Flaviviruses induce ER-specific remodelling of protein synthesis

Abstract

Flaviviruses orchestrate a unique remodelling of the endoplasmic reticulum (ER) to facilitate translation and processing of their polyprotein, giving rise to virus replication compartments. While the signal recognition particle (SRP)-dependent pathway is the canonical route for ER-targeting of nascent cellular membrane proteins, it is unknown whether flaviviruses rely on this mechanism. Here we show that Zika virus bypasses the SRP receptor via extensive interactions between the viral non-structural proteins and the host translational machinery. Remarkably, Zika virus appears to maintain ER-localised translation via NS3-SRP54 interaction instead, unlike other viruses such as influenza. Viral proteins engage SRP54 and the translocon, selectively enriching for factors supporting membrane expansion and lipid metabolism while excluding RNA binding and antiviral stress granule proteins. Our findings reveal a sophisticated viral strategy to rewire host protein synthesis pathways and create a replication-favourable subcellular niche, providing insights into viral adaptation.

Fig 1. ZIKV polyprotein biogenesis requires the SRP-dependent co-translational ER translocation pathway.

(A) Schematic of the ZIKV polyprotein showing the structural (C, prM, E) and non-structural (NS1-NS5) proteins, with transmembrane domains indicated. (B) Overview of the different ER-targeting pathways tested for involvement in ZIKV polyprotein synthesis, including the SRP-dependent co-translational translocation pathway (via Sec61 translocon), SRP-independent post-translational translocation (via Sec62 and Sec61), chaperone-mediated pathway, tail-anchored protein insertion (via TRC40 and TMCO1), mRNA targeting to ER (via RRBP1), and multi-pass transmembrane protein insertion (via EMC). Schematic was generated using Biorender. (C) Immunoblots validating siRNA-mediated depletion of factors from the ER targeting pathways in Huh7 cells, with GAPDH as loading control; NT is non-targeting control. Samples were collected 72 hours post-transfection with siRNAs to allow for optimal protein depletion. For TMCO1, samples were collected at 24 hours post-transfection. (D) Plaque assay quantification of ZIKV titres from supernatants of infected Huh7 cells at 48 hours post infection, following depletion of the indicated ER targeting factors. Data represent mean ± SD (n = 3). (E-G) Time course of infectious ZIKV production in HeLa cells depleted of indicated genes in comparison to siNT control. Viral copy numbers were determined by RT-qPCR quantification. Data represent mean ± SD (n = 3). (H) Dose-dependent inhibition of ZIKV production in Huh7 cells by Eeyarestatin I (ESI). Viral copy numbers were determined by RT-qPCR quantification at the indicated times post infection. Statistical significance was determined by two-tailed unpaired Student’s t-test, with * p < 0.05, ** p < 0.01, *** p < 0.001. (See also S1 Fig).

Fig 2. Zika virus polyprotein biogenesis is SRP54-dependent but SRα/SRβ-independent.

(A) Schematic showing the interaction between the signal recognition particle (SRP) containing the SRP54 subunit and the SRP receptor (SR) composed of the SRα and SRβ subunits on the ER membranes. Schematic was generated using Biorender. (B-C) HeLa cells depleted of SRα, SRβ or SRP54 by siRNA were transfected with plasmids encoding (B) FLAG-tagged ZIKV prME protein or (C) FLAG-tagged KPNA2, a cytosolic protein, and analysed by immunoblotting. GAPDH served as loading control; siNT represents non-targeting control siRNA. (D) Quantification of relative ZIKV E and KPNA2 protein levels from (B-C). Bars represent mean ± SD (n = 3), with protein levels normalised to siNT control. (E) ZIKV growth kinetics in HeLa cells depleted of SRα, SRβ or SRP54. Viral RNA in supernatants was quantified by RT-qPCR. Data represent mean ± SD (n = 3). (F) Immunoblot analysis of ZIKV NS5 protein levels at 24 and 48 hours post infection (hpi) in HeLa cells depleted of SRα, SRβ or SRP54. GAPDH served as loading control. (G) Influenza A virus (H1N1 strain) growth kinetics in cells depleted of SRα, SRβ or SRP54. Viral titres in supernatants were measured by plaque assays on MDCK cells. Data represent mean ± SD (n = 3). (H) Immunoblot analysis of influenza hemagglutinin (HA) protein levels at 24 and 48 hpi cells depleted of SRα, SRβ or SRP54. GAPDH served as loading control. Statistical significance was determined by two-tailed unpaired Student’s t-test, with * p < 0.05, ** p < 0.01, *** p < 0.001. See also S2 Fig

Fig 3. ZIKV infection alters host membrane protein synthesis in a SR-independent manner.

