Download Repression by RAZ of Epstein-Barr virus bZIP transcription factor

Repression by RAZ of Epstein-Barr virus bZIP transcription
factor EB1 is dimerization independent
Carine S e g o u f f i n , H e n r i G r u f f a t a n d Alain Sergeant
Unite de Virologie Hurnaine, ENS-INSERM, U412, Ecole Normale Sup4rieure de Lyon, 46 Allee d'Italie, 69364 Lyon Cedex 07, France
The hallmark of Epstein-Barr virus (EBV) infection is
the establishment of a viral genome transcription
pattern called latency. The EBV BZLF 1 gene product
EB1 (also known as ZEBRA or Zta) is a transcription
factor which is essential for the switch from latency
to the lytic cycle. It has been proposed that latency
is maintained (~ by the inhibition of EB1 translation via antisense hybridization of EBNA1 and
EB1 hnRNAs, or (iO by the inactivation of the EB1
activating function via the direct interaction of EB1
with RelA, the retinoic acid receptor and p53, or via
the titration of EBI in RAZ:EBI inactive heterodimers that are unable to bind to DNA. RAZ, a fusion
protein which contains the EB1 C-terminal dimeri-
Introduction
Epstein-Barr virus (EBV) is a human herpesvirus that
persists latently for the lifetime of the infected host. EBV is also
associated with human malignancies such as Burkitt's lymphoma, nasopharyngeal carcinoma, Hodgkin's disease, and B and T
cell lymphomas in immunocompromised individuals. EBVinfected B lymphocytes from the peripheral blood or in vitroinfected B lymphocytes can proliferate continuously when
cultured (immortalization). In such cells, only a minority of the
viral genes are expressed, a state defined as type III latency
(Liebowitz & Kieff, 1993). Data are accumulating on the
mechanisms by which the productive cycle is reactivated in
such latently infected B cells, as well as the mechanisms by
which reactivation can be suppressed.
Various chemical agents, including the tumour promoter
12-O-tetradecanoylphorbol-13-acetate (TPA) (zur Hausen et
al., 1978), can provoke the switch from a latent to a productive
viral cycle in vitro. This switch is due to the expression of two
EBV-encoded transcription factors: the BZLFl-encoded factor
EBI (also called Z, Zta or ZEBRA) (Chevallier-Greco et al.,
1986; Countryman & Miller, 1985; Countryman et al., 1989;
Authorfor correspondence:Alain Sergeant.
Fax + 33 72 72 86 86. e-mail [email protected]
0001-3756 © 1996 SGM
zation and DNA-binding domains fused to the Nterminal 86 amino acids of the EBV BRLF1 gene
product R, has been described as a trans-dominant
negative regulator of EBl-activated transcription.
We demonstrate here that although RAZ efficiently
represses EBl-mediated transcriptional activation,
the amount of RAZ protein expressed is incompatible with repression through the titration of EB1
in inactive EB1 : RAZ heterodimers. Furthermore, we
also demonstrate that RAZ efficiently represses
transcription activated by an EB1 mutant carrying
the GCN4 homodimerization domain (EBI~"4), despite the inability of RAZ and EBlgCn4to form stable
heterodimers.
Takada et al., 1986) and the BRLFl-encoded factor, called R (or
Rta) (Hardwick et al., 1988) (Fig. la). Indeed, when EB1 and R
expression vectors are transfected into latently infected B cells,
both factors are able to activate all the EBV early genes
(Buisson et al., 1989; Chevallier-Greco et al., 1986; Countryman
& Miller, 1985; Flemington & Speck, 1990b; Rooney et al.,
1989), probably by binding as homodimers to specific DNAbinding sites located in early EBV promoters (Chang et al.,
1990; Farrell et al., 1989; Flemington & Speck, 1990a; Gruffat
et al., 1990; Kouzarides et al., 1991). Moreover, only EB1
induces a productive cycle, since EB1, but not R, transactivates
DNA replication from OriLyt, the origin of replication active
during the lyric cycle (Cho & Tran, 1993; Fixman et al., 1995 ;
Rooney et al., 1989, Schepers et al., 1993).
Suppression of reactivation has been proposed to occur via
antisense hybridization of EBNA1 and EB1 hnRNAs (Prang et
al., 1995). It has also been proposed to occur through repression
of EBl-mediated transcriptional activation, via direct interaction of EB1 with the NF-;B p65 subunit (Gutsch et al., 1994),
the retinoic acid receptor (Sista et al., 1993, 1995) or p53
(Zhang et al., 1994). A trans-dominant repressor of EB1 called
RAZ has also been described (Furnari et al., I994; Kelleher et
al., 1995). RAZ is translated in vitro from a transcript
characterized originally as a cDNA called Z8 (Manet et al.,
1989), initiated at promoter PR and generated by facultative
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
52!
iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii!i iiiii iiii iiii ii i i iii i iiiiiiii
(a)
PROTEIN
TRANSLATED
R (EBI?)
relative amounts of RAZ and EB1 proteins expressed are
incompatible with a repression through the titration of EB1 in
inactive RAZ:EB1 heterodimers. Furthermore, RAZ efficiently
repressed transcription activated by an EB1 mutant carrying
the GCN4 homodimerization domain (EBlgen4), despite the
inability of RAZ and EB1 gen4 to form stable heterodimers.
cDNAs
AUG
,m~
AUG
Z13
~ ~ z t s
AUG
R (EBI?)
RAZ
BZLF1
I IH l
BRLFt
i:il Pz V////////////////,."]
]
Methods
PR
• Plasmids
Expressionvectors. The cytomegalovirus (CMV)-based expression vec-
AUG
EOl
(b)
N
Activation
1
8s
D8
Oi
170 195
C
221
1
.
