TOSV infects human iPSC-derived neurons

TOSV causes meningoencephalitis in the most severe cases. Therefore, we sought to test the sensitivity of brain cells to TOSV infection. To this end, functional, human glutamatergic neurons were generated from iPSCs through expression of the transcription factor neurogenin-2 (NGN2) [17] and exposed to different multiplicities of infection (MOIs) of TOSV for 48 h. The susceptibility of cells was assessed by flow cytometric analysis after immunofluorescence staining with antibodies (Abs) directed against all TOSV structural proteins, i.e., the nucleoprotein N and the glycoproteins Gn and Gc. Nearly 70% of the neuronal cells were infected at the highest MOI (Fig 1A). The percentage of infected cells increased over time and reached a plateau within 16 h post-infection (hpi) (Fig 1B), indicating that the fluorescence signal detected in this assay corresponded to viral replication and not to the input virions.

Download:

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

Fig 1. Toscana virus (TOSV) productively infects iPSC-derived human neurons.

(A) Induced human pluripotent stem cell (iPSC)-derived neurons were infected with TOSV at the indicated multiplicities of infection (MOIs). Intracellular expression of viral proteins was detected by immunostaining and flow cytometry analysis 48 h later. (B) TOSV infection was monitored in iPSC-derived neurons over 48 h assay as described in A. (C) A549 cells were infected with TOSV for 24 h, and the expression of viral proteins was analyzed by immunostaining and flow cytometry. (D) The titer of TOSV stocks produced on BHK-21 cells was assessed on A549 cells and BHK-21 cells by a plaque-forming unit (pfu) assay. Examples are shown for 10⁻⁸ dilutions of virus productions. (E) TOSV infection was monitored in A549 cells over 24 h and analyzed as described in C and S1 Fig. (F) Supernatants from A549 cells challenged with TOSV at an MOI 2 were assessed for the production of infectious viral progeny over 48 h by the pfu assay described in D.

https://doi.org/10.1371/journal.ppat.1011562.g001

Several cell types of various species have been reported to support productive TOSV infection [5, 14], suggesting broad host range potential and wide tissue tropism of the virus. We evaluated 18 further cell lines from different arthropod and vertebrate species and found that 16 were susceptible and permissive to infection and seven of eight tested cell lines produced infectious viral particles, as determined by our flow cytometric and plaque-forming unit (pfu) assays (S1 Table and Fig 1C and 1D). Of note, myeloid or lymphatic lineages were poorly infected, if at all. The three sand fly cell lines, namely LLE/LULS40, LLE/LULS45, and PPL/LULS49, allowed a complete viral cycle, but their sensitivity to virus infection was low compared to that of most of the mammalian cells.

To evaluate the production and release of infectious viral particles, we infected A549 cells at low MOIs and quantified virus infection and production up to 48 hpi (Figs 1E and 1F and S1). Infectious progeny was found to be released from infected cells as early as 9–16 hpi. Collectively, our analysis indicated that TOSV completes one round of infection, from virus binding and penetration to replication and release of infectious progeny within 9 h in A549 cells. Similar results were obtained in other cell lines tested (S1 Table). As we aimed to analyze TOSV entry mechanisms and restrict infection to a single round in the selected cell lines, we limited our assays to 6 hpi in all further experiments. In addition, we used MOIs for each cell line allowing the infection of approximately 20% of cells. This range of infection generally avoids saturation of infection in cells and thus, allows the detection of potential inhibitory or enhancing effects of a perturbant. Together, this series of data indicates that TOSV efficiently infects human neurons and various mammalian cell lines. In most of the following experiments, we opted to use A549 cells as they are easier to handle than the complex iPSC-derived neuron system.

Labeling of TOSV with fluorescent dyes

To visualize and quantify the early stages of TOSV entry, we purified and labeled the virus chemically with fluorescent probes. Hydroxysuccinimidyl (NHS) ester dyes with an excitation wavelength of 488 nm (Alexa Fluor [AF] 488) or 647 nm (ATTO647N) were coupled to free lysines in the glycoproteins Gn and Gc at a dye-to-glycoprotein ratio between 1:1 and 2:1. At this ratio, we assumed that all virions were labeled. Alternatively, TOSV was labeled with the lipid dye R18 primarily for the analysis of viral fusion. A high concentration of R18 molecules in viral envelopes results in autoquenching of the fluorescence signal [18]. Viral fusion allows the release of R18 into the target cell membranes leading to the dilution of the dye, and, thus, dequenching of the fluorescence.

Characterization of the fluorescent virions is shown in S2 Fig. Briefly, labeled particles were purified through a sucrose gradient so that unbound dyes were removed (S2A Fig). Analysis by SDS-PAGE and Coomassie blue staining showed that the purity of the labeled TOSV preparations was greater than 90% (S2B Fig). The only proteins labeled with NHS ester dyes were Gn and Gc (S2B Fig). The observation that N was not labeled demonstrated that the viral envelope was intact during the labeling procedure. The different labeled particles could be visualized as single spots by confocal microscopy and super-resolution stimulated emission depletion (STED) microscopy (S2C Fig). We did not notice a significant impact of labeling on TOSV infectivity. The titers were similar to those of non-labeled particles (S2D Fig).

TOSV particles were then imaged by cryo-electron microscopy (EM) after fixation with paraformaldehyde and vitrification (Fig 2A). Virions appeared roughly spherical with a diameter of 121 ± 11 nm (n = 96) and protrusions of 9 ± 2 nm (n = 96) (Fig 2B). The measured roundness coefficient of virions was close to 1, i.e., the ratio between their perpendicular width and length was of 0.9 ± 0.1 (n = 96) (Fig 2B). This reflected the nearly spherical shape of the viral particles. Overall, TOSV particles displayed the typical morphology known for other phenuiviruses such as RVFV and Uukuniemi virus (UUKV) [8, 9].

Download:

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

Fig 2. Toscana virus (TOSV) binds to human A549 cells, and its underlying internalization is slow.

(A) TOSV virions were purified and imaged by cryo-electron microscopy (EM). Scale bar, 50 nm. (B) The membrane-to-membrane and spike-to-spike diameter of virions was measured from EM images. The sphericity of each TOSV particle was determined as the ratio of the width to the perpendicular length (n = 96). (C) R18-labeled TOSV was bound to A549 cells on ice at MOI 10, and samples were analyzed by fluorometry. The virus input was measured in cell-free suspension before binding. (D) ATTO647N-labeled TOSV was bound to A549 cells at MOI 1. Nuclei were stained with Hoechst before imaging with a confocal microscope. White spots are cell-associated virions seen in a series of z-stacks merged to one plane (upper panel) and in one focal plane (lower panel). The blue dashed line indicates the nucleus. Scale bar, 5 μm. (E) Various amounts of AF488-labeled TOSV (AF488-TOSV) were bound to A549 cells on ice before analysis by flow cytometry. (F) A549 cells were pre-incubated with increasing amounts of unlabeled TOSV and Semliki Forest virus (SFV) on ice before the addition of AF488-TOSV (15 nM of glycoproteins) and flow cytometry analysis. Data were normalized to those in cells not pre-exposed to the unlabeled virus. T-test with Welch’s correction was applied. **, p<0.01. (G) AF488-TOSV was bound to A549 cells on ice at MOI 10 before warming to 37°C for 30 min. Samples were treated with trypan blue before confocal imaging. Nuclei were stained with Hoechst. Arrowheads show some fluorescent particles at the cell surface that are quenched upon trypan blue addition. Scale bar, 10 μm. (H) Internalization of AF488-TOSV (MOI ∼10) was assessed in A549 cells at 30 min with the trypan blue-based assay shown in G, but using flow cytometry for analysis instead of confocal microscopy. RU, relative unit. (I) Internalization of AF488-TOSV (MOI ∼10) was monitored in A549 cells over 50 min as assayed in H. Internalization is given as the percentage of fluorescence quantified in samples treated with trypan blue compared to that in untreated samples. The fluorescence signal measured in cells not exposed to AF488-TOSV was considered the background signal and subtracted from the other values.

