| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
in Single Living Cells Using Green Fluorescent Protein Color Variants
Department of Anatomy and Neurobiology, Kyoto Prefectural University of Medicine, Kyoto 602-8566, Japan
Address all correspondence and requests for reprints to: Mitsuhiro Kawata, Department of Anatomy and Neurobiology, Kyoto Prefectural University of Medicine, Kawaramachi Hirokoji, Kamigyo-ku, Kyoto 602-8566, Japan. E-mail: mkawata{at}basic.kpu-m.ac.jp.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(ER), we investigated the subcellular dynamics of AR and ER fused with green fluorescent protein color variants in single living cells using time-lapse microscopy and the technique of fluorescence recovery after photobleaching. AR and ER showed punctate colocalization in the nucleus with estrogen, but not androgen. N-terminal AR deletion mutant did not form a nuclear punctate pattern with either androgen or estrogen. In the presence of AR, but not ER, N-terminal AR deletion mutant formed a punctate nuclear pattern with androgen. AR had different mobility depending on the ligand and the presence of ER. On the other hand, AR had little effect on the stability of ER. ER mutant that does not bind coactivators did not alter the mobility of AR. Taken together, using an imaging technique, we clarified that possible homo/hetero dimerization between AR and ER could be attributed to androgen-estrogen interaction in living cells. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Androgen and estrogen are secreted in both sexes, and the balance between androgen and estrogen actions modulates the biological responses of target tissues (2). Some roles of estrogen in the male are now known. Much evidence has been reported showing that the production of estrogen from androgen through aromatase activity is crucial not only for the differentiation of male organs but also for male sexual functions (3, 4, 5). It has been shown that two subtypes of ER, ER and ERß, are expressed, alone or together, throughout the male reproductive tract (6, 7, 8), and male mice lacking ER are infertile (9). In addition, the expression of ER or ERß in non-sex-specific tissues, including the brain, bone, and visceral organs, suggests that estrogen has widespread functions in the male (10).
Recent evidence suggests that estrogen represents another important natural ligand for AR and may play an essential role in the development of the male reproductive system through AR (11).
Immunohistochemical studies using specific antibodies revealed broad overlap of the regions of AR and ER expression in various tissues. Colocalization of AR and ER in the same cell is observed in many brain regions (12, 13). In addition, the interaction of these receptors in disease development is now accepted. AR and ER are frequently coexpressed in breast cancer (14, 15). Androgen and estrogen work synergistically to induce a higher incidence of breast cancer than either hormone treatment alone (16). AR and ER/ß have been identified in normal and cancerous prostate tissues (17, 18). Medical antiandrogenic therapy for prostate cancer using pharmacological doses of estrogen can also induce paradoxical proliferation of neoplastic cells (19).
Elucidation of the subcellular localization of the steroid hormone receptors and determination of the trafficking and interaction between the receptors provide important information about steroid hormone actions. Studies on the subcellular localization of steroid hormone receptors in living cells using fusion constructs of green fluorescent protein (GFP) with the glucocorticoid receptor (GR) (20, 21, 22, 23), mineralocorticoid receptor (MR) (23, 24), ER/ß (25, 26, 27, 28), progesterone receptor (29), and AR (30, 31, 32) have been reported. Those studies have shown that steroid hormone receptors can be divided into three categories based on their unliganded distribution: ER/ß are located primarily in the nucleus, GR and AR are located primarily in the cytoplasm, and MR and progesterone receptor have mixed distribution in both the cytoplasm and nucleus. In all cases, addition of ligand leads to almost complete nuclear translocation. Recent studies using the fluorescence recovery after photobleaching (FRAP) technique have shown that fluorescently tagged steroid hormone receptors such as GR and ER are highly mobile and dynamic in the unliganded state, but liganded forms are less mobile (33, 34, 35).
Multicolor variants of GFP have enabled us to perform multicolor imaging of at least two different molecules in a single cell in a real time analysis. Dual color imaging of GR and MR (23), and of ER and ERß (27), has recently been reported and their interactions have been investigated.
