Different De Novo Initiation Strategies Are Used b
http://www.100md.com
病菌学杂志 2006年第5期
Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom
ABSTRACT
Various mechanisms are used by single-stranded RNA viruses to initiate and control their replication via the synthesis of replicative intermediates. In general, the same virus-encoded polymerase is responsible for both genome and antigenome strand synthesis from two different, although related promoters. Here we aimed to elucidate the mechanism of initiation of replication by influenza virus RNA polymerase and establish whether initiation of cRNA and viral RNA (vRNA) differed. To do this, two in vitro replication assays, which generated transcripts that had 5' triphosphate end groups characteristic of authentic replication products, were developed. Surprisingly, mutagenesis screening suggested that the polymerase initiated pppApG synthesis internally on the model cRNA promoter, whereas it initiated pppApG synthesis terminally on the model vRNA promoter. The internally synthesized pppApG could subsequently be used as a primer to realign, by base pairing, to the terminal residues of both the model cRNA and vRNA promoters. In vivo evidence, based on the correction of a mutated or deleted residue 1 of a cRNA chloramphenicol acetyltransferase reporter construct, supported this internal initiation and realignment model. Thus, influenza virus RNA polymerase uses different initiation strategies on its cRNA and vRNA promoters. To our knowledge, this is novel and has not previously been described for any viral RNA-dependent RNA polymerase. Such a mechanism may have evolved to maintain genome integrity and to control the level of replicative intermediates in infected cells.
INTRODUCTION
Most RNA viruses encode their own RNA-dependent RNA polymerases in order to catalyze viral genome transcription and replication (3). The structures solved to date show that all viral RNA-dependent RNA polymerases form a similar basic right-hand-like structure with fingers, palm, and thumb subdomains (45). However, different viral RNA polymerases also have unique structural features which are important for efficient and accurate initiation of RNA-dependent RNA synthesis (25, 45). An amazing variety of initiation mechanisms for viral RNA-dependent RNA polymerization have been demonstrated. Any given RNA virus may use either one or more of these initiation mechanisms (25, 45).
In general, initiation mechanisms can be classified into two different categories, primer-dependent and primer-independent (de novo) initiation (25, 45). For primer-dependent initiation, the most often used primers are either (i) an oligonucleotide covalently linked to a protein, as used by picornaviruses (e.g., poliovirus) (38) or (ii) a capped primer cleaved from the 5' end of the host mRNA by a cap-snatching mechanism, which is used for viral genome transcription by some segmented negative-stranded RNA viruses such as influenza virus (4) and Dugbe nairovirus (24). In general, viral RNAs without cap-snatched primers or terminal proteins are initiated by a de novo initiation mechanism, which is common for many positive, negative, double-stranded, and ambisense RNA viruses (25). This mechanism is, clearly, a critical step to ensure efficient production of new viral RNA and to maintain the integrity of the viral genome (25, 45).
In most cases, de novo initiation starts at residue 1 of the template and is immediately followed by elongation. However, in some cases, e.g., Tacaribe arenavirus, it has been observed that short RNA oligonucleotides may be produced by successive rounds of abortive de novo transcription directed by either terminal nucleotides or internal nucleotides of the template. These oligonucleotides are subsequently used as primers for further elongation, as in the so-called prime-and-realign mechanism (7, 17-19, 26). A similar mechanism has also been proposed to explain 3' end repair in turnip crinkle virus (37).
Influenza A virus RNA polymerase is a heterotrimeric complex composed of three subunits, PB1, PB2, and PA. No crystal structure is available for this complex, although low-resolution electron microscopic images have been described (1). The PB1 subunit is the key component of the complex, containing conserved motifs of RNA-dependent RNA polymerases (13, 30). The PA and PB2 subunits are also required for both viral RNA transcription and replication (13, 30). The polymerase catalyzes both primer-dependent RNA transcription (vRNA to mRNA synthesis) and primer-independent replication (vRNA to cRNA and cRNA to vRNA synthesis) in the nucleus of infected cells (13, 30, 46). The mechanism of transcription (vRNA to mRNA synthesis) has been well studied since both in vivo and in vitro assays are available (10, 14, 41). In contrast, the mechanism of replication remains unclear. The reason for this is that it is difficult to distinguish replication directed by the vRNA versus the cRNA promoters in vivo because of the interdependence of the two promoters in the replication cycle. Existing in vitro primer-independent replication assays (22, 32) are inefficient and may, therefore, not be authentic.
The current model for the secondary structure of both the vRNA and cRNA promoters is the so-called "corkscrew" model (2, 8, 12). This model proposes a partially duplex region formed by the conserved noncoding sequences at both the 5' and 3' ends of each gene segment and two hairpin-loop structures, each with a stem of 2 bp and a tetraloop formed by residues close to the 3' and 5' termini. The secondary structure of the cRNA promoter proposed by Crow et al. (8) was based on the results of both a mutagenic analysis of the cRNA promoter in an in vivo minireplicon system, and in vitro promoter binding studies. However, there remained unexplained differences between the in vivo and in vitro results. In particular, point mutations at positions 4, 5, and 7 of the 3' strand of the cRNA promoter inhibited vRNA, cRNA, and mRNA synthesis in vivo, whereas in vitro approaches showed that RNA polymerase can bind all three of these point mutants in a UV cross-linking assay (8) and transcribe them in an ApG-primed transcription assay (40). According to the "corkscrew" structure of the cRNA promoter, the nucleotides at positions 4, 5, and 7 of the 3' strand of the promoter are located in a hairpin loop region, and point mutations at these positions would not be expected to disrupt the secondary structure and functions of the promoter in vivo. To resolve differences between the in vivo and in vitro data, an authentic in vitro primer-independent replication assay is needed.
Here, we developed two in vitro primer-independent RNA synthesis assays to study primer-independent RNA replication on model promoters. By mutational analyses of the two promoters, we found that influenza virus RNA polymerase uses internal initiation and realignment on its cRNA promoter but terminal initiation and elongation on its vRNA promoter during replication. These results shed new light on previous in vivo data on the cRNA promoter (8).
MATERIALS AND METHODS
Plasmids. The pcDNA-PB1, pcDNA-PB2, pcDNA-PA, and pcDNA-NP protein expression plasmids for the three polymerase subunits and NP of influenza A/WSN/33 virus have been described (14). The pPOLI-cCAT-RT (8), pcDNA-PB2tap (9), and active-site mutant (ASM) pcDNA-PB1-D445A/D446A (46) plasmids have been previously described. The pcDNA-PB1, pcDNA-PB2tap, and pcDNA-PA protein expression plasmids for the three polymerase subunits of influenza A/Turkey/England/5092/91 virus polymerase were provided by Ruth Harvey.
RNA oligonucleotides. RNA oligonucleotides (Dharmacon) for the vRNA promoter were 5' AGUAGAAACAAGGCC 3' and 5' GGCCUGCUUUUGCU 3' and for the cRNA promoter were 5' AGCAAAAGCAGGCC 3' and 5' GGCCUUGUUUCUACU 3'. They contained the conserved sequence (underlined) plus two additional residues. Mutant oligonucleotides based on the above vRNA and cRNA promoters were also from Dharmacon.
Preparation of partially purified recombinant influenza virus RNA polymerase. 293T cells were transfected with plasmids (see above) using Lipofectamine 2000. After 48 h the cells were lysed and TAP-tagged influenza virus polymerase (3P) complex with or without NP was purified by affinity purification with immunoglobulin G-Sepharose, and released by tobacco etch virus protease as before (9). Partially purified proteins in tobacco etch virus protease cleavage buffer (0.15 M NaCl, 0.01 M HEPES, pH 8.0, 0.1% NP-40, 1 mM dithiothreitol, 10% glycerol) were stored at –20°C in 35% glycerol. Partially purified His-tagged influenza virus polymerase, used in the UV cross-linking competition assay, was prepared as described previously (8).
In vitro cRNA replication assay. We performed 3-μl reactions with 1.5 μl partially purified 3P/NP and 0.7 μM synthetic model cRNA promoter (an equimolar mixture of the 5' and 3' strands) in the presence of 5 mM MgCl2, 1 mM dithiothreitol, 1 mM ATP, 0.5 mM GTP, 0.5 mM UTP, 0.1 μM [-32P]CTP (>3,000 Ci/mmol, Amersham), and 1 U/μl RNasin (Promega). Where indicated, 0.5 mM ApG (Sigma) was added to the reaction. Reactions were incubated at 30°C overnight for convenience, although significant activity was obtained within 1 to 3 h. An equal volume of formamide/bromophenol blue/EDTA was added and the mixture was heated at 95°C for 5 min. An aliquot, usually 4 μl, was analyzed by 25% polyacrylamide gel electrophoresis (PAGE) in 1x Tris-borate-EDTA and 6 M urea. Autoradiography was performed with Kodak Biomax MS films and Trans screen HE intensifying screens at –80°C.
UV cross-linking assay. UV cross-linking assays were performed as described previously (8).
Dinucleotide replication initiation assay and extension to trinucleotide assay. We performed 3-μl incubations with (normally) 1.5 μl polymerase, 0.02 μM [-32P]GTP (>3,000 Ci/mmol, Amersham) in the presence of 5 mM MgCl2, 1 U/μl RNasin (Promega). and 1 mM dithiothreitol with the addition of 1 mM ATP and 0.7 μM model vRNA (an equimolar mixture of the 5' and 3' strands) or 0.7 μM model cRNA (an equimolar mixture of the 5' and 3' strands) promoters.
To assay for extension to trinucleotides, either 1 mM UTP or 1 mM CTP, as specified in the individual experiments, were added. In the "transplant" experiment (see Fig. 3), the 5' strand was added first and incubated for 10 min at 30°C prior to addition of the 3' strand to avoid self-annealing between the two strands. In the "template-switching" experiment (Fig. 6), a preincubation of either the vRNA or the cRNA promoter with polymerase was performed at 30°C for 10 min before mixing. Incubations were at 30°C for various times (usually 1 to 3 h) up to 16 h. Formamide/bromophenol blue/EDTA (10 μl) was added and the mixture was heated at 95°C for 2 min. Analysis, usually of a 5 μl aliquot, was by 25% PAGE in 1x Tris-borate-EDTA and 6 M urea, followed by autoradiography.
Gel elution, enzymatic digestion, and analysis of oligoribonucleotides. The excised band of gel containing oligoribonucleotides <5 residues in length was rinsed for 1 min in 1 ml deionized water and then crushed in 0.5 ml deionized water and shaken overnight at 4°C. After centrifugation at 10,000 x g to remove gel particles, the supernatant was freeze-dried. Finally, the lyophilized sample was dissolved in a minimum volume (5 to 10 μl) of deionized water. T2 RNase (Sigma), which degrades RNA to give 3'-phosphorylated mononucleotides, was used essentially as recommended by the manufacturers. After digestion, 0.5 μl 10 mM ATP was normally added to the sample (usually 3 to 5 μl) prior to loading onto a polyethyleneimine-cellulose (Merck PEI-cellulose F) thin layer, which was developed for 4 cm above the origin in 0.5 M ammonium sulfate, and then further developed in 0.7 M ammonium sulfate.
The mobility of pppA32p on PEI-cellulose thin-layer chromatography (TLC) was deduced because the product (pA32p) of its further digestion with tobacco acid pyrophosphatase comigrated (RF = 0.50) with unlabeled, marker pAp (Sigma) (results not shown). The mobility of pp32pGp on PEI-cellulose TLC was deduced from T2 RNase digestion of pp32pGpU, which had been characterized as possessing a pppG end group because it gave rise to a product, after P1 nuclease digestion, that comigrated with GTP on PEI-cellulose TLC (results not shown). Oligonucleotides >5 residues were gel-eluted in 0.5 M ammonium acetate and ethanol precipitated by standard methods. Calf intestinal phosphatase (Roche), tobacco acid pyrophosphatase (Epicenter) and polynucleotide kinase (Roche) treatments were carried out for 1 h at 37°C in the supplied buffer. Digested products were analyzed by 25% PAGE as above.
Primer extension assay. Primer extension assays were performed as described previously (8).
Terminal sequencing of viral RNA. Sequencing of the 5' and 3' termini of viral RNA transcripts was carried out essentially as described previously (43). Briefly, total cellular RNA from transfected 293T cells was extracted with Trizol (Invitrogen) and treated with TAP (Epicenter) prior to being reextracted with Trizol. The RNA was then circularized with T4 RNA ligase (Epicenter). cDNA copies of the ligated 3' and 5' termini were synthesized by reverse transcription (8) with separate vRNA- or cRNA-specific primers. The cDNA was amplified by PCR using Taq polymerase (Promega) and both vRNA- and cRNA-specific primers, cloned using the TOPO TA cloning kit (Invitrogen), and sequenced.
RESULTS
Mutational analysis of the 3' end of the influenza virus cRNA promoter using a primer-independent in vitro replication assay. An ApG-primed replication assay has been used previously to study the sequence requirements of the 3' end of the influenza virus cRNA promoter for RNA replication with nonrecombinant RNA polymerase (40). However, the exact nature of this assay, whether it mimics replication or transcription, was not clear. Here we developed an in vitro primer-independent replication assay based on the incorporation of [-32P]CTP (0.1 μM) into a 15-nucleotide transcript of a model cRNA promoter (Fig. 2A) in the presence of 0.5 mM GTP and UTP and 1 mM ATP. Partially purified recombinant influenza virus RNA polymerase isolated from transiently transfected 293T cells was used as a source of enzyme (5). Figure 1A (lane 1, upper panel) shows that a 15-nucleotide full-length transcript was synthesized from the model wild-type cRNA promoter.
