Figure 2: Selection of zinc fingers that can discriminate between similar targets.

(a) Diagram depicts the selection process. The desired and counter targets are placed in front of the promoters that drive HIS3 and URA3 expression, respectively. Zinc-finger pools previously selected to bind each 3 bp subsite of the desired target are used as PCR templates to assemble a four-finger library, illustrated as rainbow-coloured ovals. This four-finger library is expressed as an omega fusion. To select four-finger members of this library, which are able to discriminate between the desired targets, cells are grown under conditions that are inhibited by URA3 expression but required HIS3 activation. A workflow of the procedure is shown below. (b) Selection conditions influence the enriched amino acids that correspond to the target mismatch. Using the library described in a, selection for HIS3 activation but against URA3 activation increases the fraction of the population in the GFP-positive (Pos), mCherry-negative (Neg) quadrant 4 in comparison with a HIS3-positive selection alone (top). Using the same library, selection for both HIS3 and URA3 activations increases the fraction of the population in the GFP-positive, mCherry-positive quadrant 2 in comparison to a HIS3-positive selection alone (bottom). Sequencing the zinc fingers recovered from stringent populations of these selection conditions reveal a stark difference in the amino acids enriched in the helix that corresponds to the difference in the desired and counter target (green box). (c) The binding attributes are confirmed for zinc-finger candidates that represent the recovered pools in b. Zinc-finger candidates are paired with the CCR5 versus CCR2 reporter vector and grown without selection. The GFP versus mCherry attributes are complementary to the selection conditions from which the candidate protein was derived. The sequence of the tested zinc-finger helices are shown (N to C) above the dot plots with the discriminator helix shown in colour.