Evolution of proteomics technologies for understanding Respiratory Syncytial Virus pathogenesis

Abstract

Introduction:

Respiratory syncytial virus (RSV) is a major human pathogen associated with long term morbidity. RSV replication occurs primarily in the epithelium, producing a complex cellular response associated with acute inflammation and long-lived changes in pulmonary function and allergic disease. Proteomics approaches provide important insights into post-transcriptional regulatory processes including alterations in cellular complexes regulating the coordinated innate response and epigenome.

Areas covered:

Peer-reviewed proteomics studies of host responses to RSV infections and proteomics techniques were analyzed. Methodologies identified include 1).” bottom-up” discovery proteomics, 2). Organellar proteomics by LC-gel fractionation; 3). Dynamic changes in protein interaction networks by LC-MS; and 4). selective reaction monitoring MS. We introduce recent developments in single-cell proteomics, top-down mass spectrometry, and photo-cleavable surfactant chemistries that will have impact on understanding how RSV induces extracellular matrix (ECM) composition and airway remodeling.

Expert Opinion:

RSV replication induces global changes in the cellular proteome, dynamic shifts in nuclear proteins, and remodeling of epigenetic regulatory complexes linked to the innate response. Pathways discovered by proteomics technologies have led to deeper mechanistic understanding of the roles of heat shock proteins, redox response, transcriptional elongation complex remodeling and ECM secretion remodeling in host responses to RSV infections and pathological sequelae.

1. Introduction.

1.1. Respiratory Syncytial Virus.

The goal of this review is to examine the contribution of mass spectrometry-based proteomics technologies to the study of respiratory syncytial virus (RSV) infection, and to describe new methods and technologies in this field which could be applied to improve our understanding of this pathogen and its host response. RSV is an enveloped negative-sense RNA virus that circulates seasonally worldwide. Most often, RSV causes mild, cold-like symptoms, but in the very old and very young, the virus spreads into the lower airways, causing bronchiolitis or pneumonia. Worldwide, RSV was associated with 33.1 million episodes of lower respiratory tract infections (LRTIs), 3.2 million RSV-related hospital admissions, and 118,000 deaths in children less than 5 years of age, predominantly in developing countries [1]. Consequently, RSV infection is responsible for the majority of pediatric hospitalizations of young children [2].

RSV is an Orthopneumovirus belonging to the Pneumoviridae family [3], that forms a filamentous enveloped virion encapsulating a single-stranded, negative sense genome of approximately 15.2 kb in length [4]. Molecular epidemiological studies have shown that epidemic RSV infections occur seasonally, and represent co-circulation of two principal strains of RSV- known as RSV-A and RSV-B viruses [5]. These strains are due to antigenic variations of a hypervariable region of the G glycoprotein. Within each subtype are multiple clades of genetic variants whose abundance and composition are determined by local fitness, transmissability and immune-driven evolution [5]. Recently, rapidly transmissible RSV genotypes have been identified and represent the dominant forms of RSV strains. These include an RSV-A subtype with a 72 nt duplication of G (ON genotype) and an RSV-B subtype with a 60 nt duplication (BA genotype) [6,7]. Although these variants are highly transmissible, it is unclear whether these are more pathogenic than others; however, other genetic RSV variants elicit distinct patterns of immune responses, such as Line 19, a variant that elicits a more robust IL-13 driven T-helper type 2 response than that produced by RSV-A2 and Long strains [8]. The lifecycle of RSV strains are thought to be identical and illustrated in Figure 1.

Figure 1. Life cycle and cellular responses to RSV infection.

RSV infects epithelial cells in the respiratory tract. Upon binding to extracellular receptors, RSV fuses with the epithelial membrane, forming a fusion pore, permitting the capside entry into the cytoplasm. The RSV replication complex forms on cellular membranes triggering formation of inclusion bodies. After transcription, genome replication results in mature virus that buds from the cell cytoplasm. In the infected cell, pattern recognition receptors trigger an innate immune response that involves the production of reactive oxygen species, as well as the secretion of protective IFNs, cytokines, and exosomes. The application of proteomics technologies has elucidated key cellular responses to RSV infection, including activation of anti-oxidant homeostasis, dynamic regulation of transcriptional complexes in the IIR, and cell type differences in secreted proteins (secretome).

RSV virions are covered in attachment (G) and fusion (F) glycoproteins with a lesser amount of small hydrophobic (SH) proteins within the envelope. RSV F and G mediate viral attachment and entry into cells; RSV primarily infects epithelial cells of the respiratory tract as its initial target [3]. A number of cellular receptors have been identified for RSV including the CX3 chemokine receptor, CXC3CR1[9], Toll like receptor-4 [10], nucleolin [11] and IGF1R [12]. RSV F mediates fusion between the viral envelope and host membrane; because of its relatively invariant structure and is a target of neutralizing antibody formation, RSV F is one of the major targets for anti-viral drug development [13]. Membrane mixing permits the formation of a fusion pore, resulting in delivery of the virion capsid into the host cell, consisting of the nucleocapsid (N), L and P proteins (Figure 1). The RNA genome encodes for the key internal structural proteins (matrix protein [M] and nucleoprotein [N]), proteins required for a functional polymerase complex (phosphoprotein [P] and polymerase [L]), A viral replication complex then forms on internal cellular membranes, consisting of the polymerase (L), P and M2–1 proteins encoded by RSV, along with N protein-encapsulated viral genomic RNA.

The polymerase complex transcribes the nonstructural proteins (NS-1 and NS-2) involved in evasion of the innate immune response, transmembrane glycoproteins (small hydrophobic protein [SH], glycoprotein [G], and fusion protein [F]), and the regulatory M2 proteins (M2–1 antitermination protein and M2–2, involved in transcription/replication regulation [4]. After transcription of the RSV-encoded proteins, the polymerase switches modes and produces genome. The assembled virus then buds from the infected cell, incorporating components of the host plasma membrane.

Productive infection of the respiratory epithelium induces multiple cellular responses, including the formation of inclusion bodies, reactive oxygen stress, and rapid activation of the innate immune response (IIR). IBs appear to be both a site of RSV replication as well a mechanism for modulating the intracellular IIR [14]. A hallmark of the IIR involves resulting in secretion of cytokines [15], interferons [16], exosomes [17], and damage-associated patterns [18] that mediate mucous production and leukocytic inflammation. After viral clearance, immunity is partial and incomplete. Consequently, reinfections occur throughout life, typically causing upper respiratory tract infections. In the elderly, nosocomial RSV infections cause pneumonia. With repetitive infections, innate inflammatory signaling can result in airway remodeling [19]. Airway remodeling involves physical changes to the airways, including enhanced production of extracellular matrix (ECM) and mucous production, producing obstruction, reducing effective airflow and lung function [20]. These changes are significant and irreversible [21].

