Progesterone receptor synthesis and degradation in MCF-7 human breast cancer cells as studied by dense amino acid incorporation. Evidence for a non-hormone binding receptor precursor.

We have used the technique of density labeling of proteins by biosynthetic incorporation of 2H, 13C, 15N (dense) amino acids to study the synthesis and degradation rates of the progesterone receptor in MCF-7 human breast cancer cells. In cells grown in the absence of progestin, sucrose gradient shift analyses reveal that it takes 17 h for the normal density progesterone receptor levels to be reduced to half the initial value, whereas in the presence of 10 nM of the synthetic progestin [3H]R5020, the receptor turns over more rapidly, such that the normal density R5020-occupied progesterone receptor complexes are reduced to half in 12 h. The accelerated progesterone receptor turnover in the presence of [3H]R5020 reflects increased turnover rates of both the A (Mr-85,000) and B (Mr-115,000) subunits, as determined by sodium dodecyl sulfate gel analyses of dense and light receptors photoaffinity labeled with [3H]R5020. In both control and progestin-exposed cells, the time course of progesterone receptor turnover shows a lag of approximately 6 h after dense (15N, 13C, 2H) amino acid exposure, before dense hormone binding receptor species are seen and before normal density progestin binding activity starts decreasing. Since our evaluations of progesterone receptor depend upon its binding of radiolabeled ligand ([3H]R5020), this lag in the density shift kinetics would be consistent with the presence of a non-hormone binding biosynthetic precursor, from which the hormone-binding form of progesterone receptor is derived. A kinetic model is used to analyze the lag-decay profiles and to determine the rate constants for progesterone receptor synthesis, activation to the hormone-binding form, and degradation.

We have used the technique of density labeling of proteins by biosynthetic incorporation of 2H, 13C, lSN (dense) amino acids to study the synthesis and degradation rates of the progesterone receptor in MCF-7 human breast cancer cells. In cells grown in the absence of progestin, sucrose gradient shift analyses reveal that it takes 17 h for the normal density progesterone receptor levels to be reduced to half the initial value, whereas in the presence of 10 nM of the synthetic progestin [3H]R5020, the receptor turns over more rapidly, such that the normal density R5OZO-occupied progesterone receptor complexes are reduced to half in 12 h. The accelerated progesterone receptor turnover in the presence of [3H]R5020 reflects increased turnover rates of both the A (Mr-85,000) and B (Mr- 115,000) subunits, as determined by sodium dodecyl sulfate gel analyses of dense and light receptors photoaffinity labeled with [3H]R5020.
In both control and progestin-exposed cells, the time course of progesterone receptor turnover shows a lag of approximately 6 h after dense ("N, 13C, 2H) amino acid exposure, before dense hormone binding receptor species are seen and before normal density progestin binding activity starts decreasing. Since our evaluations of progesterone receptor depend upon its binding of radiolabeled ligand ([3H]R5020), this lag in the density shift kinetics would be consistent with the presence of a non-hormone binding biosynthetic precursor, from which the hormone-binding form of progesterone receptor is derived. A kinetic model is used to analyze the lag-decay profiles and to determine the rate constants for progesterone receptor synthesis, activation to the hormone-binding form, and degradation.
Progestins mediate their effects by binding to an intracellular protein, the progesterone receptor (PR'). While P R has been studied extensively with respect to its structure and * This work was supported by National Institutes of Health Grants HD21524 and CA18119. A preliminary report of portions of this work was presented at the 67th Annual Endocrine Society Meeting, June, 1985 (21). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
function (1)(2)(3)(4)(5)(6), little is known of the rates of its synthesis and degradation, and of factors that might influence these rates. This study was undertaken to examine P R synthesis and degradation, using the density labeling method.
We have used a well characterized human breast cancer cell line, , for this investigation. The progesterone receptors in these cells appear to be functional, and their level is under estrogen control (8). Whereas cells grown without estradiol have only low, basal P R levels (9), cells treated with estradiol have markedly increased P R levels (8), and with continued estradiol treatment, P R remains elevated at a constant, steady-state level.
We have used the density labeling technique to measure the rates of receptor synthesis and degradation under these steady-state conditions: cells incubated with medium containing normal 14N, "C, 'H amino acids (NAA) are shifted to medium containing dense 15N, 13C, 'H amino acids (DAA); since the newly synthesized receptor is of a higher density than the pre-existing protein, it can be separated and quantitated by gradient centrifugation techniques (10, 11). We have found that the turnover of P R occupied with the high affinity progestin R5020 is considerably faster then the turnover of unoccupied PR. In addition, a lag that characterizes the turnover kinetics of both occupied and unoccupied P R suggests that there is a non-hormone binding biosynthetic precursor to the receptor. 17,, 87 Ci/mmol) and radioinert R5020 (New England Nuclear Corp.); l7fl-estradiol, progesterone, cortisol, leupeptin, phenylmethylsulfonyl flouride, soybean trypsin inhibitor, Tris, EDTA, thioglycerol (Sigma). All chemicals for the SDS-polyacrylamide gel electrophoresis analysis were obtained from Bio-Rad. The Triton/xylene scintillation fluid contained 3 g/ liter 2,5-diphenyloxazole, 0.2 g/literp-bis-[2-(5-phenyloxozoyl)]-benzine, and 25% Triton X-114 in xylene.

