The multiple roles of viral 3D^pol protein in picornavirus infections

ABSTRACT

The Picornaviridae are a large group of positive-sense, single-stranded RNA viruses, and most research has focused on the Enterovirus genus, given they present a severe health risk to humans. Other picornaviruses, such as foot-and-mouth disease virus (FMDV) and senecavirus A (SVA), affect agricultural production with high animal mortality to cause huge economic losses. The 3D^pol protein of picornaviruses is widely known to be used for genome replication; however, a growing number of studies have demonstrated its non-polymerase roles, including modulation of host cell biological processes, viral replication complex assembly and localization, autophagy, and innate immune responses. Currently, there is no effective vaccine to control picornavirus diseases widely, and clinical therapeutic strategies have limited efficiency in combating infections. Many efforts have been made to develop different types of drugs to prohibit virus survival; the most important target for drug development is the virus polymerase, a necessary element for virus replication. For picornaviruses, there are also active efforts in targeted 3D^pol drug development. This paper reviews the interaction of 3D^pol proteins with the host and the progress of drug development targeting 3D^pol.

KEYWORDS:

• Picornaviruses
• 3D^pol
• RNA-dependent RNA polymerase
• virus-host interaction
• RdRp inhibitor

Introduction

The Picornaviridae family comprises 158 species organized into 68 genera (https://www.picornaviridae.com). Human enteroviruses have been classified into four categories: A, B, C, and D, based on their molecular and biological properties. Nearly 100 subtypes have been found such as enterovirus A71 (EV-A71), enterovirus D68 (EV-D68), coxsackievirus A6 (CVA6), coxsackievirus B3 (CVB3), echovirus, poliovirus (PV), etc [1,2]. Although poliovirus has been successfully eradicated in most countries because of effective vaccination programs, other members of the Picornaviridae family still pose a severe threat to human health, especially EV-A71, coxsackievirus, and hepatitis A virus (HAV) [3–7]. Due to the large number of enterovirus serotypes, the cross-protective efficacy of current vaccines appear to be insufficient in the face of simultaneous transmission of multiple serotypes [8]. In addition, several picornaviruses are significant threats to the farming industry. Foot-and-mouth disease virus (FMDV) infection causes vesicular disease in cloven-hoofed animals, with high mortality rates in young animals [9]. Senecavirus A (SVA), another agent of vesicular disease in swine, has indistinguishable clinical symptoms with illness caused by FMDV [10]. Both viruses generate substantial economic losses and remind us of the emerging risks that must be taken seriously in farmed livestock.

The genomes of the Picornaviridae family have similar backbones, with a positive-stranded RNA genome between 6.7 kb and 10.1 kb in size and a single open reading frame (ORF) flanked by untranslated regions (UTRs) [11,12]. As shown in Figure 1, the 5” end of the genome is linked to VPg, which can be uridylated by the 3D^pol protein to act as a primer for transcription [1,13]. The 5”-UTR contains the internal ribosome entry site (IRES), which directs the cap-independent initiation of protein synthesis [14]. The 3’-UTR has a stem-loop structure and polyA tail, which is essential for picornaviruses [15]. All Picornaviridae members have a single ORF, which translates a polyprotein that is cleaved by the viral proteases 2A^pro, 3C^pro, and 3 CD^pro into the precursor proteins P1, P2, and P3 [14,16]. The P1 precursor protein is then cleaved into four structural proteins, which we have designated VP1, VP2, VP3, and VP4; the P2 precursor protein is processed into nonstructural proteins 2A^pro, 2B, and 2C; and the P3 precursor protein undergoes proteolytic cleavage into nonstructural proteins 3A, 3B, 3C^pro, and 3D^pol [12]. Some picornaviruses, such as aphthoviruses, cardioviruses, and kobuviruses, additionally express a leader protein (L^pro) at the N-terminal end of the polyprotein, after which the polyprotein is cleaved to form the L^pro, P1, P2, and P3 precursor proteins [17]. In addition, the coding region of the picornavirus genome has an essential element, the internal cis-acting replication element (cre), which serves as a template for VPg uridylation [18,19].

Figure 1. A schematic representation of the EV-A71 genome and current understanding of the picornaviruses 3D^pro protein. The enterovirus 5” end is linked to a VPg followed by an IRES. After translation of the single OFR, the polyprotein is cascade cleaved to form VP1, VP2, VP3, VP4, 2A^pro, 2B, 2C, 3A, 3B, 3C^pro, and 3D^pol. at the 3” end of the EV-A71 genome is a polyA sequence. 3D^pol proteins catalyse viral RNA synthesis, catalyse VPg uridylation, interact with host proteins, and act as drug targets.

Over the past few years, we have gained a deeper understanding of members of the Picornaviridae family. The picornavirus 3D^pro has been recognized as a critical component for viral genome replication, but other functions are progressively being identified. Here, we review the recent advances in understanding the biological functions of picornavirus 3D^pro during virus infection and briefly summarize the current strategies and progress of drug design and development targeting picornavirus 3D^pro.

Characterization of the picornavirus 3D^pol

The picornavirus RNA-dependent RNA polymerase (RdRp), also known as 3D^pol, is located at the C-terminus of the viral polyprotein. Only following a series of cleavages from the precursor protein does 3D^pol exhibit a polymerase activity [20]. Several picornavirus 3D^pol structures have been solved, which have shown highly conserved structural and functional elements [20–23](Figure 2). RdRp is the critical protein responsible for viral replication; the overall structure of all RdRps resembles a cupped right hand with palm, finger, and thumb domains. The finger and thumb contact each other to form the NTP entry channel, and becomes catalytic by a subtle conformational change in the palm domain, which cooperates to perform highly specialized functions including nucleotide recognition, catalysis of nucleotide binding, and initiation of RNA replication [24,25]. RdRp also has VPg uridylation activity; following nucleotidylation by RdRp, uridylated VPg can act as a primer for virus genome replication [26–28](Figure 1).