Presently, no vaccines are available to prevent RSV infection. Both formalin-inactivated and live-attenuated viruses have failed to induce protective responses in clinical trials, and in-fact, have generated exaggerated immune responses [22–25]. In the case of a 1969 formalin-inactivated RSV vaccine trial, this resulted in a 80% hospitalization rate in the vaccine-treated infants compared to only 5% in controls [26]. Monoclonal antibodies and pharmaceutical prophylactics are available for high-risk infants, but they are expensive and have questionable efficacy [27,28]. In light of the high prevalence of RSV and severe consequences of repeated infection, it is clear that additional research is required to identify long-term solutions. In this review, we will describe the application of proteomics technologies and how these studies have elucidated key cellular responses to RSV infection, including activation of anti-oxidant homeostasis, dynamic regulation of transcriptional complexes in the IIR, and cell type differences in secreted proteins (secretome).

1.2. Mass Spectrometry & Proteomics.

RSV-infection alters cell physiology through complex and multifaceted mechanisms, including activation of the innate immune reaction through TLR3 Retinoic Acid Inducible Gene I (RIG-I) and other pattern-recognition receptors [16,29], resulting in altered transcription, post-translational modification of regulatory proteins, and epigenetic chromatin regulation [30]. Additionally, perturbations to the extracellular environment and non-infected cells play significant roles in the development of disease pathology [31]. While genomic and transcriptomic approaches can probe some of these changes at the level of gene expression, they are unsuitable for studying the true abundance of regulated proteins and post-translational modifications [32].

In contrast, Mass Spectrometry (MS)-based proteomics methodologies are capable of unambiguously and precisely quantifying changes to both protein expression and the relative abundance of most post-translational modifications (PTMs). In traditional “bottom-up” proteomic analysis, proteins are isolated and digested using trypsin or another protease [33]. The resulting peptides are then sequenced using tandem mass spectrometry, generally coupled to liquid chromatography (LC-MS) to reduce sample complexity. This method is sensitive and highly specific. More importantly, it is unbiased and high-throughput, which makes MS-based proteomics ideal for examining the cellular effects of RSV-infection at a systems biology perspective.

1.3. Survey of Shotgun Proteomics Approaches.

Bottom-up or “Shotgun” proteomics is highly versatile, and when combined with different sample preparation and molecular biology techniques, can be applied to extract an incredible diversity of information from biological systems. Most relevant to the discussion of RSV biology are its applications to protein and PTM quantitation from whole cell lysates, subcellular (e.g. nuclear and cytoplasmic) fractions, tissue matrices, and protein interactomes.

The simplest of these methods involves the direct extraction and digestion of protein from whole cells or tissue, which are then submitted to MS analysis for identification and quantitation. For convenience, we will refer to all such applications as “global bottom-up proteomics” (GBUP). The first RSV-oriented GBUP studies used 2-Dimensional gel electrophoresis (2DE) to separate proteins based on intact mass and isoelectric properties [34–36]. Protein stains, (e.g. Coomassie R-250, Sypro-Ruby) were then used to identify differentially abundant proteins, which were subsequently excised, digested in-gel, and identified via Matrix-assisted Laser Desorption Ionization – Time of Flight (MALDI-TOF) mass spectrometry and Peptide Mass Fingerprinting (PMF). Notably, PMF identifies proteins by matching the intact masses of tryptic peptides – not by peptide sequencing; resulting in generally high false-identification rates [37]. With advances in liquid chromatography and mass spectrometers over the last decade, such gel-based methods are no longer the primary tool for quantitative studies [38], nevertheless, they contributed important insights to the field of RSV biology.

More recently, reverse-phase liquid chromatography (RPLC), coupled to high-resolution mass spectrometry has become the standard methodology for bottom-up proteomics. This enables the detection of significantly more proteins and peptides than gel-based methods by improving the front-end separation as compared to 2DE; highly-reproducible LC also facilitates direct, or “label free” quantitation of peptides and proteins via extracted ion profiles [39]. Alternative methods of quantitation have also been applied to RSV research, most notably isobaric quantitation methods such as Stable Isotope Labeling by Amino Acids in Cell Culture (SILAC) [40], which enables direct binary comparisons within a single mass spectrum by altering the mass of otherwise identical peptides from different samples.

GBUP strategies are easily applied to study subcellular and organellar proteomes. Airway tissue is similarly accessible [41,42], but extracellular matrix proteins require additional processing steps for efficient extraction and digestion [43–46]. Finally, GBUP can be coupled to common molecular biology affinity purification techniques; Affinity-purification- Mass Spectrometry (AP-MS) allows for deep and unbiased profiling of protein interactomes [47]. We summarize these approaches in Figure 2.

Figure 2. Schematic of Bottom-up Proteomics approaches for studying RSV biology.

Post-infection with RSV, airway cells, tissue, and extracellular fluids (including Bronchoalveolar Lavage Fluid and Nasopharyngeal Aspirates) are subjected to extracellular matrix (ECM) enrichment, immunoprecipitation, subcellular enrichment, or lysed whole to harvest protein. The extracts are digested using trypsin or another protease, and the resulting peptides are analyzed via LC-MS to identify proteins and post-translational modifications and facilitate quantitation.

One challenge of GBUP is the quantitation of low-abundance proteins. Most commonly, mass spectrometers acquire a limited number of peptides for sequencing, fragmenting only the most abundant precursor ions within a single mass spectrum. In this process (termed “Data-dependent Acquisition”), low abundance peptides and proteins may be missed. To address this, two alternative strategies may be employed: Data-independent acquisition (DIA), which attempts to fragment all precursor ions in a given mass spectrum [48], but greatly increases the complexity of data analysis; and Selected- or Multiple-Reaction Monitoring (SRM/MRM), which targets specific precursor ions for fragmentation and sequencing regardless of their relative abundance in the sample [49], but has relatively low throughput and requires prior knowledge of the target proteins and peptides in the sample. Isotope labeled peptide standards may also be employed to ensure specific identification and accurate quantitation.

