Functional Analysis of the Noncoding Regions of the Uukuniemi Virus (Bunyaviridae) RNA Segments

Abstract

The role of the variable portion of the noncoding regions (NCRs) of the three Bunyaviridae RNA segments (L, M, S) in transcription, replication, and packaging was studied using the recently developed plasmid-driven RNA polymerase I minigenome system for Uukuniemi (UUK) virus, genus Phlebovirus (11), as a model. Comparison of the different segments showed that all NCRs were sufficient to mediate transcription/replication of a minigenome but demonstrated decreased promoter strength in the order M > L > S. Chimeric minigenomes with flanking NCRs from different genome segments revealed that the number of total base pairs within the inverted, partially complementary ends was important for transcription and replication. Point mutations increasing the base-pairing potential produced increased reporter expression, indicating that complementarity between the 5′ and 3′ ends is crucial for promoter activity. The role of the intergenic region (IGR) located between the two open reading frames of the ambisense UUK virus S segment was analyzed by inserting this sequence element downstream of the reporter genes. The presence of the IGR was found to enhance reporter expression, demonstrating that efficient transcription termination, regulated by the IGR, is important for optimal minigenome mRNA translation. Finally, genome packaging efficacy varied for different NCRs and was strongest for L followed by M and S. Strong reporter gene activity was still observed after seven consecutive cell culture passages, indicating a selective rather than random genome-packaging mechanism. In summary, our results demonstrate that the NCRs from all three segments contain the necessary signals to initiate transcription and replication as well as packaging. Based on promoter strength, M-segment NCRs may be the preferred choice for the development of reverse genetics and minigenome rescue systems for bunyaviruses.

Members of the Bunyaviridae family, which comprises more than 300 viruses grouped into the five genera (Orthobunyavirus, Hantavirus, Nairovirus, Phlebovirus, and Tospovirus), are enveloped viruses with a tripartite, single-stranded RNA genome of negative polarity. The L segment encodes the RNA-dependent RNA polymerase (L) (9), the M segment encodes the two glycoproteins (G_N and G_C) and, in some viruses, a nonstructural protein (NS_M) (25), and the S segment encodes the nucleoprotein (N) and, in some viruses, a nonstructural protein (NSs) (8, 26, 29).

Viruses belonging to this family share several common features with other negative-strand RNA viruses. The templates for viral polymerase-catalyzed transcription and replication are the ribonucleoproteins (RNP), which for bunyaviruses consist of the full-length RNA segments associated with the nucleoprotein N and the viral polymerase L. The genomic viral RNA (vRNA) segments contain genes in antisense orientation flanked by stretches of noncoding nucleotides (26). The regulatory elements for viral transcription and replication, as well as encapsidation and packaging signals, are thought to be located within these noncoding regions (NCRs) (26). The terminal nucleotides are genus specific and highly conserved, and because of their partial inverted complementarities they can form double-stranded regions leading to circular RNAs (17), providing the functional promoter region for interaction with the viral polymerase (12). The function of the remaining nucleotides of the NCRs is still not well understood, despite the fact that several reverse genetics and minigenome rescue systems for different bunyaviruses have recently been developed (3, 4, 7, 10, 11, 13, 18, 24). However, an encapsidation site has been characterized within the 5′ NCR of the S vRNA segment, using the Orthobunyavirus Bunyamwera as a model (21). Studies using influenza virus reverse genetics recently revealed a packaging signal within the highly conserved parts of the NCRs (30) and parts of the open reading frames (ORFs) (15, 32), as well as nucleotides which influence segment transcription levels (14).

A special feature of viruses belonging to the genus Phlebovirus is the ambisense S segment, which contains two genes located in different orientations: the N gene in antisense and the S RNA-encoded nonstructural protein (NSs) gene in sense orientation. Therefore, only the N gene can be transcribed directly from the genomic vRNA template, whereas the complementary replication intermediate RNA (cRNA) serves as template for NSs gene transcription (26). In Uukuniemi (UUK) virus (a phlebovirus) the two S-segment-encoded genes are separated by a 75-nucleotide (nt) intergenic region (IGR), which harbors transcription termination signals for the synthesis of the two subsegmental N and NSs mRNAs (29).

We used a recently developed reverse genetics system for the UUK virus (11) to analyze the function of the NCRs as well as the influence of the IGR on viral transcription and replication. Furthermore, we have compared the promoter activities as well as the packaging efficiencies of the L, M, and S vRNA-based minigenomes and investigated the influence of complementarity between segment NCRs on transcription and replication.

MATERIALS AND METHODS

Cells and virus.

BHK-21 cells (American Type Culture Collection) were grown on plastic dishes in Glasgow minimal essential medium (GMEM) supplemented with 5 to 10% fetal calf serum, 2 mM l-glutamine, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml (Invitrogen/Life Technologies).

Construction of plasmids.

Pol I-driven UUK virus minigenome constructs were generated by using pRF108 (12) as a murine (m) pol I promoter- and terminator-containing plasmid for inserting different PCR fragments. This plasmid was used to deliver the pol I transcription cassette (Fig. 1).

FIG. 1.

Schematic representation of the different UUK virus minigenomes used for reverse genetics studies. PCR-amplified UUK virus minigenome fragments were inserted into pRF108, a plasmid containing the murine pol I promoter and terminator for pol I-driven transcription (12). To facilitate readout of minigenome expression levels, the UUK virus RNA segment ORFs were replaced with the gene encoding CAT (left panel) or GFP (right panel). The reporter genes are flanked by the NCRs of the UUK virus S, M, and L segments. The lengths of the 5′ and 3′ NCRs are presented in Fig. 3A. Insertion of the intergenic region sequence is marked with IGR.

UUK virus S-reporter gene constructs.

