cis-Acting Packaging Signals in the Influenza Viru
http://www.100md.com
病菌学杂志 2005年第16期
Department of Pathology and Laboratory Medicine, Emory University, Atlanta, Georgia 30322
ABSTRACT
The influenza A virus genome consists of eight negative-sense RNA segments. The cis-acting signals that allow these viral RNA segments (vRNAs) to be packaged into influenza virus particles have not been fully elucidated, although the 5' and 3' untranslated regions (UTRs) of each vRNA are known to be required. Efficient packaging of the NA, HA, and NS segments also requires coding sequences immediately adjacent to the UTRs, but it is not yet known whether the same is true of other vRNAs. By assaying packaging of genetically tagged vRNA reporters during plasmid-directed influenza virus assembly in cells, we have now mapped cis-acting sequences that are sufficient for packaging of the PA, PB1, and PB2 segments. We find that each involves portions of the distal coding regions. Efficient packaging of the PA or PB1 vRNAs requires at least 40 bases of 5' and 66 bases of 3' coding sequences, whereas packaging of the PB2 segment requires at least 80 bases of 5' coding region but is independent of coding sequences at the 3' end. Interestingly, artificial reporter vRNAs carrying mismatched ends (i.e., whose 5' and 3' ends are derived from different vRNA segments) were poorly packaged, implying that the two ends of any given vRNA may collaborate in forming specific structures to be recognized by the viral packaging machinery.
INTRODUCTION
The genome of influenza A virus consists of eight negative-sense RNA segments (called vRNAs) that together encode the 11 known viral proteins (2, 14, 17). During its assembly, a nascent influenza virus particle must incorporate (package) at least one copy of each vRNA in order to become infectious. The mechanisms that govern influenza virus RNA packaging are poorly understood. Although the virus packages its genomic segments selectively in preference to most cellular RNAs, it is not clear whether it discriminates among individual segments or packages them at random, and the molecular attributes of the vRNAs that allow them to be targeted into virions remain largely unknown.
Each vRNA consists predominantly of coding sequences (in antisense orientation), flanked at both ends by untranslated regions (UTRs) that range from 19 to 58 bases long. Within these UTRs, the distal 12 and 13 noncoding bases that form the extreme 3' and 5' termini, respectively, of every segment are highly conserved among viral strains and among the eight segments themselves (4, 20). These distal conserved sequences are partially complementary to each other and so can anneal to form a bulged duplex structure (1, 7, 9, 13) that is essential for transcription and replication of the segment (3, 8, 15, 18, 19). The UTRs are believed to harbor cis-acting signals that contribute to RNA packaging, since the attachment of authentic UTRs onto a heterologous RNA can enable it to be packaged into, and transduced by, influenza virus particles (16). Packaging mediated solely by the UTRs is inefficient, however, and it has been difficult to distinguish the sequences responsible for packaging from those needed for other critical aspects of viral gene expression and replication. Recent studies have revealed that optimal packaging of at least some segments requires not only both UTRs but also short portions of the coding region. In particular, deletion analysis of reporter constructs derived from the NA, HA, and NS segments indicates that the minimal sequences needed for efficient packaging extend beyond each UTR to include 9 to 80 bases of adjacent coding sequence at either end of the segment (10, 11, 22). For each of these three vRNAs, sequences at the 3' end of the coding region appear to exert a greater quantitative effect than those at the 5' end. Little is known about the specific attributes responsible for the packaging activity of these sequences.
In this study, we used a highly quantitative transduction assay to map the cis-acting packaging signals of three other influenza virus vRNAs, designated PA, PB1, and PB2. These are the three longest segments of the influenza virus genome and together encode the three protein subunits of the RNA-dependent RNA polymerase complex that is responsible for transcribing and replicating all eight genomic RNA segments. We found that, in these three vRNAs too, packaging signals extend into the distal coding regions adjacent to the UTRs. In contrast to the vRNAs previously analyzed, we found that the 3' coding sequences in the PA and PB1 segments are less critical than those near the 5' ends and, in the case of PB2, are entirely dispensable for packaging. We also created hybrid reporter vRNAs that combined the 5' signals of one RNA segment and the 3' signals from another. All such hybrids were packaged inefficiently, suggesting that sequences at the two ends of a given vRNA may participate in specific long-range interactions that are necessary for recognition by the viral packaging apparatus.
MATERIALS AND METHODS
Cells. 293T cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (FBS). MDCK and MDBK cells were maintained in Eagle's minimal essential medium supplemented with 5% FBS. After infection by influenza virus or virus-like particles, MDCK and MDBK cells were grown in L-15 medium (15 mM HEPES, pH 7.5, nonessential amino acids, 0.75 g of NaHCO3 per liter, and 0.125% [wt/vol] of bovine serum albumin).
Plasmids. The 17-plasmid and the 8-plasmid influenza A virus reverse genetic systems were obtained from Y. Kawaoka (University of Wisconsin) and G. Hobom (Justus Liebig University, Germany), respectively. To create the reporter construct of PA, PB1, or PB2, viral RNA in the pHH21 vector was digested with two convenient restriction enzymes (MscI and PstI for PA, MscI and EcoRI for PB1, and BglII and PvuII for PB2) to replace the internal coding region with the cDNA of the green fluorescent protein (GFP) gene in the same reading frame. The resulted plasmids, PA 474-G-239, PB1 433-G-627, and PB2 455-G-230, were named after the original vRNA segment, with a G (GFP) gene flanked by two numbers indicating the sizes of the retained coding regions at the 3' and 5' ends, respectively.
The GFP genes in all reporter constructs are flanked by unique restriction enzyme sites (MscI and BamHI in PA 474-G-239, MscI and SphI in PB1 433-G-627, and PstI and SphI in PB2 455-G-230) to facilitate the construction of serial deletions within the coding regions. To create deletions at the 3' end of the coding region in each GFP reporter construct, the fragment between the 5' end of the GFP open reading frame and the upstream NheI site in the vector sequence (the MscI-NheI fragments for the PA and PB1 constructs and the PstI-NheI fragment for the PB2 construct) was replaced with a series of PCR products, which were amplified with a 5' primer paired to the vector sequence at the NheI site and a series of 3' primers targeting different coding regions. These 3' primers also contain the engineered restriction enzyme sites (MscI for PA and PB1 and PstI for PB2) for cloning purposes. A similar strategy was used to create deletions at the 5' end of the coding region in each GFP reporter construct. The fragment between the 3' end of the GFP open reading frame and the downstream ApaI site in the vector sequence (the BamHI-ApaI fragment for PA and the SphI-ApaI fragments for PB1 and PB2) was replaced with a series of PCR products, which were amplified with 5' primers targeting different coding sequences with designed restriction enzyme sites (BamHI for PA and SphI for PB1 and PB2) and a 3' primer located at the ApaI site. The primer sequences will be provided upon request.
To exchange the 5' and 3' ends of different segments, each reporter construct was digested into two fragments with two unique restriction enzymes, NheI (in the pHH21 vector) and BrsGI (in the GFP-coding sequence), followed by a religation with swapped fragments. The resultant construct, PA (66)-G-PB1 (50), contains the 3' PA sequences including the 66-nucleotide coding region and the 5' PB1 sequences including the 50-nucleotide coding sequence. Likewise, PA (66)-G-PB2 (100), PB1 (66)-G-PA (50), and PB1 (66)-G-PB2 (100) each contain the 5' and 3' sequences from different segments as indicated in their names.
