Protein Design

Computational protein design is the core thrust of the Bystroff lab. We use an approach called "leave-one-out" biosensor design, in which a sequence is designed to fit a protein of known structure with a segment omitted (the target site). The sequence of the omitted segment is replaced by the target sequence for sensing. Then protein design builds a complementary shape into the pocket of the omitted segment, creating a binding site. The computational design approach is based on the Dead End Elimination algorithm, with our own modifications. Initial experimental studies have shown that leave-one-out design works and gives us biosensor with sub-micromolar affinity for the target sequence.

A peptide biosensor based on green fluorescent protein (GFP).

click for a closer look
We are developing a version of green fluorescent protein (GFP) that will glow only in the presence of a selected peptide sequence, such as a peptide unique to a pathogen sch as the avian flu virus. Studies are being done to explore how GFP can be "re-wired" to place the peptide binding site at different places in the sequence. Re-wired GFP is also useful for basic understanding of the effect of topology on protein stability and kinetics, with an eye towards designing kinetic stability into proteins.

Back to research interests