Figure 2. Structures of the RdRps of picornaviruses. a: EV-A71 3D^pol protein (PDB: 3N6L); b: EV-D68 3D^pol protein (PDB: 6L4R); c: CVA16 3D^pol protein (PDB: 5Y6Z); d: CVB3 3D^pol protein (PDB: 3DDK); e: PV 3D^pol protein (PDB: 4K4S); and f: FMDV 3D^pol protein (PDB: 6S2L). The palm, thumb and finger subdomains are shown in blue, cyan and magenta, respectively.

Advances in picornavirus 3D^pol-host interactions

As the structures of 3D^pol and RNA complexes have been solved, the mechanism of picornavirus RNA synthesis, in which 3D^pol functions as a polymerase, has been well explored. Given that 3D^pol is critical for virus replication, it has been suggested that 3D^pol-host interactions affect the formation of the 3D^pol-RNA replication complex. On the other hand, 3D^pol-host interactions enrich our knowledge of 3D^pol functions beyond its polymerase role and provide insights about antiviral strategies targeting protein-protein interaction sites. Here, we summarize the advances in 3D^pol-host interactions in enterovirus, FMDV, and SVA (Figure 3).

Figure 3. Diverse functions of the picornavirus 3D^pol protein and interactions with multiple host factors.

Legend: 1) Host factors promote sumoylation and ubiquitination modifications of the 3D^pol. 2) Host factors inhibit 3D^pol acetylation and RdRp activity. 3) Host factors regulate virus and host translation. 4) Host factor promotes viral RNA replication. 5) 3D^pol blocks host pre-mRNA splicing. 6) 3D^pol and host factors enhance assembly of the viral replication complex. 7) 3D^pol activates the inflammatory response. 8) 3D^pol is involved in the regulation of type I interferon response. The display in the figure does not mean a direct interaction. Dashed indicator marks represent the presumed function.

Enterovirus

Enteroviruses are a group of RNA viruses belonging to the family Picornaviridae. Their 3D^pol proteins perform diverse and complex functions in their interactions with host proteins.

Ubiquitination and SUMOylation are widely studied post-translational modifications (PTMs) that are crucial in various biological processes [29–31]. The protein Ubc9 can transfer SUMO to target proteins [32]. EV-A71 encoded 3D^pol can be SUMOylated and K63-linked ubiquitinated in a SUMO-dependent manner through interacting with Ubc9, which contributes to 3D^pol intracellular stability and enhances viral replication [33]. The host factor methyltransferase METTL3 interacts with the 3D^pol of EV-A71 and regulates SUMOylation and ubiquitination of 3D^pol [34]. The nuclear NAD-dependent deacetylase Sirtuin 1 (SIRT1) protein deacetylates many substrates, such as p53, and Notch1 (NICD) [35–37]. EV-A71 infection induces upregulation of SIRT1 expression and translocation from the nucleus to the cytoplasm, where SIRT1 interacts with the EV-A71 3D^pol pinky finger domain to reverse 3D^pol acetylation and inhibit RdRp activity [38]. In addition to this, EV-A71 3D^pol interacts with several cellular proteins to regulate cellular transcription-translation processes during infection. Picornavirus genome synthesis occurs in the cytoplasm, whereas 3D^pol has a nuclear localization signal that enters the nucleus and targets the pre-mRNA processing factor 8 (Prp8) to block host pre-mRNA splicing and mRNA synthesis [39]. EV-A71 3D^pol can interact with the ribosomal protein S6 (RPS6) to promote IRES-dependent and cap-dependent translation, positively regulating viral and host translation [40].

Picornaviruses relocate cellular regulators associated with membrane metabolism and modulate their activity to rearrange intracellular membranes to form a genome replication scaffold. These are also called replication organelles (ROs) [41,42]. Enteroviruses 3A and 3D^pol recruit phosphatidylinositol 4-kinase III (PI4KB) to secretory organelle membranes to promote RO formation [43–45]. In addition, PI4KB catalyzes the production of phosphatidylinositol-4-phosphate (PI4P), and PI4P binding with 3D^pol promotes 3D^pol binding to RO membranes and regulates RdRp activity [45]. Annexin A2 (ANXA2), a member of the annexin family, positively regulates EV-A71 replication. ANXA2 promotes PI4KB and 3D^pol interactions and forms an ANXA2-PI4KB-3D^pol complex localized to ROs to promote the formation of the EV-A71 RNA replication complex [46]. Recent studies found that EV-A71 likely utilizes the Ragulator-Rag to regulate cell death in order to facilitate its replication. Ragulator is a heteropentameric protein complex composed of LAMTOR1, LAMTOR2, LAMTOR3, LAMTOR4, and LAMTOR5, with LAMTOR1 anchoring the Ragulator to the lysosomal membrane [47]. Rag GTPase comprises of heterodimers RagA/B and RagC/D interacting with the Ragulator [48]. The mechanism of the regulation of enterovirus replication by Ragulator-Rag is complex. Knockout or knockdown of Ragulator components inhibits EV-A71 infection-induced apoptosis and pyroptosis. Interestingly, EV-A71 replication requires the Ragulator-Rag complex, as genomic replication and viral titres are significantly reduced in LAMTOR1-KO, LAMTOR3-KO, and RagB-KO cells [49]. For 3D^pol functions in ROs, previous studies showed the regulation of the PI4KB-3D^pol complex in ROs [46], while further studies from another group indicated that 3D^pol directly interacts with RagB, facilitating PI4KB and 3D^pol localization at the lysosome [49]. The complexes on the lysosomal membrane formed by Ragulator, Rag, PI4KB, and 3D^pol through protein-protein interactions help virus replication, as breaking the localization of the complexes in the lysosome reduces EV-A71 production [49]. Based on the Ragulator-Rag-3D^pol axis, the inhibitor ZHSI-1 was screened to target 3D^pol and significantly inhibit the localization of 3D^pol and PI4KB on the lysosomal surface to limit enterovirus replication in vitro and in vivo [49]. However, it is uncertain whether the Ragulator, Rag, PI4KB, and 3D^pol complexes are translocated into ROs upon EV-A71 infection-induced rearrangement of intracellular membranes. Ragulator, Rag, PI4KB, and 3D^pol complexes in lysosomal aggregates may act as viral replication sites or promote the formation of ROs, which requires further study. Additionally, the endoplasmic reticulum (ER) protein UDP-glucose glycoprotein glucosyltransferase 1 (UGGT1) is a crucial regulator of the unfolded protein response and has been shown to promote the formation of EV-A71 replication complexes. Immunoprecipitation analysis revealed that enterovirus nonstructural protein 3A acts as a bridge that allows UGGT1 and 3D^pol to establish indirect interactions to form replication complexes on RO membranes. Although UGGT1 does not interact directly with 3D^pol, its facilitation of 3A-3D^pol interactions enhances viral RNA synthesis and positively regulates the replication of EV-A71 and EV-D68 [50]. It is clear that 3D^pol localization of ROs is essential for RNA virus replication and that picornaviruses achieve this by hijacking host proteins or utilizing self-encoded proteins through a complex protein-protein interaction process. Blocking the interplay between 3D^pol and host factors also provides new ideas for drug development.

