Affinity Purification, Peptide Analysis, and cDNA Sequence of the Mouse Interferon y Receptor*

The receptor for mouse interferon gamma (IFN-gamma) was purified from detergent-solubilized plasma membranes of EL-4, a thymoma cell line which expresses a high number of receptors on its cell surface. The purification was carried out by immunoaffinity chromatography using an anti-receptor monoclonal antibody. The purified receptor was subjected to NH2-terminal sequence analysis as well as sequencing of endopeptidase-generated peptides. One of the peptides was found to be identical to a portion of the published amino acid sequence of the human IFN-gamma receptor deduced from cDNA. This information was utilized to construct a mixed-sequence oligodeoxynucleotide probe which permitted the isolation of a full-length cDNA clone coding for the mouse IFN-gamma receptor. The mouse IFN-gamma receptor cDNA is comprised of 105 base pairs of the 5'-untranslated region, an open reading frame coding for a 477-amino acid serine-rich protein having calculated Mr 52,276, and a 3'-untranslated region of 539 base pairs. The receptor is first synthesized as a pre-protein from which a 25-amino acid signal peptide is cleaved. The receptor contains a hydrophobic transmembrane portion near the center of the molecule. Northern blot analysis of various cell lines showed that each contained a single 2.0-kilobase mRNA. A direct correlation between the amount of IFN-gamma receptor mRNA and the level of receptor expressed on the cell surface was observed. The mouse and human IFN-gamma receptors are structurally similar, showing 51% over-all homology in amino acid sequence. Mouse IFN-gamma receptor cDNA when inserted in a mammalian shuttle vector and transfected into COS-7 monkey cells was able to direct the expression of specific binding activity for mouse IFN-gamma.

The receptor for mouse interferon y (IFN-7) was purified from detergent-solubilized plasma membranes of EL-4, a thymoma cell line which expresses a high number of receptors on its cell surface. The purification was carried out by immunoaffinity chromatography using an anti-receptor monoclonal antibody. The purified receptor was subjected to NHz-terminal sequence analysis as well as sequencing of endopeptidase-generated peptides.
One of the peptides was found to be identical to a portion of the published amino acid sequence of the human IFN-7 receptor deduced from cDNA.
This information was utilized to construct a mixed-sequence oligodeoxynucleotide probe which permitted the isolation of a full-length cDNA clone coding for the mouse IFN--, receptor.
The and transfected into COS-7 monkey cells was able to direct the expression of specific binding activity for mouse IFN-y.

Interferons
(a, p, y) are polypeptide hormones involved in the modulation of cellular growth and differentiation, and in the release of certain cytokines, which in turn have a pleio- The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL. Data Bank with accession number(s) 505265.
11 To whom correspondence should be addressed: NCI, NIH, Bldg. 37, Rm. lB04, Bethesda, MD 20892. Tel.: 301-496-4549. tropic effect on target cells (l-6). IFNs' are involved in the induction of an antiviral response in the inhibition of cell proliferation and cellular mobility (7), and in the expression of membrane class I and II antigens, certain cell surfaces receptors (8,9), and fibronectin (10). IFN-7 and to a lesser extent IFN-a and $3 are known to be immunomodulatory factors (11). IFN-7 strongly inhibits mitogen activation of B cells, but promotes growth and final differentiation of these into immunoglobulin producing cells (12). In addition, it activates macrophages, boosts cytotoxicity of natural killer cells and stimulates T cell cytotoxicity (13,14). IFN-7 is synthesized and released from immune effector cells (activated T lymphocytes and NK cells) upon stimulation by various agents (alloantigens, tumors, mitogens, etc.). The biological response to IFN-7 is induced by its binding to a specific receptor on the surface of target cells (ll), which is distinct from that recognized by IFN-(Y and IFN-/3 (15). IFNy receptors are generally expressed at low levels, except on certain tumor cells, which exhibit up to 50,000 receptors/cell (16,17). The IFN-7 receptor binds noncooperatively to the ligand with high affinity, with & ranging from lo-' to 10-l' M (16, 18-20).
The human IFN-7 receptor was purified by sequential ligand-affinity chromatography and immobilized anti-receptor monoclonal antibodies (21). The receptor consists of two protein species with an apparent M, of 90,000 and 50,000 of which the 50-kDa component was shown to be a proteolytic degradation product of the 90-kDa species. Recently, a cDNA coding for the human IFN-7 receptor was isolated from an expression library using a polyclonal anti-receptor antibody (22). The deduced amino acid sequence revealed structural features typical of cell surface receptors, such as a putative signal peptide and a hydrophobic, transmembrane-like portion.
