Discovery and characterization of a novel extremely acidic bacterial N-glycanase with combined advantages of PNGase F and A

Peptide-N4-(N-acetyl-β-glucosaminyl) asparagine amidases [PNGases (peptide N-glycosidases), N-glycanases, EC 3.5.1.52] are essential tools in the release of N-glycans from glycoproteins. We hereby report the discovery and characterization of a novel bacterial N-glycanase from Terriglobus roseus with an extremely low pH optimum of 2.6, and annotated it therefore as PNGase H⁺. The gene of PNGase H⁺ was cloned and the recombinant protein was successfully expressed in Escherichia coli. The recombinant PNGase H⁺ could liberate high mannose-, hybrid- and complex-type N-glycans including core α1,3-fucosylated oligosaccharides from both glycoproteins and glycopeptides. In addition, PNGase H⁺ exhibited better release efficiency over N-glycans without core α1,3-fucose compared with PNGase A. The facile expression, non-glycosylated nature, unusual pH optimum and broad substrate specificity of this novel type of N-glycanase makes recombinant PNGase H⁺ a versatile tool in N-glycan analysis.     

High-throughput quantitative analysis of plant N-glycan using a DNA sequencer

High-throughput quantitative analytical method for plant N-glycan has been developed. All steps, including peptide N-glycosidase (PNGase) A treatment, glycan preparation, and exoglycosidase digestion, were optimized for high-throughput applications using 96-well format procedures and automatic analysis on a DNA sequencer. The glycans of horseradish peroxidase with plant-specific core α(1,3)-fucose can be distinguished by the comparison of the glycan profiles obtained via PNGase A and F treatments. The peaks of the glycans with (91%) and without (1.2%) α(1,3)-fucose could be readily quantified and shown to harbor bisecting β(1,2)-xylose via simultaneous treatment with α(1,3)-mannosidase and β(1,2)-xylosidase. This optimized method was successfully applied to analyze N-glycans of plant-expressed recombinant antibody, which was engineered to contain a minor amount of glycan harboring β(1,2)-xylose. These results indicate that our DNA sequencer-based method provides quantitative information for plant-specific N-glycan analysis in a high-throughput manner, which has not previously been achieved by glycan profiling based on mass spectrometry.     

Plant N-glycan profiling of minute amounts of material

Development of convenient strategies for identification of plant N-glycan profiles has been driven by the emergence of plants as an expression system for therapeutic proteins. In this article, we reinvestigated qualitative and quantitative aspects of plant N-glycan profiling. The extraction of plant proteins through a phenol/ammonium acetate procedure followed by deglycosylation with peptide N-glycosidase A (PNGase A) and coupling to 2-aminobenzamide provides an oligosaccharide preparation containing reduced amounts of contaminants from plant cell wall polysaccharides. Such a preparation was also suitable for accurate qualitative and quantitative evaluation of the N-glycan content by mass spectrometry. Combining these approaches allows the profiling to be carried out from as low as 500 mg of fresh leaf material. We also demonstrated that collision-induced dissociation (CID) mass spectrometry in negative mode of N-glycans harboring α(1,3)- or α(1,6)-fucose residue on the proximal GlcNAc leads to specific fragmentation patterns, thereby allowing the discrimination of plant N-glycans from those arising from mammalian contamination.     

Jack bean α-mannosidase digestion profile of hybrid-type N-glycans: effect of reaction pH on substrate preference

Jack bean α-mannosidase (JBM) is a well-studied plant vacuolar α-mannosidase, and is widely used as a tool for the enzymatic analysis of sugar chains of glycoproteins. In this study, the JBM digestion profile of hybrid-type N-glycans was examined using pyridylamino (PA-) sugar chains. The digestion efficiencies of the PA-labeled hybrid-type N-glycans Manα1,6(Manα1,3)Manα1,6(GlcNAcβ1,2Manα1,3)Manβ1,4GlcNAcβ1,4GlcNAc-PA (GNM5-PA) and Manα1,6(Manα1,3)Manα1,6(Galβ1,4GlcNAcβ1,2Manα1,3)Manβ1,4GlcNAcβ1,4GlcNAc-PA (GalGNM5-PA) were significantly lower than that of the oligomannose-type N-glycan Manα1,6(Manα1,3)Manα1,6Manβ1,4GlcNAcβ1,4GlcNAc-PA (M4-PA), and the trimming pathways of GNM5-PA and GalGNM5-PA were different from that of M4-PA, suggesting a steric hindrance to the JBM activity caused by GlcNAcβ1-2Man(α) residues of the hybrid-type N-glycans. We also found that the substrate preference of JBM for the terminal Manα1-6Man(α) and Manα1-3Man(α) linkages in the hybrid-type N-glycans was altered by the change in reaction pH, suggesting a pH-dependent change in the enzyme-substrate interaction.     

