Journal of General Virology (2005), 86, 171–180 DOI 10.1099/vir.0.80435-0 Nature of a paramyxovirus replication promoter influences a nearby transcription signal Diane Vulliémoz, Samuel Cordey, Geneviève Mottet-Osman and Laurent Roux Correspondence Laurent Roux Department of Microbiology and Molecular Medicine, University of Geneva Medical School, CMU, 1 rue Michel-Servet, CH-1211 Geneva 4, Switzerland Laurent.Roux@medecine. unige.ch Received 9 July 2004 Accepted 14 October 2004 The genomic and antigenomic 39 ends of the Sendai virus replication promoters are bi-partite in nature. They are symmetrically composed of leader or trailer sequences, a gene start (gs) or gene end (ge) site, respectively, and a simple hexameric repeat. Studies of how mRNA synthesis initiates from the first gene start site (gs1) have been hampered by the fact that gs1 is located between two essential elements of the replication promoter. Transcription initiation, then, is separated from the replication initiation site by only 56 nt on the genome, so that transcription and replication may sterically interfere with each other. In order to study the initiation of Sendai virus mRNAs without this possible interference, Sendai virus mini-genomes were prepared having tandem promoters in which replication takes place from the external one, whereas mRNA synthesis occurs from the internal one. Transcription now initiates at position 146 rather than position 56 relative to the genome 39 end. Under these conditions, it was found that the frequency with which mRNA synthesis initiates depends, in an inverse fashion, on the strength of the external replication promoter. It was also found that the sequences essential for replication are not required for basic mRNA synthesis as long as there is an external replication promoter at which viral RNA polymerase can enter the nucleocapsid template. The manner in which transcription and replication initiations influence each other is discussed. INTRODUCTION The first step in the Sendai virus (SeV) multiplication cycle is the production of mRNA from a helical subviral nucleocapsid, in which the viral genome RNA is tightly and stoichiometrically associated with the viral nucleocapsid (N) protein. This N-subunit assembly, together with multiple attached viral polymerases (a complex of the P and L proteins) is the minimum subviral unit that is thought to retain infectivity (Lamb & Kolakofsky, 2001). The synthesis of negative-stranded RNA genomes (or antigenomes) and their assembly with N protein is coupled, and these viral RNAs are only found as nucleocapsids (Gubbay et al., 2001). SeV nucleocapsid appears as a flexible helical assembly complex of variable pitch, with 13 N-subunits per turn, in which each N-subunit binds 6 nt (Bhella et al., 2002; Egelman et al., 1989). For viruses of the subfamily Paramyxovirinae, efficient replication of model mini-genomes in transfected cells requires that their total length be a multiple of six (Calain & Roux, 1993; Vulliémoz & Roux, 2001). The requirement for hexameric genome length for viruses of the subfamily Paramyxovirinae has recently been underscored using reverse genetic systems for Human parainfluenza virus 2, a rubulavirus, and measles virus, Morbillivirus (Rager et al., 2002; Skiadopoulos et al., 2002). 0008-0435 G 2005 SGM The genomic and antigenomic replication promoters (G/Pr and AG/Pr) of paramyxoviruses are found within the terminal 96 nt or 16 N-subunits of each RNA and are bi-partite in nature (Murphy et al., 1998; Pelet et al., 1996; Tapparel et al., 1998). Approximately 30 nt at the 39 end constitute the first element (PrE-I) in which the first 12 nt are conserved between genomes and antigenomes, and across each genus. A second downstream element (PrE-II) is found within the 59-UTR of the N gene or the 39-UTR of the L gene (Fig. 1). For SeV and Human parainfluenza virus 3, PrE-II is a simple but phased hexameric sequence repeat (39 [C1n2n3n4n5n6]3) bound to the fourteenth, fifteenth and sixteenth N protein subunits (Hoffman & Banerjee, 2000; Tapparel et al., 1998). For simian virus 5, [n1n2n3n4G5C6]3 is repeated in subunits 13, 14 and 15 (Murphy et al., 1998). PrE-II, then, is adjacent to PrE-I in the helical nucleocapsid, forming a common or contiguous surface on two turns of the helix. The hexamer (or N-subunit) phase of at least PrEII is known to be critical for its function (Tapparel et al., 1998; Murphy et al., 1998). This phase effect is thought to be due to the different chemical environments of each of the 6 nt bases associated with each N-subunit, as revealed by chemical attack studies of resting viral nucleocapsids (Iseni et al., 2002). Adenosines in any hexamer phase are largely protected from dimethylsulfate, whereas cytosine Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Printed in Great Britain 171 D. Vulliémoz and others Fig. 1. Primary structure of the SeV replication promoters. The 96 nt of the genomic (G/Pr) and antigenomic (AG/Pr) are presented as RNA sequence in ‘hexamers’, numbered 1–16 from the 39 end (-OH 39). PrE-I and PrE-II refer to the two elements described as essential for viral RNA replication (see text), with the three times repeated motif in PrE-II indicated as C*. In G/Pr, the leader coding sequence (leaderC, 55 nt) is outlined, as well as the N gene transcription start signal (nt 56–65), which constitutes the first transcription start signal (N gs1) seen by the RdRp. In AG/Pr, the trailer coding sequence (trailerc, 57 nt) is shown, as well as