O-GlcNAcylation of Keratin 18 coordinates TCA cycle to promote cholangiocarcinoma progression

doi:10.21203/rs.3.rs-3325874/v1

Download PDF

Article

O-GlcNAcylation of Keratin 18 coordinates TCA cycle to promote cholangiocarcinoma progression

https://doi.org/10.21203/rs.3.rs-3325874/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Glycosylation in human cholangiocarcinoma (CCA) actively contributes to pathophysiological steps of tumor progression. Of note is the dynamic modification of proteins by O-linked β-N-acetyl-glucosamine (O-GlcNAcylation) that modulates various tumor-associated biological activities. By using a cutting-edge chemical proteomic methodology for intact glycopeptide analysis, we show herein that O-GlcNAcylation of Keratin 18 (K18) coordinates the tricarboxylic acid (TCA) cycle enzymes, namely isocitrate dehydrogenases (IDHs), to promote CCA progression. Mechanistically, site-specific O-GlcNAcylation of K18 on Ser 30 stabilizes K18, which benefits the expression of cell cycle checkpoints to enhance cell cycle progression and cell growth. Interaction with IDHs down-regulates the level of citrate and isocitrate, while up-regulates the level of α-ketoglutarate (α-KG). Our study thus expands the current understanding of protein O-GlcNAcylation, and adds another dimension of complexity to post-translational control over metabolism and tumorigenesis.

Cholangiocarcinoma (CCA), also known as bile duct cancer, constitutes a constellation of malignancies emerging in the biliary tree. CCA is the second most common primary hepatic malignancy after hepatocellular carcinoma (HCC), accounting for approximately 15% of all primary liver tumors, and its incidence is increasing worldwide.¹ CCA is typically asymptomatic in the early stages and difficult to cure at late stages, which highly compromises therapeutic options and leads to a dismal prognosis.^2,3 The discoveries of early diagnostic markers are, therefore, of great significance for improving CCA diagnosis and therapeutics.²

A growing appreciation for the physiological roles of glycans, either in glycoconjugates (e.g., glycoproteins, glycolipids, proteoglycans) or in their free form, have resulted in increasing attention in their functional elucidation.^4,5 Of note, β-O-linked N-acetylglucosamine (GlcNAc) is a dynamic glycosylation attached to serine and threonine of nucleocytoplasmic and mitochondrial proteins. This modification is dynamically concerted by a pair of enzymes, O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA).⁶O-GlcNAcylation has been identified in more than 5,000 human proteins and exhibits complex crosstalk with protein phosphorylation.^7,8 Accumulating evidence has shown that O-GlcNAcylation appears to modulate various biological activities (e.g., epigenetics, transcription, cellular metabolism) in response to environmental cues, and alternation in O-GlcNAcylation underlies neurodegenerative diseases, diabetes and various cancers.^9–11 Aberrant O-GlcNAcylation actively contributes to pathophysiological steps of tumor progression, by regulating cancer cell signaling, proliferation, invasion, angiogenesis, metastasis, cell-matrix interaction and immune modulation both in vitro and in vivo.^5,12 However, the mechanistic elucidation by which protein-specific O-GlcNAcylation contributes to CCA metabolic reprogramming and tumorigenesis remains largely elusive. Herein we present the systematic perception of O-GlcNAcylation, OGT and OGA in human CCA samples, and then perturbed the CCA cells with a metabolically-based intact glycopeptide analysis strategy, to profile global O-GlcNAcylation with glycan composition and glycosylation site resolution. We also provide mechanistic insight as to how Keratin 18 O-GlcNAcylation affects and choreographs the tricarboxylic acid (TCA) cycle in CCA progression.

O -GlcNAcylation is up-regulated in CCA

We started by investigating the clinical relevance of O-GlcNAcylation in CCA. We first analyzed the O-GlcNAc, and the expression of OGT and OGA in 21 pairs of human resected tumor tissues and adjacent normal bile duct using immunohistochemistry (IHC) (Fig. 1A). The levels of O-GlcNAc and OGT were distinctively elevated in tumor tissues compared with normal tissues, consistent with the IHC staining scores in these CCA patients (P < 0.001, Fig. 1B). However, the level of OGA did not show distinctive changes between the two groups (Fig. 1A, 1B). We then analyzed the O-GlcNAc expression levels in additional 15 peritumoral/tumor tissue pairs from CCA patients. Again, O-GlcNAcylation and OGT were identified to be up-regulated in tumor tissues (Fig. 1C and Fig. S1), and quantitative analysis of the signals confirmed that the increases are statistically significant (P < 0.01 and 0.05, respectively by Student’s t-tests, two-tailed) (Fig. 1D). Notably, mRNA expression of OGT and OGA was escalated from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) datasets (GSE32879, GSE107943, GSE119336, GSE76297) (Fig. S2A, 2B). CCA patients with high levels of OGT and low levels of OGA gene expression displayed worse overall survival (OS) according to the Kaplan-Meier survival analysis, indicating the pivotal role of O-GlcNAcylation in CCA (Fig. S2C, 2D). We further assessed the O-GlcNAc, OGT and OGA expressions in three human CCA cell lines, namely HCCC-9810, RBE and HuCCT1, and the Human Intrahepatic Biliary Epithelial Cell (HIBEpiC) control cell line (Fig. 1E). O-GlcNAcylation level increased in all CCA cells compared with HIBEpiC cells, while mRNA expression of OGT and OGA, as measured by quantitative real-time PCR (qRT-PCR), exhibited varied expression levels (Fig. 1F). These data altogether pinpointed the up-regulation of O-GlcNAcylation as a common event in CCA samples.

Global O -GlcNAcylation contributes to CCA cell proliferation

Previous reports by the Wongkham group indicated the significance of O-GlcNAcylation in controlling the metastatic ability of CCA cells via nuclear translocation of NF-κB and heterogeneous nuclear ribonucleoprotein-K (hnRNP-K).^13–15 We therefore postulated that modulation of O-GlcNAc levels would alter CCA oncology phenotypes (i.e., proliferation, apoptosis, cell cycle), by targeting pathways known to regulate CCA progression. To do so, we first used chemical tools to inhibit OGT and OGA. Ac₄5SGlcNAc (designated as 5S thereof) is a known inhibitor of OGT that acts as a metabolic precursor to form uridine diphosphate activated precursor-5SGlcNAc (UDP-5SGlcNAc), and OGA can be selectively and effectively inhibited by Thiamet-G (designated as TMG thereof) (Fig. 2A).^16,17 Treatment of all three CCA cell lines with 5S or TMG led to the robust decrease or increase of O-GlcNAc modification level on proteins, in a time- and dose-dependent manner (Fig. 2B, Fig. S3A, S4A and S5), in accordance with previous observations. At the same time, cellular viability was assessed using a cell counting kit-8 (CCK-8). The overall cytotoxicity in HuCCT1 cells was governed by the inhibition of OGT rather than OGA, suggesting that lowering O-GlcNAcylation will induce cell death (Fig. 2C). Similar effects were also observed in RBE and HCCC-9810 cells in a dose-dependent manner (Fig. S3B, S4B). Moreover, suppression of OGT promoted CCA cell apoptosis, as evidenced by the increased percentage of both early apoptosis (FITC⁺/PE⁻) and late apoptosis (FITC⁺/PE⁺) after OGT silencing upon 5S incubation (Fig. 2D, Fig. S3C). No remarkable effects of TMG as an anti-apoptotic inhibitor were detected (Fig. 2D, Fig. S3C). Mounting reports implied that O-GlcNAc plays a multifaceted role during the cell cycle, and incongruousness arises partly due to different physiological cues.¹⁸ We examined the cell cycle distribution of HuCCT1 and RBE cells using flow cytometric analysis after incubation with 5S or TMG at varied concentrations (Fig. 2E, Fig. S3D). Interestingly, cell cycle distribution of HuCCT1 cells was majorly found in the G2/M phase, while RBE cells were arrested in the S phase when treated with 5S, but not under TMG-treated scenarios (Fig. 2E, Fig. S3D). It is therefore sufficient to conclude that an appropriate O-GlcNAc level is critical for all cell cycle phases. Additionally, we blotted the classical biomarkers for cell cycle and apoptosis after perturbation of O-GlcNAcylation using 5S and TMG, respectively (Fig. 2F, 2J, Fig. S3E). There were no notable differences in these biomarkers for the TMG-treated group, yet the 5S-treated group presented global up-regulation on apoptosis markers and down-regulation on the cell cycle checkpoints (Fig. 2F, 2J, Fig. S3E). To further elucidate the correlation between O-GlcNAcylation and cancer progression, we complemented three independent small interfering RNAs (siRNAs) for OGT or OGA knockdown (Fig. 2G, Fig. S3F), all of which showed potent knockdown efficacy with the corresponding O-GlcNAc processing enzymes (i.e., OGT or OGA), meanwhile maintaining the desired catalytic activities (Fig. 2H, Fig. S3G). Clonogenic assay both qualitatively and quantitatively validated that knockdown of OGT in HuCCT1 and RBE cells clearly reduced cell proliferation and presented an anti-tumor effect (Fig. 2I, Fig. S3H). These biological characterizations reciprocally integrate with chemical tools, emphasizing that global O-GlcNAcylation may promote the development and progression of cholangiocarcinoma.

