Assessment of Life Cycle Modeling Systems as Prediction Tools for a Possible Attenuation of Recombinant Ebola Viruses

Ebola virus (EBOV) causes hemorrhagic fever in humans with high case fatality rates. In the past, a number of recombinant EBOVs expressing different reporters from additional transcription units or as fusion proteins have been rescued. These viruses are important tools for the study of EBOV, and their uses include high throughput screening approaches, the analysis of intercellular localization of viral proteins and of tissue distribution of viruses, and the study of pathogenesis in vivo. However, they all show, at least in vivo, attenuation compared to wild type virus, and the basis of this attenuation is only poorly understood. Unfortunately, rescue of these viruses is a lengthy and not always successful process, and working with them is restricted to biosafety level (BSL)-4 laboratories, so that the search for non-attenuated reporter-expressing EBOVs remains challenging. However, several life cycle modeling systems have been developed to mimic different aspects of the filovirus life cycle under BSL-1 or -2 conditions, but it remains unclear whether these systems can be used to predict the viability and possible attenuation of recombinant EBOVs. To address this question, we systematically fused N- or C-terminally either a flag-HA tag or a green fluorescent protein (GFP) to different EBOV proteins, and analyzed the impact of these additions with respect to protein function in life cycle modeling systems. Based on these results, selected recombinant EBOVs encoding these tags/proteins were then rescued and characterized for a possible attenuation in vitro, and results compared with data from the life cycle modeling systems. While the results for the small molecular tags showed mostly good concordance, GFP-expressing viruses were more attenuated than expected based on the results from the life cycle modeling system, demonstrating a limitation of these systems and emphasizing the importance of work with infectious virus. Nevertheless, life cycle modeling system remain useful tools to exclude non-viable tagging strategies.