Two Molecular Forms of the Human Interferon-? Receptor LIGAND BINDING, INTERNALIZATION, AND DOWN-REGULATION*

The receptors for human interferon-y (IFN-y) on peripheral blood monocytes and various cells of non- hematopoietic origin were thoroughly characterized and compared. The receptors of all cell types exhibited a similar affinity for IFN-y (Kd - 1 X M), and in all cases receptor-mediated endocytosis and ligand degradation were demonstrated. However, the recep- tors differed in their molecular weights (95,000 in HeLa cells and 140,000 in monocytes, assuming a 1:l ligand to receptor ratio) as concluded from experi- ments of cross-linking to lZ5I-IFN-y. Lower molecular weight species were obtained as well, particularly in monocytes. Such species could represent either degra- dation products or subunit structures. The monocyte and HeLa receptor responded differently to an excess of ligand. A significant receptor down-regulation was observed when monocytes were incubated with an ex- cess of ‘261-IFN-y, whereas no such down-regulation was observed in HeLa cells or in normal fibroblasts. This differential response was observed both in the presence or in the absence of a protein synthesis inhib-itor. The receptor on monocytes was found to be acid- labile whereas that on HeLa cells was resistant to acid treatment. These and additional experiments indicate that the monocyte receptor is inactivated following internalization, whereas the HeLa receptor retains its structure and recycles back to the cell surface. The difference in the properties and fate of these two recep- tor subtypes is probably related to the differential functions of IFN-y in various cell types.

a similar affinity for IFN-y (Kd -1 X M), and in all cases receptor-mediated endocytosis and ligand degradation were demonstrated. However, the receptors differed in their molecular weights (95,000 in HeLa cells and 140,000 in monocytes, assuming a 1:l ligand to receptor ratio) as concluded from experiments of cross-linking to lZ5I-IFN-y. Lower molecular weight species were obtained as well, particularly in monocytes. Such species could represent either degradation products or subunit structures. The monocyte and HeLa receptor responded differently to an excess of ligand. A significant receptor down-regulation was observed when monocytes were incubated with an excess of '261-IFN-y, whereas no such down-regulation was observed in HeLa cells or in normal fibroblasts. This differential response was observed both in the presence or in the absence of a protein synthesis inhibitor. The receptor on monocytes was found to be acidlabile whereas that on HeLa cells was resistant to acid treatment. These and additional experiments indicate that the monocyte receptor is inactivated following internalization, whereas the HeLa receptor retains its structure and recycles back to the cell surface. The difference in the properties and fate of these two receptor subtypes is probably related to the differential functions of IFN-y in various cell types.
Interferon-? (IFN-7)' is a lymphokine produced by activated T-lymphocytes. It exerts antiviral activity, growth inhibitory effect, and several immunoregulatory activities on a variety of cell types (1). IFN-7 elicits its various activities via a specific cell-surface receptor. The existence of a specific receptor for IFN-7 on various human and mouse cells was demonstrated in several studies (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). In all cases, specific receptors were demonstrated by high affinity binding of lZ5Ilabeled IFN--y (Kd < lo-' M ) and by specific competition with unlabeled IFN-7. The exclusivity of the receptor for 1FN"y was concluded from the finding that other types of IFN did not compete for binding to the receptor (4,6,7,10,19). Crosslinking experiments of labeled IFN--y to intact cells yielded in most studies complexes of apparent molecular weights of * This work was supported by a grant from InterYeda Ltd., Ness Ziona, Israel. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Has the Edna and Maurice Weiss Chair in Interferon Research.
We have recently presented evidence that human peripheral blood monocytes have an IFN--y receptor (IFN--yR) that is structurally and functionally different from its counterpart in cells of nonhematopoietic origin (11). In the present study, we have further characterized and compared the IFN--yR of human monocytes to the receptor from cells of nonhematopoietic origin. We demonstrate here that both forms of the IFN-yR have a similar affinity for IFN--y and in both cases the ligand is internalized and degraded. However, these receptors differ significantly with respect to their acid lability and their fate following internalization.

