Apr 25, 1983 - Four RNA polymerase II promoters have been mapped in the. DNA sequence of the EcoRI-H and -Dhet fragments of B95-8. Epstein-Barr virus.
The EMBO Journal Vol.2 No.8 pp. 1331-1338, 1983
Latent and lytic cycle promoters of Epstein-Barr virus
Paul J.Farrell*l, Alan Bankier, Carl Skguin, Prescott Deininger2 and Bart G.Barrell Laboratory of Molecular Biology, Medical Research Council Centre, Hills Road, Cambridge CB2 2QH, UK Communicated by A.A. Travers Received on 25 April 1983
Four RNA polymerase II promoters have been mapped in the DNA sequence of the EcoRI-H and -Dhet fragments of B95-8 Epstein-Barr virus. RNAs transcribed from three of these promoters are dramatically induced by treatment of B95-8 cells with 12-O-tetradecanoylphorbol-13-acetate (TPA). The other promoter is active with or without TPA treatment of the cells and is thus active in the latent virus cycle. Deletion mapping suggsts that DNA sequence homologies between some of the promoters lie in the same region as essential upstream promoter elements. Key words: Epstein-Barr virus/transcription Introduction Epstein-Barr virus (EBV) is a human herpesvirus which is the infectious agent in infectious mononucleosis and is closely associated with Burkitt's lymphoma and nasopharyngeal carcinoma (reviewed in Epstein and Achong, 1979; Tooze, 1980). In tissue culture, the virus transforms B lymphocytes from humans and some primates. We are studying the B95-8 line of EBV-transformed cells; this line was generated by infecting marmoset lymphocytes with EBV initially derived from a patient with infectious mononucleosis (Miller et al., 1972). Virus production from this cell line is stimulated up to 50-fold (zur Hausen et al., 1978) by treating the cells with 12-O-tetradecanoylphorbol-13-acetate (TPA). EBV has a double-stranded DNA genome of 170 kb in length. The EcoRI and BamHI restriction fragments have been cloned and the restriction map of these sites is known (Arrand et al., 1981; see Figure la). The viral mRNAs are thought to be transcribed by the host cell RNA polymerase II. Some viral mRNAs have been mapped on to the viral genome by a Northern blotting analysis (Hummel and Kieff, 1982a). Also some of the viral proteins have been mapped by hybrid-selected translation (Hummel and Kieff, 1982b). We are analyzing EBV gene expression by determining the DNA sequence of the virus and mapping RNA polymerase II promoters on to the sequence. We previously sequenced the EcoRI-C fragment of the virus (Bankier et al., 1983) and mapped three RNA polymerase II promoters in this region of the virus (Farrell et al., 1983). RNAs from these promoters were all dramatically induced in response to TPA treatment of the cells. We have now determined the DNA sequence of the EcoRI-H and EcoRI-Dhet fragments of B95-8 EBV and mapped four more RNA poly-
'Present address: Ludwig Institute for Cancer Research, MRC Centre, Hills Road, Cambridge CB2 2QH, UK. 2Present address: LSU Medical Center, 1901 Perdido St., New Orleans, LA 70112, USA. *To whom reprint request should be sent.
©) IRL Press Limited, Oxford, England.
merase II promoters in these regions of the virus. Three of these promoters are substantially induced by TPA treatment of the cells but RNA from one of them is already present prior to TPA treatment. This promoter is therefore apparently active in the latent cycle of the virus as well as the productive growth cycle. The DNA sequence of the EcoRI-H and -Dhet regions and a detailed analysis of their coding properties will be published elsewhere.
