The Neuron-Restrictive Silencer Element-Neuron-Restrictive Silencer Factor System Regulates Basal and Endothelin 1-Inducible Atrial Natriuretic ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
MOLECULAR AND CELLULAR BIOLOGY, Mar. 2001, p. 2085–2097 Vol. 21, No. 6
0270-7306/01/$04.00⫹0 DOI: 10.1128/MCB.21.6.2085–2097.2001
Copyright © 2001, American Society for Microbiology. All Rights Reserved.
The Neuron-Restrictive Silencer Element–Neuron-Restrictive Silencer
Factor System Regulates Basal and Endothelin 1-Inducible Atrial
Natriuretic Peptide Gene Expression in Ventricular Myocytes
KOICHIRO KUWAHARA,1 YOSHIHIKO SAITO,1* EMIKO OGAWA,1 NOBUKI TAKAHASHI,1
YASUAKI NAKAGAWA,1 YOSHIHISA NARUSE,2 MASAKI HARADA,1 ICHIRO HAMANAKA,1
TAKEHIKO IZUMI,1 YOSHIHIRO MIYAMOTO,1 ICHIRO KISHIMOTO,1 RIKA KAWAKAMI,1
MICHIO NAKANISHI,1 NOZOMU MORI,2,3 AND KAZUWA NAKAO1
Department of Medicine and Clinical Science, Kyoto University Graduate School of Medicine, Kyoto 606-8397,1 Department
of Molecular Genetic Research, National Institute for Longevity Sciences, Oobu, Aichi 474-8522,2 and CREST,
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
Science and Technology Corporation of Japan (JST), Kawaguchi, Saitama 332-0012,3 Japan
Received 19 October 2000/Accepted 19 December 2000
Induction of the atrial natriuretic peptide (ANP) gene is a common feature of ventricular hypertrophy. A
number of cis-acting enhancer elements for several transcriptional activators have been shown to play central
roles in the regulation of ANP gene expression, but much less is known about contributions made by
transcriptional repressors. The neuron-restrictive silencer element (NRSE), also known as repressor element
1, mediates repression of neuronal gene expression in nonneuronal cells. We found that NRSE, which is located
in the 3ⴕ untranslated region of the ANP gene, mediated repression of ANP promoter activity in ventricular
myocytes and was also involved in the endothelin 1-induced increase in ANP gene transcription. The repression
was conferred by a repressor protein, neuron-restrictive silencer factor (NRSF). NRSF associated with the
transcriptional corepressor mSin3 and formed a complex with histone deacetylase (HDAC) in ventricular
myocytes. Trichostatin A (TSA), a specific HDAC inhibitor, relieved NRSE-mediated repression of ANP
promoter activity, and chromatin immunoprecipitation assays revealed the involvement of histone deacetyla-
tion in NRSE-mediated repression of ANP gene expression. Furthermore, in myocytes infected with recombi-
nant adenovirus expressing a dominant-negative form of NRSF, the basal level of endogenous ANP gene
expression was increased and a TSA-induced increase in ANP gene expression was apparently attenuated,
compared with those in myocytes infected with control adenovirus. Our findings show that an NRSE-NRSF
system plays a key role in the regulation of ANP gene expression by HDAC in ventricular myocytes and provide
a new insight into the role of the NRSE-NRSF system outside the nervous system.
Cardiac hypertrophy is an adaptive response of the heart to molecular mechanisms involved in the establishment and
mechanical stress, tissue injury or neurohumoral activation, maintenance of the phenotypes of terminally differentiated
but while hypertrophy initially acts as a compensatory mecha- cardiac myocytes.
nism, when prolonged it can lead to heart failure. In most Many of the features seen with hypertrophy in vivo can be
forms of cardiac hypertrophy, there is an increase in the ex- duplicated using in vitro cardiomyocyte models, the most com-
pression of embryonic genes, including those encoding natri- monly used of which employ primary culture of neonatal rat
uretic peptides and fetal contractile proteins (6). Expression of cardiac myocytes (6). Using these systems, it has been shown
such embryonic genes in the ventricular myocardium normally that the 5⬘-flanking region (FR) of the ANP gene contains a
decreases during the perinatal period and remains quiescent in number of cis-acting enhancer elements for several transcrip-
the adult. For instance, the gene encoding atrial natriuretic tion factors, including Csx/Nkx2.5, GATA4, serum response
peptide (ANP) is expressed in both the atrium and the ventri- factor (SRF), AP-1, and Sp-1, which play important roles in the
cle during embryonic development, but shortly after birth its regulation of ANP gene expression under various conditions
expression is downregulated in the ventricle, leaving the atrium (3, 10, 14, 16, 27, 30, 45, 51, 52). Nevertheless, its expression
as the primary site of ANP synthesis (56, 58). However, when profile in the postnatal heart clearly suggests that ANP gene
ventricles are subjected to hemodynamic overload, expression expression in ventricles depends on negative as well as positive
of the ventricular ANP gene is reactivated (2, 29, 46). This transcriptional controls, but much less is known about the
induction of the ANP gene is a common feature of ventricular mechanisms which specifically mediate the repression of ANP
hypertrophy in all mammalian species and is a prognostic in- gene transcription.
dicator of clinical severity (2, 9, 11, 15, 21, 25, 29, 46, 53). The neuron-restrictive silencer element (NRSE), also known
Accordingly, elucidation of the mechanisms regulating ANP as repressor element 1 (RE-1), has been defined as a negative-
gene expression should enable one to better understand the acting DNA regulatory element that prevents expression of
neuronal genes in nonneuronal cell types and in undifferenti-
ated neuronal cells (28, 34). To date, several studies have
* Corresponding author. Mailing address: Department of Medicine
and Clinical Science, Kyoto University Graduate School of Medicine,
reported that NRSE-like sequences, present in the regulatory
54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606, Japan. Phone: 81-75- regions of multiple neuronal genes, are important for their
751-4287. Fax: 81-75-771-9452. E-mail: yssaito@kuhp.kyoto-u.ac.jp. neuron-specific expression (4, 12, 20, 22, 32, 33, 36, 42, 47, 48,
20852086 KUWAHARA ET AL. MOL. CELL. BIOL.
55). The neuron-restrictive silencer factor (NRSF), also known Plasmids encoding myc-tagged NRSF and a dominant-negative form of NRSF
as the RE-1 silencing transcription factor (REST), has been were kindly provided by David. J. Anderson (California Institute of Technology,
Pasadena). A GAL4-NRSF expression plasmid was constructed within the
identified as an NRSE-binding protein (7, 49). NRSF is a novel pBIND vector (Promega, Madison, Wis.). The full-length NRSF cDNA, con-
member of the Krüppel-like zinc finger transcriptional factor structed based on the plasmid encoding myc-tagged NRSF, was cloned in frame
family and is widely expressed in most nonneuronal tissues, with the GAL4 DNA-binding domain of the pBIND vector (pBIND-NRSF).
including the heart (7, 38, 48). NRSF is also strongly expressed Generation of recombinant adenovirus. The system used for introducing
cDNA into the viral genome was described in detail by Kanegae et al. (23). In
in undifferentiated neuronal progenitors in the nervous system,
brief, a cDNA encoding myc-tagged dominant-negative NRSF was generated by
and its downregulation during development enables the ex- PCR using the expression plasmid encoding myc-tagged dominant-negative
pression of neuron-specific terminal differentiation genes (5). NRSF as a template: an upstream sense oligonucleotide (5⬘-GGATC CCATC
Not only neuron-specific genes but also some non neuronal GATTT AAAGC TATGG AGC-3⬘), which incorporates a ClaI site at the 5⬘ end
genes have been reported to contain sequences similar to and a downstream antisense oligonucleotide (5⬘-CAGAT CGATT ACTCG
TTGCT GACGG CGTT-3⬘), which incorporates a ClaI site at the 3⬘ end, were
NRSE (48), but the function of NRSE in nonneuronal genes is used to amplify. The amplified products were verified by sequencing and then
not known yet. inserted into the ClaI site located upstream of the rabbit -globin polyadenyla-
In this study, we found that NRSE, which is located in the 3⬘
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
tion signal and downstream of the chicken -actin promoter-cytomegalovirus
untranslated region (UTR) of the ANP gene, mediated repres- enhancer (CAG promoter) (39) of the cosmid carrying the adenovirus vector.
sion of ANP expression in ventricular myocytes and was in- The recombinant virus, named Ad/DNNR, was purified and concentrated as
described previously (24).
