Identification of two regions within the cytoplasmic domain of the human interferon-gamma receptor required for function.

Functionally active human interferon-gamma (IFN gamma) receptors require the presence of at least two polypeptides: the IFN gamma receptor and an accessory molecule encoded by a gene on human chromosome 21. Here we have used a murine L cell line that stably contains human chromosome 21 (SCC16-5) to determine whether the receptor's cytoplasmic domain is important for receptor function. SCC16-5 stably transfected with the full-length human IFN gamma receptor cDNA bound, internalized, and responded to human IFN gamma. In contrast, SCC16-5 expressing human IFN gamma receptors lacking a cytoplasmic domain bound human IFN gamma but did not internalize or respond to it. Using a family of IFN gamma receptor deletion mutants, two functionally important regions within the intracellular domain were identified: (a) a membrane proximal region (residues 256-303) required for ligand processing and biologic responsiveness and (b) the carboxyl-terminal 39 amino acids (residues 434-472) needed exclusively for biologic responses.


Identification of Two Regions within the Cytoplasmic Domain of the Human Interferon-y Receptor Required for Function*
Functionally active human interferon-y (IFNy) receptors require the presence of at least two polypeptides: the IFNy receptor and an accessory molecule encoded by a gene on human chromosome 21. Here we have used a murine L cell line that stably contains human chromosome 2 1 (SCC16-6) to determine whether the receptor's cytoplasmic domain is important for receptor function. SCC16-5 stably transfected with the full-length human IFNy receptor cDNA bound, internalized, and responded to human IFNy. In contrast, SCC16-5 expressing human IFNy receptors lacking a cytoplasmic domain bound human IFNy but did not internalize or respond to it. Using a family of IFNy receptor deletion mutants, two functionally important regions within the intracellular domain were identified: (a) a membrane proximal region (residues 266-303) required for ligand processing and biologic responsiveness and ( b ) the carboxyl-terminal 39 amino acids (residues 434-472) needed exclusively for biologic responses.
Interferon-y (IFNy)' is a cytokine produced by T cells and natural killer cells that plays an important role in regulating host defense and immunopathologic processes (1, 2). IFNy mediates its pleotropic effects on cells following interaction with a specific receptor expressed at the target cell surface (3). During the past 2 years a considerable amount of information has been obtained that has defined the IFNy receptor at the molecular level. Human and murine IFNy receptors have been purified to homogeneity (4-6), their cDNAs cloned and expressed (7-12), and the proteins characterized as 85-95-kDa single chain polypeptides that display a modest amount of molecular weight heterogeneity due to cell specific differences in glycosylation. Human and murine receptors * This work was supported by United States Public Health Service Grants CA43059 and A124854 and a grant from Genentech. 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.
3 To whom correspondence should be addressed Dept. of Pathology, Box 8118, Washington University School of Medicine, 660 S. Euclid Ave., St. Louis, MO 63110.
The abbreviations used are: IFNy, interferon-y; rHuIFNy, recombinant human interferon-gamma; rMuIFNy, recombinant murine interferon-gamma; IFNa, interferon alpha; DMEM, Dulbecco's modified Eagle medium; FCS, fetal calf serum; PCR, polymerase chain reaction; PBS, phosphate buffered saline; rHuIFNa, recombinant human interferon-alpha; IU, international reference units; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; MHC, major histocompatibility complex; IL, interleukin; hgR, human IFNy receptor; bp, base pair(s). share a similar molecular organization but display only 52% homology at the amino acid sequence level and bind their respective ligands in a strictly species specific manner. They consist of 472 (human) and 450 (murine) amino acids and contain a single 23-amino acid transmembrane domain located in the approximate middle of the molecule. The extracellular domains of the human and murine proteins are 50% homologous and are each composed of 228 amino acids. Both proteins contain large intracellular domains (221 (human) and 200 (murine) amino acids) which display only 55% sequence homology. Interestingly, the intracellular domains of the two proteins contain an unusually high percentage (25%) of serine and threonine residues which are more conserved than the other amino acids within this molecular region.
Recently, Jung et al. (13) used murine-human somatic cell hybrids to show that expression of functionally active human IFNy receptors in rodent cells required the presence of at least two human gene products. The first was the IFNy receptor encoded by a gene on human chromosome 6. All hybrids that contained human chromosome 6 bound human IFNy. The second was encoded by a gene on human chromosome 21. Hybrids that contained only human chromosome 21 did not bind human IFNy, but hybrids that contained both human chromosomes 6 and 21 responded to the human ligand. This concept was subsequently confirmed by transfection experiments. Murine cells transfected with the human IFNy receptor cDNA bound human IFNy but did not respond to it (7). In contrast, expression of the human receptor in rodent cells which contained a copy of human chromosome 21 produced cells that bound and responded to human IFNy (14, 15). Thus the current concept is that functionally active human IFNy receptors require the presence of at least two human proteins, the IFNy binding protein and an as yet undefined accessory molecule encoded by a gene located on human chromosome 21.
The structural analysis of the IFNy receptor has shown that it is unusual since it contains a relatively large, serine/ threonine-rich intracellular domain, yet shows no significant homology to any known protein with kinase or phosphatase activity. This observation has led to the speculation that the receptor's intracellular domain may be important in effecting IFNy-dependent cellular responses. This hypothesis has been strengthened by the recent reports that the IFNy receptor in intact human cells is phosphorylated on serine and threonine residues following interaction with IFNy (16,17). The degree of receptor phosphorylation tightly correlated with the magnitude of the biologic response induced. In addition, other studies have shown that the IFNy receptor directs the internalization and degradation of IFNy. However, no data is currently available that directly documents a role of the receptor's intracellular domain in mediating cellular responses to ligand.