Results and Discussion We have used in vitro transcription (Manley et al., 1980) to screen the EcoRI-H and -Dhet fragments (see Figure la for restriction map) for promoter activity. We used run-off transcription assays (Weil et al., 1979) to transcribe various restriction fragments from the EcoRI-H and -Dhet fragments. With this assay the start of transcription is located at a distance equal to the length of the observed transcript band from the restriction site marking the end of the template fragment. The direction of transcription can be established by transcribing an overlapping array of DNA templates. Transcription from promoters in the EcoRI-H fragment Five separate restriction fragments (indicated in Figure lb) from the EcoRI-H region of the virus were transcribed in the HeLa cell extract and the products analysed on agarose gels (Figure 2, tracks 1 -6). The products from the same transcription reactions were also analysed on polyacrylamide gels (Figure 2, tracks 7- 12). The only substantial transcript observed on the agarose gel analysis was from the H3c fragment (Figure 2, track 2) and was 460 bases in length. This band can also be seen on the polyacrylamide gel of this transcription reaction (Figure 2, track 8, arrowed). The Xho b fragment overlaps the H3c fragment, being 280 bases shorter at the left hand end and 483 bases longer at the right hand end. It can be seen (Figure 2, track 11 arrowed) that the Xho b template yields a similar transcript exactly 280 bases shorter than that derived from the H3c template. This demonstrates that the promoter giving rise to this transcript reads leftwards from base 460 of the EcoRI-H sequence. We call this promoter EH-LI (see Bankier et al., 1983 for nomenclature). There are a number of other faint bands in these transcription reactions but we have not been able to locate convincingly any other promoters in this region. Transcription from promoters in the Dhet region Transcription of the large BamA-R1 fragment yielded a prominent transcript of -1100 bases in length (Figure 3, track 1). This is from a promoter reading rightwards because transcription of the BamA DNA, which shares its right end with large BamA-R1 but not its left end, also gives the 1100 base RNA (Figure 3, track 2). This promoter, which we call ED-RI, therefore has a transcription start at about base 5650 of the EcoRI-Dhet sequence. The fainter band in both these transcriptions and also in lane 3 at - 1.9 kb was present even without added DNA with the particular extract used for these experiments. The band in the BamA transcription products at -4.6 kb may represent transcription from the EC-LI promoter (Farrell et al., 1983) in the EcoRI-C fragment. No -
1331
P.J.Farrell et al.
a
,
I
I
I
0
10
20
30
J a I
50
60
A I
I
I
40
I
I
70 80 90
*
I lI
*
_
T
I
E
B
C
H
A
r _I
C WWWWWW H F QUPO MSL E ZRK B G D TXV cb d WWWWWY e a
H
XS H3X
P lm
s
I
-d H3S l
s
lm
l
H3c
t
_//
I
l
IH3b
H3a
H3B I ..Bgl .. 1
_
Dhet
A
C
It
I
Nhet
Bam Hl
Bgl HRb
I
BamA-RI
-
Cla a
CI b Cla b DH2 2
Ri
Bam A
Ssta
1
Eco
_
I----
Xhob
0
-
I
Dhet
F
(C)
-.
I
100 110 120 130 140 150 160 170 kb
G1LG2 FNMVK *|*|
I
3
4
5
6
0
1
a
a
2 ,
3
4 I
5 .
DH4 6 I
7 .
8 *
a
9 10 11 12 13 I
I
I
I
Kb
Fig. 1. (a) Restriction map of EcoRI and BamHl sites in B95-8 EBV (Arrand et al., 1981). (b) Restriction fragments from EcoRl-H and EcoRl-Dhet segments used for in vitro transcription. Restriction sites for BglIl (Bgl), HindIll (H3), BamHI (B), Clal (C), SstI (S) and Xhol (X) are marked -Dhet regions. The sequences of the EcoRI-H and -Dhet fragments will be published separately. The numbering of nucleotides in each EcoRI in the -H and fragment is from the left-hand end of that fragment on the conventional map of the linear virus genome.
specific transcripts were detected with the Cla b fragment as template (Figure 3, track 5). Transcription of the DH2 clone cut with HindIII and analysis of the transcription products of the ED-RI promoter on a polyacrylamide gel (Figure 3, track 6) permitted a more accurate sizing of the transcript. The transcript migrates with a size of 990 bases, consistent with the ED-RI promoter transcribing rightwards from a start at 5650 of the Dhet sequence. Transcription of the HR b template yielded two RNAs of interest (Figure 3, track 3), one of 1000 bases and the other of 3.1 kb. Neither of these transcripts was obtained with the C/a a template (Figure 3, track 4), which shares its right end with HR b but not its left end, implying that both the promoters detected in HR b transcribe leftwards. The transcription starts are therefore located at 7650 (promoter EDL2) and 9660 (promoter ED-L1). Consistent with this, the Cla a template did give a faint transcript of 4.2 kb which could come from the ED-L2 promoter. The predicted transcript from the ED-LI promoter would be >6 kb long and such long transcripts are rarely made in the in vitro system. The ED-L2 promoter was located more accurately by transcribing DH4 cut with EcoRI and Sall (the BglII end was destroyed in the construction of this clone, see Materials and methods). A transcript of 750 bases was obtained (Figure 3, track 7) confirming that the ED-LI promoter transcribes leftwards from -
-
-
-
-
-
9660.
