An intrinsic loop-mediated structural stability modulating inhibitor potency in the SADS-CoV and SARS-CoV-2 main proteases
Rui Zeng, Shizhan Cui, Xiaoyan Xia, Chong Huang, Jingxuan Sun, Xinyue Deng, Qiaoling Gui, Huahao Fan, Xiaojuan Liu, Yang Yu, Shengyong Yang, Jian Lei
TL;DR
Researchers found structural differences in a key viral enzyme from SADS-CoV compared to SARS-CoV-2, which affect how well inhibitors work, and tested a new inhibitor that reduces virus replication in pig cells.
Contribution
The study reveals a novel inhibition mechanism and structural dynamics in SADS-CoV Mpro that influence inhibitor potency and guide broad-spectrum antiviral design.
Findings
SADS-CoV Mpro has a single-helix region instead of a coiled coil seen in SARS-CoV-2 Mpro, affecting inhibitor binding.
Inhibitor 27h effectively suppresses SADS-CoV replication in cell-based assays and porcine intestinal organoids.
The absence of residue 51 in α-CoV Mpros drives structural differences and modulates inhibitor efficiency.
Abstract
Swine acute diarrhea syndrome coronavirus (SADS-CoV) poses a significant zoonotic risk. The absence of the structure of SADS-CoV main protease (Mpro) severely impedes the development of effective antiviral therapeutics. Here, we present the high-resolution structures of SADS-CoV Mpro and its complexes with inhibitors 27h and SY110, respectively. These two compounds inhibit SADS-CoV Mpro through a novel inhibition mechanism. Residues 40–53 of SADS-CoV Mpro adopt a single-helix conformation, in contrast to a coiled coil formed by two consecutive alpha-helices observed in SARS-CoV-2 Mpro. These structural differences contribute to the varying inhibitor potency between Alphacoronavirus (α-CoV) and Betacoronavirus (β-CoV) Mpros. We subsequently demonstrate that the absence of residue ‘51’ in α-CoV Mpros plays a key role in these conformational changes. Furthermore, 27h was proved to…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig 1
Fig 2
Fig 3
Fig 4
Fig 5
Fig 6- —http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China
- —http://dx.doi.org/10.13039/501100018542Natural Science Foundation of Sichuan Province
- —the grant of 1.3.5 project, West China Hospital, Sichuan University
- —http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China
- —the Major Project of Guangzhou National Laboratory
- —the grant from Frontiers Medical Center, Tianfu Jincheng Laboratory Foundation
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAnimal Virus Infections Studies · SARS-CoV-2 and COVID-19 Research · Viral gastroenteritis research and epidemiology
Introduction
Coronaviruses (CoVs), belonging to the order Nidovirales, pose ongoing threats to global health and economies due to their broad host range, high transmissibility, and pathogenicity [1,2]. CoVs are divided into four genera (Alphacoronavirus (α-CoV), Betacoronavirus (β-CoV), *Gammacoronavirus (γ-CoV), and Deltacoronavirus (δ-CoV)), with α- and β-CoVs presenting greater risks to human health by causing respiratory, gastrointestinal, and neurological diseases [3–5]. To date, seven human CoVs (HCoVs) have been identified, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), SARS-CoV, Middle East respiratory syndrome coronavirus (MERS-CoV), HCoV-OC43, and HCoV-HKU1, which belong to β-CoV, and HCoV-229E and HCoV-NL63, which are classified into α-*CoV. SARS-CoV, MERS-CoV, and SARS-CoV-2 are frequently associated with severe clinical outcomes and fatalities [6,7]. While HCoV-OC43, HCoV-HKU1, HCoV-229E, and HCoV-NL63 typically cause mild respiratory illness; however, these four HCoVs can also lead to severe and life-threatening lower respiratory tract infections, particularly in the elderly and infants [8].
The emergence of HCoVs is commonly linked to zoonotic transmission from animal reservoirs. SARS-CoV is believed to have crossed into humans via civets, which acted as intermediate hosts between bats and humans [9], while MERS-CoV was transmitted to humans through dromedary camels [10]. Recently, the risk of cross-species transmission of novel animal-origin coronaviruses has become a growing concern [11,12]. For instance, a canine coronavirus (CCoV)-HUPN-2018 (an α-CoV) was identified in hospitalized pneumonia patients in Malaysia [11]. In addition, a porcine deltacoronavirus (PDCoV, a δ-CoV) has demonstrated the ability to infect children [12]. Particularly, current evidence increasingly indicates that a swine acute diarrhea syndrome coronavirus (SADS-CoV, also known as swine enteric alphacoronavirus (SeACoV) or porcine enteric alphacoronavirus (PEAV)) can replicate efficiently in a variety of human and other mammalian cell lines [13–15], highlighting its significant zoonotic potential. SADS-CoV, first isolated in 2017 in Guangdong province of China, caused severe acute diarrhea and vomiting in newborn piglets, with a mortality rate of 90–100% [16,17]. Regional epidemics were subsequently reported in Fujian, Guangxi, Jiangxi, and Henan provinces of China between 2018 and 2023, leading to sustained substantial economic losses in the swine industry [18–22]. Recently, SADS-CoV and SADS-CoV-related bat CoVs have been identified in Vietnam [23,24], suggesting the potential for a SADS-CoV outbreak outside China. SADS-CoV shares 95% sequence identity with Rhinolophus bat coronavirus HKU2. Although phylogenetic analysis classifies SADS-CoV as an α*-CoV [17], its spike protein displays molecular features characteristic of β-*CoVs [25,26], implying possible cross-species transmission acquired through α- and β-CoVs genomic recombination [26]. Currently, all CoVs responsible for large-scale human infections have originated from the α- and β-CoV genera. Remarkably, the SADS-CoV genome harbors genetic features from both [19,25,26] and presents efficient infection in multiple human cell lines [13–15]. These characteristics confer a markedly increased risk of its cross-species transmission, highlighting the urgent need for antiviral development targeting SADS-CoV. However, no specific vaccines or antiviral therapies are available for SADS-CoV thus far.
The CoV main protease (M^pro^, also known as 3C-like protease) cleaves the viral polyproteins (pp1a / pp1ab) into functional non-structural proteins essential for assembling the replication–transcription complex. Owing to its indispensable role in the viral life cycle and the absence of closely related homologs in humans, M^pro^ represents an attractive target for anti-CoV drug development [27–30]. Moreover, the high conservation of M^pro^ across diverse CoVs highlights its potential as a broad-spectrum therapeutic target, particularly against emerging and re-emerging zoonotic CoVs. Inhibitors targeting HCoV M^pro^s, particularly SARS-CoV-2 M^pro^, have been well developed [27–29,31–33]. However, the structure of SADS-CoV M^pro^ and its interaction modes with inhibitors remain poorly understood, posing a major obstacle to the development of effective antiviral therapies. Meanwhile, existing inhibitors targeting SARS-CoV-2 M^pro^ exhibit reduced inhibitory activity against α-CoV M^pro^s [27,34–36], thus necessitating re-evaluation of their anti-SADS-CoV efficacy. We and our colleagues previously reported two inhibitors, 27h (initial name: FT544) [33] and SY110 [31], targeting SARS-CoV-2 M^pro^. The former exhibits favorable pharmacokinetic properties, while the latter has received clinical trial approval from the Chinese National Medical Products Administration (NMPA, No.2023LP01491).
Here, we present the first crystal structure of SADS-CoV M^pro^ and identify two inhibitors—27h and SY110—through screening of in-house and commercial libraries. We characterize their inhibition mechanisms and reveal that conformational dynamics in residues 40–53/54 (in particular, residues 45–50/51) of α- and β-CoV M^pro^s modulate inhibitor potency. We further elucidate the underlying regulatory mechanism. The anti-SADS-CoV efficacy of 27h was validated in both cell-based models and porcine intestinal organoids—representing, to the best of our knowledge, the first use of an organoid system to evaluate anti-SADS-CoV M^pro^ agents. Notably, 27h shows promise as a broad-spectrum M^pro^ inhibitor. Overall, our study links M^pro^ dynamics to inhibitor efficiency and offers valuable insights for pan-CoV inhibitor development.
Results
Structure and inhibitors of the SADS-CoV Mpro
To obtain the crystal structure of the SADS-CoV M^pro^, we initially performed extensive crystallization screening using wild-type (WT) protein. After testing over thousands of conditions across temperature gradients (e.g., 4°C and 18°C) and different protein concentrations (~ 5 – 15 mg/mL), no visible crystals were obtained for diffraction study. Protein-surface-modification with rational mutagenesis could promote preparation of suitable X-ray quality crystals [37]. Thus, we employed a series of mutations targeting surface-exposed cysteine residues to reduce the non-specific disulfide bond formation, as well as entropy-reducing mutations (e.g., Lys, Arg to Ala or Val) to minimize conformational flexibility. Among dozens of engineered variants, only the M^pro^-Lys35Val mutant successfully yielded the crystals. The protein quality of this mutant is shown in S1A Fig. Subsequently, the structure of SADS-CoV M^pro^ was determined at ~ 2.1 Å resolution. The overall structure of SADS-CoV M^pro^ belongs to the space group C2, with two protomers (named as A and B) in the asymmetric unit, forming the typical homodimer of M^pro^ (Fig 1A). Each protomer comprises three domains: Domain I (residues 1–97) consists of four α-helices and seven antiparallel β-strands; Domain II (residues 98–183) is formed by seven antiparallel β-strands; and Domain III (residues 200–302), connected to Domain II through an extended loop, folds into a helical bundle of five α-helices. The catalytic dyad (Cys144-His41) is located within the substrate-binding cleft between Domains I and II (Fig 1A and 1B). Analysis of the solved structure indicates that Lys35 is not involved in substrate or inhibitor binding to M^pro^ (S1B Fig). Consistently, the substrate-binding affinity and inhibitory potency of compound 27h (see below) against the Lys35Val mutant are comparable to those observed for the WT M^pro^ (S1C–S1F Fig). Sequence alignment (S2 Fig) revealed that SADS-CoV M^pro^ shares ~ 44.8%, 43.8%, 47.4%, 45.9%, 43.9%, 64.2%, and 64.0% homology with human CoVs, including SARS-CoV-2, SARS-CoV, MERS-CoV, HCoV-HKU1, HCoV-OC43, HCoV-229E, and HCoV-NL63, respectively. Structural superposition analysis revealed that evolutionary divergences are primarily localized to the residues 40–53 region (related to the substrate-binding pocket) of Domain I (Figs 1C and S2) and the region of residues 240–249 of Domain III (Figs 1D and S2). Notably, the residues 40–53 region exhibits marked conformational flexibility, with significantly higher B-factor values compared to adjacent regions (Fig 1A and 1C).
