Hitting Coronaviruses Where It Hurts: Antiviral Strategies Targeting Viral Proteins

Abstract

Human coronaviruses include both endemic respiratory pathogens and highly pathogenic viruses with pandemic potential, underscoring the need for antiviral agents with activity beyond single viral species. This review focuses on selected, representative antiviral inhibitors directed against viral proteins of human coronaviruses. Accordingly, the review examines representative inhibitors targeting viral proteins essential for coronavirus infection and propagation. The discussion includes inhibitors acting at different functional levels of the viral life cycle, such as entry, genome replication, RNA processing and particle assembly, and reflects the range of molecular entities reported to date. By organizing known inhibitors according to their viral protein targets, this review provides an integrated overview of current antiviral candidates and illustrates the central role of viral proteins as points of intervention. Overall, the review aims to consolidate existing knowledge on antiviral inhibitors across human coronaviruses, offering a coherent framework for appreciating their scope and diversity.

Viral proteins represent key vulnerabilities in human coronaviruses (HCoVs). This review outlines inhibitors targeting spike (S), main protease (Mpro), papain‐like protease (PLpro), and RNA‐dependent RNA polymerase (RdRp), mapping chemically diverse strategies that disrupt entry and replication, and provides an integrated view of current candidates, supporting the design of broad‐spectrum antiviral strategies.

1. Introduction

Historically, coronaviruses were primarily associated with mild upper respiratory tract infections in humans, as well as various pathologies in different animal species. However, the emergence of three highly pathogenic strains over the past two decades (SARS, MERS, and SARS‐Cov‐2) has highlighted the ability of coronaviruses to cause severe respiratory infections and their pandemic potential [1].

To date, seven human coronaviruses (HCoVs) have been identified exclusively within the genera alpha‐coronavirus and beta‐coronavirus, whereas other orthocoronavirus genera have so far only been detected in animal hosts (Figure 1) [2]. Four of them, HCoV‐229E and HCoV‐NL63 (alpha‐coronavirus) and HCoV‐OC43 and HCoV‐HKU1 (beta‐coronavirus), are primarily associated with mild upper respiratory tract infections and account for approximately 15%–30% of common colds [3]. In contrast, the remaining three HCoVs, all belonging to the genus beta‐coronavirus, have demonstrated the ability to cross species barriers and cause severe epidemic or pandemic disease in humans, as exemplified by SARS‐CoV, MERS‐CoV, and SARS‐CoV‐2 [4, 5, 6, 7]. Infections caused by these viruses present a broad clinical spectrum ranging from mild respiratory symptoms to severe pneumonia, which in the most critical cases may progress to acute respiratory distress syndrome (ARDS), multiorgan failure, and death [8]. Moreover, SARS‐CoV‐2 infection has been associated with systemic manifestations, including anosmia, fatigue, thrombotic events, and cardiovascular complications, highlighting a multisystemic pathogenesis [9]. Although coronaviruses infect a wide range of animal species, including members of the genera gamma‐coronavirus and delta‐coronavirus, as shown in Figure 1, this review focuses exclusively on those known to infect humans.

FIGURE 1.

Classification of human coronaviruses according to the International Committee on Taxonomy of Viruses.

Coronaviruses are capable of infecting a wide range of animal species, with bats recognized as natural reservoirs for a large diversity of coronavirus lineages. Their unique immune system enables persistent viral infection without severe pathology. Phylogenetic and genomic analyses indicate that several HCoVs, including SARS‐CoV, MERS‐CoV, and SARS‐CoV‐2, originated in bats [10, 11, 12]. Zoonotic transmission to humans typically involves an intermediate host. Civets (Paguma larvata) were identified as the intermediate host for SARS‐CoV [13], while dromedary camels (Camelus dromedarius) serve as the primary reservoir for MERS‐CoV [14]. For SARS‐CoV‐2, pangolins (Manis spp.) have been proposed as potential intermediate hosts based on sequence similarity, although this remains under investigation [15] (Figure 2).

FIGURE 2.

Schematic diagram of the structure of HCoVs. Created in https://BioRender.com.

HCoVs are mainly transmitted via respiratory droplets (Flügge droplets) generated during coughing, sneezing, speaking, or breathing, and can also spread through aerosols under certain conditions [16, 17]. Viral transmission via contaminated surfaces may occur, as coronaviruses can persist outside the host for extended periods depending on environmental factors [18]. The incubation period generally ranges from 2 to 14 days [19].

1.1. Structure of Coronaviruses

Coronaviruses (CoVs) are enveloped, positive‐sense single‐stranded RNA viruses belonging to the family Coronaviridae and the order Nidovirales. Within this family, coronaviruses are classified into four genera: alpha‐coronavirus, beta‐coronavirus, gamma‐coronavirus, and delta‐coronavirus [2] (Figure 1). A defining characteristic of coronaviruses is their unusually large RNA genome, which ranges from approximately 26–32 kb and represents one of the largest genomes among RNA viruses [20].

Coronaviruses are viral particles enveloped by a lipid bilayer derived from host cellular membranes. Embedded within this envelope are three viral structural proteins: the membrane protein (M), the envelope protein (E), and the spike glycoprotein (S) [20] (Figure 2). The spike glycoprotein forms prominent surface projections that confer the characteristic crown‐like appearance of coronaviruses, from which their name is derived, and mediate interactions with host cell receptors [20, 21]. Enclosed within the viral envelope is the nucleocapsid, composed of multiple copies of the nucleocapsid protein (N) complexed with the viral RNA genome [22, 23]. This ribonucleoprotein complex adopts a flexible, continuous “beads on a string” conformation that allows efficient accommodation of the long RNA genome within the confined space of the virion [24]. The nucleocapsid, together with the surrounding envelope and associated structural proteins, provides structural integrity and protects the viral genome during extracellular transmission [25].

Despite sharing a common architectural organization, HCoVs display structural variability, particularly among different genera and between mildly pathogenic and highly pathogenic strains. In coronaviruses associated with mild respiratory infections, such as HCoV‐229E and HCoV‐NL63 (alpha‐coronaviruses) and HCoV‐OC43 and HCoV‐HKU1 (beta‐coronaviruses), the overall virion structure and most structural proteins are relatively conserved. However, sequence variations in the spike protein are observed and contribute to differences in host range and cellular tropism [26]. In contrast, highly pathogenic coronaviruses, such as SARS‐CoV, MERS‐CoV, and SARS‐CoV‐2, belonging to the genus beta‐coronavirus, exhibit more pronounced structural differences, particularly within the spike protein and in the presence of additional accessory proteins [27]. In SARS‐CoV‐2, for example, the presence of a polybasic furin cleavage site within the spike protein (absent in other closely related coronaviruses) has been identified as a distinctive structural feature [27]. Variability in accessory proteins, such as ORF3b and ORF6, further contributes to structural and functional diversity among coronaviruses and has been implicated in differences in viral behavior and host interactions [28].

1.2. Replicative Cycle of HCoVs

The replicative cycle of HCoVs begins with the attachment of the viral particle to the host cell surface through interactions between the viral S glycoprotein and a specific cellular receptor (Figure 3). Different HCoVs employ distinct host receptors, reflecting differences in viral tropism. HCoV‐229E and HCoV‐NL63 recognize CD13 and angiotensin‐converting enzyme 2 (ACE2), respectively, whereas HCoV‐OC43 and HCoV‐HKU1 bind to sialic acid residues on the cell surface. Highly pathogenic coronaviruses such as SARS‐CoV and SARS‐CoV‐2 also use ACE2 as the primary receptor, while MERS‐CoV engages dipeptidyl peptidase‐4 (DPP4) [26, 29]. Following receptor engagement, conformational rearrangements of the spike protein enable fusion between the viral envelope and the host cell membrane or entry via endocytic pathways, depending on the virus and cell type [30]. After entry, the viral envelope is removed and the positive‐sense RNA genome is released into the cytoplasm.

FIGURE 3.

Schematic representation of the complete coronavirus replication cycle in human cells. Created in https://BioRender.com.

Once in the cytosol, the positive‐sense viral RNA genome directly functions as messenger RNA (mRNA) and is immediately translated by the host ribosomal machinery. This initial translation yields two large viral polyproteins, pp1a and pp1ab, which serve as precursors of the viral nonstructural proteins (NSPs). These polyproteins are proteolytically processed by virus‐encoded proteases, including the main protease 3CLpro (Mpro) and the papain‐like protease (PLpro), generating at least sixteen mature NSPs [31, 32].

