Early enterovirus translation deficits extend viral RNA replication and elicit sustained MDA5-directed innate signaling

ABSTRACT Enteroviruses (EVs)—positive-sense single-strand RNA viruses of Picornaviridae—are among the most common human viral pathogens. Unlike other cytoplasmic viral RNAs (vRNAs), which are detected by RIG-I, sensing of EV RNAs is mediated primarily by MDA5. EVs evolved with rapid, cytotoxic life cycles as a means of host cell interference, ostensibly to suppress MDA5-orchestrated antiviral type-I interferon (IFN) responses. At the core of this strategy are viral 2A protease (2Apro)-directed cleavage of the eukaryotic initiation factor (eIF)4G—the central translation initiation scaffold and ribosome adaptor—and degradation of nuclear pores. EV MDA5 agonism has an intriguing innate footprint, characterized by polar TBK1-IFN regulatory factor 3 (IRF3) signaling, which fuels sustained type-I IFN release to provide context for antitumor CD8+ T-cell priming after in situ cancer vaccination. Here we compared EV-host interactions between wild-type EVs and the highly attenuated recombinant polio:rhinovirus chimera and cancer immunotherapy agent (PVSRIPO) to test how targeted inhibition of discrete steps in the viral life cycle affects the innate antiviral response. Our investigations demonstrate that sustained MDA5-TBK1-IRF3 signaling elicited by PVSRIPO is the result of inefficient immediate 2Apro-directed eIF4G degradation, while wild-type EVs counter host innate type-I IFN defenses via rapid eIF4G cleavage. The MDA5-directed innate signature of profoundly attenuated EVs may provide efficient innate immune-stimulatory assets, for example, in in situ cancer vaccination strategies. IMPORTANCE Multiple pattern recognition receptors sense vRNAs and initiate downstream innate signaling: endosomal Toll-like receptors (TLRs) 3, 7, and 8 and cytoplasmic RIG-I-like receptors (RLRs) RIG-I, and MDA5. They engage distinct signaling scaffolds: mitochondrial antiviral signaling protein (RLR), MyD88, and TLR-adaptor interacting with SLC15A4 on the lysosome (TLR7 and TLR8) and toll/IL-1R domain-containing adaptor inducing IFN (TLR3). By virtue of their unusual vRNA structure and direct host cell entry path, the innate response to EVs uniquely is orchestrated by MDA5. We reported that PVSRIPO’s profound attenuation and loss of cytopathogenicity triggers MDA5-directed polar TBK1-IRF3 signaling that generates priming of polyfunctional antitumor CD8+ T-cell responses and durable antitumor surveillance in vivo. Here we unraveled EV-host relations that control suppression of host type-I IFN responses and show that PVSRIPO’s deficient immediate host eIF4G cleavage generates unopposed MDA5-directed downstream signaling cascades resulting in sustained type-I IFN release.

human pathogenic RNA viruses evolved with sophisticated ploys to prevent, intercept, or counteract type-I IFN signaling at every conceivable level.
Enterovirus (EV)-host innate relationships are unique among human pathogenic RNA viruses.Sensing of EV RNA is orchestrated principally by MDA5 rather than RIG-I, which detects 5′ppp ends in vRNAs of most RNA virus families (1)(2)(3).EV RNA 5′ ends are covalently bound to the genome-linked protein VPg (4), which precludes RIG-I sensing and canonical translation initiation via eukaryotic initiation factor (eIF)4E binding to a 5′ 7-methyl-guanosine (m 7 G) "cap" (5).EVs prevent innate antiviral IFN release at least in part through drastic, cytopathogenic interference with host cell protein synthesis via 2A protease (2A pro )-directed cleavage of the translation initiation scaffold eIF4G.Two eIF4G isoforms (eIF4G1 and eIF4G2) enable m 7 G-cap-dependent translation initiation by bridging mRNA 5′ ends to 40S ribosomal subunits via binding to eIF4E and eIF3 (6).eIF4G1 and eIF4G2 are cleaved by EVs (7), inhibiting cap-dependent host protein synthesis by separating the N-terminal eIF4E-binding sites from the C-terminal eIF3-bind ing sites (8).This does not affect viral translation, as the EV internal ribosomal entry sites (IRESs) recruit ribosomes by binding eIF4G, or the C-terminal eIF4G cleavage fragment generated by 2A pro -directed cleavage, without eIF4E involvement (9).
