Fig 1.
Interaction network of DENV NS5 and human proteins from infected cells.
(A) Construction of recombinant DENVs. Viruses with Strep-tag fused to NS5 in different positions: N-terminus (r1-DV), inter-domain linker (r2-DV), low conserved loop region (r3-DV) and C-terminus (r4-DV) are shown. Immunofluorescence assays after RNA transfection using specific anti-E antibodies show replication of each virus. (B) Flow chart of experimental procedure for NS5-containing protein complex purification. Human cells (HEK 293T or Huh7) were mock infected, DENV WT infected, or DENV NS5-tag infected in parallel. (C) Coomassie blue staining showing the recovery of NS5 binding partners. (D) Results of the mass spectrometry analysis, the major functional categories of host proteins identified as NS5 interactors are depicted. (E) NS5-host proteins interaction network. Lines depict interactions detected in this study between NS5 and human proteins. Also, interaction between human proteins extracted from database (http://string-db.org) is shown. Components of the spliceosomal U5 snRNP are shown in red. Previously reported NS5 interactors are indicated in green.
Fig 2.
NS5 interacts with splicing proteins in the context of active spliceosomes.
(A) NS5 binds CD2BP2, DDX23, and EFTUD2. Pull down assays and western blot analysis of indicated proteins are shown. (B) NS5 binds to assembled snRNPs containing the U1, U2, U4, U5, and U6 snRNAs. (C) Detection of indicated pre-mRNAs in the NS5 associated complexes. (D) Subcellular fractionation showing that NS5 accumulates in the insoluble chromatin fraction, similar to CD2BP2, a U5 snRNP component (Nuc: nucleoplasm, Cyt: cytoplasm, Chro: chromatin). Fractionation of cells expressing GFP is also shown as control, indicating that this protein distributes along different cellular compartments.
Fig 3.
DENV infection modifies the splicing landscape of host cells.
(A) Strategy to detect and quantify alternative splicing isoforms in mock infected or DENV infected cells is depicted on the left. Representative autoradiographs and quantification of the inclusion/exclusion ratio for two alternative exon cassette events are shown (duplicates, mean ± SD). (B) Fold-change of inclusion/exclusion ratios for DENV or mock infected cells. A battery of endogenous alternative exon-cassette events (duplicates, mean ± SD) in A549 and Huh7 cells is shown. (C) DENV-induced changes in the splicing pattern of IKKε and MxA mRNAs. Quantification of alternative splicing events is shown for each case.
Fig 4.
NS5 alone interferes with alternative splicing.
(A) NS5 shows a dose dependent effect on the splicing reporter minigenes CFTR, EDI, and Bclx. Radiolabeled amplification products corresponding to different splicing isoforms derived from the indicated minigenes are shown. (B) Quantitative analysis of the changes induced by NS5 (0.9 μg of plasmid). (C) Expression of NS5 mutants with impaired methyltransferase (Mut-MTase) or polymerase (Mut-RdRp) enzymatic activity also alters alternative splicing of reporter genes.
Fig 5.
Abundance and localization of spliceosomal factors during DENV infection.
(A) Western blots showing the levels of splicing proteins CD2BP2, DDX23, EFTUD2, SNRNP40 or SF3A2 and STAT2 used as a control, at 48 h post-infection. (B) Nuclear localization of CD2BP2 or DDX23 and NS5 in DENV infected or mock infected cells. (C) Determination of relative abundance of U2, U4, U5, and U6 snRNAs in DENV infected or mock A549 and Huh-7 cells.
Fig 6.
NS5 interferes with an in vitro splicing reaction and inhibits pre-mRNA maturation.
(A) Time course of standard in vitro reaction showing radiolabeled substrate, intermediates and products of the splicing reaction. (B) Purified recombinant NS5 inhibits mature mRNA accumulation. A native or heat denature preparation of NS5 was added to the splicing reaction as indicated on the top of the gel. On the right, a graph showing dose-dependent inhibition of the splicing reaction by NS5. (C) Quantification of pre-mRNA and mature mRNA as a function of time in the presence of native NS5 or heated NS5 at 1μM.
Fig 7.
NS5 reduces the splicing efficiency of endogenous RIG-I mRNA.
(A) Schematic representation of RIG-I pre-mRNA and mRNA. The abundance of these forms was measured by quantifying the immature and mature RNA transcripts using the region corresponding to exon 10, exon 11 and the intron between them, as indicated. The relative positions of the primers designed for RT-qPCR are depicted. (B) NS5 expressed as a mature protein significantly reduces the RIG-I splicing efficiency compared with GFP or an empty vector control. Data corresponding to four independent experiments is shown, mean ± SD. (C) RIG-I splicing efficiency (estimated as the RIG-I mRNA/pre-mRNA ratio) was not altered by α-IFN. (D) NS5 expressed as a mature protein does not induce STAT2 protein degradation.
Fig 8.
Decreased splicing efficiency in DENV infected cells.
(A) Schematic representation of splicing analysis. The four splicing events analyzed are shown: Intron retention (IR), Exon skipping (ES), Alternative splice site donor/acceptor (Alt5’SS and Alt3’-SS). Constitutive exon and alternative 5’ and 3’SS regions are shown in grey and intron regions in pink. For intron retention the information of the junctions (E1-I, IE-2 and E1E2) was used to calculate the PIR metric. For ES and AltSS the PSI metric was used (see Materials and methods). (B) Data of splicing analysis of DENV or mock infected cells. On the upper panels, total amount of bins analyzed and altered splicing events for each time point post infection (24 and 36 hours). The percentage of altered events (ES, ALtSS and IR) is shown in pie charts. On the lower panel, increase or decrease retention of introns for each time point is shown.
Fig 9.
Silencing U5 snRNP components enhances DENV replication.
(A) Reporter DENV replication in A549 cells transfected with siRNA directed to spliceosomal proteins (CD2BP2, DDX23, SNRNP40, PRPF8, EFTUD2, SNRNP200 or SF3A2). An infectious DENV carrying the luciferase gene was used. Three controls were included: a non-related siRNA (NR, black bar), siRNAs directed to Renilla luciferase (control for inhibition, red bar) and siRNA directed to STAT2 (control of NS5 binder with antiviral activity, green bar). Results are representative of three independent experiments (duplicates, mean ± SD). At the bottom, immunoblots indicate the levels of silenced proteins. (B) DENV replication in A549 cells transfected with a non-related siRNA (NR), or siRNAs directed to CD2BP2, DDX23 or EFTUD2, measured by quantitative real time PCR (mean ± SD) at 10hpi for EFTUD2 and 24hpi for CD2BP2 and DDX23. (C) Induced levels of mRNAs corresponding to RIGI, ISG15 and IL8 in A549 cells infected with NDV, which were previously silenced with siRNAs directed to CD2BP2 or EFTUD2 as indicated for each case. RNA levels were measured by quantitative real time PCR (mean ± SD). On the right, immunoblots indicate the levels of silenced proteins and loading controls.