2. Applications of MS-based proteomics to cell culture-based RSV systems.

Many RSV-related proteomics studies have been conducted using in-vitro culture systems with malignant, KRAS-mutation transformed cells – most commonly A549 type 2 alveolar epithelial cells. A549 cells are highly differentiated adenocarcinoma cells that maintain characteristics of alveolar type II epithelial cells [50]. These cells are highly permissive for RSV replication [51], express type I and III interferons and cytokines upon RSV infection [15], have endogenous TLR3 and RIG-I pattern recognition receptors [16], and elaborate an integrated interferon response factor (IRF)-nuclear factor kB (NFκB) pathway [52–54]. These cells therefore are a valid model system of understanding pathogenesis of RSV-induced lower respiratory tract infections that has provided information on cell stress response validated in animal models and human infections. Other cell types, such as Hep2, have been used. Hep2 are derived from laryngeal carcinoma cells but have been shown to have HeLa contamination [55]. Fewer studies have used primary airway epithelial cultures, despite the benefits of the more human-like model, as historically, their added expense and relatively short lifespan in culture have imposed severe limitations on many typical workflows [56]. In recent years, LC-MS sensitivity has increased to the point where efficient label-free quantitation of primary cells is possible [57–59], but these benefits have not been applied to RSV models. For many of the same reasons, multicellular lung organoids – which may recapitulate multicellular interactions seen in pulmonary biology [60,61] - have yet to be applied to RSV proteomics. Nevertheless, some findings from transformed cell lines have been validated in primary cell cultures [17,62] and in-vivo [63,64], lending credence to these models. Another important consideration in the interpretation of these studies is that the cellular models are typically done with high multiplicity of infection (MOIs), so that most cells are actively infected or repetitively infected at the time of analysis.

2.1. Heat Shock Proteins are regulated by RSV infection and are required for viral polymerase activity.

Understanding the protein composition of cellular organelles and how these change in response to RSV infection provides an added dimension of information into the homeostatic response to virus and how the virus may manipulate these changes to enhance its replication. Consequently, substantial effort has been devoted to the characterization of airway epithelial cell nuclear proteome.

In one early approach to address this, subcellular fractionation, high-performance liquid chromatography, 2DE, and MALDI-TOF MS were combined to deeply probe the nuclear proteome of A549 cells and quantify its response to RSV infection [65]. Overall, RSV induced changes to 24 nuclear proteins controlling chromatin remodeling, protein refolding, cytoskeletal structure, membrane function, metabolic processes, mitochondrial function, RNA binding, protein synthesis, signaling, and transcription factor activities. These included large upregulations in the abundance of cytokeratins, heat shock proteins (Hsp60/Hsp70), and components of Nuclear Domain 10 (ND10) structures. These results were used to inform confocal immunofluorescence microscopy experiments which demonstrated that RSV infection disrupted nuclear ND10 structures while simultaneously inducing cytoplasmic aggregation and nuclear accumulation of Hsp70.

In short succession, some of these results were confirmed and given biological significance. McDonald, et. al., also in 2004, demonstrated that the RSV viral polymerase associates with lipid rafts on the cell’s plasma membrane [66]. They accomplished this by isolating lipid rafts via a flotation gradient, and analyzing tryptic peptides via 2-Dimensional nanoflow LC-MS (2D-nLC-MS); this strategy utilized two complementary chromatography methods (Strong Anion Exchange and Reverse-phase Liquid Chromatography) to reduce sample complexity and bypass the low sensitivity of gel-based methods; this maximized the number of proteins that they could identify in their analysis, and enabled them to detect the RSV-L protein, which lacks a suitable antibody for immunoblotting. In 2005, the same group leveraged their 2D-nLC-MS method to show that Hsp70 associated with these raft/polymerase complexes [67]. Furthermore, they demonstrated that anti-Hsp70 antibodies could reduce viral polymerase activity, indicating that the interaction was functional and required for efficient viral replication.

In 2015, this result was further fleshed out by Munday, et. al., who used affinity purification via EGFP-labeled viral proteins, coupled to nanoflow LC-MS, to discover host-cell protein interactors of the viral replication complex [68]. In this AP-MS analysis, they identified numerous heat shock proteins that interacted with the RSV L-protein, including Hsp70 and Hsp90. When they treated their cell culture system with 17-AAG (an Hsp90 inhibitor) they found that the global abundance of cellular RSV proteins and corresponding mRNA plummeted. Finally, they tested the necessity of Hsp70 for viral transcription using an Hsp70 inhibitor and a gel-based cell extract assay, and observed that HSP70 inhibition was required for RSV polymerase activity on an intact nucleocapsid template. These results collectively demonstrated that the RSV infection induces increased expression of Heat Shock Proteins which are in-turn required for viral polymerase activity.

2.2. Elucidation of oxidative stress response mechanisms by subcellular enrichment and MALDI-TOF MS.

RSV replication induces an oxidative stress response in vitro and in vivo, a phenomenon linked to activation of transcription factors important in the innate inflammatory response. In particular, RSV-infection has been shown to inhibit expression of cytoplasmic antioxidants, including Glutathione S-transferase (GST), superoxide dismutase (SOD), and catalase. Despite a deep body of transcriptional regulation, a holistic understanding including post-transcriptional regulation of the viral response has been lacking.

To address this, Jamaluddin, et. al. combined in-solution Isoelectric Focusing (EIF) followed by 2DE and MALDI-TOF peptide mass fingerprinting to examine the A549 nuclear proteome in greater depth and determine its responses to RSV-infection [35]. To enable more accurate quantitation, free cysteine side chains were labeled with fluorescent dyes. In addition to reproducing the results of Brasier, et. al., these authors found that RSV induced shifts to the isoelectric points of several peroxiredoxin (Prdx) isoforms, without changing their apparent abundance. Notably, an siRNA knockdown model demonstrated that cells lacking Prdx-1, Prdx-4, or both, showed increased levels of reactive oxygen species (ROS) and protein carbonylation.

To follow up on this, a novel proteomics labeling technique was developed and applied to covalently link free cysteine side chains with fluorescent dyes to systematically identify proteins sensitive to RSV-induced cysteinyl oxidation in the presence or absence of Prdx-1, Prdx-4, or both. Saturation fluorescence labeling of unoxidized cysteine side chains followed by quantification using 2DE fractionation identified 15 unique proteins, including Lam A, annexin, and centromere E-associated protein that had enhanced oxidative modifications in the Prdx-depleted nuclei. These studies concluded that Prdx-1 and Prdx-4 are essential for preventing oxidative damage in a subset of nuclear intermediate filament and actin binding proteins in RSV-infected epithelial cells.

