Estradiol-binding Kinetics of the Activated and Nonactivated Estrogen Receptor*

The dissociation of PHlestradiol from the estrogen receptor of calf uteri followed a two-component exponential function. The first component rate constant, kl, was 0.12 + 0.01 min-I, while the second component, k,, had a rate constant of 4.0 f 0.3 x 10m3 min-’ at 25”. Increasing the fractional activation of the receptor before initiating the [3H1estradiol dissociation measurement subsequently pro-duced a corresponding decrease in the fraction of the first component without a change in either rate constant. Su- crose gradient centrifugation analysis indicated that incubation of the estrogen receptor at 25” in 0.4 M sodium thiocyanate inhibited the 4 to 5 S receptor transformation; only the 4 S monomer was observed. The release of PHlestradiol from the 4 S receptor at 25” in 0.4 M sodium thiocyanate follows a single exponential function with a rate constant of 0.17 min-I, similar to the first component of the biphasic dissociation curve in the absence of sodium thiocyanate. Sucrose gradient centrifugation analysis of the 5 S receptor dimer showed that the PHlestradiol dissociation at 25” in 0.4 M KC1 was a single exponential function

The dissociation of PHlestradiol from the estrogen receptor of calf uteri followed a two-component exponential function.
The first component rate constant, kl, was 0.12 + 0.01 min-I, while the second component, k,, had a rate constant of 4.0 f 0.3 x 10m3 min-' at 25". Increasing the fractional activation of the receptor before initiating the [3H1estradiol dissociation measurement subsequently produced a corresponding decrease in the fraction of the first component without a change in either rate constant. Sucrose gradient centrifugation analysis indicated that incubation of the estrogen receptor at 25" in 0.4 M sodium thiocyanate inhibited the 4 to 5 S receptor transformation; only the 4 S monomer was observed. The release of PHlestradiol from the 4 S receptor at 25" in 0.4 M sodium thiocyanate follows a single exponential function with a rate constant of 0.17 min-I, similar to the first component of the biphasic dissociation curve in the absence of sodium thiocyanate.
Sucrose gradient centrifugation analysis of the 5 S receptor dimer showed that the PHlestradiol dissociation at 25" in 0.4 M KC1 was a single exponential function with a rate constant of 4.4 x 10e3 min'. The [3Hlestradiol association rate constants, k,,, were approximately 1 X 10' Mm1 min-' at 25", whether measured in the presence of 0.4 M KCl, in 0.4 M sodium thiocyanate, or without salt. The association constants, K,, calculated from the rate constants (k+,/k_, and k+,/k-,) indicated that the estrogen receptor has two affinity states. The inactive 4 S monomer has a lower affinity for estradiol (1 x lo9 M-l) than the activated 5 S receptor dimer (2.8 x lOlo M-l). Consequently, estradiol-binding drives the monomer-dimer equilibrium toward the higher affinity 5 S dimer state of the receptor. These data demonstrate that the 4 to 5 S receptor transformation is a biochemical regulatory mechanism modulated by estradiol.
Two forms of the estrogen receptor have been described, a cytoplasmic form with a sedimentation coefficient of 4 S and a nuclear 5 S species. The nuclear 5 S form is generated from * This work was supported by Grant HD 06707 from the National Institutes of Health. A preliminary report was presented at the Meeting of the Endocrine Society in Chicago, Illinois, June 1977 Endocrinology 100, 210A). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
the cytoplasmic 4 S receptor by an estradiol-and temperaturedependent reaction (l-3).
