Ethics statement

Participants gave written informed consent. Ethics approval was obtained from the relevant national authorities in Cameroon (the Ministry of Health and the National Ethics Committee) and France (Commission Nationale de l’Informatique et des Libertés, Comité de Protection des Personnes Ile de France IV). The study was registered at www.clinicaltrials.gov, https://clinicaltrials.gov/ct2/show/NCT03225794/.

Human plasma samples

Blood samples were drawn from adult populations living in villages and settlements across the rainforest in Cameroon or Gabon. SFV infection was diagnosed by a clearly positive Gag doublet on western blots using plasma from the participants and the amplification of the integrase gene and/or LTR DNA fragments by PCR using cellular DNA isolated from blood buffy-coats [42]. The origin of the SFV was identified by phylogenetic analysis of the sequence of the integrase gene [42]. Plasma samples from 29 participants were used for this study (S1 Table). Nine participants were not infected with SFV and 20 were infected with a gorilla SFV. Their neutralization profile and genotype-specific PCR were performed as described [17].

Cells, viruses and plasmids

BHK-21 (ATCC-CLL-10, hamster kidney fibroblast) cells were cultivated in DMEM supplemented with 5% fetal bovine serum (FBS, PAA Laboratories) and HT1080 cells (ECACC 85111505, human fibrosarcoma) in Eagle’s Minimum Essential Medium with Earle’s Balanced Salts and L-glutamine (EMEM-EBSS, Lonza) supplemented with 10% FBS+ 1% L-glutamine (Gibco). Human embryonic kidney 293T cells (Cat. N° 12022001, Sigma) were cultured in DMEM-10% FBS. We used two zoonotic gorilla SFV strains, SFVggo_huBAD468 (GI-D468) and SFVggo_huBAK74 (GII-K74), isolated from Cameroonian hunters, with limited in vitro passage [41] and the laboratory adapted CI-PFV derived from SFVpsc_huHSRV13 and SFVpve_Pan2 (CII-SFV7) [43]. Viral stocks were produced in BHK-21 cells and titrated with gorilla foamy virus activated β-galactosidase (GFAB) indicator cells [44]. The CI-PFV-GFP molecular clone was derived the HSVR13 molecular clone that encodes infectious CI-PFV [45,46]. The 5’ U3 region was replaced by the immediate/early CMV promoter to obtain Tas-independent constitutive expression (pc13 plasmid, [47]). We inserted the human codon-optimized sequence of gfp 5’ of LP, as previously done for FV vectors [33]. The env and pol genes overlap by 53 nucleotides. We inactivated the 3 ATG sites 5’ of env by nucleotide substitution, resulting in a silent mutation in the pol coding sequence. We inserted the gfp gene after the pol stop codon fused to the full env gene, in which we introduced silent mutations over the first 53 nucleotides to avoid homologous recombination. Infectious CI-PFV-GFP was produced by transfecting 293T cells with 2.3 μg plasmid DNA per 5.10⁵ cells in the presence of LipoD293 In Vitro DNA Transfection Reagent (Signagen Laboratories) at a 3:1 ratio. After 48 h of culture, supernatants were harvested, filtered and used to infect BHK-21 cells. The cells were passed twice a week. Once the cytopathic effect (CPE) was visible, uninfected BHK-21 cells were added at each passage (2:1) to propagate the virus. GFP expression was stable over five months of culture.

We generated the sENV-GFP plasmid encoding GI-D468 fused to GFP in C-term in which we mutated K15, K34, and K55 and deleted the last six amino acids of TM to prevent intracellular retention and enhance its expression at the cell surface, as described for CI-PFV [39,48]. The original plasmid was that used to produce the foamy vector expressing GI-D468 Env [17]. Mutagenesis and cloning were performed by Genscript, The Netherlands.

Neutralization assays

Before use in neutralization assays, plasma samples were diluted 1 to 10 in DMEM + 1 mM MgCl₂, heated 30 min at 56°C to inactivate complement proteins and frozen as single-use aliquots. Transmitter cells (BHK-21 or HT1080) were seeded (10⁵ cells/25 cm² flasks) and infected the following day at a multiplicity of infection (moi) of 0.05. Ruxolitinib (5 μM) was added to gorilla SFV -infected cultures [49]. At 72 h post-infection, SFV-infected cells were seeded in 96-well plates (5 x 10³/well) in triplicates. The next day, serial dilutions of plasma samples were added to transmitter cells and the plates incubated for 1 h before the addition of 5 x 10³ uninfected GFAB target cells. After 72 h of culture, β-galactosidase expression was detected by X-gal staining [44]. An Ultimate UV Image analyzer (CTL Europe, Bonn, Germany) was used to count X-Gal-stained cells. One infectious unit (IU) was defined as a blue cell or syncytia. Despite using the same moi on day 0 and the same transmitter-to-target cell ratio on day 4, the number of IUs was higher for chimpanzee than gorilla SFV strains and higher for HT1080 than BHK-21 cells on day 7. Data were therefore expressed as relative infectivity for easier comparison of plasma titration curves. Cells cultured without the addition of plasma samples provided the reference value, which is indicated in the figure legend. Relative infectivity was calculated for the wells treated with plasma samples and is expressed as a percentage of the reference value. Cell-free virus neutralization by the same plasma samples was carried out as described [17]. Briefly, assays were performed in triplicate in 96-well plates in which 5 x 10³ GFAB cells were infected with 100 IU of virus preincubated with serial dilutions of plasma samples. Neutralization titers were defined as the inverse of the plasma dilution required to reduce viral infectivity by half (IC₅₀).

Fluorescence microscopy analysis

BHK-21 or HT1080 cells were seeded (10⁵ cells/25-cm² flask) and infected the following day at a moi of 0.05. At the second passage, cells were seeded on glass coverslips (borosilicate glass D263M, Marienfeld) at a density of 3 x 10⁴ cells/ml in 24-well plates. Cells were cultured until the appearance of CPE or for a maximum of three days. Cells were fixed by removing the culture medium before the addition of PBS containing 2% paraformaldehyde (PFA) and incubation for 10 min at room temperature (RT). Intercellular connections were preserved by directly adding PBS containing 16% PFA to culture medium to reach a final PFA concentration of 2%. Cells were washed with PBS and stored at +4°C in PBS.

