Results

Wet Lab Results

SDS-PAGE

Deglycosylation under native conditions revealed partial deglycosylation of glycosylated lectins (A) based on visible mobility shifts and no change for non-glycosylated lectins (B). However, deglycosylation under denaturing conditions (C) revealed more complete deglycosylation. This indicates that some glycans may be inaccessible in the native form of some lectins. Full deglycosylation may require denaturing conditions or longer incubation. Whether glycans play a role in maintaining tertiary structure of lectins must be further investigated.

Lectin ELISA

ELISA assay results for non-deglycosylated (ND) control, 1 hour deglycosylation (D1), and 16 hr deglycosylation (D16) conditions reveal no significant difference (NS) in binding to the Fetuin-B model protein for most lectins. However, RCA-1 shows a significant (*) improvement in Fetuin-B binding from 1 hour of deglycosylation. PHA-E, on the other hand, shows a decrease in Fetuin-B binding from 16 hours of deglycosylation. These results reveal that contributions of lectin glycosylation to lectin-glycan binding may be lectin- and glycoprotein-dependent. Further experimental characterization is required.

Glycoprofile Predictions

Lectin-Binding Profiles

Fetuin-B lectin-binding profile outputs from the GlycoSeq pipeline indicate the relative amounts or strengths of binding for each lectin of interest. Profiles differed greatly between (A) expected, (B) non-deglycosylated, and (C) 1 hr deglycosylation conditions, especially for RCA-1. While RCA-1 is not expected to contribute significantly to Fetuin-B binding, we found significant lectin-glycan binding which was further increased when RCA-1 was deglycosylated. This reveals the need to update the GlycoSeq model's assumptions and/or train the model on a larger, more representative dataset.

GlycoSeq Fetuin-B Profile Predictions

Selecting the two lectins which showed significant differences in lectin-glycan binding based on the ELISA results, we studied how deglycosylation would impact the GlycoSeq pipeline's ability to predict the expected Fetuin-B glycoprofile. (A) PHA-E showed some improvements for all three of our experimental conditions compared to prior predictions. Deglycosylation also allowed for some recapturing of glycan 38 as well as a slight reduction in the incorrectly predicted glycans 4, 5, 6, 7, and 40. (B) RCA-1, however, did not show a clear improvement in achieving the expected Fetuin-B glycoprofile prediction. Nonetheless, predicted glycoprofiles do reveal differences when using data from deglycosylated lectins.

Interpretations

• Some glycans may be inaccessible in a lectin’s native conformation

  • Glycans may contribute to lectin tertiary structure

• PHA-E and RCA-1 have significantly different binding activity to Fetuin B with and without deglycosylation while other lectins show no significant difference in binding affinity

  • The role of lectin glycosylation may be lectin-dependent and requires further investigation

• Deglycosylated PHA-E came closer to model predictions for at least one glycan compared to non-deglycosylated PHA-E but the model still struggles to predict accurate profiles

Page Leader: Emma Finburgh