INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The prototypic rhabdovirus, vesicular stomatitis virus (VSV), is an enveloped, nonsegmented, negative-strand RNA virus with a genomic RNA that is 11,161 nucleotides long (38). Within the virion core, the genomic RNA is tightly wrapped around by the RNA-binding nucleocapsid protein (N) forming ribonuclease-resistant nucleocapsid structure (N-RNA) that serves as the template for transcription and replication of the genome by the viral RNA-dependent RNA polymerase. The viral RNA polymerase is a complex of the phosphoprotein (P) and the large protein (L) (11, 13, 32). Genetic and biochemical studies have suggested that the L protein carries all the enzymatic activities necessary for generation of mature viral mRNAs, i.e., ribonucleotide polymerization activity, methyl- and guanylytransferase activity, and poly(A) polymerase activity (19, 20, 42). P serves as an accessory protein required for the functions of the L protein, and differential phosphorylation of the P protein at different domains appears to influence the transcriptase and replicase activities of the L protein (8, 10, 35, 37).

Following entry of the virus into susceptible cells and uncoating of the viral nucleocapsid in the cytoplasm, the negative-sense nucleocapsid template is first transcribed by the template-associated viral RNA polymerase complex. The polymerase is thought to initiate transcription at the extreme 3' end of the genome (12) and generates a small 47-nucleotide-long uncapped and nonpolyadenylated leader RNA and five capped and polyadenylated mRNAs for the five structural proteins of VSV, namely, N, P, M, G, and L. However, recent studies have also suggested that the viral RNA polymerase may initiate transcription by entering at the internal sites in the genome (9). Transcription from the viral genome is sequential, which reflects the physical order of the genes from the 3' end of the genome (1, 3). Furthermore, transcription is attenuated at each of the gene junctions, resulting in the generation of a gradient in the molar amounts of the mRNAs which also follows the gene order (22, 49). Following translation of the mRNAs, the negative-sense nucleocapsid template is used by the viral polymerase for replication, resulting in the synthesis of a full-length positive-sense antigenomic RNA in the form of nucleocapsid. The antigenomic RNA is subsequently used for further rounds of replication to generate the genomic-sense nucleocapsids.

During the replicative cycle of VSV, the RNA synthetic events, such as transcription, replication, and encapsidation, are controlled by various interactions between the RNA template, the nucleocapsid protein (N), and the RNA polymerase complex. It has been proposed that the interaction between the N protein and the nascent RNA strand is critical for the switch from transcription to replication (6). The cis-acting signals that mediate the RNA synthetic events are located at the termini as well as at the intergenic junctions in the viral genome. With the recent development of methods that allow genetic manipulation of the genomes of VSV and its defective interfering (DI) particles (26, 30, 33, 47, 51), it has been possible to address many of the long-standing questions relating to the role(s) and requirements of various sequence elements at the termini and intergenic junctions in transcription and replication of the viral genome. With cDNAs encoding transcription- and replication-competent minigenomes or minireplicons of VSV, it has been shown by mutational analysis that the first three nucleotides of the conserved sequence 3' UUGUC 5', found at each of intergenic junctions, are required for efficient transcription (48); the 3' AUAC(U)₇ 5' sequence element is required for polyadenylation and subsequent transcription termination (4, 21); the dinucleotide 3' (G/C)A 5' may be required for transcription termination (48), whereas other studies suggest that the dinucleotide may function as a spacer element between the transcription termination and reinitiation signals (5, 21). It has also been shown that the termini of VSV and its DI particle genome contain all the necessary signals for replication (34, 50) and that the termini of DI particle genome and the 3' terminus of VSV antigenome contain a replication enhancer sequence (RES) that upregulates replication (30). The presence of RES element in the DI genome has been proposed to be a key factor for efficient replication of DI RNA as well as for asymmetric levels of replication of genomic and antigenomic RNAs of VSV (30).

