Received: 23 March 2021 | Revised: 27 January 2022 | Accepted: 28 January 2022 DOI: 10.1002/bit.28051 ARTICLE Differential N‐ and O‐glycosylation signatures of HIV‐1 Gag virus‐like particles and coproduced extracellular vesicles Jesús Lavado‐García1 | Tao Zhang2 | Laura Cervera1 | Francesc Gòdia1 | 2 Manfred Wuhrer 1 Grup d'Enginyeria Cellular i Bioprocessos, Departament d'Enginyeria Química, Biològica i Abstract Ambiental, Escola d'Enginyeria, Universitat Human immunodeficiency virus 1 (HIV‐1) virus‐like particles (VLPs) are nanos- Autònoma de Barcelona, Barcelona, Spain 2 Center for Proteomics and Metabolomics, tructures derived from the self‐assembly and cell budding of Gag polyprotein. Leiden University Medical Center, Leiden, The Mimicking the native structure of the virus and being noninfectious, they represent Netherlands promising candidates for the development of new vaccines as they elicit a strong Correspondence immune response. In addition to this, the bounding membrane can be functionalized Jesús Lavado García, Grup d'Enginyeria with exogenous antigens to target different diseases. Protein glycosylation depends Cel·lular i Bioprocessos, Escola d'Enginyeria, Universitat Autònoma de Barcelona, Campus strictly on the production platform and expression system used and the displayed de Bellaterra, Cerdanyola del Vallès, Barcelona glycosylation patterns may influence downstream processing as well as the immune 08193, Spain. Email: jesus.lavado@uab.cat response. One of the main challenges for the development of Gag VLP production bioprocess is the separation of VLPs and coproduced extracellular vesicles (EVs). In Funding information this study, porous graphitized carbon separation method coupled with mass spec- European Molecular Biology Organization, Grant/Award Number: Short‐Term Fellowship trometry was used to characterize the N‐ and O‐ glycosylation profiles of Gag VLPs Number 8613; “la Caixa” Foundation, produced in HEK293 cells. We identified differential glycan signatures between Grant/Award Number: LCF/BQ/ES17/ 11600003 VLPs and EVs that could pave the way for further separation and purification strategies to optimize downstream processing and move forward in VLP‐based vaccine production technology. KEYWORDS HIV‐1, N‐glycans, O‐glycans, porous graphitized carbon, vaccine, virus‐like particles 1 | INTRODUCTION transport regulation, pathogen recognition, immune responses, and allergic reactions (Johannssen & Lepenies, 2017; Varki, 2017; Protein glycosylation represents a crucial feature in the design and Vigerust & Shepherd, 2007). Therefore, the design and character- development of new vaccines. Glycans play essential roles in cellular ization of the different glycoforms present in a biopharmaceutical processes such as signaling, cell‐to‐cell communication, vesicle product have tremendous impact in its development and further Abbreviations: APC, antigen‐presenting cells; CHO, Chinese hamster ovary; CLR, C‐type lectin receptor; EIC, extracted ion chromatogram; EV, extracellular vesicle; Gag::eGFP, translational fusion of HIV‐1 Gag polyprotein and enhanced GFP; GlcNAc, N‐acetylglucosamine; GnT‐III, N‐acetylglucosaminyltransferase‐III; HEK, human embryonic kidney; Hex, hexose; HexNAc, N‐acetylhexosamine; HIV‐1, human immunodeficiency virus 1; LC‐MS/MS, liquid chromatography coupled to tandem mass spectrometry; Man, mannose; NeuAc, N‐acetylneuraminic acid; NeuGc, N‐glycolylneuraminic acid; NTA, nanoparticle tracking analysis; PBS, phosphate‐buffered saline; PEI, polyethylenimine; PGC, porous graphitized carbon; RFU, relative fluorescence unit; sEV, small extracellular vesicles; VLP, virus‐like particle. This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. © 2022 The Authors. Biotechnology and Bioengineering published by Wiley Periodicals LLC Biotechnol Bioeng. 2022;119:1207–1221. wileyonlinelibrary.com/journal/bit | 1207 1208 | LAVADO‐GARCÍA ET AL. safety, efficacy, tolerability, and bioavailability optimiza- cell membrane glycoproteins. Understanding the glycosylation dis- tion (Duivelshof et al., 2019; Yehuda & Padler‐Karavani, 2020). played in VLPs produced in HEK293 cells could help assess and study Harboring the correct glycosylation of biopharmaceuticals like their immunogenic potential. Moreover, characterizing the glycan monoclonal antibodies has been demonstrated to improve the sta- profile of Gag VLPs produced in HEK293 cells could offer relevant bility, solubility and it is important for safety and efficacy (Gupta & information that could be used to design purification strategies and Shukla, 2018; Mastrangeli et al., 2018). In the case of influenza viral new methods to separate VLPs and extracellular vesicles (EVs), a particles, high titers of IgG were produced when HA proteins carried current remaining challenge (Effio & Hubbuch, 2015; Seifalian high mannose glycans while HA inhibition and neutralizing antibody et al., 2010). titers were higher with complex glycans (Lin et al., 2013). The use- In this study, we report on the characterization of the N‐ and fulness of polysaccharides in eliciting a complete immune response is O‐glycan profile of Gag VLPs produced in HEK293 cells. To do this, widely valued in conjugate vaccines where they are used to interact the analysis by porous graphitized carbon (PGC) liquid chromato- directly with B cells through a T cell‐independent pathway to produce graphy coupled to tandem mass spectrometry (LC‐MS/MS) has been a rapid and restricted antibody response (Goldblatt, 2000). Thus, selected due to the advantages of the method. PGC‐LC‐MS/MS is future glycoengineered viral particles could lead the way to self‐ able to separate structural and linkage isomers, providing a detailed adjuvanted viral vaccines (Schön et al., 2020). However, the most and specific glycosylation fingerprint (Jensen et al., 2012; important element representing implications in vaccine development Stavenhagen et al., 2014; Thomsson et al., 2010; Zhang et al., 2020). is choosing the best‐suited cell‐based platform for viral particle The insight of these analyses may help further optimize the design of production, one of the most relevant challenges nowadays. Viruses VLP‐based vaccines via glycoengineering approaches and provide depend almost entirely on the host cell glycosylation machin- new ideas for purification steps that could pave the way towards ery (Watanabe et al., 2019). Thus, the recombinant production of better bioprocesses, vaccines, and therapies. viral particles based on viral component, such as human im- munodeficiency virus 1 (HIV‐1) Gag‐based virus‐like particles (VLPs), relies on the host cell biosynthesis pathways for post‐translational 2 | EXPERIMENTAL SE CTION modifications (Roldão et al., 2010). Different expression systems like plant or insect cells are among the most used platforms for viral 2.1 | HEK 293 MAMMALIAN CELL LINE, vaccine production due to high titers (Cervera et al., 2019) while CULTURE CONDITIONS producing glycans that are substantially different from human gly- cosylation, influencing the immunological properties of the sub- The cell line used in this study is a serum‐free suspension‐adapted sequent viral particles (Lancaster et al., 2016). Although displaying HEK 293 cell line (HEK293SF‐3F6, NRC) kindly provided by Prof. specific exogenous glycan forms, like fungal glycans, can activate Amine Kamen from McGill University (Montreal, Canada). Cells were C‐type lectin receptors (CLRs) and promote an immune re- cultured in disposable polycarbonate 250 ml flasks with vent cap sponse (Lepenies et al., 2013), the use of biopharmaceuticals pro- (Corning®) at 37°C, 5% of CO2 and 85% RH at 130 rpm in an LT‐X duced in nonhuman expression systems for clinical applications Kuhner shaker (LT‐X Kuhner). Cell culture media was HyCell™ requires special attention. Exogenous immunologically active glycans TransFx‐H media from HyClone™ (GE Healthcare) supplemented may cause allergy reactions, therapy rejection, or other different with 4 mM GlutaMAX™ (Gibco, Life Technologies, ThermoFisher) and side effects (Md et al., 2021; Tretter et al., 1993). Still within mam- 0.1% Pluronic™ F‐68 nonionic surfactant (Gibco, Life Technologies). malian cell systems, nonhuman platforms like Chinese hamster Cell concentration and viability were determined using the ovary (CHO) cells present complex N‐glycans capped with NucleoCounter®NC‐3000 automatic cell counter (Chemometec) ac- N‐glycolylneuraminic acid (NeuGc), largely absent in human cells cording to manufacturer's instructions. (Dhar et al., 2019; Ghaderi et al., 2010). Also, human cell lines can generate very different glycosylation patterns depending on their progeny, culture conditions, or immortalization methods (Colgrave 2.2 | TRANSIENT TRANSFECTION et al., 2012). Altogether these considerations make the glycomic characterization of the biotherapeutic derived from the chosen pro- Transfections were carried out at a cell density of 2 × 106 cells/ml duction platform a requirement for further clinical applications. In the using a final DNA concentration of 1 µg/ml. PEI/DNA complexes case of HIV‐1 Gag VLPs, the human embryonic kidney (HEK) cell line were formed by adding PEI to plasmid DNA diluted in fresh culture is considered as a reference for its production, since they are easily medium (10% of the total culture volume to be transfected). transfected and can be cultured in suspension, resulting in good Transfection reagent PEIpro® (Polyplus‐transfection) was used. scalability (Ausubel et al., 2012; Lavado‐García, Cervera, et al., 2020). Briefly, the corresponding plasmid was diluted with supplemented VLPs are nanostructures resembling the nature confirmation of the HyCell™ culture media and vortexed for 10 s. Then PEI was added in virus but lacking genetic information, hence they are noninfective 1:2 (w/w) DNA:PEI ratio and vortexed three times, incubated for while carrying viral antigens to elicit an immune response. Gag VLPs 15 min at room temperature (RT) and then added to the cell culture. are produced via budding from the cell membrane, carrying the host The plasmid used encoded an HIV‐Gag protein fused in frame to LAVADO‐GARCÍA ET AL. | 1209 eGFP (Gag::eGFP). As a transfection control, a plasmid sharing the 0.45 μm, Millipore) was preconditioned with three times 200 μl 70% same backbone but lacking the Gag::eGFP gene was used and noted ethanol followed by equilibration three times with 200 μl water. as mock. Previously prepared lysates as well as intact EVs and VLPs were applied to the PVDF membrane directly. Purified fetuin glycoprotein (Sigma‐Aldrich) was used as a standard and also directly applied to 2.3 | HIV‐1 GAG VLP QUANTIFICATION BY the PVDF membrane followed by horizontal shaking for 30 min. FLUORIMETRY Protein denaturation was achieved by applying 75 μl of denaturation mix, 72.5 μl 8 M GuHCl (ThermoFisher Scientific), and 2.5 μl 200 mM The extracellular concentration of HIV‐1 Gag VLPs was assessed by DTT (Sigma‐Aldrich) in each well, followed by 10 min shaking and fluorimetry using an ad hocdeveloped and validated quantification incubating at 60°C in a moisture box for 30 min. Subsequently, the assay (Gutiérrez‐Granados et al., 2013). VLP‐containing supernatants unbound material was removed by washing three times with water were recovered by cell culture centrifugation at ×1000g for 5 min. each time followed by centrifugation at ×500g for 1 min. Any residual Relative fluorescence unit values were calculated by subtracting liquid was subsequently removed from the membrane by pipetting fluorescence unit values of nontransfected negative control samples. and discarded. 2.4 | ULTRACENTRIFUGATION 2.7 | ENZYMATIC N‐GLYCAN RELEASE, REDUCTION, AND PURIFICATION VLP‐containing supernatants were recovered by cell culture cen- trifugation at ×1000g for 5 min. Then, concentrated and purified N‐glycan release was performed as previously described (Holst HIV‐1 Gag VLPs were obtained by single sucrose cushion ultra- et al., 2017) with small modifications. PNGase F (Roche Diagnostics, centrifugation. Briefly, a volume of 35 ml of clarified supernatant 2 U of enzyme diluted with water up to 15 μl) was added to each well from HEK293 transiently transfected cell culture was layered on top followed by 10 min shaking at RT. The plate was incubated in a hu- of a 30% sucrose cushion of 3 ml and centrifuged at 31,000 rpm for midified box for 15 min at 37°C. Subsequently, another 15 μl of water 2.5 h at 4°C using an SW32 rotor in a Beckman Optima L100XP was added in each well, including 10 ng of maltoheptaose (DP7; centrifuge. After ultracentrifugation, supernatant was discarded and M7753, Sigma‐Aldrich) as an internal standard. The plate was then the obtained pellet was resuspended in 1 ml of PBS (Hyclone, GE incubated overnight in a humidified plastic box at 37°C to avoid HeathCare). The concentrated material was stored at −80°C for evaporation of the digestion solution. Released N‐glycans were re- future studies. covered from the PVDF plate by three rounds of centrifugation at ×1000g for 2 min and washing with 40 μl of water. To hydrolyze N‐glycan glycosylamine forms, 10 μl of 100 mM ammonium acetate 2.5 | EVs AND VLPs LYSIS (Sigma‐Aldrich) (pH 5) were added followed by a 1 h incubation at RT. Samples were subsequently dried in a SpeedVac concentrator 5301 Total protein concentration was quantified using Micro BCA™ (Eppendorf) at 40°C. Collected N‐glycans were then reduced and Protein Assay Kit (23235, ThermoFisher Scientific) according to desalted followed by PGC cleanup using a 96‐well plate‐based pro- manufacturer's instructions. Membrane lysis was performed using tocol previously reported (Jensen et al., 2012). Briefly, for the re- lysis buffer without detergents: 50 mM Tris‐HCl (Sigma‐Aldrich), duction, 20 μl of 1 M NaBH4 (Sigma‐Aldrich) in 50 mM KOH (Sigma‐ 100 mM NaCl (Sigma‐Aldrich), 1 mM EDTA (Sigma‐Aldrich), and one Aldrich) was added to each well followed by incubation for 3 h at tablet of protease inhibitor cocktail (Complete ULTRA Tablets Mini, 60°C in a humidified plastic box. To quench the reaction, 3 μl of Roche). Briefly, the corresponding volume of purified samples con- glacial acetic acid (Honeywell Fluka) was added to each sample. taining 20 μg of protein was mixed at a ratio 9:1 with concentrated Desalting of the samples was performed using a strong acidic cation lysis buffer, vortexed for 30 s and followed by two cycles of heating exchange resin Dowex 50W‐X8 (Sigma‐Aldrich) which was self‐ at 60°C for 15 min and sonication for 30 min in an ultrasound bath packed into 96‐well filter plates. Briefly, 100 μl of resin slurry in at RT. methanol (50/50, v/v) was added to each well in the filter plate. The columns were preconditioned by 3 × 100 μl of 1 M HCl, followed by 3 × 100 μl methanol and 3 × 100 μl water each time removed by 2.6 | GLYCOPROTEIN AND VESICLE LYSATE centrifuging at ×500g. The samples containing N‐glycans were loaded IMMOBILIZATION, DENATURATION, AND onto the columns and eluted two times with 40 μl of water followed REDUCTION by centrifugation at ×500g. The combined flow‐through and wash were collected and dried in a SpeedVac concentrator at 40°C. The A specific method for glycan release and reduction was carried out remaining borate was removed by several rounds of co‐evaporation (Zhang et al., 2020). A 96‐well plate with hydrophobic Immobilon‐P using 100 μl methanol in the SpeedVac concentrator at 40°C. SPE PVDF membrane (MultiScreen® HTS 96 multiwell plates, pore size cleaning step was performed by packing 60 μl (approximately 6 mg) 1210 | LAVADO‐GARCÍA ET AL. of bulk sorbent Carbograph slurry (Grace Discovery Sciences) in 10 min wash step using 95% of B at a flow of rate of 0.6 μl/min. The methanol into a 96‐well filter plate. The columns were precondi- column was held at a constant temperature of 45°C. Ionization was tioned by 3 × 100 μl of 80% acetonitrile in water containing 0.1% achieved using the nanoBooster source (Bruker) with a capillary TFA, and 3 × 100 μl water with 0.1% TFA. After loading the sample, voltage of 1000 V applied, dry gas temperature of 280°C at 5 L/min, the columns were washed three times with 80 μl of 0.1% TFA, fol- and nebulizer at 3 psi. Isopropanol enriched dopant nitrogen was lowed by N‐glycan elution by 3 × 40 μl of 60% acetonitrile in water used. MS spectra were acquired from m/z 500–1850 for N‐glycans containing 0.1% TFA. Samples were dried in a SpeedVac con- and 380–1850 for O‐glycans in enhanced mode, smart parameter centrator directly in polymerase chain reaction plates for injection setting was set to m/z 1200 for N‐glycans and m/z 900 for O‐glycans; and re‐dissolved in 10 μl of water before PGC nano‐LC‐ESI‐MS/MS ion charge control (ICC) to 4 × 103 and maximum acquisition time to analysis. 