Live imaging and interactome analysis of Zika and chikungunya viral RNAs via dual-action aptamer tag

Abstract

Arthropod-borne, positive-sense RNA viruses encompass many human pathogens posing significant health threats worldwide. Understanding intracellular dynamics and host-factor interactions of the viral genome is essential for devising effective antiviral strategies. Here, we engineer dual-labeled Zika and chikungunya viruses incorporating a Mango-II aptamer for RNA pulldown and real-time imaging, achieving high-resolution imaging of flavivirus RNA in live cells. Tandem imaging with viral replication proteins revealed intracellular dynamics of replication complexes and RNA molecules of these viruses. We identify >200 high-confidence host interactors of viral RNA and corroborated 355 hits from existing data sets. Spliceosome factor SFPQ is as a common interactor decreasing RNA production in both viruses when depleted. Intracellular colocalization of SFPQ-ZIKV and SFPQ-CHIKV RNA is confirmed through confocal microscopy imaging of infected cells. Thus, we demonstrate Mango-II aptamer tagging as an innovative strategy for identifying spatiotemporal dynamics and virus-host interactions of viral genomes in positive-strand RNA viruses.

The study describes the implementation of an RNA tag that enables visualization of Zika and chikungunya virus genomes in live cells and identifies critical host proteins that interact with the genomes for these viruses to replicate efficiently.

Introduction

Viruses from the Flaviviridae and Togaviridae families efficiently infect vertebrate and arthropod cells by hijacking various host factors to translate, replicate, and package their positive-strand (+) genomic RNA. Orthoflaviviruses such as Zika virus (ZIKV), dengue virus (DENV), and Powassan virus (POWV) from Flaviviridae, as well as alphaviruses like chikungunya virus (CHIKV) and Eastern equine encephalitis virus (EEEV) from Togaviridae, are known to cause severe diseases in humans, including polyarthritis, hemorrhagic fever, and fatal encephalitis^1–3. Zika virus infections have been reported in over 80 countries, with the Americas experiencing a large outbreak from 2015-2016 that directly caused an increased incidence of microcephaly in infants^4,5. An estimated 390 million dengue virus infections occur annually, with ~21,000 deaths attributed to the disease^6,7. Since 2005, over 2 million cases of CHIKV infection have been reported worldwide, with outbreaks occurring in over 110 countries across Africa, Asia, Europe, and the Americas⁸. Major CHIKV outbreaks are currently ongoing in the Americas, with Brazil reporting 71,578 cases, while La Réunion, France, has reported over 47,500 cases and 12 deaths as of May 4, 2025 (WHO). Alphaviruses and Orthoflaviviruses are small, enveloped, spherical virions. A single copy of their (+) strand RNA genome is packaged by several copies of a capsid protein within a lipid bilayer embedded with transmembrane envelope glycoproteins arranged in icosahedral symmetry. Both virions enter host cells via receptor-mediated endocytosis and undergo fusion of the viral and host membranes in the late endosome, releasing their viral genomic RNA (vgRNA) into the cytoplasm.

In infected cells, the orthoflavivirus vgRNA is translated by ribosomes on the endoplasmic reticulum (ER) into a polyprotein, which is cotranslationally translocated onto the ER membrane. Cleavage by host and viral proteases produces three structural proteins [capsid (C), premembrane (prM), and envelope (E)] as well as seven nonstructural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5)⁹. Replication of vgRNA is mediated by the RNA-dependent RNA polymerase (RdRp) NS5 from the replicase complex found in virus-induced replication organelles, which are formed by ER membrane invaginations where the positive-sense vgRNA and its negative-sense complement create a pseudo-circular double-stranded intermediate^10,11. Nascent (+) RNA that exits the replication compartments is designated for translation, replication, or packaging by the C protein to form virions. Live cell imaging of capsid protein traffic during orthoflavivirus infections has been performed in some published studies; however, the spatiotemporal distribution of vgRNA is not well understood^9,10,12–14. A subset of the nascent (+) RNA molecules forms subgenomic flavivirus RNA (sfRNA) fragments through incomplete degradation by host exoribonucleases, contributing to pathogenesis through interaction with host proteins^15,16. The packaged nucleocapsid of the vgRNA-C protein complex buds into the ER lumen, where it acquires a prM/E envelope, forming the immature virus, which traffics through the secretory pathway, undergoing maturation, which includes low pH-mediated prM/E rearrangement and a furin cleavage of prM before finally exiting the cell as mature virions.

Unlike orthoflaviviruses, the released alphavirus vgRNA is translated by polysomes into two nonstructural polyproteins (nsP1-3 and nsP1-4), which are processed by the nsP2 protease into four nonstructural proteins, nsP1, nsP2, nsP3, and nsP4^17,18. Initial RNA replication occurs in virus-induced replication spherules on the plasma membrane, followed by late replication on endosomal and lysosomal membranes that form cytopathic vacuole type I (CPV-I) vesicles at the perinuclear region^19–22. Following replication through a dsRNA intermediate mediated by the RdRP nsP4 from the replicase complex, a subgenomic 26S RNA is formed from the (-) RNA, which encodes the structural polyprotein CP-E3-E2-6K-E1 processed to form the capsid protein (CP), Envelope proteins E3, E2, E1 and 6 K ion channel protein, as well as a transframe protein (TF) formed by ribosomal frameshifting on 6 K²³. The N-terminal CP is released from the structural polyprotein by autoproteolytic cleavage, the remaining structural proteins are translocated and processed in the secretory pathway, and trimeric spikes of E2/E1 heterodimers are transported to the plasma membrane. The viral genomic RNA (vgRNA) assembles into nucleocapsid cores through the interaction of CP with a packaging signal and subsequently buds from the plasma membrane through a specific interaction with the cytoplasmic domain of E2, releasing virions^24–26.

Cryogenic electron microscopy (cryo-EM) and electron tomography studies have unveiled the structural aspects and life cycle stages of orthoflaviviruses and alphaviruses. Near-atomic cryoEM structures of mature and immature orthoflaviviruses, including ZIKV, DENV, tick borne encephalitis virus (TBEV), Binjari virus, Spondweni virus^24–32, and alphaviruses such as CHIKV, Sindbis virus (SINV), Mayaro virus, WEEV^33–36, have revealed details of the structural organization of virions. Resin-embedded transmission electron microscopy (TEM) and cryoelectron tomography studies have shown the organization of virus-induced membrane modifications, cytoskeletal rearrangements, and assembly sites, demonstrating extensive interactions with the host in orthoflavivirus- and alphavirus-infected cells^10,37–41.

Investigating host interactions with viral RNA using mass spectrometry and live and fixed imaging has expanded in recent years. Leveraging tools from gene knockout studies has enabled the identification of relevant interactors critical for the virus life cycle⁴². Pulldown studies targeting the viral RNA of ZIKV, DENV, CHIKV, and SINV revealed a complex network of interactions with known and non-canonical RNA binding proteins (RBPs)^43–47. Notably, previous work has relied on UV or chemical crosslinking to capture RBPs, thereby obscuring the distinction between stable and transient interactions⁴⁸. Yet identifying RBPs through the pulldown of RNA in its native context has not been feasible, which is necessary to determine the complex interactions of vgRNA during infection. The dynamic nature of viral proteins and virions in the context of virus infection has been explored using super-resolution and live microscopy^47–49; however, the lack of effective tools has impeded the investigation of vgRNA dynamics in live infected cells. Intracellular RNA visualization has been established using fluorescence in situ hybridization (FISH) in orthoflaviviruses and alphaviruses; however, this technique is limited to fixed-cell imaging, lacking temporal resolution^50–56. Live-cell imaging of RNA can be achieved by RNA aptamers, which bind to fluorescent proteins or dyes to enable imaging. This strategy was used effectively in a replicon system of the flavivirus TBEV; however, the 24x aptamer repeats required for sufficient signal, combined with the steric bulk of the fused binding protein, have prevented application in infectious viruses⁵⁷. More recently, the RNA aptamer Broccoli was used to image alphavirus RNA from SINV in live cells; however, this approach was unable to achieve single-molecule resolution⁵⁸.

The Mango aptamer, developed in 2014 by Dolgosheina et al., is a small ( < 60 nt) G quadruplex structure that binds to the fluorogenic dye thiazole orange 1-biotin (TO1-B) to generate a strong green fluorescent signal⁵⁹. In addition to the high-quality imaging properties, Mango aptamers are compatible with streptavidin column pulldown, allowing both intracellular localization and protein interactome data to be gathered from the same construct^60,61. Second-generation Mango-II-IV aptamers have been developed with stronger fluorescence and improved binding affinity for TO1-B, resulting in stable arrays of Mango-II triplets for enhanced imaging capabilities^62,63. We used the dual-action Mango-II triplet tag for high-resolution live imaging and proteomics of native viral genomic RNA in two unrelated RNA viruses, ZIKV and CHIKV, to investigate their infectious life cycles. We resolved the trafficking of viral RNA in association with modified ER for ZIKV and endo-lysosomal vesicles in CHIKV. Among several interacting partners identified and validated, the nucleic acid binding protein SFPQ (splicing factor proline and glutamine rich) emerged as a host interactor of both ZIKV and CHIKV vgRNA, which showed a significant impact on viral RNA replication upon knockdown in infected cells. Subsequent imaging revealed that SFPQ colocalizes with vgRNA in the perinuclear region in ZIKV-infected cells and in the nsP3-containing vesicles in CHIKV-infected cells. This study presents a two-pronged approach to investigating viral RNA, applicable to diverse positive single-stranded RNA viruses.

Results

Generation of Mango-II aptamer-tagged viruses

Advancements in live-cell compatible RNA imaging techniques have been significantly hindered by the existing tags, which are bulky, and the challenges in identifying non-lethal insertion sites in RNA, limiting the ability to establish viral RNA dynamics in live cells for most positive single-stranded RNA viruses. Based on the relatively small size (159 nt) and the ability to generate single-molecule resolution signal in optical microscopy, we selected the fluorogenic Mango-II aptamer triplet tag for insertion into the genome of representative members of the Flaviviridae (ZIKV) and Togaviridae (CHIKV) families. We amplified a Mango-II cassette from the pLenti-mCherry-Mango II × 24 plasmid (Addgene, #127587) containing three TO1-B binding sites for inserting into cDNA clones of CHIKV and ZIKV⁶³. We inserted the Mango-II triplet sequence, devoid of stop codons, into the 5’ end of ZIKV NS2A, a site that has been shown to tolerate a large fluorescent protein tag (ZIKV-FP)⁶⁴, (Supplementary data 1). We engineered two ZIKV clones expressing Mango-II, first with a single Mango-II triplet inserted directly onto NS2A as a peptide tag consisting of 53 amino acids (ZIKV-M), and the second where the DNA sequence corresponding to the Mango-II triplet is linked to an mCherry fluorescent protein gene, generating a dual-labeled ZIKV expressing Mango-II-mCherry fluorescent protein-tagged NS2A (ZIKV-MFP) (Fig. 1A and Supplementary data 1). We performed growth kinetic analysis for wild-type ZIKV (WT-ZIKV), ZIKV-M, and ZIKV-MFP and determined the virus titer. The maximum titer of WT-ZIKV was 2.43 × 10⁷ PFU/ml, comparable to ZIKV-M, which reached a maximum titer of 9.60 × 10⁶ PFU/ml at 96 h post infection (h.p.i.) (Fig. 1B). ZIKV-MFP was attenuated, showing a significant ~2 log reduction compared to WT-ZIKV at early 24 and 48 h.p.i. time points before reaching a maximum titer of 6.13 × 10⁶ PFU/ml at 96 h.p.i. (Fig. 1B). Together, these data show that Mango-II tag insertions support infection and can be used to obtain biologically relevant data.

