The Molecular Mechanism of the in Vitro 4 S to 5 S Transformation of the Uterine Estrogen Receptor*

SUMMARY The rat uterine cytosol contains two estrogen-binding proteins (EBP) referred to as the ‘4 S” and “5 S” EBP. The 5 S EBP was produced by incubating the uterine cytosol-[aH]estradiol mixture at 28’ for 30 min to promote the “4 S to 5 S transformation” as described by Jensen et al. ((1971) Biochem. Sot. Symp. 32,133). These two EBP species were characterized in high ionic strength buffers using sucrose gradient centrifugation and Sephadex G-200 gel chromatography. With these analytical methods the sedimentation coefficient, molecular Stokes radii, and molecular weights were determined. In the presence of 0.4 M KC1 at pH 7.4 the 4 S EBP has a sedimentation coefficient of 4.2 f 0.04 S, a molecular Stokes radius of 44.0 f 0.4 A, and a molecular weight of 76,200. While in the identical buffers the 5 S EBP has a sedimentation coefficient of 5.5 f 0.02 S, a molecular Stokes radius of 58.5 + 0.5 A, and a molecular weight of 132,700. In the presence of 0.4 M KC1 and 3 M urea at pH 7.4 the 4 S EBP shows a decrease in


The Molecular
Mechanism of the in Vitro 4 S to 5 S Transformation of the Uterine Estrogen Receptor* (Received for publication, August 30, 1973) ANGELO C. NOTIDES

AND SUSAN NIELSEN~
From the Department of Pharmacology and Toxicology, University of Rochester School of Medicine and Dentistry, Rochester, New York l&Q?

SUMMARY
The rat uterine cytosol contains two estrogen-binding proteins (EBP) referred to as the '4 S" and "5 S" EBP. The 5 S EBP was produced by incubating the uterine cytosol- [aH]estradiol mixture at 28' for 30 min to promote the "4 S to 5 S transformation" as described by Jensen et al. ((1971) Biochem. Sot. Symp. 32,133).
These two EBP species were characterized in high ionic strength buffers using sucrose gradient centrifugation and Sephadex G-200 gel chromatography.
With these analytical methods the sedimentation coefficient, molecular Stokes radii, and molecular weights were determined.
In the presence of 0.4 M KC1 at pH 7.4 the 4 S EBP has a sedimentation coefficient of 4.2 f 0.04 S, a molecular Stokes radius of 44.0 f 0.4 A, and a molecular weight of 76,200. While in the identical buffers the 5 S EBP has a sedimentation coefficient of 5.5 f 0.02 S, a molecular Stokes radius of 58.5 + 0.5 A, and a molecular weight of 132,700.
In the presence of 0.4 M KC1 and 3 M urea at pH 7.4 the 4 S EBP shows a decrease in its sedimentation to 3.6 & 0.04 S, but an increase in its molecular Stokes radius to 53.8 & 0.9 A. The estimated molecular weight is 79,900. The 5 S EBP has a sedimentation coefficient of 4.6 f 0.9 S and a molecular Stokes radius of 70.6 =t 1.0 A. The molecular weight is 133,900.
In addition, sucrose gradient centrifugation analyses, in the presence of 0.4 M KCl, 3 M urea at pH 7.4 showed that a moderate fraction (25 to 50%) of the 5 S EBP is dissociated to the 4 S EBP.
In the presence of 0.4 M KC1 at pH 6.8 the 4 S and 5 S EBPs sediment at 4.7 =t 0.04 and 5.6 f 0.05 S, respectively.
The 5 S EBP also shows a tendency to revert to the 4 S EBP.
Consequently, it was observed that acombination of 0.4 M KC1 and 3 M urea at pH 6.8 is extremely effective, in contrast to using each of these reagents alone, in dissociating the 5 S EBP to the 4 S EBP without causing a loss of [3H]estradiol binding. The dissociation process is first order with a half-life of approximately 5 hours. These results indicate that the in vitro "4 S to 5 S transformation" is an association of the 4 S EBP, having a molecular weight of -80,000, with a second component or subunit of -50,000 to form the 5 S EBP (~130,000). Concurrent with the association process, the 4 S and 5 S EBPs are also capable of marked conformational changes that were discerned when the EBPs were compared to the protein standards under identical conditions.
