Abstract

   [40]graphic file with name au4c00214_0007.jpg

   Core fucosylation, a special type of N-linked glycosylation, is
   important in tumor proliferation, invasion, metastatic potential, and
   therapy resistance. However, the core-fucosylated glycoproteome has not
   been extensively profiled due to the low abundance and poor ionization
   efficiency of glycosylated peptides. Here, a “one-step” strategy has
   been described for protein core-fucosylation characterization in
   biological samples. Core-fucosylated peptides can be selectively
   labeled with a glycosylated probe, which is linked with a
   temperature-sensitive poly(N-isopropylacrylamide) (PNIPAM) polymer, by
   mutant endoglycosidase (EndoF3-D165A). The labeled probe can be further
   removed by wild-type endoglycosidase (EndoF3) in a traceless manner for
   mass spectrometry (MS) analysis. The feasibility and effectiveness of
   the “one-step” strategy are evaluated in bovine serum albumin (BSA)
   spiked with standard core-fucosylated peptides, H1299, and Jurkat cell
   lines. The “one-step” strategy is then employed to characterize
   core-fucosylated sites in human lung adenocarcinoma, resulting in the
   identification of 2494 core-fucosylated sites distributed on 1176
   glycoproteins. Further data analysis reveals that 196 core-fucosylated
   sites are significantly upregulated in tumors, which may serve as
   potential drug development targets or diagnostic biomarkers. Together,
   this “one-step” strategy has great potential for use in global and
   in-depth analysis of the core-fucosylated glycoproteome to promote its
   mechanism research.

   Keywords: glycosylation, glycoproteomics, core-fucosylation,
   chemoenzymatic labeling, one-step strategy, lung adenocarcinoma,
   biomarker

Introduction

   Protein glycosylation (N-linked, O-linked, or C-linked) is one of the
   most ubiquitous and complex posttranslational modifications (PTMs),
   which is involved in various biological processes.^[41]1,[42]2 Core
   fucosylation is a special type of N-linked glycosylation, in which
   α1,6-fucose residue attaches to the innermost N-acetylglucosamine
   (GlcNAc) residue of N-linked glycans.^[43]3 The formation of core
   fucosylation is catalyzed by a single enzyme_fucosyltransferase 8
   (FUT8) in humans.^[44]4 Emerging evidence indicates that core
   fucosylation is associated with inflammation and cancer
   aggressiveness.^[45]5 An aberrant increase in core fucosylation is
   observed during tumorigenesis in various cancers, such as
   hepatocellular carcinoma (HCC), non-small-cell lung cancer (NSCLC),
   pancreatic ductal adenocarcinoma (PDAC), etc.^[46]6−[47]8 Many studies
   indicate that increased core fucosylation promotes proliferation,
   invasion, metastatic potential, and therapy resistance.^[48]9−[49]11
   The core-fucosylated alpha-fetoprotein (AFP) has been approved by the
   Food and Drug Administration (FDA) as a clinical HCC diagnostic marker.
   It is worth noting that the increase in core-fucosylated AFP in HCC
   patients can indicate cancer progression more specially than the
   increase in total AFP.^[50]12 Recently, another interesting discovery
   by Zhang and coworkers showed that multiple viral envelop proteins,
   such as hepatitis C virus (HCV)-E2 and severe acute respiratory
   syndrome coronavirus 2 (SARS-CoV-2)-spike, enhance FUT8 expression and
   core fucosylation levels.^[51]13 As a consequence, the development of
   efficient strategies to identify and quantify aberrant core
   fucosylation would provide a means for the discovery of new drug
   targets or diagnostic biomarkers.

   Due to the low abundance and poor ionization efficiency of glycosylated
   peptides, the detection of core-fucosylated sites in complex mixtures
   is challenging. To improve the coverage of core-fucosylated sites, it
   is essential to efficiently isolate core-fucosylated peptides from the
   digests of complex mixtures. Currently, diverse strategies have been
   developed for the enrichment of core-fucosylated peptides before liquid
   chromatography with tandem mass spectrometry (LC-MS/MS) analysis, such
   as lectin binding,^[52]14,[53]15 hydrophilic interaction chromatography
   (HILIC),^[54]16,[55]17 mixed-mode polymeric sorbent,^[56]18,[57]19
   metabolic labeling,^[58]20,[59]21 and chemoenzymatic
   labeling.^[60]22,[61]23 Lectin binding is widely used for purifying
   core-fucosylated peptides, owing to its affordability and ease of use.
   However, this method often suffers from weak binding affinity, cross
   reactivity, and waste of time.^[62]24 Recently, chemoenzymatic labeling
   has been developed to profile core fucosylation for its good
   specificity and few side reactions.^[63]22,[64]23,[65]25 There are two
   types of probes that have been widely used in chemoenzymatic labeling
   methods namely “two-step probes” and “one-step probes.”^[66]26 A
   “two-step probe” contains a biorthogonal reactive group, such as an
   alkynyl or azido group, and additional chemical reactions are required
   to introduce a reporter group. In contrast, a “one-step probe” is
   conjugated with a reporter group, such as a fluorescent or biotin
   group, via a covalent bond. Accumulating evidence shows that the
   “one-step probe” is much more sensitive than the “two-step probe” as it
   only relies on the enzymatic transfer and does not need additional
   chemical reaction.^[67]27−[68]29 These results have encouraged us to
   develop a “one-step” strategy for the global characterization of
   core-fucosylated glycoproteome in complex mixtures.

   In this work, a “one-step” strategy is described for protein
   core-fucosylation characterization in biological samples.
   Core-fucosylated peptides from complex biological samples were
   selectively labeled with a glycosylated probe (1), which is linked with
   a temperature-sensitive poly(N-isopropylacrylamide) (PNIPAM) polymer,
   by mutant endoglycosidase (EndoF3-D165A) ([69]Scheme [70]1
   ). The tagged probes can be reversibly removed by wild-type
   endoglycosidase (EndoF3) in a traceless manner for LC-MS/MS analysis.
   The performance of “one-step” strategy was evaluated in bovine serum
   albumin (BSA) spiked with standard core-fucosylated peptides, H1299,
   and Jurkat cell lines. Finally, the “one-step” strategy was applied to
   characterize core-fucosylated sites in human lung adenocarcinoma
   (LUAD).

Scheme 1. A “One-Step” Probe for the Global Profiling of Core-Fucosylated
Glycoproteome.

   Scheme 1
   [71]Open in a new tab

Results and Discussion

   Enrichment methods are often required before LC-MS/MS analysis, as core
   fucosylation has a low abundance in cells. Recently, we have developed
   a reversible labeling strategy for core fucosylation analysis on cell
   surface by using a biotinylated probe and EndoF3-D165A from
   Elizabethkingia meningoseptica.^[72]22 Nevertheless, this strategy
   cannot be used to label cell lysates or human samples as a strong
   nonspecific labeling signal caused by the oxazoline group was observed.
   Therefore, we have developed a “two birds one stone” method for
   core-fucosylated peptides and O-GlcNAcylated peptides analysis by
   employing a temperature sensitive PNIPAM.^[73]23 This method depended
   on a “two-step” enrichment strategy. First, mutant endoglycosidase
   (EndoF3-D165A) added an oxazoline probe (2) bearing an azido group to
   the target glycopeptides, and then the azido modified glycopeptides
   were linked with dibenzocyclooctyne functionalized PNIPAM (3) through
   strain-promoted azide–alkyne cycloaddition (SPAAC) reaction. Following
   a temperature dependent enrichment process, the target glycopeptides
   were enriched and released with wild-type endoglycosidase (EndoF3) for
   LC-MS/MS analysis. These two works indicated that endoglycosidases are
   powerful tools for glycosylation study. Inspired by these works, we
   developed a glycosylated probe (1) for efficient enrichment of core
   fucosylation in a “one-step” manner and in-depth site-specific analysis
   in large scale ([74]Scheme [75]1
   ).