Chemical enrichment and profiling of intact O -GlcNAcylated glycopeptides in CCA

Since imbalanced O-GlcNAcylation impacts the process of cancer progression in CCA, we next explored to systematically enrich, identify and profile intact O-GlcNAcylated glycopeptides with matched information on glycosylation sites and glycan compositions. We adopted Click-iG, a comprehensive platform that amalgamates metabolic oligosaccharide engineering (MOE) of selected, clickable unnatural sugar probes for O-GlcNAcylated protein enrichment, and a customized pGlyco3 search engine for intact glycopeptide annotation.^19,20 In brief, azidosugars were metabolically incorporated into various glycans. The azido-containing glycoproteins were reacted with a three-module alkyne-photocleavable linker-biotin tag (alkyne-PC-biotin), digested with trypsin, and enriched using streptavidin beads. After photocleavage with 365 nm ultraviolet (UV), the released click-labeled glycopeptides were subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis, with preferred glycopeptide fragmentation strategies, namely, stepped collision energy based higher-energy collisional dissociation (sceHCD) followed by product-dependent electron transfer/higher-energy dissociation (sceHCD-pd-EThcD) (Fig. 3A). 1,6-di-O-propionyl-N-azidoacetylgalactosamine (1,6-Pr₂GalNAz), an optimized monosaccharide O-GlcNAc chemical reporter with minimal nonspecific S-glyco-modification and cytotoxicity, was employed in the experiment (Fig. S6).²¹ We first evaluated the metabolic efficacy of 1,6-Pr₂GalNAz in the CCA and HIBEpiC cell lines and observed azidosugar incorporation in a dose- and time-dependent manner (Fig. 3B, Fig. S7A). Significant fluorescence labeling at nucleocytoplasmic regions was achieved when 1,6-Pr₂GalNAz-treated cells were permeabilized and reacted with alkyne-AZDye-488, via the ligand-assisted copper (I)-catalyzed azide-alkyne cycloaddition (CuAAC), in agreement with the localization of O-GlcNAc glycosylation (Fig. 3C, Fig. S7B). Similar quantitative data on all four cell lines were observed by flow cytometry (Fig. 3D). In view of these results, HuCCT1 and HIBEpiC cells incubated with 200 µM 1,6-Pr₂GalNAz for 48 h were used as the standardized condition for glycoproteomic analysis.

By using the established streamlined MS data analysis and annotation procedure (Fig. S8), we then mapped a total of 1170 O-GlcNAc sites on 368 O-GlcNAcylated proteins from three replicate experiments using the Click-iG workflow (Fig. 3A, 3E, Fig. S9A, S10A). In combination, the Click-iG strategy yielded pan-scale intact glycopeptide information (i.e., intact glycosites, glycosites, glycoproteins) on HuCCT1 and HIBEpiC cells (Fig. S10B). For instance, a total of 1872 intact glycosites were identified in HuCCT1 cells, of which 1094 overlapped with those identified in HIBEpiC cells, indicating the cell-type specific feature of protein glycosylation (Fig. S10B, left panel). By glycan classification, we identified a total of 3164 intact glycosites, which also consisted of 1421 intact N-linked glycosites and 573 intact mucin-type O-linked glycosites (Fig. S10C). Sequence visualized using a probabilistic approach around the identified glycosites revealed the typical sequon/motif for O-GlcNAcylation, mucin-type O-linked glycosylation and N-linked glycosylation (Fig. S9B, S10D). Remarkably, the Click-iG also provided in-depth glycan type and composition information from intact glycosites (Fig. S11A, S11B, S12A and S12B). Gene Ontology (GO) analysis showed that the identified O-GlcNAcylated proteins concentrated in the nucleocytoplasmic region (Fig. S9C). In addition, proteins carrying 23 up-regulated (fold change > 1.50, P < 0.05) and 36 down-regulated (fold change < 0.67, P < 0.05) O-GlcNAc sites in HuCCT1 cells were potently enriched using Click-iG (Fig. 3F, Table S1, S2), of which many protein regulators involved in the cell cycle and growth, as well as transcriptional processes, were identified to be O-GlcNAcylated (Fig. 3G, Fig. S13). These results, collectively, demonstrate that Click-iG enables simultaneous and comprehensive profiling of O-GlcNAcylation (and other types of glycosylation) in the CCA and HIBEpiC cell lines, with intact glycosite level resolution.

Keratin 18 is mainly O -GlcNAcylated at Ser 30

Keratins play a pivotal role in differentiation and tissue specialization and are highly regulated among various epithelia in a cell-specific manner.²² We noticed that Keratin 18 (K18) is on the list of up-regulated O-GlcNAcylated proteins in HuCCT1 cells compared to HIBEpiC control cells (Fig. 3F, Table S1, S2). Keratin 18 plays various roles in intracellular scaffolding and cellular processes and is highly associated with malignant phenotype in digestive epithelia.^23,24 Post-translational modification (PTM) including sumoylation,²⁵ acetylation/methylation,²⁶O-GlcNAcylation²⁷ and its reciprocal cross-talk with phosphorylation²⁸ on K18 have been reported in previous works. To confirm the O-GlcNAcylation of K18, we incubated HuCCT1 and HIBEpiC cells with 1,6-Pr₂GalNAz for 48 h, reacted with alkyne-biotin using CuAAC, and conducted pull-down procedures with streptavidin beads. Immunoblotting with anti-K18 demonstrated that K18 was modified with azidosugars (Fig. 4A, left panel). To further verify the labeling results, we treated both cell lysates with a permissive β-1,4-galactosyltransferase mutant (Y289L GalT1), which transfers N-azidoacetylgalactosamine (GalNAz) from its uridine diphosphate activated precursor (UDP-GalNAz) to O-GlcNAc residues.²⁹ Subsequent chemoselective click reaction with alkyne-biotin followed by streptavidin enrichment proved O-GlcNAcylation occurred on endogenous K18 (Fig. 4A, right panel). The extent of O-GlcNAc modification on K18 in HuCCT1 cells was semi-quantitatively measured by labeling azides with alkynylated polyethylene glycol 5000 (alkyne-PEG_5KD) in a mass shift assay. The stoichiometric ratio for O-GlcNAcylation, whether in O-β-N-azidoacetylglucosamine (O-β-GlcNAz) or in O-β-GlcNAc-β-1,4-GalNAz form, was approximately 50% (Fig. 4B). Based on Click-iG, we mapped four O-GlcNAc modification sites (Ser 15, Ser 18, Ser 30 and Ser 31) on K18, of which Ser 30 and Ser 31 were previously reported. Interestingly, O-GlcNAcylation on Ser 49 was previously reported yet not found in our experiment setting.²⁷ Of note, all four sites identified are well conserved in Homo sapiens, Mus musculus and Rattus norvegicus. (Fig. S14A, S14B). To further determine the relative abundance of O-GlcNAc modification on these sites, the FLAG-tagged K18 mutants, K18^S15A, K18^S18A, K18^S15/18A, K18^S30A, K18^S31A, K18^S30/31A and K18^{S15/18/30/31A} (K18^4A), were transfected into HEK293T cells. The incorporated azides, either via MOE with 1,6-Pr₂GalNAz, or Y289L GalT1 chemoenzymatic labeling, were reacted with alkyne-biotin for streptavidin capture and immunoblot analysis. Ser 30 appears as the major O-GlcNAcylation site in K18 (Fig. 4C, 4D), in line with our chemical proteomic analysis results (Fig. 3F). Alternatively, we pulled down with anti-FLAG beads for K18 enrichment, and then measured the O-GlcNAc level for FLAG-K18^WT or FLAG-K18^S30A, using biotinylated signals introduced via the above-mentioned glyco-analytical methods. A significant loss of O-GlcNAcylation for FLAG-K18^S30A compared with FLAG-K18^WT was observed (Fig. 4E and 4F). In addition, sceHCD-pd-EThcD based LC-MS/MS analysis annotated the K18 peptide with amino acid 28–45 (PVSSAASVYAGAGGSGSR) as an O-GlcNAcylated peptide at Ser 30 (Fig. 4G).

O -GlcNAcylation of K18 promotes CCA proliferation and progression in vitro and in vivo

With the detailed glycosylation information for K18 at hand, we next asked whether K18 O-GlcNAc modification would impact cholangiocarcinoma phenotype(s). We first examined the K18 O-GlcNAcylation levels among typical human CCA cell lines and the normal HIBEpiC cell line. Elevated K18 O-GlcNAcylation was observed in all three CCA cell lines but not in HIBEpiC cells (Fig. 5A). To scrutinize the importance of Ser 30 O-GlcNAcylation, we generated stable CCA cell lines with three independent targeting-resistant short hairpin RNA (shRNA) for K18 knockdown (Fig. S15A, S15B, S16A and S17A), and then recovered K18 expression using either FLAG-K18-WT or FLAG-K18-S30A (Fig. 5B, Fig. S15C, S16B and S17B). A systematic evaluation of KRT18 mRNA expression and knockdown efficiency led to shK18-2 as the optimal construct (designated as shK18 thereof). CCK-8 and colony formation assays indicated that depletion of K18 gravely inhibited cell proliferation in HuCCT1 cells, which was rescued by re-expression of FLAG-K18-WT rather than FLAG-K18-S30A (Fig. 5C, 5D). Similar results were also observed in RBE and HCCC-9810 cells (Fig. S16C, S16D, S17C and S17D). These data imply that Ser 30 O-GlcNAcylation is functionally crucial in CCA proliferation in vitro. We next assessed the cell cycle distribution for each rescue cell line and found that FLAG-K18-S30A-rescued HuCCT1, RBE and HCCC-9810 cells showed an increase in S and G2/M phase arrest, partially in conformity with a previous report that O-GlcNAc on K18 is known to increase upon G2/M phase arrest (Fig. 5E, Fig. S16E, S17E).³⁰ Furthermore, immunoblot analysis on cell cycle biomarkers for these rescued cell lines displayed a similar correlation (Fig. 5F, Fig. S16F, S17F).

