For most proteins the amino acid sequence determines the tertiary structure. The relative importance of the individual amino acids in specifying the fold, however, remains unclear. To highlight this, Creamer and Rose put forth the 'Paracelsus challenge': Design a protein with 50% sequence identity to a protein with a different fold. We have met this challenge by designing a sequence which retains 50% identity to a predominantly beta-sheet protein, but which now adopts a four helix bundle conformation and possesses the attributes of a native protein. Our results emphasize that a subset of the amino acid sequence is sufficient to specify a fold, and have implications both for structure prediction and design.