A subset of these NSPs assembles to form the replication–transcription complex (RTC), which associates with host‐derived membranes, primarily originating from the endoplasmic reticulum, leading to the formation of characteristic double membrane vesicles (DMVs) [33]. These membrane‐associated replication compartments provide a protected environment in which viral RNA synthesis and processing take place. The RTC includes enzymatic activities required for RNA replication and transcription, such as RNA‐dependent RNA polymerase and associated RNA processing functions. Within the DMVs, viral RNA synthesis is initiated by the production of a complementary negative‐sense RNA intermediate from the incoming positive‐sense genome. This negative‐sense RNA serves as a template for the synthesis of new full‐length positive‐sense genomic RNA molecules, which are subsequently packaged into progeny virions, as well as for the generation of a set of subgenomic RNAs [34].

In parallel with genome replication, the RTC transcribes a series of smaller subgenomic RNAs from the negative‐sense RNA template. These subgenomic RNAs encode the structural proteins (S, E, M, and N) as well as accessory proteins, specific to each coronavirus. This transcription process is discontinuous, allowing the generation of multiple subgenomic RNAs of different lengths [35]. In contrast to continuous transcription, in which the RNA polymerase synthesizes a full‐length RNA copy without interruption, coronavirus transcription involves template switching events in which the polymerase skips specific genomic regions while copying defined segments. This discontinuous transcription strategy enables a single viral genome to produce a nested set of subgenomic RNAs, each directing the expression of a specific viral protein and facilitating coordinated regulation of viral gene expression [14].

The structural proteins S, M, and E are synthesized at the rough endoplasmic reticulum and are subsequently transported to the endoplasmic reticulum–Golgi intermediate compartment (ERGIC). Meanwhile, the nucleocapsid protein (N) is translated in the cytosol, where it associates with newly synthesized genomic RNA to form ribonucleoprotein complexes [36].

Assembly of new viral particles occurs within the ERGIC, where nucleocapsids are incorporated into membranes containing the viral structural proteins. Fully assembled virions are then transported in secretory vesicles through the Golgi apparatus and released from the host cell by exocytosis. This release of infectious particles enables subsequent rounds of infection within the host and facilitates transmission to new hosts [37].

2. Inhibitors of HCoV Replication

The sequential steps of the coronavirus replicative cycle offer multiple opportunities for therapeutic intervention, as viral entry, genome replication, transcription, and virion assembly rely on distinct viral and host factors.

This review does not aim to provide an exhaustive catalog of all compounds described to date. Instead, selected representative examples were chosen on the basis of their biological targets, relevance, and/or inhibition strategy.

2.1. Inhibition of Viral Attachment and Entry

Inhibition of viral attachment and entry represents the earliest opportunity to block coronavirus infection. This stage of the viral life cycle is mediated by the S glycoprotein, which is responsible for receptor recognition, activation by host proteases, and membrane fusion [38]. Because these events occur upstream of viral genome replication, entry inhibitors can prevent infection at its initial step and may offer prophylactic as well as therapeutic benefits across different HCoVs.

From a mechanistic standpoint, spike‐mediated entry can be inhibited through different intervention strategies. These include direct targeting of viral components, most notably the spike protein itself, as well as interference with host cellular factors required for viral attachment, activation, or internalization. Both virus‐directed and host‐directed approaches have been explored, each associated with specific advantages and limitations with respect to selectivity, resistance, and safety profiles.

Direct targeting of the spike protein offers high mechanistic specificity but presents notable challenges for small‐molecule drug discovery. In particular, the identification of low‐molecular‐weight compounds capable of directly inhibiting spike‐mediated receptor engagement is intrinsically difficult due to the extended and relatively flat nature of the Spike‐receptor protein–protein interface [39]. In contrast, host‐directed strategies often target proteins that fulfill essential physiological functions; therefore, their antiviral efficacy must be balanced against potential cytotoxicity, off‐target effects, and limited therapeutic windows. In addition, the contribution of individual host factors to viral entry may vary depending on cell type, tissue, and disease stage, which can complicate translation and clinical applicability [40].

2.1.1. Prevention of Receptor Binding

The initial step of coronavirus entry involves binding of the receptor‐binding domain (RBD) within the S1 subunit of the spike protein to a host‐cell receptor [30]. This interaction determines viral tropism and host specificity and represents a critical checkpoint in the infection process. Although the overall architecture of the spike protein is conserved, the RBD exhibits substantial sequence variability among HCoVs, influencing receptor usage and susceptibility to neutralization.

Inhibition of receptor binding can be achieved either by directly targeting the receptor‐binding domain (RBD) of the spike protein or by interfering with host receptors involved in viral attachment. Several monoclonal antibodies (mAbs), such as m336, m337, and m338, have been described as effective in inhibiting MERS‐CoV [41], while MAb201 showed activity against SARS‐CoV [42]. Additionally, peptides mimicking portions of cellular receptor sequences have been identified, which compete for binding to the RBD. Likewise, the protein griffithsin (GRFT) [43], isolated from the red alga Griffithsia, has been shown to block SARS‐CoV infection both in vitro and in vivo, reducing mortality in mouse models. This protein specifically binds to glycosylation sites on the S protein. More recently, GRFT‐based formulations have progressed to early clinical evaluation, further supporting the translational potential of this entry‐inhibition strategy [44]. Among clinically authorized mAbs against SARS‐CoV‐2, sotrovimab (Xevudy), approved in 2021, targets a highly conserved epitope within the RBD of the spike protein and demonstrated robust neutralizing activity against early Omicron subvariants, significantly reducing hospitalization and mortality in high‐risk outpatients; however, its clinical use was later restricted due to reduced activity against newer Omicron lineages [45, 46, 47].

Following an alternative strategy, small molecules capable of disrupting receptor engagement have also been explored; however, this approach remains considerably less developed than antibody‐based interventions. The identification of low‐molecular‐weight compounds that directly inhibit the Spike–receptor interaction is particularly challenging due to the extended and relatively flat nature of the protein–protein interface. To date, only a limited number of small molecules have been reported to directly interfere with Spike‐ACE2 binding, most of them remaining at an early “proof‐of‐concept” stage.

In this context, DRI‐C23041 [48] (Figure 4) represents a well‐characterized example of a rationally designed small‐molecule inhibitor of coronavirus receptor binding. Identified through structure‐guided optimization, this compound directly inhibits the interaction between the spike protein and the ACE2 receptor, displaying submicromolar to low‐micromolar activity in biochemical binding assays against SARS‐CoV‐2. Consistent with this mechanism, DRI‐C23041 also inhibits spike‐mediated viral entry in pseudovirus‐based cell assays with micromolar potency and a favorable selectivity profile. Importantly, inhibitory activity is retained against SARS‐CoV and the human α‐coronavirus HCoV‐NL63, both of which also utilize ACE2 for cellular entry, supporting the feasibility of targeting conserved features of the Spike‐ACE2 interface with small molecules. Although its antiviral potency remains in the micromolar range, DRI‐C23041 provides a “proof‐of‐concept” that direct prevention of receptor binding is achievable using low‐molecular‐weight compounds [48].

FIGURE 4.

Chemical structures of DRI‐C23041 and ASF‐006.

A related small‐molecule strategy has been described for compound ASF‐006 (Figure 4), a tetrapodal tryptophan‐based derivative that shows inhibitory activity against several SARS‐CoV‐2 Omicron subvariants [49]. This compound binds directly to the Omicron spike protein and inhibits spike‐mediated viral entry at submicromolar concentrations, while displaying little or no activity against the ancestral Wuhan‐Hu‐1 strain. Biophysical and structural analyses indicate that ASF‐006 stabilizes the spike trimer in a closed three‐RBD‐down conformation, thereby reducing RBD exposure and indirectly diminishing ACE2 engagement. Consistent with this mechanism, antiviral activity was observed in cell‐based assays and further supported in an in vivo murine model of Omicron infection [49].

2.1.2. Inhibition of Membrane Fusion and Endocytosis

After receptor binding, coronavirus entry requires activation of the spike glycoprotein and fusion between the viral and cellular membranes. This step is mediated by the S2 subunit of the spike protein, which undergoes conformational changes that lead to the formation of a six‐helix bundle (6‐HB) and drive membrane fusion [38]. The membrane fusion machinery located in the S2 subunit is highly conserved among both alpha and beta‐coronaviruses, in contrast to the variable receptor binding domain within the S1 subunit. This structural conservation underpins the rationale for targeting the S2 fusion machinery as a strategy to achieve broader antiviral coverage [50].

Several strategies have been explored to interfere with membrane fusion and associated entry processes. A series of repurposing efforts identified clinically used drugs that interfere with clathrin‐mediated endocytosis or endosomal acidification, thereby inhibiting coronavirus entry. Among these, chloroquine, chlorpromazine, and loperamide (Figure 5) showed low‐micromolar inhibition of MERS‐CoV, SARS‐CoV, and HCoV‐229E replication (EC₅₀ 2.5–8 μM) by blocking early endocytic steps or postentry processes [51]. However, these compounds display limited selectivity and are associated with adverse effects, including cardiotoxicity and neurotoxicity.