Chemicals
Cells and Cell Culture Conditions-MCF-7 cells, obtained from Dr. Charles McGrath of the Michigan Cancer Foundation, were maintained at 37 "C in closed Corning T-75 flasks (Corning Glassworks, Corning, NY) and passaged in logarithmic growth phase. Growth medium was Eagle's minimal essential medium containing Hanks' balanced salts (Grand Island Biological Co. (Gibco), catalog number 410-1200), supplemented with 0.006 pg/ml insulin (Sigma), 10 nM hydrocortisone (Sigma), 0.01 M HEPES buffer (Gibco), 50 pg/ml gentamicin (Schering), 100 units/ml penicillin (Gibco), 100 pg/ml streptomycin (Gibco), and 5% charcoal-dextran-stripped calf serum (CDCS) prepared as described previously (12). At 4 days prior to use, 10 nM estradiol was added to media to increase PR levels. All density shift experiments were performed with near confluent cultures. For experiments, cells were grown in normal amino acid medium supplemented with 15% dialyzed charcoal-dextran-treated serum, 0.5% charcoal-dextran treated but nondialyzed serum, and 10 nM estradiol (NAA medium + estradiol). At times ranging from 1 to 36 h prior to Synthesis and Degradation 13237 harvesting the cells, the normal medium was replaced with medium containing dense amino acids. Dense Aminn Acid Medium-The composition and preparation of the dense amino acid medium was exactly as described previously (13,14). The medium contained 0.5 mg/ml of I3C, 15N, 'H-amino acids (dense amino acids, DAA, Merck Sharp and Dohme; >96% enriched in the dense isotopes). It was necessary to use dialyzed serum for these studies in order to remove amino acids present in the serum. However, in the presence of dialyzed CDCS (CDCS exhaustively dialyzed against Hanks' balanced salt solution), cellular PR levels dropped substantially. Addition of 0.5% CDCS elminated this problem. Thus 100 ml of DAA medium contained 15 ml of dialyzed CDCS plus 0.5 ml of CDCS.
Preparation of Cytosol and Nuclear Extracts-MCF-7 cells from a single near confluent T-75 flask were incubated at 37 "C for the indicated times in NAA medium + estradiol or in DAA medium + estradiol. The medium was decanted and the cells were rinsed once with Hanks' balanced salt solution and harvested with Hanks' balanced salt solution containing l mM EDTA. The cells were collected by centrifugation, washed gently in 2 ml of 10 mM Tris, 1.5 mM EDTA, 10 mM thioglycerol, 10% glycerol, pH 7.4, at 4 "C (TETG buffer) and homogenized in 300 pl of TETG buffer containing protease inhibitors (soybean trypsin inhibitor, 5 mg/ml; leupeptin, 1 mg/ ml, and phenylmethylsulfonyl fluoride 1 mg/ml) by 40 strokes of a Dounce homogenizer (B pestle). The homogenate was centrifuged at 800 X g for 10 min. The supernatant thus obtained was centrifuged at 180,000 X g for 30 min. A 300-pl aliquot of the high-speed supernatant (cytosol) was transferred to a minifuge tube containing [3H] R5020 in 3 pl of absolute ethanol such that the final [3H]R5020 concentration was 10 nM and incubated for 2 h at 0-4 "C. Parallel tubes contained a 100-fold excess of radioinert R5020. At the end of the 2-h incubation, charcoal-dextran slurry (5% acid washed Norit, and 0.5% dextran C in 10 mM Tris-HC1, pH 7.4, at 4 "C, containing 0.02% sodium azide) was added at 1 part slurry to 9 parts cytosol and incubated for 2 min at 0-4 "C followed by centrifugation for 2 min a t 15,600 X g. An aliquot (50 pl) was removed for determination of recovery and 250 pl were layered onto linear 5-20% sucrose gradients.
