key: cord-0956387-gkg1059v
authors: JEONG, YONG SEOK; REPASS, JOHN F.; KIM, YOUNG-NAM; HWANG, SUN-MIN; MAKINO, SHINJI
title: Coronavirus Transcription Mediated by Sequences Flanking the Transcription Consensus Sequence
date: 1996-03-01
journal: Virology
DOI: 10.1006/viro.1996.0118
sha: 121d6115779d455aa909ca67331aa248b6ebf7f9
doc_id: 956387
cord_uid: gkg1059v

Abstract In our studies of murine coronavirus transcription, we continue to use defective interfering (DI) RNAs of mouse hepatitis virus (MHV) in which we insert a transcription consensus sequence in order to mimic subgenomic RNA synthesis from the nondefective genome. Using our subgenomic DI system, we have studied the effects of sequences flanking the MHV transcription consensus sequence on subgenomic RNA transcription. We obtained the following results. (i) Insertion of a 12-nucleotide-long sequence including the UCUAAAC transcription consensus sequence at different locations of the DI RNA resulted in different efficiencies of subgenomic DI RNA synthesis. (ii) Differences in the amount of subgenomic DI RNA were defined by the sequences that flanked the 12-nucleotide-long sequence and were not affected by the location of the 12-nucleotide-long sequence on the DI RNA. (iii) Naturally occurring flanking sequences of intergenic sequences at gene 6–7, but not at genes 1–2 and 2–3, contained a transcription suppressive element(s). (iv) Each of three naturally occurring flanking sequences of an MHV genomic cryptic transcription consensus sequence from MHV gene 1 also contained a transcription suppressive element(s). These data showed that sequences flanking the transcription consensus sequence affected MHV transcription.

intergenic consensus sequence, which marks the start of the gene (Joo and Lai et al., 1984 ; Ma-Mouse hepatitis virus (MHV), a coronavirus, is an envekino et al., 1988b; Shieh et al., 1987; Spaan et al., 1983) . loped virus with a single-stranded, positive-sense RNA MHV subgenomic RNAs are detected in MHV-infected genome of approximately 31 kb (Lee et al., 1991; Pachuk cells but not in MHV virion Stohlman, 1978). et al., 1989) . MHV-infected cells synthesize genomic-Therefore, subgenomic-sized RNAs must be synthesized length virus-specific mRNA and six or seven species of from a genomic-sized RNA. Synthesis of subgenomic virus-specific subgenomic mRNAs. The viral mRNAs mRNAs involves a discontinuous transcription step; durhave a 3-coterminal nested-set structure (Lai et al., 1981;  ing subgenomic-sized RNA synthesis independently Leibowitz et al., 1981) and are numbered 1 to 7, in detranscribed leader RNA species possessing a trans-actcreasing order of size (Lai et al., 1981; Leibowitz et al., ing property fuse with the body sequences of subgeno-1981) . None of the mRNAs are packaged into MHV virimic-sized RNA (Jeong and Makino, 1994; Zhang et al., ons, except for mRNA 1, which is efficiently packaged 1994). There are at least two stages in coronavirus subdue to the presence of a packaging signal (Fosmire et genomic RNA synthesis: we call the first stage primary al., 1992). The 5 ends of the MHV genomic RNA and the transcription, during which subgenomic-sized RNA is subgenomic mRNAs start with a leader sequence that is synthesized from a genomic-sized template RNA; the approximately 72 to 77 nucleotides long other stage is called secondary transcription, during 1984; Spaan et al., 1983) . The leader sequence is enwhich subgenomic-sized RNA serves as template (Jeong coded only once in the genomic RNA at the 5 end. The and . MHV-specific genes, which are downstream from the When an intergenic region from MHV is inserted into leader, are separated from one another by an intergenic a location in an MHV defective interfering (DI) RNA, a region. Each intergenic region, located upstream of a novel subgenomic DI RNA is transcribed in helper virusgene essential for MHV replication, includes the unique infected cells (Makino et al., 1991) . We used this subgeconsensus sequence of UCUAAAC, or a very similar senomic DI RNA system to study how MHV transcription quence (Lai et al., 1984; Spaan et al., 1983) . On the subis flexible enough to recognize a mutated transcription genomic mRNAs, the leader sequence is fused to the consensus sequence. We constructed a series of MHV DI RNAs that contain one UCUAAAC consensus se-RNAs showed that the MHV transcription mechanism is DNA construction flexible enough to recognize mutated transcription con-MHV DIssE-specific cDNA clone DE5-w3 (Makino and sensus sequences; subgenomic DI RNAs are synthe- ) was used as a parental clone for DNA consized from most of the mutated consensus sequences struction. Conventional methods of DNA manipulation (Joo and . In another study we showed that (Sambrook et al., 1989) were used. By using PCR-based sequences flanking the same intergenic region between site-directed mutagenesis, a 12-nucleotide-long segenes 6 and 7 do not affect the efficiency of subgenomic quence, TCTAATCTAAAC, was inserted into DI cDNA DI RNA transcription (Makino and Joo, 1993) . These two (Joo and Makino, 1992 , and these recombinant PCR was also employed for construction regions are not utilized for transcription.