Generation of influenza viruses and virus-like particles. 293T cells in six-well plates were transfected with all eight influenza virus plasmids, together with 1 μg of a reporter construct, using the transfection reagent TransIT-LT1 (Panvera, Madison, Wisconsin) according to the manufacturer's protocol. After 6 h, the transfection media were replaced by fresh Opti-MEM supplemented with 0.1% FBS and 1% bovine serum albumin, and 293T cells were further incubated at 37°C for 48 h. The supernatants were then harvested, filtered through a 0.2-μm filter, and stored at –80°C.
Determination of relative packaging efficiency. Aliquots of supernatants from 293T cell transfection were used to infect MDBK or MDCK cells in six-well plates at 37°C for 1 h. After the infection medium was replaced with fresh L-15 medium, cells were further incubated at 37°C for 15 h and harvested for immunostaining. Cells were fixed in 4% paraformaldehyde and incubated first with mouse anti-influenza A virus NP antibody (Serotec) and then with a secondary R-phycoerythrin-conjugated anti-mouse immunoglobulin G polyclonal antibody (BD PharMingen). After extensive washing with phosphate-buffered saline, the cells were subjected to flow cytometric analysis to quantify the expression of both GFP (green) and NP (red) proteins. The percentage of doubly GFP- and NP-positive cells was used to indicate the relative packaging efficiency of each individual reporter.
Immunofluorescence assay. MDBK or MDCK cells seeded on chamber slides were infected with 293T cell supernatants for 15 h. The cells were fixed with paraformaldehyde and incubated with mouse anti-influenza A virus NP antibody (Serotec), followed by a secondary Alexa Fluor 594 goat anti-mouse immunoglobulin G antibody (Molecular Probes). After extensive washing with phosphate-buffered saline, the cells were mounted with Vectashield mounting medium with DAPI (4',6'-diamidino-2-phenylindole) (Vector Laboratories) and observed under fluorescence microscopy.
Quantitative real-time RT-PCR. Total RNA was extracted from plasmid-transfected 293T cells at 48 h posttransfection by using RNAbee reagent (TEL-TEST, Inc.) according to the manufacturer's protocol. The total RNA was cleared of plasmid DNA contamination by incubation for 30 min at 37°C with DNase I, which was then inactivated by heating to 85°C for 15 min. Reverse transcription was conducted using strand-specific primers for NP (5' CAGGATGTGCTCACTGATGC 3') and GFP (5' CAGAAGAACGGCATCAAGCG 3') according to the manufacturer's protocol for the SuperScript III first-strand synthesis system for reverse transcriptase PCR (RT-PCR) (Invitrogen). The quantitative real-time PCR was carried out in a 20-μl reaction mixture with gene-specific primers for NP (5' CAGGATGTGCTCACTGATGC 3' and 5' TTCTCCGTCCATTCTCACCC 3') or for GFP (5' CAGAAGAACGGCATCAAGCG 3' and 5' TGGGTGCTCAGGTAGTGGTTG 3'), using SYBR green DNA dye (Invitrogen). The PCR conditions were 50°C for 2 min, 95°C for 2 min, and 45 cycles of 95°C for 15 s, 55°C for 30 s, and 72°C for 30 s. The plasmids pPolI-NP and PA 474-G-239 were used as standards for the NP and GFP genes, respectively.
RESULTS
We created reporter plasmids that expressed PA, PB1, or PB2 vRNAs identical to those found in virions except that the bulk of the native coding region was replaced by in-frame fusion of a cDNA encoding GFP. The resulting fusions were designed to preserve the entire 3' and 5' UTRs along with at least 230 bases of adjacent viral coding sequence at each end, as studies of other segments had suggested that up to 80 bases of retained coding sequence might be needed to support maximally efficient packaging (10, 11, 22). The reporters were named by the identity of the parental vRNA, followed by the number of bases of viral coding sequence at the 3' end, the letter G (indicating GFP), and the number of bases of viral coding sequence at the 5' end. The longest versions of each construct tested were PA 474-G-239, PB1 433-G-627, and PB2 455-G-230 (Fig. 1A).
We used these reporters in conjunction with a plasmid-based reverse genetic system that generates influenza virus particles in transfected cells (Fig. 1B). This system comprises eight plasmid vectors that are transcribed in both orientations by opposed promoters to yield the eight wild-type vRNAs and all viral proteins needed to produce complete virions (12). When transfected into 293T cells, this set of plasmids directs the assembly and release of infectious virions that accumulate to levels of 106 to 107 PFU/ml in the supernatant after 48 h. For the present study, we transfected 293T cells with all eight wild-type plasmids together with one of our vRNA reporter vectors. When examined 48 h later under fluorescence microscopy, nearly all of the transfected 293T cells exhibited strong GFP fluorescence, indicating that vRNAs expressed from the reporter vectors could be transcribed by the viral polymerase to produce functional mRNA encoding GFP (data not shown). To determine whether the reporter vRNAs could be packaged and transduced by influenza virus particles, we harvested supernatant from the 293T cells and used it to infect MDBK cells, which were collected 15 h later and stained by indirect immunofluorescence to detect the influenza virus nucleoprotein (NP) as a marker for virally infected cells. The immunostained cell populations were then examined by fluorescence microscopy (Fig. 2A). For each reporter, nearly all MDBK cells were strongly positive for NP expression (red), an indication of productive viral infection. Most infected cells also expressed GFP (green), indicating that the reporter vRNA had been successfully transduced and expressed. Those results were quantified more precisely using two-color flow cytometry (Fig. 2B), which enabled us to rapidly score NP and GFP expression in 20,000 cells from each infected population. Under the conditions of this assay, we defined the relative packaging efficiency of a given reporter as the percentage of NP-positive MDBK target cells that were also GFP-positive. We found that our three full-length reporter vRNAs (PA 474-G-239, PB1 433-G-627, and PB2 455-G-230) were each incorporated at relative packaging efficiencies of 70 to 95%.
In the assay we used, both a reporter vRNA and its wild-type counterpart are expressed in the 293T cells where viral assembly occurs. Preliminary studies indicated that omitting any one of those three wild-type vRNAs dramatically reduced the titer of hemagglutinating particles released into the supernatant and completely eliminated plaque-forming activity, even when a vector encoding the corresponding viral protein (PA, PB1, or PB2) was added to the transfection mixture (data not shown). It is expected that all three subunits of the viral polymerase must be produced in an infected target cell in order to amplify and detectably express the reporter vRNA. In principle, the wild-type segments encoding those subunits could be delivered to the target cell either by the virion that carried the reporter or by coinfection with a helper virion from the same supernatant. We observed that, for a given reporter, near-maximal frequency of target cells coexpressing GFP and NP was obtained by infecting at a ratio of at least 1.5 PFU per MDBK cell (Fig. 2C), and that ratio was consequently used in our standard assay. Using much lower ratios, we found that a small percentage (typically 0.3 to 1.0%) of plaques were composed of GFP-expressing cells (Fig. 2D), indicating that they had arisen from individual virions that stably carried the reporter vRNA along with the eight wild-type segments needed for multiple rounds of infection. This confirms that some influenza virus virions can stably carry at least nine vRNAs, but it also indicates that the great majority of particles that packaged a reporter vRNA and delivered it into a target cell were not fully competent for infectivity and might require a helper virion in order to express GFP. With those caveats, the flow cytometric assay provided a means to map the sequences in each reporter vRNA that are required for efficient packaging.
