James W Gillespie
Auburn University, USA
Title: Title: In silico prediction of influenza interaction sites with host cells using an artificial phage display evolution system
Biography
Biography: James W Gillespie
Abstract
Viruses lack the ability to replicate without a host. Therefore, to ensure replication and continued persistence in an environment they must acquire mutations in their capsid proteins, through natural selection, to allow specific interactions with receptors expressed on a host cell. These interaction sites are ideal targets for vaccine development or therapeutic drug development, but identification can be time consuming or highly variable due to antigenic drift and rapid mutation rates of the virus. The filamentous bacteriophage, fd, has no natural tissue tropism to mammalian cells, but can be engineered to display short peptides fused to the 4,000 copies of its major coat protein. We hypothesize that these engineered phages can be used to predict interaction sites of natural viruses with a host. Here, we enriched for a sublibrary of phage clones that interact with small airway epithelial (SAE) cells from a multi-billion phage library and identified the recovered sequences by next-generation sequencing (NGS). Representative consensus sequences for influenza hemagglutinin (HA) and neuraminidase (NA) proteins were generated using the NCBI Influenza Virus Resource. Using blastp with settings optimized for short peptides, the resulting sequences were searched against our recovered phage sublibrary interacting with SAE cells. Several peptides with high structural homology to either influenza structural proteins were identified. The recovered peptides were found near previously identified functional domains including, the membrane fusion domain and the HA0 cleavage site of HA. Additional domains were identified suggesting residues that may be involved with a co-receptor binding site. Here, we justify the use of phage display as an artificial evolution system in combination with next generation sequencing datasets to identify virus-host interaction sites based on the protein sequence of the virus. This technique can be extended to broader applications to rapidly identify interaction sites of novel pandemic or high-risk viral pathogens.