of plasmid DI-D20 and DI-TA7 (Higuchi, 1990) . For the One of the possible reasons why transcription does construction of FDI-1/2wt, FDI-2/3wt, FDI-6/7wt, FDI-M1, not occur in these 19 MHV genomic regions is that the FDI-M2, and FDI-M3, MHV-JHM-specific cDNA was iniflanking sequences of these regions suppress transcriptially synthesized by incubating MHV-JHM genomic RNA tion. In both of the previous studies we used DI RNAs with specific primers (Makino et al., 1988a) , the desired which contain the intergenic region between genes 6 MHV-JHM-specific RT-PCR products were obtained after and 7; we used that intergenic region because mRNA 7, incubating MHV-JHM cDNA with two specific oligonuclewhich is synthesized from this intergenic region, is the otides, as described previously (Makino and Joo, 1993) . most abundant MHV mRNA species and we expected For construction of FDI-1/2M, FDI-2/3M, and FDI-6/7M that a large amount of subgenomic DI RNA would be the 12-nucleotide-long sequence, TCTAATCTAAAC, was synthesized from this inserted intergenic region. Howinserted into the RT-PCR products using the recombinant ever, characterization of only this intergenic region may PCR procedure. The recombinant PCR procedure was overlook the possibility that the sequences flanking these also used to insert a TCTTAAC sequence into FDI-M5. 19 regions may suppress transcription from these re-

The resulting RT-PCR product was inserted into the AflIIgions.

SacII site of DE5-w3. For all of the constructs used in Deletion analysis of those MHV DI RNAs that contain this study we sequenced the inserts that were derived the intergenic sequence from gene 6-7 with its naturally from PCR products to confirm the presence of specific occurring flanking sequences showed that reducing the mutations and the absence of extraneous mutations. number of base pairs between the genomic leader sequence and the intergenic region decreases the tran-RNA transcription and transfection scription efficiency (Makino et al., 1991) . However, when van der Most et al. (1994) used another MHV DI RNA, Plasmid DNAs were linearized by XbaI digestion and those authors reported that the extent of base pairing transcribed with T7 RNA polymerase as previously debetween the leader RNA and the intergenic sequence scribed (Makino and Lai, 1989) . The lipofection procedoes not control subgenomic RNA abundance; they used dure (Makino et al., 1991) was used for RNA transfection DI RNAs which lacked the naturally occurring sequences into DBT cells. that flank that intergenic sequence. If the flanking sequences of the transcription consensus sequence affect Preparation of virus-specific intracellular RNA and transcription, then the different results from these two Northern (RNA) blotting studies may be explained by the differences in the nature of sequences flanking the inserted intergenic sequence.

Virus-specific RNAs in virus-infected cells were extracted as previously described (Makino et al., 1984) . For We have studied the possible influence of the flanking sequences on transcription initiated at a transcription each sample, 1.5 mg of intracellular RNA was denatured and electrophoresed through a 1% agarose gel con-consensus sequences by using a subgenomic DI system. Our data indicated that indeed some flanking se-taining formaldehyde, and the separated RNA was blotted onto nylon filters as described previously (Jeong and quences affected transcription. . In some experiments poly(A) containing

RNAs that were selected by oligo (dT)-cellulose column Viruses and cells chromatography were used (Makino et al., 1984) . The nylon filter was soaked in a prehybridization buffer, and The plaque-cloned A59 strain of MHV (MHV-A59) (Lai et al., 1981) was used as a helper virus. Mouse DBT Northern blot hybridization was performed (Jeong and Makino, 1992) . The 32 P-labeled probes were prepared by cells (Hirano et al., 1974) were used for growth of viruses. a random-priming procedure (Sambrook et al., 1989) . For the densitometric analysis, autoradiograms were scanned using a scanner (RELI 4816 scanner, Relysis) and the intensity of each band was quantitated using Scan Analysis program (Biosoft, Cambridge, UK).

Primer extension products were purified from the gel and amplified by PCR under the same conditions described above. The gel-purified RT-PCR products were separated by agarose gel electrophoresis. Direct PCR sequencing was performed according to the procedure established by Winship (1989) .

