The subgenomic flaviviral RNA suppresses RNA interference through competing with siRNAs for binding RISC components

ABSTRACT

During the life cycle of mosquito-borne flaviviruses, substantial subgenomic flaviviral RNA (sfRNA) is produced via incomplete degradation of viral genomic RNA by host XRN1. Zika virus (ZIKV) sfRNA has been detected in mosquito and mammalian somatic cells. Human neural progenitor cells (hNPCs) in the developing brain are the major target cells of ZIKV, and antiviral RNA interference (RNAi) plays a critical role in hNPCs. However, whether ZIKV sfRNA was produced in ZIKV-infected hNPCs as well as its function remains not known. In this study, we demonstrate that abundant sfRNA was produced in ZIKV-infected hNPCs. RNA pulldown and mass spectrum assays showed ZIKV sfRNA interacted with host proteins RHA and PACT, both of which are RNA-induced silencing complex (RISC) components. Functionally, ZIKV sfRNA can antagonize RNAi by outcompeting small interfering RNAs (siRNAs) in binding to RHA and PACT. Furthermore, the 3′ stem loop (3′SL) of sfRNA was responsible for RISC components binding and RNAi inhibition, and 3′SL can enhance the replication of a viral suppressor of RNAi (VSR)-deficient virus in a RHA- and PACT-dependent manner. More importantly, the ability of binding to RISC components is conversed among multiple flaviviral 3′SLs. Together, our results identified flavivirus 3′SL as a potent VSR in RNA format, highlighting the complexity in virus-host interaction during flavivirus infection.

IMPORTANCE

Zika virus (ZIKV) infection mainly targets human neural progenitor cells (hNPCs) and induces cell death and dysregulated cell-cycle progression, leading to microcephaly and other central nervous system abnormalities. RNA interference (RNAi) plays critical roles during ZIKV infections in hNPCs, and ZIKV has evolved to encode specific viral proteins to antagonize RNAi. Herein, we first show that abundant sfRNA was produced in ZIKV-infected hNPCs in a similar pattern to that in other cells. Importantly, ZIKV sfRNA acts as a potent viral suppressor of RNAi (VSR) by competing with siRNAs for binding RISC components, RHA and PACT. The 3′SL of sfRNA is responsible for binding RISC components, which is a conserved feature among mosquito-borne flaviviruses. As most known VSRs are viral proteins, our findings highlight the importance of viral non-coding RNAs during the antagonism of host RNAi-based antiviral innate immunity.

INTRODUCTION

Zika virus (ZIKV) is a mosquito-borne flavivirus that is closely related to dengue virus (DENV), West Nile virus (WNV), Japanese encephalitis virus (JEV), and yellow fever virus (YFV). Most ZIKV infections are asymptomatic or only develop mild symptoms such as fever, rash, arthralgia, and/or conjunctivitis (1). Notably, ZIKV infection during pregnancy can cause fetal loss, microcephaly, and other brain abnormalities (2–4). Investigation with neurospheres, organoids, and animal models have demonstrated that ZIKV directly targets human neural progenitor cells (hNPCs) (5–7), the proliferation and differentiation of which are vital for normal brain development. More importantly, ZIKV shows a strong tropism for hNPCs, whereas the more differentiated mature neurons are less sensitive to ZIKV (5, 6, 8). ZIKV infection suppresses NPC proliferation and differentiation, resulting in cortical thinning and thereby causing microcephaly and other central nervous system (CNS) abnormalities (5, 9).

The flaviviral genome is a single-stranded, positive-sense RNA encoding a 5′untranslated region (UTR), a single open reading frame (ORF), and a 3′UTR. During flaviviruses infection, the 5′–3′ exoribonuclease (XRN1) can degrade viral genomic RNA (gRNA) (10). However, the 3′UTR resists XRN1 digestion by tightly folded RNA structures, resulting in the production of subgenomic flavivirus RNA (sfRNA) (11, 12). sfRNA has been found abundant in infected mosquito and mammalian somatic cells (10, 11) and well documented with multiple functions including pathogenicity in mice and cytopathicity in cell culture (10, 13). The sfRNA has been described to exert functions through binding host proteins. DENV-2 sfRNA binds G3BP1, G3BP2, CAPRIN1 to inhibit their activities in regulation of IFN-stimulated genes (ISGs) translation, leading to ISGs downregulation, thus protecting DENV-2 replication against the IFN antiviral effects. DENV-2 sfRNA also binds ubiquitin ligase TRIM25 to interfere with its deubiquitylation, thus preventing amplified and sustained retinoic acid-inducible gene 1 (RIG-I) signaling for antiviral IFN induction (14, 15). ZIKV sfRNA has been shown to bind to FMRP in Hela cells. FMRP can repress ZIKV infection by inhibiting viral translation and sfRNA can antagonize its activity to facilitate viral infection (16). In addition, ZIKV and WNV sfRNA bind ME31B to promote viral replication in Aag2 cells (17). However, the characteristics and functions of ZIKV sfRNA in hNPCs have not been investigated.

