E protein interactions in flavivirus membrane fusion – flavistem

Molecualr biology, cloning of deletion mutants into expression vectors for insect cells. Biochemistry, purification of the proteins in large quantities and biophyscial characterization. Crystallization using a robotized facility. Diffraction data collection at synchrotron sources. Computer processing of the resutling diffraction data, model building into the experimentally determined electron density maps

In progress, too soon to say.

None

None yet

Flaviviruses enter cells by endosomal uptake followed by fusion of viral and endosomal membranes. The main player in these essential entry events is the envelope protein E, a class II viral fusion protein. Over the years, important structural information was obtained on protein E and its role in driving membrane fusion. This includes the crystal structure of the E ectodomain in its native, pre-fusion form; its organization at the surface of infectious particles, and the structure of E trimers in the post-fusion form. These data show that E is an essentially all ß-sheet protein, composed of three domains that line up to make an extended rod in the native, pre-fusion conformation. Domain I lies at the center, with domains II and III at either end. Domain II has a “fusion loop” (FL) at its distal end, and domain III connects to the C-terminal trans-membrane (TM) segments via a region termed the “stem”. Exposure to the acidic endosomal environment leads to E dimer dissociation with concomitant exposure of the FL, which can then insert into the endosomal membrane. After dimer dissociation, the individual E monomers trimerize and adopt a hairpin conformation, in which the C-terminus of domain III is directed towards the fusion loops. However, no structure of a flavivirus E protein containing the stem is available. The current models postulate that the stem displays key interactions with the body of the trimer, such that the TM segments become juxtaposed to the fusion loops. A hairpin conformation has been observed in all the viral fusion proteins that have been studied to date. However, in spite of serious attempts to obtain the crystal structure of the intact ectodomain of flavivirus E in its post-fusion conformation, no structural data of this region is available today.
Recent data on dengue virus E suggest that the upstream region of the stem is disordered (Klein 2012). Compared to the significantly shorter stem of class II fusion proteins (E1) from the Togaviridae family, flaviviruses have an insertion of about 40 amino acids. Recent structural data on the rubella (Dubois et al, Nature, 2013) and Chikungunya (unpublished) virus E1 glycoprotein provide details of the interactions of the stem with domain II at the level of the fusion loop. In addition, the EFF-1 fusion protein of C elegans – a distant homolog of flavivirus E (unpublished data) – also has a long stem, with about the same number of amino acids as in flavivirus E. Its structure in shows that the upstream half of its stem is disordered, as observed with dengue virus E, while the downstream region shows the same type of interactions as in the Togaviridae. Furthermore, there is a striking sequence similarity between the EFF-1 and the flavivirus E protein in this region of the stem, possibly suggesting similar interactions.
Because the interactions of the stem region are essential in providing the energy to bring together the FLs and the TM segments during fusion, the structure of this region could provide an important target for antiviral strategies. In this project, we therefore propose to bring structural information on the stem region in order to gain novel insights into the mechanism of flavivirus fusion. With this aim, we propose two parallel strategies. The first is to introduce deletions based on the recent structural data mentioned above, eliminating the upstream part of the stem. The second is to study E from a subset of flaviruses that infect exclusively mosquitos and are potential biological control weapons against transmission of mosquito-borne diseases. The C-terminal part of their E protein, containing domain III and the stem, could be more similar to that of alphaviruses. Indeed, domain III has not only one, but three disulfides and is followed by a much shorter stem region, as in alphaviruses. This second approach has the additional potential of bringing forth new information on the evolution of class II fusion proteins and their functional features.