Sep 10, 1992 - The EMBO Journal vol. 12 no. 1 pp. 167 - 175, 1993. The Epstein Barr virus nuclear antigen 2 interacts with an EBNA2 responsive cis-element ...
The EMBO Journal vol. 12 no. 1 pp. 167 - 175, 1993
The Epstein Barr virus nuclear antigen 2 interacts with an EBNA2 responsive cis-element of the terminal protein 1 gene promoter Ursula Zimber-Strobi, Elisabeth Kremmer1, Friedrich Grasser2, Gabriele Marschall, Gerhard Laux and Georg W.Bornkamm Institut fOir Klinische Molekularbiologie und Tumorgenetik, Institut fiir Immunologie, im Forschungszentrum fiir Umwelt und Gesundheit, GSF, Marchioninistrasse 25, 8000 Munchen and 2lnstitut fOir Medizinische Mikrobiologie, Universititsklinikum, Haus 47, 6650 Homburg/Saar, Germany Communicated by W.Schaffner
The Epstein-Barr virus protein EBNA2 acts as a transcriptional activator of cellular and viral genes and plays a crucial role in the immortalization of human primary B-cells by EBV. We have shown previously that EBNA2 transactivates the promoters of the latent membrane antigens LM1P, TP1 and TP2. The promoter of the TPI gene was chosen as a model system to study the molecular mechanism of EBNA2 mediated transactivation. To identify an EBNA2 dependent cisacting element, various TPI promoter-reporter gene constructs were transfected in the absence and presence of an EBNA2 expression vector into the established Bcell line BL41-P3HR1. We were able to delineate an 81 bp EBNA2 responsive region between -258 and -177 relative to the TPI RNA start site. The element worked in either orientation and could mediate EBNA2 dependent transactivation on a heterologous promoter. Electrophoretic mobility shift assays revealed three specific protein-DNA complexes formed with sequences of the EBNA2 responsive element. Two of these were not cell type specific, but the third was detected only in EBNA2 positive cell extracts. Gel-shift analysis in the presence of EBNA2 specific monoclonal antibodies revealed that EBNA2 is a component of the third complex. Thus, these experiments demonstrate that EBNA2 interacts with an EBNA2 responsive cis-element of the TPI promoter. Key words: DNA -protein interaction/EBNA2/Epstein Barr virus/terminal protein 1 /transactivation
Introduction Epstein-Barr virus (EBV), a widespread human herpesvirus, is the aetiological agent of infectious mononucleosis, a self-limiting lymphoproliferative disease. The virus is associated with two human malignancies, namely Burkitt's lymphoma (a B-cell neoplasia) and nasopharyngeal carcinoma, which is composed of poorly differentiated epithelial cells. EBV has a dual tropism: it infects primary resting B-cells and epithelial cells. Virus production occurs only in epithelial cells, whereas B-cells are latently infected. In vitro infection of resting Blymphocytes leads to the outgrowth of permanently virus© Oxford University Press
transformed lymphoblastoid cell lines. In these immortalized cell lines the virus is maintained as an episomal molecule at high copy number (Lindahl et al., 1976) and only a few viral genes are expressed. These genes code for six nuclear antigens, EBNA1, 2, 3a, 3b, 3c and -LP, and three membrane proteins, the latent membrane protein (LMP) and terminal proteins (TPs) 1 and 2, also denoted by LMP2A and B (for review see Kieff and Liebowitz, 1990). These gene products seem to be important for induction and maintenance of B-cell immortalization. However, the individual functions of these gene products in latent infection and growth transformation are only poorly understood. EBNA2, one of the best studied latent gene products, is absolutely necessary for B-cell immortalization. One viral mutant (P3HR1) carrying a 6.6 kb deletion (Bornkamm et al., 1982; Jeang and Hayward, 1983) that removes EBNA2 and part of EBNA-LP, has lost the ability to transform primary B-cells. Reintroduction of EBNA2 into the P3HRl virus by homologous recombination reconstituted the full transforming capacity (Cohen et al., 1989; Hammerschmidt and Sudgen, 1989). In natural EBV isolates, two alleles of the EBNA2 gene exist coding for two proteins (EBNA2A and EBNA2B) with 57% homology (Dambaugh et al., 1984; Adldinger et al., 1985; Zimber et al., 1986). There have been some reports that EBV strains carrying EBNA2A could transform primary B-cells more efficiently than those containing an EBNA2B allele (Rickinson et al., 1987; Cohen et al., 1989). The EBNA2 gene codes for a 487 amino acid phoshorylated polypeptide (Grasser et al., 1991). Characteristic features of the primary structure of the protein include a proline stretch of -40 amino acids near the N-terminus, a negatively charged C-terminus and a glycine/arginine-rich region. It has recently been shown that EBNA2 acts as a transcriptional activator: it stimulates the transcription of the cellular genes CD21, CD23 and c-fgr (Calender et al., 1987; Wang et al., 1987; Abbot et al., 1990; Cordier et al., 1990; Knutson, 1990) as well as the expression of the viral genes LMP, TPI and TP2 (Fahraeus et al., 1990; Ghosh and Kieff, 1990; Zimber-Strobl et al., 1991). Although a transactivation domain has been mapped in the C-terminus of EBNA2 (Cohen and Kieff, 1991), it is not yet known whether EBNA2 transactivates several genes by interacting with EBNA2 responsive cis-elements or whether it does so by modulating the expression or modification of transcription factors. A consensus sequence in the promoter region of the EBNA2-regulated genes has not been described until now. To elucidate the mechanism of EBNA2 mediated transactivation, we focused our interest on the TPI promoter region. The TP gene is transcribed across the terminal repeats of the EBV genome (Laux et al., 1988; Sample et al., 1989). A functional transcription unit can only be created if the linear molecule is circularized. The TATA-box of the TPI promoter is located at the right hand end of the linear viral 1 67
U.Zimber-Strobl et al.
