The major mRNA species in the E1 region of the genome of bovine adenovirus type 3 (BAV3) have been defined by using a combination of PCR, 5h RACE, ...
Journal of General Virology (1999), 80, 1735–1742. Printed in Great Britain ...................................................................................................................................................................................................................................................................................
Transcription units of E1a, E1b and pIX regions of bovine adenovirus type 3 B. J. Zheng,1† F. L. Graham1, 2 and L. Prevec1, 2 Departments of Biology1 and Pathology2, McMaster University, Hamilton, Ontario, Canada L8S 4K1
The major mRNA species in the E1 region of the genome of bovine adenovirus type 3 (BAV3) have been defined by using a combination of PCR, 5h RACE, Northern analysis and DNA sequencing. Independent transcription initiation sites were identified for each of the E1a, E1b and protein IX (pIX) transcription units, but all mRNA species terminated at the same poly(A) addition site immediately downstream of the pIX open reading frame. Thus, the BAV3 E1 region, which consists of the E1a and E1b genes together with that for pIX, functions as a nested overlapping transcription unit. One major mRNA species encoding the E1a protein was found and two mRNAs encoding E1b species, the smaller of which encodes the E1b 17K protein alone and the larger encodes both 17K and 47K E1b proteins, were identified. One mRNA species encodes pIX. The E1a transcript, encoding the predicted 214 residue E1a protein, has four exons. The smaller E1b mRNA has two exons, the second of which corresponds to the last exon of E1a. No introns were detected in the larger E1b mRNA that encodes both the E1b 17K and 47K proteins nor in the mRNA encoding pIX. The relative times of appearance of the mRNAs from the E1–pIX gene region following infection of bovine cells with BAV3 was determined.
Introduction The left end of all known mammalian adenoviruses encodes the two early gene regions E1a and E1b, as well as the region encoding delayed-early virion structural protein IX (pIX) (for a recent review of adenovirus replication see Shenk, 1996). The E1a transcription unit is equivalent to the pre-early region of DNA viruses, in that transcription is independent of prior virus protein synthesis and occurs as one of the first events in the infected cell. At least one protein translated from the E1a transcripts functions to facilitate transcription from other early virus genes, including E1b. The best studied adenoviruses, those of the human type C adenovirus group, produce at least five distinct mRNA species from the E1a transcription region. Of these, the major 289R E1a protein, which is the product of the 13s mRNA and contains the conserved amino acid region CR3, is required for optimal transactivation of other virus genes. E1b transcripts are processed to yield at least two major mRNA species, which together produce two E1b proteins (19K Author for correspondence : Ludvik Prevec (at the Department of Biology, LSB429). Fax j1 905 521 2955. e-mail prevec!mcmaster.ca † Present address : Department of Microbiology, University of Hong Kong, Queen Mary Hospital, Hong Kong.
0001-6270 # 1999 SGM
and 58K for the human type C adenoviruses). The specific mRNA molecules used for the synthesis of these proteins differ in different adenovirus systems. Prior to the work described in this paper, 3h-end processing of transcripts from the combined E1–pIX region of adenovirus genomes was known to occur by one of two different pathways. In human adenoviruses (Ad2, 5, 7, 12) and simian adenovirus type 7, the E1a transcripts are polyadenylated at sequences between the E1a and E1b coding regions. E1b transcripts, on the other hand, are polyadenylated at sequences downstream of the pIX coding region. Thus, the pIX transcription unit is nested within the E1b transcription unit. This nesting had been suggested as one possible mechanism for ensuring that pIX expression is delayed in adenovirus infection (Vales & Darnell, 1989). An alternative transcription scheme for the E1–pIX region is that described for murine adenovirus type 1 (MAV-1) (Ball et al., 1989). In this virus system, the E1a and E1b transcription units, each with independent promoters and initiation sites, have been shown to share a common 3h poly(A) addition site. Thus, the E1b transcription unit is nested within the E1a transcription unit in this virus, while the transcription unit for pIX of MAV-1 has an independent promoter and termination site. BHDF
B. J. Zheng, F. L. Graham and L. Prevec and washed a further four times with low-salt buffer. After transferring the spin column to a new microcentrifuge tube, mRNA was eluted and collected in 200 µl elution buffer. The mRNA was precipitated with 0n15 vol. 2 M sodium acetate and 2n5 vol. 100 % ethanol. BAV3 mRNA (1 µg) was first copied into single-stranded cDNA with Superscript reverse transcriptase (BRL), using as primer the poly(T)-rich AB3822 (0n2 µg). The reaction was carried out at 37 mC for 1 h in firststrand buffer (BRL) containing 10 mM DTT, 1n5 mM dNTP and 100 U reverse transcriptase. The appropriate primers (0n2 µg) were added to a portion of the above cDNA in PCR buffer (BRL), together with 0n2 mM dNTP, 2n5 mM MgCl and 2n5 U Taq DNA polymerase (BRL). The reaction was cycled # through 33 cycles of 95 mC for 1 min, 45 mC for 1n5 min and 72 mC for 3 min in an DNA thermal cycler (Perkin-Elmer). The products of RT–PCR were electrophoresed on 1n5 % agarose gels at 50 V and room temperature for about 3 h, stained with ethidium bromide and photographed, as shown in Fig. 2.
