The Orthobunyavirus Germiston Enters Host Cells from Late Endosomes

ABSTRACT

With more than 80 members worldwide, the Orthobunyavirus genus in the Peribunyaviridae family is a large genus of enveloped RNA viruses, many of which are emerging pathogens in humans and livestock. How orthobunyaviruses (OBVs) penetrate and infect mammalian host cells remains poorly characterized. Here, we investigated the entry mechanisms of the OBV Germiston (GERV). Viral particles were visualized by cryo-electron microscopy and appeared roughly spherical with an average diameter of 98 nm. Labeling of the virus with fluorescent dyes did not adversely affect its infectivity and allowed the monitoring of single particles in fixed and live cells. Using this approach, we found that endocytic internalization of bound viruses was asynchronous and occurred within 30 to 40 min. The virus entered Rab5a-positive (Rab5a⁺) early endosomes and, subsequently, late endosomal vacuoles containing Rab7a but not LAMP-1. Infectious entry did not require proteolytic cleavage, and endosomal acidification was sufficient and necessary for viral fusion. Acid-activated penetration began 15 to 25 min after initiation of virus internalization and relied on maturation of early endosomes to late endosomes. The optimal pH for viral membrane fusion was slightly below 6.0, and penetration was hampered when the potassium influx was abolished. Overall, our study provides real-time visualization of GERV entry into host cells and demonstrates the importance of late endosomal maturation in facilitating OBV penetration.

IMPORTANCE Orthobunyaviruses (OBVs), which include La Crosse, Oropouche, and Schmallenberg viruses, represent a growing threat to humans and domestic animals worldwide. Ideally, preventing OBV spread requires approaches that target early stages of infection, i.e., virus entry. However, little is known about the molecular and cellular mechanisms by which OBVs enter and infect host cells. Here, we developed accurate, sensitive tools and assays to investigate the penetration process of GERV. Our data emphasize the central role of late endosomal maturation in GERV entry, providing a comprehensive overview of the early stages of an OBV infection. Our study also brings a complete toolbox of innovative methods to study each step of the OBV entry program in fixed and living cells, from virus binding and endocytosis to fusion and penetration. The information gained herein lays the foundation for the development of antiviral strategies aiming to block OBV entry.

INTRODUCTION

Orthobunyavirus is a large genus of RNA viruses comprising more than 80 isolates distributed worldwide that, along with the genera Herbevirus, Pacuvirus, and Shangavirus, constitute the family Peribunyaviridae in the order Bunyavirales (1). Orthobunyaviruses (OBVs) are transmitted by arthropod vectors, most by mosquitoes and midges but a few by ticks and bed bugs, and, consequently, belong to the group of arthropod-borne viruses (2). These viruses constitute a global threat to human and veterinary public health. Many cause severe diseases in humans, ranging from acute but self-limiting febrile illnesses (e.g., Oropouche virus [OROV] in South America) to neurological disorders (e.g., La Crosse virus [LACV] and California encephalitis virus [CEV] in North America) (3). In livestock, OBV infections often result in abortion, congenital malformations in offspring, and stillbirth. A good example is Schmallenberg virus (SBV), which emerged in Northern Europe in the early 2010s and is now distributed throughout the European continent (4, 5). Due to their mode of transmission, global warming, and the spread of arthropod vectors to new regions, OBVs are considered potential agents of emerging diseases. Recent examples are Cristoli and Umbre viruses, two OBVs identified in France that cause lethal encephalitis in immunocompromised patients (6, 7). No vaccine or treatment for OBV infection is currently approved for human use.

Similar to those of other peribunyaviruses, OBV virions are enveloped and roughly spherical, with a diameter of approximately 80 to 120 nm (8). The viral genome consists of three single-stranded negative-sense RNA segments present as pseudohelical ribonucleoproteins (RNPs) in the viral particles (3). The RNA segments exclusively replicate in the cytosol and encode up to two nonstructural proteins and four structural proteins, namely, the RNA-dependent RNA polymerase L, the nucleoprotein N, and the two transmembrane glycoproteins Gn and Gc. The N protein binds to genomic RNA to form, together with the viral polymerase, RNPs. In the envelope, the glycoproteins form spike-like projections of up to 20 nm arranged in a nonicosahedral lattice, with surface glycoprotein protrusions exhibiting a unique tripod-like arrangement (8). Gn and Gc are thought to be responsible for the attachment of viral particles to the surface of host cells and the subsequent penetration into the cytosol (9). A recent structural study on several OBVs showed that the fusion machinery underlying the OBV spike likely consists of the C-terminal half of Gc together with Gn, which connects three adjacent spikes on the viral particles (10).

The receptors, cellular factors, and entry pathways used by OBVs to infect host cells remain largely uncharacterized. Sulfation and heparan sulfates have been found to be involved in the binding of SBV and Akabane virus to the cell surface (11, 12). The human C-type lectin dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) has been reported to promote LACV infection (13). Other C-type lectins, namely, Mincle, Dectin-1, and Dectin-2, have been shown to facilitate LACV binding but not productive infection (14). Inhibitor studies have indicated that many OBVs rely on endosomal acidification for infection (9, 15–18). The functional disruption of dynamin-2 and clathrin implies a role of clathrin-mediated endocytosis in OBV internalization (13, 15, 16, 19). In addition, LACV and OROV have been proposed to enter early endosomes (EEs) after uptake (16, 19). Considering these findings collectively, it is apparent that OBVs make use of the endocytic machinery to enter cells, but the exact pathway involved and the identity of the endosomal vacuoles from which OBVs enter the cytosol remain elusive (1), which is precisely the aim of our study.

We focus here on Germiston virus (GERV), which was first isolated from Culex mosquitoes in South Africa in the 1950s (20). To date, GERV has been associated with the infection of two laboratory workers (20). The virus is, however, classified as a biosafety level 2 (BSL2) pathogen in Europe and thus can be studied by approaches such as live cell imaging that are nearly impossible for the study of OBVs of higher biosafety classification. To analyze GERV entry into cultured mammalian cells, we developed sensitive and quantitative tools and assays that allowed us to determine general properties of the virus entry process, such as binding, internalization, membrane trafficking, acid-activated fusion, and penetration. The results showed that GERV penetrates host cells via acid-activated membrane fusion early in the degradative pathway of the endocytic machinery.

RESULTS

GERV life cycle in A549 cells.