Unlike the 3' terminus of the viral antigenomic RNA, which contains only the signals for replication, the 3' terminus of the viral genome must contain the signals for both transcription and replication. By using reconstituted in vitro transcription system and synthetic VSV nucleocapsids (31), it was shown that only the 3'-terminal 22 nucleotides were sufficient to serve as excellent transcription templates (44). To examine the transcription and replication signals at the 3' terminus in the context of VSV genome, we have undertaken a deletion mutational analysis of the first 47 nucleotides corresponding to the leader gene. We have analyzed the mutant templates for their ability to replicate to generate the genomic- and antigenomic-sense RNA products as well as to transcribe. Our results show that the first 24 nucleotides contain overlapping signals for both transcription and replication. Downstream sequences from nucleotides 25 to 47 do not influence replication appreciably;however, they appear to be required for optimal levels of transcription. These results suggest that sequences from nucleotides 25 to 47 within the leader gene may play a more important role in optimal transcription than in replication. Furthermore, deletions within the 3' terminus of VSV genome do not influence the site at which transcription of mRNA is reinitiated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells and viruses. Baby hamster kidney (BHK-21) cells were maintained as monolayer cultures in Eagle's minimal essential medium (MEM) containing 7.5% heat-inactivated fetal bovine serum (FBS) and the antibiotics penicillin G (100 U/ml), streptomycin (20 µg/ml), and kanamycin (20 µg/ml). Thymidine kinase-negative human 143B cells were also maintained as monolayers in Eagle's MEM supplemented with 5% FBS. Stocks of VSV (Indiana serotype, San Juan strain) and the recombinant vaccinia virus (vTF7-3) carrying the bacteriophage T7 RNA polymerase gene (15) were prepared in BHK-21 cells, and infectivity titers of these viruses were determined by plaque assay with BHK-21 and 143B cells, respectively.

Minigenome and protein expression plasmids. Plasmids pN, pP, and pL, encoding the VSV nucleocapsid protein (N), the phosphoprotein (P), and the large protein (L), respectively, under the control of the T7 RNA polymerase promoter have been described previously (36).

Site-directed mutagenesis. A series of deletion mutants (Fig. 1B) spanning the first 47-nucleotide region of the 5' terminus of the VSV antigenomic RNA (which corresponds to the complementary sequences of the 3' terminus of VSV genome) was generated, with p9BN as the template. These plasmids were designated p9BN3'1-6, p9BN3'7-12, p9BN3'13-18, p9BN3'19-24, p9BN3'25-30, p9BN3'31-36, p9BN3'37-42, p9BN3'43-47, p9BN3'25-36, and p9BN3'37-47, which contained deletion of nucleotides 1 to 6, 7 to 12, 13 to 18, 19 to 24, 25 to 30, 31 to 36, 37 to 42, 43 to 47, 25 to 36, and 37 to 47 at the 3' terminus of the VSV genome, respectively. Two more substitution mutants, p9BN-m1 and p9BN-m2, encoding minireplicons (9BN-m1 and 9BN-m2) in which sequences from 25 to 36 and 37 to 47 were replaced by random heterologous sequences GUCAAGCUACGU and GUCAAGCUAGC, respectively, were also generated. All these mutants were generated by the PCR megaprimer method (40). A negative-sense primer containing the desired sequence deletions or substitutions and a primer containing the unique FspI site that annealed to sequences within the -lactamase gene of the vector were used to amplify a fragment of approximately 1,200 bp by PCR, with p9BN as the template. The fragment was then used as megaprimer in a second round of PCR amplification, with p9BN as the template, along with another negative-sense primer containing a unique BglII site (at position 210 of the VSV genome) that annealed to nucleotides 226 to 209 of the N gene. The second PCR product (~1.4 kbp) was digested with BglII and FspI and subcloned into p9BN plasmid digested with the same enzymes. After transformation of competent DH5 cells, bacterial colonies carrying the recombinant plasmids were screened and the mutant plasmids were identified by nucleotide sequencing. Standard methods of plasmid subcloning and preparation (2, 39) were used.

Virus infections and DNA transfections. BHK-21 cells were grown in 60-mm-diameter plates or 35-mm-diameter 6-well plates to about 90% confluencey. These cells were infected with the recombinant vaccinia virus (vTF7-3) at a multiplicity of infection of 10 PFU per cell. Forty-five minutes after virus adsorption at 37°C, inoculum was removed, the cells were washed twice with Dulbecco's modified MEM (DMEM) without serum and transfected with various plasmid DNAs with lipofectin (Gibco/BRL, Gaithersburg, Md.) according to the manufacturer's specifications or with a transfection reagent prepared in the laboratory as described previously (7). At 4 to 5 h posttransfection, medium from transfected cells was removed, cells were washed twice with DMEM containing 2% FBS, and incubated with the appropriate volume of the same medium. For RNA replication and transcription assays in 60-mm-diameter plates, 3 µg of pN, 2 µg of pP, 1 µg of pL, and 5 µg of p9BN or mutant p9BN plasmids were used. These plasmid amounts were reduced by half when 6-well plates were used in the experiments.

