The angiotensin-converting enzyme II (ACE2) is a multifunctional protein in both health and disease conditions, which serves as a counterregulatory component of RAS function in a cardioprotective role. ACE2 modulation may also have relevance to ovarian cancer, diabetes, acute lung injury, fibrotic diseases, etc. Furthermore, since the outbreak of the coronavirus disease in 2019 (COVID-19), ACE2 has been recognized as the host receptor of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The receptor binding domain of the SARS-CoV-2 S-protein has a strong interaction with ACE2, so ACE2 may be a potent drug target to prevent the virus from invading host cells for anti-COVID-19 drug discovery. In this study, structure- and property-based virtual screening methods were combined to filter natural product databases from ChemDiv, TargetMol, and InterBioScreen to find potential ACE2 inhibitors. The binding affinity between protein and ligands was predicted using both Glide SP and XP scoring functions and the MM-GBSA method. ADME properties were also calculated to evaluate chemical drug-likeness. Then, molecular dynamics (MD) simulations were performed to further explore the binding modes between the highest-potential compounds and ACE2. Results showed that the compounds 154-23-4 and STOCK1N-07141 possess potential ACE2 inhibition activities and deserve further study.
Angiotensin-converting enzyme (ACE) is a highly glycosylated transmembrane protein existing in two differentially spliced forms: the two-domain somatic ACE (ACE1, N- and C-domains) with similar, though not identical, substrate specificities and the single-domain testicular form (ACE2). Donoghue et al. identified the ACE2 gene as one that was upregulated in a human heart failure cDNA library [1]. ACE2 actually serves a multiplicity of functions and plays vital roles in different diseases, such as ovarian cancer, diabetes, acute lung injury, fibrotic diseases, etc. [2]. ACE2 is very common in ovarian cancer with amplification mutations. High expression of ACE2 promotes the prognosis of patients with ovarian cancer [3]. Recently, it has been proven that human ACE2 is a host receptor of severe acute respiratory syndrome coronavirus (SARS-CoV), which could specifically bind to SARS-CoV spike protein with high affinity [4,5]. A newly published paper reported that the novel coronavirus could enter ACE2-expressing cells, but not the cells that did not express ACE2, so ACE2 is also the host receptor for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [6].
SARS-CoV-2 is a well-known novel coronavirus that causes an acute infectious pneumonia disease, coronavirus disease 2019 (COVID-19) [7,8]. Symptoms of the infection include respiratory symptoms, fever, cough, shortness of breath, and breathing difficulties. In more severe cases, infections can cause pneumonia, severe acute respiratory syndrome, kidney failure, and even death [9,10,11]. It has been reported that bats might be the original host of this virus [12,13,14] and some animals sold at seafood markets may be the intermediate hosts for novel coronavirus [15]. The World Health Organization (WHO) announced that the novel coronavirus epidemic was listed as a Public Health Emergency of International Concern (PHEIC) on 30 January 2020. According to the COVID-19 Dashboard by the Center for Systems Science and Engineering at Johns Hopkins University, up to 1 October 2021, COVID-19 had caused more than 233 million confirmed cases and 4.7 million deaths across 192 countries and areas.
The common structure of coronavirus consists of spike (S), envelope (E), membrane (M), and nucleoprotein (N) [16,17,18,19]. The genus of coronavirus has been divided into four genera: α-coronavirus, β-coronavirus, γ-coronavirus, and δ-coronavirus [13,20]. As with SARS-CoV, the SARS-CoV-2 sequence also belongs to β-coronavirus [21]. The sequence similarities reach 76.04%, 73.33%, and 50.00% for the whole protein, receptor-binding domain (RBD), and receptor-binding motif (RBM) between the SARS-CoV-2 spike and the SARS-CoV spike (isolated from human), respectively [22]. Receptor recognition is the first essential step in the viral infection of host cells, and spike protein has been reported to mediate the entry of the virus into host cells [23,24,25]. Paxlovid was authorized by the FDA as the first oral antiviral method for the treatment of COVID-19 on 22 December 2021. Paxlovid is a mixture of nirmatrelvir and ritonavir, which can be used to treat mild to moderate COVID-19 adult and child patients (12 years and older and at least 40 kg) [26]. Therefore, it is extremely urgent to discover new drugs to inhibit novel coronavirus as soon as possible.