Previous studies have shown that GERV can infect various cell types originating from different species (21). To investigate GERV infectious entry, we selected the human epithelial cell line A549, as it represents the lungs, a tissue that is affected during GERV-induced febrile diseases (20). We also selected the human hepatocytic cell line Huh-7 and the glioblastoma, astrocytoma cell line U-87, as they represent liver and brain tissues, respectively, for which OBVs generally show significant tropism (3). We first determined the timing for a single round of infection in the selected cell lines. The susceptibility of cells to GERV infection was assessed by flow cytometric analysis after immunostaining with the anti-GERV antibody GR1 (Fig. 1A). GR1 is a polyclonal antibody raised against purified, lysed GERV particles in guinea pigs that recognizes the viral proteins N, Gn, and Gc (22). Approximately 20% of cells infected at a multiplicity of infection (MOI) of 5 were positive for the GERV structural proteins at 8 h postinfection (hpi), regardless of the infected cell line (Fig. 1B; Fig. S1A and B in the supplemental material). The fluorescence signal increased over time and plateaued value at 24 hpi, indicating that the signal detected in this assay corresponded to viral replication and not to the input virus.

FIG 1.

Quantification of GERV infection in A549 cells. (A) A549 cells were exposed to GERV at an MOI of 5 and harvested 8 h later. After fixation and permeabilization, infected cells were stained with the polyclonal guinea pig antibody GR1, which recognizes the GERV structural proteins N, Gn, and Gc. Infection was then analyzed by flow cytometry. (B) A549 cells were infected with GERV at the indicated MOIs, and infection was monitored over 30 h using the flow cytometry-based assay described in panel A. (C) Supernatant harvested from A549 cells at 30 hpi was assessed for progeny virion production by a PFU assay in BHK-21 cells. (D) A549 cells were exposed to GERV, and the supernatant from infected cells was collected at time points up to 30 hpi and analyzed with the PFU assay described in panel C. hpi, hours postinfection.

To examine the release of infectious viral particles, cells were infected at MOIs of up to 5, and virus production was quantified up to 30 hpi by a PFU assay. After crystal violet staining, plaques were clearly visible (Fig. 1C). Infectious progeny virions were released from cells infected at an MOI of 0.2 from 8 hpi (Fig. 1D; Fig. S1C and D). Collectively, our results showed that GERV completes one round of infection comprising virus binding and penetration to replication and egress of infectious particles within 8 to 10 hpi in all three cell lines. In subsequent experiments, as we aimed to characterize the process of GERV entry, we used an MOI of 5 and limited our assays to 8 hpi, which led to a cell infection rate of approximately 20%. This range of infection allowed the detection of potential inhibitory or enhancing effects of perturbants.

GERV particles.

In this study, large-scale preparations of GERV were produced and pelleted by ultracentrifugation through a 30% sucrose cushion. Viral particles were subjected to SDS-PAGE under nonreducing conditions followed by Coomassie blue staining. Bands were observed at 28, 35, and 98 kDa corresponding to the nucleoprotein N and the two envelope glycoproteins Gn and Gc, respectively (Fig. 2A). The purity of the GERV stocks was estimated to be greater than 90% based on the densities of Gn, Gc, and N divided by the total density of the lane. Using the polyclonal anti-GERV antibody GR1, we confirmed the identity of all three major structural proteins by Western blotting (Fig. 2B). The viral particles were visualized by cryo-EM after fixation with paraformaldehyde and vitrification (Fig. 2C). Cryo-electron micrographs showed roughly spherical viral particles with spike-like projections of 11 ± 3 nm (n = 28) and a roundness coefficient, i.e., the ratio between the perpendicular length and width, of approximately 0.9 ± 0.1 (n = 100) (Fig. 2D). The observed roundness coefficient was close to 1, reflecting the nearly spherical shape of the particles. The diameter of the GERV particles was approximately 98 ± 11 nm (n = 100) (Fig. 2E).

FIG 2.

Structural organization of GERV particles. (A) GERV was pelleted through a 30% sucrose cushion and was then purified through a 15 to 60% sucrose gradient by ultracentrifugation. Proteins of purified viral particles were subsequently separated by nonreducing SDS-PAGE and stained with Coomassie blue. (B) GERV particles were analyzed by SDS-PAGE and Western blotting under nonreducing conditions. Viral proteins were visualized by using the GR1 antibody. (C) Purified GERV particles were fixed with paraformaldehyde, vitrified, and analyzed by cryo-EM. The top panel shows an example of a cryo-electron micrograph of GERV particles, and the bottom panel shows an enlarged image of one particle. White arrowheads indicate spike-like projections. Scale bar, 50 nm. (D) The perpendicular length and width of GERV particles were measured. The roundness coefficient (the ratio of the width to the length) was then calculated for each particle (n = 100). (E) The diameter distribution of GERV particles (n = 100) was analyzed, with virion diameter defined as the mean between length and width.

Labeling of GERV with fluorescent dyes.

To quantify and visualize the early steps of GERV entry, we purified virus particles labeled with fluorescent probes following a procedure previously established in our laboratory for other Bunyavirales members (23). To render the viral particles fluorescent, we exploited the numerous lysine residues in the glycoproteins Gn and Gc, to which we conjugated hydroxysuccinimidyl (NHS) ester dyes with excitation wavelengths of 647 nm (Atto 647N) and 488 nm (Atto 488) at a dye/glycoprotein ratio of 3:1. At this ratio, we assumed that all viral particles were labeled. Analysis by SDS-PAGE and Coomassie blue staining showed that the fluorescently labeled GERV stocks were pure (Fig. 3A). The only fluorescently labeled proteins in the Atto 647N- and Atto 488-conjugated GERV preparations (Atto 647N- and Atto 488-GERV, respectively) were Gn and Gc (Fig. 3B). Gn labeling appeared to be slightly more effective with the Atto 647N dye. The observation that the nucleoprotein N was not labeled indicated that the viral envelope was intact and, thus, that the viral particles remained intact during the labeling procedure. The labeled particles could be visualized as single spots by confocal fluorescence microscopy (Fig. 3C). Alternatively, GERV was labeled with the lipid dye R18 for the analysis of viral fusion shown in the following sections. A high concentration of R18 molecules in the envelopes of labeled viral particles results in autoquenching of the fluorescence signal (24). Viral fusion allows the release of R18 in the target cellular membrane, resulting in dilution and dequenching of the fluorescence. Overall, we found that labeling with Atto and R18 dyes did not significantly impact GERV infectivity. In the PFU assay, the titers of labeled particles were similar to those of nonlabeled particles (Fig. 3D).

