Baby Hamster Kidney cells (BHK-21, Kerafast, Cat # EH1011), Lenti-X Human Embryonic Kidney cells (Lenti-X HEK-293T, Clontech, Cat # 632180), Normal Human Bronchial Epithelial cells (NHBE, Lonza, Cat # CC-2541) and HEp-2 cells (ATCC, Cat # CCL-23) were maintained in the respective growth medium and incubated at 37°C, 5% CO₂ atmosphere and 95% relative humidity.

BHK-T7 cells constitutively expressing T7-RNA polymerase were developed using lentiviral transduction. Cells were used for the production of the rgRSV-P-BlaM recombinant virus. BHK-21, BHK-T7 and Lenti-X HEK-293T were grown in DMEM supplemented with 10% heat-inactivated Fetal Bovine Serum (FBS) (Gibco BRL, Carlsbad, CA, United States), 100 U/mL penicillin, 100 μg/mL streptomycin. BHK-21 and BHK-T7 required an additional supplement of 2 mM L-glutamine; while the culture medium of Lenti-X HEK-293T required to be supplemented with 1 mM sodium pyruvate.

HEp-2 Cells were obtained from American Type Culture Collection (ATCC number: CCL-23) and maintained in Opti-MEM (Gibco BRL, Carlsbad, CA, United States) supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin.

Undifferentiated bronchial epithelial cells (NHBE) were maintained in Bronchial Epithelial Growth Medium (BEGM) (Lonza, Walkersville, MD, United States) following the supplier’s instructions.

The plasmid MDA-P-BlaM containing the RSV cDNA with the coding sequence for P-BlaM was derived from plasmid HC123 (Hallak et al., 2000). The P-BlaM as an extra sequence was constructed from gBlocks (Integrated DNA Technologies). The BlaM sequence is an optimized version of beta lactamase Y105W (Wolf et al., 2009). Using Gibson Assembly (New England Biolabs), gBlocks were cloned between P gene and M gene using AvrII (nt. 2930) and PvuI (nt. 6573) restriction sites of RSV HC123 cDNA plasmid. The plasmid MDA-P-BlaM is designed with a GFP gene positioned as the initial gene preceding the NS1 gene, and the entire vector is regulated by the T7 promoter. Plasmid pLVX-Hyg-T7 resulted by inserting the coding sequence T7 RNA polymerase (gift of Dr. Guy Lemay, University of Montreal, Canada) between XhoI and NotI restriction sites of plasmid pLVX-IRES-Hyg (Clontech Inc., Mountain View, CA, United States). The pLVX-Hyg-T7 plasmid was used to generate BHK-T7 cells. pLVX-Hyg-P-BlaM was constructed by inserting the P-BlaM coding sequence using gBlocks (Integrated DNA Technologies) into XhoI and NotI restriction sites of the pLVX-IRES-Hyg construct (Clontech Inc., Mountain View, CA, United States).

Plasmids were electroporated into Escherichia coli EPI300 (Lucigen Corp., Middleton, WI, United States, Catalog EC300110) for amplification. Plasmid purification was performed using QIAprep Spin Miniprep (QIAGEN GmbH, Hilden, Germany) and HiPure Plasmid Maxiprep (Invitrogen, Carlsbad, CA, United States).

Plasmids containing sequences of interest were mixed with Lenti-X HTX Packaging mix (Clontech Inc., Mountain View, CA, United States), followed by transfection into Lenti-X HEK-293T cells in the presence of Xfect (Clontech Inc., Mountain View, CA, United States), following the manufacturer’s instructions. The lentiviral particles pVLX-T7 and pVLX-P-BlaM were recovered and frozen at −150°C for later use.

