Exploring phenolics against penicillin-binding proteins as druggable targets in antibacterial therapy : a structure-activity relationship study

Abstract

Alterations in penicillin binding proteins (PBPs) coupled with inactivation of beta-lactam scaffold by beta-lactamase remain the major basis for the broad clinical resistance to beta– lactam antibiotics. While enzyme modification is most common with Gram-negative bacteria, alteration in key PBPs is prevalent in Gram-positive organisms and increasingly being reported in Gram-negative bacteria. Modification of PBPs ensures that the interaction with beta-lactams occurs at a considerably higher antibiotic doses than with PBPs from susceptible strains, thus substantially reducing biological activity. Consequently, developing potent therapeutics with a different scaffold against PBP druggable targets represents a viable alternative strategy against resistance to beta-lactams. Phenolics, aside having a different scaffold from conventional betalactams, have been reported to be effective against various multidrug-resistant Gram-negative and Gram-positive bacteria. The characteristic hydroxyl group of phenolics are implicated in strong PBPs binding while having the capability to auto-oxidize in the presence of transition metals to generating reactive oxygen species (ROS) which have been implicated in the lethality of most conventional antibiotics, including beta-lactams. These intriguing characteristics of phenolics position them as a viable source of drug candidates that can be developed as alternative beta-lactams. Hence, this study employed a multifaceted in silico and in vitro approach in bioprespection for phenolics as modulators of PBPs and ROS in an effort to manage resistance to beta-lactams due to PBP modifications. Using a structure-based drug design (SBDD) approach involving the use of pharmacophore screening and molecular docking, the over 10,000 currently known phenolics were screened at the active site of key PBPs of common and clinically important Gram-positive (Staphylococcus aureus and Streptococcus pneumoniae) and Gram-negative (Escherichia coli and Pseudomonas aeruginosa) bacteria. The identified top-20 phenolics against each of the investigated key PBPs (PBP2a of S. aureus, PBP2x of S. pneumoniae, PBP3 of P. aeruginosa and PBP3 and 5 of E. coli) had higher negative docking scores relative to the four beta-lactams (amoxicillin, cefotaxime, aztreonam and doripenem) reference standards suggesting effectiveness of the adopted SBDD approach. Subsequent filtering of the top-20 leads in each case based on their physicochemical, pharmacokinetic and toxicological profiles afforded topfive phenolics that are chemically diverse, orally bioavailable and with good synthetic feasibility scores (<5) appropriate for drug development. The profiled top-five compounds against the key PBPs in each case were further subjected to 120-ns molecular dynamic (MD) simulations using the in-house HEAL1361 programme with Amber 18 package of the CHPC (Centre for High Performance Computing), South Africa. The identified lead phenolics with the highest negative binding free energy (ΔGbind) in each case of the essential PBPs, which include silicristin (PBP2a active site: −25.61 kcal/mol and PBP2x allosteric site: -50.54 kcal/mol), epicatechin gallate (PBP2a allosteric site: −47.65 kcal/mol), lysidicichin ( PBP3 of E. coli: −41.66 kcal/mol), epigallocatechin 4-benzylthioether (PBP5 of E. coli −38.97 kcal/mol), catechin 3-rhamside (PBP3 of P. aeruginosa: -28.99 kcal/mol), and epicatechin 3- O-(3-O-methylgallate (PBP2x active site: -42.18 Kcal/mol), all had higher negative binding free energy than the reference standards. Except for silicristin with lesser stability [higher root means square deviation (RMSD)] against three of the essential PBPs [PBP2a (5.65 Å) of S. aureus, PBP3 (3.97 Å) of E. coli and PBP2x (4.90 Å), of S. pneumoniae], all the lead phenolics had better stability (< 3 Å), compactness and flexibility relative to the referenced standards and other top-five phenolics in each case of the investigated PBPs. However, silicrsitin remain the only lead phenolics with broad spectrum ΔGbind for all the essential PBPs, suggesting that the lesser stability of the compound with three of the essential PBPs could be improved to enhance stability. Moreso, as the focus of this study is to identify lead phenolic with wide spectrum activity against the investigated bacteria. Thus, for subsequent analysis, silicristin was selected for structural optimization against the essential PBPs for the possibility of yielding novel drug candidates with improved antibacterial efficiency, higher stability, reduced toxicity and favorable pharmacokinetic properties relative to silicristin. Improved thermodynamic stability has been demonstrated to be critical in affinity/potency of a drug for a target and its subsequent efficacy in vitro and in vivo. This study employed hybridization of pharmacophore and structural simplification techniques in the enhancement of thermodynamic stability of silicristin against the key PBPs of the investigated bacteria. The knowledge of the interactions of silicristin with the key PBPs of the organisms was employed in designing 25 novel derivatives of chalcone bis-coumarin from easily synthesizable starting materials of coumarin moieties. Following molecular docking, pharmacokinetic screening, MD simulation and energy refinement of the 25 derivatives, RD2c (PBP3 of E. coli: -46.26 kcal/mol), RD1c (PBP3 of P. aeruginosa: -39.42 kcal/mol and PBP2x active site: -42.90 kcal/mol), RD1a (PBP2a active site: -37.56 kcal/mol), D1a (PBP5 of E. coli: -37.19 kcal/mol), D1h (allosteric site of PBP2a: -53.09 kcal/mol and PBP2x of S. pneumoniae:-38.53 kcal/mol), were identified as best derivatives with enhanced ΔGbind, stability, and pharmacokinetics properties relative to silicristin. Quantum calculations using DFT B3LYP/6-31G+ (dp) techniques further showed that these derivatives are more reactive and with