Fig. 1. Design and engineering of Mango-II aptamer-tagged ZIKV and CHIKV.

A Schematic of viral genomes of WT ZIKV, ZIKV-MFP, and ZIKV-M indicating the location of Mango-II aptamer (green) and mCherry tags (red). B Graph of growth kinetic analysis of WT ZIKV, ZIKV-MFP, and ZIKV-M (MOI = 1.0) obtained from infected JEG-3 cells. (n = 3 technical replicates; error bars represent mean ± SEM). Statistical significance was determined using a two-tailed multiple lognormal t test and Holm–Sidak method compared to WT control. P values for ZIKV-MFP were 0.001569 at 24 h.p.i., 0.000421 at 48 h.p.i., and 0.006753 at 72 h.p.i. For ZIKV-M the P value was 0.003852 at 72 h.p.i. C Schematic of the organization of WT CHIKV and CHIKV-MFP viral genome RNA indicating the position of Mango-II aptamer (green) and mCherry tags (red). D Graph of growth kinetic analysis of viruses produced by WT CHIKV and CHIKV-MFP (MOI = 1.0) from infected U2OS cells. (n = 3 technical replicates; error bars represent mean ± SEM). Statistical significance was determined using a two-tailed multiple lognormal t test and Holm-Sidak method compared to WT control. P values for CHIKV-MFP were 0.000643 at 24 12 h.p.i., 0.0206 at 24 h.p.i., 0.000371 at 48 h.p.i., and 0.0215 at 72 h.p.i. Tag sequences and locations are shown in Supplementary Data 1. Source data are provided as a Source Data file.

To generate Mango-II tagged CHIKV, we used a cDNA clone of CHIKV strain 181/25, a gift from Naomi Forrester, University of Texas Medical Branch at Galveston, TX^65,66. Like ZIKV-MFP, we engineered a dual-labeled 181/25 cDNA clone containing a Mango-II triplet sequence linked to an mCherry fluorescent protein inserted into a flexible region of nonstructural protein 3 (nsP3), which has been previously shown to tolerate large insertions without affecting virus growth in other alphaviruses (CHIKV-MFP) (Fig.1C)⁶⁷. We performed growth kinetic analysis of WT CHIKV and CHIKV-MFP and quantified the production of infectious virus at indicated time points hours post infection by plaque assay. The WT CHIKV titer reached an average maximum of 4.00 × 10⁷ PFU/ml at 24 h.p.i., then decreased at later time points due to high pathogenicity and cell death (Fig. 1D). The CHIKV-MFP produced infectious virus to an average titer of 8.00 × 10⁶ PFU/ml at 24 h.p.i., and the titer continued to increase significantly compared to WT CHIKV, reaching a peak of 1.73 × 10⁸ at 48 h.p.i. (Fig. 1D)^68,69. Similar to ZIKV-FP, we also engineered a previously established fluorescent protein-tagged nsP3 CHIKV without the Mango-II triplet sequence (CHIKV-FP) as a control for imaging (Supplementary Fig. 1A–C)⁷⁰. This confirms that virus assembly and release are unaffected by the insertion of tags in the CHIKV-MFP. The stability of the tag insertions was assessed by passaging the viruses and performing RT-PCR and sequencing of viral RNA extracted from released virus particles at each passage. Sequencing results of the tagged viruses indicated that the tagged ZIKV were stable up to passage three, and tagged CHIKV were stable up to passage six (Supplementary data 1). Throughout this study, we used virus stock at passage 2 or lower to maintain retention of the Mango-II and FP tags.

Mango-II imaging of single-stranded vgRNA is validated by FISH

During replication, both alphavirus and orthoflavivirus vgRNA enter a double-stranded (dsRNA) intermediate state as the negative-sense strand is generated and becomes the template for new positive-sense vgRNA molecules. To determine if the Mango-II tag can be detected in this dsRNA replication state or exclusively as single-stranded molecules, we designed FISH probes to target the single-stranded vgRNA specifically. Starting with ZIKV, human placental cells (JEG-3) were infected with ZIKV-WT, ZIKV-MFP, or ZIKV-M, then fixed at 36 h.p.i. and permeabilized. These cells were then treated with fluorescently tagged FISH probes targeting ZIKV (+) single-stranded RNA and subsequently with TO1-B and Hoechst stains and imaged using confocal microscopy. In the ZIKV-FP control, which contains no Mango-II tag, FISH detection showed RNA in association with NS2A along the ER and in large virus-induced perinuclear compartments (Fig. 2A top); however, there is minimal fluorescence in the green channel, showing that TO1-B does not bind to untagged vgRNA. Conversely, in ZIKV-M and ZIKV-MFP, distinct punctate structures with green fluorescence were observed. This represents TO1-B bound to the Mango-II aptamer tag on vgRNA (Fig. 2A, middle, bottom). Uninfected cells showed minimal fluorescence in the green and far-red channels, indicating low background from TO1-B staining and FISH probes (Supplementary Fig. 2A, B). Pearson’s correlation (PC) analysis was performed by comparing the green (TO1-B) and far-red (FISH) channels, which revealed low overall colocalization in the negative control with an average coefficient of 0.112. In contrast, both tagged viruses, ZIKV-M and ZIKV-MFP, had average PC values over 0.600 (Fig. 2B). This confirms that Mango-II-tagged ZIKV constructs produce a sufficient signal for visualization and that the foci observed are consistent with the established localization of ZIKV vgRNA by FISH.

Fig. 2. Mango-II tagged vgRNA colocalizes with positive-strand viral RNA in fluorescence in situ hybridization (FISH).

A Fluorescent micrographs of JEG-3 cells infected with fluorescent protein-tagged ZIKV-FP (mCherry-NS2A, red), ZIKV-MFP, or ZIKV-M fixed at 36 h.p.i. and treated with (+) strand ZIKV-specific FISH probes (far red, pseudo-colored white) followed by TO1-B staining (green) and Hoechst nuclear stain (blue). B Graph of Pearson’s correlation values showing colocalization of green (535 nm) and far-red (670 nm) channels for each micrograph in (A) (n = 10 technical replicates; error bars represent mean ± SEM). P values of ZIKV-M was 1.92 × 10⁻¹³ and ZIKV-MFP was 1.82 × 10⁻¹² compared to ZIKV-FP control. C Fluorescent micrographs of U2OS cells infected with WT CHIKV or CHIKV-MFP fixed at 16 h.p.i. and treated with (+) strand CHIKV-specific FISH probes (far red, pseudo-colored white) followed by TO1-B staining (green) and Hoechst nuclear stain (blue). D Graph of Pearson’s correlation values showing colocalization of green (535 nm) and far-red (670 nm) channels for each micrograph in (C) (WT CHIKV n = 9 technical replicates; CHIKV-MFP n = 10 technical replicates, error bars represent mean ± SEM). P value of CHIKV-MFP was 1.55 × 10⁻⁷ when compared to WT CHIKV. All scale bars in (A) and (C) correspond to 20 μm. Significance in (B) was calculated by One-way ANOVA and (D) was calculated by two-tailed unpaired t test with Welch’s correction using GraphPad Prism software. Source data are provided as a Source Data file.

Next, we infected U2OS cells with CHIKV or CHIKV-MFP, fixed at 16 h.p.i., and permeabilized using 0.05% Triton X-100. The cells were then treated with FISH probes targeting CHIKV (+) single-stranded vgRNA, followed by TO1-B and Hoechst staining. In cells infected with WT CHIKV, we observed fluorescence from the FISH throughout the cytoplasm and in vesicular structures localized to the perinuclear region (Fig. 2C, top). In cells infected with CHIKV-MFP, green fluorescence from TO1-B is observed in the cytoplasm, perinuclear region, and on the plasma membrane, where early replication of alphaviruses occurs (Fig. 2C, bottom)⁷¹. Minimal fluorescence was observed in the green channel for the negative control WT CHIKV, indicating that TO1-B does not bind to the vgRNA of untagged WT CHIKV. Similarly, uninfected U2OS cells showed minimal background fluorescence from TO1-B and FISH probes (Supplementary Fig. 3A, B). Pearson’s correlation analysis showed that colocalization between FISH and TO1-B is high in CHIKV-MFP, with an average PC value of 0.57 compared to CHIKV-WT, which has minimal FISH/TO1-B colocalization with an average PC value of 0.09 (Fig. 2D). Therefore, the RNA localization detected by Mango in CHIKV is consistent with FISH, specific for the (+) single-stranded vgRNA.

The Mango-II tag does not detect double-stranded RNA

To detect dsRNA replication sites, we performed an immunofluorescence assay (IFA) using an anti-dsRNA antibody, along with TO1-B staining, in cells infected with WT-ZIKV, ZIKV-M, or ZIKV-MFP (Fig. 3A). In samples infected with WT-ZIKV, there was minimal background fluorescence in the green channel, indicating low levels of TO1-B background staining. In both Mango-II-tagged virus-infected samples, green fluorescence was observed and accumulated in the perinuclear region (Fig. 3A, middle, bottom). Green fluorescence of TO1-B in the ZIKV-MFP was observed with the red fluorescence of mCherry-NS2A, indicating that the vgRNA is likely localized to the ER. In contrast, dsRNA is observed in distinct puncta representing replication sites formed during ZIKV infection (Fig. 3A). Uninfected JEG-3 cells showed minimal background fluorescence from TO1-B and the dsRNA antibody (Supplementary Fig. 4A, B). Pearson’s correlation analysis revealed limited colocalization between TO1-B and dsRNA in ZIKV-infected samples, as all three viruses exhibited average Pearson correlation values below 0.20 (Fig. 3B).

Fig. 3. Mango-II imaging exclusively detects single-stranded vgRNA, and not dsRNA replication intermediate.

