Detection of two high molecular weight hydrophobic forms of the human estrogen receptor

J. sreroid &x-hem. Vol. 33, No. 5, pp. 965-970, 1989 Printed in Great Britain. All rights reserved

0022-4731/89 $3.00 + 0.00 Copyright f(J 1989 Pergamon Press plc

DETECTION OF TWO HIGH MOLECULAR WEIGHT HYDROPHOBIC FORMS OF THE HUMAN ESTROGEN RECEPTOR SALMAN M. HYDER and JAMES L. WITTLIFF* Hormone Receptor Laboratory, James Graham Brown Cancer Center, University of Louisville, KY 40292. U.S.A. (Received 3 October 1988; received for publication 2 June 1989)

Summary-The human estrogen receptor gene encodes a single protein of molecular weight 65,000 daltons. However, using a sensitive and rapid technique of high-performance hydrophobic interaction chromatography we have detected two distinct estrogen receptor species both of which are high molecular weight proteins (cu. 60A) as determined by high-performance size-exclusion chromatography. These are detected either in the presence or absence of sodium molybdate; rechromatography of individual isoform indicates that the two protein complexes have independent hydrophobic contact points. Co~istent elution patterns of the two receptor species indicates they are formed selectively. We conclude that different ~st-tr~slation~ m~ifi~tions of the estrogen receptor protein could allow their specific interaction with non-receptor components resulting in the formation of two distinct high molecular weight complexes which would be rapidly resolved by high-performance hydrophobic interaction chromatography.

INTRODUCI’ION

The estrogen receptor (ER) from several sex-steroid target tissues has been cloned [l, 21. The gene sequences from all the different sources share homology indicating an evolutionary relationship of steroid receptors between different species [1, 21. Several functional domains have been assigned to the predicted protein structure of the ER (mol. wt 65,080 daltons) in combination with mutagenesis studies in vitro. These include the cysteine-rich DNA binding domain, the N-terminus portion with no clear definition but thought to be involved in protein-protein interaction and finally the steroid binding domain near the C-terminus which is rich in hydrophobic amino acid residues 121. In soluble fractions from target cells, ER and other steroid receptors are associated with macromolecules such as heat-shock proteins (hsp90) which promote their s~imentation as 8s complexes on sucrose density gradients [3]. Progesterone receptor (PR) also exists as two different 8s complexes [4]. The possibie functional role of these 8s complexes has been implicated from studies with glucocorticoid receptor (GR) where it was shown clearly that RU 486, an antiprogestin, abolished GR activation by preventing disaggregation of 8s complex [5]. Thus large molecular weight receptor proteins may have physiological significance due to association with non-receptor *Address for correspondence: Dr James L. Wittliff, Hormone Receptor Laboratory, James G. Brown Cancer Center, University of Louisvifie, Louisville, KY 40292, U.S.A.

proteins some of which may have regulatory function. Purification of these high molecular weight isoforms should provide greater insight into the importance of their various components in hormone action. In this report we present evidence that high-performance hydrophobic interaction chromatography (HPHIC) resolves two high moiecular weight forms of ER of approximately 6OA each. The receptor isoforms do not appear randomly since their presence is consistent in different tumor biopsies. Each isoform has an independent contact point with the chemicahy bonded phase of the HPHIC column since reanalysis of individual peaks resulted in elution of receptor with the same retention time. HPHIC of each high molecular weight complex eluted from HPSEC either in the absence or presence of sodium molybdate {MOO:-) resulted in HPHIC chromatographic patterns similar to those obtained with original cytosol. This is the first report of two different high molecular weight complexes of ER detected either in the absence or presence of molybdate with distinct hydrophobic properties. EXPERIMENTAL

Materials and methods HPLC-grade ammonium sulfate was obtained from Bio-Rad Laboratories (Richmond, Calif.). The ligand [16a-“*‘I]iodoestradiol-178 (ca 2200 Ci/mmol) (IE) was obtained from New England Nuclear/ DuPont (Boston, Mass.). Na, MoQ disodium ethylen~i~netetraa~tic acid (EDTA), and glycerol

965

SALMANM. HYDER and JAMESL. WI~LEFF

966

were purchased from Fisher Scientific (Louisville, Ky), Unlabeled diethylstilbestrol (DES) Norite A, Dextran T-70, dithiothreitol (DTT), were obtained from Sigma (St Louis, MO.). Human breast tumor tissues were provided by the various surgeons and pathologists at local hospitals, cooperating with the Hormone Receptor Laboratory. The tissues were brought to the laboratory on dry ice and kept frozen at -86°C until analyzed. Only residual tissue from clinical receptor analyses was used in this study.

