ABSTRACT
Rev-erbA
[alpha]
is an orphan nuclear receptor that functions as a dominant transcriptional
repressor. Tissue culture and in situ
hybridisation studies indicated that Rev-erbA[alpha]
plays an important role in mammalian differentiation and development. Previous
studies have localised the silencing domain of Rev-erbA[alpha]
to the D/E region of the orphan receptor. This study utilised the GAL4 hybrid
system to demonstrate that efficient repression is mediated by 34 amino acids
(aa) between aa 455 and 488 in the E region of the receptor. This domain
contains the ligand binding domain (LBD)- signature motif [(F/W)AKXXXXFXXLXXXDQXXLL] and a region that, according to the recently published crystal structures
of steroid receptors, would be predicted to form helix 5 of the canonical LBD
structure. Fine deletions and site-specific mutagenesis indicated that both the LBD signature motif and helix
5 were necessary for efficient silencing. Utilising mammalian two hybrid
technology, we have also demonstrated that Rev-erbA[alpha]
does not associate with the interaction domain (aa 2218-2451) of the nuclear receptor co-repressor, N-CoR, that is known to interact with the thyroid hormone and
retinoic acid receptors. This suggested that transcriptional repression by Rev-erbA[alpha] is not
mediated through an interaction with N-CoR. In conclusion, we have identified and characterised the minimal
domain of Rev-erbA[alpha], that mediates transcriptional repression by this orphan receptor.
The Rev-erb family of proteins are members of the type III/IV steroid/thyroid
hormone nuclear receptor superfamily that act as ligand-inducible transcriptional regulators (
1
). This superfamily binds DNA through two highly conserved zinc-co-ordinated modules that contact specific bases within
cis
-acting regulatory elements of target genes, termed hormone response
elements (HREs) (
2
). Binding specificity is determined by a short sequence within the DNA-binding domain (DBD), termed the P-box (
3
). Type II, III and IV steroid/thyroid receptors recognise the hexamer motif,
AGGTCA. These nuclear receptors bind to a vast array of characterised
hydrophobic ligands, such as thyroid hormone (T
3
), all-
trans
-retinoic acid (all-
trans
-RA) and vitamin D, providing a direct link between extracellular hormonal
signals and transcriptional responses. The Rev-erb family of receptors belong to a rapidly expanding subclass of this
superfamily termed `orphan receptors' which, although structurally and
functionally related, have no identified ligand (
4
).
Two isoforms of the Rev-erb family have been isolated from mammalian genotypes: Rev-erbA[alpha] (
5
) [also known as Ear-1; (
6
)] and RVR (
7
) [also known as Rev-erb[beta]; (
8
,
9
) and BD73; (
10
)]. Major differences between the two isoforms occur within the hyper-variable A/B and D regions of the proteins (
10
). Both isoforms are expressed in a wide range of tissues and are present in all
major organs. Rev-erbA[alpha] mRNA is upregulated during adipocyte differentiation but repressed
during myogenesis (
11
,
12
). Addition of planar aromatic antioxidants (e.g. butylated hydroxytoluene) to
HepG2 cells induces the expression of both isoforms of the Rev-erb family, with a profile similar to that of immediate early response
genes (e.g. c-
fos
and c-
jun
) (
10
). The physiological and developmental role of the Rev-erb family during embryogenesis is largely unknown. Insight into their
possible role and function, however, has been gained by the observation that
constitutive expression of Rev-erbA[alpha] cDNA can block morphological and biochemical differentiation of
mouse skeletal muscle cells (
12
).
Both receptors have been shown to bind as monomers to an asymmetric 11 base pair
(bp), (
A
/
T
)
6
AGGTCA motif (
9
,
10
,
13
). Rev-erbA[alpha] has also been demonstrated to bind to a novel HRE preferentially
as a homodimer. The HRE consists of two tandemly arranged AGGTCA motifs
separated by 2 bp with unique 5' flanking and spacer nucleotides (RevDR-2) (
14
). Recently, a naturally occurring RevDR-2 has also been identified in the promoter of Rev-erbA[alpha] suggesting auto-regulation of its own transcript (
15
).
Reports on the transcriptional properties of the Rev-erb family were initially conflicting. Rev-erbA[alpha] was first reported to act as a constitutive activator of
transcription through its cognate monomeric asymmetric motif (
13
). More recently, we and other groups have demonstrated that members of the Rev-erb family are, in fact, constitutive repressors of transcription (
7
,
9
,
10
,
12
,
14
). The Rev-erb family does not activate transcription and can repress
trans
-activation by RORs (retinoic acid-related orphan receptors) and RARs (retinoic acid receptors).