RNA interference (RNAi) is a mechanism of eukaryotic posttranscriptional gene regulation mediated by small interfering RNA (siRNA)-induced sequence-specific RNA degradation (18). It is also well known that RNAi acts as an important antiviral defense mechanism in a wide range of organisms (19). During virus infection, RNAIII-like endonuclease (named Dicer) recognizes and cleaves the virus-derived long double-stranded RNA (dsRNA) into siRNAs of approximately 21–23nucleotides (nts). These siRNAs are subsequently incorporated into the RNA-induced silencing complex (RISC) which guides the cleavage of target mRNAs with perfect sequence complementarity to the viral siRNAs (vsiRNAs) (20). RISC fractionates in large molecular mass complexes containing multiple proteins, such as RHA, TRBP, and PACT, which are important for RISC function in gene silencing (21–24). To counter the RNAi-mediated antiviral defense, viruses have developed a variety of viral suppressor of RNAi (VSR). For example, a number of viral proteins have been identified as VSRs, such as flock house virus (FHV) B2 protein (25), Wuhan nodavirus (WhNv) B2 (26), Criket paralysis virus (CrPv) 1A (27), vaccinia virus E3L, Nodamura virus (NoV) B2, human enterovirus 71 (HEV71) 3A (28), Ebola virus (EBOV) VP35 (29), and Influenza A virus (IAV) NS1 (30, 31). In addition to these viral proteins acting as VSR, some viral RNAs can also act as VSRs, such as adenoviral VA1 noncoding RNA (32), human immunodeficiency virus-1 (HIV-1) TAR RNA, and Rev-Response Element (RRE) RNA (33, 34). As for flaviviruses, sfRNAs of WNV, DENV, and Kunjin virus (KUNV) have been implicated as putative suppressors of RNAi (35, 36), but the molecular details need to be elucidated. RNAi-mediated antiviral response plays critical roles during ZIKV infection in hNPCs (37). However, whether ZIKV sfRNA has the RNAi suppression activity and the underlying mechanism remains unknown.

Recently, we have shown that ZIKV and some other flaviviruses suppress RNAi via NS2A (38). In this study, we first found that ZIKV-infected hNPCs produced amount of sfRNA of the same size as that in differentiated cells. Then, we identified that ZIKV sfRNA as a potent VSR in RNA format, which competes with siRNA for binding to RISC components RHA or PACT to interfere RNAi and prompt ZIKV replication in hNPCs. We further found the 3′ stem loop (3′SL) of sfRNA is responsible for binding RISC components and inhibiting RNAi. Moreover, the ability of binding to RISC components is conversed among multiple flaviviral 3′SLs. Overall, our findings demonstrated a novel role of flavivirus 3′SL in evading RNAi.

RESULTS

ZIKV sfRNA involves in a variety of biological processes in hNPCs

Herein, to determine whether sfRNA was produced in ZIKV-infected hNPCs, ZIKV strains FSS13025 (a historical Asian strain) and GZ01 (a contemporary American strain) were used to infect hNPCs and BHK-21 cells. Total cell RNA were subjected to Northern Blot analysis. The result showed both two ZIKV strains produced abundant sfRNA in hNPCs, and the size of sfRNAs in hNPCs was similar to that in BHK-21 cells (Fig. 1A).

Fig 1.

Identification of ZIKV sfRNA-binding proteins in hNPCs. (A) hNPCs and BHK-21 cells were infected with ZIKV FSS13025 or ZIKV GZ01. Thirty-six hours post-infection, cell-associated RNA was harvested and levels of gRNA and sfRNA were analyzed by Northern Blot. (B) Workflow of RNA pulldown and LC-MS/MS analysis. Biotinylated RNAs containing FSS13025 sfRNA, GZ01sfRNA, or a size-matched control transcript were incubated with hNPC lysates. Proteins that co-purified with the RNAs on streptavidin beads were identified by chromatography mass spectropmetry. (C) Volcano plots of eluted proteins. The x axis shows the mean 2Log difference in protein abundance between the FSS13025 sfRNA (left) or GZ01 sfRNA (right) and control samples from three independent biological replicates. The y axis shows the 10Log of the P value by Student’s t test from the comparison of the protein abundance in the FSS13025 sfRNA (left) or GZ01 sfRNA (right) with the control samples. Significantly enriched proteins are shown as red. The 12 proteins that bond by both FSS13025 sfRNA and GZ01 sfRNA are shown as red diamonds, and their names are labeled aside. The names of proteins associated with RISC assembly are shown in blue. (D) Go enrichment classification of the 12 proteins that bond by both FSS13025 sfRNA and GZ01 sfRNA. (E) Biotinylated RNAs containing FSS13025 sfRNA, GZ01 sfRNA, or a size-matched control transcript were incubated with hNPC lysates. Associated proteins were eluted and analyzed by Western blot.

Then, RNA pulldown and liquid chromatograph tandem mass spectrometry (LC-MS/MS) assays were performed to identify host proteins that bind to ZIKV sfRNA in hNPCs. The full-length sfRNA from FSS13025 and GZ01 was used to bind with the hNPC cytoplasmic extracts (Fig. 1B). Volcano plots were generated for both FSS13025 and GZ01 sfRNA where proteins with a ≥4-fold relative enrichment in the sfRNA samples and a P value of ≤0.05 were considered significant (Fig. 1C). A total of 12 target proteins were identified with sfRNAs from both FSS13025 and GZ01 but not with the size-matched control RNA. Gene ontology (GO) classification of these result revealed that the most significant enriched pathway is RISC complex assembly, followed by viral regulation, cellular response to oxidative stress, RNA splicing, and cellular response to stress (Fig. 1D). These data provide a valuable resource to test for pro- or anti-viral activities during ZIKV infection.

To validate the interactions, we chose five representative proteins and performed direct pulldown assay and probed proteins by Western Blot (WB) (Fig. 1E). As expected, RHA, EIF2AK2, STAU1, and PACT directed binds to both FSS13025 and GZ01 sfRNA. Meanwhile, TRBP did not bind to FSS13025 sfRNA or GZ01 sfRNA, indicating TRBP may interact with ZIKV sfRNA indirectly. Nevertheless, these data demonstrated that ZIKV sfRNA binds to a variety of host proteins and may sequester them in an effector to regulate viral replication in hNPCs.

