BANAL-236 and SARS-CoV-2 failed to replicate in RFe cells

To better understand the biology of BANAL-236 in Rhinolophus-derived cells, we first established a primary lung fibroblast cultures from the lung tissues of a R. ferrumequinum (RFe) bat collected in Spain [28]. The cells were then immortalized by expressing SV40-LT antigen [28]. To assess the expression of key viral entry factors, we performed RT-qPCR analysis to quantify endogenous ACE2 and TMPRSS2 mRNA levels in RFe cells, as well as in 293T and Caco-2 cells as comparisons. Consistent with previous reports [29], 293T and Caco-2 cells expressed moderate levels of ACE2, approximately10² copies per μg of total RNA (Fig 1A), whereas ACE2 mRNA was undetectable in RFe cells. Both 293T and RFe cells expressed less than 10 copies of TMPRSS2 per μg of total RNA. By contrast, Caco-2 cells expressed high levels of the protease, over 10³ copies per μg of total RNA (Fig 1A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. SARS-CoV-2 and BANAL-236 replicate in Caco-2 cells, but not in RFe cells.

(A) Quantification of copy numbers per μg of total cellular RNA of endogenously expressed ACE2 and TMPRSS2 in indicated cell lines via RT-qPCR analysis. Data are means ± SD of three independent experiments. Caco-2 cells (pink lines) were infected with SARS-CoV-2 (B-D) or BANAL-236 (C-E) at a multiplicity of infection (MOI) of 0.0002 and 0.02 respectively. Cells were treated or not with 1μg/ml of trypsin TPCK (dotted lines). RFe cells (dark green lines) were infected with SARS-CoV-2 (B-D) or BANAL-236 (C-E) at a MOI of 0.2 and 0.5 respectively and treated or not with 1μg/ml of trypsin TPCK (dotted lines). (B-C) The relative amounts of cell-associated viral RNA were measured by RT-qPCR analysis at different times post-infection and were expressed as genome equivalents (GE) per µg of total cellular RNAs. Data are the means ± SD of three independent experiments. (D-E) The percentages of cells expressing the viral N protein were determined by flow cytometry analysis. Data are means ± SD of three independent experiments.

https://doi.org/10.1371/journal.ppat.1013573.g001

Next, we examined the replication kinetics of BANAL-236 and SARS-CoV-2 (Wuhan strain) in RFe cells. Caco-2 cells, which are susceptible to both viruses [1,5], served as positive controls. Based on pilot experiments, Caco-2 cells were infected with 100 more infectious particles of BANAL-236 than SARS-CoV-2 to achieve comparable viral RNA yields. RT-qPCR analysis revealed that SARS-CoV-2 RNA levels in Caco-2 cells increased until 48 hpi, reaching about 10⁷ genome copies per μg of total RNA and remained stable through 72 hpi (Fig 1B). The growth kinetics of BANAL-236 were slower than those of SARS-CoV-2 in Caco-2 cells, showing a less rapid increase in RNA yield from the first to the second time point (Fig 1C). Flow cytometry analyses were used to further evaluate viral replication. Around 50% of Caco-2 cells were positive for the SARS-CoV-2 N protein at 48 hpi (Fig 1D), while around 30% of Caco-2 cells were expressing the N protein of BANAL-236 at 72 hpi (Fig 1E). Thus, BANAL-236 replicated less efficiently than SARS-CoV-2 in Caco-2 cells, despite the higher inoculum, a finding that contrasts with a previous study reporting comparable replication of BANAL-236 and SARS-CoV-2 (WK-521 strain) in these cells [5].

In RFe cells, infection with SARS-CoV-2 or BANAL-236 at a MOI of 0.2 and 0.5, respectively, did not result in increased viral RNA levels over time, suggesting an absence of replication (Fig 1B-1C). In agreement with these data, no N positive cells were detected by flow cytometric analysis (Fig 1D-1E). The lack of viral replication may be due to the absence of endogenous ACE2 and/or a low TMPRSS2 (Fig 1A). To activate the S protein and allow viral fusion independently of TMPRSS2 and other proteases, viral input was treated with the serine-protease trypsin [29]. Pre-activation of SARS-CoV-2 or BANAL-236 with trypsin did not enhance viral RNA yield or the percentage of RFe cells expressing the N protein (Fig 1B-1E), suggesting that S protein cleavage is not the limiting factor for infection in RFe cells. The absence of viral RNA and protein production in RFe cells may therefore result from a lack of pro-viral factors, such as ACE2 (Fig 1A) and/or the presence of potent antiviral mechanisms.

Efficient entry of BANAL-236 and BANAL-52 in RFe cells requires both hACE2 and hTMPRSS2

To investigate viral entry mechanisms and ensure reliable detection of human and Rhinolophus ACE2 and TMPRSS2, plasmids expressing tagged versions of the four proteins (r/hACE2-V5 and myc-h/rTMPRSS2) were produced. Since BANAL-236 and BANAL-52 were isolated from R. marshalli and R. malayanus, respectively [1] and the R. marshalli genome is unavailable, rACE2 was synthetized from R. malayanus sequence. Given the endogenous TMPRSS2 expression in RFe cells (Fig 1A), rTMPRSS2 was cloned from R. ferrumequinum cDNA to avoid having cells expressing TMPRSS2 from 2 species. The expression levels of the tagged proteins were assessed by western blot analysis in transfected 293T cells using antibodies against hACE2, hTMPRSS2, V5 or myc. Antibodies against hACE2 recognized as efficiently rACE2 and hACE2 (S1A Fig), while antibodies against hTMPRSS2 [30] did not detect its Rhinolophus ortholog (S1B Fig). Immunofluorescence experiments performed in transfected 293T cells using antibodies against hACE2 and/or myc revealed that the entry factors were well expressed (S1C Fig), consistent with our western blot analysis (S1A-S1B Fig). The 4 proteins localised to the cytoplasm. An overlap between the signals of the 2 specifies-specific pairs of proteins was observed, suggesting a distribution consistent with a potential interaction.

To assess whether the V5 tag interferes with ACE2 receptor function, we performed SARS-CoV-2 infection experiments in human A549 lung cells. SARS-CoV-2 does not replicate in wild-type A549 cells unless they express hACE2 [25]. We established A549 cells stably expressing hACE2-V5 (S1D Fig). Flow cytometry analysis using an anti-S antibody at 24 hpi (MOI 1) revealed that approximately 12% of hACE2-V5-expressing cells were positive for the S protein (S1E Fig), indicating that viral S protein binds hACE2-V5. Our previous experiments showed that around 3% of A549 expressing untagged hACE2 were expressing the SARS-CoV-2 S protein after 24 hours of infection with the same MOI [25]. Thus, the V5 tag does not interfere with hACE2 receptor function.

Since hACE2 is a known substrate of hTMPRSS2 [31,32], we performed Western blot analysis to determine whether myc-tagged hTMPRSS2 and rTMPRSS2 retain their ability to cleave hACE2. 293T cells were transfected with hACE2 and increasing amount of myc-TMPRSS2. In the absence of the protease, hACE2-V5 appeared as 2 bands of about ~125-kDa (S1F Fig), likely representing differentially glycosylated forms. In the presence of myc-hTMPRSS2, the heavier form of hACE2-V5 was not detectable anymore, while the ~120-kDa form remained visible. Additionally, a ~20-kDa ACE2 fragment appeared (S1F Fig), consistent with cleavage by myc-hTMPRSS2. These results align with previous studies performed in 293T cells expressing untagged hACE2 and hTMPRSS2 [31,32]. Identical hACE2 patterns were observed with myc-rTMPRSS2 (S1F Fig). Together, these results indicate that myc-tagged human and Rhinolophus TMPRSS2 retain protease activity and efficiently process hACE2.