FIGURE 5.

Chemical structures membrane fusion and endocytosis.

The antiviral arbidol (umifenovir, Figure 5) was also reported to inhibit coronavirus entry [52]. In vitro studies showed reduced SARS‐CoV‐2 infection, and molecular simulations suggest that arbidol binds at the RBD‐ACE2 interface, increasing spike rigidity and preventing the conformational rearrangements required for membrane fusion [53]. A systematic review reported activity against SARS‐CoV, MERS‐CoV, and SARS‐CoV‐2, suggesting a potential broad‐spectrum profile [54]. However, despite early clinical use during the COVID‐19 pandemic, its efficacy has remained inconsistent.

Peptide‐based inhibitors targeting the viral fusion machinery have also shown potent antiviral activity. For HCoV‐229E, the HR‐derived peptides 229E‐HR1P and 229E‐HR2P (Figure 5) form a stable six‐helix bundle and inhibit spike‐mediated membrane fusion, showing submicromolar activity in cell‐based assays with minimal cytotoxicity [55]. These studies supported the feasibility of targeting the conserved fusion machinery of coronaviruses. Building on this strategy, the pan‐coronavirus fusion inhibitor EK1 (Figure 5), derived from the HR2 region of HCoV‐OC43/HCoV‐229E, was designed to bind the conserved HR1 domain and block 6‐HB formation. EK1 displays broad‐spectrum antiviral activity against multiple beta‐coronaviruses (SARS‐CoV, MERS‐CoV, and SARS‐CoV‐2) as well as selected alpha‐coronaviruses such as HCoV‐229E, reflecting the high conservation of the HR1 fusion core [56]. Optimization through cholesterol conjugation produced EK1C4 (Figure 5), a lipopeptide which displays markedly enhanced potency and robust inhibition across SARS‐CoV, MERS‐CoV, and seasonal coronaviruses, as well as live virus inhibition in vitro. Intranasal administration of EK1C4 protects mice from HCoV‐OC43 infection, highlighting its potential as a prophylactic respiratory antiviral [57].

In parallel, host‐directed approaches have aimed to inhibit cellular processes required for spike activation and entry. Targeting host proteases involved in spike activation has also been explored. The serine protease inhibitors camostat and nafamostat (Figure 5) effectively block TMPRSS2‐mediated cleavage of the S protein [58], thereby preventing membrane fusion at the cell surface and reducing viral entry. TMPRSS2 is highly expressed in human airway epithelial cells and participates in the activation of several HCoVs. Both compounds were subsequently evaluated in clinical studies for COVID‐19; however, clinical trials failed to demonstrate a consistent therapeutic benefit [59, 60].

2.2. Inhibition of Replication Complex

The coronavirus replicative machinery arises from the viral polyproteins pp1a and pp1ab, which are proteolytically processed into a set of nonstructural proteins (NSPs). These NSPs include essential enzymatic components (such as viral proteases and RNA‐processing factors) that coordinate genome replication and transcription. Because many of these proteins are structurally conserved across HCoVs, they represent attractive targets for broad‐spectrum antiviral inhibition [61].

2.2.1. PLpro (nsp3) as a Therapeutic Target

The PLpro is one of the two essential viral cysteine proteases required for coronavirus polyprotein maturation, acting in concert with the main protease (Mpro). PLpro is conserved across alpha‐ and beta‐coronaviruses, although sequence divergence outside the catalytic core influences substrate specificity and inhibitor binding [62]. In addition to its role in processing the viral polyproteins pp1a and pp1ab, PLpro removes ubiquitin and ISG15 modifications from host proteins, thereby suppressing innate immune signaling. Inhibition of PLpro therefore simultaneously disrupts viral replication and counteracts immune evasion, conferring a dual antiviral mechanism [61].

Early efforts to inhibit PLpro led to the identification of GRL‐0617 (Figure 6), a noncovalent small‐molecule inhibitor originally developed against SARS‐CoV PLpro [63] and subsequently shown to inhibit SARS‐CoV‐2 PLpro [64] with low‐micromolar potency in biochemical assays. GRL‐0617 binds to a pocket adjacent to the catalytic site, blocking substrate access and inhibiting both proteolytic and deISGylating activities. In cell‐based systems, GRL‐0617 reduces viral replication with limited cytotoxicity, and its cocrystal structures established the BL2 loop region as a druggable hotspot for PLpro inhibition [65].

FIGURE 6.

Structure of PLpro inhibitors.

A number of repurposed compounds have also been reported to inhibit PLpro activity. Thiopurine analogs such as 6‐thioguanine and 6‐mercaptopurine (Figure 6) reduce PLpro enzymatic activity in vitro, likely through interaction with the active site, and attenuate its deubiquitinating and deISGylating functions toward host substrates [66]. Similarly, disulfiram (Figure 6), a thiol‐reactive compound used clinically for the treatment of alcoholism, inhibits PLpro from SARS‐CoV and MERS‐CoV through covalent modification of catalytic cysteine residues, thereby blocking both deubiquitination and deISGylation (removal of the ISG15 posttranslational modification) [67]. However, these repurposing hits generally display modest potency and limited selectivity, and their broad reactivity with host cysteine‐containing proteins constrains their translational potential as antiviral therapeutics.

Despite the strong biological rationale for targeting PLpro, progress in the development of potent and selective small‐molecule inhibitors has historically lagged behind that of Mpro. This is largely attributable to the structural and functional resemblance of PLpro to human deubiquitinating enzymes, as well as to the extended and relatively shallow nature of its substrate recognition surfaces, which complicate the identification of drug‐like inhibitors with adequate selectivity and cellular activity [68, 69].

Building on the GRL‐0617 scaffold, recent structure‐guided optimization efforts have yielded more potent and selective PLpro inhibitors with improved drug‐like properties. The Pfizer‐developed compound PF‐07957472 (Figure 6) which exhibits nanomolar enzymatic inhibition, submicromolar antiviral activity in cell‐based assays, and efficacy in a murine model of SARS‐CoV‐2 infection, providing the first in vivo validation of PLpro as a therapeutic target [70]. Parallel strategies have also produced covalent PLpro inhibitors that incorporate tailored electrophilic warheads to selectively engage the catalytic Cys111 residue while maintaining high selectivity over human deubiquitinating enzymes [71]. Together, these advances establish PLpro as a viable antiviral target and underscore the potential of both reversible and covalent inhibition strategies for the development of next‐generation coronavirus therapeutics [70, 71].

2.2.2. RNA‐Dependent RNA‐Polymerase (RdRp, nsp12) y Helicase (nsp13)

The RdRp (nsp12) is the central enzymatic component of the coronavirus replication machinery, responsible for the synthesis of viral RNA with the assistance of several cofactors [72]. The RNA‐dependent RNA polymerase (nsp12) is one of the most conserved proteins across both alpha‐ and beta‐coronaviruses, underpinning the broad‐spectrum activity of nucleoside analogs [73, 74]. Together with the helicase nsp13 (which unwinds RNA structures and provides the necessary 5′ → 3′ helicase and NTPase activities) these proteins form the core of the RTC [75]. Both enzymes are highly conserved among HCoVs and are indispensable for genome replication, making them prime targets for antiviral drug development. Inhibition of nsp12 or nsp13 disrupts viral RNA synthesis at fundamental steps, offering strong potential for broad‐spectrum therapeutic strategies.

Remdesivir (Figure 7) has been reported to act as an inhibitor of the coronavirus RdRp (nsp12) [76]. This compound is a phosphoramidate prodrug capable of crossing the cellular membrane and release the corresponding monophosphate nucleotide. Following intracellular activation to its triphosphate form (the active metabolite), remdesivir triphosphate functions as an ATP analog and competes with the natural substrate (ATP) at the active site of nsp12. The compound is incorporated into nascent viral RNA strands, resulting in delayed chain termination during viral RNA replication [76]. Remdesivir, marketed under the trade name Veklury, is a broad‐spectrum antiviral agent developed by the biopharmaceutical company Gilead [77]. It has demonstrated activity against Ebola virus, SARS‐CoV, SARS‐CoV‐2 (in the nanomolar range), MERS‐CoV, as well as HCoV‐OC43 and HCoV‐229E [75]. Remdesivir was authorized for compassionate use during the COVID‐19 pandemic [78].

FIGURE 7.

Structures of nucleoside prodrugs targeting the viral RdRp.