In our previous analysis of sequences in the left end of bovine adenovirus type 3 (BAV3) (Zheng et al., 1994), we suggested possible open reading frames (ORFs) for the major E1a protein, two E1b proteins and pIX. We pointed out that the proximal TATA sequence upstream of the first E1b ORF might be located too close to the initiating ATG to serve as a promoter binding site. We also observed that useful consensus poly(A) addition sequences signals were present in the sequence only downstream of the pIX ORF. The work in this paper confirms some of our predictions and defines the transcription start sites of the E1a, E1b and pIX transcripts.
Methods
Reagents, primers, cells and virus. Reagents for Northern blotting, RT–PCR and 5h RACE were purchased from BRL, Dupont, Pharmacia, Sigma or Boehringer Mannheim. All oligonucleotides used as primers were synthesized by the Central Facility of the Molecular Biology and Biotechnology Institute (MOBIX) at McMaster University, Hamilton, Ont. The primers used in this work, their sequences and corresponding locations on the BAV3 genome are all presented in Table 1. Madin–Darby bovine kidney (MDBK) cells were maintained in medium α-MEM with 10 % newborn calf serum (BRL). BAV3 (strain WBR-1) (Darbyshire et al., 1965) was kindly provided by B. Derbyshire, University of Guelph, Guelph, Canada.
5h RACE of BAV3 mRNA. 5h RACE of BAV3 mRNA was performed by using a commercial kit (5h RACE system ; BRL) following the protocol provided by the manufacturer. Briefly, the cDNA product of the first-strand reverse transcription described above was tailed by the addition of C residues. This was then amplified by PCR by using a commercially supplied primer (AP) that hybridized with the oligo(C) tail and a second primer indicated in the legend to Fig. 3 that was complementary to a region within the expected mRNA molecule. In the case of E1a mRNA, a second amplifying PCR utilized a commercial primer (UAP) that nested within the first commercial primer and a second selected primer complementary to expected mRNA sequences within the first PCR product. The DNA products were electrophoresed on 1n5 % agarose gels at 50 V and room temperature for about 3 h, stained with ethidium bromide and photographed. If required, specific DNA fragments were purified from the gel by using the Wizard DNA purification system (Promega) following the protocol described by the manufacturer. In some instances, the DNA products were subsequently cloned into pTZ plasmids (Pharmacia) prior to sequencing.
Isolation of BAV3 mRNA and its use in RT–PCR. BAV3 mRNA was isolated from MDBK cells after infection with BAV3 by using the FasTrack mRNA isolation kit (Invitrogen) following the protocol provided. Briefly, MDBK cells infected with 50 p.f.u. BAV3 per cell were harvested 16–20 h later in trypsin–EDTA buffer. After washing with PBS, about 5i10( infected cells were lysed by resuspension in 15 ml lysis buffer and incubation at 45 mC for 10 min. The lysate was passed through a 21G needle and incubated at 45 mC for a further 40 min and the NaCl concentration was then adjusted to 0n5 M. Oligo(dT) cellulose was added to the lysate and incubated for 60 min at room temperature with gentle mixing. The oligo(dT) cellulose was washed twice with binding buffer, four times with low-salt wash buffer, transferred to a spin column
Sequencing. Sequencing of gel-purified PCR products or of plasmids, which were first purified by standard CsCl protocols, was
Table 1. BAV3 sequence primers used for PCR and 5h RACE Primers were designed on the basis of the sequence data of Zheng et al. (1994). Primer
Sequence (5h–3h)
Genome site
AB2646 AB2647 AB2820 AB2953 AB2977 AB3107 AB3822
TGATCCCTGGCTAAATG GGATCACTTAAGCGTTC GTAACTGATGTGCTCGA TGTACAAAGTGCTGAGA GGCAGGGCTACAGTCCC CCACCATGAAGTACCTG GGGTCGACTTTTTTTTTTTTTTTTT
AB4706 AB5969 AB5970 AB6220 AB6221
GTAGCTCTTTATGTCCTC GCCAGGGATCAGGAAACAGC ATAGCCGCTCCGAAGCAACC GGGGTTAACTTGGCCGC CCACAAGGTTAGCCTGC
965–981 1478–1494 2023–2007 2946–2962 2853–2837 601–617 Complementary to poly(A) 2101–2084 975–956 1682–1663 3568–3552 3551–3535
BHDG
Direction
Transcript map of BAV3 E1–pIX region performed by automated sequencing on a 373A DNA Sequencer (Applied Biosystems) with Taq-Dye terminators by B. Allore (MOBIX, McMaster University).