In the present study, to elucidate possible interactions between AR and ER in single living cells, we report here, for the first time, the imaging results of the visualization of AR and ER simultaneously in single cells using two different spectral variants of GFP, yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP). We found that the subcellular localization of YFP-AR changed in response to estradiol (E2) as well as testosterone (T) and that E2 induced the colocalization of YFP-AR and CFP-ER in the nucleus with a discrete punctate formation. We established the specificity of AR and ER interaction using N-terminal AR deletion mutant (
NAR) and C-terminal ER deletion mutant. Furthermore, to investigate the mobility of receptors in the nucleus in the presence vs. the absence of the ligand, we performed FRAP using YFP-AR or YFP-ER. AR had different mobility depending on the ligand and the presence of ER. On the other hand, AR had little effect on the stability of ER. ER mutant that does not bind coactivators did not alter the mobility of AR. We discuss the mechanism by which AR might interact with ER within the same cell.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
expression vector was transiently transfected into COS-1 cells. The molecular weights of AR and ER are 110 kDa and 68 kDa, respectively, and the expected sizes of fusion proteins with the 27-kDa GFP color variants YFP and CFP are 137 kDa and 95 kDa, respectively. Immunoblotting of proteins of the transfected cells with anti-GFP antibody, which cross-reacts with CFP and YFP, revealed the production of proteins of the expected sizes (Fig. 1B
). To ensure that the GFP portion of our construct did not interfere with AR or ER activity, the constructs were cotransfected with a mouse mammary tumor virus (MMTV)-luciferase (Luc) reporter plasmid or an ER-responsive reporter plasmid, estrogen response element (ERE)-Luc, into CV-1 cells. As an internal standard, the ß-galactosidase gene was also cotransfected. When treated with 10-6 M T or 10-7 M E2 for 30 h, YFP-AR and CFP-ER were able to induce significant activation of the reporter gene (Fig. 1, C and D). This confirmed that the receptor fusion proteins used in this study are effectively transfected into living cells and maintain the capacity to function in ligand-dependent transcription.
|
expression vectors and AR- or ER-reporter gene constructs were used to identify and measure potential functional effects of AR-ER interaction. Coexpression of CFP-ER with YFP-AR decreased YFP-AR transactivation in the presence of both of T and E2 (Fig. 1C
). On the other hand, coexpression of YFP-AR with CFP-ER did not significantly modulate CFP-ER transactivation (Fig. 1D).
Subcellular Distribution of YFP-AR or CFP-ER in Singly Transfected Living Cells
In COS-1 cells, which express neither endogenous AR nor ER, YFP-AR showed a diffuse predominantly cytoplasmic fluorescence in the absence of hormone (Fig. 2A, E, G, and K). When the cells were incubated with 10-6 M T or 10-9 M T, the nuclear accumulation of YFP-AR started, and the fluorescence became primarily or exclusively nuclear within 60 min (Fig. 2, B and F). A subnuclear punctate pattern of YFP-AR was evident at around 60 min, and the production of nuclear fluorescent foci induced by 10-9 M T was more moderate than that induced by 10-6 M T. In all experiments, the nucleoli demonstrated almost no fluorescence.
|
). The fluorescent foci in the nucleus were more moderate than that of foci induced by T. In the presence of 10-10 M E2, only slight nuclear import of YFP-AR was observed at 60 min (Fig. 2L) and nuclear clusters were not induced.
CFP-ER transfected in COS-1 cells before ligand treatment had a diffuse distribution throughout the nucleoplasm but was excluded from the nucleoli. Little fluorescence was observed in the cytoplasm (Fig. 2, C, I, and M). After 10-7 M E2 treatment, CFP-ER was found in a punctate nuclear distribution (Fig. 2J). The punctate pattern appeared within 10 min (data not shown) and reached its full extent within 60 min after ligand treatment. In contrast, 10-10 M E2 did not give rise to any punctate nuclear redistribution of CFP-ER (Fig. 2N).
Addition of 10-6 M T had little effect on CFP-ER redistribution in transfected COS-1 cells. CFP-ER showed a diffusely homogeneous pattern in the nucleus without accumulation in the nucleoli (Fig. 2D).
Dynamics of YFP-AR and CFP-ER Distribution in Response to Ligands in the Cotransfected Cells
In cells cotransfected with YFP-AR and CFP-ER, before hormone addition, YFP-AR had a diffuse cytoplasmic distribution (Fig. 3, A, E, I, and M), whereas CFP-ER showed a homogeneous nuclear pattern, as observed in singly transfected cells (Fig. 3, C, G, K, and O). After addition of 10-6 M T, YFP-AR was translocated from the cytoplasm into the nucleus with a clear punctate distribution pattern (Fig. 3B), whereas CFP-ER continued to have diffuse distribution throughout the nucleoplasm and to be absent from the nucleoli (Fig. 3D). In contrast, in the presence of 10-7 M E2, both YFP-AR and CFP-ER showed a punctate pattern in the nucleoplasm (Fig. 3, F and H). After addition of 10-8 M E2, even distribution of YFP-AR in both the cytoplasm and the nucleus without a nuclear punctate pattern was observed (Fig. 3J), although CFP-ER showed a nuclear punctate pattern in the nucleoplasm (Fig. 3L). On the other hand, 10-10 M E2 caused no significant change of the subcellular localization of either receptor: YFP-AR showed only slight nuclear translocation, and CFP-ER was not redistributed into a punctate nuclear pattern (Fig. 3, N and P).
