Identification of potential inhibitors against Staphylococcus aureus shikimate dehydrogenase through virtual screening and susceptibility test

Abstract

Staphylococcus aureus shikimate dehydrogenase (SaSDH) plays a crucial role in the growth of Staphylococcus aureus (S. aureus), but absent in mammals and therefore a potential target for antibacterial drugs to treat drug-resistant S. aureus infection. In this study, a 3D model of SaSDH was constructed by homology modelling and inhibitors of SaSDH were screened through virtual screening. (-)-Gallocatechin gallate and rhodiosin were identified as inhibitors with K_is of 2.47 μM and 73.38 μM, respectively. Molecular docking and isothermal titration calorimetry showed that both inhibitors interact with SaSDH with a K_D of 44.65 μM for (-)-gallocatechin gallate and 16.45 μM for rhodiosin. Both inhibitors had antibacterial activity, showing MICs of 50 μg/mL for (-)-gallocatechin gallate and 250 μg/mL for rhodiosin against S. aureus. The current findings have the potential for identification of drugs to treat S. aureus infections by targeting SaSDH.

Keywords:

• Shikimate dehydrogenase
• Staphylococcus aureus
• virtual screening
• SaSDH inhibitors

Introduction

Antimicrobial resistance (AMR) threatens public health and methicillin-resistant Staphylococcus aureus (MRSA) is on the WHO list of antibiotic-resistant "priority pathogens". The average MRSA infection rate among 76 countries was about 35% in 2021, an increase of 15% from 2017. Resistance to the first-line drug, vancomycin, and to the new-generation drugs, linezolid and daptomycin has been reported^1–3. There is an urgent need to develop anti-MRSA drugs using innovative approaches to overcome resistance.

The shikimate pathway is responsible for the synthesis of aromatic amino acids, flavonoids and other aromatic compounds in plants and microorganisms but is absent from mammals⁴. The pathway begins with erythrose phosphate and phosphoenolpyruvate, eventually synthesises chorismite through seven enzymatically catalysed steps by 3-deoxyarabinoheptanulose-7-phosphate synthetase, dehydroquinic acid synthetase, 3-dehydroquinic acid dehydratase, shikimate dehydrogenase, shikimate kinase, 5-enolpyruvyl shikimate-3-phosphate synthetase and chorismite synthetase. Knockout of genes encoding these enzymes decreases bacterial proliferation and virulence^5,6. Enzymes in shikimate pathway are potential new antimicrobial targets. Shikimate dehydrogenase (SDH) converts 3-dehydroshikimate to shikimate (SHK) in the 4^th step. The SDH protein structure includes an α/ß fold which is highly conserved among plants and microbes with subtle variations at the active site⁷. SDH is vital for bacterial survival and SDH inhibitors^8–11 are attractive antimicrobial targets.

Staphylococcus aureus shikimate dehydrogenase (SaSDH) is a 29 kDa monomer for which antibacterial inhibitors have yet to be identified. SaSDH was expressed and purified during the current work and high-throughput virtual screening against a structurally diverse, commercially available compound library performed to identify SaSDH inhibitors. Antimicrobial potency was determined using an in vitro susceptibility test and binding mode investigated.

Materials and methods

SaSDH expression and purification

The SaSDH gene was amplified from the genomic DNA of Staphylococcus aureus ATCC49775 strain. Primers were F:5′-CTCGGATCCATGAAATTTGCAGTTA-3′ and R:5′-CCGCTCGAGTTATTCTCCTTTTAA-3′ with underlining representing the restriction sites for XHoI and BamHI. The PCR product was cloned into the pET-28a(+) vector (Youbio, China) and the recombinant plasmid transfected into E. coli BL21(DE3)pLysS cells (TianGen, China). Recombinant bacteria were cultured in 500 ml LB medium containing 50 μg/mL kanamycin at 37 °C until the absorbance at 600 nm reached 0.6–0.8 when protein expression was induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 25 °C for 8 h. Cells were harvested by centrifugation at 8,000 g for 5 min at 4 °C, resuspended in 20 mM Tris-HCI, pH 8.0 containing 300 mM NaCI (Buffer A) and lysed by 800 bar high pressure. Cell debris was removed by centrifugation at 10,000 g for 30 min and the supernatant filtered through a 0.45 μm membrane. SaSDH was eluted over a 10–300 mM imidizole gradient in Buffer A through a Ni-NTA affinity column (Cytiva, America) using AKTA pure system (GE, America).