The essential feature of an authentic primer-independent replication product is the presence of a triphosphate moiety at its 5' end. In order to confirm that unprimed transcripts had a 5' triphosphate end group, we compared the mobilities of the 15-nucleotide unprimed transcript with the ApG-primed transcript, variously treated with calf intestinal phosphatase, tobacco acid pyrophosphatase, or polynucleotide kinase on a 25% PAGE-urea gel (Fig. 1B). The 15-nucleotide unprimed transcripts (5' pppA) migrated faster than the ApG-primed transcripts (5'OHA) (Fig. 1B, compare lane 4 with 1 and 2). Unprimed transcripts treated with calf intestinal phosphatase (5' pppA5' OHA), comigrated with the ApG-primed transcripts (5' OHA) (Fig. 1B, compare lanes 3 with 1). When the unprimed transcripts were treated with tobacco acid pyrophosphatase (5' pppA5' pA), they comigrated with the polynucleotide kinase-treated ApG-primed transcripts (5' OHA5' pA) (Fig. 1B, compare lanes 5 and 6). In addition, a faint band migrated slightly faster than the unprimed transcripts in lane 4. This faint band is likely to be a 5' diphosphate-ended transcript since it had an intermediate mobility between the 5' p- and 5' ppp-ended transcripts. Its origin is uncertain, but it may be due to a contaminant nucleotidyl pyrophosphatase in our enzyme preparation. Therefore, a majority of transcripts had a 5' triphosphate end group.
Having validated the authenticity of this primer-independent transcription assay, we compared the effects of point mutations at positions 1 to 9 from the 3' end of the 3' strand of the model cRNA promoter in this primer-independent replication assay (Fig. 1A, upper panel) with an ApG-primed transcription assay (Fig. 1A, lower panel) under the same conditions. In the unprimed replication assay, point mutants of the promoter at positions 1, 6, 7, and 8 showed transcription activities similar to the wild type (Fig. 1A, upper panel, compare lanes 3 and 8 to 10 with lane 1). Lower yields were obtained with the mutant at position 3, and only residual levels were obtained for the mutant at position 9. Interestingly, however, no products were detected with the mutants at positions 2 and 4 to 5. In the ApG-primed transcription assay, however, significant activity was observed with the mutants at positions 3 to 4 and 6 to 8 (Fig. 1A, lower panel, compare lanes 5 to 6 and 8 to 10 with lane 1), lower activities were observed with the mutants at positions 1 and 5, residual activity was observed with the mutant at position 9, but no activity was observed with the mutant at position 2. The most obvious difference between the results of the two assays is at positions 4 and 5. Mutations here interfered with unprimed transcription completely, but not with the ApG-primed transcription (Fig. 1A, compare lanes 6 and 7 in the upper and lower panels).
The polymerase binding properties of these two mutants had been previously studied in a UV cross-linking competition assay where it was shown that both mutants compete with the wild-type model cRNA promoter (8). A more detailed UV cross-linking competition analysis of these two mutants is shown here (Fig. 1C), confirming that the 4UA and 5CA mutants were not significantly different from the wild-type model cRNA promoter in competing with cross-linking of a radiolabeled wild-type promoter to the subunits (PB1, PB2, and PA) of the polymerase (Fig. 1C, compare lanes 5 to 8 and 9 to 12 with lanes 1 to 4). Thus, we excluded the possibility that the absence of any replication products observed for the mutants at positions 4 and 5 of the 3' strand of the model cRNA promoter in the primer-independent replication assay (Fig. 1A, upper panel) was due to their weak binding to the polymerase.
pppApG is copied from nucleotides at positions 4 and 5 of the 3' end of the cRNA promoter in contrast with positions 1 and 2 of the vRNA promoter. Influenza virus ribonucleoprotein isolated from virions is known to synthesize the dinucleotide pppApG in vitro by copying the terminal 3' OHU-C 5' residues of the vRNA template under conditions where only ATP and GTP are present (28). Here we have adapted this dinucleotide synthesis assay to investigate initial events in replication (i.e., synthesis of pppApG) in the absence of added primer with recombinant influenza virus RNA polymerase and model vRNA and cRNA promoters (Fig. 2A and B).
Wild-type and mutant promoters were used to test whether the observed lack of replication (Fig. 1A) was caused by events early in replication. We found that mutations at positions 4 and 5 of the 3' strand of the model cRNA promoter (Fig. 2C, lanes 6 to 7) inhibited pppApG synthesis dramatically, whereas mutants at positions 1 and 2 (Fig. 2C, lanes 3 and 4) showed activity similar to that of the wild type. By contrast, mutants at positions 1 and 2 of the model vRNA promoter (Fig. 2D, lanes 3 to 4) were inactive in pppApG synthesis. These results indicated that the vRNA and cRNA promoters have distinctive sites of pppApG synthesis i.e., the 3' OHU-C 5' at positions 1 to 2 of the 3' end of the vRNA promoter, but the 3' U-C 5' at positions 4 to 5, from the 3' end of the cRNA promoter.
In addition, mutations elsewhere within the promoters can influence pppApG synthesis activity (Fig. 2C and D). In the case of the cRNA promoter, mutants at positions 6, 7 and 8 (Fig. 2C, lanes 8, 9, and 10) retained significant pppApG synthesis activity, while the mutant at position 9 showed only residual activity (Fig. 2C, lane 11). The dinucleotide synthesized from the position 3 (AU) mutant (Fig. 2C, lane 5) showed a slightly slower mobility and was found to consist of both pppApG and pppGpA by T2 RNase digestion (results not shown). In the case of the vRNA promoter, mutants at positions 4 to 6 and 8 showed weak activity (Fig. 2D, lanes 6 to 8 and 10), although mutants at positions 7 and 9 were inactive (Fig. 2D, lanes 9 and 11 to 12). Interestingly, the mutant at position 3 (GA) was more active than the wild type (Fig. 2D, compare lanes 2 with 5). In summary, the pattern of pppApG synthesis observed in the mutagenesis screening (Fig. 2C and D) differs significantly between the two promoters. This further suggests that the RNA polymerase initiates pppApG synthesis on the vRNA and cRNA promoters in different ways.
We next asked if a model vRNA promoter, mutated at position 5 to a C residue, thereby generating a 3' U-C 5' sequence at positions 4 and 5, could allow pppApG synthesis at positions 4 and 5. To do this we had to introduce an additional mutation at position 1 (UA) to block any possibility of terminal initiation. Figure 2E shows that the double mutant (5UC, 1UA) (lane 4) was inactive in pppApG synthesis, while the 5UC mutant (lane 3) still synthesized pppApG as we had observed with 5UA mutant in Fig. 2D (lane 7). This suggested that simply mutating position 5 UC in the vRNA promoter does not permit the synthesis of pppApG internally. Overall, these results indicate that intrinsic differences in the vRNA and cRNA promoters may control the site of pppApG synthesis.
Absence of a hinge in the 5' strand and presence of a hinge and identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter are the major determinants of internal pppApG synthesis. The cRNA promoter is complementary to the vRNA promoter, and a very similar corkscrew secondary structure model has been proposed for both promoters (2, 8, 12) (Fig. 2A and B). Nevertheless, there are obvious differences between the two promoters. First, the lengths of the 5' and 3' arms differ due to the presence or absence of a hinge, an extra nucleotide located between the base-paired duplex region and the hairpin-loop structure (Fig. 2A and B). In the case of the 5' arm of the promoters, a hinge A residue is present in the vRNA promoter but is absent in the cRNA promoter. In the case of the 3' arm of the promoters, a hinge U residue is present in the cRNA promoter but is absent in the vRNA promoter.
Second, the identities of the 3 to 8 base pair and the nucleotide at position 5 of the hairpin loop structures of both 5' and 3' arms vary. Specifically, a C-G base pair at positions 3' and 8' and an A at position 5' in the 5' hairpin loop of cRNA are replaced by a U-A base pair and a G at the corresponding positions of vRNA. In the 3' hairpin loop, an A-U base pair at positions 3 and 8 and a C at position 5 of cRNA are replaced by G-C and U, respectively, in vRNA (Fig. 2A and B).
In order to test which of these sequences or presumed secondary structural elements in cRNA are responsible for internal pppApG synthesis, a series of model 5'-strand and 3'-strand promoter mutants that transplant the cRNA-like elements one by one into the vRNA promoter were designed (Fig. 3A and C). In addition, a mutation was introduced at position 1 (UA) of the 3' strand to exclude any possibility of terminal pppApG synthesis (Fig. 2D, lane 3).
The 5'-arm vRNA promoter mutants transplanted with different 5'-strand cRNA-like elements were first tested in combination with the 3' arm of the cRNA promoter with the additional 1UA mutation (see above) to evaluate which elements of the 5' arm were needed for internal initiation on the 3' arm of cRNA. Figure 3B shows that no internal pppApG synthesis occurred when the nucleotide 5'A of cRNA replaced 5'G of vRNA (lane 3), and only residual activity was obtained when the 3'C-8'G base pair of cRNA replaced the 3'U-8'A base pair of vRNA (lane 4). However, significant activity was detected when 10'A was deleted from vRNA (lane 5). No further significant increase in activity was observed if this 10'A deletion was combined with either the 5'A element (lane 6) or the 3'C-8'G element (lane 7). This result suggests that the absence of a 10A hinge in the 5' arm of the cRNA promoter is the key structural element controlling internal pppApG synthesis.
Second, we tested which 3'-arm cRNA elements were required for internal pppApG synthesis by again transplanting cRNA elements into the vRNA promoter. In this case, the residue at position 5 (a U residue) of the vRNA promoter had to be mutated to a C residue to allow pppApG transcription from positions 4 and 5. The 1(UA) mutation was also used throughout to block pppApG synthesis from nucleotides 1 and 2. The cRNA elements 3A and 8U were treated separately since point mutants at these positions showed significant activity (Fig. 2C). These elements and the 10U hinge were then transplanted individually or in different combinations into the 3' arm of the modified vRNA backbone and tested in combination with the 5' arm of the cRNA promoter for pppApG synthesis (Fig. 3C). No significant activity was obtained with the mutants bearing either the 3A or the 8U, or both the 3A and 8U cRNA elements (Fig. 3D, lanes 6, 7, and 8), whereas residual activity was observed when the 10U hinge was inserted (lane 9). However, when the 10U hinge was introduced in combination with the 3A element, wild-type activity was observed (Fig. 3D, compare lanes 10 and 2). In contrast, only partial activity was obtained when the 10U hinge was introduced in combination with the 8U element (Fig. 3D, lane 11). As expected, wild-type activity was obtained when all three elements (3A, 8U, and 10U) were present (Fig. 3D, lane 12). These results suggest that the combination of the 3A element and the 10U hinge in the 3' end of the cRNA promoter is required for full internal pppApG synthesis activity. Overall, we concluded that the absence of the hinge in the 5' strand and the presence of the hinge and the identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter are the major structural elements that determine internal pppApG synthesis.
Internally synthesized pppApG anneals to position 1 and 2 of the 3' strand of the cRNA promoter for extension into trinucleotides in a template-directed manner. Theoretically, to maintain the integrity of the viral genome, internally synthesized pppApG from nucleotides 4 and 5 of the 3' strand of the cRNA promoter would have to realign to nucleotides 1 and 2 for elongation. To test this model, we performed the pppApG synthesis reaction in the presence of UTP in addition to ATP and radiolabeled GTP. Under these conditions, we observed pppApG and pppGpU synthesis (see Discussion) (Fig. 4A and B, band A) as well as elongation to a trinucleotide, although elongation was inefficient (Fig. 4A, lane 2). This trinucleotide (Fig. 4B, band E) was characterized as pppApGpX by gel elution and T2 RNase digestion, followed by PEI-cellulose TLC. The identity of X was deduced as a U residue because no trinucleotide product was observed when the UTP was omitted or CTP was added instead (results not shown). However, we failed to observe elongation to either a 5-nucleotide product, pppAGAAA, derived from residues 4 to 8 of the cRNA promoter (Fig. 2A) or an 8-nucleotide product, pppAGUAGAAA, derived from residues 1 to 8 (results not shown). This failure, and the low yield of the trinucleotide pppApGpU, could be due to the low concentration of GTP used in this reaction.
To further test the properties of the cRNA promoter required for realignment, we studied the 1UA and 2CA mutations, together with the 4UA and 5CA mutants as controls. The previous result that mutations at positions 4 and 5 inhibited pppApG synthesis was confirmed (Fig. 4A, lanes 5 and 6), whereas pppGpU (Fig. 4B, band D) was the major product synthesized from the 4UA mutant. Lanes 3 and 4 showed that the 1UA mutation but not the 2CA mutation allows trinucleotide pppApGpU synthesis. The presence of pppApGpU, derived from the wild type and the mutant at position 1 suggested that pppApG, internally synthesized from residues 4 and 5 of the model cRNA promoter, was able to realign with wild-type residues 1 and 2 of the promoter and extend by one residue, albeit inefficiently.
