Abstract
The orf63 gene resides in a region of the lambda bacteriophage genome between the exo and xis genes and is among the earliest genes transcribed during infection. In lambda phage and Shiga toxin (Stx) producing phages found in enterohemorrhagic Escherichia coli (EHEC) associated with food poisoning, Orf63 expression reduces the host survival and hastens the period between infection and lysis thereby giving it pro-lytic qualities. The NMR structure of dimeric Orf63 reveals a fold consisting of two helices and one strand that all make extensive intermolecular contacts. Structure-based data mining failed to identify any Orf63 homolog beyond the family of temperate bacteriophages. A machine learning approach was used to design an amphipathic helical ligand that bound a hydrophobic cleft on Orf63 with micromolar affinity. This approach may open a new path towards designing therapeutics that antagonize the contributions of Stx phages in EHEC outbreaks.