The search for an ACE2 inhibitor or activator could promote drug discovery for several diseases. Specifically, the inhibition of ACE2 may prevent the invasion of SARS-CoV-2. In this study, we carried out a combined virtual screening protocol to search for potential ACE2 inhibitors. Molecular docking was used to screen three natural compound product databases (ChemDiv, TargetMol, and InterBioScreen) targeting ACE2. The molecular mechanics-generalized Born surface (MM-GBSA) was also calculated to evaluate the binding of chemicals to ACE2. The ADME properties were used to measure chemical drug-likeness. Clustering analysis based solely on the structural information was performed to aid the selection of potential ACE2 inhibitors with various skeletons. Molecular dynamics simulation was used to investigate the binding mode of the inhibitors with ACE2.
The structures of the human angiotensin-converting enzyme ACE2 (PDB ID: 1R42) and the ligand binding sites are shown in Figure 1. A total of 70,902 natural compounds were combined as a ligand database, including 398 compounds from pure natural products of ChemDiv, 2131 compounds from pure natural products of TargetMol, and 68,373 compounds from the InterBioScreen natural subset. These chemicals were neatened, minimized, and prepared for docking screening to predict their binding affinities and molecular recognition using Glide SP (standard precision), with an output 42,614 molecules. These compounds had SP Glide scores ranging from −0.023 to −7.626 kcal/mol. Higher negative docking score values indicate higher affinity between the receptor and the ligands [27].
Glide XP (extra precision) is a docking method superior to SP docking [28]. A total of 10,451 out of 42,614 molecules (about the top-ranked 25%) were subjected to Glide XP calculation, and 9678 molecules were successfully docked to the receptor with Glide scores ranging from 2.951 to −7.997.
Then, the 500 top-ranked complexes of the XP docking score were chosen to calculate the MM-GBSA values (ΔG), which were used to assess the binding abilities of the receptor and ligands [29]. The obtained MM-GBSA ΔG values ranged from 0.839 to −60.737 kcal/mol.
ADME properties were calculated for the 500 top-ranked ligands after MM-GBSA treatment. The criteria contained in the rule of five were: (I) molecular mass less than 500 dalton, (II) partition coefficient (QPlogPo/w) not greater than 5, (III) hydrogen bond donors less than 5, (IV) hydrogen bond acceptors less than 10, (V) PSA less than 140 Å 2, and (VI) percent human oral absorption more than 25. These rules were selected to evaluate the drug-likeness of compounds or to determine if a compound had pharmacological or biological potency [30]. A total of 298 ligands remained after deleting molecules that did not meet these standards.
Cluster analysis was executed based on the molecular structural information. The remaining 298 molecules were clustered into 12 different categories, and the compound number contained in each category is shown in Table 1. The chemical with the best docking score in every category was chosen for further analysis.
The selected representative structures of the 12 ligands with their corresponding ADME properties are shown in Table 2. The interactions between the receptor and top 12 ligands are shown in Figure 2. Their docking results and binding residues are listed in Table 3 (The figures shown in this table were depicted using Maestro 10.1 (Schrödinger Inc., LLC, New York, NY, USA)). The structures of the 12 ligands are mostly phenol, ketone, and amine compounds.
Four of the twelve ligands were found to have clear sources and medicinal effects as known drug ingredients; these were compounds 154-23-4, 132-98-9, STOCK1N−53429, and STOCK1N-07141. With the increase in confirmed COVID-19 cases and deaths, it is a good strategy to find novel effects from known drugs or ingredients to prevent the invasion of SARS-CoV-2 because drug repositioning possesses several advantages when considering time, research cost, and safety [31]. Therefore, compounds 154-23-4, 132-98-9, STOCK1N−53429, and STOCK1N-07141 were selected for further analysis, and their interactions and distance are shown in Figure 3 (figures were depicted using PyMOL Molecular Graphics System Version 2.5.2 (Schrödinger, Inc., LLC)).
