Nature of nuclear acceptor sites for glucocorticoid- and estrogen-receptor complexes.

Abstract In cell-free systems, glucocorticoid receptor-dexamethasone complexes from hepatoma tissue culture (HTC) cells bind to HTC cell nuclei, and estrogen receptor-estradiol complexes from immature uteri bind to uterine nuclei. The binding is specific in that it is of high affinity (apparent Kd = 2 to 3 x 10-10 m) and involves a limited number of acceptor sites (about 3850 per haploid genome). Furthermore, nonradioactive receptor-steroid complexes inhibit the binding of the homologous receptors complexed with tritiated steroid. In each case, the maximum number of complexes bound to isolated nuclei is very similar to the number found in the nuclear fraction of intact cells exposed to saturating concentrations of steroid. Receptor-glucocorticoid complexes also bind with high affinity (apparent Kd = 2.6 x 10-10 m) to about 2150 acceptor sites (per haploid genome) in isolated uterine nuclei, even though the immature uterus is devoid of glucocorticoid receptors. The uterine acceptor sites for glucocorticoid receptors are different from those for estrogen receptors. Thus, the binding of one class of receptor-steroid complex is not inhibited by the other type. The acceptor sites for glucocorticoid receptors in both uterine and HTC cell nuclei are destroyed by DNase, whereas those for estrogen receptors are resistant. Nuclear-bound receptor-dexamethasone complexes are released from both uterine and HTC cell nuclei at a lower concentration of NaCl than is required to release bound receptor-estradiol complexes. Finally, acceptors for estrogen receptors are not found in HTC cell nuclei. It is concluded that nuclear acceptors for glucocorticoid receptors are not restricted to target tissues and therefore nuclei from certain cells are able to bind more than one class of steroid receptor. Where this occurs, however, the results show that each type of nuclear acceptor site binds only one of the two classes of receptor-steroid complex examined. Glucocorticoid and estrogen acceptor sites differ chemically and physically, possibly reflecting differences in the molecular mechanisms of action of these hormones.

In each case, the maximum number of complexes bound to isolated nuclei is very similar to the number found in the nuclear fraction of intact cells exposed to saturating concentrations of steroid.
Receptor-glucocorticoid complexes also bind with high affinity (apparent Kd = 2.6 x lo-lo M) to about 2150 acceptor sites (per haploid genome) in isolated uterine nuclei, even though the immature uterus is devoid of glucocorticoid receptors. The uterine acceptor sites for glucocorticoid receptors are different from those for estrogen receptors. Thus, the binding of one class of receptor-steroid complex is not inhibited by the other type. The acceptor sites for glucocorticoid receptors in both uterine and HTC cell nuclei are destroyed by DNase, whereas those for estrogen receptors are resistant.
Nuclear-bound receptor-dexamethasone complexes are released from both uterine and HTC cell nuclei at a lower concentration of NaCl than is required to release bound receptor-estradiol complexes. Finally, acceptors for estrogen receptors are not found in HTC cell nuclei.
It is concluded that nuclear acceptors for glucocorticoid receptors are not restricted to target tissues and therefore nuclei from certain cells are able to bind more than one class * This research was supported by Grant  of steroid receptor. Where this occurs, however, the results show that each type of nuclear acceptor site binds only one of the two classes of receptor-steroid complex examined. Glucocorticoid and estrogen acceptor sites differ chemically and physically, possibly reflecting differences in the molecular mechanisms of action of these hormones.
Steroid hormones are involved in the regulation of many aspects of cellular development and metabolism (l-4) ; indeed it appears that all tissues of the body are directly influenced by at least one of the classes of steroid hormones.
In all "target organs," the initial events in steroid hormone action are broadly similar (5, 6). The steroid readily enters the cell and binds with high affinity and stereo-specificity to cytoplasmic receptor proteins.
Conformational changes in the receptor-steroid complex then occur and the complex becomes associated with chromatin in the nucleus.