https://doi.org/10.1371/journal.ppat.1011562.g002

TOSV binding to cells is specific

To evaluate TOSV binding to cells, R18-labeled virions were first allowed to bind to A549 cells in suspension at 4°C for 90 min. Incubation on ice allows virions to bind to the cell surface but prevents their internalization into cells. The unbound virions were washed away, and the total fluorescence associated with cell-bound viral particles was determined using a fluorimeter. The results showed that the fraction associated with the cells was about 25% of the virus input (Fig 2C). After binding of ATTO647N-labeled particles to A549 cells at 4°C, virus particles were imaged by confocal microscopy and could be detected on the cell surface (Fig 2D). That the spots had varied sizes suggested that not only individual virions were attached to cells. The largest clusters were probably formed by 2–3 virions but were only seen as a single spot due to the limitation of confocal resolution. The fact that the number of spots per cell was 10.5 ± 8.5 (n = 151) at MOI only 1, i.e., one infectious virion initially added per cell, indicated that the ratio between infectious and noninfectious bound virions was around 1:10.

Flow cytometry analysis allowed detection and quantification of TOSV-AF488 from MOI 3 and above (Fig 2E). Binding of TOSV-AF488 was observed to be abrogated by pre-binding of non-labeled TOSV in a dose-dependent manner (Fig 2F). A 33-fold higher concentration of non-labeled virions reduced TOSV-AF488 binding by one-fourth. Complete inhibition was not achieved because the necessary concentrations of the non-labeled virus were not reached under our experimental settings. However, the pre-binding of Semliki Forest virus (SFV) did not affect TOSV-AF488 binding under the same conditions. Combined, the data indicated that TOSV binding to cells is specific and likely involves one or more attachment factors or receptors.

TOSV enters early and late endosomal compartments

The phenuiviruses analyzed for entry have been shown to enter host cells by endocytosis [7]. We, therefore, aimed to determine whether TOSV is internalized and sorted into early endosomes (EEs) following virus binding to the cell surface. To analyze TOSV internalization by fluorescence microscopy, we synchronized virus entry. To this end, we first allowed TOSV-AF488 to bind to A549 cells on ice at high MOI (∼10). The cells were washed to remove unbound viruses, then rapidly warmed to 37°C to trigger endocytosis, and finally placed back on the ice after 30 min to stop further endocytosis. To discriminate between internalized and surface-bound virions, cells were treated with trypan blue for 10 sec before imaging or flow cytometry analysis. Trypan blue is a membrane-impermeable dye that quenches green-emitting dyes such as AF488 and thus quenches the fluorescence emitted by TOSV-AF488 particles exposed on the cell surface while leaving intracellular virions unquenched (Fig 2G and 2H). Time-course analysis of the generation of trypan blue-resistant fluorescence of cell-associated TOSV-AF488 revealed that internalization into A549 cells started within the first 5 min and increased over time to reach the half-maximal level (t_1/2) within 10 ± 2 min and the plateau within 10–15 min later (Fig 2I). Evidently, TOSV uptake occurred rather synchronously.

To assess whether internalization leads to the sorting of viral particles into EEs, we used A549 cells transiently expressing the small GTPase Rab5a, a marker of EEs, tagged with a monomeric enhanced green fluorescent protein (EGFP). After synchronization of TOSV-ATTO647N binding to A549 cells expressing EGFP-Rab5a on ice, the temperature was rapidly shifted to 37°C for periods of up to 40 min. Confocal microscopy showed TOSV co-localizing with EGFP-Rab5a-positive (+) vesicles 5 min post-warming (Fig 3A). The amount of co-localizing virions reached a maximum within 5–10 min post-warming and decreased thereafter (Fig 3B). In live A549 cells, confocal microscopy showed that TOSV-ATTO647N moved together within EGFP-Rab5a+ endosomal vacuoles (S1 Movie).

Download:

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

Fig 3. Toscana virus (TOSV) enters early endosomes and then late endosomal organelles.

(A) A549 cells transiently expressing EGFP-Rab5a were exposed to ATTO647N-TOSV at MOI 1 on ice and then rapidly warmed for 5 min at 37°C to allow the internalization of virions. TOSV and Rab5a were imaged by confocal microscopy. One focal plane is shown. Higher magnifications of association between TOSV and Rab5a-positive vesicles (white arrowhead) are shown on the right side as a z-stack series. Yellow numbers indicate the position of the stack in the series, and the original plane is marked with 0. Scale bar, 5 μm. (B) Internalization of ATTO647N-TOSV was monitored in cells transiently expressing EGFP-Rab5a wild-type (WT) or its constitutively active mutant Q79L over 40 min as described in A. Co-localization is expressed as the percentage of bound TOSV associated with Rab5a+ vesicles at different times post-warming. A minimum of 6 cells were analyzed per time point. (C to F) Prebound ATTO647N-TOSV was internalized into A549 cells transiently expressing EGFP-Rab7a (C and D) or LAMP1-EGFP (E and F) for up to 40 min before analysis as described in A and B. At least 9 cells were analyzed per condition. Scale bar, 5 μm.

https://doi.org/10.1371/journal.ppat.1011562.g003

In addition, TOSV was observed within vesicles containing the late endosomal markers Rab7a and LAMP-1 tagged with EGFP (EGFP-Rab7a and LAMP-1-EGFP) but at later time points. Co-localization and coordinated motion with EGFP-Rab7a+ endosomes in live A549 cells were maximal 20–40 min after uptake (Fig 3C and 3D and S2 Movie). Co-localization with the lysosome marker LAMP-1 (LAMP-1-EGFP) was somewhat delayed (Fig 3E and 3F). Of note, some virions were located in the middle of the vesicles and had probably not yet undergone fusion with the limiting endosomal membrane. Overall, the results showed that TOSV is sorted into the endocytic machinery after attachment to the cell surface, and the temporal overlap of co-localizations with Rab7 and LAMP-1 suggested that viral particles reach late endosomes (LEs) rather than lysosomes.

