Rainbow trout (Oncorhynchus mykiss) individuals approximately 6-7 cm in size were acquired from a certified commercial farm (Piszolla S.L., Cimballa Fish Farm, Zaragoza, Spain). The fish supplier followed strict biosecurity protocols that include the regular monitoring of stock health and compliance with veterinary standards to ensure that the fish are free of pathogens. Fish were maintained at 14°C with a recirculating, dechlorinated water system at the Miguel Hernández University (UMH) animal house facilities. Prior to experimentation, fish were acclimatized to laboratory conditions for two weeks. The number of fish samples used is expressed by an “n” in each experiment’s figure legend.

The protocol used to obtain and purify rainbow trout RBCs has been previously described (5, 30). In brief, fish were bled from the caudal vein using insulin syringes (NIPRO, Bridgewater). Collected blood samples were placed in a 2 mL eppendorf tube with RPMI-1640 medium (Dutch modification) (Gibco, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) gamma irradiated (Cultek), 1 mM pyruvate (Gibco), 2 mM L-glutamine (Gibco), 50 μg/mL gentamicin (Gibco), 2 μg/mL fungizone (Gibco), and 100 U/mL penicillin/streptomycin (Sigma-Aldrich). These samples were layered out onto 2 mL of Histopaque-1077 Ficoll gradient solution (Sigma-Aldrich), a sterile-filtered, ready-to-use solution with a density of 1.077 g/mL. Two consecutive density gradient centrifugations were carried out for purification. Then, RBCs were washed twice with RPMI 2% FBS. Finally, RBCs were cultured with RPMI 10% FBS in 25 cm² flasks (Nunc Roskilde) at 14°C for 24 hours before experimentation.

The VHSV strain 07.71 (31) was purchased from the ATCC (VR-1388) and cultured in fathead minnow Epithelioma papulosum cyprini (EPC) cell line at 14°C as previously reported (32). VHSV is known to replicate efficiently in EPC cells with a high degree of reproducibility (33).

VHSV (10⁸ tissue culture infectious dose 50% [TCID₅₀]/mL) was inactivated by two rounds of UV-B irradiation at 1 J/cm² using a Bio-Link crosslinker BLX E312 (Vilber Lourmat, BLX-E312), following previously validated methods (9). The viral suspension was placed in a shallow multi-well culture plate, to ensure minimal liquid height to maximize UV light penetration and effectiveness. Ficoll-purified rainbow trout RBCs were exposed to VHSV_UV in RPMI 2% FBS, at a multiplicity of infection (MOI) of 10. RBCs unexposed to VHSV_UV were used as a control. At 12 hours post-exposure (hpe), RBCs were placed on top of Quantifoil holey film copper TEM grids (Au-G200F1) coated with poly-L-lysine (Merck, Germany), with fiducial gold markers (100 nm; BBI Solutions, UK) and vitrified by plunge freezing in liquid ethane in a Leica EM-CPC. We selected the 12 hpe time point based on previous studies indicating that VHSV replication decreases between 6- and 24-hpe (5). This time frame was chosen to capture a critical phase for evaluating cellular responses, while considering the viral replication dynamics.

Vitrified grids were transferred in liquid nitrogen to the cryo-correlative cooling stage (CMS196 stage; Linkam Scientific Instruments, UK) to hold samples stably at 190°C during analysis. The cryo-stage was inserted into a Zeiss Axio Scope (Carl Zeiss, Germany) epifluorescence microscope to visualize the frozen grids. Selected samples were transferred under cryogenic conditions to the MISTRAL beamline (ALBA light source) at ALBA Synchrotron (34, 35). Tomographic data were collected at 520 eV, irradiating the samples for 1-2 s per projection. 520 eV is the energy range in which water is transparent for X-rays and they are mostly absorbed by carbon, allowing the visualization of biological material. In tomographic setup, images obtained at different sample orientations are computationally combined to produce a three-dimensional image, permitting the three-dimensional representation of the sub-cellular ultrastructure of whole, intact cells. A tilt series was acquired for each cell using an angular step of 1° on a 70° angular range with a Fresnel Zone plate (FZP) with a 40 nm outermost zone width and effective pixel size of 13 nm.