Given that O-GlcNAcylation on K18 was reported to regulate its solubility, filament organization, and stability,³¹ we tested whether the stability of K18 is modulated by O-GlcNAcylation at Ser 30. Previous reports from the Rajiv lab exhibited that the half-life for K18 is regulated by O-GlcNAcylation in human hepatocytes (Chang) cell line.^28,31 The CCA cell lines were treated with cycloheximide (CHX) to block protein synthesis for an immunoblot chase assay. OGT inhibition by 5S considerably accelerated the K18 degradation, while silencing of OGA by TMG poses minimal effect on its rate of decay (Fig. 5G, Fig. S16G, S17G). Likewise, FLAG-K18-WT was more stable than FLAG-K18-S30A after 8 h of protein lifespan (Fig. 5H, Fig. S16H, S17H). Correlatively, an increased level of ubiquitination was observed in FLAG-K18-S30A (Fig. 5I, Fig. S16I, S17I). These findings indicate that K18 stability was enhanced through up-regulation of its O-GlcNAcylation at Ser 30, as well as the inhibition of the corresponding ubiquitination.

To decipher the effect of K18 O-GlcNAcylation on tumor growth in vivo, we injected BALB/c nude mice with shRNA negative control with an empty vector (shNC + Mock), shK18, shK18 + WT, shK18 + S30A stable HuCCT1 cell lines, and quantitatively measured tumor formation. K18 depletion and its mutant at Ser 30 greatly repressed tumor growth rate, tumor size/weight, and the Ki67 positive percentage (a marker to determine cancer cell proliferation) (Fig. 5J-M, Fig. S18A, S18B). In parallel, tumor tissues dissected from HuCCT1 cells expressing FLAG-K18-WT showed a higher level of K18 O-GlcNAcylation than the corresponding cells expressing FLAG-K18-S30A (Fig. 5N). Luckily, we observed that both the protein expression level and the O-GlcNAcylation level for K18 were significantly enhanced when advancing toward the clinical CCA tumor tissues compared with adjacent normal tissues (Fig. 5O, Fig. S19). These results are able to recapitulate most of the effects observed in vitro and solidify the hypothesis that O-GlcNAcylation of K18 promotes CCA progression in vivo.

K18 O -GlcNAcylation promotes K18-Isocitrate dehydrogenase interaction to regulate the TCA cycle in CCA.

O-GlcNAcylation serves as a nutrient rheostat in a myriad of physiological contexts, especially in metabolically active organs such as the liver. The tight interconnection between O-GlcNAcylation and glucose metabolism prompted us to explore the mechanistic insight as to how K18 O-GlcNAcylation affects its interaction.³² We transfected HuCCT1 shK18 stable cell line with FLAG-K18^WT or FLAG-K18^S30A with equivalent protein expression level, co-immunoprecipitated the K18-interacting proteins with anti-FLAG beads and subjected to LC-MS/MS analysis (Fig. S20A, S20B). We identified 858 up-regulated (fold change > 1.50, P < 0.05) interacting proteins in FLAG-K18^WT HuCCT1 cells compared to FLAG-K18^S30A cells (Fig. 6A), the majority of which were closely related to the TCA cycle shown by the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (Fig. 6B). To further elucidate the K18 binding partners in the TCA cycle, we analyzed 8 out of 13 up-regulated TCA enzymes, by immunoblotting with their corresponding antibodies (Table S3). We noticed an apparent signal decrease in isocitrate dehydrogenase (IDH) families, including IDH2, IDH3A, IDH3B and IDH3G, when O-GlcNAc modification at Ser 30 of K18 was functionally amputated (Fig. 6C). Similar results were observed in RBE-rescued cells (Fig. S21). The IDH enzymes are responsible for the oxidative decarboxylation of isocitrate. To further corroborate these results, we performed relative quantification of cellular metabolites using LC-MS/MS analysis. We found that in the absence of O-GlcNAcylation at Ser 30 on K18, the rescued cells accumulated higher levels of citrate, aconitate and isocitrate, but lower levels of pyruvate, α-ketoglutarate, succinate, fumarate and malate, compared to K18 wild type rescued cells (Fig. 6D). These data suggest a strong correlation between K18 O-GlcNAcylation and IDH catalytic activities. In toto, we conclude that in cholangiocarcinoma, up-regulated OGT level will inevitably enhance the K18 O-GlcNAcylation, which increases its interaction with isocitrate dehydrogenases in the TCA cycle, one of the major metabolic pathways to regulate glucose metabolism, to coordinate the cell proliferation and CCA progression (Fig. 6E).

The glycosylation landscape of CCA, and how it affects CCA diagnosis and prognosis, has recently emerged as an appealing research field.³³ Glyco-biomarkers such as CA19-9 and CEA are up-regulated in CCA and exhibit diagnostic value.³⁴ Glyco-analytical methodologies including lectin blot,³⁵ glycan microarray³⁶ and glycoproteomic analysis^37,38 have been utilized to probe the glycobiology in CCA. In addition, abnormal sialylation,³⁹ fucosylation⁴⁰ and O-linked N-acetylgalactosamine (GalNAc) modification (O-GalNAcylation)⁴¹ have also been implicated in prognostic outcome in human cholangiocarcinoma. Changes in O-GlcNAc levels and the cycling enzyme pair OGT/OGA in a myriad of tumor types have also been reviewed.^33,42 However, a thorough investigation on how O-GlcNAcylation affects CCA progression, based on chemical glycoproteomic analysis has never been previously reported, to the best of our knowledge. In this study, we utilized a chemical glycoproteomic strategy Click-iG, to mechanistically investigate the molecular basis of O-GlcNAcylation in CCA, as exemplified by K18 O-GlcNAcylation at Ser 30. Alternative CCA glyco-biomarker candidates are also included within the intact glycopeptide analysis data (Table S1, S2) that await validation. It is expected to enhance our current knowledge of CCA progression and may reveal novel targets for this highly lethal malignancy.

We conduct the systematic analysis of O-GlcNAcylation, OGT and OGA in cholangiocarcinoma from clinical samples, and demonstrate a universal increase in O-GlcNAcylated proteins, in correlation with OGT up-regulation. Contrarily, we do not observe a down-regulation of OGA protein level in CCA as previously reported,¹³ further illustrating the individual variability and complexity of clinical samples. Hence, we choose the CCA and HIBEpiC cell lines, which offer an easy-to-manipulate, cost-effective and stable platform. The roles of O-GlcNAcylation on metastasis, migration and invasion abilities in CCA cells were previously emphasized.^14,15 Thus, we focus on the biological influence of O-GlcNAcylation on cell apoptosis, cell cycle and proliferation in the CCA cells. OGT and OGA are essential in mammals, the mutation in OGT and/or OGA requires conditional alleles, and studies with regard to these proteins are majorly restricted in isolated mutant cell strains.^43,44 We selected 5S and TMG as the chemical tools for perturbing OGT and OGA because these compounds avoid some of the problems associated with genetic models.⁴⁵ In addition, global O-GlcNAcylation is also regulated using siRNA against OGT and/or OGA mRNA to guarantee and compensate for the chemical regulation via mRNA suppression. Overall, these findings strongly suggest that a high level of O-GlcNAcylation supports progressive phenotypes of CCA cells in several aspects.

For the Click-iG strategy, 1,6-Pr₂GalNAz can cross cell membranes, and undergo deacetylation by non-specific esterases to generate cell-active GalNAz, which can be metabolically converted to UDP-GalNAz with high efficiency in cells via the GalNAc salvage pathway.⁴⁶ The nicotinamide adenine dinucleotide (NAD)-dependent epimerase UDP-galactose-4-epimerase (GALE) converts UDP-GalNAz to UDP-GlcNAz that serves as the substrate for O-GlcNAcylation by the Leloir-type glycosyltransferase OGT. However, cautions are needed in glycoproteomic data interpretation, because the resulting UDP-GalNAz and UDP-GlcNAz are general nucleotide sugar donors, and can be readily incorporated into N-linked-, mucin-type O-linked-, and O-GlcNAcylated-glycoproteins. Indeed, the precise annotation of glycan structures is still a major challenge in intact glycopeptide analysis, so we manually annotate the O-GlcNAcylation site with the aid of glycoprotein subcellular location and the presence of N-(4-aminomethyl)-triazolylacetylglucosamine (GlcNAt).¹⁹ Fortunately, Click-iG enables comprehensive coverage of the protein glycosylation landscape, which provides a blueprint for interrogating the crosstalk between different glycosylation pathways. Other O-GlcNAc chemical reporters such as peracetylated azido- or alkynyl-modified GalNAc or GlcNAc counterparts (i.e., Ac₄GalNAz, Ac₄GalNAlk, Ac₄GlcNAz, Ac₃6AzGlcNAc) are possible candidates for Click-iG application, although non-specific S-glyco-modification should be taken into consideration.^47,48

Keratins have long been recognized as diagnostic markers and prognostic indicators in a variety of epithelial malignancies. For instance, Keratin 19 has been identified as the CCA proliferative marker for postoperative intrahepatic cholangiocarcinoma patients and is of use in distinguishing cholangiocarcinoma from hepatocellular carcinomas.⁴⁹ Our chemical glycoproteomic analysis linked Keratin 18 to its O-GlcNAcylation with detailed glycosite information. Nevertheless, how O-GlcNAcylation of K18 underscores the cellular proliferation and tumor growth has been incongruous.⁵⁰ A previous report from the Rajiv lab indicated that O-GlcNAcylation decreases K18 stability and elevation of O-GlcNAcylation promotes ubiquitinoylation in human hepatocytes,³¹ in contradict with our findings in this study. We verify that the removal of O-GlcNAc modification at Ser 30 of K18 negatively affected CCA cell proliferation and xenograft tumor growth, adding a novel mechanistic insight into the regulation of K18, and highlighting the potential to intervene in K18 O-GlcNAcylation as a therapeutic strategy against CCA. Doubtless, site-specific phosphorylation of K18 is well characterized for its biological roles.⁵¹ The reciprocal interplay between O-GlcNAcylation and phosphorylation is also a critical event in modulating protein-protein interaction, subcellular localization and protein degradation.^52–54 Competitive site blocking between O-GlcNAcylation and phosphorylation (e.g., Ser 49), as well as their synergism (e.g., Ser 31 O-GlcNAcylation and Ser 34 phosphorylation), poise functional modulation in K18 solubility, filament organization, and stability.^28,55 It will be interesting to investigate whether K18 O-GlcNAcylation has crosstalk with phosphorylation or other PTM.

