Development and characterization of monoclonal antibodies to a specific domain of human estrogen receptor Abdulmaged M. Traish, * Rachel Ettinger,? and Herbert H. Wotiz* Departments

of Biochemistry”

Noel Kim,* Ann Marshak-Rothstein,?

and Microbiology,

f Boston

University

School of Medicine,

Boston, MA, USA

We have synthesized three peptides with amino acid sequences identical to those spanning amino acids 201-215, 231-245, and 247-261 of the human estrogen receptor (hER). These peptides were conjugated to keyhole limpet hemocyanin and used as immunogens to develop monoclonal antibodies (MoAbs) to hER. Antibody responses were only elicited by the peptide with amino acid sequence 247-261. Splenocytes from immunized mice were used for hybridoma production. Of the seven MoAbs that recognized the native (functional) form of the ER, four (MoAbs 16, 33, 114, and 213) recognized the ER with high af$nity, as demonstrated by the increased sedimentation coefjcient of the antibody-complexed ER in sucrose density gradients. Antibodies 318,35, and 36 bound to ER with low affinity since they immunoprecipitated ER, but the ER-antibody complex appeared to dissociate on sucrose density gradients. The high-affinity MoAbs appear to be site-specific since the peptide competed effectively for binding oj the receptor by the antibody. The fact that they reacted with ER from human breast cancer and calf, rat, and mouse uterine tissues suggests that this epitope of the receptor is conserved in these species. Although the DNA-binding region appears to be conserved among the various steroid receptors, these MoAbs did not recognize the native forms of progesterone, androgen, or glucocorticoid receptors. These MoAbs bound to the KCl-activated 4s ER and heat-transformed SS ER, suggesting that the antibody-binding site is accessible in the monomeric and dimeric forms of ER. The antibodies did not recognize the untransformed 8s ER in the presence of molybdate and without KCl, suggesting that the antibody-binding site in the oligomeric form of ER is inaccessible. The fact that the antibodies did bind to the unoccupied4S ER was demonstrated by the data obtained with sucrose density gradient analysis followed by postlabeling of ER with [3H]estradiol. The antibodies bound to ERs with high affinity (Kn = 0.4 to 1.8 nM). At a$xed concentration of antibody, ERs ranging from 20 to I ,OOOfmol were detectable. These MoAbs did not inhibit nuclear or DNA binding of ER in vitro. This can be attributed to the dissociation of the antibodies from ER when the latter interacts with its acceptor site. These results demonstrate the development of site-spec$c MoAbs to the native form of the hER using synthetic peptides as immunogens. (Steroids 55196-208, 1990)

Keywords:

steroids; estrogen receptor; monoclonal antibodies

Introduction Steroid hormone receptors represent a family of ligand-dependent transacting regulatory factors that stimulate transcription of specific genes by binding to specific DNA sequences termed hormone responsive

Address reprint requests to Dr. Abdulmaged M. Traish, Department of Biochemistry, Boston University School of Medicine. Boston. MA 02118, USA. Received September 13, 1989; revised December 19, 1989.

196

Steroids,

1990, vol. 55, May

elements. * Steroid receptors are intracellular proteins that are biologically inactive in the absence of their respective hormones. Subsequent to hormone binding, a process, termed activation and/or transformation, ensues which leads to the formation of functional hormone receptor complexes with high affinity for nuclear components. 2-5 Despite numerous studies, however, the precise mechanism of receptor binding to nuclear components remains poorly understood. Recent investigations, in which the complementary DNAs (cDNAs) for steroid receptors have been cloned, have generated new information concerning 0 1990 Butterworth

Publishers

Monoclonal

the functional domains of steroid receptors.“i2 To date, all nuclear receptors appear to have similar structural organization13 in that they exhibit a variable N-terminal region (A/B), a short and well-conserved cysteine-rich region, and a relatively conserved C-terminal region. These three domains are separated by less-defined regions. Functional studies have delineated the specific roles of these domains.i3 Thus, it was proposed that the N-terminal region is important for the activation of specific genes and is required for full functional activity. The central domain is thought to be responsible for DNA binding and transactivation. The steroid-binding region appears to have various domains with specific functions, such as the nuclear translocation signal, transactivation, dimerization, and binding to heat-shock proteins.13 Comparison of the human estrogen receptor (hER) sequence with other steroid receptors revealed the presence of two regions, termed “C” and “E,” which are conserved among the various steroid and related nuclear receptor proteins . 1~6-12Region “C” contains the putative DNA binding domain; it is rich in arginyl, lysyl, and cysteine residues. The presence of these conserved cysteine residues may allow the formation of structures similar to the DNA-binding zinc fingers proposed for eukaryotic transcription factor TFIIIA.i4 Region “E” contains the steroid-binding domain. Our research over the past several years has focused on the reactions involving conversion of the ER from a state with low affinity for nuclei to a state with high affinity.2 It has been suggested that the hydrophilic region of the receptor, spanning amino acids 185 to 263, represents the DNA-binding domain and is probably exposed to the surface of the protein on activation6Ti5 To further analyze the structure-function relationship of the DNA-binding domain, we produced monoclonal antibodies (MoAbs) directed against preselected regions of the ER protein. We synthesized three peptides with sequences identical to amino acids 201-215 (AT,), 231-245 (ATz), and 247-261 (AT3) of the hER. These peptides, after linking to an antigenic protein, were used as immunogens to produce antibodies specific for this domain of hER.i6 Several polyclonal antibodies and MoAbs to calf and human ER have been produced and characterized.*7-20 None, however, has been specifically directed toward the DNA-binding region of ER. Although some of these antibodies were shown to inhibit DNA binding partially or to interfere with receptor activation, the precise epitope was not known and the effects of the antibodies may be the result of recognition of a site distal to the DNA-binding domain.21*22In this report, we describe the development and initial characterization of these site-specific MoAbs and their interactions with ER. Experimental Isotopes and chemicals The following isotopes and chemicals were used: 6,7[3H]Estradiol (40 to 60 Ci/mmol) ([3H]E2), 7a-

antibodies

to human estrogen receptors:

Traish et al.

methyl-17a-[3H]methyl-19-nortestosterone (70-85 Ci/ mmol) ([3H]DMNT), 16a-ethyl-21-hydroxy-9-nor[6,7-3H]-pregn-4-ene-3,20-dione (40 to 60 Ci/mmol) ([3H]ORG 2058), and [3H]dexamethasone (95 Gil mmol) ([3H]DEX). Unlabeled DMNT and ORG 2058 were obtained from Amersham (Arlington Heights, IL, USA). Unlabeled diethylstilbestrol (DES) and E2 were obtained from Steraloids (Wilton, NH, USA). Unlabeled DEX was obtained from Sigma Chemical Co. (St. Louis, MO, USA). All other chemicals were reagent grade and were obtained from commercial sources.

