METHOD FOR INHIBITING INFECTION AND ACTIVATION OF VIRUS

Abstract

The present invention relates to a technology for preventing or treating virus infection and infectious disease by inducing a mixed tailing regarding an RNA virus.

Claims

1.-6. (canceled)

7. A method for screening a pharmaceutical composition for prevention of viral infection or treatment of viral infection symptoms, the method comprising the steps of: (a) transfecting a virus into host cells; (b) treating the host cells with a drug candidate material; and (c) analyzing an RNA tailing level of a test group treated with the drug candidate material and comparing the RNA tailing level with that of a control treated without the drug candidate material, wherein the drug candidate material is selected as an effective candidate when the RNA tailing level of the test group is relatively reduced compared to the non-treated control.

8. A method for preparing a virus-resistant cell, the method comprising a step of knocking out expression of at least one of TANT4A and TENT4B in a cell isolated from a living body.

9. A method for stabilizing an RNA sequence, the method comprising a step of inserting a stem-loop sequence including a pentaloop structure consisting of the RNA sequence represented by the following general formula into a target RNA sequence: TABLE-US-00006 [General Formula] 5′-CNGGN-3′, wherein N's are each independently selected from adenosine (A), uracil (U), guanine (G) and cytosine (C).

10. A method for preventing viral infection or treating viral infection symptoms, the method comprising administering to a subject in need thereof, a pharmaceutical composition comprising an inhibitor against at least one of TENT4A and TENT4B, both responsible for mixed tailing of viral RNA, wherein the inhibitor is siRNA or shRNA inhibiting expression of at least one of TENT4A and TENT4B, or an antibody or an antigen binding fragment thereof inhibiting activity of at least one of TENT4A and TENT4B.

11. The method of claim 10, wherein TENT4A and TENT4B are derived from a subject which is at risk of viral infection or has been infected with virus.

12. The method of claim 10, wherein the virus is a virus which undergoes RNA tailing induced by at least one of TENT4A and TENT4B.

13. The method of claim 12, wherein the virus is a virus belonging to the family Hepadnaviridae or Herpesviridae.

14. The method of claim 13, wherein the virus belonging to the family Hepadnaviridae is any one selected from the group consisting of hepatitis B virus, ground squirrel hepatitis B virus, woodchuck hepatitis B virus, duck hepatitis B virus, and heron hepatitis B virus.

15. The method of claim 13, wherein the virus belonging to the family Herpesviridae is any one selected from the group consisting of Iltovirus, Mardivirus, Scutavirus, Simplexvirus, Varicellovirus, Cytomegalovirus, Muromegalovirus, Proboscivirus, Roseolovirus, Lymphocryptovirus, Macavirus, Percavirus, and Rhadinovirus.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0062] FIGS. 1a, 1b, 1c, 1d, 1e, 1f, 1g and 1h show extensive mixed tailing of viral RNAs. a, The fraction of guanylated tails of cellular and viral mRNAs was calculated for mRNAs with a poly(A) tail length of 25 nt or more (n=2 TAIL-seq experiments). b, The fraction of guanylated tails of cellular RNAs, HCMV RNA and RNA2.7 (VRNA2.7) was calculated for RNAs with a poly(A) tail length of 25 nt or more (n=1 TAIL-seq experiments). c, Global poly(A) tail length distributions of cellular (gray) and HBV (yellow) mRNAs. Medians are marked by dashed vertical lines and the median length is shown in parentheses. d, Global poly(A) tail length distributions of cellular (gray), HCMV (yellow) and VRNA2.7 (orange) RNAs. Medians are marked by dashed vertical lines and the median length is shown in parentheses. e, Fraction of guanylated tails of HBV mRNAs with a poly(A) tail length of 25 nt or more in TENT4-depleted HepG2.2.15 cells (n=1 TAIL-seq experiment). f, Fraction of guanylated tails of HCMV RNAs with a poly(A) tail length of 25 nt or more in TENT4-depleted, HCMV-infected HFF cells (n=1 TAIL-seq experiment). g, HBV mRNA levels in TENT4-depleted HepG2.2.15 cells were measured by RT-qPCR using two separate primer sets. Data are represented as mean±s.e.m. (n=3 independent experiments). **P<0.01; two-sided Student's t-test. h, Half-lives of HBV mRNAs were measured by RT-qPCR using two separate primer sets. Data are represented as mean±s.e.m. (n=3 independent experiments). GAPDH mRNA was used for normalization.

[0063] FIGS. 2a and 2b show data obtained from experiments demonstrating that HBV RNAs are the major mRNA substrates of TENT4 via the PRE of HBV. a, Standardized read coverage of fCLIP-seq libraries across the HBV genome (NCBI U95551.1). The PRE region of HBV is highlighted in yellow. The HBV transcripts and their respective coding sequence regions are shown below as a reference. Input and NMG libraries were used as negative controls. b, Scatter plot of abundance (x axis) and enrichment score (y axis) of fCLIP-seq peak clusters. HBV PRE is shown as a red dot. A size-matched input library was used as a negative control for this analysis.

[0064] FIGS. 3a, 3b, 3c, 3d, 3e, 3f, 3g and 3h shows that the stem-loop structure of PRE is necessary but may not be sufficient for TENT4-dependent RNA tailing of HBV mRNA. a, Schematic view of firefly luciferase reporter constructs harboring subregions and mutants of PRE in their 3′ UTRs. Blue, La-binding motif; green, stem-loop alpha; and red asterisk, region of point mutation. b, Firefly luciferase activity of reporters containing subregions of PRE in parental HeLaT (gray) and TENT4 KO (red) cells. Data are represented as mean±s.e.m. (n=7 independent experiments for PRE, PREα and PREβ, n=6 independent experiments for all others). Luciferase activities of both renilla and control vector were used for normalization. *P<0.01, **P<0.001; two-sided t-test. c, Distributions of poly(A) tail length of reporter RNAs measured by Hire-PAT assay. The reporter constructs containing PREα and PREβ sequences were transfected into HEK293T cells after TENT4 depletion. d, Firefly luciferase activity of PRE reporter constructs in parental or TENT4 KO cells with ectopically expressed wild-type TENT4 or a catalytically inactive mutant. Data are represented as mean±s.e.m. (n=3 independent experiments). Both renilla and control vector were used for normalization. *P<0.05, **P<0.01, ***P<0.001; two-sided Student's t-test. e, Schematic diagram of firefly luciferase reporter constructs harboring subregions of WPRE in their 3′ UTRs. f, Firefly luciferase activity of WPRE reporter constructs in TENT4 KO cells. Data are represented as mean±s.e.m. (n=3 independent experiments for Wβ, n=7 independent experiments for all others). Luciferase activities of both renilla and control vector were used for normalization. *P<0.05, **P<0.01; two-sided Student's t-test. g, Stem-loop structures of HBV PRE, mutant HBV PRE and WPRE. h, Firefly luciferase activity of mutant PRE reporter constructs shown in a in TENT4 KO cells. Data are represented as mean±s.e.m. (n=4 independent experiments). Luciferase activities of both renilla and control vector were used for normalization. *P<0.05, **P<0.01; two-sided Student's t-test.

[0065] FIGS. 4a, 4b, 4c, 4d and 4e show that the stem-loop structure of HCMV RNA2.7 is the cis-acting RNA element responsible for TENT4-dependent regulation. (a) RT-qPCR of indicated RNAs after formaldehyde crosslinking and immunoprecipitation with anti-TENT4A antibody in RNA2.7-transfected HEK293T cells. Data are represented as mean±s.e.m. (n=3 independent experiments). Immunoprecipitation with NMG was used for normalization. **P<0.01, ***P<0.001; two-sided Student's t-test. (b) Schematic view of firefly luciferase reporter constructs possessing subregions and mutants of HCMV RNA2.7 in their 3′ UTRs. Green, stem-loop and red asterisk, point mutations. Stem-loop structure of HCMV RNA2.7 and the mutation are shown on the left. In the 1E mut (1E mutation) construct, the red nucleotides in the loop and the C at the base of the loop are mutated as indicated. FRG, fragment. c-e, Firefly luciferase activity of reporters in parental and TENT4 KO cells. Data are represented as mean±s.e.m. Reporters with subregions of RNA2.7 (n=3 independent experiments) (c), partial deletion mutants of fragment 1 (1-513) of RNA2.7 (n=3 independent experiments) (d) and construct 1D, 1E and stem-loop mutant of 1E (n=3 independent experiments) (e). Luciferase activities of both renilla and control vector were used for normalization. *P<0.05, **P<0.01; two-sided Student's t-test.