Molecular identification and characterization of peptide: N-glycanase from Schizosaccharomyces pombe

Peptide:N-glycanase (PNGase) is an enzyme responsible for deglycosylation of misfolded glycoproteins in so-called endoplasmic reticulum-associated degradation (ERAD) system. In this study, we reported the molecular identification and characterization of SpPNGase (Schizosaccharomyces pombe PNGase). Enzymatic analysis revealed that SpPNGase deglycosylated the misfolded glycoproteins and distinguished native and denatured high-mannose glycoproteins in vitro. The deglycosylation activity was lost with the addition of chelating agent EDTA and was not restored by re-addition of metal ions. By construction of deletion mutant, we confirmed that N-terminal α-helix of SpPNGase was responsible for the protein-protein interaction. Combining the results from ternary structure prediction and dendrogram analysis, we suggested that the N-terminal α-helices of PNGase are derived from evolutionary motif/peptide fusion.     

Malonic acid suppresses mucin-type O-glycan degradation during hydrazine treatment of glycoproteins

Hydrazine treatment is frequently used for releasing mucin-type O-glycans (O-glycans) from glycoproteins because the method provides O-glycans that retain a reducible GalNAc at their reducing end, which is available for fluorescent labeling. However, many O-glycans are degraded by “peeling” during this treatment. In the current study, it was found that malonic acid suppressed O-glycan degradation during hydrazine treatment of bovine fetuin or porcine gastric mucin in both the gas and liquid phases. This is paradoxical because the release of O-glycans from glycoproteins occurs under alkaline conditions. However, malonic acid seems to prevent the degradation through its acidic property given that other weak acids also prevented the degradation. Accordingly, disodium malonate did not suppress O-glycan degradation. Application of this method to rat gastric mucin demonstrated that the majority of the major O-glycans obtained in the presence of malonic acid were intact, whereas those obtained in the absence of malonic acid were degraded. These results suggest that hydrazine treatment in the presence of malonic acid would allow glycomic analysis of native mucin glycoproteins.     

Comparison of procainamide and 2-aminobenzamide labeling for profiling and identification of glycans by liquid chromatography with fluorescence detection coupled to electrospray ionization–mass spectrometry

One of the most widely used methods for glycan analysis is fluorescent labeling of released glycans followed by hydrophilic interaction chromatography–(ultra-)high-performance liquid chromatography [HILIC–(U)HPLC]. Here, we compare the data obtained by (U)HPLC–fluorescence (FLR) coupled to electrospray ionization–mass spectrometry (ESI–MS) for procainamide and 2-aminobenzamide (2-AB)-labeled N-glycans released from human immunoglobulin G (IgG). Fluorescence profiles from procainamide show comparable chromatographic separation to those obtained for 2-AB but gave higher fluorescence intensity as well as significantly improved ESI efficiency (up to 30 times that of 2-AB). Thus, labeling with procainamide increases the ability to identify minor glycan species that may have significant biological activity.     