Before intracellular staining, cells were permeabilized with 0.5% triton X-100 (Sigma-Aldrich) in PBS for 5 min, washed in PBS-0.1% Triton X-100 (PBST) and incubated with PBST-1% BSA for 20 min to block non-specific binding. SFV proteins were detected using murine monoclonal antibodies (mAbs) generated by M.L. Linial. P6G11G11 is specific for the intracytoplasmic domain of the Env leader peptide and was used conjugated to AF647 dye (anti-LP-AF647 [49], 5μg/mL). P3E10 is specific for the Env SU [50]. In ELISA assays [24], it binds to the 206-VMIDFEIPLGDPRDQ-220 peptide (OD = 1.61 vs 0.07 against the adjacent peptide for 1:100 diluted hybridoma supernatant). P3E10 was used either unconjugated (1:250 dilution of hybridoma supernatant) or conjugated to biotin (5 μg/mL). Accordingly, the secondary reagents were donkey anti murine-IgG (H+L)-AF488 (A21202, Invitrogen, 2 μg/mL) and Streptavidin-AF488 (S32354, Life Technologies, 1 μg/mL). Phalloidin-conjugated to rhodamine (R415, Lifesciences technologies, 1/250) was added to cells at the same time as the secondary reagents. We used an anti-CD98-FITC antibody (556076, BD Biosciences clone UM7F8) to stain the plasma membrane in combination with biotinylated P3E10 and Streptavidin-AF647 (S32357, Life Technologies, 1 μg/mL). The antibodies were diluted in PBST-1% BSA, the staining performed for 1 h at RT and the coverslips washed three times with PBST between each step. DAPI (0.6 μM, D9542, Sigma-Aldrich) was added followed by a 5-min incubation. Coverslips were washed with PBS, rapidly rinsed with H₂0 and mounted on SuperFrost slides (Thermo Fisher) with one drop of SlowFade diamond antifade mountant (S36967 Invitrogen). Nail polish was used to fix the coverslips and avoid desiccation.

Before surface labelling, cells were fixed by the addition of PBS containing 16% PFA to the culture medium to reach a final PFA concentration of 2% and incubated for 1 h at RT. Staining was performed without Triton X-100 and the coverslips were incubated with the antibodies for 30 min. The SU-specific reagents were the same as for the intracellular labelling. Human plasma samples were diluted 1/100 to stain permeabilized or nonpermeabilized cells and anti-human IgG-A-M-FITC (0.2 μg/mL, 74511, Bio-Rad) was used to detect bound human antibodies.

Images were acquired using an inverted (Axio observer Z1, Zeiss) or upright (Apotome 2) widefield microscope equipped with an apotome grid and the following objectives: Plan-Neofluar 10x/0.3, 20x/0.8, 25x/0.8, 100x:1.3 Oil; Plan-Apochromat 40x/1.4 Oil and 63x/1.4 Oil. Acquired serial Z-plane frames at various Z depths with the same XY position (Z-stack) were used to build three-dimensional (3D) reconstitution images with Imaris software (Oxford instruments). Low magnification (x10) images were captured to check that the infection frequency was in the expected range (roughly 0.2 to 2%) and that cell distribution was homogenous on the 0.5 cm² coverslip. For each item of interest, images from at least five different fields were acquired at high magnification (x63 or x100) for the two cell types (BHK-21 and HT1080), and three viral strains (CI-PFV, GI-D468 and GII-K74). All images were independently checked by at least two of the authors before selection of the representative ones.

For the quantification of infected cells, coverslips were stained with anti-SU-biotin, Streptavidin-AF488 and DAPI. Automatic mosaic acquisition was performed at low magnification (10x) on an area covering at least 37.5 mm². All coverslips were analyzed the same day using fixed illumination parameters. The STARDIST method was used on the DAPI channel for nuclei segmentation [51] using the pre-trained versatile fluo 2D model. Maximal and minimal intensities of anti-SU staining were recorded for each segmented nucleus and then normalized (from 0 to 1) using values acquired for all samples from a given experiment. An infection status was then determined for each nucleus using the Otsu automatic threshold [52] on the normalized maximum anti-SU intensity distribution. The analysis pipeline was implemented in Python 3.7 with the scipy 3.2.4, sci-kit image 0.18.1, numpy 1.20 and pandas 1.2.4 libraries and is available at https://doi.org/10.5281/zenodo.6478121.

Flow cytometry staining of SFV-infected cells with plasma samples

Cells (BHK-21 or HT1080) were seeded (10⁵ cells/25-cm² flasks) and infected the following day at a moi of 0.05. At 72 h post-infection, cells were passed at 1:10 and 1:20 in 25-cm² flasks. Starting from observation of the first syncytia in culture, infected cultures were labelled daily to assess intracellular Env expression. Cells were fixed with 2% paraformaldehyde (PFA) diluted in PBS for 10 min at RT, washed in PBS supplemented with 0.1% bovine serum albumin (BSA), permeabilized with 0.1% Triton X-100 in PBS-0.1% BSA for 10 min at RT and washed before the addition of anti-SU for 30 min at RT. Cells were washed with PBS-0.1% BSA, donkey anti murine-IgG (H+L)-AF488 was added and the cells incubated for 30 min at RT. Cells were then washed and resuspended in 300 μl PBS-2% PFA. When infection rates were above 20%, cells were used the following day for surface labelling, performed in the absence of Triton. Plasma samples (diluted 1:10) were added to cells resuspended in PBS-0.1% BSA for 30 min at RT. Cells were washed and incubated with anti-human Ig-G-BV421 (2 μg) for 30 min at RT and fixed with 300 μl PBS-2% PFA. The CI-PFV-GFP-infected BHK-21 cells were passed at 1:2 with uninfected BHK-21 and stained 2 to 4 days later using the procedures described for cells infected with primary strains. BHK-21 cells were transfected with the sENV-GFP plasmid (2.3 μg DNA per 5.10⁵ cells in the presence of LipoD293 at a 3:1 ratio) and stained two days later using the procedures described for cells infected with the primary strain.

