Nicholas J. DeBono

N -glycolylneuraminic acid (Neu5Gc) has been repeatedly associated with human cancer, but reliable detection has remained elusive, generating controversy regarding its presence in human samples. To address this, matrix-matched calibration curves, which have been pioneered in proteomics and metabolomics for assessing changes in complex mixtures, were measured of released N -glycans at four orders of magnitude dynamic range in defined mixtures, systematically benchmarking Neu5Gc-containing N -glycan detection across multiple LC-MS platforms and sites. Orthogonally, the gold-standard analytical method, consisting of fluorescence detection of labelled monosaccharides separated by LC, was applied to the same samples, yielding absolute concentrations of Neu5Gc. LC-MS demonstrated an extended detection range of three or more orders of magnitude while retaining intact N -glycan measurement, improving assay specificity and enabling detection of the variety of Neu5Gc-bearing N -glycans. By combining orthogonal dimensions of evidence, including chromatographic separation, isotopic distribution matching, and composition-confirming MS/MS, LC-MS confidently resolved Neu5Gc signals from noise, even at low abundance. In comparison, DMB-LC-FLR was limited to two orders of magnitude dynamic range, insufficient for detection of Neu5Gc in commercially available pooled human sera. These findings strongly support that DMB-LC-FLR assay specificity and sensitivity are insufficient for Neu5Gc detection in human samples due to noise overwhelming the Neu5Gc signal. By establishing a reusable benchmarking framework for future glycomic studies, we aim to use LC-MS to improve the measurement of Neu5Gc in clinical samples.

Controlling protein glycosylation as a critical quality attribute of biopharmaceuticals remains challenging when glycosylation is coupled to cellular production systems. Here, we present a proof-of-concept glycosyltransferase immobilised enzyme reactor (IMER) housed within a 3D-printed column that enables directed post-production glycan modification of purified glycoproteins. Using β-1,4-galactosyltransferase (β4GalT1-IMER) and α-2,6-sialyltransferase (ST6Gal1-IMER) immobilised on Ni-NTA resin, the IMER achieved near-complete galactosylation and substantial sialylation of partially deglycosylated bovine fetuin N-glycans with their respective substrates with a maximum substrate-enzyme contact time of four minutes. Isomeric-level analysis revealed arm-specific addition preferences for both enzymes, consistent with known specificities. The modular IMER design permits sequential connection of individual enzyme chambers, potentially offering a scalable, plug-and-play platform for constructing defined glycan structures on recombinant glycoprotein therapeutics.

: Human soluble CD52 is a short glycopeptide comprising 12 amino acids (GQNDTSQTSSPS) which functions as an immune regulator by sequestering the pro-inflammatory high mobility group box protein 1 (HMGB1) and supressing immune responses. Recombinant CD52 has been shown to act as a broad anti-inflammatory agent, dampening both adaptive and innate immune responses. This short glycopeptide is heavily glycosylated, with a complex sialylated N-linked glycan at N3 and reported O- linked glycosylation possible on several serine and threonine residues. Previously we demonstrated that specific glycosylation features of CD52 are essential for its immunosuppressive function, with terminal α-2,3-linked sialic acids required for binding to the inhibitory SIGLEC-10 receptor leading to T-cell suppression. Using high resolution mass spectrometry, we have further characterised the N- and O-linked glycosylation of Expi293 recombinantly produced CD52 at a glycopeptide and released glycan level, accurately determining glycan heterogeneity of both N- and O-linked glycosylation, and localising the site of O-glycosylation to T8 with high confidence and direct spectral evidence. This detailed knowledge of CD52 glycosylation informed the construction of a model system, which we analysed by molecular dynamics simulations to understand the mechanism of recognition and define interactions between bioactive CD52, HMGB1 and the SIGLEC-10 receptor. Our results confirm the essential role of glycosylation, more specifically hyper-sialylation, in the function of CD52, and identify at the atomistic level specific interactions between CD52 glycans and the Box B domain of HMGB1 that determine recognition, and the stability of the CD52/HMGB1 complex. These insights will inform the development of synthetic CD52 as an immunotherapeutic agent.

