Estriol and estradiol interactions with the estrogen receptor in vivo and in vitro

$3.00+ 0.00 0022-4731/84 Copyright0 1984PergamonPressLtd

1984 J. steroid Biochem. Vol. 20, No. 4B, PP. 1039-1046, Printedin Great Britain. All rights reserved

ESTRIOL AND ESTRADIOL INTERACTIONS WITH THE ESTROGEN RECEPTOR IN VIVO AND IN IX’XO R. ENZIO MUELLER, DAVID M. BEEBE,EVA BERCEL,ABDULMAGED M. TRAISH and HERBERTH. WOTIZ Boston University School of Medicine, Department of Bi~hemistry, 80 B. Concord Street, Boston, MA 02118, U.S.A. Summary-_The cytosolic estrogen receptor (calf uterus) bound to estradiol (I$) at 0°C changes from a state with fast into a state with slow I& dissociation rates when placed at 28°C. This temperature accelerated transition in receptor affinity for its ligand takes place within 10min at 28°C. Similarly, receptor bound to estriol (E& at 0°C changes, when heated, from a state with fast into a state with slow E, dissociation. The main difference between RE, and RE, was that E, dissociates from unheated 8s RE, and heat-transformed 5s RE, at a much faster rate than E, from RE,. In the mature ovariectomized rat a slow dissociating 5s receptor estrogen complex is found in nuclei

I h after injection of [3H]Er or [‘H]E,. In vim dissociation of these 2 estrogens from this nuclear bound

receptor formed in vivatakes place at rates similar to those from heat-transformed

cytosolic RE, or RE,

complexes. Addition of pyridoxal 5’-phosphate (PLP) to the slowdiss~iating hot-transform 5s estrogen receptor complexes causes rapid dissociation of E, or E,; this effect is dose-dependent and is not due to disruption of 5s dimers, since after PLP addition RE, and RE, sediment unchanged as 5s dimers. The presence of a large excess of non-radioactive 4s RE, does not interfere with the temperature induced rapid transition of 4s R[‘H]E, complexes from the state with fast into a state with slow E, dissociation kinetics. A model is presented to explain the temperature induced biphasic estrogen dissociation from the receptor. It is proposed that the low affinity 4s RI!&monomer undergoes a temperature and estrogen dependent conformation change, such that the ligand is “locked” into the receptor’s binding site. This conformational change results in the formation of a high affinity 4s monomer from which estrogen dissociates at a slower rate. This reaction is independent from subsequent 4s to 5s dimerization (transformation). The different rates of &and dissociation from the low and high affinity 4s receptors reflect the different interactions (hydrophobic and hydrogen bonding) of E, and E, with the estrogen binding domain.

INTRODUCTION The biological activity of estrogens is directly related to their affinity for the estrogen receptor and to their ability to induce those changes in receptor properties required for prolonged interaction of the estrogen receptor complexes with nuclear acceptor sites [l-3]. The following three reactions are probably related to the formation of biologically active receptor estrogen complexes. Reaction 1: activation, defined as the change in RE, which results in increased binding capacity for nuclei and is associated with the appearance of positively charged residues [4-6] on the surface of the receptor. It has been proposed that activation precedes transformation and represents dissociation of the 8s aggregate into 4s subunits with concomitant conformational changes of the recep tor’s nuclear binding domain [7, 81. Reaction 2: the transition of RE2 from a state with fast Ez dissociation into a state with slow E, dissociation [9]; it

Supported by Grant HD15213 and CA28856. This is publi~tion No. 114 from the Hubert H. Humphrey Cancer Research Center of Boston University.

is independent from transformation [IO, 111. Reuction 3: transformation, represents dimerization of 4s RE, complexes to form a stable 5s entity [ 12, 131. It is well established that estriol is a partial agonist of estradiol. Thus, estriol has weak uterotrophic activity. This has been related to the inability of the RE, complexes to reside in the nucleus for a time period adequate to induce gene expression [1,2]. Consistent with this view is the finding that at 35°C E, dissociates more rapidly than E2 from the unheated as well as from the heat-transformed calf uterine cytosol estrogen receptor [14]. Thus, upon in vivo administration of an equal dose of & or E,, a higher concentration of RE2 complexes will interact for a prolonged time period with the nucleus of the target cell. This could explain why estradiol in admixture with estriol is less estrogenic than an equal dose of estradiol alone [15] and why EJ acts as a potent estrogen when administered in multiple daily doses 1161. It has recently been shown 1171 that equilibrium binding of E2 to the calf uterine estrogen receptor is an interaction showing positive cooperativity; such cooperativity was no longer observed when clomiphene was admixed with E,. One can

1039

1040

R.

