2.2 Proteomics
Another commonly employed set of methodologies for systems analysis in innate immunity are proteomics-based methods. Bottom-up proteomics using mass spectrometry (MS), commonly coupled with liquid chromatography (LC-MS), has been traditionally used to assay changes in the proteomic, phosphoproteomic, kinomic and secretomic innate immune landscape. Various omics strategies offer diverse biological conclusions and when combined and can elucidate the complex signaling environment that occurs in host-pathogen interactions. In the simplest of terms, bottom-up proteomics refers to the digestion of proteins into peptide components using suitable enzymes before analysis with LC-MS. Bottom-up proteomics is also the most widely used approach of proteomics-based methodologies since it provides information on protein identity and post-translational modifications (PTMs), both qualitatively and quantitatively. Bottom-up proteomics can be performed in numerous ways; data independent acquisition (DIA), data dependent acquisition (DDA) such as the case of shotgun MS proteomic profiling, and targeted LC-MS. DDA is usually utilized as a baseline experiment, or a first approach, in what is known as discovery level proteomics. It involves a semi-random selection of target analytes for fragmentation based on set cutoffs for relative intensities, which can be very useful for identifying unknown proteins without any prior knowledge about them. In DIA experiments, non-random selection of precursor ions is curated with a wide precursor ion isolation window. This accounts for more specificity and quantification accuracy for more peptides in comparison to DDA, with the tradeoff of a more challenging spectral analysis. In targeted LC-MS, the parameters describing target analytes to be fragmented are set by the user and include precursor mass to charge ratio (m/z ), LC elution time, and collision energy needed to drive fragmentation for each analyte. This makes targeted MS a useful tool for validation experiments or absolute quantification.12 Targeted proteomics has two main variations: selected reaction monitoring (SRM) (also known as multiple reaction monitoring or MRM when applied to more than one precursor and fragment ions) and parallel reaction monitoring (PRM). SRM allows for precise quantification across multiple sample groups and is typically used with LC-MS instruments that have a triple quadrupole (QQQ) where the first quadrupole (Q1) act as a mass filter and selectively monitors analyte precursor, fragmentation is carried out in the second quadrupole Q2, and guided to the third quadrupole Q3 for analysis. Untargeted peptides and their fragment ions are discarded leaving only the ions coming from the analyte of interest to be quantified by the detector over time.13,14 Like SRM, PRM also uses MS2 data as input, but rather than selectively monitoring a single target analyte’s fragmented ions, the full MS2 spectra is scanned at high resolution and the intensity of multiple fragment ions is monitored in parallel.14
Bottom-up proteomics approaches also allow for both relative and absolute quantification of a target analyte across multiple biological sample groups using different labeling techniques. Relative quantification can also be achieved with a label free approach (LFQ), where LC-MS is done on each sample separately. When labeling is done prior to LC-MS, such as stable isotope labeling, samples can be pooled together (both labeled and unlabeled) and run in a single LC-MS run to quantify the target analyte based on the ratio of labeled to unlabeled forms of the analyte. Absolute quantification can be done by comparing to known standards of the stable isotopes. There are many different types of labeling strategies available that can be employed for metabolic labeling; stable isotope labeling by amino acids in cell culture (SILAC) using heavy isotopes of carbon and nitrogen is frequently used.15,16 There is also an abundance of chemical labels that are utilized such as tandem mass tags (TMT), isobaric tags, 18O labels, dimethyl labeling, and others.17,18,19 All of these labeling methods, when compared to known peptide standards, provide highly accurate and reproducible workflows for both relative and absolute quantification of target analytes across multiple biological sample groups.12 It should also be noted that bottom-up proteomics workflows can be adapted for the detection of PTMs.20 PTMs are chemical modifications on the side chains of amino acids within a peptide. Thus, detection methods rely on tandem MS to compare the mass shift of the unmodified tryptic peptide to the modified peptide. The immune system signaling is tightly controlled by changes in post translation modifications of protein complexes, which are highly dynamic and often labile. Lower abundance of modified peptides compared to the non-modified pool adds another layer of complexity in the analysis