Molecular Human Reproduction, Vol. 6, No. 7, 648-660,
July 2000
© 2000 European Society of Human Reproduction and Embryology
Pregnancy |
Expression of the cyclic AMP-dependent transcription factors, CREB, CREM and ATF2, in the human myometrium during pregnancy and labour
Department of Obstetrics and Gynaecology, University of Newcastle upon Tyne, Royal Victoria Infirmary, 4th Floor Leazes Wing, Newcastle upon Tyne NE1 4LP, UK
Abstract
Elevated concentrations of cyclic AMP (cAMP) in the human myometrium may promote uterine quiescence during pregnancy by protein kinase A (PKA)-mediated phosphorylation and subsequent inactivation of myosin light-chain kinase, as well as by the phosphorylation and activation of cAMP-dependent transcription factors. In this context, we show that the altered expression of cAMP response-element binding protein (CREB), cAMP response-element modulator protein (CREM) and activating transcription factor 2 (ATF2) are implicated in the maintenance of myometrial quiescence during fetal maturation and the switch to uterine activation at term. Using electrophoretic mobility shift and super shift assays, as well as immunoblotting of paired myometrial tissue samples from non-pregnant, pregnant non-labouring and spontaneous labouring women, we defined the patterns of expression of various isoforms of these proteins in the human uterus. Here, we report spatio–temporal changes in the expression of a 43 kDa form of CREB, a 28 kDa CREM-like protein, and a novel 28 kDa ATF2-like protein which are differentially expressed, depending on the gestational state of the uterus. Changes in the pattern of expression of these potent transcription factors may have an important role in the control of uterine activity throughout pregnancy.
bZIP/cAMP/CREM/myometrium/transcription factors
Introduction
Elucidation of the myometrial mechanisms that regulate uterine activity during human pregnancy remains a high priority in the search for better therapeutic agents to control the incidence of premature labour. Although a great deal of effort has recently been focused on the study of nitric oxide synthase (NOS) in the maintenance of uterine quiescence during pregnancy, there is at present only controversial evidence to support a role for nitric oxide (NO) as a major relaxatory agent of human uterine smooth muscle cells (Weiner et al., 1994
; Kuenzli et al., 1996; Nakanishi et al., 1996; Jones and Poston, 1997; Sladek et al., 1997; Thomson et al., 1997; Hennan and Diamond, 1998). However, the role of cyclic AMP (cAMP) in the regulation of myometrial quiescence is supported by several studies indicating that increased formation of cAMP during pregnancy is potentiated by up-regulation of isoforms of the GTP-binding protein (G-protein) G
s (Europe-Finner et al., 1993, 1994; Lopez Bernal et al., 1995); this is coupled to increased human chorionic gonadotrophin (HCG/LH) receptor expression/HCG concentrations (Zuo et al., 1994) and the progesterone-induced decrease in cAMP phosphodiesterase activity observed during pregnancy (Kofinas et al., 1990). Increased cAMP accumulation during gestation may potentially affect the expression of other key genes involved in regulating the activity of the human uterus throughout fetal maturation and labour. This hypothesis is supported by previous findings (Ambrus and Rao, 1994) which showed that incubation of human myometrial cells in culture with HCG for 4–8 h directly decreases expression of connexin-43 gap junction mRNA and protein concentrations via activation of protein kinase A (PKA). Connexin-43 plays an important part in vivo in regulating uterine contractions during labour and its expression increases in the myometrium prior to the onset of parturition (Chow and Lye, 1994), and this may be related to the reciprocal decrease in concentrations of expression of HCG receptors and Gs. Moreover, observations by Yang et al. (1997) indicate that over-expression of Gs in transfected cells increases events distal to cAMP accumulation, resulting in phosphorylation of cAMP-dependent transcription factors by PKA and the subsequent activation of gene expression in these cells. The most studied cAMP-dependent transcription factors include the cAMP response-element binding protein (CREB; Hoeffler et al., 1988) and modulator protein (CREM; Foulkes et al., 1991), and members of the activating transcription factor (ATF) sub-family. These factors bind to specific regulatory sequences in the promoters of the genes that they control (Lalli and Sassone-Corsi, 1994), and are all members of the basic region/leucine zipper (bZIP) superfamily of proteins (Landschulz et al., 1988; Ziff, 1990). The best characterized of these binding motifs is the cAMP-response element (CRE), a consensus palindromic sequence in the form 5'-TGACGTCA-3' to which proteins from the CREB/CREM/ATF family of transcription factors bind as homo- or heterodimers (Sassone-Corsi, 1988; Borrelli et al., 1992). Both CREB and CREM are derived from multi-exonic genes, the alternative splicing of which generates a complex array of isoproteins that can act as both activators and repressors of transcription (Sun et al., 1992; Walker et al., 1994; Habener et al., 1995; Sanborn et al., 1997). While ATF2 and ATF3 are products of distinct genes (Maekawa et al., 1989; Gaire et al., 1990; Kara et al., 1990; Adam et al., 1996; Liang et al., 1996), ATF1 is known to be highly homologous to CREB lacking a glutanide rich region just upstream of the phosphorylation regulatory domain (P-box; Reyfuss et al., 1999). Furthermore, ATF4 (also known as CREB2) is homologous with ATF1, but also lacks a transcription activation domain (TAD) and a PKA site to act as a repressor (Karpinski et al., 1992). Interestingly, ATF2 is regulated by mitogen-activated protein kinase (MAPK) rather than PKA (Abdel-Hafiz et al., 1992), which adds another dimension to the control of expression of genes with CRE-containing promoters in tissues in which ATF2 is expressed.
