Comparative proteomics of human naso-oropharyngeal swabs from COVID19 patients for identification of potential biomarker candidates Authors: Sabyasachi Bandyopadhyay, Megha Brijwal, Arhaan Haridas, Aashish Choudhary, Punit Kaur, Lalit Dar, Gururao Hariprasad Published: 2026 ABSTRACT AbstractCOVID19 continues to be a public health issue with subvariants of the main strains causing infections globally. This study looks to identify potential biomarker candidates that can help to understand molecular etio-pathogenesis. Proteins from naso-oropharyngeal samples of healthy Control and COVID19 subjects were trypsin digested and subjected to liquid chromatography assisted mass spectrometry analysis on Orbitrap Fusion Lumos (ThermoFisher). Proteomic analysis was carried out using Proteome Discoverer 2.4 (ThermoFisher) using human and SARS-CoV-2 FASTA databases from UniProtKB. Proteins with fold Change ≤0.5, or ≥2 were designated as differentially expressed and taken up for interactomic analysis using STRING and Cytoscape, and pathway analysis using KEGG, Reactome, and Wiki pathway. Area under curve for receiver operating characteristic was done in MetaboAnalyst (v6.0) to estimate diagnostic parameters of sensitivity and specificity. Proteomic analysis of naso-oropharyngeal swabs from COVID19 patients identified 26 differentially expressed human proteins of which 11 are potential biomarker candidate proteins. In addition, 29 SARS-CoV-2 specific peptides from spike protein, nucleocapsid, membrane, ORF1ab, ORF6, ORF7a, ORF8 and ORF9b proteins have been identified of which 6 were mutated. These SARS-CoV-2 proteins make an array of interactions with the identified human proteins establishing their roles in pathogenesis of viral infection. Human biomarker candidate proteins are associated with alterations in glycolytic pathways and immune dysregulation, which are known to facilitate viral invasion, replication, and contribute to disease severity. Identified biomarker candidates have a minimum of 70% and a maximum of 100%, sensitivity and specificity thereby having translational value in developing protein-based tests that can complement currently available diagnostics for COVID19.Keywords: LC-MS/MS proteomics, COVID19, Biomarkers, Glycolysis and immunity dysregulation, SARS-CoV-2 peptides Introduction Coronavirus disease 2019 (COVID19) is a pandemic caused by the novel Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) that has affected millions of people worldwide since December, 2019 [1]. While the majority of patients experienced mild to moderate symptoms such as fever, body aches, and sore throat, a significant number developed severe complications including pneumonia, sepsis, respiratory failure, acute respiratory distress syndrome (ARDS), and multiorgan dysfunction [2,3]. High morbidity and mortality, particularly among the elderly and individuals with comorbidities such as diabetes, hypertension, and chronic organ diseases, posed major challenges to patient management and treatment [[1], [2], [3], [4]]. COVID19 can mimic a variety of viral and bacterial infections, including influenza, dengue, other viral haemorrhagic fevers, bacterial pneumonia, and tuberculosis as they share common symptoms [5]. Owing to the varied clinical presentation, it also resembles certain autoimmune and inflammatory disorders such as systemic lupus erythematosus, Kawasaki disease, vasculitis, and multisystem inflammatory syndrome [6]. In addition, COVID19 affects different organ systems thereby presenting with organ-specific disorders such as myocardial infarction, myocarditis, acute coronary syndromes, pulmonary embolism, gastrointestinal, and neurological disorders making differential diagnosis challenging [3,4]. Burden of the disease therefore makes it essential to map the molecular events at the site of infection to have a better understanding of virus-host interaction in causing infection. The gold standard for detecting active infection is the reverse transcription polymerase chain reaction (RT-PCR) test [7]. However, its sensitivity varies between 60% and 90%, depending on the timing of the test, sample type, and quality of sample collection [8,9]. This variability can result in false-negative outcomes, especially in cases with low viral loads or improper sample handling. In contrast, serology or antibody tests are used to detect past infection or immune response to SRAS-CoV-2. These tests typically measure, IgM (appears within 5 to 10 days of infection) and IgG (appears after 10 days of infection) or combined IgM/IgG antibodies that suggests late stage or recovery from infection [8,9]. Optimal sensitivity and specificity of these serology-based assays generally range from 70% to 90% after 14 days of symptom onset, however, results are susceptible to false positives due to cross-reactivity with antibodies from other respiratory or influenza viruses [8,9]. Molecular pathogenesis involves SARS-CoV-2 invading the host through interaction of viral Spike (S) protein with Angiotensin Converting Enzyme-2 (ACE-2) receptor on host cells present in naso-oropharyngeal cells and lungs [10]. Proteases, particularly TMPRSS2 and cathepsins, cleave and activate the S protein, facilitating viral and cellular