Interactions of Glucocorticoid Receptor l Steroid Complexes with Acceptor Sites*

The binding of the “activated” receptor .glucocorticoid complexes of cultured rat hepatoma cells to nuclei, chromatin, and DNA has been studied under cell-free conditions. A critical factor in determining the shape of the binding curve is shown to be an inhibitory material which is present in crude cytosol and which can be removed without destroying the receptor .steroid complex. These and other results argue that the apparent saturation observed in earlier experiments may have been due to the inhibitors. Thus, the actual number of acceptor sites in hepatoma tissue culture cell nuclei is much larger than previously estimated and their affinity for the complex is lower. Nuclear binding experiments indicate that the inhibitory material interacts with the receptor .steroid complex. The inhibitors appear to be macromolec-ular; but their effects cannot be mimicked by albumin or hemoglobin. The acceptor capacity at low ionic strength for binding receptor .glucocorticoid complexes increases when proceeding from nuclei to DNA. An analysis of the kinetics of association and dissociation and of the relative binding behavior of nucle‘i and DNA argues that the affinity of complex for nuclei is much greater than for DNA. DNA-associated histones reduce the amount of complex that binds to DNA. These and perhaps other chromosomal proteins may be responsible for the ordering of acceptor capacity. Evidence

late in the nucleus (4, 5, 8, 10-13) where a sizable portion is associated with acceptor sites in the chromatin (10, 12, 14, 15). In an as yet undefined manner, formation of this ternary chromatin receptor. steroid complex is thought to alter transcription to produce increased amounts of specific messenger RNAs (10, 16, 17). The translation products of these molecules give rise to the observed biological effects (10, 12, 18).
In view of the central role of the nucleus in steroid action much effort has been expended trying to determine both the number and nature of the nuclear acceptor sites. Earlier experiments had suggested that there may be about 4000 such sites per haploid genome with affinities in the neighborhood of 10" M-' (4,5,19,20). Other investigations indicate a considerably larger number of potential nuclear sites with a much lower affinity (21)(22)(23). ' The present studies were undertaken to examine these questions in cell-free systems using glucocorticoid-responsive, rat hepatoma tissue culture cells (24). We report that material in cytosol inhibits receptor. steroid complex association with isolated nuclei, chromatin, or DNA. This inhibitory material influences the binding experiments with all acceptors such that a small number of high affinity sites are observed only when increasing concentrations of complex and inhibitor are employed. We have also investigated the chemical nature of the nuclear acceptor sites. Previous biochemical and genetic evidence has indicated that DNA plays an important role in these sites (25,26). Other studies, however, have implicated chromosomal proteins (10, 27,28), ribonucleoprotein particles (29), and the nuclear membrane (30).

Fractionation
of HTC? Cells HTC cells were grown and harvested (32,33), and cytosol (asteroid) was prepared as described previously (4), except that the homogenization buffer also contained lo-' M dithiothreitol. Alternatively, if the 600 x g supernatant of the crude cell homogenate was centrifuged at 35,000 x g for 15 min before adding any steroid and then spun at 105,000 x g for 90 to 105 min, the final cytosol solution was much clearer. In all the properties investigated (see below) these two cytosol preparations were identical. The preparation of nuclei in sucrose buffer (above homogenization buffer with 0.25 M sucrose) has been described (4). Variations in the quality of the nuclei had no apparent effect on nuclear binding. To prepare HTC nuclei labeled with ['lC]DNA, 1.0 ml of [methyl-"Clthymidine (1 pCi/ml) in homogenization buffer was added to each B-liter culture of HTC cells. After labeling for 16 to 20 hours the cells were harvested as usual (4). For the preparation of chromatin, nuclei labeled with 10 times the above amount of [methyl-"Clthymidine were washed once in about 10 volumes of 0.01 M Tris, pH 8.0 at 0", and then allowed to swell in 3.5 volumes of the same buffer for 1 to 2 hours. LNuclear breakage was achieved by sonication (Bronwill Biosonic IV, 0.9 cm diameter probe, 3 x 6.5 s at 50% output on HIGH).
After centrifugation at 20,000 x g for 15 min, the supernatant layer containing chromatin had 88 to 98% of the total nuclear DNA at about 2.5 mg/ml.
