Conformational preferences of plumbagin with phenyl-1-thioglucoside conjugates in solution and bound to MshB determined by aromatic association
Graphical abstract
Highlights
► Pharmacophore shows 95% inhibition of the mycothiol biosynthesis enzyme MshB. ► Docking calculations are inconsistent with experimental binding pattern. ► Free energy of the preferred solution conformations correspond with experiment. ► Preferred in solution ligand conformations are preserved upon binding to the enzyme. ► This in the context of Fischer’s ‘lock-and key’ and Koshland’s induced-fit model.
Introduction
Mycothiol (MSH), 1-O-[2-[(2R)-2-acetamido-3-mercaptopropanamido]-2-deoxy-α-d-glucopyranosyl]-d-myo-inositol, is the major low molecular weight thiol protecting Mycobacterium tuberculosis (M. tuberculosis) against oxidative stress.1, 2 This ensures not only the survival of the pathogen, but allows it to flourish. There is no mammalian counterpart to the MSH biosynthetic pathway. Therefore, interrupting the biosynthetic pathway of this pseudo-disaccharide has become a focus for the design of anti TB drugs.
MshB (EC# 3.5.1.103), a GlcNAc-Ins N-deacetylase, is the third enzyme in the MSH biosynthetic pathway and catalyses hydrolysis of the N-acetyl amide bond of GlcNAc-Ins (N-acetyl-glucosaminylinositol, see Supplementary data Fig. S1) to give the free amino sugar GlcN-Ins. We have recently shown that a series of substrate analogues, based on a pyranose ring scaffold to which two aromatic hydrophobic functional groups were tethered, inhibit enzyme MshB with increasing effectiveness as the methylene chain linking one of the hydrophobic moieties to the sugar was lengthened (Fig. 1).3
A rationale for this inhibition is still outstanding, and is key to a programme led by rational drug design of MshB inhibitors. Here we aim to investigate a developing pharmacophore profile of a series of ligands with equivalent chemical identities that ranged from 57% (short methylene chain) to 95% (long methylene chain) inhibition of MshB. To explain the difference in competitive inhibition, we explore the molecular recognition of the ligands and MshB.
Induced conformational changes to ligands and enzymes that lead to their optimised molecular recognition of each other are critical to the design of small molecule drugs. In the case of a substrate binding to an enzyme, the active site is often observed to undergo conformational changes. These changes alter the local environment in the catalytic domain to align the reactive groups to more efficiently react, resulting in the transition state.4 On the other hand the ligand may deform from its low energy solution conformation on binding to the enzyme active site as has been shown for peptides.5 Three classic ligand–enzyme complex models are used to interpret the conservation or deformation of the ligand or the active site conformation during the process of binding. The oldest of these is Fischer’s ‘lock-and-key’ model6, where both the ligand and catalytic binding site conformations are preserved upon binding. Koshland’s induced-fit model7 has been around for more than 50 years and has been used as a textbook explanation for molecular recognition events where both the ligand and catalytic binding site conformations undergo changes upon binding. The more recent selected-shift model, or conformational selection mechanism, describes the dynamic response of the protein resulting in the initial binding of a preconfigured ligand conformation to an ensemble of ligand binding sites.8
Although the set of ligands we investigate here is small, the dramatic difference in their inhibitory behaviour begs attention simply because it is due to an apparent minor alteration from the least to most active elements of the set. That is that a small extension to the methylene chain, linking a carbohydrate and an aromatic functional group, switches inhibition of MshB on. Simple inspection of the binding pocket using a lock-and-key tool kit offers no explanation, as there are no differences between the hydrogen bonding sites of the active and relatively inactive ligands. Further the orientations of the ligands within the binding pocket are uniform due to the requisite coordination of their common pyranose to a Zn2+ ion (present in the natural substrate) that is central to the catalytic success of MshB.
