A cytoplasmic thyroid hormone binding protein: characterization using

Transcriptional stimulation by thyroid hormone of a cytosolic thyroid hormone binding protein which is homologous to a subunit of pyruvate kinase M1...
0 downloads 0 Views 3MB Size
Biochemistry 1989, 28, 617-623

617

A Cytoplasmic Thyroid Hormone Binding Protein: Characterization Using Monoclonal Antibodies Toru Obsta,$ Takaaki Fukuda,$ Mark C. Willingham,t Chi-Ming Liang,s and Sheue-yann Cheng*.* Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda. Maryland 20892, and Division of Blood and Blood Products, Food and Drug Administration, Bethesda, Maryland 20892 Received May 31, 1988; Revised Manuscript Received September 14, 1988

ABSTRACT: W e have previously purified a cellular thyroid hormone binding protein (p58) from a human carcinoma cell line [Kitagawa, S., Obata, T., Hasumura, S., Pastan, I., & Cheng, S.-y. (1987) J. Biol. Chem. 262, 3903-39083. In the present study, the binding characteristics, the molecular properties, and subcellular localization of p58 were further characterized. Binding of the purified p58 to thyroid hormones was examined. Analysis of binding data indicates that p58 binds to 3,3’,5-triiodo-~-thyronine (T,) with a Kd of 24.3 f 0.3 nM and n = 0.7 1. p58 binds to L-thyroxine similarly as to T3. However, D-T, and reverse-T3 bind to p58 with an affinity 4- and 20-fold less than that of T3, respectively. By use of the purified p58 as an immunogen, two hybridomas, J11 and 512, secreting monoclonal antibodies to p58 were isolated; both antibodies belong to the IgGIK subclass. 512 recognizes p58 from human, monkey, dog, hamster, and rat, but not mouse. J11 exhibits a similar species specificity except that it does not react with p58 from hamster. With these antibodies, p58 was found to be not posttranslationally modified by glycosylation, sulfation, or phosphorylation. It has a cellular degradation rate tllZ= 2.1 h. Immunocytochemical studies indicate that p58 is located in the nonmembranous cytoplasm (cytosol). These results are consistent with subcellular fractionation studies which show that >95% of 511 and 512 reactivity and T3 binding activity can be found in the llOOOOg supernatant.

N u m e r o u s studies have shown that the entry of thyroid hormones into cells is a carrier-mediated process (Krenning et al., 1978; Rao et al., 1981; Horiuchi et al., 1982a,b; Cheng, 1983; Pontecorvi & Robbins, 1986; Pontecorvi et al., 1987). This process can be blocked by various inhibitors of ATP production, cytoskeletal integrity, endocytosis, and Na+,K+-ATPase, such as oligomycin, antimycin, bacitracin, and monodansylcadaverine (Cheng, 1983; Pontecorvi & Robbins, 1986; Pontecorvi et al., 1987; Horiuchi et al., 1982a,b). This inhibition leads to a reduction not only in the cellular uptake but also in the accumulation of thyroid hormones in nuclei. Furthermore, antibodies against a membrane protein of rat hepatocytes were shown to inhibit the uptake of 3,3’,5-triiodo-~-thyronine (T3)l and L-thyroxine by rat hepatocytes (Mol et al., 1986). In addition, an energy-dependent transport system for translocation of T3 from cytoplasm to nucleus has also been demonstrated in hepatocytes and several other tissues (Mooradian et al., 1985; Oppenheimer & Schwartz, 1985). These studies all suggested that specific T3 binding proteins are involved in the transmembrane and transnuclear uptake of T,. However, up to the present time, the proteins responsible for such events have not been isolated and purified. We have recently purified a cellular T, binding protein from a human carcinoma cell line to apparent homogeneity (Kitagawa et al., 1987). This protein has an apparent molecular weight of 58 000 (p58) as determined by SDS-PAGE. The purified protein retains specific T3 binding activity. To understand its subcellular localization and further characterize its properties, we have developed monoclonal antibodies against p58. Two hybridomas secreting mAbs against p58 were iso-

* Address correspondenceto this author at Building 37, Room 4B09, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892. *National Institutes of Health. Food and Drug Administration.

lated. Using these mAbs, we found that p58 is not a posttranslationally modified protein and is exclusively localized in the cytoplasm. MATERIALS AND METHODS Materials. [1251]T3(2200 Ci/mmol), carrier-free [,*PIphosphoric acid, and [35S]sulfatewere purchased from Du Pont New England Nuclear. CHAPS, sodium salts of T,, D-T,, and L - T ~ phenylmethanesulfonyl , fluoride, aprotinin, and leupeptin were from Sigma. Endoglycosidase H and Nglycanase were purchased from Genzyme Corp. (Boston, MA). [3SS]Methionine(1200 Ci/mmol) was from Amersham; tunicamycin was from Calbiochem. Dulbecco’s modified Eagle’s medium was from Gibco; hypoxanthine/aminopterin/thymidine was obtained from Bethesda Research Laboratory. Heat-inactivated fetal bovine serum and Iscor’s (modified) Dulbecco’s medium were from Hazleton Dutchland, Inc. Affinity-purified rabbit and goat anti-mouse (H and L chain) antibodies were from Jackson Immunoresearch, Inc. Poly(ethylene glycol) ( M , 3000-3700) was from T. J. Baker Co. Cell Lines. A431, GH,, mouse N I H 3T3, and Chinese hamster ovary cells were propagated as described (Cheng, 1983). Human KB, HepG2, dog MDCK, and monkey Vero cells were obtained from American Type Culture Collection (Rockville, MD). MCF-7 cells were provided by Dr. R. Evans of the Salk Institute for Biological Studies. Purification of the T3 Binding Protein from A431 Cells. p58 was purified from A431 cells as described by Kitagawa et al. (1987) except for the following modification. After QAE ion-exchange column chromatography the flow-through (-90 I Abbreviations: T,, 3,3’,5-triiodo-~-thyronine; T,,L-thyroxine; PBS, phosphate-buffered saline; CHAPS, 3-[(3-~holamidopropyl)dimethylammoniol-1-propanesulfonate;SDS, sodium dodecyl sulfate; SDSPAGE, SDS-polyacrylamide gel electrophoresis; Tris, tris(hydroxymethy1)aminomethane;EDTA, ethylenediaminetetraacetic acid; mAbs, monoclonal antibodies.

