A related approach to directed evolution is to deliberately recombine functional domains from different proteins. An example is the creation of novel restriction enzymes by linking the cleavage domain from the restriction enzyme FokI with different sequence-specific DNA binding domains. FokI is a type II restriction enzyme with distinct N-terminal and C-terminal domains that function in DNA recognition and DNA cutting, respectively.
By itself the endonuclease domain cuts DNA nonspecifically. However, when the nuclease domain is attached to a DNA-binding domain, this domain determines the sequence specificity of the hybrid protein. The two domains may be joined via a sequence encoding a linker peptide such as (GlyGlyGlyGlySer)3 (Fig. 11.8). Cleavage of the DNA occurs several bases to the side of the recognition sequence, as in the native FokI restriction enzyme.
Several different DNA binding domains have been combined with the nuclease domain of FokI. For example, the Gal4 protein of yeast is a transcriptional activator that recognizes a 17 base-pain consensus sequence. Gal4 has two domains, a DNA binding domain and a transcription activating domain. The N-terminal 147 amino acids of Gal4 can be fused to the nuclease domain of FokI, giving a hybrid protein that binds to the Gal4 consensus sequence and cleaves the DNA at that location.
Zinc finger domains have also been joined to the nuclease domain of FokI. The zinc finger is a common DNA binding motif found in many regulatory proteins. The zinc finger consists of 25 to 30 amino acids arranged around a Zn ion, which is held in place by binding to conserved cysteines and histidines. Each zinc finger motif binds three base pairs, and a zinc finger domain may possess several motifs. In this example, domains approximately
90 amino acids long and comprising three zinc finger motifs, and which therefore specifically recognize nine base DNA sequences, were connected to the FokI nuclease domain (Fig. 11.9).
Both the amino acid sequence and the corresponding DNA recognition specificity are now known for a wide range of zinc fingers. This information allows zinc finger domains to be designed to read any chosen DNA sequence. DNA segments that encode zinc finger domains may be derived either from naturally occurring DNA binding proteins or by artificial synthesis, because individual zinc finger motifs are about 25 amino acids (75 nucleotides) in length. In principle, engineered proteins can now be provided with a zinc finger domain that will enable them to bind specifically to any chosen DNA sequence.