Thermodynamic analysis of mutant lac repressors

Abstract

The lactose (lac) repressor is an allosteric protein that can respond to environmental changes. Mutations introduced into the DNA binding domain and the effector binding pocket affect the repressor's ability to respond to its environment. We have demonstrated how the observed phenotype is a consequence of altering the thermodynamic equilibrium constants. We discuss mutant repressors, which (1) show tighter repression; (2) induce with a previously noninducing species, orthonitrophenyl-β-D-galactoside; and (3) transform an inducible switch to one that is corepressed. The ability of point mutations to change multiple thermodynamic constants, and hence drastically alter the repressor's phenotype, shows how allosteric proteins can perform a wide array of similar yet distinct functions such as that exhibited in the Lac/Gal family of bacterial repressors.

Fig. 1

The diagram illustrates the linked equilibria that result from effector (E) and operator (O) binding. R and R* correspond to the active and inactive conformations of the repressor and are in equilibrium (K_RR*). Each conformation has a distinct binding affinity for both the effector ligand and the operator ligand.

Fig. 2

Fractional expression was measured for the wild-type repressor (green triangles) and for the two mutants Q18M (orange circles) and Q18A (purple squares). The experimental data and the fit for each mutant are included. Error bars are included from replicate measurements. Q18M (the tight-binding mutant) is less leaky but induces poorly. Q18A induces well but is leakier than the wild-type repressor and the Q18M mutant.

Fig. 3

Structural view of the ligand binding pocket. (a) Structure of IPTG bound to the repressor, illustrating the hydrogen-bonding network. Light blue corresponds to the N-terminal domain, and dark blue corresponds to the C-terminal domain. (b) The residues in yellow interact with the constituent group off the C1 carbon of the galactoside effectors and were mutated to produce 115 mutant repressors. Note that the aqua residues in (a) interact with the inducer ligand but were not altered because they anchor the galactose ring or are implicated in the allosteric signaling pathway.

Fig. 4

The phenotype of the mutant repressors is plotted as a function of their leakiness (x-axis) and dynamic range (y-axis). The mutant repressors broadly fit into four classical phenotypes.

Fig. 5

Changes in thermodynamic parameters affect GFP expression. Simulated plots of fractional induction and experimental data are shown as a function of effector concentration (plotted on a log axis). Using the determined thermodynamic parameters for the wild-type repressor, we simulated theoretical plots for stepwise changes in the following: (a) the ratio of inducer binding affinity for active repressor conformation to inducer binding affinity for inactive repressor conformation (X = K_ER*/K_ER). The dark blue curve describes the wild-type repressor (see Fig. 2). The experimental values for the F161T mutant show an increase in the ratio compared to the wild type: (b) the conformational equilibrium K_RR*. The mutant Q291I illustrates that decreasing K_RR* produces a less leaky repressor; the curve is right-shifted compared to the wild type and does not induce completely. (c) The absolute values of the effector binding affinities K_ER and K_ER* at a fixed ratio X. The mutant Q291K is an example of a mutant that results in a right shift of the binding curve but does not alter the leakiness or the dynamic range.

Fig. 6

Different effectors paired with mutant repressors have unique phenotypes. (a) Addition of anti-inducers to the wild-type repressor results in a decrease in the fractional expression level by several percentages. (b) Nine mutants demonstrated a change in effector specificity when incubated in the presence of 10 mM ONPG (e₂). While induction of many mutants is slightly greater than twofold, one mutant (L296W) that was capable of a much larger induction was identified. (c) When 10 mM ONPF is added to several mutants displaying the I⁻ phenotype (e₂), repression of the natural OR1 operator is restored. While ONPF acts as a corepressor with both wild type and I⁻ mutants, a more dramatic change in expression occurs with the I⁻ mutants.