Inhibition of Cysteine Proteases by 6,6′-Dihydroxythiobinupharidine (DTBN) from Nuphar lutea

The specificity of inhibition by 6,6′-dihydroxythiobinupharidine (DTBN) on cysteine proteases was demonstrated in this work. There were differences in the extent of inhibition, reflecting active site structural-steric and biochemical differences. Cathepsin S (IC50 = 3.2 μM) was most sensitive to inhibition by DTBN compared to Cathepsin B, L and papain (IC50 = 1359.4, 13.2 and 70.4 μM respectively). DTBN is inactive for the inhibition of Mpro of SARS-CoV-2. Docking simulations suggested a mechanism of interaction that was further supported by the biochemical results. In the docking results, it was shown that the cysteine sulphur of Cathepsin S, L and B was in close proximity to the DTBN thiaspirane ring, potentially forming the necessary conditions for a nucleophilic attack to form a disulfide bond. Covalent docking and molecular dynamic simulations were performed to validate disulfide bond formation and to determine the stability of Cathepsins-DTBN complexes, respectively. The lack of reactivity of DTBN against SARS-CoV-2 Mpro was attributed to a mismatch of the binding conformation of DTBN to the catalytic binding site of Mpro. Thus, gradations in reactivity among the tested Cathepsins may be conducive for a mechanism-based search for derivatives of nupharidine against COVID-19. This could be an alternative strategy to the large-scale screening of electrophilic inhibitors.

In view of the current worldwide COVID-19 pandemic, it was challenging to test the effect of DTBN on the cysteine protease, M pro , of the new coronavirus, SARS-CoV-2 [20]. It is a large enveloped, positive sense, single-stranded RNA betacoronavirus. The viral RNA encodes two open reading frames that, through ribosome frame-shifting, generate two polyproteins [21]. These polyproteins produce most of the proteins of the replicase-transcriptase complex [22]. The polyproteins are processed by two viral cysteine proteases: a Papain-like protease (PLpro), which cleaves three sites, releasing non-structural proteins nsp 1-3, and a 3CL protease, also referred to as the main protease (M pro ), (EC 3.4.22.69), that cleaves at 11 sites to release non-structural proteins (nsp [4][5][6][7][8][9][10][11][12][13][14][15][16], which play a key role in the infectivity of the virus. We tested the DTBN against this protease. To do so, we used the COVID Moonshot initiative (https://postera.ai/covid and https: //www.nature.com/articles/s41557-020-0496-2 [23]), which assayed the inhibitory effect of the DTBN against SARS-CoV-2 M pro [23]. Our strategy yielded a more comprehensive view of the mode of action and provided evidence for a plausible pleiotropic effect exhibited by DTBN. Insight into the structure-function relationship that leads to the high inhibitory selectivity among cysteine proteases could form a foundation for the development of new therapeutics against various diseases. We trust that DTBN is an important addition to the plethora of small-molecule inhibitors of cysteine proteases.

Computaitonal Analyses
Nupharidines are a class of terpene alkaloid known to exhibit antimetastatic activity [2], high cytotoxic activity against B16F10, U937 and HT 1080 cell lines [24], and inhibit human type II topoisomerase [25] and B16 melanoma cells [24]. The pleotropic effect of nupharidines was ascribed to electrophilic sulphur of thiaspirane, which targets nucleophilic sulphur at the active site of proteases. Since cathepsins belong to a family of proteases, we anticipated that 6,6 -dihydroxythiobinupharidine (DTBN) may exhibit an inhibitory action via the hypothesis proposed above. To better understand the mode of action of DTBN and the role active site cysteine plays in Cathepsin affinity, a series of molecular docking simulations and molecular dynamics simulations were initially performed to predict the binding mode of DTBN when docked to three types of Cathepsin protein (B, L and S) and SARS-CoV-2 M pro . This was further validated by in vitro study.

