Methicillin-Resistant Staphylococcus aureus: Docking-Based Virtual Screening and Molecular Dynamics Simulations to Identify Potential Penicillin-Binding Protein 2a Inhibitors from Natural Flavonoids

Staphylococcus aureus (S. aureus) is responsible for several disorders including skin and soft tissue infections, bacteremia, pulmonary infections, septic arthritis, osteomyelitis, meningitis, gastroenteritis, toxic-shock syndrome, and urinary tract infections. Methicillin-resistant S. aureus (MRSA) contains penicillin-binding protein 2a (SauPBP2a) responsible for catalyzing the peptidoglycan production within the bacterial cell wall. The binding affinity of SauPBP2a to beta-lactam antibiotics is low, and thus, it is necessary to discover new effective SauPBP2a inhibitors to combat mortality and morbidity in patients affected by MRSA. The binding affinity of 46 natural flavonoids to the SauPBP2a active site was examined via molecular docking analysis. The stability of docked poses associated with the top-ranked flavonoids was tested by performing molecular dynamics (MD) in 10 nanoseconds (ns) computer simulations. Kaempferol 3-rutinoside-7-sophoroside and rutin demonstrated a considerable binding affinity to the SauPBP2a active site (ΔGbinding < −11 kcal/mol). Their docked poses were found to be stable for 10 ns MD simulations. Kaempferol 3-rutinoside-7-sophoroside and rutin also exhibited salient binding affinity to the enzyme's allosteric site. This study suggests that kaempferol 3-rutinoside-7-sophoroside and rutin may be considered as drug candidates for therapeutic aims in several human infections associated with MRSA. Nevertheless, in vitro and in vivo confirmations are warranted.


Introduction
Staphylococcus aureus (S. aureus), a Gram-positive bacteria, is the leading cause of severe invasive and toxin-mediated disorders including skin and soft tissue infections, bacteremia, pulmonary infections, septic arthritis, osteomyelitis, meningitis, gastroenteritis, toxic-shock syndrome, and urinary tract infections. Methicillin-resistant Staphylococcus aureus (MRSA) refers to any strains of S. aureus that carry the mecA gene encoding penicillin-binding protein 2a (PBP2a), responsible for catalyzing the peptidoglycan production within the bacterial cell wall. As PBP2a has a lower binding affinity to beta-lactam-containing antibiotics compared to other PBPs, MRSA continues to catalyze bacterial cell wall synthesis in the presence of penicillinderived antibiotics such as methicillin, oxacillin, nafcillin, and cephalosporins [1][2][3][4][5][6][7].
us, PBP2a in S. aureus (SauPBP2a) has been considered a potential target for developing specific inhibitors to combat MRSA, which will be accelerated by analyzing the three-dimensional structure of the SauPBP2a active site [8]. Furthermore, the allosteric site of SauPBP2a has also been considered for negative regulation of the enzyme's activity [9].
Organic compounds with antibacterial properties have sparked public interest in recent years for treating infections [10]. Flavonoids are one of the most important active compounds that naturally occur in photosynthesizing plants with several valuable properties in humans [11,12]. Fruits and vegetables containing flavonoids have been widely used to treat human disorders [13]. Several flavonoids have been shown to possess antibacterial properties in previous research. In this regard, chrysin has been demonstrated to have bacteriostatic activity against Escherichia coli (E. coli) and Pseudomonas aeruginosa (P. aeruginosa).
In the present study, we hypothesized that inhibition of PBP2a is a possible mechanism flavonoids use to combat MRSA growth.
us, a docking-based virtual screening study was executed to examine the binding affinity of several natural flavonoids to the SauPBP2a active site. Molecular dynamics (MD) simulations were conducted to evaluate the stability of docked poses of top-ranked SauPBP2a active site inhibitors. e binding affinity of salient SauPBP2a active site inhibitors to the enzyme's allosteric site was also tested.

