Functional Analysis of the Genomic and Antigenomic Promoters of Human Respiratory Syncytial Virus

Abstract

The promoters involved in transcription and RNA replication by respiratory syncytial virus (RSV) were examined by using a plasmid-based minireplicon system. The 3′ ends of the genome and antigenome, which, respectively, contain the 44-nucleotide (nt) leader (Le) and 155-nt trailer-complement (TrC) regions, should each contain a promoter for RNA replication. The 3′ genome end also should have the promoter for transcription. Substitution for the Le with various lengths of TrC demonstrated that the 3′-terminal 36 nt of TrC are sufficient for extensive (but not maximal) replication and that when juxtaposed with a transcription gene-start (GS) signal, this sequence was also able to direct transcription. It was also shown that the region of Le immediately preceding the GS signal of the first gene could be deleted with either no effect or with a slight decrease in transcription initiation. Thus, the TrC is competent to direct transcription even though it does not do so in nature, and the partial sequence identity it shares with the 3′ end of the genome likely represents the important elements of a conserved promoter active in both replication and transcription. Increasing the length of the introduced TrC sequence incrementally to 147 nt resulted in a fourfold increase in replication and a nearly complete inhibition of transcription. These two effects were unrelated, implying that transcription and replication are not interconvertible processes mediated by a common polymerase, but rather are independent processes. The increase in replication was specific to the TrC sequence, implying the presence of a nonessential, replication-enhancing cis-acting element. In contrast, the inhibitory effect on transcription was due solely to the altered spacing between the 3′ end of the genome and GS signal, which implies that the transcriptase recognizes the first GS signal as a promoter element. Neither the enhancement of replication nor the inhibition of transcription was due to increased base-pairing potential between the 3′ and 5′ ends. The relative strengths of the Le and TrC promoters for directing RNA synthesis were compared and found to be very similar. Thus, these findings highlighted a high degree of functional similarity between the RSV antigenomic and genomic promoters, but provided a further distinction between promoter requirements for transcription and replication.

Human respiratory syncytial virus (RSV) is a member of the family Paramyxoviridae of the order Mononegavirales, the nonsegmented negative-strand RNA viruses (35). The genome of RSV (strain A2) is 15,222 nucleotides (nt) in length and encodes 11 proteins. Three proteins are associated with the nucleocapsid: the major RNA-binding nucleocapsid N protein, the P phosphoprotein, and the major polymerase subunit L (21, 32, 42). The RSV N, P, and L proteins together with the RNA genome are the virus-specific components required for RNA replication (19, 43). Processive transcription requires, in addition, the transcription antitermination protein, M2-1 (11, 15, 20).

In some aspects of transcription and replication, RSV resembles prototype mononegaviruses such as Sendai virus (SeV) and vesicular stomatitis virus (VSV) (reviewed in references 12, 26, and 39). The genome is tightly bound by N protein to form the nucleocapsid, which is the template for the viral polymerase. The 3′ and 5′ ends of the genome consist of short extragenic leader (Le) and trailer (Tr) regions, respectively (29). Genome transcription is initiated at the genomic promoter located at the 3′ (Le) end (13) and involves a sequential stop-start mechanism in which the polymerase is guided by conserved cis-acting signals present at the ends of each gene to produce a series of subgenomic mRNAs (24). The RSV transcription signals are the 10-nt gene start (GS) and 12- to 13-nt gene end (GE) signals found at the beginning and end, respectively, of each gene (25). It is not known whether the RSV Le region is transcribed into a short positive-sense Le RNA, comparable to those of VSV and SeV.

In RNA replication, the genome is copied into a complete positive-sense encapsidated intermediate called the antigenome; hence, the 3′ and 5′ ends of the antigenome are the Tr complement (TrC) and Le complement (LeC), respectively. The antigenome is the template for the synthesis of progeny genome. In the case of RSV, the Le and TrC regions are 81% identical for the first 26 nt, after which there is no apparent relatedness (Fig. 1) (29). This likely represents a conserved promoter at the 3′ end of the genome and antigenome. It might also represent a conserved encapsidation signal at the 5′ end of these molecules, although as yet there is no evidence for this signal for RSV (33).

FIG. 1.

Sequences of the 3′ termini of RSV genomic and antigenomic RNA (Le and TrC, respectively). The first 52 nt of the TrC sequence are shown with gaps introduced to maximize sequence alignment with the Le sequence, and nucleotide assignments that are identical in Le and TrC are shown in boldface. The last 10 nt of Le are underlined, the GS signal is italicized, and the transcription initiation site is indicated with an arrow. Note that the Le contains a C residue at position 4. Both the C and G assignments have been identified at this position in biologically derived virus.

The early events in mononegavirus transcription and RNA replication are not well understood. A widely held view is that a common polymerase initiates at a single promoter, copies the Le region, and somehow commits to either stop-start transcription or readthrough replication. The availability of soluble N protein to direct cosynthetic encapsidation of the nascent positive-sense product is thought to switch the polymerase to readthrough replication (4, 23). According to this model, transcription and RNA replication are interconvertible processes. However, an alternative possibility is that transcription and replication are independent processes that involve different versions of the polymerase and/or different cis-acting initiation signals. In this case, transcription need not necessarily initiate at the 3′ genomic terminus. In this respect, there is some evidence that transcription can be initiated directly at the GS signal of the first gene of VSV (7).

