Abstract
Transient transfection experiments in which three different estrogen response element-containing reporter genes were cotransfected into HeLa cells, together with constitutively expressed estrogen receptor (ER) constructs, demonstrate that activation of the transcription of the reporter genes by epidermal growth factor (EGF) and by cholera toxin with 3-isobutyl-1-methyl-xanthine, which elevate cellular cAMP, is dependent upon the presence of functional ER. Cotransfection of the reporter genes with truncated versions of the ER shows that the two non-ligand activators of ER require different regions of the receptor to produce their effects on transcription. EGF acts primarily by means of transactivation domain AF-1, whereas cAMP acts via transactivation domain AF-2 of the ER. A point mutation that removes a major site of inducible phosphorylation within the AF-1 domain of the ER abolishes the response to EGF, but the response to estradiol and cAMP is retained. Specific inhibition of cAMP-activated protein kinase (protein kinase A) prevents the response to elevated cAMP but does not affect EGF or estradiol responses. Overexpression of the protein kinase A catalytic subunit in HeLa cells results in an amplified response to estradiol, similar to that induced by cholera toxin with 3-isobutyl-1-methyl-xanthine. Comparable experiments performed using COS-1 cells produce similar results but also reveal cell type- and promoter-specific aspects of the activation mechanisms. Apparently, the ER may be activated by three different signal molecules, estradiol, EGF, and cAMP, each using a mechanism that is distinguishable from that of the others.
INTRODUCTION
The estrogen receptor (ER) is a member of a superfamily of nuclear receptors that function as ligand-activated transcription factors (1–3). Binding the hormone activates the receptor protein so that it may bind to discrete estrogen response element (ERE) sequences in the genomic DNA and stimulate the transcription of specific structural genes. The ER, in common with other members of the nuclear receptor superfamily, possesses a modular structure in which various aspects of receptor function are associated with specific regions of the peptide sequence (4, 5). In addition to well defined DNA-binding and ligand-binding domains, nuclear receptors possess two separate regions that are required for optimal transcriptional activation. An amino-terminal transcription activation function (AF-1) operates in a manner that is independent of ligand binding (6, 7). A second activation function (AF-2) is located in the ligand-binding domain of the receptor toward the carboxy terminus of the molecule, and its activity is dependent upon the binding of an agonistic ligand (6, 7). Both AF-1 and AF-2 are required for optimal stimulation of transcription, but the relative contributions of the two varies in a promoter- and cell type-specific manner (6–8).
However, more recent evidence has suggested that the ER, in particular, may be activated in a ligand-independent manner by a variety of agents that include several growth factors (9–11), the neurotransmitter dopamine (12), and cAMP (13, 14). The expression of a variety of estrogen-responsive genes has been shown to be stimulated by these agents by a mechanism that is inhibited by the presence of the pure antiestrogen, ICI 164384 (9, 11, 13). Because ICI 164384 acts by binding to the ER (15), this is taken as evidence that transcriptional activation by these alternative agents involves the ER. Stimulation by epidermal growth factor (EGF) of the expression of estrogen-responsive reporter gene constructs in HeLa cells, which do not express endogenous ER, has been shown to be dependent upon coexpression of ER (9). In addition, it has been shown that the estrogen-like effects of EGF upon the mouse uterus do not occur in ER-deficient transgenic mice (10).
We now present evidence that the activation of the human ER by EGF and by cAMP involves mechanisms distinct from that of estradiol and distinguishable from each other. Furthermore, the transactivation effect of EGF is not dependent upon estradiol effects upon receptor activity.
