Exploring the phytochemicals of Platycodon grandiflorus for TMPRSS2 inhibition in the search for SARS-CoV-2 entry inhibitors

1 Introduction

Platycodon grandiflorus (Jacq.) A. DC. is the only single species of the Platycodon genus, (Campanulaceae), commonly known as balloon flower. China, Korea, Japan, Mongolia, and Russia are all home to this species. P. grandiflorus is quite common in China, and its cultivation sites may be found throughout the country (Zhang et al., 2015). The rhizomes of P. grandiflorus are popularly known as Jiegeng or Lingdanghua in China, Doraji in North Korea, Kikyo in Japan, and Huridunzhaga in Mongolia (China Pharmacopoeia Committee, 2005). P. grandiflorus is a perennial plant that grows to a height of 20–120 cm. The leaves are green, ovate, elliptic, or lanceolate, and measure 2–7 0.5–3.5 cm². It has a blue or purple bloom that is 1.5–4.5 cm² in size. Because of its importance as traditional medicine and food resource, P. grandiflorus roots and rhizomes are usually collected around August (Huang, 2008). P. grandiflorus thrives in sunny herbal communities in thickets and is only seldom seen in woods below 2000 m in height, owing to its high ecological flexibility (Zhang et al., 2015). The rhizomes of P. grandiflorus have been widely utilized in traditional Chinese medicine (TCM) with outstanding therapeutic benefits for the treatment of cough, excessive phlegm, and sore throat (China Pharmacopoeia Committee, 2005). Its rhizomes are traditionally used in various Japanese prescriptions to treat suppuration, chronic rhinitis, chronic tonsillitis, and other ailments (Zhang et al., 2015). The roots of P. grandiflorus were utilized to treat bronchitis, asthma, pulmonary tuberculosis (TB), diabetes, and inflammatory disorders in Korea after being grown for four years (Lee, 1973; Takagi and Lee, 1972). Furthermore, this plant may be utilized as a food source and is prepared into delectable meals in East Asian nations including China, Japan, and Korea (China Pharmacopoeia Committee, 2005).

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) depends on the host protein angiotensin-converting enzyme 2 (ACE2) receptor for entry into the host cells (Hoffmann et al., 2020). The viral entry necessitates not only binding to the ACE2 receptor but also priming of the virus's spike (S) protein by cleavage of the S proteins at the S1/S2 and S2 sites by a multi-domain type II transmembrane serine protease, transmembrane protease serine 2 (TMPRSS2) (Baughn et al., 2020). This cleavage process is necessary for the virus to fuse with the host cell membrane and enter the cell (Hoffmann et al., 2020; Matsuyama et al., 2020). TMPRSS2 and ACE2 are co-expressed in several tissues such as the lung, heart, stomach, smooth muscle, liver, kidney, neurons, and immune cells (Fathema et al., 2021; Fuentes-Prior, 2021; Gemmati et al., 2020). Besides, both are abundantly present in type II pneumocyte cells (Kaur et al., 2021). Given the critical function of TMPRSS2 in viral entry and the lack of approved treatments for addressing the current coronavirus disease 2019 (COVID-19) pandemic, drug repurposing strategies to block this protease have received much consideration. Camostat, nafamostat, and gabexate, which are clinically authorized pharmacologic treatments of pancreatitis in Japan, can also suppress the activity of TMPRSS2 (Yamaya et al., 2020). Given the importance of host TMPRSS2 in the pathogenesis of SARS-CoV-2 infections, our current research focuses on the binding potential of a few selected active compounds of Platycodon grandiflorus to human TMPRSS2 using molecular docking, as well as probing the structural changes in the target protein induced by binding of the active compounds using molecular dynamics simulations approach. The phytochemicals proposed in the study are most likely to suppress TMPRSS2 function, preventing the SARS-CoV-2 virus from infecting host cells.