This article not subject to U.S. Copyright. Published 1989 by the American Chemical Society

618

Biochemistry, Vol. 28, No. 2, 1989

mL) was concentrated to -20 mL by ultrafiltration (PM-30) and dialyzed against 1-2 L of buffer D (50 mM NaC1,20 mM phosphate, 0.5 mM CHAPS, and the protease inhibitors 1 mM EDTA, 1 pg/mL leupeptin, and 0.5 mM phenylmethanesulfonyl fluoride, pH 6.0) for 18-20 h with three changes at 4 "C. After dialysis, the dialyzate was applied to an SPSephadex C-50 column (1.2 X 8.8 cm) which was preequilibrated with buffer D. After being washed with 50 mL of buffer D, p58 was eluted with 50 mL of 100 mM phosphate, pH 6.5, at a flow rate of 50 mL/h. The T3 binding fractions were pooled (20 mL) and diluted with 3 volumes of buffer E (10 mM phosphate, 0.5 mM CHAPS, and the same protease inhibitors as in buffer D, pH 6.8). The pooled fractions were applied to a hydroxylapatite column (0.9 X 6 cm) which was preequilibrated with buffer E. The column was washed successively with 15 mL of 30 mM phosphate, 5 mL of 50 mM phosphate, 20 mL of 65 mM phosphate, and 20 mL of 70 mM phosphate. The T3 binding activity was eluted with 30 mL of 80 mM phosphate followed by 10 mL of 120 mM phosphate at a flow rate of 10 mL/h. The purity of p58 was examined by 10% SDS-PAGE. Compared with that of the method used before (Kitagawa et al., 1987), the yield in this simplified protocol is increased by 3-4-fold from 50 to 200 pg per 2.5 X lo9 cells. Furthermore, the specific activity of the purified protein is also increased by 2-fold from 37.6 f 12.5 to 90.1 f 1.5 fmol of T3 bound per pg of protein. Binding Assay. Binding assays were carried out by incubating the CHAPS extracts or purified protein with 0.2 nM [1251]T,for 30 rnin at 4 OC. Protein-bound [1251]T,was separated from unbound radioligand on a Sephadex G-25 (medium) column as described (Bolger & Jorgensen, 1980). Binding data were analyzed by least-squares analysis as described previously (Horiuchi et al., 1982a,b). Production of Monoclonal Antibodies. Mice were immunized with purified p58 according to Obata et al. (1988). The last boost ( 5 g of purified p58 in 0.5 mL of phosphatebuffered saline) was administered half by intravenous and half by interperitoneal injection 72 h before fusion. Splenocytes from the immune mice were fused with p3x 63Ag8653 myeloma cells by the methods described previously (Liang et al., 1985). The positive hybridomas were identified by (A) Elisa, (B) strip-comb dot immunobinding (Obata & Cheng, 1988), and (C) immunoprecipitation of the [35S]methionine-labeled lysate of A431 cells as described by Obata et al. (1988). The hybridomas were cloned by a limiting dilution method. Degradation Rate of p58 in A431 Cells. The degradation rate of p58 was evaluated by pulse-chase experiments as described previously (Hasumura et al., 1986). After the cells were pulsed with [35S]methioninefor 15 min, the medium was aspirated. The cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, and cells were harvested at intervals of 0, 1, 2, 4, 7, and 24 h. Immunoprecipitation of cellular extracts was carried out as previously described (Obata et al., 1988). The immunoprecipitated bands were quantified by densitometry. Posttranslational Modification of p58 in A431 Cells. ( I ) Tunicamycin Experiment. A43 1 cells (3 X 1O6 cells/ 100-mm dish) were preincubated with 10 pg/mL tunicamycin at 37 "C for 5 h. [%]Methionine (1 mCi/dish) was added and incubated for 3 h at 37 "C. At the end of the incubation, cells were washed and harvested with a rubber policeman. Preparation of cellular extracts and immunoprecipitation with J 1 1 were carried out as described in Hasumura et al. (1986). Inhibition in the incorporation of carbohydrate moieties into the receptor for epidermal growth factor was used as a positive