|
are distributed in the intranuclear region and whether the two receptors are colocalized, the images of fluorescence of both receptor proteins were captured and were pseudocolored with red and green, respectively. Confocal laser scanning microscopy revealed that after treatment with 10-6 M T, YFP-AR (red) and CFP-ER (green) showed almost no punctate colocalization (Fig. 4A). In contrast, confocal images clearly showed the partial colocalization of the punctate fluorescence of YFP-AR and CFP-ER in the nucleus after treatment with 10-7 M E2 (Fig. 4C). We analyzed the degree of colocalization in more detail by plotting fluorescence intensity along a single line drawn through an optical section (line scan) of the dual-labeled nucleus. A line scan of the nuclear profile demonstrated that none of the peaks of YFP-AR and CFP-ER overlapped with 10-6 M T treatment, whereas some of the peaks of YFP-AR and CFP-ER overlapped with 10-7 M E2 (Fig. 4, B and D). The peak patterns of some other punctate cluster pairs were coincident, although the relative intensities of YFP-AR and CFP-ER were different (Fig. 4D). This indicates that the distributions of the YFP-AR and CFP-ER clusters are positively correlated after addition of 10-7 M E2.
|
, A and B). Optical scan analysis showed that YFP-AR and CFP-ER clusters in the nucleus were positively correlated (Fig. 5C). When 10-8 M E2, which caused a nuclear punctate localization of CFP-ER but not YFP-AR, was added in combination with 10-6 M T, both receptors showed positively correlated nuclear punctate clusters (Fig. 5, D–F).
|
Mutant using the yeast and mammalian two-hybrid systems (36) was reported. To examine the specificity of the positive correlation between YFP-AR and CFP-ER, we exploited YFP- or CFP-fused N-terminal deletion AR mutant (YFP-NAR, CFP-NAR) and YFP or CFP fused C-terminal deletion ER mutant (YFP-ERC341, CFP-ERC341) (Fig. 6, A and B).
|
NAR expressed in COS-1 cells was present in both the cytoplasm and the nucleus but was excluded from the nucleoli. The large number of the cells showed more nuclear fluorescence (Fig. 7, A, D, and E). Exposure to 10-6 M T resulted in a more rapid transfer of YFP-NAR to the nucleus than that of YFP-AR. In addition, YFP-NAR formed nuclear foci, and many of them coincided with the nucleolar region within 5 min (Fig. 7Ai). T (10-9 M) also induced the rapid nucleolar-nuclear diffuse distribution of YFP-NAR, but the nucleolar accumulation was not observed (data not shown). The subnuclear punctate pattern of YFP-NAR, as was shown in the YFP-AR transfected cells, was not induced even 60 min after 10-6 M T treatment (Fig. 7Ai.)
|
NAR, which contained the dimerization domain, was changed by dimerizing with wild-type AR, we produced CFP-AR (Fig. 6, A and B) and cotransfected with YFP-NAR. In YFP-NAR- and CFP-AR-coexpressing cells, CFP-AR was present in both the cytoplasm and the nucleus in the absence of ligand treatment, whereas YFP-NAR had the same distribution pattern as was shown in the singly transfected cells. These results suggest that the subcellular localization pattern of full-length AR was altered in the presence of N-terminal AR mutant. After 10-6 M T treatment, CFP-AR began to move into the nucleus and formed a punctate nuclear pattern within 60 min. In contrast to singly transfected cells, YFP-NAR in the presence of CFP-AR did not show a nucleolar accumulation or nucleolar distribution even after 10-6 M T incubation (Fig. 7Aii). Moreover, YFP-NAR was found in a punctate nuclear distribution, which was nearly coincident with that of CFP-AR (Fig. 7, B and C).
Effect of ER on a Nucleolar Accumulation of N-Terminal AR Mutant with Androgen
When CFP-ER was coexpressed with YFP-NAR, YFP-NAR had predominantly nuclear, but with significant cytoplasmic, distribution in the absence of hormone, as we observed in singly transfected cells. After addition of 10-6 M T, YFP-NAR was translocated into the nucleus, including the nucleolus, although neither nucleolar accumulation nor any punctate nuclear distribution of YFP-NAR were observed. CFP-ER continued to have diffuse distribution in the nucleus except for the nucleolus both before and after 10-6 M T treatment (Fig. 7Di).
In singly transfected cells, as we previously reported (28), CFP-ERC341, which lacks almost all the ligand binding domain, did not show punctate pattern in the presence of T or E2. In doubly transfected cells with YFP-NAR and CFP-ERC341, in the presence of 10-6 M T, YFP-NAR was imported into the nucleus and found in nucleolar accumulation. As expected, CFP-ERC341 did not show ligand-dependent relocalization even in the presence of YFP-NAR (Fig. 7Dii).
