Fig. 4: Computational design of protein components to create chemically controlled ON-switches.

a Computational design approach and architecture of the AIR-GEMS system. Starting from the CDH components (drug receptor (beige) and protein binder (green) we used a multistate design approach to search for drug-receptor variants that retained binding to the protein binder (positive design) and become resistant to the drug (negative design). We then assembled the AIR-GEMS, were the drug triggers the expression of SEAP by activating the JAK-STAT pathway. b Structural representation of Drug-1 binding pocket in Bcl-XL. Drug binding pocket (white surface) where the mutations R102E (green sticks) and F105I (blue sticks) were performed to obtain the variant iBcl-XL_v3. Drug-1 is shown in sticks representation and colored in brown. Four other designable residues are highlighted on the surface, E98 in red, T109 in cyan, S145 in orange and A149 in yellow. c Apparent IC₅₀s for Drug-1 induce dissociation of Bcl-XL:LD3 and iBcl-XL_v3:LD3 determined by SPR drug competition assay. d Structural representation of Drug-1 binding pocket in Bcl2. Drug binding pocket (white surface) where the mutations A100V (red sticks), D103N (green sticks) and Y201H (orange sticks) were performed to obtain the variant iBcl2_v4. Drug-2 is shown in sticks representation and colored in green. Two other designable resides are highlighted on the surface, V148 in blue, V156 in yellow. e Apparent IC₅₀s of Drug-2 induce dissociation of Bcl2:LD3 and iBcl2_v4:LD3 determined by SPR drug competition assay. f AIR-GEMS fold-change activity between drug (1 µM) versus no drug treatment (DMSO), showing the quantification of SEAP expression after 24 h. Each bar represents the mean of three biological replicates ± s.d, overlaid with a scatter dot plot of the original data points. g Drug dose-dependent responses in engineered cells expressing the AIR-GEMS. Each data point represents the mean ± s.d. of three replicates and the EC₅₀s were calculated using four-parameter nonlinear regression. h Schematic representation of the AIRs utilized to control transcription regulation with the GAL4/UAS system (AIR-TFs). i AIR-1-TF fold-change activity between drug treated (1 µM) versus untreated (DMSO), showing the quantification of SEAP expression after 24 h drug treatment. Each bar shows the mean of three biological replicates ± s.d, overlaid the original data points. j Drug dose-dependent responses of AIR-1-TF quantified by SEAP expression after 24 h drug treatment. Each data point represents the mean of n = 3 biological replicates, and the EC₅₀s were calculated using four-parameter nonlinear regression. k AIR-1-TF induces CAR expression comparison between treated (1 µM) versus untreated (DMSO). Histogram of flow cytometry (up) and quantification of CAR expressing cells (bottom). f, g, i, j, k Source data are available in the source data file.