| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Institut für Pharmakologie und Toxikologie, Universität Münster, D-48129 Münster, Germany
b Abteilung für Molekulare Pharmakologie, Universität Göttingen, D-37070 Göttingen, Germany
c Klinik und Poliklinik für Thorax-, Herz- und Gefäßchirurgie, Universität Münster, D-48129 Münster, Germany
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C-X, in the human heart. Overexpression of CREM-IbC-X decreased cAMP response element (CRE) -mediated gene transcription in HIT-T15 cells, and this activity was assigned to the part of the sequence encoding putative internally translated proteins. Two of three possible internally translated proteins were immunologically identified in cells overexpressing CREM-IbC-X tagged with the hemagglutinin epitope of the influenza virus. Both proteins were expressed in bacteria and showed CRE-specific DNA binding, formation of heterodimers with the cAMP response element binding protein (CREB), and inhibition of CREB's binding to the CRE. CREM expression was detected on the mRNA and protein levels in the human heart. We conclude that CREM-IbC-X generates internally translated repressors of CRE-mediated gene transcription, suggesting the first example for the existence and function of human cardiac CREM.—Müller, F. U., Bokník, P., Knapp, J., Neumann, J., Vahlensieck, U., Oetjen, E., Scheld, H. H., Schmitz, W. Identification and expression of a novel isoform of cAMP response element modulator in the human heart. FASEB J. 12, 1191–1199 (1998)
Key Words: cyclic AMP • transcription factor • DNA binding proteins • base sequence
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In end-stage human heart failure, excessive ß-adrenergic stimulation by enhanced endogenous catecholamines is hypothesized to contribute to expressional alterations of myocardial regulatory proteins (2). In patients with heart failure, long-term therapy with ß-adrenoceptor antagonists was able to restore expressional changes (3, 4) and to improve prognosis and/or clinical signs of heart failure (5, 6). In rats, chronic infusion of the ß-adrenoceptor agonist isoproterenol led to similar myocardial expressional changes on the protein, mRNA, and transcriptional levels as found in human heart failure (ref 7 and references therein). To investigate cAMP-dependent transcription factors that might contribute to human cardiac expressional control after ß-adrenergic stimulation, we studied the expression of CREM in the human heart. We amplified a novel CREM isoform (CREM-IbC-X) by reverse transcription-polymerase chain reaction (RT-PCR) from human cardiac RNA. We show that internally translated repressors of CRE-mediated transcription can be generated from CREM-IbC-X, which can inhibit CREB's binding to the CRE and form heterodimers with CREB. CREM-IbC-X is the first example of the expression and function of CREM in the human heart suggesting a physiological role in the human myocardium.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C-X mRNA
Plasmid construction
CREM-IbC-X (BHIb, bases 4–425) and deletional constructs were subcloned in the vector pBK-CMV (Stratagene, La Jolla, Calif.) for eukaryotic expression of inserts controlled by the cytomegalovirus immediate early promoter. Subcloning was performed using the restriction sites NheI and PstI introduced by the PCR primers for amplification of the respective inserts. One construct, BH, had a deletion of bases 190–425 leading to exclusive expression of the truncated protein BH (see Results and 
Fig. 1C).
The construct HIb with a deletion of bases 4–134 only allowed expression of putative internally initiated proteins ( Fig. 1C). A further deletion of bases 153–162 of HIb was achieved by linearization at the SacI restriction site at base 159, digestion with Mung bean exonuclease I, and blunt-end religation, resulting in the construct HIb. This construct permitted complete internal translation initiated at base 189 whereas translation from AUG 135 was blocked at base 189. For the immunological detection of internal translated proteins, the nucleotide sequence encoding the carboxy-terminal 10 amino acids of the HA epitope (YPYDVPDYASL) (9) was inserted in frame between the last amino acid of exon Ib (=Y) and the stop codon in the full-length construct of BHIb in pBK-CMV. For expression of different translation products in bacteria, cDNAs corresponding to BH (bases 4–189), HIbI (bases 135–425), and HIbII (bases 189–425) were subcloned in the bacterial expression vector pSE420 (10) using primers introducing restriction sites for NcoI and XbaI. Subcloning, transformation of Escherichia coli cells, and plasmid isolation were performed according to standard protocols. All constructs were sequenced in both directions to confirm the identity and orientation of the inserts.
