(Received for publication, April 24, 1997)
From the Departments of Medicine (Cardiology) and
§ Cell Biology, Albert Einstein College of Medicine,
Bronx, New York 10461
Terminal differentiation is characterized by cell
cycle arrest and the expression of cell type-specific genes. Previous
work has suggested that the p300 family of transcriptional coactivators plays an important role in preventing the re-initiation of DNA synthesis in terminally differentiated cardiac myocytes. In this study,
we investigated whether p300 proteins are also involved in the
transcriptional activation of cell type-specific genes in these cells.
Since p300 function can be abrogated through direct binding by the
adenovirus E1A protein, we overexpressed E1A in cardiac myocytes using
recombinant adenoviral vectors. The expression of transfected reporter
genes driven by - or
-myosin heavy chain promoters was markedly
diminished by expression of the 12 S E1A protein. In contrast, the
activity of a promoter derived from the ubiquitously expressed
-actin gene was affected only modestly. While an E1A mutant unable
to bind members of the retinoblastoma family of pocket proteins
decreased the activity of
- and
-myosin heavy chain promoters to
nearly the same extent as wild type 12 S E1A, transcriptional
repression by a mutant defective for p300 binding was severely
impaired. Furthermore, overexpression of p300 and, to an even greater
extent, p300del33, a mutant lacking residues required for binding by
E1A, relieved E1A's repression of
-myosin heavy chain promoter
activity while having no effect on the activity of the
-actin
promoter. Thus, E1A's transcriptional repression of cell type-specific
genes in cardiac myocytes is mediated through its binding of p300
proteins, and these proteins appear to be involved in maintaining both
cell type-specific gene expression and cell cycle arrest in cardiac
myocytes.
Cardiac myocytes are striated muscle cells whose structural and functional properties are especially suited to maintain normal blood flow. These cells differentiate from progenitors in the lateral plate mesoderm (Ref. 1; reviewed in Ref. 2) ~7.5-8.0 days post-coitum in the mouse. Differentiation is marked by the activation of muscle-specific genes, formation of sarcomeres, and onset of rhythmic contractions. In contrast to skeletal myogenesis, differentiation and proliferation in the cardiac myocyte lineage are not mutually exclusive; at least a subset of differentiated cardiac myocytes continues to undergo cell division (3). Beginning in mid-gestation, however, increasing proportions of these cells arrest irreversibly in G1 of the cell cycle. DNA synthesis is undetectable in most cardiac myocytes by the end of fetal life and occurs rarely, if at all, after the third postnatal week in the rat (4, 5). Although cardiac myocytes retain considerable plasticity to undergo further modifications in gene expression, morphology, and function (reviewed in Ref. 6), their identity as post-mitotic heart muscle cells is fixed following terminal differentiation. The molecular mechanisms that induce pre-cardiac mesoderm to differentiate into cardiac myocytes and subsequently maintain these cells in a differentiated state are poorly understood.
Clues about cellular proteins that regulate aspects of striated muscle differentiation have been provided by the effects of the adenovirus E1A oncoprotein in these cells. E1A blocks the differentiation of skeletal myocytes from their myoblast precursors (7-12). In addition, when expressed in already differentiated skeletal (13) or cardiac (14, 15) muscle cells, it stimulates cellular DNA synthesis and represses muscle-specific gene expression. E1A interacts with several cellular proteins including members of the p300 family of transcriptional coactivators and the retinoblastoma (Rb) family of pocket proteins (reviewed in Ref. 16). E1A mutants defective for binding these proteins are less efficient than wild type at stimulating DNA synthesis and repressing muscle-specific gene expression in cardiac myocytes (14, 15). These experiments suggest that E1A's effects on cardiac myocyte differentiation are mediated through its interactions with these cellular proteins.