Northern blotting of BAV3 mRNA. After electrophoresis of BAV3 mRNA in 1n5 % agarose gels containing 0n66 M formaldehyde (Lehrach et al., 1977) at 50 V and room temperature for 3 h, the mRNA, depurinated in 0n25 M HCl for 4 min, was transferred to a GeneScreen hybridization transfer membrane (Dupont) in 20i SSC by using an LKB 2016-100 VacuGene vacuum blotting unit (Pharmacia). The transferred mRNA was fixed to the membrane by UV cross-linking at a wavelength of 254 nm and a dose of 1200 mW\cm#. Hybridization probes labelled with fluorescein-12-dUTP were prepared by PCR. A plasmid containing 11 % of the left end of the BAV3 genome (Zheng et al., 1994) was used as the template. Primers AB3107 and AB5969 produced a probe for E1a, AB2647 and AB5970 for E1b 17K, AB2898 and AB2977 for E1b 47K, AB2647 and AB2820 for total E1 and AB2953 and AB6220 for the pIX gene. The PCR mixture was subjected to 25 cycles of 95 mC for 1 min, 50 mC for 1 min and 72 mC for 2 min in a DNA thermal cycler (Perkin-Elmer). After pre-hybridization at 65 mC for 3 h with 0n5 % blocking reagent (Dupont) and 5 % dextran sulphate (Pharmacia) in 5i SSC containing 0n5 % SDS (Sigma), 25 ng\ml probe (boiled for 5 min and then chilled on ice before use) and 50 µg\ml salmon sperm DNA was annealed to the membrane-fixed mRNA at 65 mC overnight in pre-hybridization buffer. The membrane was then washed with 2i SSC, 0n1 % SDS for 15 min followed by 0n2i SSC, 0n1 % SDS for 15 min at 65 mC. After washing twice with buffer 1 (0n1 M Tris–HCl, pH 7n5, 0n15 M NaCl) at room temperature, the membrane was blocked with 0n5 % blocking reagent (Dupont) for 1 h at room temperature and then incubated in antibodyconjugate solution [anti-fluorescein HRP conjugate (Dupont) diluted 1 : 1000] for another 1 h at room temperature. The membrane was washed, incubated in chemiluminescence reagent solution [1 : 1 mixture of the enhanced luminol reagent and oxidizing reagent (Dupont)] for 1 min and exposed to X-ray film (Kodak) for from 1 to 20 min.
Results Defining exons, poly(A) addition and transcription start sites
cDNA was synthesized from mRNA extracted from BAV3infected MDBK cells by using a primer (AB3822) that is complementary in part to the poly(A) tail of mRNA. Based on our sequence data (Zheng et al., 1994), PCR primers were constructed to correspond to coding sequences within the E1a and E1b 17K regions (Table 1). These primers were used in separate PCRs, as indicated in Fig. 2, to amplify portions of the cDNA product, and allowed us to determine the location of introns and the position of 3h-terminal poly(A) addition. To determine transcription start sites, 5h RACE was carried out with appropriate primers for cDNA synthesis. The results of these analyses, which led to the transcript processing scheme presented in Fig. 1, are described in some detail in the following section. Exons and poly(A) addition site for mRNAs containing the predicted E1a ORF
The only significant product of a PCR with cDNA as template and using the primers AB3822 [oligo(dT)] and
AB2646 (within the E1a coding region) was some 950 bp in length. When sequenced, this product indicated that E1a mRNA consisted of the four exons depicted for this mRNA in Fig. 1. The intron between nt 1215 and 1323, a correlate of which is found in all adenovirus E1a mRNAs, had been predicted by us on the basis of comparisons with Ad5 E1a mRNAs (Zheng et al., 1994). The translation stop signal for the E1a protein is located in the second exon, within sequences in the first of two 35 bp repeats that we have previously described in this virus. This exon terminates within the second 35 bp repeat, at nt 1379, and is spliced to the third exon, which consists of nt 1807–2073. This exon is in turn spliced to the fourth exon, which begins at nt 3325 within the pIX gene and terminates with a poly(A) sequence after nt 3638, 26 nucleotides downstream of the poly(A) addition signal beginning at nt 3612. With the cDNA as the template in a PCR, primers AB2646 and AB6220 should generate a DNA product of 815 bp. As seen in lane 1 of Fig. 2, this expected result was observed. There was also a slightly larger band [expected to be 905 bp or larger, depending on the length of poly(A) included] that, as stated above, is due to PCR amplification between primers AB2646 and AB3822, which remained in the reaction mixture. The minor PCR product of about 500 bp is of unknown origin, but also contained the third exon when sequenced and therefore did not represent a mRNA