatants were removed, and the modified peptides were eluted with 55 µL and 50 µL of 0.15% TFA for 10 minutes, respectively. The eluates were desalted in C₁₈ tips (3M, USA) and lyophilized.

Peptide sequencing by LC–MS

Peptides were redissolved in mobile phase A (0.1% formic acid (FA, Sigma-Aldrich, USA)), loaded onto a trapping column (150 µm × 10 mm, ReproSil-Pur C₁₈-AQ, 3 µm; Dr Maisch, GmbH, Germany) and then separated on an analytical column (150 µm × 150 mm, ReproSil-Pur C₁₈-AQ, 1.9 µm; Dr Maisch, GmbH, Germany) at a flow rate of 600 nL min⁻¹.

The 16 HCC samples used for the systematic assessment of the effect of FFPE tissue storage duration on proteomics stability were analyzed using an Ultimate 3000 RSLCnano (Thermo Fisher Scientific) in tandem with a Q Exactive HF-X (Thermo Fisher Scientific) in data-independent acquisition (DIA) mode. The gradients used for proteomics and ubiquitomics were as follows: 10–15 min, 6–10% mobile phase B (0.1% FA, 99.9% ACN); 15–70 min, 10–30% mobile phase B; 70–80 min, 30–40% mobile phase B; 80–80.1 min, 40–95% mobile phase B; 80.1–85 min, 95–95% mobile phase B; 85–85.1 min, 95–6% mobile phase B; and 85.1–88 min, 6–6% mobile phase B. The mass spectrometry parameters for proteomics were as follows: MS1 scans were performed in the m/z range 300–1400 with a resolution of 120 000 and an AGC target of 3 × 10⁶. The maximum injection time was 60 ms. MS/MS scans of the spectra had a resolution of 30 000 with an AGC target of 3 × 10⁶ and a maximum injection time of auto. The mass spectrometry parameters used for ubiquitomics were as follows: parameters for MS1 were the same as those used for proteomics. The resolution of the MS/MS scanned spectrum was 30 000, the AGC target was 3 × 10⁶, and the maximum injection time was set to auto. For the window information, please refer to Table S2.

All other samples were analyzed via DIA of peptides using an Easy-nLC 1200 (Thermo Fisher Scientific) in tandem with a Q Exactive HF (Thermo Fisher Scientific) system. The parameters were the same for proteomics and ubiquitomics, with a liquid phase gradient of 0–6 min, 5–12% mobile phase B (0.1% FA, 99.9% ACN), 6–56 min, 12–30% mobile phase B, 56–68 min, 30–40% mobile phase B, 68–70 min, 40–95% mobile phase B, 70–78 min, 95–95% mobile phase B. MS parameters were as follows: MS1 scanning m/z range of 400–1200, resolution of 120 000, AGC target of 3 × 10⁶, and maximum injection time of 80 ms. MS/MS scans of the spectra were performed with a resolution of 30 000, an AGC target of 3 × 10⁶, and a maximum injection time of 41 ms. All raw data were processed using Spectronaut (version 19.0) software.

Results

Evaluation and comparison of different solutions for extraction of GlyGly proteins from FFPE mouse liver tissues

Highly efficient and unbiased protein extraction and digestion are pivotal in proteomics research. However, the use of formalin when preparing FFPE tissues results in the formation of crosslinked proteins, which in turn reduces protein solubility and impedes protein extraction and digestion. Therefore, selecting the most appropriate protein lysis buffers and digestion methods for proteomics analysis of FFPE tissues is highly important. Due to the unique structure of ubiquitin, existing proteomics extraction and digestion methods for FFPE tissues may not be suitable for ubiquitomics studies. Therefore, we compared several commonly used protocols in proteomics profiling studies to identify the most appropriate one for ubiquitomics research. Here, we used mouse liver FFPE tissues for proteomics and ubiquitomics analyses to evaluate the efficacy of seven published protein extraction and digestion protocols and two technical replicates were performed for each method. These protocols, including RapiGest–ISD,²⁶ UA–ISD,³⁰ SDS–SP3,³³ TFE–SP3,³⁵ SDC–ISD,¹⁸ TFE–ISD and SDS–FASP,^35,36 encompass a range of different lysis buffers and protein digestion methods (Table 1). Comparisons were made among these different protocols in terms of protein and GlyGly protein extraction efficiency, protein digestion efficiency, and the numbers of proteins and GlyGly sites identified to determine the optimal lysis buffer and digestion method for ubiquitomics. Finally, a streamlined technique for proteomics and ubiquitomics analysis of FFPE tissues was established. The results revealed that the highest protein and peptide coverage was reached when SDS was used as the lysis buffer (Fig. 1A). Furthermore, the protein digestion method had no effect on proteomics identification when SDS was utilized as the lysis buffer. The missed cleavage rate for FASP was 13.7%, which was lower than the 16.7% observed for SP3. Additionally, the numbers of proteins and peptides identified with FASP were marginally greater than those identified with SP3 (Fig. 1A).