Synthesis and Characterization of Probe 1

   PNIPAM (Mn ∼ 10 000), a temperature-sensitive polymer of
   N-isopropylacrylamide, displays an outstanding water solubility at room
   temperature, resulting in excellent compatibility with enzyme-catalyzed
   reactions. Therefore, a glycosylated probe (1) was synthesized by
   combining glycans (precursor of the enzymatic substrate) with PNIPAM
   via a hydrophobic linker. The alkyne modified linker (8) was prepared
   from 11-((tert-butoxycarbonyl)amino)undecanoic acid (6) through two
   step reactions. An alkyne linker bearing an amino group was reacted
   with carboxylated PNIPAM (9) and was linked with azido-modified glycans
   (10) through a copper-catalyzed azide–alkyne cycloaddition reaction to
   give glycosylated polymer 11. Probe 1 was obtained by treating 11 with
   2-chloro-1,3-dimethylimidazolinium chloride (DMC) in the presence of
   triethylamine ([76]Scheme [77]2
   , [78]Figure S1; for synthesis details, see the [79]Materials and
   Methods). Compound 1 bearing an oxazoline group was a donor of
   EndoF3-D165A and could be added to core-fucose (Fucα1,6GlcNAc
   structure) modified substrates through enzyme-catalyzed reaction.

Scheme 2. Chemical Synthesis of Probe 1.

   [80]Scheme 2
   [81]Open in a new tab

The Enrichment Effectiveness of a Core-Fucosylated Peptide with Two
Strategies

   A core-fucosylated peptide (4) was enriched with two strategies
   following the procedure described in [82]Scheme [83]1
   , respectively. For the “one-step” procedure, probe 1 was added to the
   solution of 4 for three times every 1 h in the presence of EndoF3-D165A
   to capture 4. For the “two-step” procedure, probe 2 was added to the
   solution of 4 for five times every 0.5 h in the presence of
   EndoF3-D165A to produce an azido-peptide (5). 5 was linked with
   compound 2 at room temperature for 1 h. Then, the temperature of the
   above mixtures was raised to 40 °C to precipitate the PNIPAM compounds.
   After centrifugation, the supernatant was analyzed by HPLC to obtain
   the decrease of the core-fucosylated peptide (after enrichment with 1)
   or the azido peptide (after enrichment with 3). The precipitate was
   washed to remove unreacted compounds and incubated with wild-type
   EndoF3 to release the enriched glycopeptide (4). HPLC analysis showed
   that 4 could be effectively enriched with probe 1, indicating that
   glycosylated polymer 1 can be accepted by EndoF3-D165A ([84]Figure
   [85]1A). Furthermore, the enriched core-fucosylated peptide was well
   released from probes 1 and 3. This is due to excellent substrate
   tolerance of EndoF3-D165A and EndoF3 and the PNIPAM polymer’s
   outstanding compatibility with enzymatic reactions in aqueous
   solutions.^[86]30,[87]31

Figure 1.

   Figure 1
   [88]Open in a new tab

   Enrichment and detection of core-fucosylated peptide (4) by the
   “one-step” and “two-step” methods. (A) HPLC analysis of the “catch and
   release” process of 4 following the procedure shown in [89]Scheme [90]1
   . (i) HPLC profile of standard glycopeptide 4 and control peptide. (ii)
   Glycopeptide 4 and control peptide were incubated with probe 2 and
   EndoF3-D165A at 30 °C for 2.5 h to produce azido-peptide 5. (iii)
   Supernatant of the reaction after enrichment of 4 with probe 1 by
   centrifugation. (iv) Supernatant of the reaction after enrichment of 5
   with probe 3 by centrifugation. (v,vi) Supernatant of the control group
   (only containing the enriched polymers 1 and 3). (vii,viii) Release of
   the glycopeptides from the enriched polymers 1 and 3 by EndoF3,
   respectively. During the repeat experiments, the peptide peaks were
   further collected and confirmed by MS analysis (see the [91]Materials
   and Methods for experimental details). (B) Mass spectra and (C) peak
   area of LC–MS analysis of the glycopeptide (4) and BSA peptide mixture
   (w/w 1:500) before and after enrichment by the “one-step” (probe 1) and
   “two-step” (probes 2 and 3) methods. The red asterisk indicates
   glycopeptide (4). (D) HCD MS/MS spectrum of the released glycopeptides.

   Next, we compared the effectiveness of two strategies in enriching the
   core-fucosylated peptide from a mixture of core-fucosylated peptides
   and BSA digests. Core-fucosylated peptide 4 was mixed with
   trypsin-digested peptides from BSA at a ratio of up to 1:500 (w/w). The
   mixture was enriched as described above using probe 1 or probes 2 and
   3, respectively. After intensive washes, the captured core-fucosylated
   peptide was released with EndoF3 and analyzed by MS. It was difficult
   to identify the peak of core-fucosylated peptide in the original
   mixture by MS as it was merged with other signals. After the enrichment
   procedure, the target peaks became dominant ([92]Figures [93]1
   B and [94]S2). Then, we calculated the peak area of the target
   glycopeptide (4) to evaluate the effectiveness of the two probes. The
   peak area of 4 after enrichment by probe 1 was approximately 1.9 times
   greater than that after enrichment by probes 2 and 3 ([95]Figure
   [96]1C). The further LC-MS/MS analysis gave information on where the
   glycosylation site occurred ([97]Figure [98]1D). These findings
   indicate that probe 1 is more effective than probes 2 and 3 for
   performing site-specific analysis of core fucosylation in a mixture of
   core-fucosylated peptides and BSA digests.

Identification of Core-Fucosylated Sites in Human Cell Lines

   We next employed probe 1 and probes 2 and 3 to profile the
   core-fucosylated glycoproteome in two human cancer cell lines, a lung
   cancer cell line (H1299) and a lymphocyte cell line (Jurkat). Proteins
   (10 mg) extracted from cancer cells were digested into peptides by
   trypsin. The peptide mixture (3 mg) was treated with EndoF3 to expose
   the reactive group (Fucα1,6GlcNAc structure) before the labeling
   reaction. The peptide mixture was labeled and enriched as described
   above. After intensive washing, the captured glycopeptides were
   released from the PNIPAM polymers (1 and 3) by EndoF3. In total,
   approximately 9.4 μg (0.31%) of glycopeptides was obtained using probe
   1, whereas about 7.5 μg (0.25%) of glycopeptides was obtained using
   probes 2 and 3. Owing to the better enrichment effectiveness of probe
   1, more glycopeptides was obtained. The enriched glycopeptides were
   further fractionated into six fractions with a tip-based concatenated
   high-pH reversed phase to improve the coverage of core fucosylation
   analysis. The fractionated peptides were analyzed by LC-MS/MS, and the
   MS data were processed by MaxQuant software.^[99]32 The identified
   core-fucosylated sites were considered positive results only if they
   occurred at the consensus motif of N-X-S/T/C (X ≠ P).