Although there is a substantial amount of data indicating an increase in cAMP concentrations in the human myometrium during pregnancy, at present there is little evidence implicating a role for cAMP-dependent transcription factors in regulating myometrial gene expression during fetal maturation. However, several gene products specifically involved in controlling myometrial quiescence during pregnancy and the switch to uterine activation at the onset of parturition have been found to contain symmetrical and asymmetrical CREs within their promoter regions; these include the ß2-adrenoceptor, oxytocin receptor (OTR), cyclo-oxygenase 2 (COX-2) and connexin-43 genes. Moreover, previous studies (Fuchs et al., 1984; Moonen et al., 1986; Sparey et al., 1999) have shown that myometrial connexin-43, OTR and COX-2 proteins appear to be spatially regulated within different regions of the uterus during fetal maturation; therefore a primary objective of this present study was to investigate the spatio-temporal expression of total cellular concentrations of CREB, CREM and ATF proteins in the uterus. This was accomplished using paired myometrial tissues sampled from the upper (corpus) and lower uterine segments from non-pregnant, pregnant non-labouring and spontaneous labouring women in immunoblotting, immunoprecipitation and two-dimensional sodium dodecyl sulphate–polyacrylamide gel electrophoresis (2-D SDS–PAGE) experiments with antisera to various isoforms of CREB, CREM and ATF. In addition a further objective was to demonstrate the ability of these myometrial proteins, by employing tissue and cell lysates in electrophoretic mobility shift and super shift assays (EMSAs), to bind DNA oligonucleotide primers containing the CREs within the ß2-adrenoceptor, OTR, COX-2, connexin-43, HCG and rat somatostatin (as a control palindromic CRE sequence) genes. The antibodies used in these experiments, and the degree of their specificity, are described below.
Materials and methods
Reagents
All electrophoretic reagents, including low molecular weight pre-stained and unstained standards, were of the highest grade available and obtained from Bio-Rad (Richmond, CA, USA). Antibodies to CREB (06–504) and phosphoCREB (06–519) were obtained from Upstate Biotechnology (Lake Placid, NY, USA). Antibodies to CREB-1 (CREB-specific, directed towards an epitope corresponding to amino-acids 254–327; X-12), CREM-1 (CREB/CREM/ATF specific; X-12), ATF1-4 (C41-5.1; F2BR-1; C-19; C-19 respectively) c-Fos, c-Jun and full length CREM protein were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Similarly electrophoretic mobility super shift antibodies to CREB-1, CREM-1 and ATF1-4 were also obtained from Santa Cruz Biotechnology. The Gß antibody (SW/1) was from NEN (USA). Horseradish peroxidase conjugated goat anti-rabbit immunoglobulin G (IgG)/goat anti-mouse IgG were from Dako Ltd (High Wycombe, Bucks, UK). The enhanced chemiluminesence assay system (ECL) was obtained from Amersham International plc (Aylesbury, Buckinghamshire, UK). Protein A agarose was obtained from Calbiochem (Nottingham, UK). Ampholytes for 2-D SDS–PAGE were from BioRad.
Selection of patients and tissue collection
Non-pregnant myometrium
Samples of myometrium from non-pregnant premenopausal women (aged 32–46 years) 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 ~5–10 mm away from the endometrial or serosal surfaces. Samples obtained in both the follicular and luteal phases of the cycle were used. For comparison of upper and lower uterine segments, paired samples of myometrium were taken from the corpus and the lower region close to the cervix, although it must be realized that the lower uterine segment is poorly defined in the non-pregnant uterus from individual women.
Pregnant non-labouring and spontaneous labouring myometrium
A total of 25 pregnant non-labouring women undergoing elective lower segment Caesarean section at 38–39 weeks gestation were recruited from the antenatal ward prior to surgery. The second group consisted of 25 women in spontaneous labour at term who required emergency Caesarean section for fetal distress or failure to progress. Spontaneous labour was defined as regular uterine contractions and cervical dilatation of >3 cm on admission. Augmentation of labour with oxytocin was not a reason for exclusion, provided the actual onset of labour had been spontaneous. Routine Caesarean section was carried out under sub-arachnoid block or general anaesthetic. After delivery of the infant, placental site was confirmed manually and the placenta then delivered. Five non-fundal (corpus), upper segment biopsies (100–250 mg) were then taken from within the uterine cavity using laparoscopic biopsy forceps (Wolf) introduced through the lower segment incision. Biopsies were taken from the side opposite to the placental bed to help minimize the inclusion of decidua. In addition, a 1–2 cm sample of myometrium was taken from the upper margin of the lower uterine segment incision using tissue forceps and scissors. Great care was taken so as not to sample cervical tissue. Note that paired upper and lower segment tissue samples were taken from all women in each patient group. The myometrial samples were snap frozen in liquid nitrogen and stored at –700C. 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.
Preparation of myometrial homogenates and cell lysates for Western blotting/mobility shift assays
All procedures were carried out on ice. (i) Tissue samples were homogenized at a ratio of 1:10 in 25 mmol/l Tris buffer, pH 7.6, containing 0.25 mol/l sucrose and 1 mmol/l EDTA in the presence of pepstatin, leupeptin, aprotinin and phenylmethylsulphonyl fluoide (PMSF) protease inhibitors. Homogenates were then centrifuged at 1000 g to remove tissue debris, the genomic DNA sheared by centrifugation through QIAshredder columns (QIAgen, Crawley, UK) and the supernatants stored at –70°C. (ii) Primary myometrial cell cultures were prepared from non-pregnant/pregnant non-labouring myometrial biopsies as described previously (Phaneuf et al., 1993) and used prior to passage three. Whole cell lysates were prepared by resuspending harvested and washed cells in 150 µl buffer 1 per confluent T75 flask [10 mmol/l HEPES (pH 7.9), 10 mmol/l KCl, 1.5 mmol/l MgCl2, 0.5 mmol/l dithiothreitol (DTT), 1 mmol/l sodium orthovanadate, 1 mmol/l NaF, 1% NP-40, plus protease inhibitors as above]. The samples were vortexed, incubated on ice for 10 min, and micro-centrifuged at 7000 rpm (3000 g) for 1 min. Pellets were resuspended in 30 µl of buffer 2 per original T75 flask [20 mmol/l HEPES (pH 7.9), 0.42 mol/l NaCl, 1.2 mmol/l MgCl2, 0.5 mmol/l DTT, 0.3 mmol/l EDTA, 25% glycerol, plus protease inhibitors as above]. After incubation on ice for 1 h with occasional vortexing, the samples were micro-centrifuged at 13 000 rpm (12 000 g) for 30 min at 4°C, and the supernatants stored at –70°C.
