The EMBO Journal vol. 12 no. 1 1 pp.4269 - 4278, 1993
Functional and biochemical interaction of the HTLV-1
Taxl transactivator with TBP
Cecile Caron1, Raphael Rousset1,
Christophe Beraud"3, Vincent Moncollin2,
Jean-Marc Egly2 and Pierre Jalinot1
'Laboratoire de Biologie Moleculaire et Cellulaire, Ecole Normale
Superieure de Lyon, UMR 49, CNRS, 46 Allee d'Italie, 69364 Lyon
cedex 07 and 2Laboratoire de Genetique Mol6culaire des Eucaryotes,
CNRS, Unite 184 de Biologie Moleculaire et de Genie Gen6tique,
INSERM, Institut de Chimie Biologique, Faculte de Medecine, 67085
Strasbourg cedex, France
3Present address: Gladstone Institute of Virology and Immunology,
University of California at San Francisco, San Francisco General
Hospital, PO Box 419100, San Francisco, CA 94141-9100, USA
Communicated by C.Kedinger
The human T-celi leukemia virus type I (HTLV-I) codes
for the potent transcriptional activator, Taxl, which
induces the enhancer activity of various enhancer
elements. In the case ofthe 21 bp enhancer ofthe HTLV-I
provirus, this induction is correlated with the association
of Taxl with this DNA element via a specific cellular
factor. That the indirect association of Taxl with DNA
can lead to transcriptional activation has also been
supported by the study ofchimeric GAL4-Taxl proteins.
The GAL4-Taxl stimulatory effect exhibits a strong self-
squelching. In order to determine whether Taxl interacts
directly with the general transcription factors or via
intermediary molecules, we have analyzed how over-
expression oftheTATA binding protein (TBP) and TFIB
protein affects the squelching curve ofGAL4-Taxl. The
data presented here show that overexpression of TBP
strongly increases the stimulatory effect ofGAL4-Taxl,
causes a displacement ofthe maximum of the squelching
curve and partially alleviates the squelching. Under
similar conditions TFlIB exhibited little effect. From
these results we conclude that Taxl can increase the
recruitment of TBP by directly interacting with this
protein. Biochemical experiments with purified proteins
produced in bacteria confirmed that Taxl can interact
with TBP but not with TFIIB. Taxl interacts with the
conserved C-terminal part ofTBP. Analysis ofthe ability
of different mutants of Taxl fused to the GAL4 DNA
binding domain to activate transcription and to associate
with TBP, showed that these activities are correlated.
However, since one transcriptionally inactive mutant was
able to interact efficiently with TBP in vitro, it would
appear that an event other than the Taxl-TBP contact
also intervenes in the activation oftranscription by Taxl.
Key words: DNA binding proteins/HTLV-I/Taxl/TBP trans-
The human T-cell leukemia virus type I (HTLV-I) Taxl
transactivator exhibits pleiotropic activities.
(C) Oxford University Press
several viral and cellular transcriptional promoters through
specific DNA elements by activating their enhancer proper-
ties (Fujisawa et al., 1986; Inoue et al., 1986; Fujii et al.,
1988; Alexandre and Verrier, 1991; Green, 1991). Taxl-
responsive enhancer motifs that have been well defined
include the xB sites present in the HIV-1 provirus long
terminal repeat (LTR) and those in the promoter ofthe gene
coding for the interleukin 2 (IL-2) a-chain receptor (Inoue
et al., 1986; Bohnlein et al., 1988). It is firmly established
that Taxi acts by inducing the nuclear translocation ofseveral
members of the rel family which are normally sequestered
in the cytoplasm (Ballard et al., 1988; Leung and Nabel,
1988; Ruben et al., 1988). Another target sequence for Taxl
is a 21 bp motif repeated three times in the HTLV-I promoter
(Fujisawa et al.,
1986). The enhancer activity of this
sequence is strongly induced by Taxl (Montagne et al.,
1990). By using a DNA affinity precipitation assay we have
previously shown that Taxl forms a ternary complex with
this DNA element and a cellular factor, HEB1, which
specifically binds to this sequence (Beraud et al., 1991). In
agreement with this model of indirect interaction of Taxl
with DNA, it has been established that this protein contains
a transcriptional activation domain by analyzing the activity
ofa fusion protein in which the GAL4 DNA binding domain
is fused to the entire TaxI coding sequence (Fujisawa et al.,
Association of a viral transcriptional activator with DNA
elements through a specific cellular factor is not without
precedent. A detailed genetic study has shown that the
adenovirus Ela protein functionally interacts with ATF2 (Liu
and Green, 1990) and interaction ofthe herpes simplex VP16
protein with Octl has been established biochemically
(Preston et al., 1988). These two viral proteins can also
activate transcription when linked to the GAL4 DNA binding
domain (Martin et al., 1990). Interaction with DNA-bound
cellular factors is therefore likely to be a common means
for viral transcriptional activators to modify the activity of
specific regulatory promoter elements. Another question is,
how do these viral proteins act on the transcriptional
machinery? Transcriptional initiation by RNA polymerase
II (pol II) is a complex molecular process (for a review see
Sawadogo and Sentenac, 1990). Association of pol II with
the template results from a cascade of interactions between
several proteins called the general transcription factors. The
first step is the binding of TFIID to the TATA element
(Davison et al., 1983; Buratowski et al., 1989). This allows
TFIIB to bind (Ha et al., 1991; Moncollin et al., 1992a).
TFIIF mediates association ofpol II with this TFID --TFU1B
complex (Killeen et al., 1992). Addition to this complexof
the TFIIE and TFIIH activities enables pol II to initiate
transcription. The complexity of the whole mechanism
indicates that transcriptional activators can in principle
intervene at different levels. The question of the specific
molecular event stimulated by a given transcriptional
activator has already been addressed by different studies.
C.Caron et al.
In the case of VP16, in which the activator domain is of
the acidic type, Lin and Green (1991), using in vitro
transcription of an immobilized DNA matrix, showed that
the association of TFIIB, but not that of TFIID, is a rate-
limiting step which is specifically stimulated by this viral
transcriptional activator. This hypothesis is supported by the
observation that cloned TFIIB strongly and specifically
interacts with VP16 (Lin et al., 1991). By performing
genetic experiments, Colgan et al. (1993) have recently
shown that the target factor of the glutamine-rich activation
domain of fushi tarazu is TFIIB. It is now well established
that human TFIID is a multi-subunit complex containing
several tightly associated polypeptides (Peterson et al., 1990;
Dynlacht et al., 1991). The core protein of this complex is
a 38 kDa polypeptide which can bind to the TATA box
(Hoffmann et al., 1990; Kao et al., 1990; Peterson et al.,
1990). This TATA binding protein is now called TBP.
