In silico study of potential anti-SARS cell entry phytoligands from Phlomis aurea: a promising avenue for prophylaxis

Aim: The severity of COVID-19 has raised a great public health concern evoking an urgency for developing multitargeted therapeutics. Phlomis species was ethno-pharmacologically practiced for respiratory ailments. Materials & methods: An array of 15 phytoligands previously isolated from Phlomis aurea were subjected to molecular docking to explore their potential SARS-CoV-Spike-angiotensin-converting enzyme 2 complex inhibition, that is essential for virus entry to host cell. Results: Acteoside (11) showed the most potent in silico inhibition with an additional merit, over hesperidin (16), of not binding to angiotensin-converting enzyme 2 with well proven in vivo pulmonary protective role in acute lung injury, followed by chrysoeriol-7-O-β-glucopyranoside (12) and luteolin-7-O-β-glucopyranoside (14). Conclusion: Phytoligands (11, 12 and 14) were posed as promising candidates with potential prophylactic action against COVID-19. These phytoligands were prioritized for further biological experimentation because of their acceptable predicted ADME and drug-likeness parameters. Moreover, they could aid in developing multitargeted strategy for better management of COVID-19 using phytomedicines.


Results
Molecular docking analysis of Phlomis aurea phytoligands Viral infections are generally initiated with the binding of viral particles to specific host-cell surface receptors. Hence, receptor recognition is considered as a critical determinant of the viral entry and cell invasion. Targeting such process is; therefore, considered an inspiring prophylactic approach against viral infection. In SARS-CoV-2, the viral entry is mediated by an envelope-embedded surface-located spike (S) glycoprotein [20] which utilizes hACE2 for cell entry [21]. The S glycoprotein is cleaved, in most cases, by host cell proteases into S1 and S2 subunits for receptor recognition and cell membrane fusion, respectively. S1 is further subdivided into N-terminal domain and C-terminal domain (CTD), both can function as a receptor-binding entity. Most recently, a pioneer study utilized immunostaining and flow cytometry assays to first identify the S1 CTD of SARS-CoV-2 as the key region that interacts with the hACE2 receptor. A 2.5-A • crystal structure of SARS-CoV-2-CTD in complex with a single hACE2 molecule in asymmetric unit (PDB ID: 6LZG) [16] was solved revealing a clear receptor-binding mode. Further analysis of the virus-receptor interaction on structural basis was performed to identify the key aminoacids involved in complex formation. Accordingly, a series of hydrophilic residues forming an H-bonds network and salt bridge interactions were located along the binding interface [16].
In light of the mentioned information we utilized docking simulation to probe the ability of the studied phytoligands to destabilize the virus-receptor complex or even prevent its formation, where possible accommodation/fitting of such phytochemicals at the SARS-CoV-2-CTD-h2ACE interface and their interaction with the complex key amino acids may provide preliminary promising insights for further biological investigation. Several in silico studies displayed the possible inhibitory mechanisms of various phytochemicals on the ACE2-Spike complex of SARS-CoV-2, and highlighted the structural determinants of important interactions [22][23][24][25][26]. In absence of a cocrystalized inhibitor at the interface of the studied complex, we employed the 'site finder' feature of MOE 2015.10 to locate the most suitable site for docking the studied phytochemicals into the SARS-CoV-2-CTD-2hACE binding interface taking in consideration the key residues involved in the complex formation. Flexible docking of the 15 phytochemicals under investigation was performed several times in the defined site at the binding interface with hesperidin as a reference viral entry inhibitor [27].
Docking simulations results (Table 1) showed that most of the studied phytoligands displayed good binding affinities compared with hesperidin (16). Acteoside (11) came at the top of the list recording the best binding affinity (-7.75 kcal/mol) among the studied phytoligands, even better than hesperidin (-7.10 kcal/mol).
Interaction of Phlomis phytoligands with human SARS-CoV-Spike-ACE2 complex Some of the studied phytoligands showed promising in silico results upon inspecting their binding modes with human SARS-CoV-Spike-ACE2 complex. They were able to efficiently accommodate into the interface and interact with the key aminoacids mostly via hydrogen bonding (Figures 3 & 4). Hence, they can destabilize the virus-receptor complexation which is dominated by polar bonding interactions with key hydrophilic amino acid residues [16]. Acteoside (11), a phenolic glycoside recording the best binding affinity in the current study, exhibited hydrogen bonding interactions with Gly496 of the SARS-CoV-2-CTD engaged in complex formation ( Figure 3K & L). Additionally, it displayed hydrogen bonding with the nearby amino acids, in other words, Glu406, Arg408, Gln493 and Ser494 on the SARS-CoV-2-CTD side of the interface. These interactions were posed by the sugar hydroxyl groups for Glu406 and Arg408; whereas, Gln493 and Ser494 interacted with the phenolic hydroxyl groups. This highlights that sugar hydroxyl groups and the phenolic hydroxyl groups are the important structural features of the phytoligand. Obviously, acteoside (11) did not show any interactions with hACE2 residues of SARS-CoV-Spike-ACE2 complex as compared with the reference hesperidin (16) which displayed π-π and hydrogen bonding interactions with hACE2 His34 and Ala387, respectively, in addition to its binding to SARS-CoV-2-CTD Gln409, Lys417, Ser494 ( Figure 3U & V).

