Mucinomics Enables Characterization of Mucins in the Tumor Microenvironment

Date of Award

Spring 1-1-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Chemistry

First Advisor

Malaker, Stacy

Abstract

Glycosylation is a highly prevalent and heterogeneous post-translational modification (PTM), modifying 80% of the human proteome. While N-linked glycosylation occurs in a conserved motif, O-glycosylation can modify any Ser or Thr across the protein backbone. Many proteins are O-glycosylated, but mucins have the highest density of O-glycosylation within their PTS "mucin" domains. Mucins and their glycosylation are key for protection of epithelium, immune signaling, and homeostasis, and are dysregulated in most malignancies. Unfortunately, most biochemical characterization and structural prediction techniques are often unsuccessful in uncovering the glycan heterogeneity at a molecular level. Mass spectrometry is a highly sensitive and versatile technique that has been used to identify PTMs during the past three decades. That said, there are many challenges associated with mucin glycoproteomic analysis, some of which will be addressed in this dissertation. Most mucin O-glycoproteomic-centric workflows include the addition of N-deglycosidase PNGaseF. We reported that these released "free" N-glycans could be identified during data acquisition, hampering detection of O-glycopeptides. Additionally, these glycans can adduct onto unmodified peptides, incorrectly being identified as O-glycopeptides by search algorithms. We demonstrated that these were present in several of our experiments, as well as data published by others in the field. To solve this problem, we showed that free N-glycans can be removed using a 10 kDa molecular weight cut off filter following PNGaseF treatment. O-glycopeptides are suppressed by unmodified peptides due to their hydrophilicity and low abundance. Thus, most workflows include a glycan-focused enrichment method prior to MS analysis. High-field asymmetric waveform ion mobility spectrometry (FAIMS) can separate glycopeptides in the gas phase prior to MS analysis, improving glycopeptide detection. We showed that FAIMS separation was beneficial for N- and O- glycopeptide analysis for highly complex samples. Importantly, we noticed - in FAIMS fragmentation - (IFF), a phenomenon where glycans fragment during FAIMS separation, leading to an artificially high number of identified glycopeptides. FAIMS experiments exhibited a 2- to 5-fold increase in source fragmentation compared with control experiments, which was also observed in publicly available data. Our work indicated that this was likely a systemic occurrence when using FAIMS. Database searching of glycoproteomic RAW data remains an arduous challenge present in the field. To solve this issue, several glycoproteomic search algorithms have emerged in the last decade, with a repertoire of glycan scoring and filtering techniques. To determine the efficacy of these new tools, we analyzed three common glycoproteomic sample types with a variety of search algorithms and manually validated the output. Our work highlighted the continued need for manual validation of glycan identifications, especially for O-glycopeptides. Based on our results, we provide a series of recommendations to the field with regards to which search algorithm to use based on the analysis type. With these tools in hand, we thus sought to understand the role of LAMP-1, LAMP-2, and CD68 glycosylation in the tumor microenvironment. LAMP-1 and -2 are lysosomal proteins that are translocated to the cell membrane in cancer cells. While we know that their glycans interact with metastasis-promoting glycan-binding proteins, little is understood about their glycosylation in the molecular level. Here, we performed a comprehensive glycoproteomic mapping of LAMP-1 and LAMP-2 using a combination of trypsin, GluC, and thermolysin. Our results showed that trypsin and thermolysin allowed for the mapping of complementary N-glycan structures and N-glycosites. We confirmed occupancy of 17 and 15 N-glycosites for LAMP-1 and LAMP-2, and 14 O-glycosites in both proteins. While LAMP-1 and LAMP-2 share 40% sequence homology, the N- and O- glycans decorating these proteins are vastly different. CD68 is a mucin-domain glycoprotein that is overexpressed in macrophages, including tumor-associated macrophages. While we know that its glycosylation gets remodeled during macrophage phagocytosis, and that it is overexpressed in tumor-associated macrophages, little is known about its implication in the tumor microenvironment. Here, used a combination of glycoproteomics and lectin staining to understand CD68 glycosylation at the molecular level, potential CD68-interacting proteins that could lead to immune inhibition, and glycosylation changes across macrophage function and polarization. We identified differential sialylation and N-glycosylation between monocytes and macrophages, and uncovered two CD68 proteoforms with differential expression in polarized macrophage subtypes.

This document is currently not available here.

Share

COinS