Table 1 Details of the seven protocols for protein extraction and digestion from FFPE tissues

Method	Buffer components	Method of digestion	GlyGly enrichment peptide input
RapiGest–ISD	0.2% RapiGest, 20 mM DTT, 50 mM ABC	ISD	200 µg
UA–ISD	50% ACN, 100 mM ABC+8 M UA, 100 mM ABC	ISD	200 µg
SDS–SP3	2% SDS, 100 mM Tris–HCl, pH 8.0	SP3	200 µg
TFE–SP3	50% TFE, 300 mM Tris–HCl, pH 8.0	SP3	200 µg
SDC–ISD	2% SDC, 100 mM Tris–HCl, pH 8.0	ISD	200 µg
TFE–ISD	50% TFE, 300 mM Tris–HCl, pH 8.0	ISD	200 µg
SDS–FASP	2% SDS, 100 mM Tris–HCl, pH 8.0	FASP	200 µg

---

Fig. 1 Comparison of the seven protocols for protein extraction and digestion from FFPE tissues (n = 2). (A) Number of proteins identified. (B) Ratio of the number of peptides with a C-terminal K residue to the number of peptides with a C-terminal R residue. (C) Distribution of hydrophobicity scores of the peptides identified. (D) Number of GlyGly sites identified from peptides generated by different protocols. (E) Pearson correlation analyses between 2 replicates for the three protocols with the top performances.

With the SDS-FASP protocol, the C-terminal K/R ratio (lysine to arginine in the C-terminus of the peptides) of the identified peptides in FFPE tissue was similar to that in FrFr tissue (0.98 vs. 1.06) (Fig. 1B, and S1A), indicating that this method could effectively release protein crosslinking in FFPE tissues. The distributions of the hydrophobicity scores for the peptides identified using these protocols were similar, indicating that there was no significant difference in peptide hydrophobicity (Fig. 1C). GlyGly peptide enrichment experiments were subsequently conducted at the peptide level. Significantly more GlyGly peptides and GlyGly sites (over 4600) were identified via SDS–FASP than via the other protocols (Fig. 1D). The SDS–FASP protocol also exhibited optimal reproducibility (Pearson r = 0.9344) (Fig. 1E). In conclusion, the SDS–FASP protocol demonstrated optimal performance for both proteomics and ubiquitomics analyses when it was applied to FFPE tissues.

Finalization of the streamlined workflow for extraction of GlyGly proteins from FFPE mouse liver tissues

The efficiency of protein extraction, denaturation and decrosslinking are affected by the components and working conditions of the extraction buffer. Treating FFPE tissues with a compound that can reduce disulfide bonds can increase protein recovery.¹⁴ Therefore, we sought to determine whether the incorporation of disulfide-reducing compounds (DTT, 50 mM) into lysis buffer would improve the extraction efficiency of GlyGly proteins. Additionally, studies have shown that the formation of formalin adducts is reversible under high-temperature or high-pressure conditions.^21,37 Thus we also evaluated the effects of different heating durations on the efficiency of decrosslinking and the stability of the proteins and GlyGly proteins. All experiments for proteomics and ubiquitomics analyses described above were conducted using mouse liver FFPE tissues and included two technical replicates each. The results revealed that GlyGly site identification was improved by approximately 25% in the samples treated with lysis buffer containing DTT (DTT+) compared with the samples treated in the absence of DTT (DTT−) (Fig. 2A), indicating that incorporating reducing reagents in the lysis buffer was advantageous for ubiquitomics analysis. Mouse liver FFPE tissues of the same size were used for the heating duration experiments, and the results revealed that the quantity of peptide recovered increased significantly when the heating duration was extended from 90 min to 150 min (Fig. 2B). However, the numbers of proteins and peptides identified did not notably change as the heating duration varied (Fig. 2C). The ubiquitomics analysis results indicated that the numbers of identified GlyGly peptides and GlyGly sites increased with increasing heating duration, peaked at 120 min (Fig. 2D), and began to decrease after heating for 150 min (∼8%), suggesting that GlyGly proteins may remain stable when heated at 99 °C for 120 min, but extending the heating duration further may result in breakage of the K-ε-GG bond.