We previously reported that polio:rhinovirus chimera and cancer immunotherapy agent (PVSRIPO), the live attenuated type-1 poliovirus (Sabin) vaccine replicating under control of a human rhinovirus (HRV) type 2 IRES (10), bolsters tumor immune surveillance due to sustained type-I IFN dominant inflammation in the tumor microenvironment (11,12).In this work, we compared innate antiviral immunity to PVSRIPO with the (wild-type) cytopathogenic EV coxsackievirus A21 (CAV21) and human rhinovirus type 16 (HRV16).We tested the impact of targeted inhibitors of specific steps in the EV life cycle on the host innate antiviral response.We show that in wild-type EV infection, incipient viral translation of incoming vRNA causes 2A pro -directed eIF4G1 cleavage and prevents MDA5 downstream signaling and early type-I IFN synthesis, thus intercepting the transcriptional network driving sustained IFN-α release.Our observations indicate that PVSRIPO's encumbered early host interference resulting in an extreme level of attenuation is optimal for sustained type-I IFN release required for effective in situ cancer vaccination.

Failure of PVSRIPO to induce immediate eIF4G cleavage is associated with inefficient early translation
Ever since transitioning to immortal HeLa cell lines in the 1950 s (13), poliovirus (PV) research has been dominated by this in vitro model.In lieu of wild-type PV, we employ CAV21-a member of the "C-CAVs" (receptor: intercellular adhesion molecule 1 [ICAM1]) -which, together with the PVs, are part of the HEV-C species of Enteroviridae.The C-CAVs likely gave rise to the PVs (14); CAV21 causes paralytic poliomyelitis in human ICAM1-transgenic mice (15), analogous to the PVs in human CD155-transgenic mice (16).
Yet, HeLa cells do not reliably model PVSRIPO relations with authentic non-malignant PV host cells, e.g., myeloid cells.Probing PVSRIPO susceptibility in monocyte-derived dendritic cells (DCs), monocyte-derived macrophages (MDMs), or ex vivo malignant glioma slices revealed a failure to execute eIF4G cleavage with absent cytopathogenicity (11,12,19,20).We reported PVSRIPO cytotoxicity with eIF4G cleavage in many in vitro tumor models (10,11), including after IFN-α pre-treatment of cells (21).We identified a panel of tumor models, A375 melanoma and Hs683/M059J malignant glioma cell lines, that mirrors loss of cytopathogenicity/eIF4G cleavage activity observed in slice culture assays (Fig. 1B through D).Exemplifying the cytopathogenic host relations of wild-type EVs, CAV21 infection yielded efficient eIF4G cleavage in our panel at par with HeLa cells (Fig. 1A through D).The CAV21:PVSRIPO differential is reflected in puromycylation assays of protein synthesis activity, documenting suppression of host translation in CAV21-but not PVSRIPO-infected A375 cells (Fig. S1).

EV-directed eIF4G cleavage prevents IRF7 induction/activation
vRNA sensing by MDA5/RIG-I initiates a signaling cascade that drives transcriptional programs leading to type-I IFN (IFN-α/IFN-β) production, coordinated in part by sequential and overlapping activation of IFN regulatory factors (IRF)3 and IRF7.IRF3/IRF7 control IFNA/B transcription yielding specific temporal release patterns of IFN-β and the 12 IFN-α isoforms (23).To unravel the relation of cytopathogenic profile with the antiviral type-I IFN response, we compared PVSRIPO to EVs on the opposite ends of a spectrum of intrinsic cytopathogenic potential: CAV21, exemplifying the cytotoxic HEV-C (Fig. 2A) and HRV16 (Fig. 2B), which both use ICAM1 for host cell entry.HRV16 has a naturally non-cytopathogenic phenotype in human airway epithelial cells, its principal human host target cells (24).Viruses were subjected to long-range kinetic analyses of viral translation/cytopathogenicity and the innate host response in A375 cells (Fig. 2A  through E).