species that terminated between the E1a and E1b ORFs. Thus, unlike most other adenoviruses, the major E1a transcript is not polyadenylated between the E1a and E1b genes. In lane 2, the PCR product primed by AB3107 and AB6220 was expected to be 1182 bp, confirming the presence of the postulated E1a-initiating ATG within this transcript. Again, the presence of a slightly larger band (1273 bp or larger) is due to priming by excess AB3822 in the reaction mixture. In lane 3, the principal product primed by AB2646 and AB2820 is no smaller than the expected 524 bp. This suggests strongly that there is no prevalent E1a mRNA species in BAV3 equivalent to the 12s mRNA of human adenoviruses, from which the CR3 coding region has been deleted. Exons and poly(A) addition sites for mRNAs containing the predicted E1b 17K ORF
As seen in Fig. 2, lanes 4 and 5, the principal PCR product corresponding to the region between primer AB6220 or AB6221, respectively, and primer AB2647 (within the expected 17K coding region) is about 850 bp in size. The product due to the reaction between excess primer AB3822 and AB2647 should contain 950 bp and is not separated from the lower band or is present in a much smaller amount. A minor product of more than 2000 bp was also produced in this PCR. As described below, this corresponds to the larger E1b-related mRNA species. Sequencing of the smaller products revealed the transcript structure shown for the E1b 17K protein in Fig. 1. The first exon, which contains the predicted 17K-initiating BHDH
B. J. Zheng, F. L. Graham and L. Prevec BAV3 E1 region
ITR
ORF
poly (A)
mRNA E1a E1b small
poly (A)
E1b large
poly (A) poly (A)
pIX
Primers Fig. 1. A schematic map of ORFs and mRNA molecules resulting from transcription of the E1 region of BAV3. The left end of the BAV3 genome (4000 nucleotides, as determined by Zheng et al., 1994) is represented on the top line with the inverted terminal repeat (ITR) and the internal 35 bp repeat indicated. The predicted ORFs for the E1a, E1b 17K, E1b 47K and pIX virion proteins are indicated by open rectangles with initiating and last coding nucleotides. The mRNA species determined in this paper and designated E1a, E1b small, E1b large and pIX are indicated by the positions of nucleotides bordering each exon. All mRNA species are polyadenylated at a common site (nt 3640). When nucleotide positions are not given they are identical to the position shown directly above. The location and direction of primers (‘ leftward ’ primers numbered above the line) used are indicated on the bottom line of the figure.
1
2
3
4
5
6
7
1
bp
bp
1500
1500
2
3
4
600
600
100
100
Fig. 2
Fig. 3
Fig. 2. PCR products resulting from amplification of cDNA sequences of BAV3 mRNA. RNA was extracted from BAV3-infected bovine cells in culture and, after cDNA synthesis with reverse transcriptase and a primer (AB3822) complementary to poly(A) sequences, the cDNA was further amplified in PCR by the addition of primers corresponding to selected regions within the BAV3 E1 region. The PCR products were then analysed on an agarose gel. The two outside lanes are 100 bp ladders. Lanes : 1, primers AB6220 (nt 3567–3551) and AB2646 (965–981) ; 2, primers AB6220 and AB3107 (598–616) ; 3, primers AB2820 (2022–2006) and AB2646 ; 4, primers AB6221 (3550–3534) and AB2647 (1477–1493) ; 5, primers AB6220 and AB2647 ; 6, primers AB2977 (2852–2836) and AB2647 ; 7, primers AB6221 (3550–3534) and AB2953 (2945–2961). Fig. 3. PCR products containing the 5h-terminal sequences of BAV3 E1 mRNA molecules. After extraction of mRNA from BAV3infected cells and synthesis of the cDNA as described in Methods, the cDNA was extended with C residues. A commercial primer, AP, was used in conjunction with a primer corresponding to a selected sequence within the E1 region. In one case, a second PCR amplification was conducted with another commercial primer (UAP), which primes from within the AP sequences, in conjunction with a second selected internal primer. The products were analysed on an agarose gel and the major band was isolated and sequenced to determine the 5h nucleotide start site. Lanes : 1, E1a ORF-derived product of second PCR with primers AB5969 (nt 975–956) and UAP after the first PCR with primers AB2820 (2022–2006) and AP ; 2, E1b 17K ORFderived product of PCR with primers AB5970 (1681–1662) and AP ; 3, E1b 47K ORF-derived product of PCR with primers AB4706 (2100–2083) and AP ; 4, pIX ORF-derived product of PCR with primers AB6220 (3567–3551) and AP.