To examine this issue in more detail we conducted a struc-ture/function analysis of the intracellular domain of the human IFNy receptor. In the current report we demonstrate that the intracellular domain of the receptor is absolutely required for internalization and degradation of ligand and for induction of IFNy-dependent biologic responses. Moreover, through the use of receptor deletion mutants we have identified two specific regions within the intracellular domain that are differentially required for internalization of ligand and/or induction of cellular responses.

EXPERIMENTAL PROCEDURES
Reagents-Purified recombinant human and murine IFNy were generously provided by Dr. Susan Kramer of Genentech, Inc. (South San Francisco, CA). The rHuIFNy and rMuIFNy used in this study displayed specific antiviral activities of 1.9 X lo7 and 5 X lo6 IU/mg, respectively. Purified rHuIFNy and rMu1FN-y were radioiodinated, using Bolton-Hunter reagent (ICN Chemicals, Radioisotope Division, Irvine, CA) to specific activities of 6.5-17.8 pCi/pg as previously described (24). Purified forms of recombinant human IFNa2. (specific activity = 3 X 10' IU/mg) and recombinant 1FNaA.D (specific activity = 1.1 X 1 0 ' IU/mg) were generously provided by Dr. Gianni Garotta, Hoffmann-LaRoche AG (Basel, Switzerland). Recombinant human IFNaz (specific activity = 1.7 X 10' IU/mg) was generously provided by Dr. Marvin Siege1 (Schering-Plough, Bloomfield, NJ). IFNa2 and IFNorzs are species specific forms of human IFNor, whereas 1FNaA-D induces biologic responses in both human and murine cells. G418 was obtained from Sigma.
Antibody Reagents-GIR-208 is a murine IgGl monoclonal antibody specific for the human IFNy receptor. It reacts with an epitope that is identical with or closely linked to the ligand binding site of the receptor. The antibody was purified from culture supernatants by affinity chromatography on protein A-Sepharose and biotinylated using the ENZO biotinylating reagent (ENZO Biochemicals) as described previously (18). Goat anti-murine Ig-Sepharose was prepared as previously described (19). The Sepharose preparation used in these experiments contained 7.8 mg of antibody/ml of gel. GR-20, a rat IgG,. monoclonal antibody specific for the murine IFNy receptor (6) was kindly provided by Dr. Stephen Russell (University of Kansas Medical Center, Kansas City, KS) and was purified by affinity chromatography on protein G-Sepharose (Genzyme Corp., Boston, MA). Monoclonal antibody 11-4.1 recognizes murine H-2Kk and was purified from spent hybridoma culture supernatants by protein A-Sepharose chromatography. Cells were obtained from ATCC (Rockville, MD).
Cells and Cell Culture-L929, a murine fibroblast cell line, was obtained from ATCC (Rockville, MD). SCC16-5 is a murine L cell line that contains a single copy of human chromosome 21 (20). The line was generously provided by Dr. David Cox (University of California, San Francisco). Both cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 2% L-glutamine, 1 mM sodium pyruvate, 50 units/ml penicillin, and 50 pg/ml streptomycin.
Primers-The primers listed below were synthesized on a Pharmacia gene assembler. All primers were based on the nucleotide sequence of the human IFNy receptor cDNA.*

4 5 ' C T C A C G C A G A A G G A A G A T G A T 3 ' 5 0 5 4 5 ' A T T C A T C C G C T T A T T A T C A C 3 '
J. Calderon, P. W. Gray, J. T. Pingel, W. Henzel, and R. D.
Primer 673 is complementary to the noncoding strand at positions -28 to -7 relative to the ATG start site. Primer 723 is complementary to the coding strand at positions 800-830 and was modified such that a stop codon was introduced at positions 817-819. Primer 805 is complementary to the noncoding strand at positions -28 to -7 and has been engineered to contain a restriction site for ClaI. Primer 8202 is complementary to the coding strand at positions 1332-1352 which was modified to introduce a stop codon at positions 1350-1352 followed by a ClaI restriction site. Primer 817 is a hybrid primer whose 5' half is composed of 15 nucleotides complementary to the noncoding strand at positions 801-815, whereas its 3' half is composed of 15 nucleotides complementary to the noncoding strand at positions 960-974. Primer 818 spans the same region as primer 817 but is complementary to the coding strand. Primer 829 is a hybrid primer whose 5' half is composed of 15 nucleotides complementary to the noncoding strand at positions 945-959, whereas its 3' half is composed of 15 nucleotides complementary to the noncoding strand at positions 1188-1202. Primer 830 spans the same region as primer 829 but is complementary to the coding strand. Primer 833 is a hybrid primer whose 5' half is composed of 15 nucleotides complementary to the noncoding strand at positions 1173-1187 whereas its 3' half is composed of 15 nucleotides complementary to the noncoding strand at positions 1350-1364. Primer 834 spans the same region as primer 833 but is complementary to the coding strand. Primers 949 and 5054 are primers used to amplify the transformed human IFNr receptor cDNAs by PCR from cell lysates of the transfected cells. Primer 949 is complementary to the noncoding strand sequence at positions 553-572. Primer 5054 is a primer derived from the pSFFV vector sequence itself located 50 nucleotides 3' of the cloning site.
Plasmid Construction-The cDNA for the human IFNy receptor (hgR) was cloned from a human placental library prepared in X g t l O and was originally inserted into the expression vector pRK5.' For these studies, the cDNA was cloned into the plasmid vector pSFFV, which contains the neomycin resistance gene (21), since we have found that it directs more consistent expression of the human IFNy receptor in murine cells. The full-length hgR cDNA was excised from pRK5.hgR by digesting with XbaI and SfiI. The 2.1-kilobase pair cDNA insert was treated with T4 DNA polymerase to blunt the ends and then methylated with EcoRI methylase. EcoRI linkers (pGGAATTCC, New England Biolabs, Inc.) were then added and the resulting DNA fragment digested with EcoRI and inserted into the EcoRI cloning site of pSFFV to generate the plasmid pSFFV.hgR.
