Robustness of RapiFluor-MS N-Glycan Sample Preparations and Glycan BEH Amide HILIC Chromatographic Separations

In this application note, we demonstrate the robustness of RapiFluor-MS N-glycan sample preparation used together with Glycan BEH Amide HILIC Column chromatography.

The GlycoWorks RapiFluor-MS N-Glycan Kit enables analysts to perform a high yielding sample preparation with quantitative recovery that ensures accurate and repeatable profiling of N-glycans that is highly comparable to HPLC, 2-AB based methodologies. Moreover, it has been demonstrated that Glycan BEH Amide Columns afford exemplary resolving power and ruggedness for separations of RapiFluor-MS labeled N-glycans.

This separation can be readily transferred between UPLC and HPLC platforms. To further ensure success with these new methodologies, two standards have been commercialized, and their use to facilitate RapiFluor-MS analyses has been demonstrated. The RapiFluor-MS Glycan Performance Test Standard has been used for benchmarking studies, while the novel RapiFluor-MS Dextran Calibration Ladder has been employed to enhance the reproducibility of chromatographic retention time data.

Used together, the GlycoWorks RapiFluor-MS N-Glycan Kit and its supporting standards and columns can significantly reduce the burdens associated with N-glycan profiling while providing accurate, reproducible, and sensitive analyses.

Benefits

• High yield sample preparation with quantitative recovery to ensure accurate and repeatable profiling of N-glycans
• Comparability to historical 2-AB based released glycan analysis approaches
• RapiFluor-MS Glycan Performance Test Standard for method familiarization, troubleshooting, and benchmarking
• Robust Glycan BEH Amide HILIC separations supported by GU calibration with the novel RapiFluor-MS Dextran Calibration Ladder

N-glycosylation of proteins is routinely characterized and monitored because of its significance to the detection of disease states^1-3 and the manufacturing of biopharmaceuticals.^4-5 Glycosylation profiles are most often assessed by means of released glycan analyses, wherein samples are often prepared by techniques that are notoriously time-consuming or lead to compromises in MS sensitivity.^6-7 With the development of the GlycoWorks RapiFluor-MS N-Glycan Kit, we have addressed these shortcomings by enabling unprecedented sensitivity for glycan detection while also improving the throughput of N-glycan sample preparation.⁸ Using the GlycoWorks RapiFluor-MS N-Glycan Kit, glycoproteins are deglycosylated in 10 minutes to produce N-glycosylamines that are then rapidly reacted with the novel RapiFluor-MS labeling reagent (Figure 1). In a final step, the resulting labeled glycans are extracted from reaction byproducts by means of an SPE method that facilitates immediate analysis of samples. As a result, an analyst can now complete an N-glycan sample preparation, from glycoprotein to ready-to-analyze sample, in just 30 minutes and be poised to perform high sensitivity N-glycan profiling using hydrophilic interaction chromatography (HILIC) and mass spectrometric (MS) or fluorescence (FLR) detection.

Equally important as the efficiency and sensitivity gains afforded by this new sample preparation approach is its robustness and its ability to produce results consistent with historical N-glycan profiling. Within this application note, we will discuss these attributes of the RapiFluor-MS based sample preparation and the corresponding HILIC-based LC analyses.

Intact mass analysis was performed by LC-MS with a Xevo G2-QTof

Sample description

RapiFluor-MS labeled N-glycans were prepared from glycoproteins, including Intact mAb Mass Check Standard (p/n: 186006552), using a GlycoWorks RapiFluor-MS N-Glycan Kit (p/n: 176003606) according to the guidelines provided in its Care and Use Manual (715004793). 

2-AB labeled N-glycans were prepared using a Prozyme GlykoPrep Rapid N-Glycan Preparation with 2-AB Kit according to the manufacturer’s recommended protocol. In addition, 2-AB labeled N-glycans were also prepared using an approach combining the use of a Prozyme GlykoPrep Digestion Module, an in-house optimized 2-AB labeling protocol, and a GlykoPrep Cleanup Module (Prozyme, Hayward, CA).