BHK-21 cells and HEp-2 cells were transduced with respective lentiviruses to generate the BHK-T7 and Hep-2-P-BlaM cell lines, respectively. The transduction was performed using spinoculation; where, 100 μL of the lentiviral suspension was added to the cells followed by centrifugation at 1,200 g for 90 min at 32°C. The inoculum was removed. Cell culture was washed with DPBS and then maintained in DMEM supplemented with 2% FBS for 24 h. Transduced cells were selected for their resistance to hygromycin B (200 μg/mL). The selection process took 15 days, changing the medium supplemented with hygromycin every 48 h. BHK-T7 and Hep-2-P-BlaM cells were cryopreserved in DMEM medium supplemented with 10% DMSO and 10% FBS.

BHK-T7 cells were transfected with the plasmid MDA-P-BlaM and helper plasmids pA2-Nopt, pA2-M2-1opt, pA2-Lopt, pA2-Popt (Hotard et al., 2012), using Lipofectamine 3,000 (Thermo Fisher Scientific). After 48 h, the transfected cells were subcultured in a 25 cm² flask (Corning^®). Once the syncytia reached out 60% of the monolayer, the cells were separated using a scraper in a volume of 5 mL of culture medium. Cell debris was removed by centrifugation at 1,800 g for 10 min at 4°C. The virions present in the supernatant were stabilized by the dropwise of MgSO₄ (0.1 M). Virions were cryopreserved at −150°C.

The aliquots containing recovered virions were used to propagate the rgRSV-P-BlaM virions according to Hallak et al. (2000). One 1 milliliter aliquot of rgRSV-P-BlaM was used to infect HEp-2 cells grown in a 75 cm² flask. After 2 h of incubation at 37°C, the medium was replaced by DMEM supplemented with 2% FBS and 1% Penicillin-Streptomycin, with medium changes every 48 h until fluorescent syncytia were observed that reached 60% of the monolayer. The supernatant containing the recombinant virus was cryopreserved at −150°C.

Hep-2-P-BlaM cells were infected following the previously described methodology. Viral titer by GFP and P-BlaM activity were characterized by flow cytometry and western blot, respectively.

The viral titer was determined following the protocol of Techaarpornkul et al. (2001). Briefly, NHBE cells grown in 24-well dishes were infected with serial dilutions of aliquots containing the infectious inoculum for 2 h at 37°C. then, the inoculum was removed and cells were incubated for 16 h at 37°C. Infected cells were detected by expression of the GFP signal; while the cells where there was release of the ribonucleoprotein content of the virus was estimated by the signal of the fluorescence shift towards the hydroxycoumarin channel by indicating the beta-lactam ring cleavage into CCF2 fluorochrome. The viral titer was estimated considering that dilution of the aliquot that generated a signal in the range of 0.05 to 0.1 of the total cells that were analyzed by flow cytometry.

The rgRSV-P-BlaM viral suspension was centrifuged at 20,000 g for 2 h at 4°C to concentrate the virions in a smaller volume. The proteins present in the virions were obtained by lysis. Used as a control, lysate from a culture of cells infected with rgRSV-P-BlaM was used as a positive control. Nupage LDS sample buffer (Thermo Fisher) was used for the lysate in the presence of Halt^™ Protease Inhibitor Cocktail (Thermo Fisher). Proteins were separated by SDS-PAGE under reducing conditions (+βME) and transferred to a nitrocellulose membrane. The P-BlaM protein was detected with the Anti-BlaM antibody [clone 8A5.A10] (ab12251). The binding of anti-BlaM was revealed by secondary antibodies labeled with near-infrared fluorochromes (800cw goat anti mouse IgG1) and visualized with Odyssey (LI-COR, Lincoln, NE, United States).

The P-BlaM protein was modeled in silico using the Colab server (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynbaccessed 01/06/2022). The generated structure was compared to protein P (Uniprot code P03421) and BlaM (Uniprot code C5I4X2). The reliability of the system predictions was assessed by the Local Distance Difference Test (LDDT) score according to the CASP14 assessment steps published by Jumper et al. (2021). The intrinsically disordered regions identified by AlphaFold2 were reviewed using DISOPRED3 (PSIPRED Workbench ucl.ac.uk) (Jones and Cozzetto, 2015). In addition, the amino acid sequence of the P protein was also modeled in AlphaFold2 and compared with the modeled structure of the P-BlaM chimera. The structure of the P protein was also reviewed using the derived model from cryo-EM studies that analyzed the P-protein tetramer in interaction with the RSV-L protein (PDB: 6PZK) (Gundenon et al., 1987).

