key: cord-0753672-u9fmtmqm
authors: Hijikata, Atsushi; Shionyu‐Mitsuyama, Clara; Nakae, Setsu; Shionyu, Masafumi; Ota, Motonori; Kanaya, Shigehiko; Hirokawa, Takatsugu; Nakajima, Shogo; Watashi, Koichi; Shirai, Tsuyoshi
title: Evaluating cepharanthine analogues as natural drugs against SARS‐CoV‐2
date: 2021-12-09
journal: FEBS Open Bio
DOI: 10.1002/2211-5463.13337
sha: 2ea9ce4583071bc4650374e0ceea486ec9ba5ec4
doc_id: 753672
cord_uid: u9fmtmqm

Cepharanthine (CEP) is a natural biscoclaurine alkaloid of plant origin and was recently demonstrated to have anti‐severe acute respiratory syndrome coronavirus 2 (anti‐SARS‐CoV‐2) activity. In this study, we evaluated whether natural analogues of CEP may act as potential anti‐coronavirus disease 2019 drugs. A total of 24 compounds resembling CEP were extracted from the KNApSAcK database, and their binding affinities to target proteins, including the spike protein and main protease of SARS‐CoV‐2, NPC1 and TPC2 in humans, were predicted via molecular docking simulations. Selected analogues were further evaluated by a cell‐based SARS‐CoV‐2 infection assay. In addition, the efficacies of CEP and its analogue tetrandrine were assessed. A comparison of the docking conformations of these compounds suggested that the diphenyl ester moiety of the molecules was a putative pharmacophore of the CEP analogues.

underline the still-ongoing requirement of effective anti-SARS-CoV-2 therapeutics.

Some of the authors recently reported that a combined treatment of nelfinavir and cepharanthine (CEP) was highly effective for COVID-19 by screening a panel of the approved drugs in a SARS-CoV-2 cell culture model [7] . Nelfinavir and CEP are the approved drugs for anti-AIDS and anti-leukopenia, respectively. CEP hinders SARS-CoV-2 entry into cells, and nelfinavir inhibited the catalytic activity of viral main protease (M-pro) to suppress viral replication. In vitro assay confirmed a synergistic effect of the combined treatment to suppress SARS-CoV-2 proliferation. Nelfinavir was predicted to shorten the period before viral clearance by~5 days by treatment at early days postinfection, and the additional treatment of CEP enhanced the efficacy significantly.

CEP is a natural product isolated from the plant Stephania cephalantha Hayata and has been approved for leukopenia, xerostomia and alopecia to date [8] . Natural products such as CEP are expected to be promising sources of anti-COVID-19 therapeutics [9] [10] [11] [12] [13] . It is well known that natural drugs exhibit antiviral activities against notable viral pathogens, including coronavirus, dengue virus, coxsackievirus, hepatitis B virus, hepatitis C virus and influenza virus [14, 15] . Therefore, so far, many studies for nature-derived drug discovery have been conducted against SARS-CoV-2 by in silico virtual screenings [16] [17] [18] , as well as in vitro assays [19, 20] . For instance, quercetin, which is a plant flavonoid and known to have antiviral properties, has been confirmed to block the proteolytic activity of SARS-CoV-2 3CLpro in vitro [19] and is currently in a phase 3 clinical trial (ClinicalTrials.gov identifier: NCT04578158). The other flavonoids, such as baicalein, herbacetin and pectolinarin, have also been discovered as potent inhibitors for the enzyme [20] .

Thus, the efficacies of natural CEP analogues against COVID-19 would be worth further investigating. In this study, an in silico docking study of natural CEP analogues and a cell-based assay of anti-SARS-CoV-2 activity for selected analogues have been attempted.

The CEP analogues were sought in the KNApSAcK database [21] . The CEP structure was exhaustively compared with the compounds in the database by using COMPLIG [22] . COMPLIG matches molecular graphs and evaluates the similarity score of two molecules, A and B, as min{M bonds in molecules A and B, and the total number of atoms and bonds  matched between molecules A and B, respectively. Both element and chirality should be identical for atoms, and bond  order should be identical for bonds to be matched, if applicable. A total of 24 compounds with more than 0.90 similarity score were extracted (Table 1) . COMPLIG was also used to superimpose the compound structures according to the graph matching.