To perform viral entry assays, we generated 293T cells stably expressing hACE2-V5 (hA), myc-hTMPRSS2 (hT), or both factors (hAT). Similarly, we generated 293T cells expressing rACE2-V5 (rA), myc-rTMPRSS2 (rT), or both factors (rAT). Flow cytometry analysis revealed that 20–40% of 293T cells were expressing hACE2-V5 (hA) at their surface (Fig 2A) while 40–80% were expressing myc-hTMPRSS2 (hT) at their surface (Fig 2B). The surface expression level of both factors was below the detection limit in wild-type cells (Fig 2A-2B). RT-qPCR analysis of the levels of exogenous hACE2 and hTMPRSS2 mRNAs in 293T cells were consistent with the flow cytometric analysis (S2A-S2B Fig). Flow cytometry showed that ~70% of 293T-rA cells expressed rACE2-V5 and ~40% of 293T-rAT expressed rACE2-V5 (Fig 2C). Around 40% of 293T-rT and 293T-rAT expressed myc-rTMPRSS2 (Fig 2D). RT-qPCR analysis of the abundance of rACE2 and rTMPRSS2 mRNAs in 293T cells were consistent with the flow cytometric analysis (S2D-E Fig). ACE2 expression in 293T cells was reduced in the presence of TMPRSS2, as compared to when it was expressed on its own (Figs 2A, 2C, S2A and S2D), likely due to the ability of the protease to cleave ACE2 (S1F Fig) [31,32].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Pseudovirus assays revealed that efficient entry of BANAL-236 and BANAL-52 in RFe cells requires both hACE2 and hTMPRSS2.

(A-B) Cell surface expression of hACE2-V5 (A) and myc-hTMPRSS2 (B) were determined by flow cytometry analysis prior to the entry assays shown in (E) in wt 293T cells (-) or expressing hACE2-V5 (hA), Myc-hTMPRSS2 (hT) or both entry factors (hAT). Data are the means ± SD of three independent experiments. (C-D). Cell surface expression of rACE2-V5 (C) or Myc-rTMPRSS2 (D) were determined by flow cytometry analysis prior to the entry assays shown in (E) in wt 293T cells (-) or expressing rACE2-V5 (rA), myc-rTMPRSS2 (rT) or both entry factors (hAT). Data are the means ± SD of three independent experiments. (E) Entry assays were performed in wt 293T cells (-) or stably expressing hACE2-V5 (hA), myc-hTMPRSS2 (hT), rACE2-V5 (rA), myc-rTMPRSS2 (rT), or both entry factors (hAT or rAT). Cells were transduced with the same HIV-1 p24 quantity of pseudo-lentiviruses bearing the S proteins of SARS-CoV-2 (purple), BANAL-236 (beige) or BANAL-52 (grey). Results are expressed in relative luminescence units (RLU). The dashed line indicates the average RLU obtained with cells transduced with empty vectors (white). Data are means ± SEM of three independent experiments. (F-H) Cell surface expression of hACE2 (F), hTMPRSS2 (G) or rACE2-V5 (H) were determined by flow cytometry analysis prior to the entry assays shown in (I) in wt RFe cells (-) or expressing hACE2 (hA), hTMPRSS2 (hT), both entry factors (hAT) or rACE2-V5. Data are the means ± SD. (I) Entry assays were performed in wt RFe cells (-) or stably expressing hACE2 (hA), hTMPRSS2 (hT), both human entry factors (hAT), or rACE2-V5 (rA). Cells were transduced with the same HIV-1 p24 quantity of pseudo-lentiviruses bearing the S proteins of SARS-CoV-2 (purple), BANAL-236 (beige) or BANAL-52 (grey). Results are expressed in relative luminescence units (RLU). The dashed line indicates the average RLU obtained with cells transduced with empty vectors (white). Data are the means ± SEM of three independent experiments.

https://doi.org/10.1371/journal.ppat.1013573.g002

Lentiviruses expressing the S proteins of SARS-CoV-2 (Wuhan strain), BANAL-236 or BANAL-52 were generated, as well as lentiviruses with no S, which served as negative controls. All three S proteins facilitated entry of pseudovirions into 293T-hACE2 cells (Fig 2E), but not into wild-type 293T. Expression of hTMPRSS2 alone was insufficient for efficient viral entry, although co-expression with hACE2 slightly enhanced S-mediated entry (Fig 2E). Expression of Rhinolophus entry factors, alone or in combination, did not support SARS-CoV-2 or BANAL-236 S-mediated entry in 293T cells (Fig 2E). Expression of rACE2 alone allowed entry of lentiviruses expressing the S proteins of BANAL-52 in 293T (Fig 2E), although not as efficiently as hACE2. This result was unexpected given that BANAL-52 was originally detected in R. malayanus. When myc-rTMPRSS2 was expressed together with rACE2-V5 in 293T cells, S-mediated entry of BANAL-52 was less efficient than in 293T cells expressing rACE2 alone (Fig 2E). This is likely due to the lower expression of rACE2 in 293T-rAT cells compared to 293T-rA cells (Figs 2C and S2D).

To investigate viral entry mechanisms in another cellular models, experiments were conducted in hamster BHK-21 kidney cells, which are commonly used for these assays [33]. Cells were transfected with the tagged versions of the human and Rhinolophus entry factors and then transduced with S-bearing lentiviruses (S2E Fig). Pilot experiments conducted with plasmids expressing GFP revealed that only a small percentage (around 5%) of cells were GFP positive. This discrepancy may explain why luciferase activity was approximately 2 logs lower in BHK-21 cells compared to 293T cells (S2E Fig). Despite a low transfection efficiency, the results showed that expression of hACE2 in BHK-21 cells promoted efficient entry of lentiviruses bearing the S proteins of all 3 viruses (S2E Fig), consistent with the data obtained in 293T cells (Fig 2E). However, unlike in 293T cells (Fig 2E), co-expression of hTMPRSS2 with hACE2 in BHK-21 cells did not enhance the entry of the lentiviruses, as compared to cells expressing hACE2 alone (S2E Fig). Additionally, expression of rACE2 alone did not facilitate entry of lentiviruses expressing the S protein of BANAL-52 in BHK-21 cells. These results are likely due to a low transfection (and possibly transduction) efficiencies in BHK-21 cells. Nevertheless, despite the limitations of this cellular model, its use revealed that expression of hACE2 is necessary and sufficient to promote the entry of lentiviruses expressing the S proteins of the 3 sarbecoviruses in 293T and BHK-21 cells. Of note, the inability of rACE2-V5 to mediate the entry of lentiviruses bearing the SARS-CoV-2 S protein in 293T (Fig 2E) and BHK-21 cells (S2E Fig) explained why expression of rACE2-V5 did not support SARS-CoV-2 replication in A549 cells (S1D-S1E Fig).

Finally, to perform entry assays in Rhinolophus cells, we generated RFe cells stably expressing hACE2, rACE2, hTMPRSS2 or both human factors. Unfortunately, repeated attempts to produce RFe cells stably expressing rTMPRSS2, alone or in combination with rACE2, were unsuccessful. Untagged versions of the human entry factors were used in these experiments. Flow cytometric analysis revealed that approximately 30% of RFe-hA and RFe-hAT cells expressed hACE2 at their surface (Fig 2F), despite RFe-hAT cells expressing slightly higher levels of hACE2 mRNA than RFe-hA cells (S2F Fig). Around 40% of RFe-hT and 90% of RFe-hAT expressed hTMPRSS2 at their surface (Fig 2G), despite exhibiting similar abundance of hTMPRSS2 mRNA (S2G Fig), suggesting a more efficient plasma membrane addressing of hTMPRSS2 in RFe-hAT than in RFe-hT. Finaly, about 90% of RFe-rACE2 cells expressed rACE2-V5 (Figs 2H and S2H). Expression of hACE2 or hTMPRSS2 alone in RFe cells did not support efficient viral entry (Fig 2I). Similarly, no viral entry was observed in RFe-rA cells (Fig 2I), suggesting that expression of rACE2 alone is insufficient to mediate S-dependent entry in RFe cells, as observed in 293T cells (Fig 2E). Only co-expression of both human entry factors enabled efficient viral entry in RFe cells (Fig 2I). These results suggest that the undetectable levels of ACE2 and low TMPRSS2 levels in RFe cells (Fig 1A) contribute to the absence of viral replication.