|
Cell culture and transient transfection of HIT-T15 cells
HIT-T15 (hamster insulinoma tumor) cells (11) were cultured and transfected as described (12). The indicator plasmid 4xSomCRE was a luciferase reporter gene construct controlled by a tetramerized rat somatostatin CRE sequence (12). Cells were transiently transfected with 2 µg indicator and 6 µg expression plasmid per 6 cm dish. Transfection efficiency was controlled by cotransfection with 0.4 µg per dish RSV-CAT. Cells were treated with either 10 µM forskolin or solvent (dimethylsulfoxide) as control 6 h before cell harvesting. CAT and luciferase assays were performed as described (12). There were no significant differences in the transfection efficiency between the different groups. Values are means from three or four independent transfection experiments, each performed in duplicate. Luciferase activity was expressed relative to the appropriate control. In a second set of experiments, cells were transfected as described above with the construct HIb in comparison with the control. A value of P < 0.05 vs. control in analysis of variance, followed by Bonferroni's t test, was considered significant. To obtain a concentration response curve, 1, 3, 6, or 9 µg effector DNA (pBKCMV without insert or containing CREM-IbC-X) was cotransfected with indicator plasmid and RSV-CAT as above. Appropriate amounts of pBluescript were added to equal the total amount of DNA transfected. Here, the ratio of luciferase activities (standardized to protein content) with expression of CREM-IbC-X and of the control without insert was plotted vs. the amount of effector DNA transfected. Values are means from three independent transfections in duplicate. For the immunological detection of internal translation products, HIT-T15 cells were grown overnight on 16 cm plates and transiently transfected with 40 µg of the HA-tagged CREM-IbC-X construct or with pBK-CMV without insert using the Superfect reagent (Qiagen, Hilden, Germany) according to manufacturer's specifications. Here, cells were harvested 24 h after transfection.
Bacterial expression and electrophoretic mobility shift assays
E. coli (JM109 strain) cells were transformed with constructs containing CREM-IbC-X or its partial sequences in pSE420 (10). Cells were induced with 1 mM isopropyl-ß-D-thiogalactopyranoside 3 h before they were pelleted and boiled in 1/10 vol TE (pH 7.5). Recombinant CREB was prepared as described (13) from E. coli transfected with a bacterial CREB expression plasmid, which was a kind gift from Dr. M. Vallejo. Protein content was determined according to Bradford (14), with bovine serum albumin as standard. Electrophoretic mobility shift assays were performed as described previously (15) using a 32P-labeled CRE-containing DNA oligonucleotide of 30 bp derived from the human chorion gonadotropin 
gene promoter.
Detection of CREM mRNA
The insert of the construct BH was labeled with [-32P]dCTP using the random primer labeling technique (Megaprime Kit, Amersham, Braunschweig, Germany) according to manufacturer's specifications. A Northern blot with 2 µg per lane poly A+-RNA from different human tissues was obtained (Clontech, Palo Alto, Calif.) and hybridized with the radiolabeled cDNA according to standard protocols. The blot was washed at a final stringency of 0.2x SSC with 0.1% sodium dodecyl sulfate (SDS) at 60°C before radioactivity was visualized using the PhosphorImager System (Molecular Dynamics, Sunnyvale, Calif.).
Immunological detection of CREM protein
Human ventricular homogenates were prepared according to Rockman et al. (16). CREM proteins were immunoprecipitated from 500 µg homogenate by using a rabbit polyclonal antibody raised against purified mouse CREM-
(1:200, Upstate Biotechnology Inc., N.Y.) as described (17). The precipitated proteins were separated by electrophoresis on 10% SDS-polyacrylamide gels, transferred onto nitrocellulose, and visualized with the same antibody (1:1000), a polyclonal anti-CREB antibody (1:1000, Upstate Biotechnology Inc.), or a monoclonal anti-ATF-1 antibody (1 µg/ml; Santa Cruz, Santa Cruz, Calif.) and [125I]-labeled protein A, using the PhosphorImager system. Moreover, signal specificity was tested by preincubation of the anti-CREM antibody with 1 µg of the peptide EAAKECRRRKKEYVK (amino acids 1–15 of CREM exon Ib, peptide synthesized and purified to >95% by Genosys Biotechnologies, Pampisford, U.K.) or with the same amount of a nonrelated peptide (amino acids 436–454 of the transcription factor SP-1, Santa Cruz). To identify internally translated proteins in transfected cells, cell lysates were immunoprecipitated as described above with a polyclonal antibody against the HA epitope (Santa Cruz) and visualized with the anti-CREM antibody (1:1000) and [125I]-labeled protein A.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fig. 1A). Exons B, H, and Ib were as described (18, 19), except for one patient with a deviation in base 238 (GAAGAA
GAAGGA); following a common nomenclature (20), we named the novel isoform CREMIbC-X.
