Appropriate aglycone modification significantly expands the glycan substrate acceptability of α1,6-fucosyltransferase (FUT8)

The α1,6-fucosyltransferase, FUT8, is the sole enzyme catalyzing the core-fucosylation of N-glycoproteins in mammalian systems. Previous studies using free N-glycans as acceptor substrates indicated that a terminal β1,2-GlcNAc moiety on the Man-α1,3-Man arm of N-glycan substrates is required for efficient FUT8-catalyzed core-fucosylation. In contrast, we recently demonstrated that, in a proper protein context, FUT8 could also fucosylate Man 5 GlcNAc 2 without a GlcNAc at the non-reducing end. We describe here a further study of the substrate specificity of FUT8 using a range of N-glycans containing different aglycones. We found that FUT8 could fucosylate most of high-mannose and complex-type N-glycans, including highly branched N-glycans from chicken ovalbumin, when the aglycone moiety is modified with a 9-fluorenylmethyloxycarbonyl (Fmoc) moiety or in a suitable peptide/protein context, even if they lack the terminal GlcNAc moiety on the Man-α1,3-Man arm. FUT8 could also fucosylate paucimannose structures when they are on glycoprotein substrates. Such core-fucosylated paucimannosylation is a prominent feature of lysosomal proteins of human neutrophils and several types of cancers. We also found that sialylation of N-glycans significantly reduced their activity as substrate of FUT8. Kinetic analysis demonstrated that Fmoc aglycone modification could either improve the turnover rate or decrease the Km value depending on the nature of the substrates, thus significantly enhancing the overall efficiency of FUT8 catalyzed fucosylation. Our results indicate that an appropriate aglycone context of N-glycans could significantly broaden the acceptor substrate specificity of FUT8 beyond what has previously been thought. AutoFlex Mass Spectrometer (MALDI) with the N, N-dimethylaniline supplied 2,5-dihydroxybenzoic acid (DHB) matrix (47). Peaks were assigned by GlycoWorkbench. Kinetic analysis of the substrates using FUT8 enzyme. The relative enzyme activity of the substrates, A2Asn, A2Asn-Fmoc, GlcNAc 1 Man 5 GlcNAc 2 -Asn, GlcNAc1Man 5 GlcNAc 2 -Asn-Fmoc, Man 5 GlcNAc 2 -Asn and Man 5 GlcNAc 2 -Asn-Fmoc, were determined by titrating with wild type Fut8 enzyme purified as reported earlier (27) . Enzyme kinetics were performed using GDP-Glo TM Glycosyltransferase assay (Promega) for the substrates at a concentration range of 0-1mM (0-1.5 mM for Man 5 GlcNAc 2 -Asn) along with GDP-fucose (0.2 mM final concentration, pre-treated with Calf Intestinal alkaline-phosphatase (Promega)) as donor sugar (27). Reactions were carried out in a 10 μl reaction volume consisting of a universal buffer (200 mM each of Tris, MES, MOPS, pH 7.5) with the purified wild type enzyme at 37°C for 30 min. Reactions were stopped using 5 μl of GDP detection reagent and an equal volume of reaction mix in a polystyrene, white 384-well plate and incubating in dark for 1 h at room temperature. The luminescence values were measured using a GloMax Multi detection plate reader (Promega) and compared with a GDP standard curve to quantify the final released GDP product. The steady state parameters of K M , k cat , and k cat /K M values were determined using nonlinear curve fitting in GraphPad Prism 6 software. α1,6-fucosyltransferase; erythropoietin; granulocyte-macrophage ovalbumin;