A III
.zzzL--45 rzzzzzzzzzz4llll
-
124
- 177
B
- 258
"IlllzzzriI
CAT-constructs -45
- 432 -
-
543
804
Stimulation
BL41-P3HR1/ 1
2.1
2.1
-124
0.8
1.4
1.75
-177
1.5
1.2
0.8
-258 -432
1.0 0.7 0.3
10.5 7.2
10.5 10.4
5.3
17.7
-543
A -1 47/-258
77777M,,,,{,,,,zz//d ..
%acetylation BL41-B95-8
BL41-P3HR1
C RLU 100000
10000 -
1 000 E
tHi I1 I1 0
I1
l
n.,I
-45
-124
-177
-258
-432
-543
-804
1,7
1,1
1,2
25,6
50,9
57,2
71
A
-147/ -258
transfected LUC-constructs
0,72
fold induction
by EBNA2
Fig. 1. Transfection of different TPI promoter-reporter gene constructs in the absence and presence of EBNA2. (A) Schematic representation of the TPJ promoter-reporter gene constructs. The restriction enzymes used to construct the recombinant plasmids are indicated. The positions of the TPJ promoter fragments denote their 5'-end relative to the TPJ transcription start site. Open boxes represent the reporter gene, hatched boxes the TPI promoter region. (B) CAT activities after transfection of the TPI promoter-CAT constructs into EBNA2 negative BL41-P3HR1 cells and EBNA2 positive BL41-B95-8 cells. Median activities of two independent experiments and the fold inductions by EBNA2 are indicated. (C) Relative light units (RLU) measured after transfections of TPI promoter-LUC constructs in the absence and presence of EBNA2 expression vectors into BL41-P3HR1 cells. Open boxes, transfections without EBNA2; hatched boxes, cotransfection with the EBNA2 expression vector pU294-6, coding for wild-type EBNA2; dotted and black boxes, cotransfections with EBNA2 expression vectors p604 and pU812/9 respectively, coding for mutated EBNA2. The median activities of two to five independent experiments and standard deviations are indicated. genome at position 166 469 of sequence (Baer et al., 1984). In this report we describe an
the B95-8 virus DNA
element that mediates transactivation by EBNA2 to the TPI promoter. In addition we demonstrate specific protein-DNA interactions within this region and provide evidence that EBNA2 itself is a component of this protein-DNA complex. 168
Results Identification of an EBNA2 responsive cis-element within the TP1 promoter region EBNA2 transactivated a TPJ promoter-CAT construct containing 804 bp upstream of the TPI cap site (ZimberStrobl et al., 1991). Plasmids containing progressive
EBNA2 interaction with TP1 promoter
A
B 0
*
B.
..*
*---------@@
-
-
WT Mut
WT Mut
-
W1 MLIt
-
81 DP 2
135bp 2
135bp1t
-
W7
VT Mut
mrinimal
8l0E
promoter 1,2
17
0.9
1 8
37
1.0
3
22
1.8
0.5
2
0.2
0.2
EDNA2
-
CMV-
CAT-
LTR
corstructs
0,2
100
%
2
zf
...
.glob.TATA 25
0,7
EBNA2
CMV-
CAT-
LTP
conStructs
100
I/. acetylation
acetylation
Fig. 2. 81 bp (-177 to -258) of the TPI promoter are sufficient to mediate transactivation by EBNA2. (A) Results of CAT assays of transfections of CAT constructs containing the EBNA2 responsive region (pTPl-CAT/135bp and pTP1-CAT/81bp) in front of the TPI minimal promoter (pTPI-CAT/-45) without EBNA2 (-), with pU294-6 coding for wild-type EBNA2 (WT) and with p604 coding for mutated EBNA2 (Mut) into BL41-P3HR1 cells. The element was cloned in the correct (135bp/l, 81bp/1) and inverse (135bp/2, 81bp/2) orientations. (B) CAT assays after transfections of CAT constructs containing the TPI promoter region from position - 124 to -258 in front of the ,3-globin minimal promoter into BL41-P3HR1 cells without EBNA2 (-) and with the EBNA2 expression vector pU294-6 (+). The element is cloned in either orientation (f135bp/1, correct; ,B135bp/2, inverse).