Obata et al. control. Under these conditions and with Ab 2913 for immunoprecipitation (Beguinot et al., 1986), receptor for epidermal growth factor migrated as bands with molecular weights of 133000 and 75 000 in contrast to the intact molecule with a molecular weight of 170000 (Mayes & Waterfield, 1984). ( I I ) Digestion of the Purified p58 with N-Glycanase and Endoglycosidase H . The digestion of the purified p58 (5-10 pg) by either enzyme was carried as described in Hasumura et al. (1986). Ovalbumin (1 pg) was used as a positive control. At the end of digestion, the enzymatic mixture was analyzed by a 10% SDS-PAGE followed by Coomassie blue staining. (111)Phosphorylation and Sulfation Experiments. A43 1 cells (3 X lo6 cells/60-mm dish) were incubated with [32P]phosphoric acid (2.5 mCi/dish) in phosphate-free medium or with [ 3 5 ] ~ ~ l f a(1t emCi/dish) in sulfate-free medium for 2, 4,6, or 24 h. The medium was removed, and cells were washed with PBS. CHAPS extracts were prepared and immunoprecipitated as described above. Localization of p58 by Subcellular Fractionation. Overnight cultures of A43 1 cells (1 X lo7 cells/ 150-mm dish) were washed with 30 mL of PBS (Ca*+and Mgz+free). Cells were harvested with a rubber policeman and pelleted. The cell pellet was resuspended in 2 mL of homogenization buffer (0.25 M sucrose, 20 mM Tris, 1.1 mM MgCl,, pH 7.85) and homogenized in a tight Dounce homogenizer. After centrifugation at 800g for 10 min, the supernatant was further centrifuged at 1lOOOOg for 90 min at 4 "C (supernatant A). The pellet was resuspended by pipetting in 1 mL of the homogenizing buffer. The suspension as centrifuged at 1lOOOOg for 90 rnin (supernatant B). The pellet was extracted with 1 mL of 3 mM CHAPS by stirring for 30 min. The suspension was centrifuged at 1lOOOOg for 90 min (supernatant c ) . T3binding was carried out by using 50 pg of protein from supernatant A, B, and C. In a separate experiment, A431 cells ( 5 X lo6 cells/lOO-mm dish) were labeled with [35S]methi~nine for 20 h at 37 "C. Similar cell fractionation was carried out. A total of 25 pg of protein from supernatant A, B, and C was immunoprecipitated with J1 1. Immunocytochemistry. A431 cells, as well as other cell types, in 35-mm plastic dishes were fixed with 3.7% formaldehyde for 10 min at room temperature, washed in PBS, and then incubated with J l l or 512 at 10 pg/mL in a diluent composed of 4 mg/mL normal goat globulin, 0.1% saponin, and PBS (NGG-sapPBS) as previously described (Willingham & Pastan, 1985). The cells were incubated in this step for 30 min at 23 "C, washed in PBS, and then incubated with rhodamine-labeled affinity-purified goat anti-mouse IgG (Jackson ImmunoResearch) (50 pg/mL in NGG-sapPBS for 30 min, 23 "C). The cells were washed and mounted under a coverslip in buffered glycerol. Other experiments were performed by permeabilization with 0.1% Triton X-100 for 5 rnin at 23 "C or by 80% acetone treatment for 5 rnin after the formaldehyde fixation step. For electron microscopic immunocytochemistry, A43 1 cells were fixed in plastic dishes according to the EGS procedure with 0.17% glutaraldehyde and labeled according to the ferritin bridge method as previously described (Willingham, 1980). The steps of the ferritin bridge sequence included (1) mouse monoclonal antibodies (J1 1 or nonreactive control mouse monoclonal; 10 pg/mL), (2) goat anti-mouse IgG (1:20 diluted whole IgG), (3) affinity-purified mouse anti-ferritin (Jackson ImmunoResearch; 50 pg/mL), and (4)horse spleen ferritin (200 pg/mL). The cells were then postfixed in glutaraldehyde

Biochemistry. Vol. 28. No. 2, 1989 619

Cytoplasmic T, Binding Protein

Table I Reactivitv of Monoclonal Antibodies JI I and J12 with Cultured Cells of Different Soecies human hamster

antibody JI I

assay A431 KB HepG2 MCF-7 monkey Vero CHO ratGH, mouse3T3 dog MDCK immunofluorescence + + + + + + + immunoprecipitation' 14 44 5 20 37 0 I 0 ND' + JI2 immunofluorescence + + + + + + + immunoprecipitationa 22 I87 15 46 80 6 0.5 0 N D ~ .The immunoprecipitationwas carried out as described under Materials and Methods. An equal amount of protein (25 pg) was used in immunoprecipitation. The intensity of the radioactive p58 bands was determined by densitometry. The values are relative intensity with p58 from GH, immunoprecipitated by J I I as reference. bND = not determined.

A. Immunoprecipitation

z1

$

r

-

1; \I

J11

KDa,

1 2

3 4

J12

5 6

7 8 9 10 11 12 13 14 15 16

'.

,

-..--

b

-.-_-_ --__ 0

8

7

6

5

31

Ligand I-Log,,MI

1: Thyroid hormone binding to purified p58. Purified p58 (0.4 fig) was incubated with 0.2 nM ["sI]Tl in the absence or in the presence of increasing concentrations of unlabeld TI (0). T4 (O),DT, (A), or reverse-T (A)for O S h at 4 "C. At the end of incubation, FIGURE

B. Strip-comb Dot lmmunobinding

free and bound [IbI]T3were separated as described under Materials and Methods. (Inset) The data from TI binding were replotted as bound vs total free T,. The line is the theoretical curve calculated from the equation described in Horiuchi et al. (1982a,b).

and osmium and embedded in situ (Willingham, 1980). Quantitation was performed by measuring demarcated areas of extramembranous cytoplasm in random prints of the same magnification with the MACMEASURE program (Hook & Rasband, 1987) on a MacPlus T M computer. The numbers of ferritin cores in these areas were counted, and by assuming a section thickness of 1000 A, the levels of p58 localization were calculated as numbers of ferritin cores per cubic micrometer of extramembranous cytoplasm (cytosol).