membrane fusion releasing its positive-sense RNA genome into the cytoplasm [11]. While in some cases, SARS-CoV-2 can delay or suppress type I interferon responses, allowing the virus to replicate efficiently in the early stages of infection [12]. In a few severe cases, uncontrolled cytokine release, known as a cytokine storm leads to systemic inflammation, acute respiratory distress syndrome (ARDS), and multi-organ failure [6,13]. Proteomics has emerged as a powerful tool in understanding the molecular and pathophysiological mechanisms to unravel host–virus interactions, immune responses, and disease progression of COVID19 [3,14]. Liquid Chromatography assisted tandem Mass Spectrometry (LC-MS/MS) based proteomics has been extensively used for biomarker identification and validation, drug repurposing efforts, vaccine development, identification of therapeutic targets by uncovering viral and host protein interactions [[15], [16], [17]]. Targeted assays LC-MS/MS have been employed to quantify viral and host peptides with high sensitivity [18]. Mass spectrometry has also been extensively used to characterize post-translational modifications (PTMs) of SARS-CoV-2 proteins, in particular, site-specific glycan profiling of the spike glycoprotein [19]. Integrated analyses of the phosphoproteome, glycoproteome, and ubiquitinome has provided critical insights into complex virus and host interactions and mechanisms of pathogenesis [20]. Furthermore, automated sample preparation combined with high-speed liquid chromatography enables high-throughput MS has enabled analysis of hundreds of clinical samples per day [21]. In this study, comparative proteomics of naso-oropharyngeal swabs has been used as a platform to identify differentially expressed proteins in COVID19. This paves the way to understand molecular pathways implicated in the pathogenesis at the site of infection. Materials and methods Ethics, patient screening, and sample collection The study received ethical approval from the Institute Ethics Committee, All India Institute of Medical Sciences, New Delhi, India (IEC-558/19.06.2020). Methodology followed in this study is provided as an illustration (Fig. 1). Detailed clinical history was taken, examination was conducted, and pharmacological interventions were noted. For screening COVID19 patients Naso-oropharyngeal samples were collected in viral transport medium (VTM, HiMedia, Mumbai, India) and Reverse transcription polymerase chain reaction (RT-PCR) tests were conducted at the virology laboratory. RT-PCR for SARS-CoV-2 targeting ORF1ab and Envelope genes (COVIDsure Multiplex Real Time RT-PCR kit, Trivitron Healthcare Pvt. Ltd., Chennai, India) was performed following RNA extraction and cycle threshold (CT) values were noted. Human RNAse P gene served as an internal control. Subjects who tested negative for RT-PCR were designated as Healthy controls while subjects with CT value of ≤22 and with either severe pneumonia or acute respiratory distress syndrome (ARDS) were designated as COVID19 patients. For this study, 2 ml of VTM was collected from each 10 Healthy Control (HC) subjects and 10 COVID19 patients. All samples collected from 20 subjects were partially inactivated by sterilizing at 56°C for 30 min in the biosafety cabinet in virology laboratory of Department of Microbiology, AIIMS, New Delhi and aliquoted equally in two microcentrifuge tubes (MCTs) and stored at −80°C until further use [22]. Overview of the methodology followed in this study. Details of reagents and solvents used Ammonium bicarbonate (solid, purity ≥99%), Urea (solid, purity >99%), Dithiothreitol (solid, purity ≥99%), Iodoacetamide (crystalline, purity ≥99%) and Benzonase Nuclease (Purity >99%) were purchased from Sigma Aldrich – Merck, India and used without any further purification. Mass spectrometry grade Trypsin (lyophilized powder) was purchased from Thermo Fisher Scientific, USA. HPLC and LCMS grade solvents were either purchased from Sigma Aldrich – Merck, India or Thermo Fisher Scientific, USA. Unless otherwise mentioned all reactions were performed at room temperature. Protein extraction, estimation and trypsin digestion For protein extraction experiment one MCT (containing 1 ml of VTM) was taken out from −80°C freezer and the viruses were inactivated completely by heating at 90°C for 15 min [22]. Samples were then treated with 50 mM ammonium bicarbonate buffer (pH7.8) containing 0.5% ProteaseMax (Promega, USA) and 8 M urea. Furthermore, 250 U of Nucleases were added to solubilize the proteins trapped in the mucous membrane. Proteins were estimated using nano-drop technique in BioTek Cytation 1 cell imaging multimode reader (Agilent, USA) at a wavelength of 280 nm. 100 μg of protein was incubated with 10 mM Dithiothreitol (DTT) at 60°C for 30 min, followed by treatment with 50 mM Iodoacetamide (IAA) at room temperature in the dark for 30 min. The solution was diluted with 50 mM Ammonium Bicarbonate (pH ∼8.0) to reduce the urea concentration below 1 M and digested using Trypsin proteolytic enzyme (1:25 Trypsin to protein ratio) by incubating at 370C for overnight. The digestion was quenched by 1% trifluoroacetic acid (TFA) treatment and the resultant peptide mixture was desalted using C18 Reverse Phase (RP) column (Pierce, Thermo Scientific, USA). The purified peptide mixture was concentrated in Labconco CentriVap Benchtop Vacuum Concentrators (Labconco, USA) and reconstituted in 0.1% FA containing LC-MS grade water (loading buffer). Liquid chromatography assisted tandem mass spectrometry (LC-MS/MS) LC-MS/MS acquisition was carried out on an Orbitrap Fusion Lumos mass spectrometer coupled with an Easy-nLC 1200 nano-flow LC system (both Thermo Scientific). 