Taken together, our observations show that deficient early eIF4G cleavage of PVSRIPO is associated with MDA5-TBK1-IRF3 signaling culminating in sustained IRF7 induction and IFN-α synthesis.They indicate that A375 cells accurately reflect host type-I IFN responses to PVSRIPO and wild-type EVs and, thus, are a suitable model for mechanistic investigations of innate defenses elicited by CAV21 vs PVSRIPO.

EV eIF4G cleavage occurs immediately after vRNA entry, prior to the onset of processive viral translation
Our studies suggest immediate eIF4G cleavage as a pivotal factor in shaping innate defenses against EVs.CAV21/HRV16 efficiently thwart type-I IFN responses, while PVSRIPO's failure to cleave eIF4G is associated with sustained IFN-α release.EV eIF4G cleavage is directed by 2A pro (28), which approaches its substrate via the eIF4G binding partner eIF3 (29).Upon entry, incoming vRNA is translated and 2A pro is auto-catalytically released from the nascent polyprotein by co-translational cis-cleavage at its N-terminus (30, 31) (Fig. 3A).PVSRIPO fails to cleave eIF4G efficiently at any stage of the infection in A375 cells, despite abundant viral proteins being generated at 6-48 hpi (Fig. 2A and B).Meanwhile, complete CAV21-mediated eIF4G cleavage occurs even before viral protein synthesis is detectable (Fig. 1A through D).This suggests that eIF4G cleavage depends on the specific circumstances of viral translation at incoming vRNA, e.g., subcellular localization and/or eIF4G/eIF3 recruitment.To test this hypothesis, we investigated eIF4G cleavage in the presence of an inhibitor of viral polyprotein processing.

Inhibiting EV polyprotein processing does not prevent eIF4G cleavage
While 2A pro accounts for proteolytic processing of EV polyproteins at two sites (31), all remaining cleavages but one are carried out by 3C pro .Inhibiting 3C pro with the selective, irreversible inhibitor rupintrivir (34) does not interfere with entry, translation of incoming vRNA or 2A pro auto-catalytic release (Fig. 3A).However, it prevents all subsequent steps in the EV life cycle leading to hijacking of phosphatidyl-inositol 4-kinase IIIβ for generating PI4-phosphate-enriched membrane structures that serve as vRNA replication platforms (Fig. 3B).
We used rupintrivir to investigate 2A pro cleavages in CAV21 vs PVSRIPO-infected A375 cells at a dose of 100 nM (see Fig. S2 for dose titration [Fig.S2A and B] and estimates of inhibitory constants for viral translation [Fig.S2C] and dsRNA accumulation [Fig.S2D]).We analyzed viral polyprotein processing in infected A375 cells to confirm that rupintrivir (100 nM) leads to the accumulation of unprocessed P3 precursor in CAV21infected cells, and a decrease in (final processing product) 3D pol in CAV21/PVSRIPO-infec ted cells (Fig. S2E and F).In addition to eIF4G, we also tested the second prominent host feature targeted for immediate 2A pro cleavage, the nuclear pore (35), and NUP98 (29) specifically (Fig. 4).In CAV21-infected A375 cells, rupintrivir profoundly diminished viral translation and vRNA replication, as shown by dot blot detection of the ds-vRNA replicative intermediate (Fig. 4A).However, eIF4G and NUP98 cleavages were only minimally delayed, indicating that they occur prior to complete polyprotein processing (Fig. 4A).