For construction of S-segment-based minigenomes, primers containing the complete 5′ (25 nt, positions 1 to 25) and 3′ (34 nt, positions 1687 to 1720) vRNA NCRs of the UUK virus S segment were used to amplify the chloramphenicol acetyltransferase (CAT) or green fluorescent protein (GFP) reporter genes in antisense (CAT, primers RF289 and RF291; GFP, primers RF290 and RF292) or sense (CAT, primers RF294 and RF295; GFP, primers RF296 and RF297) orientation (Fig. 2). PCR fragments were inserted into the BsmBI/BbsI fragment of the pRF108 (12) plasmid (Fig. 1). This resulted in four constructs, two with an antisense (−) reporter gene (pRF287, pol I [m] UUK S-CAT [−]; pRF288, pol I [m] UUK S-GFP [−]) and two with a reporter gene in sense (+) orientation (pRF289, pol I [m] UUK S-CAT [+]; pRF290, pol I [m] UUK S-GFP [+]). It should be noted that all generated minigenomes, independent of reporter gene orientation, contained similar Kozak sequences (purine at positions −3 and +1) to avoid differences in translation initiation efficiencies (pRF287 S-CAT [−], AGATCATGG; pRF289 S-CAT [+], TAAGCATGG; pRF288 S-GFP [−], AGATCATGG; pRF290 S-GFP [+], TAAGCATGG [CAT gene start codons are underlined]).

FIG. 2.

Oligonucleotide primers used to construct plasmids. Italics, UUK-specific sequences; shading, reporter or viral gene start and stop codons; boldface, restriction endonuclease recognition sequences; underlining, introduced nucleotide changes.

UUK virus S-reporter gene constructs containing the IGR.

For further analyses of the UUK virus S-segment NCRs, the IGR sequence (UUK virus S vRNA, positions 848 to 921) was inserted into the four different UUK virus S-reporter gene minigenomes as described above (pRF287, pRF288, pRF289, and pRF290). PCR primers were designed containing the 5′ UUK virus S NCR (RF327) and the 3′ UUK virus S NCR (RF326) adjacent to the IGR sequence, respectively (Fig. 2). By using the BbsI-restricted PCR fragment RF327/291 or RF326/RF294 for insertion into the BsmBI/BbsI fragment of the pRF108 vector (Fig. 1), CAT reporter gene-containing minigenome constructs were generated and named pRF312 (pol I UUK virus S-CAT_IGR [−]) and pRF310 (pol I UUK virus S-CAT_IGR [+]), respectively. Similar constructs containing the GFP reporter gene flanked by the UUK virus S-segment NCRs and the IGR sequence were generated by combining BbsI-restricted PCR product RF327/292 or RF326/RF296 with the BsmBI/BbsI fragment of pRF108, resulting in pRF313 (pol I UUK virus S-GFP [−] IGR) and pRF311 (pol I UUK virus S-GFP [+] IGR), respectively.

UUK virus L-reporter gene constructs.

For analyzing UUK virus minigenomes encoding reporter genes flanked by the UUK virus L-segment NCRs, the 5′ (95 nt, vRNA positions 1 to 95) and 3′ (18 nt, vRNA positions 6406 to 6423) NCRs were added to the CAT or GFP reporter genes in two consecutive PCRs. Primers used in the first PCR provided the complete 3′ UUK virus L-NCR (CAT, RF299; GFP, RF301) and the last 35 nt of the 5′ UUK virus L-NCR (CAT, RF316; GFP, RF317), whereas the second PCR, using the first amplification product PCR RF316/RF299 or PCR RF317/RF301 as a template (diluted 1:100), added the remaining first 60 nt of the 5′ UUK virus L-NCR (CAT, RF298B/RF299; GFP, RF298B/RF301). Insertion of the BbsI-restricted PCR products into the BsmBI/BbsI fragment of pRF108 resulted in two different UUK virus L-segment-based minigenome constructs, where the L ORF is exactly replaced by either the CAT or GFP reporter genes pRF293 (pol I UUK virus L-CAT) and pRF294 (pol I UUK L-GFP).

Chimeric UUK virus minigenomes.

For analyzing UUK virus minigenomes containing NCRs derived from different genome segments, CAT gene-containing UUK virus L, M, and S minigenome constructs pRF293, pRF200, and pRF312, respectively, were treated with EcoRI and NheI restriction endonucleases, resulting in two fragments of approximately 3,350 bp (containing the plasmid backbone, including the pol I promoter and the 5′ NCR) and 450 bp (containing the pol I terminator and the 3′ NCR) in length. By combining the larger fragment from one plasmid with the smaller fragment from another construct, UUK virus minigenomes with NCRs derived from different UUK virus RNA segments were generated: pRF367 (UUK S-CAT-M), pRF368 (UUK M-CAT-S), pRF369 (UUK L-CAT-M), pRF370 (UUK M-CAT-L), pRF371 (UUK L-CAT-S), and pRF372 (UUK S-CAT-L).

Transfection and superinfection with UUK virus.

BHK-21 and BSR cell lines were seeded in 6-cm-diameter tissue culture dishes and were transfected with the different minigenome plasmid DNAs using Lipofectamine 2000 reagent (Invitrogen/Life Technologies). Transfections were performed as described previously (12). To determine the efficiency of transfection, the plasmid pHL2823 containing an enhanced GFP under the control of the cytomegalovirus (CMV) promoter (R. Flick and G. Hobom, unpublished data) was similarly transfected. Transfected cells were superinfected 24 h posttransfection with UUK virus at a multiplicity of infection (MOI) of 1 to 3. Briefly, 150 μl of virus stock, diluted in 350 μl of cell culture medium (without fetal calf serum [FCS] and antibiotics), was incubated for 1 h in 6-cm-diameter cell culture plates (37°C, 5% CO₂). Unadsorbed virus was removed and 5 ml of cell culture medium (with 5% FCS and antibiotics) was added.

Passaging of recombinant UUK virus.

