A Robust α-l-Fucosidase from Prevotella nigrescens for Glycoengineering Therapeutic Antibodies

Eliminating the core fucose from the N-glycans of the Fc antibody segment by pathway engineering or enzymatic methods has been shown to enhance the potency of therapeutic antibodies, especially in the context of antibody-dependent cytotoxicity (ADCC). However, there is a significant challenge due to the limited defucosylation efficiency of commercially available α-l-fucosidases. In this study, we report a unique α-l-fucosidase (PnfucA) from the bacterium Prevotella nigrescens that has a low sequence identity compared with all other known α-l-fucosidases and is highly reactive toward a core disaccharide substrate with fucose α(1,3)-, α (1,4)-and α(1,6)-linked to GlcNAc, and is less reactive toward the Fuc-α(1,2)-Gal on the terminal trisaccharide of the oligosaccharide Globo H (Bb3). The kinetic properties of the enzyme, such as its Km and kcat, were determined and the optimized expression of PnfucA gave a yield exceeding 30 mg/L. The recombinant enzyme retained its full activity even after being incubated for 6 h at 37 °C. Moreover, it retained 92 and 87% of its activity after freezing and freeze-drying treatments, respectively, for over 28 days. In a representative glycoengineering of adalimumab (Humira), PnfucA showed remarkable hydrolytic efficiency in cleaving the α(1,6)-linked core fucose from FucGlcNAc on the antibody with a quantitative yield. This enabled the seamless incorporation of biantennary sialylglycans by Endo-S2 D184 M in a one-pot fashion to yield adalimumab in a homogeneous afucosylated glycoform with an improved binding affinity toward Fcγ receptor IIIa.


Expression and purification of recombinant α-L-fucosidases
The coding sequence of α-L-fucosidase was amplified using Polymerase Chain Reaction (PCR) with the Fusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific Baltics, Lithuania), following the manufacturer's recommendations.It was then cloned into the expression vector after linearization with appropriate restriction enzymes, which were purchased from New England BioLabs (NEB, UK).The resulting expression vectors were transformed into E. coli DH5α strain (Yeastern Biotech, Taiwan) for plasmid preservation and subsequently into the E. coli BL21 strain (Yeastern Biotech, Taiwan) for recombinant protein expression.Transformants were selected from Luria-Bertani (LB) medium agar plates supplemented with 100 μg/mL ampicillin.
For recombinant enzyme production, a single bacterial colony was inoculated into LB medium supplemented with ampicillin and cultured at 37°C overnight.The next day, fresh LB medium was inoculated with the overnight culture to get a starting optical density (OD600) of 0.1 and was cultured at 37°C until the OD600 reached between 0.6 and 0.8.The culture was then induced with 0.5 mM isopropyl 1-thio-β-D-galactopyranoside (IPTG, AK Scientific, US) and incubation continued at 16°C for an additional 16 hours.Cells were collected by centrifugation at 6000 g for 5 min.The resulting cell pellet was resuspended in binding buffer for protein purification, and cell lysis was performed with the One Shot Cell Disruptor equipment (Constant Systems, UK).
The purification of recombinant α-L-fucosidases was carried out by affinity chromatography with an AKTA Purifier FPLC System (GE Healthcare, Sweden).Histagged recombinant enzymes were purified with a HisTrap FF column (Cytiva, Sweden) and a phosphate binding buffer (20 mM sodium phosphate, 500 mM NaCl and 20 mM imidazole, pH 7.4).After the sample injection, the column was washed with 15 column volumes (CV) of binding buffer and subsequently eluted with 20 CV of elution buffer (20 mM sodium phosphate, 500 mM NaCl and 500 mM imidazole, pH 7.4) with a linear gradient from 0 to 100%.For MBP-fusionned proteins, purification was performed with an MBPTrap HP column (Cytiva, Sweden).The binding buffer contained 20 mM Tris-HCl, 200 mM NaCl and 1 mM Ethylenediaminetetraacetic acid (EDTA) at pH 7.4, and the elution buffer was prepared with the same buffer supplemented with 10 mM maltose.The purity of proteins in the elution fractions were verified by SDS-PAGE analysis using the TGX Stain-Free FastCast Acrylamide Kit (Bio-Rad, US).Fractions showing a strong band of the target protein were gathered and concentrated using an Amicon Ultra Centrifugal Filter Unit (Merck, Germany).Following buffer exchange with 50 mM sodium phosphate buffer (pH 6.5), the protein concentration of the purified enzyme was determined by the Bradford assay 1 using Bio-Rad Protein Assay (Bio-Rad, US), with bovine serum albumin (BSA) serving as the standard.

Enzymatic activity experiment
The enzyme activity of α-L-fucosidases was determined with 0.03 ug of purified Δ20PnfucA and 2 mM of the artificial substrate p-nitrophenyl-α-L-fucopyranoside (pNPfucose, Carbosynth, UK).The reaction was carried out in 50 mM sodium phosphate buffer (pH 6.5) for 5 min at 37°C.The hydrolysis product, pNP, was detected by spectrophotometry using a BioDrop Duo+ (Biochrom, UK) at OD405.Its concentration was quantified using a standard curve generated from serially diluted pNP solutions, ranging from 0.016 to 0.5 mM.One enzyme unit (U) was defined as 1 mol of pNP released per minute.
The effects of pH were determined using 50 mM sodium acetate (pH 4 to 5.5), sodium phosphate (pH 6 to 7.5) and Tris (pH 8 to 9.5) buffers.The effects of temperature were determined across a range of temperatures from 20 to 70°C.Thermostability was determined by incubating 0.045 ug of purified enzyme at temperatures ranging from 25 to 70°C before testing the enzymatic activity.Relative activity was determined in comparison with the highest activity value measured among different temperatures or pH.
The effects of divalent ions were determined in 50 mM sodium phosphate buffer (pH 7) containing 5 mM various divalent ions (Mg 2+ , Co 2+ , Ni 2+ , Cu 2+ , Zn 2+ or Ca 2+ ) or EDTA.Relative activity was determined in comparison with the enzyme activity measured in the sample without addition of divalent ions (i.e.buffer only).