It is thought that the complex acts as a genetic regulatory element, directly or indirectly altering the transcription of specific parts of the genome to produce ultimately the observed biological effects. Although they have completely different steroid specificities (7-9), sex hormone receptors and glucocorticoid receptors have many features in common (5,6). It is thus probable that their interaction with the cell nucleus would also be similar.
However, studies on the binding of receptor-progesterone complexes to chick oviduct chromatin have led to the proposal that tissue-specific nuclear acidic proteins are essential for this process (lo), whereas a preliminary investigation of the nuclear acceptor sites for glucocorticoid receptors in hepatoma cell nuclei suggests that DNA itself may be the acceptor for these receptors (II).
These results taken together suggest that nuclear acceptor sites for different receptor-steroid complexes may have different chemical and physical characteristics.
This communication describes an attempt to gain further insight into the nature of the nuclear acceptor sites for glucocorticoid receptors by comparing their nuclear binding properties in a cell-free system with those of another extensively studied hormone receptor, the receptor-estradiol complex of the immature uterus (5,6).
Specific binding of the estrogen receptor was observed with uterine nuclei but not with HTC' cell nuclei, whereas the glucocorticoid receptor was bound specifically by qualitatively similar acceptor sites in the nuclei .of both cell types. The chemical and physical properties of the uterine acceptor sites for the estrogen receptor differ significantly from those for the glucocorticoid receptors. METHODS AXD MATERIALS Chemicals and Reagents-'l'he sources of all chemicals and reagents and the composition of all buffers not mentioned below have been described previously (12, 13).
Adhering adipose and vascular tissue was dissected away. Subsequent isolation procedures were at O-4". The tissue was removed from the PBS, quickly blotted on tissue paper, placed in Tricine-T\'aCl buffer (0.01 M N-tris(hydroxymethyl)methylglycine (pH 7.6)-0.15 M NaCl, 0.1 to 0.2 ml per uterus), finely minced with scissors and disrupted in an all-glass homogenizer (Duall, Kontes Glass Co.). Ten-second periods of homogenization were separated by 30-s periods for cooling.
As required for the particular experiment, furt'her portions of the supernatant fraction were left steroid-free or received 1OP M nonradioactive estradiol. They were then centrifuged at 105,000 x g for 1 hour and the soluble fractions (cytosols) were stored at 0" until required (up to 4 hours).
To prepare the nuclear fraction, the pellet from the first, lowspeed centrifugation was resuspended in fresh Tricine-NaCl buffer (10 volumes) and gently rehomogenized.
The suspension was filtered through a pad of washed glass wool and the filtrate was centrifuged at 800 x g for 10 min. The pellet, resuspended in fresh buffer (10 volumes), was centrifuged again and the final pellet was resuspended in the same buffer (1 to 2 ml per uterus) with the final molarity adjusted to 0.25 M with sucrose (2 M). The preparation consisted mainly of intact uterine nuclei and was stored at 0" until required (up to 3 hours).
As required for the particular experiment, furt'her portions were left steroid-free or received 5 X lop8 M nonradioactive dexamethasone.
They were then centrifuged at 105,000 x g for 1 hour and the supernatant fractions (cytosols) were kept at 0" until required (up to 4 hours).
Binding of Steroids by Whole Uteri and Intact HTC Cells-Whole uteri were incubated at 37" for 1 to 2 hours in Dulbecco's modified Eagle's medium (Grand Island Biological). Each conical flask (50 ml) contained up to 10 uteri (1 uterus per ml of medium) and was gassed with 95% O2 + SC/b CO*. In each esperiment, the uteri were divided into two groups, one of which was incubated with the tritiated steroid and the other with the tritiated steroid together with a 500-to lOOO-fold escess of the same nonradioactive steroid.
[KH]Desamethasone was used at 5 X 10V8 M and [3H]estradiol was used at 1 to 2 X 10P8 M.