Biochemical studies have shown that VP16 can interact with
TBP, although at a lower affinity than with TFIIB (Stringler
et al., 1990; Lin et al., 1991). In in vitro transcription
experiments, cloned TBP can efficiently substitute for the
TFIID fraction for basal transcription, but appears to be only
very weakly stimulated by upstream or enhancer factors
(Peterson et al., 1990). In these in vitro conditions, the
stimulatory activity ofthe upstream/enhancer factors depends
on proteins which in some cases are tightly associated with
TBP (Pugh and Tjian,
1990; Dynlacht et al.,
Meisterernst et al., :1991). It has been shown that the
transcription factor Spi acts by contacting TAFi 1O of
Drosophila TFIID (Hoey et al., 1993). The existence of
intermediary molecules linking the upstream/enhancer factors
to the general transcription factors has also been inferred
from activator interference experiments (Berger et al., 1990;
Kelleher et al., 1990; Tasset et al., 1990). These proteins
have been variously named adaptors, coactivators, cofactors
or transcriptional intermediate factors. Such molecules have
been identified for VP16 (Flanagan et al., 1991; White
et al., 1991) but in the case of Ela, the molecular process
involved in the stimulation of transcription has been less
extensively studied. Lee et al. (1991), however, have shown
by various biochemical experiments that Ela interacts
strongly with TBP.
In order to achieve a better understanding of the activity
of TaxI on the transcriptional machinery, the effect of this
activator on the recruitment of the TBP and TFIIB proteins
was analyzed. This was done in the living cell by taking
advantage of the strong self-squelching exhibited by a
GALA-Taxi chimeric protein. The effect of increasing
amounts of TBP and TFIIB on this squelching effect was
evaluated. The results ofthese experiments support the notion
that TaxI stimulates transcription by increasing the recruit-
ment of TBP but not that of TFIIB. In vitro biochemical
experiments and analysis of several TaxI mutants indicate
that this recruitment involves a direct protein-protein
Effect of TBP overexpression on Tax 1 transcriptional
The GAL4-TaxI protein was constructed by inserting the
TaxI coding sequence between amino acids 2 and 353 down-
stream of the first 148 amino acids of GAL4 (pSG4-Taxl
Tran%feuted plasmzid (ng)
Fig. 1. Squelching effect of GAL4-Taxl. HeLa cells were
cotransfected by the calcium phosphate coprecipitation method with
plasmid pG4G3CAT (3 Ag) and either pG4M or pSG4-Taxl. The
amounts ofpG4M and pSG4-Taxl were 0, 10, 50, 100, 500 and
1000ng. The CAT assay was performed aspreviously described
(Chevallier-Greco et al., 1989) and the conversionpercentage values
were plotted against the amount of transfected pG4M and pSG4-Taxl
plasmids. The exact numbers corresponding to the different points are
given in the table under the plot.
expression vector). Theactivity of this GAL4-Taxl protein
was testedby performing cotransfectionexperiments in HeLa
cellsusing areporter constructbearing four GALAbinding
sitesupstream ofthe c-globin TATA box linked to the CAT
gene (plasmid pG4G3CAT). As already reported by
Fujisawa etal. (1991), the GAL 4-Taxl protein clearly
activates transcription of the reporter gene (Figure 1). In
similar conditions the GALA DNA binding domain did
not activate the basal activity of this construct (pG4M
expression vector, Figure 1). The effect of increasing
amounts of GALA-Taxbe
was evaluated by cotransfecting
different ahiounts of the corresponding expression vector
(pSG4-Taxoi). Thestimulatory effect of GALt-Taxl reached
a maximum at 50ng of transfectedpSG4-Taxl. Increasing
this aorount further strongly weakened the effect which
becamenegligible at 1 ig
has already been reported for many other transcriptional
activators and has been calqed'squelching' (Ptashne, 1988).
This inhibition isthought to result from titration of atarget
factormediatingthe activator effect on the basal transcription
overexpression ofwild-type Taxl protein also inhibits the
activity of GAL4-Taxl. Similar results have been obtained
in ourlaboratory (C.Be'raud, unpublished results). Several
studiesanalyzing thesquelching effect of different activators
have predicted the existence of intermediary molecules
between activators andgeneral transcription factors (Ptashne,
1988; Berger et al., 1990; Keheller et al., 1990; Tasset
etal.,. 1990). The human cDNAs corresponding to the
factors involved in the first two steps ofthe assembly of the
initiation complex, namely TBP and TFIIB, have now been
et al. (1991) have reported that
Interaction of HTLV-1 Taxl transactivator with TBP
'5Lg 0.1jig 0.5,g 2ig
CAT concentration(pgf*(X) plI
Fig. 2. Modifications of the GAL4-TaxI squelching curve by
transfection of increasing amounts of a TBP expression vector.
Plasmid pG4G3CAT (1Ag)was cotransfected with increasing amounts
of pSG4-Taxl (0, 2, 10, 20, 40, 70 and 100 ng) and pSG-TBP (50,
500 and 2000 ng). Quantification of the CAT enzyme was carried out
by using CAT ELISA. The CAT concentration, expressed in pg for
200A1 of cellular extract, was plotted against the amount of transfected
pSG4-Taxl for the different amounts of cotransfected pSG-TBP. The
exact numbers corresponding to each point are given in the table under
cloned (Kao et al., 1990; Peterson et al., 1990; Hoffmann
et al., 1990; Ha et al., 1991; Malik et al., 1991). The effect
of increasing amounts of these two proteins
GAL4-Taxl squelching curve has been analyzed. The
squelching effect is due to limiting amounts of the activator
target protein. Consequently, if the Taxi target protein is
TBP or TFIIB, overexpression of these proteins should
intermediary molecule contacts both TaxI and TBP or
TFIIB, increasing the amounts of these latter two proteins
should increase squelching since the intermediary factor
would then be titered out by both the activator and the general
transcription factor. The GAL4-Taxl squelching curve was
evaluated in the presence ofTBP at three different conditions
(Figure 2). This was achieved by cotransfecting 50 ng,
500 ng or 2 ,ug of a TBP eukaryotic expression vector
(pSG-TBP). The amount of TBP present in the cell when
increasing quantities of expression vector were cotransfected
was monitored by immunoprecipitation (Figure 3). The
results indicate that the TBP concentration indeed increases
approximately linearly with respect to the amount of
transfected expression vector (Figure 3A). Measurement of
the CAT enzyme quantity expressed at different GAL4-Taxl
and TBP concentrations showed that the overexpression of
this factor markedly increased the stimulatory effect of
GAL4-Taxl (up to 7-fold, Figure 2), whereas it did not affect
the basal transcriptional activity. The increase of the
transcription level can be explained by assuming that,
although all of the templates bound to GAL4-Taxl, onlya
ug UlJ Ipg 0.5}ig
Fig. 3. Analysis of TBP and TFIIB factors expressed in the cell for
different amounts of transfected pSG-TBP and pSG-TFIIB plasmids.
