|Home | About | Journals | Submit | Contact Us | Français|
We aimed to produce an estrogen-responsive reporter plasmid that would permit monitoring of estrogen receptor function in the uterus in vivo. The plasmid pBL-tk-CAT(+)ERE was induced by estrogen in bovine endometrial stromal cells. When the CAT gene was replaced by the secreted alkaline phosphatase SeAP, the resulting construct pBL-tk-SeAP(+)ERE remained estrogen responsive. However when the tk promoter was replaced by the cytomegalovirus (cmv) promoter, the resulting plasmid (pBL-cmv-SeAP(+)ERE) was not estrogen responsive. Inhibition of ERE function was not due to an effect in trans or due to lack of estrogen receptor. It was not due to an interaction between the cmv promoter and the SeAP gene. cmv promoter function was dependent on NF-κB, and mutagenesis in the NF-κB sites reduced basal reporter expression without imparting responsiveness to estrogen. A mutation in the TATA box also failed to impart estrogen responsiveness. Modeling of DNA accessibility indicated the ERE was inserted at a site accessible to transcription factors. We conclude that the cmv promoter inhibits ERE function in cis when the two sequences are located in the same construct, and that this effect does not involve an interaction between cmv and reporter gene, NF-κB sites or the TATA box, or DNA inaccessibility.
Steroid hormone receptors are transcription factors that, together with co-activators, control gene expression in target tissues. Among the techniques developed to assess the function of transcription factors is the use of reporter constructs. These are DNA plasmids containing a reporter gene under the transcriptional control of an enhancer sensitive to the transcription factor it is required to test. On transfection into cells, the endogenous transcription factor activates expression of the reporter gene; levels of the gene product reflect the rate of transcription, and hence transcription factor activity. Specificity for a particular transcription factor is imparted by limiting the number of effective enhancers in the reporter plasmid promoter region.
In common with other transcription factors, the activity of a steroid hormone receptor is potentially affected by a number of processes. One of these is binding of the steroid hormone to the receptor molecule. In addition, steroid hormone receptors may be activated by phosphorylation through protein kinase cascades induced by extra-cellular signals.
Reporter constructs have been used widely to investigate steroid hormone receptor function in cells in tissue culture. However, they have not been generally applied to tissues in vivo. The uterus, being comparatively accessible through endoscopy or with an artificial-insemination cannula, is a tissue in which use of this technique in vivo would be beneficial.1 All uterine tissues are targets for steroid hormones; however, although we know a good deal about the control of steroid hormone receptors in tissue-culture systems, it has proved difficult to confirm these findings in vivo. It is established that plasmid expression vectors are transcribed following administration to the uterine lumen.2,3 However, no experiments have been carried out with steroid-responsive reporter plasmids.
The aim of the work described here was to produce a reporter plasmid expressed in an estrogen-dependent manner on administration into the uterine lumen. Because previous experiments had shown that expression plasmids containing the immediate early cmv promoter are expressed in the endometrium,2,3 we utilized the cmv promoter. Furthermore, we used a reporter gene with a product that is secreted from cells, so that it could be measured by flushing fluid through the uterine lumen. The strategy therefore was, firstly, starting with an estrogen-responsive plasmid with a thymidine kinase (tk) promoter, to replace the chloramphenicol acetyl transferase (CAT) reporter gene with the gene coding for a modified form of the heat-stable, L-homoarginine-resistant placental alkaline phosphatase (SeAP). This truncated form of the phosphatase, which is not synthesized in mammalian tissues, is engineered to ensure its secretion by cells expressing it. Subsequently, the thymidine kinase promoter would be replaced with the cmv promoter. To confirm the functions of these plasmids, we proposed to test them in bovine stromal (BST) cell culture (bovine endometrial stromal cells) before administration in vivo.
Structures of plasmids used are shown diagrammatically in Figure 11.