The mouse IFN-7 receptor has been solubilized by various detergents from plasma membranes of myelomonocytic (WHEI-3) and thymoma (EL-4) cell lines and identified by direct radiolabeled 24). Detergent-solubilized membranes showed only a single class of binding activity with values comparable to that obtained using intact cells. IFN-r binding was demonstrated to be specifically competed by unlabeled IFN-7, as well as by monoclonal antibodies (23-25). The receptor was shown to be a glycoprotein of 85-95 kDa.
To determine the structure of the mouse IFN-7 receptor we have purified the protein on a large scale by affinity chromatography with immobilized monoclonal anti-receptor antibodies.
Chemical sequencing of peptides obtained by digestion of the purified receptor with a lysine-specific endopeptidase has permitted the construction of an oligodeoxynucleotide probe which was used to isolate a full-length mouse receptor cDNA clone. in the presence or absence of lOO-fold'exckss of-unlabeled IFN-7 and incubated for 1 h at 35 "C. After washing the cells 3 x with PBS, 1 ml of trypsin was added, the cells collected, and radioactivity measured in a y-counter.

IFN-7
Receptor Purification-To purify the mouse IFN-7 receptor on a large scale, we tested the effectiveness of both ligand and monoclonal antibody affinity chromatography. Although mouse recombinant IFN-7 linked to Sepharose 4B retained negligible amounts of receptor activity, immu-Comparison of Mouse and Human IFN-7 Receptors noaffinity column chromatography proved to be effective. Recovery of the mouse receptor from the immunoaffinity column was found to be optimal under conditions of mild acid treatment with glycine-HCl buffer, pH 3, containing 8 mM CHAPS. Before elution of the receptor, however, extensive washes under stringent conditions were required (see "Materials and Methods").
Following acid elution, it was necessary to immediately neutralize the preparation to maintain integrity of the receptor. The receptor preparation was found to be predominantly a single band of -85 kDa by silver staining of SDS-PAGE gels (Fig. 1A). The band itself was somewhat diffuse, typical of a glycosylated protein. A band of the same apparent M, was also recognized with the GR-20 monoclonal antibody by Western blotting (Fig. 1B). From 2 x 10'" cells, the yield of the receptor purified by this procedure was determined by amino acid analysis to be 150 pmol.
Amino Acid Sequence Analysis of IFN-7 Receptor-About 50 pmol of pure receptor were digested with Achromobacter lyticus Protease I, a lysine-specific endopeptidase, and the products separated by HPLC. The chromatographic profile consisted of 17 peaks (Fig. 2). Fractions corresponding to four of the peaks were subjected to automated Edman degradation and each of these was found to give a single sequence. The sequence of the peptide in the fraction corresponding to peak 2 showed 100% identity with a stretch of amino acids deduced for the human IFN-7 receptor. The sequences of the peptides in fractions corresponding to peaks 4 and I showed 60 and 40% identity, respectively, to the human receptor, whereas peptide from peak 3 showed no similarity.
The sequences of these peptides and yield at each cycle are shown in Table I.
The undigested receptor was also subjected to automated Edman degradation to determine the NH1-terminal amino acid sequence. The sequence obtained was found to match that determined for the peptide from peak 7, which was generated by endopeptidase digestion. cDNA Cloning and Nucleotide Sequence Analysis-A mouse cDNA library consisting of approximately 6 X lo5 recombinant phage, constructed with poly(A)+ RNA from the cell line ABPL-2, was screened at low stringency of hybridization with an 24-base mixed-sequence oligodeoxynucleotide probe. The probe was constructed on the basis of the amino acid sequence of the peptide from peak 2 (see "Materials and Methods"). Seven positive signals were observed on autoradiographs of duplicate plaque lifts (-0.001% abundance). One of these Lane A, an aliquot of affinity purified  receptor was subjected to IO-15% gradient SDS-PAGE and visualized by silver staining. Lane B, a sample from the same fraction was subjected to 7.5% SDS-PAGE, transferred to nitrocellulose filter, and reacted with GR-20, a rat antimouse IFN-+y receptor monoclonal antibody. Peptides were separated by HPLC on an RP 300 reversephase column with a O-70% linear gradient of acetonitrile in 0.05% trifluoroacetic acid at a flow rate of 0.1 ml/min. a Numbers in parentheses are picomoles of phenylthiohydantoinderivatives detected.
b Not detected.
