Expansion of the genetic alphabet to include a third base pair not only has immediate utility for a number of applications, such as site-specific oligonucleotide labeling, but also serves as the foundation for an organism with an expanded genetic code. Toward this goal, we have examined a large number of different unnatural nucleotides bearing mainly hydrophobic nucleobase analogs that pair based on packing and hydrophobic interactions rather than H-bonding. Optimization based on extensive structure-activity relationship studies and two screens resulted in the identification of a class of unnatural base pairs that are well recognized by DNA and RNA polymerases. More recently, we have engineered E. coli to import the requisite unnatural triphosphates and shown that DNA containing the unnatural base pair is efficiently rep licated within the cell, resulting in the first semi-synthetic organism that stores increased information in its genome.