(Received for publication, March 10, 1997, and in revised form, May 5, 1997)
From the Center for Cardiovascular and Muscle Research, Department of Anatomy and Cell Biology, State University of New York, Brooklyn, New York 11203
The expression of the cardiac myosin light chain
2 (MLC2) gene is repressed in skeletal muscle as a result of the
negative regulation of its transcription. Two regulatory elements, the cardiac specific sequence (CSS) located upstream (360 base pairs) and
a downstream negative modulatory sequence (NMS), which function in
concert with each other, are required for repression of the MLC2
promoter activity in skeletal muscle. Individually, CSS and NMS have no
effect. Transient transfection analysis with recombinant plasmids
indicated that CSS- and NMS-mediated repression of transcription is
position- and orientation-dependent and is transferable to heterologous promoters. A minimal conserved motif, GAAG/CTTC, present
in both CSS and NMS, is responsible for repression as the mutation in
the core CTTC sequence alone was sufficient to abrogate its repressor
activity. The DNA binding assay by gel mobility shift analysis revealed
that one of the two complexes, CSSBP2, is significantly enriched in
embryonic skeletal muscle relative to cardiac muscle. In extracts from
adult skeletal muscle, where the cardiac MLC2 expression is suppressed,
both complexes, CSSBP1 and CSSBP2, were present, whereas the cardiac
muscle extracts contained CSSBP1 alone, suggesting that the protein(s)
in the CSSBP2 complex accounts for the negative regulation of cardiac MLC2 in skeletal muscle. A partial cDNA clone (Nished) specific for
the candidate repressor factor was isolated by expression screening of
the skeletal muscle cDNA library by multimerized CSS-DNA as probe.
The recombinant Nished protein binds to the CSS-DNA, but not to
CSS-DNA where the core CTTC sequence was mutated. The amino acid
sequence of Nished showed a significant structural similarity to the
sequence of transcription factor "runt," a known repressor of gap
and pair-rule gene expression in Drosophila.
The acquisition of the differentiated phenotype of eukaryotic cells is a consequence of activation of tissue-specific genes and repression of other genes, both of which are precisely controlled during development of multicellular organisms (1). Since only a small population of genes is expressed at any given time in the differentiated cell, it is becoming increasingly clear that the mechanisms by which genes are repressed are as important as those that activate them (2-4). Repression of transcription is commonly achieved via binding of the negative regulators to cis-elements where the degree of repression is controlled by the location and/or orientation (5-9), structure (10), or copy number (11) of the regulatory elements. Other mechanisms of transcriptional repression involve occupation of the activator binding site (12-14), squelching of factors via protein-protein interaction (15), and formation of a dimeric complex that binds DNA but lacks the activation domain (16-18).
The coordinate expression of muscle-specific genes during myogenesis in differentiated myocytes suggests the existence of a tightly controlled regulatory program involving a cascade of expression of specific positive and negative transcription factors (19). We have shown previously that the tissue-specific expression of the chicken cardiac myosin light chain 2 (MLC2)1 gene is regulated by both positively and negatively acting cis-elements and their cognate DNA-binding factors (20-24). The regulation is due to the activators CArG box (20, 25) and the myocyte enhancer factor 2 binding sites (26), and a negative regulatory region, cardiac specific sequence (CSS), responsible for repression of cardiac MLC2 transcription in skeletal muscle (21). Removal of CSS alone restores cardiac MLC2 expression in skeletal muscle without impairing its function in cardiac muscle cells (21). An upstream negative regulatory domain distinct from CSS also exists in the rat MLC2 gene promoter, mutation of which led to ectopic expression of the gene in transgenic animals (27).
