- Email: [email protected]
The intranuclear distribution of rat uterine estrogen receptors determined after nuclease treatment and chromatin fractionation
Molecular and CellularEndocrinology, 26 (1982) 201-216 Elsevier/North-Holland Scientific Publishers, Ltd.
201
THE INTRANUCLEAR DISTRIBUTION OF RAT UTERINE ESTROGEN RECEPTORS DETERMINED AFTER NUCLEASE TREATMENT AND CHROMATIN FRACTIONATION Edward J. PAVLIK * and Benita S. KATZENELLENBOGEN
**
Department of Physiology and Biophysics, Universityof Illinois, and School of Basic Medical Sciences, Universityof Illinois, College of Medicine, Urbana,IL 61801 (U.S.A.) Received 6 July 1981; revision received 3 December 1981; accepted 29 December 1981
The intranuclear locations at which rat uterine estrogen receptors interact with chromatin have been probed using digestions performed with DNAase I and micrococcal nuclease. Exposure to nuclease has been controlled to effect limited to extensive digestion of nuclear DNA under conditions which maintain the integrity of the [sH] estradiol-receptor complex. The effect of divalent cation concentration on the release of estrogen receptors from nuclease-treated chromatin was examined and found to be of consequence above 2 mM. Exposure to nuclease released nuclear estrogen receptors from chromatm, with DNAase I being more efficient than micrococcal nuclease in mediating this release. The release of the bulk of nuclear estrogen receptors closely paralleled the nuclease-mediated digestion of chromatin DNA. At 1 h after exposure to estrogen, substantial quantities of uterine estrogen receptors (80-90%) were distributed in chromatin fractions which, on the basis of fractionation terminology, have been termed ‘transcriptionally inactive’by convention. Enrichment of estrogen receptors in chromatin which has been termed ‘transcriptionally active’ only occurred with lo-20% of the estrogen receptors. Hence, our findings support a model where, at early times after estrogen exposure, receptors from the rat uterus are enriched to only a minor extent in chromatin to which ‘transcriptional activity’ is generally assigned while the bulk of receptors are localized in chromatin which is generally considered ‘transcriptionally inactive’. Keywords: nuclear estrogen receptors; tionation; uterus.
DNAase I; micrococcal nuclease; chromatin frac-
Steroid hormones exert their effects upon target tissues after passing through the cell membrane and interacting with a cytoplasmic binding protein which as an activated binding complex translocates to the nucleus. The ensuing intranuclear
* Present address: Departments of Obstetrics and Gynecology and Biochemistry, University of Kentucky, Lexington, KY 40503 (U.S.A.). ** Address correspondence to: Dr. Benita S. Katzenellenbogen, Department of Physiology and Biophysics, 524 Burrill Hall, 407 South Goodwin Avenue, University of Illinois, Urbana, IL 61801 (U.S.A.). 0303-7207/g2/0000-0000/$02.75
0 Elsevier/North-Holland
Scientific Publishers, Ltd.
Edward J. Pavlik,Benita S. Katzenellenbogen
202
interactions presumably initiate certain molecular events which ultimately are responsible for alterations in the rates of transcription, translation and replication (for reviews see Jensen and DeSombre, 1972, 1973; O’Malley and Means, 1974; Katzenellenbogen and Gorski, 1975; Yamamoto and Alberts, 1976). At present very little is known about the intranuclear interactions in which estrogen receptors are involved; even less is known about the intranuclear location of these interactions. Attempts to resolve the intranuclear location of receptor-chromatin associations have utilized fractionations which define transcriptionally active and inactive regions using either mechanical (Levy and Baxter, 1976; deBoer et al., 1978; Franeschi and Kim, 1979) or enzymatic (Hemminki, 1977; Hemminki and Vauhkonen, 1977; Senior and Frankel, 1978; Rennie, 1979; Scott and Frankel, 1980) disruptions of chromatin; however, the results have not been unambiguous. Estrogen receptors from chick oviduct (Franeschi and Kim, 1979) and rat uterus (deBoer et al., 1978) have been reported to associate predominately with fast sedimenting, transcriptionally inactive fractions. Similar observations have been reported for. androgen receptors in prostatic chromatin (Rennie, 1979) and glucocorticoid receptors in pituitary chromatin (Levy and Baxter, 1976). Conversely, other investigators have reported that transcriptionally active regions of chromatin are enriched in estrogen receptors (Hemminki, 1977; Hemminki and Vauhkonen, 1977; Scott and Frankel, 1980). Since these opposing reports can be the result of differences in methodology or in the application of a methodology, this report examines the intranuclear distribution of rat uterine estrogen receptors using nuclease-mediated chromatin fractionation techniques. In these studies, nuclease-mediated hydrolysis of DNA has been finely controlled so that the release of estrogen receptors from rat uterine chromatin was studied under a spectrum of nucleolytic DNA hydrolyses, ranging from mild to extensive digestion of nuclear DNA. In particular, this approach has resulted in rapid chromatin fractionations which minimize the opportunities for estrogen receptors to reassociate during preparation.
MATERIALS
AND METHODS
Preparation of nuclei
Immature female Sprague-Dawley rats were obtained from Holtzman, Madison, WI, and used between ages 20-23 days. Uteri were debrided of fat and mesentery and placed in HeLa medium (DIFCO), pH 7.4-7.6 adjusted with CO2 gassing at 0-4°C. Estrogen receptors were translocated to the nucleus by exposing the uteri in vitro to receptor saturating concentrations (50 nM) of [3H] estradiol and to [3H] estradiol in the presence of 200-fold excess diethylstilbestrol (DES) to assess nonspecific binding (Williams and Gorski, 1973; Pavlik et al., 1979). After 60 min at 37°C the translocation reaction was terminated by transferring the uteri to iced buffer (5 ml) containing 50 mM Tris, 25 mM KCl, 0.5 M sucrose, pH 7.4 (0.5 M S-TK). After cooling on ice for 5 min, the uteri were homogenized at 0-4”C with
Intranuclear distribution of estrogen receptors
203
seven 3-set bursts of a Polytron sonic/shear homogenizer (Brinkman Industries) set at 40% of maximum speed. All subsequent manipulations were performed at 4°C. The homogenate was centrifuged in an HB-4 swinging bucket rotor (Sorvall) at 6500 Xg for 5 min in a RC-2B centrifuge (Sorvall). The supernatant (SA) was aspirated and the pellet was resuspended in TK buffer containing 1.5 M sucrose (1.5 M S-TK) using a loose-fitting, round bottom !eflon homogenizer. Final resuspension was performed by 10 aspirations through the bore of a standard Pasteur pipette. The resuspended material was centrifuged at 0-4”C for 20 min at 13 000 X g. The pelleted material was suspended in TK buffer containing 0.25 M sucrose (0.25 M S-TK) and inspected under a phase-contrast microscope. The resulting supernatant (Sn) was saved. Nuclei purified in this manner and judged to be free of major non-nuclear contaminants by microscopic examination were used in studies employing nucleases. If the purity of the preparation was questionable, the suspension was underlaid in 1.5 M S-TK and centrifuged at 1300 X g for 20 min. Nuclear yields by this protocol were 40-60%. Agarose gel electrophoresis Nuclei were suspended in buffer containing EDTA (15 mM), EGTA (15 mM), sodium dodecylsulfate (2%) and NaCl (2.3 M). DNA was extracted 3 times with chloroform : isoamylalcohol (24 : 1) and precipitated overnight with 2.5 vol. of ethanol at -20°C. DNA was pelleted by centrifugation at 16 000 Xg for 10 min and the pellet was lyophilized and stored at -20°C. Before use, the DNA was dissolved in 5 mM EDTA and the DNA concentration determined by absorption at 260 nm. Agarose (1.5%) in TANE buffer (40 mM Tris, 20 mM sodium acetate, 18 mM NaCl, 2 mM NasEDTA, pH 8 .O) was dissolved by brief boiling and was poured into glass beds. Wells were formed with Teflon combs, and after gelling cloth wicks were applied to both ends thereby connecting the gel to the buffer reservoirs, both of which contained TANE buffer (Hayward and Smith, 1972). Gels were run at room temperature at 10 mA/gel. DNA was visualized by UV-induced fluorescence after staining with ethidium bromide (10 mg/ml). Sized DNA fragments were generated by Hind II and Hind III restriction endonuclease cuts of plasmid p NO1001 which carries 7 ribosomal protein genes (Fallon et al., 1979). Determination of estrogen binding Soluble estrogen receptors were quantitatively determined by adsorption on hydroxylapatite (Pavlik and Coulson, 1976). In general, receptors were exposed to saturating concentrations of ( [3H]estradiol, 50 r&l) in the presence and absence of 200-fold excess nonradioactive DES, and were bound to hydroxylapatite which was washed 3 times with Tris buffer (10 mM), pH 7.4 to remove free steroid. After extraction with ethanol, saturable and competitive binding was taken to represent the binding by the limited capacity estrogen receptor and was determined as the difference between total binding measured in the presence of labeled ligand alone and nonsaturable binding measured in the presence of labeled ligand mixed with
204
Edward J. Pavlik, Benita S. Katzenellenbogen
excess competitor. Chromatinbound estrogen receptors were determined as described (Katzenellenbogen et al., 1980a,b) after adding hydroxylapatite to the preparation to bind any receptor, which might dissociate from chromatin, and by centrifugally washing chromatin pellets 3 times with Tris buffer followed by ethanol extraction of bound steroid. Specific binding was determined as outlined above. Estrogen receptor was normalized relative to chromatin, and DNA was determined as described by Burton (1956). Nuclease digestions DNAase I (Sigma Chemical and Worthington Biochemical Corp.) and micrococcal nuclease (Worthington Biochemical Corp.) were aliquoted to 80-100 pg portions and stored at -20°C. On the day of use, a nuclease diluent was prepared which contained Mg2+ and Ca2+ (4 : 1) to a final concentration as specified in individual experiments. The nuclease digestion was performed in the presence of 0.15 M sucrose, 30 mM Tris, and 15 mM KCl, pH 7.4. Unless otherwise stated, digestions were performed at 0-4°C and terminated after 4 min by bringing the digest to 5 mM EDTA/S mM EGTA, pH 7.4. Chromatin was then pelleted at 6500 Xg for 5 min and the preparation processed as described. Nuclease activity was verified spectrophotometrically with one Kunitz unit (KU) of activity defined as an increase in base-pair absorbance due to enzymatic hydrolysis of 1000A2e0 units/mm/l at pH 7.4,25”C. Chromatin fractionation terminology DNA associated with nuclei or chromatin which had been washed free of sucrose and pelleted centrifugally is referred to as ‘chromatin DNA’. All material having a UV absorbance at 260 nm which is not pelleted but found soluble in the supernatant after 0.3 N perchlorate treatment is referred to as ‘acid-soluble DNA’ and very likely consists of mostly small size classes of DNA. Material which is released from chromatin after lysis (5 min) with hypotonic Tris (10 mM) is referred to as ‘A 260 material released by lysis.’ Chromatin fractionation Fractionation described by Scott and Frankel (1980) was employed, After nuclei were prepared and exposed to nuclease, the reaction was stopped by bringing the suspension to 5 mM EDTA/EGTA and the chromatin pelleted at 6500 Xg for 5 min (2°C). The fractionation procedure is outlined in Fig. 1. The chromatin was resuspended in TloNloEl [0.5 ml: Tris (10 mM), NaCl (10 mM), EDTA (1 mM), pH 7.41. After 20 min at 2”C, lysis was completed by adding 1.5 ml of TloNlo buffer and centrifuging at 6500X g for 5 min. The pellet (Pl) was processed as described earlier. The supernatant (Sl) was brought to 3 mM MgCL for 30 min at 2°C and centrifuged at 12 000 Xg for 10 min in order to pellet Mg’+-insoluble material (P2). The resulting supernatant (S2) contained Mg’+-soluble material. Estrogen receptor activity was determined in Sl , S2, Pl, l?2 fractions. A260, as well
Intranuclear distribution
of estrogenreceptors
205
as perchlorate (0.3 N) soluble Ash0 activity, was determined for all supernatants. _4s6e was determined in fraction P2 by redissolving the precipitate in TloNloEl buffer, while DNA was assayed in fraction Pl as described by Burton (1956).