RESULTS
HeLa cells from human cervical carcinoma do not contain ERs, but their growth is stimulated by EGF (16). We transfected eukaryotic expression vectors containing cDNAs derived from the human ER (4) into HeLa cells. HEG0 contains an insert coding for the full-length wild type human ER (amino acids 1–595) (17). HE0 codes for a full-length ER that contains a single point mutation resulting in an amino acid substitution, Gly400, which is replaced by a Val residue (17). HE457 codes for a full-length ER containing a point mutation resulting in the replacement of Ser118 by an Ala residue (18). HE15 contains an insert coding for a carboxyl-terminal truncated receptor (amino acids 1–282). HEG19 codes for an amino-terminal truncated receptor (amino acids 179–595). As indicated in Fig. 1a, the product of HE15 consists of the A/B and C domains and therefore retains AF-1, but not AF-2, transactivation function. The HEG19 product consists of domains C, D, and E/F and therefore has ligand-binding activity and retains AF-2 but not AF-1 transactivation function. We investigated the ability of these ER constructs to stimulate the expression of chloramphenicol acetyltransferase (CAT) reporter gene constructs, in response to estradiol, EGF, and elevated cAMP. The relative contributions of AF-1 and AF-2 to overall transcriptional activation are reported to be promoter specific (6–8), and we therefore used three reporter constructs, each with a different promoter structure (Fig. 1b).
DNA Constructs Transfected into HeLa and COS-1 Cells a, ER derivatives. Amino acid positions are numbered from the amino terminus of the ER and indicate the boundaries of the functional domains of the receptor and of the truncated receptors. Positions of point mutations (amino acid substitutions) in HE0 and HE457 are indicated. b, Reporter constructs (not drawn to scale). Solid squares indicate consensus EREs. ERE.VIT contains two contiguous (nonconsensus) endogenous EREs together with an additional consensus ERE inserted upstream of the endogenous EREs (31). VA, An NF1-related vitellogenin activator element (33). 2ERE.TATA is a synthetic enhancer/promotor sequence containing the elements indicated (31). ERE.TK consists of a consensus ERE linked to the promotor sequence (nucleotides −150 to +51) of the herpes simplex virus thymidine kinase gene (6).
In the absence of ER, the expression, in HeLa cells, of an estrogen-responsive reporter gene is not stimulated by estradiol, EGF, or cholera toxin + 3-isobutyl-1-methylxanthine (CT/IBMX) (Fig. 2). CT causes irreversible stimulation of adenylate cyclase, and IBMX inhibits cAMP phosphodiesterase; together they cause a major increase in intracellular cAMP (13). Cotransfection of the full length ER (HEG0) with the reporter gene enables all three agents to stimulate CAT expression (Fig. 2). EGF alone causes an increase in CAT expression that is less than that seen with estradiol; however, simultaneous treatment of the cells with EGF and estradiol results in a response that is the sum of the two separate responses. CT/IBMX alone causes only a weak stimulation of reporter gene activity but has a synergistic effect when combined with estradiol. The Gly->Val mutation in HE0 has been reported to prevent the ligand-independent activation of the ER by dopamine (12). However, we found that HE0 is able to support a pattern of response to estradiol, EGF, and CT/IBMX, alone and in combination, similar to that seen in the presence of the wild type receptor. The pattern of response to the inducers is similar for the different reporter gene constructs, although the scale of the increase varies.
ER Dependence of Reporter Gene Response to Estradiol, EGF, and Elevated cAMP in HeLa Cells HeLa cells were withdrawn from estradiol and transferred to serum-free medium before being transfected with reporter plasmid DNA with or without ER plasmid (HEG0 or HE0) DNA as indicated. All cells were cotransfected with pSV-β-galactosidase plasmid DNA. After 24 h, cells were transferred to serum-free medium with additions (E, EGF, CT/IBMX) or without addition (C) as indicated. Cells were harvested after 24 h of treatment and assayed for CAT and β-galactosidase activity. CAT activity was normalized relative to thr β-galactosidase activity and is expressed as a percentage of the activity in HEGO-transfected, estradiol-treated cells. Results are the means± sd of three experiments.
Cotransfection of the truncated receptors, HE15 and HEG19, with the reporter genes reveals differences between the three inducers and their relationships with each other. EGF is able to induce an increase in CAT expression in the presence of HE15 but not of HEG19 (Fig. 3). The scale of increase produced by EGF with HE15 (e.g. 2ERE.TATA, 5.6 ± 0.8-fold increase) is equivalent to that seen with the complete receptor (2ERE.TATA, 4.8± 0.7-fold increase). The product of HE15 does not possess a hormone-binding site and is therefore unresponsive to both estradiol and ICI 164384. HEG19 can respond to estradiol, but there is no further increase seen when EGF is combined with the hormone.