Cells treated with [3H]R5020 were rinsed, harvested, and homogenized exactly as described above. The 800 X g crude nuclear pellet was resuspended in 300 pl of 10 mM Tris, 1.5 mM EDTA, 10 mM thioglycerol, 10% glycerol, 0.8 M KCl, pH 8.5, at 4 'C (TETGK buffer) containing protease inhibitors. The suspension was incubated for 1 h at 0-4 "C with resuspension every 15 min, and then centrifuged at 180,000 X g for 30 min. 50 pl of supernatant thus obtained (nuclear extract) were retained for determination of recovery and 250 pl were layered onto linear 5-20% sucrose gradients.
Sucrose Gradient Centrifugation and Analysis of Gradient Profiles to Determine the Proportion of Normal ('H, "C, "N) and Dense ('H, 13C, "N) Receptors-Gradients contained 5-20% sucrose and 0.4 M KC1 in 10 mM Tris, 1.5 mM EDTA, 10 mM thioglycerol, 10% glycerol buffer, pH 7.4, at 4 "C, and were prepared in deuterium oxide. Samples were layered onto 3.6-ml gradients. Centrifugation was for 40 h at 357,000 X g (59,000 rpm, Beckman SW 60 Ti rotor) at 0-4 "C and two-drop fractions were collected and counted in Triton/xylene scintillation fluid. Centrifugation for shorter and longer times was tested, but the 40-h centrifugation provided the best separation of normal and dense receptor peaks. ["C]Ovalbumin (3.5 S) and ["Cly-globulin (6.6 S), methylated by the method of Rice and Means (15), were utilized as sedimentation markers. Quantitation of normal and dense gradient peaks was accomplished by use of a general mathematical function plotting program available on the University of Illinois PLAT0 computer system, essentially according to the method of Devroetes et al. (16). A linear descending background was subtracted from the gradient data points prior to data analysis. The normal density peak (IH, lZC, "N-receptor complexes, Oh profiles, no dense amino acid exposure) was used as a "template" for the normal (light) receptor complex and was fitted to a gaussian function. A template for the dense receptor peak was derived by fitting a gaussian function to the right hand side of the 36-h gradient profile (to avoid interference with light receptor). The centers of the dense and light template functions were fixed at their original position. At each time point, the heights of the 2 s was adjusted such that their sums best fit the experimental data, the ratio of the areas of the dense and light templates giving the proportion of dense and light receptors. By this method, we were able to obtain ratios reproducible within +5%.
Photoactivated Covalent Binding of rHjR5020 to PR-Photolysis was performed in a Rayonet photochemical reactor (RPR-100) as described (17) using seven 8-W black light bulbs with a 350-nm emission. Samples in quartz tubes were placed in a lucite holder which is rotated at 50 rpm within a 1600-ml quartz beaker filled with an ice-water slurry during the 60-min irradiation. The efficiency of R5020 covalent attachment was typically 12%.

RESULTS
Stimulation of PR Levels in MCF-7 Cells-It is most convenient to apply the density shift technique to measure receptor turnover when the system is at steady-state with respect to PR synthesis and degradation rates. Therefore, we studied the time course of PR stimulation by treating MCF-7 cells with 10 nM estradiol for 1-6 days before determining PR levels. As seen in Fig. 1, PR levels rose from low basal levels in the controls (0.58 t 0.01 pmol/mg DNA) to reach maximal PR levels which were achieved by day 2, and these levels (4.1 k 0.1 pmol/mg DNA; 6.5-7.8-fold increase above the controls) remained constant for the remaining period assayed.
Dense Amino Acid Labeling of Unoccupied PR-To determine the rates of synthesis and degradation of the unoccupied PR, MCF-7 cells that had been treated with 10 nM estradiol for 4 days were incubated at 37 "C in normal amino acidcontaining medium with estradiol (NAA + estradiol medium) for 36 h or until replacement with dense amino acid medium plus estradiol (DAA + estradiol medium) at various times ranging from 6 to 36 h before cell harvest and fractionation. Cytosol was then prepared, labeled in vitro with [3H]R5020, and layered onto 5-20% sucrose gradients containing 0.4 M KCI, as described under "Materials and Methods." Fig. 2 shows that the "new-dense'' (newly synthesized) receptor and the "old-light" (pre-existing) receptor are well Estradiol-stimulated increase in the cellular PR content of MCF-'7 cells. MCF-7 cells in parallel 75-cmZ flasks were exposed to 10 nM estradiol or control vehicle (0.1% ethanol ) for 6,5, 4,3, 2, or 1 day before cells were harvested and assayed for PR levels by charcoal-dextran assay. Estradiol and fresh media were renewed daily. DNA was determined by the Burton assay (36). 0, ., and A represent different experiments having control levels of PR of 0.60, 0.55, and 0.59 pmol/mg DNA, respectively.

FIG. 2. Density shift of the unoccupied MCF-7 cytosolic PR
labeled for various times with dense amino acids. Near confluent 75-cm2 flasks of MCF-7 cells that had been treated with 10 nM estradiol for 4 days were transferred to NAA + estradiol media and at the times indicated were transferred to DAA + estradiol media.