NHBE cell cultures were exposed to an infectious inoculum of rgRSV-P-BlaM at an infectious dose of 3 FFU per 10 cells, which is approximately 0.3 MOI, for 1 h at 22°C This temperature allowed the adsorption but not the fusion (Henderson and Hope, 2006). The fusion was triggered by switching the temperature to 37°C, which consequently was defined as time “0.” Just before switching the temperature, we replaced medium in all cultures. For the set labeled as time “0.” A set of 3 dishes were incubated with quercetin (125 μg/mL) or with palivizumab (200 μg/mL) or mock. At each intervention time, the medium in the respective set of cultures was replaced with the one with the experimental condition. This was done until we reached 120 min. After 120 min, all the dishes were incubated with HBSS supplemented with 2 μM CCF2-AM (GeneBLAzer in vivo Detection Kit, Thermo Fisher Scientific) and the culture was incubated for an additional 1 h at 17°C. The cells were processed by flow cytometry (BD FACSCanto II) and hydroxycoumarin signal was measured for each condition.

NHBE cells grown in 24-well dishes to 80% confluence, were treated with different concentrations of ulixertinib for 1 h at 37°C. Cultures were inoculated with rgRSV-P-BlaM at an equivalent dose of 1 virion per 2 cells, in the presence of the inhibitor. The virions were adsorbed on the plasma membrane for 1 h at 22°C, to synchronize the viral entry at 37°C. After 2 h, all wells were incubated with HBSS supplemented with 2 μM CCF2-AM (GeneBLAzer in vivo Detection Kit, Thermo Fisher Scientific), for 1 h at 17°C. The cells were processed by flow cytometry and hydroxycoumarin signal was measured for each condition.

The MDA-P-BlaM plasmid containing the RSV cDNA was modified by inserting the recombinant P-BlaM sequence between the P and M genes (Figure 1A). The tertiary structure of the P-BlaM chimera was modeled using the Colab server running the AlphaFold2 algorithm (https://colab.research.google.com/github/sokrypton/Colab/blob/main/AlphaFold2.ipynb accessed 01/06/2022) (Figure 1B). The modeled structure was visualized in PyMOL. The modeled structure of the chimera RSV-P phosphoprotein has a 5 α helices and 5 loops in common as modeled sequence from the RSV-P phosphoprotein. Regions that share a loop-like structure include: 1M-10G, T29-N92, N111-N124, D212-N217, and S232-S237; while those that has similar α-helices are: E11-I24, S99-F109, I131-A155, T160-K208 and P218-L227 (Figure 1C). Comparing the modeled structure of the chimera with the regions to which we were able to access one of the monomers reported by Gilman et al. (2019), and which is deposited in the protein databank (PDB: 6PZK), we found homology in the respective alpha-helice zones; among these: 131ITARLDRIDEKLSEILGMLH151, 174REEMIEKIRT183, 190DRLEAMARLR199, 203SEKMAKDT210 and 217NPTSEKLNNLLE228. Likewise, the loop region 211SDEVSL216 is also shared. The regions that show disparity when both structures are compared as 161SARDGIRDAMVGL173 and 184EALMTN189. In the chimera P protein, these regions adopt an α-helical structure, while in the P protein elucidated by cryo-MS they have a loop shape.

In a study performed by Gilman et al. (2019), it is observed that phosphoprotein P shows structural plasticity of such magnitude that it has been proposed as an intrinsically disordered protein. The P protein has a low representation of hydrophobic amino acids (19.9%) and aromatic amino acids (5%), which is compatible with what is expected for intrinsically disordered proteins. In addition, Gly and Pro constitute 8.3% of the amino acids of the RSV-P protein, which are characterized by destroying the α-helices and β-sheet secondary structures. The presence of regions without a defined secondary structure in the RSV-P protein of the chimera is compatible with the characteristics of an intrinsically disordered protein (Knauer et al., 2012), like what was previously described for the RSV NS1 and NS2 proteins.