volved in endothelin 1 (ET-1)-induced reactivation of ANP Ventricular myocyte culture and transfection. Neonatal (2- to 4-day-old) rat
gene expression. NRSF bound to the NRSE, and the repres- ventricular myocytes were prepared on a Percoll gradient (37), and their tran-
sion was mediated in part through the association of NRSF sient transfection was carried out by electroporation at 280 V and 300 mF (26).
with the transcriptional corepressor mSin3, resulting in the When transfected with reporter plasmid only, 10 g of promoter-reporter con-
struct was transfected per 3 ⫻ 106 cells. In cotransfection experiments, the DNA
recruitment of histone deacetylase (HDAC). This is the first
mixture contained 10 g of reporter and 0 to 10 g of the appropriate expression
report showing the involvement of the NRSE-NRSF system vector; pCDNA3.1 was then used to adjust the total amount of DNA to 20 g per
and HDAC in the regulation of cardiac embryonic gene ex- 3 ⫻ 106 cells. In addition, in all experiments, 4 g of pRL-TK (TOYO INC Co.,
pression in ventricular myocytes. Furthermore, this study also Ltd.), in which the herpes simplex virus thymidine kinase (TK) promoter was
provides new evidence that the NRSE-NRSF system, which is fused to the Renilla luciferase gene, was cotransfected and used to normalize
luciferase activity. Transfected cells were initially plated for 24 h in gelatin-
known to function as a silencer to determine cell-type-specific coated six-well plates (5 ⫻ 105 cells per well) in Dulbecco modified Eagle
expression of neuronal genes, also participates in the regula- medium (DMEM) supplemented with 10% fetal calf serum (FCS). The medium
tion of nonneuronal gene expression as a transcriptional re- was then changed to serum-free DMEM for 12 h and finally to serum-free
pressor in nonneuronal cells. DMEM containing 0.1% bovine serum albumin, with or without ET-1 or TSA,
for 48 h. The cells were then harvested, lysed, and assayed for luciferase activity
using a luminometer (Lumat LB 9507; Berthold, Wildbad, Germany) according
MATERIALS AND METHODS to the manufacturer’s protocol (TOYO INC Co., Ltd.). In each experiment,
Agents. Human ET-1 was purchased from Peptide Institute, Inc. (Osaka, aliquots of cell lysate from triplicate wells were assayed, and the luciferase
Japan), and trichostatin A (TSA) was from Wako Pure Chemical industries, Ltd. activities were normalized to pRL-TK luciferase activities.
(Osaka, Japan). For immunoprecipitation experiments, ventricular myocytes were transfected
Plasmid constructs. A luciferase gene driven by the proximal enhancer-pro- with 10 g of expression vector encoding myc-tagged NRSF alone or cotrans-
moter of the ANP gene, ⫺452hANPLuc, was generated by PCR using genomic fected with a combination of 10 g of the vector encoding myc-tagged NRSF and
DNA as a template: an upstream sense oligonucleotide (5⬘-GAAGA TCTGG FLAG-tagged mSin3B. Cells were then maintained with DMEM supplemented
CTGCC TGCCA TTTCC TCCTC TCCAC CCTTA TTTGG-3⬘), which incor- with 10% FCS for 72 h.
porates a BglII site at the 5⬘ end, and a downstream antisense oligonucleotide Adenoviral infection. One day after plating, ventricular myocytes were infected
(5⬘-CGCAA GCTTG CTGGC GTCGT CAAGG AGC-3⬘), which incorporates with adenovirus at a multiplicity of infection of five particles/cell for 24 h in
a HindIII site at the 3⬘ end, were used to amplify the relevant fragments from DMEM with 10% FCS. The cells were cultured in serum-free medium for an
⫺452 to ⫹97 of the ANP gene. The amplified products were then cloned into the additional 24 h. The efficiency of expression, examined by determining lacZ gene
BglII/HindIII sites of a luciferase reporter construct PGV-B2 (TOYO INC Co., expression (Ad/lacZ) in cultured ventricular myocytes, is consistently more than
Ltd., Tokyo, Japan). The fragment from ⫹1769 to ⫹2228 of the ANP gene, 90% by this protocol.
which encompasses the 3⬘-UTR and the 3⬘ FR, was obtained by PCR using EMSAs. Nuclear extract from P19 cells was prepared as previously described
an upstream sense oligonucleotide (5⬘-AAGGA GCTCG CAGTA CTGAA (8). Double-stranded oligonucleotides containing two copies of NRSEANP (5⬘-
GATAA CAGC-3⬘), which incorporates a SacI site at the 5⬘ end, and a down- CACTT CAGCA CCATG ACAG AAGA-3⬘) or mutant NRSE (5⬘-CACTT
stream antisense oligonucleotide (5⬘-GCGAG ATCTA CTAGA TGAAT CAGCA CA ATTTA CAGAA GACAA A-3⬘) were synthesized and used as a
TAGGC CTCC-3⬘), which incorporates a BglII site at the 3⬘ end. This fragment probe for electrophoretic mobility shift assays (EMSAs). DNA-protein binding
was cloned into the SacI/BglII sites located upstream of the ANP promoter in reactions were carried out in a 20-l final volume of reaction buffer containing
⫺452hANPLuc, and the resultant plasmid was designated ⫹1769/⫺452hANPLuc. A 20 mM HEPES (pH 7.9), 125 mM KCl, 5 mM MgCl2, 10% glycerol, 125 g of
second fragment from positions ⫹1769 to ⫹2228 that included a BamHI site at poly(dI-dC) per ml, and 1 mM dithiothreitol. The nuclear extract (15 g of
the 5⬘ end and a SalI site at the 3⬘ end was amplified by PCR using an upstream protein) was added to the reaction buffer and preincubated for 10 min on ice.
sense oligonucleotide (5⬘-TACGG ATCCG CAGTA CTGAA GATAA CAGC- Radiolabeled DNA probe was then added, and the nuclear extract was incubated
3⬘), which incorporates a BamHI site at the 5⬘ end, and a downstream antisense for another 30 min at room temperature. Electrophoresis was then performed in
oligonucleotide (5⬘-ATAGT CGACA CTAGA TGAAT TAGGC CTCC-3⬘), 4% polyacrylamide gels in 0.25⫻ Tris-borate-EDTA for 1 h at 150 V. For
which incorporates a SalI site at the 3⬘ end. This fragment was cloned into the competition assays, 50⫻ cold double-stranded NRSEANP, mutant NRSEANP,
BamHI/SalI sites located downstream of the luciferase gene in ⫺452hANPLuc, NRSE from the SCG10 promoter (NRSESCG10), or the EBNA-binding sequence
and the resultant plasmid was designated ⫺452/⫹1769hANPLuc. Mutations in was coincubated in reaction mixtures with double-stranded radiolabeled
the ANP gene NRSE (NRSEANP) were generated by PCR using either wild-type NRSEANP probe as described above.
⫹1769/⫺452hANPLuc or ⫺452/⫹1769hANPLuc and their homologous NRSE Monoclonal antibody raised against NRSF (3B7) used in the supershift exper-
mutants as templates. The sequence of the mutated NRSEANP was 5⬘-CTTCA iment was generated using the NRSF N-terminal domain (positions 1 to 156)
GCACA ATTTA CAGAA GACAA A-3⬘, which has been reported to result fused with glutathione S-transferase as an antigen and purified through BALB/c
in the loss of NRSF binding to the sequence (34). The plasmids carrying ascites.
the mutation were designated mtNR/⫹1769/⫺452hANPLuc and mtNR/⫺452/ Coimmunoprecipitation. Ventricular myocytes transfected only with expres-
⫹1769hANPLuc. All plasmid constructions were verified by sequencing. sion vector encoding myc-tagged NRSF were cultured for 72 h with DMEMVOL. 21, 2001 NRSF AND CARDIAC EMBRYONIC GENE 2087
supplemented with 10% FCS, collected in lysis buffer, and sonicated. The cell enhancer-promoter region of the ANP gene is known to me-
debris was cleared by centrifugation, after which the supernatant (1 ml) was diate cardiac myocyte-specific gene expression, it is not suffi-
mixed overnight at 4°C with monoclonal anti-myc antibodies (9E7, 1:100 dilu-
tion; Santa Cruz Biotech, Inc., Santa Cruz, Calif.). Immunoprecipitates were
cient to regulate the expression of the ANP gene under several
subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- conditions (25, 44). To determine the regulatory roles of the
PAGE) and analyzed by Western blotting using the following dilutions of anti- ANP gene 3⬘-UTR and 3⬘-FR, the sequence from positions
bodies as probes (Santa Cruz Biotech): anti-myc (1:500), anti-mSin3A (1:100), ⫹1769 to ⫹2228 of the ANP gene was inserted either upstream
anti-mSin3B (1:100), anti-HDAC1 (1:100), and anti-HDAC2 (1:100).
or downstream of a reporter gene driven by the ANP proximal
Alternatively, the supernatant obtained from ventricular myocytes transfected
with the vector combination encoding myc-tagged NRSF and FLAG-tagged enhancer-promoter (Fig. 1A). To our surprise, the fragment
mSin3B was mixed overnight at 4°C with anti-myc (1:100) or anti-FLAG (M2, markedly repressed the activity of the ANP promoter in both
1:100; Sigma Chemical, St. Louis, Mo.) antibodies. Immunoprecipitates were cases (Fig. 1B). Using a computer search, we identified se-
subjected to SDS-PAGE and analyzed by Western blotting using anti-myc (1: quences in this region that were similar to NRSE. Such NRSE-
1,000) and anti-FLAG (1:500).
