Optimizing HILIC-based Analyses of RapiFluor-MS Labeled Sialylated N-Glycans

This application note describes applied hydrophilic interaction chromatography for the analysis of RapiFluor-MS labeled sialylated N-glycans.

Here we have demonstrated an improvement in chromatographic resolution of sialylated N-glycans through the use of higher ionic strength mobile phases with a wide pore ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 μm Column. Moreover, it has been shown that these methods can be effectively combined with neuraminidase digestion to elucidate linkage specific information about the sialic acid content of a sample. Together, the GlycoWorks RapiFluor-MS N-Glycan Kit, RapiFluor-MS Sialylated Glycan Performance Test Standard, and wide-pore ACQUITY UPLC Glycoprotein BEH Amide Column establish a very useful collection of tools for characterizing sialylated N-glycans.    

Benefits

• Tuned source parameters and acquisition settings to improve the intensity and quality of MS information for RapiFluor-MS. labeled sialylated N-glycans
• Optimized mobile phases used along with a wide-pore ACQUITY UPLC Glycoprotein BEH Amide Column to enhance HILIC chromatographic resolution
• Linkage-specific neuraminidase profiling of RapiFluor-MS Sialylated Glycan Performance Test Standard to demonstrate a cursory sialic acid linkage analysis

Sialic acids, a family of acidic monosaccharides, are often found as terminal residues of glycan structures and they play important roles in many physiological and pathological processes.¹ With respect to therapeutics, sialic acid residues are of interest because they can impact serum half-lives^2,3 and they can serve as indicators of cell culture conditions.^4,5 N-Glycolylneuraminic acid, a non-human mammalian sialic acid with a structure similar to the human sialic acid N-acetylneuraminic acid, can contaminate biotherapeutic glycoproteins via production in non-human cell lines or from animal-derived media,^6,7 which will result in different levels of immune responses in humans.^8,9 Accordingly, sialylated glycan profiling is often needed to monitor a critical quality attribute related either to drug stability, activity, or immunogenicity.¹⁰ Due to the lability of sialic acid residues, accurate mass spectrometric analysis of sialylated glycans has sometimes been elusive or required neutralization of the sialic acids. The recent development of kits for the rapid release and labeling of N-glycans via GlycoWorks RapiFluor-MS enables enhanced sensitivity for fluorescence (FLR) and mass spectrometry (MS) based detection of all types of N-glycan species, including those that are sialylated.¹¹

In this work, we applied hydrophilic interaction chromatography (HILIC)-FLR-MS analyses to the study of RapiFluor-MS labeled sialylated N-glycans. A wide-pore (300Å) amide HILIC stationary phase was used with high ionic strength mobile phases to improve chromatographic resolution for sialylated N-glycans. Combined with optimized electrospray ionization (ESI) source parameters, a high resolution separation was achieved along with high MS sensitivity. This analysis produces data of unparalleled quality, yet, as we additionally show, it is also readily amenable to being paired with neuraminidase treatments as a means to perform a cursory level sialic acid linkage analysis.

It is important to note that the HILIC conditions optimized in this study involving a wide-pore amide HILIC stationary phase and high ionic strength mobile phases are different from what have been used for generating the RapiFluor-MS Glycan GU Scientific Library, thus the retention time obtained under these conditions cannot be searched against the library for glycan identification. In addition, extended use of higher ionic strength mobile phases can quicken the fouling of the source/ion optics of a mass spectrometer, so system sensitivity monitoring and routine maintenance is recommended to ensure optimal results. 

Sample description

RapiFluor-MS Sialylated Glycan Performance Test Standard (p/n 186008660) was reconstituted in 50 µL of water. 

α2–3 Neuraminidase S digestion

One vial of RapiFluor-MS Sialylated Glycan Performance Test Standard was reconstituted in 9 µL of water, followed by sequential addition of 1 µL of GlycoBuffer 1 (10×) and 1 µL of α2–3 Neuraminidase S (New England BioLabs, p/n P0743). After overnight incubation at 37 °C, 39 µL of water was added to the solution.