The 5' pppA end group of band F derived from the 1UA mutant (Fig. 4A, band F), as confirmed by T2 RNase digestion and PEI-cellulose TLC (Fig. 4B), shows that the trinucleotide transcript had effectively corrected the 1UA mutation of the cRNA template by realignment. This result further suggests that base-pairing between the first nucleotide of pppApG and the terminal residue of the 3' end of the cRNA promoter template is not critical for realignment. On the other hand, no trinucleotide was observed from the mutant at position 2 (Fig. 4A, lane 4), suggesting that base-pairing between the second nucleotide of pppApG and the C residue at position 2 of the 3' end of the model cRNA promoter template is essential and sufficient for extension in vitro. Based on these results, we conclude that the influenza virus cRNA promoter might use an internal initiation and realignment mechanism to initiate viral RNA synthesis.
Mutated or deleted nucleotide at position 1 of the 3' end of the cRNA promoter is corrected to the wild-type nucleotide in a minireplicon system. As was shown in Fig. 4A (lane 3, band F) above, a wild-type trinucleotide transcript had been generated in vitro from a cRNA template mutated at position 1. In vivo, it has been shown that a mutation at position 1 of the 3' end of the cRNA promoter was compatible with the synthesis of significant levels of mRNA, vRNA and cRNA from a CAT reporter (8). Thus, it was of interest to test whether the 1UA mutant previously studied in vivo was corrected to the wild-type sequence.
Two additional mutants with a 1UC or a deleted residue 1 (1U) were generated and, together with the previously isolated 1UA mutant, were tested in a primer extension assay (8). This assay measures the steady-state levels of vRNA, mRNA and cRNA in a minireplicon system in which 293T cells were transiently transfected with 4 protein expression plasmids encoding the three subunits of influenza virus RNA polymerase and the nucleoprotein, together with a pPOLIcCAT plasmid which encodes a CAT reporter gene flanked by the influenza virus cRNA promoter (8).
Figure 5A shows that templates with a mutated or deleted residue 1 were replicated (vRNA and cRNA) and transcribed (mRNA) to wild-type levels (Fig. 5A, lanes 1, 3, 5, and 7). As controls, the background levels of unamplified polymerase I (POLI) transcripts of the wild-type and the three mutant templates (Fig. 5A, lanes 2, 4, 6, and 8) were obtained by transfection of an active-site polymerase mutant (replacing wild-type PB1 with mutant PB1-D445A/D446A, abbreviated 3P-ASM) (46). Sequencing of the RNA (see materials and methods) showed that all three mutants (i.e., the two mutated and the one deleted nucleotide at position 1) had reverted to the wild-type sequence (Fig. 5B).
pppApG derived from the cRNA promoter can be released and realigned onto a vRNA promoter in vitro. According to our model, pppApG initiates internally and is extended by realignment onto terminal residues of the cRNA template (cis-realignment). It was of interest to establish whether pppApG synthesized internally on a cRNA template could dissociate and realign to nucleotides 1 and 2 of a vRNA template for extension (trans-realignment). To test this, we designed an in vitro template-switching experiment in which a cRNA promoter mutant, which allows pppApG synthesis internally but does not allow extension to a trinucleotide, is mixed with a vRNA promoter mutant, which does not allow pppApG synthesis terminally but would allow extension to a trinucleotide if pppApG were available. Thus, if the pppApG dissociated from the cRNA promoter mutant, it would be available for the vRNA promoter mutant to use for extension into trinucleotide. If, however, the pppApG remained associated with the cRNA promoter mutant, no extension to a trinucleotide would occur. From previous mutational analyses of the cRNA and vRNA promoters (Fig. 2C and D), the 3'-strand cRNA promoter mutant 2 (CA) and the 3'-strand vRNA promoter mutant 1 (UA) fit these criteria. Also, by taking advantage of the different mobilities of pppApGpC (Fig. 6A, lane 4) directed by a model vRNA template and pppApGpU (Fig. 6A, lane 1) directed by a model cRNA template, we could distinguish whether the template of the synthesized trinucleotides was the vRNA or the cRNA promoter.
As controls, we used model wild-type cRNA and vRNA in the presence of either UTP or CTP to synthesize trinucleotides as markers (Fig. 6A, lanes 1 to 4). As expected, pppApGpU (band D) was only synthesized from the cRNA template and only in the presence of UTP (lane 1). Similarly, pppApGpC (Fig. 6A, band E) was only synthesized from the vRNA template and only in presence of CTP (Fig. 6A, lane 4). The cRNA and vRNA promoter mutants mentioned above were also included as further controls to confirm that the 2CA cRNA mutant can generate pppApG internally but cannot be extended into a trinucleotide in the presence of either UTP or CTP (Fig. 6A, lanes 5 and 6), and the 1UA vRNA promoter mutant cannot generate pppApG or trinucleotides (Fig. 6A, lanes 7 and 8).
The test for template switching was carried out by mixing the model cRNA promoter mutant (2CA) and the model vRNA promoter mutant (1UA). The result showed that the trinucleotide pppApGpC (Fig. 6A, band F) was only observed when CTP, but not UTP, was provided in the reaction (Fig. 6A, compare lanes 9 and 10). This suggested that pppApG synthesized by the cRNA promoter mutant (2CA) had realigned to the vRNA promoter (1UA) and extended into the trinucleotide pppApGpC. We were also able to confirm that template switching occurred when the wild-type cRNA promoter was mixed with the mutant vRNA promoter (1UA) (Fig. 6A, lanes 11 to 12). In this case, both pppApGpU (band G), which was synthesized from the wild-type cRNA promoter template, and pppApGpC, (band H), which was synthesized by template switching of pppApG from the wild-type cRNA promoter to the vRNA promoter mutant, were observed. The pppA end group of these trinucleotides (bands D to H) was confirmed by T2 RNase digestion and comparison with T2 RNase-digested dinucleotides on PEI-cellulose TLC (Fig. 6B). Overall, we conclude that pppApG synthesized internally from the cRNA template can be realigned onto a cRNA or a vRNA template for elongation.
DISCUSSION
De novo RNA synthesis is a common strategy used by viral RNA-dependent RNA polymerases for viral RNA replication to maintain the integrity of the viral genome (25, 45). However, the mechanisms for primer-independent initiation vary among different RNA viruses (25, 45). The aim of this work was to study the mechanism by which influenza virus polymerase initiates replication on its vRNA and cRNA promoters.
Here, we have developed two in vitro primer-independent RNA synthesis assays with recombinant influenza virus RNA polymerase and model promoters. The first assays influenza virus cRNA promoter-specific replication by primer-independent transcription of a 15-nucleotide template. The transcripts were shown to possess 5'-triphosphate end groups, demonstrating the authenticity of the assay. By mutagenesis screening, we found that point mutations at positions 2, 4, 5, and 9 from the 3' end of the cRNA promoter largely inhibited replication in this assay. In contrast, ApG-primed transcription was only inhibited by mutations at positions 2 and 9, with partial inhibition at position 5 (Fig. 1A). This suggested that pppApG might be synthesized from residues 4 and 5.
We therefore investigated the site of de novo pppApG synthesis by a second assay, an in vitro replication initiation assay. We found that the nucleotides at positions 4 and 5 from the 3' end of the cRNA promoter were important for directing pppApG synthesis in contrast with the nucleotides at positions 1 and 2 at the 3' end of the vRNA promoter (Fig. 2). As residues 4 and 5 from the 3' end of the cRNA promoter are 3' U-C 5', we hypothesized that pppApG may be templated internally from these residues and then realigned to the terminal 3' OHU-C 5' for elongation. Accordingly, we showed that pppApGpU was synthesized from the wild-type cRNA promoter and the mutant at position 1, but not from the mutant at position 2 (Fig. 4), suggesting that internally synthesized pppApG can realign onto positions 1 and 2 for extension into trinucleotide pppApGpU in a template-directed manner (Fig. 4). This further suggests that base-pairing of pppApG with the C residue at position 2 alone was sufficient for realignment of pppApG and elongation, and that a mutation at position 1 of the template would be corrected to the wild-type base in the transcript. This realignment would explain why mutations at positions 2, 4, and 5 in the model cRNA failed to synthesize a 15 nucleotide unprimed transcript (Fig. 1).
These data are also consistent with the previous unexplained in vivo observations that nucleotides at positions 4 and 5 and the identity of the 2 to 9 base pair of the 3' hairpin loop were crucial for cRNA promoter activity (8, 33). In addition, we also showed here that a mutated or deleted residue at position 1 at the 3' end of the cRNA promoter in a minireplicon system can revert to a wild-type sequence in vivo (Fig. 5). Therefore, taking the in vitro and in vivo data together, we conclude that the influenza virus polymerase uses an internal initiation and realignment mechanism for de novo RNA synthesis on the cRNA promoter. This contrasts with a simpler terminal initiation mechanism used by the polymerase on the vRNA promoter.
To investigate which structural and sequence properties of the cRNA promoter were required for internal pppApG initiation, we performed a transplant experiment (Fig. 3). This experiment analyzed the extent of internal pppApG synthesis after transplanting cRNA-like promoter elements into the model vRNA promoter one at a time. The results suggested that internal pppApG synthesis was mainly determined by the absence of a hinge in the 5' strand and the presence of a hinge and the identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter. These results indicate that the spatial arrangement between the 5'- or 3'-strand hairpin-loop and the duplex region in the corkscrew configuration are important determinants controlling whether internal or terminal pppApG synthesis occurs on the vRNA and cRNA promoters.
Differences in the hinge residues between the vRNA and cRNA promoters have also been found to be essential for endonuclease cleavage activity (31) and to act as the nuclear export signals for selective packaging of vRNA rather than cRNA into the progeny virions (44). It has previously been suggested that the PB1 subunit of the polymerase might adopt a different conformation upon binding the cRNA and vRNA promoters (20). Differences in vRNA and cRNA promoter structures, in particular in hinge residues, could influence their binding properties to the influenza virus RNA polymerase, thereby altering the site of initiation of pppApG synthesis. Specifically, we propose that there might be differences in the precise binding of the vRNA and cRNA promoters, so as to initially prevent access of incoming nucleotides to positions 1 and 2 of the cRNA promoter. Such access, however, would be required for positions 1 and 2 of the vRNA promoter. Proof of this hypothesis must await the successful crystallization and X-ray analysis of the influenza virus polymerase.
To further understand the mechanism of internal initiation and realignment, an in vitro template-switching experiment was performed to test whether internally synthesized pppApG, instead of realigning to the residues 1 and 2 of a cRNA promoter (cis-realignment), could realign to the terminal nucleotides of a vRNA template (trans-realignment). The results (Fig. 6) support the trans-realignment model in vitro. However, as we have not been able to design an experiment to unambiguously prove trans-alignment in vivo, we can only speculate that both cis- and trans-realignment might occur in vivo.
A question that arose from this work was the significance (if any) of pppGpU synthesized from the model cRNA promoter (Fig. 4 and 6, bands A) and pppGpC synthesized from the model vRNA promoter (Fig. 6, band C), in addition to pppApG. The potential extension of pppGpU and pppGpC (complementary to residues 2 and 3 of the cRNA and vRNA promoter, respectively) would result in replication products lacking a 5'A end. Because the 5'A ends of both vRNA and cRNA are required for polymerase binding, replication products lacking the 5' A residue would probably not be used efficiently in subsequent rounds of transcription and replication (8, 16, 40). It is possible, however, that pppGpU and pppGpC may not form full-length transcripts as efficiently as pppApG, since pppApG was preferred over pppGpU or pppGpC for elongation to a trinucleotide under our experimental conditions (Fig. 4 and 6). Thus, the significance of the observed abortive synthesis of the dinucleotides pppGpU and pppGpC remains uncertain.
Other potential explanations of our data should be considered. For example, the in vivo data for reversion to wild type of a mutated or deleted residue at position 1 at the 3' end of the cRNA promoter (Fig. 5) could, theoretically, be explained by a model similar to that proposed for the correction of the 3'-end nucleotide of hepatitis C virus RNA (6). These authors proposed that ATP might be preferentially selected over other nucleotides and might be prebound to the RNA polymerase for initiation in a non-template-directed manner. However, the inhibitory effect of a mutation at position 1 from the 3' end of the vRNA promoter for in vitro pppApG synthesis (Fig. 2), and for in vivo activity (11) would argue against this possibility. In addition, it might be argued that our model is based, in part, on the synthesis of abortive transcripts and not on full-length replication. Abortive synthesis of short oligonucleotides has been previously observed for influenza virus (42), and is, in fact, common for both prokaryotic and eukaryotic RNA- and DNA-dependent RNA polymerases (34). These abortive transcripts, predominantly di- or trinucleotides, can then be used by polymerases to initiate new rounds of RNA synthesis. Such transcripts may significantly improve the efficiency of the RNA synthesis by circumventing the first rate-limiting nucleotidyl transfer reaction (23, 36). Therefore, we believe that the internal initiation and realignment model best fits our in vitro and in vivo data.