   In H1299, 1200 core-fucosylated sites corresponding to 652
   glycoproteins and 828 core-fucosylated sites corresponding to 491
   glycoproteins were identified by using probe 1 and probes 2 and 3,
   respectively ([100]Figure [101]2
   A and [102]Table S1). Almost all of the core-fucosylated sites (95.7%)
   identified by using probes 2 and 3 were detected by using probe 1. The
   motif of the core-fucosylated sites showed that T was more frequently
   found in position 2 compared to S with a frequency rate of 53.5% (probe
   1) and 54.2% (probes 2 and 3), which was in accordance with the
   previous report ([103]Figures [104]2B and [105]S3A).^[106]33 The
   distribution of core-fucosylated sites on proteins was also
   investigated, and at least 9% (probe 1) and 7% (probes 2 and 3) of the
   proteins had more than four sites ([107]Figures [108]2C and [109]S3B).
   Protein clustering results showed that the identified core-fucosylated
   proteins were mainly enriched in the biological processes of cell
   adhesion and cell migration, the cellular components of membrane and
   extracellular exosome, and the molecular functions of integrin binding
   and signaling receptor activity ([110]Figures [111]2D and [112]S3C). In
   Jurkat, 1277 core-fucosylated sites corresponding to 694 glycoproteins
   and 951 core-fucosylated sites corresponding to 550 glycoproteins were
   identified by employing probe 1 and probes 2 and 3, respectively
   ([113]Figures S4 and S5 and Table S2). These results indicated that
   probe 1 showed better enrichment effectiveness of core-fucosylated
   sites from human cell lines and could promote core fucosylation
   analysis from complex samples.

Figure 2.

   [114]Figure 2
   [115]Open in a new tab

   Core-fucosylated sites identified by the “one-step” (probe 1) method in
   H1299 cells. (A) Overlapped core-fucosylated sites identified by
   “one-step” (probe 1) and “two-step” (probes 2 and 3) methods. (B) Motif
   alignment of amino acids adjacent to core-fucosylated sites. (C)
   Distribution of core-fucosylated sites. (D) Clustering of identified
   core-fucosylated proteins based on the biological process, cellular
   component, and molecular function.

   In total, 1236 core-fucosylated sites corresponding to 671
   glycoproteins and 1345 core-fucosylated sites corresponding to 722
   glycoproteins were identified in H1299 and Jurkat in this work
   ([116]Figure S6). The function and domain of these glycoproteins were
   similar between H1299 and Jurkat, which was consistent with previous
   reports ([117]Figures S7–S9).^[118]16,[119]17,[120]22 Notably, up to
   785 core-fucosylated sites were overlapped in the two cell lines, and
   these conservative glycosylated sites potentially play essential roles
   in cellular function as the evolutionary selection ([121]Table S3). In
   our previous work, 119 conservative core-fucosylated sites were
   identified in HepG2, MCF7, and HeLa, 99 (83.2%) of which were detected
   in this work ([122]Figure S6). These results indicated the high
   coverage and confidence of the core fucosylation in this work. Then,
   the relative position of the 785 conservative sites in protein sequence
   was plotted and showed that proteins prolow-density lipoprotein
   receptor-related protein 1 (LRP1) and insulin-like growth factor 2
   receptor (IGF2R) had the highest number of conservative sites reaching
   18 and 13 sites with distinct peptide sequences surrounding the sites
   ([123]Figure S10). These heavily core-fucosylated proteins were
   reported to be associated with cancer invasion.^[124]34,[125]35
   Furthermore, it was observed that these 785 sites had relatively even
   distribution across the protein sequence but less frequent at the
   protein C-terminus ([126]Figure S10). This was in accordance with
   previous studies that the glycosylation efficiency of N-X-S/T/C (X ≠ P)
   motif decreases toward the C terminus.^[127]36,[128]37

Mapping Aberrant Core-Fucosylated Sites Associated with Human LUAD

   LUAD is one of the most common and fatal lung cancers in the
   world.^[129]38 Here, we performed proteomic and “one-step” method-based
   core-fucosylated glycoproteomic analysis of 4 pairs of LUAD tumors and
   normal adjacent tissues (NATs). This analysis of tumors and NATs was
   performed at the same time to ensure a similar enrichment recovery bias
   between tumors and NATs. A total of 2494 core-fucosylated sites (2371
   sites in tumors; 2161 sites in NATs) corresponding to 1176
   glycoproteins were identified ([130]Figure [131]3
   A and [132]Table S4). The abundance of these glycoproteins was further
   investigated, and it was observed that many glycoproteins with low
   abundance could also be identified by the “one-step” (probe 1) method
   ([133]Figure [134]3B). The motif, distribution, and function of these
   core-fucosylated proteins were similar to that of glycoproteins in cell
   lines ([135]Figures S11 and S12). However, owing to the complexity of
   tissue samples, about two-thirds of core-fucosylated sites detected in
   cell lines were also identified in the LUAD sample ([136]Figure S13A).
   The results from cell lines and LUAD sample were combined, 3081
   core-fucosylated sites were identified in this work, and approximately
   61.7% (1900 sites) were newly identified ([137]Figure S13B).

Figure 3.

   Figure 3
   [138]Open in a new tab

   Profiling of core fucosylation in human LUAD by the “one-step” (probe
   1) method. (A) Number of identified core-fucosylated sites in tumors
   and the paired NATs. (B) Protein abundance distribution (Log[2](protein
   intensity)) of identified glycoproteins compared with that in the total
   proteomic data set. (C) Comparison between tumors and the paired NATs
   using the total intensity of core-fucosylated sites in each sample. (D)
   Scatterplot depicting comparison of abundance changes between
   core-fucosylated sites and their corresponding proteins.

   Next, the total intensity of core-fucosylated sites was summed in each
   sample, and it was found that the total intensity of core-fucosylated
   sites was significantly higher in tumors (p = 0.037) ([139]Figure
   [140]3
   C). Among the 2494 core-fucosylated sites, 1660 sites corresponding to
   848 glycoproteins were quantified in at least 50% of samples in tumors
   ([141]Figure S14). In addition, 690 glycoproteins that contained 1448
   core-fucosylated sites were quantified in the proteomic data set.
   Interestingly, 1208 core-fucosylated sites (83.4%) had greater changes
   in core fucosylation abundance than those in the corresponding
   glycoprotein abundance ([142]Figure [143]3D). These results indicated
   that the level of core fucosylation upregulated in tumors, which
   accorded well with the upregulation of FUT8 in tumors (p = 0.019, fold
   change = 1.86) ([144]Figure S15).