The ability of steroid receptors to repress basal transcription has long been
established (
16
-
18
), although the mechanism by which steroid receptors achieve repression still
remains unclear. The recent characterisation of two co-repressors, N-CoR and SMRT, that interact with unliganded thyroid hormone receptor
(TR) and RAR has shed some light on the mechanism of nuclear receptor
repression (
19
-
21
). We have demonstrated previously, using the GAL4 hybrid system, that Rev-erbA[alpha] contains an active transcriptional silencing domain located in the
hinge/LBD (
12
). Therefore, we characterised further the repression domain of the Rev-erbA[alpha] receptor to a minimal region (34 amino acids) that is conserved
within the Rev-erb family (97% homologous to RVR) and located in the E domain of the
protein. The identified silencing domain was then examined in the context of
the recently published crystal structures of steroid receptor LBDs (TR, RAR and
the retinoid x receptor: RXR). Possible interaction between the recently
characterised steroid receptor co-repressor, N-CoR, and Rev-erbA[alpha] was also studied utilising mammalian two hybrid
technology.
GMUQ142 5'-TCCAGCGGATCCTCCCCCAGCCGGACC-3'
GMUQ132 5'-CGGGATCCGGGCGGGTCACTGGGCGTCC-3'
GMUQ171 5'-GCGCGTCGACATATGGAAGACAGCAGCCGA-3'
GMUQ172 5'-GGCGTCGACTATTGCCGGGCGGGTCACTGGGCGTCC-3'
GMUQ173 5'-GCGGAATTCACCATGTATGGAAATTATTCCCAC-3'
GMUQ176 5'-GCGGAATTCCTCAGGTGATCTGCAGTGGGGTCT-3'
GMUQ296 5'-GCGAATTCACCATGGTNAA
A
/
G
A
/
T
CNAA
G
/
A
-
AA
G
/
A
CA-3'
GMUQ297 5'-GCGAATTCACCNC
A
/
T
A
/
G
TCN
G
/
C
A
/
T
NA
A
/
G
N- GT
T
/
C
TC
G
/
A
TA
T
/
C
TG-3'
GMUQ301 5'-GCGCGTCGACATATG
T
/
A
CTG
G
/
T
A
/
G
CA
T
/
G
G- A
A
/
G
ATCTGGGAAG-3'
GMUQ302 5'-GCGTCTAGATGA
A
/
C
GCAAA
T
/
G
CG
T
/
C
ACCAT-
T
/
C
A
A
/
G
A
/
C
A-3'
GMUQ303 5'-GCGCGTCGACATATGTTTGC
A
/
C
AA
G
/
A
A
/
C
G
/
A-
GAT
T
/
C
CC
T
/
C
GGC-3'
GMUQ304 5'-GCGTCTAGAAGC
T
/
C
TT
T
/
A
A
A
/
G
CAG
A
/
G
T
/
G
T-
G
/
C
ACCTG-3'
GMUQ307 5'-GCGCGTCGACATATGGCTGGTGCT
A
/
C
G
/AG- AT
T
/
C
CC
T
/
C
GGCTTC-3'
GMUQ308 5'-GCGCGTCGACATATGGCTGATGCT
A
/
C
G
/
A
G- AT
T
/
C
CC
T
/
C
GGCTTC-3'
GMUQ327 5'-GCGCGTCGACATATGTTTGCCGCACACATC
GMUQ328 5'-GCGCGTCGACATATGTTTGCCAAACACATCCCCGGCGCCCGTGACC
GMUQ329 5'-GCGCGTCGACATATGTTTGCCAAACACATCGCCGGCGCCCGTGACC
All GAL; VP16 and GALVP16-Rev-ebA[alpha] and GAL-N-CoR constructs were sequenced to confirm the
reading frame using a Pharmacia sequencing kit. Amplification using these
primers was with Ultima DNA polymerase (Perkin Elmer) or Pfu DNA polymerase
(Stratagene) according to the manufacturer's protocols.
The expression plasmids pGALO (
22
), pNLVP16 (
23
), GAL4/VP16 (
12
), pGEX-cTR[alpha] (
24
) and pG5E1b-CAT (
25
) have been described elsewhere. pGAL0 contains the GAL4 DBD and pGAL4/VP16
contains the GAL4 DBD linked to the acidic activation domain of VP16.
