Universal Identification of Pathogenic Viruses by Liquid Chromatography Coupled with Tandem Mass Spectrometry Proteotyping

Abstract

Accurate and rapid identification of viruses is crucial for an effective medical diagnosis when dealing with infections. Conventional methods, including DNA amplification techniques or lateral-flow assays, are constrained to a specific set of targets to search for. In this study, we introduce a novel tandem mass spectrometry proteotyping-based method that offers a universal approach for the identification of pathogenic viruses and other components, eliminating the need for a priori knowledge of the sample composition. Our protocol relies on a time and cost-efficient peptide sample preparation, followed by an analysis with liquid chromatography coupled to high-resolution tandem mass spectrometry. As a proof of concept, we first assessed our method on publicly available shotgun proteomics datasets obtained from virus preparations and fecal samples of infected individuals. Successful virus identification was achieved with 53 public datasets, spanning 23 distinct viral species. Furthermore, we illustrated the method's capability to discriminate closely related viruses within the same sample, using alphaviruses as an example. The clinical applicability of our method was demonstrated by the accurate detection of the vaccinia virus in spiked saliva, a matrix of paramount clinical significance due to its non-invasive and easily obtainable nature. This innovative approach represents a significant advancement in pathogen detection and paves the way for enhanced diagnostic capabilities.

Keywords: diagnosis; mass spectrometry; proteomics; proteotyping; virus.

Graphical abstract

Fig. 1

Database construction along the cascade search strategy for a priori free proteotyping-based organism identification. Branches of the phylogenetic tree highlighted in red show the taxa included in the database at each of the three steps of the cascade search. During the first step, a subset of NCBInr harboring one representative per species (NCBInrS) is queried. Based on the results of this first iteration, the database used for the second step is constructed, including all identified genera and their descendants. Similarly, the database used for the third step is constructed based on the species detected during step 2, and their descendants. Taxa that are identified and used to construct the database of the next step are indicated by blue arrows.

Fig. 2

Proteotyping of clinical samples harboring complex communities. Taxonomical analysis of two feces samples at the phylum (A), class level (B) from the PXD036663 dataset (Balvers et al., 2023), and schematic representation of rotavirus A (credit: ViralZone, SIB Swiss Institute of Bioinformatics), detected in sample 11 highlighting the number of peptides, TSMs and length in amino acid (aa) for each structural protein (C).

Fig. 3

Schematic representation of the sample preparation and data acquisition workflow. Healthy donors were asked to spit saliva in tubes. Per donor, 100 μl of saliva were spiked with inactivated vaccinia virus and 100 μl were kept as control. Proteins were extracted followed by SP3 digestion in microplate. Peptides were quantified and 200 ng were injected on a LC-M/MS system.