Thyrotropin Receptors in Thyroid Plasma Membranes CHARACTERISTICS OF THYROTROPIN BINDING AND SOLUBILIZATION OF THYROTROPIN RECEPTOR ACTIVITY BY TRYPTIC DIGESTION

Biologically active bovine “%thyrotropin preparations have been prepared, characterized, and used to evaluate the optimal conditions for thryotropin binding to bovine thyroid plasma membranes in uitro. Binding of lz5 I-TSH has a pH optimum around 6.0 and is sensitive to the choice and concentration of buffer. Binding is inhibited by salts, especially those containing magnesium and calcium ions; magnesium concentrations optimal for adenylate cyclase assays (2 to 5 mM) result in 85 to 98% inhibition of binding. Binding is temperature sensitive. At 3rj” binding its highest initial level; however, instability of the membrane at this temperature a rapid of binding activity. Binding at 0” is optimal in 30 min and at the same level as initial binding at 37”; since there is no decrease in binding activity, it has been chosen as the optimal temperature. Thyrotropin, the p subunit of the


Biologically
active bovine "%thyrotropin preparations have been prepared, characterized, and used to evaluate the optimal conditions for thryotropin binding to bovine thyroid plasma membranes in uitro. Binding of lz5 I-TSH has a pH optimum around 6.0 and is sensitive to the choice and concentration of buffer. Binding is inhibited by salts, especially those containing magnesium and calcium ions; magnesium concentrations optimal for adenylate cyclase assays (2 to 5 mM) result in 85 to 98% inhibition of binding.
Binding is temperature sensitive. At 3rj" binding has its highest initial level; however, instability of the membrane at this temperature causes a rapid loss of binding activity. Binding at 0" is optimal in 30 min and at the same level as initial binding at 37"; since there is no decrease in binding activity, it has been chosen as the optimal temperature.
Thyrotropin, luteinizing hormone, the p subunit of thyrotropin, and the cy subunit of thyrotropin have relative binding affinities for the thyrotropin receptors of 100, 10, 2, and <0.5, respectively. In all of these characteristics, ""I-TSH binding as a function of hormone concentration results in curved Scatchard plots; however, Hill plots of these same binding data are linear and have a slope of 0.65. Taken together, these data suggest that the heterogeneity in thyrotropin binding constants which is evident in the Scatchard plot reflects a negatively cooperative relationship among the thyrotropin receptor sites, i.e. decreased hormonal affinity as hormone concentrations increase. Adenylate cyclase studies yield kinetic plots which also exhibit negative cooperativity; corrections for thyrotropin bound under the adverse binding conditions of the adenylate cyclase assays suggest that K, values for thyrotropin in this enzymatic assay are compatible with binding constants measured by the Y-thyrotropin preparations. Tryptic digestion destroys binding activity on the thyroid plasma membrane but releases specific thyrotropin receptor activity into the supernatant phase. Chromatography on Sephadex G-100 indicates that this solubilized receptor fragment has a molecular weight between 15,000 and 30,000.
In a previous report (1) we described the binding of bovine [3H]thyrotropin to receptors on isolated bovine thyroid membranes. The specificity of binding and its relationship to biologic function were indicated by the following observations: unlabeled TSH' was able to compete with or chase the [3H]TSH binding to the receptors whereas other hormones and albumin could not; there was a near absence of specific binding 1 The abbreviations used are: TSH, thyroid-stimulating hormone or thyrotropin; LH, luteinizing hormone. to muscle and liver plasma membranes; and [3H]TSH binding correlated with adenylate cyclase activation.