Structural characterization of SADS-CoV Mpro and inhibition assays with compounds 27h and SY110.(A) Overall structure of SADS-CoV Mpro. Protomer A is shown as a surface colored according to B-factor values (max(red)–min(blue)). Protomer B is depicted as a cartoon overlaid with surface, where Domain I is colored pink, Domain II green, and Domain III purple. The catalytic residue Cys144 is displayed as a yellow sphere, and His41 as a blue sphere. The N- and C-termini of protomer B are marked. (B) The substrate-binding pocket of SADS-CoV Mpro. Cys144 and His41 are shown in yellow and blue, respectively. The subsites of the Mpro substrate pocket are labeled. (C and D) Structural alignment of SADS-CoV Mpro (magenta, PDB code: 9VUT) Domain I, II (C) and Domain III (D) with seven HCoVs (gray), including SARS-CoV-2 (PDB code: 7MHH), SARS-CoV (PDB code: 1UJ1), MERS-CoV (PDB code: 4WME), HCoV-HKU1 (PDB code: 3D23), HCoV-OC43 (PDB code: 9C7W), HCoV-229E (PDB code: 2ZU2), and HCoV-NL63 (PDB code: 7E6M). The residues 40–53 and 240–249 regions are highlighted with red dashed boxes. The N- and C-termini are marked. (E–I) Inhibition evaluation of 27h and SY110. Chemical structures of compounds 27h (E) and SY110 (F). P1’, P1, P2, and P3 are labeled. The warheads are marked with asterisks. (G) Dose-dependent inhibition of SADS-CoV Mpro by 27h and SY110 in FRET assays. (H) The Ki values of 27h and SY110 against SADS-CoV Mpro were determined using the Morrison Ki assays. Data of (G) and (H) are presented as mean ± SD from three independent experiments (n = 3). (I) Thermal stabilization of SADS-CoV Mpro by 27h and SY110 in DSF assay. Curves represent the average values from three independent experiments (n = 3).
To initiate antiviral research targeting SADS-CoV M^pro^, we screened an in-house chemical library comprising hundreds of SARS-CoV-2 M^pro^ inhibitors, along with approximately 10,000 compounds from commercial libraries, using fluorescence resonance energy transfer (FRET) and differential scanning fluorimetry (DSF) assays. Totally, three compounds—N5084 (S3A Fig), 27h (Fig 1E), and SY110 (Fig 1F)—from our in-house library exhibited inhibition against SADS-CoV M^pro^. N5084 exhibits low IC_50_ (50% inhibition concentration) value (about 1.9 μM, S3B Fig), additionally, this compound displayed the poor metabolic stability [38], which limits its potential for clinical application. In contrast, our previously reported 27h [33] and SY110 [31] display IC_50_ values within the nanomolar range (Fig 1G). Meanwhile, 27h possesses favorable pharmacokinetic properties [33], and SY110 has received clinical trial approval from the Chinese NMPA (No:2023LP01491) [31]. Therefore, we selected 27h and SY110 for further studies. The Ki values of 27h and SY110 against SADS-CoV M^pro^ were subsequently measured to be approximately 130 and 260 nM, respectively (Fig 1H). DSF assays further validated that both 27h and SY110 enhance the thermal stability of SADS-CoV M^pro^ (Fig 1I).
Inhibitory mechanisms of 27h and SY110 against SADS-CoV Mpro
To elucidate the binding modes of inhibitors 27h and SY110 with SADS-CoV M^pro^, we determined crystal structures of the SADS-CoV M^pro^ in complex with 27h (~ 2.4 Å) and SY110 (~ 2.0 Å), respectively (Fig 2A–2D). The electron density maps clearly demonstrate that both inhibitors adopt an inverted conformation within the pocket of M^pro^, with their P1’ groups occupying the S1 pocket and P1 moieties extending into the S1’ subsite (Fig 2A and 2C), which differs from the binding mode of the typical M^pro^ inhibitors, i.e., Nirmatrelvir [27]. For 27h, the α-keto group forms a ~ 1.8-Å covalent bond with the SG atom of Cys144 (Figs 1E and 2B). The hemithioacetal hydroxyl group establishes a ~ 2.6-Å hydrogen bond with the NE2 of His41 (S1 Table). The amide oxygen directs toward the oxyanion hole, forming hydrogen bonds with the backbone NH groups of Cys144 (~ 2.9 Å) and Gly142 (~ 3.0 Å). The amide nitrogen further interacts with the backbone O of Gln163 via a ~ 2.9-Å hydrogen bond. The P1’ pyridinyl ring anchors in the S1 pocket through a ~ 3.2-Å hydrogen bond with the NE2 of His162, and the P1 benzyl group is stabilized in the S1’ pocket via hydrophobic interactions with Thr25 and Leu27. The fluorine of P2 group forms a ~ 3.6-Å hydrogen bond with the OG1 of Thr47 instead of pointing towards the S2 subsite. The P3 amide oxygen makes a hydrogen bond with the backbone N of Glu165 via a water molecule, and its difluorocyclohexyl group engages in sandwich-like hydrophobic interactions with His41 and Pro188 (Fig 2B), supplemented by additional hydrophobic contacts with Thr47 and Ile164.
Crystal structures of SADS-CoV Mpro in complex with 27h and SY110.(A) Inhibitor 27h in the substrate pocket of SADS-CoV Mpro (PDB code: 9VUU). SADS-CoV Mpro is shown in gray, with catalytic dyad residues Cys144 (yellow) and His41 (blue). Compound 27h is displayed in cyan. The four subsites (S1’–S4) of the Mpro and the four moieties (P1’ – P3) of 27h are labeled. The Fo - Fc electron density map (σ = 2.5) for 27h is shown as a black mesh. (B) Interaction network between 27h and SADS-CoV Mpro. Hydrogen bonds are represented as yellow dashed lines, and a water molecule is shown as a red sphere. (C) Inhibitor SY110 in the SADS-CoV Mpro substrate pocket (PDB code: 9VUV). SY110 is colored purple. The Fo - Fc electron density map (σ = 2.5) for SY110 is shown as a black mesh. (D) Interaction network between SY110 and Mpro. (E and F) Structural comparison of Mpro-27h (blue, hydrogen bonds in blue) and Mpro-SY110 (pink, hydrogen bonds in brick red) complexes. The changed direction and distance of the P3 group are indicated.
SY110 exhibits a binding mode similar to that of 27h and detailed interactions are presented in Fig 2D and S1 Table. However, there are still two obvious structural differences between SADS-CoV M^pro^-27h and M^pro^-SY110 structures (Fig 2E and 2F). First, the P2 difluoroproline of 27h directly forms a ~ 3.6-Å hydrogen bond with the OG1 of Thr47, whereas in SY110, this interaction is replaced by a water-mediated hydrogen-bond network, in which the water molecule bridges the backbone nitrogen of P2 and the OG1 atom of Thr47, resulting in a total distance of ~ 5.6 Å. Second, the para-fluorine substitution on the P3 cyclohexane ring in 27h shifts to a meta position in SY110, inducing an upward ~ 2.0-Å conformational shift of the P3 ring and abolishing the sandwich-like hydrophobic interaction with Pro188 in latter. These differences may explain the 2-fold lower Ki value of 27h compared to that of SY110.