[0066] FIGS. 5a, 5b, 5c and 5d show that SAM domain-containing proteins, including ZCCHC14, bind to the stem-loop structure and regulate HBV mRNAs. a, Mass spectrometry analysis of SL2.7 RNA pulldowns from HEK293T cell lysates (n=2 mass spectrometry experiments). The SL2.7 mutant and “bead only” samples were used as negative controls. Protein enrichment score was calculated as the log.sub.2 fold change of the LFQ (label-free quantification) intensity from the SL2.7 sample over the mutant data. Spectra counts were used to filter for candidates of enrichment. b, Top, phylogenetic tree of proteins with putative RNA-binding SAM domains. Saccharomyces cerevisiae (Sc), Schizosaccharomyces pombe (Sp), Candida albicans (Ca), C. elegans (Ce), Drosophila melanogaster (Dm), Danio rerio (Dr), Xenopus laevis (XI), Mus musculus (Mm), and Homo sapiens (Hs). The asterisks indicate manual gene name assignment based on the phylogenetic tree. UniProt IDs are listed in the method section. Bottom, domain architecture of three human proteins with SAM domains. Orange, SAM domain; purple, CCHC zinc finger (ZnF) domain. c, Pulldown of biotinylated SL2.7 from HEK293T cells, followed by western blot using the indicated antibodies. The asterisk indicates a cross-reacting band. d, HBV mRNA levels in ZCCHC14-depleted and SAMD4A/B-depleted HepG2.2.15 cells were measured by RT-qPCR using two sets of primers. Data are represented as mean±s.e.m. (n=4 independent experiments). *P<0.05, **P<0.01, ***P<0.001; two-sided t-test. Uncropped blots for c and source data for graphs in d are available online.

[0067] FIGS. 6a, 6b, 6c, 6d, 6e, 6f and 6g show that cytoplasmic ZCCHC14 recruits TENT4 to protect RNAs with the stem-loop structure. a, The fraction of guanylated tails of HBV mRNAs was calculated for mRNAs with a poly(A) tail length of 25 nt or more in ZCCHC14-depleted HepG2.2.15 cells. b, Global poly(A) tail length distributions of HBV mRNAs in control (gray) and ZCCHC14-depleted HepG2.2.15 cells (red). Medians are marked by dashed vertical lines and the median length is shown in parentheses. c, Firefly luciferase activity of reporters harboring HCMV RNA2.7 1E, HBV PREα, WHV WPRE (Wγ+Wα) and their respective mutants (1E mut, α mut2 and Wγ+Wα mut) in ZCCHC14-depleted HEK293T cells. Data are represented as mean±s.e.m. (n=4 independent experiments for PREα and a mut2, n=3 independent experiments for all others). Importantly, the WPRE mutant reporter (Wγ+Wα mut) consists of point mutations in both WPREα stem-loop structures (Supplementary FIG. 6b). Luciferase activities of both renilla and control vector were used for normalization. *P<0.05; two-sided Student's t-test. d, RT-qPCR of indicated RNAs after immunoprecipitation with anti-ZCCHC14 antibody in HepG2.2.15 cells. Data are represented as mean±s.e.m. (n=3 independent experiments). Immunoprecipitation with normal rabbit IgG was used for normalization. *P<0.05; two-sided Student's t-test. e, Localization of ZCCHC14 protein was examined by subcellular fractionation and western blotting in HeLaT cells: cytoplasm (Cyto), membrane (Memb) and nucleus (Nuc). GAPDH (cytoplasm), GM130 (Golgi) and Histone H3 (nucleus) were used as fraction markers. f, Immunoprecipitation with anti-TENT4A and anti-TENT4B followed by western blot was used to measure the physical interaction between TENT4A/B and ZCCHC14 in HeLaT cytosol extracts. Cell extracts were treated with RNase A. GAPDH was used as a loading control. g, Proposed model. HBV mRNAs and HCMV RNA2.7 recruit the TENT4-ZCCHC14 complex via the CNGGN pentaloop to induce targeted mixed tailing, which subsequently protects the viral RNAs from cellular decay factors.

[0068] FIGS. 7a, 7b, 7c, 7d, 7e, 7f, 7g, 7h, 7i, 7j, 7k and 7m show extensive 3′ end tail modification of viral RNAs. a, HBV mRNAs are highly mixed tailed (G: guanylation, U: uridylation, C: cytidylation) in HepG2.2.15 cells. Fraction of each modification includes their respective terminal and internal modifications. Fraction of cytidylated and uridylated tails of viral mRNAs were calculated for those with poly(A) tail length 25 nt (n=2 TAIL-seq experiments). b, HCMV RNAs are also highly mixed tailed in HCMV-infected HFF cells. HCMV RNA2.7 (VRNA2.7) is substantially mixed tailed in particular. (n=1 TAIL-seq experiment) c, Gene-level analysis of guanylated tails of cellular (grey) and viral (yellow) mRNAs in HCMV-infected HFF cells. d, siRNA knockdown confirmation by RT-qPCR. e, Decrease in the fraction of mixed tail of HBV mRNAs in TENT4-depleted HepG2.2.15 cells. f, TENT4A/B expression is induced in HCMV-infected HFF cells. TENT4A and TENT4B RNA levels in HCMV-infected HFF cells were measured by RT-qPCR (n=1). g, Decrease in the fraction of mixed tail of HCMV RNA2.7 in TENT4-depleted HCMV-infected HFF cells. h, HBV mRNAs exhibit shorter poly(A) tail length in TENT4-depleted HepG2.2.15 cells. The medians are marked by dashed vertical lines and shown in parentheses. i, Global poly(A) tail length distribution of viral mRNAs after TENT4 depletion in HCMV-infected HFF cells. j, Poly(A) tail length distribution of HCMV RNA2.7 after TENT4 depletion. k, HBV transcripts and their respective CDS regions are shown. Bars indicate the position of RT-qPCR amplicons (#1-3). l, Tail modifications of HBV mRNAs across poly(A) tail length in TENT4-depleted HepG2.2.15 cells. m, Half-life of HCMV RNA2.7 was measured by RT-qPCR (n=3 independent experiments) in HCMV-infected HFF cells after TENT4 depletion. GAPDH mRNA was used for normalization. Mean was calculated. Data for graphs in a-b, d-g, and m are available as source data.

[0069] FIGS. 8a and 8b show analysis and confirmation of TENT4A/B and HBV mRNA interaction. a, TENT4B-bound RNA fragments are again enriched within the PRE region of HBV (highlighted in yellow). Standardized read coverage of fCLIP-seq libraries across the HBV genome (NCBI: U95551.1). X-axis scale is the same as in FIG. 2a. b, Immunoprecipitation with anti-TENT4A and anti-TENT4B followed by RT-qPCR was used to measure the enrichment of TENT4-bound RNA in HepG2.2.15 cells (n=3 independent experiments). Immunoprecipitation with normal mouse IgG was used for normalization. Mean was calculated and Error bars represent SEM. **P<0.01, ****P<0.0001; two-sided t test. Data for graphs in b is available as source data.

[0070] FIGS. 9a, 9b, 9c, 9d, 9e and 9f show confirmation of TENT4-dependent regulation of PRE reporters. a, Knockout confirmation by western blotting. GAPDH was used as a loading control. b, RNA levels of firefly reporters harboring subregions of PRE were measured by RT-qPCR (PRE: PREα: PREβ, n=4; αΔ1, n=2; αΔ2, n=1 independent experiment). mRNA levels of both renilla and control vector were used for normalization. Mean was calculated and Error bars represent SEM. *P<0.05; two-sided t test. c, siRNA knockdown confirmation by western blotting in HEK293T cells. GAPDH was used as a loading control. d, Firefly luciferase activity of PRE reporter constructs in TENT4-depleted HEK293T cells (PRE: PREα: PREβ, n=7; αΔ1, n=6; αΔ2, n=5 independent experiments). Luciferase activities of both renilla and control vector were used for normalization. Mean was calculated and Error bars represent SEM. *P<0.05, **P<0.01, ***P<0.001; two-sided t test. e, Over-expression confirmation by western blotting in KO cells. GAPDH was used as a loading control. f, RNA levels of firefly reporters used in rescue experiments in TENT4 KO cells with ectopically expressed TENT4 wild-type or mutant (n=3 independent experiments) were measured by RT-qPCR (n=3 independent experiments). mRNA levels of both renilla and control vector were used for normalization. Mean was calculated and Error bars represent SEM. *P<0.05, **P<0.01; two-sided t test. Uncropped blots for panels a, c, and e and data for graphs in b, d, and f are available as source data.