BHK-21 cells were transfected as described above and were superinfected 24 h later with UUK virus at an MOI of 1 to 3. Cells were analyzed for reporter gene expression 72 hours postinfection (hpi), and the corresponding supernatants were used for virus passaging. Cell debris was removed by centrifugation at 3,000 × g for 10 min, and cells (approximately 3 × 10⁶ BHK-21) were infected with 2 ml or 200 μl of undiluted supernatant. After a 1-h incubation period (37°C, 5% CO₂) the inoculum was replaced by fresh medium (GMEM, 5% FCS, antibiotics) and cells were incubated for 72 h. This procedure was repeated as indicated in successive passages.

CAT assays.

Cell extracts were prepared as described by Gorman et al. (16), and CAT activity was determined using a commercially available Flash Cat kit (Molecular Probes, Eugene, Oreg.) as described previously (11, 12). The reaction products were visualized by UV illumination, documented by photography, and evaluated using WinCam software (Cybertech, Berlin, Germany) or Quantity One (Bio-Rad). Ratios of activities were calculated based on at least three independent sets of serial dilutions of cell lysates down to a level of 30 to 50% product formation for better quantification within a linear range.

Cell fixation and UV microscopy.

Cells transfected with GFP-containing minigenome constructs and either cotransfected with pCMV UUK-L and pCMV UUK-N (expression plasmids under CMV control for UUK virus L and N proteins [11]) or superinfected with UUK virus were fixed with 4% paraformaldehyde. GFP expression was visualized using an Axioplan 2 microscope (Zeiss) and documented using a 3CDC color video camera (DXC-970MD; Sony) and the imaging software Northern Eclipse 6 (Empix Imaging, Inc.). Alternatively, cells were trypsinized before fixation with 4% paraformaldehyde and were analyzed using fluorescence-activated cell sorting (FACSCalibur; Becton Dickinson).

RESULTS

Analysis of the terminal NCRs of the ambisense UUK virus S segment.

To analyze the NCRs of the ambisense UUK virus S segment which are responsible not only for the S segment replication but also for transcription of the N and the NSs genes, minigenomes containing the CAT or GFP reporter gene in either sense or antisense orientation and replacing either the NSs or N ORF were generated (Fig. 1). This allowed us to study the cis-acting sequences for the two S-segment-carrying genes separately. In a first series of experiments, constructs were generated where the CAT or GFP reporter gene was inserted in sense or antisense orientation between the 25 nt of the 5′ NCR and the 34 nt of the 3′ NCR of the UUK virus S vRNA. Antisense-oriented reporter gene constructs (UUK S-CAT [−], pRF287; UUK S-GFP [−], pRF288) (Fig. 1) simulate transcription of the N gene (vRNA→ [cRNA → vRNA] → mRNA), whereas transcription of the NSs gene (vRNA → cRNA → mRNA) was analyzed using constructs containing sense-oriented reporter genes (UUK S-CAT [+], pRF289; UUK S-GFP [+], pRF290) (Fig. 1). Pol I-driven (20, 33) minigenome constructs were transfected into BHK-21 cells and were cotransfected with pCMV UUK-L and pCMV UUK-N, which provide the viral L and N proteins. After pol I transcription in the nucleus, the artificial UUK virus minigenomes are transported to the cytoplasm, where they are encapsidated by the N protein and are replicated and transcribed by the L polymerase. The level of reporter gene expression reflects the efficiency of transcription and replication from the promoter regions located in the flanking NCRs. The analysis of all tested minigenome constructs resulted in reporter gene expression (Fig. 3A, panels c, d, g, and h; B, columns 3, 5, 7, and 9; insert, lanes 2 and 6), confirming that the terminal NCRs of the S RNA contain all regulatory elements for encapsidation, replication, and transcription of viral genome S segments. No reporter gene activity could be detected by omitting the UUK-L and UUK-N expression plasmids (Fig. 3A, panels a and b; insert, lanes 1 and 5). Similarly, CAT and GFP reporter gene expression levels demonstrated equal promoter activity of the 3′ cRNA (5′ vRNA) NCR (pRF289/pRF290), responsible for the NSs gene transcription start, compared to the 3′ vRNA NCR (pRF287/pRF288) serving as the transcription start point for the N gene (Fig. 3A, panels c and d versus g and h; B, column 3 versus 5 and column 7 versus 9; insert, lane 2 versus 6).

FIG. 3.

Role of the intergenic region for minigenome expression levels. Different UUK virus S-segment-based minigenome constructs were analyzed for reporter gene expression to define the role of the IGR of the ambisense UUK virus S segment. Minigenomes were transfected into BHK-21 cells and were analyzed 24 or 48 h posttransfection for CAT or GFP reporter gene expression, respectively. (A) Analyses of UUK virus GFP minigenome expression by FACS. Transfection efficiency was determined, and average GFP intensity was measured. In the dot plots each dot represents 1 out of 10,000 analyzed cells; signals in the lower left corner represent the cells without GFP expression, and signals in the upper left corner represent the GFP-expressing cells. In the FACS histogram the y axis represents the cell numbers (%) and the x axis represents the GFP intensities. (B) Quantification of GFP and CAT expression. Column 1, reporter gene background activity in BHK-21 cells upon pRF311 (UUK S-GFP_IGR) transfection; column 2, reporter gene background activity after transfection of pRF310 (UUK S-CAT_IGR); columns 3 to 6, average GFP intensities after transfection with different GFP-containing UUK virus S-segment-based minigenomes and cotransfected UUK-L and -N expression plasmids; columns 7 to 10, CAT activity in BHK-21 cells transfected with different CAT-containing UUK virus S-segment-based minigenomes and cotransfected with pCMV UUK-L and pCMV UUK-N. Insert lanes 1, 3, 5, and 7 show CAT reporter gene background activity after transfection with pol I-driven UUK virus S-CAT minigenomes, omitting the UUK-L and -N expression plasmids. Lanes 2, 4, 6, and 8 of the insert show CAT activity after transfection with the different UUK virus S-CAT constructs cotransfected with pCMV UUK-L and -N plasmids.