Rupintrivir (100 nM) abolished viral translation/RNA replication of PVSRIPO.eIF4G cleavage did not occur in untreated cells, despite abundant viral protein synthesis in the processive phase, implying the presence of 2A pro (Fig. 4B).However, late NUP98 cleavage in-step with processive viral translation at 4-24 hpi was observed (Fig. 4B).This divergence may be explained by the finding that, in contrast to eIF4G, 2A pro binds NUP98 directly (29), and cleavage may not depend on the specific context of eIF3-eIF4G assembly (joining 2A pro with its eIF4G substrate [29]) during translation of incoming vRNA.Thus, CAV21 eIF4G cleavage occurs during translation of incoming vRNA, prior to complete polyprotein processing and vRNA replication (<3 hpi, Fig. 4A).PVSRIPO's deficit defining its host relationship is the inability to carry out this early eIF4G cleavage.Even during processive translation in cells with rampant type-I IFN responses (~4-24 hpi), eIF4G remains cleavage resistant (Fig. 4B).Nuclear pore degradation during processive viral translation in infected A375 cells (Fig. 4B) does not interfere with the robust antiviral type-I IFN response elicited by PVSRIPO (Fig. 2A and B).We tested whether intrinsic deficits of PVSRIPO's 2A pro mediate its eIF4G/NUP98 cleavage phenotype using a mix-and-match recombinant featuring the PVSRIPO 2A pro coding sequence within a CAV21 background (Fig. 4C and D).This chimeric virus was viable with efficient growth and cytotoxicity in HeLa cells, albeit a very small plaque phenotype (Fig. 4C).PVSRIPO 2A pro in a CAV21 background-mediated eIF4G/NUP98 cleavage with kinetics resembling wild-type CAV21 (Fig. 4D).Thus, PVSRIPO's host protein cleavage phenotype is the result of early viral translation deficits mediated by the foreign HRV2 IRES and not due to performance of its cognate 2A pro .

Inhibiting PI4KIIIβ or ENT1/ENT2 rescues type-I IFN responses to wild-type EV infection
While deficient immediate eIF4G cleavage defines PVSRIPO's accentuated type-I IFN imprint, this deficit does not explain the source and unusual kinetics of this response.We used the PI4KIIIβ inhibitor PI4KIIIβ-IN-10 to block the formation of PI4P-enriched replication organelles/vRNA replication (Fig. 3B) and assessed the kinetics of dsRNA accumulation, eIF4G cleavage, viral translation, innate signaling response, and host type-I IFN release in A375 cells (Fig. 5).Dose titration studies of PI4KIIIβ-IN-10 revealed higher sensitivity of dsRNA accumulation for PVSRIPO (IC 50 , ~5 nM) vs CAV21 (IC 50 , ~22 nM) (Fig. S3).We deliberately used a PI4KIIIβ-IN-10 dose with incomplete suppres sion of vRNA replication (10 nM, Fig. S3) to gauge the ability of CAV21 vs PVSRIPO to withstand PI4K inhibition and to induce innate host type-I IFN responses.
PI4K inhibition had a minor effect on eIF4G cleavage by CAV21, consistent with this event occurring prior to PI4P enrichment, upon translation of incoming vRNA (Fig. 5A).In contrast, CAV21 dsRNA accumulation (up to 24 hpi) and processive viral translation were profoundly suppressed (Fig. 5A, C, and E).As shown in Fig. 2, CAV21 infection of mock-treated cells led to blunted p-STAT1(Y701) without downstream IRF7 induction or type-I IFN release (Fig. 5A and E).P-STAT1(Y701) at 24-40 hpi in mock-treated cells likely does not reflect type-I IFN signaling (IFN-α/IFN-β were not detected; Fig. 5A) and may be due to cytotoxicity.