RFSULTS Binding of T, and Its Analogues lo Purified p58. Previously, the binding characteristics of p58 in total cellular extracts were examined, and p58 was found to bind T, and T, similarly. To evaluate whether the binding characteristics have been altered by the chromatographic steps during purification, the binding of T, and its analogues to the purified p58 was examined. Binding of ["SI]T, to p58 was rapid and reached equilibrium after 30 min at 4 OC (data not shown). Figure I shows the binding of ['2sI]T1 to p58 with increasing concentrations of unlabeled T,, T,, D-T,, and reverse-T, under equilibrium conditions. T, and T, hind to p58 with similar affinity. However, D-T, and reverse-T, bind to p58 with an affinity 4- and 20-fold less than that of T,, respectively. The inset in Figure 1 shows a plot of TI bound vs increasing free T, concentration. The line is the theoretical curve calculated from the equation described in Horiuchi et al. (1982a.b). The data fit the calculated curves to give one class of binding sites with a Kd = 24.3 f 0.3 nM and n = 0.71 (moles of T, bound per mole of p58). Earlier, the Kd for the binding of T, to p58 in the crude cellular extracts was found to be 17 f 3 nM. These results indicate that the binding

Jll 1

2

J12

3

4

5

6

7

8

.. (A) Autoradiogram of immunoprecipitates from [lsS]methioninolabeled CHAPS extracts. Cellular extracts from cultured cells of diffmnt species [(l-2) X I@ cells/6omm dish] were extracted with 3 mM CHAPS, and 25 fig of cellular extracts was immunoprecipitated with JII (lanes 1-8) or 512 (lanes 9-16). Lanes 1 and 9, A431; lanes 2 and IO, MCF-'I; lanes 3 and I I, HepGZ; lanes 4 and 12, K B lanes 5 and 13, Vero; lanes 6 and 14. CHO lanes 7 and IS, CHI; lanes 8 and 16, 3T3 cells. (B)Stripcomb dot immunobinding of p58 to JI 1 or 112. Purified pS8 (20 ng) was spotted onto the tips of strip combs. After being dried, the combs were incubated with SO fiL of 111 (lanes I and 2) or 112 (lanes 5 and 6) supernatant or negative control medium (lanes 3,4,7, and 8). The washing, blacking. and color developing were carried out as described (Obata & Cheng, 1988). FIGURE 2

activity of p58 is stable under the conditions used for puri-

fication. Characterization of Monoclonal Antibodies against p58. Purified p58 which retains T, binding activity was used as an immunogen to produce antibodies. By use of a special immunization protocol (Obata et al., 1988). two hybridomas, J11 and JI2, were found to secrete antibodies that immunoprecipitated a 58K protein in A431 cells (lanes 1 and 9 in Figure 2A). Furthermore, as shown in Figure 2B, both antibodies reacted with purified p58 as shown by stripcomb dot immunobinding (Obata & Cheng, 1988). To determine the species specificity of J11 and J12 and to compare the abundance of p58 in different culture cell lines, p58 was extracted and immunoprecipitated by both antibodies. As shown in Figure 2A and Table I, both antibodies recognize p58 from human, monkey, dog, and rat. None of the two

620 Biochemistry. Vol. 28. No. 2. I989 antibodies recognizes p58 from mouse. 512 recognizes p58 from hamster, but JI I does not. Furthermore, comparison of the relative intensity of the bands showed that KB cells, a carcinoma cell line, have the highest abundance of p58. Interestingly, 512 has a higher affinity for human p58 but is less reactive to rat p58 than 512. These results suggest that JI 1 and 312 are not the same mAb and do not recognize the same epitopes on p58. Immunodepletion experiments were also carried out to see whether the T, binding activity of purified p58 can be recognized by JI 1 and 512. p58 was first reacted with Jll or 512. The immunocomplex was precipitated by Staph A. The binding activity in the supernatant was examined. When HBZI, a monoclonal antibody against the human transferrin receptor was used, - I X IO5dpm of [1251]T,binding activity was detected in the supernatant. This binding activity was specific, since in the presence of 20 pM unlabeled T, only 5% of the activity remained. However, Jll or 512 removed 9&95% of this T,binding activity from the supernatant (data not shown). These results indicate that JI 1 and 312 recognized the component responsible for the observed T, binding activity. Thus, it is clear that the T, binding activity resides in p58 and the binding activity does not derive from a trace contaminant present in the p58 preparation. To determine the class and subclass of the antibodies, the clones were labeled with ['5S]methionine; the culture supernatants were harvested and immunoprecipitated with affinity-purified rabbit anti-mouse IgG subclass specific antibodies. Both Jll and 512 were found to be IgG,K antibodies. Immunocytochemical Localization of p58. To identify the subcellular location of p58, cultured cells were fixed and permeabilized for immunofluorescence. Following formaldehyde fixation and permeabilization with either saponin, acetone, or Triton X-100,the cells showed a bright pattern of fluorescence confined to the extramembranous cytoplasm as shown in Figure 38. Both J11 and 512 showed identical patterns in A431 cells. No localization was seen in the interior of the nucleus under any of these fixation permeabilization conditions. Typical of such a cytosolic pattern, surface ruffles showed bright localization, and larger cytoplasmic organelles could be seen as dark areas of exclusion in the cytoplasm. The species specificity of mAbs JI I and 512 was also evaluated by immunofluorescence. As shown in Table I, the results are consistent with the findings by immunoprecipitation. The localization of p58 was investigated further by electron microscopic immunocytochemistry in A431 cells. As shown in Figure 3C, the distribution of p58 seen with this approach agreed with the results seen with immunofluorescence, in that all of the localization found was in the extramembranous cytoplasm, with no particular concentration in any one part of the cell. No localization was seen on the external plasma membrane surface or in the nucleus. The cytosolic localization of p58 was then quantitated morphologically by counting ferritin cores per cubic micrometer. From these measurements, the minimum number of molecules of p58 present per cell can be estimated (>300000 molecules per cell). This estimation is consistent with the biochemical estimation in which A431 cells contained (1-2) X lo6 molecules of pS8/cell. Localization of p58 by Subcellular Fraction. To identify the subcellular localization of p58 biochemically, A431 cells were homogenized under isotonic conditions. After removal of unbroken cells and nuclei, the 1 IOOOOg supernatant was prepared. After the pellet was washed once with the homogenization buffer, the pellet was further extracted with 3 mM CHAPS. The results show that 9 4 9 7 % of T, binding activity

Obata et al.