A Fluorescent micrographs of JEG-3 cells infected with WT ZIKV (top), ZIKV-M (middle), or ZIKV-MFP (bottom) at an MOI = 0.1 and fixed at 36 h.p.i. Cells were treated with J2 dsRNA antibody (far red, pseudo-colored white) followed by TO1-B staining (green) and Hoechst nuclear stain (blue). Red corresponds to mCherry-NS2A. B Colocalization of green (535 nm) and far-red (670 nm) channels for (A) measured by Pearson’s correlation (n = 10 technical replicates; error bars represent mean ± SEM). Significance was calculated by One-way ANOVA using GraphPad Prism software. n.s. p > 0.05. C Fluorescent micrograph of U2OS cells infected with WT CHIKV (top) or CHIKV-MFP (bottom) at an MOI = 0.1 and fixed at 16 h.p.i. Cells were treated with J2 dsRNA antibody (far red, pseudo-colored white) followed by TO1-B stain (green) and Hoechst nuclear stain (blue). Red corresponds to mCherry-nsP3. D Colocalization of green (535 nm) and far-red (670 nm) channels for (C) measured by Pearson’s correlation (n = 10 technical replicates; error bars represent mean ± SEM). Significance was calculated by two-tailed unpaired t test with Welch’s correction using GraphPad Prism software. n.s. p > 0.05. All scale bars in (A) and (C) correspond to 20 μm. Source data are provided as a Source Data file.

Similarly, U2OS cells infected with WT CHIKV or CHIKV-MFP were fixed and stained with dsRNA antibodies. We detected dsRNA in distinct puncta, indicating replication sites throughout the cytoplasm, in contrast to Mango-II staining, which was more diffuse (Fig. 3C). We also observed minimal background staining of TO1-B and dsRNA in uninfected U2OS cells (Supplementary Fig. 5A, B). Pearson’s correlation analysis demonstrated minimal colocalization between TO1-B and dsRNA in CHIKV-infected samples, with no statistically significant difference observed between WT CHIKV and CHIKV-MFP (Fig. 3D). The uniformly low Pearson correlation coefficients across all viruses indicate that TO1-B staining of RNA labeled with the Mango-II aptamer tag does not effectively detect the double-stranded RNA replication intermediate.

Live-cell imaging of ZIKV and CHIKV vgRNA using Mango-II aptamer tag

The primary advantage of using a fluorogenic aptamer tag is the ability to image vgRNA in live cells. During ZIKV infection, the transmembrane NS2A protein localizes broadly to the ER membrane at early time points (Supplementary Fig. 6A). To validate the subcellular localization of NS2A, we performed live imaging of ZIKV-FP with ER or Golgi markers and observed that mCherry-NS2A localized to ER in infected cells, but not to Golgi (Supplementary Fig. 6A–C, and Supplementary Movie 1, 2). As ZIKV infection modifies the intracellular membranes, NS2A localizes on the tubular ER membrane, in large ER-associated vesicles, and accumulates in the perinuclear region during late infection³⁷. In contrast, it has been previously established that CHIKV nsP3 can be observed on the plasma membrane where early-stage (3–6 h.p.i.) replication and translation occur, as well as on endosomal membranes, lysosomes, and large CPV-Is during later time points^21,72,73.

Having established the ability of Mango-II to visualize RNA molecules in infected cells, we set out to image the spatial and temporal dynamics of vgRNA molecules relative to viral proteins and associated cellular organelles during active infection. We infected Huh7.5 cells with ZIKV-MFP and stained the cells with TO1-B at 72 h.p.i. Infected cells showed NS2A-FP associated with the entire ER network of flattened membrane sheets, known as cisternae, located near the nucleus and corresponding to the rough ER (Fig. 4A and Supplementary Movie 3). In these cells, green foci indicating TO1-B bound vgRNA can be seen in distinct puncta (Fig. 4A, ROI) throughout the ER and in the perinuclear region. Notably, there is a large cluster of TO1-B-associated foci at the center of the perinuclear structure where NS2A and other viral proteins are known to accumulate. These RNA foci are not colocalized with NS2A-FP (Fig. 4A, ROI). Cells were also identified as representing tubular ER based on the red fluorescence of NS2A-mCherry localization (Fig. 4B and Supplementary Movie 4). We observed vgRNA puncta moving along the tubule structures of the ER (Fig. 4B, ROI). Large NS2A-containing vesicles were also detected to be associated with the modified ER (Fig. 4B, ROI); however, the vesicles were largely devoid of TO1-B-bound RNA. Much of the NS2A in these cells accumulates at the perinuclear region of a distinctly kidney-shaped nucleus, a documented modification of ZIKV-infected cells (Fig. 4B)³⁷. Our results show, for the first time, that ZIKV vgRNA is highly dynamic in live cells and is associated with the ER membrane during infection.

Fig. 4. Live-imaging of vgRNA in ZIKV and CHIKV infected cells.

A, B Representative images in merge (left) green (center) and red (right) channels from live-imaging experiments (Supplementary Movies 3-4) of ZIKV-MFP infected (MOI = 0.1) Huh7.5 cells stained with TO1-B (green). The white box indicates ROI shown enlarged in the top right corner. Videos started at 72 h.p.i. Scale bar indicates 20 μm. C, D Representative images in the merge (left) green (center) and red (right) channels from live-imaging experiments (Supplementary Movies 5-6) of CHIKV-MFP infected (MOI = 0.1) U2OS cells stained with TO1-B (green). Videos started at 6 h.p.i. (C) or 8 h.p.i. (D). For (A–D), the white box indicates the ROI shown enlarged in the top right corner. The scale bar in full-size images indicates 20 μm, and the scale bar in zoomed ROI panels indicates 5 μm. All movies were performed in 3 biological replicates and have a frame rate of 7 frames per second (Supplementary Movies 3–6). See also Supplementary Fig. 3 and Supplementary Fig. 4. Source data are provided as a Source Data file.

To image CHIKV-MFP, we infected U2OS cells and stained the cells with TO1-B at 6–8 h.p.i. to evaluate the dynamics of vgRNA in relation to the labeled viral protein, nsP3-mCherry. Infected cells were identified by the red fluorescence of nsP3-mCherry localization on the plasma membrane, where early replication and translation occur (Fig. 4C and Supplementary Movie 5). The green fluorescence corresponding to TO1-B-bound CHIKV vgRNA was observed on the plasma membrane and the limiting membranes of cytoplasmic vesicles, indicating possible colocalization with endosomes, lysosomes, and CPV-I, and membraneless stress granules (Fig. 4C). The vgRNA was observed colocalizing with nsP3-containing vesicles, as well as on dynamic foci containing no nsP3 and large filamentous structures (Fig. 4C, ROI). In addition to vgRNA on nspP3-associated vesicles, we also observed cells containing vgRNA in filamentous formations throughout the cytoplasm (Fig. 4D and Supplementary Movie 6). Notably, unlike ZIKV vgRNA, much of the CHIKV vgRNA was not associated with the ER membrane.

Subsequently, we performed long-term live-imaging experiments to assess the photostability of the TO1-B fluorophore and evaluate the suitability of the Mango-II aptamer for determining the vgRNA dynamics during virus infection. We infected Huh7.5 cells with ZIKV-MFP or U2OS cells with CHIKV-MFP, stained the cells with TO1-B at 48 h.p.i. or 8 h.p.i., respectively, and imaged the infected cells for 12 h. Throughout the 12-h duration of the live-cell imaging, Mango-II-labeled vgRNA remained detectable and provided sufficient signal for extended imaging (Supplementary Movie 7–8). Particle tracking of Mango-II-labeled vgRNA puncta revealed average velocities of 0.0149 microns/sec in ZIKV-MFP and 0.0241 microns/sec in CHIKV-MFP-infected cells (Supplementary Fig. 7A and 7D). The relative fluorescent intensity of TO1-B signal decreased gradually, indicating low levels of photobleaching of the TO1-B fluorophore (Supplementary Fig. 7B, C).

We next imaged infected live cells in the presence of LysoTracker, a cellular stain that stains acidic cellular vesicles, including lysosomes and late endosomes. We observed limited colocalization between vgRNA-lysosomes and NS2A-lysosomes in ZIKV-infected cells, whereas CHIKV-infected cells showed higher colocalization values of vgRNA with lysosomes (Supplementary Fig. 8A–D). This suggests that lysosomes play an important role in viral RNA interactions and trafficking for CHIKV, but not ZIKV. Our data show the colocalization of alphavirus vgRNA with a mCherry-tagged viral protein, nsP3, in live cells and reveal the potential for either their presence on the same cellular structure or a direct interaction between vgRNA and nsP3-mCherry in real time. These experiments provide proof-of-concept evidence that Mango-II-tagged viral RNA can be used to track the transport of viral genomes in infected cells, even in unrelated viruses. Furthermore, we have shown differences in vgRNA localization of the same virus, which could represent early (Fig. 4A: ZIKV and 4 C: CHIKV) and late (Fig. 4B: ZIKV and 4D: CHIKV) stages of infection, demonstrating the progression of infection. Variability in stages of infection was also observed at similar time points, potentially due to asynchronous infection. A comparison of CHIKV with ZIKV vgRNA reveals marked differences in vgRNA localization and trafficking, shedding light on the functional diversity of these RNA viruses.

Host interactome of ZIKV and CHIKV vgRNA using Mango-II tag and AP-MS

We utilized the RNA pulldown ability of the Mango-II tag via affinity purification to determine the host protein interactome of Mango-II-tagged ZIKV vgRNA. We infected JEG-3 cells in 150 mm culture plates with WT ZIKV, ZIKV-M, ZIKV-MFP, or mock infection for 36 h, then harvested the cell lysate, clarified by ultracentrifugation, and incubated with streptavidin resin derivatized with desthiobiotin to allow Mango-II aptamer binding (Fig. 5A). The bound resin was then washed and subjected to LC/MS using a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific). For ZIKV-M-infected cells, 138 proteins were enriched with fold change (FC) values ≥ 1.5 compared to mock, and 129 proteins were enriched compared to the negative control ZIKV-WT with no Mango-II tag (Fig. 5B, C). The Mango-II pulldown of ZIKV-MFP had a greater abundance of total hits with 814 proteins enriched compared to mock and 393 proteins enriched over untagged WT ZIKV (Fig. 5D, E). Comparison of the datasets showed consistent results, with 180 out of 191 total ZIKV-M hits also detected in ZIKV-MFP. ZIKV NS5 was enriched in all Mango-tagged samples, indicating a strong interaction between the RdRP and vgRNA. Gene Ontology analysis was performed on the top 200 hits, revealing that RNA binding was the most enriched function, with a false discovery rate (FDR) of 1.27 × 10⁻⁵⁰ (Fig. 6A). We also observed significant enrichment in ATP hydrolysis activity (FDR = 1.53 × 10⁻²⁵), isopeptide bond association (FDR = 2.27 × 10⁻²⁵), RNA splicing (FDR = 6.52 × 10⁻¹⁵), and host-virus interaction (FDR = 7.21 × 10⁻¹⁰), among others.