The protein standards used were myoglobin, trypsin, ovalbumin, serum albumin, alkaline phosphatase, aldolase, y-globulin, and ferritin. The conformational changes of the EBPs were shown by their sedimentation coefficients and molecular Stokes radii varying in a reciprocating manner while their molecular weights remained constant.
The target tissues of the estrogenic hormones contain within their cytoplasm specific estrogen-binding proteins (EBPs) referred to as "recept,ors" (1). The association of estradiol with the receptor is accompanied by a temperature-dependent redistribution of the estrogen receptor from the cytoplasm to the nucleus (2)(3)(4)(5).
Subsequently, with the localization of the estrogen recept.or complex in the nucleus, enhanced nuclear biosynthetic activities are observed (6-9).
However, the exact molecular activity of the estrogen receptor has not been shown. Numerous investigators have shown that the cytoplasmic estrogen receptor sediments in sucrose gradients at approximately 4 S in the presence of 0.4 M KCl, while the estrogen receptor isolated from the nuclells sediments at approximately 5 S (10-12). Several studies have centered upon resolving the molecular basis for the differences in the sedimentation behavior of the cytoplasmic and nuclear forms of the estrogen receptor based upon the supposition that the molecular mechanism for the estrogen receptor's action may then be resolved. Jensen et al. (13,14) have reported that a cell-free "4 S to 5 S transformation" of the estrogen receptor produces a 5 S EBPr similar to the isolated nuclear 5 S eskogen receptor.
Based upon sucrose gradient analysis and gel chromatography 1 The abbreviations used are: EBP, estrogen-binding protein; TEK buffer, 40 mM Tris-2 mM Naz EDTA-0.4 M KCl. nH 7.4: Heaes. N -2 -hydrdxyethylpiperazine -XT' -2 -ethanesulfo& acid; HEK buffer, 40 mM Hepes-2 mM Nas EDTA-0.4 M KCl, pH 6.8. studies of the 4 S EBP and the 5 S EBP produced by the in vitro the above Triton X-lOO-toluene scintillation fluid for each l-ml _ "4 S to 5 S transformation" procedure, we report in this paper the fraction* molecular characteristicsof, and relationship between, the 4 S and Dialysis and concentration of [3H]estradiol peaks from the gel 5 S EBPs. The data are consistent with a model indicating that columns were carried out with a Diaflo ultrafiltration apparatus and an XM-50 membrane maintained at 0". the 4 S EBP (-80,000) associates with a second component of -50,000 to form the 5 S EBP with a molecular weight of approximately 130,000.

EXPERIMENTAL PROCEDURES
Preparation oj Rat Uterine Cytosol and J$ S to 6 S EBP Transformation-uteri from 20-to 24-day-old Holtaman rats were excised and collected in 40 mM Tris, pH 7.4, 0". The uteri were rinsed in a fresh quantity of the same buffer and homogenized (five uteri per ml of 40 mM Tris, pH 7.4, 0") with three 10-s bursts from a Polytron PT-10 (Brinkmann Instruments) at a power setting of 4, at 50-s intervals.
The homogenate was centrifuged at 220,000 X g for 30 min to obtain a supernatant fraction that is referred to as the "cytosol." The protein concentration of the cytosol was determined by the method of Lowry el al. (23). The [aH]estradiol was immediately added to the cytosol, which was then incubated at 0" for the time noted in each figure. Aliquots of the [3H]estradiol-cytosol were further incubated at 28 f 0.5" for 30 min to promote the 4 S to 5 S EBP transformation of the estrogen receptor as described previously (13).
Sucrose Gradient Centrifugation (.24)-Linear sucrose gradients (3.8 ml) were prepared with a Buchler Gradient Former. The concentrations of the sucrose gradients, the buffers used, and the centrifugation conditions are noted with each figure. The gradients were allowed to stand 3 to 5 hours at 4O before layering a 0.2-ml sample onto the gradient.
All samples contained myoglobin (2.0 S) and alkaline phosphatase (6.2 S) as internal sedimentation markers.
Frequently, myoglobin, alkaline phosphatase, and bovine serum albumin (4.6 S) were layered onto another gradient as an additional control.
The samples were centrifuged in a Beckman LZ-65 ultracentrifuge using an SW 56 rotor at 4".