Positive-sense RNA viruses can utilize the membrane structure of autophagosomes to promote viral replication [42]. Recent studies have identified a strong interaction between EV-A71 3D^pol and autophagy. Beclin1 is an autophagy protein that regulates innate immune pathways [51,52]. RNA interference and overexpression experiments showed that the expression level of Beclin1 and EV-A71 replication were positively correlated [53], suggesting that EV-A71 infection can induce autophagy to enhance viral replication [54,55]. Previous studies showed that Beclin1 promotes the formation of double membrane structures [56,57], while the mechanisms of 3D^pol-Beclin1 interactions that promote EV-A71 replication need to be better understood. Further studies revealed that two structural domains of Beclin 1, the coiled-coiled domain (CCD) and evolutionarily conserved domain (ECD), are responsible for interacting with 3D^pol and are essential for facilitating EV-A71 replication [53]. This interaction may contribute to the formation of ROs and enable 3D^pol localization to the membrane and/or participate in regulating innate immunity to promote viral replication. Additionally, peroxisomes are dynamic organelles that play an essential role in multiple metabolic pathways and redox homoeostasis [58,59]. EV-A71 3D^pol interacts with peroxisomal protein acyl-CoA oxidase 1 (ACOX1) and degrades ACOX1 via the proteasome pathway. Moreover, infection with EV-A71 or knockdown of ACOX1 reduced the number of peroxisomes, ultimately inducing apoptosis and autophagy [60].

Inflammatory cytokines and interferons are essential for fighting microbial infections [61–63]. EV-A71 activates interleukin-1beta (IL-1β) by regulating the NLR family PYRIN domain containing-3 (NLRP3) inflammasome, and the RdRp activity of EV-A71 3D^pol is critical for EV-A71-induced NLRP3 inflammasome activation. Further studies revealed that 3D^pol interacts with NLRP3 to form a 3D^pol-NLRP3-ASC ring-like structure that promotes the assembly of the inflammasome complex [64]. In addition, 3D^pol is involved in the regulation of type I interferon responses. Melanoma differentiation-associated protein 5 (MDA5) is a crucial pathogen recognition receptor that initiates type I interferon transcription [65,66]. 3D^pol of EV-A71 interacts with the caspase activation and recruitment domains (CARDs) of MDA5 to inhibit MDA5-mediated type I interferon activation signalling. Similarly, the coxsackievirus B3 3D^pol interacts with MDA5 to inhibit type I interferon activation [67]. The 3D^pol of EV-D68 interacts with PGAM family member 5 (PGMA5), resulting in the downregulation of PGMA5 and upregulation of mitofusin 2 (MFN2) expression. Furthermore, MFN2 inhibits the activation of the RIG-I-like receptor signal pathway to suppress interferon beta (IFN-β) production, thereby promoting viral replication [68]. EV-A71 infection induces upregulation of the expression of the interferon-stimulating factor shiftless antiviral inhibitor of ribosomal frameshifting gene (SHFL), and EV-A71 3D^pol interacts with SHFL after 3D^pol is degraded via the ubiquitin-proteasome pathway [69]. It is apparent that enterovirus 3D^pol inhibits the type I interferon antiviral response by different mechanisms.

FMDV and SVA

FMDV and SVA are severe threats to agricultural production. Both viruses belong to the family Picornaviridae and are similar in many respects, including viral invasion and replication. Therefore, studies on their interactions with the host are of interest to sustainable agriculture and human health. Currently, several studies are dedicated to revealing the interactions of FMDV and SVA 3D^pol with the host.

As described above, EV-A71 3D^pol SUMOylation and ubiquitination modifications improve 3D^pol stability in cells [33,34]. A recent study found that overexpression of ubiquitin-conjugating enzyme E2 L6 (UBE2L6) promotes SVA replication, and conversely, knockout of UBE2L6 inhibits SVA replication; further studies showed that UBE2L6 interacts with 3D^pol to mediate 3D^pol ubiquitination and inhibit 3D^pol degradation [70]. The host nuclear protein 68 kDa Src-associated substrate during mitosis (Sam68) is an RNA-binding protein involved in cellular processes within viral infections [71,72]. A previous study found that Sam68 interacts with 3D^pol and then relocates from the nucleus to the cytoplasm to regulate viral replication [73]. Redistribution of Sam68 was similarly observed upon FMDV or bovine enterovirus 1 (BEV-1) infection, and Sam68 directly interacts with FMDV IRES. Knockdown of Sam68 using RNAi significantly inhibited FMDV replication and IRES activity [71]. Further research suggests that Sam68 May be involved in FMDV RNA synthesis. Experimental studies coupled with computational predictions have shown direct binding between Sam68 and FMDV-encoded 3D^pol [74]. Although the significance of the Sam68-3D^pol interaction on viral infection has yet to be determined and needs to be further explored, the above findings suggest Sam68 could act as a mediator to make FMDV RNA and 3D^pol spatially closer, thus promoting daughter RNA synthesis.