truncations of the 5' region of the TPI promoter linked to the reporter genes chloramphenicol acetyltransferase (CAT) or luciferase (LUC) were constructed (Figure IA) to identify an EBNA2 responsive cis-element within this promoter. These constructs were transiently transfected into the EBNA2 negative BL41-P3HR1 cell line and into EBNA2 positive BL41-B95-8 cells. A cytomegalovirus enhancer-long terninal repeat-CAT construct (CMV-LTR-CAT) was used as a positive control. Transfections of the TPI promoter-CAT constructs into BL41-P3HR1 cells resulted in very low or undetectable CAT activities (Figure 1B). Transfections of constructs containing -45, -124 and -177 bp upstream of the TPI cap site into BL41-B95-8 cells resulted in low CAT activities similar to those obtained with transfections into BL41-P3HRI cells. However, CAT activities significantly increased after transfections of CAT constructs containing sequences upstream of position -177 relative to the TPI cap site (Figure 1B). These results suggested that an EBNA2 responsive region is located between -177 and -258. We performed cotransfection experiments in order to prove that the increased TPI promoter activity in BL41-B95-8 cells is mediated by EBNA2. TPI promoter deletion constructs driving the reporter gene luciferase were transfected into BIAl-P3HRI cells either without EBNA2 or with an EBNA2A (pU294-6) expression vector or with EBNA2A expression vectors carrying mutations within the EBNA2 open reading frame (ORF). The EBNA2 mutation p604 contains a linker insertion after one-third of the EBNA2 ORF, leading to a truncated EBNA2 gene product. In the EBNA2 mutation pU812/9, amino acids 248-382 of EBNA2 are deleted. The EBNA2 gene and its mutants are under the control of the natural promoter of EBNA2. The obtained relative light units (RLU) of the LUC assays confirmed the CAT data described above (Figure IC). Only
LUC constructs containing at least 258 bp of the TPI 5' flanking region were inducible by EBNA2 (by factors of 25-71), whereas the three smallest constructs were unresponsive to EBNA2. In order to demonstrate that a specific EBNA2 responsive element has been identified, we removed the TPI promoter region between positions -258 and -147 in the TPI promoter reporter gene construct pTP1-LUC/-543, leaving the boundaries of the TPI promoter intact. This construct was transfected into BL41-P3HR1 cells together with the EBNA2 expression vectors mentioned above. As shown in Figure IC, induction of LUC activities by EBNA2 could not be observed after transfection of the TPI promoter LUC construct lacking the TI' promoter region between -258 and -147. This was confirmed by transfection of the same construct with the CAT reporter gene (data not shown). These data show that sequences between -258 and -147 are required for EBNA2 mediated transactivation. The EBNA2 responsive element is orientation independent and active on a heterologous promoter The putative element was cloned in front of the TPI minimal promoter (+60/-45) and the CAT reporter gene to study whether the EBNA2 responsive region is active even when separated from its cognate TPI promoter region. Two DNA constructs were made carrying a larger (-124 to -258) and a smaller fragment (-177 to -258) in both orientations. In these constructs the suggested EBNA2 responsive element is surrounded by vector sequences. The distance of the element from the TPI TATA box is comparable to the native situation. All DNA constructs were transfected in the absence and presence of EBNA2 expression vectors into BL41-P3HR1 cells. pU294-6 and p604, coding for wild-type EBNA2 and truncated EBNA2, respectively, were used as EBNA2 expression vectors. Transfection of all constructs 169
U.Zimber-Strobi et al.
.,PI~~~~k ..wsU+bStWof~~~~~~~~~~~~~~ I':, Bf
. --
n.
^^~~~~~~~-
^
to - 0
fa
-i
mm
. is ..
&
I
Fig. 3. Analysis of DNA-protein interactions within the TPI promoter. Autoradiogram of an EMSA is shown made with nuclear extracts of BL41-P3HRI cells which were incubated with a 32plabelled fragment spanning the TPI promoter region from position -12 to -258 relative to the TPI transcription start. Competition studies were performed with oligonucleotides covering the EBNA2 responsive region identified in the CAT and LUC analyses. The positions of the oligonucleotides in the EBV genome are described in Materials and methods. Unlabelled oligonucleotides were added to the gel-shift reaction in a 100-fold molar excess.
resulted in a significant higher CAT activity in the presence of the EBNA2 expression vector pU294-6, regardless of the orientation of the element relative to the promoter (Figure 2A). To determine whether the EBNA2 responsive region contains sequences that can regulate a heterologous promoter, we cloned the TPI promoter region from positions -124 to -258 upstream of the 3-globin minimal promoter (Westin et al., 1987). Cotransfections of the EBNA2 expression vector pU294-6 resulted in significantly increased CAT activities in the presence of reporter plasmids carrying the EBNA2 responsive region (Figure 2B). In this assay, the element conferred EBNA2 inducibility in an orientation independent fashion as has been demonstrated by the TPI promoter constructs. From these data we conclude that an 81 bp region within the TPI promoter harbours an EBNA2 responsive ciselement, which we will call EBNA2RE. EBNA2RE is active in both orientations and able to regulate a heterologous promoter. Specific DNA -protein interactions within the EBNA2 responsive region Gel-retardation assays were performed to identify specific DNA-protein interactions within the TPI promoter. A DNA fragment spanning the TPI promoter from position -12 to -258 was end-labelled and incubated with nuclear extracts of BL41-P3HR1 cells. The result of the electrophoretic mobility shift (EMSA) is shown in Figure 3. EMSA revealed one prominent DNA-protein interaction (complex I) and two weaker complexes (II and III). To investigate whether some of these complexes are formed with sequences located in the EBNA2 responsive region (-258
170
_
^
free oligo
Fig. 4. EMSAs using 054bp (EBV positions 166 236-166 289 according to the EBV sequence of Baer et al., 1984) as radioactively labelled probe and nuclear extracts of the promyelocyte line HL60, cervical carcinoma line HeLa, the EBV negative B-cell line BL41 and four EBV positive B-cell lines: BL41-P3HRI (EBNA2 negative), HH514 (EBNA2 negative), Jijoye (EBNA2B) and M-ABA (EBNA2A).