An internal initiation and realignment mechanism is well known for many viral RNA and DNA polymerases although the mechanisms vary in detail. In the case of protein-primed transcription, as used by poliovirus RNA polymerase, the VPg primer is uridylylated from an internal loop of the template. The VPgpUpU subsequently anneals to terminal residues for positive-stranded RNA synthesis (39) Bacteriophage 29 DNA polymerase, on the other hand, uses a sliding-back mechanism, in which protein-primed transcription starts at residue 2 and then slides back to residue 1 (35). In the case of de novo RNA synthesis, the RNA polymerases of some viruses, e.g., bunya-, arena-, and nairoviruses, generate oligonucleotides internally from the 3' end of their templates for terminal realignment. However, the mechanism differs from the prime and realign mechanism proposed here because a nontemplated C residue is added (17-19, 24, 35).
Prime and realign models have been proposed as the mechanism for repair of the 3' ends of the viral templates since the 3' ends are vulnerable to degradation by host 3' exonucleases (29). In such 3'-end repair models, as reported for turnip crinkle virus, short abortive oligonucleotides serve as a "patch" to repair one or more nucleotide deletions at the 3' end (19, 21, 25, 37). In the case of the influenza virus RNA polymerase, our in vitro (Fig. 4) and in vivo (Fig. 5) results suggest that a mutation at position 1 of the 3' strand of the cRNA promoter could revert to the wild-type sequence by the internal initiation and realignment mechanism suggested here. It is also possible that a mutation at position 1 of the vRNA promoter could be corrected to the wild-type sequence by internal initiation on the cRNA promoter, followed by trans-realignment onto the vRNA promoter. If so, initiation of pppApG from the cRNA promoter may have a selective advantage because it would allow mutations at the 3' end of either cRNA or vRNA to be repaired. This mechanism could also correct transcripts to which a nontemplated nucleotide was added at its 3' end (15).
Another interesting question is why influenza virus polymerase might employ different initiation strategies for viral RNA replication on its vRNA and cRNA promoters (Fig. 7). It is well known that, late in virus infection, more vRNA than cRNA is synthesized for packaging (13, 30). Thus, vRNA and cRNA levels are differentially regulated. Due to the fact that de novo initiation of viral RNA synthesis is likely to be the rate-limiting step for replication (25, 27, 45), we speculate that the different strategies used by influenza virus RNA polymerase for initiation of replication on the vRNA and cRNA promoters may regulate the levels of cRNA and vRNA in infected cells. Therefore, it would be interesting to compare the efficiencies of the vRNA and cRNA promoters for initiation of replication.
In summary, we now know that influenza A virus RNA polymerase uses not only different initiation mechanism for viral RNA transcription (cap-snatching initiation mechanism) and replication (de novo initiation), but also different de novo initiation strategies for viral RNA replication. For the vRNA promoter, primer-independent initiation occurs at the 3' terminus of the viral genome giving rise initially to the dinucleotide pppApG immediately followed by elongation (Fig. 7B). For the cRNA promoter, pppApG is synthesized internally from positions 4 and 5, and then realigns to the 3' terminal residues 1 and 2 for subsequent elongation (Fig. 7A). We speculate that these different strategies may have been selected in the evolution of influenza A virus in order to maintain the integrity of the viral genome and to regulate vRNA and cRNA synthesis in infected cells.
In addition, the absolute sequence conservation of nucleotides 4 and 5 in the cRNA promoter of all eight segments of all influenza A, B, and C viruses (13, 30) suggests that this internal initiation and realignment mechanism is conserved among all influenza viruses. Interestingly, the sequence of nucleotides 1 to 5 of the cRNA of orthobunyaviruses and hantaviruses (19) is identical to the cRNA of influenza virus. Thus, it would be interesting to test whether an internal initiation and realignment mechanism is also used by orthobunyaviruses and hantaviruses. Finally, our internal initiation and realignment model is critically dependent on the synthesis of the dinucleotide pppApG. Nucleotide analogues or competitive substrates for the synthesis of pppApG could be considered potential novel antiviral agents.
ACKNOWLEDGMENTS
We thank Wendy Barclay for generously providing clones of influenza virus A/Turkey/England polymerase genes used in parts of this study, Ervin Fodor for helpful discussion, and Julian Robinson for DNA sequencing.
This study was supported by MRC program grant G9523972 to G.G.B., MRC cooperative group grant G9901312, and an ORS award to T.D.
T. Deng and F. T. Vreede contributed equally to this work.
REFERENCES
Area, E., J. Martin-Benito, P. Gastaminza, E. Torreira, J. M. Valpuesta, J. L. Carrascosa, and J. Ortín. 2004. 3D structure of the influenza virus polymerase complex: localization of subunit domains. Proc. Natl. Acad. Sci. USA 101:308-313.
Azzeh, M., R. Flick, and G. Hobom. 2001. Functional analysis of the influenza A virus cRNA promoter and construction of an ambisense transcription system. Virology 289:400-410.
Ball, L. A. 2001. Replication strategies of RNA viruses, p. 105-132. In D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed., vol. 1. Lippincott Williams & Wilkins, Philadelphia, Pa.
Bouloy, M., S. J. Plotch, and R. M. Krug. 1980. Both the 7-methyl and the 2'-O-methyl groups in the cap of mRNA strongly influence its ability to act as primer for influenza virus RNA transcription. Proc. Natl. Acad. Sci. USA 77:3952-3956.
Brownlee, G. G., and J. L. Sharps. 2002. The RNA polymerase of influenza A virus is stabilized by interaction with its viral RNA promoter. J. Virol. 76:7103-7113.
Cai, Z., T. J. Liang, and G. Luo. 2004. Effects of mutations of the initiation nucleotides on hepatitis C virus RNA replication in the cell. J. Virol. 78:3633-3643.
Chen, D., and J. T. Patton. 2000. De novo synthesis of minus strand RNA by the rotavirus RNA polymerase in a cell-free system involves a novel mechanism of initiation. RNA 6:1455-1467.
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.
Deng, T., J. Sharps, E. Fodor, and G. G. Brownlee. 2005. In vitro assembly of PB2 with a PB1-PA dimer supports a new model of assembly of influenza A virus polymerase subunits into a functional trimeric complex. J. Virol. 79:8669-8674.
Fechter, P., L. Mingay, J. Sharps, A. Chambers, E. Fodor, and G. G. Brownlee. 2003. Two aromatic residues in the PB2 subunit of influenza A RNA polymerase are crucial for cap binding. J. Biol. Chem. 278:20381-20388.
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., and G. G. Brownlee. 2002. Influenza virus replication, p. 1-29. In C. W. Potter (ed.), Influenza. Elsevier Science, New York, N.Y.
Fodor, E., M. Crow, L. J. Mingay, T. Deng, J. Sharps, P. Fechter, and G. G. Brownlee. 2002. A single amino acid mutation in the PA subunit of the influenza virus RNA polymerase inhibits endonucleolytic cleavage of capped RNAs. J. Virol. 76:8989-9001.
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.
Fodor, E., B. L. Seong, and G. G. Brownlee. 1993. Photochemical cross-linking of influenza A polymerase to its virion RNA promoter defines a polymerase binding site at residue 9 to residue 12 of the promoter. J. Gen. Virol. 74:1327-1333.
Garcin, D., and D. Kolakofsky. 1990. A novel mechanism for the initiation of Tacaribe arenavirus genome replication. J. Virol. 64:6196-6203.
Garcin, D., and D. Kolakofsky. 1992. Tacaribe arenavirus RNA synthesis in vitro is primer dependent and suggests an unusual model for the initiation of genome replication. J. Virol. 66:1370-1376.
Garcin, D., M. Lezzi, M. Dobbs, R. M. Elliott, C. Schmaljohn, C. Y. Kang, and D. Kolakofsky. 1995. The 5' ends of Hantaan virus (Bunyaviridae) RNAs suggest a prime-and-realign mechanism for the initiation of RNA synthesis. J. Virol. 69:5754-5762.
González, S., and J. Ortín. 1999. Distinct regions of influenza virus PB1 polymerase subunit recognize vRNA and cRNA templates. EMBO J. 18:3767-3775.
Guan, H., and A. E. Simon. 2000. Polymerization of nontemplate bases before transcription initiation at the 3' ends of templates by an RNA-dependent RNA polymerase: an activity involved in 3' end repair of viral RNAs. Proc. Natl. Acad. Sci. USA 97:12451-12456.
Honda, A., K. Mizumoto, and A. Ishihama. 2002. Minimum molecular architectures for transcription and replication of the influenza virus. Proc. Natl. Acad. Sci. USA 99:13166-131671.
Jia, Y., and S. S. Patel. 1997. Kinetic mechanism of transcription initiation by bacteriophage T7 RNA polymerase. Biochemistry 36:4223-4232.
Jin, H., and R. M. Elliott. 1993. Non-viral sequences at the 5' ends of Dugbe nairovirus S mRNAs. J. Gen. Virol. 74:2293-2297.
Kao, C. C., P. Singh, and D. J. Ecker. 2001. De novo initiation of viral RNA-dependent RNA synthesis. Virology 287:251-260.
Kao, C. C., and J. H. Sun. 1996. Initiation of minus-strand RNA synthesis by the brome mosaic virus RNA-dependent RNA polymerase: use of oligoribonucleotide primers. J. Virol. 70:6826-6830.
Kim, M. J., W. Zhong, Z. Hong, and C. C. Kao. 2000. Template nucleotide moieties required for de novo initiation of RNA synthesis by a recombinant viral RNA-dependent RNA polymerase. J. Virol. 74:10312-10322.
Kolpashchikov, D. M., A. Honda, and A. Ishihama. 2004. Structure-function relationship of the influenza virus RNA polymerase: primer-binding site on the PB1 subunit. Biochemistry 43:5882-5887.
LaCava, J., J. Houseley, C. Saveanu, E. Petfalski, E. Thompson, A. Jacquier, and D. Tollervey. 2005. RNA degradation by the exosome is promoted by a nuclear polyadenylation complex. Cell 121:713-724.
Lamb, R. A., and R. M. Krug. 2001. Orthomyxoviridae: the viruses and their replication, p. 1487-1530. In D. M. Knipe, P. M. Howley, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. R. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Leahy, M. B., G. Zecchin, and G. G. Brownlee. 2002. Differential activation of influenza A virus endonuclease activity is dependent on multiple sequence differences between the virion RNA and cRNA promoters. J. Virol. 76:2019-2023.
Lee, M. T., K. Bishop, L. Medcalf, D. Elton, P. Digard, and L. Tiley. 2002. Definition of the minimal viral components required for the initiation of unprimed RNA synthesis by influenza virus RNA polymerase. Nucleic Acids Res. 30:429-438.
Li, X., and P. Palese. 1992. Mutational analysis of the promoter required for influenza virus virion RNA synthesis. J. Virol. 66:4331-4338.
Matsumoto, J. 1994. Evolutionary role of abortive transcript as a primer for DNA replication. J. Mol. Evol. 39:620-624.
Mendez, J., L. Blanco, J. A. Esteban, A. Bernad, and M. Salas. 1992. Initiation of 29 DNA replication occurs at the second 3' nucleotide of the linear template: a sliding-back mechanism for protein-primed DNA replication. Proc. Natl. Acad. Sci. USA 89:9579-9583.
Moroney, S. E., and J. A. Piccirilli. 1991. Abortive products as initiating nucleotides during transcription by T7 RNA polymerase. Biochemistry 30:10343-10349.
Nagy, P. D., C. D. Carpenter, and A. E. Simon. 1997. A novel 3'-end repair mechanism in an RNA virus. Proc. Natl. Acad. Sci. USA 94:1113-1118.
Paul, A. V. 2002. Possible unifying mechanism of picornavirus genome replication, p. 227-246. In B. L. Semler and E. Wimmer (ed.), Molecular biology of picornaviruses. American Society for Microbiology, Washington, D.C.
Paul, A. V., J. Yin, J. Mugavero, E. Rieder, Y. Liu, and E. Wimmer. 2003. A "slide-back" mechanism for the initiation of protein-primed RNA synthesis by the RNA polymerase of poliovirus. J. Biol. Chem. 278:43951-43960.
Pritlove, D. C., E. Fodor, B. L. Seong, and G. G. Brownlee. 1995. In vitro transcription and polymerase binding-studies of the termini of influenza A virus cRNA - evidence for a cRNA panhandle. J. Gen. Virol. 76:2205-2213.
Rao, P., W. Yuan, and R. M. Krug. 2003. Crucial role of CA cleavage sites in the cap-snatching mechanism for initiating viral mRNA synthesis. EMBO J. 22:1188-1198.
Seong, B. L., and G. G. Brownlee. 1992. Nucleotides 9 to 11 of the influenza A virion RNA promoter are crucial for activity in vitro. J. Gen. Virol. 73:3115-3124.
Szymkowiak, C., W. S. Kwan, Q. Su, T. J. Toner, A. R. Shaw, and R. Youil. 2003. Rapid method for the characterization of 3' and 5' UTRs of influenza viruses. J. Virol. Methods 107:15-20.
Tchatalbachev, S., R. Flick, and G. Hobom. 2001. The packaging signal of influenza viral RNA molecules. RNA 7:979-989.
van Dijk, A. A., E. V. Makeyev, and D. H. Bamford. 2004. Initiation of viral RNA-dependent RNA polymerization. J. Gen. Virol. 85:1077-1093.