Analysis of the IGR.

To determine the role of the IGR in the transcription and replication processes of the UUK virus S-derived minigenomes, the 75-nt-long sequence located between the two S-segment-encoded N and NSs genes was inserted into the UUK virus S-segment-based minigenome constructs described above (pRF287, pRF288, pRF289, and pRF290). For antisense constructs the IGR was inserted exactly between the stop codon of the reporter gene and the 5′ vRNA NCR (Fig. 1). For UUK virus minigenomes with a reporter gene in sense orientation, the IGR was inserted between the 3′ vRNA NCR and the stop codon of the reporter gene (Fig. 1). This resulted in two CAT-containing (UUK S-CAT_IGR [−], pRF312; UUK S-CAT_IGR [+], pRF310) and two GFP-containing (UUK S-GFP_IGR [−], pRF313; UUK S-GFP_IGR [+], pRF311) minigenome constructs. Cotransfection of each of these plasmids with the expression plasmids for the UUK virus L and N proteins (pCMV UUK-L and pCMV UUK-N) led to reporter gene expression, which was compared to the corresponding constructs without the inserted IGR sequence (Fig. 3). No reporter gene expression could be detected in the absence of cotransfected UUK virus L and N expression plasmids (Fig. 3B, lanes 1 and 2; insert, lanes 3 and 7), confirming that there is no background for any of the different IGR-containing minigenomes. For CAT analysis, cells were harvested 24 h posttransfection and CAT activity was determined from cell lysates. IGR-containing minigenomes both showed a strong increase in reporter gene expression levels compared to that of corresponding constructs without the IGR sequence for the antisense- and the sense-oriented reporter gene constructs (Fig. 3B, column 7 versus 8 and column 9 versus 10; insert, lane 2 versus 4 and lane 6 versus 8).

GFP expression levels, determined 48 h posttransfection by using UV microscopy and FACS analysis, were significantly higher (average GFP intensity was 3 to 6 times stronger) after transfection of IGR-containing minigenomes compared to that of constructs lacking the IGR sequence (Fig. 3A, panel c, panel d versus e, f, and g, panel h versus i, and panel j; B, column 3 versus 4 and column 5 versus 6). These results suggest that gene expression from UUK virus minigenomes based on the ambisense S segment is improved by inserting the IGR sequence downstream of the ORF.

Comparison of promoter activities between the three UUK virus genome segment NCRs.

In order to compare the activity and efficiency of cis-acting elements within the terminal NCRs of the three UUK virus RNA genome segments (Fig. 4A), we generated eight different pol I-driven UUK virus minigenomes by replacing the viral genes N, NSs, M, and L with CAT or GFP reporter genes (Fig. 1, pRF200 [M-CAT] [11], pRF312 [S_N-CAT], pRF310 [S_NSs-CAT], pRF293 [L-CAT], pRF31 [M-GFP] [11], pRF313 [S_N-GFP], pRF311 [S_NSs-GFP], pRF294 [L-GFP]). Each UUK virus minigenome construct was transfected into BHK-21 or BSR cells, and cells were cotransfected with the CMV-driven expression plasmids pCMV UUK-L and pCMV UUK-N. Cells transfected with the CAT-containing minigenomes were harvested 24 h posttransfection, and CAT activity was assayed from cell lysates (Fig. 4B, left panel). In the case of UUK virus GFP minigenomes, cells were fixed 48 h posttransfection with paraformaldehyde and GFP expression was determined using UV microscopy (data not shown) and FACS analysis (Fig. 4B, right panel). Transfection with all tested minigenome constructs resulted in reporter gene expression, confirming that the NCRs of each segment contain all regulatory elements for encapsidation, replication, and transcription of the artificial viral genome segments. Similar CAT activities were measured for each construct in BHK-21 versus BSR cells (comparison not shown). No reporter gene activity could be detected by omitting the UUK-L and UUK-N expression plasmids (data not shown).

FIG. 4.

Comparison of promoter strengths of the three UUK virus RNA segments. RNA pol I-driven UUK virus minigenome constructs containing the CAT or GFP reporter gene flanked by the NCRs of the three different RNA genome segments L, M, and S were transfected into BHK-21 cells. Cotransfection of expression plasmids pCMV UUK-L and pCMV UUK-N provides the L and N proteins required for transcription and replication of the minigenomes. (A) Schematic representation of the terminal nucleotides of UUK virus genome segments. Highly conserved nucleotides within all UUK virus genomic segments are indicated. Underlined nucleotides represent the gene start codon (antisense) in the L and M vRNAs. Nucleotide numbers within the flanking 5′ or 3′ NCRs are shown to the right. Calculated free energy values (kilocalories/mole) of predicted secondary structures are shown in the column on the right. (B) UUK virus CAT and GFP minigenomes. Twenty-four or 48 h posttransfection cells were harvested and analyzed for CAT activity or GFP expression, respectively. Lanes 1 to 4, CAT activity in BHK-21 cells after cotransfection of different UUK virus CAT minigenome constructs and UUK virus L and N expression plasmids; lanes 5 to 10, comparison of CAT activity from the plasmids S_(NSs) (pol I UUK S-CAT_IGR [+]) and S_(N) (pol I UUK S-CAT_IGR [−]) to determine promoter activities of the ambisense UUK virus S segment for NSs versus N transcription by using dilution series of extracts from BHK-21-transfected cells; lanes 11 to 14, FACS analysis of cells 48 h after transfection with different UUK virus GFP minigenomes. Average GFP intensities (from three independent experiments) of BHK-21 cells after cotransfection of UUK virus GFP minigenome constructs and UUK virus L and N expression plasmids are presented. (C) Time course experiment. BHK-21 cells were transfected with UUK virus genome segment-based reporter plasmids and were cotransfected with UUK virus L and N expression plasmids. Cells were harvested at the indicated time points posttransfection and were analyzed for CAT activity. BHK-21 cells only transfected with minigenome served as a negative control (MOCK).