The gradients were centrifuged for 40 h at 357,000 X g and two-drop fractions were collected. Radioactivity was measured in each fraction using a Triton/xylene based scintillation fluid. ['4C]Ovalbumin (3.5 S) and ["C]y-globulin (6.6 S) were used as internal markers in all gradients. The dashed lines give only an approximate indication of the normal and dense receptor peaks which were quantitated as described under "Materials and Methods." (NAA, normal amino acids; DAA, dense amino acids).
resolved by this velocity sedimentation technique. PR from cells grown in NAA + estradiol medium sediments as a single 4 S moiety. After 6 h of exposure to DAA + estradiol medium, there is little change in the [3H]R5020-labeled PR profiles on sucrose gradients. Thereafter, however, turnover appears to accelerate, such that by 12 h, nearly 35% of the total PR has turned over (and is thus represented by a progressive increase in the amount of dense receptor, with a concomitant and proportional decrease in the light receptor), and by 36 h, nearly all the [3H]R5020-labeled receptor is seen at the position of the newly synthesized, dense receptor. Thus, there appears to be a lag in the appearance of newly synthesized receptor. This lag is more apparent in the kinetic plot (Fig.  5A), and the significance of this observation is discussed later.
As we fractionate MCF-7 cells, PR that we obtain in the cytosol fraction represents only about 30% of total PR in cells. Therefore, in order to examine whether turnover of total cell PR was the same as that for cytosolic PR, we performed a similar time course of dense amino acid labeling. Cells were incubated in NAA + estradiol medium with subsequent transfer to DAA + estradiol medium, but at 1 h prior to harvesting of the cells, [3H]R5020 was added to the growth medium to label total cellular PR. This total cell PR was then extracted with 0.6 M salt and analyzed on density gradients (Fig. 3). Fig. 3 shows that the rate of disappearance of normal density R5020 was added to the medium. Nuclear salt extracts were then prepared and a 25O-pl aliquot of the nuclear extract was layered onto 0.4 M KC1 containing 5-20% sucrose gradients prepared in TETG; pH 7.4, buffered DzO. The gradients were centrifuged for 40 h at 357,000 X g and two-drop fractions were collected at the end of the centrifugation. Radioactivity was measured in each fraction using a Triton/xylene based scintillation fluid. ["C]Ovalbumin (3.5 S ) and ['4C]y-globulin (6.6 S) were used as internal markers in all gradients. receptor and rate of appearance of dense receptor are virtually identical to that of the unoccupied cytosol PR (Fig. 2), suggesting a similar turnover for unoccupied receptor observed in the cytosolic or total cellular extract (cf. Fig. 5

, A uersus B).
It is of note that over the period of 36 h there was no significant decrease in total PR labeled in cells. Also, the newly synthesized, dense receptor was indistinguishable from the normal density receptor with respect to affinity for [3H] R5020 (data not shown). Thus, it was possible to quantitate relative amounts of the normal and dense receptors and thereby to determine the rate of decrease of the normal density receptor (see Fig. 5 ) .
Dense Amino Acid Labeling of PR Occupied by R5020"We next wished to determine whether occupancy of PR by the potent progestin R5020 influenced receptor turnover. To perform these experiments, cells were treated with 10 nM [3H] R5020 for 24 h prior to DAA exposure. This concentration of hormone was found to be adequate to saturate the receptor, and it resulted in localization of over 90% of PR in the nucleus (19). Furthermore, 10 nM R5020 does not affect the rate of cell growth over this brief time, although it does reduce cell growth rate over a more prolonged time period. In addition, our measurement of total cell protein synthesis rate, as monitored by [3H]leucine incorporation, showed no alteration over a 72 h period of exposure to 10 nM R5020. With this concentration of R5020, cytosol receptors were depleted, and there was a corresponding increase in the level of nuclear receptor that remained elevated for at least 72 h.
Thus, to measure the turnover of R5020 occupied PR, cells treated with 10 nM estradiol for 4 days to elevate PR were exposed for 24 h to 10 nM [3H]R5020. At various times over the next 36-h interval, separate flasks of cells were transferred to DAA + estradiol medium also supplemented with 10 nM [3H]R5020 (DAA medium + estradiol + [3H]R5020). At the 36-h point, cells were harvested and fractionated, and nuclear extracts were prepared and analyzed on density gradients. Fig. 4 shows the PR profiles obtained. As before, PR extracted from cells that had been only in normal amino acidcontaining medium (0 h), sedimented as a single 4 S species.
