the interaction site for tamoxifen aziridine with the ... · chromatography purification. after...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264, No. 23. Issue of August 15, pp. 13453-13459,1989 Printed in U.S.A.

The Interaction Site for Tamoxifen Aziridine with the Bovine Estrogen Receptor*

(Received for publication, April 10, 1989)

Thomas RatajczakS, Steven P. Wilkinson, Martine J. Brockway, Roland Hahnel, Robert L. Moritzt, Geoffrey S. Beggg), and Richard J. Simpsong) From the Endocrine Research Laboratory, Department of Obstetrics and Gynaecology, University of Western Australia, King Edward Memorial Hospital for Women, Subiaco, Western Australia 6008, Australia and the $Joint Protein Structure Laboratom. Ludwip Institute for Cancer Research and The Walter and Eliza Hall Institute for Medical Research, Parkville, Victoria ;OS,, Australia

Calf uterine estrogen receptor was covalently labeled with [3H]tamoxifen aziridine during affinity chromatography purification. After carboxymethyla- tion, affinity labeled receptor was digested with trypsin under limit conditions and the labeled peptides were fractionated by reversed-phase high performance liquid chromatography into one major and two minor components. Sequence analysis of the dominant labeled fragment indicated the facile cleavage of label during Edman degradation but identified two peptides, both derived from the extreme carboxyl terminus of the steroid-binding domain. The 17 residues of one peptide were fully conserved in all estrogen receptors. This fragment contained five nucleophilic amino acids and was considered as the more favored interaction site for tamoxifen aziridine. A corresponding region of the glucocorticoid receptor has recently been identified as one of three major contact sites for glucocorticoids (Carlstedt-Duke, J., Stromstedt2 P.-E., Persson, B., Cederlund, E., Gustafsson, J.-A., and Jornvall, H. (1988) J. Biol. Chem. 263,6842-6846). A comparison of amino acid physical characteristics in the hormone- binding domains of human estrogen and glucocorticoid receptors demonstrated an excellent structural correlation between the two regions and delineated elements in the estrogen receptor which may be directly involved in estradiol binding.

The amino acid sequences for estrogen receptors from a number of species have been derived from their corresponding receptor cDNAs (1-6). All estrogen receptors (M, = 66,000) have a structural organization similar to that initially described for human and chicken estrogen receptors by Krust et al. (3). The two functional regions, the DNA-binding domain and the steroid-binding domain, are highly conserved in all of these proteins (1-6). Deletion studies have defined a stretch of approximately 250 amino acids, toward the COOH terminus, as essential for the formation of the hydrophobic cavity involved in hormone binding (7). The size of this domain suggests the requirement of a correct tertiary structure for estrogen binding activity.

* This work was supported in part by the National Health and Medical Research Council of Australia. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should he addressed.

For the human estrogen and glucocorticoid receptors, considerable amino acid sequence similarity (30% identity) exists between the hydrophobic regions thought most likely to contain the hormone-binding domain (3, 8). Within the steroid- binding domain of the glucocorticoid receptor three amino acids, which interact with the steroid, have recently been identified by radiosequence analysis of photoaffinity labeled and affinity labeled receptor-hormone complexes (8-10). Pho- toinduced coupling located the interaction site for the A-ring of the steroid to 2 residues (methionine and cysteine) which, although proximal in the steroid-binding mode, are widely spaced (134 amino acids) along the polypeptide chain (8). The results suggest that the protein folds to accomodate this favored binding conformation (8). A cysteine residue, positioned between these photoaffinity labeled amino acids, was identified by alkylation with dexamethazone 21-mesylate (11) as the interaction site for the D-ring of the steroid (8-10).

The non-steroidal antiestrogen tamoxifen aziridine’ is an efficient labeling reagent for estrogen receptors (12). Evidence suggests that antiestrogens may compete allosterically rather than directly with estradiol for the hormone-binding site (13). An alternative mode of action may result in antiestrogen occupying the binding cavity but interacting with additional features in the hormone-binding domain to give a receptor with different conformational properties to that induced by estradiol (14). Despite the uncertainties in the mechanism of action of antiestrogens the highly selective, covalent interaction between tamoxifen aziridine and estrogen receptors has been of considerable utility in studies of receptor structure and function (15-19).