~'
/¢
DNA binding (DB)
Dlmericatlon
EB1
245
RAZ
Prollne rich
Acidic rich
C
I
Niiiiiiiilllll]lllllllllllllllllm
35o
Actlv|tlon
sos
R
(01)
Fig, 1. Schematic representation of the EBV genome region containing the
BZLF1 and BRLF1 genes. (e) Structure of full-length cDNAs
corresponding to the mRNAs originating from the BomHI Z/R region, Intron
sequences are indicated by thin lines. The cDNAs are listed on the right
and the proteins translated from the mRNAs are listed on the left. (b)
Functional domains of the EB1, R and RAZ proteins with activation, DNA
binding (DB) and dimerization (Di) domains indicated,
splicing (Fig. la). RAZ is a fusion protein containing the EB1
C-terminal dimerization and DNA-binding domains fused to
the N-terminal 86 amino acids of R (Fig. I b). In vitro-translated
RAZ does not bind stably to D N A and has been shown to
recruit EB1 into RAZ:EB1 heterodimers that are unable to bind
D N A (Furnari et aI., 1994). Accordingly, in co-transfection
assays done in HeLa cells, RAZ expressed from plasmid
pCMV-RAZ represses transcription activated by EB1 expressed from the plasmid pCMV-Z (Furnari et al., 1994).
Additionally, in co-transfection assays done in B cells latently
infected with EBV, RAZ expressed from plasmid pCMV-RAZ
impairs the induction of EA-D (early antigen diffuse) by EB1
and R translated from bicistronic mRNAs initiated at the CMV
promoter in plasmid pEBV-R/Z (Furnari et al., 1994). However,
the RAZ protein has never been visualized in the repression
assays, nor in EBV-infected cells. More importantly, the relative
amounts of EBI and RAZ proteins were not evaluated in the
repression assays, and repression through heterodimerization
was not assayed directly. We have therefore evaluated the
relative amounts of RAZ and EBI proteins expressed in our
repression assays and have investigated whether RAZ:EB1
heterodimerization is essential for RAZ-repressed EB1transcriptional activation.
Our results clearly demonstrate that although RAZ efficiently represses EBI-mediated transcriptional activation, the
53C
tors for EB1 (pCMV-EB1), EB1go"4 (pCMV-EB1~en4) and RAZ (pCMVZS) contain the EB1, EB1gcn4and Z8 cDNAs cloned into the EcoRI site
of the pRc-CMV vector (Invitrogen). Plasmid pRc-CMV contains the
cytomegalovirus immediate-early (CMV) promoter, the SV40 early polyadenylation signal and the T7 promoter. The SV40-based expression
vectors for EBI (pSV-Z41), for RAZ (pSV-Z8) and for the EB1 deletion
mutant called Z59-140 (pSV-Z59-140) have been described elsewhere
(Manet et al., I989; Mikaelian ef al., 1993a). The EB1 and Z8 cDNAs
have been described elsewhere (Manet et al., 1989). The EB1g°n4 cDNA
was excised from plasmid pAAC-EB1Cen4(Giot et al., 1991; Mikaelian
et al., 1993a). The FLAG-RAZ expression plasmid pCMV-FLAG-Z8
was generated by PCR. The 5' primer contained sequences coding both
for the FLAG peptide (IBI Flag system, Kodak) and for part of the N
terminus of RAZ: 5' GCCCCGGATCCCACCATGGACTACAAGGA
CGACGATGACAAGCATATGAGGCCTAAAAAGGATGGC3'.
The 3' primer, 5' GCAAGCTTCGGTAGTGCTGCAGCAG 3', was complementary to the Z8 cDNA sequences at 622 bp from the ATG. The
PCR amplified product was digested by BamHI and XmaI, and cloned
into plasmid pRc-CMV to generate pCMV-FLAG-ZS.
Reporter genes. The pZ-CAT reporter gene (see Fig. 2a) has been described elsewhere (Urier et al., 1989). Plasmid pCMV-CAT was made
by inserting the CAT gene into the HindIII site of plasmid pCEP4
(Invitrogen).
• Transfections. Plasmids were prepared by the alkaline lysis method
and purified through two sequential caesium chloride gradients. The
DNAs were in the same topological state as assayed by agarose gel
electrophoresis. HeLa cells were grown in DMEM (Gibco) supplemented
with 10 % (v/v) fetal calf serum and were seeded at I x I06 cells per
100 mm Petri dish 8 h prior to transfection. Transfections were performed
by the calcium phosphate precipitation method. Cells were mixed with
the appropriate DNAs : typically 15 ,g of DNA was used which included
the expressing vector and plasmids carrying the reporter genes.
Transfected cells were washed and collected 48 h after transfection.
• CAT assays. These were performed essentially as described
previously (Gorman el' al., 1982). Sonication of the cells, however, was
replaced by lysis in a buffer containing 0"25 M-Tris-HC1pH 8 and 0"05%
SDS. Acetylation of chloramphenicol was quantified by thin-layer
chromatography followed by scintillation counting.
• Immunoblots. Cells were washed with cold PBS and incubated with
100 ,l of lysis buffer as already described (Mikaelian et al., 1993 a). Equal
amounts of protein were loaded and separated on a 10 % polyacrylamideSDS gel, and transferred to a nitrocellulose membrane (ScbIeicher and
Schuell) by electroblotting. The membrane was then incubated with the
primary antibody. The monoclonal antibody mAbZ125 detects both EB1
and EB1g°n4 proteins (see Figs 2a and 6a). The monoclonal antibody
mAbZ130 detects EB1, Z59-140 and RAZ (see Fig. 3a). The monoclonal
antibody mAbM2 (IBIFlag system, Kodak) detecfs FLAG-RAZ (see Figs
2 a and 3 a). The rabbit polyclonal antibody to bacterially synthesized EB1
leucine zipper, AbLZebl, detects EB1 and RAZ and does not detect the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
EB1gcn4 protein (see Fig. 3a, AbLZebl). The membranes were then
incubated with horseradish peroxidase-conjugatedgoat anti-mouse (for
mAbZ125 and mAbM2 antibodies) or goat anti-rabbit (for AbLZebl)
immunoglobulins (Amersham).The proteins were visualizedwith an ECL
kit (Amersham) as specifiedby the supplier. When aSS-labelledproteins
were to be visualized by ECL, a sheet of transparent plastic paper was
placed between the X-ray film and the membrane in order to selectively
absorb radioactiveemission from asS.