[0071] FIGS. 10a, 10b, 10c, 10d and 10e show confirmation of TENT4-dependent regulation of RNA2.7 reporters. a, Immunoprecipitation with anti-TENT4A was confirmed by western blotting in RNA2.7-transfected HEK293T cells. Normal mouse IgG (NMG) was used as a negative control. RNA2.7 transcript is shown with bar indicating the position of RT-qPCR amplicon. b-e, RNA levels of firefly reporters harboring b, subregions of RNA2.7 (n=3 independent experiments), c, deletion constructs of fragment 1 (1-513) of RNA2.7 (n=3 independent experiments), d, construct 1D (314-413), 1E (414-513) and stem-loop mutant 1E (n=3 independent experiments) and e, construct SL2.7 (414-463) and controls (n=3 independent experiments) in parental and TENT4 KO cells. mRNA levels of both renilla and control vector were used for normalization. Mean was calculated and error bars represent SEM. *P<0.05, **P<0.00.0; two-sided t test. Uncropped blots for panel a and data for graphs in b-e are available as source data.

[0072] FIG. 11 shows confirmation of ZCCHC14 and SAMD4A/B knockdown. siRNA knockdown confirmation by RT-qPCR (n=4 independent experiments) and western blotting in HepG2.2.15 cells. Mean was calculated and Error bars represent SEM. ****P<0.0001; two-sided t test. GAPDH was used as a loading control.

[0073] FIGS. 12a, 12b, 12c, 12d and 12e show confirmation of ZCCHC14 knockdown, stem-loop-dependent tail regulation, and immunoprecipitation. a, siRNA knockdown confirmation by western blotting in ZCCHC14-depleted HEK293T cells. GAPDH was used as a loading control. The dashed line indicates discontinuous lanes from the same gel. b, Stem-loop structures of WPREα and their respective mutants. c, Distributions of poly(A) tail length of reporter RNAs measured by Hire-PAT assay. The reporter constructs were transfected into HEK293T cells after ZCCHC14 depletion. d, Immunoprecipitation with anti-ZCCHC14 was confirmed by western blotting in HepG2.2.15 cells. Normal rabbit IgG (NRG) was used as a negative control. e, Immunoprecipitation with anti-TENT4A and anti-TENT4B followed by western blotting was used to measure the physical interaction between TENT4A/B and ZCCHC14 in HepG2.2.15 and primary HFF cells. Cell extracts were treated with RNase A. Normal mouse IgG (NMG) was used as a negative control. The asterisks indicate cross-reacting bands. Uncropped blots for panel a and d-e are available as source data.

[0074] FIG. 13 shows a stem-loop structure containing a pentaloop.

[0075] The insertion of the stem-loop structure of the present disclosure into a target RNA sequence may be performed using conventionally well-known methods without particular limitations.

DETAILED DESCRIPTION

[0076] A better understanding of the present disclosure may be obtained via the following examples which are set forth to illustrate, but are not to be construed as limiting the present disclosure.

EXAMPLES

Experimental Methods

Example 1: TAIL-seq Library Preparation and Data Processing

[0077] TAIL-seq was conducted as previously described [3]. In brief, ˜50 μg of DNasel (Takara, 2270A) treated total RNA (>200 nt) was used to deplete ribosomal RNA twice by Ribo-Zero kit (Epicentre, MRZH11124 (discontinued) for primary HFF library or Illumina, TruSeq Stranded Total RNA Library Prep Human/Mouse/Rat, 20020596 for HepG2.2.15 library). The rRNA-depleted RNAs were ligated to the 3′ adapter and partially fragmented by RNase T1 (Ambion). The fragmented RNAs were pulled-down with streptavidin beads (Invitrogen), 5′ phosphorylated and purified by 6% urea-PAGE gel (500-1,000 nt). The purified RNAs were ligated to the 5′ adapter, reverse-transcribed and amplified by PCR. The libraries were sequenced by paired-end run (51×251 cycles) on the Illumina platform (MiSeq) with PhiX control library v.3 (Illumina) and spike-in mixture. TAIL-seq sequencing data have been deposited to the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database with accession number GSE146600. TAIL-seq analysis was carried out using Tailseeker v.3.1.5 [9]. Briefly, read 1 was used for gene identification, and read 2 was used to detect 3′ end modifications and measure the length of the poly(A) tail. For each TAIL-seq library, read 1 was mapped to human genome GRCh38 with STAR 2.5.2b [50] and either HBV genome NCBI U95551.1 or HCMV genome NCBI GU937742.2 with bowtie2.2.6 [51]. Only TAIL-seq fragments with poly(A) tail length of at least 25 nt were used as done previously[6]. Of read 2, 3′ end 10-mer sequences with terminal or internal mono-guanylation (or -cytidylation, -uridylation) were examined to estimate the fraction of nonadenosine incorporation of the poly(A) tail.

Example 2: fCLIP-seq Library Preparation and Data Processing

[0078] fCLIP-seq was performed as previously described with minor modifications [16,52]. In brief, HepG2.2.15 cells on two 150 mm dishes were crosslinked with 0.1% paraformaldehyde (Pierce, 28906), collected and lysed. Fifteen μg of each antibody (NMG, Santa Cruz, sc-2025; TENT4A, laboratory made; TENT4B, laboratory made) was conjugated to protein A and G Sepharose beads (1:1 mix, total 20 μl, GE Healthcare, 17-5138-01 and 17-0618-01, respectively). Lysates were incubated with antibody-conjugated beads, and RNAs were purified from the eluates followed by DNasel treatment and washing. Then, 1 μμg of input RNAs or RNAs from each sample were prepared for the libraries. rRNAs were depleted twice by Ribo-Zero kit (Illumina, TruSeq Stranded Total RNA Library Prep Human/Mouse/Rat, 20020596). The rRNA-depleted RNAs were ligated to the 3′ adapter and purified by 6% urea-PAGE gel (80-500 nt, corresponding to 50-470 nt RNAs that were fragmented by sonication) followed by 5′ phosphorylation, 5′ adapter ligation, reverse-transcription, and PCR amplification. The fCLIP-seq libraries were sequenced by paired-end run (151×151 cycles) on Illumina platform (MiSeq) with PhiX control library v.3 (Illumina). fCLIP-seq sequencing data have been deposited to the NCBI GEO database with accession number GSE146597.

[0079] For each fCLIP-seq library, pair-end reads were assembled with pear v.0.9.10[53], and aligned to human genome GRCh38 with STAR 2.5.2b[50] and HBV genome NCBI U95551.1 with bowtie2.2.6[51]. Raw read coverage of fCLIP-seq libraries were computed with bedtools v.2.26.0. fCLIP-seq peak clusters of cellular mRNAs were estimated using Piranha v1.2.1 [54] with a cluster size of 200 nt (−z 200), internal normalization (−n), log transformation (−l), and the NMG library as the covariate. Of note, HBV PRE spans approximately 400 nt. Enrichment scores of peak clusters were calculated by the log 2 ratio of the reads of the TENT4 fCLIP-seq library and of the NMG library. The input library was used as a negative control of this analysis[55]. To account for technical background of the read coverage, the Hodges-Lehmann estimator over the HBV genome was used. Specifically, given the HBV genome of length n, the median of the following set is computed:

[00001]$\left\{\left(\mathrm{Xi}+\mathrm{Xj}\right)/2:1\le i\le j\le n\right\}$

[0080] where Xi and Xj are the raw read coverages at positions i and j, respectively. The raw read coverages were then normalized by this Hodges-Lehmann estimator and visualized using the ggplot2 R package.

Example 3: Cell Culture, Transfection and Actinomycin D Treatment

[0081] All cell lines used in this study were tested negative for mycoplasma. HeLaT was derived from HeLa (gift from C.-H. Chung at Seoul National University) with a null mutation in the TUT4 gene. HeLaT and HEK293T (gift from S. Kim at Seoul National University) cells were authenticated by ATCC (STR profiling). HepG2.2.15, HeLaT and HEK293T cells were grown in DMEM (Welgene, LM 001-05) containing 10% FBS (Welgene, S001-01). Primary HFF (ATCC SCRC-1041) cells were grown in DMEM (HyClone) containing 10% FBS (HyClone), GlutaMAX-I (Gibco), and penicillin-streptomycin (Gibco). Before transfection, primary HFF were grown in antibiotics-free media.