The estrogen response element (ERE) construct, pBL-tk-CAT(+)ERE, was kindly provided by Professor M.G. Parker (Institute of Reproductive and Developmental Biology, London, UK). It was engineered from the vector pBLCAT2, by inserting 38 bp containing 13 bp vitellogenin ERE (5′-CAGGTCAcagTGACCTG-3′)4 at an XbaI site. The Herpes simplex virus thymidine kinase (tk) promoter (restriction sites BamHI/XhoI) and the downstream chloramphenicol acetyltransferase (CAT) reporter gene (sites XhoI/BmgI) were individually replaced with the cytomegalovirus immediate/early promoter (cmv) and SeAP reporter gene in a series of cloning experiments. The cmv promoter was obtained by polymerase chain reaction (PCR) from the expression vector pCMV5 (sequence: Accession no AF239429; annealing temperature 56°C, forward primer 5′-AATAGTAATCAATTACGGGG-3′; reverse primer 5′- GATCTGACGGTTCACTAA-3′), cloned into T-easy vector (Promega, Madison, WI) and sequenced. The SeAP gene, together with a SV40 poly A element (EcoRI/SalI fragment), was retrieved from the plasmid pSEAP2-Basic (Clontech, Mountain View, CA). The ERE sequence was left in or removed (XbaI digestion) from constructs to form a ±ERE series of plasmids (Figure 11).). Additional constructs used were: pCMV-β (Clontech, Mountain View, CA) with the cmv promoter and a downstream β-galactosidase reporter gene, and pCMV-hER, kindly provided by Prof. B. Katzenellenbogen (University of Illinois), with a cmv promoter and human estrogen receptor α gene.
The cmv immediate/early promoter contains four NFκB sites, numbered here in descending order from 4 to 1 as they approach the TATA box. The GGG motif in all of these was changed to TGC as described by Sun et al.5 and five variations of the mutated sites were created: single mutations in each of sites 4, 3, or 2, a double mutation in sites 2 and 3, and a triple change in sites 4, 3, and 2. Additionally, the consensus TATA box was mutated to CACACA. Changes in DNA sequence were introduced using QuikChange Multi Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) and confirmed by sequencing.
The BST cell line was established from the endometrium of a single uterine horn, from a day 16 post-oestrus cyclic cow, older than 30 mo.6 Cells were cultured in Dulbecco modified Eagle medium (DMEM) with 1% (v/v) antibiotic/antimycotic (Sigma) and 10% (v/v) foetal bovine serum (FBS, Gibco, Paisley, UK) at 37°C, 95% humidity, and 5% CO2. BST cells were electroporated at 80% confluence with 5 μg of plasmid DNA or (for electroporation controls, EC) without DNA at 300 V, 1650 μF. In co-transfection experiments, a total of 10 μg pDNA was used (5 μg of each construct). Co-transfection with pCMV-β was used to control for electroporation transfection efficiency where appropriate. Transfected BST cells were transferred into DMEM containing 1% (v/v) antibiotic/antimycotic and 10% (w/v) dextran-charcoalstripped bovine fetal serum, and 3.2 × 104 cells were cultured per well. Estradiol-17β was added at concentrations ranging from 0.01 to 100 nM, depending on the experimental design. ICI 182,780 (500 nM) was used as an estradiol receptor antagonist. Cells were cultured at 37°C, for 24 h (in the case of cmv promoter constructs) or 48 h (for tk promoter constructs), reflecting the difference between these promoters in level of transcription. The use of different times did not affect the interpretation of the data, as responses to estradiol were independent of time of cell culture. The post-culture medium was collected for SeAP assay. Remaining cells were lysed for CAT assay, β-galactosidase assay, or Bradford total protein assay. Estradiol had no effect on cell viability or proliferation when added either for 24 or 48 h.
Two inhibitors of NFκB protein were used. MG132 (Z-Leu-Leu-Leu-CHO) (Calbiochem) is a proteosome inhibitor, preventing IκB degradation, and SN50 prevents nuclear translocation of the activated NFκB complex. The SN50 peptide is composed of two distinct regions—the hydrophobic signal peptide of Kaposi fibroblast growth factor, providing cell-permeability, and the nuclear localization sequence (NLS) of the human transcription factor NFκB p50. The bovine NFκB NLS differs from that in the human and therefore a peptide based on the bovine NFκB p100/p49 (Accession no 1715333B), referred to here as bSN50, was synthesized (School of Chemistry, Nottingham). The amino acid sequence of bSN50 was AAVALLPAVLLALLAP (hydrophobic region)-VQRKRRKALP (bovine NF-κB p100/p49 nuclear translocation signal). After electroporation, BST cells were treated with 200 μg/mL bSN50 in DMEM containing 1% antibiotic/antimycotic and 10% dextran-charcoal-stripped serum with or without 10 nM estradiol and cultured for 24 h. Alternatively, BST cells were pre-incubated with 10 μM MG132 at 37°C, 1 h before electroporation. Transfected BST cells were treated with 10 μM MG132 with DMEM containing 1% antibiotic/antimycotic and 10% dextran-charcoal-stripped serum with or without 10 nM estradiol, and cultured for 24 h.
