Main

The coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has affected nearly 290 million patients with more than 5 million deaths as of late December 2021 (https://covid19.who.int/). Despite great success in vaccines, unexpected mutations of the SARS-CoV-2 virus genome bring substantial challenges to the effectiveness of current vaccines1. Particularly, the Omicron2 variant has been demonstrated to be able to escape the protection of current vaccines3. Moreover, the likelihood of SARS-CoV-2 to exist for a long time4 highlights the critical importance of safe and effective antivirals. Unfortunately, treatment options for COVID-19 remain limited5. Considering the current situation of the pandemic and possible future needs, drug discovery for SARS-CoV-2 must be expedited as much as possible.

The main protease (Mpro, also named 3CL protease) of SARS-CoV-2 plays a key role in the viral life cycle and is conserved among various variants of concern (Supplementary Fig. 1)6,7,8. Thus, Mpro is a promising target for antiviral drug development. Currently, a number of SARS-CoV-2 Mpro inhibitors have been reported9,10,11,12,13,14,15,16,17,18,19,20,21. The majority of them belong to peptidomimetics, which are often linked to poor pharmacokinetic (PK) properties. Recently, several research efforts have been made to improve the PK properties of peptidomimetic Mpro inhibitors9,10,11,12,13,14,15,16,17,18, but limited progress has been achieved for optimal oral bioavailability13,17, with only one orally administrated compound (PF-07321332) just being approved for clinical use by the US Food and Drug Administration (FDA)17, and one (S-217622) still in late clinical trials19. Orally available antivirals are highly desirable as patients can be treated as outpatients without overburdening the healthcare system. We herein report an orally available SARS-CoV-2 Mpro inhibitor with potent in vivo antiviral activity against wild-type (WT) and emerging variants of SARS-CoV-2.

Results

Discovery of SARS-CoV-2 Mpro inhibitors

To rapidly obtain lead compounds, we utilized the classical one-pot Ugi four-component reaction (Ugi-4CR, Fig. 1a)22,23. This reaction fuses a carboxylic acid (R-CO2H), an amine (R1-NH2), an aldehyde (R2-CHO) and an isocyanide (R3-NC) to generate diamine derivatives with R, R1, R2 and R3 linking the terminal acyl, the middle chiral carbon and amino nitrogen, and the terminal amino nitrogen, respectively. The Ugi-4CR has been previously adopted to produce non-covalent Mpro inhibitors15,24. Despite many efforts, the potency of generated diamine derivatives is still at the micromolar or sub-micromolar level15,24. Considering the advantages of covalent inhibitors in prolonging residence time and their ability to compete with high-affinity natural substrates25, we decided to design covalent inhibitors on the basis of the Ugi-4CR.

Fig. 1: Schematic diagram of the design of SARS-CoV-2 Mpro inhibitors.
figure 1

a, The classical one-pot Ugi four-component reaction (Ugi-4CR). b, Fragments R, R1, R2 and R3 in Ugi-4CR occupy the S1′, S1, S2 and S4 pockets of Mpro, respectively.

Full size image

According to the reported co-crystal structure of SARS-CoV-2 Mpro and diamines15, fragments R, R1, R2 and R3 occupy the S1′, S1, S2 and S4 pockets of Mpro, respectively (Fig. 1b). In the first step, we managed to find a proper electrophilic group (warhead) at the R position because it binds next to a cysteine residue (Cys145). To reduce potential toxicity, we chose warheads with mild or moderate electrophilicity. Here, α-ketoamide (R = acetyl) and acrylamide (R = vinyl) were selected as candidates because compounds carrying these warheads have been approved for use in the clinic25, in addition to their lower reactivity. Two compounds (‘1a’ and ‘1b’, Extended Data Fig. 1a) were synthesized with R being acetyl or vinyl, and R1, R2 and R3 fixed as pyridine, tert-butylbenzene and tert-butyl, respectively, similar to groups in ML188, a non-covalent diamine derivative Mpro inhibitor24. Bioactivities of the two compounds against SARS-CoV-2 Mpro were determined by a fluorescence resonance energy transfer (FRET) assay and a differential scanning fluorimetry (DSF) assay. The α-ketoamide derivative ‘1a’ (racemic mixture) showed good activity with a 50% inhibitory concentration (IC50) of 3.11 μM in the FRET assay and a thermal shift (∆Tm) of 4.7 °C in the DSF assay (Extended Data Fig. 1a). However, the acrylamide derivative ‘1b’ displayed no activity (IC50 > 500 μM). A possible reason could be that the thiol of the Cys145 side chain prefers to react with the α-position carbon due to spatial location restriction, rather than the β-position carbon that is favoured in the common Michael addition reaction15,16. We then synthesized another 3 compounds with a bulkier α-ketoamide warhead, which exhibited substantially reduced activity (Extended Data Fig. 1a), implying that a bulky α-ketoamide warhead is not favourable. Therefore, the α-ketoamide with R = acetyl was finally selected as the warhead. Subsequently, racemic mixture ‘1a’ was separated by chiral high performance liquid chromatography (HPLC). (R)-‘1a’ showed an IC50 of 2.38 μM, but (S)-‘1a’ displayed much weaker activity with an IC50 of 67.67 μM (Extended Data Fig. 1b). We solved the co-crystal structure of (R)-‘1a’-Mpro (PDB entry 7FAY), which confirms the (R)-configuration of the active stereoisomer of ‘1a’ (Extended Data Fig. 1c). Meanwhile, this co-crystal structure shows that the α-ketoamide group forms a covalent bond with Cys145, and pyridine (R1), tert-butylbenzene (R2) and tert-butyl (R3), occupy the S1, S2 and S4 sites of Mpro, respectively, consistent with our expectation.

Then, a stepwise optimization of R1, R2 and R3 of ‘1a’ was carried out, which provided a number of optimal fragments with one fragment for R1, four for R2 and four for R3 (Extended Data Fig. 2a and Supplementary Figs. 2–4). Although the pyrazinyl group at R1 was found to have a better performance in improving potency (3f, Supplementary Fig. 3), it was not included because compounds containing this group at R1 are unstable in alcohol solutions (Supplementary Fig. 5). To find better compounds, we further synthesized 16 compounds with various combinations of optimal fragments at R2 and R3 (Extended Data Fig. 2b). All these compounds showed potent activity against SARS-CoV-2 Mpro; they are epimeric mixtures due to the introduction of another stereocenter with a fixed S-configuration in R3. The epimeric mixtures for these 16 compounds were subsequently separated by chiral HPLC. (R)-epimers (for simplicity, a simple nomenclature is used hereafter: (R)-/(S)-epimer, where (R)-/(S)- indicate the configurations of the chiral carbon linking the two amides) for all these compounds displayed much higher potency than the corresponding (S)-epimers. Among these compounds, (R)-‘5b’ showed the highest potency against Mpro with an IC50 of 11.3 nM. Again the (S)-epimer (S)-‘5b’ displayed weaker activity (IC50 = 751.9 nM). We then examined whether (R)-‘5b’ is prone to epimerization in vivo. Our results showed that the most active epimer (R)-‘5b’ could rapidly convert to the less active epimer (S)-‘5b’ in vivo (Extended Data Fig. 3); this could be due to the presence of an exchangeable hydrogen in the chiral carbon centre linking the two amides26.

