Allosteric Modulation Of Ligandgated Channels

The first allosteric models, originally proposed to describe enzymatic reactions (74,75), were quickly recognized for their relevance and possible applications to other heteromeric proteins, such as LGCs (76). An important extension to ligand-gated receptors with the introduction of the concept of desensitization was deduced from a series of experiments designed to resolve the receptor kinetics (77-79). It was concluded from these studies that a four-state model was sufficient to describe most of the features of the muscular nAChR (Fig. 5A) (80). The main relevance of such models concerns their capacity to predict the effects induced by different compounds. Any drug that stabilizes the active state (A) will promote a physiological response, whereas compounds that stabilize the (R) or (D) state will reduce the response amplitude that can be evoked by agonist. Another distinction that can be derived from this model is the difference between competitive inhibitors and negative allosteric effectors. Predictions made for competitive inhibitors are that for high concentrations of the inhibitor, the current is reduced to zero and that increase of the agonist concentration removes the inhibition (81). Furthermore, the partial inhibition is consistent with an allosteric mechanism but not with a competitive mechanism or with an open-channel blockade. Negative allosteric effectors do not necessarily abolish the agonist-evoked current, but the inhibition might level off even for very high concentrations of the effector.

Examination of the human a4^2 PROG dose-response curve reveals that a strong reduction of the ACh-evoked current is observed for low concentrations of PROG, but that even at its highest concentration, this steroid fails to abolish the response to agonists (Fig. 2B). Best fits obtained either with a negative allosteric effector (thick lines) or a competitive inhibitor (dashed lines) clearly illustrate the differences in prediction between these two modes of action. In addition, the effects of a competitive inhibitor should be reduced when the agonist concentration is increased. Comparison of the inhibition obtained at two ACh concentrations differing by 10-fold further illustrates that PROG does not act as a competitive inhibitor of the human neuronal nAChR but rather as a negative allosteric effector at both concentrations. The best fits were obtained with the allosteric model using two sites for the effector.

When comparable experiments are performed with the homomeric human a7 receptor, a distinct pattern of inhibition is observed (Fig. 2D). In contrast to the inhibition observed for a4^2, the a7 nAChR is almost fully inhibited by 100 ^M progesterone. The a7 dose-response inhibition is also characterized by a steeper slope of the curve (larger apparent Hill coefficient). Fits of this inhibition curve using the allosteric coefficient obtained from the a7 dose-response curve (Fig. 2C) reveal that data are adequately described using a value of five for the number of putative effector binding sites (thick line). The power of five, which corresponds to the number of negative effector sites, is necessary for the description of both the curve steepness and the amount of inhibition observed for high PROG concentrations. Results of computations corresponding to identical coefficients but using only a power of two are significantly different from the observed inhibition (thin line in Fig. 2D). Previous analysis of the properties of the a7 nAChR have revealed that this homomeric receptor must result from the assembly of five subunits (48,49,53). Thus, it is tempting to conclude that each of the a7 nAChR subunits possibly displays one binding site for PROG, which, as for the ACh-binding site, might lay at the interface between the a and its adjacent subunit.

Using the same allosteric model, predictions can also be made for positive allosteric effectors that promote the stabilization of the active (open) state. Interestingly, neuronal nAChRs are modulated by the extracellular calcium and, from the analysis of the dose-response curves recorded at different calcium concentrations, it can be shown that this divalent cation exerts a positive allosteric modulation of the neuronal nAChRs. Namely, addition of calcium in the extracellular medium induces a shift to the left of the dose-response curve, which is accompanied by an increase of the Hill coefficient together with an augmentation of the current evoked at saturating ACh concentrations (82-86). These three observations are in perfect agreement with the predictions made on the basis of allosteric modeling as shown quantitatively by Edelstein et al. (87).

Simulations computed using the allosteric model are presented in Fig. 5. The effects of a fixed concentration of positive or negative allosteric effector on the agonist-dose response profiles are illustrated in Fig. 5B. Figure 5C illustrates the dose-response profiles of a positive or negative allosteric effector at a fixed concentration of agonist. Simulations obtained with either two or five binding sites for the effectors show the difference in the steepness of the dose-response profiles and the lack of complete inhibition (m = 2).

Fig. 5. Allosteric model and theoretical prediction. (A) Schematic diagram illustrating the four-state allosteric model developed for the neuromuscular junction (79) with: R, resting state, A, active state, I, inactive state and D, desensitized state. The only open state is A. (B) Simulation of the agonist dose-response curve. The fraction of the receptor in the open state was computed using Eq. 1 (see below) (75). Dashed lines represent the effects induced by the presence of a fixed concentration of positive or negative allosteric effector. (C) Simulation of the dose-response curve of an allosteric effector. Values were computed using Eq. 1, but with a fixed concentration of agonist and several effector concentrations. Simulation of the effects of a positive or negative effector are represented. Continuous lines illustrate the results obtained assuming two binding sites for the effector, and dashed lines are the results predicted for five sites.

Fig. 5. Allosteric model and theoretical prediction. (A) Schematic diagram illustrating the four-state allosteric model developed for the neuromuscular junction (79) with: R, resting state, A, active state, I, inactive state and D, desensitized state. The only open state is A. (B) Simulation of the agonist dose-response curve. The fraction of the receptor in the open state was computed using Eq. 1 (see below) (75). Dashed lines represent the effects induced by the presence of a fixed concentration of positive or negative allosteric effector. (C) Simulation of the dose-response curve of an allosteric effector. Values were computed using Eq. 1, but with a fixed concentration of agonist and several effector concentrations. Simulation of the effects of a positive or negative effector are represented. Continuous lines illustrate the results obtained assuming two binding sites for the effector, and dashed lines are the results predicted for five sites.

0 0

Post a comment