FIG 3.

Labeling of GERV with Atto fluorescent dyes. (A and B) GERV was labeled with either a Atto 488 or Atto 647N dye at a 1:3 molar ratio of GERV glycoproteins to dye. Both unlabeled and purified fluorescently labeled viral particles were analyzed by nonreducing SDS-PAGE and Coomassie blue staining (A) or fluorography (B). Note that the GERV glycoproteins Gn and Gc were the only fluorescently labeled proteins. (C) Confocal imaging of Atto 488-GERV showed individual fluorescent spots. Scale bar, 2 µm. (D) Fluorescently labeled GERV particles were analyzed by a PFU assay, and the titers were normalized to the amount of the glycoprotein Gc. (E) Increasing amounts of Atto 488-GERV were bound to A549 cells for 1 h on ice before fixation and analysis by flow cytometry. Virus binding is expressed as the relative fluorescence intensity associated with the cells, as measured by flow cytometry. (F) Atto 488-GERV was bound to A549 cells at an MOI of 10 and imaged with a confocal microscope. Nuclei were stained with DAPI. The top panel shows one focal plane with viral particles in green and nuclei in blue, while the bottom panel shows cell-associated virus particles (white spots) seen in maximum z-projection images acquired in the 488-nm channel. Scale bar, 5 µm.

GERV binding to cells.

To analyze binding to cells, increasing amounts of Atto 488-GERV were allowed to bind to A549 cells on ice for 30 min. Unbound viral particles were removed by washing, and the remaining cell-associated fluorescence was evaluated by flow cytometry. This assay allowed the detection of Atto 488-GERV binding at MOIs of 1 and higher (Fig. 3E). Total binding increased with increasing concentrations of input GERV. After binding of Atto 488-GERV to A549 cells on ice, viral particles were observed on the cell surface by confocal microscopy (Fig. 3F). The number of spots per cell was quantified and found to be greater than 214 ± 79 (n = 3) at an MOI of 10. The fact that the spots had different sizes suggests that not only individual viral particles were bound to cells. Larger clusters may be formed by multiple virions. The results indicated that the ratio of infectious to noninfectious particles was, however, likely high.

GERV traffics through EEs.

We next sought to determine whether GERV is internalized into EEs after binding of viral particles to the cell surface, as previously reported for LACV and OROV (16, 19). To monitor virus endocytosis, we first exposed A549 cells to Atto 488-GERV, washed them to remove unbound particles, and rapidly warmed the samples to allow virus internalization. To distinguish between internalized and surface-bound particles, cells were treated with trypan blue before flow cytometric analysis according to a standard procedure (24, 25). Trypan blue is membrane impermeable and thus quenches only the fluorescence emitted by surface-exposed Atto 488-GERV, whereas fluorescence from intracellular virus remains unquenched (Fig. 4A). Time-course analysis of the generation of trypan blue-resistant fluorescence of cell-associated Atto 488-GERV revealed that internalization into all three cell lines started after a 5 to 10 min lag and attained the half-maximal level (t_1/2) within 20 to 25 min (n = 3) (Fig. 4B). Evidently, GERV uptake occurred asynchronously over a span of 5 to 40 min in all cell lines.

FIG 4.

GERV enters host cells through endocytic internalization and is sorted into Rab5a⁺ EEs. (A) Atto 488-GERV was bound to A549 cells at an MOI of 10 on ice before warming to 37°C for 30 min. Cells were then fixed and treated with trypan blue before flow cytometric analysis. Trypan blue treatment allowed specific quenching of the fluorescence of cell surface-bound viruses. RU, relative unit. (B) A549, Huh-7, and U87 cells were exposed to Atto 488-GERV on ice and rapidly shifted to 37°C to allow virus internalization for up to 60 min. Endocytic uptake of viral particles was analyzed by flow cytometry after cell fixation and trypan blue treatment. Internalization is expressed as the percentage of fluorescence measured in samples treated with trypan blue compared to that in untreated samples. The fluorescence signal observed for cells not exposed to Atto 488-GERV was considered the background signal and was subtracted from the other values. (C) Atto 647N-GERV at an MOI of 30 was bound to A549 cells expressing EGFP-Rab5a for 30 min on ice. Cells were then washed and rapidly warmed for 30 min. GERV (red) and Rab5a (green) were imaged by confocal microscopy. One focal plane is shown, and higher-magnification images of viral particles within Rab5a⁺ endosomes are displayed on the right. Scale bar, 10 µm. (D) Atto 647N-GERV intracellular trafficking was monitored over 60 min as shown in panel C. Colocalization is expressed as the percentage of GERV colocalized with Rab5a⁺ vesicles at different times postwarming. A minimum of 5 cells per time point were analyzed. (E) Same experiments described in panel C, but the cells were kept alive and imaged with a spinning disc confocal microscope after warming. The time-lapse series shows the coordinated motion of Atto 647N-GERV (red) with Rab5a⁺ (green) endosomal vacuoles. The white arrow shows a coordinated motion of a GERV particle with a Rab5a⁺ vesicle. The timing is indicated (min/s) at the top right of each image. The corresponding movie (Movie S1) is available online as supplemental data. Scale bar, 2 µm. (F) EGFP-Rab5a WT and S34N (DN mutant) were transiently expressed in A549 cells. The cells were then infected with GERV at an MOI of ∼12. Cell populations were selected by flow cytometry to isolate populations expressing 1-log incremental levels of EGFP at 8 hpi (low-medium-high), and infection was quantified within each population as described in Fig. 1A. Data were normalized to those in cell populations expressing EGFP-Rab5a WT.

To define the endocytic route taken by GERV, we next determined whether uptake results in sorting of the virus in the EEs. To this end, A549 cells were transduced with the LX307 lentiviral vector encoding Rab5 tagged with a monomeric version of enhanced green fluorescent protein (EGFP) (EGFP-Rab5A) (26). Rab5 is a small GTPase required for the trafficking and maturation of EEs (27). Cells were then selected for moderate, homogeneous expression by fluorescence-activated cell sorting and maintained in culture under selection pressure with puromycin. After binding of Atto 647N-GERV (MOI of ∼30) to cells on ice, the temperature was rapidly shifted to 37°C for up to 60 min. Confocal microscopy showed GERV in cytoplasmic vesicles positive (+) for EGFP-Rab5a (Fig. 4C), and colocalization peaked 30 min postwarming (Fig. 4D). Consistent with this finding, coordinated motion of Atto 647N-GERV with EGFP-Rab5a⁺ EEs was observed in live cells beginning at 15 min after warming, i.e., approximately 10 min after initiation of virus internalization into A549 cells (Movie S1; Fig. 4B and E).