the complement of the L gene polyadenylation site (L gec). For more details see text of Introduction. reactivity is high only in hexamer positions one and six, precisely the positions of the conserved cytidines in the PrE-II element of G/Pr and AG/Pr. Besides the bi-partite nature of G/Pr and AG/Pr, another feature conserved among the viruses of the subfamily Paramyxovirinae (respiro-, morbilli- and rubulaviruses) is that the first (N) mRNA always starts precisely opposite U56, within the conserved decamer mRNA start signal 39 55-AU2CCCA NUUUCN-66 (for SeV, phase indicated). Curiously, none of the conserved cytidines of this cis-acting signal are in hexamer positions one and six. The N gene mRNA start site (abbreviated as gs1) is also on another face of the helix than the bi-partite replication promoter in resting nucleocapsids. Since virus RNA-dependent RNA polymerase (vRdRp) initiates chains at two closely spaced start sites on the N : RNA template, one would expect that these two events would interfere with each other, at least under some conditions, and evidence of the negative influence of gs1 on SeV replication promoter strength has recently been presented (Le Mercier et al., 2003). The study of the influence of the replication promoters on mRNA synthesis from gs1, however, is complicated by its location between the two essential elements of the replication promoter, and the fact that the two RNA initiation events are only 56 nt apart on the N : RNA. In order to study the initiation of SeV mRNA in the absence of the above complications, we have prepared SeV mini-genomes with tandem 96 nt long G/Pr in which replication takes place 172 from the external G/Pr, whereas mRNA synthesis occurs from the internal G/Pr, having initiated at position 146 rather than position 56. METHODS Virus and cells. BSR-T7 cells, a BHK cell line constitutively expressing a T7 RNA polymerase, a gift from K.-K. Conzelmann (Max-von-Pettenkofer Institute & Gene Center, LudwigMaximilians-University Munich, Munich, Germany) (Buchholz et al., 1999), were propagated in Dulbecco minimal Eagle medium (DMEM) supplemented with 5 % fetal calf serum (FCS) in a 5 % CO2 atmosphere. The AGP-55 recombinant Sendai virus (rSeVAGP55) was constructed and rescued as described previously (Le Mercier et al., 2002). AG/Pr of this virus has the first 55 nt of the trailer sequence replaced with the 55 nt of the leader sequence (see Fig. 1). Virus stocks, prepared in 9-day-old embryonated chicken eggs from three times plaque purified virus, reached titres ranging between 56108 and 109 p.f.u. ml21. Sequence and plasmids. All the plasmids harbouring the mini- replicons expressing the green fluorescent protein (GFP) are derived from pSV-DI-H4D96 described in Vulliémoz & Roux (2001, 2002). In all derivatives used in this study, the 39 end of the mini-genome RNA complementary to the T7 RNA transcript (intermediate replicon for GFP template RNA replication) contains an AG/Pr (see Fig. 2a of the present work, construct [AGP]). The GFP open reading frame (ORF), flanked by the SeV transcription start and stop signals, was introduced between the CelII and MunI restriction sites (as in Vulliémoz & Roux, 2001), and the sequence located between MunI and DraIII was deleted to generate [GP]. A cassette containing the adequate promoter(s) between DraIII and BamHI generated constructs [GP-GP], [AGP] and [AGP-GP]. Site-directed substitutions Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Journal of General Virology 86 Paramyxovirus replication and transcription Fig. 2. Generation of the ectopic transcription start signal. (a) Schematic representation of the mini-replicons expressing GFP. First line: derivative of pSP65 plasmid containing the mini-genome encoding sequences, flanked by the Hepatitis delta virus ribozyme (Rbz) and the T7 RNA polymerase promoter (T7p). Encapsidated T7 RNA transcripts which constitute proper mini-genome templates for synthesis of GFP mRNAs are portrayed below (lines 1–5) in a 39 to 59 orientation. In the double promoter constructs, a 6 nt intervening sequence (XhoI site) is inserted between the two promoters (short black line). The 59 end extremity of the mini-replicons is invariably the AG/Pr complementary sequence (AG/Prc), meaning that the anti minigenome RNAs invariably harbours an AG/Pr. GFP, green fluorescent protein ORF. Right-angle arrow, viral transcription start signal (in text, gs1). White cross and flush end right angle bar in G/Pr, C58A mutation portraying transcription arrest. Open triangle, deletion at the 39 end of G/Pr. The extent of the deletion is indicated by the denomination of the construct to the left: d12 means a 12 nt deletion. (b) Representative Northern blot analysis of the encapsidated RNAs produced in the rescue of the mini-replicons presented in (a). About 107 BSR T7 cells were infected with rSeV-AGP55 (helper virus) and transfected with the plasmids harbouring the corresponding mini-genome sequences as described in Methods. Encapsidated RNAs were purified and analysed by Northern blotting using a specific 32P-labelled riboprobe of positive polarity (see Methods). ND, rSeV-AGP55 genome. DI-RNA, mini-replicons RNAs, templates for GFP messages. The numbers below refer to levels of replication (arbitrary units, measured as described in Methods). Note the relative position of the mini-replicons DI-RNAs consistent with replication initiation at the external promoter only. (c) Relative GFP fluorescence normalized to the corrected replication (see Methods) averaged from three independent experiments with [AGP-GP] taken as the series standard (100 %). Deviation of the mean is indicated by bars. (C58A) and deletions in the genomic promoter of [GP58A-GP], [GP58A], [GP58A-GPd1258A], [GPd12] and [AGP-GPd1258A] were introduced by fusion PCR with adequate oligonucleotides and primers. Constructs [GP58A-GPd12] as well as [AGP-GPd12] and [AGP GPd48] were derived from constructs [GP58A-GP] and [AGP-GP], respectively, with specific substitutions. Plasmids pTM1-N, -P/Cstop and -L were constructed by introducing in the pTM1 vector the SeV-N, -P/Cstop and -L genes as described in Calain et al. (1992) by using the NcoI site following the Encephalomyocarditis virus IRES. http://vir.sgmjournals.org Rescue of mini-replicons in the presence of the helper rSeV-AGP55. BSR-T7 cells were seeded at 16106 cells in 9 cm diameter Petri dishes. The next day, 10 p.f.u. per cell of rSeV-AGP55 in 1 ml basic salt solution (BSS) was added (33 uC, 1 h). The infectious medium was replaced drop by drop with a pre-prepared mix containing 12 mg plasmid harbouring GFP-mini-genome and 24 ml Fugene (Roche) in a total of 400 ml BSS. After 6 h, 10 ml of DMEM-10 % FCS replaced the transfection mix. The next day, fresh FCS-free DMEM was added. At day 4, medium and cells were Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 173 D. Vulliémoz and others collected after trypsinization (3 mg ml21 acetylated trypsine), and overlaid onto subconfluent fresh BSR-T7 cells, infected with 10 p.f.u. per cell of rSeV-AGP55. After 4 h, floating cells and medium were replaced by 10 ml FCS-free DMEM. At day 6, the cells were trypsinized, resuspended in DMEM-5 % FCS to inactivate the trypsin, pelleted and finally resuspended in 1 ml PBS. A fraction of the cell samples was used for analysing GFP expression by flow cytometry. The remainder was pelleted and used to characterize the extent of replication or transcription by Northern blotting and primer extension, respectively. distribution among the different templates as follows: Replication/ %gated6100=Corrected Replication. In the end GFP fluorescence was expressed as: Mean fluorescence/Corrected Replication = GFP fluorescence. To be able to integrate the data of more than one experiment, the results were expressed as a percentage relative to one template taken as the reference for the series; this reference is indicated in the figure legend, as is the number of independent experiments (three or four) performed, and the mean of these values (see also text of Results). Replication of mini-replicons in the presence of support plasmids. Confluent BSR-T7, seeded as described above the day RESULTS before, were transfected with a mixture of plasmids including the plasmid harbouring the mini-genome (5 mg), the pTM1-N (1?5 mg), pTM1-P/Cstop (1?5 mg), pTM1-L (0?5 mg), 20 ml Fugene (Roche). Thirty-six hours post-transfection, the cells were collected and treated as above for GFP expression analysis by flow cytometry and for Northern blot analysis. Recovery of the encapsidated and non-encapsidated viral RNAs. Infected/transfected BSR-T7 cells were collected as described above. Nine-tenths was pelleted and resuspended in 1 ml lysis buffer (0?6 % NP40, 50 mM Tris/HCl pH 8?0, 10 mM NaCl; Mottet & Roux, 1989). Post nuclei supernatants were made 5 mM in EDTA and loaded onto linear 20–40 % w/w CsCl gradients (Beckman SW60). After centrifugation (40 000 r.p.m., 12 uC, overnight), the nucleocapsids banding in the CsCl gradient and the non-encapsidated cellular and viral RNAs in the pellet were separately collected as described previously (Calain & Roux, 1995). Poly adenylated RNAs were selected in the non-encapsidated RNA fraction using the Oligotex mRNA mini kit (Qiagen), according to the supplier’s instructions. Northern blot analysis. Northern blots were performed as described previously (Calain & Roux, 1995). To score replication the 32 P-labelled riboprobe contained promoter and GFP-specific sequences, which allow detection of the helper virus genome. To score nonencapsidated RNAs present in the CsCl gradient pellets, a GFPspecific probe was prepared. The blotted membranes were exposed to Kodak X-Omat films. The autoriadiographs were scanned and the intensity of the replication signal was measured using ONE-Dscan version 1.0 (Scananalytics; CSP). Primer extensions. Primer extensions were done as described pre- viously (Vulliémoz & Roux, 2001). An infrared dye-labelled oligonucleotide (IRD 800, 59-CAGCTTGCCGTAGGTGGCATCGCCC-39) of negative polarity positioned in the GFP ORF was used. Due to built in properties of the automated sequencer (LI-COR DNA 4000; MWG-Biotech), the results can only be interpreted semi-quantitatively and to define the position of RNA synthesis initiation. Analysis of the GFP expression. Infected/transfected cells were collected in PBS. Flow cytometry was performed on a BectonDickinson FACSCan2. R1 and M1 parameters were adjusted on BSR-T7 cells infected with rSeV-AGP55 and transfected with the plasmid harbouring [GPd12]. Data analysis was performed with a Becton Dickinson software. GFP relative expression. Estimation of GFP gene transcription by measures of mean GFP fluorescence depended on the amount of the template available responsible for GFP expression. However, due to differences in replication ability of the different templates, the fraction of the cells harbouring these templates varied. The flow cytometry measures, however, allow one to estimate, for each template, its