While it appears that certain features of transcription and replication are shared among the mononegaviruses, RSV has its own distinct features. For example, processive transcription requires an antitermination factor, the M2-1 protein, which is not found in most other mononegaviruses. Two additional proteins, NS1 and M2-2, which exist only for the Pneumovirus genus, have been implicated in regulating RNA synthesis (2, 3, 22). Deletion of the M2-2 gene reduces RNA replication and augments RNA transcription. Unlike other paramyxoviruses, RSV replication does not require the genome nucleotide length to be a multiple of 6 (36). Furthermore, the cis-acting signals involved in initiation of replication and transcription appear to be mostly or entirely confined to the extragenic regions and the first GS signal (10, 24), whereas for other paramyxoviruses, these signals extend into the adjacent genes (30, 38).

The present study investigates the cis-acting sequences involved in RSV transcription and RNA replication and, in particular, examines and compares the genomic and antigenomic promoters contained in the Le and TrC regions, respectively.

MATERIALS AND METHODS

cDNAs.

Minigenome plasmids C41, C2, and 2G have been described previously (15, 19, 33). Each encoded minigenome contains in 3′-to-5′ order the 44-nt RSV Le, the 10-nt NS1 GS signal, the upstream 29 nt of the nontranslated region of the NS1 gene, a 669-nt negative-sense copy of the chloramphenicol acetyltransferase (CAT) open reading frame (ORF), the last 12 nt of the nontranslated region of the L gene, the 12-nt L GE signal, and the 155-nt Tr region. Minigenome 2G differs in that it contains a C-to-G mutation (negative sense) at position 2 relative to the 5′ end of the Tr. Each cDNA is bordered at the 5′ end relative to the encoded minigenome by three G residues and the T7 RNA polymerase promoter (the G residues improve efficiency of initiation by the T7 RNA polymerase) and at the 3′ end with a self-cleaving ribozyme: the hammerhead ribozyme in C2 and 2G (19) and the hepatitis delta virus ribozyme in C41 (34). Plasmid C4 contains minigenome C2 in the opposite orientation with respect to the ribozyme and T7 RNA polymerase promoter and so encodes a positive-sense miniantigenome. Minigenome plasmids A36 to A147 were prepared by PCR-amplifying portions of the RSV Tr sequence with C41 plasmid as a template. The negative-sense primer contained a portion of the hepatitis delta ribozyme sequence, including an RsrII site, and hybridized to the end of the Tr region. The positive-sense primer contained a BstXI site and hybridized within the Tr region to generate Tr fragments of 36, 57, 77, 97, 117, or 147 nt. The PCR products were digested with RsrII and BstXI and inserted into the RsrII-BstXI window of minigenome C41, which contains an RsrII site within the ribozyme and a naturally occurring BstXI site within the Le (the BstXI recognition sequence spans nt 35 to 46). Plasmids B36 to B147 were constructed by the method of Byrappa et al. (5): minigenome plasmids A36 to A147 were used as templates for PCR amplification by using a positive-sense phosphorylated primer whose 5′ end lay 42 nt from the Tr terminus and a negative-sense phosphorylated primer whose 5′ end lay at the end of the L GE signal. The PCR product was gel purified and ligated. Plasmids C61, C81, and C101 were prepared by inserting oligonucleotide duplexes into the BstXI site of plasmid C2, resulting in heterologous insertions of 17, 37, or 57 nt. These insertions regenerated the end of the Le such that the heterologous sequence was placed between the Le and the NS1 GS signal. Each insert included a unique AflII site. Minigenome plasmids C121 and C151 were prepared by inserting 20 and 50 nt into the AflII site within C101 by using oligonucleotide duplexes consisting of sequence randomly chosen from the RSV N gene. Plasmids D36 to D97 were generated by using A36 to A97 as templates in a PCR (5), a phosphorylated positive-sense primer whose 5′ end lay at the first G residue of the NS1 GS signal, and a phosphorylated negative-sense primer whose 5′ end lay at position 34 of the Le. Plasmids E36 to E97 were constructed by transferring the Tr region of plasmid 2G into plasmids A36 to A97 by using NcoI, which restricts within the CAT ORF, and HindIII, which restricts between the plasmid backbone and the T7 RNA promoter. Minigenome plasmid F1 was generated from plasmid C41 by PCR mutagenesis (5) to insert a hammerhead ribozyme (1) and a sequence GGGACGG, which allows optimal transcription by the T7 RNA polymerase, between the Tr and the T7 RNA polymerase promoter. Minigenome F3 was constructed in a similar manner with a version of C41 that contains a G rather than a C at position 4 of the Le (negative sense). Minigenome plasmids F2 and F4 were prepared from F1 and F3, respectively, by using PCR to replace the Tr region with LeC sequence. The LeC sequence contained a G at position 4 relative to the 3′ end of the antigenome.

Transfections.

Monolayers of HEp-2 cells in six-well dishes were simultaneously infected with 10 PFU (per cell) of vaccinia virus vTF7-3 (provided by Thomas Fuerst and Bernard Moss), which expresses the T7 RNA polymerase (18), and transfected with the following mixture of plasmids per well of a six-well dish: 0.2 μg of minigenome DNA, 0.4 μg of pTM1 N, 0.2 μg of pTM1 P, 0.1 μg of pTM1 M2-1, and 0.1 μg of pTM1 L (Fig. 2 to 7) or 0.2 μg of minigenome DNA, 0.4 μg of pTM1 N, 0.2 μg of pTM1 P, and 0.1 μg of pTM1 L (Fig. 8), as described previously (19). Control transfections lacking L or all support plasmids received pTM1 plasmid with no insert so that the amount of transfected DNA was equivalent in each well. Twenty-four hours later, the transfection-infection mixture was replaced with OptiMem containing 2% fetal bovine serum and actinomycin D (Calbiochem) at 2 μg/ml. The actinomycin D-containing medium was removed after 2 h, replaced with fresh OptiMem containing 2% fetal bovine serum, and incubated for a further 24 h. Each transfection reaction was set up in duplicate; RNA was directly extracted from cells from one of the wells, and the cells in the other well were lysed with nonionic detergent and incubated with micrococcal nuclease (MCN) prior to RNA purification to digest unencapsidated RNA, as described previously (14).