ENZIO

speculate that the antagonistic effects of E, on E, induced uterine growth [15] may be related to an inhibiting effect of E, on positive coooperativity (cf. paper by S. Sasson and A. Notides in this issue). We have examined the interactions of E,, and E, with the estrogen receptor in cell free systems (calf uterine cytosol), and in the mature ovariectomized rat. MATERIALS AND METHODS

MCILLER

rl

ul.

and confirm the original observations of Weichman and Notides[9]. At 28 ‘C the dissociation of [‘H]E, from heat-transformed 5s R[‘H]E, is slow and follows a single exponential curve; from the semilogarithmic plot in Fig. 1, we calculate a half time (t,J of 130 min, and a dissociation rate constant Kdlss= 5.3 x 10 3min ‘. The half-time of the heattransformed 5s R[‘H]E, is much shorter (50min), with a dissociation rate constant Kdlbi= 14 x lo-’ min ‘. When unheated non-transformed

Animals Mature (150-200 g body wt) ovariectomized (14 days p.o.) female Sprague-Dawley rats were used. Studies with the calf uterine estrogen receptor were performed with tissue from immature calves (5-15 g uterine wet wt) procured and stored in liquid N, as described elsewhere [8].

6Ok! :: 1:

Chemicals and (2,4,6,7[3H])estradiol (90-105 Ci/mMol) (2,4,6,7[3H])estrio1 (85-105 Ci/mmol) were purchased from New England Nuclear Corporation. All other reagents were analytical grade and obtained from commercial sources.

\ \

Buffers Buffer TD: 40mM Tris, 1 mM DTT, pH 7.2 at 28°C. Buffer BSM: 200 mM borate, 125 mM sucrose, 5 mM MgCl,, pH 8.0 at 0°C. The procedures for preparation of uterine cytosol and nuclei, formation of R[3H]E, and R[3H]E3 complexes and their measurement by the dextran coated charcoal (DCC) or by the hydroxylapatite (HAP) technique have been described elsewhere [8]. In uiuo labeling of uterine estrogen receptors was obtained by injection of [‘HIestrogen in 0.9% NaCl containing the steroid dissolved in ethanol as previously described [18]. Sucrose density gradient analysis and measurement of dissociation kinetics of E, and E, from the cytoplasmic and nuclear estrogen receptor were performed as described elsewhere [8]. The experiments with pyridoxal 5’-phosphate were carried out in borate buffer to allow SchitT’s base formation [5]. RESULTS

Dissociation kinetics of E2 and E3 jLom receptor estrogen complexes formed in vitro The dissociation of E, from RE, displays biphasic kinetics [19,20]. Weichman and Notides[9] demonstrated that E2 dissociates rapidly from unheated, non-transformed 8s or 4s RE, complexes; in contrast, Ez dissociation from heat-transformed 5s RE2 dimers occurs at a slow rate. We proposed [3] that the weak estrogenicity of E, is possibly related to its fast dissociation rate from the receptor; this would result in inadequate nuclear retention of receptor with lesser estrogenicity. The data in Fig. 1 support this concept

60

30 Time

90

120

(mln)

Fig. 1. Dissociation kinetics of [‘HJE, and [jH]Ej from cytosolic estrogen receptor. Calf uterine cytosol (7mg protein/ml) prepared in TD buffer was incubated for 3 h at 0°C with [‘H]E* (5 nM) or [)H]EX (20 nM). Parallel incubations with [‘HIsteroid and 200-fold excess of unlabeled DES were carried out to determine nonspecific binding. All samples were treated with a pellet of dextran coated charcoal[8] and mixed with 1/lOvol of DES (final cont. 1 x 10-6M). The samples were then placed at 28°C to induce exchange of receptor-bound [‘H]E, or [3H]E, with DES. A parallel incubation of cytosol labeled with [‘HIsteroid, treated with DCC and reconstituted to 5 nM [jH]E, or 20 nM [lH]E, was performed to measure receptor stability at 28°C; as previously reported [8] receptor loss does not exceed 10% over a period of 120 min. At the indicated times aliquots of cytosol (0.2 ml) were added to siliconized test tubes containing 0.5 ml of a 70% hydroxylapatite (HAP) slurry and kept at 0°C for 30 min with intermittent blending on a Vortex mixer. HAP washing and radioactivity counting was done as described [8]. All data have been corrected for nonspecific binding (less than 15%) and for receptor degradation (cf. stability incubation). The inset represents the dissociation of [3H]E, and [sH]E, from the low affinity receptor state after correcting for the contribution of the slow dissociating component [9]. The monophasic slow dissociating entity (solid lines) represents [3H]steroid dissociation from the estrogen receptor which, after 3 h at O”C, had been heat-transformed for 30min at 28’C. The sedimentation coefficients of unheated and heated R [3H]estrogen were 4.4s and 5.2s in 0.4 M KC1 sucrose gradients (data not shown).