BHDI
Transcript map of BAV3 E1–pIX region
ATG, terminates at nt 2073 and from that point on resembles the E1a transcript described above. This mRNA contains the 17K ORF but only a small portion of the predicted 47K ORF. The larger PCR product shown in lanes 4 and 5 of Fig. 2 represents a transcript that is polyadenylated at the same site as the E1a and E1b transcripts described above but has no introns detected by sequencing. This mRNA molecule encodes both the 17K protein and the 47K protein and may be analogous to the corresponding E1b mRNA of the human type C adenoviruses (Bos et al., 1981). Translation of the 47K protein would depend on internal initiation of translation at the fourth AUG of this mRNA. Interestingly, this same internal AUG is present in both the E1a and shorter E1b mRNA species described above. The resulting ORF in those mRNA species could encode a protein of 12n3 kDa or 106 amino acids, of which the 75 amino-terminal residues would be common to the predicted 47K protein. Lanes 6 and 7 of Fig. 2 confirm the expected size of PCR products with at least one primer that is unique to the expected 47K coding region. Thus, primers AB2977 and AB2647 produce a major fragment consistent with the expected size of 1375 bp and primers AB2953 and AB6221 produce a product consistent with the expected size of 605 bp. No such products were produced if the intervening oligo(dT) cDNA step was omitted from the reaction (not shown), showing that the unspliced product is not simply amplified from contaminating viral DNA. Determining the transcription start sites by 5h RACE
After synthesis of a cDNA template by using an oligo(dT) primer, four distinct RACE reactions were carried out, as described in Methods. The reactions differed only in that one primer for the subsequent internal PCR was chosen to allow specificity in detecting the transcription starts for each of E1a, E1b 17K (E1b small), E1b 17Kj47K (E1b large) and pIX. The size of the final products, which were subsequently sequenced to determine the exact transcription starts, are shown in Fig. 3. The transcription starts determined by sequencing are shown in Fig. 1 and indicate that E1a, E1b and pIX are each initiated independently. E1a transcription is initiated at nt 554, E1b at nt 1447 and pIX at nt 3167. As anticipated, the transcription initiation sites for both of the E1b-containing mRNA species are identical. Time-course of appearance of mRNA species transcribed from the BAV3 E1 region
RNA was extracted from MDBK cells at 12, 18, 24 and 48 h after infection with BAV3 at an m.o.i. of 50 p.f.u. per cell as described in Methods. A modified Northern analysis was carried out on this RNA with the cDNA sequences indicated in Fig. 4 as probes. Probe 1 was designed to detect E1a-encoding transcripts only, while probe 2 was specific for the E1b 17K-
encoding region and probe 3 was specific for the E1b 47Kencoding region. Probe 4 should detect the two E1b mRNA species with equal efficiency and may also detect the E1a transcript, although with reduced efficiency under these conditions. Probe 5 should detect E1a, E1b and pIX transcripts. Probe 1 detected levels of E1a mRNA very near the limit of resolution of this system. E1a mRNA was first detected at 18 h post-infection (p.i.) and was present at approximately the same level at 24 and 48 h p.i. For reasons which are still unclear to us, but which might indicate that relatively low levels of this mRNA were present, E1a mRNA was detected poorly by the different E1a probes that we used. For example, both probes 4 and 5, which should have been capable of detecting some E1a transcript, showed very little of this RNA relative to E1b or pIX transcripts, respectively, and a totally distinct E1a-specific probe showed the same low level of detection (data not shown). The RNA species detected by the E1a probe at approximately 4 kb could represent the unspliced E1a transcript, but this was not determined. The nature of the material observed below the band designated as E1a (expected) is unknown, although it is unlikely to represent an E1a-encoding mRNA species, on the basis of our inability to amplify by PCR any smaller poly(A)-containing species that might correspond to this material. On repeating the Northern blotting for E1a and E1b, using restriction enzyme-derived DNA sequences purified from plasmid DNA as probes, the same principal E1a and E1b species were detected at 18 h p.i., although larger amounts of a small (100 to 200 nucleotides) species were also detected (data not shown). Northern blotting of RNA from uninfected bovine cells resulted in the detection of some minor species that did not correspond to the virus-specific species shown here (result not shown). The results obtained with both probe 2 and probe 4 indicate that, except for the earliest