Estrogen-Induced Nuclear Translocation of N-Terminal AR Deletion Mutant in the Presence or Absence of ER
E2 (10-7 M) also induced the rapid nuclear translocation of YFP-NAR, but the nucleolar accumulation was not observed. The subnuclear punctate pattern of YFP-NAR, as was shown in the YFP-AR-transfected cells, was not induced even 60 min after 10-7 M E2 treatment (Fig. 7Ei). In YFP-NAR and CFP-AR doubly transfected cells, 10-7 M E2 also induced nuclear translocation of both YFP-NAR and CFP-AR, although a nuclear punctate pattern of both receptors was not observed (Fig. 7Eii). When CFP-ER was coexpressed with YFP-NAR, after addition of 10-7 M E2, YFP-NAR was translocated into the nucleus without forming a nucleolar accumulation and an intranuclear punctate pattern, and CFP-ER was found in a punctate distribution in the nucleoplasm (Fig. 7Eiii). In doubly transfected cells with YFP-NAR and CFP-ERC341, in the presence of 10-7 M E2, YFP-NAR was imported into the nucleus without a nucleolar accumulation and a nuclear punctate pattern, and CFP-ERC341 did not show the ligand-dependent relocalization (Fig. 7Eiv).
Ligand-Induced Nuclear Punctate Distribution of Full-Length AR in the Presence of C-Terminal ER Deletion Mutant
YFP-ERC341 did not have the ligand-dependent punctate pattern formation activity even in the presence of full-length CFP-AR, whereas CFP-AR was translocated into the nucleus and found in a punctate distribution in the presence of 10-6 M T or 10-7 M E2 (Fig. 7F).
Mobility of Nuclearly Distributed AR after Ligand Incubation
Although both T and E2 induced a punctate nuclear distribution of AR, the degree of transcriptional activation of AR by T is much higher than that by E2 (31). We therefore hypothesized that the stability of AR in the nucleus after ligand incubation is different between T and E2. To test this hypothesis, we carried out FRAP analysis of YFP-AR-transfected COS-1 cells. We analyzed the mobility of nuclear YFP-AR 2 h after the addition of 10-6 M T. After the bleaching, we observed a clear fluorescence-lacking area in the bleach zone and gradual recovery with a half-recovery time (t1/2) of 16.1 ± 1.30 sec (n = 30 cells) (Fig. 8, Ai and B). As we expected, cells treated with 10-7 M E2 showed more rapid recovery (t1/2 = 8.44 ± 0.404 sec; n = 30, P = 6.06 x 10-7) than cells treated with 10-6 M T targeted cells (Fig. 8, Aii and B).
|
on the Stability of AR in the Nucleus on AR dynamics, we carried out a FRAP analysis of cells cotransfected with YFP-AR and CFP-ER. Pretreatment with 10-6 M T or 10-7 M E2 induced a punctate nuclear pattern of YFP-AR. The recovery of YFP-AR was not significantly different (10-6 M T: t1/2 = 15.0 ± 1.23 sec; n = 30, 10-7 M E2: t1/2 = 6.72 ± 0.39 sec; n = 30) from that observed in cells transfected with YFP-AR alone (Fig. 8B). CFP-ERC341 also did not have significant effects on the recovery of YFP-AR (10-6 M T: t1/2 = 17.3 ± 1.72 sec; n = 30, 10-7 M E2: t1/2 = 7.14 ± 0.461 sec; n = 30) (Fig. 8B). In both cases, cells treated with 10-7 M E2 showed more rapid recovery of YFP-AR than cells treated with 10-6 M T targeted cells (Fig. 8B). On the other hand, the recovery of the T + E2-induced punctate nuclear pattern of YFP-AR showed a significant difference in CFP-ERC341 cotransfected cells (10-6 M T + 10-7 M E2: t1/2 = 16.1 ± 1.16 sec; n = 30, 10-6 M T + 10-8 M E2: t1/2 = 15.9 ± 1.11 sec; n = 30) compared with CFP-ER cotransfected cells (10-6 M T + 10-7 M E2: t1/2 = 11.7 ± 0.749 sec; n = 30, P = 2.33 x 10-2, 10-6 M T + 10-8 M E2: t1/2 = 12.4 ± 1.02 sec; n = 30, P = 0.219 x 10-2) or cells in the absence of either ER form (10-6 M T + 10-7 M E2: t1/2 = 16.4 ± 1.18 sec; n = 30, P = 0.138 x 10-2, 10-6 M T + 10-8 M E2: t1/2 = 15.9 ± 1.03 sec; n = 30, P = 1.83 x 10-2) (Fig. 8B).
Effect of AR on the Stability of ER in the Nucleus
We also carried out a FRAP analysis of YFP-ER transfected into COS-1 cells. In the absence of ligand, or in the presence of 10-6 M T, YFP-ER was extremely mobile, and most of the receptor passed through the bleach zone even within bleaching time (4.45 sec) (Fig. 8Aiii). In contrast, when YFP-ER-transfected cells were treated with 10-7 M E2 for 2 h, the fluorescence recovery could be detected with a t1/2 of 15.2 ± 1.25 sec (n = 30) (Fig. 8, Aiv and C).
