Mechanism of Microbial Metabolite Leupeptin in the Treatment of COVID-19 by Traditional Chinese Medicine Herbs

ABSTRACT Coronavirus disease 2019 (COVID-19) has caused huge deaths and economic losses worldwide in the current pandemic. The main protease (Mpro) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is thought to be an ideal drug target for treating COVID-19. Leupeptin, a broad-spectrum covalent inhibitor of serine, cysteine, and threonine proteases, showed inhibitory activity against Mpro, with a 50% inhibitory concentration (IC50) value of 127.2 μM in vitro in our study here. In addition, leupeptin can also inhibit SARS-CoV-2 in Vero cells, with 50% effective concentration (EC50) values of 42.34 μM. More importantly, various strains of streptomyces that have a broad symbiotic relationship with medicinal plants can produce leupeptin and leupeptin analogs to regulate autogenous proteases. Fingerprinting and structure elucidation using high-performance liquid chromatography (HPLC) and high-resolution mass spectrometry (HRMS), respectively, further proved that the Qing-Fei-Pai-Du (QFPD) decoction, a traditional Chinese medicine (TCM) formula for the effective treatment of COVID-19 during the period of the Wuhan outbreak, contains leupeptin. All these results indicate that leupeptin at least contributes to the antiviral activity of the QFPD decoction against SARS-CoV-2. This also reminds us to pay attention to the microbiomes in TCM herbs as streptomyces in the soil might produce leupeptin that will later infiltrate the medicinal plant. We propose that plants, microbiome, and microbial metabolites form an ecosystem for the effective components of TCM herbs.

C oronavirus disease 2019 (COVID- 19) has caused a huge threat to global public health (34,35). Moreover, the constant emergence of variants of concern (VOCs) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may bring the risk of vaccine and antibody failure, such as VOC 501Y.V1 (a variant) or 501Y.V2 (b variant) (1). Thus, there is a demand to develop antiviral drugs. At present, many HIV and hepatitis C virus (HCV) drugs targeting the viral protease have been approved. SARS-CoV-2 can also produce two proteases, main protease (M pro ) and papain-like protease (PL pro ), which hydrolyze polyproteins 1a and 1ab for virus replication and maturation. M pro cuts 11 sites of viral polyproteins 1a and 1ab to obtain 12 proteins, such as the nsp7-nsp8-nsp12 complex RNA-dependent RNA polymerase (RdRp), nsp13 helicase-triphosphatase, and nsp14-nsp10 complex S-adenosyl methionine (SAM)-dependent (guanine-N7) methyltransferase (2). Therefore, M pro is thought to be an ideal target for treating COVID- 19.
Leupeptin (Fig. 1A) is a secondary metabolite produced by actinomycetes (3)(4)(5)(6), which has broad-spectrum protease-inhibitory activities, including serine, cysteine, and threonine proteases (7)(8)(9)(10)(11). Leupeptin treatment can prevent or delay muscular dystrophy in mouse or chicken models (12). Rabbits treated with leupeptin did not develop experimental autoimmune myasthenia gravis despite repeated acetylcholine receptor injections (13). These studies showed that leupeptin has a certain clinical value. In addition, human coronavirus strain 229E can be inhibited by leupeptin in plaque reduction tests (14). Here, we found that leupeptin could also inhibit M pro of SARS-CoV-2. Interestingly, high-performance liquid chromatography (HPLC) fingerprinting and high-resolution mass spectrometry (HRMS) showed that the Qing-Fei-Pai-Du (QFPD) decoction, a traditional Chinese medicine (TCM) formula, contains leupeptin. These results reveal that leupeptin may contribute to the QFPD decoction in the treatment of COVID-19, which had been popularly used in Wuhan in the first wave of the outbreak.