.

'

'

.... . . .- /-.>/7,-

'

"

,-./-. -,-3., ,

,,.,;*:,:,--...

6

.,,

. .:.,

" .l

-* !.

N, .

_..I

-

Immunocytochemical localization of p58 using monoclonal antibody J l l . A431 cells in culture dishes were fixed in situ in formaldehyde (A and B) or acwrding to the EGS fixation protocol ( C ) as described under Materials and Methods. The cells were subsequently incubated with monoclonal antibody JI I (B and C)or a control nonreactive monoclonal antibody ( A ) in the presence of saponin. The cells were then further incubated in a rhodamineanti-mouse IgG wnjugate (A and B) or with the steps of the ferritin bridge method (C). The immunofluorescence pattern shown in (B) is typical of a diffuse distribution in the nonmembranous cytoplasm, although increased levels at the margins of membranous elements cannot be completely led out. Note the lack of labeling in the nucleus (n), the dark areas representing lack of labeling in large cytoplasmic organelles (B, arrows), and the labeling of ruffles at the cell margin (B, white arrowheads). By electron microscopy (C), the ferritin labeling (arrowheads show examples) shows large amounts of pS8 in the cytosolic compartment, with no label detected on the lumenal side of the endoplasmic reticulum (er), in mitochondria (m), or in the nucleus (N). (A and B) Magnification 300% bar = IO pm. (C) Magnification 36000X; bar = 0.1 pm. FIGURE 3:

and JI 1 immunoreactivity are present in the 1 IOOOOg supernatant. These findings are consistent with the findings by morphological methods that p58 is located in cytosol. The 3-6% T, binding activity that remained with the llOOOOg pellet may be due to p58 trapped in vesicles during homogenization. In these and earlier preliminary studies, it was found that the T, binding activity is 2-fold more stable in the presence of 0.5-3 mM CHAPS (data not shown). Therefore, all binding studies and the chromatographic steps were carried out in the presence of CHAPS. Posttranslational Modification and Turnover of p58. To evaluate whether p58 is posttranslationally modified by glymylation, three different experiments were carried out. A431 cells were pulsed with ['5S]methionine for 15 min followed by a chase with unlabeled methionine for 1, 2, 4, 7, and 24 h. Figure 4 shows the autoradiogram of a SDS-PAGE of the immunoprecipitates with JI 1. No specific radioactive bands with molecular weight higher than 58K could be detected at all time points examined. These results indicated that, within the detection sensitivity of the size change in a 10% gel, p58 is not processed. The p58 bands were quantified by densitometry, and the decay curve was plotted. The data fit into a model in which there is a fast decaying component and a stable

Cytoplasmic T, Binding Protein

Biochemistry, Vol. 28, No. 2, 1989 621

"t €6

I

L lime lhouml 0

1 2 4 7 24

n c w 4 Dcgradationrateofp58. A4!1 ylls (6.8 X l@cells/60mm dish) were incubated with [I S]methlonine (0.95mCi/mL) for 15 min. The dishes were cooled to 4 OC was washed with 1 mL of 4 mM methionine. The cells were incubated in serumcontaining medium for 0 (lane I), I (lane 2), 2 (lane 3). 4 (lane 4), 7 (lane 5). and 24 (lane 6) h. p58 was extracted and immunoprecipitated with JI I as described under Materials and Methods. The radioactive bands at various time points were quantified by densitometry, and the decay curve was plotted by computer fitting using the equation y = A + Be-k', where y is the total activity at time 1, A is the activity at 24 h, B is the activity at time 1, k is the rate constant. and ll/z = 0.693Jk.

pool. As shown in the inset of Figure 4, the fast decaying pool was degraded with a r l j 2 of 2.11 1.08 h. Medium was collected and assayed for its content of p58 by immunoprecipitation. No detectable p58 was found; thus, p58 was not secreted from the cell. Two additional experiments were also carried out to confirm that p58 is not posttranslationally modified. A431 cells were pretreated with tunicamycin for 5 h and labeled with [%Imethionine for 3 h. Cellular extracts were prepared and immunoprecipitated with J l l . No difference in the molecular weight of immunoprecipitable bands was observed whether cells were treated with tunicamycin or not (data not shown). These results indicate that p58 is mast likely not glycosylated. In addition, purified p58 was treated with N-glycanase or endoglycosidase H. In the same experiment, ovalbumin and transthyretin were treated similarly. The sugar residues were removed from ovalbumin by either enzyme to give a protein band with a molecular weight of -38K. while transthyretin was not affected. Under these conditions, no detection of lower molecular weight species was seen for p58. The results from these three experiments indicate that p58 is not likely to be a glycosylated protein. To determine whether p58 is posttranslationally modified by phosphorylation or sulfation, A431 cells were labeled with [J2P]phosphoricacid or [JSS]sulfuricacid for I , 2.4, and 16 h. Cell lysates were immunoprecipitated with J I 1. No incorporation of ,*P or % was seen in p58. These results indicate that p58 probably is not a phosphorylated or sulfated protein.