To prevent or reduce configuration conversion, we used deuterium to replace the exchangeable hydrogen. Deuterium has previously been explored to stabilize interconverting enantiomers, and has been shown to neither affect the pharmacological properties of a compound nor present a safety concern27,28. The generated compound, d-(R)-‘6a’ (a compound name with a ‘d-’ prefix indicates deuterium other than hydrogen attached to the chiral carbon linking the two amines; Extended Data Fig. 4), showed similar potency as (R)-‘5b’. Importantly, conversion from (R)- to (S)-epimers was substantially reduced (Extended Data Fig. 3). However, the PK properties of d-(R)-‘6a’ are poor, particularly the low oral bioavailability (10.1%) (Supplementary Table 1). We thus further optimized the chemical structure to improve oral bioavailability. A total of 11 compounds with deuterium were synthesized (Extended Data Fig. 4). All these compounds showed potent activity, with IC50 values ranging from 8.1 to 25.4 nM. The PK properties of these compounds were then evaluated in rats (Supplementary Table 1). Two compounds, d-(R)-‘6c’ and d-(R)-‘6j’, displayed good PK properties with oral bioavailabilities >30%. Because d-(R)-‘6c’ showed more potent activity than d-(R)-‘6j’ in both enzymatic (8.1 nM vs 20.4 nM, Extended Data Fig. 4) and cellular antiviral assays (Extended Data Fig. 5), d-(R)-‘6c’ (hereafter called Y180) was finally selected to carry out further evaluation; this compound also showed low conversion rate (from (R)- to (S)-epimers) in vivo (Extended Data Fig. 3).

As Y180 is a covalent inhibitor, we next determined the equilibrium-binding constant Ki and the inactivation rate constant kinac to further verify its binding with Mpro. The measured Ki and Kinac values of Y180 are 1 nM and 2.6 × 10−4 s−1, respectively (the corresponding values for PF-07321332, which is also a covalent Mpro inhibitor, are 2 nM and 5.1 × 10−4 s−1, respectively, in the same assay; Supplementary Table 2). These data confirmed that Y180 is a potent Mpro inhibitor. Additionally, Y180 showed almost no activity against several common mammalian proteases with similar structure, including caspase 2, chymotrypsin, thrombin, cathepsin B, cathepsin D and cathepsin L (IC50 > 100 μM; Supplementary Table 3). Collectively, all these data indicate that Y180 is a potent and selective Mpro inhibitor.

Crystal structure of SARS-CoV-2 Mpro in complex with Y180

To illustrate the binding mode of Mpro-Y180, we determined the crystal structure of Mpro with Y180 at 2.1 Å (PDB entry 7FAZ). In this structure, two protomers per asymmetric unit form the biological dimer of Mpro (Fig. 2a). The detailed interaction mode of Mpro with Y180 is discussed below (Fig. 2b–d). The terminal carbonyl moiety (warhead) of inhibitor Y180 (Fig. 2b) forms a 1.8 Å covalent bond with the sulfur atom of catalytic residue Cys145 (Fig. 2c). The hydroxyl group of this thiohemiketal points into the traditional oxyanion hole stabilized by a hydrogen bond with the main-chain amide of Cys145 (Fig. 2d). In addition, the backbone amide of the residue Gly143 forms another hydrogen bond with the oxygen of the α-keto moiety. The dibenzo(b,d)furyl group of Y180 inserts deeply into the S2 pocket. Importantly, this dibenzo(b,d)furyl moiety displays parallel-displaced π-π stacking interactions with His41 and an intramolecular 4-fluorophenyl ring. Moreover, the dibenzo(b,d)furyl group is further stabilized through hydrophobic interactions with Cβ of His41 and the side chain of Met49 (Fig. 2d). The pyridyl ring of inhibitor Y180 occupies the S1 pocket, forming a 2.9 Å hydrogen bond with a side chain of His163. Meanwhile, the backbone oxygen between the pyridyl ring and the 1-(4-fluorophenyl)ethyl group makes a hydrogen bond with the main-chain amide of Glu166. In the last moiety of Y180, the 4-fluorophenyl ring is stabilized through the π-π stacking interactions with the dibenzo(b,d)furyl group (mentioned above) and the hydrophobic interactions with Cε of Met49, and the Cβ and Cγ of Gln189.

Fig. 2: Overall structure of SARS-CoV-2 Mpro in complex with Y180.
figure 2

a, One protomer of the Mpro dimer is shown in surface view in wheat, the other displayed in cartoon view in magenta. The active site Cys145 is presented in yellow in the former protomer. Y180 in both molecules is shown with a stick model in cyan. The N and C termini of the cartoon-view protomer are labelled. b, The chemical structure of Y180. The warhead carbon is marked with a red triangle. Deuterium is shown in blue. c, The compound Y180 and Cys145 forming a covalent bond are displayed with a ball-and-stick model in cyan and yellow, respectively. Fo − Fc density map (σ = 2.5) is shown in grey. d, Detailed interactions of Mpro and Y180. Hydrogen bonds between them are shown as black dotted lines. The residues interacting with Y180 are labelled. a, c and d were prepared using PyMOL (https://pymol.org).

Full size image

Although Mpro is highly conserved in the emerging SARS-CoV-2 variants, there are still a number of mutated amino acids found in several variants (Supplementary Fig. 1), for example, Lys90Arg in B.1.351 (Beta), Lys90Arg and Ala193Val in B.1.351.2 (Beta), Lue205Val in P.2 (Zeta) and Pro132His in B.1.529 (Omicron). However, all these residues are far away from the Y180 binding site (>10 Å; see Supplementary Fig. 6), thus not affecting the inhibition efficiency of Y180 against Mpro (Supplementary Table 4).