to - 177), four overlapping oligonucleotides [01 (-262/-230), 02 (-239/-209), 03 (-219/-181) and 04 (-196/-158)] covering these sequences were synthesized and added to BL41-P3HR1 extracts. Complexes I and III could be competed with 01 and 02 covering sequences between -262 and -209. None of the complexes could be competed by the adjacent oligonucleotides 03 and 04. In order to study the interaction of TPI promoter sequences from position -262 to -209 with cellular proteins in more detail, the 54 bp fragment was end-labelled and incubated with nuclear extracts of several cell lines (HL60, HeLa, BL41, BL4l-P3HR1, HH514, Jijoye, M-ABA and Raji) (Figures 4 and SA). Similar complexes were formed with the 54 bp fragment as with the 247 bp fragment used in the EMSA described above. Again one prominent (I) and two weaker complexes (II and III) could be detected. Complexes I -III were formed in all cell extracts tested and therefore seem to be cell-type unspecific. Notably, additional complexes (IVA and IVB) could be detected with extracts of the EBNA2 positive cell lines Jijoye, M-ABA and Raji, which were not found with extracts of EBNA2 negative cell lines. The specificity of the DNA-protein interactions was analysed in competition experiments following addition of the unlabelled 54 bp oligonucleotide to EMSAs with Raji, HH514, Jijoye and M-ABA extracts (Figure 5A and B). Only complex II could not be competed and appears to be unspecific. Additionally, we tested the ability of 0 1 and 02 to prevent formation of protein-DNA complexes formed with the 54 bp oligonucleotide and nuclear proteins of Raji cells. All specific complexes could be competed by 0 1 and 02 with the same efficiency as by 054bp (Figure 5A). The
EBNA2 interaction with TP1 promoter
B
A 54bp -..
01 02 a__.
llrn
-
23
a-IV -
aS -a
-
IVA
--.
_
'l1~'
:1
;t'
b
d
--
*
IVb
I- III
W a
F
*. X
£
*
free
_
oligo free ol!go C 262
-
~
~
209
-:1 --~
-2-2 6
Fig. 5. Competition assays demonstrate the specificity of complexes I, III and IVA. (A) Autoradiogram of an EMSA. Nuclear extracts of Raji cells (EBNA2A positive B-cell line) were incubated with the labelled oligonucleotide 054bp. Unlabelled oligonucleotides 054bp, 01, 02 and 1 lmut were added to the reaction mixture in 10-, 50- and 100-fold excesses. The DNA-protein complexes were separated on a 5% polyacrylamide gel. (B) Nuclear extracts of HH514, Jijoye and M-ABA cells (indicated by 1, 2 and 3 respectively) were incubated with the radioactively labelled 054bp oligonucleotide. For competition unlabelled oligonucleotide 054bp was added in 100-fold excess to the reaction mixture. The resulting products were separated on a 4% polyacrylamide gel. (C) Schematic representation of the oligonucleotides used in the EMSA. The indicated positions correspond to the 7PJ transcription start site. An 11 bp motif duplicated in 054bp is marked by dotted boxes, the 10 bp overlapping region of 01 and 02 by hatched boxes.
two oligonucleotides share a 10 bp overlapping region and a common motif of 11 bp (Figure 5C). To test whether the
I1 bp motif was involved in the protein -DNA interaction, these sequences were truncated (oligo 1 imut, Figure 5C). As shown in Figure 5A, the truncated oligonucleotide has completely lost the ability to prevent formation of the DNA-protein complexes. This suggests that the 11 bp motif is important for protein -DNA interaction within the EBNA2 responsive region.
EBNA2 is present in DNA -protein complexes formed with the EBNA2 responsive region Differences between EBNA2 positive and negative cell lines detected in the EMSAs described above were low mobility complexes IVA and IVB. To test whether EBNA2 was present in any of the complexes described above, DNA binding assays were performed with HH5 14, Jijoye and MABA extracts (EBNA2 negative, EBNA2B positive and EBNA2A positive, respectively) in the presence and absence of monoclonal antibodies raised against EBNA2A. The monoclonal antibody EBNA2-R1 is able to detect EBNA2A as well as EBNA2B, whereas anti-EBNA2-R2 recognizes
EBNA2A but has only a very low affinity for EBNA2B in immunoprecipitation (Kremmer et al., in preparation). Two nonspecific antibodies, raised against murine CD45, of the same isotype as the anti-EBNA2-Rl and anti-EBNA2-R2 antibodies were used as a control. The presence of the antiEBNA2-R1 antibody resulted in a supershift of complex IVB in Jijoye and complex IVA in M-ABA extracts (Figure 6A). By contrast, anti-EBNA2-R2, which recognizes EBNA2A with a high affinity but EBNA2B with a very low affinity, led to a supershift of complex IVA but not of complex IVB. Both control antibodies had no effect on the mobility of complexes IVA and IVB. We have repeated the experiments with two further EBNA2 monoclonals, anti-EBNA2-R3 and antiEBNA2-R1 1. Both antibodies recognized EBNA2A, whereas EBNA2B was recognized only by EBNA2-R3. We got the same results as with the antibodies described above. Both antibodies caused a supershift of complex IVa, but only anti-EBNA2-R3 was able to supershift complex IVb (data not shown). These results imply that complex IVA contains EBNA2A and complex IVB EBNA2B. None of the other complexes 171
U.Zimber-Strobi et al.
.-
!; . I-
Er F !
S
a 6
ii:
0
W
i.:
BIX
:.Z
A'..-,
A
a as i".f..i...