In a subsequent study (2) we showed that luteinizing hormone, a structural analog of TSH, could at high concentrations compete with TSH for the thyrotropin receptor despite a negligible or significantly lower effect on thyroid biologic activity and that the absence of significant effects on thyroid function by the cy and fi subunits of TSH could be correlated with their very low levels of binding to the TSH receptor. These studies of TSH binding to bovine thyroid plasma membranes were initiated as a consequence of our work on experimental exophthalmos. In these studies (3-lo), we had observed (a) that specific TSH receptors could be detected on plasma membranes derived from the retro-orbital tissue of the guinea pig, a mammalian model of exophthalmos (9, 10); (b) that the sera of exophthalmic patients contained a y-globulin which could induce experimental exophthalmos in fish (7,8); and (c) that this y-globulin significantly increased the in vitro binding of TSH to TSH receptors on plasma membranes of retro-orbital tissue but not to TSH receptors on thyroid plasma membranes (1,9,10). These observations suggested that specific TSH receptors not only existed in retro-orbital as well as thyroid tissue but that there was a functional or structural difference between these receptors which could account for the increased TSH binding to the retro-orbital tissue receptors in the presence of this "autoimmune" y-globulin. In the present report, we have further characterized the properties of the TSH receptor in bovine thyroid plasma membranes, and in an accompanying report (10) we have further characterized the properties of the TSH receptor on guinea pig retro-orbital tissue plasma membranes in an effort to uncover functional or structural differences which might exist. In both studies we have used both a tritiated TSH preparation which has been previously described (l-5) and an iodinated TSH preparat,ion which is characterized in this report. In contrast to the [3H]TSH which is used at much higher concentrations, the 1251-TSH has been used at hormone concentrations in vitro which could be expected in uiuo. Although we show that the binding properties of TSH receptors on thyroid and retro-orbital tissue plasma membranes are effectively the same except for the y-globulin effect, we show that a fragment of the TSH receptor released by tryptic digestion of thyroid plasma membranes is structurally different from a component of the TSH receptor which can be released by tryptic digestion of retro-orbital tissue plasma membranes (10).

MATERIALS AND METHODS
Hormone and y-Globulin Preparations-Purified bovine TSH was prepared as described in an accompanying report (10) and purified [3H]TSH was prepared and characterized as previously described (1 and used a modification of the procedure previously reported (l), this decrease in binding activity at 37" was not a function of degradation of the TSH by membrane enzymes. Instead it appeared to be a function of membrane stability since preincubation of the membranes at 37" resulted in an analogous loss of binding ability whereas preincubation of the TSH at 37" was without effect on binding (Table II). This effect of temperature was partially prevented by the presence of TSH in the incubation medium and was reversible by returning the membranes to 0" for 1 hour prior to assay (Table  II) The Y-TSH and [3HJTSH concentrations were the same as in Fig. 2   time-dependent with an optimum of 30 min (Fig. 5). At 0" and 5 x 10-l' M, "%TSH time dependence increased to over 1 hour (not shown) whereas at 5 x 10m6 M, both iz51-TSH and [3H]TSH binding at 0" was immediate (Fig. 5). Although these data are compatible with a simple second order binding reaction (A + B z AB) which exhibits time dependence as the concentration of one reactant (TSH) decreases, they in no way prove that the TSH-receptor interaction has such a simple mechanism.
Since there was no loss of binding activity at O", this temperature was chosen as the standard condition of incubation.
As was previously reported for [3H]TSH (l), binding with 'YTSH was a linear function of the concentration of membrane protein.