Fig. 5. Pulldown of RNA from virus-infected cells using Mango-II aptamer tag.

A Schematic illustration of pulldown protocol. Dtb = desthiobiotin and S = streptavidin bead. B, C Volcano plots showing enrichment of proteins pulled down by ZIKV-M vgRNA. B shows the P value and fold change over mock-infected cells, and (C) shows the P value and fold change over WT ZIKV-infected cells. D, E Volcano plots showing enrichment of proteins pulled down by ZIKV-MFP vgRNA. D shows the P value and fold change over mock-infected cells, and (E) shows the P value and fold change over ZIKV-infected cells. Purple circles indicate proteins of interest, and green circles indicate viral proteins. F, G Volcano plots showing enrichment of proteins pulled down by CHIKV-MFP vgRNA. F shows the P value and fold change over mock-infected cells, and (G) shows the P value and fold change over WT CHIKV-infected cells. Blue circles indicate proteins of interest, and orange circles indicate viral proteins. The threshold values (indicated by gray dashed lines) used for all volcano plots were a fold change >1.5 of and a P value of <0.05, as calculated by two-tailed multiple t test analysis followed by correction for multiple comparisons using the two-stage step-up method of Benjamini, Krieger, and Yekutieli with a desired false discovery rate of 1% using GraphPad Prism software. The technical replicates were n = 3 for all samples. The total mass spectrometry results are provided in Supplementary data 2. Comparison with the available literature data sets is shown in Supplementary data 3. See also Supplementary Fig. 9. Source data are provided as a Source Data file.

Fig. 6. The highest fold change of the top 200 hits was identified from the vgRNA pulldown results.

A Interactome map of ZIKV vgRNA with the top 200 high-confidence hits. Color is mapped to the relative log₂ of the fold change value, as indicated in the legend at the top left. B Interactome map of CHIKV vgRNA interactions with the top 200 high-confidence hits. Color is mapped to the relative log₂ of the fold change value, as indicated in the legend at the top left. The black border indicates RNA-binding proteins. Ovals indicate common hits identified in both ZIKV and CHIKV vgRNA pulldowns. Protein clusters were determined by STRING enrichment in Cytoscape.

We performed a similar Mango-II RNA pulldown and AP-MS experiment using WT CHIKV, CHIKV-MFP, and mock-infected U2OS cells to determine the host protein interactome of CHIKV vgRNA. Due to the initial low abundance of proteins from U2OS cells, the CHIKV samples were processed with peptide purification and labeling, followed by Nano-LC-MS/MS using an EASY-nLC 1200 HPLC system (SCR: 014993, Thermo Fisher Scientific), which increased the total number of detected proteins from 100 to 759. In CHIKV-MFP-infected cells, 275 proteins were enriched with fold change values ≥ 1.5 compared to mock, and 277 proteins were enriched compared to WT (Fig. 5F, G). Viral structural proteins CP, E1, and E2 were enriched over WT CHIKV-infected cells, indicating that these structural proteins interact directly with the viral RNA. Nonstructural proteins nsP1, nsP2, and nsP3 were also detected with enrichment compared to mock-infected cells; however, only nsP3 was enriched with an average fold change of 7.09 over WT CHIKV. Among the host proteins identified, RNA binding was determined to be the most enriched function by GO analysis of the top 200 hits with an FDR of 9.63 × 10⁻³⁴ (Fig. 6B), as observed in ZIKV AP-MS. We also observed significant enrichment in Intracellular non-membrane bound organelle association (FDR = 5.53 × 10⁻²²), Isopeptide bond association (FDR = 2.19 × 10⁻³³), and Nucleoplasm (FDR = 9.34 × 10⁻¹⁸), as well as enrichment in host-virus interaction (FDR = 2.39 × 10⁻⁶).

Analysis of proteins interacting with both ZIKV and CHIKV vgRNA

We compared the MFP-infected enrichment over mock from ZIKV and CHIKV vgRNA pulldowns and identified 132 high-confidence overlapping protein hits (Supplementary Fig. 9A, B). We performed a String analysis of protein-protein interaction and Gene Ontology (GO) enrichment on the list of common proteins using Cytoscape⁷⁴. The most significantly enriched molecular function was RNA binding (FDR = 2.71 × 10⁻³³). However, non-canonical RBPs were also detected (Supplementary Fig. 9). Virus-host interaction encompassed a large subgroup of these proteins (FDR = 1.25 × 10⁻⁹), indicating that common ZIKV/CHIKV interacting proteins may be necessary across many viruses (Supplementary Fig. 9). Similarly, we identified several immune response proteins (FDR = 5.89 × 10⁻⁸) (Supplementary Fig. 9). A significant number of nuclear proteins, including many splicing factors (FDR = 4.00 × 10⁻⁸) and proteins found in the nuclear periphery (FDR = 3.73 × 10⁻¹¹), were identified for both viruses despite their life cycles taking place outside the nucleus. We also observed enrichment in methylation (FDR = 1.97 × 10⁻¹⁹), ribonucleoprotein complex formation (FDR = 3.09 × 10⁻¹²), and ATP binding (FDR = 6.18 × 10⁻⁷) (Supplementary Fig. 9).

Stable knockdown of identified proteins impacts viral replication

To validate the hits identified based on the increased detection in the Mango-II pulldown experiments, we selected proteins of interest (POIs) to evaluate in more detail. Specifically, we identified protein hits with known functions in RNA binding, viral replication, or host-viral RNA interactions. We selected proteins unique to ZIKV (FXR1, CTNND1, PABPC1, PACSIN2, PUM2, QARS, TJP2, YTHDF1, RBM47, ARID3A) or CHIKV (DDX5, DDX21, TMPO, COPA, TUBB3, FASN) as well as proteins found in both vgRNA pulldown datasets (BRD4, EIF4G1, PLEC, SFPQ). All POIs for ZIKV and CHIKV had an FC ≥ 1.5 when MFP-infected virus was compared to WT ZIKV, except for PABPC1, which met this requirement when compared to the mock. The POIs found in both pulldown datasets had a FC ≥ 1.5 either in comparison to mock or WT pulldowns, except for BRD4, which had a slightly lower FC in ZIKV vgRNA pulldowns (Supplementary data 2).

To determine the impact of vgRNA-interacting proteins on viral infection, we depleted POIs from human cells through short hairpin RNA (shRNA)-driven knockdown. For ZIKV, stable cell lines were generated in JEG-3 cells. Relative gene expressions were evaluated by qRT-PCR, which confirmed all knocked-down (KD) cell lines showed a reduction of ≥35% as compared to the scrambled (SCR) control (Fig. 7A). These stable cell lines were infected with ZIKV, and cell lysates were collected at 36 h.p.i. to determine ZIKV vgRNA levels by qRT-PCR (Fig. 7B). Knockdown of FXR1 increased ZIKV RNA with a FC of 2.01 (Fig. 7B). Knockdowns of EIF4G1, BRD4, TJP2, PLEC, and PUM2 each increased the viral RNA copy number relative to SCR more than 1.45-fold as determined by qRT-PCR, demonstrating the KD had a positive impact on viral RNA production (Fig. 7B). Knockdown of ARID3A, PACSIN2, and CTNND1, had a minimal impact on vgRNA production, with RNA fold change values of 1.19, 1.17, and 0.92 respectively (Fig. 7B). In contrast, YTHDF1and QARS knockdown cell lines resulted in a decrease in viral RNA copy number relative to SCR, with fold change values greater than 0.72 (Fig. 7B). Finally, the knockdown of PABPC1, RBM47, and SFPQ resulted in the strongest negative impact on viral RNA production with FC < 0.5 (Fig. 7B). The most substantial effect was caused by the depletion of SFPQ in JEG-3 cells (FC = 0.13) (Fig. 7B). In addition to evaluating RNA copy numbers by qRT-PCR, we also tested infectious virus production. Cell culture supernatants were collected at 20 h.p.i. and virus titers were determined by plaque formation assay (Fig. 7C). ZIKV in SCR-JEG-3 cells produced a titer of 5.33 × 10³ PFU/ml at 20 h.p.i. We observed a slight increase in titer from EIF4G1-KD cells (1.20 × 10⁴ PFU/ml). Out of the 14 total KD cell lines tested, the most significant impact was observed with the depletion of RBM47 (4.26 × 10² PFU/ml) and SFPQ (8.53 × 10² PFU/ml).

Fig. 7. Protein of interest knockdown and screening reveal a significant impact of host factor SFPQ.

A qRT-PCR measuring RNA of proteins of interest in JEG-3 KD cells as the percentage relative to JEG-3 scrambled (SCR) cells (n = 3 technical replicates; mean ± SEM). B ZIKV vgRNA detected by qRT-PCR from supernatants of infected JEG-3 KD cells (MOI = 0.1) at 36 h.p.i. as fold change relative to vgRNA of infected JEG-3 SCR cells (n = 3 technical replicates; mean ± SEM). C Viral titers at 20 h.p.i. determined by plaque assay from ZIKV infected JEG-3 KD cells (MOI = 0.1) (n = 3 biological replicates; mean ± SEM, significance determined by One-way ANOVA, n.s. p > 0.05). P value of RBM47 was 0.0269 and SFPQ was 0.0403 when compared to SCR. D qRT-PCR measuring RNA for proteins of interest in U2OS KD cells as a percentage relative to U2OS SCR cells (n = 3 technical replicates; mean ± SEM). E CHIKV vgRNA detected by qRT-PCR from infected U2OS KD cells (MOI = 0.1) at 16 h.p.i. as fold change relative to vgRNA of infected U2OS SCR cells (n = 3 technical replicates; mean ± SEM). F Viral titers at 8 h.p.i. determined by plaque assay from CHIKV infected U2OS KD cells (MOI = 0.1) (n = 3 biological replicates; mean ± SEM, significance determined by One-way ANOVA, n.s. p > 0.05). G Viral titers at 24 h.p.i. from SCR or SFPQ KD JEG-3 cells infected with CHIKV and SINV (MOI = 0.1). (n = 3 biological replicates; mean ± SEM, significance determined by two-tailed unpaired t test). P value for ZIKV was 0.0091 and YFV was 0.015. H Viral titers at 24 h.p.i. from SCR or SFPQ KD JEG-3 cells infected with ZIKV and YFV (MOI = 0.1). (n = 3 biological replicates; mean ± SEM, significance determined by Student’s t test). I, J Fluorescent micrographs of ZIKV-MFP fixed 36 h.p.i. I or CHIKV-MFP fixed 16 h.p.i. J infected JEG-3 cells treated with TO1-B (green) and anti-SFPQ antibody (far-red, pseudo-colored white). Red shows mCherry-tagged ZIKV NS2A (I) or CHIKV nsP3 (J). Scale bars = 20 μm. See also Supplementary Fig. 6 and Supplementary Fig. 7. Source data are provided as a Source Data file.