TOSV depends on the passage through EEs and LEs for infectivity

To examine whether passage through the endosomal compartments was required for infectivity, we first assessed TOSV internalization and infection in A549 cells transfected with DNA plasmids to express a constitutively-active mutant of Rab5a tagged with EGFP (EGFP-Rab5a Q79L). The expression of this mutant typically results in the enlargement of EEs (Fig 4A), compromising the maturation of LEs and transport of cargo to lysosomes [19]. Infection was measured in populations of cells expressing identical levels of EGFP as selected by flow cytometry. Expression of EGFP-Rab5a Q79L reduced TOSV infection by 80% in comparison to EGFP-Rab5a wild type (wt) (Fig 4B). In contrast to EGFP-Rab5a, the number of virions co-localizing with EGFP-Rab5a Q79L+ vacuoles remained high even after 20–40 min (Fig 3B). Expression of the dominant-negative mutant Rab5a S34N (EGFP-Rab5a S34N), which abrogates the maturation of newly-formed EEs [20], also resulted in a large decrease in infection, i.e., more than 60% (Fig 4B). Altogether, these results indicated that the infectious entry pathway involves the passage of TOSV in Rab5a+ EEs, though the transport to downstream endosomal vesicles is also needed for productive infection.

Download:

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

Fig 4. Toscana virus (TOSV) relies on late endosomal maturation for infectious entry.

(A) ATTO647N-TOSV (MOI ∼1) was internalized for 20 min in A549 cells transiently expressing the constitutively active mutant Q79L of EGFP-Rab5a and then imaged by confocal microscopy as described in Fig 3. One focal plane is shown. Higher magnifications of association between TOSV and Rab5a Q79L-positive vesicles (white arrowhead) are shown on the right side as a z-stack series. Yellow numbers indicate the position of the stack in the series, and the original plane is marked with 0. Scale bar, 5 μm. (B) A549 cells transfected to transiently express EGFP-Rab5 WT, its constitutively active mutant Q79L, and its dominant-negative mutant S34N were challenged with TOSV at MOI 4 for 6 h. Using flow cytometry, cell populations with similar levels of Rab5a were selected and analyzed for infection. Infection was normalized to that in the cell population expressing EGFP-Rab5 WT. (C) The mutants Q67L (constitutively active) and T22N (dominant-negative) of Rab7a were assessed for their effect on TOSV infection in A549 cells as described in B. (D and E) iPSC-derived neurons (D) and A549 cells (E) were pretreated with nocodazole, colcemid, MG-132, or taxol and then infected with TOSV at MOI 10 and 2, respectively, in the continuous presence of inhibitors. Infection was quantified by flow cytometry, and data were normalized to those in control samples without inhibitor treatment.

https://doi.org/10.1371/journal.ppat.1011562.g004

As a block in the maturation of EEs to LEs impeded infection, TOSV potentially needs to pass through LEs for its productive entry into the cytosol. To test this possibility, we evaluate the role of the small GTPase Rab7a in TOSV infection. Rab7a is a key player in LE maturation and functions [21]. When A549 cells transiently expressed an EGFP-tagged dominant negative mutant of Rab7a (EGFP-Rab7a T22N), TOSV infection was severely impaired, but not when the cells expressed the constitutively active Q67L mutant of Rab7a (Fig 4C). In some cell types, LE maturation relies on microtubule (MT)-mediated transport of endosomes to the perinuclear region of cells and on free cytosolic ubiquitin [22, 23]. Treatment of neurons and A549 cells with either nocodazole or colcemid, two drugs that hamper MT polymerization, resulted in a 40–60% decrease in infection (Fig 4D and 4E). When free ubiquitin was depleted by the proteasome inhibitor MG-132, TOSV infection was reduced in a dose-dependent manner in both neurons and A549 cells. Conversely, taxol, an MT-stabilizing drug, did not affect the infection of A549 cells. Taken together, these experiments show that the transport of virions to LEs is required for infectivity.

Low pH is sufficient and necessary for TOSV fusion

To examine whether the acidic pH in endosomal vesicles is important in TOSV infection, the virus was added to neurons and A549 cells in the presence of agents that neutralize vacuolar pH. The lysosomotropic weak bases ammonium chloride (NH₄Cl) and chloroquine induced a dose-dependent inhibition of TOSV infection (Fig 5A–5C). In these experiments, we monitored and regulated the medium for a pH above 7.0 ensuring the effectiveness of the base. Two inhibitors of vacuolar-type H+-ATPases, bafilomycin A1 and concanamycin B, gave similar results. Taken together, these experiments showed that TOSV depends on vacuolar acidification for infection in both neurons and A549 cells. These results also suggested that the virus penetrates host cells by acid-activated membrane fusion.

Download:

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

Fig 5. Toscana virus (TOSV) entry into iPSC-derived human neurons depends on endosomal acidification.

(A to C) iPSC-derived neurons (A) and A549 cells (B and C) were pretreated with agents that elevate endosomal pH at the indicated concentrations and were infected with TOSV at MOI 10 for 8 h and MOI 2 for 6 h, respectively, in the continuous presence of ammonium chloride (NH₄Cl), chloroquine, bafilomycin A1, and concanamycin B. Infection was analyzed by flow cytometry, and data were normalized to those of control samples without inhibitor treatment.

https://doi.org/10.1371/journal.ppat.1011562.g005

To define the pH threshold and link fusion with infection, we tested the capacity of TOSV to fuse at the plasma membrane of cells as originally described for SFV [24]. In such a scenario, the virus bypasses the need for endocytosis during productive infection. Briefly, TOSV was bound to A549 cells at MOI 10 at 4°C, the temperature was rapidly elevated to 37°C for 1.5 min in buffers with different pH values, and NH₄Cl-containing medium at neutral pH was then added for the remaining period of infection to prevent further infection through endosomes. The bypass resulted in efficient infection at pH values of 5.7 and below (Fig 6A). 50% of the maximal infection was reached at a pH of 5.6. These data demonstrated that a reduction in pH is sufficient to trigger infectious penetration of viral RNPs from the plasma membrane to the cytosol. Additional processing steps were apparently not required in the endosomes to activate fusion or infection by TOSV.

Download:

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

Fig 6. Toscana virus (TOSV) penetrates host cells by acid-activated membrane fusion.

