TEXT

Top Abstract Text References

Transcription and replication of the RNA segments of influenza A virus are catalyzed by a virally encoded RNA-dependent RNA polymerase which contains three protein subunits, PB1, PB2, and PA. In addition, the viral nucleoprotein, or NP, which encapsidates the RNA templates, is also required for RNA synthesis. During replication, the viral RNA segments (vRNAs) are copied into cRNAs which, in turn, are templates for the synthesis of new copies of vRNA (5). By contrast, during transcription the vRNAs are copied into mRNAs. The mRNAs are not exact complementary copies of the vRNAs as they contain (i) a short 5' cap sequence derived from cellular mRNAs and (ii) a poly(A) tail at the 3' end (9). The cis-acting signals responsible for mRNA polyadenylation have been analyzed previously (3, 10, 11, 17), and it is believed that the addition of the poly(A) tail to the end of the mRNAs is achieved by stuttering of the viral polymerase at a stretch of five to seven uridine residues which are located near the 5' end of the vRNA template (11, 18). However, to date there is no direct evidence available indicating that such a stuttering mechanism exists in vivo during synthesis of influenza virus mRNAs. In this communication, we present evidence that the polymerase of influenza virus stutters at or near the poly(U) stretch during RNA synthesis in infected cells.

Variants of the NAdD virus containing insertions in the neuraminidase (NA) gene. Influenza A viruses have eight vRNA segments of negative polarity which contain at least one open reading frame (ORF) flanked by short noncoding regions. These untranslated regions contain the signals responsible for RNA replication, transcription, and polyadenylation and packaging of the RNA segments (12). Previously, we have generated a transfectant influenza virus, NAdD, in which the noncoding sequences at the 3' and 5' ends of the vRNA encoding the NA protein were modified by deleting 7 nucleotides (nt) upstream of the stop codon of the NA ORF and 4 nt downstream of the initiation codon of the NA ORF (Fig. 1) (22). The NAdD virus showed reduced replication (approximately 60-fold) of the NA-specific vRNA in infected cells compared with that in wild-type influenza A/WSN/33 virus (WSN) and grew to titers which were 2 logs lower than those of wild-type viruses (22). Five independent NAdD virus plaques were serially passaged 10 times in Madin-Darby bovine kidney (MDBK) cells. This experiment was conducted in order to investigate if compensatory mutations in the NA-specific vRNA would be selected, thus restoring wild-type levels of RNA replication. Liquid supernatants from the 10th passage were used to isolate vRNA and to clone and sequence the 3' and 5' noncoding regions of the NA vRNA. Sequences at the 3' ends of the NA genes of the viruses were amplified by reverse transcription-PCR (RT-PCR) with in vitro 3'-polyadenylated RNA (22) as template and the primers 5'-GCGCAAGCTTCTAGATTTTTTTTTTTTTT-3' and 5'-GCGCAAGCTTTATTGAGATTATATTTCC-3', the latter containing nt 115 to 98 of the NA gene of WSN. Direct sequencing of the PCR products derived from the five passaged virus plaques revealed the same sequence at the 3' noncoding region of the NA gene. This sequence was identical to that of the original NAdD transfectant virus.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Schematic representation of NA gene-specific influenza virus RNAs. The wild-type sequence of the noncoding regions of the NA gene of WSN (NA/WT) is shown on the top. The sequence of the noncoding regions of the NAdD RNA is shown in the middle. The dots indicate the deleted nucleotides of the NA gene of the transfectant virus. The rescued NA/NSm-NA RNA is shown at the bottom. Altered sequences are in italics. For convenience, the 5'- and 3'-terminal sequences are shown in a panhandle configuration (11). Fork and/or corkscrew configurations are also possible (2, 8). The start and stop codons of the NA genes are indicated in lowercase. The lines represent the coding sequence of the NA gene.

Isolation of viable NAdD variants. In order to obtain clonal populations of infectious viruses from the analyzed supernatants, viruses from the supernatants after 10 liquid passages of five independent NAdD plaques (S1-10, S2-10, S3-10, S4-10, and S5-10 samples) were plaque purified in MDBK cells overlaid with agar. Two plaques from each sample were used for liquid infections of MDBK cells, and the 5' ends of the NA genes present in the obtained samples were subjected to RT-PCR and sequenced as described above. The results are shown in Table 2. Of the 10 clones, only one sequence was identical to that of the 5' end of the original NAdD virus. The other sequences contained the following insertions at the poly(U) stretch of the 5' end of the NA vRNA: U (one sequence), UU (three sequences), UUCUUU (one sequence), UCUUCU (one sequence), and UUCUUCU (three sequences). The last three insertions seem to be originated by polymerase stuttering at a region of the template which includes not only the poly(U) stretch but also the adjacent C residue. These results suggest that viral polymerase stuttering may involve a template realignment mechanism similar to that proposed for RNA editing in paramyxoviruses (6).