Two recent studies on the structural analysis of the hormone binding region of the estrogen receptor have used con- trolled proteolysis of [3H]tamoxifen aziridine-labeled receptors (17, 18). The attachment site for the reagent was shown to be restricted to a small region of the protein near the carboxyl terminus (17). In this paper we describe an extension of these studies and report on the sequence analysis of [3H]

’ The abbreviations and trivial names used are: tamoxifen aziridine, (Z)-1-(4-[2-N-aziridinyl)ethoxy]phenyl]-l,2-diphenyl-l-butene; dexamethazone 21-mesylate, 1,4-pregnadien-9a-fluoro-l6a-methyl- ll~,17a,21-triol-3,20-dione 21-methylsulfonate; triamcinolone acetonide, 1,4-pregnadien-9a-fluoro-ll~,l6a,l7~,Zl-tetrol-3,2O-dione-l6, 17-acetonide; SDS, sodium dodecyl sulfate; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; HPLC, high performance liquid chromatography; hER, human estrogen receptor; bER, bovine estrogen receptor; PTH, phenylthiohydantoin; hsp 90, heat shock protein of M, = 90,000; TRTPCK, ~-l-(tosylamido)-2- phenylethyl chloromethyl ketone.

13453

13454 Hormone-binding Site Structure of the Estrogen Receptor

tamoxifen aziridine-labeled peptides derived by limit trypsin digestion from pure affinity labeled receptors. Our results indicate that the major interaction site for [3H]tamoxifen aziridine with the bovine estrogen receptor lies close to the COOH terminus of the hormone-binding domain.

MATERIALS AND METHODS*

RESULTS

Purification of PHITamoxifen Aziridine-labeled Estrogen Receptors-Two different approaches were used for the preparation of highly purified [3H]tamoxifen aziridine-labeled estrogen receptors prior to enzyme digestion and sequence analysis. Bovine estrogen receptor was purified initially by sequential chromatography on cellulose phosphate and hep- arin-Sepharose followed by affinity chromatography using [3H]tamoxifen aziridine for receptor recovery (19). This method, which includes sodium molybdate in purification buffers, isolates the M, = 65,000 estrogen receptor associated with bovine heat shock protein hsp 90 (19-21). Small amounts of proteolyzed receptor, ranging in size from M, 50,000- 60,000, are also present in the purified extracts, with the M, - 50,000 component usually being most abundant (19). After electrophoretic separation by preparative SDS-PAGE, bands corresponding to the M , - 65,000 and 50,000 receptors were excised and the proteins were recovered by electroelution (20). Assuming a single binding site/receptor monomer, tritium estimation revealed the recovered fractions to contain 388 and 470 pmol of the labeled M, - 65,000 unit and the M, - 50,000 receptor fragment, respectively. Silver staining of analytical SDS-PAGE gels showed single bands for both receptor forms which coincided with the migration of protein- bound [3H]tamoxifen aziridine (Fig. 1, A and B) .

The affinity gel used in the above isolation experiments has recently been incorporated in a rapid, single-step procedure which, in the absence of molybdate buffers, provides untrans- formed, hsp 90-associated estrogen receptor of 5-15% p ~ r i t y . ~ In batchwise experiments, using [3H]tamoxifen aziridine as eluting ligand, this alternative purification method gave 1092 pmol of covalently labeled estrogen receptor, as determined by hydroxyapatite exchange assay using receptor extracts recovered in parallel with estradiol. Coomassie Blue staining of an analytical SDS-PAGE gel showed dominant bands attributed to the M , - 65,000 receptor and hsp 90 proteins (Fig. 2 A , left lane). Fluorography of the same gel revealed the M , - 65,000 receptor as the major labeled component (Fig. 2A,right lane). Consistent with our earlier observations (19), bands of radioactivity corresponding to M, - 60,000, 53,000, and 50,000 receptor fragments were also evident (Fig. 2 A , right lane).

Trypsin Digestion of pH1Tamorifen Aziridine-labeled Es- trogen Receptors-Our strategy for processing the pure labeled M, - 65,000 and 50,000 receptors was to carboxymethylate the dithiothreitol-reduced proteins with iodoacetic acid. After limit digestion with trypsin, highly pure [3H]tamoxifen aziridine-labeled peptide(s) would be isolated by reversed-phase HPLC ready for sequence analysis. A similar experimental design was proposed for the partially purified, affinity labeled receptor. This would eliminate the need for further receptor purification by repeated preparative SDS-PAGE/electroelu-

Portions of this paper (including “Materials and Methods,” Table I, and Figs. 1, 2, and 5) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

3T. Ratajczak, M. J. Brockway, and R. Hahnel, manuscript in preparation.

tion but at the same time be more reliant on the high resolving power of “multidimensional” reversed-phase HPLC (22, 23) to provide the 3H-labeled peptide(s) in pure form.