[0082] For combinatorial knockdown, equal amounts of small-interfering RNAs or GAPmer against target genes were mixed to have a final concentration denoted as discussed in the following. For TENT4A and TENT4B knockdown, HepG2.2.15 cells were transfected with 100 nM siRNAs by Lipofectamine RNAiMAX (Invitrogen) at days 0, 2 and 4 and collected on day 6. For ZCCHC14, SAMD4A and SAMD4B knockdown, HepG2.2.15 cells were transfected with 100 nM siRNAs or GAPmer by Lipofectamine RNAiMAX (Invitrogen) at days 0 and 2, and collected on day 4. HEK293T cells were transfected with the RNA2.7-expressing pCK vector for RNA immunoprecipitation by Lipofectamine 3000 at day 0 and collected on day 2. Primary HFF cells were transfected with 20 nM siRNAs by DharmaFECT 1 (Dharmacon) at days 0 and 2 and collected on day 5. The sequence information of siRNAs is shown in Table 1.

TABLE-US-00001 TABLE 1 siRNAs (sense strand) siCont only for CCUACGCCACCAAUUUCGU primary HFF siCont AccuTarget ™ Negative Control siRNA from Bioneer Inc (Seoul, South Korea) siTENT4A-1 CUACGGUACCAAUAAUAAA siTENT4A-2 GGAAGAAUCAUCAAAGUAA siTENT4A-3 CCAAACAGAGACGCCGAAA siTENT4A-4 GCGAAUAGCCACAUGCAAU siTENT4A only for ON-TARGETplus SMART pool HepG2.2.15 (Dharmacon) siTENT4B-1 GGACGACACUUCAAUUAUU siTENT4B-2 GGCCUUUGAUUAUGCCUAC siTENT4B-3 GCGCUGACGUCCAGAUAUU siTENT4B-4 CCUAUUGCAGAGGGACCUU siZCCHC14-1 GCAUUUUAUGUGGAGCGAA siZCCHC14-2 CCUUCUCACGUGUUGAAAA siZCCHC14-3 GAAUAAAUUUGAGUCUCUU siSAMD4A-1 GGAUAUCGACAGCAAAGAA siSAMD4A-2 CAUCAUGAAACAAGGAAGA siSAMD4B-1 CACUAGAGAUGCAGAACUA siSAMD4B-2 CCUACUCAAUCGAGAGCAA siSAMD4B-3 AGGAGAACAUCACCAGUUA

[0083] The present inventors purchased siRNAs against TENT4A from ON-TARGETplus SMARTpool (Dharmacon) and GAPmer against SAMD4A from Antisense LNA GapmeR Standard (Qiagen, 339511 LG00236046-DDA) for better knockdown efficiency in HepG2.2.15. Actinomycin D (Sigma, A9415, 4 μg ml-1) was added to HepG2.2.15 cells to block transcription and cells were gathered at the denoted time.

Example 4: HCMV Infection

[0084] Infectious HCMV particles were generated by transfecting HCMV Toledo BAC DNAs (received from T. Shenk, Princeton University) into primary HFF cells by electroporation (Invitrogen, Neon). When a 100% cytopathic effect was observed, cell culture supernatants were gathered, centrifuged to remove cell debris, and stored at −80° C. in 1 ml aliquots. To titrate the viral stocks, primary HFFs grown on cover glass were inoculated with diluted virus stocks for 1 hour and fixed with 3.7% formaldehyde at 24 hours post inoculation. Cells were permeabilized using 0.1% Triton X-100, incubated with blocking buffer (2% bovine serum albumin in phosphate-buffered saline), stained with HCMV 1E1 antibody (MAB810R; Millipore) followed by FITC-conjugated anti-mouse antibody (115-095-146; Jackson Laboratories), and mounted with DAPI-containing solution (H-1200; Vector Laboratory). The number of HCMV IE1-positive cells were counted and the multiplicity of infection (MOI) was determined by calculating the ratio of IE1-positive cells to total cells. For HCMV infection, primary HFF cells were incubated with HCMV (MOI, 2) diluted in serum-free DMEM for 1 hour, washed with PBS, and incubated in DMEM for further culture.

Example 5: Plasmid Construction

[0085] Previously described wild-type TENT4A and TENT4B isoforms (792 and 666 aa, respectively)[6] except the FLAG tag were used. A point mutation for catalytic dead mutant was introduced (D352A for TENT4A, D326A for TENT4B, respectively). For PRE reporter constructs, PRE (1,154-1,687 bp), PREα (1,154-1,351 bp), PRE (1,352-1,687 bp), αΔ1 (1,276-1,351 bp), and αΔ2 (1,154-1,275 bp) were amplified from HepG2.2.15 complementary DNA and introduced into the 3′ UTR region of firefly luciferase mRNA sequence in pmirGLO-3XmiR-1 vector[6]. Point mutations for the La-binding motif and stem-loop a were the same as the previously described mutants[30]. α mutt, α mut2, and a mut3 were generated by PCR-directed mutagenesis. For WPRE reporter constructs, WPRE (1,095-1,670 bp), Wγ+Wα (1,095-1,507 bp), Wα+Wβ (1,300-1,670 bp), Wγ (1,093-1,299 bp), Wα (1,300-1,507 bp), and Wβ (1,508-1,670 bp) were amplified from pLenti-CMV vector (Addgene, 17492) and introduced into 3′ UTR of firefly luciferase in pmirGLO-3Xmir-1 vector. Point mutations for the CNGGN pentaloops of WPREα were designed in a similar manner to the PRE mutants. For RNA2.7 construct, FRG1 (1-513 bp), FRG2 (414-913 bp), FRG3 (814-1,313 bp), FRG4 (1,214-1,713 bp), FRG5 (1,614-2,113 bp), FRG6 (2,014-2,513 bp), 1ΔA (1-513 bp, Δ1-100 bp), 1ΔB (1-513 bp, Δ101-200 bp), 1ΔC (1-513 bp, Δ201-300 bp), 1ΔD (1-513 bp, Δ301-400 bp), 1ΔE (1-513 bp, Δ401-513 bp), 1D (301-400 bp), 1E (401-513 bp), and SL2.7 (414-463 bp) were amplified from HCMV-infected primary HFF cDNA and introduced into the 3′ UTR region in pmirGLO-3Xmir-1 vector. The 1E and SL2.7 mutant constructs were generated in the same way as the PRE mutants. For RNA immunoprecipitation, RNA2.7 (1-2,513 bp) were amplified from pmirGLO-3Xmir-1 vector and subcloned into a pCK vector.

Example 6: TENT4A and TENT4B Knockout Cell Preparation

[0086] For ablation of TENT4A and TENT4B, CRISPR-Cas genome editing was carried out as previously described[56] with minor adaptations that T7 Endonuclease l (NEB, M0302) was used instead of Surveyor nuclease. The 6E4 HeLaT cells on 24-well plate were transfected by using Metafectene (Biontex, T020) with 300 ng pSpCas9(BB)-2A-GFP-px458 plasmid (Addgene no. 48138) with single-guide RNA (agccacgttgtgcttccgcg, PAM sequence is GGG) against TENT4A first. This HeLaT parental cell was derived from HeLa and contained a null mutation in the TUT4 gene. After single cell screening and Sanger sequencing to confirm knockout, parental and changed genomic sequences are listed in Table 2, and inserted sequences are marked in bold and underlined.

TABLE-US-00002 TABLE 2 Genomic sequences of TENT4A and TENT4B KO HeLaT cells Parental GAGCAAGCCCTGCGGAAGCACAACGTGGCT TENT4A GAGCAAGCCCTGCGGGAAGCACAACGTGGCT KO GAGCAAGCCCTGCGCAAGATGGACGGCACCGAGGAACTGC TCGTGAAGCTGAACAGAGAGGACCTGCTGCGGAAGCAGCG GACCTTCGACAACGGCAGCATCCCCCACCAGATCCACCTG GGAGAGCTGCACGCCATTCTGCGGCGGCAGGAAGAAGCAC AACGTGGCT GAGCAAGCCCTGCGCCGGTTCTTCCGTCTGGTGTATCTTC TTCTGGCGGTTCTCTTCAGCCGGGTGGCCTCGGCTGTTTC GCCGCTGTCGAACAGCAGGGCTCCGATCAGGTTCTTCTTG ATGCTGTGCCGGTCGGTGTTGCCCAGCACCTTGAATTTCT TGCTGGGCACCTTGTACTCGTCGGAAGCACAACGTGGCT Parental TAGTGACATCGACCTAGTGGTGTTTGGGAA TENT4B TAGTGACATCGACCTAGT--TGTTTGGGAA KO TAGTGACATCGACCTAGT-A-AGTTTGGGAA TAGTGACATCGACCTAGTGGTTGTTTGGGAA

[0087] Using TENT4A knockout cells, TENT4B was further ablated with sgRNA (gacatcgacctagtggtgtt, PAM sequence is TGG) against TENT4B. Changed genomic sequences are also listed in Table 2, and inserted and deleted sequences are marked in red and dashes, respectively.