1 μg of peptides were loaded onto a C18 trap column (length, diameter, particle size, pore size; 2 cm × 75 μm × 3 μm, 100 Å, respectively; Thermo Scientific) and then separated in a C18 analytical column (length, diameter, particle size, pore size; 15 cm × 75 μm × 3 μm × 100 Å, respectively; Thermo Scientific, USA) and injected through a glass spray tip in Nano-Flex ion source (Thermo Scientific, USA). Mobile phase A (0.1% formic acid, loading buffer) and mobile phase B (80% acetonitrile, 0.1% formic acid, elution buffer) were used to establish a 130 min separation gradient (3 min - 5% B; 120 min - 45% B; 123 min - 90% B; 125 min - 90% B; 130 min - 5% B) with 300 nL/min flow rate. For the analysis in data-dependent acquisition (DDA) mode, each scan cycle consisted of one full-scan mass spectrum in Orbitrap analyzer, MS1 (R = 120 K, AGC = 3e6, max IT = 50 ms, scan range of 350 – 1800 m/z) followed by 20 MS/MS events in Linear Ion Trap (Normalized AGC target = 200%, max IT = 50 ms and data type = centroid). High energy collision induced dissociation (HCD) energy was set to 30 to generate fragment ions. Isolation window for precursor selection was set to 1.2 Da and dynamic exclusion was employed for 60 s excluding all charge states for a given precursor. During the LC-MS/MS experiment loading buffer and 500 ng of HeLa digest were included in between the experimental runs to wash the column and to check the QC/QA over the run, respectively. Raw files generated from LC-MS/MS experiment were analyzed in Proteome Discoverer (V2.4, Thermo Scientific, USA) software. Proteins were searched in Sequest algorithm against human proteome database contained 20,430 reviewed proteins and SARS-CoV-2 database contained 17 reviewed and 7024 unreviewed proteins. The databases were downloaded from UniprotKB (https://www.uniprot.org/) [23]. Peptide Spectrum Matches (PSMs) were searched in ‘Sequest’ algorithm with full tryptic specificity, allowing maximum of 2 missed cleavages. Precursor and fragment mass tolerances were set at 10 ppm and 0.6 Da, respectively. Carbamidomethylation of cysteine was set as fixed modification, and oxidation of methionine, and N-terminal acetylation were considered as variable modifications during database search. For human proteins the searched peptides were validated in ‘Percolator’ validation node with false discovery rate (FDR) of 0.01 for peptide spectrum match (PSM) and protein identification. Identified peptides were quantified by Label Free Quantification (LFQ) approach where peptide intensity was measured in “Precursor Ions Quantifier” node which was followed by protein quantification considering both unique and razor peptides. Filter criteria of ‘two unique peptides’ and ‘found in all samples’ were applied to identify the proteins. For SARS-CoV-2 proteins the searched peptides were validated in ‘Fixed value PSM validator’ node keeping other parameters same. Bioinformatics analysis Human proteins having at least two unique peptides and present in all samples were considered and protein abundance data was exported as CSV to conduct statistical analysis in MetaboAnalyst (V6.0; https://www.metaboanalyst.ca/) [24]. The samples were median normalized, Log (base 10) transformed and outliers (if any) were identified by Principal Component Analysis (PCA) with 95% confidence limit. Up- and down-regulated proteins were mapped on a volcano plot by applying p-value threshold of ≤0.05 using Student's t-test approach and Fold Change (FC) threshold of 0.5 ≥ FC ≥ 2. Fold Change (FC) was calculated dividing normalized protein abundance of ten COVID19 patients with normalized protein abundance of ten HC. Differential protein expression was visualized in heatmap with enrichment score ranging from +2 to −2. Variable in Importance Projection (VIP) was obtained for individual proteins that had VIP scores of ≥1. Area Under the Curve of classical univariate Receiver Operating Characteristic (AUC-ROC) analysis was generated for individual proteins by plotting True Positive Rate against False Positive Rate with 95% confidence intervals (CIs) with threshold criteria of 0.7. Identification of SARS-CoV-2 specific peptides SARS-CoV-2 peptides found with abundance value more than 15,000 in at least 5 COVID19 subjects, and either completely absent or present with abundance value ≤ 3000 in HC were manually curated and exported in Excel format. In addition to the endogenous SARS-CoV-2 peptides, mutated peptides were also identified using unreviewed UniprotKB database. The mutation sites were confirmed by aligning the mutated sequences with the wild type sequences (NCBI sequence id – NC_045512.2) in Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo) [25]. All the SARS-CoV-2 peptides were verified by performing BLAST