Alkylation and trypsin digestion of [3H]tamoxifen aziridine-labeled estrogen receptors was monitored by analytical SDS-PAGE and fluorography. For the partially purified receptor preparation a slight increase was evident in the apparent molecular size of all radiolabeled species after car- boxymethylation (Fig. 2B, lane 2). Similar results were observed separately with the pure M, - 65,000 and 50,000 receptor species (not shown). Methanol precipitation of the alkylated product in the presence of trypsin gave a solubilized fraction in which the 3H-labeled receptor was already partially proteolyzed to radiolabeled peptides of M, 20,000-22,000 (Fig. 2B, lane 3) . After extended exposure to trypsin, limit digestion was demonstrated by the retention of radioactivity only by small peptides migrating with the dye front (Fig. 2B, lane 4). Variable losses of [3H]tamoxifen aziridine-labeled material were sustained during these manipulations. Tritium estimation showed that 204 pmol of labeled peptides (yield 53%) were recovered from the pure M, - 65,000 receptor preparation. The corresponding results for digests derived from the pure M, - 50,000 receptor and partially purified receptor extracts were 388 (81%) and 580 pmol (53%), respectively.

Reversed-phase HPLC Purification of pH]Tamoxifen Azir- idine-labeled Tryptic Peptides-Tryptic peptides were sepa- rated initially by reversed-phase HPLC using a short microbore column (2.1-mm internal diameter) and a linear 0-60% (v/v) acetonitrile/water gradient delivered over 40 min at a flow rate of 1 ml/min. Trifluoroacetic acid (0.1% v/v) was included in the mobile phase. The radioactivity profiles obtained from the three digestion products were all closely comparable. Fig. 3A shows the profiles generated from the pure M, - 65,000 and 50,000 receptor digests. These profiles were characterized by a major peak of radioactivity (Peak 1 ) which was eluted with -42% acetonitrile (Fig. 3A). A second, much smaller peak (Peak 2) was observed at slightly higher acetonitrile concentrations (Fig. 3A). Extended column elution with 60% acetonitrile, during fractionation of the M, - 50,000 protein digest, led to the isolation of a third radiolabeled component (Peak 3 ) (Fig. 3A). A similar profile pattern was produced during analytical reversed-phase HPLC of the trypsin digest derived from partially pure 3H-labeled receptor (Fig. 3B). Column elution conditions were altered to a linear 0-100% (v/v) acetonitrile/water gradient applied over 60 min at 1 ml/min. The steeper gradient caused decreased resolution between the labeled peptides of Peaks 1 and 2. A broad peak of radioactivity (Peak 3 ) was attributed to the third labeled component which was recovered with -70% acetonitrile (Fig. 3B). These elution conditions for the Peak 3 component suggested the presence of a 3H-labeled peptide with strong hydrophobic interaction properties.

For the most abundant peptide fraction (Peak 1) tritium estimation indicated yields of 70, 80, and 120 pmol isolated from the pure M, - 65,000 and 50,000 receptor preparations and the partially purified receptor extract, respectively. Much lower quantities of the Peak 2 labeled component were recovered, 17 pmol were derived from the M, - 65,000 receptor monomer and 20 pmol from the M, - 50,000 receptor fragment. The Peak 3-labeled peptide of M, - 50,000 receptor origin amounted to 35 pmol.

Further purification of the isolated 3H-labeled peptides was carried out using microbore reversed-phase HPLC with decreased flow rates and alternative mobile phase conditions. The peptides of Peak 1 (Mr - 65,000 receptor) were fractionated using a low pH mobile phase (0.1% v/v aqueous trifluo-