The UUK virus CAT-containing reporter gene flanked by the UUK virus M segment NCRs (pRF200) showed the highest reporter gene expression level compared to those of the CAT reporter flanked by the L (pRF293) and S segment NCRs (pRF310 and pRF312) (Fig. 4B, lane 1 versus 2, 3, and 4). The L NCRs displayed a higher promoter activity compared to that of the UUK virus S-CAT minigenomes (Fig. 4B, lane 2 versus 3 and 4).

These results were confirmed by using the UUK virus minigenomes containing GFP as the reporter gene. The M NCR-flanked GFP minigenome (pRF31) clearly showed the strongest GFP expression level, followed by the UUK virus L-GFP (pRF294) and the UUK virus S-GFP constructs (pRF311 and pRF313) (Fig. 4B, lane 11 versus 12 versus 13 and 14). It is noteworthy that the different expression levels from the analyzed UUK virus minigenomes were not based on different pol I minigenome transcription levels (data not shown).

To determine the expression kinetics for each UUK virus segment-based minigenome, a time course experiment was performed. Pol I-driven minigenome plasmids were transfected together with UUK-L and -N expression plasmids. Cells were harvested at 10 to 70 h posttransfection, and CAT activity was determined from cell lysates (Fig. 4C). At 10 h posttransfection M- and L-segment-based minigenomes showed very weak CAT activities, whereas no reporter gene activity could be detected at this time for the two S-segment-based constructs. A strong increase in CAT expression could be observed 22 h posttransfection. Minigenomes flanked by the L- or S-segment NCRs reached their maximum expression levels at 28 h, whereas the M minigenome expression did not reach a plateau until 46 h, after which it remained high up to 70 h posttransfection (Fig. 4C).

For a more accurate comparison of the S-NCR-derived minigenomes, dilution series of minigenome-transfected cell lysates were used and analyzed for CAT expression. A minor but significant difference between the promoters responsible for the N or NSs gene expression could be observed (Fig. 4B, middle panel, lane 5 versus 6, lane 7 versus 8, and lane 9 versus 10), demonstrating that the S-segment-based minigenome with the N gene replaced by a CAT gene is somewhat more efficiently expressed than the constructs in which the NSs gene is replaced.

Taken together, comparison of UUK virus minigenomes based on all three different UUK virus genome segments revealed that the strongest promoter activity is located within the 185 nt of the 5′ NCR and 17 nt of the 3′ NCR of the UUK virus M segment.

Passaging of recombinant UUK viruses.

We next analyzed whether the minigenomes could be packaged into infectious UUK virus particles and serially passaged to fresh cell cultures. As was recently described for the M-segment-based minigenome (11), we expected that the reporter gene expression level would decrease with the number of transfers, due to the fact that there is no pressure for the virus to keep an additional reporter gene-encoding segment.

To determine the influence of the different NCRs on packaging efficiency, BHK-21 cells were transfected with pol I-driven minigenomes containing the NCRs of any of the three genomic RNA segments and were cotransfected with UUK-L and UUK-N expression plasmids. To provide the necessary packaging machinery, cells were superinfected with UUK virus 24 h posttransfection and aliquots (2 ml or 200 μl) from supernatants were transferred 72 hpi to fresh cell cultures. From the transfected and superinfected cells, as well as from each consecutive passage, reporter gene activity was determined at 72 hpi. This was repeated six times, resulting in a total of seven passages.

All minigenomes could be efficiently rescued by serially passaging tissue culture medium once (Fig. 5A, lanes 2, 6, 10, and 14), demonstrating that cis-active signals responsible for RNA incorporation into UUK progeny viruses are located within the flanking NCRs of all three genomic RNA segments. Minigenome constructs based on the M (pRF200) or S segment (pRF310 and pRF312) showed rapidly decreasing levels of reporter gene expression during the successive transfers (Fig. 5A, lanes 1 to 4, 9 to 12, and 13 to 16, respectively; see also panel B). After three passages, only weak CAT activities were detectable for pol I UUK M-CAT (for pRF200, see Fig. 5A, lane 4, and B) and pol I UUK S_NSs-CAT (for pRF310, see Fig. 5A, lane 12, and B), whereas no reporter gene expression was measurable for pol I UUK S_N-CAT (for pRF312, see Fig. 5A, lane 16, and B). In contrast, the L-segment-based artificial vRNA (pol I UUK L-CAT [pRF293]) showed slightly increased CAT activities during the first four passages (Fig. 5A, lanes 5 to 8, and B), indicating a more efficient packaging event for this minigenome than for that of the M- and S-segment-based RNAs. Even after seven passages, about 75% of the CAT reporter expression level of the primary transfected and superinfected cells was observed (Fig. 5B).

FIG. 5.

Analysis of CAT activity in BHK-21 cells after serial passages of supernatants containing recombinant UUK virus. BHK-21 cells were transfected with pol I-driven minigenomes based on all three genomic RNA segments [M, pRF200; L, pRF293; S_(NSs), pRF310; and S_(N), pRF312] and were cotransfected with UUK-L and UUK-N expression plasmids. Cells were superinfected 24 h posttransfection with UUK virus and aliquots (A, 2 ml; B, 2 ml and 200 μl) from supernatants were transferred 72 hpi to fresh cell cultures. From the transfected and superinfected cells and from each successive passage reporter gene activity was determined 72 hpi. This was repeated six times, resulting in a total of seven passages (p1 to p7). (A) Lanes 1, 5, 9, and 13, CAT activity after transfection of different pol I-driven minigenome constructs together with UUK-L and -N expression plasmids. All reporter gene activities were set to 100% to compare signal changes after transfer to fresh cell cultures. Lanes 2 to 4, reporter gene analysis (p1 to p3) of recombinant M-CAT UUK virus; lanes 6 to 8, reporter gene analysis (p1 to p3) of recombinant L-CAT UUK virus; lanes 10 to 12, reporter gene analysis (p1 to p3) of recombinant S_(NSs)-CAT UUK virus; lanes 14 to 16, reporter gene analysis (p1 to p3) of recombinant S_(N)-CAT UUK virus. (B) Comparison of CAT activity after passaging minigenome-containing UUK virus. CAT activities were compared between each passage of transfected and UUK virus-superinfected cells. Reporter gene expression was compared to the first transfected and superinfected cultures (set to 100%).