In vitro antiviral activity of Y180

To evaluate the antiviral activity of Y180 in vitro, we used the VeroE6 cells stably expressing transmembrane serine protease 2 (VeroE6-TMPRSS2) and the human lung epithelial cell line, Calu3. The cells were challenged with WT SARS-CoV-2 and three other emerging variants B.1.1.7 (Alpha), B.1.617.1 (Kappa) and P.3 (Theta)29. Our data showed that Y180 did not cause observable cytotoxicity in VeroE6-TMPRSS2 and Calu3 cells within our treatment time frame (24 h) (Extended Data Fig. 6). Importantly, 20 µM Y180 decreased virus replication by 4.44-, 5.83-, 4.74- and 6.03-log against WT, B.1.1.7, B.1.617.1 and P.3, respectively, in VeroE6-TMPRSS2 cells at 24 h post infection (Fig. 3a–d). Similarly, Y180 at 20 µM demonstrated robust antiviral efficacy against SARS-CoV-2 in Calu3 cells with 2.14-, 3.13-, 1.77- and 2.29-log reduction of viral gene copies against SARS-CoV-2 WT, B.1.1.7, B.1.617.1 and P.3 variants, respectively (Fig. 3e–h). We next evaluated the efficacy of Y180 in inhibiting infectious virus progeny production with plaque reduction assays. Our results demonstrated that Y180 effectively reduced plaque formation of SARS-CoV-2 in both number and size, with EC50 of 11.4 nM, 20.3 nM, 34.4 nM and 23.7 nM against WT SARS-CoV-2, B.1.1.7, B.1.617.1 and P.3, respectively (Fig. 3i–l). To investigate whether suppression of virus replication would benefit the infection outcome of host cells, luciferase-based cell viability assay was carried out. Our results showed that Y180 substantially improved cell viability starting at 0.8 µM (Fig. 3m–p). At this concentration, Y180 demonstrated potent effects in protecting against SARS-CoV-2-induced cell death and the protection efficiency was higher than that of remdesivir (RDV) for WT (60.7% vs 43.9%, P < 0.0001), B.1.1.7 (106.6% vs 32.3%, P < 0.0001), B.1.617.1 (87.1% vs 14.6%, P < 0.0001) and P.3 (81.4% vs 67.1%, P = 0.0136). In addition to the comparison with RDV, we similarly side-by-side compared the Y180 antiviral efficacy with that of PF-07321332. When compared with the mock-treated control, Y180 effectively reduced viral titre by 5.7-log at 0.8 µM, while PF-07321332 decreased the viral load by 0.2-log at this concentration (Y180 vs PF-07321332, P < 0.0001; Extended Data Fig. 7). We further tested the efficacy of Y180 against the B.1.1.529 (Omicron) variant in vitro. Y180 decreased B.1.1.529 viral gene copies by 3.92-log at 20 µM compared with the vehicle control (Extended Data Fig. 8). Together, our data demonstrated that Y180 rendered superb cross-protection against WT and different emerging SARS-CoV-2 variants, including the latest Omicron variant, by potently abolishing virus replication.

Fig. 3: Evaluation of in vitro antiviral activity of Y180.
figure 3

a–d VeroE6-TMPRSS2 cells were infected with SARS-CoV-2 at 0.01 MOI. Lysates were collected at 24 h.p.i. for the detection of viral titre with RT–qPCR (n = 4). Viral RNA-dependent RNA polymerase (RdRp) gene copies in VeroE6-TMPRSS2 infected with WT (a), B.1.1.7 (b), B.1.617.1 (c) and P.3 (d). e–h, Calu3 cells were infected with SARS-CoV-2 at 1 MOI. Lysates were collected at 24 h.p.i. for the detection of viral titre with RT–qPCR (n = 4). Viral RdRp gene copies in cells infected with WT (e), B.1.1.7 (f), B.1.617.1 (g) and P.3 (h). i–l, VeroE6-TMPRSS2 cells were infected with 50–70 PFU SARS-CoV-2 followed by Y180 addition at 2 h.p.i. and fixation at 48 h.p.i. (n = 3). Plaque formation was counted in cells infected with WT (i), B.1.1.7 (j), B.1.617.1 (k) and P.3 (l). Inhibition of plaque formation was normalized to vehicle controls. m–p, VeroE6-TMPRSS2 cells were infected with SARS-CoV-2 at 1 MOI. Cell viability was determined at 48 h.p.i. by measuring luciferase activity in cells (n = 4) infected with WT (m), B.1.1.7 (n), B.1.617.1 (o) and P.3 (p). Luciferase readings were normalized to mock-treated controls without virus infection. Data were obtained from three independent experiments. All data are shown as mean ± s.d. Statistical significance was assessed by one-way ANOVA compared with the vehicle control group; Student’s t-test compared with RDV treatment group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; NS, not significant.

Source data

Full size image

Evaluation of PK properties and safety of Y180

Before evaluating the in vivo PK properties of Y180, we first measured its lipid/water partition coefficient (logP), apparent permeability coefficients (Papp), water solubility and in vitro metabolic stability in human liver microsomes. The measured logP/ Papp/solubility values of Y180 are 2.35/1.52 × 10−6 (cm s−1)/3.06 (μg ml−1) (the corresponding values of PF-07321332 are 0.91/0.58 × 10−6 (cm s−1)/5.72 (μg ml−1); Supplementary Table 5), indicating low water solubility (without any auxiliary materials as solubilizer) and good membrane permeability. Since ritonavir was recently reported to increase the stability of PF-0732133217 in vivo, we examined whether the pharmacokinetic metabolism of Y180 can also be improved by ritonavir co-administration. To this end, we performed the human liver microsome assay. The data suggested that the half-life of Y180 was substantially increased from 13.86 min to >120 min by ritonavir co-administration (Supplementary Table 6).

Next, the in vivo PK properties of Y180 were evaluated. Y180 displayed satisfying PK properties in mice, rats and dogs, with oral bioavailabilities of 92.9%, 31.9% and 85.7%, respectively (Supplementary Tables 1 and 7). Since mice were used in the subsequent in vivo animal experiments, key PK parameters of Y180 in mice are summarized as follows. When administered orally (p.o.) (150 mg kg−1), Y180 showed an area under the curve (AUC) value of 29,201 h×ng ml–1. It displayed a half-life (T1/2) of 1.42 h and an oral bioavailability of 92.9%. On the basis of the EC50/EC90 values from B.1.1.7- or B.1.617.1-infected VeroE6-TMPRSS2 cells (Fig. 3j,k), a single oral (p.o.) dose of 150 mg kg−1 d−1 Y180 maintained the plasma levels at EC50 (20.3 nM) or EC50 (34.4 nM), and EC90 (73.1 nM) or EC90 (157.0 nM) for about 18 or 16 h, and 11 or 12 h, respectively (Extended Data Fig. 9). In vitro cytotoxicity of compound Y180 and (S)-‘6c’ ((S)-epimer, without deuterium) (Supplementary Fig. 7), a potential configuration conversion product of Y180, was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay in various cell lines, including A549, VeroE6, Huh7, BEAS-2B and HUVEC. Both compounds showed low cytotoxicity in these cell lines (Supplementary Table 8). Preliminary in vivo toxicity of Y180 and (S)-‘6c’ was then evaluated in rats. In an acute toxicity experiment, no rats died after p.o. treatment with 600 mg kg−1 of Y180 or (S)-‘6c’ (Supplementary Table 9). In a repeated-dose toxicity study, treatment with p.o. Y180 or (S)-‘6c’ at 400 mg kg−1 twice daily for 7 consecutive days did not result in notable body weight change or pathological changes in heart, liver, lung, kidney and spleen (Extended Data Fig. 6e,f), suggesting no toxicity of Y180 and (S)-‘6c’ in rats at therapeutic concentrations.

