Generation of Replication-Competent Recombinant Influenza A Viruses Carrying a Reporter Gene Harbored in the Neuraminidase Segment

Abstract

Replication-competent influenza viruses carrying reporter genes are of great use for basic research, screening of antiviral drugs, and neutralizing of antibodies. In this study, two recombinant influenza A viruses with a neuraminidase (NA) segment harboring enhanced green fluorescent protein (EGFP) in the background of A/PR/8/34 (PR8) were generated. The viral RNA (vRNA)-specific packaging signals for NA were largely retained for efficient packaging. An “autocleave” 2A peptide sequence, which was inserted at the N terminus or the COOH terminus of NA to link with EGFP, enabled NA and EGFP to be expressed monocistronically. Further analysis demonstrated that both viruses, named rPR8-EGFP+NA and rPR8-NA+EGPF, although with some characteristic differences in growth and EGFP expression, could replicate in noncomplementary cells and propagate to large quantities while maintaining genome stability after multiple passages in embryonated eggs. These replication-competent influenza viruses carrying reporter genes are a great addition to the tool set for developing antiviral therapeutics and vaccines and for in vivo studies of viral dissemination and pathogenicity.

The worldwide transmission of the high-pathogenicity avian influenza virus (HPAI) H5N1 and its occasional jumping to humans have posed great public health challenges and caused significant economic losses since 1996. In 2009, the “swine” H1N1 virus spread in 208 countries and claimed more than 12,220 lives (WHO, 27 December 2009). Live recombinant influenza viruses containing reporter genes, such as those encoding enhanced green fluorescent protein (EGFP) or luciferase, could be of great use in studying influenza virus and developing vaccines and antiviral drugs. The influenza A viral genome is constituted by 8 single-stranded, negative-sense RNA segments. Many efforts have been made to insert extra genes into its genome. Small peptides have been successfully inserted in-frame into the functional nonessential region of the hemagglutinin (HA) (8, 13) and neuraminidase (NA) (2, 3, 19) proteins, but the insertion of full-length proteins, such as the EGFP reporter, have been less successful. A ninth synthetic segment has been added to the viral genome (14, 15), but its lack of competitiveness in viral RNA (vRNA) packaging means that it can get lost eventually and therefore is infeasible for extensive propagation and large-scale usage. The G glycoprotein of the vesicular stomatitis virus (VSV-G) and the HA-esterase fusion (HEF) glycoprotein of the influenza C virus have been substituted for HA and NA in pseudoinfluenza viruses (5, 9, 25). These viruses could not be considered exactly the same as influenza A viruses, since HA was replaced by HEF of influenza C virus or VSV-G. The establishment of cell lines stably expressing HA proteins provides an alternative means to produce pseudotyped and single-round-replication influenza A virus for drug screening (12, 16, 17). However, these recombinant viruses can grow only in cells providing HA in trans and therefore cannot be produced in embryonated eggs, which makes them unsuitable for scale-up production. Attempts have been made to generate replication-competent influenza viruses by directly incorporating green fluorescent protein (GFP) or chloramphenicol acetyltransferase (CAT) genes as the dicistronic NA segment (24), but the incorporated GFP gene was unstable and often silenced. So far, replication-competent influenza virus carrying a gene such as that encoding GFP has not been generated successfully on a large scale for practical use.

Recently, the coding region in each vRNA has been found to play critical roles in the selective incorporation of vRNAs into mature progeny virions, as referenced (7, 18). These segment-specific packaging sequences allow the creation of better influenza vectors with high efficiency of packaging of foreign gene-containing segments (5, 7, 21, 23, 24). One example was that during the revision of this article, a paper was published in which a recombinant influenza A virus harboring a ninth segment was generated, containing a modified PB1 gene with its original packaging sequences exchanged for those of the NA segment (6).

