JEV binding, entry and infection in neuronal cells is clathrin-independent

Kinetic analysis of the virus replication cycle in SH-SY5Y cells showed that virus uptake in cells peaks between 30 min- 1hour (h) (S1a Fig). Following release from endosomes through a low pH mediated uncoating of the envelope, viral RNA initiates replication on ER-derived membranes and a rapid increase in viral RNA levels is observed in the infected cells (S1a Fig). Virus egress in the extracellular milieu can be detected from 12 h post-infection (hpi) (S1b Fig).

JEV endocytosis was analyzed through quantitative estimation of viral RNA copy number at early times of infection using two different assay conditions (Fig 1a). One was a widely used ice-synchronized technique that involves virus binding to cells on ice for 1h, followed by washing and further incubation in complete serum medium at 37^OC for 1h. However, since ice incubation has been reported to reduce membrane fluidity and can potentially inhibit fast endocytic processes [23], we also employed an alternate assay where virus binding and entry was performed at 37^OC (Fig 1a). MOI 5 (ice-binding) and MOI 10 (physiological conditions) we determined to be ideal for quantitative estimation of viral RNA to analyze virus binding and endocytosis. At these MOIs nearly every cell receives at least one infectious virus particle and enables an accurate measurement of virus endocytosis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. JEV binding, entry and replication in neuronal cells is clathrin-independent.

(a) Schematic representation of virus binding/entry assay. Created in BioRender. Kalia, M. (2025) https://BioRender.com/22xa6tp. (b-h) SH-SY5Y cells were transfected with siNT/si CLTB for 72 h. (b) Western blot shows CLTB protein levels and bar graph shows the relative level of CLC mRNA post-siRNA transfection. (c-f) The siRNA-transfected cells were infected with JEV (5/10 MOI), and virus binding/entry assays were performed as described. Absolute virus envelope copies were detected with qRT-PCR. (g-h) The siRNA-transfected cells were infected with JEV (1 MOI) for 24 h. Intracellular viral RNA levels were determined by qRT-PCR (g), and the extracellular virus particles were detected with plaque assays (h). (i-l) Cells seeded in nu-serum containing media were pre-treated with either DMSO or 80µM dynasore for 1 h 37^Oc and were infected with purified JEV as described previously. (i-j) Bound and internalized virus load was estimated by qRT-PCR, when incubated on ice. (k-l) Viral RNA levels and titres at 24 hpi. All values are represented from two or more independent experiments with mean ± S.D. Statistical analysis was determined with Mann Whitney test with a 95% confidence level, NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g001

To examine the role of CME in JEV entry in the human neuroblastoma cell line SH-SY5Y, an RNA interference-based knockdown approach was utilized. The effect of Clathrin light chain (CLC) knockdown was first tested on the internalization CME cargo: Transferrin (Tf) and Epidermal growth factor (EGF). While Tf endocytosis is strictly through CME, EGF follows different endocytic routes, with a lower dose resulting in uptake through CME, whereas, at a higher dose, a major fraction of EGF gets internalized through CIE [27,28]. Treatment of SH-SY5Y cells with CLC siRNA resulted in a significant decrease in the protein and transcript levels compared to siNT, indicating efficient knockdown (Fig 1b). These cells were given a pulse of Tf (20 ng/ml) and different concentrations of EGF (10, 20, 50, & 180 ng/ml) for 5 min at 37^Oc. As expected under conditions of CME inhibition, in comparison to short interfering non-targeting (siNT), the si Clathrin light chain B (CLTB) transfected cells showed a significant decrease in the uptake of Tf (S2a and S2b Fig). The knockdown of CLTB also decreased the internalization of EGF at a concentration range between 10–50 ng/ml, suggesting a role of CME in EGF uptake at these concentrations. However, a higher dose of 180 ng/ml EGF showed no perturbation in uptake, confirming its endocytosis via CIE (S2a and S2b Fig).

After the cargo uptake inhibition was established upon CLTB knockdown conditions, we next checked its effect on JEV endocytosis. In one set of experiments, the virus (5 MOI) was allowed to bind on the cell surface at 4^Oc (ice) for 1 h, which synchronizes virus binding on cells but prevents internalization. After 1 h of binding, virus entry was initiated by adding complete media and shifting the temperature to 37^Oc for 1 h. Cells are then harvested following trypsin treatment, which removes the extracellular bound virus particles to enable quantification of internalized viral load only (Fig 1a, upper panel). Virus binding (ice) remained unaltered, indicating that virus attachment is unaffected by CLC depletion (Fig 1c). Virus endocytosis also remained unaltered (Fig 1d).

In a parallel set of experiments, virus (10 MOI) binding was performed at 37^Oc for 1 h, followed by further incubation in complete medium for 1 h to permit virus entry (Fig 1a, lower panel). These physiological or FEME conditions were also used since ice incubation is known to change membrane fluidity, which can inhibit certain CIE pathways [24]. The virus binding at physiological/FEME conditions showed no significant change (Fig 1e), and further virus internalization under these conditions was also unperturbed (Fig 1f). The depletion of CLTB also had no effect on virus replication as no change in the viral RNA levels was observed at 24 h (Fig 1g), along with unaltered virus titers (Fig 1h). These experiments indicate that JEV internalization in neuronal cells can follow a clathrin-independent pathway.

To test the requirement of dynamin for the endocytosis of JEV in SH-SY5Y cells, Dynasore, a cell-permeable and reversible GTPase inhibitor that inhibits the activity of dynamin was used [29]. Cells were treated with 80 µM of dynasore and given a pulse of Tf (20 ng/ml) and EGF (50 ng/ml) for 5 min at 37^Oc. As expected, dynasore inhibited both EGF and Tf uptake (S2c Fig). To confirm the role of dynamin in JEV endocytosis, virus entry assays was performed in the presence of the dynasore. While there was no difference in virus binding (Fig 1i), nearly ~90% decrease in virus internalization (Fig 1j), and a very significant inhibition of virus replication (~80% reduction in viral RNA levels, and ~99% reduction in virus titers at 24 hpi) were observed compared to DMSO treated cells (Fig 1k-1l). We also checked for the localization of dynamin with capsid labelled virions at early times of infection (S3 Fig). At 10 minutes post-infection (mpi) ~75% of the capsid puncta overlapped with dynamin (S3a and S3b Fig). In agreement with earlier published work from our lab [16], these results demonstrate that JEV entry is dynamin-dependent.