Given that RFe-hAT cells supported efficient entry of lentiviruses pseudotyped with the BANAL-236 S protein (Fig 2I), we selected them for further investigation.

BANAL-236 and SARS-CoV-2 replicate in RFe-ATC cells

We assessed the ability of BANAL-236, and, for comparison, SARS-CoV-2, to replicate in RFe-AT cells. Entry assays showed that these cells supported entry of lentiviruses bearing the S proteins of both viruses (Fig 2I). Despite maintaining RFe cells under antibiotic selection, the expression of hACE2 and hTMPRSS2 was unstable. Therefore, the cell surface expression of hACE2 and hTMPRSS2 was evaluated during each experiment by flow cytometry (Fig 3A). Around 20% of RFe-AT cells were positive for hACE2 and 50% for hTMPRSS2 (Fig 3A). RFe-AT cells were infected with SARS-CoV-2 or BANAL-236 at a MOI of 0.2. RT-qPCR analysis revealed that viral RNA levels did not increase over time (Fig 3B-3C), suggesting an absence of viral replication. Although not significant, a modest increase of BANAL-236 and SARS-CoV-2 RNA was observed at late time post-infection in RFe-AT cells treated with TPCK compared to untreated cells (Fig 3B-3C), suggesting that trypsin activation of the S protein may enable a small fraction of viruses to enter cells. Consistent with the RT-qPCR results (Fig 3B-3C), flow cytometry detected no N protein expression in RFe-AT cells, even in the presence of trypsin (Fig 3D-3E). This suggests that any viruses that entered the cells were blocked at an early stage of the replication cycle. Collectively, these data demonstrate that, although hACE2 and hTMPRSS2 expression permits S-mediated entry in RFe-AT cells (Fig 2I), it is insufficient to support viral replication. Of note, SARS-CoV-2 rapidly adapted to replicate in RFe-AT cells by acquiring a single point mutation in N [28].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. RFe-ATC cells, but not RFe cells, are permissive to SARS-CoV-2 and BANAL-236.

(A) Expression of hACE2 at the cell surface of RFe-hACE2-hTMPRSS2 cells (RFe-AT, in blue) or clonal cells (RFe-ATC, in light green) were determined by flow cytometry analysis. Data are means ± SD of three independent experiments. RFe-AT (blue lines) and RFe-ATC (green lines) cells were infected with SARS-CoV-2 (B-D) or BANAL-236 (C-E) at a multiplicity of infection (MOI) of 0.2 and 0.5, treated or not with 1μg/ml of trypsin TPCK (dotted lines). Cells were collected at the indicated hours post-infection (HPI). (B-C) The relative amounts of cell-associated viral RNA were measured by RT-qPCR analysis and were expressed as genome equivalents (GE) per µg of total cellular RNAs. Data are the means ± SD of three independent experiments. (D-E) The percentages of cells expressing the viral N protein were determined by flow cytometry analysis. Data are means ± SD of three independent experiments. (F) hACE2 and hTMPRSS2 expression in RFe, RFe-AT, and RFe-ATC cells. Whole-cell lysates were analysed by western blotting with antibodies against the indicated proteins. Data are representative of two independent experiments. (G) RFe, RFe-AT and RFe-ATC cells were infected with VSV at a MOI of 0.05 for 16 hours. The percentages of cells expressing the viral G protein were determined by flow cytometry analysis. Data are means ± SD of three independent experiments.

https://doi.org/10.1371/journal.ppat.1013573.g003

Due to the unstable expression of hACE2 and hTMPRSS2 in RFe-AT cells, we produced over 30 clonal cell lines. Among these, only one clone (RFe-ATC) exhibited a cytopathic effect following SARS-CoV-2 infection. Flow cytometry analysis revealed that approximately 25% of RFe-ATC cells expressed hACE2 at their surface, a proportion comparable to that observed in RFe-AT cells (Fig 3A). In contrast, a higher proportion of RFe-ATC cells (around 90%) expressed hTMPRSS2 at their surface (Fig 3A), despite similar overall expression levels (Fig 3F). RFe-ATC cells were infected with SARS-CoV-2 or BANAL-236 at a MOI of 0.2 and 0.5, respectively. RT-qPCR analyzes showed that SARS-CoV-2 RNA levels increased by approximately 1-log during the first 20 hours of infection, followed by a plateau (Fig 3B). In contrast, BANAL-236 RNA levels rose steadily between 20- and 72-hours post-infection in RFe-ATC cells (Fig 3C). Flow cytometry analysis showed that a small percentage of RFe-ATC cells were positive for SARS-CoV-2 N protein at 24 hpi, increasing to 60% at 72 hpi (Fig 3D). Consistently with the viral RNA yield data (Fig 3C), BANAL-236 replication in RFe-ATC cells was delayed compared to SARS-CoV-2 (Fig 3E). Taken together, these results demonstrate that both SARS-CoV-2 and BANAL-236 are capable of replicating in RFe-ATC cells.

To determine whether RFe parental and RFe-AT were generally resistant to viral infection, we infected them with Vesicular Stomatitis virus (VSV), a negative-strand RNA virus from the Rhabdoviridae family that is known for its broad cell tropism [34]. Flow cytometry analysis using an antibody against the VSV G protein revealed that VSV replicated similarly in RFe, RFe-AT and RFe-ATC cells (Fig 3G), suggesting that the inability of SARS-CoV-2 and BANAL-236 to replicate in RFe and RFe-AT cells is not due to a pre-existing antiviral state.

The high surface expression of hTMPRSS2 in RFe-ATC cells (Fig 3A) may underlie their susceptibility to SARS-CoV-2 and BANAL-236. To test this hypothesis, we used siRNAs to knock-down hTMPRSS2 expression, selecting a concentration that reduced the proportion of surfaced hTMPRSS2-positive RFe-ATC cells (Fig 4A), to levels comparable to those in RFe-AT cells (Fig 3A). Under these conditions, BANAL-236 and BANAL-52 S-mediated entry was modestly, but reproducibly, less efficient in RFe-ATC cells expressing reduced levels of surface hTMPRSS2, as compared to cells transfected with control siRNAs (Fig 4B). Viral replication in the presence of reduced levels of hTMPRSS2 was then assessed by flow cytometry by measuring the number of N protein-positive cells at 72 hpi. Viral replication was completely abolished in these cells, as compared to cells treated with control siRNAs (Fig 4C), suggesting a strong dependence on hTMPRSS2. This pronounced effect likely reflects the cumulative impact of reduced hTMPRSS2 expression over multiple rounds of replication, in contrast to the modest effect observed during a single-entry event. Thus, the susceptibility of the RFe-ATC cells to viral infection may be attributed to their high level of surface hTMPRSS2 expression, as compared to RFe-AT cells.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. BANAL-236 replication in RFe-ATC cells is dependent on high cell surface TMPRSS2 expression.

