Abstract:

Due to their restricted genome size, viruses have developed various strategies to maximize their genetic diversity through mechanisms such as alternative splicing (AS) or discontinuous transcription (DT). Viral proteins themselves can have multiple functions during the replication cycle. Concurrently, viruses have evolved to hijack numerous cellular proteins to support their replication. Among them, a large number of RNA binding proteins (RBPs) play an essential role at different stages of the cycle. My thesis project aims to study certain RBPs, viral or cellular, that participate in the fate of viral RNAs during HIV-1 and SARS-CoV-2 infection. My project revolves around three axes: 1-Characterize the diversity of viral genomes and transcriptomes by developing a bioinformatics tool for full-length sequencing analysis: NanoViZer, 2-Characterize the unexpected role of the viral protein Tat in the regulation of HIV-1 AS, and 3-Identify cellular RBPs and characterize the mechanisms involved in SARS-CoV-2 virion assembly.

Full-length sequencing has revolutionized transcriptome studies, but bioinformatics tools require genome annotation files and are limited to studying specific mechanisms. In the context of viruses, annotation files are often incomplete or missing. To analyze the architecture of viral genomes and transcriptomes without a priori on the mechanisms involved, we developed NanoViZer, a fast and user-friendly tool that does not require any annotation files. After coding the algorithm, I validated this tool on literature data to analyze the transcriptome of HIV-1, HSV-1, and SARS-CoV-2. Thanks to this tool, I revealed the presence of deletions in the genomes of hepatitis D virus in the circulating blood of patients.

In the second part of my thesis, I explored a new role of the HIV-1 Tat protein in AS regulation. During its replication, HIV-1 generates more than 50 mRNAs through AS of a single pre-messenger RNA. An imbalance affecting virion production, this process is finely regulated. Three studies suggested a role of Tat in AS, but its effect and the mechanisms involved were controversial. Using a viral model in which HIV-1 transcriptional transactivation is independent of Tat (HIV-VP16), and taking advantage of full-length sequencing and NanoViZer, we demonstrated that Tat destabilizes the entire viral AS program and particularly favors the production of Tat mRNAs, thereby promoting its own protein expression. By exploring the activity of several Tat mutants, we then showed that Tat's activity in AS is uncoupled from its transcriptional activity. Finally, analysis of the replication of these different mutants suggests that Tat's role in AS promotes virus replication, independently of its transcriptional role.

Lastly, amidst the COVID-19 pandemic, I focused on RBPs involved in late stages of SARS-CoV-2 replication. For this purpose, SARS-CoV-2 virions were isolated from two infected pulmonary cell lines and analyzed by mass spectrometry. Through computational analyses, I identified 92 cellular proteins associated with virions, including many RBPs. In particular, 27 proteins were found in cytoplasmic structures called stress granules. Interestingly, G3BP proteins, core proteins of stress granules, were particularly enriched. By silencing their expression, we revealed that G3BPs interact with the viral nucleoprotein N and SARS-CoV-2 RNA, promoting SARS-CoV-2 virion assembly in the cell and the production of infectious viral particles.

In conclusion, my work has led to the development of a bioinformatic tool, NanoViZer, enabling the analysis of deletions in viral genomes and transcriptomes without prior assumptions through full-length sequencing. Additionally, I have demonstrated new roles for RNA-binding proteins in viral replication. The viral Tat protein of HIV-1, which regulates the splicing of viral RNAs, and the cellular G3BP proteins, which promote the assembly of SARS-CoV-2 viral particles.

Key words: HIV-1, SARS-CoV-2, NanoViZer, RBP, Full-length sequencing, transcriptomic, proteomic, Tat protein, G3BP protein, bioinformatic