HeLa cells were transfected with increasing amounts of pSG-TBP (A)
or pSG-TFIIB (B). The expressed proteins were metabolically labelled
by incubating the transfected cells with a mixture of [35S]methionine
and [35S]cysteine. After immunoprecipitation with specific monoclonal
antibodies, the proteins were analyzed by SDS-PAGE. The
autoradiograms obtained for TBP (A) and TFIIB (B) are shown. The
positions of molecular weight markers run in parallel are indicated on
the left hand side of the gels. The positions of the signals
corresponding to TBP and TFIIB are indicated.
fraction of them are transcriptionally active, due to the
limiting amount of the Taxl target factor. Increasing this
the number of active templates and
consequently the whole level of transcription observed. From
the shape of the GAL4-Taxl squelching curves obtained at
increasing TBP concentrations, two main comments can be
First, the squelching observed with increasing
GAL4-Taxl concentrations was weaker in the presence of
TBP. This was clear at an intermediate TBP concentration
(500 ng curve, Figure 2). When the amount of TBP was
higher the squelching was again quite pronounced (2 Ag
curve, Figure 2). This weakening of the squelching in
response to an increase of the TBP concentration supports
the model of a direct interaction of TaxI and TBP. That the
squelching again increased at elevated TBP concentrations
was probably the consequence of the titration of another
factor by high
Taxi -TBP complex. That a direct interaction occurs
between Taxl and TBP is also supported by the fact that
C.Caron et al.
the maximum level of transcription was obtained
increasingly elevated GAL4-Taxl concentrations when the
amount of TBP was raised. In the case of a rate-limiting
intermediary factor, the maxima of the different curves
should occur at the same, or weaker (in the case ofa titration
of such a protein by TBP) GAL4-Taxi concentration. These
experiments therefore indicate that in the presence of
neighboring Taxl protein, association with TBP is a limiting
event that can be stimulated by increasing the cellular TBP
concentration. Moreover, since TBP can partially alleviate
the squelching resulting from increasing the GAL4-Taxl
concentration and since it can also displace the curve
maximum, Taxl would appear to activate functionally this
transcription factor by directly contacting it.
The ability of various GAL4-Taxl mutants to stimulate
transcription, either with or without TBP overexpression,
was also analyzed. Several mis-sense mutants spanning the
entire Taxl protein (kindly provided by Dr W.C.Greene),
and previously characterized as abrogating its ability to
activate the HTLV-I promoter (Smith and Greene, 1990),
were selected and cloned downstream of the GAL4 DNA
M pSG4-TaxI (10 ng) + pSG-TBP (1 Ug)
0 pSG4-Tax(40 ng)
* pSG4-Tax 1 (4(1 ng) + pSG-TBP (1 Ug)
binding domain. This series of constructs was transfected
into HeLa cells together with either pSG5 or pSG-TBP
(1p/g).The experiment was performed using two different
amounts of pSG4-Taxl (10 ng and 40 ng). Of the seven
selected mutants, five showed a reduced basal activity
20% of the wild-type activity) which was not stimulated
by the overexpression of TBP (M9, M26, M35, M41 and
M47, Figure 4). These results indicated that some Taxl
mutants which fail to activate the HTLV-I promoter also
failed to cause the functional recruitment of TBP. The two
others also exhibited
however, was activated by the TBP overexpression (M5,
M18, Figure 4). As compared with the wild-type construct,
the level of activity reached by these mutants when TBP was
overexpressed was lower, but the fold induction was similar.
In contrast to what was observed with the wild-type
construct, the TBP stimulatory effect was higher for 40 ng
of transfected pSG4-Taxl plasmid than for 10 ng.
a reduced basal activity which,
Effect of TFIIB overexpression on Tax
Overexpression experiments were also performed with a
eukaryotic TFIIB expression vector (Figure 5). As for TBP,
the cellular concentration of TFIHB increased as a function
ofthe amount oftransfected expression vector (Figure 3B).
The effects of increasing cellular TFIIB concentration on
transcription of the reporter gene were moderate compared
with those obtained with TBP. The strongest stimulatory
GAL4-Taxi squelching curve was slightly affected by a
The shape of the
CAT concenlration (pa/2(X) U1)
Fig. 4. Analysis of activating properties and of the sensitivity to TBP
overexpression of various Taxl mutants as GAL4-Taxl fusion
proteins. The Taxl coding sequence including different mutations was
cloned downstream of the GAL4 DNA binding domain. These
mutations, which correspond to the modification of two consecutive
amino acids to either AlaSer or ArgSer, have been previously analyzed
in detail for their ability to activate the HTLV-I and HIV-1 promoters
(see Table 1 in Smith and Greene, 1990). The following mutants were
used: M5 (22AspCys-AlaSer), M9 (41HisArg-AlaSer), M18
(123ThrLeu-AlaSer), M26 (161ProPro-AlaSer), M35
(206MetIle-AlaSer), M41 (287HisPro-AlaSer) and M47
(319LeuLeu-ArgSer). None of these mutants could activate the
HTLV-I promoter. The test plasmid, pG4G3CAT (2 ug), was
cotransfected with either 10 ng or 40 ng of pSG4-Taxl, wild-type and
including the mutations previously cited, and either with or without
pSG-TBP (1 tog). Quantification of the CAT enzyme was done as
described in the legend to Figure 2. The exact numbers corresponding
to each point are given in the table under the plot.
CAT concentration (pg/20()gil)
Fig. 5. Modifications of the GALA-Taxl squelching curve by
transfection of increasing amounts of a TFIIB expression vector. The
test plasmid, pG4G3CAT (1 jig), was cotransfected with increasing
amounts of pSG4-Taxl (0, 2, 10, 20, 40, 70 and 100 ng) and pSG-
TBP (0, 50, 100, 500 and 2000 ng). Quantification of the CAT
enzyme and the representation of results was done as described in the
legend to Figure 2.
