| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Department of Obstetrics and Gynaecology, University of Newcastle upon Tyne, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, United Kingdom
Address all correspondence and requests for reprints to: Dr. Jarrod Bailey, Department of Obstetrics and Gynaecology, University of Newcastle upon Tyne, Royal Victoria Infirmary, Fourth Floor Leazes Wing, Newcastle upon Tyne NE1 4LP, United Kingdom. E-mail: ; or . g.n.europe-finner{at}ncl.ac.uk
Abstract
There is now extensive evidence to indicate that components of the cAMP signaling pathway are up-regulated in the human myometrium during pregnancy so as to potentiate the maintenance of uterine quiescence until term. In many tissue and cell types, increased signaling of the cAMP pathway results in profound changes in gene expression that are catalyzed via stimulation of PKA and activation of cAMP-dependent transcription factors that bind cAMP response elements (CREs) within the promoter regions of affected genes. In the myometrium, these CRE containing genes include ß2-adrenoceptor, cyclo-oxygenase 2, oxytocin receptor, and connexin-43. In preliminary investigations, we reported the differential expression of members of the cAMP bZIP protein family in the myometrium during pregnancy and labor. In this present study, we have now identified and functionally characterized these proteins with respect to myometrial gene expression. We report the identification of a 39,000 mol wt CRE response element modulator protein (CREM)
2
protein having both transactivation and transrepressor properties whose expression is sequentially decreased in the myometrium during gestation and parturition. In contrast, expression of a myometrial 28,000 mol wt CREM protein having only transrepressor actions progressively increased in the myometrium during pregnancy and labor. Similarly, we have isolated two ATF2 proteins of 60,000 and 28,000 mol wts, which represent full-length ATF2 and a novel small isoform of ATF2 that we have termed ATF2-small (ATF2-sm). These proteins are potent transactivators of gene expression and appear to be spatially expressed within the myometrium of the upper and lower uterine regions. The identification and functional characterization of these basic region/leucine zipper proteins in the myometrium may provide further insight into the molecular mechanisms regulating uterine activity during fetal maturation and parturition.
ELUCIDATION OF THE mechanisms controlling myometrial contractility during human pregnancy is a major priority in the search for more effective pharmacological agents to control the incidence of premature labor, which remains a significant cause of perinatal mortality and infant morbidity worldwide. During fetal maturation, many physiological adaptations take place within the uterus, including enlargement to contain the growing fetus and the maintenance of myometrial relaxation. However, at term the myometrium must rapidly switch from quiescence to a state capable of producing a series of forceful contractions so as to expel the infant. It is becoming increasingly apparent that these processes involve the differential expression of specific genes that are responsible for controlling the activity of the uterus during pregnancy and parturition, the abnormal regulation of which could result in premature delivery or the inability to initiate or progress labor. There is now extensive evidence to indicate that components of the cAMP signaling pathway are differentially regulated in the human myometrium during pregnancy so as to potentiate the maintenance of uterine quiescence until term. These include human CG/LH (1) and calcitonin gene-related peptide (2) receptors and the adenylyl cyclase stimulatory G protein Gs (3, 4, 5), whose levels of expression are substantially increased within the myometrium during gestation resulting in the increased production of cAMP, which is further amplified by the progesterone-induced down-regulation of myometrial cAMP phosphodiesterase activity (6). Changes in the expression of specific genes brought about by increased cAMP formation and activation of PKA in most tissues and cell types are known to be mediated by phosphorylation of the transcription factors cAMP response element binding protein (CREB), modulator protein (CREM) and the activating transcription factor family (ATF1–4), which interact with the palindromic cAMP response element (CRE) 5'-TGACGTCA-3' in the promoter regions of the genes affected (7, 8, 9). In this context, several specific genes involved in regulating myometrial activity during pregnancy and labor have from database searches been found to contain palindromic and hemipalindromic CREs within their promoter regions these include ß2-adrenoceptor, cyclo-oxygenase 2, oxytocin receptor, and connexin-43.
CREB, CREM, and ATF1–4 are members of a large class of transcription factors known as bZIP (basic region/leucine zipper) proteins (10, 11), which bind as homo- or selective heterodimers to DNA (12, 13, 14, 15) and promote or inhibit gene transcription (16, 17, 18, 19). The genes that code for these proteins are multiexonic in structure and alternative splicing of exons generates a complex array of isoproteins that are either transactivators or transrepressors (16, 20, 21, 22). However, while the exonic structures of most genes in this family have been elucidated, and their alternatively spliced products have been well characterized, there is a paucity of information about the exonic structure of the ATF2 gene.
