Specific Residues of the Influenza A Virus Hemagglutinin Viral RNA Are Important for Efficient Packaging into Budding Virions

Abstract

A final step in the influenza virus replication cycle is the assembly of the viral structural proteins and the packaging of the eight segments of viral RNA (vRNA) into a fully infectious virion. The process by which the RNA genome is packaged efficiently remains poorly understood. In an approach to analyze how vRNA is packaged, we rescued a seven-segmented virus lacking the hemagglutinin (HA) vRNA (deltaHA virus). This virus could be passaged in cells constitutively expressing HA protein, but it was attenuated in comparison to wild-type A/WSN/33 virus. Supplementing the deltaHA virus with an artificial segment containing green fluorescent protein (GFP) or red fluorescent protein (RFP) with HA packaging regions (45 3′ and 80 5′ nucleotides) partially restored the growth of this virus to wild-type levels. The absence of the HA vRNA in the deltaHA virus resulted in a 40 to 60% reduction in the packaging of the PA, NP, NA, M, and NS vRNAs, as measured by quantitative PCR (qPCR), and the packaging of these vRNAs was partially restored in the presence of GFP/RFP packaging constructs. To further define nucleotides of the HA coding sequence which are important for vRNA packaging, synonymous mutations were introduced into the full-length HA cDNA of influenza A/WSN/33 and A/Puerto Rico/8/34 viruses, and mutant viruses were rescued. qPCR analysis of vRNAs packaged in these mutant viruses identified a key region of the open reading frame (nucleotides 1659 to 1671) that is critical for the efficient packaging of an influenza virus H1 HA segment.

The influenza A virus genome is composed of eight negative-sense RNA segments (22, 26), which can encode up to 11 viral proteins (3, 23). Upon infection, the viral RNA (vRNA) is transported to the nucleus, the site of vRNA transcription and replication. At late stages of the infectious cycle, the viral ribonucleoprotein complex, composed of the three influenza polymerase proteins, the nucleoprotein, and vRNA, is exported from the nucleus in association with the influenza virus matrix (M1) and nuclear export (NEP) proteins (5). The final step of the virus replication cycle involves the assembly of the viral structural proteins and the packaging of the viral genome. For a virus particle to be fully infectious, it must contain a full complement of the eight vRNA segments. The process by which this packaging occurs is not well understood. However, it is known that vRNA is specifically packaged in preference to cellular RNAs and that the different vRNAs are present in an equimolar ratio within a population of virus particles (21).

Two models have been proposed for the packaging of vRNA into budding virions: the random incorporation model and the selective incorporation model (23). The random incorporation model assumes that a common structural feature is present on all vRNAs which enables them to be randomly incorporated into budding virions. In other words, the virus randomly incorporates vRNA into budding virions and does not differentiate among the different segments (4). This means that the likelihood of a virion obtaining a full complement of the eight vRNAs is determined entirely by chance. This model is supported by evidence that infectious viruses may possess more than eight vRNAs (8). Mathematical modeling suggests that if eight vRNAs are packaged randomly, 0.24% of released virus particles would contain a full complement of vRNAs and be infectious (1, 8). If 12 vRNAs were packaged randomly, the mathematical models suggest that infectivity increases to approximately 10%, which is comparable with the experimental data (6). As only 1 to 2% of the weight of the influenza virus particle is vRNA, it is difficult to accurately quantify the exact number of vRNA segments packaged. The earliest evidence suggesting that influenza vRNAs have packaging signals was from Luytjes et al. (14). Utilizing both the 3′ and 5′ terminal untranslated regions of the NS gene, it was possible to package a chloramphenicol acetyltransferase gene into influenza virus particles. This foreign gene was packaged into infectious particles and passaged several times, suggesting that the terminal 22 5′ and 26 3′ nucleotides are sufficient to provide the signals for RNA transcription and replication as well as for the packaging of RNA into influenza virus particles.

The selective incorporation model suggests that each vRNA segment contains a unique “packaging signal” allowing it to act independently, with each vRNA segment being packaged selectively. Evidence supporting this model comes from several reports demonstrating that defective interfering vRNA mutants can competitively inhibit the packaging of their normal counterparts but not that of other vRNAs (7, 17, 18). This selective incorporation model is supported by thin-section electron microscopy images of influenza virus particles that show eight and only eight distinct “dots,” presumably viral RNPs within virions (19).

There is now increasing evidence to support the theory of a specific packaging signal which includes nucleotides in the coding regions at both the 5′ and 3′ ends of the vRNAs. Such signals have been reported for all segments (10, 11, 13, 16, 20, 27). The exact mechanism by which individual vRNA segments are packaged is not known, but it has been hypothesized that packaging occurs via specific RNA-RNA or protein-RNA interactions (16).

Understanding the role of these packaging signals is complicated by sequence variation within the regions identified. A prime example of this is the packaging regions identified for the hemagglutinin (HA) vRNA (27). The packaging signals reported for the HA vRNA from the WSN strain of influenza A virus were 9 nucleotides at the 3′ end (vRNA sense) and 80 nucleotides at the 5′ end of the coding sequence (27). These regions are not highly conserved across the H1 HA subtype and are even less conserved across all 16 HA subtypes. However, it is well established that the reassortment of viruses between different subtypes with diverse sequences is possible, suggesting that there must be flexibility within these signals (or that smaller conserved regions are contained within the larger packaging signals).

The aim of the present work is to further characterize the sequences that are important for the packaging of the HA vRNA into replication-competent influenza A viruses. We demonstrate that although a seven-segmented virus can be rescued, replication is impaired, even when the missing HA protein is supplied by a complementing cell line. In the absence of an HA vRNA, the packaging of other vRNAs is also reduced. Constructs with coding sequences from both ends of the HA open reading frame (ORF) are sufficient for the efficient incorporation and passage of a green fluorescent protein (GFP)-/red fluorescent protein (RFP)-containing packaging construct. Incorporation of these constructs partially restored growth and packaging of vRNAs to wild-type (wt) levels. Mutational analysis of the HA packaging regions identified a 15-nucleotide region at the 5′ end of the HA ORF which must be conserved for efficient packaging. Synonymous changes introduced into this region resulted in a threefold reduction in HA vRNA packaging.