• Synthesis of in vitro-translated proteins and EMSA. EBI,
EB1g°n4, RAZ and Z59-93 (EB1 deleted from aa 59 to 93; Giot el a].,
I991) proteins were synthesizedalone or in combinationin a reticulocyte
[ysate-coupled transcription-translation system (Promega). 14C-or a~Slabelled proteins were visualizedby SDS-PAGE and autoradiographyof
the dried gel. Electrophoreticmobility shift assays (EMSA) were performed by incubating 4 x I04 c.p.m, of 5' a2p-labelleddouble-stranded
DNA probes with 2,1 of in vitro-translated proteins for 30 rain at room
temperature in 20 mM-HEPES (pH 7"9), 100 mM-KC1, 1 mM-MgCI2,
0"5 mM-DTT,10% glycerol and 1 ~g of poly(dI-dC) in a final volume of
20,1. After incubation, the mixture was loaded onto a 4"5% (w/v)
polyacrylamide gel (29 to I cross-linked), 0"2 x TBE, and run at room
temperature at 10 V/cm for 3 h. The protein-DNA complexes were
visualized by autoradiography.
Results
Efficient repression by RAZ of EBl-mediated
transcriptional activation cannot be explained by the
formation of RAZ: EB1 heterodimers
RAZ and EB1 cannot be separated in size by SDS-PAGE
(not shown). Thus, in order to specifically differentiate the
RAZ protein from the EB1 protein in transient repression
assays, the coding sequence for a peptide (FLAG) recognized
by monoclonal antibody mAbM2 was inserted 5' to the Z8
cDNA in plasmid pCMV-Z8 to generate plasmid pCMVFLAG-Z8 (Fig. 2a). The FLAG-RAZ protein should be
recognized specifically by antibody mAbM2, whereas the EBI
protein should be recognized specifically by monoclonal
antibody mAbZ125 directed against an epitope located
between amino acids 59 and 86 of EB1, but which is absent in
the RAZ protein (Fig. 2a). This was found to be the case (Fig.
2c).
We then evaluated the relative amounts of EB1 and RAZ
protein expressed in a transient repression assay. The BZLF1
gene promoter PZ carrying two EBl-binding sites and linked
to the CAT gene (pZ-CAT, Fig. 2a) was co-transfected into
HeLa cells together with the EBI expression vector pCMVEB1. As expected, transcription of the CAT gene from plasmid
pZ-CAT was activated by EB1 (Fig. 2b, lane 2), and this
activation was strongly repressed by the addition of increasing
amounts of plasmid pCMV-FLAG-Z8 to the DNA used for
transfection (Fig. 2 b, lanes 3 to 5). Surprisingly, the decrease in
the EBI-mediated transactivation of promoter PZ (Fig. 2b,
lanes 3 to 5) correlated with a decrease in the amount of EBI
protein expressed from plasmid pCMV-EBI (Fig. 2c,
mAbZI25, lanes 3 to 5). More importantly, although the
amount of FLAG-RAZ protein increased (Fig. 2 c, mAbM2,
lanes 3 to 5) when plasmid pCMV-FLAG-Z8 was transfected in
increasing amounts, the amount of FLAG-RAZ protein was far
less than the amount of EB1 protein, even in the transfections
where the amount of EB1 was reduced (Fig. 2c, compare
mAbZ125 with mAbM2, lane 5). It should be stressed that 5,1
of total cell protein extract was used to visualize the amount of
EB1 protein present in the transfected cells whereas 20 I11of the
same extract was used to visualize the FLAG-RAZ protein.
Since using different antibodies to compare the amounts of
EB1 and RAZ is not state of the art, we had to show that at the
dilution used, antibodies mAbM2 and mAbZ125 detect EBI
and RAZ with comparable efficiencies. To do this, EB1 and
RAZ were in vitro-translated and labelled with [aSS]methionine
in rabbit reticulocyte lysates (RRL), size-separated by
SDS-PAGE, transferred to nitrocellulose, and visualized by
autoradiography. As shown on the autoradiogram in Fig. 3 (b),
equivalent amounts of RAZ (lanes I to 3) and EBI (lanes 6 to
7) proteins were present on the membrane (EB1 contains three
methionines and RAZ four). As shown on the immunoblot in
Fig. 3 (b), the antibodies mAbM2 (dil.: 1/500 of the original
solution) and mAbZI25 (dil.: I / I 0 0 of the original solution)
detected the in vitro-translated RAZ (lanes 1 to 3) and EB1
proteins (lanes 6 to 8) with comparable efficiencies. By using
the two antibodies at the dilutions indicated above, we
observed that when 150 ng of pCMV-EB1 was co-transfected
in HeLa cells with 300 ng of pCMV-FLAG-Z8, more EBI
protein (Fig. 3 b, lane 5) than RAZ protein (Fig. 3 b, lane 4) was
detected. It must be noted that under the conditions of EB1 and
RAZ transfection described above, EB1 activation was strongly
repressed by RAZ (Fig. 2 b, lane 5).
We also evaluated the relative levels of RAZ and EB1
proteins in HeLa cells by using as an internal reference a shorter
EB1 mutant called Z59-I40, and the antibodies mAbZ130 and
AbLZ ebl directed against epitopes present in EB1, RAZ and
Z59-140 (Fig. 3 a). Firstly, and as shown in Fig. 3 (c), comparable
amounts of RAZ (autoradiogram, lanes 1 to 3) and EBI
(autoradiogram, lanes 6 to 8) proteins translated in vitro and
labelled with [14C]leucine were separated by SDS-PAGE and
transferred to nitrocellulose. The antibodies mAbZ130
(immunoblot mAbZI30) and AbLZ e61 (immunoblot AbLZ ebl)
efficiently detected the in vitro-translated RAZ (lanes 1 to 3)
and EB1 (lanes 6 to 8) proteins. When HeLa cells were cotransfected either with i ~g of pSV~EB1 and I ~g of pSV-Z59140, or with 1 lag of pSV-FLAG-Z8 and I ~tg of pSV-Z59-140,
both mAbZI30 and AbLZ ebl antibodies clearly detected Z59140 expressed at comparable levels in both co-transfections.