in UniprotKB (https://www.uniprot.org/blast) to confirm the sequence identity [23]. Protein-protein interactome (PPI) and functional enrichment analyses Differentially expressed proteins were studied in Search Tool for the Retrieval of Interacting Genes/Proteins (STRING, version 11.5, https://string-db.org) [26] database for protein-protein interactomes (PPIs) using homo sapiens as background organism. This was followed by functional enrichment analyses, such as, Gene Ontology for Biological Processes (GO_BP, https://www.geneontology.org/), Kyoto Encyclopaedia of Genes and Genomes (KEGG, https://www.kegg.jp/kegg/pathway.html), Reactome pathways (https://reactome.org/) and Wiki pathway analysis (https://www.wikipathways.org/) considering FDR threshold of ≤0.05 [26]. The PPI of host protein with SARS-CoV-2 viral protein was also studied in Cytoscape (v 3.4, https://cytoscape.org/) [27] using Biogrid plug-in (https://thebiogrid.org/) [28]. All SARS-CoV-2 proteins found in this study were considered to build the interactome with differentially expressed human proteins. Non-interacting SARS-CoV-2 and human proteins were removed from the PPI analysis. Final set of potential biomarker candidates were identified using a Venn diagram in InteractiVenn (https://www.interactivenn.net/) [29] that incorporated differentially expressed proteins, VIP score, AUC_ROC, represented in pathways, and interacting with SARS-CoV-2 proteins. Results Clinical phenotyping Twenty naso-oropharyngeal samples were collected that included ten from healthy controls and ten from COVID19 patients. Healthy controls had a mix of 4 males and 6 females with median age of 32.5 years, and COVID19 patients had a mix of 6 males and 4 females with a median age of 51.5 years (Table 1). Amongst COVID19 patients, six were diagnosed to be having pneumonia and the rest were diagnosed to be having acute respiratory distress syndrome. Six of the ten patients had CT values for ORF1ab and Envelop gene below 20. Table 1Demography of Healthy Control and COVID19 subjects.Table 1S.N.Patient IdAge (yrs)GenderSampleClinical PresentationCT values of ORF1ab; Envelop geneRT-PCRClinical Phenotype111955435FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC1)211955536MNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC2)311955621FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC3)411955849FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC4)511961028MNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC5)611961535FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC6)711961825FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC7)811962030MNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC8)911962265FNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC9)1011962417MNaso-oropharyngeal swabAsymptomaticUndeterminedNegativeHealthy Control (HC10)1110969945FNaso-oropharyngeal swabPneumonia18; 18PositiveCOVID-19 (COV1)129572446MNaso-oropharyngeal swabPneumonia21; 22PositiveCOVID-19 (COV2)139394453FNaso-oropharyngeal swabPneumonia15; 16PositiveCOVID-19 (COV3)149392872MNaso-oropharyngeal swabPneumonia19; 20PositiveCOVID-19 (COV4)159570542MNaso-oropharyngeal swabARDS22; 22PositiveCOVID-19 (COV5)169484064MNaso-oropharyngeal swabARDS, Septic Shock21; 22PositiveCOVID-19 (COV6)1711901650FNaso-oropharyngeal swabARDS18; 18PositiveCOVID-19 (COV7)189584856MNaso-oropharyngeal swabPneumonia22; 22PositiveCOVID-19 (COV8)199584987FNaso-oropharyngeal swabPneumonia12; 12PositiveCOVID-19 (COV9)2011936930MNaso-oropharyngeal swabPneumonia15; 15PositiveCOVID-19 (COV10) Demography of Healthy Control and COVID19 subjects. Proteomic analysis Proteomic analysis of twenty naso-oropharyngeal samples identified 298 unique human protein groups of which 135 proteins were shortlisted and taken up for further analysis. Principal Component Analysis (PCA) confirms that all the samples are confined in 95% confidence region of two distinguishable clusters of controls and diseased without any outliers (Fig. 2A). Volcano plot for the 135 proteins showed 26 differentially expressed proteins of which 12 were down-regulated and 14 were up-regulated in COVID19 patients compared to the controls (Fig. 2B). Of the differentially expressed proteins 12 proteins, Alpha-enolase (ENO1), Pyruvate Kinase PKM (PKM), Actin, cytoplasmic 1 (ACTB), Fructose-bisphosphate aldolase A (ALDOA), Cofilin-1 (CFL1), Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), H4 histone 16 (H4-16), Peptidyl-prolyl cis-trans isomerase FKBP1A (FKBP1A), Eukaryotic translation initiation factor 2A (EIF2A), Histone H2B type 2-E (H2BC21), Prolactin-inducible protein (PIP) and neuroblast differentiation-associated protein AHNAK (AHNAK) were down-regulated while 14 proteins, Heat shock cognate 71 kDa (HSPA8), l-lactate dehydrogenase B chain (LDHB), Myosin-9 (MYH9), Haptoglobin (HP), Hemoglobin subunit beta (HBB), Alpha-1-acid glycoprotein 1 (ORM1), Myristoylated alanine-rich C-kinase substrate (MARCKS), Non-histone chromosomal protein HMG-17 (HMGN2), Haemoglobin subunit alpha 1 (HBA1), Non-histone chromosomal protein HMG-14 (HMGN1), CD166 antigen (ALCAM), Statherin (STATH), Eukaryotic translation initiation factor 2 subunit 2 (EIF2S2), and Hsp90 co-chaperone Cdc37 (CDC37) were up-regulated