To evaluate the importance of the POI identified by CHIKV vgRNA pulldown, knockdown cell lines were generated in U2OS cells and infected with CHIKV. Due to the faster life cycle of alphaviruses, supernatant for infectious CHIKV production determination was collected at 8 h.p.i., and total RNA was purified from cell lysate at 16 h.p.i. for qRT-PCR. All knockdown cell lines showed a reduction in RNA of over 30% for the POI (Fig. 7D). Depletion of EIF4G1 increased the RNA copy number by more than 2-fold as compared to SCR- U2OS cells, indicating a strong positive impact on viral RNA production (Fig. 7E). We also observed a positive effect on viral RNA production for DDX21 (FC = 1.3) and FASN (FC = 1.2) knockdown cells (Fig. 7E). Knockdown of, TUBB3, PLEC, TMPO, and BRD4 caused minimal impact on viral RNA production with fold change values ranging from 0.87-1.02 (Fig. 7E). SFPQ-KD had a negative impact on viral RNA production, much like ZIKV, with a reduction in RNA copy number (FC = 0.32) (Fig. 7E). The most substantial negative impact on viral RNA production was seen in the depletion of COPA and DDX5 with FC values of 0.24 and 0.18, respectively. Finally, titer determination by plaque assay was performed on the supernatant from infected knockdown cells collected at 8 h.p.i. The titer of WT CHIKV in U2OS-SCR cells was (2.86 × 10⁶ PFU/ml). SFPQ-KD in U2OS cells resulted in a slight reduction (1.20 × 10⁶ PFU/ml). We also observed a decrease in viral titer in TUBB3 (4.33 × 10⁵ PFU/ml) and TMPO (8.80 × 10⁵ PFU/ml) knockdown cells (Fig. 7F).

Due to the significant impact on viral RNA levels in both viruses and the reduction in titer for ZIKV in JEG-3 cells, we selected SFPQ for further characterization. SFPQ was detected above our threshold limits for fold change and P value in ZIKV-MFP over mock and ZIKV-M over WT infected cells. Furthermore, SFPQ is a canonical RNA binding protein with implications in neuronal development and neurodegenerative disease and has also been identified to play a role in the replication of other (+) ssRNA viruses^75–78. Primarily, SFPQ was selected because it had the most significant impact on viral titer out of all protein knockdowns tested.

SFPQ colocalizes with vgRNA in ZIKV- and CHIKV-infected cells

To determine if SFPQ knockdown would impact alphavirus titers when infected in JEG-3 cells, we infected JEG-3-SCR or JEG-3-SFPQ-KD cells with CHIKV and SINV and harvested the virus-containing supernatant at 24 h.p.i. (Fig. 7G). The reduction in SFPQ expression demonstrated by qRT-PCR (Fig. 7A) was further confirmed by confocal microscopy and by western bot analysis using an antibody detecting SFPQ (Supplementary Fig. 10A–D). Both viruses showed a significant decrease in viral titer of ~1 log and ~3 log PFU/ml in the knockdown cells compared to SCR cells, indicating that SFPQ is important for alphavirus production during infection in JEG-3 cells, but not in U2OS. This data indicated that there is a cell-type-dependent KD effect of SFPQ on CHIKV infection. To test whether the knockdown of SFPQ in JEG-3 cells impacted other orthoflaviviruses, we infected JEG-3-SCR or JEG-3-SFPQ-KD cells with YFV as well as ZIKV as a control and harvested virus-containing supernatant at 24 h.p.i. (Fig. 7H). In both cases, we saw a significant decrease of ~1.5 log PFU/ml in virus produced in the SFPQ knockdown cells compared to SCR cells, although the difference in titer was greater for ZIKV. Growth kinetic analysis was performed in SCR and SFPQ-KD JEG-3 cells for ZIKV, YFV, CHIKV, and SINV (Supplementary Fig. 10E–H). ZIKV and YFV showed a greater reduction in titer at earlier time points (Supplementary Fig. 10E, F). In comparison, CHIKV and SINV showed greater reduction in titer at late time points (Supplementary Fig. 10G, H), indicating different functionality of the vgRNA-SFPQ interaction in flaviviruses vs. alphaviruses. To determine whether SFPQ KD affects cell viability, we performed an alamarBlue cell viability assay on WT, SCR, and SFPQ KD JEG-3 and U2OS cells. We observed no statistically significant difference in the percent viability when comparing SCR and SFPQ-KD cells to WT JEG-3 (Supplementary Fig. 11A) or U2OS (Supplementary Fig. 11C) cells. Additionally, cell morphology and growth rates of JEG-3 (Supplementary Fig. 11B) and U2OS (Supplementary Fig. 11D) cells remained unaffected.

Next, we evaluated the intracellular localization of SPFQ with viral RNA and proteins using IFA. Glass-bottomed dishes were seeded with JEG-3 cells and infected at a multiplicity of infection (MOI) of 0.1 with either ZIKV-MFP or CHIKV-MFP. The cells were then fixed at 36 h.p.i. or 16 h.p.i. respectively and treated with an antibody targeting SFPQ. Being a predominantly nuclear protein, in both ZIKV-MFP and CHIKV-MFP-infected cells (Fig. 7I-J) and uninfected cells (Supplementary Fig. 10A), a majority of SFPQ was detected in the nucleus. In ZIKV-MFP, a small amount of SFPQ was detected in the perinuclear region, colocalizing with TO1-B stain, indicating a possible interaction with vgRNA (Fig. 7I). This experiment was repeated in ZIKV-MFP-infected JEG-3-SFPQ-KD or JEG-3-SCR-KD cells (Supplementary Fig. 10D). Confocal imaging indicated a similar extranuclear localization pattern of SFPQ and ZIKV viral RNA indicating that even with decreased SFPQ levels, the available SFPQ is still capable of interacting with the virus (Supplementary Fig. 10D). In cells infected with CHIKV-MFP, SFPQ was observed to colocalize with vgRNA in nsP3-containing vesicles (Fig. 7J). Pearson’s correlation analysis showed that colocalization between SFPQ and TO1-B is high in ZIKV-MFP (0.74) and CHIKV-MFP (0.79) (Supplementary Fig. 12A–C).

SIM reveals SFPQ-ZIKV vgRNA interaction in high-resolution

The images obtained by confocal microscopy showed the region of SFPQ-ZIKV vgRNA interaction; however, the resolution ( ~ 0.2–0.3 μm) is insufficient to show details within this perinuclear structure. We extended our imaging analysis by performing structured illumination microscopy (SIM) (resolution ~120 nm) to further examine this phenomenon. In uninfected cells, SFPQ is observed exclusively in the nucleus (Supplementary Fig. 13A). In ZIKV-M-infected cells, we observed perinuclear structures morphologically similar to those seen by confocal microscopy containing SFPQ (red) and vgRNA (green) (Supplementary Fig. 13B). This colocalization is oriented in the center of a kidney-shaped indentation of the nucleus (Supplementary Fig. 13B), and nuclear reshaping is not observed in uninfected cells (Supplementary Fig. 13A). Within the observed structure, SFPQ and vgRNA are highly colocalized, with intensity peaks overlapping within <0.5 um (Supplementary Fig. 13C, D). Pearson’s correlation analysis showed an average PC value of 0.89 when analyzing SFPQ-TO1-B colocalization outside of the nucleus (Supplementary Fig. 13E).

The dataset identified here provides information on many vgRNA-binding proteins that interact with CHIKV and ZIKV. Further exploration of proteins involved in both viral families will elucidate the common and unique strategies that distinct positive-strand RNA viruses employ to effectively hijack cellular pathways. The Mango-II aptamer tagging system, applied here to viral RNA, provides a sensitive and robust method for evaluating intracellular spatial and temporal dynamics of viral genomes. Our approach, which incorporates fluorescently tagged viral proteins, provides context for the extensive modifications these viruses induce in infected cells. Most importantly, the dual-action functionality of the Mango-II aptamer tag facilitates the identification of key RNA-binding proteins such as SFPQ, paired with validation by imaging using a single elegantly designed construct, which will provide insights into mechanisms of viral RNA interactions as well as mechanisms of proviral and antiviral strategies.

Discussion

In this study, we have developed a dual-action approach to investigate viral RNA from both imaging and proteomics perspectives by generating Mango-II aptamer-tagged genomes in both ZIKV and CHIKV. Our dual-labeled viruses were designed to maintain viral function, thus preserving the biologically relevant context of active infection. We showed that the TO1-B stain detects single stranded Mango-II tagged vgRNA but does not detect the dsRNA replication intermediate. This lack of detection is likely due to the unfolding of the Mango-II secondary structure by the viral helicases, NS3 in ZIKV and nsP2 in CHIKV. The unfolding of viral RNA secondary structure, combined with base pairing to the negative-strand RNA during replication, would potentially prevent TO1-B binding. Importantly, the differentiation of (+) strand RNA recognition provides a potential method of spatiotemporally distinguishing replicating RNA from single-stranded translating or packaging viral RNA in infected cells. The stable integration of the Mango-II tag into live infectious viruses has enabled the visualization of the viral RNA using a fluorogenic aptamer. This advancement has allowed the detection of viral RNA dynamics during infection by two positive-sense, single-stranded RNA viruses.

Viral RNA imaging in live cells has been a long-standing challenge in the field due to the limited availability of RNA tools for visualization. Our approach combines a small, state-of-the-art RNA aptamer with established fluorescent protein tagging sites in both ZIKV and CHIKV to visualize the simultaneous intracellular dynamics of viral RNA and proteins. The ZIKV vgRNA was observed mainly on and around the ER, which agrees with its highly membrane-associated life cycle. In contrast, CHIKV vgRNA exhibited more cytoplasmic and vesicular localization, which reflects the known characteristics of translation and replication in alphaviruses. The live-imaging compatibility of Mango II tag allowed us to observe ZIKV and CHIKV vgRNA dynamics along the ER or in cytoplasmic vesicles, respectively. Notably, the comparison of two disparate viruses provides rigorous validation of our results by demonstrating biological differences in TO1-B localization that are dependent on specific virus infections. Future studies can adapt this system to additional single-stranded RNA virus families to further compare the functions of different viral genomes.