*

DISCUSSION By the use of monoclonal antibodies, the present studies demonstrate that the previously isolated cellular thyroid hormone binding protein (Kitagawa et al., 1987) is located in the cytosol. With crude cytosol preparations, a cytosolic binding protein for thyroid hormone has been reported to be present in liver (Yoshida et al., 1979; Davis et al., 1974;

Hamada et al., 197% Dillman et al., 1974; Defer et al., 1975). kidney (Davis et al., 1974; Dillman et al., 1974). red cells (Yoshida & Davis, 1977; Michelot et al., 1979). brain (Geel, 1977; Dozin-Van Roye & De Nayer, 1978; Gee1 et al., 1981; Lennon et al., 1980). and GH, cells (Samuels et al., 1974). The reported biochemical properties, however, vary greatly. " (Yoshida et al., 1979) The molecular weights range from 6 to >IO0000 (Hamada et al., 1970); the Kd's of T, and T4 binding vary from 6.6 X lo4 to 4 X lo-" M (Davis et al., 1974; Hamada et al., 1970 Dillman et al., 1974; Michelot et al., 1979; Dozin-Van Roye & De Nayer, 1978; Gee1 et al., 1981; Lennon et al., 1983, 1980; Samules et al., 1974) and 1.2 X to 2.9 X lWl0 M (Davis et al., 1974; Dillman et al., 1974; Michelot et al., 1979; Lennon et al., 1980; Samuels et al., 1974), respectively. Furthermore, one and two classes of binding sites for thyroid hormone were also reported (Davis et al., 1974; Dillman et al., 1974; Michelot et al., 1979; Dozin-Van Roye & De Nayer, 1978; Lennon et al., 1983, 1980; Samuels et al., 1974). In the present study, using a purified preparation, one class of binding sites with a Kd of 24.3 0.3 nM was found. T4 binds to p58 with an affinity similar to that of T,. These binding characteristics were similar to those found in our earlier studies in which a crude preparation was used (Kitagawa et al., 1987). It is unclear whether these variations reflect the difference in the cytosolic preparations under different experimental conditions or the differences in the entity of protein molecules from different species and tissues. Recently, it has been found that there is more than one single T, nuclear receptor. T, nuclear receptor exists in different forms in different tissues of the same species (Weinberger et al., 1986; Sap et al., 1986; Thompson et al., 1987; Benbrook & Pfahl, 1987). A question frequently raised is the structural relationship between the cytosolic binding protein and nuclear receptors for T,. Earlier, by use of binding characteristics (Dillman et al., 1974; Defer et al., 1975; Samuels et al., 1974) and migration rates on sucrose gradients and polyacrylamide gels (Defer et al., 1975). cytosolic T, binding protein has been suggested to be different from nuclear receptors. With the availability of the purified p58 and its mAbs, the present study provided direct evidence to show that they are structurally unrelated: (1) Immunofluorescence and morphological studies indicate that p58 is exclusively present in cytosol and not in nuclei. (2) T, binding activity of p58 can be absorbed by J1 I and 312, whereas the T, binding activity of nuclear extracts of A431 cells cannot be removed by J l l and J12 (data not shown). (3) It has been reported that c-erb-A protein is the T, nuclear receptor (Weinberger et al., 1986; Sap et al., 1986; Thompson et al., 1987). When the ["S]methionine-labeled in vitro translation products of human placental c-erb-A (Weinberger et al., 1986) were incubated with J1 I or 512, no immunoprecipitable bands were detected (data not shown). (4) Comparison of the peptide maps of V8 protease digestion between p58 and c-erb-A translation products showed no common fragments (data not shown). These results indicate that p58 is distinct from T, nuclear receptors. In this regard, thyroid hormone may be more similar to retinoic acid than to the steroid hormones. Recently, cDNA sequences for the retinoic acid nuclear receptor have been reported (Giguere et al., 1987; Petkovich et al., 1987). Comparison of the deduced sequences indicates that the retinoic acid nuclear receptor and cellular retinoic acid binding protein are structurally unrelated (Shubeita et al., 1987; Wei et al., 1987). The functional roles of the cytosolic thyroid hormone binding protein remain to be established. The cytosolic binding

*

622

Biochemistry, Vol. 28, No. 2, 1989

protein has been reported to be present in many tissues of several species. The present study also showed that it is present in many cultured cells of various species. Its role as an intracellular regulator (Hamada et al., 1970), buffer (Yoshida et al., 1979), or a transporter (Spaulding & Davis, 1971) for thyroid hormones has been proposed. This last proposal was supported by the studies reported by Mooradian et al. in which an energy-dependent and stereospecific transporter system for translocation of T3 from cytoplasm to nucleus was demonstrated in hepatocytes (Mooradian et al., 1985). Therefore, p58 could be the intracellular transport protein or part of the transport system. However, p58 could also have additional functions. Thyroid hormones were found to increase erythrocyte production in animals (Donati et al., 1966) and cause erythroid proliferation (Popviz et al., 1977). Cytosolic T3 binding protein was shown to be present in erythrocytes (Defer et al., 1975; Yoshida & Davis, 1977; Michelot et al., 1979). In addition, the number of cytosolic T4 binding sites in rat brain changes during development. A decrease of T3 binding sites in the cortex from day 3 to day 35 and near disappearance of binding sites in cerebellum of adult brain were found (Lennon et al., 1980). The low number of cytosolic thyroid hormone binding sites in adult brain has been implicated in the unresponsiveness to thyroid hormones as measured by oxygen consumption (Fazekas et al., 1951; Sokoloff et al., 1953). Therefore, the functions of cytosolic T, binding protein may not be limited to the roles mentioned above. With the availability of the monoclonal antibodies to p58, it has become possible to investigate some of these questions. ACKNOWLEDGMENTS We thank Dr. P. McPhie for computer fitting of the T3 binding and p58 degradation data, Dr. Jay S. Epstein for help in culturing hybridomas, and Dr. Hirokazu Kat0 for help in screening hybridomas. Registry No. T,,6893-02-3; T4,51-48-9; D-T,, 5714-08-9; reverse-T,, 5817-39-0.