The cycling of O-GlcNAc on nucleocytoplasmic proteins is tightly associated with glucose metabolism and nutrient flux, which is often dysregulated and reprogrammed in cancer.^56,57 A recent report emphasized the down-regulation proteins in the glycolysis/gluconeogenesis pathway of large-duct type intrahepatic cholangiocarcinoma.⁵⁸ Here we identified a previously unknown mechanism for the regulation of the TCA cycle through O-GlcNAc modification on K18, which interacts with TCA enzymes, namely IDH2, IDH3A, IDH3B and IDH3G. Deregulation of IDHs catalytic activities is associated with cholangiocarcinoma and other human diseases. The heterozygous gain-of-function mutation in IDHs, which catalyzes the oxidation of isocitrate to α-ketoglutarate (α-KG), generates oncometabolite 2-hydroxyglutarate (2-HG), by NADPH reduction of α-KG. 2-HG competitively inhibits the α-KG-dependent dioxygenases that catalyze the transformation from α-KG to succinate, resulting in dysregulation of epigenetic and gene expression profiles, apoptosis resistance and increased migration and invasion. Therefore, IDHs are considered as proto-oncogenes in cancer metabolic derangement.^56,59 Given the dependence of CCA and other cancer cells on O-GlcNAc, targeting OGT and other components of the TCA cycle may reduce the aggressiveness of CCA cells while sensitizing them to chemotherapeutic drugs.

Cellular metabolism of 1,6-Pr₂GalNAz inevitably alters the endogenous level of UDP-GlcNAc and UDP-GalNAc, therefore, it is not ruled out that UDP-GlcNAz or UDP-GalNAz may interfere with the catalytic activity of glycosyltransferases in CCA cell lines. Due to the accessibility of bioorthogonal reaction and the dynamic feature of O-GlcNAc modification, it is still challenging to bypass the abundance bias and generate MS spectra for glycosite identification, with precise annotation. As O-GlcNAcylation occurs on thousands of protein substrates, targeting OGT and OGA using either chemical or biological methodologies might compromise normal biological processes. In addition, other functional consequences of K18 O-GlcNAcylation at other sites still remain ambiguous, because only a limited number of tools exist to study its site-specific O-GlcNAcylation effects.⁶⁰ Further preclinical and clinical studies are needed to confirm the trueness of CCA-associated O-GlcNAcylated proteins as diagnosis/prognosis indicators. The integration of glycomics and other “omics” such as genomics or transcriptomics in CCA cell lines or tissues from patients, will provide an avenue for greater impact on developing therapeutics. Additional studies are required to address these questions in the future.

In summary, we find that O-GlcNAcylation is dysregulated in human cholangiocarcinoma, majorly attributed to the up-regulation of OGT. Global up-regulation of O-GlcNAcylation benefits the expression of cell cycle checkpoints such as BUB1, FOXM1 and cMyc, enhancing cell cycle progression and cell growth. To elucidate the function of O-GlcNAcylation in CCA progression, we used chemical glycoproteomic analysis Click-iG for systematic profiling of intact glycopeptides in CCA and HIBEpiC cell lines. O-GlcNAcylation of K18 at Ser 30 increases its stability, while the orchestration between glycosylated K18 and IDH proteins in the TCA cycle reveals a positive regulation of CCA progression. How knowledge of O-GlcNAcylation has led to the aberrant state of cell proliferation and tumorigenesis in cholangiocarcinoma will be interesting to pursue.

Patient samples and statement

Human CCA tumor tissues and adjacent normal tissues were collected from patients undergoing surgery at the Drum Tower Hospital Affiliated to the Medical School of Nanjing University (Nanjing, China). Written consent was obtained from all patients, and all experiments in this study were conducted in accordance with official guidelines (Declaration of Helsinki), approved by the Medical Ethics Review Committee of Nanjing Drum Tower Hospital (Nanjing, China).

Cell lines and cell culture

Human CCA cells HCCC-9810 and RBE, and HEK293T cells were purchased from the Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China. Human CCA cells HuCCT1 were purchased from the Japanese Collection of Research Bioresources Cell Bank (JCRB, Osaka, Japan). HIBEpiC cells were purchased from Shanghai Zhong Qiao Xin Zhou Biotechnology (Shanghai, China). All cell lines were identified using short tandem repeat profiling, and tested negative for mycoplasma contamination. HCCC-9810, RBE, and HuCCT1 cells were cultured in RPMI 1640 medium (Gibco, CA, USA). HEK293T cells were cultured in DMEM medium (Gibco, CA, USA). HIBEpiC cells were cultured in Epithelial Cell Medium supplementing with 1% Epithelial Cell Growth Supplement (EpiCGS), which was purchased from Shanghai Zhong Qiao Xin Zhou Biotechnology (Shanghai, China). All cell lines were cultured with 10% fetal bovine serum (FBS, Gibco, USA), and 100 U/mL of penicillin and streptomycin (Gibco, USA) in a cell incubator at 37 ℃ with 5% CO₂. Cells were transfected with siRNAs or plasmid in this study using jetPRIME® transfection reagent (Polyplus) when they reached approximately 80% confluence.

Reagents

Dimethyl sulfoxide (DMSO, Cat# D2650), CuSO₄ (Cat# 931071), sodium ascorbate (Cat# A7631), methanol (Cat# 439193), urea (Cat# U5378), ammonium bicarbonate (ABC, Cat# A6141), dithiothreitol (DTT, Cat# D9779), iodoacetamide (IAA, Cat# I1149), N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid (HEPES, Cat# H4034), MnCl₂ (Cat# 429449), chloroform (Cat# 650498) and G418 disulfate salt (Cat# A1720) were purchased from Sigma-Aldrich. Thiamet-G (Cat# S7213) and puromycin 2HCl (Cat# S7417) were purchased from Selleck. Alkyne-PC-biotin (Cat# BB-28) and 2-(4-((bis((1-tert-butyl-1H-1,2,3-triazol-4-yl)methyl)amino)methyl)-1H-1,2,3-triazol-1-yl) acetic acid (BTTAA, Cat# BDJ-4) were purchased from Confluore Biotech (Xi’an, China). The mass spectra grade trypsin (Cat# HLS TRY001C) was purchased from Beijing Life Proteomic (Beijing, China). Alkyne-biotin (Cat# 1137) and alkyne-AZDye-488 (Cat# 1277) were purchased from Click Chemistry Tools. Nonidet P-40 (NP-40, Cat# 20103ES) was purchased from Yeasen (Shanghai, China). UDP-GalNAz (Cat# CLK-077) was purchased from Jena Bioscience. Alkyne-PEG_5KD (Cat# HWG26053) was purchased from HWRK CHEM (Beijing, China). Cycloheximide (CHX, Cat# HY-12320) and MG132 (Cat# HY-13259) were purchased from MedChemExpress. Ac₄5SGlcNAc, 1,6-Pr₂GalNAz, Y289L GalT1 and HA-Ubiquitin Vector were Gift from Dr. Xing Chen’s lab. Pierce™ bicinchoninic acid (BCA) Kit (Cat# 23227) and Hoechst 33342 (Cat# H3570) were purchased from Thermo Fisher Scientific. FDbio-Dura enhanced chemiluminescence (ECL) Kit (Cat# FD8020) was purchased from Fdbio science (Hangzhou, China). jetPRIME® transfection reagent (Cat# 101000046) was purchased from Polyplus. RNA isolater Total RNA Extraction Reagent (Cat# R401-01), HiScript III RT SuperMix for qPCR (+ gDNA wiper, Cat# R323-01), ChamQ Universal SYBR qPCR Master Mix (Cat# Q711-02), Cell Counting Kit-8 (Cat# A311-02) and ClonExpress Ultra One Step Cloning Kit (Cat# C115-02) were purchased from Vazyme (Nanjing, China). BeyoBlue™ Coomassie Brilliant Blue Ultrafast Staining Solution (CBB, Cat# P0017F), Annexin V-FITC Cell Apoptosis Detection Kit (Cat# C1062L) and Cell Cycle Detection Kit (Cat# C1052) were purchased from Beyotime (Shanghai, China).