200 ms. MS/MS spectra were generated using collision‐induced dis- sociation over an m/z range from 100 to 2500 on the top three most abundant precursors, applying an isolation width of 3 Thomson. The 2.8 | O‐GLYCAN RELEASE AND PURIFICATION fragmentation cut‐off was set to 27% with 100% fragmentation amplitude using the Enhanced SmartFrag option from 30% to 120% After removal of N‐glycans, O‐glycans were released from the same in 32 ms and ICC was set to 150,000. PVDF membrane immobilized sample via reductive β‐elimination. Briefly, 50 μl of 0.5 M NaBH4 in 50 mM KOH was applied onto each PVDF membrane well after rewetting with 3 μl of methanol. 10 ng of 2.10 | DATA PROCESSING maltopentaose (DP5; M8128, Sigma‐Aldrich) was added as an inter- nal standard. Plates were then shaken for 15 min on a horizontal Glycan structures were assigned based on glycan composition ob- shaker and incubated in a humidified plastic box for 16 h at 55°C. tained from accurate mass, relative PGC elution position, MS/MS After incubation and cooling to RT, released O‐glycans were re- fragmentation pattern in negative‐ion mode, and general glycobio- covered by centrifugation at ×1000g for 2 min into 96‐well collection logical knowledge (Karlsson, Schulz, et al., 2004; Karlsson, Wilson, plates. The wells were rewetted by 3 μl of methanol and washed et al., 2004) with help of Glycoworkbench (Ceroni et al., 2008) and three times with 50 μl of water with 10 min incubation steps on a Glycomod (Cooper et al., 2001) software tools. Extracted ion chro- horizontal shaker before centrifugation at ×500g for 2 min. Before matograms were used to integrate the area under the curve for each desalting, the collected samples were concentrated to approximately individual glycan isomer using Compass Data Analysis software v.5.0 30 μl under vacuum in a SpeedVac concentrator at 40°C for 2 h. (Bruker). Glycan peaks in the profile were manually picked and in- Subsequently, 3 μl of glacial acetic acid was added to acidify the tegrated. Relative quantitation of individual glycans was performed reaction followed by brief centrifugation to collect the sample at the on the total area of all included glycans within one sample normalized bottom of the well. Subsequently, high throughput desalting and PGC to 100%. SPE purification were performed as described in the N‐glycan pre- paration section. The purified O‐glycan alditols were re‐suspended in 10 μl of water before PGC nano‐LC‐ESI‐MS/MS analysis. 2.11 | PARTICLE SIZE MEASUREMENT Nanoparticle tracking analysis (NTA) was performed with a 2.9 | ANALYSIS OF RELEASED N‐ AND NanoSight® LM20 device (NanoSight Ltd.) equipped with a blue laser O‐GLYCAN ALDITOLS USING PGC NANO‐LC‐ESI‐ module (488 nm) to quantify HIV‐1 Gag::eGFP VLPs and neutral MS/MS density filter for total particle by light scattering. Data were analyzed with NanoSight® NTA 3.2 software. Briefly, samples were injected, The analysis of N‐glycans and O‐glycans was performed on an and three technical replicate analyses were carried out. Subsequently, Ultimate 3000 UHPLC system (Thermo Fisher Scientific) equipped particles were identified and tracked by their Brownian motion at RT. with a Hypercarb PGC trap column (5 μm Hypercarb Kappa, Capture settings were recorded with an sCMOS camera (camera level 320 μm × 30 mm, Thermo Fisher Scientific) and a Hypercarb PGC of 8 for Gag::eGFP VLP samples, and 11 for controls, viscosity: 0.9 cP) nano‐column (3 μm Hypercarb 75 μm × 100 mm, Thermo Fisher and analyzed with a detection threshold of 4. Scientific) coupled to an amaZon ETD speed ion trap (Bruker). Mobile phase A consisted of 10 mM ammonium bicarbonate (ABC), while mobile phase B was 60% (v/v) acetonitrile/10 mM ABC. To analyze 2.12 | EXPERIMENTAL DESIGN AND glycans, 2 μl of prepared sample was injected and trapped on the trap STATISTICAL RATIONALE column using a 6 μl/min loading flow in 2% buffer B for N‐glycan and 1% buffer B for O‐glycans for 5 min. Separation was achieved with a For the comparative relative quantification glycomic experiments, multi‐step gradient of B: 2%–9% in 1 min and 9%–49% in 80 min for three conditions were tested, with two biological replicates (in- N‐glycan and 1%–52% over 72 min for O‐glycans followed by a dependent cell cultures) from each: no transfection (N), transfection LAVADO‐GARCÍA ET AL. | 1211 with mock plasmid (M), and transfection with the plasmid coding for 3 | RESULTS Gag::eGFP protein following the standard protocol previously de- scribed (S). Samples of each condition were taken at 72 h post‐ 3.1 | HEK293 VLP production transfection (hpt) followed by an ultracentrifugation to purify HIV‐1 Gag VLPs and EVs with the same density. As it is depicted in After transfecting HEK293 cells with PEI:DNA complexes containing Figure 1a, three technical replicates were analyzed from each sample. the plasmid encoding Gag::eGFP polyprotein, cell cultures were Fold changes were calculated from the average relative abundance harvested at 72 hpt. The clarified supernatants containing VLPs were value of each glycan species in the three studied conditions, noted as then concentrated using analytical ultracentrifugation. The resulting N, M, and S as previously described. pellet was resuspended in PBS, increasing VLP concentration in F I G U R E 1 Experimental design and virus‐like particle (VLP) production analysis. (a) Experimental workflow. Two biological replicates of HEK293 cell cultures were grown under three different conditions: no transfected condition, transfected with an empty plasmid denoted as “mock,” and transfected using the standard protocol with the plasmid coding for the translational fusion of HIV‐1 Gag polyprotein and enhanced GFP (Gag::eGFP) polyprotein. Both transfections were performed at the same cell density of 2 × 106 cells/ml. At the time point of 72 h post‐transfection (hpt), cultured samples were taken and centrifuged and supernatants were ultracentrifuged at 31,000 rpm using a 30% sucrose cushion. The resulting pellet was resuspended in 1 ml of phosphate‐buffered saline (PBS). The concentrated sample was subjected to N‐ and O‐glycan release and reduction. Resulted alditols were analyzed by porous grafitized carbon liquid chromatography with electrospray ionization tandem mass spectrometry (PGC‐LC‐ESI‐MS/MS) to assign the glycosylation structures. (b) Concentration via ultracentrifugation measured in relative fluorescence units (RFUs). Harvested VLPs were concentrated 12.2‐fold in the analytical ultracentrifugation step. (c) Extracted ion chromatograms profiles of samples that underwent lysis treatment and samples that contained intact VLPs for glycan analysis (no lysis treatment) 1212 | LAVADO‐GARCÍA ET AL. 12.2‐fold (Lavado‐García et al., 2021) (Figure 1b). These concentrated GlcNAcβ1‐ structures. The Gal residue can be sialylated, sulfated, VLP samples were used for further glycan analyses. also fucosylated or remain unsubstituted. Of the total 51 identified species, seven were high mannose N‐glycans. In these structures, the α(1,3) and α(1,6) arms contained only mannoses whose specific 3.2 | Analysis of N‐glycans branching pattern was not resolved. Additionally, hybrid N‐glycans were detected, which is a group of glycans with a mannosidic α(1,6) To analyze the N‐glycans species present in the Gag VLP membrane, arm and a complex α(1,3) arm containing GlcNAc, Gal, and NeuAc. some samples underwent a lysis