The hgR deletion constructs used in this study were created through the use of the polymerase chain reaction with pSFFV.hgR serving as the template. The plasmid pSFFV.hgRAIC was made using primers 673 and 723. The PCR reaction was carried out using a Perkin-Elmer Cetus DNA thermal cycler and the Perkin-Elmer Cetus GeneAmp reagent kit. Fifty nanograms of pSFFV.hgR was added to a reaction mixture containing 200 mM dNTPs, 1 X PCR buffer, 1 mM primer 673, 1 mM primer 723, and 2.5 units of AmpliTaq polymerase in a total volume of 100 pl. The mixture was heated to 94 "C for 5 min to completely denature the plasmid DNA and then subjected to 30 cycles of amplification using the following protocol: (a) 1 min at 94 "C, ( b ) 1 min at 59 "C, and ( c ) 2 min at 72 "C (standardprotocol). The amplified fragment was purified by electrophoresis on a 1% low melting temperature agarose gel and extracted from the agarose with acid phenol. The purified PCR fragment was treated with T4 polymerase to blunt the ends. ClaI linkers were added, the resulting fragment digested with ClaI, and then inserted into pSFFV, which had been modified by placing a ClaI linker (pCCCATCGATGGG, New England Biolabs, Inc.) into the EcoRI cloning site of the vector.
The plasmid pSFFV.hgRA434-472 was made using primers 805 and 8202. Twenty cycles of amplification were utilized following the standard protocol. Following amplification, the fragment was purified by agarose gel electrophoresis as above and digested with ClaI since the primers used had been engineered to contain this restriction site. The digested fragment was inserted into pSFFV.
The deletion constructs pSFFV.hgRA256-303, pSFFV.hgRA304-379, and pSFFV.hgRA380-433 were made using a two-step PCR procedure. The plasmid pSFFV.hgRA256-303 was constructed using primers 805 and 818 to generate the PCR fragment 5' of the desired deletion. In a separate reaction, primers 817 and 809 were used to generate the PCR fragment 3' of the desired deletion. Twenty-five cycles of amplification following the standard protocol were used to produce each fragment. After purification by gel electrophoresis, small aliquots of the 5' and 3' fragments were combined with 200 mM dNTPs, 1 X PCR buffer, and 2.5 units of AmpliTaq polymerase, Since the 5' and 3' fragments share a 30-base pair overlap which bridges the desired deletion, each strand can act as a primer for the other. Three cycles of PCR were carried out using the standard protocol, thereby generating the full-length construct containing the desired deletion. Primers 805 and 809 were then added, and an additional 15 cycles were carried out following the standard protocol to amplify the deletion construct. The PCR fragment was then inserted into the ClaI site of pSFFV. The plasmids pSFFV.hgRA304-379 and pSFFV.hgRA380-433 were generated following the same procedure but substituting primers 830 and 829 and primers 834 and 833, respectively, for primers 818 and 817. (22).
The sequence of each construct was verified by dideoxy sequencing DNA Transfection-SCC16-5 and L929 cells (5 X 1 0 ' ) were transfected with 25 pg of plasmid using the calcium phosphate precipitation method (23). Selection was begun 48 h after transfection using G418 at a final concentration of 1 mg/ml active ingredient. After selection was completed, cells were sorted based on their expression of the human IFNr receptor (see below) and maintained in culture in the presence of 0.5 mg/ml active G418.

Demonstration of Human and Murine ZFNr Receptors on Trans-
fected Murine CeUs-Adherent cell cultures were washed with 0.02% EDTA solution, and cells were harvested by gentle scraping using a Costar cell scraper. Cells were washed once with the appropriate buffer and resuspended based on the type of assay to be run.
For flow cytometry, 1 X lo6 cells were suspended in 200 p1 of phosphate buffered physiologic saline, pH 7.4 (PBS), supplemented with 10% heat-inactivated fetal calf serum (PBS-FCS), 1% normal mouse serum in the presence or absence of 10 pg of rHuIFNr, and incubated for 60 min at 4 "C. Five pg of biotinylated GIR-208 (antihuman IFNr receptor) or GR-20 (anti-murine IFNr receptor) were added to each sample and the incubation continued for 1 h. Cells were washed with PBS lacking serum and resuspended in 200 pl of streptavidin-phycoerythrin conjugate (Chromoprobe, Redwood City, CA) diluted 1:lO in PBS. After incubation for 30 min at 4 "C, cells were washed, resuspended in 400 pl of PBS, and analyzed on a Becton Dickinson FACScan (Becton Dickinson, Braintree, MA) equipped with a single 15-milliwatt argon laser. Cells pretreated with human or murine IFNr (which thereby blocks the epitopes recognized by either GIR-208 or GR-20, respectively) were used as the negative control. Human IFNr receptor positive cells were sorted by selecting the brightest 10% viable cells in the sample on a Coulter EPICS 753 (Coulter Electronics, Hialeah, FL) equipped with a single coherent 5watt argon laser.
Receptor expression on transfected cells was quantitated by radioligand binding analysis performed with '%I-labeled rHuIFNr using 2-5 X 10' cells in suspension in a total volume of 100 p1 as described (24). Binding data was analyzed by the method of Scatchard (25).