distribution which corresponds to the percentage of gated cells (%gated). Replication was then corrected for the variable 174 A transcription start site positioned downstream of a replication-only promoter In this study, SeV transcription was followed using a series of mini-replicons expressing GFP, and GFP fluorescence was measured by flow cytometry. Since GFP expression depends on the efficiency of mini-genome transcription, but also on template availability, it was normalized to mini-genome abundance as estimated by Northern blotting. Moreover, the fraction of the cells harbouring these templates in the culture will vary according to the replication ability of the mini-genomes. Since the mean fluorescence of the FACS analysis is that of the cells that are gated, these results should also be normalized for the different distribution. For example, when a certain amount of template (Northern blot signal) is distributed in 20 or 90 % of the cells, the amount of template per cell is 4?5 times higher in the former case. Consider these two situations: 100 templates distributed in 100 cells (one template per cell), or 100 templates distributed in 20 cells (five templates per cell); if, in both situations, the mean fluorescence of the gated cells is the same, then the five templates in the 20 cells transcribe fivefold less efficiently than the template in the 100 cells. In the end, crude GFP measure by flow cytometry was standardized to the amount of template corrected for their distribution in the culture (see also Methods). This approach was, finally, validated by direct measurement of GFP mRNAs (see below). Transcription downstream of AG/Pr To create mini-replicons in which GFP mRNAs do not initiate from within the elements used for replication, minireplicons with tandem 96 nt promoters were used, in which the external promoter was AG/Pr preceding a G/Pr, harbouring the active gs1 (Fig. 2). The series of GFP minireplicons used to characterize this general configuration is shown in Fig. 2(a). [AGP] is a single promoter construct with AG/Pr directly upstream of the GFP ORF. These double promoter constructs initiate replication only from the external AG/Pr (AG/Prext) (Vulliémoz & Roux, 2001, 2002). Whether they are transcription competent was open to question, either when the internal GP is integer [AGP-GP] or when it carries a 12 nt deletion in PrE-I. Construct [AGPGPd1258A], further carries a C58A mutation on the internal G/Pr (G/Print) that strongly decreases transcription from gs1 (Le Mercier et al., 2002). It was introduced here to verify Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Journal of General Virology 86 Paramyxovirus replication and transcription whether transcription from [AGP-GPd12] can be suppressed. Finally, [GPd12] was designed to test the effect of the 12 nt deletion on [GP] transcription. Fig. 2(b) shows that all the constructs were replication competent, within the same range, with the exception of [GPd12] for which the 12 nt deletion totally suppresses replication. As indicated by the difference in migration between the [AGP] and constructs 2, 3 and 4, replication of the double promoter constructs initiated only at AG/Prext as expected (see also Fig. 6). As shown in Fig. 2(c), construct [AGP] only expressed background levels of GFP, indicating that GFP expression depends on the presence of gs1. Construct [AGP-GP], on the other hand, did so significantly. The 12 nt deletion of G/Print, which results in replication suppression (see [GPd12]) had no deleterious effect on GFP expression (see [AGP-GPd12]). Inhibition of this expression by the C58A mutation (construct [AGP-GPd1258A]) indicates that mRNA initiation takes place at gs1. In conclusion, an ectopic gs1 within an internal, replication-incompetent G/Pr, is functional downstream of AG/Pr. This raises the question of how vRdRp reaches gs1. Does it enter the template at AG/ Prext and proceed to gs1, or is direct entry at G/Pr possible? Internal gs1 still functions without the two replication promoter elements The 12 nt deletion at the 39 end of PrE-I of G/Print allows transcription from gs1146 (see Fig. 2); this part of PrE-I that is essential for replication is then dispensable for transcription. When the 12 nt deletion was extended to 48 nt ([AGPGPd48], Fig. 3), transcription was decreased ca. fivefold, indicating that nucleotides 12–48 play a role in transcription efficiency. Nevertheless, transcription can take place with a completely truncated PrE-I. We then examined the requirement for PrE-II, the other element essential for replication. [AGP-GPd48dBB], which also lacks PrE-II, showed no further penalty in transcription than [AGP-GPd48] (Fig. 3c). In conclusion, significant transcription takes place from gs1 that lies downstream of AG/Prext and which is removed from the totality of the G/Pr sequences known to be essential for replication. Assuming that PrE-I and PrE-II together directly recruit vRdRp to initiate at the genome 39 end, it appears unlikely that a G/Pr devoid of these two elements can directly recruit vRdRp to initiate at gs1. The most likely explanation for GFP expression from the ectopic gs1 is that vRdRp is recruited to the template by AGPext before it can initiate at the downstream gs1. Presumably, trailer RNA synthesis would occur first, and vRdRp could then scan the template for the downstream gs1. Transcription downstream of G/Pr As transcription from gs1 can initiate downstream of AG/ Pr, we next asked whether the nature of the