time-point, the amount of the E1b transcript that encodes only the 17K protein (E1b small) was greater than the amount of the transcript that encodes both the 17K and 47K proteins (E1b large). Examination of the results from probe 3, which detects only the larger E1b message, and that of probe 4, which detects both, shows that, while both messages could be detected at 12 h p.i., the amount of the smaller E1b message increased more rapidly until 24 h p.i. Since the smaller mRNA is a spliced product of the larger transcript, it is perhaps significant that the appearance of the larger mRNA species slightly preceded that of the smaller, as seen by the results with both probes 2 and 4 at the 12 h time-point. The results obtained with probe 5 indicate that pIX transcripts were first detected at 12 h p.i. but levels continued to increase significantly between 24 and 48 h p.i. This is in keeping with later transcription of pIX. This experiment failed to detect significant amounts of E1a and E1b mRNA with this probe ; however, other experiments with the same probe (not shown) detected these species, albeit at much lower concentrations than that of pIX mRNA. BHDJ
B. J. Zheng, F. L. Graham and L. Prevec
kb 7.46 4.40 2.37 1.35
0.24
Probes for BAV3 E1 mRNA
Poly (A) Poly (A) Poly (A) Poly (A) Fig. 4. Northern blot indicating the time-course of appearance of mRNA species transcribed from the BAV3 E1 region. Biotintagged probes corresponding to regions of the BAV3 E1 genome were synthesized by PCR with selected primers as indicated in the lower portion of the figure. The primer number and the approximate location along the genome are shown for each of the five probes used in this study. RNA was extracted from BAV3-infected cells at 12 h (lanes 1) ; 18 h (lanes 2) ; 24 h (lanes 3) and 42 h (lanes 4) p.i. The RNA was separated by electrophoresis on 1n5 % agarose–formaldehyde gels, transferred to GeneScreen membranes and probed as described in Methods. The expected positions for the mRNA species corresponding to the E1a, small E1b, large E1b and pIX products are indicated to the left and the locations of standard RNA markers are shown on the right.
Discussion The transcription pattern of the BAV3 E1 region as defined in this work differs from the adenovirus E1 transcription patterns described previously in that all three independently initiated regions, E1a, E1b and pIX, share a common poly(A) addition site. E1a transcripts therefore overlap not only the E1b region, as in MAV-1, but also the pIX region. It is interesting, and perhaps significant in this context, that the MAV-1 E1a region appears to give rise to just one dominant transcript, comparable to the situation in BAV3 but different from that seen in the human type C adenoviruses. The lower level of transcription from the BAV3 E1a region suggested by our results may be in keeping with the need to minimize levels of E1a transcripts to allow optimal E1b expression in this type of transcription system. If this idea has some validity, it could perhaps help to explain why only one mRNA species is produced from the BAV3 E1a region. It has been suggested by Spector et al. (1993) that different DNA molecules may be used for E1a and E1b transcription, to avoid conflicts between transcribing complexes and promoter elements. The work of Falck-Pedersen et al. (1985) suggests, however, that efficient BHEA
expression of E1b may require upstream (E1a) transcripts to continue into the promoter region of E1b. E1a–E1b co-transcripts have been described as minor species in cells infected with Ad2 (Berk & Sharp, 1978 ; Kitchingman & Westphal, 1980) and Ad12 (Saito et al., 1983), although the specific function, if any, of these transcripts remains unknown. An unusual transcription product initiating at the E1a start site and terminating at the pIX poly(A) addition site was reported by Steinthorsdottir & Mautner (1991) and confirmed by Ishida et al. (1994) in Ad40-infected cells. This transcript contains the first 40 E1a codons spliced to a single codon and a translation stop codon in the E1b region without the use of known splice-site sequences (Hall & Padgett, 1994 ; Mount, 1996). The significance of this particular transcription product is unknown but it suggests that transcription may occur across the whole E1–pIX region in this virus system, although splicing between independent transcripts cannot be ruled out. The internal intron of E1a, which is a consistent feature of all adenovirus E1a transcripts and the position of which we had predicted correctly from protein sequence comparisons (Zheng et al., 1994), is removed from the mRNA encoding the BAV3
Transcript map of BAV3 E1–pIX region