The first natural product was 154-23-4 (catechin), a polyphenolic compound found in the bark and twigs of plants [32], which possesses antioxidant, anti-inflammatory, antibacterial, antifungal, antiviral, and anticancer properties [33,34,35,36]. Catechin also possesses virus inhibition activity, and the EC50 for the influenza A (H1N1) virus is 18.4 μg/mL [36]. The Glide docking results showed strong interaction between catechin and ACE2. The three OH from chromane in 154-23-4 formed an OH-O hydrogen bonding interaction with the backbone carbonyl atom from Ala387 with a distance of 1.9 Å. The five OH from chromane formed an OH-O interaction with the backbone carbonyl atom from Ala386 with a distance of 1.9 Å. The three OH from phenyl in 154-23-4 formed an OH-N interaction with the amide -NH from Asn33 with a distance of 2.2 Å and formed an OH-O interaction with the oxygen from carboxyl in Asp30 with a distance of 1.8 Å. The four OH from phenyl formed an OH-O interaction with the oxygen from carboxyl in Asp30 with a distance of 1.7 Å. All these results indicated a strong interaction between the ligand and protein with a Glide score of −5.418 and a value of −37.592 kcal/mol.
The second natural product was 132-98-9 (Penicillin V Potassium), which is useful for the treatment of bacterial infections [37]. The docking results indicated that Penicillin V Potassium could bind to ACE2 very well. The oxygen from carboxyl in 132-98-9 formed an O-HN interaction with the -NH from Lys26 with a distance of 1.8 Å. The carbonyl from carboxyl formed an O-HN interaction with the -NH from Asn90 with a distance 1.6 Å. The carbonyl from β-lactam in 132-98-9 formed an O-HN interaction with the -NH from Gln96 with a distance 2.0 Å. The oxygen from phenoxy in 132-98-9 formed an O-HN interaction with the -NH from Asn33 with a distance of 2.3 Å. The Glide score and ΔG value were −4.335 and −18.899 kcal/mol, respectively.
The third natural compound was STOCK1N−53429 (quinic acid), a widely presented natural product found in plants [38] Quinic acid has antioxidants, increases urinary excretion, and enhances DNA repair and immunity properties [39,40,41,42]. Glide docking results showed good affinity between quinic acid and ACE2 with four hydrogen bonding interactions and a salt-bridge interaction. The three OH from cyclohexane in STOCK1N−53429 formed an OH-O interaction with amide carbonyl from Gln76 with a distance of 1.8 Å. The OH from cyclohexane formed an O-HN interaction with the -NH from Gln76 with a distance of 2.1 Å, as well as an OH-O interaction with oxygen from carboxyl in Glu35 with a distance of 1.8Å. Carbonyl in STOCK1N−53429 formed an O-HN interaction with the -NH from Lys31 with a distance of 1.8 Å. Furthermore, there was a salt-bridge interaction between STOCK1N−53429 and Lys 31 with a distance of 4.4 Å. The Glide score and ΔG value were −5.923 and −19.312 kcal/mol, respectively.
The fourth natural compound was STOCK1N-07141 (arbutin), which was first discovered in the leaves of the bearberry plant and is widely found in animals, plants, and microbes [43,44]. Arbutin has been shown to have antioxidant, anti-inflammatory, and antibacterial properties [45,46,47]. Hydroxymethyl in STOCK1N-07141 formed an OH-O interaction with the oxygen from carboxyl in Asp30 with a distance of 2.4 Å and formed an O-HN interaction with the -NH from Asn33 with a distance of 1.9 Å. The two OH in STOCK1N-07141 formed an O-HN interaction with the -NH from Arg393 with a distance of 2.0 Å and formed an OH-O interaction with the oxygen from carboxyl in Glu37 with a distance of 1.7 Å. The Glide score was −4.898, and the ΔG value was −27.518 kcal/mol.
These four chemicals have been found to possess medicinal effects, including enhanced immunity, anti-inflammatory, and antibacterial properties, etc. In addition, chemical 154-23-4 has antiviral effect