Fig. 2 Optimization of the FFPE tissue sample pretreatment conditions (n = 2). (A) Number of GlyGly sites identified under (DTT+) versus (DTT−) conditions. (B) Quantity of peptides recovered after different heating durations. (C) Numbers of proteins and peptides identified after different heating durations. (D) Number of GlyGly sites identified after different heating durations.

In conclusion, adding DTT to lysis buffer increases GlyGly site identification by approximately 25%, and heating for 120 min facilitates in-depth proteomics and ubiquitomics analyses while maintaining a high peptide recovery rate. It is therefore recommended that the SDS–FASP protocol be modified by adding 50 mM DTT to the lysis buffer and subsequently heating for 120 min to guarantee sufficient release of GlyGly proteins.

Comparison of ubiquitomics between fresh frozen (FrFr) and FFPE tissues

We demonstrated that the optimized, streamlined technique allowed sensitive ubiquitomics analysis of FFPE tissues. However, it remains unclear whether the ubiquitomics information in FFPE tissues could duplicate the identification results obtained from FrFr tissues. We thus employed FFPE and FrFr tissues derived from the same mouse liver source for proteomics and ubiquitomics analyses, and included two technical replicates each. The hydrophobicity of the peptides from the FFPE tissue obtained using this workflow was highly consistent with that of the peptides from the FrFr tissue (Fig. 3A). There was no significant difference between them in terms of protein identification and GlyGly peptide identification (Fig. S1B and C), and proteins identified from FFPE tissue overlapped 93% with those identified from FrFr tissue (Fig. 3B), which is greater than the proportion reported in other studies (∼65%).²² Furthermore, the FrFr tissue and FFPE tissue exhibited an overlap of GlyGly peptides of 67% (Fig. 3C), in contrast to the 95–99% overlap for the FrFr replicates and 97–99% overlap for the FFPE replicates, indicating that fixation with formalin affected the ubiquitomics of the tissue samples. Finally, we performed a cellular compositional analysis of GlyGly proteins identified in FFPE and FrFr tissues, which revealed no significant differences between FFPE and FrFr tissues (Fig. 3D). In summary, formalin fixation had a minor effect on proteomics, whereas more varieties could be introduced at the ubiquitomics level.

Fig. 3 Ubiquitomics landscape of mouse liver FFPE and FrFr tissues and FFPE tissues from clinical HCC samples. (A) Distribution of the hydrophobicity scores for the proteins obtained from mouse liver FFPE tissue and FrFr tissue. (B) Overlap of proteins identified from mouse liver FFPE tissue and FrFr tissue. (C) Overlap of GlyGly peptides identified from mouse liver FFPE tissue and FrFr tissue. (D) Number of proteins corresponding to specified keywords in cellular components. (E) Numbers of proteins and peptides identified in the clinical HCC samples. (F) Numbers of GlyGly peptides and GlyGly sites identified in the clinical HCC samples.

Ubiquitomics of clinical FFPE samples

The results on the mouse liver FFPE tissues demonstrated that our workflow could obtain ubiquitomics information comparable to that of FrFr tissues. Subsequent efforts were dedicated to evaluating the clinical applicability of this method using FFPE tissue sections from hepatocellular carcinoma (HCC) patients. Over 4600 proteins and 56 000 peptides were identified (Fig. 3E). Two hundred micrograms of peptide material was used for the GlyGly peptide enrichment experiments, resulting in the identification of 7000–10 000 GlyGly sites (Fig. 3F). In conclusion, this streamlined technique allows for the consistent and reliable capture of proteomics and ubiquitomics information from FFPE tissues and shows excellent clinical applicability and reproducibility.