Fig. 6. EBNA2 is a component of complex IV. (A) Effect of monoclonal EBNA2 antibodies on TPJ promoter DNA-protein complexes. Gel-shift reactions were incubated with the indicated antibodies. Anti-EBNA2-Rl and anti-EBNA2-R2 are monoclonal rat antibodies raised against EBNA2A. Anti-EBNA2-RI recognizes EBNA2A as well as EBNA2B; anti-EBNA2-R2 recognizes EBNA2A with the same affinity as anti-EBNA2-Rl, but only recognizes EBNA2B with a very low affinity. Incubation reactions were separated on a 4% polyacrylamide gel. Only complexed DNA is visible on the autoradiogram since the free oligonucleotide was run off the gel. 1, 2 and 3 indicate gel-shift reactions with extracts of cell lines HH514 (EBNA2 negative), Jijoye (EBNA2B) and M-ABA (EBNA2A) respectively. Anti-mouse CD45 IgGI and IgG2a are two control monoclonal antibodies of the same Ig subclasses as anti-EBNA2-RI and anti-EBNA2-R2, respectively, and recognize neither EBNA2A nor EBNA2B. (B) Effect of in vitro translated EBNA2A on the reconstitution of complex IV. Gel-shift reactions were incubated with the indicated nuclear extracts (NE). 3 Al reticulocyte lysate (IVT) programmed with EBNA2 RNA or BMV RNA and 1 1d supernatant of monoclonal antibodies (anti-EBNA2-R3 or antimouse CD45) were added as indicated above. The reactions were preincubated for 5 min at room temperature before adding the radioactively labelled oligonucleotide 054bp. The DNA protein complexes were analysed on a 4% polyacrylamide gel. The free oligonucleotide was run off the gel.
Discussion
were affected by the monoclonal EBNA2 antibodies, suggesting that they do not contain EBNA2. We confirmed that EBNA2 is really a component of the protein -DNA complex IV by studying the ability of in vitro synthesized EBNA2 to form this complex. The EBNA2 ORF was translated in a reticulocyte lysate. The in vitro synthesized EBNA2 was added to the EBNA2 negative BL41-P3HR1 extracts and EMSAs were performed. As shown in Figure 6B the in vitro translated EBNA2 formed a more slowly migrating complex. In contrast, addition of a control reticulocyte lysate containing in vitro translated brome mosaic virus (BMV) RNA did not induce this complex. That the complex induced by in vitro translated EBNA2 did contain the added EBNA2 was confirmed by assaying the effect of monoclonal antibodies. AntiEBNA2-R3 but not the control antibody (anti-CD45) induced a supershift of the newly formed complex. An additional complex (*) appeared in the presence of nuclear extracts,
It is assumed that the contribution of EBNA2 to B-cell immortalization is conferred by its ability to transactivate viral and cellular genes. In order to understand the process of immortalization at a molecular level, it is essential to elucidate the mechanism of EBNA2 mediated transactivation. We chose the TPI promoter as a model system for studying EBNA2 dependent transcription. As shown previously, TPI transcription was dependent on the presence of EBNA2 in Burkitt's lymphoma cells (Zimber-Strobl et al., 1991). Our first aim was to identify a cis-acting element mediating EBNA2 responsiveness. Therefore we transfected different TPJ promoter-reporter gene constructs in the presence and absence of EBNA2. By this approach we were able to identify an 81 bp element between positions 177 and -258 which is sufficient to confer EBNA2
EBNA2 and monoclonal anti-EBNA2 antibodies in this experiment. This complex was not always reproducible and is therefore not further considered. Besides the in vitro translated EBNA2 diminished formation of complexes I and III. Whether this effect is specific or not is still unclear. The in vitro translated EBNA2 was not able to bind DNA in the absence of nuclear extracts, suggesting that EBNA2 interacts with DNA not directly but indirectly.
The EBNA2 responsive element (EBNA2RE) works in either orientation and is able to regulate a heterologous promoter. These results imply that EBNA2 responsiveness is conferred by an enhancer-like cis-element. This observation is in agreement with previous reports. EBNA2 responsive elements were also identified within the LMP, BamHI-C and CD23 promoters and it has been reported that they work in an orientation independent fashion and regulate
172
-
responsiveness.
EBNA2 interaction with TP1 promoter
a heterologous promoter (Sung et al., 1991; Tsang et al., 1991; Wang et al., 1991; Woisetschlaeger et al., 1991). We performed EMSAs to identify specific protein -DNA interactions within the TPJ promoter. One prominent DNA-protein complex (I) and two weaker complexes (II and III) were detected with a radioactively labelled fragment spanning the TPI promoter from position -12 to -258 (5'-end of the EBNA2RE). It was possible to compete specifically complexes I and III with two oligonucleotides spanning the region -262 to -230 and -239 to -209. When extracts were mixed with a radioactive oligonucleotide from position -262 to -209 and EMSAs were performed, the same pattern of complex formation occurred. These experiments indicate that protein(s) interact with sequences located in the EBNA2RE. Formation of complexes I, II and III appeared to be independent of EBNA2 and even independent of the cellular background of B-cells, since they were detected in all cell lines studied. This implies binding of a universal transcription factor. We were not able to identify this factor by comparing the sequence with known consensus sequences of transcription factors (Ghosh, 1990). The only obvious cognate nuclear factor binding sequence in the 54 bp fragment is one SPI consensus sequence at position -240. Addition of a 500-fold excess of an unlabelled oligonucleotide representing the SPI binding site failed to compete the shift (data not shown). It is therefore unlikely that this SPI site plays an essential role in the observed DNA -protein interaction. Formation of all complexes except complex II was competed by an excess of unlabelled 01 as well as 02. These two oligonucleotides share a 10 bp overlapping region and a common motif of 11 bp. The 11 bp motif seems to play an essential role in DNA -protein interactions, since an oligonucleotide containing the complete 10 bp overlap, but two truncated 11 bp motifs, has lost the ability to inhibit the protein binding of the 54 bp oligonucleotide. Realizing the importance of the 11 bp motif we searched for this sequence in the EBNA2 responsive regions of the BamHI-C and LMP