The SeAP assay was performed by chemiluminescence in culture medium using the Great EscAPe SeAP Kit (Clontech). Centrifuged medium was mixed with 1 vol of dilution buffer and incubated at 65°C, 30 min to inactive endogenous alkaline phosphatase, and then treated with l-homoarginine. After incubation, the mixture was equilibrated to room temperature and mixed with 2 vol of assay buffer. Finally, 1.25 mM CSPD chemiluminescent substrate was added to samples and incubated 10 min at room temperature. Light emission was then measured in a MicrolumatPlus LB 96V plate reader (Berthold, Bad Wild-bad, Germany).
The assay reagent was mixed with 100 μL 5 M MgCl2, 7 mL 0.1 M sodium phosphate buffer (pH 7.4), and 1 mL 4 mg/mL chlorophenol red-B-D-galactopyranoside (CPRG, Boehringer Mannheim) in 0.1 M sodium phosphate buffer (pH 7.4). Cleared cell lysate (10 μL) was mixed with 60 μL β-galactosidase assay reagent and incubated at 37°C. Absorbance was measured at λ = 595 nm in a Benchmark microplate reader (Bio-Rad).
Cleared cell lysate (30 μL) from each sample was incubated with 10 μL 5 mM chloramphenicol at 37°C for 5 min before 10 μL [3H] acetyl CoA, specific activity 15-74 GBq/ mmol (Amersham Biosciences) was added. After incubation at 37°C for 3 h, 48 μL of reaction mixture was transferred to a plastic vial with 1 mL 7 M urea and 10 mL toluene/PPO (0.8%) and shaken for 15 sec to ensure transfer of acetylchloramphenicol to the organic phase. [3H]Acetylchloramphenicol was then measured using a Tri-Carb 2100TR liquid scintillation analyzer (Packard Bioscience).
Protein concentrations were measured in cell lysates by the method of Bradford.7
Each transfection treatment was carried out in triplicate and all experiments were carried out between two and four times. Values are presented as means ± S.D. Statistical significance was analyzed using Student’s t-test by Excel 97 analysis ToolPak.
To analyze DNA accessibility, we used the Recon interface (http://www.mgs.bionet.nsc.ru/mgs/programs/recon/),8 a tool for eukariotic promoter analysis. This software package constructs a profile of nucleosome formation potential for promoter sequences, predicting promoter activity (e.g., tissue-specific promoters are generally less active and their profile has higher values than house keeping genes) and identifying potential transcription factor binding sites within promoter sequences.
pBL-tk-CAT(+)ERE was estrogen-responsive in BST cells ( p < 0.001; Figure 22).). Maximum response (1.8-fold) was obtained with 1 nM estradiol, and the response to 10 mM estradiol was significantly reduced by the antiestrogen ICI 182,780. The plasmid was also estrogen responsive when the CAT reporter gene was replaced by SeAP ( p < 0.001; Figure 33).). Maximal response was again achieved with 1 nM estradiol (Figure 3B3B);); the response was blocked by ICI 182,780 (Figure 3A3A);); and there was no estrogen response when the ERE was removed from the construct (Figure 3A3A).). Other experiments (Figures 55 and 66)) also demonstrated the estradiol responsiveness of pBL-tk-SeAP (+)ERE; the difference between experiments in the level of response to estradiol presumably reflected experimental variation.