FIG. 2.

Effect of replacing the 3′ 34 nt of Le with increasing lengths of TrC. (A) Structures (not to scale) of minigenome C41, representing wild-type RSV, and minigenomes A36 to A147, which are designated according to the length of the added TrC sequence. GS or GE signals are indicated with open or solid boxes, respectively. (B and C) Northern blots of positive-sense RNAs synthesized by the reconstituted RSV polymerase. HEp-2 cells were infected with vaccinia virus vTF7-3 and simultaneously transfected with plasmids that encode minigenome C41 (lanes 1, 2, and 3) or minigenomes A36 to A147 (lanes 4 to 9, as indicated) together with pTM1 support plasmids expressing N, P, and M2-1 (lane 2) or N, P, M2-1, and L (lanes 3 to 9) proteins. Lane 1 received empty pTM1 expression plasmid. Forty-eight hours later, the cells were processed directly for RNA purification (B), or lysates were prepared and treated with MCN to destroy unencapsidated RNA (C). The blots were hybridized with a negative-sense, CAT-specific riboprobe. (D and E) Northern blot analyses of plasmid-supplied minigenome template. HEp-2 cells were infected with vTF7-3 and transfected with plasmids that encode minigenome C41 (lanes 1 and 2) or minigenomes A36 to A147 (lanes 3 to 8, as indicated) together with plasmids expressing N, P, and M2-1 proteins (lanes 2 to 8) or with empty pTM1 plasmid (lane 1). L plasmid was omitted from all reactions, and hence the only source of minigenome RNA was plasmid. Forty-eight hours later, the cells were processed directly for RNA purification (D) or lysed and treated with MCN followed by RNA purification (E). The RNAs were detected with a positive-sense, CAT-specific riboprobe.

FIG. 7.

Analysis of the promoter strength of Le versus TrC under conditions in which the minigenome template was not amplified. (A) Diagram (not to scale) of a series of minigenomes, E36 to E77, in which the 3′ 34 nt of Le have been replaced by the indicated length of TrC sequence (as in series A, Fig. 2) and the penultimate nucleotide of the Tr region has been changed from a C to a G residue. (B) Northern blot analysis of positive-sense RNAs synthesized from minigenome 2G, which contains the complete Le (lanes 1 and 2), or minigenomes E36 to E77 (lanes 3 to 5). For comparison, the same nucleotide substitution was introduced into minigenome C41 containing the wild-type Le (Fig. 2), creating minigenome 2G. Cells were transfected with plasmids expressing the indicated minigenome and plasmids expressing N, P, and M2-1 (lane 1) or N, P, M2-1, and L (lanes 2 to 5). The RNAs were analyzed by Northern blotting with a negative-sense, CAT-specific riboprobe. The total amount of RNA in each lane is indicated in arbitrary PhosphorImager units.

FIG. 8.

Transcription and replication of minigenomes in which the Tr was replaced by the complement of the Le (LeC). (A) Structures (not to scale) of minigenomes F1 to F4. Each minigenome contains a 44-nt Le at the 3′ terminus, minigenomes F1 and F3 contain the complete 155-nt Tr at the 5′ terminus, and minigenomes F2 and F4 contain the 44-nt LeC sequence at the 5′ terminus. The 3′ Le sequence of minigenomes F1 and F2 differs from that of F3 and F4 by a single nucleotide substitution at position 4; hence, both naturally occurring assignments are represented. (B and C) Northern blot analyses of positive-sense RNAs synthesized by the reconstituted RSV polymerase. Parallel wells of cells were transfected with minigenome F1 (lanes 1 to 3), F2 (lanes 4 to 6), F3 (lanes 7 to 9), or F4 (lanes 10 to 12) together with plasmids expressing N and P (lanes 2, 5, 8, and 11) or N, P, and L (lanes 3, 6, 9, and 12). Lanes 1, 4, 7, and 10 received empty pTM1 expression plasmid. The blots show total (B) or MCN-resistant (C) RNA detected by hybridization with a negative-sense CAT probe. (D) Northern blot analysis of MCN-resistant minigenome RNA, detected by hybridization with a positive-sense CAT probe.

RNA isolation, oligo(dT) chromatography, and Northern blot hybridization.

RNA was extracted by dissolving cell pellets or MCN-treated cell lysates in Trizol reagent (Life Technologies) according to the supplier's protocol, except that the RNAs were extracted with phenol-chloroform and ethanol precipitated after the isopropanol precipitation. Oligo(dT) chromatography was performed with an Oligotex mRNA mini kit (Qiagen) according to the manufacturer's instructions, except that the RNA was denatured by being heated to 95°C prior to addition of the Oligotex suspension. RNA representing 1/10 of one well of cells was analyzed by electrophoresis in a 1.5% agarose gel containing 0.44 M formaldehyde, transferred to nitrocellulose (Schleicher & Schuell), and fixed by UV cross-linking (Stratagene). Negative-sense or positive-sense ³²P-labelled CAT-specific riboprobe was synthesized by T7 RNA polymerase from XbaI-digested C2 cDNA or NcoI-digested C4 cDNA, respectively and hybridized to the Northern blot in a mixture of 6× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 5× Denhardt's solution, 0.5% sodium dodecyl sulfate (SDS), and 200 μg of sheared DNA per ml at 65°C for 12 h. The blots were washed in 2× SSC–0.1% SDS at room temperature for 30 min and then at 65°C for 2 h and then in 0.1× SSC–0.1% SDS at 65°C for 15 to 30 min. The blot shown in Fig. 4D was hybridized with a 5′ ³²P-labelled, negative-sense, Le-specific oligonucleotide probe (5′-GGTTTATGCAAGTTTGTTGTACGCATTTTTTCCCGT) in a solution of 6× SSC, 5× Denhardt's solution, 0.1% SDS, 0.05% sodium pyrophosphate at 52°C for 12 h. The blot was washed in 6× SSC for 30 min. PhosphorImager analysis was carried out with a PhosphorImager 445 SI (Molecular Dynamics).