MATERIALS AND METHODS

Cell lines.

293T and MDCK cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and were maintained in Dulbecco's modified Eagle's medium and minimal essential medium, respectively (GIBCO, Carlsbad, CA), supplemented with 10% fetal calf serum (HyClone, Logan, UT) and penicillin-streptomycin (GIBCO).

Constructs and cloning.

The plasmids used for the rescue of recombinant influenza A/WSN/33 virus (WSN virus) have been described previously (9). Packaging constructs utilizing WSN HA packaging signals were prepared by subjecting the WSN HA vRNA rescue plasmid pPOLI-HA to site-directed mutagenesis to create an NheI restriction site at the end of the 3′ packaging region (45 nucleotides of coding sequence) and an XhoI restriction site at the end of the 5′ packaging region (80 nucleotides of coding sequence). The ORFs of GFP from pEGFP-C1 (BD Biosciences-Clontech, Palo Alto, CA) and RFP from pRSETb-mRFP (kindly provided by Roger Tsien) (2) were PCR amplified, incorporating the NheI and XhoI restriction sites, and cloned with the packaging regions flanking. The GFP-GFP insert was prepared by amplifying the GFP ORF with two different primer pairs, one incorporating NheI and EcoRI restriction sites and the other EcoRI and XhoI. The two PCR products were digested and cloned as described above in a three-way ligation. The protein expressed from this fusion construct is two GFP molecules with a hexaglycine linker between them.

To generate recombinant WSN viruses with mutated HA vRNA segments, a pPOLI plasmid containing the HA gene of WSN virus was subjected to site-directed mutagenesis using a Stratagene QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). Sequences of each mutated construct were confirmed by automated sequencing. To generate recombinant influenza A/Puerto Rico/8/34 viruses (PR/8 viruses) with mutated HA vRNA segments, a pGEM plasmid containing the HA gene of PR/8 virus was subjected to site-directed mutagenesis using a Stratagene QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). Sequences of each mutated construct were confirmed by automated sequencing. Correctly mutated PR/8 HA genes were then subcloned to pDZ (24). All primer sequences and vector maps are available upon request.

The plasmid pCAGGS-WSN-HA-HygR contains the WSN HA ORF, an internal ribosome entry site (IRES) element, and a hygromycin resistance gene. The IRES and hygromycin resistance genes were both PCR amplified from pQCXIH (Clontech, Mountain View, CA).

Generation of HA-MDCK cells.

MDCK cells were transfected with pCAGGS-WSN-HA-HygR by use of Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Twenty-four hours posttransfection, cells were plated into 150-mm dishes at a low density, and transfected cells were selected for by the addition of 0.2 mg/ml hygromycin B (Invitrogen). Cells were cloned utilizing cloning cylinders and screened for HA expression by immunofluorescence using an HA-specific monoclonal antibody, 2G9, and by Western blotting using a cross-reacting PR/8 polyclonal antiserum. The stably expressing HA-expressing MDCK (HA-MDCK) cells were maintained in medium containing 0.2 mg/ml hygromycin B.

Reverse genetics for recombinant viruses.

The reverse genetics system for the generation of recombinant influenza WSN (9) and PR/8 (25) viruses has been described previously. Briefly, for the generation of recombinant WSN viruses, a mix of 293T and MDCK cells was transfected with eight pPOLI vectors containing the viral genomic RNA segments transcribed using a human RNA polymerase I promoter and hepatitis D virus ribozyme, giving rise to correct 3′ and 5′ ends, and four pCAGGS protein expression vectors encoding the subunits of the viral polymerase and the nucleoprotein. After a 48-h incubation, the supernatant was passaged to fresh MDCK cells to amplify the generated viruses. For the rescue of HA vRNA-deficient viruses, HA-MDCK cells were used in the place of MDCK, and cells were transfected without the inclusion of the pPOLI-WSN HA plasmid. HA protein expression in 293T cells was facilitated by the inclusion of pCAGGS-WSN HA. For the generation of recombinant PR/8 viruses, 293T cells were transfected with eight pDZ plasmids (24), bidirectional plasmids containing the negative-sense viral genomic RNA segments and expressing viral protein from a positive-sense mRNA. After a 24-h incubation, cells were collected and inoculated into the allantoic cavities of 10-day-old embryonated chicken eggs and incubated at 37°C for 2 days. Each of the rescued viruses was further plaque purified, and mutations were confirmed by sequencing the HA genes of the newly generated viruses.

Viral growth kinetics.

A comparison of the viral growth kinetics of the HA vRNA-deficient recombinant viruses was undertaken with HA-MDCK cells. Nearly confluent six-well plates of HA-MDCK cells were infected with each of the viruses at a multiplicity of infection (MOI) of 0.001. Infected cells were maintained in minimal essential medium supplemented with 0.3% bovine albumin, 1% penicillin-streptomycin, and 1 μg/ml TPCK (l-1-tosylamide-2-phenylethyl chloromethyl ketone)-trypsin. The viral titer in the supernatant of infected cells was determined by plaque titration on HA-MDCK cells. The growth of mutant recombinant WSN viruses was evaluated by inoculation of MDCK cells at an MOI of 0.001 and incubation of cells for 48 h, at which time the supernatant was collected and the titer determined by plaque titration on MDCK cells. Growth kinetics of PR/8 viruses were determined by inoculation of 10-day-old eggs with 100 PFU of virus. Forty-eight hours postinoculation, allantoic fluid was harvested and the titer determined by plaque titration on MDCK cells.

