BIOINFORMATIC ANALYSIS OF VIRUS GENOMES AND PROTEIN MODELLING IN RATIONAL VACCINE DESIGN
Human immunodeficiency virus (HIV) is the infectious agent responsible for the Acquired Immune Deficiency Syndrome (AIDS). A vaccine against HIV will most likely require the elicitation of broadly neutralizing antibodies. In this work, I computationally designed and experimentally optimized a new class of protein antigens, termed epitope-scaffolds, aimed at eliciting 2F5 and bl2, two broadly neutralizing antibodies against HIV. Using the framework of Rosetta protein modeling software, the bound conformations of the 2F5 and bl2 epitopes are transferred to heterologous proteins. Epitope-scaffolds could subsequently be used in vivo to attempt re-elicitation of the respective broadly neutralizing antibodies they were designed to interact with. Different protein design strategies, involving both fixed and flexible backbone design, were employed to engineer complex antigens that structurally stabilize the linear epitope of 2F5, that provide contacts to the long CDR H3 loop of 2F5, or that stabilize the discontinuous epitope of bl2. In vitro evolution on the surface of yeast was used together with the computational approaches to further optimize the designed epitope-scaffolds. The methods described here are general and useful for other protein design applications such as the design of novel enzymes or protein inhibitors.
Use of FutureSystems
FutureGrid provides a testbed for this experiment in HPC/GPU computing.
Scale of Use
We need to use computer nodes with large RAM for a week.