Following centrifugation, the gradients were fractionated with an Isco Gradient Fractionator; 40 O.l-ml fractions were collected. The positions of the sedimentation markers were noted by their absorbance at 280 nm with an Isco UA-2 Analyzer.
The position of alkaline phosphatase was also assessed by its enzymatic activity (25). Following the addition of 5 ml of scintillation fluid (toluene-Triton X-lOO-2,5-diphenyloxazole (PPO)-1,4-bis [2-(5phenyloxazolyl)]benzene (POPOP) in the proportions 666:333: 5.5:0.12, v/v/w/w), the radioactivity was measured in a Nuclear Chicago Mark II spectrometer; the counting efficiency was 40 to 43yc. The sedimentation coefficient was estimated for the apex position of the [aH]estradiol profile by extrapolating from the linear relationship between fraction number and the sedimentation coefficient of the standards.
Gel Chromatography-Sephadex G-200 was swollen in TEK buffer, TEK-3 M urea buffer, or HEK buffer for 1 to 4 weeks at room temperature and the fines removed by aspiration.
The gel was packed into columns, 2.5 X 45 or 2 X 40 cm, extensively washed with buffer at a flow rate of 14 to 17 ml per hour, and maintained at 4'. The distribution coefficients (Kd) were measured according to the equation: where V,, the void volume, was determined with blue dextran; Vi, the internal volume within the gel, was determined with potassium chromate; and V, was the elution volume at the apex of each protein peak, monitored by absorbance at 280 nm or the PHIestradiol radioactivity peak. The data were plotted as the elution volume to void volume ratio (V,:V,). The columns were also calibrated according to a method previously described for molecular Stokes radius (26). The molecular Stokes radii of the protein standards have been published (27,28). The radioactivity was measured with an efficiency of 25 to 35% using 10 ml of Molecular Weight Determinations-The method used for determining the molecular weights was originally described by Siegel and Monty (28), in which the molecular Stokes radius (a) determined by gel chromatography, and the sedimentation coefficient (S) determined by sucrose gradient analysis are used together to estimate accurately the molecular weight (M) and the frictional ratio (f:fa) of a protein with the following relationships.

M=-
(1 -VP) (2) f:fo = a [ 3M(v?~6,p)ll'a where n is the viscosity, p is the density of the solvent, and N is Avogadro's number. The molecular weights and frictional ratios were calculated by assuming a partial specific volume (6) of 0.725 cm3 per g and a solvation factor (6) of 0.2 g of solvent per g of protein; which are typical values (29,30). -_ -_ Muterials-The 178-16.7-3Hlestradiol (48 Ci ner mmole) was obtained from New England Nuclear Carp: Theradiochemical purity was verified by thin layer chromatography.
The proteins used as standards for molecular Stokes radii and sedimentation coefficient determinations were the best available grades and were obtained as follows: human r-globulin, ovalbumin, sperm whale myoglobin, and Escherichia coli alkaline phosphatase were from Sigma; rabbit muscle aldolase and bovine pancreas trypsin were from the Worthington Corp.; bovine serum albumin and horse liver ferritin were from Miles Laboratories, Inc. The sucrose, urea, and Tris were ultrapure grades from Schware-Mann.
All other reagents used were analytical grade. The Sephadex G-200 and blue dextran were from Pharmacia.

Transformation of Estrogen Receptor from 4 S to 5 S EBP-In
sucrose gradient analysis, the rat uterine estrogen receptor appeared as two estrogen-binding proteins (EBPs) : 4.2 f 0.04 S, referred to as the 4 S EBP, and 5.5 f 0.02 S, referred to as the 5 S EBP.
The quantity of the 4 S EBP was always greater than that of the 5 S EBP.
Occasionally only the 4 S EBP was seen in the cytosol-[3H]estradiol mixture that was incubated 30 or 60 min at 0' prior to sucrose gradient centrifugation.
Incubation of the cytosol-[aH]estradiol mixture at 0" for 30 min, followed by 28" for 30 min, reduced the quantity of the 4 S EBP and increased that of the 5 S EBP to 90 to 100% of the total [3H]estradiol-binding capacity present (Fig. 1A). No significant loss of the [%estradiol-binding capacity was observed during the in vitro transformation procedure.
The addition of a 200-fold excess of unlabeled estradiol to the cytosol-[zH]estradiol mixture (i.e. 4 S EBP) immediately before the 28" incubation had little effect on the appearance of the labeled 5 S EBP.