In vivo antiviral activity of Y180

Finally, we evaluated the in vivo antiviral efficacy of Y180 against B.1.1.7 and B.1.617.1 variants in an established K18-human angiotensin-converting enzyme 2 (K18-hACE2) transgenic mouse model. Compared with other currently available small-animal models for COVID-19, the K18-hACE2 mouse model supports robust SARS-CoV-2 replication with lethal outcome30,31,32, which makes it a stringent model for the evaluation of drug candidates in ameliorating severe SARS-CoV-2 infection. Y180 was orally administered to the mice starting from 1 h before infection and 6 h post infection (h.p.i.) in the prophylactic group and therapeutic group, respectively (Fig. 4a). In the B.1.1.7-infected K18-hACE2 mice, prophylactic Y180 significantly decreased viral gene copies in the nasal turbinate by 12.0- (P < 0.0001) and 31.5-fold (P < 0.0001) at 2 d.p.i. and 4 d.p.i. compared with the vehicle group, respectively (Fig. 4b). Similarly, even when treatment was postponed to begin at 6 h.p.i., therapeutic administration of Y180 effectively inhibited virus replication, as evidenced by 3.3- (P < 0.0001) and 5.4-fold (P < 0.0001) decrease in nasal turbinate viral gene copies at 2 d.p.i. and 4 d.p.i. (Fig. 4b). In addition, virus replication in the lung was also inhibited by Y180 treatment, with decreases in viral gene copies ranging from 2.7- to 8.8-fold (Fig. 4c). Consistent with the findings in viral gene copies, infectious viral titres in both the lung and the nasal turbinate of Y180-treated mice were substantially diminished compared with the vehicle group (Fig. 4d,e). In agreement with these observations, both prophylactic and therapeutic Y180 significantly ameliorated the expression of representative virus-induced pro-inflammatory cytokines, including interferon gamma (IFNγ) and interferon-inducible protein 10 (IP10) (Fig. 4f). Next, immunohistochemistry assays revealed that abundant expression of viral antigen was identified in both lung and nasal turbinate of the mock-treated mice at 2 d.p.i. (Fig. 4g, black arrows). In contrast, Y180 treatment, even when administered after virus challenge, markedly suppressed viral nucleocapsid protein expression in both lung and nasal turbinate (Fig. 4g). To examine whether Y180 treatment contributed to alleviating disease severity in the transgenic mice, we performed hematoxylin-eosin (H&E) staining to assess virus-induced tissue damage in both lung and nasal turbinate. In the lung of vehicle-treated mice, severe inflammatory infiltrations were frequently observed in the alveoli and alveoli septa, leading to alveolar congestion at 2 d.p.i. (Fig. 4h, vehicle, enlarged image 1, yellow arrowheads). In the small airways, debris of necrotic epithelium was detected in the bronchiole cavities (Fig. 4h, vehicle, enlarged image 2, yellow arrowheads). In the blood vessels, mononuclear cell infiltrations in the endothelium were observed (Fig. 4h, vehicle, enlarged image 3, yellow arrowheads). In sharp contrast, the overall histological architecture of mice receiving either prophylactic or therapeutic Y180 treatment was remarkably preserved (Fig. 4h), except for mild infiltrations of inflammatory cells in the alveolar septa. Consistently, severe tissue damage, including loss of integrity of the nasal epithelium, haemorrhage in the lamina propria and necrotic cells in the nasal cavities, were detected in the nasal turbinate of vehicle-treated mice, while morphology of these tissues in the Y180-treated mice remained largely intact at 2 d.p.i. (Fig. 4h). Since our data demonstrated that Y180 is effective against multiple SARS-CoV-2 variants, we performed additional in vivo experiments with B.1.617.1 to evaluate the infection outcome of K18-hACE2 transgenic mice upon lethal SARS-CoV-2 infection. Our results showed that Y180 treatment significantly improved the survival of mice against B.1.617.1 infection (Fig. 4i; vehicle: 0 of 8 or 0% vs prophylactic treatment: 2 of 9 or 22.2% vs therapeutic treatment: 4 of 9 or 44.4%; P = 0.0402 and P = 0.0086, respectively). Corroborating the survival findings, Y180 also significantly improved the body weight of B.1.617.1-infected mice starting from 4 d.p.i. (Fig. 4j).

Fig. 4: Evaluation of in vivo antiviral activity of prophylactic and therapeutic Y180 in the lethal K18-hACE2 transgenic mouse model for COVID-19.
figure 4

a, Schematic illustration of experiment design. K18-hACE2 transgenic mice were challenged with either 250 PFU B.1.1.7 or 1 × 105 PFU B.1.617.1. Oral administration of either 150 mg kg−1 dose−1 Y180 or vehicle solution was performed twice daily (bis in die; b.i.d.) for 8 d or until death, whichever was earlier. For mice with prophylactic treatment, Y180 treatment was initiated at 1 h before virus challenge, whereas administration of Y180 was postponed until 6 h.p.i. for mice receiving therapeutic treatment. b,c, Mice infected with B.1.1.7 were killed at 2 d.p.i. and 4 d.p.i. for detection of viral RdRp gene copies in nasal turbinate (b) and lung (c) by RT–qPCR (n = 5 at 2 d.p.i.; n = 6 at 4 d.p.i.). d,e, Mice infected with B.1.1.7 were killed at 2 d.p.i. and 4 d.p.i. for detection of infectious viral titre in nasal turbinate (d) and lung (e) by plaque assay (n = 5 at 2 d.p.i.; n = 6 at 4 d.p.i.). f, Gene expression of IFNγ and IP10 in the lung of mice infected with B.1.1.7 at 4 d.p.i. were determined with RT–qPCR (n = 5 at 2 d.p.i.; n = 6 at 4 d.p.i.). g, Immunohistochemistry staining visualizing the nucleocapsid protein (brown, indicated by black arrows) in the lung and nasal turbinate of mice infected with B.1.1.7 at 2 d.p.i. Scale bar, 100 µm. h, Histology analysis visualizing the virus-induced pathology in the lung and nasal turbinate of mice infected with B.1.1.7 at 2 d.p.i. Representative images of alveoli, bronchioles and blood vessels of the lung are indicated by yellow dotted circles with numbers, and enlarged in images 1, 2 and 3, respectively. Yellow arrowheads indicate inflammatory infiltrations in the alveoli, dead cell debris in the bronchioles and mononuclear cell infiltrations in the blood vessels in the enlarged images 1–3. Black arrowheads, haemorrhage; black arrows, epithelium detachment; #, dead cell debris. Scale bar, 100 µm for enlarged images 1–3, 200 µm for nasal turbinate image. Three mice were sampled in each group and 4–6 sections from each animal were used for histology analysis. i,j, Mice challenged with 1 × 105 PFU B.1.617.1 were treated with either 150 mg kg−1 dose−1 Y180 or vehicle solution twice daily (n = 8–9). Survival (i) and body weight (j) of mice were monitored daily until 14 d.p.i. Data were obtained from two independent experiments. All data are shown as mean ± s.d. Statistical differences were determined by two-way ANOVA in b–e and j, one-way ANOVA in f or log-rank (Mantel-Cox) test in i. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Source data