Another nucleoside analog evaluated as a coronavirus RdRp inhibitor is favipiravir (Figure 7), a prodrug originally developed for the treatment of influenza virus infections [79]. Upon intracellular activation, favipiravir is converted into its ribonucleoside triphosphate form, which is recognized by nsp12 and incorporated into viral RNA. Rather than acting as a classical chain terminator, favipiravir induces error accumulation during RNA synthesis, leading to lethal mutagenesis of the viral genome [80]. Although favipiravir has shown antiviral activity against SARS‐CoV‐2 in vitro, its clinical utility has been limited by low plasma exposure, rapid clearance, and inefficient conversion to the active triphosphate metabolite, resulting in inconsistent antiviral efficacy in vivo [81].

Building on this mutagenesis‐based antiviral strategy, molnupiravir (Figure 7) was developed as an orally bioavailable nucleoside analog specifically optimized to enhance error induction during viral RNA replication [82]. Following intracellular activation to its triphosphate form (NHC‐TP), molnupiravir is efficiently incorporated by nsp12 and induces base‐pairing ambiguity rather than chain termination. Once incorporated into the nascent RNA strand, the compound gives rise to template‐directed misincorporation events in subsequent replication cycles, thereby amplifying mutational errors and driving the virus beyond its error threshold [83]. This optimized lethal‐mutagenesis mechanism explains the broad antiviral activity of molnupiravir against SARS‐CoV‐2 and other RNA viruses. Molnupiravir was authorized for the treatment of COVID‐19 in several countries, becoming one of the first orally available antiviral approved for this indication [84].

Another key protein in the coronavirus replicative cycle is nsp13. Although nsp13 is an RNA‐dependent helicase whose primary function is to resolve secondary structures of the viral RNA during replication, it also exhibits triphosphatase (TPase) activity within the same catalytic site [85]. The involvement of nsp13 in multiple stages of the viral replication cycle, together with its high degree of conservation across different coronavirus species, underscores its essential role in viral viability and highlights its potential as a therapeutic target. Consistent with this broad conservation, nsp13 is highly conserved across both alpha‐ and beta‐coronaviruses at the level of sequence and functional domains, supporting its classification as an attractive yet comparatively underexplored antiviral target [86, 87].

Accordingly, several efforts have focused on identifying small‐molecule inhibitors of nsp13 targeting its ATPase or helicase functions. Although no clinically approved compound currently acts directly on its catalytic ATPase/helicase site, several inhibitors have been identified that bind to or interfere with this highly conserved functional region. These include diketoacid derivatives initially reported to inhibit coronavirus helicase activity [88], as well as the repurposed drug IOWH‐032 (Figure 8), which suppresses both the ATPase and helicase activities of SARS‐CoV‐2 nsp13 by interfering with RNA binding [87]. Together, these observations support nsp13 as a complementary antiviral target to nsp12 within the RTC and rationalize ongoing efforts aimed at identifying and optimizing small‐molecule helicase inhibitors.

FIGURE 8.

Structure of Helicase inhibitors.

Beyond direct inhibition of the ATPase/helicase active site, several additional pockets within nsp13 have been explored as potential druggable regions. SSYA10‐001 (Figure 8) was identified as an inhibitor of the nsp13 helicase from MERS‐CoV and SARS‐CoV, exhibiting EC₅₀ values of approximately 25 μM and no apparent toxicity against either virus [89]. Docking studies based on this scaffold led to the identification of sixteen halogenated triazole derivatives predicted to interact with the MERS‐CoV nsp13 helicase [23].

Using a homology model of the SARS‐CoV‐2 nsp13 protein, a virtual screening campaign was conducted to identify novel inhibitors [90]. Among these, lumacaftor and nilotinib (Figure 8), emerged as top candidates, displaying the most favorable docking scores in both the apo model and the ATP‐bound form. These in silico predictions were subsequently validated through biochemical assays measuring inhibition of the ATPase activity of purified recombinant SARS‐CoV‐2 nsp13, where lumacaftor and nilotinib exhibited IC₅₀ values of 0.3 and 0.4 μM, respectively.

2.2.3. Mpro (3C‐Like Protease, nsp5) as a Therapeutic Target

During early stages of coronavirus replication, the main protease (Mpro or 3CLpro) mediates most polyprotein cleavage events required for the generation of functional NSPs. Mpro is a highly conserved cysteine protease essential for coronavirus replication, as it processes the bulk of cleavage sites within the viral polyproteins pp1a and pp1ab [91]. Notably, this conservation extends across both alpha‐ and beta‐coronaviruses, particularly in key residues of the active site involved in catalysis and substrate recognition, supporting the use of Mpro as a target for broad‐spectrum antiviral inhibitors [92]. Its functional importance is enhanced by the absence of human homologs, making it an especially attractive target for selective antiviral inhibition [93, 94, 95].

Mpro contains a catalytically active pocket formed by four subsites (S1′, S1, S2, and S4) that harbor the catalytic dyad Cys145‐His41 within a cleft between domains I and II [96, 97, 98, 99]. The thiol group of Cys145, positioned in the S1′ subsite, serves as the primary anchoring point for covalent inhibitors, a structural feature considered critical for the antiviral potency of Mpro‐targeting compounds [91].

The high degree of sequence and structural conservation of Mpro among alpha‐ and beta‐coronaviruses, particularly within the active site, has facilitated the rapid development of small‐molecule inhibitors targeting this protease.

2.2.3.1. Covalent Inhibitors of Mpro

A considerable number of SARS‐Cov and SARS‐Cov‐2 Mpro inhibitors rely on covalent mechanisms of action, typically involving nucleophilic attack of the catalytic Cys145 on an electrophilic warhead [97]. These compounds include peptidomimetics, aldehydes, α‐ketoamides, and other electrophilic scaffolds that form either irreversible or reversible covalent adducts within the substrate‐binding cleft [100, 93]. Owing to their ability to achieve high potency and prolonged target engagement, covalent inhibitors have played a central role in antiviral drug‐discovery efforts against coronaviral proteases [93].

The peptide N3 (Figure 9) was initially identified as an inhibitor of the Mpro enzymes from SARS‐CoV and MERS‐CoV, and it also displays antiviral activity against SARS‐CoV‐2 in Vero cells (EC₅₀ = 16.77 μM) [101]. As one of the earliest structurally characterized covalent inhibitors of coronavirus Mpro, peptide N3 provided a useful reference scaffold for subsequent inhibitor design. Crystallographic analysis of the Mpro‐N3 complex demonstrated that N3 acts as a Michael‐type covalent inhibitor, forming a bond between its vinyl group and the catalytic residue Cys145 [101]. Within the active site, N3 engages the S1′, S1, S2, and S4 subsites in a manner characteristic of peptidic substrate analogs. The γ‐lactam ring anchors the inhibitor in the S1 pocket, while the remaining residues occupy the hydrophobic subsites [101].

FIGURE 9.

Structure of covalent inhibitors of Mpro.

Pfizer identified the peptidic inhibitor PF00835231 (Figure 9) as a potent SARS‐CoV and SARS‐CoV‐2 Mpro inhibitor (EC₅₀ = 0.23 μM) [102]. Its phosphate prodrug PF07304814 (lufotrelvir, Figure 9) was developed to facilitate intravenous administration being efficiently converted into the active compound in vivo. However, the low oral bioavailability of PF00835231 prompted further structural optimization. Replacement of the hydroxymethyl ketone warhead by a nitrile group, together with additional modifications at the P1 and P2 positions, led to PF07321332 (Figure 9), a reversible covalent Mpro inhibitor with improved potency, solubility, and stereochemical stability. This compound was ultimately marketed as nirmatrelvir in combination with ritonavir, under the trade name Paxlovid [103].

As a result of the screening of a structurally diverse chemical library against multiple proteases, the dipeptidyl aldehyde prodrug GC376 (Figure 9) was identified as a potent inhibitor of multiple coronaviral Mpro enzymes [104, 105]. Upon release of the active aldehyde GC373, the inhibitor forms a reversible hemithioacetal with the catalytic residue Cys145. Its γ‐lactam moiety anchors the compound within the S1 subsite, while the hydrophobic P2 residue occupies the S2 pocket and the backbone aligns along the substrate binding cleft in a substrate mimetic fashion. GC376 exhibits broad‐spectrum potency, displaying submicromolar inhibition of SARS‐CoV‐2 Mpro (IC₅₀ ≈ 0.19 μM) as well as SARS‐CoVMpro (K_i ≈ 20 nM), together with a favorable toxicity profile [101, 102].