In preliminary studies employing Western immunodetection, immunoprecipitation, and 2D-gel isoelectric focusing, we have described dramatic changes in the expression of protein isoforms of CREB, CREM, and ATF2 in the human myometrium during pregnancy and parturition (23). Essentially a 43,000 mol wt CREB protein was positively identified and found to be primarily expressed in nonpregnant myometrium compared with tissues sampled from pregnant nonlaboring and spontaneous laboring women. Expression of an apparent 39,000 mol wt CREM protein isoform was also observed to progressively decrease in myometrial tissues during pregnancy and labor compared with nonpregnant controls. In contrast, a 28,000 mol wt CREM-like protein was found to be elevated severalfold in pregnant nonlaboring tissues compared with the nonpregnant state, where it was only weakly detected, and then subsequently increased even further during parturition. Similarly two potential isoforms of ATF2 with apparent molecular weights of 60,000 and 28,000 were also observed to be present in the myometrium. The smaller 28,000 mol wt ATF2 isoform was found to be expressed at much greater levels in tissue sampled from the upper (corpus) compared with the lower uterine segment during pregnancy/labor, whereas expression of the larger 60,000 mol wt protein appeared to be sequentially decreased during pregnancy and labor compared with nonpregnant samples.
In this present study, we have identified and functionally characterized these CREM and ATF2 protein isoforms. We report the identification of the 39,000 and 28,000 mol wt CREM proteins as CREM2 and CREM, respectively. Using transfection of primary human myometrial cells in culture and COS-7 cells with CRE reporter gene constructs and plasmids coding for CREM2 and CREM, we demonstrate that CREM2 can act as either a transactivator or transrepressor depending on promoter context, whereas CREM acts as a potent transrepressor in the myometrium. We have also characterized the putative 60,000 and 28,000 mol wt ATF2 proteins, respectively, as full-length ATF2 and a novel alternatively spliced isoform of ATF2 having no intrinsic histone acetyltransferase (HAT) activity but with potent transactivation properties in common with the full-length isoform from which it is derived. The elucidation of these transcriptional regulators may implicate them in an important role in the modulation of uterine activity throughout pregnancy and labor.
Materials and Methods
Collection of myometrial tissues
Samples of myometrium from nonpregnant premenopausal women (ages 32–46 yr) were obtained at hysterectomies performed for benign gynaecological disorders such as menorrhagia or dysmenorrhoea. The uteri were excised longitudinally, and samples of myometrium were taken from the middle of the uterine wall approximately 5 mm away from the endometrial or serosal surfaces. Samples obtained in both the follicular and luteal phases of the cycle were used. Pregnant nonlaboring and spontaneous laboring lower uterine segment myometrial samples were respectively collected from women undergoing elective caesarean section at 38–39 wk gestation and at emergency caesarean section for fetal distress or failure to progress. The myometrial samples were snap frozen in liquid nitrogen and stored at -70 C until required. Written consent was obtained from all women, and ethics approval for the study was granted by the Newcastle and North Tyneside Health Authority Ethics Committee.
Western immunodetection and RT-PCR
Western blotting of nonpregnant, pregnant nonlaboring, and spontaneous laboring tissues with the anti-CREB/anti-CREM antibody and the ATF-2 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was as described by Bailey et al. (23) and is illustrated in this present study employing pools of 4 individual patient samples in each tissue group. mRNA was isolated from pooled frozen myometrial term pregnant nonlaboring (n = 6 patient samples) tissues (1 g) respectively using the Poly(A)Pure mRNA Isolation Kit [Ambion (Europe) Ltd., Huntingdon, Cambridgeshire, UK]. An electrical homogenizer was used for tissue disruption, and no changes were made to the published protocol. Fifty nanograms of the mRNA were used as a template for RT primed with 2 pmol of gene-specific reverse primer, using Superscript RNase H- (Invitrogen, Paisley, UK) as the polymerase. The primers used were:
CREM(R) 5'-CGTCGACATTCTTTGGCAGC-3'
ATF2(R) 5'-GAAACTGGTCTTTCCTTGATT-3'.
The Ambion, Inc. protocol was followed and 5 µl of the reaction mixture subsequently used as a template for the PCR reaction. In each case, Pfu polymerase (Promega Corp., Madison, WI) was used to ensure fidelity. Primers used were as follows:
CREM(F) 5'-ATGACCATGGAAACAGTTGAATCCCA-3'
ATF2(F) 5'-CACCATGAAATTCAAGTTACATGTG-3'.
Reverse primers were the same as those used in the RT reactions detailed above; the forward primer for ATF2 contained a CACC-5' addition for the purpose of directional cloning (see Constructs). Note that the primers used in these RT-PCR reactions are specific for the 5' and 3' terminal exons of the CREM and ATF2 genes (predicted exons in the case of ATF2), and incorporate the start and stop codons. This ensured the detection of any alternatively spliced isoforms derived from these genes. RT-PCR products were visualized on EtBr-stained 1.2% agarose gels, and gel-purified using the QIAEX II kit (QIAGEN, Crawley, UK), eluting in 20 µl of water. Seven microliters of this eluate were A-tailed for TA-cloning by incubation with 5U Taq polymerase in 1x Taq buffer in the presence of deoxy-ATP at 72 C for 15 min.