Molecular Docking Analysis
DTBN was docked to the active site of Cathepsins B, L, S and SARS-CoV-2 M pro . Binding energy metrics revealed that DTBN possess low binding affinity against Cathepsin B, L and SARS-CoV-2 M pro , whereas it has a high binding affinity towards Cathepsin S.
An analysis of molecular docking of DTBN with Cathepsin B, which shows the lowest binding energy (−42.57 Kcal/mol) of the Cathepsins, revealed that the 20 amino acids loop, which is missing in both Cathepsin S and L, is in close proximity to the binding site gorge ( Figure 2). The loop creates a wall-like structure, blocking one side of the binding site and reducing the binding site volume. This leads to a steric clash between amino acid residue, the His110 present in the concluding loop and the furan ring of DTBN (Figure 3a), which might contribute to its lower binding energy.  However, at the active site of Cathepsin B, DTBN was stabilized by a hydrogen bonding interaction between -O of the furan ring in DTBN with -NH of His111 at the protein backbone (Table 1). In addition, the DTBN was stabilized by two other hydrogen bonds (i) between the hydroxy group of DTBN and -C(O) of Thr120 at protein backbone, (ii) between the hydroxy group of DTBN and side chain carboxylic -OH of Glu122 (Table 1). This gain in interactions may contribute to the final docking score (Figure 3). In comparison, the Cathepsin B inhibitor at the same active site is stabilized by three hydrogen bond interactions, ((i) between the side chain carboxylic -OH of Asp69 and carboxylic hydrogen of the inhibitor ((O)  Similarly, DTBN at the active site of Cathepsin L is stabilized by two hydrogen bond interactions (i) between hydroxy oxygen of DTBN and -NH of Gln21 at the protein back bone, and (ii) between hydroxy hydrogen of DTBN and -C(O) of Gln21 at the protein back bone. In addition, the DTBN was stabilized by the π . . . π stacking interaction between the furan ring of DTBN and indole ring of Trp189 ( Figure 4). However, DTBN experiences a steric clash with amino acid residues at the active site. A steric clash is observed between the furan ring in DTBN and amino acid residue, Leu144 (Figure 4a), which might contribute to its low binding energy. In comparison, the Cathepsin L inhibitor at the same active site is stabilized by a hydrogen bond interaction (Asp162C(O) . . . HN of the inhibitor (1.72 Å)) and by three aromatic C-H . . . O hydrogen bond interactions. These stabilizing interactions likely contributed to its high binding energy and reactivity. An analysis of the docking results of DTBN with Cathepsin S revealed that the ligand, DTBN, is stabilized by one hydrogen bond interaction at the active site ( Figure 5). A hydrogen bond is observed between the hydroxy group of DTBN and -C(O) of Val162 at the protein backbone. Unlike Cathepsins B and L, in Cathepsin S there is no steric clash observed between protein and ligand, probably because the absence of a destabilizing interaction resulted in the observed high binding energy (Table 1). In comparison, the Cathepsin S inhibitor at the active site is stabilized by a hydrogen bond interaction between the side chain amide -NH of Asn161 and cyano group of the inhibitor ((O)C-NH 2 . . . NC, (2.02 Å)). The molecular docking studies shown above predicted that the selectivity and specificity of the active site amino acids surrounding the cysteine would affect the binding site gorge size and shape. The active site is sandwiched between two β-barrel-shape polypeptides in the cysteine proteases, and the shape of the pace between them is important when fitting the inhibitor to the site, and for its effective inhibition. A profound difference is seen between Cathepsin B and the other tested cathepsins, where a loop of 20 amino acids is located in close proximity to the binding site, which may hinder the structural fitting of DTBN to the active site.
The main Protease (M pro ) of SARS-CoV-2 in the active state exists as a homodimer. The homodimer is essential for the catalytic activity of M pro and in shaping the S1 pocket of the substrate-binding site. The active site of M pro features a cysteine amino acid residue, similar to the Cathepsins active site [30][31][32]. Therefore, we performed in silico docking analysis of DTBN to the substrate binding cleft (between Domain-1 and II) of M pro . Docking analysis revealed that the ligand, DTBN, is in close proximity to Cys145. At the active site (S1 subsite), a hydroxy group proximal to the tetrahydrothiophene ring of DTBN is hydrogen bonded with -NH hydrogen of Glu166 (2.13 Å) ( Figure 6). Similarly, the hydroxy hydrogen is hydrogen-bonded with carbonyl oxygen of the α-carboxylic acid of Glu166. However, the binding conformation of DTBN leads to three steric clashes, (i) between the tetrahydrothiophene ring of nupharidine and Cys145 sulphur, (ii) between the quinolizidine ring of nupharidine and imidazole ring of His41 and (iii) between tetrahydrothiophene sulfur in DTBN and the OH group of Met165. These destabilizing interactions likely contributed to the observed low binding energy (−35.87 kcal/mol). In comparison, the M pro inhibitor at the active active site is stabilized by 8 hydrogen bond interactions between co-crystalized inhibitor and amino acid residues, Glu166, His164, His163, Phe140, Gly143, Cys145, His41. The interactions of the co-crytalized inhibitor post docking were found to be similar to the crystal structure interaction, thus validating our docking protocol with deviation of RMSD 0.40 Å. In M pro , although DTBN is in close proximity to Cys145, the binding conformation of DTBN led to three steric clashes, which restricted effective access to M pro catalytic site. Thus, a mismatch of the binding conformation of DTBN with the M pro active site and associated steric clashes with important amino acid residues (catalytic dyad Cys145 and His41) of the active site contributed to its low binding energy (−35.87 Kcal/mol). Therefore, we may observe a low inhibitory effect against M pro .
The above in silico molecular docking results show that DTBN experiences a steric clash at the active site of Cathepsins B, L and M pro , whereas, at the active site of Cathepsin S, DTBN did not experience any such steric clashes. Binding energy metrics and the interactions at the active site indicate that DTBN has a relatively higher binding affinity towards Cathepsin S in comparison to Cathepsins B, L and M pro . In brief, DTBN may form a relatively more stable complex with Cathepsin S than with Cathepsin B, L and M pro . This was further validated by molecular dynamic (MD) simulation.