Since previous experiments showed that reporter constructs driven by the cmv promoter are highly expressed in the endometrium, we replaced the thymidine kinase promoter in pBL-tk-SeAP(+)ERE with the cmv promoter. In contrast to the response obtained with pBL-tk-SeAP(+)ERE, pBL-cmv-SeAP(+)ERE was not estrogen responsive (Figure 44),), although the level of SeAP expression obtained exceeded 30-fold that with pBL-tk-SeAP(+)ERE. There was no effect of ICI 182,780 on cmvdriven expression. Measurement of SeAP in the culture media and in the cells remaining after removal of medium showed that SeAP was secreted by the cells, the ratio of SeAP/mg protein in medium exceeding that in the cells by a factor of 20. The lack of estrogen responsiveness by pBL-cmv-SeAP(+)ERE was not due to the SeAP gene, as it was also absent with CAT as a reporter gene in pBL-cmv-CAT(+)ERE (Figure 4B4B).
To determine whether the lack of estrogen response observed with pBL-cmv-SeAP(+)ERE reflected a low level of ER in BST cells, the construct was co-transfected with 5 μg pCMV-hER, to raise intracellular levels of ERα. Co-transfection had no effect on the estrogen responsiveness of pBL-cmv-SeAP(+)ERE (Figure 5A5A),), although it did increase SeAP expression by pBL-tk-SeAP(+)ERE twofold ( p < 0.001; Figure 5B5B).). Therefore, we concluded that lack of expression was not due to too low a level of ER in BST cells. Expression of pBL-cmv-SeAP(+)ERE was not stimulated by phenol red in the culture media, as shown by lack of effect of ICI 182,780 (Figure 4 A,BA,B).
Co-transfection of pBL-tk-SeAP(+)ERE with pCMV-hER suggested that the cmv promoter in pCMV-hER did not block ERE function in trans. To confirm this result with a cmv-driven plasmid that did not affect ERE function through overexpression of the ERα, we repeated the experiment with the β-galactosidase expression vector pCMV-β in place of pCMV-hER (Figure 6A6A).). In these experiments, co-transfection with pCMV-β had no effect on the estrogen responsiveness of pBL-tk-SeAP(+)ERE or the effect of ICI 182,780. Expression of pCMV-β was not affected by estradiol, and as a result the ratio of SeAP to β-galactosidase was increased by estradiol in co-transfected cells (Figure 6B6B).
Antagonism occurs between ER and NF-κB through a number of different mechanisms.9–11 Furthermore, three of the four NF-κB sites in the cmv promoter are responsible for the transcriptional activity of the promoter.5 Therefore, we investigated whether NF-κB binding to its response elements in the cmv promoter was involved in the effect on ERE function. To confirm that the cmv promoter in pBL-cmv-SeAP(+)ERE was activated by NF-κB, we treated transfected BST cells with the NF-κB blockers bSN50 and MG132. Both compounds reduced SeAP expression by between 25 and 50% ( p < 0.05 in both cases; Figure 7A and BB).). However, these effects were not specific for NF-κB, as these compounds also reduced SeAP expression in cells transfected with pBL-tk-SeAP(+)ERE, by approximately the same margin (Figure 7C and DD).). Neither compound prevented the estrogen responsiveness of pBL-tk-SeAP(+)ERE, induced estrogen responsiveness in cmv promoter constructs (Figure 77),), or had any effect on protein content of the cultures (data not shown).
To further clarify the role of NF-κB in cmv promoter function, we synthesized a series of mutations in the three functional NF-κB sites5 in the cmv promoter of pBL-cmv-SeAP(+)ERE. When transfected into BST cells, these mutant constructs showed that mutation in each of the three sites affected SeAP expression by about 40%, and that simultaneous mutation in all three sites was required to reduce expression by more than 80% (Figure 8A8A).). Removal of NF-κB sites did not restore estrogen responsiveness to pBL-cmv-SeAP(+)ERE.
As mutations in the NF-κB sites of the cmv promoter failed to impart estrogen responsiveness to pBL-cmv- SeAP(+)ERE, the possibility was considered that TATA box binding factors may be responsible for ERE inhibition. However, mutation in the TATA box did not recover ERE function, although it was effective in reducing SeAP expression (Figure 8B8B).
Although the steps towards production of an estrogen-responsive reporter construct were completed successfully, the final product did not meet the requirement of estrogen responsiveness. Thus the SeAP reporter gene product was confirmed to be secreted by cells in culture and was expressed in an estrogen-responsive manner when driven by the tk promoter, and the cmv promoter was incorporated into the construct to increase SeAP expression relative to the thymidine kinase promoter. However the principal finding was that the cmv promoter inhibited ERE function so that the resulting construct pBL-CMV-SeAP(+) ERE was not responsive to estrogen.