Enzyme activity and kinetic parameters

SaSDH activity was measured as previously reported¹². Briefly, the reaction medium contained 2 mM NADP, 8 mM SHK and 1.54 ng purified recombinant SaSDH in 100 mM Tris-HCI, pH 9.0, at 37 °C and was initiated by SaSDH addition and absorbance monitored at 340 nm every minute for 60 min using SpectraMax Paradigm (Molecular Devices, America). Vmax and Km for NADP and SKH were determined by varying the concentration of one substrate, 0.4, 0.5, 1, 2, 4 mM NADP and 8, 4, 2, 1, 0.5, 0.25 mM SKH, while the other substrate was kept at constant saturation. Kinetic data were analysed by double reciprocal plots.

Homology modelling and virtual screening

Homology modelling was performed using SWISS-MODEL referring to the crystal structure of Staphylococcus epidermidis shikimate dehydrogenase (SeSDH: PDB ID:3DON) which has 70% sequence homology with SaSDH¹³. The SaSDH model was optimised by the Protein Preparation Wizard module in Schrödinger Maestro 11.4. The amino acid residues, SER13/SER15/THR60/LYS64/ASN85/ASP100/PHE236, form the binding site of SaSDH¹⁴ and inhibitors were identified by molecular docking in Glide module in Schrödinger Maestro 11.4 using compounds from the Life Chemicals HTS Compound Collection, Specs HTS Compounds Library and MCE Bioactive Compound Library PlusA. Three-dimensional low-energy conformations of compounds and atomic partial charges were obtained using the LigPrep Module and compounds were selected according to docking scores.

Enzyme inhibition

Enzyme inhibition potency of inhibitors was measured, as above, in a medium containing 1 mM NADP, 1 mM SHK, 1.54 ng SaSDH and inhibitor of different concentration in 100 mM Tris-HCI, pH 9.0, at 37 °C and absorbance at 340 nm measured at t = 0 and t = 30 min. Negative controls lacking inhibitor compounds were included. Inhibition was calculated according to the following equation: (1) $$\begin{array}{c}V=\frac{T{30}_{\mathit{\text{Absorption}}}-T{0}_{\mathit{\text{Absorption}}}}{T30-T0}\\ \end{array}$$(1) (2) $$\mathit{\text{Inhibition}}\left(\mathrm{\%}\right)=\left(1-\frac{V_{\mathit{\text{Sample}}}}{V_{\mathit{\text{Negative control}}}}\right)\times 100\mathrm{\%}$$(2)

The 50% inhibition concentration(IC₅₀) was calculated using GraphPad prism. Inhibition constants (Ki) were measured with 4, 2, 1, 0.5 and 0.25 mM SHK and 3 inhibitors concentrations with absorbance at 340 nm measured at 1 min intervals for 60 min. Kis were calculated from Lineweaver-Burke plots using GraphPad prism.

Isothermal titration calorimetry (ITC)

ITC was performed in 100 mM Tris-HCI, containing inhibitors, pH 9.0, at 25 °C with constant stirring at 200 rpm on a MicroCal iTC200 system (MicroCal, England), as previously reported¹⁵. The cell contained 100 μM compound and the syringe 1000 μM SaSDH. 5 μL SaSDH was injected into the cell after equilibration to trigger binding. Data analysis, including baseline correction and evaluation, was carried out with Nano ITC and all SaSDH injections fitted, except the first, to calculate maximum heat and the equilibrium dissociation constant (K_D).

