Figure 6.

A simple model captures predictive power of the scMIC

1. Cefotaxime diffuses into the periplasmic space of the cell, where the enzyme β-lactamase hydrolyzes cefotaxime. a_in and a_out correspond to the cefotaxime concentrations in the periplasmic space and outside of the cell, respectively.
2. At steady state, the diffusion rate of cefotaxime into the cell equals the Michaelis–Menten hydrolysis rate of cefotaxime within the cell. The corresponding cefotaxime concentration inside the periplasmic space is therefore smaller than the concentration outside the cell. C is a permeability parameter; V_max and K_M characterize the hydrolytic activity of the enzyme.
3. Bacterial growth curves with the same initial antibiotic concentrations but different starting densities. The cells die until the external concentration of cefotaxime reaches the scMIC of the strain.
4. The growth/death rate is a step function of the external cefotaxime concentration, as this determines the periplasmic concentration. Strains with different versions of TEM enzyme will have different scMIC values, which is the external antibiotic concentration at which the growth rate becomes negative (death).
5. The fits of the inoculum effect curves of the reference and mutant strains (dark regions correspond to the fitting interval). The error bars are the maximum of a discretization error and the standard error of the mean of three measurements.
6. The model prediction for competition experiments, with a 1% initial fraction of the mutant. At the scMIC of the reference strain, the final fraction of the mutant strain starts to increase, indicating that selection for the more resistant mutant starts near the scMIC. Different colors correspond to different initial cell densities (labeled in CFU/ml). The gray bar corresponds to the scMIC of reference strain. For the model, parameter values are provided in the Supplementary Information.

Source data are available online for this figure.