Buffers and solutions The following buffers and solutions were used: phosphate-buffered saline (PBS) (0.2g KH2P04, 8g NaCl, 2.16g Na2HP04 * 7 H20 in 1 liter of distilled water, pH 7.5); buffer TGET (50 mM Tris-HCl, 10% [vol/vol] glycerol, 1 mM EDTA, 10 mM monothioglycerol, 0.02% NaN3, pH 7.4 at 2C); buffer TGET/Mo (buffer TGET containing 10 mM sodium molybdate); buffer TGET/Mo KC1 (buffer TGET/Mo containing 0.4M KCl); buffer TT (50 mM Tris-HCl, 10 mM monothioglycerol pH 7.4 at 2C); buffer TT/Mo (buffer TT with 10 mM sodium molybdate); and buffer TT/Mo/KCl (buffer TT/Mo containing 0.4 M KCl). Human and animal tissues Rats. Male Sprague-Dawley rats (300 to 350 g body weight) and female immature 22- to 23-day-old rats were obtained from Charles River Breeding Laboratories (Wilmington, MA, USA). Male rats were castrated as described23; 24 hours later, they were killed, and the prostates were excised and stored at -70 C until use. Rat uterine tissue was obtained from immature female animals and stored frozen at -70 C until use. Mice. Five- to eight-week-old female (BALB/c x A/J) mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). Human breast cancer tissue. Human breast cancer tissues were obtained from patients undergoing surgery for breast cancer as described previously.24 The tissue was placed on dry ice and transported to the laboratory where it was stored frozen at -70 C until experimentation. Calf uterine tissue. Calf uterine tissues were obtained from a local slaughter house, placed on dry ice, and transported to the laboratory as described elsewhere.25 Preparation

of tissue cytosols

Cytosol fractions from each tissue type were prepared in the appropriate buffer as described previously.23-25 Briefly, fresh tissue or frozen tissue powder was homogenized (1 g/4 ml) in buffer TGET/Mo, pH 7.4, at 2 C. The homogenate was then centrifuged at 105,000 x Steroids,

1990, vol. 55, May

197

Papers g for 45 minutes at 2 C and the supernatant fraction (cytosol) was used for receptor-binding studies. Labeling

of cytosols with steroid hormones

To label the ER, aliquots of the tissue cytosols were incubated at 2 C for 4 hours with 5 nM [3H]E2 in the absence (total binding) or presence (nonspecific binding) of a lOO-fold molar excess of unlabeled DES.24s25 Progesterone receptors were labeled by incubating calf uterine cytosol with 15 nM [3H]ORG 2058 in the absence or presence of unlabeled ORG 2058 as described previously.26 Androgen receptors (ARs) were labeled by incubating rat prostatic cytosol with 10 nM [3H]DMNT in the absence or presence of unlabeled DHT as described elsewhere.23 Glucocorticoid receptors were labeled by incubating calf uterine cytosol with 10 nM t3H]DEX in the absence or presence of unlabeled DEX as described previously.2’ At the end of the incubation, free radioactivity was removed with dextran-coated charcoal (DCC) pellets and the supernatant was used for antibody receptor interactions. Sucrose

density gradient analysis

Sucrose density gradients (5% to 20%) were prepared with TGET/Mo buffer containing 0.4 M KCl. In some experiments (as indicated), the gradients were made 5% to 20% in TGET/Mo buffer with or without 0.4 M KCI. Samples to be analyzed were layered on the gradient together with t4C-labeled bovine serum albumin (4.6s) and human gamma-globulin (7s) as sedimentation markers. The gradients were centrifuged at 50,~O rpm in an SW60 rotor (297,~ x g) for 18 hours at 2 C. Gradients were fractionated into individual 0. l-ml fractions, scintillation fluid was added to each sample, and radioactivity was counted. Nuclear binding assay of activated t3H]E2R complexes Nuclear fractions of calf uterine tissue were prepared, and the nuclear binding assay was carried out as described.25 Briefly, aliquots of cytosol (0.5 to 1.0 ml) were labeled with 5 nM [3H]E2 in the absence or presence of DES for 16 hours at 0 C and treated with DCC to remove free radioactivity. Samples were then incubated with (to induce activation) or without (to maintain in the unactivated state) 0.4 M KCl, added in If10 volume. Monoclonal antibodies (50 pg) were then added to the test samples and kept on ice for an additional 6 to 16 hours. Controls received buffer or equal amounts of unrelated MoAbs against rat heart fatty acid-binding protein. All samples were then diluted with buffer TGET/Mo to reduce the KC1 concentration from 0.4 to 0.1 M, a concentration shown not to interfere with nuclear or DNA binding of r3H]E2R in vitro.2s One volume of diluted labeled cytosol was reconstituted with one volume of nuclear suspension (125 to 250 pg DNA) and incubated for 1 hour at 0 C with frequent blending on a vortex mixer. The nuclei were then separated by centrifugation at 1,000 x g for 198

Steroids,

1990, vol. 55, May

10 minutes at 2 C and washed three times by resuspending the pellets in 3 ml of ice-cold buffer followed by centrifugation. Some of the nuclear pellets were extracted with 2 ml of absolute ethanol for 16 hours at room temperature, and the total extract (residue and ethanol) was transferred to scintillation vials, scintillation tluid was added, and radioactivity was counted. The rest of the nuclear pellets were subjected to extraction with buffer containing 0.4 M KC1 for 1 hour at 0 C, then centrifuged at 100,000 x g for 30 minutes. The saltextracted [3H]EzR complexes were further analyzed by sucrose density gradient (SDG) after reconstitution with or without MoAb. Nuclear-bound [3H]EzR was calculated by subtracting the binding to nuclei obtained with cytosol which had been labeled with [3H]E2 in the presence of DES (nonspecific binding) from that obtained with cytosol labeled with [3H]E~ only. DNA-cellulose-binding [3H]E2R complexes

assay of activated

The binding of i3H}E2R to DNA-cellulose was carried out as described previously.25 DNA-celIulose was suspended in buffer TGET containing 1 M KC1 and 0.02% sodium azide for 16 hours at 0 C. The KC1 was then removed from the resin by five washes in buffer without KCl. The washed DNA-cellulose resin was then resuspended in buffer TGET/Mo and dispensed into siliconized test tubes (100 pg DNA equivalent/sample) and used for binding assay. The binding of [3H]EzR complexes to DNA-cellulose was carried out as described for the nuclear-binding assay. Synthesis and purijication

of peptides

Peptides (AT,, AT2, and AT3) were synthesized as described elsewhere.Z7s28 The synthetic peptides were purified by gel filtration and analyzed by high-performance liquid chromatography. An analysis of the amino acid composition correlated well with the primary sequence. Each peptide contained one 3H-labeled amino acid as a tracer. This provided a means of determining the efficiency of coupling to the carrier protein. Conjugation hemocya~in

of the peptides to keyhole limpet and bovine serum albumin

Keyhole limpet hemocyanin (KLH) and bovine serum albumin (BSA) were dissolved in PBS to give a final concentration of 1 mg/ml. One hundred milligrams of each peptide were then dissolved in 10 ml of the KLH solution and 50 mg was dissolved in 5 ml of the BSA solution. The pH of the mixtures was adjusted to 9 with 0.1 M LiOH. The coupling of the peptides to the carrier proteins was initiated by dropwise addition of 6.25% glutaraldehyde to achieve a final concentration of 1%. Each mixture was then incubated at 0 to 4 C for 1 hour with gentle agitation. Aliquots (50 to 200 ~1) were then removed and used to determine total pep-

Monoclonal

tide concentration by radioactivity counting. The remainder of each mixture was then transferred to dialysis tubes and dialyzed extensively against four changes of PBS. Aliquots were removed after dialysis and counted to determine the efficiency of coupling. The remaining dialysate was divided into l-ml aliquots and frozen at -80 C until needed.

Immunization Three groups of female mice (five animals per group), 6 to 8 weeks of age, were immunized by subcutaneous injection with 100 pg of the designated peptides emulsified in Freund’s complete adjuvant. Two booster injections were given at 3-week intervals. Serum samples were collected 7 days after immunization and tested for hER-specific antibodies by SDG analysis.27 Immunized mice were injected intraperitoneally 1 month later with 100 pug of the appropriate antigen in PBS; after 3 days, their spleens were removed and used for fusion. Cells and media Cultures of mouse myeloma cell line Sp2/0 were maintained in Dulbecco’s modified Eagle’s medium (DME) (GIBCO, Grand Island, NY, USA) supplemented with fetal calf serum (FCS, 20%) (Hazelton, Lenexa, KS, USA), penicillin (10 U/ml), streptomycin (100 pg/ml), nonessential amino acids (GIBCO), and glutamine (580 pg/ml). After fusion, the cells were plated in DME supplemented with hypoxanthine (0.1 mM), aminopterin (0.4 PM), thymidine (3 PM) and concanavalin A (ConA)-conditioned medium (10%) (HAT medium). Dulbecco’s modified Eagle’s medium supplemented with hypoxanthine and thymidine (HT medium) was also prepared. Concanavalin A-conditioned medium was obtained by incubating BALB/c mouse spleen cells (3 x lo6 cells/ml) with ConA (2 pglml) at 37 C in DME under 5.6% CO2 for 24 hours. The tissue culture supernatants were separated by centrifugation and stored at 4 C.