Whether the storage duration affects the stability of PTMs in FFPE tissue

Susceptibility to long durations of storage at RT and rich clinical information pertaining to disease progression or treatment response are significant advantages of FFPE tissues.³⁸ Prior studies have indicated that proteins remain stable in FFPE adenocarcinoma tissues after 6–9 years of storage, whereas peptide identification from FFPE tissues that were stored for 14–20 years was attenuated by 16%.³⁵ We aimed to evaluate the stability of GlyGly proteins in FFPE tissues with increasing storage duration. The experiments were conducted on 10 µm thick sections of FFPE tissues obtained from 16 HCC patients. Samples were collected across 4 time points (2012, 2017, 2019, and 2021), with 4 patients enrolled at each time point. The results revealed no significant differences in the intensities of the proteomics among the 16 samples. A total of 6400 proteins were identified and quantified from the 16 HCC samples, with more than 6000 proteins identified in each sample. Furthermore, the number of proteins identified did not change with increasing storage duration (Fig. 4A), and there is a high degree of overlap (>90%) in the identified proteins across all 4 groups (Fig. 4B). Principal component analysis (PCA) was subsequently performed on the quantified protein data obtained from the 16 samples. No significant grouping of samples from different years was observed (Fig. 4E), indicating that the proteomics information of the stored FFPE tissues did not significantly differ as a function of storage duration.

Fig. 4 Relationships between the storage duration and protein stability in FFPE tissues. (A) Plot comparing the proteins identified from the 4 groups of samples with different ages. (B) Upset plot of proteins identified from the 4 groups. (C) Plot comparing the GlyGly sites identified from the 4 groups of samples with different ages. (D) Upset plot of GlyGly sites identified from the 4 groups. (E) PCA of the quantitative protein information from the 16 HCC samples. (F) PCA of the quantitative GlyGly site information from the 16 HCC samples. (G) Clustering analysis of the quantitative GlyGly sites from the 16 HCC samples.

A total of 9284 GlyGly sites were identified and quantified from the 16 HCC samples. Furthermore, there was a notable decrease in the number of GlyGly sites identified from the 2012 samples compared with those identified from the other years (Fig. 4C). Analysis of the GlyGly sites identified from the 4 groups of samples revealed that 96% of the GlyGly sites identified in the 2012 samples were also identified in the other three groups of samples (Fig. 4D). PCA was performed using the quantified GlyGly site information, and the analysis revealed that the samples from 2012 were significantly different from those from the other years (Fig. 4F), but no significant differences were found among the samples from the other years. Subsequently, we selected the top 500 GlyGly sites with the most significant differences for clustering analysis. The results showed that the 2012 samples formed a distinct cluster (Fig. 4G). Subsequently, we performed differential analysis on the GlyGly sites jointly identified in the 2012 group and the 2021 group (with the shortest storage duration). The results demonstrate that the expression of KLHL22_K260_M1 and RIGI_K115_M1 in 2012 was downregulated notably, while the expression of AP2S1_K45_M1 was upregulated, indicating that prolonged storage (>7 years) may affect the stability of GlyGly modifications. In conclusion, the results demonstrate that ubiquitomics information can be retained in FFPE tissues for at least 7 years (e.g., those from 2017, 2019, and 2021) without significant changes and that GlyGly proteins remain stable. Upon extending the storage period to 12 years, a decrease in the number of the GlyGly protein identifications was observed.

To further understand the varied stabilities of other types of PTMs in FFPE tissues, we leveraged the peptide solution collected after proteomics and ubiquitomics heating-duration experiments to evaluate the effects of different heating durations on the efficiency of decrosslinking and the stability of phosphorylated proteins. We found that the number of phosphosites identified decreased with increasing thermal incubation time ranging through 60–150 min (Fig. 5A and B). In addition, we also leveraged the peptide solution collected after proteomics and ubiquitomics analyses of the 16 HCC samples to perform phosphopeptide enrichment. Our data demonstrated that phosphoproteomics information in FFPE tissue samples stored for up to 12 years did not change with storage duration. In conclusion, our experimental results show that a thermal incubation duration of 60 min is most conducive to the analysis of the phosphoproteomics of FFPE tissues and that the phosphorylated protein information remains stable when FFPE tissues are stored for up to 12 years (Fig. 5C and D).