1041

Estriol and estradioi interactions with estrogen receptor

8s R[3H]E, (low salt buffer) was placed at 28”C, biphasic exponential dissociation kinetics were observed. The rate constant, K_,, for the [)H]E, dissociation from the low affinity R13H]E, state is 0.19 min - 1(inset of Fig. 1); during heating, transition to the slow dissociating state was completed within approx. 10 min. Under identical conditions [‘HJE, dissociated from non-transformed R[3H]E3 complexes with a rate constant, K _ ,, of 0.15 min - ’ (inset of Fig. I), a value which is not significantly different from that observed with the receptor estradiol complex. In a separate experiment using a different cytosol preparation (Fig. 4), the rate constants of estrogen ~ss~iation from non-~ansfo~ed RE3 and RE, at 28°C were 0.29min-’ and 0.28min-’ respectively. As with RE,, the transition of R[-‘H]E, into the slow dissociating state was completed within the first 10min. These data are in very good agreement with a previous report [14], in which estrogen was shown to dissociate from the non-transformed receptor with a dissociation rate constant, K_,, of 0.12 + 0.01 min-’ for E, and 0.16 + 0.02 for E,; the estrogen dissociation rate constants from the heattransformed RE complex were 4 + 0.13 x 1O-3 min-’ with RE, and 17 +0.2 x 10-3min-’ with RE,. ~is~oc~~tion kinetics of E2 and E3 from estrogen complexes formed in vivo

exists in the dissociation rate constant of Ej from the heat-transform calf uterine 5s RE, (Kdis = 14 x 10W3min-’ in Fig. 1 and jr(diss= 15 x IOw3min-’ in Fig. 4) and that measured with nuclear bound RE, formed in vivo in the rat. This may be related to the animal species or to differences in the experimental conditions. For instance, the buffer composition used in the cell free experiment in Fig. 1 was different from that used in this experiment. Not shown in the present report are the properties of the RE, and RE, complexes formed in viuo which were recovered in the cytosolic fraction (approximately 30% of the total RE complexes labeled in vivo). Preliminary data analysis suggests that cytoplasmic RE2 and RE, was not transformed into

receptor

So far, the transition of RE2 from a state with fast into a state with slow E, dissociation has been analyzed only in vitro; the systems used were cytosolic RE, [9], nuclear bound RE, formed upon in vitro translocation of cytosolic salt-activated 4s or heatactived 5s RE, [8] and nuclear bound RE, formed upon in vitro incubation of uterine tissue with E, 1211. The data presented in Fig. 2 demonstrate that also in z&o a slow dissociating receptor-estrogen complex is observed. For this experiment ovarie~tomi~d rats were injected with t3H]E2 or [3H]E3 and sacrificed after 1 h. Since the sensitivity of in viva formed RE, to pyridoxal S-phosphate was also to be determined, tissue was homogenized in BSM buffer [5] containing unlabeled DES to prevent secondary [3H]E, labeling of unoccupied receptor sites during homogenization. The dissociation kinetics of [3H]E2 or [3H]E3 were measured by incubating the nuclear suspension at 28°C as described previously [8]. As indicated in Fig. 2, the nuclei contained only the slow-dissociating receptor estrogen complex; [-‘H]E, dissociated from Ri3H]E, with a t,,, = 140 min and a dissociation rate constant of 4.9 x 10-3min-‘. These data confirm that the slow dissociating RE2 state formed in vitro (Fig. 1) is not an experimental artifact since the E2 dissociation rate is similar to that of nuclear RE, formed in vivo. Figure 2 shows that [3H]E3dissociates from nuclear bound receptor formed in vivo with a t,,2 of 19min and a Giss = 36 x 10V3min-‘, confirming that also in vivo receptor interacts with E, with lower affinity than with Ez. Some discrepancy

ae t

I

I

I

I

30

60

90

120

i min) Fig. 2. In vitro dissociation kinetics of estrogen from nuclear hormone-receptor complexes formed in viva. Mature ovariectomized rats (7 animals/group) were injected ip with 1 pg [3H]EZ(10.4 Ci/mmol) or 1 pg [‘H]E, (9.2 Ci/mmol). Time

Steroid was dissolved in ethanol and mixed with saline (1% ethanol); animals were injected with 0.5 ml of this solution. One h after injection the animals were killed by cervical dislocation. The uteri were homogenized in ice-cold BSM buffer containing I x 10m6M nonradioactive DES. The washed nuclear pellet was resuspended in this buffer, aliquoted into siliconized tubes and rH]steroid exchanged with unlabeled DES by placing the tubes into a 28°C water bath. Nonspecific binding was measured with nuclear suspensions which had been heated at 60°C for 45min to destroy the estrogen receptor. At the indicated time points the tubes (duplicates) were placed in an ice bath, mixed with 0.5 mi of hydroxylapatite slurry and further incubated for 30min. Under these conditions nuclear [3H]steroid-receptor complexes which could have leaked out from the nuclei are rebound to HAP[8]. The HAP-nuclei suspension was washed 3 times, radioactivity extracted with ethanol and counted[8]. The data presented in this figure have been corrected for non-specific binding (less than 10% of total binding).