in COVID19 patients as compared to healthy controls. While the down-regulated proteins had a fold change difference between 2 and 100, up-regulated proteins had a fold change between 2 and 130. Heatmap analysis helped to map concentrations for the 26 differentially expressed proteins across twenty individuals recruited in the study (Fig. 2C). Variable Importance in Projection identified 22 proteins with threshold score of >1 (Fig. 2D). Along with human proteins a total of 23 endogenous and 6 mutated SARS-CoV-2 peptides were also identified, originating from three structural proteins, spike (S), nucleoprotein (N), and membrane (M) and five open reading frames (ORFs): ORF1ab, ORF6, ORF7a, ORF8, and ORF9b (Table 2). Of these, 11 peptides were exclusively detected in SARS-CoV-2-infected samples. Among these 11 peptides, one peptide each was derived from S, M, ORF7a, and ORF8, while ORF6 and ORF1ab contributed 2 and 5 peptides, respectively. In COV1, COV2, COV3, COV6, and COV8 that had CT values ≤ 18, at least 15 viral peptides were identified that have mass spectrometric intensities of 60,000 or higher. Notably, COV1, COV3, and COV8, that had CT values of ≤15, at least 20 viral peptides were identified that have maximum mass spectrometric intensities. In contrast, three control samples HC6, HC7, and HC10 showed no detectable SARS-CoV-2 peptides, while three others, HC2, HC3, and HC 5 each contained only a single peptide. In addition, 6 mutated peptides were also identified (Table 2). Mass shifts (Δm) of parent peptide ion along with their fragmented ions y and b were picked up on MS1 and MS2, respectively (Fig. 3). Each of these mutated peptides were detected in at least five COVID19 samples. Two mutations, T345S and R346K, were found on a single peptide of spike protein (Receptor Binding Domain); two mutations, S197L and P207L, were detected on nucleoprotein (linker chain); two mutations, P2110S and L2609F, were identified on two separate peptides of ORF1ab (Non-structural protein 3) and one mutation, V6624L was identified on ORF1ab (Non-structural protein 15). Comparative proteomics of naso-oropharangeal tissues from COVID19 patients and healthy controls. (A) Principal Component Analysis Oval represents 95% confidence region. (B) Volcano plot showing differentially expressed proteins with fold change (FC) and p value cut off of 0.5 ≥ FC ≥ 2 and 0.05, respectively. Blue and brown describe down-regulated and upregulated proteins, respectively. (C) Heat map showing the differential expression pattern of top 30 proteins with p value of <0.05. Brown and blue indicate high and low expression level, respectively. (D) Variable importance in projection (VIP) scores of proteins. Bold colour indicates significant protein with VIP score of >1. Green indicates healthy control and Red indicates COVID19 patients. Table 2Heatmap depicts the SARS-CoV-2 peptide concentration in individual sample.Table 2Residues in blue indicate mutation. Colour green to red refers to the peptide abundance. Heatmap depicts the SARS-CoV-2 peptide concentration in individual sample. Residues in blue indicate mutation. Colour green to red refers to the peptide abundance. Mass spectra analysis of SARS-CoV-2 mutated peptides. In each peptide spectrum both MS1 (left panel) and MS2 (right panel with y and b ions) are shown for the six mutated peptides A-F. The amino acids in bracket represent the static and dynamic modifications where (C) represents the Carbamidomethylation of Cysteine/+57.02 (static modification) and (M) represents the oxidation of methionine/+16.0 (dynamic modification). Pink and green highlights in the MS1 spectra depict the instrument determined and selected monoisotopic peptide mass, respectively. However, the MS2 spectra have been generated from the green highlighted MS1 peak only. Yellow highlight is 1.2 Da precursor isolation window. Red and blue peaks in MS2 spectra corresponds to b and y ions, respectively. –NH3, –H2O and –O indicate the neutral losses which reduce the mass of fragment ions by 17, 18, and 64.1 respectively. Interactomics, functional enrichment and pathway analysis Interactomics of 12 down-regulated and 14 up-regulated proteins contains 70 edges with the average clustering ecoefficiency of 0.57 and (Protein-Protein Interaction) PPI enrichment with p-value of 1.96−10 (Fig. 4A). It is represented by one central core network consisting of 13 strongly interacting proteins, while 13 other proteins on the outside have a moderate or weak interactions with the core or amongst themselves. Fig. 4B shows the functional enrichment processes of the dysregulated proteins in the interactome. Glycolytic and immune mediated dysregulations are the main pathways featuring in the analysis. Apart from these two, SARS-CoV-2 signalling too has an implication in the pathogenesis. SARS-CoV-2 protein interaction with human proteins shows seven SARS-CoV-2 structural and ORF proteins interacting with 20 differentially expressed human proteins. HSPA8, MARCKS, and ALDOA make 6 interactions, while ENO1, LDHB, CFL1, EIF2S2 and MYH9 make 5 interactions with SARS-CoV-2 proteins. Amongst the SARS-CoV-2 proteins, the nucleocapsid protein (N) makes the maximum of 17 interactions, followed by spike protein (S) makes 15 interactions and membrane (M) makes 14 interactions with the identified human proteins (Fig. 5). Interactome and functional analysis of differentially expressed human proteins. (A) Human protein-protein interaction. Up-regulated proteins are highlighted in red; down-regulated proteins are highlighted in green; semi circles on each node depict functional enrichment where glycolysis/metabolism (blue), immune regulatory (cyan) SARS-CoV-2 signalling (pink) are represented. (B) Bar diagram illustrating functional enrichment processes. Functions are shown in x axis; bar height represents number of proteins associated to functions (left y-axis) and black diamonds represents False Discovery Rate of those functions (right y-axis). Interactome analysis between SARS-CoV-2 and differentially expressed human proteins. Green indicates down-regulated proteins, red indicates up-regulated proteins, and yellow indicates SARS-CoV-2 virus. Identification of potential biomarker candidate proteins Based on certain essential parameters such as VIP score, AUC-ROC, Pathway representation, SARS-CoV-2 interaction and differential expression 11 proteins were identified as potential biomarker candidates (Fig. 6). Functions and relevance of these proteins is discussed in Table 3. While translational utility of these proteins was obtained by plotting AUC_ROC curves, diagnostic accuracy was obtained by Box-Whisker plots (Fig. 6B–L). They show a minimum of 87% of sensitivity and 70% of specificity. ENO1 is the only protein that has 100% sensitivity and 100% specificity. Identification of potential biomarker candidates and their translational utility. (A) Venn diagram illustrating the criteria used for shortlisting of biomarker candidate proteins; Number in brackets indicates total number of proteins (B) Receiver Operating Characteristics (AUC_ROC) and Box-Whisker plot for the biomarker candidate proteins. Blue indicates area under the curve and red dot indicates optimal cut-off for the best sensitivity and specificity values. Green and red boxes indicate the range of first and third quartiles for healthy controls and COVID-19 patients, respectively. Black dots represent individual readings, and red line represents the median of box plot. Fold Change and p-values are mentioned alongside the Box-Whisker plot. ∗ indicates p-value ≤0.05; ∗∗ indicates p-value ≤0.01; ∗∗∗ indicates p-value ≤0.001. Table 3Function of differentially expressed proteins and their relevance in COVID-19 pathogenicity.Table 3S.N.ProteinRegulationFold ChangeVIPFunction and Relevance in this studyReference1Alpha-enolase (ENO1)Down0.022.3Converts 2-phosphoglycerate to phosphoenolpyruvate, encouraging usage of glycolysis for ATP production[30]Downregulates ATP production in host cells which is crucial for cellular energy production to fight against viral replication.2Pyruvate kinase PKM (PKM)Down0.091.9Transfers phosphate group from phosphoenolpyruvate to ADP, resulting in pyruvate + ATP, a key step of glycolysis[31,32]SARS-CoV-2 infection downregulates PKM mediated glycolysis, which induces the host cell apoptosis through the production of mitochondrial-derived ROS.3Actin, cytoplasmic 1 (ACTB)Down0.181.4A cytoskeletal protein involved in maintaining cell structure, intracellular transport, and cellular signaling.[33]Its downregulation during viral infection leads to cytoskeletal disruption and cellular apoptosis, resulting in altered host cell architecture, impaired immune responses, and enhanced viral replication.4Fructose-bisphosphate aldolase A (ALDOA)Down0.331.5A central enzyme in the glycolytic pathway, which is essential for generating ATP, the primary energy currency of the cell.[34]Downregulates glycolytic pathway and ATP production favours nucleotide synthesis for viral replication leading to the immune suppression which contributes to viral immune evasion.5Cofilin-1 (CFL1)Down0.201.4Actin-binding protein that regulates cytoskeletal dynamics by depolymerising actin filaments, affecting cell motility, morphology and cytokinesis.[35]Downregulation stabilizes actin and intracellular transport of pathogens that alter immune response of host and facilitate viral entry via endocytosis.6Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)Down0.131.5Enzyme phosphorylates glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate.Metabolic reprogramming in infected immune and lung epithelial cells promotes aerobic glycolysis and thereby pathogenicity.[36]7Heat shock cognate 71 kDa (HSPA8)Up24.91.8Chaperone protein assists in correct folding of polypeptides, protein transport across membranes.[37]Interacts with multiple viral proteins thereby facilitating entry and replication8l-lactate dehydrogenase B chain (LDHB)Up7.71.5Enzyme catalyzes the reversible conversion between pyruvate and lactate to replenish NAD + for glycolysis.[14,38]Promotes glycolysis for viral replication even in aerobic conditions and can also alter of host's immune-metabolic responses and cause severe inflammation during infection.9Myosin heavy chain 9 (MYH9)Up7.71.5MYH9 is an actin-based motor protein present in proximity of cell surface that has vital role in maintaining cell-cell and cell-extracellular matrix adhesion and cell migration, division, and signal transduction.