Full size image

Moreover, we performed head-to-head comparison to evaluate the antiviral efficacy of Y180/ritonavir and PAXLOVID (PF-7321332/ritonavir), the FDA-approved orally available Mpro inhibitor, against B.1.1.529 infection. Our data demonstrated that therapeutic treatment with Y180/ritonavir potently reduced viral RdRp gene copies by 14.9-fold (P = 0.0429) and 27.7-fold (P = 0.0056) in nasal turbinate and lung, respectively, of the infected mice at 2 dpi. (Extended Data Fig. 10a,b). In comparison, PF-07321332/ritonavir reduced viral gene copies by 0.8-fold (P = NS) and 14.1-fold (P = 0.0070) in nasal turbinate and lung, respectively, under the same therapeutic conditions (Extended Data Fig. 10a,b). Similarly, Y180/ritonavir suppressed the production of infectious viral progenies in lung more efficiently than PF-07321332/ritonavir (Extended Data Fig. 10c). Therefore, our data suggested that the in vivo antiviral efficacy of Y180/ritonavir combined treatment is more potent than that of PF-07321332/ritonavir in the K18-hACE2 mouse model. Collectively, our results demonstrate that Y180 is highly effective against WT SARS-CoV-2 and emerging SARS-CoV-2 variants including Omicron both in vitro and in vivo by reducing virus replication, which results in an attenuated disease severity and improves the host survival outcomes upon SARS-CoV-2 infection.

Discussion

The emergence of SARS-CoV-2 variants heavily increased the challenges for curbing the COVID-19 pandemic. For example, the B.1.1.7, B.1.617.1, P.3 and the recently emerged B.1.1.529 variants2 were shown to have enhanced transmissibility and/or ability to escape from currently available vaccines and therapeutic monoclonal antibodies1,2,3, which target the spike protein. Effective antivirals targeting the Mpro, which is highly conserved among different SARS-CoV-2 variants, is an effective strategy to deal with this challenging task.

In this study, to discover Mpro inhibitors with potential for rapid translation into clinical application, we designed a series of diamide derivatives, which can be easily synthesized using Ugi-4CR. In our designed compounds, we introduced an acetyl group at the terminal acyl of the amide, which allowed the formation of an α-ketoamide warhead in the S1′ pocket. Further, to reduce configuration conversion of the most active epimer ((R)-epimer) to its less active epimer (S-configuration), deuterium was used to replace the exchangeable hydrogen attached to the chiral carbon linking the two amides.

The selected compound Y180 showed potent SARS-CoV-2 Mpro inhibitory activity with an IC50 of 8.1 nM. Importantly, the introduction of deuterium at the chiral carbon substantially reduced configuration conversion of Y180 from the active (R)-epimer to the less active (S)-epimer. Y180 displayed good PK properties in mice, rats and dogs, with oral bioavailabilities of 92.9%, 31.9% and 85.7%, respectively. It displayed outstanding in vitro antiviral potency against WT SARS-CoV-2 and the B.1.1.7, B.1.617.1, P.3 and B.1.1.529 variants. Importantly, Y180 demonstrated remarkable in vivo antiviral potency and improved the disease outcome against B.1.1.7, B.1.617.1 and B.1.1.529 infection both prophylactically and therapeutically as evidenced by the significantly decreased viral burden, attenuated tissue damage and improved survival among the infected animals. Together, our data provided strong evidence for the therapeutic potential of Y180 against SARS-CoV-2 infection in a lethal infection model.

Compared with other reported Mpro inhibitors9,10,11,12,13,14,15,16,17,18,19,20,21, Y180: (1) is the most potent non-peptidomimetic Mpro inhibitors reported thus far, inhibiting SARS-CoV-2 Mpro activity at single-digit nanomolar ranges, (2) demonstrates the best EC50 values against WT SARS-CoV-2 and its variants at low double-digit nanomolar ranges in vitro and (3) is the only reported orally administered Mpro inhibitor that can rescue the survival of animals upon infection with SARS-CoV-2 variants in a lethal COVID-19 mouse model. Importantly, in a head-to-head in vivo antiviral assay, therapeutic treatment with Y180/ritonavir was more potent than PF-07321332/ritonavir against SARS-CoV-2 Omicron in K18-hACE2 transgenic mice. Meanwhile, the simplicity of the chemical synthesis of Y180 is also unrivalled compared with that of PF-07321332, which is a peptidomimetic Mpro inhibitor. As a covalent inhibitor, the safety of Y180 may be of concern. However, Y180 has obvious safety advantages because: (1) it contains a mild warhead, α-ketoamide, which reduces toxicity (a number of compounds containing α-ketoamide warheads have been approved for clinical use, such as antivirals against hepatitis C virus (HCV) infection, Boceprevir, Telaprevir and Narlaprevir); (2) it showed excellent selectivity for Mpro compared with other common mammalian proteases with similar structure (see Supplementary Table 3); and (3) in our preliminary preclinical safety evaluation, Y180 did not show obvious toxicity both in vitro and in vivo. Of course, the safety of Y180 in humans remains to be determined in clinical trials. Collectively, we identified an orally available Mpro inhibitor that demonstrated potent in vitro and in vivo antiviral activity against WT SARS-CoV-2 and emerging variants. Our work advances the development of highly specific antiviral therapeutics against COVID-19.

Methods

Cell lines

Huh7 (Cell Bank, Chinese Academy of Sciences, Shanghai, SCSP-526), BEAS-2B (Cell Bank, Chinese Academy of Sciences, Shanghai, GNHu27), A549 (ATCC, CRM-CCL-185), HUVEC (CCTCC, Wuhan, GDC0635), VeroE6 (ATCC, CRL-1586) and VeroE6-TMPRSS2 (JCRB 1819) cells were maintained in Dulbecco’s modified Eagle medium (DMEM, Gibco) supplemented with 10% heat-inactivated foetal bovine serum (FBS), 100 U ml−1 penicillin and 100 μg ml−1 streptomycin. Calu3 cells (ATCC, HTB-55) were maintained in DMEM/F12 supplemented with 10% heat-inactivated FBS, 100 U ml−1 penicillin and 100 μg ml−1 streptomycin.