However, in contrast to the situation seen with unoccupied PR (cf. Figs. 2 and 3), the [3H]R5020 occupied receptor showed a more rapid shift to the dense species. At 6 h, the newly synthesized receptor represented approximately 10% of the total PR, by 12 h, approximately 50% of the old, normal density receptor had been degraded and replaced by an equal amount of dense receptor, and by 18 h, this process was 90% complete. At 24 and 36 h (not shown), there was no change in the profile from that obtained at 18 h. Thus, exposure to 10 nM [3H]R5020 results in a more rapid turnover of PR in these cells. In addition, there again seems to be a lag before newly synthesized receptor is detected (cf. Fig. 5C). It is very unlikely this lag represents a problem in dense amino acid equilibration into amino acid pools, since studies in the same MCF-7 cells on estrogen receptor turnover indicated no lag separated by sucrose gradient centrifugation, by covalent photolabeling with [3H]R5020, and electrophoretic separation of a marked difference in the turnover rates of the two subunits, one would expect that their relative proportions in the normal and dense regions of the sucrose gradient would change with time. For example, if one subunit were turning over very DAA rapidly in comparison with the other, it would disappear faster from the normal density region of the gradient and appear more rapidly in the dense region.
With the above rationale in mind, we exposed cells to 14

FIG. 4. Density shift of MCF-7 [SH]RS020 occupied PR labeled for various times with dense amino acids.
Near confluent 75-cm2 flasks of MCF-7 cells that had been treated with 10 nM estradiol for 4 days were then exposed to 10 nM 13H]R5020 (in the continued presence of estradiol) for 24 h before cells were transferred to NAA medium + estradiol + [3H]R5020 or DAA medium + estradiol + [3H]R5020 as indicated in thepanels. At the end of the dense amino acid labeling period, cells were harvested. Nuclear salt extracts were prepared and a 250-pl aliquot of the nuclear extract was layered onto 5-2076 sucrose gradients prepared in TETG, pH 7.4, buffered D20 containing 0.4 M KCl. The gradients were centrifuged for 40 h at 357,000 X g and two-drop fractions were collected. Radioactivity was measured in each fraction using a Triton/xylene based scintillation fluid. ["C]Ovalbumin (3.5 S) and ["Cly-globulin (6.6 S ) were used as internal markers in all gradients. dense receptor and be very low in the normal density region of the sucrose gradient.) Cells in the control flask were incubated in NAA medium + estradiol + [3H]R5020. The cells were harvested and fractionated, and nuclear extracts were layered onto sucrose gradients containing 0.4 M KC1 in buffered deuterium oxide. After centrifugation, the gradients were fractionated into minifuge tubes containing 10 gl of 10 mg/ ml leupeptin to reduce the possibility of proteolysis. A small portion of each fraction was counted to determine the regions of the gradient that contained normal density and heavy PR (Fig. 6, upper panels).
As shown in Fig. 6A (top), fractions 20 and 21 from the gradient of the control experiment were selected and pooled Near confluent 75-cmZ flasks of MCF-7 cells that had been treated with 10 nM estradiol were then exposed to 10 nM [3H]R5020 (in the continued presence of estradiol) for 24 h prior to transfer to NAA medium + estradiol + [3H]R5020 (A) or DAA medium + estradiol + [3H]R5020 (B) for 14 h. Following this 14-h period, cells were harvested and fractionated. Nuclear extracts were prepared and a 250-pl aliquot was layered onto 0.4 M KC1 containing 5-20% sucrose gradients prepared in buffered D20. The gradients were centrifuged for 40 h at 357,000 X g and at the end of the centrifugation, two-drop fractions were collected into cooled microfuge tubes containing a IO-fold concentrated solution of leupeptin so that the final concentration in the fraction was 1 mg/ ml. Aliquots (IO pl) of each of these fractions were counted. Peak fractions from A (20 and 21) and from B (20 and 21 representing normal density receptor and 26 and 27 representing dense PR) were separately pooled and irradiated for 60 min at 0-4 "C using a 350-nm wavelength light source. These photoaffinity labeled PR preparations were then analyzed on SDS-polyacrylamide gels (lower panels). The M, values were determined by comparison with the mobilities of known protein standards.
to represent normal density, control receptor, and as is shown in Fig. 6B (top), from the 14-h density shift experiment, fractions 20 and 21 were pooled to represent the normal density PR, and fractions 26 and 27 were pooled to represent the dense receptor. These three samples were irradiated for 1 h a t 0-4 "C a t 350 nm to effect R5020-PR photocross-linking, and were then prepared for analysis on SDS-polyacrylamide gels. The SDS gel profiles (Fig. 6, lower panels) revealed that the B:A (Mr-115,00085,000) ratio remained very similar during the course of PR turnover. Thus, assuming equal efficiency of photoactivated cross-linking of R5020, we can conclude that the turnover of the two PR subunits occupied by R5020 is very nearly the same.