treatment to disrupt the particles This group accounted for 6 of the 51 N‐glycans (Figure 2b) and 3 out and others maintained the complete structures before N‐glycan re- of these 6 hybrid glycans presented NeuAc as terminal modifications. lease and LC‐MS/MS analysis. These first measurements revealed The measured glycans differed greatly in relative abundance. The highly comparable N‐glycan profiles of lysed versus intact particles most abundant glycan, accounting for ∼13% of relative abundance (Figure 1c). Relative abundance and elution patterns of glycans were was H5N4F1S1a observed at m/z 1038.89 (charge 2−) which may be maintained, with differences in signal intensity. All identified represented as Hex2 HexNAc2 Deoxyhexose1 NeuAc1 Man3 GlcNAc2 N‐glycans were observed in both lysed and intact sample conditions. (Figure 2a–c). A total of 51 N‐glycan species were identified and 48 quantified. All characteristics and quantitative data relating to these 51 species are presented in Table S1. Overall, 38 glycan species were complex‐type. 3.3 | Analysis of O‐glycans Of these, 31 were sialylated, 1 contained a sulfate group and 7 presented Lewis epitopes (Figure 2a,b). Lewis epitopes or antigens The O‐glycan profile of VLP sample is exemplified in Figure 1c. A are fucosylated oligosaccharides derived from Galβ1‐3/4(Fucα1,4/3) total of nine O‐glycans were identified, from which eight were F I G U R E 2 Analysis of N‐glycan species from Gag virus‐like particles (VLPs). (a) Combined extracted ion chromatograms of 51 N‐glycan species released from VLP samples. Structures of the 17 most abundant glycans are depicted. Complex glycans are represented in blue, hybrid glycans represented in red, and oligomannoses or high mannose glycans represented in green. (b) Classification of the different types of N‐glycan species identified in the sample. Glycans with sulfate and Lewis epitopes represent a subdivision within complex glycans. (c) Relative abundance distribution of the 12 most abundant N‐glycan species identified in the VLP sample. Error bars represent standard deviation. The average value of all 51 glycans (1.96) is represented by the dotted line. F, fucose; H, hexose; N, N‐acetylhexosamine; S, N‐acetylneuraminic acid, Su: sulfate LAVADO‐GARCÍA ET AL. | 1213 F I G U R E 3 Analysis of O‐glycan species from Gag virus‐like particles (VLPs). (a) Combined extracted ion chromatograms of nine O‐glycan species released from VLP samples in which the top five most abundant glycans account for 94% of relative abundance. (b) Classification of the different types of O‐glycan species identified in the sample. (c) Relative abundance of the nine O‐glycan species identified in the VLP sample. Error bars represent standard deviation. Significance calculated by two‐way analysis of variance. H, hexose; N, N‐acetylhexosamine; S, N‐acetylneuraminic acid; Su, sulfate sialylated, presenting core 1, 2, and 4‐type structures. The only 3.4 | Changes in glycan signature upon transient nonsialylated O‐glycan presented core 2‐type structure and one of transfection and Gag VLP production the lowest abundances (Figure 3a,b). Isobaric O‐glycan structural isomer differentiation was successfully carried out by PCG‐LC The most abundant N‐ and O‐glycan species identified in the analyzed separation method coupled with negative ion tandem mass spec- samples did not show significant changes in the nontransfected, trometry (Figure 4). For instance, this method allowed to resolve transfected with mock, and transfected with the plasmid coding for the position of the sialic acid in the m/z 1040.521− based on the Gag::eGFP studied conditions. When assessing the major glycan com- identification of characteristic diagnostic ions (Figure 4b,c). Within positions, the N‐glycosylation patterns appeared rather similar for the the total of nine identified O‐glycans species, the top two most different samples (Figure 2c). Still, some N‐glycans were found to show abundant species accounted for the 76% of relative abundance, small, significant differences in relative abundance. One of the four being the most abundant the tetrasaccharide H1N1S2 with an isomers (H7N6F1S4a) of N‐glycan H7N6F1S4 showed elevated levels observed m/z of 966.4 followed by the hexasaccharide H2N2S2 of as indicated by the relative abundance fold change in M and S condition m/z 1331.6 (Figure 3c). These two species are common (Figure 5a and Figure S1). This sialylated tetrantennary N‐glycan ac- O‐glycosylations found in many human proteins. No significant counted for ∼0.5% of relative abundance among all N‐glycans in N and differences were observed between nontransfected, transfected M conditions shifted to ∼2%. Isomer H7N6F1S4c showed lower in- with mock, and transfected with Gag::eGFP conditions tensities in M and S compared to N condition. A shift was observed in (Figure 3b,c). Similar to N‐glycans, O‐glycan profiles were not in- H7N6F1S4c (retention time 78.2 min) from ∼1.0% to ∼0.5% upon fluenced by lysis (Figure 1c). All the identified and quantified transfection. Isomers H7N6F1S4b and H7N6F1S4d hardly differed in O‐glycans are presented in Table S2. all studied conditions. It has been reported that upon recombinant 1214 | LAVADO‐GARCÍA ET AL. F I G U R E 4 Example of the porous graphitized carbon‐liquid chromatography separation of O‐glycan structural isomers and MS2 analysis. (a) Extracted ion chromatogram (EIC) of the m/z 1040.521− (Hex2 HexNAc2 NeuAc1) O‐glycan isomers. (b) MS2 spectrum of the m/z 1040.521− isomer observed at RT 58.1 min. (c) MS2 spectrum of the m/z 1040.521− isomer observed at RT 66.1 min. Diagnostic ions are denoted in red. H, hexose; N, N‐acetylhexosamine; S, N‐acetylneuraminic acid production of human Protein C in HEK293 cells, the structure of abundance in S condition were three of the identified high mannose H7N6F1S4 glycan shifted from presenting all α(2‐3) linkage to a N‐glycans; H7N2a, H9N2, and H8N2. High mannose N‐glycans re- combination of α(2‐3) and α(2‐6) linkage of sialic acid (Yan et al., 1993). sidues have been reported to influence pharmacokinetics of ther- The structural differences of isomers H7N6F1S4a and H7N6F1S4c apeutic molecules like IgGs, reducing their half‐life (Alessandri might also be related to a change in sialic acid linkage combinations. et al., 2012; Goetze et al., 2011). Interestingly, the other N‐glycan in- Also, the isomer H3N5F1b showed a significant increase in S and M creasing in S condition and therefore in the VLP fraction, was H5N5 conditions (Table S1). Moreover, different N‐glycans were observed to (Figure 5a), a glycan found on glycoprotein CD59 (Meri et al., 1996; have elevated relative abundance only in S condition (Figure 5a and Rudd et al., 1997). Glycoprotein CD59 is commonly present in HIV‐1 Figure S1). The biantennary glycan H3N5F1a, bearing a bisecting virions preventing the action of the complement cascade and the lysis GlcNAc, a β(1,4)GlcNAc attached to the innermost mannose residue, of the viral particles (Amet et al., 2016). There were also N‐glycan was observed to slightly increase in relative abundance in VLPs, from species that decreased upon transfection and VLP production. For in- ∼0.4% to ∼0.7%. Structures presenting a bisecting GlcNAc have been stance, the second most abundant N‐glycan, H5N4F1S2 (Figure 5a and reported to be involved in relevant immunological processes like im- Figure S1). Among the N‐glycans that showed lower intensity only mune tolerance and immune suppression (Chen et al., 2020). The se- upon VLP production, the greatest decrease, 0.3‐fold, was found in paration and identification of structural isomers H3N5F1a and H5N5F1S1b, having an α(1,3) sialic acid linkage. The hybrid H7N3F1 H3N5F1b is illustrated in Figure 6. The characteristic diagnostic ions for glycan also presented a reduction of 0.5‐fold in relative abundance in S the bisecting structure (D‐221 ion) and from the nonbisecting isomer condition. As for the O‐glycan species, no significant differences were (D and D‐18 ions) supported the assignment and identification observed between