Cell fusion Cell fusion was carried out by the method of MarshakRothstein et a1.29Briefly, spleens were excised, fat and mesenteric tissues were removed quickly, and a single cell suspension was made by squeezing the spleen between two glass slides in Hank’s balanced salt solution buffered with 0.01 M phosphate, pH 7.2. Red blood cells were lysed by brief incubation in ammonium chloride lysis buffer. Spleen cells (5 x lo7 cells) were mixed with Sp2/0 cells (5 x lo6 cells) in round-bottom tubes and pelleted at 700 X g for 5 minutes at 22 C. The cells were resuspended in serum-free DME and centrifuged. After removal of the supernatants, the cell pellets were resuspended for 6 minutes in 0.5 ml of polyethylene glycol 1500 (PEG, 30% vol/vol) (Baker Chemical Co., Phillipsburg, NJ, USA), followed by addition of 4 ml of serum-free DME to dilute out the PEG. The cell suspensions were transferred to petri

antibodies

to human

estrogen

receptors:

Traish et al.

dishes (100 x 17 mm) and DME containing 20% FCS was added. The cultures were incubated at 37 C for 24 hours under 5.6% CO*. The cells were then pelleted and resuspended in HAT-conditioned medium (1 x lo6 cells/ml). Aliquots of the cell suspension (0.1 ml) were dispensed into 96-well flat-bottom microtiter dishes and incubated at 37 C. Seven days later, the cultures were fed with 0.1 ml of conditioned media (DME, HT). After 2 more days, clones were screened by enzymelinked immunosorbent assay (ELISA) on BSA-conjugated peptides.

Cloning by limiting dilution Cells from wells that tested positive by ELISA were cloned by limiting dilution. Cells were diluted to 1,0.3, and 0.1 cell equivalent/l00 ~1 in DME containing 20% FCS and BALB/c peritoneal exudate cells (5 x 104 cells/ml) and then plated in 96-well microtiter plates. After 10 days, wells with single clones were identified by microscopic examinations and tested for the presence of antibodies by ELISA. Clones that tested positive were expanded in large flasks; spent media were then collected and the cells either were used for ascites production or frozen for later use.

Ascites production Female mice (BALB/c x A/J)FI were injected intraperitoneally with pristane (0.5 ml) and, 7 days later, the animals were injected with 1 x lo6 hybridoma cells in 0.2 ml PBS. Ascites fluid was collected 7 to 10 days later by insertion of a needle into the peritoneal cavity. The fluid was clarified by centrifugation at 700 X g for 10 minutes, divided into several vials and kept frozen at -70 C until use.

Immunoprecipitation of 13H]E2R by monoclonal antibodies Aliquots of cytosol labeled with C3H]E2were treated with DCC, then incubated at 0 C for 16 hours with normal mouse serum (10 pl), ascites fluid (10 pl), or tissue culture supernatant (200 ~1). All samples received 0.1 volume of buffer containing 4 M KCl. At the end of the incubation, rabbit anti-mouse IgGi (250 ~1) was added and samples were kept for an additional 4 hours at 0 C. To determine nonspecific binding, parallel incubations were made with cytosol, labeled with [3H]E2 in the presence of excess unlabeled DES, and treated with DCC. The precipitates were collected by centrifugation and washed three times with PBS; they were then extracted with 1 ml of absolute ethanol. The total mixture was transferred to counting vials, scintillation fluid was added, and radioactivity was counted. To further ensure that the immunoprecipitated [‘H]E2R was a result of ER binding to the antipeptide antibodies, we incubated cytosol, labeled with [3H]E2 in the absence or presence of unlabeled DES, with a MoAb against rat heart fatty acid-binding protein (obtained from Dr. P. Brecher, Boston University School of Medicine) and an antifluoranthene MoAb (provided Steroids,

1990, vol. 55, May

199

Papers

by Dr. J. Groopman, Boston University School of Public Health). The nonspecifically precipitated radioactivity was subtracted from total bound radioactivity immunoprecipitated by the antipeptide antibodies.

Coupling of monoclonal antibodies CNBr-activated sepharose

Test for antibody class The isotype of each MoAb was determined by ELISA. Microtiter plates coated with the immunogenic peptide were incubated with aliquots of the spent media from the hybridoma clones. Bound antibody heavy chain class was determined by the addition of goat antimouse isotyping reagents (Southern Biotech, Birmingham, AL, USA) diluted 1 :I,000 in PBS-0.2% BSA, followed by alkaline phosphatase-conjugated rabbit anti-goat secondary antibody and the substrate.

Results Figure 1 depicts the amino acid sequence thought to represent the putative DNA-binding domain of the hER. The three peptides used for antibody production are represented by the underlined amino acids. The nine conserved cysteine residues, noted by asterisks, may play a role in the tertiary structure of the putative zinc-binding fingers.i4 N-terminals of all three peptides were kept acetylated, while the carboxyl terminals remained amidated. Peptides linked to KLH were used

REGION OF RECEPTOR 111111111111111111

90 kD HSP r,,,,,,,,,I n

Dlmenzanon TlnNac@,vr.mnNuclearrnN!nCIpDII

’

EAVENDYASGYHYGYW-

”,,,,,,,,,,,,,,,,,,N,,,,N,,,,,,

n

-

------I

SEEGEKAPPKRSIQGHNDYMEPA~Q~~DKNRRKSEQA~RLRK~Y~VGMMKGGIRKDRRG peptide 1

Peptide 2

Pepudc 3

Figure 1 The predicted amino acid sequence of the DNA-binding domain of the hER from MCF-7 human breast cancer cells is depicted. The underlined amino acids represent the sequences chosen to synthesize the peptides used for immunization. The cysteine residues with the asterisks represent those thought to be conserved and may be involved in the formation of the putative zinc-binding fingers.

200

Steroids,

1990, vol. 55, May

4.6 S

to

Monoclonal antibodies in the ascites fluid were precipitated with 40% ammonium sulfate and coupled to CNBr-activated Sepharose gel (Pharmacia) according to the manufacturer’s instructions. The antibody-coupled Sepharose was washed with 0.1 M acetate/O.5 M NaCl buffer, pH 4, followed by Tris-HCl buffer/O.5 M NaCl, pH 8. The resin was resuspended in PBS containing 0.02% NaN3 and stored at 4 C.