When the homogenates/lysates were to be used in EMSAs, they were dialysed prior to storage against a 1000x volume of 20 mmol/l HEPES (pH 7.9), 5 mmol/l MgCl2, 1 mmol/l EDTA, 70 mmol/l KCl, 5 mmol/l DTT, 10% glycerol at 4°C overnight. In all cases, the total protein content of each sample was assayed using the DC protein assay kit (BioRad) with bovine serum albumin (BSA) as standard.
Western blot immunodetection of myometrial proteins
Total protein (500 µg) from each myometrial homogenate was solubilized in sample buffer and resolved on 12.5% polyacrylamide gels containing 0.0625% bis-acrylamide for 4 h at 50 mA and then elecrotransferred onto a polyvinylidine difluoride (PVDF) membrane (Hybond-P; Amersham) at 90 V for 2 h using a Trans-Blot cell (BioRad). These membranes were subjected to a blocking step for 1 h at room temperature in 5% non-fat milk. Primary antibodies to CREB and phospho–CREB were used at 4 µg/ml, and to CREM (anti-CREM-1) and ATF1-4 at 0.15 µg/ml, all in the presence of 5% non-fat milk and 0.05% Tween 20 in 1x phosphate buffered saline (PBS) overnight at 4°C. The primary antiserum was then removed, and blots washed thoroughly with PBS. Blots were then incubated with goat anti-rabbit IgG or goat anti-mouse IgG coupled to horseradish peroxidase at a 1:2000 dilution in PBS for l h at room temperature, then rinsed thoroughly with PBS as before. Antibody complexes were detected by ECL using X-O-Mat X-ray film. Data were obtained under conditions where a linear relationship existed between the amount of protein loaded and the intensity of the ECL signal from the immunoblots.
In all cases, the loading of equivalent amounts of total protein for each sample was confirmed by Ponceau S staining of the PVDF membrane after electroblotting and subsequent densitometric scanning. For blots involving myometrial tissue from the upper and lower uterine segments, where the latter may contain a higher amount of connective tissue, equivalent smooth-muscle protein loading was also confirmed by immunodetection with an antibody common to all Gß subunits, which are known to be expressed to similar values in the myometrial cells of the non-pregnant, pregnant non-labouring and spontaneous labouring uterus (Europe-Finner et al., 1994; Hattachote et al., 1998; Sparey et al., 1999).
Immunoprecipitation of proteins
Cleared myometrial homogenate was dialysed against a 1000x volume of 1x PBS, 0.5% deoxycholic acid, 0.1% SDS, 0.1% Triton X-100 (all from Sigma), overnight at 4°C. This was pre-cleared to minimize non-specific reactions with the precipitating primary serum by incubation with 5 µg of normal IgG from the same species as the antibody and 20 µl of protein A agarose suspension with agitation at 4°C for 30 min. The agarose was pelleted by microcentrifugation at 3000 rpm (700 g) for 5 min, and the supernatant was transferred to a fresh tube. 10 µg of the precipitating primary antibody was added per mg of total protein, and incubated at 4°C for 1 h with agitation. 15 µl of the protein A agarose suspension was then added, and incubation continued as above overnight. The sample was centrifuged as above, and the pellet washed with 1 ml of cold PBS, followed by another centrifugation step. This was repeated four times, after which the precipitated protein was eluted from the agarose by incubating with an equal volume of 2x Laemmli buffer at 80°C for 5 min. The sample was microcentrifuged at 13 000 rpm (12 000 g) for 5 min to pellet the agarose, and the eluted protein subjected to SDS–PAGE and Western blotting analysis as above.
Two-dimensional SDS–PAGE
Immunoprecipitated proteins were subjected to 2-D SDS–PAGE using the BioRad Mini-PROTEAN II 2-D cell and tube module, according to the published protocol. Briefly, 0.5 mg of total protein from cleared myometrial homogenate was precipitated using 0.5 µg of primary antibody and 50 µl of protein A agarose as above, and eluted in 50 µl of 1st dimension sample buffer (9.5 mol/l urea, 2% de-ionized Triton X-100, 5% ß-mercaptoethanol, 1.6% 5/7 ampholyte and 0.4% 3/10 ampholyte). Following pre-electrophoresis of the 1st dimension tube gels using 100 mmol/l NaOH and 10 mmol/l H3PO4 as the upper and lower chamber buffers respectively, 5 µl of this was loaded onto the surface of the tube gels, overlayed with 20 µl of sample overlay buffer (9 mol/l urea, 0.8% 5/7 ampholyte, 0.2% 3/10 ampholyte, Bromophenol Blue; diluted 5x prior to use), and electrophoresed at 750 V for 3.5 h. The gels were then extracted from their capillary tubes, and layered onto the surface of mini slab gels for the 2nd dimension run. The electrophoresed proteins were then blotted onto PVDF membranes for immunodetection, as above.
Electrophoretic mobility shift and super shift assays
Sense and antisense oligonucleotides corresponding to the CREs of the HCG (5'-ATGGTAAAAATTGACGTCATGGTAATTACA-3'), human ß2 adrenoceptor (hß2-AR; 5'-CGAAAGTTCCCGTACGTCACGGCGAGGGCA-3'), rat somatostatin (rat-SS; 5'-AGAGATTGCCTGACGTCAGAGAGCTAG-3'), COX-2 (5'-CCTATTAAGCGTCGTCACTAAAACATA-3'), connexin-43 (Cx-43; 5'-CAAGTAGAGGCGTTCATGTCCCTAAT-3') and OTR (5'-GGAGCGCACGCGTCACTGGGGCCGT-3') genes were end-labelled with [
32P]-ATP, annealed and purified using ProbeQuant G-50 microcolumns (Amersham-Pharmacia Biotech); OCT-1 sequence-containing oligonucleotides for use as a negative control (OCT-1 5'-TTCTAGTGATTTGCATTCGACA-3') were treated similarly. Approximately 40 000 cpm of these substrates were then used in each binding reaction, in a total volume of 15 µl with 1 µg of poly (dI-dC) (Pharmacia), 30 µg of total cellular protein, 25 mmol/l HEPES (pH 7.5), 10% glycerol, 5 mmol/l MgCl2, 100 mmol/l KCl, 1 mmol/l DTT, 0.2 mmol/l EDTA, 0.5 mg/ml BSA, and 200-fold cold competitor oligonucleotide where applicable in competition experiments. Competitors were added to the reactions 15 min prior to the addition of the probe, after which they were incubated for a further 15 min at room temperature. In the case of the super shift experiments, 2 µl of the appropriate antibody was incubated with the protein extract for 30 min at room temperature before the addition of the probe, and the reactions were continued as above. The samples were then subjected to 4% non-denaturing PAGE at 150 V for 3 h, after which time the gels were dried and exposed to X-ray film.