promoters which were shown to be activated by EBNA2. Sung et al. (1991) reported an EBNA2 responsive region of 98 bp within the BamHI-C promoter between positions -367 and -465. In the LMP promoter, Tsang et al. (1991) described an element of 30 bp between positions -234 and -205 which is essential although not sufficient for EBNA2 responsiveness. Remarkably, the BamHI-C promoter contains, at position -371, a 10 bp sequence that is identical to the 11 bp motif, and the LMP promoter contains, at position -218, a sequence that shares 7 bp (GTGGGAA) of the 11 bp motif. Recently it was shown by footprint analyses that parts of the 10 bp motif of the BamHI-C promoter interact with a DNA binding protein (Jin and Speck, 1992). Taking this into account, it seems likely that the 11 bp motif plays an essential role in EBNA2 mediated transactivation. Is EBNA2 a component of the DNA -protein complexes? As mentioned above, EBNA2 positive and negative cell extracts showed almost identical binding patterns in EMSA. Only a very slowly migrating band (IVA or IVB) could be detected in Raji, M-ABA and Jijoye extracts, which was missing in all EBNA2 negative cell extracts. We mixed EBNA2 specific monoclonal antibodies in the EMSA to
study whether one of the retarded bands contained EBNA2. Addition of EBNA2 antibodies resulted in a supershift of complexes IVA and IVB. An antibody recognizing only EBNA2A induced a supershift of complex IVA only, whereas an antibody specific for EBNA2A and EBNA2B retarded migration of both complexes IVA and IVB. This indicates that EBNA2B is a component of complex IVB and EBNA2A a component of complex IVA. To prove that EBNA2 is indeed a component of complex IV, we added in vitro translated EBNA2 to EBNA2 negative BL41-P3HR1 extracts and performed EMSAs. The in vitro translated EBNA2 induced the formation of a more slowly migrating complex. By adding monoclonal antibodies we showed that the EBNA2-induced complex contained the added EBNA2. The complex containing the in vitro translated EBNA2 migrated slightly faster than complex IVa visible in M-ABA extracts. This phenomenon could be caused by different modifications of the in vivo and in vitro synthesized EBNA2. Although the 54 bp fragment is sufficient to bind EBNA2, it is not able to mediate EBNA2 responsiveness (data not shown), suggesting that sequences between -209 and -177 are necessary for EBNA2 dependent transactivation. It is conceivable that an additional protein binds to these sequences and contributes to EBNA2 responsiveness. Here we have shown for the first time that EBNA2 is present in complexes of specific DNA sequences and protein(s). Our experiments suggest that EBNA2 interacts with DNA indirectly via protein -protein interactions. Such interactions have been shown to be responsible for transactivation by EIA and VP16 (Lillie and Green, 1989; Stern et al., 1989; Liu and Green, 1990) transactivators of adenovirus and herpes simplex virus, respectively. VP16 binds to the universal transcription factor Oct 1, juxtaposing the transactivation domain of VP16 to the promoter. It seems likely that EBNA2 also interacts with a transcription factor, perhaps contained in complexI and/or III. We are in the process of identifying the factors interacting with the EBNA2RE of the TPI promoter, which should further elucidate EBNA2 mediated transactivation.
Materials and methods Cell lines and culture conditions Cell lines M-ABA, Raji and BL41 have been described elsewhere (Crawford et al., 1979; Pulvertaft, 1965; Lenoir et al., 1985). HH514 is a single cell clone of Jijoye, carrying a deletion in the EBNA2 gene (Hinuma et al., 1967; Miller et al., 1974). BL41-P3HR1 and BL41-B95-8 were obtained after infection of the EBV negative Burkitt's lymphoma cell line BL41 with virus strains P3HR1 and B95-8, respectively (Calender et al., 1987). HL60 is a promyelocyte cell line described by Collins et al. (1978). All cell lines were grown in RPMI 1640 medium supplemented with 10% fetal calf serum, 100 U/mI penicillin and 100 [tg/mn streptomycin. Cultures were incubated at 37°C in an atmosphere of 5% CO2. Cells were diluted 1:3 with fresh medium twice a week. Plasmids
The TPJ promoter-CAT construct described previously (Zimber-Strobl et al., 1991) was cut with either SacH (-45), SnaI (-124), Pvull (-177), NruI (-258), HindU (-432) or Nsil (-543) and protruding ends were filled in with T4 polymerase. After digestion with PstI the fragments containing the TPI promoter region were isolated and inserted into the filled in HindIII and PstI sites of the vector pBLCAT5 (P.Herrlich, personal communication). pTPl-CAT/A - 147/-258 was constructed by BclI and NruI digestion of pTPI-CAT/-543. pTPI-CAT/135bp and pTPI-CAT/81bp were generated by SmaI +NruI and PvuII +NruI digestion of pTP1-CAT/-543. The 135 bp and 81 bp fragments were isolated and cloned into the StuI
173
U.Zimber-Strobi
et al.
site of pTP1-CAT/-45. The 135 bp fragment was also cloned into the filled ,B-globin minimal promoter in in SalI site of (-glob-CAT containing the front of the CAT reporter gene (Laux et al., in preparation). All LUC constructs were obtained by digesting pBLLUC5 (Laux et al., in preparation) with BamHI and KpnI. The fragment containing the luc gene was isolated and ligated to the corresponding pTP1-CAT construct also digested with BamHI and KpnI. The EBNA2 expression vector pU294-6 contains an EBV fragment from position 35 448 to 54 360 according to the B95-8 EBV sequence (Baer et al., 1984). The clone was generated by deleting the NotI repeats of pM780-28 described by Polack et al. (1984). pU812/9 was generated by digesting pU294-6 with SphI and religation. p604 was kindly provided by W.Hammerschmidt. It was constructed by inserting an XbaI linker in p135.15 (Hammerschmidt and Sudgen, 1989) at position 48 950 according to the B95-8 EBV sequence published by Baer et al. (1984). The CMV-LTRCAT construct was obtained from I.S.Y.Chen (Cann et al., 1988). The plasmid used in the in vitro translation of EBNA2 was constructed by inserting the genomic B95-8 EBV sequence from position 48 039 to 54 360 in a Bluescript vector digested with HindIII and BamHI. Just in front of the translation start of EBNA2, an EcoRI site was inserted by site directed mutagenesis (pBSmutC). pBSmutC was cut with HindIII and EcoRI and an oligonucleotide representing a ribosome binding site was inserted
(pGa206/6).
