Comparison of denaturing agent effects in enzymatic N-glycan release for human plasma N-glycan analysis

Glycosylation is an essential posttranslational modification observed in the living proteome. Glycosylation profiles in glycoproteins can change in commonly observed diseases such as cancer. Identifying these changes is crucial in discovering new biomarkers for the early diagnosis of cancer. One of the main steps of N-glycan analysis is to release N-glycans from glycoproteins by specific enzymes. The study compares common denaturing agent combinations used in N-glycan release methods. In the study, human plasma was used to test the release methods of N-glycans containing different detergent combinations. The released N-glycans were labeled with the procainamide tag, purified using cellulose-containing solid-phase extraction cartridges, and analyzed by high-performance liquid chromatography-hydrophilic interaction liquid chromatography equipped with fluorescence detection (HPLC-HILIC-FLD). The results showed that the sodium dodecyl sulfate and sodium deoxycholate (SDS + SDC) detergent combination provided the highest average FLD signal areas and intensities in the N-glycan analysis. The protocol with SDS resulted in more reproducible average FLD signal areas and intensities. It was also found that the average signal FLD signal areas and intensities of the detected N-glycans were reduced when SDS and SDC were used with 1,4-dithiothreitol (DTT) reducing agents. In addition, deglycosylation of glycoproteins with various denaturing agents resulted in relatively minor variation in human plasma N-glycosylation profiles.


Results and discussion
In this study, the human plasma N-glycans were released using the methods including various detergent combinations, labeled by the procainamide tag, and purified via cellulose-containing solid-phase extraction cartridges. Then, N-glycan analysis was performed by an HPLC-HILIC-FLD. The detected N-glycan peak areas and intensities were processed by software and subsequently evaluated. The experiments were achieved on two different days with three replicates (n = 6). In the study, we compared the six detergent combinations commonly used in the N-glycan release methods as shown in Table. Since most manufacturers of PNGase F offer SDS and DTT as denaturing agents, and the use of SDC in glycoproteomics applications has increased, these denaturing agents and their combinations were selected in the study for comparison. Figure 1 presents an example FLD chromatogram obtained from the analysis of the procainamide labeled N-glycans by the HPLC-HILIC-FLD. We detected 22 N-glycan peaks belonging to human plasma N-glycome in this study. The peaks were annotated based on the literature knowledge and our previous work (18,19). Then, the obtained data were compared among N-glycan release methods with detergent combinations.

Comparison of detergent combinations in N-glycan release methods based on peak areas and peak intensities
The detected peaks belonging to N-glycans of human plasma were found in each FLD chromatogram obtained from N-glycan release methods containing different detergent combinations ( Figure 2). The resolution and shape of the N-glycan peaks were not influenced by the N-glycan release methods with different detergent combinations.
The N-glycan release methods with different detergent combinations were first compared by evaluating the peak areas of N-glycans extracted from the FLD chromatograms. Figure 3a shows the comparison of data based on total peak areas of the detected N-glycan peaks. The highest peak areas were calculated for each detergent containing N-glycan release methods in the FLD chromatogram. The SDS + SDC detergent combination used in the N-glycan release methods was found to have the highest total peak area. The N-glycan release approach that used SDS was found to have the secondhighest area. The N-glycan release methods with different detergents were also compared by evaluating the peak intensity of N-glycans extracted from the FLD chromatogram. Figure 3b displays the comparison of the data based on peak intensities of the detected N-glycan peaks. As expected, the SDS + SDC detergent combination was the highest total peak intensities in the FLD chromatogram when the obtained intensity values were compared in the N-glycan release methods. The SDS detergent containing the N-glycan release method was found to have the second-highest intensity. The data was ordered based on total areas and intensities as follows: SDS + SDC > SDS > SDC > SDS + SDC + 0.5 M DTT > SDS + DTT > SDC + 0.5 M DTT. On the other hand, the average areas and intensities obtained from each method were evaluated to test the efficiency of N-glycan analysis. The results obtained from the total area and intensity values of the N-glycans were matched with the results from average areas and intensities (Figures 3c and 3d). The N-glycan release method with SDS detergent provided more reproducibility based on total and average peak areas and intensities.

Comparison of detergent combinations in N-glycan release methods based on relative abundances
The total area normalization approach was used for the calculation of the relative abundances of the N-glycans. The calculated relative abundances of each N-glycan based on peak areas and intensities are shown in Figures 4a and 4b, respectively. The relative abundances of the detected peaks are also shown in Tables S1 and S2 based on peak areas and peak intensities, respectively. The data resulted in relatively minor changes in relative abundances of human plasma N-glycans among N-glycan release methods, including different detergent combinations. However, the N-glycan G14 (H5N4S2, diantennary sialylated type) abundantly found in human plasma differed among protocols. It had a higher abundance in DTT-containing protocols (Tables S1 and S2). In addition, the relative abundances of tri-antennary sialylated species were determined higher in DTT reducing agent-containing methods than other methods worked in this study. For human plasma glycoproteins, the protein structures in their native form may form steric hindrance, which restricts the access of any PNGase F enzyme to defined glycosylation sites. Methods involving reducing agents can remove the steric hindrance found in proteins and allow the PNGase F enzyme to more efficiently reach the glycosylation sites. The area-based CV% (coefficient of variation) of the peak abundances was also investigated for N-glycan release methods with different detergent combinations ( Figure S1). It was found that SDS + DTT and SDC + DTT containing N-glycan release methods had higher mean CV% (it was 32% and 63% for SDS + DTT and SDC + DTT, respectively). However, the mean CV% values of plasma N-glycan abundances for SDS, SDC, SDS + SDC and, SDS + SDC + DTT detergent containing N-glycan release method were detected to be relatively low. The average CV% of these N-glycan release methods was 26%, 18%, 24%, and 20%, respectively.
The main limitation of this study was to compare three different detergents and their combinations, whereas many others are available in N-glycan release methods. Many manufacturers producing the PNGase F enzyme offer SDS-based N-glycan release methods. These methods were applied with DTT reducing agents. Therefore, the most widely used chemicals for that purpose were evaluated in the study. On the other hand, SDC has been recently evaluated in releasing N-glycans from glycoproteins extracted from biological samples [20]. They have determined that SDC assisted approach was found to be more efficient compared with filter-aided sample preparation. On the other hand, the purification of procainamide labeled N-glycans may also affect the data obtained from N-glycan analysis. In the literature, many purification methods, including different interactions between N-glycans and adsorbents, have been introduced [21, 22]. For HILIC-FLD-based N-glycosylation analysis, hydrophilic-interaction-based sorbents have commonly been employed for the purifications [23]. It has been detected that HILIC-based sorbents showed good reproducibility [24]. Therefore, we used the self-packed cellulose solid-phase extraction cartridges for that purpose in this study. In our study, the N-glycan profiles of human plasma were found to have minor changes in different detergent combinations containing N-glycan release methods. It could be concluded that the denaturation agents influence the efficiency of N-glycan release. In addition, the removal of N-glycan types was found to differ based on N-glycan release methods with different detergent combinations. A recent study evaluated the PNGase F enzymes produced by three manufacturers regarding the N-glycan profiles. They have found that deglycosylation with PNGase F enzymes manufactured by different companies resulted in different IgG and plasma N-glycosylation HILIC-FLD profiles [12]. These results indicated that the applied N-glycan release method, including denaturing agents and PNGase F enzymes, provided different N-glycosylation HILIC-FLD profiles.