Joanna Brunel

The human genome comprises 8% of endogenous retroviruses (HERVs). Though HERVS contribute to physiological functions, copies retained pathogenic potential. The HERV-W ENV protein was shown expressed in patients with worse COVID-19 symptoms and post-COVID syndrome. A significant detection of the mRNA encoding HERV-W ENV from patients with COVID-19 in B cells from RNAseq reads obtained from peripheral blond mononuclear cells. This data stratified with increased COVID-19 symptoms or with post-acute sequelae of COVID-19 (long COVID) after 3 months. The HERV-W ENV-U3R RNA was confirmed to display the best alignment with chromosome X ERVWE2 locus. However, a stop codon precluding its translation was re-addressed after recent understandings of ribosome readthrough mechanisms. Experimental results evidenced that this HERV gene can effectively express a full-length protein in the presence of molecules allowing translation via a readthrough mechanism at the ribosome level. Results not only confirm HERV-W ENV RNA origin in these patients, but show for the first time how a defective HERV copy can be translated into a complete protein when specific factors make it possible at the ribosome level. The present proof of concept now requires further studies to identify the factors involved in this newly understood mechanism, following SARS-CoV-2 exposure.

Epidemiology has repeatedly associated certain infections with a risk of further developing psychiatric diseases. Such infections can activate retro-transposable genetic elements (HERV) known to trigger immune receptors and impair synaptic plasticity of neuroreceptors. Since the HERV-W ENV protein was recently shown to co-cluster with pro-inflammatory cytokines in a subgroup of patients with schizophrenia or bipolar disorder, we questioned the influence of the COVID-19 pandemic on patients with psychosis spectrum disorders (PSD). Present results revealed that (i) SARS-CoV-2 serology shows high prevalence and titers of antibodies in PSD, (ii) HERV-W ENV is detected in seropositive individuals only and (iii) SARS-CoV-2 and HERV-W ENV positivity co-clustered with high serum levels of pro-inflammatory cytokines in psychotic patients. These results thus suggest that SARS-CoV-2 infection in many patients with psychotic disorders now admitted in the psychiatry department did not cause severe COVID-19. They also confirm the previously reported association of elevated serum pro-inflammatory cytokines and HERV-W ENV in a subgroup of psychotic patients. In the context of the COVID-19 pandemic, this cluster is only found in SARS-CoV-2 seropositive PSD cases, suggesting a dominant influence of this virus on HERV-W ENV and cytokine expression, and/or patients’ greater susceptibility to SARS-CoV-2 infection. Further investigation on an interplay between this viral infection and the clinical evolution of such PSD patients is needed. However, this repeatedly defined subgroup of psychotic patients with a pro-inflammatory phenotype and HERV expression calls for a differential therapeutic approach in psychoses, therefore for further precision medicine development.

Patients with COVID-19 may develop abnormal inflammatory response and lymphopenia, followed in some cases by delayed-onset syndromes, often long-lasting after the initial SARS-CoV-2 infection. As viral infections may activate human endogenous retroviral elements (HERV), we studied the effect of SARS-CoV-2 on HERV-W and HERV-K envelope (ENV) expression, known to be involved in immunological and neurological pathogenesis of human diseases. Our results have showed that the exposure to SARS-CoV-2 virus activates early HERV-W and K transcription but only HERV-W ENV protein expression, in an infection- and ACE2-independent way within peripheral blood mononuclear cell cultures from one-third of healthy donors. Moreover, HERV-W ENV protein was significantly increased in serum and plasma of COVID-19 patients, correlating with its expression in CD3 ⁺ lymphocytes and with disease severity. Finally, HERV-W ENV was found expressed in post-mortem tissues of lungs, heart, brain olfactory bulb and nasal mucosa from acute COVID-19 patients in cell-types relevant for COVID-19-associated pathogenesis within affected organs, but different from those expressing of SARS-CoV-2 antigens. Altogether, the present study revealed that SARS-CoV-2 can induce HERV-W ENV expression in cells from individuals with symptomatic and severe COVID-19. Our data suggest that HERV-W ENV is likely to be involved in pathogenic features underlying symptoms of acute and post-acute COVID. It highlights the importance to further understand patients’ genetic susceptibility to HERV-W activation and the relevance of this pathogenic element as a prognostic marker and a therapeutic target in COVID-19 associated syndromes. Graphical

Effective host defence against viruses depends on the rapid triggering of innate immunity through the induction of a type I interferon (IFN) response. To this end, microbe-associated molecular patterns are detected by dedicated receptors. Among them, the RIG-I-like receptors RIG-I and MDA5 activate IFN gene expression upon sensing viral RNA in the cytoplasm. While MDA5 forms long filaments in vitro upon activation, RIG-I is believed to oligomerize after RNA binding in order to transduce a signal. Here, we show that in vitro binding of synthetic RNA mimicking that of Mononegavirales (Ebola, rabies and measles viruses) leader sequences to purified RIG-I does not induce RIG-I oligomerization. Furthermore, in cells devoid of endogenous functional RIG-I-like receptors, after activation of exogenous Flag-RIG-I by a 62-mer-5′ppp-dsRNA or by polyinosinic:polycytidylic acid, a dsRNA analogue, or by measles virus infection, anti-Flag immunoprecipitation and specific elution with Flag peptide indicated a monomeric form of RIG-I. Accordingly, when using the Gaussia Luciferase-Based Protein Complementation Assay (PCA), a more sensitive in cellula assay, no RIG-I oligomerization could be detected upon RNA stimulation. Altogether our data indicate that the need for self-oligomerization of RIG-I for signal transduction is either dispensable or very transient.

The genome of nonsegmented negative-strand RNA viruses is tightly embedded within a nucleocapsid made of a nucleoprotein (N) homopolymer. To ensure processive RNA synthesis, the viral polymerase L in complex with its cofactor phosphoprotein (P) binds the nucleocapsid that constitutes the functional template. Measles virus P and N interact through two binding sites. While binding of the P amino terminus with the core of N (N _CORE ) prevents illegitimate encapsidation of cellular RNA, the interaction between their C-terminal domains, P _XD and N _TAIL is required for viral RNA synthesis. To investigate the binding dynamics between the two latter domains, the P _XD F497 residue that makes multiple hydrophobic intramolecular interactions was mutated. Using a quantitative mammalian protein complementation assay and recombinant viruses, we found that an increase in P _XD -to-N _TAIL binding strength is associated with a slower transcript accumulation rate and that abolishing the interaction renders the polymerase nonfunctional. The use of a newly developed system allowing conditional expression of wild-type or mutated P genes, revealed that the loss of the P _XD -N _TAIL interaction results in reduced transcription by preformed transcriptases, suggesting reduced engagement on the genomic template. These intracellular data indicate that the viral polymerase entry into and progression along its genomic template relies on a protein-protein interaction that serves as a tightly controlled dynamic anchor. IMPORTANCE Mononegavirales have a unique machinery to replicate RNA. Processivity of their polymerase is only achieved when the genome template is entirely embedded into a helical homopolymer of nucleoproteins that constitutes the nucleocapsid. The polymerase binds to the nucleocapsid template through the phosphoprotein. How the polymerase complex enters and travels along the nucleocapsid template to ensure uninterrupted synthesis of up to ∼6,700-nucleotide messenger RNAs from six to ten consecutive genes is unknown. Using a quantitative protein complementation assay and a biGene-biSilencing system allowing conditional expression of two P genes copies, the role of the P-to-N interaction in polymerase function was further characterized. We report here a dynamic protein anchoring mechanism that differs from all other known polymerases that rely only onto a sustained and direct binding to their nucleic acid template.