Isolation of variants of NA/NSm transfectant virus. We have obtained further evidence for viral polymerase stuttering at the poly(U) stretch of the vRNA template by using a different transfectant influenza virus. The NA gene of the NA/NSm-NA virus is derived from an engineered WSN NA gene in which the the 3' noncoding sequence was modified (Fig. 1). In this engineered NA gene, positions 13, 14, and 15 at the 3' end were changed to those present in the wild-type NS gene. In addition, we replaced the noncoding nucleotide after position 15 with a novel sequence of 30 nt. One single virus plaque was obtained from the tissue culture medium after ribonucleoprotein transfection of the NA/NSm-NA gene (1). After three plaque-to-plaque passages of the NA/NSm-NA transfectant virus, one single plaque was used for infecting four 35-mm-diameter dishes of MDBK cells. The resulting virus preparation was used for further analysis. The identity of the NA/NSm-NA transfectant virus was confirmed by sequencing the 3' end of its NA gene. The sequence was identical to that of the transfected gene. However, RNA sequencing of the 5' end of the NA gene of the NA/NSm-NA virus revealed the incorporation of one more U residue in the poly(U) stretch (Fig. 1). The NA/NSm-NA virus has a phenotype similar to that of the NAdD virus in MDBK cells (22). Specifically, the levels of NA-specific vRNA are reduced approximately fourfold in viruses and in infected cells compared to those of wild-type virus (4a).

2

Variants of the NAdD virus generated during multicycle replication. We then analyzed whether different variants of the NA segment of NAdD virus could be detected after one single passage in MDBK cells with a low multiplicity of infection. MDBK cells were infected at a multiplicity of infection of 0.01 with NAdD virus or with wild-type WSN virus. Two days postinfection, cells were harvested and washed with phosphate-buffered saline, and total RNA was extracted with an RNAzol isolation kit (Tel-Test, Inc., Friendswood, Tex.). The RNA was subjected to RT-PCR as described above, in order to amplify the 5' ends of the NA-specific RNA segments. As a control for potential errors during the RT-PCR, we used RNA synthesized in vitro by T3 RNA polymerase from plasmid pT3NAdD, containing the cDNA of the NA segment of NAdD virus. PCR products were cloned into pUC19 with SpeI and HindIII restriction sites, and several clones were used for sequencing reactions with dideoxyribosylthymine termination mix and a primer annealing downstream from the polylinker of pUC19. This experiment allowed us to determine the presence of insertions or deletions in the 5' noncoding region of the NA segments. No changes were seen in 25 clones derived from T3 RNA and in 28 clones derived from WSN virus-infected cells. By contrast, the analysis of 20 clones derived from NAdD virus-infected cells revealed the insertion of two uridine residues at the poly(U) stretch in 14 clones, the insertion of seven uridine residues in one clone, and no changes in the remaining five clones.

Discussion. A poly(U) stretch near the 5' end of the vRNAs of influenza virus is required for the polyadenylation of the synthesized mRNAs (10). Interestingly, polyadenylation has not been observed before during cRNA synthesis. It is possible that the use of a capped primer for the initiation of RNA synthesis during transcription determines the subsequent polyadenylation of the newly synthesized RNA. Nevertheless, the influenza virus RNA polymerase must choose to ignore the polyadenylation signal during replication in order to synthesize a correct cRNA. Our sequencing data suggest that, in some instances, the viral polymerase makes a wrong choice and stutters at the poly(U) stretch of the vRNA during replication. It is possible that this happens more frequently than previously thought, since in most cases such polymerase errors will result in a nonfunctional cRNA template that will not be amplified. We propose that the insertions described in Tables 1 and 2 are due to the inherent ability of the influenza virus polymerase to stutter at the poly(U) stretch of the vRNAs, supporting a stuttering mechanism for the polyadenylation of the mRNAs of influenza virus.