We examined whether sequences outside of the consensus sequence could affect subgenomic RNA transcription efficiency by using the subgenomic DI RNA system. We initially constructed eight different MHV DI cDNA clones, each with a 12-nucleotide-long TCTAATCTAAAC insert (12-nt sequence) placed at a different site; that inserted 12-nt sequence is perfectly complementary to the 3 region of the genomic leader sequence and in- (Makino et al., 1988a) and intide numbering begins from the 5-end of DE5-w3, and the inserted 12cludes cis-acting replication signals, which are essential nucleotide sequences are boxed.

for MHV-JHM DI RNA replication; these regions were the 5-most 0.47 kb, the 3-most 0.46 kb, and about 60 nucleotides near the 5 end of domain II (Kim et al., 1993;  labeled by random-priming with 32 P. This probe detects only helper virus genomic RNA, genomic DI RNAs, and Kim and Makino, 1995b; Lin and Lai, 1993) (Fig. 1A) . We used PCR-based site-directed mutagenesis to insert the subgenomic DI RNAs ( Fig. 2A) . For the detection of the remaining DI mutants, we used the EagI-SphI fragment 12-nt sequence into DE5-w3. The locations of the 12-nt sequences in these eight clones, DI-AflIII, DI-StuI, DI-of DE5-w3 as a probe (Fig. 2C) , and a probe corresponding to a region 18 to 262 nucleotides from the 3-end of EagI, DI-SphI, DI-SpeI, DI-D20, DI-NruI, and DI-TA 7, are shown in Fig. 1B . ; the former probe detects only helper virus mRNA 1 and genomic DI RNAs and the latter We examined replication and transcription of these DI RNAs in DI RNA-transfected, MHV-infected cells. DI binds to all MHV RNA species. To directly compare the efficiency of subgenomic DI RNA transcription, we used RNAs were synthesized in vitro and transfected by lipofection into DBT cell monolayers that were infected with the same membrane in the experiments documented in Figs. 2B and 2C. All eight DI RNAs replicated efficiently MHV-A59 helper virus 1 hr prior to transfection. At 7 hr postinfection, intracellular RNAs were extracted, sepa-and synthesized a subgenomic DI RNA of expected size; however, the amounts of the subgenomic DI RNAs dif-rated by formaldehyde-agarose gel electrophoresis, and analyzed by Northern blot. For the analysis of DI-AflIII, DI-fered (Fig. 2) . Subgenomic DI RNA was most efficiently transcribed from DI-SpeI DI RNA. DI-SphI and DI-D20 StuI, and DI-EagI, we electrophoresed poly(A)-containing intracellular RNAs and probed Northern blots with the also supported subgenomic DI RNA transcription. DI-NruI supported a low level of subgenomic DI RNA tran-EagI-SphI DNA fragment of DE5-w3 (Fig. 1A ) that was scription, which was not immediately apparent in one quence insertion differed. If the sequences flanking the 12-nt sequence are important for subgenomic DI RNA experiment, as shown in Fig. 2B , in which it could only transcriptional regulation, then the efficiency of subgenobe detected after prolonged exposure of the gel (data mic DI RNA transcription from the newly constructed DI not shown); in another experiment, DI-NruI subgenomic RNAs should be similar to that of their parental DI RNAs. DI RNA synthesis was observed more readily, because

We examined synthesis of subgenomic DI RNA in DI the autoradiogram had a lower background (Fig. 2D) . A RNA-transfected, MHV-A59-infected cells by Northern lower level of subgenomic DI RNA synthesis was also blot analysis in which we applied only poly(A)-containing observed in DI-StuI and DI-EagI. Although synthesis of RNAs onto the gels. We used probe 1 (see Fig. 3 ) to DI-AflIII and DI-TA 7 subgenomic DI RNAs was not eviestimate the ratio of genomic to subgenomic DI RNA for dent by Northern blot analysis, synthesis of both subgemost of the DI RNAs; the exception was that we used nomic DI RNAs was detected by RT-PCR analysis (Jeong probe 2 (see Fig. 3 ) for the analysis of FDI-D20. Probe 1 and Makino, 1994) (data not shown). Direct sequencing was not suitable for analysis of FDI-D20, because while of genomic DI RNA-specific RT-PCR products demonthis probe hybridizes with genomic FDI-D20 RNA at two strated that the inserted 12-nt sequence and its flanking different sites (one is downstream of the AflII site and regions was maintained for all the mutants in DI RNAthe other is downstream of the SacII site), it hybridizes replicating cells (data not shown).