Vreede, F. T., T. E. Jung, and G. G. Brownlee. 2004. Model suggesting that replication of influenza virus is regulated by stabilization of replicative intermediates. J. Virol. 78:9568-9572.(Tao Deng, Frank T. Vreede)
ABSTRACT
Various mechanisms are used by single-stranded RNA viruses to initiate and control their replication via the synthesis of replicative intermediates. In general, the same virus-encoded polymerase is responsible for both genome and antigenome strand synthesis from two different, although related promoters. Here we aimed to elucidate the mechanism of initiation of replication by influenza virus RNA polymerase and establish whether initiation of cRNA and viral RNA (vRNA) differed. To do this, two in vitro replication assays, which generated transcripts that had 5' triphosphate end groups characteristic of authentic replication products, were developed. Surprisingly, mutagenesis screening suggested that the polymerase initiated pppApG synthesis internally on the model cRNA promoter, whereas it initiated pppApG synthesis terminally on the model vRNA promoter. The internally synthesized pppApG could subsequently be used as a primer to realign, by base pairing, to the terminal residues of both the model cRNA and vRNA promoters. In vivo evidence, based on the correction of a mutated or deleted residue 1 of a cRNA chloramphenicol acetyltransferase reporter construct, supported this internal initiation and realignment model. Thus, influenza virus RNA polymerase uses different initiation strategies on its cRNA and vRNA promoters. To our knowledge, this is novel and has not previously been described for any viral RNA-dependent RNA polymerase. Such a mechanism may have evolved to maintain genome integrity and to control the level of replicative intermediates in infected cells.
INTRODUCTION
Most RNA viruses encode their own RNA-dependent RNA polymerases in order to catalyze viral genome transcription and replication (3). The structures solved to date show that all viral RNA-dependent RNA polymerases form a similar basic right-hand-like structure with fingers, palm, and thumb subdomains (45). However, different viral RNA polymerases also have unique structural features which are important for efficient and accurate initiation of RNA-dependent RNA synthesis (25, 45). An amazing variety of initiation mechanisms for viral RNA-dependent RNA polymerization have been demonstrated. Any given RNA virus may use either one or more of these initiation mechanisms (25, 45).
In general, initiation mechanisms can be classified into two different categories, primer-dependent and primer-independent (de novo) initiation (25, 45). For primer-dependent initiation, the most often used primers are either (i) an oligonucleotide covalently linked to a protein, as used by picornaviruses (e.g., poliovirus) (38) or (ii) a capped primer cleaved from the 5' end of the host mRNA by a cap-snatching mechanism, which is used for viral genome transcription by some segmented negative-stranded RNA viruses such as influenza virus (4) and Dugbe nairovirus (24). In general, viral RNAs without cap-snatched primers or terminal proteins are initiated by a de novo initiation mechanism, which is common for many positive, negative, double-stranded, and ambisense RNA viruses (25). This mechanism is, clearly, a critical step to ensure efficient production of new viral RNA and to maintain the integrity of the viral genome (25, 45).
In most cases, de novo initiation starts at residue 1 of the template and is immediately followed by elongation. However, in some cases, e.g., Tacaribe arenavirus, it has been observed that short RNA oligonucleotides may be produced by successive rounds of abortive de novo transcription directed by either terminal nucleotides or internal nucleotides of the template. These oligonucleotides are subsequently used as primers for further elongation, as in the so-called prime-and-realign mechanism (7, 17-19, 26). A similar mechanism has also been proposed to explain 3' end repair in turnip crinkle virus (37).
Influenza A virus RNA polymerase is a heterotrimeric complex composed of three subunits, PB1, PB2, and PA. No crystal structure is available for this complex, although low-resolution electron microscopic images have been described (1). The PB1 subunit is the key component of the complex, containing conserved motifs of RNA-dependent RNA polymerases (13, 30). The PA and PB2 subunits are also required for both viral RNA transcription and replication (13, 30). The polymerase catalyzes both primer-dependent RNA transcription (vRNA to mRNA synthesis) and primer-independent replication (vRNA to cRNA and cRNA to vRNA synthesis) in the nucleus of infected cells (13, 30, 46). The mechanism of transcription (vRNA to mRNA synthesis) has been well studied since both in vivo and in vitro assays are available (10, 14, 41). In contrast, the mechanism of replication remains unclear. The reason for this is that it is difficult to distinguish replication directed by the vRNA versus the cRNA promoters in vivo because of the interdependence of the two promoters in the replication cycle. Existing in vitro primer-independent replication assays (22, 32) are inefficient and may, therefore, not be authentic.
The current model for the secondary structure of both the vRNA and cRNA promoters is the so-called "corkscrew" model (2, 8, 12). This model proposes a partially duplex region formed by the conserved noncoding sequences at both the 5' and 3' ends of each gene segment and two hairpin-loop structures, each with a stem of 2 bp and a tetraloop formed by residues close to the 3' and 5' termini. The secondary structure of the cRNA promoter proposed by Crow et al. (8) was based on the results of both a mutagenic analysis of the cRNA promoter in an in vivo minireplicon system, and in vitro promoter binding studies. However, there remained unexplained differences between the in vivo and in vitro results. In particular, point mutations at positions 4, 5, and 7 of the 3' strand of the cRNA promoter inhibited vRNA, cRNA, and mRNA synthesis in vivo, whereas in vitro approaches showed that RNA polymerase can bind all three of these point mutants in a UV cross-linking assay (8) and transcribe them in an ApG-primed transcription assay (40). According to the "corkscrew" structure of the cRNA promoter, the nucleotides at positions 4, 5, and 7 of the 3' strand of the promoter are located in a hairpin loop region, and point mutations at these positions would not be expected to disrupt the secondary structure and functions of the promoter in vivo. To resolve differences between the in vivo and in vitro data, an authentic in vitro primer-independent replication assay is needed.
Here, we developed two in vitro primer-independent RNA synthesis assays to study primer-independent RNA replication on model promoters. By mutational analyses of the two promoters, we found that influenza virus RNA polymerase uses internal initiation and realignment on its cRNA promoter but terminal initiation and elongation on its vRNA promoter during replication. These results shed new light on previous in vivo data on the cRNA promoter (8).
MATERIALS AND METHODS
Plasmids. The pcDNA-PB1, pcDNA-PB2, pcDNA-PA, and pcDNA-NP protein expression plasmids for the three polymerase subunits and NP of influenza A/WSN/33 virus have been described (14). The pPOLI-cCAT-RT (8), pcDNA-PB2tap (9), and active-site mutant (ASM) pcDNA-PB1-D445A/D446A (46) plasmids have been previously described. The pcDNA-PB1, pcDNA-PB2tap, and pcDNA-PA protein expression plasmids for the three polymerase subunits of influenza A/Turkey/England/5092/91 virus polymerase were provided by Ruth Harvey.
RNA oligonucleotides. RNA oligonucleotides (Dharmacon) for the vRNA promoter were 5' AGUAGAAACAAGGCC 3' and 5' GGCCUGCUUUUGCU 3' and for the cRNA promoter were 5' AGCAAAAGCAGGCC 3' and 5' GGCCUUGUUUCUACU 3'. They contained the conserved sequence (underlined) plus two additional residues. Mutant oligonucleotides based on the above vRNA and cRNA promoters were also from Dharmacon.
Preparation of partially purified recombinant influenza virus RNA polymerase. 293T cells were transfected with plasmids (see above) using Lipofectamine 2000. After 48 h the cells were lysed and TAP-tagged influenza virus polymerase (3P) complex with or without NP was purified by affinity purification with immunoglobulin G-Sepharose, and released by tobacco etch virus protease as before (9). Partially purified proteins in tobacco etch virus protease cleavage buffer (0.15 M NaCl, 0.01 M HEPES, pH 8.0, 0.1% NP-40, 1 mM dithiothreitol, 10% glycerol) were stored at –20°C in 35% glycerol. Partially purified His-tagged influenza virus polymerase, used in the UV cross-linking competition assay, was prepared as described previously (8).
In vitro cRNA replication assay. We performed 3-μl reactions with 1.5 μl partially purified 3P/NP and 0.7 μM synthetic model cRNA promoter (an equimolar mixture of the 5' and 3' strands) in the presence of 5 mM MgCl2, 1 mM dithiothreitol, 1 mM ATP, 0.5 mM GTP, 0.5 mM UTP, 0.1 μM [-32P]CTP (>3,000 Ci/mmol, Amersham), and 1 U/μl RNasin (Promega). Where indicated, 0.5 mM ApG (Sigma) was added to the reaction. Reactions were incubated at 30°C overnight for convenience, although significant activity was obtained within 1 to 3 h. An equal volume of formamide/bromophenol blue/EDTA was added and the mixture was heated at 95°C for 5 min. An aliquot, usually 4 μl, was analyzed by 25% polyacrylamide gel electrophoresis (PAGE) in 1x Tris-borate-EDTA and 6 M urea. Autoradiography was performed with Kodak Biomax MS films and Trans screen HE intensifying screens at –80°C.
UV cross-linking assay. UV cross-linking assays were performed as described previously (8).
Dinucleotide replication initiation assay and extension to trinucleotide assay. We performed 3-μl incubations with (normally) 1.5 μl polymerase, 0.02 μM [-32P]GTP (>3,000 Ci/mmol, Amersham) in the presence of 5 mM MgCl2, 1 U/μl RNasin (Promega). and 1 mM dithiothreitol with the addition of 1 mM ATP and 0.7 μM model vRNA (an equimolar mixture of the 5' and 3' strands) or 0.7 μM model cRNA (an equimolar mixture of the 5' and 3' strands) promoters.
To assay for extension to trinucleotides, either 1 mM UTP or 1 mM CTP, as specified in the individual experiments, were added. In the "transplant" experiment (see Fig. 3), the 5' strand was added first and incubated for 10 min at 30°C prior to addition of the 3' strand to avoid self-annealing between the two strands. In the "template-switching" experiment (Fig. 6), a preincubation of either the vRNA or the cRNA promoter with polymerase was performed at 30°C for 10 min before mixing. Incubations were at 30°C for various times (usually 1 to 3 h) up to 16 h. Formamide/bromophenol blue/EDTA (10 μl) was added and the mixture was heated at 95°C for 2 min. Analysis, usually of a 5 μl aliquot, was by 25% PAGE in 1x Tris-borate-EDTA and 6 M urea, followed by autoradiography.
Gel elution, enzymatic digestion, and analysis of oligoribonucleotides. The excised band of gel containing oligoribonucleotides <5 residues in length was rinsed for 1 min in 1 ml deionized water and then crushed in 0.5 ml deionized water and shaken overnight at 4°C. After centrifugation at 10,000 x g to remove gel particles, the supernatant was freeze-dried. Finally, the lyophilized sample was dissolved in a minimum volume (5 to 10 μl) of deionized water. T2 RNase (Sigma), which degrades RNA to give 3'-phosphorylated mononucleotides, was used essentially as recommended by the manufacturers. After digestion, 0.5 μl 10 mM ATP was normally added to the sample (usually 3 to 5 μl) prior to loading onto a polyethyleneimine-cellulose (Merck PEI-cellulose F) thin layer, which was developed for 4 cm above the origin in 0.5 M ammonium sulfate, and then further developed in 0.7 M ammonium sulfate.
The mobility of pppA32p on PEI-cellulose thin-layer chromatography (TLC) was deduced because the product (pA32p) of its further digestion with tobacco acid pyrophosphatase comigrated (RF = 0.50) with unlabeled, marker pAp (Sigma) (results not shown). The mobility of pp32pGp on PEI-cellulose TLC was deduced from T2 RNase digestion of pp32pGpU, which had been characterized as possessing a pppG end group because it gave rise to a product, after P1 nuclease digestion, that comigrated with GTP on PEI-cellulose TLC (results not shown). Oligonucleotides >5 residues were gel-eluted in 0.5 M ammonium acetate and ethanol precipitated by standard methods. Calf intestinal phosphatase (Roche), tobacco acid pyrophosphatase (Epicenter) and polynucleotide kinase (Roche) treatments were carried out for 1 h at 37°C in the supplied buffer. Digested products were analyzed by 25% PAGE as above.
Primer extension assay. Primer extension assays were performed as described previously (8).
Terminal sequencing of viral RNA. Sequencing of the 5' and 3' termini of viral RNA transcripts was carried out essentially as described previously (43). Briefly, total cellular RNA from transfected 293T cells was extracted with Trizol (Invitrogen) and treated with TAP (Epicenter) prior to being reextracted with Trizol. The RNA was then circularized with T4 RNA ligase (Epicenter). cDNA copies of the ligated 3' and 5' termini were synthesized by reverse transcription (8) with separate vRNA- or cRNA-specific primers. The cDNA was amplified by PCR using Taq polymerase (Promega) and both vRNA- and cRNA-specific primers, cloned using the TOPO TA cloning kit (Invitrogen), and sequenced.
RESULTS
Mutational analysis of the 3' end of the influenza virus cRNA promoter using a primer-independent in vitro replication assay. An ApG-primed replication assay has been used previously to study the sequence requirements of the 3' end of the influenza virus cRNA promoter for RNA replication with nonrecombinant RNA polymerase (40). However, the exact nature of this assay, whether it mimics replication or transcription, was not clear. Here we developed an in vitro primer-independent replication assay based on the incorporation of [-32P]CTP (0.1 μM) into a 15-nucleotide transcript of a model cRNA promoter (Fig. 2A) in the presence of 0.5 mM GTP and UTP and 1 mM ATP. Partially purified recombinant influenza virus RNA polymerase isolated from transiently transfected 293T cells was used as a source of enzyme (5). Figure 1A (lane 1, upper panel) shows that a 15-nucleotide full-length transcript was synthesized from the model wild-type cRNA promoter.