(A) RFe-ATC cells were transfected with siRNAs targeting TMPRSS2 or control siRNAs (siCTL) for 3 days. The percentages of cells expressing hTMPRSS2 at their surface were determined by flow cytometry analysis. Data are the means ± SD of four independent experiments. T-tests two-tailed were performed **: p < 0.01. (B) Entry assays were performed with lentiviruses pseudotyped with the S proteins of SARS-CoV-2 (purple), BANAL-236 (beige) or BANAL-52 (grey). Cells were transduced with the same HIV-1 p24 quantity of pseudo-lentiviruses. Results are expressed in relative luminescence units (RLU). The dashed line indicates the average RLU obtained with cells transduced with empty vectors. (C) Cells were infected with SARS-CoV-2 and BANAL-236 at a multiplicity of infection (MOI) of 0.02 and 0.2, respectively. Seventy-two hours later, cells were collected and the percentages of cells expressing the viral N protein were determined by flow cytometry analysis. Data are means ± SD of at least three independent experiments.

https://doi.org/10.1371/journal.ppat.1013573.g004

SARS-CoV-2 and BANAL-236 complete their replication cycles in RFe-ATC cells

To further characterize SARS-CoV-2 and BANAL-236 replication in RFe-ATC cells, we performed transmission electron microscopy on cells infected for 72 h, with Caco-2 cells included for comparison and non-infected cells as negative controls (S3 Fig). In infected Caco-2 cells, we observed numerous viral replication factories composed of double-membrane vesicles (DMVs) (Fig 5A-5B), as previously described in SARS-CoV-2 infected Vero cells [35], although the DMVs in these cells displayed slight structural differences consistent with variations reported across different cell lines [36]. These DMVs, which likely derive from ER membranes and autophagic processes, serve as sites for viral RNA synthesis [36,37]. Fully assembled virus particles were detected within large intracellular vesicles (Fig 5C), some of which fused with the plasma membrane (Fig 5D), suggesting virion release via exocytosis. Many virions remained attached to the surface of infected cells (Fig 5D), possibly due to high hACE2 expression at the plasma membrane. RFe-ATC cells infected with SARS-CoV-2 exhibited similar ultrastructural changes (Fig 5E-H), including numerous DMVs. Large vacuoles containing virions (Fig 5G) and surface-bound virions (Fig 5H) were also present in RFe-ATC cells. In contrast, Caco-2 and RFe-ATC cells infected with BANAL-236 contained very few DMVs (Fig 5I-5P), with only occasional DMV-like structures observed (Fig 5N, white arrows). In addition, large vacuoles filled with dozens of BANAL-236 virions were observed in both cell lines (Fig 5J and 5O), some of which were connected to the plasma membrane (Fig 5P). The presence of virions at the surface of RFe-ATC cells (Fig 5N and 5P) suggested that BANAL-236 completed its replication cycle in these bat cell line.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. SARS-CoV-2 and BANAL-236 complete their replication cycle in RFe-ATC cells.

(A-P) Caco-2 and RFe-ATC cells were infected for 72 h. Caco-2 cells were infected with SARS-CoV-2 (A-D) or BANAL-236 (I-L) at an MOI 0.001 and 0.01 respectively. RFe-ATC cells were infected with SARS-CoV-2 (E-H) and BANAL-236 (M-P) at MOI 0.02 and 0.2 respectively. Cells were then subjected to transmission electron microscopy analysis. DMV-like structures are pointed by white arrows. (Q-R) At 48 and 72 h post-infection supernatants from Caco-2 (Q) and RFe-ATC (R) cells were collected. Titration of clarified supernatants by TCID₅₀ assays was performed on Vero-E6 cells infected for 5 days. Data are the means ± SEM of at least two independent experiments.

https://doi.org/10.1371/journal.ppat.1013573.g005

To assess the infectivity of virions observed at the cell surface, supernatants from infected RFe-ATC and Caco-2 cells were titrated on Vero-E6 cells (Fig 5Q-R). Supernatants from Caco-2 cells contained approximately 5.10⁵ TCID₅₀/mL of infectious BANAL-236 or SARS-CoV-2 particles (Fig 5Q). Similarly, about 10⁵ TCID₅₀/mL were retrieved from the supernatant of infected RFe-ATC cells (Fig 5R). These results confirmed that both viruses complete their replication cycles in RFe-ATC cells.

BANAL-236 antagonizes the IFN response in Rhinolophus and human cells

Given the critical role of viral IFN antagonism in overcoming species barrier, we evaluated ISG induction in RFe cells and their derivatives following infection with BANAL-236 or SARS-CoV-2. We compared mRNA abundance of OAS1 and ISG20, two ISGs that are conserved across vertebrate species [38], in infected cells. No upregulation of OAS1 or ISG20 expression was observed in RFe nor RFe-AT cells at 72 hpi (Fig 6A-6B), which was consistent with the absence of viral replication in these cells (Figs 1 and 3). Despite robust replication of BANAL-236 and SARS-CoV-2 in Caco-2 and RFe-ATC cells (Figs 1 and 3), no induction of OAS1 or ISG20 expression was observed either at 72 hpi (Fig 6A-6B). Similar results were obtained when the mRNA abundance of these two ISGs were evaluated in cells infected for 4, 24, and 48 hours (S4 Fig). These results suggest that both viruses efficiently counteract the IFN response in these cell lines.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. SARS-CoV-2 and BANAL-236 replication failed to induce the expression of Interferon-stimulated genes in Caco-2 and RFe-ATC cells.

(A-B) Caco-2 (pink), RFe (dark green), RFe-AT (blue) and RFe-ATC (grey) cells were infected with SARS-CoV-2 or BANAL-236 at a multiplicity of infection (MOI) of 0.2 and 0.5, respectively, for 72 hours. Alternatively, cells were treated with 100-500 ng of poly I:C for 16 hours. The relative amounts of ISG20 (A) and OAS1 (B) mRNAs were determined by RT-qPCR analysis. Results were first normalized to GAPDH mRNA and then to mRNA levels of control cells (mock-treated or non-treated cells), which were set at 1. Data are means ± SD of at least two independent experiments. (C-D) Volcano plots representing differentially expressed genes in RFe-AT (C) and RFe-ATC (D) cells upon polyI:C treatment (100 ng for 16 hours) compared to mock (PBS) transfected cells. Differentially expressed genes were selected using a log2 fold-change > 1.5 with and adjusted pvalue ≤ 0.05. (E-F) Top 10 enriched GO terms for Biological Process ranked by the adjusted p-value, for upregulated genes identified in RNA-Seq comparison comparing poly I:C treated cells to mock (PBS) transfected condition for RFe-AT (E) and RFe-ATC (F). Genes were defined as upregulated if the log2 fold change was equal or above 1.5 with a significance of adjusted pvalue below 0.05 compared to mock (PBS) transfected condition. (G) Caco-2 cells were infected for 72 hours with SARS-CoV-2 or BANAL-236 at a MOI of 0.0002 and 0.02, respectively, or left non-infected (NI). They were stimulated with IFN⍺2 at 1 000 IU/ml (lower panel) or mock-treated (upper panel) 16 hours before fixation. They were stained with antibodies recognizing IFITM3 (green) and the viral nucleocapsid N (red). Nucleus were counterstained using Hoechst dye (blue). Images are representative of three independent experiments. Scale bars, 40 μm. (H) Percentages of IFITM3‐positive cells among cells expressing the viral N proteins were estimated by analysing at least 40 cells per condition. For uninfected (NI) cells, at least 60 cells were randomly picked. Data are mean ± SD. One‐way ANOVA tests with Dunnett’s correction were performed. ****P < 0.0001.