Results
Immunodetection and expression of myometrial CREB, CREM and ATF1-4 proteins
Immunoblotting of total protein extracts from non-pregnant, pregnant non-labouring and spontaneous labouring myometrial tissues sampled from the lower uterine segment with an anti-CREM-1 (CREB/CREM/ATF1 and ATF4) antibody showed a significant difference in the isoforms of these proteins expressed in the three types of tissue (Figure 1a). A protein of apparent molecular mass 43 kDa was prevalent almost exclusively in non-pregnant (NP) myometrium (Figure 1b), whereas a protein of ~28 kDa was detected primarily in pregnant non-labouring myometrium (P) which was subsequently expressed at much higher concentrations in spontaneous labouring (SL) tissues (Figure 1c). In all three tissue types, a 39 kDa protein was detected, the expression of which was significantly decreased in P and SL tissues, compared with NP myometrium (Figure 1d). The 43 and 39 kDa proteins reported here have similar apparent molecular weights to those reported previously (Habener et al., 1995; Müller et al., 1997) for CREB/CREM/ATF transcription factors. Furthermore, immunoblotting of a HeLa cell extract as a standard control (for identification of nuclear factors) also confirmed the identity of the proteins detected in the myometrial homogenates as CREB/CREM/ATF proteins.
|
Due to the cross-reactivity of the anti-CREM-1 (CREB/CREM/ATF) antibody with the CREB, CREM and ATF proteins (Santa Cruz product information), immunoblots were repeated using a specific anti-CREB antibody directed towards epitopes in the highly variable 200 amino-terminal residues of the protein; this antibody is designed to specifically recognize inactive and active forms of CREB but not CREM or ATF (Ginty et al., 1993
; Baler et al., 1997). Figure 2a shows the detection of the 43 kDa full-length protein primarily in non-pregnant tissues, with much reduced concentrations in pregnant and labouring tissue samples. This identifies this protein as an isoform of CREB; parallel blots using an anti-phospho-CREB antibody (directed towards an epitope comprising the phosphorylated Ser133 residue of full-length CREB and consequently recognising activated CREB/CREM/ATF proteins) suggest that the full-length 43 kDa CREB protein detected mainly in non-pregnant tissue is in its phosphorylated and active form (Figure 2b), although this result is not definitive due to the documented cross-reactivity of this antibody with other closely related proteins.
|
The specific recognition of only the 43 kDa protein species by the anti-CREB antibody identified the 39 kDa and 28 kDa proteins detected in Figure 1
as CREM/ATF transcription factors; these were further investigated by blotting appropriate protein samples from non-pregnant, pregnant and spontaneous labouring myometrial samples alongside purified full-length CREM protein and rat kidney lysate followed by immunodetection with anti-CREM antibody. Figure 3 shows the detection of a common 39 kDa protein in each tissue type (lanes 3–5), which co-migrates with the full-length CREM protein in lane 1. It was therefore concluded that the 39 kDa protein detected in the myometrial homogenates was the full-length CREM protein, whilst the 28 kDa protein that was not detected in the rat kidney lysate (lane 2) might be a derivative of CREM, one of the ATF proteins, or a novel protein which appears to be related to the bZIP family of transcription factors. In light of the following results we have termed this protein CREM-28.
|
Identical immunoblotting experiments were also performed using primary antibodies to the ATF proteins 1–4. Although 500 µg of tissue homogenate protein was loaded per lane, ATF1, ATF3 and ATF4 proteins remained undetected using antibodies specific to these proteins. In contrast, bands corresponding to proteins of approximate molecular mass 60 kDa and 28 kDa were identified using the antibody to ATF2 (Figure 4a
). The 60 kDa protein, ostensibly the full-length 60–68 kDa ATF2 protein (Maekawa et al., 1989; Gaire et al., 1990; Kara et al., 1990; Adam et al., 1996; Liang et al., 1996), was found to be expressed mainly in non-pregnant myometrium at concentrations significantly higher than in pregnant non-labouring and spontaneous labouring tissues (Figure 4b). The 28 kDa protein (hereafter referred to as ATF2-28) was expressed in all tissue types, with concentrations of this protein decreasing significantly from the non-pregnant to the pregnant non-labouring/spontaneous labouring state (Figure 4c). In control experiments all antibodies were substituted with non-immune sera from animals in which the antibodies were raised (data not shown). In all cases immunodetected bands were found to be specific. The specific expression of CREB, CREM and ATF2 proteins in uterine smooth-muscle cells was also confirmed by immunohistochemical staining of paraffin-embedded myometrial sections prepared from the three tissue types (data not shown).
|
Expression of myometrial CREM (28 kDa) and ATF2-like (28 kDa) proteins in the upper and lower uterine segments during pregnancy and labour
From Western blot analysis the two main cAMP-dependent transcription factors that are expressed in the human myometrium during gestation and parturition appear to be the CREM 28 kDa and ATF2-like 28 kDa proteins. Consequently, further immunodetection experiments were performed to monitor the spatial expression of these proteins in myometrial tissues taken from the upper (corpus) and lower uterine regions using paired samples obtained from individual women (Sparey et al., 1999
). In all cases, the loading of equivalent amounts of total protein for each sample was confirmed by staining of the blotted membranes with Ponceau S solution (Sigma, UK) and scanning prior to immunodetection, and by re-probing the blots with an anti-Gß antibody. The Gß protein subunits have consistently been shown to be expressed to similar levels in the upper and lower uterine segment myometrium from pregnant non-labouring and spontaneous labouring patients (Sparey et al., 1999), and this is confirmed in Figure 5b,e. Immunoblotting of homogenized upper and lower myometrial samples using the anti-CREM-1 (CREB/CREM/ATF) and anti-ATF2 antibodies revealed that while myometrial CREM-28 was expressed to similar values in the upper and lower uterine regions during pregnancy and at increased concentrations in both uterine regions during labour as described previously, there was a consistent and striking difference between concentrations of expression of myometrial ATF2-28 in upper and lower uterine samples (Figure 5a,c,d). In this respect, a comparison of the optical density (OD) values that are a measure of the levels of expression of this protein indicated a decreased level of myometrial expression in the lower uterine segment (0.5 OD units) compared with the upper (1.5–2.0 OD units) during pregnancy and labour.