Covalent Docking Analysis
In order to evaluate the possible working mechanistic hypothesis, we performed a covalent docking of DTBN with Cathepsins of interest and M pro .
For covalent docking, possible intermediate DTBN (iDTBN) was generated from DTBN ( Figure 1), docked to the active site of Cathepsins B, L and S and evaluated for potential interactions with nucleophilic cysteine sulphur at the active site. A disulfide (S-S) covalent bond formation was observed between the iDTBN and nucelophilic cysteine sulphur at the active sites. This indicates that DTBN may possibly inhibit Cathepsins via the proposed working mechanistic hypothesis. However, other possibilities cannot be ignored. In addition to the covalent linkage between iDTBN and Cathepsin, the iDTBN was stabilized by various weak interactions, which contribute to the observed binding energy ( Table 2).  In Cathepsin B-iDTBN covalent adduct, iDTBN was stabilized by one hydrogen bond interaction and two C-H . . . O aromatic hydrogen bond interactions (Table 2)   In the Cathepsin S-iDTBN covalent adduct, iDTBN was stabilized by one hydrogen bond interaction, three C-H . . . O aromatic hydrogen bond interactions and one π-cation interaction ( Table 2). Unlike Cathepsin B and L-iDTBN covalent adduct, there no steric clash was observed in the Cathepsin S-iDTBN complexes. These stabilizing interactions likely contributed to the observed higher binding energy (Figure 9) as compared to Cathepsins B and L.  The above covalent docking results show that the elctrophilic sulphur of iDTBN and cysteine sulphur at the active site of Cathpesins and M pro can potentially form a disulfide bond, which could be considered as supporting the proposed working mechanistic hypothesis. We presume that DTBN may mediate its inhibitory action via iDTBN. Binding energy metrics and interactions show that DTBN may exhibit a gradient inhbitory effect against Cathepsins B, L, S and M pro , with activity towards Cathepsin S.

Molecular Dynamics (MD) Simulation
To determine the stability of docked poses of DTBN with Cathepsins and iDTBN-Cathepsins covalent adduct, and to find the stable interactions, a molecular dynamic (MD) simulation was performed. Both DTBN and iDTBN were subjected to 100 ns MD simulation with Cathepsins B, L and S and M pro .