The effects of incubation time was determined by incubating the purified
Δ20PnfucA in 50 mM sodium phosphate buffer (pH 7) up to 6 h.Relative activity was determined in comparison with the time point zero (i.e.t = 0 h).
To examine the enzyme's stability, the purified Δ20PnfucA, maintained at a concentration of 2 mg/mL in sodium phosphate buffer (pH 7), underwent various storage conditions: 1) Stored at 4°C for 28 days.2) Stored at -20°C for 28 days, followed by direct enzyme assay after thawing.3) It underwent freeze-drying before being stored at 4°C for 28 days.Relative activity was determined in comparison with the enzyme activity measured with freshly purified enzyme on day 1.
The kinetics study was conducted with pNP-fucose concentrations ranging from 0.0625 to 16 mM.The Km and Vmax values were determined through nonlinear curve fitting utilizing the Michaelis-Menton model in Prism 10 (GraphPad, US).

Protein deglycosylation experiment
The deglycosylation by endo-β-N-acetylglucosaminidase (Endo-S2) was performed with 50 μg protein substrate and 5 μg Endo-S2 wild-type or mutant D184M enzymes in 50 mM sodium phosphate buffer (pH 7) at 37 °C for 3 hours.The complete removal of N-glycans was performed with 10 μg protein substrate and the Peptide-N-Glycosidase F (PNGase F, NEB) according to the manufacturer's recommendation.Fetuin from fetal bovine serum and BSA served as control proteins.

Endo-S2 mutant D184M glycosynthase activity experiment
To evaluate the glycosyntase efficiency of Endo-S2 mutant D184M, the hydrolysis reaction was monitored by incubating 100 g original adalimumab (HumOri) with 1 g purified Endo-S2 mutant in a Tris buffer (50 mM Tris-HCl, 50 mM NaCl and 1 mM CaCl2, pH 7.6) at 37 °C for 1 h, while the transglycosylation rate was determined by incubating 100 g deglycosylated adalimumab with 1 g purified Endo-S2 mutant in the presence of 10 mM sialyl-complex-type N-glycan-oxazoline (SCT-ox) under the same reaction conditions.To evaluate the SCT-ox concentration effects on the transglycosylation efficiency, 100 g HumOri was treated with 10 g Endo-S2 mutant D184M and 10 g Δ20PnfucA in 50 mM Tris buffer (supplemented with 50 mM NaCl and 1 mM CaCl2, pH 7.6) at 37 °C for 3 h.Then, the transglycosylation reaction was initiated by adding 0.1, 1 or 10 mM SCT-ox in the reaction mix, and the incubation continued at 37 °C for 6 h.The reaction products were taken after 1, 2, 4 and 6 hours, purified by Protein A chromatography, and then digested by thermolysin treatmenment.The N-glycan composition was determined by LC-ESI-MS analysis.

Determination of glycan-engineered adalimumab binding to FcRIIIa
The binding activity was determined by ELISA assay.The FcRIIIa V158 (GST-Fusion, AB Bioscience) (0.5g/mL) was coated on a 96F MaxiSorp ELISA plate (NUNC) at 4 °C overnight in 100 L of coating buffer composed of 50 mM carbonate-bicarbonate (pH 9.6).Then, the plate was washed three times with a washing buffer composed of PBS with 0.05% Tween 20 (PBST), and incubated at room temperature for 2 h after the addition of 200 L of blocking buffer (2% BSA in PBS buffer).After the blocking step, the plate was washed with PBST before adding 200 L of original adalimumab (HumOri), deglycosylated and afucosylated adlimumab (HumGlcNAc) and homogeneous and afucosylated SCT-glycan adalimumab (HumSCT).Those sample were serially diluted in assay buffer (1% BSA in PBS) at concentrations ranging from 4000 to 0.49 ng/mL, and the plate was incubated at 4 °C overnight.On the third day, after washing the plate three times with PBST, 100 L of Goat (Fab')2 anti-human IgG Fcγ-HRP (Jackson ImmunoResearch Laboratories) diluted 1:10000 in assay buffer were added per well, and the plate was incubated at room temperature for 1 h.After a last washing step, 100 L of TMB ELISA substrate (Abcam) were added.After at least 5 min in the dark, the color formation was stopped with 100 L of stop solution (2N H2SO4), and the ELISA plate was read with a spectrophotometer (Varioskan Lux, Thermo Scientific) for the absorbance at OD450.The half maximal effective concentration (EC50) values were determined through nonlinear curve fitting utilizing the Sigmoidal 4PL model in Prism 10 (GraphPad, US).The phylogenetic tree was constructed using the Neighbor-Joining method in MEGA 11.

Supplement Figure 1
Fucosidases that exhibit linkage selectivity towards α(1,3)and α(1,4)-linked substrates were classified in the GH29B subgroup, while those with broad substrate specificity were classified in the GH29A subgroup.The bacterial GH 95 α-L-fucosidase from Bifidobacterium longum strain ATCC 15697 was used as the outgroup.All enzyme sequences were selected from CAZy 2 , NCBI and PDB databases (listed in Table S3).

Supplement Table 3
Table S3: List of α-L-fucosidases mentioned in the Figure S1.

Figure
Figure S8: LC-MS analysis of fucosylated substrates hydrolyzed by 20PnfucA.

Figure S11 :
Figure S11: Moniting the transglycosylation products (HumSCT) in the presence of different