However, much less RAZ protein than EB1 protein was
detected in transfected cells suggesting that RAZ was either
less stable than EB1 and/or that RAZ was less efficiently
translated than EB1. Indeed, equal amounts of cytoplasmic
RAZ and EB1 poly(A) + RNAs were found in HeLa cells
transfected as described above (not shown), confirming that
the difference in the amounts of RAZ and EB1 proteins
expressed in HeLa cells was also not transcriptional.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
53
iiiiiiiiiiiiiiiiiiiiiiiiiii
iii iii iiii !i!iiiiiiiii!i
(a)
(a)
-22
pZ-CAT
TATAm
-72
P-1
-86
ZREZ1
-225
mAbZ125
s,X
EB1
mAbZ125
A
N
EB1
1
/
140
170
195
221
Z59-140
¥
ss
X
I
40
t't~o~''' 195
221
245
1
140
1to lss
~r
2~
s'6
14o
,;'~,Sy
~.,
245
-"->-4
"
FLAG
FLAG-RAZ
111111
y
245
8'6
FLAG
rm
FLAG-RAZ
C
86
mAbZ130
ZREZ2
mAbM2
AbLzebl
mAbM2
(b)
(b)
Exlracls (MI)
FIAG-RAZ
EB1
[2,5%
O~
lanes
10.0%
in vitro
translated
HeLa cells
extract
~
+
1
in vitro
translated
+
+
1"-5~'5l ~
+
+
÷
+ +
+
+
2
3
4
7
8
5
6
Autoradiogram
~' ~' 75,%
~ ~
Immunoblot
<
- -
mAbM2--
--mAbZ12$--
5.o~
2,5%
(c)
0,0%
lanes
pCMV-EB1
(ng)
pCMV-FLAG-Z8
(rig)
I
I
I
I
I
1
/
2
3
4
5
~ 1 5 0
/
0
75 150
300
Extracts (,ul)
FLAG-RAZ
EB1
Z59-140
lanes
in vitro
translated
I ' - ' ~
+
+ +
1
2
3
HeLa cells
in vitro
extract
translated
1"3'-0-""'3'~ I " ~
+
+ +
+
+
+
+
4
5
6
7
8
Autoradiogram
(c)
RAZ -,~
1
2
3
4
5
Extract (HI)
Z59-140-~
mAbZ130
mAbZ125
5
mAbM2
20
mmunoblot
RAZ
Z59-140
AbLZ e b l
Fig. 2
Fig. 3
Fig. 2. RAZ repression of EBI-mediated transactivation cannot be explained by the formation of RAZ; EBI heterodimers. (a)
Schematic structure of the pZCAT reporter gene construct. Regions in the EBI and FLAG-RAZ proteins specifically
immunodetected by antibodies mAbZ125 and mAbM2 are indicated. (b) The reporter plasmid pZCAT was cotransfected in
HeLa cells alone (lane I ) or cotransfected with pCMV-EBI (lane 2) or cotransfected with pCNV-EBI and increasing amounts
of pCMV-FLAG-Z8 (lanes 3 to 5), as indicated. The amount of CNV promoter was kept constant by co-transfecting, where
necessary, with plasmid pCMV-O. CAT activity is expressed as percentage chloramphenicol acetylation. (c) Visualization of the
EBI (mAbZ125) and FLAG-RAZ (mAbM2) proteins expressed in HeLa cells transfected as described in (b), lanes I to 5. The
volume of total cell extract analysed is indicated on the right.
Fig. 3. The difference in the amount of EBI and RAZ proteins detected in HeLa cells is not due to a difference in the relative
detection efficiencies of the antibodies used. (o) Schematic representation of the proteins used in the immunoblotting
experiments and of the epitopes recognized by mAb7125, mAbZ130, mAbM2 and AbLZebl. (b) The FLAG-RAZ (lanes I to 3)
and EBI (lanes 6 to 8) proteins, translated in vitro and labelled with [3SS]methionine (autoradiogram), were detected by
antibodies mAbN2 and mAbZ125 with similar efficiencies (immunoblot). Visualization by mAbM2 and mAbZ125 of the FLAGRAZ (lane 4) and EBI (lane 5) proteins expressed in HeLa cells cotransfected with 150 ng of pCMV-EBI and 300 ng of
pCMV-FLAG-Z8 (immunoblot). The volume of extracts analysed and which protein they contained are indicated. (c) The FLAGRAZ (lanes I to 3) and EBI (lanes 6 to 8) proteins, translated in vitro and labelled with [14C]leucine (autoradiogram), were
efficiently detected by antibodies mAbZ130 and AbLZebl (immunoblot). Visualization by mAbZ130 and AbLZebl antibodies of
the FLAG-RAZ (lane 4) and EBI (lane 5) proteins expressed in HeLa cells cotransfected with I llg of pSV-Z8 and I pg of
pSV-Z59-140, or cotransfected with I l~g of pSV-Z41 and I pg of pSV-Z59-140 (immunoblot). The volume of extracts
analysed and which protein they contained are indicated.
i3;
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
(a)
~,0,0 % -
70,0%
-
60,0%
-
50,0%
-
40.0%
-
30.0%
-
,<
20,0%
-
] 0.0%
-
0 " 0 ¢'¢~
l
I
t
I
I
2
I
3
4
5
6
/
0.07
0,15
0,3
/
/
/
I
/
6
7
lanes
pCMV-CAT (~g)
I
7
0.25
pCMV-FLAG.Z8 (l~g)
pCMV-EB1 (pg)
0.6
/
0.6
0.15
0.15
(b)
1 2
m AbZ l 2 5
mAbM2
3
4
5
~;~!:i!i:,;:
Extract (~1)
5
20
Fig. 4. RAZ repression of EB1 -mediated transactivation of the CMV
promoter cannot be explained by the formation of RAZ: EB1 heterodimers.