Metabolic labeling and analyses of RNA. At 15 to 16 h posttransfection, cells were pretreated with 15 µg of actinomycin D per ml of DMEM at 37°C for 45 min and subsequently exposed to 15 µCi each of [³H]uridine and/or [³H]adenosine per ml of DMEM containing 2% FBS and the same concentrations of actinomycin D for 6 to 8 h. After labeling, cytoplasmic extracts were prepared in lysis buffer as described previously (30, 36). Replicated RNAs in nucleocapsids were immunoprecipitated by polyclonal anti-VSV antibodies. RNAs were recovered from immunoprecipitated nucleocapsids by extraction with phenol and chloroform and precipitation with ethanol. The RNAs were subsequently resolved by electrophoresis in agarose-urea gels (28) and detected by fluorography (25). For transcription studies, cytoplasmic extracts were collected as described above. Labeled RNAs present in extracts were purified by phenol-chloroform extraction, analyzed by agarose-urea gel electrophoresis, and detected by fluorography as described above.

Primer extension analysis. At 24 h posttransfection, cytoplasmic extracts from transfected cells were harvested. Total unlabeled RNA from the extracts was recovered by extraction with phenol and chloroform and subjected to RNase-free DNase (Promega Biotech, Madison, Wis.) digestion to remove residual transfected plasmid DNAs that may have been present in the RNA preparations. After digestion, RNA was extracted with phenol-chloroform and recovered by ethanol precipitation. Dried RNA pellet was resuspended in 5 µl of H₂O, boiled for 1 min, and quick-chilled ice-water. Annealing was performed by mixing the template RNA with 1 µl (20 ng/µl) of minus-sense N gene primer 5' CCTCATTTGCAGG 3' (which anneals to N mRNA at a site that is 80 nucleotides from its 5' terminus) and 2 µl of 5× first-strand synthesis buffer (Gibco/BRL). The mixture was heated at 65°C for 5 min and slowly cooled to room temperature. The following reaction components were added at the final concentrations of 2 mM (each) dGTP, dCTP, and dTTP; 0.1 mM dATP; 10 mM dithiothreitol; 0.1 mM Tris (pH 8.0); 0.01 mM EDTA; 0.75 mCi of [³⁵S]dATP (1,250 Ci/mmol); and 20,000 U of Moloney murine leukemia virus reverse transcriptase per ml of reaction volume. The primer extension reaction was performed at 37°C for 90 min. The reaction was terminated by adding 6 µl of sequencing reaction stop solution (United States Biochemicals, Cleveland, Ohio). The samples were denatured at 85°C for 3 min, and 4 µl of the reaction products was electrophoresed in a 6% polyacrylamide sequencing gel alongside a sequence ladder generated with the same primer and p9BN template. The primer extension products were detected by autoradiography.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

In order to identify the replication and transcription signals at the 3' terminus of the VSV genome, we used the plasmid p9BN (30) encoding an antigenomic minireplicon RNA (9BN) containing the VSV N gene and part of the L gene flanked by the 3'- and 5'-terminal sequences from the VSV antigenome. The processes of encapsidation, replication, and transcription of 9BN RNA in transfected cells in the presence of the viral proteins are schematically depicted in Fig. 1A. Briefly, in transfected cells, the antigenomic RNA template generated from p9BN plasmid by T7 RNA polymerase is encapsidated by the N protein and replicated by the VSV RNA polymerase (P and L proteins) to produce the full-length minus-sense RNA, which in turn serves as the template for further rounds of replication as well as for transcription to generate NL mRNA (30).