The Ability of Carboxyl-Terminal, but not Amino-Terminal, Truncated ER to Support EGF Stimulation of Reporter Gene Expression in HeLa Cells HeLa cells were transfected with the indicated reporter plasmid DNA plus expression vector DNA containing either HE15- or HEG19-truncated ER cDNA. Experiments were conducted as described for Fig. 2 (ICI, antiestrogen ICI 164384). CAT activity is expressed as a percentage of the activity measured in HEG0-transfected, estradiol-treated cells. Results are the mean ± sd of three separate experiments.
In contrast, CT/IBMX is able to stimulate reporter gene activity in the presence of HEG19 but not of HE15 (Fig. 4). The pattern of response to CT/IBMX, alone or in combination with estradiol or with ICI 164384, in the presence of HEG19 is similar to that seen with HEG0. The response to estradiol in the presence of HEG19 is reduced (e.g. ERE.VIT, 6.1-fold increase) compared with that seen with HEG0 (11.1-fold increase), but CT/IBMX in combination with estradiol still produces a synergistic response with ERE.VIT and 2ERE.TATA reporters, although the ERE.TK response is merely additive (data not presented). Cells transfected with HEG19 were rigorously depleted of estrogen before they were treated with CT/IBMX, making it unlikely that the response seen in the absence of exogenous estradiol involves an interaction between cAMP and ligand-occupied receptor.
The Ability of Amino-Terminal, but not Carboxyl-Terminal, Truncated ER to Support cAMP Stimulation of Reporter Gene Expression in HeLa Cells HeLa cells were transfected with the indicated reporter plasmid DNA plus expression vector DNA containing either HE15- or HEG19-truncated ER cDNA. Experiments were conducted as described for Fig. 2. CAT activity is expressed as a percentage of the activity measured in HEG0-transfected, estradiol-treated cells. Results are the means± sd of three separate experiments.
The simultaneous presence of HE15 and HEG19 enables the ERE.VIT and ERE.TATA reporter genes to respond to both estradiol and EGF (Fig. 5a). In addition, the response to EGF is now inhibited by ICI 164384, indicating that the two truncated receptors interact with each other. Phosphorylation of Ser118 has been shown to be necessary for full activation of AF-1 (18) and to occur via mitogen-activated protein kinase in EGF-stimulated cells (19, 20). The HE457 mutant ER cannot be phosphorylated at this residue and does not support stimulation of the ERE.VIT or 2ERE.TATA reporter genes by EGF, although their responsiveness to estradiol and to increased cAMP is retained (Fig. 5b).
Interaction between Amino-Terminal and Carboxyl-Terminal Truncated ERs in HeLa Cells and the Effect of a Point Mutation in AF-1 of the ER upon the Stimulation of Reporter Gene Expression a, HeLa cells were transfected with both HE15- and HEG19-truncated ER plasmid DNA plus the indicated reporter plasmid DNA. b, HeLa cells were transfected with HE457-mutated ER plasmid DNA plus the indicated reporter plasmid DNA. Experiments were conducted as described for Fig. 2. CAT activity is expressed as a percentage of the activity measured in estradiol-treated cells. Results are the mean ± sd of three separate experiments.