external replication promoter influenced this transcription. We therefore http://vir.sgmjournals.org Fig. 3. GFP expression in the absence of the PrE-I and PrE-II elements. (a) Schematic representation of the mini-replicons used as in Fig. 2(a). dBB, deletion of PrE-II element in the internal G/Pr. (b) A representative Northern blot analysis of the negative polarity encapsidated RNAs of the mini-replicons presented in (a) (rescue and purification as in Fig. 2b). (c) Relative GFP fluorescence averaged from three independent experiments as in Fig. 2(c). [AGP-GP] was taken as the series standard. replaced AG/Prext with G/Pr-58A, as this minimal mutation has little or no effect on promoter strength but strongly suppresses mRNA synthesis from gs1 so that transcription from this minimally mutated gs1 was suppressed (Le Mercier et al., 2003). The second series of constructs carrying G/Pr-58Aext is presented in Fig. 4(a). Fig. 4(b) illustrates their replication ability as measured by Northern blotting. Two points are noteworthy. First, the difference in migration of the single or double G/Pr mini-replicons is visible, and shows that only G/Prext is used for replication (compare lanes 1 and 3, with 2, 4 and 5), as shown before (Vulliémoz & Roux, 2002). Fig. 4(c) presents the level of GFP fluorescence normalized to replication relative to the single GP mini-genome. Comparison of [GP] and [GP58A] shows that the C58A mutation strongly inhibits GFP expression and has no deleterious effect on replication [Fig. 4(b), compare 1 and 3]. Second, the comparison of [GP-GP] and [GP58A-GPd12] shows, as in the case of the AG/Prext series, that G/Print deleted of its first 12 nt remains competent for transcription. This observation was confirmed by [GP58A-GPd1258A], in which GFP expression was suppressed upon mutation of the internal gs1 as well. By analogy with the AG/Prext constructs, G/Pr-58Aext of [GP58A-GPd12] can apparently recruit vRdRp that initiates at the genome 39 end, and which in this case leads to leader RNA synthesis. Upon release of leader RNA, vRdRp would Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 175 D. Vulliémoz and others Fig. 4. GFP expression downstream of G/Pr. (a) Mini-genomes used, portrayed as in Fig. 2(a). In this series, the external promoter is G/Pr or its mutated version. (b) A representative Northern blot analysis of the replicated RNAs of negative polarity of the mini-replicons portrayed in (a). The mini-replicons were rescued in BSR-T7 cells with the help of rSeV-AGP55 and the encapsidated RNAs purified as described in Fig. 2(b). Note the relative position of the mini-replicons DI-RNAs consistent with replication initiation at the external promoter only. (c) Relative GFP fluorescence averaged from three independent experiments as in Fig. 2(c). [GP] was taken as the series standard. again be free to scan the template for a downstream, ectopic gs1 (see also Discussion). AG/Pr and G/Pr are not equivalent in promoting GFP expression To examine whether the nature of the external replication promoter influenced the efficiency of GFP expression from an ectopic gs1, we directly compared two series of minireplicons which contain either AG/Prext or G/Pr -C58Aext (Fig. 5a). The three AGP constructs were amplified (20–50fold) better than their GP counterparts, as expected, as AG/Pr is known to be the stronger of the two replication promoters (Fig. 5b). Despite their lower level of replication, the G/Prext mini-replicons constantly exhibited a higher mean GFP fluorescence (data not shown). After 176 Fig. 5. Comparison of GFP expression downstream of AG/Pr or G/Pr in the presence of helper virus rSeV-AGP55. (a) Schematic representation of the mini-replicons used as in Fig. 2(a). (b) A representative Northern blot analysis of the negative polarity encapsidated RNAs of the mini-replicons presented in (a) (rescue and purification as in Fig. 2b). (c) Relative GFP fluorescence averaged from four independent experiments. [GP58A-GP] was taken as the series standard. (d) Northern blot analysis of the total non-encapsidated RNAs recovered from the CsCl gradient pellets as described in Methods, using a probe of negative polarity positioned in the GFP ORF. Lanes 1–6 as in (a). Lanes 7 and 8, rSeV-AGP55 and mock infected BSR-T7 cells, respectively. (e) As in (d), but after selection for the poly adenylated RNAs (see Methods). normalization to template levels, the G/Prext constructs expressed 20–50-fold more GFP than the AG/Prext counterparts (Fig. 5c). Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Journal of General Virology 86 Paramyxovirus replication and transcription Direct estimation of the various levels of GFP mRNA present intracellularly was also carried out, by Northern blotting and primer extension. The pellets of the CsCl gradients used to purify the encapsidated mini-replicons (Fig. 5b) contain the bulk of cellular and non-encapsidated viral RNAs. These CsCl pellet RNAs were first analysed by Northern blotting (using a GFP probe of negative polarity) either directly (Fig. 5d) or after oligo-dT selection (Fig. 5e). The specificity of this probe is underlined in both cases by the absence of signal for the RNAs isolated from uninfected cultures (lane 8), or infected by the helper virus alone (lane 7). The interesting point is that the level of GFP mRNAs detected is similar regardless of the nature of the external replication promoter (crude, Fig. 5d, and olig-dT selected, Fig. 5e, compare lanes 1 and 2 with 4 and 5). Given the 