R. E%zto M~LLEK

1042

the slow dissociating state, since steroid dissociation kinetics were biphasic with rates similar to those (Fig. 1) obtained with in Gtro labeled cytosolic calf uterine estrogen receptor. Sensitivity of’ RE? and RE, to p~rida~a~ S-phosphate We have previously shown that the lysine modifying reagent pyridoxal S-phosphate (PLP) inhibits receptor activation (nuclear binding), 4s to 5s transformation and the transition of RE, from a state with fast into a state with slow E, dissociation [S, 22.231. More recently [lo] we have shown that addition of PLP to previously heat-activated and heat transformed RE, (6s in borate buffer) reduces the affinity of RE? such that EZ dissociates rapidly, as if the receptor was converted to the non-transformed state. This treatment, however, does not alter the sedimentation characteristic of heat-transformed RE,, which still sediments as the 6s dimer [22,24]. The experiments presented below were performed to determine whether RE, and RE, have different sensitivity to PLP. Such differences might provide some insight into the conformational state of the estrogen binding site when occupied by E2 or E,. Figure 3, panel A, illustrates the effects of 5 and 10mM PLP on E, dissociation from the heattransformed RE, dimer. In the absence of this reagent [3H]Ez dissociates slowly from the estrogen receptor with a t,,>of 130 min, a value similar to that obtained in the experiment in Fig. I. Addition of PLP 12 min after E, dissociation at 28°C had begun, causes a sudden decrease in RE, affinity, as reflected by the very rapid dissociation of E,: this effect is dose dependent, as indicated by the larger fraction of RE, modified to the fast dissociating state at 10 mM PLP. PLP Addddlfion

CY ul

From the experimental points shown in Fig. 3A. we conclude that at approximately 30min all those receptors which had been modified by PLP lost their hgand. The remaining slow dissociating component represents RE, which had not been modified due to insufficient concentrations of PLP. A dose-response effect of PLP on receptor transformation has previously been demonstrated [22]. Furthermore, the effects of PLP have been shown [lo, 231 to be reversible upon transschiffization through lysine addition; thus, the increased rates of E? dissociation are not due to receptor instability nor to receptor denaturation induced by PLP treatment. The effects of PLP on R[‘H]E, are shown in Fig. 3, panel B. [‘H]E, dissociation from the heat-transformed R[‘H]E, complex (tlfz = 38 min) is significantly faster than that of E2 from RE2. Upon addition of PLP the receptor is modified to a state of even lower affinity for E,, as indicated by the dose dependent increase of the fast dissociating entity. Comparison of the PLP effects on RE2 and RE, can only be made by taking into account that these two hormone receptor complexes release the ligand at different rates. Thus, at the time point of PLP addition to RE,, 80% of the hormone-receptor complexes were still undissociated; at 36 min, binding had further decreased to 54’j/, in the absence of PLP and to 25% in the presence of 5 mM PLP; thus the net effect of PLP was to increase dissociation of RE, complexes by 29 percentage points. In comparison, 930,; of RE, complexes were still undissociated at the time of PLP addition; the fraction of receptors modified by 5 mM PLP was completely dissociated at 24 min, at which time binding had decreased to 65”1,. Binding in the control incubation was SSU/;at 24 mitt; rc

(A)

-*

PLP

(

Addifmo

B)

+10mMf?_Po\ \

0

PLP

~~t3HlE3+ IO m&l I

I

I

I

I

30

60

90

(20

30

Time

60

90

PLP

420

(min)

Fig. 3. Effects of pyridoxal S-phosphate (PLP) on [-‘H]estrogendissociation from heat-transformed cytosohc SS R-[3Hfestrogen complexes. Calf uterine cytosol prepared in BSM buffer was incubated with [“H]E, (10nM) or [3H]E3(20 nM); experimental conditions were as described in Fig. 1. Samples were subjected to exchange with DES at 28°C; after 12min freshly prepared solutions of PLP were added to the R13H]E2 (panel A) and Rk’H]$ (pane1 B). The final concentration of PLP in the incubation were 5 and 10 mM. At the indicated times aliquots (duplicates) were removed and assayed for bound radioactivity with HAP at 0°C. As described in Fig. I, data represent only specifically bound [‘HIsteroid after correction for receptor degradation (less than 10% in 120 min at WC). All data have been normalized to eliminate the effect of receptor dilution which took place upon PLP addition (20”/, dilution),