To examine the effect of AR on ER dynamics, we also carried out a FRAP analysis of cells cotransfected with YFP-ER and CFP-AR or CFP-NAR. In the presence of CFP-AR or CFP-NAR, pretreatment with 10-6 M T did not slow immediate recovery of YFP-ER (t1/2 < 4.45 sec) (data not shown). As we expected, in the presence of CFP-NAR, after 10-6 M T + 10-7 M E2 or 10-6 M T + 10-8 M E2 treatment, YFP-ER recovered on a time scale statistically similar to that observed when treated with 10-7 M E2 alone (10-7 M E2; t1/2 = 15.9 ± 0.894 sec; n = 30, 10-6 M T + 10-7 M E2: t1/2 = 15.3 ± 0.881 sec; n = 30, 10-6 M T + 10-8 M E2: t1/2 = 14.2 ± 0.775 sec; n = 30), and contrary to expectation, CFP-AR also had no effect on YFP-ER mobility (10-7 M E2; t1/2 = 14.0 ± 0.727 sec; n = 30, 10-6 M T + 10-7 M E2: t1/2 = 13.5 ± 1.01 sec; n = 30, 10-6 M T + 10-8 M E2: t1/2 = 14.3 ± 1.01 sec; n = 30) (Fig. 8C).
Effect of ER Coactivators on the Stability of AR in the Nucleus
To examine the effect of ER coactivators on the stability of AR, we generated ER mutant (ERK367A) that does not bind coactivators by replacing lysine-367 with alanine residue (37). ERK367A showed almost the same intranuclear distribution pattern as that of wild-type ER. After E2 treatment, ERK367A was found in a punctate nuclear distribution. Partial colocalization of the punctate fluorescence of AR and ERK367A in the nucleus after treatment with E2 was observed (data not shown). We carried out a FRAP analysis of YFP-AR in the presence of CFP-ERK367A. CFP-ERK367A did not have significant effects on the recovery of YFP-AR (10-6 M T: t1/2 = 18.1 ± 1.27 sec; n = 30, 10-7 M E2: t1/2 = 7.56 ± 0.55 sec; n = 30, 10-6 M T + 10-7 M E2: t1/2 = 16.6 ± 1.46 sec; n = 30, 10-6 M T + 10-8 M E2: t1/2 = 17.2 ± 1.43 sec; n = 30).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
to examine the interaction of AR and ER in a single cell. We confirmed that the YFP and CFP fusion proteins were transcriptionally active using MMTV-Luc and ERE-Luc reporter assays, respectively, and had the expected molecular mass in immunoblots. Using these two plasmid constructs, we first performed dual-color imaging to reveal the dynamics of YFP-AR and CFP-ER movement in a single living cell. The subcellular localization of AR in the absence of ligands was quite controversial. Unliganded AR was found in the nucleus in target tissues (38, 39), whereas in transfected cells overexpressing AR, the AR was located in both the cytoplasm and the nucleus, or exclusively in the nucleus, depending on the cell type (40, 41). Our results obtained here showed that in COS-1 cells YFP-AR was distributed mainly in the cytoplasm in the absence of ligands. In the majority of transfected cells, upon androgen exposure YFP-AR rapidly migrated into the nucleus. These results are generally consistent with those of previous studies that also employed GFP-AR fusion proteins (30, 31, 32). In addition to T, other natural steroid hormones, E2 and progesterone, transactivate AR (11, 31). Furthermore, E2 induces abnormal growth of prostate cells (42, 53). The effectiveness of E2 for translocating AR to the nucleus and inducing a punctate nuclear distribution suggests that E2 represents another important natural ligand for AR that may play an essential role in AR function. Estrogen is produced by aromatization of T, and the AR-E2 complex may be important for natural sexual differentiation (11).
The nuclear localization of some steroid receptors is dependent upon ligand binding when the receptor translocates from the cytoplasm to the nucleus. This has been shown in live cells using GFP-fused constructs of GR (20, 21, 22, 23) and MR (23, 24). This provides a very obvious and efficient mechanism for preventing transcription by unliganded receptors. However, unliganded and liganded ER, ERß, and progesterone receptor are located predominantly in the nucleus (25, 26, 27, 28, 29). The structural features that differentiate these two types are presently unknown.
In some sex steroid hormone target tissues, AR and ER are expressed in the same cell (12, 13). Interactions between sex steroid hormones and their receptor plays an important role in the mechanisms of development, growth, homeostasis, physiologic responses, and pathologic disturbances. It is thus important to examine the interaction between AR and ER in living cells. In the present study, we investigated the subcellular distribution of AR and ER in single living cells. Before hormone addition, YFP-AR was distributed in the cytoplasm and CFP-ER had a diffuse intranuclear distribution, as observed in singly transfected cells. After addition of T, both YFP-AR and CFP-ER were localized within the nucleus. However, only YFP-AR was found to have a punctate distribution. E2 (10-7 M) treatment also induced a punctate YFP-AR nuclear distribution regardless of the presence of CFP-ER. CFP-ER was redistributed in a punctate pattern. Partial overlapping of YFP-AR and CFP-ER was observed. Line scan analysis indicated that the visual colocalization was not due merely to occasional overlapping of randomly distributed YFP-AR and CFP-ER. The spatial distribution of YFP-AR and CFP-ER was positively correlated.