RapiFluor-MS Glycan Performance Test Standard (p/n: 186007983) was reconstituted in 50 µL of water and injected as a 1 µL volume for chromatographic benchmarking and lifetime testing experiments. RapiFluor-MS Dextran Calibration Ladder (p/n: 186007982) was reconstituted in 100 µL of water and injected as a 1 µL volume for retention time calibrations.  

Percent yields for the sample preparation workflows were determined by means of quantitative analyses. Column loads were calibrated using external quantitative standards of 2-AB labeled triacetyl chitotriose  and RapiFluor-MS derivatized propylamine obtained in high purity (confirmed by HPLC and 1H NMR).  

To determine percent yields, the measured quantities of FA2 glycan from Intact mAb Mass Check Standard (p/n: 186006552) were compared to theoretical yields calculated for the preparation. For example, the theoretical yield for the FA2 glycan resulting from the GlycoWorks RapiFluor-MS N-Glycan Kit was calculated as follows:

Robust sample preparation: Deglycosylation Each procedural step in the GlycoWorks RapiFluor-MS  N-Glycan Kit has been optimized to be high yielding and to minimize the introduction of bias to an N-glycan profile. Previous work based on SDS PAGE gel shift assays has demonstrated that the rapid deglycosylation procedure developed for this kit produces complete deglycosylation of a diverse set of glycoproteins.⁸ This completeness of deglycosylation is also supported by intact mass analysis using LC-MS, where the deglycosylation of a monoclonal antibody (mAb) can be readily tracked. Figure 2 presents deconvoluted ESI mass spectra for Intact mAb Mass Check Standard, a murine IgG1 mAb. The top spectrum shows the mAb before it had been subjected to rapid deglycosylation (Figure 2A). The bottom spectrum meanwhile presents the mAb after it was processed according to the approach specified in the GlycoWorks RapiFluor-MS N-Glycan Kit, wherein glycoproteins are subjected to 1% (w/v) RapiGest SF Surfactant-assisted heat denaturation followed by incubation with Rapid PNGase F at 50 °C for 5 minutes (Figure 2B). The masses observed in these spectra confirm that these samples differ in terms of glycan occupancy. The control sample contains the mAb in its doubly glycosylated, native form (one glycan on each heavy chain). In contrast, the sample subjected to the proposed 2-step rapid deglycosylation procedure is homogenous with an observed molecular weight that is in agreement with the predicted molecular weight of the fully deglycosylated mAb (145.3 kDa). And although high temperatures are employed in this method for the purpose of heat denaturation, no detrimental effects on an N-glycan profile have been observed. To this point, notice that there are no differences in an N-glycan profile prepared from pooled human IgG when using an excessive 20 minute heat denaturation at 90 °C versus the rapid 3 minute procedure (Figure 3).