(A) A549 cells were exposed to TOSV (MOI ∼10) on ice, subjected to various pH values at 37°C for 90 sec to trigger the virus fusion at the plasma membrane, and then incubated for 7 h at 37°C in the presence of 50 mM NH₄Cl to prevent viral penetration from endosomes. Infection was quantified by flow cytometry, and the data were normalized to those from samples where the infection was triggered with a buffer at pH 5.0. (B and C) TOSV particles were bound to A549 cells (B) and iPSC-derived neurons (C) at MOIs 1 and 15, respectively, on ice and then rapidly shifted to 37°C to allow virus internalization. 50 mM NH₄Cl was added at the indicated times to block further viral penetration. Infected cells were quantified by flow cytometry, and data were normalized to the samples where NH₄Cl was added 80 min (B) and 90 min (C) post-warming. (D) R18-TOSV was bound at MOI 10 to A549 cells on ice and rapidly warmed to 37°C. The increase in fluorescence resulted from the dequenching of the lipid dye R18 after virus fusion with cell membranes, as measured by fluorometry (black line). NH₄Cl was used to block virus fusion by neutralizing endosomal pH and, thus, to define the fluorescence background due to spontaneous translocation of the R18 dyes between the viral envelope and the neighboring cell membrane (grey line). The red line shows the virus fusion-specific R18 release, i.e., the black line (fusion + free diffusion) minus the grey line (free diffusion). RU, relative unit. (E) TOSV and Semliki Forest virus (SFV) were bound to A549 cells on ice, and samples were shifted to indicated temperatures for 50 min. Infected cells were then incubated at 37°C for 6 h in the presence of NH₄Cl and quantified by flow cytometry. Infection was normalized to that in samples incubated throughout at 37°C. (F) TOSV fusion was assessed in A549 cells for efficiency at various temperatures using the assay in A. Data were normalized to those of samples incubated throughout at 37°C. (G) The binding of R18-TOSV (MOI ∼10) to A549 cells was synchronized on ice. Samples were then warmed to 37°C in a fluorometer, and the fluorescence signal was monitored over 10 min. “+pH” indicates when buffers at pH 5.0 or 7.4 were added to trigger virus fusion. Data were normalized to those at the time point 0. The red line shows the virus fusion-specific R18 release at pH ∼5.0. RU, relative unit. (H) R18-TOSV penetration into A549 cells at MOI 10 was recorded in real time for 90 min using the protocol in D, and triton X-100 was added at the end to induce the dequenching of all R18 molecules associated with bound and internalized virions. The data show the ratio between the fluorescence resulting from viral fusion and that associated with all virions in the cells.

https://doi.org/10.1371/journal.ppat.1011562.g006

Acid-activated penetration occurs in late endosomal compartments

To determine the timing of the acid-requiring step following virus internalization, we took advantage of the fact that the rise in endosomal pH is almost instantaneous when NH₄Cl is added to the extracellular medium [25]. To synchronize virus entry, virions were first allowed to bind to neurons and A549 cells on ice at MOIs of 15 and 1, respectively. Virus entry was then allowed by rapid warming to 37°C, and NH₄Cl was added at different times following the temperature switch. NH₄Cl was used at 50 mM to ensure that infection was completely abolished after adding the weak base. Infectious penetration started after a 5-min delay in A549 cells and reached a t_1/2 within 15 min and a plateau 25–45 min later (Fig 6B). It was apparent that individual viral particles had completed the NH₄Cl-sensitive step non-synchronously in neurons, most likely due to the heterogeneity of cell preparations typical of iPSC-derived neurons (Fig 6C).

To further analyze acid-activated membrane fusion in late endosomal vacuoles, we relied on TOSV-R18 to monitor viral fusion in living cells with a fluorimeter. In this assay, the increase in fluorescence signal results from dequenching of the fluorescence lipid dye R18 upon activation of viral fusion. Though the subsequent release of the dye into the target cell membranes corresponds to the hemifusion mixture of outer leaflets and not to fusion pore formation, it is a good correlate for fusion. TOSV-R18 was bound to A549 cells on ice, and virus endocytosis was synchronized by switching the cells rapidly to 37°C (Fig 6D). The kinetics measured in this assay were very similar to those determined with the above procedure based on NH₄Cl addition. The fluorescence signal started to increase after a 6 min lag and reached a t_1/2 within 18 min ± 2 min post-warming with a plateau value about 20 min later.

The kinetics of penetration closely resembled the time course of endolysosomal maturation, which usually lasts 30–60 min [22]. To challenge the hypothesis that TOSV penetration occurs in LEs, we examined the temperature dependence of entry. The transport of cargo from EEs to LEs is known to be inhibited at temperatures below 20°C [26]. TOSV binding to A549 cells was synchronized on ice, and cells rapidly warmed to different temperatures for 1 h before incubation at 37°C for 6 h in the presence of NH₄Cl to prevent further penetration. The infection was greatly reduced at 30°C and below (Fig 6E). In contrast to TOSV, a temperature of 30°C had no noticeable effect on infection by SFV, which is acid-activated in EEs [27]. SFV infection was still detected at temperatures as low as 16°C.

To rule out that the fusion process was altered by temperature, we also analyzed the temperature dependence of TOSV fusion using the bypass protocol described above. At 25°C, fusion corresponded to 40% of the 37°C control (Fig 6F), whereas infection via the normal route was lowered by 85% (Fig 6E). At 21°C, fusion was still 10% of the 37°C control, whereas infection via the normal route could no longer be detected. As expected of a virus capable of replicating in insect hosts, this suggested that fusion was not a bottleneck for penetration at lower temperatures. Most likely the viral particles did not infect at 25°C and below because they did not reach LEs with low enough pH.

Next, we analyzed the dynamics of viral fusion. To this end, we induced fusion of TOSV-R18 at the plasma membrane of cells, as described in the bypass assay, but instead measured the increase in fluorescence associated with the dequenching of R18 dye in real time. In brief, R18-labeled virions were first allowed to attach to A549 cells on ice, and viral fusion was then triggered by the addition of buffer at pH ∼5.0. We found that the release of R18 molecules into target cell membranes reached a t_1/2 of 27 ± 16 sec and was completed 25–50 sec later (Fig 6G). Almost two-thirds of the plasma membrane-bound virions entered the cells and fused (Fig 6H). Together, our data indicate that acid-activated penetration of TOSV involves most cell surface-bound virions and is achieved through a fast and efficient fusion process from late endosomal compartments.

The TOSV envelope glycoprotein Gc is a class-II fusion protein

A growing body of evidence suggests that the Gc envelope glycoprotein of phenuiviruses belongs to the group of class-II membrane fusion proteins [7]. As a result, phenuiviruses are thought to follow an acid-activated penetration process in line with this class of fusion proteins [28, 29]. No structural data are currently available for TOSV glycoproteins, but others have resolved the crystal structure of RVFV Gc [10]. Analysis of the M polypeptide sequence with the blastp algorithm showed that TOSV Gc shares about 48% amino acid identity and 67% similarity with RVFV Gc (Figs 7A and S3). The fusion domain of RVFV reaches 83% amino acid identity with the corresponding sequence in TOSV Gc. This high degree of conservation suggests a fairly recent evolutionary divergence between TOSV and RVFV.

Download:

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

Fig 7. Toscana virus (TOSV) Gc belongs to the group of class-II fusion proteins.