Hormone-binding Site Structure of the Estrogen Receptor I

10 20 30 40 50 60 70 80 Fraction Nu.ber

Peak 1 I

Peak 3

2 l .- .-

0 - 10 20 30 40 50

0

Fract ion k .ber

FIG. 3. Reversed-phase HPLC of [3H]tamoxifen aziridine- labeled peptides. A , separation of radiolabeled peptides obtained from limit trypsin digestion of the pure M , - 65,000 (-1 and M , - 50,000 ( . . . .) affinity labeled receptors. Chromatographic conditions: column, Brownlee RP-300 (30 X 2.1-mm inner diameter), linear 40-min gradient from 0 to 100% B, where solvent A was 0.1% (v/v) trifluoroacetic acid and solvent B was 60% (v/v) acetonitrile, 40% water containing 0.1% trifluoroacetic acid, flow rate, 1 ml/min. Frac- tions were collected every 30 s, and 5 - ~ 1 aliquots were withdrawn from each for tritium estimation. B, analytical separation of 3H- labeled peptides derived by limit trypsin digestion of affinity chromatography purified estrogen receptors covalently labeled with [3H] tamoxifen aziridine. Chromatographic conditions: column, Brownlee RP-300 (30 X 2.1-mm inner diameter), linear 60-min gradient from 0 to 100% B, where solvent A was 0.1% (v/v) trifluoroacetic acid and solvent B was acetonitrile containing 0.1% (v/v) trifluoroacetic acid, flow rate, 1 ml/min. Fractions were collected manually into counting vials every 30 s for tritium estimation.

roacetic acid, pH 2.1) and gave a complex peptide profile (Fig. 4A). Ninety two % of the applied radioactivity was recovered and eluted in two consecutive fractions (retention time 44.59 min), synchronous with a major peak of peptide absorbance (Fig. 4A). Under the same chromatographic conditions, fractionation of Peak 1 peptides, derived from partially pure estrogen receptor, gave a radioactivity profile which was su- perimposable on that obtained for the corresponding M, - 65,000 receptor fraction (Fig. 4B). The peptides of Peak 1 (MI - 50,000 receptor) were selected for further purification using a linear water-acetonitrile gradient with unbuffered sodium chloride (1%) as the mobile phase. Two major peptide peaks, eluting at 37.59 and 38.19 min, were resolved with this chromatographic dimension and the recovered radioactivity (61% of that applied to the column) was found to be evenly distributed between them (Fig. 5).

0.02

T

0.4

0.2

0 10 20 30 40 50 60 70

B

0 10 20 30 40 50 60 70 - Retention Time (min)

FIG. 4. Microbore reversed-phase HPLC of Peak 1 radioactivity fractions isolated in Fig. 3 from limit trypsin digests of affinity labeled homogeneous M, - 65,000 estrogen receptor ( A ) and affinity chromatography purified receptors ( B ) . Chromatographic conditions: column, Brownlee RP-300 (30 X 2.1- mm inner diameter), linear 70-min gradient from 0 to 100% B, where solvent A was 0.1% (v/v) aqueous trifluoroacetic acid and solvent B was 60% (v/v) acetonitrile, 40% water containing 0.09% (v/v) trifluoroacetic acid, flow rate, 100 Fl/min, column temperature, 45 'C. Fractions were collected manually at 1-min intervals, and 5-pl aliquots were withdrawn from each for radioactivity estimation. The fractions indicated by the horizontal bar were subjected to sequence analysis.

The above separation systems were also employed for the microbore HPLC fractionation of Peak 2 peptides combined from both radiolabeled M, - 65,000 and 50,000 estrogen receptors. The small, observed peak of radioactivity corre- sponded to only a minor peptide in the 215 nm absorbance profile (not shown) and the low levels of material precluded further analysis of the labeled fragment by Edman degradation. Similar results were obtained with Peak 3 peptides isolated from the M, - 50,000 proteolyzed receptor form.

Sequence Analysis of PHITarnoxifen Aziridine-labeled Pep-


tides-HPLC-purified, radiolabeled peptides were loaded onto a polybrene-treated sequenator sample disc. Prior to Edman degradation the sample was exposed to trifluoroacetic acid, extracted with butyl chloride, precoupled with phenylisothi- ocyanate and trimethylamine and extracted with heptane and ethylacetate. Analysis of the major 3H-labeled component isolated from the Peak 1 (MI - 65,000 receptor) fraction (see Fig. 4A) identified a mixture of two peptides (TI, T2) (Table I and Fig. 6). The amino acid sequences of the two fragments showed close correspondence with two tryptic peptides predicted from the cDNA-derived amino acid sequence of the human estrogen receptor (1, 2) (Fig. 6). The residues of peptide T2 are fully conserved in all estrogen receptors (1-6). However, for peptide TI, the single substitution of phenylal- anine for isoleucine differentiates the sequenced residues from the corresponding human receptor sequence which is identical among all other estrogen receptors (1-6) (Fig. 6). Alignment of the TI and T2 sequences with the cDNA-deduced amino acid sequence of the human estrogen receptor indicated that both peptides reside close to the COOH terminus of the hormone binding region of the protein (7).