Tight binding between 27h (or SY110) and CoV Mpros does not always correlate with potent inhibitory activity
To evaluate the potential of 27h and SY110 as broad-spectrum inhibitors and to obtain a comprehensive understanding of their inhibitory mechanisms against various CoV M^pro^s, we compared binding modes of these two compounds with those our previously reported for SARS-CoV-2 M^pro^ [31,33] and observed notable structural differences in the residues 40–53/54 region of M^pro^s. In SARS-CoV-2 M^pro^-27h structure (Fig 3A), this region (residues 40–54) adopts a coiled coil conformation formed by two consecutive alpha-helices (designated as H1: residues 40–44 and H2: residues 45–51). The H1 helix is present in the SADS-CoV M^pro^-27h structure; however, the H2 helix is replaced by a flexible loop (residues 45–50) (Fig 3A). Similar observations are also found in the apo forms of SARS-CoV-2 and SADS-CoV M^pro^s (S4A Fig). Notably, due to the increased flexibility resulting from the replacement of the H2 helix with a loop, the Cα position of Thr47 in SADS-CoV M^pro^ shifts approximately 6.6 Å closer to the inhibitor compared to the corresponding residue Glu47 (residues of SARS-CoV-2 are underlined) in SARS-CoV-2 M^pro^ (Fig 3A). As a result, the direct interactions between Thr47 of SADS-CoV M^pro^ and the P2 as well as the P3 moieties of 27h, are completely lost in the SARS-CoV-2 M^pro^-27h complex. Meanwhile, the residue Gln189 in SARS-CoV-2 M^pro^ is replaced by Pro188 in SADS-CoV M^pro^, accompanied by a ~ 2.2-Å shift in the position of their Cα atoms, thus, the hydrophobic interaction between Pro188 and the P3 moiety of 27h in the SADS-CoV M^pro^-27h complex is also absent in the SARS-CoV-2 M^pro^-27h counterpart (Fig 3A). Above similar structural deviations were observed in the superimposition of SADS-CoV M^pro^-SY110 and SARS-CoV-2 M^pro^-SY110 complexes as well (Fig 3B). These observations indicate that 27h (or SY110) exhibits a stronger binding affinity to SADS-CoV M^pro^ compared to SARS-CoV-2 M^pro^. Subsequently, we determined the Ki values of 27h (~ 11.1 nM) and SY110 (~ 2.3 nM) against SARS-CoV-2 M^pro^ (S4B Fig). Remarkably, these values show over 10- and 100-fold stronger inhibition compared to their activity against SADS-CoV M^pro^, despite 27h and SY110 establishing stronger interactions with the latter protease. Theoretically, inhibition of CoV M^pro^ is more effective when the inhibitor binds more tightly to this protease [27,29,31–33,39]. However, our results appear to contradict this correlation.
Structural dynamics of residues 40–53/54 in Mpros regulate inhibitor potency.(A) Structural comparison of SADS-CoV Mpro-27h (PDB code: 9VUU) with SARS-CoV-2 Mpro-27h (PDB code: 8I30). SARS-CoV-2 Mpro-27h is colored wheat, with residue labels underlined. The shift of the Cα of Glu47 (SARS-CoV-2) relative to Thr47 (SADS-CoV) is highlighted. The H1 and H2 regions have been outlined. The H2 of SARS-CoV-2 Mpro is shown in orange, and the residues 45–51 of SADS-CoV Mpro are shown in cyan. (B) Structural comparison of SADS-CoV Mpro-SY110 (PDB code: 9VUV, purple) with SARS-CoV-2 Mpro-SY110 (PDB code: 8HHU, wheat). (C) The B-factor values of residues 45–51 for SARS-CoV-2 Mpro (PDB code: 7MHH), Mpro-27h (PDB code: 8I30), Mpro-SY110 (PDB code: 8HHU), and SADS-CoV Mpro (PDB code: 9VUT), Mpro-27h (PDB code: 9VUU), Mpro-SY110 (PDB code: 9VUV). (D) Sequence alignment of residues 40–54 from five β-CoV Mpros and residues 40–53 from three α-CoV Mpros. (E) Structural alignment of residues 40–54 from SARS-CoV-2 (PDB code: 7MHH), SARS-CoV (PDB code: 1UJ1), MERS-CoV (PDB code: 4WME), HCoV-HKU1 (PDB code: 3D23), HCoV-OC43 (PDB code: 9C7W), and residues 40–53 from HCoV-229E (PDB code: 2ZU2), HCoV-NL63 (PDB code: 7E6M), and SADS-CoV Mpro (PDB code: 9VUT). The H1 and H2 helices are highlighted with cyan and red dashed boxes, respectively. (F) Backbone hydrogen bonds in the H2 helix of SARS-CoV-2 Mpro. (G) Structural comparison of the effect of residue '51' on the H2 helix. (H) Substrate affinity of SARS-CoV-2 Mpro WT and its mutants, along with inhibition of 27h. Data are presented as mean ± SD from three independent experiments (n=3). LoopSADS-CoV (and so on): SARS-CoV-2 Mpro mutant in which its residues 40–54 was replaced by residues 40–53 of SADS-CoV Mpro; Mpro-del: Asn51 was deleted in SARS-CoV-2 Mpro; Mpro-TT: Glu47 and Asp48 in SARS-CoV-2 Mpro were mutated to Thr47 and Thr48 in SADS-CoV Mpro, as control.
To understand this discrepancy, we further investigated the regions with dominant structural differences (mentioned above) between SARS-CoV-2 (residues 40–54) and SADS-CoV (residues 40–53) M^pro^s. In the apo forms of these two M^pro^s, two H1 helices (residues 40–44) display the comparable B-factor values (~ 34 Ų in SARS-CoV-2 M^pro^ vs ~ 35 Ų in SADS-CoV M^pro^, S4C Fig). However, the average B-factor of H2 in SARS-CoV-2 (residues 45–51, ~ 45 Ų) is approximately 20 Ų lower than the corresponding loop region of SADS-CoV M^pro^ (residues 45–51, ~ 63 Ų; Figs 3C and S4C). Upon binding of inhibitor 27h or SY110, the average B-factor of the SARS-CoV-2 H2 increases approximately 2-fold (from ~ 45 to ~ 90 Ų, Fig 3C). Strikingly, this value of the corresponding loop region in SADS-CoV M^pro^ exhibits a decrease in B-factor upon binding. Of note, residue Thr47 within this loop forms direct interactions with the inhibitor, resulting in a ~ 5 – 15 Ų reduction in its B-factor (Fig 3C). These findings suggest that inhibitor binding stabilizes the residues 45–51 flexible loop region in SADS-CoV M^pro^.
Then, what is the reason that stronger binding to M^pro^ is associated with lower inhibitory potency? Based on above findings, we hypothesize that the inherent flexibility of the loop region (residues 45–51) in SADS-CoV M^pro^ likely hinders efficient inhibitor engagement during the binding process. The dynamic motion of this loop could interfere with the approach or proper positioning of the inhibitor. However, once binding is established, this loop appears to adopt a more stable conformation, contributing to the overall binding affinity. In contrast, the H2 helix in SARS-CoV-2 M^pro^ is more rigid, likely providing a pre-formed (or larger) binding pocket that accommodates the inhibitor more efficiently—though not necessarily with tighter binding—thereby facilitating stronger inhibitory activity. To support our hypothesis, we measured the kinact/Ki (a rate constant describing the overall inhibition efficiency of the covalent inhibitor, combining both inhibitor reactivity and affinity) values of 27h against SARS-CoV-2 and SADS-CoV M^pro^s (S2 Table), this value of SARS-CoV-2 M^pro^ (~ 12.4 mM^-1^ s^-1^) is about 4.6-fold higher than that of SADS-CoV M^pro^ (~ 2.7 mM^-1^ s^-1^). Furthermore, we determined the kinact/Ki value of a chimeric SARS-CoV-2 M^pro^ variant in which its residues 40–54 were replaced by residues 40–53 of SADS-CoV M^pro^ (designated as Loop_SADS-CoV_). Thus, the kinact/Ki value of the Loop_SADS-CoV_ is ~ 1.2 mM^-1^ s^-1^, similar to that of WT SADS-CoV M^pro^ (S2 Table), indicating that the loop region does affect inhibitor efficiency.
Taken together, these findings could, to some extent, explain why the binding strength between the inhibitor and CoV M^pro^ does not necessarily correlate with its inhibitory potency, and highlight the significant impact of the intrinsic structural stability of M^pro^ on inhibitor efficiency.
Mechanism underlying the formation of variable conformations in the residues 40–53/54 region of α- and β-CoV Mpros
Based on above observations, the residues 40–53/54, particularly, the residues 45–51 (the H2 region), play a critical role in regulating inhibitor potency against SADS-CoV and SARS-CoV-2 M^pro^s. However, it remains unclear whether the conformational changes in the residues 40–53/54 region reflect broader lineage-specific characteristics between α- and β-CoVs, as well as what the exact underlying mechanism is for the formation of these variable conformations.
To this end, we performed a detailed structural analysis of the residues 40–53/54 region across SADS-CoV plus seven HCoVs M^pro^s. Results demonstrated that the N- and C-termini of these regions are anchored through a conserved interaction network in all eight CoVs (S4D Fig): a salt bridge between Arg40 and Asp186, and a hydrogen bond between Tyr53 and Asp186. The H1 helix (residues 40–44), stabilizing by backbone atoms between residues Arg40 and Leu43, and His41 and Ala44, is conserved (S4E Fig). However, the topology of the residues 45–51 region shows clear differences between α- and β-CoV M^pro^s among these eight CoVs (Fig 3D and 3E). In all five β-CoV M^pro^s, this region adopts the H2 helix, which is maintained by backbone hydrogen bonds between residues Thr45 and Met49, Glu47 and Leu50, and Asp48 and Asn51 (Figs 3F and S5). By contrast, in all three α-CoV M^pro^s, this region exhibits a flexible loop conformation. We found that the absence of the corresponding residue ‘51’ in α-CoV M^pro^s introduces a spatial gap of ~ 4.2 Å compared to β-CoV M^pro^s (Fig 3D and 3G). Simultaneously, we noticed that the backbone nitrogen and oxygen atoms of ‘Ile51’ (corresponding to Pro52 in β-CoV M^pro^s, Fig 3D) interact with the side chain of Gln187 in α-CoV M^pro^s, stabilizing two flexible regions (residues 40–53 and 187–190) to sustain the proper S2 pocket (S6A Fig). In SARS-CoV-2 M^pro^, these two regions are stabilized by interactions between Met49 and Gln189 [40]. Thus, due to the involvement of the backbone atoms (N and O) of residue 51 in above mentioned interactions, the corresponding ‘48 and 51’ hydrogen bond of the H2 helix cannot be formed in α-CoV M^pro^s. These findings suggest that the absence of the residue ‘51’ very likely leads to loop formation in the 45–51 region of α-CoV M^pro^s. To validate this conclusion, we determined the crystal structure of SARS-CoV-2 M^pro^ with a deletion of Asn51 (designated as SARS-CoV-2 M^pro^-del) at ~ 1.6 Å (S6B Fig), and found that the H2 helix was indeed replaced by a flexible loop (S6C Fig), with an average B-factor ~ 72 Ų (S6D Fig). In addition, the Ki value of 27h against SARS-CoV-2 M^pro^-del (~ 53 nM) is about 5-fold higher than that for WT SARS-CoV-2 M^pro^ (Fig 3H), further supporting the role of the 45–51 loop region in modulating inhibitor potency. In parallel, we determined the structure of a SADS-CoV M^pro^ variant in which ‘Ile51’ was replaced with ‘Asn51Ile52’ (designated as SADS-CoV M^pro^-add), at a resolution of ~ 2.8 Å (S6E Fig). The results indicated that the H2 helix was indeed restored in this variant (S6F Fig).