A small decrease (10 to 30%) in the total [SH]estradiol-binding capacity was observed, but this was presumably due to an exchange of [3H]estradiol with the unlabeled estradiol during the 28" incubation.
This suggesk that most of the [3H]estradiol was still associated with the 4 S EBP (or a modified form of the 4 S EBP) and that the [SH]estradiol did not dissociate from the 4 S EBP to reassociate during the 28" incubation with a "de novo 5 S EBP." The selective transfer of the [aH]estradiol from the 4 S to the 5 S EBP during incubation was also excluded by the observation that the binding affinity of the 5 S EBP was not greater than that of the 4 S EBP. 2 The similarity in the binding affinities of the different forms of the EBPs has been reported previously (12). (3 nM) were incubated for 60 min at 0" ( l ), or 30 min at 0" following by 30 min at 28" (0 ). The samples were then subjected to sucrose gradient centrifugation for 22 hours at 394,000 X gmax on a 10 to 3070 sucrose gradient with TEK buffer (A), or a 5 to 2070 sucrose gradient with TEK-3 M urea buffer (B).
ELUTION VOLUME voiDKtuME A l-ml aliquot of uterine cytosol was incubated with [aH]estradiol (5 nM) for 60 min at 0" (A). A second aliquot of the cytosol-[3H]estradio1 mixture was incubated for 30 min at 0" followed by 30 min at 28" (B).
The samples were then chromatographed on a Sephadex G-200 column (2.5 X 45 cm) that was equilibrated with TEK buffer.
The flow rate was 14 to 17 ml per hour.
Sedimentation Behavior in Urea-Sucrose gradient analysis of the two forms of the estrogen receptor (4 S EBP and 5 S EBP) in the presence of urea showed decreased sedimentation coefficients, while gel chromatography using identical urea solutions showed an increase in their molecular radii; urea has altered the conformations but not the molecular weights.
Although the two EBPs showed a gradual reduction in their sedimentation coefficients with increasing urea concentration, the larger 5 S EBP in addition showed a tendency to revert to the smaller 4 S EBP species. In sucrose gradients containing TEK-1 M urea buffer the two EBPs were 3.8 f 0.03 S and 5.3 + 0.08 S with insignificant reversion of the 5 S EBP to the smaller species. Using TEKS M urea buffer the EBPs were 3.7 + 0.04 S and 5.0 f 0.05 S with slight (~10 to 20%) reversion of the larger EBP to the smaller EBP.
Using TEK-3 M urea buffer the EBPs were 3.6 f 0.04 S and 4.6 + 0.09 S with moderate (-25 to 50%) reversion of the larger EBP to the smaller EBP (Fig. 1B) producible elution positions after Sephadex G-200 chromatography with TEK buffer. The cytosol-[3H]estradiol mixture incubated 30 to 60 min at 0" prior to gel chromatography contained an EBP with a V, : Vu of 1.59 (Fig, 2, Curve A). The molecular Stokes radii of the EBPs were estimated by comparing their elution parameters on Sephadex G-200 with those of other standard proteins whose molecular Stokes radii are known.
Calibration of the Sephadex G-200 columns by the method of Porath (26) indicated that the molecular Stokes radius of the labeled 4 S EBP was 44.0 + 0.4 A (Fig. 3A).
Incubation of the cytoso-[3H]estradiol mixture at 0" for 30 min followed by 28" for 30 min resulted in an EBP with a greater molecular Stokes radius; the V, : VU was 1.30 (Fig. 2, Curve B) and the molecular radius was 58.5 + 0.5 A (Fig. 3C). To insure that the estrogen receptor was either predominantly the 4 S or 5 S EBP species, aliquots of the cytosol-[3H]estradio1 subjected to Sephadex G-200 chromatography were analyzed concurrently by sucrose gradient centrifugation. Analyses by sucrose gradient centrifugation of the [3H]estradiol peaks eluted from the Sephadex G-200 indicated that their sedimentation coefficients were identical with those of the EI3Ps applied to the column.
In two separate experiments the [3Hlestradiol peaks from the Sephadex G-ZOO columns were collected, concentrated by ultrafiltration, and subjected to sucrose gradient analyses.