Fig. 5 Stability assessment of phosphorylated proteins. (A) Number of phosphopeptides identified after different heating durations. (B) Number of phosphosites identified after different heating durations. (C) Plot comparing the phosphosites identified from the 4 groups of samples with different storage durations. (D) PCA of the quantitative phosphosite information from the 16 HCC samples.

Discussion

Previously, proteomics and PTMomics (glycosylation, phosphorylation, and acetylation) studies have been successfully conducted on FFPE tissues. This study presents the first ubiquitomics investigation of FFPE tissues. An experimental comparison revealed that the SDS–FASP protocol was the most effective for extracting GlyGly proteins from FFPE tissues. This may be due to the pronounced dissolving properties of SDS.³⁹ TFE is another lysis buffer commonly used for proteomics analysis of FFPE tissue, although it may result in residual tissue precipitation.³⁵ The number of GlyGly sites identified by the SDS–FASP protocol was significantly greater than that identified by SDS–SP3, which may be attributed to the superior cleavage efficacy of FASP in comparison with that of SP3, particularly with large sample inputs exceeding 20 µg.⁴⁰ Furthermore, SP3 may lead to the enrichment of hydrophilic proteins and may result in the loss of some hydrophobic proteins.⁴¹

It has been demonstrated that extending the heating duration at 95 °C from 60 to 120 min results in a slight increase in total lysed peptide identification but a slight decrease in phosphopeptide identification.³³ In this study, we assessed the stabilization of GlyGly proteins after heating. The total number of identified GlyGly peptides slightly increased when the heating duration was extended from 90 min to 120 min. However, the number of GlyGly peptides identified slightly decreased after 150 min of heat treatment. These findings suggest that 120 min is the optimal heating duration for the ubiquitomics analysis of FFPE tissues.

Our study revealed that the proteins in FFPE tissues stored for up to 12 years remained unaltered. However, the ubiquitomics data remained stable in tissues stored for up to 7 years. Thus, storing FFPE tissue for up to 7 years is recommended for ubiquitomics analysis because excessive storage durations may result in the loss of ubiquitomics information and a reduction in the reliability of the experimental data.

The stability of phosphorylated proteins was not in concordance with that of GlyGly proteins, which underlines that the stability of different types of modifications varies in FFPE tissue and that sample pretreatment protocols need to be optimized according to the properties of focused PTMomics.

Author contributions

Shifang Liang: conceptualization, methodology, formal analysis, visualization, writing-original draft; Bo Zheng: conceptualization, methodology investigation; Dongying Huang: methodology, data curation; Jian Zhang: methodology; Haiyang Li: methodology; Yingyi Zhao: methodology; Xiaoxia Gao: conceptualization, supervision, writing-review & editing; Ting Xiao: conceptualization, supervision, writing-review & editing; Wantao Ying: conceptualization, methodology validation, funding acquisition, writing-original draft, writing-review & editing.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

All data are available in the article or the supplemental information (SI). The MS data of this study have been deposited in the ProteomeXchange Consortium repository with the dataset identifier PXD061954.

Supplementary information is available. See DOI: https://doi.org/10.1039/d5ay01727k.

Acknowledgements

This work was supported by the National Key Research and Development Project of China (2021YFA1300201), and the Open Project Program of the State Key Laboratory of Medical Proteomics (SKLP-O202410). The authors acknowledge the mass spectrometry facility and the functional technology platform facility of the National Center for Protein Sciences (Beijing).