The differences in subcellular organization, organellar morphology, and localization of viral components observed in this study likely represent distinct stages of flavivirus (ZIKV) and alphavirus (CHIKV) infection. As positive-stranded RNA viruses that replicate via double-stranded RNA intermediates, both families are known to induce membrane remodeling, significant changes in organellar morphology, and the formation of virus-induced structures. In flaviviruses, predominantly membrane-anchored viral replication proteins, including NS2A, are localized to both tubular and sheet ER regions that are significantly modified as infection progresses to the late stage. For alphaviruses, the replication protein nsP3 bound to RNA is associated with endosomes and lysosomes; endosome-associated nsP3 is dynamic during the early stages of infection; and in the late stage, it becomes part of lysosomes and forms CPV-I, which is largely static in the perinuclear region⁷⁹. We found that Mango-II-labeled CHIKV vgRNA exhibited a greater average velocity than ZIKV vgRNA during infection, consistent with previously reported differences in the subcellular localization and trafficking of viral genomes in alphaviruses and flaviviruses^37,80. Collectively, these results demonstrate that the Mango-II-tagged viruses faithfully recapitulate the canonical processes during viral infection and highlight the biological differences in the localization and dynamics of vgRNA. Future studies utilizing this Mango-II imaging system are needed to determine vgRNA and organelle associations throughout the virus life cycle.

The compatibility of Mango-II with streptavidin pulldown provides a significant advantage that we leveraged to identify host proteins that interact with viral RNA in the context of infected cells. We performed RNA pulldowns using the Mango-II aptamer tag for ZIKV and CHIKV. The most significantly enriched protein pulled down with ZIKV vgRNA was NS5. This protein acts as the viral RNA-dependent RNA polymerase (RdRp) with methyltransferase activity responsible for genome capping. We also identified the structural protein prM, which has not been previously shown to interact with vgRNA. Previous work from our lab and others have indicated an essential role for prM in the assembly of immature viral particles, which is supported by the data shown here^81–84. In the CHIKV RNA pulldown, non-structural proteins nsP1, nsP2, and nsP3 were detected, all of which are involved in genome replication. Unlike flavivirus NS5, the alphavirus RdRp nsP4 is stoichiometrically less abundant and rapidly degraded which agrees with the absence of nsP4 in our proteomic analysis^40,85. The most strongly enriched protein over WT CHIKV was nsP3, which was observed colocalizing with the Mango-II aptamer-tagged vgRNA in large dynamic vesicles in live-infected cells. Additionally, structural proteins E1, E2, and CP were detected, indicating that the vgRNA might interact with these proteins during nucleocapsid core assembly and virus budding. Notably the capsid protein of ZIKV was not identified by our proteomic analysis, likely due to fundamental differences in nucleocapsid core assembly and genome packaging which are not well understood for orthoflaviviruses. The ZIKV capsid-RNA binding mechanism may disrupt the RNA secondary structures in a way that is markedly different from alphavirus nucleocapsid formation.

In addition to viral proteins, we identified 814 host protein interactions for ZIKV and 275 host protein interactions for CHIKV. A cross-reference of this data with previous viral RNA pulldown studies showed appreciable agreement with the literature (Supplementary data 3). One example of an RNA-binding protein identified by ZIKV vgRNA pulldown is RBM47, which was previously demonstrated by RNA-seq to be upregulated in human cells infected with DENV⁸⁶. In this study, RBM47 knockdown in JEG-3 human placental cells led to a significant decrease in ZIKV production, suggesting a proviral role. Similarly, one RNA binding protein identified by CHIKV pulldown was DDX5 (RNA Dead-box helicase 5). DDX5 was shown to be a proviral factor, relocalizing from the nucleus to the cytoplasm in association with viral RNA in SINV⁸⁷. More broadly, DDX5 serves a known proviral function by suppressing the IFN response, thereby promoting virus infection in several viruses, including SARS-CoV, HCV, JEV, and IFV⁸⁸. Here, DDX5 knockdown in U2OS cells reduced virus replication, indicating a proviral effect. Despite having distinctly different life cycles, the vgRNA of ZIKV and CHIKV showed interactions with many of the same proteins, suggesting either similar protein interactions are being hijacked for various purposes or unrelated viruses are using identical strategies in infected cells.

Proteins of interest were selected using an inclusive approach that emphasized biological relevance and statistical enrichment. Given that many functionally important RNA-binding proteins have more transient interactions or lower enrichment in the cytoplasm, the detected interactions using our current methods can fall below the conventional cutoffs. Thus, candidate proteins were prioritized for downstream applications based on known RNA-binding functions and involvement in viral replication or overlapping interactions between CHIKV and ZIKV. Of the selected host protein interactions identified by Mango-II aptamer pulldown and characterized in this study, SFPQ showed the most significant impact on titer for both viruses when knocked down in JEG-3 cells, suggesting a cell-type dependency. SFPQ is a predominantly nuclear splicing factor and RNA-binding protein that has been shown to interact with positive single-stranded RNA viruses, such as SARS-CoV-2⁸⁹. Both ZIKV and CHIKV are neurotropic viruses that affect the human brain. It is essential to note that SFPQ is related to neuronal development and has been linked to Alzheimer’s disease as well as amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD)^75,77. The knockdown of SPFQ in JEG-3 placental cells led to a significant reduction in titer for ZIKV and CHIKV, as well as additional representative members of the Flaviviridae and Togaviridae families. We also observed the redirection of SFPQ from the nucleus in association with vgRNA to the cytoplasmic, nsP3-containing vesicles in CHIKV-infected cells and the perinuclear region in ZIKV-infected cells, which was further validated by super-resolution imaging. Although SIM analysis was limited to a small number of cells and regions of interest, the observed vgRNA-SFPQ colocalization was consistent with the perinuclear association detected by confocal microscopy in ZIKV-MFP infection.

Previous studies have demonstrated that SFPQ translocated outside the nucleus to aid in replication and translation of other (+) ssRNA viruses^90–92. SFPQ was shown to interact with initiation factors and ribosomal proteins, shifting replication toward an internal ribosome entry site-dependent mechanism⁹². We also demonstrate that SFPQ can translocate from the nucleus and potentially interact with the vgRNA. The vgRNA of ZIKV and CHIKV contains a canonical type I cap, allowing for host cap-dependent translation of viral RNAs. Given that SFPQ KD affects ZIKV and CHIKV infection, we can speculate that SFPQ interacts with vgRNA and other host proteins to modulate replication and translation or participate in the formation of replication complexes. Collectively, our data suggest that SFPQ plays a significant role in the biogenesis of both alphaviruses and orthoflaviviruses through a coordinated interaction with the vgRNA.

Limitations of this study, as evidenced by the relatively low abundance of host proteins detected in the CHIKV-infected U2OS pulldown compared to the ZIKV-infected JEG-3 pulldown, we report here that the Mango-II pulldown protocol does not work equally well in all cell types. One explanation is varying concentrations of biotin in cell growth media, which inhibit streptavidin resin binding. The U2OS cells used for CHIKV pulldown and host protein interaction screening also exhibit an attenuated immune response, which may obscure the impact of host protein knockdowns as seen with SFPQ. The results presented here do not inform the mechanism of SFPQ proviral activity, which will require future study. Notably, some of the identified interacting proteins evaluated here did not meet our statistical thresholds but were selected for functional and biological relevance.

Additionally, the imaging results here are insufficient to determine which population of vgRNA is detected. For live imaging, we did not perform z-stacks, thus the TO1-B labeled vgRNA could be moving out of the plane of focus thereby decreasing the reliability of quantification and particle tracking of the live-images. While our long-term imaging showed TO1-B signal was retained for 12 h (frames acquired every 5 min), there was evidence of photobleaching occurring during the quicker acquisition times of one min or less. Another limitation of this system is that background TO1-B fluorescence may be elevated in certain infection contexts, such as CHIKV infection, potentially reducing the specificity of viral RNA detection and necessitating validation with independent markers of infection or viral RNA localization. With this, our current labeling system is specific to the viral RNA but limited to a single viral nonstructural protein for each virus. Future studies are required to track viral structural proteins to study entry, uncoating, and early RNA trafficking during the eclipse period as 3x Mango-II repeats are likely not enough to achieve single-molecule resolution.

Live imaging of the viral structural proteins prM and E has not yet been established for orthoflaviviruses. AP-MS captured a single time-point for each virus, eliminating potential interactions exclusive to early or late infection. All SFPQ imaging was performed through immunofluorescence assay which is limited to fixed cells. While SIM was performed to assess colocalization of vgRNA and SFPQ, the spatial resolution was not greatly improved compared to other techniques available. Addressing these limitations could be achieved by coupling our system with other imaging modalities, including cryoelectron tomography (cryo-ET), correlative light, and correlative light and electron microscopy (CLEM).

In summary, this study employs a state-of-the-art aptamer tagging system to investigate the viral genomes of ZIKV and CHIKV. Our proof-of-concept strategy facilitates complementary assays through both live-imaging and AP-MS that can be used to identify and validate host protein interactions. Future work will combine this approach with other tools, such as the Sun Tag system, to study viral RNA during translation⁹³. The dual-action approach established here has the potential to revolutionize the study of genome interactions and dynamics in diverse RNA viruses.

Methods

Cell lines and viruses

JEG-3 (ATCC, #HTB-36) cells were maintained in DMEM: Nutrient Mixture F-12 (DMEM/F-12) (Thermo Fisher Scientific, #12400-024). HEK-293T (ATCC, #CRL-1573), VeroE6 (ATCC, #CRL-1586), and Huh 7.5 (a kind gift from Dr. Charles M. Rice, Rockefeller University) were maintained in Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher Scientific, #12800-028). U2OS (ATCC, #HTB-96) cells were maintained in McCoy’s 5 A Medium Modified (Sigma Aldrich, #M4892). BHK-15 (a kind gift from Dr. Richard J. Kuhn, Purdue University) was maintained in Minimum Essential Medium (MEM) (Thermo Fisher Scientific #41500-018). All media were supplemented with 10% fetal bovine serum (FBS) (Avantor Seradigm, #97068–085) and Penicillin-Streptomycin antibiotic (Corning Inc., #30-002-CI). All cells were grown at 37 °C incubation with 5% CO₂ unless otherwise noted.

ZIKV MR766 cDNA clones under a CMV promoter were used to produce DNA-launched wild-type and tagged ZIKV⁶⁴. CHIKV 181/25, SINV (Toto64), and YFV-17D cDNA clones under the SP6 promoter were in vitro transcribed and used to produce wild-type and tagged CHIKV, SINV, and YFV, respectively^65,66. Transfections were performed in HEK-293Tcells to produce virus stocks of all CHIKV and ZIKV viruses using Polyethylenimine Hydrochloride as the transfection reagent (Polysciences, #49553-93-7). Cell culture supernatants containing ZIKV were collected at 4 days post-transfection, and supernatants containing CHIKV were collected at 24 h post-transfection.