(A) Workflow for quantifying nascent protein synthesis using L-azidohomoalanine (AHA) labelling. (B) Analysis of host and viral protein synthesis in ZIKV-infected HeLa cells depleted of SRα, SRβ or SRP54. Cells were pulsed with AHA for 4 hours at 24 hpi, and nascent proteins were captured on streptavidin beads and immunoblotted as described in (A). GAPDH and EGFR were probed as representative cytosolic and membrane proteins, respectively. (C) [³⁵S]-labelled mock- or ZIKV-infected HeLa cells depleted of SRα or SRβ were separated into cytosolic and membrane fractions and analysed by autoradiography. Red arrowheads indicate proteins specifically inhibited in SRP54, but not SRα or SRβ-depleted cells. (D) Quantification of [³⁵S]-labelled proteins in cytosolic and membrane fractions using scintillation counting. Graphs show relative total protein levels over time, normalised to the mock siNT control. Data represent mean ± SD (n = 3).

Fig 4. ZIKV non-structural proteins interact with SRP and the translocon to facilitate ER targeting.

(A) FLAG-tagged ZIKV non-structural proteins NS2A, NS2B-NS3, NS4A, NS4B, and NS5 were expressed in HeLa cells and immunoprecipitated using anti-FLAG antibodies. Lysates and co-precipitating proteins were analysed by immunoblotting. GAPDH served as a loading control in lysates. (B) Schematic representation of altered interaction of SRP with the viral non-structural proteins in virus infected cells. Schematic was generated using Biorender. (C) Mock- or ZIKV-infected HeLa cells were metabolically labelled with [³⁵S]-cysteine/methionine for 30 min and chased for the indicated times. Endogenous SRα was immunoprecipitated and co-precipitating proteins were analysed by autoradiography. (D) Immunoblot analysis of SRα immunoprecipitates from mock (M) and virus-infected (Z) cells. (E, F) Structural models of SRP54 interactions with ZIKV NS2B-NS3 (E) generated using AlphaFold. The SRP54 NG domain forms a key interface (F, inset) for binding to the viral proteins. (G) HeLa cells ectopically expressing inducible viral NS2B-NS3 were labelled with [³⁵S]-cysteine/methionine. Lysates were immunoprecipitated with anti-SRα and co-precipitating proteins were analysed by autoradiography. (H) Control or SRα depleted HeLa cells were transfected with prME-FLAG and NS2B-NS3 plasmids, metabolically labelled with [³⁵S]-cysteine/methionine and chased for the indicated times. E-protein and NS3 were immunoprecipitated on anti-FLAG and anti-NS3 antibodies and visualised by autoradiography. See also S3 Fig.

Fig 5. Time course of SRα and NS4A localisation during ZIKV infection.

(A) Immunofluorescence analysis of SRα (green) and ZIKV NS4A (red) localisation in HeLa cells at 10, 24, and 48 hours post-infection (hpi) with ZIKV. Nuclei were stained with DAPI (blue). At 10 hpi, both SRα and NS4A were diffusely distributed throughout the cell, with no apparent co-localisation. At 24 hpi, NS4A began to form distinct puncta, likely representing viral replication complexes, which partially co-localised with SRα. By 48 hpi, NS4A puncta became more prominent and numerous, while SRα co-localisation decreased. Scale bar, 10 μm. (B) Quantification of the co-localisation coefficient between SRα and NS4A at different time points post-infection. The co-localisation coefficient was calculated using Pearson’s correlation and represents the fraction of SRα that colocalises with NS4A. Data represent mean ± SEM (n = 30 cells per time point), with * p < 0.05 and ** p < 0.01 by one-way ANOVA with Tukey’s post hoc test.

Fig 6. Overexpression of SRα-ΔSRx inhibits ZIKV polyprotein synthesis and viral replication.