higher ability to form hydrogen bonds due to their higher electrostatic potential compared to silicristin. Except for PBP2x, this observation is consistent with the bond analysis of each complex post-MD simulation as the best derivative in each case of the PBP formed higher number of hydrogen bonds compared to silicristin. The exploration of catechol, resorcinol, and hydroquinone either alone or in combination with each other at either end of the derivatives promotes good pharmacological properties, while the presence of a pyrogallol and/or a benzene ring with more than two hydroxy group at one or both ends of the derivatives caused reduced oral bioavailability relative to silicristin. The synthetic accessibility of the derivatives was exploited by synthesizing the parent compound of the best derivative using Knoevenagel condensation, Vilsmeiers Haack, and Aldol condensation reactions. The successfully synthesized parent compound was subjected to nuclear magnetic resonance (NMR) for structural elucidation and was further validated using the gauge independent atomic orbital (GIAO) method. Expectedly, significant improvement in binding capacity and pharmacokinetic properties was observed with the best derivatives relative to silicristin; an observation that could translate to enhanced potency in vitro and in vivo. Silicristin is one of the bioactive compounds of silymarin, a standardized extract from the dried seeds of milk thistle which has been shown for its PBP modulatory properties in silico. Thus, to understand the mechanism of action of silicristin in vitro, this study investigates the antibacterial properties of silicristin and the synthesized parent compounds of the best derivatives of chalcone bis coumarin (PSCB) derived from silicristin against sensitive and resistant strains of Gram-positive (S. aureus and S. pneumoniae) and Gram-negative bacteria (E. coli and P. aeruginosa). Also, the possible involvement of ROS in the lethality of silicristin, PSCB and reference beta-lactams (amoxicillin against Gram-positive organisms and cefotaxime against Gram-negative bacteria) was also investigated. The minimum inhibitory concentration (MIC) assessment showed that silicristin, PSCB and reference beta-lactams had higher MIC values against the resistant strains [cefotaxime (2 – 16 µg/mL), amoxicillin (4 – 32 µg/mL, silicristin (8 – 64 µg/mL) and PSCB (16 – 128 µg/mL)]. The highest MIC for silicristin, PSCB and reference beta-lactams was observed against the resistant and sensitive strains of P. aeruginosa [sensitive strains (8 – 64 µg/mL), resistant strains (16 – 128 µg/mL)], while the lowest MIC was found in S. pneumoniae [sensitive strains (0.5 – 8 µg/mL), resistant strains (2 – 16 µg/mL)]. Compared to silicristin (4 – 64 µg/mL), PSCB (8 – 128 µg/mL) exhibited higher MIC values against all the test strains and the antibacterial activities were observed to occur in a concentration- and time-dependent manner. Except against the sensitive and resistant strains of E. coli where silicristin had a comparable killing effect as cefotaxime, silicristin had a better killing rate than PSCB and reference standards against all the test organisms with the best effect observed between 8 – 14 h after treatment. Except for the exploration of silicristin and amoxicillin against the resistant strain of E. coli, which was found to be addictive, the combinatorial therapy showed that silicristin had a synergistic effect with both amoxicillin and cefotaxime against all test organisms. This could mean that silicristin had other mechanism of antibacterial effect different from beta-lactams. In contrast, PSCB with cefotaxime and amoxicillin had only synergistic effects against resistant strains of Grampositive organisms suggestive of a similar mechanism of action for PSCB and the test betalactams. A further probe into the ROS generating potential of the antibacterial agents using ROS-specific dyes, showed that while PSCB and reference beta-lactams caused significant (P < 0.05) higher generation of superoxide ion (O2 •−), silicristin conversely enhanced higher (P < 0.05) generation of the most toxic ROS form, the hydroxy radicals ( •OH), which is often produced via the Fenton reaction. This observation with silicristin suggests its enhanced modulatory effect on Fenton reaction relative to PSCB and the standards and further facilitated increased NAD/NADH ratio, ADP/ATP ratio, lipid peroxidation, carbonyl content, propidium iodide (PI) uptake, and depletion of reduced glutathione as observed with silicristin-treated cells. This observation points to the enhanced involvement of oxidative stress in the modulatory effect of silicristin against all the test organisms compared to PSCB and the reference standards. This enhanced oxidative stress in silicristin-treated cells was substantiated with scanning electron microscopy (SEM) and transmission electron microscopy (TEM) micrographs with cell walls with more blisters, eruptions, and sores compared to the observed effects with other treatments and the untreated cells. Oxidative stress was also noted to facilitate SOS (Save Our Souls) response in upregulating SordA and SoxR, with the most prominent effect observed with silicristin in a manner that is more evident in the sensitive strains relative to the resistant strains. Taken together, these observations points to the enhanced redox effect of silicristin on the sensitive and resistant strains of both Gram-positive and Gram-negative organisms relative to PSCB and reference beta-lactams and could be explored to develop novel alternative therapeutics to beta-lactams in the treatment of infections caused by the test organisms. Additionally, since the best derivative in each case of the essential PBPs had a similar synthetic route as PSCB, it is recommended that the synthetic route be adopted in the synthesis of the best derivatives to enable the investigation of their in vitro and in vivo antibacterial activities for the possibility of subsequent development as beta-lactams.