Identifying the interaction between nuclear oxidant stress response and cytoskeletal changes contributed the following insights to the field: 1). Cytoskeletal reorganization of filamentous actin is required for RSV transcription, assembly and budding. 2). RSV disrupts cellular antioxidant proteins. 3). Subcellular isolation followed by HPLC prefractionation significantly enhanced identification of nuclear processes affected by virus replication.

2.3. The RSV-Nonstructural protein 1 interferes with type I and II interferon expression by suppressing RIG-I activation.

The RSV nonstructural (NS) proteins are small, non-packaged viral proteins that are well-established to inhibit host type I (e.g. IFN-α, IFN-ß) and type III interferon (e.g. IFN-λ) signaling and regulate NF-κB signaling [69–76]. However, their mechanisms of action were incompletely understood until recently. To address this knowledge gap, Hastie and colleagues used 2DE and MALDI-TOF/TOF mass spectrometry to identify the effects of a recombinant, NS1-deficient RSV virus on the whole cell proteome of A549 cells [77]. They found that 22 proteins that were differentially abundant between the NS1-deficient and wild-type viruses, including Superoxide Dismutase-2 (SOD2) and several other known type II IFN-inducible proteins. Further investigation using cytokine-treated A549 cells showed that these same proteins were induced by both type I IFNs and type II IFNs (e.g. IFN-γ), but not TNF-α or IL1ß.

Using a type I IFN-deficient Vero cell model, the investigators went on to show that NS1-deficient RSV-induced changes to SOD2 and other markers could be induced in a type I IFN-independent manner, thereby implicating NS1 in blocking type II IFN regulation. The authors additionally showed that this mechanism was independent of the NRF2 oxidative stress response pathway (see Section 3.1), which could also induce expression of SOD2 alongside catalase and Thioredoxin reductase. The authors speculated that this behavior was consistent with targeting of the RIG PRR by NS1 – an observation that has been borne out by later mechanistic studies [78].

2.4. LC-MS analysis indicates mitochondrial disfunction in ROS production and broadens the understanding of anti-viral responses to RSV infection.

In addition to the 2DE/MALDI-TOF studies of RSV infection, several studies have been conducted using online LC-MS. Unfortunately, most such studies omitted biological and/or technical replicates, making their interpretation difficult in relation to other studies. Nevertheless, some specific results were confirmed using additional methods, such as immunoblotting and confocal microscopy, and are worth acknowledging.

In 2010, Munday, et. al. published two papers employing SILAC to quantify RSV-induced changes to both the cytoplasmic and nuclear proteomes of Hep2 cells [79,80]. In these studies, the authors identified concerted reductions to the abundances of mitochondrial proteins such as the Translocase of the Outer Mitochondrial membrane (TOM) complex (e.g. Tom20, Tom22, Tom40, Tom70), as well as Voltage-dependent Anion Channels (VDACs; e.g. VDAC-1, VDAC2, VDAC-3). These changes were induced by both RSV subtypes, although no direct comparison of the two was conducted. Their findings led the investigators to propose that RSV infection altered the permeability of mitochondrial transition pores; a hypothesis that they then confirmed using a live-cell mitochondrial permeability assay. These results suggested that mitochondrial dysfunction may contribute to RSV-induced oxidative damage. A third study from this group applied SILAC to mitochondrial extracts from RSV-infected A549 cells, and confirmed that infection altered the abundances of TOM proteins [81]. The authors demonstrated that knockdown of Tom70 increased viral titer, further indicating a role for these proteins during viral infection.

Somewhat later, Ternette et. al. contributed to these experiments by combining in-solution IEF with LC-MS and Data-independent Analysis (DIA) more deeply probe the whole-cell proteome of RSV-infected A549 cells [82]. Their data quantified over 1000 proteins, reproduced the findings of many earlier studies, and highlighted the cellular accumulation of the Interferon-induced protein with tetratricopeptide repeats 3 (IFIT3) protein, supporting the activation of the interferon response. They also observed differential isoelectric mobility of the 5’-3’-exoribonuclease 2 (XRN2) protein, which they attributed to post-translational modification.

In 2014, Dave et. al. conducted a rigorous examination of A549 whole cell lysates post-RSV infection, utilizing five biological replicates and contrasting the performance of IEF pre-separation with a fractionation-free workflow prior to LC-MS [83]. This study flagged extensive alterations to IFN-induced proteins, and pathway analysis demonstrated that the vast majority (83%) of cellular proteins with differential regulation were controlled by IFNs, Tumor-necrosis Factor Alpha, and NF-κB. Notably, this work indicated that RSV infection induced an IFN-γ response in A549 cells, despite the prevailing belief that A549 epithelial cells should not produce that cytokine. Finally, this study also demonstrated the degree to which proteomics technologies had advanced over the years; using their fractionation-free workflow, the investigators quantified over 5000 proteins from whole cell lysates and five biological replicates; a staggering improvement to what was possible only a few years before.

2.5. Multiplexed quantitative proteomics measurements of the innate immune response.

It is well-established that RSV infection induces type I and III mucosal interferon (IFN) production with paracrine activation of protective, anti-viral interferon stimulated genes (ISGs) [84–86]. This results in dynamic nuclear translocation and cytoplasmic shuttling of innate transcription factors and regulatory kinases, as well as expression of pathway inhibitors. However, quantification of this response is extremely challenging. First, the major regulatory components of the IIR, consisting of PRRs, kinases, and transcription factors, are all low-abundance proteins [87,88]. Second, few high affinity detection reagents are available that can reliably detect them in complex cellular samples. Third, classical “shotgun” proteomics is limited by stochastic sampling of a fraction of the proteome in a manner that is usually biased toward higher abundance proteins [33]. Consequently, many important low-abundance proteins are usually not consistently identified across samples. Such fragmentary data sets are not satisfactory for understanding the signaling pathway at the systems level.