Plasmids and cloning

The pLenti-mCherry-Mango-II x 24 plasmid was obtained from Addgene (#127587), and Mango-II triplet inserts were amplified by PCR (Supplementary data 1) using Q5 High-Fidelity 2X Master mix (New England Biolabs, M0492S) with primers containing BssHII or BamHI restriction enzyme digestion sites (Supplementary data 1) and purified using a 10% acrylamide gel. A BssHII site was introduced into ZIKV cDNA between NS1-NS2A and a BamHI site was introduced into 181/25 cDNA between nsP3_385-386 (Supplementary data 1) by site-directed mutagenesis using Phusion DNA polymerase (New England Biolabs, #MO530), and Mango-II triplet was cloned into the cDNA clone by restriction digestion followed by ligation using T4 DNA ligase (New England Biolabs, 0202S). The cDNA plasmid of CHIKV-FP was generated by removing the Mango-II triplet from CHIKV-MFP via site-directed mutagenesis using Phusion DNA polymerase (New England Biolabs, #MO530). The cDNA plasmid sequences containing the Mango-II triplet and/or mCherry inserts were confirmed by Nanopore whole plasmid sequencing at the Penn State Genomics Core facility. Resulting sequences corresponding to the final tagged clones are shown in Supplementary data 1. The plasmids pHAGE2 GFP-Atl1 (#86679) and EGFP-GalT (#11929) for stable cell production were obtained from Addgene.

Plaque assays

Viral titers were determined by plaque assays in Vero E6 cells for all ZIKV constructs and BHK-15 for CHIKV, YFV, and SINV. Cells were plated into 24-well plates one day prior to infection to form monolayers. On the day of infection, media was removed, and virus stocks were diluted in DMEM (ZIKV) or MEM (CHIKV, YFV, and SINV) containing 10 mM HEPES (N-2-hydroxyethylpiperazine-N’−2-ethanesulfonic acid; Sigma Aldrich, #HO887) and pipetted directly onto cells. The infection continued for 1 h at 37 °C with rocking and was overlaid with 500 μL of growth media containing 1.5% colloidal cellulose (Sigma-Aldrich, #435244) and 1% FBS. Plates were incubated for 6 days (ZIKV) or 4 days (CHIKV, YFV, and SINV). Plaques formed were counted after fixing the cells with neutral-buffered formalin and staining with Crystal violet (Thermo Fisher Scientific, #21932)⁸¹.

Growth kinetics

For the Growth kinetics of ZIKV, ZIKV-M, and ZIKV-MFP, JEG-3 cells grown on 6-well plates were infected with the virus at MOI = 1.0, and 200 μL of supernatant was collected and replenished with fresh media every 24 h for 4 days. For wild-type, CHIKV, CHIKV-MFP, and CHIKV-FP growth kinetic analysis, U2OS cells were infected with the virus at MOI = 1.0, and 200 uL of supernatant was collected and replenished with fresh media at 12, 24, 48, and 72 h.p.i. Virus titers were determined by plaque assays. For growth kinetic analysis in JEG-3 cells of ZIKV, CHIKV, YFV-17D, and SINV (Toto64), cells were infected with the viruses at MOI = 0.1. Supernatants (200 μL) were harvested at the time points indicated in the graphs, and virus titers were determined by plaque assays as described above.

Confocal and super-resolution microscopy

All confocal micrographs and videos were acquired using a Nikon A1R confocal microscope with a heated 60x oil immersion objective and 1.4 numerical aperture. For live imaging, an environmentally controlled microscope stage-top imaging chamber (Tokai Hit, Fujinomiya, Shizuoka Prefecture, Japan) was used to maintain temperature (37 °C) and CO₂ concentration (5%) conditions. Cells were imaged in either FluoroBrite DMEM (Thermo Fisher, #A1896701) for live-cell imaging or FluoroSave (Millipore Sigma, #345789) for fixed-cell imaging, both supplemented with 140 mM KCl (Millipore Sigma, PX1405-1) and 1 mM MgCl₂ (New England Biolabs, #B0510A) to facilitate aptamer folding. For live imaging of Huh7.5 cells, single frames were captured every 30 seconds for 8.5–16.5 minutes (Videos S3, S4, and S6) or every 15 seconds for 10 minutes (Video S5). Videos were set to play back at 10 frames per second. For structured illumination microscopy (SIM) of fixed cells, the motorized inverted Nikon ECLIPSE Ti-E TIRF microscope with the perfect focus system was used with CFI SR Apochromat TIRF 100×oil objective (NA1.49). Laser and emission band-passes used were as follows: blue, excitation of 405 nm and emission of 425–475 nm; green, excitation of 488 nm and emission of 500–550 nm; red, excitation of 561 nm and emission of 570–620 nm; and far-red, excitation of 640 nm and emission of 660–740 nm. Nikon NIS Elements software was used for all image acquisition and analysis. Linear lookup tables (LUTs) were used to adjust pixel brightness to improve visualization and contrast for clarity Gaussian convolution processing was used to reduce image noise. Videos were processed using ImageJ software⁹⁴.

Mango-II aptamer tag staining and imaging

Cells used for Mango-II imaging were stained in 250 nM TO1-B Mango-II aptamer stain (Applied Biological Materials, #G955). For fixed cells, the stain was diluted in PBS, and for live cells, the stain was diluted in growth media. The respective diluted stain was supplemented with 140 mM KCl and 1 mM MgCl₂ to facilitate aptamer folding for both fixed and live-cell imaging. Cells were incubated with stain for 45–60 min at 37 °C and washed twice with PBS supplemented with 140 mM KCl and 1 mM MgCl₂ prior to imaging.

Fluorescence in situ hybridization (FISH)

Custom Stellaris™ RNA FISH Probes were designed against the viral genomes of ZIKV and CHIKV by using the Stellaris RNA FISH Probe Designer (LGC, Biosearch Technologies, Petaluma, CA) available online at www.biosearchtech.com/stellarisdesigner (version 4.2). Probes set containing 48 unique probes targeting the viral RNA were labeled with Quasar 670 (Biosearch Technologies), following the manufacturer’s instructions available online at www.biosearchtech.com/stellarisprotocols. For imaging, ZIKV-infected JEG-3 cells or CHIKV-infected U2OS cells grown in glass-bottomed dishes (Ibidi, #81218-200) were fixed in 3.7% formaldehyde (VWR, #10790-710) and permeabilized using 70% ethanol according to the Stellaris FISH protocol. Cells were washed with Stellaris Wash Buffer A (LGC Biosearch Technologies, #A-SFM-WAI-60) and then incubated in a hybridization solution containing FISH probes diluted 1:100 in Stellaris Hybridization Buffer (LGC Biosearch Technologies, #SMF-HBI-10). After overnight incubation at 4 °C, cells were washed in Stellaris Wash Buffer A for 30 min, then stained with Hoechst and TO1-B Mango-II aptamer stain. Finally, cells were washed with Stellaris wash buffer B (LGC Biosearch Technologies, #SMF-WBI-20).

Immunofluorescence Assay (IFA)

JEG-3 or U2OS cells were grown to a confluency of 25–50% in glass-bottomed dishes (Ibidi, #81218-200), then infected with ZIKV or CHIKV, respectively, at an MOI of 0.1 and incubated at 37˚C and 5% CO₂ for 48–72 h. At indicated time points, the cells were fixed with 3.7% paraformaldehyde for 15 min and permeabilized with 0.5% Triton X-100 diluted in PBS. After permeabilization, cells were washed thrice with PBS and blocked in 10 mg/mL bovine serum albumin (BSA) (Sigma, #A7906) diluted in PBS for 1 h before adding the primary antibody. The Primary antibodies used were SCICONS anti-dsRNA antibody (J2) (1:400 dilution) (Exalpha #10010500) and Anti-SFPQ antibody (1:200 dilution) [EPR11847] (Cell Signaling, #ab177149). After overnight incubation at 4 °C, the primary antibody was removed, and cells were washed thrice with PBS. Cells were then treated with a secondary antibody. IRDye® 680RD Goat anti-Mouse IgG Secondary Antibody (LI-COR, # 926-68070) (1:800 dilution) was used to detect dsRNA, and IRDye® 680RD Goat anti-Rabbit IgG Secondary Antibody (LI-COR, #926-68071) (1:800 dilution) or TRITC Goat anti-Rabbit IgG (H + L) Secondary Antibody (Thermo Fisher Scientific, 31670) (1:200 dilution) to detect SFPQ. Secondary antibodies were diluted 1:200 in 10 mg/mL BSA for 1 h, followed by staining with Hoechst stain and TO1-B Mango-II aptamer stain (Applied Biological Materials #G955) diluted in PBS containing 140 mM KCl and 1 mM MgCl₂. Cells were subsequently washed thrice with PBS containing 140 mM KCl and 1 mM MgCl₂ before imaging using confocal or SIM.

Affinity-purification mass-spectrometry

The RNA pulldown protocol was adapted from Panchapakesan 2018⁵⁸. All buffers were supplemented with 140 mM KCl and 1 mM MgCl₂. JEG-3 or U2OS cells were grown in 150 mm culture dishes (Thermo Fisher Scientific, #150468) to produce monolayers and infected with wild-type and Mango-II-tagged ZIKV or CHIKV, respectively, at an MOI of 1.0. Infection was allowed to continue at 37 °C for 48 h for ZIKV and 16 h for CHIKV. Cells were washed with cold PBS and harvested using a cell scraper. Cells were then resuspended in cold PBS supplemented with Protease Inhibitor Cocktail (Millipore Sigma, #P8340) and sonicated using a microtip (Branson) to lyse the cells. The lysate was clarified by centrifugation at 100,000 × g for one h using a TLA 120.2 rotor in an Optima TLX centrifuge (Beckman Colter). Strep-Tactin Sepharose beads (IBA Lifesciences, #2-1201-002) were prepared according to the manufacturer’s protocol and then derivatized with 5.98 nM TO1-3PEG-Desthiobiotin (Dtb) from Applied Biological Materials, #G7956 for 15 min at room temperature. The beads were washed once to remove excess Dtb. The supernatant from the clarified sample cell lysate was then incubated with Dtb-derivatized beads for 60 min, washed twice with PBS, and resuspended in wash buffer (IBA Lifesciences, #2-1003-100). The beads were subjected to Mass Spectrometry analysis at the University of Indiana Proteomics Core facility.