   Then, we analyzed aberrant core-fucosylated sites in LUAD. It should be
   mentioned that the differences of PTM sites could be caused by the
   changes of either protein abundance or PTM occupation.^[145]39,[146]40
   Thus, to identify the relative changes in core-fucosylated sites
   arising from tumors regardless of total protein levels, we normalized
   the intensities of core-fucosylated sites to total protein levels
   ([147]Table S4). After normalization, we observed 196 upregulated and 6
   downregulated core-fucosylated sites in tumors relative to paired NATs
   (p < 0.05 and fold change >1.5) ([148]Figure [149]4
   A and [150]Table S5). Further pathway enrichment analysis by the Kyoto
   Encyclopedia of Genes and Genomes (KEGG) showed that extracellular
   matrix (ECM)-receptor interaction was significantly enriched by
   upregulated core fucosylation events ([151]Figure [152]4B). Remodeling
   of the ECM in tumors might be critical to supporting tumorigenesis and
   tumor progression.^[153]41−[154]43 Besides, upregulated
   core-fucosylated sites were also associated with the PI3K-Akt signaling
   pathway ([155]Figures [156]4B and[157]4C). Hyperactivation of the
   PI3K-Akt signaling pathway was associated with tumor progression,
   invasion, metastasis, and poor prognosis.^[158]44−[159]46
   Interestingly, we identified 10 putatively druggable glycoproteins with
   FDA-approved drugs, such as epidermal growth factor receptor (EGFR),
   mesenchymal-epithelial transition factor (MET), etc. ([160]Table S6).
   We also identified 2 putatively druggable glycoproteins (laminin alpha5
   (LAMA5) and laminin gamma 1 (LAMC1)) with known but as yet non-FDA
   approved molecules. It was worth noting that most core-fucosylated
   sites were localized in the extracellular domains of these
   glycoproteins, which played an important role in mediating the
   functions of proteins.^[161]47,[162]48 For example, the
   core-fucosylated site N196 on EGFR and N785 on MET were localized in
   the extracellular domains that were essential for the activation of
   EGFR and MET, respectively.^[163]48,[164]49 Furthermore, the
   core-fucosylated site N196 on EGFR was near the monoclonal antibody
   (such as cetuximab) binding sites.^[165]50,[166]51 In addition, 73
   upregulated core-fucosylated sites were observed on 64 secreted
   glycoproteins, among which 27 glycoproteins were annotated as plasma
   proteins ([167]Figure S16). These upregulated core-fucosylated sites
   might have the potential to serve as biomarkers for early detection of
   LUAD, e.g., VWF-N2585, TIMP1-N53, SERPINA1-N271, CP-N358, and CLU-N86
   ([168]Figure [169]4D). Taken together, these results suggested that
   core-fucosylated sites could serve as potential drug development
   targets and biomarker candidates, although the sample size was limited
   in this work. A larger cohort and various sample types (e.g., blood
   plasma) are needed to verify these findings.

Figure 4.

   Figure 4
   [170]Open in a new tab

   Comparative analysis of the core-fucosylated sites in tumors and the
   paired NATs. (A) Volcano plot of differentially regulated
   core-fucosylated sites between tumors and the paired NATs. (B) KEGG
   analysis of 196 upregulated proteins. (C) Heatmap showing the
   upregulated sites in the PI3K-Akt signaling pathway. Glycoproteins
   annotated as drug targets (as defined by https://go.drugbank.com/) are
   highlighted by vertical bars on the left side of the figure. (D) Box
   plots showing the relative abundance of upregulated sites in secreted
   proteins. The p value is calculated using student’s t test.

   Investigating the location and occupation of glycosylated sites on
   glycoproteins is important for understanding the biological functions
   of protein glycosylation. Here, our glycosylated probe 1 enabled
   efficient enrichment of core fucosylation in a “one-step” manner.
   Compared with previous work, this probe can be used to label cell
   lysates and human samples.^[171]22 A total of 1200, 1277, and 2494
   core-fucosylated sites were identified in H1299, Jurkat, and human LUAD
   samples using probe 1, respectively. There are many works on global
   analysis of core fucosylation events in cancer cell lines and complex
   human samples. These works mainly employed lectin binding, HILIC, or
   mixed-mode polymeric sorbent to enrich intact core-fucosylated peptides
   or glycopeptides, and then EndoF3 was used to simplify these intact
   glycopeptides.^[172]14,[173]16,[174]18,[175]19 There are also several
   works identified core-fucosylated sites by chemoenzymatic labeling
   methods.^[176]22,[177]23,[178]52 On average approximately 500
   core-fucosylated sites were identified in cancer cell lines, while
   about 1000 core-fucosylated sites were detected in complex human
   samples in these
   works.^[179]14,[180]16,[181]18,[182]19,[183]22,[184]23,[185]52 Though
   different cancer cell lines and specimens were used, we were able to
   identify a total of 1200, 1277, and 2494 core-fucosylated sites in
   H1299, Jurkat, and human LUAD, indicating the effectiveness and
   robustness of the current method. However, it is worth mentioning that
   EndoF3 shows the highest activity only toward core-fucosylated
   N-glycans with bi- and triantennary structures. Hence, this method
   cannot identify and quantify core-fucosylated sites with more complex
   N-glycan structures, such as tetra-antennary glycan structures.

Conclusion

   In summary, we have successfully developed a “one-step” strategy for
   the global analysis of core fucosylation. Taking advantage of the
   substrate specificity of wild-type EndoF3 and mutant endoglycosidase
   EndoF3-D165A, a “one-step” probe linked with the PNIPAM polymer can be
   specifically installed on the core fucose structure without further
   need for an additional chemical reaction. Owing to the thermosensitive
   property of PNIPAM, the labeled peptides can be captured from the
   complex samples. In addition, the traceless cleavage by EndoF3 makes it
   ideal for the site-specific analysis of core fucosylation by MS
   technology. The feasibility and effectiveness of our method were
   evaluated in the BSA spiked with standard core-fucosylated peptides,
   H1299, and Jurkat cell lines, and it showed a significant improvement
   over the “two-step” strategy. The “one-step” strategy was further
   employed to characterize core-fucosylated sites in human LUAD, and 196
   upregulated sites were identified in tumors. These upregulated sites
   might promote the discovery of new drug targets or diagnostic
   biomarkers. We anticipate that this work will accelerate the study of
   core fucosylation function in important physiological and pathological
   processes such as tumorigenesis.

Materials and Methods

Materials

   All chemical reagents were purchased from commercial sources and used
   without further purification unless otherwise stated. A
   poly(N-isopropylacrylamide) (PNIPAM) dendrimer (Mn ∼ 10 000) and bovine
   serum albumin (BSA) were purchased from Sigma-Aldrich.
   Biotin-PEG[4]-alkyne and THPTA were purchased from Click Chemistry
   Tools. DBCO acid was purchased from Bidepharm (Shanghai, China).
   Sep-Pak tC18 cartridges were from Waters. Trypsin (0.25% EDTA) was
   purchased from Yeasen Biotechnology. BCA protein assay kit and
   streptavidin agarose resin were from Invitrogen. Sequencing grade
   Modified Trypsin was purchased from Promega. An EDTA-free protease
   inhibitor cocktail was purchased from New Cell and Molecular (NCM)
   Biotech. Peptides were purchased from Biochem (Shanghai, China).

