Send to printer »

Tutorials : May 15, 2010 (Vol. 30, No. 10)

N-Linked Oligosaccharide Separation

New Method Developed to Improve the Resolution of Neutral and Sialylated Glycans
  • Ting Zheng, Ph.D.
  • ,
  • Jeffrey Rohrer, Ph.D.
  • ,
  • Srinivasa Rao, Ph.D.

The analysis of glycosylation patterns on glycoprotein therapeutics such as antibodies is important as glycans are critical to glycoprotein function and stability. Among the many techniques that have been used for glycan analysis, high-performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD) is significant as it provides high-resolution separation and high-sensitivity detection for oligosaccharides without requiring sample labeling.

HPAE-PAD separations can be performed on Dionex CarboPac® columns, a family of anion-exchange columns developed for carbohydrate analysis. The newest member of the CarboPac column family—the CarboPac PA200 column—is designed for high-resolution oligosaccharide separations.

The CarboPac PA200 typically uses a NaOAc gradient of up to 250 mM in 100 mM NaOH. Some glycobiologists have observed that this gradient is not always effective for separating neutral N-linked oligosaccharides. 

We have found that the eluting NaOH concentration has significant impact on neutral oligosaccharide separations. Decreasing the NaOH concentration from 100 to 50 mM greatly improved resolution.

Six neutral and two sialylated oligosaccharide standards were used in this study (Figure 1). These oligosaccharide structures are commonly found on monoclonal and polyclonal antibodies. All separations were performed on a Dionex ICS-3000 system, which includes an ICS-3000 DP gradient pump, an AS autosampler, and an ICS-3000 DC column compartment with an electrochemical detector cell.

The carbohydrates were detected by pulsed amperometry with a gold working electrode and an Ag/AgCl reference electrode using the standard 4-potential waveform developed at Dionex (t1=0.0s, E1= +0.1V; t2=0.2s, E2= +0.1V; t3=0.4s, E3= +0.1V; t4=0.41s, E4= -2.0V; t5=0.42s, E5= -2.0V; t6=0.43s, E6= +0.6V; t7=0.44s, E7= -0.1V; t8=0.5s, E8= -0.1V). Chromatography was controlled by Chromeleon® Chromatography Data System software.

The effect of NaOH concentration on oligosaccharide separation using the CarboPac PA200 column is shown in Figure 2A. With the typically used 100 mM NaOH, two pairs of oligosaccharides co-eluted and another pair was only partially resolved.

When 50 mM NaOH was used, the resolution was greatly improved and all the six neutral oligosaccharides were well resolved. The sodium acetate gradient was 0–5 mM NaOAc in 40 min for both. Adjusting the sodium acetate gradient did not significantly improve the resolution. For this set of neutral oligosaccharides, NaOH concentrations higher than 100 mM resulted in faster elution and more severe co-elution. Decreasing the NaOH concentration further to 30 mM again resulted in co-elution of some peaks.

Separation of Linkage Isomers

The oligosaccharide standard G1 showed two peaks, presumably the two positional isomers (Figure 1). To confirm the identities of these peaks, the G1 standard was digested with β 1,4-galactosidase and then analyzed to compare with G0 and G1.

Results showed that, after digestion, only a minor peak appeared at the position of one of the two original G1 peaks, and a major peak eluted at the same time as the G0 standard (Figure 2B). These results show that the two original G1 peaks had terminal galactose groups and after the galactose cleavage both were converted to G0, which strongly suggests that the two original G1 peaks are the positional isomers shown in Figure 1.

When 50 mM NaOH was used, the oxygen dip (~33 min in Chromatogram 4 in Figure 2A) interfered with the peak detection. The dip in the baseline is due to the dissolved oxygen introduced by sample injection being greater than the eluent’s oxygen concentration. The intensity of the dip is usually 1–2 nC, and can be reduced by setting a 25 µL cut volume in the sample injection command, which can reduce the amount of air injected.

To move the dip away from the sample peaks, flow rate and sodium acetate gradient were adjusted. The oxygen dip was eluted later than the sample peaks when the flow rate was decreased from 0.5 mL/min to 0.35 mL/min, with an acetate gradient of 1–6 mM in 45 min. The good resolution of the oligosaccharide standards was maintained with this gradient.

The optimized gradient can be extended to include elution with higher concentration of NaOAc, allowing separation of both neutral and sialylated oligosaccharides in the same injection.

The extended gradient was used to profile the N-linked glycans released from polyclonal human IgG and a monoclonal antibody (Figure 2C). The proteins were digested with PNGase F. After 24 hours of  incubation at 37°C, the digestion mix was directly injected onto the CarboPac PA200 column and analyzed with the extended optimal gradient.

Salts in the digestion buffer could affect the binding of the samples to the column and cause them to elute slightly earlier. Therefore, standards were dissolved in the same digestion buffer and analyzed under the same conditions. Profiling results showed that the polyclonal antibody had more neutral-type oligosaccharides than the monoclonal antibody.

The monoclonal antibody had fewer mono-sialylated glycans, while the polyclonal human IgG had more mono-sialylated glycans than di-sialylated glycans. Comparison with the standard profile showed that the major neutral N-glycans of polyclonal human IgG were G0, G1, and G2. The two peaks labeled with an asterisk eluted later than G1 and G2 and could be G1-F and G2-F, as fucosylated oligosaccharides usually elute earlier than the same structure without the core fucose. The neutral N-glycans of the monoclonal antibody were mostly G0 and G1. In conclusion, an efficient method was developed for high-resolution separation of both neutral and sialylated N-linked oligosaccharides on the CarboPac PA200 column.