with the subgenomic DI RNA at just a single site. All the tested DI RNAs replicated efficiently in the DI RNA-Flanking sequences of the 12-nt sequence affected transfected, MHV-A59-infected cells (Fig. 4) . A more transcription slowly migrating band appeared in FDI-SpeI DI RNA-To test whether the flanking sequences of the inserted replicating cells, in FDI-D20 DI RNA-replicating cells and 12-nt sequence affected the efficiency of transcription, in FDI-NruI DI RNA-replicating cells (Fig. 4 , lanes 5-7); we constructed another set of DI cDNAs. From DI cDNAs, this band probably represented a DI RNA that was newly DI-StuI, DI-EagI, DI-SphI, DI-SpeI, DI-D20, and DI-NruI, generated in DI RNA-transfected cells. All the DI RNAwe removed a 0.4-kb-long PCR fragment, which carried replicating cells synthesized subgenomic DI RNA with the 12-nt sequence and 0.2 kb of its upstream and downdifferent transcription efficiencies. The difference in the stream flanking sequences, and inserted this into the radioactivity ratios of subgenomic DI RNA to genomic DI AflII-SacII site of DE5-w3 to produce FDI-StuI, FDI-EagI, RNA for these five DI RNAs was roughly comparable to FDI-SphI, FDI-SpeI, FDI-D20, and FDI-NruI, respectively those of the parental DI RNAs (Figs. 2 and 4) . The 0.2- (Fig. 3) . These six cDNAs were the same size, and each kb-long upstream and downstream sequences flanking carried the same 12-nt sequence in the middle of a samethe 12-nt sequence clearly affected subgenomic DI RNA transcription efficiency. length insertion; only the regions flanking the 12-nt se- FIG. 3 . Schematic diagram of the structure of DE5-w3 and DE5-w3-derived insertion mutants with flanking sequences. Fragments for FDIs contained a 0.4-kb-long PCR product consisting of the 12-nucleotide sequence (solid box) and 0.2 kb from an upstream and a downstream flanking sequence; the flanking sequences were derived from DI-StuI, DI-EagI, DI-SphI, DI-SpeI, DI-D20, and DI-NruI genomic fragments. The 0.4-kb PCR fragment of DI-StuI, DI-EagI, DI-SphI, DI-SpeI, DI-D20, and DI-NruI was inserted into the AflII-SacII site of the DE5-w3 to produce FDI-StuI, FDI-EagI, FDI-SphI, FDI-SpeI, FDI-D20, and FDI-NruI, respectively. Solid boxes represent the 12-nucleotide sequence. Locations of probe 1 and probe 2 used for Northern blot analysis (see Fig. 4 ) are also shown.

Effect of the 12-nt sequence location on transcription probe 1 (Fig. 5A ). FDI-StuI/Sp, FDI-EagI/Sp, FDI-SphI/Sp, FDI-AflII/Sp, and FDI-SacII/Sp replicated and transcribed We next examined whether the location of the 12-nt (Fig. 5B) . The molar ratio of the genomic to subgenomic sequence could affect transcription by constructing a new series of DI cDNAs with the 12-nt sequence positioned at different sites in DE5-w3; each construct was similar to the others in that they all contained the 0.4-kb inserted region that carried the 12-nt sequence and its upstream and downstream 0.2-kb sequences; each construct differed only in the location of the insertion within the DI genome. We chose the 0.4-kb fragment used for construction of FDI-SpeI, because FDI-SpeI showed the most efficient subgenomic DI RNA transcription (Fig. 4) . We inserted this 0.4-kb fragment at the AflIII, StuI, EagI, SphI, AflII, SacII, and NruI sites of DE5-w3 to produce FDI-AflIII/Sp, FDI-StuI/Sp, FDI-EagI/Sp, FDI-SphI/Sp, FDI-AflII/Sp, FDI-SacII/Sp and FDI-NruI/Sp, respectively (Fig.  5A) . We expected that this set of DI constructs would overcome the transcriptional suppressive effect that structure of DE5-w3 and DE5-w3-derived insertion mutants with flanking sequences. Fragment for FDIs contained a 0.4-kb-long PCR product consisting of the 12-nucleotide sequence (solid box) and 0.2 kb from an upstream and a downstream flanking sequence of DI-SpeI genomic fragments. This 0.4-kb PCR fragment was inserted into the AflIII site, StuI site, EagI site, SphI site, AflII site, SacII site, and NruI site of the DE5-w3 to produce FDI-AflIII/Sp, FDI-StuI/Sp, FDI-EagI/Sp, FDI-SphI/Sp, FDI-AflII/Sp, FDI-SacII/Sp, and FDI-NruI/Sp, respectively. Solid boxes represent the 12-nucleotide sequence. Location of probe 1 for Northern blot analysis is also shown. (B) Northern blot analysis of FDI RNAs. Intracellular RNAs were extracted from DI RNA-transfected, MHV-infected cells (lanes 6-10) or DI RNA-transfected, mock-infected cells (lanes 1-5), and poly(A)-containing RNAs were analyzed by Northern blot analysis. Lane 11 represents RNA from MHV-infected cells. 32 P-labeled probe 1 (see A) was used as a probe. Arrow represents genomic DI RNAs.