To determine whether passage through EEs is required for infectivity, we assessed GERV infection in A549 cells expressing a dominant negative (DN) mutant of Rab5a, S34N. Expression of this mutant abrogates the maturation of newly formed EEs (26). In this assay, A549 cells were transfected with DNA plasmids encoding the wild-type (WT) and DN forms of Rab5a tagged with EGFP (EGFP-Rab5a WT and EGFP-Rab5a S34N, respectively). Transfected cells exhibiting different levels of EGFP expression were selected and analyzed for infection. Expression of the Rab5 mutant resulted in a decrease in infection of up to 50% (Fig. 4F). Taken together, these results indicated that productive GERV infection involves passage through Rab5a⁺ endosomal vesicles.

GERV enters late endosomal compartments.

We next sought to determine whether the transport of viral particles to downstream organelles is required for infectious GERV entry. For this purpose, A549 cells were transduced with the LX307 lentiviral vector and selected for homogeneous expression of EGFP as described above, but instead of encoding Rab5, the vector encoded either the small GTPase Rab7a or LAMP-1, both of which were tagged with EGFP (EGFP-Rab7a and LAMP-1-EGFP) (26). Rab7 is a key player in late endosome (LE) maturation and function, and LAMP-1 is localized mainly within endolysosomes (27). A549 cells expressing EGFP-Rab7a and LAMP-1-EGFP were exposed to Atto 647N-GERV for 30 min on ice and were then rapidly warmed to allow endocytosis and virus internalization. GERV could be observed within Rab7a⁺ intracellular vesicles (Fig. 5A). In live cells, spinning disc confocal microscopy showed that Atto 647N-GERV moved in the cytoplasm with EGFP-Rab7a⁺ vesicles approximately 25 to 30 min postwarming (Fig. 5B and Movie S2). After 60 min, 39% ± 16% (n = 100) of viral particles were associated with Rab7⁺ vesicles (Fig. 5C). Colocalization of GERV with endosomal vacuoles containing LAMP-1 was also observed (Fig. 5D). However, the number of these events was minimal (Fig. 5E), and the fluorescence signal appeared indistinct and diffuse, not as expected for intact viral particles, as shown in Fig. 4C and E and Fig. 5A and B. These observations suggested that the virus had already fused with the limiting membrane at these late time points. Overall, our results indicated that after binding to the cell surface, GERV enters cells via endocytosis and continues its journey along the degradative pathway until it reaches Rab7a⁺ LEs.

FIG 5.

GERV enters Rab7⁺ late endosomal organelles. (A) Atto 647N-GERV was bound to A549 cells expressing EGFP-Rab7a at an MOI of 30 for 30 min on ice. Cells were then washed and rapidly warmed for 60 min. GERV (red) and Rab7a (green) were imaged by confocal microscopy. One focal plane is shown, and higher-magnification images of viral particles within Rab7a⁺ endosomes are displayed on the right. Scale bar, 10 µm. (B) Atto 647N-GERV intracellular trafficking was monitored in A549 cells by live confocal imaging over 60 min after warming. The time-lapse series shows the coordinated motion of Atto 647N-GERV (red) with Rab7a⁺ (green) endosomal vacuoles. The white arrow shows coordinated motion between a GERV particle and a Rab7a⁺ vesicle. The timing is indicated (min/s) at the top left of each image. The corresponding movie (Movie S2) is available online as supplemental data. Scale bar, 1 µm. (C) Atto 647N-GERV intracellular trafficking was monitored in A549 cells over 120 min as shown in panel A. Colocalization is expressed as the percentage of GERV colocalized with Rab7a⁺ vesicles at different times postwarming. A minimum of 5 cells per time point were analyzed. (D) Same experiment described in panel A but with A549 cells expressing LAMP-1-EGFP instead of EGFP-Rab7a and imaged 60 min postwarming. Scale bar, 10 µm. (E) Same experiment described in panel C but showing the percentage of GERV colocalized with LAMP-1⁺ vesicles during virus intracellular trafficking. (F) EGFP-Rab7a WT and T22N (DN) were transiently expressed in A549 and U87 cells. The cells were then infected with GERV at an MOI of ∼12. Cell populations were selected by flow cytometry to isolate populations expressing high levels of EGFP at 8 hpi, and infection was quantified in the selected population as described in Fig. 1A. Data were normalized to those in the cell population expressing EGFP-Rab7a WT.

GERV infection depends on intact LEs.

To further investigate whether GERV requires late endosomal compartments for productive infection, we exploited the small GTPase Rab7a. To examine the role of Rab7a in GERV penetration, we infected A549 and U87 cells transiently expressing an EGFP-tagged DN mutant of the Rab7a small GTPase, T22N (26). Similar to the procedure for Rab5a DN mutant-expressing cells, cells were selected for different levels of EGFP expression and subsequently monitored for infection. High expression of EGFP-Rab7a T22N hampered GERV infection in both cell lines (Fig. 5F), indicating that GERV infectious entry relies on functional LEs in A549 and U87 cells. Infection of Huh-7 cells could not be analyzed with this approach because too few cells survived expression of the Rab7 DN mutant. Taken together, these results confirmed that GERV depends on late endosomal maturation for infectious entry but probably does not need to reach the most downstream compartments of the degradative endocytic machinery.

Proteolytic activation is not required for GERV infectious entry.

Host cell proteases are sometimes required for virus activation and fusion (28). For example, serine proteases have been proposed to prime and activate infection by LACV (13). These proteases include furin in producer cells and trypsin-like proteases generally localized at or near the plasma membrane in target cells. To evaluate the role of endogenous proteases in the mechanisms of GERV entry, we first assessed infection in the continuous presence of broad-spectrum inhibitors of serine proteases, i.e., aprotinin and 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF). Pretreatment of cells with the highest concentration of aprotinin not toxic to cells, i.e., 20 µM, reduced GERV infection by 30% (Fig. 6A). Inhibition of infection was also minimal when AEBSF was used instead of aprotinin (Fig. 6B). For comparison, coronavirus activation and infection rely on serine proteases in target cells, and SARS-CoV-2 infection decreases by 90% when cells are pretreated with 5 µM aprotinin (29).