Transcription from the ED-LI, ED-L2, ED-RI and EH-L1 promoters was prevented by 0.5 jig/ml ca-amanitin (data not shown), indicating that RNA polymerase II was responsible for the transcription in each case. The promoters work in B95-8 cells To demonstrate that the promoters detected by in vitro transcription function in B95-8 cells, SI mapping experiments (Berk and Sharp, 1978) were performed for each promoter. Probes for these SI mapping reactions were made by primed synthesis on suitable M1 3 clones spanning the sites of trans1332
cription initiation. The probe for mapping the EH-LI transcription start was made from M13 clone CS186.RlH and extended from the sequencing primer through the start of the clone (base 361 of the H sequence) to a HinfI site at base 496 of the H sequence. With a transcription start at base 460 we would predict a protected fragment of 100 bases. In fact a major protected fragment of 97 bases is seen (arrowed) with RNA from TPA-treated B95-8 cells (Figure 4, tracks 5,6), suggesting an in vivo start at base 457 for the EH-L1 promoter. Only a very low signal was obtained with RNA from B95-8 cells not treated with TPA, (Figure 4, tracks 3,4). An approximately equal signal in the SI mapping assay was obtained with poly(A) + RNA (Figure 4, track 5) as with poly(A)- RNA (Figure 4, track 6) even though 50 times as much RNA was used in the poly(A)- RNA hybridisation. This indicates that the RNA transcribed from the EH-L 1 promoter is a polyadenylated RNA. RNA from the ED-L2 promoter was mapped using a probe made from M13 clone B503.HET, running from the primer to a Hinfl site at 7858 in the Dhet sequence. Therefore, the probe is 350 bases long including the primer and contains 303 bases complementary to the EBV sequence running from base 7556 to 7858. With a transcription start at 7645, a protected fragment of - 89 bases would be predicted. A major protected fragment of - 89 bases is indeed observed (Figure 4, tracks 11,12 arrowed), indicating an in vivo start for the EDL2 promoter at 7645. A substantial number of larger bands are seen in this assay indicating some degree of upstream starts or other RNAs around this region. The RNA from the ED-LI promoter is strongly induced by TPA (compare Figure 4, tracks 9, 10 with 11,12) and is a polyadenylated RNA (compare tracks 11,12). RNA from the ED-RI promoter was difficult to map because, although it is induced by TPA, it still seems to be present at only a very low level in B95-8 cells. The probe used was made from M13 clone D134.HET and is 405 bases long including the M13 primer, extending up to the EcoRI site of -
Epstein-Baf virus promoters
1
2
3
4
5
9 11
6
8_
10
_...
460
-
529
-
406
-
182
-
16 2
-
W.
~
12
-
Fig. 2. In vitro transcription of EcoRl-H fragments. Template DNAs from appropriate restriction digests were isolated from LGT agarose gels and transcribed as described previously. Products were analysed on agarose gels (tracks 1-6) or polyacrylamide gels (tracks 7- 12). Radioactive size markers on the agarose gels were Xenopus laevis rRNAs and a HindIIl digest of SV40 DNA and on the polyacrylamide gel (M) an MspI digest of pBR322 DNA. Template DNAs were (I) no DNA, (2) H3c, (3) H3a, (4) H3b, (5) Xho b, (6) Sst a, (7) no DNA, (8) H3c, (9) H3a, (10) H3b, (11) Xho b, (12) Sst a.
