Madin-Darby bovine kidney (MDBK) cells (a kind gift from the Kimron Veterinary Institute, Beit-Dagan, Israel). Cells were checked for absence of Bovine Viral Diarrhea Virus (BVDV) and were cultured in Eagle’s Minimum Essential Medium (EMEM) supplemented with 10% of non-inactivated horse serum (HS), 2 mM L-Glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and nonessential amino acids (Cat. number: 01-340-1B) (All from Biological Industries, Israel). Spontaneously immortalized ovine kidney (OK) cells and Vero African Green Monkey Kidney cells were cultured and employed in viral passaging and plaque assays as previously described in (Shai et al., 2013; Dellac et al., 2021).

The inhibitors’ sources and final concentrations are indicated in brackets: ammonium chloride (Sigma-Aldrich, A9434; 25 mM), dynasore (Sigma-Aldrich, D7693; 80 μM), dynole 34-2 (Abcam, ab120463; 40 μM), ikarugamycin (Cayman Chemical, 15386; 0.5 μM, 2 μM), methyl-β-cyclodextrin (Sigma-Aldrich, 332615; 15 mM), amiloride (Sigma-Aldrich, 1019701; 1 mM), latrunculin-B (Abcam, ab144291; 1 μM).

4’,6-diamidino-2-phenylindole (DAPI, Invitrogen, D1306), Alexa Fluor 568 Phalloidin (Invitrogen, A12380), Lysotracker deep red (Invitrogen, L12492).

Epizootic hemorrhagic disease virus serotype 2-EHDV2-Ibaraki was obtained and passaged as in (Shai et al., 2013). We have previously described the generation of the VSV mutant VSV-ΔM51 (Dellac et al., 2021). The same procedure was used to recover the GFP-expressing VSV clone (VSV-GFP) (Andreu-Moreno and Sanjuán, 2018), generously provided by Dr. Ron Geller (Institute for Integrative Systems Biology (I2SysBio), Universitat de Valencia-CSIC, Valencia, Spain). Bluetongue virus serotype 8 (BTV-8) was a kind gift from Prof. Eyal Klement (Koret School of Veterinary Medicine, Hebrew University of Jerusalem, Israel).

EHDV2-Ibaraki ( 1.2 × 10 7 PFU/ml) was irradiated as described in (Shai et al., 2013).

For EHDV2-Ibaraki and BTV-8, viral titers were quantified by plaque assays previously described (Shai et al., 2013). For VSV infections, plaque assay was performed as previously described (Dellac et al., 2021).

The cell pellet was resuspended in cold RIPA solution (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% Deoxycholate, 0.1% SDS, 140 µM NaCl) supplemented with 1:25 protease inhibitor cocktail (Sigma-Aldrich) and 1:100 phosphatase inhibitor was cocktails I and II (Sigma-Aldrich) and incubated at 4°C for 1 hour. Lysed cells were centrifuged at 13,000 rpm, 4°C for 13 minutes. Supernatants were collected, and protein concentration was determined via Bicinchoninic acid (BCA) assay (Cyanogen). Equal amounts of protein extracts (30 µg) in sample buffer containing β-mercaptoethanol were separated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) through a 10% polyacrylamide gel and subsequently transferred onto a nitrocellulose membrane (Bio-Rad). Membranes were blocked in 5% skim milk dissolved in TTBS (20 mM Tris-HCl pH 7.5, 15 mM NaCl, 0.1% TWEEN20) for 1 hour at room temperature and then incubated with a primary antibody for 1 hour at room temperature. Next, membranes were incubated with a secondary antibody linked to HRP in 5% skimmed milk for 1 hour at room temperature. Immunoreactive bands were detected by chemiluminescence. Densitometry of the immunoblots was quantified using the Image Lab program.