[39]Binding of Spike protein to ACE2 receptor requires a cofactor like MYH9 to facilitate the viral entry through host cell membrane.10Haptoglobin (HP)Up2.51.2Acute phase plasma protein that binds to free haemoglobin to prevent oxidative damage and reduce inflammation.[40]Promotes CD163 signalling that initiates immunoregulatory pathways during infection11Hemoglobin subunit beta (HBB)Up16.92.3Is part of haemoglobin tetramer that contains haem, allowing oxygen transport via erythrocytes.[40]Binding of viral proteins to this protein Decreases Free Haemoglobin to facilitate viral replication Function of differentially expressed proteins and their relevance in COVID-19 pathogenicity. Discussion All the COVID19 patients included in this study either presented with pneumonia or acute respiratory distress syndrome. Representation of higher age group individuals in the patient category reflects the predilection of old people to COVID19 infection. All the patients had a CT value of <22 for ORF1ab and Envelop genes indicating a high viral load [41]. PCA analysis showed intra-group homogeneity and inter-group variability among the clusters of two different phenotypes without the trace of any outlier. Differentially expressed proteins are a cumulative effect of body's systematic response to SARS-CoV-2 virus. While identification of DE immune regulatory proteins (PKM, ACTB, ALDOA, CFL1, HSPA8, MYH9, HP and HBB) is indicative of systemic inflammatory response, identification of DE glycolytic and metabolic proteins (GAPDH, LDHB, ENO1, and HSPA8) is a clear indicative of the glycolytic alteration triggered by the viral invasion. All these differentially expressed proteins which showed consistent regulation pattern across the HC and COVID19 phenotypes is a clear indicative of causal association between protein expression profiling and clinical outcomes in COVID19 patients. Moreover, the VIP analysis highlights the relative importance of the human proteins to be predicting biomarker candidates. Protein analysis revealed the presence of SARS-CoV-2 peptides in both patient and healthy control samples. Interestingly, multiple SARS-CoV-2 peptides were detected in HC1, HC4, HC8, and HC9, suggesting that these individuals may have been infected with the virus but had viral loads below the RT-PCR detection threshold, resulting in undetermined outcomes. High intensity of viral peptides in certain patient samples with low CT values of RT-PCR highlights the counter validation between the proteomic and genomic data. Emergence of various mutations in SARS-CoV-2 proteins is one of the key factors contributing to COVID19 severity. Mutations enable the virus to evade neutralizing antibodies, resulted enhanced transmissibility, and reduced vaccine efficacy [11,42]. T345S and R346K mutations on spike protein was identified in 70% of COVID patients. These two mutations reside in the Receptor Binding Domain which is responsible for engaging the primary ACE2 receptor to enable viral entry into host cells [2,43]. Mutations S197L and P207L were identified on nucleocapsid protein in at least 50% of the patient samples. These mutations are located in the linker region of nucleocapsid protein, that regulates its function through interactions with NSP3 protein and plays a crucial role in regulating the virus interaction with the host cell [44,45]. Mutations P2110S, L2609F, V6624L on ORF1ab can affect both the structure and function of the encoded NSPs, potentially influencing the virus ability to replicate and interact with the host immune system [46,47]. Human protein-protein interactomics was done to reflect the cumulative physiological responses to SARS-CoV-2 infection. The salient features of the analysis are: (1) LDHB that interacts strongly with GAPDH, ALDOA, ENO1 and PKM catalyzes the conversion of lactate to pyruvate, favouring oxidative metabolism over anaerobic glycolysis suggesting a metabolic shift toward oxidative phosphorylation [14]; (2) GAPDH and ALDOA are two interacting proteins that are central enzymes in glycolytic pathway that influence oxidative phosphorylation which affects downstream ATP production and energy generation necessary for cellular processes in host [12,48]; (3) ENO1-PKM converts 2-phosphoglycerate to pyruvate thereby influencing ATP production [30,31]; (4) PKM and ALDOA are key regulators of immune cell metabolism, where PKM regulates macrophages and activate T cells, ALDOA through its interaction with ACTB influences cell migration essential for immune surveillance [49]; (5) ACTB and CFL1 are involved in actin depolymerization and cytoskeletal reconstitution which is crucial for many immune cell functions [35,50] (6) ACTB interacts with MYH9 which is a motor protein and essential for actin cytoskeleton formation found in immune cells [51]; (7) ORM1, HP, HBB and HBA1 are four acute phase proteins strongly interacting with each other and implicated dampening of immune responses mostly through haemoglobin mediated functions [[52], [53], [54]]; (8) HSPA8 interacts with key glycolytic enzymes, GAPDH and PKM2 to ensure proper folding and prevents aggregation during cellular stress or high metabolic demand. The protein acts as a binding and internalizing