Analysis of RMSD Value of Proteins and Ligands
The obtained MD simulation results were analysed using Cα and ligand RMSD values. The Root Mean Square Deviation (RMSD) value indicates the conformational changes in the ligand and protein from the initial structure during the simulation. For globular proteins, any RMSD value within 3 Å is acceptable [33]. Greater deviation from 3 Å implies that the protein in the protein-ligand complex is less stable due to the significant conformational changes that occurred during MD simulation. All seven protein-ligand complexes, Cathepsins B, L and S-DTBN, Cathepsins B, L and S-iDTBN and M pro -DTBN showed deviations in RMSD value well below 3 Å when compared to the initial frame throughout the trajectory (Figures 11-14). This indicates that the Cathepsins and M pro are stable in their respective protein-ligand complexes. Analysis of the ligand (DTBN and iDTBN) RMSD value against Cathepsins indicates that DTBN and iDTBN did not show a large deviation from their initial position during the MD run, signifying that their binding is stable. In contrast, DTBN shows a larger RMSD against M pro during MD simulation ( Figure 14). This indicates that it forms a less stable complex with M pro .    In Cathepsin L In Cathepsin L-DTBN MD simulation, there was one π . . . π stacking interaction that exceeded 30% of the minimum contact strength. The π . . . π stacking interaction was observed between the furan ring of DTBN and phenyl ring of amino acid residue, Trp189 ( Figure S16a in the Supplementary Materials). In the Cathepsin L-iDTBN covalent adduct, one hydrogen bond was found to persistently have 70% contact strength ( Figure S16b in the Supplementary Materials), indicating a strong interaction between Cathepsin L-iDTBN covalent adduct.
In Cathepsin S At the active site of Cathepsin S, DTBN was stabilized by a prominent H-bond interaction (71%) between the amino acid residues Val162 and hydroxy group of DTBN, and by a π . . . π stacking interaction (36%) between the phenyl group of amino acid residues Phe211 and furan ring of DTBN ( Figure S17a in the Supplementary Materials). At the active site of the S-iDTBN complex, iDTBN was stabilized by two interactions: (i) water bridged H-bond interaction (64%) between amino acid residues Asn163, water and the hydroxy group of iDTBN, and (ii) π . . . π stacking interaction (53%) between the phenyl group of amino acid residues Phe70 and furan ring of iDTBN ( Figure S17b in the Supplementary Materials).
In Mpro During the MD run of M pro -DTBN, DTBN was stabilized by a hydrogen bond interaction between the hydroxy group in DTBN and carbonyl oxygen α-carboxylic acid of amino acid residue, Glu166 which exceeded 40% of the minimum contact strength ( Figure S18 in the Supplementary Materials).
The above MD simulation results suggest that DTBN and iDTBN form a stable complex with Cathepsins B, L and S. An analysis of the existing interactions between protein-ligand complexes during the MD run show that DTBN and iDTBN can form a more stable complex with Cathepsin S in comparison with Cathepsins B and L. On the other hand, the RMSD of DTBN and molecular interactions between the M pro -DTBN complex in MD simulation show that DTBN binds less stably to M pro .
In the docking results, it was shown that the cysteine sulphurs of M pro and Cathepsin B, S and L are in close proximity to the nupharidine thiophene ring, potentially creating the conditions for a nucleophilic attack. The results presented here are compatible with the plausible mechanism of action of the 6,6 -dihydroxythiobinupharidine (DTBN), with the cysteine proteases proposed in Figure 1. This provides further support to the demonstration by Shenvi's group that the sulfur atom of the Nuphar thiaspirane iminium pharmacophore reacts as an electrophile with nucleophilic thiols at ambient temperature. In dihydroxy nupharidine dimers, although not in monohydroxydimers, this reactivity causes a rapid retrodimerization to electrophilic, unsaturated iminium ions, which was ascribed to retrodimerization and, potentially, a highly reactive covalent binder [19]. It was proposed that the sulfur electrophilicity in the iminium thiaspirane could lead to the design of new therapeutics based on the Nuphar pharmacophore and new tools for the selective capture of free thiols in a cellular setting [19].
The above in silico molecular docking analysis and MD simulation studies show the DTBN's potential ability to inhibit Cathepsins (B, L and S) and M pro . However, the protein-ligand stability and molecular interactions show that DTBN may exhibit gradient inhibitory activity against Cathepsins and M pro , with better activity against Cathepsin S in comparison to Cathepsin B and L and the least inhibitory activity against M pro . To validate the obtained molecular docking and MD simulation results, an in vitro assay was conducted, and the results were compared.