|
Expression of the myometrial ATF2 28 kDa protein in the upper and lower regions of the non-pregnant uterus
While it seems clear that ATF2-28 is expressed differentially in different regions of the human myometrium throughout pregnancy and labour, the question arose as to whether this gradient of ATF2-28 concentrations from the corpus to the lower segment of the uterus was induced at some point following conception, or whether it also existed in the non-pregnant uterus. To address this, further anti-ATF2 immunoblots were performed using samples of homogenized upper and lower myometrium from non-pregnant uteri. In all cases, results indicated the uniform expression of this protein in non-pregnant myometrial tissues (1.5–2.0 OD units; Figure 6a,b
). It is also worthy of note that the levels of expression of ATF2-28 in both the upper and lower non-pregnant uterine segments, as determined by OD values, were equivalent to the values observed for non-pregnant tissues in Figure 4c (where n = 25), and for upper pregnant/labouring myometrial segments in Figure 5d, thus suggesting that ATF2-28 protein expression is down-regulated specifically in the lower pregnant/labouring myometrial segment compared with non-pregnant tissue.
|
Further characterization of the myometrial CREM-28 and ATF2-28 proteins
As detailed above, the potentially novel 28 kDa CREM and ATF2-like proteins appear to be the most prevalent proteins of this type in the pregnant non-labouring and spontaneous labouring myometrium. An important property of these types of protein is their ability to form homo- and heterodimers, an attribute that enables such complexes to exert alternate effects upon the genes that they regulate. To determine the extent of any interaction between the CREM-28 and ATF2-28 proteins, and also to provide further evidence for the ATF2-28 protein being a possible isoform derived from the ATF2 gene, the anti-CREM-1 (CREB/CREM/ATF) and anti-ATF2 antibodies were used in separate immunoprecipitation reactions with pooled myometrial homogenates from lower segment biopsies of pregnant non-labouring and spontaneous labouring tissues. Subsequent SDS–PAGE and immunoblotting of the precipitated proteins with the same antibodies revealed only a slight interaction between CREM-28 and ATF2-28 as indicated by the low-level of co-precipitation of the non-antibody-specific protein. This is demonstrated in Figure 7
where immunoprecipitation with the anti-ATF2 antibody resulted in the weak co-precipitation of CREM-28 protein as detected by the anti-CREM-1 antibody; note that only very small amounts of CREB-43 and CREM-39 were precipitated in comparison to the high amounts of CREM-28. In parallel experiments the anti-CREM-1 antibody was also able to co-immunoprecipitate a small amount of ATF2-28, and in each case the specificity of the immunoprecipitation reactions was confirmed by the inclusion of appropriate negative control reactions (data not shown).
|
Further immunoprecipitation experiments were performed to investigate whether the ATF-28 protein is a product of a splice-variant of the ATF2 gene. These were in the form of co-immunoprecipitations with antibodies to the related bZIP transcription factors c-Jun and c-Fos; ATF2 has previously been shown by co-immunoprecipitation and Western blotting to interact with the former but not the latter (Maekawa et al., 1989
; Benbrook and Jones, 1990; Ivashkiv et al., 1990; Macgregor et al., 1990). Figure 8 shows an anti-ATF2 Western blot of the proteins immunoprecipitated by anti-ATF2, anti-c-Fos and anti-c-Jun antibodies (lanes 1–3 respectively); this demonstrates the formation of ATF2/c-Jun and ATF2-28/c-Jun heterodimers by the co-precipitation of ATF2 and ATF2-28 by the anti-c-Jun antibody (lane 3), but reveals no interaction of these proteins with c-Fos (lane 2). Parallel blots with anti-c-Jun and anti-c-Fos antibodies ratified this result (data not shown). Whilst it cannot be concluded that either the ATF2-60 or ATF-28 proteins do not interact with c-Fos (far-Western blotting experiments have demonstrated heterodimer formation between these two proteins; Hoeffler et al., 1991), these results provide further evidence of similarity between ATF2 and ATF2-28, and this may suggest a similar genetic basis.
|
Due to the multiple-exonic structure of the genes and the differential splicing of their primary transcripts (Foulkes et al., 1991
; Ruppert et al., 1992; Laoide et al., 1993; Meyer and Habener, 1993; Habener et al., 1995; Girardet et al., 1996; Walker et al., 1996; Gellersen et al., 1997), CREM and ATF proteins are known to exist in several distinct isoforms. To determine the degree of homogeneity of the constituent proteins of the bands representing CREM-28 and ATF2-28 following normal SDS–PAGE, the immunoprecipitates, which predominantly contain CREM-28 and ATF2-28 respectively, were subjected to 2-D PAGE. This technique, involving the separation of proteins according to their iso-electric points prior to separation in terms of size via conventional slab-gel electrophoresis, produced the results shown in Figure 9 after subsequent electroblotting and immunodetection. Figure 9a indicates the presence of three proteins specifically precipitated and detected by the anti-CREM-1 antibody. While the major upper band (apparent molecular weight 28 kDa) clearly represents the CREM-28 protein, the less prevalent accompanying two proteins with slightly increased electrophoretic mobilities may represent the same protein in altered phosphorylation states. Figure 9b shows the result of the 2-D PAGE immunoblot using the anti-ATF2 antibody, which identified the presence of six differentially-focussed proteins, each with a molecular weight of 28 kDa.