The essential feature of an authentic primer-independent replication product is the presence of a triphosphate moiety at its 5' end. In order to confirm that unprimed transcripts had a 5' triphosphate end group, we compared the mobilities of the 15-nucleotide unprimed transcript with the ApG-primed transcript, variously treated with calf intestinal phosphatase, tobacco acid pyrophosphatase, or polynucleotide kinase on a 25% PAGE-urea gel (Fig. 1B). The 15-nucleotide unprimed transcripts (5' pppA) migrated faster than the ApG-primed transcripts (5'OHA) (Fig. 1B, compare lane 4 with 1 and 2). Unprimed transcripts treated with calf intestinal phosphatase (5' pppA5' OHA), comigrated with the ApG-primed transcripts (5' OHA) (Fig. 1B, compare lanes 3 with 1). When the unprimed transcripts were treated with tobacco acid pyrophosphatase (5' pppA5' pA), they comigrated with the polynucleotide kinase-treated ApG-primed transcripts (5' OHA5' pA) (Fig. 1B, compare lanes 5 and 6). In addition, a faint band migrated slightly faster than the unprimed transcripts in lane 4. This faint band is likely to be a 5' diphosphate-ended transcript since it had an intermediate mobility between the 5' p- and 5' ppp-ended transcripts. Its origin is uncertain, but it may be due to a contaminant nucleotidyl pyrophosphatase in our enzyme preparation. Therefore, a majority of transcripts had a 5' triphosphate end group.
Having validated the authenticity of this primer-independent transcription assay, we compared the effects of point mutations at positions 1 to 9 from the 3' end of the 3' strand of the model cRNA promoter in this primer-independent replication assay (Fig. 1A, upper panel) with an ApG-primed transcription assay (Fig. 1A, lower panel) under the same conditions. In the unprimed replication assay, point mutants of the promoter at positions 1, 6, 7, and 8 showed transcription activities similar to the wild type (Fig. 1A, upper panel, compare lanes 3 and 8 to 10 with lane 1). Lower yields were obtained with the mutant at position 3, and only residual levels were obtained for the mutant at position 9. Interestingly, however, no products were detected with the mutants at positions 2 and 4 to 5. In the ApG-primed transcription assay, however, significant activity was observed with the mutants at positions 3 to 4 and 6 to 8 (Fig. 1A, lower panel, compare lanes 5 to 6 and 8 to 10 with lane 1), lower activities were observed with the mutants at positions 1 and 5, residual activity was observed with the mutant at position 9, but no activity was observed with the mutant at position 2. The most obvious difference between the results of the two assays is at positions 4 and 5. Mutations here interfered with unprimed transcription completely, but not with the ApG-primed transcription (Fig. 1A, compare lanes 6 and 7 in the upper and lower panels).
The polymerase binding properties of these two mutants had been previously studied in a UV cross-linking competition assay where it was shown that both mutants compete with the wild-type model cRNA promoter (8). A more detailed UV cross-linking competition analysis of these two mutants is shown here (Fig. 1C), confirming that the 4UA and 5CA mutants were not significantly different from the wild-type model cRNA promoter in competing with cross-linking of a radiolabeled wild-type promoter to the subunits (PB1, PB2, and PA) of the polymerase (Fig. 1C, compare lanes 5 to 8 and 9 to 12 with lanes 1 to 4). Thus, we excluded the possibility that the absence of any replication products observed for the mutants at positions 4 and 5 of the 3' strand of the model cRNA promoter in the primer-independent replication assay (Fig. 1A, upper panel) was due to their weak binding to the polymerase.
pppApG is copied from nucleotides at positions 4 and 5 of the 3' end of the cRNA promoter in contrast with positions 1 and 2 of the vRNA promoter. Influenza virus ribonucleoprotein isolated from virions is known to synthesize the dinucleotide pppApG in vitro by copying the terminal 3' OHU-C 5' residues of the vRNA template under conditions where only ATP and GTP are present (28). Here we have adapted this dinucleotide synthesis assay to investigate initial events in replication (i.e., synthesis of pppApG) in the absence of added primer with recombinant influenza virus RNA polymerase and model vRNA and cRNA promoters (Fig. 2A and B).
Wild-type and mutant promoters were used to test whether the observed lack of replication (Fig. 1A) was caused by events early in replication. We found that mutations at positions 4 and 5 of the 3' strand of the model cRNA promoter (Fig. 2C, lanes 6 to 7) inhibited pppApG synthesis dramatically, whereas mutants at positions 1 and 2 (Fig. 2C, lanes 3 and 4) showed activity similar to that of the wild type. By contrast, mutants at positions 1 and 2 of the model vRNA promoter (Fig. 2D, lanes 3 to 4) were inactive in pppApG synthesis. These results indicated that the vRNA and cRNA promoters have distinctive sites of pppApG synthesis i.e., the 3' OHU-C 5' at positions 1 to 2 of the 3' end of the vRNA promoter, but the 3' U-C 5' at positions 4 to 5, from the 3' end of the cRNA promoter.
In addition, mutations elsewhere within the promoters can influence pppApG synthesis activity (Fig. 2C and D). In the case of the cRNA promoter, mutants at positions 6, 7 and 8 (Fig. 2C, lanes 8, 9, and 10) retained significant pppApG synthesis activity, while the mutant at position 9 showed only residual activity (Fig. 2C, lane 11). The dinucleotide synthesized from the position 3 (AU) mutant (Fig. 2C, lane 5) showed a slightly slower mobility and was found to consist of both pppApG and pppGpA by T2 RNase digestion (results not shown). In the case of the vRNA promoter, mutants at positions 4 to 6 and 8 showed weak activity (Fig. 2D, lanes 6 to 8 and 10), although mutants at positions 7 and 9 were inactive (Fig. 2D, lanes 9 and 11 to 12). Interestingly, the mutant at position 3 (GA) was more active than the wild type (Fig. 2D, compare lanes 2 with 5). In summary, the pattern of pppApG synthesis observed in the mutagenesis screening (Fig. 2C and D) differs significantly between the two promoters. This further suggests that the RNA polymerase initiates pppApG synthesis on the vRNA and cRNA promoters in different ways.
We next asked if a model vRNA promoter, mutated at position 5 to a C residue, thereby generating a 3' U-C 5' sequence at positions 4 and 5, could allow pppApG synthesis at positions 4 and 5. To do this we had to introduce an additional mutation at position 1 (UA) to block any possibility of terminal initiation. Figure 2E shows that the double mutant (5UC, 1UA) (lane 4) was inactive in pppApG synthesis, while the 5UC mutant (lane 3) still synthesized pppApG as we had observed with 5UA mutant in Fig. 2D (lane 7). This suggested that simply mutating position 5 UC in the vRNA promoter does not permit the synthesis of pppApG internally. Overall, these results indicate that intrinsic differences in the vRNA and cRNA promoters may control the site of pppApG synthesis.
Absence of a hinge in the 5' strand and presence of a hinge and identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter are the major determinants of internal pppApG synthesis. The cRNA promoter is complementary to the vRNA promoter, and a very similar corkscrew secondary structure model has been proposed for both promoters (2, 8, 12) (Fig. 2A and B). Nevertheless, there are obvious differences between the two promoters. First, the lengths of the 5' and 3' arms differ due to the presence or absence of a hinge, an extra nucleotide located between the base-paired duplex region and the hairpin-loop structure (Fig. 2A and B). In the case of the 5' arm of the promoters, a hinge A residue is present in the vRNA promoter but is absent in the cRNA promoter. In the case of the 3' arm of the promoters, a hinge U residue is present in the cRNA promoter but is absent in the vRNA promoter.
Second, the identities of the 3 to 8 base pair and the nucleotide at position 5 of the hairpin loop structures of both 5' and 3' arms vary. Specifically, a C-G base pair at positions 3' and 8' and an A at position 5' in the 5' hairpin loop of cRNA are replaced by a U-A base pair and a G at the corresponding positions of vRNA. In the 3' hairpin loop, an A-U base pair at positions 3 and 8 and a C at position 5 of cRNA are replaced by G-C and U, respectively, in vRNA (Fig. 2A and B).
In order to test which of these sequences or presumed secondary structural elements in cRNA are responsible for internal pppApG synthesis, a series of model 5'-strand and 3'-strand promoter mutants that transplant the cRNA-like elements one by one into the vRNA promoter were designed (Fig. 3A and C). In addition, a mutation was introduced at position 1 (UA) of the 3' strand to exclude any possibility of terminal pppApG synthesis (Fig. 2D, lane 3).
The 5'-arm vRNA promoter mutants transplanted with different 5'-strand cRNA-like elements were first tested in combination with the 3' arm of the cRNA promoter with the additional 1UA mutation (see above) to evaluate which elements of the 5' arm were needed for internal initiation on the 3' arm of cRNA. Figure 3B shows that no internal pppApG synthesis occurred when the nucleotide 5'A of cRNA replaced 5'G of vRNA (lane 3), and only residual activity was obtained when the 3'C-8'G base pair of cRNA replaced the 3'U-8'A base pair of vRNA (lane 4). However, significant activity was detected when 10'A was deleted from vRNA (lane 5). No further significant increase in activity was observed if this 10'A deletion was combined with either the 5'A element (lane 6) or the 3'C-8'G element (lane 7). This result suggests that the absence of a 10A hinge in the 5' arm of the cRNA promoter is the key structural element controlling internal pppApG synthesis.
Second, we tested which 3'-arm cRNA elements were required for internal pppApG synthesis by again transplanting cRNA elements into the vRNA promoter. In this case, the residue at position 5 (a U residue) of the vRNA promoter had to be mutated to a C residue to allow pppApG transcription from positions 4 and 5. The 1(UA) mutation was also used throughout to block pppApG synthesis from nucleotides 1 and 2. The cRNA elements 3A and 8U were treated separately since point mutants at these positions showed significant activity (Fig. 2C). These elements and the 10U hinge were then transplanted individually or in different combinations into the 3' arm of the modified vRNA backbone and tested in combination with the 5' arm of the cRNA promoter for pppApG synthesis (Fig. 3C). No significant activity was obtained with the mutants bearing either the 3A or the 8U, or both the 3A and 8U cRNA elements (Fig. 3D, lanes 6, 7, and 8), whereas residual activity was observed when the 10U hinge was inserted (lane 9). However, when the 10U hinge was introduced in combination with the 3A element, wild-type activity was observed (Fig. 3D, compare lanes 10 and 2). In contrast, only partial activity was obtained when the 10U hinge was introduced in combination with the 8U element (Fig. 3D, lane 11). As expected, wild-type activity was obtained when all three elements (3A, 8U, and 10U) were present (Fig. 3D, lane 12). These results suggest that the combination of the 3A element and the 10U hinge in the 3' end of the cRNA promoter is required for full internal pppApG synthesis activity. Overall, we concluded that the absence of the hinge in the 5' strand and the presence of the hinge and the identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter are the major structural elements that determine internal pppApG synthesis.
Internally synthesized pppApG anneals to position 1 and 2 of the 3' strand of the cRNA promoter for extension into trinucleotides in a template-directed manner. Theoretically, to maintain the integrity of the viral genome, internally synthesized pppApG from nucleotides 4 and 5 of the 3' strand of the cRNA promoter would have to realign to nucleotides 1 and 2 for elongation. To test this model, we performed the pppApG synthesis reaction in the presence of UTP in addition to ATP and radiolabeled GTP. Under these conditions, we observed pppApG and pppGpU synthesis (see Discussion) (Fig. 4A and B, band A) as well as elongation to a trinucleotide, although elongation was inefficient (Fig. 4A, lane 2). This trinucleotide (Fig. 4B, band E) was characterized as pppApGpX by gel elution and T2 RNase digestion, followed by PEI-cellulose TLC. The identity of X was deduced as a U residue because no trinucleotide product was observed when the UTP was omitted or CTP was added instead (results not shown). However, we failed to observe elongation to either a 5-nucleotide product, pppAGAAA, derived from residues 4 to 8 of the cRNA promoter (Fig. 2A) or an 8-nucleotide product, pppAGUAGAAA, derived from residues 1 to 8 (results not shown). This failure, and the low yield of the trinucleotide pppApGpU, could be due to the low concentration of GTP used in this reaction.
To further test the properties of the cRNA promoter required for realignment, we studied the 1UA and 2CA mutations, together with the 4UA and 5CA mutants as controls. The previous result that mutations at positions 4 and 5 inhibited pppApG synthesis was confirmed (Fig. 4A, lanes 5 and 6), whereas pppGpU (Fig. 4B, band D) was the major product synthesized from the 4UA mutant. Lanes 3 and 4 showed that the 1UA mutation but not the 2CA mutation allows trinucleotide pppApGpU synthesis. The presence of pppApGpU, derived from the wild type and the mutant at position 1 suggested that pppApG, internally synthesized from residues 4 and 5 of the model cRNA promoter, was able to realign with wild-type residues 1 and 2 of the promoter and extend by one residue, albeit inefficiently.