Degradation of PR in Cells Exposed to Cycloheximide-Because of the unusual turnover kinetics of PR determined by density labeling, we studied the rate of degradation of PR by a different, more widely used method, wherein biosynthetic inhibitors are used to reduce cellular protein synthesis to insignificant levels, and the rate of disappearance of preexistent proteins is monitored. For these studies, MCF-7 cells were exposed to M cycloheximide (a concentration which inhibited protein synthesis over 97%) at the indicated times prior to cell harvest and fractionation. A t1h of approximately 18 h was observed (Fig. 7). Interestingly, after cycloheximide treatment, no lag was seen in the degradation of PR.

DISCUSSION
Since the actions of progestins appear to be mediated via interaction with PR, there has been tremendous interest in understanding the basic biochemistry of this receptor and its role in regulating the function of reproductive tissues and the response of human breast and endometrial cancers to hormone treatment. PR measurements, along with estrogen re- Near confluent 75-cm2 flasks of MCF-7 cells that had been treated with 10 nM estradiol for 4 days were exposed to lo" M cycloheximide at the indicated times prior to cell hamest, and in the last 30 min of the cycloheximide treatment, cells were exposed to 10 nM [3H]R5020 zk 100-fold excess unlabeled R5020. Nuclear extracts were prepared, charcoal-dextran treated to remove free ligand and aliquots were counted using a Triton/xylene based scintillation fluid. Values represent specific [3H]R5020 binding (total [3H]R5020 binding minus binding in the presence of 100-fold excess [3H]R5020). ceptor measurements, are now widely used in assessing the hormonal dependence of human breast cancer and in predicting the therapeutic utility of endocrine therapy and the disease-free survival of patients (22, 23).
As a step toward understanding some fundamental aspects of the regulation of PR levels in cells, we undertook a study to determine the rates of synthesis and degradation of this molecule and to study the factors that might influence these rates. We have used a powerful density shift technique to address the question of rates of synthesis and degradation of this protein. Cells were exposed to DM-containing media such that any newly synthesized proteins incorporating the dense amino acids were distinguishable from the pre-existent pool of proteins by their faster sedimentation on sucrose

Progesterone Receptor Synthesis and Degradation 13241
gradients. PR labeled with tritiated ligand can thus be studied with respect to rates of synthesis (appearance of dense PR) or degradation (disappearance of normal density PR). This technique has the advantage of causing little perturbation to the system being studied when compared with other methods frequently employed (use of inhibitors of protein synthesis or of transcription). We found that after DAA addition, the time required for the level of the unoccupied form of the PR to be reduced to half of the initial value is approximately 17 h, while this time is reduced to 12 h for PR occupied by the progestin R5020. These observations of an effect of R5020 Occupancy On receptor turnover brought us into a very interesting, although rather controversial subject in the field of PR studies, namely, whether the PR is constituted by one or two proteins and the relationship between the observed A and B subunits. Until recently, it was well accepted that PR was composed of two structurally and functionally dissimilar subunits; their molecular weights were estimated to be 79,000 and 108,000 on the basis of studies performed on the chick oviduct (2), and two subunits of similar or somewhat larger size were found in T47D human breast cancer cells (5). In contrast, work done on the rabbit PR system seems to indicate that a single M,-110,000 protein may function as the PR in this species (6). In our studies, we have always seen both MJ35,OOO and 115,000 proteins during SDS-polyacrylamide gel electrophoretic analysis of R5020 binding to cytosolic or nuclear PR. (The M, values we have obtained in 4 separate analyses are 85,000 f 3,000 for the A unit and 113,500 -+ 2,200 for the B unit).
However, that does not rule out the possibility that the smaller species is being generated by proteolysis of the larger during cell fractionation, even though protease inhibitors (leupeptin, phenylmethylsulfonyl fluoride, and soybean trypsin inhibitor) are present in the homogenization and nuclear extraction buffers, and care is taken that the samples remain at 0-4 "C. Our studies show that both subunits have the same turnover rate, and that they are not linked in a precursor-product relationship in uiuo. This would be consistent either with a model in which the subunits arise independently from distinct precursors, or a model in which there is in uiuo only one form of PR (B) with A being generated only upon proteolysis during cell fractionation.