One-ml fractions were collected, and the free and protein-bound steroid were detected radiometrically in a Micromedics 4/600 gamma radioisotope detector (Rohm & Haas, Cleveland, Ohio), having a counting efficiency of 65%. Since the nonspecific binding (radioacti~ty eluted from cytosols labeled in the presence of DES) showed mainly base levels and represented no more than 5-10% of the total binding, these are not shown in the figures. Recovery of total radioactivity and injected protein was usually 75--100%.

Preparation and labeling of soluble estrogen receptor

High performance (HPSEC)

All procedures were performed at 4°C. Human breast tumors (ca 200-400 mg/ml were homogenized in P,,EDG [IOmM phosphate/l.5 mM EDTA/l mM EDTA/l mM DTT/lO% (v/v), glycerol (PH 7.411. Homoge~~tion was performed with two IO-set bursts in a Brinkman (Westbury, N.Y.) Polytron homogenizer. Soluble fractions were prepared by centrifugation of the homogenate for 30min at 40,OOOrpm in a Beckman Ti 70.1 rotor (San Ramon, Calif.). The su~matant was carefully removed, avoiding the layer of fat at the top. The soluble fractions were labeled with 2-3 nM IE in the presence and absence of a 200-fold excess of DES for 2-4 h at 4°C. Free steroid was removed with 1% (w/v) destran-coated charcoal (DCC). DCC was then removed by centrifuging the sample for 5 min at 1OOOg. Cytosol protein con~ntrations were determined by the method of Bradford[6], using bovine serum albumin as the standard. The protein concentrations generally ranged from 4 to 8 mg/ml. Highperformance ~fograp~y

hydrophobic

interaction

chro-

in a Chromatography was performed Puffer-Hubbard cold box (Ashville, N.C.) at 4°C. All buffers were filtered under vacuum through Millipore 0.45 pm HAWP filters (Bedford, Mass.) before use. Free steroid or estrogen receptor complexes were applied to the SynChropak propyl column (300A pore size) obtained from SynChrom (LaFayette, Ind.) using an A1te.xModel 210 sample injection valve. All samples were adjusted to 1.5 M (NH,),SO, prior to injection. Elution was carried out with a Beckman Model 114 solvent delivery module including a Model 421 system controller. Unless otherwise stated, the gradient program consisted of a preliminary wash with Eluent A (P,,EDG), containing 2 M (NH4)2S04 (pH 7.4) at flow-rate of 1 ml/min. Following sample injection, a descending salt gradient was developed to reach 75% P,,EDG (Eluent B) in 10 min followed by a gradient reaching 100% B in 30 min. Eluent B was maintained at a flow rate of 1 ml/min for the next 20 min before re-equilibration with Eluent A. In experiments which required Na,MoO, in the mobile phase, both Eluents A and B contained 10 mM Naz MOO,.

size excfwion

chromatography

Analytical size-exclusion columns (Spherogel TSK3000 SW), particle size 10 pm (7.5 x 600 mm) from Beckman Altex Instruments, were used for steroid receptor separation, as described ~~~0~1~~~. HPLC was performed at 4°C with a Beckman 114 solvent delivery module, including a Model 421 system controller and injector block. Cytosols were applied in lOO-250-~1 volumes with a Hamilton syringe. Estrogen receptor peaks previously separated by size exclusion chromatography were reinjected onto the hydrophobic column in 500-700 ,ul volumes. The elution buffer (PH 7.4) at 4”C, was PEDGK,, (10 mM phosphate buffer/l.5 mM EDTA/l mM DTT/lO% (v/v) glycerol/l00 mM KCl). All buffers were filtered through a 0.45pm filter (Millipore). Elution was carried out at a flow-rate of 0.7 ml/min. Fractions were collected at OS-min intervals in 12 x 75 mm tubes. Recoveries were in the range of 7.5-100%. RESULTS AND DISCUSSION The receptor itself and its associated proteins provide several potential contact points for interaction with the bonded phase of hydrophobic matrices. Since HPHIC is a mild separation procedure, we utilized this property to isolate soluble ER isoforms in their native states. Large molecular weight ER complexes have previously been detected employing other types of chromatography [7,8]. HPHIC ap pears to be more effective for separation of various enzymes with virtually 100% recoveries of their activity [9]. Previously we reported that estrogen receptors (ER) from either human breast tumors or rat uterus can be separated into two isoforms based on their surface hydrophobic or ionic properties [lO-171. Receptor polymorphism appears to have biological significance in that the distribution of receptor isoforms changed specifically with tissue differentiation [18, 191 and appears to be related to endocrine responsiveness in breast cancer [IS]. To ascertain the origin of recep tor hydrophobic polymorphism, a study was undertaken using the sensitive HPLC techniques we developed earlier in conjunction with a high gamma emitting radioisotope, [lz5I]iodoestrodiol-178 [ 12, 141.