Example 7: TENT4A and TENT4B Antibody Preparation

[0088] Mice were immunized with TENT4A antigen (126-571 aa from 792 aa isoform, NCBI NP_008930.2) or TENT4B antigen (106-533 aa from 666 aa isoform, NCBI XP 006721308.1) and killed. Collected splenocytes were fused with Sp2/O myeloma, and hybridoma cells producing high affinity antibodies were selected. Ascites was obtained after intraperitoneal injection of hybridoma cells to mice (Youngin Frontier Inc.), and used for the immunoprecipitation experiment.

Example 8: Western Blotting

[0089] HepG2.2.15 cells were lysed in RIPA buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 10% NP-40 (Sigma, 74385), 1% sodium deoxycholate (Sigma, D6750), 1% SDS (Ambion, AM9823)). Other cells were lysed in Buffer D (200 mM KCl, 10 mM Tris-HCl (pH 8.0), 0.2 mM EDTA, 0.1% Triton X-100 (Promega, H5142)). Roughly 60 μg of lysate was loaded on 10% or 8-16% (Novex) SDS-PAGE gel with the ladder (Thermo, 26616 and 26619). After transferring to a methanol-activated polyvinylidene difluoride membrane (Millipore), the membrane was blocked in PBS-T containing 5% skim milk, probed with primary antibodies and washed three times with PBS-T. Anti-mouse or anti-rabbit HRP-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) were incubated and washed three times with PBS-T. Chemiluminescence was conducted with West Pico or Femto Luminol reagents (Thermo), and the signals were detected by ChemiDoc XRS+ System (BioRad). Anti-TENT4A (1:500, Invitrogen, PA5-61302; 1:250-500, Atlas Antibodies, HPA045487), anti-TENT4B (1:500, Invitrogen, PA5-60177; 1:500, Atlas Antibodies, HPA042968; 1:500, laboratory made), anti-ZCCHC14 (1:1,000-2,000, Bethyl Laboratories, A303-096A), anti-GAPDH (1:1,000-2,000, Santa Cruz, sc-32233), anti-SAMD4B (1:500, Invitrogen, PA5-53490), anti-GM130 (1:500, BD Bioscience, 610822) and anti-Histone H3 (1:2,000, Cell signaling, 4499) were used as the primary antibodies.

Example 9: RT-qPCR

[0090] Total RNA was extracted using Trizol reagent (Invitrogen, 15596018) or Maxwell 16 LEV simplyRNA Tissue kit (Promega, AS1280), treated with DNasel, purified with RNeasy MinElute Cleanup Kit (Qiagen), and reverse-transcribed with RevertAid reverse transcriptase (Thermo) and oligo dT for luciferase assay samples or random hexamer (Invitrogen) for other samples. mRNA levels were measured with SYBR Green assays (Thermo) and StepOnePlus Real-Time PCR System (Applied Biosystems) or QuantStudio 3 (Applied Biosystems). The list of RT-qPCR primers is given in Table 3.

TABLE-US-00003 TABLE 3 qPCR primers qPCR-ZCCHC14-F ACCCCGTCTTTAAGCAGCTC qPCR-ZCCHC14-R GGCACCGTCTCTCTGACTTC qPCR-TENT4A-F CCCACCACTTCCAGAACACT qPCR-TENT4A-R GCTTTCAAAGACGCAGTTCC qPCR-TENT4B-F TCGCAGATGAGGATTCG qPCR-TENT4B-R CTGCTCTCACGCCATTCT qPCR-GAPDH-F CTCTCTGCTCCTCCTGTTCGAC qPCR-GAPDH-R TGAGCGATGTGGCTCGGCT qPCR-FireflyLuc-F CCCATCTTCGGCAACCAGAT qPCR-FireflyLuc-R GTACATGAGCACGACCCGAA qPCR-RenillaLuc-F CTGGACGAAGAGCATCAGG qPCR-RenillaLuc-R TGATATTCGGCAAGCAGGCA qPCR-SAMD4A-F CTGAGCAGCTGCGATGGG qPCR-SAMD4A-R GGTCTGGAGACCAAGAGCTG qPCR-SAMD4B-F TGCAATGACATCCACCTGCT qPCR-SAMD4B-R GTACTCCGACTTGGCCTCTG qPCR-HBV#1-F AACGGCCAGGTCTGTGCCAA qPCR-HBV#1-R CCGCAGTATGGATCGGCAGAG qPCR-HBV#2-F TACATAAGAGGACTCTTGG qPCR-HBV#2-R GCAGACCAATTTATGCCTAC qPCR-HBV#3-F TACTGCGGAACTCCTAGCCG qPCR-HBV#3-R TTGCGGGAGAGGACAACAGAG qPCR-RNA2.7-F CAGCTTTTCTCCCAAACCTCG qPCR-RNA2.7-R AGGAATAATCCGTGCGACCG qPCR-HSPA8-F ACCTACTCTTGTGTGGGTGTT qPCR-HSPA8-R GACATAGCTTGGAGTGGTTCG qPCR-ACTB-F CCTGTACGCCAACACAGTGC qPCR-ACTB-R ATACTCCTGCTTGCTGATCC qPCR-FireflyLuc-F used for AAGGTTGTGGATCTGGATAC RIP-qPCR spike-in qPCR-FireflyLuc-R used for GATTGTTTACATAACCGGAC RIP-qPCR spike-in

Example 10: RNA Immunoprecipitation

[0091] For TENT4A and TENT4B immunoprecipitation with HepG2.2.15 cells, cells from one 150 mm dish were gathered, lysed with lysis buffer (20 mM HEPES (pH 7.0-7.6) (Sigma, H0887), 4% NP-40, 100 mM KCl, 0.1 mM EDTA, 10% glycerol, 1 mM DTT, 20 U ml-1 RNase inhibitor (Ambion, AM2696), 1× protease inhibitor (Calbiochem, 535140)) on ice and then centrifuged. 0.375 μg of each antibody (NMG, Santa Cruz, sc-2025; anti-TENT4A, laboratory made; anti-TENT4B, laboratory made) was conjugated to protein A and G sepharose beads (1:1 mix, total 10 μl). Lysates were incubated with antibody-conjugated beads and then washed with wash buffer (the same lysis buffer but with 2% NP-40). After adding 10 ng of firefly luciferase mRNA to the sample as a spike-in used for normalization, RNAs were purified by TRIzol reagent, treated with DNasel and used for RT-qPCR. A similar method was used for ZCCHC14 immunoprecipitation with HepG2.2.15 cells but also with Turbo DNasel (Ambion, AM2239) treatment during lysis and each 10 μg of antibody (Normal rabbit IgG (Cell signaling, 2729S) and anti-ZCCHC14 (Bethyl Laboratories, A303-096A)). For TENT4A immunoprecipitation with HEK293T cells (one 100 mm dish), a similar method was used with each 6 μg of antibody (NMG and anti-TENT4A) and the cells were also crosslinked (0.1% paraformaldehyde) before collection.

Example 11: Coimmunoprecipitation

[0092] Similar steps were used as in RNA immunoprecipitation but with the addition of Turbo DNasel and RNase A (Thermo, EN0531). Briefly, HepG2.2.15 cells from five 150 mm dishes were lysed and used for immunoprecipitation with 17.5 μg of each antibody (NMG, anti-TENT4A, anti-TENT4B), which was conjugated to protein A and G sepharose beads (1:1 mix, total 25 μl). Primary HFF cells from two 150 mm dishes were lysed and used for immunoprecipitation with each 6 μg of antibody (NMG, anti-TENT4A, anti-TENT4B), which was conjugated to protein A and G sepharose beads (1:1 mix, total 20 μl). RNase A was added to lysates to have 0.2-0.3 μg/μl final concentration. For the coimmunoprecipitation experiment, cytosol extracts from 1 and 1/3 150 mm dishes of HeLaT cells were obtained by subcellular fractionation as described in the following, and then used for immunoprecipitation with 18 μg of each antibody (NMG, anti-TENT4A, anti-TENT4B).

Example 12: Luciferase Assay and Transfection

[0093] The 1E5 HeLaT cells on 24-well plate were transfected with 200 ng pmirGLO plasmids for luciferase reporter assay by Lipofectamine 3000 at day 0 and collected on day 2. For the rescue experiment, 1.5E5 HeLaT parental and TENT4 KO cells on a 24-well plate were transfected with 50 ng of pmirGLO plasmids along with 20 ng of null vector or TENT4A/B (1:1 mix) plasmids by Lipofectamine 3000 at day 0 and collected on day 2. HEK293T cells were transfected with 100 nM siRNAs at Day 0 and 1.5E5 cells on 24-well plate were further transfected with 100 nM siRNAs at day 2 along with 100 or 200 ng of pmirGLO plasmids by Lipofectamine 3000. HEK293T cells were collected on day 4. For luciferase assay, cells were lysed and analyzed with the dual-luciferase reporter assay system (Promega) according to the manufacturer's instructions.