FUNCTION

-

m

1.o

w”

0.5

I P)

0.0 0

10

FRACTION

20

30

40

NUMBER

Figure 2 Interactions of antisera from mice immunized with peptides AT,, AT?, and AT3 with ER. Calf uterine cytosol was prepared and incubated for 3 hours at 0 C with 5 IJM PH]E2 with or without 5 ELMunlabeled DES in buffer TesiGME (0.02M N-Tris hydroxymethyl-methyl-2-amino ethane sulfonic acid, 10% voli vol glycerol, 10 mM sodium molybdate, and 1 mM EDTA, pH7.4, at 2 C). After DCC treatment, samples of the incubations (0.15 ml) were mixed with 50 ~1 of the preimmune serum or the antisera against the peptides. Fifty microliters of 2 M KCI were added to each sample to give final concentrations of 0.4 M, and all samples were kept on ice for an additional 4 hours. The samples were then analyzed on SDG prepared in the same buffer containing 0.4 M KCI. Open circles represent control ER only, closed circles represent ER incubated with preimmune mouse serum, open triangles represent ER incubated with antiserum against peptide AT,, solid triangles represent antiserum against peptide AT,, and open squares represent antiserum against peptide AT,.

as immunogens

and those linked to BSA were used for screening assays. Antisera from immunized (BALB/c x A/J) Fl mice were tested for their ability to increase the sedimentation rate of ER in SDG containing 0.4M KCl. Only antisera from animals immunized with peptide AT3 (amino acids 247-261) tested positive against native ER (Figure 2). Because of the intent to screen for antibodies that recognize the native ER, only those animals that produced antibodies reactive with the native ER were used for the development of MoAbs. After cell fusion, the tissue culture supernatants from the various clones were initially screened for MoAbs against the peptide conjugated to BSA (AT3BSA). Fifteen clones derived from two separate fusions appeared to contain MoAbs against peptide AT3. Further assessment of these antibodies for their ability to bind [3H]E2R by immunoprecipitation showed that clones 16, 33, 36, 35, 114, 213, and 318 recognized the native ER molecule. Of the clones tested for the ability to increase sedimentation coefficient of ER by SDG, only four (16, 33, 114, and 213) bound to ER with high aflinity (Figure 3A through F), as demonstrated by the increased sedimentation value of ER from 4-5s to 7S in the presence of 0.4 M KC1 (nonequilibrium conditions). Three others (36, 35, and 318) bound to the native ER complex with low affinity, since they did not cause significant changes in the sedimentation of the

Monoclonal 76

B

antibodies

to human estrogen receptors:

Traish et al.

4.6.6

11

12 ~

t 0

I 14 i*.

6 3

f t 1'r\5 1 i

0

w

0

0

10

20

30

9

0

6

6

3

3

5’ n u6

3 00

10

20

30 0 0

10

20

30

FRACTION

0

10

20

30

Oo

10

20

36

NlJMBf3

Figure 3 Analysis of ER interaction with MoAbs produced against peptide AT3 by SDGs. Calf uterine cytosol was prepared in buffer TGET/Mo and labeled with 13H]E2at 0 C for 4 hours. To determine nonspecific binding, aliquots were labeled with [3H]E2 in the presence of 5 PM unlabeled DES. At the end of the incubation, free radioactivity was removed with DCC pellets, and samples were removed and placed in propylene microfuge tubes. Aliquots (50 to 100 pg equivalent of antibody) of the ascites fluid from the various clones were then added and the samples were reincubated at 0 C for 16 hours. Additional samples either were treated with an equal mixture of ascites from four clones or remained untreated (control). Samples were then analyzed on 5% to 20% SDGs made in TGET/Mo buffer containing 0.4 M KCI as described in the Experimental section. All data presented are corrected for nonspecific binding. The arrows represent the sedimentation of the 14C-labeled protein markers included in each gradient. Fraction number 30 represents the top of the gradient. (A) Mixture of 213,114,33, and 16, closed triangles; (B) 213, solid circles; (C) 33, solid triangles; (D) 114, open squares; (E) 16, open diamonds. Control (solid squares) is shown in A through F. (F) Antiprogesterone receptor antibody, open circles. (G) Incubation of ER with MoAb 320, open diamonds; MoAb 318, closed squares; MoAb 36, closed circles; no addition of antibody, closed triangles, (H) Incubation of ER with MoAb 31, open squares; MoAb 110, upward closed triangles; MoAb 35, inverted closed triangles; closed diamonds represent a mixture of MoAbs 110,35, and 31.

native ER on SDG (Figure 3G and H). All but one of these MoAbs were IgGi antibodies, while 67 was an IgM antibody. An equal mixture of MoAbs 213, 114,33, and 16 did not result in any additional increase in the sedimentation rate of ER when compared with the increase obtained by any individual MoAb’, indicating that these antibodies bound to the same epitope on ER complexes (Figure 3A). To verify that the increased sedimentation rate observed in Figure 3 with the antipeptide MoAbs was due to specific interaction with ER, we incubated labeled cytosol under identical conditions with MoAb (IgGi, clone 414) raised against a peptide representing amino acids 533-547 of human progesterone receptor (unpublished data). As shown in Figure 3F (open circles), there was no increase in the sedimentation rate of ER in the presence of MoAb against the human progesterone receptor. Similar results were obtained with the MoAbs against rat heart fatty acid-binding protein and fluoranthene (data not shown). These results suggest that the antipeptide MoAbs are specific for ER. These results also suggest

that the four MoAbs are site-specific. To test this premise, we incubated [3H]E2R with MoAbs alone or with MoAbs preincubated with 50 pg of the peptide AT+ As shown in Figure 4, [3H]E2R alone sedimented as a 4.6s complex on SDG/0.4M KCl, while in the presence of MoAbs, the [3H]E~R complexes sedimented in the 7-8s region of the gradient. When the MoAbs were preincubated with the peptide, they did not increase the sedimentation coefficient of [3H]E2R, suggesting that the peptide had occupied the antibodybinding site. Experiments in which [3H]E2R was first incubated with the peptide and then reincubated with the antibodies produced results similar to those obtained when the antibodies were first incubated with the peptide. This is likely due to the presence of much greater amounts of peptide compared with [3H]E2R. Incubation of [3H]E2R with free peptide and subsequent analysis on SDG/0.4M KC1 did not alter the sedimentation of ER. Thus, the peptide had no effect on ER. These observations reinforce the suggestion that these MoAbs are site-specific with respect to the domain-spanning amino acids 247-261 of the ER. Steroids,

1990, vol. 55, May

201

Papers

8 i

6

6

0

to

20

30

FRACTION

10

0

20

30

NUMBER

Figure 4 Site specificity of the MoAbs as determined by SDG. Calf uterine cytosol was prepared and labeled as described in Figure 3. Aliquots of the MoAbs to be tested (100 pg) were first incubated at 0 C for 5 hours in the absence (control) or presence of 100 pg of ATs peptide in PBS or with an aliquot of calf uterine ER complexed to unlabeled E2 (200 fmol]. At this point, all samples received an aliquot of the [3H]E2-labeled cytosol and were kept at 0 C for 16 hours. Antibody-receptor interactions were analyzed by SDG, as described in Figure 3. (A) MoAb 213, (8) 33, (C) 114, and (D] 16. Open circles represent sedimentation of [3H]E2R only; solid triangles represent sedimentation of the [3H]E2R incubated with antibodies bound to the peptide. Solid diamonds represent the sedimentation of [3H]E2R incubated with antibodies preincubated with buffer only.