Isolation of packaged vRNAs.

To analyze packaged vRNA, three 150-mm dishes of MDCK or HA-MDCK cells were infected with recombinant viruses. When maximal cytopathic effect was visible (approximately 36 to 48 h postinfection), the supernatant was collected and clarified by low-speed centrifugation. The supernatant was then further clarified by centrifugation at 10,000 rpm using a Beckman SW20 rotor (Beckman Coulter, Fullerton, CA). Clarified supernatant was then layered on a 30% sucrose cushion and further centrifuged at 25,000 rpm for 2.5 h. Pelleted virus was resuspended in TMK (10 mM Tris-HCl, pH 7.5, 1.5 mM MgCl₂, 10 mM KCl), and vRNA was extracted using TRIzol LS reagent (Invitrogen). Precipitated vRNA was resuspended in a final volume of 15 μl of 10 mM Tris-HCl, pH 8.0, and stored at −80°C. To analyze packaged vRNA for PR/8 mutant viruses, 10-day-old eggs were inoculated with approximately 1,000 PFU and incubated for 2 days. Allantoic fluid was harvested and processed as described above.

qPCR analysis of packaged vRNAs.

Extracted vRNAs (approximately 200 ng) were reverse transcribed using a universal 3′ primer (5′-AGGGCTCTTCGGCCAGCRAAAGCAGG) and Superscript II reverse transcriptase (RT) (Invitrogen). RT product was then diluted 10,000-fold and used as a template for quantitative PCR (qPCR). Separate PCRs were then carried out with segment-specific primers (Table 1) by use of a LightCycler 480 (Roche, Nutley, NJ). The 10-μl reaction mixture contained 1 μl of diluted RT product, SYBR green I (Molecular Probes), 0.5 μM of each primer, 200 μM deoxynucleoside triphosphate, 3 mM MgCl₂, 1 μl of 10× PCR buffer II, and 1U of AmpliTaq Gold enzyme (Applied Biosystems, Foster City, CA). Relative concentrations of vRNA were determined by analysis of cycle threshold values, normalizing for the total vRNA amount by equalizing the level of PB2 vRNA and then calculating the percentage of incorporation relative to the levels of wt vRNA packaging. Results are presented as the average incorporations of vRNA ± standard deviations, with results derived from two independent virus purifications and with vRNA levels quantified in triplicate.

TABLE 1.

Primers used for the qPCR analysis of packaged vRNA^a

Segment	Forward primer	Reverse primer
WSN PB2	GATCAGACCGAGTGATGGTA	ACATGGTCCATTCCTATTCC
WSN PB1	CTGCCAGAAGACAATGAACC	GGCCATTGCTTCCAATACAC
WSN PA	AGCAAGGCGAGTCAATAGTC	CAGGCTATTGTGCGATCTCT
WSN HA	TAACCTGCTCGAAGACAGAC	AGAGCCATCCGGTGATGTTA
WSN NP	GCGCCAAGCTAATAATGGTG	GGAGTGCCAGATCATCATGT
WSN NA	TTGGTCAGCAAGTGCATGTC	ACAGCCACTGCTCCATCATC
WSN M	TAGCCAGCACTACAGCTAAG	GGCCTGACTAGCAATATCCA
WSN NS	CAGAATGGACCAGGCGATCA	TAGAGTCTCCAGCCGGTCAA
WSN HA packaging	GGTGCTTTTGGTCTCCC	GCATATTCTGCACTGC
PR8 PB2	GGAGGCACAGGATGTAATCA	CGTTAGTTGCGATTCCGATG
PR8 PB1	ACACAAGGCCGACAGACCTA	ATGCTGTTGCAGCAGGTTGG
PR8 PA	GCACTCACTTGGAAGTATGC	TTGACTCGCCTTGCTCATTG
PR8 HA	CTGCTCGAAGACAGCCACAA	GAGCCATCCGGCGATGTTAC
PR8 NP	ACGGCTGGTCTGACTCACAT	TCCATTCCGGTGCGAACAAG
PR8 NA	TTGGTCAGCAAGTGCATGTC	ACAGCCACTGCTCCATTATC
PR8 M	CAAGCAGCAGAGGCCATGGA	GACCAGCACTGGAGCTAGGA
PR8 NS	AGCAGATAGTGGAGCGGATT	GTACAGAGGCCATGGTCATT

^aDifferent primer pairs were used for analysis of vRNA obtained from WSN and PR/8 viruses.

Analysis of vRNA replication.

To determine whether silent mutations were affecting vRNA replication rates, 293T cells were cotransfected with pCAGGS-WSN PB1, pCAGGS-WSN PB2, pCAGGS-WSN PA, pCAGGS-WSN NP, pPOLI-WSN NA, and the different pPOLI-WSN HA mutant plasmids. Forty-eight hours posttransfection, total RNA was extracted from cells by use of TRIzol reagent (Invitrogen). The RNA (approximately 2 μg) was then analyzed by qPCR as described above, with levels of HA vRNA normalized to levels of NA vRNA.

RESULTS

Preparation of HA-MDCK cells.

In order to complement the growth of an influenza virus lacking an HA vRNA, MDCK cells constitutively expressing WSN HA glycoprotein were generated. Cells were transfected with a protein-expressing pCAGGS-WSN HA plasmid which was modified by the insertion of an IRES element and a hygromycin resistance gene to provide selection. MDCK cells were transfected with this construct, and HA-expressing cells were selected by the addition of 0.2 mg/ml hygromycin B. Cells were screened for HA expression by immunofluorescence and Western blotting (Fig. 1).

FIG. 1.

Characterization of the HA-MDCK cell line. Stable expression of the HA glycoprotein in MDCK cells was assessed in comparison to the parental MDCK cell line. (A) Immunofluorescence using an HA-specific monoclonal antibody (2G9). Nuclei were stained using DAPI (4′,6′-diamidino-2-phenylindole). (B) Western blot using polyclonal anti-PR/8 antiserum. The blot was also probed using monoclonal anti-actin (Sigma) as a loading control.

