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