Several studies have suggested that the ER protein has two conserved regions, namely, the DNA-binding region (region “C”) and the steroid-binding region (region “E”). Thus, it seemed likely that the MoAbs produced against the DNA-binding domain of hER might interact with ER from other species. Figure 5 shows that when the L3H]E2R complexes derived from different species were incubated with MoAbs and analyzed by SDG10.4 M KCl, an increase in the S values was observed, suggesting that [3H]E2R from these species cross-reacted with the MoAbs. It has also been proposed that the DNA-binding domain of receptor proteins in the steroid hormone family share 42% to 95% homology with each other.’ Thus, production of MoAbs to this region of hER could result in MoAbs that might cross-react with progesterone, glucoco~icoid, and androgen receptors. To test this possibility, MoAbs were incubated with progesterone receptors from calf uterus labeled with [3H]ORG 2058 in buffer without molybdate and ana202

Steroids,

1990, vol. 55, May

lyzed in SDG containing 0.4 hf KCI. No increase in the sedimentation coefficient was observed, demonstrating that these MoAbs did not cross-react with the native form of the progesterone receptor. Similarly, when rat prostatic androgen receptors, labeled with [“HIDMNT, or glucocorticoid receptors from calf uterine tissue, labeled with [>H]DEX, were incubated with MoAbs and analyzed in SDGI0.4 M KCl, no increase in the sedimentation coefficient of these receptor proteins was observed (Figure 6). Under identical conditions, however, these MoAbs bound to and increased the S value of the [‘H]E2R complexes (Figure 3). These data strongly suggest that these MoAbs are specific for the native form of ER and do not bind to the native forms of other steroid receptors. When amino acid sequences of hGR, hAR, and hPR corresponding to that of peptide AT3 of hER were aligned and compared (Table I), no striking homology was observed with hER, indicating that the lack of cross-reactivity with other receptors is due to the lack of homology in this sequence. We then examined the interactions of these MoAbs with [3H]EzR under conditions in which the concentration of the MoAbs was kept constant while the concentration of [“H]EZR was increased, At all concentrations

A 7s

5

I

4.85 I

JS

4.6s

1

1

Calf UtsnAa

4

6

3

0 0

10

20

30

FRACTION

0

10

20

30

NUMBER

Figure 5 Interaction of MoAbs prepared against peptide from hER with [3H]EzR from various mammalian species. Cytosols from human breast cancer tissue (A), calf (B], rat K), and mouse fD] uteri were labeled with f3H]E2 for 16 hours at 0 C. After removal of free radioactivity with DCC, samples were incubated without (control, open triangles] or with ascites from clones 213 (solid circles], 114 (open squares), 33 (solid triangles), and 16 (open diamonds). After 16 hours at 0 C, the samples were analyzed by SDG as described in Figure 3.

Monoclonal

antibodies

to human estrogen receptors:

Traish et al.

2.5

-

0

20

10

30

0

10

FRACTION

20

30

0

10

20

30

NUMBER

Figure 6

Receptor specificity of MoAbs prepared against a peptide from hER. Progesterone receptor of calf uterine cytosol was labeled with [3H]ORG 2056, glucocorticoid receptor was labeled by incubating calf uterine cytosol with [3H]DEX, and androgen receptor of rat ventral prostatic cytosol was labeled with [3H]DMNT. Aliquots of these receptor preparations were then incubated at 0 C for 16 hours with an aliquot of each of the ascites from the various clones. As control, samples were incubated with buffer only. All samples were then analyzed on SDG as described in Figure 3. (A) Progesterone receptor, (B) androgen receptors, (C) glucocorticoid receptor. Symbols are identical to those in Figure 5.

presence of 0.4 M KCl, the bound [3H]E~R was determined after washing the Sepharose-MoAb resin three times with ice-cold buffer; the bound radioactivity was extracted and counted. Increasing concentrations of ER bound to unlabeled DES competed effectively for [3H]E2R binding. These data were then analyzed according to the isotope dilution method30 after correction for nonspecific binding. The estimated equilibrium dissociation constants, as calculated from Scatchard plots,31 ranged from 0.4 nM to 1.8 nM, suggesting that the MoAbs have high affinity for the receptor (data not shown). Similarly, since SDG requires 16 to 20 hours of centrifugation under nonequilibrium conditions, the stability of the [3H]E2R MoAb complexes in these experiments further points to the high affmity of the MoAbs. The antibodies produced in this study are site-directed to a domain encompassing amino acids 247-261 of hER. This region is thought to be a part of the DNAbinding domain and presumably exposed on the surface of the protein due to its hydrophilicity.6 Therefore, one would expect these site-specific antibodies to react with the native 8s ER complexes. To test this premise, we incubated [3H]EtR complexes, prepared

of ER tested (20 to 1,000 fmol), all four MoAbs caused an increase in the S value of [3H]EzR (data not shown). We have also tested the ability of MoAb 213 to bind to the unoccupied ER. The latter was incubated with the MoAb and analyzed on SDG containing 0.4 M KCl. The fractions collected were incubated with r3H]E2, in the absence or presence of unlabeled DES, and the specifically bound radioactivity was determined and compared with the SDG profiles of [3H]E2R, or [3H]E2R preincubated with the same MoAb. As shown in Figure 7, MoAb 213 bound to the unoccupied receptor. Since incubation with antibody and the subsequent analysis on SDG were performed with the unoccupied receptor form, the partial loss of ER observed was most likely due to the known instability of the ER in the unoccupied state. To determine the apparent equilibrium association and dissociation constants, we first conjugated each MoAb with Sepharose and then incubated them with [3H]E2R in the absence (total binding) or presence of increasing concentrations of DES-bound ER. Nonspecific binding was determined by using Sepharose conjugated to MoAb against the rat heart fatty acid-binding protein. After incubation for 16 to 20 hours in the

Table 1 Amino acid sequences hER

C

Y

E

V

hGR hAR hPR

C

L

Q

A A A

:

:

G

M

M

K

M

N T V

L L L

:

G

G

I

R

K

N

R

R

G

R R R

K K K

T L F

K

K

K L F

I

: G

A A G

I:

Amino acid sequences of subregions of C-terminal of the DNA-binding domain of hER, GR, AR, and PR starting with the last conserved cysteine residue in the DNA-binding domain. 1~7*8,10.1* The underlined sequence of hER represents peptide AT3, to which antibodies have been raised.

Steroids,

1990,

vol.

55, May

203

Papers

2

I

7S

4.6s

1

1

since there was no demonstrable increase in the S value of ER. These data clearly suggest that the antibody-binding site in the 8s ER complex is either masked (inaccessible) by binding of other proteins (e.g., heat shock protein) to this site, or the epitope conformation is presented to the antibody as a nonbinding site. The possibility remained, however, that the 8s ER-antibody complexes dissociated during centrifugation (18 hours, 2 C) due to the lower affinity of interaction of these MoAbs with 8s ER complexes. To test this, we incubated [3H]EzR for 16 hours at 0 C in the presence of 10 mM molybdate (without KCI) with the MoAbs coupled to Sepharose. No immunoprecipitable C3H]E2R complexes were observed, suggesting that these antibodies did not react with the untransformed ER complexes (data not shown). As shown in Figure 8B, when [‘H]E2R complexes prepared in buffer containing 10 mM sodium molybdate were incubated for 16 hours at 0 C with 500 pg of each antibody in 0.4 mM KC1 and analyzed on SDG containing 0.4 M KCI, an increase in the sedimentation rate of the ER complexes from 4.6s to 7s was observed. These data suggest that KC1 (0.4 M) induced the dissociation of the unactivated 8s ER complexes into the activated 4.6s species, which were recognized by the MoAbs. Thus, it is possible that the dissociation of the 8s ER complex into the 4.6s subunit exposes the antibody-binding site or presents it as a high affinity site with respect to antibody interaction. From previous experiments, it appears that the MoAbs did not react with [3H]E2R during the incubation at 0 C in the absence of KC1 and in the presence of molybdate. Thus, the observed antibody-recentor interactions must ha ve taken place during centrifugation when the

I

R

6

FRACTION

NUMBER

Figure 7

Interaction of MoAb 213 with unoccupied ER. Calf uterine cytosol (solid circles) was incubated in the presence of MoAb 213 for 16 hours at 0 C. Another aliquot of the cytosol was first labeled with f3H]E2, then incubated in the absence (open squares) or presence (open triangles) of MoAb 213. Samples were layered on SDGs as described in Figure 3. The fractions from the unoccupied ER sample were then reincubated at 0 C for 16 hours with 13HlE2 in the absence or presence of 5 FM DES. Bound radioactivity was determined by DCC.

and labeled in buffer TGET in the presence of 10 mM molybdate, with or without (control) 50 pg of each antibody. Subsequent analysis of the reaction mixture on SDGs prepared in the same buffer but without 0.4 M KC1 did not reveal anv interaction of the antibodies with nonactivated (8s) ER complexes (Fig ur e g-41,

rA

75 4.6s 1

1

C

B 7s 4.6s 1

1

7s 4.6s 1 .