RESULTS AND DISCUSSION
SARS-CoV-2 M pro -inhibitory activity. In order to verify whether leupeptin has inhibitory activity against M pro , we expressed and purified M pro of SARS-CoV-2 according to our previous work (15,16). Fluorescence resonance energy transfer (FRET) enzymatic assays showed that leupeptin has some inhibitory activity against M pro , with a 50% inhibitory concentration (IC 50 ) value of 127.2 mM (Fig. 1B). However, the progress curve of substrate hydrolysis did not indicate a clear time-dependent inhibition of leupeptin against M pro (see Fig. S1 in the supplemental material). Therefore, whether leupeptin is a covalent inhibitor of M pro still needs to be further evaluated, which is addressed in the following experiments.
Inhibitory mechanisms of leupeptin. We further determined the crystal structures of M pro -leupeptin at resolutions of 1.7 Å to figure out the inhibitory mechanisms of leupeptin ( Fig. 1C and Table S1). A room-temperature crystal structure of M pro -leupeptin at a lower resolution of 2.2 Å has also been solved (17), and during the preparation of the manuscript, a similar complex structure, also at a 1.7-Å resolution, was reported (18). We found that the mode of binding of leupeptin with M pro was almost the same in these three structures except for the conformation of the P1 guanidine residue. The unambiguous electron density maps show that leupeptin binds in the active site of M pro (Fig. S2). The mode of binding of leupeptin with M pro is different from that of leupeptin with other cysteine proteases, such as papain (Fig. 1E) or human calpain-3 (Fig. 1F). In the complex structure, the Sg atom of the nucleophilic Cys145 in M pro forms a C-S covalent bond with the aldehyde carbon of leupeptin ( Fig. 1D and Fig. S2). The hemiacetal oxygen of leupeptin faces the "oxyanion hole" with the formation of two hydrogen bonds with the main-chain amides of Cys145, Asn141, and Gly142, which is similar to the trypsin-leupeptin complex (11) (Fig. 1D). But this result is inconsistent with the results of the substrate hydrolysis progress curve. The reason for this may be that the aldehyde group of leupeptin can reversibly react with Cys145 of M pro by an attack of water or other nucleophiles. The negatively charged guanidine residue can be tolerated by the S1 subsite and form one hydrogen-bonding interaction with Glu166. The backbone of leupeptin forms two other hydrogen-bonding interactions with Glu166 and His164 (Fig. 1D). The hydrophobic P2 isobutyl (i-Bu) group of leupeptin can fit into the hydrophobic S2 subsite well. On the contrary, the hydrophobic P3 i-Bu group was exposed to the solvent, which seems to have no interactions with the surrounding amino acids of the S3 subsite. In addition, the P4 acetyl residue cannot fully occupy the S4 subsite. All of these factors are responsible for the relatively weak inhibitory activity.
Antiviral potency of leupeptin against SARS-CoV-2. The IC 50 value of leupeptin in plaque reduction tests was 0.4 mg/ml (about 1 mM) against human coronavirus strain 229E (14). Combined with the SARS-CoV-2 M pro activity inhibition results, this indicates that leupeptin may inhibit the replication of SARS-CoV-2 in cells. Subsequent antiviral assays showed that the RNA levels of SARS-CoV-2 can be inhibited by leupeptin in Vero  Fig. 2A). Moreover, viral RNA (vRNA) copy numbers dropped significantly in the presence of 40 mM leupeptin (Fig. 2B). Leupeptin can inhibit human coronavirus strain 229E better than SARS-CoV-2 in cells. The possible reason for this is that leupeptin inhibits trypsin and cathepsin L involved in the 229E virus entry process (19). However, the entry of SARS-CoV-2 needs TMPRSS2 and furin-like proteases instead of cathepsin (20).