Each transmission projection image of the tilt series was normalized using flat-field. This process considers the possibility of different exposure times and the slight decrease of the electron beam current during the acquisition. To increase the image quality, wiener deconvolution considering the experimental impulse response of the optical system was applied to the normalized data (36). The Naperian logarithm was used to reconstruct the linear absorption coefficient (LAC). The resulting stacks were loaded into IMOD software (RRID: SCR_003297), and individual projections were aligned to the common tilt-axis using the internal cellular structures as markers (37). Then, the aligned stacks were reconstructed with algebraic reconstruction techniques (38). The visualization, feature segmentation, and quantification of the LACs and volumes were carried out using Amira 3D software (Thermo Fisher, RRID: SCR_007353). Cellular organelles are distinguished by image contrast, which is due to differential and quantitative absorption of X-ray photons in the ‘water window’, between the absorption edges of carbon (284 eV) and oxygen (543 eV) (39). The interaction of X-rays is element-specific and is due to the relative concentration of organic material and water. We used previous knowledge on the appearance of cellular organelles under cryo-SXT (40) together with representative TEM images (7, 41–45).

The euchromatin/heterochromatin (EU/HET) ratio was calculated taking into account the absorbance of the segmented material that is the corresponding LAC for each voxel of the volume reconstructed with SXT (46). For this purpose, the nuclei were analyzed and compared in terms of volume, absorbance, and shape. As a result, a difference in density can be observed, demonstrating a bimodal distribution. This allows the selection of regions of lower absorption (i.e., light areas) and regions of higher absorption (i.e., dark areas) within each nucleus. Regions with low LAC values correspond to euchromatin regions with low chromatin condensation and more transcriptional machinery. In contrast, regions with high LAC values correspond to heterochromatin, a condensed form of DNA with less transcriptional activity (47).

Ficoll-purified RBCs from the peripheral blood of rainbow trout individuals (5 x 10⁵ cells/well) were exposed to VHSV_UV at MOI 10 and VHSV at MOI 1 and 10, in RPMI 2% FBS at 14°C, and incubated for 6 and 24 hours. Samples were resuspended in TRK lysis buffer (Omega Bio-Tek, Inc.) and stored for RNA extraction and quantitative real-time PCR (qPCR) analysis as previously described, and detailed below (5). For differential expression analyses, we examined a set of genes that are well-known representative components of the UPR (22, 48–50). Evaluated genes and primers used can be found in Table 1 .

To examine the cytotoxicity of 4-phenylbutyric acid (4-PBA) (Sigma-Aldrich) or niclosamide (Sigma-Aldrich) on rainbow trout RBCs, propidium-iodide (PI) (Sigma-Aldrich) staining was performed to discriminate between live and dead cells using flow cytometric analysis. Briefly, RBCs were treated with 4-PBA or niclosamide according to the times and conditions used in the assays. For 4-PBA, RBCs were incubated for 24 and 72 hours at the maximum concentration used (8 mM), whereas for niclosamide, RBCs were incubated for 24 hours at 10 µM. As a positive control to determine cell death, 10 μL of 50% hydrogen peroxide (H₂O₂) (an apoptosis inducer) was added to untreated RBCs for 10 minutes. Afterward, cells were stained with 10 µL PI (1 mg/mL). Possible cell damage was detected by measuring the fluorescence intensity of PI-stained RBCs using a FACS Canto II flow cytometer (BD Biosciences).

To assess the functional involvement of the ER in viral defense, we used the ER stress inhibitor 4-PBA and tested the susceptibility of cells to VHSV exposure. 4-PBA acts as a chemical chaperone whose hydrophobic regions interact with the hydrophobic segments of the unfolded protein, thereby promoting protein folding and reducing ER stress (51). To evaluate the effect of ER stress inhibition on viral replication, cells were treated with 4-PBA prior to virus exposure as previously described (52). For this purpose, Ficoll-purified rainbow trout RBCs (5 x 10⁵ cells/well) were incubated for 24 hours at 14°C with 2, 4, or 8 mM 4-PBA in RPMI 10% FBS. RBCs were washed and exposed to VHSV at MOI 1 for 6 or 72 hours in RPMI 2% FBS at 14°C. Samples were stored in TRK lysis buffer for RNA extraction and qPCR analysis.

To evaluate the effect of ER stress inhibition on autophagy, in the context of viral infection, Ficoll-purified RBCs (5 x 10⁵ cells/well) were pre-treated with 8 mM of 4-PBA for 24 hours at 14°C and then exposed to VHSV at MOI 10 for 6 hours. In parallel, RBCs were exposed to VHSV for 3 hours and then incubated 24 hours with 10 µM of niclosamide, an autophagic flux blocker. Then, we analyzed the expression of ubiquitin-binding protein p62 to monitor autophagic flux as well as the N-VHSV protein to quantify viral replication, by means of flow cytometry, as explained below.

