Background

Current mass spectrometry (MS)-based glycomic mapping remains insufficient to delineate the full complexity of the glycome. Although MALDI-MS mapping and MS/MS sequencing would efficiently afford a very useful first impression glycomic profile, it is nearly impossible to acquire MALDI-MS/MS on every putative glycan signal detected, especially those of low intensities and/or occurring at higher masses. In that respect, LC-ESI-MS/MS analysis in an automated and data-dependent acquisition (DDA) mode provides a far more comprehensive MS/MS data coverage (Cheng et al., 2013; Cheng et al., 2015; Hsiao et al., 2017; Yu et al., 2018). We have demonstrated that permethylation in conjunction with 2 steps C18/amine SPE or a single step mixed mode MAX SPE fractionation can yield fully methylated non-sulfated, mono-sulfated, and di-sulfated glycans in separate pools (Yu et al., 2020) for MALDI-MS screening. We found that loss of sodium sulfite from permethylated di-sulfated glycans occurred readily during MALDI-ionization (Figure 1A), while the two sulfates on permethylated glycans were fully retained when analyzed by nanoESI-MS in negative ion mode, allowing them to be detected as [M-2H]^2- (Figure 1B), and further selected for MS/MS analyses at high sensitivity. Moreover, we showed that the negative mode nanoLC-MS/MS analysis of permethylated sulfated glycans on a C18 reverse phase capillary column could be efficiently carried out using the common acetonitrile/formic acid/water solvent system, and hence be fully compatible with the normal set-up of an analytical laboratory devoted to proteomics.

By virtue of a panel of synthetic sulfated glycans, we have shown that the beam-type higher energy collision dissociation (HCD) MS/MS as implemented on the hybrid Orbitrap series would afford a range of low mass fragment ions. These diagnostic ions would define the location of sulfate on which glycosyl residue at which position (Cheng et al., 2015). Ion trap-based CID MS², on the other hand, would not retain these very useful ions due to its one third low mass cut-off (Patnode et al., 2013). It does, however, offer a higher sensitivity and acquisition rate, allowing definitive assignment of target sulfated glycotopes via MS² product ion-dependent MS³. Occasionally, when a di-sulfated O-glycan may carry a different permutation of sulfated glycotopes on either the 6-arm or the extended 3-arm, additional targeted MSⁿ analyses can help identify the existence of isomers (Yu et al., 2018). Moving from the original linear ion trap-Orbitrap hybrid systems to more recent tribrid Orbitrap Fusion systems, the current MS systems afford a higher degree of flexibility for different combination of single to multiple stages of HCD versus ion trap-based CID MS/MS to be acquired either in Orbitrap for greater resolution and mass accuracy, or the ion trap for better sensitivity and speed. These aspects will not be further dealt with here. Suffice to point out that the basic principle remains the same, namely to acquire as many HCD/CID MS² within an elution time window-compatible DDA duty cycle and to couple each, if possible and desirable, to a pre-determined list of product-dependent MS³ for greater depth of structural details. With or without MS³, it is important that the low mass ions produced in negative mode MS² be retained and detected at reasonably high resolution and mass accuracy (< 5 ppm, if possible).

Online LC and data-dependent MS/MS acquisition will produce a huge dataset which is near impossible to manually analyze systematically. Unlike proteomics, neither the glycan database nor the MS/MS sequencing algorithm is well developed to allow direct MS² ions or spectral matching search for unambiguous identification. An in-house computational tool, GlyPick, was developed to filter out bona fide glycan MS² spectra by user-defined criteria, usually by presence of at least 2 to 3 diagnostic glycan fragment ions. It can also extract out and compute the occurrence and summed intensity of a list of user-input diagnostic MS² ions that will define the presence and relative abundance of those important glycotopes such as 6-sulfated GlcNAc, 3’-sulfated Gal, 6’-sulfated Gal, sulfated LacNAc, sulfated Le, etc, from these MS² spectra (Hsiao et al., 2017; Yu et al., 2018). Results are output in CSV format, which can be conveniently interrogated using Excel for further data mining and graph plotting.


Figure 1. MS spectra of di-sulfated permethylated N-glycans acquired by different ionizations, which are (A) MALDI-MS and (B) nanospray ESI-MS, both in the negative ion mode.