Example 13: Subcellular Fractionation

[0094] The 2E6 HeLaT cells were collected and washed with cold PBS. To obtain cytosolic fraction, cells were lysed with 200 μl of cytosol lysis buffer (0.2 μg/μl digitonin (Sigma, D141), 150 mM NaCl, 50 mM HEPES (pH 7.0-7.6), 0.1 mM EDTA, 1 mM DTT, 20 U/ml RNase inhibitor, 1× protease inhibitor, 1× phosphatase inhibitor). After centrifugation at 2,000 relative centrifugal force at 4° C. for 5 min, supernatant was collected and used as a cytosol fraction. For the membrane and nucleus fractions, subcellular protein fractionation kit (Thermo Scientific, 78840) was used according to the manufacturer's instructions.

Example 14: Hire-PAT Assay

[0095] Hire-PAT assay and signal processing of capillary electrophoresis data were conducted as previously described[3,57] with minor adaptations that total RNAs were G/I tailed by yeast poly(A) polymerase (Thermo Scientific, 74225Z25KU). Poly(A) site of firefly luciferase gene was confirmed by Sanger sequencing, and the forward PCR primer is listed in Table 4.

TABLE-US-00004 TABLE 4 Hire-PAT PCR primer Hire-PAT-FireflyLuc-F GGACAAACCACAACTAGAATG

Example 15: Phylogenetic Analysis

[0096] To identify putative stem-loop binding proteins, homologs of the RNA-binding SAM domain of Drosophila Smaug (UniProtKB Q23972) and human SAMD4A/B (UniProtKB Q9UPU9 and Q5PRF9) were searched using UniProt BLAST with an E value threshold of ten and not gapped[58]. Protein sequences used for the analysis: UniProtKB Q08831 (S. cerevisiae), UniProtKB Q9P6R7 (S. pombe), UniProtKB Q5AI80 (C. albicans), UniProtKB O76699 (C. elegans), UniProtKB Q23972 (D. melanogaster), UniProtKB E7F857, E7FBA1, A0A0R4IUM4 (D. rerio), UniProtKB Q6GLT9, Q5FWP2, A0A1L8GL36 (X. laevis), UniProtKB Q8CBY1, Q80XS6, Q8VIGO (M. musculus) and UniProtKB Q9UPU9, Q5PRF9, Q8WYQ9 (H. sapiens). MUSCLE v.3.8.31[59) with default parameters was used for multiple sequence alignment of the resulting 17 protein sequences. The phylogenetic tree was reconstructed using PhyML v.3.1/3.0 aLRT60 and visualized using TreeDyn v.198.3 [61) on the Phylogeny.fr platform[62]. The edge and leaves of the reconstructed phylogenetic tree were reordered and rooted manually for visualization.

Example 16: RNA Pulldown Assay

[0097] The 3′ TEG (tetraethylene glycol spacer)-biotinylated RNAs of RNA2.7 stem-loop (SL2.7) and stem-loop mutant (SL2.7 mut) were synthesized (Bioneer Inc.), and the sequences are listed in Table 5.

TABLE-US-00005 TABLE 5 RNA oligos for pull-down mass SL 2.7 CACCGCGUUAUCCAUUCCUCGUAGGCUGGU CCUGGGGAACGGGUCGGCGG[Biotin-TEG] SL-2.7 mutant CACCGCGUUAUCCAUUCCUCGUAGGAUUUU CCUGGGGAACGGGUCGGCGG[Biotin-TEG]

[0098] Streptavidin M-270 beads (Thermo Scientific, 65305) for set 1 and streptavidin magnetic beads (NEB, S1420S) for set 2 were washed twice with pulldown buffer (50 mM Tris (pH 8.0), 150 mM NaCl, 5% glycerol, 1 mM DTT, 100 U/ml RNase inhibitor, 1× Protease inhibitor). Next, 10 μg of oligonucleotides of SL2.7 or its mutant were conjugated to streptavidin beads in pulldown buffer overnight at 4° C. and then washed with pulldown buffer. HEK293T cells from four 150 mm dishes were gathered and then lysed with 1.5 ml pulldown buffer followed by sonication. After centrifugation, supernatant was incubated with beads and washed with wash buffer no. 1 (50 mM Tris (pH 8.0), 300 mM NaCl, 5% glycerol), wash buffer no. 2 (50 mM Tris (pH 8.0), 150 mM NaCl, 5% glycerol, 0.1% triton X-100) and pulldown buffer. For elution, the beads were incubated with 60 μl of elution buffer (100 mM (pH 7.5), 4% SDS, 100 mM DTT) at 1,600 r.p.m. in thermomixer for 10 min at 65° C. Then, 10 and 50 μl of eluates were used for western blotting and mass spectrometry, respectively.

Example 17: Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) Analysis

[0099] The immunoprecipitated protein samples containing SDS were subjected to filter-aided sample preparation digestion. Briefly, protein samples were first reduced and alkylated with denaturing ABC buffer (8 M urea in 50 mM ammonium bicarbonate (ABC)). The alkylated samples were placed on preconditioned 30 kDa MWCO Amicon filter (Millipore, UFC5030) and centrifuged for 30 min at 14,000 g. Serial washing was carried out with 200 μl of 8 M urea or 50 mM ABC buffer and centrifuging for 15 min at 14,000 g to remove SDS in the filter unit. Then, protein samples were digested with 2% (w/w) trypsin at 37° C. for overnight. The resulting peptide samples were subject to C18 ziptip clean-up (Millipore, Z720070) and LC-MS/MS analysis. In-house packed long capillary columns (100 cm×75 μm inside diameter) and trap columns (3 cm×150 μm inside diameter) with 3-μm Jupiter C18 particles (Phenomenex) were used for peptide separation. A flow rate of 300 nl/min and a linear gradient ranging from 95% solvent A (water with 0.1% formic acid) to 40% of solvent B (acetonitrile with 0.1% formic acid) for 100 min were applied on nanoACQUITY UPLC (Waters) coupled with Orbitrap Fusion Lumos mass spectrometer (Thermo Scientific), which was operated using the following parameters: m/z 300-1,800 of precursor scan range, 1.4 Th of precursor isolation window, 30% of normalized collision energy for higher-energy collisional dissociation (HCD), 30 s of dynamic exclusion duration and a 60,000 or 75,000 resolution at m/z 200 for full MS or MS/MS scan, respectively.

[0100] Mass spectrometric raw data files were processed with MaxQuant (v.1.5.3.30) searched using the built-in Andromeda search engine[65] in MaxQuant against the human Uniprot database[58] (v.12/5/2018) at default settings (20 or 6 ppm of precursor ion mass tolerances for initial or main search, respectively, and 0.5 Da for fragment ion masses). Enzyme specificity was set to trypsin/P, and a maximum of two missed cleavages were allowed. Cysteine carbamidomethylation and methionine oxidation were selected as fixed and variable modifications, respectively. A 1% FDR was required at both the protein and the peptide level. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE[66] partner repository with the dataset identifier PXD018061 and 10.6019/PXD018061. To account for spurious or nonspecific binding to the SL2.7 construct, the results from the “bead only” and the mutant construct were considered as technical background and designed a customized statistical test for protein spectra count enrichment. Let Np be the number of identified spectra count for protein p from the SL2.7 construct, and Mp be the count number from the background construct. For each protein p with Np 1, the statistical significance (or P value) of enrichment was computed:

[00002]$P\left(X\ge N_p\right)=\underset{k=N_p}{\overset{N}{.Math.}}B\left(k;N,\rho \right)$

[0101] where N=ΣPNP is the total spectra count, ρ=(Mp+1)/ΣP(MP+1) is the background probability, and B(k; n, p) is the binomial distribution of getting k successes in n trials with success probability of P. Finally, P values were adjusted via the Benjamini-Hochberg method. Proteins with an adjusted P<0.01 when using a technical background (‘bead only’ or mutant) were considered statistically significant. Statistically significant proteins in either set 1 or set 2 datasets were considered for subsequent protein intensity analysis. To handle missing values, the minimum nonzero intensity value of the dataset was imputed. Protein enrichment score was then calculated as the log 2 fold change of the protein intensity of the SL2.7 over the mutant data.