Apart from the absence of residue ‘51’, sequence alignment of the residues 45–51/52 region indicated that Pro52 is conserved in five β-CoV M^pro^s, whereas Ile51 is conserved in three α-CoV M^pro^s (Fig 3D). To investigate whether these two residues contribute to conformational changes in this region and thereby affect inhibitor potency, we constructed a SARS-CoV-2 M^pro^-Pro52Ile mutant and evaluated the inhibitory activity of 27h against this mutant. Thus, the Ki and kinact/Ki values of the M^pro^ Pro52Ile are ~ 37.4 nM and ~ 8.6 mM^-1^ s^-1^ (S7 Fig), respectively, compared with ~ 53.4 nM and ~ 3.2 mM^-1^ s^-1^ for the SARS-CoV-2 M^pro^-del mutant and ~ 11.1 nM and ~ 12.4 mM^-1^ s^-1^ for the WT protease (S2 Table). These results suggest that substitution of Pro with Ile could, to some extent, affect the flexibility of the residues 45–52 (or 40–52) region and interfere with inhibitor activity, even though Pro52 is not directly involved in H2 helix formation. This observation is likely attributable to the intrinsic rigidity conferred by proline residues.
Collectively, these findings demonstrate that variable conformations in the residues 45–51 region exist across α- and β-CoV M^pro^s, and suggest that the absence of residue ‘51’ in α-CoV M^pro^s is a key determinant of these conformational changes. Additionally, the conserved Pro52 in β-CoV M^pro^s affects local flexibility to a certain extent.
Clade-specific functional conservation of residues 40–53/54 in α- and β-CoV Mpros
To comprehensively and systematically investigate the role of the 40–53/54 region in M^pro^ substrate affinity and inhibitor susceptibility, we further engineered seven SARS-CoV-2 M^pro^ mutants by replacing its 40–54 loop region with homologous regions from other β-CoVs (SARS-CoV, MERS-CoV, HCoV-HKU1, and HCoV-OC43: residues 40–54) and α-CoVs (HCoV-229E, HCoV-NL63, and SADS-CoV: residues 40–53), designated as Loop_SARS-CoV_, Loop_MERS-CoV_, Loop_HCoV-HKU1_, Loop_HCoV-OC43_, Loop_HCoV-229E_, Loop_HCoV-NL63_, and Loop_SADS-CoV_. FRET assays revealed that chimeric SARS-CoV-2 M^pro^ variants containing analogous regions from α-CoVs (Loop_HCoV-229E_, Loop_HCoV-NL63_, and Loop_SADS-CoV_) showed significantly reduced substrate binding efficiency (KM values: ~ 52.2 – 127.1 μM, Fig 3H), whereas variants incorporating the corresponding regions from other β-CoVs (Loop_SARS-CoV_, Loop_MERS-CoV_, Loop_HCoV-HKU1_, Loop_HCoV-OC43_) exhibited similar binding affinities (~ 28.2 – 39.5 μM). Next, we investigated the inhibitor susceptibility of these variants. Due to the stronger inhibitory activity of 27h compared to SY110 (Fig 1G and 1H), subsequent experiments prioritized 27h for further study. The results showed that 27h exhibited markedly reduced potency against Loop_HCoV-229E_, Loop_HCoV-NL63_, and Loop_SADS-CoV_ (α-CoV) mutants (Ki: ~ 45.6 – 93.9 nM, Fig 3H), but maintained high potency against Loop_SARS-CoV_, Loop_MERS-CoV_, Loop_HCoV-HKU1_, and Loop_HCoV-OC43_ (β-CoV) mutants (~ 6.8 – 21.0 nM). Overall, these results suggest that the 40–53/54 region of M^pro^ is functionally conserved in both substrate binding affinity and the regulation of inhibitor potency within each CoV genus (α or β), with little intra-genus variation.
27h significantly suppresses SADS-CoV replication in vitro
Currently, no anti-SADS-CoV therapeutic is available. To evaluate the antiviral efficacy of compound 27h against SADS-CoV in cell-based assays, we tested three treatment regimens in Huh7 cells: pre-treatment (2 h before infection), co-treatment (at infection), and post-treatment (2 h post-infection). Our data showed that 27h effectively suppressed viral replication with EC_50_ (50% effective concentration) values of ~ 10 μM, 9.2 μM, and 7.8 μM, respectively (Fig 4A–4C). Western blot (WB) analysis confirmed that treatment with 2.5 μM 27h nearly abolished viral nucleocapsid (N) protein expression (Fig 4D–4F). Due to comparable inhibitory activity across treatment timings, the post-treatment regimen was selected for subsequent experiments. We next assessed the anti-SADS-CoV activity of compound 27h in two porcine-derived cell lines: swine testicular (ST) and porcine small intestinal epithelial (IPEC-J2) cells. In ST cells, 27h effectively inhibited SADS-CoV replication (EC_50_: ~ 4.8 μM, Fig 4G), with 2.5 μM almost completely blocking N protein expression (Fig 4H). In IPEC-J2 cells, no clear dose dependency was observed within the 2.5 – 10 μM range (Fig 4I), however, 2.5 μM 27h still significantly reduced N protein levels (Fig 4J). Collectively, 27h effectively inhibits SADS-CoV replication in multiple cell lines, highlighting its potential as an antiviral candidate.
Anti-SADS-CoV activity of compound 27h in vitro.(A–C) RT-qPCR quantification of viral genome copies in Huh7 cells treated with 27h at pre-treatment (A), co-treatment (B), and post-treatment (C) regimens. Data are presented as mean ± SD from three independent experiments (n = 3). (D–F) WB analysis of viral N protein expression in Huh7 cells treated with 27h at pre-treatment (D), co-treatment (E), and post-treatment (F). (G) RT-qPCR quantification of viral genome copies in ST cells treated with 27h at post-treatment. Data are presented as mean ± SD from four independent experiments (n = 4). (H) WB analysis of viral N protein expression in ST cells treated with 27h at post-infection. (I) RT-qPCR quantification of viral genome copies in IPEC-J2 cells treated with 27h at post-treatment. Data are presented as mean ± SD from four independent experiments (n = 4). (J) WB analysis of viral N protein expression in IPEC-J2 cells treated with 27h at post-treatment.
27h inhibits the replication of SADS-CoV in porcine small intestinal organoids
The pig intestinal organoid infection system closely mimics physiological SADS-CoV infection [41]. However, to the best of our knowledge, no any anti-SADS-CoV M^pro^ agent was evaluated using this system currently. Thus, using a human intestinal organoid protocol [42], we established porcine small intestinal organoids from 6-month-old pigs (Fig 5A). Undifferentiated organoids rapidly proliferated in amplification medium with R-spondin-1, Wnt3a, Noggin, and EGF (epidermal growth factor), and were passaged at a 1:4 ratio every 5 days (Fig 5B), sustaining over 20 generations across 6 months. Differentiation was induced by switching to differentiation medium on day 2 for 4 days, leading to luminal morphological changes (S8A Fig), LYZ (Lysozyme) and LGR5 (Leucine-rich repeat-containing GPCR5) downregulation (S8B Fig), and upregulation of ALPI (Intestinal alkaline phosphatase), MUC2 (Mucin 2), and CHGA (Chromogranin A). Immunofluorescence confirmed the presence of mature cell types (S8C Fig). We next assessed the susceptibility of differentiated porcine intestinal organoids to SADS-CoV. After mechanical shearing and viral inoculation, increasing viral RNA and progeny virus were detected in supernatants at 24, 48, and 72 h post-infection (Fig 5C). Immunofluorescence showed abundant SADS-CoV N protein in villous epithelial cells (Fig 5D), indicating effective infection in the organoid model.