Rescue of HA-deficient influenza A virus and supplementation with a reporter construct.

To generate a recombinant influenza WSN virus deficient in the HA vRNA, a mixture of 293T and HA-MDCK cells was transfected with only seven rescue plasmids encoding the WSN vRNA for PB2, PB1, PA, NP, NA, M, and NS, with the plasmid encoding the HA vRNA being left out. To make up for the lack of the HA protein, an HA protein-expressing plasmid, pCAGGS-WSN HA, was included in the transfection mix. Within 48 h postpassage of the transfection supernatant to fresh HA-MDCK cells, cytopathic effect was evident in the culture, and the supernatant tested positive for hemagglutination activity, confirming the rescue of a seven-segmented influenza A virus (deltaHA virus). To confirm that this virus was deficient in HA vRNA and therefore in expression of HA protein, the virus was used to inoculate parental MDCK cells. After incubation, no evidence of influenza virus infection was detected. However, the deltaHA virus grew in HA-MDCK cells and was successfully passaged six times in these cells without losing its ability to replicate.

To analyze the packaging signal requirements of HA vRNA, three constructs were created with fluorescent protein genes flanked by packaging regions (45 3′ and 80 5′ nucleotides) of the WSN HA vRNA segment (Fig. 2A). In earlier experiments in the laboratory, we were not able to rescue a virus with GFP flanked by the minimal HA packaging regions (9 3′ and 80 5′ nucleotides) identified by Watanabe et al. (27). Constructs containing 80 5′ nucleotides and either 9, 15, 30, 45, or 60 nucleotides from the 3′ end were tested. GFP-containing constructs were efficiently incorporated only with either 45 or 60 nucleotides derived from the 3′ end (data not shown). In order to determine whether fewer than 80 nucleotides from the 5′ end were sufficient for packaging, constructs were tested with 40 and 60 nucleotides. However, neither of these constructs packaged efficiently, as demonstrated by the lack of fluorescent protein expression following the passage of recombinant viruses in the transfection supernatant (data not shown). Therefore, in all subsequent work, the reporter constructs included 45 3′ and 80 5′ nucleotides from the HA vRNA. In order to determine whether the length of the vRNA or the sequence of reporter construct had any effect on the packaging of a vRNA with HA packaging regions, constructs were created either with a single GFP gene, with two tandem GFP genes, or with an RFP gene (Fig. 2A). By use of each of these constructs in the place of the HA vRNA, recombinant viruses were rescued and passaged to HA-MDCK cells.

FIG. 2.

Rescue and characterization of HA vRNA-deficient influenza A viruses with complementing GFP/RFP packaging constructs. (A) Schematic representation of the packaging constructs tested. Each construct consists of a GFP/RFP flanked by the terminal untranslated regions (black) along with 45 3′ and 80 5′ nucleotides (gray) required for efficient incorporation derived from the WSN HA vRNA. (B) Immunofluorescence of HA-MDCK cells infected with wt WSN, deltaHA, and HA(45)GFP-GFP(80) viruses. Infected cells were detected using an anti-M1 monoclonal antibody (mAb). Nuclei of cells were stained using DAPI. (C) Kinetics of growth of the following HA vRNA-deficient viruses in HA-MDCK cells infected at an MOI of 0.001 and titrated by plaque assay on HA-MDCK cells: wt WSN (wt), deltaHA (dHA), HA(45)GFP(80) (GFP), HA(45)GFP-GFP(80) (GFP/GFP), and HA(45)RFP(80) (RFP). Results represent two independent infections titrated in duplicate.

Immunofluorescence of rescued HA-deficient viruses.

To evaluate the growth and expression of the recombinant viruses, HA-MDCK cells were infected at an MOI of 0.001 and incubated for 24 h, at which time cells were fixed with methanol and stained for the expression of influenza A virus matrix protein. Figure 2B demonstrates the growth and spread of deltaHA and HA(45)GFP-GFP(80) packaged viruses. The HA(45)RFP(80) virus did not express RFP, although by RT-PCR the RFP gene could be detected (data not shown). Six independent rescues were performed for the latter virus, with loss of RFP expression occurring at each occasion. RFP genes from two rescues were sequenced. In one construct, nucleotides 25 to 27 (residue 9) of the RFP were changed from AAG (K) to a stop codon (TAG). In the other sequence, a guanidine nucleotide (nucleotides 57 to 59 [GGG] of the RFP) was lost, resulting in a 21-amino-acid truncated protein. This suggests that the RFP protein interfered with influenza virus replication and that only in cases where expression was lost could the virus replicate efficiently.

Growth kinetics of recombinant HA-deficient viruses.

The growth of recombinant viruses was then assessed in multicycle replication in HA-MDCK cells and compared to that of wt WSN virus. In a low-MOI infection, the growth of the deltaHA virus was reduced by approximately 3 log units compared to that of wt WSN (maximum titer of 4.6 × 10⁴ versus 2.5 × 10⁷; Fig. 2C). The inclusion of a GFP or RFP packaging construct in HA vRNA-deficient viruses increased the replication to 1 to 2 log units higher than the maximum titer of deltaHA. However the maximum titers obtained were still 1 to 2 log units lower than that for the wt WSN virus (Fig. 2C). No significant differences were observed in the growth kinetics of viruses with the different packaging constructs.

Analysis of the packaged vRNA.