Constructs
Plasmid constructs expressing CREM2, CREM, ATF2-fl, and ATF2-sm isoforms were assembled by cloning the A-tailed RT-PCR products into the eukaryotic expression vectors pcDNA3.1/V5-His TOPO or pcDNA3.1D/V5-His TOPO (Invitrogen). Clones were produced by TA insertion according to the accompanying protocols, and were verified by sequencing. The expression vectors for the wild-type and mutant catalytic subunits of PKA (pRSV-Cß/pRSV-Cßmutant) were generously provided by Dr. R. Maurer (Oregon Health Sciences University, Portland OR). The CRE-containing luciferase reporter construct pCRE/-36rPRL/luc3 and the cAMP responsive dPRL-332/luc3 construct were kindly supplied by Dr. B. Gellersen (University of Hamburg, Hamburg, Germany), and respectively consisted of a classical CRE element linked to a minimal promoter and the firefly luciferase gene and a cAMP-responsive (in a delayed fashion) portion of the PRL promoter linked to a minimal promoter plus the firefly luciferase gene.
In vitro transcription/translation
The expression of CREM2, CREM, ATF2-fl, and ATF2-sm was examined in vitro by adding 0.5 µg of the expression plasmid (pCREM2, pCREM, pATF2-fl, and pATF2-sm, respectively) to TnT-T7 quick-coupled transcription/translation mix (Promega Corp.), in the presence of 35S-methionine in a final volume of 25 µl. Reactions were incubated at 30 C for 75 min, then 2 µl analyzed by SDS-PAGE and subsequent autoradiography.
EMSAs
EMSAs were performed using a [
32P]-ATP labeled consensus CRE oligonucleotide (5'-AGAGATTGCCTGACGTCAGAGAGCTAG-3'), and a negative control OCT-1 oligonucleotide (5'-TGTCGAATGCAAATCACTAGAA-3') (Promega Corp.) in nonspecific competition reactions. Labeled oligonucleotides were annealed and purified using ProbeQuant G-50 microcolumns (Amersham Pharmacia Biotech, Bucks, UK), and then incubated with 2 µl of in vitro translation product in the presence of 1 µg of poly (deoxyinosine-deoxycytidine) (Amersham Pharmacia Biotech), 0.1 µg poly-L-lysine, 1x EMSA buffer [25 mM HEPES (pH 7.5), 10% glycerol, 5 mM MgCl2, 100 mM KCl, 1 mM DTT, 0.2 mM EDTA, 0.5 mg/ml BSA] and 200-fold competitor oligonucleotide where appropriate, in a total volume of 15 µl. Reactions were incubated at room temperature minus labeled oligo for 30 min, and for a further 30 min following addition of approximately 40,000 cpm. The reactions were then subjected to 4% SDS-PAGE for 3 h at 150 V, then the gels dried and exposed to x-ray film.
Transient transfections and CRE-reporter gene studies
Primary cultures of human myometrial cells were prepared from nonpregnant or pregnant nonlaboring myometrium as described by Phaneuf et al. (24). Transfections were performed upon myometrial cells before passage 3 at 70% confluence or COS-7 cells at 40–60% confluence in 35-mm plates, using 4 µl of LT-1 lipid-based transfection reagent (Cambridge Bioscience, Cambridge, UK) per 1 µg of plasmid. Typically, the amounts of plasmids used for COS-7 transfections were as follows: 0.8 µg CREM/ATF2 expression plasmid, 0.6 µg pRSV-Cß/Cßmutant, 1.4 µg pCRE/-36rPRL/luc3 luciferase reporter plasmid, 1.4 µg dPRL-332/luc3 luciferase reporter plasmid, and 28 ng pRL-TK Renilla luciferase control reporter vector (Promega Corp.); these amounts were doubled for transfection of myometrial cells. The plasmids were combined with 11.3 µl for COS-7 cells or 22.6 µl for myometrial cells of LT-1 reagent and 200 µl of serum-free medium and incubated at RT for 5 min.
This complex was added to the cells and incubated at 37 C for 4 h, after which time it was removed and fresh medium (+ serum) added. Incubation at 37 C was continued for 48 h, then the cells harvested and assayed for reporter activity using the Dual Luciferase Assay System (Promega Corp.). Each transfection and reporter assay was done in duplicate.
Results
We have previously described the expression of protein isoforms of CREB, CREM, and ATF2 in the human myometrium during pregnancy and labor (23), which is summarized in Fig. 1
. Briefly, a 43,000 mol wt phosphorylated form of CREB was positively identified and found to be primarily expressed in nonpregnant (NP) tissues, a putative 39,000 mol wt CREM protein was found to be progressively decreased in term pregnant and laboring tissues, respectively, in comparison to the nonpregnant controls, and a 28,000 mol wt CREM-like protein, absent in NP myometrium, was detected in pregnant nonlaboring (P) myometrium and observed to be substantially increased in laboring (SL) tissues. Two forms of ATF2 were also detected in P myometrial tissues as well as NP and SL tissues (data not shown); constituting 60,000 mol wt and 28,000 mol wt forms which are represented in Fig. 1B. From our detailed study on levels of expression of these proteins in the myometrium, described in (23), the smaller ATF2 protein was found to be present at much greater levels in the upper compared with the lower uterine segment during pregnancy and labor, whereas the larger isoform progressively decreased in both uterine regions compared with nonpregnant controls (23). All of these proteins were shown to possess CRE-binding abilities in the form of EMSAs using NP, P, and SL myometrial tissue lysates (23). This study indicated that two isoforms of CREM and ATF2 appeared to be the predominant bZIP proteins expressed in the myometrium throughout gestation and at the onset of parturition.