Inhibition of ERE response by the cmv promoter depended on the two sequences being in the same plasmid. A similar observation has been made by Ishikawa et al.,12 who showed that transfer of partial DNA sequences occurs between co-transfected plasmids, as suggested previously for the SV 40 origin of relication in COS-7 cells.13 The non-homologous end joining process proposed by Ishikawa et al.13 occurred most efficiently when non-circular DNA was transfected, whereas in the current work the plasmids used were circular. This difference may account for the lack of ERE inhibition in trans by co-transfected pCMVβ in the present study, if transcription from pBL-tk-SeAP (+)ERE commenced before the endonuclease reactions required to open the circular pCMVβ entering the cell. In the experiments of Ishikawa et al.,12 reporter plasmids with multiple copies of the ERE were inhibited by the cmv promoter, suggesting that the current results were unlikely to reflect use of a single ERE.
The cmv promoter was more active than the tk promoter when driving SeAP expression, by a factor of approximately 30, although when driving CAT expression this difference was much reduced (to about 9-fold without and 1.5-fold with estradiol; Figure 4B4B).). The absence of estrogen induction of SeAP with the cmv promoter did not appear to reflect expression at a rate that could not be further stimulated, because when the function of the cmv promoter was reduced by mutation of the NF-κB or TATA box sites, SeAP expression was not induced by estrogen. Mutation of the NF-κB sites in the cmv promoter reduced promoter function by up to 80%. This is consistent with the results obtained using the NF-κB inhibitors MG132 and bSN50. The lack of specificity of those compounds demonstrated here by an effect on the tk promoter, which lacks NF-κB sites, presumably reflects their ability to block the action or nuclear uptake/degradation of other transcription modulators, in addition to NF-κB.
One possible mechanism for ERE inhibition by the cmv promoter is the sequestration of an ER-dependent coactivator by an enhancer in the cmv sequence. One candidate compound is steroid receptor coactivator-1, which is bound by both the ER-ERE complex and NF-κB.14 In isolation, this mechanism appears unlikely at first sight, because inhibition occurred only when the two sequences were co-located in the same construct, and sequestration might be expected to occur if the sequences were on separate plasmids. If sequestration depended on the binding sequences being close together, on the other hand, possibly through bending of the DNA by bound transcription factors,15,16 such a mechanism may apply. It is unlikely to involve NF-κB or proteins bound at the TATA box, however, as mutations in these sites did not reverse ERE inhibition. Mutation of the CRE sites in the cmv promoter was not attempted in the present experiments, because it has been shown5 that these sites were not functional, but it remains possible that these sites are involved in sequestration of co-activators.
The positioning of the ERE relative to cmv promoter may be critical in determining its function. Reese and Katzenellenbogen17 showed that EREs positioned between the TATA box of the cmv promoter and the transcription start site act as transcription disruptors, due to ER binding to EREs forming a physical obstruction to movement of the transcription start complex. In the experiments of Ishikawa et al.,12 the distance between the ERE and the cmv promoter was larger than in the constructs tested here (more than 1.6 kb versus 600 bp, respectively). The effects observed by Reese and Katzenellenbogen17 and Ishikawa et al.12 differ in that the former authors showed inhibition of basal cmv function, whereas Ishikawa et al. Showed that the cmv promoter prevented estrogen-induced ERE function, as demonstrated in the present work. This suggests that the cis-interaction involved is sufficiently strong to operate over a considerable distance. Reese and Katzenellenbogen17 also reported lack of an effect on basal cmv function where multiple EREs were placed upstream from the transcription start site.
Analysis of DNA accessibility8 showed that the ERE resided at a location which was likely to be accessible to protein factors binding the duplex (Figure 99).). Introduction of the ERE sequence lowered nucleosome formation potential in the region 150–200bp (Figure 99),), where it already was low, indicating that this particular region of the promoter is highly accessible to transcription factors. On the basis of this analysis, the presence of the ERE would improve accessibility of DNA and the arrangement of cis elements. Therefore, a decrease in accessibility through insertion of the ERE consensus sequence is not likely to account for the absence of an ERE response.
We are grateful to E.L. Sheldrick for assistance during the early stages of this work and to P. Fisher for isolation of BST cells.