Also as previously reported (I), [3H]TSH binding at 0" appeared to be a linear function of hormone concentration with an apparent binding constant of 0.5 x lo* Mm' at pH 7.5 in 0.02 M Tris-chloride. At pH 6.0, in 0.025 M Tris-acetate, [3H]TSH binding at 0" yielded an apparent binding constant of 0.25 x lo* M I. In contrast, '*%TSH binding yielded a complex curve (Fig. 6) which suggested that the TSH receptor sites were heterogeneous in their ability to bind TSH. If this heterogeneity were explained by the existence of a group of "high affinity" and "low affinity" binding sites, which did not interact, i.e. a discrete group with a binding constant of 0.25 x 10"' M-' and a separate and independent group with a binding constant of 0.25 x 10" M -I as suggested in Fig. 6, inset, a Hill plot of these data should be linear and the line should have a slope of unity (32)(33)(34). A Hill plot with a value greater than unity would indicate that the receptor sites were not independent but had a positively cooperative relationship whereas a plot with a value less than unity would indicate a negatively cooperative relationship (32)(33)(34). As noted in Fig. 7, the Hill plot of these data yields a line with a slope of 0.65, i.e. the TSH receptors yield complex Scatchard plots not because of the existence of discrete groups of sites with different binding constants but because of a relationship among the sites such that the affinity for hormone It is notable, therefore, that both classes of sites have the same pH optima, the same buffer optima, the same salt inhibition phenomena, and the same temperature effects at 37". 'Y-TSH binding was a reversible process as previously reported for [3H]TSH binding (l), i.e. unlabeled TSH at a lOO,OOO-fold excess could displace '9-TSH bound during a prior incubation (Table III). There was negligible specific ""I-TSH binding to plasma membranes from either bovine skeletal muscle or bovine adrenocortical tissue (Table III). kinetics is thus performed at saturating hormone levels where effectively all of the unlabeled hormone will act as a competing agent. In the case of ""I-TSH, however, inhibition is performed under nonsaturating conditions where as much as 40 to 50% of the total hormone added to the incubation can be bound even after the addition of unlabeled hormone; accordingly, nonlinear inhibition curves will be obtained. This is evident in Fig. 8, where lz51-binding is measured in the presence of unlabeled TSH, unlabeled luteinizing hormone (LH), and the unlabeled subunits of TSH. When, however, corrections are made for the 12'I-TSH bound by considering the total of the labeled and unlabeled TSH concentration present in the assay and the percentage of this total that should be bound, a theoretic curve can be derived. As noted in Fig. 8, the theoretic curve (dashed line) derived for unlabeled TSH inhibition of lZ51-TSH binding encompassed the actual data points reasonably well.
By comparing the concentration of LH, P-TSH, or cu-TSH necessary to achieve 50% inhibition of 'Y-TSH binding, a table of their relative binding affinities could be calculated (Table IV). The similarity of these values obtained at a 1.5 x 1Om'o M concentration of lZ51-TSH and those obtained at a 5 x 10m6 M concentration of [3H]TSH (2) again suggest that, despite the different binding constants of the TSH receptors being examined under these two conditions, the characteristics of the receptors are the same.
Lissitzky et al. (37) have reported the displacement of bound 'Y-TSH from TSH receptors by the subunits of TSH and by LH. In that study they used whole porcine cells, an approximately 0.1 M medium with a pH of 7.4, and 15.min incubations at 35". Despite these differences in conditions and the major difference of whole cells as opposed to plasma membranes, the factors calculated in Table IV can be applied to their data reasonably well. Thus, the conversion of their LH or subunit concentrations to equivalent TSH concentrations using the factors of Table IV and the calculation  of the inhibition  which  this equivalent TSH concentration actually exhibited in their assay yields inhibition values very close to their experimental determinations.
As noted earlier, studies of [3H]TSH binding to plasma membranes of retro-orbital tissue have shown that -y-globulin from patients with malignant exophthalmos could increase the [3H]TSH binding (9, 10). This effect was not seen with normal y-globulin or y-globulin from patients with Graves' disease without exophthalmos (9, lo), and this effect was not seen when [3H]TSH binding was examined in thyroid plasma membranes (1). In the present study '%TSH binding to thyroid plasma membranes also was not enhanced by yglobulin from patients with malignant exophthalmos when the y-globulin was tested at concentrations which enhanced lZ51-TSH binding in retro-orbital tissue membranes (9, 10). Effect of Trypsin on TSH Receptor of Bovine Thyroid Plasma Membranes-During the course of our studies on the solubilization of the TSH receptor from thyroid membranes (38), it was noted that trypsin did not destroy binding activity as did other proteases; in contrast, trypsin did destroy the binding of TSH to receptors on intact thyroid plasma membranes (Fig. 9A). This discrepancy was resolved when binding in the supernatant phase was measured as well as binding to the plasma membranes (Fig. 9A). As noted, the loss in receptor binding activity on the plasma membranes correlates with the appearance of receptor binding activity released by trypsin into the supernatant.