Data were acquired using a Cytoflex cytometer (Beckman Coulter) and analyzed with Kaluza software. For the analysis, single cells were selected by gating on a FSC-A/SSC-A dot-plot. Env-specific staining was quantified by the percentage of Env-expressing cells for intracellular labelling or median fluorescence intensity (mfi) for surface labelling. Staining of SFV-infected cultures is expressed as the ratio of mfi from infected to mock-infected cultures, or ratio of mfi from GFP expressing (GFP^pos) to GFP negative (GFP^neg) cells to correct for nonspecific binding of plasma samples to cells.

Live-cell imaging

CI-PFV-GFP-infected cells were mixed with uninfected cells to obtain a GFP-expressing cell frequency of 5% and seeded in 96-well plates (3000 cells/well) in triplicate. Plates were cultivated in an IncuCyte Live-Cell Analysis device and phase and green fluorescence images acquired every hour over four days with a 4x or 20x objective. Cell confluence was quantified by the Incucyte software using in-phase contrast images. We manually counted intercellular connections and measured their length using the software measurement tool.

Statistics

We used the paired t test to compare the relative infectivity of cells and virus exposed by plasma samples. The level of significance was set to 0.05.

Setting up of the cell-to-cell SFV transmission assay

Two cell lines, one epithelial (BHK-21) and one fibroblastic (HT1080), with the highest susceptibility to primary zoonotic gorilla SFV were used as virus-transmitter cells [49]. They were infected at a low moi for a short period of time to minimize viral particle release associated with CPE. The target cells were the indicator GFAB cell line, which produces β-galactosidase under the control of the LTR from a gorilla SFV [44]. After coculture with the transmitter cells, the infected target cells were specifically stained for the colorimetric reaction, with no signal detected in the transmitter cells. Gorilla SFV-infected transmitter BHK-21 cells were seeded at a low density to reach a final level of infection in the same range as that used in the cell-free neutralization assay [17].

We verified that SFV infection detected in the microtitration assay required cell contact (Fig 1). No infection was detected when transmitter and target cells were separated by a membrane permeable to cell-free virus (transwell). The supernatants of transmitter cells carried little infectious virus relative to the cells themselves. Cell-to-cell and cell-free infection led to clusters of infected cells and homogenous distribution throughout the well, respectively (Fig 1A). Thus, we defined culture conditions in which GFAB cells were infected through cell-to-cell transmission, with no or minimal contribution of infection by released free viral particles. Infection with GI-D468 and GII-K74 strains was tested in the presence of a plasma sample from an uninfected individual (SFV^neg) or neutralizing plasma samples (anti-GI and anti-GII, respectively) in the four conditions of infection: contact with infected cells, infected cells in transwells, infected cell supernatants, and cell-free virus). The neutralizing plasma samples blocked cell-free transmission and had no or a partial effect on cell-to-cell transmission (Fig 1B and 1C).

Download:

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

Fig 1. Settings of the SFV cell-to-cell transmission assay.

A. Schematic description of the experiments to verify that SFV infection requires cell contact. On day 0, transmitter BHK-21 cells were infected at a moi of 0.05 for 72 h in T25 cm². On day 3, uninfected indicator GFAB cells were seeded in 24-well plates (2 x 10⁴ cells/well) and infected cells were added to GFAB cells on day 4. These were placed in direct contact with the indicator cell line or in the insert of transwell devices (polycarbonate membrane-insert pore size = 0.4 μm). In addition, we quantified the infectious virus present in the supernatants of infected cells. As a control, de novo infection with cell-free virus was performed using the same moi as for standard neutralization assays of cell-free virus (moi of 0.02). Plasma samples were added to the culture medium at a final dilution of 1:100 on day 4, immediately after infection. On day 7, β-galactosidase expression under control of the SFV LTR was detected by addition of its chromogenic substrate (X-gal), leading to blue-stained infected GFAB cells. Representative images of wells cultivated in the absence of plasma samples illustrate the foci of infected cells after cell-cell transmission, the efficient blockade of infection by the transwell device, the low dose of infectious virus present in the supernatant of transmitter cells on day 4, and the homogenous spread of infected cells after cell-free virus infection. B. GI-D468-infected BHK-21 cells and GI-D468 virus were used to infect GFAB cells via the four transmission modes described in panel A. On day4, we added 1:100 diluted plasma from one uninfected individual (SFV^neg, MEBAK195, gray symbols) and one neutralizing plasma (anti-GI, LOBAK2, blue symbols). The number of infectious units (i.e., cells or syncytia) per well is presented for independent experiments. C. GII-K74-infected BHK-21 cells and GII-K74 virus were used to infect GFAB cells via the four transmission modes described in panel A. On day4, we added 1:100 diluted plasma from one uninfected individual (SFV^neg, MEBAK195, gray symbols) and neutralizing plasma (anti-GII, MEBAK88, red symbols). The number of infected cells per well is presented for independent experiments. Panel A was created with Biorender.com.

https://doi.org/10.1371/journal.ppat.1010470.g001

Plasma samples from SFV-infected individuals do not neutralize SFV cell-to-cell transmission

Then, we used the microtitration assay to quantify inhibition of cell-associated and cell-free virus by serial dilution of plasma samples (Fig 2A). Two plasma samples that neutralize the cell-free GI-D468 strain (BAD463, LOBAK2) failed to neutralize GI-D468-infected BHK-21 cells (Fig 2B and 2C). Similarly, two plasma samples that neutralize cell-free GII-K74 (BAD551, BAK232) did not neutralize GII-K74-infected BHK-21 cells (Fig 2D and 2E). Importantly, without the addition of plasma samples, SFV-infected BHK-21 cells transmitted a similar or lower number of IUs than in the cell-free virus condition: GI-D468: 33 and 70 IU/well for infected cells and virus, respectively; GII-K74: 67 and 80 IU/well. Therefore, the action of neutralizing antibodies was assessed against comparable cell-free and cell-associated infectious doses.