Analysis of Transfected Murine Cells for Responsiveness to Human ZFNr-The ability of human IFNr to enhance MHC Class I antigen expression on transfected murine fibroblasts was examined by culturing 5 X 1 0 ' cells in 100-mm petri dishes for 72 h in 10 ml of either medium alone or rHuIFNr, rMuIFNr, rHuIFNaa, rHuIFNap, or rHuIFNaA-D at a final concentration of 1000 IU/ml. MHC Class I expression was assessed by flow cytometry on cells stained sequentially with the 11-4.1 H2-Kk-specific monoclonal antibody in the presence of 1% normal goat serum and fluorescein isothiocyanatelabeled goat anti-murine Ig (Chromoprobe, Redwood City, CAI. Surface Labeling and Immunoprecipitation of Human ZFNr Receptors-Cells were harvested from culture by scraping as described above, and 1 X IO7 cells were washed in PBS and resuspended in 1 ml of PBS containing 50 mM D-glucose. Cells were incubated with 200 pg of lactoperoxidase, 2.5 pg of glucose oxidase, and 1 mCi of NalZ61 for 45 min at room temperature. The reaction was stopped by addition of 10 ml of ice-cold DMEM containing 10 mM NaN3. Cells were washed three times with ice-cold DMEM, 10 mM NaN3, lysed by addition of 0.5 ml of lysis buffer, and labeled receptor analyzed by SDS-PAGE following immunoprecipitation as described (19).
Internalization and Degradation of ZFNy-These studies were performed in 6-well tissue culture plates (Costar, Cambridge, MA) that had been preincubated overnight with DMEM containing 10% FCS (D-10) so as to reduce nonspecific binding of ligand. Cells were grown in the wells to confluency and then incubated with saturating doses (5-30 ng) of '261-rHuIFNy at 4 "C for 1 h to allow binding of ligand to the receptor. The cell monolayers were then rapidly washed with ice-cold DMEM containing 2% FCS to remove unbound IFNr. One ml of D-10 was added to each well and the plates placed at 37 "C. After various periods of time at 37 "C, the cell supernatants were harvested, subjected to precipitation with 10% (final concentration) trichloroacetic acid, and the trichloroacetic acid-soluble counts quantitated to determine the amount of degraded 1FN-y. Cell surface bound radioactivity was determined by treating duplicate samples with either 1 ml of D-10 or acid (0.25 M acetic acid, 0.125 M NaC1, pH 2.7) for 5 min at 4 "C. Supernatants were collected and internalized counts determined by lysing the cell monolayer with 1% Nonidet P-40. Total internalized counts were then determined as the sum of internalized plus degraded (trichloroacetic acid-soluble) counts. Nonspecific binding was determined by preincubating selected cell monolayers with 20 pg of unlabeled human IFNr. Cells treated in this manner showed no time-dependent increase of trichloroacetic acid-soluble counts.
To determine cellular degradation of internalized ligand, cell monolayers were pretreated with saturating doses of 'Z61-rHuIFNy at 4 "C for 1 h and then washed to remove unbound counts. Cell monolayers were incubated at 37 "C for various periods of time, culture supernatants harvested, and the level of trichloroacetic acid-soluble counts determined as previously described (24). Calculation of percent trichloroacetic acid-soluble counts was determined by dividing the total trichloroacetic acid-soluble counts in a sample with the sum of the trichloroacetic-soluble counts plus the cell-associated counts (surface receptor bound counts plus internalized counts).
Amplification by PCR of Human ZFNr Receptor cDNAs Obtained from Transfected Cell Lysates-Five million cells were washed once in PBS and then lysed in 1 ml of PCR buffer (10 mM Tris-HC1, pH 8.3, containing 50 mM KCl, 2.5 mM MgC12, 0.1 mg/ml gelatin, 0.45% Nonidet P-40, and 0.45% Tween 20). Sixty micrograms of proteinase K (BRL, Gaithersburg, MD) was then added and the cell lysates heated to 56 "C for 1 h. The proteinase K was inactivated by heating the lysates to 95 "C for 10 min and the lysates stored at 4 "C before use. For amplification, 17 pl of the cell lysate was brought up to 50 pl in PCR buffer. The desired primers (1 mM) were added along with a final concentration of 200 mM dNTPs, 1 X PCR buffer, and 2.5 units of AmpliTaq polymerase and the reaction mixture brought up to a final volume of 100 pl with water. Thirty rounds of amplification were carried out as follows: ( a ) 30 s at 94 "C, (b) 30 s at 59 "C, and (c) 1 min at 72 "C. PCR products were analyzed on a 1% agarose gel MD).
run along with a 123-base pair DNA ladder (BRL, Gaithersburg,

RESULTS
Characterization of a Murine Cell Line Containing Human Chromosome 21:SCC16-5-Previous studies have established that formation of a functionally active human IFNy receptor in murine cells requires the expression of both the human receptor protein and a second human component encoded by a gene on human chromosome 21 (13, 15). Fig. 1 shows the characteristics of the murine L cell line, denoted used in the current study, that stably contains a single copy of human chromosome 21 (20). As evidenced by flow cytometry using receptor-specific monoclonal antibodies, SCC16-5 expresses only its endogenous murine receptor and not human IFNy receptors (Fig. 1, panels a and b). This result was confirmed by quantitating IFNy receptor expression on the cells using radioligand binding techniques. Whereas SCC16-5 specifically bound 3700 molecules of murine IFNylcell, no specific binding of the human ligand was detected (data not shown).