In this report, we have delineated the regulatory domains within CSS essential for repression of the cardiac MLC2 promoter in skeletal muscle. There are three distinct protein binding sites, CSS-A, CSS-B, and CSS-C, each of which contains a common sequence motif, GAAG/CTTC. Mutation in the CTTC motif in CSS-B alone was sufficient to abrogate totally both DNA-protein interaction and the inhibitory function of CSS. CSS-mediated repression, however, requires the presence of another downstream sequence element, the negative modulatory sequence (NMS), which also contains the conserved GAAG motif and serves as the binding site for nuclear proteins. Neither of the two motifs alone can repress transcription. The CSS/NMS-binding protein complex, CSSBP2, is present at a significantly higher level in nuclear extracts from skeletal muscle relative to cardiac muscle, suggesting that CSSBP2 binding to CSS/NMS accounts for repression of the MLC2 promoter function in skeletal muscle. In an attempt to identify the protein(s) involved in the repression mechanism, we have isolated a partial cDNA clone (Nished; Sanskrit for "negative") by expression screening of the cDNA library derived from chicken skeletal muscle mRNA by multimerized CSS-DNA as a probe. The predicted amino acid sequence of Nished shows a significant similarity to two previously described repressors, runt from Drosophila (28) and SP3 from human (29).
To construct CSS-containing muscle
creatine kinase (MCK) promoter/reporter recombinant, the proximal MLC2
promoter (160 to +158) in pMLC410CAT was replaced by the 246-bp
HindIII fragment containing the MCK basal promoter from
plasmid E4 (30). MCK enhancer contained within a 300-bp
BamHI restriction fragment was then introduced into the
unique BamHI site to generate MCKCSS. Likewise, the proximal
MLC2 promoter (
160 to +158) of pMLC410CAT was replaced by the
2,500-bp MCK promoter/enhancer fragment (30) to construct MCK2.5CSS.
The HindIII-PstI restriction fragment from
pMLC371CAT that contains the CSS domain was cloned upstream to the
pANG700 promoter (31) and designated as pANG700CSS. To construct
pANGCSS, the HindIII-PstI fragment from
pMLC371CAT was cloned into the plasmid pBasic CAT (Promega) in the
HindIII and PstI sites in the polylinker. The
264-bp NciI-XbaI fragment of ANG proximal
promoter was then cloned downstream to CSS in the polylinker at
SalI-XbaI to produce pANGCSS. The PstI
fragment (
130 bp to +40 bp) from the MLC2 promoter was cloned into
the PstI site in pBasic CAT to construct PST. A 160-bp
HindIII-PstI fragment containing the CSS domain
was spliced in the corresponding sites in the polylinker of pBasic CAT;
the resultant plasmid was then linearized with PstI and
ligated to the PstI fragment (
160 to +158) containing the
MLC2 promoter to create plasmid CSSPST. A synthetic oligonucleotide
harboring the NMS sequence (GAAG) flanked by XbaI was
inserted downstream of the MLC promoter in plasmid PST to create
CSSPSTNMSCAT. MLC2 proximal promoter (
125 to + 158 bp) obtained as an
HindIII fragment was cloned into the HindIII site
in the polylinker of the basic luciferase reporter plasmid (Promega) to
create MLCLUC. An oligonucleotide with HindIII ends,
spanning the 50-bp CSS element
(5
-agcttccattgtgaaggacgagggggtacttctaccctgaagcaaaagga-3
) was cloned
into the HindIII site of polylinker of pBluescript-SK+. Plasmid with the CSS domain in forward and reversed orientations was
digested with restriction enzymes SacI and XhoI,
and the restriction fragments were cloned in the corresponding sites in
the polylinker in MLCLUC. The resultant plasmids were designated
MLCCSSLUC and MLC3
CSSLUC, respectively. An oligonucleotide containing
the CSS sequence with nucleotides altered to introduce specific
mutations and flanked by HindIII and XhoI sites
was synthesized. The resultant oligonucleotides were ligated to
SmaI and XhoI sites in the polylinker in MLCLUC
to ensure directed ligation. The resultant plasmid was designated
MLC
BCSS.