Download:

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

Fig 2. The neutralization capacity of plasma samples from SFV-infected individuals depends on the infection route.

A. Schematic description of the neutralization experiments. On day 0, transmitter cells (BHK-21 or HT1080) were infected at a moi of 0.05 and seeded in 96-well microtitration plates (5 x 10³ cells/well). On day 1, infected cells were incubated with serial dilutions of plasma samples for 1 h before the addition of 5 x 10³ uninfected GFAB cells. On day 4, β-galactosidase expression by infected GFAB cells was detected by X-gal staining. For cell free virus neutralization, GFAB cells were seeded in 96-well microtitration plates at day 3. On day 4, the viral inoculum (moi of 0.02) was incubated with serial dilutions of plasma samples for one hour before addition to GFAB cells. At day 7, β-galactosidase expression by infected GFAB cells was detected by X-gal staining. B-E: Neutralization curves of BHK-21 SFV-infected cells (open symbols) and cell-free virus (filled symbols) are shown for four plasma samples against the matched genotype strain. Individuals BAD463 (B) and LOBAK2 (C) are infected with a GI strain and their plasma samples were tested against the GI-D468 strain; individuals BAD551 (D) and BAK232 (E) are infected with a GII strain and their plasma samples were tested against the GII-K74 strain. F-I: Plasma samples from BAD463, LOBAK2, BAD551 and BAK232 SFV-infected individuals and from one uninfected control (MEBAK189) were tested for the neutralization of BHK-21 (F, H) or HT1080 (G, I) cells infected with GI-D468 (F, G) or GII-K74 (H, I). In panels B-E, results are presented as IU/well to show that the infectious loads were comparable between cell-transmitted and cell-free virus conditions. IU/well values of 0 were replaced by 1 to allow visualization on the graph. In panels F to K, the results are expressed as the infectivity relative to that of cells cultivated in the absence of plasma sample (untreated cells) to show data from each virus–plasma pair on the same scale. Experiments have been carried out in triplicates and mean and standard deviation are shown. The mean number of IU/well transmitted by untreated cells were 24 (GI-D468 infected-BHK-21, panel F), 83 (GII-K74 infected-BHK-21, panel H), 978 (GI-D468 infected-HT1080, panel G) and 629 (GII-K74 infected-HT1080, panel I). In each panel, the inverse of the plasma sample dilution is presented on the x axis and the mean and standard errors from triplicates are shown. J-K. Twenty-two plasma samples diluted 1:80 were tested for their capacity to neutralize BHK-21 transmitter cells infected with GI-D468 (J) or GII-K74 (K). Results are expressed as the infectivity relative to that of cells cultivated in the absence of plasma samples. The reference values for cell-transmitted virus were 25 IU/well for GI-D468 and 83 IU/well for GII-K74 and ≈ 100 IU/well for cell-free virus according to the experimental design [17]. The relative infectivity of cell-transmitted virus in the presence of plasma samples is shown by open squares and labelled as “cells” on the x axis. Relative infectivity of cell-free virus in the presence of the same plasma sample is shown for comparison by filled squares, labelled as “virus” on the x axis. Data are presented for four plasma samples from uninfected controls (grey symbols), nine samples from SFV-infected individuals that neutralize the GI-D468 strain (blue symbols), five samples that neutralize the GII-K74 strain (red symbols) and four samples that neutralize both the GI-D468 and GII-K74 strains (purple symbols). P values from the paired t test are indicated in panels J and K. Panel A was created with Biorender.com.

https://doi.org/10.1371/journal.ppat.1010470.g002

These four samples and a fifth one drawn from an uninfected control were tested at dilutions ranging from 1:20 to 1:5,120 against two cell lines (BHK-21 or HT1080) infected with strains belonging to the two gorilla SFV genotypes (i.e., GI-D468, GII-K74). BHK-21 cells transmitted less virus than HT1080 (as indicated in the figure legend) and the results are expressed as the infectivity relative to untreated cells for easier representation. Both cell lines transmitted gorilla SFV that was resistant to neutralizing activity (Fig 2F and 2I). In addition, plasma samples failed to block infection transmitted by cells infected with the antigenically related chimpanzee SFV strains, CI-PFV and CII-SFV7 (S1A–S1D Fig).

We generalized our findings by testing additional plasma samples at a single dilution (1:80) against BHK-21 cells infected with the gorilla or chimpanzee SFV (Figs 2J, 2K, S1E and S1F). The data obtained with cell free virus are shown for comparison. None of the 22 samples neutralized the SFV infection transmitted by infected cells, despite inhibition of cell-free virus from one or both genotypes. In conclusion, plasma antibodies neutralize cell-free SFV infection but not SFV cell-to-cell transmission under closely related in vitro culture conditions.

Early addition of neutralizing antibodies before de novo production of viral proteins reduces but does not prevent the cell-to-cell spread of SFV

In the microtitration assay, infected transmitter cells were cultured for three days before incubation with plasma samples and coculture with uninfected target cells. Such kinetics may allow Env expression to reach levels at which anti-Env antibodies present in the samples are all bound to Env, leaving a sufficient amount of free Env protein to mediate viral entry. We thus tested another experimental setting (S2A Fig) in which we added plasma antibodies (SFV^neg, anti-GI, anti-GII and anti-(GI+GII)) on BHK-21 cells two hours after exposure to cell-free virus. We detected SFV infection by immunofluorescence. Cells were labelled with anti-SU and DAPI and 45,000 to 140,000 cells per condition were analyzed (S3 Fig). Plasma samples with high neutralizing titers against the genotype-matched strain reduced the rate of infected cells by 47% (GI-D468 + anti-GI plasma), 13% (GI-D468+antiGI+GII plasma) and 70% (GII-K74 + anti-GII plasma) (Fig 3A and 3B). These reductions were inferior to those observed in the cell-free neutralization assay (> 99% at 1:100). Furthermore, the presence of plasma induced no obvious change in the size or morphology of mono- or multicellular infected foci (Fig 3C). In conclusion, neutralizing plasma antibodies added early after exposure to the virus and before de novo viral protein production had a very modest effect on cell-to-cell spreading of SFV.