SCC16-5 also displays the appropriate responsiveness to different forms of interferon. In these experiments, cellular responsiveness was assessed by monitoring expression of MHC Class I antigens on cells. SCC16-5 constitutively expressed MHC Class I antigens (Fig. 1, panels c-j). As expected, murine IFNy enhanced (a 60-channel increase) MHC Class I expression on these cells (Fig. 1, panel c), whereas human IFNy was completely inactive (Fig. 1, p a n e l d ) . Thus, on the basis of receptor expression and IFNy responsiveness, SCC16-5 behaved like normal murine L cells. In order to confirm that the SCC16-5 line maintained in our laboratory contained human chromosome 21 (and thus presumably expressed the human accessory molecule) we took advantage of two experimentally established findings. (1) The gene for the human IFNa receptor resided on human chromosome 21 (26,  (2) different isoforms of human IFNa differed in their ability to bind to and activate human and murine cells (28). A species-nonspecific form of human IFNa ( I F N~A -~) up-regulated MHC Class I expression on either SCC16-5 ( Fig.  1, panel e ) or normal murine L cells (data not shown). MHC Class I up-regulation was also observed when SCC16-5 were exposed to recombinant human IFNaZ,, a species-specific form that reacts only with human and not murine IFNa receptors (Fig. 1, panel f ) . No up-regulation of MHC Class I was observed when normal L cells were incubated with HuIFNa2. under identical conditions (see Fig. 4). These results thus confirm the presence of human chromosome 21 in the murine SCC16-5 cell line and thereby establish the suitability of these cells for use in the structure-function analysis of the human IFNy receptor.

Generation and characterization of a Human IFNy Receptor
Truncation Mutant That Lacks the Intracellular Domain-In order to determine whether the intracellular domain of the human 1FN-y receptor was involved in either ligand processing or biologic response induction, we stably expressed in SCC16-5 two human IFNy receptor cDNAs: (1) one that encoded the full-length receptor and (2) a second that encoded a truncated form of the receptor that lacked all but the first 4 amino acids of the intracellular domain. As a control, the full-length human receptor cDNA was also stably expressed in normal murine L929 cells. Following selection, homogeneous populations of human IFNy receptor-positive cells were derived by repeated rounds of cell sorting using receptor-specific monoclonal antibodies and expansion. This protocol resulted in the establishment of the three cell lines denoted SCC.hgR, SCC.hgRAIC, and L929.hgR representing SCC16-5 express-.. c Fluorescence Intensity ing full-length or truncated human IFNy receptors and L cells expressing the full-length human IFNy receptor, respectively. Expression of receptors on the surface of the resulting cell lines was confirmed qualitatively by flow cytometry (Fig.  2) and quantitatively by radioligand binding analysis. Based on at least four independent determinations, SCC.hgR, SCC.hgRAIC, and L929.hgR expressed 1,300 f 300,81,000 f 39,000, and 5,000 f 2,000 receptors/cell, respectively. Whereas the ligand binding affinities of the full-length receptors in either SCC16-5 or L cells (7.9 +. 3.3 X lo9 "' and 5.9 f 3.1 X lo9 ?VI-', respectively) were indistinguishable from that of natural human IFNy receptors in authentic human WISH cells (6.7 X lo9 M-'), the truncated receptor in SCC16-5 displayed a 10-fold lower KO (0.8 f 0.4 x IO9 M-'). Surface labeling of the transfected cell lines followed by immunoprecipitation and SDS-PAGE analysis confirmed that the recombinant receptors displayed the appropriate molecular sizes (Fig. 3). Full-length receptors expressed on SCC.hgR and L929.hgR migrated as proteins of 86 kDa, whereas the truncated receptor expressed on SCC.hgRAIC displayed an M, of 52,000.

Demonstration That the Intracellular Domain of the IFNy Receptor Is Required for IFNy-dependent Induction of Biologic
Responses-The three transfectants were tested for their ability to respond to human or murine IFNr by monitoring MHC Class I expression. As expected, all three lines responded to their homologous murine ligand and showed increases of 60-80 channels when examined for H-2Kk expression by flow cytometry (Fig. 4, middle panels). In contrast, only SCC.hgR responded to human IFNy showing a 56-channel increase in MHC Class I expression (Fig. 4, top panels). The lack of

FIG. 2. Expression of full-length and truncated human IFNr receptors on murine fibroblasts. Fulllength and truncated
IFNr receptor cDNAs were constructed as described under "Experimental Procedures." The products encoded by the resulting cDNAs are schematically represented on the left side of the figure. The cDNAs were cloned into the mammalian cell expression vector pSFFV and stably transfected into SCC16-5 or L929 cells. Receptor expression on the transfected cell lines as determined by flow cytometry is depicted in the right panels. SCC16-5, SCC.hgR, SCC.hgRAIC, and L929.hgR were stained with 5 pg of biotinylated human IFNr receptor specific monoclonal antibody (GIR-208) and streptavidin-phycoerythrin. Dotted lines represent background staining determined as in Fig. 1. responsiveness of SCC.hgRAIC to human IFNy could not be attributed to a spontaneous loss of chromosome 21, since both SCC.hgRAIC and SCC.hgR responded to human IFNah (Fig.  4, bottom panels). The species specificity of the IFNaz. preparation used in these studies was confirmed by demonstrating its total lack of activity of L929.hgR cells (Fig. 4, bottom right  panel). Thus, these results confirm that a human chromosome 21 gene product is required to form a functionally active human IFNy receptor in murine cells and establish that the receptor's intracellular domain plays an obligatory role in promoting biologic responses.

Demonstration That the Intracellular Domain of the ZFNy Receptor Is Required for Ligand Internalization and Process-
ing-IFNy has been shown to be internalized and degraded after binding to its receptor on a variety of different primary cells and cultured cell lines (24, 29). Moreover, others have .. .. ..

Fluorescence Intensity (Human IFNyR)
suggested that intracellular IFNy plays an important role in inducing cellular responses (30-32). Therefore, we examined the role of both chromosome 21 and the intracellular domain on ligand internalization and degradation. As shown in Fig.  5, both L929.hgR and SCC.hgR internalized and degraded human IFNy. After 1 h, L929.hgR and SCC.hgR internalized 43 and 31%, respectively, of the total surface bound ligand. This result demonstrates that the human accessory molecule is not required for the internalization of human ligand in the transfected cells. In contrast, SCC.hgRAIC internalized only 7% of cell bound human ligand. Moreover, L929.hgR and SCC.hgR degraded 79 and 68% of cell associated lZ5I-HuIFNy, respectively, whereas SCC.hgRAIC degraded less than 6% of bound ligand. Thus, although the receptor's intracellular domain is required for high efficiency internalization and degradation of ligand, its role in these processes is not speciesspecific.

