Evasion of innate antiviral responses by picornavirus security proteins

Visser, Linda

Promoter:
Prof.dr F.J.M. (Frank) Van Kuppeveld
Co-promoter:
Dr R.J. (Raoul) de Groot & dr M.A. (Martijn) Langereis
Research group:
Kuppeveld
Date:
November 11, 2020
Time:
11:00 h

Summary

The Picornaviridae are a large virus family consisting of 47 genera and 110 species and amongst these viruses are several human and veterinary pathogens. Viral infection is recognized by the host and triggers an antiviral response, which is centered around the production of type I interferons and the antiviral effects of interferon stimulated genes (ISGs). Meanwhile, viruses try to evade the antiviral effects of this response. Picornaviruses are known to predominantly depend on their security proteins (2A and L) to evade antiviral responses, although many of the underlying molecular mechanisms remain to be elucidated. In this thesis, we studied how several picornaviruses evade the induction of type I interferons and the antiviral activities of the ISG PKR. PKR acts as a dsRNA sensor and activator of the cellular stress response.

In chapter 2, we investigated whether and how members of the Aphthovirus genus suppress the cellular stress response. We found that FMDV efficiently suppresses stress granule (SG) formation, which we found to be dependent on the catalytic activity of the Leader protease (Lpro ). We generated chimeric EMCVs that encoded either FMDV Lpro or ERAV Lpro (FMDV’s closest relative), to demonstrate that Lpro does not suppress PKR signaling but affects SG formation. We also demonstrated that Lpro cleaves the SG scaffolding protein G3BP1 and G3BP2, and we showed that G3BP1 cleavage occurs upon FMDV infection.

In chapter 3, we identified FMDV Lpro as a deISGylase and showed that Lpro has very little deubiquitinase activity. We demonstrate this preference of Lpro for ISG15 in vitro, but also in infected cells. Furthermore, we demonstrate that Lpro is a unique deISGylase, leaving a Gly-Gly modification on the lysine of target proteins, whereas canonical deISGylases hydrolyze the isopeptide linkage between ISG15 and the lysine of the target protein. The presence of uniquely modified lysines in proteins from FMDV-infected cells can potentially be used to develop new FMDV detection strategies, which can efficiently distinguish between infected and vaccinated animals.

The findings in chapter 3 also initiated the research described in chapter 4. FMDV suppresses the production of type I interferons by suppressing RLR signaling but also by directly affecting translation of IFN-a/b. FMDV Lpro inhibits cap-dependent translation by cleaving eIF4G and this likely affects IFN-a/b translation. FMDV also requires Lpro to suppress RLR signaling but the underlying mechanism remains to be fully elucidated as several mechanisms have been proposed over the years. Lpro has been suggested to degrade the transcription factors IRF3, IRF7 and NFkB. Additionally, Lpro’s DUB activity has been suggested to interfere with RLR signaling. While Lpro does have DUB activity upon overexpression, we found no evidence of significant DUB activity during viral infection (chapter 3). In chapter 4, we investigated whether the deISGylase activity of Lpro contributes to the suppression of RLR signaling or whether alternative mechanisms contribute to this ability of Lpro. First, we demonstrated that Lpro cleaves the RLR signaling proteins TBK1 and MAVS. Introduction of specific mutations in Lpro subsequently allowed us to identify amino acids that are essential for either Lpro’s deISGylase/DUB activity or its proteolytic activity towards RLR signaling proteins. We demonstrated that a Lpro mutant that is severely impaired in its deISGylase activity can still suppress IFN-β gene transcription, while a Lpro mutant that is severely impaired in the cleavage and degradation of RLR signaling proteins cannot. Our combined observations suggest that Lpro suppresses RLR signaling independently from its deISGylase activity and point to a role for the cleavage of RLR signaling proteins.

In chapter 5, we focused on members of the Enterovirus genus and investigated their ability to evade the induction of IFN-α/β gene expression and the cellular stress response. The two enterovirus proteases (i.e. 2Apro and 3Cpro ) have been suggested to help the viruses evade antiviral responses. We compared the ability of individual enterovirus proteases of representative members of the enterovirus species EV-A to EV-D to interfere with the induction of IFN-α/β gene expression as well as cellular stress response activation. This comprehensive analysis revealed that enterovirus 2Apro of multiple enteroviruses is able to suppress IFN-α/β gene expression and inhibit SG formation. 2Apro inhibits SG formation but does not affect cellular stress response signaling (e.g. counteract the phosphorylation of PKR or eIF2a), suggesting that 2Apro targets SG assembly. SGs themselves have been suggested to act as a signaling platform that is critical for the induction of IFN-α/β expression. Therefore, we investigated whether the effects of 2Apro on RLR signaling depend on 2Apro ’s ability to counteract SG assembly. Using cells that are deficient in SG formation (i.e. PKR k.o. MEFs and G3BP1/G3BP2 double k.o. HeLa cells), we demonstrated that the inability of cells to form SGs reduces IFN-α/β gene expression ~10-fold. Thus, we provide evidence that SG formation contributes to, but is not critical for, RLR signaling. In addition, we demonstrated that 2Apro suppresses IFN-α/β gene expression ~500-fold, suggesting that 2Apro has additional mechanisms to directly target RLR signaling.

In chapter 6 we investigated how members of the Kobuvirus genus evade the cellular stress response. Thus far, viruses counteract the stress response by directly antagonizing SG formation, or by preventing and/or reversing eIF2 phosphorylation. Aichivirus, through the activity of L, allows ongoing translation in the presence of high levels of p-eIF2. Aichivirus L binds eIF2B and this prevents association of p-eIF2 to eIF2B. Hereby, p-eIF2 no longer interferes with the interaction between non-phosporylated eIF2 and eIF2B, and translation can continue. Mouse kobuvirus L, which is the closest relative of aichivirus L, was the only other kobuvirus L that rescued  translation upon high levels of p-eIF2. The inability of other kobuviruses leaders to interfere with the cellular stress response suggests that this ability was a recent gain-of-function event and that kobuvirus leader likely has another primary function.

Full text