The binding affinities of the CEP analogues and M-pro inhibitors to the receptor binding domain of spike protein (S-pro) and M-pro of SARS-CoV-2, Niemann-Pick type C intracellular cholesterol transporter 1 (NPC1) and two pore segment channel 2 (TPC2) of human were predicted via molecular docking simulations. The target atom coordinates of S-pro and M-pro were adapted from the previously reported complex models [12] based on the Protein Data Bank (PDB) [23] codes 6M0J [24] and 6LU7 [9] , respectively. The NPC1-NPC2 complex model was constructed based on PDB code 6W5V [25] . The TPC2 coordinates were based on PDB code 6NQ0 [26] . Note that the resolutions of the latter two structures were lower than the recommended threshold for a docking simulation, 2.5 A, specifically, 4.00 and 3.70

A for 6W5V and 6NQ0, respectively. The target coordinates in complex with the bestscored CEP are available from https://harrier.nagahama-ibio.ac.jp/dtx/SARS-CoV-2/.

In the previous study, knowledge-based modeling, in which the similar ligands in the known complex structure were used to predict the interactions, was applied for modeling the drug-SARS-CoV-2 protein complexes [12] . However, no complex structure binding a CEP analogue was detected among the known homologous structures. Thus, a docking simulation was used in this study. AUTO-DOCK VINA (ADV) version 1.1.2 was used to dock the model compounds to the proteins [27] . AUTODOCKTOOLS version 1.5.6 was used to prepare PDBQT files of the target proteins and compounds [28] . The concaved areas on the protein surface (pockets) were detected using fpocket2 [29] . The pockets overlapping the angiotensin-converting enzyme 2 (ACE2) interaction site of S-pro and the active site of Mpro were selected for the target site. For NPC1 and TPC2, preliminary docking simulations of CEP were executed, and two and one pockets, respectively, with the highest docking score were selected for the target sites. The centers of the docking grids (28 9 28 9 28 A) were defined as the centroid coordinates of amino acid residues composing each pocket (Fig. S2) performed with the options exhaustiveness of 1024 and seed of 1024. To verify the results, we also executed the docking simulations using AUTODOCK 4.2.6 (AD4) [30] with the same grid setting (center and size) and PDBQT files as ADV docking. To set the docking grid size to 28 9 28 9 28 A, the grid point spacing was changed from 0.375 to 0.2 A. The AD4 docking simulations were executed using the Lamarckian genetic algorithm with default parameters.

The molecular graphics were prepared by using UCSF Chimera [31] . The schematic diagrams of the interactions in the binding sites were prepared by using MOE [32] . The principal component (PC) analysis (PCA) of the docking scores was executed by using the R package [33] . The matrix of the best docking scores of the compounds for S-pro receptor binding domain (SproS), main protease pocket site (MproS), NPC1S1, NPC1S2 and TPC2S sites ( Table 1 ) was applied to the prcomp function of R with scaling for the PCA. To examine the contribution of each target site/protein to PC axes, we calculated the factor loadings (vector of correlation coefficients between the docking score axes and the PC axes). The docking scores were sign inverted so that the factor vectors pointed to the higher-affinity directions. The anti-SARS-CoV-2 activities of the compounds were assayed in the cell culture model as reported previously [7] . In brief, VeroE6/TMPRSS2 cells (VeroE6 cells overexpressing transmembrane protease, serine 2 [34] ) were inoculated with the SARS-CoV-2 Wk-521 strain at a multiplicity of infection of 0.003 in Dulbecco's modified Eagle's medium (Fujifilm Wako Pure Chemical, Osaka, Japan) supplemented with 10% fetal bovine serum (Sigma Aldrich), 10 UÁmL À1 penicillin, 10 mgÁmL À1 streptomycin and 10 mM HEPES (pH 7.4) at 37°C in 5% CO 2 for 1 h, and unbound viruses were removed by washing. The compounds were treated in nine serial dilutions (30.00, 10.00, 3.30, 1.10, 0.36 and 0.12 µM, and three technical replicates were also performed for 5.00, 1.00 and 0.20 µM) for 24 h, and the amounts of extracellular viral RNA were measured. Viral RNA was extracted with a MagMAX Viral/ Pathogen II Nucleic Acid Isolation kit (Thermo Fisher Scientific, Waltham, MA, USA) and quantified by real-time RT-PCR analysis with a one-step quantitative RT-PCR kit (THUNDERBIRD Probe One-step qRT-PCR kit; TOYOBO, Osaka, Japan) using the SARS-CoV-2-specific primers and probe [35] . The relative viral RNA levels were calculated by setting that for DMSO control treatment as 1.0 and were plotted against the concentrations of each compound. The cell culture infection assay was handled in a biosafety level 3.