Establishment of porcine small intestinal organoids and 27h inhibition assay.(A) A schematic graph of the porcine intestinal organoids culture system was created in BioRender. Cui, S. (2026) https://BioRender.com/lzxisaq, https://BioRender.com/472459m, https://BioRender.com/f3l8lyx, and https://BioRender.com/46d1swl. (B) Photomicrographs show growing organoids on day 1–5 (D1-5). (C) Culture media were harvested from the infected porcine intestinal organoids and subjected to viral load detection and viral titration by TCID50 assay. (D) SADS-CoV-infected or mock-infected porcine intestinal organoids were fixed and immunostained to identify viral N protein (NP, green)-positive cells. Villin, nuclei and actin filaments were counterstained with Villin (orange), DAPI (blue) and F-actin (purple), respectively. (E) Porcine intestinal organoids infected with SADS-CoV were treated with 50 μM 27h for 24 hours. The supernatants were subsequently collected to determine the viral copy number. (F) The EC50 curve depicts the inhibition of SADS-CoV genome copies by 27h in cell lysate, with a maximum drug concentration of 10 μM. Data are presented as mean ± SD from three independent experiments (n = 3).
To assess the antiviral effect of 27h in organoid model, we inoculated the organoids with SADS-CoV at a multiplicity of infection (MOI) of 0.5, and added 27h two hours post-infection. The results demonstrated that, compared to the control group, 27h significantly reduced the SADS-CoV copy number in the supernatant (Fig 5E), and exerted its antiviral effects in a dose-dependent manner in cell lysate, with an EC_50_ value of ~ 896 nM (Fig 5F).
Overall, 27h effectively inhibits SADS-CoV replication in the organoid model, supporting the feasibility of using porcine organoids for screening and evaluating anti-SADS-CoV M^pro^ inhibitors.
27h potently inhibits the activity of multiple coronavirus Mpros
To evaluate the broad-spectrum anti-coronavirus potential of compound 27h, we assessed its inhibitory activity against M^pro^s of all other HCoVs. Results demonstrated that 27h exhibits nanomolar-level inhibition potency against β-CoV (SARS-CoV-2, SARS-CoV, MERS-CoV, HCoV-HKU1, and HCoV-OC43) M^pro^s, with Ki values ranging from ~ 1.7 to 115.5 nM (Fig 6) and the corresponding kinact/Ki values of ~ 9.8 – 21.8 mM^-1^ s^-1^ (S2 Table). For α-CoV (HCoV-229E, HCoV-NL63, and SADS-CoV) M^pro^s, 27h exhibited slightly attenuated inhibitory activity, with Ki values spanning ~ 129 – 1675 nM (Fig 6) and kinact/Ki values of ~ 0.9 – 3.6 mM^-1^ s^-1^ (S2 Table). Overall, 27h exhibits higher inhibitory activity against β-CoV M^pro^s than against α-CoV M^pro^s. Meanwhile, these results indicate that 27h possesses broad-spectrum anti-CoV activity against both α- and β-CoV M^pro^s.
Inhibition of Mpro from eight CoVs by 27h.(A–H) The Ki values of 27h against Mpros of SARS-CoV-2 (A), SARS-CoV (B), MERS-CoV (C), HCoV-HKU1 (D), HCoV-OC43 (E), HCoV-229E (F), HCoV-NL63 (G), and SADS-CoV (H). The Ki curves for α- and β-CoV Mpros are shown in orange and cyan, respectively. Data are presented as mean ± SD from three independent experiments (n = 3).
Discussion
The cross-species transmission of coronaviruses has become a persistent threat to global public health. SADS-CoV has exhibited a recurrent regional outbreak pattern in pig farms [18–22]. Particularly, its ability to infect various human along with other mammalian cell lines underscores its significant zoonotic risk [14,15]. However, no approved drug or vaccine targeting SADS-CoV is available currently. The lack of structural information on the key anti-CoV target, M^pro^, in SADS-CoV, severely hampers the development of antiviral inhibitors. Meanwhile, reevaluating reported SARS-CoV-2 M^pro^ drug candidates against SADS-CoV M^pro^ can offer effective antiviral strategies.
In this study, to fill the gap in structural studies of SARS-CoV M^pro^, we determined the high-resolution structure of this enzyme. Upon high-throughput screening and reevaluation, two previously reported SARS-CoV-2 inhibitors 27h [33] and SY110 [31] were found to possess proper inhibitory activity against SADS-CoV M^pro^. We observed that the H2 region (residues 45–51) of SARS-CoV-2 M^pro^ is replaced by a flexible loop in the corresponding residues 45–50 region of SADS-CoV M^pro^. These distinct structural differences are conserved between α- and β-CoV M^pro^s. Through mutagenesis and structural approaches, we confirmed that the absence of residue ‘51’ is a critical factor responsible for these structural differences. Of note, these structural conformational differences lead to a lack of correlation between the binding affinity of the inhibitor to CoV M^pro^ and its inhibitory potency. These findings contrast with earlier views and suggest that the dynamic behavior of this loop should be considered in the design of α-CoV (or broad-spectrum) inhibitors.
Organoid-based infection models offer more physiologically relevant data than traditional cell lines for assessing antiviral activity and toxicity. They also support high-throughput, mechanistic studies while minimizing interspecies differences and ethical issues of animal models. Thus, we established a porcine intestinal organoid-based infection model and, for the first time, demonstrated its applicability for the evaluation of SADS-CoV M^pro^ inhibitors. 27h demonstrated potent antiviral efficacy with an EC_50_ of ~ 896 nM in this model. Nirmatrelvir is currently the most successful anti-CoV drug. No studies have yet evaluated its efficacy against SADS-CoV, thereby, we tested its enzymatic activity, yielding a Ki of 16.1 nM and a kinact of 1.26 × 10^-^⁴ s^-^¹ against SADS-CoV M^pro^ (S9A–S9C Fig). In cellular assays as determined by RT-qPCR (S9D–S9F Fig), the EC_50_ values of Nirmatrelvir were ~ 2.5 μM in Huh7 cells and ~ 1.5 μM in ST cells. Western blot analysis demonstrated that ~ 1.25 μM Nirmatrelvir significantly reduced the expression of the viral N protein (S9G–S9I Fig). In porcine intestinal organoids, it exhibited potent antiviral activity with an EC_50_ of ~ 150 nM (S9J Fig). At present, the greatest challenge to Nirmatrelvir lies in the emergence of resistant variants, notably the Glu166-to-Val mutation [43], reported to decrease its antiviral efficacy by > 100-fold. Structural analysis reveals that, in contrast to Nirmatrelvir, the side chain of Glu166 (Glu165 in SADS-CoV) does not participate in binding with 27h (S9K and S9L Fig), thereby conferring resistance to the Glu166-to-Val mutation. Furthermore, 27h exhibits comparable pharmacokinetics to those of Nirmatrelvir (half-life: 2.83 h vs. 2.80 h; oral bioavailability: 30% vs. 34%) [27,33], positioning it as a promising candidate against Nirmatrelvir-resistant strains. Nirmatrelvir synthesis requires cryogenic conditions and an excess lithium bis(trimethylsilyl)amide, increasing production costs for large-scale manufacturing [44]. In contrast, 27h is synthesized via a simpler and more cost-effective process [33], making it better suited for widespread use—particularly in treating CoV-infected animals. Thus, although Nirmatrelvir showed higher anti-SADS-CoV activity than compound 27h, the latter exhibited greater resistance to viral variants, lower production costs, and broader host applicability.
The absence of inhibitors targeting M^pro^s across all four CoV genera (α, β, γ, and δ) highlights the challenge in developing pan-CoV antivirals. Since only α- and β-CoVs infect humans currently, prioritizing broad-spectrum inhibitors against α/β-CoV M^pro^s are more practical and clinically relevant. 27h provides a scaffold for developing α/β-CoV M^pro^ inhibitors, due to its inhibitory activity against the M^pro^ of eight CoVs (Fig 6). 27h showed the strongest inhibition against HCoV-OC43 M^pro^ and the weakest against HCoV-NL63 M^pro^. Interestingly, HCoV-OC43 M^pro^ exhibits the lowest substrate binding affinity (S3 Table), while HCoV-NL63 M^pro^ displays an affinity comparable to SADS-CoV M^pro^. Subsequently, we superimposed HCoV-OC43 M^pro^ onto SARS-CoV-2 M^pro^-27h and HCoV-NL63 M^pro^ onto SADS-CoV M^pro^-27h, and found that their inhibitor-binding regions are nearly identical with each other (S10 Fig). These findings suggest that the design of CoV M^pro^ inhibitors should go beyond traditional structure-based affinity optimization and incorporate conformational dynamics and/or topological features to improve antiviral efficacy.
In summary, we present the SADS-CoV M^pro^ crystal structure, addressing a key structural gap. We elucidate that residues 40–53/54 (in particular, residues 45–50/51) modulate inhibitor potency in α- and β-CoV M^pro^s, and uncover the underlying mechanism. We also demonstrate the anti-SADS-CoV activity of 27h in cells and porcine organoids, highlighting its potential as a broad-spectrum M^pro^ inhibitor. Overall, our work reveals how intrinsic M^pro^ dynamics modulate inhibitor activity, offering new insights for future anti-CoV inhibitor development.
Materials and methods
Ethical statement
All animal experiments were conducted by the guidelines approved by the Animal Ethics Committee of Peking University Third Hospital (Approval No. SA20250653).
Cell lines and virus
Huh7 (Haixing Biosciences, China, RRID: CVCL_0336), ST (Shanghai Sur Biotech Co., Ltd., China, RRID: CVCL_2204), and IPEC-J2 (Zhejiang Meisen Cell Technology Co., Ltd., China, RRID: CVCL_2246) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Biosciences, C3113) supplemented with 10% (v/v) fetal bovine serum (Vistech, SE100) and 1% (v/v) penicillin/streptomycin (Yeasen Biotechnology Co., Ltd., China). Cells were cultured in an incubator at 37°C with 5% CO_2_. The SADS-CoV strain CN/GDWT/2017 (GenBank accession no. MG557844, provided by Prof. Huahao Fan, Tianjin university), was propagated in Huh7 cells, and viral titers were determined by TCID_50_ assay on Huh7 cells.