clones, designated XMIR3b1, was found by nucleotide sequence analysis to contain a full-length cDNA. All of the peptide sequences determined for the mouse IFN-7 receptor were in complete agreement with the determined cDNA sequence, thereby verifying the reading frame of the full-length cDNA clone. The remaining clones contained the 3'-noncoding portion together with a truncated coding region. The structural map and sequencing strategy for mouse IFN--, receptor cDNA are shown in Fig. 3. The cDNA was found to be unique, having no homology to entries in the GenBank other than that of the human IFN-7 receptor (22). A comparison of the nucleotide sequence of the full-length cDNA and deduced protein sequence for the mouse IFN-7 receptor with those of the human receptor is shown in Fig. 4 and 3'-untranslated regions having 73, 49, and 42% G + C content, respectively.
The first ATG triplet encountered from the 5' end of the cloned cDNA is surrounded by the sequence GCAGGAATGG, which matches 7 of 10 positions in the sequence believed to be optimal for translation initiation, GCC(A,G)CCAUGG (39). Interestingly, an additional in-frame AUG triplet occurs nine codons downstream from the first, AUG, which coincides with the proposed start site for translation of the human receptor protein and is surrounded by a sequence which matches 6 of 10 positions of the canonical translation initiation motif. However, in accordance with the scanning model for translational initiation by the eucaryotic ribosome (39), very little, if any, protein product would be expected to originate from the second AUG. The mouse receptor mRNA has the termination codon UAA and contains a variant polyadenylation signal AGUAAA, located 12 bases upstream from the poly(A) tail, compared to the typical signal, AAUAAA . The open reading frame in the full-length cDNA codes for a 477-amino acid protein having M, = 52,276. The discrepancy between the size of the cloned product and the apparent size of the purified receptor isolated from membranes is apparently the result of the known glycosylated nature of the receptor (25).* The NH2 terminus of the mature receptor starts at Ala-26. The receptor is therefore synthesized as a preprotein from which a 25-amino acid signal peptide is cleaved. The mouse receptor contains two portions having considerable hydropathic character. One of these is located between residues 5 and 20 at the NH, terminus, typical of a hydrophobic core of a signal sequence. The other is located at amino acids residues 254-275, suggestive of a transmembrane region.
The mouse and human receptor cDNAs share a number of common structural features. The two cDNAs show an overall similarity of 60%; there is 51% homology between the 5'untranslated region of the human receptor cDNA and the tors as deduced from their respective cDNAs show an overall similarity of 51%. The mouse receptor protein contains a nine-amino acid extension in the NH*-terminal signal sequence which is absent from the human receptor. Despite this extention, the mouse receptor is still 12 amino acids shorter than the human sequence as a result of several internal stretches which are missing in the mouse sequence but present in the human receptor. One portion of the mouse receptor sequence shows a striking similarity to the human receptor; there is 88% identity in amino acid residues 286-311 to the corresponding portion of the human receptor (Fig. 4). This stretch also contains a sequence of 12 consecutive identical residues, the longest region of identity between the two molecules.
In Vitro Transcription/Translation-Capped RNA was synthesized in uitro from the T7 promoter contained on phagemid DNA derived from clone XMIR3bl.
The T7 promoter in this phagemid is contained in the polylinker portion of the vector and is located upstream of the 5' end of the mouse IFN-7 receptor cDNA insert. Prior to transcription, the phagemid was cleaved with HindIII, which cuts the template in the vector portion downstream of the poly(A) tail of the cDNA insert, producing a single 2.2-kilobase transcript (not shown). The length of this transcript is consistent with the distance between the T7 promoter and the site of scission of the phagemid. The RNA was translated in uitro, in the presence or absence of microsomal membranes, and the products were immunoprecipitated with the monoclonal antibody GR-20. The monoclonal antibody recognized a -75-kDa in vitro translation product derived from lysates supplemented with microsomal membranes (Fig. 5, lane 3). The size of the product corresponded to the lower portion of the M, range of authentic glycosylated mouse IFN-7 receptor immunoprecipitated from EL-4 cells metabolically labeled with [35S]methionine (Fig. 5, lane 2). The monoclonal antibody was unable to recognize the translation product when synthesized in the absence of glycosylating activity (Fig. 5, lane 4). Expression of Mouse IFN-7 Receptor in Transfected COS-7 Cells-For transfection experiments, a 2.2-kilobase PstI-KpnI segment from the phagemid, containing the full-length mouse IFN-7 receptor cDNA, was inserted into these same sites in the mammilian shuttle vector pCD-PS (42). Various amounts of DNA from the resulting construct, designated pCD-PS-MIR3b1, were transfected into COS-7 cells and the expression of 1251-mouse IFN-7 binding activity was measured 72 h after transfection. Table II shows that specific binding increased with an increasing amount of transfected DNA. No binding was observed with nontransfected COS-7 cells or transfectants with the vector itself.