FIG. 4.

Insertion of a heterologous spacer into the Le region has differentiated effects on transcription and replication. (A) Structures (not to scale) of the C series of minigenomes, in which a spacer sequence was inserted at the indicated position in the Le region, resulting in Le regions of 61 to 151 nt (C61 to C151, designated according to the final length of the Le). (B, C, and D) Northern blot analyses of positive-sense RNAs synthesized by the RSV polymerase. Cells were transfected with plasmids expressing minigenome C2 (lanes 1 and 2), minigenomes C61 to C151 (lanes 3 to 7, as indicated), or minigenome A147 (lane 8), together with plasmids expressing N, P, and M2-1 (lane 1) or N, P, M2-1, and L (lanes 2 to 8). Panels B and D show total intracellular RNA, and panel C shows MCN-resistant intracellular RNA. The RNAs were detected with a negative-sense, CAT-specific riboprobe (B and C) or negative-sense, Le-specific oligonucleotide probe (D).

Primer extension analysis of RNA.

Oligo(dT)-purified or total RNA representing one-third to one-half of a well of cells or 5 pmol of RNA transcribed in vitro from plasmid C4 by T7 RNA polymerase was annealed to an excess of 5′ ³²P-labelled, negative-sense, CAT-specific oligonucleotide probe (5′-GGGATATATCAACGGTGGTATATCCAGTG) in 1× SuperScript II buffer (Life Technologies) by heating the mixture to 95°C for 5 min and placing it at room temperature for 15 min. One-half of the RNA-DNA hybrid was utilized as a template in a reverse transcriptase reaction using SuperScript II reverse transcriptase (Life Technologies) according to the supplier's reaction conditions, except that the reaction was carried out at 37°C for 1 h. The cDNA was extracted with phenol-chloroform and precipitated with ethanol and resuspended in 20 μl of a mixture of 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, and 0.05% xylene cyanol. Ten microliters of this cDNA was electrophoresed on a 5% LongRanger (J. T. Baker) polyacrylamide gel and analyzed by autoradiography and phosphorimaging. As a molecular length marker, a dideoxy C sequencing reaction was carried out with the same oligonucleotide as a primer and with C41 plasmid as a template for T7 Sequenase (Amersham).

RESULTS

Effects of replacing Le with TrC.

RNA synthesis from the RSV genomic and antigenomic promoters was compared by using a plasmid-based, intracellular system in which a genome analog (minigenome) containing a CAT reporter gene is coexpressed with RSV nucleocapsid and polymerase proteins. The RNAs synthesized by the reconstituted RSV polymerase are detected by Northern blotting. Minigenome C41 represents the wild-type genome and contains at its 3′ end the first 86 nt of the RSV genome, including the 44-nt Le, the NS1 GS signal, and the nontranslated region of the NS1 gene, and at its 5′ terminus, the last 179 nt of the genome, including the nontranslated region of the L gene, the L GE signal, and the 155-nt Tr (Fig. 2A). To examine the promoter activity of the TrC region, we constructed a series of minigenomes in which the first 34 nt of Le sequence were replaced with various lengths of TrC sequence ranging from 36 to 147 nt (minigenomes A36 to A147). The 10 nt that lie immediately upstream of the NS1 GS signal were left intact to avoid disruption of possible cis-acting elements preceding this GS signal (Fig. 2A).

Plasmids encoding minigenome C41 or A36 to A147 were transfected together with support plasmids into cells which had been infected with a vaccinia virus encoding T7 RNA polymerase to drive plasmid expression. Total intracellular RNA was harvested at 48 h and analyzed by Northern blot hybridization with a negative-sense riboprobe to detect antigenome and mRNA generated by the RSV polymerase. To further distinguish the encapsidated antigenome, cell extracts from duplicate transfection wells were treated with MCN prior to RNA purification. As described previously (15), minigenome C41 generated a large amount of mRNA and a small amount of antigenomic RNA (Fig. 2B and C, lane 3). Similarly minigenome A36 generated both antigenome and mRNA (Fig. 2B and C, lane 4). Thus the first 36 nt of TrC contain a promoter that can direct transcription in addition to replication when juxtaposed with the Le-NS1 junction sequence. As the length of TrC sequence was increased (minigenomes A57 to A147), the level of antigenome increased, although minigenomes containing more than 97 nt of TrC sequence yielded somewhat less than the maximal level (Fig. 2B and C). In contrast, mRNA synthesis decreased significantly, such that the mRNA synthesized from minigenomes A97 to A147 was barely detectable (Fig. 2B, lanes 7 to 9). Thus, increasing the amount of TrC sequence augmented replication and reduced transcription.