The construct VP16-mRXR[gamma] was amplified using primers GMUQ173 and GMUQ176 to amplify the
coding region from pSG5-RXR[gamma]. The resulting fragment was cleaved with
Eco
RI, end filled with Klenow then cloned into
Nde
I-digested, Klenow end- filled pNLVP16.
The construct VP16-cTR[alpha] was created by digestion of pGEX-cTR[alpha] with
Eco
RI. The resulting fragment was end filled with Klenow and ligated into
Nde
I-cleaved, Klenow end-filled pNLVP16.
GAL-NCoR (ID) was created by amplifying the NCoR interaction domain from a
C2C12 muscle cell cDNA library in [lambda]ZAP II using the primers GMUQ296 and GMUQ297. The resulting product was
cleaved with
Eco
RI and ligated with
Eco
RI-cleaved pGAL0.
pGAL4/VP16-Rev-erbA[alpha] (GV-Rev) and VP16-Rev-erbA[alpha] (VP16-Rev) chimeras were created by
inserting fragments of Rev-erbA[alpha] into the pGAL4/VP16 and pNLVP16 vectors respectively. To construct
VP16-Rev aa 107-614, the primers GMUQ171 and GMUQ172 were used to amplify a region
from pBS-rRev-erbA[alpha]. The resulting fragment was cleaved with
Nde
I and ligated with
Nde
I-digested pNLVP16. GV-Rev aa 437-614 and GV-Rev aa 557-614 were constructed by cleaving Rev-erb cDNA with
Bgl
II-
Bam
HI and
Eco
RI-
Bam
HI respectively. The resulting fragments were end filled with Klenow and ligated
into
Nde
I-cleaved, Klenow end-filled GAL4/VP16.
VP16-Rev aa 437-614 was constructed by cleaving Rev-erb cDNA with
Bgl
II-
Bam
HI end filling the resulting fragment with Klenow and ligating to
Xho
I-cleaved, Klenow end-filled pNLVP16. To construct VP16 Rev aa 509-614, GAL-Rev aa 21-614 (
12
) was digested with
Sac
I, end filled with T4 DNA polymerase and ligated into
Xho
I-cleaved, Klenow end-filled pNLVP16. To construct GV-Rev aa 509-614, VP16 Rev aa 509-614 was digested with
Sal
I/
Xba
I, the resulting fragment was then ligated into
Sal
I/
Xba
I-cleaved pGAL4/VP16. To construct VP16 Rev aa 21-614, the primers GMUQ142 and GMUQ132 were used to amplify a region
from 63 to 1847 bp. The resulting fragment was cleaved with
Bam
HI, end filled with Klenow and ligated into
Nde
I-cleaved, Klenow end-filled pNLVP16. To construct VP16 Rev aa 290-614, Rev-erbA[alpha] cDNA was digested with
Eco
RV/
Bam
HI, end filled with Klenow and ligated to
Xho
I-cleaved, Klenow end-filled pNLVP16.
To construct VP-Rev aa 437-509 and GV-Rev aa 437-509, VP16-Rev aa 437-614 was cleaved with
Sac
I, and ligated into
Sac
I-cleaved pGAL4/VP16. The resulting vector was cleaved with
Sal
I/
Xba
I and the Rev-erbA[alpha] fragment was cloned into
Sal
I/
Xba
I-digested pNLVP16 respectively.
For construction of the following GV-Rev chimeras, the following primers were used to amplify regions of Rev-erbA[alpha] from GV-Rev aa 437-614; GV-Rev aa 437-476 (GMUQ301 and GMUQ304); GV-Rev aa 455-476 (GMUQ303 and
GMUQ304) and GV-Rev aa 455-488 (GMUQ303 and GMUQ302). The following primers were used to
create mutations in the FAKH and L3-4 regions of GV-Rev aa 455-488; GV-Rev AGAR (GMUQ307 and GMUQ302); GV-Rev ADAQ (GMUQ308 and GMUQ302); (GV-Rev FAA (GMUQ 327 and GMUQ 302); GV-Rev L3-4 A (GMUQ 328 and GMUQ 302)
and GV-Rev L3-4 AA (GMUQ 329 and GMUQ 302). These fragments containing primer-derived 5'
Sal
I and 3'
Xba
I sites were digested with
Sal
I/
Xba
I and ligated to
Sal
I/
Xba
I-digested pGAL4/VP16. VP16 Rev aa 455-488 was constructed by amplifying this region of Rev-erbA[alpha] from GV-Rev aa 437-614 using the primers GMUQ303 and GMUQ302, digesting the
product with
Sal
I/
Xba
I and ligating to
Sal
I/
Xba
I-digested pNLVP16.
