Fig. 3: Optimization of nitazene sensor breadth and sensitivity using deep mutational scanning.

A Overview of the protein engineering workflow. B Heatmap of calculated enrichment ratios of single-point mutations to the PYR1^nita sensor in binding nitazene relative to the reference control. 16 residues relevant to ligand binding were scanned. Purple indicates an increased enrichment ratio and orange indicates a decreased enrichment ratio relative to the parental sensor sequence. C Correlation plot of the enrichment ratios of PYR1^nita binding nitazene versus menitazene. The color scale correlates to that of the heatmap. Open purple squares indicate mutations that result in constitutive binding. D Model of the PYR1^nitav2 computationally designed structure. Structure of the PYR1^nitav2 binding pocket, highlighting the histidine residue at position 159 which is hypothesized to delocalize the partial charge on the nitro group. Selected mutated residues from PYR1 are shown as cornflower blue sticks. Black dashes represent potential polar contacts between the bound water, ligand, and protein. E Sequence differences between WT PYR1, computationally designed PYR1^nita, PYR1^nitav2, and PAN^nita. New mutations added in each step of the optimization are shown in orange (PYR1^nita), purple (deep mutational scan), and green (PAN sensor engineering). F Yeast surface display titrations of PYR1^nitav2 against nitazene (black circles), menitazene (purple circles), 4-hydroxy nitazene (blue circles), and benzylfentanyl (open red circles). The EC₅₀ was 2.5 nM (95% c.i. 2.2–2.9 nM), 3.4 nM (95% c.i. 3.0–4.0 nM), and 2.9 nM (95% c.i. 2.2–3.9 nM) for nitazene, menitazene, and 4-hydroxy nitazene, respectively, using 500 nM of biotinylated ΔN-HAB1^T+. Data are presented as mean values + SD. G Yeast surface display binding measured for PAN^nita against indicated ligands (BF indicates the benzylfentanyl concentration). Red striped bars indicate 0 μM ligand, orange striped bars indicate 0.01 μM ligand, green striped bars indicate 0.5 uM ligand, and purple striped bars indicate 24 μM of the nitazenes and 10 μM of benzylfentanyl. Analysis of variance was calculated using an ordinary 2-way ANOVA and Dunnett’s multiple comparison test, with a single pooled variance (**** p < 0.0001, ** p < 0.002, * p < 0.05, ns not statistically significant). For all measurements, n = 4 comprising two technical replicates performed on different days. Error bars represent 1 s.d. Source Data are provided with this paper as Source Data. Data are presented as mean values + SD.