In accordance with our dose-titration assays (Fig. S3), PI4K inhibition with 10 nM PI4KIIIβ-IN-10 did not block CAV21 vRNA replication permanently.DsRNA accumulation was undetectable before 24 hpi in the presence of drug but gradually increased at 24-40 hpi (Fig. 5A and C).At 40 hpi in the presence of 10 nM PI4KIIIβ-IN-10, a dose that blocked processive viral translation (at all times) and vRNA replication (up to 24 hpi), IRF7 induction, and type-I IFN production occurred (Fig. 5A).Thus, profound suppression of vRNA replication protects cells from CAV21 cytopathogenicity, enabling delayed dsRNA accumulation that triggers innate antiviral type-I IFN defenses (akin to PVSRIPO).This response may be possible because of the presence of intact eIF4G in CAV21-infected, PI4KIIIβ-IN-10-treated cells, due to either remnant protein or the induction of eIF4G biosynthesis in infected cells in the presence of PI4K inhibition (Fig. 5A).Indeed, IRES activity has been shown to drive expression of eIF4G1 isoforms generated by a series of in-frame initiation codons (36).Thus, in CAV21-infected cells protected from viral processive translation/cytopathogenicity by PI4KIIIβ-IN-10, active cap-independent translation of eIF4G1 may enable type-I IFN responses.In contrast to CAV21, PI4K inhibition abolished all PVSRIPO activities other than a weak, late type-I IFN response (Fig. 5B and D).Due to deficient translation of incoming vRNA in A375 cells, PVSRIPO cannot overcome PI4K inhibition in the manner observed with CAV21 at the same dose of PI4KIIIβ-IN-10.Thus, no dsRNA was detected in the presence of PI4K inhibitor (1-64 hpi) (Fig. 5D), and the innate STAT1-IRF7 antiviral response was almost completely blocked (Fig. 5B and D).In untreated cells, the PVSRIPOinduced sustained p-IRF7(S477)/IFN-α/IFN-β response (Fig. 5B and F) correlated with ongoing dsRNA accumulation (Fig. 5D).Thus, type-I IFN responses to CAV21 (induced by PI4K inhibition) and to PVSRIPO are enabled by preservation of intact eIF4G and the accumulation of viral dsRNA; they do not correlate with the kinetics of viral translation.
PI4KIIIβ-IN-10 treatment of HeLa cells caused delayed PVSRIPO dsRNA accumulation and IRF3-STAT1 phosphorylation/activation (Fig. S4A) in a similar manner as observed for CAV21 in A375 cells (Fig. 5A).Thus, PI4K inhibition converted PVSRIPO's cytopathogenic profile in HeLa cells (Fig. 1A) toward the accentuated MDA5 signaling phenotype in A375 cells (Fig. S4A).Type-I IFN responses to CAV21 and PVSRIPO mirrored the pace of dsRNA accrual in all models/conditions, consistent with evidence that the EV replicative intermediate is the pattern sensed by MDA5 (1).Conversely, inhibiting TBK1 with Bx795 (19) in A375 cells converted the accentuated MDA5 signaling phenotype of PVSRIPO into a cytopathogenic program with enhanced early viral translation and efficient eIF4G cleavage (Fig. S4B).
We extended our findings with inhibition of 3D pol -direted vRNA synthesis by dipyridamole (DIP) treatment of A375 cells.Broad antiviral activity of DIP, due to inhibition of nucleoside import by ENT1/ENT2 (Fig. 3B) (37), was shown in 1977 (38).DIP dose titration assays informed our decision to use a concentration of 100 µM with the intent to prevent cytotoxicity and produce host innate type-I IFN responses to CAV21 at 24 and 48 hpi (Fig. S5A).In CAV21-infected A375 cells, DIP severely curtailed processive viral translation (at any time) and dsRNA accumulation (up to 16 hpi, Fig. 6A and C).However, dsRNA accumulation in the presence of 100 µM DIP recovered by 24 hpi (Fig. 6A and C).At 30 hpi, it increased further, followed by a gradual decrease (Fig. 6A).Late dsRNA accrual in DIP-treated cells occurred in step with a robust type-I IFN response (Fig. 6A).As observed with PI4K inhibition, this may be due to partially preserved host cell integrity (due to inhibition of processive viral translation) and retention of intact eIF4G (Fig. 6A).