α2–3,6,8,9 Neuraminidase A digestion

One vial of RapiFluor-MS Sialylated Glycan Performance Test Standard was reconstituted in 9 µL of water, followed by sequential addition of 1 µL of GlycoBuffer 1 (10×) and 1 µL of α2–3,6,8,9 Neuraminidase A (New England BioLabs, p/n P0722). After overnight incubation at 37 °C, 39 µL of water was added to the solution.

To achieve 50 mM ammonium formate in aqueous mobile phases A and C (Figure 3A and 4A):

To achieve 100 mM ammonium formate in aqueous mobile phases A and C (Figure 3B):

To achieve 200 mM ammonium formate in aqueous mobile phases A and C (Figure 3C, 4B, 5, and 6):

Optimization of Esi-Ms Source Parameters to Improve the Ion Intensities of Protonated RapiFluor-Ms Labeled Sialylated N-Glycans

In a positive ion mode analysis, multiple cation adducts of glycans can be readily observed if poorly optimized conditions are used. This, in turn, can dilute the overall ion intensity across several species within the same charge state and makes interpretation of spectra challenging. Proton, ammonium, and sodium adducts are commonly encountered, as is the case with the mass spectra of RapiFluor-MS labeled sialylated N-glycans following hydrophilic interaction chromatography. Moreover, in-source fragments (ISFs) can occur during ionization, again to the detriment of sensitivity and data quality. In order to maximize the signal intensities of protonated ions of RapiFluor-MS labeled sialylated N-glycans, the effects of numerous ESI source parameters were investigated using LC-MS with an ACQUITY UPLC Glycan BEH Amide, 130Å, 1.7 μm, 2.1 x 50 mm Column coupled to a Xevo G2-XS QTof Mass Spectrometer. To facilitate this method development, we utilized a new standard, the RapiFluor-MS Sialylated Glycan Performance Test Standard (p/n 186008660) containing upward of 20 unique RapiFluor-MS labeled mono- to pentasialylated N-glycans released from bovine fetuin. We focused on manipulating the MS response for one of the most abundant glycans from this profile, an A3G3S3 species (Oxford notation)^.12,13 This is a structure comprised of a triantennary core (A3) modified with three galactose (G3) and three sialic acid (S3) residue extensions.

Of all the parameters studied herein, sampling cone voltage was found to produce the most significant effects, not only in the relative abundance of the protonated molecular ions, but also the intensities of ISFs. At lower sampling cone voltages, eg, 30 V, the signal intensities of the 3+ charge states of A3G3S3 were much higher than those of the 2+ charge states or ISFs (Figure 1A). Within the 3+ charge state window, multiple cation adducts of A3G3S3 were observed, and [A3G3S3+2H⁺+NH₄+]³⁺ gave the most abundant ion intensity (Figure 1A inset). When the sampling cone voltage was increased to 75 V, A3G3S3 ions in a 3+ charge state were still the most abundant, while the intensities of ions from in-source fragmentation slightly increased (Figure 1B). Within the 3+ charge state window, [A3G3S3+3H⁺]³⁺ gave the most abundant ion intensity (Figure 1B inset). When the sampling cone voltage was increased to 120 V, the intensities of molecular ions of A3G3S3 decreased significantly, meanwhile the mass spectrum became dominated by ions from in-source  fragmentation (Figure 1C and inset). A graph of the correlation between the normalized absolute abundance of [A3G3S3+3H⁺]³⁺ and sampling cone voltages revealed a trend with a bell-shaped distribution: as the sampling cone voltages increased from 30 V, the signal intensities of [A3G3S3+3H⁺]³⁺ increased, reached a maximum at a sampling cone voltage  of 75 V, and decreased thereafter because of increasing  levels of in-source fragmentation (Figure 1D). In terms of relative abundance, a similar trend was observed: the relative abundance of [A3G3S3+3H⁺]³⁺ was higher when the sampling cone voltage was set at 75 V versus either 30 or 120 V; when the sampling cone voltage was changed from 75 to 120 V, significantly higher relative abundances of ISFs were encountered (Figure 2A). In consideration of this, a sampling cone voltage of 75 V has been selected for the analysis of RapiFluor-MS labeled sialylated N-glycans.