Monosialylated biantennary N-glycoforms containing GalNAc-GlcNAc antennae predominate when human EPO is expressed in goat milk

Recently, our group reported the expression of recombinant human erythropoietin in goat milk (rhEPO-milk) as well as in the mammary epithelial cell line GMGE (EPO-GMGE) by cell culture using the adenoviral transduction system. N-Glycosylation characterization of rhEPO-milk by Normal-Phase HPLC profiling of the fluorophore, 4-aminobenzoic acid-labeled enzymatically released N-glycan pool from rhEPO-goat milk, combined with MALDI, ESI-MS and LC/MS, revealed that low branched, core-fucosylated, N-glycans predominate. The labeled N-glycans were separated into neutral and charged fractions by anion exchange chromatography and the charged N-glycans were found to be mostly α2,6-monosialylated with Neu5Ac or Neu5Gc in a ratio of 1:1. Unlike the N-glycans from rhEPO produced in CHO cells, where the glycans are multiantennary highly sialylated, core-fucosylated oligosaccahrides, or even in the goat mammary gland epithelial cell line cultured in vitro in which multiantennary, core- and outer-arm fucosylated, monosialylated N-glycans are the most abundant species, a large proportion of the N-glycans from rhEPO-milk were monosialylated, biantennary, antennae mostly terminating with the more unusual GalNAc-GlcNAc motive and without outer-arm fucosylation. These findings, emphasizing the difference in the N-glycan repertoire between the rhEPO-milk and EPO-GMGE, are consistent with the principle that glycosylation is cell-type dependent and that the cell environment is crucial as well.     

Building a PGC-LC-MS <Emphasis Type="Italic">N</Emphasis>-glycan retention library and elution mapping resource

Porous graphitised carbon-liquid chromatography (PGC-LC) has been proven to be a powerful technique for the analysis and characterisation of complex mixtures of isomeric and isobaric glycan structures. Here we evaluate the elution behaviour of N-glycans on PGC-LC and thereby provide the potential of using chromatographic separation properties, together with mass spectrometry (MS) fragmentation, to determine glycan structure assignments more easily. We used previously reported N-glycan structures released from the purified glycoproteins Immunoglobulin G (IgG), Immunoglobulin A (IgA), lactoferrin, α1-acid glycoprotein, Ribonuclease B (RNase B), fetuin and ovalbumin to profile their behaviour on capillary PGC-LC-MS. Over 100 glycan structures were determined by MS/MS, and together with targeted exoglycosidase digestions, created a N-glycan PGC retention library covering a full spectrum of biologically significant N-glycans from pauci mannose to sialylated tetra-antennary classes. The resultant PGC retention library (http://www.glycostore.org/showPgc) incorporates retention times and supporting fragmentation spectra including exoglycosidase digestion products, and provides detailed knowledge on the elution properties of N-glycans by PGC-LC. Consequently, this platform should serve as a valuable resource for facilitating the detailed analysis of the glycosylation of both purified recombinant, and complex mixtures of, glycoproteins using established workflows.     

The ammonia-catalyzed release of glycoprotein <Emphasis Type="Italic">N</Emphasis>-glycans

Despite the great significance of release and analysis of glycans from glycoproteins, the existing N-glycan release methods are undermined by some limitations and deficiencies. The traditional enzymatic protocols feature high N-glycan release specificity but are generally costly and inefficient for some types of N-glycans. The existing chemical methods require harsh reaction conditions or are accompanied by the remarkable formation of by-products. Herein, we describe a versatile chemical method for the release and analysis of N-glycans from glycoproteins. This method differs from the existing methods as only aqueous ammonia is used to catalyze the N-glycan release reactions. Optimization of reaction conditions was performed using RNase B as a model glycoprotein and the obtained results indicated a highest N-glycan yield in ammonia at 60 °C for 16 h. Comparison of this method with traditional enzymatic protocols and recently reported NaClO methods confirmed the good reliability and efficiency of the novel approach. We also successfully applied this method to some complex biological samples, such as Ginkgo seed protein, fetal bovine serum (FBS) and hen egg white, and demonstrated its great compatibility with various neutral N-glycans, core α-1,3-fucosylated N-glycans and sialylated N-glycans. This method is very simple and cost-effective, enabling convenient analysis and large-scale preparation of released reducing N-glycans from various biological samples for structural and functional glycomics studies.     