FIG 6.

GERV infectious entry does not rely on host cell proteases. (A and B) Cells were pretreated for 30 min with the indicated concentrations of aprotinin (A) and AEBSF (B), both of which are broad-spectrum inhibitors of serine proteases. Cells were then infected with GERV at an MOI of 5 in the continuous presence of the drugs, and infection was then quantified by flow cytometry as described in Fig. 1A. The data were normalized to those in samples without inhibitor treatment. (C and D) GERV (MOI of ∼20) was subjected to pretreatment with furin (C) and trypsin (D) at the indicated concentrations for 15 min at 37°C prior to infection of A549 cells. Infected cells were analyzed by flow cytometry as described in Fig. 1A. The data were normalized to those in samples not pretreated with proteases. (E and F) Cells were pretreated with increasing concentrations of SB412515 (E) and CA074Me (F), both of which are inhibitors of endolysosomal cathepsins. Cells were infected with GERV at an MOI of 5 in the continuous presence of the drugs, and infection was then measured by flow cytometry. The data were normalized to those in samples without inhibitor treatment.

In a complementary approach, we assessed whether increased proteolytic processing of the glycoproteins Gn and Gc results in enhanced GERV activation and infection. To this end, viral particles were exposed to exogenous furin and trypsin and were then added to A549 cells. Using our flow cytometry-based infection assay, we found that none of the pretreatments impacted infection (Fig. 6C and D). Notably, the protease inhibitors and all other drugs used in this study were evaluated over a range of concentrations for which no cytotoxicity was detected, as shown by a quantitative assay measuring the release of lactate dehydrogenase into the extracellular medium upon cell death and lysis (Fig. S2).

Our results indicated that GERV reaches LEs and relies on the degradative endocytic machinery for infectious entry. Therefore, we also examined whether the activity of endolysosomal proteases is involved in the infectious entry of GERV. For this purpose, cells were infected in the presence of increasing amounts of the inhibitors SB412515 and CA074Me, which block a broad spectrum of endolysosomal cathepsins. None of these drugs affected GERV infection at any concentration (Fig. 6E and F). Collectively, these data suggested that proteolytic processing of GERV glycoproteins is already complete after release of the virus from producer cells or is not needed for productive entry into target cells.

GERV relies on vacuolar acidification for infection.

Previous studies have shown that OBV infection is sensitive to agents that neutralize vacuolar pH (15, 16). To assess whether endosomal acidification is important for the infectious entry of GERV, the virus was added to A549 cells in the presence of increasing amounts of chloroquine and ammonium chloride (NH₄Cl), two weak lysosomotropic bases (Fig. 7A and B). Both agents induced dose-dependent inhibition of GERV infection in A549 cells. The dependence of GERV on vacuolar acidification for productive entry was confirmed by treatment with two inhibitors of vacuolar-type proton (H⁺)-ATPases, i.e., bafilomycin A1 and concanamycin B. Incubation of cells with increasing amounts of these two drugs resulted in dose-dependent inhibition of GERV infection (Fig. 7C and D). Collectively, these results demonstrated that GERV infection depends on endosomal acidification.

FIG 7.

GERV relies on acid-activated membrane fusion for infection. (A to D) Cells were pretreated with endosomal pH-disrupting drugs at the indicated concentrations and were then infected with GERV at an MOI of 5 in the continuous presence of chloroquine (A), NH₄Cl (B), bafilomycin A1 (C), or concanamycin B (D). Infection was quantified by flow cytometry as described in Fig. 1A, and the data were normalized to those in control samples without inhibitor treatment. (E) GERV was bound at an MOI of 5 to confluent monolayers of A549 cells for 1 h on ice. Cells were then washed and treated at the indicated pH values at 37°C for 1.5 min. Infected cells were subsequently incubated overnight in the presence of NH₄Cl (50 mM) at pH ∼7.4 to block virus penetration from endosomes. Consequently, the release of viral genomes only from the plasma membrane was monitored in this assay. Infection was analyzed by flow cytometry and expressed as the percentage of infected cells relative to total cells. (F) R18-labeled GERV was bound at an MOI of 10 to A549 cells on ice before warming to 37°C for 1 h. The increase in the fluorescence signal corresponded to dequenching of the lipid dye R18 after virus fusion with endosomes in living cells and was measured with a fluorimeter. NH₄Cl was used to block virus fusion by raising the endosomal pH and, hence, to determine the fluorescence background due to spontaneous translocation of the R18 molecules between the viral envelope and the adjacent cell membrane. RU, relative unit. (G and H) Cells were pretreated with KCl (G) or TEA (H), a broad-spectrum K⁺ channel blocker. Cells were infected with GERV at an MOI of 5 in the continuous presence of the perturbants. The percentage of infected cells was quantified by flow cytometry as described in Fig. 1A, and the data were normalized to those in samples not treated with the perturbants.

Low pH is sufficient and necessary for GERV fusion.

To determine the pH threshold, we assessed the capacity of GERV to fuse to the plasma membrane and thus bypass the need for endocytosis during productive infection, as previously described for unrelated viruses (26). In brief, GERV at an MOI of 5 was allowed to bind to A549 cells on ice, and the temperature was then rapidly shifted to 37°C for 1.5 min in buffers at the indicated pH values. NH₄Cl-containing medium at neutral pH was then added for the remaining period of infection to prevent infection via endosomal vesicles. As shown in Fig. 7E, bypass resulted in efficient infection and occurred at pH values of 6.0 and below, with the half-maximal infection occurring at pH values slightly below 6.0. Overall, these results demonstrated that a reduction in pH is sufficient to trigger fusion.

To further investigate the timing of viral fusion, we used an approach consisting of monitoring viral fusion in living cells. This approach relied on autoquenching of the R18 fluorescent lipid dye (24). For this assay, R18-GERV was bound to the three cell lines on ice, and virus internalization was synchronized by switching the cells rapidly to 37°C. Live monitoring of infected cells with a fluorimeter revealed that the t_1/2 of fluorescence emission occurred 35 to 40 min after warming and plateaued value approximately 40 to 80 min later (Fig. 7F; Fig. S3A and B). This timing was compatible with penetration of GERV from late endosomal compartments, and the time span observed for fusion was most likely due to asynchronous internalization of the virus (Fig. 4B). Collectively, these results established the link between endosomal acidification, viral fusion, and productive entry.

