Fig. 3

The P22N DCA-scape elucidates recognition of the Box B left element. a The P22 N protein-binding recognition landscape. Interactions are shaded according to the inset color scale. Representative interactions are enlarged to indicate predicted mutational outcomes. b Experimental confirmation of alternative specificities. The wild-type Box B left sequence is indicated with a blue star while mutants are labeled with pink stars. The U6A A12U mutation was previously described as an alternate binding mode and is predicted to result in a favorable interaction (6C–12C is observed in the wild-type Box B right motif). The A10G single mutant and C14U, G17C double mutant were analyzed using the yeast-three hybrid assay (error bars represent 1 standard deviation, three biological replicates corresponding to separate cultures were conducted for each measurement, **represents p-value < 0.01 under one-tailed t-test). Both mutations predicted to be favorable retain binding. c A mutational landscape of the Box B left sequence. The upper panel illustrates the predicted binding specificity matrix for all possible single mutations. Red stars indicate substitutions that were analyzed experimentally. The lower panel provides the experimentally observed binding activities collected based on these predictions (see Supplementary Table 1). Boxes are colored by the relative change using the wild type sequence as a point of reference. The landscape accurately captures both disruptive and favorable changes in protein–RNA recognition between protein P22 and the structured RNAs (six biological replicates corresponding to separate cultures were conducted for each measurement)