Endophilin is an essential host-factor for JEV entry and replication

To characterize other molecular players for JEV endocytosis, we tested the role of endophilin. Endophilin A exists in three isoforms: A1/A2/A3. Firstly, we checked the basal protein level expression of all three endophilin A proteins in three different cell types: (1) human neuronal cell line SH-SY5Y, (2) HeLa cell line and (3) mouse brain tissue. As shown earlier [24,30], endophilin A2 expression was ubiquitously observed in all three samples (Fig 2a, left panel). In contrast, endophilin A1 expression was only observed in brain tissue, while there was lower expression of endophilin A3 in SH-SY5Y, compared to HeLa cells and mouse brain (Fig 2a, left panel). We then utilized a genetic knockdown-based approach to establish the effect of endophilin A depletion on the uptake of Tf and EGF. We performed a triple knockdown (TKD) of all three endophilin isoforms, as in the absence of one, the other endophilin A isoforms have been shown to compensate [23]. Western blots confirmed efficient knockdown by siTKD transfection (Fig 2a, right panel).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Endophilin is an essential host-factor for JEV entry and replication.

(a) Protein lysates from SH-SY5Y, HeLa and homogenized mouse brain tissue were immunoblotted with endophilin A1/A2/A3 and GAPDH (internal control), left panel. Right panel shows levels of endophilin A/A2/A3 in untransfected, NT siRNA and siTKD post-transfected SH-SY5Y cells. (b-h) SH-SY5Y cells were transfected with all three endophilin A1/A2/A3 siRNA (siTKD) for 72 h. (b) Bar graph shows the relative endophilin A2 levels in siNT/siTKD transfected cells. (c-d) Endophilin siTKD cells were allowed to bind with 10 MOI of virus on ice, and virus binding/internalization assays were performed as described previously. Absolute virus envelope copies were detected with qRT-PCR. (e-f) The siNT and siTKD transfected cells were infected with 10 MOI of the virus at 37^Oc for 1 h, and virus binding/internalization assays were performed. Absolute virus envelope copies were detected with qRT-PCR. (g-h) Endophilin siTKD cells were infected with 1 MOI of the virus at 37°c for 1h and were washed and harvested at 24 hpi, viral RNA was determined by qRT-PCR (g), and the extracellular virus particles were detected with plaque assays (h). (i) Representative room temperature ultramicrotomy cell sections imaged under TEM. Larger cellular sections with zoomed-in insets (red boxes) show spherical JEV particles of ~50 nm (indicated by red arrows)) near the cell membrane (marked in green); scale: 500 nm. (j-m) SH-SY5Y cells (j-k), or primary cortical neurons (l-m), were pre-treated with CHIR-99021 for 1 h at 37^Oc, and then were infected with 10 MOI of JEV at 37°c for 1h, and virus binding/internalization assays were performed as described. Absolute virus envelope copies were detected with qRT-PCR. Data shown are from two or more independent experiments represented as mean ± S.D. Statistical analysis was determined with Mann Whitney test with 95% confidence level, NEJM: 0.12 (ns), 0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g002

SH-SY5Y cells transfected with siNT or siTKD were subjected to a pulse of EGF and Tf for 5 min at 37^Oc. Depletion of endophilin led to no substantial reduction in the uptake of Tf (S4a and S4b Fig). An earlier study [23] confirmed that endophilin A is not required for Tf uptake through CME. A dose-dependent pulse of the siTKD transfected SH-SY5Y cells with EGF (50 & 180 ng/ml) for 5 min at 37^Oc showed a reduction in EGF uptake at a higher concentration of 180 ng/ml but had no effect at a concentration of 50 ng/ml. These results indicated that while endophilin TKD did not affect CME, it inhibited the CIE of EGF at higher doses (S4a and S4b Fig). As reported earlier [23], siTKD increased the cell surface levels of EGFR (S4c Fig). The CIE of EGF has also been shown in several independent studies in multiple cell lines such as BSC1, RPE1 and HeLa [23,31–33].

The impact of genetic inactivation of endophilin A (siTKD) was subsequently checked on virus endocytosis with entry experiments performed at two different temperatures, as stated previously (Fig 1a). The relative mRNA expression of endophilin A2 was checked in siTKD transfected cells in every experiment to confirm the knockdown at 72 h (Fig 2b). Depletion of endophilin did not affect virus binding on ice (Fig 2c). Still, it significantly reduced virus entry (~69%), validating endophilin’s crucial role in virus internalization (Fig 2d). We also performed the binding/internalization assays at a physiological temperature of 37^Oc. Again, in the TKD cells, the virus binding at 37^Oc remained unaltered (Fig 2e), while virus internalization showed significant (~75–80%) inhibition across different MOIs (Figs 2f, S5a, and S5b). These results were also validated to check whether this observed decrease in virus entry would manifest in lower virus replication. Indeed, we observed a substantial decline (~44%) in viral RNA levels (Fig 2g) and (~28%) virus titers (Fig 2h) at 24 hpi. These results highlight the crucial role of endophilin in JEV entry and infection.

We next checked for the colocalization of endophilin A2 with virus particles stained with specific capsid antibody, that will be indicative of internalized virions at early times of infection (before initiation of virus replication). While no signal was detected in mock-infected cells, several distinct puncta were visualized in cells where virus was added (S5c Fig). A quantitative estimation of virus particles showed that ~70 puncta were detected per cell at 10 mpi, a time-point where virus endocytosis is initiated (S5c and S5d Fig). We observed colocalization of endophilin A2 with JEV particles (capsid structures) at 10 mpi (S5c Fig). Spot-spot colocalization analysis of the confocal micrographs showed ~47% of JEV was colocalized with endophilin A2 positive structures (S5d Fig). Time course analysis showed that while the number of capsid puncta gradually declined by 60 min, the colocalization with endophilin remained high (~60%). These data suggest that the earliest endocytic structures containing internalized JEV particles are also positive for endophilin. We also performed TEM imaging of sectioned SH-SY5Y cells to examine the early entry events of virus infection. Spherical virus particles of ~ 50 nm were observed on the surface of SH-SY5Y cells, showing interactions in large inundations of the cell membrane that appear distinct from the classical clathrin-coated vesicles (Fig 2i).

JEV entry is enhanced by GSK3β inhibition

To further investigate if there are any parallels between endophilin-mediated CIE of JEV and the FEME pathway, we treated SH-SY5Y cells with a small molecule compound CHIR-99021, a well-reported kinase inhibitor of GSK3β, which suppresses FEME under normal physiological conditions [34]. Treatment of cells with CHIR-99021 for 1 h at 37^Oc enhanced the uptake of EGF, further supporting the observation that inhibition of GSK3β leads to upregulation of FEME (S6a Fig). The effect of this inhibitor was then checked on virus binding/internalization at a physiological temperature of 37^Oc (FEME conditions). Treating cells with CHIR-99021 led to a significant increase in virus binding and internalization (Fig 2j and 2k). These results suggest that relieving the autoinhibition of FEME by treating cells with inhibitors of kinases that regulate the key steps of the pathway substantially increased JEV endocytosis. These findings were further corroborated in the primary culture of cortical neurons harvested from mice brain, which showed a similar enhancement of JEV entry on the CHIR-99021 treatment (Fig 2l and 2m).