Compound library
The clinical & FDA approved drug library, anti-virus compound library, covalent inhibitor custom library, mini scaffold library, and ubiquitination library were purchased from APExBIO. The natural product library and Nirmatrelvir were purchased from TargetMol.
Plasmids construction
The gene sequences encoding the main proteases (M^pro^) of SADS-CoV(GenBank number: MG557844), SARS-CoV-2 (GenBank number: NC_045512), SARS-CoV (GenBank number: NC_004718), MERS-CoV (GenBank number: NC_019843), HCoV-HKU1 (GenBank number: NC_006577), HCoV-OC43 (GenBank number: NC_006213), HCoV-229E (GenBank number: NC_002645), and HCoV-NL63 (GenBank number: NC_005831) were codon-optimized, synthesized, and cloned into the pET28b vector via NcoI and XhoI sites. To generate native M^pro^, six residues from the C-terminus of Nsp4 were added to the N-terminus to enable auto-cleavage during expression. A PreScission protease cleavage site was introduced at the C-terminus to generate the native C-terminus after cleavage. SADS-CoV M^pro^ wild-type (WT) and its mutants (M^pro^-Cys219Ser, M^pro^-Cys224Ser, M^pro^-Cys259Ser, M^pro^-Lys35Val, M^pro^-Lys69Val, M^pro^-Lys106Val, M^pro^-Lys155Val and M^pro^-Lys35Val/Cys219Ser, M^pro^-Lys35Val/Cys224Ser, M^pro^-Lys35Val/Cys259Ser, and M^pro^-add-Lys35Val/Cys224Ser) were generated via site-directed PCR using the recombinant SADS-CoV M^pro^ plasmid as a template. SARS-CoV-2 M^pro^ mutants (Loop_SARS-CoV_, Loop_MERS-CoV_, Loop_HCoV-HKU1_, Loop_HCoV-OC43_, Loop_HCoV-229E,_ Loop_HCoV-NL63_, Loop_SADS-CoV_, M^pro^-del, M^pro^-del-Lys97Val (facilitating crystallization), M^pro^-TT and M^pro^-Pro52Ile) were constructed via site-directed PCR using the SARS-CoV-2 M^pro^ plasmid as a template. Primer sequences are listed in S4 Table. PCR products were digested with DpnI, transformed into E. coli DH5α competent cells, and validated by sequencing (Tsingke Biotech Co., Ltd., China).
Protein expression and purification
Verified plasmids were transformed into E. coli BL21(DE3) competent cells. Cultures were grown in LB medium to an OD_600_ of ~ 0.6 – 0.8 and induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 18°C for ~ 18 hours. Cells were harvested by centrifugation (7,500 rpm (revolutions per minute), 5 min), resuspended in Buffer A (20 mM HEPES pH 7.5, 10 mM imidazole, 500 mM NaCl, 10% glycerol), and lysed by sonication. Lysates were centrifuged (18,000 rpm, 4°C, 1 h), filtered through 0.45 μm membranes (Biofil, China), and incubated with Ni-NTA resin (Senhui Microsphere Technology Co., Ltd., China). After washing with 5% Buffer B (20 mM HEPES pH 7.5, 500 mM imidazole, 500 mM NaCl, 10% glycerol), proteins were eluted with Buffer B. His-tags were removed using PreScission protease, and proteins were dialyzed into Buffer C (20 mM HEPES pH 7.5, 150 mM NaCl). Native M^pro^ was further purified via a second Ni-NTA column and gel filtration (Superdex 75 Increase 10/300 GL, GE Healthcare) in Buffer C (20 mM HEPES pH 7.5, 150 mM NaCl). Protein purity was confirmed by SDS-PAGE.
Crystallization, data collection, phase determination, and refinement
To obtain crystals of SADS-CoV M^pro^ WT, M^pro^-27h, M^pro^-SY110, M^pro^-add and SARS-CoV-2 M^pro^-del, purified M^pro^ WT and M^pro^ mutants (mentioned above) were concentrated at 4°C using 10-kDa molecular weight cut-off centrifugal filters (Millipore, USA) by centrifugation at 3,500 rpm, and the final protein concentrations (~ 8 mg/mL) were measured using a Nano-300 microvolume spectrophotometer (Allsheng, China). Inhibitor powders stored at -80°C were dissolved in dimethyl sulfoxide (DMSO) to prepare 100 mM stock solutions, which were then aliquoted and stored at -20°C. The concentrated proteins were then incubated with the inhibitors at a 1:10 molar ratio overnight.
Commercial crystallization screens were used. SADS-CoV M^pro^-Lys35Val crystals were obtained in JCSG-plus condition 24 (0.2 M potassium citrate tribasic monohydrate, 20% w/v PEG 3350, Molecular Dimensions). M^pro^-Lys35Val-27h crystals were obtained in JBScreen Basic 2 condition D9 (20% w/v PEG 8000, 100 mM MES pH 6.5, 200 mM magnesium acetate, Jena Bioscience), while M^pro^-Lys35Val/Cys224Ser-SY110 crystals appeared in PACT premier condition 27 (0.1 M PCTP pH 6.0, 25% w/v PEG 1500, Molecular Dimensions). SARS-CoV-2 M^pro^-del-Lys97Val crystals were obtained in Morpheus condition 1–44 (0.12 M Alcohols, 0.1 M Buffer System 2 pH 7.5, 37.5% v/v Precipitant Mix 4, MiTeGen). SADS-CoV M^pro^-add-Lys35Val crystals were obtained in Morpheus condition 2–39 (0.1 M Amino acids, 0.1 M Buffer System 1 pH 6.5, 30% v/v Precipitant Mix 3, MiTeGen). The SADS-CoV M^pro^, M^pro^-27h, and M^pro^-SY110 crystals were cryo-protected using 25% (v/v) PEG 400, while the M^pro^-add and M^pro^-del crystals were flash-frozen directly without additional cryoprotectant.
Diffraction data for SADS-CoV M^pro^-Lys35Val, M^pro^-Lys35Val-27h, M^pro^-Lys35Val/Cys224Ser-SY110, SARS-CoV-2 M^pro^-del-Lys97Val were collected at beamline BL19U1 of the Shanghai Synchrotron Radiation Facility (SSRF) (wavelength: 0.97853 Å), and M^pro^-add-Lys35Val/Cys224Ser was collected at beamline BL18U1 of the SSRF (wavelength: 0.97852 Å). All data were processed using XDS [45] and scaled with CCP4 Aimless [46], molecular replacement was carried out by CCP4 Molrep [47] with the initial model (PDB code: 8I30 [33]). Then, refinement was performed using CCP4 Refmac [48] and Phenix.refine [49], and the model was manually adjusted in Coot [50]. Diffraction data and refinement statistics are in S5 Table. Structural figures were generated using UCSF ChimeraX (https://www.cgl.ucsf.edu/chimerax/).
Enzyme activity assays
Kinetic assays were performed in optimized buffer containing 20 mM HEPES pH 6.5, 120 mM NaCl, 0.4 mM EDTA, 4 mM dithiothreitol, and 20% glycerol. M^pro^ (final concentration 0.2 μM) was incubated with MCA-STLQ↓AGLK-Lys(Dnp)-Lys-NH_2_ substrate (final concentration 3.1 – 200 μM, the cleavage site is located between Nsp4 and Nsp5 of SADS-CoV and the arrow indicates cleavage site, GL Biochem (Shanghai) Ltd., China). Fluorescence (excitation/emission: 320/405 nm) was monitored using SynergyHTX microplate reader (BioTek, USA). Initial velocities were calculated using an MCA standard curve (0.16 – 20 μM) and analyzed by linear regression. KM values were derived via Michaelis-Menten analysis in GraphPad Prism 8.0.
For 50% inhibition concentration (IC_50_) and inhibition constant (Ki) assays, M^pro^ (final concentration 0.2 μM) was preincubated with 27h, SY110, or Nirmatrelvir for 10 min at room temperature, followed by substrate addition (final concentration 35 μM). Inhibition relative to DMSO controls was calculated, and IC_50_ values were determined using dose-response models in GraphPad Prism 8.0. Initial velocities were calculated as above and Ki values were analyzed by Morrison equation in GraphPad Prism 8.0.
In the inactivation rate constant (kinact)/Ki assay, M^pro^ (final concentration 0.2 μM) was mixed with various concentrations of 27h or Nirmatrelvir. The reactions were immediately initiated by adding the fluorescent substrate (final concentration: 35 μM). The reactions were monitored by SynergyHTX microplate reader for 15 min. The observed first-order rate constant (kobs) was determined by fitting progress curves of various inhibitor concentrations to the one-phase association equation in GraphPad Prism 8.0 software. The constant kobs/[I] was obtained by linear regression analysis. The kinact/Ki value was calculated according to the equation: kinact/Ki = (1 + [S]/KM) × kobs/[I]. Subsequently, the kinact value was determined according to the equation: kinact = (kinact/Ki) × Ki.