Initial deletions showed that coding sequences contributed to packaging of all three reporters. As shown in Fig. 3A, for example, the packaging efficiency of our longest PA-derived reporter, with 474 bases of 3' and 239 bases of 5' coding sequences, was 91% (left), but this declined to roughly 7% when all 3' coding sequences were removed (right). The elimination of 5' coding sequences had profoundly deleterious effects, as it reduced the packaging efficiencies of all three reporters to less than 1% (Fig. 3B). The effects of 3' coding deletions were less pronounced: deleting all but 99 bases at this end caused no reduction in packaging of any of the three reporters, and the PA and PB1 vectors continued to be packaged at 7 to 10% efficiency when all 3' coding bases were removed. Interestingly, a PB2 reporter with only 99 bases of 3' coding sequence was packaged with reproducibly higher efficiency than one with 455 bases (see also Fig. 4C), and substantial packaging (60% efficiency) persisted even when all 3' coding bases were removed.
Further truncations of these reporters helped delineate the minimal sequences that could support packaging (Fig. 4). Setting an arbitrary threshold of 65% efficiency, we observed efficient packaging with as little as 66 bases and 40 bases of coding sequence at the 3' and 5' ends, respectively, of the PA vector (Fig. 4A), although significant residual activity was observed with as little as 15 bases of 3' or 30 bases of 5' coding sequence; the shortest fully active PA vector we tested was PA 99-G-40. Correspondingly, 66 bases of 3' and 40 bases of 5' coding sequence from PB1 proved compatible with full activity, and the shortest PB1 vector with maximal packaging was PB1 99-G-40, but intermediate activity required as few as 33 coding bases at the 3' end of this reporter (Fig. 4B). Full packaging efficiency required as few as 80 coding bases from the 5' end of PB2 and was exhibited by the vector PB2 99-G-80 (Fig. 4C), although the lack of any requirement for 3' PB2 coding sequences in other contexts suggests that further 3' truncation of that vector would be tolerated.
We asked whether the observed differences in packaging merely reflected differences in steady-state expression levels among these reporters in 293T cells. To that end, we assayed expression of selected informative reporter vRNAs at 48 h after transfection, using real-time PCR with primers that amplified a portion of the GFP cassette. Similar quantification of NP-specific vRNA was performed in parallel for each sample, and levels of GFP-specific reporter RNA were expressed in relationship to this internal standard. As indicated in Fig. 5A, we found that vRNA concentrations varied by up to sevenfold among the different truncated forms of any given reporter. PA derivatives generally were highly expressed, and PB1 expression seemed to correlate with the length of its 5' end, perhaps reflecting differences in RNA synthesis or stability. These differences could not, however, account for the salient differences in packaging efficiency that we observed. For example, PB1 99-G-40 was expressed at a slightly lower level than PB1 99-G-0 but was packaged 70-fold more efficiently, and the 4-fold-lower expression of PB2 99-G-60 compared to other PB2 constructs could not account for their 20-fold difference in packaging. Confirmatory evidence came from the finding that GFP fluorescence levels in the 293T producer cells appeared to be comparable among these reporters (Fig. 5B). We conclude that the observed variations in packaging do not simply reflect differences in vRNA concentration at the time of virion assembly.
Our findings, together with earlier reports, indicated that critical packaging signals are localized to two separate regions at the 3' and 5' ends of a given vRNA. Apart from the highly conserved 12- and 13-base distal sequences of these ends, the remaining sequences of the UTRs and adjacent coding regions that harbor these signals are unique to each segment, which led us to ask whether the signals at each end of a vRNA are themselves functionally autonomous and interchangeable. We therefore designed a series of six hybrid GFP reporters, each of which combined the 3' and 5' packaging sequences from two different genome segments (Fig. 6A). In each case, these packaging sequences comprised a complete UTR along with sufficient adjacent coding sequences to support maximal packaging of the parental vRNA. We transfected each of these hybrid vectors into 293T cells along with the eight-plasmid assembly system and assayed the resulting levels of GFP and reporter vRNA expression and the efficiency of packaging into virions. Three additional reporters, carrying these same ends in their normal, homologous pairings, were tested in parallel. As illustrated in Fig. 6B, all nine reporters gave rise to GFP fluorescence, indicating that all were competent for transcription, replication, and expression, although the GFP levels from some of the hybrid vectors were noticeably reduced. Consistent with that finding, reporter vRNA concentrations in the 293T cells (each normalized to NP vRNA as an internal control) were generally higher for homologous than for hybrid vectors. Apart from the homologous PA reporter, however (which, as in our earlier comparisons, was the most highly expressed), the concentrations of these reporters varied no more than sevenfold. In striking contrast, whereas the three homologous reporters were each packaged with high efficiency (85 to 93%), all of the hybrid vRNAs were packaged very poorly, in most cases with efficiencies of less than 0.5%. Only one hybrid reporter (combining ends from PB2 and PB1) showed appreciable packaging, at 11% efficiency. These results demonstrate that the bipartite packaging signals of influenza virus vRNAs function only in their native combinations.
DISCUSSION
By systematically truncating reporter vRNAs derived from the influenza virus PA, PB1, and PB2 genome segments, we have mapped cis-acting sequences that are necessary and sufficient to enable each of these segments to be packaged efficiently into virions. Our study utilized a plasmid-based assembly system that encodes all of the viral proteins and RNAs needed to produce live, replication-competent influenza virus in transfected 293T cells. We designed a series of reporter vRNAs encoding GFP fusion proteins flanked by various lengths of viral RNA sequence, coexpressed these along with the nascent virions in 293T producer cells, and asked whether the resulting viral particles could transduce a given GFP reporter into MDBK target cells. This approach parallels that used by Kawaoka and colleagues to map packaging signals in the NA, HA, and NS segments (10, 11, 22), except that we employed flow cytometry to quantify GFP expression in the target cells. In addition, whereas those investigators routinely omitted the wild-type counterpart of each reporter under study from their assembly reactions, we found it essential to coexpress both the reporter and the corresponding wild-type vRNA. This presumably reflects the fact that the three vRNAs we studied each code for an essential subunit of the viral polymerase complex, whose activity is necessary to support both virus production by the 293T cells and GFP expression in target cells. Thus, to the extent that influenza virus vRNAs compete for packaging, our assay measured the efficiency with which a given reporter vRNA could compete against the full complement of wild-type vRNAs.
Despite those methodological differences, our findings fully accord with earlier reports in demonstrating that sequences within the distal coding regions contribute strongly to the packaging of influenza virus genome segments. With the addition of the three segments analyzed here, six of the eight viral segments have now been characterized, and in each case maximal packaging has been shown to require up to 80 bases of coding sequence adjacent to the UTRs at either end of a segment. Although coding sequences at both the 3' and 5' ends contribute to packaging of most segments, our results suggest that the PB2 segment may be unique in requiring such sequences only at its 5' end. Taking into account the various lengths of the UTRs (which range from 19 to 58 bases among the six vRNAs studied so far), available data thus indicate that all cis-acting signals needed for efficient packaging are localized within two separate regions, each 27 to 125 bases long, at the 3' and 5' termini of each segment.
