4D-DIA Proteomics Uncovers New Insights into Host Salivary Response Following SARS-CoV-2 Omicron Infection

Abstract

Since late 2021, Omicron variants have dominated the epidemiological scenario as the most successful severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) sublineages, driving new and breakthrough infections globally over the past two years. In this study, we investigated for the first time the host salivary response of COVID-19 patients infected with Omicron variants (BA.1, BA.2, and BA.4/5) by using an untargeted four-dimensional data-independent acquisition (4D-DIA)-based proteomics approach. We identified 137 proteins whose abundance levels differed between the COVID-19 positive and negative groups. Salivary signatures were mainly enriched in ribosomal proteins, linked to mRNAviral translation, protein synthesis and processing, immune innate, and antiapoptotic signaling. The higher abundance of 14-3-3 proteins (YWHAG, YWHAQ, YWHAE, and SFN) in saliva, first reported here, may be associated with increased infectivity and improved viral replicative fitness. We also identified seven proteins (ACTN1, H2AC2, GSN, NDKA, CD109, GGH, and PCYOX) that yielded comprehension into Omicron infection and performed outstandingly in screening patients with COVID-19 in a hospital setting. This panel also presented an enhanced anti-COVID-19 and anti-inflammatory signature, providing insights into disease severity, supported by comparisons with other proteome data sets. The salivary signature provided valuable insights into the host's response to SARS-CoV-2 Omicron infection, shedding light on the pathophysiology of COVID-19, particularly in cases associated with mild disease. It also underscores the potential clinical applications of saliva for disease screening in hospital settings. Data are available via ProteomeXchange with the identifier PXD054133.

Keywords: COVID-19; DIA-PASEF; Omicron; SARS-CoV-2; mass spectrometry; proteomics; saliva.

Figure 1

Schematic representation of the study design and proteomics workflow. LC-MS/MS: Liquid Chromatography-Tandem Mass Spectrometry; DIA-PASEF: data-independent acquisition-parallel accumulation-serial fragmentation; TIC: total ion chromatogram. ROC: receiver operating characteristic.

Figure 2

Differentially expressed proteins (DEPs) in the saliva of COVID-19 positive and negative groups. (A) Volcano plot of DEPs showing 122 upregulated proteins (red dots) and 15 downregulated proteins (green dots) in the COVID-19 positive group, as determined by the Student’s t test (FDR 1%) with a log 2FC ≥ 1.2. (B) The heatmap shows clusters of samples based on protein expression patterns, with green indicating low and red indicating high differential protein abundance. Z-scores of normalized protein abundances were used, and clustering was performed using Euclidean distance and average linkage. The rows represent the DEPs, while the columns correspond to the individual samples.

Figure 3

Six densely connected protein complexes obtained from upregulated proteins in the COVID-19 positive cohort using the Molecular Complex Detection (MCODE) algorithm. The top three highly enriched pathways or processes are described for each module. M1, M2, M3, M4, M5, and M6 are modules 1, 2, 3, 4, 5, and 6 respectively.

Figure 4

Predictor selection and classification model. A) Venn Diagram indicating the 16 overlapping proteins between Boruta feature selection and differential abundance analysis. B) ROC curve of the best model based on SVM multivariate ROC curve analysis. Model 4 achieved the highest AUC (0.959, CI. 0.873-1). C) Seven selected proteins from Model 4 ranked by average importance. D) Confusion matrix and performance of Model 4 highlighting sensitivity (91%), specificity (94%), balanced accuracy (92%), PPV (95%), and NPV (88%) metrics.