S. aureus susceptibility

The minimum inhibitory concentration (MIC) for S. aureus was determined by micro-broth two-fold serial dilution. Compounds were dissolved in DMSO and diluted with MH broth medium in 96-well microtiter plates before the addition of 100 μL bacterial suspension (1 × 10⁶ CFU/ml) and incubation at 37 °C for 16 h. The negative control was compound-free and the positive control included ampicillin.

Growth curves were measured by incubating 1 × 10⁷CFU/ml bacterial suspension at 37 °C and adding the MIC of inhibitor. Absorbance at 600 nm was read for 200 μL sample at 2 h intervals and time-kill curves plotted.

Scanning electron microscopy (SEM)

Overnight S. aureus culture was diluted to 10⁷ CFU/ml and incubated with MIC and 1/2 MIC drug concentrations for 1.5 h before 10 μL samples were settled on a cationic adsorption membrane for 30 min and the medium aspirated. 10 μL 2.5% glutaraldehyde was added and fixed for 30 min, the upper fixative discarded and the membrane immersed in the fixative for 12 h. The membrane was washed twice with 10xPBS, dehydrated by ethanol gradient, dried, sprayed with gold and bacterial morphology observed by SEM (FEI Quanta x50 FEG, America).

Cell membrane integrity

Overnight S. aureus culture was diluted to 10⁷ CFU/ml and incubated until the logarithmic phase was reached. Drugs were added to a final concentration of MIC with incubation at 37 °C for 3 h before washing 3x with PBS, the addition of Calcein AM/PI and incubation for 15 min in the dark. Samples were observed under the fluorescence microscope (Leica DMi8, Germany).

Results

SaSDH expression and purification

Recombinant SaSDH with a 6His-tag at its N-terminus was expressed in E. coli BL21 (DE3) pLysS induced by 0.5 mM IPTG. The medium was purified by His TrapTM ^FF column using AKTA pure system and SaSDH eluted with imidazole. The SaSDH elution peak was seen at 100 mM imidazole (Figure 1A). A protein of molecular weight 29 kD was seen after SDS-PAGE (Figure 1B), consistent with SaSDH predicted weight, indicating the presence of recombinant SaSDH with high purity.

Figure 1. SaSDH purification. (A) Chromatogram from AKTA pure. (B) SDS-PAGE: lane 1: bacterial medium; lane 2: bacterial precipitate; lane 3: sample flow through column; lane 4 and lane 5: SaSDH eluted with 100 mM imidazole; lane 6: marker.

Enzyme activity and kinetic parameters

SaSDH activity was assayed in the reverse catalytic direction producing NADP⁺-dependent oxidation of SHK to form NADPH and 3-dehydroshikimate, accompanied by increased absorbance at 340 nm until a plateau at 60 min. No absorbance change was seen in the control assay lacking the enzyme (Figure 2A). One unit (U) of SaSDH activity was defined as the amount of enzyme required to catalyse the production of 1 μM NADPH per minute in a 100 μL mixture at 37 °C. Purified SaSDH had an activity of 913 KU/mg and a hyperbolic plot was produced, indicating that Michaelis-Menten kinetics were obeyed (Figure 2B). The Km for SHK was 0.04614 mM and Vmax 3.362 mM/min/μg. The Km for NADP was 0.09816 mM and Vmax 3.785 mM/min/μg (Table 1).

Figure 2. (A) SaSDH activity. Activity was determined by continuous monitoring of the increase of NADPH concentration at 340 nm. (B) Reciprocal plots of initial velocities under conditions where either NADP or SHK was varied at different fixed concentrations of the co-substrate.

Table 1. Apparent kinetics parameters for SaSDH.