Our results also confirm earlier reports that some influenza virus particles can package at least nine vRNAs (6, 21), giving rise to viral strains that stably transmit a reporter vRNA in addition to the eight segments needed for infectivity. Such particles appear to be relatively infrequent, however, accounting for fewer than 1% of plaque-forming virions in our study. Indeed, many of the GFP transduction events scored in our assay are likely to have involved particles that packaged reporter without the full complement of other genome segments; in such cases, expression of GFP required coinfection by helper virions, which we estimate were at least 2.5-fold more abundant than reporter-transducing particles in the viral supernatants we studied.
Additional studies will be needed to localize the packaging signals more precisely, as distinct from the signals responsible for stability, transcription, replication, and other critical properties of these viral RNAs. At present, the molecular characteristics of the packaging signals remain unknown, including the specific aspects of RNA sequence or secondary structure that confer packaging activity. A better understanding of those signals would be useful in engineering genetically modified influenza viruses, including influenza virus vectors that stably incorporate heterologous genes, and it would also shed light on the processes that give rise to interspecies recombination between influenza viruses, an important source of new pandemic strains. The finding that key packaging signals reside within portions of the UTRs and coding regions whose sequences are unique to a given vRNA offers a potential means by which the packaging mechanism could specifically discriminate one vRNA from another and so package its genome segments specifically rather than randomly. Such discrimination might be achieved though segment-specific protein-RNA or RNA-RNA interactions and might account for reports that certain defective-interfering mutant vRNAs selectively compete for packaging against their wild-type counterparts (5).
As one further step in characterizing the influenza virus packaging signals, we asked whether the two terminal packaging regions of a vRNA are functionally autonomous. We created six hybrid GFP vectors that combined 3' and 5' packaging regions from different vRNAs, and we asked whether these alternative pairings could support packaging into virions. When expressed in 293T cells along with an active viral polymerase, all six hybrid reporters gave rise to GFP fluorescence (although in some cases at lower levels than the parental reporters), indicating that they were competent to direct protein expression. Nevertheless, all six hybrids proved severely defective in packaging, with only one exhibiting even modest activity. Even the 5' terminus from the PB2 segment, which functioned independently of any 3' coding sequences in its native vRNA, was completely inert when paired with 3' sequences from the PA or PB1 segments. Thus, influenza virus packaging signals that are fully active in their normal combinations cannot be interchanged. This implies that the two separate packaging regions at the ends of an influenza virus vRNA form a uniquely interdependent pair and that they must undergo critical segment-specific physical and/or functional interactions to be recognized by the packaging machinery.
ACKNOWLEDGMENTS
We thank Y. Kawaoka and G. Hobom for the influenza virus plasmids and H. Ly for critical review of the manuscript.
This work was supported by NIH grant AI-40317.
REFERENCES
Bae, S. H., H. K. Cheong, J. H. Lee, C. Cheong, M. Kainosho, and B. S. Choi. 2001. Structural features of an influenza virus promoter and their implications for viral RNA synthesis. Proc. Natl. Acad. Sci. USA 98:10602-10607.
Chen, W., P. A. Calvo, D. Malide, J. Gibbs, U. Schubert, I. Bacik, S. Basta, R. O'Neill, J. Schickli, P. Palese, P. Henklein, J. R. Bennink, and J. W. Yewdell. 2001. A novel influenza A virus mitochondrial protein that induces cell death. Nat. Med. 7:1306-1312.
Crow, M., T. Deng, M. Addley, and G. G. Brownlee. 2004. Mutational analysis of the influenza virus cRNA promoter and identification of nucleotides critical for replication. J. Virol. 78:6263-6270.
Desselberger, U., V. R. Racaniello, J. J. Zazra, and P. Palese. 1980. The 3'- and 5'-terminal sequences of influenza A, B and C virus RNA segments are highly conserved and show partial inverted complementarity. Gene 8:315-328.
Duhaut, S. D., and J. W. McCauley. 1996. Defective RNAs inhibit the assembly of influenza virus genome segments in a segment-specific manner. Virology 216:326-337.
Enami, M., G. Sharma, C. Benham, and P. Palese. 1991. An influenza virus containing nine different RNA segments. Virology 185:291-298.
Flick, R., and G. Hobom. 1999. Interaction of influenza virus polymerase with viral RNA in the ‘corkscrew’ conformation. J. Gen. Virol. 80:2565-2572.
Flick, R., G. Neumann, E. Hoffmann, E. Neumeier, and G. Hobom. 1996. Promoter elements in the influenza vRNA terminal structure. RNA 2:1046-1057.
Fodor, E., D. C. Pritlove, and G. G. Brownlee. 1995. Characterization of the RNA-fork model of virion RNA in the initiation of transcription in influenza A virus. J. Virol. 69:4012-4019.
Fujii, K., Y. Fujii, T. Noda, Y. Muramoto, T. Watanabe, A. Takada, H. Goto, T. Horimoto, and Y. Kawaoka. 2005. Importance of both the coding and the segment-specific noncoding regions of the influenza A virus NS segment for its efficient incorporation into virions. J. Virol. 79:3766-3774.
Fujii, Y., H. Goto, T. Watanabe, T. Yoshida, and Y. Kawaoka. 2003. Selective incorporation of influenza virus RNA segments into virions. Proc. Natl. Acad. Sci. USA 100:2002-2007.
Hoffmann, E., G. Neumann, Y. Kawaoka, G. Hobom, and R. G. Webster. 2000. A DNA transfection system for generation of influenza A virus from eight plasmids. Proc. Natl. Acad. Sci. USA 97:6108-6113.
Hsu, M. T., J. D. Parvin, S. Gupta, M. Krystal, and P. Palese. 1987. Genomic RNAs of influenza viruses are held in a circular conformation in virions and in infected cells by a terminal panhandle. Proc. Natl. Acad. Sci. USA 84:8140-8144.
Lamb, R. A., and R. M. Krug. 2003. Orthomyxoviridae: the viruses and their replication, p. 1487-1531. In B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Fields virology, 4th ed., vol. 2. Lippincott-Raven Publishers, Philadelphia, Pa.
Li, X., and P. Palese. 1992. Mutational analysis of the promoter required for influenza virus virion RNA synthesis. J. Virol. 66:4331-4338.
Luytjes, W., M. Krystal, M. Enami, J. D. Pavin, and P. Palese. 1989. Amplification, expression, and packaging of foreign gene by influenza virus. Cell 59:1107-1113.
McGeoch, D., P. Fellner, and C. Newton. 1976. Influenza virus genome consists of eight distinct RNA species. Proc. Natl. Acad. Sci. USA 73:3045-3049.
Neumann, G., and G. Hobom. 1995. Mutational analysis of influenza virus promoter elements in vivo. J. Gen. Virol. 76:1709-1717.
Piccone, M. E., A. Fernandez-Sesma, and P. Palese. 1993. Mutational analysis of the influenza virus vRNA promoter. Virus Res. 28:99-112.
Robertson, J. S. 1979. 5' and 3' terminal nucleotide sequences of the RNA genome segments of influenza virus. Nucleic Acids Res. 6:3745-3757.
Scholtissek, C., W. Rohde, E. Harms, R. Rott, M. Orlich, and C. B. Boschek. 1978. A possible partial heterozygote of an influenza A virus. Virology 89:506-516.