(a) HeLa cells were transfected with pCMV-CAT alone (lane 1 ) or with
pCMV-CAT and increasing amounts of pCIvlV-FLAG-Z8 (lanes 2 to 5) or
with pCMV-EB1 (lane 6) or with both pCMV-EB1 and pCMV-FLAG-Z8
(lane 7), as indicated. CAT activity was expressed as percentage
chloramphenicol acetylation. (b) Visualization of the EB1 (mAbZ125) and
FLAG-RAZ (mAblVl2) proteins expressed in HeLa cells transfected as
indicated in (o), lanes 1 to 7. The volume of total cell extract analysed is
indicated.
The results presented above definitely exclude the possibility that the difference in the amount of EB1 and RAZ
proteins detected in HeLa cells was due to a difference in the
relative detection efficiencies of the antibodies used. They also
exclude the possibility of simple promoter interference since
the amounts of Z59-140 protein expressed were comparable in
each co-transfection. These results demonstrate that the
amount of FLAG-RAZ protein expressed is incompatible with
a repression of EBl-mediated transcriptional activation
through the titration of EB1 in inactive ILAZ:EB1 heterodimers.
Moreover, these results also suggest that the CMV promoter
driving the expression of mRNAs from which EB1 and FLAGRAZ were translated was activated by EB1 and repressed by
FLAG-RAZ.
RAZ repression of E B l - m e d i a t e d transactivation of t h e
C M V p r o m o t e r c a n n o t be e x p l a i n e d by t h e f o r m a t i o n
of RAZ: EB 1 h e t e r o d i m e r s
When HeLa cells were transfected with a plasmid carrying
the CMV promoter linked to the CAT gene (pCMV-CAT), the
CAT gene was detectably transcribed (Fig. 4a, lane 1). CAT
gene transcription was unchanged when plasmid pCMVFLAG-Z8 was added in increasing amounts to plasmid pCMVCAT in the transfections (Fig. 4a, lanes 2 to 5), and the FLAGRAZ protein was clearly detected by monoclonal antibody
mAbM2 (Fig. 4 b, mAbM2, lanes 3 to 5). However, the activity
of the CMV-CAT reporter gene was strongly increased by
EBI expressed from plasmid pCMV-EBI co-transfected into
HeLa cells with plasmid pCMV-CAT (Fig. 4a, lane 6), and EB1
protein was detected by the monoclonal antibody mAbZI25
(Fig. 4 b, lane 6). Moreover, EBI-activated CAT transcription
was strongly repressed when plasmid pCMV-FLAG-Z8 was
co-transfected with plasmids pCMV-CAT and pCMV-EB1
(Fig. 4a, lane 7). Again, the decrease in EBl-activated
transcription was associated with a decrease in the amount of
EB1 protein translated from mRNAs transcribed from plasmid
pCMV-EB1 (Fig. 4 b, mAbZ125, compare lane 6 with lane 7). It
is noteworthy that although equal amounts of plasmid pCMVFLAG-Z8 were added to the cells in the transfections depicted
in lanes 5 and 7, the amount of FLAG-RAZ protein detected
was only slightly higher in the presence of EB1 (Fig. 4b,
mAbM2, lane 7) than in the absence of EB1 (Fig. 4 b, mAbM2,
lane 5). This indicates that RAZ represses the EBl-mediated
expression of both EB1 and FLAG-RAZ from plasmids pCMVEB1 and pCMV-FLAG-Z8 respectively. But when the amount
of EB1 protein is compared to the amount of FLAG-RAZ
protein expressed, this is again not compatible with a
repressive effect of RAZ through the titration of EB1 into
inactive RAZ:EB1 heterodimers. These results clearly demonstrate that transcription initiated at the CMV promoter is
responsive to EB1, and that this EBI-mediated transcriptional
activation is repressed by FLAG-RAZ by a mechanism which
cannot be explained by invoking the formation of RAZ:EB1
heterodimers.
The f o r m a t i o n of R A Z : E B 1 h e t e r o d i m e r s is not
r e q u i r e d for RAZ to repress E B l - a c t i v a t e d
transcription
If the repressing effect of RAZ described above cannot be
explained by the titration of EB1 into inactive RAZ:EB1
heterodimers, then repression should be seen if EBI and RAZ
cannot form heterodimers. To directly evaluate this possibility,
we constructed a hybrid protein, EB1een4, in which the
homodimerization domain of EB1 is replaced by the homodimerization domain of the yeast transcription factor GCN4
(Fig. 5a). In order to demonstrate that the GCN4 and EBI
dimerization domains do not stably heterodimerize, we
performed a mobility shift assay using a 32P-labelled doublestranded oligonucleotide containing an EBl-binding site (Fig.
5a). EBI gen4 (Fig. 5b, lanes I to 3) or EB1 (Fig. 5b, lanes 5 to
7) were co-translated with increasing amounts of the EB1
deletion mutant called Z59-93 (Fig. 5 a). The formation of the
heterodimers EB1gen4 : Z5 9-93 or EBI : Z59-93 and their binding to DNA was evaluated by incubating the in vitro-translated
proteins with the 32P-labelled DNA probe described in Fig.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
;3:
(a)
(a)
~la
Ipc~bes S ' )
'
- ....
~
3'
°
S
E81
'
170
195
221
170
195
221
245
EB1 gcn4
S9
EB1
Y
ebl e b !
OB
DI
93
.....