E1a protein. In the BAV3 system, we did not detect mRNA molecules which corresponded to other mRNA molecules from human Ad5 E1a, in which larger regions of the E1a coding region are deleted by splicing. Even the use of PCR primers that might have increased the sensitivity of detection of an mRNA corresponding to the 12s message of human Ad5 failed to do so (see for example lane 3 of Fig. 2). Thus, as mentioned above, in contrast to the processing of transcripts in the human type C adenoviruses but consistent with observations for MAV-1 E1a transcripts (Ball et al., 1989), there appears to be no significant fraction of processed mRNA transcripts in BAV3 infection that lacks the CR3 region of E1a. Translation termination of the BAV3 E1a protein occurs at a TGA sequence that consists of nt 12–14 within the first of two 35 nt repeats found in this region of the BAV3 genome (Zheng et al., 1994). The 5h splice-donor site that ends the second exon is ten nucleotides within the second copy of the 35 nt repeat sequence. Since the identical sequence is found within the first repeat, it is at least theoretically possible that this first repeat site could also be used in splicing. If this site were used as the donor to the same splice acceptor site, it would add a single leucine residue to the predicted E1a protein. We found no evidence for an mRNA that was spliced from within the first repeat. Interestingly, in their sequence of the BAV3 E1 region, Elgadi et al. (1993) found three repeats of the 35 bp sequence. In view of the splicing pattern described above, the presence of additional repeats should have no effect on the nature of the final E1a mRNA. The basal promoter of human adenovirus E1b has been shown to be a TATAAA sequence preceded within 30 nt by a GC-rich Sp1-binding site (Wu et al., 1987 ; Wu & Berk, 1988). Additional regulatory elements upstream of this, within the E1a coding and transcript region, have been identified (Parks et al., 1988 ; Spector et al., 1993). The 35 bp repeat sequences of BAV3 are appropriately positioned to contain promoter elements for the E1b transcription unit, and the GC-rich sequence at the end of each repeat is a potential Sp1-binding site. The polymerase assembly sequence is probably the TTTAAA sequence beginning at nt 1416, 12 nucleotides downstream of the end of the 35 bp repeat. The E1b transcription start site at nt 1447 is located upstream of a consensus TATAAA sequence that we had previously pointed out as being too close to the predicted translation-initiating ATG to function effectively as a polymerase assembly point (Zheng et al., 1994). One E1b transcript that we have identified, and which could encode both the predicted 17K and 47K E1b proteins, contains no introns. This unspliced product could be a strictly nuclear product, but since we detected no other mRNA products that encoded the 47K ORF it seems likely that this is the mRNA that encodes both the 17K and 47K proteins. Synthesis of the 47K protein from this mRNA requires internal initiation of translation, in a manner identical to that described for the E1b proteins of Ad5 (Bos et al., 1981). A second, smaller
mRNA resulting from intron deletion from a transcript identical to the above mRNA encodes the 17K protein. This mRNA, which is the most easily detected transcript of the E1a\E1b region, accumulates rapidly for the first 24 h following infection. The intron removed to produce the 17K mRNA is identical to the third intron removed from the E1a transcript. It is of some theoretical interest that both the E1a mRNA and the 17K mRNA contain the initiating ATG of the 47K protein and a consequent ORF that encodes a potential protein of 12n3 kDa. Since this ORF could direct initiation of translation from the 17K mRNA by a mechanism identical to that used by the 47K protein, it would seem reasonable to expect that this protein may be made in virus-infected cells. The same ORF is present in the E1a mRNA molecule but in this case it represents a downstream ORF completely independent of the E1a ORF. This mRNA could by used for translation of the 12n3 kDa protein by a reinitiation process (Kozak, 1987), but the existence of this protein, from any potential mRNA source, remains to be demonstrated. Note added in proof : Because the identification of the E1a transcription start site required a second amplifying PCR and the resulting DNA sequence was not necessarily unique, the start site at nt 554 may not be the sole, or the major, transcription start site for this region. We thank Derek Cummings for general technical assistance and Carole Evelegh for repeating the Northern blots. This work was supported by grants from the National Science and Engineering Research Council of Canada (NSERC), the Medical Research Council of Canada (MRC), the National Cancer Institute of Canada and the National Institutes of Health. F. L. G. is a Terry Fox Research Scientist of the National Cancer Institute of Canada.