ZIKV sfRNA specifically interacts with RISC components RHA and PACT in the context of infection

To further validate the interaction of those proteins with ZIKV sfRNA in the context of ZIKV infection, we performed RNA immunoprecipitation (RIP) using corresponding antibodies to evaluate the interaction of these proteins with ZIKV viral RNA. Total RNA was purified from the immunoprecipitated material, and the levels of ZIKV gRNA and sfRNA precipitates were analyzed by RT-qPCR. As seen in Fig. 2A through D, all of these four proteins co-precipitated gRNA and virus-made sfRNA significantly above the IgG control background. These results demonstrated these proteins directly bond with ZIKV sfRNA in ZIKV-infected hNPCs. Importantly, RHA and PACT have a well-established association with RNAi, which is important for ZIKV infection in hNPCs (37, 39).

Fig 2.

Analyses of the interaction between ZIKV sfRNA and host proteins in the context of infection. RNA immunoprecipitation was performed on hNPCs at 48-h post-infection with FSS13025 and GZ01 (MOI = 2). Infected cells were subjected to RHA (A) or PACT (B) or EIF2AK2 (C) STAU1 (D) or IgG control pull-down followed by immunoblotting with the indicated antibodies to confirm efficacy of pulldown. Co-precipitated viral gRNA and sfRNA were detected by RT-qPCR. Data are expressed as mean ± SD from three independent experiments. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, as determined by Student’s t test.

ZIKV sfRNA suppresses RNAi

Given ZIKV sfRNA binds to RISC components, we hypothesized that ZIKV sfRNA possesses similar RNAi suppressor activity. To verify this hypothesis, we assayed the capability of ZIKV sfRNA in suppressing shRNA-induced eGFP silencing. The eGFP-specific shRNA (shGFP) caused a strong decrease of eGFP expression (Fig. 3A through C) compared to the negative control shRNA (shNC). Notably, transfection of the in vitro transcripted ZIKV sfRNA remarkably resurrected the expression of RNAi-silenced eGFP (Fig. 3A through C), indicating that ZIKV sfRNA suppressed the shRNA-induced RNA silencing of eGFP.

Fig 3.

ZIKV sfRNA inhibits shRNA-induced RNAi in hNPCs. (A) hNPCs were cotransfected with plasmids pEGFP-N1 (encoding eGFP) (100 ng), shGFP, or shNC (300 ng), and ZIKV sfRNA or control RNA (300 ng). The expression of eGFP was analyzed 48 h after cotransfection. The intensity of eGFP was observed under fluorescence microscopy. (B) Cell lysates from (A) were harvested and analyzed by Western blot. (C) The levels of GFP expression in panel A of three independent experiments were quantified by densitometry and normalized to that of actin. The expression level of GFP in the first column was set at 1.0. (D) hNPCs were infected with ZIKV or △10 ZIKV. Thirty-six hours post-infection, cell-associated RNA was harvested and levels of gRNA and sfRNA were analyzed by Northern Blot. (E) hNPCs were mock infected or infected with ZIKV FSS13025 (MOI = 2) or △10 ZIKV(MOI = 5) (different MOIs were used to achieve similar infection rates). Twelve hours after infection, cells were cotransfected with plasmids as indicated. The expression of eGFP was analyzed 36 h after cotransfection. The intensity of eGFP was observed under fluorescence microscopy. (F) Cell lysates from (D) were harvested and analyzed by Western blot. (G) The levels of GFP expression in panel F of three independent experiments were quantified by densitometry and normalized to that of actin. The expression level of GFP in the first column was set at 1.0. Scale bar, 200 μm. Data are expressed as mean ± SD from three independent experiments. *P ≤ 0.05, ***P ≤ 0.001, as determined by Student’s t test.

To assess the RNAi suppression activity of sfRNA during ZIKV infection, mock- or ZIKV-infected hNPCs were transfected with an eGFP expressing plasmid in the presence or absence of shGFP. As seen in Fig. 3E through G, ZIKV infection efficiently suppressed shRNA-induced RNA silencing of eGFP but not the mock infection. Importantly, a ZIKV mutant strain △10 which is deficient in sfRNA accumulation (16, 40) (Fig. 3D) showed only a weak effect on the suppression of eGFP RNAi (Fig. 3E through G), indicating ZIKV sfRNA could suppress RNAi in the context of infection.

ZIKV sfRNA competes with siRNA for binding to RHA or PACT

We next sought to determine the mechanism by which ZIKV sfRNA inhibits RNAi. In the RNAi pathway, after being processed from dsRNA, siRNAs needed to be appropriately incorporated into Ago2 to mediate the cleavage of cognate mRNAs (41). As RHA and PACT play important roles in siRNAs recognition and loading siRNAs onto Ago2 for RISC activation (21, 42, 43), we next investigate whether ZIKV sfRNA can compete with siRNA for binding to RHA or PACT. To do so, we obtained recombinant tandem dsRNA-binding domains (dsRBDs) of RHA (because the tandem dsRBDs of RHA is responsible for siRNA binding) and PACT proteins (Fig. 4A) and performed electrophoresis mobility shift assay (EMSA). As shown in Fig. 4B and C, the shifted complex of biotin-labeled siRNAs and protein were observed as RHA or PACT protein was used, indicating that RHA and PACT can bind to siRNA. However, the shifting amount of siRNA was decreased by sfRNA, but not the control RNA, in a dose-dependent manner, suggesting that sfRNA can displace the labeled siRNAs from RHA or PACT. These results indicated that ZIKV sfRNA could compete with siRNA for binding to essential RISC components and, thus, hinder mature RISC formation.

Fig 4.