Differential scanning fluorimetry (DSF) assay
SADS-CoV M^pro^ (final concentration 2 μM) was incubated with 40 μM 27h or SY110 for 30 min at room temperature. SYPRO Orange dye (1:10,000, Sigma-Aldrich, USA) was added, and thermal denaturation (20 – 95°C, 1.5°C/min) was monitored using a BioRad CFX96 RT-PCR system (Bio-Rad, USA). Melt temperature (Tm) and thermal shift ΔTm (Tm(compound) − Tm(DMSO)) were calculated using the Boltzmann model in GraphPad Prism 8.0.
50% tissue culture infective dose (TCID50) assay
Huh7 cells seeded in 48-well plates were infected with serially 10-fold diluted virus. Count the number of wells with cytopathic effects. Calculate the proportionate distance (PD) according to the formula: proportionate distance (PD) = (positive above 50% − 50%)/ (positive above 50% − positive below 50%). Calculate the TCID_50_ according to the formula: lgTCID_50_ = lg(dilution with > 50% positive) + PD × (-lg(dilution factor)).
Reverse transcription quantitative PCR (RT-qPCR)
Huh7, ST, and IPEC-J2 cells were seeded in 24-well plates. For Huh7 cells, 27h (final concentrations of 0.625 – 10 μM) was added 2 h pre-infection, during infection, and post-infection. Nirmatrelvir (final concentrations of 0.3125 – 5 μM) was added only post-infection. For both compounds, the multiplicity of infection (MOI) was 0.1 in Huh7 cells. In ST and IPEC-J2 cells, 27h (final concentrations of 0.625 – 10 μM) or Nirmatrelvir (final concentrations of 0.3125 – 5 μM) was added post-infection (MOI = 0.1 and 1, respectively). Total RNA was extracted (Vazyme, RC101–01) and reverse-transcribed using Hifair Advanced One Step RT-qPCR SYBR Green Kit (Yeasen, 11175ES20) according to the manufacturer’s instruction. Primers are listed in S6 Table.
Western blot (WB) analysis
Huh7, ST, and IPEC-J2 cells were seeded in 12-well plates. Inoculate with the virus and compound 27h or Nirmatrelvir as above. After 24 hours of infection, total proteins were extracted from cells using the cell lysis buffer (Beyotime, P0013). Proteins were separated by SDS-PAGE, transferred to PVDF membranes (Millipore ISEQ00010), blocked with 5% skim milk, and probed with primary antibody (anti-SADS-CoV N protein, a gift from Prof. Huahao Fan, Tianjin university) or GAPDH antibody (Proteintech, RRID: AB_2107436). HRP-conjugated Goat Anti-Mouse IgG(H + L) antibody was obtained from Proteintech Biotechnology (China, RRID: AB_2722565). Signals were detected using enhanced chemiluminescent reagent (Abbkine, BMU102) and observed using Tanon-5200 imager (Tanon, China).
Establishment, maintenance, and differentiation of porcine intestinal organoids
To establish porcine small intestinal organoids, we adapted the protocol for human intestinal organoids described by Zhou et al. [42]. In brief, small intestines from 6-month-old pigs were collected from an abattoir and preserved in saline. Sections of the intestinal segments were excised using a scalpel, and the intestinal contents and villi were removed by gently scraping the intestinal wall with a glass slide. After washing with cold PBS, the tissue was cut into small pieces (<1 mm) and digested with 2 mg/mL collagenase (Sigma Aldrich, Germany) for 15 minutes at 37°C. The digested tissue was then sheared using a glass Pasteur pipette (Brand, Germany) and filtered through a 100 μm cell filter (FALCON, USA). The resulting single cells were embedded in matrix gel (Corning, USA) and seeded in 24-well culture plates (Corning, USA). Once the Matrigel solidified, 500 μL of expansion medium (S7 Table) was added to each well, and cultures were maintained at 37°C in a humidified incubator with 5% CO_2_. The porcine intestinal organoids were maintained in expansion medium, with media changes every two days and passaging every five days. To induce differentiation, the expansion medium (S7 Table) was replaced with differentiation medium on day 2 of growth.
Viral infection and detection of porcine intestinal organoids
Differentiated porcine small intestinal organoids were sheared using a glass Pasteur pipette and incubated with SADS-CoV at a MOI of 0.5 for 2 hours at 37°C. After removal of the virus, the organoids were rinsed with PBS, re-embedded in Matrigel, and left to solidify before adding the corresponding medium. To assess replication kinetics, cell-free culture supernatants were collected at the indicated time points post-inoculation. RNA was extracted using the Nucleic Acid Extraction Kit (Vazyme) and analyzed for viral load by RT-qPCR, as well as viral titers by TCID_50_ assay. Primers are listed in S8 Table.
Immunofluorescence staining of porcine intestinal organoids
Immunofluorescence staining of porcine small intestinal organoids was performed to identify specific cell types and virus-infected cells. Briefly, the organoids were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.5% Triton X-100, and blocked with 5% bovine serum albumin (BSA). The organoids were then incubated overnight at 4°C with primary antibodies: rabbit anti-Villin (Abcam, UK), rabbit anti-MUC2 (Abcam, UK), or murine anti-SADS-CoV N protein (Abcam, UK), all diluted in 5% BSA. Afterward, the organoids were incubated with the corresponding secondary antibodies for 1 – 2 hours at room temperature. Nuclei were stained with DAPI (Invitrogen, USA) and actin filaments were stained with Phalloidin-647 (Sigma Aldrich, Germany). Confocal images were acquired using a Nikon AX laser confocal microscope (Nikon, Japan).
Statistical analysis
All statistical analyses were performed using GraphPad Prism 8.0. Statistical differences between two groups were analyzed using Student’s t-test, with considered significant at *P < 0.05, **P < 0.01, ***P < 0.001, and ns (not statistically significant) indicating P > 0.05. All data are presented as the mean ± SD from at least three independent experiments (n ≥ 3).
Supporting information
S1 FigCharacterization of the SADS-CoV M^pro^-Lys35Val mutant.(A) Size-exclusion chromatography (SEC) profile and SDS-PAGE of the M^pro^-Lys35Val mutant. (B) Cα distance between Val35 and the catalytic residue Cys144. Val35 is shown in magenta sticks and Cys144 is shown in yellow sticks. (C and D) Michaelis-Menten kinetics of the M^pro^ (C) and M^pro^-Lys35Val mutant (D) determined by a fluorescence resonance energy transfer (FRET)-based enzymatic assay. (E and F) Inhibition potency of the M^pro^ (E) and M^pro^-Lys35Val mutant (F) by compound 27h. Data of (C–F) are presented as mean ± SD (n = 3).(TIF)
S2 FigMultiple sequence alignment of M^pro^ among SARS-CoV-2, SARS-CoV, MERS-CoV, HCoV-HKU1, HCoV-OC43, HCoV-229E, HCoV-NL63, and SADS-CoV.The sequence accession numbers are: SARS-CoV-2 (GenBank number: NC_045512), SARS-CoV (GenBank number: NC_004718), MERS-CoV (GenBank number: NC_019843), HCoV-HKU1 (GenBank number: NC_006577), HCoV-OC43 (GenBank number: NC_006213), HCoV-229E (GenBank number: NC_002645), HCoV-NL63(GenBank number: NC_005831) and SADS-CoV (GenBank number: MG557844). The regions comprising residues 40–53/54 and 240–249 (or 241–252) have been highlighted with red boxes. The residues involved in anchoring the N- and C- termini of the 40–53/54 region are indicated by black arrows. The residues Ile51 and Gln187 involved in stabilizing SADS-CoV M^pro^ residues 40–53 and 187–190 are shown in red. The absent residue ‘51’ is labeled with a black triangle.(TIF)
S3 FigInhibition of SADS-CoV M^pro^ by N5084.(A) Chemical structure of N5084. The warhead is indicated with a red asterisk. (B) N5084 inhibits the activity of SADS-CoV M^pro^ in a dose-dependent manner. Data are presented as mean ± SD from three independent experiments (n = 3).(TIF)
S4 FigConformational analysis of residues 40–53/54 and inhibition of SARS-CoV-2 M^pro^ by 27h and SY110.(A) Structural superposition of M^pro^ residues 40–53/54 between SADS-CoV (cyan, PDB code: 9VUT) and SARS-CoV-2 (orange, PDB code: 7MHH). The N-terminus and C-terminus are labeled. H1: H1 helix; H2: H2 helix. (B) Inhibition of SARS-CoV-2 M^pro^ by 27h and SY110. Data are presented as mean ± SD from three independent experiments (n = 3). (C) The B-factor values of residues 40–54 of SARS-CoV-2 M^pro^, M^pro^-27h (PDB code: 8I30), M^pro^-SY110 (PDB code: 8HHU), and residues 40–53 for SADS-CoV M^pro^, M^pro^-27h (PDB code: 9VUU), M^pro^-SY110 (PDB code: 9VUV). The yellow-highlighted region indicates the residues comprising the H2 helix. (D) Conserved interactions at the N- and C-termini of residues 40–53/54. α-CoVs are colored in cyan; β-CoVs are in orange. Key interactions (hydrogen bond: yellow dashed lines, salt bridge: purple dashed lines) and loop termini are labeled. (E) Hydrogen bonds involving the main chain of M^pro^ H1 helix in SARS-CoV-2, SARS-CoV (PDB code: 1UJ1), MERS-CoV (PDB code: 4WME), HCoV-HKU1 (PDB code: 3D23), HCoV-OC43 (PDB code: 9C7W), HCoV-229E (PDB code: 2ZU2), HCoV-NL63 (PDB code: 7E6M), and SADS-CoV. Hydrogen bonds are shown with yellow dashed lines.(TIF)