Hydrophobic focms of human estrogen receptor

a-

,-

>-

,-

, Cl

10

20

30

40

50

60

0

203040Msc

2030405060

FRACTION N”kR

Fig. 1. Size-exclusion chromatography of estrogen receptor isofocms first separated by high-performance hydrophobic interaction chromatography in the presence of sodium molybdate. Estrogen receptors were extracted from human breast tumor (200-484l me/ml) using a Brinkmann Polytcon (Westbury, N.Y.). The buffer was composed of 10 mM phosphate, 1.5 mM EDTA, 1 mM D’IT and 10% v/v glycerol (P,,EDG). Cytosol (105,000 g supematant) was made 10 mM with respect to sodium molybdate. We have previously shown that addition of molybdate before or after homogenization or at the cytosol level provide similar results. Receptors ace labeled with 3 nM [16J2’I]iodoestcadiol for 3 h at 4°C in either the presence oc absence of 200-fold excess (DES). Following labeling, free. steroid was removed with 1% w/v (DCC) as previously described. (A) Cytosol (7.4mg/ml) was injected onto a SynChcopak Pcopyl 300 column obtained from SynChcorn (Linden, Ind.) which was previously equilibrated with 2 M (NH&Son in P,,EDG supplemented with 10 mM molybdate (mobile phase A). Following injection of sample, a two phase linear gradient was developed with Beckman Model 114 solvent delivery module linked to a 421 system controller, first to reach 75% mobile phase B(P,,EDG) in 10min then to reach 100% B in the next 30min. Mobile phase B was then sustained at 100% for the next 10 min. (0) represents [‘25]iodoestcadiol labeled receptor. The elution profile from DES competition showed a curve similar to the base line and is omitted from the figures for clarity. Non-specific components never exceeded 57% of the total binding. The eluted ceeeptoc isofocms (MI and MII) were then injected onto a TSK 3000 SW size-exclusion column (7.5 x 300 mm) obtained from Beekman/Altex, San Ramon, CA (B, C). The injection volume was 1 ml of the peak fractions from the HPHIC eluted fractions shown in A. Chcomatogcam was developed with mobile phase consisting of P,,EDG + 100 mM KC1 as described in out previous publication [18]. The size markers used were blue dextcan to obtain the void volume (V,), Fe (feccitin), Hb (Haemoglobin) and cytoehcome C (Cc). Fig. 1D represents the size-exclusion profile of the original cytosol (injection vol was 100 ~1). Recoveries of the samples injected were each near 100%.

In the presence of MOO:-, ER isoforcns eluted at a retention time (R,) of 13.2 + 0.4 min (n = 6) and 20 f 0.6 min (n = 6) from HPHIC column (Fig. 1).

Isoform MI was only seen in the presence of MOO:-. It appeared to be unstable, interconverting into other isoforms (detected in the absence of MOO:-) when MOO:- is absent from the mobile phase [13]. These results were consistent using more than 15 individual human breast tumors which argues against random formation of these isoforms. These data also confirm our previous observation using another type of HPHIC column which had a polyether linked bonded phase [13, 141. Subsequent analysis of ER eluted from HPHIC (Fig. 1A) by high performance size exclusion chromatography (HPSEC) revealed that both MI and MI1 were high molecular weight proteins (Fig. 1B and 1C). Fig. 1D represents the HPSEC profile of the original cytosol. Previously we reported there was little interconversion between isoform MI and MI1 suggesting these do not represent equilibrium prod-

