Crystal structures of antibodies in complex with viral proteins have illustrated vulnerable parts of the Lassa virus, elucidating key differences between functional and non-functional antibodies produced by the human immune system. In order to better understand the complex relation between the virus and the human immune system we will determine X-ray crystallographic structures of the main proteins (GPC, NP, Z) of the Lassa virus and other arenaviruses in complex with human antibodies.

Dr. Erica Ollmann Saphire focuses on the production and crystallization of glycoproteins and antibodies that play key roles in the pathogenesis and lethality of hemorrhagic fever viruses. The resulting crystal structures provide information seminal to the design of vaccines and inhibitors against the viruses as they exist naturally, and also provide structural templates that would permit us to anticipate and rapidly respond to newly emerging and man-made versions of the virus and viral proteins. Her lab seeks to understand how key, vulnerable three-dimensional epitopes may be displayed on a viral surface, and how antibodies might be targeted against them. The Ollmann Saphire lab is fully equipped for protein production, tissue culture and X-ray crystallography.

Technical Description

Crystal structures of viral surface glycoproteins offer a tremendous wealth of insights into receptor attachment, host tropism, mechanisms of infection and pathogenesis, viral assembly and immune recognition. For example, crystal structures of influenza virus HA have illustrated how point mutations in the receptor binding site allow shifts in tropism and infection of new species. For viruses like influenza, Vesicular stomatitis virus (VSV), and Ebola virus (EBOV), the availability of both pre-fusion and post-fusion crystal structures (the “before” and “after”) allows analysis of the conformational changes which must take place in infection and illustrates how antibody epitopes are gained or lost during this transformation. Crystal structures of viral glycoproteins also reveal mechanisms by which the glycoproteins (GP) evade host immune responses.

Crystal structures of antibodies in complex with these proteins have illustrated vulnerable epitopes, elucidated key differences between neutralizing and non-neutralizing antibodies when they recognize similar epitopes, and have directed the design of improved immunogens. In the absence of structural information, it is often difficult to understand why two antibodies of similar affinity may have different capacities to neutralize virus. For example, a large panel of antibodies has been raised against the CD4 binding site of HIV-1 gp120, but among them, antibodies such as IgG b12 that neutralize effectively, are quite rare. Structures of b12 in complex with gp120 explain that b12 has adopted a unique mode of access to the cryptic epitope. Thus, for gp120, although the shared epitope itself may be solvent-exposed, access to these sites is sterically limited to only certain angles of approach of bulky antibodies.

These insights were not evident from other methodologies and have now led to development of improved second-generation immunogens. Similarly, the crystal structure of EBOV GPs has illustrated the very few sites not masked by the thick carbohydrate coating, and has suggested alternate immunogen designs which might elicit antibodies against those sites. However, no structure of any arenavirus protein yet exists. As the arenavirus GPs have little or no sequence homology to any other non-arenavirus GP (outside of the expected coiled-coil and six-helix bundle regions of GP2), homology modeling is unlikely to be useful in accurately predicting the structure of an arenavirus GP. Likewise, the internal proteins of the arenavirus have not been examined by structural techniques. Therefore, we propose to map antibody footprints by X-ray crystallography that will identify B cell epitopes of Lassa virus proteins.