1043

Estriol and estradiol interactions with estrogen receptor therefore, the net effect of PLP was a 22 percentage points increase in RE2 dissociation. Clearly, a more detailed analysis with many more time points after PLP addition is necessary before one can make any definite conclusion. From this preliminary experiment, however, it appears that the 5s REI may be less sensitive to PLP than the SS RE, complex, suggesting that the binding site of RE, is “locked” into a conformation in which the lysine residues are not as easily accessible to PLP. In this experiment (data not shown) PLP had no effect on the sedimentation characteristics of heattransformed RE, or RE,, which remained unaltered at 6S (sedimentation through 0.4 M KC1 sucrose gradients prepared in borate buffer with and without PLP); this is in agreement with previous reports [lo, 22,241. Does estriol affect the temperature accelerated conversion of RE,from a state withfast into a state with slow E, dissociation?

The following scheme illustrates the experimental approach taken to answer the above question: Low R+[3H]E2

(1)

0.4M KCl_ o"c

R+[3H]E2

0.4M KCl, OOC

R+[3H]E3

(2)

*

0.4M KCl_ o"c

(3) a)

R+13HlE2

b) Rtunlabeled E3

High affinity state

4s R[3H]E,

45 R[3H]E,

/

28OC

5s [3~]E,R--R[3H]E, * transformation

0.4M KC1 o"c

R+[3H]E3

trogen binding site as observed with estradiol K_, = 4 x 10m3min-‘; vs (K_,= 15 x 10-3 Fig. 4). In reaction 3 a parallel sample of the fast dissociating 4S R[‘HjE* formed in reaction 1 was mixed at 0°C with a 19-fold excess of nonradioactive 4s RE3 and then placed at 28°C in the presence of 4 x 10e6 M unlabeled E,. In this experiment [‘H]E, dissociation was measured by exchange with E, rather than Ez, to avoid formation of nonradioactive RE2 which could then have competed with RE3 for dimerdimer formation with R[3H]E,. The results of this mixing experiment (Fig. 4, dots) indicate that the conversion of 4s R[‘H]E, from a state with fast into a state with slow estradiol dissociation took place at the same rate (i.e. within the first 10min of incubation at 28°C) as in the absence (open squares) of an excess of unlabeled RE,. Sucrose gradient analysis indicated that also 4s to 5s transformation of R[3H]E, was unimpeded by the presence of the nonradioactive RE, (data not shown). We also performed the reverse experiment, to see whether R[3H]E3 4s

0.4MlVol. o"c -19

o"c

4s

R’3H1Ez>

*8°Ctranrformatio5~

[3~]~3R--~[3~]~3

45 R[3H]E3

45

Vol. 45

In Reaction 1 [3H]E2 binds to the 4s monomer (high salt buffer) at 0°C to form the low affinity 4s R[3H]E, complex. The sample is then heated in the presence of 4 x 10m6M unlabeled estriol. As shown in Fig. 4 (open squares) thermal energy induces the conversion of R[3H]E, from a state with fast into a state with slow [3H]E2 dissociation. The kinetic parameters are similar to those reported above (Figs 1 and 3). During heating 4s to SS transformation occurred (data not shown). Reaction 2 represents the interaction of [‘H]E, with the receptor under conditions which are identical to those of reaction 1. As shown in Fig. 4 (open circles), thermal energy induces the transition of RE, into a state of higher affiinity; although 4s to 5s transformation takes place unimpeded (data not shown) the transformed 5s dimer does not provide the same high affinity interaction between estriol and the es-

monomers when mixed with nonradioactive REI may form 5s homodimers with [3H]E3 affinity higher than that of R[3H]E3 homodimers. Again, the rates of PH]E, dissociation were not altered by the addition of REX (data not shown). DISCUSSION

In order to explain the post-coital contraceptive activity of estriol and related steroids we proposed [3] that substances capable of interfering with the formation of cytosol receptor-estradiol complex could decrease the nuclear concentration of hormone-receptor complex, thereby preventing the full expression of estradiol action. This concept was further verified in studies on the effects of E,, E, and their admixtures on uterine growth and on the induction of progesterone receptors [25]. Further

1044

R. ENZIO MUELLER et al.

10 0

’

’

10

20

’ 30

8 40 Time

’

’

’

’

’

’

’

50

60

70

60

90

100

110

(mtn)