REFERENCES Beguinot, L., Werth, D., Ito, S., Richert, N., Willingham, M. C., & Pastan, I. (1986) J. Biol. Chem. 261, 1801-1807. Benbrook, D., & Pfahl, M. (1987) Science (Washington, D.C.) 238, 788-791. Bolger, M. B., & Jorgensen, E. C. (1980) J. Biol. Chem. 255, 10271-10278. Cheng, S.-Y. (1 983) Endocrinology (Baltimore) 11 2, 1754-1762. Davis, P. J., Handwerger, B. S., & Glaser, F. (1974) J . Biol. Chem. 249, 6208-6217. Defer, N., Dastugne, B., Sabatier, M. M., Thomopoulos, P., & Kruh, J. (1975) Biochem. Biophys. Res. Commun. 67, 995-1004. Dillman, W., Surks, M. Z., & Oppenheimer, J. H. (1974) Endocrinology (Philadelphia) 95, 492-498. Donati, R. M., Warnecke, M. A., & Gallagher, N. Z. (1966) Proc. SOC.Exp. Biol. Med., 1199-1201. Dozin-Van Roye, B., & De Nayer, Ph. (1978) FEBS Lett. 96, 152-1 54. Fazekas, J. F., Grases, F. B., & Altman, R. W. (1951) Endocrinology (Philadelphia) 48, 169-174. Geel, S. E. (1977) Nature (London) 269, 428-430. Geel, S. E., Gonzales, L., & Timiras, P. S. (1981) Endocr. Res. Commun. 8, 1-18. Giguere, V., Ong, E. S., Sequi, P., & Evans, R. M. (1987) Nature (London) 330, 624-629. Hamada, S., Orizuka, K., Miyake, T., & Fukase, M. (1970) Biochim. Biophys. Acta, 479-492.

Obata et al. Hasumura, S., Kitagawa, S., Lovelace, E., Willingham, M. C., Pastan, I., & Cheng, S.-Y. (1986) Biochemistry 25, 7881-7888. Hook, G. R., & Rasband, W. (1987) Proc.-Annu. Meet., Electron Microsc. SOC.Am., 920. Horiuchi, R., Cheng, S.-y., Willingham, M. C., & Pastan, I. (1982a) J . Biol. Chem. 257, 3 139-3 144. Horiuchi, R., Johnson, M. L., Willingham, M. C., Pastan, I., & Cheng, S.-y. (1982b) Proc. Natl. Acad. Sci. U.S.A.79, 5527-553 1. Kitagawa, S., Obata, T., Hasumura, S., Pastan, I., & Cheng, S.-y. (1987) J . Biol. Chem. 262, 3903-3908. Krenning, E. P., Doctor, R., Bernard, H. F., Visser, T. J., & Hennemann, G. (1978) FEBS Lett. 91, 113-1 16. Krenning, E. P., Doctor, R., Bernard, H. F., Visser, T. J., & Hennemann, G. (1981) Biochim. Biophys. Acta 676, 3 14-3 19. Lennon, A. M., Osty, J., & Nunez, J. (1980) Mol. Cell. Endocrinol. 18, 201-214. Lennon, A. M., Chantoux, F., Osty, J., & Francon, J. (1983) Biochem. Biophys. Res. Commun. 15, 901-908. Liang, C.-M., Herren, S., Sand, A,, & Jost, J. (1985) Biochem. Biophys. Res. Commun. 128, 171-178. Mayes, E. L., & Waterfield, M. D. (1984) EMBO J. 3, 53 1-5 37. Michelot, J., Dastugue, B., Defer, N., & Meyniel, G. (1979) Biochem. Biophys. Res. Commun. 88, 1368-1374. Mol, J. A., Cotro, R., Rozing, J., & Hennemann, G. (1986) J . Biol. Chem. 261, 7640-7643. Mooradian, A. D., Schwartz, H. L., Marish, C. N., & Oppenheimer, J. H. (1985) Endocrinology (Baltimore) 11 7, 2449-2456. Obata, T., & Cheng, S.-Y. (1988) BioTechniques 6,299-302. Obata, T., Fukuda, T., & Cheng, S.-Y. (1988) FEBS Lett. 230, 9-12. Oppenheimer, J. H., & Schwartz, H. L. (1985) J. Clin. Znuest. 75, 147-156. Petkovich, M., Brand, N. J., Krust, A., & Chabon, P. (1987) Nature (London) 330, 445-450. Pontecorvi, A., & Robbins, J. (1986) Endocrinology (Baltimore) 119, 2755-2761. Pontecorvi, A,, Lakshmanan, M., & Robbins, J. (1987) Endocrinology (Baltimore) 121, 2145-2152. Popviz, W. J., Brown, J. G., & Adamson, J. W. (1977) J. Clin. Invest. 6, 907-9 13. Rao, G. S., Rao, M. L., Thillmann, A,, & Quednau, H. D. (1 98 1) Biochem. J. 198, 457-466. Samuels, H. H., Tsai, J. S., Casanova, J., & Stanley, F. (1974) J . Clin. Invest. 54, 853-865. Sap, J., Munoz, A., Damm, K., Goldberg, Y., Ghysdail, J., Leutz, A., Beug, H., & Vennstrom, B. (1986) Nature (London) 324, 635-640. Shubeita, H. E., Sambrook, F. E., & McCormick, A. M. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 5645-5649. Sokoloff, L., Wechesler, R. L., Mangold, R., Balls, K., & Hety, B. S. (1953) J . Clin. Invest. 32, 202-208. Spaulding, S. W., & Davis, P. J. (1971) Biochim. Biophys. Acta 229, 279-283. Thompson, C. C., Weinberger, C., Lebo, R., & Evans, R. (1987) Science (Washington, D.C.) 237, 1610-1613. Wei, L.-N., Mertz, J. R., Goodman, D. S., & Nguyen-Huu, N. C. (1987) Mol. Endocrinol. 1 , 526-534. Weinberger, C., Thompson, C. C., Ong, E. S., Lebo, R., Grual, D. J., & Evans, R. M. (1986) Nature (London) 324, 641-646.