The efficiencies of packaging of the different constructs and the changes due to the lack of an HA vRNA were analyzed by qPCR. Viruses were amplified in HA-MDCK cells and purified from culture supernatants by pelleting through a 30% sucrose cushion. vRNA was extracted from the semipurified viruses and reverse transcribed using a universal primer specific to the 12-nucleotide conserved 3′ end of the influenza vRNAs. In individual reactions, each gene segment was amplified using specific primers, and the percentage of each vRNA was determined in comparison to the vRNA levels of the recombinant wt WSN virus. Packaging for each virus was analyzed from two different preparations and quantified in triplicate. Quantitation required the assumption that each vRNA in a population of wt virus is present in an approximately equimolar ratio (21). The packaging of the different constructs in place of the HA vRNA was efficient in all cases (Table 2). Comparison of the packaging of HA(45)GFP(80) and HA(45)GFP-GFP(80) allows a determination of the effect of vRNA length. The HA(45)GFP(80) construct is 935 nucleotides in length, whereas the HA(45)GFP-GFP(80) construct is 1,685 nucleotides in length, very close to the HA vRNA length of 1,778 nucleotides for WSN virus. Although the packaging of both constructs was efficient, the HA(45)GFP-GFP(80) showed a higher level of incorporation into virions. The HA(45)RFP(80) construct in the virus tested did not package as efficiently as the HA(45)GFP(80) and HA(45)GFP-GFP(80) constructs or the wt HA vRNA.

TABLE 2.

Packaging of individual vRNAs into HA vRNA-deficient virions either in the presence or absence of a GFP/RFP packaging construct^a

Virus	% Incorporation of indicated vRNA into progeny virions
	PB2	PB1	PA	HA	NP	NA	M	NS	HA packaging
wt WSN	100.0 ± 8.3	100.0 ± 2.8	100.0 ± 3.8	100.0 ± 0.4	100.0 ± 1.5	100.0 ± 3.8	100.0 ± 1.8	100.0 ± 2.0	100.0 ± 1.9
deltaHA	100.0 ± 5.2	86.8 ± 4.3	52.8 ± 0.5	<0.001	43.7 ± 0.6	63.5 ± 1.6	44.3 ± 1.7	44.0 ± 1.4	<0.001
HA-GFP	100.0 ± 3.7	114.7 ± 1.7	111.7 ± 2.2	<0.001	73.2 ± 2.2	81.1 ± 4.9	63.2 ± 2.5	51.5 ± 1.2	94.1 ± 8.6
HA-GFP/GFP	100.0 ± 5.1	123.7 ± 5.5	118.0 ± 2.4	<0.001	70.0 ± 2.1	101.0 ± 1.4	54.8 ± 0.4	81.3 ± 1.0	142.4 ± 14.6
HA-RFP	100.0 ± 2.7	123.2 ± 6.0	103.6 ± 1.0	<0.001	51.2 ± 2.1	77.0 ± 1.5	33.7 ± 0.5	47.2 ± 1.6	56.1 ± 3.3

^aValues given are percentages ± standard deviations for packaging of vRNA in comparison to those for the parental WSN virus. Results are from two independent experiments, with assays performed in triplicate (n = 6).

The qPCR analysis also demonstrated that the loss of packaging of the HA vRNA had an adverse effect on the packaging of other vRNAs, particularly the PA, NP, NA, M, and NS vRNAs (Table 2). This reduction in packaging of other segments was partially restored to wt levels by the addition of a construct containing HA vRNA packaging regions. However, the packaging of the other segments was not completely restored to wt levels, even with the efficient packaging of the HA(45)GFP-GFP(80) vRNA. For example the NP, M, and NS vRNA in the HA(45)GFP-GFP(80) virus did not reach wt levels. No attempts were made to determine whether the lack of the HA vRNA for packaging results in an increase in the number of virions lacking all RNAs, that is, empty particles.

Mutational analysis of the WSN HA vRNA packaging regions.

To determine whether specific regions of the HA vRNA packaging signals are critical for the efficient incorporation of vRNA into recombinant viruses, mutational analysis of the HA vRNA was performed. Synonymous mutations were inserted into the rescue plasmid pPOLI-WSN HA. In general, pyrimidines were changed to purines and vice versa. Because these changes are silent, the effect on virus replication should be due largely to changes in vRNA packaging and not due to changes in protein function. The blocks of nucleotides changed are highlighted in Fig. 3, with changes introduced into the packaging regions at both the 3′ and 5′ ends of the vRNA. By use of these mutated constructs, recombinant viruses were rescued and amplified. RNA was then extracted, and the mutated sequences were confirmed by sequencing of the RT-PCR products.

FIG. 3.

Mutational analysis of the packaging regions of the WSN HA vRNA. (A) Schematic representation of the regions of synonymous mutations (white boxes) introduced into the WSN HA vRNA. (B) Synonymous nucleotide changes introduced for each construct; the upper line shows the parental WSN virus HA sequence, with nucleotide changes presented in bold in the lower lines. The numbering of nucleotides is based on the positive-sense RNA sequence.

To evaluate the replicative abilities of the different mutated recombinant viruses, MDCK cells were infected at an MOI of approximately 0.001, and cells were incubated for 48 h. At that time, supernatants were harvested and the titer of virus was determined by plaque assay. Growth levels of all viruses were reduced compared with that of the wt WSN virus. Of the seven mutant viruses, HA52, with synonymous mutations introduced in nucleotides 1659 to 1973 (Fig. 3), showed an approximately 1.5-log unit reduction in growth (Table 3). The plaques produced by this virus in MDCK cells were also only approximately 20% of the size of wt WSN plaques (data not shown). Analysis of the packaged vRNA from each of the mutant viruses demonstrated that the mutations in the HA52 virus resulted in a threefold reduction in the packaging of the HA vRNA (Table 4). This reduction in the packaging of the HA vRNA resulted in a reduction in the packaging of other gene segments as well, especially the PB1 vRNA.

TABLE 3.