Although analyses of E1A mutants offer a starting point from which to understand the roles of cellular proteins in cardiac myocyte differentiation, these studies are limited in two respects. First, they provide only correlative information. Second, the multifunctional nature of the E1A protein makes it difficult to attribute a given effect to a specific cellular protein. Therefore, the goal of the current study was to use a more direct approach to determine whether the p300 family of proteins is involved in maintaining the cardiac myocyte phenotype. We chose to focus on p300 because, in our previous work, the E1A mutant unable to bind this protein was most defective in its ability to stimulate DNA synthesis in cardiac myocytes (15). The p300 family of transcriptional coactivators have been noted to modulate many examples of enhancer-mediated transcription (17-35). The members of this family, which, thus far, include p300 and the CREB-binding protein (CBP),1 interact directly with components of the basal transcriptional apparatus (e.g. TFIIB (19, 26) and TBP (26, 36)) and diverse enhancer-binding proteins (reviewed in Ref. 26). In skeletal muscle, MyoD and MEF-2 are included among the latter (26-29). Currently, nothing is known about which transcription factors interact with and transactivate through p300 proteins in cardiac myocytes. In addition to their possible "bridging function," p300 proteins possess intrinsic histone acetyltransferase activity (37, 38), which is hypothesized to promote a locally "open" and transcriptionally active chromatin configuration.
We demonstrate here that E1A differentially represses the expression of cell type-specific genes in cardiac myocytes and E1A's ability to bind p300 correlates with this function. Furthermore, E1A's repression is rescued partially by overexpression of p300 and to levels exceeding those observed in the absence of E1A by a p300 mutant lacking an E1A-binding domain. In contrast, E1A's mild repression of a ubiquitously active promoter is not relieved by overexpression of this construct. These data demonstrate that E1A's transcriptional repression of cell type-specific genes in cardiac myocytes is mediated through p300 protein(s) and suggest that this family of coactivators plays a role in the transcription of these genes in the absence of E1A.
Mutant human type 5 adenoviruses lacking a
functional E3 gene were used throughout these experiments.
dl520E1B and its derivatives (Ref. 39; gift of
Dr. S. T. Bayley, McMaster University, Hamilton, Ontario) contain an
E1A gene in which the 5
splice site needed to generate the 13 S
transcript has been mutated; therefore, these viruses give rise only to
the 12 S E1A transcript encoding the 243 residue protein. In addition,
dl520E1B
and derivatives lack a functional E1B
gene. Derivatives of dl520E1B
include
dl1101/520E1B
and
dl1108/520E1B
, in which E1A residues 14-25
and 124-127 have been deleted, respectively. dl70-3 (Ref.
40; gift of Dr. F. L. Graham, McMaster University, Hamilton, Ontario)
lacks functional E1A and E1B genes. Viral stocks were grown on 293 cells, purified on CsCl gradients, and titered by plaque assay on 293 cells (41).
p-MHCluc (42),
p
-MHCluc,2 and p
-actinluc (Ref. 14;
gift of Dr. M. D. Schneider, Baylor University, Houston, TX) consist of
the firefly luciferase (luc) cDNA driven by 2936, 3542, and 433 base pairs of rat
-myosin heavy chain (MHC), rat
-MHC, or avian
cytoplasmic
-actin promoter sequences. pRSVCAT contains the
bacterial chloramphenicol acetyltransferase (CAT) gene driven by Rous
sarcoma virus long terminal repeat sequences (43). p300wt and p300del33
contain the cytomegalovirus promoter/enhancer fused to a full-length
human p300 cDNA or one with an internal deletion in the E1A-binding
domain as described previously (Ref. 17; gift of Drs. R. Eckner and D. Livingston, Harvard Medical School, Boston, MA).