Oligonucleotides The positions of oligonucleotides used in relation to the EBV genomic sequence according to Baer et al. (1984) are as follows: 01, 166 236-166 268; 02, 166 259-166 289; 03, 166 279-166 317; 04, 166 302-166 340; 054bp, 166 236-166 289; and I Imut, 166 252-166 272.
Transfection of the cells Electroporation of cells was carried out by the method of Cann et al. (1988). Briefly, 107 cells were washed once and resuspended in 0.25 ml fresh icecold RPMI 1640 supplemented with 10% fetal calf serum. The cells were placed on ice in a Gene Pulser cuvette, and 20 /tg of the corresponding DNA was added. Cells were electroporated with a Bio-Rad Gene Pulser at 250 V and 960 ItF. At 10min after electroporation, cells were resuspended in 10 ml of prewarmed RPMI with 20% fetal calf serum. CAT assay CAT assays were carried out as described previously (Zimber-Strobl et al., 1991). CAT activities were quantified by densitometrically scanning the signals of an autoradiogram with the Cybertech system (Appligene). LUC assays Cells were harvested 40 h after transfection, washed once in ice-cold PBS and resuspended in 100 IA 91 mM K2HPO4, 9 mM KH2PO4, 1 mM DTT, 1% Triton X-100, pH 7.8. Debris was removed by centrifugation at 14 000 g for 10 min. 10y1 of the supernatant was mixed with 350 ,u 25 mM glycylglycine, pH 7.8; 5 mM ATP; 15 mM MgSO4 and 100 j1 1 mM luciferin; 0.5 M Tris-HCI, pH 7.8. The bioluminescence (in RLU) was measured with a Lumat LB9501 (Berthold, Wildbach). Nuclear extract preparation Nuclear extracts were prepared by a modification of the method of Dignam et al. (1983). The pellet (3-5 x 107 cells) was washed once in ice-cold PBS, resuspended with 3-4 vol 10 mM HEPES pH 7.9; 10 mM KCI, 1.5 mM MgCl2; 5 mM DTT; 0.5 mM PMSF (buffer A) and incubated on ice for 1 h. Lysis of cells was achieved by 10-20 strokes with a Dounce homogenizer, microscopically controlled in the presence of trypan blue. Nuclei were pelleted for 10 s at maximal speed in an Eppendorf minifuge, washed once in buffer A, resuspended in 3 vol of buffer B (20 mM HEPES pH 7.9; 20% glycerine; 420 mM NaCl; 1.5 mM MgCl2; 0.2 mM EDTA; 5 mM DTT; 0.5 mM PMSF) and incubated on ice for 30 min. Nuclei were removed by centrifugation at 14 000 g for 20 min. The supematant was stored in liquid nitrogen.
Gel-shift analysis Binding reactions were performed in a volume of 20 td, containing 5 /O binding buffer (41 mM HEPES pH 7.9; 200 mM KCI; 4 mM EDTA; 1.6% Ficoll; 4 mM DTT; 0.5 mM PMSF), 2 yl poly(dIdC) (2 mg/ml), 2 Id BSA (20 mg/ml), 5 jLg protein extract and 0.2-1 ng radioactively labelled DNA. After incubation at room temperature for 30 min, the reaction products were separated in a 5% polyacrylamide gel. In competition and supershift experiments either unlabelled oligonucleotide or 1 a1 tissue culture supemratant containing monoclonal antibody was added to the reaction mixture. The following were used as anti-EBNA2 monoclonal antibodies: RI (rat IgGI),
174
R2 (rat IgG2a), R3 (rat IgG2a) and RI 1 (rat IgG2a). RI and R3 recognize EBNA2A with thesame affinity as EBNA2B, whereas R2 and RI 1 recognize EBNA2A, but have a very low affinity (R2) or no affinity (RI 1) to EBNA2B. Anti-mouse CD45 (rat IgG 1 or rat IgG2a) were used as isotypic control antibodies. In vitro translation pGa206/6 was digested with NaeI. The 4.5 kb fragment containing the EBNA2 ORF and regulatory sequences was prepared and transcribed using T3 RNA polymerase and a Boehringer in vitro transcription kit. The resultant RNA was translated in vitro using a Promega reticulocyte lysate translation kit. The translation products were assayed by SDS -PAGE and Western
blotting. Radioactively labelled probes pTPl-CAT/-543 was digested with NruI and Hindmi. The isolated fragment was labelled with Klenow polymerase in the presence of [32P]dCTP. The 054bp was synthesized with 5'-protruding ends, which were filled in with Klenow polymerase in the presence of [32P]dCTP.
Acknowledgements We thank W.Hammerschmidt for providing the p604 plasmid and for critical reading of the manuscript and B.Kempkes for providing the plasmid pBSmutC. This work was supported by Die Deutsche Forschungsgemeinschaft (Forschergruppe Virus-Zellwechselwirkung) and Fonds der Chemischen Industrie.