TRANSLOCATE
ESTROGEN
RECEPTOR
HOMOGENIZE
I AND PURIFY
NUCLEI
STOP NUCLEASE
1 WITH EDTAiEGTA
t5mMJ
I CENTRIFUGE 16500 xg. 5 min.)
N”CLEAk LYSIS Suspend in 0.5 ml TIDNIOEI. After 20 min. 0% add I. 5 ml TlONlD.
1
CENTRIFUGE t65W xg, 5 min.
Bring to 3 mM Mgil2
(34 min.,
4 CENTRIFUGE IKVDD xg. IO min.
I
Mg+t
52 solubles
I
DV
I
I 2 Mg++ insoluble
Fig. 1. Flow diagram of chromatin fractionation.
RESULTS Hydrolysis of rat uterine DNA by DNAase I DNAase I is able to hydrolyze the DNA in preparations of rat uterine nuclei and hence reduces the DNA content of these preparations as shown in Fig. 2. Hydrolysis proceeds more completely at 30°C than 0°C; however, sufficient degradation does occur at O”C, a temperature which ensures maximal receptor stability. By anlyzing the size of DNA fragments generated by DNAase I hydrolysis using agarose gel electrophoresis in Fig. 3, a progressive increase in small size classes of DNA can be observed as the concentrations of nuclease increase_ Limited digestion of uterine nuclear DNA can be seen when l-10 KU of DNAase per ml was employed, while higher concentrations of nuclease more completely degrade native nuclear DNA to smaller forms. Finally, endogenous nucleases do not appear to be activated in uteri incubated for 60 min at both 0°C and 30°C as shown in the right-most lanes in Fig. 3. Thus, exogenously provided DNAase I is responsible for DNA fragmentation and not some endogenous nuclease, which becomes activated during the course of preparation or experimental manipulation.
3 OL, 0 R
100
200 DNase I,
300
400
500
K.U./ml
Fig. 2. DNAase I degradation of rat uterine nuclear DNA. Nuclei were prepared as described. One uterine equivalent of DNA was defined as 250 ug (Williams and Go&i, 1973) and yields were calculated and normalized on this basis. The yield after preparation based upon DNA content was 92%. Each aliquot contained 362 ug rat uterine DNA (-1.45 uterine equivalent) in a reaction volume of 1 ml: CaCla (05 mM) and MgClz (2 mM) in buffer containing (0.15 M sucrose, 30 mM Tris, 15 mM KCl, pH 7.4). Aliquots were treated at 0°C and 30°C for 4 min with DNAase I (2135 KU/mg), and the reaction was stopped by bringing the suspension to 5 mM EDTA, EGTA, using buffered (pH 7.4) 100 mM EDTA, EGTA. Chromatin was pelleted and brought to 0.3 N inperchlorate, held at O’C for 30 min, washed (3X) with 0.3 N perchlorate and hydrolyzed in 0.3 N perchlorate (30 min, 90°C). DNA was determined according to Burton (1956).
eiP
Fig. 3. Analysis of the degradation of rat uterine nuclear DNA by DNAase I using agarose-gel electrophoresis. Nuclei were prepared as described at 50% yield and aliquoted to a reaction mixture concentration of 100 fig DNA/ml. Nuclease treatment was conducted at O’C for 4 min and terminated as described in Fig. 1. All other preparative steps are outlined in the Methods section. Approx. 20 &g DNA were applied to each well. The stability of nuclear DNA obtained from uteri incubated for 60 min at 0°C (well No. 14) and 37°C (well No. 15) is also shown. The left-most well is well No. 1.
Intranuclear
distributzon of estrogen receptors
201
Release of nuclear estrogen receptors by nucleases and the role of DNA degradation Under conditions where low concentrations of DNAase I are causing limited digestion of DNA, nuclear estrogen receptor disappears from chromatin as shown in Fig. 4. This disappearance, however, is predominately coordinated with the disappearance of DNA from the nuclei since reexpression of these data in terms of DNA recovered after nuclease treatment reveals a relatively constant relationship between receptor and nuclear DNA. Because the disappearance of chromatin associated receptor was modest (_5-10%) when compared to the disappearance of DNA and because the range of DNAase I concentration was kept low, factors which might limit the release of nuclear estrogen receptor from nuclease-hydrolyzed chromatin, as weJ1 as the effects which might be mediated by submaximal DNA degradations, were examined. Because Mg2+ and Ca2* are required for nuclease activity, it is necessary to include these divalent cations in nuclear reaction buffers; however, divalent cations can cause estrogen receptor aggregation (Harris, 1971; Schoenberg and Clark, 1980).
I
I
,
0
5
I
IO DNose I,
I
15
I
20
I
25
K.U./ml
Fig. 4. Release of nuclear estrogen receptors by low concentrations of DNAase I. Estrogen receptor was translocated to the nucleus and nuclei were prepared as described. Estrogen receptor distrrbution: 0.71 pmole/uterine equivalent DNA was chromatin-associated (Pl), while 0.17 pmoles/uterine equivalent DNA remained soluble (SA). Based upon DNA content the yield of partially purified nuclei was -80%. Nuclei were treated with DNAase I (2135 KU/mg) after resuspension to 0.15 M sucrose, 30 mM Tris, 15 mM KCl, pH 7.4 containing CaC12 (0.5 mM final concentration) and MgCla (2 mM final concentration). The reaction volume (1 ml) contained -500 Ng rat uterine DNA. The nuclease reaction was conducted at 0°C for 4 min and terminated by bringing the aliquot to 5 mM EDTA, 5 mM EGTA by adding 50 ~1 of 100 mM EDTA/EGTA buffered at pH 7.15. After processing the preparations as described in Methods and Materials, the DNA content of both nuclease-treated and untreated aliquots was determined and the results expressed as nuclear estrogen receptor per aliquoted amount of DNA determined in untreated samples (left ordinate) or as nuclear estrogen receptor per amount of DNA remaining in the preparation after nuclease treatment (i.e., ‘recovered’ DNA and right ordinate). Untreated preparations were set at 100%.
Edward J. Pavlik, Benita S. Katzenellenbogen
208
Hence, it is important to examine the degradation of nuclear DNA and the release of nuclear estrogen receptor as a function of divalent cation concentration in order to determine if receptor, when induced to aggregate by Mg2+ and Ca2+, will consequently be unable to diffuse out of nuclease-treated chromatin. When Mg2+ (and Ca2+) are varied, maximal release of estrogen receptor and maximal degradation of nuclear DNA are achieved at Mg2+ concentrations of OS-2 mM (Fig. 5). In addition, solubilized A2e0 material is maximized in this concentration range. At very (10 mM), estrogen-receptor aggregation prevents the high Mg2+ concentrations release of all receptors from chromatin and chromatin appears stabilized against the lytic release of A260 material, while other parameters are minimally influenced. of tissue homogenates may Because the endogenous Mg2+/Ca2+ concentration
IOO-
so8070M e--o O--O w -
60B G
50-
Chromatin EzR Chromotm DNA Total APeomaterial Acid sol.A2eomat. A,eomateriol released by lys~s
_
h 40302D-
IO t k
0
I
2
I
I
8
I
4
6
8
IO
[WP]
. mM
I
Fig. 5. The effect of magnesium ion on DNAase I activity and release of estrogen receptors from chromatin. Nuclei were prepared as described at 65% yield. Estrogen-receptor distribution: 0.48 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.08 pmoles/ uterine equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -260 E.rg rat uterine DNA and 500 KU/ml DNAase I (2135 KU/mg). The reaction was conducted and terminated as described in Fig. 4 with the exception of Mga+ concentration which was varied as described. A constant relationship of Mga+:Ca2+ was maintained at 4 : 1. ‘Chromatin Ea R’ was determined as the chromatin associated estrogen receptor/aliquoted DNA and ‘chromatin DNA’ was the DNA remaining after nuclease treatment determined according to Burton (1956).