Next, we tested whether JEV endocytosis increases the fluorescence intensity of endophilin-positive structures that are a hallmark of FEME [23,24]. Since serum stimulation has been shown to increase FEME [34], we used it as a positive control along with EGF stimulation. Primary cortical neurons were stimulated with 20% serum for 15 min at 37^Oc. Compared to the control, serum stimulation increased endophilin-positive structures, which can be observed as discrete punctate structures when immunostained with endophilin A2. An EGF pulse (180 ng/ml) for 5 min at 37^Oc also increased endophilin-positive structures. JEV infection also enhanced the endophilin-positive punctate structures at 15 mpi, which further increased at 30 mpi (S6b and S6c Fig). This suggests that like serum stimulation, high EGF doses, and JEV endocytosis can increase endophilin-positive structures, likely indicative of an upregulation of the FEME pathway. Collectively, our data on the genetic depletion of endophilin and molecular upregulation of FEME suggest a prominent role of this pathway in JEV endocytosis.

Endophilin domain mutants block JEV internalization and replication

To further characterize the role of endophilin A2 for JEV internalization, we generated stable clones expressing EGFP tagged endophilin A2 full length (FL), or the domain truncated mutants ΔSH₃, ΔH₀, and ΔBAR in SH-SY5Y cells (Fig 3a). The over-expression of the endophilin domain truncated mutants will impose a dominant effect and override the endogenous protein function. The over-expression of the constructs was confirmed through immunoblotting with anti-GFP antibody. Western blot showed bands at the expected molecular weights for FL and truncated domains of endophilin A2 (Fig 3b). Confocal microscopy images show GFP expressing stable cells with EGFPC1, FL, ΔSH₃, ΔH₀, and ΔBAR overexpression (Fig 3c).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Endophilin domain mutants block JEV internalization and replication.

(a) Schematic representation of the GFP endophilin A2 with full length (FL), ΔSH₃, ΔH₀, and ΔBAR domains deletion mutants. Created in BioRender. Kalia, M. (2025) https://BioRender.com/g4etbij (b) Lysates of SH-SY5Y cells stably expressing EGFPC1, FL, ΔSH₃, ΔH₀, and ΔBAR domains were immunoblotted with anti-GFP antibody. (c) Confocal micrographs of cells stably expressing the various domains. Scale: 10 µm (d) Endophilin domain mutant cell lines were stimulated with or without 20% FBS for 1 h at 37°c, infected with 10 MOI for 1 h at 37^Oc, and harvested post 1 h with trypsin. The histogram represents the viral copy number of internalized virus particles. (e-f) Cells were infected with 1 MOI of JEV for 1 h and were harvested post-24 h. Viral RNA was determined by qRT-PCR (f), and the extracellular virus particles were detected with plaque assays (g). All values are represented as mean ± S.D from at least two independent experiments. Statistical analysis was determined using ordinary one-way ANOVA with Dunnett’s multiple comparison test with a 95% confidence level. Statistical significance: NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g003

These clones were tested for cargo uptake assays to demonstrate their effect on EGF uptake at low and high concentrations. An increase in the uptake of fluorescently labelled EGF was observed in the cells expressing FL endoA2 as compared to EGFPC1 at both concentrations, indicating upregulation of the pathway upon endophilin A2 over-expression. Cells with the domain truncated mutants of ΔSH₃, ΔH₀, and ΔBAR displayed a decrease in the uptake of EGF (S7a Fig first & second panel, S7b and S7c). These mutants specifically perturb endophilin mediated endocytosis and not CME, as the uptake of Tf remains unchanged (S7a Fig third panel, and S7d). The impact of these stably overexpressing cells was also checked on the fluid-phase uptake of FITC-dextran, which was unaltered (S7a Fig fourth panel, and S7e). These data demonstrate that the endophilin A2 domain deleted mutants exert a dominant negative effect in blocking EGF uptake in SH-SY5Y cells.

We next checked for the role of each of these domains on JEV internalization. The virus entry assays were next performed under FEME conditions: 37^Oc for 1 h to allow virus binding at 10 MOI, with subsequent internalization at 1 h at 37^Oc. Here, an increase in virus entry was observed with FL overexpressing cells compared to the empty vector. A subsequent decline in the virus entry was observed with all three truncated mutants, with a maximum decrease (~90%) with ΔBAR-expressing cells. This result highlights the crucial role of the three main domains of endophilin A2 proteins in JEV entry (Fig 3d, left panel). To further strengthen our results, we checked virus entry in these mutant cell lines when stimulated with additional serum (20% FBS) to enhance FEME. Entry experiments showed an increase in virus copy numbers with FL compared to the empty vector, while the ΔH₀ and ΔBAR mutants showed a decrease of ~ 52% and 88% respectively (Fig 3d, right panel). We also checked the effect of these endophilin mutant cell lines on virus replication. As previously observed with our entry experiments, we observed an increased viral RNA level in cells over-expressing FL compared to empty vector. The ΔSH₃ and ΔBAR-expressing cells showed a reduction in virus replication as compared with the FL, and maximum inhibition was observed with ΔBAR deletion (~50%) (Fig 3e). These results were also supported by reduced virus titers in ΔSH₃ (~38%) and ΔBAR (~43%) domain-expressing cells (Fig 3f). Collectively our data shows that over-expression of endophilin A2 enhances JEV uptake and also establishes a crucial role of the cargo receptor binding domain and the membrane curvature-inducing domains of endophilin A2 for JEV internalization.

Pharmacological inhibition of actin cytoskeleton perturbs JEV entry and replication

Actin polymerization is crucial for multiple CIE pathways. The molecular mechanism of FEME is known to be an actin-driven pathway, vital for carrier formation and driving the vesicles into the cytoplasm. We checked for the role of the actin machinery in JEV entry through pharmacological inhibition of the pathway using various known inhibitors, Latrunculin A (Lat A) which inhibits actin polymerization [35], Jasplakinolide (Jas) which binds to F-actin and blocks the actin depolymerization [36], and Cytochalasin D (CytoD) which induces the depolymerization of actin filaments through binding to the G-actin leading to dimer formation and nucleation of the newly formed filaments [37,38]. The other small molecule inhibitors used were CK-548, which binds to the Arp2/3 complex and inhibits its ability to polymerize actin, and Wiskostatin, an N-WASP-mediated actin polymerization inhibitor [39]. Cell viability assays were performed to establish non-toxic drug concentrations: LatA (1uM), Jas (1uM), CytoD (5uM), CK-548 (50uM) and Wiskostatin (10uM) (S8a Fig). Cells pre-treated with the drugs for 1 h inhibited virus binding and internalization at 37^Oc (Fig 4a and 4b). This was also observed through confocal imaging, wherein treatment of cells with the drugs led to a decrease in virus endocytosis, which can be observed with a reduced number of envelope positive immunostained structures (Fig 4c and 4d). Treatment of cells with CytoD, CK-548 and Wiskostatin inhibited virus replication as observed by a decrease in viral RNA levels (S8b Fig) and virus titers (S8c Fig), thus confirming the critical role of a functional actin machinery for JEV infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Pharmacological inhibition of actin cytoskeleton perturbs JEV entry and replication.