These problems have illustrated the need for high-accuracy, high- sensitivity multiplex assays for measuring dynamic changes in the IIR. One such approach is the selective reaction monitoring (SRM)-MS assay. SID-SRM-MS assays for 10 major regulators of IIR (including RelA, IRF3, and RIG-I) have been developed and reported [89]. Application of SID-SRM method reproducibly measured the concentrations of these proteins in cytoplasmic and nuclear compartments of A549 cells over a time course of ds-RNA stimulation, revealing differences in translocation kinetics between associated transcription factors. These studies also elucidated, for the first time, the existence of a negative NF-kB-IRF3 cross-talk pathway, and the consumption of the RIG-I/Tank Binding Kinase (TBK1) “signalsome” in IRF3 signaling.

Coupling the SID-SRM method with affinity enrichment using antibodies or single-stranded DNA aptamers results in dramatic enhancement of the assay sensitivity by enhancing the signal-to-noise ratio [90]. Using an ssDNA aptamer (termed P028F4) that binds to the activated (IκBα-dissociated) form of RelA with a K(D) of 6.4 × 10⁻¹⁰, coupled to SID-SRM MS assay was reported. This assay produced a linear response over 1,000 fold dilution range of input protein and had a 200 amol lower limit of quantification, enabling the amount of nucleoplasmic and chromatin associated RelA could be estimated in response to stimulation. The aptamer-SID-SRM-MS assay quantified the fraction of activated RelA in subcellular extracts, detecting the presence of a cytoplasmic RelA reservoir unresponsive to TNFα stimulation.

In a separate study, Tian, et. al. used these SRM and IP-SRM assays to show that RSV-induced inflammation triggers a dynamic interaction between the pro-inflammatory NF-κB RelA subunit with the epigenetic scaffold Bromodomain-containing Protein 4 [29]. Small molecule BRD4 inhibitors (JQ1+) and siRNA depletion of BRD4 further demonstrated that this interaction was necessary for RSV-induced activation of NF-κB dependent cytokines, but was independent of viral replication. Ultimately these results contributed the investigators’ novel finding that BRD4 coupled NF-κB to the IRF/RIG-I anti-viral amplification loop by contributing to the formation of active transcriptional elongation complexes on NF-κB-dependent pro-inflammatory genes.

2.6. Dynamic Protein Interaction monitoring of RSV regulated epigenetic scaffolds by PASEF-MS

In addition to mediating RSV-induced inflammation, the BRD4 epigenetic scaffold is also associated with the development of airway remodeling and fibrosis, and bromodomain inhibitors prevent downstream airway remodeling in murine models [19,91–95]. However, the mechanism of this action is currently unclear.

In a recent study, Morgan Mann et al. examined the dynamic interactome of BRD4 in an attempt to provide context for this phenomenon [96]. They tested their hypothesis that RSV-infection would trigger BRD4 to interact with transcription factors that regulate inflammation and remodeling by examining the makeup of immunopurified BRD4 complexes isolated from A549 cells infected with RSV and treated with a specific BRD4 inhibitor (ZL 0454). To facilitate deep proteomic profiling despite the low cellular abundance of BRD4, the authors employed Parallel Accumulation – Serial Fragmentation MS (PASEF-MS), which utilizes Trapped Ion Mobility separation (TIMS) as an additional mechanism to reduce sample complexity prior to MS-analysis. Additionally, the TIMS dimension increases sensitivity and throughput by focusing ion signals and excluding non-peptide contaminants from fragmentation and sequencing.

In this study, the investigators identified 557 potential BRD4 interactors, and found that a staggering 305 were responsive to RSV-infection, of which 272 demonstrated increased relative abundance on the BRD4 complex post-infection; 95 of these proteins were also disrupted from the RSV-activated complex by the BRD4 inhibitor (Figure 3). Interestingly, these proteins were highly enriched in transcriptional regulators, including members of the AP1 transcription factor complex (i.e. c-JUN, AP1), and the Wnt-signaling pathway (i.e. β-Catenin, γ-Catenin). AP1 is a transcription factor complex involved in cell survival, apoptosis, and inflammation [97,98]; critically, it also contributes to the expression of Interleukin-6 – a key pro-inflammatory cytokine induced by RSV-infection [99,100]. Similarly, the Catenin proteins are transcription factors that are highly correlated with activation of the epithelial-to-mesenchymal transition (EMT) and are linked to asthmatic airway remodeling [101–106]. Furthermore, treatment with the BRD4 inhibitor resulted in the partial eviction of c-JUN and the complete eviction of both β-Catenin and γ-Catenin from the BRD4 complex. These results indicated that RSV-infection triggers interactions between BRD4 and pro-inflammation/remodeling transcription factors and suggested the disruption of those interactions was responsible for the observed effects of BRD4 inhibitors in vivo. In a broader sense, this experiment also perfectly demonstrated the unique capability of modern proteomics technologies to analyze entire networks of proteins in a single experiment.

Figure 3. RSV-infection induces dynamic and bromodomain-dependent interactions to the host cellular protein BRD4.

Cytoscape network visualization of RSV-inducible and bromodomain-dependent BRD4 interactors. Node color represents the RSV-inducible Log₂ fold change. Edge color is keyed to the STRING score and represents interaction confidence. Low confidence interactions (STRING score < 0.4) and nodes without interactions are omitted. Reproduced with permission from [96].

3. Proteomics studies conducted in alternative models of RSV-infection.

Owing to the experimental control of the timing and extent of infection, in-vitro cell culture systems – especially A549 alveolar epithelial cells - have dominated the field of RSV-proteomics [34,35,66–68,77,82,83,107]. Nevertheless, RSV-infection is a tissue-level event, affecting numerous cell types that can act in concert, and producing marked changes to the extracellular matrix and extracellular fluids [17,30]. Accordingly, investigations focusing on models outside of the petri dish have significant value, and have produced notable data contributing to our understanding of RSV pathogenesis.

3.1. RSV depletion of antioxidant capacity is linked to disease pathogenesis.

In 2011, Hosakote, et. al. continued the work started by Jamaluddin, et. al in 2010 by demonstrating that RSV infection induced a significant decrease in the expression and/or activity of SOD, catalase, GST, and glutathione peroxidase in murine lungs and in the airways of children with severe bronchiolitis [64]. The mechanism of RSV-induced reductions to the expression of these antioxidant genes was determined by mechanistic studies that demonstrated RSV induced degradation of transcription factor Nuclear Factor, Erythroid 2 Like 2 (NRF2), a transcription factor that normally maintains expression of a cluster of antioxidant enzymes including catalase [108]. Catalase is a key enzyme for the dismutation of virus-mediated generation of hydrogen peroxide (H₂O₂).