Immunohistochemistry analysis

The formalin-fixed, paraffin-embedded samples were sliced, dewaxed, and rehydrated, followed by incubation with anti-O-GlcNAc (RL2, 1:200, MA1-072, Thermo Fisher Scientific), anti-OGT (1:200, ab177941, Abcam), anti-OGA (1:200, ab124807, Abcam) or anti-Ki67 (1:200, GB111499, Servicebio) antibody, and secondary antibody. After that, samples were incubated with diaminobenzidine (Dako, USA) and restained with hematoxylin (Sigma-Aldrich). The analysis of the IHC score (RL2 (anti-O-GlcNAc), OGT and OGA staining) of CCA tumor tissues and adjacent normal bile duct was performed as previously described.⁵⁸ Level of staining: 0-negative, 1-weakly positive, 2-positive, 3-strongly positive. For Ki-67 staining, the positive area of xenograft tumors was determined through IHC analysis.

Protein extraction and Western blot analysis

The cells were lysed using 4% sodium dodecyl sulfate (SDS, wt/vol) supplementing with protease inhibitor (Beyotime) under sonication. After centrifugation, the supernatant was quantified for protein concentration using a BCA kit (Pierce, USA).

For Western blot analysis, cell lysates supplemented with 5× loading buffer were boiled at 99 ℃ for 5 min. Equal amounts of proteins were separated by SDS-PAGE gel and blotted onto a polyvinylidene fluoride (PVDF) membrane. After blocking with 5% non-fat powdered milk (wt/vol), the membrane was incubated with specific primary antibodies for anti-O-GlcNAc (RL2, 1:2000, ab2739, Abcam), anti-OGT (1:2000, ab177941, Abcam), anti-OGA (1:2000, ab124807, Abcam), anti-β-actin (1:5000, FD0060, Fdbio), anti-Cleaved PARP (1:1000, #5625, Cell Signaling Technology), anti-Cleaved Caspase-3 (1:1000, #9664, Cell Signaling Technology), anti-Bcl2 (1:1000, A19693, ABclonal), anti-FOXM1 (1:1000, sc-376471, Santa Cruz), anti-cMyc (1:1000, AF6513, Beyotime), anti-BUB1 (1:1000, DF6698, Affinity), anti-Cyclin D1 (1:1000, AF0126, Beyotime), anti-Cyclin E1 (1:1000, AF6384, Beyotime), anti-Cyclin A2 (1:1000, AF6624, Beyotime), anti-Cyclin B1 (1:1000, AF6627, Beyotime), anti-K18 (1:1000, sc-6259, Santa Cruz), anti-FLAG (1:2000, M20008, Abmart), anti-FLAG (1:1000, AG8050, Beyotime), anti-Nup98 (1:1000, sc-74553, Santa Cruz), anti-Nup153 (1:1000, sc-101544, Santa Cruz), anti-MIDEAS (1:1000, sc-514710, Santa Cruz), anti-FOXK1 (1:1000, sc-373810, Santa Cruz), anti-CNOT2 (1:1000, sc-81229, Santa Cruz), anti-HA (1:1000, M20003, Abmart), anti-IDH2 (1:1000, sc-374476, Santa Cruz), anti-IDH3A (1:1000, sc-398021, Santa Cruz), anti-IDH3B (1:1000, A13742, ABclonal), anti-IDH3G (1:1000, sc-365489, Santa Cruz), anti-OGDH (1:1000, 15212-1-AP, Proteintech), anti-SUCLG1 (1:1000, 14923-1-AP, Proteintech), anti-PDHA1 (1:1000, 18068-1-AP, Proteintech) and anti-ACO2 (1:1000, 11134-1-AP, Proteintech), followed by secondary horseradish peroxidase (HRP)-conjugated antibodies. After washes with Tris-buffered saline with Tween 20 (TBST), blots were reacted with the ECL reagent (Fdbio science) and protein bands were detected by a chemiluminescence system (Tanon-5200, Shanghai, China).

qRT-PCR analysis

Total RNA was isolated from cultured cells by RNA isolater Total RNA Extraction Reagent (Vazyme) according to the manufacturer’s instructions. To quantify mRNAs, total RNA was converted to cDNA using the reverse transcription kit (Vazyme), followed by PCR using ChamQ Universal SYBR qPCR Master Mix (Vazyme) and gene-specific primers (Table S4). All of the reactions were run in triplicate. The expression levels of mRNAs were normalized to ACTB mRNA using the 2^−ΔΔCт method.

CCK-8 assay

The cell viability of HuCCT1, RBE and HCCC-9810 cells was determined using the CCK-8 assays (Vazyme) following the manufacturer’s instructions. Briefly, HuCCT1, RBE and HCCC-9810 cells were seeded into 96-well plates at a density of 1×10⁴ cells/well, followed by exposure to a gradient concentration of 5S and TMG for 48 h. Then, 100 µL/well RPMI 1640 medium containing 10% CCK-8 was replaced into the test well and incubated at 37 ℃ for 1 h. Absorbance was then measured at a wavelength of 450 nm.

Cell apoptosis assay

HuCCT1, RBE and HCCC-9810 cells were seeded into 6-well plates at a density of 4×10⁵ cells/well, followed by exposure to a gradient concentration of 5S and TMG. After incubation for 48 h, the cells were harvested for apoptosis analysis. Briefly, the cells were washed twice with cold PBS and resuspended in 1× binding buffer with 1×10⁶ cells/mL, following FITC-Annexin V (FITC fluorescence) and PI (PE fluorescence) were added. The cells were incubated at room temperature for 15 mins in the dark and were analyzed by flow cytometry (Agilent) within 1 h after staining.

Cell cycle analysis

Cell cycle distribution of HuCCT1, RBE and HCCC-9810 cells followed by exposure to a gradient concentration of 5S and TMG for 48 h was detected by FACS analysis. Briefly, cells were trypsinized into single-cell suspension, rinsed with ice-cold PBS and fixed in ice-cold 70% ethanol overnight. Then the cells were stained in PI added with RNase A (100 µg/ml) at 37 ℃ for 30 mins, and determined by flow cytometry (Agilent).

Colony formation assay

HuCCT1 and RBE cells transfected with ncRNA, siOGT-1, siOGT-2, siOGT-3, siOGA-1, siOGA-2 or siOGA-3 were seeded into 6-well plates at a density of 300 cells/well, and cultured in RPMI 1640 medium supplemented with 10% FBS for 14 days, during which the medium was replaced every 3 days. Colonies were then fixed with methanol for 10 mins and stained with 4% crystal violet (Solarbio) in PBS for 15 mins. Colony formation was shown by the number of stained colonies.

Metabolic oligosaccharide engineering (MOE) of living cells

HuCCT1, HIBEpiC, HCCC-9810 and RBE cells seeded at 10-cm dishes were treated with unnatural sugar (1,6-Pr₂GalNAz) at varied concentrations for 48 h or with 200 µM 1,6-Pr₂GalNAz for up to 72 h when they reached approximately 30% confluence. The cells were harvested by trypsin digestion and washed twice with PBS, and were lysed as described in the ‘Protein extraction and Western blot analysis’ section. All lysates were incubated with 500 µM premixed CuSO₄/BTTAA (molar ratio 1:2), 100 µM alkyne-biotin and 2.5 mM fresh sodium ascorbate for 2 h at room temperature, followed adding with 5× loading buffer were boiled at 99 ℃ for 5 min. Equal amounts of proteins were detected with anti-biotin (1:2000, A0303, Beyotime) by Western blot analysis. For confocal fluorescence microscopy imaging, the cells were seeded into 8-chamber at 1×10⁴ cells/well and treated with 200 µM 1,6-Pr₂GalNAz for 48 h. The cells were washed twice with PBS, and then fixed with 4% paraformaldehyde (wt/vol), and permeabilized with 0.5% Triton-X 100 (vol/vol). Then, the cells were incubated with 50 µM premixed CuSO₄/BTTAA (molar ratio 1:6), 50 µM alkyne-AZDye-488 and 2.5 mM fresh sodium ascorbate for 10 mins at room temperature. For nucleus staining, cells were incubated with 5 µg/mL Hoechst 33342 at room temperature for 20 min. The cells were washed three times after each step. Finally, the cells were imaged by Leica TCS SP5 laser scanning confocal system equipped with a ×63 oil immersion objective lens. For FACS analysis, the cells were seeded at 6-well plates and treated with 200 µM 1,6-Pr₂GalNAz for 48 h when they reached approximately 30% confluence. The cells were trypsinized into single-cell suspension, rinsed with ice-cold PBS and fixed in ice-cold 70% ethanol overnight. Then, the cells were incubated with 50 µM premixed CuSO₄/BTTAA (molar ratio 1:6), 50 µM alkyne-biotin and 2.5 mM fresh sodium ascorbate for 10 mins on ice, following incubated with streptavidin-Alexa Fluor™ 488 Conjugate (1:2000, S32354, Thermo Fisher Scientific) for 30 mins on ice, and determined by flow cytometry (BD Biosciences).