209
Intranuclear distribution of estrogen receptors
moderate the release of receptor from chromatin, all subsequent nucleate degradations were performed at 2 divalent cation concentrations: at a Mg2+ concentration well below the optima demonstrated in Fig. 5 in order to compensate for the possible inclusion of endogenous divalent cations as well as at a Mg2+ concentration (2 mM) which is optimal for nuclease activity. Both the disappearance of DNA from nuclei and the increase in DNAase I solubilized A260 material were similar when divalent cation concentrations were low (Fig. 6A) or high (Fig. 6B); however, with higher divalent cation concentrations more A 2 6Omaterial was released by lysis of nuclei and more of the DNAase I solubilized material was hydrolyzed to small, acid-soluble fragments. Under conditions of low divalent cation concentration (Fig. 6A), the disappearance of estrogen receptors from chromatin more closely paralleled
Mg+c.
100
- 0.1 0.025
Co++
mM
O-O .--a o---o H
Chromotin EpR Chromotin DNA Total A26omateriol Acid sol.A~+~mat. Aesomaterml released by lysis
_ -
B 0
I
100
I
200
I
I
300
400
I
500
I
DNose I, K.U./ml
Fig. 6. DNAase I digestion of rat uterine nuclei under low and high concentrations of Mga+ and Caa+. (A) Low divalent cation concentration. Nuclei were prepared as described at 40% yield. Estrogen*eceptor distribution: 1.05 pmole/uterine equivalent DNA was chromatm associated (pl) while 0.45 pmoles/uterme equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -130 fig rat uterine DNA. Conditions were as described in Figs. 3 and 4. (B) High divalent cation concentration. Nuclei were prepared as described at 40% yield. Estrogen-receptor distribution 0.61 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.25 pmoles/uterine equivalent DNA remained soluble (SA). All other conditions were as described above.
210
Edward .I. Pavlik, Benita S. Katzenellenbogen
the loss of DNA from chromatin than under conditions of high divalent cation concentration (Fig. 6B) where the disappearance of estrogen receptors from chromatin modestly exceeded the loss of DNA from nuclei by -5-10%. Similar experiments were performed with micrococcal nuclease (Fig. 7). In these experiments, micrococcal nuclease liberated DNA from uterine nuclei but released much less receptor from chromatin than DNAase I under both low and high concentrations of divalent cations. Moreover, micrococcal nuclease prepared with high divalent cation concentrations (Fig. 7B) had no greater activity than nuclease prepared with low divalent ion concentrations (Fig. 7A) in terms of AZ6,, material released by nuclear lysis or in terms of A 260 material hydrolyzed to small, acidsoluble fragments.
00
0 _
- 0t loo-
0
o--_/ ‘k
80-
YTA -
‘*.. _-_
60-
*-----___+___~,+-4
--__ 60-
II
0
I
-
~-_-_______-__-“---
I
200 Micrococcal
I
I
I
+irkd
400 Nuclease
(U./ml)
Fig. 7. Micrococcal nuclease digestion of rat uterine nuclei with low and high concentrations of Mga+ and Ca2+. (A) Low divalent cation concentration. Nuclei were prepared as descrrbed at 55% yield. Estrogen-receptor distribution: 0.62 pmoles/uterme equivalent DNA was chromatin associated (Pl) while 0.20 pmoles/uterine equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -170 Mg rat uterine DNA. Conditions were as descrrbed in Fig. 5. (B) High divalent cation concentration. Nuclei were prepared as described at 60% yield. Estrogen-receptor distribution: 0.38 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.20 pmoles/uterine equivalent DNA remained soluble (SA). Micrococcal nuclease (Worthington: 23882 lJ/mg) was used and conditions were as described.
0.1
2.1
(-)
85.4 (0.9) 100.0
0
-(-)
0
0
0.6
6.5
(-)
(-)
(-)
78.4 ~ (1 .O) 78.8
-
0
0
9.3
14.4 -(3.3) 4.3
5
0
0.7 (-)
(-)
(2.9)
74.0 (1 .O) 70.6
-
5.4 o(-)
J+
46
13.3
10
(0.8)
(-)
70.0 (1.6) 43.3
$
z+
9.1 +1.6)
13.7 (1.9) 7.3
50
44.8 (1.0) 44.1
4.4 __ (0.5) 9.3
+.(-)
14.9 E(l.2)
17.1 (1.6) 10.5
100
(1.6)
(0.4)
39.5 __ (1.3) 29.9
2$
14.4 - 0.8 (18)
26.2 -16.2 (1.6)
139
22.5
250
15.6 L(O.6) 25.2
8.2 -(0.3) 23.9
27.2 -(17) 1.6
27.1 1591.7)
20.9 p(1.2) 17.9
500
a DNAase treatments were as described in the text in the presence of MgCls (2 mM) and CaCls (0.5 mM) for 4 mm at 0°C. Receptor values are expressed as a percent of total recovered estrogen receptor in the absence of nuclease and are the result of averaged duplicates. (Estrogen receptor distribution: Nuclear 0.60 pmoles/uterine equivalent DNA; cytoplasmic 0.24 pmoles/uterine equivalent DNA. One uterine equivalent of DNA is 250 pg.) Values for DNA are expressed as a percent of total aliquoted DNA (270 &g/tube) and are the result of averaged duplicates. b Numbers presented in the Table are the estrogen receptor values (numerator), DNA values (denominator), and the ratio of receptor to DNA presented in parentheses.
Pl
P2
s2
(-)
1.4
Sl
0
10.9 -(2.9)b 3.8
0
DNAase I (KU/ml)
SO
Fraction
Table 1 Chromatin fractionation and estrogen receptor distribution and DNA distribution after treatment with DNAase I a
z r
< 3
5 G !j 2
s 9
5.
2 z 2 B A 3 3. P
0 3.0 (-)
0
s2
P2
-
0
114.9
86.0
0
+
7.7 T(3.7)
(0.7)
(0)
(-)
14.4 5.0 (2.9)
5
(2.1)
0 (0)
69.2 9820.7)
0
-02.6 (-)
g
10.6 T(3.4)
10
88
68.0
0.1 x(05)
(0.8)
0 3.2 (-)
5.0 - 1.9 (2.6)
12.4 4.4 (2.8)
50
(0.1)
(4.2)
57.3 89.0 (0.6)
z
T(44) 4.4
g
11.2 5.6 (2.0)
100
89
62.8 (0.7)
0.9 6.4 (0.1)
o.l3.6 (36)
6.9 26 (2.7)
12.9 90 (1.4)
250
nuclease
(14.8)
72.7 p(1.3) 55.7
3.1 -10.0 (0.3)
;
16.4 49 (3.3)
16.7 1521.1)
500
a
a Nuclease treatments were as described in Table 1. Receptor values are expressed as a percent of total recovered estrogen receptor in the absence of nuclease and are the result of averaged duplicates. (Estrogen receptor distrbution: Nuclear 0.73 pmoles/uterine equivalent DNA; cytoplasmic 0.23 pmoles/uterine equivalent DNA. One uterine equivalent of DNA is 250 Mug).Values for DNA are expressed as a percent of total aliquoted DNA (300 ug/tube) and are the result of averaged duplicates. b Numbers presented in the Table are the estrogen-receptor value (numerator), DNA value (denominator), and the ratio of receptor to DNA presented in parentheses.
80.3 (0.8) 100.0
(0)
4.5 __2.0 (2.3)
Sl
0
12.3 46 (2.7) b
0
Micrococcal nuclease (U/ml)
so
Fraction
Table 2 Chromatin fractionation and estrogen receptor distribution and DNA distribution after treatment with micrococcal
;i’ 9 3 h 9 : 2 9 2
$ e Q .S
Irmanucleardistributionof estrogenreceptors
213
Chromatin fractionation and localization of estrogen receptors after exposure to DNAase I and micrococcal nuclease The general scheme and nomenclature for chromatin fractionation is outlined in Fig. 1. Receptor localization and DNA distribution data after treatment with a wide range of concentrations of nucleases are presented in Tables 1 and 2. DNAase I appeared to release more DNA from nuclei and to generate more AZbOmaterial by nuclear lysis than micrococcal nuclease. Both DNAase I (Table 1) and micrococcal nuclease (Table 2) treatment resulted in very small amounts of AZhOmaterial in the Mg’+-soluble (S2) fraction. In comparing the effects of DNAase I and micrococcal nuclease on estrogen receptors, DNAase I appeared also to more efficiently release estrogen receptors from chromatin than micrococcal nuclease. Second, the Mg2+soluble fraction appears enriched in estrogen receptors as shown by this fractionation study of the distribution of estrogen binding (Tables 1 and 2), and the receptor specific activity (i.e., receptor/pg DNA) is highest in this fraction (ca. 9 pmoles/ 250 I-(g DNA following 500 KU/ml DNAase or 500 U/ml micrococcal m&ease exposure). This latter result, which represents an approx. 15-fold enrichment in receptors, occurs chiefly because the soluble A 260 material in fraction S2 is very low and not because a large proportion of estrogen receptors are distributed in this fraction. As indicated in the Tables, only ca. lo-25% of total estrogen receptors are found in the S2 fraction.
