No abstract

Consistent with the goals of a comprehensive carbohydrate sequencing strategy, we extend earlier reports to include the characterization of structural (constitutional) isomers. Protocols were developed around ion trap instrumentation providing sequential mass spectrometry (MS n ) and supported with automation and related computational tools. These strategies have been built on the principle that for a single structure all product spectra upon sequential fragmentation are reproducible and each stage represents a rational spectrum of its precursor; i.e., all major fragments should be accounted for. Anomalous ions at any stage are clues indicating the presence of structural isomers. Gas-phase isolation and subsequent fragmentation of such ions provide an opportunity to specifically resolve selected structures for their detailed characterization. Importantly, some isomers were not detected following MS 2 and required multiple (MS n>2 ) stages for their characterization. Derivatization remains critical to position substructures in a glycan array since product ions carry fragmentation "scars" throughout the MS n tree. Equally as important are the pathway relationships between each stage and the greater yield of fragments with the smaller number of oscillators. Applications were directed to the structural isomers in ovalbumin and IgG, where, in the latter case, several previously unreported glycans were detected. Procedures were supported with bioinformatics tools for assimilating structure from the MS n data sets.A comprehensive carbohydrate sequence is particularly challenging when considering the constitutional and stereoisomers that are abundant in glycan arrays. Constitutional isomers, also referred to as structural isomers, are defined as an identical atomic composition arranged in a different structure. Stereoisomers are exemplified with those frequently encountered in the hexoses, mannose, galactose, and glucose, while the two G1 glycans found in IgG are representative of simple structural isomers. In this latter case, an identical monomer may be linked on either of two antennae providing a different topology. In another nomenclature clarification, the term isomer is often confused with isobar. which relates to a different composition of atoms occurring at a nominal (unit) mass resolution. 1 At the MS profile stage, isobars are not usually a significant problem in carbohydrate structural characterization, but structural isomers and stereoisomers are. During fragmentation, however, isobaric fragments may be formed, but identification will require higher mass resolving power and mass accuracy.

This is the third in a sequence of reports devoted to the development of congruent strategies for carbohydrate sequencing. Two previous reports outlined the strategies for observing structural detail from MS n data and introduced tools that compile, search, and compare fragment spectra in a bottomup approach to oligosaccharide sequencing. In this third report, we introduce the operational details of an algorithm that we define as the Oligosaccharide Subtree Constraint Algorithm (OSCAR). This algorithm assimilates analyst-selected MS n ion fragmentation pathways into oligosaccharide topology (branching and linkage) using what may be considered a top-down sequencing strategy. Guided by a series of logical constraints, this de novo algorithm provides molecular topology without presumed biosynthetic constraints or external comparisons. In this introductory study, OSCAR is applied to a series of permethylated oligomers and isomeric glycans, and topologies are assigned in a few hundredths of a second.The two preceding reports in this series introduced methodologies for determining the structural details of carbohydrates from MS n data 1 and described tools that extract further details through spectral comparisons against known oligosaccharide fragments. 2 Glycobiology in general has been a fertile area for the development of bioinformatics tools, 3 and a variety of computer programs are now available to support carbohydrate analysis using mass spectrometry data. The web-based tool GlycoMod 4 accepts a glycan MS mass and returns a list of possible compositions, using literature-derived constraints to limit its output. In contrast, StrOligo 5 , 6 examines oligosaccharide MS 2 spectra to propose a set of candidate structures restricted by biosynthetic constraints.The candidates are fragmented in silico, and the resulting simulated spectra are ranked against the experimental MS 2 spectrum. Similarly, GlycosidIQ 7 compares an experimental MS 2 spectrum against simulated spectra generated from the contents of GlycoSuiteDB, 8 a curated database of known structures, to produce a ranked list of candidate glycans. In the catalog library method, 9 , 10 a catalog contains the characteristic fragmentation patterns of substructures isolated from a library of known oligosaccharides. Total structure assignment is accomplished by matching observed fragmentation patterns with the catalog motif entries. NIH Public AccessThe web-based Saccharide Topology Analysis Tool (STAT) accepts a native glycan mass as input, infers the glycan's possible compositions over a wide range of monosaccharide residues, and generates a candidate set of all possible branching topologies. 11 , 12 STAT then accepts a list of ion masses extracted from the MS n data tree, where each ion must form a connected substructure within the candidate topologies. Finally, a ranked list of branching topologies is generated. STAT supports native glycans, requires manual intervention to resolve ambiguous ion compositions, and restricts the branch points for N-linked ...

No abstract