Cell Culture

   H1299 and Jurkat cells were cultured in the RPMI 1640 medium (Gibco)
   supplemented with 10% fetal bovine serum and 1% (v/v)
   penicillin/streptomycin. All cell lines were grown under a humidified
   atmosphere supplied with 5% CO[2] at 37 °C.

Enzyme Preparation

   EndoF3 and EndoF3-D165A from Elizabethkingia meningoseptica([186]53)
   were synthesized by GenScript (Nanjing, China). All genes were cloned
   into the pET-28a vector with six histidine tags (6xHis tag) for
   purification by Ni-affinity chromatography. The confirmed constructs
   were subsequently transformed into E. coli BL21 (DE3) for protein
   expression. BL21 (DE3) cells harboring a recombinant vector of pET-28a
   were cultured in 2 L of LB medium containing 50 ug/mL kanamycin in a
   rotary shaker at 37 °C and 200 rpm, and 0.2 mM IPTG was added until OD
   was 0.8. Protein expression was allowed to proceed at 16 °C overnight.
   The cells were harvested by centrifugation at 7000 rpm for 10 min. The
   cell precipitation was resuspended in lysis buffer (50 mM Tris-HCl
   buffer, 300 mM NaCl, 10 mM imidazole; pH = 7.5). Cells were disrupted
   by a microfluidizer, and the lysate was centrifuged at 12 000 g for 10
   min to remove the cell debris. The his-tagged proteins were purified
   using a Ni-NTA agarose column. Before purification, the column was
   equilibrated with the lysis buffer (50 mM Tris-HCl, 300 mM NaCl, 10 mM
   imidazole; pH = 7.5). The column was washed with 2 column volumes of
   the lysis buffer and eluted with elution buffer (50 mM Tris-HCl, 300 mM
   NaCl, 300 mM imidazole; pH = 7.5). The enzyme was desalted by
   filtration (Amicon Ultra-5, 10 K[d]). Protein concentration was
   determined by a BCA Protein Assay Kit.

Chemical Synthesis of 1 ([187]Scheme S1)

   11-((tert-Butoxycarbonyl)amino)undecanoic acid (6, 3012 mg, 10 mmol),
   2-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium,
   hexafluorophosphate, (HATU, 7600 mg, 20 mmol), and
   N,N-diisopropylethylamine (DIPEA, 5000 μL, 30 mmol) were added to the
   solution of prop-2-yn-1-amine (825 mg, 15 mmol) in DCM (50 mL). The
   resulting mixture was stirred at room temperature under a N[2]
   atmosphere for 4 h. Then, the solvent was removed by rotary
   evaporation. The solid was purified by dialysis (MWCO 1000) to obtain
   an alkyne PNIPAM product. the residue was purified by column
   chromatography on silica gel eluted with DCM/MeOH (50:1–20:1) to give
   brown oil 7 (2710 mg, yield: 80%). ^1H NMR (400 MHz, CDCl[3]) δ 5.90
   (s, 1H), 4.54 (s, 1H), 4.03 (dd, J = 4.7, 2.4 Hz, 2H), 3.07 (t, J = 6.5
   Hz, 2H), 2.32–2.06 (m, 3H), 1.66–1.55 (m, 2H), 1.51–1.37 (m, 11H),
   1.35–1.15 (m, 12H). ^13C NMR (101 MHz, CDCl[3]) δ: 172.95, 156.14,
   79.83, 79.13, 71.55, 40.76, 36.51, 30.14, 29.50, 29.41, 29.33, 29.29,
   29.20, 28.54, 26.84, 25.63 (one peak less because of overlapping).

   To a solution of compound 7 (2700, 8 mmol) in DCM (10 mL) was added
   trifluoroacetic acid (TFA, 10 mL), and the resulting mixture was
   stirred at room temperature for 1 h. After the completion of the
   reaction (monitored by MS), the solvent was removed by rotary
   evaporation. The solid was resuspended in Et[2]O (50 mL) and filtered.
   The solid was washed with Et[2]O and dried to give white solid 8 (1760
   mg, yield: 86%). ^1H NMR (400 MHz, DMSO) δ 8.22 (t, J = 5.5 Hz, 1H),
   7.62 (s, 2H), 3.82 (dd, J = 5.5, 2.5 Hz, 2H), 3.09 (t, J = 2.5 Hz, 1H),
   2.81–2.71 (m, 2H), 2.06 (t, J = 7.4 Hz, 2H), 1.58–1.41 (m, 4H), 1.24
   (s, 12H). ^13C NMR (101 MHz, DMSO) δ 171.92, 81.41, 72.78, 38.83,
   35.05, 28.8, 28.63, 28.52, 27.70, 27.01, 25.78, 25.14 (two peak less
   because of overlapping).

   11 was prepared from the poly(N-isopropylacrylamide) (PNIPAM) dendrimer
   (Mn ∼ 10 000, 9), amino-alkyne (8), and compound 10 according to the
   reported procedure.^[188]54 Amino-alkyne (8, 48 mg, 0.2 mmol),
   2-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium,
   hexafluorophosphate, (HATU, 152 mg, 0.4 mmol), and
   N,N-diisopropylethylamine (DIPEA, 105 μL, 0.6 mmol) were added to the
   solution of compound 9 (1000 mg, 0.1 mmol) in DMF (20 mL). The
   resulting mixture was stirred at room temperature under a N[2]
   atmosphere for 8 h. Then, the solvent was removed by rotary
   evaporation. The solid was resolved by water and purified by dialysis
   (MWCO ∼ 1000) to obtain an alkyne PNIPAM product. Compound 10 (207 mg,
   0.1 mmol), CuSO[4] (3.2 mg, 0.02 mmol), l-ascorbic acid sodium salt
   (VcNa, 19.8 mg, 0.1 mmol), and THPTA (17 mg, 0.04 mmol) were added to
   the solution of the alkyne PNIPAM compound in DMF (20 mL). The
   resulting mixture was stirred at room temperature until TLC detection
   (i-PrOH:NH[4]OH:H[2]O = 7:3:2, v:v:v) indicated the completion of the
   reaction. The resulting mixture was lyophilized, and the residue was
   purified by Sephadex G50 (water) to give glycosylated PNIPAM compound
   11 (583 mg, yield: 53%). The sugar loading efficiency is about 50%
   according to the calculation by NMR spectra. The hydrogen at the C1
   position on the sugar ring was falling into the signal-silent region.
   These characteristic peaks were used to calculate loading efficiency
   and are marked with red asterisks in the spectra.

   2-Chloro-1,3-dimethylimidazolinium chloride (DMC, 3.4 mg, 20 μmol) and
   Et[3]N (11 μL, 80 μmol) were added to a solution of compound 11 (45 mg,
   2 μmol) in water. The reaction mixture was incubated in an ice bath for
   30 min. Then, 4.0 M NaCl in water (100 μL) was added. The temperature
   of the solution was raised to 40 °C to precipitate compound 1. After
   centrifugation (10 000 × g), the oxazoline PNIPAM compound 1 was
   obtained as a white solid. 1 was used for enrichment without further
   purification. The hydrogen of the C1 position at 6.06 ppm (double)
   indicated the formation of the oxazoline ring.