Additional covalent Mpro inhibitors include the α‐ketoamide series, represented by compound 1 (Figure 9) [100]. These inhibitors rely on an α‐keto warhead that undergoes nucleophilic attack by the catalytic residue Cys145 to form a reversible thiohemiketal within the active site. The crystal structures of the complexes show that the γ‐lactam moiety at P1 binds deeply in the S1 subsite, while a small hydrophobic group at P2 fits into the flexible S2 pocket and the backbone extends along the substrate binding cleft, adopting a substrate mimetic arrangement. Compound 1 exhibits inhibitory activity against several coronaviral proteases, with IC₅₀ values of 0.90 μM for the SARS‐CoV Mpro, 0.58 μM for the MERS‐CoV Mpro, and 0.67 μM for the SARS‐CoV‐2 Mpro [103].

Based on the catalytic mechanism of Mpro, in which Cys145 plays a central role in covalent inhibition, Dai et al. [93] designed the aldehyde‐containing inhibitors 2 and 3 (Figure 9). Both molecules incorporate an aldehyde warhead that forms a covalent adduct with Cys145 and an (S)‐γ lactam group to occupy the S1 subsite, adopting a substrate mimetic binding mode within the S1′, S1, S2, and S4 subsites. The compounds exhibit high potency against the SARS‐CoV‐2 Mpro, with IC₅₀ values of 0.053 μM for 2 and 0.040 μM for 3, and display antiviral activity in vitro and in a mouse model, together with low cytotoxicity and favorable pharmacokinetic properties [93].

Additional covalent inhibitors of the SARS‐CoV‐2 Mpro include the derivatives GRL1720 [106] and 4 [106] (Figure 9). GRL1720 is an indoline chloropyridinyl ester that acts as an irreversible covalent inhibitor through nucleophilic attack of Cys145 on its electrophilic ester warhead, whereas 4 contains an activated carbonyl group that forms a reversible covalent adduct with the catalytic cysteine. Both compounds efficiently block SARS‐CoV‐2 Mpro activity, with an IC₅₀ value of 0.32 μM for GRL1720, whereas 4 is a reversible tight‐binding covalent inhibitor with a K_i of 17.6 nM, consistent with its high affinity for the catalytic Cys145. Notably, 4 suppresses viral infectivity and cytopathicity without detectable toxicity even at 200 μM, and exhibits synergistic antiviral effects when combined with remdesivir, supporting the potential of dual target strategies combining Mpro and RdRp inhibition. This finding suggests that combination therapy employing inhibitors targeting distinct stages of the viral replication cycle, such as Mpro and RdRp, may represent a highly effective therapeutic strategy for COVID‐19.

Another class of covalent Mpro inhibitors is represented by selenium‐containing compounds such as ebselen (Figure 9) and related analogs. Ebselen is a clinically used organoselenium drug with known antiviral [107, 108, 109], antimicrobial [107, 110], and virucidal properties [107, 110], and displays low cytotoxicity. In a high‐throughput screening campaign, ebselen was identified as an inhibitor of the SARS‐CoV‐2 Mpro, exhibiting an IC₅₀ value of 0.67 μM and antiviral activity in Vero cells with an EC₅₀ of 4.67 μM [20]. Ebselen behaves as an irreversible inhibitor of both SARS‐CoV‐2 and SARS‐CoVMpro [101], forming a covalent selenosulfide bond with the catalytic residue Cys145. Subsequent optimization of the ebselen scaffold led to derivatives with markedly enhanced potency, with some analogs incorporating additional methoxy and ortho hydroxyl substituents on the phenyl ring achieving nanomolar inhibitory activity against SARS‐CoV‐2 Mpro [101].

2.2.3.2. Noncovalent Inhibitors of Mpro

In addition to covalent inhibitors, a diverse set of natural products and synthetic small molecules has been reported to inhibit the coronavirus Mpro through noncovalent interactions. These compounds encompass structurally unrelated scaffolds and display a wide range of binding modes within the protease, reflecting the plasticity of the Mpro active site and adjacent regions.

The indole diketopiperazine alkaloids neoechinulin A and echinulin (Figure 10), isolated from Aspergillus fumigatus MR2012 collected from the Red Sea, have been identified as noncovalent inhibitors of the SARS‐CoV‐2 Mpro [111]. Neoechinulin A exhibits submicromolar potency (IC₅₀ = 0.47 μM), whereas echinulin is less active (IC₅₀ = 3.90 μM) [111]. Docking and molecular dynamics studies indicate that both compounds bind within the active‐site cavity through their diketopiperazine scaffold, establishing hydrogen‐bond and hydrophobic interactions consistent with reversible, noncovalent inhibition [108]. Notably, neoechinulin A shows higher binding stability in steered molecular dynamics simulations, in agreement with its superior inhibitory potency [111].

FIGURE 10.

Structure of noncovalent inhibitors of Mpro.

Another natural product inhibitor is shikonin (Figure 10), a naphthoquinone pigment isolated from Alkanna tinctoria that acts as a noncovalent inhibitor of Mpro [22]. Shikonin inhibits the Mpro of multiple HCoVs‐including SARS‐CoV‐2, SARS‐CoV, MERS‐CoV, HCoV‐HKU1, HCoV‐NL63, and HCoV‐229E‐ with IC₅₀ values ranging from 1.57 to 16.91 μM, depending on the viral species [22]. Crystallographic analyses revealed that shikonin adopts a noncanonical orientation within the active site, inducing a rearrangement of the His41–Cys145 catalytic dyad and establishing aromatic stacking and hydrogen‐bond contacts with conserved residues [22]. This distinctive binding mode, observed across coronavirus proteases, positions shikonin as a promising scaffold for the development of broad‐spectrum, noncovalent Mpro inhibitors.

A further class of noncovalent inhibitors is represented by 9,10‐dihydrophenanthrene derivatives, among which compound C1 (Figure 10) emerged as the most potent member of the series [112]. C1 inhibits the Mpro of SARS‐CoV‐2 with an IC₅₀ of 1.55 μM and also displays activity against SARS‐CoV Mpro (IC₅₀ = 5.35 μM), behaving as a mixed‐type inhibitor of the protease [112]. Computational studies suggest that C1 can interact both with the active‐site region and with the dimerization interface of Mpro, potentially stabilizing alternative binding poses consistent with this mixed inhibition profile. In addition, C1 exhibits excellent gastrointestinal, plasma, and microsomal stability, supporting its potential as an orally available, drug‐like scaffold for the development of broad‐spectrum coronavirus Mpro inhibitors [112].

In addition to these chemical structures, several peptidomimetic and noncovalent scaffolds have been reported to inhibit Mpro through alternative binding modes. Among noncovalent peptidomimetic inhibitors, a series of N‐benzylpiperidine derivatives have been reported by the Camarasa group [113], consisting of peptidomimetic compounds based on an N‐benzylpiperidine scaffold (1,4,4‐trisubstituted piperidines) (Figure 10). Several derivatives inhibit the Mpro of SARS‐CoV‐2 and HCoV‐229E in the low‐micromolar range, representing a novel chemotype with antiviral potential. Among them, compound 5 (Figure 10) showed the most favorable antiviral activity against HCoV‐229E (EC₅₀ = 3.3 μM) without detectable cytotoxicity, whereas compound 6 (Figure 10) was the most potent inhibitor of SARS‐CoV‐2 Mpro (IC₅₀ = 15 μM). In silico docking and molecular dynamics simulations indicate that these inhibitors remain stably bound within the Mpro active‐site cavity, adopting comparable binding orientations, while their interaction profiles vary depending on the nature of the substituents located in the region that projects toward the S2 and S4 subsites of the protease.

A prominent noncovalent clinical inhibitor is Ensitrelvir (Figure 10), a nonpeptidic, orally available Mpro blocker that inhibits SARS‐CoV‐2 Mpro with submicromolar potency (IC₅₀ ≈ 0.05 μM) [114]. Crystallographic studies show a conserved binding mode across Mpro orthologs from SARS‐CoV‐2 variants and other HCoVs such as SARS‐CoV, MERS‐CoV, and HCoV‐NL63, although antiviral activity has only been demonstrated for SARS‐CoV‐2 [115]. Ensitrelvir also displays favorable pharmacokinetic properties and does not require ritonavir boosting. Marketed in Japan as Xocova, it received emergency approval in 2022 [116], full approval in March 2024 [117], and in 2026 became the first oral antiviral also approved for postexposure prophylaxis of COVID‐19 [118].

2.3. Inhibition of nsp14/15/16 and Structural Proteinski

Beyond the essential enzymatic components of the RTC, several additional NSPs (such as nsp14, nsp15, and nsp16) contribute to immune evasion, RNA proofreading, and viral RNA maturation, making them attractive antiviral targets [119]. In parallel, certain structural proteins, particularly the membrane (M) protein and the envelope (E) protein, play key roles in virion assembly, morphogenesis, and particle release [119, 120]. Although less commonly targeted than enzymatic NSPs, inhibition of these structural components can disrupt critical stages of the viral life cycle and impair the production of infectious virions [120].

