| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
and βDepartment of Biosciences and Nutrition at Novum (A.E.D., G.S., J.-Å.G.), Karolinska Institutet, 14157 Huddinge, Stockholm, Sweden; and Foundation for Biomedical Research (G.S.), Academy of Athens, 11527 Athens, Greece
Address all correspondence and requests for reprints to: Giannis Spyrou, Foundation for Biomedical Research, Academy of Athens, 11527 Athens, Greece. E-mail: giannisspyrou{at}bioacademy.gr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(ER) and ERβ mobility. Our results indicated that these two receptors share similarities in response to some ligands but differ significantly in response to others. Our results suggest that for ER, ligands can be classified into three distinct groups: 1) ligands that do not affect the mobility of the receptor, 2) ligands that cause a moderate effect, and 3) ligands that strongly impact mobility of ER. Interestingly, we found that for ERβ such a classification was not possible because ERβ ligands caused a wider spectrum of responses. One of the main differences between the two receptors was the response toward the antiestrogens ICI and raloxifene, which was not attributable to differential subnuclear localization or different conformations of helix 12 in the C-terminal domain. We showed that both of these ligands caused a robust phenotype, leading to an almost total immobilization of ER, whereas ERβ retained its mobility; we provide evidence that the mobility of the two receptors depends upon the function of the proteasome machinery. This novel finding that ERβ retains its mobility in the presence of antiestrogens could be important for its ability to regulate genes that do not contain classic estrogen response element sites and do not require DNA binding and could be used in the investigation of ligands that show ER subtype specificity. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
So far, two ERs have been identified, designated ER and ERβ, which are products of different genes located on different chromosomes (2, 3). In mammals, both ER and ERβ have conserved DNA-binding domains (96%), but they differ in their ligand-binding domain, showing only 58% homology (4). ER has two distinct transcriptional activation functions (AFs): AF-1 and AF-2. AF-1, located at the N terminus, is ligand independent, constitutively active and the region that contributes to the transcriptional activity of the receptor by recruiting coactivator proteins such as GRIP1 and steroid receptor coactivator-1 and the histone acetyl-transferases p300/cAMP response element binding protein-binding protein and pCAF (5, 6). It is known that p300 enhances transcriptional activity through several different mechanisms, including bridging and scaffolding functions to join together other transcription factors with the basal transcriptional apparatus as well as histone acetylation activity resulting in acetylation of specific lysines contained within histone proteins. In ERβ, the activity of AF-1 is negligible (7). The AF-2 domain in both ER and ERβ is under the control of ligands. Variations in the phenotypes of knockout mice lacking ER or ERβ suggest that these two proteins have different biological activities (8, 9, 10). This was further supported by in vitro and in vivo studies on uteri in ERβ knockout mice, indicating that ERβ is a modulator of ER activity because it is able to reverse the effects of ER and inhibit estradiol (E2)-dependent proliferation (11, 12, 13, 14).
Recently, the use of green fluorescent protein (GFP) and advanced imaging techniques in live cells has revealed a very dynamic process in which nuclear receptors rather than statically binding to the chromatin are very mobile, and a hit-and-run model has been proposed as the mode of action for nuclear receptor-dependent transcription regulation (15, 16, 17). By the use of the fluorescence recovery after photobleaching (FRAP), ER has been shown to be a highly mobile protein within the nucleus and the addition of ligands affected the mobility of the receptor as well as its subnuclear distribution (18). The presence of E2 caused a redistribution of its subnuclear localization and a slow down in its mobility compared with unliganded receptor. However, the presence of the antiestrogen ICI caused an almost complete immobilization of ER, an effect that the presence of E2 could not prevent. Thus far, no studies have been conducted on the subnuclear mobility of ERβ, and with the distinct biological differences between the ERs, we set out to perform a comparative study between ER and ERβ and analyze the effect of a wide array of ligands on their subnuclear localizations. Our studies provide new insights of the effect of ligands on both ER and ERβ, in particular regarding the differential response by the two receptors upon ligand treatment.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(595 amino acids) and ERβ (530 amino acids) were cloned into the pEGFP-C1 (Clontech, Mountain View, CA) vectors, employing conventional molecular biology tools and methods, resulting in fusion proteins of ER/β with GFP located at the N terminus of the ERs. The mutants where the helix 12 domains of the two receptors were exchanged contain amino acids 1–534 of ER and 486–530 of ERβ for the ERH12β constructs, whereas amino acids 1–486 of ERβ and 535–595 of ER were used for the ERβH12 construct.