Of great interest is our finding of a pronounced lag in density shift kinetics observed in experiments employing dense amino acid labeling of the PR. This lag can best be appreciated in Fig. 5, which summarizes data from the experiments presented in Figs. 2-4 plus several additional experiments; these semi-logarithmic plots reveal that the decay of normal density receptor is clearly not first order, but has a pronounced lag, regardless of whether unoccupied or R5020occupied PR is being observed. This lag in the decay profile would be consistent with the existence of a biosynthetic intermediate between amino acids and the ligand-binding form of the receptor that is assayed by this method, This precursor form would not bind hormone, but would be converted into the binding form by some activation process. The lag in the decay of the hormone binding form (receptor) after density shift would result from the fact that the normal density hormone binding form of the receptor continues to be produced for as long as the pool of normal density precursor persists; only as the pool of normal density precursor becomes depleted, does the decay rate of normal density receptor increase to maximum. It is extremely unlikely that this lag represents a problem in dense amino acid equilibration in amino acid pools, since studies in the same MCF-7 cells on estrogen receptor turnover indicated no lag before incorporation into estrogen receptors (13,14,20,21). In addition, it is also unlikely that this lag represents a precursor-product relationship between the B and A subunits of PR. Our experiments analyzing the subunit composition of PR during the density shift (Fig. 6) indicate that the turnover rates of the B and A subunit are very similar. Also, a kinetic model involving a precursor-product relationship between subunits B and A would predict only a modest change in the slope of the normal density receptor decay curve, not a pronounced lag as we have observed.
If the hormone-binding form of PR is derived from a nonhormone-binding precursor by an activation process, then analysis of the non-linear (lag-decay) kinetics requires a more complex kinetic model than is typically used to analyze the kinetics of protein turnover. The ligand binding form is designatid R (receptor), and the precursor form pre-R. The zero-order biosynthetic rate constant is k, and kl, and k2 are the first order rate constants for the activation process and the degradation process, respectively. A complete mathematical presentation of this model is given in the Appendix; a summarized version will be given here, together with the conclusions that can be drawn by the application of this model to the density-shift turnover kinetics we have determined for PR. Just prior to the shift to dense amino acids, the system is in a steady state, i.e. the concentration of pre-R and R are constant, since each is being formed and degraded at the same rate. Mathematically, this initial state can be expressed as: ko = kl.pre-Ro = k2.Ro (1) where pre-Ro and Ro are the concentrations of the receptor precursor and receptor at time 0. To express this proportionality relationship, it is convenient to define a proportionality constant a (Equation 2), as the ratio of the precursor to receptor concentrations, or, what is equivalent, the ratio of the rate constants of degradation to activation.

Pre-R, -kz
Ro kl (2) Very soon after the normal density amino acids are replaced by heavy amino acids, the biosynthesis of the normal density forms of the receptor ceases (& goes to 0). The following differential equations then describe the decay kinetics of the normal density species pre-R and R.

_ _ _ -
The first differential equation is easily integrated (Equation This form of the differential equation for R (Equation 6) can be integrated using Laplace transforms (as described in the Appendix), and the integrated form can be expressed solely in terms of Ro, k2, and the proportionality factor a.
Different integrated rate equations are used to describe the situations where a # 1 (precursor and receptor pool sizes different; Equation 7) and a = 1 (precursor and receptor pool sizes the same; Equation 8). It is interesting that because of the initial steady-state boundary condition (Equation l), the decay kinetics of the system are fully defined by specifying only two terms, the degradation rate constant (k2) and the steady-state proportionality constant ( a ) . Ro, the initial concentration of R, is determined by direct measurement, and the other terms (pre-Ro, kl, and k,,) can be derived from kp, a, and Ro, using Equations 1 and 2.
The general character of these functions (that is, the shape of the decay curves for pre-R and R as a function of a, for a fixed value of Ro and k,) is explored in greater detail in the Appendix. Either a non-linear curve fitting procedure or a graphical analysis method (that utilizes the measurement of limiting slopes and intercepts) can be used to extract values of kz and a from our experimental kinetic data (Fig. 5 ) . The results of this analysis are given in Table I. As is explained in greater detail in the Appendix, three cases must be considered: Case I (a < 1):

TABLE I Determination of kinetic parameters for progesterone receptor synthesis, activation, and degradation in the unoccupied and R5020
occupied state from kinetic modeling analyses Values for k, and a have been determined from the data presented in Fig. 5, using the graphical analysis method described in the Appendix. The standard error of the mean for the measurement of Ro is given in the Table ( n = 6). Hence the coefficient of variation at the 95% confidence limit for the measurement of Ro (unoccupied) is 10% and for Ro (R5020 occupied) is 14%. Error (not calculated) is also introduced in determining the limiting slope and intercept from the lag-decay profiles and is propagated into the calculations of the other kinetic parameters. Therefore the values in the Table are [3.51 Ro ( Case ZII (a = 1): In -= -k*t + ln(l + k2t) ["df' l and for each case, there is a convenient form of the integrated rate expression (Equations 9-11). Kinetic analysis of the lagdecay profile alone does not permit one to distinguish between Cases I and I1 (cf. Appendix); independent methods are needed to determine whether the precursor pool is larger or smaller than the receptor pool. Therefore, two sets of estimates for the pool sizes and kinetic rate constants, that correspond to Cases I and 11, are presented in Table I. (We are justified in not considering case I11 ( a = l), since in all cases the values estimated for a differ substantially from I).