In summary, the passaging experiments demonstrated that cis-acting signals responsible for the incorporation of RNA segments into UUK virus particles are located within the two flanking NCRs of each genomic segment but that there are clear differences in packaging efficiencies (L > M > S_NSs > S_N).

Chimeric UUK virus minigenomes.

To further examine the basis for the differing promoter activities within the three genomic RNA segments, we generated chimeric constructs with the reporter gene flanked by NCRs derived from different UUK virus segments. The CAT gene-containing UUK virus L, M, and S minigenome constructs pRF293, pRF200, and pRF312, respectively, were treated with EcoRI and NheI restriction endonucleases, resulting in two fragments approximately 3,350 and 450 bp in length. By combining the larger fragment from one plasmid with the smaller fragment from another construct, the following UUK virus minigenomes with NCRs derived from different UUK virus RNA segments were generated: pRF367 (UUK S-CAT-M), pRF368 (UUK M-CAT-S), pRF369 (UUK L-CAT-M), pRF370 (UUK M-CAT-L), pRF371 (UUK L-CAT-S), and pRF372 (UUK S-CAT-L) (Fig. 6A). These plasmids were individually transfected into BHK-21 cells and cotransfected UUK-L and -N expression plasmids (pCMV UUK-L and pCMV UUK-N). Comparison of these chimeric UUK virus minigenomes revealed that combining NCRs derived from different UUK virus segments resulted in an almost complete loss of reporter gene expression (Fig. 6B, lanes 4 to 9) compared to that of the three wild-type UUK virus constructs (Fig. 6B, lanes 1 to 3). No significant difference in CAT expression from the various chimeric constructs could be determined.

FIG. 6.

Analysis of chimeric UUK virus minigenomes. (A) Schematic representation of chimeric UUK virus minigenomes. Highly conserved nucleotides within all UUK virus genomic segments are indicated. Underlined nucleotides represent the CAT gene start codon (antisense) in the L and M vRNAs. Potential base pair numbers within the partial complementary inverted terminal nucleotides are shown to the right. Calculated free energy values (kilocalories/mole) of predicted secondary structures are shown in the column on the right. (B) Reporter gene analysis. Chimeric constructs with CAT reporter genes flanked by NCRs derived from different UUK virus segments were transfected into BHK-21 cells together with expression plasmids pCMV UUK-L and pCMV UUK-N. CAT activities were analyzed 48 h posttransfection. Lanes 1 to 3, CAT activities of control plasmids (M-, L-, and S-segment-based UUK virus minigenomes) for direct comparison with chimeric minigenome constructs. Lanes 4 to 9, reporter gene activity after transfection of different chimeric UUK virus CAT-minigenome constructs.

For structural comparison, the first and last 25 nt of the chimeric minigenome sequence, connected by a stretch of 20 uracils, were used as a basis for secondary structure prediction (GeneBee) (5, 6). The calculated total free energy (kilocalories/mole) revealed a substantially reduced predicted stability of the chimeric constructs compared to that of the wild-type NCRs (Fig. 4A and 6A). The first 20 nt of both vRNA ends, which play a major role during the transcriptional start, can form 18 to 20 bp for the UUK virus L, M, and S segments (Fig. 4A). However, the NCRs of the chimeric minigenomes have the potential for only 10 to 16 bp (Fig. 6A). The missing interaction of nucleotides from the segment ends can cause dramatic changes in the secondary structure of the NCRs. In particular, the predicted length of stem structures and interrupting loops (data not shown) and, therefore, the corresponding free energy values (kilocalories/mole) of the RNA secondary structures differ substantially between the chimeric and wild-type minigenomes (Fig. 4A and 6A). Based on these analyses we introduced point mutations into the first or last 20 nt of each chimeric minigenomes by using oligonucleotide-directed mutagenesis to increase the number of potential base pairs within the terminal segment regions. We either introduced multiple mutations (exchanges and deletions) into the 5′ NCR, adjusting the sequence to that of the 3′ NCR of the chimeric construct, or vice versa (Fig. 7A). The resulting constructs contained 16 to 20 potential base pairs within the first 20 terminal nucleotides (Fig. 7A). Cotransfection of these pol I-driven chimeric minigenome plasmids with restored base pairing, together with expression plasmids encoding the viral L and N proteins, resulted in substantially increased reporter gene expression for three of the constructs compared to that of the original chimeric constructs containing 5′ NCR mutations (Fig. 7B, lanes 2, 4, and 6). Therefore, adjusting the two ends of the chimeric constructs in order to increase the potential for base pairing led to more efficient minigenome expression. However, minigenomes with mutations within the 3′ NCRs showed no change in rescue efficiency compared to those of the corresponding basic chimeric constructs (Fig. 7B, lanes 3, 5, and 7).

FIG. 7.