nontransfected, transfected with mock, and trans- (Figure 6b,c). Other N‐glycans presenting an increased relative fected with Gag::eGFP conditions (Figure 5b). LAVADO‐GARCÍA ET AL. | 1215 F I G U R E 5 Relative abundance changes in glycans of produced particles upon transfection and virus‐like particle production. (a) Heatmap illustrating relative abundance and fold change ratios of N‐glycans showing changes in the different studied conditions (left). Replicates are denoted as I, II, III, and IV. (b) Heatmap illustrating relative abundance and fold change ratios of O‐glycans showing changes in the different studied conditions (left). Replicates are denoted as I, II, III, and IV. F, fucose; H, hexose; N, N‐acetylhexosamine; S, N‐acetylneuraminic acid; Su, sulfate 3.5 | VLPs present lower glycan density per conditions (Figure 7a). From the total 2.88 × 1011 particles/ml mea- particle than EVs sured in S condition, 82.3% corresponded to fluorescent particles (Gag::eGFP VLPs). The remaining 17.7% corresponded to coproduced Particle concentration and the size distribution curve in each condi- EVs. We next aimed at estimating the glycan content per particle. To tion were measured by NTA showing a 5.4 times increased con- get an estimate of the glycan amount in each condition, glycan in- centration of particles in the S condition compared to N and M tensity was normalized to the intensity of the internal standard DP7. 1216 | LAVADO‐GARCÍA ET AL. F I G U R E 6 Example of the porous graphitized carbon‐liquid chromatography separation of N‐glycan structural isomers and MS2 analysis. (a) Extracted ion chromatogram (EIC) of the m/z 832.812− (Hex3 HexNAc5 Fuc1) N‐glycan isomers. (b) MS2 spectrum of the m/z 832.812− isomer observed at RT 49.8 min. (c) MS2 spectrum of the m/z 832.812− isomer observed at RT 59.9 min. Diagnostic ions are denoted in red. F, fucose; GlcNAc, N‐Acetylglucosamine; H, hexose; N, N‐acetylhexosamine Then, assuming complete release of glycans and similar response produced. For human application, it is fundamental to assure that factors between condition and DP7 standard, the number of glycans the final bioproduct does not present exogenous and allergenic per particle was estimated (Figure 7b). The condition containing the glycoforms that could neutralize and handicap the envisaged func- VLP fraction (S condition) presented a significantly lower glycan tion of the biopharmaceutical. Although some xeno‐glycomotifs can density than N and M conditions. Interestingly, M condition also have a desirable effect of boosting the immune response, others can presented a reduction of glycan density compared to N condition cause allergy reactions (C. Jin et al., 2008). HIV‐1 Gag VLPs con- (Figure 7b). A reduction from an average of 2400 and 1620 N‐glycan ceived as vaccine candidates are commonly produced in insect and molecules per particle for the N and M condition respectively to 380 plant cells expression systems to obtain higher titers. However, the for the S condition was assessed. For O‐glycans, it dropped from presence of α1,3‐fucosylations in the core of N‐glycans produced in 6830 to 4300 O‐glycans/particle for the N and M condition re- insect cells or the presence of both allergenic α1,3‐fucosylations spectively to 1150 for the S condition. This is schematically re- and β1,2‐xyloses in N‐glycans produced in plant cells increment the presented in Figure 7c. complexity of the bioprocesses using these expression systems. Additional steps to remove these glycoforms are needed as the final purified VLPs cannot contain them for human applications. In this 4 | DISC US SION study, the glycosylation profile of Gag VLPs produced in HEK293 cells has been characterized, showing that the displayed glycans The analysis and characterization of the glycosylation signature of presented structures of high mannose, hybrid and mostly complex biopharmaceuticals is essential to understand their effect on eli- glycans. For instance, Chikungunya VLPs produced in Sf9 insect citing the expected immune response. The glycosylation pattern cells presented predominantly high mannose structures, (Lancaster depends exclusively on the cell‐based platform where it has been et al., 2016) needing to complement the VLPs with adjuvants to LAVADO‐GARCÍA ET AL. | 1217 F I G U R E 7 Analysis of diffracting particles. (a) The size distribution curve of diffracting particles in each condition measured by nanoparticle tracking analysis (NTA). The vertical gray line represents 145 nm. N: nontransfected condition, M: condition transfected with mock plasmid, S: condition transfected with plasmid encoding translational fusion of HIV‐1 Gag polyprotein and enhanced GFP (Gag::eGFP) polyprotein, VLPs: virus‐like particles, EVs: extracellular vesicles. (b) Glycan density per particle. The total intensity area of identified glycans in each condition was normalized by their corresponding internal standard DP7 intensity area and divided by the corresponding number of particles in the sample. The number of particles was calculated using the particle concentration value obtained by NTA. Significance was calculated using two‐way analysis of variance. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns: non significant. (c) Schematic representation of the glycan density in EVs (left) and Gag VLPs (right). The mean value of the particle diameter measured by NTA is represented within the particle. The represented glycan species in EVs (left) are the most abundant and the ones showing downregulation in S condition. The represented glycans in VLPs (right) are the ones showing an increase in S condition. The values showing the number glycan/particle are the approximate value calculated assuming the same response factor of the standard DP7 with known concentration, used in each condition achieve the same neutralizing antibody titers as VLPs produced in detection from the immune system. This is one of the main reasons mammalian cells (Wagner et al., 2014). High mannose structures for vaccine strategies to target the surface protein, not considering have been described to produce fewer neutralizing antibody titers the glycan shield and leading to low neutralizing titers. Displaying than complex glycans (Lin et al., 2013). The lower neutralizing an- not only high mannose but also hybrid and complex glycans con- tibody titers generated by high mannose glycans may be explained tributes to a stronger immunological response (Lin et al., 2013). by the glycan shielding effect that they provide to viral particles. Most importantly, VLPs produced in HEK293 cells harbored sialic Often, immunogenic surface proteins from viruses like Ebola GP, acid terminations in 63.5% of all identified glycans, between N‐ and (Beniac & Timothy, 2017) Influenza HA, (An et al., 2015) HIV‐1 Env O‐glycans. This feature may potentially improve the recognition of glycoprotein (Lee et al., 2016; Seabright et al., 2019; Wei VLPs by macrophages and further presentation to antigen‐ et al., 2003), or SARS‐CoV‐2 S protein (Casalino et al., 2020; Grant presenting cells. This opens a potential research line for VLPs in the et al., 2020) are masked in high mannose structures preventing future, as macrophage recognition is promoted via Siglec‐1 1218 | LAVADO‐GARCÍA ET AL. receptor, which selectively binds to α(2,3)‐linked sialic acids, en- is the use of multivalent glycan display for cell‐specific CLR tar- hancing the immune response (Jobe et al., 2018). geting. Specific fungal high mannose or Lewis‐B or Lewis‐X antigens The relative abundance of N‐ and O‐glycans identified in this selectively interact with CLR CD209, eliciting a robust CD4+ and study has been observed to change upon transfection and VLP CD8+ immune response (Unger et al., 2012). This strategy has production. The changes observed upon transfection correlated proved to achieve immunization using lentiviral nanoparticles (Yang with the generation of small EVs (sEVs), previously reported et al., 2008), an approach that could be applied to Gag VLPs in (Lavado‐García, González‐Domínguez, et al., 