(a-b) SH-SY5Y cells were pre-treated with DMSO control, LatA (1uM), Jas (1uM), CytoD (5uM), CK-548 (50uM) and Wiskostatin (10uM) for 1 h at 37^Oc, followed by 10 MOI JEV at 37^Oc for virus binding. Cells were harvested with trypsin treatment to remove extracellular virus particles. Absolute virus envelope copies were detected with qRT-PCR. (c) SH-SY5Y cells were pre-treated with DMSO control/ drug for 1 h at 37^Oc subsequently followed with 100 MOI virus binding at 4^Oc. Cells were then shifted to 37^Oc for virus internalization for 15 min, fixed, immunostained with Alexa Fluor 546-phalloidin (cyan), and viral envelope particles (magenta) and imaged with 63x objective. The representative confocal micrograph shows the phalloidin-stained F-actin and JEV envelope puncta in the presence of actin inhibitors, scale: 10 µm. (d) The histogram quantifies the number of envelope positive puncta per cell analyzed through Image J. All values are represented as mean ± S.D./S.E.M; statistical analysis was determined with Ordinary one-way ANOVA/Kruskal-Wallis test with Dunn’s multiple comparisons test. Statistical significance: NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g004

Through confocal microscopy, we monitored changes in the actin cytoskeleton upon JEV entry at different time points, starting from 5 min to 60 min pi. Mock-infected cells showed cortical actin meshwork fibers with different well-aligned stress fibers, including dorsal and peripheral fibers (S9a Fig, uninfected panel). In the JEV-infected cells, by 5 min pi, the F-actin redistribution changed, and dense condensed F-actin filaments could be observed with long protrusions in the filopodia (S9a Fig, 5 min panel). By 10 min post-virus entry, increased and dense lamellipodia structures were observed (S9a Fig, 10 min panel). Long filopodia extensions were observed at 15 min and 30 min post- virus entry (S9a Fig, 15- and 30 min panel). The number of protrusions decreased by 60 min post-virus entry (S9a Fig, 60 min panel). We also checked actin rearrangements in primary cortical neurons infected with JEV, which showed multiple long protrusions with denser lamellipodium (S9b Fig).

EGFR signaling is activated and essential for JEV entry

Several viruses such as ZIKV, PEDV, and IAV, can activate EGFR signaling cascade [40–42]. JEV has also been shown to activate EGFR-PI3K signaling during entry [19]. To test the same in our experimental setup, SH-SY5Y cells were infected with purified JEV, and the levels of phosphorylated EGFR were analyzed through western blotting at 0, 5, 10, 15, 30, and 60 min of virus internalization (Fig 5a). Compared to time-matched mock infected cells, autophosphorylated EGFR (Tyr1068) was detected in JEV-infected cells peaking at 15 min pi. EGF treatment (100 ng/ml) was also used as a positive control (Fig 5a). These observations were also confirmed with confocal microscopy, wherein EGF treatment or JEV entry resulted in the activation of p-EGFR compared to mock-infected cells (Fig 5b).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. EGFR signaling is activated and essential for JEV entry.

(a) SH-SY5Y cells were first allowed to bind with purified JEV at 4^Oc for 1h and were subsequently shifted to 37^Oc for 0, 5, 10, 15-, 30-, 60-, and 120-min. Cell lysates were analyzed through western blotting using p-EGFR, EGFR and GAPDH (loading control) antibodies. (b) Representative confocal images of SH-SY5Y cells: mock, treated with EGF (100 ng/ml), or JEV (100 MOI, 4^Oc, 1h for attachment, and then shifted to 37^Oc for 15 min), immunostained with p-EGFR antibody. (c-f) SH-SY5Y cells, (g-h) primary cortical neurons, were serum starved overnight and treated with genistein (400 ug/ml) or gefitinib (400 ug/ml) for 5 h before infection. (c) Virus attachment assay (10 MOI, 4°c, 1 h). (d) Virus entry assay (10 MOI, 4°c, 1 h binding and then shifted to 37^Oc for 1h). (e-f) Virus replication assay (1 MOI, 37°c, 24 hpi). Viral RNA was determined by qRT-PCR (c-e), and the virus titers were quantified with plaque assays (f). (g-h) JEV RNA levels and titers in primary cortical neurons (1 MOI, 37°c, 24 hpi). (i-j) SH-SY5Y cells (i), primary cortical neurons (j), were pre-treated with erlotinib for 4 h at 37°c and then infected with JEV (1 MOI, 37°c, 24 hpi). Viral RNA was determined by qRT-PCR. Data shown are from two or more independent experiments represented as mean ± S.D. Statistical analysis was determined by Ordinary one-way ANOVA with Dunnett’s multiple comparison tests or the Mann-Whitney test with 95% confidence level. Statistical significance: NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g005

To examine the role of EGFR signaling in the context of JEV endocytosis, we utilized the broad-spectrum RTK inhibitor genistein, and the EGFR-specific inhibitors: gefitinib and erlotinib. To ensure that the drugs do not exert any cytotoxic effect, a cell viability assay was performed (S10a and S10b Fig). Serum-starved SH-SY5Y cells were pre-treated with 400 ug/ml of genistein or gefitinib for 5 h at 37^Oc. This did not alter virus binding or attachment (Fig 5c) but significantly reduced virus entry (Figs 5d and S10c), replication (Fig 5d), and production of infectious virions (Fig 5f). We further confirmed our results in primary cortical neurons and observed a significant reduction in JEV RNA levels and titers with genistein and gefitinib treatment (Fig 5g and 5h).

These observations were validated with erlotinib, another known small molecule inhibitor of EGFR tyrosine kinase activity. Erlotinib showed antiviral activity in SH-SY5Y cells with ~50% reduction in viral RNA levels at 24 hpi (Fig 5i). A similar inhibition in virus replication was also observed in primary cortical neurons (Fig 5j). Collectively, pharmacological inhibition of EGFR with both broad-spectrum and specific inhibitors was detrimental for virus entry and infection.