https://doi.org/10.1371/journal.ppat.1013573.g006

To ensure that RFe cells were immunocompetent, we first assessed ISG induction upon stimulation with polyI:C, a synthetic dsRNA analog. The mRNA levels of OAS1 and ISG20 increased upon stimulation in all four cell lines (Fig 6A-6B), suggesting intact IFN-induction and -signaling pathways. We then conducted a transcriptomic analysis of RFe-AT and RFe-ATC cells following stimulation with polyI:C. The analysis identified 181 differentially expressed genes (DEGs)(fold change ≥1.5 log2, p-value < 0.05) in stimulated RFe-AT cells compared to unstimulated control cells, and 166 DEGs in stimulated ATC cells (Figs 6C-6D and S5A, and S1 Table). A substantial proportion of the upregulated genes were shared between both cell lines, including ISGs that are conserved among mammalian species (such as BST2, MX1, and IRFs) (S5A Fig and S1 Table). Notably, OAS1 and ISG20 were among the most highly upregulated genes (Fig 6C-6D). To identify pathways affected by poly(I:C) treatment, we performed Gene Ontology term analysis on the upregulated genes. As expected, both cell lines exhibited significant enrichment in pathways related to viral infection and IFN responses (Fig 6E-6F). These findings indicate that both cell lines are similarly immunocompetent. Finally, we infected both RFe-ATC and RFe-AT cells with Japanese Encephalitis Virus (JEV), an orthoflavivirus known to circulate in multiple bat species, including Rhinolophus species [39]. Cells were infected with JEV at an MOI of 10, and flow cytometric analysis revealed that approximately 30% of cells were positive for the viral E protein at 24 hpi (S5B Fig). RT-qPCR analysis showed a modest, yet reproducible, increase in the mRNA levels of OAS1 and ISG20 following JEV infection in both cell lines (S5C-S5D Fig). This modest increase, compared to that observed in polyI:C-treated cells (Fig 6A-6B), is likely due to the ability of JEV to counteract the IFN response [40,41]. Together, these results provide evidence that both cell lines exhibited similar immunocompetence following a viral infection or polyI:C treatment. Thus, the absence of IFN induction during BANAL-236 replication is not due to a dysregulation of the IFN response in RFE-ATC cells, but rather reflects the virus’s ability to actively counteract this response.

To investigate the ability of the 2 viruses to antagonize IFN signalling in individual cells, we assessed IFITM3 expression, another ISG that is conserved across vertebrate species [38], using fluorescent microscopic assays in Caco-2 cells (due to the lack of suitable antibodies for Rhinolophus ISGs). IFITM3 was detected in about 10% of unstimulated Caco-2 cells, but in approximately 45% of cells treated with IFN⍺2 (Fig 6G-6H). In cells infected for 72 hours, IFITM3 was detected in fewer than 5% of N-positive cells (Fig 6G-6H). When infected cells were treated with IFN⍺2 16 hours before fixation (Fig 6G-6H), fewer than 10% of cells were positive for both N and IFITM3, confirming that SARS-CoV-2 or BANAL-236 replication limits ISG expression, as observed in the RT-qPCR analysis (Fig 6A-6B).

To identify BANAL-236 proteins capable of antagonizing the IFN response in human cells, the open reading frames (ORFs) of all 30 viral proteins were cloned downstream of a N‐terminal 3X‐FLAG tag and expressed in 293T cells, which are easy-to-transfect cells. Western blot analyses confirmed the expression of 27 BANAL-236 proteins at their expected sizes (Fig 7A), including NSP3, the largest coronavirus protein (~220 kDa). A second ~80 kDa NSP3 band was detected, as previously reported for SARS-CoV-2 NSP3 [19]. NSP3 is composed of around 15 domains, including the papain-like protease [42], and this extra band could represent an autoproteolytic product. FLAG-ORF8 was poorly expressed, and FLAG-S was unproperly processed. We thus generated ORF8 and S sequences with a C-terminal STREP-tag. ORF8-STREP was detected at the correct size (Fig 7A). Both the uncleaved precursor and the dissociated S2 fragment of S-STREP were detected, as expected [43]. Due to its small size (1.5 kDa), NSP11 was undetectable and was thus excluded from the analysis.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. BANAL-236 antagonizes the IFN response in human cells.

(A) Viral protein expression in 293T cells. 293T cells were mock‐transfected (NT), transfected with empty plasmids (EV) or with plasmids encoding FLAG or STREP‐tagged BANAL-236 viral proteins. Cells were harvested 24 h post‐transfection and protein expression was assessed by Western blotting with anti‐FLAG, anti-STREP and anti‐GAPDH antibodies. Data are representative of three biological replicates. (B) 293T cells were co‐transfected for 24h with the Firefly luciferase reporter plasmid p‐ISRE‐luc, TK Renilla luciferase control plasmid phRluc‐TK, △RIGI plasmid and plasmids encoding individual viral proteins of BANAL-236. Empty vectors (EV) and plasmids encoding the NS1 of Flu protein were used as negative and positive controls, respectively. The data were analyzed by first normalizing the Firefly luciferase activity to the Renilla luciferase (Rluc) activity and then to EV samples, which were set at 100%. Data are means ± SD of three biological replicates. One‐way ANOVA tests with Dunnett’s correction were performed compared to EV non‐significant are not shown, *P < 0.1, **P < 0.01, ***P < 0.001, ****P < 0.0001 (C) 293T cells were co‐transfected with the Firefly luciferase reporter plasmid p‐ISRE‐luc, TK Renilla luciferase control plasmid phRluc‐TK, and plasmids encoding individual viral proteins of BANAL-236. Empty vectors (EV) and plasmids encoding the NS5 of TBEV protein were used as negative and positive controls, respectively. Cells were stimulated 7 h post‐transfection with IFN⍺2 at 200 IU/ml and assayed for luciferase activity 16 h later. The data were analysed by first normalizing the Firefly luciferase activity to the Renilla luciferase (Rluc) activity and then to EV samples, which were set at 100%. Data are means ± SD of three biological replicates. One‐way ANOVA tests with Dunnett’s correction were performed compared to EV non‐significant are not shown, *P < 0.1, **P < 0.01, ****P < 0.0001. (D) RFe cells stably expressing BANAL-236 NSP13 and ORF6 proteins upon lentiviral transduction. Cells transduced with empty lentiviruses (mock) or lentiviruses expressing GFP were used as negative controls. Whole-cell lysates were analyzed by western blotting with antibodies against the indicated proteins. Data are representative of two independent experiments. (E-F) The relative amounts of ISG20 (E) and OAS1 (F) mRNAs levels determined by RT-qPCR analysis after poly I:C treatment for 16 hours. Results were first normalized to GAPDH mRNA and then to mRNA levels of mock (PBS) transfected control cells. Data are means ± SD of at least three biological replicates. One‐way ANOVA tests with Tukey’s correction were performed. ns: non-significant, * P < 0.1, ****P < 0.0001.

https://doi.org/10.1371/journal.ppat.1013573.g007

The effect of BANAL-236 proteins on the activation of the IFN-stimulated response elements (ISREs) was investigated in 293T cells overexpressing a constitutively active form of RIG-I (ΔRIG-I), a potent inducer of IFN production [44]. ISREs can be activated directly by transcription factors of the IFN induction pathway, like IRF3, in an IFN-independent manner [45]. Cells transfected with an empty plasmid (EV) served as references. Cells expressing a FLAG‐tagged version of the NS1 of influenza A virus (IAV), a known IFN antagonist [46], served as positive controls. Its expression indeed reduced ISRE activation by over 100-fold (Fig 7B). TBEV NS5, which inhibits the JAK-STAT pathway [47], had no effect on ISRE activation (Fig 7B), confirming that ISRE was activated in an IFN-independent manner in our experimental settings. Four BANAL-236 proteins diminished the activity of the ISRE by at least10-fold: NSP1, NSP6, NSP13, and ORF6 (Fig 7B), mirroring the antagonistic effects of their SARS-CoV-2 counterparts detected in similar experiments [16,17,20], despite sequence differences (S2 Table). To determine whether BANAL-236 proteins could inhibit IFN-I signalling, we repeated the screen in 293T cells stimulated with IFN⍺2 (Fig 7C). As expected [47], TBEV NS5 inhibited ISRE activation by around 10-fold (Fig 7C). NSP1, NSP3, NSP15 and S modestly suppressed the JAK-STAT pathway (Fig 7C). By contrast, SARS-CoV-2 expresses several proteins that potently block ISRE activation in IFN-I treated 293T cells, such as NSP1, NSP6, NSP13, NSP14, ORF6 and ORF7b [16,17,19,20]. Together, these findings showed that BANAL-236 is armed with mechanisms to evade the IFN response in human cells, albeit with a potential narrower repertoire of IFN-I signalling antagonists compared to SARS-CoV-2.