2020). The increase in the future. the generation of vesicles smaller than 200 nm upon transient Moreover, the presence of N‐glycans presenting structures transfection in HEK293 cells correlates with the increase of large containing a bisecting GlcNAc has been reported to influence im- tetrantennary glycans like H7N6F1S4b. This is hypothesized to be mune tolerance, shielding the cells that possessed bisecting GlcNAc related to the generation of EVs via de multivesicular bodies path- structures from the immune system (Link‐Lenczowski et al., 2018; way and release as exosomes. This phenomenon could also be re- Miwa et al., 2012). In addition to this, bisecting GlcNAc has been lated to the reduction in glycan density in mock‐derived EVs and reported to suppress further terminal modification such as sialic acid, VLPs. Since sEVs and VLPs are smaller particles, they might deplete fucose, or Lewis epitopes. Losing the bisecting GlcNAc made N‐gly- the cell of more membrane surface at a higher rate, reducing the cell cans better substrates for many glycosyltransferases, increasing the capacity of replenishing the glycan reservoir. In fact, the VLP spe- number of terminal modification (Nakano et al., 2019). Therefore, to cific productivity is 4.5 times higher producing VLPs than in mock increase potential immunogenic terminal modifications and reduce condition, from (3.5 ± 0.1) × 103 to (0.8 ± 0.1) × 103 VLP cell−1 day−1 shielding the VLPs from the immune system, a HEK293 N‐ respectively (Lavado‐García et al., 2021). Secreting vesicles at a acetylglucosaminyltransferase‐III (GnT‐III) knock out production lower rate might provide more time for glycoproteins to be trans- platform could be established. This would provide VLPs rich in po- ported to the membrane. The reduction in vesicle size and increase tential immunogenic terminal modifications as better vaccine candi- in production rate of the generated vesicles might account for the dates. Another identified feature of the analyzed VLPs is the reduction in glycan density. More interestingly, the specific changes presence of high mannose N‐glycans. A high concentration of this observed in VLP glycans can shed some light on one of the greatest type of N‐glycan has also been reported to influence immune challenges of Gag‐VLP bioprocess development: the separation of shielding. One relevant and illustrating example is HIV‐1 Env glyco- VLPs from the coproduced EVs. This is an essential part of the protein. This is one of the most heavily glycosylated proteins ever downstream purification process that is still under research. The studied, containing almost solely high mannose N‐glycans (Bonomelli usual purification process of VLPs start with clarifying the recovered et al., 2011; Doores et al., 2010). This high concentration of high supernatant to remove cells and debris followed by capture and mannose glycans offers shielding to HIV‐1 Env glycoprotein from the concentration, typically performed by chromatography. The current recognition of antibodies and hamper the immune response. High capture technologies that are being used for VLPs are anion‐ mannose to complex N‐glycans ratio has also been described to di- exchange monolith or Capto™‐Heparin columns, yielding from 2% rectly influence the stimulation of antibody synthesis and recogni- to 13% at elution (Pereira Aguilar et al., 2020; Reiter et al., 2019). tion (Raska et al., 2010; Wei et al., 2003). Consequently, to reduce These current methods do not efficiently separate VLPs and EVs. high mannose glycans in the VLP surface, particles can be treated in The study of VLPs and coproduced EVs glycosylations can con- vitro with α‐mannosidases to hydrolyze terminal mannoses. How- tribute to the design of more efficient affinity methods in the future ever, a better option reducing the bioprocess final costs and com- for the downstream processing step of capture, separating both plexity could be to establish an engineered HEK293 platform population of nanoparticles. In this study, we have described dif- overexpressing Golgi α‐mannosidases coded by genes MAN1A1 and ferences in glycosylation abundance in VLPs and EVs, highlighting MAN1A2. Knocking out these genes has been reported to generate the difference in glycoforms of CD59. H5N5 N‐glycans from CD59 cells with only high mannose surface N‐glycans (Z. C. Jin et al., 2018) significantly increased in VLPs. These changes could open the door and overexpressing them would trim mannoses from uncapped gly- to the design of affinity method based on these glycans to chro- cans with no terminal modifications, increasing complex species and matographically separate both EV and VLP populations. In addition reducing the high mannose to complex N‐glycans ratio. A different to this, the glycans that significantly decreased in VLPs, like glycoengineering approach to enhance VLP immunogenicity is the H5N5F1S1b with α(2,3) sialic acid linkage could be used to create expression of the α‐Gal epitope. The trisaccharide structure Gal‐α affinity separation methods to allow the flow‐through of VLPs while (1,3)‐Gal‐β(1,4)‐GlcNAc‐R is an epitope found in all nonprimate retaining EVs. Either using affinity resins to capture VLPs for latter mammals and absent in humans. The presence of α‐Gal epitope elicits elution or retaining EVs while filtering VLPs could provide more a strong immune response and has been used to boost the elicited efficient solutions for VLP purification. Apart from designing a se- immune response of vaccines against Influenza (Abdel‐Motal paration method for downstream processing, knowing the glyco- et al., 2007), HIV‐1 (Abdel‐Motal et al., 2010), and SARS‐CoV‐2 sylation signature of Gag VLPs paves the way for glycoengineering (Galili, 2020). These modulations based on glycoengineering could be strategies to enhance VLP immunogenicity towards the creation of implemented towards the development of an improved vaccine self‐adjuvanted and cell‐specific VLP‐based vaccines. One example platform based on HIV‐1 Gag VLPs. LAVADO‐GARCÍA ET AL. | 1219 5 | C ONC LUS I ON S Amet, T., Lan, J., Shepherd, N., Yang, K., Byrd, D., Xing, Y., & Yu, Q. (2016). Glycosylphosphatidylinositol anchor deficiency attenuates the production of infectious HIV‐1 and renders virions sensitive to The characterization of protein glycosylation in biopharmaceuticals is complement attack. AIDS Research and Human Retroviruses, 32, essential for the correct development of further clinical trials to reach 1100–1112. human applications. HIV‐1 Gag‐based VLPs are promising candidates An, Y., McCullers, J. A., Alymova, I., Parsons, L. M., & Cipollo, J. F. (2015). for the development of new vaccine therapies. In this study, we have Glycosylation analysis of engineered H3N2 influenza A virus hemagglutinins with sequentially added historically relevant characterized the glycome of VLPs, showing that they present the glycosylation sites. Journal of Proteome Research, 14, 3957–3969. required complex glycosylations for the envisaged interaction with Ausubel, L. J., Hall, C., Sharma, A., Shakeley, R., Lopez, P., Quezada, V., the immune system of this type of biopharmaceuticals. Moreover, to Couture, S., Laderman, K., McMahon, R., Huang, P., Hsu, D., & design an integrated production bioprocess, key protein glycosyla- Couture, L. (2012). Production of CGMP‐grade lentiviral vectors. BioProcess International, 10, 32–43. tions have been identified for purification and separation strategies Beniac, D. R., & Timothy, B. F. (2017). Structure of the Ebola virus of VLPs and coproduced EVs. N‐glycans like H5N5 and H5N5F1S1b glycoprotein spike within the virion envelope at 11 Å resolution. showed significant differences in VLPs compared to EVs. This could Scientific Reports, 7, 46374. pave the way for affinity‐based separations leading to specific pur- Bonomelli, C., Doores, K. J., Dunlop, D. C., Thaney, V., Dwek, R. A., ification methods that would allow the optimization of downstream Burton, D. R., Crispin, M., & Scanlan, C. N. (2011). The glycan shield of HIV is predominantly oligomannose