In silico prediction of physicochemical properties, ADMET & drug-likeness parameters of the most promising phytoligands
Recently, in silico prediction studies of physicochemical properties, ADMET and drug-likeness parameters are utilized for the identification of the most promising leads. Herein, SwissADME software was employed to compute the physicochemical properties formulating drug-likeness parameters of the hit phytochemicals (Table 2) [18,28].

Discussion
As revealed form our results, acteoside (11) did not inhibit ACE2, which is an added value to its promising anti-SARS-CoV2 potential as the ACE2 inhibition is unfavorable to COVID-19 patients with already developed His A34 Tyr B453 Arg B403 Asp B405 Tyr B495 Gln B409 Arg B408 Asp A30 Pro A389 Arg A393 Glu A37

Lys B417
Arg B403 His A34 Glu A35 Gly B496 Tyr B453 Tyr B495 Thr B415 Tyr B505 Glu A37 Lys A353 Asp A38  Table 2. In silico predicted physicochemical properties, ADMET and drug-likeness parameters of the promising phytoligands (11, 12 & 14).  symptoms. These symptoms develop as a consequence to the decreased production of angiotensin 1-7, which exhibits antifibrotic, anti-inflammatory, vasodilatory actions via Mas receptor [38]. Furthermore, there are preliminary data proving that patients taking angiotensin-II inhibitors (ACE-I) exhibit severe symptoms with a higher mortality rate as compared with their counterparts not taking these medications [39]. Accordingly, ACE2 plays a protective role in the animal models of acute respiratory distress syndrome and acute lung injury [40]. Acteoside (11) occurs frequently in several botanical families viz. Scrophulariacea, Lamiacea and Verbenacea [41]. It was reported to exhibit a strong in vivo antiviral activity against influenza and antirespiratory syncytial virus which causes the lower respiratory tract infection in infants and young children in addition to other biological actions including antihepatotoxic, anti-inflammatory, antinociceptive and antioxidant effects [42].
It is worth noting that acacetin-7-O-β-glucopyranoside (13) has been reported to exhibit anti-inflammatory action [43,44] with well-proven efficacy against chronic obstructive pulmonary disease via inhibiting neutrophilic lung inflammation in a murine model of chronic obstructive pulmonary disease [45].
The phytoligands 'chrysoeriol-7-O-β-glucopyranoside (12) and luteolin-7-O-β-glucopyranoside (14)' showed a slight deviation from the ideal Lipinski's [47] and Veber's [37] drug-like bioavailability parameters. According to Lipinski's rule of five, phytoligands (12 and 14) violated the number of hydrogen bond acceptors and donors. Acteoside (11) recorded an extra Lipinski violation due to its high molecular weight. On the other hand, the studied phytochemicals violated the ideal total polar surface area of the drug-like molecule according to Veber's parameters (number of rotatable bonds ≤10 and total polar surface area ≤140). Again, 11 showed one more violation regarding the number of rotor bonds. Pre-ADMET software [19] was employed for prediction of absorption, distribution, metabolism, excretion and toxicity (ADMET) properties of the studied natural products. 12 and 14 were predicted to display moderate human intestinal absorption. 12 recorded higher predicted absorption percentage (42%) than 14 (25%); whereas, 11 was predicted to be poorly absorbed recording only 7.6% intestinal absorption. The three phytochemicals recorded good aqueous solubility. 11 was predicted to be the most readily soluble compound among the group (S = 290.5 mg/l), followed by 14 (S = 85.91 mg/l) and 12 (S = 40.22 mg/l), respectively. This refers to the expected feasibility of the studied phytochemicals to be formulated in various pharmaceutical dosage forms. The three phytochemicals displayed comparable low CNS absorption as detected by their predicted blood-brain barrier penetration values (around 0.03). Thus, the compounds are expected to display limited possible CNS side effects. The compounds displayed medium Caco-2 model and low MDCK model permeabilities. 11 recorded the highest Caco-2 model permeability value (almost two-fold) compared with 12 and 14. On the other hand, 14 was predicted to be two-fold more absorbable by MDCK than 12. 11 displayed the least MDCK permeability value among the group. Interestingly, all compounds were considered poorly bound to plasma proteins (PPB ranges from 62 to 73%) indicating that much of the unbound compound will be available for transport across various membranes to display its pharmacological activities. They were predicted to be devoid of cytochromes P450 2D6 (CYP2D6) inhibition activities but not CYP3A4, addressing the possibility of limited predicted drug interactions. Finally, the median lethal dose (LD 50 ; mg/kg) of the studied compounds in rodents was predicted employing ProTox [35], the toxicity predictor program, to be 5000 mg/kg; thus, classified according to the Globally Harmonized System of Classification and Labeling of Chemicals (GHS) as class V concerning acute oral toxicity.