140
mAbZ125
I
245
ZSg*g3
RAZ
I
D ~ ¢n4
EBlg en4
I
AbLZ
(b)
(b)
EB1 g©n4
+
+
+
4-
EB1
4-
+
=
Z59-93
lanes
i
KDa M
46
1
2
3
4
5
6
7
.<
EB1 gcn4 I~
3O
(c)
lanes
L 1
2
3
4
5
6
~ 8
7
plasmids (lag)
pZ-CAT
gcn4
gcn4
EB1
: EB1
Z59-93:Z59-93
~ _ _ . EB1 :EB1
I~
11
EB1:Z59-93
Z59-93:Z59-93
9 lo 11
s
pCMV-EB1
pCMV-EB1 g©n4
/
pCMV-Z8
/
o . l s ~
O.15
/
(c)
0.07 o.ls
7
0.3 o.s
8
9
I
10
0.07 O.lS 0.3 0.6
11
mAbZ12S
AbLZ ebl
Fig. 5
Fig. 6
Fig. 5. The GCN4 and EB1 leucine-zippers do not stably heterodimerize in vitro. (o) DNA sequence of the double-stranded
oligonucleotide used to evaluate specific binding of EB1, EB1 gcn4,Z59-93 and their dimeric combinations (the EB1 -binding site
is underlined), and schematic representation of the proteins used in the electrophoretic mobility shift assay (EMSA)
experiments. (b) [~4C]Leucine-labelled EB1, EB1 gcn4and Z59-93 proteins used in the EMSAs; lane M, molecular mass markers.
(c) EMSAs were performed with the proteins described in (b), lanes 1 to 7, or with the reticulocyte lysate (lane L). The
protein-DNA complexes are indicated on the left and right of (c).
Fig. 6. Repression of EBl-mediated activation by RAZ is dimerization-independent. (o) Schematic representation of the
proteins and epitopes recognized by mAbZ125 and AbI_Zebl. (b) The reporter plasmid pZCAT was cotransfected in HeLa cells
with the expression plasmids pCMV-EB1, pCMV-EBlgCn4 and pCMV-Z8 as indicated. CAT activity is expressed as percentage
chloramphenicol acetylation. (c) Visualization of the EBlgCn4 (mAbZ125) and RAZ (AbI_Zebl) proteins expressed in HeLa cells
transfected as in (b) lanes 7 to 11.
4 (a). As expected, the EB1 : EB1/DNA complexes (Fig. 5 c, lane
5) formed as efficiently as the EBlgen4:EBlgen4/DNA complexes (Fig. 5c, lane I) and could be separated from the Z5993 :Z59-93/DNA complexes (Fig. 5 c, lane 4). However, while
EB1 : Z59-93 :DNA complexes formed efficiently (Fig. 5 c, lanes
6 and 7), no EBIgen4:Z59-93/DNA complexes formed (Fig.
5 c, lanes 2 and 3). From these results we conclude that EB1 and
EBI gen4 dimers bind the DNA probe in vitro with the same
apparent efficiency and that RAZ cannot form heterodimers
with EBI gen4.We then considered whether RAZ would repress
53 z
EBIgen4-activated transcription. As expected, CAT transcription from plasmid pZ-CAT in HeLa cells was very inefficient
(Fig. 6 b, lane 1), but was strongly increased, to the same extent,
by EBI (Fig. 6b, lane 2) and by EBI gen4 (Fig. 6b, lane 7).
Moreover, strong repression of both EBl-activated transcription (Fig. 6b, lanes 3 to 6) and EBlgen4-activated
transcription (Fig. 6b, lanes 8 to 11) was observed when
plasmid pCMV-Z8 was added in increasing amounts to the
DNA used for transfection. Therefore, despite the inability of
EB1gen4 and RAZ to form stable heterodimers, RAZ repressed
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
EBlgen4-activated transcription. Again, the amount of EBI gen4
protein expressed from pJasmid pCMV-EB1 gen4 decreased
(Fig. 6c, mAbZ125, lanes 8 to 11) as increasing amounts of
RAZ proteins were expressed (Fig. 6c, AbLZ ebl, lanes 8 to 11).
Discussion
Transient expression assays have been used to show that
the EBV protein RAZ is a trans-dominant negative regulator of
EBl-activated transcription, and can be a potential regulator of
the switch from EBV latency to the lytic cycle (Fumari et al.,
1994). Furthermore, as RAZ does not bind stably to DNA in
vitro, Furnari et al. suggested that RAZ represses EBl-activated
transcription by titrating the EB1 protein in RAZ:EB1
heterodimers which are unable to bind to DNA. However,
the effect of RAZ on EBl-mediated transcriptional activation
can be considered to be repression by heterodimerization only
if the amount of RAZ protein is equal in each transfection to the
amount of EB1 protein. We show here that this is clearly not
the case in transient expression assays where the CMV-based
expression vectors used are similar to those used in the
repression assays published by Fumari eta].
Indeed, we show that in transient assays where repression
by RAZ of EBl-activated transcription is clearly seen, the
amount of RAZ protein expressed is much lower than that of
EBI, and therefore is not compatible with the titration of EB1
in inactive EBI:RAZ heterodimers. The difference in the
amounts of RAZ and EB1 proteins detected is not due to the
use of different antibodies (Fig. 3c). The difference in the
amount of RAZ and EB1 proteins detected is also not
transcriptional, since RAZ and EB1 RNAs were transcribed
with similar efficiency from the SV40 promoter (not shown).
However, although both RAZ and EB1 were located in the
nuclei, much less RAZ protein was found in transfected cells
suggesting that RAZ was either less stable than EB1 and/or
that RAZ was less efficiently translated than EB1. Upon cotransfection of pCMV-FLAG-Z8 and pCMV-EBI in the EBVpositive Raji B cell line, no FLAG-RAZ protein could be
detected whereas large amounts of EB1 protein were present in
transfected cells (not shown), suggesting that the RAZ protein
is even less stable in B cells than in HeLa (epithelial) cells. These
last observations cast serious doubts on the validity of the
model in which RAZ would be a negative regulator of the
switch from latency to the lyric cycle. It must be emphasized at
this point that the Z8 cDNA was very poorly represented (one
copy) in the cDNA libraries originally screened, and could
simply represent a product of aberrant splicing (Manet et al.,
1989).