DI RNA from these replication-competent DI RNAs was inserted within the cis-acting replication signals in these DI RNAs (Kim et al., 1993) (Fig. 1) . Probably the insertion essentially the same (Fig. 5B) ; the ratio was approximately 0.3, a value that was very close to that of FDI-of the 0.4 kb disrupted the structure of the cis-acting replication signals in regions that were essential for DI SpeI. We did not observe replication of FDI-AflIII/Sp and FDI-NruI/SP (data not shown). The 0.4-kb region was RNA replication, resulting in failure of the DI RNA to replicate. These data indicated that the location of 12-nt 2 and genes 2-3 demonstrated subgenomic DI RNA transcription activity that was similar to the 0.4-kb-long sequence on DE5-w3 was not crucial for the regulation of subgenomic DI RNA transcription.

flanking sequences of the intergenic region of genes 6-7, indicating that no transcriptionally suppressive element existed in the naturally occurring flanking se-Effect of the naturally occurring sequences that flank quences adjacent to the intergenic regions at genes 1the MHV intergenic regions on transcription 2 and 2-3. Interestingly, sequence alteration from a wildtype intergenic sequence to the 12-nt sequence affected The data presented above demonstrated that some sequences flanking the 12-nt sequence could suppress transcription efficiency only in the DI RNA containing the intergenic region from genes 6-7, but not in DI RNAs subgenomic DI RNA transcriptional efficiency. This data contrasted with our previous observation that the effi-containing intergenic regions from genes 1-2 and 2-3. These data demonstrated that the sequence(s) sur-ciency of subgenomic DI RNA transcription is not regulated by sequences flanking the genes 6-7 intergenic rounding the intergenic region of genes 6-7 contained a transcriptionally suppressive element(s) that suppressed sequence (Makino and Joo, 1993) . Here we present that not all sequences flanking the 12-nt sequence inhibited transcription from the 12-nt sequence, but not from the naturally occurring 18-nucleotide-long intergenic se-the transcription; some sequences flanking the 12-nt sequence did not suppress transcription (see Fig. 4 ). We quence. Less subgenomic DI RNA synthesis in FDI-6/ 7M than in FDI-6/7wt was consistent with our previous interpreted these data as indicating that nonnaturally occurring sequences flanking the 12-nt sequence could study; in that study a mutant like FDI-6/7M, with the same deletion in the intergenic region between genes 6 and suppress subgenomic DI RNA transcription, whereas naturally occurring sequences flanking the genes 6-7 7, synthesized significantly less subgenomic DI RNA than did a DI RNA with an intact intergenic region (Makino intergenic sequence do not suppress transcription.

We examined naturally occurring flanking sequences and Joo, 1993). from other MHV intergenic regions. For this analysis we chose sequences surrounding the genes 1-2 intergenic Sequences surrounding genomic cryptic consensus sequence and the genes 2-3 intergenic sequence. The sequences suppressed transcription amounts of mRNA 2 and mRNA 3, synthesized from the intergenic regions between genes 1-2 and genes 2-3, MHV transcription regulation is flexible enough to recognize altered consensus sequences in which just one respectively, are about 50 and 30 times lower than the amount of mRNA 7 in MHV-infected cells (Leibowitz et of seven consensus sequence nucleotides is changed to any of the three possible alternative bases (Joo and al., 1981) . Currently we do not know why the amount of mRNA 2 and mRNA 3 is significantly lower than mRNA . Indeed some MHV sequences that differ only by one nucleotide from the consensus sequence 7. We constructed six different DI cDNAs, all of which had an insertion of a 0.4-kb fragment at the AflII-SacII are transcriptionally active (Schaad and Baric, 1993) . However, many genomic cryptic consensus sequences site of DE5-w3 (see Figs. 3 and 6) . The inserted 0.4-kb fragment of FDI-1/2wt, FDI-2/3wt, and FDI-6/7wt, was a that also differ from UCUAAAC by only one nucleotide do not act as sites for initiation of subgenomic RNA syn-RT-PCR fragment of the MHV-JHM sequence at genes between 1-2, 2-3, and 6-7, respectively (Fig. 6) . The thesis (Joo and . Because some of the sequences flanking the inserted 12-nt sequence negatively 0.4-kb fragment consisted of the intergenic sequence and its natural occurring 0.2-kb flanking sequences. FDI-affected subgenomic DI RNA transcription, subgenomic RNA transcription from a sequence that differs slightly 1/2M, FDI-2/3M, and FDI-6/7M had structures that, respectively, were very similar to FDI-1/2wt, FDI-2/3wt, and from UCUAAAC may possibly be suppressed by the sequences flanking that sequence; this could explain the FDI-6/7wt, except that these DI cDNAs contained the 12nt sequences instead of the naturally occurring in-flexibility of the consensus sequence.

To test this possibility we constructed a new series of tergenic sequences (Fig. 6) .