Interaction of HTLV-I Taxi transactivator with TBP
weaker inhibitory effect of high GAL4-Taxi concentrations
(2 ug curve, Figure 5). However, contrary to what was
observed with TBP, the maxima ofthe different curves were
always observed at the same GAL4-Taxl concentration. This
indicates that TFIIB is not likely to be a target molecule for
TaxI. The effect ofthe coexpression ofboth TFIIB and TBP
was also analyzed (Figure 6). In the presence of the two
factors the shape of the GAL4-Taxl squelching curve was
not modified although the different points were shifted to
values -35% higher. These data indicate that association
of TFIIB is probably not a rate-limiting step for Taxl-acti-
vated transcription. The stimulatory effect of TFIIB on the
basal and TBP-stimulated activation by Taxl could result
from a better recruitment ofTFIIB to the initiation complex
when its concentration is raised over that normally existing
in the cell. However, this effect appeared to be of limited
Tax1 - TBP interaction
The results obtained with the transient expression experi-
ments are most plausibly explained by a direct interaction
between TaxI and TBP. However, an indirect contact
between both factors mediated by a non-limiting factor could
not be ruled out. As a consequence, the existence of the
Taxi -TBP interaction was probed biochemically. For this
purpose the TaxI and TBP proteins were produced in
Escherichia coli and purified. TaxI was expressed using the
glutathione S-transferase gene fusion system (Smith and
Johnson, 1988). The GST-Taxl fusion protein was produced
in E. coli grown at low temperature (28°C) and bound to
glutathione-agarose beads. After elution by incubation with
free glutathione, bands of molecular weight 65 and 26 kDa
were observed (Figure 7C). The lower bands which co-
migrated with the GST protein probably resulted from
degradation of the 65 kDa GST-Taxl protein. The retention
of purified TBP and TFIIB on GST-
Taxi -agarose beads was evaluated. The beads were
incubated with the proteins in the presence of BSA. After
several washes, GST or GST-Taxl was eluted by treatment
with 5 mM glutathione and the presence of TBP or TFIIB
in the eluate analyzed by Western blotting. The results of
this experiment clearly indicate that TBP was retained on
the GST-Taxl beads but not on those coupled to GST (Figure
7A, lanes 1-4). Since this retention was observed in the
presence of a large amount of BSA, following a wash with
buffer containing 300 mM KCI and with an elution process
which specifically eluted GST-Taxl, we conclude that the
interaction of TBP with Taxl was strong and specific. By
comparison, in similar experimental conditions no retention
of TFIIB could be observed, even when less stringent KCl
washes were performed (Figure 7B, lanes 1-4 and data not
shown). This experiment, using purified proteins, rules out
the possibility of bridge molecules being tightly associated
with either Taxl or TBP and clearly supports the notion of
a direct interaction between the two proteins. It also confirms
that the Taxl viral transactivator is not able to interact with
TFIIB. In order to evaluate the biological significance of this
Taxi -TBP interaction, which was observed in vitro, the
various Taxl mutants tested as fusion proteins with the
GAL4 DNA binding domain were also expressed in bacteria
using the GST system. The different GST-Taxl constructs
were produced in bacteria and selectively adsorbed onto
glutathione-agarose beads. The proteins recovered after
PS(C-Tl-IIB (I pgl)
CATcon<cnuaion pg2(E pla
Fig. 6. Modifications of the GAL4-TaxI squelching curve by
cotransfection of TBP and TFIIB expression vectors. The test plasmid
pG4G3CAT (1 jg)was cotransfected with increasing amounts of
pSG4-Taxl (0, 2, 10, 20, 40 and 100 ng), pSG-TBP (1 Ag)and pSG-
TFIIB (1 Lg). Quantification of the CAT enzyme and the
representation of results was done as described in the legend to
L)- 4l- 1
Fig. 7. TaxI -TBP interaction. The GST and GST-Taxl proteins were
expressed in E. coli and selectively coupled to glutathione-agarose
beads. The GST and GST-Taxl beads were incubated with purified
TBP (A, lanes 1-4) or TFIIB (B, lanes 1-4). The GST and GST-
Taxl proteins were uncoupled from the beads by incubation with free
glutathione and the eluted complexes were loaded on to a 10%
SDS-polyacrylamide gel. The presence of TBP and TFIIB in the
eluate was analyzed by Western blotting using specific monoclonal
antibodies. As a control the proteins eluted from GST and GST-Taxl
beads prior to incubation were analyzed in a protein gel stained using
silver nitrate (C). The TBP and TFUB fractions were also analyzed in
protein gels stained either using silver nitrate (TBP, D) or Coomassie
brilliant blue (TFIIB, E). For each gel the positions of molecular
weight markers run in parallel are indicated on the left hand side.
C.Caron et al.
SDS-PAGE. This experiment showed that the different
mutants were all stable and produced in similar quantities
(Figure 8A). The retention of TBP protein produced by
in vitro translation on agarose beads coupled to these various
mutated proteins was analyzed as described above. As
observed using TBP protein produced in bacteria, the wild-
type Taxl protein interacted with the in vitro-translated TBP
(Figure 8B, lane 1). The mutants M5, M9, M18, M26, M35
and M41 exhibited a reduced affinity for Taxi, whereas the
mutant M47 retained TBP at least as efficiently and possibly
slightly better than the wild-type (Figure 8B). In order to
quantify these different affinities, a densitometric analysis
of the autoradiograms of two independent experiments was
performed, and the amount ofTBP protein retained by each
mutant was expressed as a fraction of that retained by the
wild-type (Figure 8E). The mutants M9, M26, M35 and
M41 exhibited the weakest affinity for TBP. The mutant M5
was slightly more active. The mutant M18 had an affinity
close to that of the wild-type. With the notable exception
of M47, these results indicate that a parallel exists between
the ability of these different mutants to interact with TBP
and their ability to stimulate transcription at GAL4-Taxl
fusion proteins. M5 and M18, which were still activated by
the overexpression ofTBP, interacted with this protein more
efficiently than M9, M26, M35 and M41. The observation
that the M47 mutant was inactive as a GAL4-TaxI fusion
protein but efficiently interacted with TBP in vitro, indicates
were analyzed using
Fig. 8. Analysis of the ability of Taxl mutants to interact with TBP.