Cultures of primary rat cardiac myocytes were prepared from the hearts of embryonic day 20 Sprague-Dawley fetuses (Taconic Farms) as described (15). After cells were pre-plated for 1 h to minimize fibroblast contamination (typically 5-10% following this step), myocytes were plated at a density of 700/mm2 onto 60-mm dishes (Primaria, Falcon) and cultured at 37 °C, 5% CO2 in media containing 10% (v/v) fetal bovine serum (Hyclone). Thirty-six hours after plating, transfections were performed using DEAE-dextran as described previously (44) with minor modification. Briefly, cells were incubated for 1 h in serum-free media with DEAE-dextran (500 µg/ml) and the DNAs to be transfected (see figure legends for combinations and doses) following which they were incubated for 2 min with 10% dimethyl sulfoxide and for 3 h with 100 nM chloroquine. Twenty-four hours after transfection, cells were infected with adenovirus at a multiplicity of infection (m.o.i) of 1 or 10 plaque forming units (pfu) per cell as specified in a volume of 1 ml for 60-mm plates at 37 °C, 5% CO2 for 1.5 h (45). Twenty-four hours after infection (48 h after transfection), cells were harvested and luciferase and CAT activities were determined as described (46). The relative luciferase activity was calculated from the ratio of luciferase minus background/CAT minus background and expressed as the mean ± S.E. Groups were compared using the unpaired, two-tailed t test and differences considered significant if p < 0.05.
ImmunoprecipitationExtracts from primary cultures of cardiac myocytes were immunoprecipitated using either anti-human p300 monoclonal antibody (Upstate Biotechnology) or normal mouse serum in low stringency buffer (50 mM Tris (pH 7.4), 0.15 M NaCl, 0.5% Nonidet P-40, 1 mM EDTA, 10 µg/ml aprotinin (Sigma) and leupeptin (Sigma), and 0.5 mM phenylmethylsulfonyl fluoride (Sigma)) for 16 h at 4 °C and incubated with protein A beads for 1 h at 4 °C. The precipitate was washed four times in the same buffer, resuspended in 50 µl of SDS lysis buffer (20 mM Tris (pH 7.5), 50 mM NaCl, 0.5% SDS, 1 mM dithiothreitol), heated to 95 °C for 2 min, electrophoresed by SDS-polyacrylamide gel (12%), transferred to Immobilon membranes, reacted with adenovirus E1A monoclonal antibody (Oncogene Science) which was subsequently detected using horseradish peroxidase-conjugated anti-mouse IgG. Signals were detected using the ECL Western blotting detection system (Amersham) according to the manufacture's instruction.
RNA AnalysisNorthern blot analysis of 10 µg of total RNA
was performed as described previously (47). An isoform specific
antisense deoxyoligonucleotide complementary to nucleotides 5846-5869
of the rat 3-untranslated region (48) was used to detect
-MHC
mRNA as described previously (47). To detect
-actin mRNA, a
0.4-kilobase pair cDNA probe consisting of the HinfI
fragment of human
-actin gene (49) was used.
Previously, we demonstrated that expression of E1A
is sufficient to stimulate DNA synthesis in embryonic day 20 (E20)
cardiac myocytes, most of which had already undergone cell cycle arrest (15). This occurs in essentially all successfully infected cells within
24 h of the addition of adenovirus to cultures. E1A can also
induce apoptosis in these cells; this process, however, is not evident
until at least 36 h after infection. Previous studies in a variety
of cell types, including immortal skeletal muscle lines and primary
cultures of neonatal cardiac myocytes, have shown that expression of
E1A represses the transcription of cell type-specific genes (7-14). To
determine if it has a similar effect in our system, we transfected E20
cardiac myocytes with luciferase constructs driven by promoters derived
from the - or
-myosin heavy chain (MHC) genes, whose expression
is limited to striated muscle cells, or the
-actin gene, which is
ubiquitously expressed. A CAT reporter driven by the constitutively
active RSV promoter was co-transfected to control for transfection
efficiency. Twenty-four hours after transfection, plates were infected
with one of two adenoviruses: dl520E1B
, which
expresses the wild type 12 S E1A or dl70-3, which does not.