The differences in their responses to mutations in the ER indicate that the two ligand-independent activators of the ER, EGF and cAMP, achieve their effects by different mechanisms. We investigated further the involvement of protein kinases in these two mechanisms. The compound N-\[2-((p-bromocinnamyl)amino)ethyl\]-5-isoquinolinesulfonamide, 2 HCl (H-89) selectively inhibits cAMP-activated protein kinase[ protein kinase A (PKA)] (21) whereas bisindolylmaleimide I (BIMD) is a highly selective inhibitor of protein kinase C (22). The response to cAMP of the ERE.VIT reporter gene, cotransfected into HeLa cells with the HEG0 expression vector, is abolished by H-89 but is unaffected by the presence of BIMD (Fig. 6). Interestingly, the response of the reporter gene to estradiol or to EGF is unaffected by either of these protein kinase inhibitors. The involvement of PKA in the activation of the ER by cAMP is supported by the results shown in Fig. 7. HeLa cells were cotransfected with the ERE.VIT reporter gene, with the HEG0 ER expression vector and, as indicated, with an expression vector coding for either the α- or the β-isoform of the human PKA catalytic subunit (23). Overexpression of PKA alone had no detectable effect upon reporter gene expression but, in the presence of estradiol, there was a nearly 4-fold increase in the response above that seen with estradiol alone. Both PKAα and PKAβ produced a synergistic increase, equivalent to that seen upon elevation of cellular cAMP levels with CT/IBMX. The effects of PKA upon ERE.VIT expression were abolished by the presence of ICI 164384, indicating their requirement for ER.
The Ability of Inhibitors of Protein Kinases to Influence the Induction of Reporter Gene Expression in HeLa Cells HeLa cells were transfected with HEG0 ER plasmid DNA, ERE.VIT reporter plasmid DNA, and pSV-β-galactosidase plasmid DNA. Experiments were conducted as described for Fig. 2. Cells were treated for 24 h with the indicated agents, including H-89 (20 μm, PKA inhibitor) or BIMD 1 (10 μm, protein kinase C inhibitor). CAT activity is expressed as a percentage of the activity measured in estradiol-treated cells. Results are the mean ± sd of three separate experiments.
The Influence of Overexpression of PKA Catalytic Subunit upon the Induction of Reporter Gene Expression in HeLa Cells HeLa cells were transfected with HEG0 ER plasmid DNA, ERE.VIT reporter plasmid DNA, pSVβ-galactosidase plasmid DNA and, where indicated, with either pCαEV (PKAa) or pCβEV (PKAb) expression plasmid DNA. Experiments were conducted as described for Fig. 2. Cells were treated for 24 h with the indicated agents in medium containing 80μ m ZnSO4. CAT activity is expressed as a percentage of the activity measured in estradiol-treated cells. Results are the mean ± sd of three separate experiments.
Because responses of transfected estrogen-induced reporter genes have been reported to be cell-specific (6–8), we repeated some of our experiments using COS-1 cells from monkey kidney. The full-length ER (HEG0) supported a response to estradiol comparable to that seen in HeLa cells (Fig. 8). Neither reporter gene showed a significant response to EGF alone, although the CAT activity with EGF + estradiol (ERE.VIT, 121 ± 4%; ERE.TK, 125 ± 5%) was marginally higher than that with estradiol alone. Both reporter genes responded to CT/IBMX, but only ERE.VIT showed a synergistic response when CT/IBMX was combined with estradiol (CT/IBMX, 2-fold increase; estradiol, 14-fold increase; estradiol + CT/IBMX 36-fold increase).
The Ability of Estradiol, EGF, and cAMP to Stimulate Reporter Gene Expression in the Presence of ER in COS-1 Cells COS-1 cells were withdrawn from estradiol and transferred to serum-free medium before being transfected with reporter plasmid DNA and ER plasmid (HEG0) DNA as indicated. Experiments were conducted as described for Fig. 2. CAT activity is expressed as a percentage of the activity measured in estradiol-treated cells. Results are the mean± sd of three separate experiments.
In the presence of HE15, ERE.VIT showed a pattern of responses similar to that seen in HeLa cells (Fig. 9), i.e. induction by EGF but not by estradiol or CT/IBMX. However, ERE.TK responded to EGF and failed to respond to estradiol, but also showed a clear response to CT/IBMX. ERE.VIT and ERE.TK, in the presence of HEG19, reproduced the pattern of responses to the three inducers that was seen in HeLa cells, i.e. induction by estradiol and by CT/IBMX, but only ERE.TK showed a synergistic response to their simultaneous presence.
