standing of

this high-affinity interaction (Stayton et al., 1999). In ual compartments can directly associate with the genes that t functional, facilitating tion. Unlike off rate of use the se-ntially have a gene en-ing protein est this ap-ger binding ity, just as eads haveaddition, the streptavidin:biotin couple has been reengineered, yielding designed streptavidin variants with altered affinities for biotin (Qureshi et al., 2001; Reznik et al., 1998), designed mono-meric and dimeric streptavidins (Aslan et al., 2005; Qureshi et al., 2001; Sano et al., 1997; Wu and Wong, 2005), streptavidins with altered valency (Howarth et al., 2006), and alternatives to biotin as a ligand (Dixon et al., 2002; Hidalgo-Fernandez et al., 2006; Weber et al., 1994). Surprisingly, however, there have been few attempts to alter the couple using directed evolution (Aslan et al., 2005; Voss and Skerra, 1997). In vitro compartmentalization (IVC) technologies have been developed as a means of linking proteins and protein pheno-encode them. Those genes that produce the mos proteins will best bind their tagged DNA templates subsequent copurification, amplification, and selec other methods, this approach relies solely on the the interaction formed within the compartment. Beca lected interaction is noncovalent, receptors can pote extremely high affinities. Sepp and Choo (2005) used coding zinc finger binding sites and a zinc finger bind known to have picomolar affinity for these sites to t proach. Though this was successful, multiple zinc fin sites were used to increase binding through avid multiple receptors displayed on the surfaces of bChemistry & Biolog