The EBP eluted with a molecular Stokes radius corresponding to 44 A had sedimentation coefficients of 4.3 and 4.0 S, whereas the [3H]estradiol peak eluted with a molecular Stokes radius corresponding to 58.5 A had sedimentation coefficients of 5.5 and 5.4 S.
Incubation of the cytosol at 28' for 30 min in the absence of the estradiol, followed by [3H]estradiol at 0" for 30 min, showed insignificant transformation of 4 S to 5 S EBP during warming, as previously reported by Jensen et al. (13). Sephadex G-200 chromatography provided ident'ical results: with prior incubation at 28" before the addition of the [3H]estradiol, the estrogen receptor remained predominantly t,he 44 A EBP (i.e. 4 S) with only a trace of the 58.5 A EBP (i.e. 5 S) (Fig. 2  This suggests that the sedimentation determination of the estrogen receptor in sucrose gradients was not altered by the sucrose, that the elution positions of the 4 S and 5 S EBPs from the Sephadex G-200 in 0.4 M KC1 were not due to a differential retention of the proteins by the gel, or that a concentration-dependent association of proteins was not occurring on gel chromatography.

Effect of urea on [sH]estradiol-binding
activity of rat uterine estrogen receptor Gel Chromatography with S M Urea-Sephadex G-200 chromatography with TEK-3 M urea buffer revealed that the 4 S and 5 S EBPs were eluted as separable proteins distinguished by a marked increase in their molecular radii. Upon removal of the 3 M urea the 4 S and 5 S EBPs reassumed their original molecular parameters: the 4 S EBP was eluted with a V,: VO of 1.40 (Fig. 4A, Curve b) and a molecular Stokes radius of 53.8 + 0.9 A (Fig. 3B). Peak b, after dialysis with TEK buffer and concentration by ultrafiltration, revealed an EBP on sucrose gradient analysis with a sedimentation coefficient of 4.4 S (Fig. 4B, Curve b). Sephadex G-260 chromatography with TEK-3 M urea buffer of the 5 S EBP indicated a V, : Vo of 1.16 (Fig. 4A, Curve a) and a molecular Stokes radius of 70.6 f 1 .O A (Fig. 30). Peak a, after dialysis and concentration, showed an EBP with a sedimentation coefficient of 5.5 S (Fig. 4B, Curve a) Efects of Urea on Estimations of Molecular Parameters--It was important to determine whether the effects of urea on the-estrogen receptor were caused by the increased density or denaturing activity of the urea, or by alteration of the conformational state, or dissociation of the estrogen receptor, or both.
In contrast to the estrogen receptor, various protein standards showed excellent agreement in their sedimentation behavior and distribution coefficients in chromatographic experiments in the absence or presence of 3 M urea. The molecular parameters of the protein standards were invariable with or without urea; under identical conditions the observed changes in the estrogen receptor's molecular parameters relative to the protein standards were significantly greater. protein standards which sedimented in a linear relationship, as described by Martin and Ames (24) ; estimates of sedimentation c0efficient.s were reliable in linear sucrose gradients of 5 to 20%, 5 to 20% with 3 M urea, 10 t.o 30%, or 12 to 33%. The absolute distances of migration of the standards were a function of the density of the gradients, in the order: 5 to 20% < 5 to 20% + 3 M urea = 10 to 30% < 12 to 33% sucrose gradients (Fig. 5). (b) The distribution coefficients (Kd) of the protein standards in Sephadex G-200 chromatography with or without 3 M urea were identical, indicating that the 3 M urea did not alter the molecular Stokes radius (Fig. 3). (c) The alkaline phosphate standards, after sucrose gradient centrifugation with or without 3 M urea, showed identical enzymatic activities.
(G?) The [3H]estradiolbinding capacity of the estrogen receptor was not decreased in 1 to 3 M urea (Table I). Several observations suggest that valid comparisons of the ex-Effect of pH MZ Molecular Parameters of Estrogen Recepforperimental data obtained with or without urea can be made.