Chemical and enzymatic N-glycan release comparison for N-glycan profiling of monoclonal antibodies expressed in plants

Plants synthesize N-glycans containing the antigenic sugars α(1,3)-fucose and β(1,2)-xylose. Therefore it is important to monitor these N-glycans in monoclonal antibodies produced in plants (plantibodies). We evaluated several techniques to characterize the N-glycosylation of a plantibody produced in tobacco plants with and without the KDEL tetrapeptide endoplasmic reticulum retention signal which should inhibit or drastically reduce the addition of α(1,3)-fucose and β(1,2)-xylose. Ammonium hydroxide/carbonate-based chemical deglycosylation and PNGase A enzymatic release were investigated giving similar 2-aminobenzamide-labeled N-glycan HPLC profiles. The chemical release does not generate peptides which is convenient for MS analysis of unlabeled pool but its main drawback is that it induces degradation of α1,3-fucosylated N-glycan reducing terminal sugar. Three analytical methods for N-glycan characterization were evaluated: (i) MALDI-MS of glycopeptides from tryptic digestion; (ii) negative-ion ESI-MS/MS of released N-glycans; (iii) normal-phase HPLC of fluorescently labeled glycans in combination with exoglycosidase sequencing. The MS methods identified the major glycans, but the HPLC method was best for identification and relative quantitation of N-glycans. Negative-mode ESI-MS/MS permitted also the correct identification of the linkage position of the fucose residue linked to the inner core N-acteylglucosamine (GlcNAc) in complex N-glycans.     

Systematic synthesis and inhibitory activity of haloacetamidyl oligosaccharide derivatives toward cytoplasmic peptide:<Emphasis Type="Italic">N</Emphasis>-glycanase

A series of glycosyl haloacetamides were synthesized as potential inhibitors of cytoplasmic peptide:N-glycanase (PNGase), an enzyme that removes N-glycans from misfolded glycoproteins. Chloro-, bromo-, and iodoacetamidyl chitobiose and chitotetraose derivatives exhibited a significant inhibitory activity. No inhibitory activity was observed with of fluoroacetamididyl derivatives. Moreover, N-acetylglucosamine derivatives, β-chloropropionamidyl chitobiose, and chloroacetamidyl cellooligosaccharide derivatives did not show any activity. These results underscore the importance of the N-acetyl groups of chitobiose for PNGase recognition. In addition, reactivity and position of the leaving group at the reducing end are also important factors.     

ABH blood group antigens in N-glycan of human glycophorin A

We previously showed that a small proportion of the O-linked oligosaccharide chains of human glycophorin A (GPA) contains blood group A, B or H antigens, relevant to the ABO phenotype of the donor. The structures of these minor O-glycans have been established (Podbielska et al. (2004) [20]). By the use of immunochemical methods we obtained results indicating that ABH blood group epitopes are also present in N-glycan of human GPA (Podbielska and Krotkiewski (2000) [22]). In the present paper we report a detailed analysis of GPA N-glycans using nanoflow electrospray ionization tandem mass spectrometry. N-glycans containing A-, B- and H-related sequences were identified in GPA preparations obtained from erythrocytes of blood group A, B and O donors, respectively. The ABH blood group epitopes are present on one antenna of the N-glycan, whereas a known sialylated sequence NeuAcα2-6Galβ1-4GlcNAc- occurs on the other antenna and other details are in agreement with the known major structure of the GPA N-glycan. In the bulk of the biantennary sialylated N-glycans released from GPA preparations, the blood group ABH epitopes-containing N-glycans, similarly O-glycans, constituted only a minor part. The amount relative to other N-glycans was estimated to 2-6% of blood group H epitope-containing glycans released from GPA-O preparations and 1-2% of blood group A and B epitope-containing glycans, released from GPA-A and GPA-B, respectively.     