SauPBP2a and Preparation of Flavonoids' Structures.
e three-dimensional coordinate of SauPBP2a was obtained from the Protein Data Bank (PDB) database, which is available at https://www.rcsb.org. Two PDB codes including 1MWT (X-ray resolution � 2.45Å [8,16]) and 4CJN (X-ray resolution � 1.95Å [17]) were selected for docking analyses with the active and allosteric sites of the enzyme, respectively. e 1MWT and 4CJN files included two polypeptide chains as A and B. Chain B was selected for further analysis, which contained 646 and 642 residues in the 1MWT and 4CJN files, respectively. e chain B also included penicillin G (PDB ID: PNM) and (E)-3-(2-(4-cyanostyryl)-4-oxoquinazolin-3(4H)-yl) benzoic acid (PDB ID: QNZ) in the 1MWT and 4CJN PDB files, respectively; these ligands were removed from the SauPBP2a structure before molecular docking and dynamics studies.
Energy minimizing (EM) of SauPBP2a with two different codes was carried out before docking operations using the Swiss-PDB Viewer version 4.1.0, which is available at https:// spdbv.unil.ch/ [18]. A total of 46 natural flavonoids were docked with the active site of SauPBP2a to identify potential inhibitors of the enzyme's catalytic domain. Next, molecular docking analysis was done using the known PBP inhibitors, including penicillin G and methicillin. Due to the low binding affinity of SauPBP2a to the beta-lactam-containing antibiotics, penicillin G and methicillin were considered control negative in this study. Furthermore, previous studies have reported that "oxadiazole" [19] and "ceftobiprole" [20] had exhibited high binding affinity to the SauPBP2a active site and hence were considered control positive in our work. e two-dimensional (2D) coordinates of flavonoids, methicillin, and ceftobiprole were principally collected from the PubChem database (https://pubchem.ncbi.nlm.nih.gov) as a structure data file (SDF). Subsequently, the Cactus web server was used to achieve PDB formats from the SDF files, available at https://cactus.nci.nih.gov/translate/. e threedimensional structure of penicillin G was collected from the 1MWT file, while the construction of oxadiazole was drawn by the ChemDraw software version 12.0.2.1076. e EM was administered for all small molecules using the HyperChem software version 8.0.10 [21].

Molecular Docking with the SauPBP2a Active Site.
A windows-based program system was used with the following features: system type, 64 bit; processor, Intel Core i7; installed memory, 32 GB. Docking-based virtual screening was executed using AutoDock 4.0 (https://autodock.scripps.edu) [22]. e AutoDock software applies limited flexibility in the receptor and uses the Lamarckian genetic algorithm to predict the docked pose of the ligand [23]. e SauPBP2a catalytic site was considered the receptor for executing docking analyses with flavonoids, control negative, and control positive compounds. After analyzing the interactions between nitrocefin, penicillin G, methicillin, and the residues within the SauPBP2a catalytic domain from the study by Lim and Strynadka [8], a total of 14 main residues were identified inside the PBP2a active site including Ser337, Lys340, Ser403, Lys406, Tyr446, Ser462, Asn464, r500, Ser548, Gly549, Ser598, Gly599, r600, and Met681. us, these amino acids were considered as a receptor site for docking analysis to identify potential inhibitors of the enzyme's active site.
e grid box settings were as follows: X-dimension, 60; Y-dimension, 66; Z-dimension, 60; X-center, −34.264; Y-center, 44.65; Z-center, 66.778; spacing, 0.375Å. e number of docked poses for each component was set at 50, where the docked model with the lowest ∆G binding involved in the largest cluster along with the largest number of models using a root mean square deviation (RMSD), tolerance of 2.0Å, was considered as a binding affinity of the ligand to the receptor site.

Cross-Validation Study.
Cross-validation of docking studies was executed for top-ranked SauPBP2a active site inhibitors based on the ΔG binding values achieved from the AutoDock tool (ΔG binding < −10 kcal/mol).
is was performed using the Schrödinger Maestro docking software version 10.2 [24,25]. In this regard, docking scores were calculated, and the prime MM-GBSA method was used to determine the relative binding energies of ligands. e Glide docking system was employed to evaluate the binding affinity of compounds to the SauPBP2a active site. e lowest Glide score (dock score) for each ligand was considered the best-docked model. More details regarding settings in the Schrödinger Maestro docking software have been reported in a previous study by Azadian et al. [26].