DISCUSSION The data presented in this report have been collected by making use of a very well characterized estrogen-receptor system and by utilizing nuclease-mediated chromatin fractionation. This fractionation methodology has been reported to differentiate transcriptionally active DNA from transcriptionally inactive DNA (Weintraub and Groudine, 1976; Senior and Frankel, 1978; Bloom and Anderson, 1979; Scott and Frankel, 1980). The cogent observations reported here are as follows. (1) Estrogen receptors are released from chromatin by treatment with nucleases, and DNAase I appears more efficient than micrococcal nuclease in mediating this release. (2) Nucleaskmediated release of estrogen receptors closely parallels the nucleolytic disappearance of chromatin DNA. (3) There is some enrichment of estrogen receptors in chromatin fraction S2 which is herein designated as ‘presumed to be transcriptionally active’, but the bulk of estrogen receptors (ca. SO-90%) appear associated with regions presumed to be transcriptionally inactive. The studies described in this report have utilized both DNAase I and micrococcal nuclease because it has been shown that selective digestion by DNAase I may in certain systems be more indicative of a property imposed by chromatin conformation than by the potential for transcriptional activity (Weintraub and Groudine,
214
Edward J. Pavlik, Benita S. Katzenellenbogen
1976; Garel et al., 1977; Miller et al., 1978; Palmiter et al., 1978; Weintraub, 1979). It is important to stress though that, at best, only limited enrichment of transcriptional activity is generally achieved through chromatin fractionation methodology (Levin, 1980). Although appropriate cDNA probes are not presently available in the rat uterine system to enable clear correlations between transcriptional and chromatin fractionation properties, the fractionation methodology employed is capable of reproducibly separating chromatin components on the basis of different physical properties and hence, its application in a widely studied estrogen target tissue should provide useful information. Previous investigators have used nucleases to release bound estradiol from mouse (Harris, 1971) and rat (King and Gordon, 1972) uterine nuclei. DNAase I appeared to be more efficient than micrococcal nuclease at releasing nuclear bound estradiol and an inhibition of receptor release by Mg2+ and Ca2+ (10 mM; Harris, 1971) was indicated. Recently other investigators have reported that Ca2+ (5 mM) aggregates nuclear estrogen receptors and prevents release from chromatin (Schoenberg and Clark, 1980). Extensive digestion of lamb endometrial nuclei with micrococcal nuclease has resulted in the release of 80% of the nuclear estrogen receptors as two molecular forms (Andre et al., 1978; Rochefort and Andre, 1978) while DNAase I treatment of rat uterine nuclei only released about half of the nuclear receptors (Katzenellenbogen et al., 1980a,b). Very recently, it has been reported that estrogen receptors are enriched 3-4fold in monosomes and 7-19-fold in di- and tri-nuclebsomes of the S2 fraction prepared from the MCF-7 human breast cancer cell line (Scott and Frankel, 1980). This enrichment is based upon assessment of estrogen binding per unit DNA as determined by measuring [r4C]thymidine incorporated into DNA during cell growth (Scott and Frankel, 1980). In the studies reported here, we too have observed an enrichment in the specific activity of estrogen receptor in the S2 fraction which is here presumed to contain transcriptionally active DNA; however, the bulk of estrogen receptors are found elsewhere, in a fraction (Pl) presumed to contain transcriptionally inactive DNA. In our rat uterine studies, the enrichment of estrogen-receptor specific activity in fraction S2 is associated with the small amounts of DNA in this fraction. Our observations are compatible with those of others who have found chick oviduct estrogen receptors (Franeschi and Kim, 1979) and glucocorticoid (Levy and Baxter, 1976) and androgen receptors (Rennie, 1979) predominantly associated with transcriptionally inactive chromatin. Our observations support a model in which some chromatin-associated estrogen receptors from the rat uterus are ~distributed in chromatin fractions which have often been associated with transcriptional activity (S2) but in which the bulk of receptors are located in chromatin regions presumed to be transcriptionally inactive at early times after estrogen’ exposure, as investigated here. Since a close correlation exists between the concentration of nuclear-bound estrogen receptors and synand thesis of estrogen-induced protein, ‘IP’ in the rat uterus (Katzenellenbogen Gorski, 1972, 1975) and the number of RNA initiation sites in the chick oviduct (Kalimi et al., 1976), it would seem reasonable that the bulk of the estrogen recep-
Intranuclear
distribution of estrogen receptors
215
tors in the nucleus have a productive role in the mechanism of steroid action. Conversely, it has been suggested that only a small number of estrogen receptors productively bind to high affinity sites on the genome while the remainder bind to lower affmity, nonproductive sites (Yamamoto and Alberts, 1975). Receptors associated with productive sites could possibly be those demonstrated in fraction S2 or those receptors (ca. lo-20%) which are released from chromatin more easily than DNA (Le., Figs. 4 and 6B) is liberated. Furthermore, a dynamic interpretation of intranuclear receptor interactions has been suggested by the demonstration of nuclear receptor turnover or ‘processing’ (Clark and Peck, 1976; Juliano and Stancel, 1976;Horwitz and McGuire, 1978a,b). Hence, in moving through chromatin from sites of mechanistic initiation to sites of receptor processing, estrogen receptors may only transiently be found in ‘transcriptionally active’ fractions, while accumulating for processing in ‘transcriptionally inactive’ regions. At present, additional investigation is required to define the nuclear locations where receptors are processed and to define the temporal continuity of estrogen receptors with transcriptionally active and inactive regions.
ACKNOWLEDGEMENTS This work was supported by National Institutes of Healthgrants USPH HD06726 and CA 18119. We thank Brenda Lovett and the School of Basic Medical Sciences Word Processing Center for assistance with manuscript preparation.
REFERENCES Andre, J., Raynaud, A., and Rochefort, H. (1978) Biochemistry 17,3619-3626. Bellard, M., Gannon, F., and Chambon, P. (1978) Cold Spring Harbor Symp. Quant. Biol. 42, 779-791. Bloom, KS., and Anderson, J.N. (1978) Cell 15,141-150. Bloom, KS., and Anderson, J.N. (1979) J. Biol. Chem. 254,10532-10539. Burton, K. (1956) Biochem. J. 62,315-323. Clark, J.H., and Peck, E.J. (1976) Nature (London) 260,635-637. deBoer, W., devries, J., Mulder, E., and van der Molen, H.J. (1978) Nucleic Acids Res. 5,87103. Fallon, A.M., Jinks, C.S., Yamamoto, M., and Nomura, M. (1979) J. Bacterial. 38,383-396. Franeschi, R.T., and Kim,K.H. (1979) J. Biol. Chem. 254,3637-3646. Garel, A., ZoJan, M., and Axel, R. (1977) Proc. Natl. Acad. Sci. (U.S.A.) 74,4867-4871. Harris, G.S. (1971) Nature (London), New Biol. 231,246-248. Hayward, G.S., and Smith,M.G. (1972) J. Mol. Biol. 63,383-395. Hemminki, K. (1977) Acta Endocrinol. 84,215-224. Hemminki, K., and Vauhkonen, M. (1977) Biochim. Biophys. Acta 474,109-116. Horwitz, K.B., and McGuire, W.L. (1978a) J. Biol. Chem. 253,2223-2228. Horwitz, K.B., and McGuire, W.L. (197813) J. Biol. Chem. 253,6319-6322. Jensen, E.V., and DeSombre, E.R. (1972) Annu. Rev. Biochem. 41,203-230.
216
Edward J. Pavlik, Benita S. Katzenellenbogen
Jensen, E.V., and DeSombre, E.R. (1973) Science 182,126-134. Juliano, J.V., and Stancel, GM. (1976) Biochemistry 15,916-920. Kalimi, M., Tsai, S.Y., Tsai, M-J., Clark, J.H., and O’Malley, B.W. (1976) J. Biol. Chem. 251, 516-523. Katzenellenbogen, B.S., and Gorski, J. (1972) J. Biol. Chem. 247,1299-1305. Katzenellenbogen, B.S., and Gorski, J. (1975) in: Biochemical Actions of Hormones (Ed. G. Litwack), Vol. 3, Academic Press, New York, pp. 187-243. Katzenellenbogen, B.S., Lan, N.C., and Rutledge, S.K. (1980a) J. Steroid Biochem. 13, 113122. Katzenellenbogen, B.S., Pavlik, E.J., Lan, N.C., and Eckert, R.L. (1980b) in: The Endometrium (Ed. F.A. Kimball), Spectrum, New York, pp. 107-126. King, R.J.B., and Gordon, J. (1972) Nature (London), New Biol. 240,185-187. Levin, B.J. (1980) in: Gene Expression, Vol. 2, Wiley, New York, pp. 944-949. Levy, B.W., and Baxter, J.D. (1976) Biochem. Biophys. Res. Commun. 68,1045-1051. Miller, D.M., Turner, P., Nienhuis, A.W., Axelrod, D.E., and Gopalakrishnan, T.U. (1978) Cell 14,511-521. O’Malley, B.W., and Means, A.R. (1974) Science 183,610-620. Palmiter, R.D., MulvihiB, E.R., McKnight, G.S., and Senear, A.W. (1978) Cold Spring Harbor Symp. Quant. Biol. 42,639-647. Pavlik, E.J., and Coulson, P.B. (1976) J. Steroid Biochem. 7,357-367. Pavlik, E.J., Rutledge, S., Eckert, R.L., and KatzenelJenbogen, B.S. (1979) Exp. Cell Res. 123, 177-189. Rennie, P.S. (1979) J. Biol. Chem. 254, 3947-3952. Rochefort, H., and Andre, J. (1978) Cancer Res. 38,4229-4232. Schoenberg, D.R., and Clark, J.H. (1980) Endocrinology, 106,56-60. Scott, R.W., and Frankel, F.R. (1980) Proc. Natl. Acad. Sci. (U.S.A.) 77, 1291-1295. Senior, M.B., and Frankel, F.R. (1978) Cell 14,857-863. Weintraub, H. (1979) Nucleic Acids Res. 7,781-792. Weintraub, H., and Groudine, M. (1976) Science 193, 848-856. Williams, D., and Gorski, J. (1973) Biochemistry 12,297-306. Yamamoto, K.R., and Alberts, B.W. (1975) Cell 4,301-310. Yamamoto, K.R., and Alberts, B.W. (1976) Annu. Rev. Biochem. 45,721-746.