This report describes an improved strategy for generating replication-competent recombinant influenza A viruses containing the EGFP gene within the NA vRNA segment. The NA vRNA packaging signals, including both the 3′ noncoding region (NCR) and the adjacent 183 nucleotides (nt) of the coding region and the 5′ NCR and the adjacent 157 nt of the coding region, were maintained for the efficient packaging of recombinant NA vRNAs. A short sequence that codes for a 2A “autocleave” peptide was inserted between the EGFP and NA genes, allowing EGFP and NA to be expressed monocistronically (Fig. 1A). Two plasmids, named pM-PR8-NA(183)-EGFP-2A-NA and pM-PR8-NA-2A-EGFP-NA(157), were constructed by inserting the EGFP gene into the 3′ and 5′ ends of NA vRNA in a bidirectional expression pM vector (4) for translation before and after the NA gene, respectively.

FIG. 1.

Generation of influenza A viruses rPR8-EGFP+NA and rPR8-NA+EGFP. (A) Schematic representation of wild-type NA (upper), chimera EGFP +NA (middle), and NA+EGFP vRNA (lower). The 3′- and 5′-end noncoding regions (NCRs; gray), and 183 nt and 157 nt in the NA-coding regions (black) were responsible for the efficient incorporation of vRNA into virus particles. An “autocleave” 2A peptide-coding sequence (gray and white text) was inserted between the EGFP (green) and NA (blank) genes. Protein translation start points (→) and stop points (▾) are shown. (B) Eight-plasmid virus rescue system. The bidirectional cytomegalovirus promoter (CMV)-bovine growth hormone poly(A) signal (pA) and Pol I-mTer transcription cassettes in the pM plasmid and the whole eight plasmids are shown. (C) RT-PCR detection of NA (upper), EGFP (middle), and NP (lower) from allantoic fluids. (D) EGFP expression in MDCK cells infected with influenza A virus rPR8-EGFP+NA or rPR8-NA+EGFP (scale bar = 100 μm). (E and F) Viral genome stability analysis. Recombinant influenza A viruses were consecutively inoculated into 10-day-old embryonated eggs for 7 passages. NA and EGFP genes were detected using RT-PCR. pM-PR8-EGFP+NA and pM-PR8-NA+EGFP plasmids were used as positive controls. DNA markers: for DL2000, 2,000, 1,000, 750, 500, 250, and 100 bp, and for DL15000, 15,000, 10,000, 7,500, 5,000, 2,500, 1,000, and 250 bp.

Recombinant influenza A viruses rPR8-EGFP+NA and rPR8-NA+EGFP were successfully rescued in the PR8 background by cotransfection with the other seven plasmids into 293T cells (Fig. 1B). After 48 h, culture supernatants were inoculated into 10-day-old embryonated chicken eggs. Two days later, progeny viruses were harvested from the allantoic fluid. Both rPR8-EGFP+NA and rPR8-NA+EGFP viruses were consecutively propagated for seven passages in 10-day-old embryonated eggs at about 50 PFU per egg. The genomic vRNAs were extracted with Trizol reagent (Invitrogen). Chimeric NA-, EGFP-, and nucleoprotein (NP)-specific bands were amplified by reverse transcription-PCR (RT-PCR) (Fig. 1C) using specific primers (10). Expression of EGFP mediated by influenza viruses was confirmed with MDCK cells at 24 h postinfection (p.i.) (Fig. 1D). Specific EGFP and NA recombinant vRNAs could be readily detected from passages 1 to 7 for both rPR8-EGFP+NA and rPR8-NA+EGFP viruses (Fig. 1E and F). These results indicated that recombinant influenza A viruses carrying EGFP in NA vRNA, as either rPR8-EGFP+NA or rPR8-NA+EGFP viruses, could be successfully generated and were stable throughout the course of our study.