|
Electrophoretic mobility shift assays
Electrophoretic mobility shifts were observed with whole-cell lysates of cultured human myometrial cells and several symmetrical (5'-TGACGTCA-3') and asymmetrical (5'-CGTCA-3') CRE-containing 32P-labelled oligonucleotides. Figure 10a
shows an example of a mobility shift obtained with the palindromic-CRE containing rat somatostatin (rat-SS) oligonucleotide; similar shifts were also found with the palindromic HCG–CRE and hß2AR–CRE oligonucleotides, but not with the negative control non CRE-containing OCT-1 oligonucleotide (data not shown). Band shifts were also observed using the asymmetrical OTR–CRE (Figure 10b), as well as COX2–CRE (data not shown). For both symmetrical and asymmetrical CRE [32P]-labelled olignucleotides the observed shifts were efficiently competed away with a 200-fold excess of non-labelled CRE-containing oligonucleotides (Figure 10a, lanes 3 and 4; Figure 10b, lane 2), but not with the non-CRE containing OCT-1 sequence (Figure 10a, lane 5; Figure 10b, lane 3). Super shifts were also detected with antibodies specific for CREB, CREM and ATF 1–4 transcription factors (Figure 10a, lanes 7–12; Fig 10b, lanes 4–6), note that incubation with the ATF3 antibody causes a marked reduction in the amount of the labelled band shift obtained, compared with the other antibodies used, indicating a CRE-specific DNA-binding activity of these proteins in human myometrial cells. No super shifts were observed in controls using pre-immune serum in place of antibody (data not shown). In contrast to Western immunoblotting, ATF1, ATF3 and ATF4 antibodies were capable of super shifting bands and consequently their detection may be due to the greater sensitivity of the EMSA. Similar shift and super shifts were also obtained using whole tissue extracts from a pool of pregnant non-labouring/spontaneous labouring myometrial samples containing two or three individual samples from each patient group. Figure 10c shows a typical EMSA using the [32P]-labelled CRE containing rat-SS oligonucleotide with the super shift antibodies detailed previously; note that two super shift complexes were observed, complex I having a high intensity whereas complex II was of a lesser intensity (Figure 10c, lanes 3–8). The existence of one extra and one higher intensity complex with tissue-derived proteins may be due to higher levels of expression and/or the occurrence of different homo-/heterodimeric complexes, revealed by their interaction with the super shift antibodies. When using either tissue or myometrial whole cell lysates in conjunction with the labelled asymmetrical CRE-containing connexin-43 oligonucleotide, only non-specific bands and no shifts or super shifts were observed (data not shown).
|
Discussion
The results presented here demonstrate the ability of CREB, CREM and ATF1–4 proteins from both myometrial homogenates and cultured cells to bind a variety of symmetrical and asymmetrical CRE-containing DNA oligonucleotides, and more importantly reveal a significant switch in the isoforms of these proteins expressed in the human myometrium at the beginning or during the course of pregnancy. In summary; levels of CREB isoform of apparent molecular weight 43 kDa, the full-length and ubiquitously expressed 341 amino acid CREB-341 protein (Sun et al., 1992), were found to be significantly lower in pregnant non-labouring tissue compared to non-pregnant tissue; this protein was detected only at very low concentrations in spontaneous labouring myometrium. A CREM protein with an apparent molecular weight of 39 kDa followed a similar pattern of expression, although concentrations were not found to decrease in the pregnant non-labouring and spontaneous labouring tissues quite so dramatically. A putative CREM-like protein of apparent molecular weight 28 kDa, completely absent in the non-pregnant myometrium, was expressed to a high level in pregnant non-labouring tissues and up-regulated even further in the spontaneous labouring samples. Expression of the closely related transcription factors ATF1, ATF3 and ATF4 was revealed only by EMSA experiments and was undetected in all tissues by Western blotting, but two apparent isoforms of ATF2 were identified; one protein of apparent molecular weight 60 kDa, ostensibly the full-length ATF2 protein, was found to be expressed at a high level in the non-pregnant myometrium, at lower levels in pregnant non-labouring tissues, and at greatly decreased levels in spontaneous labouring samples. A previously unidentified protein of apparent molecular weight 28 kDa was also expressed at a high level in non-pregnant tissue, but appeared to be down-regulated to similar levels of expression in both the pregnant non-labouring and spontaneous labouring myometrium.
Further investigation revealed the differential expression of the ATF2-28 protein in the upper and lower regions of the uterus during pregnancy and labour. Comparison with the pattern of expression in non-pregnant tissues confirmed that the spatial change in concentrations of ATF2-28 occurred at some point following conception, evidently brought about by the down-regulation of ATF2-28 in the lower segment of the uterus. In contrast, expression of the prevalent CREM-28 protein remained at equivalent levels in the upper and lower uterine segments during pregnancy but was substantially increased in both regions of the uterus during parturition. Co-immunoprecipitation experiments demonstrated the weak ability of the CREM-28 and ATF2-28 proteins to form heterodimers, whereas ATF2-60 and ATF2-28 were both shown to form strong heterodimer complexes with c-Jun as has been previously described (Maekawa et al., 1989). Furthermore, 2-D PAGE analysis revealed the existence of six apparent forms of the ATF2-28 protein.
Although this present investigation may only define the capability of the above cAMP-dependent transcription factors to bind DNA containing CRE sequences and not their functionality in regulating specific myometrial genes, it is tempting to speculate about the possible roles of these proteins in modulating myometrial gene expression during pregnancy and the onset of labour at term. The switch between the expression of the CREB-43 and CREM-28 proteins in the transition from the non-pregnant to the pregnant non-labouring state suggests that these proteins, both potent activators and/or repressors of the transcription of target genes containing CRE elements in their promoters, may play an important role in the control of uterine activity throughout the 37–40 weeks of pregnancy. For example, the substantial reduction during pregnancy in the concentration of the putative full-length CREB protein may have far-reaching effects in terms of gene expression in human uterine smooth muscle cells; CREB is known not only to activate the transcription of CRE-containing genes (Brindle and Montminy, 1992), but also to be a potent inhibitor of the transcription of a diverse range of transcription factors (Lemaigre et al., 1993).
