English

• Coronaviruses are single-stranded, positive-sense RNA enveloped viruses that have posed a significant threat to human health over the past few decades, particularly severe acute respiratory syndrome coronavirus (SARS-CoV), Middle East respiratory syndrome coronavirus (MERS-CoV), and SARS-CoV-2. These viruses have caused widespread infections and fatalities, with profound impacts on global economies, social life, and public health systems. Due to their broad host range in natural settings and the consequent high potential for zoonotic spillover events, a thorough investigation of the common viral mechanisms and the identification of druggable targets for pan-coronavirus antiviral development are of utmost importance.

  The spike (S) protein of coronaviruses, a trimeric glycoprotein presented on the coronavirus surface, plays a critical role in host cell recognition and membrane fusion. Its central role in viral entry, coupled with the shared mechanisms observed among diverse coronaviruses, make it a key target for the development of broad-spectrum antiviral therapies with both therapeutic and preventive effects. The S protein is divided into two subunits, S1 and S2. In β-coronaviruses, including SARS-CoV-2, SARS-CoV, as well as in the α-coronavirus human coronavirus (HCoV)-NL63, the S1 subunit initiates membrane fusion by recognizing the angiotensin-converting enzyme 2 (ACE2) receptor. Current membrane fusion models (Fig. 1) suggest that after ACE2 binding, the S1 subunit dissociates, exposing the S2′ site, which is then cleaved by host cell surface proteases such as transmembrane protease serine 2 (TMPRSS2) or the lysosomal protease cathepsin L (CTSL). Subsequently, the compressed heptad repeat 1 (HR1) and central helix (CH) regions of S2 are released, forming a central helix that facilitates the insertion of the fusion peptide (FP) into the host cell membrane. The HR1 domain then exposes three hydrophobic grooves that bind HR2, forming a stable six-helix bundle (6-HB) that drives membrane fusion, enabling viral genome release into the host cell cytoplasm [1]. Although key structures of the S protein have been resolved, the transition from the ACE2-bound pre-fusion state to the late-stage fusion intermediate remains unclear. This gap hinders a full understanding of S protein dynamics and the discovery of broad-spectrum antiviral targets.

  Figure 1

  

  Figure 1.  Schematic representation of S protein-driven membrane fusion.

  DownLoad: Full-Size Img PowerPoint

  Driven by the identified knowledge gaps in S protein conformational transitions Lu et al. carried out in-depth research [2]. Using flow cytometry, they observed that transmembrane-expressed S protein, upon stimulation with the extracellular domain of ACE2, could bind to the HR1 antibody or its ligand EK1 (a broad-spectrum coronavirus peptide in the clinic) peptide without the S1 subunit dissociation, which suggested the existence of early fusion intermediate conformation (E-FIC), deviating from the traditional understanding of the fusion process. To capture and visualize this intermediate conformation, the researchers co-incubated the S protein with the extracellular domain of ACE2 and performed cryo-electron microscopy (cryo-EM) particle screening. After refinement, they successfully determined a high-resolution (3.45 Å) structure of the E-FIC. By comparing the three conformations of the S protein during fusion—RB3, E-FIC, and post-fusion—the researchers discovered that ACE2 binding induces a rotation of both the S1 and S2 subunits towards the cell membrane. This rotation brings the viral and host membranes closer together, facilitating the insertion of the fusion peptide into the host membrane.

  Structural analysis of the E-FIC provided a rationale for the design of dual-functional antiviral agents. The spatial proximity observed between the receptor-binding domain (RBD, target 1) of S1 and the HR1 (target 2) of S2 enables ACE2 to trigger the E-FIC conformation in a cell-free system, allowing the HR1 ligand to bind to HR1 faster than the viral HR2, thereby locking the S protein in the E-FIC state. Building on this, they selected the ACE2 protein, a clinically validated RBD binder, and the HR1-binding EK1 peptide as drug candidates. The two components were then linked via a G₄S (Gly-Gly-Gly-Gly-Ser) peptide linker to generate a series of dual-functional drug candidate. Enzyme-linked immunosorbent assay (ELISA) demonstrated that the protein with five linker repeats (AL5E, with a structure of ACE2-(G₄S)₅-EK1) exhibited optimal binding activity (ACE2: half maximal inhibitory concentration (IC₅₀) = 29.23 nmol/L; HR1: IC₅₀ = 299.5 nmol/L) (Fig. 2).