tants-The data presented thus far clearly established the importance of the IFNy receptor's intracellular domain for both ligand internalization and biologic response induction.
We next attempted to more precisely localize the relevant functional portions of the intracellular domain. Four cDNAs were constructed tllat encoded receptor mutants with sequential deletions that spanned the entire length of the receptor's intracellular domain (Fig. 6). The cDNA constructs were expressed stably in SCC16-5 and expression confirmed by ( a ) flow cytometry (Fig. 6) and ( b ) radioligand binding analysis ( Table I). The latter demonstrated that the ligand binding affinities exhibited by the receptor deletion mutants were indistinguishable from that displayed by the full-length receptor.
To confirm that the resulting cell lines expressed only the expected deletion mutants, a PCR assay was developed based on amplification of the nucleotide sequence that encoded the intracellular domain of the transfected cDNAs. Fig. 7 (left  panel, lanes a -d ) depicts the PCR products obtained from were incubated with saturating doses of '261-HuIFN-y for 1 h at 4 "C and then shifted to 37 "C for various periods of time. Ligand internalization was defined as the amount of ligand which remained cell associated following treatment of the cells with 0.25 M acetic acid, pH 2.7, plus the amount of degraded ligand. Degraded IFNr was defined as the number of counts that could not be precipitated with trichloroacetic acid. Nonspecific binding as determined using cells pretreated with a 1000-fold excess of unlabeled IFN-y has been subtracted.
lysates of the parental SCC16-5, SCC.hgR, SCC.hgRAIC, and L929.hgR cell lines, respectively. A 930-bp component representing the amplification product of the complete intracellular domain was observed for both SCC.hgR and L929.hgR, while no bands were observed for SCC16-5 and SCC.hgRAIC. The latter result was expected since SCC16-5 lacks the human IFNr receptor cDNA template for amplification and SCC.hgRAIC lacks the human receptor 3' intracellular domain sequences required for amplification in this PCR assay. Lanes e-h depict the PCR analyses of lysates from cell lines containing the A256-303, A304-379, A380-433, and A434-472 deletion mutants. These constructs gave rise to the expected products of 777,693,759, and 0 bp, respectively. Using another set of 3' primers for PCR we were able to distinguish between A434-472 and AIC mutants, which gave rise to 850-and 315-bp products, respectively (Fig. 7, right panel, lanes a  and b), and between A256-303 and A380-433, which gave rise to products of 654 or 0 bp, respectively (lanes c and d).
The physical state of the receptor deletion mutant proteins expressed at the surface of the transfected cells was analyzed by SDS-PAGE following surface labeling and immunoprecipitation (Fig. 8). This analysis showed that all deletion mutants Lysates were subjected to PCR as described, using primers 949 and 5054 (lanes a and b ) or primers 949  and 8202 (lanes c and d ) . 5,000,000 SCC16-5 (a), SCC.hgR ( b ) , SCC.hgRAIC ( c ) , L929.hgR (d),

46.0-
four receptor deletion mutants were tested for their ability to internalize and degrade ligand. Deletion mutants A304-379, A380-433, and A434-472 were all able to direct the degradation of 60-80% of cell-associated '"I-HuIFNy during a 4-h incubation (Fig. 9). This result is indistinguishable from that obtained with the full-length receptor expressed on SCC.hgR (68% degraded) or L929.hgR (79% degraded). However, under identical conditions the deletion mutant A256-303 degraded only 12% of bound ligand. The ligand processing defect displayed by SCC.hgRA256-303 was nearly identical to that displayed by the fully truncated receptor (SCC.hgRAIC, 7% ligand degradation). This result thus localizes the region of the intracellular domain required for internalization to the transmembrane-proximal portion of the molecule.
Two Restricted Regions of the Intracellular Domain Are Required for Functional Activity-The SCC cell lines expressing the receptor deletion mutants were then tested for responsiveness to human IFNy. The two deletion mutants lacking interior regions of the intracellular domain, A304-379 and A380-433, responded to human IFNy and showed MHC Class I increases of 52 and 53 channels, respectively (Fig. 10, panels  b and c). The observed enhancement of MHC Class I expression in these two mutant cell lines was equivalent to that seen when the mutants were stimulated with murine IFNy (panels f and g, 62-and 57-channel increases, respectively) or when SCC.hgR were stimulated with human IFNy (56-channel hgRA380-433 (c), SCC.hgRA434-472 ( d ) , or SCC16-5 ( e ) were surface-iodinated using lactoperoxidase and Na1z51 for 45 min at room temperature as described under "Experimental Procedures." Cells were washed and lysed, and receptors immunoprecipitated with GIR-208. Radiolabeled receptors were analyzed as in Fig. 3. Lane A represents a shorter exposure of the same autorad because of the higher expression of the A256-303 mutant.
In contrast, deletion of either the transmembrane-proximal 48 amino acids, A256-303 (Fig. 10, panel a ) , or the carboxylterminal 39 amino acids, A434-472 (panel d ) of the intracellular domain completely abrogated functional responses of the cells to human IFNy. This loss of functional activity was not due to a general loss of MHC Class I inducibility, since both deletion mutant expressing cell lines responded to murine IFNy (panels e and h, 65-and 71-channel increases, respectively). Moreover, the two cell lines also responded to human IFNaz (panels i and l ) , thereby ruling out the possibility that they had spontaneously lost human chromosome 21. These results thus identify two distinct, functionally important regions in the IFNr receptor's intracellular domain: a membrane-proximal region of 48 amino acids that contains elements required for both ligand internalization and biologic response induction and a carboxyl-terminal region of 39 amino acids that contains elements required selectively for induction of cellular responses.