We also observed in the repression assays that the amount
of EB1 protein expressed from plasmid pCMV-EB1 decreased
as the amount of FLAG-RAZ protein expressed from plasmid
pCMV-FLAG-Z8 increased. Since the total amount of CMV
promoter sequences was equal in each transfection, this
observation cannot be explained by what is called 'promoter
interference'. However, this observation can be explained by
our finding that EB1 strongly activates transcription initiated
at the CMV promoter. Thus, when plasmids pCMV-EB1 and
pCMV-FLAG-Z8 are co-transfected, EB1 activates EB1 and
FLAG-RAZ expression from plasmids pCMV-EBI and pCMVFLAG-Z8 respectively whereas FLAG-RAZ represses EB1activated expression of EB1 and FLAG-RAZ, leading to a net
decrease in EB1 production. In these experiments, the amount
of FLAG-RAZ protein produced increased, probably because
the increase in the amounts of plasmid pCMV-FLAG-Z8
transfected partially compensated for the repressive effect of
RAZ. Nevertheless, it is clear that in the repression assays
described here, RAZ represses EBl-mediated transcriptional
activation of both the PZ promoter and the CMV promoter,
probably by mechanisms independent of the formation of
inactive RAZ:EB1 heterodimers. We directly confirmed this
hypothesis by demonstrating that RAZ similarly efficiently
repressed EB1- and EBlgen4-mediated transcriptional activation, although RAZ and EB1gena could not form stable
heterodimers.
A probable mechanism to explain how RAZ represses EB1and EBlgen4-activated transcription is that RAZ titrates a
cellular factor essential for EB1 function which is present in
limiting amounts in the cell. Indeed, RAZ did not detectably
repress transcription initiated at the CMV promoter unless EB1
was expressed, demonstrating that RAZ interferes specifically
with the function of EB1, rather than with a factor required for
the activity of the CMV promoter. The domain of RAZ
involved in repression is likely to be the basic domain, and/or
the R protein moiety present at the N terminus of RAZ, since
the repressive effect of RAZ was also seen on EB1gen4activated transcription. It is noteworthy that the EB1 basic
domain has been shown to interact with cellular factors
important for EB1 transcriptional activation (Mikaelian el al.,
I993 b), and to titrate a cellular factor required for R-activated
transcription (Giot et al., 1991). Moreover, in some efficient
transfections where pCMV-EB1 was transfected in increasing
amounts together with fixed amounts of the reporter gene pZCAT, EBl-activated CAT transcription decreased as the
amount of EBI protein increased (not shown). This last result
strongly suggests that a cellular factor required for EB1activated transcription is present in limiting amount in the
cells.
In conclusion, it must be stressed that although RAZ
represses EBl-activated transcription in transient expression
assays, such a situation has not been yet documented in EBV
infected cells, and it remains to be seen if and when RAZ is
efficiently expressed upon activation of the EBV switch from
latency to the lyric cycle.
We thank Conrad B. Bluinkfor helpfuldiscussionsand commentson
the manuscript.Thisworkwas supportedby INSERM,by the Association
pour la recherche sur le cancer (contract no. 6810 to A.S) (contract no.
2049 to H. G), and by FNCLCC.C.S is a recipientof an MRT fellowship.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23
13!
References
Buisson, M., Manet, E., TrescoI-Biemont, M.-C., Gruffat, H., Durand, B.
& Sergeant, A. (1989). The Epstein-Barr virus (EBV) early protein EB2
is a posttranscriptional activator expressed under the control of EBV
transcription factors EB1 and R. Journal of Virology 63, 5276-5284.
Chang, Y., Dong, D., Hayward, G. & Hayward, S. (1990). The
Epstein-Barr virus Zta transactivator: a member of the bZIP family with
unique DNA-binding specificity and a dimerization domain that lacks the
characteristic heptad leucine zipper motif. Journal of Virology 64,
3358-3369.
Chevallier-Greco, A., Manet, E., Chavrier, P., Mosnier, C., Daillie, J. &
Sergeant, A. (1986). Both Epstein-Barr virus (EBV)-encoded transacting factors, EB1 and EB2, are required to activate transcription from an
EBV early promoter. EMBO Journal 5, 3243-3249.
Cho, M.-S. & Tran, V.-M. (1993). A concatenated form of Epstein-Barr
viral DNA in lymphoblastoid cell lines induced by transfection with
BZLF1. Virology 194, 838-842.
Countryman, J. & Miller, G. (1985). Activation of expression of latent
Epstein-Barr herpesvirus after gene transfer with a small cloned
subfragment of heterogeneous viral DNA. Proceedings of the National
Academy of Sciences, USA 82, 4085-4089.
Countryman, J. K., Jenson, H., Grogan, E. & Miller, G. (1989). A 2"7-kb
rearranged DNA fragment from Epstein-Barr virus capable of disruption
of latency. Cancer Cells 4, 517-523.
Farrell, P.J., Rowe, D.T., Rooney, C. M. & Kouzarides, T. (1989).
Epstein-Barr virus BZLF1 transactivator specifically binds to a consensus
AP-1 site and is related to cFos. EMBO Journal 8, 127-132.
Fixman, E. D., Hayward, G. S. & Hayward, S. D. (1995). Replication of
Epstein-Barr virus oriLyt: lack of a dedicated virally encoded originbinding protein and dependence on Zta in cotransfection assays. Journal
of Virology 69, 2998-3006.
Flemington, E. & Speck, S. H. (1990o). Evidence for coiled-coil dimer
formation by an Epstein-Barr virus transactivator that lacks a heptad
repeat of leucine residues. Proceedingsof the National Academy of Sciences,
USA 87, 9459-9463.