(A) The mutants of Taxl previously tested as GAL4-Taxl fusion
proteins (see legend to Figure 4) were produced in bacteria using the
glutathione S-transferase systeli. The proteins were selectively bound
to glutathione-agarose beads, eluted by incubation with free
glutathione and analyzed by SDS-PAGE. The gel stained with
Coomassie brilliant blue is represented. The black circle indicates the
position of the GST protein (lane 1) and the plus sign that of the GST-
Taxl protein, either wild-type (lane 2) or including the mutation M5
(lane 3), M9 (lane 4), M18 (lane 5), M26 (lane 6), M35 (lane 7),
M41 (lane 8) or M47 (lane 9). (B) TBP was produced and labelled
with [35S]methionine by in vitro translation; it was incubated with
agarose beads coupled to GST-Taxl protein, either wild-type (lane 1)
or including the mutation M5 (lane 2), M9 (lane 3), M18 (lane 4),
M26 (lane 5), M35 (lane 6), M41 (lane 7) or M47 (lane 8). The
proteins eluted by incubation with free glutathione were analyzed by
SDS-PAGE. The autoradiogram of the protein gel is represented. The
black arrow indicates the position of the TBP protein. (C) The TBP-N
(lane 1) and TBP (lane 2) proteins were produced by in vitro
translation. The former corresponds to the N-terminal part of TBP
between amino acids 1 and 161. An aliquot of the in vitro translation
products was analyzed by SDS-PAGE. The autoradiogram of the gel
is shown Oanes 1 and 2). TBP-N was incubated with agarose beads
coupled to GST protein (lane 3) and GST-Taxl, either wild-type
(lane 5) or including the mutation M5 (lane 6), M9 (lane 7), M18
(lane 8), M26 (lane 9), M35 (lane 10), M41 (lane 11) or M47
(lane 12). As a control the retention of the entire TBP molecule was
tested with agarose beads coupled to wild-type GST-Taxl protein
(lane 4). The black arrow indicates the position of TBP. (D) The
interaction of TBP-C, which corresponds to the C-terminal part of
TBP between amino acids 159 and 339, with wild-type and mutated
TaxI proteins was tested exactly as described for TBP-N. The white
arrow indicates the position of TBP-C. (E) and (F) The
autoradiograms corresponding to two independent experiments of
interaction of TBP (E) or TBP-C (F) with wild-type and mutated Taxl
proteins were analyzed by densitometry. The signal measured for each
mutant was divided by that corresponding to the wild-type protein. The
plots represent the mean of the values obtained for the different
that in the living cell another as yet unidentified event is
required to induce transcriptional activation.
Comparison of the TBP sequence of different species
clearly indicates that it has a bipartite organization. The N-
terminal domain is of variable size and is poorly conserved.
The C-terminal 180 amino acids are well conserved and
contain all the activity necessary for pol II basal transcrip-
tion, as evaluated by in vitro transcription experiments. In
order to determine which part of TBP interacts with Taxl,
the retention ofeither the N-terminal or C-terminal domains
of the human TBP protein on the wild-type and mutated
GST-Taxl proteins was analyzed and compared with that
of the wild-type protein. Proteins TBP-N (amino acids
1-161) and TBP-C (amino acids 159-339) were produced
by in vitro translation (Figure 8C and D, lanes 1). TBP-N
was not detectably retained by TaxI (Figure 8C, lane 5),
whereas TBP-C clearly was (Figure 8D, lane 5). Comparison
of the intensity of the signals corresponding to the initial
in vitro translation products and to the fractions retained by
TaxI shows that wild-type TBP and TBP-C interact with
TaxI with a similar efficiency (Figure 8D, lanes 1, 2, 4 and
5). The different TaxI mutants exhibited the same pattern
of relative affmnities for TBP-C as for the entire TBP
(Figure 8E and F). These results clearly indicate that TaxI
interacts with the C-terminal domain of TBP. In order to
confirm that the part ofTBP which is functionally involved
in the TaxI activation process indeed corresponds to the 180
C-terminal amino acids, the effect of overexpression ofTBP-
Interaction of HTLV-I Taxi transactivator with TBP
Tpa smx GG1A
evalats e dto
loivingtcell on the GAL4-Taxi squelching curve
pn re vios lyg
pS4TBP 0 2 0,
andre tondisplace theC-emai
mumpar of The
60we hs w
examined whether Taxl can stimulate tran
t hansfc e
Tn hebinding ofepl
involvingd aspeveadifferen aectivities.
DNA Is-Tax vqery
cm leur oew ss
DNAhitheractions TBPhe difrenutsofteps followrameftinediorder
tand are- likeyt able
the TATubovi the firstr event.s
The corenroteiof TFlIItoDNA
notfeonly ienethe formancition oftrsit
stdeahvcsabihdthaitiTBP psemlasaky rolte
thosend of RNariu prolyerasesotiand and(Mrgottin-
Comack adfned Struhe,
sciptio eNyitherf bya
are likely to be
betweenfothese tof posiiite,Ah
throg fraiontermtedir mntitolecules.x
a GAL4-Taxl chimeric protein was examined. The results
obtained indicate that TBP is not only able partiallyto relieve
the squelching but also strongly stimulates the activating
effect of the GAL4-TaxI construct. This activity is specific
since it was not obtained with GAL4 DNA binding domain
(C.Caron, unpublished results) and because TBP alone had
no effect on the basal transcription. Another interesting
observation was that the maximum of transcription was
obtained at higher GAL4-Taxl concentrations when the
intracellular concentration ofTBP was raised. These results
are most simply explained by a direct interaction between
TaxI and TBP. Such an interaction was indeed observed by
testing bacterially produced purified proteins. This interaction
was also observed with TBP obtained by in vitro transla-
tion and can be detected with TBP partially purified from
nuclear extracts of Jurkat cells (R.Rousset, unpublished
results). The Taxi -TBP interaction and the squelching
correction were also observed using only the C-terminal part
of this protein. The weaker effects observed in the over-
expression experiments were probably the consequence of
a reduced stability or of a lower concentration of the
truncated protein in the cell. Although the C-terminal part
of TBP is well conserved between different species, Taxl
is not able to interact with yeast TBP (C.Caron, unpublished
results). Taken together these results indicate that Taxi can
activate the formation ofthe transcriptional initiation complex
by contacting TBP. In agreement with this statement, the
ability of different Taxl mutants to stimulate transcription
in response to TBP overexpression can be correlated with
their in vitro affinity for TBP. These different mutants, which
spanned the entire Taxl molecule, did not allow the defini-
tion of a precise region which contacts TBP. This is probably
due to disruption of the general conformation of Taxl in
several of these mutants.
It is clear from this study, however, that the mechanisms
by which Taxl stimulates transcription are more complex
than a single Taxi-TBP contact. Many questions remain
and several points will be interesting to study. In particular,
the mutant M47 was unable to activate transcription as a
GAL4-TaxI fusion protein either with TBP overexpression
or without, but still interacted strongly with TBP in vitro.
This indicates that an event other than the Taxi -TBP
interaction is necessary to induce activation oftranscription.
A very similar observation has been made in the case of the
Ela transactivator (Lee et al., 1991). In this perspective,
the role ofthe proteins associated with TBP will be important
to determine. It is possible that TaxI contacts both TBP and
a TAF. The cloning of the cDNAs coding for the various
TAFs, which is in progress (Dynlacht et al., 1993; Hisatake
et al., 1993; Hoey et al., 1993; Ruppert et al., 1993),
should greatly facilitate the characterization of such inter-
actions. Since in the case of pol II TBP intervenes as the
multisubunit complex TFIID, one important question to
address was the variation in the amount of TFIID observed
in response to TBP overexpression. This point was evaluated
by preparing nuclear extracts of cells transfected either with
a control plasmid or the TBP expression vector, pSG-TBP.