The Ability of Truncated ERs to Support the Stimulation of Reporter Gene Expression by Estradiol, EGF, and cAMP in COS-1 Cells COS-1 cells were withdrawn from estradiol and transferred to serum-free medium before being transfected with reporter plasmid DNA and either HE15- or HEG19-truncated ER plasmid DNA, as indicated. Experiments were conducted as described for Fig. 2. CAT activity is expressed as a percentage of the activity measured in HEG0-transfected, estradiol-treated cells. Results are the mean ± sd of three separate experiments.
DISCUSSION
Our results demonstrate that the expression of estrogen-responsive reporter genes may also be stimulated by a polypeptide growth factor, EGF, and by elevated cAMP, in the absence of exogenous estradiol and in a manner that is dependent upon the presence of the ER. The simultaneous presence of an optimally inducing concentration of estradiol (24) and either EGF or CT/IBMX results in a level of reporter gene expression that is higher than that seen with estradiol alone. This suggests that the mechanism of ER activation by EGF or cAMP differs from that used by estradiol. Furthermore, the fact that EGF + estradiol causes an additive increase whereas CT/IBMX + estradiol causes a synergistic increase suggests that the two non-ligand activators differ from each other in their mechanisms of action.
This latter conclusion is supported by our experiments using truncated versions of the ER, which demonstrate that EGF and cAMP require different regions of the receptor to achieve their effects upon transcription. The fact that EGF can cause stimulation of transcription in the absence of the carboxyl-terminal AF-2 transactivation domain suggests that its effects are mediated via AF-1. EGF can stimulate reporter gene transcription via HE15, a truncated version of the ER that completely lacks the ligand-binding domain of the receptor, clearly establishing that EGF does not simply enhance the action of estradiol, but is able to activate the ER de novo. On the other hand, elevated cellular cAMP levels, produced by treating cells with CT/IBMX, appear to act by increasing the transactivation activity of AF-2, located in the carboxyl-terminal half (E domain) of the receptor. In this connection it is interesting to note that the only difference we observed between the activity of the wild type receptor (HEG0) and the receptor with a point mutation in the E domain (HE0) is that the latter showed a diminished response to CT/IBMX (Fig. 2). Because the HEG19 construct retains the ligand-binding domain of the receptor, we cannot determine whether or not the cAMP effect requires the occupation of the LBD by an agonistic ligand. The fact that stimulation of transcription by CT/IBMX can be seen in cells that have been rigorously depleted of estrogen makes this unlikely.
The role played by phosphorylation of the ER in the ability of the receptor to stimulate transcription is unclear. The ER protein is phosphorylated at a basal level in the absence of estradiol, and numerous reports (e.g. Refs. 18, 25, and 26) have shown that the level of phosphorylation is increased upon hormone binding. The identity of the amino acids phosphorylated, their relative levels of phosphorylation, and the functional consequences of their phosphorylation are, however, the subjects of conflicting reports (18, 25–27). Nevertheless, the fact that ER phosphorylation is also increased, at apparently identical sites, by antiestrogen binding (18, 25) indicates that ligand-induced transactivation activity of the receptor is not determined solely by its level of phosphorylation. The evidence we present here supports the idea that ligand-independent activation of the ER requires phosphorylation of the receptor protein at sites that differ with the identity of the activator. Ser118, which is situated within the AF-1 domain (28), has been identified as a major site of ligand-stimulated phosphorylation (18, 26) and has also been shown to be phosphorylated by mitogen-activated protein kinase in response to EGF stimulation of cells (19, 20). Conversion of this residue to a nonphosphorylatable alanine (HE457) abolishes activation of the ER by EGF (Fig. 5b). However, HE457 is still activated, albeit to a reduced extent, by estradiol and by elevated cAMP. Elevation of cellular cAMP has been reported to stimulate ER phosphorylation at a site(s) outside the A/B domain and distinct from those responsive to ligand binding (25). We show that inhibition of cAMP-stimulated protein kinase (PKA), by treating cells with the specific inhibitor H-89, prevents stimulation of reporter gene activity by CT/IBMX and, most strikingly, prevents the synergism between estradiol and CT/IBMX (Fig. 6). The involvement of PKA in the activation of ER by elevated cAMP is also supported by our observation that increasing kinase activity by overexpression of the PKA catalytic subunit duplicates the synergistic response to estradiol seen with CT/IBMX treatment (Fig. 7). This stimulation of reporter gene expression by elevated PKA level was prevented by the antiestrogen, ICI 164384, indicating the requirement for ER.