The radioactivity monitored in PTH amino acid extracts, after each degradation cycle, represented only a small fraction of the peptide-associated radioactivity applied to the sequencer. The bulk of the tritium label was located in the waste solvent container of the sequenator. These results suggested that the covalent link between the peptide and labeled tamoxifen aziridine was labile to the strongly acidic and/or basic conditions encountered during the Edman degradation. A small quantity of the equivalent, rechromatographed Peak 1 (partially purified receptor) fraction (see Fig. 4B) was loaded onto the sequencer and radioactivity was sampled from washes recovered after each predegradation treatment. Little radioactivity was collected after trifluoroacetic acid and butyl chloride extraction. However, approximately 19% of the applied tritium label was released on exposure of the labeled peptide to trimethylamine vapor followed by heptane and ethylacetate extraction. This result demonstrated the sensi- tivity of the aziridine-peptide linkage to base-mediated cleavage.

Sequence analysis of the two major radiolabeled peptides isolated on rechromatography of the Peak 1 (MI - 50,000 receptor) fractions (see Fig. 5) gave only low level signals unrelated to estrogen receptor sequences. Radioactivity was distributed as before with the major quantity appearing in sample disc washes and residual tritium levels being observed only in early cycles after PTH amino acid analysis. Similar results were obtained with Edman degradation of the rechromatographed Peak 1 peptides derived from the partially pure receptor. The appearance, in early cycles, of a high back- ground of PTH amino acids made the data uninterpretable.

Evidence suggests that the highly selective coupling of

P o p t I d . T 1 bER hER 564 L A Q L L L I L S H m

L A Q L L L I L S H F

P a p t l d o 12 bER N V V P L Y D L L L E H L D A H C R ) hER 5 3 2 N V V P L Y D L L L E I I L D A H R

FIG. 6. Sequence analysis of affinity labeled peptides. Bo- vine estrogen receptor (bER) peptides TI and T1 were identified by sequence analysis of the major radiolabeled fraction isolated by microbore HPLC of the Peak l component (M, - 65,000 receptor) (see Fig. 4A). Also shown is a structural comparison between these sequenced residues and tryptic peptides predicted from the hormone- binding domain of the human estrogen receptor (hER) (7). Amino acids are identified by their single letter code. Numbers represent the position of the amino acid residue in the human estrogen receptor sequence.

tamoxifen aziridine to estrogen receptors is facilitated by a site-activated process that may involve initial protonation of the aziridinyl group followed by alkylation of a suitably placed nucleophilic residue (12). Such a mechansim might require an identical microenvironment for the aziridine function within the active site of estrogen receptors. The above observations, taken together, favor the totally conserved peptide Tz over the variant fragment TI as the affinity labeled peptide. In peptide T2 tyrosine, glutamic acid, aspartic acid, methionine, histidine, and possibly arginine all have nucleophilic functions with a potential for alkylation by the reactive aziridine group. Preliminary hydrolytic experiments in which the label was retained by the peptide after a 2 h, 120 “C treatment with 2 M potassium hydroxide solution, mitigate against glutamate and aspartate as the labeled amino acid residues.

Structural Homology of Receptor Steroid-binding Do- mains-The physical properties of amino acid residues con- trol the folding of proteins (32). Argos et al. (33) have used comparisons of smoothed plots of amino acid physical characteristics versus residue sequence number to demonstrate structural similarities between functionally related proteins which have no apparent sequence homology. A comparison of the primary structures of the human estrogen and glucocorticoid receptors showed only a 30% positional identity over the 250-residue long steroid-binding domain (3). Application of the physical parametric approach (33) to the sequence comparison of the two proteins in this region indicated an excellent predicted structural similarity between both receptors (Fig. 7).