We examined synthesis of genomic DI RNA and sub-mutant DI RNAs, all of which contained an insertion of a 0.4-kb fragment at the AflII-SacII site of DE5-w3. The genomic DI RNA in DI RNA-transfected, MHV-A59-infected cells by Northern blot analysis using the 32 P-la-inserted 0.4-kb fragment of FDI-M1, FDI-M2, and FDI-M3 corresponded to MHV-JHM sequences at about 0.7-1.1 beled SacII-SpeI fragment of DE5-w3 as a probe (probe 2) (see Fig. 3 ). The molar ratio of genomic DI RNA to kb, 7.9-8.3 kb, and 10.7-11.1 kb from the 5-end, respectively. The center of these 0.4-kb fragments had a natu-subgenomic DI RNA in the tested DI RNAs was approximately the same, except that FDI-6/7M synthesized less rally occurring UCUUAAC sequence, which differed by only one nucleotide from the UCUAAAC consensus se-subgenomic DI RNA (Fig. 7) . In spite of the significantly lower amounts of mRNA 2 and mRNA 3 relative to mRNA quence. These naturally occurring UCUUAAC sites in the MHV genome are not transcriptionally active. We also 7 that are inherent in MHV-infected cells, the 0.4-kb-long sequences flanking the intergenic region of genes 1-constructed FDI-M5 as a control clone; FDI-M5 was simi- lar to FDI-1/2wt except that FDI-M5 contained a UCU-We examined the synthesis of genomic DI RNA and subgenomic DI RNA in DI RNA-transfected, MHV-A59-UAAC sequence (Fig. 8) in place of the FDI-1/2wt AAU-CUAUAC sequence (see Fig. 6B ). FDI-M5 deleted the 5 infected cells by Northern blot analysis using the 32 Plabeled SacII-SpeI fragment of DE5-w3 as a probe two As of the intergenic sequence of FDI-1/2wt; FDI-M5 was a more appropriate control than FDI-1/2wt, because (probe 2, see Fig. 3 ). All the DI RNAs replicated efficiently, whereas subgenomic DI RNA synthesis occurred only in the presence of two As immediately upstream of the consensus sequence sometimes increases the level of FDI-M5 (Fig. 9) . Replication of FDI-M5 was lower than that of the other DI RNAs, which was consistent with subgenomic DI RNA transcription (Makino et al., 1991) .

our previous study showing that DI RNAs that do not synthesize subgenomic DI RNA replicate more efficiently than those that synthesize subgenomic DI RNA (Jeong and Makino, 1992) . Here we clearly showed that flanking sequences of UCUUAAC in FDI-M1, FDI-M2, and FDI-M3 suppressed subgenomic DI RNA transcription, whereas the flanking sequences of UCUUAAC in FDI-M5 did not. Transcription from some of the genomic cryptic consensus sequences was suppressed by the sequences flanking these regions.

We have investigated whether different sequences flanking a 12-nt constructed sequence that carried an MHV transcription consensus region might affect sub- sus sequence suppressed transcription. We found that transcriptionally active when its flanking sequences were some synthetic sequences flanking the 12-nt sequence replaced with the flanking sequences of the intergenic did suppress subgenomic DI RNA transcription. Placesequence at genes 1-2, indicating that transcription from ment of the 12-nt sequence with fixed flanking sethe UCUUAAC sequence in MHV was suppressed by its quences within different regions of the DI genome did flanking sequences. These assembled data indicated not significantly affect subgenomic DI RNA transcription.

that the influence the flanking sequence(s) exert on tran-Naturally occurring intergenic sequences from genes 1scription from a transcription consensus sequence is a 2, 2-3, and 6-7 inserted into MHV DI-RNA along with point of regulatory control in coronavirus transcription. their flanking sequences did not suppress transcription, Although our present study clearly showed that some whereas a DI RNA containing the 12-nt sequence plus flanking sequences of the transcription consensus sethe genes 6-7 naturally occurring flanking sequences quence suppressed transcription, we do not know how showed reduced transcription efficiency. Subgenomic DI that suppression occurred. MHV transcription generally RNA synthesis did not occur from a UCUUAAC cryptic occurs from the UCUAAAC sequence or very closely retranscription consensus sequence with its naturally oclated sequence and some nucleotide curring flanking sequences, whereas that sequence was changes within the UCUAAAC affect transcription efficiency (Joo and . Noncoronavirus RNAs containing negative sense transcription consensus sequence of coronavirus transmissible gastroenteritis virus (TGEV) serve as templates for transcription in TGEV-infected cells (Hiscox et al., 1995) . These data suggest that the coronavirus transcription machinery most probably recognizes transcription consensus sequences for transcription. Recognition of the intergenic region by MHV transcription machinery may involve host proteins (Zhang and Lai, 1995) . One possible mechanism of flanking-sequence-mediated transcription-suppression is that the transcription consensus sequence and its flanking sequences may form a stable RNA structure that may prevent accessibility of the MHV transcription mechanism to the consensus sequence. Many of the first series of FIG. 9 . Northern blot analysis of DI RNAs. Intracellular RNAs were extracted from DI RNA-transfected, MHV-infected cells (lanes 6-9) or DI RNAs had 12-nt sequences flanked by MHV cis-acting DI RNA-transfected, mock-infected cells (lanes 2-5) and analyzed by replication signals (Fig. 1A) . These cis-acting replication Northern blot analysis. Lane 1 represents RNA from MHV-infected cells. signals may form stable secondary or tertiary struc-32 P-labeled probe 2 (see Fig. 3 ) was used as a probe. Arrow and ture(s), which are essential for the recognition by viral arrowhead represent genomic DI RNAs and subgenomic DI RNA, respectively.