MD Simulations.
To examine the stability of docked poses among top-ranked flavonoids and the SauPBP2a active sites, MD was simulated by applying the Discovery Studio Client software (version 16.1.0.15350). Only ligands that demonstrated considerable affinity to the enzyme's active site using the AutoDock tool and Schrödinger docking software were considered for MD analysis. e advanced settings for computer simulations were as follows: MD time: 10 nanoseconds (10 ns), solvation model: explicit periodic boundary; cell shape: orthorhombic; minimum distance from the boundary: 10Å; solvent: water; target temperature: 310 K; force field: CHARMm; charge distribution: point. In addition, the time evolution of root mean square deviation (RMSD) of backbone atoms and root mean square fluctuation (RMSF) of SauPBP2a complexed with the best inhibitors was also analyzed. e BIOVIA Discovery Studio Visualizer version 19.1.0.18287 was then utilized to demonstrate the interactions among top-ranked flavonoids and the residues inside the SauPBP2a catalytic site and to superimpose protein-ligand complexes before and after MD simulations.

Molecular Docking with the SauPBP2a Allosteric Site.
is study considered the best SauPBP2a active site inhibitors to evaluate their binding affinity to the enzyme's allosteric site. e grid box options at this stage followed the settings of the study by Ibrahim et al. [9]: X-dimension: 52; Y-dimension: 52; Z-dimension: 52; X-center: 9.658; Ycenter: −1.662; Z-center: −70.269; spacing: 0.375Å.

Molecular
Docking with the SauPBP2a Active Site. It was found that 9 of the flavonoids can potentially bind to the SauPBP2a catalytic site on the nanomolar scale (nM). A total of 37 compounds revealed an inhibitory effect on SauPBP2a active site at the micromolar (uM) scale. Four components, including kaempferol 3-rutinoside-7-sophoroside, rutin, amentoflavone, and quercetin, bounded to the SauPBP2a active site with the salient affinity of ∆G binding < −10 kcal/mol (Supplementary Table 1). us, these compounds found the top SauPBP2a active site inhibitors among the tested flavonoids based on the AutoDock tool and were considered for further analyses using the Schrödinger Maestro docking software. Ceftobiprole and oxadiazole exhibited a mild binding affinity to the enzyme's active site with the ∆G binding values of −8.79 and −7.87 kcal/mol, respectively. Figure 1 shows the two-dimensional structures of kaempferol 3rutinoside-7-sophoroside, rutin, amentoflavone, quercetin, and control negative and positive compounds in this study, which were achieved by the ChemDraw Ultra version 12.0.2.1076. Of note, the ∆G binding between penicillin G as well as methicillin and SauPBP2a active site was calculated to be −6.35 and −5.51 kcal/mol, respectively. Figure 2 shows the ∆G binding values between top-ranked flavonoids, control inhibitors, and the SauPBP2a active site. In comparison, the details of energies between top-ranked flavonoids, control inhibitors, and the enzyme's catalytic domain are given in Table 1. All hydrogen, hydrophobic, electrostatic, and miscellaneous interactions between kaempferol 3-rutinoside-7-sophoroside, rutin, amentoflavone, quercetin, and residues inside the SauPBP2a catalytic site were analyzed before and after MD simulations (Table 2 and Figure 3). In this regard, kaempferol 3-rutinoside-7-sophoroside revealed the most H-bonds before and after MD analyses. e H-bonds with a distance of >5Å were not considered significant interactions. Furthermore, Figure 4 shows the interaction modes between control compounds and the residues inside the enzyme's active site. All possible interactions between kaempferol 3-rutinoside-7-sophoroside, rutin, amentoflavone, quercetin, and the residues within the SauPBP2a catalytic site are shown in Figure 5(a) as a single graph.
is network was achieved using the Cytoscape software (3.8.0; https://www.cytoscape.org) [27]. Subsequently, the degree of each residue was calculated using the Cytoscape network analyzer tool, illustrating the number of interactions between each residue and the top-ranked flavonoids. Lys430 indicated the highest degree between the residues ( Figure 5(b)).

3.2.
Cross-Validation and MM-GBSA Analyses. Schrödinger Maestro docking scores were calculated for kaempferol 3-rutinoside-7-sophoroside, rutin, amentoflavone, and quercetin. Kaempferol 3-rutinoside-7-sophoroside revealed the highest binding affinity to the active site of SauPBP2a, followed by rutin with the dock scores of −12.697 and −9.341 kcal/mol, respectively (Table 3). us, these two compounds were considered salient inhibitors of the SauPBP2a active site and were selected for further MD analysis. Meanwhile, MD simulations were performed based on the complexes achieved from the AutoDock software. Furthermore, the interaction modes between kaempferol 3rutinoside-7-sophoroside, rutin, and residues within the sauPBP2a active site were also studied based on the complexes achieved from the Schrödinger Maestro docking tool ( Table 2, Figures 3(g) and 3(h)). e prime MM-GBSA analysis showed each compound's relative binding-free energy (ΔG bind), with the results given in Table 4. e expanded formula is given as follows [28]: ΔG (bind) � ΔG (solv) + ΔE (MM) + ΔG (SA).