2.3.1. Inhibitors of nsp14

Nsp14 is a multifunctional protein encoded by all coronaviruses, containing an N‐terminal 3′‐5′ exoribonuclease (ExoN) domain that is essential for high‐fidelity viral RNA synthesis and for maintaining resistance to nucleoside analogs, as well as a C‐terminal N7‐methyltransferase domain required for viral RNA processing [121]. Its ExoN activity is markedly enhanced through interaction with the cofactor nsp10, and this nsp14‐nsp10 interface is critical for efficient viral replication, RNA proofreading, and recombination [122]. Because these functions are indispensable for coronavirus replication, nsp14 has emerged as a key component of the replication machinery and a potential point of vulnerability for antiviral intervention.

S‐Adenosylmethionine (SAM)‐pocket inhibitors constitute the largest group of nsp14 N7‐methyltransferase inhibitors identified so far [123]. Among them, C10 (Figure 11) currently represents the most advanced candidate, being the only one with demonstrated in vitro, cellular, and in vivo efficacy [124]. C10 is a selective non‐nucleoside inhibitor of the SAM‐binding pocket of nsp14, showing potent inhibition of its N7‐methyltransferase activity while sparing the ExoN domain and human methyltransferases. It displays strong antiviral activity against SARS‐CoV‐2, including several variants, with EC₅₀ values in the low‐nanomolar to submicromolar range (Wuhan‐Hu‐1: 0.31 μM; Delta: 0.50 μM; Omicron: 64 nM), and also inhibits other beta‐coronaviruses nsp14 such as those of MERS‐CoV (IC₅₀ = 671.7 nM) and HCoV‐OC43 (IC₅₀ = 1.02 μM), while showing much weaker inhibition of the alpha‐coronavirus HCoV‐229E (IC₅₀ = 9.11 μM). Preliminary in vivo studies show that C10 reduces viral load and ameliorates lung pathology in SARS‐CoV‐2–infected K18‐hACE2 mice [124].

FIGURE 11.

Inhibitors of nsp14.

In addition to non‐nucleoside SAM‐pocket inhibitors, several nucleoside‐based scaffolds have also been explored as nsp14 N7‐methyltransferase inhibitors. Among these, adenosine‐derived analogs constitute a distinct class designed to mimic features of the natural SAM/SAH substrate while improving selectivity and drug‐like properties. HK370 (Figure 11) is a representative member of the adenosine‐5′‐carboxamide series developed as nsp14 N7‐methyltransferase inhibitors [125]. This nucleoside‐based analog incorporates a 5′‐carboxamide moiety that is essential for activity and shows markedly improved selectivity for nsp14 over human methyltransferases (RNMT). HK370 inhibits SARS‐CoV‐2 nsp14 with nanomolar potency (IC₅₀ = 31 nM) while showing no activity against RNMTat 25 μM, and displays moderate antiviral efficacy in Calu‐3 cells (EC₅₀ = 12 μM). HK370 also exhibits a favorable in vitro pharmacokinetic profile, showing stability in human and mouse plasma and liver microsomes, together with good transepithelial permeability in Caco‐2 monolayers. These features, together with the essential role of the 5′‐carboxamide moiety for activity, support HK370 as an optimized adenosine‐derived scaffold within this chemical class [125].

An alternative strategy has focused on the design of bisubstrate molecules that simultaneously engage the SAM‐binding pocket and the RNA‐cap site of nsp14. These compounds aim to increase binding affinity and selectivity by mimicking both the cofactor and the natural RNA substrate. A representative example is compound 7 (Figure 11), identified through a structure‐guided campaign and validated by multiple high‐resolution cocrystal structures with SARS‐CoV‐2 nsp14 [126]. These structural studies show that this derivative span both the SAM‐binding site and the RNA‐cap pocket, establishing interactions that stabilize the catalytic cavity and rationalize its bisubstrate behavior. Compound 7 showed nanomolar potency against SARS‐CoV‐2 nsp14 and a strong inhibition of SARS‐CoV as well as the alpha‐coronaviruses HCoV‐NL63 and HCoV‐229E, while retaining submicromolar activity against MERS‐CoV and sparing human RNMT and flaviviral MTases [126]. Although compound 7 shows favorable metabolic stability, passive permeability, and no cytotoxicity in HepG2 cells, its antiviral activity in infected cells remains modest, indicating that further optimization is required to improve cellular engagement.

Additional bisubstrate scaffolds have been developed by Debart et al. [127], based on sulfonamide‐linked SAM mimetics designed to bridge the SAM and RNA‐cap binding sites of nsp14. Among this series, compound 8 (Figure 11) emerged as the most potent representative, displaying low‐nanomolar biochemical inhibition and high selectivity over human and flaviviral methyltransferases. However, despite strong target engagement, its antiviral activity in infected cells remained modest.

2.3.2. Inhibitors of nsp15

Nsp15 is a conserved coronavirus endoribonuclease (EndoU) that cleaves viral RNA at uridylate residues and plays a key role in evading host innate immune responses [128]. Genetic inactivation of nsp15 results in attenuated viral replication and enhanced immune recognition, underscoring its importance for coronavirus fitness and pathogenesis. Together, these features position nsp15 as a modulator of host–virus interactions and a potential antiviral target.

The first functional evidence that nsp15 is a druggable antiviral target came from the identification of the guanidine‐substituted diphenylurea EPB‐113 (Figure 12) [129]. This compound suppresses the replication of HCoV‐229E and SARS‐CoV‐2 in cell culture at low micromolar concentrations via direct engagement of the nsp15 EndoU domain. Selection of mutant viruses under EPB‐113 pressure revealed substitutions at or near the conserved catalytic residue His250. Subsequent studies using reverse genetics demonstrated that introduction of the His250 mutation into a recombinant SARS‐CoV‐2 virus confers resistance to EPB‐113, thereby validating nsp15 as the molecular target of this compound [130]. Furthermore, mechanistic studies have shown that EPB‐113 does not act as a classical catalytic inhibitor but instead induces an altered EndoU state that interferes with viral RNA synthesis [129].

FIGURE 12.

Inhibitors of nsp15.

While EPB‐113 enabled functional validation of nsp15 as an antiviral target, tipiracil (Figure 12) demonstrated that the uridine‐binding pocket of nsp15 is druggable [129]. Tipiracil, an FDA‐approved uracil analog used clinically in oncology as part of combination chemotherapy regimens, binds to the uridine‐recognition site of nsp15 and inhibits EndoU activity in the low‐micromolar range, with only modest antiviral effects in cell‐based assays. Although not suitable for direct antiviral use, these findings support uracil‐based scaffolds as starting points for structure‐guided nsp15 inhibitor design.

A high‐throughput screening of a library of more than 100,000 compounds led to the identification of three additional nsp15 inhibitors: hexachlorophene, CID5675221, and IPA‐3 (Figure 12) [131]. Hexachlorophene acts as a poorly specific and highly promiscuous inhibitor, while CID5675221 displays a reversible mode of inhibition that has not been fully characterized. In contrast, the mechanism of IPA‐3 has been investigated in greater detail and corresponds to an irreversible inhibition, likely mediated by covalent modification of cysteine residues located near the catalytic center of nsp15. Although IPA‐3 provided valuable mechanistic insight by revealing a covalent mode of nsp15 inhibition, its intrinsic chemical reactivity and lack of selectivity toward other targets, such as PAK1, limit its relevance beyond that of a tool compound.

Together, these studies establish nsp15 as a valid antiviral target, although the development of potent and selective inhibitors remains at an early stage.

2.3.3. Inhibitors of nsp16

Nsp16 is a coronavirus‐encoded 2′‐O‐methyltransferase (2′‐O‐MTase) responsible for the methylation of the 5′‐cap structure of viral mRNA, a modification that is essential for efficient viral translation and evasion of host innate immune recognition. This enzymatic activity converts the viral RNA cap into a host‐like structure, preventing its detection by cytosolic RNA sensors. Nsp16 is catalytically active only upon association with its cofactor nsp10, which stabilizes the enzyme and promotes substrate binding. Because cap 2′‐O‐methylation is indispensable for coronavirus replication and immune evasion, nsp16 represents a well‐defined enzymatic step amenable to antiviral intervention. Accordingly, inhibition strategies have focused either on direct blockade of its 2′‐O‐MTase activity or on disruption of the nsp10–nsp16 interaction [132].

The use of S‐adenosyl‐L‐homocysteine (SAH) and the SAM analog sinefungin (Figure 13) facilitated the initial biochemical and structural characterization of nsp16 [133]. These ligands allowed to establish that nsp16 functions as a canonical SAM‐dependent methyltransferase and helped to define the architecture of its SAM‐binding pocket in the postcatalytic and inhibited states. In particular, they provided direct experimental evidence that the SAM pocket is preformed, conformationally stable, and accessible to small molecules.