The binding of the hormone causes the receptor protein to undergo a conformational change and receptor activation, which increases the binding of the receptor to the nuclear matrix. The nuclear matrix has been reported to associate with transcription sites, including steroid receptor binding sites (known as hormone response elements) in the regulatory region of the target genes. The androgen response element has a distinct sequence motif of GGTACAnnnTGTTCT, whereas the ERE sequence motif is AGGTCAnnnTGACCT. The palindrome motif suggests that the corresponding receptors interact with DNA as homodimers (1). One possible explanation of the partial colocalization of AR and ER after 10-7 M E2 treatment is that these elements are localized closely on genomic DNA.
Heterodimerization between members of the nuclear hormone receptor superfamily is not uncommon and provides one basis for the diversity of steroid signaling pathways. For instance, the 9-cis-retinoid X receptor is known to serve as a common partner for several nuclear receptors, including the retinoic acid receptor, the thyroid hormone receptor, the vitamin D receptor, and the peroxisome proliferator-activated receptor (44, 45, 46, 47, 48, 49, 50, 51). Many other receptors, once thought to bind only as homodimers, are now known to interact with trans-regulatory proteins. In a recent report, the possibility of AR/ER heterodimerization was demonstrated using yeast and mammalian two-hybrid systems (36). Thus, the ligand-induced partial colocalization of AR and ER may be due to AR/ER heterodimer formation in the nucleus. Another possibility is that they may be translocated to the same sites individually to play similar roles, e.g. structural remodeling of chromatin, because many punctate structures produced by the ligands corresponded well with the sites of Brg-1, which is involved in chromatin remodeling (27, 52, 53, 54, 55). The close localizations of AR and ER in the nucleus may affect their functional modulation.
On the other hand, low E2 concentration less than 10-7 M could not induce the nuclear colocalization of YFP-AR and CFP-ER. It is intriguing to consider that E2 gradient in tissues, which has aromatase enzyme activity, may play an important role in cellular functions.
To establish the specificity of AR and ER interaction, we utilize YFP-NAR and CFP-NAR and YFP-ERC341 and CFP-ERC341, which lack domains involved in a direct interaction between the receptors (36). YFP-NAR was defective in maintaining the correct intracellular distribution of AR. YFP-NAR was primarily both cytoplasmic and nucleic in the absence of 10-6 M T. Upon addition of 10-6 M T, this mutant translocated to the nucleus and formed the nucleolar accumulation. However, in the presence of CFP-AR, YFP-NAR, which was intact in the domain involved in the homodimerization, was excluded from the nucleoli and was found in a punctate distribution, which was nearly coincident with that of CFP-AR. We provide, for the first time, support for the homodimerization of AR in living cells by an imaging technique (56). On the other hand, 10-7 M E2 treatment, which induced the punctate nuclear colocalization of YFP-AR and CFP-ER, resulted in the punctate nuclear relocalization of CFP-ER, but not YFP-NAR. One possible explanation of this result is that the heterodimerization between AR and ER was disturbed by deleting the N-terminal part of AR. In the present study, it was clear that the degree of colocalization of YFP-AR and CFP-ER was much less than that of YFP-NAR and CFP-AR. That result could mean that ER homodimers were preferentially formed rather than heterodimers of AR/ER by 10-7 M E2 treatment (28). Contrary to our expectation, in the presence of CFP-ER, 10-6 M T-induced nucleolar accumulation of YFP-NAR, which was observed in singly transfected cells, was suppressed. Cotransfection with CFP-ERC341 resulted in nucleolar accumulation of YFP-NAR after 10-6 M T incubation. These results suggest the possible interaction between C-terminal AR and C-terminal ER, which could not be identified in in vitro protein interaction analysis (36).