In the rat glucocorticoid receptor, residues 622, 656, and 754 were found to be alkylated either by photoinduced reaction with the A-ring of triamcinolone acetonide (Met-622 and Cys-754) (8) or by electrophilic affinity labeling via the D- ring mesylate functionality of dexamethazone 21-mesylate (Cys-656) (8,9). The labeled residues correspond to Met-604, Cys-638, and Cys-736 in the human glucocorticoid receptor (8, 34). As seen in Fig. 7, segments of the glucocorticoid

---hGR 550 600 650 , 70: 750

I e t - 6 0 4 cys-736

1

9 7.0

u)

L

a2 3 6 . 5 I ’ 6.0

5 .5

I ‘TI 12 ”

I - 350 400 450 500 550

T ” 1

- hER

FIG. 7. The sum of the smoothed amino acid physical characteristics plots for the human estrogen receptor (solid curve) and glucocorticoid receptor (dashed curve) versus the amino acid sequence number of steroid-binding domain residues. Receptor sequences were aligned as described by Krust et ai. (3) with gaps introduced for maximum alignment. Numbers represent the position of the amino acid residues in each sequence. The estrogen receptor sequence begins at residue 302 and that for the glucocorticoid receptor at residue 518. The positions of bovine estrogen receptor tryptic peptides TI and T2 are shown to correspond to human estrogen receptor segments 504-515 and 532-548, respectively. hGR, human glucocorticoid receptor.

Hormone-binding Site Structure of the Estrogen Receptor 13457

receptor which included these 3 residues gave parametric curves which were closely correlated with those of the estrogen receptor. The result suggests that these corresponding regions in the estrogen-binding protein may represent the equivalent contact domains for estradiol. Consistent with this proposal, human estrogen receptor peptides, which correspond to the sequenced fragments T1 (equivalent to hER residues 504-515) and Tz (hER residues 532-548), are shown in Fig. 7 to be located proximal to Cys-736 of human glucocorticoid receptor, the expected site for A-ring steroid interaction (8). The predicted secondary structure (35) in this region for both receptors is strikingly similar. Cys-736 in the glucocorticoid receptor marks the start of a P-sheet. For the estrogen receptor a predicted P-sheet structure begins at position 532 which cor- responds to the first amino acid residue (asparagine) of peptide Tz.

DISCUSSION

In the present study we have shown that limit trypsin digestion of homogeneous and partially purified estrogen receptors, labeled covalently with t3H]tamoxifen aziridine, generates a consistent pattern of one major labeled fragment and two additional radiolabeled peptides present in lower concentrations. With the more abundant component, our strategy of using multidimensional microbore HPLC to isolate labeled peptides prior to radiosequence and amino acid sequence analysis was partially successful and provided sequence information on two tryptic fragments (Tl and Tz) derived from the steroid-binding domain. While it is possible that both peptides play a role in the interaction of tamoxifen aziridine with the receptor only one (peptide Tz) is favored to contain the affinity labeled residue. Based on homologous sequences in steroidogenic enzymes, steroid receptors, and a steroid-binding protein a consenus sequence has recently been proposed which identifies the steroid binding region in these proteins (36). It is of note that the linear sequence of peptide T2 overlaps this consensus site (36). Our evidence suggests that the interaction site for tamoxifen aziridine with the estrogen receptor is positioned close to the carboxyl terminus of the protein. This result is compatible with the photoinduced, covalent interaction of the corresponding region in the glucocorticoid receptor with the A-ring of the glucocorticoid derivative triamcinolone acetonide (8).

The covalent link between tamoxifen aziridine and estrogen receptors is stable to a wide range of denaturing conditions including hot solvent extraction, boiling in the presence of mercaptoethanol/sodium dodecyl sulfate, and acid fixation after polyacrylamide gel electrophoresis (12, 19). Our present results indicate the facile cleavage of the affinity label-peptide coupling by trimethylamine during routine precycling treat- ments carried out prior to Edman sequencing. It appears that the cleavage reaction generates an unaltered amino acid, and the previously labeled residue can then be sequenced in the normal mode. Several nucleophiles including aspartic acid, cysteine, histidine, lysine, methionine, and tyrosine have been implicated in the interaction of steroids with enzymes and steroid hormone-binding proteins (8, 37-44). In the estrogen receptor an early study by Ikeda (37) indicated that a sulfhydryl group proximates the D-ring of the steroid, and recent evidence suggests that a histidine residue or an unusually reactive tyrosine may also be important for steroid binding (45). Five nucleophilic residues, aspartic acid, glutamic acid, histidine, methionine, and tyrosine (all with a potential for affinity alkylation by tamoxifen aziridine) are present in peptide Tz, the fragment thought most likely to contain the labeled amino acid. Our initial hydrolytic experiments dis-

count aspartic acid or glutamic acid as the labeled residue. Current efforts by our group, aimed at identifying the labeled amino acid, are being directed on the assumption that the active residue is histidine (equivalent to His-547 in the human estrogen receptor).