These extracts were immunoprecipitated using an anti-
CCG1/TAF 250 antibody [kindly provided by Drs Sekiguchi
and Nishimoto (Sekiguchi et al., 1991)] and the precipitated
proteins analyzed by Western blotting with an anti-TBP
antibody. Alternatively, these extracts were sedimented
through a glycerol gradient and the different fractions
analyzed by Western blotting using both the anti-CCGl/TAF
250 and anti-TBP antibodies. These experiments indicated
that overexpression ofTBP indeed causes an increase of the
amount of TFIID, but this was less importantthan that of
C.Caron at al.
unpublished results)]. This was probably due to limiting
amounts of TAFs. A model of dual interaction of TaxI with
TBP and one of the TAFs is therefore compatible with our
observations. In this regard, it is interesting to note that it
has been shown by in vitro transcription experiments using
purified factors that various activators, in particular Ela,
function more efficiently with TFIID than with TBP (Zhou
et al., 1992). Interestingly,
association with the different TAFs requires only the
conserved C-terminal domain of TBP (Zhou et al., 1993).
However, it is also possible that the function affected by the
M47 mutant is at another level. A non-limiting factor, other
than a TAF, could play an important role in the process.
Alternatively, binding of TFIID to the TATA box could be
blocked by the chromatin structure. This would explain why
overexpression of TBP, and consequently of TFIID, does
not affect the basal level of transcription under our
experimental conditions. Conceptually, Taxl could relieve
this negative effect by exerting an anti-repressive effect and
stimulate the TFIID recruitment by contacting TBP. In order
to discriminate between these different possibilities it will
be necessary to perform in vitro transcription experiments
with purified TFIID factor in the presence and absence of
chromatin components that inhibit formation of the transcrip-
tional initiation complex (Croston et al., 1991; Laybourn
and Kadonaga, 1991; Workman et al., 1991).
Assembly of the transcriptional initiation complex takes
place in several steps. The second event of the cascade is
the association ofTFIIB. The DNA-TBP-TFIIB complex
represents a molecular structure which is recognized by the
pol II associated with TFIIF (Greenblatt, 1991; Killeen et al.,
1992). An increase in the cellular TFIIB concentration did
not markedly modify the GAL4-Taxl
Moreover, no interaction of Taxl with TFIIB was detected
using purified proteins. Therefore, the entry of TFIIB into
the complex is not likely to be stimulated by Taxl. After
being attenuated at intermediate TBP concentrations, the
squelching effect of GAL4-Taxi was again clearly observed
at high TBP concentrations. This probably resulted from the
titration of another general transcription factor by the
Taxi -TBP complex. Since this effect was not corrected by
TFUB, this latter protein does not correspond to the limiting
factor involved in this inhibition. As already mentioned,
association of TFIIB has been described as the target step
of the acidic VP16 activator. In this regard, it is interesting
to note that the acidic domain contained in the C-terminal
part of Taxl is not required for transcriptional activity (Smith
1990; Fujisawa et al.,
interesting to determine whether Taxl and VP16 can co-
operatively induce transcription or not. Indeed, the capacity
of different transcriptional activators to function syner-
gistically or antagonistically probably results from their
respective ability to stimulate different or identical events.
In conclusion, the data presented in this paper indicate that
biochemically interact with the TATA box binding protein,
TBP. In the living cell, the entry of TFIID into the
transcriptional initiation complex appears to be a rate-limiting
step which is stimulated by Taxi. Since Taxi does not
interact specifically with DNA but binds to the HTLV-I
21 bp enhancer sequence through a specific cellular factor
(Beraud et al., 1991), this protein offers an example of a
-5-fold for 2 ,ug of transfected pSG-TBP (C.Caron,
it has also been shown that
It will be
bridge molecule linking enhancer/upstream DNA-bound
factors to the transcriptional machinery. Taxl
the enhancer activity of different types of sequences. In the
case of the xB site it does not bind to the DNA motif but
induces the nuclear translocation of several factors of the
rel family (B6hnlein et al., 1988; Beraud, 1992). However,
it has recently been shown that Taxi is able to bind to
activity of the SRE elements located in the promoters of
genes such as c-fos, egrl or egr2 (Fujii et al., 1992). The
ability of this viral transactivator to function as a coactivator
for different cellular enhancer factors would explain the
diversity of its activities.
this interaction causing the induction of the enhancer
Materials and methods
Transfection and plasmids
HeLa cells, grown in monolayers to 40% confluence, were transfected by
the calcium phosphate coprecipitation method. The level of CAT protein
was measured either as previously described (Chevallier-Greco et al., 1989)
or using an ELISA (Boehringer Mannheim) which was performed according
to the manufacturer's instructions. Transfections were done without an
internal control (Farr and Roman, 1992) but were repeated three times or
more with at least two different plasmid preparations. Plasmid pG4G3CAT,
which contains four GAL4 binding sites and the 3-globin TATA box, was
constructed as follows: theXhoI-SmaI
pBLCAT2 (Luckow and Schuitz, 1987) was inserted between the PvuII and
NdeI restriction sites of plasmid pG3, giving plasmid pG3CAT. Both insert
and vector were filled in. pG4G3CAT was obtained by inserting the
SmiaI-HindI restriction fragment of plasmid pG4GI (Manetetal., 1993)
between the SmnaI and HindIll restriction sites of plasmid pG3CAT. The
plasmid pSG4-Taxl, which expresses the GALA-Taxl fusion protein, is
a derivative of the pG4MpolyEI plasmid (Webster et al., 1989). The Taxl
coding sequence was generated by PCR amplification using specific
oligonucleotides. The PCR fragment was cut withXhoI and BamHl restriction
enzymes andinserted between theXhoI andBgl
pG4Mpolyl. The GALA-Taxl
contains the first 148 amino acids ofGALA and the Taxl coding sequence
between amino acids 2 and 353. The GAL4-Taxl constructs including
mutations in the TaxI coding sequence were generated in a similar manner
by performing the PCR amplification using the Taxi expression vector,
pcTaxl, with mutations M5, M9, M18, M26, M35, M41 and M47 (Smith
and Greene, 1990). The plasmids pSG-TBP and pSG-TFIIB are derivatives
of the pSG5 expression vector (Green et al., 1988). The entire coding
sequence of the human TBP and TFIIB proteins was inserted between the
EcoRI and BamHI restriction sites of pSG5. The plasmids pSG-TBPN and
pSG-TBPC were generated by inserting between the EcoRI and BamHI
restriction sites of pSG5 a DNA fragment generated by PCR amplification
using specific oligonucleotides. For pSG-TBPN this fragment included the
sequence coding for the 161 first amino acids of TBP. For pSG-TBPC this
fragment included the TBP coding sequence between amino acids 159 and
339. The amounts of the different plasmids used in the differenttransfections
are indicated in the legends to the figures. The total amount of SV40
promoter-containing plasmids was adjusted to a constant level with pSG5
and the total amount of transfected DNA was adjusted to 15
restriction fragment of plasmid
restriction sites of plasmid
fusion protein expressed by this vector
tg with pUC
Immunoprecipitations of TBP and TFIHB were performed as described by
Harlow and Lane (1988) with NP40 lysis buffer complemented with1mM
DTT. The transfected cells were incubated for 3 h with 100yCi of amixture
of [35S]methionine and [35S]cysteine (Translabel, Dupont). After two
washes with PBS the cells were incubated in1ml of NP40 buffer and the
lysate was centrifuged at 200 000 g for 30 min. The supernatant was
precleared by incubation with protein A-Sepharose preswollen in NP40
buffer. After collection of the beads by brief centrifugation, the lysate was
incubated with monoclonal antibodies dirted against TBP (3G3, Brou et al.,
1993) or TFIIB (2G8 and 4AIO; V.Moncollin and J.-M.Egly, submitted)
for 1h at 4°C. After addition ofprotein A-Sepharose and a further 30 min
incubation, the beads were collected by brief centrifugation and washed
twice in NP40buffer. The proteins were resuspended in
loading buffer and were separated in a 10% SDS protein gel.