Download:

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

Fig 3. The addition of plasma samples from SFV-infected individuals before the production of viral proteins does not block cell-to-cell transmission of gorilla SFV.

BHK-21 cells were infected at a moi of 0.05 with GI-D468 (A) or GII-K74 (B) for 2 h, washed and cultured in the presence of 1:100 diluted plasma samples (S2A Fig). At the last passage, cells were seeded on glass coverslips and fixed 72 h later. Cells were permeabilized with 0.5% Triton X-100 and stained with anti-SU-biotin+Streptavidin-AF488, phalloidin (actin) and DAPI (nuclei) and analyzed by brightfield microscopy at a magnification of 10x. Images covering a 37.5-mm² square in the center of the coverslip were acquired. Nuclei and infected cells were quantified as described in Materials and Methods from 25 images per coverslip, corresponding to 48,000 to 140,000 analyzed cells/condition. Data are presented as the infection rate per image to reflect the variability of the infection rate across the coverslip. Plasma samples were derived from an uninfected individual (MEBAK189, grey symbols) and from SFV-infected individuals for whom the plasma neutralized GI-D468 (LOBAK2, blue symbols), GII-K74 (MEBAK88, red symbols), or both strains (BAK177, purple symbols). Their neutralization titers (IC₅₀) are indicated on the figure. The Mann-Whitney test was used to compare cultures treated with samples from uninfected and SFV-infected individuals and the P values are indicated above the graphs. C. Representative images captured at magnification of 63x showing that the addition of non-neutralizing (MEBAK189, upper line) or neutralizing plasma samples (LOBAK2 and MEBAK88, bottom line) had no effect on the morphology of the BHK-21-infected cells, which consisted of either isolated elongated cells (white arrows) or multicellular infected foci (yellow arrows). Env, actin and nuclei are presented in green, red and blue, respectively; scale bar = 20 μm.

https://doi.org/10.1371/journal.ppat.1010470.g003

Cellular localization of gorilla SFV Env

Env protein mediates the fusion of cellular membranes, allowing cell-to-cell transmission of SFV [39]. We hypothesized that neutralizing antibodies failed to inhibit SFV cell-to-cell transmission because of reduced availability of Env at the cell surface. FV buds both intracellularly and at the plasma membrane and viral strains differ in their capacity to bud at the cell surface [25,36–38,53]. Gorilla SFV has not yet been characterized in this respect. We, therefore, performed fluorescence microscopy analysis to visualize the cellular distribution of Env in BHK-21 and HT1080 cells infected with GI-D468 and GII-K74 (S2B Fig). We used the CI-PFV strain at the same moi for comparison. Cells were permeabilized before Env, actin and nuclei staining. The BHK-21 cells grew at a higher density than HT1080 cells (Fig 4A and 4A’) and were less susceptible to CPE induced by SFV (Fig 4B–4J and 4B’–4J’). Env was present surrounding the nuclei and in the cytoplasm in isolated SFV-infected cells (Fig 4B, 4E and 4B’, white arrows). Some SFV-infected cells adopted a fusiform shape (Fig 4E, 4C’ and 4D’, yellow arrows) and others showed intercellular connections containing Env (Fig 4E–4G, and 4E’–4G’, blue arrows). Both GI-D468 and GII-K74 induced syncytia of varying size (Fig 4H–4J and 4H’–4J’). Among BHK-21 cells, the largest contained more than 20 nuclei. In general, the observed syncytia were smaller in HT1080 cells, probably due to a greater CPE. Diffuse cytoplasmic Env staining was found within a crown formed by nuclei or surrounding the nuclei (Fig 4H–4J and 4I’–4J’, white and yellow triangles, respectively). The Env distribution, morphological changes and formation of multinucleated cells were similar in cultures infected with gorilla SFV strains and CI-PFV (Fig 4).

Download:

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

Fig 4. Env localization in gorilla SFV-infected cells.

Cells were infected with GI-D468, GII-K74, or CI-PFV, at a moi of 0.05. At the second passage, cells were seeded on glass coverslips and fixed within 1 to 3 days, depending on the cytopathic effect (S2B Fig). Cells were permeabilized with 0.5% Triton X-100 and stained with anti-SU-biotin+Streptavidin-AF488, phalloidin (actin) and DAPI (nuclei) and analyzed by brightfield microscopy with optical sectioning. Top panels: BHK-21 cells; bottom panels: HT1080 cells (indicated as A’ to J’ for conditions matching A to J). A: Mock infected cells, B-E-H: GI-D468 infected cells, C-F-I: GII-K74 infected cells, D-G-J: CI-PFV infected cells. B-C-D: SFV Env in isolated cells is located in the cytoplasm and surrounding the nuclei and some infected cells acquired fusiform shapes. E-F-G: Intercellular connections are visible between two infected cells or between infected and uninfected cells. Env labelling is present in the intercellular connections; H-I-J: Syncytia with diffuse Env staining within a crown formed by nuclei or in the cytoplasmic space around the nuclei; both patterns were observed for the three strains and both cell lines. Representative examples of Env staining are indicated with white arrows for cytoplasmic labelling, yellow arrows for fusiform shapes, blue arrows for intercellular connections, white triangles for cytoplasmic labelling inside a crown of nuclei in syncytia and yellow triangles for cytoplasmic labelling around the nuclei in syncytia. Env, actin and nuclei are presented in green, red and blue, respectively; scale bar = 20 μm.

https://doi.org/10.1371/journal.ppat.1010470.g004

Three-dimensional reconstitution images showed the distribution of Env in cytoplasmic vesicles and the perinuclear region (Fig 5A), the absence of polarized SFV Env staining at the contact zones between infected and uninfected cells (Fig 5B) and the formation of long intercellular connections that did not adhere to the coverslip and formed continuous cytoplasm (Fig 5C). Some Env was localized close to the actin network beyond the cell surface (yellow arrows), suggesting that Env may be expressed at the cell surface (Fig 5A and 5C). We confirmed this hypothesis by co-staining the SFV-infected HT1080 cells with anti-SU and anti-CD98 (large neutral amino acid transporter (LAT1), commonly used as plasma membrane markers) and observed colocalization of the two molecules (Fig 5D, yellow arrows).