MD Simulations.
MD simulation was done to assess the stability of docked poses of kaempferol 3-rutinoside-7sophoroside and rutin. Figure 6 shows the superimposed structures of SauPBP2a complexed with kaempferol 3rutinoside-7-sophoroside and rutin before and after MD simulations. Furthermore, the RMSD of backbone atoms and RMSF for SauPBP2a complexed with kaempferol 3rutinoside-7-sophoroside and rutin are shown in Figure 7.   As shown in Figure 7, it seems that the helix structures are more stable than the other secondary structures in SauPBP2a.

Molecular Docking with the SauPBP2a Allosteric Site.
e results showed that the potential inhibitors of SauPBP2a active site including kaempferol 3-rutinoside-7-sophoroside and rutin also had an excellent binding affinity to the allosteric site of the enzyme with the criteria of ΔG binding values as −14.35 and −9.68 kcal/mol, respectively (Table 1).

Discussion
Antibiotic resistance has emerged in response to the overuse of currently available antibiotics [10]. us, it is necessary to discover and design novel antibacterial compounds with new formulations to overcome serious infections [29]. MRSA infection is one of the most frequent causes of hospital-acquired conditions and is currently associated with poor prognosis and increased mortality/morbidity [30]. In the present study, molecular docking and dynamics simulations were performed to identify potential inhibitors  Kaempferol is a flavonoid compound with various health benefits such as anti-inflammatory, anticancer, liver-protective, antiobesity, antidiabetes, and heart-protective  International Journal of Microbiology effects. It is found in fruits and vegetables, tea, beans, cabbage, broccoli, barriers, grapes, apples, and citrus fruits [31]. Previous research has shown that kaempferol and its many derivatives have antimicrobial properties. Qui et al. [32] demonstrated that two of the kaempferol derivatives  Rutin is an organic flavonol with several pharmacological features such as antioxidant and antibacterial properties [34][35][36][37]. In a previous report, a combination of rutin and florfenicol exhibited enhanced antibacterial activity against multidrug-resistant (MDR) Aeromonas hydrophila, suggesting that the use of organic compounds in combination with traditional antibiotics may result in synergistic inhibition of the growth of MDR bacteria [29]. Recently, Rodríguez-Valdovinos and Salgado-Garciglia [38] designed a study to examine the antibacterial and antioxidant effects of Verbesina sphaerocephala (V. sphaerocephala) leaf and flower extracts, in which rutin is the main phenolic component. Rodríguez-Valdovinos and Salgado-Garciglia [38] reported that V. sphaerocephala extracts demonstrated significant antibacterial effects on S. aureus and E. coli. Our AutoDock results found that rutin can bind to the SauPBPa2 catalytic domain at the nanomolar concentration (Ki � 436.10 nM) with an estimated Gibbs free energy of −11.14 kcal/mol. Furthermore, the Schrödinger Maestro docking software estimated that the dock score between rutin and SauPBP2a active site was −9.341 kcal/mol. Before MD simulations, rutin formed 3 hydrogen and 1 hydrophobic interaction with Lys430, Tyr446, and r600 within the SauPBPa2 active site. is flavonoid showed 2 hydrogen and 3 hydrophobic interactions with Tyr444, Arg445, and Ser643 inside the SauPBP2a active site after 10 ns MD simulation.
Amentoflavone is a bioflavonoid metabolite with miscellaneous beneficial properties including anti-inflammatory, neuroprotective, antiviral, and anticancer effects [39][40][41][42]. It has been widely purified from Hypericum perforatum, Ginkgo biloba, and Selaginella tamariscina [43]. Hwang et al. [44] performed a study to evaluate the antibacterial effects of amentoflavone alone, isolated from Selaginella tamariscina, and in combination with other antibiotics including chloramphenicol, cefotaxime, and ampicillin. e authors reported synergistic results of combined amentoflavone with other traditional antibiotics on various Gram-positive and Gram-negative bacteria including S. aureus, E. coli, Enterococcus faecium, and Pseudomonas aeruginosa (P. aeruginosa). is was carried out by applying the Clinical and Laboratory Standards Institute guideline [45]. e authors suggested that the enhanced production of hydroxyl radicals by amentoflavone is potentially the main reason for the synergistic outcome. Our results found that amentoflavone can bind to the SauPBPa2 active site at the nanomolar concentration (Ki � 31.63 nM) and with an estimated free binding energy of −10.23 kcal/ mol. Our docking results indicated that amentoflavone revealed 3 hydrogen, 4 hydrophobic, 2 electrostatic, and 2 miscellaneous interactions with Lys430, r444, Tyr446, Val448, Asn464, r600, Glu602, and Met641 within the SauPBPa2 active site.