Since NSP1, NSP13 and ORF6 were the most efficient at inhibiting the IFN response in 293T cells (Fig 7B), we tested their impact on IFN response in RFe cells. Pilot experiments performed with plasmids encoding for GFP and several transfection reagents showed that the highest percentage of GFP-positive cells achieved was only 5%, which is insufficient to reliably assess the impact of viral proteins on IFN responses in these cells. We thus generated RFe cells stably expressing FLAG-NSP1, FLAG-NSP13, FLAG-ORF6, or GFP as a control. Western blots analysis of stable RFe cells revealed that NSP13 was expressed at a decent level, but ORF6 expression was low (Fig 7D). Most cells expressing NSP1 died, which is consistent with its known ability to block translation [12], and these cells were thus excluded from the analysis. RFe cells stably expressing FLAG-NSP13, FLAG-ORF6, or GFP were stimulated with polyI:C or left untreated for 16 hours. RT-qPCR analysis showed that the relative amounts of ISG20 and OAS1 mRNAs were significantly lower in RFe cells expressing FLAG-NSP13, compared to control cells transduced with empty lentiviruses (mock) or lentiviruses expressing GFP (Fig 7E-7F). These results, combined with the ones obtained with 293T cells expressing the BANAL-236 proteins (Fig 7B), suggest that NSP13 inhibits the IFN response in human and Rhinolophus cells. The inability of ORF6 to modulate ISG expression in RFe cells is likely due to its low level of expression.

Cell lines

Human colorectal adenocarcinoma Caco-2 cells (kind gift from Nathalie Sauvonnet, Institut Pasteur, Paris), African green monkey kidney epithelial Vero-E6 cells (ATCC CRL-1586), A549 lung cancer cells (ATCC CCL-185), and human embryonic kidney (HEK) 293T cells (ATCC CRL-3216), hereinafter referred to as 293T, were maintained in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco) containing GlutaMAX I, sodium pyruvate (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Dutscher) and 1% penicillin and streptomycin (10 000 IU/ml; Thermo Fisher Scientific). BHK-21 golden hamster kidney fibroblasts were cultivated in Eagle’s Minimum Essential Medium (EMEM) (Gibco) supplemented with 10% heat-inactivated FBS (Dutscher), 1% penicillin and streptomycin (10 000 IU/ml; Thermo Fisher Scientific) and 1x non-essential amino acids. The Rhinolophus ferrumequinum cells (RFe) were derived from lung tissue of a bat captured in Jerez, in Southern Spain. Primary lung fibroblasts were immortalized using lentiviruses expressing the Simian Vacuolating Virus 40 large T antigen (SV40T) [28]. RFe cells were cultured in medium containing 50% DMEM with GlutaMAX I, sodium pyruvate (Invitrogen) and 50% F-12 (Ham) medium (Gibco), supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Dutscher) and 1% penicillin and streptomycin (10.000 IU/ml; Thermo Fisher Scientific). All cells were maintained at 37°C in a humidified atmosphere with 5% CO₂.

RFe and 293T cell lines stably expressing hACE2, hTMRSS2, rACE2-V5 or C-myc-rTMPRSS2 were generated by lentiviral transduction. 2.10⁵ cells were seeded in 6-well plates. Cells were transduced with 300 μL of lentivirus in 1 ml of medium containing 1x Polybrene the following day. After 3 days of incubation at 37°C, cells were either directly used for experiments or placed under antibiotic selection (6 μg/ml of puromycin and/or 6 μg/ml of blasticidin for RFe cells; 1 μg/ml of puromycin and 10 μg/ml of blasticidin for 293T cells). RFe cells stably expressing hACE2 and hTMPRSS2 (RFe-AT) were selected with 5 μg/ml blasticidin and 250 μg/ml hygromycin (Invitrogen). RFe-AT cells were used to derivate clones (RFe-ATC) using limiting dilution.

Viruses and infections

Experiments with BANAL-236 and SARS-CoV-2 viruses were conducted in a Biosafety Level 3 (BSL-3) laboratory, following safety protocols approved by the Institut Pasteur’s risk prevention service. The SARS-CoV-2 strain BetaCoV/France/IDF0372/2020 and BANAL-236 were obtained from the National Reference Center for Respiratory Viruses hosted by Institut Pasteur. Viral stocks were produced by amplification on Vero-E6 cells, for 72 h in DMEM supplemented with 2% FBS and 1% P/S.Cells were exposed to the virus for 3 hours in a low volume of FBS-free medium. Cells were infected with the viruses at different MOIs as indicated in each figure For infection in the presence of TPCK-trypsin, it was added after 3 hours of infection, at a concentration of 1 μg/ml, in the medium described above or in DMEM containing 2% FBS.

Experiments with Japanese encephalitis virus (JEV) strain UVE/JEV/UNK/TW/RP9 (kind gift from Philippe Desprès, UMR PIMIT, Reunion Island) were conducted in a Biosafety Level 3 (BSL-3) laboratory, following safety protocols approved by the Institut Pasteur’s risk prevention service. Cells were exposed to the virus for 2 hours in a low volume of 2% FBS containing medium. RFe-AT and RFe ATC cells were infected at a MOI of 10. The percentage of infection was assessed 24 hours post infection using the 4G2 antibody, an pan-flavivirus antibody raised against the viral envelope (E) protein by flow-cytometry.

Experiments with Vesicular stomatitis virus (VSV) and lentiviruses were performed in a BSL-2 + setting following biosafety regulations of the Institut Pasteur, Paris. The VSV Indiana strain was kindly provided by N. Escriou (Institut Pasteur).

TCID₅₀ assays

Supernatants of infected cells were first cleared of cell debris by centrifugation at 500 rpm for 5 minutes at 4°C. They were 10-fold serially diluted in DMEM supplemented with 2% FBS and 1% P/S. Vero-E6 cells in suspension were infected with 10-fold serial dilutions in 96-well plates. Cells were seeded at a concentration of 80 000 cells/ml and incubated for 5 days in DMEM medium containing 2% FBS. At the end of the incubation period, cells were washed twice with PBS and then fixed with a crystal violet solution containing 3% formaldehyde for 30 minutes at room temperature. Cytopathic effects (CPE) were assessed by calculating the 50% tissue culture infective dose (TCID₅₀) using the Spearman-Karber method [66].

Cloning of host proteins and production of pseudoviruses

The sequence of Rhinolphus malayanus ACE2 (rACE2) was synthesized with a C-terminal V5-6 His sequence tag and was amplified using primers designed to add a 5’ BclI restriction site (rACE2 for: 5’-TCTAGTTGATCAGCCACCATGTCAGGCTCTAGCTGGC) and 3’ XhoI restriction site (rACE2-V5 rev: 5’-GAGAGGCTCGAGTCTATCAATGGTGATGGTG). Human ACE (hACE2) coding sequence was amplified from pLenti6-hACE2 previously described [67] with the same primer design strategy as for rACE2 (hACE2 for: 5’-TCTAGTTGATCAGCCACCATGTCAAGCTCTTCCTGGCTCCTTC, hACE2 rev: 5’- GAGAGGCTCGAGTCGAGTTAAAAGGAGGTCTGAACATCATCA). PCR products were digested using BclI/XhoI restriction enzymes and cloned in the pFlap-Ubc-nLuc-IRES-Puro lentiviral vector (kind gift from Pierre Charneau, Institut Pasteur, Paris, France) digested with BamHI/XhoI.