The activated form of the receptor.steroid complex (i.e. RS** or the form that is capable of binding to nuclei) was quantified in an "excess nuclei" assay containing 0.1 ml of activated cytosol and 0.6 ml of sucrose buffer with nuclei from 800 ml of cells at 5 x 10' cells/ml (4). The nuclei were suspended with a glass rod and incubated and washed as previously described (4). The pellet was digested overnight at room temperature and then at 45" for 2 hours in 1.0 ml of NCS tissue solubilizer (Amersham) before being assayed for radioactivity (4 Up to 0.7 ml of sucrose buffer or steroid-free cytosol and partially purified activated complex or crude activated cytosol (all in the same sucrose buffer), was added to the pellets of 0.5.ml aliquots of suspended nuclei (4) for a total volume of about 0.8 ml. Other details are indicated in the text and legends. The nuclei were suspended with a glass rod and binding was measured after 2 to 3 hours as described (4) except that the final nuclear pellet was digested in 0.4 ml of NCS. If the nuclei are not digested before being counted, the 3H counts per min are 5 to 10% and the "C counts per min are 20 to 35% lower. Differences in quenching in similarly treated samples were negligible.
In most cases, about 20% of the total complexes bound to nuclei under conditions of "Constant Protein" (see "Results"), a, to the pellets of 0.5 ml of suspended nuclei in sucrose buffer and the indicated amount of steroid-free cytosol in the same buffer (total volume of added solutions = 0.6 ml) was added 0.1 ml of crude receptor 'steroid complex.
The samples were incubated and worked up as described under "Materials and Methods." Nuclear bound Wsteroid was expressed as per cent of control (i.e. the 3H-steroid bound in the absence of added steroidfree cytosol). b, steroid-free cytosol (0.0 or 0.7 ml) and Tris/glycerol buffer (0.7 or 0.0 ml) were added to DNA-cellulose and cellulose pellets (8 ~1) followed by 0.  (0)). Increasing amounts of partially purified complex (0.05 to 0.7 ml) were also added to the same amount of nuclei and the total assay volume completed with buffer (Purified RS Complex (0)). After incubation and work-up ("Materials and Methods"), the free receptor .steroid complex (including that which will not bind to nuclei) was calculated as (total added complex) -(nuclear bound complex).  (Fig. 86). This is exactly as predicted for Model II but is inconsistent with Model I (see Fig. 6 inhibitor (Fig. 9). Neither bovine serum albumin nor hemoglobin caused a significant reduction in the binding of complex to DNA. In contrast, lysozyme and the histones (Fig. 9, a and b) significantly inhibited the binding of receptor. steroid complex. The magnitude of this inhibition was variable, however, with the effect of histones always greater than that of lysozyme, Higher concentrations (Le. 2600 fig/ml) of both lysozyme and histone precipitated up to 65% of the total receptor. steroid complexes. These results raised the possibility that the formation of soluble aggregates of the complex and lower concentrations of histones or lysozyme could account for the observed inhibition (e.g. Model II, Fig. 6). However, this explanation appears unlikely since histone is a much more effective inhibitor of binding than lysozyme and yet both proteins precipitate the complex equally well. Furthermore, when DNA was preincubated with either histone or lysozyme and the nonabsorbed protein removed, the pretreated DNA had a diminished complex-binding activity equivalent to that seen when the excess basic protein was not removed.3Thus, only the histone or lysozyme bound to DNA is responsible for the inhibition of complex binding. At low concentrations (2 10 pg/ml), both histone and lysozyme increase complex binding to DNA (Fig. 9). Again the enhancement was variable (up to a 50% increase) and due to DNA-bound proteins; and, histones were always more effective than lysozyme.