20–50-fold difference to which the templates for these messengers had accumulated intracellularly (Fig. 5b), it is clear that the G/Prext constructs produce significantly more GFP mRNAs (20–50-fold) per mini-genome than their AG/Prext counterparts. Primer extensions were also performed to precisely determine the position of the RNA 59 ends. For instance, the strong decrease in GFP expression due to the C58A mutation introduced in the internal G/Pr (Fig. 2, [GP58AGPd1258A]; Fig. 3, [AGP-GPd1258A]) suggests that the internal gs1 is the site of mRNA initiation. If so, this will be undoubtedly demonstrated by primer extension. Fig. 6(a) shows the results of extending a primer of negative polarity situated in the GFP ORF on the various CsCl pellet RNAs. A double band (presumably due to the presence of a 59 cap group) is present only at the transcription start site of the internal gs1. When primer extensions were performed with the encapsidated (CsCl band) RNAs, the 59 ends of the mini-antigenomes were found to be displaced according to the nucleotide deletion in G/Print (Fig. 6b). Identical results were obtained with the three constructs of the GP58A series (data not shown). These results confirm that GFP expression is due to mRNAs that have initiated at the ectopic gs1. GFP expression in the absence of the helper virus The above mini-replicon system uses a helper virus to provide the replication and transcription functions. As the various mini-replicons compete more or less effectively with the helper genome, the amounts of replication substrates available could vary and this could in some way affect the relative use of vRdRp as a transcriptase or replicase. It was therefore of interest to also examine the two series of mini-replicons of Fig. 5 when they were amplified and expressed by the N, P and L proteins derived from plasmids (Fig. 7). vRdRp availability here only depends on the extent of plasmid transfection. In contrast to the helper virus mediated mini-genome amplification (Fig. 5), there was only a two- to fivefold difference between the levels of AGPext and GPext mini-replicons amplified by plasmid-expressed N, P, and L (Fig. 7b). The absence of competition with the helper genomes is probably responsible for this difference, and vRdRp provided by plasmids may be less limiting as well. More importantly, the mean fluorescence of the GP58A-GP constructs was Fig. 6. Identification of the transcription and replication start positions. (a) Primer extension analysis using the non-encapsidated RNAs presented in Fig. 5(d) as templates. An infrared dye labelled (IRD) primer of negative polarity positioned in the GFP ORF was used (see Methods). Above and below, schematic representation of relevant portions of the two promoters present at the 39 end of the GP58A-GP and AGP-GP minireplicons, respectively. The corresponding portions of the nucleotide sequence, performed on plasmid DNA with the same IRD primer as the one used for primer extension, are presented and serve as size markers. In the sequencing gel, the 39 end extremities of the external promoters, and the gene starts (mutated, gs1m or not, gs1) are highlighted. (b) The encapsidated RNAs purified from CsCl gradients, as in Fig. 5(b), were used as templates for primer extensions using the IRD primer of negative polarity cells as in Fig. 5(a). In the sequencing gel, the 39 end extremities of the external promoters, and the gene starts (mutated, gs1m or not, gs1) are highlighted. http://vir.sgmjournals.org Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 177 D. Vulliémoz and others DISCUSSION We have prepared mini-replicons in which mRNA synthesis initiates not from its normal position gs156 within G/Pr, but at a gs1 downstream of the promoter used for replication, at position 146. In this manner, we have examined the minimum sequence requirements for this ectopic gs1 to function, and the effect of nearby cis-acting sequences on the frequency with which this ectopic gs1 is used to express GFP. There are two views of how non-segmented negativestranded vRdRp gains access to gs1. vRdRp can either directly interact with gs1 without having first initiated leader RNA, as recently proposed for vesicular stomatitis virus (VSV) (Whelan & Wertz, 2002). Alternatively, vRdRp can reach gs1 after having entered the nucleocapsid at its 39 end. Our results suggest that this latter mechanism applies for SeV. Although in most experiments gs1 is part of an internal replication-incompetent G/Pr (due to deleting the conserved first 12 nt), it is possible to delete up to the first 48 nt of G/Print and its PrE-II element as well without eliminating GFP expression (Fig. 3). In these latter cases, it is unlikely that there are sufficient cis-acting sequences remaining for SeV RdRp to enter the template directly at gs1, given that RNA initiation at the genome 39 end requires at least two essential sequence elements spread over 96 nt of G/Pr. If, on the other hand, vRdRp arrives at gs1 after having entered the nucleocapsid at the 39 end, gs1 of the highly deleted G/Print would still function. Fig. 7. Comparison of GFP expression downstream from AG/Pr or G/Pr in the presence of support plasmids. The minireplicons presented in Fig. 5(a) were replicated in the presence of the helper functions L, P and N expressed from plasmids (see Methods). (a) Mean GFP fluorescence expressed as the mean of data from duplicate samples, shown with variation of the mean. (b) Northern blot analysis of the purified encapsidated RNA as in Fig. 2(b). The cells of the duplicate samples presented in (a) were pooled. (c) Relative GFP expression corrected for template availability as measured in (b). [GP58AGP] was taken as the series standard. again higher than that of the AGP-GP constructs (Fig. 7a). After correction for template levels, G/Prext constructs expressed GFP four-to 40-fold more efficiently than corresponding AG/Prext constructs (Fig. 7c). 