receptor for viruses, facilitating their entry into host cells through clathrin-dependent endocytosis [55]. Interactomics data further helped to understand the functional roles of the proteins and their associated pathways. Firstly, LDHB, GAPDH, ALDOA, PKM, HSPA8 and ENO1 are majorly involved in glycolytic functions and pathways that includes ATP and hexose metabolism, glycolysis, gluconeogenesis and HIF-1 signaling pathway. Secondly, CFL1, ALDOA, MYH9, HBB, ORM1, PKM, HSPA8, ACTB, HP and FKBP1A in the interactome that are majorly contributing to immune regulatory functions and immune related pathways. Human protein and SARS-CoV-2 protein interaction was done to decipher the possible mechanisms of virus infection in host. Notable among the interactions are: (1) HSPA8, a molecular chaperone, forms a stable complex with viral spike protein and assists in protein folding, thereby stabilizing viral proteins and facilitating viral replication [55]; (2) ALDOA binds to ORF proteins of the virus facilitating in viral invasion [36]; (3) MYH9 acts as an ACE2 coreceptor and colocalizes with the viral spike protein through S2 subunit and supports viral entry through endocytic pathway [39]; (4) GAPDH binds to spike protein at its receptor-binding domain and interferes with viral docking to ACE2 receptors [13,56,57]. It is therefore clear that the identified differentially expressed proteins, through interactions amongst themselves compromises the glycolytic metabolism and immune response, while interactions with viral proteins facilitates viral invasion and replication. Seven mutations were identified from the mass spectrometric analysis. These mutations were seen to be on Receptor Binding Domain of spike protein, linker chain of nucleoprotein, and NSP3 and NSP15 of ORF1ab. These proteins have been known to influence strength of viral interaction, modulate host immunity, and play crucial role in viral replication, respectively [2,44,45]. This therefore highlights the role of clinical proteomics and mass spectrometry in identification of viral mutations and mapping of dys-regulated host proteins that dictate virus-human response in COVID19 infection Implication and limitations This proposed study will help to develop a proteomic platform that will subsequently help in designing diagnostics for COVID19 from naso-oropharyngeal samples. This will help clinicians in early detection of COVID19 pathogenicity in patients and to streamline patient management. However, there a few limitations such as: (1) limited number of samples have been assayed; (2) not supported by the validation experiments; and (3) some endogenous peptides and two SARS-CoV-2 mutated peptides were identified in the MS analysis of control subjects, suggesting a potential false-positive identification of the viral peptide. Conclusions Proteomic analysis of naso-oropharyngeal swabs from COVID19 patients helped to identify 26 differentially expressed human proteins of which 11 are potential biomarker candidate proteins. In addition, 29 SARS-CoV-2 specific peptides from spike protein, nucleocapsid, membrane, ORF1ab, ORF6, ORF7a, ORF8 and ORF9b proteins have been identified of which 6 were mutated. These SARS-CoV-2 proteins make an array of interactions with the identified human proteins establishing their roles in pathogenesis of viral infection. Human biomarker candidate proteins are primarily associated with alterations in glycolytic pathways and immune dysregulation, which are known to facilitate viral invasion, replication, and contribute to disease severity. This study lays a platform to develop protein-based tests that can supplement RT-PCR; detect host protein changes at primary site of infection; understand host-virus interaction; identify pathways involved in viral replication and immune evasion; and lastly identify mutational sites on Spike protein as therapeutic targets. Ethics approval and consent to participate This study was conducted after approval was obtained from the Institute Ethics Committee at All India Institute of Medical Sciences New Delhi, India, and is in accordance with the Declaration of Helsinki. Informed consent was obtained from all subjects and/or their legal guardian(s). Funding sources Research Section, AIIMS, New Delhi for the Intramural research grant (A-COVID-53). CRediT authorship contribution statement Sabyasachi Bandyopadhyay: Data curation, Formal analysis, Methodology, Writing – original draft. Megha Brijwal: Investigation, Methodology. Arhaan Haridas: Formal analysis. Aashish Choudhary: Formal analysis, Methodology. Punit Kaur: Conceptualization, Funding acquisition, Project administration, Resources, Supervision. Lalit Dar: Investigation, Methodology, Resources. Gururao Hariprasad: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Declaration of competing interest Authors report no conflict of interest in this work. Acknowledgements Proteomics lab at Centralized Core Research Facility (CCRF), AIIMS for the assistance. Data availability The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with Project accession: PXD067986 [58] References Associated Data Data Availability Statement The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with Project accession: PXD067986 [58]