Analysis of chimeric UUK virus minigenomes with restored complementary ends. (A) Schematic representation of mutated chimeric UUK virus minigenomes. Point mutations were introduced into different chimeric UUK virus minigenomes using oligonucleotide-directed mutagenesis (RF508, RF512, and RF514 for 5′ vRNA modifications; RF509, RF513, and RF515 for 3′ vRNA modifications) (Fig. 2) to increase the total number of base pairs within the terminal nucleotides. Highly conserved nucleotides within all UUK virus genomic segments are highlighted. Underlined nucleotides represent the CAT gene start codon (antisense) in the L and M vRNAs. Introduced point mutations are marked with a box. (B) Reporter gene analysis. BHK-21 cells were transfected with the chimeric reporter plasmids and cotransfected with UUK-L and -N expression plasmids. Reporter gene activity was analyzed 48 h posttransfection. Lane 1, CAT activity of cells transfected with S-segment-based UUK virus minigenomes (set at 100%); lanes 2, 4, and 6, chimeric UUK virus minigenomes with point mutations introduced in the 5′ NCRs to restore base pairing with the 3′-terminal nucleotides; lanes 3, 5, and 7, chimeric UUK virus minigenomes with nucleotide changes in the 3′ NCR to restore base pairing with 5′-terminal nucleotides.

DISCUSSION

NCRs of segmented negative-stranded RNA viruses contain cis-acting sequences important for regulating RNA encapsidation, transcription, replication, and packaging of the viral genome segments. The analysis of these sequences and the determination of the location and nature of cis-regulating elements will help in understanding the molecular mechanisms by which a segmented RNA genome is processed during the infectious life cycle. Furthermore, more detailed knowledge about these processes could reveal strategies for the development of specific and efficient antiviral approaches.

The NCRs of Bunyaviridae genomes can be separated into three parts: (i) the highly conserved terminal nucleotides, which play an important role as a promoter element(s) for viral polymerase binding, transcription, and replication initiation (12, 26); (ii) the remaining portion of the NCRs which flank the ORFs and are thought to contain a nucleoprotein encapsidation signal(s) (21), regulatory sequences for high/low and early/late expression of the corresponding viral gene(s), and packaging signal(s) for the incorporation of genomic segments into virus particles (11); and (iii) the IGR of the ambisense S segment, which provides transcription termination signals (27).

To further dissect the function of the NCRs of Bunyaviridae genus members, we have used a recently developed pol I-based reverse genetics system for the UUK virus (11). Minigenomes of UUK virus, a model for phleboviruses, under the control of the eukaryotic RNA pol I, are transcribed into vRNA-like RNA molecules, which are then recognized and used as templates by the viral N and L proteins for encapsidation, transcription, replication, and packaging.

Different rescue efficiency of the three UUK virus segment-based minigenomes.

Reporter genes flanked by any of the three different genomic segment NCRs were all expressed after transfection of pol I-driven minigenomes into BHK-21 cells expressing the viral L and N proteins (Fig. 4). This demonstrated that all necessary signals for RNA encapsidation, transcription, and replication are located within these flanking NCRs. The use of minigenomes of similar lengths facilitated the comparison of expression efficiencies (promoter strengths) between the three segment-derived RNAs without the interference of various RNA sizes. The M-segment-based constructs showed the strongest reporter gene expression (CAT and GFP) followed by the L-derived constructs, whereas the rescue efficiencies of minigenomes containing S-segment NCRs were the lowest (Fig. 4B). This result is surprising considering that the N protein, a gene product of the S segment, is the most abundant protein in UUK virus-infected cells (31). The reason for this result could be partly due to the different lengths of the wild-type genomic segments (L, 6,423 bp; M, 3,229 bp; S, 1,720 bp) compared to the sizes of the analyzed minigenomes (L-CAT, 772 bp; M-CAT, 863 bp; S-CAT, 800 bp; L-GFP, 854 bp; M-GFP, 947 bp; S-GFP, 884 bp). In the wild-type RNA segments the overall length could influence viral gene expression, as shown in similar studies for influenza virus (M. Azzeh, R. Flick, and G. Hobom, unpublished data). Our results here suggest that the longer M and L segments need stronger promoter activities than the S segment to compensate for their larger size. Interestingly, in a time course experiment comparison of reporter gene expression efficiency revealed that M minigenomes showed the strongest CAT activity at all time points examined, followed by the L-segment-based construct and the two S-segment-based constructs. However, the different UUK virus minigenomes did not follow the same kinetics. All UUK virus minigenomes showed a strong increase of reporter gene activity between 10 and 22 h posttransfection. The M-segment-based minigenome showed a further increased reporter gene expression up to 46 h (Fig. 4C). In contrast, L- and S-segment-based constructs reached a plateau at 22 to 28 h posttransfection. This is in agreement with the situation in infected cells where M-segment-directed glycoprotein expression increases during the late stage of infection for efficient assembly of progeny viruses. The ratio of vRNA segments (L:M:S) in purified virions has been found to be about 1:4:2 (22, 23), and similar ratios have been observed in infected cells (31). This also supports the conclusion that the M RNA is the most efficiently replicated segment. Using a different reverse genetics system, Barr et al. (1) recently obtained very similar results for BUN virus (Orthobunyavirus genus). In their study, a luciferase reporter gene was flanked by the NCRs from the three RNA segments. The relative replication efficiency (promoter strength) of the three minigenomes was also found to be M > L > S.