178 We found that the frequency with which gs1146 initiated mRNA synthesis appeared to depend, in an inverse fashion, on the strength of the upstream replication promoter. When a minimal (10 nt) gs1 was introduced into AG/Pr at position 56, this diminished the use of this 39 end promoter for replication (Le Mercier et al., 2003). In those experiments, there was an inverse relationship between the presence of gs1 and the relative strength of the replication promoter. In the present study, there was an inverse relationship between the relative efficiency of gs1 positioned downstream of the replication promoter and the relative strength of that promoter. Thus, when transcription and replication start sites were in close proximity, each form of viral RNA synthesis negatively affected the other. There are two relatively straightforward, but very different, interpretations for these observations. The first is that mRNA start sites and 39 end replication promoters compete for a common pool of vRdRp (Le Mercier et al., 2003). The second is that, given the proximity of these two RNA start sites, the interaction of a transcriptase with gs1 interferes with a replicase starting RNA synthesis from the genome 39 end. In this case, the apparent competition would be due to steric interference of the vRdRp that initiates mRNA synthesis with that which synthesizes RNA from the genome 39 end, and vice-versa. In mini-replicons, in which gs1 was displaced from position 56 to 68, the gene start site was equally effective in reducing Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Journal of General Virology 86 Paramyxovirus replication and transcription replication (Le Mercier et al., 2003). This limited displacement, which maintains gs1 between PrE-I and PrE-II of G/Pr, however, is probably insufficient to alter possible steric interference. In the present experiments the transcription start site of GP58A-GPd12 had been displaced 90 nt downstream to position 146 relative to the genome 39 end. Nevertheless, the apparent competition between gs1 and the 39 end replication promoter could still be observed. It is then less likely that this competition operates via some form of steric interference. In this case, it would appear that the two viral RNA start sites within G/Pr negatively influence each other by competing for a common pool of vRdRp. Remarkably, the competition between transcription and replication promoters occurred even when vRdRp and N protein were provided by plasmids in the absence of helper virus (Fig. 7). This competition thus appears to be direct, and not due to secondary effects affecting the provision of replication substrates that could occur during helper-virus mediated mini-genome expression. There is good evidence that the negative-stranded vRdRp can scan the nucleocapsid template for new gene start sites once they have released the mature mRNA (Stillman & Whitt, 1998; Fearns & Collins, 1999). These vRdRp may also be able to scan the template for gs1. If so, the competition we envision would occur in large part during this template scanning: the presence of a gs near the 39 end replication promoter would divert vRdRp from scanning back to the genome 39 end, and the presence of a strong 39 end promoter would disfavour vRdRp scanning to gs1. It will be of interest to displace a minimal gs1 progressively away from the genome 39 end, and to determine whether the intervening distance affects the efficiency of mRNA synthesis from gs1. One complication in interpreting our data is that we do not know how the C58A mutation inhibits GFP mRNA synthesis from gs1. It is possible that this mutation simply inhibits mRNA initiation, or it inhibits productive mRNA synthesis. Indeed, in a study of the transcription start signal of VSV, mutations were found that would allow initiation of mRNA synthesis without proper capping (Stillman & Whitt, 1999). In this case, the uncapped transcripts were prematurely terminated and degraded, a finding that could only be made in in vitro studies, and thus more difficult with SeV. Even in this case, however, SeV RdRp would, in effect, simply recapitulate leader RNA synthesis, i.e. synthesize a short, uncapped transcript without being committed to transcription. It is difficult to see how synthesis of an abortive transcript from gs156 would so strongly enhance productive mRNA synthesis from a downstream gs1. Moreover, when [GP58A-GPd12] and [GP-GP] were compared (Fig. 4), there was no evidence that the C58A mutation had enhanced expression from gs1146. Finally, we recently produced [GPgsm-GPd12] constructs in which nt 56–65 of GPext were all substituted with the corresponding nucleotides of AGP. These replicated two- to threefold better than their [GP58A-GPd12] counterparts, but generated http://vir.sgmjournals.org similar high mean GFP fluorescence (data not shown). In the end the more likely possibility is that vRdRp initiates at gs1146 only after releasing the short leader RNAs. It must then scan the template for a new RNA synthesis start site, and it is during this process that the various