Viruses

SARS-CoV-2 HKU-001a (GenBank accession number MT230904), B.1.1.7 (GISAID: EPI_ISL_1273444), B.1.617.1 (GISAID: EPI_ISL_2423557), P.3 (GISAID: EPI_ISL_1660475) and B.1.529 (GenBank accession number OM212472) were isolated from laboratory-confirmed COVID-19 patients in Hong Kong33. WT SARS-CoV-2 and its variants were cultured in VeroE6-TMPRSS2 and titrated by plaque assays. All infectious experiments followed the approved standard operating procedures of the Biosafety Level 3 facility at the Department of Microbiology, HKU.

Production of authentic SARS-CoV-2 and Omicron variant’s main protease

Using the wild-type SARS-CoV-2 (GenBank: MN908947.3) pET28b-main protease (Mpro) plasmid13 as template, SARS-CoV-2 Omicron variant (B.1.1.529, GISAID accession ID: EPI_ISL_6590782.2) Mpro (Pro132His mutation) was obtained by site-directed mutagenesis PCR using the two primers: 5’-GCAATGCGTCATAATTTTACCATTAAGGGTAGTTTTCTG-3’ and 5’-GGTAAAATTATGACGCATTGCACACTGATAAACGCC-3’ (mutated codons shown in italic). The PCR products were then digested by DpnI (TakaRa) and directly transformed to Escherichia coli DH5α competent cells (Novagen). The positive clones were incubated in Luria broth medium overnight and the corresponding plasmids were extracted using the Omega E.Z.N.A. Plasmid Mini Kit I (Omega Bio-Tek). All DNA plasmids were sequenced and the correctness of the Pro132His mutation was verified. Subsequently, the WT SARS-CoV-2 or omicron variant pET28b-Mpro plasmid was expressed in E. coli BL21(DE3) cells (Novagen). The cell pellets were resuspended in buffer A (20 mM Tris-HCl pH 8.0, 10 mM imidazole, 500 mM NaCl and 5% glycerol), lysed via ultrasonication on ice and centrifuged at 38,759 × g at 4 °C. The supernatants were loaded onto HisTrap FF column (GE Healthcare) and eluted with buffer B (20 mM Tris-HCl pH 8.0, 500 mM imidazole, 500 mM NaCl and 5% glycerol). The target protein (wild-type or Omicron variant Mpro) was processed by PreScission protease at 4 °C overnight to remove the His-tag. The mixture was loaded onto GSTrap FF (GE Healthcare) and HisTrap FF columns again to remove PreScission and uncleaved Mpro. The target protein was further purified by gel filtration (Superdex 75 Increase 10/300 GL, GE Healthcare) in buffer C (20 mM Tris-HCl pH 8.0, 150 mM NaCl). The quality of the purified wild-type or Omicron variant Mpro was checked by SDS–PAGE.

Crystallization of SARS-CoV-2 Mpro in complex with (R)-1a or Y180

The SARS-CoV-2 Mpro was concentrated to ~5 mg ml−1 and incubated with each compound at a molar ratio of 1:10 at 4 °C overnight. Subsequently, crystallization was performed at 291 K using the sitting-drop vapour-diffusion method by mixing 1 μl protein solution plus 1 μl reservoir to equilibrate against 70 μl reservoir solution. The commercial screen kits—Index, Crystal Screen 1/2 (Hampton Research) and JBScreen Basic 1-4 (Jena Bioscience)—were used. The crystals of Mpro-(R)-‘1a’ were obtained within 5 d under condition No. 80 of Index: 0.2 M ammonium acetate, 0.1 M HEPES pH 7.5, 25% w/v PEG3350. The crystals of Mpro-Y180 were obtained within 3 d under condition No. 22 of Crystal Screen 1/2: 0.2 M sodium acetate trihydrate, 0.1 M TRIS hydrochloride pH 8.5, 30% w/v PEG4000. All crystals were flash-cooled in liquid nitrogen with cryo-protectant solution (50% w/v PEG3350).

Data collection, phase determination and refinement

The diffraction datasets of both Mpro-(R)-‘1a’ and Mpro-Y180 were collected at an X-ray wavelength 0.97852 Å at Shanghai Synchrotron Radiation Facility (SSRF) beamline BL19U1 (Shanghai, China). The dataset of Mpro-(R)-‘1a’ was processed at a resolution of 2.1 Å and in space group ‘C2’ by ‘XDS’34, and scaled with ‘Aimless’35 in CCP4. The dataset of Mpro-Y180 was also processed at a resolution 2.1 Å, but with space group ‘P212121’. The initial phases of the SARS-CoV-2 Mpro was determined by molecular replacement method with the programme PHASER36, using the Mpro structure from PDB entry 7D3I13 as the search model. Compound (R)-‘1a’ or Y180 was then built into the model using the programme ‘Coot’37. Refinement of the complex Mpro-(R)-‘1a’ was performed using ‘BUSTER’38, while refinement of Mpro-Y180 was processed using ‘Phenix REFINE’39. In the final structure of Mpro-(R)-‘1a’ or Mpro-Y180, >97% of residues are in the preferred regions of the ‘Ramachandran plot’ statistics. The statistics of the diffraction dataset and the final refinement are presented in Supplementary Table 10.

Enzyme kinetics, correction of inner-filter effect, IC50, K i and k inac measurements

Enzyme kinetics assay was performed in 20 mM HEPES buffer, 120 mM NaCl, 0.4 mM EDTA, 4 mM dithiothreitol and 20% glycerol at pH 6.5. The reaction was initiated by adding 25 μl of the MCA-AVLQSGFR-Lys(DNP)-Lys-NH2 fluorescent substrate (1.6~200 μM) into 25 μl of the WT SARS-CoV-2 or Omicron variant Mpro (final concentration: 100 nM). The fluorescence signal of unquenched MCA was monitored at emission/excitation wavelengths of 405/320 nm using a CLARIOstar microplate reader (BMG Labtech). The initial velocities were calculated from the linear section of the reaction curves and the amounts of the cleaved substrate were obtained by fitting the fluorescence to a calibration curve of free MCA (0.01~20 μM).

The inner-filter effect of the fluorescent substrate was corrected. The measured fluorescence values of background and substrate with different concentrations were defined as f(0) and f(S). Free MCA (final concentration: 0.5 μM) was then added into each well. The second fluorescence measurements were taken as f(0+MCA) and f(S + MCA). The inner-filter effect of the substrate was corrected according to the equation

$${{{\mathrm{Corr}}}}\% = 100\% \times \frac{{{{{f}}}\left( {{{{\mathrm{S}}}} + {{{\mathrm{MCA}}}}} \right) - {{{f}}}({{{\mathrm{S}}}})}}{{{{{f}}}\left( {0 + {{{\mathrm{MCA}}}}} \right) - {{{f}}}(0)}}.$$

The corrected initial velocity of the reaction in the enzyme kinetics assay was calculated as

$${{{V}}} = {{{V}}}_{{{{\mathrm{initial}}}}\,{{{\mathrm{velocity}}}}}/({{{\mathrm{Corr}}}}\% ).$$

For the determination of IC50 values, 22.5 μl of the recombinant authentic WT SARS-CoV-2 or Omicron variant Mpro (final concentration: 100 nM) was pre-incubated with 2.5 μl of various concentrations of each compound for 10 min. Then the reaction was initiated by adding 25 μl of the fluorescent substrate (final concentration: 20 μM) and monitored at 405 nm with excitation at 320 nm in the kinetic mode using the CLARIOstar microplate reader (BMG Labtech). The initial velocities were calculated from the linear section of the reaction curves. The IC50 values were calculated using a dose-response model in GraphPad Prism 8.0 software.