Skeletal
muscle tissue was collected from the leg (soleus) muscle of 13-day-old
chicken embryo, digested with 0.1% pancreatin, and cells were
suspended in complete medium (60% Waymouth, 40% Hanks' balanced salt
solution medium with 15% horse serum, 2% chicken serum) and filtered
through sieves of 90 µm and 45 µm sequentially. Preplating was done
three times, for 1 h each, to facilitate differential removal of
fibroblasts. Cells were plated at a density of 2 × 106 cells/10-cm dish. Cultures were re-fed with fresh
medium before transfection. Transfections were done using calcium
phosphate precipitation (32). CAT enzyme activity in the extracts was assayed as described previously (20, 21, 23, 33). Plasmid SV40
luciferase (Promega) was used to normalize against fluctuations in
plasmid uptake and expression. Cells were lysed in 1 × lysis buffer (Promega), and the luciferase activity was assessed according to
instructions provided by Promega in a monolight luminometer. For
normalization in the uptake of luciferase reporter plasmids, -galactosidase activity under the pCMV promoter was used as an internal control.
Tissues from embryonic and adult heart and skeletal muscles were minced finely, and the dissociated cells were lysed in lysis buffer (20 mM Hepes, pH 7.6, 20% glycerol, 10 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.1% Triton X-100), 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and leupeptin (10 µg/ml) with a Wheaton dounce homogenizer. Nuclei were collected, and protein extracts were prepared as described previously (23, 33).
Gel Electrophoretic Mobility Shift Assay (GMSA)Double-stranded oligonucleotide was end labeled with
[-32P]ATP and 0.5 ng of labeled oligonucleotide,
incubated with 2 µg of poly(dI·dC), 1-12 µg of protein in 20 mM Hepes, pH 7.5, 3% glycerol, 1.5 mM
MgCl2, 1 mM dithiothreitol, 2 mM
EDTA, and 50 mM KCl at 4 °C for 30 min. Competitor DNA
was added in a 100-fold excess following incubation for 30 min on ice,
and the reaction mixtures were analyzed by electrophoresis as described
previously (23).
A footprinting assay was performed using nuclear extract from embryonic and adult skeletal muscle essentially as described earlier (24). A 160-bp EcoRI-XhoI fragment containing the CSS domain was incubated with nuclear extract (10-60 µg of protein) from embryonic and adult cardiac and skeletal muscle in a 50-µl reaction buffer containing 20 mM Hepes, pH 7.9) 5 mM MgCl2, 0.1 mM EDTA, 50 mM KCl, 0.5 mM dithiothreitol, and 10% glycerol. Following incubation at room temperature freshly diluted DNase-I (1 µg/ml) was added and then allowed to incubate for 60 s. DNA was extracted with phenol/chloroform and analyzed on an 8% sequencing gel. For footprinting of the NMS domain, a PstI-HindIII fragment spanning the NMS region was labeled by end filling.
Screening of Expression Library from Skeletal Muscle with CSS-50Chicken skeletal muscle cDNA expression library cloned
in the Zap expression vector obtained from Stratagene was screened according to Singh et al. (34) using the multimer (4 ×) of
CSS-50. For expression of the recombinant protein in Escherichia
coli, expression vector pET29B (Novagen) was digested with
BamHI and XhoI ligated with cDNA insert
obtained as a BamHI-XhoI insert from phagemid
Bluescript and transformed into host cells BL(21)DE3. Induction was
done according to manufacturer's (Novagen) instructions.
Poly(A)+ was made from cardiac and skeletal muscle of an adult chicken following the FastTrack mRNA isolation kit (Invitrogen). Briefly, 1 g of tissue was homogenized in 15 ml of lysis buffer containing RNase protein degrader and incubated at 450 °C for 60 min. The lysate was centrifuged at 4,000 × g for 5 min at room temperature. The supernatant was recovered and 950 µl of 5 M NaCl added. The DNA was sheared using a 18-21-gauge needle, and oligo(dT) cellulose pellet was added and incubated with gentle rocking for 60 min at room temperature. Then, the samples were centrifuged at 3,000 × g for 5 min, and the supernatant was carefully removed from the oligo(dT) bed. The beds were washed three times with 20 ml of binding buffer; the supernatant was removed after a spin of 3,000 × g for 5 min, followed by a wash in low salt buffer (three times). The poly(A)+ was eluted in 200 µl of elution buffer and then precipitated with cold ethanol and 0.1 volume of 3 M NaOAc. 10 µg of poly(A)+ from cardiac and skeletal muscle was loaded in 1.3% agarose/formaldehyde in 1 × MOPS, as described previously (32, 33).