On bead digests of pulldown samples

After washing, the beads were treated with 8 M Urea, 100 mM Tris hydrochloride, pH 8.5, and then reduced with 5 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP, Sigma-Aldrich Cat No: C4706) for 30 min at room temperature to break the disulfide bonds. The resulting free cysteine thiols were alkylated using 10 mM chloroacetamide (CAA, Sigma-Aldrich, Cat No: C0267) for 30 min at room temperature (RT), protected from light. Samples were diluted to 2 M Urea with 50 mM Tris, pH 8.5, and proteolytic digestion was carried out overnight at 35 °C with Trypsin/LysC Gold (0.4 µg, Mass Spectrometry grade, Promega Corporation, Cat No: V5072). After digestion, samples were quenched with 0.4% trifluoroacetic acid (TFA, v/v, Fluka Cat No: 91699).

LC-MS of ZIKV vgRNA pulldown samples

Approximately 1/15th of each IP sample was loaded onto a 5 cm C18 trap column Acclaim™ PepMap™ 100 (3 µm particle size, 75 µm diameter; Thermo Scientific, Cat No: 164946), followed by a 25 cm EASY-Spray column (Thermo Scientific, Cat No: ES902) and analyzed using a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific) operated in positive ion mode. Solvent B was increased from 5–35% over 100 min to 90% over 2 min, back to 3% over 2 min (Solvent A: 95% water, 5% acetonitrile, 0.1% formic acid; Solvent B: 100% acetonitrile, 0.1% formic acid). A data-dependent top 15 method was employed, utilizing an MS scan range of 350–2000 m/z, a resolution of 70,000, an AGC target of 3e6, and a maximum IT of 200 ms. MS2 resolution of 17,500, scan range of 200–2000 m/z, normalized collision energy of 30, isolation window of 4 m/z, target AGC of 1e5, and maximum IT of 150 ms. Dynamic exclusion of 10 sec, charge exclusion of 1, 7, 8, >8, and isotopic exclusion parameters were used.

Sample preparation and Nano-LC-MS/MS of CHIKV vgRNA pulldown samples

Following bead digestion, samples were acidified with trifluoroacetic acid (TFA, 0.5% v/v) and desalted on Pierce C18 spin columns (Thermo Fisher, Cat No: 89870) using a wash of 0.5% TFA followed by elution in 70% acetonitrile and 0.1% formic acid (FA). Mass spectrometry was performed utilizing an EASY-nLC 1200 HPLC system (SCR: 014993, Thermo Fisher Scientific) coupled to an Exploris 480™ mass spectrometer with FAIMSpro interface (Thermo Fisher Scientific). 1/5th of each fraction was loaded onto a 25 cm Aurora column (Ion Opticks) at 350 nL/min. The gradient was held at 5% B for 5 min (Mobile phases A: 0.1% formic acid (FA), water; B: 0.1% FA, 80% Acetonitrile (Thermo Fisher Scientific Cat No: LS122500)), then increased from 4-30%B over 98 min; 30–80% B over 10 mins; held at 80% for 2 min; and dropping from 80–4% B over the final 5 min. The mass spectrometer was operated in positive ion mode, with a default charge state of 2, advanced peak determination, and lock mass of 445.12003. Three FAIMS CVs were utilized (−40 CV, −55 CV, −70CV), each with a cycle time of 1.3 s and with identical MS and MS2 parameters. Precursor scans (m/z 375-1500) were done with an orbitrap resolution of 120000, RF lens% 40, automatic maximum inject time, standard AGC target, minimum MS2 intensity threshold of 5e3, MIPS mode to peptide, including charges of 2 to 7 for fragmentation with 30 S dynamic exclusion. MS2 scans were performed with a quadrupole isolation window of 4 m/z, 30% HCD CE, 15000 resolution, standard AGC target, automatic maximum IT, and fixed first mass of 110 m/z.

Data analysis

Data were analyzed using Proteome Discoverer 2.5.0.400 (Thermo Fisher Scientific). A Homo sapiens reference proteome database (UniProtKB/TrEMBL; 78806 sequences downloaded 05132022), plus ZIKV or CHIKV with mCherry sequences plus common laboratory contaminants (73 sequences including streptavidin) was searched using SEQUEST HT. Precursor mass tolerance was set to 10 ppm, and fragment mass tolerance was set at 0.02 Da with a maximum of 3 missed cleavages. A maximum of 3 modifications were allowed per peptide. Dynamic modifications include methionine oxidation phosphorylation on serine, threonine, and tyrosine, and dynamic protein terminus modifications were acetylation, met-loss, and met-loss plus acetylation. Static modifications were carbamidomethylation on cysteines. Percolator false discovery rate (FDR) filtration of 1% was applied to both the peptide-spectrum match and protein levels. Search results were loaded into Scaffold Q + S Software (version 5.2.2, Proteome Software, Inc.) for visualization. Fold change values were calculated as the ratio of detected peptide counts in Mango-containing samples relative to the reference sample, either the WT virus or mock-infected cells. Statistical analysis of protein enrichment was performed in GraphPad Prism using a two-tailed unpaired parametric t test with individual variance tested for each protein across three technical replicates. To account for multiple testing across the proteomics datasets, p values were corrected using the two-stage step-up method of Benjamini, Krieger, and Yekutieli with a desired false discovery rate of 1.00%. Adjusted p values were used to determine the statistical significance. Data points with no calculable P value were removed.

Generation of lentiviral vectors and stable cell lines

HEK-293T cells were grown to a confluency of 50–75% in DMEM media supplemented with 10% FBS and then cotransfected with pSPAX2 (Addgene, plasmid #12260), pMD2.G (Addgene, plasmid #12259) and the shRNA corresponding to the gene of interest (Millipore Sigma, Mission)(Supplementary data 1) in OptiMEM (Thermo Fisher Scientific, #22600-050) at an approximate concentration ratio of 200 ng: 350 ng: 200 ng respectively using lipofectamine 2000 (Invitrogen, #11668–019). After overnight incubation, the media were replaced with fresh DMEM supplemented with 10% FBS, and the supernatant containing lentiviruses was harvested at 24 and 48 h.p.t and stored at −80 °C until use.

JEG-3 or U2OS cells were transduced with 50 μl of lentivirus per well and rocked at room temperature for 5 min. Plates were then spun at 841 × g for 20 min at 20 °C to increase the efficiency of transduction, a technique known as Spinoculation. After overnight incubation, the media were replaced with fresh growth media and incubated for an additional 24 h. Transduced cells were selected in media containing 1.25 μg/ml puromycin (selection media), which was replaced every 48 h for at least 6 days to obtain stable cell lines

Quantitative reverse transcription PCR

Stable cell lines were grown to 90% confluency for infection. The knockdown JEG-3 stable cells were infected with ZIKV, and U2OS stable cells were infected with CHIKV. Cell culture supernatants containing released virus particles were collected for plaque assays, and total RNA from cell lysates was harvested from ZIKV-infected JEG-3 cells at 36 h.p.i. and from CHIKV-infected U2OS cells at 16 h.p.i. RNA was purified using a Quick-RNA miniprep kit (Zymo, #ZR1055) from cell lysates to generate total cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, #4374967). The qRT-PCR reaction was prepared in triplicate using PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific, #A25742) for each sample. Primers aligning to the shRNA target gene, the ZIKV NS2A gene, the CHIKV nsP1 gene, and the host beta-actin gene were used to determine the relative intracellular RNA levels (Supplementary data 1). Knockdown efficiency and change in viral RNA levels were measured using the comparative Ct method (∆∆Ct).

Mango-II tag stability (reverse transcription PCR)

VeroE6 cells or BHK-15 were grown to 90% confluence for infection. WT ZIKV, ZIKV-FP, ZIKV-M, and ZIKV-MFP were passaged in VeroE6 and WT-CHIKV, CHIKV-MPF, and CHIKV-FP were passaged in BHK-15 cells. Cell culture supernatant containing released virus particles were collected at 3 or 4 days post-infection for CHIKV and ZIKV, respectively. RNA was purified using a Quick-RNA miniprep kit (Zymo, #ZR1055) from supernatant to generate cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, #4374967). Primers were designed to amplify the region containing the Mango-II tag (Supplementary data 1). The amplified cDNA was purified using the GenCatch™ Advanced PCR extraction kit (Epoch Life Sciences, #23-60250) and the presence of the tag was analyzed by Sanger sequencing.

Western blot

Stable cell lines were grown in 6-well plates to 100% confluency and lysed using 250 μL of 1X buffer (LI-COR, #928-40004). Cell lysates were run on SDS-PAGE and then transferred to the nitrocellulose membrane (BioRad #162-0115). After the transfer was complete, membranes were allowed to dry overnight, then rehydrated in TBST (10 mM Tris, 100 mM NaCl, 0.1% Tween 20) and blocked using LI-COR blocking buffer (#927-60001) for 1 h. Primary antibodies targeting SFPQ (Cell Signaling, #ab177149), and control β-actin Mouse mAb #3700 (Cell Signaling, #8H10D10) were diluted 1:2000 in LI-COR buffer supplemented with 0.2% Tween-20 and incubated with membranes overnight with gentle rocking at 4 °C. Membranes were then washed 3 times in TBST and incubated with IRDye® 800CW Goat anti-Rabbit Ig, (LI-COR, #827-08365) and IRDye® 680RD Goat anti-Mouse IgG (LI-COR, #926-68070) secondary antibodies were diluted 1:20000 in LI-COR buffer supplemented with 0.2% Tween-20 for 45 min with gentle rocking at room temperature. Membranes were washed again in TBST and then imaged using the LI-COR Odyssey F series 10-channel image scanner.

Cell viability assay

All experimental conditions were performed with six replicates in a 96-well plate. Cells were seeded at a density of 1.0 × 10⁴ cells per well and allowed to adhere for 16 h. After, 90 µl of OptiMEM and 10 µL of alamarBlue cell Viability Reagent (Thermo Fisher Scientific, #A50100) was added to each well and incubated for 3 h at 37 °C and 5% CO₂. Reduction of alamarBlue dye was determined by measuring the absorbance at 570 nm using 600 nm as a reference wavelength in a Spectramax M5 plate reader. Cell viability was calculated by subtracting background fluorescence (blank wells containing OptiMEM plus alamarBlue with no cells) and expressing each condition as a percentage of the WT control cells to the SCR and SFPQ KD cells using the following equation: % Cell viability = (Absorbance of KD cells - Absorbance of blank)/(Absorbance of WT cells – Absorbance of blank) x 100.

Author contributions

A.J. and B.M.K. performed the experimental studies and manuscript writing. A.N. supervised experimental studies. J.J. supervised all work and edited the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

J.J. discloses support for the research of this work from the National Institutes of Health grants 1R01AI176505, 1R01AI168169, 1R21AI178533, 1R01AI177352 and United States Department of Agriculture for the NACA 58-3022-2-038 award from USDA.

Data availability

The mass spectrometry data generated in this study have been deposited in the ProteomeXchange partner MassIVE under accession code MXV000098425, cross-referenced as PXD065747. The data supporting the findings of this study are provided in the Supplementary Information and Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77830-4.