Salt-Adaptation of Tobacco BY2 Cells Induces Change in Glycoform of N-Glycans: Enhancement of Exo- and Endo-Glycosidase Activities by Salt-Adaptation

In this report, we describe that a salt adaptation of plant cells induces glycoform changes in N-glycoproteins. Intracellular and cell-wall glycopeptides were prepared from glycoproteins expressed in wild-type BY2 cells and salt-adapted cells. N-Glycans were liberated from those glycopeptides by hydrazinolysis, and the released oligosaccharides were N-acetylated and pyridylaminated. The structures of pyridylaminated (PA-) N-glycans were analyzed by a combination of two-dimensional sugar-chain mapping, MS analysis, and exoglycosidase digestion. In both wild-type cells and salt-adapted cells, the plant complex type structure was predominant among N-glycans expressed on glycoproteins, but we found that the Man2Xyl1Fuc1GlcNAc2 structure was significantly expressed on intracellular and cell-wall glycoproteins of the salt-adapted cells. Furthermore, enhancement of the specific activities of α-mannosidase and β-N-acetylglucosaminidase was observed in the salt-adapted BY2 cells, suggesting that the glycoform changes are due to changes in glycosidase activities.     

Improved nonreductive O-glycan release by hydrazinolysis with ethylenediaminetetraacetic acid addition

The study of protein O-glycosylation is receiving increasing attention in biological, medical, and biopharmaceutical research. Improved techniques are required to allow reproducible and quantitative analysis of O-glycans. An established approach for O-glycan analysis relies on their chemical release in high yield by hydrazinolysis, followed by fluorescent labeling at the reducing terminus and high-performance liquid chromatography (HPLC) profiling. However, an unwanted degradation known as “peeling” often compromises hydrazinolysis for O-glycan analysis. Here we addressed this problem using low-molarity solutions of ethylenediaminetetraacetic acid (EDTA) in hydrazine for O-glycan release. O-linked glycans from a range of different glycoproteins were analyzed, including bovine fetuin, bovine submaxillary gland mucin, and serum immunoglobulin A (IgA). The data for the O-glycans released by hydrazine with anhydrous EDTA or disodium salt dihydrate EDTA show high yields of the native O-glycans compared with the peeled product, resulting in a markedly increased robustness of the O-glycan profiling method. The presented method for O-glycan release demonstrates significant reduction in peeling and reduces the number of sample handling steps prior to release.     

Tracing the history of Galα1–4Gal on glycoproteins in modern birds

Galα1–4Gal is typically found in mammalian glycolipids in small quantities, and recognized by some pathogens, such as uropathogenic Escherichia coli. In contrast, glycoproteins containing Galα1–4Gal were rarely found in vertebrates except in a few species of birds and amphibians until recently. However, we had previously reported that pigeon (Columba livia) egg white and serum glycoproteins are rich in N-glycans with Galα1–4Gal at non-reducing termini. Our investigation with egg white glycoproteins from 181 avian species also revealed that the distribution of (Galα1–4Gal)-containing glycoproteins was not rare among avians, and is correlated with the phylogeny of birds. The differentiated expression was most likely emerged at earlier stage of diversification of modern birds, but some birds might have lost the facility for the expression relatively recently.     

Purification of Derivatized Oligosaccharides by Solid Phase Extraction for Glycomic Analysis

Profiling of glycans released from proteins is very complex and important. To enhance the detection sensitivity, chemical derivatization is required for the analysis of carbohydrates. Due to the interference of excess reagents, a simple and reliable purification method is usually necessary for the derivatized oligosaccharides. Various SPE based methods have been applied for the clean-up process. To demonstrate the differences among these methods, seven types of self-packed SPE cartridges were systematically compared in this study. The optimized conditions were determined for each type of cartridge and it was found that microcrystalline cellulose was the most appropriate SPE material for the purification of derivatized oligosaccharide. Normal phase HPLC analysis of the derivatized maltoheptaose was realized with a detection limit of 0.12 pmol (S N⁻¹ = 3) and a recovery over 70%. With the optimized SPE method, relative quantification analysis of N-glycans from model glycoproteins were carried out accurately and over 40 N-glycans from human serum samples were determined regardless of the isomers. Due to the high stability and sensitivity, microcrystalline cellulose cartridge showed potential applications in glycomics analysis.     