the Ml 3 vector on the far side of the insert. The probe contains EBV sequences from 5768 back to 5420 (349 bases). With a transcription start at 5650, a protected fragment of 120 bases would be predicted. There is a very heavy background of bands visible in these SI mapping reactions, largely because a much longer exposure than normal has been required to visualise the protected fragment. It can be seen that the only place (arrowed) where a band is induced in response to TPA treatment in the poly(A) + RNA (compare Figure 4, tracks 15 and 17) is indeed at 121 bases on the gel. This confirms the transcription start of the ED-RI promoter at 5648 going rightwards. Comparison of Figure 4 tracks 15 and 17 indicates that this RNA is induced by TPA and comparison of tracks 15 and 16 shows that it is a polyadenylated RNA. Clearly the strongest feature of the SI mapping experiment is the prominent band in the assays of poly(A) - RNA 135 bases long (Figure 4, tracks 16 and 18). We do not understand the significance of this band; we have not seen it -
-
-
in all our SI mapping experiments with probes in this region. It is either an SI artefact or perhaps reflects a chance homology of the probe with rRNA (since it is only in the poly(A) - reactions). Finally, it could represent a major constitutive poly(A) - RNA transcribed from this region of EBV but it is hard to believe that this would not have been observed previously. The ED-L1 promoter was mapped used a probe made from M13 clone B1.HET. The probe is 168 bases long including the M13 primer and extends to a base site at 9741. It contains 119 bases complementary to EBV extending from bases 9623 to 9741. A transcription start at 9665 would give a predicted protected fragment of -40 bases. Protected fragments of 35-41 bases were observed in the SI mapping reactions (Figure 4, tracks 21-24) indicating a transcription start for the ED-L I promoter leftwards from - 9660. In each case the signal obtained with poly(A) + RNA is about the same as that with poly(A) - RNA (compare tracks 21,22 and 1333 -
P.J.Farrell et al.
*
M
1:
-=1 q U
ows
so -I40
.'.9
3
w di
Fig. 3. In vitro transcription of the EcoRl-Dhet fragments. Templates were prepared as described in Figure 2. Tracks (1-5) agarose gels, tracks (6-11) polyacrylamide gels. Size markers for the agarose gels were as for Figure 2. Template DNAs were (1) BamA-Rl, (2) BamA, (3) HR b, (4) Cla a, (5) Cla b, (6) DH2 cut with HindIII, (7) DH4 cut with EcoRI and at a Sall site in the vector (because the BgII end was destroyed in the construction). Track (8) no DNA. Tracks (9) and (10) contain end-labelled DNA size markers (9) MspI digest of pBR322 (10) Hinfl, EcoRl digest of pBR322.
tracks 23,24), even though 50 times as much RNA was used in the poly(A) - RNA hybridisations. This indicates that the RNA from the ED-LI promoter is polyadenylated. A significant difference between this promoter and the others mapped so far lies in its response to TPA treatment of the cells. Whereas for other promoters (see above and Farrell et al., 1983) a dramatic increase in the amount of RNA present is seen in response to TPA treatment of the cells, for the ED-LI promoter only a very modest increase is seen (densitometry of the autoradiographs suggests about a 3-fold increase). The RNA from the ED-LI promoter is thus largely constitutive with respect to TPA and the ED-LI promoter appears to be one of a small subclass of promoters significantly active in the latent virus cycle. It has been known for some time that one of the major RNAs of the latently infected cells maps to this region of the virus (Hummel and Kieff, 1982a; Arrand and Rymo, 1982; Weigel and Miller, 1983). In all the SI mapping reactions, significant protection of 1334
the full length of EBV sequence in the probe was seen. This either represents RNA transcribed right through the entire region covered by all the probes or else hybridisation to EBV DNA contaminating the RNA preparations.
Organisation of reading frames in EcoRI-H and -Dhet We have previously mapped promoters in the EcoRI-C region of B95-8 EBV and correlated them with major open reading frames present in the DNA sequence (Farrell et al., 1983). We have performed similar analyses of the DNA sequence in the EcoRL-H and -Dhet regions and we have combined all the information in Figure 5. We know the complete DNA sequence of the region indicated (from EcoRI-E through EcoRI-Dhet) and that all the EcoRI fragments are contiguous. Although we have not reported any promoter analysis of EcoRI-E, part of this fragment (sequenced by R.Baer, personal communication) is included in Figure 5 to show the extent of the open reading frame downstream of the
Epstein-Barr virus promoters
I a -p
-p ow qb
_:., _O 111. I
_ 4^ _
"Ii
SF _1 '.
