Replication-Competent Influenza B Reporter Viruses as Tools for Screening Antivirals and Antibodies

Abstract

Influenza B virus is a human pathogen responsible for significant health and economic burden. Research into this pathogen has been limited by the lack of reporter viruses. Here we describe the development of both a replication-competent fluorescent influenza B reporter virus and bioluminescent influenza B reporter virus. Furthermore, we demonstrate these reporter viruses can be used to quickly monitor viral growth and permit the rapid screening of antiviral compounds and neutralizing antibodies.

TEXT

Influenza B virus (IBV) is a common human pathogen that causes yearly epidemics and significant disease in the pediatric population (1–4). Despite the importance of this virus, far less is known about IBV relative to influenza A virus (IAV). Research into IAV has been greatly enhanced by several replication-competent reporter viruses (5–17). These reporter viruses have been crucial for identifying host factors, understanding basic viral pathogenesis, screening for antiviral compounds, and characterizing broadly reactive antibodies (6–10, 12–19). Lack of viruses encoding reporters has delayed similar progress for IBV.

Development of a fluorescent influenza B reporter virus.

We have previously published a description of an IAV reporter virus that contains the Gaussia luciferase gene directly behind the PB2 open reading frame (ORF) in segment 1 of A/Puerto Rico/8/1934 virus (PR8) (7). A 2A proteolytic cleavage site inserted between the PB2 ORF and the Gaussia luciferase gene allows for the cotranslational separation of the two proteins (20).

Utilizing a similar approach, an influenza virus codon-optimized ORF of mNeonGreen (mNeon), a gene encoding a bright monomeric fluorescent protein (21), was cloned behind each of the three polymerase segments of the B/Yamagata/16/1988 (Ya88) virus (Fig. 1A). The reporter segments were rescued individually utilizing standard protocols (22–24) to generate the PB1 mNeon, PB2 mNeon, and PA mNeon viruses. Madin-Darby canine kidney (MDCK) cells were infected at a multiplicity of infection (MOI) of 0.5 for 18 h with the recombinant wild-type (rWT) Ya88 or one of the three reporter viruses (with all experiments carried out at 33°C). Tosyl phenylalanyl chloromethyl ketone (TPCK) trypsin, which is required for spread of viral progeny, was excluded from the infection medium. The cells were imaged or subjected to flow cytometry after being labeled with Hoechst stain (Thermo Scientific) or LIVE/DEAD stain (Life Technologies), respectively. By epifluorescence microscopy, the PB1 mNeon and PB2 mNeon viruses were bright and easily observed by 18 h postinfection, but the PA mNeon virus was less detectable (Fig. 1B to E). Flow cytometry (BD LSRIIA) recapitulated the microscopy results (Fig. 1F to I). Quantification of the flow cytometry data with FlowJo indicated that both the PB1 mNeon and PB2 mNeon viruses yielded the brightest signals (Fig. 1J and K), and the PB1 mNeon virus was chosen for further study.

FIG 1.

The PB1 polymerase segment of the Ya88 virus allows for the highest expression of mNeon. (A) Graphical representation of the Ya88 mNeon reporter segments. The primers and sequences used to clone these viruses are available upon request. Abbreviations: UTRs, untranslated regions; Mut Pkg, 50 nucleotides of silently mutated packaging sequence; PTV 2A, porcine teschovirus 2A cleavage site with GSG linker; mNeon, mNeon ORF; Pkg, 150 bp of repeated packaging sequence behind the mNeon ORF. (B to I) MDCK cells were infected individually at an MOI of 0.5 with rWT Ya88 or the Ya88 mNeon virus. Infected cells were either imaged at 18 h postinfection after staining of nuclei with Hoechst stain (scale bars, 400 μm) (B to E) or harvested, stained with LIVE/DEAD (Life Technologies), and analyzed by flow cytometry (F to I). (J) The intensities of the green signals of all three reporter virus-infected populations (and the rWT-negative population) are displayed as histograms. (K) Average brightness of three independent infected cell populations for each virus. Error bars indicate standard errors of the means (SEM). Statistical significance of fluorescent intensity was determined by pairwise analysis with Student's t test. (All data were analyzed by GraphPad Prism.) ****, P < 0.0001; NS, not significant.

The PB1 mNeon virus can rapidly monitor the growth of IBV in vitro.

The growth characteristics and reporter stability of PB1 mNeon were next assessed. Stocks of the PB1 mNeon virus grew to titers comparable to those of the rWT virus in eggs (Fig. 2A). In a multicycle growth comparison, the PB1 mNeon virus grew at a similar rate to the rWT virus, indicating minimal attenuation (Fig. 2B). Furthermore, the reporter signal was stable after dilution of the virus to 10⁻⁷ for four serial passages (Fig. 2C). We postulate the influenza virus codon optimization greatly enhanced the stability of the PB1 mNeon virus.

FIG 2.