Conclusion
In the current study, we conducted docking simulation to predict the in silico inhibitory potential of a set of 15 phytoligands previously isolated Phlomis aurea, a wild Sinai peninsula plant, against the functional activity of SARS-CoV-Spike-ACE2 complex. Among the studied phytoligands, the phenolic glycoside 'acteoside (11)' showed the most potent in silico inhibitory action with an additional merit, over the reference hesperidin (16), of not binding to ACE2 which was reported recently to possess a pulmonary protective role in acute lung injury and acute respiratory distress syndrome in vivo. The second most active phytoligands were flavonoids viz. acacetin-7-O-βglucopyranoside (13), chrysoeriol-7-O-β-glucopyranoside (12), followed by the lignan 'liriodendrin (15)' and the iridoid 'lamiide (6)'.
Our results provide promising leads from Phlomis aurea plant for designing and developing drug candidates with phytoprophylactic potential against COVID-19. Besides, pulmonary inflammation and fibrosis are recognized now as the first death causes of COVID-19 patients [48].
The role of anti-inflammatory agents in the effective management of symptoms during COVID-19 has been suggested by clinical practitioners [2]. Interestingly, three of the most active phytoligands in our study have well reported anti-inflammatory actions such as acteoside (11), acacetin-7-O-β-glucopyranoside (13), liriodendrin (15), with the former compound possessed antifibrotic action additionally. Accordingly, these phytoligands could aid in developing multitargeted strategy for better management and reducing the likelihood of COVID-19 using phytomedicines. Moreover, preparation of quality controlled P. aurea leaf extract aiming to possess highest levels of these phytoligands is warranted. However, in vitro and in vivo experimentation now follows in order to validate the predicted antiviral potential of the most promising phytoligands in P. aurea and/or its leaf extract standardized to the therapeutic levels of these phytoligands. The predicted ADME and drug-likeness parameters were computed for the three most promising phytochemicals (11, 12 & 14). Results showed that chrysoeriol-7-O-β-glucopyranoside (12) and luteolin-7-O-β-glucopyranoside (14) recorded relatively better predicted drug-like criteria compared with acteoside (11). Thus, these phytochemicals deserve to be further subjected to in vivo pharmacokinetic studies on experimental animals.
• Acteoside possessed the most potent in silico inhibition without binding to ACE2.
• The study posed some phytoligands with prophylactic potential against COVID-19.