Previous studies in our laboratory (21) have identified
a negative regulatory region, CSS, located between 371 and
282 bp
in the chicken cardiac MLC2 gene promoter, which is required for
repression of cardiac MLC2 gene transcription in skeletal muscle cells.
To test the potential role of CSS in repressing the transcription of
heterologous promoters such as skeletal MCK (30) and the non-muscle rat
ANG (31) promoters, we used plasmid pMCKCSS and pANGCSS containing CSS
in the respective promoters in a transient transfection assay in
skeletal muscle cells in culture. pMCKCSS expression was repressed
effectively (80%) compared with that of parent plasmid pMCK without
CSS (Fig. 1A). When CSS was placed 2.5 kilobases upstream to the MCK promoter (see "Materials and
Methods"), the activity of the resultant plasmid, MCK2.5CSS, was the
same as the parent plasmid pMCK2.5, suggesting that repression of
transcription by CSS is position-dependent (Fig.
1A). Similar results were obtained when we tested the rat
ANG promoter, which is expressed optimally in liver and at a lower
level in skeletal muscle. As shown in Fig. 1B, pANGCSS
activity was repressed significantly (52%) relative to the level of
the ANG promoter lacking CSS (pANG). The expression of plasmid
containing CSS placed 700 bp upstream in the ANG promoter (pANG700CSS)
was not repressed. When CSS was fused to the reporter plasmid carrying
the thymidine kinase promoter, no repression was obtained (Fig.
1C), suggesting that the repression mechanism might require
other regulatory sequence(s) involved in the CSS-mediated inhibition of
transcription (see below).
Identification of Protein Binding Domains in CSS
To identify
the nucleotide sequence within CSS involved in protein-DNA interaction,
a DNase-I footprinting assay was performed using a 160-bp fragment that
contains the CSS domain (410 to
250) and nuclear extracts from
adult and embryonic (13-day-old) chicken skeletal (soleus) muscle (Fig.
2). Three protected regions (CSS-A:
5
-CCATTGTGAAGGAC-3
; CSS-B: 5
-GATACTTC-3
; and CSS-C: 5
-CTGAAGCAAAGG-3
), which together span between
360 and
310, and
at least one hypersensitive site were detected. A common nucleotide sequence motif, GAAG, is present in regions A and C and the
complementary sequence, CTTC, in region B. We have denoted this region
(
360 to
310) as CSS-50. The sequence GAAG/CTTC was also identified as a regulatory motif in the negative regulatory domains of several other genes (27, 35-41), suggesting that the motif plays a conserved role in negative regulation of transcription.
To define further the role of CSS-50, GMSAs were done with nuclear
extracts from embryonic heart and skeletal muscles (Fig. 3A). Two well defined complexes, CSSBP1 and
CSSBP2, were formed with extracts from both embryonic tissues. The
intensity of the fast migrating complex, CSSBP2, was markedly higher in
skeletal muscle compared with the corresponding complex in cardiac
muscle. Both complexes were competed out by CSS and by an
oligonucleotide IRE (intron-responsive) containing an activator element
in the first intron of the MLC2 gene, which also contains the GAAG/CTTC motif. When nuclear extracts from the adult (4 weeks old) skeletal and
cardiac muscle tissues were compared as above, only CSSBP1 binding
activity was present in the cardiac muscle, whereas the skeletal muscle
extract contained both CSSBP1 and CSSBP2 complexes (Fig.
3A). Since cardiac MLC2 is down-regulated in skeletal
tissue, one may speculate that the relative abundance of CSSBP2
accounts for negative regulation of MLC2 gene in skeletal muscle. A
barely visible complex was observed when CTTC in CSS-B was mutated to GGTC (CSSB) (Fig. 3B), suggesting that the binding
activity of CSS-50 is primarily due to the CSS-B sequence.