Monoclonal mouse anti-nonstructural protein 3 of EHDV2-Ibaraki antibody (anti-NS3) was previously described (Shai et al., 2013)). The α-Myc tag hybridoma (9E10) was a generous gift from Prof. Yoav Henis (Tel Aviv University). Commercial antibodies included monoclonal mouse anti-HSP-70 (Sigma-Aldrich, H5147), monoclonal mouse anti-β-Actin (Merck, A2228), and monoclonal rabbit anti-LAMTOR4 (Cell Signaling, 13140). Secondary antibodies included HRP-conjugated goat anti-mouse IgG (Jackson ImmunoResearch, 115-035-146), goat-anti-mouse and goat-anti-rabbit antibodies labeled with Alexa-488 or Alexa-647 respectively (Invitrogen).

Total RNA was extracted from cells using the EZ-RNA kit (Biological Industries Beit Haemek, 20-400-100), followed by reverse transcription using the q-PCR-BIO cDNA Synthesis Kit (q-PCR-BIO, PB30.11-10) according to the manufacturer’s instructions, with additional no-RNA control. Real-Time PCR analysis of the mRNA levels of IFNβ, NS3, and GAPDH was done in triplicates, using Fast SYBR-green master mix (Bio-Rad CFX connect) with BIO-RAD Real-Time PCR System (Applied Biosystems, 4376600). The following RT-qPCR primers were used: bovine IFNβfw 5’AAGACTCAGCTTCAGCACCTAC3’, bovine IFNβrev 5’ACACTCTTTAAGGCTCTGACG3’, NS3fw 5’CACGCCAACATCAATGCCAA3’, NS3rev 5’TGACGCATACGCAACCTTCT3’, BTV8fw 5’GCATAAAATACTAGAGGATGGCG3’, BTV8rev 5’AGCYGTTCCAATCACAACC3’ bovine GAPDHfw 5’AAGGTCGGAGTGAACGGATTC3’, bovine GAPDHrev 5’ATGGCGACGATGTCCACTTT3’.

Imaging was performed with two systems: 1- Axiovert 200M (Carl Zeiss MicroImaging) controlled by SlideBook 5.0 (Intelligent Imaging Innovations), employing either a 63x oil immersion objective (Plan Apochromat, NA 1.4) or a 10x air objective (Plan Apochromat, NA 0.25) and registered on an EZ camera (Photometrics). 2- EVOS M5000 Imaging System (Invitrogen) employing either 10x air or 60x oil immersion objectives.

Transferrin internalization. 2x10⁵ MDBK cells were plated on 13 mm-glass coverslips. Cells were pre-treated (or not) with the different inhibitors (30 min, in serum-free medium). Incubation with transferrin Alexa Fluor 488 Conjugate (Invitrogen, T13342, 50 μg/ml) was for 20 min at 37°C, in the same medium as the pre-treatment. Fixation was with 4% paraformaldehyde (PFA), after which cells were permeabilized in PBT solution (2% BSA, 0.1% Triton X-100 in PBS, 5 min RT) and stained with DAPI (45 minutes, in PBT). Mounting was with the fluorescent mounting medium (Getter; E18-18).

Endocytosis of Myc-labeled type II transforming growth factor-β receptor (Myc-TβRII). COS7 cells were grown on glass coverslips and transfected with Myc-TβRII in pcDNA3. 24 h after transfection cells were labeled in the cold with anti-Myc IgG [as in (Chetrit et al., 2011)]. Cells were either kept on ice or warmed to 37°C for different time periods, after which they were fixed and permeabilized as described above. Subsequently, cells were permeabilized and stained with secondary antibodies and DAPI prior to mounting.

Staining of LAMTOR4. 2x10⁵ MDBK cells grown on glass coverslips were treated or not with IKA (1 µM, 20 h) prior to fixation, permeabilization, and staining as described above.

For staining of acidic organelles, 2x10⁵ MDBK cells were plated in a 12-well plate, treated or not with IKA (2 μM, 24 h), and incubated with lysotracker deep red (Invitrogen, L12492, 30 μM, 25 min, 37°C).

Acid wash, for the removal of membrane-localized transferrin or α-myc antibodies, was performed by repeated cold washes with acidic glycine buffer (150 mM NaCl, 0.1M Glycine, pH 2.5) prior to fixation.