It was important to confirm that the level of plasmid-supplied minigenome RNA was equivalent in each transfection reaction. Therefore parallel transfections were carried out in the absence of the L expression plasmid, and the accumulation of plasmid-supplied minigenome was analyzed by Northern blotting with a positive-sense probe. This analysis showed that a similar amount of plasmid-supplied minigenome RNA accumulated in each transfection (Fig. 2D). However, examination of MCN-treated RNA showed that there was a progressive change in the pattern of MCN-resistant input minigenome with increasing length of added TrC sequence (Fig. 2E, lanes 3 to 8). Minigenomes C41, A36, and A57 could be observed as abundant single bands (Fig. 2E, lanes 2 to 4), but minigenomes A77 to A147 appeared as less-abundant multiple bands, mostly of slightly reduced length (Fig. 2E, lanes 5 to 8). This suggested that one or both termini of these RNA molecules were incompletely encapsidated and hence subject to digestion by MCN and that encapsidation overall was much less efficient. The results presented below indicated that this was due to the high degree of terminal complementarity of these minigenomes, which probably promoted panhandle structure formation and inhibited encapsidation of the termini. The relatively lower levels of full-length, encapsidated minigenome generated from plasmids A97 to A147 would account for the relatively lower levels of antigenome produced in these transfections, as described above (Fig. 2B and C).

The amounts of antigenome and mRNA generated from the minigenomes that were encapsidated efficiently (C41, A36, and A57) were quantitated by PhosphorImager analysis. This analysis showed that minigenome A36 produced amounts of antigenome and mRNA similar to those of minigenome C41, whereas minigenome A57 generated approximately twofold more antigenome and one-third less mRNA.

TrC nt 77 to 97 enhance replication independently of increased terminal complementarity.

Increasing the length of TrC sequence used to replace Le might have affected the pattern of positive-sense RNA synthesis either directly, as a consequence of its primary sequence or increased length compared to that of Le, or indirectly, by increasing the degree of terminal complementarity and potential interaction between the genome ends, or both. Terminal complementarity has been shown to affect transcription and replication by VSV (40, 41). To distinguish between these possibilities, minigenomes were constructed that were similar to those described above, but contained only the 5′-proximal 42 nt of Tr sequence (B series of minigenomes [Fig. 3A]). Thus, these minigenomes have increasing amounts of TrC sequence at the 3′ terminus, but share the same degree of terminal complementarity.

FIG. 3.

Effect of replacing Le with increasing lengths of TrC under conditions in which terminal complementarity is limited to 42 nt. (A) Structure of a series of minigenomes, B36 to B147, in which the 3′ 34 nt of Le have been replaced by the indicated length of TrC sequence (as in minigenome series A shown in Fig. 2) and the Tr has been truncated to the 5′ 42 nt, thus limiting terminal complementarity. (B and C) Northern blots of positive-sense RNAs synthesized by the reconstituted RSV polymerase. Parallel wells of cells were transfected with plasmid C41 (lanes 1 to 3) or minigenomes B36 to B147 (lanes 4 to 9, as indicated) together with plasmids expressing N, P, and M2-1 (lane 2) or N, P, M2-1, and L (lanes 3 to 9). Lane 1 received empty pTM1 expression plasmid. The blots show total (B) or MCN-resistant (C) RNA, detected by hybridization with a negative-sense CAT probe. (D and E) Northern blot analyses of plasmid-derived minigenome template. Cells were transfected with plasmid C41 (lanes 1 and 2) or plasmids encoding minigenomes B36 to B147 (lanes 3 to 8, as indicated) without support plasmids (lane 1) or with N, P, and M2-1 plasmids (lanes 2 to 8). The blots show total (C) or MCN-resistant (D) RNA, detected by hybridization with a positive-sense CAT probe.

Figure 3B and C show Northern blots of total and MCN-resistant, positive-sense RNA generated by the RSV polymerase. Note that in panel B, lane 5 is underloaded in this particular experiment, as was confirmed by repeat experiments. This analysis showed that, as with minigenomes containing complete Tr, antigenome levels increased and mRNA levels decreased with increasing length of TrC. For minigenomes B57 to B147, these effects cannot be attributed to increasing terminal complementarity and therefore must be due to the primary sequence or increased length of the introduced TrC segment.

Figure 3D and E are Northern blots of the control transfections showing the accumulation of total and MCN-resistant, plasmid-supplied minigenome, respectively. In contrast to the situation seen with minigenomes containing complete Tr, each minigenome was observed as a single, discrete band following MCN treatment, indicating that each of these minigenomes was completely encapsidated. This result indicated that it was the high level of terminal complementarity of minigenomes A77 to A147 that inhibited encapsidation (Fig. 2E).

Because each transfection received completely encapsidated minigenome, this experiment allowed us to measure the effect of different lengths of TrC sequence on antigenome accumulation. The RNA bands in panel C were quantitated by using a PhosphorImager and adjusted according to the values for panel E to account for the minor variation in input encapsidated RNA. This analysis showed that increasing the length of TrC from 36 to 97 nt augmented antigenome levels approximately fourfold and that increasing the length of TrC from 97 to 147 nt caused a further minor increase in antigenome synthesis.

The effect of TrC on RNA replication is independent of its effect on transcription.

To determine if the effects of increasing TrC were sequence or spacing dependent, minigenomes were constructed in which a spacer sequence was inserted at the end of the Le region, such that the length of the 3′ extragenic region was similar to that in the minigenomes used in Fig. 2 and 3 (C series, Fig. 4A). These minigenomes differ slightly from those used in the experiments described above, because their 3′ terminus is generated by a hammerhead ribozyme rather than the hepatitis delta virus ribozyme, a technical point which would not influence the results. However, so that the minigenome backbones within the experiment were consistent, minigenome C2 was used as a positive control instead of minigenome C41.

Analysis of the positive-sense RNAs generated from these minigenomes showed that, similar to the findings with increasing the length of TrC, increasing the length of the 3′ extragenic region caused a decrease in mRNA synthesis, particularly if the length was increased above 101 nt (Fig. 4B). This result indicates that the inhibition of transcription due to increasing the length of TrC is not sequence specific, but rather is an effect of increased length of the 3′ extragenic region.