HTC cells in growth medium were exposed to 5 X 1OF M [3H]desamethasone with and without 5 x 10-j M nonradioactive dexamethasone at 37" for 30 to 60 min.2 At the end of the incubations, the uteri or the HTC cells were washed with cold PUS and fractionated to prepare washed nuclei (see above) which were assayed for DNA and radioactivity (15). The amount of receptor-steroid complex bound by the nuclei was calculated by subtracting the radioactivity bound in the presence of the competing unlabeled steroid (nonspecific binding) from the radioactivity retained in its absence.2 Assay of Specific Binding in Cytosol Preparations-Samples (0.1 ml) of cytosol were assayed before activation (see below) for specific desamethasone or estradiol binding by the charcoal adsorption method or by gel filtration over Sephades G-25 (12). r\Tonspccific binding of steroid n-as estimated with cytosol in which an excess of nonradioactive steroid prevented tritiated steroid binding to the specific receptor proteins but not to the nonspecific binding components.
Specific binding due to the receptor was thus obtained by subtracting the radioactivity resistant to charcoal or excluded from the Sephadex in the presence of the competitor from that obtained with rytosol containing only the tritiated steroid (12).

Activation of Receptor-Steroid
Complexes-The HTC cell receptor-dexamethasone complex does not bind to HTC cell nuclei unless it is "activated" (15). This is achieved conveniently by incubating cytosol containing the steroid at 20" for 30 min with 0.15 M NaCl.
Uterus cytosol, already prepared in 0.15 M NaCl, was merely incubated at 20" for 30 min although it n-as not determined whether this procedure is obligatory for specific binding of the estrogen receptor to isolated uterine nuclei. After activation, the cytosols were passed through Sephadex (G-25) equilibrated with Tricine-KaCl buffer to remove free steroid.
The macromolecular fraction was collected and immediately incubated with nuclei (see below).
In each case, the active form of the receptor-steroid complex (capable of binding to nuclei) was estimated by incubating samples (0.05 to 0.1 ml) of the activated cytosol with a large excess of nuclei sufficient to bind all of the active species. Using this assay, it was found that the uterine system and the HTC cell system have many features in common.
First, no further nuclear binding occurred when cytosol recovered from such an assay was exposed to fresh nuclei.
Second, only part of the specifically bound steroid in activated cytosol was in the form of active receptor-steroid complex, 40 f 6.5% (S.E.M.) in the case of uterine cytosol containing estradiol.
Third, this proportion did not change on subsequent dilution, and, lastly, activated cytosol containing excess competin, 0' nonradioactive steroid was devoid of active tritiated receptor-steroid complex. Binding of Active Receptor-Steroid Comple.xes to h'zlclei-Samples (0.5 ml) of the suspension of washed nuclei (see above) were centrifuged (800 x g for 5 min) and the supcrnatant fractions were discarded.
To the drained nuclear pellet were added 0.1 ml of sucrose (0.8 M), a variable amount of activated gel-filtered cytosol (see legends), and enough Tricine-NaC1 buffer to make the final incubation volume 0.8 ml. The nuclei were resuspended gently with a Pasteur pipette and the suspensions were incubated at 0" for 2 hours (sufficient for equilibrium binding). The nuclei were then removed by centrifugation, washed twice with Tricine-NaCl buffer (1 ml), resuspended (Vortex mixer) in 0.5 ml of water, and assayed for radioactivity and DNA (15). In these experiments, the concentration of the active receptorsteroid complex remaining unbound at equilibrium was calculated by subtracting the nuclear-bound active complex from the amount of active complex added at the start of the incubation.
The latter was determined by incubating in parallel samples of the activated cytosol with excess nuclei (see above). In both systems, the calculated concentration of the active complex was in close agreement with that determined directly (15).
Incubation of Nuclei with DNase-Washed nuclei, resuspended in homogenization buffer with 0.25 M sucrose, were digested with bovine pancreatic DBase I (Worthington, RNasefree). The enzyme was used at 100 pg per ml at 0" for 1 hour or at 500 pg per ml at 25" for 20 min. After digestion, the nuclear suspension was centrifuged, the DNA released into the supernatant fraction was assayed, and the ability of the nuclei to bind receptor-steroid complexes was determined (see above). Release of Receptor-Steroid Complexes from Nuclei by NaCl-After binding active receptor-steroid complexes, the nuclei were washed twice and resuspended in homogenization buffer (0.5 ml) containing NaCl (0 to 0.4 M) for 30 min at 0". The nuclei were removed by centrifugation and the supernatant fraction was assayed for total radioactivity and receptor-steroid complexes (gel filtration method).
