The ZCCHC14/TENT4 complex is required for hepatitis A virus RNA synthesis

Abstract

Despite excellent vaccines, resurgent outbreaks of hepatitis A have caused thousands of hospitalizations and hundreds of deaths within the United States in recent years. There is no effective antiviral therapy for hepatitis A, and many aspects of the hepatitis A virus (HAV) replication cycle remain to be elucidated. Replication requires the zinc finger protein ZCCHC14 and noncanonical TENT4 poly(A) polymerases with which it associates, but the underlying mechanism is unknown. Here, we show that ZCCHC14 and TENT4A/B are required for viral RNA synthesis following translation of the viral genome in infected cells. Cross-linking immunoprecipitation sequencing (CLIP-seq) experiments revealed that ZCCHC14 binds a small stem-loop in the HAV 5' untranslated RNA possessing a Smaug recognition-like pentaloop to which it recruits TENT4. TENT4 polymerases lengthen and stabilize the 3' poly(A) tails of some cellular and viral mRNAs, but the chemical inhibition of TENT4A/B with the dihydroquinolizinone RG7834 had no impact on the length of the HAV 3' poly(A) tail, stability of HAV RNA, or cap-independent translation of the viral genome. By contrast, RG7834 inhibited the incorporation of 5-ethynyl uridine into nascent HAV RNA, indicating that TENT4A/B function in viral RNA synthesis. Consistent with potent in vitro antiviral activity against HAV (IC₅₀ 6.11 nM), orally administered RG7834 completely blocked HAV infection in Ifnar1^-/- mice, and sharply reduced serum alanine aminotransferase activities, hepatocyte apoptosis, and intrahepatic inflammatory cell infiltrates in mice with acute hepatitis A. These results reveal requirements for ZCCHC14-TENT4A/B in hepatovirus RNA synthesis, and suggest that TENT4A/B inhibitors may be useful for preventing or treating hepatitis A in humans.

Keywords: RNA-binding protein; animal model; antiviral therapy; hepatitis A; picornavirus.

Fig. 1.

Hepatovirus RNA synthesis requires ZCCHC14 and TENT4A/B. (A) Organization of the HAV genome, subgenomic 18f-FLuc replicon, and 18f-NLuc reporter virus genome. (B) ZCCHC14 immunoblot of ZCCHC14-KO and control (sgCtrl) cells. (C) HAV RNA abundance in ZCCHC14-KO and sgCtrl cells 4 and 6 d postinfection (p.i.) with p16 virus. (D) Firefly luciferase (FLuc) expressed by ZCCHC14-KO and sgCtrl cells transfected with HAV-FLuc replicon RNA. Data are representative of 2 independent experiments, each with 3 technical replicates, and are shown as means ± SDs relative to FLuc expressed by the replication-incompetent HAV-FLuc/GAA RNA mutant. **P = 0.0012, ****P < 0.0001 by 2-way ANOVA with Šídák’s multiple comparisons test. (E) NLuc expressed by ZCCHC14-KO cells relative to sgCtrl cells 4 h p.i. ± SD with 18f-NLuc virus. Mean NLuc expression was 12.7-fold above background. n = 3 independent experiments, each with 3 technical replicates. (F, Left) circRNA HAV IRES reporter assay showing GFP synthesized from back-spliced RNA in ZCCHC14-KO and sgCtrl cells. (Right) Quantitation of GFP normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) protein in immunoblots and back-spliced RNA measured by specific RT-PCR assay. Data shown are technical replicates from 1 of 2 independent experiments with similar results. (G) Polysome analysis of ZCCHC14-KO and sgCtrl cells 6 h after 18f virus infection, showing distribution of HAV RNA among ribosomes and polysomes separated by gradient centrifugation. Actin mRNA distribution is shown in SI Appendix, Fig. S1 I. (H, Left) Immunoblots of TENT4 proteins in Huh-7.5 cells transduced with lentiviruses expressing guide RNAs targeting TENT4A, TENT4B, or both TENT4A and 4B (TENT4A/B) or a scrambled sgRNA (sgCtrl). (Right) NLuc expressed 48 h after infection of CRISPR-edited TENT4 knockout Huh-7.5 cells shown in the blots on the left. Data shown are mean percent NLuc relative to infected sgCtrl cells from 2 experiments, each with 3 technical replicates. P values by 1-way ANOVA with Geisser-Greenhouse correction and Dunnett’s multiple comparison test. (I) NLuc expressed by 18f-NLuc virus in Huh-7.5 cells 48 h p.i. with increasing concentrations of RG7834. Data shown are technical replicates from 1 of 2 experiments with similar results. (J) NLuc expressed by 18f-NLuc virus in Huh-7.5 cells 2–8 h p.i. in the presence or absence of 200 nM RG7834. Data from 1 of 2 independent experiments with similar results, each with 3 technical replicates. (K) Nascent HAV RNA synthesis assay. 18f virus-infected cells were metabolically labeled with 5-ethynyl uridine (5EU) for 24 h following the addition of RG7834 (200 nM) or DMSO. RNA was isolated, click-labeled with biotin, and captured on streptavidin beads for quantitation of nascent HAV and actin mRNA. Nascent RNA in DMSO-treated cells was arbitrarily set to 100. Results shown are means ± SEMs from 3 independent experiments, 1 with 2 technical replicates. (L) HAV poly(A) tail length assay. 18f virus-infected Huh-7.5 cells were treated with RG7834 (200 nM) or vehicle (DMSO) for 3 d. Total RNAs were extracted, guanosine and inosine (G/I) tailed, reverse transcribed (RT) and PCR amplified using genome-specific (GS) forward (F) and either GS or G/I-specific reverse (R) primers. PCR products were analyzed by agarose gel electrophoresis. Results shown are representative of two independent experiments. (M) HAV poly(A) tail length determined by sequencing 50 individual cDNA clones generated from RG7834 and vehicle-treated samples. Graph depicts median and quartile lengths; mean was 35.80 ± 11.2 SD nt for RG7834 versus 35.48 ± 10.71 SD for vehicle.