The effects of desolvation temperature on signal intensities of protonated RapiFluor-MS labeled sialylated N-glycans was investigated in a similar way. The relative abundance of [A3G3S3+3H⁺]³⁺ slightly increased as the desolvation temperature was decreased from 650 to 350 °C, while no significant difference was observed regarding the relative intensities of ISFs (Figure 2B). A desolvation temperature between and including 350 and 500 °C appears optimal, although the absolute ion intensities near a 500 °C desolvation temperature were higher than those from a 350 °C desolvation temperature. Therefore, a desolvation temperature of 500 °C has been chosen for the analysis of RapiFluor-MS labeled sialylated N-glycans. Altering three other ESI source parameters, capillary voltage (Figure 2C), source temperature (Figure 2D), and desolvation gas flow (data not shown), did not result in any significant changes, neither in the relative abundance of [A3G3S3+3H⁺]³⁺ nor in the level of in-source fragmentation.

Based on the above observations, a sampling cone voltage of 75 V, desolvation temperature of 500 °C, capillary voltage of 2200 V, source temperature of 120 °C, and desolvation gas flow of 600 L/Hr have been chosen for the analysis of RapiFluor-MS labeled sialylated N-glycans, when using a Xevo G2-XS QTof Mass Spectrometer. Under these conditions, the protonated ions of A3G3S3 were optimized to be the most abundant molecular ions, and, at the same time, the formation of ISFs was minimized.

Optimized Mobile Phases and the use of Wide-Pore Amide Hilic for Enhanced Chromatographic Resolution of Rapifluor-Ms Labeled Sialylated N-Glycans

Just as it was used to optimize the quality of MS data, the RapiFluor-MS Sialylated Glycan Performance Test Standard has also been used for the development of an improved chromatographic separation of sialylated N-glycans. To begin, we selected a wide-pore amide HILIC stationary phase, specifically an ACQUITY UPLC Glycoprotein BEH Amide  300Å, 1.7 µm phase, because of its demonstrated applicability to resolving large, released N-glycan structures.^14,15 The  HILIC separation obtained with such a phase, in the form of a 2.1 x 150 mm column, is shown in Figure 3A. Here, a typical aqueous mobile phase comprised of 50 mM ammonium formate buffer was employed. As can be seen, the peak shape and resolution obtained with this mobile phase was not ideal, particularly for the most strongly retained, tetrasiaylated species. HILIC is intended to be a separation driven by the hydrophilicity of analytes, yet it can be complicated by ionic secondary interactions. With this in mind, we explored the use of higher mobile phase ionic strengths. As the ammonium formate concentration of the aqueous mobile phase was increased from 50 to 200 mM, the HILIC retentivity of the glycans increased, as shown in the higher retention time for all the components. A similar trend was observed in previous research.¹⁶ Moreover, better peak shape and chromatographic resolution was observed with higher ionic strength mobile phases. For example, peaks 5 and 6, corresponding to isomers of A3S1G3S3 and A3G3S3, co-eluted with 50 mM ammonium formate, but were near baseline resolved when the ionic strength was increased to 200 mM ammonium formate (half-height resolution 1.32).

It is noteworthy that the benefit of improved chromatographic resolution coming from higher ionic strength mobile phases came at a price of a slight increase in the intensity of background ions. Analyzed under the same mass spectrometric parameters, the absolute signal intensities of [A3G3S3+3H⁺]³⁺ were marginally suppressed with the shift from 50 to 200 mM ammonium formate (Figure 4). More notably,  the intensity of a 1+ ion at m/z 708 was higher  in the mass spectrum acquired with 200 versus 50 mM ammonium formate. As such, the increased background ion intensities from the higher ionic strength mobile phase obscured the glycan profile, as shown in the observed base peak intensity (BPI) chromatogram (Figure 5A). To circumvent this, the acquisition window was narrowed to 715 to 2000 m/z. In turn, a HILIC-FLR-MS method for sialylated N-glycans was developed that provides effective chromatographic resolution along with high quality MS information, even for very low abundant glycans (Figure 5B).