Double-stranded sequencing of ligation junctions confirmed authenticity and that
the foreign proteins were being expressed in frame.
COS-1 or JEG-3 cells were cultured for 24 h in Dulbecco's Modified Eagle's Medium
(DMEM) containing 5% (v/v) fetal calf serum (FCS) before transfection. Each 35
mm dish (Falcon) of COS-1 cells (60-70% confluence) was transiently transfected with 2.5 [mu]g of reporter plasmid DNA (G5E1b-CAT) expressing chloramphenicol acetyl transferase (CAT),
mixed with the appropriate amount of expression vector (1 [mu]g GAL4/VP16 chimeras) by the DOTAP (Boehringer Mannheim) mediated procedure
as described previously (
24
,
26
). The DNA/DOTAP mixture was added to the cells in 3 ml of fresh medium. After a
period of 24 h, fresh medium was added to the cells. Cells were harvested for
the assay of CAT enzyme activity 24-72 h after the transfection period. Each transfection was performed at
least three times to overcome the variability inherent in transfections.
Mouse myogenic C2C12 cells (
27
,
28
) were grown in DMEM supplemented with 20% (v/v) FCS in 6% CO
2
. Prior to and during transfection, this cell line was induced to biochemically
and morphologically differentiate into multinucleated myotubes by serum
withdrawal medium [DMEM supplemented with 2% (v/v) FCS]. Each 35 mm dish of
myogenic C2C12 cells (90-100% confluence) was transiently transfected as described above. After 24
h, fresh medium was added to the cells. Cells were harvested for the assay of CAT enzyme activity 24 h after the addition of fresh medium.
Each transfection experiment was performed at least three times to overcome the
variability inherent in transfections.
Mammalian two hybrid assay
Plasmids ([G5E1bCAT reporter, 3 [mu]g) and GAL NCoR (ID: interaction domain: aa 2218-2446) were co-transfected/expressed into JEG-3 cells cultured for 24 h in 3 ml DMEM containing 5%
charcoal stripped FCS, in 6% CO
2
at 37oC prior to transfection (using DOTAP as described previously) with one of
the following vectors; VP16, VP16-cTR[alpha], VP16-mRXR[gamma], VP16-Rev aa 21-614, VP16-Rev aa 107-614, VP16-Rev aa 290-614, VP16-Rev aa
437-614, VP16-Rev aa 509-614 or VP16-Rev aa 455-488. The DNA/DOTAP mixture was added to the
cells in 3 ml of fresh medium. After a period of 24 h, fresh medium was added.
Cells were harvested after a further 24 h and assayed for the ability to
trans
-activate the reporter (GE1bCAT). Each transfection was performed at least
three times in order to overcome the variability inherent in transfections.
Cells were harvested, normalised to protein concentration (
29
) and CAT activity measured as previously described (
30
). Aliquots of the cell extracts were incubated at 37oC, with 0.1-0.4 mCi [
14
C]chloramphenicol (ICN) in the presence of 5 mM Acetyl CoA in 0.25 M Tris-HCl pH 7.8. After a 1-4 h incubation period, the reaction was stopped by the addition of
1 ml ethyl acetate which was used to extract the chloramphenicol and its
acetylated forms. Extracted materials were analysed on Silica gel thin layer
chromatography plates as described previously (
30
). Quantitation of all CAT assays (except Fig.
2
D) was performed by an AMBIS [beta]-scanner. Please note that Figure
2
D was quantitated on a BioRad Molecular Image System, GS-363, which has increased sensitivity. This helped to discriminate between
the subtle effects of different point mutations.
Cells were washed in PBS, scraped into 1% SDS and protein content was determined
using the BCA Protein Assay (Pierce). Total cell extract (30 [mu]g) was boiled in Laemmli sample buffer and electrophoresed on 10%
polyacrylamide resolving gels before being electrophoretically transferred onto
nitrocellulose (Hybond-C Super, Amersham). Non-specific binding sites were blocked by immersing membranes in 5%
skim milk in PBS for 1-2 h at room temperature followed by incubation with Rabbit anti-GAL 4 antisera (Santa Cruz # sc-428) in (1 [mu]g/ml) 1% skim milk in PBS overnight at 4oC. Membranes were then washed three times in PBS
containing 0.05% Tween 20. Further steps involving HRP secondary antibodies and
the catalysed oxidation of luminol were carried out with Amersham ECL Western
blotting detection reagents according to the manufacturer's protocols.