(A) The M polyproteins of TOSV (strain H4906) and Rift Valley fever virus (RVFV) (strain 35/74) were aligned with blastp suite-2sequences and Multiple Sequence Alignment Viewer. The black boxes indicate the transmembrane domains and the blue lines indicate the insertions into TOSV and RVFV M polyproteins. The grey lines correspond to identical amino acids, the light red lines to similar amino acids, and the dark red lines to different amino acids. (B and C) Pre- (B) and post-fusion (C) conformations of TOSV Gc were predicted using the AlphaFold2-derived algorithm ColabFold, and the Gc structures available for RVFV as modeling models (PDB, 6F9F and 6EGT). The structural predictions were visualized using PyMOL. Domains I, II, and III are typical of class-II membrane fusion proteins with the fusion loop at the end of domain II and appear in red, yellow, and blue, respectively. Flexible stem-loops are shown in purple. (D and E) Pre- (D) and post-fusion (E) conformations of TOSV Gc were compared with those of RVFV Gc using UCSF ChimeraX. In the Gc structure overlay, TOSV and RVFV Gc appear in blue and orange, respectively. Otherwise, the color code for domains I, II, and III is the same as that used in B and C. The right boxes in E show magnifications of the fusion unit with corresponding loops and amino acids (upper panel) and the position of A₀B₀ and C₀D₀ strands (lower panel).

https://doi.org/10.1371/journal.ppat.1011562.g007

The level of identity and similarity in amino acids supports the idea that the Gc glycoproteins of TOSV and RVFV resemble each other structurally. To further test this possibility, we utilized AlphaFold to predict the structure of the TOSV Gc ectodomain based on the information available for the RVFV Gc structure (S1 File). A bent conformation was obtained by comparison with intact RVFV virions (Protein Data Bank [PDB], 6F9F) [16], likely corresponding to the canonical pre-fusion orientation of the glycoprotein (Fig 7B). The Gc AlphaFold prediction exhibited domains I, II, and III typical of its phenuivirus orthologs and other viral class-II fusion proteins. The acid-activated conformation of TOSV Gc (S2 File) was predicted using the X-ray post-fusion structure of RVFV Gc (PDB, 6EGT) [30] (Fig 7C). The extent of confidence in the modeling was assessed using an average predicted local distance difference test (pLDDT). The closer the value is to 100, the more likely the prediction is to be close to the real structure. A value below 50 is categorized as low confidence. Both structural models of TOSV Gc achieved a high overall confidence score, with pLDDT values higher than 90 for the individual domains I, II, and III.

The AlphaFold pre-fusion conformation of TOSV Gc was virtually identical to the experimentally-resolved pre-fusion structure for RVFV Gc, except for a 19-degree greater angle between domains I and II (Fig 7D). This difference could be due to the preference of AlphaFold for more energetically stable conformations and not to the original pre-fusion Gc structure itself. The post-fusion Gc forms of the two viruses differed further. The position of A₀B₀ and C₀D₀ strands in domain I did not undergo major changes after TOSV Gc activation (Fig 7E and S3 Movie). Both A₀B₀ and C₀D₀ strands had deviations of 101 and 34 degrees, respectively, from the same strands in the RVFV Gc post-fusion structure. A second distinction was the barrel sheet in domain III, which was lower in position for TOSV than for RVFV, and the third difference was the presence of an isoleucine residue (I967) in the fusion loop instead of a tryptophan (W821). This latter difference is consistent with the observation made by others that isoleucine at this position allows the distinction between phenuiviruses transmitted by sand flies and ticks and other phenuiviruses [30]. Overall, our computational approaches showed that TOSV Gc can be confidently considered a viral class-II fusion protein.

TOSV remains infectious even after acid exposure below the fusion threshold

The activation and priming of class-II viral fusion proteins are described as irreversible steps, and viral fusion proteins act only once [29]. To check whether TOSV fusion is an irreversible process, we first evaluated the possibility of inactivating the virus, by applying acidic buffers in the absence of target-cell membranes, before infection under neutral-pH conditions. In such an assay, the virus undergoes a transition toward the post-fusion state at the optimal pH. If the transition is irreversible, the fusion protein is no longer able to fuse with target-cell membranes, and, thus, the viral particles are rendered noninfectious. Using this approach, we did not observe any negative effect on TOSV infectivity in A549 cells by exposing virions to buffers ranging from pH ∼5.0 to 7.5 for 5 min (Fig 8A). Infection was even greater when the viral particles were subjected to buffers at pH ∼6.0 and 6.5.

Download:

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

Fig 8. Low pH does not inactivate Toscana virus (TOSV) or other bunyaviruses.

(A) TOSV (MOI ∼2) was first pretreated at various pH for 5 min and then buffered at pH 7.4 before the infection of A549 cells and quantification by flow cytometry. Data were normalized to that of samples pretreated at pH 7.4. T-test with Welch’s correction was applied. ***, p<0.001; ns, nonsignificant. (B to E) The sensitivity of Rift Valley fever virus (RVFV) construct RVFVΔNSs:EGFP, Germiston virus (GERV), Semliki Forest virus (SFV), Uukuniemi virus (UUKV), TOSV, and influenza A virus (IAV) to acidic pretreatments was evaluated as described in A. (F) ATTO647N-TOSV at MOI 1 was pretreated at low pH and then allowed to bind to A549 cells on ice before confocal imaging. Nuclei were stained by Hoechst. Scale bar, 10 μm. (G) Depicted is the quantification of viral particles bound to A549, BHK-21, and Vero cells as described in F. n > 124 cells. (H) Alexa Fluor 647-labeled UUKV was pretreated at low pH and bound to BHK-21 cells at MOI 0.3 as described in F. Virus binding was quantified by flow cytometry, and data were normalized to samples pretreated at pH 7.0.

https://doi.org/10.1371/journal.ppat.1011562.g008

To determine whether low pH resistance is a hallmark of TOSV or has larger implications, we expanded our study to related and unrelated viruses with class-I or class-II fusion proteins. Briefly, RVFV, Germiston virus (GERV), and UUKV are bunyaviruses with a class-II fusion protein and related to TOSV. In addition, SFV is an alphavirus with a class-II membrane fusion protein, and the unrelated influenza A virus (IAV) has a class-I fusion protein [29]. SFV enters host cells from EEs with a fusion threshold at pH ∼6.2 while the others are late-penetrating viruses (L-PV) with a pH-activation threshold ranging from 5.0 to 6.0 depending on the virus species [19, 31–34]. Like TOSV, we found that the three other bunyaviruses remained infectious in A549 and BHK-21 cells after exposure to buffers of different pH values (Fig 8B–8D). In contrast to bunyaviruses, the infectivity of SFV and IAV was dampened by 90–95% after exposure to pH below 6.0 (Fig 8D and 8E). We noted that although TOSV infectivity appeared to be lower in BHK-21 cells after exposure to pH ∼5.5, 20–30% of virions remained infectious, whereas SFV inactivation was virtually complete (Fig 8D).

To investigate whether the fusion process as such was affected by greater virus binding to cells following exposure to low pH, ATTO647N-TOSV was first subjected to buffers ranging in pH from 5.0 to 7.4. The fluorescently-labeled virions were then allowed to attach to A549, BHK-21, and Vero cells under neutral pH conditions on ice for 1.5 h before imaging (Fig 8F and 8G). Lowering the pH to 6.0 in the binding medium resulted in a higher efficiency of virion attachment to most cell lines, from ∼10 to ∼20 virions per cell. Similar results were obtained when UUKV binding to BHK-21 cells was analyzed by flow cytometry (Fig 8H). More acidic pH probably caused the unmasking of epitopes in viral glycoproteins that promote interactions with one or more cell attachment factors or receptors. The number of cell-bound virions decreased slightly at the lowest pH, i.e., 5–10 virions per cell at pH ∼5.0. Overall, virus binding could not account for enhanced fusion at pH ∼5.5 and below. Combined, the results indicated that TOSV is not inactivated at pH values below the fusion threshold. This suggested that virions remain infectious in late endosomal vacuoles even when the intraluminal acidity was inferior to the fusion threshold.