These viruses are otherwise identical. Activities of reporter genes
were assayed 24 h following infection. As shown in Fig. 1A, activities of
- and
-MHC promoters
were 95 and 98% lower, respectively, in cardiac myocytes infected with
dl520/E1B
as compared with dl70-3
(both m.o.i. 10). In contrast, the activity of the
-actin promoter
was ~35% lower in cells infected with dl520/E1B
as compared with dl70-3
(both m.o.i. 10). Thus, E1A represses the activity of both cell
type-specific and ubiquitously active promoters in E20 cardiac
myocytes; the magnitude of repression, however, is considerably more
marked in the case of cell type-specific promoters.
To investigate further whether -MHC and
-actin promoters have
different sensitivities to E1A's transcriptional repression, we
assessed whether differences in this parameter would be even more
exaggerated when viruses were delivered at a lower m.o.i. We chose
m.o.i. 1 because the percentage of cardiac myocytes exhibiting positive
immunostaining for E1A following infection with
dl520E1B
at this m.o.i., 80-90%, is similar
to that observed at m.o.i. 10 (data not shown). At m.o.i. 1,
-MHC
promoter activity was repressed 67% while that of the
-actin
promoter was affected only minimally (Fig. 1B). Thus, the
muscle specific
-MHC promoter is more sensitive than the
ubiquitously active
-actin promoter to transcriptional repression by
E1A.
Previously, we had observed that an E1A
mutant defective for binding members of the p300 family was
considerably less efficient than wild type 12 S E1A at inducing DNA
synthesis in E20 cardiac myocytes (15). In contrast, a second E1A
mutant that was defective for pocket protein binding was as active as
wild type in this function. To determine whether the ability to bind
p300 family members also correlates with E1A's repression of cell
type-specific promoters, we compared the effects of wild type 12 S E1A
and each of the above mutants on the transcriptional activities of -
and
-MHC promoters in E20 cardiac myocytes.
To confirm that the mutants exhibited the appropriate p300 binding
properties in cardiac myocytes, lysates of cells infected with each of
the viruses were immunoprecipitated with an antibody that reacts with
both p300 and CBP. Western blots of these immunoprecipitates were then
reacted with an antibody monospecific for E1A (Fig. 2A). As expected, E1A co-immunoprecipitated
with p300 proteins in plates infected with
dl520E1B, encoding wild type 12 S E1A
(lane 2), and dl1108/520E1B
,
encoding the mutant unable to bind pocket proteins (lane 4). In contrast, E1A was not present in plates infected with
dl1101/520E1B
, encoding the E1A mutant
defective for p300/CBP binding (lane 3). E1A was also absent
from plates infected with dl70-3, which does not produce an
E1A protein (lane 1), and from plates infected with
dl520E1B
when normal mouse serum was
substituted for anti-p300 in the immunoprecipitation (lane
5). Thus, dl1101/520E1B
, but not
dl520E1B
or
dl1108/520E1B
, is defective for binding p300
family proteins in cardiac myocytes. Immunoblots were also performed on
cellular lysates (without prior p300 immunoprecipitation) to confirm
that similar E1A levels resulted from infection with each of the
viruses (lanes 6-9). Since the infection efficiencies of
each of the viruses in cardiac myocytes are also similar (data not
shown), it follows that the levels of E1A per myocyte are similar.