ucts [l l] or that this interconversion is retarded by MOO:-. Thus two distinct high molecular weight species of ER exist with different hydrophobic pcopet-ties. Reanalysis of these isoforms separately yielded distinct ER peaks reaffirming the separate identity of isoforms MI and MI1 (Fig. 2). HPHIC of ER in the absence of MOO:- also yielded two isoforms (Fig. 3A) with retention times of 21.2+ l.Omin (n = 18) and 31.7+0.8min (n = 18). Based on the R, values, it appears that isoforms MI1 and I are similar. Although analysis of isoforms I and II indicated high molecular weight proteins (Fig. 3B and C) similar to that obtained in the presence of sodium molybdate, these profiles were not always consistent. Other complicated profiles were seen implying that in the absence of sodium molybdate, both forms I and II are unstable macromolecules which dissociate into smaller components upon analysis by HPSEC. This suggests a column induced effect on receptor stability. A stabilizing factor such as an associated RNA molecule [20] may have been lost on

SALMAN

M.

HYDER and

Jm

L.

WIITLIFF

-t :. RECHROMAlCORAPHI

IO 20

0

30

40

30

YD

60

FRACTtON NUMBER Fig. 2. R~hromato~aphy of lsoforms MI and MI1 by HPHIC. (A) Cytosol was tirst injected onto the synchropak propyl column and eluted with buffers containing sodium molybdate as described in Fig. 1. Each of the eluted isoforms MI and MI1 was then reinjected onto the Synchropak propyl column (B, C) and eluted under similar conditions. Recoveries were near 100% in all cases.

leaving the receptor unstable during HPSEC. The hydrophobic properties of individ~l receptor forms, however, were conserved when HPHIC eluted ER isoforms were reanalyzed (Fig. 4). This indicates that individual receptor isofonns (I and II) have distinct contact points with the bonded phase and there was no interconversion of the two within the experimental time required to complete the assay. Presence of MOO:- may promote specific association of nonreceptor components with ER thereby altering the hydrophobicity of the complex. Alternatively the molybdate ions may directly alter receptor complex conformation which influences receptor elution from the hydrophobic matrice. HPHIC

To confirm that the high molecular weight components of ER were ~l~o~hic we partially purified ER from human breast cancer by HPSEC (Fig. 5A and C) both in the absence and presence of molybdate. Each of the high molecular weight protein complexes was then analyzed by HPHIC. The results showed that each homogeneous HPSEC peak was actually composed of heterogeneous hydrophobic forms with the elution profile being identical to that obtained with original cytosol. These data indicate that the two types of hydrophobic peaks arise due to highly specific surface characteristics of the receptor complexes. Reasons for this phenomenon include (1) direct interaction of the receptor with the bonded

- c. yc,

6. HPSEC I

A HPHlC

Z*

1

I

to-

IO

0.“PSEc

1,

II

i

1

ORlOlNAL C”TOSOL

o-

o-

‘-

~

0 02eso4oMoo

02030405060

FRACTION NUMBER Fig. 3. Size-exclusion chromatography of estrogen receptor isoforms first separated by HPHIC in the absence of sodium molybdate, Cytosol(7.0 mg/ml) was injected onto the Synchropak propyl300 column (A) and eluted as described in the legend to Fig. 1. Procedures for sample preparation and receptor elution were as described in the legend to Fig. 1. (A) HPHIC profile of original sample labeled with [1251]iodoestradiol (0); (B, C) Size-exclusion chromatography of HPHIC eluted isoforms I and II respectively; (D) HPSEC profile of original cytosol (injection volume 100 ~1).

Hydrophobic forms of human estrogen receptor o “1

A. ORIOINAL

B. RECNROMATOGRAPHY

CYTOSOL

32

P

0

IO

20

30

969

40

60

60

IO

0

20

30

40

50

60

FRACTION NUMBER

Fig. 4. Rechromatography of Isoforms I and II by HPHIC. (A) Cytosol(8.1 mg/ml) was first injected onto the SynChropak Propyl 300 column and eluted with buffers lacking sodium molybdate as described in

Fig. 1. Each of the eluted isoforms I and II was then rcinjected separately onto the SynChropak Propyl Column 300 (B) and elutcd under similar conditions. Recoveries were approximately 100% in all cases.