|
Identification of CREM and ATF2 mRNA species in the myometrium
Employing RT-PCR with mRNA extracted from pregnant nonlaboring (P) myometrial tissues, two CREM PCR products of 1133 bp and 948 bp were observed as detailed in Fig. 2A. These products were subsequently cloned and sequenced and found to represent mRNAs encoding the alternative spliced variants CREM2 and CREM (with calculated mol wts of 31,900 and 24,300, respectively) as described in (16, 25) and depicted in Fig. 3. These data therefore implied that the 39,000 and 28,000 mol wt proteins detected in vivo were CREM 2 and CREM, respectively, which was subsequently confirmed as detailed below. RT-PCR using the ATF2 primers and extracted myometrial tissue mRNA also resulted in only two products of approximately 1500 bp and 500 bp (Fig. 2B), which were subsequently cloned and sequenced. The base sequences of these clones and the corresponding amino acid sequences of the translated products are depicted in Fig. 4. Note that the only reported exonic structure of the human ATF2 gene is in the form of a predicted model (Ensembl gene ID ENSG00000115966) based upon protein/genomic sequence alignments that gives rise to an expected 12-exon arrangement. The larger product was thus found to be full-length ATF2, termed ATF2 full-length (ATF2-fl) for clarity and having a calculated mol wt of 54035, a potent transactivator of transcription possessing intrinsic HAT activity (26) and a substrate of p38 MAPK. The smaller ATF2 product detected (GenBank accession no. AY029364) consisted of the first two and last two putative exons (1, 2, 11, and 12) and translates to a polypeptide of 144 amino acids with a calculated mol wt of 15,408, which is approximately half of the 28,000 mol wt ATF2 like protein detected in vivo (this discrepancy is explained below). This product appears to be devoid of the proline-rich domain (involved in protein-protein interactions), one of the two glutamine-rich domains (involved in trans-activation of transcription), the domain known to be responsible for HAT activity, and part of the bZIP domain that binds to CREs and hereafter is termed ATF-small (ATF2-sm).
|
|
|
To further identify the myometrial CREM and ATF2 isoforms detected in the myometrium during pregnancy and labor, in vitro transcription/translation experiments were performed employing the CREM and ATF2 expression constructs derived from RT-PCR of myometrial tissue mRNA.
Figure 5, lanes 1 and 2 show in vitro transcriptions/translations of the CREM2 (pCREM2) and CREM (pCREM) constructs, respectively. The principal product from pCREM2 was found to be of apparent mol wt 32,500 with shorter polypeptides of 20,000 and 16,000 also being formed. The primary product of pCREM was a 27,000 mol wt protein, with a shorter product present at 16,000. In both cases, expression of the smaller mol wt 20,000 and 16,000 protein products arises from alternative translation initiation sites as has been reported previously (25). Although the apparent mol wts of the primary in vitro proteins differ from those found for these proteins detected in vivo in myometrial tissues by Western blotting, these discrepancies (39,000 in vivo to 32,500 in vitro = 6,500 or 20% for CREM2 and 28,000 in vivo to 27,000 in vitro = 1,000 or 4% for CREM) may arise as a result of different phosphorylation and/or other posttranslational modification events occurring in the myometrium in vivo. Evidence for this was further provided by transfection of myometrial cells with the pCREM2 and pCREM plasmids, which only resulted in proteins of mol wt 39,000 and 28,000 as detected in Western blotting by the CREM-1 antibody from Santa Cruz Biotechnology, Inc. (data not shown).
|
, lanes 3 and 4, shows an autoradiograph of the 35S-labeled ATF2 proteins produced in vitro. The construct containing the full-length ATF2 isoform (pATF2-fl) gave rise to a major protein of apparent mol wt 63,000 (lane 3), in agreement with previous results (27, 28, 29), and a second band of slightly lower mol wt, which is produced due to an alternative translation initiation site that is known to be present. The ATF2-sm expression vector (pATF2-sm) produced a protein of apparent mol wt 19,000 (lane 4), with another alternative translation initiation product below in common with the full-length form, but also produced a protein of apparent mol wt 28,000, which is in accordance with the smaller ATF2 protein detected in myometrial tissues by Western blotting. Transfection of myometrial cells with the ATF2-sm expression vector and subsequent Western blotting of cell extracts with an ATF2 monoclonal antibody (Santa Cruz Biotechnology, Inc.) only resulted in the detection of two proteins of mol wts 28,000 and 60,000 (data not shown). The absence of a 19,000 mol wt protein indicated that the 28,000 mol wt protein appeared to be the predominant moiety expressed in vivo. The possible nature/composition of the 28,000 mol wt protein detected in SDS-PAGE gels is discussed below.