Use of Neuraminidases for Cursory Level Sialic Acid Linkage Elucidation

The improvements in the HILIC-FLR-MS method outlined above are beneficial to interrogating the compositions of a glycan profile containing sialylated species. Linkage analysis of terminal sialic acid residues is of great importance due to their critical roles in binding specificities. Avian influenza preferentially recognizes α2–3-linked over α2–6-linked sialic acids. Interestingly, humans have innate protection against the virus since α2–6-linked sialic acids are enriched on the airway epithelium of humans.¹⁷ Sialic acid linkages are also of interest in the development of therapeutics. Therapeutic glycoproteins produced in Chinese hamster ovary (CHO) cells only possess α2–3-linked sialic acid residues due to the lack of α2,6-sialyltransferases. Efforts have been made to engineer cell lines for the production of glycoproteins with both α2–3- and α2–6-linked sialic acid residues, as such is more similar to human glycosylation.¹⁸

As mentioned before, the RapiFluor-MS Sialylated Glycan Performance Test Standard (p/n 186008660) is comprised of a series of RapiFluor-MS labeled sialylated N-glycans released from bovine fetuin. Though it contains various glycans at a wide range of relative abundances, it clearly has 7 predominant species from 3 glycan compositions, as labeled and displayed in Figure 6A. The glycan profile of bovine fetuin also contains a mix of both α2–3- and α2–6-linked sialic acids, which have been reported to be present at ratios ranging from 38:62 to 51:49, depending on the detection and sample preparation techniques used.^19-21 Exoglycosidases, enzymes that break the glycosidic bonds at the non-reducing ends of glycans, have been widely applied for structural elucidation of complex glycans. In this study, α2–3 Neuraminidase S and α2–3,6,8,9 Neuraminidase A were used as exoglycosidases to selectively release sialic acid residues from the glycans of the RapiFluor-MS Sialylated Glycan Performance Test Standard. With this, a cursory level linkage elucidation assay can be demonstrated.

Before neuraminidase digestion, the majority of glycans in the RapiFluor-MS Sialylated Glycan Performance Test Standard  were found to be trisialylated (72%) and tetrasialylated (17%) triantennary glycans or disialylated biantennary glycans (11%) (Figure 6A). Overnight digestion with a neuraminidase specific to α2–3-linked sialic acids (α2–3 Neuraminidase S) resulted in a chromatographic profile with six abundant peaks from five glycan compositions: two with biantennary structures (12%; A2G2S1 and A2G2S2), and the other three with triantennary structures (88%; A3G3S1, A3G3S2, and A3G3S3) (Figure 6B). No tetrasialylated glycan peaks were seen to remain after α2–3 Neuraminidase S digestion. Under the assumption that α2–3 Neuraminidase S performs specific and complete cleavage of α2–3-linked sialic acids, it can be proposed that about two thirds of A2G2S2 contain one α2–3- and one α2–6-linked sialic acid residue and that one third contains two α2–6-linked sialic acids. The amounts of A2G2S2 with two α2–3-linked sialic acids appeared to be minimal. For triantennary N-glycans, it seems that the majority contain one (43%) or two (40%) α2–6-linked sialic acids. After overnight digestion with the broad specificity α2–3,6,8,9 Neuraminidase A, three chromatographic peaks from two glycan compositions were observed, biantennary A2G2 (11%) and triantennary A3G3 (89%) (Figure 6C). In sum, the overall ratio of α2–3- to α2–6-linked sialic acids observed herein is in  agreement with the range reported in the literature, and these chromatographic profiles provide a reference for future studies with this method.

Using the RapiFluor-MS Sialylated Glycan Performance Test Standard, we have optimized ESI-MS source parameters for improved ionization and detection of sialylated N-glycans. In addition, we have demonstrated an improvement in chromatographic resolution of sialylated N-glycans through the use of higher ionic strength mobile phases with a wide pore ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 µm Column. Moreover, it has been shown that these methods can be effectively combined with neuraminidase digestion to elucidate linkage specific information about the sialic acid content of a sample. Together, the GlycoWorks RapiFluor-MS N-Glycan Kit, RapiFluor-MS Sialylated Glycan Performance Test Standard, and wide-pore ACQUITY UPLC Glycoprotein BEH Amide Column establish a very useful collection of tools for characterizing sialylated N-glycans.