Biochemistry 1989, 28, 623-628

Willingham, M. C. (1980) Histochem. J . 12, 419-434. Willingham, M. C., & Pastan, I. (1985) in An Atlas of Immunofluorescence in Cultured Cells, pp 1-13, Academic, Orlando, FL.

623

Yoshida, K., & Davis, P. J. (1977) Biochem. Biophys. Res. Commun. 78, 697-705. Yoshida, K., Davis, P. J., & Schoenl, M. (1979) Biochim. Biophys. Acta 582, 332-345.

Kinetics and in Vitro Origin of the Temperature-Dependent Transition of the Estrogen Receptor Monomer? Jeffrey C. Hansent and Jack Gorski* Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, 420 Henry Mall, Madison, Wisconsin 53706 Received March 11, 1988; Revised Manuscript Received August 17, 1988

ABSTRACT: Partitioning of estrogen receptors in aqueous two-phase polymer systems has provided the basis

for a detailed kinetic analysis of the effects of temperature on estrogen receptor (ER) structure in vitro. Exposure to temperatures of 0-30 OC increased the rate of change in E R partition coefficients by up to 100-fold but did not affect the final extent of the process. The temperature-dependent change in E R partition coefficients was characterized by a linear Arrhenius plot and an activation energy of 25 kcal/mol. The rate of the temperature-dependent ER transition (28 “ C ) was found to be unaffected by greater than 50-fold changes in receptor concentration, which indicates that the temperature-dependent change in partition coefficients reflects a first-order process. The partition coefficients of heated E R were unaffected by subsequent 18-h incubations at 0 OC,indicating that the temperature-dependent E R transition is irreversible in vitro. Direct heating of the unoccupied E R resulted in both a change in E R partition coefficients and a loss of ER binding sites. The temperature-dependent change in unoccupied E R partition coefficients was complete within 30 min at 28 OC and yielded a first-order rate constant that was the same as that obtained for heating the receptor-estradiol complex at 28 OC. In contrast, the loss of unoccupied E R binding sites that occurred during 28 OC incubations did not reach completion after 150 min of heating and was found to behave as a second-order process. Thus, the temperature-dependent change in unoccupied receptor structure is a separate process from the temperature-dependent loss of unoccupied E R binding sites. Taken together, these results indicate that the temperature-dependent change in E R partition coefficients represents an irreversible, hormone-independent change in the conformation of the ER monomer that originates entirely as a consequence of tissue homogenization and subsequent in vitro extraction of the receptor into the cytosol.

I t is becoming increasingly clear that the estrogen receptor monomer1 is a dynamic protein composed of multiple functional domains (Green et al., 1986; Krust et al., 1986; Kumar et al., 1986). In general, all steroid hormone receptors share the same organization of these functional domains, in which the hormone binding site is located in the hydrophobic carboxy-terminal region of the receptor, the nuclear association domains in the middle of the primary sequence, and a transcription “modulating” function near the amino terminus of the protein (Carlstedt-Duke et al., 1983; Giguere et al., 1986; Green & Chambon, 1986; Hollenberg et al., 1987). Recent mutational analyses of estrogen receptor (ER2 and glucocorticoid receptor cDNAs indicate that hormone-dependent induction of gene transcription in vivo at least in part involves conformational rearrangement of steroid receptor monomers upon hormone binding (Green & Chambon, 1987; Godowski et al., 1987; Hollenberg et al., 1987). In vitro, it has been demonstrated that the structure of the ER monomer is influenced by both hormone binding (Hansen & Gorski, 1985, 1986) and exposure to elevated temperature (Bailly et al., 1980; Sakai & Gorski, 1984; Muller et al., 1985; Hansen & ‘This work was supported by National Institutes of Health Grant HD 08192 and by the National Foundation for Cancer Research. J.C.H. was a recipient of the Evelyn Steenbock predoctoral fellowship. * T o whom correspondence should be addressed. *Present address: Department of Biochemistry and Biophysics, Oregon State University, Corvallis, OR 97331.

0006-2960/89/0428-0623$01.50/0

Gorski, 1986; Redeuilh et al., 1987). Hormone binding has also recently been demonstrated to alter the structure of progesterone (Moudgil & Hurd, 1987) and glucocorticoid (Moudgil et al., 1987) receptors in vitro, suggesting that a hormone-induced conformational change is a general property of all steriod receptors. Although temperature-dependent changes in steroid receptor structure have been routinely and reproducibly observed with virtually every experimental approach [see Gorski and Gannon (1976), Grody et al. (1982), Sherman and Stevens (1 984), and Andreasen (1 987) for reviews], the temperature-dependent transition of the ER monomer occurs in vitro in the absence of bound hormone (Hansen & Gorski, 1986), and thus it is not clear what biological role can be attributed to this hormone-independent phenomenon. Perhaps least understood is the origin of the temperature-dependent ER transition in vitro; Le., why does temperature influence the structure of a protein recently isolated from the 37 “C in vivo environment? In an attempt to better understand the nature and role of the in vitro effects of temperature on the structure of the ER I The estrogen receptor “monomer” is defined as the M , 65 000 steroid binding protein encoded by the estrogen receptor gene. The ER monomer has been referred to elsewhere as the “ER steroid binding subunit” and the “4sestrogen receptor”. Abbreviations: ER, estrogen receptor(s); KoW. partition coefficient; PEG, poly(ethy1ene glycol); TED, 10 mM Tris-HCI, 1.5 mM EDTA, and 0.5 mM dithiothreitol.

0 1989 American Chemical Society