Cytosolic recognition of DNA has emerged as a critical cellular mechanism of host immune activation upon pathogen invasion. The central cytosolic DNA sensor cGAS activates STING, which is phosphorylated, dimerises and translocates from the ER to a perinuclear region to mediate IRF-3 activation. Poxviruses are dsDNA viruses replicating in the cytosol and hence likely to trigger cytosolic DNA sensing. Here we investigated the activation of innate immune signalling by 4 different strains of the prototypic poxvirus vaccinia virus (VACV) in a cell line proficient in DNA sensing. Infection with the attenuated VACV strain MVA activated IRF-3 via cGAS and STING, and accordingly STING dimerised and was phosphorylated during MVA infection. Conversely, VACV strains Copenhagen and Western Reserve inhibited STING dimerisation and phosphorylation during infection and in response to transfected DNA and cGAMP, thus efficiently suppressing DNA sensing and IRF-3 activation. A VACV deletion mutant lacking protein C16, thought to be the only viral DNA sensing inhibitor acting upstream of STING, retained the ability to block STING activation. Similar inhibition of DNA-induced STING activation was also observed for cowpox and ectromelia viruses. Our data demonstrate that virulent poxviruses possess mechanisms for targeting DNA sensing at the level of the cGAS-STING axis and that these mechanisms do not operate in replication-defective strains such as MVA. These findings shed light on the role of cellular DNA sensing in poxvirus-host interactions and will open new avenues to determine its impact on VACV immunogenicity and virulence.
Vaccinia virus produces two types of virions known as single-membraned intracellular mature virus (MV) and double-membraned extracellular enveloped virus (EV). EV production peaks earlier when initial MV are further wrapped and secreted to spread infection within the host. However, late during infection MV accumulate intracellularly and become important for host-to-host transmission. The process that regulates this switch remains elusive and is thought to be influenced by host factors. Here we examined the hypothesis that EV and MV production are regulated by the virus through expression of F13 and the MV-specific protein A26. By switching the promoters and altering the expression kinetics of F13 and A26, we demonstrate that A26 expression downregulates EV production and plaque size, thus limiting viral spread. This process correlates with A26 association with the MV surface protein A27 and exclusion of F13, thus reducing EV titres. Thus, MV maturation is controlled by the abundance of the viral A26 protein, independently of other factors, and is rate-limiting for EV production. The A26 gene is conserved within vertebrate poxviruses, but strikingly lost in poxviruses known to be transmitted exclusively by biting arthropods. A26-mediated virus maturation thus has the appearance to be an ancient evolutionary adaptation to enhance transmission of poxviruses that has subsequently been lost from vector-adapted species, for which it may serve as a genetic signature. The existence of virus-regulated mechanisms to produce virions adapted to fulfil different functions represents a novel level of complexity in mammalian viruses with major impact on evolution, adaptation and transmission.
Ubiquitylation is a covalent post-translational modification that regulates protein stability and is involved in many biological functions. Proteins may be modified with mono-ubiquitin or ubiquitin chains. Viruses have evolved multiple mechanisms to perturb the cell ubiquitin system and manipulate it to their own benefit. Here, we report ubiquitylation of vaccinia virus (VACV) protein N1. N1 is an inhibitor of the nuclear factor NF-κB and apoptosis that contributes to virulence, has a Bcl-2-like fold, and is highly conserved amongst orthopoxviruses. The interaction between N1 and ubiquitin occurs at endogenous protein levels during VACV infection and following ectopic expression of N1. Biochemical analysis demonstrated that N1 is covalently ubiquitylated, and heterodimers of ubiquitylated and non-ubiquitylated N1 monomers were identified, suggesting that ubiquitylation does not inhibit N1 dimerization. Studies with other VACV Bcl-2 proteins, such as C6 or B14, revealed that although these proteins also interact with ubiquitin, these interactions are non-covalent. Finally, mutagenesis of N1 showed that ubiquitylation occurs in a conventional lysine-dependent manner at multiple acceptor sites because only an N1 allele devoid of lysine residues remained unmodified. Taken together, we described a previously uncharacterized modification of the VACV protein N1 that provided a new layer of complexity to the biology of this virulence factor, and provided another example of the intricate interplay between poxviruses and the host ubiquitin system.