DISCUSSION
This paper demonstrates that the intracellular domain of the IFNy receptor plays an obligatory role in mediating the protein's functional activity. Moreover, the results specifically identify two distinct functionally important regions within this domain. The first is comprised of 48 amino acids, proxir I 1 FIG. 9. Receptor-mediated internalization/degradation of ligand requires a 48-amino acid region of the receptor's intracellular domain. Cells were incubated with saturating doses of '"I-HuIFNy at 4 "C for 1 h. Unbound IFNy was then removed by washing and the cells shifted to 37 'C. After 4 h at 37 "C, cell supernatants were harvested and degraded IFNr determined as described under "Experimental Procedures." Nonspecific binding has been corrected for using cells pretreated with a 1000-fold excess of unlabeled HuIFNy.  -303 ( a , e , a n d i), f ,and j ) , g, and k ) , and SCC.hgRA434-472 (d, h, and I ) were cultured in 10 ml of media for 72 h with 1000 IU/ml of either human IFNy (ad ) , murine IFNy (e-h), or human IFNolz (c, j , k, and I ) . Cells were harvested, washed, and then stained for MHC Class I expression using the 11-4.1 monoclonal antibody. Expression was quantitated by flow cytometry. Dotted lines represent the constitutive level of MHC Class I expressed on cells incubated in medium alone. Background staining was determined as in Fig. 1. mal to the receptor's transmembrane domain (amino acids 256-303), and contains elements required for both receptormediated ligand internalization and biologic response induction (enhancement of MHC Class I expression). The second is comprised of the 39 carboxyl-terminal amino acids (amino acids 434-472). This region is not important for ligand internalization but contains elements absolutely required for induction of IFNy-mediated biologic responses. These results thus demonstrate the critical importance of the IFNy receptor's intracellular domain and show that ligand internalization is not sufficient for induction of IFNy-dependent biologic responses.
The current study was dependent upon an ability to express an active human IFNy receptor in murine cells. Recent reports have indicated that functional human IFNy receptors require the presence of at least two human proteins: the human IFNy receptor polypeptide and a second, as yet uncharacterized, protein encoded by a gene on human chromosome 21 (7,8,(13)(14)(15). This observation was confirmed herein using murine fibroblast lines either lacking or containing human chromosome 21. Normal murine L cells transfected with the human IFNy receptor cDNA bound but did not respond to human IFNy. In contrast, a murine L cell derivative that stably expressed a single copy of human chromosome 21 (SCC16-5) became responsive to human IFNy following transfection with the human receptor cDNA. As a measurement of cellular responsiveness, we chose to quantitate IFNy's ability to enhance MHC Class I antigen expression. This read-out was selected because (a) it represents a documented physiologic response of murine fibroblasts to interferon (2) and (6) it was of sufficient magnitude to provide unequivocal and reproducible results. All the L cells used in this study expressed increased levels of MHC Class I antigens following exposure to either murine IFNy or a species-nonspecific form of human IFNa (IFN~A-D). In contrast, only murine L cells containing human chromosome 21 (which also contains the gene that encodes the human IFNa receptor) responded to human IFNa2, a species-specific form of IFNa. Thus the same assay system could be used to confirm the presence of chromosome 21 in the transfected cell lines. Taken together, these results validate the use of SCC16-5 and its derivatives to study the structure/function relationships that exist within the human IFNy receptor. Using this assay system we showed first that the intracellular domain of the IFNy receptor was required for receptor . . . . . . mediated ligand internalization and cellular responses. Subsequently, two restricted regions within this domain were identified which contained elements that were required for these activities. To rule out uncertainties concerning the receptor cDNA constructs, the identity of each construct was confirmed by sequencing and the presence of the appropriate receptor cDNA in each cell established by PCR. To rule out the possibility that the mutant receptor proteins were postsynthetically cleaved we documented their molecular masses by surface labeling and SDS-PAGE analysis. Thus, it was possible to show that each cell line expressed on its surface only the expected receptor-related polypeptide. The two functionally important regions of the intracellular domain (amino acids 256-303 and 434-472) do not show a high degree of sequence homology to other known proteins including kinases, phosphatases, or other cytokine receptors. Thus it is not yet clear what their exact roles are in the signaling/internalization process. However, the 256-303 region of the human IFNy receptor contains a consensus motif found in the intracellular domains of a variety of rapidly internalized receptors including, among others, the transferrin receptor (33) and the mannose-6-phosphate receptor (34). This consensus motif forms a Type I turn (33), which appears to be the critical structural determinant recognized in the internalization process. The internalization motif is a tetrapeptide that consists of an aromatic amino acid separated by 2 residues from a hydrophobic residue that occurs within 35 amino acids of the membrane. Interestingly, both the human and murine IFNy receptors have sequences which exactly fit this motif (YVSL, residues 287-290, and YSLV, residues 285-288, respectively). Based on hydropathy analysis of this region in the receptor using the method of Emini et al. (35), these sequences have a high probability of being exposed at the protein's surface (data not shown). These observations may explain the inability of receptor mutants lacking either the entire intracellular domain or specifically the 256-303 region to mediate internalization and degradation of ligand. This possibility is currently being assessed using site-directed mutagenesis approaches.