Watanabe, T., S. Watanabe, T. Noda, Y. Fujii, and Y. Kawaoka. 2003. Exploitation of nucleic acid packaging signals to generate a novel influenza virus-based vector stably expressing two foreign genes. J. Virol. 77:10575-10583.(Yuying Liang, Ying Hong, )
ABSTRACT
The influenza A virus genome consists of eight negative-sense RNA segments. The cis-acting signals that allow these viral RNA segments (vRNAs) to be packaged into influenza virus particles have not been fully elucidated, although the 5' and 3' untranslated regions (UTRs) of each vRNA are known to be required. Efficient packaging of the NA, HA, and NS segments also requires coding sequences immediately adjacent to the UTRs, but it is not yet known whether the same is true of other vRNAs. By assaying packaging of genetically tagged vRNA reporters during plasmid-directed influenza virus assembly in cells, we have now mapped cis-acting sequences that are sufficient for packaging of the PA, PB1, and PB2 segments. We find that each involves portions of the distal coding regions. Efficient packaging of the PA or PB1 vRNAs requires at least 40 bases of 5' and 66 bases of 3' coding sequences, whereas packaging of the PB2 segment requires at least 80 bases of 5' coding region but is independent of coding sequences at the 3' end. Interestingly, artificial reporter vRNAs carrying mismatched ends (i.e., whose 5' and 3' ends are derived from different vRNA segments) were poorly packaged, implying that the two ends of any given vRNA may collaborate in forming specific structures to be recognized by the viral packaging machinery.
INTRODUCTION
The genome of influenza A virus consists of eight negative-sense RNA segments (called vRNAs) that together encode the 11 known viral proteins (2, 14, 17). During its assembly, a nascent influenza virus particle must incorporate (package) at least one copy of each vRNA in order to become infectious. The mechanisms that govern influenza virus RNA packaging are poorly understood. Although the virus packages its genomic segments selectively in preference to most cellular RNAs, it is not clear whether it discriminates among individual segments or packages them at random, and the molecular attributes of the vRNAs that allow them to be targeted into virions remain largely unknown.
Each vRNA consists predominantly of coding sequences (in antisense orientation), flanked at both ends by untranslated regions (UTRs) that range from 19 to 58 bases long. Within these UTRs, the distal 12 and 13 noncoding bases that form the extreme 3' and 5' termini, respectively, of every segment are highly conserved among viral strains and among the eight segments themselves (4, 20). These distal conserved sequences are partially complementary to each other and so can anneal to form a bulged duplex structure (1, 7, 9, 13) that is essential for transcription and replication of the segment (3, 8, 15, 18, 19). The UTRs are believed to harbor cis-acting signals that contribute to RNA packaging, since the attachment of authentic UTRs onto a heterologous RNA can enable it to be packaged into, and transduced by, influenza virus particles (16). Packaging mediated solely by the UTRs is inefficient, however, and it has been difficult to distinguish the sequences responsible for packaging from those needed for other critical aspects of viral gene expression and replication. Recent studies have revealed that optimal packaging of at least some segments requires not only both UTRs but also short portions of the coding region. In particular, deletion analysis of reporter constructs derived from the NA, HA, and NS segments indicates that the minimal sequences needed for efficient packaging extend beyond each UTR to include 9 to 80 bases of adjacent coding sequence at either end of the segment (10, 11, 22). For each of these three vRNAs, sequences at the 3' end of the coding region appear to exert a greater quantitative effect than those at the 5' end. Little is known about the specific attributes responsible for the packaging activity of these sequences.
In this study, we used a highly quantitative transduction assay to map the cis-acting packaging signals of three other influenza virus vRNAs, designated PA, PB1, and PB2. These are the three longest segments of the influenza virus genome and together encode the three protein subunits of the RNA-dependent RNA polymerase complex that is responsible for transcribing and replicating all eight genomic RNA segments. We found that, in these three vRNAs too, packaging signals extend into the distal coding regions adjacent to the UTRs. In contrast to the vRNAs previously analyzed, we found that the 3' coding sequences in the PA and PB1 segments are less critical than those near the 5' ends and, in the case of PB2, are entirely dispensable for packaging. We also created hybrid reporter vRNAs that combined the 5' signals of one RNA segment and the 3' signals from another. All such hybrids were packaged inefficiently, suggesting that sequences at the two ends of a given vRNA may participate in specific long-range interactions that are necessary for recognition by the viral packaging apparatus.
MATERIALS AND METHODS
Cells. 293T cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (FBS). MDCK and MDBK cells were maintained in Eagle's minimal essential medium supplemented with 5% FBS. After infection by influenza virus or virus-like particles, MDCK and MDBK cells were grown in L-15 medium (15 mM HEPES, pH 7.5, nonessential amino acids, 0.75 g of NaHCO3 per liter, and 0.125% [wt/vol] of bovine serum albumin).
Plasmids. The 17-plasmid and the 8-plasmid influenza A virus reverse genetic systems were obtained from Y. Kawaoka (University of Wisconsin) and G. Hobom (Justus Liebig University, Germany), respectively. To create the reporter construct of PA, PB1, or PB2, viral RNA in the pHH21 vector was digested with two convenient restriction enzymes (MscI and PstI for PA, MscI and EcoRI for PB1, and BglII and PvuII for PB2) to replace the internal coding region with the cDNA of the green fluorescent protein (GFP) gene in the same reading frame. The resulted plasmids, PA 474-G-239, PB1 433-G-627, and PB2 455-G-230, were named after the original vRNA segment, with a G (GFP) gene flanked by two numbers indicating the sizes of the retained coding regions at the 3' and 5' ends, respectively.
The GFP genes in all reporter constructs are flanked by unique restriction enzyme sites (MscI and BamHI in PA 474-G-239, MscI and SphI in PB1 433-G-627, and PstI and SphI in PB2 455-G-230) to facilitate the construction of serial deletions within the coding regions. To create deletions at the 3' end of the coding region in each GFP reporter construct, the fragment between the 5' end of the GFP open reading frame and the upstream NheI site in the vector sequence (the MscI-NheI fragments for the PA and PB1 constructs and the PstI-NheI fragment for the PB2 construct) was replaced with a series of PCR products, which were amplified with a 5' primer paired to the vector sequence at the NheI site and a series of 3' primers targeting different coding regions. These 3' primers also contain the engineered restriction enzyme sites (MscI for PA and PB1 and PstI for PB2) for cloning purposes. A similar strategy was used to create deletions at the 5' end of the coding region in each GFP reporter construct. The fragment between the 3' end of the GFP open reading frame and the downstream ApaI site in the vector sequence (the BamHI-ApaI fragment for PA and the SphI-ApaI fragments for PB1 and PB2) was replaced with a series of PCR products, which were amplified with 5' primers targeting different coding sequences with designed restriction enzyme sites (BamHI for PA and SphI for PB1 and PB2) and a 3' primer located at the ApaI site. The primer sequences will be provided upon request.
To exchange the 5' and 3' ends of different segments, each reporter construct was digested into two fragments with two unique restriction enzymes, NheI (in the pHH21 vector) and BrsGI (in the GFP-coding sequence), followed by a religation with swapped fragments. The resultant construct, PA (66)-G-PB1 (50), contains the 3' PA sequences including the 66-nucleotide coding region and the 5' PB1 sequences including the 50-nucleotide coding sequence. Likewise, PA (66)-G-PB2 (100), PB1 (66)-G-PA (50), and PB1 (66)-G-PB2 (100) each contain the 5' and 3' sequences from different segments as indicated in their names.