With minor changes in buffer pH, the sedimentation coefficient (a) All sucrose gradient analyses contained two or three internal of the estrogen receptor was readily altered. These changes re-fleet conformational changes rather than significant changes in mass, Sucrose gradient analyses of the 4 S and 5 S EBPs with TEK buffer, pH 9.0, indicated the two EBPs were 3.9 S and 5.3 S (not illustrated); using HEK buffer, pH 6.8, the two EBPs were 4.7 h 0.04 S and 5.6 + 0.05 S (Fig. 6). During sucrose gradient analysis with HEK buffer the 5 S EBP showed less change in its sedimentation coefficient than the 4 S EBP. In addition, the 5 S EBP was observed to show a tendency to revert from the 5 S to the 4 S EBP on HEK gradients (Fig. 6; compare A and B). Sephadex G-200 chromatography in HEK buffer indicated that the 4 S EBP has a molecular Stokes radius of 43.2 + 0.6 A. The 5 S EBP on Sephadex G-206 chromatography with HEK buffer showed a [3H]estradiol-binding peak equivalent to a molecular Stokes radius of approximately 55 to 60 A, but 20 to 50% of the radioactivity appeared in the void volume. Because the radioactivity in the void volume may indicate a tendency of the labeled EBP to aggregate under these conditions, the analysis of this 5 S EBP chromatography data was not included in Table II The 4 9 and 5 S EBP (formation of the 5 S EBP is described in Fig. 1) were subjected to sucrose gradient centrifugation for 18 hours at 326,000 X gmax on a 5 to 207, sucrose gradient with TRK buffer, pH 7.4 (A), or a 5 to 207c sucrose gradient with HEK buffer, pH 6.8 (B). The increased sedimentation coefficient of the EBP in HEK buffer was observed when compared to the internal standards in the HEK-sucrose gradients or when compared to the standards in sucrose gradients containing TEK buffer. molecular Stokes radius from 43.2 to 53.8 A was observed by gel chromatography.
However, the molecular weight of the 4 S EBP (estimated with Equation 2) remained approximately 80,000 in each medium, indicating that the change occurred in the conformational state and not in the mass of the 4 S EBP.
The large values of the frictional ratio (f:fo, estimated with Equation 3) indicated that the 4 S EBP is an asymmetrical protein whose shape increased to a more asymmetrical form with conditions (3 M urea) favoring a decreased sedimentation coefficient.
The estimated molecular weight of the 5 S EBP was 133,000 (with or without 3 M urea), a 65% increase in molecular weight as compared to the 4 S EBP (-80,060), Table II. The data suggest that the 5 S EBP increased its molecular weight by association with a second component or subunit of approximately 50,000. The calculation of the molecular weights of the 4 S and 5 S EBP with Equation 2 was based upon the assumption that I? was 0.725 cm* per g for both EBP species.
Since the exact partial specific volume (6) of the 4 S and 5 S EBP was not known, we considered whether the difference in the estimated molecular weights of the 4 S EBP and 5 S EBP (80,000 versus 133,000) resulted from each having a different part#ia! specific volume (i,) induced during the in vitro transformation (e.g. by, acetylation, phosphorylation, peptide or lipid release, conformational changes, etc.). Direct estimations of the densities of the labeled 4 S and 5 S EBP, as well as theoretical considerations suggest that this was extremely improbabIe and support the concept of an increase in mass as a result of an association of two subunits.
The sedimentation rates of t.he 4 S and 5 S EBPs in very dense sucrose gradients (38 to 50y0) for 20, 60, and 90 hours showed no change with time, as compared to each other, or to the protein standards, thus indicating that the densities of the 4 S and 5 S EBP were not discernibly different.
Similarly, samples of the 4 S or 5 S EBPs layered on a 50% (w/v) sucrose solution in TEK buffer (density of 1.238 g per cm3 at 0') or on a 75% (w/v) sucrose solution in TEK buffer (density of 1.389 g per cm3 at 0") and centrifuged for 36 hours showed no difference in their migration into the dense sucrose solutions.
The 4 S and 5 S EBPs each migrated 0.7 ml below the meniscus in the 50% sucrose solution and showed no migration below the meniscus in the 75% sucrose solution.
