The COVID-19 pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in approximately 7 million deaths by October 2024 [1], posing an ongoing threat to global health. While various strategies, such as vaccines and antiviral drugs (i.e., Paxlovid®), have mitigated severe outcomes, the rapid emergence of immune-evading variants or serotypes, as well as the limitations of virus-directed therapies (including the compromised efficacy against mutated variants), has spurred the necessity of therapeutic innovations. This challenge calls for host-targeted strategies that disrupt conserved viral dependence mechanisms, offering broad-spectrum protections against evolving coronaviruses, thereby mitigating future epidemics and pandemics [2].

Enveloped viruses, including SARS-CoV-2, rely on the host cell endoplasmic reticulum quality control (ERQC) machinery for glycoprotein folding and virion maturation (Fig. 1) [3,4]. The ER α-glucosidases I/II are central components of the ERQC machinery [5]. Wherein, ER α-Glu I is the gatekeeper of the ERQC pathway. It removes the terminal glucose residue of Glc3Man9GlcNAc2N-linked glycans attached to nascent glycoproteins. The product is in turn hydrolyzed by ER α-Glu II to cleave the second α-1,3-glucose residue of N-glycan and generate the trimmed glycoprotein containing Glc1Man9GlcNAc2, which is recognized and retained by ER chaperones calnexin and calreticulin to ensure correct folding and processing to the Golgi apparatus for secretion [6]. If misfolded, removal of a mannose residue marks the glycoprotein for degradation by the ER-associated degradation pathway [3]. Inhibition of host ER glucosidases I and II interfere with the correct folding of viral glycoprotein in ER and prevents the formation of infectious virions [7]. Genetic evidence further supports this target: two siblings carrying MOGS mutations (encoding α-glucosidases I) showed resistance to viral infections despite their severe hypogammaglobulinemia [8]. Crucially, viral glycoproteins exhibit heightened sensitivity to the ER α-glycosidase inhibition compared to the host proteins. Thus, there is a therapeutic window within which the virus's glycoprotein processing can be effectively suppressed without affecting the host normal protein processing [9]. Therefore, host ER α-glucosidases I/II are promising host-targeted therapeutic targets for the development of broad-spectrum antivirals against viral strains which depend on ERQC for infectivity.

1-Deoxynojirimycin (DNJ), a glucose-mimicking iminosugar from Morus alba (Fig. 2) [10], emerged as a prototypic α-glucosidase inhibitor with anti-viral potential [11,12]. While two DNJ-derived drugs, miglitol (Glyset™) and miglustat (Zavesca™), have been developed for the treatment of type 2 diabetes and type 1 Gaucher diseases, respectively [13,14], their antiviral utility has not been fully explored. Since Taylor and co-workers first disclosed that DNJ exerted its antiviral effect against Human cytomegalovirus (CMV), [15] several DNJ derivatives have been shown with broad-spectrum of in vitro antiviral activity against both DNA and RNA viruses, including hepatitis B and C viruses, Dengue and other flaviviruses, HIV and Ebola virus [16,17]. Moreover, DNJ derivatives have also been shown to inhibit DENV [18] and Japanese encephalitis virus [19] in mice and hepatitis virus [20] in woodchucks. Recently, a single dose of N-9′-methoxynonyl 1-deoxynojirimycin (UV-4), the best-studied inhibitors of α-Glu I and α-Glu II, prevented the death of mice infected with lethal doses of influenza or dengue virus, even when treatment was begun as late as 48 h post-infection [21]. Two DNJ derivatives UV-4 and NB-DNJ have been advanced to Phase 1 or Phase 1/2 clinical trials for the antiviral therapy [22]. Recently, a series of DNJ-valiolamine derivatives have been developed as α-glucosidase inhibitors for anti-SARS-CoV-2 therapy, which showed promising activity [23].

However, the development of iminosugars as antivirals is limited by their suboptimal efficacy. Two prototype iminosugars DNJ and NB-DNJ have 50 % effective concentration (EC50) values at millimolar level. DNJ derivative with longer alkylated N1-side chain (i.e., N1-nonyl DNJ, NN-DNJ)) seems dramatically improve antiviral efficacy, with EC50 values in low micromolar range, while also increased the cytotoxicity of the iminosugars. No candidate has achieved optimal balance between efficacy and safety.

Given our constant interests in the development of novel natural product-based antiviral agents [24], we herein attempted to bridge this gap through a Traditional Chinese Medicine (TCM) theory-inspired design strategy. Specifically, we designed and synthesized series of DNJ-flavonoid conjugates as novel α-glucosidase inhibitors targeting ERQC-dependent glycoprotein folding. Remarkably, DNJ-20 not only demonstrated remarkable inhibition activity against α-glucosidase and viral entry process, but also exerted potent activity against several SARS-CoV-2 PsV and variants, as well as HCoV-229E and HCoV-OC43, with EC50 values up to 1.49 μM, which is more potent than the benchmark α-glucosidase inhibitor UV-4. Besides, it had no observable cytotoxicity in HeLa-ACE2, HEK-293T and Beas-2B cells. This work not only validates ERQC as a druggable target, but also pioneers a TCM-guided approach to develop broad-spectrum antivirals with enhanced therapeutic indices.