View larger version (48K): [in this window] [in a new window]   FIG. 1.   (A) The plasmid p9BN encoding the antigenomic plus-sense minireplicon. Only the relevant regions of the plasmid are shown. 10, T, and  represent the T7 RNA polymerase promoter, the terminator, and the hepatitis delta virus ribozyme sequences, respectively; le' and tr' represent complementary sequences of leader and trailer regions of the VSV genome; N and L represent the coding sequences of the N and part of the L gene of VSV. In plasmid-transfected cells, the antigenomic minireplicon, synthesized from p9BN by T7 RNA polymerase, is encapsidated by the viral N protein to generate plus-sense nucleocapsid with two additional guanosine residues at the 5' terminus. Replication of this nucleocapsid by VSV RNA polymerase generates the genomic minus-sense nucleocapsid, which serves as the template for transcription by VSV RNA polymerase to synthesize NL mRNA and also for replication to generate the antigenomic plus-sense nucleocapsid. The plus-sense RNA synthesized by VSV RNA polymerase differs from the plus-sense RNA synthesized by T7 RNA polymerase by the absence of the two 5'-terminal guanosine residues. (B) Various mutant minireplicons with deletion or substitution of nucleotides (as shown) at the 5' terminus of the antigenomic minireplicon (9BN), which corresponds to the leader gene sequences at the 3' terminus of genomic-sense RNA. Sequences of the first 55 nucleotides at the 5' terminus of the antigenomic RNA are shown. They correspond to 47 nucleotides of leader RNA sequence, three nontranscribed intergenic nucleotides (UUU, bold-faced), and the first five nucleotides (AACAG, outlined) of N mRNA. Underlined sequences represent nucleotides that are complementary to the sequences at the 3' terminus of the antigenome. Deleted nucleotides within the leader region are represented by dots. Random heterologous sequences that replace leader sequences are shown in boxes.

Replication signal(s) at the 3' terminus of VSV genome. A series of six- and five-nucleotide deletions spanning the 5'-terminal 47-nucleotide region of the antigenomic-sense minireplicon RNA was generated (Fig. 1B). These templates, when replicated by VSV polymerase, generate genomic-sense templates with deletion of sequences at the 3' terminus, and therefore the effect of these deletions on transcription and replication could be examined. These mutants were designated 9BN3'1-6, 9BN3'7-12, 9BN3'13-18, 9BN3'19-24, 9BN3'25-30, 9BN3'31-36, 9BN3'37-42, and 9BN3'43-47 (Fig. 1B). Since the wild-type and mutant minireplicon RNAs must be first encapsidated by the N protein to serve as templates for replication, we initially examined the ability of mutant RNAs to be encapsidated by the N protein. When cells expressing the N protein and the wild-type or mutant minireplicon templates were labeled with [³H]uridine, almost similar amounts of labeled wild-type or mutant minireplicons were immunoprecipitated from the cell extracts (data not shown), suggesting that encapsidation of mutant minireplicons was unaffected by deletion of sequences at the 5' terminus. This was surprising, since the encapsidation signal(s) is presumed to reside at the 5' terminus of the RNA. It is possible that the six-nucleotide deletions are not large enough to disrupt the encapsidation signal or that deletion of small regions could be functionally replaced by adjacent sequences.

View larger version (76K): [in this window] [in a new window]   FIG. 2.   (A) Replication of various mutant antigenomic minireplicons to generate genomic-sense 9BN() RNA. Cells infected with vTF7-3 were transfected with plasmids encoding the N, P, and L (or without L in the negative control, lane 2) proteins and either p9BN or deletion mutant plasmids. At 16 h posttransfection, cells were labeled for 6 h with [3H]uridine in the presence of actinomycin D. Replicated RNAs in the nucleocapsids were immunoprecipitated, purified, and analyzed by agarose-urea gel electrophoresis as described in Materials and Methods. G and N represent the G mRNA and N mRNA of VSV isolated from infected cells. The migration position of plus-sense RNA is indicated by an arrow. 9BN() is the negative-sense replication product. An asterisk indicates the band of NL mRNA, which is sometimes immunoprecipitated (30, 50) by anti-VSV antibodies. (B) Relative levels of replication of deletion mutant minireplicons to produce genomic-sense products. Histograms show averages and range of levels of replication of various deletion mutants (described at the top of panel A) from three independent experiments.

View larger version (71K): [in this window] [in a new window]   FIG. 3.   (A) Primer extension analysis to detect 5' termini of various plus-sense RNA products in transfected cells. Lane 6, extension products of RNAs from cells after transcription and replication; lane 5, extension products of RNAs without transcription or replication. The extension products labeled a, b, and c are described in the text. The 5' terminus of the replication product (indicated by a dot in lane 6) maps to the T residue (identified by the top arrow on the left) corresponding to the first nucleotide (U) in the VSV genome. The 5' terminus of the NL mRNA transcription product maps to the first T residue (identified by the bottom arrow on the left) of the transcription initiation signal UUGUC in the VSV genome. The more intense top bands in doublets c and b most likely represent the extension product of capped mRNA. (B) Analysis of plus-sense RNA replication products from mutant templates by primer extension. Replication products are identified by dots in the lanes. Only the top portion of the gel is shown. (C) Average normalized replication of various mutants (shown at the top of panel B) from two separate experiments as determined with the following formula: Normalized levels of plus-sense replication products (%) = relative levels of plus-sense replication products/relative levels of minus-sense template × 100.