There is a reciprocity in the equations that describe Cases I and 11, such that analysis of the same data by the two expressions produce values for a that are reciprocals of one another. As a result, as can be seen in Table I, the kinetic rate constants k, and k,, determined according to Case I, become the values for kz and k,, respectively, determined in Case 11. The pool size for R (R,) is the same for both cases, since it is determined independently, by assay of the receptor content at time 0. However, estimates for the pool size of pre-R (pre-Ro) and for the biosynthetic rate constant (kJ differ, as these both depend on the value of a. It is apparent from the ka values ( Table I) occupied receptor is degraded more than 50% faster than unoccupied receptor; the rate constant for precursor activation (kl) and for overall biosynthesis (h) for R5020 occupied receptor are also larger, but the precursor pool size (pre-Ro) is smaller. Hence, it is of note that R5020 accelerates PR turnover. Ligand binding has also been shown to accelerate turnover of the glucocorticoid receptor and thyroid hormone receptor (24,25), although we observed little effect of estrogen on the rapid rate of turnover (tlh = 3-4 h) of the estrogen receptor in MCF-7 cells (13, 26). Furthermore, it is of note that the turnover rates of steroid and thyroid hormone receptors cover a considerable range, having half-lives in the unoccupied state from 3-4 h for estrogen receptors (13, 14,20,21,26) and androgen receptors (27), and 5 h for thyroid hormone receptors (25), to approximately 19 h for glucocorticoid receptors (24). From this study, which is the first report on turnover of the PR using density shift analyses, we find rather rapid rates of degradation of the hormone binding form of PR (tlA of 4.5 or 8.0 h), with conversion of prereceptor to the hormone binding form occurring with a tl,2 of 8.0 or 4.5 h, depending upon estimates of the precursor pool size.
It is of note that no lag is observed in the degradation of PR in MCF-7 cells when protein synthesis is blocked by the use of cycloheximide and that a longer apparent half-life for PR (18 h) is then estimated. The hypothesized precursor pool thus appears no longer capable of generating hormone binding moieties. This would be explicable on the basis of the presence of a rapidly turning over activating enzyme: inhibition of protein synthesis would cause a dramatic decrease in the enzymatic activity, thus precluding receptor activation to a hormone binding form. In addition, degradation of PR could be slowed if a rapidly turning over enzyme is involved in the PR degradation. Thus, the tlh for PR observed in the studies using cycloheximide is considerably longer (18 h) than that Synthesis and Degradation 13243 measured by the density shift procedure. In T47D human breast cancer cells, the use of cycloheximide also revealed an apparent half-life of 16 h for the PR (28). Hence, cycloheximide results in what is most likely an artifactually prolonged half-life for PR. Such an effect of cycloheximide on degradation has been observed for other proteins (29, 30), including the estrogen receptor in these MCF-7 cells and in uterine cells (21). The observation that some cells produce immunoreactive non-hormone binding proteins (presumably receptor-related forms), that specifically react with highly purified antireceptor monoclonal antibodies, may be particularly relevant to our findings. Several groups have reported recently on the presence of immunoreactive 94,000 molecular weight material (the same molecular weight as the normal hormone binding glucocorticoid receptor) that fails to bind hormone in some mutant rat hepatoma (HTC) and mouse lymphoma (S49 and WEHI7) cells (31-33). Interestingly, both wild type and mutant (nuclear transfer deficient and "receptorless") S49 cells contain significantly more immunoreactive material than hormone binding activity. Indeed, the receptorless mutant, which virtually lacks hormone binding activity, nevertheless produces 20-50% as much 94,000 molecular weight immunoreactive material as the wild type parent (31-33). In addition, the recent use of cDNA probes for the glucocorticoid receptor have predicted the occurrence of hormone-binding and nonhormone binding glucocorticoid receptor forms in the same cell (34). In the case of the PR, monoclonal antibodies generated to the subunit B protein from chick oviduct (35) have been able to detect antibody-antigen interactions only under denaturing conditions, and it has been concluded at this point that the antibodies react either with a denatured conformation of receptor that does not bind hormone or with a modified, possibly precursor form of receptor unable to bind hormone.
In conclusion, the turnover of PR in MCF-7 cells, as we have studied it by the density shift technique, appears to be complex, with a pronounced lag in decay kinetics being consistent with the existence of a non-hormone binding biosynthetic precursor to receptor. Further biochemical studies and the use of antireceptor immunochemical probes should enable us to learn more about these progesterone receptor precursors.