Myometrial CREM2, CREM, ATF2-fl, and ATF2-sm proteins functionally bind CREs
To verify that the proteins derived from the CREM2, CREM, ATF2-fl, and ATF2-sm constructs were capable of binding CREs, in vitro expressed proteins were used in EMSAs employing a labeled consensus CRE oligonucleotide (5'-AGAGATTGCCTGACGTCAGAGAGCTAG-3') and a negative control OCT-1 oligonucleotide (5'-TGTCGAATGCAAATCACTAGAA-3'). Results from these experiments are detailed in Fig. 6, A and B, and show that the two isoforms of CREM and ATF2 caused shifts in mobility, thus indicating the ability of these proteins to bind CREs. In brief, the CREM2 protein produced three shifted bands having different mobilities as previously reported (25 ; Fig. 6A, lane 3), which arise due to alternative translation initiation sites. Similarly the CREM protein also produced three shifted bands (Fig. 6A, lane 6), two of which had lower mobilities than that observed for the CREM2 protein. Cotranslated CREM2 and CREM proteins when used in EMSAs also produced three shifted bands, two of which correlated with the two higher bands observed for CREM2 alone and a band (labeled I, lane 9) that was intermediate between the upper two bands observed for CREM alone, which presumably arises due to the formation of CREM2/CREM heterodimers.
|
, lane 3) the lowest of which also appeared to be present in the control reaction using reticulocyte lysate alone but at much lower intensity. Note that the intensity of all three bands was substantially reduced using a specific competitor (Fig. 6B, lane 4) but not by a nonspecific competitor (Fig. 6B, lane 5). ATF2-sm gave rise to two shifted bands (Fig. 6B, lane 6), which appeared to comigrate with the lower two bands observed for ATF2-fl. The intensity of both bands was also significantly reduced using the specific competitor (Fig. 6B, lane 7) but not by the nonspecific competitor (Fig. 6B, lane 8). Because the lowest band observed for ATF2-fl and ATF2-sm can be competed away in both instances, this may indicate that the DNA-protein complex found in the control reaction arises due to endogenous amounts of ATF2 proteins present in the reticulocyte lysate. ATF2-fl and ATF2-sm proteins used together produced three complexes (Fig. 6B, lane 9) identical to ATF2-fl alone, which could be competed away with the specific competitor (Fig. 6B, lane 10) but not by the nonspecific competitor (Fig. 6B, lane 11). It also appeared that the presence of ATF2-sm in this reaction both reduced and increased, respectively, the intensity of the upper complex and lower two complexes produced by ATF2-fl alone. These data indicate that the ATF2-sm protein although devoid of a large proportion of amino acids present in the ATF2-fl protein, which include residues comprising various functional domains and even part of the DNA-binding bZIP domain, still has the ability to bind CRE-containing DNA. CREM and ATF2 proteins display potent transactivation and/or transrepression activities in vivo
To determine the functionality of the myometrial CREM and ATF2 proteins as transactivators or transrepressors, primary human myometrial cell cultures and COS-7 cells were cotransfected with plasmids expressing CREM2, CREM, ATF2-fl, and ATF2-sm either singly or in combinations, in the presence of a plasmid expressing the PKA catalytic Cß subunit (to potentiate the cAMP system) and the reporter vector pCRE/-36rPRL/luc3 containing the firefly luciferase reporter gene under the control of a CRE-containing promoter; this was replaced in certain cases with the cAMP responsive dPRL-332/luc3 construct (see below). All experiments also included the transfection of an additional reporter plasmid in the form of a Renilla luciferase vector (Promega Corp., Dual luciferase system) with which to standardize readings to a baseline level, to account for differences in transfection efficiencies. Forty-eight hours post transfection cells were lysed, and the luciferase activities of each lysate measured to determine the relative level of expression of the CRE-reporter gene.
Figure 7, A–F, depicts graphically the results of these transfection experiments, with all reporter activities normalized to a background pRSV-Cßmutant activity of 1. Note that in each case the pRSV-Cßmutant control was obtained in the presence of the CREM/ATF2 plasmid under investigation. Similarly, the transactivation or transrepressor properties of these plasmids was observed in the presence of the wild-type pRSV-Cß vector. Statistical analysis of these results was by way of one-way ANOVA followed by Bonferroni’s multiple comparison post test. Although there were subtle differences in luciferase activities between duplicate experiments involving both myometrial and COS-7 cells, the basic trends in affecting reporter activity remained the same. Essentially, CREM2 and CREM individually repressed reporter activity, as did a combination of both proteins (Fig. 7, A and B). This effect was most pronounced in myometrial cells, where luciferase activity dropped 4- to 6-fold to a level at or below background, in contrast to a reduction of 50% to a level twice that of background in COS-7 cells.
|
2 can exhibit both activator and repressor functions depending upon promoter context and cell type as detailed in (25), the effect of this protein was examined using the alternative dPRL-332/luc3 reporter vector, which contains a portion of the PRL promoter that is cAMP responsive in a delayed fashion. Figure 7C shows the effect of CREM2 protein on this reporter construct in COS-7 cells and indicates that in this instance reporter activity is substantially increased. Similar results were also obtained using myometrial cells (data not shown).
The effects of ATF2-fl and ATF2-sm on pCRE/-36rPRL/luc3 reporter activity are shown in Fig. 7, D and E, and indicate that similar responses are seen in both myometrial and COS-7 cells. As expected, the ATF2-fl protein was found to possess potent transactivation activity and remarkably so was the ATF2-sm protein; in each case, reporter activity was increased by approximately 80–90%. A combination of both ATF2 isoforms also activated CRE-mediated transcription but was not found to be additive.