The activation or transformation of the receptor from the 4 S to the 5 S state has also been investigated in cell-free systems; activation of the receptor is contingent upon brief incubation of the estradiol receptor complex at 20"-37" (4,5). The homogenate was centrifuged at 20,000 x g for 10 min, followed by centrifugation of the supernatant at 220,000 x g for 30 min; the supernatant fraction is referred to as the "cytosol." The protein concentration of the cytosol, determined by the method of Lowry et al. (12), was 6 to 10 mglml. The dissociation rate constants were: k-,, 0.09 min-'; k-,, 6.5 x 1O--3 mined'.
showed a slower PHlestradiol dissociation, similar to K-, of the two-component dissociation curve. Uterine cytosol was equilibrated with 10 nM [3H]estradiol, then made 0.4 M with respect to sodium thiocyanate (NaSCN) to retard the hydrophobic interactions between the 4 S estrogen receptor and its complementary monomeric unit, thereby preventing the formation of the 5 S dimer during incubation at 25". The PHlestradiol receptor was incubated at 25" with or without an excess of unlabeled estradiol for 0, 2, 5, 10, and 15 min. The dissociation of the radioactive steroid was stopped by cooling the aliquots to 0" and adding charcoal/dextran. Sucrose gradient centrifugation analysis showed that the predominant form of the receptor was 4 S and that the bound [3H]estradiol was exchanged for unlabeled estradiol during incubation at 25" (Fig. 3A). A single first order dissociation curve was seen when the replacement of radioactivity of the 4 S peak was plotted (Fig. 3B). The rate constant was 0.17 min', in good agreement with the 0.12 min-' value observed for k-, of the two-component 13Hlestradiol dissociation kinetics at 25" (Table  I). When the 5S estrogen receptor was subjected to 0.4 M NaSCN for 2 h at O", before the [3H]estradiol dissociation was measured, sucrose gradient analysis indicated that it had completely dissociated to the 4 S receptor and that it had a rapid rate of [3H]estradiol dissociation. After removal of the 0.4 M NaSCN by Sephadex G-25 filtration, the receptor resumed the slower PHlestradiol dissociation kinetics (data not shown).
To bring about the dissociation of [3H]estradiol from the 5 S estrogen receptor, uterine cytosol was equilibrated with 10 nM [3H]estradiol, then incubated for 30 min at 28" (inducing receptor activation and the 4 S to 5 S transformation).
The cytosol was made 0.4 M with respect to KC1 or NaCl, then incubated at 25" with and without the excess unlabeled estradiol for 0, 0.5, 1, 2, and 2.5 h. Sucrose gradient analysis showed that incubation of the estrogen receptor with an excess of unlabeled estradiol in the presence of 0.4 M KC1 or 0.4 M NaCl for 2.5 h at 25" dissociated the [3Hlestradiol-5 S receptor complex. Incubation of the 5 S receptor without the excess estrogen demonstrated the stability and lack of aggregation of the receptor during a 2.5-h incubation at 25" ( (Table I) The association rate constants in Buffer TD, Buffer TD with 0.4 M KCl, or 0.4 M NaSCN were not significantly different (Table III), which indicates that the NaSCN only affected the dissociation rate of [3H]estradiol. Using either the fast (k-J or the slower (k-,1 dissociation rate constant, an estimation of the association constants (K,) from the ratios k+&, and k+,/lz-, indicates a 25-to 30-fold difference in the affinity between the two states of the receptor at 25".