promoters compete with each other for scanning vRdRp. ACKNOWLEDGEMENTS We are indebted to Dominique Garcin, Philippe Le Mercier and Daniel Kolakofsky for critical discussions in the course of the work. We are moreover grateful to Daniel Kolakofsky for his substantial contribution to the writing of the paper. This work was supported by a grant from the Swiss National Foundation for Scientific Research and from the Commission of the European Communities specific RTD programme ‘Quality of Life and Management of Living Resources’, QLK2-CT200101225, ‘Towards the design of new potent antiviral drugs: structure and function analysis of Paramyxoviridae RNA polymerase’. It does not necessary reflect the views of the Commission and in no way anticipates future policy in this area. REFERENCES Bhella, D., Ralph, A., Murphy, L. B. & Yeo, R. P. (2002). Significant differences in nucleocapsid morphology within the Paramyxoviridae. J Gen Virol 83, 1831–1839. Buchholz, U. J., Finke, S. & Conzelmann, K.-K. (1999). Generation of bovine respiratory syncytial virus (BRSV) from cDNA: BRSV NS2 is not essential for virus replication in tissue culture, and the human RSV leader region acts as a functional BRSV genome promoter. J Virol 73, 251–259. Calain, P. & Roux, L. (1993). The rule of six, a basic feature for efficient replication of Sendai virus defective interfering RNA. J Virol 67, 4822–4830. Calain, P. & Roux, L. (1995). Functional characterisation of the genomic and antigenomic promoter of Sendai virus. Virology 212, 163–173. Calain, P., Curran, J., Kolakofsky, D. & Roux, L. (1992). Molecular cloning of natural paramyxovirus copy-back defective interfering RNAs and their expression from DNA. Virology 191, 62–71. Egelman, E. H., Wu, S.-S., Amrein, M., Portner, A. & Murti, G. (1989). The Sendai virus nucleocapsid exists in at least four different helical states. J Virol 63, 2233–2243. Fearns, R. & Collins, P. L. (1999). Model for polymerase access to the overlapped L gene of respiratory syncytial virus. J Virol 73, 388–397. Gubbay, O., Curran, J. & Kolakofsky, D. (2001). Sendai virus genome synthesis and assembly are coupled: a possible mechanism to promote viral RNA polymerase processivity. J Gen Virol 82, 2895–2903. Hoffman, M. A. & Banerjee, A. K. (2000). Precise mapping of the replication and transcription promoters of human parainfluenza virus type 3. Virology 269, 201–211. Iseni, F., Baudin, F., Garcin, D., Marq, J. B., Ruigrok, R. W. H. & Kolakofsky, D. (2002). Chemical modification of nucleotide bases and mRNA editing depend on hexamer or nucleoprotein phase in Sendai virus nucleocapsids. RNA 8, 1056–1067. Lamb, R. A. & Kolakofsky, D. (2001). Paramyxoviridae: the viruses and their replication. In Fields Virology, 4th edn, pp. 1305–1340. Edited by D. M. Knipe & P. M. Howley. Lippincott, Williams & Wilkins. Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 179 D. Vulliémoz and others Le Mercier, P., Garcin, D., Hausmann, S. & Kolakofsky, D. (2002). Ambisense Sendai viruses are inherently unstable but are useful to study viral RNA synthesis. J Virol 76, 5492–5502. Le Mercier, P., Garcin, D., Garcia, E. & Kolakofsky, D. (2003). parainfluenza virus type 2 follows the rule of six, and recombinant viruses recovered from non-polyhexameric-length antigenomic cDNAs contain a biased distribution of correcting mutations. J Virol 77, 270–279. Competition between the Sendai virus N mRNA start site and the genome 39-end promoter for viral RNA polymerase. J Virol 77, 9147–9155. Stillman, E. A. & Whitt, M. A. (1998). The length and sequence com- Mottet, G. & Roux, L. (1989). Budding efficiency of Sendai virus Stillman, E. A. & Whitt, M. A. (1999). Transcript initiation and 59-end nucleocapsids: influence of size and ends of the RNA. Virus Res 14, 175–187. modifications are separable events during vesicular stomatitis virus transcription. J Virol 73, 7199–7209. Murphy, S. K., Ito, Y. & Parks, G. D. (1998). A functional Tapparel, C., Maurice, D. & Roux, L. (1998). The activity of Sendai antigenomic promoter for the Paramyxovirus Simian virus 5 requires proper spacing between an essential internal segment and the 39 terminus. J Virol 72, 10–19. virus genomic and antigenomic promoters requires a second element past the leader template regions: a motif (GNNNNN)3 is essential for replication. J Virol 72, 3117–3128. Pelet, T., Delenda, C., Gubbay, O., Garcin, D. & Kolakofsky, D. (1996). Partial characterization of a Sendai virus replication Vulliémoz, D. & Roux, L. (2001). ‘Rule of Six’: how does the Sendai promoter and the rule of six. Virology 224, 405–414. Vulliémoz, D. & Roux, L. (2002). Given the opportunity, the Sendai Rager, M., Vongpunsawad, S., Duprex, W. P. & Cattaneo, R. (2002). virus RNA-dependent RNA polymerase could as well enter its template internally. J Virol 76, 7987–7995. Polyploid measles virus with hexameric genome length. EMBO J 21, 2364–2372. Skiadopoulos, M. H., Vogel, L., Riggs, J. M., Surman, S. R., Collins, P. L. & Murphy, B. R. (2003). The genome length of human 180 position of vesicular stomatitis virus intergenic regions affect mRNA levels and the site of transcript initiation. J Virol 72, 5565–5572. virus RNA polymerase keep count? J Virol 75, 4506–4518. Whelan, S. P. J. & Wertz, G. W. (2002). Transcription and replication initiate at separate sites on the vesicular stomatitis virus genome. Proc Natl Acad Sci U S A 99, 9178–9183. Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Thu, 03 Jan 2019 01:30:16 Journal of General Virology 86