FIGURE 13.

Inhibitors of nsp16.

Early attempts to exploit the SAM pocket for inhibition, focused on adenine‐based and adenosine‐derived scaffolds beyond classical SAM mimetics. Cocrystal structures revealed that noncanonical nucleoside analogs such as toyocamycin and sangivamycin [134] (Figure 13) bind within the SAM pocket in alternative orientations, highlighting a significant degree of structural plasticity of the catalytic site. Despite well‐defined binding modes, binding of these compounds did not translate into efficient enzymatic inhibition or consistent antiviral activity. Similar limitations were observed with non‐nucleoside chemotypes, including coumarin derivatives (e.g., ZINC12880820) [135] (Figure 13), which inhibited nsp16 2′‐O‐methyltransferase activity only at micromolar concentrations and without evident antiviral activity in cell‐based assays.

More recent efforts have focused on the rational design of small‐molecules targeting the nsp16 SAM‐binding pocket, yielding compounds with measurable enzymatic inhibition. In this context, the SAM‐competitive inhibitors SS148 and WZ16 (Figure 13), identified through cross‐screening of methyltransferase inhibitors, were shown to inhibit SARS‐CoV‐2 nsp16 2′‐O‐methyltransferase activity in biochemical assays at low‐micromolar concentrations and were cocrystallized with the nsp10–nsp16 complex [136]. Structural analyses confirmed binding of both compounds within the SAM pocket. Although their inhibitory potency remains modest, these studies provide a clear proof‐of‐principle that the nsp16 SAM pocket can be functionally targeted by rationally designed small molecules.

In addition to direct interaction at the nsp16 SAM‐binding site, disruption of the nsp10–nsp16 interface has been explored as an alternative strategy to interfere with coronavirus 2′‐O‐methyltransferase activity. A peptide derived from the nsp16‐interacting region of nsp10 (TP29) [137] (Figure 13) was shown to inhibit nsp16 activity in biochemical assays and to reduce viral replication in cell‐based models of beta‐coronaviruses, including a murine beta‐coronavirus model (mouse hepatitis virus, MHV) and a SARS‐CoV replicon system [137]. In vivo administration of TP29 in an MHV infection model further demonstrated a reduced viral replication and disease severity, providing experimental evidence that interference with the nsp10–nsp16 complex can reduce coronavirus replication in a mammalian host. In parallel, recent structure‐based virtual screening studies have identified small‐molecule inhibitors targeting either the nsp16 SAM‐binding pocket or the nsp10–nsp16 interface, such as compounds B1–B3 [138], (Figure 13) which exhibit moderate inhibition of nsp16 2′‐O‐methyltransferase activity (≈45%–55% at 50 µM) in biochemical assays. Although these compounds remain at an early, exploratory stage, they support to continue the investigation of the nsp10–nsp16 complex as a potential antiviral target.

2.3.4. Inhibitors of Nucleocapsid (N) Protein

The coronavirus nucleocapsid (N) protein is a major structural component that plays essential roles in viral RNA packaging, ribonucleoprotein (RNP) formation, and regulation of the replication‐transcription process. The protein contains an N‐terminal RNA‐binding domain (NTD) and a C‐terminal oligomerization domain (CTD), connected by intrinsically disordered regions that contribute to its functional versatility. Beyond its structural roles, N protein is involved in multiple interactions with viral and host factors, including components of the replicase complex and host innate immune pathways [139].

As a relatively conserved coronavirus protein with essential roles in genome packaging and replication, the N protein has attracted interest as a potential antiviral target. Despite its essential biological functions, the N protein has so far proven difficult to target pharmacologically. Structural and computational studies have suggested that specific regions of the N protein, in particular the NTD, may be amenable to small‐molecule targeting; however, these proposals largely remain at a conceptual or preliminary level and lack of robust experimental validation [140]. To date, no small‐molecule inhibitors have been conclusively demonstrated to directly and specifically target the N protein with a clear antiviral mechanism, underscoring the current limitations of nucleocapsid N protein as a druggable target [141].

2.3.5. Inhibitors of the M Protein

The coronavirus membrane (M) protein is the most abundant structural protein in the virion and functions as the central organizer of virus assembly and morphogenesis. M coordinates interactions with the spike (S), envelope (E), and nucleocapsid (N) proteins and plays a key role in shaping virion architecture. The protein is relatively conserved among coronaviruses, which is consistent with its essential structural role throughout the viral life cycle [142].

Although M is largely membrane‐associated, recent studies have demonstrated that its conformational dynamics are critical for productive virion assembly. Notably, small‐molecule modulation of M conformation has been shown to effectively block coronavirus assembly and transmission in vivo, establishing the M protein as a functionally validated, nonenzymatic antiviral target [143].

CIM‐834 (Figure 14) was identified through a large‐scale phenotypic antiviral screening followed by a medicinal‐chemistry optimization and target elucidation program. The compound represents the first small‐molecule inhibitor shown to directly target the coronavirus M protein [143]. The compound exhibits potent antiviral activity against SARS‐CoV‐2, including multiple variants of concern, with half‐maximal effective concentrations in the nanomolar range (≈80–110 nM), and retains activity against SARS‐CoV with submicromolar potency, while lacking activity against MERS‐CoV and several alpha‐coronaviruses. Mechanistic studies demonstrated that CIM‐834 acts at a late stage of the viral replication cycle, selectively inhibiting virion assembly without affecting viral RNA synthesis. Structural and biochemical analyses revealed that CIM‐834 binds to the short conformation of the M protein at a pocket located near the dimer interface and stabilizes this nonproductive state, thereby preventing the conformational transition required for M oligomerization and membrane curvature during virion formation [143]. Importantly, CIM‐834 displayed good antiviral efficacy in vivo, significantly reducing levels of infectious virus in the lung and preventing transmission in treated animals. These results characterized M‐protein‐mediated assembly inhibition as a viable antiviral strategy.

FIGURE 14.

Inhibitors of the M protein.

A second M‐protein‐targeting small molecule, JNJ‐9676 (Figure 14) [144], was identified as a potent inhibitor of SARS‐CoV‐2 replication. The compound exhibits nanomolar antiviral activity in vitro, with EC₅₀ values in the 14–30 nM range among SARS‐CoV‐2 variants, including Omicron lineages, it also retains activity against SARS‐CoV as well as other zoonotic SARS‐like coronaviruses. Time‐of‐addition and resistance‐selection studies [144] support a late mechanism of action, consistent with inhibition of virion assembly rather than viral RNA synthesis. Structural characterization by cryo‐electron microscopy revealed that upon binding the compound stabilizes the M protein dimer in a distinct conformational state through interactions within a transmembrane pocket involving both protomers. In vivo efficacy was demonstrated in Syrian hamster models. Treatment resulted in reduced levels of infectious virus and improved lung pathology in both pre‐exposure and postexposure settings [144].

2.3.6. Inhibitors of the E Protein

The coronavirus envelope (E) protein is a small integral membrane protein that localizes predominantly to the endoplasmic reticulum‐Golgi intermediate compartment and contributes to late stages of virion assembly and release [142, 143]. Although present at low abundance in mature virions, E protein has been implicated in coronavirus fitness and pathogenicity. The protein oligomerizes in host membranes to form an ion‐conducting channel. This activity has been associated with alterations in host cell homeostasis and inflammatory responses. These features have led to interest in E protein as a potential antiviral target, although pharmacological validation has so far been confined mainly by biochemical and cell‐based studies [145, 146].

Among E protein channel blockers, the hexamethylene amiloride (HMA, Figure 15) represents a widely used reference compound [147]. HMA was shown to inhibit ion‐channel activity of the E protein from several coronaviruses in planar lipid bilayer systems, including the alpha‐coronavirus HCoV‐229E and the beta‐coronavirus MHV, and to suppress replication of the corresponding viruses in cell‐based assays at low‐micromolar concentrations. Importantly, genetic deletion of E eliminated the antiviral effect of HMA; this provides direct evidence that E is the relevant target in these systems. Subsequent mechanistic studies [148] confirmed the direct binding of HMA to the E protein in lipid membranes and supported a channel‐blocking mode of action, although with limited selectivity and unfavorable pharmacological properties. These observations support the use of HMA and related acylguanidines as experimental modulators of E protein ion‐channel activity in different coronaviruses.

FIGURE 15.

Inhibitors of the E protein.