JEV entry is impaired in EGFR depleted cells

We used siRNA targeting of the human EGFR gene to specifically knockdown its expression. SH-SY5Y cells were transfected with siNT or siEGFR, and a significant reduction in the EGFR mRNA levels was observed at 72 h of transfection compared to the control siRNA transfected cells (Fig 6a). As expected, the silencing of EGFR led to a decrease in the cargo uptake of fluorescently labelled EGF (S11a and S11b Fig). EGFR silencing resulted in a significant decrease in the virus binding (Fig 6b), and entry under both ice synchronization (Fig 6c), and FEME conditions (Fig 6d), suggesting the potential role of EGFR in virus attachment and entry. EGFR silencing also reduced viral RNA levels by ~40% at 24 hpi, associated with decreased virus titers (Fig 6e and 6f). These results indicate the role of EGFR as an essential entry co-factor for JEV internalization in neuronal cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Silencing of EGFR decreases JEV entry.

SH-SY5Y cells were transfected with transfected with siNT/si EGFR for 72 h. (a) Western blot shows protein level and Histogram shows the relative level of EGFR knockdown at the mRNA level post-siRNA transfection. (b) JEV attachment on cells (10 MOI, 4°c, 1 h). (c) JEV internalization (10 MOI, 4°c, 1h binding and internalization at 37°c for 1 h). (d) JEV internalization under FEME conditions (10 MOI, 37°c, 1 h). (e-f) JEV replication (1 MOI, 24 hpi). JEV RNA levels were determined by qRT-PCR (b-e), and titers plaque assays (f). All values are represented from two or more independent experiments with mean ± S.D. Statistical analysis was determined with Mann Whitney test with a 95% confidence level, NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g006

Ligand binding domain of EGFR plays a critical role in JEV entry

Since EGFR gene silencing inhibited virus attachment, we next explored the role of the EGFR ligand binding domain in virus entry. An antibody inhibition-based approach was used to specifically block the ligand binding domain of the receptor. Incubation of cells with an EGFR epitope binding monoclonal antibody significantly decreased virus attachment (Fig 7a) and entry (Fig 7b), as compared to isotype control. This further manifested as reduced virus replication and egress at 24 hpi (Fig 7c and 7d). Thus, the EGFR ligand binding domain appeared to be critical for virus attachment on cells and further infection.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Ligand binding domain of EGFR is essential for JEV entry.

(a-d) SH-SY5Y cells were pre-incubated on ice with an antibody mapping to a cell surface epitope of EGFR for 1 h. (a) JEV attachment (10 MOI on ice, 30 min). (b) JEV entry (10 MOI virus at 4°c for 30 min and then shifted to 37°c for 1 h). (c-d) JEV replication (1 MOI, 24 hpi). Viral RNA levels were determined by qRT-PCR (a-c) and titres by plaque assays (d). (e-f) SH-SY5Y cells were left untreated or incubated with IgG isotype control or cetuximab at 100 ug/ml concentration for 3 h at 37^Oc. Cells were then pulsed with Alexa fluor 555 EGF (100 ng/ml) for 5 mins at 37°c. (e) Representative images, scale: 20 µm. (f) Histogram quantifies the total fluorescent units of EGF uptake under untreated, IgG isotype control and cetuximab-treated conditions. Analysis was performed using image J software with ~100 cells/coverslip, represented as mean ± S.E.M. (g-j) Cells were pre-treated with either IgG isotype control or with cetuximab (50 and 100 ng/ml) at 37°c for 3 h. (g-i) Virus attachment (10 MOI, ice, 1h) in SH-SY5Y cells (g), (h) Virus entry (10 MOI, ice 1 h, shifted to 37°c for 1 h). (i) Virus replication (1 MOI, 24 hpi), (j) Virus attachment (10 MOI, ice, 1h) in primary cortical neurons. Viral RNA levels were quantified through qRT-PCR. (k) SH-SY5Y cells were serum stimulated with 20% FBS for 15 min at 37°c, and were allowed to bind with 10 MOI virus along with different doses of EGF at 4°c for 1 h. The histogram shows the bound viral copies detected through qRT-PCR. (l) Cells were serum-stimulated with 20% FBS, followed by EGF stimulation at 37°c for 15 min, and then 1 MOI virus at 4°c for 1 h. (m) Serum and EGF-stimulated cells were infected with 10 MOI viruses and different EGF concentrations for 1 h on ice and were harvested post 1 h with trypsin treatment. (n) Intracellular RNA levels were detected at 24 hpi, and (o) the extracellular viral load was detected through plaque assay. All values are represented as mean ± S.D from at least two independent experiments. Statistical analysis was determined by unpaired students’ t-test or ordinary one-way ANOVA, with Dunnett’s multiple comparison test, with a 95% confidence level. Statistical significance: NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***).

https://doi.org/10.1371/journal.ppat.1013790.g007

This was confirmed by treatment of cells with Cetuximab (Erbitux), a recombinant human chimeric monoclonal IgG1 antibody which binds to the EGF binding domain [43]. Compared to the isotype control, cetuximab significantly decreased EGF uptake (Fig 7e and 7f). Consistent with the previous results, we also observed a reduction in JEV attachment, entry and replication in cetuximab-treated cells as compared to isotype control (Fig 7g-7i). This was also validated in primary cortical neurons, which showed similar inhibition of virus binding (Fig 7j). These results indicated the possibility that JEV could be binding to EGFR at a site that overlaps with the EGF binding.

To further elucidate this, we stimulated cells with EGF and observed its effect on the early events of the virus life cycle. The binding of EGF with its receptor EGFR leads to receptor dimerization and activation of the downstream signaling cascade. However, prolonged stimulation with EGF results in ligand-induced receptor degradation [44]. Surface EGFR levels were significantly reduced in cells upon dose dependent EGF pulse (S12a and S12b Fig). The effect of adding different doses of EGF was first checked on virus attachment. A stimulation step with 20% additional serum was also performed to activate FEME. Here, we observed a decrease in virus binding starting from 50 ng/ml of EGF (Fig 7k). These results suggested that incubation of both EGF and JEV could result in a competitive inhibition for binding to EGFR. Next, we treated the serum-stimulated cells with the above-mentioned doses of EGF before infection. This step was performed to activate the receptor internalization before infection. A decrease in virus binding was observed starting from 10 ng/ml of EGF, which decreased further with an increase in the EGF dose (Fig 7l). This might result from a reduced level of surface EGFR, which is essential for virus attachment. The EGF treatment also correlated with a reduction in the virus entry at 1 hpi starting from 20 ng/ml EGF dose, which decreased further with an increase in EGF dose (Fig 7m). Reduced virus replication and titers at 24 hpi was also observed (Fig 7n and 7o). Altogether, our results highlight two crucial observations: first, the importance of the accessibility of EGFR ligand binding domain during the initial virus attachment, and second, the availability of the surface receptor for virus entry.