   Compounds 2 and 3 were prepared according to the reported work.^[189]23

The Washing Process

   The precipitate was dissolved in water (600 μL) in an ice bath. Then,
   4.0 M NaCl in water (200 μL) was added. The mixture’s temperature was
   raised to 40 °C to precipitate the PNIPAM compounds. The solid was
   collected by centrifugation (10 000 × g).

Enrichment of Standard Peptides (4) with 1

   The standard peptide (4, 20 μg) was reacted with 1 in the presence of
   EndoF3-D165A (0.2 mg/mL) in PBS buffer (100 mM, pH = 7.5, 600 μL) at
   room temperature, and 1 (15 mg) was added every 1 h three times. Then,
   4.0 M NaCl in water (200 μL) was added. The mixture’s temperature was
   raised to 40 °C to precipitate the PNIPAM compounds. The supernatant
   and the solid were separated by centrifugation (10 000 × g). The
   supernatant was analyzed by HPLC to obtain a decrease of standard
   peptides (after enrichment with 1). After three washes (see [190]The
   Washing Process section), the precipitate was dissolved in a PBS buffer
   (100 mM, pH = 7.0, 600 μL) in an ice bath. Then, EndoF3 (0.1 mg/mL)
   were added. The mixture was reacted at room temperature for 12 h. Then,
   4.0 M NaCl in water (200 μL) was added. The mixture’s temperature was
   raised to 40 °C to precipitate the PNIPAM compounds. The supernatant
   was collected by centrifugation (10 000 × g) and analyzed by HPLC to
   determine the release of standard peptides (release of 1). The
   enrichment and release procedure were monitored at 214 nm by reverse
   phase HPLC. The gradient was set as follows: 10–18% B in 11 min, 18–31%
   in 9 min.

Enrichment of Standard Peptides (4) with 2 and 3

   Reactions were performed in PBS buffer (200 mM, pH = 7.5, 300 μL)
   containing the standard peptide (4, 0.5 mM) and EndoF3-D165A (0.2
   mg/mL) at 30 °C. Then, 2 (1 mM) wasadded to the reaction every 30 min
   five times. Then, the reaction was terminated and analyzed with HPLC to
   determine the formation of azido-peptides 5 (reacted with 2). The
   solution of 5 (corresponding to 20 μg of 4) was reacted with compound 3
   (45 mg) in PBS buffer (100 mM, pH = 7.5, 600 μL) at room temperature
   for 1 h. Then, 4.0 M NaCl in water (200 μL) was added. The mixture’s
   temperature was raised to 40 °C to precipitate the PNIPAM compounds.
   The supernatant and the solid were separated by centrifugation (10 000
   × g). The supernatant was analyzed by HPLC to obtain the decrease of
   azido-peptides (after enrichment with 3). After three washes (see
   [191]The Washing Process section), the precipitate was dissolved in PBS
   buffer (100 mM, pH = 7.0, 600 μL) in an ice bath. Then, EndoF3 (0.1
   mg/mL) were added. The mixture was reacted at room temperature for 12
   h. Then, 4.0 M NaCl in water (200 μL) was added. The mixture’s
   temperature was raised to 40 °C to precipitate the PNIPAM compounds.
   The supernatant was collected by centrifugation (10 000 × g) and
   analyzed by HPLC to determine the release of standard peptides (release
   of 3). The enrichment and release procedure were monitored at 214 nm by
   reverse phase HPLC. The gradient was set as follows: 10–18% B in 11 min
   and 18–31% in 9 min.

Enrichment of the Standard Peptide and BSA Peptide Mixture with 1, 2, and 3

   BSA (100 mg) in a NH[4]HCO[3] buffer solution (100 mM, 10 mL) was
   digested with trypsin at 37 °C for 24 h. After the completion of the
   reaction, the digested peptides were desalted by using a C18
   solid-phase column (Sep-Pak tC18 cartridges), lyophilized, and stored
   at −80 °C for further use. The concentration of digested peptides was
   determined by the BCA Protein Assay Kit.

Enrichment with 1

   The standard peptide (4, 20 μg) was mixed with digested BSA peptides
   (10 mg) in a ratio of 1:500 (w/w). The resulting peptides were reacted
   with enzymes and 1 (45 mg) according to the procedure described above.
   Then, 4.0 M NaCl (200 μL) was added. The mixture’s temperature was
   raised to 40 °C to precipitate the PNIPAM compounds. After six washes
   (see [192]The Washing Process section), the precipitate was dissolved
   in PBS buffer (100 mM, pH = 7.0, 600 μL) in an ice bath. Then,
   endoglycosidases (EndoF3, 0.1 mg/mL) were added. The mixture was
   reacted at room temperature for 12 h. Then, 4.0 M NaCl (200 μL) was
   added. The temperature of the mixture was raised to 40 °C to
   precipitate the PNIPAM compounds. The supernatant was collected by
   centrifugation (10 000 × g). The release of the standard peptides (4)
   in the supernatant was subjected to SDB-RPS solid-phase extraction
   column desalting (3 M Empore) and analyzed by LC-MS/MS.

Enrichment with 2 and 3

   The standard peptide (4, 20 μg) was mixed with digested BSA peptides
   (10 mg) in a ratio of 1:500 (w/w). The resulting peptides were reacted
   with enzymes and 2 to obtain the azido-peptides according to the
   procedure described above. The azido-peptides were subjected to C18
   solid-phase extraction desalting (Sep-Pak tC18 cartridges) and
   lyophilized. The azido-peptide solids were dissolved in PBS buffer (100
   mM, pH = 7.5, 600 μL) and reacted with alkyne PINPAM compounds (3, 45
   mg) at room temperature for 1 h. Then, 4.0 M NaCl (200 μL) was added.
   The mixture’s temperature was raised to 40 °C to precipitate the PNIPAM
   compounds. After six washes (see [193]The Washing Process section), the
   precipitate was dissolved in PBS buffer (100 mM, pH = 7.0, 600 μL) in
   an ice bath. Then, endoglycosidases (EndoF3, 0.1 mg/mL) were added. The
   mixture was reacted at room temperature for 12 h. Then, 4.0 M NaCl (200
   μL) was added. The temperature of the mixture was raised to 40 °C to
   precipitate the PNIPAM compounds. The supernatant was collected by
   centrifugation (10 000 × g). The release of the standard peptides (4)
   in the supernatant was subjected to SDB-RPS solid-phase extraction
   column desalting (3 M Empore) and analyzed by LC-MS/MS.