Flow cytometry analysis was carried out using a FACS Canto II flow cytometer (BD Biosciences). RBCs were fixed with 4% paraformaldehyde (PFA) and 0.008% glutaraldehyde diluted in RPMI medium for 45 minutes. Permeabilization of RBCs was carried out in 0.05% saponin (Sigma-Aldrich) buffer for 15 minutes. We used rabbit anti-p62/SQSTM1 antibody (www.antibodiesonline.com; Ref #ABIN2854836, RRID: AB_3096920) diluted 1/250 in permeabilization buffer to track the autophagy process. We used murine 2C9 anti-NVHSV (53) (RRID: AB_2716276) diluted 1/1000 in permeabilization buffer to track viral replication. Primary antibodies were incubated for 60 minutes. Goat anti-rabbit IgG (H+L) CF647 and goat anti-mouse IgG (H+L) CF488 (Sigma-Aldrich) diluted 1/200 in permeabilization buffer, and incubated for 30 minutes, were used as secondary antibodies. RBCs were stored in 1% PFA in PBS.

Individual rainbow trout were challenged with 50 µL VHSV (10⁸ TCID₅₀/mL) resuspended in RPMI 2% FBS by intramuscular injection. Negative controls were injected with 50 µL RPMI 2% FBS. Animals were kept at 14°C over the course of the infection. Fish were sacrificed at 2 days post challenge (dpc), and peripheral blood collected from the caudal vein was purified to obtain the RBCs as described above. RBCs were stored in TRK lysis buffer for RNA extraction, and gene expression was analyzed by qPCR. We analyzed the genes involved in UPR and related cellular pathways listed in Table 1 .

Samples were first digested with Proteinase K solution (Omega Bio-Tek) and then RNA was isolated using the E.Z.N.A. Total RNA Kit (Omega Bio-Tek) according to the manufacturer’s instructions. Briefly, samples were applied to the HiBind RNA Mini Column, where total RNA binds to the matrix. Cell debris and contaminants were removed by a series of wash steps. Finally, high-quality RNA was eluted in DEPC-treated water. TURBO DNase (Ambion, Thermo Fisher Scientific Inc.) was used to remove residual genomic DNA as previously described (5). RNA was quantified with a NanoDrop spectrophotometer (Nanodrop Technologies). cDNA was synthesized from RNA using M-MLV reverse transcriptase (Invitrogen, Thermo Fisher Scientific) as previously described (54). qPCR reactions were performed in a total volume of 20 μL containing cDNA, 900 nM of each primer, 10 μL of TaqMan Universal PCR master mix (Thermo Fisher Scientific) with 300 nM probe, or 10 μL of SYBR Green master mix (Thermo Fisher Scientific). qPCR was carried out using the QUANTSTUDIO 3 system (Applied Biosystems, Thermo Fisher Scientific Inc.). Cycling conditions were 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min (5). Gene expression was analyzed by the 2^-ΔCt or 2^−ΔΔCt method (55) as indicated in each experiment, and the ef1α gene was used as an endogenous control. Primers and probes sequences are listed in Table 1 . The number of samples analyzed in each experiment is indicated by an “n” in the figure legends.

To visualize the ER changes in RBCs in response to VHSV, the maximum concentration of 4-PBA (8 mM) was chosen. RBCs were incubated with 4-PBA for 24 hours prior to VHSV exposure. Then, RBCs were exposed with VHSV at MOI 1 for 6 hours. Cells were washed with fresh medium and stained with ER-Tracker Red reagent (BODIPY TR Glibenclamide, Thermo Fisher Scientific Inc.) for 30 minutes, CellMask Green Plasma Membrane Stain (Thermo Fisher Scientific Inc.) for 15 minutes, and Hoechst reagent (Thermo Fisher Scientific Inc.) for 5 minutes. Dyes were diluted 1:1000. Images were taken with a Zeiss LSM900 with Airyscan 2 and analyzed with ZEN 3.2 (Blue Edition) software (Carl Zeiss Microscopy, RRID: SCR_013672). This software was also used to measure the fluorescence intensity in the Red Channel (“ER staining”) and to represent and quantify ER volume at multi-depth along the z-axis (Z-stacks). Volume rendering was performed using IMARIS software v9.3 (Bitplane, RRID: SCR_007370).

GraphPad Prism 6 software (GraphPad Software Inc., RRID: SCR_002798) was used for graphic creation and statistical analysis. Statistical tests and P values are indicated for each assay. Flowing software 2.5.1 (www.flowingsoftware.com/, RRID: SCR_015781) was used to process and analyze flow cytometry data, and Floreada.io software (RRID: SCR_025286) was applied to generate histograms and dot plots for cytotoxicity assays. ZEN 3.2 (Blue Edition) software and IMARIS v9.3 were used to analyze and produce confocal microscopy images. Amira Thermo Scientific software (Thermo Fisher) was used for the analysis and processing of images obtained by cryo-SXT.