ICI 164384 is termed a “pure” antiestrogen (15) because, unlike non-steroid antiestrogens such as tamoxifen, it does not exhibit any agonist activity but inhibits both the AF-1 and the AF-2 transactivation functions of the receptor (29). The mechanism by which ICI 164384 occupation of the ligand-binding domain, in the carboxy-terminal half of the receptor, may interfere with the activity of the amino-terminal AF-1 function is unclear. Our experiments (Fig. 5a), in which both truncated versions of the ER (HE15+HEG19) were simultaneously expressed in HeLa cells, show that the AF-1 domain and the ligand-binding domain do not need to be situated in the same molecule for inhibition to occur. ICI 164384 was able to inhibit the action of EGF, previously shown to operate by stimulating AF-1 but not AF-2, indicating that the two truncated receptors interacted with each other rather than acting independently. The mechanism of this intermolecular interaction is unclear. Both HE15 and HEG19 possess DNA-binding domains so that they may be capable of binding to the ERE as a heterodimer, and the association of HEG19, in an inactive configuration, with HE15 may inactivate HE15 also.
Both AF-1 and AF-2 contribute to the overall transactivation activity of the ligand-occupied ER, but the nature of their contributions differs. Truncated receptors have shown that AF-1 can exhibit transactivation in the absence of estrogen binding to the ER whereas AF-2 activity is dependent upon hormone binding (6, 7). We have now demonstrated that AF-1 activity, although ligand-independent, is not constitutive but may be indirectly regulated by signal molecules that bind to plasma membrane receptors. The relative contributions of AF-1 and AF-2 to gene activation are markedly influenced both by the structure of the target gene promoter and by the cell type containing the target gene (6–8). Differences in their extent of response were observed between the reporter genes, although the pattern of response to the inducers was similar for all three reporter gene constructs. However, it is difficult to detect any consistent pattern in the relative responsiveness of the reporter genes. In general, the ERE.VIT reporter gene, which has the largest number of copies of the ERE and the most complex promoter, gave the largest response, but there were exceptions to this, e.g. with HEG19 in COS-1 cells (Fig. 9). We assume that these differences between promoters arise from differences in requirements for possible intermediary factors and differences in the availability of these molecules in different cell types (30). We compared the responses of the ERE.VIT and ERE.TK reporter genes in COS-1 cells with those in HeLa cells. The only significant difference between the results with the two cell lines was that ERE.TK was able to respond to CT/IBMX in the presence of HE15, in COS-1 cells. Apparently, in the presence of certain promoters and in a cooperative cellular context, cAMP may also work through AF-1. A potential PKA phosphorylation site does exist within the HE15 sequence at Ser236 (25).
Our results show that the ER can be stimulated to activate transcription by EGF and cAMP as well as by binding estradiol. Transcriptional activation by estradiol is believed to involve both the AF-1 and AF-2 transactivation functions. Our experiments indicate that, at least in certain promoter/cell type contexts, EGF and cAMP effects each involve only one transactivation function: AF-1 for EGF and AF-2 for cAMP. The fact that addition of either EGF or CT/IBMX in the presence of a saturating concentration of estradiol results in a further increase in reporter gene expression indicates that there are basic differences between the mechanisms of ligand-independent and ligand-dependent activation. We have presented evidence that specific protein kinases are involved in the mechanisms triggered by the two non-ligand activators of the ER. It is increasingly recognized that the many intracellular signal transduction pathways in eukaryotic cells do not operate independently of each other but that cross-talk between pathways is possible and potentially important. It appears that the ER provides a site for the integration of three well recognized signal transduction pathways, i.e. those of steroid hormones, cAMP, and tyrosine kinase receptors.