References

1. J. Park, J. Cho and E. J. Song, Arch. Pharmacal Res., 2020, 43, 1144–1161.
2. F. M. Hansen, M. C. Tanzer, F. Brüning, I. Bludau, C. Stafford, B. A. Schulman, M. S. Robles, O. Karayel and M. Mann, Nat. Commun., 2021, 12, 254.
3. D. A. Cruz Walma, Z. Chen, A. N. Bullock and K. M. Yamada, Nat. Rev. Mol. Cell Biol., 2022, 23, 350–367.
4. D. B. Beck, A. Werner, D. L. Kastner and I. Aksentijevich, Nat. Rev. Rheumatol., 2022, 18, 435–447.
5. D. Chirnomas, K. R. Hornberger and C. M. Crews, Nat. Rev. Clin. Oncol., 2023, 20, 265–278.
6. D. Popovic, D. Vucic and I. Dikic, Nat. Med., 2014, 20, 1242–1253.
7. L. Deng, T. Meng, L. Chen, W. Wei and P. Wang, Signal Transduction Targeted Ther., 2020, 5, 11.
8. Y. Zhu, T. Weiss, Q. Zhang, R. Sun, B. Wang, X. Yi, Z. Wu, H. Gao, X. Cai, G. Ruan, T. Zhu, C. Xu, S. Lou, X. Yu, L. Gillet, P. Blattmann, K. Saba, C. D. Fankhauser, M. B. Schmid, D. Rutishauser, J. Ljubicic, A. Christiansen, C. Fritz, N. J. Rupp, C. Poyet, E. Rushing, M. Weller, P. Roth, E. Haralambieva, S. Hofer, C. Chen, W. Jochum, X. Gao, X. Teng, L. Chen, Q. Zhong, P. J. Wild, R. Aebersold and T. Guo, Mol. Oncol., 2019, 13, 2305–2328.
9. H. Chen, Y. Zhang, H. Zhou, W. Chen, J. Peng, Y. Feng, L. Fan, J. Li, J. Zi, Y. Ren, Q. Li and S. Liu, Anal. Chem., 2023, 95, 16733–16743.
10. S. Amatori and M. Fanelli, Int. J. Mol. Sci., 2022, 23, 1103.
11. Z. W. Lai, J. Weisser, L. Nilse, F. Costa, E. Keller, M. Tholen, J. N. Kizhakkedathu, M. Biniossek, P. Bronsert and O. Schilling, Mol. Cell. Proteomics, 2016, 15, 2203–2213.
12. G. P. Pujari, K. K. Mangalaparthi, B. J. Madden, F. A. Bhat, M. C. Charlesworth, A. J. French, G. Sachdeva, E. Daviso, U. Thomann, P. McCarthy, S. Vasantgadkar, D. Bhattacharyya and A. Pandey, J. Am. Soc. Mass Spectrom., 2023, 34, 1225–1229.
13. P. D. Piehowski, V. A. Petyuk, R. L. Sontag, M. A. Gritsenko, K. K. Weitz, T. L. Fillmore, J. Moon, H. Makhlouf, R. F. Chuaqui, E. S. Boja, H. Rodriguez, J. S. H. Lee, R. D. Smith, D. M. Carrick, T. Liu and K. D. Rodland, Clin. Proteomics, 2018, 15, 26.
14. O. Azimzadeh, Z. Barjaktarovic, M. Aubele, J. Calzada-Wack, H. Sarioglu, M. J. Atkinson and S. Tapio, J. Proteome Res., 2010, 9, 4710–4720.
15. P. Ostasiewicz, D. F. Zielinska, M. Mann and J. R. Wiśniewski, J. Proteome Res., 2010, 9, 3688–3700.
16. D. M. Carrick, M. G. Mehaffey, M. C. Sachs, S. Altekruse, C. Camalier, R. Chuaqui, W. Cozen, B. Das, B. Y. Hernandez, C. J. Lih, C. F. Lynch, H. Makhlouf, P. McGregor, L. M. McShane, J. Phillips Rohan, W. D. Walsh, P. M. Williams, E. M. Gillanders, L. E. Mechanic and S. D. Schully, PLoS One, 2015, 10, e0127353.
17. R. Longuespée, D. Baiwir, G. Mazzucchelli, N. Smargiasso and E. De Pauw, Methods Mol. Biol., 2018, 1723, 19–31.
18. M. Wakabayashi, H. Yoshihara, T. Masuda, M. Tsukahara, N. Sugiyama and Y. Ishihama, J. Proteome Res., 2014, 13, 915–924.
19. K. Buczak, J. M. Kirkpatrick, F. Truckenmueller, D. Santinha, L. Ferreira, S. Roessler, S. Singer, M. Beck and A. Ori, Nat. Protoc., 2020, 15, 2956–2979.