We previously described that EV-A71 3D^pol binding to NLRP3 promotes inflammasome complex assembly; SVA 3D^pol interacts with NLRP3 to activate the inflammasome assembly. In addition, SVA 3D^pol interacts with IKKα and IKKβ to activate NF-κB, thereby promoting pro-IL-1β transcription [75]. DEAD-box RNA helicase 1 (DDX1) regulates RNA metabolism and can control viral replication [76,77]. A recent study identified DDX1 interacting with 3D^pol of FMDV by two-hybrid experiments. DDX1 inhibits the replication of FMDV, which is partly related to the activity of the DDX1 ATPase/helicase. In addition, DDX1 stimulates the activation of the IFN-β pathway in FMDV-infected cells, and experiments showed that FMDV infection upregulated DDX1 mRNA expression. Yet, the protein expression level of DDX1 was reduced [78]. However, the functional details of DDX1-3D^pol interaction are unclear. Beyond indirect IFN-β effects operated by DDX1 under FMDV infection, whether 3D^pol interaction with DDX1 leads to DDX1 degradation is an exciting line of research. The IFN pathway is one of the primary mechanisms for host cells to fight pathogens [62]. For 3D^pol regulation of the IFN pathway, a recent study showed that FMDV-encoded 3D^pol inhibits activation of the IFN pathway by interacting with multiple elements involved in the IFN pathway, including IKKα, IKKε, IRF3, IRF7, NEMO, MDA5, and MAVS. Interestingly, species differences exist in these protein-protein interactions, such as 3D^pol only interacting with porcine-derived MAVS [79]. Obviously, the 3D^pol protein of FMDV and SVA plays a non-negligible role in viral RNA replication and is involved in diverse host cell biological processes by interacting with host factors to impose effects on virus pathogenicity.

Drugs acting on 3D^pol

Picornavirus genome replication is dependent on 3D^pol. Considering the critical role of 3D^pol proteins in the viral life cycle, 3D^pol is an attractive target for the design of antiviral strategies [16,80–82]. Indeed, viral RdRp has been one of the targets for developing antiviral drugs. For example, FDA-approved Remdesivir is used for COVID-19 and Ebola virus infection [83], and Ribavirin is used to combat hepatitis C virus (HCV) infection [84]. Polymerase inhibitors can be divided into two categories. The first category belongs to nucleoside and nucleotide substrate analog inhibitors, which target the active site of the viral polymerase. The second category is allosteric inhibitors, which inhibit the polymerase by binding to allosteric sites. Here, we summarize the drugs targeting picornaviruses 3D^pol.

Nucleoside/Nucleotide analogs

Nucleoside analog inhibitors are usually present in vivo as prodrugs that undergo metabolic activation to nucleotide forms. Nucleoside analogs mimic the entry of natural nucleotides into RdRp and inhibit RdRp by causing mutagenic effects and chain extension termination. The excessive rate of genetic mutation leads to genetic information loss, which is dangerous for organisms [85]. Currently, multiple nucleoside analogs are under development for use against picornaviruses (Table 1) (Figure 4).

Figure 4. Structures of nucleoside/nucleotide analog inhibitors.

Table 1. RNA-dependent RNA polymerase nucleoside/nucleotide analog inhibitors.

Drug names	Virus	Cell lines/animal model	EC50/IC50	Main mechanisms	References
Nucleoside/nucleotide analogs
Ribavirin	FMDV	BHK-21, and IBRS-2	EC50 = 28.09 μM	Mutagenicity	[111–113],
		Mice	Not tested
	PV Ribavirin	HeLa	EC50 = 100 μM		[85, 114]
5-nitrocytidine	PV, and CVB3	HeLa S3	Not tested	Inhibition of RdRp activity	[88]
NITD008	EV-A71	RD, 293T, and Vero	EC50 = 0.49 to 4.93 μM	Inhibit RNA chain elongation	[90, 91]
		Mouse	Not tested
	CVA16	Vero	EC50 = 0.64 μM
15 l	EV-A71	RD	IC50 = 0.19 ± 0.27 μM	Unclear	[92]
	EV-D68	RD	IC50 = 0.17 ± 0.16 μM
15f	EV-A71	RD	IC50 = 0.13 ± 0.08 μM	Unclear
Gemcitabine	PV	HeLa	IC50 = 0.3 μM	Inhibit RNA chain elongation	[93]
FNC	CVA6	RD	EC50 = 13.43 nM	Inhibit RNA synthesis	[96]
	CVA16		EC50 = 52.12 nM
	CVB3		EC50 = 33,78 nM
	EV-A71		EC50 = 16.87 nM
	EV-D68		EC50 = 1.548 nM
	EV-A71, and EV-D68	Mice	Not tested
GS-5734	EV-A71	HeLa	EC50 = 0.203 to 0.991 μM	Inhibit RNA synthesis	[99]
	CVB3		EC50 = 0.097 μM
	EV70		EC50 = 0.026 μM
	PV	Human	Not tested		[100]
T-1105	FMDV	BHK-21	EC50 = 0.26 to 1.31 μg/ml	Mutagenicity and inhibit chain elongation.	[101]
		Pig	Not tested
T-705	HAV	Huh7	IC50 = 18.8 μM	Mutagenicity	[104]
FU	FMDV	BHK-21	Not tested	Inhibit VPg uridylation and RNA chain elongation	[106, 107]
7-deaza-6-methyl-9-β-D-ribofuranosylpurine	PV	HeLa	IC50 = 11 nM	Destabilization of the PV RdRp-RNA primer/template complex or by interfering with the host state	[116]
2-C-MetCyt	FMDV	BHK-21	EC50 = 6.4 ± 3.8 μM	Unclear	[118]
	SVDV	SK6	EC50 = 45.2 ± 0.5 μM
	HPeV1	BGM	EC50 = 21 ± 4 μM		[119]
	HPeV3		EC50 = 4 ± 1 μM
	EV-A71	RD	EC50 = 1.3 ± 0.04 μM		[120]
			EC50 = 2 ± 0.4 μM
		HIOs	EC50 = 1 ± 0.3 μM
			EC50 = 1.5 ± 0.3 μM
	CVA16		Not tested
	EV-A71	C57 mice	Not tested