ChIP assay. Chromatin immunoprecipitation (ChIP) assays were carried out
like sequences (NRSEANP) are well conserved in the ANP
according to the protocols supplied by the manufacturer (Upstate Biotech, Lake genes of mammalian species (Fig. 2A), and we therefore hy-
pothesized that repression by ANP gene fragment from ⫹1769
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
Placid, N.Y.). Briefly, cells were fixed in 1% formaldehyde for 15 min at 37°C.
They were then collected, resuspended in SDS lysis buffer (1% SDS, 10 mM to ⫹2228 is mediated by NRSEANP.
EDTA, 50 mM Tris-HCl [pH 8.1]) containing 1 mM phenylmethylsulfonyl flu-
To verify that NRSEANP mediated the observed repression,
oride, 1 g of aprotinin per ml, and 1 g of pepstatin A per ml; sonicated three
times for 10 s each time; and cleared by centrifugation. One aliquot of the lysate we generated a mutant NRSEANP, as previously reported
(10 l) was removed and served as a control; the remainder was incubated (34), and examined the expression of the mutant constructs.
overnight at 4°C in ChIP dilution buffer (16.7 mM Tris [pH 8.1], 167 mM NaCl, As shown in Fig. 2B, introduction of a mutation into the
1.1% Triton X-100, 0.01% SDS, 1.2 mM EDTA, 1 mM phenylmethylsulfonyl NRSEANP completely abolished repression mediated by the
fluoride, 1 g/m of aprotinin per ml, and 1 g of pepstatin A per ml) with 5 l
of anti-acetylated H4 antibody. Immunocomplexes were then recovered by add-
fragment from ⫹1769 to ⫹2228, confirming that NRSEANP
ing 60 l of salmon sperm DNA and a protein A-agarose bead suspension, mediated the repressor activity of this fragment. The fact that
followed by incubation at 4°C for 4 h. The beads were then sequentially washed the mutant constructs showed greater luciferase activities than
for 5 min each time in 1 ml of buffer containing 150 mM or 500 mM NaCl plus the construct containing the proximal promoter alone suggests
0.1% SDS, 1% Triton X-100, 20 mM Tris-HCl (pH 8.1), 2 mM EDTA, and
that other enhancer elements exist in the region whose activ-
Tris-EDTA (TE) (pH 8.0). The immunocomplexes were eluted by incubating the
beads with 200 l of 1% SDS plus 100 mM NaHCO3. After the addition of 5 M ities are normally repressed in the presence of NRSEANP.
NaCl, the eluates were heated to 65°C for 6 h to reverse any formaldehyde When the fragment from ⫹1769 to ⫹2228, with or without the
cross-linking. DNA was recovered using proteinase K treatment, phenol-chloro- mutation, was fused upstream of the TK or simian virus 40 (SV
form extraction, and ethanol precipitation. The resultant pellets were resus- 40) promoter (mtNR/⫹1769hA-TKLuc or ⫹1769hA-TKLuc
pended in 50 l of TE. Quantitative PCR was then carried out for 23 to 28 cycles
using 3-l samples of the DNA. Primers amplifying the DNA fragments encom-
[for TK] or mtNR/⫹1769hA-SV40Luc or ⫹1769hA-SV40Luc
passing the 3⬘-UTR of the human ANP gene and the PGV-B2 sequence were [for SV40]), the reporter constructs containing the mutant
used to amplify ⫺452/⫹1769hANPLuc and mtNR/⫺452/⫹1769hANPLuc, while NRSEANP showed greater luciferase activities than those con-
primers amplifying the DNA fragment encompassing the 3⬘-UTR and 3⬘-FR of taining wild-type NRSEANP, thus demonstrating that NRSEANP
the rat ANP gene were used to amplify the endogenous rat ANP gene. PCR
can repress the activities of multiple promoters in ventricular
products were resolved by electrophoresis and visualized with ethidium bromide.
Images were recorded and quantified using NIH Image 1.5 Software. Aliquots of myocytes (Fig. 2C and D).
chromatin before immunoprecipitation were also analyzed by PCR (INPUT). To further confirm that NRSE is sufficient to repress the
Northern blot analysis. Total RNA was prepared from ventricular myocytes promoters in ventricular myocytes, we generated reporter
using Trizol (Life Technologies, Inc., Rockville, Md.), and 10-g samples were genes in which only two copies of NRSEANP were inserted
separated and transferred to Biodyne membranes (Pall Corp., Glen Cove, N.Y.).
Randomly labeled ANP and glyceraldehyde-3-phosphate dehydrogenase
upstream of ⫺452hANPLuc or TKLuc (2NR-hALuc and 2NR-
(GAPDH) cDNA probes were then hybridized to the membranes as previously TKLuc, respectively). As shown in Fig. 2E and F, the luciferase
described (13, 37). activities of 2NR-hALuc or 2NR-TKLuc were significantly
RT-PCR analysis. Single-stranded cDNA was synthesized with the Thermo- lower than those of ⫺452hANPLuc or TKLuc, respectively.
script reverse transcription-PCR (RT-PCR) kit (Life Technologies, Inc.) using 3
NRSF represses transcription of the ANP gene through
g of total RNA prepared from cultured ventricular myocytes incubated with or
without 10 nM ET-1 for 24 h. The mRNA level of NRSF and REST4, an NRSEANP. NRSF (or REST), a zinc finger transcriptional re-
altenative spliced variant of NRSF, in cultured ventricular cells was determined pressor protein, is known to bind to NRSE (7, 49). To test
by PCR using an upstream sense oligonucleotide (5⬘-GCTAC AGTTA TGGCC whether NRSF binds to NRSEANP, EMSAs were carried out
ACCCA GGTGA T-3⬘) and a downstream antisense oligonucleotide (5⬘-GC using nuclear extract from P19 cells, which is known to be
GTA GTCAC ACACG GGGCA GTTGA AC-3⬘) for NRSF and an upstream
sense oligonucleotide (5⬘-CTACA TGGCA CACCT GAAGC ACCAC-3⬘) and
abundant in NRSF, and a shifted band was observed when
a downstream antisense oligonucleotide (5⬘-GGCTT CTCAC CCATC TAGAT radiolabeled NRSEANP was used as a probe (Fig. 3A, arrow,
CACAC T-3⬘) for REST4 (40). The PCR products were analyzed on a 2% lane 2). The shift was completely blocked by the addition of
agarose gel. cold NRSEANP (Fig. 3A, lane 3) or NRSE from the SCG10
RIA. The concentrations of ANP in the culture medium were measured by
promoter (NRSESCG10; Fig. 3A, lane 4) but was unaffected by
using our specific radioimmunoassay (RIA) as previously reported (13, 37).
Statistical analysis. Data are presented as the mean ⫾ the standard error of a cold probe for EBNA binding sequences (Fig. 3A, lane 5).
the mean (SEM). Analysis of variance with post hoc Fisher’s tests was used to When the mutant NRSEANP was used as a probe, no shifted
determine significant differences. P values of ⬍0.05 were considered significant. band was observed (Fig. 3B, lane 7, intact NRSEANP as a
probe, versus lane 8, mutant NRSEANP as a probe). In addi-
RESULTS tion, when intact NRSEANP was used as a probe (Fig. 3B, lane
9), cold mutant NRSEANP probe failed to block the band shift,
NRSEANP in the 3ⴕ-UTR mediates repression of ANP gene in contrast to the cold intact NRSEANP probe (Fig. 3B, lane
transcription in ventricular myocytes. Although the proximal 10). Furthermore, the band was supershifted in the presence of2088 KUWAHARA ET AL. MOL. CELL. BIOL.