..
a
Fig. 4. SI mapping reactions. Probes are described in the text and Materials and methods. Tracks 1-6, mapping EH-LI promoter. Tracks 7- 12, ED-L2, Tracks 13- 18, ED-RI and tracks 19- 24, ED-LI In each case 1 yg of the relevant poly(A) + RNA or 50 Ag of poly(A) - RNA was used. Control RNA is from untreated B95-8 cells and TPA RNA is from B95-8 cells treated with TPA (see Materials and methods). Tracks 1, 7, 13, 19 are untreated probes. Tracks 2- 6, 8-12, 14-18 and 20-24 are SI mapping reactions. Tracks 2, 8, 14, 20, no RNA in hybridisation. Tracks 3, 9, 17, 21, poly(A)+ RNA from control cells. Tracks 4, 12, 18, 22, poly(A)- RNA from control cells. Tracks 5, 13, 15, 23, poly(A)+ RNA from TPA cells. Tracks 6, 14, 16, 24, poly(A)RNA from TPA cells. Note the unusual order of tracks in the ED-RI panel. Size markers (M) are an MspI digest of pBR322 DNA, end-labelled with Klenow DNA polymerase. .
EH-LI promoter. It can be seen that several of the promoters in front of open reading frames which are terminated close to polyadenylation sites (AATAAA). Because we have not yet established the splicing patterns in this region we are not able to connect exons to indicate complete mRNAs but we believe that the major reading frames we have indicated are all expressed by the virus at some stage of its life cycle. We have no evidence for any promoter transcribing through the terminal repeated portion of the EcoRI-Dhet fragment but this could still be accomplished by transcription leftwards from, for example, EcoRI-I or J or alternatively from a rightwards promoter in Dhet which we did not detect by in vitro transcription. It seems likely that there will be other promoters in the region we have examined which we have overlooked, probably because they may not transcribe efficiently in the HeLa cell extract. are
Sequence comparison of promoters
By a combination of in vitro transcription and SI mapping
5' ends of RNA from infected cells, we have identified a further four promoters in the EcoRI-H and -Dhet regions of B95-8 EBV. All these promoters gave rise to polyadenylated RNAs. Three of the four promoters apparently function during the productive virus cycle (they are induced by TPA) but one is active during the latent and productive virus cycles (the ED-LI promoter). We prevously identified three promoters in the EcoRI-C fragment of the virus and noted that they had remarkable sequence homology blocks in their upstream regions. We advanced the argument that different combinations of these homologous sequences, which lie in the region which would be predicted to contain upstream promoter elements, might coordinately regulate the promoter activity. The DNA sequences around the four new promoters found in the EcoRI-H and -Dhet regions are shown in Figure 6 along with the sequences of the three promoters in EcoRI-C. It can be seen that the pattern of sequence homologies upstream of the transcription starts is continued. As expected, each promoter has a sequence homologous to the 1335
P.J.Farrell et al.