In the inhibition constant (Ki) assay, 22.5 μl of the Mpro (final concentration: 100 nM) was mixed with 2.5 μl of various concentrations (4.9~312.5 nM) of Y180 or PF-07321332. The reactions were immediately initiated by adding 25 μl of the fluorescent substrate (final concentration: 30 μM). The fluorescence signal was monitored using CLARIOstar microplate reader (BMG Labtech). The initial velocities were calculated from the linear section of the reaction curves. The Ki values for compounds Y180 and PF-07321332 were determined with the Morrison equation in GraphPad Prism 8.0 software.

In the inactivation rate constant (kinac) assay, 22.5 μl of the Mpro (final concentration: 100 nM) was mixed with 2.5 μl of different concentrations (3.3~25.0 nM) of Y180 or PF-07321332. The reactions were monitored by addition 25 μl of the fluorescent substrate (final concentration: 20 μM) for 15 min. The observed first-order rate constant kobs was obtained by fitting the progress curves of different concentrations of inhibitor to the one-phase association equation in GraphPad Prism 8.0 software. The constant kobs/[I] ([I] is the concentration of the inhibitor) was obtained by linear regression analysis. The mimic kinac value was calculated using the equation40

$$k_{\mathrm{inac}}/K_\mathrm{i} = \left( {1 + \frac{{\left[ S \right]}}{{K_\mathrm{m}}}} \right) \times k_{\mathrm{obs}}/[I].$$

Differential scanning fluorimetry assay

Differential scanning fluorimetry assays were performed using BioRad CFX96 real‐time PCR detection system. Mpro was added to a final concentration of 2 μM and incubated with 40 μM test compounds for 30 min. Orange dye (5×SYPRO, Sigma) was added and the thermal denaturation was measured with a temperature gradient of 1.5 °C min−1 from 20 °C to 95 °C. The melt temperature (Tm) was calculated by using a Boltzmann model in GraphPad Prism 8.0 software. The thermal shift (ΔTm) was calculated using the equation ΔTm = Tm(compound) − Tm(DMSO). All experiments were performed in triplicate, and the values are presented as mean ± s.d.

Biochemical activity assay against mammalian proteases

The biochemical activity against mammalian proteases was determined by Human Proteases assay kit (Biovision) and performed following the manufacturer’s instructions. All experiments were performed in triplicate.

Cytotoxicity assay

For VeroE6, HUVEC, A549, BEAS-2B and Huh7 cells, cells were seeded in 96-well plates and grown overnight. Various concentrations (0.41~100 μM) of Y180 or (S)-6c were then added to each well. After incubation for 48 h, cell viability was evaluated using MTT (Sigma) according to the manufacturer’s protocol. All experiments were performed in triplicate.

For VeroE6-TMPRSS2 and Calu3 cells, cells were treated with either Y180 in 5-fold serial dilution, ranging from 0.0013 μM to 500 μM, or with vehicle (dimethyl sulfoxide; DMSO) for 72 h. Cell viability was quantified by CellTiter-Glo luminescent cell viability assay kit (G7572, Promega) following the manufacturer’s instructions, with a multilabel plate reader Victor X3 (Perkin Elmer)41.

PK properties and chiral inversion in vivo

All procedures related to animal handling, care and treatment in PK studies were performed according to approved guidelines. The PK studies were approved by the Ethics Committee of Shanghai Medicilon Inc., ZLA (Beijing) Pharmaceutical Technology Co. Ltd. and Sichuan Greentech Biotechnology Co. Ltd. All animals used in this study were chosen randomly.

In PK properties and chiral inversion studies, compounds, doses and administration schemes are shown in Supplementary Tables 1 and 7. Briefly, male Sprague Dawley (SD) rats, male ICR (Institute of Cancer Research) mice or beagle dogs (n = 3 per group, aged 1~2 yr and weighing 9~12 kg) were treated with the test compounds at the indicated doses by oral gavage or intravenously. In the case of the male SD rats, the compounds were dissolved in 10% DMSO, 40% PEG300, 5% Tween-80 and 45% saline. As for the ICR mice and beagle dogs, the compounds were dissolved in 10% DMSO, 42% PEG300, 3% polyethylene-polypropylene glycol and 45% saline.

Blood samples were collected at 0.083, 0.25, 0.5, 1, 2, 4, 6, 8 and 24 h after administration. Serum samples were obtained by centrifugation and the concentrations of compounds in the serum were analysed by liquid chromatography tandem mass spectrometry (LC–MS/MS) (Agilent 1200 series HPLC (Agilent) coupled to an API4000 mass spectrometer (Applied Biosystems/MDS Sciex)). The PK parameters were calculated using Phoenix WinNonlin7.0.

Serum samples from rats at 0.25, 1 and 6 h were selected to evaluate chiral inversion of the compounds in vivo by LC–MS/MS. The LC–MS/MS separation of R- and S-epimers of compounds was performed on a Lux Cellulose 3 chiral column (4.6 × 250 mm, 5.0 μm particle size) and the percentage of S-epimers was used to estimate the degree of chiral conversion.

Metabolic stability in human liver microsomes (HLM)

Test compound (Y180, Y180 with Ritonavir, PF-07321332 or PF-07321332 with Ritonavir) incubations were conducted with 0.5 mg ml−1 human liver microsomes at 37 °C open to ambient air at various timepoints (0, 5, 15, 35, 45 and 60 min). Samples were collected and analysed by LC–MS/MS. The substrate depletion half-life (T1/2) and intrinsic clearance (CLint) were calculated using: T1/2 = 0.693/K (K is the rate constant from a plot of ln[concentration] vs incubation time) and CLint = (0.693/T1/2) × {1/[microsomal protein concentration (0.5 mg ml−1)]} × scaling factors (1,254.2).

In vivo toxicity study

All procedures related to animal handling, care and treatment in in vivo toxicity studies were performed according to approved guidelines. This study was approved by the Ethics and Animal Welfare Committee of West China Hospital, Sichuan University (20211063A). All animals used in this study were chosen randomly.

SD rats (aged 7–11 weeks) consisted of half male (weight: 200–230 g) and half female (weight: 190–220 g) were used to assess toxicity of Y180 and (S)-‘6c’ in vivo. Compounds were dissolved in 10% DMSO, 42% PEG300, 3% polyethylene-polypropylene glycol and 45% saline.