of PTMs. Therefore, enrichment of post translationally modified peptides is almost always needed for their detection by LC-MS. Some of the most common PTM enrichment techniques include immobilized metal affinity chromatography (IMAC)21 and metal oxide affinity chromatography (MOAC)22 for the detection of phosphorylated sites, di-glycine remnant (KεGG) immunoaffinity enrichment for ubiquitination site identification,23 and hydrophilic interaction chromatography (HILIC) for glycopeptide enrichment.24Quantitative measurement of PTMs can be performed along with their identification by mass spectrometry, as discussed here for the phosphoproteome analysis in immune cell signaling.25As with any quantification method, data analysis software is essential for making useful interpretations of the data. Comprehensive overview of the software available for targeted proteomics analysis and the advantages and disadvantages to using each application can be found elsewhere.26 Bottom-up proteomics methodologies are widely used within systems biology for their ability to investigate functional signaling dynamics and represent a fundamental area of research in the battles against new and emerging diseases. For example, in a study conducted by Wendisch et al., using a multiomics approach including shotgun-based proteomics, it was found that monocyte derived macrophages accumulate in the lungs during acute respiratory distress syndrome (ARDS) leading to a phenotype like the one observed in patients with pulmonary fibrosis. Their data supports the hypothesis that SARS-CoV-2 induces a profibrotic transcriptome and proteome profile within macrophages. Usually, these damage repair pathways when activated within macrophages help to control inflammatory mediated tissue damage. However, when left unchecked, these pathways can also lead to dysregulated fibroproliferation as well as protracted respiratory failure.27 This study is one example of how proteomics-based methods can further our understanding of host-pathogen interactions and demonstrates the need to develop more proteomics-based workflows that will broaden our knowledge of immune cell pathogen interactions, and aid in the development of effective treatments to limit the emerging viral diseases on human health.
While perhaps the most common approach, bottom-up proteomics is not the only set of proteomics-based methodologies used to quantify relative and absolute abundance of target analytes within a given proteome. Both top-down and middle-up/down proteomics can provide valuable information on the higher order structure and intact topology of the target protein, where bottom-up falls short owing to a completely different sample processing approach. Top-down proteomics permits analysis of intact proteins and even native-like protein complexes, without the need for proteolysis, and is useful for characterizing functional and genetic isoforms of target proteins. Since the proteins stay undigested, top-down MS also preserves PTM sites and allows for the characterization of entire proteoforms.28 With sufficient number of useful fragments, top down can provide complete primary sequence characterization and at the same time reveal modifications on the protein.29 The sample preparation methods are however, not as high throughput or standardized as in bottom up proteomics, high-resolution mass analyzers (also more expensive and somewhat less user friendly compared to the bench-top instruments) are needed to resolve charges on high mass precursors, and the data analysis is needs specialized knowledge, which keeps top-down rather out of reach, although recent developments are promising. There have also been many attempts at adapting bottom-up targeted proteomics labeling techniques to top-down workflows, such as TMT and SILAC labeling.30 These labeling techniques can be used for absolute quantification of target peptides or for entire proteoforms and negates the protein inference problem in bottom-up proteomics. Conclusions for protein identification in a bottom-up experiment rely on inferring the identification of a protein based on the peptides that compose the protein. In top-down proteomics, the data is representative of the parent protein which allows for direct protein and proteoform quantification, and when combined with the appropriate fragmentation technique, has the potential to achieve near complete identification of individual proteoforms. In concordance with the developments in targeted top-down proteomics, untargeted approaches with LFQ have also been developed and are especially useful for characterizing changes in proteoform abundance between different samples. For a more comprehensive overview on the quantitative proteomics-based methodologies for quantification of target proteins, Neagu et al. have provided a review summarizing various MS-based approaches and applications of tandem MS for protein analysis in biomedical research.31