Enrichment of the Cellular Peptides with 1, 2, and 3

   Cells were routinely cultured in an RPMI 1640 medium, supplemented with
   10% FBS, penicillin (100 U/mL), and streptomycin (100 U/mL) in a
   humidified cell incubator at 37 °C under an atmosphere of 5% CO[2]. The
   cells were collected and lysed at 4 °C with RIPA lysis buffer (50 mM
   Tris, 150 mM NaCl, 0.1% (w/v) SDS, and pH 7.4) containing EDTA-free
   protease inhibitor cocktail (NCM) and 2% SDS for 20 min. The mixture
   was centrifuged at 4 °C to collect the supernatant. After reduction
   with dithiothreitol (DTT, 20 mM) at 37 °C for 2 h and alkylation with
   iodoacetamide (IAA, 40 mM) at 25 °C for 40 min in the dark, proteins
   were purified and pelleted with the methanol/chloroform/water (4:1:4)
   precipitation. The protein solid was dissolved and digested with
   sequencing grade modified trypsin (Promega, enzyme: substrate ratio of
   ∼1:100) in NH[4]HCO[3] buffer (100 mM containing 1 M urea) at 37 °C for
   24 h. After digestion, the peptides were subjected to C18 solid-phase
   extraction desalting (Sep-Pak tC18 cartridges) and lyophilized. The
   digested peptides (5 mg/mL) were dissolved in PBS buffer (100 mM, pH =
   7.5), and EndoF3 (0.1 mg/mL) were added. The mixture was reacted at 37
   °C for 2 h. Then, the deglycosylated peptides were subjected to C18
   solid-phase extraction desalting (Sep-Pak tC18 cartridges) and
   lyophilized. The deglycosylated peptides solid were dissolved in water
   and stored at −80 °C. The concentration of deglycosylated peptides was
   determined by BCA Protein Assay Kit.

Enrichment with 1

   Reactions were performed in PBS buffer (200 mM, pH = 7.5, 300 μL)
   containing 3 mg of deglycosylated peptides (10 mg/mL) and EndoF3-D165A
   (0.2 mg/mL) at room temperature. Then, 1 (15 mg) were added to the
   reaction every 1 h for three times. Then, 4.0 M NaCl (200 μL) was
   added. The mixture’s temperature was raised to 40 °C to precipitate the
   PNIPAM compounds. The precipitate was collected by centrifugation (10
   000 × g). After six washes (see [194]The Washing Process section), the
   precipitate was dissolved in PBS buffer (100 mM, pH = 7.0, 600 μL) in
   an ice bath. Then, endoglycosidases (EndoF3, 0.1 mg/mL) were added. The
   mixture was reacted at room temperature for 12 h. Then, 4.0 M NaCl (200
   μL) was added. The mixture’s temperature was raised to 40 °C to
   precipitate the PNIPAM compounds. The glycopeptides in the supernatant
   were collected by centrifugation (10 000 × g) and lyophilized.

Enrichment with 2 and 3

   Reactions were performed in PBS buffer (200 mM, pH = 7.5, 300 μL)
   containing 3 mg of deglycosylated peptides (10 mg/mL) and EndoF3-D165A
   (0.2 mg/mL) at room temperature. Then, 2 (1 mM) were added to the
   reaction every 30 min for five times. Then, the azido-peptides were
   subjected to C18 solid-phase extraction desalting (Sep-Pak tC18
   cartridges) and lyophilized. The azido-peptides were dissolved in PBS
   buffer (100 mM, pH = 7.5, 600 μL) and reacted with alkyne PINPAM
   compounds (3, 45 mg) at room temperature for 2 h. Then, 4.0 M NaCl (200
   μL) was added. The mixture’s temperature was raised to 40 °C to
   precipitate the PNIPAM compounds. The precipitate was collected by
   centrifugation (10 000 × g). After six washes (see [195]The Washing
   Process section), the precipitate was dissolved in PBS buffer (100 mM,
   pH = 7.0, 600 μL) in an ice bath. Then, endoglycosidases (EndoF3, 0.1
   mg/mL) were added. The mixture was reacted at room temperature for 12
   h. Then, 4.0 M NaCl (200 μL) was added. The mixture’s temperature was
   raised to 40 °C to precipitate the PNIPAM compounds. The glycopeptides
   in the supernatant were collected by centrifugation (10 000 × g) and
   lyophilized.

Enrichment of the Human Lung Adenocarcinoma (LUAD) Sample Peptides with 1

   Frozen tissues were homogenized in SDT lysis buffer (4% SDS (m/v), 100
   mM DTT, 100 mM Tris-HCl, pH = 7.6) (Homogenizer). Then, the proteins
   were heated at 95 ^◦C for 5 min and centrifuged at 12 000 g for 10 min.
   The supernatants were collected and the protein concentration was
   determined by tryptophan fluorescence emission as described
   previously.^[196]55 After alkylation with IAA (200 mM) at 25 °C for 40
   min in the dark, proteins digested with sequencing grade modified
   trypsin (Promega, enzyme: substrate ratio of ∼1:100) following the
   solvent precipitation SP3 (SP4) procedure.^[197]56 After digestion, the
   peptides were subjected to C18 solid-phase extraction desalting
   (Sep-Pak tC18 cartridges) and lyophilized. A portion of these dried
   peptides was used for whole proteome analysis.The digested peptides (5
   mg/mL) were dissolved in PBS buffer (100 mM, pH = 7.5), and EndoF3 (0.1
   mg/mL) were added. The mixture was reacted at 37 °C for 2 h. Then, the
   deglycosylated peptides were subjected to C18 solid-phase extraction
   desalting (Sep-Pak tC18 cartridges) and lyophilized. The deglycosylated
   peptide solids were dissolved in water and stored at −80 °C. The
   concentration of deglycosylated peptides was determined by the BCA
   Protein Assay Kit.

Enrichment with 1

   Reactions were performed in PBS buffer (200 mM, pH = 7.5, 300 μL)
   containing 3 mg of deglycosylated peptides (10 mg/mL) and EndoF3-D165A
   (0.2 mg/mL) at room temperature. Then, 1 (15 mg) were added to the
   reaction every 1 h for three times. Then, 4.0 M NaCl (200 μL) was
   added. The mixture’s temperature was raised to 40 °C to precipitate the
   PNIPAM compounds. The precipitate was collected by centrifugation (10
   000 × g). After six washes (see [198]The Washing Process section), the
   precipitate was dissolved in PBS buffer (100 mM, pH = 7.0, 600 μL) in
   an ice bath. Then, endoglycosidases (EndoF3, 0.1 mg/mL) were added. The
   mixture was reacted at room temperature for 12 h. Then, 4.0 M NaCl (200
   μL) was added. The mixture’s temperature was raised to 40 °C to
   precipitate the PNIPAM compounds. The glycopeptides in the supernatant
   were collected by centrifugation (10 000 × g) and lyophilized.

Sample Preparation for LC-MS/MS Analysis

   The dried glycopeptides were resuspended in 500 μL of 0.1% FA and
   fractionated with a self-made 10 μm C18 column. Phase A consisted of 10
   mM ammonium formate in water (pH = 10.0), and phase B consisted of 10
   mM ammonium formate in 90% ACN (pH = 10.0). 2 mg of 10 μm C18 resin
   were sonicated in ACN and transferred into a 200 μL tip column. The tip
   column was then sequentially equilibrated with ACN and 0.1% FA. The
   resuspended glycopeptides were then loaded onto the tip column. The tip
   column was then washed with 0.1% FA and phase A, respectively.
   Subsequent fractions were eluted sequentially using 200 μL of 3% phase
   B/97% phase A (v/v) (1), 6% phase B/94% phase A (v/v) (2), 9% phase
   B/91% phase A (v/v) (3), 12% phase B/88% phase A (v/v) (4), 15% phase
   B/85% phase A (v/v) (5), 18% phase B/82% phase A (v/v) (6), 21% phase
   B/79% phase A (v/v) (7), 25% phase B/75% phase A (v/v) (8), 30% phase
   B/70% phase A (v/v) (9), 35% phase B/65% phase A (v/v) (10), 50% phase
   B/50% phase A (v/v) (11), and 100% phase B (12). The fraction (1) was
   mixed with fraction (7), the fraction (2) was mixed with fraction (8),
   the fraction (3) was mixed with fraction (9), the fraction (4) was
   mixed with fraction (10), the fraction (5) was mixed with fraction
   (11), and the fraction (6) was mixed with fraction (12). These six
   fractions were evaporated with a Speed Vac. Finally, the fractionated
   peptides were subjected to a SDB-RPS solid-phase extraction column
   desalting (3 M Empore).