Similarly, the increase in expression of the CREM-28 protein prior to the onset of labour coupled with a gradient of expression of ATF2-28 from the corpus to the cervix of the uterus suggest that these transcription factors are possibly involved in defining the spatial regulation of genes specifically involved in regulating uterine activity during fetal maturation. In this context, previous studies (Fuchs et al., 1984; Sparey et al., 1999) have respectively shown that expression of OTRs, myometrial gap-junction protein connexin-43 and COX-2 are also spatially regulated in different regions of the uterus during pregnancy and labour. Not only were these gene products found to be spatially expressed in the pregnant uterus in common with ATF2-28, but their expression may be directly influenced by the bZIP family of transcription factors via the multiple CRE elements present in their promoter regions (Appleby et al., 1994; Chen et al., 1995,1998; Lefebvre et al., 1995; Miller et al., 1998). Although this present study did not indicate specific DNA binding to the labelled connexin-43 oligonucleotide by the CREB, CREM or ATF1–4 proteins, it is known that the connexin-43 promoter contains multiple binding sites for the c-Jun/c-Fos group of bZIP transcription factors (AP1 binding sites). These bZIP proteins are known to form heterodimeric complexes (Benbrook and Jones, 1990; Ivashkiv et al., 1990; Macgregor et al., 1990) with members of the CREB/CREM/ATF group of bZIP factors, therefore the formation of such complexes may indirectly affect the expression of the myometrial connexin-43 gene as well as other genes containing AP-1 and CRE elements (Angel and Karin, 1992; Geimonen et al., 1996). Like ATF2-28, Sparey et al. (1999) found the gradient of connexin-43 protein expression to be from the corpus to the cervix, whereas the gradient of expression of COX-1 and COX-2 was found to be the converse with the highest concentrations of both enzymes observed in the lower, compared with the upper, uterine segment.
These investigations are consistent with the idea that the differential expression of COX-1 and COX-2 and connexin-43 in the uterus may allow cervical ripening prior to and dilatation during labour and facilitate effective propagation of contractions from corpus to cervix, as suggested by Sparey et al. (1999). This pattern of myometrial gene expression in the uterus prior to and during labour may well be aided in part by the differential expression of CREB, CREM and ATF2 transcription factors described here. Recently, it has been shown (Stevens et al., 1998) that the myometrial corticotrophin-releasing hormone receptor sub-type 1 (CRH-R1) is also differentially expressed within the human uterus during gestation and parturition with the highest levels expressed in the lower, compared with the upper, uterine segment.
In summary, we present further evidence to support a role for cAMP in the control of uterine activity throughout weeks 37–40 of pregnancy and labour. It appears that the CREM-28 and ATF2-28 proteins identified here may be important due to the differences in their spatial and temporal expression. It is therefore our immediate intention to clone and characterize these factors, and to further elucidate their roles in the control of uterine activity during pregnancy.
Acknowledgments
This work was funded by the Wellcome Trust (grant number 053563).
Notes
1 To whom correspondence should be addressed at: Department of Obstetrics and Gynaecology, University of Newcastle upon Tyne, Royal Victoria Infirmary, 4th Floor Leazes Wing, Newcastle upon Tyne NE1 4LP, UK. E-mail: jarrod.bailey{at}ncl.ac.uk 
References
Abdel-Hafiz, H.A., Heasley, L.E., Kyriakis, J.M. et al. (1992) Activating transcription factor-2 DNA-binding activity is stimulated by phosphorylation catalyzed by p42 and p54 microtubule-associated protein kinases. Mol. Endocrinol., 6, 2079–2089.[Abstract]
Adam, E., Kerkhofs, P., Mammerickx, M. et al. (1996) The CREB, ATF-1, and ATF-2 transcription factors from bovine leukemia virus-infected B lymphocytes activate viral expression. J. Virol., 70, 1990–1999.[Abstract]
Ambrus, G. and Rao, C.V. (1994) Novel regulation of pregnant human myometrial smooth muscle cell gap junctions by human chorionic gonadotropin. Endocrinology, 135, 2772–2779.[Abstract]
Angel, P. and Karin, M. (1992) Specific members of the Jun protein family regulate collagenase expression in response to various extracellular stimuli. Matrix, 1 (Suppl.), 156–164.[Medline]
Appleby, S.B., Ristimaki, A., Neilson, K. et al. (1994) Structure of the human cyclo-oxygenase-2 gene. Biochem. J., 302, 723–727.
Baler, R., Covington, S. and Klein, D.C. (1997) The rat arylalkylamine N-acetyltransferase gene promoter. cAMP activation via a cAMP-responsive element-CCAAT complex. J. Biol. Chem., 272, 6979–85.
Benbrook, D.M. and Jones, N.C. (1990) Heterodimer formation between CREB and JUN proteins. Oncogene, 5, 295–302.[ISI][Medline]
Borrelli, E., Montmayeur, J.P., Foulkes, N.S. et al. (1992) Signal transduction and gene control: the cAMP pathway. Crit. Rev. Oncog., 3, 321–338.[Medline]
Brindle, P.K. and Montminy, M.R. (1992) The CREB family of transcription activators. Curr. Opin. Genet. Dev., 2, 199–204.[Medline]
Chen, Z.Q., Lefebvre, D., Bai, X.H. et al. (1995) Identification of two regulatory elements within the promoter region of the mouse connexin 43 gene. J. Biol. Chem., 270, 3863–3868.
Chen, G., Kamal, M., Hannon, R. et al. (1998) Regulation of cyclo-oxygenase gene expression in rat smooth muscle cells by catalase. Biochem. Pharmacol., 55, 1621–1631.[ISI][Medline]
Chow, L. and Lye, S.J. (1994) Expression of the gap junction protein connexin-43 is increased in the human myometrium toward term and with the onset of labor. Am. J. Obstet. Gynecol., 170, 788–795.[ISI][Medline]
Europe-Finner, G.N., Phaneuf, S., Watson, S.P. et al. (1993) Identification and expression of G-proteins in human myometrium: up- regulation of G alpha s in pregnancy. Endocrinology, 132, 2484–2490.[Abstract]
Europe-Finner, G.N., Phaneuf, S., Tolkovsky, A.M. et al. (1994) Down-regulation of G alpha s in human myometrium in term and preterm labour: a mechanism for parturition. J. Clin. Endocrinol. Metab., 79, 1835–1839.[Abstract]
Foulkes, N.S., Borrelli, E. and Sassone-Corsi, P. (1991) CREM gene: use of alternative DNA-binding domains generates multiple antagonists of cAMP-induced transcription. Cell, 64, 739–749.[ISI][Medline]
Fuchs, A., Fuchs, F., Husslein, P. et al. (1984) Oxytocin receptors in the human uterus during pregnancy and parturition. Am. J. Obstet. Gynecol., 150, 734–741.[ISI][Medline]
Gaire, M., Chatton, B. and Kedinger, C. (1990) Isolation and characterization of two novel, closely related ATF cDNA clones from HeLa cells. Nucleic Acids Res., 18, 3467–3473.