1

w

FRACTION Figure 8

NUMBER

Interaction of antipeptide MoAbs with the untransformed 8s ER and the salt-activated 4s ER. Calf uterine cytosol prepared in buffer TGET containing 10 mM sodium molybdate was incubated at 0 C for 16 hours with 5 nM 13H]E2 in the absence or presence of unlabeled DES. After treatment of the incubations with DCC, the samples were divided and incubated with the MoAbs as specified. (A) Aliquots of the labeled cytosol were reincubated at 0 C for 16 hours without (control, solid triangles) or with 50 pg of MoAb 213 (solid circles). MoAb 114 (open squares), MoAb 33 (upward solid triangles), or MoAb 16 (open diamonds). All samples were then analyzed on SDG in buffer with molybdate but without KCI. (B) Aliquots of the labeled cytosol were incubated with the MoAbs as in panel A, but in the presence of 0.4 M KCI. The samples were analyzed on SDG containing molybdate and 0.4 M KCI. Only half of the control sample was layered on the gradient. The symbols are identical to those in panel A. (C) Aliquots of the labeled cytosol were layered simultaneously (without preincubation) with the indicated MoAbs on SDG in buffer containing molybdate and 0.4 M KCI. The symbols are identical to those in panel A.

204

Steroids,

1990, vol. 55, May

~o~o~~on~i

ER complexes were exposed to 0.4 M KC1 during centrifugation. To test this possibility, we layered l’H]EzR complexes, prepared and labeled with [3H]E2 in the presence of 10 mM sodium molybdate, with the MoAbs on SDG prepared in buffer containing 0.4 M KC1 and 10 mM molybdate. As shown in Figure 8C, the ER-antibody complexes sedimented at a greater rate than the receptor alone. We suggest that the MoAb interacted with the ER complex during centrifugation since no preincubations of ER with the antibodies occurred. These observations suggest that in the gradient, the 8s ER dissociates in the presence of KC1 into 4S ER, which binds to the MoAb. The MoAbs also bound to PH]EzR which was prepared in buffer without molybdate and heat-activated at 28 C for 30 minutes to transform ER into the 5S complex (data not shown). Several antipeptide polyclonal region-specific antibodies to steroid receptors have been produced which inhibit the binding of activated receptor complexes to DNA.32-35To test if the MoAbs produced in this study inhibit binding of salt-activated r3H]E2R complexes to nuclei and DNA-cellulose, two experiments were carried out (Figure 9). [3H]E2R complexes were prepared in buffer without sodium molybdate. Aliquots of these ER complexes were kept at 0 C (unactivated) in the absence or presence of MoAb 213; additional aliquots were incubated in the absence or presence of 0.4 M KC1 (activation) with or without 50 lug of MoAb 213. After 16 hours at 0 C, the samples were diluted with buffer containing 10 mM sodium molybdate to reduce the KC1 concentration to 0.1 M and then incubated with nuclear suspension or DNA-ce~ulose for 1 hour at 0 C. The nuclear or DNA-cellulose-bound [3H]EzR complexes were assayed as described in the Experimental section. As shown in Figure 9A (open bars), nonactivated [3H]E2R complexes bound poorly to nuclei (5% to 8%); incubation of the nonactivated ER complexes with MoAb 213 did not interfere with binding to nuclei (Figure 9B). KCl-activated [3H]E2R complexes bound appreciably to nuclei (40% to 45%) (Figure 9C). KCl-activated 13H]EzR, incubated with MoAb 213, bound to nuclei to the same extent as the KCl-activated ER in the absence of antibody. These observations suggest that MoAb 213 did not inhibit the binding of activated ER to nuclei. Even increasing the concentration of the MoAb to 75 hg (Figure 9E) or 100 pg (Figure 9F) did not produce any inhibition of ER binding to nuclei. Thus, it is unlikely that the MoAb concentrations used were not sufficient to inhibit ER binding to nuclei, since these concentrations were shown to complex all the ER on SDG, Identical results were obtained with the other three MoAbs (16,33, and 114) (data not shown). One can argue that the binding of [3H]E2R to nuclei in vitro may be disrupted by the interaction of [3H]E~R with nuclear proteins, in addition to DNA. Thus, we used DNA-ce~ulose as the acceptor for the activated ER. As shown in Figure 9 (cross-hatched bars), no inhibition of activated ER binding to DNA-cellulose was observed in the presence of the MoAb. These data

a~tjbo~~es to human estrogen receptors:

Tr~ish et al. -I

6TilDNA-Cellulose

A

B

C

D

E

F

figure 9 Effects of antipeptide MoAbs on interaction of saftactivated ER complexes kiih nuclei and DNA-cellulose. Nonactivated (A and B) and KCI-activated IC throuah F) f3HIE9R complexes were incubated for 16 hours at 0 C wkhout‘(A’and C) or with 50 Fg (B and 0). 75 pg (E), or 100 pg (F) of MoAb 213. All samples were then diluted with buffer TGETlMo to reduce the KCI concentration and incubated with nuclei (open bars) or DNA-cellulose as described in the Experimental section. The specifically bound rad~oa~ivi~ was determined after correction for nonspecific binding and expressed as percent of specifically added 13H]E2Rcomplexes. Each point is carried out in triplicate, and the data are the average of two separate experiments.

are in contrast to those obtained with polyclonal antibodies to estrogen35 and other steroid receptors.32-34 It is possible that on dilution of the cytosol incubation to reduce the KC1 concentration, the antibody dissociated from the ER, thus allowing nuclear and DNA binding. To test this possibility, we analyzed the diluted cytosol by SDG in the presence of 0.4 M KCl. As shown in Figure lOC, the MoAb remained bound to ER subsequent to cytosol dilutions. Whether the ~3H]E2R-MoAb bound to nuclei and DNA as a complex or whether the [3H]E2R-MoAb complex dissociated, releasing the antibody on binding of ER to nuclei or DNA, needed to be investigated. Therefore, [3H]E2R complexes were incubated for 16 hours at 0 C with 0.4 M KC1 in the absence or presence of 50 pg of MoAb 213. Samples were then diluted and incubated with nuclei for 1 hour at 0 C, and the nuclear suspension was centrifuged to remove unbound 13H]E2R. The nuclear pellets were then washed three times with buffer and extracted with buffer containing 0.4 M KCl. The KC1 extract was then divided and incubated for 3 hours with or without 50 pg of MoAb 213 and was subsequently analyzed on SDG. As shown in Figure IOA and B (solid circles), KCl-extractable [3H]E2R complexes sedimented as 5.6s complexes whether the labeled cytosol was incubated with (Figure 10B) or without (Figure 1OA) MoAb. To eliminate the possibility that the antibody may remain absorbed to the nuclei or that KC1 extraction might interfere with the subsequent analysis, we carried out Steroids,

1990, vol. 55, May

205

Papers 5 A 4.6

S

4.6

1

S

4.6

I

w”

tI

S

1

l

9

i

tc,

.