Because the novel ATF2-sm, CREM2 and CREM proteins appeared to be the predominant factors expressed in the myometrium during pregnancy/labor, the effect of combinations of these proteins on CRE reporter activity was determined. COS-7 cells were cotransfected with CREM2/ATF2-sm and CREM/ATF2-sm employing the pCRE/-36rPRL/luc3 CRE-reporter vector. Figure 7F indicates that when the ATF2-sm protein is coexpressed with the CREM2 factor, CRE-mediated transcription of the luciferase reporter remains repressed to a degree similar to that obtained by CREM2 alone; the presence of ATF2-sm thus appears to have no effect on the repressive action of CREM2. Conversely, when ATF2-sm is cotransfected with CREM, the effect obtained is intermediate between the results obtained with these proteins when transfected singly; the repressive effect of CREM appears to be cancelled out by the activating effect of ATF2-sm. Note that myometrial cells gave similar effects (data not shown).
Discussion
In this study employing RT-PCR with specific CREM and ATF2 primers and mRNA extracted from myometrial tissues sampled from pregnant nonlaboring women, we have isolated two isoforms of CREM and ATF2, respectively. Cloning and sequencing of these mRNAs indicated that the two CREM species were encoded by CREM2 (having a calculated mass of 31,900) and CREM (having a calculated mass of 24,300), whereas the two ATF2 moieties were encoded by full-length ATF2 (termed ATF2-fl and having a calculated mass of 54,035) and a novel small variant of ATF2 (termed ATF2-sm and having a calculated mass of 15,408) which is generated via a 1083 bp internal deletion within ATF2-fl. In vitro transcription/translation experiments were subsequently performed to determine whether CREM2/CREM and ATF2-fl/ATF2-sm mRNAs coded for the 39,000/28,000 mol wt CREM and the 60,000/28,000 mol wt ATF2 proteins, which were observed to be differentially expressed in the myometrium during pregnancy and labor (23). These in vitro assays resulted in the CREM2 and CREM constructs expressing proteins of apparent mol wts of 32,500 and 27,000, respectively. The apparent mol wts of these in vitro expressed proteins are in good agreement with the apparent mol wt CREM 39,000 and 28,000 proteins observed in vivo. Note also that CREM2 has also been found in vivo to have apparent higher mol wts in several cultured cell lines, for example 36,000 in adrenocortical cancer cells (30). Moreover, transfection of myometrial cells with the CREM2 and CREM constructs only resulted in the expression of proteins with apparent mol wts of 39,000 and 28,000. Consequently, the discrepancies that are observed between in vitro and in vivo expressed proteins may be the result of posttranslational modification events including phosphorylation that occurs in the myometrium and not in the in vitro assay.
In vitro translation of ATF2-fl gave rise to a protein of apparent mol wt 63,000, consistent with the apparent mol wt 60,000 ATF2 protein detected by Western blotting of myometrial tissues (23) and by other investigators (27, 28, 29). Expression of the ATF2-sm construct in vitro gave rise to two protein products of apparent mol wts 19,000 (which is slightly larger than the calculated mol wt of 15,400) and 28,000. Furthermore, transfection of this plasmid in myometrial cells did not result in the detection of an ATF2 species of apparent mol wt 19,000 but only in proteins with mol wts of 60,000 and 28,000, as would be expected from endogenous production/over expression of the 28,000 mol wt species in this in vivo system. The lack of expression of a 19,000 mol wt protein in vivo coupled with the detection of a 28,000 mol wt protein in vitro, which is nearly double the calculated mol wt of 15,400 for this protein, may suggest that the 28,000 mol wt ATF2-sm protein detected in vivo represents a dimer of two 15,400 mol wt monomers. Although rare there are other instances whereby protein dimers remain undenatured under SDS-PAGE denaturing conditions. These include dimers of ß-amyloid, which occurs via dityrosine bridge formation induced by peroxidase reaction (31), dimers of tubulin and tektin (32), and dimers of pilin (33). However, the possibility exists that the detection of a 28,000 mol wt protein in vivo/in vitro under SDS-PAGE conditions may also arise due to the aberrant mobility of the 19,000 mol wt protein as a result of the presence of reduced sulfhydryl residues on the three cysteines within ATF2-sm (note that ATF-fl has five cysteines). This phenomenon has been observed for recombinant interferon-inducible protein-10, which has four cysteines that are reduced and results in a 18,000 mol wt protein in SDS-PAGE gels instead of a 10,000 mol wt protein (34). Iakoucheva et al. (35) have also suggested that not only are reduced sulfhydryl residues involved in aberrant mobility but that glutamine-rich regions within proteins may also be responsible for shifts in mobility in SDS-PAGE gels. Analysis of the ATF2-sm sequence indicates that this protein has six glutamine residues, which appear not to be concentrated in any one region. Klenova et al. (36) have also implicated a high negative charge at the N or the C terminus in causing anomalous electrophoretic migration of the CTCF transcription factor. This has also been observed for the papillomavirus 16 E7 protein (37) and the Escherichia coli ams protein (38). In this context analysis of the ATF-sm protein sequence indicates that the protein is composed of 10.5% acidic amino acids, which are dispersed within the internal region and not at the N or C terminals. In light of the above observations, further investigation is therefore required to resolve the nature and/or composition of the 28,000 mol wt ATF2-sm protein. Moreover, if this protein does prove to be a dimer of two ATF2-sm monomers, then the possibility arises that the nondetection of a 19,000 mol wt moiety in vivo may be due to exclusivity of the monoclonal ATF2 antibody employed in Western blotting for epitopes within the homodimer.