The CTTC/GAAG Motif Is Essential for Repression
To
investigate whether the sequence contained in CSS-50 alone is
sufficient for repression or whether it acts in concert with other
elements, we used the minimal basal MLC2 promoter contained in the
121 to +158 fragment containing the three cis-regulatory elements, CArG box, myocyte enhancer factor 2 site, and TATA box, which
are required for its optimal expression in cardiac muscle cells (21,
23). As shown in Fig. 4, the presence of the CSS-50 alone produced an effective (70%) inhibition of MLC2 transcription, as
measured by the luciferase assay. Recombinant plasmid MLCCSS3
LUC, with
CSS-50 in reverse orientation, was totally ineffective, suggesting that
CSS is not a conventional silencer, as it is both position- and
orientation-dependent. Since it was reported that the
regulatory sequences of myocyte enhancer factor 2 (44), SP1 (45) and E
box (25) have an additive effect on transcription when present in
multiple copies, we made recombinants with four copies of CSS, arranged
in tandem, placed upstream in plasmid pMLCLUC. However, the presence of
multiple copies of CSS in plasmid pMLC4XCSSLUC did not cause additional
repression of transcription compared with repression observed with the
single CSS copy in pMLCCSSLUC. When a substitution mutation (CTTC
GGTC) was introduced in window B in plasmid pMLC
BCSSLUC, the
repression due to CSS was disrupted, and the expression level of the
mutant plasmid reached 80% of the activity of the parent plasmid
pMLCLUC. Additional mutations in the reverse complement GAAG sequence
in window C exhibit no further loss of repression (data not shown).
Since mutation in CSS-B window alone caused a loss of protein binding
in CSS-50, it would appear that the conserved sequence CTTC in CSS-B is
the primary target for binding the repressor protein(s), which also accounts for its loss of function upon mutation.
CSS-mediated Function Is Dependent upon a Downstream NMS
The
plasmid pMLCLUC, used in the experiments above, contains the downstream
sequence up to +158. An examination of the downstream sequence revealed
that the conserved GAAG/CTTC motif is also present once in the
5-untranslated region at +60 in the MLC2 gene. To test the potential
involvement of this motif in repression, we constructed the plasmids
PST and CSSPST, containing the basal promoter alone extending to +42,
without and with CSS, respectively (see "Materials and Methods")
(Fig. 5). Plasmid CSSPST, which lacks the downstream
sequence containing the GAAG motif but has the upstream CSS element,
was surprisingly as active as PST, which lacks both the upstream and
downstream elements. This would mean that CSS alone was unable to
repress MLC2 transcription in skeletal muscle cells and suggests the
requirement of the downstream GAAG motif, which was absent in both
constructs. To test this possibility, we inserted the GAAG contained in
a short (17-mer) oligonucleotide (5
-tctagacctagaagacttctaga-3
)
downstream of the MLC2 promoter in CSSPST (see "Materials and
Methods"). The resultant plasmid CSSPSTNMS was active in repression
of the promoter activity almost to the same extent as the native
sequence NMS containing the motif in plasmid pMLC371. The downstream
motif alone (PSTNMS) was inactive as repressor. These findings thus
confirm the requirement of the downstream GAAG sequence which was
supported further when insertion of another cis-regulatory
sequence, the IRE region, present in the first intron of the MLC2 gene
containing GAAGCTTC motif, also led to repression of the
promoter.2
To test whether NMS is recognized by DNA-binding proteins, we performed
the DNase-I footprinting assay with the
PstI-HindIII fragment, which includes the NMS
region spanning from +40 to +158. As shown in Fig.
6A, there was a protected window
corresponding to CTTCATG in the noncoding strand. Although a partial
protection also occurred in regions flanking the CTTCTGA sequence, the
repression activity in plasmid CSSPSTNMS containing only the protected
GAAG sequence (see Fig. 5) was sufficient to restore repression of the
MLC2 promoter activity. The sequence flanking the GAAG motif has no
resemblance to the native sequence flanking GAAG in the MLC2 gene.