Introduction
Protein glycosylation is a prevalent post-translational modification that regulates the structure and functions of proteins in different ways (1). Core-fucosylation, the addition of an α1,6-linked fucose at the innermost N-acetyl-glucosamine (GlcNAc) of N-glycans, plays essential roles in many biological recognition processes, including antibody's immune functions, cell adhesion, signal transduction, and tumor metastasis. For example, core-fucosylation regulates the activities of many cell surface receptors involved in ligand recognition and cell signaling (2)(3)(4)(5)(6)(7); elevated levels of core-fucosylation are often associated with cancers such as hepatocellular carcinoma, breast carcinoma, and lung squamous cell carcinoma (8)(9)(10)(11); and core fucosylation of Fc glycans adversely affects the affinity of antibodies for FcγRIIIa and, as a result, significantly reduces antibody-dependent cellular cytotoxicity (ADCC) (12,13).
The adverse effect of core-fucosylation on ADCC has prompted glycoengineering to produce low or nonfucosylated monoclonal antibodies as more effective therapeutics for the treatment of cancer (14,15).
The α1,6-fucosyltransferase, FUT8, is the sole enzyme responsible for the core-fucosylation of N-glycans in mammalian systems (16,17). Intensive previous studies have demonstrated that FUT8 has a strict acceptor substrate specificity and requires the presence of an unmodified β-1,2-linked GlcNAc moiety on the Man-α1,3-Man arm of the N-glycans for efficient core-fucosylation (16,(18)(19)(20)(21)(22)(23)(24)(25)(26)(27). Thus, lack of a terminal GlcNAc moiety on the Man-α1,3-Man arm, or modification of the GlcNAc residue by galactosylation, would block core-fucosylation of the N-glycans by FUT8 (23,24). Nevertheless, some glycan characterization from isolated tissues and recombinant glycoproteins expressed from mammalian cells has demonstrated the existence of core-fucosylated high-mannose type glycans (28,29). Such discrepancy implicates a broader substrate specificity of FUT8 towards natural glycoproteins. Recently, we have revisited the substrate specificity of FUT8 and have found that FUT8 could efficiently fucosylate Man 5 GlcNAc 2 glycan in cells and in vitro when the glycan is presented in the context of a suitable aglycone moiety (17,30). For example, the Man 5 GlcNAc 2 on human erythropoietin could be efficiently core-fucosylated by FUT8 in HEK293 GnTI -/cells when FUT8 is overexpressed, and a Man 5 GlcNAc 2 Downloaded from http://portlandpress.com/biochemj/article-pdf/doi/10.1042/BCJ20210138/905983/bcj-2021-0138.pdf by guest on 21 March 2021 Biochemical Journal. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BCJ20210138 glycan could be fucosylated by FUT8 in vitro when the aglycone portion is modified by a hydrophobic 9fluorenylmethyloxycarbonyl (Fmoc) moiety, even when both lack the free GlcNAc moiety on the Man-α1,3-Man arm (17,30). We describe here an expanded study of the substrate specificity of FUT8 with a range of high-mannose type and complex type N-glycans, in the context of a modified aglycone portion.
We found that FUT8 could fucosylate a wide range of high-mannose, paucimannose, and complex type N-glycans tested, including highly branched high mannose and complex type N-glycans from chicken ovalbumin, when the aglycone moiety is modified with an Fmoc tag or with a suitable peptide/protein moiety. The efficiency of the fucosylation of paucimannose structures within a glycoprotein context is relatively high. Such glycoforms are widely detected in lysosome enzymes from neutrophils and several types of cancer (31)(32)(33). In comparison, we found that terminal sialylation of N-glycans significantly reduced its substrate activity toward FUT8. Our experimental data indicate that an appropriate aglycone context could significantly enhance the substrate activity of N-glycans toward FUT8, suggesting that FUT8 has a much more broadened substrate specificity than what was previously implicated.