Fig. 2.

ZCCHC14 binds stem-loop Vb and recruits TENT4A to the HAV 5′ UTR. (A) Coimmunoprecipitation of HAV RNA with ZCCHC14 from lysates of 18f-infected Huh-7.5 cells. HAV genomic RNA and actin mRNA were quantified by qRT-PCR in immunoprecipitates and normalized to spiked-in GLuc RNA with RNA abundance in isotype control (Ctrl) precipitates arbitrarily set to 1.0. n = 3 independent experiments. ZCCHC14 immunoblot shown below. (B) FLAG-ZCCHC14 CLIP-seq result in 18f-infected 293T cells showing coverage of HAV-specific reads in 2 independent assays (blue and red tracks). Data are shown as log₂ of the ratio of reads recovered from the immunoprecipitate to input RNA, aligned to the HAV 18f genome. An expanded view of reads mapping to nucleotides 500–1,000 is shown at the top. (C) Secondary structure of domain V (nucleotides 596–709) of the wild-type HAV 5′ UTR (35), showing stem-loops Va, Vb, and Vc. Dashed lines indicate potential pseudoknot structure (35). pY2, pyrimidine-rich tract 2. (D) RNA pull down using full-length synthetic 3′ biotinylated HAV 5′ UTR and 5′ biotinylated 3′ UTR baits. RNA baits were incubated with Huh-7.5 cell lysate, then affinity purified on streptavidin beads. Copurified ZCCHC14 and TENT4A proteins were detected by immunoblotting. Relative fluorescence intensities within each blot are noted in italicized text. (E) ZCCHC14 and TENT4A pulldown with a 3′ biotinylated synthetic probe representing stem-loop Vb (nucleotides 644–661) and Vb-mut; nucleotide substitutions in Vb-mut (blue) ablate the CNGGN loop. (F) ZCCHC14 and TENT4A proteins pulled down with the Vb-wt probe from lysates of Huh-7.5 cells, in the presence and absence of 500 nM RG7834, and ZCCHC14-KO cells. *Nonspecific band. (G) RNA pull-down assays using (Top) RNA probes representing sequences upstream of domain V (nucleotides 1–95, 1–156, or 1–254) or (Bottom) full-length 5′ UTR sequence with either Vb loop substitutions (Vb-mut, see panel E) or complete deletion of Vb (ΔVb). (H) NLuc expressed by Huh-7.5 cells transfected with 18f-NLuc virus RNA with wild-type (Vb-wt), mutant (Vb-mut) or deleted (ΔVb) stem-loop Vb. Replication-incompetent RNA (Δ3D) was included as a control. (I) NLuc expressed 72 h after transfection of the 18f-NLuc Vb reporter virus mutants in sgCtrl versus ZCCHC14-KO cells. GAA, replication-incompetent virus mutant.