1. Varki, A. Sialic acids in human health and disease. Trends in Molecular Medicine 2008, 14 (8), 351–360.
2. Ashwell, G. & Harford, J. Carbohydrate-specific receptors of the liver. Annual Review of Biochemistry 1982, 51 (1), 531–554.
3. Weigel, P. H. & Yik, J. H. Glycans as endocytosis signals: the cases of the asialoglycoprotein and hyaluronan/chondroitin sulfate receptors. Biochimica et Biophysica Acta (BBA)-General Subjects 2002, 1572 (2), 341–363.
4. Yuen, C. T. et al. Relationships between the N‐glycan structures and biological activities of recombinant human erythropoietins produced using different culture conditions and purification procedures. British Journal ofHaematology 2003, 121 (3), 511–526.
5. Joosten, C. E. & Shuler, M. L. Effect of culture conditions on the degree of sialylation of a recombinant glycoprotein expressed in insect cells. Biotechnology Progress 2003, 19 (3), 739–749.
6. Martin, M. J., Muotri, A., Gage, F. & Varki, A. Human embryonic stem cells express an immunogenic nonhuman sialic acid. Nature Medicine 2005, 11 (2), 228–232.
7. Bardor, M., Nguyen, D. H., Diaz, S. & Varki, A. Mechanism of uptake and incorporation of the non-human sialic acid N-glycolylneuraminic acid into human cells. Journal of Biological Chemistry 2005, 280 (6), 4228–4237.
8. Zhu, A. & Hurst, R. Anti‐N‐glycolylneuraminic acid antibodies identified in healthy human serum. Xenotransplantation 2002, 9 (6), 376–381.
9. Nguyen, D. H., Tangvoranuntakul, P. & Varki, A. Effects of natural human antibodies against a nonhuman sialic acid that metabolically incorporates into activated and malignant immune cells. The Journal of Immunology 2005, 175 (1), 228–236.
10. Fournier, J. A Review of Glycan Analysis Requirements. BioPharm International 2015, 28 (10), 32–37.
11. Lauber, M. A. et al. Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Labeling Reagent that Facilitates Sensitive Fluorescence and ESI-MS Detection. Analytical Chemistry 2015, 87 (10), 5401–5409.
12. Harvey, D. J. et al. Proposal for a standard system for drawing structural diagrams of N‐and O‐linked carbohydrates and related compounds. Proteomics 2009, 9 (15), 3796–3801.
13. Glycobase 3.2., <http://glycobase.nibrt.ie> (accessed 6 January 2015).
14. Lauber, M. A. & Koza, S. M. Enhancing the Peak Capacity of High Molecular Weight N-Glycan HILIC Separations with a Wide-Pore Amide Bonded Stationary Phase. Waters Tech Brief 720005381en 2015.
15. Lauber, M. A., Koza, S. M. & Chambers, E. E. Comprehensive Characterization of the N and O-Linked Glycosylation of a Recombinant Human EPO. Waters Application Note 720005462en 2015.
16. Ahn, J., Bones, J., Yu, Y. Q., Rudd, P. M. & Gilar, M. Separation of 2-aminobenzamide labeled glycans using hydrophilic interaction chromatography columns packed with 1.7 μm sorbent. Journal of Chromatography B 2010, 878 (3), 403–408.
17. Stencel-Baerenwald, J. E., Reiss, K., Reiter, D. M., Stehle, T. & Dermody, T. S. The sweet spot: defining virus-sialic acid interactions. Nature Reviews Microbiology 2014, 12 (11), 739–749.
18. Lin, N. et al. Chinese hamster ovary (CHO) host cell engineering to increase sialylation of recombinant therapeutic proteins by modulating sialyltransferase expression. Biotechnology Progress 2015, 31 (2), 334–346.
19. Cointe, D., Leroy, Y. & Chirat, F. Determination of the sialylation level and of the ratio α-(2→ 3)/α-(2→ 6) sialyl linkages of N-glycans by methylation and GC/MS analysis. Carbohydrate Research 1998, 311 (1), 51–59.
20. Guttman, M. & Lee, K. K. Site-specific mapping of sialic acid linkage isomers by ion mobility spectrometry. Analytical Chemistry 2016, 88 (10), 5212–5217.
21. Green, E. D., Adelt, G., Baenziger, J., Wilson, S. & Van Halbeek, H. The asparagine-linked oligosaccharides on bovine fetuin. Structural analysis of N-glycanase-released oligosaccharides by 500-megahertz 1H NMR spectroscopy. Journal of Biological Chemistry 1988, 263 (34), 18253–18268.