Targeted glycomics by selected reaction monitoring for highly sensitive glycan compositional analysis

Abstract

The development of glycomics increasingly requires the detection and quantification of large numbers of glycans, which is only partially achieved by current glycomics approaches. Taking advantage of selected reaction monitoring to enhance both sensitivity and selectivity, we report here a strategy termed targeted glycomics that enables highly sensitive and consistent identification and quantification of diverse glycans across multiple samples at the same time. In this proof-of-principle study, we validated the method by analyzing global N-glycans expressed in different systems: single proteins, cancer cells, and serum samples. A dynamic range of three orders of magnitude was obtained for the detection of all five glycans released from ribonuclease B. The limit of detection of 80 attomole for Man(9)GlcNAc(2) demonstrated the excellent sensitivity of the method. The capability of the strategy to identify diverse glycans was demonstrated by identification and detection of 162 different glycans and isomers from pancreatic cancer cells. The sensitivity of the method was illustrated further by the ability to detect eight glycans from 250 cancer cells and five glycans released from 100 cancer cells. In serum obtained from rabbits fed control diet or diet enriched with 2% cholesterol, differences to 42 glycans were accurately measured and this indicates that this strategy might find use in studies of biomarker discovery and validation.

Figure 1

Outline of targeted glycomics by LC-SRM workflow. Targeted glycomics is composed of six stages, including preparation of glycans, derivatization of glycans, selection of a glycan set, determination of SRM transitions, identification of glycans by LC-SRM, further validation of glycan identified, and quantification of glycans.

Figure 2

The MS/MS spectra of Man₅GlcNAc₂ labeled with aniline (a) or aniline-d₅ (b). The total N-glycans released from 50 μg RNase B were split into two aliquots for aniline or aniline-d₅ derivatization.

Figure 3

The extracted ion chromatograms of LC-SRM showing separation and detection of high mannose glycans (Man_5-9GlcNAc₂) released from 50 μg RNase B.

Figure 4

The extracted ion chromatogram demonstrating the limit of detection of 80 attomole on column for targeted glycomics analysis of Man₉GlcNAc₂. The concentration of Man₉GlcNAc₂ was 1.6 ×10⁻¹¹ M, and injection volume was 5 μL.

Figure 5

The extracted ion chromatograms of the glycans from 250 cancer cells (a) and 100 cancer cells (b). Peak 1: Hex₈HexNAc₂; Peak 2: Hex₉HexNAc₂, Peak 3: Hex₇HexNAc₂; Peak 4: Hex₆HexNAc₂; Peak 5: Hex₅dHex₁HexNAc₄; Peak 6: Hex₅HexNAc₂; Peak 7: Hex₅NeuAc₁HexNAc₄; Peak 8: Hex₅dHex₁NeuAc₁HexNAc₄.

Figure 6

Plot of relative amounts of 42 glycans between the replicate analyses of the same sample. The replicate analyses were conducted by using four aliquots of 10 μL serum sample. Four aliquots were divided into two pairs and each pair was subjected to enzymatic release of N-glycans, derivatization of N-glycans with aniline or aniline-d₅, and LC-SRM analysis. The relative amounts of 42 glycans from one analysis were plotted versus corresponding relative amounts of those glycans from the other analysis.

Figure 7

Bar graphs of the ion intensity ratios of 42 glycan compositions to their corresponding isotope internal standards. Rabbits (n=5) in control group were fed with a normal diet, and rabbits (n=5) in treated group were fed with a diet supplemented with 2% cholesterol. T-1 to T-5 and C-1 to C-5 correspond to sera collected every three weeks up to 12 weeks from the same rabbit. The glycan compositions were summarized in Table 2. Triplicate analyses were conducted for each serum sample and standard deviation was indicated by the error bar. The glycans with changes in abundance upon the treatment of 2% cholesterol diet were highlighted by circles, including Hex₃HexNAc₄, Hex₃dHex₁HexNAc_4, Hex₃dHex₁HexNAc₅ and Hex₄dHex₁HexNAc_5.