Fig. 3.

RG7834 treatment of acute hepatitis A in Ifnar1^−/− mice. (A) (Top) Serum ALT activities (***P < 0.001) and (Bottom) fecal HAV RNA quantified by qRT-PCR in mice treated with RG7834 10 mg/kg per os (p.o.) b.i.d. or vehicle only between days 6 and 12 p.i. with 2 × 10⁶ genome equivalents (GEs) of wild-type HM175-mp6 virus. n = 12–15 mice in each group between days 6–16, 8 on day 19, and 2–3 at day 33. *P = 0.016 by t test; ***P ≤ 0.001. (B) Geometric mean HAV RNA copies per microgram total RNA, ±95% confidence interval (CI), in liver tissue from mice in panel A on days 12 and 19 p.i. *P = 0.040 by Mann-Whitney test. (C and D) High (C) and low (B) magnification views of hematoxylin and eosin (H & E)–stained liver sections from mice in panel A. Tissues were collected on 12 (end of therapy) and 19 d p.i. Apoptotic hepatocytes are surrounded by inflammatory infiltrates (arrows). (E) Numbers of inflammatory foci per square millimeter in H & E-stained sections of livers. (F) Interferon-β and proinflammatory chemokine transcript levels determined by RT-PCR in livers harvested from mice in panel A on day 12, relative to levels expressed in naïve mice. P values by t test with Welch correction and Holm-Šídák test for multiple comparisons. (G) H & E-stained liver from an RG7834-treated animal in panel A on day 19. (H) Serum HAV-neutralizing antibody titers in HAV-infected mice in panel A. (I) Total CD8⁺ T cells and (J) HAV-specific tetramer-positive CD8⁺ T cells enumerated by flow cytometry in liver tissues 32 d p.i. in mice treated with RG7834 as in panel A. (K) Serum ALT activities in mice infected with high-titer HAV (2 × 10⁷ GE) and treated with RG7834, 10 mg/kg or 2.5 mg/kg b.i.d., or vehicle only (n = 5 in each group) between days 5 and 9 p.i. ***P < 0.0001 versus sham treated by 2-way ANOVA with Dunnett’s multiple comparison test. (L) Fecal HAV shedding by mice shown in panel H. Statistical testing as in panel H. Dashed horizontal lines indicate limits of detection. Error bars represent SDs in all panels, unless noted otherwise.

Fig. 4.

Chemoprevention of hepatitis A in Ifnar1^−/− mice. (A) Experimental plan, showing groups of mice (n = 5 each) given postexposure prophylaxis with RG7834 10 mg/kg p.o. b.i.d. or sham treated with vehicle only for 5 d commencing immediately after i.v. challenge with 2 × 10⁷ GE HAV. Mice were followed twice weekly for serum ALT elevation and fecal virus shedding until necropsy on day 19 p.i. (B) Maximum ALT elevation and (C) maximum fecal HAV RNA shedding in mice between virus challenge and necropsy at day 19 p.i. (D) Intrahepatic HAV RNA at necropsy on day 19 p.i. (E) Neutralization of HM175/18f-NLuc virus by fourfold dilutions of serum collected on day 19 p.i. from mice receiving RG7834 or vehicle only. Neutralization is shown as inhibition of NLuc expressed by Huh-7.5 cells 72 h p.i. with serum-virus mixtures, relative to sera from 4 HAV-naïve mice (mean 6.6 × 10² light units). (F) Representative H & E-stained liver sections from mice receiving postexposure prophylaxis with RG7834 or vehicle only as in panel A. Arrows indicate apoptotic hepatocytes. Statistical testing of ALT and HAV RNA levels by Mann-Whitney test. Dashed horizontal lines indicate limits of detection. Error bars represent SDs.

Fig. 5.

Model explaining RG7834 inhibition of hepatovirus RNA synthesis. TENT4A/B is recruited to ZCCHC14 bound to stem-loop Vb within the 5′ UTR, and recognizes the 3′ end of the polyadenylated HAV genome, providing a protein bridge that functionally circularizes the genome, thereby promoting viral RNA synthesis. RG7834 binds to and interrupts the interaction of TENT4A/B with ZCCHC14, disrupting circularization of the genome and impeding replication.