Allosteric Inhibitors of the Enzyme IspD of the Non-Mevalonate Pathway for Isoprenoid Biosynthesis

Abstract

Malaria is a major burden and cause of death in developing contries with over 200 million cases reported worldwide each year. The emergence of multi drug resistant malaria parasites, in particular against the widely used artemisinin based therapy, requires a timely response. The development of drugs with new modes of action is among the most efficient ways in combating resistance and necessitates the exploartion of new molecular targets of the malaria parasite. The discovery of the non-mevalonate pathway of isoprenoid synthesis has opened up a diverse set of molecular targets for the treatment of malaria but also for antibacterial agents and herbicides. This pathway fundamentally differs from the mevalonate-dependent pathway by which humans and mammals derive isoprenoids, and therefore promises a good selectivity for new treatments. Of the seven enzymes comprising the non-mevalonate pathway, IspD stands out due to its allosteric regulatory site. Inhibition through allosteric sites is highly advantageous as inhibitors do not need to compete with substrates present at high concentrations. Preceeding work has established azolopyrimidines (IC50 (1) = 0.18 uM, IC50 (2) = 0.029 uM) and the ion chelating pseudilins (IC50 (3) = 40 uM and 5.8 uM with Cd2+) as allosteric inhibitors of IspD of Arabidopsis thaliana (AtIspD) and has elucidated their binding mode. A more detailed insight into the allosteric mechanism was given by the discovery of phenylisoxazole ligand 4 (IC50 = 9.2 uM) and mutation studies on the enzyme. Binding of allosteric ligands leads to unwinding of an alpha helix and a large displacement of a loop region of the enzyme. The loop is displaced to a different extend depending on the ligand class, and the side chains of aspartate residues 261 and 262 of the loop were found to mediate inhibition, likely by coulombic repulsion of the substrate. New allosteric ligands to investigate the specific requirements for allosteric inhibition were designed and synthesized, based on the structural information and mechanistic understanding of the allosteric mechanism. Substitution of the dichlorophenol of pseudilin 3 by a motif that replaces the halogens by spirocyclopropanes highlighted the contribution of halogen bonding interactions of 3 with the enzyme (IC50 (5) = 98 uM). Difluoroindoles such as 6 (IC50 = 60 uM) provide a replacement for the potentially unstable tribromopyrrole of the pseudilins. Unlike the lead structure 3, binding of 6 however occurs independent of metal ions. The characteristic allosteric loop conformation of each ligand class has raised the question if the conformation of the displaced loop is linked to the degree of inhibition and if different conformations are possible for a given ligand class. Low activity of the extended azolopyrimidine ligand 7 designed to stabilize a loop conformation as observed for pseudilin 3, suggests specificity of a loop conformation to their ligand class. A common feature of inhibitors 1-4 is the acidic hydroxy moiety which allows the ligand to bind the enzyme in an anionic form at physiological pH. The effect of ligand acidity on binding affinity was shown by ligand 8 (IC50 = 2.9 uM) and 9 (IC50 = 0.13 uM), which are isomeric to the lead structures 1 and 2. The shifted nitrogen of the heterocyclic scaffold results in a less acidic ligand, while shape and binding interactions are unaffected. The higher pKa values of 8 and 9 correlate with lower inhibition and have identified ligand acidity as a key component for the design of new inhibitors. The azolopyrimidine ligand class has shown a strong impact of small substituents on the benzyl group on inhibition. As a consequence, the allosteric pocket of AtIspD was recognized as a responsive environment to evaluate bioisosteres of benzene. DFT calculations have identified quadricyclane as an isostere of benzene which, unlike other aliphatic replacements, also mimics its quadrupole moment. In this context, the norbornadiene-quadricyclane photoswitch is proposed as a compact switching element for photopharmacological applications.