RESULTS
Surface Binding of IFN-7-Various cell types of both hematopoietic and nonhematopoietic origin were found to exhibit high affinity for 1'51-IFN--y. In all cases, one class of binding sites was deduced from the linear Scatchard plots of 1251-IFN-r binding (Figs. 1 and 2, insets). Dissociation constants of 1.3 X 10"' M f 7% and 1 X 10"' M k 13% were determined by competition studies with unlabeled ligands for HeLa cells and monocytes, respectively (Figs. 1 and 2). Similar values were obtained for binding to many other cell types such as foreskin fibroblasts (FSll), epithelial cell lines (WISH), and cell lines of hematopoietic origin (e.g. U-937, KG-1, and Daudi; data not shown). The number of receptors per cell was found to be 8400 f 200 and 1400 f 100 for HeLa cells and monocytes respectively (Figs. 1 and 2). In all cases, nonspecific binding was less than 1% as found by the LI-GAND program and less than 5% by using an excess of unlabeled IFN-r.
Cross-linking of 1251-IFN-y to Cells-Cross-linking of lZ5I-IFN-r to various cells revealed the presence of different molecular forms of IFN--yR. In HeLa cells, a complex of molecular weight 120,000 was seen in autoradiographs of SDS-polyacrylamide gel electrophoresis. Additional minor bands corresponding to molecular weights of 90,000 and 70,000 were also seen (Fig. 3). A similar pattern of bands was obtained in other cells of nonhematopoietic origin such as FSll and WISH. Cross-linking of 1251-IFN-7 to monocytes gave a faint band corresponding to a molecular weight of 165,000 as well as intense broad bands of molecular weights 60,000-80,000 and 90,000 (Fig. 3). The 120,000 and the 165,000 bands were mutually exclusive in the various cell of an excess of unlabeled IFN-y (2 X 10" units). Following crosslinking with disuccinimidyl suberate, the cells were washed and solubilized. '"I-IFN-y and its cross-linked products were immunoprecipitated by rabbit anti-IFN-y and protein-A-agarose. The agarose beads were suspended in a sample buffer, and the supernatant was analyzed by SDS-polyacrylamide gel electrophoresis (7.5%) followed by autoradiography. Lane a, "C-labeled molecular weight markers (indicated on the left in thousands); lane f, '2sII-IFN-y cross-linked to itself in the absence of cells.
types. These bands were abolished when the cross-linking experiments were performed in the presence of an excess of unlabeled IFN-y. A control cross-linking experiment of ' "1-IFN-y to itself revealed a major band of M, 45,000, corresponding to IFN-y dimer and two very faint bands corresponding to molecular weights 73,000 and 90,000. (It should be emphasized at this point that protease inhibitors were present in all extraction buffers.) Internalization and Degradation of 'Z'I-IFN-y-Internalization of "'I-IFN-y was studied in both HeLa cells and monocytes. The cells were incubated with an excess of ""I-IFN-y a t 37 "C for different time periods, surface-bound IFNy was removed a t low pH, and the level of internalized ' "1-IFN-y was measured (Fig. 4). A rapid initial rate of internalization corresponding to 7000 molecules/cell/h in HeLa cells and 2000 molecules/cell/h in monocytes was observed in the first 30 min. Thereafter, the level of internalized 12"I-IFN-y continued to rise for another 3 h in HeLa cells and for 2 h in monocytes, but at a reduced rate. The amount of internalized 1251-IFN-y in monocytes declined during the third hour. When the experiment was repeated in the presence of chloroquine, a much larger accumulation of internalized 12'I-IFN-y was obtained both in HeLa cells and in monocytes (Fig. 4). The rate of IFN-y accumulation in HeLa cells was found to be linear during the entire incubation period (3 h) in the presence of an excess of ligand. In monocytes, the rate of accumulation decreased after the first 2 h.
The fate of cell-associated IFN-y was then studied both in HeLa cells and in monocytes. Both cell types were preincubated a t 4 "C for 2 h with an excess of ""I-IFN-y, washed, and further incubated in the absence of IFN-y for different time periods a t 37 "C. In both cell types, a rapid internalization of the surface-bound '*"I-IFN-y was observed (Fig. 5).