ZIKV sfRNA competes with siRNA for binding to RHA and PACT. (A) SDS-PAGE of purified recombinant RHA (tandem dsRBDs, 1–264aa) and PACT. (B) Increasing amounts of unlabeled ZIKV sfRNA [left panel: 0.5 (lane 3), 2 (lane 4), 9 (lane5), or 27 (lane 6) pmoles; right panel: 0.2 (lane 3), 0.5 (lane 4), 1 (lane5), or 3 (lane 6) pmoles] (B) or control RNA (C) were incubated with 4 µM protein and 10 nM biotin-labeled siRNA as indicated in a 20 μL reaction. Samples were separated on a 6% native-TBE polyacrylamide gel electrophoresis (PAGE), transferred to membranes, and then incubated with Horseradish-conjugated Streptavidin.

ZIKV sfRNA antagonizes antiviral function of RNAi through binding RHA and PACT

RISC components RHA and PACT play important roles in RNAi pathway, as depletion of RHA or PACT inhibits siRNA-and shRNA-mediated gene silencing (21, 24, 44). Given that ZIKV sfRNA could compete with siRNA for binding to RHA and PACT, we next sought to test the hypothesis that ZIKV sfRNA is capable of attenuating RNAi-mediated antiviral response. For this purpose, we asked whether WT ZIKV and sfRNA-deficient △10 ZIKV exhibit different sensitivity to RHA or PACT depletion in hNPCs. We observed that RHA or PACT knockdown (Fig. 5A) significantly increased RNA replication (Fig. 5B) of both WT ZIKV and △10 ZIKV. More importantly, RHA or PACT knockdown disproportionately enhanced △10 ZIKV replication compared to WT ZIKV (Fig. 5C): whereas WT virus was enhanced by approximately twofold, the infection for △10 ZIKV rose by fourfold. Additionally, similar effects were observed on levels of ZIKV infection rate (Fig. 5D). Together, these results suggest that RISC components RHA as well as PACT act as restriction factors for ZIKV and sfRNA could antagonize this antiviral response.

Fig 5.

ZIKV sfRNA antagonizes antiviral function of RNAi. hNPCs were transfected with control siRNA (siCtrl) and siRNAs targeting RHA (siRHA) or PACT(siPACT). Thirty-six hours after transfection, cells were infected with ZIKV FSS13025 at MOI = 1 for 36 h. (A) Representative Western blot of RHA or PACT knockdown efficiency. (B) Viral RNA copies in supernatants were measured by RT-qPCR. (C) Fold change of (B). (D) The expression of viral envelope protein from (B) was detected by immunostaining. Viral infection rates were measured by envelope protein signal. Scale bar, 50 μm. Data are expressed as mean ± SD from three independent experiments. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, as determined by Student’s t test.

The 3′SL of ZIKV sfRNA is responsible for the RNAi suppression activity

After confirming the VSR activity of ZIKV sfRNA, we aimed to ascertain the critical region(s) responsible for its RNAi inhibition activity. We first mapped the region(s) of ZIKV sfRNA that mediate binding of RHA or PACT. Based on the previously proposed RNA structure (45) (Fig. 6A), we constructed and transcribed a series of truncated sfRNA and examined their ability to interact with RHA or PACT by RNA pulldown assay. Interestingly, among these truncated sfRNAs, only 3′SL retained the ability to bind to RHA or PACT (Fig. 6B), suggesting that 3′SL, the last region of ZIKV sfRNA, is responsible for RHA and PACT binding. Then, we detect whether the 3′SL can inhibit RNAi by the reversal-of-silencing assay in hNPCs, as shown in Fig. 6C through E, 3′SL alone remarkably resurrected the expression of RNAi silenced eGFP. These data demonstrated that the 3′SL of sfRNA is responsible for inhibiting RNAi through binding to RHA and PACT.

Fig 6.

Mapping the region(s) of sfRNA responsible for RHA or PACT binding and RNAi repression activity. (A) Schematic representation of the truncate RNAs. (B) Biotinylated truncate RNAs were incubated with hNPC lysates. Associated proteins were eluted, separated on a 10% SDS-PAGE and probed with the antibodies indicated on the left by Western blot. (C) hNPCs were cotransfected with plasmids pEGFP-N1(encoding eGFP) (100 ng), shGFP, or shNC (300 ng) and 3′SL RNA or control RNA (300 ng). The expression of eGFP was analyzed 48 h after cotransfection. The intensity of eGFP was observed under fluorescence microscopy. (D) Cell lysates from (C) were harvested and analyzed by Western Blot. (E) The levels of GFP expression in panel D of three independent experiments were quantified by densitometry and normalized to that of actin. The expression level of GFP in the first column was set at 1.0. Scale bar, 200 μm. Data are expressed as mean ± SD from three independent experiments. **P ≤ 0.01, as determined by Student’s t test.

The 3′SL of ZIKV rescues the replication of the VSR-deficient human enterovirus 71

We sought to further examine whether the VSR activity of ZIKV 3′SL can rescue the replication of the VSR-disabled virus. To this end, we detected the effect of ZIKV 3′SL on the replication of HEV71 mutant HEV71_D23A, which lacks VSR function and is restricted by RNAi (28). Human 293T cells were transfected with ZIKV 3′SL RNA or the control RNA and then infected with HEV71_D23A. Our results show that ZIKV 3′SL RNA significantly facilitated HEV71_D23A RNA accumulation and virion production (Fig. 7A and B), suggesting that ZIKV 3′SL can rescue the replication of VSR-deficient virus. More importantly, this facilitation was not observed when RHA and PACT were knocked down (Fig. 7A through C). Together, these results indicated that ZIKV 3′SL can antagonize host RNAi suppression through sequestering RISC components.

Fig 7.

