Y-Box Binding Protein 1 Interacts with Dengue Virus Nucleocapsid and Mediates Viral Assembly

Abstract

Infection with dengue virus (DENV) induces vast rearrangements of the endoplasmic reticulum, which allows the compartmentalization of viral RNA replication and particle assembly. Both processes occur in concert with viral and cellular proteins. Prior studies from our group suggest that the host RNA-binding protein (RBP) Y-box binding protein 1 (YBX1) is required for a late step in the DENV replication cycle. Here we report that YBX1 interacts with the viral nucleocapsid, distributes to DENV assembly sites and is required for efficient assembly of intracellular infectious virions and their secretion. Genetic ablation of YBX1 decreased the spatial proximity between capsid and envelope, increased the susceptibility of envelope to proteinase K mediated degradation, resulted in the formation of rough empty-looking particles, and decreased the secretion of viral particles. We propose a model wherein YBX1 enables the interaction between the viral nucleocapsid with the structural protein E, which is required for proper assembly of intracellular virus particles and their secretion. IMPORTANCE The global incidence of dengue virus (DENV) infections has steadily increased over the past decades representing an enormous challenge for public health. During infection, DENV viral RNA interacts with numerous host RNA binding proteins (RBPs) that aid viral replication and thus constitute potential molecular targets to curb infection. We recently reported that Y-box-binding protein 1 (YBX1) interacts with DENV RNA and is required at a late step of the replication cycle. Here we describe the molecular mechanism by which YBX1 mediates DENV infection. We show that YBX1 interacts with the viral nucleocapsid, distributes to DENV assembly sites and is required for efficient assembly of intracellular infectious virions. These results provide important insights into DENV assembly, revealing novel functions of host RBPs during viral infection and opening new avenues for antiviral intervention.

Keywords: RNA; YBX1; dengue; virus assembly.

FIG 1

YBX1 is required for production of DENV progeny. (A) CRISPR/Cas9-derived YBX1 knockout (KO) clones (KO1 and KO2) do not express YBX1 protein as determined by Western blotting. (A-E) Huh7 wild type (WT) and YBX1 KO cells were infected with DENV at multiplicity of infection (MOI) of 1. (B) Supernatant virus titers were determined by plaque assay and expressed as PFU per mL. The number of vRNA copies was determined by RT-qPCR in supernatants (C) and intracellularly (D). (E) Mock-infected and DENV-infected cells were processed for double stranded RNA (dsRNA) immunofluorescence at 24 hpi (green). Nuclei were stained with DAPI (blue). dsRNA foci per cell was determined in at least 30 cells per experiment using an in-house macro for ImageJ. Data are presented as mean ± SEM (B–D) and median ± IQR (E) from at least three independent experiments.

FIG 2

YBX1 localizes to DENV assembly sites. (A, B) Representative immunofluorescence of Huh7 cells mock-infected or infected with DENV at MOI 5 for 24 hpi and stained for the indicated markers. (C) Pearson's and Manders' correlation coefficients were extracted from the data set generated by the coloc2 plugin for imageJ. (D) At 24 hpi, Huh7 cells were processed for in situ proximity ligation assay (PLA) of E and YBX1 (upper panels) and dsRNA-YBX1 (lower panels). Nuclei were stained with DAPI (blue) and PLA signals are detected as red puncta. (E) The number of PLA signals per cell were counted in at least 30 cells per experiment using an in-house macro for ImageJ. Data are presented as median ± IQR from two independent experiments.

FIG 3

YBX1 interacts with DENV vRNA and capsid protein but is dispensable for nucleocapsid formation. Huh7 WT and YBX1 KO cells were infected with DENV at MOI 5 and processed for RNA-immunoprecipitation (RIP) or Co-IPs at 24 hpi. Cell lysates were incubated with an isotype control (IgG) and an antibody against YBX1 or capsid. Bound material was captured on Dynabeads protein G. (A) YBX1-RIP. Representative immunoblot of YBX1-pull down and total number of vRNA copies as determined in the IP by RT-qPCR. (B) IgG and YBX1-IPs were left untreated or treated with RNase A/T1 and immunoblots were performed for YBX1, capsid and envelope. (C) Densitometry analysis was performed to determine the levels of YBX1 and capsid in the IP sample treated with RNase relative to the nontreated control (lane 7 versus 6). (D) Capsid-RIP. Representative immunoblot of capsid-pull down and total number of vRNA copies as determined in the IP by RT-qPCR in WT and YBX1 KO cells. Data are presented as mean ± SEM from three independent experiments.

