A computational study of the substrate conversion and selective inhibition of aldosterone synthase

When a functional or structural impairment of cardiac output has occurred, the cardiovascular system will attempt to compensate for the reduced blood flow. Unfortunately, many of the resulting processes, such as the renin angiotensin aldosterone system, will progressively weaken the heart, resulting in the condition called heart failure. The renin angiotensin aldosterone regulatory system is currently targeted with medicine for heart failure. Many successes for the prolongation of patient age have been achieved by inhibition of angiotensin II synthesis and action. It has become apparent that this approach is suboptimal. Antagonists of aldosterone have provided better treatment options, however, side-effects are still observed. In the search for an alternative therapeutic application, we have studied a novel treatment involving the selective inhibition of aldosterone biosynthesis. The scope of this study has involved the in silico design and prediction of novel inhibitors, the synthesis of these inhibitors and analogues, and finally the in vitro measurement of their potency. The biosynthesis of aldosterone is performed by two cytochrome p450 enzymes, 11B1 and 11B2, denoted as CYP11B1 and CYP11B2, respectively. From these two family members, only CYP11B2 can perform the final synthesis step that converts 18-hydroxycorticosterone into aldosterone. CYP11B1 performs the synthesis of glucocorticoids that are responsible for metabolic, immunologic and homeostatic functions. Because these glucocorticoid actions should not be inhibited, the newly designed medicine must be CYP11B2 selective. Since CYP11B1 is highly homologous to CYP11B2, we have performed an in silico study that allows us to model the interactions of substrates and inhibitors in both the active sites of CYP11B1 and CYP11B2. Using comparative modelling, we have constructed models for the three dimensional architecture of both proteins. These models have been validated by investigating the torsional properties of the protein backbone and residue side chains, the overall protein packing and the dynamic behaviour of the protein models. Subsequently, the models have been used to evaluate the binding mechanisms and conversion mechanisms for the natural steroidal ligands of CYP11B1 and CYP11B2. A hypothetical binding mode has been proposed for 18-hydroxycorticosterone in CYP11B2, featuring the presence of stabilising hydrogen bonding interactions required for its conversion. Quantum mechanical analyses on the conversion of the steroids involved have shown a favourable conversion for this conformation, thereby supporting our hypothesis. In addition, the quantum mechanical analyses have provided insights on steroid conformations in the active sites during conversion. The suitability of the protein models for inhibitor design has been tested by subjecting the models to a case study with four known inhibitors of CYP11B1 and CYP11B2. Using molecular dynamics and molecular docking, the inhibitor potencies for CYP11B1 and CYP11B2 have been predicted, and their interactions with the proteins have been evaluated. The trends in inhibitor potency found by these computational methods have been confirmed by in vitro inhibition measurements. As a next step, the molecular docking study has been expanded to improve the confidence in the predictive power of the models. Using the protein states evaluated by the molecular dynamics study, the molecular docking results of inhibitor analogues have been investigated and the predictive power of the models has been qualitatively improved. In a final approach, we have performed a ligand-based investigation of the inhibitor analogues to determine which ligand characteristics are important for the potency for CYP11B1 and CYP11B2. To this end, we have conducted decision tree analyses on the physico-chemical properties of inhibitor substituents, resulting in a collection of descriptors that can be used for the prediction and design of novel inhibitors. We have shown that a combination of synthesis, molecular modelling and experimental measurements form a promising approach towards the design of potentially new inhibitors.