  Figure 2

  

  Figure 2.  Graphical representation of the discovery of the SARS-CoV-2 S-targeting drug candidate AL5E.

  DownLoad: Full-Size Img PowerPoint

  In the following antiviral assays, AL5E demonstrated potent inhibitory activity against 20 SARS-CoV-2 variants, with half maximal effective concentration (EC₅₀) values ranging from 0.03 nmol/L to 1.96 nmol/L. This represents a 26- to 623-fold improvement over ACE2 and a 636- to 33, 162-fold increase in potency compared to EK1. Additionally, AL5E exhibited broad-spectrum inhibition of ACE2-utilizing coronaviruses including SARS-CoV, HCoV-NL63, SARSr-CoVs WIV1 and Rs3367, with EC₅₀ values of 0.69–5.90 nmol/L, which is 13- to 70-fold more potent than ACE2 and 227- to 1298-fold more effective than EK1. These results indicate that AL5E inherits the broad-spectrum activity of its parent compounds while exhibiting significantly stronger inhibition, suggesting that its dual-targeting of the RBD and HR1 results in a synergistic antiviral effect. To validate whether AL5E locks into the E-FIC conformation to exert its virus-inactivating ability, the team purified virus particles treated with the antiviral protein and assessed their ability to inactivate the virus. Results showed that AL5E (EC₅₀ = 1.29–8.72 nmol/L) exhibited 21- to 754-fold higher inactivation activity against free viral particles from 20 SARS-CoV-2 variants compared to ACE2. In ACE2-using viruses, AL5E (EC₅₀ = 1.08–18.13 nmol/L) demonstrated 25- to 539-fold greater inactivation than ACE2, while EK1 showed no inactivation ability at a concentration of 10, 000 nmol/L. This fact further confirms the importance of targeting the E-FIC conformation for antiviral activity.

  To assess the in vivo efficacy of the dual-functional molecule, the researchers used mouse models infected with the Omicron BA.5.2 variant of SARS-CoV-2 and HCoV-NL63. In these models, inhaled AL5E significantly reduced viral load in BA.5.2-infected mice, showing a decrease of 2.3 and 3.0 logs in N and E gene copies, respectively, outperforming ACE2. Intriguingly, AL5E treatment led to fewer activated microglia in the brain compared to remdesivir (a Food and Drug Administration-approved antiviral drug for coronavirus disease 2019 treatment) pretreated virus, suggesting a potential reduction in neuroinflammation. In NL63-infected mice, AL5E demonstrated strong preventive and therapeutic effects, reducing viral load and alleviating lung damage. In conclusion, AL5E locks the S protein in its E-FIC conformation to exert potent virus-inactivating activity. It is a promising dual-function therapeutic candidate for both infection inhibition and virus inactivation, potentially offering a good start in that direction for early prevention and treatment strategies to mitigate virus-induced brain damage.

  A recent study in Cell Host & Microbe has significant implications for coronavirus research [3]. They found that SARS-CoV-2 S protein can persist in the brain's protective meninges and skull bone marrow for up to four years post-infection. This can trigger chronic inflammation and increase the risk of neurodegenerative diseases. This highlights the importance of research on coronaviruses S protein. The discovery of the E-FIC of S protein offers new insights into membrane fusion. Based on this, the dual-functional antiviral protein AL5E was designed. AL5E shows broad-spectrum inhibition and viral inactivation both in vitro and in vivo, holding great promise for preventing and treating ACE2-dependent coronavirus infections. It provides a new drug candidate and novel targets and strategies for developing antiviral drugs against coronaviruses, which is crucial for addressing the challenges and long-term impacts of coronaviruses on human health.