Transcription signal(s) at the 3' terminus of VSV genome. We next examined the 3' terminal sequences of VSV genome that are required for mRNA transcription. Total RNA from transfected cells radiolabeled in the presence of actinomycin D was analyzed in agarose-urea gel. Results (Fig. 4A) show that no NL mRNA was detected from the first three deletion mutants, 3'1-6, 3'7-12, or 3'13-18 (lanes 4 to 6, respectively), even after much longer exposure of the fluorogram. The levels of NL mRNA from 3'25-30, 3'31-36, 3'37-42, and 3'43-47 (lanes 8 to 11, respectively) were 25% to 35% of that of the wild-type template (lane 3), whereas 3'19-24 template supported transcription at a level of only about 10% (lane 7). When the levels of transcription were normalized by considering the levels of minus-sense template (as shown in Fig. 2), the results (Fig. 4B) suggested that transcription activity of the mutants 3'25-30, 3'31-36, 3'37-42, and 3'43-47 were 35 to 45% of the level of wild-type template. The mutant 3'19-24 was 12 to 15% as active as the wild-type template, whereas the three deletion mutants 3'1-6, 3'7-12, and 3'13-18 were completely inactive in transcription. Furthermore, it must be pointed out that the transcription signal(s) at the 3' terminus of minus-sense template is much stronger than the replication signal(s), since the transcription products represented greater than 95% of the total RNA synthesized from the minus-sense template.

View larger version (67K): [in this window] [in a new window]   FIG. 4.   (A) Analysis of transcription from the mutant templates. Cells were infected with vTF7-3 and transfected with plasmids encoding N, P, and L and wild-type or mutant minireplicons. At 16 h posttransfection, cells were treated with actinomycin D and labeled with [3H]uridine for 6 h. Total labeled RNA from these cells was analyzed by electrophoresis in agarose-urea gel. Lane 2 shows RNA from the negative control (no L plasmid in transfection) sample. An arrow indicates the NL mRNA transcription product. (B) Normalized levels of transcription from the mutant templates. Values were obtained by using the formula: Nomarlized transcription (%) = relative levels of transcription/relative levels of minus-sense template × 100. Histograms represent the averages and ranges of values from three separate experiments.

Do downstream sequences play role in replication and/or transcription? Results from previous experiments (Fig. 2 to 4) suggested that the first 24 nucleotides at the 3' terminus of VSV genome contained the essential signal(s) for replication and transcription. Furthermore, the six-nucleotide deletions spanning nucleotides 25 to 47 did not appear to have any major negative effects on transcription or replication. It is possible that this region contains functionally redundant signals and that the six-nucleotide deletions may have only a partial effect on replication and/or transcription. To investigate if larger deletions within this region would have a more dramatic effect on replication and/or transcription, we generated two mutant plasmids, p9BN 3'25-36 and p9BN 3'37-47 (Fig. 1B), that encode plus-sense minireplicon templates with deletions of nucleotides 25 to 36 and 37 to 47 at their 5' termini. The initial round of replication of these RNAs would generate minus-sense genomic RNA templates with the corresponding deletions at the 3' end.

View larger version (62K): [in this window] [in a new window]   FIG. 5.   Transcription and replication activities of larger deletion mutant templates. (A) Analysis of negative-sense replication products [p9BN()] as described in Fig. 2A. Lane 2 shows replication products from the negative control (no L plasmid in transfection) sample. (B) Analysis of plus-sense replication products (identified by dots in lanes 2 to 4) by primer extension analysis as described in the legend for Fig. 3B. (C) Analysis of NL mRNA transcription products (indicated by arrow) as described in the legend for Fig. 4A. (D) Normalized levels of transcription and replication from minus-sense templates, as described in the legends for Fig. 3C and 4B.