AR has long been known to associate with the nuclear matrix (57). A recent report also showed that punctate clusters of GFP-AR are associated with the nuclear matrix (31, 58). E2 is capable of cross-reacting with AR and effective for translocating AR into the nucleus and for forming punctate foci; however, the transactivational ability of AR is much less than that of T (31). In the present study, FRAP analysis revealed that the punctate nuclear distribution of AR induced by E2 was much less stable than that induced by T. Our results may provide insights into the biological effects of stable ligand-receptor-nuclear matrix interactions that lead to effective transactivation. In previous reports, ER was shown to be capable of inhibiting transcriptional activation of androgen-induced AR transactivation (36, 59). In the present study, the T-induced punctate nuclear distribution of AR in the presence of E2-bound ER was less stable than that in the presence of ER mutant, which lacked ligand binding domain. This may explain, at least in part, the ability of ER to inhibit T-induced AR transactivation as we showed in Fig. 1C. The mechanism by which E2-bound ER increases AR mobility is unknown. On the other hand, AR had little effect on the stability of ER. A possible explanation for this unsymmetrical correlation between AR and ER in their mobility may be related to the difference in their molecular size. Transcriptional regulation by members of the nuclear hormone receptor superfamily requires the mediation of distinct subclasses of coregulators. We can not exclude the possibility that ER may be related to restrictions imposed by immobile coregulators.
While the punctate nuclear pattern of ligand-associated steroid hormone receptors has long been recognized, the functional significance of this pattern has remained unclear. There have been numerous reports linking transcription to the nuclear architecture. However, a recent report demonstrated that the number and appearance of ER foci did not overlap with RNA polymerase II immunoreactive foci (25). Transcription complexes may be in a dynamic state between assembly and disassembly, and the time they spend actively transcribing genes may be minimal. Furthermore, the transcriptional activity of steroid receptors may depend not only on their ability to enter the nucleus and bind DNA but also on their interactions with coactivator or corepressor proteins (25). In the present study, ERK367A, which is defective for binding to coactivators, including steroid receptor coactivator-1, transcriptional intermediary factor-1, and -2, was found not to decrease the stability of AR. These results suggest that efficient recruitment of coactivators is necessary for possible interaction between AR and ER because this mutation does not affect steroid or DNA binding of ER (37). Although it is currently difficult to assess which of the many coactivators that have been shown to interact with AR or ER are actually recruited in intact cells, it is tempting to speculate that AR-ER interaction may be modulated in each cell or tissue depending on the levels or affinities of coactivators.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
NAR or CFP-NAR, were generated by a similar method to the YFP-AR or CFP-AR construction. In this case, a BglII restriction enzyme site was inserted into just upstream of amino acid 539 of AR.
Construction of YFP-ER, CFP-ER, YFP-ERC341, and CFP-ER341 has previously been described previously (28). An XhoI restriction enzyme site was inserted just upstream of the first ATG of rat ER cDNA cloned into pUC118 vector (provided by Dr. M. Muramatsu, Department of Biochemistry, Saitama Medical School, Saitama, Japan). This mutated cDNA was cloned into the XhoI-EcoRI site of the MCS of pEYFP-C1 and pECFP-C1 plasmid (CLONTECH). The C terminus of EYFP or ECFP was fused to the N terminus of rat ER in the resulting fusion protein, YFP-ER or CFP-ER. YFP-ERC341 or CFP-ERC341, C-terminal ER deletion mutant, and YFP or CFP chimeras, respectively, were generated by creating a stop codon after amino acid 341 in the YFP-ER or CFP-ER expression construct. The ligated fragments in the constructions were sequenced to verify that the correct reading frames had been constructed. Point mutation in the codon coding for lysine 367 of rat ER was generated by QuikChange site-directed mutagenesis system in pECFP-ER using desired oligonucleotide primer sets. Mutagenesis was verified by automated DNA sequencing.
Cell Culture and Transfection
COS-1 cells were maintained in DMEM (Invitrogen, Carlsbad, CA), supplemented with 10% fetal bovine serum overnight in a four-well multidish with 16-mm diameter wells (Nunc A/S, Roskilde, Denmark) at an initial plating density of 3 x 104 cells per well in 500 µl of medium. Cells were transfected by the liposome-mediated method using LipofectAMINE PLUS (Invitrogen) with plasmid DNA in 150 µl of OPTI-MEM (Invitrogen) plus 1.5 µl of LipofectAMINE solution and 250 ng of plasmid DNA per well for 3 h and then cultured in 500 µl DMEM for 5 h at 37 C, according to the manufacturer’s instruction. The medium was replaced with 1 ml OPTI-MEM for 14 h before observation.
For ligand stimulation, cells were treated with various concentrations of T (10-6 M and 10-9 M) and/or E2 (10-7 M, 10-8 M, and 10-10 M) at 37 C.
Immunoblotting
For immunoblot analysis, the cells transfected with pEYFPC1-AR, pECFPC1-ER, pECFPC1-AR, pEYFPC1-NAR, pECFPC1-NAR, and pECFPC1-ERC341, respectively, were solubilized in the lysis buffer. Proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Immunobilon-P, Millipore Corp., Bedford, MA) using a semidry blotting apparatus (Transblot-SD, Bio-Rad Laboratories, Inc., Hercules, CA). Immunoblotting was performed with anti-GFP antiserum (CLONTECH, catalog no. 8367–1), which cross-reacts with CFP and YFP, diluted 1:100, and visualized using chemiluminescent detection (ECL; Amersham Pharmacia Biotech, Buckinghamshire, UK).