A mildly acidic environment primes low-pH-activation of TOSV fusion in LEs

To further examine the effect of pH on TOSV activation and fusion, we induced TOSV-R18 fusion on the A549 cell surface with buffers ranging in pH from 5.0 to 5.8 as described above and monitored the dye dequenching in real time. Viral fusion could be detected for pH as high as 5.8 and the lower the activation pH, the faster the fusion dynamics (Fig 9A). In the pH range of 5.0 to 5.5, half of the bound virions had completed fusion within 27–57 sec (Fig 9B). At pH of 5.8, the t_1/2 was reached within 164 sec, i.e., the process took longer than at lower pH values. When TOSV was first pre-exposed to mildly acidic pH values such as those prevalent in EEs, we found that virions were then able to fuse markedly faster at a pH of 5.8 (Fig 9C), a pH value typically found in LEs at the beginning of their maturation. The t_1/2 of the fusion decreased from 168 to 62 sec, which is somewhat similar to the result obtained with an activation pH of 5.5 without pretreatment. Overall, the data demonstrated that TOSV can achieve fusion at higher pHs when exposed to mildly acidic pH for longer periods. The passage through EEs and exposure to a mildly acidic environment most likely favor the activation of TOSV fusion at lower pHs in LEs.

Download:

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

Fig 9. Toscana virus (TOSV) shows remarkable adaptability to the endosomal environment to penetrate cells.

(A) The binding of R18-TOSV (MOI ∼10) to A549 cells was synchronized on ice. Cells were then rapidly warmed to 37°C, and virus fusion was triggered by adding acidic buffers within a fluorometer while recording the fluorescence signal. The virus fusion-specific R18 release is shown. Data were normalized to those at the time point 0. RU, relative unit. (B) The half-maximal fluorescence intensity (t_1/2) was measured in the series of data obtained in A. n > 3. (C) R18-TOSV particles were pretreated at pH 6.0 or 7.4 and then neutralized as described in Fig 8A before being assessed and analyzed as in panel A. The t_1/2 of fusion was calculated from n > 7. T-test with Welch’s correction was applied. *, p<0.05. (D) After the synchronization of TOSV binding at MOI 1 on ice, A549 cells were rapidly warmed to 37°C in the presence of NH₄Cl (50 mM). NH₄Cl was then washed out at the indicated times to allow endosomal acidification and the acid-activated penetration of infectious TOSV particles. Samples were harvested 6 h later, and infection was analyzed by flow cytometry. Values were normalized to those from samples for which NH₄Cl was removed at t0.

https://doi.org/10.1371/journal.ppat.1011562.g009

TOSV remains acid-activable in endosomal vesicles for long periods

The resistance of TOSV to low acidity and the ability of virions to fuse at various pH values led us to postulate that the virus is less prone to inactivation in the endocytic machinery. To examine how long TOSV remains acid-activable in endosomal vesicles, we reversed the approach of adding NH₄Cl. Virus binding to A549 cells was synchronized on ice, and cells were rapidly warmed in the presence of NH₄Cl before the weak base was washed out at different times. This assay relies on the fact that the neutralization of endosomal pH by NH₄Cl is reversible after washing. In other words, we determined the time at which TOSV acid activation was no longer possible. In A549 cells, infection decreased by 60% during the first 30 min and then more slowly until it reached an 80% decrease after 80 min (Fig 9D). Altogether, these experiments indicated that TOSV infectivity remains high in endosomal vesicles if the virus is not allowed to enter the cytosol after a long period.

Together, our results showed that TOSV resembles late-penetrating viruses in that its entry depends on a normal maturation of LEs. It is transported from EEs to LEs, and its pH of fusion corresponds to that prevailing in LEs. It cannot infect cells in which cargo transport into the degradative branch of the endocytic pathway is blocked. TOSV, and possibly other bunyaviruses, differs remarkedly from other acid-activated viruses in its use of endosomal pH as a cue to enter host cells.

Cell lines

All reagents used for cell culture were obtained from Thermo Fisher Scientific or Merck. Briefly, human A549 and HeLa epithelial cells, HEK293T embryonic kidney cells, Huh-7 hepatocellular carcinoma cells, and U87 glioblastoma cells, as well as chicken DF-1 embryonic fibroblast cells, murine L929 fibroblast cells, canine MDCK kidney epithelial cells, and African green monkey Vero kidney epithelial cells, were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). In addition, 1X non-essential amino acids were added to the culture medium of A549 cells. Baby hamster kidney BHK-21 cells were cultured in Glasgow’s minimal essential medium (GMEM) supplemented with 10% tryptose phosphate broth and 5% FBS. Human Jurkat and SUP-T1 T lymphoblast cells, raji B cells, and THP-1 monocyte cells were grown in Roswell Park Memorial Institute (RPMI 1640) containing 10% FBS and the human SH-SY5Y neuroblast cell line in Minimum Essential Medium (MEM)/F12 (Ham’s F12) supplemented with 10% serum. The sand fly cells LLE/LULS40 and LLE/LULS45 were derived from embryos of Lutzomyia longipalpis and PPL/LULS49 from Phlebotomus papatasi. All sand fly cells were cultured in an L-15-based medium in sealed, flat-sided tubes (Nunc) in ambient air at 28°C as reported elsewhere [45, 46]. All cell lines were grown in the presence of 100 units.mL^-1 penicillin and 100 mg.mL^-1 streptomycin.

iPSC-derived neurons

Human induced pluripotent stem cells (iPSCs) derived from a healthy donor (HD6, Heidelberg University) were cultured on Matrigel-coated (Corning) dishes in mTeSR plus medium (STEMCELL Technologies) at 37°C with 5% CO₂. Cells were split after 3–5 d, depending on colony size, using EDTA (Sigma). Colonies were scraped off and transferred to Matrigel-coated dishes. The medium was changed every other day. Primary mouse glial cells were prepared as described by Patzke and colleagues [47]. Briefly, newborn (p0) mouse cortices were isolated and digested with papain for 20 min, cells were dissociated by trituration using a thin pipette tip and passed through a cell strainer. Cells were then plated onto T75 flasks in DMEM supplemented with 10% FBS. Upon reaching confluence, glial cells were trypsinized and re-seeded twice to remove potential trace amounts of mouse neurons before the glial cell cultures were used for co-culture with induced neuron cells. All procedures involving animals were approved by the Governmental Council Karlsruhe, Germany, and were carried out in strict compliance with German Animal Protection Law (TierSCHG) at the Heidelberg University, Germany. Induced human glutamatergic neurons were generated from HD6 iPSCs as previously described [17]. Briefly, iPSCs were treated with Accutase (Sigma), plated, and simultaneously infected with two lentiviruses: one designed to express rtTA driven by the ubiquitin promoter and another one designed to express, in an inducible manner, NGN2 and puromycin driven by the rtTA promoter. One day later, doxycycline was added to the medium at a concentration of 2 μg.mL^-1 to drive NGN2 and puromycin expression. Two days later, 1 μg.mL^-1 puromycin was added to the medium during 24h for selection. After selection, the remaining cells were detached with Accutase and re-plated on Matrigel-coated coverslips along with mouse glia. Half of the medium was then changed every second day for eight days, and 2.5% FBS was added to support astrocyte viability. After day 10, induced neurons were cultured in B27/Neurobasal medium containing Glutamax (Gibco) and 5% FBS for a minimum of 21 days before infection with TOSV.