The 5' pppA end group of band F derived from the 1UA mutant (Fig. 4A, band F), as confirmed by T2 RNase digestion and PEI-cellulose TLC (Fig. 4B), shows that the trinucleotide transcript had effectively corrected the 1UA mutation of the cRNA template by realignment. This result further suggests that base-pairing between the first nucleotide of pppApG and the terminal residue of the 3' end of the cRNA promoter template is not critical for realignment. On the other hand, no trinucleotide was observed from the mutant at position 2 (Fig. 4A, lane 4), suggesting that base-pairing between the second nucleotide of pppApG and the C residue at position 2 of the 3' end of the model cRNA promoter template is essential and sufficient for extension in vitro. Based on these results, we conclude that the influenza virus cRNA promoter might use an internal initiation and realignment mechanism to initiate viral RNA synthesis.
Mutated or deleted nucleotide at position 1 of the 3' end of the cRNA promoter is corrected to the wild-type nucleotide in a minireplicon system. As was shown in Fig. 4A (lane 3, band F) above, a wild-type trinucleotide transcript had been generated in vitro from a cRNA template mutated at position 1. In vivo, it has been shown that a mutation at position 1 of the 3' end of the cRNA promoter was compatible with the synthesis of significant levels of mRNA, vRNA and cRNA from a CAT reporter (8). Thus, it was of interest to test whether the 1UA mutant previously studied in vivo was corrected to the wild-type sequence.
Two additional mutants with a 1UC or a deleted residue 1 (1U) were generated and, together with the previously isolated 1UA mutant, were tested in a primer extension assay (8). This assay measures the steady-state levels of vRNA, mRNA and cRNA in a minireplicon system in which 293T cells were transiently transfected with 4 protein expression plasmids encoding the three subunits of influenza virus RNA polymerase and the nucleoprotein, together with a pPOLIcCAT plasmid which encodes a CAT reporter gene flanked by the influenza virus cRNA promoter (8).
Figure 5A shows that templates with a mutated or deleted residue 1 were replicated (vRNA and cRNA) and transcribed (mRNA) to wild-type levels (Fig. 5A, lanes 1, 3, 5, and 7). As controls, the background levels of unamplified polymerase I (POLI) transcripts of the wild-type and the three mutant templates (Fig. 5A, lanes 2, 4, 6, and 8) were obtained by transfection of an active-site polymerase mutant (replacing wild-type PB1 with mutant PB1-D445A/D446A, abbreviated 3P-ASM) (46). Sequencing of the RNA (see materials and methods) showed that all three mutants (i.e., the two mutated and the one deleted nucleotide at position 1) had reverted to the wild-type sequence (Fig. 5B).
pppApG derived from the cRNA promoter can be released and realigned onto a vRNA promoter in vitro. According to our model, pppApG initiates internally and is extended by realignment onto terminal residues of the cRNA template (cis-realignment). It was of interest to establish whether pppApG synthesized internally on a cRNA template could dissociate and realign to nucleotides 1 and 2 of a vRNA template for extension (trans-realignment). To test this, we designed an in vitro template-switching experiment in which a cRNA promoter mutant, which allows pppApG synthesis internally but does not allow extension to a trinucleotide, is mixed with a vRNA promoter mutant, which does not allow pppApG synthesis terminally but would allow extension to a trinucleotide if pppApG were available. Thus, if the pppApG dissociated from the cRNA promoter mutant, it would be available for the vRNA promoter mutant to use for extension into trinucleotide. If, however, the pppApG remained associated with the cRNA promoter mutant, no extension to a trinucleotide would occur. From previous mutational analyses of the cRNA and vRNA promoters (Fig. 2C and D), the 3'-strand cRNA promoter mutant 2 (CA) and the 3'-strand vRNA promoter mutant 1 (UA) fit these criteria. Also, by taking advantage of the different mobilities of pppApGpC (Fig. 6A, lane 4) directed by a model vRNA template and pppApGpU (Fig. 6A, lane 1) directed by a model cRNA template, we could distinguish whether the template of the synthesized trinucleotides was the vRNA or the cRNA promoter.
As controls, we used model wild-type cRNA and vRNA in the presence of either UTP or CTP to synthesize trinucleotides as markers (Fig. 6A, lanes 1 to 4). As expected, pppApGpU (band D) was only synthesized from the cRNA template and only in the presence of UTP (lane 1). Similarly, pppApGpC (Fig. 6A, band E) was only synthesized from the vRNA template and only in presence of CTP (Fig. 6A, lane 4). The cRNA and vRNA promoter mutants mentioned above were also included as further controls to confirm that the 2CA cRNA mutant can generate pppApG internally but cannot be extended into a trinucleotide in the presence of either UTP or CTP (Fig. 6A, lanes 5 and 6), and the 1UA vRNA promoter mutant cannot generate pppApG or trinucleotides (Fig. 6A, lanes 7 and 8).
The test for template switching was carried out by mixing the model cRNA promoter mutant (2CA) and the model vRNA promoter mutant (1UA). The result showed that the trinucleotide pppApGpC (Fig. 6A, band F) was only observed when CTP, but not UTP, was provided in the reaction (Fig. 6A, compare lanes 9 and 10). This suggested that pppApG synthesized by the cRNA promoter mutant (2CA) had realigned to the vRNA promoter (1UA) and extended into the trinucleotide pppApGpC. We were also able to confirm that template switching occurred when the wild-type cRNA promoter was mixed with the mutant vRNA promoter (1UA) (Fig. 6A, lanes 11 to 12). In this case, both pppApGpU (band G), which was synthesized from the wild-type cRNA promoter template, and pppApGpC, (band H), which was synthesized by template switching of pppApG from the wild-type cRNA promoter to the vRNA promoter mutant, were observed. The pppA end group of these trinucleotides (bands D to H) was confirmed by T2 RNase digestion and comparison with T2 RNase-digested dinucleotides on PEI-cellulose TLC (Fig. 6B). Overall, we conclude that pppApG synthesized internally from the cRNA template can be realigned onto a cRNA or a vRNA template for elongation.
DISCUSSION
De novo RNA synthesis is a common strategy used by viral RNA-dependent RNA polymerases for viral RNA replication to maintain the integrity of the viral genome (25, 45). However, the mechanisms for primer-independent initiation vary among different RNA viruses (25, 45). The aim of this work was to study the mechanism by which influenza virus polymerase initiates replication on its vRNA and cRNA promoters.
Here, we have developed two in vitro primer-independent RNA synthesis assays with recombinant influenza virus RNA polymerase and model promoters. The first assays influenza virus cRNA promoter-specific replication by primer-independent transcription of a 15-nucleotide template. The transcripts were shown to possess 5'-triphosphate end groups, demonstrating the authenticity of the assay. By mutagenesis screening, we found that point mutations at positions 2, 4, 5, and 9 from the 3' end of the cRNA promoter largely inhibited replication in this assay. In contrast, ApG-primed transcription was only inhibited by mutations at positions 2 and 9, with partial inhibition at position 5 (Fig. 1A). This suggested that pppApG might be synthesized from residues 4 and 5.
We therefore investigated the site of de novo pppApG synthesis by a second assay, an in vitro replication initiation assay. We found that the nucleotides at positions 4 and 5 from the 3' end of the cRNA promoter were important for directing pppApG synthesis in contrast with the nucleotides at positions 1 and 2 at the 3' end of the vRNA promoter (Fig. 2). As residues 4 and 5 from the 3' end of the cRNA promoter are 3' U-C 5', we hypothesized that pppApG may be templated internally from these residues and then realigned to the terminal 3' OHU-C 5' for elongation. Accordingly, we showed that pppApGpU was synthesized from the wild-type cRNA promoter and the mutant at position 1, but not from the mutant at position 2 (Fig. 4), suggesting that internally synthesized pppApG can realign onto positions 1 and 2 for extension into trinucleotide pppApGpU in a template-directed manner (Fig. 4). This further suggests that base-pairing of pppApG with the C residue at position 2 alone was sufficient for realignment of pppApG and elongation, and that a mutation at position 1 of the template would be corrected to the wild-type base in the transcript. This realignment would explain why mutations at positions 2, 4, and 5 in the model cRNA failed to synthesize a 15 nucleotide unprimed transcript (Fig. 1).
These data are also consistent with the previous unexplained in vivo observations that nucleotides at positions 4 and 5 and the identity of the 2 to 9 base pair of the 3' hairpin loop were crucial for cRNA promoter activity (8, 33). In addition, we also showed here that a mutated or deleted residue at position 1 at the 3' end of the cRNA promoter in a minireplicon system can revert to a wild-type sequence in vivo (Fig. 5). Therefore, taking the in vitro and in vivo data together, we conclude that the influenza virus polymerase uses an internal initiation and realignment mechanism for de novo RNA synthesis on the cRNA promoter. This contrasts with a simpler terminal initiation mechanism used by the polymerase on the vRNA promoter.
To investigate which structural and sequence properties of the cRNA promoter were required for internal pppApG initiation, we performed a transplant experiment (Fig. 3). This experiment analyzed the extent of internal pppApG synthesis after transplanting cRNA-like promoter elements into the model vRNA promoter one at a time. The results suggested that internal pppApG synthesis was mainly determined by the absence of a hinge in the 5' strand and the presence of a hinge and the identity of the nucleotide at position 3 in the 3' strand of the cRNA promoter. These results indicate that the spatial arrangement between the 5'- or 3'-strand hairpin-loop and the duplex region in the corkscrew configuration are important determinants controlling whether internal or terminal pppApG synthesis occurs on the vRNA and cRNA promoters.
Differences in the hinge residues between the vRNA and cRNA promoters have also been found to be essential for endonuclease cleavage activity (31) and to act as the nuclear export signals for selective packaging of vRNA rather than cRNA into the progeny virions (44). It has previously been suggested that the PB1 subunit of the polymerase might adopt a different conformation upon binding the cRNA and vRNA promoters (20). Differences in vRNA and cRNA promoter structures, in particular in hinge residues, could influence their binding properties to the influenza virus RNA polymerase, thereby altering the site of initiation of pppApG synthesis. Specifically, we propose that there might be differences in the precise binding of the vRNA and cRNA promoters, so as to initially prevent access of incoming nucleotides to positions 1 and 2 of the cRNA promoter. Such access, however, would be required for positions 1 and 2 of the vRNA promoter. Proof of this hypothesis must await the successful crystallization and X-ray analysis of the influenza virus polymerase.
To further understand the mechanism of internal initiation and realignment, an in vitro template-switching experiment was performed to test whether internally synthesized pppApG, instead of realigning to the residues 1 and 2 of a cRNA promoter (cis-realignment), could realign to the terminal nucleotides of a vRNA template (trans-realignment). The results (Fig. 6) support the trans-realignment model in vitro. However, as we have not been able to design an experiment to unambiguously prove trans-alignment in vivo, we can only speculate that both cis- and trans-realignment might occur in vivo.
A question that arose from this work was the significance (if any) of pppGpU synthesized from the model cRNA promoter (Fig. 4 and 6, bands A) and pppGpC synthesized from the model vRNA promoter (Fig. 6, band C), in addition to pppApG. The potential extension of pppGpU and pppGpC (complementary to residues 2 and 3 of the cRNA and vRNA promoter, respectively) would result in replication products lacking a 5'A end. Because the 5'A ends of both vRNA and cRNA are required for polymerase binding, replication products lacking the 5' A residue would probably not be used efficiently in subsequent rounds of transcription and replication (8, 16, 40). It is possible, however, that pppGpU and pppGpC may not form full-length transcripts as efficiently as pppApG, since pppApG was preferred over pppGpU or pppGpC for elongation to a trinucleotide under our experimental conditions (Fig. 4 and 6). Thus, the significance of the observed abortive synthesis of the dinucleotides pppGpU and pppGpC remains uncertain.
Other potential explanations of our data should be considered. For example, the in vivo data for reversion to wild type of a mutated or deleted residue at position 1 at the 3' end of the cRNA promoter (Fig. 5) could, theoretically, be explained by a model similar to that proposed for the correction of the 3'-end nucleotide of hepatitis C virus RNA (6). These authors proposed that ATP might be preferentially selected over other nucleotides and might be prebound to the RNA polymerase for initiation in a non-template-directed manner. However, the inhibitory effect of a mutation at position 1 from the 3' end of the vRNA promoter for in vitro pppApG synthesis (Fig. 2), and for in vivo activity (11) would argue against this possibility. In addition, it might be argued that our model is based, in part, on the synthesis of abortive transcripts and not on full-length replication. Abortive synthesis of short oligonucleotides has been previously observed for influenza virus (42), and is, in fact, common for both prokaryotic and eukaryotic RNA- and DNA-dependent RNA polymerases (34). These abortive transcripts, predominantly di- or trinucleotides, can then be used by polymerases to initiate new rounds of RNA synthesis. Such transcripts may significantly improve the efficiency of the RNA synthesis by circumventing the first rate-limiting nucleotidyl transfer reaction (23, 36). Therefore, we believe that the internal initiation and realignment model best fits our in vitro and in vivo data.