  Lu et al.'s work represents a paradigm shift in viral inactivation, leveraging the induction of conformational changes in viral proteins to effectively render the virus non-infectious. This mechanism, independent of the viral life cycle, shows significant promise for pre-exposure prophylaxis (PrEP). However, structural biology limitations hinder the elucidation of protein conformational changes, leaving a gap in the field. Integrating computational modeling and artificial intelligence (AI) offers a promising path forward. Molecular dynamics simulations, particularly long-timescale and coarse-grained models, are highly effective in predicting protein dynamics and identifying allosteric sites. Combined with AI-driven molecular generation, virtual screening, and de novo protein design, this approach could accelerate the discovery of novel viral inactivators.

  Beyond viral inactivators, the design of dual-functional antiviral molecules offers significant potential [4]. Similar strategies have been demonstrated in drug design across other fields, such as proteolysis-targeting chimeras (PROTAC) technology for protein degradation, antibody-drug conjugates for cancer, and the creation of dual-target drugs through linkers. These molecules, designed via careful medicinal chemistry, can achieve synergistic effects ("1 + 1 > 2"). It is also worth noting that Jiang et al. previously constructed dual-functional molecules by linking human immunodeficiency virus (HIV) gp120 and gp41 inhibitors to exert antiviral effects, but their study was limited by a lack of crystallographic guidance and an underdeveloped linker design [5]. Future work in the viral inactivation field could benefit from the development of "multifunctional combination molecules", particularly by integrating PROTAC linker strategies to enhance antiviral efficacy.

  Furthermore, the work of Lu et al. highlights the vast potential of interdisciplinary integration in structural-driven drug discovery. The development of antiviral drugs is crucial in the prevention and control of major infectious diseases caused by viruses. Currently, antiviral drug development primarily focuses on targeting viral non-structural proteins (e.g., proteases, polymerases) and structural proteins (e.g., envelope proteins). Discovering new drug targets based on existing targets and developing efficient, broad-spectrum antiviral drugs remains a core research direction in the field. The research paradigm of this study is worth emulating: capturing transient conformations using cryo-EM and extreme ultraviolet laser dissociation mass spectrometry to expand the range of potential drug targets. By combining these advanced techniques with strategies like virtual screening and protein degradation, this research provides a valuable blueprint for the development of next-generation antiviral agents.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Heng Gao: Writing – review & editing, Writing – original draft. Jiwei Zhang: Writing – review & editing. Peng Zhan: Writing – review & editing, Conceptualization. Xinyong Liu: Writing – review & editing.

  Acknowledgments

  The authors are supported by the Key Research and Development Program, Ministry of Science and Technology of the People's Republic of China (Nos. 2023YFC2606500, 2023YFE0206500).

  
  
• 1. [1]

     C.B. Jackson, M. Farzan, B. Chen, et al., Nat. Rev. Mol. Cell Biol. 23 (2022) 3–20. doi: 10.1038/s41580-021-00418-x

  2. [2]

     L. Xing, Z. Liu, X. Wang, et al., Cell 188 (2025) 1297–1314. e24.

  3. [3]

     Z. Rong, H. Mai, G. Ebert, et al., Cell Host Microbe 32 (2024) 2112–2130. e2110.

  4. [4]

     Q. Gong, J. Song, Y. Song, et al., Chin. Chem. Lett. 36 (2025) 110456.

  5. [5]

     J. Pu, Y. Dai, Q. Wang, et al., Front. Pharmacol. 11 (2021) 613361.

• 

  Figure 1  Schematic representation of S protein-driven membrane fusion.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Graphical representation of the discovery of the SARS-CoV-2 S-targeting drug candidate AL5E.

  下载: 全尺寸图片 幻灯片
•