The PB1 mNeon virus reporter activity is stable and can be used to quickly monitor viral replication. (A) Titers determined by plaque assay of rWT or PB1 mNeon viruses propagated in eggs in triplicate. (B) MDCK cells were mock infected or infected with the rWT virus or the PB1 mNeon virus at an MOI of 0.001. At the indicated time points, the viral titer was measured by standard plaque assay (25). (C) The PB1 mNeon virus was passaged four times in eggs after being diluted to 10⁻⁷ prior to each passage (P1 through P4). Each passage was plaqued in triplicate, immunostained with anti-IBV sera, and labeled with a red Alexa Fluor 594-conjugated secondary antibody (Life Technologies). Fifty immunostained plaques from each replicate were analyzed for mNeon fluorescent activity. (D) MDCK cells were infected at an MOI of 0.5 with the PB1 mNeon virus. At 0, 12, and 24 h postinfection, cells were labeled with Hoechst stain and imaged at the same fluorescent exposures. Cells infected with the rWT virus were also imaged at 24 h. Scale bars, 400 μm. (E) MDCK cells were mock infected or infected with the rWT virus or the PB1 mNeon virus at an MOI of 0.001. At the indicated time points, viral growth was measured with an Acumen laser scanning imaging cytometer. All data are representative of four independent samples, except as noted. Error bars indicate standard deviations (SD). Statistical significance was determined by pairwise analysis with Student's t test. *, P < 0.05; NS, not significant; ND, not detected.

The PB1 mNeon virus was next used to infect MDCK cells at an MOI of 0.5 without TPCK in a single-cycle growth assay (Fig. 2D), and signal was easily detected by 12 h postinfection until 24 h before major cytoplasmic effect (CPE). Logarithmic growth of the PB1 mNeon virus in MDCK cells infected at an MOI of 0.001 with TPCK could be monitored rapidly with an Acumen laser scanning imaging cytometer at the indicated times (Fig. 2E) with comparable kinetics to standard growth curves. Thus, the PB1 mNeon virus can be utilized to monitor multicyclic growth in real time, in comparison to a standard growth curve, which takes an additional 3 to 4 days to determine the titer at each time point (25). Together, these data indicate that the PB1 mNeon virus is suitable to rapidly monitor the growth of IBV by simple microscopy.

A PB1 NanoLuc virus can sensitively monitor virus infection.

Exploiting a similar design to the PB1 mNeon virus, the gene encoding NanoLuciferase (NanoLuc), a bright 19-kDa bioluminescent protein (9), was inserted into the PB1 segment with a KDEL motif to ensure intracellular localization and rescued to make the PB1 NanoLuc virus (Fig. 3A). The KDEL motif was added based on our previous experience suggesting that this motif prevented release of the reporter into the supernatant (7). To determine if luciferase signal correlated with viral titer, MDCK cells were infected with 10-fold dilutions of PB1 NanoLuc starting at an MOI of 10. At 5 h postinfection, the cells were lysed (NEB luciferase cell lysis buffer), and the luciferase signal was assayed (Promega NanoLuc kit) (Fig. 3B). The dilution containing 5 PFU of virus (MOI of 0.0001) could be significantly detected. In addition, a clear relationship between viral titer and signal was observed (Fig. 3C). Again, as with the PB1 mNeon virus, logarithmic multicyclic replication of the PB1 NanoLuc virus could be monitored rapidly by its luciferase signal (Fig. 3D) and displayed kinetics similar to those of the actual viral titer (Fig. 3E). These results indicate that the PB1 NanoLuc reporter virus can be sensitively and rapidly detected, even in short time course experiments, and that the reporter signal directly correlates with viral replication.

FIG 3.

The PB1 NanoLuc virus can be used to sensitively and rapidly determine viral titer. (A) Schematic of the PB1 NanoLuc segment. Abbreviations: UTRs, untranslated regions; Mut Pkg, 50 nucleotides of silently mutated packaging sequence; PTV 2A, porcine teschovirus 2A cleavage site with GSG linker; NanoLuc, NanoLuc ORF; KDEL, endoplasmic reticulum retention signal; Pkg, 150 bp of repeated packaging sequence behind the NanoLuc ORF. (B) MDCK cells (50,000 cells/well) were mock infected, infected with rWT at an MOI of 10, or infected with 10-fold serial dilutions of PB1 NanoLuc starting at an MOI of 10 (500,000 PFU) to 0.0001 (5 PFU) in 96-well plates. At 5 h postinfection, the cells were lysed with 125 μl lysis buffer, 50 μl of lysate was mixed with 50 μl NanoLuc substrate, and activity was measured. (C) Data from the luciferase assays for PB1 NanoLuc were graphed against the PFU from each representative MOI and analyzed by nonlinear regression. (D and E) MDCK cells in 24-well plates were infected at an MOI of 0.001 with the rWT or PB1 NanoLuc virus with 1 μg/ml TPCK. At the indicated times, cells and supernatants were collected separately. (D) Cells were lysed with 250 μl lysis buffer, and luciferase activity was measured as before. (E) Titers in supernatants were determined by standard plaque assay. All data are representative of four independent samples. Error bars indicate standard deviation (SD). Statistical significance was determined with pairwise Student's t tests. *, P < 0.05; **, P < 0.005; ***; P < 0.001; NS, not significant; RLU, relative light units.