Substrate	Vmax (mM/min/μg)	Km (mM)
NADP^+	3.785	0.09816
SHK	3.362	0.04614

Homology modelling and virtual screening

A 3D SaSDH model was generated using SeSDH as template by homology modelling for virtual screening (Figure 3A). The stereochemical quality of the model was revisited and validated by two different programs. Only 0.4% of the residues were in the disallowed regions in Ramachandran plot which is used to reflect whether the protein is conformationally sound or not, 3.8% were in the allowed region and 95.8% in the favoured region of the plot (Figure 3B). Q-MEAN reported a normalised QMEAN4 score of 0.84 (close to 1 is the ideal) and a Z-score of −0.85 (close to zero is the ideal) (Figure 3B). Virtual screening was performed using a library of 720,000 small molecule compounds from three different commercial compound libraries. Three rounds of screening, involving high-throughput screening pattern, standard screening pattern and high-precision screening pattern, enabled the identification of the 200 compounds with the highest docking scores.

Figure 3. Homology modelling. (A) 3D model of SaSDH. (B) Ramachandran plot obtained from model protein geometry evaluation (dark green, light green, and green correspond to the favoured, allowed and outlier regions respectively). (C) Normalised QMEAN4 score graphic showing the position (red cross) of the model in the set of PDB structures used for the evaluation and the Z-score value.

Twelve compounds were selected from the MCE Bioactive Compound Library PlusA for inhibition assays. The docking scores and IC₅₀s of these compounds against SaSDH were showed in Table 2. Pentagalloylglucose, (-) - gallocatechin gallate, troxerutin, specneuzhenide, rhodiosin and hesperidin all had IC₅₀ less than 100 μM and (-)-gallocatechin gallate showed strongest inhibitory potency against SaSDH with IC₅₀ was 0.85 μM.

Table 2. Inhibition of SaSDH and S. aureus MICs for 12 compounds.

Compound	Molecular Weight	Docking Score	IC50 (μM)	MIC (μg/mL)
Pentagalloylglucose	940.68	−16.751	3.168	100
(-)-Gallocatechin gallate	458.37	−11.448	0.85	50
Troxerutin	742.68	−12.03	69.36	>1024
Rhodiosin	610.52	−12.929	91.2	250
Specneuzhenide	686.65	−12.691	72.82	>100
Davunetide	824.92	−9.282	198.32	>200
Epitalon (TFA)	504.37	−10.154	1097.01	>100
Hesperidin	610.56	−10.798	91.8	>100
SennosideD	848.76	−12.124	217.73	>250
Isoquercetin	464.38	−11.09	301.05	>100
Narirutin	580.53	−10.2	197.72	>250
Miquelianin	478.36	−10.964	141.02	>250

MICs for S. aureus were determined for these compounds (Table 2). (-)-Gallocatechin gallate had a MIC of 50 μg/mL, pentagalloylglucose of 100 μg/mL and rhodiosin, 250 μg/mL. As the structure of pentagalloylglucose is similar to that of (-)-gallocatechin gallate, (-)-gallocatechin gallate and rhodiosin were selected for further kinetic and mechanistic studies.

Kinetic study and docking of the SaSDH-inhibitor complex

Kis were calculated from Lineweaver-Burke curves to assess the nature of the inhibitory effect of (-)-gallocatechin gallate and rhodiosin (Figure 4). Rhodiosin had characteristics of competitive inhibition with SHK giving a Ki of 73.38 μM and (-)-gallocatechin gallate of uncompetitive inhibition with a Ki of 2.47 μM.

Figure 4. Double-reciprocal plot of initial velocities versus substrate concentration in the presence of different concentrations of (-)-Gallocatechin gallate and Rhodiosin.