Interaction of HTLV-I Taxi transactivator with TBP
Protein -protein interactions
The TBP and TFIIB protein were produced in E.coli and purified by
chromatography (Moncollin et al., 1992b). The Taxl protein was produced
by using the glutathione S-transferase (GST) gene fusion system. Briefly,
the TaxI coding sequence was inserted in the SmaI restriction site of plasmid
pGEX-2T (Smith and Johnson, 1988). The resulting plasmid, pGST-Taxl,
produces a fusion protein including GST and the Taxl coding sequence
between amino acids 2 and 353. The bacterial culture and the IPTG induction
of GST-Taxl expression were perfonned at 28°C in order to avoid complete
precipitation of this protein in inclusion bodies. Bacteria were resuspended
in MTPBS buffer (150 mM NaCl, 12.5 mM Na2HPO4, 2.5 mM KH2PO4,
100mM EDTA, pH 7.3), and lysed by treatment with lysozyme (0.1 mg/ml)
for 10 min at 4°C followed by ultrasonication. The lysate was incubated
with glutathione-agarose beads in the presence of 1% Triton X-100. The
beads were washed in buffer D 80 (20 mM Tris pH 7.9, 10% glycerol,
80 mM KCI, 1 mM MgC12, 0.2 mM EDTA, 10 yM ZnCl2, 0.5 mM DTT,
0.5 mM PMSF) plus 1% Triton X-100 and resuspended in buffer D 80
plus 0.25% Triton X-100. TBP or TFIIB proteins were incubated with the
GST or GST-TaxI beads for 2 h at 4°C in the presence ofBSA (1 mg/ml).
After collection by brief centrifugation, the beads were washed with buffer D
300 (including 300 mM KCI) and D 80. The GST or GST-Taxl proteins
were eluted by treatment for 10 min on ice with a buffer containing 5 mM
glutathione, 50 mM Tris, pH 8.0, 0.2 mM EDTA, 10 /LM ZnCl2 and 300
mM KCI. The eluted proteins were separated in a 10% SDS protein gel
and the presence of the TBP and TFIIB proteins was analyzed by Western
blotting using monoclonal antibodies specific for these proteins and the ECL
system (Amersham). The experiments with in vitro-translated TBP were
performned under the same conditions. TBP, TBP-N and TBP-C were
produced and labelled with [35S]methionine using the TNT Coupled
Reticulocyte Lysate System (Promega) and plasmids pSG-TBP, pSG-TBPN
and pSG-TBPC. These proteins were revealed by autoradiography. Before
exposure to film the protein gel was treated with Amplify (Amersham) and
We wish to thank I.Mikaelian and A.Sergeant for the generous gift of
plasmids pG4GI and T.Sekiguchi and T.Nishimoto for providing us with
anti-CCGI antibody. Special thanks go to W.C.Greene for giving the Taxl
mutants. We are very grateful to C.B.Bluink for critical reading of the
manuscript. This work was supported by the Agence Nationale de Recherches
sur le SIDA (ANRS) and the Association pour la Recherche sur le Cancer
Alexandre,C. and Verrier,B. (1991) Oncogene, 6, 543-551.
Ballard,D.W., B6hnlein,E., Lowenthal,J.W., Wano,Y., Franza,B.R. and
Greene,W.C. (1988) Science, 241, 1652-1655.
Beraud,C. (1992) PhD Thesis. Universite C.Bernard, Lyon.
Beraud,C., Lombard-Platet,G., Michal,Y. and Jalinot,P. (1991) EMBO J.,
Berger,S.L., Cress,W.D., Cress,A., Triezenberg,S.J. and Guarente,L.
(1990) Cell, 61, 1199-1208.
Bohnlein,E., Lowenthal,J.W., Siekevitz,M., Ballard,D.W., Franza,B.R.
and Greene,W.C. (1988) Cell, 53, 827-836.
Brou,C., Chaudhary,S., Davison,I., Lutz,Y., Wu,J., Egly,J.-M., Tora,L.
and Chambon,P. (1993) EMBO J., 12, 489-499.
Buratowski,S., Hahn,S., Guarente,L. and Sharp,P.A. (1989) Cell, 56,
Chevallier-Greco,A., Gruffat,H., Manet,E., Calender,A. and Sergeant,A.
(1989) J. Virol., 63, 615-623.
Colgan,J., Wampler,S. and Manley,J.L., (1993) Nature, 362, 549-553.
Comack,B.P. and Struhl,K. (1992) Cell, 69, 685-696.
Comai,L., Tanese,N. and Tjian,R. (1992) Cell, 68, 965-976.
Croston,G.E., Kerrigan,L.A., Lira,L.M., Marshak,D.R. and Kadonaga,J.T.
(1991) Science, 251, 643-649.
Davison,B.L., Egly,J.-M., Mulvihill,E.R. and Chambon,P. (1983)Nature,
Dynlacht,B.D., Hoey,T. and Tjian,R. (1991) Cell, 65, 563-576.
Dynlacht,B.D., Weinzierl,R.O.J., Admon,A.andTjian,R. (1993)Nature,
Farr,A. and Roman,A. (1992)Nucleic Acids Res., 20, 920.