In Vitro Analysis Enzyme Activities
The inhibitory activity of DTBN against various proteases is shown in Table 3. There was a high specificity towards papain and cysteine proteases of the Cathepsins family B, L and S. However, there was no inhibition of calpain, which belongs to the family of calcium-dependent, non-lysosomal cysteine proteases, or towards trypsin, which is a serine protease and does not act against the SARS-CoV-2 M pro protease. Within the Cathepsins, there was a significant gradation of DTBN inhibitory activity; while Cathepsin S was strongly inhibited in the low micromolar range, with an IC 50 value of 3.2 µM, Cathepsins L was less inhibited, with an IC 50 values of 13.2 µM. Cathepsin B was least inhibited with an IC 50 value of 1359.4 µM ( Table 3). The specificity of inhibition by DTBN on cysteine proteases was demonstrated in this work. There were differences in the extent of inhibition, reflecting structural-steric and biochemical differences in the active site. Cathepsin S was most sensitive to inhibition by DTBN, compared to cathepsin B, L and papain. DTBN was inactive for the inhibition of SARS-CoV-2 M pro and non-cysteine proteases, Trypsin and Calpain.
A series of molecular docking simulations suggested a mechanism of interaction that supported the biochemical results. The binding energy metrics and protein-ligand interactions of the docked poses of DTBN with Cathepsins and M pro obtained from molecular docking studies clearly show that DTBN has a variable affinity towards M pro and Cathepesins B, L and S, although belonging to the cysteine proteases family. The difference in affinity was attributed to structural-steric and biochemical differences in the active site. The proposed possible working mechanistic hypothesis for DTBN is that the electrophilic sulphur of DTBN targets nucleophilic cysteine sulphur at the active site of Cathepsins (B, Land S) or M pro to form a disulphide bond. This was validated by a covalent docking study. The binding energy and molecular interaction suggested that DTBN has a stronger interaction with Cathepsin S in comparison with Cathepsins B, L and M pro . Further, a molecular dynamic simulation was performed to evaluate the stability of M pro -DTBN, the Cathepsins-DTBN complex, and their covalent adducts and the associated molecular interactions during the MD run. The stability of the protein-ligand complex and associated molecular interactions suggested that DTBN formed teh most stable complex with S, followed by Cathepsins L, B and, finally, M pro . The results obtained from molecular docking, covalent docking and molecular dynamic simulation studies revealed the potential ability of DTBN to inhibit Cathepsins B, L and S and Mpro. However, its activity may incline towards Cathepsin S in comparison to Cathepsins B, L and Mpro. This was validated by in vitro assay. The experimental results are in accordance with the predicted computational docking studies. ). Inconsistencies in the structure, such as missing hydrogen, incorrect bond orders and orientation of the different functional groups of amino acids, were rectified during the optimization process [34]. The prepared proteins were then used for molecular docking. The amino acid numbering is same as that of the crystal structure.

Ligand Preparation
The structure of 6,6 -dihydroxythiobinupharidine ligand was prepared prior to docking using the LigPrep application in Schrödinger Maestro Suite 2020 (LigPrep, Schrödinger, LLC, New York, NY, USA, 2020-3). LigPrep was used to perform the 2-dimensional to 3-dimensional conversion of structures, correct improper bond distances and bond orders, and generate ionization states and energy minimization processes. This ligand structure prepared by LigPrep was then used for further studies.

Molecular Docking and MM-GBSA Refinement
Molecular docking studies were performed using Schrödinger Maestro Suite 2020-3 (Schrödinger, LLC, New York, NY). DTBN was docked to the active site of Cathepsins B, L and S. Cathepsins B and L the center of the grid was generated against the centroid of previously known active site residue given in Table 4. For Cathepsin S, the co-crystallized ligand was used as the centeroid of the grid. In Mpro, the crystal structure of the original ligand was removed from the complex and Cys145 residues were selected to function as the centroid of ligand binding site definition, as they play a crucial role in catalytic dyad. This was further used to generate the receptor grid with an internal ligand diameter midpoint box of the size 15 × 15 × 15 Å 3 . A postdocking simulation MM-GBSA refinement was carried out on the docked poses with a flexible residue distance of 5 Å.

Covalent Docking
To establish a covalent bond linkage between the protein and iDTBN (Figure 1) Schordinger's CovDock application was used. For Cathepsin B, Cys29 was selected to act as a reactive residue. Similarly, for Cathepsin L and S, Cys25 was selected to act as a reactive residue in both cases. In M pro , Cys145 was selected to act as the reactive residue. The aforementioned cysteine residues were used to act as the center of the docking grid and docking ligands, with a length of less than 20 Å. The pose-prediction docking mode was utilized, and post-docking MMGBSA scoring was also performed.