Previously, we have mapped an active transcriptional silencing function of Rev-erbA[alpha] to the C-terminal putative ligand binding domain (LBD) or D/E region
between aa 290 and 614 (
12
). To characterise further the transcriptional silencing properties of Rev-erbA[alpha] and to delimit the repression domain, we investigated the ability
of this orphan receptor to repress
trans
-activation by the GAL4/VP16 chimera, a potent activator of gene
expression. We sub-cloned segments of the Rev-erbA[alpha] cDNA into the GAL4/VP16 expression vector and examined the
activation of the CAT reporter gene linked to GAL4 binding sites. A similar
investigative approach utilising the GAL4/VP16 chimera has been utilised to
analyse the thyroid hormone receptor (
31
) and Rev-erbA[alpha] (
12
). The GAL4/VP16 protein, which contains the yeast GAL4 DBD and
trans
-activating domain of the herpes simplex virus, VP16, is a potent
transcriptional activator of the GAL4 DNA binding sites linked to CAT.
A total of four chimeric GAL4/VP16-Rev-erbA[alpha] (GV-Rev) receptor expression plasmids were constructed; GV-Rev aa 437-614, GV-Rev aa 509-614, GV-Rev aa 557-614, GV-Rev aa 437-509. The previously described GV-Rev aa 290-614 (
12
) was used as a comparison in these experiments. These constructs were co-transfected with the reporter, G5E1b-CAT, into COS-1 cells and the CAT activity assayed.
Efficient transcriptional repression (~60-80-fold) was observed when the regions of Rev-erbA[alpha] aa 437-614 or 437-509 were fused to the GAL4/VP16
chimera (Fig.
1
A). In contrast, the aa 509-614 and 557-614 of Rev-erbA[alpha] had no effect on the ability of the GAL4/VP16
protein to
trans
-activate gene expression (Fig.
1
A). The aa 290-614 portion of Rev-erbA[alpha] linked to GAL4/VP16 as published previously (
12
), also repressed (~40-60-fold) the transcription of GAL4/VP16. In summary, this
deletion analysis indicated that repression was mediated by a short region
between aa 437 and 509 in Rev-erbA[alpha].
Recent publication of the crystal structures for the LBD of three members of the
steroid/thyroid receptor superfamily, TR, RAR and RXR revealed a conserved
canonical structure in the LBD, consisting of 12 [alpha]-helices (
3
-
34
). The availability of this structural data prompted us to align Rev-erbA[alpha] with TR, RAR and RXR. The smallest identified repression domain of
the Rev-erbA[alpha] (aa 438-509) encompasses a region that forms [alpha]-helices 3 (H3), Loop3-4 (L3-4), 4 (H4), 5 (H5) and the [beta]-sheet formation in
TR, RAR and RXR (see Fig.
2
A). This region includes the `LBD-specific signature motif', [(F/W)AKXXXXFXX- LXXXDQXXLL] that is highly conserved among the nuclear hormone receptors,
and lies within the E-region of most steroid receptors spanning H3, L3-4 and H4 of the LBD (
35
). It has been proposed that this motif contributes to stabilisation the
canonical structure of the LBD (
35
). Interestingly, this region is found in the Rev-erbA[alpha] repression domain. Based on this information, we decided to
investigate the role of H3, L3-4, H4 (that span the LBD-specific signature motif), H5 and the [beta]-sheet in repression of transcription by Rev-erbA[alpha].
We constructed three chimeric GV-Rev receptor expression plasmids; GV-Rev aa 455-488 (FAK, L3-4, H4 and H5), GV-Rev aa 437-476 (H3, L3-4 and H4) and GV-Rev aa 455-476 (FAK, L3-4 and H4), the
previously described GV-Rev aa 437-509 (H3 to [beta]-sheet) was used as a control in this experiment. These
were co-transfected with the reporter G5E1b-CAT, and the CAT activity assayed (Fig.
2
B).
The region of Rev-erbA[alpha] between aa 455 and 488 (FAK, L3-4, H4 and H5) very efficiently repressed (~40-fold) GAL4/VP16 mediated
trans
-activation (Fig.
2
B). This suggested that most of helix 3 and the [beta]-sheet were not required for repression of transcription by Rev-erbA[alpha]. In contrast, the aa 437-476 (H3, L3-4 and H4) and 455-476 (FAK, L3-4 and H4) regions of Rev-erbA[alpha] had no effect
on the ability of the GAL4/VP16 protein to
trans
-activate gene expression (Fig.