MATERIALS AND METHODS
Chemicals and Materials
Tissue culture medium, newborn calf serum (NBCS), FCS, antibiotics, and trypsin-EDTA were purchased from GIBCO BRL (Paisley, Scotland). 17β-Estradiol, insulin, CT, IBMX, EGF (human, recombinant), proteinase K, and vanadyl complex were obtained from Sigma (Poole, England). ICI 164384 was kindly provided by Dr. A. E. Wakeling, Zeneca Pharmaceuticals (Macclesfield, England). H-89 and BIMD were obtained from Calbiochem-Novabiochem (U.K.) Ltd (Nottingham, England). Other reagents were obtained from Boehringer Mannheim UK (Lewes, England).
The ER plasmids, reporter constructs, and PKA plasmids were obtained from the laboratories in which they were constructed. ER derivatives, HE0 (4), HEG0, HEG19 (17), and HE457 (18) are contained within the eukaryotic expression vector pSG5, and HE15 is contained in pKCR2 (4). pCαEV and pCβEV consist of cDNAs coding for the α- andβ -isoforms of the PKA catalytic subunit, contained within the Zem3 vector in which transcription is driven by the mouse metallothionein promoter (23). ERE.VIT is derived from the Xenopus laevis vitellogenin B1 gene (31). 2ERE.TATA is an entirely synthetic enhancer/promoter sequence (31). ERE.TK (pERE BLCAT) consists of a consensus ERE linked to the promoter sequence (nucleotides −105 to+ 51) of the herpes simplex virus, thymidine kinase gene (6).
Cell Culture and Transient Transfections
HeLa cells and COS-1 cells were obtained from the European Collection of Animal Cell Culture (Porton Down, U.K). Both cell lines were cultured in MEM medium (GIBCO BRL) containing 100 μg/ml penicillin (GIBCO), 100 μg/ml streptomycin (GIBCO), and 10% (vol/vol) FCS (GIBCO). Cells were depleted of estrogen by culture for 6 days in phenol red-free RPMI 1640 medium (GIBCO), supplemented with 5% (vol/vol) dextran/charcoal-treated NBCS (32). Cells were further cultured for 1 day in medium supplemented with 1% (vol/vol) dextran/charcoal-treated NBCS followed by 1 day in serum-free medium[ phenol red-free RPMI 1640 supplemented with: 10 μg/ml transferin; 10 ng/ml sodium selenite; 1% (wt/vol) glutamine; 0.2% (wt/vol) BSA]. Cells were then harvested and seeded, in multiwell plates, at a density of 2 × 105 cells per 35-mm well, in serum-free medium. After 24 h, cells were transfected with DNA \[300 ng ER plasmid (HEG0, HE0, HE15, HEG19 or HE457), 1000 ng each of 2ERE.TATA, ERE.VIT, or ERE.TK plasmid, 20 ng pCαEV or pCβEV plasmid, 700 ng pSV-β-galactosidase plasmid (Promega, Madison, WI)\] using the LipofectAMINE reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer’s protocol. After 5 h an equal volume of serum-free medium was added to each well. After a further 19 h the medium was changed for fresh serum-free medium containing the inducing agents: 10−8m estradiol (E); 100 ng/ml EGF; 1μ g/ml CT, 10−4m IBMX; 10−6m ICI 164384 (ICI); or without addition (C). For transfections involving the pCαEV and pCβEV vectors, the culture medium was supplemented with 80 μm ZnSO4. Cells were incubated for 24 h and then were harvested and CAT (liquid scintillation method) and β-galactosidase activities were assayed in whole cell extracts using assay kits (Promega) according to the manufacturer’s protocols. CAT activity results were normalized relative to the β-galactosidase activity, to correct for differences in efficiency of transfection, and are expressed as a percentage of the estradiol-induced level, as indicated. Results are the mean ± sd of at least three separate experiments.
Acknowledgements
This work was supported by the Association for International Cancer Research.
We wish to thank P. Chambon for the gift of HEG0, HE0, HE15, HEG19, and HE457; G. S. McKnight for the gift of pCαEV and pCβEV; D. J. Shapiro for the gift of ERE.VIT and 2ERE.TATA; M. G. Parker for the gift of ERE.TK; and A. E. Wakeling for the gift of ICI 164384.
Le
Le