E1A's repression of muscle-specific genes
correlates with its ability to bind p300. Panel A, ability
of E1A mutants to bind p300 proteins and relative levels of E1A mutants
following infection with adenoviruses expressing each. Lanes
1-5, cardiac myocytes infected with dl70-3 (lane
1), dl520E1B (lanes 2 and
5), dl1101/520E1B
(lane
3), or dl1108/520E1B
(lane 4)
were immunoprecipitated with either normal mouse serum (NMS, lane
5) or an antibody that recognizes both p300 and CBP (lanes
1-4). Immunoprecipitates were separated by SDS-PAGE, transferred to polyvinylidine difluoride membranes, and probed with anti-E1A antibody. Lanes 6-10, lysates of cardiac myocytes infected
with dl70-3 (lane 6),
dl520E1B
(lane 7),
dl1101/520E1B
(lane 8), or
dl1108/520E1B
(lane 9) were
immunoblotted with anti-E1A antibody. Panel B, effect of E1A
mutants on the activities of
- and
-MHC promoters in cardiac
myocytes. This experiment was performed and the data expressed as
described in the legend to Fig. 1 with all adenoviruses infected at
m.o.i 10 pfu/cell. The data presented in panel B show the
combined results from three independent preparations of cells each
performed in duplicate. Panel C, Northern analysis of the effect of E1A mutants on the expression of
endogenous
-MHC and
-actin genes. RNA was harvested 24 h
after infection with dl70-3 (lane 1),
dl520E1B
(lane 2),
dl1101/520E1B
(lane 3), and
dl1108/520E1B
(lane 4) at m.o.i 10 pfu/cell. Rows labeled
-MHC and
-actin show representative
autoradiograms following sequential hybridizations of the same blot
with radiolabeled probes specific for
-MHC and
-actin
transcripts. The 28 S rRNA bands on the ethidium bromide stained gel
are shown to indicate the equivalency of loading.
Fig. 2B shows the effect of the various E1A mutants on the
transcriptional activities of transfected - and
-MHC promoters. The most striking finding is that
dl1101/520E1B
, encoding the E1A mutant
defective for binding p300 family members, is severely disabled in
repressing the activities of either promoter. In contrast,
dl1108/520E1B
, encoding the E1A mutant
defective for pocket protein binding, represses the activities of these
promoters almost as efficiently as wild type 12 S E1A. Thus, E1A's
trans-repression function correlates with its ability to bind p300
family members.
We also examined the effects of wild type and mutant forms of E1A on
the expression of the endogenous cardiac -MHC and
-actin genes.
As shown in Fig. 2C, the expression of the cell
type-specific
-MHC gene was markedly reduced in cardiac myocytes
infected with dl520E1B
(lane 2) or
dl1108/520E1B
(lane 4) as compared
with cells infected with dl70-3 (lane 1). Infection with dl1101/520E1B
(lane
3), on the other hand, affected the expression of the
-MHC gene
minimally. In contrast to the
-MHC gene, the expression of the
ubiquitously active
-actin gene was affected only moderately by the
adenoviruses. Thus, E1A-induced changes in the expression of the
endogenous
-MHC and
-actin genes closely parallel alterations in
the activities of their respective promoters.
To test whether a causal connection exists between E1A's
repression of cell type-specific promoters and its binding to p300 proteins, we assessed the effect of overexpression of p300 on this
transcriptional repression. To augment p300 levels in the cell, we
overexpressed either full-length p300 (p300wt) or a p300 mutant that
lacks residues required for E1A binding and can, therefore, avoid
sequestration by E1A (p300del33 in Ref. 17). While both p300 and
p300del33 diminish E1A's repression of -MHC promoter activity, the
effect is more marked with p300del33 (Fig. 3). Relief of
repression is incomplete when viruses are infected at m.o.i. 10; at
m.o.i. 1, however, p300 effects nearly complete rescue and p300del33
stimulates
-MHC promoter activity to levels exceeding control.
Increases in
-MHC promoter activity due to p300del33 occur in a
dose-dependent manner. In contrast to its effects on
-MHC promoter activity, p300del33 has no effect on E1A's mild repression of the
-actin promoter. Thus, exogenous p300 can bypass E1A's transcriptional repression of cell type-specific promoters and
this effect is independent of E1A binding. This result suggests that
E1A's binding of p300 family member(s) is causally related to its
transcriptional repression of the
-MHC promoter and that p300
proteins positively modulate the activity of this promoter.
The major conclusion of this study is that E1A's transcriptional repression of cell type-specific genes in cardiac myocytes is mediated by its binding of p300 and/or related proteins. Previous work has demonstrated a correlation between the ability of E1A to bind p300 and to induce DNA synthesis efficiently in cardiac myocytes (14, 15). Taken together with the present study, these data support the hypothesis that member(s) of the p300 family play important role(s) in maintaining terminal differentiation in cardiac myocytes.