2
B). This strongly indicated that helix 5 was absolutely necessary for efficient
repression. The aa 437-509 (H3 to [beta]-sheet) portion of Rev-erbA[alpha] linked to GAL4/VP16 as earlier described, also
repressed (~40-fold) the transcription of GAL4/VP16.
These fine deletion studies suggested that the LBD-specific signature motif (that spans FAK, L3-4 and H4) and H5 mediated active transcriptional silencing. The aa
of H3, N-terminal of the FAK residues and the [beta]-sheet region are not necessary for repression. Deletion of H5
abolished repression, suggesting it was absolutely required for repression:
furthermore, the data indicated that the LBD-specific signature motif (FAK, L3-4 and H4), in the absence of H5, cannot mediate active
transcriptional silencing.
The LBD-specific signature motif [
We constructed five new chimeric GV-Rev expression plasmids containing mutations in the repression domain,
changing the FAKH sequence (in H3) to
Complete mutation of the FAKH motif to either AGAR or ADAQ inhibited the ability
of Rev-erbA[alpha] to repress transcriptional activation by GAL4/VP16 (see GV-Rev AGAR and GV Rev-ADAQ; Fig.
2
C and D). Interestingly, the impact of the ADAQ versus AGAR mutation on
repression is more significant, probably due to the mutation of a neutral
alanine (A) to a charged aspartic acid (D). Furthermore, the specific mutation
of the lysine (K457) to an alanine (A) in the FAK motif significantly reduced
the ability of Rev-erbA[alpha] to repress
trans
-activation by GALVP16 (see GV-Rev FAA; Fig.
2
C and D). Specific mutation of the invariant phenylalanine (F462) in loop 3-4 or the simultaneous mutation of the invariant F and the highly
conserved proline (P460) in loop 3-4 significantly reduced the ability of Rev-erbA[alpha] to repress
trans
-activation by GALVP16 (see GV-Rev L3-4 A and GV-Rev L3-4 AA; Fig.
2
C and D).
These results demonstrated that the FAK residues and the invariant phenylalanine
(F) in L3-4 of Rev-erbA[alpha], are essential for efficient repression of
trans
-activation by the GAL4/VP16 protein. Combined with our previous
observations, we can conclude that the conserved residues in the LBD-specific signature motif (spanning FAK, L3-4 and H4) and helix 5 between aa 455 and 488 are required for
repression of transcription by Rev-erbA[alpha].
We then investigated whether co-factors involved in repression by Rev-erbA[alpha] were cell-specific. We examined whether GV-Rev aa 437-614 and GV-Rev aa 455-488, both of which exhibit
strong repression of transcription in COS-1 cells, were able to repress transcription efficiently in C2 (mouse
myogenic) and JEG-3 (choriocarcinoma) cells. In both cell lines, we observed strong
repression by both the GV-Rev aa 437-614 or the GV-Rev aa 455-488 constructs (~40-60-fold) (Fig.
3
A), suggesting that different cell types possess co-factors that interact with Rev-erbA[alpha] to mediate repression.
The molecular mechanisms responsible for nuclear receptor transcriptional
silencing are not well understood. However, the nuclear receptor co-repressor, N-CoR, a 270 kDa protein, has been characterised as a co-repressor for the steroid receptors, TR and RAR (
19
,
20
). N-CoR mediates ligand-independent inhibition of gene transcription by direct binding in
the unliganded state and regulation of AF-2 function. In an effort to understand the mechanism of repression by Rev-erbA[alpha], we used the mammalian two hybrid system to investigate the
possible interaction of Rev-erbA[alpha] and N-CoR. The two hybrid system was originally developed in yeast
to investigate protein-protein interactions and has since been adapted for use in mammalian
cells. A chimeric receptor consisting of the yeast GAL4 DNA binding domain
fused to the characterised interaction domain (ID) (aa 2218-2451) of N-CoR (
20
) was expressed in cells with a second chimeric receptor containing full length
or various deletions of the Rev-erbA[alpha] receptor linked to the
trans
-activation domain of VP16. In these assays,
trans
-activation of a CAT reporter gene downstream of GAL4 binding sites fused
to the E1b promoter is only achieved when the co-expressed receptors physically interact. In this study, we used the
chimeric VP16 constructs containing TR[alpha] and RXR[gamma] that have been shown to interact strongly and very weakly
respectively, with N-CoR as positive and negative controls respectively.