Increasing the Le length from the wild-type 44 nt to 61 nt (compare lanes 2 and 3) caused a slight increase in the level of antigenome; however, further increases in Le length had no significant effect on replication (Fig. 4C and D). This contrasted with the findings for minigenomes containing imported TrC (Fig. 2 and 3). Thus, the increase in antigenome synthesis caused by increasing the length of TrC was specific to the TrC sequence. To confirm that the lack of increase in antigenome was not due to suboptimal replication conditions (e.g., insufficient soluble N protein), a parallel reaction was carried out with minigenome A147 (Fig. 4B and C, lane 8). This minigenome produced significantly more antigenome than any of the C series of minigenomes, demonstrating that conditions were not restrictive for replication.

A subgenomic RNA species of greater size than monocistronic CAT mRNA was generated from the C series of minigenomes (Fig. 4B). Northern blotting with an oligonucleotide probe specific for the positive-sense Le transcript (Fig. 4D) identified this species as a Le-CAT readthrough mRNA, which had been described previously both with minigenomes and with authentic RSV infection (9, 24). The Le-CAT mRNA levels did not diminish with increasing Le length and thus paralleled the antigenome levels. Le-CAT mRNA was not detected in samples treated with MCN, indicating that it was not encapsidated (Fig. 4C).

The 10 nt of Le that immediately precede the GS signal are not essential for accurate transcription initiation.

As described above, TrC sequence juxtaposed to the last 10 nt of Le and the GS signal directed transcription. To examine if the 10 Le nt are necessary for transcription to occur, we compared minigenomes in which they were deleted (minigenomes D36 to D97) to minigenomes A36 to A97.

Minigenomes which contained 57, 77, or 97 nt of TrC sequence yielded similar amounts of mRNA and antigenome, irrespective of the presence of the 10 Le nt (Fig. 5B and C, compare lanes 3, 4, and 5 to lanes 8, 9, and 10). This demonstrated that these nucleotides are not required for transcription or RNA replication. However, minigenomes that contained 36 nt of TrC sequence produced significantly less mRNA and slightly more antigenome if the last 10 nt of Le were deleted, suggesting that this region does play a minor role in regulating transcription and/or replication.

FIG. 5.

The last 10 nt of Le (nt 34 to 44) are dispensable for transcription and RNA replication. (A) Structures of minigenomes A36 to A97, as described in Fig. 2, and their derivatives D36 to D97, from which the last 10 nt of the Le region between the TrC sequence and the GS signal have been deleted. (B and C) Northern blot analyses of positive-sense RNAs synthesized from minigenomes A36 (lanes 1 and 2), A57 to A97 (lanes 3 to 5), D36 (lanes 6 and 7), or D57 to D97 (lanes 8 to 10). Cells were transfected with the indicated minigenome together with plasmids expressing N, P, and M2-1 (lanes 1 and 6) or N, P, M2-1, and L (lanes 2 to 5 and 7 to 10). Panel B shows total intracellular RNA, and panel C shows MCN-resistant intracellular RNA. The RNAs were detected with a negative-sense, CAT-specific riboprobe.

It was important to confirm that the mRNA synthesized in the absence of the 10 nt of Le was initiated correctly at the NS1 GS signal. Polyadenylated RNA, isolated from the total RNA shown in Fig. 5B, lanes 2, 3, 7, and 8, was analyzed by primer extension (Fig. 6) to determine if it was initiated at the NS1 GS signal. Control reactions were carried out with total RNA derived from minigenomes A36 and D36 and antigenome RNA synthesized in vitro by T7 RNA polymerase to indicate the origin of the antigenome RNAs generated from these minigenomes (Fig. 6, lanes 2, 3, and 9).

FIG. 6.

The last 10 nt of Le (nt 34 to 44) are not necessary for correct transcription initiation at the NS1 GS signal. Primer extensions were carried out by using templates of total RNA derived from transfections with minigenome A36 (lane 2) or D36 (lane 3) or oligo(dT)-purified RNA derived from transfections with minigenome A36 (lanes 4 and 5), A57 (lane 6), D36 (lane 7), or D57 (lane 8). Lane 4 is a negative control in which the transfection reaction mixture did not contain L plasmid. Lane 9 is a positive control using miniantigenomic RNA synthesized from C4 plasmid in vitro by T7 RNA polymerase. The size of this primer extension product is 1 nt longer than that of A36, consistent with its predicted structure. Lane 1 is a ddC sequencing reaction carried out with minigenome C41 as a template; the position of the first 4 nt (CCCC) of the NS1 GS signal is indicated. Solid arrowheads indicate the positions of cDNAs generated from RNAs initiated at the minigenome 3′ terminus, which are barely detectable in lanes 5 to 8, and a large open arrowhead indicates the position of cDNAs generated from RNAs initiated at the NS1 GS signal.

This analysis showed that almost all of the polyadenylated RNA synthesized from minigenomes D36 and D57 was initiated at the NS1 GS signal (indicated by an open arrowhead), similarly to the polyadenylated RNA derived from minigenomes A36 and A57 (compare lanes 7 and 8 to 5 and 6), and only a barely detectable amount was initiated at the genome 3′ end (Le-CAT readthrough mRNA, indicated by solid arrowheads). This result demonstrates that the last 10 nt of Le are not required for accurate transcription initiation at the NS1 GS signal.

Comparison of promoter strength of TrC to that of Le.