[0102] Results

[0103] 1. HBV and HCMV Turn the Mixed Tailing Machinery to their Own Advantage.

[0104] To investigate and characterize viral RNA tailing, TAIL-seq was first applied to two viral infection models: HepG2.2.15 cells incorporated with HBV DNA and primary human foreskin fibroblast (HFF) infected with HCMV. HepG2.2.15 cells were derived from the hepatoblastoma cell line HepG2 and used to study the life cycle of HBV as they sup-port the production and assembly of HBV particles. TAIL-seq data on HepG2.2.15 cells showed exceptionally high guanylation frequency on viral RNAs (FIG. 1a and FIG. 7a). Frequency of mixed tailing (counting G, U and C) is also substantially higher on viral RNAs than cellular RNAs (FIG. 7a). Furthermore, TAIL-seq data on HCMV-infected HFF cells showed high guanylation and mixed tailing frequency on HCMV RNAs (FIG. 1b and FIG. 7b). The modification is attributed mostly to RNA2.7 (FIG. 7c, VRNA2.7), one of the four abundant and conserved long noncoding RNAs of HCMV. Both HBV and HCMV RNAs have substantially longer tails than cellular mRNAs (FIGS. 1c and 1d), indicating that these viral RNAs may undergo slower net deadenylation than cellular mRNAs.

[0105] To examine if the viral mixed tailing is mediated by the same enzymes as the host counterparts, the present inventors depleted TENT4A and TENT4B. Guanylation and mixed tailing frequency of HBV mRNAs and HCMV RNA2.7 were reduced in TENT4A- and TENT4B-depleted cells (FIGS. 1e and 1f and FIGS. 7d-7g). Knockdown of TENT4A and TENT4B also led to the shortening of the poly(A) tail of viral RNAs (FIGS. 7h-7j), demonstrating that TENT4 enzymes contribute to the elongation of viral RNA tails. Consistently, HBV mRNAs were downregulated and destabilized after TENT4 depletion (FIGS. 1g and 1h and FIG. 7k), indicating that HBV uses TENT4 to stabilize their mRNAs. Of note, no increase of tail modifications such as oligo-uridylation (FIG. 7l), excluding the possibility of any compensatory mechanism in the absence of TENT4, was found. The highly stable HCMV RNA2.7 did not show a significant change in RNA half-life within the experimental time window (FIG. 7m, P>0.05; two-sided Student's t-test), suggesting an additional downstream mechanism subsequent to deadenylation, which blocks RNA2.7 decay.

[0106] 2. HBV RNAs are Major Substrates of TENT4 in HepG2.2.15 Cells.

[0107] To elucidate the mechanism of TENT4 recruitment, formaldehyde-mediated crosslinking and immunoprecipitation sequencing (fCLIP-seq) were carried out on HepG2.2.15 cells (FIG. 2). Formaldehyde-based crosslinking captures RNA-protein inter-actions effectively and robustly even when they are transient and structure-dependent. Four libraries were prepared with the input lysates and the immunoprecipitates using normal mouse IgG (NMG), TENT4A antibody or TENT4B antibody. The input and NMG libraries were used as size-matched controls to estimate background. Standardized read coverages across the HBV genome identified substantially enriched peaks from the HBV genome with both TENT4A and TENT4B antibodies (FIG. 2a), indicating the interaction site of TENT4 on HBV RNAs. An independent experiment with TENT4B antibody showed a similar result (FIG. 8a).

[0108] HBV RNAs are generated from a circular DNA genome of 3.2 kilobases. Five overlapping viral transcripts share a common 3′ end sequence (˜700 basepairs) and a single polyadenylation site (FIG. 2a). TENT4-bound reads are enriched in the 3′ end region spanning 1,154-1,687 (FIG. 2a). This region corresponds to the post-transcriptional regulatory element (PRE), which is well known to enhance RNA nuclear export and stabilization. PRE of the woodchuck hepatitis virus (WPRE) has been widely used to increase ectopic gene expression and lentiviral gene delivery. While several RNA-binding proteins such as La, PTB, GAPDH and ZC3H18 have been reported to interact with PRE, the functional importance of these interactions remains unclear. In the fCLIP-seq experiment of the present disclosure, the PRE of HBV exhibited the highest enrichment score compared to the other putative peak clusters of the human transcriptome (FIG. 2b). Immunoprecipitation followed by quantitative PCR with reverse transcription (RT-qPCR) also confirmed the interaction between TENT4A, TENT4B and HBV mRNAs (FIG. 8b). Taken together, these results demonstrate that HBV mRNAs are the major substrates of TENT4 in HepG2.2.15 cells and that TENT4 interacts with HBV mRNAs through the PRE region.

[0109] 3. Stem-Loop a of HBV RNAs is Necessary for TENT4-Dependent Tail Regulation.

[0110] To identify the cis-acting RNA element recognized by TENT4, the present inventors generated reporters with different subregions of PRE in the 3′ untranslated region (UTR) (FIG. 3a). The PRE consists of two subelements: PREα (1154-1351) and PREβ (1352-1687) (FIG. 3a). TENT4A and TENT4B knockout (KO) cells were also generated to test TENT4-dependency (FIG. 9a).

[0111] Luciferase protein and RNA levels were significantly upregulated in a TENT4-dependent manner only when the reporter was equipped with either the full PRE or PREα (FIG. 3b and FIG. 9b). Similar results were obtained in human embryonic kidney 293T (HEK293T) cells depleted of TENT4 (FIGS. 9c and 9d), confirming that PREα is required for TENT4-dependent stabilization. Poly(A) tail length measurement by Hire-PAT assay showed that TENT4 depletion resulted in a shortening of poly(A) tails from the PREα construct but not from the PREβ construct (FIG. 3c).

[0112] This result indicates that TENT4 acts directly through tail modification in a PREα-dependent manner. Moreover, in the knockout cells, the PRE and PREα reporter expressions at both protein and RNA levels were rescued by ectopically expressed TENT4 (FIG. 3d and FIGS. 9e and 9f). The catalytically inactive mutant failed to restore the activity. Thus, the catalytic activity of TENT4 is necessary for the upregulation of mRNA containing PREα.

[0113] To examine the mechanistic conservation of PRE, use was made of the woodchuck hepatitis virus (WHV) segment (WPRE) that is homologous to the PRE of HBV (FIG. 3e). Unlike PRE of HBV, WPRE consists of three subelements: WPREγ, WPREα and WPREβ. WPREα and WPREβ are conserved between WHV and HBV whereas WPREγ is unique and thought to be responsible for the greater activity of WPRE over the PRE of HBV22. All constructs harboring WPREα (Wα) showed TENT4-dependent upregulation (FIG. 3f), suggesting that the TENT4-mediated mechanism is conserved in WHV.

[0114] The combination of WPREγ (Wγ) and WPREα (Wα) resulted in a synergistic effect (FIG. 3f), suggesting that WPREγ could act as an enhancer to TENT4-mediated regulation of WHV.

[0115] Further point mutations were introduced to specify the core motif in PREα (FIG. 3a). PREα contains a La-binding motif and stem-loop alpha (SLα), both of which were reported to be necessary for RNA export and stability. A recent mutagenesis study showed that both the La-binding site and the apical loop sequence of SLα are essential for DHQ-1-mediated HBV gene expression. Of note, the SLα contains a CAGGU pentaloop that essentially adopts the CNGG(N) family loop conformation with a single bulged G residue flanked by A-helical regions (FIG. 3g, left). The structure of SLα is also evolutionarily conserved in WHV (FIG. 3g, right). To test whether these elements are necessary for TENT4-dependent regulation, PREα mutant reporters (FIG. 3a) that disrupt the apical sequence and shape of SLα (FIG. 3g, arrows) were generated. Mutagenesis of La-binding motif (α mut1) did not affect TENT4-dependency while the SLα mutation (α mut2) and the double mutations (α mut3) abrogated the PREα function (FIG. 3h). Thus, the stem-loop of PREα (SLa) is an essential element for post-transcriptional regulation by TENT4.

[0116] 4. HCMV RNA2.7 Harbors a Similar Stem-Loop for TENT4-Dependent Regulation.