To examine eIF4G status specifically, the assay shown in Fig. 6A was repeated with measures to achieve very high eIF4G detection sensitivity (Fig. 6C).This revealed the absence of intact eIF4G in mock-treated CAV21-infected cells (after 16 hpi), while intact eIF4G was preserved, or even induced (see 42 hpi), in DIP-treated samples (Fig. 6C).These observations re-affirm our hypotheses that early EV eIF4G cleavage abolishes the antiviral type-1 IFN response and that preservation of eIF4G integrity in PVSRIPO-infec ted cells enables complex, MDA5-orchestrated TBK1-IRF3-IRF7 signaling and sustained IFN-α release.
Our investigations revealed that PVSRIPO-mediated IRF3-IRF7 activation and sustained type-I IFN release are the results of an extreme level of attenuation far beyond the naturally non-cytopathogenic (in human airway epithelium) HRV16.The failure of the virus to execute immediate eIF4G cleavage, mediating loss of cytopathogenicity in relevant PV target host cells, produces non-cytotoxic dsRNA accumulation that triggers unopposed MDA5 engagement.

DISCUSSION
EVs use a unique, highly effective strategy of thwarting innate antiviral type-I IFNs through drastic host interference via 2A pro -directed host protein cleavage.Our   type-I IFN release in EV-infected cells.Inhibiting polyprotein processing with rupintri vir and interfering with vRNA replication via PI4K or ENT1/ENT2 inhibition profoundly diminished processive viral translation/dsRNA accumulation but only modestly affected eIF4G/NUP98 cleavage in CAV21-infected cells.This demonstrates that the 2A pro -directed host protein cleavages occur upon initial translation of incoming vRNA, prior to complete polyprotein processing and the onset of vRNA replication.A persistent conundrum in the field has been the stoichiometry of the 2A pro -substrate relation, where near-complete cleavage of the entire pool of eIF4G present in the cell is executed by a limited 2A pro supply early in the infectious cycle [reviewed in reference (39)].The fact that abundant PVSRIPO translation during the processive phase of viral protein synthesis (~4 to 24 hpi) fails to execute eIF4G-but not NUP98-cleavage, as well as the finding of 2A pro :eIF3 binding as a possible requirement for cleavage (29), hints at possible explanations.We speculate that the specific conditions for initial translation of incoming vRNA, e.g., the context provided by subcellular location or the manner of eIF4G:eIF3 engagement during initial viral translation, are necessary for immediate 2A pro -directed eIF4G cleavage to occur.
Our investigations point to deficient immediate viral translation at incoming vRNA as the root cause for PVSRIPO's inability to cleave eIF4G.This phenotype was linked to stem-loop domains (SLDs) 5 and 6 in the HRV2 IRES (40).We showed previously that SLDs 5 and 6 of EV IRESs, the footprint for binding to the eIF4G translation initiation scaffold (9,19,40), bind the dsRNA-binding protein kinase (PKR) (19).Upon PKR binding to the IRES, PKR dimerization and auto-phosphorylation ensue in a manner that interferes with translation initiation at the IRES, possibly via steric hindrance of eIF4G binding to stem loop domains 5 and 6 (19).We suspect that a relative deficiency in recruiting eIF4G early exposes PVSRIPO RNA genomes to PKR sensing, thus delaying immediate translation of incoming vRNA and protecting eIF4G from 2A pro cleavage.
Comparing CAV21, HRV16, and PVSRIPO innate antiviral immunity in the same host (A375 cells) side-by-side revealed the consistency of the EV eIF4G cleavage paradigm and PVSRIPO's profound attenuation.Despite ~10-h delayed eIF4G cleavage (relative to CAV21) and efficient IRF3(S386) phosphorylation-roughly at par with PVSRIPO-a type-I IFN response to HRV16 was almost completely absent.This suggests that eIF4G cleavage, at a level and with timing commensurate with preventing host type-I IFN release, defines the Enterovirus genus.
Comparing CAV21 with PVSRIPO in the presence of PI4K or ENT1/ENT2 inhibition also revealed an extreme level of attenuation for the latter.The accentuated MDA5-IRF3-IRF7 signature of PVSRIPO in A375 cells was achieved with CAV21 only when host cell integrity was preserved, processive viral translation nearly abrogated, and vRNA replication markedly impeded by PI4KIIIβ or ENT1/ENT2 inhibition.As in the case of PVSRIPO, innate type-I IFN responses to (PI4K or ENT1/ENT2 inhibition-impeded) CAV21 correlated with the rate of viral dsRNA accumulation.
