The O-Glycome of Human Nigrostriatal Tissue and Its Alteration in Parkinson’s Disease

O-Glycosylation changes in misfolded proteins are of particular interest in understanding neurodegenerative conditions such as Parkinson’s disease (PD) and incidental Lewy body disease (ILBD). This work outlines optimizations of a microwave-assisted nonreductive release to limit glycan degradation and employs this methodology to analyze O-glycosylation on the human striatum and substantia nigra tissue in PD, ILBD, and healthy controls, working alongside well-established reductive release approaches. A total of 70 O-glycans were identified, with ILBD presenting significantly decreased levels of mannose-core (p = 0.017) and glucuronylated structures (p = 0.039) in the striatum and PD presenting an increase in sialylation (p < 0.001) and a decrease in sulfation (p = 0.001). Significant increases in sialylation (p = 0.038) in PD were also observed in the substantia nigra. This is the first study to profile the whole nigrostriatal O-glycome in healthy, PD, and ILBD tissues, outlining disease biomarkers alongside benefits of employing orthogonal techniques for O-glycan analysis.


Exoglycosidase Limitation
During data interpretation, there was growing evidence to suggest some enzymes were poorly digesting some glycan structures or simply not digesting them at all. It was determined that BTG (β1-3,4 galactosidase) had very low activity on cleaving β1-3 linked galactose if it was linked directly to a core GalNAc (the mucin-type core 1 disaccharide). This did not seem to affect other disaccharides that contained different linkages or were composed of different monosaccharides, such as GlcNAcβ1-6GalNAc or GlcNAcβ1-2man. Furthermore, there was severe inhibition of all galactosidases on βlinked galactose if the galactose was linked directly to a GlcNAc that itself was linked to a fucose.
Galactose, in this instance, would only digest once the internal fucose was removed using AMF (α1-3,4 fucosidase) or BKF (α1-2,3,4,6 fucosidase). This appeared not to be the case for ga3[ga4(fa3GL6)]GA however, which de-galactosylated without initial fucosidase treatment. Also present in some samples was a lactose (Galβ1-4Glc) contaminating peak which can be seen in some chromatograms at GU 1.96. In this case, SPG which targets specifically β1-4 linked galactose had little to no enzymatic activity, however BTG with its broader specificity for β1-4 and β1-3 could cleave the structure (this is why 2AB-glucose is present in the glycan list). Royle et al., 2002 has previously reported some of these observations 4 for O-glycans and Guile et al., 1998 5 for N-glycans. It is suggested that, for the case of the non-digestible disaccharides, the process of reductively aminating with 2AB forces the core monosaccharide into its open-ring conformation, which may inhibit recognition by specific exoglycosidases. In future studies, experiments with labelling reagents that leave the core monosaccharide in its closed-ring cyclic form (such as AQC or RapiFluor-MS) will need to be performed and compared with this method.

Microwave-assisted Non-Reductive O-Glycan Release
While results were comparable to similar ammonia-based non-reductive β-elimination techniques seen in the literature [6][7][8] , peeling from these methods remained a pressing issue. Through optimization, peeling was reduced to levels expected from a typical hydrazine based non-reductive release (approx. 15-20%), which fell more in line with non-reductively released O-glycans commercially available (Ludger TM ).
Moving forward to clinical samples in (PD/ILBD/healthy control), the major release and analysis S-5 methods were performed using this optimized ammonia-based non reductive release with 2AB labelling, however, reductive elimination was also performed for ion-trap MS experiments, which was advantageous as peeling here was negligible (no fluorescence detection however).
Due to the high number of variables that were tested in this work, many substitutions could be made to limit peeling from non-reductive β-elimination even further (whether it be identifying a more suitable alkaline reaction solution, or incorporating a more rapid tagging reagent for example). Another method to explore in the future is the creation of a flow system that removes O-glycans from the alkaline reaction solution immediately after release to limit peeling, as has been observed in the literature 9 .
Furthermore, low yields remained an issue, particularly during exoglycosidase digestion studies where larger sample sizes were required. In future work, release efficiency studies will need to be performed and the method further optimized.
Additionally, glycoproteomic approaches to glycan analysis by leaving the glycan attached to the peptide glycosylation site have proven valuable in characterizing not only the glycans but also their relative position as they appear on the native protein 10