Employing the expression constructs coding for CREM2 and CREM in transfection experiments with a reporter gene containing a minimal promoter with a classical CRE element indicates that both proteins act as repressors of gene transcription as has been previously reported (25). Similarly, our data from EMSAs also indicates that CREM2 and CREM can form heterodimers as first described by Laoide et al. (39). However, there appears to be no additive effect on CRE reporter activity when both proteins are expressed together in transfected cells. From the exonic structure of the CREM gene detailed in Fig. 3, CREM2 lacks the glutamine-rich Q1 domain (amino acids 39–87) but still retains the glutamine-rich Q2 domain (amino acids 168–229) involved in trans-activation and therefore can also function as an activator of transcription (39). This effect is further evidenced here, whereby cotransfection of the CREM2 construct and a reporter vector that contains a portion of the PRL promoter that is cAMP responsive results in a significant increase in reporter activity. A similar event was also observed by Gellersen et al. (25) using endometrial stromal cells. In contrast to the trans-activator/-repressor properties of CREM2 that are dependent on promoter context and cell type, CREM is devoid of both glutamine-rich Q1 and Q2 trans-activation domains and thus in nearly all cases can only act as a transrepressor. The exception to this involves the phosphoenolpyruvate carboxykinase gene promoter (40), in this instance CREM dimers activate transcription by association between their phosphorylated kinase-inducible domains and CREB-binding protein and CCAAT/enhancer-binding protein.
Identification of the 39,000 and 28,000 mol wt CREM-like proteins as CREM2 and CREM, respectively, in correlation with their expression in the myometrium during fetal maturation (23) indicates that as CREM2 expression decreases sequentially in the myometrium during pregnancy and labor there is a reciprocal increase in CREM expression. This may suggest that there is a switch from the production of CREM2 (a potent trans-activator and/or trans-repressor) to the predominant formation of CREM (a potent trans-repressor) that occurs via changes in alternative splicing of the CREM precursor mRNA transcript by the splicing machinery (see Fig. 3). Moreover, CREM can also dimerize with CREB; these dimers have strong binding characteristics to consensus CREs (39) and in most cases act as potent transrepressors. Therefore, because the CREB gene has multiple CREs within its promoter and is autoregulated by itself (16), the possibility exists that as CREM increases in the myometrium during pregnancy, CREM/CREB heterodimers form resulting in the observed decrease in CREB 43-kDa expression in the myometrium at term/during parturition (see Fig. 1).
In this study, we have also isolated a previously unidentified isoform of ATF2 that in common with full-length ATF2 has potent trans-activation properties. Results from EMSAs and reporter studies suggests that both ATF-fl and ATF2-sm may form interaction partners to affect gene transcription. However, although there appears to be some interaction with CREM in that the repressive effect of CREM alone appears to be cancelled out by the activating effect of ATF2-sm alone in reporter studies; this is not the case for CREM2 because repression of CRE reporter activity is only observed as if for CREM2 alone. Genomic Ensembl database searches indicate that full-length ATF2 (Ensembl gene ID ENSG00000115966) is composed of 12 exons with different spliced isoforms, generated by alternative splicing of exons, found in various mammalian species. In this context, the small ATF2-sm isoform detected in the myometrium occurs as the result of skipping of eight internal exons, which encode regions involved in protein:protein interactions, HAT activity and trans-activation of transcription. Although ATF2-sm is only composed of the first two and last two putative exons of the ATF2 gene and only contains a small part of the bZIP DNA-binding domain, it is still capable of binding CRE containing DNA as observed by EMSAs (see Fig. 6B). It also contains two threonine residues at position 69 and 71 of the protein that are primary phosphorylation substrates within ATF2-fl for stress-activated protein kinase, Jun-N-terminal kinase, and p38 MAPK. Phosphorylation of these residues relieves an intramolecular inhibitory interaction and increases the potential for transcriptional activation (41, 42, 43). In addition, ATF2 molecules phosphorylated at these sites become protected from ubiquitination and degradation, giving them a longer active half-life (44). Therefore, ATF2-sm proteins devoid of proline-rich domains involved in homo- and heterodimerization, which tends to increase ubiquitination and facilitate degradation (45), may remain in a potentially active state for a longer period of time than ATF2-fl.
It is well known that the upper and lower myometrial regions of the uterus govern contractility and dilatation, respectively. Consequently, because expression of ATF2-sm species in the myometrium is observed to be greater in the upper compared with the lower uterine region during gestation/labor (23), the possibility exists that it may promote the expression of specific genes that propagate contractions from the fundus to the cervix before and at the onset of parturition. These genes may include the oxytocin receptor (OTR), cyclo-oxygenase 2 (COX-2), and connexin-43 (Cx-43), all of which have CRE motifs in their promoter regions and have been shown to be spatially expressed within the myometrium during pregnancy/labor in common with the ATF2-sm splice variant reported here (46, 47, 48). Furthermore, as ATF2 is also a substrate for kinases unrelated to PKA, as detailed above, there would appear to be a more complex regulatory mechanism controlling activation of ATF2 species in vivo in the myometrium.