independently of production processing of VLPs. system or viral clade. PLoS One, 6, e23521. Casalino, L., Gaieb, Z., Goldsmith, J. A., Hjorth, C. K., Dommer, A. C., A C KN O W L E D G M E N T S Harbison, A. M., Fogarty, C. A., Barros, E. P., Taylor, B. C., The authors thank Prof. Amine Kamen (McGill University, Montreal, McLellan, J. S., Fadda, E., & Amaro, R. E. (2020). Beyond shielding: The roles of glycans in the SARS‐CoV‐2 spike protein. ACS Central Canada) for providing the HEK293 SF‐3SF6. The following reagent Science, 6(1722–1734), 2020–1734. was obtained through the National Institutes of Health AIDS Reagent Ceroni, A., Maass, K., Geyer, H., Geyer, R., Dell, A., & Haslam, S. M. (2008). Program, Division AIDS, National Institute of Allergy and Infectious GlycoWorkbench: A tool for the computer‐assisted annotation of Diseases, National Institutes of Health: pGag::eGFP (Cat#11468) mass spectra of glycans. Journal of Proteome Research, 7, 1650–1659. from Dr. Merilyn Resh. The support of Dr. Jose Amable Bernabé from Cervera, L., Gòdia, F., Tarrés‐Freixas, F., Aguilar‐Gurrieri, C., Carrillo, J., the Institute of Material Science of Barcelona (ICMAB‐CSIC, Barce- Blanco, J., & Gutiérrez‐Granados, S. (2019). Production of HIV‐1‐ lona, Spain) with particle tracking techniques is greatly appreciated. based virus‐like particles for vaccination: Achievements and limits. The authors declare no conflict of interest. The project that gave rise Applied Microbiology and Biotechnology, 103, 7367–7384. Chen, Q., Tan, Z., Guan, F., & Ren, Y. (2020). The essential functions and to these results received the support of a fellowship from the “la detection of bisecting GlcNAc in cell biology. Frontiers in Chemistry, Caixa” Foundation (ID 100010434). The fellowship code is LCF/BQ/ 8, 511. ES17/11600003. European Molecular Biology Organization (EMBO) Colgrave, M. L., Snelling, H. J., Shiell, B. J., Feng, Y. R., Chan, Y. P., Short‐Term Fellowship Number 8613 (2020). The research group is Bossart, K. N., Xu, K., Nikolov, D. B., Broder, C. C., & Michalski, W. P. (2012). Site occupancy and glycan compositional analysis of two recognized as 2017 SGR 898 by Generalitat de Catalunya. soluble recombinant forms of the attachment glycoprotein of Hendra virus. Glycobiology, 22, 572–584. D A TA A V A I L A B I L I T Y S T A T E M E N T Cooper, C. A., Gasteiger, E., & Packer, N. H. (2001). GlycoMod—A The raw mass spectrometry data has been submitted and is available software tool for determining glycosylation compositions from mass in GlycoPost repository with ID number GPST000169 (https:// spectrometric data. Proteomics, 1, 340–349. Dhar, C., Sasmal, A., & Varki, A. (2019). From ‘serum sickness’ to glycopost.glycosmos.org/entry/GPST000169). ‘xenosialitis’: Past, present, and future significance of the non‐human sialic acid Neu5Gc. Frontiers in Immunology, 10, 807. ORCID Doores, K. J., Bonomelli, C., Harvey, D. J., Vasiljevic, S., Dwek, R. A., Jesús Lavado‐García http://orcid.org/0000-0001-9993-6332 Burton, D. R., Crispin, M., & Scanlan, C. N. (2010). Envelope glycans of immunodeficiency virions are almost entirely oligomannose antigens. Proceedings of the National Academy of Sciences of the REFERENCES United States of America, 107, 13800–13805. Abdel‐Motal, U. M., Guay, H. M., Wigglesworth, K., Welsh, R. M., & Duivelshof, B. L., Jiskoot, W., Beck, A., Veuthey, J. L., Guillarme, D., & Galili, U. (2007). Immunogenicity of influenza virus vaccine is D′Atri, V. (2019). Glycosylation of biosimilars: Recent advances in increased by anti‐gal‐mediated targeting to antigen‐presenting analytical characterization and clinical implications. Analytica Chimica cells. Journal of Virology, 81, 9131–9141. Acta, 1089, 1–18. Abdel‐Motal, U. M., Wang, S., Awad, A., Lu, S., Wigglesworth, K., & Effio, C. L., & Hubbuch, J. (2015). Next generation vaccines and vectors: Galili, U. (2010). Increased immunogenicity of HIV‐1 p24 and gp120 Designing downstream processes for recombinant protein‐based following immunization with gp120/p24 fusion protein vaccine virus‐like particles. Biotechnology Journal, 10, 715–727. expressing α‐gal epitopes. Vaccine, 28, 1758–1765. Galili, U. (2020). Amplifying immunogenicity of prospective Covid‐19 Alessandri, L., Ouellette, D., Acquah, A., Rieser, M., Leblond, D., vaccines by glycoengineering the coronavirus glycan‐shield to Saltarelli, M., Radziejewski, C., Fujimori, T., & Correia, I. (2012). present α‐gal epitopes. Vaccine, 38, 6487–6499. Increased serum clearance of oligomannose species present on a Ghaderi, D., Taylor, R. E., Padler‐Karavani, V., Diaz, S., & Varki, A. (2010). human IgG1 molecule. mAbs, 4, 509–520. Implications of the presence of N‐glycolylneuraminic acid in 1220 | LAVADO‐GARCÍA ET AL. recombinant therapeutic glycoproteins. Nature Biotechnology, 28, Lavado‐García, J., Jorge, I., Boix‐Besora, A., Vázquez, J., Gòdia, F., & 863–867. Cervera, L. (2021). Characterization of HIV‐1 virus‐like particles and Goetze, A. M., Liu, Y. D., Zhang, Z., Shah, B., Lee, E., Bondarenko, P. V., & determination of Gag stoichiometry for different production Flynn, G. C. (2011). High‐mannose glycans on the Fc region of platforms. Biotechnology and Bioengineering, 27786, 2660–2675. therapeutic IgG antibodies increase serum clearance in humans. https://doi.org/10.1002/bit.27786 Glycobiology, 21, 949–959. Lee, J. H., Ozorowski, G., & Ward, A. B. (2016). Cryo‐EM structure of a Goldblatt, D. (2000). Conjugate vaccines. Clinical and Experimental native, fully glycosylated, cleaved HIV‐1 envelope trimer. Science, Immunology, 119, 1–3. 351, 1043–1048. Grant, O. C., Montgomery, D., Ito, K., & Woods, R. J. (2020). Analysis of Lepenies, B., Lee, J., & Sonkaria, S. (2013). Targeting C‐type lectin the SARS‐CoV‐2 spike protein glycan shield reveals implications for receptors with multivalent carbohydrate ligands. Advanced Drug immune recognition. Scientific Reports, 10, 14991. Delivery Reviews, 65, 1271–1281. Gupta, S. K., & Shukla, P. (2018). Glycosylation control technologies for Lin, S. C., Jan, J. T., Dionne, B., Butler, M., Huang, M. H., Wu, C. Y., recombinant therapeutic proteins. Applied Microbiology and Wong, C. H., & Wu, S. C. (2013). Different immunity elicited by Biotechnology, 102, 10457–10468. recombinant H5N1 hemagglutinin proteins containing pauci‐ Gutiérrez‐Granados, S., Cervera, L., Gòdia, F., Carrillo, J., & Segura, M. M. mannose, high‐mannose, or complex type N‐glycans. PLoS One, 8, (2013). Development and validation of a quantitation assay for e66719. fluorescently tagged HIV‐1 virus‐like particles. Journal of Virological Link‐Lenczowski, P., Bubka, M., Balog, C. I. A., Koeleman, C. A. M., Methods, 193, 85–95. Butters, T. D., Wuhrer, M., & Lityńska, A. (2018). The glycomic effect Holst, S., Belo, A. I., Giovannetti, E., Van Die, I., & Wuhrer, M. (2017). of N‐acetylglucosaminyltransferase III overexpression in metastatic Profiling of different pancreatic cancer cells used as models for melanoma cells. GnT‐III modifies highly branched N‐glycans. metastatic behaviour shows large variation in their N‐glycosylation. Glycoconjugate Journal, 35, 217–231. Scientific Reports, 7, 16623. Mastrangeli, R., Palinsky, W., & Bierau, H. (2018). Glycoengineered Jensen, P. H., Karlsson, N. G., Kolarich, D., & Packer, N. H. (2012). antibodies: Towards the next‐generation of immunotherapeutics. Structural analysis of N‐ and O‐glycans released from glycoproteins. Glycobiology, 29, 199–210. Nature Protocols, 7, 1299–1310. Md, A., Maeda, M., Matsui, T., Takasato, Y., Ito, K., & Kimura, Y. (2021). Jin, C., Hantusch, B., Hemmer, W., Stadlmann, J., & Altmann, F. (2008). Purification and molecular characterization of a truncated‐type Ara h Affinity of IgE and IgG against cross‐reactive carbohydrate 1, a major peanut allergen: Oligomer structure, antigenicity, and determinants on plant and insect glycoproteins. Journal of Allergy glycoform. Glycoconjugate Journal, 38, 67–76. and Clinical Immunology, 121, 185–190. Meri, S., Lehto, T., Sutton, C. W., Tyynelä, J., & Baumann, M. (1996). Structural