The level of surface-bound IFN-y decreased by 50% within 45 min in HeLa cells and within 1 h in monocytes. The level of internalized '"I-IFN-y rose rapidly during the first 30 min. However, in the absence of free '"I-IFN-y, the rate of internalization was reduced on further incubation. This reduction coincided with the disappearance of surface-bound 'TI-IFNy. The level of internally accumulated IFN-y started to decline after 2 h due to a process of degradation followed by excretion of the degradation products which was apparent after a lag of 30 min. The efficiency of the degradation process was demonstrated by an additional experiment in which the cells were preincubated at 37 "C with an excess of '2'I-IFN--y, washed, and further incubated a t 37 "C. Under these conditions, extensive internalization was allowed before excess free IFN-y was washed away. Thereafter, most of the cell-associated '2sI-IFN-y was released within 3 h as degradation products, and its level dropped to 5% and 20% of the initial values in HeLa cells and in monocytes, respectively (Fig. 6). Some of the cell-associated ""I-IFN-y was released intact as revealed by trichloroacetic acid precipitation. This release was due to the dilution effect, it corresponded to less than 15% of the total radioactivity, and it did not increase after the first 15 min of incubation (not shown). In conclusion, both cell types were capable of internalizing and degrading IFN-y with a similar efficiency.
Down-regulation of the Receptors-The level of surfacebound '2'I-IFN-y was determined after incubation of cells with an excess of '251-IFN-y for different time periods a t 37 "C. The level of surface-bound "'I-IFN-y in HeLa cells was not reduced during the entire incubation period (3 h, Fig.  7). On longer incubations (more than 6 h, data not shown), a reduction in the level of surface binding was noticed, but it was attributed to depletion of lZ'I-IFN-y from the medium. Indeed, when the medium was replaced by a fresh medium containing '2sI-IFN-y at the initial concentration, the original level of surface binding was restored. Similar results were observed in experiments with WISH cells and with fibroblasts (not shown). In contrast, the level of surface-bound '*'I-IFNy on monocytes was gradually reduced and reached 50% of  its initial value in less than 3 h. Thus, it was demonstrated that the IFN-y receptor is down-regulated by an excess of IFN-y in monocytes, but not in cells of nonhematopoietic origin.
The IFN-yR half-life was determined by incubating the cells with cycloheximide at 37 "C followed by saturation binding of lZ5I-IFN-y at 4 "C. In both HeLa cells and monocytes, the half-life was found to be about 3 h. A different response of those two types of cells was observed when the cells were incubated with an excess of Iz5I-IFN-y together with cycloheximide. In HeLa cells, excess 1251-IFN-y did not decrease the level of surface binding below the reduction obtained by cycloheximide alone. In contrast, excess lz5I-IFN-y decreased the level of the receptor in monocytes even below the reduction obtained by cycloheximide alone (Fig. 7). In other words, the effects of cycloheximide and excess lZ5I-IFN-y on the receptor level in monocytes were additive.
Acid Stability of the Receptors-Incubation of HeLa cells at pH 4.5 for 10 min at room temperature had no effect on the affinity or the number of IFN-7 receptors per cell. In contrast, a similar treatment of monocytes caused over 50% reduction in the number of available surface receptors, whereas the affinity of the remaining receptors did not change significantly (Table I).

DISCUSSION
IFN-y is a lymphokine which plays a central role in immune response (1). It acts on cells via a specific cell-surface receptor found to be present on many cell types including cells of nonhematopoietic origin (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). IFN-y has different activities on different cell types. For instance, its antiviral activity is observed on many cells of nonhematopoietic origin but not on monocytes (11). In the present study, we have characterized thoroughly the receptor of IFN-y in two cell types: HeLa, which represent cells of nonhematopoietic origin and peripheral blood monocytes of the hematopoietic lineage. Ligand binding, internalization, and degradation as well as receptor down-regulation, acid-lability, and apparent molecular weight were studied. These studies led us to conclude that there are two types of IFN-y receptors which share several properties such as binding affinity, internalization, and degradation of IFN-7. However, the two receptor types differ in their molecular weight, stability in acid, and in their regulation. Such variability may be the basis for the different activities of IFNy on different cell types.