ZIKV 3′SL rescues the replication of VSR-deficient HEV71. (A and B) 293T cells were transfected with control siRNA (siCtrl) and siRNAs targeting RHA (siRHA) and PACT (siPACT). Twenty-four hours after transfection, cells were transfected with Ctrl RNA or ZIKV 3′SL RNA. Six hours after transfection, cells were infected with HEV71_D23A at an MOI = 0.01 as indicated. At 24–72 hpi, the levels of HEV71_D23A genomic RNAs were determined via RT-qPCR. The level of HEV71_D23A RNA in cells at 24 hpi was defined as 1 (A). The viral titers at 72 hpi were determined (B). (C) The RHA and PACT knockdown efficiency was detected by Western blot. Data are expressed as mean ± SD from three independent experiments. **P ≤ 0.01, ***P ≤ 0.001, as determined by Student’s t test.

Multiple flaviviral 3′SLs can bind to RISC components and exhibit higher affinities than siRNA

Multiple flaviviral 3′SLs share highly similar structure, suggesting their commonalities in functions. Thus, we examined whether 3′SL from ZIKV, DENV, JEV, WNV, and YFV have the same ability to bind to RHA and PACT. RNA pulldown assay showed that all the 3′SLs from these viruses can strongly interact with RHA and PACT (Fig. 8A). Moreover, the flaviviral 3′SLs bind RHA and PACT with higher affinities than siRNA does (Fig. 8A). To confirm this result, the kinetics of the interaction between 3′SLs or siRNA with RHA and PACT were assayed using biolayer interferometry (BLI), and the results showed that these 3′SLs bound to RHA at an affinity of approximately 2.5 × 10⁻⁶ M, which were about fivefold stronger than the binding between siRNA and RHA (Fig. 8B and C). The similar results were found in the interactions between 3′SLs or siRNA with PACT (Fig. 8B and C). Thus, our data showed a conserved role of 3′SL as a potential VSR by competing with siRNA for binding RISC among multiply flaviviruses.

Fig 8.

Multiple flaviviral 3′SLs can bind to RISC components and exhibit higher affinities than siRNA. (A) Biotinylated 3′SL RNAs from ZIKV, DENV, WNV, JEV, YFV, or siRNA were incubated with hNPC lysates. Associated proteins were eluted, separated on a 10% SDS-PAGE and probed with the antibodies indicated on the left by Western blot. (B) BLI assay. Biotinylated RNAs were captured onto streptavidin (SA) biosensors and assayed for binding to RHA or PACT at the indicated concentrations. The data collected were processed on the Gator software. (C) The binding affinities (KD) of (B) were compared.

ZIKV sfRNA protects dsRNA from dicer cleavage

As WNV sfRNA has been shown to inhibit cleavage of dsRNA by Dicer (35), we sought to examine whether ZIKV sfRNA can also protect dsRNA from Dicer-mediated cleavage. we used a classic in vitro Dicer cleavage assay, in which purified synthetic dsRNA was incubated with recombinant human Dicer (hDicer). As shown in Fig. 9, while dsRNA was efficiently processed into siRNAs by hDicer (lanes 2), the presence of ZIKV sfRNA, but not the control RNA, effectively protected dsRNA from Dicer cleavage in a dose-dependent manner (lanes 4–6). Thus, we conclude that ZIKV sfRNA can protect dsRNA from Dicer cleavage in vitro.

Fig 9.

ZIKV sfRNA protects dsRNA from Dicer-mediated cleavage. The 300-nt-long dsRNA was incubated with the recombinant human Dicer (hDicer), in the presence or absence of ZIKV sfRNA or a size-matched control RNA at 37°C for 14 h. The RNAs were separated on 7 M urea–10% PAGE.

DISCUSSION

Here, using RNA pulldown, we identified a series of RNA-binding proteins that interact with ZIKV sfRNA. These proteins are involved in virus replication, cellular response to stress and RNAi. Among these identified sfRNA-binding proteins, RHA, DHX15, STAU1, and HNRNPUL1 have been reported as ZIKV RNA-binding proteins in a previous study (46), which validating our pulldown data. However, we did not detect FMRP or DDX6, which were identified as ZIKV sfRNA binding proteins by different studies (16, 47). This discrepancy could be explained by the exclusion parameters used to determine hits and cell line differences. STAU1 has been previously reported to bind to the 5′ trailer of Ebloa virus (EBOV) gRNA (48), hepatitis C virus (HCV) 3′UTR (49), and Enterovirus 71 (EV-71) 5′UTR (50); RHA was reported to bind DENV 3′UTR (51), HCV 5’UTR (52) and classical swine fever virus (CSFV) 3′UTR (53); PACT was reported to bind DENV sfRNA (54, 55). These results highlighted the potential of ZIKV sfRNA in regulating viral replication and cytopathicity through interaction with host RNA binding proteins (RBPs) in hNPCs. The data we present here showed that knockdown of RHA or PACT promotes ZIKV replication in hNPCs. Interestingly, RHA and PACT were previously reported to facilitate DENV replication in A549 and Huh7.5.1 cells, respectively (56, 57), suggesting that host RBPs can regulate different viruses in different ways or RBPs play different roles in different cell lines.

RNAi has been established as an important antiviral defense mechanism (37, 39, 58). To overcome the suppression of RNAi, many viruses have evolved VSRs that target different components in the RNAi pathways. As for flaviviruses, YFV capsid, NS2As from ZIKV, DENV, JEV, and WNV have been reported to inhibit Dicer to cleave dsRNAs into siRNAs (38, 59). WNV sfRNA also has been shown to inhibit cleavage of dsRNA by Dicer and KUNV sfRNA has been shown to interact with Dicer (35, 36). These reported flavivirus VSRs all function in the siRNAs production step. In this study, our results indicated ZIKV sfRNA is also a potent VSR but works downstream of siRNAs production, through displacing siRNAs from RISC components RHA and PACT. Similar to our findings, some other VSRs also antagonize RNAi in the RISC loading step. For example, EBOV VP35 protein interacts with TRBP and PACT to antagonize host RNAi (29); HIV-1 RRE RNA binds to TRBP to compete with TRBP-bounds siRNAs (34). These data demonstrated that viruses can overcome multiple steps of RNAi suppression, enriching our understanding of the strategies used by viruses to counteract host defense.