Molecular docking was performed for (-)-gallocatechin gallate and rhodiosin separately to the binding pocket of SaSDH using AutoDock Vina. As rhodiosin was competitive inhibitor and (-)-gallocatechin gallate was uncompetitive inhibitor, (-)-gallocatechin gallate was docked to enzyme-substrate complex, while rhodiosin was docked to free enzyme. (-)-Gallocatechin gallate formed hydrogen bonds with Pro 9, Ser 13, Ser15, Asn 58, Asn 85 and Glu 239. Rhodiosin formed hydrogen bonds with Asn 58, Ile 59, Asn 85, Asp 100 and Glu 239 (Figure 5A). ITC showed that both (-) -gallocatechin gallate and rhodiosin interacted with SaSDH with moderate affinity, giving a K_D for (-) -gallocatechin gallate of 44.65 μM and for rhodiosin of 16.45 μM (Figure 5B). We conclude that both (-)-gallocatechin gallate and rhodiosin achieved inhibition of SaSDH activity by binding to the enzyme.

Figure 5. (A) Molecular docking of (-)-Gallocatechin gallate and Rhodiosin with SaSDH. (B) Binding isotherm of SaSDH titrated with (-)-Gallocatechin gallate and Rhodiosin.

Antibacterial action of (-)-gallocatechin gallate and rhodiosin

Time-kill curves showed that (-)-gallocatechin gallate inhibited S.aureus growth at 50 μg/mL but did not kill the bacteria. Rhodiosin completely inhibited S.aureus growth at 250 μg/mL (Figure 6A). SEM observations of bacterial morphology after treatment with MIC and 1/2 MIC of (-)-gallocatechin gallate and rhodiosin for 1.5 h showed cell disintegration, cell surface fracturing and wrinkling compared with the smooth, intact, spherical appearance of water-treated cells (Figure 6B). Fluorescent imaging showed the appearance of PI-labeled bacteria, indicating cell membrane destruction, after treatment with MICs of (-)-gallocatechin gallate or rhodiosin compared with red fluorescence of intact cell membranes in control groups (Figure 6C). Both (-)-gallocatechin gallate and rhodiosin affected bacterial morphology by disrupting the integrity of the cell membrane.

Figure 6. Antibacterial action of (-)-Gallocatechin gallate and Rhodiosin.(A) Time-Kill curves for (-)-Gallocatechin gallate and Rhodiosin against S.aureus.(B) SEM of S.aureus treated with (-)-Gallocatechin gallate and Rhodiosin.(C) Calcein AM/PI staining of S.aureus treated with (-)-Gallocatechin gallate and Rhodiosin.

Discussion

Staphylococcus aureus is common and causes skin and soft tissue infection, infective endocarditis, osteomyelitis, bacteraemia and fatal pneumonia¹⁶. β-Lactam drugs are usually used for treatment but most were marketed half a century ago and drug resistance is rife. There is an urgent need to find new anti-Staphylococcus aureus targets to develop new drugs. Many scientists have already demonstrated that SDH is attractive target for new antibacterial drugs but no SaSDH inhibitors with antibacterial activity have been found so far. The current work explored the potential of SDH as a new drug target for anti-Staphylococcus aureus drugs. High-throughput virtual screening based on molecular docking and susceptibility were used to identify SaSDH inhibitors with antibacterial activity against Staphylococcus aureus.

494,000 compounds from Life Chemicals HTS Compound Collection, 208,000 from Specs HTS Compounds Library and 18,000 from MCE Bioactive Compound Library PlusA were screened and clogP, rotating Bonds, TPSA, HB Acceptor and HB Donor identified for the 200 most promising candidates (Supplement 2). Binding to SaSDH depended largely on hydrogen bonds from hydroxyl groups on the compounds, consistent with previous observations^17–19. All compounds in the Life Chemicals HTS Compound Collection and Specs HTS Compounds Library had absolute docking score values of less than 10 but 63 compounds from MCE Bioactive Compound Library PlusA had values greater than 10. Therefore, compounds from MCE Bioactive Compound Library PlusA were selected. We finally chose 12 compounds after multiple considerations, of which 6 had IC₅₀ less than 100 μM and 3 had antibacterial activities: pentagalloylglucose, (-)-gallocatechin gallate and rhodiosin.