It was of interest to directly compare the strengths of the genomic promoter in Le and the antigenomic promoter contained in TrC. This could not be done reliably in the preceding experiments because they employed minigenomes which were competent for amplification by the reconstituted RSV polymerase. Specifically, the miniantigenome which is produced serves in turn as a template to produce progeny minigenome. As we have described previously (14, 33), this amplifies the plasmid-supplied minigenome template 5- to 50-fold, depending on the efficiency of reconstituted replication in any particular experiment. Thus, any mutation which affects the efficiency of antigenome synthesis can drastically affect the level of minigenome template, complicating evaluation of mutations. This problem can be overcome by blocking amplification by introducing one of several point mutations into the Tr region (33), one example being a C-to-G substitution at the penultimate nucleotide (Fig. 7A). This mutation does not significantly affect encapsidation or template activity of the plasmid-supplied minigenome, but the miniantigenome it encodes is inactive as a template for the progeny minigenome.

Figure 7B shows mRNA and antigenome generated from minigenomes containing either Le or 36 to 77 nt of TrC and Tr containing the 2G mutation. Control reactions demonstrated that each reaction received a similar amount of MCN-resistant minigenome RNA (data not shown). This analysis demonstrated that the minigenome containing 36 nt of TrC generated slightly more antigenome (less than twofold) than the minigenome containing Le. However, the two minigenomes generated essentially the same amount of total positive-sense RNA, since the antigenome comprised only a small fraction. As the length of TrC was increased to 57 and 77 nt, the total amount of RNA synthesized decreased due to a reduction in mRNA transcription. There was a minor increase in antigenome levels associated with the increase in TrC length from 36 to 57 nt, but this was not reciprocal to the decrease in transcription. Thus, these data show that the promoters contained within the 3′-terminal regions of TrC and Le direct essentially the same amount of total positive-sense RNA synthesis.

Promoter activity of Le under conditions of enhanced terminal complementarity.

Previously it had been shown that increasing the complementarity within the terminal 50 nt of VSV augments replication and inhibits transcription (40, 41). Although the result shown in Fig. 3 addressed the role of terminal complementarity beyond 42 nt, this experiment did not examine the importance of complementarity within the terminal 42 nt. To examine this, minigenome C41 was modified by replacing the Tr region with the 44-nt complement of Le (LeC), creating minigenome F2 (Fig. 8A). This increased the amount of terminal complementarity from 27 of 44 nt (61%) to 43 of 44 nt (98%), with the single mismatch being at position 4 (see below). We also placed a ribozyme between the T7 promoter and the 5′ end of the minigenome so that both ends of the minigenome were generated by self-cleaving ribozymes, which leave correct ends. This made it possible to place the T7 promoter in an optimal sequence context for efficient T7-mediated RNA synthesis and obviated effects on T7 promoter efficiency due to changes in the minigenome 5′ end (33). This modification precluded the need for nonviral G residues at the 5′ end of the minigenome. A version of C41 containing the second ribozyme was constructed and designated minigenome F1 (Fig. 8A).

The F1 and F2 minigenomes were complemented with the N, P, and L plasmids, and the synthesis of positive-sense and negative-sense RNA was monitored. The two minigenomes expressed similar amounts of positive-sense RNA (Fig. 8B, lanes 3 and 6). Because the transfections did not contain M2-1, some of the mRNA was truncated and migrated as a smear below the full-length mRNA. Parallel samples which were treated with MCN prior to RNA purification confirmed the synthesis of comparable amounts of encapsidated miniantigenome (Fig. 8C, lanes 3 and 6). Encapsidated negative-sense RNA also was analyzed, which showed that F1 synthesized slightly more minigenome than F2 (Fig. 8D, lanes 3 and 6). In comparison, very little minigenome was detected when L plasmid was omitted (Fig. 8D, lanes 2 and 5), which showed that reconstituted RSV replication was very efficient and was responsible for most of the encapsidated intracellular minigenome. Thus, increasing the amount of terminal complementarity did not dramatically affect the activities of either the genomic or antigenomic promoters.

Position 4 in the Le has two naturally occurring assignments. The 4C assignment (negative sense), which was present in all of the preceding minigenomes used in this paper, results in increased antigenome synthesis in the minireplicon system (M.E.P., R.F., and P.L.C., unpublished data, and see below). The 4G assignment is more common in biologically derived viruses. We constructed minigenomes F3 and F4, which were the same as minigenomes F1 and F2, except that the Le contained the G assignment at position 4 (Fig. 8A). Minigenome F4 thus had 100% complementarity in the terminal 44 nt. Minigenome F3 synthesized slightly more positive-sense RNA and encapsidated miniantigenome and substantially more encapsidated minigenome than did F4 (Fig. 8B, C, and D, compare lanes 9 and 12). Thus, the increased complementarity did not augment synthesis. The reduced amount of antigenome synthesized by F3 and F4 (Fig. 8C, lanes 9 and 12) compared to that synthesized by F1 and F2 (lanes 3 and 6) is due to the 4C mutation.

DISCUSSION

The control of transcription and replication is a central, unresolved issue in the molecular biology of mononegaviruses. In this study, we compared the genomic and antigenomic promoters of RSV to distinguish the cis-acting requirements for replication and transcription. The promoter contained within TrC was shown to resemble that of Le in being able to direct efficient transcription in addition to replication, when juxtaposed with a GS signal (Fig. 2, 3, 5, 6, and 7). The two promoters also were essentially identical with regard to promoter strength (Fig. 7). The sequence similarity between Le and TrC lies at the 3′ termini, which share 81% nucleotide identity for the first 26 nt, after which there is no significant similarity (Fig. 1). It is likely that the 21 conserved nt include the important elements of a functionally conserved promoter, with the caveat that transcription requires, in addition, a downstream GS signal (24). The remainder of the Le and TrC appeared to have sequences which modified these activities, but were not essential and had an impact that was on the order of only several-fold. In addition, we showed that Le-specific sequence is not required for accurate transcription initiation at the first GS signal (Fig. 6).