The two molecular forms of the IFN-yR reside in different cells and are mutually exclusive. However, both receptors exhibit a similar affinity for IFN-y. We had reported that the binding of IFN-y to nonhematopoietic cells (WISH) yielded a linear scatchard plot, whereas a biphasic plot was obtained with monocytes (11). A later analysis of the same binding data by the LIGAND program indicated that there was only one class of high affinity binding sites for IFN-y in monocytes. The low affinity phase in the plot was therefore due to nonspecific binding. In this study, glycosylated IFN-y which exhibited a high specific activity (5 X 107-10R units/mg) was used for the measurement of Kd. Other studies were performed with recombinant IFN-y produced in Escherichia coli and this IFN exhibited a somewhat lower specific activity as compared to that of natural IFN-y (107-3 x lo7 units/mg, e.g. Ref. 10).
Calculation of the binding data employing glycosylated IFNy revealed that 50% saturation of the receptors occurred at 200-400 units/ml, whereas half-maximal biological effects were observed at 1 unit/ml. Thus, there is a considerable amount of spare receptors on the cell surface. Based on the specific activity and the Kd we calculated that at 1 unit/ml of IFN-y, one cell will bind 10-15 molecules of IFN-y and this quantity is sufficient for exerting half-maximal biological effect. It is therefore conceivable that signal transduction into the cell is accomplished by signal amplification.
Cross-linking experiments provided evidence for the existence of two receptor subtypes. The observed molecular weight of 120,000 for the 1251-IFN-y-receptor complex in cells of nonhematopoietic origin was in accordance with previously reported values (6,7,12). The molecular weight of the receptor was calculated as 95,000 assuming a 1:l ligand to receptor ratio in the complex. A major protein band of this molecular weight was obtained recently by ligand affinity chromatography of the IFN-yR from fibroblasts (21). Similarly, the 165,000 complex from monocytes corresponded to a molecular weight of 140,000 for the monocyte receptor. However the majority of cross-linked complex in monocyte had a molecular weight of 60,000-80,000 and 90,000. Such bands were also seen in preparations from HeLa cells but in a much lower intensity. These lower molecular weight complexes could represent degradation products or, as recently suggested, subunit structures (18). Interestingly, Ucer et al. (18) did not report the presence of the 120,000 complex in any cell type. In contrast, a complex of molecular weight 120,000, similar to that of nonhematopoietic cells was obtained by cross-linking of 1z51-IFN-y to U937 cells (9), which are a monocyte-like cell line. Such a discrepancy between U937 cells and peripheral blood monocytes may be related to differences in their state of differentiation. Since the 120,000 and the 165,000 complexes were mutually exclusive in different cell types, it is not likely that they represent different levels of degradation of the same receptor molecule. However, preliminary studies3 indicate that the receptors on these two cells are immunologically cross-reactive. The extent of the structural homology between the two subtypes will be determined by sequencing the receptors or their gene(s).