201
THE INTRANUCLEAR DISTRIBUTION OF RAT UTERINE ESTROGEN RECEPTORS DETERMINED AFTER NUCLEASE TREATMENT AND CHROMATIN FRACTIONATION Edward J. PAVLIK * and Benita S. KATZENELLENBOGEN
**
Department of Physiology and Biophysics, Universityof Illinois, and School of Basic Medical Sciences, Universityof Illinois, College of Medicine, Urbana,IL 61801 (U.S.A.) Received 6 July 1981; revision received 3 December 1981; accepted 29 December 1981
The intranuclear locations at which rat uterine estrogen receptors interact with chromatin have been probed using digestions performed with DNAase I and micrococcal nuclease. Exposure to nuclease has been controlled to effect limited to extensive digestion of nuclear DNA under conditions which maintain the integrity of the [sH] estradiol-receptor complex. The effect of divalent cation concentration on the release of estrogen receptors from nuclease-treated chromatin was examined and found to be of consequence above 2 mM. Exposure to nuclease released nuclear estrogen receptors from chromatm, with DNAase I being more efficient than micrococcal nuclease in mediating this release. The release of the bulk of nuclear estrogen receptors closely paralleled the nuclease-mediated digestion of chromatin DNA. At 1 h after exposure to estrogen, substantial quantities of uterine estrogen receptors (80-90%) were distributed in chromatin fractions which, on the basis of fractionation terminology, have been termed ‘transcriptionally inactive’by convention. Enrichment of estrogen receptors in chromatin which has been termed ‘transcriptionally active’ only occurred with lo-20% of the estrogen receptors. Hence, our findings support a model where, at early times after estrogen exposure, receptors from the rat uterus are enriched to only a minor extent in chromatin to which ‘transcriptional activity’ is generally assigned while the bulk of receptors are localized in chromatin which is generally considered ‘transcriptionally inactive’. Keywords: nuclear estrogen receptors; tionation; uterus.
DNAase I; micrococcal nuclease; chromatin frac-
Steroid hormones exert their effects upon target tissues after passing through the cell membrane and interacting with a cytoplasmic binding protein which as an activated binding complex translocates to the nucleus. The ensuing intranuclear
* Present address: Departments of Obstetrics and Gynecology and Biochemistry, University of Kentucky, Lexington, KY 40503 (U.S.A.). ** Address correspondence to: Dr. Benita S. Katzenellenbogen, Department of Physiology and Biophysics, 524 Burrill Hall, 407 South Goodwin Avenue, University of Illinois, Urbana, IL 61801 (U.S.A.). 0303-7207/g2/0000-0000/$02.75
0 Elsevier/North-Holland
Scientific Publishers, Ltd.
Edward J. Pavlik,Benita S. Katzenellenbogen
202
interactions presumably initiate certain molecular events which ultimately are responsible for alterations in the rates of transcription, translation and replication (for reviews see Jensen and DeSombre, 1972, 1973; O’Malley and Means, 1974; Katzenellenbogen and Gorski, 1975; Yamamoto and Alberts, 1976). At present very little is known about the intranuclear interactions in which estrogen receptors are involved; even less is known about the intranuclear location of these interactions. Attempts to resolve the intranuclear location of receptor-chromatin associations have utilized fractionations which define transcriptionally active and inactive regions using either mechanical (Levy and Baxter, 1976; deBoer et al., 1978; Franeschi and Kim, 1979) or enzymatic (Hemminki, 1977; Hemminki and Vauhkonen, 1977; Senior and Frankel, 1978; Rennie, 1979; Scott and Frankel, 1980) disruptions of chromatin; however, the results have not been unambiguous. Estrogen receptors from chick oviduct (Franeschi and Kim, 1979) and rat uterus (deBoer et al., 1978) have been reported to associate predominately with fast sedimenting, transcriptionally inactive fractions. Similar observations have been reported for. androgen receptors in prostatic chromatin (Rennie, 1979) and glucocorticoid receptors in pituitary chromatin (Levy and Baxter, 1976). Conversely, other investigators have reported that transcriptionally active regions of chromatin are enriched in estrogen receptors (Hemminki, 1977; Hemminki and Vauhkonen, 1977; Scott and Frankel, 1980). Since these opposing reports can be the result of differences in methodology or in the application of a methodology, this report examines the intranuclear distribution of rat uterine estrogen receptors using nuclease-mediated chromatin fractionation techniques. In these studies, nuclease-mediated hydrolysis of DNA has been finely controlled so that the release of estrogen receptors from rat uterine chromatin was studied under a spectrum of nucleolytic DNA hydrolyses, ranging from mild to extensive digestion of nuclear DNA. In particular, this approach has resulted in rapid chromatin fractionations which minimize the opportunities for estrogen receptors to reassociate during preparation.
MATERIALS
AND METHODS
Preparation of nuclei
Immature female Sprague-Dawley rats were obtained from Holtzman, Madison, WI, and used between ages 20-23 days. Uteri were debrided of fat and mesentery and placed in HeLa medium (DIFCO), pH 7.4-7.6 adjusted with CO2 gassing at 0-4°C. Estrogen receptors were translocated to the nucleus by exposing the uteri in vitro to receptor saturating concentrations (50 nM) of [3H] estradiol and to [3H] estradiol in the presence of 200-fold excess diethylstilbestrol (DES) to assess nonspecific binding (Williams and Gorski, 1973; Pavlik et al., 1979). After 60 min at 37°C the translocation reaction was terminated by transferring the uteri to iced buffer (5 ml) containing 50 mM Tris, 25 mM KCl, 0.5 M sucrose, pH 7.4 (0.5 M S-TK). After cooling on ice for 5 min, the uteri were homogenized at 0-4”C with
Intranuclear distribution of estrogen receptors
203
seven 3-set bursts of a Polytron sonic/shear homogenizer (Brinkman Industries) set at 40% of maximum speed. All subsequent manipulations were performed at 4°C. The homogenate was centrifuged in an HB-4 swinging bucket rotor (Sorvall) at 6500 Xg for 5 min in a RC-2B centrifuge (Sorvall). The supernatant (SA) was aspirated and the pellet was resuspended in TK buffer containing 1.5 M sucrose (1.5 M S-TK) using a loose-fitting, round bottom !eflon homogenizer. Final resuspension was performed by 10 aspirations through the bore of a standard Pasteur pipette. The resuspended material was centrifuged at 0-4”C for 20 min at 13 000 X g. The pelleted material was suspended in TK buffer containing 0.25 M sucrose (0.25 M S-TK) and inspected under a phase-contrast microscope. The resulting supernatant (Sn) was saved. Nuclei purified in this manner and judged to be free of major non-nuclear contaminants by microscopic examination were used in studies employing nucleases. If the purity of the preparation was questionable, the suspension was underlaid in 1.5 M S-TK and centrifuged at 1300 X g for 20 min. Nuclear yields by this protocol were 40-60%. Agarose gel electrophoresis Nuclei were suspended in buffer containing EDTA (15 mM), EGTA (15 mM), sodium dodecylsulfate (2%) and NaCl (2.3 M). DNA was extracted 3 times with chloroform : isoamylalcohol (24 : 1) and precipitated overnight with 2.5 vol. of ethanol at -20°C. DNA was pelleted by centrifugation at 16 000 Xg for 10 min and the pellet was lyophilized and stored at -20°C. Before use, the DNA was dissolved in 5 mM EDTA and the DNA concentration determined by absorption at 260 nm. Agarose (1.5%) in TANE buffer (40 mM Tris, 20 mM sodium acetate, 18 mM NaCl, 2 mM NasEDTA, pH 8 .O) was dissolved by brief boiling and was poured into glass beds. Wells were formed with Teflon combs, and after gelling cloth wicks were applied to both ends thereby connecting the gel to the buffer reservoirs, both of which contained TANE buffer (Hayward and Smith, 1972). Gels were run at room temperature at 10 mA/gel. DNA was visualized by UV-induced fluorescence after staining with ethidium bromide (10 mg/ml). Sized DNA fragments were generated by Hind II and Hind III restriction endonuclease cuts of plasmid p NO1001 which carries 7 ribosomal protein genes (Fallon et al., 1979). Determination of estrogen binding Soluble estrogen receptors were quantitatively determined by adsorption on hydroxylapatite (Pavlik and Coulson, 1976). In general, receptors were exposed to saturating concentrations of ( [3H]estradiol, 50 r&l) in the presence and absence of 200-fold excess nonradioactive DES, and were bound to hydroxylapatite which was washed 3 times with Tris buffer (10 mM), pH 7.4 to remove free steroid. After extraction with ethanol, saturable and competitive binding was taken to represent the binding by the limited capacity estrogen receptor and was determined as the difference between total binding measured in the presence of labeled ligand alone and nonsaturable binding measured in the presence of labeled ligand mixed with
204
Edward J. Pavlik, Benita S. Katzenellenbogen
excess competitor. Chromatinbound estrogen receptors were determined as described (Katzenellenbogen et al., 1980a,b) after adding hydroxylapatite to the preparation to bind any receptor, which might dissociate from chromatin, and by centrifugally washing chromatin pellets 3 times with Tris buffer followed by ethanol extraction of bound steroid. Specific binding was determined as outlined above. Estrogen receptor was normalized relative to chromatin, and DNA was determined as described by Burton (1956). Nuclease digestions DNAase I (Sigma Chemical and Worthington Biochemical Corp.) and micrococcal nuclease (Worthington Biochemical Corp.) were aliquoted to 80-100 pg portions and stored at -20°C. On the day of use, a nuclease diluent was prepared which contained Mg2+ and Ca2+ (4 : 1) to a final concentration as specified in individual experiments. The nuclease digestion was performed in the presence of 0.15 M sucrose, 30 mM Tris, and 15 mM KCl, pH 7.4. Unless otherwise stated, digestions were performed at 0-4°C and terminated after 4 min by bringing the digest to 5 mM EDTA/S mM EGTA, pH 7.4. Chromatin was then pelleted at 6500 Xg for 5 min and the preparation processed as described. Nuclease activity was verified spectrophotometrically with one Kunitz unit (KU) of activity defined as an increase in base-pair absorbance due to enzymatic hydrolysis of 1000A2e0 units/mm/l at pH 7.4,25”C. Chromatin fractionation terminology DNA associated with nuclei or chromatin which had been washed free of sucrose and pelleted centrifugally is referred to as ‘chromatin DNA’. All material having a UV absorbance at 260 nm which is not pelleted but found soluble in the supernatant after 0.3 N perchlorate treatment is referred to as ‘acid-soluble DNA’ and very likely consists of mostly small size classes of DNA. Material which is released from chromatin after lysis (5 min) with hypotonic Tris (10 mM) is referred to as ‘A 260 material released by lysis.’ Chromatin fractionation Fractionation described by Scott and Frankel (1980) was employed, After nuclei were prepared and exposed to nuclease, the reaction was stopped by bringing the suspension to 5 mM EDTA/EGTA and the chromatin pelleted at 6500 Xg for 5 min (2°C). The fractionation procedure is outlined in Fig. 1. The chromatin was resuspended in TloNloEl [0.5 ml: Tris (10 mM), NaCl (10 mM), EDTA (1 mM), pH 7.41. After 20 min at 2”C, lysis was completed by adding 1.5 ml of TloNlo buffer and centrifuging at 6500X g for 5 min. The pellet (Pl) was processed as described earlier. The supernatant (Sl) was brought to 3 mM MgCL for 30 min at 2°C and centrifuged at 12 000 Xg for 10 min in order to pellet Mg’+-insoluble material (P2). The resulting supernatant (S2) contained Mg’+-soluble material. Estrogen receptor activity was determined in Sl , S2, Pl, l?2 fractions. A260, as well
Intranuclear distribution
of estrogenreceptors
205
as perchlorate (0.3 N) soluble Ash0 activity, was determined for all supernatants. _4s6e was determined in fraction P2 by redissolving the precipitate in TloNloEl buffer, while DNA was assayed in fraction Pl as described by Burton (1956).