Two minor labeled components, isolated by reversed-phase HPLC could not be sequenced. These labeled peptides were consistently observed either with homogeneous or partially pure receptor and may represent alkylated fragments derived from differing binding modes of tamoxifen aziridine with the estrogen receptor. Martin et al. (14) have presented immu- nological evidence for antiestrogen interaction with a region of the receptor different to the estrogen-binding site. In the estradiol 17P-dehydrogenase system, evidence for an inverted substrate orientation at the active site has recently been presented (44).

Together the steroid-binding domains of the human estrogen and glucocorticoid receptors display a close correlation between predicted structural characteristics, even though these regions share only a 30% sequence identity (3). The similarity is especially significant in segments which include the interaction sites recently predicted for the glucocorticoid A- and D-rings in the human glucocorticoid receptor (8-10). The corresponding structural elements in the estrogen receptor then may also be directly involved in estradiol binding. These structural regions correspond closely to segments in both receptors which, on the basis of the sequence similarity, had previously been proposed to contain the hormone active site (46). In the human estrogen receptor, the location of Cys- 417 (1,2) in or very near the predicted interaction site for the estradiol D-ring is consistent with the reported affinity labeling of a sulfhydryl entity by a reactive D-ring substituent on the steroid (37). This residue is closely aligned to Cys-638 in the human glucocorticoid receptor, the predicted alkylation site of dexamethazone 21-mesylate (8). Both cysteinyl residues may be important for hydrogen bond formation during steroid-receptor interaction (37).

The present data allow the location of the major interaction site for tamoxifen aziridine in the estrogen receptor and define regions within the steroid-binding domain which may directly participate in the binding of estradiol. With the availiability of estrogenic affinity labeling reagents, it may now be possible to compare, at the molecular level, receptor interaction with estrogens and antiestrogens and to explain the influence of these hormones on receptor function.

REFERENCES

1. Green, S., Walter, P., Kumar, V., Krust, A., Bornert, J.-M., Argos,

2. Greene, G. L., Gilna, P., Waterfield, M., Baker, A., Hort, Y., and

3. Krust, A., Green, S., Argos, P., Kumar, V., Walter, P., Bornert,

4. Koike, S., Sakai, M., and Muramatsu, M. (1987) Nucleic Acids

5. Weiler, I. J., Lew, D., and Shapiro, D. J. (1987) Mol. Endocrinol.

6. White, R., Lees, J. A., Needham, M., Ham, J., and Parker, M.

7. Kumar, V., Green, S., Staub, A., and Chambon, P. (1986) EMBO

8. Carlstedt-Duke, J., Stromstedt, P.-E., Persson, B., Cederlund, E., Gustafsson, J.-A., and Jornvall, H. (1988) J. Biol. Chem. 2 6 3 ,

9. Simons, S. S., Jr., Pumphrey, J. G., Rudikoff, S., and Eisen, H. J. (1987) J. Biol. Chem. 262,9676-9680

10. Smith, L. I., Bodwell, J. E., Mendel, D. B., Ciardelli, T., North, W. G., and Munck, A. (1988) Biochemistry 27,3747-3753

11. Simons, S. S., and Thompson, E. B. (1981) Proc. Nutl. Acad. Sci. U. S. A. 78,3541-3545

P., and Chambon, P. (1986) Nature 320 , 134-139

Shine, J. (1986) Science 2 3 1 , 1150-1154

J.-M., and Chambon, P. (1986) EMBO J. 5,891-897

Res. 15,2499-2513

1,355-362

(1987) Mol. Endocrinol. 1, 735-744

J. 5,2231-2236

6842-6846


12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

Katzenellenbogen, J. A., Carlson, K. E., Heiman, D. F., Robert- son, D. W., Wei, L. L., and Katzenellenbogen, B. s. (1983) J. Biol. Chem. 258,3487-3495

Hahnel, R., Twaddle, E., and Ratajczak, T. (1973) J. Steroid Biochem. 4, 687-695

Martin, P. M., Berthois, Y., and Jensen, E. V. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 2533-2537

Monsma, F. J., Jr., Katzenellenbogen, B. S., Miller, M. A., Ziegler, Y. S., and Katzenellenbogen, J. A. (1984) Endocrinology 115,

Miller, M. A., Mullick, A., Greene, G. L., and Katzenellenbogen, B. S. (1985) Endocrinology 117,515-522

Katzenellenbogen, B. S., Elliston, J. F., Monsma, F. J., Jr., Springer, P. A., Ziegler, Y. S., and Greene, G. L. (1987) Bio- chemistry 26,2364-2373