Effect on replication in MDCK cells of recombinant WSN viruses with synonymous changes introduced within the packaging region of the HA vRNA^a

Virus	Plaque titer
WSN wt	2.2 × 107
WSN HA31	5.9 × 106
WSN HA51	6.7 × 106
WSN HA52	8.4 × 105
WSN HA53	8.5 × 106
WSN HA54	5.8 × 106
WSH HA55	1.1 × 107
WSH HA56	9.8 × 106

^aMDCK cells were infected at an MOI of 0.001 and incubated for 48 h. Virus produced was titrated by plaque assay in MDCK cells. Results represent the average titers from two independent infections titrated in duplicate (n = 4). Value for WSN HA52 is shown in boldface type.

TABLE 4.

Effect of synonymous changes within the packaging region of HA on the packaging of individual vRNAs into progeny WSN virions^a

Virus	% Incorporation of indicated vRNA into progeny virions
	PB2	PB1	PA	HA	NP	NA	M	NS
WSN wt	100.0 ± 6.1	100.0 ± 4.3	100.0 ± 4.3	100.0 ± 0.3	100.0 ± 2.3	100.0 ± 0.6	100.0 ± 1.1	100.0 ± 2.2
WSN HA31	100.0 ± 6.6	102.3 ± 2.6	115.4 ± 4.3	103.5 ± 2.1	88.7 ± 2.2	97.3 ± 0.3	97.2 ± 0.0	99.5 ± 0.8
WSN HA51	100.0 ± 3.9	107.9 ± 6.7	84.2 ± 3.7	96.0 ± 0.9	89.4 ± 4.8	96.4 ± 2.6	98.0 ± 1.0	97.6 ± 4.1
WSN HA52	100.0 ± 2.8	33.8 ± 1.1	69.5 ± 1.6	38.8 ± 1.7	70.6 ± 2.5	63.6 ± 1.2	118.4 ± 2.7	103.9 ± 2.5
WSN HA53	100.0 ± 1.5	110.7 ± 3.0	119.5 ± 3.7	107.8 ± 2.1	90.4 ± 5.4	103.9 ± 0.8	96.1 ± 1.0	98.7 ± 1.7
WSN HA54	100.0 ± 2.7	107.0 ± 2.0	87.1 ± 4.9	88.7 ± 3.5	89.4 ± 3.5	85.6 ± 2.1	102.7 ± 0.4	97.6 ± 1.9
WSN HA55	100.0 ± 0.9	98.6 ± 2.1	96.6 ± 1.9	101.6 ± 3.6	88.8 ± 3.8	100.0 ± 2.1	105.7 ± 0.1	98.5 ± 1.7
WSN HA56	100.0 ± 3.2	97.8 ± 1.3	102.6 ± 4.0	103.0 ± 0.6	82.9 ± 1.1	97.8 ± 2.6	98.2 ± 3.9	93.6 ± 1.3

^aValues given are percentages of packaging of vRNA in comparison to those of parental WSN virus ± standard deviations. Results are from two independent experiments, with assays performed in triplicate (n = 6). Values for WSN HA52 are shown in boldface type.

To rule out the possibility that this reduction in the packaging of the HA vRNA was a consequence of the lower replication of the mutated HA sequence, 293T cells were transfected with the different mutant pPOLI-WSN HA constructs along with plasmids expressing the influenza virus polymerases and NP. The replication of the vRNA expressed from the pPOLI constructs was assessed by qPCR and normalized to the replication of the NA vRNA from pPOLI-WSN NA as a control for transfection. The mutations introduced into the HA vRNAs had no effect on the ability of the influenza virus polymerase complex to recognize and replicate them. All mutant vRNAs were replicated to levels similar to those seen for the wt WSN HA vRNA (Table 5).

TABLE 5.

Lack of effect of synonymous changes on the replication of mutated HA vRNA by the influenza polymerase complex^a

Virus	Relative vRNA expression level
WSN wt	100
WSN HA31	106.0 ± 5.9
WSN HA51	82.0 ± 9.5
WSN HA52	118.8 ± 8.3
WSN HA53	116.0 ± 8.1
WSN HA54	108.9 ± 10.3
WSN HA55	93.8 ± 5.2
WSN HA56	112.0 ± 6.7

^aMutated pPOLI-WSN HA constructs were transfected into 293T cells along with protein expression plasmids for the influenza polymerase subunits and NP. Forty-eight hours posttransfection, total cellular RNA was extracted and vRNA replication was quantified by qPCR. Expression levels were normalized to a pPOLI-WSN NA control plasmid. Results are from two independent transfections, with assays performed in triplicate (n = 6).

Mutational analysis of the PR/8 HA vRNA packaging regions.

Previous studies aimed at identifying packaging signals for influenza A viruses all used WSN viruses. To determine whether similar regions exist for a related virus, a mutational analysis was also performed using PR/8 virus. Synonymous changes were introduced in blocks of approximately 15 nucleotides in both the 3′and 5′ ends of the PR/8 HA vRNA (Fig. 4). An extra 3′ region was mutated, extending the mutated region out to 45 nucleotides from the start codon. Each of the HA vRNA mutants was then rescued in a PR/8 background and amplified in 10-day-old embryonated chicken eggs. Growth kinetics of the rescued mutant viruses were determined for 10-day-old eggs by inoculation with 100 PFU of virus. Allantoic fluids were harvested 48 h postinoculation, and virus titers were determined on MDCK cells (Table 6). PR/8 HA52 virus was the only virus that did not grow similarly to wt PR/8, growing to a titer of 3.2 × 10⁸, approximately 1 log unit lower than that seen for the PR/8 wt control. Similar differences in virus replication were also observed at both earlier and later time points (data not shown).

FIG. 4.

Mutational analysis of the packaging regions of the PR/8 HA vRNA. (A) Schematic representation of the regions of synonymous mutations (white boxes) introduced into the PR/8 HA vRNA. (B) Synonymous nucleotide changes introduced at the 3′ and 5′ ends of the vRNA for each construct; the upper line shows the parental PR/8 virus HA sequence, with nucleotide changes presented in bold in the lower lines. The numbering of nucleotides is based on the positive-sense RNA sequence.