When the binding activity released by trypsinization of the bovine plasma membranes was eluted on Sephadex G-100, its estimated molecular weight was 15,000 to 30,000 based on standards eluted on this same column (Fig. 9B). The trypsin solubilized receptor activity eluted from the columns was specific for 12"1-TSH binding in that results analogous to those in Table III  increase in TSH binding to retro-orbital tissue receptors but not to thyroid receptors when y-globulin from patients with malignant exophthalmos was included in the incubation medium. In our present studies we had hoped that '251-TSH binding studies would allow a more detailed characterization of the TSH receptors in thyroid and retro-orbital tissue and that this characterization would uncover biochemical differences which would either explain or amplify our previous results (1,2,(7)(8)(9). However, as noted in this and an accompanying report (101, 'Y-TSH and [3H]TSH binding studies demonstrate no significant difference between the TSH receptors in these two tissues; on the contrary, they emphasize their similarity.
Both have similar pH optima, buffer optima, salt inhibition phenomena, and temperature effects; both exhibit nonlinear Scatchard plots which can be explained by negative cooperativity among the receptor sites.
This study is nevertheless interesting in three regards: the effect of trypsin on the TSH receptor; the relationship of binding to adenylate cyclase activation; and the unusual conditions which optimize in vitro binding of TSH to the TSH receptor. In regard to the former effect these results show that trypsin can release from thyroid plasma membranes a 15,000 to 30,000 molecular weight component which has specific TSH binding activity. Work by Levey et al. (39) has indicated that receptors of such small size may not be a unique phenomena, since these workers have recovered glucagon receptor activity in similarly sized units released from cat myocardium. As is pointed out in an accompanying report (38), the low molecular weight TSH receptor component described in this report appears to be a fragment of a higher molecular weight TSH receptor which can be solubilized from thyroid plasma membranes and has slightly modified TSH binding properties. Of extreme interest is our finding in a second accompanying report (lo), that analogous trypsinization experiments with retro-orbital tissue TSH receptors release a much larger receptor fragment.
In regard to the relationship of TSH binding to TSH stimulation of adenylate cyclase activity, these studies contain several important observations. TSH binding is inhibited over 85% by 2 mM magnesium and at least 95 to 98%' inhibited by 5 mM magnesium, i.e. concentrations which have been used for optimal adenylate cyclase activation of thyroid plasma membranes (15-17, 29, 30). TSH binding is also inhibited at least 2to X-fold for Tris-chloride buffers at pH values greater than 7.4. The consequence of these observations is that only 1 to 10% of the TSH added to adenylate cyclase incubations is binding to the membranes and that K, determinations relating the effect of TSH to adenylate cyclase stimulation have therefore been seriously in error in the past and must be revised. As an example, double reciprocal plots of the data in Fig. 1 yield a nonlinear curve (analogous to the Scatchard plot in Fig. 7), from which two K, values for TSH can be calculated, one at 1 x 10m6 M and one at 1.25 x 10 *M. If, however, consideration is given to the fact that no more than #o and as little as %oo the TSH is really binding to the membranes, these values can be lowered to values very close to the binding constants derived from '%TSH binding studies in this report. This observation has been confirmed by Moore and Wolff, who have used our TSH and [3H]TSH preparations to more closely study the relationship of TSH binding to adenylate cyclase activation and the effects of nucleotides on the two processes (40).
The optimal conditions for TSH binding in uitro, i.e. a pH of 6.0 to 6.5, a buffer concentration of less than 0.05 M, and a temperature of 0", are obviously far from what are presently assumed to be in uiuo conditions. The question can be raised, therefore, as to the significance of such measurements in regard to the "real" TSH receptor in uiuo. The clearest evidence that in vitro binding studies such as these at present are measuring physiologically important receptors comes from an accompanying report (23) and from the work of Macchia and Meldolesi (41). In our own studies (23) the loss and regrowth of TSH receptors on thyroid cells exposed briefly to trypsin is paralleled by the loss and return of TSH-stimulable adenylate cyclase activity in the absence of effects on basal adenylate cyclase activity. In the work by Macchia and Meldolesi (41), TSH-stimulable functions are lost in a thyroid tumor whose plasma membranes when assayed in vitro as described herein cannot bind TSH. Thus in both studies the loss of the in vitro assayable receptor is coupled with the loss of TSH-stimulable functions in uivo.