To test whether this disruption of this antioxidant pathway was relevant to disease progression, SOD mimetics and stabilized polyethylene glycol-conjugated catalase were administered in an experimental rodent model of RSV infection [109,110]. These manipulations resulted in reduced viral H₂O₂ production and produced a significant protective effect on RSV-induced inflammation, clinical disease and airway pathology. Collectively, these studies indicated that strategies to augment the host antioxidant pathway may be an important strategy to reduce manifestations of RSV infection.

3.2. Tissue proteomics indicates that RSV Vaccine-related eosinophilia acts through Chil3 and Integrin alpha-M.

In 2015, van Diepen, et. al. turned their attention to the issue of RSV-vaccine-mediated enhanced disease (VMED) [111]. They measured host proteome correlates of VMED by examining the lung proteome of mice challenged with RSV and primed with vaccine preparations based on the RSV F and G-proteins. These vaccine preparations were based on the recombinant vaccinia virus (rVV), and were known to produce phenotypically different host immune responses corresponding to VMED [112,113]. For this model, it had been shown that the rVV-G preparation resulted in the production of neutralizing antibodies and a subsequent reduction to RSV replication, but also produced severe illness and pulmonary eosinophilia. In contrast, the rVV-F preparation resulted in a more effective host defense without pulmonary eosinophilia.

When the lung tissue of the mice was harvested and analyzed, Diepen, et. al. found that rVV-G-treated mice had significantly greater levels of Chitinase-like-protein 3 (Chil3) and Integrin alpha-M (Itgam). Chil3 is a secretory protein most often produced by macrophages and neutrophils, and has chemotactic properties for T lymphocytes, bone marrow polymorphonuclear lymphocytes, and eosinophils [114,115]. Importantly, Chil3 has been demonstrated to trigger extravasation of eosinophils, and mRNA knockdown of Chil3 has been shown to block eosinophilia in a murine model of airway hyperreactiveness [116]. Similarly, Itgam is a component of the Complement Receptor Type 3 (CR3) protein, and facilitates adhesion of neutrophils and eosinophils [117]. This data suggested that Chil3 and Itgam were mechanistically responsible for the high level of eosinophilia observed in the lungs of rVV-G-treated mice.

3.3. GBUP demonstrates the presence of anti-viral and pro-ROS species in BAL fluid and nasopharyngeal aspirates.

In regard to extracellular proteomics, two recent studies have examined the proteome of bovine bronchoalveolar lavage fluid (BALF) and human nasopharyngeal aspirates, respectively. The bovine study, conducted by Häggland, et. al. used nLC-MS to examine the effect of RSV-infection on gauze-filtered BALF from RSV-vaccinated and unvaccinated cows [118]. Between all comparisons, they identified 96 proteins that were associated with the bovine RSV-disease, mostly associated with neutrophil activation and chemotaxis (e.g. Resistin, Elastase, S100-A9) and homeostatic responses to ROS (e.g. SOD, Prdx-6). Markers of lymphocyte and natural killer cell activation/chemotaxis were also enriched by RSV-infection, alongside epithelial cell markers.

The major findings in this study were that lung pathology and histological scoring correlated heavily with increased abundance of neutrophil activation markers and reduced abundance of antioxidants. A secondary finding was that bovine RSV was negatively correlated with Leucine Zipper TFL1, which has been shown to inhibit the epithelial-to-mesenchymal transition (EMT) [119] – potentially establishing a link to the inflammation-associated phenomena of airway remodeling.

Along a different angle, Aljabr, et. al., examined the proteome of nasopharyngeal aspirates from RSV-infected children, and used a combined proteomics and transcriptomics approach to characterize the disease response for potential biomarker development [120]. As has become standard, nLC-MS was applied to profile tryptic peptides derived from the samples, and the group found significant increases to numerous proteins corresponding to interferon-stimulated genes (e.g. ISG15, ICAM1, IFIT1, OAS2), and the Bactericidal/Permeability-Increasing (BPI) proteins, which have been shown to have anti-viral properties in the context of influenza [121].

3.4. Secretome Proteomics reveals cell-type differences in paracrine signaling proteins.

Recent studies have illuminated the findings that airway epithelial cells are derived from multiple lineages and are functionally distinct in their viral responses [122]. Gene-profiling experiments have suggested that lower airway epithelial cells produce greater amounts of T helper type 2 (Th2)-activating CCL-type chemokines than do epithelial cells of the conducting airways [15,123]. Because cellular proteomics undersample secreted proteins and factors Zhao et al. applied a highly sensitive unbiased secretome profiling technique to identify free and membrane-bound exosomes [17]. After developing the workflow and analysis pipeline in immortalized small airway epithelial cells, the profiling was conducted in primary cultures of human airway epithelial cells.

Approximately 1,000 high confidence protein identifications in cell supernatants at a 1% false discovery rate. These common proteins were derived from lysosomal, cytoplasmic, and nuclear compartments. Approximately one third of secretome proteins were exosomal; the remainder were from lysosomal and vacuolar compartments. These investigators also identified 577 differentially expressed proteins (by cell type) from control supernatants and 966 differentially expressed proteins from RSV-infected. Interestingly, many of these secreted proteins do not contain classic signal peptides, suggesting novel and incompletely understood mechanisms controlling innate signaling. For example, the nuclear damage-associated molecular patterns (DAMPs) HMGB1 and histone H3 undergo nuclear export and extracellular secretion in response to RSV; these DAMPs control mononuclear inflammation [18].

Focusing informatics analysis on the 103 proteins secreted by lower airway bronchiolar cells, Th2 activating cytokines (MIP1, TSLP), mucin expressing (CCL20), and fibrogenic cytokines (IL6) were identified (Figure 4). Importantly, all of these factors are dependent on NFκB, further implicating NFκB in immunopathogenesis and remodeling in RSV lower respiratory tract infection. This data underscores the cell type influences on paracrine signaling by the IIR.

Figure 4. Identification of differentially secreted proteins in response to RSV infection.