2 Materials and methods

3 Results:

We chose 34 major bioactive compounds discovered in P. grandiflorus extracts for this investigation which belong to different classes such as Triterpenoid saponins (Platycodin A, Platycodin C, Platycodin D, Platycodin D2, Deapioplatycodin D and Platycodin D3), Flavonoids (Luteolin-7-O-glucoside, Apigenin-7-O-glucoside, Apigenin, (2R,3R)-taxifolin, Luteolin, Quercetin-7-O-glucoside. Quercetin-7-O-rutinoside, Platyconin), Phenolic acids (Flavoplatycoside, Caffeic acid, 3,4-dimethoxycinnamic acid, Ferulic acid, Isoferulic acid, m-coumaric acid, p-coumaric acid, p-hydroxybenzoic acid, α-resorcylic acid, 2,3-dihydroxybenzoic acid, 2-hydroxy-4-methoxybenzoic acid, Homovanillic acid, Chlorogenic acid), Polyacetylene (Lobetyolinin, Lobetyolin) and Sterols (Betulin, β-sitosterol, δ-7-stigmastenone-3, Spinasterol, α-spinasteryl-3-O-β-D-glucoside) (Table 1). These molecules were subjected to ROF filters (molecular mass <500 Da, hydrogen bond donors <5, hydrogen bond acceptors <10, and a log P octanol–water partition coefficient <5) yielding 12 orally bioactive compounds (Table 2). The chemical structures of the filtered compounds utilised for molecular docking studies are depicted in Fig. 1. A redocking experiment shows an RMSD value of 1.060 Å between the native and docked positions of the cocrystal ligand (Fig. 2). An RMSD value <2.0 Å suggests that molecular docking parameters used in the study can precisely predict the binding poses of the compounds. The molecular docking results of the filtered bioactive compounds of P. grandiflorus as well as the control (Nafamostat) against the target receptor TMPRSS2 are displayed in Table 3. Three compounds were shown to have the best docking with TMPRSS2: apigenin, luteolin, and ferulic acid. Apigenin binds to TMPRSS2 with a binding energy of −7.47 kcal/mol and inhibition constant of 3.33 µM forming four hydrogen bonds with Gly383, Gly385, Glu389 and Cys465 and hydrophobic interactions with Gly259, Trp384, Ala386, Thr387, Glu388, Asn398, Ala399, Ala400, Asp435, Cys437, Asp440 and Ala466 (Fig. 3A). Luteolin binds to TMPRSS2 with a binding energy of −6.8 kcal/mol and inhibition constant of 10.39 µM and establishes four hydrogen bonds with Lys390, Ser436 and Ser441 and hydrophobic interactions with Glu389, Asp435, Cys437, Gln438, Thr459, Ser460, Trp461, Gly464, Cys465 and Gly472 (Fig. 3B). Ferulic acid interacts with TMPRSS2 with a binding energy of −6.62 kcal/mol and inhibition constant of 13.95 µM and establishes four hydrogen bonds with Glu389, Asp435 and Lys467 as well as hydrophobic interactions with Ala386, Thr387, Glu388, Asn433, Ser436, Cys437, Asp440, Cys465 and Ala466 (Fig. 3C). The control drug, Nafamostat binds to TMPRSS2 with a binding energy of −9.19 kcal/mol and inhibition constant of 0.184 µM and establishes six hydrogen bonds with His296, Asp345 and Gly464 and hydrophobic interactions with Val280, Cys281, Cys297, Leu302, Asp435, Ser436, Cys437, Gln438, Ser441, Ser460, Trp461, Gly462 and Cys465 (Fig. 3D).

Close

Fig. 1

The drug-like bioactive compounds from P. grandiflorus used for molecular docking experiment.

Export to PPT

Close

Fig. 2

Superposition of the native conformation (yellow) of the co-crystal ligand (GBS) with its docked pose (green) revealed an RMSD of 1.060 Å.