PVSRIPO exhibits cytopathogenic properties in many high-passage, tissue cultureadapted tumor models, such as HeLa cells.Elucidation of mitogenic signaling cascades converging on translation initiation machinery revealed that uncontrolled PKC-Raf-ERK1/2-MNK1/2 signaling-a hallmark of the malignant state (41)-spurs early viral IRES-mediated translation in PVSRIPO-infected cells and, thus, favors eIF4G cleavage (42)(43)(44)(45)(46)(47)(48)(49).It is unclear if the prevailing conditions for protein synthesis control in high-passage cancer cell lines (eg., HeLa cells) are representative of neoplastic cells in their native habitat or if they are skewed by serial passage adaptations.PVSRIPO treatment of glioblastoma patient ex vivo tumor slices-an undissociated, non-passaged authentic representation of the heterogeneous mix of cells present in the tumor-revealed the absence of cytotoxicity in all tumor compartments, including neoplastic cells (12).
Given its fundamentally altered host-innate relations, it is highly improbable that PVSRIPO could establish a gastrointestinal replication reservoir-let alone transmis sion-consistent with the lack of virus shedding in clinical trials of PVSRIPO in recurrent glioblastoma (50) and melanoma (51,52).Without the eIF4G cleavage paradigm-the defining characteristic of the Enterovirus genus shaping EV-host innate relations-the genetically engineered, non-cytopathogenic PVSRIPO may be considered axiomatically distinct from its relatives.

Immunoblot analyses
Immunoblots were carried out essentially as reported before (46,47).Briefly, cells were lysed with polysome lysis buffer (10 mM HEPES, pH 7.4, 100 mM KCl, 5 mM MgCl 2 , 0.5% Igepal CA-630 [NP-40, Sigma], supplemented with Halt protease/phosphatase inhibitor [ThermoFisher]) followed by o/n incubation at −80°C.Lysates were centrifuged 10 min at 14,000 × g, and supernatants were loaded directly on nitrocellulose for dot-blot analysis or boiled with lithium dodecyl sulfate (LDS) sample buffer (Thermo Fisher) for Western blots.Immunoblots were developed with SuperSignal West Pico-Plus, -Femto, or -Atto chemiluminescent substrates (ThermoFisher), depending on protein abundance/antibody sensitivity; RNA or protein bands were quantified using the Li-COR Odyssey FC imaging system and Image Studio software as described earlier (19).Briefly, protein levels (relative values) plotted on all graphs were calculated as follows: max imum detected value was set at 100%, each value was normalized to the GAPDH loading control at the corresponding time point.Calculations were performed from three independent experiments; means ± standard error of the mean are shown on the graphs.

FIG 1
FIG 1 Short-term time course of PVSRIPO and CAV21 infection in a panel of in vitro tumor models (multiplicity of infection [MOI], 10).(A-D) HeLa R19 (A), A375 (B), Hs683 (C), and M059J (D) cells were infected with PVSRIPO or CAV21.Lysates collected at the indicated hpi were analyzed by immunoblot for viral translation (2C) and 2A pro -directed cleavages (eIF4G, YTHDF3; asterisks indicate eIF4G/YTHDF3 cleavage products).All assays were performed at least three times, and representative immunoblots are shown.The panels in the right column depict quantification of protein levels (% max.detected, normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading control; means ± standard error of the mean), representing the average values from three independent series.(E) Expression of poliovirus receptor PVR (CD155) and CAV21 receptor ICAM1 (CD54) in the panel.(F) P-STAT1(Y701) response to transfection of increasing concentrations of poly(I:C) in the cell line panel.Cell lysates were collected at the indicated intervals and tested for p-STAT1(Y701)/STAT1 by immunoblot.