Fig. 4. Effect of nonradioactive 4S RR, complexes on the heat-induced transition of R[‘H]& into a state with slow [3H]E2 dissociation. Calf uterine cytosol was incubated in TD (0.4 M KCI) buffer (6-8 mg protein/ml; 1.50nM estrogen receptor) at 0°C with 10 nM [‘H]E, or 20 nM nonradioactive E,. A control incubation with [rH]Er (20nM) was also performed (open circles). Aliquotes of R[3H]E2 were mixed with 1: 10 volume of unlabeled E, (4 x 10-O) and placed at 28’C to induce [3H]E, exchange with unlabeled E, (control dissociation curve, open squares). Aliquots of R[3H]E, (estriol control) were similarly exchanged at 28°C (open circles). A third incubation (solid circles) consisted of I vol of R[)H]E, mixed at 0°C with 19 vol of cytosol labeled with nonradioactive E,; to this mixture I : IO volume of unlabeled E, (4 x IO --6M final concentration) was added. [‘H]E, and [3H]E, dissociation (IL ,) from the low affinity receptor state calculated as described elsewhere [9].

studies [ 141 showed that E, dissociates more rapidly than E, from non-transformed and heat-transformed receptor. This is consistent with the finding [2, 161 that the weak estrogenic action (uterine growth promotion) of E, was related to the rapid loss of nuclear bound RE, complexes. The data presented here confirm the reports [14] that E, dissociates at a rate faster than that of E, from unheated (8s state) and heat-transformed (5s state) receptor. Although in these experiments we have not measured the actual rates by which estrogen-receptor complexes change from the fast into the slow dissociating state, it appears that within IOmin at 28°C both RE, and RE, have completed this transition {Fig. 1). To our knowledge, the data in Fig. 2 represent the first report that the change in receptor affinity for its ligand so far only observed in vitro is a physiological event, since E, and E3 dissociate at similarly slow rates from the nuclear-bound and transformed 5s receptor formed in vim. We have also obtained

preliminary data suggesting that the fraction of ho~one-receptor complexes remaining in the cytoSol in vivo represents a low-affinity receptor state. In conjunction with our previous findings{261 that in uterine cell suspensions one can observe a 4S-5s transformation with rapid nuclear accumulation of the 5s RE2 species, these data suggest that the slow dissociating 5s estrogen dimer is the biologically active endproduct of receptor estrogen interaction. Whether these changes in receptor affinity towards the ligand and the formation of 5s dimers take place within the cytosol or nucleus in vivo remains to be established, although recent evidence has been obtained suggesting that in z+tro these reactions can occur within the nucieus{8]. Intranuclear 4s to 5s transformation has previously been reported by Siiteri and his collaborators (271. It has been suggested [4] that the nuclear 5s receptor observed upon in vivo administration of estradiol may represent an artifact formed upon tissue homogenization and KC1 extraction of nuclear RE,, whereby the 5s state would represent an aggregate of 4s RE2 complexes with other proteins or nucleic acids. Although a 4s RE2 nuclear binding species (activated state) is well documented [4,6,7,27, 281, we suggest that the 5s RE2 is the biological endproduct, since nuclei extracted without KC1 but with pyridoxal 5’-phosphate, an inhibitor of 4s to 5s transformation, also yield the transformed 5s RE, [S, 231. Figure 5 depicts two possible models to explain the sequence of hormone-induced changes which promote the conversion of RE2 from the low affinity into the high affinity state. Mode1 A is based on the

Fig. 5. Models for the molecular mechanism of the transition of RR, from a state with fast into a state with slow E, dissociation rates. R, RE,--low affinity 4s monomers, R’Er-high affinity monomers. Details are given in Discussion.

Estriol and &radio1 interactions with estrogen receptor proposal [9] that 4S-5s transformation is required for the formation of the slow dissociating RE, state. This model is supported by the observation that reagents which prevent 45-55 transfo~at~on or disrupt 5s dimers are also inhibitors of the formation of the slow dissociating RE2. Nowever, as discussed elsewhere 16, IO], those reagents may have a direct effect on the estrogen binding site of each 45 monomer and it may be only a coincidence that they alsa interfere with 45 to 5s transformation. Indeed several examples exist suggesting that 5s dimers can be present in the low aIIinity state: for instance, receptor complexed to estrone [14] or tamoxifen [29] forms 5S dimers in the absence of a clear transition from a state with fast into a state with slow ligand dissocation. Another argument in support of model A is found in the dependence of the rates of RE, transition into the high a&b&y state upon the intial concentration of 4s RE2 monomers. These data were obtained [30] in 0.4M KC1 buffer and such receptor concentration dependence was no longer observed at more physiological salt concentrations (0.15 M KCl). Still, such data suggest that the change in RE, affinity for E, is a bimolecular reaction (4s to 5s transfo~ation~. This point wift be discussed below, Model B is based on the proposal that the transition of RE2 into the high affinity state (symbolized as R’) is due to a conformational change within the 4s monomers while 45 to .5S transfo~at~on is a separate reaction. The evidence for this model and is based on the observation [lo] that 8s (low salt buffers) or 4s (high salt buffers) RE2 immobilized on hydroxylapatite undergoes the temperature induced transition into a slow dissociating state but, upon elution from HAP, it is still in the 4s state as revealed by sucrose gradient analysis. Similar data have been presented by Sakai and Gorski[l I]. Not shown in this model are the individual rate constants. Furthermore, we have insufficient information to state whether in models A or B the slow dissociation of Ez from the high affinity transformed E,R’--R’E, dimer is, indeed, due to Eiz dissociation from this entity, or if a slow reversaf into the low affinity RE, state must occur before E, can dissociate. Both pathways would yield the observed slow rates of E, dissociation from heat-transfo~ed 5s RE2. Based on the in viva studies (Fig. 2) we presume that the equilibrium between high affinity 4s R’I$ and 5s E,R’--R’E, is to the right, with the 5s R’E, being the therm~~amic~Iy stable endproduct. Thus, once a 4S high affinity R’E, complex is formed, subsequent 4S-5S transformation prevents the reverse reaction to the low affinity 4s state. This may also explain why in 0.4M KC1 the transition of RE, from a low afhnity into a high afhnity state appeared [30] to follow 2nd order kinetics, i.e. be equivalent to 4S to 5s dimerization; the higher the initial concentration of 45 monomers (low affinity state) the greater the chance for the intermediate 4s R’E, monomers (high affinity state) which are formed