Jin, Z. C., Kitajima, T., Dong, W., Huang, Y. F., Ren, W. W., Guan, F., composition and functional characterization of soluble CD59: Chiba, Y., Gao, X. D., & Fujita, M. (2018). Genetic disruption of Heterogeneity of the oligosaccharide and glycophosphoinositol (GPI) multiple 1,2‐mannosidases generates mammalian cells producing anchor revealed by laser‐desorption mass spectrometric analysis. recombinant proteins with high‐mannose–type N‐glycans. Journal of Biochemical Journal, 316, 923–935. Biological Chemistry, 293, 5572–5584. Miwa, H. E., Song, Y., Alvarez, R., Cummings, R. D., & Stanley, P. (2012). Jobe, O., Kim, J., & Rao, M. (2018). The role of Siglec‐1 in HIV‐1/ The bisecting GlcNAc in cell growth control and tumor progression. macrophage interaction. Macrophage, 3. e1435. Glycoconjugate Journal, 29, 609–618. Johannssen, T., & Lepenies, B. (2017). Glycan‐based cell targeting to Nakano, M., Mishra, S. K., Tokoro, Y., Sato, K., Nakajima, K., Yamaguchi, Y., modulate immune responses. Trends in Biotechnology, 35, 334–346. Taniguchi, N., & Kizuka, Y. (2019). Bisecting GlcNAc is a general Karlsson, N. G., Schulz, B. L., & Packer, N. H. (2004). Structural suppressor of terminal modification of N‐glycan. Molecular & Cellular determination of neutral O‐linked oligosaccharide alditols by Proteomics, 18, 2044–2057. negative ion LC‐electrospray‐MSn. Journal of the American Society Pereira Aguilar, P., et al (2020). Capture and purification of human for Mass Spectrometry, 15, 659–672. immunodeficiency virus‐1 virus‐like particles: Convective media vs Karlsson, N. G., Wilson, N. L., Wirth, H. J., Dawes, P., Joshi, H., & porous beads. Journal of Chromatography, A, 1627, 461378. Packer, N. H. (2004). Negative ion graphitised carbon nano‐liquid Raska, M., Takahashi, K., Czernekova, L., Zachova, K., Hall, S., chromatography/mass spectrometry increases sensitivity for Moldoveanu, Z., Elliott, M. C., Wilson, L., Brown, R., Jancova, D., glycoprotein oligosaccharide analysis. Rapid Communications in Barnes, S., Vrbkova, J., Tomana, M., Smith, P. D., Mestecky, J., Mass Spectrometry, 18, 2282–2292. Renfrow, M. B., & Novak, J. (2010). Glycosylation patterns of Lancaster, C., Pristatsky, P., Hoang, V. M., Casimiro, D. R., Schwartz, R. M., HIV‐1 gp120 depend on the type of expressing cells and affect Rustandi, R., & Ha, S. (2016). Characterization of N‐glycosylation antibody recognition. Journal of Biological Chemistry, 285, profiles from mammalian and insect cell derived chikungunya VLP. 20860–20869. Journal of Chromatography B: Analytical Technologies in the Biomedical Reiter, K., Aguilar, P. P., Wetter, V., Steppert, P., Tover, A., & Jungbauer, A. and Life Sciences, 1032, 218–223. (2019). Separation of virus‐like particles and extracellular vesicles by Lavado‐García, J., Cervera, L., & Gòdia, F. (2020). An alternative perfusion flow‐through and heparin affinity chromatography. Journal of approach for the intensification of virus‐like particle production in Chromatography, A, 1588, 77–84. HEK293 cultures. Frontiers in Bioengineering and Biotechnology, Roldão, A., Mellado, M. C., Castilho, L. R., Carrondo, M. J., & Alves, P. M. 8, 617. (2010). Virus‐like particles in vaccine development. Expert Review of Lavado‐García, J., Díaz‐Maneh, A., Canal‐Paulí, N., Pérez‐Rubio, P., Vaccines, 9, 1149–1176. Gòdia, F., & Cervera, L. (2021). Metabolic engineering of HEK293 Rudd, P. M., Morgan, B. P., Wormald, M. R., Harvey, D. J., cells to improve transient transfection and cell budding of HIV‐1 van den Berg, C. W., Davis, S. J., Ferguson, M. A., & Dwek, R. A. virus‐like particles. Biotechnology and Bioengineering, 118, (1997). The glycosylation of the complement regulatory protein, 1630–1644. https://doi.org/10.1002/bit.27679 human erythrocyte CD59‐44. Journal of Biological Chemistry, 272, Lavado‐García, J., González‐Domínguez, I., Cervera, L., Jorge, I., 7229–7244. Vázquez, J., & Gòdia, F. (2020). Molecular characterization of the Schön, K., Lepenies, B., & Goyette‐Desjardins, G. (2020). Impact of protein coproduced extracellular vesicles in HEK293 during virus‐like glycosylation on the design of viral vaccines. Advances in Biochemical particle production. Journal of Proteome Research, 19, 4532. Engineering/Biotechnology, 175, 319–354. LAVADO‐GARCÍA ET AL. | 1221 Seabright, G. E., Doores, K. J., Burton, D. R., & Crispin, M. (2019). Protein Komarova, N. L., Nowak, M. A., Hahn, B. H., Kwong, P. D., & and glycan mimicry in HIV vaccine design. Journal of Molecular Shaw, G. M. (2003). Antibody neutralization and escape by HIV‐1. Biology, 431, 2223–2247. Nature, 422, 307–312. Seifalian, A., De La Peña, H., & Seifalian, A. M. (2010). The application of Yan, S. B., Chao, Y. B., & Van Halbeek, H. (1993). Novel Asn‐linked exosomes as a nanoscale cancer vaccine. International Journal of oligosaccharides terminating in GaINAcβ(1→4)[fucα(1→3)] Nanomedicine, 5, 889. GlcNAcβ(1→•) are present in recombinant human protein C Stavenhagen, K., Kolarich, D., & Wuhrer, M. (2014). Clinical glycomics expressed in human kidney 293 cells. Glycobiology, 3, 597–608. employing graphitized carbon liquid chromatography–mass Yang, L., Yang, H., Rideout, K., Cho, T., Joo, K. I., Ziegler, L., Elliot, A., spectrometry. Chromatographia, 78, 307–320. Walls, A., Yu, D., Baltimore, D., & Wang, P. (2008). Engineered Thomsson, K. A., Bäckström, M., Holmén Larsson, J. M., Hansson, G. C., & lentivector targeting of dendritic cells for in vivo immunization. Karlsson, H. (2010). Enhanced detection of sialylated and sulfated Nature Biotechnology, 26, 326–334. glycans with negative Ion mode nanoliquid chromatography/ Yehuda, S., & Padler‐Karavani, V. (2020). Glycosylated biotherapeutics: mass spectrometry at high pH. Analytical Chemistry, 82, Immunological effects of N‐glycolylneuraminic acid. Frontiers in 1470–1477. Immunology, 11, 21. Tretter, V., Altmann, F., Kubelka, V., März, L., & Becker, W. M. (1993). Zhang, T., Madunić, K., Holst, S., Zhang, J., Jin, C., Ten Dijke, P., Fucose α1,3‐linked to the core region of glycoprotein n‐glycans Karlsson, N. G., Stavenhagen, K., & Wuhrer, M. (2020). Development creates an important epitope for IgE from honeybee venom allergic of a 96‐well plate sample preparation method for integrated: N‐ and individuals. International Archives of Allergy and Immunology, 102, O‐glycomics using porous graphitized carbon liquid chromatography‐ 259–266. mass spectrometry. Molecular Omics, 16, 355–363. Unger, W. W. J., van Beelen, A. J., Bruijns, S. C., Joshi, M., Fehres, C. M., van Bloois, L., Verstege, M. I., Ambrosini, M., Kalay, H., Nazmi, K., Bolscher, J. G., Hooijberg, E., de Gruijl, T. D., Storm, G., & van Kooyk, Y. (2012). Glycan‐modified liposomes boost CD4+ and SUPP ORTING INFO RM ATION CD8+ T‐cell responses by targeting DC‐SIGN on dendritic cells. Additional supporting information may be found in the online version Journal of Controlled Release, 160, 88–95. Varki, A. (2017). Biological roles of glycans. Glycobiology, 27, 3–49. of the article at the publisher’s website. Vigerust, D. J., & Shepherd, V. L. (2007). Virus glycosylation: Role in virulence and immune interactions. Trends in Microbiology, 15, 211–218. Wagner, J. M., Pajerowski, J. D., Daniels, C. L., McHugh, P. M., How to cite this article: Lavado‐García, J., Zhang, T., Cervera, Flynn, J. A., Balliet, J. W., Casimiro, D. R., & Subramanian, S. (2014). L., Gòdia, F., & Wuhrer, M. (2022). Differential N‐ and Enhanced production of chikungunya virus‐like particles using a O‐glycosylation signatures of HIV‐1 Gag virus‐like particles high‐pH adapted spodoptera frugiperda insect cell line. PLoS One, and coproduced extracellular vesicles. Biotechnology and 9, e94401. Watanabe, Y., Bowden, T. A., Wilson, I. A., & Crispin, M. (2019). Bioengineering, 119, 1207–1221. Exploitation of glycosylation in enveloped virus pathobiology. https://doi.org/10.1002/bit.28051 Biochimica et Biophysica Acta—General Subjects, 1863, 1480–1497. Wei, X., Decker, J. M., Wang, S., Hui, H., Kappes, J. C., Wu, X., Salazar‐ Gonzalez, J. F., Salazar, M. G., Kilby, J. M., Saag, M. S.,