Generation of influenza viruses and virus-like particles. 293T cells in six-well plates were transfected with all eight influenza virus plasmids, together with 1 μg of a reporter construct, using the transfection reagent TransIT-LT1 (Panvera, Madison, Wisconsin) according to the manufacturer's protocol. After 6 h, the transfection media were replaced by fresh Opti-MEM supplemented with 0.1% FBS and 1% bovine serum albumin, and 293T cells were further incubated at 37°C for 48 h. The supernatants were then harvested, filtered through a 0.2-μm filter, and stored at –80°C.
Determination of relative packaging efficiency. Aliquots of supernatants from 293T cell transfection were used to infect MDBK or MDCK cells in six-well plates at 37°C for 1 h. After the infection medium was replaced with fresh L-15 medium, cells were further incubated at 37°C for 15 h and harvested for immunostaining. Cells were fixed in 4% paraformaldehyde and incubated first with mouse anti-influenza A virus NP antibody (Serotec) and then with a secondary R-phycoerythrin-conjugated anti-mouse immunoglobulin G polyclonal antibody (BD PharMingen). After extensive washing with phosphate-buffered saline, the cells were subjected to flow cytometric analysis to quantify the expression of both GFP (green) and NP (red) proteins. The percentage of doubly GFP- and NP-positive cells was used to indicate the relative packaging efficiency of each individual reporter.
Immunofluorescence assay. MDBK or MDCK cells seeded on chamber slides were infected with 293T cell supernatants for 15 h. The cells were fixed with paraformaldehyde and incubated with mouse anti-influenza A virus NP antibody (Serotec), followed by a secondary Alexa Fluor 594 goat anti-mouse immunoglobulin G antibody (Molecular Probes). After extensive washing with phosphate-buffered saline, the cells were mounted with Vectashield mounting medium with DAPI (4',6'-diamidino-2-phenylindole) (Vector Laboratories) and observed under fluorescence microscopy.
Quantitative real-time RT-PCR. Total RNA was extracted from plasmid-transfected 293T cells at 48 h posttransfection by using RNAbee reagent (TEL-TEST, Inc.) according to the manufacturer's protocol. The total RNA was cleared of plasmid DNA contamination by incubation for 30 min at 37°C with DNase I, which was then inactivated by heating to 85°C for 15 min. Reverse transcription was conducted using strand-specific primers for NP (5' CAGGATGTGCTCACTGATGC 3') and GFP (5' CAGAAGAACGGCATCAAGCG 3') according to the manufacturer's protocol for the SuperScript III first-strand synthesis system for reverse transcriptase PCR (RT-PCR) (Invitrogen). The quantitative real-time PCR was carried out in a 20-μl reaction mixture with gene-specific primers for NP (5' CAGGATGTGCTCACTGATGC 3' and 5' TTCTCCGTCCATTCTCACCC 3') or for GFP (5' CAGAAGAACGGCATCAAGCG 3' and 5' TGGGTGCTCAGGTAGTGGTTG 3'), using SYBR green DNA dye (Invitrogen). The PCR conditions were 50°C for 2 min, 95°C for 2 min, and 45 cycles of 95°C for 15 s, 55°C for 30 s, and 72°C for 30 s. The plasmids pPolI-NP and PA 474-G-239 were used as standards for the NP and GFP genes, respectively.
RESULTS
We created reporter plasmids that expressed PA, PB1, or PB2 vRNAs identical to those found in virions except that the bulk of the native coding region was replaced by in-frame fusion of a cDNA encoding GFP. The resulting fusions were designed to preserve the entire 3' and 5' UTRs along with at least 230 bases of adjacent viral coding sequence at each end, as studies of other segments had suggested that up to 80 bases of retained coding sequence might be needed to support maximally efficient packaging (10, 11, 22). The reporters were named by the identity of the parental vRNA, followed by the number of bases of viral coding sequence at the 3' end, the letter G (indicating GFP), and the number of bases of viral coding sequence at the 5' end. The longest versions of each construct tested were PA 474-G-239, PB1 433-G-627, and PB2 455-G-230 (Fig. 1A).
We used these reporters in conjunction with a plasmid-based reverse genetic system that generates influenza virus particles in transfected cells (Fig. 1B). This system comprises eight plasmid vectors that are transcribed in both orientations by opposed promoters to yield the eight wild-type vRNAs and all viral proteins needed to produce complete virions (12). When transfected into 293T cells, this set of plasmids directs the assembly and release of infectious virions that accumulate to levels of 106 to 107 PFU/ml in the supernatant after 48 h. For the present study, we transfected 293T cells with all eight wild-type plasmids together with one of our vRNA reporter vectors. When examined 48 h later under fluorescence microscopy, nearly all of the transfected 293T cells exhibited strong GFP fluorescence, indicating that vRNAs expressed from the reporter vectors could be transcribed by the viral polymerase to produce functional mRNA encoding GFP (data not shown). To determine whether the reporter vRNAs could be packaged and transduced by influenza virus particles, we harvested supernatant from the 293T cells and used it to infect MDBK cells, which were collected 15 h later and stained by indirect immunofluorescence to detect the influenza virus nucleoprotein (NP) as a marker for virally infected cells. The immunostained cell populations were then examined by fluorescence microscopy (Fig. 2A). For each reporter, nearly all MDBK cells were strongly positive for NP expression (red), an indication of productive viral infection. Most infected cells also expressed GFP (green), indicating that the reporter vRNA had been successfully transduced and expressed. Those results were quantified more precisely using two-color flow cytometry (Fig. 2B), which enabled us to rapidly score NP and GFP expression in 20,000 cells from each infected population. Under the conditions of this assay, we defined the relative packaging efficiency of a given reporter as the percentage of NP-positive MDBK target cells that were also GFP-positive. We found that our three full-length reporter vRNAs (PA 474-G-239, PB1 433-G-627, and PB2 455-G-230) were each incorporated at relative packaging efficiencies of 70 to 95%.
In the assay we used, both a reporter vRNA and its wild-type counterpart are expressed in the 293T cells where viral assembly occurs. Preliminary studies indicated that omitting any one of those three wild-type vRNAs dramatically reduced the titer of hemagglutinating particles released into the supernatant and completely eliminated plaque-forming activity, even when a vector encoding the corresponding viral protein (PA, PB1, or PB2) was added to the transfection mixture (data not shown). It is expected that all three subunits of the viral polymerase must be produced in an infected target cell in order to amplify and detectably express the reporter vRNA. In principle, the wild-type segments encoding those subunits could be delivered to the target cell either by the virion that carried the reporter or by coinfection with a helper virion from the same supernatant. We observed that, for a given reporter, near-maximal frequency of target cells coexpressing GFP and NP was obtained by infecting at a ratio of at least 1.5 PFU per MDBK cell (Fig. 2C), and that ratio was consequently used in our standard assay. Using much lower ratios, we found that a small percentage (typically 0.3 to 1.0%) of plaques were composed of GFP-expressing cells (Fig. 2D), indicating that they had arisen from individual virions that stably carried the reporter vRNA along with the eight wild-type segments needed for multiple rounds of infection. This confirms that some influenza virus virions can stably carry at least nine vRNAs, but it also indicates that the great majority of particles that packaged a reporter vRNA and delivered it into a target cell were not fully competent for infectivity and might require a helper virion in order to express GFP. With those caveats, the flow cytometric assay provided a means to map the sequences in each reporter vRNA that are required for efficient packaging.