It is possible to envisage at least two models by which the interaction
between E1A and p300 might result in this effect. The most likely
model, in light of p300's recognized transcriptional coactivator
function, is one in which E1A sequesters p300 (and/or a p300-related
protein) making it unavailable to participate in the transcriptional
activation of cell type-specific genes. Strong support for the first
model is provided by the ability of p300 constructs to relieve E1A's
repression of the -MHC promoter and, in particular, by the ability
of p300del33 to effect this rescue independently of binding E1A.
Another non-mutually exclusive model is that complexes consisting of
E1A and p300 protein(s) actively participate in transcriptional
repression. Our functional data neither support nor refute this model;
additional biochemical experiments would be necessary to address this
possibility.
A corollary of the first model is that member(s) of the p300 family
participate in transcriptional activation of the - and
-MHC, and
presumably other cell type-specific genes in cardiac myocytes under
normal circumstances. In light of the functional redundancy that has
been observed between p300 family members (21, 22), proof of this
requires a cell in which the levels of all p300 family members are
present in limiting quantities. Given E1A's ability to bind both p300
and CBP (21, 22), the two members of this family recognized thus far,
it is likely that we have created such a cellular milieu by
overexpressing E1A. In this setting, overexpression of p300 constructs
augments the transcriptional activity of a
-MHC promoter. In the
case of p300del33,
-MHC promoter activity increases to levels that
exceed even those observed in the absence of E1A. These data are
consistent with p300 family member(s) playing an important role in
-MHC transcription in normal cardiac myocytes. Other approaches to
induce a deficiency of p300 family members will be useful in confirming
this hypothesis.
Given the potential role played by p300 protein(s) as coactivator(s) of
- and
-MHC transcription, the question arises as to which cardiac
transcription factor(s) interact with and transactivate through these
proteins. Included among the motifs present in the promoter sequences
used in this study are binding sites for TEF-1 (50-56), GATA-4/5/6
(57, 58), and MEF-2 (59, 60) proteins. It will be interesting to
determine whether any of these and other transcription factors interact
in a functionally significant manner with p300 family members in
cardiac myocytes. Of note, p300 and MEF-2 have been shown to interact
in skeletal myocytes (27, 28).
Recent work has demonstrated that, in some systems, p300 protein(s) are present in the nucleus in limiting concentrations (30) and, thus, may be well positioned to coordinate changes in the transcription of multiple genes in response to signals from the cytoplasm. Therefore, it will be important to determine which p300 family members, including possibly novel ones, are present in cardiac myocytes, what their levels are relative to each other and the proteins with which they interact, and how these interactions change in response to various developmental and physiological stimuli.
Our results differ somewhat from an earlier report (14) in which an E1A
mutant defective in p300 binding alone retained the ability to repress
the activity of a skeletal -actin promoter in neonatal cardiac
myocytes, while a second mutant unable to bind both p300 and
pocket proteins lacked transrepression function. The reason for this
discrepancy is unclear but might be due to differences in the E1A
mutations in the adenoviruses used in the two studies, the promoter
sequences studied or the developmental state of the cells. Although the
ability of p300del33 to effect a more than complete rescue of E1A's
repression of
-MHC promoter activity argues that additional pocket
proteins are not needed for this effect, the possibility cannot be
excluded that pocket proteins play some role in activation of the
-MHC promoter. Of note, pocket proteins are required for the
transcriptional activity of the skeletal myogenic determination factors
(61, 62), and the interaction between the coactivator hBrm and Rb
enhances transcription in other systems (63). Despite this difference
between the two studies, however, the results of both underscore the
importance of the p300 family in the activation of cell type-specific
genes in cardiac myocytes.
We thank Yan Liu for preparation of the adenoviruses and multiple helpful discussions and Soo Jin Lee for help with preparations of cardiac myocytes.