Transfections were performed with lipofectamine 2000 (Invitrogen, 11668019) in Opti-MEM (Gibco, 31985062) according to the manufacturer’s protocol. The plasmids encoding for LCA-GFP, Rab5-GFP, and Rab7-GFP were kind gifts by Tomas Kirchhausen (Harvard University). Myc-tagged TβRII has been previously described (Ehrlich et al., 2001). For GFP-fusion markers, after 12h of transfection, the cells were treated or not with IKA (1 μM) for 24 h. For endocytosis experiments with myc-TβRII, at 16 h-post transfection, cells were pre-treated (or not) with IKA (2 µM, 30 min) before labeling and internalization steps.

To quantify the fluorescence intensity of cell-associated fluorescent transferrin, cell fields were imaged in three dimensions (24 stacks, 0.16 μm spacing between stacks). Image stacks were deconvolved (NearestNeighbours option, SlideBook 5.0) and projected on a 2D image. Signals were segmented according to intensity thresholds for the DAPI and FITC channels. The sum intensity of FITC was normalized to the cell number by dividing it with the sum intensity of the DAPI signal.

To quantify the percentage of VSV-infected cells, fields of infected cells were imaged with a 10x objective. The number of cells in the field and the VSV-positive cells were assessed via intensity-based segmentation and size-based object definition, employing Slidebook™. Quantification of the percentage of cells exhibiting LAMTOR vacuoles was performed via manual inspection of multiple fields acquired with a 60x objective.

One-way ANOVA and Dunnett’s test for multiple comparisons were applied in experiments comprising a single control and multiple experimental conditions. For comparisons of one treatment and its control, a two-tailed t-test was performed, assuming equal variance between groups. A P value of ≤ 0.05 was considered significant. All statistical analysis was done using the GraphPad Prism software.

Within the orbivirus genus of the Sedoreoviridae family, Bluetongue, and Epizootic Hemorrhagic Disease viruses exhibit a considerable sequence similarity. Comparison of the sequences of EHDV2-Ibaraki with those of 15 BTV serotypes (summarized in Table 1 ) showed average identity values at the amino acid level for 10 structural and non-structural proteins (VP1-7, NS1-3) ranging from 23% (for VP2) to 80% (for VP3), and average residue similarity values ranging from 42% (for VP2) to 91% (for VP3). The entry of BTV into mammalian cells involves the VP2 and VP5 proteins (Forzan et al., 2007; Zhang et al., 2010), with critical pH-sensing functions being performed by conserved histidines (H₃₈₅-H₃₈₆ in BTV-VP5 and H₁₆₄ of BTV-VP2) and cysteines in BTV-VP2 (C₁₆₂, C₆₁₇, C₈₅₁). The latter, together with H₁₆₄, form a Zinc finger (Wu et al., 2019). These conserved residues mediate conformation alterations to VP2 and VP5 following protonation in acidic endosomal compartments (Hassan and Roy, 1999; Wu et al., 2019). Comparison of the VP2 and VP5 proteins of EHDV2-Ibaraki with those of 10 BTV serotypes (for VP5) or 7 BTV serotypes (for VP2) revealed alignment of H₃₈₇ and H₃₈₈ of EHDV2-Ibaraki-VP5, and H₁₆₄ of EHDV2-Ibaraki-VP2, align with cognate histidine residues in BTV-VP5 and BTV-VP2 ( Figure 1 ). Moreover, the three Zinc-finger-associated cysteines of VP2 (C₁₆₂, C₆₁₇ and C₈₅₁ in BTV1), are also conserved in EHDV2-Ibaraki (C₁₆₂, C₆₃₀ and C₈₇₃). Based on this, we hypothesized that EHDV2-Ibaraki requires endosome acidification for infection of MDBK cells and tested this by treating cells with NH₄Cl (25 mM) 30 minutes before and throughout infection (MOI = 1, 24 h). To visualize the levels of infection and the portion of infected cells, we labeled uninfected, infected, and infected/NH₄Cl-treated cells against the nonstructural protein 3 (NS3) of EHDV2-Ibaraki, and imaged them by fluorescence microscopy. This revealed a complete lack of NS3 expression in cells treated with NH₄Cl, indicative of a complete inhibition of infection ( Supplementary Figure 1 ).