Whereas exon B ends with a complete codon of the open reading frame, exon H begins with an incomplete codon that is kept in frame by an additional base of exon F in other CREM isoforms. Since exon F is spliced out in CREM-IbC-X, there is a frame shift at the boundary between exons B and H introducing a stop codon in exon H. This leads to predicted translation of a 60 amino acid protein (BH) initiated at base 7 and truncated at base 189 ( Fig. 1B) that lacks the DNA binding domain. However, there are potential translation start sites in exon H in frame with other CREM isoforms ( Fig. 1B). Internal initiation of translation in exon H would result in small proteins of 96, 78, or 52 amino acids (named HIbI, HIbII, and HIbIII with predicted molecular masses of 11, 9, and 6 kDa) consisting of the basic region and the leucine zipper (bZIP) of DNA binding domain Ib.
The full-length cDNA of CREM-IbC-X (=BHIb) was expressed in HIT-T15 cells cotransfected with the CRE-controlled luciferase reporter 4xSomCRE. In control cells transfected with the empty expression vector, stimulation of the cAMP signaling pathway by forskolin led to a 25-fold increase in luciferase activity ( Fig. 2,
CTR). The cAMP-stimulated luciferase activity was inhibited to about 55% by overexpression of CREM-IbC-X ( Fig. 2, BHIb). This repressor effect was dependent on the amount of effector transfected and was the same in both cAMP-induced and noninduced cells ( Fig. 3).
Overexpression of the construct HIb exclusively allowing expression of internally translation products inhibited transcription comparable to the full-length construct of CREM-IbC-X ( Fig. 2, HIb). In contrast, there was no effect of the construct BH with the sequence of the truncated protein BH initiated at base 7 ( Fig. 2, BH). Therefore, in intact cells a concentration-dependent inhibition of CRE-mediated transcription was achieved by overexpression of CREM-IbC-X, which was localized to the open reading frame of internal translation products.
|
|
To identify internally translated proteins explaining the inhibition of CRE-mediated transcription, cells were transfected with a construct of CREM-IbC-X fused to the sequence of the hemagglutinin epitope of the influenza virus (HA tag). Since the HA tag was introduced in the open reading frame of the internally translated proteins at the 3'-end of exon Ib, only internally translated proteins were tagged with the HA epitope. Two proteins were immunoprecipitated using a HA-specific antibody and visualized by an anti-CREM antibody in cells overexpressing CREM-IbC-X/HA ( Fig. 4).
No signal was obtained in control cells transfected with the empty vector. The precipitated proteins were identified as HIbI and HIbII, with predicted molecular masses (without HA epitope) of about 11 and 9 kDa, respectively. The third predicted translation product HIbIII of about 6 kDa was not detected.
|
The DNA binding properties of HIbI and HIbII were investigated in gel shift assays with lysates of bacteria expressing HIbI or HIbII. Both recombinant HIbI and HIbII showed specific binding to the CRE ( Fig. 5)
whereas there was no DNA binding of recombinant BH (data not shown). Recombinant CREB showed a CRE-specific shift representing the homodimer ( Fig. 6)
and a nonspecific signal, in line with published data (12). Increasing amounts of HIbI in the presence of constant amounts of CREB resulted in a dose-dependent loss of the CREB shift ( Fig. 6). The appearance of a new band with intermediate migration velocity was observed, likely representing the heterodimer of CREB and HIbI. Similar to HIbI, HIbII dose-dependently suppressed CREB's binding to the CRE and formed a heterodimer with CREB; however, the affinity to CREB appeared to be lower with HIbII as compared to HIbI ( Fig. 6).
|
|
Since both HIbI and HIbII were expressed in transfected cells, we investigated whether the repressor activity is blunted when translation of HIbI is suppressed. Therefore, cells were transfected with HIb, allowing expression of HIbII but not of HIbI. Overexpression of HIb inhibited CRE-mediated transcription comparable to BHIb and HIbI (luciferase activity, referred to nonstimulated control: control, stimulated with 10 µM forskolin: 8.93 ± 1.47; HIb, basal: 0.86 ± 0.03; HIb: stimulated: 4.49 ± 0.33; P<0.05), indicating that HIbII can act as a repressor of CRE-mediated gene transcription.