Inhibitory and Thermodynamic Studies on the Hydrolysis of Cyclodextrins Catalyzed by Taka-amylase A

The structures of N-glycans of total glycoproteins in royal jelly have been explored to clarify whether antigenic N-glycans occur in the famous health food. The structural feature of N-glycans linked to glycoproteins in royal jelly was first characterized by immunoblotting with an antiserum against plant complex type N-glycan and lectin-blotting with Con A and WGA. For the detail structural analysis of such N-glycans, the pyridylaminated (PA-) N-glycans were prepared from hydrazinolysates of total glycoproteins in royal jelly and each PA-sugar chain was purified by reverse-phase HPLC and size-fractionation HPLC. Each structure of the PA-sugar chains purified was identified by the combination of two-dimensional PA-sugar chain mapping, ESI-MS and MS/MS analyses, sequential exoglycosidase digestions, and 500 MHz ¹H-NMR spectrometry.

The immunoblotting and lectinblotting analyses preliminarily suggested the absence of antigenic N-glycan bearing β1-2 xylosyl and/or α1-3 fucosyl residue(s) and occurrence of β1-4GlcNAc residue in the insect glycoproteins.

The detailed structural analysis of N-glycans of total royal jelly glycoproteins revealed that the antigenic N-glycans do not occur but the typical high mannose-type structure (Man_9~4GlcNAc₂) occupies 71.6% of total N-glycan, biantennary-type structures (GlcNAc₂Man₃GlcNAc₂) 8.4%, and hybrid type structure (GlcNAc₁Man₄GlcNAc₂) 3.0%. Although the complete structures of the remaining 17% N-glycans; C4, (HexNAc₃Hex₃HexNAc₂: 3.0%), D2 (HexNAc₂Hex₅HexNAc₂: 4.5%), and D3 (HexNAc₃Hex₄HexNAc₂: 9.5%) are still obscure so far, ESI-MS analysis, exoglycosidase digestions by two kinds of β-N-acetylglucosaminidase, and WGA blotting suggested that these N-glycans might bear a β1-4 linkage N-acetylglucosaminyl residue.     

Comparison of two approaches for quantitative O-linked glycan analysis used in characterization of recombinant proteins

The principal aim of this study was to demonstrate the optimization and fine-tuning of quantitative and nonselective analysis of O-linked glycans released from therapeutic glycoproteins. Two approaches for quantitative release of O-linked glycans were examined: ammonia-based β-elimination and hydrazinolysis deglycosylation strategies. A significant discrepancy in deglycosylation activity was observed between the ammonia-based and hydrazinolysis procedures. Specifically, the release of O-glycans from glycoproteins was approximately 20 to 30 times more efficient with hydrazine compared with ammonia-based β-elimination reagent. In addition, the ammonia-based reagent demonstrated bias in the release of particular glycan species. A robust quantitative hydrazinolysis procedure was developed for characterization of O-glycans. The method performance parameters were evaluated. It was shown that this procedure is superior for quantitative nonselective release of O-glycans. Identity confirmation and structure elucidation of O-glycans from hydrophilic interaction chromatography (HILIC) fractions was also demonstrated using linear ion trap Fourier transform mass spectrometry (LTQ FT MS) with mass accuracy below 1 ppm.     

Glycan analysis of glycoprotein pharmaceuticals: Evaluation of analytical approaches to Z number determination in pharmaceutical erythropoietin products

N-Glycosylation of many glycoprotein drugs is important for biological activity and should therefore be the target of specific and quantitative analytical methods. In this study, we focus on the two N-glycan mapping approaches that are used in pharmacopoeial monograph to analyse N-glycans released from fifteen preparations of recombinant human erythropoietin supplied by ten Chinese manufacturers. Underivatised N-glycans were analysed by high performance anion-exchange chromatography with pulsed amperometric detection and fluorophore-labelled N-glycans were analysed by weak anion-exchange and normal-phase high performance liquid chromatography. N-glycans were also analysed by matrix assisted laser desorption ionisation mass spectrometry. The release of N-glycans by PNGase F was shown to be consistent. Z number, a mathematical expression of the total negatively charged N-glycans composition has provided a convenient way to summarise the complex dataset and it might be suitable for product consistency monitoring. However, this Z number reduces the information of individual acidic N-glycan structure and is also found to be method dependent. Therefore, its use requires clear specification and validation. In this study, we only found weak but positive correlation between the Z number and its bioactivity. Wide range of N-glycans yields were obtained from the fifteen preparations but the significance of their differences is unclear.