Glycopeptides enrichment

The lysates extracted from HuCCT1 and HIBEpiC cells were incubated with 500 µM premixed CuSO₄/BTTAA (molar ratio 1:2), 100 µM alkyne-PC-biotin and 2.5 mM fresh sodium ascorbate for 3 h at room temperature. The mixture was added to 8 volumes of ice-cold methanol for precipitation overnight at -80 ℃ and washed three times with ice-cold methanol. Then, the protein pellet was reconstituted by sonication using 4 M urea in 50 mM ABC, incubated with 10 mM DTT at 37 ℃ for 1 h, followed by incubation with 20 mM IAA at room temperature for 30 mins in the dark. The solution was diluted to 1 M urea in 50 mM ABC supplemented with mass spectra grade trypsin (enzyme : substrate ratio at 1 : 50) and reacted at 37 ℃ for 16 h. After that, the streptavidin agarose beads (150 µL per 40 mg, Thermo Fisher Scientific, 20359) were added to the solution above and gently rotated for 3 h at room temperature. The beads were washed five times with PBS and Milli-Q water successively and re-suspended with 200 µL 0.1% formic acid (FA, vol/vol), followed by irradiating three times under 365 nm UV light for 5 min using a UV cross-linker (CL-1000; UVP). The supernatant was collected, evaporated in a vacuum centrifuge, and subjected to LC-MS/MS analysis.

LC-MS/MS analysis and data processing

LC-MS/MS analysis of the glycopeptides enriched above was performed as previously described.¹⁹ In brief, all samples were resuspended with 0.1% FA, and subjected to analysis using a Dionex Ultimate 3000 RPLC nano system, which was connected with an Orbitrap Fusion Lumos Tribrid Mass Spectrometer with an EASY-Spray ionization source (Thermo Fisher Scientific). Survey scans of precursor were collected in Orbitrap from 350–2000 Th, under the resolution of 120,000 at 200 m/z. Monoisotopic precursors selection was enabled, and the multi-charged precursors with z = 2–8 were selected for data-dependent MS/MS scans with a cycle time of 3 s, dynamic exclusion set to 15 s, and window set to ± 10 ppm. The initial data-dependent MS/MS scans were acquired using HCD with a first mass of 100 Th, a normalized collision energy (NCE) of 30 ± 10, and a resolution of 30,000 at 200 m/z, according to the sceHCD-pd-EThcD method. Following ETD fragmentation triggered by glycan oxonium fragments, the glycan oxonium ions (i.e, m/z 168.0654, 186.0760, 204.0865, 300.1302, 366.1395, 388.1469, etc.) were detected in the sceHCD spectrum with a mass accuracy within 10 ppm, meanwhile, additional precursor isolation and EThcD acquisition were performed with supplemental activation of 35.

The raw data processing was performed using pGlyco3 (https://github.com/pFindStudio/pGlyco3/releases/tag/pGlyco3.0.rc3_build2021 0124) under the “HCD + EThcD” mode as previously described.^19,20 Briefly, MS/MS spectrum were searched against the SwissPort Mus musculus proteome database downloaded from Uniprot (https://www.uniprot.org) on 4th November 2016. The N-glycans and O-glycans were searched against pGlyco-N-glycan mode and pGlyco-O-glycan mode in the pGlyco3 software with modified glycan databases, respectively. The parameter precursor tolerance was set to ± 10 ppm and fragment tolerance ± 20 ppm. The false discovery rate (FDR) was set to less than 1%. Notably, the O-GlcNAc sites were assigned manually based on the subcellular localization. The proteins localized in the cytoplasmic side (including the nucleus, cytoplasmic, mitochondrial and cytoplasmic part of transmembrane proteins) were selected as O-GlcNAc proteins. For quantification, the O-GlcNAc sites co-identified in three independent replicates of HuCCT1 and HIBEpiC cells were defined as quantified O-GlcNAc sites. The O-GlcNAc sites with P-value < 0.05 and a fold change > 1.50 or < 0.67 were considered as up-regulated or down-regulated O-GlcNAc sites, respectively.

Chemoenzymatic labeling of O -GlcNAcylated proteins

The cells were lysed as described in the ‘Protein extraction and Western blot analysis’ section, and the lysates were added to 8 volumes of ice-cold methanol overnight at -80 ℃ and washed three times with ice-cold methanol. The proteins were reconstituted with 1% SDS (wt/vol) in 20 mM HEPES buffer (pH 7.9) using sonication, and incubated with 125 mM NaCl, 5% Nonidet P-40 (NP-40, vol/vol), 50 mM HEPES (pH 7.9), 100 mM MnCl₂, 500 µM UDP-GalNAz, and Y289L GalT1 (enzyme : substrate ratio at 1 : 40) at 4 ℃ for 20 h. Methanol, chloroform and Milli-Q water are successively added to the solution above (solution : methanol : chloroform : Milli-Q water ratio at 1 : 3 : 0.75 : 2) to obtain protein pellet, following washed three times with ice-cold methanol. The proteins were resuspended as described above.

Immunoprecipitation assay

For biotin immunoprecipitation, the cells were labeled and lysed as mentioned above. The biotinylated lysates were treated with streptavidin agarose beads (10 µL/mg), and gently rotated for 3 h at room temperature. For FLAG immunoprecipitation, the cells transfected with indicated plasmids were labeled and lysed as mentioned above. Then, the biotinylated lysates were treated with anti-FLAG beads (10 µL/mg, M20038, Abmart) at 4 ℃ for 12 h with gentle rotation. The beads were washed five times with PBS, following added with 5× loading buffer and boiled at 99 ℃ for 5 min. The final immunoprecipitated proteins were analyzed by western blot.

O -GlcNAcylation stoichiometry on K18

The cells treated with metabolic oligosaccharide engineering or chemoenzymatic labeling were incubated with 500 µM premixed CuSO₄/BTTAA (molar ratio 1:2), 100 µM alkyne-PEG_5KD and 2.5 mM fresh sodium ascorbate at 37 ℃ for 16 h, followed adding with 5× loading buffer were boiled at 99 ℃ for 5 min. Equal amounts of proteins were detected with anti-K18 (Santa Cruz) by Western blot analysis.

Plasmids construction

The FLAG-tagged K18 constructs were generated by in-frame subcloning the human KRT18 cDNA into the pFLAG-CMV-2 vector (Sigma-Aldrich). The K18 mutants (S15A, S18A, S15/18A, S30A, S31A, S30/31A and 4A) were generated using the ClonExpress Ultra One Step Cloning Kit (Vazyme) according to the manufacturer’s protocol. All constructs were confirmed by DNA sequencing (Sangon, Shanghai, China). All primers used in plasmids construction were provided in this study. (Table S4).

Lentivirus infection and stable cell line establishment

The lentiviruses with random small hairpin RNA (shNC, used as a negative control) and small hairpin RNA K18 knockdown (shK18-1 ~ 3) were generated by GenePharma (Shanghai, China) using the LV-U6-copGFP-T2A-Neo vector (GenePharma) according to the manufacturer’s instructions. The HuCCT1 cells were infected with these lentiviruses and selected for stable cell lines with 400 µg/mL G418 for two weeks. We screened the lentivirus with the highest efficiency of knockdown of endogenous K18 (shK18) by qRT-PCR and western blot analysis. To generate K18 reconstituted stable cell lines, we used the co-expressing exogenous FLAG-K18 WT or FLAG-K18 S30A lentivirus which was constructed with LV8N vector (Mock, GenePharma) to infect the shK18 HuCCT1 cells and selected for stable cell lines with 2 µg/mL puromycin for two weeks. The successful construction of stable cell lines shNC + Mock, shK18, shK18 + WT and shK18 + S30A were verified by qRT-PCR and western blot analysis. For CCK-8 analysis, the cells were seeded into 96-well plates at a density of 5×10³ cells/well, and the viability of cells was determined using the CCK-8 assays (Vazyme) following the manufacturer’s instructions. Cell cycle distribution of the cells above was by FACS analysis as described in the ‘Cell cycle analysis’ section. For colony formation, the cells were seeded into 6-well plates at a density of 200 cells/well, and detected as described in the ‘Colony formation assay’ section.

Determination of K18 half-life

The degradation of K18 in HuCCT1, RBE and HCCC-9810 cells incubated with DMSO, 200 µM 5S or 1 µM TMG for 48 h, followed by treatment with 10 µg/mL CHX for 0, 2, 4 or 8 h was determined by Western blot analysis. The stable cell lines (shK18 + WT and shK18 + S30A) were treated with 10 µg/mL CHX for 0 or 8 h. The FLAG-K18 protein levels were analyzed by western blot.

Ubiquitination assay

The stable cell lines (shK18 + WT and shK18 + S30A) were transfected with HA-ubiquitin and treated with 0 or 5 µM MG-132 for 20 h. Then, the cells were lysed and captured with anti-FLAG beads. Anti-HA blot demonstrated the ubiquitination of immunoprecipitated FLAG-K18.

Tumorigenesis in nude mice

5-week-old BALB/c nude mice (male) were purchased from GemPharmatech (Nanjing, China) and housed under specific pathogen-free (SPF) conditions. All animal experimental procedures were approved by the Animal Care and Use Committee of Nanjing University. Specifically, BALB/c nude mice as mentioned above were randomly divided into four groups (n = 6 per group) and injected subcutaneously with 5×10⁶ HuCCT1 stable cells with shNC + Mock, shK18, shK18 + WT and shK18 + S30A K18. shNC + Mock was the control group. Tumors generated by xenograft HuCCT1 stable cells were measured every six days from day 10. After 34 days, tumors were dissected, photographed, weighed, and subjected to immunohistochemistry analysis of the Ki67 protein.

Analysis of O -GlcNAcylated K18 in tissues

The tissues of xenograft tumors generated from HuCCT1 cells (shK18 + WT and shK18 + S30A) and patient samples were ground to single cells and lysed as described in the ‘Protein extraction and Western blot analysis’ section. Then, the lysates were biotinylated by chemoenzymatic labeling, and captured with streptavidin beads. The O-GlcNAcylation of K18 was determined by Western blot analysis.