HPLC fingerprinting and HRMS analysis. Leupeptin and leupeptin analogs were isolated from various strains of Streptomyces spp., such as Streptomyces roseus MA839-A1 (5), Streptomyces exfoliatus SMF13 (4), Streptomyces griseus strain 254 (3), and Streptomyces lavendulae (21). Autogenous protease inhibitors can regulate cellular turnover metabolism associated with secondary metabolism and morphogenesis by inhibiting the activity of the protease (22). S. roseus strains have been isolated from rhizospheric soil samples under cultivation in Curcuma longa L (23). S. griseus strains have been isolated from many habitats, including soils, marine sediments, and medicinal plants such as Kandelia candel (24). S. lavendulae strains have also been isolated from soils and Ginkgo biloba (25). Streptomyces accounts for 60% of endophytic actinobacteria isolated from Glycyrrhiza inflata Batal (26). In general, Streptomyces and medicinal plants have a broad symbiotic relationship. On the other hand, TCM composed of medicinal plants showed potential for the treatment of COVID-19 (27,28). Thus, we suspected that some TCM formulas for the treatment of COVID-19 may contain leupeptin. The QFPD decoction, a TCM formula, has shown therapeutic effects on mild, ordinary, and severe COVID-19 patients (29). Early treatment with QFPD was associated with the recovery of patients with COVID-19, including faster recovery, a shorter time to viral shedding, and a shorter duration of hospital stay. The overall response rates for QFPD were over 90% among 782 cases in the clinic (30). This TCM formulation was selected as one of the recommended therapeutic regimens by the National Administration of Traditional Chinese Medicine (NATCM) and is widely used in China (31,32). QFPD contains 20 kinds of herbs and a kind of mineral, which is mainly derived from 4 different classical prescriptions in the ancient work Shang Han Lun.
To characterize whether leupeptin is one of the substances of QFPD, we applied HPLC and HRMS to both the leupeptin reference standard and the QFPD decoction. Chromatogram of the leupeptin reference standard with two retention times, 5.203 and 9.255 min, at a 206-nm UV absorption, caused by the tautomeric isomers (33), is exhibited in Fig. 3A. Corresponding chromatographic peaks at retention times of 5.468 and 9.120 min were also found by chromatographic fingerprinting of the QFPD aqueous solution (Fig. 3B). HRMS was used for the further structure elucidation of leupeptin in QFPD as well as the reference standard. The observed protonated precursor ion [M 1 H] 1 of the leupeptin reference standard was m/z 427.2969 (Fig. 3C), while that of the QFPD solution was m/z 427.2968 (Fig. 3D). The mass difference was 0.2 ppm, which shows that the precursor ions were likely identical. Moreover, the major fragment ions shown in the tandem mass spectrometry (MS/MS) spectrum of the QFPD decoction highly correspond to those of the reference standard (Fig. 3C to E and Fig. S3). These results confirm that leupeptin is one of the substances in the QFPD decoction.
In a word, we have identified leupeptin, a metabolite produced by PSA, which showed antiviral activity in both enzyme assays and cell culture. Moreover, we identified that leupeptin is one of the substances in the QFPD decoction and well characterized by HPLC fingerprinting and HRMS. These results suggest that leupeptin likely contributes to the antiviral activity of the QFPD decoction against SARS-CoV-2. Also, the PSA metabolite leupeptin can transfer from microbes to medicinal plants (Fig. 4). Thus, this reminds us to take the role of microbiomes in TCM modernization seriously. Our study has Data collection and structure determination. Diffraction data were collected at Shanghai Synchrotron Radiation Facility (SSRF) BL17U (wavelength, 0.97918 Å). For data collection, the crystals were cryoprotected by briefly soaking them in reservoir solution supplemented with 20% (vol/vol) glycerol before flash-cooling in liquid nitrogen. The data set was processed with HKL2000 software. The native M pro complex structure was determined by the molecular replacement method using Phaser with the previously reported SARS-CoV-2 M pro structure (PDB accession number 7C6S). The atomic models were completed with Coot and refined with phenix.refine in Phenix, and the stereochemical qualities of the final models were assessed with MolProbity. Data collection, processing, and refinement statistics are summarized in Table S1 in the supplemental material. All structural figures were generated using PyMOL software (http://www.pymol.org).
Cells and viruses. African green monkey kidney Vero cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA), 200 mg/ml streptomycin, and 200 IU/ml penicillin at 37°C. COVID-19 virus was isolated from the Wuhan seafood market by the China CDC. All the infection experiments were performed in a biosafety level 3 (BLS-3) laboratory.