Differences in Capacity and Affinity of Acceptors for Binding Complexes- Fig.  10 shows that the binding activity decreases dramatically in going from DNA to chromatin or nuclei. These differences could mean that nuclei have either a lower affinity, or a lower capacity to bind complex than does DNA, or both. In an effort to resolve this question, it was noticed that the observed half-time of association of complex with nuclei was about 0.25 hour at 0" (4) while the half-time of dissociation of nuclear-bound complexes could easily be as long as 30 hours5 The binding reaction is considered to be at equilibrium after 2 hours at 0" since no further change occurs with time. It is more difficult to determine the nature of the dissociation reaction. However, dissociation does not occur unless the concentration of complex in solution is lowered. This suggests that an equilibrium reaction is involved. The association of complex with DNA reached equilibrium after about 2.5 hours (Fig. lla) and has an apparent tH of about 0.4 hour.61n Fig. lib, the half-time for DNA-complex dissociation was found to be about 0.35 hour.6 This reaction appears reversible since multiple washes of bound complexes remove more radioactivity than dissociates at time T, when the same sample of DNA-bound complexes is resuspended once.3 Fur-SThis figure was calculated from the data in Ref. 40 with the assumption that only one-half of the nuclear bond receptor .steroid complex would have dissociated at the time T,. @The difference between these results and those of a previous study (43) 9. Effect of protein on the binding of receptor. steroid complex to DNA. Aliquots (0.7 ml) of buffer or various proteins at different concentrations were added to 8 al pellets of either cellulose or DNA-cellulose, followed by 0.1 ml of crude receptor. steroid complex. Incubation and work-up were performed as described under "Materials and Methods." At each concentration of protein used, the radioactivity bound to DNA was plotted as per cent of control (0.7 ml of buffer added) for a, calf thymus histones; b, lysozyme; c, hemoglobin; and d, bovine serum albumin.
In each case, the cross-hatching represents the binding, as per cent of control, observed when 0.6 ml of steroidfree cytosol and 0.1 ml of buffer were added instead of 0.7 ml of buffer. thermore, no net dissociation occurs unless the free complex present at the end of the binding reaction is removed (see Fig.   lla).
Thus the reactions of receptor.steroid complex with nuclei and with DNA appear to be reversible reactions which can be described by Equations 1 and 2 of Fig. 12. The association reactions appear to follow pseudo-first order kinetics going to equilibrium since, with small amounts of added receptor. steroid complex, the bound acceptors are a small percentage of the total acceptor sites (see Figs. 3 and 5 and "Discussion"). Equations for the similar observed t, values for these reactions then give Equation 1 (all symbols defined in Fig. 12). While initial rate plots of dissociation have not been obtained, the above difference in apparent tV, values for dissociation indicates that the initial rates are quite different, in fact ( crude receptor'steroid complex (0.2 ml) and buffer (0.6 ml) were incubated with cellulose or DNA-cellulose for the indicated time, centrifuged at 600 x g for 5 min, and the radioactivity bound to cellulose (m), DNA-cellulose (O), and that specifically bound to DNA (A) was determined ("Materials and Methods"). The high cellulose "backgrounds" (see Fig. 5) are a consequence of the fact that the DNA-cellulose and cellulose pellets were not washed before being counted. However, it is better to maintain near-equilibrium conditions and the associated high backgrounds than to wash the pellets, which results in spuriously low values due to the dissociation of complexes. b, kinetics of dissociation: crude receptor. steroid complex (0.2 ml) and buffer (0.6 ml) were incubated with duplicate samples of cellulose and DNA-cellulose for 2.5 hours ("Materials and Methods"). The samples were centrifuged (1000 x g for 1 min) and the supernatants removed by aspiration.
Freshly prepared steroid-free cytosol (0.2 ml) and buffer (0.6 ml) were added to the pellets which were incubated for up to 180 min and then worked up as described under "Materials and Methods." The nonspecific, or cellulose-bound, radioactivity was released so rapidly that its level did not change during the experiment. has been ignored in this analysis since its effects appear negligible at low concentrations (see Figs. 3 and 4) and since its main effects may be only on the final levels of binding and not on the rates of reaction. The latter assumption, with regard to dissociation, was tested directly.
The observed t, values for dissociation of DNA-bound complex were the same with, or without, 0.  (Fig. 7c) argues that the reactions can be described as shown in Fig. 6 (Model II) and that, at low concentrations of protein, the inhibitor can be neglected to yield the equations of Fig. 12. Finally, the above conclusion is valid only if the concentration of nuclear sites was not much greater than DNA sites in the association experiments.
The data in Fig. 10  and they may increase the binding affinity of some, or all, of the DNA binding sites which remain exposed.