Interacting proteins analysis

The lysates from shK18 cells transfected with FLAG-K18^WT or FLAG-K18^S30A were captured with anti-FLAG beads and analyzed by LC-MS/MS. For protein quantification, the fold change from three biological replicates of the interacting proteins with K18 in FLAG-K18^WT cells was calculated relative to FLAG-K18^S30A cells. Proteins with P-value < 0.05 and a fold change > 1.50 or < 0.67 were considered as up-regulated or down-regulated interacting proteins, respectively. KEGG enrichment analysis for the upregulated interacting proteins was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) bioinformatics resources.

Analysis of metabolites by LC–MS/MS

The extraction and analysis of cell metabolic products were carried out according to the previous procedures.⁶¹ Briefly, the medium of cells seeded at the 10-cm dish was absorbed completely by a pump. The cells were rinsed with pre-cooled PBS for three times, and the dish was immediately placed on dry ice with 80% pre-cooled methanol (vol/vol), and incubated at -80°C for 1 h. Subsequently, the cells were scraped off on dry ice using a cell scraper, and the lysates were centrifuged at 14,000 g for 20 mins at 4°C to obtain the supernatant-containing metabolites. The pellets were lysed and quantified as described in the ‘Protein extraction and Western blot analysis’ section, and the supernatant containing metabolites was dried into powder using SpeedVac and lyophilize. The samples were resuspended using 100 µL 50% methanol (vol/vol). After centrifugation, 1µL supernatant was injected into the Q Exactive mass spectrometer (Thermo Fisher Scientific) for detection. For each metabolite, the standard compound was detected to ensure proper chromatographic elution time and generate a standard curve. LC-MS/MS analysis was performed as previously described.⁶² Data were acquired and processed using Tracefinder software. The quantity of the metabolite fraction was normalized to the corresponding protein.

Quantification and statistical analysis

Data from three independent experiments are shown as the mean ± standard deviation (SD). Student’s t-tests (two-tailed) were used to compare two data sets, and a P-value of < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, not significant). Otherwise, the P-value of the Kaplan-Meier survival curve (Fig. S2C, S2D) was analyzed by the log-rank (Mantel-Cox) test, which is a hypothesis test to compare the survival distributions of two samples. The adjusted P-value (adjusted by a conservative Bonferroni correction) of motif analysis in O-GlcNAc sites, N-glycosites and mucin-type O-glycosites was analyzed using the pLogo (https://plogo.uconn.edu/) (Fig. S9B, S10D). GO terms (Fig. S9C) and KEGG enrichment analysis (Fig. 6B) were performed using DAVID bioinformatics resources (https://david.ncifcrf.gov/) and the enrichment P-values were adjusted by a modified Fisher’s exact test.

Data availability

The human cholangiocarcinoma data (Fig. S2A, S2B) were derived from the GEO (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi) and TCGA Research Network (https://www.cancer.gov/ccg/research/genome-sequencing/tcga). The mass spectrometry data have been deposited to ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository43 with the dataset identifier PXD043869. All other data supporting the findings of this study are available from the corresponding author on reasonable request.

Acknowledgments

We thank Prof. Dr. Xing Chen (Peking University) for his guidance with glycoproteomics and insightful discussions. We acknowledge the support from the National Center for Protein Sciences (Beijing, China) for the LC-MS/MS experiments.

Author contributions

R.X., J.L., and X.M. conceptualized and designed the experiments; R.X., J.L., and L.W. supervised the project; X.M., Y.Z., C.W., X.T., and X.Z. performed the experiments; X.M., and Y.Z. analyzed the data; L.X., and Y.Z. collected the clinical samples; B.C., and Y.H. provided materials support; R.X., J.L., and X.M. wrote this paper. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (2207070006 and 22107005), the Natural Science Foundation of Jiangsu Province (No. BK20202299 and BK20202004), the Programs for High-level Entrepreneurial and Innovative Talents Introduction of Jiangsu Province (Individual and Group Program), the Fundamental Research Funds for the Central Universities (021414380508), and the STI2030-Major Projects (2022ZD0211804).

Competing interests

The authors declare no competing interests.