Viruses

TOSV strain H4906 (lineage B) [5], recombinant RVFVΔNSs:EGFP [48, 49], GERV [34], UUKV strain S23 [50], SFV [32] and IAV strain PR/8/34 [51] have all been described previously.

Antibodies

Polyclonal antibody against TOSV structural proteins N, Gn, and Gc was a generous gift from R.B. Tesh (University of Texas, Galveston, Texas, USA) [13]. The mouse monoclonal antibody (mAb) against SFV glycoprotein E2 was kindly provided by Prof. Margaret Kielian (Albert Einstein College of Medicine, USA). The polyclonal guinea pig antibody GR1 against N, Gn, and Gc structural proteins of GERV was recently described [34]. The mouse mAb 8B11A3, which targets a linear epitope in the UUKV nucleoprotein N, is a kind gift from Ludwig Institute for Cancer Research (Stockholm, Sweden) [19]. The mouse mAb that detects the IAV nucleoprotein was purchased from Merck (MAB8257). Secondary antibodies conjugated to AF405 and AF488 were purchased from Molecular Probes.

Plasmids and reagents

The plasmids encoding Rab5a, Rab7a, and LAMP1 wt and mutant molecules tagged with EGFP have all been described previously [19]. Stock solutions of chloroquine diphosphate (Sigma) and ammonium chloride (NH₄Cl, Sigma) were prepared in dH₂O at concentrations of 19 mM and 1 M, respectively. Bafilomycin A1 (BioViotica), colcemid (Cayman Chemical), concanamycin B (BioViotica), MG-132 (Selleck Chemicals), nocodazole (Merck), and taxol (Merck) were all dissolved in 100% dimethyl sulfoxide (DMSO) to prepare stock solutions at 100 μM, 10 mM, 50 μM, 40 mM, 20 mM, and 10 mM, respectively. Stock solutions of all drugs were diluted in DMEM at the indicated doses (Figs 4 and 5), which are known not to cause cytotoxicity [23, 34]. Both dH₂O and DMSO were included as solvent controls. The hydroxysuccinimidyl (NHS) ester dyes AF488 (Thermo Fisher Scientific) and ATTO647N (Atto-Tec) were dissolved in DMSO (10 and 5 mg.mL^-1, respectively) while octadecyl rhodamine B chloride (R18, Thermo Fisher Scientific) was dissolved in ethanol (10 mM).

Virus production, labeling, purification, and titration

TOSV, GERV, UUKV, and SFV were produced in BHK-21 cells in serum-free medium whereas RVFVΔNSs:EGFP was produced in Vero cells in 2%-containing medium and IAV in MDCK cells in serum-free medium [13, 23, 34, 51]. All viruses were purified through a sucrose cushion and then titrated. TOSV, GERV, UUKV, and SFV were titrated on BHK-21 cells and IAV on MDCK cells using a pfu assay following procedures established in the laboratory [51, 52]. The titer of RVFVΔNSs:EGFP was determined in A549 cells by quantifying EGFP-positive cells 7 hpi by flow cytometry using a protocol derived from the approach developed by Barrigua and colleagues [53]. The MOIs are therefore given according to the titers determined in BHK-21 cells for TOSV, GERV, UUKV, and SFV, in A549 cells for RVFVΔNSs:EGFP, and in MDCK cells for IAV. TOSV was fluorescently labeled using a previously-described method [52] in which one and two molecules of AF488 and ATTO647N NHS ester dye, respectively, are conjugated to one molecule of the viral glycoproteins for one of the viral glycoproteins. UUKV was labeled following the same method but with three molecules of AF647 [14]. Alternatively, TOSV (3x10⁹ pfu.mL^-1) was labeled with the lipophilic dye R18 (25 μM) [52].

Virus binding and internalization

Virions were allowed to bind to pre-cooled cells in DMEM containing 0.2% BSA and 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) at pH ∼7.4 (binding buffer) on ice for 90 min at indicated MOIs. Where indicated, virions were first buffered to different pH values in 100 mM citric acid (pH < 5.5), 2-(N-morpholino)-ethanesulfonic acid (MES) (5.5 < pH < 6.5), or HEPES (6.5 < pH < 7.4) for 5 min at 37°C before being returned to pH ∼7.4 and allowed to bind to cells. For internalization assays, cells were rapidly warmed to 37°C and incubated for the indicated periods. Both virus binding and internalization were analyzed by flow cytometry, confocal microscopy, and fluorimetry as described below. To discriminate between internalized and external virions, trypan blue (Sigma) was added to a concentration of 0.01% before the analysis. In flow cytometry- and fluorimetry-based assays, cells were detached from culture plastic by incubation with 0.5 mM EDTA, and virus binding and internalization were performed with cells in suspension in phenol-free binding buffer. In binding competition experiments, cells were first exposed to the indicated amounts of unlabeled TOSV or SFV for 45 min at 4°C and then to AF488-TOSV at a concentration of 15 nM viral glycoproteins for an additional hour in the presence of unlabeled viruses on ice.

Infection assay

Cells were exposed to viral particles at the indicated MOIs in the respective medium without serum for 1 h at 37°C. Virus inoculum was then replaced by the respective complete medium, and infection was allowed for an additional 5 h for TOSV and 7 h for RVFV, GERV, UUKV, SFV, and IAV, if not indicated otherwise. Where mentioned, virions were buffered to the indicated pH values as described in the Virus binding and internalization section before the infection of cells. For inhibition experiments, cells were pretreated for 30 min at 37°C, except for colcemid, nocodazole, and taxol. For these three drugs, cells were exposed for 3 h at 4°C before infection was carried out at 37°C. For all the drugs, cells were infected in the continuous presence of the inhibitors. To assess the dependence of TOSV entry on temperature, virus entry was first synchronized on ice. Infected cells were then incubated for 50 min at the indicated temperatures before warming to 37°C in the respective complete medium containing 50 mM NH₄Cl and buffered with 20 mM HEPES for 6 h. For NH₄Cl add-in time courses, virus binding to cells was synchronized on ice. Cells were then rapidly warmed to 37°C, and NH₄Cl (50 mM) was added at the indicated times. Cells were subsequently incubated at 37°C and harvested 6–8 h after the initial warm shift. In the reverse approach, virus binding to cells was synchronized on ice, and cells rapidly warmed to 37°C in the presence of NH₄Cl (50 mM) before the weak base was washed out at the indicated times. Cells were then incubated for an additional 6 h at 37°C. Virus infection was monitored by flow cytometry.