Human TMPRSS2 (hTMPRSS2) lentiviral vector was purchased from Addgene (pWPI-IRES-Bla-Ak-TMPRSS2; addgene plasmid #154982). Rhinolophus ferrumequinum TMPRSS2 (rTMPRSS2) coding sequence was amplified from Rhinolophus ferrumequinum cDNA using primers designed to add a N-terminal c-myc Tag and a 5’ PmeI restriction site (rTMPRSS2 fwd: 5’-TAGCCTCGAGGTTTAAACCCGGGAGCAGCACCATGGAGCAGAAACTCATCTCTGAAGAGGATCTG ATGGCTTTAAACTCAGGATC) as well as a 3’ SpeI restriction site (rTMPRSS2 rev: 5’-GTGGCCACTAGTACGTACGGTCCGCATATGGATCCCTAGCTGTTTGCCCTCATTTGTC). Amplicon were cloned into the backbone of pWPI-IRES-Bla-Ak-TMPRSS2 lentiviral vector using PmeI/SpeI restriction sites. All plasmids were propagated in Stbl3 E. coli cells and verified by sequencing (Eurofins Genomics). To generate RFe-AT and RFe-ATC cells, plasmids pTRIP-SFFV-Hygro-2A-TMRPSS2 and pWPI-IRES-Bla-Ak-ACE2 Hygro_TMPRSS2 were used. The lentiviruses coding for the S protein of BANAL-236 and SARS-CoV-2 (BetaCoV/France strain) were previously described [1]. Lentiviral pseudovirions were produced in 293T cells in 10-cm dishes. Co-transfection included the spike plasmid (5 μg), a lentiviral vector expressing luciferase (10 μg, pHAGE-CMV-Luc2-IRES-ZsGreen-W), and auxiliary plasmids (3.3 μg each) HDM-Hgpm2, HDM-tat1b, pRC-CMV-Rev1b using calcium phosphate precipitation. After 5 hours, the medium was replaced with serum-free and phenol red-free DMEM. Pseudo-particles were harvested at 48 hours, clarified by centrifugation, and frozen. Control lentiviruses without S coding sequence were produced in parallel. The amount of pseudo-particles was quantified by p24 ELISA, as described previously [68].

Cloning of BANAL-236 ORFs

BANAL-236 sequences were amplified from BANAL-236 cDNA using primers designed to add attb1 and attb2 recombination cassettes (S3 Table). Difficult to amplify ORFs were synthetized flanked by attb1 and attb2 sequences. PCR products and synthetic DNA were cloned into pDONOR221entry vector using Gateway BP clonase II Enzyme Mix (Thermo Fisher Scientific). All entry clones were transferred into gateway compatible pciNeo 3X-Flag expression vector (kind gift of Yves Jacob, Institut Pasteur). BANAL-236 NSP3 coding sequence was amplified using primers designed to add 5’ NotI and 3’ BamHI sites (S2 Table) and in-frame cloned into P3XFLAG-CMV-10 (Promega) using NotI/BamHI restriction enzymes. C-terminally STREP tagged BANAL-236 ORF8 and S expression vector were constructed cloning synthetic sequence (S3 Table) into pLVX-EF1alpha-IRES-Puro vector (Takara) using EcoRI and BamHI restriction sites. Gene synthesis was performed by Genecust.

Alternatively, SARS-CoV2 and RaTG13 viral open reading frames (ORF) previously cloned in pDONR207 [28,69] were used as templates to derive BANAL-236 sequences using site directed mutagenesis (Quick-change site directed mutagenesis kit, Agilent Technologies Inc.) with primers listed in S4 Table. Previously cloned E, Nsp9, Nsp10 and Nsp16 sequences of SARS-CoV2 in pDONOR207, which are identical to BANAL-236 were not modified.

N-terminally FLAG-tagged BANAL-236 NSP1, NSP13 and ORF6 were amplified using primers designed to add a 5’ BamH1 restriction site (BANAL-236-ORF-5’-BamH1_fwd: 5’-TCTAGTGGATCCGCCACCATGGACTACAAAGACCATGA) and a 3’ SGRD1 restriction site (BANAL-236-ORF-3’-SGRD1_rev:5’-TGATCCCGTCGACGCCACTTTGTACAAGAAAGCT) using pciNeo 3X-Flag- BANAL-ORFs as a template. Purified amplicons were digested using BamH1/SGRD1 restriction enzymes and subcloned in pFlap-Ubc-nLuc-IRES-Puro lentiviral vector digested with BamHI/XhoI.

Spike-pseudotyped lentivirus entry assays

Entry assays were conducted with 293T cells stably expressing hACE2-V5, rACE2-V5, myc-hTMPRSS2, myc-rTMPRSS2 or a combination of entry factors [25], BHK-21 cells transiently expressing hACE2-V5, rACE2-V5, myc-hTMPRSS2, myc-rTMPRSS2 or a combination of entry factors, and RFe cells stably expressing hACE2, hTMPRSS2, rACE2, or a combination of entry factors. Cells were transduced in suspension with pseudoviruses at 0.5 ng of HIV-1 p24 antigen for 2.10⁴ cells and then plated in 96-well plates. After 72h of transduction, luciferase activity was measured by adding 100 μl of Bright-Glo substrate (Promega) and quantifying luminescence with a Varioskan LUX luminometer (Thermo Fisher). The expression of the entry factors was assessed by flow cytometry at D0 for each experiment.

Transfections

Plasmid transfection in 293T was performed using Trans IT‐293 (Mirus) following the manufacturer’s protocol. For polyI:C stimulation, RFe cells were plated in 12-well plates. The next day, they were transfected with 100 ng polyI:C (InvivoGen) or PBS, respectively, using INTERFERin (Polyplus Transfection) transfection reagent. Caco-2 cells were transfected with 500 ng of polyI:C using lipofectamine RNAiMAX (Thermo Fisher). Cells were lysed 16 h after transfection.

For siRNA transfection, cells were plated in 6-well plates and transfected the next day with 100 nM of siRNA targeting human TMPRSS2 (7113, Horizon Discovery Biosciences), using Lipofectamine RNAiMAX (Thermo Fisher) transfection reagent. Cells were harvested 72 h after transfection.

Luciferase reporter assays

For luciferase experiments 293T cells were seeded in 96-well plates and transfected using Trans IT‐293 (Mirus) or using FuGENE HD (Promega) with a mixture of 20 ng of pISRE‐Luc, 2 ng of pRL‐TK‐Renilla and 20 ng of the plasmids expressing viral proteins or 20 ng of pCi‐Neo plasmids (Promega) as “empty vector” (EV) controls. NS1 Flu (kindly provided by Daniel Marc, INRAE) and NS5 TBEV [47] were used as positives controls. All plasmids were grown in TOP10 cells (Thermo Fisher Scientific) and verified by sequencing. To stimulate the RIG-I/IRF3 axis, 20ng of pdeltaRIG-I (kindly provided by Pierre Genin, Centre d’Immunologie et des Maladies Infectieuses, Paris) was added in the plasmid mix. To activate the JAK/STAT pathway, cells were stimulated with 200 IU/ml of IFNα2b (PBL assay science) 8 hours post‐transfection.Twenty‐four hours post‐transfection, cells were lysed using Passive Lysis buffer (Promega) for at least 15 min and luciferase activity was measured with Dual‐Glo Luciferase Assay System (Promega) following the manufacturer’s protocol.

Immunofluorescence microscopy

Cells were fixed with 4% paraformaldehyde (PFA) (Sigma-Aldrich) for 30 min at room temperature. Cells were blocked for 30 min with PBS containing 0.05% Saponin and 5% BSA before incubation with the indicated primary antibodies (S5 Table) for 1 hour. After incubation, cells were washed three times with PBS containing 0.05% saponin and 5% BSA. Alexa Fluor 488 or 647 conjugated secondary antibodies were added for 1 hour. After incubation, cells were washed twice with PBS 0.05% saponin and 5% BSA and once with PBS. Nuclei were stained for 20 min with Hoechst-PBS (Thermo Fisher, 33342 (20 nM)). After washing, slides were mounted with Prolong Gold imaging medium (Life Technologies, P36930). Images were acquired using a Leica SP8 confocal microscope or Zeiss LSM 700 inverted.