Results
FUT8 enzymatic activity towards various high mannose-type glycans in the context of glycoproteins. We have previously shown that FUT8 could efficiently fucosylate Man 5 GlcNAc 2 glycan on human EPO in HEK293 GnTI -/cells (17). To further analyze the enzymatic activity of FUT8 towards different highmannose type glycoforms of glycoproteins, we first engineered the HEK293T cell line by overexpression of FUT8 using lentiviral transduction. Then we transfected this cell line with the erythropoietin (EPO) and the granulocyte-macrophage colony-stimulating factor (GM-CSF) expressing plasmid DNA (17,30) in the presence of 1 μg/mL of kifunensine, an 1,2-mannosidase I inhibitor. A relatively low concentration of kifunensine was used in order to obtain a range of truncated high-mannose type glycans (Man 6-9 GlcNAc 2 ) on the glycoproteins (34). Three days after the transfection, the glycoproteins were purified, and the N-glycans were released by PNGase F treatment followed by MALDI-TOF MS analysis.
Peaks corresponding to fucosylated high mannose glycans (1403. 6 were found on both EPO and GM-CSF glycoprotein products following overexpression of FUT8 in the presence of kifunensine ( Figure 1). This result indicated that FUT8 could also fucosylate high-mannose type N-glycans, ranging from Man 5 GlcNAc 2 to Man 9 GlcNAc 2 , in the context of a suitable glycoprotein aglycone (EPO or GM-CSF). Interestingly, It appeared that more core-fucosylated products were formed in the case of smaller high-mannose N-glycans (Man 5 GlcNAc 2 , Man 6 GlcNAc 2 , and Man 7 GlcNAc 2 ) than those in the larger high-mannose N-glycans (Man 8 GlcNAc 2 and Man 9 GlcNAc 2 ) (Figure 1), suggesting that smaller high-mannose N-glycans are preferable substrates of FUT8 compared with the larger Man 8 GlcNAc 2 and Man 9 GlcNAc 2 glycoforms in the context of both human EPO and GM-CSF.
We also tested how FUT8 would act on the high-mannose glycoform of EPO in vitro. We prepared EPO with high-mannose glycoforms (EPO-HM) by transfecting wild type HEK293T cells under 1 μg/mL kifunensine. The MALDI-TOF MS characterization of released glycans showed that the recombinant EPO carried a mixture of high-mannose type N-glycans ranging from Man 5 GlcNAc 2 to Man 9 GlcNAc 2 ( Figure 2a). We then carried out in vitro enzymatic fucosylation experiments with EPO-HM using the conditions as previously described (23,30). After an overnight reaction, glycans were released from the glycoprotein by PNGase 1743.8) were 146, indicating the addition of a fucose on the glycan. We did not detect the fucosylation of the Man 9 GlcNAc 2 glycan under the condition. This result confirmed that FUT8, within the context of EPO, could core-fucosylate large high-mannose type N-glycans in vitro, but the largest Man 9 GlcNAc 2 glycan was clearly less active toward FUT8 than the smaller ones. These obeservations confirm that the protein context plays a role in the enzymatic activity of FUT8 towards N-glycan substrates. EPO provides an effective aglycone text that facilitates core-fucosylation of high-mannose type glycans. In contrast, we have previously shown that FUT8 could not fucosylate the high-mannose N-glycans on  These data demonstrated that FUT8 could indeed core-fucosylate a range of high-mannose N-glycans in the context of glycosylated EPO substrates both in vivo and in vitro, and that the smaller paucimannose glycans were better substrates than the larger high-mannose type glycoforms. On the other hand, we also evaluated the substrate activity of FUT8 towards the innermost GlcNAc monosaccharide of EPO after deglycosylation of the high-mannose EPO with the Arthrobacter endoglycosidase (Endo A). In contrast to EPO-HM and EPO-PM, no fucose transfer was detected on EPO-GlcNAc in which only the innermost GlcNAc was attached. The results suggest that even within the EPO glycoprotein context, the core GlcNAc alone is not a substrate for FUT8.  (27) and the presence of a bisecting GlcNAc moiety on the core -mannose was found to eliminate FUT8 activity (24). Here we hypothesized that a proper aglycone modification might broaden the acceptor preference of the enzyme and partially rescue the low activity of these N-glycans as acceptors for FUT8. We prepared a library of Fmoc-labeled N-glycans derived from chicken egg white ovalbumin (OVA) that is known to carry a range of high-mannose and highly branched complex-type N-glycans, including structures containing bisecting GlcNAc residues (35)(36)(37). To prepare the Fmoc-labelled Nglycans, egg white ovalbumin (OVA) was digested with pronase to generate the Asn-linked N-glycans (OVA-Asn), which was then reacted with 9-fluorenylmethyloxycarbonyl chloride (Fmoc-Cl) (38) to give a library of Fmoc-labeled chicken ovalbumin N-glycans (OVA-Asn-Fmoc, Figure 4a).
The reactivity of FUT8 on free Asn-linked and the Fmoc-labeled N-glycans from ovalbulmin was evaluated by incubation of the respective compounds with  Terminal sialylation significantly decreases the FUT8-catalyzed core-fucosylation of complex type Nglycopeptides. Our previous preliminary study has shown that FUT8 could fucosylate a complex-type biantennary N-glycan when it is present in the context of a suitable polypeptide (30). Terminal sialylation is an important modification on N-linked glycans. Thus, we sought to investigate whether terminal sialylation would affect the enzymatic activity of FUT8 on complex-type N-glycans in a peptide context.
For this purpose, we synthesized a sialylated HIV-1 glycopeptide and its corresponding asialylated glycopeptide (Figure 5a-b), using a chemoenzymatic method that we have described before (39,40). The FUT8-catalyzed enzymatic reaction on the two glycopeptide substrates was performed using GDP-Fuc as the donor substrate and a relatively high concentration (0.5 mg/ml) of FUT8. It was found that FUT8 could transfer core fucose to the asialo-glycopeptide (Figure 5c vs. 5d), albeit at a slow rate whereas only marginally detectable core-fucosylation was observed for the sialylated glycopeptide even after 3-day incubation (Figure 5e vs. 5f). The fucosylated V3-CT glycopeptide was isolated via AAL lectin affinity chromatography. The isolated product appeared as a single peak in reverse phase HPLC and its identity was confirmed by ESI-MS analysis. Our result clearly indicated that terminal sialylation significantly decreased the substrate activity of N-glycans in the FUT8-catalyzed reactions.

