Structures of Zika Virus E & NS1: Relations with Virus Infection and Host Immune Responses

Abstract

Zika virus (ZIKV), first discovered in the Zika forest in Uganda in 1947 was understudied until the recent explosive epidemic in several South American countries where it has become strongly associated with congenital birth defects leading to severe cranial malformations and neurological conditions. The increase in number of case of microcephaly in newborn children associated with ZIKV infection triggered the World Health Organization to declare the epidemic as a Public Health Emergency of International Concern in February of 2016. ZIKV is a member of the flavivirus genus and is transmitted by Aedes aegypti mosquitoes, however in the current epidemic clear evidence is emerging to suggest the virus can be sexually transmitted from human to human. The differences in epidemiology and manifestations of ZIKV infection during these outbreaks have prompted researchers to investigate mechanisms of dissemination, pathogenesis, and host immune response which contributes significantly to the control of the virus infection. The E and NS1 proteins of ZIKV are the major targets for neutralizing and protective antibodies. In this chapter, we mainly focus on recent research on the crystal structures of the ZIKV E and NS1 proteins, and their relations with virus infection and immune responses. These studies will be helpful to develop novel therapeutics and vaccines for protection and control of ZIKV infection.

Discussion of Chapter 6 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 Drs. Gao and Shi.

• Félix Rey: When you selected the antibodies, were you using the Zika virus?

• George Gao: We have had one imported case, which is the first imported case of Zika infection [in China]. We cannot label the whole virus so we only label the His-tagged E protein for antibody selection.

• Félix Rey: You said that your E protein was monomeric. So why are you surprised that you got many fusion-loop antibodies, if it is monomeric?

• George Gao: I agree that in the monomeric form you expose the fusion loop but it is possible based on the structure that it could be covered like you see in a dimer.

• Félix Rey: But your selection process is not picking those, that are at the interface.

• George Gao: If it is not that immunogenic as you are expecting, then they will not be selected and we persistently see only the fusion loop antibodies.

• Félix Rey: In Gavin Screaton’s screen, he had seven patients and the response was binary: From some patients, he only got ADE (enhancing antibodies) and from others he only got fusion loop antibodies. But he used whole viruses for selection.

• George Gao: I agree. Our selection processes may be biased.

• Aravinda de Silva: I think, we have had a lot of discussion about fusion loop antibodies and protection. And I think we are all interpreting that data. So after primary Dengue virus infection, people make a lot of fusion loop antibodies. And we know that they are not protective, because if they were protective, you would not get an infection with a different serotype, because fusion loop antibodies are broadly crossreacting. What that tells me is that in vivo, in people, when you have been exposed to a different serotype fusion loop, that antibodies are not protective. And that might mean that in vivo, the fusion loop is not exposed in the viruses that go on to mediate infection. And I think one other thing that we need to think about is when we think about virus breathing and exposure of fusion loops, how the lab-adapted strains of viruses that we work with are compared with clinical strains. And we know very little about the maturation state flexibility of the virions that circulate in people. So despite the abundance of fusion loop antibodies, there is no evidence in people that they are protective. That is different from the ADE-type antibodies, which clearly do. I do not think you can call those fusion loop antibodies, even though they are binding next to the fusion loop, they are binding to domain II, but they are not really binding to the fusion loop.

• George Gao: OK this could mean where the antibody actually binds is an important aspect but it may be minor. We use the soluble E protein of the virus for selection but end with an antibody that bind as if the protein is a dimer.

• Félix Rey: You know that the Dengue virus surface is very heterogeneous with immature patches. And you have like 60% uncleaved prM on the virion surface. So in all of those, when prM is in complex with the E protein, the fusion loop is on the side. Pr is not covering the fusion loop and it is accessible to the antibodies. I think that gives rise to the immunogenicity. I mean most of those fusion loop antibodies come from the partially immature virus.

• George Gao: We isolated at least one antibody, targeting the inner side, not on the surface of the domain I. How can we get this type of antibody that binds very strongly, of course it is non-neutralizing, but the fact is that you can isolate from our selection procedure, a new antibody that cover the inner side. I don’t see the connection with immature conformation.

• Félix Rey: That is probably because you always have some misfolded particles somewhere in the mix of your protein, even against internal proteins you can get antibodies.

• Paul Young: Have you done any lipid membrane modeling studies yet, with the Zika NS1?

• Yi Shi: No, we are working on it. We have just designed mutant proteins to see whether this region is important. We are now doing the lipid floating experiment which is a membrane-association technique.

• Subhash Vasudevan: Have you compared your Brazilian strain with the MR766 strain in the region where you think the loop structures are different? Did you compare the sequences?

• Yi Shi: You mean the sequence that is in the spike region?

• Subhash Vasudevan: Yes. There must be some differences between the strains.

• Yi Shi: No, not yet but there must be some differences.

• Aravinda de Silva: So you have suggested that the protein must have a role in helping the virus to cross the placenta. For Dengue, it is really endothelial cells that are affected by NS1. And for the placenta, the barrier is not really the endothelial cells. Because the layer between maternal and fetal blood, is the trophoblast and it forms the barrier between the mother and the baby. So if NS1 has an effect there, it has to be not on endothelial cells but on the cell that really separate the maternal and the fetal compartment in the placenta and that is the trophoblast.

• Yi Shi: Maybe I think, our proposal of NS1 helping the virus to get into the placenta is just a hypothesis. As you know, the NS1 has many effectors and there are a lot of factors that with which it can interact. It may have some function in the placenta.

• Aravinda de Silva: I think, it is an interesting hypothesis. What I am suggesting is to look at the layers of trophoblast in the placenta and not in the endothelial cells.