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 1. The sequence spanning positions ⫹1769 to ⫹2228 of the ANP gene mediates repression on the ANP promoter. (A) Schematic
representation of the reporter constructs. (B) A total of 10 g of ⫺452hANPLuc, ⫹1769/⫺452hANPLuc, or ⫺452/⫹1769hANPLuc was
cotransfected into ventricular myocytes with 4 g of pRL-TK; ⫹1769/⫺452hANPLuc and ⫺452/⫹1769hANPLuc contain the fragment from ⫹1769
to ⫹2228 inserted upstream or downstream, respectively, of ⫺452hANPLuc. Transfected cells were plated in DMEM supplemented with 10% FCS
for 24 h and then maintained in serum-free DMEM for 60 h. Luciferase activity was normalized to the activity of Renilla luciferase driven by the
TK promoter; the relative luciferase activities of ⫺452hANPLuc were assigned a value of 100. The bars represent the mean ⫾ the SEM values
of relative luciferase activities from at least three separate assays carried out in triplicate. ⴱ, P ⬍ 0.05 versus ⫺452hANPLuc.
an antibody raised against NRSF (Fig. 3C, arrowhead, lane 13). cate that NRSF represses ANP gene expression in ventricular
All of these results indicate that NRSF binds to NRSEANP. myocytes through NRSEANP.
We next assessed whether the bound NRSF was responsible The NRSE-NRSF system is involved in ET-1-induced ex-
for the NRSEANP-mediated repression of ANP expression. pression of the ANP gene. ANP gene expression is reactivated
Ventricular myocytes were cotransfected with reporter genes in ventricular myocytes during cardiac muscle cell hypertrophy.
and a dominant-negative NRSF mutant that lacked both the We examined the role of NRSEANP in hypertrophic stimulus-
N- and C-terminal regions required for repressor activities (5). induced ANP gene expression using ET-1 as a hypertrophy-
As shown in Fig. 3D, dominant-negative NRSF abolished the inducing agent (13, 37).
repression of the ANP promoter mediated by an intact 3⬘- We first evaluated the effect of NRSEANP mutation on the
UTR, whereas it had no effect on reporter gene expression by ET-1 inducibility of an ANP reporter gene. Figure 4A shows
constructs carrying the mutant NRSEANP. To further confirm that the ET-1 inducibility of mtNR/⫺ 452/⫹1769hANPLuc or
that recruitment of NRSF is sufficient to repress ANP pro- mtNR/⫹1769/⫺452hANPLuc was about 50% lower than that
moter activity, we generated 5UAS-hANPLuc, a reporter plas- of ⫺452/⫹1769hANPLuc or ⫹1769/⫺452hANPLuc, respec-
mid containing five copies of the GAL4 binding site inserted tively. This suggests that NRSEANP, at least in part, mediates
upstream of ⫺ 452hANPLuc, and cotransfected myocytes with the ET-1 inducibility of ANP gene expression.
5UAS-hANPLuc and the plasmid encoding GAL4-NRSF fu- To test whether NRSEANP could mediate the ET-1 induc-
sion protein. The marked repression of the activities of 5UAS- tion of other promoters, the ET-1 inducibility of ⫹1769hA-
hANPLuc in the presence of GAL4-NRSF shown in Fig. 3E TKLuc was compared with that of mtNR/⫹1769hA-TKLuc.
clearly illustrates that NRSF recruited by the gene represses We found that the reporter activities of ⫹1769hA-TKLuc were
ANP promoter activity. increased by ET-1 stimulation and that mutation of NRSEANP
We further examined a function of NRSF in endogenous within the construct (mtNR/⫹1769hA-TKLuc) markedly de-
ANP gene expression by using a recombinant adenovirus ex- creased the capacity of ET-1 to induce reporter activity (Fig.
pressing a dominant-negative form of NRSF (Ad/DNNR). As 4B). To further examine the role of NRSEANP in mediating the
shown in Fig. 3F and G, endogenous ANP mRNA expression induction of the ANP promoter by ET-1, we compared the
in ventricular myocytes infected with Ad/DNNR was markedly ET-1 inducibility of 2NR-hALuc to that of ⫺452hANPLuc. As
increased, compared with that in myocytes infected with the shown in Fig. 4C, the response to ET-1 of 2NR-hALuc, in
control adenovirus (Ad/lacZ). All of these lines of data indi- which only two copies of NRSE were inserted upstream of theVOL. 21, 2001 NRSF AND CARDIAC EMBRYONIC GENE 2089
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 2. NRSEANP mediates the repressor activities of the ANP gene fragment from ⫹1769 to ⫹2228. (A) NRSEANP is highly conserved among
mammalian species; shown are the NRSE sequences in the ANP gene of a human, a mouse, a rat, a horse, a guinea pig, a cow, and a sheep. (B)
NRSEANP mutation abolished the repressor activities of the fragment from ⫹1769 to ⫹2228. A total of 10 g of ⫹1769/⫺452hANPLuc,
⫺452/⫹1769hANPLuc, ⫺452hANPLuc, mtNR/⫹1769/⫺452hANPLuc, or mtNR/⫺452/⫹1769hANPLuc was cotransfected into ventricular myo-
cytes with 4 g of pRL-TK. The luciferase activity was normalized to the activity of Renilla luciferase driven by the TK promoter; the relative
luciferase activity of ⫺452hANPLuc was assigned a value 100. ⴱ, P ⬍ 0.05 versus ⫺425/⫹1769hANPLuc; †, P ⬍ 0.05 versus ⫹1769/⫺425hANPLuc.
(C and D) NRSE mediates repression of TK (C) and SV40 (D) promoters. Fragments from ⫹1769 to ⫹2228 containing either wild-type or mutant
NRSE were inserted upstream of the TK and SV40 promoters (⫹1769hA-TKLuc or mtNR/⫹1769hA-TKLuc, and ⫹1769hA-SV40Luc or
mtNR/⫹1769hA-SV40Luc, respectively). Ventricular myocytes were transfected with ⫹1769hA-TKLuc or mtNR/⫹1769hA-TKLuc and 4 g of
pRL-SV40, in which the SV40 promoter was fused to the Renilla luciferase gene (C), or with ⫹1769hA-SV40/Luc or mtNR/⫹1769hA-SV40Luc
and 4 g of pRL-TK (D). The relative luciferase activity of ⫹1769hA-TKLuc or ⫹1769hA-SV40Luc was assigned a value of 1.0. ⴱ, P ⬍ 0.05 versus
⫹1769hA-TKLuc (C) or ⫹1769hA-SV40Luc (D). (E and F) NRSE is sufficient to repress the ANP (E) or TK (F) promoter. Two copies of NRSE
were inserted upstream of ⫺452ANPLuc or TKLuc and designated 2NR-hALuc or 2NR-TKLuc, respectively. Ventricular myocytes were
transfected with ⫺452ANPLuc or 2NR-hALuc and pRL-TK (E), or with TKLuc or 2NR-TKLuc and pRL-SV40 (F). The relative luciferase activity
of ⫺452hANPLuc or TKLuc was assigned a value of 1.0. ⴱ, P ⬍ 0.05 versus ⫺452hANPLuc (E) or TKLuc (F). For all experiments, the bars
represent the mean ⫾ SEM values of relative luciferase activities from at least three separate assays carried out in triplicate.
ANP promoter, was increased in comparison with that of However, RT-PCR analysis showed that treatment of ventric-
⫺452hANPLuc (Fig. 4C). When the effect of dominant-nega- ular myocytes with ET-1 did not decrease the level of NRSF
tive NRSF on the ET-1 inducibility of ANP reporter gene mRNA (Fig. 4E). A recent study demonstrated that, in neu-
activity was examined, we found that cotransfection of domi- ronal cells, REST4, a neuron-specific splice variant of NRSF, is
nant-negative NRSF with ⫹1769/⫺452hANPLuc diminished induced in a protein kinase A-dependent manner and antag-
the capacity of ET-1 to induce reporter gene activity, although onizes the repressor function of NRSF (50). We next examined
cotransfection of dominant-negative NRSF with mtNR/⫹1769/ whether REST4 is induced by ET-1 in ventricular myocytes. As
⫺452hANPLuc did not (Fig. 4D). It thus appears that the shown in Fig. 4F, however, REST4 mRNA was neither de-
NRSF-NRSE system has a significant function in the ET-1- tected in basal condition nor induced by ET-1 in ventricular
induced signaling pathways leading to the upregulation of ANP myocytes. These results suggest that ET-1 may modify the
gene expression. function of NRSF by posttranscriptional mechanisms.