Cell culture and transfections
Cells were maintained in DMEM complemented with 10% fetal bovine serum and gentamicin and kept in a 37 C humidified incubator with 5% CO2. For the microscopy experiments, cells were grown on coverslips in DMEM without phenol red, containing stripped serum, for 1 or 2 d and transfected with the constructs for GFP-ER/β overnight using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) in medium with stripped serum but lacking antibiotics. The next day, cells were washed and supplemented with fresh media and treated overnight with the following ligands: 10 nM E2, 10 nM genistein, 1 µM tamoxifen, 1 µM raloxifene, and 20 nM ICI. For the subnuclear localization studies, cells were fixed in 4% paraformaldehyde for 30 min at room temperature and then washed three times in PBS, counterstained with 4',6-diamidino-2-phenylindole for 5 min, washed again three times in PBS, and mounted on glass slides using Fluorosave mounting medium. For the luciferase reporter assay, HEK293 cells were transfected using jetPEI (Polyplus-transfection) with a 3xERE-TATA-luc construct (19) together with one of the following constructs: pSG5-ER, pSG5-ERβ (20), pEGFPC1-ER, or pEGFPC1-ERβ. Cells were treated with E2 overnight, and the luciferase activity was measured using the luciferase assay kit (Biothema, Handen, Sweden).
Microscopy
Cells were imaged on an ASMDW multidimensional workstation microscope (Leica, Wetzlar, Germany). For each cell, a z-stack of images was acquired and deconvoluted with the Autoquant image restoration program. For the FRAP assays, a Leica TCS SP2 confocal microscope was used. Cells were imaged through a x40 dipping lens using the 488-nm laser line of an ArKr laser, and light was collected between 500 and 550 nm. Cells were point bleached for 2 sec using full laser power, and subsequently, images were collected with 0.5-sec intervals for 20 sec. The fluorescence of the bleaching area was measured before and after the bleaching, and the images obtained after bleaching were corrected for the general bleaching during image acquisition by measuring the fluorescence of nonbleached areas. Values were also standardized by setting the prebleach image as 100%. To avoid artifacts that could arise due to high overexpression of the protein, only cells with low amounts of GFP were used for the scanning, which has previously been shown to give specific results, because cells with highly overexpressed protein tend to show slow down in mobility even in the absence of ligands (18).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and -β is dependent on ligands changes in response to the presence or absence of E2, tamoxifen, and ICI (18). In this study with MCF-7 cells, we used a panel of ligands and we compared their effects on the subnuclear localization of ER and ERβ. As can be seen from deconvoluted fluorescence images, our results are in line with previous reports. In the presence of E2, tamoxifen, and ICI, ER shows a speckle-like pattern compared with a homogeneous distribution in untreated cells (Fig. 1A
). Raloxifene also induces a dotted pattern, whereas genistein only moderately, if at all, affects the subnuclear ER distribution. Interestingly, ERβ is not as homogeneously distributed in untreated cells as ER, but still addition of ligand changes the subnuclear localization (Fig. 1A). One striking difference between ER and ERβ is that genistein, although it does not cause any profound effect on localization of ER, clearly causes a change in the localization pattern of ERβ, similar to the changes imposed by E2. On the other hand, tamoxifen, raloxifene, and ICI do not seem to cause as profound changes in the subnuclear localization pattern of ERβ as they do for ER. However, in the presence of these ligands, more of the ERβ is localized around the nucleoli compared with untreated cells; in contrast, ER appears to be more abundant around the nucleoli in untreated cells than in the presence of tamoxifen, raloxifene, or ICI.
|
. Within minutes of addition of E2, ERβ is relocated to a dotted pattern.
Furthermore, to ensure the functional integrity of the ER- and ERβ-GFP chimeras, we performed transactivation experiments using a 3xERE-TATA-Luc (19) reporter plasmid cotransfected with either ER- or ERβ-GFP in the absence or presence of E2. Both receptors were able to induce the expression of luciferase in the presence of E2 compared with untreated cells, suggesting that they are functional in transcriptional assays similarly to the untagged ER and ERβ (Fig. 1C).