1045

upon heating to further react with each other to form stable 5S high affinity dimers. Conversely, at low receptor concentration, in high salt, a larger fraction of 45 R’EZ would revert to the low affinity state and release the ligand. The data presented in Fig. 4 further suggest that the biphasic Ez dissociation rates represent a conformational change of the estrogen binding domain and that this reaction is independent of 4SSS transformation. If subunit interaction was necessary to establish the slow dissociating RE, state, and knowing that E3 dissociates rapidly from 55 RE, dimers one would expect that by mixing Rf%]E, with a 19-fold excess of nonradioactive RE, of equal protein concentration the amplitude of the fast dissociating component would significantly increase, since in the second order reaction 2 (4S)-+ 5s the rate of RE, dimerization would decrease 400-fold when the 4s monomer concentration is diluted 20-fold. The absence of such effects suggests that upon heating each 4s RE, low afhnity monomer undergoes the transition into the high atlinity R’E, monomer. Futhermore, the presence of a 1Pfold excess RE, during heating will result in transfo~ation to yield R’[3H&--R’E, heter~ime~~ this prevents the reverse reaction 4s R’[3H]E2-t4SR[3H]E, (high affinity to low affinity reversal). Our proposal (Fig. 5B) that the transition of RE2 into a state with slower I$ association rates represents a conformational change of the 45 monomer is consistent with the finding that aiso the glucocorticoid [31] and the progesterone [32] receptors undergo such temperature induced affinity changes in the absence of 4C5S transformation. REFERENCES I. Anderson J. N,, Clark J. H. and peck Jr E. J.: The relationship between nuclear receptor estrogen binding and uterotrophic responses. B&hem. biiphys. Res. Commun. 48 (1972) 1460-1468. 2. Anderson J. N., Peek E. J. Jr and Clark J. H.: Bstrogen-

3. 4. 5. 6.

7.

8.

induced uterine responses and growth relationship to receptor-estrogen binding by uterine nuclei. Endocrinology % (1975) 160- 167. Mtiller R. E. and Wotiz H. H.: Post-coital contraceptive activity and estrogen receptor binding afinity of phemolic steroids Endocrin~io~y 100 (1977)5 13-5 19. Milgrom E.: Biochemical action of hormones (Edited by G. Litwack) Academic Press. New York (1981) 465-492. Meher R. E,, Tmish A. M. and Wotiz H. H. (1980) Effects of pyridoxal s’-phosphate on uterine estrogen receptor. 3: biol. Chetn.~255 ~1980) 4062-4064. Miiller R. E.. Mrabet N. I’.. Traish A. M. and Wotiz H. H.: The role of arginyl residues in estrogen receptor activation and transformation. J. biol. Chem. (198.3) Zss 11582-l1589. Bailty A., LeFevre B., Savouret J-F. and Mitgrom E. J.: Activation and changes in sad~mentation properties of steroid receptors. Biol. Chem. 255 (1980) 27292734. Miiller R. E., Traish A. M., Wotiz H. H.: Estrogen receptor activation precedes transformation. J. biol. Chem. 258 (1983)

9227-9236.