Influence of Trypsin on Kinetics of [3H]Estradiol
Dissociation -Limited proteolysis of the estrogen receptor by trypsin or endogenous proteases at 0" produces an estrogen-binding fragment of the receptor which sediments at approximately 4 S and has a molecular weight of 6 x 104, which is lower than that of the native 4 S estrogen receptor (7 to 8 x 109 (15, 16). The trypsin-treated estrogen receptor cannot undergo either the temperature-induced 4 S to 5 S receptor transformation or the temperature-enhanced nuclear-binding activity (17).  Cytosol in k-2 k-s k +t k+,lL k+,lL min-' Uterine cytosol containing 7.2 mg of protein/ml was incubated with trypsin, 0, 2, 20, and 200 pg/ml, for 60 min at 0" followed by the addition of soybean trypsin inhibitor for 30 min at 0". The [3H1estradiol dissociation at 25" in the presence or absence of soybean inhibitor, but without trypsin, showed a typical two-component dissociation curve, with a k-, of 0.11 min' and a k-, of 3.4 x 10e3 min-'. The [3H]estradiol. receptor complex incubated with trypsin, 2 pg/ml of cytosol, showed a decrease in the fraction of the first component; km1 was 0.10 min-' while K_, was slightly increased, to 5.0 x 10m3 min-' (Fig. 6A). Sucrose gradient centrifugation analysis in the absence of 0.4 M KC1 showed that the native estrogen receptor sedimented as an 8 S protein while the trypsin-treated estrogen receptor sedimented as a 4 S estrogen-binding protein. The addition of trypsin, 2 pg/ml, was not sufficient to completely convert the 8 S estrogen receptor to the 4 S sedimenting protein. Exposure to trypsin, 20 or 200 pg/ml of cytosol, completely converted the 8 S estrogen receptor to a 4 S sedimenting estrogen-binding protein (data not shown). The estrogen receptor incubated with trypsin, 20 or 200 pg/ ml of cytosol, showed [3Hlestradiol dissociation kinetics with  6B). Trypsin had an identical effect on the dissociation kinetics of the estrogen receptor whether the receptor was in the activated or nonactivated state. The receptor inactivation curves (Fig. 6) indicate that no significant specifically bound [3H]estradiol was lost during the 25" incubation; the soybean inhibitor effectively inhibited the trypsin activity. In comparison with the native receptor, tryspin-modified estrogen receptor showed a loss of biphasic [3H]estradiol dissociation kinetics suggesting a loss of the capacity to transform from a nonactive to an active state of the receptor.

DISCUSSION
The first component of the biphasic 13Hlestradiol dissociation curve is the lower affinity state of the receptor, the nonactive 4 S monomer, with a rate constant of k-,. This conclusion is based upon a number of experimentally consistent observations.
The disappearance of the amount of the first component is inversely related to the time that the [3H]estradiol. receptor complex is preincubated at 29" (receptor activation) before the [3H]estradiol dissociation from the receptor is measured. Preincubation of the receptor without estradiol fails to activate the receptor and consequently does not induce the disappearance of the K-, component (Fig. 2). Sucrose gradient centrifugation analysis demonstrates that sodium thiocyanate was effective in inhibiting the 4 S to 5 S receptor transformation and in maintaining the receptor in the 4 S state that shows a single rapid estradiol dissociation Kinetics of Estrogen Receptor Binding 8861 rate constant identical with the k-, of the biphasic dissociation curve (Fig. 3). Sucrose gradient centrifugation analysis demonstrates that the 5 S form of the receptor has a slower rate of estradiol dissociation, equal to the km, of the biphasic PHlestradiol dissociation curve (Fig. 4). The association rate constant K,, of estradiol binding by the receptor is not significantly influenced by the absence or presence of 0.4 M KC1 or 0.4 M sodium thiocyanate (Table III).
Assuming that the association rate constant k,, is an appropriate value for both states of the receptor, the association constants (K,) of the 4 S and 5 S receptor were calculated from the ratios of the rate constants K+,/k-, and k+,lkmt, respectively. This is a reasonable assumption since the formation of the 5 S dimer requires that the 4 S monomer be complexed with estradiol (4,5) and that the existence of a second association rate constant of estradiol associating with an "estradiol-free activated 5 S dime? is an unlikely pathway. The two-component exponential dissociation of estradiol from its receptor shows that the inactive state of the receptor has a faster estradiol dissociation rate and a lower affinity for estradiol (1 x lo9 M-') than the active receptor (2.8 x 10'" M-I).
Since the affinity of the 5 S receptor dimer for estradiol is-higher than that of the 4 S monomer, estradiol shifts the monomer-dimer equilibrium from the inactive to the activated state of the receptor, demonstrating that estradiol-binding is a modulator of the 4 S to 5 S receptor transformation.