EGFR colocalizes with virions and interacts with ED3 domain of envelope protein

Our data suggests that JEV could potentially be associating with EGFR for attachment and entry into cells. Next, we checked for interaction of EGFR proteins with JEV particles during entry. SH-SY5Y cells were allowed to bind with JEV (MOI 100) on ice, followed by internalization for 0, 10, 30, and 60 min. The virus particles can be visualized with JEV-E antibody as punctate structures (magenta) (Fig 8a). Maximum number of envelope-positive structures were observed at 0- and 5-min post entry, which then reduced at 30- and 60-min pi. The virions were observed to be colocalized with EGFR as soon as 0 min post-entry, which corresponds to the initial binding of the virus on the cell surface. The colocalization of EGFR with envelope structures then gradually reduced staring from 10-min post-entry onwards (Fig 8b). A similar overlap of JEV capsid structures with the EGFR protein was observed in SIM imaging (S13a Fig). We also checked the colocalization of the virus particles with EGFR at early time-points of virus entry in primary cortical neurons, and a similar pattern of colocalization at 0 min was observed (S13b Fig). EGF uptake was enhanced in cells that had been infected with JEV first, likely due to activation of signaling that preconditions cells for increased endocytosis of EGF. Interestingly, EGF uptake in virus-infected cells showed clathrin dependence at higher concentrations, suggesting that JEV can impact the endocytosis pathway of EGFR (S14a and S14b Fig)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. EGFR colocalizes with virions and interacts with ED3 domain of envelope protein.

(a-b) SH-SY5Y cells were incubated with 100 MOI of virus for 1 h on ice and were subsequently fixed post at 0, 10, 30, and 60 mins of virus internalization. (a) Cells were immunostained with JEV envelope (magenta) and EGFR (green) antibodies. Confocal images for 0 mpi is shown for reference, other time points image not included. Insets show the zoomed image of envelope-positive structures colocalized with EGFR; scale: 10 µm, inset: 1 µm. Images were acquired in Elyra PS1 (Carl Zeiss Super-resolution microscope) using 63 x objective. (b) The histogram quantifies total envelope-positive structures and the number of envelope-positive structures colocalized with EGFR per cell at different time points. Image analysis was performed using Imaris. All values are represented as mean ± S.E.M; statistical analysis was determined with Kruskal-Wallis test with Dunn’s multiple comparisons test. Statistical significance: NEJM: 0.12 (ns),0.033 (*), 0.002(**), < 0.001(***). (c-d) Bio-Layer Interferometry analysis of EGFR binding to JEV-ED3 protein. The association and dissociation curves resulting from the binding events are shown, wherein the real-time binding sensorgrams are represented in colored lines and their fitting curves are represented in black lines. (c) The JEV-ED3 protein is titrated against different concentrations of recombinant Human EGFR Protein (ECD, hFc Tag), ranging from 0-100 μg/ml. The equilibrium dissociation constants (K_D) were calculated from three independent experiments using the Octet Data Analysis software’s 2:1 (Heterologous binding model) binding model. (d) The JEV-ED3 protein is titrated against different concentrations of recombinant Human EGFR Protein (Isoform VIII, hFc Tag), ranging from 0-100 μg/ml. The binding experiment was repeated three times.

https://doi.org/10.1371/journal.ppat.1013790.g008

The immunoglobulin-like-E protein domain III (ED3) plays a major role in receptor binding and harbors key residues that bind to neutralizing antibodies [10,45]. We used the purified recombinant JEV-ED3 protein to test for specific interaction with EGFR and EGFR mutant protein vIII, which lacks the ligand binding ability with EGF or any known ligands, using BLI. Based on the obtained results, it was evident that the JEV-ED3 protein shows a significant preference for EGFR compared to the EGFR mutant protein vIII (Fig 8c and 8d). The K_D value of JEV3 against EGFR was calculated to be 283 ± 75 nM. Collectively, these results demonstrate the fundamental importance of the EGFR extracellular ligand binding domain and the intracellular kinase domain for JEV binding and entry.

Ethics statement

All animal experiments were approved by the Institutional Animal Ethics Committee of the Regional Centre for Biotechnology (RCB/IAEC/2022/114). Mice were maintained and experiments were performed as per the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Government of India.

Cell lines and virus

Human neuroblastoma cell line SH-SY5Y was obtained from ATCC. Vero and C6/36 cell lines were obtained from the cell repository of National Centre for Cell Sciences Pune, India. SH-SY5Y cells were cultured in HiGlutaXL Dulbecco’s modified eagle medium (DMEM), C6/36 cells in Leibovitz’s L-15 Medium, Vero cells in Minimum essential medium Eagle (MEM) and HeLa cells in DMEM. All media for culturing was supplemented with 10% FBS and 1X penicillin-streptomycin-glutamine (PSG). All reagents used in the study are listed in S1 Table.

JEV Vellore strain P20778 (GenBank accession no. AF080251) was generated in insect cell line C6/36 in 2% FBS with 1X penicillin-streptomycin-glutamine (PSG).

Virus generation and purification

C6/36 cells were infected with JEV 0.1 MOI at a confluency of 70–80% for 1 h and were supplemented with 2% FBS containing L-15 media. Supernatant containing the virus particles were harvested when ~80% of the cells showed cytopathic effect at approximately 72 h post infection (pi). Virus particles were collected by centrifugation at 1,000 × g for 15 min at 4^Oc. Virus concentration was done through multiple cycles of centrifugation of the collected supernatant with Amicon Ultra Centrifugal Filter, 100 kDa MWCO at 1,000 × g for 15 min at 4^Oc. For virus purification, the virus containing supernatant was first concentrated through PEG precipitation, followed with purification over 20% sucrose cushion through ultracentrifugation at 80,000 × g for 4 h at 4^Oc [16].

Virus titration

Virus titre estimation was performed by plaque assays on Vero cells. Cells were plated in 12-well plate and were infected with 10-fold serially diluted virus supernatant for 1 hr at 37^Oc. Virus inoculum was removed post 1 h and cells were washed with PBS twice after which they were overlayed with low-melting-point agarose and 2X MEM in 1:1 ratio for 5 days at 37^Oc until the plaques became visible. Cells were then fixed in 3.75% formaldehyde overnight and the agarose plug were removed followed with staining with 0.1% crystal violet blue for plaque visualization. Titres were determining by counting the number of plaques in the lowest dilutions were they were visible for counting which were represented as plaque forming units (pfu/ml).