phase, (2) interactions of the associated nonreceptor component with the bonded phase of (3) an alteration of either of the above interaction due to posttranslational modifications in the receptor molecule itself such as protein phosphorylation. Protein kinase activity has been detected associated with estrogen receptors from human breast cancer and uterus [I l-131. Collectively our results indicate that the two hydrophobic isofonns represent distinct molecular complexes. Isoforms I and II appear to acquire other specific components during incubation with molybdate which bring about their conversion to isofonns

MI and MII. Based on their retention times, isoforms MI1 and I may represent the same species of ER. MI which was observed only in the presence of MOO:-, appears to represent ER complexed with other macromolecules (e.g. hsp90). In the absence of MOO:-, dissociation occurs exposing the hydrophobic binding domain responsible for the conversion to isoform II). Our previous studies [13, 141 suggested isoform II involved a direct interaction of the DNA binding domain with the bonded phase since MOO:selectively interconverted II into MI. The physiological significance of these high molecular weight forms of receptor remains to be ex-

1 c

cl. HP6EC-“PHIC

HPSEC

+mOf-

l

1

Moo.PY”

FRACTION NUMBER

Fig. 5. Hydrophobic characteristics of high molecular weight estrogen receptors separated ftrst by size-exclusion chromatography. (A) Cytosol(8.6 mg/ml) was injected onto the TSK-3000 SW column both in the absence (A) and presence (C) of sodium molybdate and eluted with P,,EDG + 100 mM KCl. The receptor which eluted in the void volume was then analyzed by HPHIC in the absence (B) or presence (D) of sodium molybdate. The elution conditions from the hydrophobic column were the same as described in the legend to Fig. I. Recoveries in all cases were approximately 100%.

SALMANM. HYDER and JAW

970

plored. The association of hsp90 with receptor was generally thought to represent nonspecific interaction. However it now appears to have physiological significance [S]. Using monoclonal antibodies we have identified hsp90 coeluting with MI but not with MII, I or II (Dr D. Toft, Mayo Clinic, Rochester, Minn, personal communication). hsp90 has been suggested to be directly involved in receptor transport and/or the ability of receptor to undergo activation [5,21]. Certain drugs tend to retain hsp90 bound to the GR and prevent its activation [5]. It is unknown if hsp90 has comparable functions in high molecular weight ER complexes. The aggregation protential of receptors may signify their ability to interact with factors and/or other macromolecules required for efficient tanscriptional events. The availability of efficient transcription assay systems in uitro should prove effective in discerning the role of receptor isoforms in this function. Although the cloned gene sequence specifies a single 65 kDa protein molecule, molecular heterogeneity is evident [lO-191. It appears that this polymorphism arises from post-translational modification of the steroid binding molecule itself and association with other nonreceptor components such as hsp90 and protein kinases [3, 11,221. HPHIC as described in our laboratory provides a unique, rapid method of isolating intact, high molecular weight ER isoforms which appear to be native species. Using this technique, one may now study the complex assembly process reflecting the synthesis of individual components, and the influence of physiological perturbations such as endocrine status and age. Acknowledgements-This research was supported in part by USPHS Grant CA-42154 from the National Cancer Institute and by Phi Beta Psi Sorority. The authors are appreciative of the technical assistance of MS Nancy Heer. REFERENCES 1. Evans R. M.: The steroid and thyroid hormone receptor superfamily. Science 240 (1988) 889-895. 2. Green S., Walter P., Kumar V., Krust A., Bomert J. M.,

Argos P. and Chambon P.: Human oestrogen receptor cDNA: Sequence, expression and homology to v-erb A. Nature 320 (1986) 134-139. 3. Joab I., Radenyi C., Renoir M., Buchou T., Catelli M. G.. Binart N. Mester J. and Baulieu E. E.: Common non-hormone binding component in nontransformed chick oviduct receptors of four steroid hormone receptors. Nature 308 (1984) 850-853. 4. Sullivan W. P., Sullivan B. T., Vroman V. J., Bauer, R.K., Puri R. M., Riehl G. R., Pearson and Toft T. 0.: Immunological evidence that the non-hormone binding component of avian steroid receptors exist in a wide range of tissues and species. Biochemistry 24 (1985) 6586-6591. 5. Lefebvre P., Formstecher

P., Richard C. and Dautrevaux, M.: RU 486 stabilizes a high molecular weight form of the glucocorticoid receptor containing the 90K non-steroid binding protein in thymus cells. Biochem. biophys. Res. Commun. 150 (1988) 1221-1229. 6. Bradford M. M.: A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anulyt.