Differential mobility between ER and -β upon ligand treatment
We further wanted to investigate the differences between ER and ERβ in their mobility after ligand treatment. As shown in Fig. 2A, ER exhibits differential mobility depending upon which ligand is present. Genistein seems to have a minor effect on ER mobility; E2 and tamoxifen elicit a considerable change, whereas raloxifene and ICI cause almost total ER immobilization. On the other hand, in the case of ERβ, these ligands cannot be divided into similar categories. Furthermore, a comparative analysis between ER and ERβ for each ligand shows that ER and -β have similar mobilities in the absence of ligand and in the presence of E2 (Fig. 2B). Genistein seems to have a more pronounced effect on ERβ than on ER, whereas the opposite results are observed for tamoxifen where ER has a slightly slower recovery rate than ERβ. The most drastic differences between the two receptors are observed in the presence of raloxifene and ICI. Although both these ligands cause a significant decrease in ERβ recovery rate, the effects are not as pronounced as observed for ER where the protein is almost immobile, especially in the presence of ICI.
|
where the nuclear fluorescence levels are plotted, there is a decay of the nuclear fluorescence; however, similar decay can be detected in cells that have been imaged under the same conditions but have not been point bleached in the cytoplasm, suggesting that the bleaching observed is due to the imaging procedure.
We also investigated whether the cellular context may affect the ligand-dependent mobility of the receptors. Experiments carried out as described above with the non-estrogen-dependent cell line HEK293 indicate that whereas genistein has no effect, E2 and tamoxifen cause a considerable decrease in recovery of ER, and raloxifene and ICI significantly slow down ER mobility (Fig. 3A). These results follow the same pattern as described before in MCF7 cells, and the ligands can be categorized into three groups. The effect of ligands on ERβ in HEK293 cells cannot be as distinctly categorized as for ER. Furthermore, comparison of how the two receptors are affected by each ligand (Fig. 3B) shows a high similarity with the results obtained in MCF7 cells. Once again, the most distinct difference between the ER isoforms is seen after treatment with raloxifene and ICI, where ERβ is capable of retaining its mobility compared with ER. Furthermore, as shown in Fig. 3C, in cells transfected with ER, the presence of raloxifene and ICI results in almost no recovery, whereas the opposite is observed for ERβ.
|
-GFP-transfected cells treated with ICI have been followed for 10 min with an image update of 10 sec, there is almost no recovery for ER even after 10 min, whereas ERβ (supplemental movie 2) has almost fully recovered in the same time span. Similar results were also obtained under discontinued exposure to ICI. When HEK293 cells, transfected with ER- or ERβ-GFP, were treated for only 2 h with ICI and then washed and incubated in media lacking ligands for 6 h before the FRAP assay, the receptors maintained their previously described properties (supplemental Fig. 1).
The difference in mobility is independent of H12 and proteasome function
To further investigate the difference in mobility after raloxifene and ICI treatment, we cotransfected cells with plasmids encoding ERβ-GFP or ER-HcRed fusion proteins. As shown in Fig. 4A, the two proteins colocalize, suggesting that the difference in the mobility is not due to different subcellular localization.
|
caused by ICI, because a deletion mutant lacking the C terminus did not show any immobilization under ICI treatment (18). This report encouraged us to explore whether the C terminus of the two receptors is responsible for the differences observed. To test this hypothesis, two mutants were constructed, ERH12β and ERβH12, where we exchanged the C-terminal regions of ER and ERβ including the H12. These constructs were transfected into HEK293 cells, and FRAP experiments were performed in the presence of ICI. As shown (Fig. 4B), although these ER-GFP constructs have their C-terminal parts exchanged, there were no differences in their mobilities; rather, ER, which contained the H12 of ERβ (ERH12β), remained immobile, whereas ERβH12 was mobile, suggesting that H12 is not involved in the differential mobility of the ER isoforms.
To investigate whether ERβ mobility is affected by proteasomal activity, as is the case for ER (18), cells were transfected with either ER-GFP or ERβ-GFP plasmids and subsequently treated with the proteasome inhibitor MG132 for 10 h and subjected to FRAP analysis. As seen in Fig. 5, no differences could be detected between ER and ERβ mobility after proteasomal inactivation, suggesting that ERβ mobility is proteasome dependent.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B (30) and Pit-Oct-Unc-class activator Pit-1 (31) display similar properties.