Nano-LC-MS/MS Data Acquisition for Core-Fucosylated Glycoproteome

   The LC-MS/MS analysis was performed using data-dependent acquisition
   (DDA) on a nanoflow UHPLC nanoElute system (Bruker Daltonics, Germany)
   coupled with a trapped ion mobility-quadrupole time-of-flight (timsTOF
   Pro) mass spectrometer (Bruker Daltonics, Germany). The purified
   peptides were resolved using 0.1% formic acid (FA) in water and
   separated on an analytical column (75 μm × 20 cm) packed with
   reverse-phase (RP) beads (1.5 μm ReproSil Saphir C18 beads, 100-Å pore
   size; Dr. Maisch GmbH, Ammerbuch, Germany) on the nanoElute with a 90
   min gradient. The flow rate was set as 250 nL/min from 0 to 85 min and
   300 nL/min from 85 to 90 min. The column was heated at 50 °C using a
   homemade column oven. Buffer A consisted of 0.1% (v/v) FA in water, and
   buffer B consisted of 0.1% (v/v) FA in 100% ACN. The gradient was set
   as follows: 2%–4% B in 1 min; 4%–26% B in 77 min; 26%–32% B in 5 min;
   32%–90% B in 2 min; 90% B in 5 min.

   Data acquisition on the timsTOF Pro was performed using timsControl
   4.0.5. The capillary voltage was set to 1500 V. The following
   parameters were adapted: MS range m/z 100 to 1700 was scanned in
   positive electrospray mode; the mass spectrometer collected ion
   mobility MS spectra over 1/k[0] of 1.51 to 0.60 (V·s)/cm^2, and then
   performed 10 cycles of parallel accumulation-serial fragmentation
   (PASEF) MS/MS with a target intensity of 20 000 and a threshold of
   2500. The ion accumulation time and ramp time were 100 ms each. The
   duty cycle was 100%, and total cycle time was 1.17 s. Singly charged
   precursor ions were excluded with a polygon filter. The collision
   energy was decreased from 60 eV at 1/k[0] = 1.50 (V·s)/cm^2 to 54 eV at
   1/k[0] = 1.17 (V·s)/cm^2 to 25 eV at 1/k[0] = 0.85 (V·s)/cm^2 and end
   at 20 eV at 1/k[0] = 0.60 (V·s)/cm^2.

Nano-LC-MS/MS Data Acquisition for LUAD Whole Proteome

   The LC-MS/MS analysis was performed using data-independent acquisition
   (DIA) on a nanoflow UHPLC nanoElute system (Bruker Daltonics, Germany)
   coupled with a trapped ion mobility-quadrupole time-of-flight (timsTOF
   Pro) mass spectrometer (Bruker Daltonics, Germany). The purified
   peptides were resolved using 0.1% FA in water and separated on an
   analytical column (75 μm × 20 cm) packed with RP beads (1.5 μm ReproSil
   Saphir C18 beads, 100-Å pore size; Dr. Maisch GmbH, Ammerbuch, Germany)
   on the nanoElute with a 60 min gradient. The flow rate was set as 300
   nL/min at 0–55 min and 300 nL/min at 55–60 min. The column was heated
   at 50 °C using a homemade column oven. Buffer A consisted of 0.1% (v/v)
   FA in water, and buffer B consisted of 0.1% (v/v) FA in 100% ACN. The
   gradient was set as follows: 2%–4% B in 1 min; 4%–26% B in 47 min;
   26%–32% B in 5 min; 32%–90% B in 2 min; 90% B in 5 min.

   Data acquisition on timsTOF Pro was performed using timsControl 4.0.5.
   The capillary voltage was set to 1500 V. The following parameters were
   adapted: the MS range of m/z 100 to 1700 was scanned in positive
   electrospray mode; the mass spectrometer collected ion mobility MS
   spectra over 1/k[0] of 1.29 to 0.76 (V·s)/cm^2. Samples were acquired
   with a diaPASEF^[199]57 method consisting of 14 cycles, including a
   total of 28 mass-width windows (25 Da width, from 452 to 1152 Da) with
   4 mobility windows each. The ion accumulation time and ramp time were
   100 ms each. The duty cycle was 100%, and the total cycle time was 1.59
   s.

Database Searching of MS Data

   All raw files acquired from LC-MS/MS were analyzed via MaxQuant
   v2.4.2^[200]32 against the human Swiss-Prot database containing 20 600
   sequences (downloaded in June 2021). Carbamidomethyl (cysteine) was
   selected as a fixed modification, while oxidation (methionine),
   acetylation (protein N-term), deamidation (asparagine), and fucosyl
   GlcNAc (asparagine) were set as variable modifications. A maximum
   number of 5 modifications were allowed for each peptide. Trypsin/P was
   selected as the digestive enzyme. The maximum missing cleavage site was
   set as 3, and the minimal peptide length was set at 7. The tolerances
   of the first search and main search for peptides were set at 20 and 4.5
   ppm, respectively. The false discovery rate (FDR) was set to 0.01 at
   the level of peptides, proteins, and sites.

   For LUAD whole proteome, the MS data were analyzed using DIA-NN 1.8
   against the human Swiss-Prot database (downloaded in June
   2021).^[201]58 Carbamidomethylation (C) was set as a fixed
   modification, and methionine excision (protein N-terminal) was searched
   for variable modifications. Trypsin/P was selected as the digestive
   enzyme with 2 maximum missing cleavages, and the minimal peptide length
   was set at 7. The false discovery rate (FDR) was set to 0.01 at the
   level of the peptides and proteins.

Bioinformatic Analysis

   Proteins labeled as either contaminants or reverses were removed from
   analysis. Proteins quantified in at least 50% of all 8 samples in LUAD
   were included for further analysis. For analysis of differentially
   expressed core-fucosylated sites, protein-normalized core-fucosylated
   sites were first calculated by normalizing core-fucosylated sites
   abundance to their respective protein levels in each sample, and
   core-fucosylated sites without a matching protein were removed. The
   cutoff criteria of differentially expressed core-fucosylated sites were
   set as followings: fold change ≥ 1.50 or ≤ 0.67 and p value < 0.05
   (Student’s t test).

   GO and KEGG enrichment was performed by Database for Annotation,
   Visualization, and Integrated Discovery (DAVID,
   https://david.ncifcrf.gov/).^[202]59 Sequence motif analysis was
   carried out using the R package “ggseqlogo.”^[203]60 Hierarchical
   clustering heatmaps were performed using the R package
   “ComplexHeatmap.” Other visualizations not specifically indicated were
   performed using the R package “ggplot2.”

Acknowledgments