TABLE 6.

Effect on replication in embryonated eggs of recombinant PR/8 viruses with synonymous changes introduced within the packaging region of the HA vRNA^a

Virus2	Plaque titer
PR/8 wt	3.2 × 109
PR/8 HA31	1.9 × 109
PR/8 HA32	1.4 × 109
PR/8 HA51	1.5 × 109
PR/8 HA52	3.6 × 108
PR/8 HA53	1.4 × 109
PR/8 HA54	1.8 × 109
PR/8 HA55	2.0 × 109
PR/8 HA56	2.1 × 109

^aTen-day-old embryonated eggs were inoculated with 100 PFU of virus. Forty-eight hours postinoculation, virus titer in the allantoic fluid was assessed by plaque assay on MDCK cells. Results represent the average titers from two independent infections titrated in duplicate (n = 4). Value for PR/8 HA52 is shown in boldface type.

The packaging of vRNAs was analyzed by qPCR (Table 7), and like WSN HA52, PR/8 HA52 also showed an approximately threefold reduction in the packaging of the HA vRNA, suggesting that this region of the HA vRNA is important for efficient packaging in multiple virus isolates. The reduction in HA packaging also resulted in a reduction in the packaging of other gene segments, in particular the NA and M vRNAs. The reduction in the packaging of the PB1 vRNA in PR/8 HA52 was not as great as that seen for WSN HA52 (for WSN HA52 and PR/8 HA52, 33.8% and 70.7% incorporation, respectively). This may represent strain differences in packaging between WSN and PR/8. With PR/8 HA53 virus, although growth was not affected, there is a reduction in the packaging of the HA vRNA as well as that of other segments.

TABLE 7.

Effect of synonymous changes within the packaging region of PR/8 HA on the packaging of individual vRNAs into progeny virions^a

Virus	% Incorporation of indicated vRNA into progeny virions
	PB2	PB1	PA	HA	NP	NA	M	NS
PR/8 wt	100.0 ± 1.9	100.0 ± 2.7	100.0 ± 6.6	100.0 ± 5.9	100.0 ± 7.9	100.0 ± 2.5	100.0 ± 2.3	100.0 ± 3.5
PR/8 HA31	100.0 ± 3.2	91.4 ± 5.1	115.6 ± 4.9	95.9 ± 4.2	93.9 ± 2.1	103.6 ± 2.6	97.1 ± 4.1	95.2 ± 5.9
PR/8 HA32	100.0 ± 5.6	69.1 ± 3.3	100.8 ± 6.2	88.8 ± 6.3	100.3 ± 2.9	120.4 ± 6.4	99.4 ± 1.8	100.3 ± 9.6
PR/8 HA51	100.0 ± 4.5	101.6 ± 6.5	96.1 ± 7.1	108.0 ± 8.2	105.6 ± 3.4	100.5 ± 8.0	97.0 ± 1.9	108.4 ± 6.3
PR/8 HA52	100.0 ± 5.3	70.7 ± 4.2	72.9 ± 5.8	38.7 ± 3.4	80.2 ± 5.2	66.2 ± 4.9	62.2 ± 4.7	76.7 ± 1.4
PR/8 HA53	100.0 ± 3.5	64.7 ± 2.3	69.9 ± 9.1	71.3 ± 2.5	83.6 ± 3.0	82.1 ± 4.3	84.0 ± 1.2	80.7 ± 5.8
PR/8 HA54	100.0 ± 4.2	82. ± 5.3	91.7 ± 5.1	123.8 ± 7.4	96.5 ± 7.2	120.7 ± 7.9	103.4 ± 1.9	95.1 ± 9.0
PR/8 HA55	100.0 ± 5.6	93.1 ± 2.3	99.9 ± 2.5	118.9 ± 2.6	119.7 ± 4.5	118.0 ± 7.1	108.4 ± 1.6	100.5 ± 7.5
PR/8 HA56	100.0 ± 4.4	92.0 ± 6.2	99.1 ± 5.3	116.3 ± 3.4	96.3 ± 8.2	104.5 ± 10.2	94.3 ± 6.2	95.8 ± 8.7

^aValues given are percentages of packaging of vRNA in comparison to those of parental WSN virus ± standard deviations. Results are from two independent experiments, with assays performed in triplicate (n = 6). Values for PR/8 HA52 are shown in boldface type.

DISCUSSION

We created an MDCK cell line that constitutively expressed the HA glycoprotein from influenza A virus WSN. Utilizing this cell line, we rescued and passaged a recombinant influenza A virus that lacks the HA vRNA. This seven-segmented virus could be passaged in the HA-MDCK cell line without a full complement of the eight vRNAs. However, the growth of this virus was attenuated in comparison to that of the wt WSN virus. Also, the packaging of other vRNAs in the absence of the HA vRNA was affected, with reductions seen in the packaging of the PA, NP, NA, M, and NS gene segments. We do not believe that the reduction in the replication of deltaHA virus is due to insufficient levels of HA protein on the surfaces of infected cells. A virus with a mutated nonfunctional HA vRNA grew to levels similar to those seen for wt WSN virus in the HA-MDCK cells (data not shown). This finding supports the possibility that internal sequences (beyond those of the 45 3′ and 80 5′ nucleotides) of the HA vRNA are necessary for efficient incorporation of all eight of the vRNA segments.