Initial deletions showed that coding sequences contributed to packaging of all three reporters. As shown in Fig. 3A, for example, the packaging efficiency of our longest PA-derived reporter, with 474 bases of 3' and 239 bases of 5' coding sequences, was 91% (left), but this declined to roughly 7% when all 3' coding sequences were removed (right). The elimination of 5' coding sequences had profoundly deleterious effects, as it reduced the packaging efficiencies of all three reporters to less than 1% (Fig. 3B). The effects of 3' coding deletions were less pronounced: deleting all but 99 bases at this end caused no reduction in packaging of any of the three reporters, and the PA and PB1 vectors continued to be packaged at 7 to 10% efficiency when all 3' coding bases were removed. Interestingly, a PB2 reporter with only 99 bases of 3' coding sequence was packaged with reproducibly higher efficiency than one with 455 bases (see also Fig. 4C), and substantial packaging (60% efficiency) persisted even when all 3' coding bases were removed.
Further truncations of these reporters helped delineate the minimal sequences that could support packaging (Fig. 4). Setting an arbitrary threshold of 65% efficiency, we observed efficient packaging with as little as 66 bases and 40 bases of coding sequence at the 3' and 5' ends, respectively, of the PA vector (Fig. 4A), although significant residual activity was observed with as little as 15 bases of 3' or 30 bases of 5' coding sequence; the shortest fully active PA vector we tested was PA 99-G-40. Correspondingly, 66 bases of 3' and 40 bases of 5' coding sequence from PB1 proved compatible with full activity, and the shortest PB1 vector with maximal packaging was PB1 99-G-40, but intermediate activity required as few as 33 coding bases at the 3' end of this reporter (Fig. 4B). Full packaging efficiency required as few as 80 coding bases from the 5' end of PB2 and was exhibited by the vector PB2 99-G-80 (Fig. 4C), although the lack of any requirement for 3' PB2 coding sequences in other contexts suggests that further 3' truncation of that vector would be tolerated.
We asked whether the observed differences in packaging merely reflected differences in steady-state expression levels among these reporters in 293T cells. To that end, we assayed expression of selected informative reporter vRNAs at 48 h after transfection, using real-time PCR with primers that amplified a portion of the GFP cassette. Similar quantification of NP-specific vRNA was performed in parallel for each sample, and levels of GFP-specific reporter RNA were expressed in relationship to this internal standard. As indicated in Fig. 5A, we found that vRNA concentrations varied by up to sevenfold among the different truncated forms of any given reporter. PA derivatives generally were highly expressed, and PB1 expression seemed to correlate with the length of its 5' end, perhaps reflecting differences in RNA synthesis or stability. These differences could not, however, account for the salient differences in packaging efficiency that we observed. For example, PB1 99-G-40 was expressed at a slightly lower level than PB1 99-G-0 but was packaged 70-fold more efficiently, and the 4-fold-lower expression of PB2 99-G-60 compared to other PB2 constructs could not account for their 20-fold difference in packaging. Confirmatory evidence came from the finding that GFP fluorescence levels in the 293T producer cells appeared to be comparable among these reporters (Fig. 5B). We conclude that the observed variations in packaging do not simply reflect differences in vRNA concentration at the time of virion assembly.
Our findings, together with earlier reports, indicated that critical packaging signals are localized to two separate regions at the 3' and 5' ends of a given vRNA. Apart from the highly conserved 12- and 13-base distal sequences of these ends, the remaining sequences of the UTRs and adjacent coding regions that harbor these signals are unique to each segment, which led us to ask whether the signals at each end of a vRNA are themselves functionally autonomous and interchangeable. We therefore designed a series of six hybrid GFP reporters, each of which combined the 3' and 5' packaging sequences from two different genome segments (Fig. 6A). In each case, these packaging sequences comprised a complete UTR along with sufficient adjacent coding sequences to support maximal packaging of the parental vRNA. We transfected each of these hybrid vectors into 293T cells along with the eight-plasmid assembly system and assayed the resulting levels of GFP and reporter vRNA expression and the efficiency of packaging into virions. Three additional reporters, carrying these same ends in their normal, homologous pairings, were tested in parallel. As illustrated in Fig. 6B, all nine reporters gave rise to GFP fluorescence, indicating that all were competent for transcription, replication, and expression, although the GFP levels from some of the hybrid vectors were noticeably reduced. Consistent with that finding, reporter vRNA concentrations in the 293T cells (each normalized to NP vRNA as an internal control) were generally higher for homologous than for hybrid vectors. Apart from the homologous PA reporter, however (which, as in our earlier comparisons, was the most highly expressed), the concentrations of these reporters varied no more than sevenfold. In striking contrast, whereas the three homologous reporters were each packaged with high efficiency (85 to 93%), all of the hybrid vRNAs were packaged very poorly, in most cases with efficiencies of less than 0.5%. Only one hybrid reporter (combining ends from PB2 and PB1) showed appreciable packaging, at 11% efficiency. These results demonstrate that the bipartite packaging signals of influenza virus vRNAs function only in their native combinations.
DISCUSSION
By systematically truncating reporter vRNAs derived from the influenza virus PA, PB1, and PB2 genome segments, we have mapped cis-acting sequences that are necessary and sufficient to enable each of these segments to be packaged efficiently into virions. Our study utilized a plasmid-based assembly system that encodes all of the viral proteins and RNAs needed to produce live, replication-competent influenza virus in transfected 293T cells. We designed a series of reporter vRNAs encoding GFP fusion proteins flanked by various lengths of viral RNA sequence, coexpressed these along with the nascent virions in 293T producer cells, and asked whether the resulting viral particles could transduce a given GFP reporter into MDBK target cells. This approach parallels that used by Kawaoka and colleagues to map packaging signals in the NA, HA, and NS segments (10, 11, 22), except that we employed flow cytometry to quantify GFP expression in the target cells. In addition, whereas those investigators routinely omitted the wild-type counterpart of each reporter under study from their assembly reactions, we found it essential to coexpress both the reporter and the corresponding wild-type vRNA. This presumably reflects the fact that the three vRNAs we studied each code for an essential subunit of the viral polymerase complex, whose activity is necessary to support both virus production by the 293T cells and GFP expression in target cells. Thus, to the extent that influenza virus vRNAs compete for packaging, our assay measured the efficiency with which a given reporter vRNA could compete against the full complement of wild-type vRNAs.
Despite those methodological differences, our findings fully accord with earlier reports in demonstrating that sequences within the distal coding regions contribute strongly to the packaging of influenza virus genome segments. With the addition of the three segments analyzed here, six of the eight viral segments have now been characterized, and in each case maximal packaging has been shown to require up to 80 bases of coding sequence adjacent to the UTRs at either end of a segment. Although coding sequences at both the 3' and 5' ends contribute to packaging of most segments, our results suggest that the PB2 segment may be unique in requiring such sequences only at its 5' end. Taking into account the various lengths of the UTRs (which range from 19 to 58 bases among the six vRNAs studied so far), available data thus indicate that all cis-acting signals needed for efficient packaging are localized within two separate regions, each 27 to 125 bases long, at the 3' and 5' termini of each segment.