Because all these data suggest that ET-1 attenuates NRSE- NRSF associates with the mSin3-HDAC complex in ventricu-
mediated repression, we considered the possibility that ET-1 lar myocytes. Recent studies have shown that NRSF associates
downregulates the level of NRSF in ventricular myocytes. with the mSin3-HDAC complex both in vivo and in vitro and that2090 KUWAHARA ET AL. MOL. CELL. BIOL.
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 3. NRSF represses transcription of the ANP gene via NRSE. (A to C) EMSAs were carried out using nuclear extract from P19 cells and
a radiolabeled NRSE or mutant NRSE oligonucleotide, as described in Materials and Methods. (D) Dominant-negative NRSF abolished
NRSE-mediated repression. A total of 10 g of expression vector encoding dominant-negative NRSF or control vector was cotransfected into
ventricular myocytes with 10 g of either ⫹1769/⫺452hANPLuc or mtNR/⫹1769/⫺452hANPLuc and 4 g of pRL-TK. The luciferase activity was
normalized to the activity of Renilla luciferase driven by the TK promoter; the luciferase activity of ⫹1769/⫺452hALuc cotransfected with the
control vector was assigned a value of 1.0. ⴱ, P ⬍ 0.05 versus ⫹1769/⫺452hALuc without DN-NRSF. (E) Recruitment of NRSF is sufficient to
repress ANP promoter activity. 5UAS-hANPLuc, which contains five copies of the GAL4-binding site inserted upstream of ⫺452hANPLuc, was
cotransfected with the indicated doses of plasmid encoding GAL4-NRSF fusion protein (pBIND-NRSF). The relative luciferase activity of
5UAS-hANPLuc transfected with only a control vector was assigned a value of 1.0. ⴱ, P ⬍ 0.05 versus 5UAS-hANPLuc without pBIND-NRSF.
For all experiments, the bars represent the mean ⫾ the SEM value of the relative luciferase activities from at least three separate assays carried
out in triplicate. (E) Expression of dominant-negative NRSF in ventricular myocytes infected with Ad/DNNR. Whole-cell lysates from ventricular
myocytes infected with Ad/lacZ or Ad/DNNR were separated by electrophoresis and blotted with monoclonal anti-myc antibody (9E7). (F) A
dominant-negative NRSF increased endogenous rat ANP gene expression in ventricular myocytes. Myocytes were infected with control Ad/lacZ
or with Ad/DNNR at a titer of five particles/cell. Cells were incubated for 48 h and then harvested for analysis by Northern blotting. Shown are
representative blots for ANP mRNA from three independent experiments with identical results.
histone deacetylation is involved in the transcriptional repression mSin3-HDAC complex, coimmunoprecipitation was per-
mediated by the NRSE-NRSF system (18, 38, 44). We therefore formed on whole-cell lysates from ventricular myocytes trans-
examined whether the interaction between NRSF and mSin3- fected with the expression vector for myc-tagged NRSF. The
HDAC also occurs in ventricular myocytes. An expression plas- lysates were then incubated with either anti-myc antibody or
mid encoding FLAG-tagged mSin3B was transfected into ventric- control mouse immunoglobulin G (IgG). Immunocomplexes
ular myocytes, with or without the expression plasmid encoding were separated by SDS-PAGE, subjected to Western blotting,
myc-tagged NRSF, after which whole-cell lysates were immuno- and visualized using anti-mSin3A and -B or anti-HDAC1 and
precipitated using monoclonal anti-myc and anti-FLAG antibod- -2. As shown in Fig. 5B, endogenous mSin3A and -B and
ies. When immunoprecipitates pulled down by anti-FLAG anti- HDAC1 and -2 proteins were immunoprecipitated with intro-
body were separated on SDS-PAGE and subjected to Western duced myc-tagged NRSF, indicating that, as previously re-
blot analysis, FLAG-tagged mSin3B was found to precipitate ported in other cell types (18, 38), NRSF in ventricular myo-
along with myc-tagged NRSF, whose expression was confirmed by cytes recruits the mSin3-HDAC complex. These results suggest
Western blot analysis (Fig. 5A). the involvement of HDAC in NRSE-NRSF-mediated repression
To test whether NRSF associates with the endogenous of ANP gene transcription.VOL. 21, 2001 NRSF AND CARDIAC EMBRYONIC GENE 2091
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 4. Involvement of the NRSE-NRSF system in ET-1-induced ANP gene expression in ventricular myocytes. (A) NRSE mutation decreased the
ET-1 inducibility of ANP reporter genes. A total of 10 g of ⫹1769/⫺452hANPLuc, ⫺452/⫹1769hANPLuc, ⫺452hANPLuc, mtNR/⫹1769/⫺452hAN-
PLuc, or mtNR/⫺452/⫹1769hANP/Luc was cotransfected into ventricular myocytes with 4 g of pRL-TK, after which the cells were incubated for 48 h
with or without 10 nM ET-1. (B) NRSE mediates the ET-1 inducibility of the TK promoter. Cells were cotransfected with ⫹1769hA-TKLuc or
mtNR/⫹1769hA-TKLuc and 4 g of pRL-SV40 and incubated for 48 h with or without 10 nM ET-1. ⴱ, P ⬍ 0.05 versus ⫹1769hA-TKLu with ET-1. (C)
⫺452ANP Luc or 2NR-hALuc and pRL-TK were cotransfected into cells treated with or without 10 nM ET-1. ⴱ, P ⬍ 0.05 versus ⫺452hANPLuc with
ET-1. (D) Dominant-negative NRSF decreased the ET-1 inducibility of ANP gene expression. A total of 10 g of expression vector encoding
dominant-negative NRSF or control vector was cotransfected into cells with 10 g of ⫹1769/⫺452hANPLuc or mtNR/⫹1769/⫺452hANPLuc and 4 g
of pRL-TK. ⴱ, P ⬍ 0.05 versus ⫹1769/⫺452hANPLuc with ET-1 and pCS2. In panels A, C, and D the luciferase activity was normalized to the activity
of Renilla luciferase driven by the TK promoter; in panel B the luciferase activity was normalized to the activity of Renilla luciferase driven by the SV40
promoter. For all experiments, the activity of each construct transfected into myocytes treated with 10 nM ET-1 was expressed as the fold increase over
the value for myocytes treated without ET-1. The bars represent the mean ⫾ the SEM values of relative luciferase activities from at least three separate
assays carried out in triplicate. (E) RT-PCR analysis of NRSF mRNA expression in cultured ventricular myocytes treated with (⫹) or without (⫺) 10
nM ET-1 for 24 h. Total RNA (3 g) was reverse transcribed, and the indicated cycles of PCR amplification were performed. (F) RT-PCR analysis of
REST4 mRNA expression in ventricular myocytes (indicated as MC) treated with (⫹) or without (⫺) 10 nM ET-1 for 24 h and rat brain (indicated as
Brain).
NRSE-NRSF represses ANP gene transcription through his- ANP gene expression. Ventricular myocytes transfected with
tone deacetylation. To determine whether HDAC is involved reporter plasmids containing intact or mutated NRSEANP
in NRSE-mediated repression of ANP gene transcription, we were incubated with 10 ng of TSA per ml for 48 h, after which
tested the effects of TSA, a specific inhibitor of HDAC, on the cells were lysed and the luciferase activities were measured.2092 KUWAHARA ET AL. MOL. CELL. BIOL.
expression in ventricular myocytes infected with Ad/DNNR
was diminished compared with that in myocytes infected with
control Ad/lacZ (Fig. 7D and E). These results clearly indicate
that the NRSE-NRSF system represses ANP gene expression,
at least in part, through the recruitment of HDAC.
Because we found that NRSEANP is involved in ET-1-in-
duced ANP expression, our final experiments were designed to
determine the effect of ET-1 on histone acetylation. As shown
in the Fig. 7F, ET-1 increased the acetylation of histones as-
sociated with the 3⬘-UTR of the ANP gene, supporting the
notion that attenuation of NRSE-mediated repression is an
important component of the mechanism by which ET-1 in-
duces ANP gene expression.
