Pierre BrézellecPlease use the format "First name initials family name" as in "Marie S. Curie, Niels H. D. Bohr, Albert Einstein, John R. R. Tolkien, Donna T. Strickland"
<p>Since the publication of the genome of SARS-CoV-2 – the causative agent of COVID-19 – in January 2020, many bioinformatic tools have been applied to annotate its proteins. Although efficient methods have been used, such as the identification of protein domains stored in Pfam, most of the proteins of this virus have no detectable homologous protein domains outside the viral taxa. As it is now well established that some viral proteins share similarities with proteins of their hosts, we decided to explore the hypothesis that this lack of homologies could be, at least in part, the result of the documented loss of sensitivity of Pfam Hidden Markov Models (HMMs) when searching for domains in "divergent organisms". To improve the annotation of SARS-CoV-2 proteins, we used here the HHpred protein annotation tool and an available custom HH-suite database of HMMs specific to Homo sapiens proteins. To avoid "false positive predictions" as much as possible, we designed a robustness procedure to evaluate the HHpred results. In total, 6 robust similarities involving 6 distinct SARS-CoV-2 proteins were detected. Of these 6 similarities, 3 are already known and well documented, and one is in agreement with recent crystallographic results. We then examined carefully the two similarities that have not yet been reported in the literature. We first show that the C-terminal part of Spike S (the protein that binds the virion to the cell membrane by interacting with the host receptor, triggering infection) has similarities with the human prominin-1/CD133; after reviewing what is known about prominin-1/CD133, we suggest that the C-terminal part of Spike S could both improve the docking of Spike S to ACE2 (the main cell entry receptor for SARS-CoV-2) and be involved in the delivery of virions to regions where ACE2 is located in cells. Secondly, we show that the SARS-CoV-2 ORF3a protein shares similarities with human G protein-coupled receptors (GPCRs) belonging mainly to the "Rhodopsin family". We conclude that the approach described here (or similar approaches) opens up new avenues of research to better understand SARS-CoV-2 and could be used to complement virus annotations, particularly for less-studied viruses.</p>
Pfam Domains, HHpred, Hidden Markov Models (HMMs), Bioinformatics, Protein annotation, SARS-CoV-2.