Pyrimidine ribonucleoside triphosphate analog (rPTP) is a potent mutagen for the PV genome, with in vitro incorporation properties superior to those of ribavirin triphosphate [87]. The N-6-modified purine analog as a mutagen significantly inhibited poliovirus (PV) and coxsackievirus replication in vitro. At the same time, a high-fidelity PV mutant (G64S) exhibited resistance to this analog, which suggests the mutagenic effect of this analog plays an antiviral role [88].

In addition to the mutagenic impacts on the viral genome, inhibition of viral RNA synthesis is an essential antiviral strategy. 5-nitrocytidine triphosphate has a high affinity for PV 3D^pol and probably inhibits PV and CVB3 replication by competitively inhibiting RdRp activity [89].

The adenosine analog NITD008 was previously demonstrated to block flavivirus RNA synthesis. However, NITD008 is toxic to rats [90]. NITD008 also showed inhibitory effects against EV-A71, and in vivo experiments demonstrated that it protected mice from clinical signs or death [91]. The NITD008 triphosphate acts as a chain terminator to inhibit EV-A71 RdRp activity directly [92]. Mutations conferring NITD008 resistance were localized to the viral 3A and 3D^pol regions. Reverse genetics confirmed that mutations in 3A and 3D^pol resulted in resistance to NITD008, and that a combination of mutations in 3A and 3D^pol resulted in stronger resistance [91]. Considering the toxicity of NITD008 in vivo, a series of novel NITD008 derivatives have also been synthesized. The half maximal inhibitory concentration (IC₅₀) of phosphoamidate (15f) containing a hexyl ester of l-alanine against EV-A71 was 0.13 ± 0.08 μM. The derivative containing a cyclohexyl ester of l-alanine (15 l) showed a high selectivity index (SI) against EV-A71 (IC₅₀ = 0.19 ± 0.27 μM, SI = 117.00) and EV-D68 (IC₅₀ = 0.17 ± 0.16 μM, SI = 130.76). The toxicity of these NTD008 derivatives in vivo are yet to be determined [93].

2’,2’-difluoro-2’-deoxycytidine, also known as gemcitabine, had anti-poliovirus potential by high-throughput screening and inhibition of PV replication in HeLa cells at an IC₅₀ of 0.3 μM. Sensitization to gemcitabine varied between cell lines, and although 25 μM gemcitabine had no detectable cytotoxicity in HeLa cells, 0.1 μM gemcitabine was sufficient to kill Vero cells [94]. Further evaluation of the in vivo toxicity of these NITD008 derivatives and gemcitabine is important to determine their potential as anti-viral drugs.

2’-Deoxy-2’-β-fluoro-4’-azidocytidine, also known as azvudine or FNC, a cytidine analog, has been previously shown to have potent antiviral activity against HCV, human immunodeficiency virus (HIV), and human and duck hepatitis B virus (HBV) [95,96]. In 2021, this drug received marketing approval in China for the treatment of HIV infection (https://www.nmpa.gov.cn/yaowen/ypjgyw/ypyw/20210721142223181.html). Recently, FNC was found to effectively inhibit viral replication of EV-A71, EV-D68, CVA6, CVA16, and CVB3 by targeting 3D^pol, with 50% effective concentrations (EC₅₀) ranging from 1.548 nM to 52.12 nM. Animal experiments show FNC (1 mg/kg) protecting mice from lethal EV-A71 or CVA16 challenges [97]. In light of these studies, FNC is a promising drug for treating enterovirus infections.

Remdesivir (GS-5734) is a monophosphate prodrug of an adenosine nucleoside analog that is reported to have broad antiviral activity by targeting viral RdRp [98,99]. Recent studies have identified that GS-5734 inhibits RNA synthesis of EV-A71, hindering enterovirus replication, and shows broad-spectrum inhibition of a wide range of enteroviruses [100]. Similarly, current clinical data showed that vaccine-derived poliovirus infection in an immunocompromised patient was cured by GS-5734 treatment after multiple failed attempts to treat with human milk, ribavirin, and oral human immunoglobulin [101].

Some drugs have multiple mechanisms of mutagenic effect and inhibit chain elongation. The pyrazinecarboxamide derivative, T-1105 (3-hydroxy-2-pyrazinecarboxamide), provides strong in vitro antiviral activity against a wide range of FMDV serotypes and completely protects pigs (orally administered 200 mg/kg) against FMDV infection [102]. T-1105 fluorinated derivative favipiravir (6-fluoro-3-hydroxy-2-pyrazinecarboxamide), called T-705, selectively inhibits RdRp of influenza and many other RNA viruses [103]. In 2020, a clinical case was reported using favipiravir and zanamivir to clear influenza B infection in an immunocompromised child [104]. 100 μM favipiravir treatment significantly reduced HAV RNA levels by 80% in cell cultures. The present results suggest that the mutagenic effect of favipiravir in the HAV genome plays a crucial role in the antiviral process [105]. However, recombination of picornavirus genes rescues viruses from error catastrophe [106].