X
T
c |b
D
V
C
H
1
3
2
4
6
5
7
8
10
9
11
12 13
14
1516
ED-Rl
1718
19 20 21
ECL2
EHL1
EcoRI
D het
EC-Ri
0
BamHl
Nhet
A
|d
262728293031
22 2324 25
3233343536
EDL2
EC-Ll
37
8 39
EDL1
13kb DELETION
Fig. 5. Organisation of reading frames in EcoRI-E, H, C and Dhet fragments of B95-8 EBV. Major open reading frames located by computer are indicated by horizontal arrows. Promoters are shown by arrows with circles and AATAAA (polyadenylation sites) by small vertical arrows. Four terminal repeats, the number that happen to be in the particular Dhet clone which we sequenced are shown. The scale is in kb from an arbitrary point in EcoRI-E. -80
-90
-60
70
-50
-40
-30
-20
-10
0
10
EC-L2
ACAGTATGTCATCAAAGCCACA|CCCCCATAATTC|GAGTGGTGTGGCC|GAGAAGCTG TATAAG CAGCATCTTTAGTGTACTTAAGGAAGTCATACAGCC
EC-Rl
AACCGCGTGG AGACTCTGA
EC-Li
ACCTCACGT CGACTCTGA CCGCCCCCTAATTCTTCTGTATAAGGGCAGGGGPTGGG TATTTAAIGGATCTATATGCCCTTCTCTACCTGCACCTCCAAATG
EH-Ll
TGTATG
G CT C
GGCtAGGACIAGGGG,TGTGGCC]AATATCCCAGCCGAGTCG TATTTAAAGGCCAGCGACGCCCCGGACCFACAAAAGCCGGACAA
GGA
eiAGGGGGCCAGGGACGGTGCCCGGACTCAGAAT TATTAAAICCGGGTGGCAGCTCCTGGCAGTCATTCATTCGGATAA
CTT
ATTCCTCTGGCCTTCTTTCCTTGTCCTTACTG TATAAAAGTCCACGAAAACAGCTGTGCCTCACTCTCGAGATGGT
ED-L2
CACCT
ED-Ri
CGTCAGGC
ED-Li
GCAAATCCCCCCGGGCCTACATCCCAAGAAACACGCGTTACTCTGACGTAGCCGCCCKACATAAbCCTCTCACACTGCTCTGCCCCCTTCTTTCCTCAACT
C
GGGCTTCAGGCTTATATGATA*ATGGGCGTGGC]AGAATAG TATAAGAIGCGAGGCCTGGGTGAGGAGAGTCCAGAGCAATGGCC
Fig. 6. Sequence comparison of the seven promoters so far located in B95-8 EBV.
TATA box (Corden et al., 1980) at 30 bases upstream from the transcription start. The EH-LI promoter has the sequence AGACTCTGG homologous with the EC-RI and EC-LI promoters. This homology is particularly remarkable since in all three promoters the sequence is exactly the same distance from the TATA box. Other sequence homologies between the EH-LI and ED-L2 promoters and between the ED-RI promoter and the EC-L2 promoter are also indicated in Figure 6. Interestingly, the only promoter which we have not found to be substantially homologous with the others is the ED-LI promoter. This is the promoter which is active in the latent cycle of the virus while all the others are only active in the productive virus cycle. -
Upstream promoter elements lie in the region containing the homologies Our suggestion that these homologous sequences might represent conserved upstream promoter elements makes the assumption that upstream promoter elements do indeed lie in this general region in EBV genes. To provide some limited evidence that this is the case, we have constructed some deletions of the EC-RI promoter. These were deletions coming in from the 5' side of the sequence shown in Figure 6. Ideally, the deleted clones should be assayed for transcription by reintroduction to a cell containing functional EBV but so far transfection into cells of the lymphoid type has not been very sequence
1336
successful. Instead the deletion clones were injected into the nuclei of Xenopus oocytes. This system transcribes a wide range of injected promoters and is sensitive to the deletion of upstream promoter elements. Because we were unsure of the stability of EBV mRNAs in the Xenopus oocyte, the deleted promoters were all fused on to the thymidine kinase gene of herpes simplex virus (lacking its own promoter), which is known to give a stable mRNA in oocytes. The deletions, which are described in detail in Materials and methods, were from the 5' side to - 461, - 109, - 49 and - 38, numbered from the transcription start. SI mapping was used to identify transcripts from the mutated EC-RI promoters. Figure 7 shows the results of the transcription from the deleted EC-RI promoters. The pair of protected fragments 70-75 bases in length (arrowed) represent initiation at the EC-RI promoter. It can be seen (compare tracks 4,5) that deletion from the 5' side to - 109 has little effect on the transcriptional efficiency of the promoter compared with the - 461 deletion, which we are considering to be the wild-type promoter; if anything, in some experiments, transcription was a little more efficient with the - 109 template. In contrast, deletion to -48 or to - 38 dramatically reduces the amount of initiation at the ECRI promoter (Figure 7, tracks 6,7). This shows that some essential sequence element for this promoter function lies between - 109 and -48. Although this experiment in no way accurately defines the promoter upstream element, it does at