6

1

1

F r) -0

0 0

10

20

30

40

0

10

FRACTION

20

30

40

0

10

20

30

40

NUMBER

Figure IO Analysis of nuclear KCI-extractable ER complexes and cytosolic ER on SDG. (A) Aliquots of cytosol were labeled with estradiol, activated with KCI, diluted, and subjected to nuclear binding assay as described in Figure 9 (bar C). The nuclei were washed and extracted with buffer containing 0.4 M KCI. The extract was divided and reincubated for 3 hours at 0 C with (solid squares) or without (solid circles) 50 pg MoAb 213 and analyzed on SDG containing 0.4 M KCI. The nuclear-extractable ER sedimented as 5.6s in the absence of the antibody and as a 7s in the presence of the antibody. (B) Parallel aliquots of cytosol were labeled with estradiol, activated with KCI, and incubated with the MoAb as described in Figure 9 (bar 0). The nuclear-bound ER was extracted with 0.4 M KCI and divided; one sample was incubated without MoAb (solid circles) and the other sample was incubated with MoAb, as shown in panel A, and analyzed on SDG in 0.4 M KCI. (C) The sedimentation of the activated ER complexes which were incubated with MoAb 213 and analyzed on SDG containing KCI prior to (solid circles) or subsequent to threefold dilution with buffer (solid squares). Fraction 40 represents the top of the gradient.

control experiments in which the free peptide was added to the nuclear incubation to block any unbound antibody. The results obtained were identical to those shown in Figure lOB, suggesting that the antibody was removed during the nuclear washes. These observations suggest that the interaction of the [3H]E2R-MoAb complexes with nuclei altered the conformation of the ER, which affected the antibody binding site, causing the release of the antibody. To explore if this might be the case, rather than degradation of the antibody binding site (epitope), we incubated the nuclear KCl-extractable [3H]EzR complexes with MoAb 213 for 3 hours at 0 C. As shown in Figure 10A and B, the KClextractable [3H]EzR complexes bound to the MoAb, as demonstrated by their increased rate of sedimentation. These observations clearly suggest that the binding of ER to nuclei and DNA results in conformational changes sufficient to alter the binding site of the antibody to, or cause its release from, [3H]EzR complexes.

Discussion The published hER cDNA sequence has led to the deductions of the protein primary amino acid sequences.6T7 Based on this sequence, subsequent functional studies have revealed that various regions of the protein can be assigned specific functions. l6 One of the domains, designated as region “C,” spans amino acids 180-263 of hER and is thought to be the DNA-binding domain. I6 We have sought the development of site206

Steroids,

1990, vol. 55, May

specific MoAbs to the DNA-binding domain of the ER. Using a new approach, we synthesized several peptides with sequences identical to certain specific sequences of the hER. These peptides were then linked to KLH and used as immunogens to develop MoAbs to ER. Because analysis of the reactions involved in ER binding to DNA requires probes that recognize the native ER, these antibodies were selected for their ability to recognize the native form of ER. Using immunoprecipitation and SDG, the antibodies were characterized with respect to their ability to bind the functional ER. Although we used three peptides as immunogens, antibodies against native ER were detected only in hybridomas derived from animals immunized with peptide AT3. The MoAbs contained against this peptide bound to ER from human, calf, rat, and mouse tissues, suggesting that this particular domain of ER is highly conserved in many species. The DNAbinding domain of hER (amino acids 185-263) has some homology with steroid and other nuclear receptors.’ Since the homology among these proteins varies from 42% to 95%, and this region appears to be conserved, one would expect that MoAbs obtained against a specific peptide sequence from this region would cross-react with some of these receptors; nevertheless, the data clearly show that no cross-reactivity existed when these MoAbs were incubated with the native forms of progesterone, glucocorticoid, or androgen receptor. Seven MoAbs (16, 33, 35, 36, 114, 213, and 318)

~~n~c~onaf

recognized the native ER according to immunoprecipitation assays. Four of these MoAbs (16, 33, 114, and 213) remained tightly bound to ER in SDG containing 0.4 M KCl. This reflects on the high affinity of binding, since the resulting complex remained undissociated for 18 hours under nonequilibrium conditions. The antibodies appeared to share the same epitope, since an equal mixture of these antibodies did not significantly increase the S value of the antibody-ER complex when compared with single antibody-receptor complexes. The specificity of the MoAbs was tested by using three unrelated antibodies, namely, ~tifluoranthene antibody, anti-rat heart fatty acid-binding protein antibody, and anti-progesterone receptor antibody. None of these antibodies reacted with the ER under the conditions used. In contrast, four MoAbs against the peptide reacted with ER, as demonstrated by the markedly increased sedimentation rate. Site specificity was further confirmed by experiments in which the peptide was used as a competitor. All four antibodies were unable to bind to ER when incubated with the free peptide prior to the addition of labeled ER or when treated with labeled ER which was preincubated with free peptide. It appears that under the conditions used here, the antibodies, at a fixed concentration, recognized ER concentrations between 20 and 1,000 fmol. This is consistent with their high affinity for ER. The binding of the antibodies to ER is independent of steroid binding to ER under the conditions used. This suggests that these antibodies recognize the unoccupied as well as the occupied ER forms. The DNAbinding region of ER is rich in lysyl, arginyl, and cysteinyl residues6 Thus, this domain may be exposed on the surface of the protein due to its hydrophilicity.6 One would expect, therefore, that the site-directed MoAbs will interact with this domain in the various receptor forms. The data obtained suggested that MoAbs interacted only with the salt-activated (4s) and heat-transformed (5s) ERs (data not shown) but not with the unactivated (8s) ER. It appears, therefore, that the antibodies’ binding sites are inaccessible in the unactivated form of ER (8s oligomeric form) and become accessible subsequent to activation and dissociation of the ER into subunits. It should be noted that the effects of molybdate on 8s ER are reversible and can be overcome by high salt concentrations.*’ Thus, the interaction of MoAbs with ER in buffers containing both molybdate and KC1 is presumably due to the presence of the 4s ER. Under the conditions used, the ER sedimented as 4-4.6s complexes on SDG containing 0.4 M KC1 and 10 mM sodium molybdate. These data are consistent with our previous work.25 Thus, the binding obtained with MoAbs in the presence of molybdate and 0.4 M KC1 represents interaction with the 4s ER subunit during SDG analysis. The data presented show that this new approach allows generation of site-specific, high-affinity MoAbs that recognize the functional ER. We have investigated the inhibition of ER binding to nuclei and DNA by the site-specific antipeptide MoAbs. In contrast to data obtained with polyclonal

antjb~djes

to human estrogen receptors:

Traish et al.

antibo~es,32-~~ these MoAbs did not inhibit the binding of ER to nuclei and DNA. Several explanations of this observation are possible. For example, this domain may not be involved in DNA or nuclear binding and has no effect on the ER complex’s ability to interact with nuclei and DNA. Other studies have suggested that the DNA-binding region spans amino acids 185 to 251.t6 If this is the case, peptide ATs, which spans amino acids 247-261, has five amino acids in the DNAbinding region and the remaining 10 amino acids in the hinge region. Therefore, the antibodies may be directed to the amino acids in the hinge region and not to the DNA-binding region. This question is now under investigation. Another possibility is that this domain is part of the DNA-binding region and the interaction of the receptor with nuclei or DNA may induce conformational changes in this region, resulting in the release of the antibody from the ER. Other possibilities also exist, such as nonspecific binding of the ER-MoAb complexes to nuclei or DNA, or ditution of the cytosol, after activation of ER with KC1 to reduce the salt concentration, may have resulted in the release of the MoAb from ER. To examine the latter possibilities, extraction of the i3H]E2R with 0.4 M KC1 from nuclei incubated with [3H]E2R-MoAb followed by analysis of SDG demonstrated the presence of [3H]EzR not bound to the antibody. Thus, one can suggest that on binding of ER to nuclei, the MoAb dissociated from ER, probably due to conformational changes induced by such interactions. To rule out nonspecific binding of ER-MoAb complexes to nuclei, we compared the binding to nuclei of activated and nonactivated ER complexes with activated ER complexes which had been incubated with increasing concentrations of MoAb. We did not observe any increase in the binding of nonactivated ER, nor was there a difference in binding of ER when low (10 pg) or high (100 pug) concentrations of MoAbs were used in the incubation. As to the possibility that the antibody was released from the [3H]E2R on dilution of the cytosol, we have analyzed aliquots of the diluted cytosol which were preincubated with MoAb. That the antibody remained bound under these conditions was demonstrated by the presence of the 7-8s complexes on SDG containing 0.4 M KCl. To show that the antibody binding site remained intact subsequent to nuclear binding and extraction, we examined the ability of the [3H]E2R extracted from nuclei to bind the MoAb. The results obtained demonstrated that this domain was still intact, since it bound to the antibody. These observations point out the possibility that the binding of the ER to nuclei and DNA is accompanied by co~ormational changes leading to the release of the antibody from its binding site. This may be an important observation since it has never been shown that alterations in this domain of the protein must take place during binding to its acceptor sites. These new probes should be useful in future studies of receptor activation, receptor-DNA interactions, and intracellular localization of ER by immunocytochemistry. Steroids,

1990, vol. 55, May

207

Papers

Acknowledgments

18.