To investigate CREM expression in the human heart, Northern blot analysis of polyA+-RNA from different human tissues was performed with a CREM-specific cDNA. Hybridization with highly stringent washing conditions revealed CREM-specific signals in the heart that are abundant compared with other human tissues ( Fig. 7).
The same blot was hybridized to cDNAs specific for protein phosphatase isoforms, showing a similar distribution in all tissues (data not shown). The appearance of the CREM signal as a broad band at about 1.35 kb is in line with previous studies (21, 22) reflecting the existence of multiple transcripts. CREM expression was investigated on the protein level in human cardiac ventricular homogenate. An antibody against CREM- immunoprecipitated several proteins from ventricular homogenate, which were visualized in Western blots ( Fig. 8).
The same antibody detected a protein of about 9 kDa, which is the predicted molecular mass of HIbII. The 9 kDa band was CREM specific since it was not detected by antibodies against the structurally related transcription factors CREB and ATF-1, and was blocked by preadsorption with a peptide derived from exon Ib but not with a nonrelated peptide.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C-X is a novel example of the internally initiated translation from the CREM gene. The introduction of stop codons by alternative splicing has been shown for CREB, resulting in truncated proteins lacking the DNA binding domain that are developmentally regulated in the testis and involved in spermiogenesis (23). Recently, Gellersen et al. (24) described a similar feature in alternatively spliced CREM isoforms identified in endometrial stromal cells; similar to CREMC-X described therein, the combination of exons B and H in CREM-IbC-X introduces a stop codon in exon H. Another alternatively spliced CREM repressor (CREMC-G) was described in rat spermatids (20) that differs from CREM-IbC-X by the presence of the 36 bp exon X and the use of another DNA binding domain, Ia. However, CREMC-G contains an additional base G between exons B and X, conserving the open reading frame through the whole sequence leading to translation of a 17 kDa repressor protein. To exclude a reading error of the Taq polymerase, we amplified the region around the exon B/H border using different nested primers and RNA from other hearts, confirming the lack of an additional guanosine in CREM-IbC-X.
The evidence that CREM-IbC-X inhibits CRE-mediated transcription by internal translation of short repressors is given by the localization of the function to the open reading frame of the internal translation products and the immunological identification of two proteins internally translated in vivo in transfected cells. These proteins were assigned to HIbI and HIbII initiated at bases 135 and 189. In CREMC-X, AUG 189 is probably the start site for internal translation of SS-CREM (24), but no initiation at bases 135 and 267 was observed for this isoform. The sequence around AUG 189 (GUGAUGG) with guanosines in the positions -3 and +4 (base A of AUG is defined as position +1) is in homology to the consensus sequence for effective initiation of translation (25), and the analog initiation site in the CREB sequence with an almost identical context (GUUAUGG) likely is the start site for the generation of I-CREB(s) (26, 27). The 9 kDa cardiac CREM protein, the putative HIbII, would be translated from the analog AUG. Internal translation can also be initiated at AUG 135, whereas translation of HIbIII from AUG 267 was not observed. This can be explained by the context around AUG 135 (GACAUGC) with a guanosine in position -3, which is critical for translational initiation, and with the context around AUG 267 (CUAAUGA) not fulfilling any requirement of the Kozak consensus sequence (25).
Both HIbI and HIbII showed specific binding to the CRE. Moreover, our data suggest that both proteins heterodimerize and compete with CREB's binding to the CRE. This is consistent with the concept that repressors of CRE-mediated transcription act by formation of inactive homo- or heterodimers preventing activating factors from their action. As compared to HIbI, the affinity of HIbII to CREB appeared to be lower. However, overexpression of HIb (in which the open reading frame of HIbI but not of HIbII is blocked) showed a repressor activity comparable to the full-length sequence, indicating that HIbII can effectively confer transcriptional repression. Overexpression of CREM-IbC-X and of HIb inhibited CRE-mediated transcription by about 45%, which is comparable to CREMC-X and the internally translated inhibitor I-CREB(s) (24, 26, 27).
CREM mRNAs were expressed abundantly in the human heart compared to other tissues, suggesting a functional relevance of CREM for cardiac transcriptional regulation. A 9 kDa protein, the putative HIbII, was precipitated from human ventricular homogenate by anti-CREM- recognizing all known CREM isoforms, including CREM-, -ß, and -
. This 9 kDa signal was shown to be CREM specific because it was blocked by a peptide derived from the DNA binding domain Ib of CREM and was not detected by antisera against closely related factors. Two proteins of 31 and 60 kDa were also detected by anti-CREM- in the human heart. Whereas the 31 kDa protein might be assigned to CREM- and/or CREM-ß due to the apparent molecular mass, there is no known correlate for the 60 kDa protein.