Flemington, E. & Speck, S. H. (1990b). Autoregulation of EpsteinBarr virus putative lytic cycle switch gene BZLF 1. Journal of Virology 64,
1227-1232.
Furnari, F., Zacny, V., Quilivan, B., Kenney, S. & Pagano, J. (1994-).
RAZ, an Epstein-Barr virus transdominant represser that modulates the
viral reactivation mechanism. Journal of Virology 68, 1827-1836.
Giot, J.-F., Mikaelian, I., Buisson, M., Manet, E., Joab, I., Nicolas, J.-C.
& Sergeant, A. (1991). Transcriptional interference between the EBV
transcription factors EB1 and R: both DNA-binding and activation
domains of EB1 are required. Nucleic Acids Research 19, 1251-1258.
German, C., Hoffat, L. & Howard, B. (1982). Recombinant genomes
which express chloramphenicol acetyltransferase in mammalian cells.
Molecular and Cellular Biology 2, 1044-1051.
Gruffat, H., Manet, E., Rigolet, A. & Sergeant, A. (1990). The enhancer
factor R of Epstein-Barr virus (EBV) is a sequence-specific DNA binding
protein. Nucleic Acids Research 18, 6835--6843.
Gutsch, D. E., Holley-Guthrie, E. A., Zhang, Q., Stein, B., Blanar, M. A.,
Baldwin, A.S. & Kenney, S.C. (1994). The bZIP transactivator of
Epstein-Barr virus, BZLFI, functionally and physically interacts with the
p65 subunit of NF-KB. Molecular and Cellular Biology 14, 1939-1948.
Hardwick, J.M., Lieberman, P.M. & Hayward, D. (1988). A new
Epstein-Barr virus transactivator, R, induces expression of a cytoplasmic
early antigen. Journal of Virology 62, 2274-2284.
Kelleher, C., Paterson, R., Dreyfus, D., Streib, J., Xu, J., Takase, K.,
Jones, J. & Gelfand, E. (1995). Epstein-Ba~ virus replicative gene
transcription during de novo infection of human thymocytes: simultaneous early expression of BZLF-1 and its represser RAZ. Virology
208, 685--695.
Kouzarides, T., Packhman, G., Cook, A. & Farrell, P. (1991). The
BZLF1 protein of EBV has a coiled dimerisation domain without a heptad
repeat but with homology to the C/EBP Ieucine zipper. Oncogene 6,
195-204.
Liebowitz, D. & Kieff, E. (1993). Epstein-Barr virus. In The Human
Herpesviruses, pp. 107-172. Edited by B. Roizman, R. J. Whitley & C.
Lopez. New York: Raven Press.
Manet, E., Gruffat, H., TrescoI-Biemont, M.-C., Moreno, N., Chambard,
P., Giot, J.-F. & Sergeant, A. (1989). Epstein-Barr bicistronic mRNAs
generated by facultative splicing code for two transcriptional transactivators. EMBO Journal 8, 1819-1826.
Mikaelian, I., Drouet, E., Marechal, V., Denoyel, G., Nicolas, J.-C. &
Sergeant, A. (1993o). The DNA-binding domain of two bZIP
transcription factors, the Epstein-Barr virus switch gene product EB1 and
Jun, is a bipartite nuclear targeting sequence. Journal of Virology 67,
734-742.
Mikaelian, I., Manet, E. & Sergeant, A. (1993b). The bZIP motif of the
Epstein-Barr virus (EBV) transcription factor EB1 mediates a direct
interaction with TBP. Compte Rendu de l'Academie des Sciences 316,
1424-1432.
Prang, N., Wolf, H. & Schwarzmann, F. (1995). Epstein-Barr virus lytic
replication is controlled by posttranscriptional negative regulation of
BZLF1. Journal of Virology 69, 2644-2648.
Rooney, C., Rowe, D., Ragot, T. & Farrell, P. (1989). The spliced BZLF1
gene of Epstein-Barr virus (EBV) transactivates an early EBV promoter
and induces the virus productive cycle. Journal of Virology 63, 3109-3116.
Schepers, A., Pich, D. & Hammerschmidt, W. (1993). Transcription
factor with homology to the AP-I family links RNA transcription and
DNA replication in the lyric cycle of Epstein-Barr virus. EMBO Journal
I2. 3921-3929.
Sista, N. D., Pagano, J. S., Liao~ W. & Kenney, S. (1993). Retinoic acid
is a negative regulator of the Epstein-Barr virus protein (BZLF1) that
mediates disruption of latent infection. Proceedingsof the National Academy
of Sciences, USA 90, 3894-3898.
Sista, N. D., Barry, C., Sampson, K. & Pagano, J. (1995). Physical and
functional interaction of the Epstein-Barr virus BZLF1 transactivator
with the retinoic acid receptors RARe<and RXR¢~.Nucleic Acids Research
23, 1729-1736.
Takada, K., Shimizu, N., Sakuma, S. & One, Y. (1986). Transactivation
of the latent Epstein-Barr virus (EBV) genome after transfection of the
EBV DNA fragment. Journal of Virology 57, 1016-I022.
Urier, G., Buisson, M., Chambard, P. & Sergeant, A. (1989). The
Epstein-Barr virus early protein EB1 activates transcription from different
responsive elements including AP-1 binding sites. EMBO Journal 8,
1447-1453.
Zhang, Q., Gutsch, D. & Kenney, S. (1994). Functional and physical
interaction between p53 and BZLFI: implications for Epstein-Barr virus
latency. Molecular and Cellular Biology 14, 1929-1938.
zur Hausen, H., O'Neill, F.J., Freese, U. K. & Hecker, E. (1978).
Persisting oncogenic herpesvirus induced by the tumor promoter TPA.
Nature 272, 373-375.
Received 20 November 1995 ; Accepted 13 March 1996
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 12 Aug 2017 16:45:23