Kinetic analysis of the FUT8-catalyzed reaction with selectively aglycone-modified N-glycan substrates.
To determine the underlying mechanism of the aglycone-promoted enzymatic reactions, we performed kinetic analysis of FUT8-catalyzed fucosylation with selected N-glycans using the GDP-Glo assays (27 Structural insights of the aglycone enhancement. The structural basis for acceptor substrate recognition of FUT8 has been investigated in three recent X-ray crystallographic studies (25)(26)(27). Each of the studies has characterized interactions between FUT8, and the donor analog, GDP, as well as the A2-Asn acceptor. carbonyl oxygen of Gly217. However, the remainder of the Asn residue extended into the solvent with no obvious adjacent hydrophobic or highly polar regions to interact with an aglycone substituent (27). Thus, these X-ray crystallographic studies provided little information on the potential contributions of the aglycone portions to the enzymatic recognition and catalytic efficiency of FUT8.

Discussion
In this study, the acceptor substrate specificity of FUT8 was further evaluated with a library of aglycone- presents a much more broadened specificity than what was previously indicated, particularly for substrates with an extended glycone substituent.
FUT8 was observed to be efficient in fucosylating paucimannose glycans in a glycoprotein context. Corefucosylated paucimannosylation has been widely detected in lysosomal enzymes from human neutrophils, key granulocytic cells of the innate immune system (33). It is reported that neutrophil elastase modulates the immune function of cells through its core-fucosylated paucimannosylation (31). Even the released core-fucosylated paucimannose glycan from elastase could directly inhibit the growth of Pseudomonas aeruginosa, a clinically relevant bacterium, in sub-micromolar concentration (41). Our observation can provide a feasible synthetic route to prepare neutrophil lysosome enzymes with core-fucosylated paucimannosylation for functional studies. In addition to their presence on human neutrophil elastase, core-fucosylated paucimannosylation has been observed and is regarded as a signature of several types of cancers (32,33). While it is conceivable to think that the core-fucosylated paucimannose glycoforms are generated by the trimming of the core-fucosylated complex type or hybrid type glycoforms in lysosomes after they leave the Golgi (33), our experimental data imply that a GnTI-independent direct core- Taken together, our experimental data suggest that an appropriate aglycone context could significantly enhance the core-fucosylation of N-glycans by FUT8, and that FUT8 has a more broadened substrate specificity than was previously appreciated. It is expected that site-directed mutagenesis or directed evolution could potentially improve enzymatic activity toward aglycone-modified acceptors and further broaden the substrate specificity of FUT8, thus providing an efficient tool for direct core-fucosylation of a range of N-glycans or glycoproteins.

Materials and Methods
Materials and reagents. The FUT8 expression constructs were generated by inserting the FUT8 coding region (residues 41-575, Uniprot Q9BYC5) into the pGEN2-DEST vector with an 8×His-tag and a superfold GFP fusion sequence at the N-terminus. The enzyme was expressed in HEK293 cells (293-F cells, Thermo Fisher) and purified according to previously described procedures (42). The EndoM N175Q enzyme was prepared according to the previously reported procedure (43).
Cell strains and culture. The FUT8 overexpression cell line (FUT8 + ) was generated from HEK 293T by lentiviral transduction (17). The stable cell line was created by infecting HEK 293T cells with lentivirus followed by selecting with 1 μg/mL of puromycin overnight. Both the HEK 293T cells and their derivatized cells were grown in suspension culture supplied with the serum-free FreeStyle™ F17 Expression Medium (ThermoFisher Scientific), with shaking at 150 rpm/min, at 37 °C with 8% CO2.
Cells were subcultured every 3 days with a seeding density of 4x10 5 cells/mL. were purified by HisTrap™ HP Column (GE Healthcare Life Sciences). The N-glycan of both proteins were released by PNGase F treatment and analyzed by MALDI-TOF MS analysis, following the previously reported procedure (17). Preparation of V3 glycopeptides as FUT8 substrates. The HIV-1 V3 GlcNAc-Peptide, (2.0 mg, 0.48 μmol), which was synthesized as described previously (39,40), was incubated together with SCToxazoline (3.8 mg, 1.9 mmol) and EndoM N175Q (7 μg Data availability statement. All data that support the findings of this study are contained within the manuscript

Conflict of interest
The authors declare that they have no conflicts of interest with the contents of this article. and P41GM103390, P01GM107012, and R01GM130915 to K.W.M.).          Table 1