The natural products highly similar to CEP were extracted from the comprehensive species-metabolite relationship database KNApSAcK (Table 1) . A total of 24 CEP analogue compounds showed more than 90% identity in atoms/bonds to CEP by the graph-matching of chemical formula. CEP consists of two coclaurine moieties connected in a cycle. Accordingly, most of the detected analogues were biscoclaurine alkaloids of plant origins. Although the detected analogues largely complied with this molecular architecture of CEP, no one perfectly shared the stereochemistry at two chiral centers and the coclaurine moieties conjugating atom positions with CEP (Fig. S1 ). It suggested that CEP is a stereochemically unique molecule among the analogues.

Several target proteins have been proposed for the CEP analogues, such as the S-pro and M-pro of coronaviruses and NPC1 and TPC2 of human. However, identities of the target proteins have not been fully confirmed in most of the cases.

CEP is thought to interfere with the ACE2-S-pro interaction, which is essential for SARS-CoV-2 entry into cells [24] , because CEP inhibits the entry phase in viral infection [7] . Also, the interaction of CEP to NPC1 and its efficacy in NPC1 inhibition have been reported [36] [37] [38] . NPC1 is the cholesterol transporter, which acts in salvaging cholesterol from the endosomal or lysosomal compartment to the cell membrane. A dysfunction of NPC1 activity results in Niemann-Pick type C disease partly because of a disruption of membrane raft formation, which also works as an obstacle to viral entry into cells [39] [40] [41] .

TET has been shown to have anti-SARS-CoV-2 activity through inhibiting TPC2 [42] . TPC2 is an endosomal cation channel that acts in trafficking low-density lipoprotein molecules and also is known to be involved in viral entry into cells through the endosome [43] .

M-pro is the most used drug target of SARS-CoV-2, and a few of the CEP analogues, including curine, are assumed to interact with this enzyme [44] . Therefore, for a comparison purpose, three natural products, which are thought to target M-pro, have also been included in this analysis, namely, baicalein [20] , hyperoside [45] and hesperidin [46] .

ACE2 interaction site of SproS [24] , active site of Mpro (MproS) [9] , two putative CEP binding sites of NPC1 (NPC1S1 and NPC1S2) [25, 47] and the putative CEP binding site of TPC2 (TPC2S) [26] were targeted in this study (Figs 1 and S2 ). Docking simulations of the CEP analogues and the M-pro inhibitors to the hypothetical target sites were performed by using ADV [27] and AD4 [30] , and the best-scored docking poses for each pair of compound and target site were compared ( Table 1 ). The score correlation between ADV and AD4 was not high, specifically, 0.48, 0.05, 0.49, 0.17 and 0.55 for SproS, NPC1S1, NPC1S2, TPC2S and MproS, respectively. It has been proposed that the docking poses were generally reliable if root-mean-square deviation (RMSD) of the ligand atoms between the ADV and AD4 were less than 2.0 Å [48] . The RMSD values tended to be lower for the suggested targets, for example, 0.9 and 1.3 Å for CEP against SproS and NPC1S1, respectively, and 2.1 Å for TET against TPC2S.

To systematically analyze the results of docking simulation, we performed a PCA over the matrix of docking scores (Fig. 2) . The cumulative proportion of variance was 86.5% to the second PC axis (59.1% and 27.4% for PC1 and PC2, respectively) for the ADV results. For the AD4 results, the cumulative proportion of variance was 71.2% to the second PC axis (48.5% and 23.5% for PC1 and PC2, respectively).

In the PC1-PC2 planes, the distributions of the compounds were different between ADV and AD4. As a consensus feature, the distributions were determined mainly by separating the viral proteins (S-pro and Mpro) and human proteins (NPC1 and TPC2). The M-pro inhibitors (hyperoside, baicalin and hesperidin) preferred M-pro, which was consistent with the hypothetical target of the compounds. CEP relatively preferred NPC1 but was placed closer to the center of the distribution. The other compounds also, as overall, did not show a bias to a specific target. A total of 10 procurable compounds were selected so that they surround CEP in the PC1-PC2 plane as indicated with colored circles in Fig. 2 .

A cell culture SARS-CoV-2 infection assay was performed for the selected compounds. VeroE6/ TMPRSS2 cells inoculated with SARS-CoV-2 for 1 h and washed out were incubated with or without the compounds for 24 h, and viral RNA in the culture supernatant was quantified (Fig. 3) ) . Berbamine, dauricine and liensinine had even weaker activity to inhibit SARS-CoV-2. Other compounds did not show a significant effect. As a positive control, remdesivir is currently an approved anti-COVID-19 drug directly targeting a SARS-CoV-2 protein with IC 50 value of 0.99 µM [49] .