The expression of CREM in the human heart is in line with the concept that excessive ß-adrenergic stimulation contributes to expressional changes in human heart failure. This hypothesis is further supported by recent data on transgenic mice overexpressing a dominant negative CREB mutant that cannot transactivate because it cannot be phosphorylated at serin 133 (28). These mice develop signs of dilated cardiomyopathy and die from heart failure, showing that heart failure is associated with inhibition of CRE-mediated gene transcription. Therefore, it can be speculated that the novel repressors described here are examples of endogenous proteins with a function similar to that of the dominant negative CREB mutant contributing to the expressional changes observed in human heart failure.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Abbreviations: CRE, cAMP response element; CREB, cAMP response element binding protein; CREM, cAMP response element modulator; bZIP, basic region leucine zipper; HA, hemagglutinin epitope of the influenza virus; RT-PCR, reverse transcription-polymerase chain reaction; TE, Tris-EDTA buffer; SSC, standard saline-citrate; SDS, sodium dodecyl sulfate.
Received for publication March 2, 1998. Accepted for publication April 14, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-proteins in patients with congestive heart failure. Eur. J. Clin. Pharmacol. 51, 127–132[Medline]
-subunit Gi-2 in rat heart. Circ. Res. 72, 696–700This article has been cited by other articles:
|
|
 |
|
  G. S. Huggins, J. J. Lepore, S. Greytak, R. Patten, R. McNamee, M. Aronovitz, P. J. Wang, and G. L. Reed The CREB leucine zipper regulates CREB phosphorylation, cardiomyopathy, and lethality in a transgenic model of heart failure Am J Physiol Heart Circ Physiol, September 1, 2007; 293(3): H1877 - H1882. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Matus, G. Lewin, F. Stumpel, I. B. Buchwalow, M. D. Schneider, G. Schutz, W. Schmitz, and F. U. Muller Cardiomyocyte-specific inactivation of transcription factor CREB in mice FASEB J, June 1, 2007; 21(8): 1884 - 1892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. U. Muller, G. Lewin, H. A. Baba, P. Boknik, L. Fabritz, U. Kirchhefer, P. Kirchhof, K. Loser, M. Matus, J. Neumann, et al. Heart-directed Expression of a Human Cardiac Isoform of cAMP-Response Element Modulator in Transgenic Mice J. Biol. Chem., February 25, 2005; 280(8): 6906 - 6914. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Bailey, A. J Tyson-Capper, K. Gilmore, S. C Robson, and G N. Europe-Finner Identification of human myometrial target genes of the cAMP pathway: the role of cAMP-response element binding (CREB) and modulator (CREM{alpha} and CREM{tau}2{alpha}) proteins J. Mol. Endocrinol., February 1, 2005; 34(1): 1 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. Gao, T. Tang, T. Guo, S. Q. Sun, J. R. Feramisco, and H. K. Hammond Adenylyl Cyclase Type VI Gene Transfer Reduces Phospholamban Expression in Cardiac Myocytes via Activating Transcription Factor 3 J. Biol. Chem., September 10, 2004; 279(37): 38797 - 38802. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. ISODA, N. PAOLOCCI, K. HAGHIGHI, C. WANG, Y. WANG, D. GEORGAKOPOULOS, G. SERVILLO, M. A. DELLA FAZIA, E. G. KRANIAS, A. A. DEPAOLI-ROACH, et al. Novel regulation of cardiac force-frequency relation by CREM (cAMP response element modulator) FASEB J, February 1, 2003; 17(2): 144 - 151. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Gellersen, R. Kempf, R. Sandhowe, G. F. Weinbauer, and R. Behr Novel leader exons of the cyclic adenosine 3',5'-monophosphate response element modulator (CREM) gene, transcribed from promoters P3 and P4, are highly testis-specific in primates Mol. Hum. Reprod., November 1, 2002; 8(11): 965 - 976. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. U Muller, P. Boknik, J. Knapp, B. Linck, H. Luss, J. Neumann, and W. Schmitz Activation and inactivation of cAMP-response element-mediated gene transcription in cardiac myocytes Cardiovasc Res, October 1, 2001; 52(1): 95 - 102. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or with the solvent DMSO (
). *P < 0.05 vs. 1 µg DNA (DMSO); +P < 0.05 vs. 1 µg DNA (forskolin).