Based on the observations that ( a ) the intracellular domains of the human and murine IFNy receptors are only 55% identical at the protein level and ( b ) the receptor's accessory molecule is species-specific, we wondered whether the functionally important regions would reside within areas of low sequence homology. However, when the primary structures of the human and murine IFNy receptors were aligned, the functionally important regions localized to the more conserved sequences. The 256-303 and 434-472 regions displayed 67 and 64% sequence identity to the corresponding sequences in the murine IFNy receptor (residues 255-301 and 413-451, respectively). In fact, some of the highest homology between the human and murine intracellular domains are located within these two regions (88% identity between human residues 266-290 and murine residues 264-288 and 85% identity for human residues 439-452 and murine residues 418-431). In contrast, the other two intracellular domain regions (304-379 and 380-433 for human and 302-368 and 369-412 for murine) showed much less sequence identity (41 and 37%, respectively). This observation is consistent with the finding that functionally important regions on related proteins tend t o be conserved. Recent observations from Gibbs and Goedde13 have localized the region of the IFNy receptor that is required for the species-specific action of the accessory protein to the extracellular and/or transmembrane domain(s) of the molecule. Thus taken together these results demonstrate that the V. Gibbs and D. Goeddel, unpublished observations. intracellular domain of the IFNy receptor is absolutely required for function but its actions are not species-specific.
An additional observation made in the current study was that the ligand binding affinity of the fully truncated human IFNy receptor (hgRAIC) in SCC16-5 was one log lower than any of the other constructs. It is possible that this difference reflects the nonspecies-specific interaction of the intracellular domains of the IFNy receptor and the accessory molecule. This hypothesis is supported by past studies showing that solubilized forms of intact human IFNy receptors, either in detergents or following reconstitution into liposomes, bound ligand with 10-30-fold lower affinity than receptors in the context of their native membranes (5): Similar observations have been made using the soluble recombinant human IFNy receptor extracellular domain (36). However, the full-length human receptor in L cells displayed a normal K, for human IFNy despite the absence in these cells of the human accessory molecule. This result may reflect the ability of the receptor's intracellular domain to interact with other cellular components in a nonspecies-specific manner.
Our analysis of the IFNy receptor has revealed some general analogies and specific contrasts to other cytokine receptors. For example, the IL-6 receptor consists of two distinct proteins: gp85, which is mainly responsible for ligand binding, and gp130, which is responsible for signal transduction (37, 38). However, the two receptor systems are quite distinct on the basis of both structural and functional criteria. First, whereas gp85 and gp130 share structural characteristics with other members of the Type I cytokine receptor family (39) that include many of the growth factor receptors such as those for erythropoietin, granulocyte/macrophage colony stimulating factor, granulocyte colony stimulating factor, IL-2 (p75), IL-3, IL-4, IL-5, and IL-7, the IFNy receptor does not. In fact, the IFNy receptor is a member of the type I1 cytokine receptor family that also includes the IFNa receptor and tissue factor. This latter family shows only minor structural homology to the type I receptor family. Second, whereas the intracellular domain of the ligand binding component of the IL-6 receptor (gp85) is not required for induction of cellular responses, the IFNy receptor's intracellular domain is an obligatory component of the system. Finally, although a trimolecular complex of IL-6, gp85, and gp130 has been demonstrated, no such complex involving IFNy-reactive cell surface components has yet been observed. Thus although the IFNy receptor and the IL-6 receptor show a similar requirement for two polypeptides to form an active signal transduction system, they are distinguishable by the functional roles played by the individual polypeptides that comprise each receptor. This dissociation between IL-6 and IFNy receptors is not unexpected since they induce distinct (and even opposite) cellular responses. Whereas IL-6 acts, among other things, to promote cell growth (40-42), IFNy acts, among other things, to inhibit cell growth (2).
However, the intracellular domains of at least some other cytokine receptors (such as the p75 chain of the IL-2 receptor) are required for signal transduction. Hatakeyama et al. (43) used a panel of p75 deletion mutants to show that a 46-amino acid region of the intracellular domain of p75 was necessary for expression of a functionally active IL-2 receptor. However, even a truncated form of p75 lacking all but 27 residues of the intracellular domain, although functionally inactive, nevertheless was able to direct the internalization of IL-2. The results presented in the current paper therefore are similar yet distinct from those obtained for the IL-2 receptor in that the intracellular domain of the IFNy receptor mediates both cellular responsiveness and ligand internalization.
Finally, the current study adds new insights into the controversy concerning the role played by the IFNy receptor in inducing biologic responses. Several studies exist which suggest that the IFNy receptor mediates its effects on cells through classical signal transduction pathways involving calcium (44-46), serine/threonine protein kinases (16,17,47,48), and/or activation of the Na+/K+ antiporter (49). In contrast, other investigators have suggested that the receptor functions as a ligand transporter and that intracellular ligand acts as its own second messenger. Three different laboratories have tried to circumvent the species specificity of IFNy by introducing it directly into species-mismatched cells using techniques that bypass the endogenous surface IFNy receptor. These include liposome encapsulation (30), transfection (31), or microinjection (32). In all three cases, a biologic response was observed. However, the magnitude of the responses effected using receptor-bypass techniques was significantly less than those induced by homologous ligand supplied at the cell surface. Based on the latter experiments, some investigators have suggested that a major function of the IFNy receptor may be to transport ligand into the cell. The results of the current study demonstrate that ligand internalization is not sufficient for biologic response induction. This conclusion is derived from studies using both the L929.hgR cell line, which lacks human chromosome 21, and the SCC.hgRA434-472 cell line, which contains chromosome 21, thus eliminating any questions concerning the role of the receptor's accessory component in this process. Human IFNy receptors expressed on either cell line were fully capable of internalizing and directing the degradation of ligand in a manner indistinguishable from that observed for functionally active human receptors expressed on SCC16-5 cells (SCC.hgR, SCC.hgRA304-379, and SCC.hgRA380-433). The demonstration that ligand internalization is not sufficient for induction of IFNy-mediated biologic responses strengthens the concept that the receptor, and specifically its intracellular domain, plays an active role in regulating cellular function. Whether the receptor acts by directing the intracellular trafficking of ligand and/or effecting specific pathways of signal transduction remains to be determined.