Previous studies on ER suggested that, in line with other receptors, it is a highly mobile protein and that its mobility is regulated by ligands and by the proteasome machinery. So far, no data on ERβ have been available, and although the ERs have many common characteristics, they display functional differences as previously discussed. We observed that addition of both agonists and antagonists resulted in a subnuclear localization pattern of ERβ different from that in untreated cells. Furthermore, using live-cell imaging, we could show that addition of E2 leads to a highly dynamic reorganization of ERβ, occurring within minutes, in agreement with the behavior of ER. In addition, we show that even after the addition of the GFP tag, both ER and ERβ are able to respond to E2 treatment transactivating a 3xERE-TATA-driven luciferase reporter. This would suggest that the receptors are capable of retaining their functional activity and that the results obtained should reflect the properties of the untagged receptors.
In the presence of genistein, ERβ showed a subnuclear pattern similar to the one observed after addition of E2, whereas ER was as homogeneously distributed in the nucleus as in the control, suggesting that genistein functions as a ligand for ERβ but not for ER. In agreement with this finding, it has been reported that genistein favors binding to ERβ and is capable of inducing transcription of ERβ-regulated genes (32, 33).
To further investigate the differences in the nuclear mobility of ER and ERβ during ICI and raloxifene treatment, we generated mutants of the receptors where the H12s of the proteins were switched. The rationale for this experiment was that it has been reported that the C-terminal domain of ER, containing H12, is crucial for the inhibition of ER mobility by ICI (18), suggesting that differences between the H12s of ER and ERβ might account for the diverse responses of the receptors toward ICI and raloxifene. However, exchanging the H12s of ER and ERβ did not alter the inhibitory activities of ICI and raloxifene on ER mobility. Interestingly, a chimeric protein between the GR DNA-binding domain and the ER ligand-binding domain fused to the C terminus of the GFP protein was not as sensitive to ICI-mediated immobilization (34). In addition, it has been reported that even though ER and ERβ can bind to the same DNA sequences, in the presence of ICI, the ERβ interaction with the DNA is significantly more thermostable than for ER (4). Taken together, these results would implicate the DNA-binding domain and the DNA-binding properties of ER in the ICI-dependent immobilization and raise additional questions about the proposed model for ER immobilization in the presence of antiestrogens.
It has been reported that for GR, ligand affinity is a crucial determinant for the mobility of the receptor, where the ligands with higher affinity would cause higher degrees of immobility of the receptor (22). Having that in mind and taking into account that raloxifene has 20 times higher affinity for ER than ERβ (35), we performed a FRAP analysis at various concentrations (0.1, 0.5, and 1 µM) of raloxifene in HEK293 cells. However, no differences could be seen using various concentrations, and similar findings were obtained in HeLa cells (data not shown). These results suggest that the ligand concentration does not play a crucial role for the differences in mobility between the two receptors.
What has become increasingly clear is that the function of the proteasome is very important for the dynamic nature of the mobility of nuclear receptors such as GR and ER (18, 22, 36, 37). ER is able to cycle on ER-responsive promoters both in a liganded and unliganded state; however, the kinetics differ depending on the state of ER. Furthermore, clearance of ER by proteasomes is necessary for the cycling to occur, and it has been suggested that ICI might interrupt the cycling of ER and target the receptor to the proteasome (38). Thus, a possible explanation for the differences in mobility between the ER isoforms during ICI treatment could be the independence of ERβ from the proteasome machinery. To explore this possibility, we used the proteasome inhibitor MG132 and showed that similarly to ER, ERβ is dependent upon the function of the proteasome to retain its mobility because inhibition of the proteasome leads to the complete immobilization of ERβ. Hence, ERβ mobility is indeed dependent on the proteasome machinery, and most probably, the cycling mechanism suggested for ER is also applicable for ERβ. One possibility is that ICI does not disrupt the cycling of ERβ by targeting it to the proteasome, suggesting that ERβ is transcriptionally active even in the presence of ICI and raloxifene. Indeed, it has been shown that, at activator protein (AP)-1 sites, ERβ behaves in an opposite way compared with at classical estrogen response element (ERE)-containing promoters.