TRANSLOCATE
ESTROGEN
RECEPTOR
HOMOGENIZE
I AND PURIFY
NUCLEI
STOP NUCLEASE
1 WITH EDTAiEGTA
t5mMJ
I CENTRIFUGE 16500 xg. 5 min.)
N”CLEAk LYSIS Suspend in 0.5 ml TIDNIOEI. After 20 min. 0% add I. 5 ml TlONlD.
1
CENTRIFUGE t65W xg, 5 min.
Bring to 3 mM Mgil2
(34 min.,
4 CENTRIFUGE IKVDD xg. IO min.
I
Mg+t
52 solubles
I
DV
I
I 2 Mg++ insoluble
Fig. 1. Flow diagram of chromatin fractionation.
RESULTS Hydrolysis of rat uterine DNA by DNAase I DNAase I is able to hydrolyze the DNA in preparations of rat uterine nuclei and hence reduces the DNA content of these preparations as shown in Fig. 2. Hydrolysis proceeds more completely at 30°C than 0°C; however, sufficient degradation does occur at O”C, a temperature which ensures maximal receptor stability. By anlyzing the size of DNA fragments generated by DNAase I hydrolysis using agarose gel electrophoresis in Fig. 3, a progressive increase in small size classes of DNA can be observed as the concentrations of nuclease increase_ Limited digestion of uterine nuclear DNA can be seen when l-10 KU of DNAase per ml was employed, while higher concentrations of nuclease more completely degrade native nuclear DNA to smaller forms. Finally, endogenous nucleases do not appear to be activated in uteri incubated for 60 min at both 0°C and 30°C as shown in the right-most lanes in Fig. 3. Thus, exogenously provided DNAase I is responsible for DNA fragmentation and not some endogenous nuclease, which becomes activated during the course of preparation or experimental manipulation.
3 OL, 0 R
100
200 DNase I,
300
400
500
K.U./ml
Fig. 2. DNAase I degradation of rat uterine nuclear DNA. Nuclei were prepared as described. One uterine equivalent of DNA was defined as 250 ug (Williams and Go&i, 1973) and yields were calculated and normalized on this basis. The yield after preparation based upon DNA content was 92%. Each aliquot contained 362 ug rat uterine DNA (-1.45 uterine equivalent) in a reaction volume of 1 ml: CaCla (05 mM) and MgClz (2 mM) in buffer containing (0.15 M sucrose, 30 mM Tris, 15 mM KCl, pH 7.4). Aliquots were treated at 0°C and 30°C for 4 min with DNAase I (2135 KU/mg), and the reaction was stopped by bringing the suspension to 5 mM EDTA, EGTA, using buffered (pH 7.4) 100 mM EDTA, EGTA. Chromatin was pelleted and brought to 0.3 N inperchlorate, held at O’C for 30 min, washed (3X) with 0.3 N perchlorate and hydrolyzed in 0.3 N perchlorate (30 min, 90°C). DNA was determined according to Burton (1956).
eiP
Fig. 3. Analysis of the degradation of rat uterine nuclear DNA by DNAase I using agarose-gel electrophoresis. Nuclei were prepared as described at 50% yield and aliquoted to a reaction mixture concentration of 100 fig DNA/ml. Nuclease treatment was conducted at O’C for 4 min and terminated as described in Fig. 1. All other preparative steps are outlined in the Methods section. Approx. 20 &g DNA were applied to each well. The stability of nuclear DNA obtained from uteri incubated for 60 min at 0°C (well No. 14) and 37°C (well No. 15) is also shown. The left-most well is well No. 1.
Intranuclear
distributzon of estrogen receptors
201
Release of nuclear estrogen receptors by nucleases and the role of DNA degradation Under conditions where low concentrations of DNAase I are causing limited digestion of DNA, nuclear estrogen receptor disappears from chromatin as shown in Fig. 4. This disappearance, however, is predominately coordinated with the disappearance of DNA from the nuclei since reexpression of these data in terms of DNA recovered after nuclease treatment reveals a relatively constant relationship between receptor and nuclear DNA. Because the disappearance of chromatin associated receptor was modest (_5-10%) when compared to the disappearance of DNA and because the range of DNAase I concentration was kept low, factors which might limit the release of nuclear estrogen receptor from nuclease-hydrolyzed chromatin, as weJ1 as the effects which might be mediated by submaximal DNA degradations, were examined. Because Mg2+ and Ca2* are required for nuclease activity, it is necessary to include these divalent cations in nuclear reaction buffers; however, divalent cations can cause estrogen receptor aggregation (Harris, 1971; Schoenberg and Clark, 1980).
I
I
,
0
5
I
IO DNose I,
I
15
I
20
I
25
K.U./ml
Fig. 4. Release of nuclear estrogen receptors by low concentrations of DNAase I. Estrogen receptor was translocated to the nucleus and nuclei were prepared as described. Estrogen receptor distrrbution: 0.71 pmole/uterine equivalent DNA was chromatin-associated (Pl), while 0.17 pmoles/uterine equivalent DNA remained soluble (SA). Based upon DNA content the yield of partially purified nuclei was -80%. Nuclei were treated with DNAase I (2135 KU/mg) after resuspension to 0.15 M sucrose, 30 mM Tris, 15 mM KCl, pH 7.4 containing CaC12 (0.5 mM final concentration) and MgCla (2 mM final concentration). The reaction volume (1 ml) contained -500 Ng rat uterine DNA. The nuclease reaction was conducted at 0°C for 4 min and terminated by bringing the aliquot to 5 mM EDTA, 5 mM EGTA by adding 50 ~1 of 100 mM EDTA/EGTA buffered at pH 7.15. After processing the preparations as described in Methods and Materials, the DNA content of both nuclease-treated and untreated aliquots was determined and the results expressed as nuclear estrogen receptor per aliquoted amount of DNA determined in untreated samples (left ordinate) or as nuclear estrogen receptor per amount of DNA remaining in the preparation after nuclease treatment (i.e., ‘recovered’ DNA and right ordinate). Untreated preparations were set at 100%.
Edward J. Pavlik, Benita S. Katzenellenbogen
208
Hence, it is important to examine the degradation of nuclear DNA and the release of nuclear estrogen receptor as a function of divalent cation concentration in order to determine if receptor, when induced to aggregate by Mg2+ and Ca2+, will consequently be unable to diffuse out of nuclease-treated chromatin. When Mg2+ (and Ca2+) are varied, maximal release of estrogen receptor and maximal degradation of nuclear DNA are achieved at Mg2+ concentrations of OS-2 mM (Fig. 5). In addition, solubilized A2e0 material is maximized in this concentration range. At very (10 mM), estrogen-receptor aggregation prevents the high Mg2+ concentrations release of all receptors from chromatin and chromatin appears stabilized against the lytic release of A260 material, while other parameters are minimally influenced. of tissue homogenates may Because the endogenous Mg2+/Ca2+ concentration
IOO-
so8070M e--o O--O w -
60B G
50-
Chromatin EzR Chromotm DNA Total APeomaterial Acid sol.A2eomat. A,eomateriol released by lys~s
_
h 40302D-
IO t k
0
I
2
I
I
8
I
4
6
8
IO
[WP]
. mM
I
Fig. 5. The effect of magnesium ion on DNAase I activity and release of estrogen receptors from chromatin. Nuclei were prepared as described at 65% yield. Estrogen-receptor distribution: 0.48 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.08 pmoles/ uterine equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -260 E.rg rat uterine DNA and 500 KU/ml DNAase I (2135 KU/mg). The reaction was conducted and terminated as described in Fig. 4 with the exception of Mga+ concentration which was varied as described. A constant relationship of Mga+:Ca2+ was maintained at 4 : 1. ‘Chromatin Ea R’ was determined as the chromatin associated estrogen receptor/aliquoted DNA and ‘chromatin DNA’ was the DNA remaining after nuclease treatment determined according to Burton (1956).