FIG 4

YBX1 promotes the interaction between E and C proteins. (A–B) Huh7 WT and YBX1 KO cells were infected at MOI 5 and processed at 24 hpi for (A) E and capsid PLA and (B) and E and prM PLA. Nuclei were stained with DAPI (blue) and PLA signals are detected as red puncta. The number of PLA signals per cell were counted in at least 50 cells per experiment using an in-house macro for ImageJ. Data are presented as median ± IQR from two independent experiments. (C–F) Proteinase K (PK) protection assay of protein lysates from DENV-infected WT and YBX1-KO cells. Equal amounts of protein were left untreated or treated with PK in the absence and presence of Triton X-100 and samples were analyzed by Western blotting (C). The ratio of protected E protein (D) and C protein (E) was calculated by dividing the densitometry values of the PK-treated samples by that of the nontreated control. The percentage of E protein degradation (F) was calculated by dividing the densitometry value of the 70 kDa band by the densitometry value of the entire lane. Data are presented as mean ± SEM from six independent experiments.

FIG 5

Thin section TEM images of DENV-infected resin-embedded WT (A) and YBXI KO (B–C) Huh7 cells. Virus-induced structures include vesicle packets (Vp, purple arrowheads), which are the site of viral RNA replication. DENV infectious virions are identified as electron dense particles (Vi, black arrows) in arrays (A, left panel) and as individual virions (A, right panel). Empty looking particles with rough surface are detected in high abundance in the ER lumen of YBX1 KO cells (red arrows) and sporadically in WT cells (green arrows).

FIG 6

YBX1 is required for the assembly of intracellular infectious virions. Huh7 WT and YBX1 KO cells were infected with DENV at MOI 5. At 24 hpi, cell pellets were harvested and subjected to 5 cycles of freeze and thaw. Equal amounts of protein lysates were left untreated or treated with furin for 16 h at 30°C. (A) Representative Western blot analysis of prM and M expression upon furin treatment. (B) Titration of protein lysates was determined by plaque assay. Titers are expressed as PFU per milligrams of protein. Data are presented as mean ± SEM from three independent experiments.

FIG 7

YBX1 is required for viral secretion. (A) Huh7 WT and YBX1 KO cells were infected at MOI 5 and cell pellets and supernatants were analyzed via Western blot at 18 hpi. (B–D) The intracellular levels of the indicated viral proteins was normalized to that of actin and this value was used to normalize the levels detected in supernatant for each protein. In all cases, the expression relative to WT cells is shown. Data show mean ± SEM from at least three independent experiments.

FIG 8

YBX1 is required for DENV assembly: a model. The schematic shows DENV assembly and egress steps in WT (left of the dashed line) and YBX1 KO cells (right of the dashed line) (1) There is no difference in vRNA accumulation in WT and YBX1 KO cells and we interpret this to indicate normal DENV vRNA replication, which occurs in ER invaginations in close proximity to the assembly sites. (2) NS2A binds to vRNA and recruits it to the assembly sites in both WT and KO cells. (3) Processing of capsid-prM junctions allows the formation of prM-E dimers, the dimerization of capsid proteins and its association with vRNA. (4) YBX1 binds to the nucleocapsid enabling the interaction of capsid dimers with prM-E trimers. We posit that this interaction is disrupted in YBX1 KO cells. (5) The viral membrane proteins drive the budding of assembled particles into the ER lumen. In YBX1 KO cells, budded particles are empty-looking and with a rough surface. We propose that these anomalous particles arise from an assembly defect affecting the organization of E on the host membrane, likely due to the lack of interaction between this protein with the nucleocapsid, which is mediated by YBX1. (6) Assembled particles are secreted via the secretory pathway, in a process likely mediated by NS1 and other sorting mechanism involving prM-E interacting proteins. In the anomalous particles found in YBX1 KO cells, the secretion motifs are lacking or hidden, and thus particles accumulate in the ER lumen. Created with BioRender.com.