GERV penetration requires K⁺ influx.

Recent reports have suggested that in addition to H⁺ pumps, K⁺ channels play a role in the early steps of LACV and Bunyamwera virus (BUNV) infection (30, 31). Therefore, we tested whether cellular K⁺ is required for GERV infection. A549 cells were pretreated with concentrations of extracellular potassium chloride (KCl) above 20 mM to collapse K⁺ gradients and inactivate K⁺ channels. We observed a decrease in infection of 70% at 100 mM (Fig. 7G). To further investigate the role of K⁺ in OBV infectious entry, we next pretreated the three cell lines with tetraethylammonium (TEA), a broad-spectrum K⁺ channel blocker, prior to infection with GERV in the continuous presence of the inhibitor. We observed that TEA reduced GERV infection in a dose-dependent manner, with a decrease in infection of 60 to 70% at 50 mM (Fig. 7H). Taken together, these results suggested that both the switch from sodium (Na⁺) to K⁺ in maturing endosomes and the decrease in pH are needed for GERV penetration.

Collectively, our data show that GERV resembles other late-penetrating viruses (L-PVs), a large group of viruses with the common characteristic of reliance on late endosomal maturation for infection (32). GERV depends on intact LEs for infectious entry. It is transported from the plasma membrane to LEs through EEs, and its optimal pH for fusion and requirement for ions other than H⁺ correspond to these conditions prevailing in nascent LEs. GERV, however, differs from other OBVs such as LACV in that its activation and fusion do not rely on proteolytic processing.

DISCUSSION

OBVs are emerging and reemerging pathogens constituting a global threat to both human and veterinary public health. The OBV life cycle, however, remains understudied, and these viruses can be considered neglected. With this study, we developed accurate and reliable assays to examine GERV infection, endocytosis, and membrane fusion in human host cells. To this end, we applied cryo-electron microscopy (cryo-EM) to visualize viral particles. To fluorescently label virions and track GERV during each step of its entry process using flow cytometry- and fluorescence microscopy-based approaches, we used amine-reactive dyes and exploited the free amine residues in the GERV glycoproteins Gn and Gc. Alternatively, to analyze acid-activated membrane fusion, we relied on autoquenching of the fluorescence signal from the lipid dye R18.

Our cryo-electron micrographs showed that GERV particles are roughly spherical, with an average diameter of 98 nm and spike-like projections of 11 nm. Similar observations have been made for other OBVs such as LACV and BUNV (8, 33, 34). Tomographic data showed that BUNV forms nonicosahedral virions with the spikes in a unique tripod-like arrangement (8). Our results clearly indicated that GERV, although globally round, shares with LACV and BUNV a large variation in particle size, between 75 and 135 nm. However, further structural studies are required to determine whether GERV glycoproteins have a tripodal organization as well as to understand the reasons for the variation in particle size and the impact on the infectivity of OBVs.

Binding of GERV to the cell surface was readily detectable at a low MOI by flow cytometry. Our confocal images showed that the number of bound virions per cell was larger than the MOI, indicating that virus stocks prepared in BHK-21 cells contained more noninfectious than infectious viral particles. However, the infectious/noninfectious viral particle ratio was rather high compared to other Bunyavirales members, e.g., this ratio is lower than 1:1,000 for Uukuniemi virus (UUKV), a member of the Phenuiviridae family in the order Bunyavirales evolutionarily close to Peribunyaviridae and OBVs (26). Interestingly, this ratio increased significantly for UUKV when it was amplified in arthropod cells rather than mammalian cells (35). In our study, GERV stocks were prepared in BHK-21 cells. More structure-function studies are needed to determine whether GERV and other OBVs gain distinct molecular features in insect vectors to enhance their infectivity in mammalian host cells.

Our results clearly indicated that after binding, GERV relies on endocytosis and membrane transport through the endosomal machinery for infection, regardless of the infected cell line. We observed GERV in Rab5a⁺ endosomal vacuoles, and expression of Rab5a S34N, which hampers EE maturation, affected GERV infection. Thus, GERV can be concluded to pass through Rab5a⁺ EEs. Although efficient, with internalization of approximately 50 to 60% of the surface-bound particles, GERV uptake was asynchronous and completed within 40 min. For comparison, UUKV completes this step within 10 min (26). The reason that GERV endocytosis is slow is unclear. However, similar observations have been made for unrelated viruses such as human papillomavirus (36), which travels on the cell surface for several hours before internalization. Whether GERV depends on a similar process for sorting into the endosomal system remains to be examined.

Penetration of enveloped viruses implies fusion between the viral envelope and a cell membrane. Fusion is often triggered in endosomes upon acid activation of viral glycoproteins. The first indication that GERV uses this strategy was the sensitivity of infection to agents that neutralize endosomal pH, as is typical for several other OBVs (15, 16, 18). Another indication was the capacity of cell-bound viruses to fuse to the plasma membrane after exposure to a pH of 6.0 or below, with an optimum pH of 5.5. Other studies showed that cell-cell fusion occurs at similar pH values when the Gn and Gc glycoproteins of LACV and CEV are forced to the cell surface through overexpression (17, 37–39). These pH values are relatively low and comparable to the pH in nascent multivesicular bodies (MVBs), which are LE precursors (40). Taken together, these observations suggest that GERV shares with many other OBVs the need to reach acidic organelles, most likely late endosomal compartments, for infection.

The acid-induced penetration of GERV at the cell surface indicated that a low pH is sufficient to trigger fusion. Proteolytic processing in endosomes, as observed for LACV (13) and the unrelated Ebola virus (41) and coronaviruses (42), was apparently not needed. Acid activation occurred in less than 1.5 min, somewhat similar to the timing of the pH-triggered conformational changes in and oligomerization of Gc measured for LACV (43). Perturbation of K⁺ influx indicated that ions in addition to H⁺ may be needed for efficient infectious entry of GERV. Similar observations have been made for other OBVs (18, 31, 44). For the unrelated L-PV influenza A virus, the influx of K⁺ into LEs was shown to cause conformational changes in the viral matrix protein and, in turn, loss of stability in viral RNPs (45). More functional investigations are required to clarify whether a switch from Na⁺ to K⁺ is needed to prime OBV cores for efficient uncoating after viral fusion.