Next, we assessed the dependence of the inhibition of EHDV2-Ibaraki infection on the time of NH₄Cl addition. For this, MDBK cells were treated with NH₄Cl (25 mM) at 30-minute intervals relative to challenge with EHDV2-Ibaraki (MOI=1). Times of addition NH₄Cl ranged from 30 min pre-treatment and included 0, 30, 60, 90, 120, or 180 min relative to the time of addition of virus (see schematic depiction, Figure 2A ). Infection levels were assessed by examining NS3 expression at 24 hours post-infection (hpi), a time frame required for reliable detection under these experimental conditions. Figure 2B shows a representative immunoblot of this experiment, while the graph in Figure 2C summarizes multiple repeats. NS3 expression was undetectable when NH₄Cl was added 30 min before, together with, or 30 min following EHDV2-Ibaraki challenge. The addition of NH₄Cl at later time points (up to 2 h) resulted in progressive increase in infection in multiple experiments, while a lesser increase was observed between the 2 or 3-hour time point of NH₄Cl addition. These results suggest that the critical step inhibited by NH₄Cl occurs within the initial 2-3 hours of infection. We interpret these data as indicating that this is the time frame required for the virus to undergo endocytosis and pH-dependent endosomal processing.

Based on the above experiments, we termed the stages preceding the NH₄Cl-sensitive step, as early/entry stages of EHDV2-Ibaraki infection. To inquire about the dependence of these steps on the functionality of specific endocytic pathways/mechanisms, we employed a selection of inhibitors that target such pathways/mechanisms. To focus on the putative effects of such inhibitors on the early/entry stages of infection and to limit potential toxicity effects elicited by such inhibitors, we performed experiments in which cells were pretreated for 30 minutes with the inhibitors alone, followed by an hour in which cells were concomitantly exposed to virus and inhibitors, after which both inhibitor and virus were washed away before incubation of the cells with NH₄Cl (24 h), to impede any further entry while allowing for expression of viral components that serve as indicators of infection. Repeated experiments (summarized in Supplementary Figure 2 ), with different readouts of infection (e.g., expression of NS3 protein as measured by immunoblotting, or infectious virion production as measured by plaque assay) revealed a lack of effects for (i) cholesterol depletion with methyl-β-cyclodextrin (15 mM), (ii) amiloride (1 mM), (iii) latrunculin (1 μM). To confirm the action of latrunculin, we stained cells with phalloidin at 30 minutes (corresponding to the time point of virus addition) or 90 minutes (corresponding to the time of washout). This revealed marked alterations to the phalloidin staining, in accord with the proposed function of latrunculin ( Supplementary Figure 3 ). These results suggest a lack of dependence on macropinocytosis (or other actin-dependent pathways) or cholesterol-dependent endocytic pathways for EHDV2-Ibaraki entry into MDBK cells. As such, we opted to focus further analysis on inhibitors of clathrin-mediated endocytosis (CME): IKA (Elkin et al., 2016), which functions via an uncharacterized mechanism, and inhibitors of the GTPase activity of dynamin, dynasore (Macia et al., 2006) and dynole (Hill et al., 2009). The rationale for choosing two distinct dynamin inhibitors stems from their different structures. Before analyzing these inhibitors’ effects on EHDV2-Ibaraki entry, we initially assessed their effects on the internalization of fluorescently labeled transferrin, employed here as a CME marker. MDBK cells were pretreated with the inhibitors for 30 min (in serum-free medium), after which the cells were incubated with naïve or inhibitor-containing serum-free medium, supplemented with Alexa-488-labeled transferrin (50 μg/ml, 20 min, 37°C) ( Figure 3 ). Quantification of cell-associated transferrin signal (normalized to DAPI signal, as an indicator of the number of cells per field, Figure 3C ) revealed marked inhibition in the accumulation of transferrin in cells treated with the dynamin inhibitor dynole or 2 μM IKA (> 80% inhibition), suggesting their ability to potently inhibit CME. Partial reductions were induced by 0.5 μM IKA (~ 70%), dynasore (~ 60%) or NH₄Cl (~ 30%). Notably, in the presence of IKA, a transferrin signal along the contours of the cell was observed, in line with its accumulation at the plasma membrane. To focus on internalized transferrin, we probed for the effects of acid wash before fixation and imaging, which revealed that in untreated conditions, most transferrin signals corresponded to internalized transferrin. In contrast, acid wash eliminated the remaining transferrin signal in IKA-treated cells, indicative of a complete block in transferrin internalization ( Supplementary Figure 4 ). To confirm the CME-inhibitory effect of IKA, we tested its ability to inhibit the endocytosis of the Myc-tagged type II TGF-β receptor [Myc-TβRII, shown by use to occur through CME (Ehrlich et al., 2001; Hirschhorn et al., 2012)]. Here, too, 2 μM IKA sufficed to block endocytosis ( Supplementary Figure 5 ).