Low energy conformations of ligands are often loosely derived from either crystal diffraction data or vacuum-based conformational analysis computations. Frequently, docking studies are used to find leads in a binding site taken from a crystal structure in which no or unrelated ligands had resided. The most probable relaxed solution ligand structure is more often ignored in rapid search computational models. It is a misstep in analytical models that aim to evaluate the relative binding of ligands in a common catalytic domain. This is particularly the case since the correlation between the low energy ligand solution conformation and the bound ligand conformation transmits information about its underlying molecular hydration and the physical nature of binding that brings about enzyme inhibition. This information about the relationship between ligand solution conformation and the bound ligand conformation provides strong evidence for the construction of pharmacophore models.9 We show here, for a small set of ligands binding MshB, that surveying the performance of substrate solution conformation inside the enzyme binding pocket is central to evaluating their relative inhibition.
Section snippets
Computational methods
Here we investigate the interplay between the MshB binding site and the preferred solution conformations of the series of ligands. We used a number of computational techniques such as molecular docking studies, molecular dynamics and relative free energy calculations to understand the relationship between the solution ligand conformation and the bound conformation as well as the catalytic site conformation.
The CHARMM33b2 program10 was used for all macromolecular simulations. The bonded and
Most probable conformations for inhibitors in solution
The preferred conformations of the inhibitor series in solution were explored using classical molecular dynamic simulations. The starting conformation of the thioglucoside linkage was chosen from the minima of an adiabatic map generated from simulated annealing of phenyl 2-acetamido-1-thioglucoside (GlcNAc-SPh) in vacuum (Supplementary data, Fig. S2). Low energy conformations were detected at ϕ = −40°. Within this valley there are two minima at ψ −100° and 100°, separated by a barrier of 2
Conclusions
Theoretical frameworks such as the lock-and-key and Koshland’s induced-fit model have long been used to justify and predict the conformations of ligands and associated host sites within proteins. In practice techniques such as docking are often employed to rapidly screen ligand binding with the protein, without due diligence paid to the energetically favoured solution conformations limiting the pose of ligand. The results of docking investigations for families of ligands are then used to rank
Acknowledgements
This work is based upon research supported by the South African Research Chairs Initiative (SARChI) of the Department of Science and Technology and National Research Foundation awarded to K.J.N.
References (32)
- et al.
Bioorg. Med. Chem.
(2010) J. Mol. Biol.
(1994)- et al.
J. Mol. Biol.
(2004) - et al.
J. Biol. Chem.
(2003) - et al.
Curr. Opin. Chem. Biol.
(2009) - et al.
J. Comp. Phys.
(1977) - et al.
Nat. Prod. Rep.
(2008) - et al.
Microbiol. Mol. Biol. Rev.
(2008) - et al.
PLoS Biol.
(2011) Biopolymers
(1995)
Ber. Dtsch. Chem.
Proc. Natl. Acad. Sci. U.S.A.
Science
J. Comp. Chem.
J. Phys. Chem. B.
J. Comp. Chem.
Cited by (9)
Synthesis and antimicrobial evaluation of amino sugar-based naphthoquinones and isoquinoline-5,8-diones and their halogenated compounds
2018, European Journal of Medicinal ChemistryCitation Excerpt :Carbohydrate-based antibiotics such as streptamine and epistreptamine [36] are representative members of this family. Carbohydrates are important as signaling molecules and for cellular recognition events, being explored for the development of new biologically active sugar-based compounds [37,38]. Based on our experience in the field of synthetically modified quinones [16–18,26] we present the synthesis of two classes of quinones containing amino sugar groups (12a-c and 15a-c), related halogen compounds (13a-c, 14a-c, 16a-c and 17a-c) and the antimicrobial evaluation of these substances against Gram-positive and Gram-negative bacteria of clinical importance (Fig. 2).
Functional characterisation of bile metagenome: Study of metagenomic dark matter
2021, MicroorganismsIsoniazid–phytochemical conjugation: A new approach for potent and less toxic anti-TB drug development
2020, Chemical Biology and Drug DesignFree Energy Methods in Drug Design: Prospects of "alchemical Perturbation" in Medicinal Chemistry
2018, Journal of Medicinal ChemistryGreat South African molecules: The case for mycothiol
2017, South African Journal of ChemistryProgress in the synthesis of mycothiol, its biosynthetic precursors and analogues
2016, Carbohydrate Chemistry