RISC components RHA and PACT, containing 2 and 3 dsRBDs, respectively, are critical for RNAi pathway. RHA facilitates the formation of active RISC by promoting the interaction of siRNAs with Ago and the siRNA unwinding (21, 42). PACT is believed to bind siRNAs to help to position them along Dicer’s helicase domain, allowing for sensing thermodynamic asymmetry of siRNAs and RISC loading (43, 44). Our mapping experiments indicated that the 3′SL of sfRNA is responsible for RHA and PACT binding. This is not surprising because ZIKV sfRNA 3′SL folds to an approximate 30 bp dsRNA structure. The sfRNA with the same concentration can outcompete siRNA (Fig. 4), and 3′SL binds to RHA and PACT with higher affinities than siRNA does (Fig. 8), which may be attributed to the high-order structure of 3′SL. It will be interesting to determine the molecular-binding mechanism between the 3′SL and RHA or PACT to explain why 3′SL has higher binding affinity with RHA or PACT than siRNA. Combining previous findings, we proposed a model of flavivirus resistance to host antiviral RNAi. The capsid protein, NS2A protein, and sfRNA inhibit Dicer from cleaving dsRNA to siRNA, while the sfRNA also competes with siRNA for binding to RHA and PACT, leading to the reducing efficiency of mature RISC formation. These VSRs antagonize the antiviral RNAi pathway in different steps (Fig. 10).

Fig 10.

Model for the suppression of antiviral RNAi by flavivirus. During flavivirus infection, the virus-derived dsRNA is cleaved by Dicer into siRNAs, and then the dsRBPs (RHA and PACT) bind and help to reposition siRNAs for loading them onto Ago. Finally, the mature RISC forms and actives RNAi for viral RNA degradation. However, the viral proteins capsid and NS2A and sfRNA can inhibit Dicer from cleaving dsRNA to siRNA. The abundant sfRNA can also compete with siRNA for binding to RHA and PACT through its 3′SL, leading to the suppression of active RISC assembly and subsequent target cleavage.

Starting from hNPCs, we found that ZIKV sfRNA can antagonize RNAi through competing with siRNA in binding to RISC components. However, it is unknown whether sfRNA uses the similar mechanism in mammalian somatic and mosquito cells, which are important for flavivirus pathogenicity and transmissiom. The 3′SL of sfRNA can rescue the replication of the VSR-deficient HEV71_D23A in a RHA- and PACT-dependent manner in 293T cells (Fig. 7) implying that interfering with the maturation of RISC by sfRNA is a mechanism generally applicable in mammals. In mosquitos, although RHA and PACT homologs are not been reported, the loading of siRNA onto RISC also needs the help of RBPs such as R2D2 and Loquacious (Loq) (60, 61). Whether sfRNA can dysregulate these RBPs needs to be elucidated in the future studies.

Our results showed that knockdown of RHA or PACT enhanced ZIKV infection.

More importantly, sfRNA-deficient mutant △10 ZIKV infection was more efficiently increased by RHA or PACT depletion compared to WT ZIKV, suggesting that ZIKV sfRNA can antagonize antiviral RNAi. However, RHA and PACT were not only involved in RNAi pathway but also involved in interferon (IFN) antiviral pathway. RHA interacts with IPS-1 to sense dsRNA and PACT can active RIG-1 in response to virus (62–65). The RNAi and IFN pathway may have crosstalk in regulation of ZIKV replication. Thus, the more precise mechanism through which RHA and PACT regulate ZIKV replication needs further exploration. Nevertheless, our work provides a new perspective of the mechanism of how sfRNA combats host antiviral RNAi.

MATERIALS AND METHODS

Cells and viruses

Human neural progenitor cell (hNPC) line 15167 derived from fetal brains (Lonza) (66) was cultured as neurospheres in neurocult-XF basal medium (STEMCELL technologies) supplemented with neurocult-XF proliferation supplement (STEMCELL technologies), basic fibroblast growth factor (bFGF, 10 ng/mL, STEMCELL technologies), and epidermal growth factor (EGF, 20 ng/mL, STEMCELL technologies), Heparin solution (2 µg/mL, STEMCELL technologies). HEK293T and Vero cells were maintained in Dulbecco’s modified Eagle medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco). ZIKV strain FSS13025 (GenBank accession number KU955593) was originally isolated from a patient in Cambodia in 2010. The contemporary ZIKV strain GZ01 (GenBank number KU820898) was isolated from a patient returned from Venezuela in 2016. HEV71_D23A was produced in our laboratory (28).

Antibodies

Anti-RHA (ab183713, abcam), anti-PACT (ab75749, abcam), anti-TRBP (ab180947, abcam), anti-EIF2AK2 (ab32506, abcam), anti-STAU1(ab137100, abcam), anti-GFP (AE012, abclona), anti-β-Actin (AC026, abclonal). ZIKV envelope protein (mouse-BF-1176-46, Biofront technologies), SOX2 (rabbit-ab97959, Abcam).

RNA preparation

DNA fragments containing T7 promoters corresponding to the sfRNA1 full length, xrRNA1, xrRNA2, DB12, and 3′SL of ZIKV strain FSS13025 were obtained by PCR using the FSS13025 infectious cDNA clone as template. DNA fragment containing T7 promoter corresponding to the sfRNA1 full length of contemporary ZIKV strain GZ01 were obtained by RT-PCR using the GZ01 virus RNA as template. Then biotinylated RNAs were synthesized by in vitro transcription using RiboMAX Large Scale RNA Production Systems-T7 (Promega) according to the manufacturer’s instructions with the modification of that the components of rNTPs were changed to: 8 mM GTP, 5 mM ATP, 5 mM CTP, 1.3 mM UTP, 0.7 mM Bio-11-UTP (Thermo Fisher Scientific). Synthesized RNAs were purified using Purelink RNA mini kit (Thermo Fisher Scientific) and checked by agarose electrophoresis.