Other observations indicated that for GERV, the acid-activated step occurs in LEs. We found that GERV fused with the membrane of endosomal vesicles within 35 to 40 min and that penetration was completed 40 to 80 min later. The process was probably faster given the time required for GERV internalization (5 to 40 min). Despite the observed differences, the time course resembled that of L-PVs, which typically complete the acid-dependent step within 15 to 60 min (32). In stark contrast, viruses that penetrate from EEs, such as Semliki Forest virus, become insensitive within 5 min to agents that elevate endosomal pH (46). More directly, fluorescence confocal microscopy showed that the arrival of GERV particles in endosomal vesicles positive for Rab7a, a key player in LE maturation (27), coincided roughly with the time of acid activation.

Expression of the DN mutant of Rab7a, T22N, impaired GERV infection. Similarly, other studies have reported that OROV enters Rab7⁺ endosomal vesicles (16). In contrast, LACV infection was not affected by the expression of the Rab7 DN mutant (19). However, the results with Rab7 perturbants must be interpreted with caution, as the efficiency of the DN effect can depend on the expression of multiple isoforms of Rab7 and the mislocalization of the mutant (9). For instance, the expression of Rab7 T22N has not been found to affect infection with UUKV or lymphocytic choriomeningitis virus, two L-PVs (26, 46).

Previous studies on the cell entry of OBVs essentially focused on the first steps of internalization, i.e., virus uptake (15, 16, 19). With GERV, we describe here the complete entry process of an OBV, from virus binding to penetration into the cytosol. Our data indicated that GERV is an L-PV and that, as such, its penetration seems to rely on late endosomal vesicles in the early stages of maturation in the degradative endocytic machinery, most likely MVBs. This idea was further supported by the finding that GERV was not observed in LAMP-1⁺ endolysosomes. GERV provides an interesting, safe model for examining the details of the cell entry mechanisms of other OBVs with increasing expansion worldwide.

MATERIALS AND METHODS

Antibodies, plasmids, and reagents.

The guinea pig polyclonal antibody GR1 recognizes the GERV proteins N, Gn, and Gc (22). The plasmids encoding Rab5a, Rab5a S34N, Rab7a, Rab7a T22N, and LAMP-1 tagged with EGFP have been described previously (26). The pLX307-DYNLT3 (Addgene; catalog no. 98329) vectors encoding EGFP-tagged Rab5a, Rab7a, and LAMP-1 (EGFP-Rab5a, EGFP-Rab7a, and LAMP-1-EGFP, respectively) were generated by replacing the DYNLT3 sequence with fragments encoding the nucleotide sequence of the corresponding EGFP-tagged endosomal markers via the In-Fusion method (TaKaRa; catalog no. 639650). Aprotinin (Cayman Chemical), chloroquine diphosphate (Sigma), potassium chloride (KCl; Sigma), ammonium chloride (NH₄Cl; Sigma), and tetraethylammonium (TEA; Sigma) stocks were dissolved in water. AEBSF [4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride; Merck], bafilomycin A1 (BioViotica), Cat074Me (Selleck Chemicals), concanamycin B (BioViotica), SB412515 (Cayman Chemical), and DL-threo-1-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP; Cayman Chemical) were dissolved in dimethyl sulfoxide (DMSO). All drugs were assessed for cytotoxicity at the indicated concentrations using the colorimetric CytoTox96 nonradioactive cytotoxicity assay (Promega) according to the provider’s recommendations. Furin and trypsin were purchased from R&D (catalog no. 1503) and Sigma (catalog no. T1426), respectively. The hydroxysuccinimidyl (NHS) ester dyes Atto 488 and Atto 647N (Atto-Tec) were dissolved in DMSO, and R18 (Thermo Fisher Scientific) was dissolved in ethanol.

Cells and viruses.

All products used for cell culture were purchased from Thermo Fisher Scientific. Human A549 lung epithelial cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1× nonessential amino acids. Human Huh-7 hepatocellular carcinoma cells and human U87 glioblastoma cells were cultured in DMEM and 10% FBS. The baby hamster kidney cell line BHK-21 was cultured in Glasgow’s minimal essential medium supplemented with 10% tryptose phosphate broth and 5% FBS. All cell lines were grown in the presence of 1% GlutaMax, 100 units·mL⁻¹ penicillin, and 100 µg·mL⁻¹ streptomycin. A549 cells expressing EGFP-Rab5a, EGFP-Rab7a, and LAMP-1-EGFP were generated by transduction with the LX307 retroviral vector and isolated by cell sorting as previously described (47, 48). Cells were then grown in the presence of 2 µg·mL⁻¹ puromycin. The prototype strain of GERV has been described previously (49).

Virus production, purification, and titration by a PFU assay.

GERV was produced in BHK-21 cells and semipurified through a 30% sucrose cushion following a procedure established in the laboratory for other Bunyavirales members (24). GERV was labeled with fluorescent dyes using a standard protocol (23). In brief, viral glycoproteins in the virus stock and increasing amounts of bovine serum albumin (BSA) were first semiquantified with ImageJ software (v1.53c) following SDS-PAGE separation and Coomassie blue staining. BSA was used to define a standard curve correlating the amount of BSA to the relative unit values obtained with ImageJ. A linear regression was then applied to determine the molarity of Gn and Gc. The complete procedure was recently described by Hoffmann and colleagues (23). Three molecules of the Atto 488 or 647N dye were then conjugated to one molecule of GERV glycoprotein in a buffer containing 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; 20 mM), NaCl (150 mM), and EDTA (2 mM) at an approximate pH of 7.3 (HNE). Alternatively, GERV (10¹⁰ PFU·mL⁻¹) was labeled with the lipid dye R18 (25 µM; Thermo Fisher Scientific). The labeled particles were subsequently purified through a 15 to 60% sucrose gradient. For titration, confluent monolayers of BHK-21 cells were infected with 10-fold dilutions of virus in FBS-free medium and then cultured in complete medium containing 1% agarose to prevent virus spread. Plaques were stained with crystal violet 3 days postinfection. The MOI is given according to the titers determined in BHK-21 cells.

Protein analysis.