These experiments suggest that the 4 S and 5 S EBPs have densities more t.han 1.238 g per cm3, corresponding to a 0 of 0.808 cm3 per g but not less than a density of 1.389 g per cm3, or a 2! of 0.720. The partial specific volume (a) for known proteins is generally found within the narrow range of 0.69 to 0.78 cm3 per g, with lipoproteins having the higher and glycoproteins the lower ti values (29,31). For one to assume  Fig. 3. f Calculated for prolate ellipsoids (37). were subjected to sucrose gradient centrifugation on a 5 to 2O"x sucrose gradient containing TEK buffer for 18 hours at 326,000 X gmax. Sucrose gradient analvsis of the 5 S EBP that was allowed to stand for 24 hours at @ without any additions, gave radioactivity profiles identical with A.
that the apparent increase in the molecular weight of the 5 S EBP was solely due to a change in 1! during the transformation process, a 20 to 30 y0 difference between the 2: values of the 4 S and 5 S EBPs would have been necessary. Such a marked difference in the v values of the 4 S and 5 S EBPs, extending beyond the normal range of 2', would have readily been revealed by their sedimentation behavior in high density sucrose. The difference in the sedimentation and chromatography analyses of the two EBPs cannot be accounted for by a difference in the molecular parameters which contribute to their partial specific volumes.
Conditions for Reversal of 5 S EBP to 4 S EBP--Incubation of the 5 S EBP in 0.4 M KCl, 3 M urea, and 40 mM Hepes, pH 6.8, readily causes its dissociation to the 4 S EBP without a loss of [3H]estradiol-binding activity. Any one of these reagents alone, even after 24-hour incubation, is not sufficient to produce a significant reversal of the labeled 5 S EBP to the 4 S EBP (Fig.  7). Gel chromatography of the EBP that results from incubation of 5 S EUP in 0.4 M KCl, 3 M urea, 40 mM Hepes, pH 6.8, confirmed that the molecular Stokes radius was 44 A, a value identical with that of the 4 S EBP, and indicated that dissociation rather than a unique conformational change in the 5 S EBP had occurred.
The rate of dissociation of the 5 S EBP during incubation with 0.4 M KCl, 3 M urea, and 40 mM Hepes, pH 6.8, was measured by sucrose gradient analysis.
The initial dissociation process was strictly first order with a half-life of 5 hours (Fig. 8).

Measurements
of the estrogen receptor with gel chromatography and sucrose gradient centrifugation under carefully corltrolled and standardized conditions have provided highly reproducible data that show a molecular relationship between the 4 S EBP and 5 S EBP forms of the receptor.
The estrogen receptor can readily assume a range of sedimentation coefficients (3.6 S to 5.5 S, Table II)  (28") ; we have also described the dissociation from the 5 S to the 4 S EBP (Figs. 7 and 8). (b) The estrogen receptor, i.e. the 4 S or 5 S EBP species, is capable of assuming different conformational states when observed with urea or with changes in pH.
The estimated molecular weight of the 4 S EBP is 80,000 and that of the 5 S EBP is 133,000. The 4 S to 5 S transformation must result from association of the 4 S EBP and a second component or subunit of approximately 50,000, not simply from a conformation change of the 4 S EBP. The analysis of the 4 S and 5 S EPBs in urea or HEK buffer indicates that the molecular weight of each EBP is relatively constant within the range of experimental error, although their sedimentation or chromatographic behaviors are readily altered as a result of conformational changes. Numerous investigations (32)(33)(34) using purified or crude preparations have shown that sedimentation coefficients and molecular Stokes radii measurements provide molecular parameter data comparable in accuracy to standard methods, e.g. sedimentation equilibrium studies of purified proteins. A kinetic analysis of the 4 S to 5 S EBP transformation has shown that it is a second order process, consistent with a bimolecular association model indicated by this report.3 The necessity of increasing the temperature to observe an association reaction between the 4 S EBP and its subunit indicates that it is an endothermic process. This is consistent with the observation that the estrogen receptor uptake by thenucleus is a temperaturedependent process (2,3).
The molecular characterization of the 4 S and 5 S EBP permits proposing a number of models for the molecular events involved in the 4 S to 5 S transformation and excludes others. (a) The subunits of the 5 S EBP, i.e. the 4 S EBP and its 50,000 molecular weight component, are usually weakly associated in dilute or isotonic buffers. Upon [3H]estradiol binding and at an elevated temperature (28"), an increased avidity between these two subunits occurs as a result of a conformational change; with sucrose gradient centrifugation in 0.4 M KC1 the 5 S EBP is observed intact and is presumably the active form of the estrogen receptor.
In the experiments where the estrogen receptor binds estradiol but does not undergo a temperature-dependent conformational change, i.e. at 0", using sucrose gradient centrifugation in the presence of 0.4 M KCI, a 4 S EBP is observed.