polymerase and host factors (Kim et al., 1993) . RNA sec-ondary (or higher) structure of the internal cis-acting rep-ference in the flanking sequences of the inserted intergenic regions. Probably flanking sequences of the in-lication signal is indeed important for viral RNA synthesis (Kim and Makino, 1995b) . The first set of DI RNAs which serted transcription consensus sequence in the DI RNAs that were used by van der Most et al. (1994) do not contained the 12-nt sequence within or proximal to these cis-acting replication signals replicated efficiently, indi-contain a transcription suppressive element; in these DI RNAs minor changes in the intergenic sequence would cating that the putative structures made by the cis-acting replication signals were probably maintained during DI not significantly affect transcription. We use DI RNAs with an inserted intergenic region from genes 6-7 that RNA replication. To maintain active RNA structure for DI RNA replication, RNA structures formed by a 12-nucleo-contained the transcription suppressive element; in these DI RNAs, when the intergenic sequence contained tide sequence located within or near the predicted stable RNA structures of the cis-acting replication signal may nucleotide deletions, the transcription suppressive element suppressed transcription. In yet another case, DI not be transcriptionally optimal; the MHV transcription machinery may not have had easy access to a 12-nucleo-RNAs constructed to carry only the 18-nucleotide-long intergenic sequence of genes 6-7, show altered tran-tide sequence located within a stable RNA structure, resulting in lower levels of subgenomic DI RNA synthe-scription efficiency when nucleotides within this sequence are deleted (Makino and Joo, 1993) . In that study sis.

FDI-6/7M synthesized a lower amount of subgenomic we chose the insertion site of the intergenic region arbitrarily, so that the sequences surrounding the inserted DI RNA than FDI-1/2M, and FDI-2/3M, indicating that flanking sequences of the intergenic sequence at genes intergenic sequence possibly may have contained a transcription suppressive element(s); this is similar to the 6-7 could suppress subgenomic DI RNA transcription. The presence of a transcription suppressive element(s) way that the sequence around the genes 6-7 intergenic sequence suppresses transcription. This may be the rea-in the flanking sequences of the intergenic sequence at genes 6-7 was unexpected, because previous deletion son why synthesis of subgenomic DI RNA is susceptible to the sequence changes within the 18-nucleotide-long analysis of this region did not reveal a transcription suppressive element (Makino and Joo, 1993) . As shown in intergenic sequence in these DI RNAs. The naturally occurring flanking sequences of the in-this study and previous studies (Makino et al., 1991; Makino and Joo, 1993) , if DI RNAs contained the natural tergenic regions at genes 1-2 and 2-3 did not suppress subgenomic DI RNA transcription. The ns 30 protein and occurring 18 nucleotide-long intergenic sequence, then the sequences flanking this region did not suppress sub-the S protein are translated from mRNA 2 and mRNA 3, respectively. The function of ns 30 is not clear, yet this genomic DI RNA transcription; the suppressive effect was only obvious when the intergenic region contained is a conserved gene for many MHVs, except for one isolate that lacks most of this gene (Schwarz et al., 1990) . the 12-nt sequence. Most of DI RNAs we studied in our previous studies contained the 18-nucleotide-long natu-S protein is essential for MHV replication; S protein binds to the cell receptor to initiate infection (Dveksler et al., rally occurring intergenic sequence from genes 6-7 with its natural occurring flanking sequence, thus we could 1991). Expression of both of those MHV genes seems to be crucial for replication and pathogenesis of MHV. That not detect the transcriptional suppressive effect of flanking sequences at genes 6-7 (Makino et al., 1991; Makino the flanking sequences of the intergenic regions of these genes did not suppress transcription indicated that tran-and Joo, 1993).