In addition to amiloride‐based blockers, bulky polyphenolic compounds such as proanthocyanidins (e.g. 8, Figure 15) have been identified as E‐channel inhibitors through a cell‐based high‐throughput screening approach [149]. These compounds were shown to suppress SARS‐CoV‐2 replication in cell‐based assays and to directly inhibit E protein ion‐channel activity in membrane reconstitution assays. Mechanistic analyses support a pore‐blocking mode of action, providing functional evidence for direct engagement of the E protein.

2.4. Clinical Development Status of Representative Coronavirus Inhibitors

The antiviral candidates described throughout this review span a broad spectrum of development stages. Some compounds have only demonstrated activity in biochemical or cellular assays, whereas others have advanced to clinical studies or regulatory approval. To facilitate comparison between these representative inhibitors, Table 1 summarizes their principal viral targets together with their current development status.

TABLE 1.

Current development status of representative coronavirus inhibitors.

Inhibitor (s)	Current development status	References
Target: Spike (receptor binding)
m336	In vivo studies	[150]
m337, m338	Cell‐based studies	[151]
MAb201	In vivo studies	[42]
Griffithsin (GRFT)	Clinical studies (Phase I)	[44, 152]
Sotrovimab	Approved	[45, 46, 47]
DRI‐C23041	Cell‐based studies	[48]
ASF‐006	In vivo studies	[48]
Target: Spike‐mediated fusion/entry
Chloroquine	Clinical studies (Phase III)	[153]
Chlorpromazine	Clinical studies (Phase III)	[154, 155]
Loperamide	Cell‐based studies	[51]
Arbidol (Umifenovir)	Clinical studies (Phase III)	[156]
229E‐HR1P, 229E‐HR2P	Cell‐based studies	[55]
EK1	Cell‐based studies	[56]
EK1C4	In vivo studies	[57]
Camostat	Clinical studies (Phase III)	[156]
Nafamostat	Clinical studies (Phase II/III)	[157]
Target: PLpro (nsp3)
GRL0617	Cell‐based studies	[63, 64, 65]
6‐Thioguanine, 6‐Mercaptopurine	Cell‐based studies	[66]
Disulfiram	Clinical studies (Phase II)	[158]
PF‐07957472	In vivo studies	[70]
Target: RdRp (nsp12)
Remdesivir	Approved	[76, 77, 78]
Favipiravir	Clinical studies (Phase III)	[159]
Molnupiravir	Approved	[160]
Target: Helicase (nsp13)
Diketoacid derivatives	Cell‐based studies	[88]
IOWH‐032	Cell‐based studies	[87]
SSYA10‐001	Cell‐based studies	[89]
Lumacaftor	Biochemical studies	[90]
Nilotinib	Biochemical studies	[90]
Target: Mpro (nsp5)
N3	Cell‐based studies	[101]
PF00835231, PF07304814	Clinical studies (Phase I)	[161]
Nirmatrelvir (PF07321332)	Approved	[103]
GC376/GC373	In vivo studies	[162]
α‐Ketoamide compound 1	Cell‐based studies	[100]
Compound 2	In vivo studies	[93]
Compound 3	In vivo studies	[93]
GRL1720	Cell‐based studies	[106]
Compound 4	Cell‐based studies	[106]
Ebselen	Clinical studies (Phase II)	[163, 164]
Neoechinulin A	Biochemical studies	[107]
Echinulin	Biochemical studies	[111]
Shikonin	Cell‐based studies	[22]
C1	Biochemical studies	[112]
Compound 5	Cell‐based studies	[113]
Compound 6	Biochemical studies	[113]
Ensitrelvir	Approved	[117, 118]
Target: nsp14
C10	In vivo studies	[124]
HK370	Cell‐based studies	[125]
Compound 7	Cell‐based studies	[126]
Compound 8	Cell‐based studies	[127]
Target: nsp15
EPB‐113	Cell‐based studies	[129, 130]
Tipiracil	Cell‐based studies	[129]
Hexachlorophene	Cell‐based studies	[131]
CID5675221	Cell‐based studies	[131]
IPA‐3	Cell‐based studies	[131]
Target: nsp16
SAH	Biochemical studies	[133]
Sinefungin	Biochemical studies	[133]
Toyocamycin	Biochemical studies	[130]
Sangivamycin	Biochemical studies	[134]
ZINC12880820	Biochemical studies	[135]
SS148	Biochemical studies	[136]
WZ16	Biochemical studies	[136]
TP29	In vivo studies	[137]
B1–B3	Biochemical studies	[138]
Target: M protein
CIM834	In vivo studies	[143]
JNJ9676	In vivo studies	[144]
Target: E protein
HMA	Cell‐based studies	[147, 148]
Proanthocyanidin derivative 8	Cell‐based studies	[149]

3. Summary and Outlook

In this review, we have presented representative antiviral strategies targeting different stages of the coronavirus life cycle, with particular emphasis on viral proteins that are conserved across multiple coronavirus genera. Most clinically advanced antiviral approaches continue to focus on enzymatic components of the RTC, such as the viral proteases and the RNA‐dependent RNA polymerase. In parallel, increasing experimental evidence indicates that additional viral proteins involved in particle assembly, morphogenesis, and release can also be pharmacologically modulated, although the level of validation varies substantially among targets.

In this context, the membrane (M) protein has recently emerged as a functionally validated antiviral target, supported by the identification of small‐molecule inhibitors with demonstrated antiviral activity in vivo. By contrast, other nonenzymatic structural proteins, such as the envelope (E) protein, remain at a more exploratory stage, with current evidence largely limited to biochemical and cell‐based studies using tool compounds.

Looking ahead, the development of coronavirus antivirals with activity across diverse viral species is expected to continue to rely on conserved enzymatic targets, for which robust biochemical, structural, and pharmacological frameworks are already established. At the same time, complementary exploration of nonenzymatic viral proteins involved in later stages of the viral life cycle may broaden the spectrum of tractable intervention points. Continued advances in structural biology, cell‐based screening, and integrated phenotypic approaches are expected to play a key role in clarifying which of these less conventional targets can be translated into viable antiviral strategies.

Funding

This work was supported by the Ministerio de Ciencia e Innovación (SAF2015‐64629‐C2‐1R and PID2019‐104070RB‐C21), Ministerio de Economía y Competitividad (SAF2009‐13914‐C02 and SAF2012‐39760‐C02), MICIU (PID2022‐136307OB‐C21), Comunidad de Madrid (PLATESA2‐CM S2018/BAA‐4370, BIPPED‐CM S‐BIO‐0214‐2006, and BIPPED‐2 S2010/BMD‐2457), Consejo Superior de Investigaciones Científicas (PIE 201280E093, PIE‐201980E028, and PIE201980E028).

Conflicts of Interest

The authors declare no conflicts of interest.

Biographies

Álvaro de la Cruz Potenciano holds a PhD in Medicinal Chemistry from Universidad Complutense de Madrid. His doctoral research focused on the design and synthesis of antiviral compounds, with particular emphasis on therapies targeting influenza viruses and coronaviruses. In recognition of his contributions to this field, he received the Ramón Madroñero Award in 2025 from the Spanish Society of Therapeutic Chemistry. He is currently a postdoctoral researcher at the Institute of Medicinal Chemistry and Biomedicine, and has gained international research experience through his stay at the Medicinal University of Graz.

Miguel Maldonado received his BSc in Chemistry from Universidad Complutense de Madrid and his Inter‐University MSc in Drug Discovery (UCM, UAH, CEU San Pablo). He is currently in the final year of his PhD at the Institute of Medicinal Chemistry (IQM‐CSIC), supervised by Prof. María‐José Camarasa and Dr. Sonia de Castro. His doctoral research centers on medicinal chemistry strategies, specifically the design and organic synthesis of novel antiparasitic agents and biomarker tools.

María‐José Camarasa, PhD in Chemistry (Complutense University of Madrid). Full Professor at IQM‐CSIC. She leads a renowned medicinal chemistry group focused on antiviral, antiparasitic, and anticancer drug discovery. Author of over 250 papers and inventor of 18 patents, she has coordinated major national, European, and industry projects. She has held key science‐management roles, serves on multiple editorial boards, and has received prestigious awards, including the Descartes Prize and recognition among the most cited CSIC scientists in the world's top 2%.

Sonia de Castro obtained her Ph.D. in Chemistry from the Universidad Autónoma de Madrid. She conducted postdoctoral research at the National Institutes of Health (USA) as a Fulbright fellow and later joined the Instituto de Química Médica (IQM‐CSIC), where she became Tenured Scientist in 2023. Over the last 14 years, her work has focused on the design, synthesis, and evaluation of heterocycles, peptides, peptidomimetics, and PNAs as pharmacological tools or drugs to treat potentially fatal diseases, such as AIDS, influenza, or coronavirus.

Dedicated to the memory of Dr. Sonsoles Velázquez.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.