(A) Schematic of the proposed model for ZIKV non-structural proteins bypassing SRα/SRβ by directly interacting with SRP54 to facilitate ER targeting of the viral polyprotein. Overexpression of the SRP54-interacting domain of SRα (SRα-ΔSRx) is hypothesised to competitively inhibit the interaction between viral proteins and SRP54. Schematic was generated using Biorender. (B) Immunoblot analysis of SRα-ΔSR_X and Lyn kinase expression in HeLa cells transfected with empty vector, SRα, SRα-ΔSR_X and Lyn kinase plasmids. GAPDH served as a loading control. (C-D) Inhibition of ectopic prME-FLAG expression by SRα-ΔSR_X. HeLa cells transfected with empty vector, SRα-ΔSR_X or Lyn-kinase were transfected with a plasmid encoding ZIKV prME protein. (C) Immunoblot analysis and (D) quantification of relative prME levels, normalised to empty vector control. Data represent mean ± SD (n = 3). (E-F) Inhibition of ZIKV protein synthesis by SRα-ΔSR_X. HeLa cells transfected with empty vector, SRα, SRα-ΔSR_X or Lyn kinase were infected with ZIKV (MOI = 1). (E) Immunoblot analysis of ZIKV E protein levels at the indicated timepoints. (F) Quantification of relative E protein levels at 72 hpi, normalised to empty vector control. Data represent mean ± SD (n = 3). (G) ZIKV growth kinetics in HeLa cells transfected with empty vector, SRα-ΔSR_X or Lyn kinase. Supernatants were collected at the indicated timepoints, and viral RNA was quantified by RT-qPCR. Data represent mean ± SD (n = 3). Statistical significance was determined by two-tailed unpaired Student’s t-test, with * p < 0.05, ** p < 0.01.

Fig 7. ZIKV infection remodels the ER translatome in a SR-independent manner.

(A, B) Mock- or ZIKV-infected cells depleted of SRα, SRβ, or SRP54 were fractionated into cytosolic and membrane fractions, and the distribution of marker proteins was analysed by immunoblotting. (A) Immunoblots of markers in lysates (B) Immunoblots of cytosolic and membrane fractions from mock- or ZIKV-infected cells. GAPDH served as a cytosolic marker, while KDEL receptor (KDELR), calnexin (CANX), and calreticulin (CALR) were used as ER markers. RPL17 was probed to assess ribosome association with ER membranes. (C) Schematic of the polysome profiling and mass spectrometry experiment. Polysomes were isolated from mock- or ZIKV-infected HeLa cells depleted of SRα, SRβ, or SRP54 at different timepoints post-infection. Polysome-associated proteins were identified and quantified by LC-MS/MS. Schematic was generated using Biorender. (D-E) Heatmaps showing the relative abundance of polysome-associated proteins (E) ZIKV infection led to an enrichment of proteins involved in ER membrane expansion, lipid metabolism, and secretory processes, while proteins associated with RNA binding and stress granules were depleted. (F-G) Validation of polysome profiling results by [³⁵S]-cysteine/methionine labelling. Mock- or ZIKV-infected HeLa cells depleted of SRα or SRP54 were pulse-labelled for the indicated times, and newly synthesized proteins were analysed by autoradiography. (F) ER-associated proteins ACSL3 (lipid metabolism) and (G) SOAT1 (cholesterol esterification) showing increased synthesis in ZIKV-infected cells. Images are representative of at least 2 independent experiments.

Fig 8. Proposed model for ZIKV-mediated remodelling of ER targeting and altered translatome.

During infection, ZIKV non-structural proteins (NS2B-NS3, NS4A, NS4B) interact with SRP54 and the translocon to bypass the requirement for SRα/SRβ and facilitate ER targeting of the viral polyprotein. This leads to selective enrichment of host factors involved in ER membrane expansion, lipid metabolism, and secretory processes at the translating ribosomes, while RNA-binding proteins and stress granule components are excluded. This translational reprogramming supports viral replication and suppresses antiviral responses. The viral non-structural proteins play a key role in this process by hijacking the SRP-dependent targeting pathway and recruiting host factors necessary for viral replication to the ER. Schematic was generated using Biorender.