Additional analysis of the NMS-DNA-binding protein interaction was done
by GMSAs with an oligonucleotide containing the NMS domain (+40 to +70)
and nuclear extracts from the embryonic and adult heart and skeletal
muscles. As seen in Fig. 6B, the NMS-binding proteins are
present in both cardiac and skeletal embryonic tissues. As with the
CSS-binding proteins, the NMSBP1 binding activity was abundantly
present in the skeletal extracts but to a lesser extent in cardiac
extracts. The specificity of the complex formation was established by
self-competition, with CSS and another GAAG-containing regulatory
domain, IRE, all of which competed out the complex formation.
Isolation of a cDNA Clone (Nished) for CSS-binding Protein
As a prelude to identifying the CSS- and NMS-binding
proteins and delineating their role in repression of transcription, we isolated a cDNA clone by screening an expression cDNA library made from the adult chicken skeletal muscle mRNA (Stratagene) with
a high affinity DNA containing the multimerized CSS-50 element. Of
several candidate clones, one with the largest insert was identified as
a candidate for CSS-specific protein coding cDNA clone, referred to
from here on as Nished, which contains an insert of 1,217 bp in length.
The nucleotide and the predicted amino acid sequences are shown in Fig.
7. Plasmid pET29Nished, containing the Nished cDNA
in-frame with ATG, was introduced into host strain BL21(DE23) and
induced by isopropyl-1-thio--D-galactopyranoside
(IPTG). Extracts from IPTG-induced and uninduced bacteria harboring
Nished were examined by GMSAs using CSS-50 as a probe. As seen in Fig. 8, a new protein-DNA complex (see arrow) was
formed with the IPTG-induced extract which was absent in the uninduced
extract. Both extracts showed a nonspecific DNA binding activity
(top band). The specificity of binding was demonstrated when
CSS
B (
CSS) used as a probe failed to bind the
recombinant protein Nished. A Northern blot using adult skeletal
mRNA (Fig. 9) indicated that the Nished mRNA of
approximately 1,600 nucleotides was present in skeletal muscle RNA, but
a barely detectable signal of approximately the same size was seen in
cardiac muscle mRNA. Based on size estimation of the mRNA, the
cDNA clone lacks about 300 bp consisting the 5
-untranslated and
coding sequences. When analyzed against the available data bases
(SwissProt) using BLAST program, the Nished sequence showed a sequence
homology with two repressors, a zinc finger protein SP3 (29) and the
Drosophila melanogaster repressor runt protein (28). There
are three regions of similarity between Nished and runt: the amino
terminus of runt which is characteristically rich in alanine and
glutamine residues, the runt homology domain including the ATP-binding
domain, and the carboxyl region of the runt protein which is not highly
conserved among the other family members. The positively charged
residues are clustered mainly toward the carboxyl end of Nished and
overlap with the runt homology domain within Nished, suggesting that
this region could serve as the DNA binding motif in the protein.
Further characterization of Nished is presently in progress.
The physiological expression of the chicken cardiac MLC2 is restricted to cardiac muscle (44-46). In this report, we identified the functional sequence elements within the upstream negative regulatory domain (CSS) essential for repression of the cardiac MLC2 gene expression in skeletal muscle (21). In addition, a downstream sequence domain, NMS, was identified which functions in concert with CSS. The conserved CTTC motif in CSS-B alone is essential for CSS-mediated repression of the cardiac MLC2 promoter, as the substitution mutation in this motif caused the loss of DNA-protein interaction and led to a significant reduction in CSS-mediated repression of transcription. This motif is also present in the negative regulatory element (HF3) of the rat cardiac MLC2 promoter, mutation of which led to ectopic expression of the reporter luciferase activity in transgenic animals (27). A comparison of the nucleotide sequence of CSS and NMS with other genes indicated that GAAG/CTTC motif common to both elements is present in other known negative regulatory domains of several genes, such as retinoblastoma, human hypoxanthine phosphoribosyltransferase, yeast RME1, mouse albumin, rat MLC2, human neutrotropic papovavirus JC virus late promoter (27, 28, 37, 41). Thus, it appears that the negative regulation of transcription, at least in the genes tested here, is mediated by a conserved sequence element (GAAG/CTTC). The prevalence of this motif in disparate promoters suggests that a common protein, presumably a family of proteins, is responsible for down-regulation of the respective genes. In this context, CSSBP appears to be the member of an active repressor family of proteins that have intrinsic repression activity by inhibition of transcription initiation. Repression is presumably achieved with specific modulation in the level of CSSBPs, most likely of CSSBP2, which exhibits increased binding activity in the adult skeletal muscle where the expression of cardiac MLC2 is optimally repressed.