L. WI’ITLIFF

Biochem. 72 (1976) 248-254.

7. Wiehle R. D., Hofmann G. E., Fuchs A. and Wittliff, J. L.: High-performance size exclusion chromatography as a rapid method for the separation of steroid hormone receptors. J. Chromat. 307-(1984) 39-51. 8. Kortvlewicz J. B.. Nelson K.. Pavlik E. J.. Van Naaell J. R:, Gallion H: H., Donaldson E. S. and Kenidy D. E.: Identification of a very large nuclear estrogen receptor complex. Endocrinology 120(1987) 218%2191. 9. Miller N. T., Feibush B., Corina K., Lee S. P. and Karger B. L.: High-performance hydrophobic interaction chromatography: purification of rat liver carbamoyl phosphate synthetase I and omithine transcarbamoylase. Analyt. Biochem. 148 (1985) 51&517. 10. Hyder S. M., Wiehle R. D., Brandt D. W. and Wittliff J. L.: High-performance hydrophobic interaction chromatography of steroid hormone receptors. J. Chromat. 327 (1985) 237-246.

11. Hyder S. M., Sato N. and Wittliff J. L.: Characterization of estrogen receptors and associated protein kinase activity by high-performance hydrophobic interaction chromatography. J. Chromat. 397, 251-267. 12. Hyder S. M. and Wittliff J. L.: High-performance hydrophobic interaction chromatography of a labile regulatory protein: the estrogen receptor. Biochromatogruphy 2 (1987) 121-130.

13. Hyder S. M., Sato N., Hogancamp W. H. and Wittliff J. L.: High-performance hydrophobic interaction chromatography of estrogen receptors and magnesium dependent protein kinase(s): Detection of two molecular forms of estrogen receptors in the presence and absence of sodium molybdate. J. steroid Biochem. 29 (1988) 197-206.

14. Hyder S. M. and Wittliff J. L.: High-performance hydrophobic interaction chromatography as a means of identifying estrogen receptors expressing different binding domains. J. Chromut. 444 (1988) 225-237. 15. Wittliff J. L., Allegra J., Day T. J., Jr. and Hyder S. M.: Structural features and clinical significance of estrogen receptors. In Steroid Receptors in Health and Disease (Edited by V. K. Moudgil), Plenum Press, New York (1988) pp. 287-312. 16. Wittliff J. L., Shahabi N. A., Hyder S. M., van der Walt A., Myatt L., Boyle D. M. and He Y-J.: High performance liquid chromatography as a means of characterizing isoforms of steroid hormone receptor proteins. In Proteins: Structure and Function (Edited by J. J. Italian) Plenum Publishing, New York (1987) pp. 61-74. 17. Hyder S. M., Shahabi N. A. and Wittliff J. L.: Microanalysis of estrogen receptors from human uteri by multi-dimensional high-performance liquid chromatography. Biochromotog&phy 3 (1988) 216-224. 18. Hvder S. M.. Wiehle R. D. and Wittliff J. L.: Alterations in-estrogen receptor isoforms in the mammary gland and uterus of the rat during differentiation. Comp. Biochem. Phvsiol. 91B (1988) 517-525.

19. Wiehle R. D:and Wittliff J. L:: Alterations in sex steroid hormone receptors during mammary gland differentiation in the rat. Como. Biochem. Phvsiol. 76 (1983) 409417. 20. Feldman M., Kallos J. and Hollander V. P.: RNA inhibits estrogen receptor binding to DNA. J. biol. Chem. 256 (1981) 1145-1148. 21. Koyasu S., Nishida E., Kadowaki T., Matsuxaki F., Iida K., Harada F., Kasuga M., Sakai H. and Yahara I.: Two mammalian heat shock proteins, HSP-90 and HSP-100 are actin binding proteins. Proc. nom. Acad. Sci. U.S.A. 83 (1986) 8054-8058. 22. Hyder S. M. and Wittliff J. L.: Separation of two molecular forms of human estrogen receptor by hydrophobic interaction chromatography: Gradient optimization and tissue comparison. J. Chromat. (1989) In Press.