Robust sample preparation: Rapid labeling

The efficiency of the sample preparation carries over from deglycosylation to RapiFluor-MS labeling. A primary concern in this step is the relative stability of the PNGase F released N-glycosylamines, which are required for RapiFluor-MS labeling, in the pH 7.9 GlycoWorks Rapid Buffer. A time-course study involving varying delays between deglycosylation and RapiFluor-MS labeling steps has shown that the N-glycosylamines have a relatively long half-life of approximately 2 hours at 50 °C (Figure 4). That is, with our 5 minute deglycosylation step, there should be little concern over sample loss (< 3% loss) due to hydrolysis of the glycosylamine. In addition, sample losses from the labeling reaction are minimal. Many experimental parameters were explored during the development of the rapid labeling reaction specified in the RapiFluor-MS N-Glycan Kit, including pH, temperature, ionic strength, time, buffer components, and reagent molar excess. Figure 5 shows an example of optimizing the reagent molar excess as needed to maximize labeling yield. Fluorescence chromatograms for labeled, released N-glycans from Intact mAb Mass Check Standard are stacked on the left (Figure 5A). Note that with the GlycoWorks RapiFluor-MS N-Glycan Kit proteins are purposely not depleted from the sample after deglycosylation to save time and to give better control over the labeling. The RapiFluor-MS Reagent is therefore used in a molar excess over all of the nucleophiles from the glycoprotein, which for an IgG corresponds to approximately seventy five protein amines and two N-glycosylamines. Each of the corresponding samples was obtained from labeling a fixed glycoprotein concentration of 0.36 mg/mL with RapiFluor-MS Reagent at concentrations varying from 18 to 108 mM. As shown in Figure 5B, plotting of the fluorescence peak areas for the resulting N-glycan profile indicates that labeling is maximized near a RapiFluor-MS Reagent concentration of 36 mM, the conditions designed into the GlycoWorks RapiFluor-MS N-Glycan Kit. Moreover, molar excess conditions both higher and lower than the 36 mM reagent condition produced comparable fluorescence profiles, underscoring the robustness of RapiFluor-MS labeling.

Robust sample preparation: µElution HILIC SPE

The last step in the sample preparation involves extraction of the RapiFluor-MS labeled glycans from reaction byproducts using HILIC SPE. This technique has been routinely used for preparations of 2-AB labeled N-glycans and has now been optimized for RapiFluor-MS labeled species.^8-9 Previous studies have shown that RapiFluor-MS labeled glycans are obtained through this SPE processing at relatively high yields of approximately 74%.⁸ Nearly all of the observed sample losses in this step are non-specific. Figure 6 plots fluorescence peak areas for preparations of N-glycans from Intact mAb Mass Checked Standard, in which the final SPE elution volume was either 30, 90 or 180 µL. This plot shows that SPE recovery is a function of elution volume and that highest recoveries are achieved when employing large elution volumes. 

To facilitate direct analyses, however, a compromise is made such that a 90 µL elution volume is used in order to obtain a relatively concentrated glycan eluate. Regardless of the elution volume and absolute yield of glycans from the SPE sorbent, the most important characteristic of this clean-up is that the observed sample losses have been determined to be non-specific with no significant bias being introduced to a glycan profile for a wide range of glycans with diverse chemical properties, including small, neutral glycans up to large, tetrasialylated glycans (see Reference 8 for more details about GlycoWorks HILIC SPE). 

Yield of RapiFluor-MS labeled N-glycans

In another measurement of robustness, it is worth looking at the yield of N-glycans through the entire workflow. This was evaluated in order to measure the collective efficiency of combining fast  deglycosylation, rapid labeling, and HILIC SPE extraction of  RapiFluor-MS labeled glycans (Figure 7). RapiFluor-MS labeled N-glycans from Intact mAb Mass Check Standard were prepared, analyzed by HILIC-FLR, and quantified by means of an external calibration. Based on a calculated theoretical yield (see experimental) and duplicate analyses, it was determined that the percent yield through the entire RapiFluor-MS N-Glycan Kit sample preparation was approximately 73%. To provide perspective, we evaluated the yield of 2-AB labeled N-glycans from an alternative sample preparation workflow involving the use of a GlykoPrep Rapid N-Glycan Preparation with 2-AB Kit.

Quantitative analyses showed that 2-AB labeled N-glycans are prepared using this kit with a relatively low yield of approximately 35%, though it has been found that the yield of this kit can be dramatically improved by optimization and lengthening of the labeling step. Comparatively speaking, though, these results show that not only does the RapiFluor-MS approach quicken a historically time-consuming sample preparation, it also exhibits reasonably high yields.