Consistent with previous interaction studies, we saw a 15-20-fold increase in CAT activity when the VP16-cTR[alpha] construct was transfected with the GAL-N-CoR (ID) construct (
19
,
20
). In the presence of T
3
, the ligand for TR, this activity, as expected, was reduced to a level
equivalent to that seen with VP16-mRXR[gamma] (3-4-fold) indicating very weak interaction of TR with N-CoR in the presence of it cognate ligand (Fig.
3
B). Little or no CAT activity was observed when any of the VP16-Rev-erbA[alpha] (full length or delimited) constructs were co-transfected with GAL-N-CoR (ID) suggesting inefficient or no
interaction of the Rev-erbA[alpha] receptor with the characterized interaction domain (aa 2218-2451) of N-CoR in intact cells.
Transcriptional activity of Rev-erbA[alpha] is modified by agents that stimulate phosphorylation (
12
). Therefore, we also tested whether the phosphorylation of Rev-erbA[alpha] could induce the interaction of Rev-erbA[alpha] with N-CoR. In the presence of 8-Br-cAMP no significant interaction
between Rev-erbA[alpha] and N-CoR was observed in intact cells (data not shown).
These results suggest that Rev-erbA[alpha] does not physically contact/associate with the interaction domain
of N-CoR (aa 2218-2451) and that transcriptional repression by Rev-erbA[alpha] involves a mechanism that functions independently of
N-CoR.
This work reports on the characterisation of the repression domain in the Rev-erbA[alpha] receptor. We have shown that the Rev-erbA[alpha] receptor contains a potent transcriptional silencing
domain in the C-terminal region of the protein that abrogates the ability of the potent
trans
-activator, GAL4/VP16, to activate gene expression. Deletion analyses
narrowed the repression domain to a region encompassing aa 437-509 of the Rev-erbA[alpha] protein. These results are contrary, in some aspects, to
those obtained by Harding and Lazar which indicated that Rev-erbA[alpha] contains two repression domains located between aa 200 and 289 and
aa 376 and 614 that are dependent on each other for maximal repression (
14
). Furthermore, that report indicated that the region between aa 432 and 614 of
Rev-erbA[alpha] exhibited no transcriptional repression; this is not in agreement
with the results obtained from our study (aa 437-614 showed strong repression of GAL4/VP16 transcriptional activity ~60-80-fold). Differences observed in the two studies may be
accounted for by the choice of assay system used, GAL4 rather than GAL4/VP16,
and the reporter system employed, pGL2-luciferase (SV40 minimal promoter) versus G5E1b-CAT (E1b TATA box). It is unlikely however, that these differences
could account for the marked discrepancies between the two studies.
We have defined this silencing domain further in the context of the recently
published LBD consensus structure for steroid receptors (
35
). The publication of crystal structures for the LBD of TR, RAR and RXR have
revealed that the LBD forms a conserved canonical structure consisting of 12 [alpha]-helices. The repression domain identified in Rev-erbA[alpha] encompassed a region that would putatively form
helices 3-5 (H3-H5) and includes a region which contains [beta]-sheet formation. This region also contains the highly
conserved LBD-signature motif (F/W) AKXXXXFXXLXXXDQXXLL, that is present in most (>80) steroid receptors (
35
). It has been proposed that these conserved residues hold together helices H3,
H4, H5, H8 and H9, and the loops to form a hydrophobic pocket, the LBD core.
The structural analysis of TR/RAR/RXR indicates that this region is buried
inside the receptor. Whether this is the case with Rev-erbA[alpha] which does not contain helix 12 (AF-2 domain) remains to be determined. Furthermore, the
predominantly hydrophobic amino acids in this region cluster around an
invariant phenylalanine in the L3-4 sequence. The conserved aromatic amino acid, phenylalanine, and the
conserved lysine in the FAK motif have been shown to set the spatial limits of
this hydrophobic cluster and to interact with the glutamine residue in H4
respectively (aa K457 and Q470 in Rev-erbA[alpha]) (Fig.
2
A) (
35
).
Fine deletion analysis demonstrated that the LBD-signature motif (i.e. FAK, L3-4 and H4) and H5 are sufficient and necessary to mediate
transcriptional repression by Rev-erbA[alpha], delimiting the minimal region required for active transcriptional
silencing of gene transcription to a 34 amino acid region between aa 455 and
488, located in the E-region of the Rev-erbA[alpha] receptor. This domain is conserved within the Rev-erb family with only one aa difference found in the
RVR/Rev-erb[beta] receptor (aa 416-449). To date, the reported transcriptional repression
properties of both the Rev-erbA[alpha] and RVR have been identical (
9
,
10
,
15
).