Having identified conditions for CME inhibition in MDBK cells, we probed for their effect on EHDV2-Ibaraki infection. Here too, with the objective of focusing on potential effects on early/entry stages of infection, cells were pre-treated (or not) for 30 min with the different compounds, followed by co-incubation of virus ± treatments (1 h), after which cells were washed and incubated exclusively with NH₄Cl (for 24 h, Figure 4A ). When assessing NS3 RNA levels by RT-qPCR, the most prominent inhibition was obtained with NH₄Cl pretreated cells, in accord with the block in inhibition observed in Supplementary Figure 1 . IKA (at either 0.5 μM or 2 μM) yielded highly significant levels of inhibition, while treatment with dynasore or dynole exhibited a trend towards inhibition of infection, which fell short of significance ( Figure 4B ). Next, we repeated this experiment, with two additional modes of readout: assessment of NS3 protein levels ( Figures 4C, D ), and evaluation of the production of infectious virions ( Figure 4E ). These experiments confirmed the inhibitory potential of NH₄Cl and IKA while failing to detect significant inhibition of infection by the dynamin inhibitors. Together, these results indicated that inhibition of CME, as observed upon treatment with dynole (as measured by transferrin uptake, Figure 3 ), was insufficient to inhibit EHDV2-Ibaraki entry significantly.

In addition to mediating the initial steps of productive infection of different viruses, the transit through the endocytic/endosomal compartments allows for the sensing of incoming viruses. To assess the sensing of dsRNA, the genetic material of the EHDV2-Ibaraki, and a classical pathogen-associated molecular pattern (PAMP), we opted to measure the level of induction of bovine interferon β (IFN-β) by RT-qPCR. Initially, we stimulated with the dsRNA mimic polyI:C (3 µg⁄ml, 4 h) cells treated or not with NH₄Cl, dynamin inhibitors (dynasore or dynole), or IKA. This revealed a 3-orders of magnitude increase in IFN-β expression upon polyI:C stimulation, a block in such increase upon dynamin inhibition, a partial decrease in the presence of NH₄Cl, and no effects of IKA ( Figures 5A, B ). These results suggest a requirement for dynamin activity of sensing of dsRNA, a partial requirement for endosome acidification and a lack of requirement for processes/phenomena regulated by IKA. Next, we repeated the experiment, employing this time UV-inactivated EHDV2-Ibaraki (UV-EHDV2; MOI 10), which is predicted to retain its potential to induce cellular sensing responses to incoming viruses, while being unable to actively infect cells. Initially, we confirmed that UV irradiation inhibited productive EHDV2-Ibaraki infection, as shown by a complete lack of NS3 expression in cells infected with UV-EHDV2 ( Supplementary Figure 6 ). Subsequently, we treated MDBK cells as in Figure 4 or Supplementary Figure 2 , and stimulated them with UV-EHDV2 (MOI=10, 8h). Here, too, untreated cell stimulation resulted in increased IFN-β expression by more than 3 orders of magnitude ( Figures 5C, D ). NH₄Cl treatment abrogated the virus-mediated increase, indicating the requirement for endosomal pH for the viral sensing. Notably, dynasore inhibited IFN-β upregulation similarly to NH₄Cl, suggesting a strict requirement for functional dynamin-dependent pathways. While MβCD, latrunculin, or amiloride were devoid of inhibitory effects, IKA and dynole significantly inhibited IFN-β expression, albeit less efficiently than dynasore ( Figure 5D ). Together, these results support the notion that dynamin activity is required for dsRNA sensing in both forms of stimuli, polyI:C and inactivated viruses. The greater potency of inhibition observed for NH₄Cl and IKA in the case of UV-EHDV2 (relative to polyI:C) suggests an impediment to the processing of incoming viruses, which may be required for its sensing.