View larger version (44K): [in this window] [in a new window]   FIG. 6.   Transcription and replication activities of substitution mutant minireplicons. (A) Analysis of minus-sense replication products (indicated by arrow) as described in the legend for Fig. 2A. Lane 1 contains RNA from the negative control (no L) sample. (B) Analysis of NL mRNA transcription product (indicated by arrow) as described in the legend for Fig. 4A. (C) Normalized levels of transcription and replication from negative-sense templates as described in the legends for Fig. 3C and 4B.

Primer extension analysis to examine the 5' terminus of mRNAs from the mutant minireplicons. We next examined the site of reinitiation of transcription from these mutant templates by examining the 5' termini of the NL mRNAs produced from the mutant templates. Total RNA from transfected cells was isolated and subjected to primer extension analysis with the primer as described in the experiment shown in Fig. 3A. This primer hybridizes to the VSV RNA polymerase-derived NL mRNA transcription product and generates the extension products (the doublet marked c in Fig. 3A). Results (Fig. 7) show that the 5' termini of NL mRNAs from the mutant templates mapped to the first T residue at the authentic site of transcription initiation (UUGUC in the viral genome) (lanes 10 to 14) as seen for the wild-type template (lane 6). The mutant templates with deletions of sequences 1 to 6, 7 to 12, and 13 to 18 did not produce the primer extension products (lanes 7 to 9), since these templates were transcriptionally inactive (Fig. 4). Low levels of primer extension products from RNA generated from 3'19-24 mutant template (lane 10) reflect the relatively low level of transcription from this template. Furthermore, the NL mRNAs synthesized from the two larger deletion mutants 3'25-36 and 3'37-47 were also initiated from the same initiation site (data not shown). These results indicate that deletions within the leader gene do not affect the site of reinitiation of transcription from the gene immediately downstream.

View larger version (51K): [in this window] [in a new window]   FIG. 7.   Primer extension analysis to examine the 5' terminus of transcription products from the mutant minireplicons. Analysis was performed as described in the legend for Fig. 3A, and the data shown in the figure represent the extension products of NL mRNA. Lane 5 shows the extension products of RNA from a negative control experiment in which L plasmid was omitted from the transfection mixture. The 5' terminus of uncapped NL mRNA from wild-type and mutant minireplicon templates maps to the T residue (shown by the arrow in the sequence ladder) of the transcription initiation site, UUGUC, in the viral genome.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The 3' terminus of the negative-sense genomic RNA of VSV is presumed to contain the cis-acting signals for transcription and replication of the genome. In this study, using a minigenomic RNA of VSV, we have examined the role of sequences within the first 47 nucleotides (corresponding to the leader gene) at the 3' terminus in mediating transcription and replication by the viral RNA polymerase under in vivo conditions. From the data presented, the first 24 nucleotides appear to contain the signals necessary for transcription and replication. The downstream sequences from nucleotides 25 to 47 appear to be necessary for optimal levels of transcription.

The observation that mutant templates with a deletion of sequences within the first 18 nucleotides were completely defective in transcription as well as in replication, whereas the template with deletions of nucleotides 19 to 24 was partially active, suggests that the signals for transcription and replication are contained within the first 24 nucleotides and that they overlap. It is possible that both transcriptase and replicase recognize the same sequence to initiate transcription and replication. In fact, the 3'-terminal 18 nucleotides have been implicated as having a major role(s) in RNA synthesis. Our results are consistent with the proposal from sequence analysis showing that the first 18 nucleotides at the 3' termini of various serotypes and strains of VSV have strong homology (16, 23) and therefore have been thought to be involved in the initiation of RNA synthesis (transcription and replication). Furthermore, with the reconstituted synthetic VSV nucleocapsids under in vitro transcription conditions, it has been shown that the first 15 to 17 nucleotides at the 3' end of the negative-strand RNA are required for optimal transcription (44). In addition, only the first three nucleotides, 3' UGC, which are invariant in all rhabdoviruses, are absolutely essential for transcription, and nucleotides at positions 4 to 17 are not as critical as the overall length of the sequence for optimal transcription (44). However, it should be noted that in these in vitro transcription reactions, the reconstituted synthetic templates were very small (22 nucleotides), and therefore the processivity of the polymerase and the contribution of downstream nucleotides to transcription elongation by the polymerase could not be assessed. In the data present here, synthesis of mature mRNA (NL mRNA) form various mutant templates in vivo was measured. Although analysis of leader RNA synthesis is the most direct approach to assess the effects of these deletions, considering the sequential mode of transcription of VSV genes (1, 3), synthesis of mRNA could also be used for such studies. While our data are generally consistent with those obtained previously (44), we suggest that the first 19 to 24 nucleotides at the 3' terminus of negative-sense VSV genome are essential for transcription. Our data, however, cannot rule out the possibility that the mutant templates with deletions of sequences within the first 24 nucleotides were active in transcription initiation to generate short initiated transcripts.