Flow cytometry

Infection was monitored by flow cytometry as previously described [19]. Briefly, infected cells were fixed and permeabilized by 0.1% saponin before the immunofluorescence staining of newly-produced viral proteins with respective primary antibodies against TOSV, GERV, UUKV, SFV, or IAV at concentrations of 2.5 μg.mL^-1, 1:16,000, 1:1,000, 1:400, and 1:250, respectively. Cells were then washed and exposed to anti-guinea pig or anti-mouse AF405- or AF488-conjugated secondary antibodies (1:500, Thermo Fisher Scientific) for one hour. Alternatively, cells infected with RVFVΔNSs: EGFP were assayed for the EGFP signal, and in binding and internalization experiments, AF488-labeled virions were directly measured. Infection was quantified with a Celesta flow cytometer (Becton Dickinson) and FlowJo software v10.6.2 (TreeStar).

Viral protein analysis

Purified virus stocks were diluted in lithium dodecyl sulfate (LDS) sample buffer (Thermo Fisher Scientific) and separated by SDS-PAGE (NuPAGE Novex 10% Bis-Tris gel, Thermo Fisher Scientific) as previously described [5]. Viral proteins were either stained with Coomassie blue or analyzed by fluorography using an LI-COR Odyssey CLx scanner and ImageJ v1.53c [National Institute of Health (NIH), US].

Cryo-electron microscopy

Sucrose gradient-purified virus particles were washed in a buffer containing 10 mM HEPES, 150 mM NaCl, 1 mM EDTA, pH ∼7.3, pelleted by ultracentrifugation, and fixed with 4% paraformaldehyde. Subsequently, 2.5 μL of the fixed virion solution was applied to degassed Quantifoil R2/2 Cu grids that were discharged at 30 mA for 2 min before sample application. The sample was vitrified in liquid ethane using a Leica EM GP2 plunge freezer at 4°C and 90% humidity, and sensor blotting from the reverse side for 3 sec. 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. Before data acquisition, the microscope was adjusted by a comma-free alignment in SerialEM, and the gain reference was determined. Regions of interest were identified in low-magnification setups. For high-resolution data acquisition, the nominal magnification was 73,000, resulting in a pixel spacing of 2.019 Å. The camera was operated in linear mode with a dose rate of 16 e-/s/pixel. The total dose was 19.6 e-/Ų and was divided into 22 dose-fractions that were aligned and gain-corrected in SerialEM. Cryo-EM micrographs were analyzed using ImageJ v1.53c (NIH). The length and width of a viral particle were determined by measuring the largest and smallest distances between peaks in density profiles or membranes on the opposite side of the viral particle.

Fluorescence microscopy

Cells that were exposed to fluorescently labeled virions were mounted with Mowiol (Merck), and if indicated, nuclei were stained with Hoechst 33258 (0.5 μg.mL^-1, Thermo Fisher Scientific). Live cell imaging was performed in the continuous presence of viral particles. Both fixed and live samples were imaged with a Leica TCS SP8 confocal microscope equipped with an HC PL APO CS2 63x/1.4 N.A. oil immersion objective. In addition, super-resolution microscopy was used to image ATTO647N-TOSV mounted in Mowiol on PEI-coated coverslips with a 2-color-STED microscope (Abberior instruments GmbH) as described by Kummer and colleagues [51]. The STED microscope was equipped with an x100 Olympus UPlanSApo (NA 1.4) oil immersion objective, and the pixel size was set to 60 nm (confocal) and 15 nm (non-diffracted), respectively. Minor contrast and brightness adjustments of images and Richardson–Lucy deconvolution (regularization parameter of 10⁻³, stopped after 30 iterations) were carried out using Imspector software 16.1.7098 (Abberior instruments). Images were analyzed with ImageJ v1.53c software (NIH) and the Imspector software (Abberior Instruments GmbH).

DNA transfection

A549 cells (8x10⁴) were transfected with 500 ng of plasmids encoding Rab5a, Rab7a, and LAMP1 wt and mutant molecules tagged with EGFP using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s recommendations. Supernatants were replaced by fresh medium 5 h after transfection, and the cells were incubated for an additional 17 h before exposure to TOSV.

Flow cytometry-based plasma membrane virus fusion

TOSV was induced to fuse with the plasma membrane as previously described [24]. Briefly, TOSV binding to cells at the indicated MOIs was synchronized on ice, and cells were subsequently exposed to buffers of different pH values as indicated for 90 sec at 37°C. Infected cells were then washed extensively and incubated in a complete medium at pH ∼7.4 supplemented with NH₄Cl (50 mM) for 7 h. Infection was quantified by flow cytometry following the immunostaining of TOSV structural proteins.

R18-based virus fusion

The fusion of R18-TOSV with host cell membranes was performed as previously described [34]. Briefly, cells were detached from the culture surface using 0.5 mM EDTA, and binding of R18-TOSV at MOI 10 to cells in suspension was synchronized on ice in a phenol-free medium at pH ∼7.4 for 90 min. To determine the kinetics of virus penetration, virus-bound cells were rapidly warmed in an FP-8500 fluorometer (Jasco) to 37°C, and the emission of fluorescence was measured over 90 min. For virus fusion with cell membranes, virus-bound cells were rapidly warmed inside the fluorometer to 37°C, and the fusion was triggered with buffers of varying pHs as indicated. The fluorescence was measured for 600 sec. Where indicated, virions were pre-exposed to buffers at pH ∼7.4 or 6.0 as described in the Virus binding and internalization section.

Structural modeling of TOSV Gc

The amino acid sequences of the M segment of TOSV and RVFV, strains H4906 and 35/74, respectively, were first aligned and analyzed with EMBOSS Needle or blastp suite-2sequences using a BLOSUM62 matrix and Multiple Sequence Alignment Viewer 1.22.2. ColabFold v1.5.2, an algorithm that combines MMseqs2 with AlphaFold2 [54], was then used to predict the structure of TOSV Gc strain H4906 in pre- and post-fusion conformation. For this analysis, the default settings were applied, and the pre- and post-fusion structures of RVFV (PDB, 6F9F [16], and PDB, 6EGT [30], respectively) served as models. The AlphaFold predictions for TOSV Gc were visualized with the PyMOL Molecular Graphics System, v2.5.4 Schrödinger, LLC. Structural comparisons with RVFV Gc were achieved using UCSF ChimeraX, a “matchmaker” plugin to align the models, and a “morph” plugin to generate a conformation change trajectory [55]. PDB files of TOSV Gc pre- and post-fusion models are available upon request.

Statistical analysis

Graph plotting and statistics were achieved with Prism v8.0.1 (GraphPad Software). The data shown are representative of at least three independent experiments. Values are presented as the means of triplicate experiments ± standard deviation if not stated differently.