Export to PPT

Close

Fig. 3

The binding conformation and molecular interactions of the compounds-(A) Apigenin (B) Luteolin (C) Ferulic acid and (D) Nafamostat with the target receptor TMPRSS2. On the left panel, the drug-receptor is depicted as a grey ribbon and compounds-Apigenin (spring green), Luteolin (cyan), Ferulic acid (salmon) and Nafamostat (magenta) are shown in ball and stick representations. On the right panel are displayed the molecular interactions between the compounds and the receptor with the hydrogen bonds shown in green and hydrophobic interactions shown in red semi-arcs.

Export to PPT

MD simulations of the unbound TMPRSS2 and its complexes with the top three ranking compounds-Apigenin, Luteolin, Ferulic acid, and the control (Nafamostat) were performed for 100 ns, and various geometrical properties were determined from their trajectories (Table S1). The root-mean-square deviation (RMSD) is a valuable metric for assessing the atomic position structural deviation and protein structural stability. The average RMSD of backbone atoms of TMPRSS2, TMPRSS2_Apigenin, TMPRSS2_Luteolin and TMPRSS2_Ferulic acid and TMPRSS2_Nafamostat complexes were 0.33161855 ± 0.020691134 nm, 0.282429554 ± 0.043508183 nm, 0.32489471 ± 0.045273601 nm, 0.304615439 ± 0.032035141 nm and 0.300230713 ± 0.028235882 nm respectively (Table S1). The binding of the compounds induces a reduction in the structural flexibility of the target enzyme (Fig. 4). The average RMSD values of Apigenin, Luteolin, Ferulic acid and the Nafamostat were 0.247362992 ± 0.062295857 nm, 0.576125057 ± 0.214772452 nm, 0.730537652 ± 0.21814513 nm and 0.507619867 ± 0.078623572 nm indicating that the compounds have favourable binding orientations in the target enzyme's binding pocket. To analyse the local fluctuations in the target enzyme before and after compound binding, an average of the residual fluctuations in TMPRSS2 was computed and plotted as the root-mean-square fluctuation (RMSF) (Fig. 5). The RMSF figure revealed significant residual variations in several locations in the target enzyme, with high amplitudes of fluctuations in Ser204 (0.5562 nm), Gly205 (0.5075 nm), Gln253 (0.5027 nm), Ser254 (0.4391 nm), Arg255 (0.7522 nm), Ser339 (0.4562 nm), Lys340 (0.5289 nm), Glu388 (0.6266 nm), Glu389 (0.6227 nm) and Lys390 (0.6265 nm). The radius of gyration (Rg) of unbound TMPRSS2 and TMPRSS2 docked complexes were determined to measure their structural compactness (Fig. 6). The Rg values TMPRSS2, TMPRSS2_Apigenin, TMPRSS2_Luteolin and TMPRSS2_Ferulic acid and TMPRSS2_Nafamostat complexes were 2.072465355 ± 0.014329819 nm, 2.080123706 ± 0.012810064 nm, 2.06185013 ± 0.013976506 nm, 2.067286903 ± 0.020060061 nm and 2.074381249 ± 0.01411647 nm respectively. The TMPRSS2_Luteolin and TMPRSS2_Ferulic acid complexes show slightly lower Rg values whereas TMPRSS2_Apigenin and TMPRSS2_Nafamostat show higher Rg values when compared to unbound TMPRSS2. The Rg plot analysis, in this case, reveals that TMPRSS2 protein undergoes conformational changes in response to the binding of compounds, resulting in changes in structural compactness. The solvent-accessible surface area (SASA) of a protein is the part of the protein that interacts directly with the solvent molecules around it. The SASA plot for unbound TMPRSS2 and TMPRSS2 docked complexes was generated during the 100 ns MD simulation. (Fig. 7). The unbound TMPRSS2, TMPRSS2_Apigenin, TMPRSS2_Luteolin, TMPRSS2_Ferulic acid and TMPRSS2_Nafamostat complexes show average SASA values of 150.9597053 ± 3.249067344 nm², 152.919972 ± 2.82381675 nm², 152.9353017 ± 3.806978439 nm², 153.5488102 ± 5.332797744 nm² and 153.0682637 ± 3.424944207 nm² respectively. A decrease in total SASA after interaction with the compounds was observed due to the conformational changes in the protein. Intramolecular hydrogen bonding plays a significant role in determining a protein's stability and overall structure. The stability of unbound TMPRSS2 and TMPRSS2 docked complexes was confirmed by monitoring hydrogen bonds established throughout the simulation (Fig. S1.A). The average number of intramolecular hydrogen bonds in TMPRSS2, TMPRSS2_Apigenin, TMPRSS2_Luteolin, TMPRSS2_Ferulic acid and TMPRSS2_Nafamostat complexes were 225.1298701 ± 9.051691382, 215.3916084 ± 8.497675536, 218.1968032 ± 8.978431365, 225.8011988 ± 9.14677203, 223.5194805 ± 8.42637942 respectively. The compounds- Apigenin, Luteolin, Ferulic acid and the Nafamostat showed an average number of intermolecular hydrogen bonds of 1.196803197 ± 0.937139141, 1.545454545 ± 1.018912076, 0.754245754 ± 0.907493502 and 1.632367632 ± 1.051052467 respectively which aids in the stabilization of the protein–ligand complexes (Fig. S1.B).