The structure obtained by crystallography was retrieved from the Protein Databank using the keyword beta-lactamase and aligned with the structure of the P-BlaM chimera protein modeled by AlfaFold2 in PyMOL. The structure of the chimera BlaM has a high structural homology with respect to the reported structure PDB: 1XPB (Figure 1B vs. Figure 1D), where the BlaM chimera protein retains all 12 α helices, 7 folded β sheets and 5 turns. This indicates that the P-BlaM chimera protein retains the conformational structure compatible with the preservation of BlaM enzymatic activity.

Undifferentiated bronchial epithelial cell cultures were inoculated with recombinant rgRSV-P-BlaM viruses for 1 h at 22°C to synchronize the entry of the virions. Once the infection was carried out at 37°C for 2 h, the culture medium was replaced with a loading buffer containing the CCF2-AM substrate. The CCF2-AM loading was performed at 17°C for 3 h, after which the cells were analyzed using flow cytometry to evaluate the beta-lactamase activity of the recombinant viruses.

The CCF2-AM fluorescent dye is a molecule composed of hydroxycoumarin and fluorescein linked with a beta-lactam ring acts a bridge. In the absence of beta-lactamase, the excitation of hydroxycoumarin at 409 nm generates emission associated with fluorescein (520 nm) due to the Föster emission resonance transfer (FRET) phenomenon. In Figure 2A, it is observed that in uninfected cultures, all cells show green fluorescence resulting from the emission of fluorescein.

On the other hand, infected cells from cultures exposed to rgRSV-P-BlaM show blue fluorescence because the presence of beta-lactamase cleaves the beta-lactam ring, resulting in the disruption of FRET between hydroxycoumarin and fluorescein. Thus, when stimulating infected cells at 409 nm, they emit fluorescence at 447 nm (Figure 2B).

Considering that the rgRSV-P-BlaM virion also encodes the GFP fluorescent protein, we evaluated whether the percentage of cells showing beta-lactamase activity correlated with the percentage of cells expressing GFP. The shift in fluorescence signal to hydroxycoumarin was identified in cells derived from cultures exposed to the infectious aliquot for 2 h, using flow cytometry. At the same time, other cultures were infected with the same volume of aliquot, and the percentage of cells expressing EGFP was evaluated 16 h later. The percentages obtained through both approaches were identical (Figure 2C vs. Figure 2D).

Regarding the substantial diversity of GFP-expressing cells, we must emphasize the following: (1) although the RSV genome contains the genes encoding GFP and P-BlaM, the experimental design assessed different things. Regarding BlaM, we detected its activity in the cells after its delivery during the entry phase by virions carrying it as a structural protein. On the other hand, GFP was assessed 16 h later in a concurrent and parallel set of cultures using the same aliquot and infectious dose; consequently, we detected GFP expression. (2) RSV is pleomorphic and both filamentous and spherical particles have been reported. Structural analysis of RSV using cryo-electron tomography revealed that the number of genome copies varies significantly among viral particles in the same aliquot (Kiss et al., 2014). In spherical viral particles, Kiss et al. (2014) reported up to 9 genome copies per particle, while the filamentous particles were reported to carry up to 3 copies. Thus, the substantial diversity in GFP fluorescence intensity is the expected result of cells being infected by virions carrying different numbers of genome copies. In addition, viral infection follows a Poisson distribution; that is, cells are not equally infected by the same number of virions (Shabram and Aguilar-Cordova, 2000).