209
Intranuclear distribution of estrogen receptors
moderate the release of receptor from chromatin, all subsequent nucleate degradations were performed at 2 divalent cation concentrations: at a Mg2+ concentration well below the optima demonstrated in Fig. 5 in order to compensate for the possible inclusion of endogenous divalent cations as well as at a Mg2+ concentration (2 mM) which is optimal for nuclease activity. Both the disappearance of DNA from nuclei and the increase in DNAase I solubilized A260 material were similar when divalent cation concentrations were low (Fig. 6A) or high (Fig. 6B); however, with higher divalent cation concentrations more A 2 6Omaterial was released by lysis of nuclei and more of the DNAase I solubilized material was hydrolyzed to small, acid-soluble fragments. Under conditions of low divalent cation concentration (Fig. 6A), the disappearance of estrogen receptors from chromatin more closely paralleled
Mg+c.
100
- 0.1 0.025
Co++
mM
O-O .--a o---o H
Chromotin EpR Chromotin DNA Total A26omateriol Acid sol.A~+~mat. Aesomaterml released by lysis
_ -
B 0
I
100
I
200
I
I
300
400
I
500
I
DNose I, K.U./ml
Fig. 6. DNAase I digestion of rat uterine nuclei under low and high concentrations of Mga+ and Caa+. (A) Low divalent cation concentration. Nuclei were prepared as described at 40% yield. Estrogen*eceptor distribution: 1.05 pmole/uterine equivalent DNA was chromatm associated (pl) while 0.45 pmoles/uterme equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -130 fig rat uterine DNA. Conditions were as described in Figs. 3 and 4. (B) High divalent cation concentration. Nuclei were prepared as described at 40% yield. Estrogen-receptor distribution 0.61 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.25 pmoles/uterine equivalent DNA remained soluble (SA). All other conditions were as described above.
210
Edward .I. Pavlik, Benita S. Katzenellenbogen
the loss of DNA from chromatin than under conditions of high divalent cation concentration (Fig. 6B) where the disappearance of estrogen receptors from chromatin modestly exceeded the loss of DNA from nuclei by -5-10%. Similar experiments were performed with micrococcal nuclease (Fig. 7). In these experiments, micrococcal nuclease liberated DNA from uterine nuclei but released much less receptor from chromatin than DNAase I under both low and high concentrations of divalent cations. Moreover, micrococcal nuclease prepared with high divalent cation concentrations (Fig. 7B) had no greater activity than nuclease prepared with low divalent ion concentrations (Fig. 7A) in terms of AZ6,, material released by nuclear lysis or in terms of A 260 material hydrolyzed to small, acidsoluble fragments.
00
0 _
- 0t loo-
0
o--_/ ‘k
80-
YTA -
‘*.. _-_
60-
*-----___+___~,+-4
--__ 60-
II
0
I
-
~-_-_______-__-“---
I
200 Micrococcal
I
I
I
+irkd
400 Nuclease
(U./ml)
Fig. 7. Micrococcal nuclease digestion of rat uterine nuclei with low and high concentrations of Mga+ and Ca2+. (A) Low divalent cation concentration. Nuclei were prepared as descrrbed at 55% yield. Estrogen-receptor distribution: 0.62 pmoles/uterme equivalent DNA was chromatin associated (Pl) while 0.20 pmoles/uterine equivalent DNA remained soluble (SA). The reaction volume (1 ml) contained -170 Mg rat uterine DNA. Conditions were as descrrbed in Fig. 5. (B) High divalent cation concentration. Nuclei were prepared as described at 60% yield. Estrogen-receptor distribution: 0.38 pmoles/uterine equivalent DNA was chromatin associated (Pl) while 0.20 pmoles/uterine equivalent DNA remained soluble (SA). Micrococcal nuclease (Worthington: 23882 lJ/mg) was used and conditions were as described.
0.1
2.1
(-)
85.4 (0.9) 100.0
0
-(-)
0
0
0.6
6.5
(-)
(-)
(-)
78.4 ~ (1 .O) 78.8
-
0
0
9.3
14.4 -(3.3) 4.3
5
0
0.7 (-)
(-)
(2.9)
74.0 (1 .O) 70.6
-
5.4 o(-)
J+
46
13.3
10
(0.8)
(-)
70.0 (1.6) 43.3
$
z+
9.1 +1.6)
13.7 (1.9) 7.3
50
44.8 (1.0) 44.1
4.4 __ (0.5) 9.3
+.(-)
14.9 E(l.2)
17.1 (1.6) 10.5
100
(1.6)
(0.4)
39.5 __ (1.3) 29.9
2$
14.4 - 0.8 (18)
26.2 -16.2 (1.6)
139
22.5
250
15.6 L(O.6) 25.2
8.2 -(0.3) 23.9
27.2 -(17) 1.6
27.1 1591.7)
20.9 p(1.2) 17.9
500
a DNAase treatments were as described in the text in the presence of MgCls (2 mM) and CaCls (0.5 mM) for 4 mm at 0°C. Receptor values are expressed as a percent of total recovered estrogen receptor in the absence of nuclease and are the result of averaged duplicates. (Estrogen receptor distribution: Nuclear 0.60 pmoles/uterine equivalent DNA; cytoplasmic 0.24 pmoles/uterine equivalent DNA. One uterine equivalent of DNA is 250 pg.) Values for DNA are expressed as a percent of total aliquoted DNA (270 &g/tube) and are the result of averaged duplicates. b Numbers presented in the Table are the estrogen receptor values (numerator), DNA values (denominator), and the ratio of receptor to DNA presented in parentheses.
Pl
P2
s2
(-)
1.4
Sl
0
10.9 -(2.9)b 3.8
0
DNAase I (KU/ml)
SO
Fraction
Table 1 Chromatin fractionation and estrogen receptor distribution and DNA distribution after treatment with DNAase I a
z r
< 3
5 G !j 2
s 9
5.
2 z 2 B A 3 3. P
0 3.0 (-)
0
s2
P2
-
0
114.9
86.0
0
+
7.7 T(3.7)
(0.7)
(0)
(-)
14.4 5.0 (2.9)
5
(2.1)
0 (0)
69.2 9820.7)
0
-02.6 (-)
g
10.6 T(3.4)
10
88
68.0
0.1 x(05)
(0.8)
0 3.2 (-)
5.0 - 1.9 (2.6)
12.4 4.4 (2.8)
50
(0.1)
(4.2)
57.3 89.0 (0.6)
z
T(44) 4.4
g
11.2 5.6 (2.0)
100
89
62.8 (0.7)
0.9 6.4 (0.1)
o.l3.6 (36)
6.9 26 (2.7)
12.9 90 (1.4)
250
nuclease
(14.8)
72.7 p(1.3) 55.7
3.1 -10.0 (0.3)
;
16.4 49 (3.3)
16.7 1521.1)
500
a
a Nuclease treatments were as described in Table 1. Receptor values are expressed as a percent of total recovered estrogen receptor in the absence of nuclease and are the result of averaged duplicates. (Estrogen receptor distrbution: Nuclear 0.73 pmoles/uterine equivalent DNA; cytoplasmic 0.23 pmoles/uterine equivalent DNA. One uterine equivalent of DNA is 250 Mug).Values for DNA are expressed as a percent of total aliquoted DNA (300 ug/tube) and are the result of averaged duplicates. b Numbers presented in the Table are the estrogen-receptor value (numerator), DNA value (denominator), and the ratio of receptor to DNA presented in parentheses.
80.3 (0.8) 100.0
(0)
4.5 __2.0 (2.3)
Sl
0
12.3 46 (2.7) b
0
Micrococcal nuclease (U/ml)
so
Fraction
Table 2 Chromatin fractionation and estrogen receptor distribution and DNA distribution after treatment with micrococcal
;i’ 9 3 h 9 : 2 9 2
$ e Q .S
Irmanucleardistributionof estrogenreceptors
213
Chromatin fractionation and localization of estrogen receptors after exposure to DNAase I and micrococcal nuclease The general scheme and nomenclature for chromatin fractionation is outlined in Fig. 1. Receptor localization and DNA distribution data after treatment with a wide range of concentrations of nucleases are presented in Tables 1 and 2. DNAase I appeared to release more DNA from nuclei and to generate more AZbOmaterial by nuclear lysis than micrococcal nuclease. Both DNAase I (Table 1) and micrococcal nuclease (Table 2) treatment resulted in very small amounts of AZhOmaterial in the Mg’+-soluble (S2) fraction. In comparing the effects of DNAase I and micrococcal nuclease on estrogen receptors, DNAase I appeared also to more efficiently release estrogen receptors from chromatin than micrococcal nuclease. Second, the Mg2+soluble fraction appears enriched in estrogen receptors as shown by this fractionation study of the distribution of estrogen binding (Tables 1 and 2), and the receptor specific activity (i.e., receptor/pg DNA) is highest in this fraction (ca. 9 pmoles/ 250 I-(g DNA following 500 KU/ml DNAase or 500 U/ml micrococcal m&ease exposure). This latter result, which represents an approx. 15-fold enrichment in receptors, occurs chiefly because the soluble A 260 material in fraction S2 is very low and not because a large proportion of estrogen receptors are distributed in this fraction. As indicated in the Tables, only ca. lo-25% of total estrogen receptors are found in the S2 fraction.