Three-dimensional reconstruction of cryo-SXT images of control RBCs and VHSV_UV-exposed RBCs was carried out to gain insight into the structural alterations that occur at an intercellular level ( Figure 1 ). The sample preparation and data collection steps carried out for cryo-SXT are outlined in Figure 1A . We also calculated the volume of the organelles of interest and the level of structural packing of the chromatin ( Figure 2 ).

Based on 3D volume representations, we identified subcellular differences between VHSV_UV-exposed and control RBCs. We have been able to clearly identify cell organelles such as ER, mitochondria, autophagosomes and extracellular vesicles (EVs). However, cryo-SXT has limited resolution visualization of some cellular structures, such as the Golgi apparatus.

Regarding ER structures, the most striking result was the thickening of the ER in VHSV_UV-exposed RBCs ( Figure 1B ). Another distinguishing feature was the increased number and size of autophagosome-like vesicles ( Figures 1B–D ) in contrast to control RBCs ( Figure 1E ). Interestingly, these structures are often located close to the nucleus where the ER is found. Furthermore, we noted protrusions on the plasma membrane of VHSV_UV-exposed RBCs, which could correspond to the secretion of EVs ( Figure 1D ). Volumetric quantification supported these observations: we detected higher ER volume and more autophagosome and EV formation in VHSV_UV-exposed vs control RBCs ( Figure 2A ).

VHSV_UV-exposed RBCs had higher EU/HET ratios than control RBCs ( Figure 2B ), suggesting increased transcriptional activity in VHSV_UV-exposed RBCs. The selection of the LAC regions corresponding to euchromatin and heterochromatin is depicted in Supplementary Figure 1 .

We evaluated the expression of major UPR genes and their downstream targets in response to ER stress activation by VHSV_UV and live VHSV, at two time points. The goal was to identify a differential pattern of gene expression in the experimental conditions of interest: RBCs exposed to VHSV_UV ( Figure 3B ); RBCs exposed to VHSV MOI 1 and MOI 10 ( Figures 3C, D , respectively). The experimental workflow is detailed in Figure 3A .

The results showed significant upregulation of the master controller of ER stress response grp78 and ATF6 UPR branch genes, such as atf6 and their target calreticulin (calr), under VHSV_UV or live VHSV exposure compared to control RBCs (unexposed RBCs) ( Figures 3B–D ). In detail, these three genes were overexpressed in RBCs exposed to VHSV_UV at 6 hpe ( Figure 3B ), compared to control RBCs, whereas in RBCs exposed to VHSV MOI 1 and MOI 10, the response was mainly delayed, with significant upregulation observed at 24 hpe ( Figures 3C, D ), except for grp78 expression in RBCs exposed to VHSV MOI 1 at 6 hpe. Regarding PERK UPR branch, we analyzed the expression of atf4 and chop. In relation to IRE1 UPR branch, we evaluated the ER degradation enhancing alpha-mannosidase like protein 1 (edem1) gene, which is a target gene of XBP1. Neither in RBCs exposed to VHSV_UV ( Figure 3B ) nor in RBCs exposed to VHSV at MOI 1 and MOI 10 ( Figures 3C, D , respectively) was the expression of atf4 or edem1 affected, compared to control RBCs. No upregulation of chop was observed after the exposure with VHSV_UV, compared to control RBCs, and, likewise, chop expression remained unchanged at both time points ( Figure 3B ). Nevertheless, chop overexpression was observed in RBCs in response to VHSV MOI 1 and MOI 10, compared to control RBCs, and in particular at MOI 1, at 24 hpe, when it was statistically significant ( Figures 3C, D ).

Overall, the results indicated that RBCs exposed to VHSV_UV or live VHSV triggered an UPR regulation. However, it should be clarified that although VHSV_UV confirmed rapid induction of UPR mainly through the ER chaperone GRP78 and the ATF6 pathway, the expression levels of the related genes were higher in RBCs in response to live virus (VHSV MOI 1 and MOI 10).