The natural CEP analogues have been examined for anti-SARS-CoV-2 activity by combining in silico and in vitro analyses. The results indicated that CEP showed a potent activity to inhibit SARS-CoV-2 infection among the naturally occurring analogues, which confirmed the previous report [7] . TET was also shown to be effective but had lower activity than CEP. These two compounds are hypothesized to interact with human lysosomal membrane proteins: CEP to NPC1 and TET to TPC2. The natural ligands of these membrane proteins are cholesterol for NPC1 and phosphatidylinositol 3, 5-bisphosphate for TPC2, and consequently the molecular environments of target sites are hydrophobic [25, 42] . The CEP analogues are generally hydrophobic molecules with predicted logP values ranging from 5.8 to 6.7, and those for CEP and TET are 6.5 and 6.4, respectively. Usually, discriminating specificity or evaluating affinity for a highly hydrophobic interaction is difficult. Accordingly, CEP and TET did not localize in the PC1-PC2 plane of the docking scores (Fig. 2) , indicating no preference for a particular target was predicted. Thus, the result did not provide additional implications for the target protein of these compounds. CEP and TET are the natural products, and they have neither been designed nor tuned for a specific target protein. It would be probable that they weakly interact with multiple targets, for example, NPC1 and TPC2.

Like many of the CEP analogues, CEP and TET are the cyclic molecules with conjugated coclaurine moieties, and their chemical formulas are highly similar. However, because of the difference in configuration at two chiral centers and the positions of the atoms for coclaurine moieties conjugation, the 3D structures of CEP and TET considerably deviated. Therefore, the 3D structures of the best-dockingscored CEP and TET and other examined analogues are superimposed to extract common structural features (Fig. 4) . The overall structures were not superimposable between the analogues, including between CEP and TET. Most of the superimposable partial structures to CEP were centered by a diphenyl ester moiety for TET, berbamine, dauricine and trilobine as indicated with the ball-and-stick models in Fig. 4 . Between CEP and TET, the conformations of the diphenyl ester moiety (meshed in gray for CEP in Fig. 4) were similar, and none of the other examined CEP analogues without remarkable anti-SARS-CoV-2 activity had this CEP-TET structural motif in an identical manor. For example, berbamine and dauricine took a similar conformation at the corresponding parts but lacked the methyl group (meshed in magenta for CEP in Fig. 4) , which was shared between CEP and TET. Trilobine had this methyl moiety, but the conformation of the diphenyl ester moiety was slightly deviated from CEP and TET. The result suggested this common partial structure between CEP and TET as a putative pharmacophore of CEP analogues. For the CEP analogues to be used as anti-COVID-19 therapeutics, their mechanisms of action should be carefully investigated. It has been recently pointed out that the cationic amphiphilic compounds, including CEP, frequently cause cellular phospholipidosis, which is a lysosomal storage disorder characterized by the excessive accumulation of phospholipids. The extents of phospholipidosis were highly correlated with the antiviral efficacies of the compounds. Such compounds also included hydroxychloroquine, which was initially expected to be an anti-COVID-19 drug, but the clinical trials were, in some cases, terminated partly because of adverse effects [50] .

Considering the assumed action mechanisms of CEP and TET, i.e., inhibition of endosomal lipid/steroid trafficking proteins, the same adverse effect would be difficult to avoid. The drug-induced phospholipidosis is a frequently encountered hindrance in drug development, and recently it becomes clearer that ammonium cation groups in the molecules play a critical role in the symptom [51] . The structural motif suggested for CEP and TET does not contain, although exists close to, the ammonium cations of the molecules; thus, it might be useful in designing CEP derivatives with low adverse effect, for example, replacing ammonium cation without largely affecting the antiviral efficacies. In the docking poses, where higher similarity was observed between the ADV and AD4 results, the suggested pharmacophore was used for interactions with proteins, but the ammonium cations did not make remarkable bonds in all of the cases (Fig. S2B) . It might also contribute to further improving CEP or TET to obtain more target-specific (to reduce cross-reaction) or less specific (to be multitargeted) drugs.

The natural CEP analogues were examined for anti-SARS-CoV-2 activity, and the efficacies of CEP (IC 50 1.90 µM) and TET (IC 50 10.37 µM) were detected. The diphenyl ester moiety of the molecules was suggested for a putative pharmacophore, but no target protein preference was implied by the present analyses; therefore, it remained a possibility that they were weak multitargeted natural drugs.

Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig S1. Chemical formulas of CEP-analogues and Mpro inhibitors. The formulas are arranged to be comparable with that of CEP except for the M-pro inhibitors. Two chiral centers and connecting carbon atoms of coclaurine moieties of CEP are indicated with blue and green circles, respectively. The compound names are shown in red (CEP), blue (assayed CEP-analogues), green (M-pro inhibitors), or black (others). 