Supported by a grant from the Boston University Community Technology Foundation and aided by grant IN-97M from the American Cancer Society. References I. 2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13. 14.

15. 16.

17.

208

20.

Evans RM (1988). The steroid and thyroid hormone receptor superfamily. Science 240~889-894. Muller RE. Traish AM (1986). The role of Ivsyl, arginyl and sulfhydryl residues in estrogen receptor act&tion,hS-to 5s dimerization, and conversion of receptors from a state with low affinity into a state with higher affinity for estrogen. Ann NY Acnd Sci 464:202-217. Grody WW, Schrader WT, O’Malley BW (1982). Activation, transformation and subunit structure of steroid hormone receptors. Endocr Rev 3:141-163. Bailly A, LeFevre B, Savouret J, Milgrom E (1980). Activation and changes in sedimentation properties of steroid receptors. .I Biol Chem 255~2729-2734. DiSorbo DM, Phelps DS, Litwack G (1980). Probes of basic amino acid residues affect active sites of the glucocorticoid receptor. Endocrinology 106:922-929. Green S, Walter P, Kumar V, Krust A, Bonnert J-M, Argos P, Chambon P (1986). Human estrogen receptor cDNA seexpression and homology to v-erb A. Nature quence, 320:134-139. Greene GL, Gilna P, Watertield M, Baker A, Hort Y, Shine J (1986). Sequence and expression of human estrogen receptor complementary DNA. Science 231:1150-l 154. Hollenberg SM, Weinberger C, Ong ES, Cerelli G, Oro A, Lebo R, Thompson EB, Rosenfeld MG, Evans RM (1985). Primary structure and expression of a functional human glucocorticoid receptor cDNA. Nature 318:635-641. Arriza JL, Weinberger C, Cerelli G, Glaser TM, Handelin BL, Housman DE, Evans RM (1987). Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor. Science 232268-275. Misrashi M, Atger M, d’Aurio1 L, Loosfelt H, Meriel C, Fridlanskv F. Guiochon-Mantel A. Galibert F, Milgrom E (1987). Complete amino acid sequence of the humanprogesterone receptor deduced from cloned cDNA. Biochem Biophys Res Commun 143:740-748. Lubahn DB, Joseph DR, Sullivan PM, Willard HF, French FS, Wilson EM (1988). Cloning of human androgen receptor complementary DNA and localization to the X chromosome. Science 240:327-330. Chang C, Kokontis J, Liao S (1988). Molecular cloning of human and rat complementary DNA encoding androgen receptors. Science 240:324-326. Beato M (1989). Gene regulation by steroid hormones. Cell 56:335-344. Miller J, McLachin AD, Klug A (1985). Repetitive zinc-binding domains in the protein transcription factor IIIA from xenonus oocvtes. EMBO J 4:1609-1614. Chambon P (i988). Nuclear receptors as inducible enhancers. J Cell Biol 107:225. Kumar V, Green S, Stack G, Berry M, Jin J-R, Chambon P (1987). Functional domains of the human estrogen receptors. Cell 51:941-95 1. Greene GL, Gloss LE, DeSombre ER, Jensen EV (1979). Antibodies to estrophilin: comparison between rabbit and goat antisera. J Steroid Biochem 11:333-341.

Steroids,

19.

1990, vol. 55, May

21.

22.

23.

24.

25.

26.

27.

28. 29.

30.

31. 32.

33.

34.

35.

Greene GL. Gloss LE. DeSombre ER, Jensen EV (1977). Antibodies to estrogen receptor: immunochemical similarity of estrophilin from various mammalian species. Proc Nat/ Acud Sci USA 74~368 I-3685. Greene GL, Fitch F, Jensen EV (1980). Monoclonal antibodies to estrophiles: probes for the study of the estrogen receptor. Proc Nat1 Acad Sci USA 77: 157-161 Greene GL. Nolan C. Engler JP, Jensen EV (1980). Monoclonal antibodies to human estrogen receptor. Proc, Nufl Acud Sci USA 77:5515-5519. Borgna JL, Fauque J, Rochefort H (1984). A monoclonal antibody to the estrogen receptor discriminates between the nonactivated and activated estrogen and antiestrogen receptor complexes. Biochemistry 23~2162-2168. Fauque J, Borgna J-L, Rochefort H (1985). A monoclonal antibody to the estrogen receptor inhibits in vitro criteria of receptor activation by an estrogen and an antiestrogen. J Biol Chem 260:15547-15553. Traish AM, Muller RE, Wotiz HH (1986). Binding of 7a,17cudimethyl-l9-nortestosterone to androgen and progesterone receptors in human and animal tissues. Endocrinology 118:1327-1333. Muller RE, Sheard BE, Traish AM, Wotiz HH (1980). Effect of chemotherapeutic agents on the formation of estrogen receptor complexes in human breast cancer tissue cytosol. Cancer Res 40:2941-2942. Muller RE, Traish AM, Wotiz HH (1983). Estrogen receptor activation precedes transformation. J Biol Chem 258:92279236. Traish AM, Williams DF, Wotiz HH (1986). Binding of [‘HI70(.17a-dimethyl-19-nortestosterone (Mibolerone) to progesterone receptors. Comparison with binding of [‘HI R5020 and [‘HI ORG2058. Steroids 47:157-173. Traish AM, Kim N, Wotiz HH (1989). Characterization of polyclonal antibodies to preselected domains of the human estrogen receptor. Endocrinology 125:172-179. Merrifield RB (1963). Solid phase peptide synthesis. J Am Chem Sot 85:2149-2153. Marshak-Rothstein A. Fink P, Gridley T, Raulet DH, Bevan J, Geffer NL (1979). Properties and application of monoclonal antibodies directed against determinants of the Thy-l locus. J Immunol 122:2491-2497. Mukku VR (1984). Regulation of epidermal growth factor receptor levels by thyroid hormones. J Biol Chem 259:65436547. Scatchard G (1949). The attraction of proteins for small molecules and ions. Ann NY Acad Sci 51:660-673 _._. Smith DF, Lubahn DB, McCormick DJU, Wilson EM, Toft DO (1988). The production of antibodies against the conserved cysteine region of steroid receptors and their use in characterization of avian progesterone receptor. Endocrinology 122:2816-2825. Wilson EM, Lubhan DB, French FS, Jewel1 CM (1988). Antibodies to the steroid receptor deoxyribonucleic acid binding domains and their reactivity with the human glucocorticoid receptor. MO/ Endocrine/2:1018-1026. Urda LA, Yen PM, Simons SS Jr, Harmon JM (1989). Region specific antiglucocorticoid receptor antibodies selectively recognize the activated form of the ligand-occupied receptor and inhibit the binding of activated complexes to deoxyribonucleic acids. Mol Endocrinol3:251-260. Kim N, Ettinger R, Wotiz HH, Traish AM (1989). Site-directed polyclonal and monoclonal antibodies to the DNAbinding domain of the human estrogen receptor. 71st Annual Meeting of the Endocrine Society, June 21-24, Seattle, Washington (abstract 698).