At each XaCl concentration, the proportion of the complexes released after 90 min of incubation was the same as after 30 min.

Xpec$c Binding of Estrogen Receptor-Estradiol
Complex to Uterine Nuclei-h whole uteri incubated with estradiol or in intact HTC cells exposed to desamethasone, the steroid enters the cells and binds to cytoplasmic receptor proteins (9,(16)(17)(18)(19). Subsequently, steroid is bound specifically in the nuclear fraction, the receptor-steroid complex disappearing from the cytosol* (20). The results of extensive studies of these reactions in HTC cells2 and in uteri (20, 21) suggest that it is the receptor-steroid complex itself that becomes concentrated in the cell nucleus.
In a cell-free system, the active HTC cell receptor-dexamethasone complex binds specifically to HTC cell nuclei, (15), i.e. it binds with high affinity to a limited number of nuclear accept'or sites. This cell-free binding has a number of features which suggest that it represents the corresponding association in the intact cell2 (15). Isolated uterine nuclei bind the estrogen receptor-estradiol complex (5, 6, 22, 23) and this association was therefore examined using the same technique as for the HTC cell system.
After intact uteri have been incubated with estradiol at 37", an estradiol-protein complex, having in sucrose gradients a sedimentation coefficient of 5 S, can be extracted from the nuclei with KC1 (18, 19). A similar 5 S complex cau be extracted from uterine nuclei incubated at 25" in a cell-free system with uterine cytosol containing cstradiol (19,22). In fact, if cytosol coiltaining estradiol is incubated at 25", a form of the estrogen re- 1. Binding of active receptor-estradiol complex to uterine nuclei. Uterine cytosol containing [3H]estradiol was activated and gel-filtered ("Methods and Materials").
Increasing amounts (0.05 to 0.7 ml) of the macromolecular fraction were incubated with uterine nuclei (0.33 mg of DNA) at 0" for 2 hours and then the free and nuclear-bound active receptor-estradiol complex were estimated ("Methods and Materials").
A, effect of complex concent,ration on nuclear binding; B, Scatchard (27) analysis of the data. absence of nuclei (5,24). Exposure of uterine nuclei to this "activated" cytosol at 0" results in formation of 5 S nuclear complex (25). It appears that, to facilitate this cytosol conversion and to obtain 5 S nuclear binding at 0" under cell-free conditions, low ionic strength, divalent cations, and EDTA should be avoided in the preparation of the cytosol (5, 6, 26). Hence in this study, buffer containing 0.15 M NaCl but no divalent cations or EDTA was used for tissue fractionation and nuclear binding of the estrogen receptor.
It should be noted, however, that cell-free specific nuclear binding of the active estrogen receptor does occur in the absence of NaC1. 3 When uterine nuclei were incubated at 0" with increasing amounts of activated uterine cytosol containing [3H]estradiol, they appeared to have a limited capacity to bind the active receptor-estradiol complex (Fig. 1A). A Scatchard (27) analysis of the data (Fig. IB) was linear suggesting that there is a single class of nuclear acceptor sites reversibly binding the complex with high affinity.
The absence of a component parallel to the abscissa suggests that under the conditions employed the receptor-e&radio1 complex does not bind nonspecifically. Analysis of six such experiments showed that there are 7700 f 1400 sites per nucleus and that the Kd is 2.8 f 0.8 x 1O-1o M ( Table I).
The nuclear-bound estradiol is due to receptor-steroid complexes and not to estradiol alone. The concentration of free estradiol in the cytosol after gel filtration, even allowing for some dissociation of the receptor-steroid complexes, is too low for nuclear binding of free steroid to account for the observed effect (based on experiments in which nuclei were exposed to a range of estradiol concentrations in buffer). In addition, no nuclear binding of [3H]estradiol occurred when nuclei were incubated with activated uterine cytosol containing sufficient nonradioactive cstradiol to prevent the binding of the tritiated steroid to the cgtoplasmic receptors.