The biphasic curve of estradiol dissociating from the receptor depends on three parameters: the two dissociation rate constants Km, and K-, and the fraction of the estradiol dissociating with the k-, parameter. The fraction of the estradiol dissociating from the receptor with the faster rate constant km, is dependent upon the competition between the rate at which estradiol dissociates from the 4 S receptor and the rate of the 4 S monomer dimerizing to form the 5 S receptor. The 4 S receptor dimerization reaction is characterized by the second order rate constant, k,,, (Fig. 7).
These data indicate that a complex interplay exists between estradiol-binding and the state of the receptor. The PHlestradiol equilibrium binding to the receptor at O-4" has generally been accepted to indicate the presence of a single class of binding sites (9) and does not detect the transition of the receptor from the nonactivated to the activated state. The receptor's complex estradiol dissociation kinetics reveals the NON presence of two states of the receptor. The assumption based upon the Scatchard analysis of a single class of binding sites may be incorrect. The receptor's regulatory characteristics and properties that are influenced by estradiol would be very different, depending upon the structure of the receptor. At the present time there is no compelling evidence to indicate whether the 5 S receptor dimer is composed of two estrogenbinding 4 S monomers, or a 4 S estrogen binding monomer and a second, non-estrogen binding monomer (5, 6, 18). Wyman (19) and Levitzki and Schlessinger (20) have described the theoretical basis for the kinetic behavior of oligomerizing protein systems and their relationships to the Hill coefficient, a measure of cooperativity.
The dimerization of identical monomers can show positive cooperativity even in the absence of subunit-subunit interactions, provided a sufficient difference in the affinity for the ligands exists in the two states of the monomer-dimer equilibrium. If we assume that the 5 S receptor is composed of two estrogen-binding 4 S monomers, while considering the higher affinity of the 5 S dimer for estradiol, the positive cooperativity of estradiol-binding should be present. Although most studies of estradiol binding by the receptor indicate that the Scatchard or Hill plots show limited or no evidence of cooperativity (91, several investigators have reported Scatchard plots with "vaulted ceilings," which suggest positive cooperativity (10, 11). We also observed Scatchard plots with vaulted ceilings and Hill coefficients of 1.3 to 1.7.' Because of the possibility of differential receptor inactivation at lower, as compared with higher, steroid concentrations, and of small errors in separating free and bound steroids, we cannot be certain that the deviation from linearity of the Scatchard plots at low estradiol concentrations is not due to experimental artifacts rather than to positive cooperativity.
In addition, the equilibrium binding analysis of the estrogen receptor is usually carried out at O-4". Under these conditions the 4 S to 5 S receptor transformation is very slow (5), consequently limiting the equilibrium between the lower and higher estradiol affinity states of the receptor. The difference in affinity between the lower and higher affinity states is approximately 25-to 30fold, as estimated from the kinetic data (Table III), which is well within the range of the values reported and the limits of accuracy of [3H]estradiol equilibrium measurements (9). If we assume that the 5 S receptor is a heterodimer composed of an estrogen-binding monomer and a second, nonestrogen-binding monomer, then positive cooperativity of estradiol binding may not necessarily be observed. The nonestrogen-binding monomer would associate with the estrogenbinding monomer during receptor transformation, increasing the affinity for estradiol by the maintenance of the 5 S receptor. This phenomenon which is similar to the helcotropic effect (21) is distinct from the positive cooperative model of the homodimer.
A disparity between the affinities of the cytoplasmic and nuclear forms of the estrogen receptor has been previously reported. Some reports (22-24) suggest that the cytoplasmic form of the estrogen receptor has an estradiol affinity approximately lo-fold higher than the nuclear estrogen receptor.
(The estradiol affinity of the nuclear estrogen receptor was indirectly measured by the [3H]estradiol exchange method at 37" (25) and compared with an equilibrium binding assay of the cytosol at o".) This is contrary to our observation that the activated 5 S receptor or nuclear form has a higher affinity ' A. C Notides and D. E. Hamilton, unpublished observations.