Close

Fig. 4

RMSD plot of unbound TMPRSS2 (black) and docked TMPRSS2 backbone atoms (TMPRSS2_Apigenin: red; TMPRSS2_Luteolin: green; TMPRSS2_Ferulic acid: blue and TMPRSS2_Nafamostat).

Export to PPT

Close

Fig. 5

RMSF plot of unbound TMPRSS2 (black) and TMPRSS2 docked complexes (TMPRSS2_Apigenin: red; TMPRSS2_Luteolin: green; TMPRSS2_Ferulic acid: blue and TMPRSS2_Nafamostat) where residues with large amplitude of fluctuations are marked and labelled.

Export to PPT

Close

Fig. 6

Rg plot of unbound TMPRSS2 and TMPRSS2 docked complexes (TMPRSS2_Apigenin: red; TMPRSS2_Luteolin: green; TMPRSS2_Ferulic acid: blue and TMPRSS2_Nafamostat).

Export to PPT

Close

Fig. 7

Total SASA plot of unbound TMPRSS2 and TMPRSS2 docked complexes (TMPRSS2_Apigenin: red; TMPRSS2_Luteolin: green; TMPRSS2_Ferulic acid: blue and TMPRSS2_Nafamostat).

Export to PPT

The binding free energies from MD simulation studies were found to be negative for Apigenin (ΔPB = -12.37 kcal/mol, ΔGB = -12.37 kcal/mol), Luteolin (ΔPB = -6.40 kcal/mol, ΔGB = -10.05 kcal/mol), Ferulic acid (ΔPB = -2.40 kcal/mol, ΔGB = -0.24 kcal/mol) and Nafamostat (ΔPB = -9.30 kcal/mol, ΔGB = -24.91 kcal/mol) (Table S2). The van der Waals energy make a significant contribution to the binding free energy except for Ferulic acid where the electrostatic energy component has predominant contributions. The key residues contributing towards the binding interaction between Apigenin and TMPRSS2 include Gly259 (−1.25 kcal/mol), Cys437 (−1.22 kcal/mol), Asn398 (−1.06 kcal/mol), Ala466 (−1.01 kcal/mol), Val434 (−0.95 kcal/mol), Ala399 (−0.82 kcal/mol), Asp435 (−0.80 kcal/mol), Thr387 (−0.74 kcal/mol), Ile381 (−0.68 kcal/mol) and Ala400 (−0.54 kcal/mol) (Fig. S2.A). In case of TMPRSS2_Luteolin complex, residues such as Cys465 (−1.69 kcal/mol), Gln438 (−1.47 kcal/mol), Cys437 (−1.04 kcal/mol), Gly462 (−0.96 kcal/mol), Gly464 (−0.96 kcal/mol), Glu389 (−0.91 kcal/mol), Trp461 (−0.78 kcal/mol), Lys390 (−0.75 kcal/mol), Ser436 (−0.63 kcal/mol), Thr459 (−0.58 kcal/mol) contribute majorly to the total binding energy (Fig. S2.B). The top ten residues contributing towards the binding interaction between Ferulic acid and TMPRSS2 include Lys401 (−1.40 kcal/mol), Leu378 (−1.26 kcal/mol), Leu403 (−0.65 kcal/mol), Phe429 (−0.47 kcal/mol), Val402 (−0.17 kcal/mol), Arg252 (−0.07 kcal/mol), Trp380 (−0.06 kcal/mol), Lys449 (−0.06 kcal/mol), Arg255 (−0.05 kcal/mol) and Arg409 (−0.05 kcal/mol) (Fig. S2.C). In case of TMPRSS2_Nafamostat complex, residues such as Gln438 (−2.00 kcal/mol), Cys437 (−1.37 kcal/mol), Asp435 (−1.33 kcal/mol), Trp461 (−1.32 kcal/mol), Pro471 (−1.15 kcal/mol), Gly472 (−0.89 kcal/mol), Gly462 (−0.85 kcal/mol), Cys465 (−0.79 kcal/mol), Leu430 (−0.72 kcal/mol) and Gly432 (−0.68 kcal/mol) have higher contribution toward the total binding energy (Fig. S2.D).