If uterine nuclei hay-e a limited number of acceptor sites for the active receptor-estradiol complex, it would be expected that activated cytosol containing only nonradioactive estradiol should competitively inhibit the binding of active receptor-[%]estradiol complexes.
This was tested by measuring the amount of receptor-estradiol complex sedimenting at 5 S can be formed in the show that as in the HTC cell system (15) the affinity of the radioactive complex for the nuclear acceptor sites, but not the number of available sites, was decreased in the presence of the nonradioactive complex (Fig. 2). This suggests that the inhibition is competitive and that uterine nuclei contain a limited number of acceptor sites for the estrogen receptor-estradiol complex. An exchange of nonradioactive and radioactive steroid moieties between active complexes is unlikely considering the slow rate of dissociation of steroid complexes (31) and the low free steroid concentration that would be available for such an exchange reaction.
Even if exchange had occurred it could not account for the observed competition.
Although it did not occur in the experiment shown in Fig. 2 cytosol devoid of steroid sometimes competitively inhibited the nuclear binding of the radioactive complex.
On such occasions, this competition was substantially less than with cytosol containing nonradioactive estradiol.
A similar effect was seen in the HTC cell system (15). It is possible that the steroid is not absolutely necessary for the conformational changes in the cytoplasmic receptor required for nuclear binding (32) or that there is an additional element in cytosol that competitively inhibits the binding.
A number of features of the cell-free nuclear binding suggest that it resembles the binding in the intact uterus.
First, there is no steroid binding without the receptor.
Second, the association is of high affinity as it also appears to be in the intact tissue. Third, the number of estradiol molecules bound specifically in the nuclear fraction of whole uteri incubated with the steroid is very similar to the estimated capacity of isolated nuclei for the receptor-estradiol complex (Table  I). Corresponding similarities have been noted for nuclear binding in the HTC cell system (15).
Although in the present study sucrose gradient analysis of the nuclear binding has not been attempted, preliminary experi-ments3 suggest that there are no significant differences in the sucrose density gradient sedimentation patterns of estradiol complexes extracted from nuclei bound in whole uteri or in this cell-free system. Thus in both tissues, receptor-steroid complexes specifically to their homologous nuclei.
To obtain furthe:: formation about the nature of these nuclear acceptor sites, it was decided to see whether uterine and HTC cell nuclei can bind the receptor-steroid complex from the other cell type (heterologous complex) as well as the homologous complex. glucocorticoids which bind specifically to cytoplasmic receptor proteins (9, 12) and induce tyrosine aminotransferase (13) in IITC cells, estrogens, at physiological concentrations, neither induce this enzyme (13) nor are they bound specifically by soluble preparations of HTC cells.3 The immature rat uterus contains receptor proteins binding estrogens with high affinity (for review see Ref. 6). Two experimental findings make it exceedingly unlikely that this organ also contains receptors for glucocorticoids.
First, in cytosol from immature uteri no specific binding of [3H]dexamethasone could be demonstrated at concentrations in the range lo-* to 5 X 10e8 M. These concentrations are sufficient to saturate specific glucocorticoid receptors in all glucocorticoidsensitive tissues studied4 and result in maximal induction of tyrosine aminotransferase in HTC cells (13). The failure of the assay to detect specific binding of dexamethasone was not due to the presence in the cytosol of competing endogenous corticosteroids.4 Using the same charcoal assay, it was possible, on the other hand, to demonstrate binding of [3H]estradiol with high affinity to a single class of receptors with properties similar to those described by other investigators (see Ref. 6 for review). Second, after intact uteri have been incubated at 37" with [3H]estradiol (1 to 2 x lo-* M), most of the specifically bound steroid is found, as expected, in the nuclear fraction (Table I) whereas, when [3H]dexamethasone (5 x lo-* RI) was used, specific binding of this steroid by the nuclear fraction was less than 14% of that found with estradiol (Table I).