Given that the viral entry process integrates initial events at the plasma membrane with delivery and processing in endosomes, we hypothesized that part of the inhibitory effects of IKA on productive infection or sensing may stem from effects downstream of internalization. This hypothesis was strongly supported by the extensive vacuolation, in cells treated with IKA (1 µM, 20 h, quantified in Supplementary Figure 7 ). Next, we probed for effects of IKA on the intracellular distribution of clathrin light chain A-GFP (LCA-GFP, a marker of clathrin-coated pits and vesicles and clathrin-coated endosomes), Rab5-GFP (a marker of early endosomes) or Rab7-GFP (a marker of late endosomes). While no apparent IKA-induced modifications to the intracellular distribution of either LCA-GFP or Rab5-GFP were observed ( Figure 6A ), a marked portion of Rab7-GFP-expressing cells exhibited localization of this protein to the boundaries of vacuole-like structures upon IKA treatment ( Figure 6A ). This was further confirmed via staining of IKA-treated (1 µM, 20 h) or naïve cells against LAMTOR4, an endogenous marker of late endosomes/lysosomes. This revealed, marked vacuolization in IKA-treated cells ( Figure 6A ). To further characterize the vacuoles, we fed MDBK cells (treated or not with IKA) with lysotracker. This revealed typical perinuclear staining in untreated cells and a lack of accumulation in the IKA-induced vacuoles ( Figure 6B ), suggesting that either the vacuoles are not sufficiently acidic to entrap lysotracker or that they are inaccessible to this agent.

To further probe for the effects of NH₄Cl, dynasore, dynole, or IKA on the entry of different viruses, we assessed VSV or BTV-8 infection of MDBK cells, under analogous conditions to those employed for the above-described experiments with EHDV2-Ibaraki. Multiple studies (Johannsdottir et al., 2009; Cabot et al., 2022; Roche et al., 2008; Cureton et al., 2009, 2010) have established the dependency of VSV entry on clathrin-mediated endocytosis and endosomal acidification. Per the latter requirement, NH₄Cl abrogated VSV infection as measured by the marked reduction in the production of infectious virions at 24 h ( Figure 7A ). Dynasore and dynole resulted in partial but significant inhibition of VSV infection, confirming the role of dynamin-dependent processes (e.g. CME) in VSV entry. Surprisingly, IKA not only failed to inhibit VSV infection/entry but rather enhanced the production of infectious virions in a concentration-dependent manner ( Figure 7A ). This effect was further confirmed through assessment of the percentage of cells (treated or not with IKA) exhibiting GFP expression following infection with a GFP-expressing VSV clone ( Figures 7B, C ). These results suggest either an IKA-mediated enhancement of an alternative entry pathway for VSV or an enhancement of infection due to IKA-mediated alterations to the endosomal compartments. In contrast, infection of MDBK cells with BTV-8, which is expected to exhibit very similar endosomal-processing requirements as EHDV2-Ibaraki, was inhibited by NH₄Cl and IKA ( Supplementary Figure 8 ). Together, our results reinforce the prominence of endosomal processing in regulating the early steps of infection of EHDV2-Ibaraki, BTV-8, and VSV.