Site-specific mutagenesis demonstrated and verified the importance of the FAK
motif, the conserved lysine (K457) in the FAK motif and the invariant
phenylalanine (F462) in the L3-4 sequence (HIPGFRDLS) of Rev-erbA[alpha]. Interestingly, mutation of the FAK motif, to ADA versus
AGA had a more significant impact on the ability of this region to repress
transcription, suggesting that a charged amino acid (e.g. D) disrupts the
structure of this domain.
We also demonstrated that the repression domain of Rev-erbA[alpha] is functional in myogenic cells (mouse C2C12 muscle cells) and
choriocarcinoma cells adding to list of cell types (JEG-3, HepG2, NIH3T3 and 293T cells) in which transcriptional repression by
Rev-erbA[alpha] has been demonstrated (
14
). This data supports the hypothesis that co-factors involved in repression by Rev-erbA[alpha] are not cell-specific. Furthermore, we have shown that Rev-erbA[alpha] acts as a transcriptional repressor in a
cell type where it has been demonstrated to function in an antagonistic manner
(
12
).
Interestingly, both Rev-erbA[alpha] (
14
) and RVR (
15
) have been demonstrated to repress basal transcription when bound, as a
homodimer, to their cognate DR-2 element, but not however, when bound as a monomer to the characterised
asymmetric 11 bp motif. It is possible, therefore, that two copies of the
repression domain are required for function
in vivo
. Consistent with this hypothesis is the knowledge that GAL4 DBD binds to DNA as
a dimer (
36
) and therefore our chimeric constructs would presumably also bind DNA as
dimers. The repression domain we have identified as discussed earlier, contains
the LBD-signature motif which appears to be necessary for the function of the
repression domain. This motif has also been characterised as important in
protein dimerisation of steroid receptors and similar mutations to the ones
performed in this study (mutation of the aa FAK) are known to disrupt
dimerisation of other steroid receptors (
37
-
39
).
In this report a potential mechanism of repression by Rev-erbA[alpha] was also investigated utilising the mammalian two hybrid system to
study whether the recently characterised nuclear receptor co-repressor, N-CoR, interacted with the Rev-erbA[alpha] receptor. We found that, in contrast with TR which
interacted strongly with aa 2218-2451 of N-CoR in a ligand-dependent manner (
19
), Rev-erbA[alpha] did not strongly associate with the characterised interaction
domain of N-CoR (aa 2218-2451). Interaction between Rev-erbA[alpha] and N-CoR was not observed after the addition of a PKA
stimulating agent (8-Br-cAMP). These results suggest that Rev-erbA[alpha] does not mediate repression through an interaction
with N-CoR. However, we cannot rigorously rule out that other regions/domains of
N-CoR may interact with this orphan receptor. This does not discount the
possibility that novel co-repressors may mediate active transcriptional silencing by the Rev-erbA[alpha] receptor.
Whether the domain we have defined between aa 455 and 488 of Rev-erbA[alpha] directly represses transcription or functions as a repression
interface for another molecule is not clear. Another possible mechanism by
which Rev-erbA[alpha] could be mediating active transcriptional silencing of gene
transcription is by interaction with the basal transcription factors such as
TATA-binding protein or TFIIB, both of which have been demonstrated to interact with other nuclear steroid receptors in a ligand-dependent manner (
40
-
42
). It has been suggested that the interaction of these basal transcription
factors with steroid receptors may be, at least partially, responsible for
repression of basal transcription by steroid receptors (
40
-
42
). We are currently investigating whether the Rev-erbA[alpha] receptor can interact with any of these basal transcription
factors.
In conclusion, we have characterised an independent, dominant, transferable
transcriptional silencer in the E region of the orphan steroid receptor, Rev-erbA[alpha], that mediates transcriptional repression through an, as yet,
unknown mechanism.
This investigation was supported by the National Health and Medical Research
Council (NHMRC) of Australia. The Centre for Molecular and Cellular Biology is
the recipient of an Australian Research Council (ARC) special research grant.
We thank Dr W. Chin for the rRev-erbA[alpha] cDNA clone. We also wish to thank Dennis H. Dowhan and Dr Amanda
Carozzi for excellent technical assistance and helpful discussions.
REFERENCES
Return