4 Discussion

The SARS-CoV-2, the causative agent of COVID-19 depends on the host multi-domain type II transmembrane serine protease, transmembrane protease serine 2 (TMPRSS2) angiotensin-converting enzyme 2 (ACE2) receptor for the cleavage process of S proteins to enter the host cells. Platycodon grandiflorus (Jacq.) A. DC. which belongs to the Campanulaceae family has immense medicinal importance. The rhizomes of P. grandiflorus have not only been widely utilized in traditional Chinese medicine (TCM) but also has been traditionally used in various Japanese prescriptions. More recently, Kim et al. (2021) found that platycodin D (PD), a triterpenoid saponin rich in P. grandiflorus efficiently prevents SARS-CoV-2 infection via lysosome and transmembrane protease serine 2 (TMPRSS2)-driven entry. Therefore, the bioactive phytochemicals of P. grandiflorus can be explored as a source of novel chemical entities to block the replication of SARS-CoV-2. Given the importance of TMPRSS2 in viral entry and the dearth of effective therapeutics for the present COVID-19 pandemic, small compounds or the Food and Drug Administration (FDA)-approved drugs can be tested for their ability to suppress the enzyme's activity. Camostat, nafamostat, and gabexate which are clinically authorized pharmacologic treatments of pancreatitis in Japan have been also used to suppress the activity of TMPRSS2 (Yamaya et al., 2020) and these known inhibitors have IC₅₀ values of 6.2 nM, 0.27 nM and 130 nM respectively (Shrimp et al., 2020). A recent study suggested that nafamostat was protective against SARS-CoV-2 pulmonary infections in two mouse models for COVID-19 (Li et al., 2021). Given the relevance of host TMPRSS2 in the pathogenesis of SARS-CoV-2 infections, our present research focuses on identifying active Platycodon grandiflorus compounds as potential TMPRSS2 inhibitors, which would then prevent SARS-CoV-2 from infecting host cells. We have used a diverse set of 34 phytochemicals of P. grandiflorus and applied drug filters to assess their oral bioavailability. The ability of the filtered drug-like compounds to bind and interact with the target receptor TMPRSS2 was investigated. The best three molecules interacting with the drug-receptor TMPRSS2 identified were Apigenin, Luteolin and Ferulic acid which bind to the active pocket with hydrogen bonds and hydrophobic interactions. 100 ns MD simulations were performed to evaluate the dynamic behaviour of the protein–ligand complexes and their stabilities were assessed in terms of RMSD, Rg, SASA and number of hydrogen bonds. Apigenin (4′,5,7-trihydroxyflavone), a flavone family member, is a dietary flavonoid that is both nontoxic and nonmutagenic and it has hydroxyl groups in both its B and C rings (Wang et al., 2020). Apigenin has antiviral properties in vitro and in vivo against a variety of viruses, including enterovirus 71 (EV71) (Lv et al., 2014; Zhang et al., 2014), hepatitis C virus (HCV) (Ahmed-Belkacem et al., 2014), Human Immunodeficiency Virus (HIV) (Kehinde et al., 2019), and