RNA pulldown and mass spectrometry

hNPCs were lysed by a lysis buffer [50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 5 mM EDTA, 10% glycerol, 10 mM dithiothreitol, and 1× cocktail protease inhibitor (YTHX Biotechnology), 100 U/mL recombinant RNAse inhibitor (Takara)]. For the pulldown, 300 g of total protein from the cell lysates was precleared with 30 L Dynabeads MyOne Streptavidin T1 beads (Invitrogen). Biotinylated RNA was previously incubated in NEBuffer3 (NEB, B7003) at 65°C for 5 min and then slowly cooled to room temperature. The precleared cell lysate was incubated with 5 pmol RNA at 4°C for 3 h; subsequently, 30 L Dynabeads MyOne Streptavidin T1 beads were added for another 2 h. After incubation, the beads were washed five times with lysis buffer. Beads were resuspended in 50 µL 2× SDS loading buffer, boiled for 10 min at 95℃. After spin, the supernatant was collected and analyzed by SDS-PAGE. Gels were stained with Coomassie Brilliant Blue. Each lane was cut into appropriate pieces, proteins were digested by Typsin, and peptides were extracted by solid phase extraction.

LC-MS/MS analysis was conducted with an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific) combined with an EASY-nLC 1000 nano system (Waltham). Then, peptide mixtures were separated using a binary solvent system with 99.9% H₂O, 0.1% formic acid (phase A) and 80% acetonitrile, 0.1% formic acid (phase B). Linear gradients of 4%–38% B for 90 min, 38%–56% B for 20 min, 56%–100% B for 6 min, and finally 100% B for 4 min, with a flow rate of 600 nl/min. The eluent was submitted to the Orbitrap Fusion Lumos MS system. The electrospray voltage applied was 2.0 kV. The full scan MS mode was operated with the following parameters: automatic gain control (AGC) target, 5e5; resolution, 12,000 FWHM; scan range, 350–1,550 m/z; maximum injection time, 50 ms and collision energy, 35%. The MS/MS mode was set as follows: automatic gain control (AGC) target, 5e4; resolution, 15,000 FWHM and maximum injection time, 22 ms.

The MS data were analyzed using MaxQuant software (version 1.5.2.8). MS data were searched against the Swiss-Prot Human database (20,240 entries). An initial search was set at a precursor ion mass tolerances of 6 ppm for peptide masses and a mass tolerance of 20 ppm for fragment ions. The search was performed with enzyme specificity trypsin, and two missed cleavages were allowed. Carbamidomethylation of cysteines was defined as fixed modification, while protein N-terminal acetylation and methionine oxidation were defined as variable modifications for database searching. The cutoffs of global false discovery rate (FDR) for both peptide and protein identification were set to 0.01. The protein identification based on N1 unique peptides, and minimal peptide length was six amino acids. The protein abundance was calculated using intensity-based absolute quantification (iBAQ) in MaxQuant software.

Northern blot

Ten micrograms of total RNA was resuspended in 1 volume of RNA loading buffer (Takara) and then incubated 15 min at 65°C and 2 min on ice. Electrophoresis was performed in denaturing 1.5% agarose gels at 50 V. After electrophoresis, the gel was transferred to a Hybond N + nylon membrane (GE Healthcare). RNA was crosslinked to the membrane using the UV crosslinker. The membrane was pre-hybridized for 30 min with PerfectHyb plus hybridization buffer (Roche) at 68°C and hybridized for 24 h at 68°C with a Biotin-labeled RNA probe which was complementary to the complete 3′ UTR (nucleotides 10,380–10,808). After hybridization, the membrane was washed in 2× saline-sodium citrate (SSC) buffer with 0.1% SDS at 68°C (3 × 10 min) and 0.5× SSC buffer with 0.1% SDS at room temperature (3 × 10 min). Then, the membrane was incubation with 0.05% tween-20 in PBS at room temperature for 1 h, followed by incubation for 1 h at room temperature with Hrp-streptavidin(Abcam), and then washed with 0.05% tween-20 in PBS (4 × 5 min). Blots were developed using an enhanced chemiluminescence (ECL) kit.

RNA immunoprecipitation

hNPCs were infected with ZIKV. Forty-eight hours after infection, the cell lysates were subjected to immunoprecipitation with anti-RHA or PACT antibody or IgG at 4°C overnight and then mixed with 20 µL of proteinA/G Magnetic beads (Thermo Fisher Scientific) and incubated for another 3 h. Beads were washed five times with lysis buffer and resuspended in 100 µL of lysis buffer. Fifty microliters of beads was used for WB and 50 µL for RNA isolation using RNeasy Mini Kit (Qiagen). The levels of RNA in IP were determined by quantitative RT-qPCR and normalized to GAPDH

control IgG IP following the formula:

$$\begin{gathered} {\displaystyle \mathrm{F}\mathrm{o}\mathrm{l}\mathrm{d}\mathrm{e}\mathrm{n}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{h}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}=2^{-\mathrm{d}\mathrm{d}\mathrm{C}\mathrm{t}}\mathrm{w}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{d}\mathrm{C}\mathrm{T}= \\ ((\mathrm{C}{\mathrm{T}}_{\mathrm{R}\mathrm{N}\mathrm{A},\mathrm{I}\mathrm{P}}-\mathrm{C}{\mathrm{T}}_{\mathrm{G}\mathrm{A}\mathrm{P}\mathrm{D}\mathrm{H},\mathrm{I}\mathrm{P}})-(\mathrm{C}{\mathrm{T}}_{\mathrm{R}\mathrm{N}\mathrm{A},\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}\mathrm{I}\mathrm{P}}-(\mathrm{C}{\mathrm{T}}_{\mathrm{G}\mathrm{A}\mathrm{P}\mathrm{D}\mathrm{H},\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}\mathrm{I}\mathrm{P}}))} \end{gathered}$$

