Drug-target binding quantitatively predicts optimal antibiotic dose levels in quinolones
Fig 4
Prediction of relative antibiotic target molecule content from time-kill curves.
a, Dose-response curves of E. coli expressing gyrA and gyrB under the same IPTG-inducible promoter (SoA3329) grown in the presence of 10 μM IPTG (mild overexpression; red) and 100 μM IPTG (strong overexpression; yellow). A control strain (SoA3330), which expresses wild-type GyrAB levels and contains a mock plasmid, is grown in the absence of inducer (blue). The x-axis indicates the ciprofloxacin concentration, and the y-axis indicates the fold change in colony forming units over time. The dotted lines indicate experimental data, and the solid lines indicate the model fit. The best model fit was obtained for relative target molecule contents of 131% (mild overexpression) and 202% (strong overexpression) relative to the control strain (WT). b, Death rates of E. coli expressing different levels of GyrAB. The colors represent GyrAB expression conditions as in (a). The x-axis shows the percentage of bound antibiotic target normalized to the control strain; the y-axis shows the death rate δ(x). Each line represents the best fit for δ(x). c, Western blot analysis of GyrA&B in the strains/conditions shown in (a). CRP (cAMP receptor protein) was used as loading control. A representative example of six replicates is shown; see S2 Fig for full blots. d, comparison of theoretical prediction (from (b), solid colors) and GyrA2B2 tetramer levels estimated from relative GyrA&B monomer levels (quantified in (c), translucent colors). For the experimental measurements, the bars indicate the mean, and the whiskers represent the 95% confidence interval.