These findings differ in some ways from those described for model mononegaviruses. For example, the TrC of SeV can direct transcription (6), as described here for RSV, but that of VSV apparently does not (41). Furthermore, both of these model viruses appear to have Le sequence essential for transcription located immediately before the first GS signal (6, 28, 41), whereas there is no evidence for such a signal here for RSV.

Given the greater intracellular accumulation of genome compared to antigenome in cells infected with this group of viruses, it had been suggested that the antigenomic promoter is substantially more powerful than the genomic promoter. Direct comparison of the strengths of the genomic and antigenomic promoters, in a situation in which slight differences would not be exaggerated by genome amplification, showed that the RSV genomic and antigenomic core promoters directed similar amounts of positive-sense RNA synthesis (Fig. 7). In this experiment, we were comparing TrC with an Le that contained a C residue at position 4. Since there are two naturally occurring assignments at this position, we have also used the 2G minigenome backbone to compare the promoter strengths of these two Le sequences and found that a minigenome containing a C residue synthesizes twofold more antigenome RNA than a minigenome containing a G at this position (M.E.P., R.F., and P.L.C., unpublished observations), an effect that seems small, but apparently increases exponentially under conditions permissive for template amplification.

Increasing the length of inserted TrC sequence increased replication and decreased transcription. The effect on replication could not be duplicated with heterologous, non-TrC sequence (Fig. 4), indicating that it depended on a specific TrC sequence element. This appeared to lie between nt 36 and 97, but was not further defined (Fig. 3). Similarly it has been shown for VSV, SeV, and rabies virus that the primary sequence of the TrC region is more efficient at directing replication than the Le (6, 16, 17, 27, 37). There are several possible mechanisms by which sequences within TrC could enhance replication: (i) expediting polymerase binding to the promoter and/or initiation of RNA synthesis; (ii) promoting encapsidation of the nascent RNA, which could facilitate replication processivity; (iii) stabilizing naked, nascent RNA, which might be important if encapsidation lags behind RNA synthesis.

The downregulatory effect on transcription caused by increasing lengths of TrC sequence was due to the increased length of the 3′ extragenic region. This was demonstrated by inserting nonspecific spacer sequence into the Le region; mRNA synthesis decreased significantly as the length of spacer was increased (Fig. 4), similar to the situation seen with increasing length of TrC. This result suggests that efficient transcription depends on appropriate spacing between important cis-acting elements, presumably the 3′ element and the first GS signal.

The changes in transcription and replication caused by substituting for Le with TrC sequence could not be attributed to the increasing degree of terminal complementarity, because minigenomes in which terminal complementarity was consistent behaved similarly to minigenomes in which terminal complementarity was increased (compare Fig. 2 and 3). Furthermore, comparison of minigenomes containing Tr or LeC sequence at the 5′ terminus indicated that the degree of terminal complementarity had no discernible effect on transcription or replication (Fig. 8). These results are consistent with findings for SeV, which indicated that increasing terminal complementarity did not affect replication (37), but contrast with observations made for VSV (40, 41), which indicated that increasing terminal complementarity augmented replication at the expense of transcription.

In these experiments, as well as others described previously (24, 25), there was no evidence of a direct inverse relationship between transcription and RNA replication, as would be expected if these two processes were in balance (e.g., Fig. 7). Although increasing length of the TrC increased replication and decreased transcription, these effects appeared to be coincidental and unrelated. Previous studies of other mononegaviruses have not clarified whether transcription and replication are competitive; some experiments indicated that the ability of a minigenome to direct transcription was not related to its replication efficiency (6, 28), but other experiments indicated an inverse relationship between the two processes (41). The data presented here suggest that for RSV, transcription and replication are for the most part independent rather than competitive, interconvertible processes. There may be some situations in which transcription efficiency affects replication efficiency and/or vice versa (Fig. 5), but this likely is an effect of two independent processes sharing the same template.

The data presented in this paper suggest the following model for RSV transcription and replication. A single cis-acting element contained within the first 26 nt of the Le is utilized for both transcription and replication initiation. During replication, the polymerase binds to this sequence and initiates RNA synthesis directly at the 3′ end of the genome. During transcription, the polymerase contacts the 3′ element, but initiates RNA synthesis directly at the first GS signal, possibly by contacting the 3′ element and the GS signal simultaneously.

This model of transcription initiation eliminates the requirement for a transcription termination site before the first GS signal and is consistent with our finding that the Le sequence immediately upstream of the NS1 GS signal is not required for accurate transcription initiation. It also accounts for the lack of a direct inverse relationship between transcription and replication. This model also accommodates our previous finding that increasing the intracellular concentration of N protein augments replication without inhibiting transcription (14). It is possible that encapsidation is necessary for replication, at either the initiation or elongation stage, and therefore increasing the intracellular concentration of N protein does enhance replication efficiency (as determined by synthesis of complete antigenome and genome). However, since transcription is a distinct process, mRNA synthesis is unaffected by N protein concentration. If this model is correct, transcription and replication could be mediated by two separate pools of polymerase which have different conformations allowing them to bind the common promoter element at the 3′ terminus, but then recognize different initiation sites. This is consistent with evidence that VSV transcription and replication are mediated by two subsets of polymerase, differentiated by posttranslational modification (8, 31).

One important postulate of this model is that the Le 3′ terminus contains a common element for initiation of transcription and replication. We are currently testing this proposal by carrying out saturation mutagenesis of the Le region to identify the nucleotide requirements for antigenome and mRNA synthesis.