Treatment with the base analog 5-fluorouracil (FU) resulted in an increased frequency of mutations in FMDV, and replication inhibition was observed [107,108]. The FMDV MARLS mutant, which has high fitness in BHK-21 cells, showed greater resistance to FU [107]. FU is converted intracellularly to FUTP, which competitively inhibits VPg uridylation, and FUMP is inserted into the UMP or CMP site of the template to induce mutagenesis [109]. This suggests a dual action of FU-disruption of VPg uridylation and mutagenic inhibition of FMDV. However, the fact that FU is not easily absorbed by cells has hampered its application, so modifications to FU are being made [110].

Ribavirin is a guanosine nucleoside analog that has been widely used in clinical practice for decades and exhibits a broad spectrum of antiviral activity through five mechanisms, including inosine monophosphate dehydrogenase inhibition, immunomodulatory effects, interference with RNA capping, polymerase inhibition, and lethal mutagenesis [111]. Ribavirin eliminates FMDV from persistently infected cell cultures [112,113]. Treatment with ribavirin in an adult mouse model (50 mg/kg/day, intraperitoneal injection, twice daily for 3 days) had an 86.67% protective efficiency before FMDV challenge and 66.67% protective efficiency after FMDV challenge. Both pre and post-treatment with ribavirin protected suckling mice (50 mg/kg, intraperitoneal injection) from lethal challenge [86]. Vaccines take time to produce protection against viruses, and the co-administration of ribavirin (1800 mg/pig) with the FMDV vaccine enhanced early antiviral resistance [113]. Mutagenesis remains the primary mechanism of ribavirin antiviral activity. Ribavirin induces over-mutation in the poliovirus genome, forcing an “error catastrophe” that inhibits viral replication [85,114]. The mutagenic ribavirin adulterates the viral genome, leading to lethal mutagenesis, but ribavirin treatment may also lead to the emergence of resistant viral populations in the presence of low to moderate concentrations of the drug [115].

In addition to the nucleoside analog inhibitors mentioned above, other analogs have equally good antiviral activity. 7-deaza-6-methyl-9-β-d-ribofuranosylpurine effectively inhibits poliovirus, but it does not act as a mutagen or strand terminator for 3D^pol, instead potentially exerting its antiviral effects through destabilization of the PV RdRp-RNA primer/template complex or by interfering with the host state [116]. 2’-modified nucleoside analogs can compete with the substrate nucleoside triphosphate to inhibit HCV RNA synthesis [117]. Ribonucleoside analog 2’-C-methylcytidine (2’-C-MetCyt) exhibits antiviral activity against FMDV and SVDV [118]. Similarly, 2’-C-MetCyt inhibited human parechovirus A1 (HPeV1) and parechovirus A3 (HPeV3) replication in vitro [119]. The anti-EV-A71 activity of 2’-C-MetCyt was also confirmed in human small intestinal organoids (HIOs) and RD cells [120].

Non-nucleoside inhibitors

The inhibitory mechanism of non-nucleoside analog inhibitors is ambiguous, usually binding to the RdRp variable site to inhibit the enzyme conformational change or binding to the active site to impede substrate entry. Several non-nucleoside analog inhibitors have been shown to inhibit picornaviruses replication (Table 2) (Figure 5).

Figure 5. Structures of non-nucleoside inhibitors.

Table 2. RNA-dependent RNA polymerase non-nucleoside inhibitors.

Drug names	Virus	Cell lines/animal model	EC50/IC50	Main mechanisms	References
Non-nucleoside inhibitors
Amiloride	CVB3	HeLa T	IC50 = 60 μM	Inhibit VPg uridylation and chain elongation	[121, 122]
	FMDV	IBRS-2	EC50 = 304 ± 97.1 μM		[126]
			EC50 = 168.8 ± 98.57 μM
DTriP-22	EV-A71	RD	EC50 = 0.13 to 0.44 μM	Inhibition of chain elongation	[128]
	CVA, CVB, and echovirus 9		EC50 = 0.07 to 1.22 μM
ATA	EV-A71	Vero	EC50 = 2.9 μM	Inhibition of chain elongation	[129]
5D9	FMDV	BHK-21	IC50 = 12 μM	Inhibition of chain elongation	[130]
GPC-N114	EV-A71, EV-D68, EV-D70, CVA16, CVA21, CVB3, E11, PV1, PV2, PV3, HRV2, and HRV14	RD, Hela H, Vero, Hela R19, and BGM	EC50 = 0.13 to 1.73 μM	Inhibition of chain elongation	[132]
	FMDV-O, and FMDV-A	BHK-21	EC50 >25 μM
	Equine rhinitis A virus	BGM	EC50 >100 μM
	EMCV	BGM	EC50 = 5.44 ± 0.49 μM
BPR-3P0128	EV-A71	RD	EC50 = 0.029 ± 0.001 μM	Inhibit VPg uridylation and chain elongation	[135]
		RD	EC50 = 0.0029 ± 0.001 μM		[136]
	CVB3	RD	EC50 = 0.062 ± 0.002 μM		[135]
	HRV2	RD	EC50 = 0.079 ± 0.002 μM
Sorafenib	FMDV	BHK-21	EC50 = 2.03 μM	Inhibition of RdRp activity	[138]
AcTU	EV-A71	RD	EC50 = 1.0 nM	Inhibition of RdRp activity	[139]
	EV-D68		EC50 = 0.75 nM
	CVA16		EC50 = 4.96 nM
	CVA21		EC50 = 17.15 nM
	CVB1		EC50 = 5.86 nM
	EV-A71	Mice	Not tested
1, 20a, 20b, 20i, 20j, 20 l, 24a, and 24c	FMDV	IBRS-2	EC50 = 13 to 245 μM	Inhibition of RdRp activity	[140]
NSC217697, NSC670283, NSC292567, and NSC65850	FMDV	BHK-21	EC50 = 0.78 to 3.49 μM	Inhibition of RdRp activity	[141]
ZHSI-1	EV-A71	RD	IC50 = 3.272 μM	inhibit the localization of 3D^pol and PI4KB on the lysosomal surface	[49]
	CVA16		Not tested
	EV-A71	C57 mice	Not tested