(-)-Gallocatechin gallate is a naturally-occurring polyphenolic compound derived from green tea²⁰ for which antitumor²¹, antiviral^22,23, anti-depressive²⁴, antioxidant²⁵ and cardioprotective properties²⁶ have been reported. It has also been reported to have antibacterial activity with an unknown mechanism^27–29. Rhodiosin is also naturally-occurring and is a flavonoid derived from Rhodiola rosea³⁰ for which antitumor³¹, antiviral³², antioxidant and anti-inflammatory³³, but not antibacterial activities have been identified. (-)-Gallocatechin gallate was found to have a Ki for SaSDH activity of 2.47 μM and rhodiosin of 73.38 μM using Lineweaver-Burk double-reciprocal plots during the current work. Different modes of inhibition were found for the two compounds with rhodiosin being competitive with respect to SKH while (-)-gallocatechin gallate was uncompetitive, perhaps due to structural differences. Both (-)-gallocatechin gallate and rhodiosin were found to form hydrogen bonds with SaSDH by molecular docking and ITC giving respective KDs of 44.65 μM and 16.45 μM, indicating moderate affinity. Previously, Claudia et al¹⁴ firstly reported three inhibitors of SaSDH through virtual screening, two of which were competitive inhibitors with Ki were 9.53 μM and 19.76 μM respectively, another was mixed competitive having a Ki of 58.3 μM. Afterwards Claudia et al⁹ adopted a structure similarity search strategy based on these three inhibitors to find new SaSDH inhibitors, and successfully found four more inhibitors with Ki of 106.1 μM, 75.3 μM, 83.1 μM and 8.3 μM. Each of these four inhibitors had a very different behaviour in binding to SaSDH. All these inhibitors were synthetic compound with low molecular weight, not natural product. Almost all of these compounds seemed to have stronger enzymatic inhibitory activity than rhodiosin found in our work, but the antibacterial activities of these inhibitors previously reported were not determined. (-)-Gallocatechin gallate and rhodiosin we found offer new structures of SaSDH inhibitors reported actually.

Antibacterial activity MICs of 50 μg/mL for (-)-gallocatechin gallate and 250 μg/mL for rhodiosin were found, indicating fairly weak antibacterial actions, similar to most antibacterial natural products. (-)-Gallocatechin gallate was found to slow bacterial growth, unlike rhodiosin which showed complete growth inhibition, by time-kill curves. The chemical instability of (-)-gallocatechin gallate may be responsible for its more limited performance which limits its utility as an antibacterial agent. Both compounds were found to disrupt the bacterial cell wall but we believe they are unlikely to be single-target compounds and further study is necessary to determine the full extent of their antibacterial effects.

In summary, recombinant SaSDH protein was firstly expressed and purified, and then inhibitors were identified from three commercial databases through virtual screening. (-)-Gallocatechin gallate and rhodiosin were found to inhibit SaSDH catalytic activity. Both (-)-gallocatechin gallate and rhodiosin had antibacterial activity against S.aureus in vitro, disrupting the integrity of the bacterial cell membrane. The two novel SaSDH inhibitors with antimicrobial activity have potential for the development of drugs for the treatment of S.aureus infection.

Authors’ contributions

ZMF and DQ conceived and designed the experiments. ZMF and QJF prepared the materials and performed the experiments. ZMF, QJF and DQ analysed the computational data and results. The manuscript was organised and written by ZMF and DQ.

Acknowledgements

The authors would like to express their gratitude to EditSprings (https://www.editsprings.cn) for the expert linguistic services provided.

Disclosure statement

The authors report no conflicts of interest.

Additional information

Funding

The study was supported by Natural Science Foundation of Fujian Province [NO.2022J01672].