The presence of the P-BlaM protein in the virions was detected using an anti-BlaM antibody labeled with a near-infrared fluorescence tag in a western blot system (Figure 3). Routledge et al. (1988) found that the RSV-P phosphoprotein ran on a gel as a ladder of polypeptides with different molecular weights (19 kDa to 34 kDa), with the most abundant band corresponding to a molecular weight of 34 kDa. The apparent molecular weight for BlaM is 27.2 kDa. Consequently, the apparent molecular weights of the bands observed in the control lane, where the lysate of the rgRSV-P-BlaM-infected culture ran, matched the expected range for the P-BlaM polypeptide ladder (46.2 kDa to 61.2 kDa). In the lane where the lysate of virions isolated from rgRSV-P-BlaM ran, a single band at the level of 61.2 kDa was observed, which was not present in the lane with RSV virions’ lysate.

Given that the western blot suggested the presence of the P-BlaM as a structural protein in the virion, we proceeded to evaluate the ability of rgRSV virions to carry the P-BlaM protein, which had been stably expressed in HEp-2 cells transduced with the pLVX-P-BlaM lentivirus. The virions were then used to assess both entry (measured by hydroxycoumarin channel emission due to beta-lactamase activity) and infection (GFP expression) in HEp-2 cell cultures using flow cytometry (Figure 4A vs. Figure 4B). Remarkably, when using the same volume per infectious aliquot (at an equivalent dose of 1 virion per cell), 70.9% of the cells exhibited fluorescence in the hydroxycoumarin channel, providing evidence that the virions indeed carried the P-BlaM protein (Figure 4A). The viral aliquot titer, estimated through beta-lactamase activity, was determined to be 6.30 × 10⁵ FFU, while the titer estimated through GFP expression was 1.01 × 10⁶ FFU. Based on the above results, we confirmed the successful incorporation of the P-BlaM protein into the recombinant rgRSV-P-BlaM virion.

Quercetin has been evaluated as an antiviral, particularly against RSV (Machado et al., 2019; Lopes et al., 2020). There are suggestions indicating that it plays a role in various stages of RSV infection, from adsorption to the membrane, to transcription, and viral replication. The role of quercetin during the virus envelope fusion event with the cell membranes of undifferentiated bronchial epithelium cultures was evaluated using a “Time-of-addition” (ToA) assay. The rgRSV-P-BlaM virions were adsorbed at 22°C for 1 h to synchronize the viral entry. The ToA assay started when the cultures were shifted to 37°C. From that moment, and every time point (0, 10, 20, 30, 60, 90, 120 min) for a period of 2 h (Figure 5A), different triplicate sets of cultures were exposed to quercetin or the palivizumab antibody. At the end of the 2 h, the cultures were incubated with CCF2-AM loading buffer for 1 h at 17°C, followed by the detection of the percentage of cells that emitted fluorescence in the hydroxycoumarin channel using flow cytometry. Quercetin inhibited viral entry, and the kinetics that followed were like those observed when the cultures were incubated with palivizumab (Figure 5B). The impact of quercetin on cellular viability was thoroughly evaluated in our study. Despite comprehensive analyses, the results consistently demonstrated no significant effects on cell viability (data not shown).

Previous studies have proposed that the signaling pathway leading to the activation of ERK-1/2 interferes with RSV infection, but it is not known whether active ERK-1/2 is required as part of the process or mechanism that facilitates RSV entry. Ulixertinib is a competitive ATP inhibitor at the active site of ERK-1/2, interfering with the phosphorylation of its substrates. In a dose-response assay, undifferentiated epithelial cell cultures were incubated with ulixertinib for 1 h before viral inoculation, and ulixertinib remained present throughout the incubation with the virus. The rgRSV-P-BlaM virions were adsorbed to the cell membrane for 1 h at 22°C to synchronize viral entry, and then cells were incubated at 37°C for 2 h. CCF2-AM loading and detection of hydroxycoumarin signal were performed as previously described. Ulixertinib at concentrations of 12.5 μM and 25 μM reduced the number of cells emitting hydroxycoumarin signal by up to 65 and 50%, respectively, compared to the cultures treated with DMSO (Figure 6). Consequently, the ERK-1/2 pathway seems to play an important role in facilitating RSV entry in a subset of cells. The impact of ulixertinib on cellular viability was thoroughly evaluated in our study. Despite comprehensive analyses, the results consistently demonstrated no significant effects on cell viability (data not shown).