The analysis of the structures of glycans present on glycoproteins is an essential component for determining glycoprotein function; however, detailed glycan structural assignment on glycopeptides from proteomics mass spectrometric data remains challenging. Glycoproteomic analysis by mass spectrometry currently can provide significant, yet incomplete, information about the glycans present, including the glycan monosaccharide composition and in some circumstances the site(s) of glycosylation. Advancements in mass spectrometric resolution, using high-mass accuracy instrumentation and tailored MS/MS fragmentation parameters, coupled with a dedicated definition of diagnostic fragmentation ions have enabled the determination of some glycan structural features, or glycotopes, expressed on glycopeptides. Here we present a collation of diagnostic glycan fragments produced by traditional positive-ion-mode reversed-phase LC-ESI MS/MS proteomic workflows and describe the specific fragmentation energy settings required to identify specific glycotopes presented on N- or O-linked glycopeptides in a typical proteomics MS/MS experiment.

The mammalian glycome is structurally complex and diverse, composed of many glycan classes such as N- and O-linked glycans, glycosaminoglycans (GAGs), glycosphingolipids (GSLs), and other distinct glycan features such as polysialic acids (PolySia), sulfation, and proteoglycan attachment stubs. Various methods are used to analyze these different components of the glycome, but they require prefractionated/partitioned samples to target each glycan class individually. To address this need for a knowledge of the relationship between the different glycan components of a biological system, we developed a sequential release workflow for analysis of multiple conjugated glycan classes (PolySia, GAGs, GSL glycans, N-glycans, and O-glycans) from the same tissue lysate, termed SSSMuG─Same Sample Sequential Multi-Glycomics. With this sequential glycan release approach, five glycan classes were characterized (or four glycan classes plus proteomics) using enzymatic or chemical release from a single sample immobilized on a polyvinylidene difluoride membrane. The various released glycan classes were then analyzed by HPLC and MS techniques using commonly available analytical setups. Compared to single glycan class release approaches, SSSMuG was able to identify more glycans and more proteins with higher-intensity analytical peaks and provide a better comparative normalization of the different glycan classes of the complex glycome. To this end, the SSSMuG technology workflow will be a foundation for a paradigm shift in the field, transforming glycoanalytics and facilitating the push toward multiglycomics and systems glycobiology.

The structures and functions of organelles in cells depend on each other but have not been systematically explored. We established stable knockout cell lines of peroxisomal, Golgi and endoplasmic reticulum genes identified in a whole-genome CRISPR knockout screen for inducers of mitochondrial biogenesis stress, showing that defects in peroxisome, Golgi and endoplasmic reticulum metabolism disrupt mitochondrial structure and function. Our quantitative total-organelle profiling approach for focussed ion beam scanning electron microscopy revealed in unprecedented detail that specific organelle dysfunctions precipitate multi-organelle biogenesis defects, impair mitochondrial morphology and reduce respiration. Multi-omics profiling showed a unified proteome response and global shifts in lipid and glycoprotein homeostasis that are elicited when organelle biogenesis is compromised, and that the resulting mitochondrial dysfunction can be rescued with precursors for ether-glycerophospholipid metabolic pathways. This work defines metabolic and morphological interactions between organelles and how their perturbation can cause disease.

The automated design of synthetic biological circuits is an active area of research. A particularly promising area of research is the engineering of populations of communicating bacteria, in order to produce behaviour more complex than is possible with the engineering of individual bacteria. We present a computational approach to the engineering of communicating bacterial populations, using a multi-level approach. Circuits are designed using an evolutionary algorithm, at a high level of abstraction, with an agent-based model. Evolved agents can then be mapped onto previously-defined, lower-level components such as Standard Virtual Parts. This approach is applied to the evolution of a two-dimensional pattern, the French Flag.