Flow cytometry

Cells were detached with trypsin and fixed in 4% PFA for 30 min at 4°C. Staining was performed in PBS, 2% bovine serum albumin (BSA), 2 mM EDTA (Invitrogen), and 0.1% saponin (FACS buffer). Cells were washed tree time and incubated with indicated antibodies (S5 Table) during 1 hour at 4°C. Cells were washed tree time again and incubated with secondary antibody Alexa 488 or 647 during 45 min in the dark at 4°C. For cell surface staining, cells were detached with trypsin and washed with PBS, 0.5% BSA and 0.4% EDTA (surface buffer). Cells were washed tree time and incubated with indicated antibodies (hACE2-Alexa 647, hTMPRSS2 [30] or anti-N [70]) during 1 hour at 4°C. For hTMPRSS2 staining, cells were washed tree time again and incubated with secondary antibody Alexa 647 during 45 min in the dark at 4°C. Cells were wash tree times again and fixed in 1% PFA for 10 min at 4°C. Data were acquired on an Attune NxT flow cytometer (Thermo Fisher) and were analyzed using FlowJo software v10 (TriStar).

Transmission Electron Microscopy

About 10⁶ Caco-2 cells were mock-infected or infected with SARS-CoV-2 or BANAL-236 at MOI 0.001 and 0.01 respectively. RFe cells were mock-infected or infected with SARS-CoV-2 or BANAL-236 at MOI of 0.02 and 0.2, respectively. Cells were detached with trypsin and then wash twice with PBS and incubated in 1% glutaraldehyde/4% paraformaldehyde (Sigma, St. Louis, MO) in 0.1 M phosphate buffer (pH 7.2). Samples were then washed in PBS and postfixed by incubation for 1 h with 2% osmium tetroxide (Agar Scientific, Stansted, UK). The cells were then fully dehydrated in a graded series of ethanol solutions and propylene oxide. They were impregnated with a 1:1 mixture of propylene oxide/Epon resin (Sigma) and left overnight in pure resin. Samples were then embedded in Epon resin (Sigma), which was allowed to polymerize for 48 h at 60°C. Ultrathin sections (90 nm) of these blocks were obtained with a Leica EM UC7 ultramicrotome (Wetzlar, Germany). Sections were stained with 2% uranyl acetate (Agar Scientific) and 5% lead citrate (Sigma), and observations were made with a transmission electron microscope (JEM-1011; JEOL, Tokyo, Japan).

Western blot analysis

Cells were lysed in RIPA buffer (Sigma-Aldrich) supplemented with a protease and phosphatase inhibitor cocktail (Roche). Samples were denatured in 4X loading buffer (Li-Cor Bioscience) under reducing conditions (NuPAGE reducing agent, Thermo Fisher Scientific). Proteins were separated by SDS-PAGE (NuPAGE 4–12% Bis-Tris gel or NuPAGE 10–20% Tricine gel for small protein, Invitrogen) in running buffer MOPS (Invitrogen) or MES (Invitrogen) for small proteins. Proteins were transferred to nitrocellulose membranes (Bio-Rad) using a Trans-Blot Turbo system (Bio-Rad) or liquid transfer for NSP3. Membranes were blocked with PBS-Tween 0.1% containing 5% milk. After blocking, membranes were incubated overnight at 4°C with primary antibodies (S5 Table) diluted in blocking buffer or PBS-Tween 0.1%. Finally, membranes were washed and incubated for 45 minutes at room temperature with diluted secondary antibodies, then washed three time again. Images were acquired with an Odyssey CLx infrared imaging system (Li-Cor Bioscience).

Antibodies and cytokines

Details about antibodies are given in S5 Table. Cells were stimulated overnight with IFNα2b (PBL Biosciences) at a final concentration of 1 000 IU/ml.

RNA extractions and RT-qPCR assays

Total RNA was extracted from cell lysates using the NucleoSpin RNA II kit (Macherey-Nagel) and eluted in water. cDNA synthesis was performed on 1 µg of total RNA with RevertAid H Minus M-MuLV reverse transcriptase (Thermo Fisher Scientific) using random p(dN)6 primers (Roche). Real-time quantitative PCR was carried out on a Quant Studio 6 Flex system (Applied Biosystems) with SYBR green PCR mix (Life Technologies). Data were analyzed using the ΔΔCT method, with all samples normalized to GAPDH. All experiments were performed in technical triplicate. The primers used for RT-qPCR analysis are listed in S6 Table. Genome equivalent concentrations were determined by extrapolation from a standard curve generated from serial dilutions of plasmids expressing hACE2-V5, myc-hTMPRSS2, rACE2-V5, myc-rTMPRSS2 or BANAL-236 E protein.

RNA library and sequencing

Library preparation and sequencing was performed by the Biomics platform at the Institut Pasteur. cDNA libraries were prepared from 100-200 ng RNA using a Illumina Stranded mRNA library Preparation Kit (Illumina, USA) following the manufacturer’s protocol whereby adaptors are added to cDNA by A-tail mediated ligation. Index barcodes were added by PCR of 16 cycles. Unbound adaptors and indexes were eliminated by purification on AMPure magnetic beads (Beckman-Coulter). The resulting library featured an electrophoretic profile of 200–1000 bp, with a major peak of 375 bp, as visualized on a 3500 Fragment Analyzer (Agilent). A NextSeq 2000 sequencing system and a P4 50c flowcell (Illumina) were used to obtain 67-nt single-end double-indexed reads.

RNA-seq analysis

The RNA-seq analysis was performed with Sequana 0.18.0 [71]. We used the RNA-seq pipeline 0.20.0 (https://github.com/sequana/sequana_rnaseq) built on top of Snakemake 7.32.4 [72]. Briefly, reads were trimmed from adapters using Fastp 0.23.2 [73] then mapped to the corresponding genome using STAR 2.7.10a [74]. Rhinolophus ferrumequinum RFe-AT and RFe-TC samples were mapped to GCF_004115265.2 assembly with annotation GCF_004115265.2-RS_2023_02 from NCBI. FeatureCounts 2.0.1 [75] was used to produce the count matrix, assigning reads to features using corresponding annotation with strand-specificity information. Quality control statistics were summarized using MultiQC 1.17 [76]. Statistical analysis on the count matrix was performed to identify differentially regulated genes. Clustering of transcriptomic profiles were assessed using a Principal Component Analysis (PCA). Differential expression testing on each strain separately was conducted using DESeq2 library 1.38.3 [77] scripts indicating the significance (Benjamini-Hochberg adjusted p-values, false discovery rate FDR < 0.05) and the effect size (fold-change) for each comparison. DESeq2 output files (S1 Table) was used to filter and plot Differentially Expressed Genes (DEG) (Log2 fold change ≥ 1.5, adjusted p-value < 0.05). The Gene Ontology GO enrichment (Biological process) on the list of differently upregulated gene was analyzed using Mestascape [78].

Sequencing of BANAL-236 stocks was performed as previously [6] except that the SMARTer Stranded Total RNA‐Seq Kit v3 – Pico Input Mammalian library was sequenced on an Illumina NextSeq2000 in a paired‐end 200‐cycle run. Raw reads were trimmed with AlienTrimmer and further mapped onto BANAL‐236 genome (GenBank accession number MZ9370003.2) using CLC Genomics Workbench 25.0.2 with read Length fraction = 1 and Similarity fraction = 0.985. Low Frequency Variant Detection was performed with CLC Genomics Workbench 25.0.2 with the following parameters: required significance >1%, minimum read coverage 100, minimum read count 10, minimum frequency 1%, minimum central quality 20, minimum neighborhood quality 15, relative read direction filter 1%, removal of pyro-error variants. Filtering of variants hits detected by CLC were manually performed with the following parameters: minimum base quality 30, forward/reverse balance >0. Variants selected after this filtering were finally checked for position on read to remove variants positioned on read ends. Mutations harbouring an allelic frequency above 1% were selected (S7 Table) while mutations with low frequency and weak variation were discarded as putative sequencing error artifacts.

Data analysis

Data are presented and analyzed using GraphPad Prism 9. Alignments and trees were generated using CLC Genomics Workbench 22. Immunofluorescence quantification was done using Zen Blue software.