Download:

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

Fig 5. Env localization in gorilla SFV-infected cells.

The infection and labeling conditions are those described in Fig 4. Acquired serial Z-plane frames at various Z depths with the same XY position were used to build 3D views with nuclei (blue), Env (green) and actin or CD98 (red) shown in a blend-rendering mode. A. GII-K74-infected BHK-21 cells, with Env staining distributed throughout multiple vesicles and around the nuclei (3D view from the top). Yellow bars correspond to orthogonal slices showing Env staining distributed throughout multiple vesicles and around the nuclei. Grid gaps correspond to 10 μm. B. Dense network of GII-K74-infected BHK-21 cells, in which an infected cell shows an elongated intercellular connection on its right side (3D view from the side). Env staining is evenly distributed despite contacts with multiple adjacent cells. Patchy perinuclear staining is observed in one cell contacting the intercellular connection (yellow square). C. Two GI-D468-infected HT1080 cells with an intercellular connection that does not adhere to the coverslip (3D view from the top side and from the side (insert)). Yellow bars indicate the position of orthogonal slices: a. cytoplasm of a fusiform cell, b-c. Intercellular connection with Env staining (triangle), d. cytoplasm and nucleus from the second infected cell (triangle). Colocalization of Env and actin is indicated by the yellow arrows. D. GI-D468-infected HT1080 cells were permeabilized with 0.5% Triton X-100 and stained with anti-SU-biotin + Streptavidin-AF647, anti-CD98-FITC, and DAPI (nuclei) (note that the colors in the figure were matched with those of the previous panels and not the conjugated fluorochromes). Yellow bars indicate the position of orthogonal slices. Colocalization of Env and CD98 is indicated by the yellow arrows.

https://doi.org/10.1371/journal.ppat.1010470.g005

Overall, our results show that upon infection with gorilla SFV, Env is predominantly detected in the cytoplasm of infected cells. Polarized Env localization at sites of cell contact that resemble virological synapses [54] was not observed, whereas intercellular connections containing SFV Env were frequent. Microscopy staining indicates that Env is expressed at the cell surface and the presence of syncytia indicate that their quantity at the plasma membrane is sufficient to induce cell fusion.

Staining of SFV-infected cells by plasma antibodies reveals additional virus-induced changes in cell morphology

We analyzed gorilla SFV-infected BHK-21 and HT1080 cells labelled with plasma samples without permeabilization by fluorescent microscopy analysis. We observed dense staining localized either on the side of or at the cell apex (Fig 6A–6E). As a positive control, we permeabilized cells and observed a higher frequency of cells with cytoplasmic staining (Fig 6F–6J). Surface labelling by plasma samples was visualized in orthogonal views (Fig 7). Some intracellular SFV and actin labelling was observed, likely reflecting partial membrane permeabilization induced either by paraformaldehyde fixation, desiccation, or virus-induced membrane remodeling. The plasma samples offered the possibility to complete the studies performed with the anti-SU mAb (Figs 3 and 4), which led to a signal that was insufficient to detect Env at the surface of infected cells (S4 Fig). Anti-SFV staining was generally patchy at the base of the cell and continuous at the cell apex (Fig 7A). Cytoplasm protruded from the apex or the side of infected cells, forming anuclear compartments with dense anti-SFV staining (Fig 7B). Anti-SFV labelling was more intense on the side of cells from which intercellular connections started, as well as at the connection itself (Fig 7C). Finally, SFV-infected cells displayed dense surface anti-SFV staining without polarization at the contact zone with adjacent cells, despite the concomitant presence of syncytia in the culture (Fig 7D). Overall, SFV infection induced extensive morphological changes of cells and SFV proteins were expressed at the cell surface.

Download:

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

Fig 6. Plasma samples specifically stain SFV-infected cells.

BHK-21 cells were infected with GI-D468 or GII-K74, at a moi of 0.05 for seven days, seeded onto glass coverslips and fixed within 1 to 3 days, according to the CPE (S2B Fig). Cells were either untreated (A-E) or permeabilized with 0.5% Triton X-100 (F-J) before staining with anti-GI plasma sample (LOBAK 2) diluted 1:100 and anti-human IgG-A-M-FITC and DAPI (nuclei). Cells were analyzed by brightfield microscopy with optical sectioning. A, F: Mock infected cells, B, D, G, I: GI-D468 infected cells, C, E, H, J: GII-K74 infected cells. Nuclei (blue), anti-SFV plasma (green), scale bars = 100 μm (10x magnification images) and 20 μm (63x magnification images). White arrows indicated dense SFV-specific staining localized either on the side or at the cell apex; yellow arrows indicate cytoplasmic staining.

https://doi.org/10.1371/journal.ppat.1010470.g006

Download:

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

Fig 7. Plasma antibodies stain large surfaces, long intercellular connections and cell protrusions on SFV-infected cells.

