Over the last quarter century, protein microarray technology has emerged as a prominent field in scientific study.The versatility of the platform, coupled with the ability to characterize thousands of proteins in a parallel and high-throughput manner, has resulted in great strides in our knowledge database.Many clinical studies have used protein microarrays as analytical tools to identify biomarkers using analytical protein microarrays. These allow for the detection of varying expression levels of proteins in a cell lysate as well as binding affinities and specificities of a sample. One newer technique that has emerged uses cell or tissue lysates as the arrayed substance (reverse-phase protein microarray). This powerful tool allows for the determination of altered protein modifications from two different samples, allowing researchers to identify potential biomarkers in diseased tissue. Perhaps the most well-used microarray type is the functional protein microarray, or ;;target protein microarray.” The ability to performed hundreds or thousands of individual reactions in parallel becomes an unbiased and powerful tool that scientists can use to draw conclusions at a global or systematic level.As science and technology continues to advance, new approaches to old paradigms are often challenged. My thesis projects involve utilizing protein microarrays and bioinformatics to change how we view phosphorylation events and the cross-talk between phosphorylation and glycosylation. The wealth of information in databases continues to grow as larger proteomic-wide studies are carried out and deposited for all to use. This allows bioinformaticians to predict new interactions that were never possible before. One such interaction involves scaffolding proteins, proteins that can interact with at least two other proteins in signaling pathways. Scaffolding proteins have been studied for years and have been found to play critical roles in cellular signal transduction. In this study, human protein-protein interaction (PPI) and kinase-substrate relationship (KSR) networks were used to predict scaffolding proteins involved in phosphorylation signal transduction. We predicted 212 scaffolding proteins involving 612 non-redundant phosphorylation pathways. One third (359 of the 1,103 known KSRs) of the phosphorylation-mediated signaling pathways are known to be regulated by at least one scaffold protein. We examined that the predicted scaffolding proteins are enriched for protein domains known to interact with phosphorylation sites, and exhibit similar characteristics of other known scaffolding proteins. Intriguingly, the predicted scaffolding proteins tend to have large protein sizes, perhaps due to their ability to adapt to multiple interactions with other proteins. Furthermore, these proteins are more evolutionarily conserved, suggesting important roles in different biological processes across species. When comparing to other human proteins, scaffolding proteins also contain more known phosphorylation sites, indicating that the scaffolds themselves might be regulated by phosphorylation-mediated signal transduction. In order to test these predictions, microarrays were employed on the human proteome microarray, which contains over 17,000 full-length human proteins. CSNK2A1 (with predicted scaffolds PIN1 and ATF2) and MAPK9 (with predicted scaffold ATF2) were used to treat these microarrays in the presence of 33P-γ-ATP. After careful alignment and scoring, the predicted scaffolding proteins PIN1 and ATF2 were shown to mediate 28 distinct phosphorylation events. Through this initial study, we have shown that our initial predictions appear to hold merit.Mutations and dysregulation of kinases play causal roles in human disease, development, cell signaling, and metabolism. Understanding the function of kinases continues to be of interest for biomarker discovery as well as for the development of agonists and antagonists for use in disease therapy. O-linked β-N-acetylglucosamine (O-GlcNAc) is a post-translational modification known to regulate a variety of protein functions, including localization, enzyme activity, and protein stability. Like phosphorylation, O-GlcNAcylation modifies serine and threonine residues on nuclear and cytoplasmic proteins, and is a ubiquitous and reversible process that regulates cellular signaling. Recent evidence indicates that site-specific crosstalk between O-GlcNAcylation and phosphorylation as well as the O-GlcNAcylation of kinases plays an important role in regulating cell signaling. Therefore, it is very important to study the O-GlcNAcylation of the kinome. Previous studies utilizing a functional kinase array were able to identify 42 kinases as substrates of O-Glycosyltransferase (OGT) using an in vitro OGT assay with [H3] radiolabeling. While some promising results were obtained, the limitations of [H3] labeling demand a more sensitive approach if one is to examine a larger library of proteins. Herein, using multiple in vitro OGT labeling assays and immunofluorescent detection techniques, we were able to obtain a high confidence hit list based on shared hits from three different detection methods. A kinome array was fabricated, which contains 350 unique full-length human kinases representing approximately 70% of the human kinome. After in vitro labeling by OGT, arrays were treated either with one of two Cy5-labeled antibodies that recognize O-GlcNAcylated residues (RL2 and CTD110.6) or with β-1,4-galactosyltransferase (GalT1 (Y289L)), which transfers azido-modified galactose (GalNAz) from UDP-GalNAz to O-GlcNAc residues of modified proteins. A simple click chemistry reaction followed with incubation of Alexa Fluor® 647 DIBO alkyne to fluorescently label the hits. Data analysis revealed that these hits were very reproducible and a total of 104 kinases were shared between all three methods. Many hits were subsequently validated both in vitro and in vivo, further validating our methods. Among the hits, O-GlcNAcylated sites were identified via mass spectrometry for BRSK2 and PAK4, with mutagenesis studies on the latter validating the identification. Through this simple, yet sensitive strategy, we have shown with high confidence that at least 20% of the human kinome (and 31% of those kinases tested) is glycosylated and the dataset created will likely spawn many further validation studies in the future.In conclusion, we show the utility of both bioinformatics and microarrays to predict novel functions of proteins and to probe an entire family of proteins for post-translational modifications.
PDF
download
878987797
028-85220240