Amiloride and its derivatives inhibit CVB3 replication, and mutational analyses showed that either a serine-to-threonine (S299T) substitution at amino acid 299 or an alanine-to-valine (A372V) substitution at amino acid 372 of the CVB3 polymerase conferred CVB3 resistance to amiloride and its derivatives. Amino acids at positions 299 and 372 of 3D^pol are in the structural modality of catalytic activity, suggesting that amiloride and its derivatives act as CVB3 inhibitors of 3D^pol [121]. Further study showed that amiloride and its derivatives competitively inhibit the entry of nucleoside triphosphate and Mg2⁺ into CVB3 3D^pol, inhibiting VPg uridylation and chain elongation [122]. Ogram et al. also found that amiloride inhibits CVB3 and PV1 RNA replication, and computer models found that amiloride docks into the VPg binding site on 3D^pol to inhibit VPg uridylation [123]. The inhibitory activity of various amiloride derivatives on CVB3 3D^pol has been further investigated. Modification of the 5-amino group of amiloride increased the inhibitory activity, whereas 5-guanidino group substitution decreased the inhibitory activity. 3,5-diaminopyrazinyl substitution led to complete loss of the inhibitory activity, and the combination of 5-amino and guanidino group substitutions was more effective than both single substitutions [124]. Amiloride compounds alter intracellular ion concentrations, affecting polymerase fidelity and leading to indirect mutagenic activity [125]. Recent studies have shown that amiloride can also resist FMDV infection [126]. In conclusion, amiloride has a broad-spectrum resistance to picornaviruses through multiple mechanisms.

Previous studies have found that piperazine-containing pyrazolo[3,4-d]pyrimidine derivatives have antiviral activity against EV-A71 [127]. The further discoveries of piperazine-containing pyrazolo[3,4-d] pyrimidine derivatives, DTriP-22, showed broad antiviral activity against picornaviruses. DTriP-22 was found to act as an inhibitor of EV-A71 RNA polymerase and inhibit its elongation activity. Reverse genetics demonstrated that an arginine-to-lysine amino acid substitution at residue 163 (R163K) within the 3D polymerase conferred resistance to DTriP-22 [128].

Aurintricarboxylic acid (ATA) inhibits the elongation activity of viral 3D^pol to inhibit EV-A71 viral replication [129]. In addition, the study found that ATA inhibits the inflammatory response induced by EV-A71 3D^pol [64]. Further evaluation of the in vivo antiviral capacity of ATA is warranted.

Compound 5D9 is a non-competitive inhibitor of FMDV polymerase for NTP and nucleic acid substrates. Notably, 5D9 inhibited FMDV replication in a dose-dependent manner and no resistant mutants were observed to arise after consecutive passages [130,131]. De Palma et al. presented the first non-nucleoside inhibitors targeting the RNA template channel of the picornaviruses polymerase, GPC-N114, with broad-spectrum resistance to enteroviruses and cardioviruses. Mechanistically GPC-N114 targets the polymerase template receptor nucleotide site. Notably, no resistance was obtained under GPC-N114 pressure in CVB3 or poliovirus. In contrast, 3D polymerase amino acid substitutions of both M300V and I304V allowed encephalomyocarditis virus to acquire resistance. The CVB3 3D^pol with GPC-N114 X-ray crystal structure (PDB: 4Y2A) indicates that GPC-N114 is located at the bottom of the RNA-binding channel [132](Figure 6A). However, currently GPC-N114 cannot be formulated for animal administration and further optimization of its structure is necessary.

Figure 6. X-ray crystal structure of 3D^pol and viral polymerase inhibitors. a: CVB3 3D^pol in complex with GPC-N114 (PDB: 4Y2A); b: EV-D68 3D^pol in complex with NADPH (PDB: 5ZIT). The palm, thumb and finger subdomains are shown in blue, cyan, and magenta cartoons, respectively. The inhibitors are shown in red sticks.

The inhibitor NADPH occupies the RNA template-binding channel of the EV-D68 RdRp and NADPH inhibits EV-D68, poliovirus, and EV-A71 RdRp activity in vitro. Furthermore, the X-ray crystal structure of the 3D^pol-NADPH complex shows the residues involved in the NADPH-binding pocket of the EV-D68 3D^pol are highly conserved in enterovirus 3D^pol (PDB: 5ZIT), suggesting that NADPH holds promise as an inhibitor of enterovirus (Figure 6B) [133]. NADPH showed inhibition of RdRp in vitro, but related cellular and animal experiments need to be performed. Seven potential poliovirus RdRp inhibitors were identified through high-throughput screening, and five were found to bind the enzyme active sites of the polymerase extension complex through X-ray crystallography. These screened inhibitors occupy the NTP binding site and interact with template RNA, mainly blocking polymerase initiation. These compounds were ineffective in stopping the elongation process of the enzyme, probably due to the lack of the triphosphate moiety [134]. However, the in vivo inhibition of viral replication is unclear.