[0117] In addition to HBV mRNAs, HCMV RNA2.7 (VRNA2.7) also undergoes mixed tailing (FIGS. 1b and 1d and FIG. 7b). An RNA-immunoprecipitation experiment with TENT4A antibody followed by RT-qPCR showed that TENT4A interacts with RNA2.7 that is ectopically expressed in HEK293T cells (FIG. 4a and FIG. 10a), suggesting that no viral factors are required for the interaction between TENT4A and RNA2.7. To identify the functional domain of RNA2.7 responsible for TENT4 interaction, reporters that contain partially overlapping fragments of RNA2.7 (FIG. 4b, FRG1-6) were generated. Protein and RNA levels of the first two reporters (FRG1 and FRG2) were significantly higher in parental cells compared to those in TENT4 KO cells (FIG. 4c and FIG. 10b). Thus, the cis-element is present near the 5′ end. Partial deletion of FRG1 revealed that only the 1E fragment (414-513) confers the TENT4-dependent enhancement effect (FIGS. 4d and 4e and FIGS. 10c and 10d). The 1E domain harbors a stem-loop structure (FIG. 4b, lower right) that resembles the CNGG(N) family loop conformation of SLα of PREα and WPREα (FIG. 3g). Point mutations to the stem-loop of 1E (1E mut) abrogated the regulation of TENT4 (FIG. 4e and FIG. 10d). A smaller 50-nt truncated fragment that contains the stem-loop (414-463, hereafter referred to as SL2.7) has a similar activity to FRG1 and 1E (FIG. 10e). SLα of HBV PRE and WPRE are near the 3′ end of RNA, but SL2.7 of HCMV RNA2.7 is near the 5′ end, suggesting that the pentaloop may function in a position-independent manner.

[0118] Altogether, the data of present disclosure demonstrate the functional importance of the SLα-like structure in TENT4-dependent regulation.

[0119] 5. Cytoplasmic ZCCHC14 Binds to the Stem-Loop Structure and Interacts with TENT4.

[0120] The CNGGN pentaloop was first characterized as the Smaug recognition element, which binds to the sterile alpha motif (SAM) domain of yeast Vts1p and Drosophila Smaug mediates post-transcriptional repression[32-36]. Based on the structural resemblance, Schwalbe and colleagues had proposed that the HBV SLα pentaloop may serve as a binding site for a protein with the SAM domain. However, TENT4 does not contain a SAM domain[31], and the present inventor's tailing assays with recombinant TENT4 failed to detect a notable affinity to PRE in vitro, indicating an RNA-interacting cofactor.

[0121] To find the potential cofactor that recognizes SLα and SL2.7, RNA pulldown was conducted with biotinylated synthetic SL2.7 and HEK293T cell extract (FIG. 5a). The loop mutant was used as a negative control. Mass spectrometry revealed six proteins that are markedly enriched, that is, TENT4A, TENT4B, K0355, SAMD4A, SAMD4B, and ZCCHC14 (FIG. 5a). Particularly, all three human proteins with Smaug-like SAM domains are included in this list: the two homologs of Smaug (SAMD4A and SAMD4B) and ZCCHC14 (FIG. 5b). The specific enrichment of ZCCHC14 was further confirmed by RNA pulldown and western blotting (FIG. 5c). K0355 (or KIAA0355) does not have a SAM domain and it is reported to interact with SAMD4B37, indicating that K0355 may not directly interact with the pentaloop. To identify which SAM domain protein is responsible for TENT4-dependent regulation, reexamination was made of the HBV-producing HepG2.2.15 cells. Knockdown of ZCCHC14 but not that of SAMD4 proteins reduced HBV RNA levels (FIG. 5d and FIG. 11). Consistently, a recent genome-wide CRISPR screen reported ZCCHC14 as an essential host factor for HBV surface antigen production.

[0122] To understand the action mechanism of ZCCHC14, TAIL-seq experiments were performed in ZCCHC14-depleted HepG2.2.15 cells. Guanylation frequency of HBV mRNAs was reduced (FIG. 6a), and the poly(A) tail was shortened substantially on ZCCHC14 knockdown (FIG. 6b), demonstrating the critical role of ZCCHC14 in HBV RNA mixed tailing. Consistently, ZCCHC14 depletion led to the reduction of luciferase expression from reporters harboring the CNGGN stem-loop (RNA2.7 1E, PREα, Wγ+Wα) (FIG. 6c). In contrast, their respective pentaloop mutants (1E mut, a mut2, Wγ+Wα mut) did not show significant alterations after ZCCHC14 knockdown (P>0.05; two-sided Student's t-test) (FIG. 6c), indicating that the function of ZCCHC14 depends on the CNGGN motif. The present inventors also carried out Hire-PAT assay to confirm that the poly(A) tail length of reporter mRNAs decreased in ZCCHC14-depleted cells in a CNGGN motif-dependent manner (FIGS. 12a-12c).

[0123] RNA immunoprecipitation using ZCCHC14 antibody confirmed an interaction between ZCCHC14 and HBV transcripts in HepG2.2.15 cells (FIG. 6d and FIG. 12d). Moreover, in both HepG2.2.15 and primary HFF cells, ZCCHC14 co-immunoprecipitates with TENT4A and TENT4B regardless of RNase treatment (FIG. 12e), indicating an RNA-independent interaction between ZCCHC14 and TENT4. Last, the present inventors found that ZCCHC14 is mostly localized in the cytoplasm (FIG. 6e) and that ZCCHC14 coimmunoprecipitates with TENT4A and TENT4B specifically using the cytosol fraction of HeLaT cells (FIG. 6f). Collectively, the results of the present disclosure indicate that ZCCHC14 interacts with TENT4 mainly in the cytoplasm to inadvertently protect viral transcripts harboring SLα-like elements from deadenylation.

[0124] The present inventors revealed the mechanism of targeted mixed tailing in viral gene expression (FIG. 6g). HBV mRNAs and HCMV RNA2.7 contain cis-acting elements with a CNGGN pentaloop that is critical for recruiting ZCCHC14 and TENT4. The mixed tail by TENT4 counteracts the CNOT deadenylase complex, thereby protecting the viral RNAs. It is intriguing that these two distinct viruses (belonging to Hepadnaviridae and Herpesviridae, respectively) with very different life cycles and tissue specificity came up with an identical strategy for the benefit of their gene expression. This convergent evolution and the simplicity of this cis-acting element suggest that this mechanism may be used by other viruses as well.

[0125] Other than mixed tailing machinery, TENT4B works as a component of nucleolar TRAMP complex (Trf4-Air1/2-Mtr4 in yeast and TENT4B-ZCCHC7-hMTR4 in human), and is involved in the degradation of aberrant transcripts or the trimming and maturation of small noncoding RNAs in the nucleus[5, 39-43]. The present inventors previously showed that mixed tailing contributes to mRNA stability as a module distinct from TRAMP complex. Given that the knockdown of hMTR4 and ZCCHC7 does not affect the HBV protein and mRNA levels, TENT4 seems to act independently of the TRAMP complex in HBV regulation. Cytoplasmic localization of ZCCHC14 also suggests that the TENT4-ZCCHC14 complex operates in a separate subcellular compartment from the nucleolar TRAMP complex although exclusion is not made of the possibility that a minor fraction of the TENT4-ZCCHC14 complex may act in the nucleus.

[0126] Mixed tailing seems to be particularly important for HBV gene expression. Four out of five mRNA species (pgRNA, Precore, PreS1 and PreS2/S except X) contain the SLα stem-loop, and their half-lives are substantially reduced when TENT4 enzymes are depleted. Current treatment options include nucleos(t)ide analogs and interferon alpha, but they are not curative and often ineffective, leaving HBV infection as the major global medical burden causing almost 80 million deaths per year. Functional cure can be achieved by blocking the production of viral proteins, especially viral S antigen (HBsAg) that induces immune evasion. The discovery of viral mixed tailing and the involvement of TENT4 and ZCCHC14 in viral gene expression may provide mechanistic insights for the development of a new class of anti-HBV drugs. To avoid any potential side effects of targeting the TENT4 machinery, it will be important to understand the endogenous functions and action mechanisms of the TENT4 proteins and their partners.

[0127] The physical interaction between the CNGGN pentaloop and RNA-binding SAM proteins appears to be a highly conserved and widely used functional module. In fruit fly, Smaug recognizes the hairpin with a CUGGC loop of nanos mRNA to induce timely mRNA degradation, which is essential for the establishment of the anterior-posterior axis. In yeast, Vts1p binds to mRNA with the CNGG(N) pentaloop for targeted mRNA degradation. Vertebrates have three RNA-binding SAM proteins and seem to have expanded the use of this interaction module. SAMD4B, one of the vertebrate orthologs of Smaug, targets nanos1 mRNA and is essential during mammalian neural development. SAMD4A, the other vertebrate ortholog of Smaug, suppresses the translation of pentaloop-containing reporters and is associated with the formation of cytoplasmic foci. Here, the present inventors show that the same mechanism could be used for the opposite purpose via a vertebrate-specific protein ZCCHC14, in particular for the benefit of viral RNAs.