DISCUSSION The data presented in this report have been collected by making use of a very well characterized estrogen-receptor system and by utilizing nuclease-mediated chromatin fractionation. This fractionation methodology has been reported to differentiate transcriptionally active DNA from transcriptionally inactive DNA (Weintraub and Groudine, 1976; Senior and Frankel, 1978; Bloom and Anderson, 1979; Scott and Frankel, 1980). The cogent observations reported here are as follows. (1) Estrogen receptors are released from chromatin by treatment with nucleases, and DNAase I appears more efficient than micrococcal nuclease in mediating this release. (2) Nucleaskmediated release of estrogen receptors closely parallels the nucleolytic disappearance of chromatin DNA. (3) There is some enrichment of estrogen receptors in chromatin fraction S2 which is herein designated as ‘presumed to be transcriptionally active’, but the bulk of estrogen receptors (ca. SO-90%) appear associated with regions presumed to be transcriptionally inactive. The studies described in this report have utilized both DNAase I and micrococcal nuclease because it has been shown that selective digestion by DNAase I may in certain systems be more indicative of a property imposed by chromatin conformation than by the potential for transcriptional activity (Weintraub and Groudine,
214
Edward J. Pavlik, Benita S. Katzenellenbogen
1976; Garel et al., 1977; Miller et al., 1978; Palmiter et al., 1978; Weintraub, 1979). It is important to stress though that, at best, only limited enrichment of transcriptional activity is generally achieved through chromatin fractionation methodology (Levin, 1980). Although appropriate cDNA probes are not presently available in the rat uterine system to enable clear correlations between transcriptional and chromatin fractionation properties, the fractionation methodology employed is capable of reproducibly separating chromatin components on the basis of different physical properties and hence, its application in a widely studied estrogen target tissue should provide useful information. Previous investigators have used nucleases to release bound estradiol from mouse (Harris, 1971) and rat (King and Gordon, 1972) uterine nuclei. DNAase I appeared to be more efficient than micrococcal nuclease at releasing nuclear bound estradiol and an inhibition of receptor release by Mg2+ and Ca2+ (10 mM; Harris, 1971) was indicated. Recently other investigators have reported that Ca2+ (5 mM) aggregates nuclear estrogen receptors and prevents release from chromatin (Schoenberg and Clark, 1980). Extensive digestion of lamb endometrial nuclei with micrococcal nuclease has resulted in the release of 80% of the nuclear estrogen receptors as two molecular forms (Andre et al., 1978; Rochefort and Andre, 1978) while DNAase I treatment of rat uterine nuclei only released about half of the nuclear receptors (Katzenellenbogen et al., 1980a,b). Very recently, it has been reported that estrogen receptors are enriched 3-4fold in monosomes and 7-19-fold in di- and tri-nuclebsomes of the S2 fraction prepared from the MCF-7 human breast cancer cell line (Scott and Frankel, 1980). This enrichment is based upon assessment of estrogen binding per unit DNA as determined by measuring [r4C]thymidine incorporated into DNA during cell growth (Scott and Frankel, 1980). In the studies reported here, we too have observed an enrichment in the specific activity of estrogen receptor in the S2 fraction which is here presumed to contain transcriptionally active DNA; however, the bulk of estrogen receptors are found elsewhere, in a fraction (Pl) presumed to contain transcriptionally inactive DNA. In our rat uterine studies, the enrichment of estrogen-receptor specific activity in fraction S2 is associated with the small amounts of DNA in this fraction. Our observations are compatible with those of others who have found chick oviduct estrogen receptors (Franeschi and Kim, 1979) and glucocorticoid (Levy and Baxter, 1976) and androgen receptors (Rennie, 1979) predominantly associated with transcriptionally inactive chromatin. Our observations support a model in which some chromatin-associated estrogen receptors from the rat uterus are ~distributed in chromatin fractions which have often been associated with transcriptional activity (S2) but in which the bulk of receptors are located in chromatin regions presumed to be transcriptionally inactive at early times after estrogen’ exposure, as investigated here. Since a close correlation exists between the concentration of nuclear-bound estrogen receptors and synand thesis of estrogen-induced protein, ‘IP’ in the rat uterus (Katzenellenbogen Gorski, 1972, 1975) and the number of RNA initiation sites in the chick oviduct (Kalimi et al., 1976), it would seem reasonable that the bulk of the estrogen recep-
Intranuclear
distribution of estrogen receptors
215
tors in the nucleus have a productive role in the mechanism of steroid action. Conversely, it has been suggested that only a small number of estrogen receptors productively bind to high affinity sites on the genome while the remainder bind to lower affmity, nonproductive sites (Yamamoto and Alberts, 1975). Receptors associated with productive sites could possibly be those demonstrated in fraction S2 or those receptors (ca. lo-20%) which are released from chromatin more easily than DNA (Le., Figs. 4 and 6B) is liberated. Furthermore, a dynamic interpretation of intranuclear receptor interactions has been suggested by the demonstration of nuclear receptor turnover or ‘processing’ (Clark and Peck, 1976; Juliano and Stancel, 1976;Horwitz and McGuire, 1978a,b). Hence, in moving through chromatin from sites of mechanistic initiation to sites of receptor processing, estrogen receptors may only transiently be found in ‘transcriptionally active’ fractions, while accumulating for processing in ‘transcriptionally inactive’ regions. At present, additional investigation is required to define the nuclear locations where receptors are processed and to define the temporal continuity of estrogen receptors with transcriptionally active and inactive regions.
ACKNOWLEDGEMENTS This work was supported by National Institutes of Healthgrants USPH HD06726 and CA 18119. We thank Brenda Lovett and the School of Basic Medical Sciences Word Processing Center for assistance with manuscript preparation.
REFERENCES Andre, J., Raynaud, A., and Rochefort, H. (1978) Biochemistry 17,3619-3626. Bellard, M., Gannon, F., and Chambon, P. (1978) Cold Spring Harbor Symp. Quant. Biol. 42, 779-791. Bloom, KS., and Anderson, J.N. (1978) Cell 15,141-150. Bloom, KS., and Anderson, J.N. (1979) J. Biol. Chem. 254,10532-10539. Burton, K. (1956) Biochem. J. 62,315-323. Clark, J.H., and Peck, E.J. (1976) Nature (London) 260,635-637. deBoer, W., devries, J., Mulder, E., and van der Molen, H.J. (1978) Nucleic Acids Res. 5,87103. Fallon, A.M., Jinks, C.S., Yamamoto, M., and Nomura, M. (1979) J. Bacterial. 38,383-396. Franeschi, R.T., and Kim,K.H. (1979) J. Biol. Chem. 254,3637-3646. Garel, A., ZoJan, M., and Axel, R. (1977) Proc. Natl. Acad. Sci. (U.S.A.) 74,4867-4871. Harris, G.S. (1971) Nature (London), New Biol. 231,246-248. Hayward, G.S., and Smith,M.G. (1972) J. Mol. Biol. 63,383-395. Hemminki, K. (1977) Acta Endocrinol. 84,215-224. Hemminki, K., and Vauhkonen, M. (1977) Biochim. Biophys. Acta 474,109-116. Horwitz, K.B., and McGuire, W.L. (1978a) J. Biol. Chem. 253,2223-2228. Horwitz, K.B., and McGuire, W.L. (197813) J. Biol. Chem. 253,6319-6322. Jensen, E.V., and DeSombre, E.R. (1972) Annu. Rev. Biochem. 41,203-230.
216
Edward J. Pavlik, Benita S. Katzenellenbogen
Jensen, E.V., and DeSombre, E.R. (1973) Science 182,126-134. Juliano, J.V., and Stancel, GM. (1976) Biochemistry 15,916-920. Kalimi, M., Tsai, S.Y., Tsai, M-J., Clark, J.H., and O’Malley, B.W. (1976) J. Biol. Chem. 251, 516-523. Katzenellenbogen, B.S., and Gorski, J. (1972) J. Biol. Chem. 247,1299-1305. Katzenellenbogen, B.S., and Gorski, J. (1975) in: Biochemical Actions of Hormones (Ed. G. Litwack), Vol. 3, Academic Press, New York, pp. 187-243. Katzenellenbogen, B.S., Lan, N.C., and Rutledge, S.K. (1980a) J. Steroid Biochem. 13, 113122. Katzenellenbogen, B.S., Pavlik, E.J., Lan, N.C., and Eckert, R.L. (1980b) in: The Endometrium (Ed. F.A. Kimball), Spectrum, New York, pp. 107-126. King, R.J.B., and Gordon, J. (1972) Nature (London), New Biol. 240,185-187. Levin, B.J. (1980) in: Gene Expression, Vol. 2, Wiley, New York, pp. 944-949. Levy, B.W., and Baxter, J.D. (1976) Biochem. Biophys. Res. Commun. 68,1045-1051. Miller, D.M., Turner, P., Nienhuis, A.W., Axelrod, D.E., and Gopalakrishnan, T.U. (1978) Cell 14,511-521. O’Malley, B.W., and Means, A.R. (1974) Science 183,610-620. Palmiter, R.D., MulvihiB, E.R., McKnight, G.S., and Senear, A.W. (1978) Cold Spring Harbor Symp. Quant. Biol. 42,639-647. Pavlik, E.J., and Coulson, P.B. (1976) J. Steroid Biochem. 7,357-367. Pavlik, E.J., Rutledge, S., Eckert, R.L., and KatzenelJenbogen, B.S. (1979) Exp. Cell Res. 123, 177-189. Rennie, P.S. (1979) J. Biol. Chem. 254, 3947-3952. Rochefort, H., and Andre, J. (1978) Cancer Res. 38,4229-4232. Schoenberg, D.R., and Clark, J.H. (1980) Endocrinology, 106,56-60. Scott, R.W., and Frankel, F.R. (1980) Proc. Natl. Acad. Sci. (U.S.A.) 77, 1291-1295. Senior, M.B., and Frankel, F.R. (1978) Cell 14,857-863. Weintraub, H. (1979) Nucleic Acids Res. 7,781-792. Weintraub, H., and Groudine, M. (1976) Science 193, 848-856. Williams, D., and Gorski, J. (1973) Biochemistry 12,297-306. Yamamoto, K.R., and Alberts, B.W. (1975) Cell 4,301-310. Yamamoto, K.R., and Alberts, B.W. (1976) Annu. Rev. Biochem. 45,721-746.