S5 FigThe H2 helix structures of M^pro^ from four β-CoVs.(A–D) Hydrogen bonds involving the main chain of M^pro^ H2 helix in SARS-CoV (A, PDB code: 1UJ1), MERS-CoV (B, PDB code: 4WME), HCoV-HKU1 (C, PDB code: 3D23), and HCoV-OC43 (D, PDB code: 9C7W). Hydrogen bonds are shown with yellow dashed lines.(TIF)
S6 FigThe absent residue ‘51’ modulates H2 helix formation in M^pro^s.(A) The hydrogen bond interactions between Ile51 and Gln187 of M^pro^ in HCoV-229E (PDB code: 2ZU2), HCoV-NL63 (PDB code: 7E6M), and SADS-CoV (PDB code: 9VUT). (B) Structure of SARS-CoV-2 M^pro^-del (deletion of Asn51, PDB code: 9VUW) residues 40–53. Residues 40–53 are shown as cyan sticks. The 2Fo - Fc electron density map (σ = 1.0) for residues 45–51 is displayed as a black mesh. (C) Structural superposition of residues 40–53/54 in SARS-CoV-2 M^pro^ (wild-type (WT), purple, PDB code: 7MHH) and M^pro^-del (cyan, PDB code: 9VUW). (D) The B-factor values of residues 40–53 of SARS-CoV-2 M^pro^-del. The yellow-highlighted region indicates the residues comprising the H2 helix. (E) Structure of residues 40–54 of SADS-CoV M^pro^-add (residue Ile51 is replaced by Asn51Ile52, PDB code: 9VUX). Residues 40–54 are displayed in orange, and residues 45–51 are shown as sticks. The 2Fo - Fc electron density map (σ = 0.9) for residues 45–51 is displayed as a black mesh. (F) Structural superposition of residues 40–53/54 SADS-CoV M^pro^ (WT, wheat, PDB code: 9VUT) and M^pro^-add (orange, PDB code: 9VUX).(TIF)
S7 FigCharacterization of the Pro52 in modulating SARS-CoV-2 M^pro^ inhibitor efficiency.(A) Inhibition constant (Ki) curve of compound 27h against SARS-CoV-2 M^pro^-Pro52Ile. Data are presented as mean ± SD (n = 3). (B) Inhibition efficiency (kinact/Ki) curve of compound 27h against SARS-CoV-2 M^pro^-Pro52Ile. Data points represent the mean of three independent replicates.(TIF)
S8 FigEstablishment of porcine small intestinal organoids.(A) The left panel shows undifferentiated porcine small intestinal organoids, and the right panel shows differentiated porcine small intestinal organoids. Differentiation time was 4 days (magnification ×100). (B) Fold change of expression levels of cell-type markers in the organoids cultured in different differentiation media versus those cultured in expansion medium. Data show the mean and SD of three independent experiments (n = 3). (C) Confocal images of VIL1 + enterocytes, MUC2 + goblet cells in the organoids cultured in differentiation medium. Nuclei and cellular actin filaments are counterstained with DAPI (blue) and Phalloidin-647 (purple). VIL1: Villin 1; ALPI: Intestinal alkaline phosphatase; LYZ: Lysozyme; MUC2: Mucin 2; CHGA: Chromogranin A; LGR5: Leucine-rich repeat-containing GPCR5.(TIF)
S9 FigThe antiviral activity of Nirmatrelvir against SADS-CoV.(A) Chemical structure of Nirmatrelvir. (B) The Ki value of Nirmatrelvir against SADS-CoV M^pro^. (C) The kinact value of Nirmatrelvir against SADS-CoV M^pro^. Data points represent the mean of three independent replicates. (D–F) Antiviral activity of Nirmatrelvir measured by RT-qPCR in Huh7 (D), ST (E), and IPEC-J2 (F) cells. (G–I) WB analysis of viral N protein expression in Huh7 (G), ST (H), and IPEC-J2 (I) cells treated with Nirmatrelvir at post-infection. (J) EC_50_ determination of Nirmatrelvir against SADS-CoV in porcine small intestinal organoids. (K) Hydrogen bonds between Nirmatrelvir and Glu166 of SARS-CoV-2 M^pro^ (PDB code: 7RFS). Nirmatrelvir is shown in a red ball-and-stick model, with hydrogen bonds indicated by black dashed lines. (L) The side chain of Glu165 of SADS-CoV M^pro^ does not interact with 27h (PDB code: 9VUU). All experiments (B–J) are performed in three independent replicates (n = 3).(TIF)
S10 FigSuperposition analysis of 27h binding to HCoV-OC43 and HCoV-NL63 M^pro^s.(A) Superposition of HCoV-OC43 M^pro^ (orange, PDB code: 9C7W) and the SARS-CoV-2 M^pro^-27h (wheat, PDB code: 8I30) complex (RMSD = 0.9 Å for 269 Ca atoms). Residues involved in the inhibitor-interactions are labeled, with those of HCoV-OC43 M^pro^ shown in italics. Hydrogen bonds between 27h and SARS-CoV-2 M^pro^ are indicated by yellow dashed lines. (B) Superposition of HCoV-NL63 M^pro^ (cyan, PDB code: 7E6M) and the SADS-CoV M^pro^-27h (wheat, PDB code: 9VUU) complex (RMSD = 0.7 Å for 221 Ca atoms). Residues of HCoV-NL63 M^pro^ are shown in italics. Hydrogen bonds between 27h and SADS-CoV M^pro^ are indicated by yellow dashed lines. Water-mediated hydrogen bonds observed in the SADS-CoV M^pro^-27h complex are omitted as they represent indirect interactions between protease and 27h. (C) Atom numbering scheme of inhibitor 27h. (D) Table summarizing the hydrogen-bond interactions between 27h and the M^pro^s. Distances for HCoV-NL63 M^pro^ are listed as “NA” (not available) due to the flexibility of His41 and Thr47. The MatchMaker tool in UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax/docs/user/tools/matchmaker.html) was used for structural alignment.(TIF)
S1 TableInteractions between SADS-CoV M^pro^ and 27h as well as SY110.(DOCX)
S2 Tablekinact, Ki and kinact/Ki data of 27h against M^pro^s and their mutants ^a^.(DOCX)
S3 TableSubstrate-binding affinity of M^pro^ from eight CoVs ^a^.(DOCX)
S4 TablePrimers used in site-directed mutagenesis PCR for each M^pro^ mutant.(DOCX)
S5 TableData collection and structure refinement statistics.(DOCX)
S6 TablePrimer sequences of targeted genes used in RT-qPCR assays.(DOCX)
S7 TableThe components of expansion medium and differentiation medium.(DOCX)
S8 TablePrimer sequences of targeted genes used in RT-qPCR analysis in the porcine intestinal organoids experiments.(DOCX)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Ruiz-Aravena M, Mc Kee C, Gamble A, Lunn T, Morris A, Snedden CE, et al. Ecology, evolution and spillover of coronaviruses from bats. Nat Rev Microbiol. 2022;20(5):299–314. doi: 10.1038/s 41579-021-00652-2 34799704 PMC 8603903 · doi ↗ · pubmed ↗
- 2Sanche S, Lin YT, Xu C, Romero-Severson E, Hengartner N, Ke R. High Contagiousness and Rapid Spread of Severe Acute Respiratory Syndrome Coronavirus 2. Emerg Infect Dis. 2020;26(7):1470–7. doi: 10.3201/eid 2607.200282 32255761 PMC 7323562 · doi ↗ · pubmed ↗
- 3Fehr AR, Perlman S. Coronaviruses: an overview of their replication and pathogenesis. Methods Mol Biol. 2015;1282:1–23. doi: 10.1007/978-1-4939-2438-7_1 25720466 PMC 4369385 · doi ↗ · pubmed ↗
- 4Guo M, Tao W, Flavell RA, Zhu S. Potential intestinal infection and faecal-oral transmission of SARS-Co V-2. Nat Rev Gastroenterol Hepatol. 2021;18(4):269–83. doi: 10.1038/s 41575-021-00416-6 33589829 PMC 7883337 · doi ↗ · pubmed ↗
- 5Zubair AS, Mc Alpine LS, Gardin T, Farhadian S, Kuruvilla DE, Spudich S. Neuropathogenesis and Neurologic Manifestations of the Coronaviruses in the Age of Coronavirus Disease 2019: A Review. JAMA Neurol. 2020;77(8):1018–27. doi: 10.1001/jamaneurol.2020.2065 32469387 PMC 7484225 · doi ↗ · pubmed ↗
- 6Nyberg T, Ferguson NM, Nash SG, Webster HH, Flaxman S, Andrews N, et al. Comparative analysis of the risks of hospitalisation and death associated with SARS-Co V-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. Lancet. 2022;399(10332):1303–12. doi: 10.1016/S 0140-6736(22)00462-7 35305296 PMC 8926413 · doi ↗ · pubmed ↗
- 7Hu B, Guo H, Zhou P, Shi Z-L. Characteristics of SARS-Co V-2 and COVID-19. Nat Rev Microbiol. 2021;19(3):141–54. doi: 10.1038/s 41579-020-00459-7 33024307 PMC 7537588 · doi ↗ · pubmed ↗
- 8Liu DX, Liang JQ, Fung TS. Human Coronavirus-229E, -OC 43, -NL 63, and -HKU 1 (Coronaviridae). Encyclopedia of Virology. 2021. p. 428–40. 10.1016/B 978-0-12-809633-8.21501-X · doi ↗