Proteins obtained from purified virus stocks were diluted in lithium dodecyl sulfate (LDS) sample buffer (Thermo Fisher Scientific) and analyzed by SDS-PAGE (NuPAGE Novex 10% Bis-Tris gels; Thermo Fisher Scientific). Proteins were subsequently transferred to polyvinylidene difluoride membranes (iBlot transfer stacks; Thermo Fisher Scientific). Membranes were first blocked with 5% milk and then incubated with the primary antibody GR1 (1:1,000) diluted in Tris-buffered saline containing 0.1% Tween 20 and 5% milk. After extensive washing, membranes were incubated with the horseradish peroxidase-conjugated anti-guinea pig secondary antibody (1:10,000; Santa Cruz). Proteins were detected with enhanced chemiluminescence (ECL) reagents (Thermo Fisher Scientific), and an iNTAS ECL ChemoStar analyzer. Alternatively, viral proteins were analyzed by SDS-PAGE and either fluorography using a Typhoon 8600 imager or Coomassie blue staining, as recently described (22).

Cryo-electron microscopy.

Viral particles purified through sucrose gradients were sequentially washed in HNE buffer, pelleted by ultracentrifugation, and fixed with 4% paraformaldehyde diluted in HNE buffer. Then, 2.5 µL of the fixed virion solution was applied on degassed Quantifoil R2/2 Cu grids. Prior to sample deposition, grids were glow discharged for 2 min at 30 mA. The sample was vitrified in liquid ethane using a Leica EM GP2 plunge freezer at 4°C and 90% humidity, and sensor blotting was performed from the back side for 3 s. The data were acquired using SerialEM software on a Thermo Fisher Scientific Glacios transmission electron microscope operated at 200 kV and equipped with a Falcon 3 direct electron detector. Regions of interest were identified in low-magnification montages. For high-resolution data acquisition, the nominal magnification was ×73,000, resulting in a pixel spacing of 2.019 Å. Prior to data acquisition, the microscope was tuned using coma-free alignment in SerialEM, and the gain reference was determined. The camera was operated in linear mode at a dose rate of 16 e-·pixel⁻²·s⁻¹. The total dose was 19.6 e-·Å⁻² and fractionated over 22 frames, which were aligned and gain corrected on the fly in SerialEM. Cryo-EM micrographs were analyzed with ImageJ v1.53c (NIH). The length and width of a viral particle were determined by measuring the largest and smallest distance between the spikes on the opposite side of the virus particle. The virion diameter was defined as the mean between the length and width.

Virus binding and internalization.

Virions were allowed to bind to cells for 1.5 h on ice at the indicated MOI in binding buffer (DMEM, pH ∼7.4, containing 0.2% BSA and 20 mM HEPES). For internalization assays, fluorescently labeled virus-bound cells were rapidly warmed to 37°C for up to 3 h to trigger endocytosis. Binding and internalization were quantified by flow cytometry or analyzed by confocal microscopy. For flow cytometric analysis, to distinguish between internalized and external particles, samples were treated with trypan blue (0.01%; Sigma) for 15 s at room temperature (23).

DNA transfection.

U87 and A549 cells (8.10⁴) were transfected with 0.5 µg and 0.6 µg, respectively, of plasmids encoding EGFP-Rab5a, EGFP-Rab5a S34N, EGFP-Rab7a, and EGFP-Rab7a T22N using Lipofectamine 2000 (Thermo Fisher Scientific) following the manufacturer’s recommendations. Fresh culture medium was added to the cells 6 h after transfection.

Plasma membrane virus fusion.

GERV was fused to the plasma membrane as previously described (50). In brief, virus binding at an MOI of 5 was synchronized on ice, and virus-bound cells were then treated with buffers of different pH for 1.5 min at 37°C. After pH treatment, cells were extensively washed on ice and were then incubated in complete medium (pH ∼7.4) supplemented with NH₄Cl (50 mM) at 37°C overnight.

Virus infection.

Cells were exposed to viruses at the indicated MOIs in serum-free medium for 1 h at 37°C. The viral input medium was then replaced with complete culture medium, and the infected cells were incubated for up to 8 h before fixation. Cells transiently expressing EGFP-Rab5a and EGFP-Rab7a and the related mutants were infected 18 h posttransfection. For proteolytic activation, GERV was treated to furin or trypsin at the indicated concentrations for 15 min at 37°C and was then allowed to infect cells. For inhibition assays, cells were pretreated with drugs for 30 min at 37°C and were then exposed to viruses in the continuous presence of the inhibitors.

R18-based virus fusion.

Cells were collected from the culture surface with PBS-EDTA (0.5 mM), washed, and exposed to R18-labeled GERV (GERV-R18; MOI of ∼10) in binding buffer (phenol red-free DMEM [pH ∼7.4] containing 0.2% BSA and 20 mM HEPES) on ice for 1 h. Viral internalization was triggered by rapid warming to 37°C in the presence or absence of NH₄Cl (50 mM) directly inside an FP-8500 fluorometer (Jasco), and fluorescence was measured for up to 3.5 h.

Flow cytometry.

Flow cytometry-based infection assays were performed as previously described (25). In brief, infected cells were sequentially fixed and permeabilized with a saponin-based buffer prior to staining of intracellular GERV N, Gn, and Gc with the primary antibody GR1 (1:500) for 1 h. The cells were then washed and subsequently exposed to an anti-guinea pig AF488- or AF568-conjugated secondary antibody (1:500; Thermo Fisher Scientific) for 1 h. Alternatively, in the binding and internalization assays, fluorescently labeled viral particles were directly detected. Infected cells were then analyzed with a fluorescence-activated cell sorter (FACS) Celesta flow cytometer (Becton, Dickinson) using FlowJo software v10.6.2 (TreeStar).

Fluorescence microscopy.

A549 cells exposed to fluorescently labeled viral particles were mounted with Vectashield (catalog no. H-1200; Vector) and analyzed with a Zeiss LSM800 confocal microscope. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; 1.5 µg·mL⁻¹; Vector). Live cell imaging was performed in the continuous presence of viruses, and data were collected with a Leica DMI4000 inverted microscope equipped with a 60× Nikon S Fluor oil immersion objective, a spinning disk confocal head (Yokogawa CSU-W1-T1), and a scientific complementary metal oxide semiconductor (sCMOS) camera (Photometrics Prime 95B). Images were analyzed with the Zen v3.1 software and ImageJ v1.53c (NIH).

Statistical analysis.

The data shown are representative of at least three independent experiments. Values are presented as the means of triplicate experiments ± standard deviations. Graphical plotting of numerical values was performed with GraphPad Prism 9.1.2.