Minimal impact to glycan profiling with reagent batch variation

Lastly, sample preparations with the GlycoWorks RapiFluor-MS N-Glycan Kit have proven to be robust with respect to reagent manufacturing. A robustness study was performed to test the impact of changing  the batches of each reagent that plays a critical role in the preparation of RapiFluor-MS labeled N-glycans, namely RapiGest SF, GlycoWorks Rapid Buffer, GlycoWorks Rapid PNGase F, RapiFluor-MS Reagent, DMF Reagent Solvent, GlycoWorks µElution SPE Plate, and the SPE Elution Buffer. Three sets of these reagents, each varying by batch, were tested in their application to profiling the N-glycans from Intact mAb Mass Check Standard. Average relative abundances observed for the glycan species in this standard with the three different reaction sets are presented in Figure 8. Relative abundances of N-glycans were observed to be largely comparable across the different preparations with an average RSD for the labeled N-glycan species being 2.3%.

Comparability to 2-AB N-glycan profiling

Another critical aspect to the RapiFluor-MS N-glycan sample preparation is that it can be used in place of legacy 2-AB methods without requiring significant adaptations to existing analytical techniques. With the speed of the sample preparation and the enhanced method sensitivity afforded by the RapiFluor-MS tag,8 the task of analyzing N-glycosylation is in fact made significantly easier.

Just like 2-AB labeled glycans, RapiFluor-MS labeled glycans are ideally suited for HILIC separations with an amide bonded stationary  phase, such as that found in the Waters Glycan BEH Amide Columns. Figure 9 shows example separations for 2-AB and RapiFluor-MS labeled glycans obtained from Intact mAb Mass Check Standard. The 2-AB labeled glycans, in this case, were prepared using the previously mentioned GlykoPrep Kit and an approximately 3.5 hour protocol, whereas the RapiFluor-MS labeled  glycans were prepared in less than 30 minutes using a GlycoWorks RapiFluor-MS N-Glycan Kit.  

So that chromatograms exhibiting equivalent signal-to-noise could be compared, the RapiFluor-MS sample was analyzed in this study at a significantly lower mass load than the 2-AB labeled sample. Despite being prepared by different approaches, it can be seen that the labeled N-glycans are resolved by the HILIC separation into very similar profiles. For a typical mAb profile, RapiFluor-MS and 2-AB labeling both yield HILIC glycan separations with similar selectivity. However, as a result of its additional hydrogen bonding donors/acceptors, the RapiFluor-MS label introduces a slight shift of the mAb N-glycan profile to higher retention times. This change in the absolute retention window of an N-glycan profile is predictable and can therefore be easily accounted for when transitioning from 2-AB to RapiFluor-MS based methods. 

Consistency in results observed for the RapiFluor-MS-based approach compared to historical 2-AB techniques was also evaluated. N-glycan profiling of the same monoclonal IgG1 reference sample has been studied between these different methodologies. Figure 10 displays N-glycan information obtained for this mAb sample throughout 160 different profiling experiments involving 2-AB labeling and HPLC chromatography. Likewise, Figure 10 provides data from 12 recent experiments using RapiFluor-MS labeling and UPLC chromatography. Comparable relative abundances are observed for this sample in a direct comparison (Figure 10, center panel) and a control chart demonstrates the ability to transition between these two methods (Figure 10, right panel). This consistency in N-glycan profiling makes it possible to replace time-consuming 2-AB/HPLC methods with RapiFluor-MS/UPLC techniques.  

Robustness of RapiFluor-MS N-glycan separations with glycan BEH amide columns

The robustness and resolving power of the HILIC column chromatography is critically important to successfully implementing this methodology. To this end, a test standard called RapiFluor-MS Glycan Performance Test Standard is available for method familiarization, system suitability, troubleshooting, and benchmarking studies. This standard contains a complex mixture of RapiFluor-MS N-glycans from human IgG that has been isolated from pooled human serum. Its composition of approximately 20 different major constituents makes it useful for evaluating the resolving power of a separation and the sensitivity of detection methods (Figure 11).