An internal initiation and realignment mechanism is well known for many viral RNA and DNA polymerases although the mechanisms vary in detail. In the case of protein-primed transcription, as used by poliovirus RNA polymerase, the VPg primer is uridylylated from an internal loop of the template. The VPgpUpU subsequently anneals to terminal residues for positive-stranded RNA synthesis (39) Bacteriophage 29 DNA polymerase, on the other hand, uses a sliding-back mechanism, in which protein-primed transcription starts at residue 2 and then slides back to residue 1 (35). In the case of de novo RNA synthesis, the RNA polymerases of some viruses, e.g., bunya-, arena-, and nairoviruses, generate oligonucleotides internally from the 3' end of their templates for terminal realignment. However, the mechanism differs from the prime and realign mechanism proposed here because a nontemplated C residue is added (17-19, 24, 35).
Prime and realign models have been proposed as the mechanism for repair of the 3' ends of the viral templates since the 3' ends are vulnerable to degradation by host 3' exonucleases (29). In such 3'-end repair models, as reported for turnip crinkle virus, short abortive oligonucleotides serve as a "patch" to repair one or more nucleotide deletions at the 3' end (19, 21, 25, 37). In the case of the influenza virus RNA polymerase, our in vitro (Fig. 4) and in vivo (Fig. 5) results suggest that a mutation at position 1 of the 3' strand of the cRNA promoter could revert to the wild-type sequence by the internal initiation and realignment mechanism suggested here. It is also possible that a mutation at position 1 of the vRNA promoter could be corrected to the wild-type sequence by internal initiation on the cRNA promoter, followed by trans-realignment onto the vRNA promoter. If so, initiation of pppApG from the cRNA promoter may have a selective advantage because it would allow mutations at the 3' end of either cRNA or vRNA to be repaired. This mechanism could also correct transcripts to which a nontemplated nucleotide was added at its 3' end (15).
Another interesting question is why influenza virus polymerase might employ different initiation strategies for viral RNA replication on its vRNA and cRNA promoters (Fig. 7). It is well known that, late in virus infection, more vRNA than cRNA is synthesized for packaging (13, 30). Thus, vRNA and cRNA levels are differentially regulated. Due to the fact that de novo initiation of viral RNA synthesis is likely to be the rate-limiting step for replication (25, 27, 45), we speculate that the different strategies used by influenza virus RNA polymerase for initiation of replication on the vRNA and cRNA promoters may regulate the levels of cRNA and vRNA in infected cells. Therefore, it would be interesting to compare the efficiencies of the vRNA and cRNA promoters for initiation of replication.
In summary, we now know that influenza A virus RNA polymerase uses not only different initiation mechanism for viral RNA transcription (cap-snatching initiation mechanism) and replication (de novo initiation), but also different de novo initiation strategies for viral RNA replication. For the vRNA promoter, primer-independent initiation occurs at the 3' terminus of the viral genome giving rise initially to the dinucleotide pppApG immediately followed by elongation (Fig. 7B). For the cRNA promoter, pppApG is synthesized internally from positions 4 and 5, and then realigns to the 3' terminal residues 1 and 2 for subsequent elongation (Fig. 7A). We speculate that these different strategies may have been selected in the evolution of influenza A virus in order to maintain the integrity of the viral genome and to regulate vRNA and cRNA synthesis in infected cells.
In addition, the absolute sequence conservation of nucleotides 4 and 5 in the cRNA promoter of all eight segments of all influenza A, B, and C viruses (13, 30) suggests that this internal initiation and realignment mechanism is conserved among all influenza viruses. Interestingly, the sequence of nucleotides 1 to 5 of the cRNA of orthobunyaviruses and hantaviruses (19) is identical to the cRNA of influenza virus. Thus, it would be interesting to test whether an internal initiation and realignment mechanism is also used by orthobunyaviruses and hantaviruses. Finally, our internal initiation and realignment model is critically dependent on the synthesis of the dinucleotide pppApG. Nucleotide analogues or competitive substrates for the synthesis of pppApG could be considered potential novel antiviral agents.
ACKNOWLEDGMENTS
We thank Wendy Barclay for generously providing clones of influenza virus A/Turkey/England polymerase genes used in parts of this study, Ervin Fodor for helpful discussion, and Julian Robinson for DNA sequencing.
This study was supported by MRC program grant G9523972 to G.G.B., MRC cooperative group grant G9901312, and an ORS award to T.D.
T. Deng and F. T. Vreede contributed equally to this work.
REFERENCES
Area, E., J. Martin-Benito, P. Gastaminza, E. Torreira, J. M. Valpuesta, J. L. Carrascosa, and J. Ortín. 2004. 3D structure of the influenza virus polymerase complex: localization of subunit domains. Proc. Natl. Acad. Sci. USA 101:308-313.
Azzeh, M., R. Flick, and G. Hobom. 2001. Functional analysis of the influenza A virus cRNA promoter and construction of an ambisense transcription system. Virology 289:400-410.
Ball, L. A. 2001. Replication strategies of RNA viruses, p. 105-132. In D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed., vol. 1. Lippincott Williams & Wilkins, Philadelphia, Pa.
Bouloy, M., S. J. Plotch, and R. M. Krug. 1980. Both the 7-methyl and the 2'-O-methyl groups in the cap of mRNA strongly influence its ability to act as primer for influenza virus RNA transcription. Proc. Natl. Acad. Sci. USA 77:3952-3956.
Brownlee, G. G., and J. L. Sharps. 2002. The RNA polymerase of influenza A virus is stabilized by interaction with its viral RNA promoter. J. Virol. 76:7103-7113.
Cai, Z., T. J. Liang, and G. Luo. 2004. Effects of mutations of the initiation nucleotides on hepatitis C virus RNA replication in the cell. J. Virol. 78:3633-3643.
Chen, D., and J. T. Patton. 2000. De novo synthesis of minus strand RNA by the rotavirus RNA polymerase in a cell-free system involves a novel mechanism of initiation. RNA 6:1455-1467.
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.
Deng, T., J. Sharps, E. Fodor, and G. G. Brownlee. 2005. In vitro assembly of PB2 with a PB1-PA dimer supports a new model of assembly of influenza A virus polymerase subunits into a functional trimeric complex. J. Virol. 79:8669-8674.
Fechter, P., L. Mingay, J. Sharps, A. Chambers, E. Fodor, and G. G. Brownlee. 2003. Two aromatic residues in the PB2 subunit of influenza A RNA polymerase are crucial for cap binding. J. Biol. Chem. 278:20381-20388.
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., and G. G. Brownlee. 2002. Influenza virus replication, p. 1-29. In C. W. Potter (ed.), Influenza. Elsevier Science, New York, N.Y.
Fodor, E., M. Crow, L. J. Mingay, T. Deng, J. Sharps, P. Fechter, and G. G. Brownlee. 2002. A single amino acid mutation in the PA subunit of the influenza virus RNA polymerase inhibits endonucleolytic cleavage of capped RNAs. J. Virol. 76:8989-9001.
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.
Fodor, E., B. L. Seong, and G. G. Brownlee. 1993. Photochemical cross-linking of influenza A polymerase to its virion RNA promoter defines a polymerase binding site at residue 9 to residue 12 of the promoter. J. Gen. Virol. 74:1327-1333.
Garcin, D., and D. Kolakofsky. 1990. A novel mechanism for the initiation of Tacaribe arenavirus genome replication. J. Virol. 64:6196-6203.
Garcin, D., and D. Kolakofsky. 1992. Tacaribe arenavirus RNA synthesis in vitro is primer dependent and suggests an unusual model for the initiation of genome replication. J. Virol. 66:1370-1376.
Garcin, D., M. Lezzi, M. Dobbs, R. M. Elliott, C. Schmaljohn, C. Y. Kang, and D. Kolakofsky. 1995. The 5' ends of Hantaan virus (Bunyaviridae) RNAs suggest a prime-and-realign mechanism for the initiation of RNA synthesis. J. Virol. 69:5754-5762.
González, S., and J. Ortín. 1999. Distinct regions of influenza virus PB1 polymerase subunit recognize vRNA and cRNA templates. EMBO J. 18:3767-3775.
Guan, H., and A. E. Simon. 2000. Polymerization of nontemplate bases before transcription initiation at the 3' ends of templates by an RNA-dependent RNA polymerase: an activity involved in 3' end repair of viral RNAs. Proc. Natl. Acad. Sci. USA 97:12451-12456.
Honda, A., K. Mizumoto, and A. Ishihama. 2002. Minimum molecular architectures for transcription and replication of the influenza virus. Proc. Natl. Acad. Sci. USA 99:13166-131671.
Jia, Y., and S. S. Patel. 1997. Kinetic mechanism of transcription initiation by bacteriophage T7 RNA polymerase. Biochemistry 36:4223-4232.
Jin, H., and R. M. Elliott. 1993. Non-viral sequences at the 5' ends of Dugbe nairovirus S mRNAs. J. Gen. Virol. 74:2293-2297.
Kao, C. C., P. Singh, and D. J. Ecker. 2001. De novo initiation of viral RNA-dependent RNA synthesis. Virology 287:251-260.
Kao, C. C., and J. H. Sun. 1996. Initiation of minus-strand RNA synthesis by the brome mosaic virus RNA-dependent RNA polymerase: use of oligoribonucleotide primers. J. Virol. 70:6826-6830.
Kim, M. J., W. Zhong, Z. Hong, and C. C. Kao. 2000. Template nucleotide moieties required for de novo initiation of RNA synthesis by a recombinant viral RNA-dependent RNA polymerase. J. Virol. 74:10312-10322.
Kolpashchikov, D. M., A. Honda, and A. Ishihama. 2004. Structure-function relationship of the influenza virus RNA polymerase: primer-binding site on the PB1 subunit. Biochemistry 43:5882-5887.
LaCava, J., J. Houseley, C. Saveanu, E. Petfalski, E. Thompson, A. Jacquier, and D. Tollervey. 2005. RNA degradation by the exosome is promoted by a nuclear polyadenylation complex. Cell 121:713-724.
Lamb, R. A., and R. M. Krug. 2001. Orthomyxoviridae: the viruses and their replication, p. 1487-1530. In D. M. Knipe, P. M. Howley, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. R. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Leahy, M. B., G. Zecchin, and G. G. Brownlee. 2002. Differential activation of influenza A virus endonuclease activity is dependent on multiple sequence differences between the virion RNA and cRNA promoters. J. Virol. 76:2019-2023.
Lee, M. T., K. Bishop, L. Medcalf, D. Elton, P. Digard, and L. Tiley. 2002. Definition of the minimal viral components required for the initiation of unprimed RNA synthesis by influenza virus RNA polymerase. Nucleic Acids Res. 30:429-438.
Li, X., and P. Palese. 1992. Mutational analysis of the promoter required for influenza virus virion RNA synthesis. J. Virol. 66:4331-4338.
Matsumoto, J. 1994. Evolutionary role of abortive transcript as a primer for DNA replication. J. Mol. Evol. 39:620-624.
Mendez, J., L. Blanco, J. A. Esteban, A. Bernad, and M. Salas. 1992. Initiation of 29 DNA replication occurs at the second 3' nucleotide of the linear template: a sliding-back mechanism for protein-primed DNA replication. Proc. Natl. Acad. Sci. USA 89:9579-9583.
Moroney, S. E., and J. A. Piccirilli. 1991. Abortive products as initiating nucleotides during transcription by T7 RNA polymerase. Biochemistry 30:10343-10349.
Nagy, P. D., C. D. Carpenter, and A. E. Simon. 1997. A novel 3'-end repair mechanism in an RNA virus. Proc. Natl. Acad. Sci. USA 94:1113-1118.
Paul, A. V. 2002. Possible unifying mechanism of picornavirus genome replication, p. 227-246. In B. L. Semler and E. Wimmer (ed.), Molecular biology of picornaviruses. American Society for Microbiology, Washington, D.C.
Paul, A. V., J. Yin, J. Mugavero, E. Rieder, Y. Liu, and E. Wimmer. 2003. A "slide-back" mechanism for the initiation of protein-primed RNA synthesis by the RNA polymerase of poliovirus. J. Biol. Chem. 278:43951-43960.
Pritlove, D. C., E. Fodor, B. L. Seong, and G. G. Brownlee. 1995. In vitro transcription and polymerase binding-studies of the termini of influenza A virus cRNA - evidence for a cRNA panhandle. J. Gen. Virol. 76:2205-2213.
Rao, P., W. Yuan, and R. M. Krug. 2003. Crucial role of CA cleavage sites in the cap-snatching mechanism for initiating viral mRNA synthesis. EMBO J. 22:1188-1198.
Seong, B. L., and G. G. Brownlee. 1992. Nucleotides 9 to 11 of the influenza A virion RNA promoter are crucial for activity in vitro. J. Gen. Virol. 73:3115-3124.
Szymkowiak, C., W. S. Kwan, Q. Su, T. J. Toner, A. R. Shaw, and R. Youil. 2003. Rapid method for the characterization of 3' and 5' UTRs of influenza viruses. J. Virol. Methods 107:15-20.
Tchatalbachev, S., R. Flick, and G. Hobom. 2001. The packaging signal of influenza viral RNA molecules. RNA 7:979-989.
van Dijk, A. A., E. V. Makeyev, and D. H. Bamford. 2004. Initiation of viral RNA-dependent RNA polymerization. J. Gen. Virol. 85:1077-1093.
Vreede, F. T., T. E. Jung, and G. G. Brownlee. 2004. Model suggesting that replication of influenza virus is regulated by stabilization of replicative intermediates. J. Virol. 78:9568-9572.(Tao Deng, Frank T. Vreede)