Data shown in Fig. 5 and 6 suggest that deletion of larger regions spanning nucleotides 25 to 47 within the leader gene or replacement of these sequences with random heterologous sequences affected transcription more dramatically than replication. It is possible that these regions contain functionally redundant signals for replication and that deletion or substitution of one or the other region has a less adverse effect on replication. But the importance of these sequences in optimal levels of transcription could be realized from these mutants which rendered the templates transcriptionally less active (about 10 to 15% of the wild type). It seems, therefore, that the sequence requirements for optimal transcription are different from those for replication. The observation that optimal transcription requires the sequences downstream is consistent with the data from dimethyl sulfate methylation protection studies (24), which suggest a sequence-specific high-affinity binding site for the P protein at the 3'-terminal nucleotides 16 to 30 (3' ... GGUAAUAAUAGUAAU ... 5') corresponding to the middle part of leader gene. It was proposed (17) that this AU-rich sequence element may be analogous to the Goldberg-Hogness box located near the transcription initiation sites of eukaryotic genes (18). Another line of evidence that indicates the importance of this sequence in transcription is derived from studies with chimeric synthetic nucleocapsids of Indiana and New Jersey serotypes of VSV. The divergent RNA sequence of the leader gene (nucleotides 22 to 50/51) appeared to be a major determinant of serotype specificity for transcription and an indication that this sequence may be important for transcription elongation rather than initiation (45). These observations together with the data presented here support the conclusion that downstream sequences (nucleotides 25 to 47) are important for optimal levels of transcription, which may be mediated by interaction with the P protein of the polymerase complex. The findings that certain nucleotide insertions within this region of the leader gene downregulate transcription without significantly affecting replication (29, 50) further support this conclusion.

It is not known how the downstream sequences mediate optimal levels of transcription. In addition to the primary sequence, secondary structures within this AU-rich region might be involved. A potential stem-loop structure has been predicted to exist in the central region (nucleotides 16 to 33) of the leader RNA (17), suggesting that complementary sequences at the corresponding positions within the 3' terminus of minus-sense VSV genome may contain similar structures and may play an important role in regulating transcription. This is also the region at which the P protein interacts with the template (24), so it is possible that transcription activity of the RNA template is modulated by the presence of specific sequences as well as secondary structures. However, the existence of RNA secondary structures in the N protein-associated nucleocapsid templates and their involvement in signaling RNA synthetic events must await further investigation.

In a previous study (30), we showed that the 3' terminus of the VSV antigenome as well as the DI particle genome and antigenome contain a minimal promoter (within nucleotides 1 to 24) for replication and another element (RES, nucleotides 25 to 45) that enhances replication. The differential presence of this enhancing element was proposed to account for asymmetric levels of synthesis and accumulation of genomic and antigenomic RNAs in VSV-infected cells (30, 41, 43, 46). A similar situation might exist at the 3' terminus of the VSV genome for transcription in that the first 24 nucleotides contain the minimal promoter element for transcription (and replication) and that downstream sequences from nucleotides 25 to 47 are necessary for optimal transcription. Further studies on this region may provide important clues as to how these sequences mediate optimal transcription from the 3' end of the VSV genome.

Results shown in Fig. 7 suggest that the transcription reinitiation site at the leader-N gene junction remains unaffected by deletion of nucleotides within the leader gene. It is obviously important to examine synthesis and termination of leader RNA from these mutant templates. It must be pointed out that analysis of leader RNA from transfected cells has been difficult because of its small size and relative instability (27). However, preliminary data (not shown) on leader RNA analysis indicate that the minigenome templates that are transcriptionally active also generate the leader RNA. A more detailed investigation is currently under way to examine this possibility.

In conclusion, our results show that the first 24 nucleotides at the 3' terminus of the negative-sense genome of VSV contain overlapping signals for transcription and replication. The sequences downstream within the leader gene are necessary for optimal levels of transcription.