Compounds, doses and administration schemes are shown in Supplementary Table 9. Briefly, in the acute toxicity study, rats were administrated with the compounds at the indicated doses by oral gavage or via tail veins. The rats were clinically observed for 7–14 d after administration. In the repeated-dose toxicity study, rats were administered with compounds by gavage or via tail veins twice daily for 1 week. All animals were monitored for body weight, food intake and behaviours at least twice daily during the administration to detect signs of toxicity.

At the end of the experiment, samples of heart, liver, spleen, lung and kidneys were collected and processed for H&E staining for histological analysis.

Evaluation of in vitro antiviral activity

Calu3 and/or VeroE6-TMPRSS2 cells were infected with wild-type SARS-CoV-2 or variants (B.1.1.7, B.1.617.1, P.3 or B.1.1.529) at 0.01 multiplicity of infection (MOI) and 1 MOI, respectively, and treated with either Y180 or PF-07321332 in 5-fold serial dilution ranging from 0.0013 μM to 20 μM, or with vehicle or remdesivir (0.8, 4 or 20 μM). Cell lysates were collected at 24 h.p.i with RLT buffer for quantitative PCR with reverse transcription (RT–qPCR) analysis. The primer and probe sequences are provided in Supplementary Table 11.

Cell viability assay

VeroE6-TMPRSS2 cells were infected with wild-type SARS-CoV-2 or variants (B.1.1.7, B.1.617.1, P.3) at 1 MOI and treated with either Y180 in 5-fold serial dilution ranging from 0.0013 μM to 20 μM, or with vehicle or remdesivir (0.8, 4 or 20 μM). Cell viability was quantified by CellTiter-Glo luminescent cell viability assay kit (G7572, Promega) following the manufacturer’s instructions, with a multilabel plate reader Victor X3 (Perkin Elmer) at 48 h.p.i41.

Plaque reduction assay

VeroE6-TMPRSS2 cells were infected by wild-type SARS-CoV-2 or variants (B.1.1.7, B.1.617.1, P.3) with 50–70 plaque forming units (PFU) per well in a 12-well format. After inoculation, the cells were washed with PBS and covered with 2% agarose/PBS mixed with 2× DMEM/2% FBS at a 1:1 ratio, in the presence of either Y180 (5-fold dilution ranging from 0.0013 μM to 20 μM) or with vehicle only. Cells were fixed at 2 d.p.i. Fixed samples were stained with 0.5% crystal violet in 25% ethanol/distilled water for 10 min for plaque visualization.

SARS-CoV-2 infection in K18-hACE2 mice

K18-hACE2 transgenic mice aged 6–8 weeks were obtained from the Jackson Laboratory. The use of K18-hACE2 transgenic mice has received ethical approval from the Committee on the Use of Live Animals in Teaching and Research of The University of Hong Kong under Animal Ethics Committee at the University of Hong Kong (5779-21). On the day of infection, the hACE2 mice were intranasally inoculated with either 250 PFU B.1.1.7, 1 × 105 PFU B.1.617.1 or 2,000 PFU B.1.1.529, pre-diluted in 20 μl MEM. Mice were orally administered with 150 mg kg−1 dose−1 Y180 diluted in 180 μl 3% F188/45% PEG300/normal saline (for the treatment group) or PF-07321332 diluted in 180 μl 0.5% methylcellulose/2% Tween-80 or vehicle solution only (for the control group) twice daily for 8 d, or until sample collection or animal death. For the prophylactic treatment group, Y180 was administered at 1 h before virus challenge. For the therapeutic treatment group, Y180 administration was delayed until 6 h.p.i., while Y180 and ritonavir (10 mg kg−1 dose−1, twice daily) or PF and ritonavir (10 mg kg−1 dose−1, twice daily) administration was delayed until 16 h.p.i. Survival of the mice was monitored on a daily basis. Mice were killed at the designated timepoints and organ tissues were sampled for virological and histopathological analyses29.

RNA extraction and RT-qPCR

Cell lysates and tissue samples were lysed using RLT buffer (Qiagen) and extracted with the RNeasy mini kit (74106, Qiagen). After RNA extraction, RT–qPCR was performed using the Transcriptor First Strand cDNA synthesis kit (04897030001, Roche), QuantiNova SYBR Green RT–PCR kit (208154, Qiagen) or the QuantiNova Probe RT–PCR kit (208354, Qiagen) with the LightCycler 480 real-time PCR System (Roche)42. The primer and probe sequences are provided in Supplementary Table 11.

Plaque assays

VeroE6-TMPRSS2 cells were seeded in 12-well plates 1 d before infection. The collected supernatant samples were serially diluted and inoculated to the cells for 1 h at 37 °C. After inoculation, the cells were washed with PBS 3 times and covered with 2% agarose/PBS mixed with 2× DMEM/2% FBS at a 1:1 ratio. The cells were fixed after incubation at 37 °C for 48 h. Fixed samples were stained with 0.5% crystal violet in 25% ethanol/distilled water for 10 min for plaque visualization43.

Histology and immunohistochemistry staining

Histology and immunohistochemistry staining were performed using established protocols44. In brief, formic acid-soaked mice nasal turbinate and lung tissues were fixed overnight in 10% formalin. The fixed samples were then embedded in paraffin with a TP1020 Leica semi-enclosed benchtop tissue processor and sectioned at 5 μm. Tissue sections were fished and dried to fix on Superfrost Plus slides (Thermo Fisher) at 37 °C overnight. Afterwards, the sections were dewaxed and dehydrated in serially diluted xylene, ethanol and double-distilled water in sequence, co-heating together with antigen unmasking solution (H-3300, Vector Laboratories) at 85 °C for 90 s for antigen exposure, followed by blocking with 0.3% hydrogen peroxide for 30 min, and 1% BSA for 30 min. The in-house rabbit anti-SARS-CoV-2-N immune serum (1:5,000) and goat anti-rabbit IgG antibody (BA-1000-1.5, Vector Laboratories) (1:1,000) were applied as primary and secondary antibodies, respectively. The signal was developed with the 3,3’-diaminobenzidine (DAB) substrate kit (SK-4100, Vector Laboratories). Cell nuclei were labelled with Gill’s haematoxylin. The slides were mounted with antifade mounting medium with DAPI (H-1200, Vector Laboratories). For H&E staining, infected tissue sections were stained with Gill’s hematoxylin and eosin Y (Thermo Fisher). Images were taken with the Olympus BX53 light microscope (Olympus Life Science). The histology findings made in our current study represent unbiased description for the pathological damage in the tissues examined and the results were validated by an experienced pathologist in a blinded manner.

Statistical analysis

Data on figures represent mean ± s.d. All data were analysed with GraphPad Prism 7.0 and 8.0 software. Statistical comparison between different groups was performed by one-way analysis of variance (ANOVA) or log-rank (Mantel-Cox) test. All t-test performed in this study is two-tailed. Differences were considered statistically significant when P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.