As per our results, EHDV2-Ibaraki, VSV, and BTV-8 depend on acidic compartments within the endosomal-lysosomal system of the cell for productive infection of bovine cells, as demonstrated by the robust inhibition in infection observed in cells treated with NH₄Cl before and throughout infection. The requirement for luminal acidic pH for VSV or BTV entry was previously reported (Hyatt et al., 1989; Forzan et al., 2007; Beilstein et al., 2020; Cabot et al., 2022). As measured for EHDV2-Ibaraki, the alkalinization-mediated inhibition of infection was dependent on the time of addition of NH₄Cl, suggesting the pH dependence of early but not late stages of infection. This time dependence of the NH₄Cl allowed us to measure the entry window of EHDV2-Ibaraki, which completed its passage through low pH compartments within 3h, with a half-time of ~ 90 min. The NH₄Cl-mediated block in IFN-β induction, which is expected to occur upon revealing EHDV2-Ibaraki dsRNA, further supports the notion of the alkalinization-mediated interference to viral processing and/or viral degradation. However, NH₄Cl also partially inhibited IFN-β induction by polyI:C. This partial inhibition may reflect the requirement for acidic pH for toll-like receptor 3 to recognize dsRNA and activate downstream signaling (de Bouteiller et al., 2005). A partial reduction in transferrin accumulation in MDBK cells was also induced by NH₄Cl, suggesting that part of its effects may stem from interference with trafficking steps. Indeed, inhibition of endocytosis was observed upon NH₄Cl treatment of mouse tubular cells, possibly reflecting the pH-sensitive recruitment of trafficking regulators to endosomal membranes (Hurtado-Lorenzo et al., 2006; Aniento et al., 1996). However, the differences in the magnitude of the near-complete block in infection as opposed to the minor reduction in transferrin accumulation led us to conclude that the main effect of NH₄Cl reflects the requirement of endosomal pH for viral processing.

Inhibition of dynamin-, actin- or cholesterol-mediated endocytosis with dynasore/dynole, latrunculin, or methyl-β-cyclodextrin, respectively; or macropinocytosis with amiloride; all failed to significantly inhibit EHDV2-Ibaraki infection of MDBK (as measured by expression of viral mRNA, viral protein or production of infectious virions). The inhibition of dynamin is expected to affect the entirety of CME and a subset of the clathrin-independent endocytic pathways that depend on dynamin for vesicle internalization (Antonny et al., 2016; Macia et al., 2006). Accordingly, transferrin uptake ( Figure 3 ) and VSV infection were significantly inhibited by the dynole or dynasore, both of which inhibit the GTPase activity of dynamins (Macia et al., 2006; Hill et al., 2009). Cholesterol depletion is predicted to affect non-clathrin (e.g., caveolin-mediated endocytosis) and at least a subset of clathrin-mediated endocytic events, leading to the membrane retention of ligands that employ multiple entry pathways such as cholera toxin (Massol et al., 2004; Torgersen et al., 2001). Moreover, actin dynamics, which are inhibited by latrunculin (Spector et al., 1983) were proposed to be involved in several entry pathways, including a subset of clathrin-mediated events (Yarar et al., 2005; Boucrot et al., 2006; Cureton et al., 2009), fast-endophilin-mediated-endocytosis [FEME (Watanabe and Boucrot, 2017)] and entry mediated by clathrin and dynamin-independent carriers which form GPI-enriched endocytic compartments [CLIC/GEEC (Sathe et al., 2018)] pathways; in addition to being required for macropinocytosis (Salloum et al., 2023; Mylvaganam et al., 2021). The latter pathway is also predicted to be inhibited by amiloride via the acidification of the cytosol (Koivusalo et al., 2010). Thus, the lack of inhibition of EHDV2-Ibaraki infection by any of these treatments supports the notion that inhibiting any single endocytic pathway or inhibiting a selection of entry pathways does not suffice for blocking EHDV2-Ibaraki entry into MDBK cells. If the amount of virus entering through a specific pathway was constant (i.e., equal in cells treated or not with an inhibitor of a different pathway), one would expect that inhibition of this pathway would result in a proportional reduction in productive viral infection relative to the percentage of viruses entering via this specific pathway. However, this is not observed, as inhibition of specific pathways (or of pathways that depend on common requirements such as actin dynamics, dynamin activity, or membrane cholesterol content) during the window of entry of EHDV2-Ibaraki did not significantly reduce viral infection. Two possible explanations for this phenomenon are that either the virus enters through a pathway insensitive to all treatments employed in this study or cells can compensate for inhibiting certain pathways through the upregulation of others, which EHDV2-Ibaraki may also employ for entry. We favor the latter scenario, as such compensation has been reported [e.g., in the upregulation of clathrin-independent pinocytosis upon inhibition of dynamin (Damke et al., 1995)]. Similar to our results, influenza was shown to employ clathrin-coated vesicles as a non-exclusive mode of entry into cells (Bao et al., 2021; Lakadamyali et al., 2004; Mazel-Sanchez et al., 2023; Rust et al., 2004; Sun and Whittaker, 2013), but inhibition of clathrin-mediated endocytosis via expression of the Δ95-295 mutant of Eps15 or chlorpromazine failed to inhibit influenza infection in HeLa cells (Sieczkarski and Whittaker, 2002). Entry via multiple endocytic pathways may be a common characteristic of viruses that employ oligosaccharides as the element for cell binding and entry, such as influenza (Sun and Whittaker, 2013), as the oligosaccharides may be attached to receptors endowed with different endocytic determinants. Of note, BTV (which exhibits high similarity to EHDV) employs sialic acid to bind and enter cells (Wu and Roy, 2022; Zhang et al., 2010) and was reported to enter either via clathrin or macropinocytosis with dependence on cell type (Forzan et al., 2007; Gold et al., 2010; Stevens et al., 2019).