In vitro antiviral assays. A total of 20 mM leupeptin hemisulfate (catalog number T6564; TargetMol) in deionized water was diluted to 200 mM to 0.06 mM with DMEM containing 1% FBS. Vero cells cultured overnight in 96-well plates were infected by virus at a multiplicity of infection (MOI) of 0.01 for 2 h. The medium was removed, and fresh drug-containing medium was then added to the cells. After 48 h, the cells were lysed in lysis buffer. The viral RNA in 100 ml of the cell supernatant was quantified by reverse transcription-PCR (RT-PCR). Seventy-two hours later, the changes of cytopathic effect were also observed by microscopy. Experiments were performed in triplicate. The experimental results were processed using MS Excel and GraphPad Prism 8.0.
HPLC analysis. A total of 0.1 g/ml of QFPD powder (catalog number JZT-QFPDT-0318-PG-0321; Jointown Pharmaceutical Group, Wuhan, China) and 1 mg/ml leupeptin (catalog number L5793; Sigma-Aldrich, St. Louis, MO, USA) in ultrapure water were prepared. Chromatographic fingerprinting analysis of leupeptin and QFPD was performed by using an HPLC system (Agilent 1260 Infinity II Prime). The column was a Zorbax Eclipse Plus 95-Å C 18 HPLC column (2.1 by 150 mm, 3.5 mm) (catalog number 959763-902; Agilent, CA, USA). The solvent system consisted of 0.1% (vol/vol) phosphoric acid in MilliQ water as component A, while component B was acetonitrile and was run at a flow rate of 1 ml/min. The solvent gradient was begun with 10% B for 0 to 10 min, 10% to 30% B for 10 to 85 min, and 30 to 50% B for 85 to 110 min, followed by an equilibration time of 10 min with 10% B. The temperature of the column was 40°C, and the target tripeptide was successfully detected at a wavelength of 206 nm (the maximum UV absorption wavelength of leupeptin). Injection volumes of 10 ml of each sample and the standards were used for the HPLC analysis, and the supernatant liquid of the QFPD sample solution and the leupeptin standard solution was obtained by centrifugation at 20,133 Â g.
High-resolution mass spectrometry analysis. MS-grade methanol (MeOH) was purchased from Merck (Darmstadt, Germany). Approximately 0.1 g QFPD powder was added to 1.0 ml MeOH to extract leupeptin. The complex was then vortexed for 30 s and sonicated for 30 min, followed by 20 min of centrifugation at 16,000 Â g at 4°C. The extractant and 500 ng/ml of the leupeptin solution were directly infused into an electrospray ionization (ESI)-Q-orbitrap MS instrument (Thermo Q-Exactive Plus) at a flow rate of 10 ml/min for leupeptin identification. The analytes, which were monitored under a targeted data-dependent MS2 (ddMS2) mode with positive ionization, were ionized with a vaporizing temperature of 350°C and a spray voltage of 3,800 V. The MS/MS analysis was conducted with a collision energy of 35 V, a resolution of 70,000 full width at half maximum (FWHM), an automatic gain control (AGC) target of 5e4, a maximum inject time of 100 ms, and an isolation window of m/z 2.0. Xcalibur 4.1.50 software was employed to acquire and process MS and MS/MS data.
Data availability. Coordinates and structure factors have been deposited in the Protein Data Bank under accession number 7EIN.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.

ACKNOWLEDGMENTS
We thank the staff of the BL17U beamlines at the Shanghai Synchrotron Radiation Facility. We also thank Ying Huang and Minghao Liu from the Institute of Microbiology, Chinese Academy of Sciences, for providing professional advice on actinomycetes. We appreciate the kind help of Yue Gao for obtaining QFPD powder and Boxin Zhao for chromatographic testing.
This work was supported by the Strategic Priority Research Program of CAS (XDB29010202), the intramural special grant for SARS-CoV-2 research from the Chinese Academy of Sciences,