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK

Efficacy and safety of the mRNA-1273

SARS-CoV-2 Vaccine

COVID-19 vaccine breakthrough infections reported to CDC -United States

BNT162b2 vaccine breakthrough: clinical characteristics of 152 fully vaccinated hospitalized COVID-19 patients in Israel

Spike mutation D614G alters SARS-CoV-2 fitness

Will SARS-CoV-2 variants of concern affect the promise of vaccines?

Potential anti-COVID-19 agents, cepharanthine and nelfinavir, and their usage for combination treatment

Cepharanthine: an update of its mode of action, pharmacological properties and medical applications

Structure of M(pro) from SARS-CoV-2 and discovery of its inhibitors

Anti-SARS-CoV-2 activities in vitro of Shuanghuanglian preparations and bioactive ingredients

Cepharanthine: a review of the antiviral potential of a Japaneseapproved alopecia drug in COVID-19

Knowledge-based structural models of SARS-CoV-2 proteins and their complexes with potential drugs

Current status of structure-based drug repurposing against COVID-19 by targeting SARS-CoV-2 proteins

ª 2021 The Authors. FEBS Open Bio published by John Wiley & Sons Ltd on behalf of Federation of

Antiviral natural products and herbal medicines

Natural product-derived phytochemicals as potential agents against coronaviruses: a review

In silico screening of natural compounds against COVID-19 by targeting Mpro and ACE2 using molecular docking

Analysis of natural compounds against the activity of SARS-CoV-2 NSP15 protein towards an effective treatment against COVID-19: a theoretical and computational biology approach

Structure-based virtual screening of bioactive compounds from Indonesian medical plants against severe acute respiratory syndrome coronavirus

Structural stability of SARS-CoV-2 3CLpro and identification of quercetin as an inhibitor by experimental screening

Flavonoids with inhibitory activity against SARS-CoV-2 3CLpro

KNApSAcK family databases: integrated metabolite-plant species databases for multifaceted plant research

Classification of ligand molecules in PDB with fast heuristic graph match algorithm COMPLIG

Announcing the worldwide Protein Data Bank

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor

Structural basis of low-pH-dependent lysosomal cholesterol egress by NPC1 and NPC2

Structural mechanisms of phospholipid activation of the human TPC2 channel

AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading

AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility

Fpocket: an open source platform for ligand pocket detection

Automated docking of flexible ligands: applications of AutoDock

UCSF Chimera-a visualization system for exploratory research and analysis

Medicinal chemistry and the molecular operating environment (MOE): application of QSAR and molecular docking to drug discovery

R: A language and environment for statistical computing

Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR

Pharmacological blockade of cholesterol trafficking by cepharanthine in endothelial cells suppresses angiogenesis and tumor growth

Potential COVID-19 therapeutics from a rare disease: weaponizing lipid dysregulation to combat viral infectivity

The lysosome: A potential juncture between SARS-CoV-2 infectivity and Niemann-Pick disease type C, with therapeutic implications

Niemann-Pick C1 disease gene: homology to mediators of cholesterol homeostasis

Cholesterol in signal transduction

Deficiency of niemann-pick type C-1 protein impairs release of human immunodeficiency 293

ª 2021 The Authors. FEBS Open Bio published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies. virus type 1 and results in Gag accumulation in late endosomal/lysosomal compartments

Discovery of SARS-CoV-2 antiviral drugs through large-scale compound repurposing

Mining of Ebola virus entry inhibitors identifies approved drugs as two-pore channel pore blockers

Coronavirus disease 2019 drug discovery through molecular docking

Binding interaction of quercetin-3-beta-galactoside and its synthetic derivatives with SARS-CoV 3CL(pro): structure-activity relationship studies reveal salient pharmacophore features

Anti-SARS coronavirus 3C-like protease effects of Isatis indigotica root and plant-derived phenolic compounds

A structure of Niemann-Pick C1 protein reveals insights into the function of the C-terminal luminal domain in cholesterol transport

Consensus docking: improving the reliability of docking in a virtual screening context

In vitro evaluation of antiviral activity of single and combined repurposable drugs against SARS-CoV-2

Efficacy and safety of hydroxychloroquine vs placebo for pre-exposure SARS-CoV-2 prophylaxis among health care workers: a randomized clinical trial

Emerging mechanisms of druginduced phospholipidosis

The authors declare no conflict of interest.

The target coordinates in complex with the best-scored CEP are available from https://harrier.nagahama-i-bio. ac.jp/dtx/SARS-CoV-2/.