Although the dissociation of the estrogen receptor by 0.4 M KC1 and the appearance of the 4 S EBP are experimentally induced, the dissociation suggests that an "inactive" form of the estrogen receptor was present (Fig. 9). (6) In a simpler sequential model for the molecular events involved in the 4 S to 5 S transformation, the 4 S EBP forms a complex with estradiol and at 28" undergoes a conformational change leading to its "activation." The "activated" 4 S EBP by a random process associates with another specific protein.
(c) An alternative possibility is that the temperature-"activated" 4 S EBP is the active or functional form of the estrogen receptor and that its association with another protein to form the 5 S EBP is purely a fortuitous experimental reflection of the fact that the conformation change and "activation" of the 4 S EBP have taken place.
Recently, Stance1 et al. (19) have reported that the isolated nuclear 5 S and the cytoplasmic 4 S estrogen receptor could be dissociated by using 4 M urea, 1 M KCl, 50 InM mercaptoethanol, and 50 mM NaHS03, to a common 3.6 S estrogen-binding protein. In agreement with these investigators, our results have shown that the 5 S EBP, produced by the in vitro transformation method, can be dissociated to the 4 S EBP by the addition of 3 M urea, 0.4 M KCl, and 40 mM Hepes, pH 6.8. Under the experimental condition of this study, gel chromatography and sucrose gradient analyses indicate that the 4 S EBP did not, dissociate to a 3.6 S EBP as reported by Stance1 et (II. (19). We have observed that the 4 S is capable of assuming a lower sedimentation coefficient, but only as a result of a reversible conformational change.
The relationship of the 4 S or 5 S EBP observed using sucrose gradient analysis in high ionic strength buffers to the 8 S estrogen receptor observed with buffers of low ionic strength has not been elucidated here. This report has shown that the structure of the estrogen receptor is capable of assuming different conformational states, and raises the possible speculation that the 5 S EBP is the 8 S form of the estrogen receptor after it has undergone a conformational change due to the low ionic strength buffer. If one assumes that the molecular weight (-133,000) of the 5 S EBP is not altered, but its sedimentation coefficient is increased to 8 S, then the molecular Stokes radius must decrease to approximately 40 A radius. A 40 A radius would give the 8 S estrogen receptor a frictional coefficient (f:fO) of 1.1, typical of most globular proteins (29). However, molecular analysis of the 8 S estrogen receptor will not be possible until the conditions are found which will prevent the estrogen receptor from aggregating during gel chromatography with dilute buffers, or until the estrogen receptor has been purified sufficiently to use other methods. Another possible explanation may be that the 8 S form of the est.rogen receptor is a dimer of the 5s EBP that is formed in low ionic strength buffers.
A previous report from this laboratory (35) has shown the existence of a human uterine protease that is capable of "transforming" the 8 S estrogen receptor to a 4.5 S EBP.
This was observed in sucrose gradients without KCl. The 4.5 S EBP produced by the uterine protease has a molecular Stokes radius of 31 A, a molecular weight of 60,400, and a frictional coefficient (f:fO) of 1. 2. Puca et al. (15) have reported the presence in calf uteri of a "Ca2+-activated transforming factor" that appears to produce a 4.5 S EBP with molecular properties similar to that produced by the action of the human uterine protease. The molecular mechanism of the "Ca*+-activated transforming factor" has not been described.
It is evident that the "transformation" produced by the uterine protease yields a 4.5 S EBP (~60,400 mol wt), observed using sucrose gradients with low ionic strength buffers, which is different in its molecular parameters and behavior from the salt-dissociated 4.2 S EBP (-80,000 mol wt) described here. The action of the uterine protease is inhibit.ed by diisopropyl fluorophosphate (35), whereas the 4 S to 5 S EBP transformation is not, further indicating the dissimilarity of the two processes.' The in vitro induced 4 S to 5 S EBP transformation produces a 5 S EBP that is similar to the in &JO form of the nuclear 5 S EBP (13,36).
Recently, the in vitro induced 5 S EBP and the isolated nuclear 5 S EBP have been report.ed to activate RNA polymerase, while the 4 S EBP does not (9). These measurements of the molecular relationship between the two EBP, their conformational changes, and protein-protein interactions may provide an insight into the molecular mechanisms involved in the activation of the estrogen receptor.