The PB1 mNeon and PB1 NanoLuc viruses can be used to rapidly screen for antivirals and neutralizing antibodies.

Novel IBV antivirals are needed. There are only two approved antivirals (oseltamivir and zanamivir) for treatment of pediatric influenza virus infections (26, 27). Additionally, as with IAV, antiviral resistance is present in IBV isolates (28–30). We utilized the PB1 mNeon virus to test if we could screen for antivirals since growth evaluation based on reporter signal is the least labor-intensive method and may be ideal for rapid screens of chemical libraries.

For a benchmark comparison in determining the utility of our viruses as screening tools, we used zanamivir (Sigma-Aldrich SML0492) to inhibit the growth of the rWT virus in a standard plaque reduction assay (Fig. 4A) (17, 25, 31). Utilizing the PB1 mNeon virus, we were able to measure reductions in fluorescent output of this virus far more rapidly with less effort using the Acumen plate reader than the standard plaque reduction assay (Fig. 4B). The half-maximal inhibitory concentration values (IC₅₀s) calculated for both were not statistically different (Fig. 4C). The decreased replication of the PB1 mNeon virus in the presence of zanamivir detected by the Acumen reader was confirmed by microscopy (Fig. 4D). This indicates that PB1 mNeon has the potential to be used to rapidly screen for viral inhibitors.

FIG 4.

The PB1 mNeon and NanoLuc viruses can be used to rapidly screen for antivirals and neutralizing antibodies. (A) MDCK cells were infected at an MOI of 0.00005 with rWT virus in 6-well plates in triplicate. After infection, cells were overlaid with 5-fold serial dilutions of zanamivir starting at 500 nM to 0.16 nM with 1 μg/ml TPCK. At 48 h, plaques were fixed, immunolabeled, and counted without bias with ImageJ. Data were normalized to rWT plaqued without the inhibitor (i.e., 100%). (B) MDCK cells were infected with PB1 mNeon in 96-well plates at an MOI of 0.01. After infection, cells were incubated with 5-fold serial dilutions of zanamivir starting at 500 nM to 0.16 nM in infection medium with 1 μg/ml TPCK. At 36 h, the PB1 mNeon virus-infected MDCK cells were read on the Acumen plate reader without fixation or antibody labeling. (C) IC₅₀ values for PB1 mNeon were calculated from three sets of duplicate samples and three individual replicates for the rWT virus. (D) After the PB1 mNeon plate was read on the Acumen plate reader, cells were stained with Hoechst stain and imaged by fluorescence microscopy. Scale bars, 400 μm. (E and F) ELISA plates were coated with 50 μl WT PR8 or Ya88 virus overnight. Five individual serum samples from each condition were pooled, treated with receptor-destroying enzyme (RDE), and then tested for reactivity to PR8 or Ya88 virus by ELISA using standard protocols (32). (G and H) One thousand PFU of the rWT virus or PB1 NanoLuc virus was subjected to a microneutralization assay as previously described (32). (G) At 36 h postinfection, rWT virus-infected cells were fixed and labeled by a traditional microneutralization assay protocol (32). (H) At 24 h postinfection, PB1 NanoLuc-infected cells were lysed with 125 μl lysis buffer, and luciferase activity was measured as before. The ELISA and microneutralization assays are representative of three independent samples. Error bars indicate standard errors of the means (SEM). Statistical significance determined with Student's t test. NS, not significant.

Next we determined if the PB1 NanoLuc virus could be used to screen for virus neutralizing antibodies. The enzymatic nature of the reporter makes this virus exceptionally sensitive, and therefore it can accurately report the presence of low viral titers. Enzyme-linked immunosorbent assays (ELISAs) showed polyclonal sera from Yam88- or PR8 virus-infected mice were reactive against the respective strain (Fig. 4E and F). Traditional microneutralization assays (32) could detect that the Ya88 serum inhibited the rWT virus by 36 h (Fig. 4G). Utilizing the PB1 NanoLuc virus, antibody neutralization was detected at 24 h postinfection (Fig. 4H). Due to the decreased assay time and the fact that assays with the PB1 NanoLuc virus also require less effort, this approach has the potential for significant improvement over traditional methods.

In summary, we report the development and characterization of two replication-competent IBV reporter viruses. These viruses have minimal attenuation, are stable over serial passages, and have been shown to be effective tools to screen antiviral compounds and neutralizing antibodies. In future studies, these viruses can be used to identify influenza B virus-required host factors, screen for novel antivirals, and evaluate vaccination strategies. Furthermore, the strategy used to generate these viruses can likely be adapted to other strains of IBV for future studies, including the study of IBV pathogenesis in animal models.