Cells were infected with GI-D468 or GII-K74 at a moi of 0.05. At the second passage, cells were seeded on glass coverslips and fixed within 1 to 3 days, according to the CPE (S2B Fig). Cells stained with anti-GI plasma sample (LOBAK2) and anti-human IgG-A-M-FITC, phalloidin (actin) and DAPI (nuclei) without incubation with Triton X-100. Labelled cells were analyzed by brightfield microscopy with optical sectioning. Acquired serial Z-plane frames at various Z depths with the same XY position were used to build 3D views with nuclei (blue), Env (green) and actin (red) shown in a blend-rendering mode. Env and actin costaining is indicated by yellow color. A. GII-K74-infected BHK-21 cell, 3D view from the side. The infected cell displays patchy SFV staining at its base, dense staining at the top and a small protrusion at the back. Yellow bars indicate the position of orthogonal slices. B. GII-K74-infected BHK-21 cells, 3D view from the top. The infected cell in the center of the image has a large cytoplasmic protrusion on its right side (yellow triangle). Yellow bars indicate the position of orthogonal slices. The protrusion has no nuclei (slices b-e), is connected to the adjacent cell (slides c and d) and does not adhere to the glass slide (slices b-e). C. GI-D468-infected HT1080 cells, 3D view from the top. An infected cell displays strong SFV staining at its surface, polarized in the direction of the intercellular connection and away from the nuclei. One densely stained protrusion is located at the center of the cell (white arrow). The long intercellular connection adheres to and wraps around the neighboring cell (yellow triangles), with extracellular SFV staining (white triangles). D. GII-K74-infected BHK-21 cells, 3D view from the top. Punctate SFV staining is observed at the basis of a syncytia (yellow triangle) and dense SFV staining at the top of cells (white triangles). Densely stained SFV protrusions (blue arrows) and several unstained cells adjacent to cells strongly expressing SFV (white arrows) are highlighted.

https://doi.org/10.1371/journal.ppat.1010470.g007

Plasma antibodies from SFV-infected individuals recognize SFV proteins expressed at the surface of infected cells

Then, we quantified the plasma antibodies bound to the surface of nonpermeabilized infected cells by labeling and fixing before flow cytometry analysis (Figs 8 and S2C). A global shift of fluorescence was observed in infected versus uninfected cultures. Therefore, the results are expressed as the ratio of mfi of infected versus mock-infected cultures. Immune plasma samples had ratios ranging from 1.7 to 6.2. Ratios were higher against GII-K74 than GI-D468, probably reflecting the higher infection rate. Modest background staining was observed with the plasma from an uninfected individual (ratio of 1.6). We confirmed nonspecific binding of plasma samples on uninfected cells (S5 Fig). However, samples from SFV-infected individuals showed higher staining intensity on infected cultures than on mock-infected cell cultures (Fig 8) and specific binding was dependent upon the moi of infection (as tested with CI-PFV, S5 Fig).

Download:

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

Fig 8. Plasma samples bind to the surface of SFV-infected cells.

BHK-21 cells were infected with GI-D468 or GII-K74 at a moi of 0.05, passed twice and stained when a CPE was visible (S2C Fig). A. Cells were permeabilized with 0.1% Triton X-100 and stained with anti-SU. Env-labelled and unlabeled cells appeared as distinct peaks and Env-labelled cells were quantified as their percentage among all cells. B. Nonpermeabilized cells were incubated with four 1:10 diluted plasma samples and anti-hIgG-BV421. Staining obtained with SFV^neg (MEBAK189), anti-GI (LOBAK2), anti-GII (MEBAK88) and anti-(GI+GII) (BAK177) plasma samples are shown on the histogram overlay: mfi is presented on the x-axis and frequency is expressed as the percentage of gated events on the y-axis. Extracellular staining with plasma samples corresponds to an increase in fluorescence intensity of the whole cell population. Therefore, SFV-specific staining was quantified by the ratio of mfi from infected to uninfected cultures.

https://doi.org/10.1371/journal.ppat.1010470.g008

We used a GFP-expressing CI-PFV molecular clone (S6 Fig) to distinguish infected from uninfected cells present in the same culture. Plasma samples were added to nonpermeabilized cells and SFV-specific binding was quantified by the ratio of the mfi of infected (GFP^pos) versus uninfected (GFP^neg) cells (S4D Fig). Plasma from SFV-infected individuals stained GFP-expressing cells, with stronger staining of cells expressing the highest level of GFP (Fig 9A). Twelve plasma samples were tested in at least three independent experiments. Only those from SFV-infected individuals bound to GFP^pos cells. The median mfi_GFPpos/mfi_GFPneg ratio ranged from 1.8 to 5.0 for anti-GI samples, 1.6 to 3.3 for anti-GII samples and 1 to 1.2 for uninfected samples (Fig 9B), supporting binding of the antibodies from plasma samples to SFV proteins on the cell surface.

Download:

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

Fig 9. Plasma samples bind to Env expressed at the surface of infected cells.

A. CI-PFV-GFP-infected BHK-21 cells were stained with 1:10 diluted plasma samples from SFV-infected individuals and anti-hIgG-BV421. Staining with human plasma samples was quantified by the mfi of GFP^neg and GFP^pos cells and the results are expressed as mfi_GFPpos/mfi_GFPneg ratios (S2D Fig). B. mfi_GFPpos/mfi_GFPneg ratios obtained at a 1:10 dilution for the plasma from 12 individuals: n = 4 SFV^neg (grey symbols), n = 8 infected with gorilla SFV belonging to genotype I (blue symbols) or genotype II (red symbols). Three to five experiments are presented per plasma sample; bars indicate median values. C. sENV-GFP-transduced BHK-21 cells were stained with 1:10 diluted plasma samples from SFV-infected individuals and anti-hIgG-BV421. Staining with human plasma samples was quantified by the mfi of GFP^neg and GFP^pos cells and the results are expressed as mfi_GFPpos/mfi_GFPneg ratios (S2E Fig). D. mfi_GFPpos/mfi_GFPneg ratios obtained at a 1:10 dilution for the plasma from 12 individuals: n = 4 SFV^neg (grey symbols), n = 8 infected with gorilla SFV belonging to genotype I (blue symbols) or genotype II (red symbols). Three to four experiments are presented per plasma sample; bars indicate median values.

https://doi.org/10.1371/journal.ppat.1010470.g009

Finally, we transduced BHK-21 cells with a plasmid encoding GI-D468 Env fused to GFP. We introduced mutations into the LP and TM subdomains that are homologous to those that overcome CI-PFV Env intracellular retention and enhance its expression at the plasma membrane (S2E and S7 Figs [39,48]). Plasma samples from SFV-individuals stained nonpermeabilized Env-transduced cells (i.e., GFP^pos cells, Fig 9C). The median mfi_GFPpos/mfi_GFPneg ratios ranged from 2.0 to 3.6 for anti-GI samples, 2.6 to 3.3 for anti-GII samples, and 1 to 1.1 for uninfected samples (Fig 9D). These experiments directly demonstrate the capacity of immune plasma samples to specifically bind to SFV Env expressed at the cell surface.