Quercetin is predominantly found in vegetables and fruits such as apples, broccoli, berries, onions, kales, tee, and red grape. Quercetin has several valuable characteristics, including antioxidant, anti-inflammatory, anticarcinogenic, and antiviral properties. It also reduces the risk of infectious diseases and promotes mitochondrial biogenesis [46][47][48][49][50][51]. Wang et al. [52] performed in vivo and in vitro studies to examine the antibacterial effects of quercetin on cecal microbiota of Arbor Acre broiler chickens, as well as its mechanism of action. Wang et al. [52] demonstrated that quercetin significantly reduced the total amount of S. aureus (P < 0.01), P. aeruginosa (P < 0.05), Salmonella enterica serotype Typhimurium (P < 0.01), and E. coli (P < 0.01) compared to the negative control. Furthermore, the authors provided the in vitro clue that quercetin damages the membrane and cell walls of S. aureus and E. coli, indicating that it could be used as an alternative antibacterial feed for animals. Based on our findings, quercetin revealed 1 hydrogen and 6 hydrophobic interactions with Lys430, Val443, Tyr446, Tyr519, and Met641 within the PBPa2 active site.   e stability of docked poses of kaempferol 3-rutinoside-7-sophoroside and rutin was studied by performing MD in 10 ns computer simulations. It was suggested that the docked poses of these 2 flavonoids were stable in 10 ns simulation. In addition, kaempferol 3-rutinoside-7-sophoroside and rutin also revealed considerable binding affinity to the allosteric site of the enzyme. us, these compounds may be considered effective inhibitors of SauPBP2a, though this needs in vitro and in vivo confirmations in future studies. e present study had a principal limitation. MD is a computer simulation approach analyzing the physical movements of atoms and molecules in an environment similar to reality. us, a supercomputer with a powerful processor and random access memory (RAM) is an indispensable tool for executing these types of time-consuming procedures. Unfortunately, the supercomputer was unavailable for our research team to run MD simulations for more significant periods. Each MD simulation (10 ns) took   VAL45  THR63  GLN81  THR103  GLY121  ASP139  ARG157  PRO175  SER193  PHE211  LYS229  LYS247  LYS265  LEU283  ILE301  LYS319  TYR337  GLY355  TYR373  LEU391  THR409  ILE427  ARG445  ASP463  LYS481  TYR499  GLN521  GLU539  VAL557  MET575  ASN593  SER619  GLN637  about one month with the computer we had. We strongly believe that simulating MD for a greater period of times results in more reliable results.

Conclusion
46 flavonoids were docked to the SauPBP2a active site using the AutoDock tool to estimate the binding affinity of the selected herbal flavonoids to the target protein. A crossvalidation study was performed for compounds with the criteria of ΔG binding < −10 kcal/mol. Next, MD simulation was executed for compounds that exhibited considerable binding affinity to the enzyme's active site using AutoDock and Schrödinger Maestro docking tools. e results suggest that kaempferol 3-rutinoside-7-sophoroside and rutin have a high affinity of binding to the SauPBP2a active site. eir docked poses were stable in the 10 ns simulation. ey also exhibited excellent binding affinity to the enzyme's allosteric site. us, it may be speculated that kaempferol 3-rutinoside-7-sophoroside and rutin could be helpful in the therapeutic aims of MRSA infections. Nevertheless, further in vivo and in vitro inhibition experiments are warranted in the future.

Consent
Not applicable.

Conflicts of Interest
e authors declare that they have no conflicts of interest.