Epstein-Barr virus promoters
1 2 3 4 5 6 7
shorter protected fragments -45 bases in length in this experiment is uncertain. They were observed in every oocyte injection experiment with these templates but corresponding bands are not found when SI mapping B95-8 cell RNA or RNA from tissue culture (COS) cells transfected with the same set of DNA clones. To some extent their relative intensities paralleled those of the 70-75 base protected fragments. It is possible that they represent adventitious initiation from a TATA box in the DNA sequence -30 bases downstream from the EC-RI TATA box. In summary, four RNA polymerase II promoters have been mapped in the EcoRI-H and -Dhet regons of B95-8 EBV. All these promoters function in B95-8 cells induced to a productive virus cycle with TPA but one of the promoters (ED-LI) is also active in cells not treated with TPA. It seems to be a promoter active in the latent virus cycle. The EH-L1, ED-L2 and ED-RI promoters all lie just upstream of the start of substantial open reading frames (Figure 5) and these all have poly(A) addition sites (AATAAA) at their 3' ends. Probably these promoters give rise to simple unspliced mRNAs. The ED-LI promoter does not lead into a simple open reading frame and we believe that up to three exons may be spliced together to make the mRNA from this promoter. The pattern of finding homologous DNA sequences in the upstream promoter regions of EBV promoters (Farrell et al., 1983) has been extended in the EcoRI-H and -Dhet regions. It would be interesting to exchange upstream promoter elements between the ED-LI promoter and some of the other promoters and test whether this caused the proteins specified by the other promoters to be expressed during the latent virus cycle. Materials and methods The EcoRI-H and EcoRl-Dhet clones were isolated from a cosmid library (Arrand et al., 1981). Both fragments were sequenced by the M13/dideoxynucleotide method using random clones generated by sonication. The DNA sequence and a detailed analysis of its coding properties will be published
Fig. 7. SI mapping reactions of RNA transcribed in Xenopus oocytes upon injection with 5' deletions of the EC-RI promoter. The detailed structure of the deletions is described in Materials and methods. The SI mapping probe (from LEG 156) is described in the text. Track 1, untreated probe. Tracks 2-7, SI mapping. Track 2, hybridization to no RNA, track 3, uninjected oocyte RNA, tracks 4-7, oocytes injected with (4) -461 deletion, (5) 109 deletion, (6) -49 deletion, (7) -38 deletion. Markers were an Mspl digest of pBR322 DNA. -
least confirm that sequences important for the function of this promoter lie in the same region as the sequences conserved between the different EBV promoters. The origin of the
elsewhere. Restriction fragments for transcription in the EcoRl-H region were prepared by appropriate digestion of the cosmid clone. Fragments for transcription in the EcoRl-Dhet regions were isolated both from the cosmid clone and from the subclones DH2 and DH4. The subcloning of Dhet involved cutting the Dhet EcoRI fragment into four pieces with BgIII and Hindlll (see Figure lb). These four fragments were then cloned into either the large EcoRI-BamHI fragment or the large HindIII-BamHI fragment of pAT153 as appropriate. The M13 subclones used for in vitro transcription and for preparation of S1 mapping probes were isolated during the random sequencing programme. In vitro transcription in HeLa (S3) whole cell extracts (Manley et al., 1980) and analysis of the RNA products by electrophoresis through denaturing polyacrylamide gels or agarose gels after glyoxylation was as described previously (Farrell et al., 1983). Single-stranded probes for mapping 5' ends of B95-8 cell RNA were prepared by the 'prime cut' method (D.Bentley, personal communication) as described previously (Farrell et al., 1983). The M13 clones used for preparation of SI mapping probes were called