adenoviruses (Kanerva et al., 2007). Apigenin inhibited HCV replication by lowering the levels of mature miR122 expression (Shibata et al., 2014) and suppressed FMDV (Foot and Mouth Disease Virus) infection at the post-entry stage by reducing internal ribosome entry site (IRES)-driven translational activity (Qian et al., 2015). Epstein-Barr virus (EBV) reactivation into the lytic cycle and virion generation by EBV-positive nasopharyngeal carcinoma (NPC) cells are both inhibited by apigenin (Wu et al., 2017). Luteolin (3,4,5,7-tetrahydroxyflavone) is a clear yellow crystal that belongs to the bioflavonoid family (Wang et al., 2020). By blocking the host proprotein convertase furin, luteolin can obstruct the later stages of the dengue virus (DENV) viral life cycle in infected cells (Peng et al., 2017), Epstein-Barr virus (Wu et al., 2016), Japanese encephalitis virus (Fan et al., 2016), HIV-1 (Mehla et al., 2011), Hepatitis B virus (Bai et al., 2016) and influenza A virus (Yan et al., 2019) are all inhibited by luteolin. The previous study suggests that the flavonoids-Apigenin and Luteolin inhibit SARS-CoV 3CL^pro activity as well with IC₅₀ values of 280.8 µM and 20.2 µM respectively (Ryu et al., 2010). Besides, luteolin has been found to block the entry of SARS-CoV (EC₅₀ = 10.6 µM) into host cells by binding to the surface spike protein (Yi et al., 2004). While our findings are encouraging, additional in vitro or in vivo studies are needed to determine the anti-SARS-CoV-2 efficacy of the compounds. Further, in TMPRSS2¯cells, severe acute respiratory syndrome coronavirus 2 (SARS-CoV) S can use the endosomal cysteine proteases cathepsin B and L (CatB/L) for S protein priming (Simmons et al., 2005) and therefore it would be fascinating to see if these compounds have similar inhibitory activity against them. Because TMPRSS2 functions as a viral protein processing protease in the pathogenesis of other coronaviruses-SARS-CoV, Middle East respiratory syndrome coronavirus (MERS-CoV), as well as influenza viruses (Mahoney et al., 2021), these lead compounds may be quite promising to be developed as broad-spectrum antivirals. The compounds could be subjected to structure–activity studies in order to enhance their potency and specificity.

5 Conclusion

The paucity of effective therapeutic medications for SARS-CoV-2 infections, as well as the rising mortality rate, necessitate the development of innovative drug candidate molecules with few adverse effects. Here, we explored the bioactive phytochemicals from a traditionally important medicinal plant Platycodon grandiflorus as a source of novel molecules for the inhibition of cellular TMPRSS2 activity, a key pharmacological target for preventing the SARS-CoV-2 entry into the host cells. Our study concludes that Apigenin, Luteolin and Ferulic acid are the most effective compounds for inhibition of the TMPRSS2 activity demonstrating that these agents may be used in COVID-19 treatment. These compounds should be subjected to further experimental studies to confirm their efficacies.