In line with its intended purpose, we have used the RapiFluor-MS Glycan Performance Test Standard to benchmark the chromatographic performance of four different columns containing amide bonded stationary phases designed for glycan separations. Two of the columns were UPLC-based and contained sub-2-µm particles while the remaining two were intended for use on HPLC instrumentation and contained 2.5 µm and 2.6 µm particles. Figure 12 shows representative chromatograms obtained with each of these columns run under equivalent conditions and linear velocities. Four glycan species spread across these separations were monitored to measure retention windows, average peak widths, and peak capacities. Notice that whether performing a separation with a phase intended for UPLC or HPLC chromatography, Glycan BEH Amide Columns provide exemplary resolving power and comparable selectivities thereby enabling the seamless transfer of this glycan separation between HPLC and UPLC platforms.¹⁰

Separations of RapiFluor-MS labeled glycans with glycan BEH amide columns have also proven to be very robust. In demonstration of this, a single Glycan BEH Amide, 130Å, 1.7 μm Column was subjected to lifetime testing and 300 sequential runs. At every 20th run, RapiFluor-MS Glycan Performance Test Standard was separated in order to track any changes in the retentivity and selectivity of the column. 

Chromatograms corresponding to the 1st and 300th runs are provided in Figures 13A and 13B, respectively. Quite clearly, near identical separations were obtained at the onset as well as at the end of this approximately 2-week constant use scenario, with no significant shifts in retention times of the labeled N-glycans having been observed throughout the testing (Figure 13C).

In this testing, LC calibrations were performed after every separation of the glycan mixture through application of a dextran ladder and assignment of glucose unit (GU) values. Separations with glycan BEH amide columns can be used in conjunction with glucose unit (GU) values as a means to calibrate HILIC-based glycan separations. Use of GU values minimizes subtle retention time variations between runs and between different instruments by expressing chromatographic retention in terms of  standardized GU values.¹¹ To assign GU values, a dextran ladder (comprised of glucose multimers of increasing length) is used as an external calibrant. The retention times of the glucose multimers are then used via cubic spline fitting to convert glycan retention times  into GU values.  

The development of a dextran calibration ladder suitable for use with RapiFluor-MS labeled glycans was essential yet technically challenging. Given that dextran is a reducing sugar without a strong nucleophile, it cannot, unlike N-glycosylamines, be readily labeled with RapiFluor-MS Reagent. Because of the distinctive urea linkage imparted to N-glycans upon their derivatization with rapid tagging reagents, RapiFluor-MS labeled N-glycans have very unique fluorescence maxima at approximately 265 nm (excitation) and 425 nm (emission) (Figure 14A). In a novel labeling approach, we have prepared a RapiFluor-MS Dextran Calibration Ladder by first reductively aminating dextran with ethanolamine and then labeling it with RapiFluor-MS. 

The resulting urea-linked dextran derivatives exhibit identical fluorescence properties to those of RapiFluor-MS labeled N-glycans. Furthermore, the obtained dextran is tuned for desired HILIC retention because of the hydroxyl group being incorporated through ethanolamine. A representative fluorescence chromatogram for this novel dextran ladder is provided in Figure 14B, and an example cubic spline fit of the retention data is shown in Figure 14C.

The impact of implementing GU value calibration is exemplified in Figure 13D, where the retention time data throughout the Glycan BEH Amide lifetime testing are reported in GU values. In comparing the retention time data shown in Figure 13C to the GU data in Figure 13D, one can see that the subtle fluctuations in retention times across the 2-week lifetime testing are compensated for by the GU calibration. In fact, RSDs for the GU value data are reduced by a factor of 2 compared to the RSDs in the retention time data. 

Also, linear regression analysis of the GU value data shows that there is essentially no drifting in the HILIC retention data once calibrated using a dextran ladder. This analysis therefore clearly demonstrates the value of GU calibration with respect to improving the quality of reported data.  

