Mass Spectrometry-Centered Multi-Omic Applications In The Analysis Of Inflammation And Exposure

Abstract

Bottom-up proteomics represents an exciting technology which has found great utility across multiple fields of biological research. Using high-resolution mass spectrometry coupled with sophisticated bioinformatic software applications, bottom-up proteomics affords qualitative and quantitative information that reflects the actual molecular phenotype of a system in ways that next-generation sequencing technologies do not. Despite this, there are also blind spots in conventional bottom-up proteomics experiments; many of these limitations can be abrogated via the integration of bottom-up proteomics with other forms of ‘omics technologies and data. Through supplemental bioinformatic workflows, putative identifications of non-canonical peptides or non-host peptides (e.g microbial, viral) can be validated. The use of RNA-Seq data can be used to generate protein sequence databases files for proteogenomics, where non-canonical peptide sequences arising from genomic mutations, translocations, aberrant splicing events, etc. which are invisible to conventional proteomics experiments can be readily detected. In addition, by integrating and directly comparing proteomics data with transcriptomic data, levels of epigenetic and/or post-transcriptional control can be examined in a system in response to stimuli of interest that are invisible to both technologies. These supplemental approaches expand the power of bottom-up proteomics to where it becomes a highly useful tool for studying systems in which multiple levels of gene product expression response are regulated, including viral infections, tissues undergoing long-term inflammation, and exposure to endogenous and exogenous electrophiles.The first chapter of this thesis provides an overview of the current state of proteomics-centered multi-omic technologies and their potential utility in biological research. The review begins with outlining the improvements to bottom-up proteomics technologies which have enabled a greater depth of information, such as isobaric peptide tagging, data-independent acquisition, ion mobility applications, and other instrumental advances. From there, bioinformatic tools are discussed that are of use in the analysis of proteomics data, with a focus on integrating mass spectrometry-based data with other forms of ‘omics data. Specific applications of using RNA-Seq data to inform the data analysis of proteomics, proteogenomics, are also addressed. The chapter concludes with notable instances of proteomics-centered multi-omics analysis as well as potential future applications of these technologies. The second chapter of this thesis addresses the analysis of open-source proteomics datasets with customized multi-omic bioinformatic tools to determine the optimal targets for the detection of SARS-CoV-2 infections in patients. Through the use of in vitro and patient datasets, a panel of potential viral peptides were established, was used to search patient datasets. Ultimately, we found four peptides in the viral nucleocapsid which were reliably detected in patients and were unique to the SARS-CoV-2 virus. The third chapter of this thesis utilizes proteogenomics workflows to examine the consequences of long-term inflammation in the proximal colon tissue of a murine model of inflammatory bowel disease. In this model, Rag2-/-Il-10-/- mice were subjected to five months of Helicobacter hepaticus infection in their colon to trigger chronic infection. For these analyses, RNA-Seq data acquired from these test subjects in an earlier study were converted into a FASTA protein sequence database containing variant sequences stemming from this treatment. Through quantitative proteogenomic analysis we noted significant changes in abundances of proteins consistent with an inflammatory response; through bioinformatic analysis of our data, we also validated and confirmed the presence of 39 non-canonical peptides across our infected and control samples, demonstrating the importance of validation of targets of interest in proteogenomic studies. The fourth chapter of this thesis integrates multiple levels of ‘omic analysis to examine the effects of inflammation on murine Type II pneumocytes, a constituent cell within alveoli which serve as the source of lung adenocarcinomas. Mice were exposed to intranasal dosages of LPS or to whole-body cigarette smoke exposure for variable amounts of time before being sacrificed and the Type II cells isolated for analysis. Bottom-up proteomics of cells subjected to LPS for 3 weeks revealed a phenotype consistent with inflammation; this was reinforced when compared to transcriptomic data from the same cells, as these showed. Global proteomics analyses of Type II pneumocytes of mice subjected to exposure to cigarette smoke revealed significant changes in protein abundances occurring after after 10 weeks of exposure with a 4-week recovery period post exposure, encompassing biological processes such as nucleotide and amide metabolism as well as synthesis and acetyl CoA synthesis, which demonstrated a greater degree of disjuncture with the associated RNA-Seq data as compared to our LPS study. The fifth chapter of this thesis examines the utility of bottom-up proteomics in examining the formation of amino acid adducts in hemoglobin, which serves as a valuable reservoir for exposome studies due to its longevity and high concentration within the blood. We were able to validate the presence of 4-hydroxybenzyl adducts at the N-terminal valine of hemoglobin and demonstrate their formation at nucleophilic side chains within the protein. In addition, we compared bottom-up proteomics to the FIRE method, an experimental procedure which serves to isolate N-terminal adducts in hemoglobin for LC-MS detection, with a panel of electrophilic compounds incubated with blood at various concentrations and incubation times. Ultimately, we found that a proteomics-based approach to untargeted adductomics allowed for the detection of novel adducts at a number of sites within hemoglobin. In this thesis we have applied mass spectrometry-based ‘omics technologies to complicated biological systems. We have utilized publicly available proteomics datasets to determine the optimal targets for MS-based detection of SARS-CoV-2 in patient samples. Using RNA-Seq data, we performed quantitative proteogenomic analysis of a murine model of IBD and validated the presence of several non-canonical peptide sequences. We also used multi-omic analyses to compare LPS-driven and cigarette smoke-driven inflammation of murine Type II pneumocytes. Finally, we demonstrated the utility of bottom-up proteomics in detecting and characterizing adducts in human hemoglobin as a record of the exposome. Overall, this work expands the utility of proteomics-centered analyses in characterizing systems subjected to viral infection, inflammatory stimuli, and exposure to environmental contaminants.