Banales JM, Marin JJG, Lamarca A, Rodrigues PM, Khan SA, Roberts LR, et al. Cholangiocarcinoma 2020: the next horizon in mechanisms and management. Nat Rev Gastroenterol Hepatol. 2020; 17: 557-588.
Rizvi S, Khan SA, Hallemeier CL, Kelley RK & Gores GJ. Cholangiocarcinoma-evolving concepts and therapeutic strategies. Nat Rev Clin Oncol. 2018; 15: 95-111.
Baria K, Toni END, Yu B, Jiang Z, Kabadi SM & Malvezzi M. Worldwide incidence and mortality of biliary tract cancer. Gastro Hep Adv. 2022; 1: 618-626.
Varki A, Cummings RD, Esko JD, Stanley P, Hart GW, Aebi M, et al. Essentials of glycobiology. 4th edn. Ch. 46-47 (Cold Spring Harbor Laboratory Press, 2022).
Pinho SS & Reis CA. Glycosylation in cancer: mechanisms and clinical implications. Nat Rev Cancer. 2015; 15: 540-555.
Hart GW, Housley MP & Slawson C. Cycling of O-linked β-N-acetylglucosamine on nucleocytoplasmic proteins. Nature. 2007; 446: 1017-1022.
Wulff-Fuentes E, Berendt RR, Massman L, Danner L, Malard F, Vora J, et al. The human O-GlcNAcome database and meta-analysis. Sci. Data. 2021; 8: 25.
Leney AC, Atmioui DE, Wu W, Ovaa H & Heck AJR. Elucidating crosstalk mechanisms between phosphorylation and O-GlcNAcylation. Proc Natl Acad Sci. 2017; 114: E7255-E7261.
Lee JB, Pyo KH & Kim HR. Role and function of O-GlcNAcylation in cancer. Cancers. 2021; 13: 5365.
Lee BE, Suh PG & Kim JI. O-GlcNAcylation in health and neurodegenerative diseases. Exp Mol Med. 2021; 53: 1674-1682.
Gonzalez-Rellan MJ, Fondevila MF, Dieguez C & Nogueiras R. O-GlcNAcylation: a sweet hub in the regulation of glucose metabolism in health and disease. Front Endocrinol. 2022; 13: 873513.
Slawson C & Hart GW. O-GlcNAc signalling: implications for cancer cell biology. Nat Rev Cancer. 2011; 11: 678-684.
Phoomak C, Silsirivanit A, Wongkham C, Sripa B, Puapairoj A, & Wongkham S. Overexpression of O-GlcNAc-transferase associates with aggressiveness of mass-forming cholangiocarcinoma. Asian Pac J cancer Prev. 2012; APJCP 13 Suppl. : 101-105.
Phoomak C, Vaeteewoottacharn K, Sawanyawisuth K, Seubwai W, Wongkham C, Silsirivanit A, et al. Mechanistic insights of O-GlcNAcylation that promote progression of cholangiocarcinoma cells via nuclear translocation of NF-κB. Sci Rep. 2016; 6: 27853.
Phoomak C, Park D, Silsirivanit A, Sawanyawisuth K, Vaeteewoottacharn K, Detarya M, et al. O‐GlcNAc‐induced nuclear translocation of hnRNP‐K is associated with progression and metastasis of cholangiocarcinoma. Mol Oncol. 2019; 13: 338-357.
Gloster TM, Zandberg WF, Heinonen JE, Shen DL, Deng L & Vocadlo DJ. Hijacking a biosynthetic pathway yields a glycosyltransferase inhibitor within cells. Nat Chem Biol. 2011; 7: 174-181.
Yuzwa SA, Macauley MS, Heinonen JE, Shan X, Dennis RJ, He Y, et al. A potent mechanism-inspired O-GlcNAcase inhibitor that blocks phosphorylation of tau in vivo. Nat Chem Biol. 2008; 4: 483-490.
Liu C & Li J. O-GlcNAc: a sweetheart of the cell cycle and DNA damage response. Front Endocrinol. 2018; 9: 415.
Liu J, Cheng B, Fan X, Zhou X, Wang J, Zhou W, et al. Click-iG: simultaneous enrichment and profiling of intact N‐linked, O‐GalNAc, and O‐GlcNAcylated glycopeptides. Angew Chem Int Ed. 2023; 62: e202303410.
Zeng WF, Cao WQ, Liu MQ, He SM & Yang PY. Precise, fast and comprehensive analysis of intact glycopeptides and modified glycans with pGlyco3. Nat Methods. 2021; 18: 1515-1523.
Cheng B, Wang C, Hao Y, Wang J, Xia X, Zhang H, et al. Facile synthesis of clickable unnatural sugars in the unprotected and 1,6‐Di‐O‐Acylated forms for metabolic glycan labeling. Chem A Eur J. 2023; 29: e202203054.
Jacob JT, Coulombe PA, Kwan R & Omary MB. Types I and II keratin intermediate filaments. Cold Spring Harb Perspect Biol. 2018; 10: a018275.
Ku N, Strnad P, Bantel H & Omary MB. Keratins: biomarkers and modulators of apoptotic and necrotic cell death in the liver. Hepatolo. 2016; 64: 966-976.
Weng YR, Cui Y & Fang JY. Biological functions of cytokeratin 18 in cancer. Mol Cancer Res. 2012; 10: 485-493.
Snider NT, Weerasinghe SVW, Iñiguez-Lluhí JA, Herrmann H & Omary MB. Keratin hypersumoylation alters filament dynamics and is a marker for human liver disease and keratin mutation. J Biol Chem. 2011; 286: 2273-2284.
Jang K, Yoon H, Lee J, Yi H, Park S, Lee S, et al. Liver disease–associated keratin 8 and 18 mutations modulate keratin acetylation and methylation. FASEB J. 2019; 33: 9030-9043.
Ku NO & Omary MB. Identification and mutational analysis of the glycosylation sites of human keratin 18. J Biol Chem. 1995; 270: 11820-11827.
Kakade PS, Budnar S, Kalraiya RD & Vaidya MM. Functional implications of O-GlcNAcylation-dependent phosphorylation at a proximal site on keratin 18. J Biol Chem. 2016; 291: 12003-12013.
Khidekel N, Ficarro SB, Clark PM, Bryan MC, Swaney DL, Rexach JE, et al. Probing the dynamics of O-GlcNAc glycosylation in the brain using quantitative proteomics. Nat Chem Biol. 2007; 3: 339-348.
Chou CF & Omary MB. Mitotic arrest-associated enhancement of O-linked glycosylation and phosphorylation of human keratins 8 and 18. J. Biol Chem. 1993; 268: 4465-4472.
Srikanth B, Vaidya MM & Kalraiya RD. O-GlcNAcylation determines the solubility, filament organization, and stability of keratins 8 and 18. J Biol Chem. 2010; 285: 34062-34071.
Phoomak C, Vaeteewoottacharn K, Silsirivanit A, Saengboonmee C, Seubwai W, Sawanyawisuth K, et al. High glucose levels boost the aggressiveness of highly metastatic cholangiocarcinoma cells via O-GlcNAcylation. Sci Rep. 2017; 7: 43842.
Silsirivanit A, Phoomak C & Wongkham S. Diagnosis and management of cholangiocarcinoma. Ch. 25 (Springer, Cham. Press, 2021).
Moro A, Mehta R, Sahara K, Tsilimigras DI, Paredes AZ, Farooq A, et al. The impact of preoperative CA19-9 and CEA on outcomes of patients with intrahepatic cholangiocarcinoma. Ann Surg Oncol. 2020; 27: 2888-2901.
Indramanee S, Silsirivanit A, Pairojkul C, Wongkham C & Wongkham S. Aberrant glycosylation in cholangiocarcinoma demonstrated by lectin-histochemistry. Asian Pac J cancer Prev. 2012; APJCP 13 Suppl. : 119–124.
Matsuda A, Kuno A, Nakagawa T, Ikehara Y, Irimura T, Yamamoto M, et al. Lectin microarray-based sero-biomarker verification targeting aberrant O‑linked glycosylation on mucin 1. Anal Chem. 2015; 87: 7274-7281.
Chang T, Cheng J, Tsai H, Young K, Hsieh S & Ho C. Plasma proteome plus site‐specific N‐glycoprofiling for hepatobiliary carcinomas. The J Pathol Clin Res. 2019; 5: 199-212.
Li J, Zhao T, Li J, Shen J, Jia L, Zhu B, et al. Precision N‐glycoproteomics reveals elevated LacdiNAc as a novel signature of intrahepatic cholangiocarcinoma. Mol Oncol. 2022; 16: 2135-2152.
Wattanavises S, Silsirivanit A, Sawanyawisuth K, Cha’on U, Waraasawapati S, Saentaweesuk W, et al. Increase of MAL-II binding alpha-2,3-aialylated glycan is associated with 5-FU resistance and short survival of cholangiocarcinoma patients. Medicina. 2019; 55: 761.
Ament CE, Steinmann S, Evert K, Pes GM, Ribback S, Gigante I, et al. Aberrant fucosylation sustains the NOTCH and EGFR/NF-κB pathways and has a prognostic value in human intrahepatic cholangiocarcinoma. Hepatol. (2023). 10.1097/hep.0000000000000322.
Detarya M, Lert-Itthiporn W, Mahalapbutr P, Klaewkla M, Sorin S, Sawanyawisuth K, et al. Emerging roles of GALNT5 on promoting EGFR activation in cholangiocarcinoma: a mechanistic insight. Am J cancer Res. 2022; 12: 4140-4159.
Queiroz RM de, Carvalho É & Dias WB. O-GlcNAcylation: the sweet side of the cancer. Front Oncol. 2014; 4: 132.
O’Donnell N, Zachara NE, Hart GW & Marth JD. OGT-dependent X-chromosome-linked protein glycosylation is a requisite modification in somatic cell function and embryo viability. Mol Cell Biol. 2004; 24: 1680-1690.
Muha V, Authier F, Szoke-Kovacs Z, Johnson S, Gallagher J, McNeilly A, et al. Loss of O-GlcNAcase catalytic activity leads to defects in mouse embryogenesis. J Biol Chem. 2021; 296: 100439.
Alteen MG, Tan HY & Vocadlo DJ. Monitoring and modulating O-GlcNAcylation: assays and inhibitors of O-GlcNAc processing enzymes. Curr Opin Struct Biol. 2021; 68: 157–165.
Boyce M, Carrico IS, Ganguli AS, Yu SH, Hangauer MJ, Hubbard SC, et al. Metabolic cross-talk allows labeling of O-linked β-N-acetylglucosamine-modified proteins via the N-acetylgalactosamine salvage pathway. Proc Natl Acad Sci. 2011; 108: 3141-3146.
Zhu Q & Yi W. Chemistry-assisted proteomic profiling of O-GlcNAcylation. Front Chem. 2021; 9: 702260.
Qin K, Zhang H, Zhao Z & Chen X. Protein S‑glyco-modification through an elimination–addition mechanism. J Am Chem Soc. 2020; 142: 9382-9388.
Jain R, Fischer S, Serra S & Chetty R. The use of cytokeratin 19 (CK19) immunohistochemistry in lesions of the pancreas, gastrointestinal tract, and liver. Appl Immunohistochem Mol Morphol. 2010; 18: 9-15.
Ku NO, Toivola DM, Strnad P & Omary MB. Cytoskeletal keratin glycosylation protects epithelial tissue from injury. Nat Cell Biol. 2010; 12: 876-885.
Ku NO, Liao J & Omary MB. Phosphorylation of human keratin 18 serine 33 regulates binding to 14‐3‐3 proteins. EMBO J. 1998; 17: 1892-1906.
Konzman D, Abramowitz LK, Steenackers A, Mukherjee MM, Na HJ & Hanover JA. O-GlcNAc: regulator of signaling and epigenetics linked to X-linked intellectual disability. Front Genet. 2020; 11: 605263.
Zhao Y, Yue S, Zhou X, Guo J, Ma S & Chen Q. O-GlcNAc transferase promotes the nuclear localization of the focal adhesion–associated protein Zyxin to regulate UV-induced cell death. J Biol Chem. 2022; 298: 101776.
Ruan HB, Nie Y & Yang X. Regulation of protein degradation by O-GlcNAcylation: crosstalk with ubiquitination. Mol Cell Proteom. 2013; 12: 3489-3497.
Zhou H, Palma SD, Preisinger C, Peng M, Polat AN, Heck AJR, et al. Toward a comprehensive characterization of a human cancer cell phosphoproteome. J Proteome Res. 2013; 12: 260-271.
Raggi C, Taddei ML, Rae C, Braconi C & Marra F. Metabolic reprogramming in cholangiocarcinoma. J Hepatol. 2022; 77: 849-864.
Zhu Q, Zhou H, Wu L, Lai Z, Geng D, Yang W, et al. O-GlcNAcylation promotes pancreatic tumor growth by regulating malate dehydrogenase 1. Nat Chem Biol. 2022; 18: 1087-1095.
Guo Y, Li Q, Ren W, Wu H, Wang C, Li X, et al. Quantitative proteomics reveals down-regulated glycolysis/gluconeogenesis in the large-duct type intrahepatic cholangiocarcinoma. J Proteome Res. 2022; 21: 2504-2514.
Wu M, Shi L, Merritt J, Zhu AX & Bardeesy N. Biology of IDH mutant cholangiocarcinoma. Hepatol. 2022; 75: 1322-1337.
Gorelik A & van Aalten DMF. Tools for functional dissection of site-specific O-GlcNAcylation. Rsc Chem Biol. 2020; 1: 98-109.
Yuan M, Breitkopf SB, Yang X & Asara JM. A positive/negative ion-switching, targeted mass spectrometry-based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. Nat Protoc. 2012; 7: 872-881.
Che N, Ma Y, Ruan H, Xu L, Wang X, Yang X, et al. Integrated semi-targeted metabolomics analysis reveals distinct metabolic dysregulation in pleural effusion caused by tuberculosis and malignancy. Clin Chim Acta. 2018; 477: 81-88.

There is no duality of interest

Download PDF

Version 1

posted

You are reading this latest preprint version

O-GlcNAcylation of Keratin 18 coordinates TCA cycle to promote cholangiocarcinoma progression

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Materials and methods

Patient samples and statement

Cell lines and cell culture

Reagents

Immunohistochemistry analysis

Protein extraction and Western blot analysis

qRT-PCR analysis

CCK-8 assay

Cell apoptosis assay

Cell cycle analysis

Colony formation assay

Metabolic oligosaccharide engineering (MOE) of living cells

Glycopeptides enrichment

LC-MS/MS analysis and data processing

Immunoprecipitation assay

Plasmids construction

Lentivirus infection and stable cell line establishment

Determination of K18 half-life

Ubiquitination assay

Tumorigenesis in nude mice

Interacting proteins analysis

Analysis of metabolites by LC–MS/MS

Quantification and statistical analysis

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1