(A). Statistical analysis of microarray (SAM) for secretome proteins whose expression differs by cell type in the basal state. X axis, expectation score; Y axis, observation score. The diagonal line shows where false discovery rate = 0.01. Points above (in red) outside the threshold are those with FDR < 0.01, and points below the threshold (in green) are those proteins with FDR > 0.01. Proteins with increased expression in hSAECs are indicated by red points; those decreased are indicated in green. (B). SAM for secretome proteins whose expression di ers due to RSV infection. (C). Expression values of proteins were z-score-normalized data of log2-transformed expression values for biological replicates. Two-dimensional hierarchical clustering was performed, with columns representing cell samples and rows representing individual proteins (green, low expression, red, high expression). Cluster 3 is unique to the bronchiolar-derived human small airway epithelial cells (HSAECs). HSAECs differentially express CCL20, TSLP, IL6, and MIP1. Reproduced with permission from [17].

4.0. Underutilized and unexplored proteomics approaches for RSV-research.

Thus far, we have discussed proteomics technologies and approaches that have been well applied to RSV-research. However, the field of proteomics is fast-moving, and several new technologies and recent advances merit discussion here for their potential applications towards RSV-biology. In this section, we discuss Single-cell proteomics, extracellular matrix proteomics, and Top-down mass spectrometry.

4.1. Single-cell proteomics.

GBUP approaches have provided detailed overviews of RSV-induced changes to cellular protein abundances. However, these measurements represent bulk protein, averaged over millions of cells [124]. Even in cell culture systems with identical genomic backgrounds, stochastic influences can result in significant heterogeneity at the transcript and protein level, and there is a growing consensus that this heterogeneity is an important metric for understanding biological systems.

Historically, proteomics techniques have lagged behind similar genomic and transcriptomic methods [125], and while single-cell sequencing has been available for many years, single-cell proteomics has yet to reach its stride. Nevertheless, rapid advances have been made in the field, aided by improvements in sample preparation, LC-separation, and the application of isobaric labeling and quantitation techniques [126,127]. At present, the state of the art is capable of identifying and quantifying over 1500 proteins from single HeLa cells [128], which already eclipse the number of proteins quantified in early RSV proteomics studies. Notably, Hela cells have similar total protein content to A549 cells, opening the door to analysis of airway biological conditions, such as RSV-infection [129,130]. Further technical advances in the next several years should be anticipated to increase sensitivity even more, potentially allowing analysis of primary cells and stem cells.

While single-cell proteomics is certainly out of reach to most laboratories, the findings so far emphasize the importance of these measurements. A recent preprint from the Matthias Mann group compared single cell transcriptomics to their recently developed single cell proteomics strategy, and found significant disagreement between the two methods [128]; in general, the proteome was far more stable than the transcriptome. Given that viral infection modulates both transcriptional activity and protein abundance, both types of information will likely be required to decipher the actual effects of RSV-infection at single-cell resolution. This will be especially important for investigations into the proteomic landscape of partially infected cells and tissues, as opposed to current cell culture models that use very high MOIs.

4.2. FFPE Proteomics

Formalin-fixed Parrafin-embedded (FFPE) tissue is commonly collected in both clinical and research settings to observe pathological changes to tissue [131]. Owing to the high stability of parrafin-embedded tissue and its ease of storage, significant archives of FFPE tissue have accumulated in the past century. Nevertheless, these samples were historically considered inaccessible to proteomic analysis, owing to the molecular crosslinking induced by formaldehyde. However, this assumption has proven false in the last decade [132,133], and numerous groups since then have developed sophisticated extraction techniques to perform high-throughput characterization of FFPE tissue sections [134–136].

In the context of RSV biology, FFPE lung sections have been used in mouse models to characterize the extent of virus-mediated damage and airway remodeling [137], and human samples are sometimes obtained during autopsy [138,139]. These samples represent a rich, but as-of-yet untapped source of data regarding the mechanisms of RSV-mediated remodeling and pathogenesis. Furthermore, in combination with laser-capture microdissection - which has been coupled to FFPE proteomic analysis [132,140] - it is possible to isolate specific cell populations for analysis, and probe differences between tissues with high- and low-grade pathology within a single tissue sample. This development opens the door for sophisticated and high-power experiments to characterize how cellular heterogeneity contributes to RSV pathogenesis.

4.3. Proteomics of extracellular matrices.

Chronic RSV-infection leads to irreversible, fibrotic changes to airway structures and extracellular matrices (ECM), collectively referred to as airway remodeling [21]. However, measuring ECM proteins via proteomics is far more difficult than measuring cellular or secreted proteins, as the various ECM-components form complex macromolecular structures with low solubility in aqueous solution [141–143]. In addition, many have proven resistant to typical proteolytic digestion strategies employed [144]. Meanwhile, the workflows developed to alleviate these issues are labor intensive compared to cellular GBUP experiments [43–45]. This has resulted in an unfortunate knowledge gap, and to our knowledge, no RSV-oriented ECM proteomics study has been published.

Fortunately, recent advances in surfactant chemistries and sample preparation methodologies show promise in alleviating these issues. In 2020, Knott, et. al. published a streamlined and improved ECM proteomics strategy employing a photocleavable surfactant (4-Hexylphenylazosulfonate) [46]. Using this strategy in combination with 1- and 2-dimensional liquid chromatography, they examined the ECM proteome of mouse mammary tumors, and identified multiple novel PTMs. Importantly, they rigorously established the reproducibility of the method to ensure applicability in biological contexts.

4.4. Post-translational Modifications and Top-down mass spectrometry.

Protein post-translational modifications provide an additional degree of complexity to the discussion of RSV biology, as phosphorylation, ubiquitination, and numerous other PTMs can modulate the function and abundance of key signaling molecules [145]. While the bottom-up proteomics studies described thus far have generally provided useful quantitative data regarding protein abundance, the same cannot be said regarding RSV-relevant post-translational modifications. This is especially glaring given that several studies indirectly detected RSV-induced changes to PTMs via altered protein isoelectric points, and yet the topic has seen no additional attention.

This deficiency most likely reflects the generally low cellular abundance of post-translationally modified proteins compared to their unmodified variants, and the chemical difficulties involved in preserving them throughout long sample-handling and LC-MS workflows. The most commonly applied fragmentation method for peptide sequencing further complicates the matter, as Collision-induced Dissociation (CID) can remove labile PTMs, including phosphorylation, from peptides in a neutral-loss event that is not always obvious in a tandem mass spectrum [146]. In studies where phosphopeptides are expected, the use of Electron-capture or Electon-transfer Fragmentation is recommended (ETD & ETF).