Molecular Dynamics (MD) Simulation
MD simulation was performed using a DESMOND module by Schordinger (Schordinger relase 2021). Through the system builder panel, DTBN and iDTBN with a protein complex were immersed in an orthorhombic box of TIP3P solvent model. The solvated system was neutralized using counter ions and physiological NaCl salt concentration of 0.15 M. OPLS4 force field was utilized. The simulation was 100 ns using NPT assemble class at a temperature of 300 K and atmospheric pressure of 1.013 bar.

In Vitro Methodology Enzymes Assays
Enzyme inhibition assays were performed in triplicates using transparent 96-well plates for Calpain, Trypsin, Papain, Cathepsin L and Cathepsin B (167008; Nunc, Roskilde, Denmark), or black ones for Cathepsin S, (Greiner 96 Polystyrol GRE96fb), according to manufacturers' instructions. Briefly, a reaction mixture of 100 µL contained: 45µL of enzyme in assay buffer, 5 µL of inhibitor diluted in DMSO. 50 µL substrate prepared in assay buffer according to manufacturers' instructions. The inhibitor was serially diluted in DMSO to concentrations from 0.0075 µg/µL to 0.0000075 µg/µL. A total of 50 µL DMSO was added to the control sample. The enzyme substrate mix was added last, just before starting the fluorometric measurement. Data were collected for 1.5 h at 60 sec intervals.
The assay was conducted according to [35]. Compounds were seeded into assayready plates (Greiner 384 low volume 784900) using an Echo 555 acoustic dispenser and DMSO was back-filled for a uniform concentration in assay plates (maximum 1%). Reagents for M pro assay were dispensed into the assay plate in 10 µL volumes for a total of 20 µL. Final reaction concentrations were 20 mM HEPES pH = 7.3, 1 mM TCEP, 50 mM NaCl, 0.01% Tween-20, 10% glycerol, 5 nM M pro 375 nM fluorogenic peptide substrate ([5-FAM]-AVLQSGFR-[Lys(Dabcyl)]-K-amide). M pro was pre-incubated for 15 min at room temperature with compound before the addition of substrate. Protease reaction was measured continuously in a BMG Pherastar FS with a 480/520 ex/em filter set. Data were mapped and normalized in the Genedata Screener.

Conclusions
Cathepsins, a family of proteases expressed in all living organisms, including humans, participate in protein degradation, cell signaling and protein-processing. Abnormal expression and/or activity of Cathepsins has been associated with a variety of human diseases, including cancer, inflammatory and cardiovascular diseases, neurodegenerative disorders, viral diseases, and many others. 6,6 -Dihydroxythiobinupharidine (DTBN), a bioactive sesquiterpene thioalkaloid, is present in yellow water lily, Nuphar lutea (L.) Sm.
(Nymphaeaceae) was investigated as an inhibitor of Cathepsins and other proteases for comparative purposes. DTBN shows differences in the extent of Cathepsins inhibition, reflecting structural-steric and biochemical differences in the active site. Cathepsin S was most sensitive to inhibition by DTBN compared to Cathepsin B, L and papain. DTBN was inactive for the inhibition of SARS-CoV-2 M pro of COVID-19 and non-cysteine proteases, Trypsin and Calpain. Docking simulations suggested a mechanism of interaction that was supported by the biochemical result. In the docking results, it was shown that the active site cysteine sulphur of Cathepsin S, L and B is in close proximity to the nupharidine tetrahydrothiophene ring, potentially creating the necessary conditions for a nucleophilic attack. Covalent docking of intermediate DTBN (iDTBN) to the active site of Cathepsins led to a disulfide bond between the electrophilic sulphur of iDTBN and nucleophilic cysteine sulphur of Cathepsins. This explains the potential of electrophilic sulphur in DTBN as active warhead that inhibits nucleophilic cysteine sulphur containing enzymes. Further, molecular dynamic (MD) simulation revealed the stability of protein-ligand complexes, where Cathepsin S formed the more stable complex with DTBN and iDTBN in comparison with Cathespsins B and L. The order of stability was in accordance with in vitro results, emphasizing the importance of the stability of the protein-ligand complex for inhibition. DTBN's lack of reactivity against SARS-CoV-2 M pro was attributed to a mismatch of binding conformation of DTBN to the catalytic binding site of SARS-CoV-2 M pro . Our results may be conducive for a mechanism-based search of derivatives of nupharidine against COVID-19. This could be an alternative strategy to the large-scale screening of electrophilic inhibitors.