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
DISCUSSION
In the regulation of ANP gene expression in ventricular
myocytes, a number of cis-acting enhancer elements have been
shown to play central roles (3, 10, 16, 30, 51, 52). Ventricular
ANP gene expression is elevated during embryonic life, but it
remains low in the postnatal heart unless the ventricle is sub-
FIG. 5. Interaction between NRSF and mSin3-HDAC in ventricu- jected to hemodynamic stress. This profile of ventricular ANP
lar myocytes. (A) Whole-cell lysates of ventricular myocytes trans-
fected with either FLAG-tagged mSin3B alone or in combination with gene expression suggests that negative as well as positive tran-
myc-tagged NRSF were subjected to immunoprecipitation (IP). The scriptional controls play an important role in ANP gene ex-
top panel shows the bands immunoprecipitated by monoclonal anti- pression, but the contribution made by transcriptional repres-
FLAG antibody (M2), separated by electrophoresis, and blotted with sors remains unclear. Here we show that NRSE, which is
monoclonal anti-myc antibody (9E5). The middle panel shows a West-
located in the 3⬘-UTR of the ANP gene, mediates repression
ern blot of an M2 immunoprecipitate probed with M2, while the
bottom panel shows a Western blot of a 9E5 immunoprecipitate of ANP gene expression in ventricular myocytes and is also
probed with 9E5. (B) Whole-cell lysates from ventricular myocytes involved in the ET-1-induced increase in ANP gene expres-
transfected with myc-tagged NRSF were subjected to immunoprecipi- sion. The repression is conferred by NRSF, and HDAC is
tation (IP). The lysates were incubated with either the 9E5 or control involved in the NRSF-mediated repression of ANP gene tran-
IgG and precipitated with protein G-agarose. The panels depict the
immunocomplex separated by electrophoresis and blotted with either scription. NRSF-mSin3-HDAC complex-mediated repression
anti-myc, anti-mSi3A, anti-mSin3B, anti-HDAC1, or anti-HDAC2 an- under the basal condition and several sources of hypertrophic
tibody (␣), as indicated. stimulus-induced inhibition of the repression may contribute
to the dynamic changes of ANP gene expression in ventricular
myocytes. Our findings show a novel regulatory mechanisms of
TSA increased the transcriptional activities of reporter genes ANP gene expression in ventricular myocytes and provide a
containing intact NRSE to a greater degree than those con- new insight into the role of the NRSE-NRSF system outside
taining mutated NRSEANP or those lacking NRSEANP alto- the nervous system.
gether (Fig. 6A). The ANP 3⬘-UTR containing the NRSE repressed the ANP
Then, to test whether NRSE mutation correlates with proximal promoter, whether it was situated upstream or down-
changes in the status of histone acetylation, ChIP assays were stream of the promoter. NRSEANP mutation completely abol-
performed using primers spanning NRSEANP. These assays ished the repression by the 3⬘-UTR, confirming the role of
revealed that NRSEANP mutation markedly increases the level NRSE in mediating the repression. NRSEANP also mediated
of acetylated histones associated with the 3⬘-UTR of the ANP repression of the TK and SV40 promoters, indicating that
gene (Fig. 6B). Moreover, it was found that, in the endogenous NRSEANP is capable of repressing multiple promoters in ven-
ANP gene, the acetylation level of histones around NRSEANP tricular myocytes. In fact, two copies of NRSEANP were suffi-
is low under basal conditions (Fig. 7A) and that TSA treatment cient to repress both the ANP and TK promoters. Because
increases the acetylation level of the histones, suggesting that NRSF is expressed in nonneuronal tissues, including the heart,
HDAC is largely responsible for maintaining basal hypoacety- and has been reported to bind to NRSE, we tested whether
lation of histones around NRSEANP. This TSA-induced in- NRSF binds to the NRSEANP. EMSAs carried out using ra-
crease in the acetylation level of histones around NRSEANP diolabeled NRSEANP as a probe demonstrated that NRSF
correlates with increased endogenous ANP gene expression does bind to the sequence. A dominant-negative mutant of
(Fig. 7A and B). Consistent with the increased expression of NRSF introduced into ventricular myocytes with the ANP re-
ANP mRNA, the level of secreted ANP protein was also porter gene restored the NRSE-mediated repression of ANP
higher in TSA-treated myocytes than in untreated cells (Fig. promoter activity. Furthermore, a GAL4-NRSF fusion protein
7C). TSA had no effect on the expression of the GAPDH gene, significantly repressed the activities of an ANP promoter car-
which is constitutively expressed in ventricular myocytes, indi- rying the tandem GAL4-binding sites in its upstream region.
cating that its effect on ANP expression is specific. Further- Thus, recruitment of NRSF to the ANP gene is apparently
more, the TSA-induced increase in endogenous ANP gene responsible for NRSEANP-mediated repression of ANP geneVOL. 21, 2001 NRSF AND CARDIAC EMBRYONIC GENE 2093
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 6. HDAC is involved in NRSE-mediated repression of the ANP promoter. (A) NRSE mutation decreases TSA inducibility of ANP
reporter genes. A total of 10 g of ⫹1769/⫺452hANP/Luc, ⫺452/⫹1769hANPLuc, ⫺452hANPLuc, mtNR/⫹1769/⫺452hANPLuc, or mtNR/
⫺452/⫹1769hANPLuc was cotransfected into ventricular myocytes with 4 g of pRL-TK, after which the cells were incubated for 48 h with or
without 10 ng of TSA per ml. The relative luciferase activities were normalized to the activity of Renilla luciferase driven by the TK promoter. The
relative luciferase activity of each construct transfected into myocytes treated with TSA was expressed as the fold increase over the value for
constructs transfected into myocytes not treated with TSA. The bars represent the mean ⫾ the SEM values of relative luciferase activities from
at least three separate assays carried out in triplicate. (B) NRSE mutation causes a marked increase in the level of acetylated histones associated
with the 3⬘-UTR of the ANP gene. Soluble chromatin preparations from myocytes transfected with ⫺452/⫹1769hANPLuc or mtNR/⫺452/
⫹1769hANPLuc were immunoprecipitated with an antibody raised against acetylated histone H3 or H4 (␣-AcH3 or ␣-AcH4, respectively) and
analyzed by quantitative PCR. Aliquots of chromatin obtained before immunoprecipitation were also analyzed (INPUT). Representative PCR
results are shown in the lower three panels. The graphs show the relative amounts of PCR product coimmunoprecipitated with acetylated H3 or
H4, corrected for the total amount of PCR product. The data are the mean ⫾ the SEM values from three independent experiments carried out
in duplicate. ⴱ, P ⬍ 0.05 versus mtNR/⫹1769/⫺452hANPLuc. The bars and numbers above the target genes represent the relative positions within
the genes and the lengths of the products amplified in the PCR analysis.2094 KUWAHARA ET AL. MOL. CELL. BIOL.
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 7. HDAC is involved in the regulation of endogenous ANP gene expression in ventricular myocytes. (A) TSA induces the acetylation of
histones associated with the 3⬘-UTR of the endogenous ANP gene. Soluble chromatin preparations from ventricular myocytes treated with or
without TSA for 24 h were immunoprecipitated using an anti-acetylated histone H3 or H4 antibody (␣-AcH3 or ␣-AcH4) and analyzed by
quantitative PCR. Aliquots of chromatin obtained before immunoprecipitation were also analyzed (INPUT). Representative PCR results are
shown in the lower three panels. The results shown in the top panel are expressed as the relative amount of PCR product coimmunoprecipitated
with acetylated H3 or H4 corrected for the total amount of PCR product. The data are the mean ⫾ the SEM values from three independent
experiments carried out in duplicate. ⴱ, P ⬍ 0.05 versus TSA(⫺). The bar and numbers above the target gene represent the relative position in
the ANP gene and the length of the product amplified in the PCR analysis. (B) TSA induces the endogenous ANP gene expression in ventricular
myocytes. Cells were incubated for 24 h with or without 10 ng of TSA per ml and then harvested for analysis by Northern blotting. Shown are
representative blots for ANP and GAPDH (G3PDH) mRNA from two independent experiments with identical results. (C) TSA increased the
secretion of ANP. Ventricular myocytes were incubated for 48 h with or without 10 ng of TSA per ml, and the concentration of ANP in the culture
medium was measured by RIA. The data are the mean ⫾ the SEM values from two independent experiments carried out in triplicate. ⴱ, P ⬍ 0.05
versus the control. (D) TSA-induced increase in endogenous ANP gene expression is attenuated in ventricular myocytes infected with Ad/DNNR.
