The Transactions of NS3 and NS5 in Flaviviral RNA Replication

Abstract

Dengue virus (DENV) replication occurs in virus-induced vesicles that contain the replication complex (RC) where viral RNA, viral proteins and host proteins participate in RNA-RNA, RNA-protein and protein-protein interactions to ensure viral genome synthesis. However, the details of the multitude of interactions involved in the biogenesis of the infectious virion are not fully understood. In this review, we will focus on the interaction between non-structural (NS) proteins NS3 and NS5, as well as their interactions with viral RNA and briefly also the interaction of NS5 with the host nuclear transport receptor protein importin-α. The multifunctional NS3 protease/helicase and NS5 methyltransferase (MTase)/RNA-dependent RNA polymerase (RdRp) contain all the enzymatic activities required to synthesize the viral RNA genome. The success stories of drug discovery and development with Hepatitis C virus (HCV), a member of the Flaviviridae family, has led to the view that DENV NS3 and NS5 may be attractive antiviral drug targets. However, more than 10 years of intensive research effort by Novatis has revealed that they are not “low hanging fruits” and therefore, the search for potent directly acting antivirals (DAAs) remains a pipeline goal for several medium to large drug discovery enterprises. The effort to discover DAAs for DENV has been boosted by the epidemic outbreak of the closely related flavivirus member – Zika virus (ZIKV). Because the viral RNA replication occurs within a molecular machine that is composed several viral and host proteins, much interest has turned to characterising functionally essential protein-protein interactions in order to identify potential allosteric inhibitor binding sites within the RC.

Discussion of Chapter 11 in Dengue and Zika: Control and Antiviral Treatment Strategies

This discussion was held at the 2nd Advanced Study Week on Emerging Viral Diseases at Praia do Tofo, Mozambique.Transcribed by Hilgenfeld R and Vasudevan SG (Eds); approved by Dr. Subhash Vasudevan.

• Félix Rey: Very nice story! What did you make of the other NLSs that were postulated before? I’m referring to the NLS in the folded part of RdRp domain.

• Subhash Vasudevan: Right! You’re talking about what we call the conventional α/βNLS between residues 369 and 405. These are sequences that were actually idenified based on work that was carried out with peptide fragments fused to β-galactosidase or to GFP. However, for DENV, these do not appear to be functional in the context of the full-length NS5 in the cell-based assesments. I think the caveat is that one really has to look at localization in the context of full-length NS5. Presentation of the NLS is very important as we show in our current work with the P884T mutation. But having said that, I still believe that the thumb subdomain region of NS5 where the α/βNLS in the folded part occurs is very dynamic and it’s possible that in the context of some viruses like ZIKV NS5 for instance, it may be important. What I haven´t discussed here is that ZIKV NS5 is also localized to the nucleus much more strongly than some of the DENVs – and moreover the NS5 forms distinct nuclear structures. Why these structures form is not known but the C-terminal part of Zika virus NS5 does not contain a NLS like what we have described for DENV NS5. And there also doesn’t appear to be any kind of potential phosphorylation regulation as we have suggested for DENV 4 C-terminal NLS. I should add that this suggestion still needs to be verified experimentally.

• Paul Young: I know you don’t know the answer, but can you speculate on why at least the vast majority of NS5 actually shows up in the nucleus? And what’s the chronology of that? How early in an infection does NS5 transfer? Is it late in infection? Is it during when you are expecting most replication?

• Subhash Vasudevan: DENV2 NS5 and ZIKV NS5 start to go into the nucleus as early as about 12 hours post infection. And what is the role? While the NS5 P884T mutant virus (in DENV) suggested that it is replicating like a wild-type even though it is mostly localized to the cytoplasm. We have not clearly investigated what the downstream effect could be of the mislocalization – I think, as Julien Lescar pointed out in his talk, these NS5s may have evolved to engage different host factors a spart of their pathogenesis mechanism. It is evident when you compare JEV NS5 structure with our DENV 3 NS5. The DENV 3 NS5 seems to form this very nice compact structure that shows extensive interaction between the methyltransferase and the RdRp domains, whereas in the JEV NS5, there is not much of that. It is possible that these might interact with host proteins that could modulate events that results in the varying pathogenesis mechanisms but more work needs to be done in this area. {Note: Two publications since the meeting have shown that NS5 can cause splice variations that could modulate genes involved in innate immune pathway or sequester host nuclear transport receptors so that cellular transport required for normal host innate immune response is affected.}

• David Jans: So if I can also add a little bit. I think the simplest model at the moment would be that clearly the C-terminus plays a big role in localization, but also I think the conventional α/βNLS in the folded part does as well. All the evidence would suggest that both are important.

• Félix Rey: So you think that the folded NLS of NS5 is exposed?

• David Jans: I think that is quite possible and if you have the situation between dimers and monomers of NS5, a lot of things can happen. Don’t forget, and this is something that even I hardly mention in talks, that there is also an export signal present in DENV NS5. And if you inhibit nuclear export, NS5 level is increased in the nucleus by a lot. So we think that this is true for all of the DENV serotypes – even DENV1 NS5. Clearly NS5 is cycling all the time in a cell. And if you inhibit one or the other you can change where it goes. So it’s dyamic. And so i think probably we should be thinking that the dynamics may be the key thing anyway for what it does in the nucleus or in the cytosol for that matter.

• Ok the reverse genetics experiments are quite crystalline and the P884T has some in the cytoplasm and the virus still grow and so on. But I just want to raise a potential caveat to something that we hadn’t thought about – the dynamics of the nuclear import/export process. So the reverse genetics is done in DENV2?

• Subhash Vasudevan: Yes.

• David Jans: So not in DENV1 or DENV3. The crystal structures of full-length NS5 were from DENV3, right?

• Subhash Vasudevan: But DENV2 NS5 is the one that is in the nucleus and so is DENV3 NS5.

• David Jans: We also have to admit that we’re limited by the crystal structures. So that’s DENV3 we have the full-length NS5 structure of. We have the reverse genetic side of DENV2. And maybe we should do some reverse genetics in DENV1? I think that the 884 mutation to convert threonine into proline would be quite interesting to do. Also what we want really is the crystal structure of the full-length NS5 bound to importin-α. If we had the full-length, we’d all be happy and Félix would know what’s going on.

• Subhash Vasudevan: Agree that the mutation should be done in DENV1 infectious clone and also in DENV4.