In terms of post-binding events, both receptors behaved as typical polypeptide hormone receptors (22). Receptor-mediated endocytosis followed by ligand degradation and release of degradation products were demonstrated in both cell types in accordance with previous studies (3,19). An unusual effect of chloroquine was, however, observed. Chloroquine was found to prevent the degradation of various ligands by increasing lysosomal pH, thereby causing an accumulation of intact ligands within the cell. However, the elevation of pH by   chloroquine should also prevent the uncoupling of ligandreceptor complexes in the endosomes, thereby preventing the receptors from recycling to the cell surface and subsequently leading to a reduced rate of internalization. In our studies, chloroquine did not decrease the rate of lZ51-IFN--, uptake by either HeLa cells or monocytes. Instead, it prolonged the linear phase of lZ51-IFN--, uptake. In the case of HeLa cells, this linear uptake persisted for more than 4 h. It seems, therefore, that fibroblasts have alternative ways for restoring the level of IFN--, receptors on cell surface. A similar phenomenon was observed in some other ligand-receptor system (e.g. Ref. 24). The two receptor types differed most significantly when tested for receptor down-regulation by an excess of the specific ligand. Such down-regulation was obtained in the case of monocytes, but not in the case of HeLa cells. Consequently, HeLa cells could internalize IFN--, at high rates over long time periods, whereas this process in monocytes was limited. When protein synthesis was blocked by cycloheximide, incubation with an excess of ligand did not cause an additional decrease in the level of receptors in HeLa cells. It was therefore concluded that the steady level of the receptor obtained in the presence of an excess of ligand was not due to an increase in the rate of receptor biosynthesis. Instead, it is proposed that the receptor in HeLa cells acts as a "shuttle" which brings in its ligand, releases it, and then recycles intact to the cell surface. In contrast, cycloheximide and an excess of IFN--, had an additive effect on the extent of reduction of IFN--, surface receptors in monocytes. Therefore, it was concluded that the IFN--, receptor is not recycled in monocytes. Down-regulation of IFN--, receptors was previously reported in HeLa cells (7). However, the cells were preincubated with excess unlabeled IFN--, before the labeled ligand was added. Therefore, in that case a distinction between real downregulation and receptor preoccupancy by residual unlabeled ligand was not possible.
Receptors that recycle should resist the low endosomal pH (22). Indeed, we demonstrated that the receptor in HeLa cells is acid-resistant, whereas the monocyte receptor is rapidly inactivated at low pH (Table I). It appears therefore that following internalization the monocyte receptor is inactivated by the low endosomal pH even before digestion by lysosomal proteases. In contrast, the receptor in HeLa cells is resistant to the low endosomal pH. The properties of the two receptors are summarized in Table 11.
-1FN"y exhibits antiviral activity in nonhematopoietic cells such as fibroblasts, WISH, and HeLa, but not in monocytes (11). Similarly, IFN--, induces ( 2 ' 4 ' ) oligo A synthetase in the nonhematopoietic cells, but not in monocyte^.^ In this study, we demonstrated that the receptors of IFN--, in monocytes and in cells of nonhematopoietic origin differ in their structure and in their stability. These differences determine the fate of the receptor molecules following internalization and thereby the ability of the cells to internalize IFN--, over long time periods. It is possible that these differences correlate with the differential responses to IFN--, as observed in cells of hematopoietic and nonhematopoietic origins. combined and their radioactivity counted. The amount of radioactivity in t h e fractions wbs taken m surface bound 1251-lFN-7. The eel15 were then detached with trypsin and the eeil -dated radioactivity was eounkd. These counts were t&en m the internalized ligand. Non-specific binding and internalization WI I C dclermlned by repealing the sxperi-men1 in the presence of an e x r s (1wO fold) or unlabeled IFN-7.
The effect of ehlomquine on the internalization of 1251.1FN-7 was measured m fol-io%: mnfluent monolayem or adherent cella wore preineubated with rhloiaqvine (200 UM) lor 30 min at 37% lZ5l-IFN-7 ( S m uniklml) was added (final volume pl) sad incubation s t 37% was continued. A1 the indicated times the cells were washed with cold PBS and wilh dilute acetic acid ao described. detached with trypsin and counted. T h e counts were considered as internalized 1FN.7 Determination or murlaee bound, internalized and degraded
Four tubes were used for each time point.
N t e r incubation the tubas were *pun ( l a g , 5 mi". 4%) and the e x e s or 1251-IFN-7 wss aspirated. The eel1 pellets were washed

The cuiturcs
were then emlrd to 4%. washed with wid PBS (3x300 PI). and surface bound 1251.~N-7 was determined as dncribcd prcviousiy. Measurement of receptor halfcycloheximide (35 OM) was sddad at different time points. A1 the end of the ioeubation life wm done as lollown: Cell culture8 ware ioeubatcd lor 3 hr at 37% during which the CUI~YIF. were cmlrd lo I%, the medium was rcplsccd by a medium containing 1251-1FN-7 (500 units/ml). the cuIturea were iclt at 4% lor 2.5 hr and then pmelsed lor the determination of surface-bound 1251-1FN-7 as previously described.