Development of a chemical biology approach for glycoprotein discovery in Caenorhabditis elegans reveals glycosylated isoforms of multiple proteins with mitochondrial function

Abstract

Proper glycosylation is critical for cellular physiology and organism health. But understanding the cellular pathways that produce glycoproteins and the functional roles these glycoproteins play within the cell depends on having an adequate quiver of experimental approaches available. Together with my dissertation advisor, Dr. Patricia Berninsone and the assistance of excellent colleagues listed in the appropriate Chapters of this dissertation, I have developed several experimental methods to address these issues. Furthermore, the data gained from experiments using these methods have contributed new insights into our understanding of the glycosylation of proteins with mitochondrial function.The first such method is a chemical biology approach for the discovery of glycoproteins in Caenorhabditis elegans (C. elegans). Despite the classic strengths of C. elegans for cell biology and genetics, very few chemical biology approaches are effective in this system. To address this issue, I developed a method using C. elegans primary embryonic cells for azido-sugar metabolic labeling and then used this method to identify novel C. elegans glycoprotein candidates (Chapter 2). Additionally, in a project led by another student, Dr. Casey Snodgrass, and in which I played a supporting role, our lab adapted this approach for the discovery of phosphorylcholine modified N-linked glycoproteins in C. elegans (Appendix A). Thus we believe that azido-substrate metabolic labeling of primary embryonic cells from the relevant <italic>C. elegans</italic> mutants will be broadly useful as a method for identifying multiple types of post-translationally modified proteins and analyzing their synthesis pathways with the power of C. elegans genetic analysis.The second method that I developed is an experimental strategy for identifying proteins that contain different post-translational modifications between samples (Chapter 5). Currently, there are few tools available that enable the facile detection of differentially expressed post-translational modifications in a specific and high-throughput manner. To resolve this issue, I developed an experimental approach that combines the sensitivity and specificity of Click Chemistry with the high-throughput capabilities of two-dimensional differential gel electrophoresis (DIGE). In this approach, proteins in different samples are metabolically labeled with azido-substrate analogs, the azido-labeled proteins are marked with custom-synthesized Cy3- or Cy5-alkyne probes via Click Chemistry, and then differences in the post-translational modifications between the samples are visualized by multiplexing the fluorescent samples and electrophoresing them in two-dimensional gels using the established DIGE format. We demonstrate the utility of this approach by using it to visualize multiple differentially expressed glycoproteins between a mutant cell line defective in UDP-Galactose transport and the parental cell line. Because there are many azido-substrates already commercially available, we anticipate that the Click-DIGE approach could be compatible with a variety of post-translational modifications and thus could also be broadly useful to the research community.Furthermore, our experiments using these approaches have provided an intriguing biological insight. Namely, I have determined that multiple proteins with established and essential mitochondrial function have low-abundance glycosylated isoforms (Chapters 3 and 4). Using a small-scale candidate-driven approach, I first identified five such proteins (Chapter 3). Then by using three complementary, large-scale proteomic approaches for glycoprotein enrichment and detection from bovine heart mitochondrial fractions, I identified candidate glycosylated isoforms for 88 proteins with mitochondrial function, of which 67 are novel (Chapter 4). This represents a substantial contribution to the 101 such proteins previously reported. Because reports of glycosylated isoforms of proteins with mitochondrially function are historically rare, these results suggest that glycosylation of these proteins plays an unexplored role in the biology of mammalian and nematode cells. Furthermore, a significant contribution of our study is the finding that glycosylated isoforms of these mitochondrial proteins constitute a small fraction of the total protein pool. This observation may explain the low frequency with which such glycosylated isoforms have been previously identified in large-scale studies. It also suggests that glycosylation of such proteins is not constitutive and instead could potentially play a regulatory role. I also tested the hypothesis that O-GlcNAc glycosylated isoforms of mitochondrial proteins are glycosylated at a consensus motif that is different from that of another sub-group of O-GlcNAc glycoproteins. By performing a meta-analysis on previously published glycosylation sites, I found that mitochondrial proteins carrying O-GlcNAc on serine have significant position-specific differences in amino acid composition flanking the glycosylated site as compared to transcription factors or nuclear pore proteins that are also O-GlcNAcylated on serine. Furthermore, for one of the mitochondrial proteins for which I identified a glycosylated isoform (PDHE1&alpha;), I tested the hypothesis that the proportion of glycosylated PDHE1&alpha; increases when cells are grown under high environmental glucose conditions (Appendix B). I found no such detectable increase, thus the functional role(s) of this glycoprotein and all the other glycosylated isoforms we identified remain unknown.