An additional avenue worth exploring is the application of intact protein mass spectrometry to the problem. Intact protein MS, or “Top-down” MS contrasts with the usual “bottom-up” approach in that proteins are not digested with a protease prior to analysis [147,148]. This often requires different extraction techniques, chromatography, and MS-instrumentation, but can simplify sample preparation workflows and preserve sensitive post-translational modifications [149] . In addition, modified protein variants (“proteoforms”) are chemically similar, which enables direct quantitation of PTM stoichiometry. Accordingly, top-down mass spectrometry has been applied to deeply characterize the post-translational modifications of numerous proteins, including muscle tissue proteins [150,151], phosphoproteins [152], and glycoproteins [153–155]. For identical reasons, Top-down may also prove useful in examining proteins which are difficult to detect in bottom-up workflows, such as the RSV-G protein.

5.0. Expert Opinion

RSV alters cell biology through complex and multi-faceted mechanisms, with the common phenotypes representing a convergence of numerous factors and signaling pathways. In the face of such complexity, traditional biochemical methods for protein quantitation are ill-suited to make new discoveries, due to their low sensitivity and throughput. MS-based proteomics technologies, in contrast, have proven highly useful in filling in the blanks of RSV-biology. Over the course of nearly twenty years, RSV-oriented proteomics studies have cast a wide net over the cellular protein landscape, resulting in significant findings related to RSV viral replication and transcription. Additional studies have also shed light on host protein correlates of the disease, implicating neutrophils and oxidative damage as key drivers of the disease.

Over this period, MS and related technologies have advanced rapidly, progressing from crude gel-based methods with relatively low sensitivity and throughput to sophisticated unions of mass spectrometry, liquid chromatography, and gas-phase separation, that maximize the proteomic depth and coverage of any given analysis. We expect that this pace will continue, incorporating new technologies and methods such as single-cell and top-down proteomics.

5.1. Strengths

Compared to traditional biochemical approaches for protein quantitation, MS-based proteomics methodologies are incredibly specific, sensitive, and high-throughput. This enables effective quantitation of low-abundance transcription factors and cellular regulators, as well as proteins (e.g. RSV-L) that do not have specific antibodies available for immunodetection. Furthermore, the multiplexed measurements obtained lend themselves well to bioinformatics analysis, enabling recognition of entire biological pathways and protein complexes that are regulated during RSV infection. At present, this combination of advantages is entirely unique to MS-based proteomics.

MS analysis can also serve as a detection endpoint for numerous biological assays; bottom-up proteomics has been robustly applied to global measurement of both cellular and extracellular protein abundances, with interactome analysis also facilitated by affinity purification. Many post-translational modifications can also be preserved and quantified in an unbiased proteomic analysis.

5.2. Weaknesses

Although the mass spectrometer can be coupled to most sample preparation workflows, this often requires additional, labor-intensive steps. While automation and recent developments in sample preparation methodologies have made great progress in overcoming these difficulties, the additional workload may still be intimidating to some researchers.

In addition, the bottom-up proteomics workflows covered by this review are often insufficient to comprehensively study post-translational modifications, owing to their sometimes-uncooperative chemical properties and typically low relative abundance compared to unmodified peptides. While top-down mass spectrometry offers a powerful tool to circumvent the usual limitations, it is technically complicated and low throughput in comparison to bottom-up analysis. Furthermore, it may be overwhelmed by highly modified and diverse proteoforms.

Although many of the cellular models of RSV infection have advanced understanding of host responses to RSV infection used alveolar epithelial-like cells, such as A549, applications of high resolution and sensitive proteomics to primary cells and organoids will lead to additional discoveries from these more differentiated systems.

Finally, the cellular studies reviewed here were performed at high MOIs, an experimental design intentionally to reveal response to viral replication. Little is understood about the paracrine response of uninfected cells in the airways. Single cell proteomic designs using low MOIs will be illuminating studies.

5.3. Potential Opportunities

Immediate research opportunities largely represent applications of new technologies and approaches to older problems. PASEF and similar ion mobility-mass spectrometry workflows facilitates a massive increase to sensitivity and proteome coverage and may enable high-throughput quantitation of RSV-relevant proteins with low sample abundances, or from samples derived from primary cells and other protein sources traditionally avoided due to high cost or low protein content. Meanwhile, laser microdissection and FFPE proteomics workflows may enable nuanced analysis of the effect of tissue and cell-type heterogeneity on RSV-pathogenesis in-vivo.

Additionally, no study has directly compared the effects of the RSV subgroups or different genotypes within a subgroup on the cellular proteome. It is speculated that genomic differences contribute to altered virulence between strains, and proteomic changes may offer insights into the drivers of increased virulence.

5.4. Five-year view.

Over the next five years, we anticipate that single-cell proteomics studies will become sufficiently comprehensive and applicable to tissue to allow unbiased analysis of RSV-induced protein changes with cell-type resolution, and dissection of effects of direct replication vs paracrine signaling. ECM proteomics will be highly complementary to these studies and allow analysis of changes to airway structure and fibrosis. Finally, as top-down proteomics becomes more common and approachable, more comprehensive identification of changes in phosphorylation, oxidation, s-nitrosylation and other pathways will be illuminating.

Collectively, we believe that these developments will enable future studies to deconvolute the complex web of inter-cell signaling present in-vivo RSV-infections. The prospect of more completely deciphering the mechanisms of RSV-induced airway remodeling, for example, is extremely exciting, and appears to be on the horizon. Vaccine development will also benefit, as top-down MS facilitates deeper analysis of RSV-proteins and their relevant post-translational modifications

Article Highlights:

• Unbiased proteomic profiling studies have discovered viral replication effects on nuclear dot 10 (ND10) structures, heat shock responses and perturbations of anti-oxidant defenses important in disease pathogenesis. These post-translationally regulated homeostatic responses are invisible to gene expression profiling.

• Affinity purification-MS enables focused studies of dynamic changes in protein-protein interaction networks controlling innate immune response (IIR).

• Advances selective reaction monitoring to profile of dynamic changes in abundance and subcellular distribution of IIR regulators opens opportunities for understanding how RSV subverts anti-viral responses.

• Emerging studies implicate RSV replication has substantial effects the formation and remodeling of the extracellular matrix (ECM). New developments in photocleavable Azo dyes are described with the potential to unlock discoveries in ECM remodeling.

• Advances in proteomics, including top down and single-cell proteomics will advance understanding of post-translational modifications and enable dissection of direct-versus-paracrine effects of viral replication.