Pre-treatment with the ER stress inhibitor 4-PBA at different concentrations (2, 4, and 8 mM) for 24 hours was carried out prior to VHSV exposure to investigate the involvement of ER stress in the response to the virus in rainbow trout RBCs ( Figure 4A ). Then, viral replication was analyzed by qPCR of the N-VHSV gene in RBCs at 6 and 72 hpe. N-VHSV gene expression increased in a dose-dependent manner after ER stress inhibition with 4-PBA ( Figure 4B ). It should be highlighted that, as 4-PBA concentration increased, so did N-VHSV gene transcription, even at 72 hpe it was statistically significant for the maximum concentration used. This rise of viral replication at 72 hpe is surprising, since previous reports found that the VHSV replication cycle is halted in RBCs beyond 24 hours (5). The effect of 4-PBA on UPR modulation was also assessed in RBCs ( Supplementary Figure 2 ). grp78 and calr expression levels were significantly lowered in 4-PBA-treated RBCs at 6 hours, relative to their control. In contrast, at this time point, both genes were significantly upregulated in 4-PBA-treated and VHSV-exposed RBCs, when compared to 4-PBA-treated RBCs. Thus, 4-PBA appeared to attenuate UPR in RBCs while VHSV exposure acted as a stress condition that 4-PBA could not resolve. Prior to experiments, the absence of cytotoxicity of 4-PBA on RBCs was confirmed using PI staining method and measured by flow cytometry ( Supplementary Figure 3 ).

To achieve consistency in our observations on the response of RBCs to the virus under ER stress inhibition, we performed morphological monitoring of the ER under 4-PBA treatment (8mM) and VHSV exposure. For this, we combined live-cell fluorescence microscopy through ER-Tracker with quantification of ER volume based on confocal microscopy Z-stack images. As shown in confocal microscopy images ( Figure 5A ), VHSV-exposed RBCs displayed a strong fluorescent ER signal followed by RBCs pre-treated with 4-PBA and exposed to VHSV. We calculated the mean fluorescence intensity (MFI) of the ER signal and the results showed that it significantly increased in VHSV-exposed RBCs whereas, in those RBCs treated with 4-PBA (8 mM), ER signal decreased ( Figure 5B ). Further, as shown in 3D reconstructed images from confocal microscopy, VHSV exposure led to an enlargement of the membranous network of the ER, together with ER expansion ( Figure 5C ). We next quantified ER volume in the presence or absence of 4-PBA. Analogously with MFI results, a significant increase in ER volume was visible when RBCs were exposed to VHSV but when pre-treated with 4-PBA the ER volume decreased ( Figure 5D ). It is worth pointing out here that, when considering the results of cryo-SXT, both VHSV and VHSV_UV shared a similar pattern in RBCs, expanding the ER in response to the stimulus.

We decided to investigate whether autophagy may be altered by ER stress inhibition. For this, 4-PBA was used as an ER stress inhibitor, and niclosamide was used to block autophagic flux due to inhibition of autolysosomes formation (56). Previously, we evaluated whether niclosamide treatment had a cytotoxic effect on RBCs and results showed that niclosamide treatment at 10 µM did not induce cell damage ( Supplementary Figure 4 ). As summarized in Figure 6A , for ER stress inhibition, RBCs were preincubated with 4-PBA and then exposed to VHSV. In the case of autophagy blockade, RBCs were first exposed to VHSV and then treated with niclosamide. By means of flow cytometry, we analyzed the ubiquitin-binding protein p62 to monitor autophagic flux as well as the N-VHSV protein to quantify viral replication under these conditions. When VHSV-exposed RBCs were treated with 4-PBA or niclosamide, we found a slight (non-significant) increase in p62 protein levels accompanied by an increase in N-VHSV protein levels, compared to untreated RBCs ( Figures 6B, C ). Although the differences were not significant, these results suggest that while both ER stress inhibition and autophagy flux blockade appear to affect VHSV replication in RBCs similarly, we cannot say whether ER stress directly affects autophagy or vice versa.

We evaluated the possible correlation between the expression of UPR-associated genes (grp78, chop, atf6, and calr), and genes related to antigen presentation and vesicular transport pathways (mhcI, mhcII, exocyst complex component 1 [exoc1], and SEC13 homolog, nuclear pore and COPII coat complex component [sec13]) in RBCs of VHSV-challenged rainbow trout. Additionally, we evaluated the gene tumor necrosis factor receptor-associated protein 1 (trap1), which is known to be involved in ER stress by modulating the UPR and assisting in protein folding and mitochondrial stress (57, 58) ( Figure 7A ). At 2 dpc, rainbow trout RBCs showed upregulation of the genes studied, with statistically significant upregulation of grp78, calr, mhcI, and sec13 genes ( Figure 7B ). The gene heatmap showed a marked differential ER stress, antigen presentation and vesicular transport response in RBCs from VHSV-challenged individuals ( Figure 7C ). Although we cannot definitively conclude that these pathways are directly interconnected, the observed simultaneous activation suggested a coordinated response to VHSV infection. A more comprehensive analysis is required to explore potential mechanistic links.