CS186.RIH, B503.HET, D134.HET and B1.HET. CS186.RIH was an M13mp7 clone with the insert at the Hincll site and the others were all M13mp8 clones with the insert at the Sma site. B95-8 cells were grown in RPM1 1640 containing 10°7o foetal calf serum. Permissive infection was induced by adding 30 ng/ml TPA and harvesting the cells 3 days later. Cytoplasmic RNA was extracted and chromatographed on oligo(dT)-cellulose as described previously (Farrell et al., 1980). The deletions for the oocyte injection experiment were constructed using an M 13 clone (LEG 156) spanning the region of the EC-R I promoter. This clone contains EBV DNA from base 5506 to 6033 of the EcoRl-C sequence (Bankier et al., 1983) inserted at the HinclI site of M13mp7. The - 461 clone was made by isolating the small BamHl fragment from LEG 156 and cloning it into the large BamHI/Bgtll fragment of the vector pXTKIO (Pelham,
1337
P.J.Farrell et al. 1982). A clone with the EC-RI promoter aiming towards the pXTKI0 thymidine kinase gene was selected. The - 109, -49 and -38 deletions were made by isolating the EcoRI insert out of LEG 156, and cutting it with either Clal, Ball or Hinfl, respectively. The ends were repaired with Klenow polymerase and synthetic HindlII linkers (BRL, 10 mer) were ligated onto the DNA fragments. After digestion with HindIlI and BamHI, the appropriate HindlII/BamHl fragments containing the deleted promoter sequences were isolated from a gel. These were then cloned into the large HindII/BgIII fragment of pXTKI0. All the clones were then characterised by sequencing; the deletion to the BalI site would have been expected to go to - 53 but in fact went to -49, perhaps due to some exonuclease in the Ball enzyme used. DNA was injected into Xenopus oocytes and total RNA was isolated as described (Bienz and Pelham, 1982). The probe for SI mapping of the transcribed RNA came from M13 clone LEG 156 and contained EBV sequence from base 6033 to the Ball site at base 5914 (Bankier et al., 1983). The probe is 155 bases long and contains 120 bases complementary to EBV. Because a few bases from the M13 linker are present between the EBV sequences and the pXTK10 sequence in the templates injected into the oocytes, the fragments protected by the LEG 156/Ball probe here are a little longer than those protected by B95-8 cell RNA (Farrell et al., 1983). SI mapping reactions with oocyte RNA were as for mapping B95-8 cell DNA except that 3 jg of total oocyte RNA was used.
Acknowledgements We would like to thank Mariann Bienz for performing the oocyte injections and Beverly Griffin and John Arrand for the cosmid library. PJF was supported by an MRC training fellowship, CS by a fellowship of the MRC of Canada and PD by a NATO fellowship.
References Arrand,J.R., Rymo,L., Walsh,J.E., Bjork,E., Lindahl,T. and Griffin,B.E.
(1981) Nucleic Acids Res., 9, 2999-3014. Arrand,J.R. and Rymo,L. (1982) J. Virol., 41, 376-389. Bankier,A., Deininger,P., Farrell,P.J. and Barrell,B. (1983) Mol. Biol. Med.,
in press. Berk,A.J. and Sharp,P.A. (1978) Proc. Nati. Acad. Sci. USA, 75, 1274-1278. Bienz,M. and Pelham,H.R.B. (1982) EMBO J., 1, 1583-1588. Corden,J., Wasylyk,B., Buchwalder,A., Sassone-Corsi,P. and Kedinger,L. (1980) Science (Wash.), 209, 1406-1414. Epstein,M.A. and Achong,B.G., eds. (1979) The Epstein-Barr Virus, published by Springer Verlag. Farrell,P.J., Broeze,R.J. and Lengyel,P. (1980) Nature, 279, 523-525. Farrell,P.J., Deininger,P., Bankier,A. and Barrell,B. (1983) Proc. Nati. Acad. Sci. USA, 80, 1565-1569. Hummel,M. and Kieff,E. (1982a) J. Virol., 43, 262-272. Hummel,M. and Kieff,E. (1982b) Proc. Natl. Acad. Sci. USA, 79, 56985702. Manley,J.L., Fire,A., Cano,A., Sharp,P. and Gefter,M. (1980) Proc. Natl. Acad. Sci. USA, 77, 3855-3859. Miller,G., Shope,T., Lisco,H., Stitt,D. and Lipman,M. (1972) Proc. Natl. Acad. Sci. USA, 69, 383-387. Pelham,H.R.B. (1982) Cell, 30, 517-528. Tooze,J., ed. (1980) DNA Tumor Viruses, published by Cold Spring Harbor Laboratory Press, NY. Weigel,R. and Miller,G. (1983) Virology, 125, 287-298. Weil,P.A., Luse,D.S., Segall,J. and Roeder,R.G. (1979) Cell, 18, 469-484. zur Hausen,H., Hecker,E., O'Neill,F.J. and Freese,U.K. (1978) Nature, 272, 373-375.
1338