Interestingly, expression of EGFP in MDCK cells infected with these two influenza A viruses showed different patterns in cellular distribution. Confocal laser scanning analyses were carried out using cells stained with 4′,6-diamidino-2-phenylindole (DAPI) fluorescent nuclear dye and NP-specific antibody (1:500; Southern Biotech) at 36 h p.i. EGFP was confirmed mainly located in the membrane in rPR8-EGFP+NA-infected MDCK cells, while EGFP was evenly distributed in rPR8-NA+EGFP-infected MDCK cells, including nuclei (Fig. 2AA). The difference in distribution could be attributed to the fact that in rPR8-EGFP+NA virus, fusion of the first 183 nt of NA mRNA (encoding 61 amino acids [aa], comprising the translocation signal and the transmembrane region) to the NH₂ terminus of EGFP (Fig. 2B) could lead to the targeting of the EGFP protein to the cell membrane. These viruses obtained not only maintained their genomic stability (Fig. 1E and F) but also were consistent for the patterns of EGFP expression in MDCK cells infected with viruses from passages 1, 4, and 7 (Fig. 2C).

FIG. 2.

Expression pattern of EGFP in recombinant influenza A virus-infected MDCK cells. (A) Confocal laser scanning analysis of EGFP localization. MDCK cells were infected with wtPR8, rPR8-EGFP+NA, and rPR8-NA+EGFP viruses. Cells transfected with the pEGFP plasmid were used as a reference control for original EGFP expression. Cells were fixed with formalin and stained with 4′,6-diamidino-2-phenylindole (DAPI) and NP-specific antibody (dilution = 1:500) 36 h later. Scale bar = 10 μm. (B) Schematic presentation of 2A self-cleavage in protein translation. EGFP and NA fusion proteins are translated from monocistronic mRNA, and the 2A peptide self-cleaves the fusion proteins into individual ones. The translation start and stop points are shown as in Fig. 1. (C) EGFP localization in MDCK cells infected with influenza viruses from passages 1, 4, and 7. Cells were fixed and stained with DAPI at 24 h p.i. and subjected to confocal laser scanning analysis. Scale bar = 10 μm.

Western blot analysis using an EGFP-specific antibody (diluted 1:1,000; Southern Biotech) showed specific EGFP protein bands in virus-infected MDCK cells, indicating that EGFP was efficiently cleaved by the 2A peptide from the fusion with NA (Fig. 3A). As expected, EGFP expressed by rPR8-EGFP+NA virus was larger than that expressed by rPR8-NA+EGFP, as the result of both the fusion of an additional 61-aa peptide from NA to the NH₂ terminus of EGFP and the fusion of the 2A peptide to the COOH terminus of EGFP (Fig. 2B).

FIG. 3.

Virological characteristics of influenza A viruses rPR8-EGFP+NA and rPR8-NA+EGFP. (A) Western blot analysis of EGFP in virus-infected MDCK cells. (B) Western blot analysis of EGFP in purified viral particles. Viral NP and M1 proteins were utilized as a normalization control. Cells transfected with the pEGFP plasmid were used as a positive control (Con.) for original EGFP. (C) Measurement of NA enzymatic activities in the purified viral particles. The enzymatic activity of the NA protein was measured using the NA substrate 2′-(4-methylumbelliferyl)-α-d-acetylneuraminic acid (MUNANA) (Sigma) and normalized to the level for the parental wtPR8 virus by use of HA titers. (D) Efficiencies of packaging of NP, PB2, and NA vRNA into viral particles. The abundances of the NP, PB2, and NA segments were qualified by SYBR green I-based quantitative PCR method with segment-specific primer pairs, and their relative abundances were normalized to the level for the wtPR8 virus. (E) Comparison of plaque sizes in MDCK cells infected with rPR8-EGFP+NA, rPR8-NA+EGFP and wtPR8 viruses. Cells were infected with diluted viruses and overlaid with 0.8% agarose gel. Plaques were immunostained with anti-NP monoclonal antibody (dilution = 1:500) at 48 h p.i. (F) Curves for virus growth in embryonated eggs. Ten-day-old embryonated chicken eggs were inoculated with 100 PFU of rPR8-EGFP+NA, rPR8-NA+EGFP, and wtPR8 and incubated at 37°C. The average titers of the virus from triplicate eggs at each time point are shown (mean ± standard deviation [SD]). Lg, log10. (G) Virus growth curves in MDCK cells. (H) Virus growth curves in A549 cells. (I) NA activities in MDCK cells. (J) NA activities in A549 cells. MDCK and A549 cells were infected at a multiplicity of infection (MOI) of 0.001. Cell culture media were collected at the different time points. The virus titers and NA activities (relative fluorescence units [RFU]) were measured. The data shown are means ± SD for triplicate wells at each time point.