Geimonen, E., Jiang, W., Ali, M. et al. (1996) Activation of protein kinase C in human uterine smooth muscle induces connexin-43 gene transcription through an AP-1 site in the promoter sequence. J. Biol. Chem., 271, 23667–23674.
Gellersen, B., Kempf, R. and Telgmann, R. (1997) Human endometrial stromal cells express novel isoforms of the transcriptional modulator CREM and up-regulate ICER in the course of decidualization. Mol. Endocrinol., 11, 97–113.
Ginty, D.D., Kornhauser, J.M., Thompson, M.A. et al. (1993) Regulation of CREB phosphorylation in the suprachiasmatic nucleus by light and a circadian clock. Science, 260, 238–241.
Girardet, C., Walker, W.H. and Habener, J.F. (1996) An alternatively spliced polycistronic mRNA encoding cyclic adenosine 3',5'-monophosphate (cAMP)-responsive transcription factor CREB (cAMP response element-binding protein) in human testis extinguishes expression of an internally translated inhibitor CREB isoform. Mol. Endocrinol., 10, 879–891.[Abstract]
Habener, J.F., Miller, C.P. and Vallejo, M. (1995) cAMP-dependent regulation of gene transcription by cAMP response element-binding protein and cAMP response element modulator. Vitam. Horm., 51, 1–57.[ISI][Medline]
Hatthachote, P., Morgan, J., Dunlop, W. et al. (1998) Gestational changes in the levels of transforming growth factor-beta1 (TGFbeta1) and TGFbeta receptor types I and II in the human myometrium. J. Clin. Endocrinol. Metab., 83, 2987–2992.
Hennan, J.K. and Diamond, J. (1998) Evidence that spontaneous contractile activity in the rat myometrium is not inhibited by NO-mediated increases in tissue levels of cyclic GMP. Br. J. Pharmacol., 123, 959–967.[ISI][Medline]
Hoeffler, J.P., Meyer, T.E., Yun, Y. et al. (1988) Cyclic AMP-responsive DNA-binding protein: structure based on a cloned placental cDNA. Science, 242, 1430–1433.
Hoeffler, J.P., Lustbader, J.W. and Chen, C.Y. (1991) Identification of multiple nuclear factors that interact with cyclic adenosine 3',5'-monophosphate response element-binding protein and activating transcription factor-2 by protein-protein interactions. Mol. Endocrinol., 5, 256–266.[Abstract]
Ivashkiv, L.B., Liou, H.C., Kara, C.J. et al. (1990) mXBP/CRE-BP2 and c-Jun form a complex which binds to the cyclic AMP, but not to the 12-O-tetradecanoylphorbol-13-acetate response element. Mol. Cell. Biol., 10, 1609–1621.
Jones, G.D. and Poston, L. (1997) The role of endogenous nitric oxide synthesis in contractility of term or preterm human myometrium. Br. J. Obstet. Gynaecol., 104, 241–245.[ISI][Medline]
Kara, C.J., Liou, H.C., Ivashkiv, L.B. et al. (1990) A cDNA for a human cyclic AMP response element-binding protein which is distinct from CREB and expressed preferentially in brain. Mol. Cell. Biol., 10, 1347–1357.
Karpinski, B.A., Morle, G.D., Huggenvik, J. et al. (1992) Molecular cloning of human CREB-2: an ATF/CREB transcription factor that can negatively regulate transcription from the cAMP response element. Proc. Natl Acad. Sci. USA, 89, 4820–4824.
Kofinas, A.D., Rose, J.C., Koritnik, D.R. et al. (1990) Progesterone and estradiol concentrations in non-pregnant and pregnant human myometrium. Effect of progesterone and estradiol on cyclic adenosine monophosphate-phosphodiesterase activity. J. Reprod. Med., 35, 1045–1050.[ISI][Medline]
Kuenzli, K.A., Bradley, M.E. and Buxton, I.L. (1996) Cyclic GMP-independent effects of nitric oxide on guinea-pig uterine contractility. Br. J. Pharmacol., 119, 737–743.[ISI][Medline]
Lalli, E. and Sassone-Corsi, P. (1994) Signal transduction and gene regulation: the nuclear response to cAMP. J. Biol. Chem., 269, 17359–17362.
Landschulz, W.H., Johnson, P.F. and McKnight, S.L. (1988) The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins. Science, 240, 1759–1764.
Laoide, B.M., Foulkes, N.S., Schlotter, F. et al. (1993) The functional versatility of CREM is determined by its modular structure. Embo J., 12, 1179–1191.[ISI][Medline]
Liang, G., Wolfgang, C.D., Chen, B.P. et al. (1996) ATF3 gene. Genomic organization, promoter, and regulation. J. Biol. Chem., 271, 1695–1701.
Lefebvre, D.L., Piersanti, M., Bai, X.H. et al. (1995) Myometrial transcriptional regulation of the gap junction gene, connexin-43. Reprod. Fertil. Dev., 7, 603–611.[Medline]
Lemaigre, F.P., Ace, C.I. and Green, M.R. (1993) The cAMP response element binding protein, CREB, is a potent inhibitor of diverse transcriptional activators. Nucleic Acids Res., 21, 2907–2911.
Lopez Bernal, A., Europe-Finner, G.N., Phaneuf, S. et al. (1995) Preterm labour: a pharmacological challenge. Trends Pharmacol. Sci., 16, 129–133.[Medline]
Macgregor, P.F., Abate, C. and Curran, T. (1990) Direct cloning of leucine zipper proteins: Jun binds cooperatively to the CRE with CRE-BP1. Oncogene, 5, 451–458.[ISI][Medline]
Maekawa, T., Sakura, H., Kanei-Ishii, C. et al. (1989) Leucine zipper structure of the protein CRE-BP1 binding to the cyclic AMP response element in brain. EMBO J., 8, 2023–2028.[ISI]