Packaging of the GFP or RFP gene constructs into HA vRNA-deficient viruses was made possible by utilizing terminal nucleotides of both the 3′ and 5′ coding sequence of the WSN HA vRNA. Initially, constructs contained the 9 3′ and 80 5′ nucleotides reported by Watanabe et al. (27) as necessary and sufficient for the efficient incorporation and passage of a GFP for one cycle of replication. However, we were unable to obtain efficient incorporation of constructs containing only these packaging regions. In an attempt to boost the packaging of the GFP/RFP constructs, we prepared constructs with 15, 30, 45, and 60 nucleotides from the 3′ end of the HA vRNA. Although viruses were rescued on each attempt, efficient incorporation of the packaging construct was achieved only if 45 or 60 nucleotides were included. Constructs with 40 and 60 nucleotides from the 5′ end were also tested, but packaging was found to be inefficient. Using 45 3′ and 80 5′ nucleotides from the HA vRNA, we were able to passage the rescued viruses for multiple rounds, demonstrating the stable incorporation of the packaging constructs. The requirement for a longer 3′ packaging region for effective incorporation and multiple-cycle passaging of a GFP gene is different from that reported by Watanabe et al. (27) for the stable incorporation of a vesicular stomatitis virus G glycoprotein construct. The vesicular stomatitis virus G glycoprotein gene was packaged in place of the HA glycoprotein gene with 9 3′ and 45 5′ nucleotides. However, as the G glycoprotein was required for attachment and infection, the packaging of the gene was under selection. In the present study, the packaging of the GFP gene was not under any positive selection for retention. The requirement for a longer 3′ packaging region in our system may be due to different requirements for vRNA incorporation in viruses undergoing single-cycle (27) versus multiple-cycle (the present system) replication.

Incorporation of the packaging constructs into the deltaHA virus allowed a partial recovery of the wt replication phenotype and an increase in the packaging of other gene segments (Table 2). However, even with the efficient incorporation of the HA(45)GFP-GFP(80) construct, the packaging of other segments was not fully restored, suggesting that internal sequences of the HA vRNA may also be important for the optimal packaging of other segments. Alternatively the GFP sequences may have an inhibitory effect on packaging.

While it is possible to successfully passage a seven-vRNA-containing influenza virus in the (complementing) HA-MDCK cell line, we are unable to define a molecular model which accounts for the attenuation of this virus and the concomitant reduction in the packaging of several of the other segments. It is not clear whether the packaging signals studied here reflect RNA-RNA or RNA-protein interactions. Additional work is necessary to define the precise nature of these signals.

Further attempts to characterize the packaging signals were carried out by mutational analysis of the HA vRNA of WSN and PR/8 viruses. Synonymous changes were introduced into short regions of 12 to 18 nucleotides at both the 3′ and 5′ ends of the HA gene. The mutations of the HA vRNA creating the recombinant HA52 viruses resulted in a reduced ability to replicate and an approximately threefold reduction in the packaging of the HA vRNA. The region mutated in these viruses (nucleotides 1659 to 1673) is close to the end of the 80-nucleotide 5′ region identified by Watanabe et al. (27) as important for GFP incorporation. This region would be nucleotides 58 to 72 of the 80-nucleotide 5′ packaging region, confirming the essential role of this region described by Watanabe et al. (27).

The packaging signal identified in the HA52 viruses is conserved in HA sequences of human, swine, and avian influenza viruses of the H1 subtype (>90% nucleotide conservation in the region of 1662 to 1681). In contrast, neighboring regions of the H1 sequences show less than 70% sequence conservation. However, it should be noted that the packaging signal identified in the H1 HA vRNAs shows no sequence conservation across the 16 different subtypes of HA.

In an attempt to select for reversion of the attenuated phenotype seen with both the WSN HA52 and PR/8 HA52 viruses, they were passaged in MDCK cells and embryonated eggs, respectively. However, even after six sequential passages, neither virus had changed its growth potential and reverted to a wt level of replication, and sequence analysis of the HA vRNA showed no reversion back to the wt sequence (data not shown). These data suggest that multiple sequence changes may be required for the virus to restore vRNA packaging back to an equimolar ratio.

While this work was being completed, a study by Gog et al. (12) analyzed conserved regions of the PB2 and NA vRNAs. In their analysis of conserved regions, they noted that most clusters of conserved residues are located at the termini of the vRNAs, within previously identified packaging signals (11, 13, 16). They were able to demonstrate that changing the codon usage of individual highly conserved residues resulted in a reduced transduction of the GFP reporter construct when passaged to fresh MDCK cells. The role of these highly conserved residues in the context of the full-length vRNAs in infectious viruses is yet to be studied.

Conservation of sequence could be important for the maintenance of RNA secondary structures. The introduction of synonymous mutations could prevent correct folding by blocking intrasegmental base pairing or could prevent intersegmental interactions. Analysis of computer-predicted RNA folding using the Mfold program (15, 28) for HA52 for both WSN and PR/8 HA vRNAs in comparison to the parental sequence did not show any obvious structural changes (data not shown). However, the computer-predicted structures are based on naked RNA, and the structures of the influenza vRNA could be influenced by the binding of NP and the polymerase subunits.

For efficient packaging of GFP/RFP constructs into recombinant viruses, 45 coding nucleotides of the 3′ end of the HA vRNA were required. The introduction of synonymous changes within this region in the context of the full-length HA vRNA did not result in changes in packaging. The precise molecular mechanism by which these sequences facilitate packaging is not known.

In conclusion, we have rescued and passaged a seven-segmented influenza virus (deltaHA virus) in a complementing cell line. The lack of an HA vRNA resulted in reduced packaging of the PA, NP, NA, M, and NS vRNAs. Supplementing this virus with a GFP/RFP packaging construct partially restored replication and packaging to wt levels. Mutational analysis of the packaging regions in the HA vRNA identified a key sequence (nucleotides 1659 to 1671) which is necessary for efficient incorporation. While these data support a segment-specific packaging signal in the HA vRNAs of two H1 influenza viruses (WSN and PR/8 viruses), further studies of the packaging regions in other RNA segments of a variety of subtype viruses will be necessary in order to better understand the precise molecular mechanism by which vRNAs are incorporated into influenza virions.