Our results also confirm earlier reports that some influenza virus particles can package at least nine vRNAs (6, 21), giving rise to viral strains that stably transmit a reporter vRNA in addition to the eight segments needed for infectivity. Such particles appear to be relatively infrequent, however, accounting for fewer than 1% of plaque-forming virions in our study. Indeed, many of the GFP transduction events scored in our assay are likely to have involved particles that packaged reporter without the full complement of other genome segments; in such cases, expression of GFP required coinfection by helper virions, which we estimate were at least 2.5-fold more abundant than reporter-transducing particles in the viral supernatants we studied.
Additional studies will be needed to localize the packaging signals more precisely, as distinct from the signals responsible for stability, transcription, replication, and other critical properties of these viral RNAs. At present, the molecular characteristics of the packaging signals remain unknown, including the specific aspects of RNA sequence or secondary structure that confer packaging activity. A better understanding of those signals would be useful in engineering genetically modified influenza viruses, including influenza virus vectors that stably incorporate heterologous genes, and it would also shed light on the processes that give rise to interspecies recombination between influenza viruses, an important source of new pandemic strains. The finding that key packaging signals reside within portions of the UTRs and coding regions whose sequences are unique to a given vRNA offers a potential means by which the packaging mechanism could specifically discriminate one vRNA from another and so package its genome segments specifically rather than randomly. Such discrimination might be achieved though segment-specific protein-RNA or RNA-RNA interactions and might account for reports that certain defective-interfering mutant vRNAs selectively compete for packaging against their wild-type counterparts (5).
As one further step in characterizing the influenza virus packaging signals, we asked whether the two terminal packaging regions of a vRNA are functionally autonomous. We created six hybrid GFP vectors that combined 3' and 5' packaging regions from different vRNAs, and we asked whether these alternative pairings could support packaging into virions. When expressed in 293T cells along with an active viral polymerase, all six hybrid reporters gave rise to GFP fluorescence (although in some cases at lower levels than the parental reporters), indicating that they were competent to direct protein expression. Nevertheless, all six hybrids proved severely defective in packaging, with only one exhibiting even modest activity. Even the 5' terminus from the PB2 segment, which functioned independently of any 3' coding sequences in its native vRNA, was completely inert when paired with 3' sequences from the PA or PB1 segments. Thus, influenza virus packaging signals that are fully active in their normal combinations cannot be interchanged. This implies that the two separate packaging regions at the ends of an influenza virus vRNA form a uniquely interdependent pair and that they must undergo critical segment-specific physical and/or functional interactions to be recognized by the packaging machinery.
ACKNOWLEDGMENTS
We thank Y. Kawaoka and G. Hobom for the influenza virus plasmids and H. Ly for critical review of the manuscript.
This work was supported by NIH grant AI-40317.
REFERENCES
Bae, S. H., H. K. Cheong, J. H. Lee, C. Cheong, M. Kainosho, and B. S. Choi. 2001. Structural features of an influenza virus promoter and their implications for viral RNA synthesis. Proc. Natl. Acad. Sci. USA 98:10602-10607.
Chen, W., P. A. Calvo, D. Malide, J. Gibbs, U. Schubert, I. Bacik, S. Basta, R. O'Neill, J. Schickli, P. Palese, P. Henklein, J. R. Bennink, and J. W. Yewdell. 2001. A novel influenza A virus mitochondrial protein that induces cell death. Nat. Med. 7:1306-1312.
Crow, M., T. Deng, M. Addley, and G. G. Brownlee. 2004. Mutational analysis of the influenza virus cRNA promoter and identification of nucleotides critical for replication. J. Virol. 78:6263-6270.
Desselberger, U., V. R. Racaniello, J. J. Zazra, and P. Palese. 1980. The 3'- and 5'-terminal sequences of influenza A, B and C virus RNA segments are highly conserved and show partial inverted complementarity. Gene 8:315-328.
Duhaut, S. D., and J. W. McCauley. 1996. Defective RNAs inhibit the assembly of influenza virus genome segments in a segment-specific manner. Virology 216:326-337.
Enami, M., G. Sharma, C. Benham, and P. Palese. 1991. An influenza virus containing nine different RNA segments. Virology 185:291-298.
Flick, R., and G. Hobom. 1999. Interaction of influenza virus polymerase with viral RNA in the ‘corkscrew’ conformation. J. Gen. Virol. 80:2565-2572.
Flick, R., G. Neumann, E. Hoffmann, E. Neumeier, and G. Hobom. 1996. Promoter elements in the influenza vRNA terminal structure. RNA 2:1046-1057.
Fodor, E., D. C. Pritlove, and G. G. Brownlee. 1995. Characterization of the RNA-fork model of virion RNA in the initiation of transcription in influenza A virus. J. Virol. 69:4012-4019.
Fujii, K., Y. Fujii, T. Noda, Y. Muramoto, T. Watanabe, A. Takada, H. Goto, T. Horimoto, and Y. Kawaoka. 2005. Importance of both the coding and the segment-specific noncoding regions of the influenza A virus NS segment for its efficient incorporation into virions. J. Virol. 79:3766-3774.
Fujii, Y., H. Goto, T. Watanabe, T. Yoshida, and Y. Kawaoka. 2003. Selective incorporation of influenza virus RNA segments into virions. Proc. Natl. Acad. Sci. USA 100:2002-2007.
Hoffmann, E., G. Neumann, Y. Kawaoka, G. Hobom, and R. G. Webster. 2000. A DNA transfection system for generation of influenza A virus from eight plasmids. Proc. Natl. Acad. Sci. USA 97:6108-6113.
Hsu, M. T., J. D. Parvin, S. Gupta, M. Krystal, and P. Palese. 1987. Genomic RNAs of influenza viruses are held in a circular conformation in virions and in infected cells by a terminal panhandle. Proc. Natl. Acad. Sci. USA 84:8140-8144.
Lamb, R. A., and R. M. Krug. 2003. Orthomyxoviridae: the viruses and their replication, p. 1487-1531. In B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Fields virology, 4th ed., vol. 2. Lippincott-Raven Publishers, Philadelphia, Pa.
Li, X., and P. Palese. 1992. Mutational analysis of the promoter required for influenza virus virion RNA synthesis. J. Virol. 66:4331-4338.
Luytjes, W., M. Krystal, M. Enami, J. D. Pavin, and P. Palese. 1989. Amplification, expression, and packaging of foreign gene by influenza virus. Cell 59:1107-1113.
McGeoch, D., P. Fellner, and C. Newton. 1976. Influenza virus genome consists of eight distinct RNA species. Proc. Natl. Acad. Sci. USA 73:3045-3049.
Neumann, G., and G. Hobom. 1995. Mutational analysis of influenza virus promoter elements in vivo. J. Gen. Virol. 76:1709-1717.
Piccone, M. E., A. Fernandez-Sesma, and P. Palese. 1993. Mutational analysis of the influenza virus vRNA promoter. Virus Res. 28:99-112.
Robertson, J. S. 1979. 5' and 3' terminal nucleotide sequences of the RNA genome segments of influenza virus. Nucleic Acids Res. 6:3745-3757.
Scholtissek, C., W. Rohde, E. Harms, R. Rott, M. Orlich, and C. B. Boschek. 1978. A possible partial heterozygote of an influenza A virus. Virology 89:506-516.
Watanabe, T., S. Watanabe, T. Noda, Y. Fujii, and Y. Kawaoka. 2003. Exploitation of nucleic acid packaging signals to generate a novel influenza virus-based vector stably expressing two foreign genes. J. Virol. 77:10575-10583.(Yuying Liang, Ying Hong, )