Cells infected with control Ad/lacZ or Ad/DNNR were incubated for 24 h with or without 10 ng of TSA per ml and then harvested for analysis
by Northern blotting. Shown are representative blots for ANP mRNA from two independent experiments with identical results. (E) The graphs
show the relative ANP mRNA levels, which were determined by quantitation of the density of ANP mRNA bands using NIH image software and
by correction with the amount of 28S. The relative ANP mRNA level in ventricular myocytes infected with Ad/lacZ and treated without TSA was
assigned a value of 1.0. The data are the mean ⫾ the SEM values obtained from two independent experiments carried out in triplicate. The
numbers above each bar indicate the fold increase induced by TSA (10 ng/ml). (F) ET-1 increases the level of acetylation of histones associated
with the 3⬘-UTR of the endogenous ANP gene. Soluble chromatin preparations from cells treated with or without ET-1 for 24 h were
immunoprecipitated with ␣-AcH3 or ␣-AcH4 and analyzed by quantitative PCR. Aliquots of chromatin obtained before immunoprecipitation were
also analyzed (INPUT). Representative PCR results are shown in the lower three panels. The results shown in the top panel are expressed as the
relative amount of PCR product coimmunoprecipitated with acetylated H3 or H4, corrected for the total amount of PCR product. The data are
the mean ⫾ the SEM values obtained from three independent experiments carried out in duplicate. ⴱ, P ⬍ 0.05 ET (⫺).
expression. In fact, adenovirus-mediated expression of domi- tional NRSE (48). However, the functional role of NRSE in
nant-negative NRSF in ventricular myocytes increased endog- the expression of nonneuronal genes has not been elucidated.
enous ANP gene expression. Since ANP gene expression and synthesis of the protein are
NRSE was initially identified as a silencer element regulat- predominantly higher in the heart than in other tissues, our
ing neuron-specific expression of the SCG10 and type II so- data represent the first evidence that NRSE also functions as a
dium channel (NaII) genes in rat (28, 34) and is now known to repressor of nonneuronal genes, as well as of neuron-restricted
regulate neuron-specific expression of many neuronal genes (4, genes, in certain tissues. Of particular interest to us is the fact
12, 20, 22, 32, 33, 36, 42, 47, 48, 55). In addition, Schoenherr et that ANP is expressed in the same cardiac myocytes in which
al. described five nonneuronal genes that also contain func- NRSF is expressed, suggesting that repression by NRSF can beVOL. 21, 2001 NRSF AND CARDIAC EMBRYONIC GENE 2095
overcome by strong, cardiac-myocyte-specific activators. In- that whereas the mSin3-HDAC complex associates with the
deed, the ANP promoter has a number of heart-specific en- N-terminus repressor domain, the C-terminus repressor do-
hancer elements for several transcription factors, including main of NRSF acts by a mechanism unrelated to histone
Csx/Nkx2.5, SRF, and GATA4 (16, 30, 52). The balance be- deacetylation (18, 38). In that regard, CoREST, a novel core-
tween the activities of these enhancers and repressors, such as pressor, was recently reported to interact with the C-termi-
NRSF, should determine the level of ANP gene expression in nal domain (1). CoREST contains two SANT (SW13/ADA2/
ventricular myocytes. In addition, though the heart is a primary NcoR/TFIIIB) domains, the same structures found in the
site of the ANP synthesis, modest levels of ANP gene expres- NcoR/SMRT corepressors of steroid hormone receptors (17,
sion have also been detected in the hypothalamus. Whether 35, 41). It will be of great interest to delineate the mechanisms
NRSE plays a role in regulating the expression of the neuronal by which CoREST could function as a corepressor to com-
ANP gene remains to be determined. pletely resolve the mechanism by which NRSF represses ANP
NRSF has been reported to contain repressor domains at gene transcription in ventricular cells.
both its N and its C termini (38, 54). mSin3 associates with the NRSE is also apparently involved in the reactivation of ANP
N terminus and forms a complex with HDAC, thereby medi- gene expression during hypertrophic responses, since NRSE
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
ating repression through histone deacetylation (18, 39). Since mutation or expression of a dominant-negative NRSF substan-
NRSF was shown in C6 glioma and NIH 3T3 cells to repress tially reduced the capacity of ET-1 to induce ANP gene ex-
the expression of target genes in part via recruitment of the pression. This suggests that attenuation of NRSE-mediated
mSin3-HDAC complex (18, 39), we examined whether mSin3- repression of the ANP gene is an important component of the
HDAC is also involved in NRSE-NRSF-mediated repression ET-1-induced increase in ANP gene expression. Indeed, ET-1
in cardiac myocytes. Cotransfection and immunoprecipitation was found to increase the level of acetylation of histones
of myc-tagged NRSF and FLAG-tagged mSin3B showed that around NRSEANP, supporting the notion that ET-1 inhibits
the two proteins do associate in ventricular myocytes. Further, NRSE-NRSF-mediated repression of ANP gene transcription.
an introduced NRSF associated with endogenous mSin3 and The fact that ET-1 did not completely abolish NRSF-NRSE-
HDAC, which confirms that NRSF recruits mSin3-HDAC mediated repression means its inhibition of NRSF is incom-
complex in ventricular myocytes and suggests that NRSF re- plete. The details of the mechanisms by which ET-1 attenuates
presses ANP gene transcription, at least in part, via histone the repressor function of NRSF remain to be determined,
deacetylation. Consistent with that idea, inhibition of HDAC however. In neuronal cells, REST4, an alternatively spliced
using TSA increased the luciferase activity of ANP reporter variant of NRSF, was recently reported to be induced depen-
constructs carrying intact NRSEANP to a significantly greater dent on protein kinase A and to act in a dominant-negative
degree than those carrying mutant NRSEANP or lacking fashion (50). However, because we could not detect REST4
NRSEANP. Finally, ChIP assays showed that the level of acet- mRNA expression in ventricular myocytes in both basal and
ylation of histones associated with a reporter gene containing ET-1-stimulated condition, it is unlikely that ET-1 induces
intact NRSEANP was lower than that of histones associated REST4 and inhibits the repressor function of NRSF in ven-
with the gene containing mutant NRSEANP. tricular myocytes (Fig. 4F). In addition, ET-1 did not decrease
We also assessed the level of acetylation of the histones the expression level of NRSF mRNA in ventricular myocytes
around the NRSE of the endogenous ANP gene. In this case, (Fig. 4E). ET-1 may regulate the function of NRSF by post-
ChIP assays revealed that the level of histone acetylation was transcriptional mechanisms. In that regard, it was reported that
low under basal conditions and was significantly increased by Eve, a transcriptional repressor, is negatively regulated by
TSA. Corresponding Northern blots showed that the expres- phosphorylation (31). Phosphorylated Eve was unable to in-
sion of the endogenous ANP gene, but not the GAPDH gene, teract with the TATA-binding protein, a target for repression.
was upregulated with TSA. The upregulation of endogenous ET-1 may phosphorylate NRSF and attenuates its repressor
ANP gene expression by TSA was attenuated in ventricular function through the allosteric interference with a protein-
myocytes infected with Ad/DNNR. In summary, these results protein or protein-DNA interaction.
are consistent with a scenario in which the NRSE-NRSF sys- That ANP reporter constructs carrying the mutant NRSEANP
tem represses ANP gene expression in ventricular myocytes were still induced indicates that ET-1 also exerts effects on
through the recruitment of HDAC and the deacetylation of other regions. As shown in this study and by others, the ANP
histones. promoter region can also mediate ET-1-inducible ANP gene
Although HDAC is certainly involved in the mechanism by expression. The ANP promoter region contains elements for
which NRSF-NRSE represses ANP gene expression, the de- several transcriptional enhancers, including SP-1 and SRF,
tails of the mechanism by which recruited NRSF represses the which were reportedly involved in the induction of ANP gene
ANP promoter is not well understood. Relocation of the expression under various conditions (3, 16, 52). Thus, ET-1
NRSF-mSin3-HDAC complex to the core promoter region, appears to induce ANP gene expression through both upregu-
which enables the deacetylation of histones and, maybe, the lation of enhancer activities and downregulation of repressor
transcription factors bound there, would be essential for re- activities.
pression by the complex (19). It is known that mSin3A is able NRSE also exists in several other cardiac embryonic genes,
to bind the general transcription factor TFIIB, indicating the including those for BNP, skeletal ␣-actin, and the Na,K-
possibility that mSin3 may recruit the NRSF repressor complex ATPase ␣3 subunit (Fig. 8). Although the role of NRSE in the
to the core promoter region (57). Another possibility is that expression of the Na,K-ATPase ␣3 subunit gene is controver-
factors associating with the C-terminus region of NRSF inter- sial (42, 48), we recently confirmed that NRSE acts as a re-
acts directly with core promoter factors. It has been proposed pressor of BNP gene expression (E. Ogawa and Y. Saito,2096 KUWAHARA ET AL. MOL. CELL. BIOL.