Live cell imaging of CI-PFV-tagged cell-to-cell transmission

Finally, we used the CI-PFV-GFP virus and live-cell imaging to obtain a dynamic view of SFV cell-to-cell transmission (S2F Fig). Contacts between infected and uninfected cells showed three major outcomes: cell fusion, transmission of productive infection to the uninfected cell without fusion, and neither fusion nor productive infection (S1 and S2 Movies). Selected sequential frames show representative fusion and non-fusion events between infected and uninfected cells (Figs 10A for BHK-21 cells and S8A for HT1080 cells; note the difference in scale and greater cell displacement in HT1080 than BHK21-cells). We readily observed SFV production after a nonfusogenic contact with an infected cell for BHK-21 cells (Fig 10B). Such an event was not observed for HT1080 cells, which show the highest susceptibility to SFV infection [30] and the highest cytopathic response (see below).

Download:

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

Fig 10. CI-PFV-GFP cell-to-cell transmission visualized by live-cell imaging.

CI-PFV-GFP infected BHK-21 cells were mixed with uninfected cells to obtain a GFP^pos cell frequency of 5% and cultivated in 96-well plates (3000 cells/well) inside an Incucyte device for 96 h (S2F Fig). Hourly acquisition was started within 30 min after seeding. We selected sequentially acquired images (one per hour) showing representative events in CI-PFV-GFP-infected cultures. A. Acquisition between 0D-14H and 0D-20H, 250 μm x 250 μm images with a superimposed 5 x 5 grid (50 μm between lines). A GFP^bright syncytium has established contact with uninfected cells; its orientation frequently shifted with the lamellipodium, often oriented towards an adjacent cell. The contacts between the syncytium and cellsto fusion (white triangles indicate cells that will be fused to the syncytium in the subsequent time frame) or not (blue triangles). B. Acquisition between 1D-0H to 1D-3H, 400 μm x 400 μm images with a superimposed 5 x 5 grid (80 μm between lines). A round GFP^bright cell is in contact with a layer of uninfected cells at 1D-0H. One hour later, one adjacent cell expressed GFP (white triangle) and the cell cluster has split (arrow). Three hours later, among the cluster which left the GFP^pos cells, one cell becomes GFP^low (yellow triangle). C. Acquisition between 1D-8H and 1D-15H, 250 μm x 250 μm images with a superimposed 5 x 5 grid (50 μm between lines). Morphology of a migrating infected mesenchymal, with a large lamellipodium that often splits, leading to changes in direction. The uropod appears to remain anchored, either to a cell that was in contact hours before (white dot) or to the substratum (pink dot). The former interaction can generate structures that resemble intercellular connections. One cell has remained in contact with the syncytium without fusion nor productive infection (blue triangle).

https://doi.org/10.1371/journal.ppat.1010470.g010

SFV-infected cells showed higher mobility than uninfected cells, in particular at high cell density. The net displacement of BHK-21 and HT1080 cells was < 50 μm over one hour and < 100–150 μm over 12 hours at low density (S9 Fig and S3 and S4 Movies). Similarly, most isolated SFV-infected cells showed motility but little net movement (Fig 10 and S1 and S2 Movies). In subconfluent cultures, syncytia moved rapidly over a distance of > 200 μm as they fused to other cells. Intercellular connections were mostly formed after cell contact and while cells were moving apart: infected cells maintained adhesion to focal points while their bodies moved away, leading to the formation of trailing membrane extensions that persisted over 1 to 4 hours (Figs 10C and S8B, and S1 and S2 Movies). In addition, cellular extensions appeared to correspond to areas of tight attachment of the uropod to the substrate (Figs 10C and S8B, dots). When quantified at the culture level, the epithelial BHK-21 cells formed few intercellular connections and SFV infection led to an increase in their number and length (Fig 11A–11C). Cell confluence was almost similar for infected and uninfected BHK-21 cells (Fig 11E). Fibroblastic HT1080 cells spontaneously formed intercellular connections (Fig 11B) and SFV-infected HT1080 cells emitted elongated structures (S8B Fig and S2 Movie). At the culture level, the confluence of infected cells was less than that of uninfected cultures, as was the number of intercellular connections (Fig 11B and 11F). The mean length of intercellular connections appeared to be similar for uninfected and SFV-infected cell cultures (Fig 11D). These data highlight the high susceptibility of HT1080 cells to the cytopathic effect induced by SFV. Overall, live-cell imaging showed that SFV is transmitted when adjacent cells establish contact over large surfaces, whereas intercellular connections are formed posterior to cell contacts.

Download:

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

Fig 11. Number and length of intercellular connections in SFV-infected BHK-21 and HT1080 cells.

CI-PFV-GFP-infected BHK-21 and HT1080 cells were mixed with uninfected cells to obtain a GFP^pos cell frequency of 5% and cultivated in 96-well plates (3000 cells/well) inside an Incucyte device for 96 h. Uninfected cells were seeded at the same density. For each condition, one 57 mm² field acquired at 20x magnification, at 4-h intervals over 96 h, was randomly selected to manually count intercellular connections and measure their size. Cell confluence in the same wells was quantified by Incucyte software hourly from 0 to 96 h and expressed as the percentage of the total well surface. Intercellular connections counted per field (A-B), their mean ± SD length (μm, C-D), and cell confluence (percentage, E-F) are presented for BHK-21 (A, C, E) and HT1080 cells (B, D, F). Mock infected cultures are presented in grey and SFV infected cultures in blue.

https://doi.org/10.1371/journal.ppat.1010470.g011