In this application note, we have demonstrated the robustness of RapiFluor-MS N-glycan preparations and Glycan BEH Amide HILIC Column chromatography. The RapiFluor-MS N-Glycan Kit enables analysts to perform a high yielding sample preparation with quantitative recovery that ensures accurate and repeatable profiling of N-glycans that is highly comparable to HPLC, 2-AB based methodologies. Moreover, it has been demonstrated that Glycan BEH Amide Columns afford exemplary resolving power and ruggedness for separations of RapiFluor-MS labeled N-glycans. Additionally, this separation can be readily transferred between UPLC and HPLC platforms. To further ensure success with these new methodologies, two standards have been commercialized, and their use to facilitate RapiFluor-MS analyses has been demonstrated. The RapiFluor-MS Glycan Performance Test Standard has been used for benchmarking studies, while the novel RapiFluor-MS Dextran Calibration Ladder has been employed to enhance the reproducibility of chromatographic retention time data. In summary, the GlycoWorks RapiFluor-MS N-Glycan Kit and supporting standards and columns can significantly reduce the burdens associated with N-glycan profiling while providing accurate, reproducible, and sensitive analyses.

Acknowledgement

We would like to thank Peter De Vreugd and Mark Eggink from Synthon Biopharmaceuticals BV for providing data on the comparability of released N-glycan analyses with RapiFluor-MS and UPLC versus 2-AB and HPLC.

1. Ohtsubo, K.; Marth, J. D., Glycosylation in cellular mechanisms of health and disease. Cell 2006, 126 (5), 855–67.
2. Mechref, Y.; Hu, Y.; Garcia, A.; Zhou, S.; Desantos-Garcia, J. L.; Hussein, A., Defining putative glycan cancer biomarkers by MS. Bioanalysis 2012, 4 (20), 2457–69.
3. Ruhaak, L. R.; Miyamoto, S.; Lebrilla, C. B., Developments in the identification of glycan biomarkers for the detection of cancer. Mol Cell Proteomics 2013, 12 (4), 846–55.
4. Dalziel, M.; Crispin, M.; Scanlan, C. N.; Zitzmann, N.; Dwek, R. A., Emerging principles for the therapeutic exploitation of glycosylation. Science 2014, 343 (6166), 1235681.
5. Beck, A.; Wagner-Rousset, E.; Ayoub, D.; Van Dorsselaer, A.; Sanglier-Cianferani, S., Characterization of therapeutic antibodies and related products. Anal Chem 2013, 85 (2), 715–36.
6. Mechref, Y.; Hu, Y.; Desantos-Garcia, J. L.; Hussein, A.; Tang, H., Quantitative glycomics strategies. Mol Cell Proteomics 2013, 12 (4), 874–84.
7. Ruhaak, L. R.; Zauner, G.; Huhn, C.; Bruggink, C.; Deelder, A. M.; Wuhrer, M., Glycan labeling strategies and their use in identification and quantification. Anal Bioanal Chem 2010, 397 (8), 3457–81.
8. Lauber, M. A.; Brousmiche, D. W.; Hua, Z.; Koza, S. M.; Guthrie, E.; Magnelli, P.; Taron, C. H.; Fountain, K. J., Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent. Waters Application Note 720005275EN 2015.
9. Yu, Y. Q.; Gilar, M.; Kaska, J.; Gebler, J. C., A rapid sample preparation method for mass spectrometric characterization of N-linked glycans. Rapid Commun Mass Spectrom 2005, 19 (16), 2331–6.
10. Koza, S. M.; Lauber, M. A.; Fountain, K. J., Successful Transfer of Size-Exclusion Separations between HPLC and UPLC. Waters Application Note 720005214EN 2015.
11. Campbell, M. P.; Royle, L.; Radcliffe, C. M.; Dwek, R. A.; Rudd, P. M., GlycoBase and autoGU: tools for HPLC-based glycan analysis. Bioinformatics 2008, 24 (9), 1214–6.