Structure and Functional Consequences of Serine vs. Threonine Mucin-Type O-linked Glycosylation

Division Of Basic Sciences - Nci

Investigators

Abstract

In previous studies (which have been discontinued for this cycle) we have studied the properties of a unique glycopeptide called antiproliferative factor (APF), that is found in the urine of patients with a bladder disease called interstitial cystitis/painful bladder syndrome (IC/PBS). APF is a potent antiproliferative agent, inhibiting normal bladder epithelial proliferation at nanomolar concentrations. APF was also found to be equally potent in inhibiting bladder tumor cell proliferation. Coincidentally, this nine-residue glycopeptide has a covalently attached trisaccharide on a Thr residue which is the sialylated version of the TF antigen described above. We discovered that the sialic acid is not needed for activity, and most analogues were prepared simply with O-linked TF antigen. One of our analogues with a TF antigen on a Ser in place of the natural Thr lowered activity by 4 orders of magnitude. We have defined the high-resolution structural features of APF, where we showed unique conformations about the glycosylated residue depending on whether the glycan was linked to serine or threonine. We have extended this structural analysis to our MUC4 glycopeptide described above, since we also observed specific properties of the glycopeptides depending on whether the TF antigen was Ser- or Thr-linked. We want to exploit these differences to inform our design of immunogens that are increasingly specific to MUC4-positive pancreatic tumor cells. Specific Aim 1. Detailed Structure/Function to Ser vs. Thr linked O-glycopeptides. In the last cycle, we have prepared glycopeptidomimetics of APF and showed that the structure can tolerate certain sugar modifications. However, the intriguing difference in activity of Ser vs. Thr analogues has prompted us to further study this structural detail in our MUC4 glycopeptides. The primary goal of this work is an attempt to further define the structural features that are most relevant to the presentation of these motifs on cancer cells. The more details we can determine, the better antigens we can synthesize to be used as immunogens and tools for immunotherapy tools. We have prepared various TF-Ser and TF-Thr-linked MUC4 glycopeptides, where a natural Ser/Thr was mutated to its "counterpart" (e.g., Ser-to-Thr, Thr-to- Ser) and studied their High-Resolution NMR structures. We have collaborated with the developer of a specific NMR program to attempt to automate the assignments and analysis of these glycopeptides from standard 2D NMR data. Specific Aim 2. Substrate Specificity of Mucin Peptides to N-Acetylgalactosaminyl Transferases (GALNTs) as a guide to Antigen Design. While several studies have shown that various TACAs are present on MUC4, no one has yet to define the actual glycans present through glycoproteomics. This aim proposes two ways to approach this issue: 1) Determine substrate specificity of MUC4 peptide sequences with various GALNTs that are overexpressed in pancreatic cancer. This may offer clues as to which Ser and Thr residues may be glycosylated, suggesting that synthetic constructs of glycopeptides with glycans at these positions would be tumor-selective antigens. 2) Perform modern full glycoproteomic analysis of tumor-associated MUC4 by high level mass spectrometry studies to determine directly where the glycans are covalently linked and what specific glycan structures are present at those sites. Immunization with these exact structures would hopefully more closely mimic the presentation of MUC4 to the immune system and offer the "most selective" immunogens for this target. We plan on synthesizing these constructs with our robust linker to both immunize animals for antibody production and use them in our latest vaccine designs for an active immunotherapeutic agent. We have already defined the substrate specificities of at least 10 peptide sequences from the TR of MUC4. This guided antigen design project 1 leading to a very selective antibody. With this hypothesis validated, we are now trying to define the glycan repertoire of tumor-associated MUC4 with our collaborators at Yale.