Komnberg,R.D. (1991) Nature, 350, 436-438.
Fujii,M., Sassone-Corsi,P. and Verma,I.M. (1988)Proc. NatlAcad. Sci.
USA, 85, 8526-8530.
Fujii,M., Tsuchiya,H., Chuhjo,T., Akizawa,T. and Seiki,M. (1992) Genes
Dev., 6, 2066-2076.
Fujisawa,J., Seiki,M., Sato,M. and Yoshida,M. (1986) EMBO J., 5,
Fujisawa,J., Toita,M., Yoshimura,T. and Yoshida,M. (1991)J. Virol., 65,
Green,J.E. (1991) Mol. Cell. Biol., 11, 4635-4641.
Green,S., Issemann,I. and Sheer,E. (1988) Nucleic Acids Res., 16, 369.
Greenblatt,J. (1991) Cell, 66, 1067-1070.
Ha,I., Lane,W.S. and Reinberg,D. (1991) Nature, 352, 689-695.
Harlow,E. and Lane,D. (1988) Antibodies: A Laboratory Manual. Cold
Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Hisatake,K., Hasegawa,S., Takada,R., Nakatani,Y., Horikoshi,M. and
Roeder,R.G. (1993) Nature, 362, 179-181.
Hoey,T., Weinzierl,R.O., Gill,G., Chen,J.-L., Dynlacht,B.D.and Tjian,R.
(1993) Cell, 72, 247-260.
Hoffmann,A., Sinn,E., Yamamoto,T., Wang,J., Roy,A., Horikoshi,M.
and Roeder,R.G. (1990) Nature, 346, 387-390.
Inoue,J., Seiki,M., Taniguchi,T., Tsuru,S. and Yoshida,M. (1986)EMBO
J., 5, 2883-2888.
Kao,C.C., Lieberman,P.M., Schmidt,M.C., Zhou,Q., Pei,R. and Berk,A.J.
(1990) Science, 248, 1646-1650.
Kelleher,R.J.,III, Flanagan,P.M. and Komberg,R.D. (1990) Cell, 61,
Killeen,M., Coulombe,B. and Greenblatt,J. (1992) J. Biol. Chem., 267,
Laybourn,P.J. and Kadonaga,J.T. (1991) Science, 254, 238-245.
Lee,W.S., Kao,C., Bryant,G.O., Xuan,L. and Berk,A.J. (1991) Cell, 67,
Leung,K. and Nabel,G.J. (1988) Nature, 333, 776-778.
Lieberman,P.M., Schmidt,M.C., Kao,C.C. and Berk,A.J. (1991) Mol. Cell.
Biol., 11, 63-74.
Lin,Y.S. and Green,M.R. (1991) Cell, 64, 971-981.
Lin,Y.S., Ha,I., Maldonado,E., Reinberg,D. and Green,M.R. (1991)
Nature, 353, 569-571.
Liu,F. and Green,M.R. (1990) Cell, 61, 1217-1224.
Luckow,B. and Schutz,G. (1987) Nucleic Acids Res., 15, 5490.
Malik,S., Hisatake,K., Sumimoto,H., Horikoshi,M. and Roeder,R.G.
(1991) Proc. Natl Acad. Sci. USA, 88, 9553-9557.
Manet,E., Allera,C., Gruffat,H., Mikaelian,I., Rigolet,A. and Sergeant,A.
(1993) Gene Expr., 3, 49-59.
Margottin,F., Dujardin,G., Gerard,M., Egly,J.-M., Huet,J. and Sentenac,A.
(1991) Science, 251, 424-426.
Martin,K.J., Lillie,J.W. and Green,M.R. (1990) Nature, 346, 147-152.
Meisterernst,M., Roy,A.L., Lieu,H.M. and Roeder,R.G. (1991) Cell, 66,
Moncollin,V., Fischer,L., Cavallini,B., Egly,J.-M.and Chambon,P. (1992a)
Proc. Natl Acad. Sci. USA, 89, 397-401.
Moncollin,V., Schaeffer,L., Chalut,C. and Egly,J.-M. (1992b) Prot. Expr.
Pur., 3, 374-379.
Montagne,J., Beraud,C., Crenon,I., Lombard-Platet,G., Gazzolo,L.,
Sergeant,A. and Jalinot,P. (1990) EMBO J., 9, 957-964.
Peterson,M.G., Tanese,N., Pugh,B.F. and Tjian,R. (1990) Science, 248,
Preston,C., Frame,M.M.C. and Campbell,M.E.M. (1988) Cell, 52,
Ptashne,M. (1988) Nature, 335, 683-689.
Pugh,B.F. and Tjian,R. (1990) Cell, 61, 1187-1197.
Reeves,R., Gorman,C.M.and Howard,B. (1985)NucleicAcids Res., 13,
Ruben,S., Poteat,H., Tan,T., Kawakami,K., Roeder,R.G., Haseltine,W.
and Rosen,C.A. (1988) Science, 241, 89-91.
Ruppert,S., Wang,E.H.andTjian,R. (1993) Nature, 362, 175-179.
Sawadogo,M.and Sentenac,A. (1990)Annu. Rev. Biochem., 59,711-754.
Schultz,M.C., Reeder,R.H.and Hahn,S. (1992) Cell, 69, 697-702.
Sekiguchi,T., Nohiro,Y., Nakamura,Y., Hisamoto,N.and Nishimoto,T.
(1991) Mol. Cell. Biol., 11, 3317-3325.
Simmnen,K.A., Bernus,J., Parry,H.D., Stunnenberg,H.G.,Berkenstam,A.,
Smith,D.B.and Johnson,K.S. (1988) Gene, 67, 31.
Smith,M.R.and Greene,W.C. (1990)Genes Dev., 4, 1875-1885.
Stringler,K.F., Ingles,C.J.and Greenblatt,J. (1990)Nature, 345,783-786.
Tasset,D., Tora,L., Fromental,C., Scheer,E.and Chambon,P. (1990) Cell,
and Mattaj,I.W. (1991) EMBO J., 10,
C.Caron et al.
Chambon,P. (1989) EMBO J., 8, 1441-1446.
White,J.H., Brou,C., Wu,J., Burton,N., Egly,J.-M. and Chambon,P. (1991)
Proc. Natl Acad. Sci. USA, 88, 7674-7678.
Workman,J.L., Taylor,I.C.A. and Kingston,R.E. (1991) Cel, 64, 533-544.
Zhou,Q., Liebermian,P.M., Boyer,T.G. and Berk,A.J. (1992) Genes Dev.,
Zhou,Q., Boyer,T.G. and Berk,A.J. (1993) Genes Dev., 7, 180-187.
Received on July 5, 1993