Small interfering RNA (siRNA) depletion and virus assays

Cells seeded in 24-well plate were double transfected (day 1 and day 2) with 25 nM siRNA on each day with DharmaFECT 1 transfection reagent according to the manufacturers instruction. Cells were transfected for 72 h before virus infection assays. Knockdown efficiency was checked through qRT-PCR and western blotting. Virus entry assays were performed by allowing the virus 10 MOI to bind on ice for 1 h followed with removing the virus inoculum. Cells were shifted to 37^Oc to allow virus internalization for 1 h and were washed with 1X PBS twice followed with treatment with trypsin to remove extracellular bound virus particles. Cells were then washed with PBS and were harvested with TRIzol for RNA extraction and qRT-PCR. For virus entry assay under FEME conditions, cells were infected with virus 10 MOI for 1 h at 37^Oc followed with virus inoculum removal. Cells were washed with PBS and were either harvested immediately with trypsin treatment (0 hpi) or kept for another 1 h in complete media (1 hpi) followed with trypsin treatment to harvest cells with TRIzol. qRT-PCR was performed with primers against JEV envelope region to measure the absolute viral copy numbers. For virus binding or attachment studies, cells were infected with virus 10 MOI on ice for 1 h to allow virus particles to bind. Cells were washed with chilled PBS and were harvested with TRIzol for RNA isolation and qRT-PCR. For virus infection assays, cells were infected with JEV 1 MOI for 1 h and were washed with PBS. Cells were then kept in complete media for 24 h and were harvested with TRIzol. qRT-PCR was done to measure virus RNA levels. For immunofluorescence based virus entry assays, cells were infected with virus of 100 MOI on ice and were shifted to 37^Oc for different time intervals subsequently fixing it with 4% PFA.

RNA isolation and Quantitative Real Time (qRT)-PCR

RNA extraction was performed with TRIzol RNAiso Plus reagent using a phenol-chloroform based method. Briefly, cells were harvested with TRIzol and were homogenized by pipetting multiple times. 100 ul of chloroform was added per sample and was incubated for 2 mins at RT followed with centrifugation at 12000 rpm for 15 min at 4^Oc to allow phase separation. The clear aqueous phase was separated and 250 ul isopropanol was added. The samples were incubated at room temperature for 15 min to allow RNA precipitation and were centrifuged at 12000 rpm for 15 min at 4^Oc. The supernatant was discarded and the pellet was washed with chilled 75% ethanol with centrifugation at 12000 rpm for 15 min at 4^Oc. Samples were air dried and were resuspended in nuclease free water by heating at 55^Oc for 10 min. cDNA preparation was done using Prime-script Transcription kit and real- time PCR was performed using Quant Studio 6 (Applied Biosystems). All the primer sequences used are mentioned in Table A in S2 Table.

Western blotting

Cell lysis was performed in lysis buffer containing 150 mM NaCl, 1% Triton X-100, 50 mM Tris-HCl pH 7.5, with 1 mM PMSF, and protease inhibitor cocktail for 60 min on ice. Supernatant collected were used for protein quantification using BCA assay kit. 5X loading was added to the cell lysates and these were heated at 95^Oc for 10 min. Equal concentration of protein lysates were run on a SDS-PAGE gel and transferred to PVDF membrane for immunoblotting. Membranes were cut according to the size of the desired protein and were blocked in 5% not fat milk in 1X PBS/TBS for 1 h at RT. Primary antibody incubation was done in the blocking solution at the desired antibody dilution at 4^Oc overnight. Membrane were washed 3 times with 1x PBST/TBST (1% tween in PBS/TBS). Secondary HRP-conjugated antibody was diluted in milk and was incubated for 1 h at RT. Membrane was washed thrice with 1x PBST/TBST, and bands were revealed with chemiluminescence using HRP substrate.

Transmission electron microscopy

SH-SY5Y cells were infected with JEV (MOI 200), incubated in ice for 2 hours and were fixed at 30 seconds post infection in 2.5% glutaraldehyde and 2% PFA in 0.1M sodium cacodylate buffer, followed by post-fixation in 1% osmium tetroxide, dehydration in ethanol, and embedded in LR white resin (Ted Pella). Blocks were sectioned using Leica UC7 Ultramicrotome 5-glass knives. Sections were examined at RT with a transmission electron microscope (Talos, L120C) at 11kx.

Gene cloning and stable mutant cell line generation

Full-length (FL) Human Endophilin A2 (Endo-A2)/SH3GL1(Gene ID: 6455, NP_003016.1) and different functional domain mutants were cloned into mammalian expression vector pEGFP_C1 to obtain N-terminus GFP-tagged fusion protein. Domain specific regions were classified based on the NCBI database. Briefly, full length Endo-A2 (368 aa) was generated by PCR as BglII/EcoRI fragments (FL-F and FL-R) and subcloned into pEGFP_C1 at the same sites. Similarly, fragments expressing 22–368 aa and 1–308 aa were PCR amplified from cDNA, using primers (ΔH_O-F and R and ΔSH₃-F and R, respectively) with 5’ and 3’ tail containing BglII and EcoRI sites, for directional cloning into the pEGFP_C1 vector. These constructs were referred to as EGFP-ΔHO (22–368 aa) and EGFP-ΔSH3 (1–308 aa), respectively.

Overlap extension PCR was used to construct the BAR-domain mutant. This chimera consists of N-terminal 1–24 residues fused to the C-terminal 248–368 aa of Endo A2. In the first round PCR, fragment 1 viz. 1–24 aa was amplified using (ΔH_O-F and ΔBAR_1-R) primers whereas fragment 2 viz. 248–368 aa was amplified using the primers (ΔBAR_2-F and ΔSH₃-R), and both the fragments were then gel purified. The chimeric primers viz. ΔBAR_2-F and ΔBAR_1-R contains ~12 bp overhangs at the 5’ end, which are complementary to fragment 1 and fragment 2, respectively. In a 2-step overlap extension, two purified fragments (fragment 1 and 2) were mixed in equimolar concentration (0.1 pmol each) for overlap extension PCR performed as follows: thirteen cycles of 10 sec at 92^Oc (denaturation), 1 min at 60^Oc (annealing) and 72^Oc for 50 sec (extension). In step 2, resulting PCR product was taken as a template for the nested PCR, where extreme end primers viz. ΔBAR_1-F and ΔBAR_2-R were used for the amplification of the full length (147aa) hybrid fragment. The fusion product was then digested and ligated into pEGFP_C1 vector at BglII-EcoRI sites to generate EGFP-ΔBAR. The recombinant plasmids were transformed into DH5α E. coli strains and screened for the positive clones. All constructs were verified by sequencing. To generate the stable cell lines expressing different domain mutants of Endo-A2, SH-SY5Y cells were seeded in 24 well plate overnight and transfected with the respective plasmids viz. pEGFP_C1 (empty vector), EGFP-FL (full length), EGFP-ΔHO, EGFP-ΔSH3 and EGFP-ΔBAR. Briefly, cells were transfected with plasmid concentration of 2 µg using Lipofectamine™ 3000 Transfection Reagent according to the manufacturer’s protocol. Stably expressing cells with EGFPC1 constructs were selected with neomycin (800 µg/ml) for 4-weeks and were sorted using flow cytometry for varied GFP intensities. Cells with high GFP signals were grown in antibiotic selection pressure and expression of tagged-protein variants was confirmed by immunofluorescence and immunoblotting. All the primer sequences used for cloning are mentioned in Table B in S2 Table.