20. Y. Kawashima, Y. Kodera, A. Singh, M. Matsumoto and H. Matsumoto, Clin. Proteomics, 2014, 11, 4.
21. S. Magdeldin and T. Yamamoto, Proteomics, 2012, 12, 1045–1058.
22. Y. Zhang, M. Muller, B. Xu, Y. Yoshida, O. Horlacher, F. Nikitin, S. Garessus, S. Magdeldin, N. Kinoshita, H. Fujinaka, E. Yaoita, M. Hasegawa, F. Lisacek and T. Yamamoto, Proteomics, 2015, 15, 2568–2579.
23. V. Broeckx, K. Boonen, L. Pringels, X. Sagaert, H. Prenen, B. Landuyt, L. Schoofs and E. Maes, Mol. BioSyst., 2016, 12, 553–565.
24. G. D. Barnabas, V. Goebeler, J. Tsui, J. W. Bush and P. F. Lange, Anal. Chem., 2023, 95, 3291–3299.
25. G. Mitsa, Q. Guo, C. Goncalves, S. E. J. Preston, V. Lacasse, A. Aguilar-Mahecha, N. Benlimame, M. Basik, A. Spatz, G. Batist, W. H. Miller Jr, S. V. Del Rincon, R. P. Zahedi and C. H. Borchers, Int. J. Mol. Sci., 2022, 23, 4443.
26. M. C. Föll, M. Fahrner, V. O. Oria, M. Kühs, M. L. Biniossek, M. Werner, P. Bronsert and O. Schilling, Clin. Proteomics, 2018, 15, 11.
27. C. B. Fowler, R. E. Cunningham, T. J. Waybright, J. Blonder, T. D. Veenstra, T. J. O'Leary and J. T. Mason, Lab. Invest., 2008, 88, 185–195.
28. T. Handa, H. Sasaki, M. Takao, M. Tano and Y. Uchida, Fluids Barriers CNS, 2022, 19, 56.
29. N. J. Nirmalan, P. Harnden, P. J. Selby and R. E. Banks, J. Pathol., 2009, 217, 497–506.
30. K. Weke, S. Kote, J. Faktor, S. Al Shboul, N. Uwugiaren, P. M. Brennan, D. R. Goodlett, T. R. Hupp and I. Dapic, Anal. Chim. Acta, 2022, 1204, 339695.
31. R. W. Sprung, Jr., J. W. Brock, J. P. Tanksley, M. Li, M. K. Washington, R. J. Slebos and D. C. Liebler, Mol. Cell. Proteomics, 2009, 8, 1988–1998.
32. S. M. Thompson, R. A. Craven, N. J. Nirmalan, P. Harnden, P. J. Selby and R. E. Banks, Proteomics:Clin. Appl., 2013, 7, 241–251.
33. C. Friedrich, S. Schallenberg, M. Kirchner, M. Ziehm, S. Niquet, M. Haji, C. Beier, J. Neudecker, F. Klauschen and P. Mertins, Nat. Commun., 2021, 12, 3576.
34. D. Huang, Y. Leng, X. Zhang, M. Xing, W. Ying and X. Gao, J. Proteomics, 2023, 288, 104993.
35. F. Coscia, S. Doll, J. M. Bech, L. Schweizer, A. Mund, E. Lengyel, J. Lindebjerg, G. I. Madsen, J. M. Moreira and M. Mann, J. Pathol., 2020, 251, 100–112.
36. M. W. Ni, L. Wang, W. Chen, H. Z. Mou, J. Zhou and Z. G. Zheng, Rapid Commun. Mass Spectrom., 2017, 31, 171–178.
37. C. B. Fowler, R. E. Cunningham, T. J. O'Leary and J. T. Mason, Lab. Invest., 2007, 87, 836–846.
38. R. W. Sprung, M. A. Martinez, K. L. Carpenter, A. J. Ham, M. K. Washington, C. L. Arteaga, M. E. Sanders and D. C. Liebler, J. Proteome Res., 2012, 11, 3498–3505.
39. Y. Q. Yu, M. Gilar, P. J. Lee, E. S. Bouvier and J. C. Gebler, Anal. Chem., 2003, 75, 6023–6028.
40. M. Sielaff, J. Kuharev, T. Bohn, J. Hahlbrock, T. Bopp, S. Tenzer and U. Distler, J. Proteome Res., 2017, 16, 4060–4072.
41. J. M. Conforti, A. M. Ziegler, C. S. Worth, A. M. Nambiar, J. T. Bailey, J. H. Taube and E. S. Gallagher, J. Proteome Res., 2024, 23, 3877–3889.

---

Footnote

† These authors contributed equally to this work.

---