Western blot

The samples were fractionated by electrophoresis on 10% SDS-polyacrylamide gels, and resolved proteins were transferred onto PVDF membranes. After blocking with 5% skimmed milk, the membranes were incubated with in the corresponding primary antibody, and then washed with 0.05% tween-20 in PBS (4 × 5 min), followed by appropriate horseradish peroxidase-conjugated secondary antibodies, and then washed with 0.05% tween-20 in PBS (4 × 5 min). Blots were developed using an enhanced chemiluminescence (ECL) kit.

Electrophoretic mobility shift assay

The recombinant proteins His&SUMO-RHA(1–264aa) and His&GST-PACT were expressed and purified by Sino Biological Inc. (Beijing, China). Biotin-labeled siRNA (sense 5′-GCAAGCUGACCCUGAAGUU-3′, antisense 5′-AACUUCAGGGUCAGCUUGC-biotin-3′) was synthesized by Sangon Biotech (Shanghai, China). EMSA was performed using LightShift Chemiluminescent RNA EMSA Kit (Thermo Scientific, 20158) according to the manufacturer’s instructions. Briefly, sfRNA was previously incubated in NEBuffer3 (NEB, B7003) at 65°C for 5 min and then slowly cooled to room temperature. Four micromolar of RHA or PACT protein was reacted with different concentrations of sfRNA or control RNA and 10 nM Biotin-labeled siRNA in 1× RMESA-binding buffer at 25°C for 40 min, and then 10 nM biotin-labeled siRNA was added for another 25 min at 25°C. The RNA and protein-binding reactions were resolved through a native 6% polyacrylamide gel and then transferred to a Hybond N + nylon membrane (GE Healthcare). The membrane was UV cross-linked to fixate transferred RNA to the membrane. After blocking and incubation with Stabilized Streptavidin-Horseradish Peroxidase Conjugate buffer, the membrane was washed four times. Blots were developed using an enhanced chemiluminescence (ECL) kit.

Sirna transfection

30 pmoles RHA siRNA (sc-45706 santa cruz) or PACT siRNA (sc-36175 santa cruz) or control siRNA (sc-37007 santa cruz) per well was transfected in a 24-well plate using RNAiMAX reagent (Thermo Fisher Scientific) as recommended by the manufacturer. The medium was changed at 6 h after transfection. Cells were infected as noted in the figure legends.

RNA quantitation

Viral gRNA levels were quantified by RT-qPCR using one-step RT-PCR kit (Takara). Primers used were: forward 5′GGTCAGCGTCCTCTCTAATAAACG3′; reverse 5′GCACCCTAGTGTCCACTTTTTCC3′; probe: FAM-AGCCATGACCGACACCACACCGT-BQ1.

sfRNA levels were quantified using a RT-qPCR assay designed based on the DeSCo-PCR developed by KANODIA et al. (40). One step TB green RT-PCR kit (Takara) was used for RT-qPCR reactions with forward primer(FP) plus reverse primer(RP) plus block primer(BP). A 20µL PCR reaction mix was prepared with 2 µL template and final concentration of each of the primers were as follows: 0.2 µM FP, 0.2 µM RP, 4 µM BP. The primers are: FP: 5′GTTGTCAGGCCTGCTAG3′; RP: 5′GTCCTCTGAGGGGCTCAC3′ BP: 5′GGGTCCACACCTGGAGTGCTATAAGCACCAATCTTAGTGTTGTCAGGCCACGATC3′.

Immunofluorescence analysis

Cells were fixed with 1% paraformaldehyde for 15 min and permeabilized with 0.2% Triton X-100 for 15 min, blocked with 3% bovine serum albumin for 2 h at 37°C, incubated in the corresponding primary antibody, and then washed with PBS (3 × 5 min), followed by incubating in the secondary antibody at 37°C for 1 h. After three washes with PBS, the cell nuclei were stained with DAPI. Cells were photographed under an Olympus IX73 microscope.

Biolayer interferometry

Biolayer interferometry measurements were performed on the Gator BLI system using the Gator bioanalysis software (Gator Bio). The Q buffer [PBS (10 mM PH7.4) with 0.02% Tween 20 and 0.2% BSA] was used as running buffer. Biotinylated RNAs at 50 nM were immobilized on the tips (1–2 nm immobilized). The loaded tips were then dipped into serial dilutions of either RHA protein or PACT. Curves were fit using a 1:1 model on the Gator software after subtracting the background.

in vitro dicer cleavage assay

The 300-nt-long dsRNA was prepared from the first 300 nucleotides of GFP. One hundred twenty-five nanomolar of dsRNA was incubated with 300 nM recombinant human Dicer protein (hDicer) (Origene, TP319214) and different concentrations of sfRNA in the Dicing buffer (25 mM Tris (pH 7.0), 25 mM NaCl, 2 mM DTT, 1.5 mM MgCl₂, 1% glycerol) at 37 ℃ for 14 h. The Dicer-treated RNAs were added with RNA loading buffer (Takara, 9169) and separated on 7 M urea–10% PAGE.

Statistical analysis

All results are presented as mean ± SD of three independent experiments. Data were analyzed unpaired, Student’s t-test and considered significant if P < 0.05 (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).