Primary cortical neurons isolation and culture

Primary cortical neurons were isolated from a previously described protocol [69]. Briefly,. embryos were collected from pregnant mice at embryonic day E.16.5 from C57/BL6 mice through decapitation from the pregnant mice in ice-cold dissociation media, HBSS (1X sodium pyruvate, 20% glucose, 1 M HEPES, pH 7.3). The cortices were gently dissected from the brain and were collected in the dissociation media. Tissues were washed and digested with trypsin and DNAse I at 37^Oc for 20 min. Single cell suspension were collected, washed twice and finally resuspended in neurobasal medium supplemented with 10% FBS, 20% glucose, 1X Sodium pyruvate, and antibiotics. The neurons isolated were plated on poly-l-lysine coated plates. The media was changed the following day upon adherence of the neurons on the plate with maintenance media (neurobasal B-27, 1X glutamine, penicillin-streptomycin solution).

Antibody inhibition assays

Cells were seeded in 96-well plate and were allowed to bind on ice for 1 h with 50 ug/ml of the control (negative) mouse IgG and with specific EGFR antibody against the epitope binding domain (sc-120). This was followed with chilled ice cold PBS washes twice and were subsequently infected with JEV (1 MOI) for 30 min on ice. Cells were washed with chilled PBS and were incubated with complete media for 24 h. Cells were harvested post 24 h and were proceeded with RNA extraction and qRT-PCR. Plaque assays was performed to check virus titres collected from the supernatant. For virus binding/entry assays, post incubation with antibodies, cells were infected with JEV (10 MOI) for 30 min on ice and were washed with chilled PBS twice before harvesting with TRIzol (0 hpi) or adding complete media for 1 h at 37^Oc (1 hpi). Cells were washed and were treated with trypsin to remove non-internalized virus particles followed with PBS wash and were harvested with TRIzol reagent for RNA extraction and qRT-PCR.

EGF stimulation and virus infection assay

Cells were seeded in 24-well plates overnight and were stimulated with 20% serum for 15 min at 37^Oc to activate FEME. For virus binding assay, cells were either EGF stimulated or not, followed with incubation with both virus (10 MOI) and increasing concentration of EGF at 4^Oc for 1 h. Cells were harvested followed with RNA extraction, cDNA preparation and qRT-PCR. For entry assay, cells were stimulated with increasing concentration of EGF for 15 mins at 37^Oc and were subsequently allowed to bind with JEV (10 MOI) and with the increasing EGF concentration on ice for 1 h. Cells were washed with ice cold PBS and were incubated at 37^Oc for 1 h with complete media. Cells were harvested with TRIzol post 1 h after trypsin treatment. For studies of virus replication, the serum and EGF stimulated cells were infected with 1 MOI virus followed with harvesting at 24 hpi.

Cargo uptake assay

For endocytic cargo uptake assays, cells were seeded in coverslip coated 24- well plate overnight. Post-transfection with siRNA for 72 h, for Tf uptake assay, cells were serum starved for 30 mins at 37^Oc. Cells were pulsed with Alexa fluor labelled Tf of concentration 20 ng/well for 5 min at 37^Oc. Cells were washed with acid wash buffer and subsequently with PBS twice to remove uninternalized Tf. Cells were fixed with 4% PFA for 15 min at 37^Oc. For EGF uptake assay, cells were never serum starved, and were pulsed with different EGF concentrations of 10, 20, 50, or 180 ng/ml for 5 min at 37^Oc. Cells were then washed with PBS twice to remove uninternalized EGF and were subsequently fixed with 4% PFA for 15 min at 37^Oc.

Cell viability assay

SH-SY5Y cells were seeded in 96-well plate and were either left untreated (DMSO control) or treated with the drugs with the different concentrations. After 5 h post treatment, MTT was added at a final concentration of 0.5 mg/ml for 3 h followed by addition of solubilization solution to stop the reaction. OD was measured at 570 nm using a plate reader. The percentage cell viability was then calculated and normalized to DMSO treated control cells.

Immunofluorescence studies

JEV ED3 protein purification

JEV ED3 protein purification was performed from a previously described protocol from [10]. Briefly, E. coli BL-21(DE3) cells transformed with pET28a plasmid carrying the ED3 cDNA (aa 303–398) with a c-terminal His-tag were grown at 37^Oc till the O.D₆₀₀ reaches 0.6-0.8. The protein expression was induced by adding 1 mM IPTG (Isopropyl β-D-1-thiogalactopyranoside) at 30°C for 4h with constant shaking at 200 rpm. The induced bacterial pellet was lysed under denaturing conditions in lysis buffer for 1h with continuous swirling at 37^OC (100 mM NaH₂PO₄, 10 mM Tris-HCl, 8 M urea, pH 8.0, 1:50 PI and 1:50 PMSF). Ni-NTA beads were added to the cell supernatant for protein purification and was eluted in 8 M urea containing 200mM imidazole at pH 8. The elute was dialyzed against 8 M, 6 M, 4 M, 3 M, 2 M, 1M urea and finally in PBS. Dialyzed protein was collected and centrifuged at 12000 rpm for 90 min at 4^Oc and supernatant was concentrated using 3 KDa Amicon filter. The protein concentration was determined using BCA and ED3 protein bands were confirmed using western blots as a single band at ~10 KDa molecular weight.

Biolayer Interferometry (BLI)

All the binding experiments were performed using an Octet RED96 (Forte Bio) at 30°C with shaking at 800 rpm. Prior to use, the Ni-NTA biosensors were loaded into the columns of a biosensor holding plate and pre-hydrated in assay buffer for 15 min. Protein JEV-ED3 with a c-terminal His-tag was immobilized on Ni-NTA biosensor tips (Forte Bio) at a concentration of 20 μg/mL in an assay buffer containing PBS. Binding interactions were measured by submerging the JEV-ED3 loaded biosensors in solutions containing different concentrations of recombinant Human EGFR Protein (ECD, hFc Tag) or (Isoform Viii, hFc Tag) in a two-fold dilution series. The association and dissociation were measured for 400 s. The biosensors were regenerated for 3–4 cycles using a 10 mM glycine solution, pH 2.0, and then neutralized in the assay buffer. After regeneration, the biosensors were recharged using a 10 mM NiCl₂ solution for 60 s. The sensor with 0 μg/mL recombinant EGFR was used for reference subtraction. The binding data were fitted using a 2:1 (Heterologous binding model) binding model using Octet Data Analysis software to evaluate the equilibrium dissociation constants (K_D).

Statistical analysis and software

GraphPad Prism version 9 was used to perform all the statistical analysis. Statistical analysis was done using paired/unpaired Student’s t-test, Kruskal-Wallis test with Dunn’s multiple comparisons test or one-way ANOVA followed by Dunnett’s/Šídák’s multiple comparisons multiple comparison test. Error bar indicates means ± SD/SEM. Microscopic image analysis was performed on Image J or Imaris version 9.9.