Examination of Transcriptional Activity
The procedure for the transcription assay has been published elsewhere (25, 28). CV-1 cells were transfected in 35-mm dishes with 10 ng of plasmid containing the effecter gene (YFP-AR and/or CFP-ER) and 1 µg of plasmid encoding MMTV-Luc or ERE-Luc. pAct-ß-galactosidase (1 µg), which is a ß-actin promoter-driven ß-galactosidase expression construct, was used as an internal control, and the relative luciferase activity was obtained by normalizing by the ß-galactosidase activity. Cells were transfected using 1 ml OPTI-MEM with a LipofectAMINE-DNA complex for 3 h. After removal of the transfection mixtures, the cells were maintained in DMEM with 10% fetal bovine serum for 5 h, and then the medium was changed to OPTI-MEM. Cells were treated with 10-6 M T and/or 10-7 M E2 for 30 h and then harvested in lysis buffer. Cell lysates were centrifuged at 12,000 rpm for 2 min at 4 C, and the luciferase activity of the resulting supernatants was assayed at 25 C using the luciferase assay system Pica Gene (Toyo Inki, Tokyo, Japan). Three independent experiments were carried out. Results were reported as means ± SE. Statistical analysis was performed and values of P < 0.05 were considered significant.
Fluorescence Microscopy Analysis
As previously reported by Nishi et al. (23, 25), transfected cells were analyzed under a fluorescence microscope (IX-70, Olympus Corp., Tokyo, Japan) in a room with the temperature controlled at 37 C. Fluorescence images were captured using a cooled charge-coupled device camera (Photometrics, Tucson, AZ). Cells were observed with a x40 objective lens. YFP fluorescence was viewed using a filter set with 500-nm excitation and 545-nm emission, and a 525-nm dichroic mirror (Omega Optical Inc., Brattleboro, VT); and CFP fluorescence was observed using a filter set with 440-nm excitation and 480-nm emission, and a 455-nm dichroic mirror. Data were evaluated with the image analysis software program Meta Morph (Universal Imaging Corp., West Chester, PA). For high-resolution analysis, an image deconvolution procedure was applied to a series of images.
Confocal Laser Scanning
Twenty-five thousand cells transfected with 250 ng YFP chimera construct and 250 ng CFP chimera construct were plated on a poly-L-lysine-coated 35-mm glass bottom dish (Matsunami Glass, Inc., Ltd., Japan), and cultured in OPTI-MEM for 24 h. After 60 min of hormone treatment, cells were fixed with 4% paraformaldehyde in PBS and then observed with a x60 oil-immersion lens (Nikon, Tokyo, Japan). Images were collected with a confocal laser microscope (LSM 510, Carl Zeiss, Inc., Oberkochen, Germany). YFP fluorescence was viewed using a filter set with 514-nm excitation and 530- to 600-nm emission, and a 458/514 nm dichroic mirror; and CFP fluorescence was observed using a filter set with 458-nm excitation and 475- to 525-nm emission, and a 458/514 nm dichroic mirror.
FRAP Analysis
For living cell FRAP imaging, we used an LSM 510 (Carl Zeiss) system equipped with a CO2-controlled on-stage heating chamber. All experiments were done at 37 C. Fluorescent images of a single optical section were taken at 1.0-sec intervals after bleaching. Imaging scans were acquired with the laser power attenuated to 5% of the bleaching intensity. The region of the cell to be bleached was demarcated circularly (
= 2 µm) and scanned 100 times in 4.45 sec using a laser beam (488 nm and 514 nm) at maximum power through an oil immersion x40 objective lens. For quantitative analysis of FRAP, fluorescence intensities of the bleached region were measured at each time point using LSM 510 software. Averaged half-recovery time (t1/2), the time for fluorescence to recover by 50% due to diffusion, was determined from the fluorescence recovery curve generated from 30 individual cells of three independent experiments. Results were reported as means ± SE. Statistical analysis was performed and values of P < 0.05 were considered significant.
|
ACKNOWLEDGMENTS |
|---|
cDNA. In addition, we would like to thank H. Nakauchi for technical assistance. |
FOOTNOTES |
|---|
Abbreviations: AR, Androgen receptor; CFP, cyan fluorescent protein; E2, estradiol; ER, estrogen receptor; ERE, estrogen response element; FRAP, fluorescence recovery after photobleaching; GFP, green fluorescent protein; GR, glucocorticoid receptor; Luc, luciferase; MMTV, mouse mammary tumor virus; MR, mineralocorticoid receptor; NAR, N-terminal AR deletion mutant; T, testosterone; YFP, yellow fluorescent protein.
Received for publication July 27, 2002. Accepted for publication October 9, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
within the testis and excurrent ducts of the rat and marmoset monkey from perinatal life to adulthood. J Endocrinol 153:485–495[Abstract]