It was reported that the fusion of the transmembrane domain of NA to bioactive cytokines could lead to their incorporation into viral particles (26). We next carried out an experiment to assess the presence of EGFP in the virions. Recombinant rPR8-EGFP+NA and rPR8-NA+EGFP influenza viruses were purified from allantoic fluid through two-step glucose cushion ultraspeed centrifugation (Beckman SW28 rotor) (5). We were able to detect the presence of the EGFP protein in rPR8-EGFP+NA viral particles but not in rPR8-NA+EGFP viral particles (Fig. 3B), confirming that the EGFP protein was incorporated into virions. The NA enzymatic activity from purified virus was measured using an NA substrate, 2′-(4-methylumbelliferyl)-α-d-acetylneuraminic acid (MUNANA) (Sigma), as described previously (1). The amounts of NA in rPR8-EGFP+NA and rPR8-NA+EGFP viruses were compared with those in wtPR8 virus under the same amount of virion (Fig. 3C). The rPR8-NA+EGFP virion possessed nearly as much NA activity (∼90%) as the wtPR8 virus, while the rPR8-EGFP+NA virion possessed much less NA activity (∼17%), which may be the result of EGFP interference and the truncation of the N-terminal region of the NA protein.

We wondered if the insertion of the EGFP gene into NA vRNA may affect the efficiency of packaging of NA vRNA into the virions. A SYBR green I-based quantitative PCR assay using segment-specific primers for NA, NP, and PB2 (16) was used to determine the efficiency of packaging of the NA segment into the virions. Compared wtPR8, nearly 80% of NA+EGFP vRNA segments were packaged into the rPR8-NA+EGFP virions, while fewer than 50% of EGFP+NA vRNA segments were packaged into the rPR8-EGFP+NA virions (Fig. 3D).

We next evaluated the growth characteristics of these recombinant viruses, as these viruses are different in fusion arrangement and in the amounts of NA vRNAs and NA proteins in their virions. The plaque-forming capabilities in infected MDCK cells were tested by immune staining with an anti-NP monoclonal antibody as described previously (5). The results revealed that the plaques formed by rPR8-EGFP+NA were smaller and that the plaques formed by rPR8-NA+EGFP were similar to wtPR8 (Fig. 3E), indicating a significant attenuation of the rPR8-EGFP+NA virus. We then compared the growth curves of the rPR8-EGFP+NA, rPR8-NA+EGFP, and wtPR8 viruses in embryonated eggs (Fig. 3F), in MDCK cells (Fig. 3G), and in A549 cells (Fig. 3H). Viral titers in allantoic fluid samples and cell culture medium collected at various time intervals were measured using a plaque formation assay. The rPR8-EGFP+NA virus was significantly attenuated, and its virus yield was lower than that of wtPR8. The rPR8-NA+EGFP virus maintained a growth rate similar to that for wtPR8. The NA activities measured for cell culture medium from MDCK and A549 cells infected with respective viruses also correlated with the growth characteristics of these viruses, with the rPR8-EGFP+NA virus having the lower NA activity level (Fig. 3I and J).