IQ46

R. ENZIO MtiLtER

9. Weichman B. M. and Notides A. C.: Eslradiol-binding kinetics of the activated and non-activated estrogen receptor. J. biol. Chen?. 252 (1977) 8856-8862. 10. Miller R. E., Hirota T., Traish A. M. and Wotiz H. H.: RE2 transition from a state with fast info a state with slow E, dissociation is independent of 45 to 5s transformation. ~npubiis~ed data It. Sakai D. and Gorski J.: Estrogen receptor transformation in the absence of 4s to 5s conversion. The Endocrine Society Qfh ~~~~1 ~eetjr~g, (1983) Abstract 1042. I2 Notides A. C. and Nielsen S.: The moiecuIar mechanism of the in ttirro 4s to 5s transfo~ati~n of the uterine estrogen receptor. J. bioi. Chem. 249 (1974) 1S661s73._ 13. Notides A. C.: In Receptors aad Hormone Action. Vol. 2 fEdited bv B. W. O’Mallev and L. Birnbaumer) ’ (1078) pp. 33-61. Academic Press, New York. 14. Wcichman B. M. and Notides A. C.: Estrogen receptor activation and the dissociation kinetics of estradioi, estrioi and estrone. ~~docr~~o~ogy106 (19SO)434439. 15 Hisaw F. L. Jr: Comparative effectiveness of estrogens an fluid inhibition and growth of the rat uterus. &docrinolffgy 64 (1959) 276289. 16. Clark J. H., Paszko Z. and Peck E. J. Jr: Nuclear binding and retention of the estrogen receptor complex: relation to the agonistic and antagonistic properties of estriol. Endocrin&g~ 100 (1977) 91-94. 17. Sasson S. and Notides A. C.: The inhibition of the estrogen receptor’s positive cooperative [‘Hlestradiol binding by the antagonist, clomiphene. J. biol. Chem. 257 (1982) 11540-l 1545. 18. Muller R. E., Traish A. M. and Wotiz H. H.: lnteraction of receptor estrogen complex with uterine nuclei. J. biol. Chem. 252 (1977) 8206-8211. 19 cast-~elpomme M., Fies J. and Erdos T.: K~teractio~s entre I’oestradiol et des sites recepteurs uterines. Errr. J. Biachem. 17 (1970) 4255432. 20 Sanborn 3. M., Rao B. R. and Korenman S. G.: Interactions of 17B estradiol and its specific uterine receptor. Evidence for complex kinetic and equil;brium behaviour. ~iochern~~tr~10 0971) 49.554962. 21. DeBoer W. and Notides A. e.: D&ciation kinetics of

et

ui.

the nuclear estrogen receptor. Biochemisq 20 (1981) 1290-1294. 22. Traish A. M., Miiller R. E. and Wotiz H. H.: Effects of pyridoxal 5’-phosphate on uterine estrogen receptor. J. biol. Chem. 255 (I 980) 4068-4072. 23. Miiller R. E., Traish A. M. and Wotiz H. H.: In Hormones and Cujlcer (Edited by S. Jacobetli, R. J. B. King, H. R. Lindver and M. E. Lippman). Raven Press, New York (1980) 241-254. 24. Muldoon T. G. and Cidlowski J. A.: Specific rn~i~cat~on of rat uterine estrogen receptor by pyridoxal S-uhosphatc. J. hiof. Ckm. 255 f1980) _ , 3lO&3107: . 25. Wotiz H. H., Chattoraj S. C,, Kudisch M. and Mu&r R. E.: fmpeding estrogens and the etiology of breast cancer. Cancer Res. 38 (1978) 40124020. 26. Traish A. M., Miiller R. E. and Wotiz H. H.: Comparison of formation, activation and nuclear translocation of receptor estradiol (RE,) complex at 0°C and 37°C in intact uterine cells. J. biol. Chem. 254 (1979) 656~6563. 27. Siiteri P. K., Schwarz B. E., Moriyama J., Ashby R., Linkie D. and MacDonald P. C.: Estrogen binding in the rat and human. A&. fxp. M&i. Biol. 36 (1973) 97-112. 28. Sato B., Nishizawa Y., Noma K., Mat~moto K. and Yamamoro Y.: Estrogen inde~ndent nuclear binding of receptor protein of rat uterine cytosoi by removal of tow molecular weight inhibitor. ~n~~~r~~~~Qg~i 104 (1979) 14761479. 29. Rochefort H. and Borgna J. L.: Differences between estrogen receptor activation by estrogen and antiestrogen. Nature 292 (1981) 257-259. 30. Weichman B. M. and Notides A. C.: Analysis of estrogen receptor activation by its [‘Hlestradiol dissociation kinetics. ~jochern~,~rry18 (1979) 220-225. 31. McBlain W. A., Toft D. O., Shyamala G.: Transformation of mammary cytoplasmic gfucocorticoid receptor under all-free conditions. Bioe~~emjstry20 (1981) 679c-6798. 32. Wolfson A., Mester J., Yang C. R. and Baulieu E-E.: Nonactivated form of the progesterone receptor from chick oviduct: charact~ri~tion. Biochmr, b~o~h~~. Res. &o~~rn~~.9-S (I 980) 1577--I 584.