BPR-3P0128 inhibits the replication of various RNA viruses, including influenza, human rhinovirus, coxsackievirus, and enterovirus [135]. Velu et al. showed that BPR-3P0128 inhibited EV-A71 replication in a dose-dependent manner; mechanistically, BPR-3P0128 inhibited EV-A71 RdRp activity and VPg uridylation. Moreover, no resistant strains were observed after 11 serial passages [136]. Sorafenib, a c-RAF and multi-kinase inhibitor, is approved for treating liver cancer [137]. Drug repurposing of sorafenib has been shown to inhibit viral replication in a wide range of viruses, and recent studies have found that sorafenib targets the FMDV 3D^pol active site, inhibiting polymerase activity and modulating the c-RAF and AKT/PI3K pathways to inhibit FMDV replication [138].

Tenovin-1 derivative AcTU showed broad-spectrum resistance against enteroviruses such as EV-A71, EV-D68, CVA16, CVA21, and CVB1. Mechanistically, AcTU inhibited viral replication by targeting 3D^pol, and excitingly, in vivo experiments demonstrated that oral administration of AcTU at 0.6 mg/kg/d to mice was highly protective against lethal EV-A71 attack [139]. Jeong et al. found 2-amino-4-arylthiazole targets FMDV 3D^pol by high throughput screening (HTS) adapting fluorescence polarization (FP), and in vitro evaluations of its derivatives 20i and 24a showed more effective inhibition of FMDV [140]. Theerawatanasirikul et al. screened compounds and found NSC217697 (quinoline), NSC670283 (spiro compound), NSC292567 (nigericin), and NSC65850 target 3D^pol to inhibit FMDV replication in vitro [141].

Based on 3D^pol and host protein interactions, a recent screen identified a novel inhibitor, ZHSI-1, that suppresses the replication of EV-A71 and CAV by inhibiting 3D^pol localization on the lysosomal membrane, providing adequate protection against EV-A71 infection in mice [49]. These compounds provide a basis for the design of more efficient antiviral drugs.

Conclusion

The Picornaviridae family of viruses are severe threats to human and animal health. With recent advances, the picornavirus genome and protein functions are becoming more apparent. The research on EV-A71 3D^pol has been relatively in-depth compared to several other viruses of the same family. Current knowledge suggests the mechanisms of modification and regulation of host signalling pathways by 3D^pol are similar in picornaviruses, providing more information for the study of picornaviruses. The PTM of 3D^pol is critical for its stability and activity, and 3D^pol activates inflammatory responses. This offers new insights for the development of drugs that have antiviral activity or treat virus-induced inflammatory reactions. Interestingly, 3D^pol carries a nuclear localization signal that regulates host RNA splicing [39]. An exciting direction to investigate would be additional roles for 3D^pol within the nucleus. In addition, further studies of 3D^pol localization at intracellular membranes are essential to understanding picornavirus replication. Finally, an increasing amount of evidence on the interplay between 3D^pol and type I interferon responses suggests a role for 3D^pol in modulating the host immune response. In summary, the 3D^pol protein of picornaviruses plays a vital role in viral replication and cell biological processes involved in the viral lifecycle, such as autophagy, inflammatory response formation, and type I interferon response caused by 3D^pol-host interactions.

The 3D^pol protein plays an indispensable role in the replication of picornaviruses and is highly conserved in structure and core functional elements, therefore it is a promising drug target. We have summarized the current discovery of drugs that most effectively target 3D^pol and inhibit of picornavirus replication. The two main categories of drugs are nucleotide analog inhibitors and non-nucleotide analog inhibitors. These inhibitors have exhibited good antiviral activity in vitro.

However, several issues need to be considered during the development of 3D^pol-targeting drugs. It is worth noting that different cells have different sensitivities to the drugs. Therefore, more attention needs to be paid to the issue of drug cytotoxicity [94]. Several drugs, such as Ribavirin, NITD008, FNC, T-1105, GS-5734, and AcTU, showed antiviral activity in vivo, and their broad spectrum protection against picornaviruses deserves to be further investigated. Additionally, several drugs have already been approved for treating other diseases, so repurposing of these drugs will significantly shorten the drug development cycle. The picornavirus polymerase has low fidelity, and a single drug tends to induce the production of drug-resistant mutants. Cocktail chemotherapy has been widely used in cancer therapy and similar attempts in the treatment of EV-A71 have been shown to delay the development of drug-resistant mutants [142,143]. Viruses resist drugs by rapidly mutating and as more and more targeted drugs are developed, drug resistance needs to be addressed. Notably, in vitro passages with two non-nucleoside analog drugs, GPC-N114 and BPR-3P0128, did not produce resistant mutants. Remarkably, the interactions between picornavirus 3D^pol and the host facilitate viral replication. Studies have identified a novel inhibitor that blocks 3D^pol-host interactions, thereby inhibiting viral replication [49]. These interactions provide a new direction for targeted 3D^pol drug screening.

In conclusion, there are still many undiscovered areas in the interactions between 3D^pol proteins and host organisms. Further studies of 3D^pol-host interactions are necessary, particularly on replication complex assembly and antagonism of innate immunity, which will be critical for understanding picornavirus replication and pathogenicity. Importantly, targeting 3D^pol-host interactions is also an exciting direction for new drug development.

Author contributions

J.P. and H.Z. wrote the initial draft with input from Z.N., F.Z., and H.Z., Z.N., F.Z., H.Z., H.Z., and J.P. edited the final draft.

Availability of data

Data sharing is not applicable to this article as no new data were created or analysed in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

We acknowledge support from the National Key R&D Program of China [2021YFD1800300]; the Hatch Project of State Key Laboratory for Animal Disease Control and Prevention [SKLADCP2023HP02]; the Key Project of National Natural Sciences Foundation of China [32330107]; the Science and Technology Major Project of Gansu Province [22ZD6NA001]; and the Innovative Group for Basic Research on the Interaction Between Important Pathogens and Livestock [23JRRA561].