References Ball, A. O., Beard, C. W., Redick, S. D. & Spindler, K. R. (1989).
Genome organization of mouse adenovirus type 1 early region 1 : a novel transcription map. Virology 170, 523–536. Berk, A. J. & Sharp, P. A. (1977). Sizing and mapping of early adenovirus mRNAs by gel electrophoresis of S1 endonuclease-digested hybrids. Cell 12, 721–732. Bos, J. L., Polder, L. J., Bernards, R., Schrier, P. I., van den Elsen, P. J., van der Eb, A. J. & van Ormondt, H. (1981). The 2n2 kb E1b
mRNA of human Ad12 and Ad5 codes for two tumor antigens starting at different AUG triplets. Cell 27, 121–131. Darbyshire, J. H., Dawson, P. S., Lamont, R. H., Ostler, D. C. & Pereira, G. H. (1965). A new adenovirus serotype of bovine origin. Journal of
Comparative Pathology 75, 327–330. Elgadi, M., Rghei, N. & Haj-Ahmad, Y. (1993). Sequence and sequence
analysis of E1 and pIX regions of the BAV3 genome. Intervirology 36, 113–120. Falck-Pedersen, E., Logan, J., Shenk, T. & Darnell, J. E., Jr (1985).
Transcription termination within the E1A gene of adenovirus induced by insertion of the mouse beta-major globin terminator element. Cell 40, 897–905. BHEB
B. J. Zheng, F. L. Graham and L. Prevec Hall, S. L. & Padgett, R. A. (1994). Conserved sequences in a class of rare eukaryotic nuclear introns with non-consensus splice sites. Journal of Molecular Biology 239, 357–365. Ishida, S., Fujinaga, Y., Fujinaga, K., Sakamoto, N. & Hashimoto, S. (1994). Unusual splice sites in the E1A-E1B cotranscripts synthesized in
adenovirus type 40-infected A549 cells. Archives of Virology 139, 389–402. Kitchingman, G. R. & Westphal, H. (1980). The structure of adenovirus 2 early nuclear and cytoplasmic RNAs. Journal of Molecular Biology 137, 23–48. Kozak, M. (1987). Effects of intercistronic length on the efficiency of reinitiation by eucaryotic ribosomes. Molecular and Cellular Biology 7, 3438–3445. Lehrach, M., Diamond, D., Wozney, J. M. & Boedtker, H. (1977). RNA molecular weight determinations by gel electrophoresis under denaturing conditions, a critical reexamination. Biochemistry 16, 4743–4751. Mount, S. M. (1996). Perspective : AT–AC introns : an ATtACk on dogma. Science 271, 1690–1692. Parks, C. L., Banerjee, S. & Spector, D. J. (1988). Organization of the transcriptional control region of the E1b gene of adenovirus type 5. Journal of Virology 62, 54–67. Saito, I., Shiroki, K. & Shimojo, H. (1983). mRNA species and proteins of adenovirus type 12 transforming regions : identification of proteins translated from multiple coding stretches in 2n2 kb region 1B mRNA in vitro and in vivo. Virology 127, 272–289.
BHEC
Shenk, T. (1996). Adenoviridae : the viruses and their replication. In Fields Virology, 3rd edn, pp. 2111–2148. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia : Lippincott-Raven. Spector, D. J., Parks, C. L. & Knittle, R. A. (1993). A multicomponent cis-activator of transcription of the E1b gene of adenovirus type 5. Virology 194, 128–136. Steinthorsdottir, V. & Mautner, V. (1991). Enteric adenovirus type 40 : E1B transcription map and identification of novel E1A-E1B cotranscripts in lytically infected cells. Virology 181, 139–149. Vales, L. D. & Darnell, J. E., Jr (1989). Promoter occlusion prevents transcription of adenovirus polypeptide IX mRNA until after DNA replication. Genes & Development 3, 49–59. Wu, L. & Berk, A. (1988). Constraints on spacing between transcription factor binding sites in a simple adenovirus promoter. Genes & Development 2, 403–411. Wu, L., Rosser, D. S., Schmidt, M. C. & Berk, A. (1987). A TATA box implicated in E1A transcriptional activation of a simple adenovirus 2 promoter. Nature 326, 512–515. Zheng, B., Mittal, S. K., Graham, F. L. & Prevec, L. (1994). The E1 sequence of bovine adenovirus type 3 and complementation of human adenovirus type 5 E1A function in bovine cells. Virus Research 31, 163–186.
Received 23 February 1999 ; Accepted 3 March 1999