Although antiestrogens abolish the transactivation activity of ERβ at ERE sites, at AP-1 sites, antagonists result in higher activation of reporter constructs, and whereas E2 activates ERβ at ERE sites, it inhibits ERβ transactivation at AP-1 sites (20). Furthermore, it has been shown that ERβ is a potent transactivator of the human retinoic acid receptor promoter a1 in the presence of the ER antagonists tamoxifen, raloxifene, and ICI, whereas E2 abolishes this activity. In contrast, ER acted as an activator only in the presence of agonists and not in the presence of antagonists. Promoter analysis, deletion mutants, and site-specific mutations led to the identification of Sp1 elements as responsible for the ERβ activity on the promoter (35). Taken together, this might suggest that the ability of ERβ to retain its mobility in the presence of antiestrogens could be a crucial factor for its capacity to regulate genes that do not contain classic ERE sites and do not require DNA binding such as AP-1 and Sp1 sites.
Finally, the present novel finding that antagonists appear to differentially modify the mobility of the two ER isoforms could be used to search for ligands with ER subtype specificity. It has previously been proposed that the GR-ER chimera could be used in standardized screenings to isolate molecules that can act as ligands for the ERs (34). The ability of the GR-ER chimera to translocate from the cytoplasm to the nucleus in the presence of ER ligands provides a possible mechanism by which candidate molecules could be assayed for their ability to act as ligands for ER. However, this chimera does not correctly mimic the intranuclear mobility properties of the native ER and therefore could not be used to distinguish between potential estrogens and antiestrogens, respectively. Hence, our findings could be used to identify ligands that could potentially function as ER agonists or antagonists.
|
Footnotes |
|---|
This work was supported by the Swedish Cancer Fund, the Swedish Research Council Project K99-13X-10370, and Karolinska Institutet.
Disclosure Statement: A.E.D. and G.S. have nothing to disclose. J.-Å.G. is shareholder, research grant receiver, and consultant of KaroBio AB.
First Published Online September 20, 2007
Abbreviations: AF, Activator function; AP, activator protein; E2, estradiol; ER, estrogen receptor; ERE, estrogen response element; FRAP, fluorescence recovery after photobleaching; GFP, green fluorescent protein; GR, glucocorticoid receptor.
Received February 14, 2007.
Accepted for publication September 13, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. J Biol Chem 272:25832–25838 and ERβ. J Steroid Biochem Mol Biol 69:165–175[CrossRef][Medline] and β in reproductive tissues. J Steroid Biochem Mol Biol 74:287–296[CrossRef][Medline] in the uterus. Proc Natl Acad Sci USA 97:5936–5941 transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens. Endocrinology 140:5566–5578 and β on cyclin D1 gene expression. J Biol Chem 277:24353–24360 is ligand- and proteasome-dependent. Nat Cell Biol 3:15–23[CrossRef][Medline] and ERβ at AP1 sites. Science 277:1508–1510 and steroid receptor coactivator-1. Mol Endocrinol 14:518–534 reveals a role for ligand in the nuclear distribution of the receptor. Mol Biol Cell 10:471–486B-dependent gene activity. EMBO J 25:798–810[CrossRef][Medline] exchange and chromatin dynamics are ligand- and domain-dependent. J Cell Sci 119:4101–4116 and β. Endocrinology 138:863–870 and estrogen receptor β to partial estrogen agonists/antagonists. Mol Pharmacol 54:105–112-1 promoter in response to tamoxifen and other estrogen receptor antagonists, but not in response to estrogen. Mol Endocrinol 13:418–430 on responsive promoters is an integral feature of estrogen signaling. Mol Cell 11:695–707[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  H. A. Molenda-Figueira, S. D. Murphy, K. L. Shea, N. K. Siegal, Y. Zhao, J. G. Chadwick Jr., L. A. Denner, and M. J. Tetel Steroid Receptor Coactivator-1 from Brain Physically Interacts Differentially with Steroid Receptor Subtypes Endocrinology, October 1, 2008; 149(10): 5272 - 5279. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, ER
, ERβ. C, Fluorescence loss in photobleaching (FLIP) analysis of the nuclear shuttling of ER
, ER