Thus the immature uterus, which responds to estrogens (33) and has estrogen receptors (B), has no glucocorticoid receptors. Nuclear Binding of Heterologous Receptor-Steroid Complexes-HTC cell nuclei, incubated at 0" with activated uterine cytosol containing [%H]estradiol, bind the estrogen receptor-estradiol complex.
Ho\\-ever, this binding was not specific since there was no evidence for a limited number of nuclear acceptor sites or high affinity binding (Fig. 3, A and B).
In contrast, uterine nuclei do contain specific acceptor sites for the glucocorticoid receptor (Fig. 4A).
Considerations similar to those applied to the binding by HTC cell and uterine nuclei of their homologous receptors (see Ref. 15 and above) discount the possibility that the nuclear binding of dexamethasone by uterine nuclei is due to free dexamethasone and not to the glucocorticoid receptor. Thus binding of the glucocorticoid receptor complex to uterine nuclei is remarkably similar to the binding of the homologous estrogen receptor (see Fig. 1, A and B, Table I) and must therefore also bc regarded as specific. There do appear to be fewer uterine acceptor sites for the glucocorticoid receptor than for the estrogen receptor, although the ranges for each class overlap.
The possibility was therefore considered t'hat the uterine acceptor sites for the glucocorticoid receptor might be similar to and even coincide with those for the estrogen receptor.
Estrogen and Glucocorticoid Receptors Do Not Compete with Each Other for Nuclear Acceptor Sites-The effect of the glurocorticoid receptor-dexamethasone complex on the binding of the estrogen reccptor-[3H]estradiol complex was examined using the same experimental method as was used to show that active receptor-estradiol complex bound with nonradioactive estradiol competitively inhibits the binding of active rcceptor-[3H]estradiol to uterine nuclei (see Fig. 2).
Increasing amounts of activated, gel-filtered uterine cytosol containing [3l :]estradiol were incubated at 0" with uterine nuclei in the presence and absence of a constant amount of activated HTC cell cytosol bound with nonradioactiv dexamethasone. As shown in Fig. 5, binding of estrogen receptors was not affected by glucocorticoid receptors even though in the same experiment receptor bound with nonradioactive estradiol did inhibit the binding of the radioactive estrogen receptor. This suggests that the glucocorticoid and estrogen receptors bind to spatially distinct uterine acceptor sites. This conclusion is reinforced by the results shown in Fig. 6   Receptor-estradiol complex released from uterine nuclei ( l ); receptor-dexamethasone complex released from uterine (0) or HTC cell (A) nuclei.
for glucocorticoid receptor complexes in HTC cell nuclei in that they bind the receptor-desamethasone complex with the same affinity, they have the same DBase sensitivity and the clution of receptor-dexamethasone complexes from them occurs at similar XaCl concentrations.
These results confirm the previous proposal that DKh is a major component of the nuclear acceptor sites for glucocorticoid receptors. Indeed it is possible that these acceptor sites consist of specific base sequences in the DNA.
Regardless of their chemical nature, these sites probably occur in the nuclei of all cells. First, very many tissues contain c\-toplasmic receptors for glucocorticoids and, second, at least one tissue, the immature uterus, which is devoid of these receptors nonetheless has glucocorticoidspecific nuclear acceptors.
From these considerations, it seems possible that these acceptor sites might be essential elements in all nuclei, but are susceptible to modulation by glucocorticoid receptors only in tissues containing such receptors. hlternatively, it is conceivable that at some time during uterine development, cytoplasmic receptors for glucocorticoids, and thus sensitivity to these hormones, might be present.
In contrast, several lines of evidence show that the nuclear acceptor sites for glucocorticoid receptors are different from those for estrogen receptors.
First, estrogen acceptor sites are absent from HTC cell nuclei.
Second, estrogen receptors and glucocorticoid receptors do not compete for nuclear acceptor sites in uterine nuclei.
Third, treatment of uterine nuclei with DNase assumed that the binding of receptor-steroid complexes to nuclear acceptor sites influences transcription of the genome. If this is the case, the results may imply that the molecular mechanisms of action of estrogens and glucocorticoids are dissimilar.