Downloaded from http://mcb.asm.org/ on February 1, 2021 by guest
FIG. 8. NRSE-like sequences contained within other cardiac embryonic genes. The cardiac embryonic genes containing NRSE are listed.
unpublished data), and others have shown that, when trans- gene expression to neurons. Cell 80:949–957.
8. Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcrip-
lated in vitro, NRSF binds to the NRSE in the skeletal ␣-actin tion initiation by RNA polymerase II in a soluble extract from isolated
gene (48). Thus, NRSF may be an important regulator of the mammalian nuclei. Nucleic Acids Res. 11:1475–1489.
expression of various cardiac embryonic genes. Indeed, by con- 9. Drexler, H., J. Hänze, M. Finckh, W. Lu, H. Just, and R. E. Lang. 1989.
Atrial natriuretic peptide in a rat model of cardiac failure. Atrial and ven-
trolling the expression of several embryonic genes, NRSF may tricular mRNA, atrial content, plasma levels, and effect of volume loading.
participate in the establishment and maintenance of the ter- Circulation 79:620–633.
minally differentiated phenotype of ventricular myocytes. 10. Durocher, D., F. Charron, R. Warren, R. J. Schwartz, and M. Nemer. 1997.
The cardiac transcription factors Nkx2-5 and GATA-4 are mutual cofactors.
In conclusion, our results demonstrate the involvement of EMBO J. 16:5687–5696.
the NRSE-NRSF system in the regulation of basal and induc- 11. Edwards, B. S., D. M. Ackermann, M. E. Lee, G. S. Reeder, L. E. Wold, and
J. C. J. Burnett. 1988. Identification of atrial natriuretic factor within ven-
ible ANP gene expression in ventricular myocytes. This study tricular tissue in hamsters and humans with congestive heart failure. J. Clin.
also provides a new insight into a potential contribution of the Investig. 81:82–86.
NRSE-NRSF system in the normal physiology and pathophys- 12. Hahm, S. H., L. Chen, C. Patel, J. Erickson, T. I. Bonner, E. Weihe, M. K.
Schafer, and L. E. Eiden. 1997. Upstream sequencing and functional char-
iology of nonneuronal tissues. acterization of the human cholinergic gene locus. J. Mol. Neurosci. 9:223–
236.
13. Harada, M., H. Itoh, O. Nakagawa, Y. Ogawa, Y. Miyamoto, K. Kuwahara,
ACKNOWLEDGMENTS
E. Ogawa, T. Igaki, J. Yamashita, I. Masuda, T. Yoshimasa, I. Tanaka, Y.
We thank M. Okumura, C. Ohta, and Y. Arikuni for their excellent Saito, and K. Nakao. 1997. Significance of ventricular myocytes and non-
secretarial work. We also thank D. J. Anderson for giving us the myocytes interaction during cardiocyte hypertrophy: evidence for endothe-
lin-1 as a paracrine hypertrophic factor from cardiac nonmyocytes. Circula-
plasmids encoding myc-tagged NRSF and the dominant-negative form tion 96:3737–3744.
of NRSF and D. G. Gardner for his pertinent advice on the transfec- 14. Harsdorf, R., J. G. Edwards, Y. T. Shen, R. K. Kudej, R. Dietz, L. A.
tion method. Leinwand, B. Nadal-Ginard, and S. F. Vatner. 1997. Identification of a
This work was supported in part by research grants from the Japa- cis-acting regulatory element conferring inducibility of the atrial natriuretic
nese Ministry of Education, Science, and Culture; the Japanese Min- factor gene in acute pressure overload. J. Clin. Investig. 100:1294–1304.
istry of Health and Welfare; the Japanese Society for the Promotion of 15. Hasebe, N., L. Hittinger, S. Kohin, Y. T. Shen, R. M. Graham, and S. F.
Science Research for the Future program (JSPS-RFTF96I00204 and Vatner. 1995. Effects of hypertrophy on left atrial and ventricular compliance
JSPS-RFTF98L00801); and the Smoking Research Foundation. and plasma ANF levels in conscious dogs. Am. J. Physiol. 268:H781–H787.
16. Hines, W. A., J. Thorburn, and A. Thorburn. 1999. A low-affinity serum
response element allows other transcription factors to activate inducible
REFERENCES gene expression in cardiac myocytes. Mol. Cell. Biol. 19:1841–1852.
1. Andrés, M. E., C. Burger, M. J. Peral-Rubio, E. Battaglioli, M. E. Anderson, 17. Hörlein, A. J., A. M. Näär, T. Heinzel, J. Torchia, B. Gloss, R. Kurokawa, A.
J. Grimes, J. Dallman, N. Ballas, and G. Mandel. 1999. CoREST: a func- Ryan, Y. Kamei, M. Söerström, C. K. Glass, et al. 1995. Ligand-independent
tional corepressor required for regulation of neural-specific gene expression. repression by the thyroid hormone receptor mediated by a nuclear receptor
Proc. Natl. Acad. Sci. USA 96:9873–9878. co-repressor. Nature 377:397–404.
2. Arai, H., K. Nakao, Y. Saito, N. Morii, A. Sugawara, T. Yamada, H. Itoh, S. 18. Huang, Y., S. J. Myers, and R. Dingledine. 1999. Transcriptional repression
Shiono, M. Mukoyama, H. Ohkubo, et al. 1988. Augmented expression of by REST: recruitment of Sin3A and histone deacetylase to neuronal genes.
atrial natriuretic polypeptide gene in ventricles of spontaneously hyperten- Nat. Neurosci. 2:867–872.
sive rats (SHR) and SHR-stroke prone. Circ. Res. 62:926–930. 19. Imhof, A., and A. P. Wolffe. 1998. Transcription: gene control by targeted
3. Ardati, A., and M. Nemer. 1993. A nuclear pathway for alpha 1-adrenergic histone acetylation. Curr. Biol. 8:R422–R424.
receptor signaling in cardiac cells. EMBO J. 12:5131–5139. 20. Ishiguro, H., K. S. Kim, and T. H. Joh. 1995. Identification of a negative
4. Bessis, A., N. Champtiaux, L. Chatelin, and J. P. Changeux. 1997. The regulatory element in the 5⬘-flanking region of the human dopamine beta-
neuron-restrictive silencer element: a dual enhancer/silencer crucial for pat- hydroxylase gene. Brain Res. Mol. Brain Res. 34:251–261.
terned expression of a nicotinic receptor gene in the brain. Proc. Natl. Acad. 21. Izumo, S., B. Nadal-Ginard, and V. Mahdavi. 1988. Protooncogene induc-
Sci. USA 94:5906–5911. tion and reprogramming of cardiac gene expression produced by pressure
5. Chen, Z. F., A. J. Paquette, and D. J. Anderson. 1998. NRSF/REST is overload. Proc. Natl. Acad. Sci. USA 85:339–343.
required in vivo for repression of multiple neuronal target genes during 22. Kallunki, P., G. M. Edelman, and F. S. Jones. 1997. Tissue-specific expres-
embryogenesis. Nat. Genet. 20:136–142. sion of the L1 cell adhesion molecule is modulated by the neural restrictive
6. Chien, K. R., K. U. Knowlton, H. Zhu, and S. Chien. 1991. Regulation of silencer element. J. Cell Biol. 138:1343–1354.
cardiac gene expression during myocardial growth and hypertrophy: molec- 23. Kanegae, Y., G. Lee, Y. Sato, M. Tanaka, M. Nakai, T. Sakaki, S. Sugano,
ular studies of an adaptive physiologic response. FASEB J. 5:3037–3046. and I. Saito. 1995. Efficient gene activation in mammalian cells by using
7. Chong, J. A., J. Tapia-Ramirez, S. Kim, J. J. Toledo-Aral, Y. Zheng, M. C. recombinant adenovirus expressing specific Cre recombinase. Nucleic Acids
Boutros, Y. M. Altshuller, M. A. Frohman, S. D. Kraner, and G. Mandel. Res. 23:3816–3821.
1995. REST: a mammalian silencer protein that restricts sodium channel 24. Kanegae, Y., M. Makimura, and I. Saito. 1994. A simple and efficientYou can also read