4 FIG 2
FIG 2 Long-term time course of EV infection in A375-and myeloid antigen-presenting cells.(A and B) A375 cells were infected with CAV21 or PVSRIPO (MOI, 10 [A]) and with HRV16 or PVSRIPO (MOI, 10 [B]).HRV16 infections were carried out at 33°C.Lysates were analyzed by immunoblot for viral translation (eIF4G cleavage, 2C/2BC); TBK1 activation [p-IRF3(S386)/IRF3]; and type-I IFN signaling [p-STAT1(Y701)/STAT1, p-IRF7(S477)/IRF7, MDA5] over time (*eIF4 G1 cleavage products).(C) IFN-α/IFN-β immunoblots.(D and E) Expression levels of viral 2C and the indicated (phospho)-proteins in PVSRIPO-infected cells (D) or in CAV21-infected cells (E) were quantified.The values represent the average abundance of proteins in three independent series (% max.detected, normalized to GAPDH; mean ± standard error of the mean).The assays were performed in at least three independent series; representative results are shown.(F and G) Long-term time course of CAV21 (MOI, 10 [F]) and PVSRIPO (MOI, 10 [G]) infection in human monocyte-derived antigen-presenting cells.Lysates collected at the indicated intervals were analyzed by immunoblot for the parameters shown.Cells from two distinct donors were tested with CAV21 infection; #, loss of signal with the tubulin loading control (and all other proteins) at 24-48 hpi are due to rampant cytolysis (F).PVSRIPO infection of MDMs (G) was analyzed by immunoblot of samples collected at the indicated intervals.Cells from three distinct donors were tested in five independent series; representative results are shown with samples from one donor (the arrow indicates weak signal for viral 2C; # indicates prominent non-specific background bands).

FIG 5
FIG 5 Effect of PI4KIIIβ inhibition on the innate antiviral type-I IFN response to CAV21 and PVSRIPO (see Fig. S3 for extended data).A375 cells were infected with CAV21 (A) or PVSRIPO (B) as described for Fig. 2 in the presence or absence of 10 nM PI4KIIIβ-IN-10.Cell lysates collected at the indicated hpi were analyzed by dot blot for accumulation of viral dsRNA (upper panels) or by immunoblot for viral protein synthesis (2C), eIF4G cleavage, and the innate antiviral response [p-STAT1(Y701), p-IRF7(S477), IFN-α/IFN-β, and MDA5] (bottom panels; *eIF4G1 cleavage products).All assays were performed in at least three independent series; representative results are shown.Levels of dsRNA accumulation (C, CAV21; D, PVSRIPO), levels of p-STAT1(Y701)/2C in the presence of absence of inhibitor (E, CAV21), and levels of pIRF7(S477)/p-STAT1/2C in mock-treated cells (F, PVSRIPO) were quantified and plotted in the graphs (% max.normalized to GAPDH, means ± standard error of the mean).

FIG 6
FIG 6 The effect of inhibiting vRNA replication with dipyrimadole (DIP) on innate type-I IFN responses to CAV21 and PVSRIPO.Time course of A375 infection with CAV21 (A) or PVSRIPO (B) performed in a manner similar to that in Fig. 5 in the absence (dimethyl sulfoxide[DMSO]) or presence of 100 μM DIP.Cell lysates were analyzed by dot blot for vRNA replication (dsRNA, top panels) and by immunoblot for eIF4G cleavage, viral translation (2C), and innate type-I IFN response signatures (middle panels; *eIF4G1 cleavage products, **p-IRF7 with decreased electrophoretic mobility detected with IRF7 antibody).Assays were performed in at least three independent series; representative results are shown.Accumulation of dsRNA and 2C protein in CAV21-infected cells was quantified from three independent experiments and plotted in the bottom graph (% max.normalized to GAPDH, means ± standard error of the mean).(C) A repeat series of the assay shown in panel A with high detection sensitivity was conducted to rigorously document eIF4G1 status in CAV21-infected cells in the presence of DIP.
investiga tions indicate that immediate eIF4G cleavage is necessary and sufficient to abrogate