Koike, S., Nii, A., Sakai, M., and Muramatsu, M. (1987) Biochem- istry 26,2563-2568

Ratajczak, T., Comber, M., and Hahnel, R. (1988) J. Steroid Biochem. 31, 909-916

Ratajczak, T., Comber, M., Moir, R., Hahnel, R., Grego, B., Rubira, M. R., and Simpson, R. J. (1987) Biochem. Biophys. Res. Commun. 143,218-224

Atrache, V., Ratajczak, T., Senafi, S., and Hahnel, R. (1985) J. Biol. Chem. 260,5936-5941

Simpson, R. J., Moritz, R. L., Lloyd, C. J., Fabri, L. J., Nice, E. C., Rubira, M. R., and Burgess, A. W. (1987) FEBS Lett. 224,

Simpson, R. J., Moritz, R. L., Rubira, M. R., and Van Snick, J.

Laemmli, V. K. (1970) Nature 227,680-685 Titani, K., Koide, A., Hermann, J., Ericsson, L. H., Kumar, S.,

Wade, R. D., Walsh, K. A., Neurath, H., and Fisher, E. H. (1977) Proc. Natl. Acad. Sci. U. S. A. 85, 2533-2537

Squire, P. G., Moser, P., and O'Konski, C. T. (1968) Biochemistry 7,4261-4272

143-153

128-132

(1988) Eur. J . Biochem. 176, 187-197

27.

28.

29. 30.

31.

32. 33.

34.

35. 36.

37. 38.

39.

40.

41.

42.

43.

44.

45.

46.

Nisbet, A. D., Saudry, R. H., Moir, A. J. G., Fothergill, L. A., and

Reynaud, J., Luccioni, F., Bouthier, M., Savary, J., and Derrien,

Koide, T., and Ikenaka, T. (1973) Eur. J. Biochem. 32, 401-407 Brew, K., Vanaman, T. C., and Hill, R. L. (1967) J. Bid. Chem.

Merril, C. R., Goldman, D., Sedman, S. A., and Ebert, M. H.

Creighton, T. E. (1978) Biophys. Mol. Biol. 33,231-297 Argos, P., Hanei, M., Wilson, J. M., and Kelley, W. N. (1983) J.

Weinberger, C., Hollenberg, S. M., Rosenfeld, M. G., and Evans,

Chou, P. Y., and Fasman, G . D. (1974) Biochemistry 13,222-245 Picado-Leonard, J., and Miller, W. L. (1988) Mol. Endocrinol. 2,

Ikeda, M. (1982) Biochem. Biophys. Acta 718, 66-73 Kuhn, R. W., and Raymoure, W. J. (1982) J. Steroid Biochern.

Thomas, J. L., and Strickler, R. C. (1983) J. Biol. Chem. 258,

Holmes, S. D., and Smith, R. G. (1983) Biochemistry 22, 1729-

Thomas, J. L., La Rochelle, M. C., Asibey-Berko, E., and Strick-

Hearne, M., and Benisek, W. F. (1985) Biochemistry 24, 7511-

Onoda, M., Haniu, M., Yanagibashi, K., Sweet, F., Shively, J. E.,

Murdock, G. L., Warren, J. C., and Sweet, F. (1988) Biochemistry

Baker, M. E., Sklar, D. H., Terry, L. S., and Hedges, M. R. (1985)

Baker, M. E. (1986) Biochem. Biophys. Res. Commun. 139,281-

Fothergill, J. E. (1981) Eur. J. Biochem. 115, 335-345

Y. (1971) Biochemie 53, 1095-1098

242,3747-3749

(1981) Science 211, 1437-1438

Biol. Chem. 258, 6450-6457

R. M. (1985) Nature 318,670-672

1145-1150

17,459-460

1587-1590

1734

ler, R. C. (1985) Biochemistry 2 4 , 5361-5367

7516

and Hall, P. F. (1987) Biochemistry 26,657-662

27,4452-4458

Biochem. Int. 11,233-238

286

for radicdctivity "t.

dlqeet im.

Hormone-binding Site Structure of the Estrogen Receptor 13459

A B 1 2 3 4

1.K-

8P8"

B K -

0.10 r

0.05

I

6 7 8 9

10 11 11 13 14 13 16 17

the interaction site for tamoxifen aziridine with the ... · chromatography purification. after...

Documents