Different mechanisms are used for the negative regulation of eukaryotic genes (for review, see Ref. 36). One of them involves interference with the binding of activators that recognize sequences that overlap the binding sequences of negative factors. The fact that CSS and NMS do not overlap with any known activator site in the MLC2 gene, nor do they serve as binding sites for other known regulatory factors, suggests that the mechanism of negative regulation based on binding of factors to the same sequence, but with the opposite effect (47-50), or via steric hindrance due to overlapping sites (51), can be excluded. For repressors, as for activators, interaction with the components of the basal transcription machinery is common, although not essential (52). Some factors can repress transcription of certain promoters but not of others, suggesting the involvement of specific intermediary factors or cofactors (16). It is not uncommon for transcription factors to have different effects on transcription based on different locations within the same promoter or in different promoters. Some repressors become active only when they are overexpressed (17, 36, 54-56). CSS appears to function when positioned proximal to the promoter but requires, at the same time, the downstream GAAG motif. Thus the tissue-specific expression of the MLC2 gene is not the function of CSS alone; it depends on cooperative interaction with NMS.
Cooperation among multiple regulatory domains is, in general, the basis of negative regulation in several genes (36, 52, 55, 57-59), although bipartite promoter elements have also been identified in negative regulation (38, 60, 61). Interaction between distinct domains, presumably by virtue of the folding of DNA, leads to repression of transcription, such as of the insulin-like growth factor I receptor under negative regulation by Wilms' tumor-1 gene (62). The effect of Wilms' tumor-1 depends on the number of Wilms' tumor-1 binding sites located both upstream and downstream of the insulin-like growth factor I receptor transcription initiation site (63). Repressors that pose interference with the function and assembly of initiation complex without binding DNA have also been described (47-50). However, in our case, at least two protein-binding DNA motifs, located upstream and downstream of the transcription initiation site, are needed for repression, a situation analogous to the requirement for multiple sites for silencing of insulin-like growth factor I receptor or vimentin gene expression (63, 64). Interaction of the putative repressor protein(s) with multiple sites (CSS and NMS) located on both sides of the transcription initiation site might be a requirement for the physical and structural rigidity needed for the negative control of transcription initiation.
To understand further the CSS-mediated repression, we obtained a partial 1.2-kilobase cDNA encoding the CSS binding factor, Nished. Although the evidence on its involvement in repression of transcription is not yet available, the primary sequence analysis of the Nished cDNA revealed a similarity to other known repressors, such as runt of Drosophila melanogaster (28, 65), and SP3 (29). Runt is required for suppression of certain abdominal genes to limit the domains of engrailed expression in parasegmental pattern in Drosophila (16, 53). It has a conserved domain RHD, with an ATP binding box located within, which is essential for DNA binding and heterodimerization. Thus, Nished, based on its sequence similarity to these functional domains, appears to belong to a family of transcriptional inhibitors. Activation of Nished binding activity in skeletal muscle may be an important prerequisite for suppression of cardiac MLC2 gene expression in skeletal muscle. It is not yet known whether Nished is one of the two CSSBPs observed in skeletal nuclear extracts. The generation of Nished-specific antibodies, presently in progress in our laboratory, may help in elucidation of its role in repression mechanism(s) underlying the tissue-specific expression of cardiac MLC2.
The sequence reported in this paper has been submitted to NCBI BankIt113173 with accession number AF003093.