References
Gordon,J., Wang,F., Rymo,L. and Rickinson,A.B. (1990) J. Virol., 64, 2126-2134. Adldinger,H.K., Delius,H., Freese,U.K., Clarke,J. and Bornkamm,G.W. (1985) Virology, 141, 221-234. Baer,R. et al. Nature, 310, 207-211. Bornkamm,G.W., Hudewentz,J., Freese,U.K. and Zimber,U. (1982) J. Virol., 43, 952-968. Calender,A., Billaud,M., Aubry,J.P., Banchereau,J., Vuillaume,M. and Lenoir, G.M. (1987) Proc. Natl. Acad. Sci. USA, 84, 8060-8064. Cann,A.J., Chen,I.S.Y. (1988) Oncogene, 3, 123-128. Cohen,J. and Kieff,E. (1991) J. Virol., 65, 5880-5885. Cohen,J.I., Wang,F., Mannick,J. and Kieff, E. (1989) Proc. Natl. Acad. Sci. USA, 86, 9558-9562. Collins,S.J., Ruscetti,F.W., Gallagher,R.E. and Gallo,R.C. (1978) Proc. Natl. Acad. Sci. USA, 75, 2458-2462. Cordier,M. (1990) J. Virol., 64, 1002-1013. Crawford,D.H., Epstein M.A., Bomkamm,G.W., Achong,B.G., Finerty,S. and Thompson, J. (1979) Int. J. Cancer, 24, 294-302. (1984) Proc. Dambaugh,T., Hennessy,K., Chamnankit,L. and Acad. Sci. USA, 81, 7632-7636. Dignam,J.D., Lebovitz,R.M. and Roeder,R.G. (1983) Nucleic Acids Res., 11, 1475-1489. Fahraeus,R.. Jansson,A., Ricksten,A., Sjoblom,A. and Rymo,L. (1990) Proc. Natl. Acad. Sci. USA, 87, 7390-7394. Ghosh,D. (1990) Nucleic Acids Res., 18, 1749-1756. Ghosh,D and Kieff,E. (1990) J. Virol., 64, 1855-1858. and Muller-Lantzsch,N. (1991) J. Virol., Grasser,F.A., Haiss,P., 65, 3779-3788. Abbot,S.D., Rowe,M., Cadwallader,K., Ricksten,A.,
(1984)
Koyanagi,J.,
Kieff,E.
Natl.
Gottel,S.
Hammerschmidt,W. and Sugden,B. (1989) Nature, 340, 393-397. Hinuma,Y., Konn,M., Yamaguchi,J., Wudarski,D.J., Blakeslee,J.R.J. and
Grace,J.T.J. (1967) J. Virol., 1, 1045-1051. Jeang,K.T. and Hayward,S.D. (1983) J. Virol., 48, 135-148. Jin,X.W. and Speck,S.H. (1992) J. Virol., 66, 2846-2852. Kieff,E. and Liebowitz,D. (1990) In Fields,B. and Knipe,D. (eds), Virology. Raven Press, New York, pp. 1889-1920. Knutson,J.C. (1990) J. Virol., 64, 2530-2536. Laux,G., Perricaudet,M. and Farrell,P.J. (1988) EMBO J., 7, 769-774. Lenoir,G.M., Vuillaume,M. and Bonnardel,C. (1985) IARC Sci.
60, 309-318. Lillie,J.W. and Green,M.R.
Publ.,
(1989) Nature, 338, 39-44. Lindahl,T., Adams,A., Bjursell,G., Bornkamm,G.W., Kaschka-Dierich,C.
and Jehn,U. (1976) J. Mol. Biol., 102, 511-530. Liu,F. and Green,M. (1990) Cell, 61, 1217-1224. Miller,G., Robinson,J., Heston, L. and Lipman,M. (1974) Proc. Nat!. Acad. Sci USA, 71, 4006-4010.
EBNA2 interaction with TP1 promoter Polack,A., Hartl,G., Zimber,U., Freese,U.K., Laux,G., Takaki,K., Hohn,B., Gissmann,L. and Bomkamm,G.W. (1984) Gene, 27, 279-288. Pulvertaft,R.J.V. (1964) Lancet, 1, 238-240. Rickinson,A.B., Young,L.S. and Rowe,M. (1987) J. Virol., 61, 1310-1317. Sample,J., Liebowitz,D. and Kieff,E. (1989) J. Virol., 63, 933-937. Stern,S., Tanaka,M. and Herr,W. (1989) Nature, 341, 624-630. Sung,N.S., Kenney,S., Gutsch,D. and Pagano,J.S. (1991) J. Virol., 65, 6765 -6771. Tsang,S.F., Wang,F., Kenneth,M.I. and Kieff, E. (1991) J. Virol., 65, 6765-6771. Wang,F., Gregory,C.D., Rowe,M., Rickinson,A.B., Wang,D., Birkenbach,M., Kikutani,H., Kishimoto,T. and Kieff,E. (1987) Proc. Natl. Acad. Sci. USA, 84, 3452-3456. Wang,F., Hitoshi,K., Tsang,S.F., Kishimoto,T. and Kieff,E. (1991) J. Virol., 65, 4101-4106. Westin,G., Gerster,T., Muller,M.M., Schaffner,G. and Schaffner,W. (1987) Nucleic Acids Res., 15, 6787-6789. Woisetschlaeger,M., Jin,X.W., Yandava,C.N., Furmanski,L.A., Strominger,J.L. and Speck,S.H (1991) Proc. Natl. Acad. Sci. USA, 88, 3942 -3946. Zimber,U. et al. (1986) Virology, 154, 56-66. Zimber-Strobl,U., Suentzenich,K.O., Laux,G., Eick,D., Cordier,M., Calender,A., Billaud,M., Lenoir,G.M. and Bornkamm,G.W. (1991) J. Virol., 65, 415-423. Received on July 20, 1992; revised on September 10, 1992
175