In conclusion, we have characterized the predominant cAMP-related bZIP transcription factors that are expressed in the human myometrium during gestation and parturition. These proteins may exert profound effects on the expression of myometrial genes involved in regulating uterine activity during fetal maturation not only by binding to CREs but also to AP1 sites within the promoter regions of affected genes, the latter occurring as a result of dimerization with the Jun and Fos bZIP families. Similar transcriptional regulation may also be elicited by the cAMP bZIP family via CREB-binding protein-independent routes involving the recently reported FHL (four-and-a-half LIM domain) tissue-specific coactivator proteins such as ACT (activator of CREM in testis; 49). Consequently, to further comprehend the role of these CREM and ATF2 isoforms in regulating activity of the uterus during pregnancy and labor will necessitate the determination of their potential target genes in the myometrium.
Acknowledgments
We thank Birgit Gellersen (University of Hamburg, Hamburg, Germany) for her kind gifts of the pCRE/-36rPRL/luc3 and dPRL-332/luc3 reporter plasmids, and Richard Maurer (Oregon Health Sciences University, Portland, OR) for the pRSV-Cß PKA expression plasmids.
Footnotes
This work was funded by a grant made available from the Wellcome Trust (Grant 053563).
Abbreviations: ATF1–4, Activating transcription factor family; bZIP, basic region/leucine zipper; CRE, cAMP response element; CREB, cAMP response element binding protein; CREM, cAMP response element modulator protein; HAT, histone acetyltransferase; NP, not pregnant; P, pregnant nonlaboring; S, spontaneous laboring.
Received October 22, 2001.
Accepted December 15, 2001.
References
s in prenancy. Endocrinology 132:2484–2490s in human myometrium in term and preterm labour: a mechansim for parturition. J Clin Endocrinol Metab 79:1835–1839[Abstract]
in the human adrenocortical cancer cell line H295R. J Clin Endocrinol Metab 85:345–354s proteins in the upper and lower segments of the human uterus during pregnancy and labour. J Clin Endocrinol Metab 84:1705–1710This article has been cited by other articles:
 |
|
 |
|
  S. KNIPPEN, T. LONING, V. MULLER, C. SCHRODER, F. JANICKE, and K. MILDE-LANGOSCH Expression and Prognostic Value of Activating Transcription Factor 2 (ATF2) and its Phosphorylated Form in Mammary Carcinomas Anticancer Res, January 1, 2009; 29(1): 183 - 189. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. R. Chapman, I. Smyrnias, D. O. C. Anumba, G. N. Europe-Finner, and S. C. Robson Expression of the GTP-Binding Protein (G{alpha}s) Is Repressed by the Nuclear Factor {kappa}B RelA Subunit in Human Myometrium Endocrinology, November 1, 2005; 146(11): 4994 - 5002. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J. Tyson-Capper, J. Bailey, A. R. Krainer, S. C. Robson, and G. N. Europe-Finner The Switch in Alternative Splicing of Cyclic AMP-response Element Modulator Protein CREM{tau}2{alpha} (Activator) to CREM{alpha} (Repressor) in Human Myometrial Cells Is Mediated by SRp40 J. Biol. Chem., October 14, 2005; 280(41): 34521 - 34529. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. Otun, M. W.J. MacDougall, J. Bailey, G. N. Europe-Finner, and S. C. Robson Spatial and Temporal Expression of the Myometrial Mitogen-Activated Protein Kinases p38 and ERK1/2 in the Human Uterus During Pregnancy and Labor Reproductive Sciences, April 1, 2005; 12(3): 185 - 190. [Abstract] [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]
|
|
|
|
|
|
J. Bailey and G N. Europe-Finner Identification of human myometrial target genes of the c-Jun NH2-terminal kinase (JNK) pathway: the role of activating transcription factor 2 (ATF2) and a novel spliced isoform ATF2-small J. Mol. Endocrinol., February 1, 2005; 34(1): 19 - 35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Yao, N. Hiramatsu, Y. Zhu, T. Morioka, M. Takeda, T. Oite, and M. Kitamura Nitric Oxide-Mediated Regulation of Connexin43 Expression and Gap Junctional Intercellular Communication in Mesangial Cells J. Am. Soc. Nephrol., January 1, 2005; 16(1): 58 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. I. Azzam, S. M. de Toledo, and J. B. Little Expression of CONNEXIN43 Is Highly Sensitive to Ionizing Radiation and Other Environmental Stresses Cancer Res., November 1, 2003; 63(21): 7128 - 7135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. SAEZ, V. M. BERTHOUD, M. C. BRANES, A. D. MARTINEZ, and E. C. BEYER Plasma Membrane Channels Formed by Connexins: Their Regulation and Functions Physiol Rev, October 1, 2003; 83(4): 1359 - 1400. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