Replication-competent recombinant influenza A viruses expressing a reporter gene can be used in many applications for studying influenza A virus and developing vaccines and antiviral therapeutics. These viruses should have advantages for in vitro and in vivo studies over the pseudotyped lentivirus (20) and pseudotyped influenza reporter virus (17). We thus tested a mouse anti-PR8 serum by incubating a series of 5-fold diluted sera with the rPR8-NA+EGFP virus for 2 h, followed by infection of MDCK cells. EGFP-expressing cells were observed at 24 h p.i. under a fluorescence microscope (Fig. 4A), and mean fluorescence intensity was measured by fluorescence-activated cell sorting (FACS) (Fig. 4B). With increases in the titer of anti-PR8 serum, there were decreases in the number of EFGP-positive cells and the fluorescence intensity of EGFP. This proof-of-principle experiment indicated that replication-competent influenza A viruses with a reporter can be used as a rapid-readout system for screening neutralizing antibodies or antiviral drugs for influenza A virus.

FIG. 4.

Application of influenza A virus rPR8-NA+EGFP virus for evaluating antiviral neutralizing antibody. The rPR8-NA+EGFP virus was incubated in 5-fold-diluted mouse serum against the wtPR8 virus and negative-control serum (Con. Serum) and in 1× phosphate-buffered saline (PBS) buffer for 2 h before application to MDCK cells. At 24 h p.i., EGFP was observed under inverted fluorescence microscopy (A), and its expression levels were analyzed with FACS from triplicate wells (mean ± SD) (B). MFI, mean fluorescence intensity; mock, no virus infection control. Scale bar = 100 μm.

In this study, we demonstrated that the combination of 183 and 157 nt in the NA-coding region flanked by 3′ and 5′ NCRs and the use of a synthetic “autocleave” 2A peptide between EGFP and NA resulted in a great outcome. Previous reports indicate that neither the insertion of the EGFP gene into the influenza viral genome in the NS1- or NA-encoding segment nor the use of dicistronic expression cassettes was successful, as the EGFP expression was often unstable or the expression level was very low or even undetectable (11, 14, 24). An earlier report indicates that the insertion of the CAT gene with the 2A sequence into NA vRNA was exploited, but there were no follow-up reports for its application (22). We speculate that the disruption of NA vRNA packaging signals at the 3′ end in that study may have resulted in significant attenuation of the virus, which rendered scale-up production and application difficult. During the revision of this article, a newly published paper described the generation of a nine-segmented influenza A virus in which the PB1 open reading frame was flanked by the NA segment-specific packaging sequences, including the NCRs and the coding region packaging signals (6), indicating the importance of NA's coding region.

Our strategy of including 183 and 157 nt in the NA-coding region improved chimera NA vRNA packaging efficiency. Instead of using a different promoter (24) or internal ribosome entry site sequence, introduction of the 2A fusion peptide is the most simple and effective way to insert an exogenous gene into a vRNA segment. The presence of EGFP in the context of replication-competent influenza virus has several advantages. (i) The recombinant influenza viruses harboring the EGFP gene in the NA segment in both formats can be rescued and propagated to a large quantity in embryonated eggs and, importantly, are stable for multiple passages. (ii) The expression of EGFP allows direct measurement of viral infection and replication in target cells without the need for using chemical (e.g., NA enzyme assay), biological (e.g., plaque assay or hemagglutinin inhibition assay), and molecular (e.g., PCR) approaches which in general are more time-consuming and costly. (iii) For the rPR8-EGFP+NA virus, the fusion of the NA-anchoring transmembrane region to EGFP results in the localization of the EGFP protein to the cell membrane, thus enabling live monitoring the influenza virus in vivo. (iv) The rPR8-NA+EGFP virus has biological characteristics similar to those of the parental virus and thus has great potential for use in evaluating anti-influenza drugs and neutralizing antibodies.

In summary, we demonstrated a new strategy for generating replication-competent recombinant influenza A viruses containing a reporter gene, such as that encoding EGFP, in NA vRNA while maintaining all native viral genes. These reporter viruses can replicate and propagate without the need for complementary cells to provide NA proteins in trans and are stable for multiple passages of propagation in embryonated eggs. Future studies for incorporating luciferase into this format should enable more-powerful assessment of influenza virus infection and migration in animal models.