To further analyze the reasons for the different rescue efficiencies of the three genome segments, chimeric minigenomes were generated containing NCRs derived from different UUK virus RNA segments. Surprisingly, all six possible combinations (L-S, S-L, M-S, S-M, L-M, and M-L) showed only very low CAT expression levels (Fig. 6B), demonstrating that NCRs derived from different genome segments do not interact in a functionally efficient way to regulate RNA encapsidation, transcription, and replication processes. Further analysis of the chimeric minigenomes using RNA secondary structure prediction software and stability calculation (GeneBee) (5, 6) of the base-paired terminal nucleotides suggested that a decreased base-pairing potential between the terminal nucleotides of both NCRs could be the reason for the inefficient rescue of the chimeric minigenomes. We therefore introduced multiple point mutations (exchanges and deletions) to adjust the first 20 nt of the 5′ NCR to the last 20 nt of the 3′ NCR to increase the base-pairing potential between the ends (Fig. 7A). These base-paired chimeric constructs showed significantly increased CAT signals (Fig. 7B, lanes 2, 4, and 6) compared to the original constructs, indicating that interaction between the terminal nucleotides of each segment is crucial for efficient cis-regulatory processes. Reporter gene activities from constructs with 5′-mutated NCRs were restored to S-segment-based CAT levels (Fig. 7B). In contrast, UUK virus minigenomes with multiple nucleotide mutations within the 3′ NCR showed only low rescue efficiencies, comparable to those of the corresponding chimeric constructs (Fig. 7B, lanes 3, 5, and 7). One explanation could be that the mutations influenced the Kozak sequence of the M and L 3′ NCRs, thus interfering with translation initiation. Furthermore, transcription and replication initiation takes place at the terminal 3′ nucleotides of vRNA molecules. It was shown previously that not only base pairing but also the nucleotide sequence at the conserved highly base-paired region at the 3′ end of the M RNA NCR were important for promoter activity (12). Our present results show that multiple mutations in the adjacent 3′ NCR also affect promoter function, and they do so more severely than mutations in the 5′ end. It is noteworthy that restoration of the CAT activity to S-segment-based minigenomes was achieved by rather drastic changes of the sequences adjacent to the highly conserved promoter region of the viral RNA segments by multiple (two to four) nucleotide exchanges and/or deletions (two) at the 5′ end. The results clearly demonstrate that base pairing between the terminal nucleotides of the nonconserved NCRs is necessary for efficient transcription and replication processes.

Results similar to ours were recently reported for BUN virus (2). Using hybrid 3′ and 5′ ends flanking a luciferase reporter, no transcription or replication of reporter RNAs was observed. When base pairing was restored, RNA synthesis was again readily detected. The authors concluded that the BUN virus NCRs do not act as independent RNA synthesis promoters; rather, the 3′ and 5′ sequences cooperate to form the functional promoter for both RNA replication and mRNA synthesis. Our results are in good agreement with these conclusions.

Analysis of S-segment-based minigenomes.

To analyze the cis-acting signals responsible for viral N or NSs gene transcription and replication, we generated minigenomes by inserting reporter genes in sense or antisense orientation between the two terminal NCRs of the UUK virus S segment. Surprisingly, CAT and GFP levels were similar for all the constructs, demonstrating equal promoter activities for N and NSs gene transcription and replication (Fig. 4). However, dilution series revealed that promoter elements in the vRNA orientation, responsible for N gene mRNA transcription, showed moderately stronger activities compared to those of cRNA NCRs responsible for the NSs gene transcription (Fig. 4B). This is in agreement with protein levels during UUK virus infection, where the N protein is more abundant than NSs. In addition, NSs is much less stable than the N protein, contributing to the lower level of NSs observed in infected cells (28).

Additional S-segment-based minigenomes were generated to analyze the effect of the IGR on minigenome rescue efficiency. The IGR harbors the transcription termination signals for N and NSs mRNA synthesis (27). The 75-nt-long IGR was inserted downstream of the reporter gene stop codon. Comparison of minigenomes with and without IGR revealed an enhancing effect of the IGR on reporter gene expression levels. This was independent of the orientation of the CAT/GFP gene, demonstrating that cis-acting signals located within the IGR play an important role for transcription of the N and NSs genes.

Our present working hypothesis is that proper transcription termination, regulated by signal elements within the IGR (27, 28, 29), eliminates the encapsidation signal located in the 5′ vRNA NCR (21) in the mRNA transcript, preventing the binding of N proteins to the mRNA transcripts. In minigenomes without IGR no termination signal is available, resulting in mRNA transcripts containing the complete vRNA 5′ NCR, including the N binding site. This leads to an encapsidated mRNA transcript and subsequent inhibition of efficient translation. In addition, mRNA transcripts with both complete flanking NCRs (as in the case of constructs without IGR) can form terminal base-paired regions (panhandle-like structures) interfering with translation initiation.

Based on these results, the IGR of ambisense RNA segments should be included when developing minigenome rescue systems based on the S-segment NCRs in order to improve reporter gene expression levels.

Analysis of packaging efficiencies by passaging experiments.

To determine if cis-acting signals responsible for RNA segment incorporation into virions are located within the two flanking NCRs, minigenomes derived from the three UUK virus genomic segments were used in passaging experiments. All minigenomes were packaged into progeny virus particles as demonstrated by the successful transfer of CAT activity from transfected and UUK virus-superinfected cells to fresh cell lines. This showed that signals responsible for segment incorporation are located within the terminal NCRs; however, packaging efficiencies differed between the minigenomes. M- and S-segment-based RNAs could be detected for two to three passages, whereas after subsequent passages no reporter gene activity could be measured (Fig. 5A and B). This was expected, because no system was used to select for the recombinant viruses containing the CAT minigenome against the wild-type virus used for the superinfection experiment. Similar results were originally reported for influenza virus segments. Passaging of a reporter gene flanked by just the 3′- and 5′-terminal NCRs of vRNAs resulted in the rapid loss of the transgene from the pool of progeny virus particles (19). Surprisingly, minigenomes containing UUK virus L-segment-derived NCRs could be passaged three times with increasing CAT signals, whereas subsequent passages (passages 4 to 7) resulted in only a 20 to 30% decrease in reporter gene expression levels. This outcome suggests that a stable pool of recombinant UUK viruses containing the L-segment-derived minigenomes was generated. Whether the minigenome is incorporated as an additional fourth segment or replaces the wild-type L segment remains to be determined. By analogy with influenza virus vRNA segments (15, 32), it is possible that additional packaging signals are present within the coding regions of each UUK virus segment. The pol I system should make it possible to identify such putative signals.