For mutants FDI-1/2wt, FDI-1/2M, FDI-2/3wt, and FDI-scriptionally suppressive sequences were eliminated during evolution of MHV. During the evolution of MHV, 2/3M that lacked the transcription suppressive element found in the flanking sequences, the sequence comple-many mutations must have occurred within the sequences flanking the intergenic region of these genes; mentarily between the intergenic sequence and the 3end of leader sequence did not directly correlate with variant viruses whose flanking sequences did not suppress transcription might have had a selective advan-transcription efficiency. In a similar report, van der Most et al. (1994) concluded that the extent of base pairing tage. We showed that flanking sequences of the intergenic between the leader RNA and the intergenic sequence does not control subgenomic RNA abundance; they ana-sequence at genes 6-7 contained the transcription suppressive element. MHV mRNA 7 synthesized from the lyzed several MHV-A59-derived DI RNAs with inserted transcription consensus sequences. On the contrary, de-intergenic region at genes 6-7 encodes the N protein that is essential for MHV replication; N protein forms creasing complementarily between the intergenic sequence and the 3-end of the leader sequence results in helical nucleocapsid and is most likely required for viral RNA synthesis (Compton et al., 1987;  Kim and Makino, a decreasing level of subgenomic DI RNA transcription in DI RNAs with an inserted genes 6-7 region (Makino 1995a) . Probably sequences that correspond to the transcription suppressive element encode an important func-et al., 1991; Makino and Joo, 1993) . We speculate that the different conclusions between van der Most et al.

tion for either M protein or N protein, thus this transcription suppressive element was not eliminated during MHV (1994) and our previous studies probably reflect the dif-

evolved to contain the 18-nucleotide-long intergenic se-Analysis of intracellular small RNAs of mouse hepatitis virus: Eviquence that can overcome the transcription suppression dence for discontinuous transcription

In vitro replication of mouse hepatitis One curious, unanswered question from this study is virus strain A59

Cloning of the 0.4-kb region from genes 1-2, 2-3, and 6-7 synthesized mouse hepatitis virus (MHV) receptor: Expression in human and similar amounts of subgenomic DI RNA. This inconsishamster cell lines confers susceptibility to MHV

Identification and mic DI system and actual MHV transcription. The major characterization of a coronavirus packaging signal

RNA for subgenomic RNA synthesis; MHV genomic RNA

Recombinant PCR. In ''PCR Protocols'' (M. A. Innis, is 31 kb, whereas genomic DI RNA is only 2.6 kb

Replication kb from the 3-end of genome, respectively, whereas our and plaque formation of mouse hepatitis virus (MHV-2) in mouse experimental intergenic regions were located only 1.2 kb cell line DBT culture

Investi-RNA has multiple intergenic regions, whereas genomic gation of the control of coronavirus subgenomic mRNA transcription by using T7-generated negative-sense RNA transcripts

Mechanism of coronavirus transcripbetween the amounts of subgenomic DI RNA and real tion: Duration of primary transcription initiation activity and effect of MHV mRNAs? Coronavirus RNA synthesis is considered subgenomic RNA transcription on RNA replication

Characterization of a murine coronavirus defective interfering RNA internal cis-acting replication sig-1995). The transcription factor may dissociate from the nal

template during scanning

is longer, the chances of the loss of a factor during scan-Characterization of leader RNA sequences on the virion and mRNAs ning should be higher. Furthermore some scanning facof mouse hepatitis virus, a cytoplasmic RNA virus

Mouse hepatitis virus A59: mRNA structure intergenic regions, each of which is separated by about and genetic localization of the sequence divergence from hepato-125 nt, into DI RNA results in synthesis of two subgenotropic strain MHV-3

which is transcribed from the upstream in-Presence of leader sequences in the mRNA of mouse hepatitis virus. tergenic region (on the positive-sense RNA)

RNA of mouse hepatitis virus. of the smaller subgenomic DI RNA (Joo and Makino

MHV genomic RNA contains many intergenic re

gions, many transcription scanning factors may fall off

The complete the template during scanning and this may result in the sequence (22 kilobases) of murine coronavirus gene 1 encoding the lower level of synthesis of the larger subgenomic putative proteases and RNA polymerase

The virus-mRNAs. specific intracellular RNA species of two murine coronaviruses: ACKNOWLEDGMENTS MHV-A59 and MHV-JHM

Deletion mapping of a mouse hepati-This work was supported by Public Health Service Grants AI29984 and AI32591 from the National Institutes of Health. tis virus defective interfering RNA reveals the requirement of an internal and discontinuous sequence for replication

Evidence for new transcription units encoded at the 3 end of the mouse hepatitis virus genome. flanking sequences on coronavirus transcription

A system for study of coronavirus mRNA synthesis: A regulated, expressed subgenomic rus nonstructural protein ns2 is not essential for viral replication in transformed cells

The 5-end sequence of the murine coronavirus

genome: Implications for multiple fusion sites in leader-primed transcription

Coronavirus switching during coronavirus defective interfering RNA replication

Subgenomic RNA synthesis directed by a synthetic defective interfering RNA RNA of murine coronavirus

uous transcription generates heterogeneity at the leader fusion sites of coronavirus mRNAs

An improved method for directly sequencing PCR material using dimethyl sulfoxide

Interaction between the cytoplasmic proteins and the intergenic (promoter) sequence of mouse hepatitis mouse hepatitis virus

and virus RNA: Correlation with the amounts of subgenomic mRNA transcribed

Molecular cloning of the gene encoding the putative polymerase of mouse hepatitis virus, strain

Molecular Cloning