An apparent discrepancy is revealed when comparing the effects of treatments within the NH₄Cl-sensitive window with dynamin inhibitors or IKA on EHDV2-Ibaraki infection. While both inhibit the CME-dependent accumulation of transferrin (particularly notable for dynole and IKA), only IKA inhibits EHDV2-Ibaraki infection. A plausible explanation for this is that CME inhibition does not suffice to block EHDV2-Ibaraki infection, and IKA’s ability to do so depends on other effects of this compound. Indeed, cells treated with IKA exhibited a prominent accumulation of vacuole-like structures, decorated with Rab7 or LAMTOR4 and which failed to accumulate lysotracker in their interior. This suggests that such structural (and likely function) modification of Rab7/LAMTOR4 compartments may have functional implications for EHDV2-Ibaraki infection. Interestingly, a reverse discrepancy was observed for VSV. Here, dynamin inhibitors significantly reduced infection, while IKA increased infection by an order of magnitude. This lack of inhibition of VSV by IKA suggests that in MDBK cells VSV may employ clathrin-independent carriers for entry and employ distinct endosomal compartments as compared with EHDV2-Ibaraki (or BTV-8). In line with this putative difference in the endosomal compartment employed for entry, the pH requirements for viral entry were reported to differ between VSV [~6.2 pH (Johannsdottir et al., 2009)] and BTV [~5.5 pH (Wu et al., 2019)]. Of note, IKA was shown to affect autophagic flux and induce TFEB signaling (Wang et al., 2017), which may be related to its effects on late endosomes. One feature of IKA that may endow it with the ability to alter endosomes is its strong affinity for sodium ions (Paquette et al., 1990), which are required for pH regulation of endosomes/lysosomes (Steinberg et al., 2010; Chadwick et al., 2021; Scott and Gruenberg, 2011). Notably, conditions that induce similar cell vacuolation, such as treatment with the phosphatidylinositol-3-phosphate 5-kinase inhibitor apilimod or deficient expression of VPS29, also inhibited viral infection (Kreutzberger et al., 2021; Nelson et al., 2017; Poston et al., 2022). The case of VPS29 deficiency is of particular interest as there too the effects were virus-specific, causing inhibition of HCoV-OC43 or SARS-CoV-2, and facilitation of influenza infection (Poston et al., 2022). Thus, while the hampering of endosome functions is an attractive target for antiviral therapies, the generality of such approaches towards different viruses may be restricted by the heterogeneity in endosomal requirements of such viruses.