Northern Blot Analysis-Poly(A) mRNA was isolated from seven different cell lines known to exhibit varying levels of IFN-7 receptor on their cell surface. The RNAs were subjected to Northern blot analysis using the full-length cDNA as probe under conditions of high stringency of hybridization. As shown in Fig. 6, a single band corresponding to a 2.0-kilobase mRNA was observed in all cell lines tested. The levels of mRNA in the various cell lines was seen to vary strikingly, Alignment of the S'-untranslated region of the mouse (M) cDNA with the 5' end of the cDNA sequence from human (H) is shown on the first two lines of the diagram. In the coding regions of the two cDNAs (lines 3-54), the comparison is shown in sets of four lines, the upper two lines being an alignment of the deduced protein sequences (single-letter code) and the lower two lines being an alignment of the cDNAs. Dots indicate identical residues. Dashes indicate gaps inserted to maximize alignment between the sequences. Numbers on the left indicate the position in the sequence of the first residue of the corresponding line. In the protein sequences, the NH*-terminal signal peptides and the putative transmembrane regions are underlined. Potential asparagine-linked glycosylation sites are underlined with a dashed line. In the cDNA sequences, the triplets for translation initiation and translation termination, as well as the putative polyadenylation signals, are underlined. Data for the human receptor sequence are from Phagemid DNA derived from clone XMIR3bl was linearized with Hind111 and used as template to generate capped RNA transcripts of mouse IFN-7 receptor cDNA. Transcription was from the T7 promoter contained in the polylinker portion of the vector located upstream of the 5' end of the cDNA insert. RNA samples were translated in a reticulocvte lvsate in the presence of [""Slmethionine. Shown is an autoradiograph of a Western blot of proteins immunoprecipitated by monoclonal antibody GR-20 from CHAPS-solubilized cells labeled with [""Slmethionine and from in vitro translation reactions. Separation was by 12.5% SDS-PAGE. Immunoprecipitates were from the following sources: Lane I, solubilized CV-1 monkey cells; lane 2, solubilized EL-4 mouse cells; lane 3, in vitro translation products derived from recombinant mouse IFN--y receptor transcripts in which the reticulocyte lysate was supplemented with canine pancreatic microsomal membranes: lane 4. as in lane 3 except that translation was performed in the absence of microsomal membranes.  "Specific binding is expressed as the difference in counts/min between the total and the nonspecific binding of ""I-IFN-7 to whole cells, the total and nonspecific binding being shown in parentheses. Nonspecific binding was the amount of  bound in the presence of a loo-fold excess of unlabeled IFN-7. single-column immunoaffinity purification step was sufficient to obtain pure IFN-7 receptor. The purified receptor preparation consisted predominantly of the 85-kDa species. However, a minor component of about 200 kDa was occasionally detected by Western blotting.
The 200-kDa component was detected with rabbit anti-rat antibodies alone and is most likely the result of the leaching of a small amount of monoclonal antibody from the affinity column. Also, a 60-kDa minor component was detected on silver-stained gels. This component is presumably a degradation product of the receptor since it increases upon prolonged storage of the receptor preparation.
receptor for both IFN--y and the GR-20 monoclonal antibody was found to be preserved in membranes solubilized with this particular detergent. Also, a higher specific activity (receptor concentration/mg of total solubilized protein) was obtained with CHAPS, as compared with other detergents such as Triton X-100 and Nonidet P-40. Under these conditions, a Partial Amino Acid Sequence-The NH2 terminus and lysine-specific endopeptidase-generated peptides of the IFN--, receptor were subjected to sequence analysis. All of the endopeptidase-generated peptides, which gave a single sequence, were in agreement with the protein sequence deduced from the cDNA clone, thereby providing additional evidence of the purity of the receptor preparation. One of these peptides showed 100% identity to a portion of the deduced amino acid sequence of the human IFN-7 receptor. This peptide was chosen to synthesize an oligodeoxynucleotide probe for screening a cDNA library. Despite the degeneracy of this Comparison of Mouse and Human IFN--y Receptors probe, we found that it could be used successfully to detect IFN-7 receptor cDNA clones with no false signals.
Structural Features of the Receptor and Its cDNA-The cloned cDNA was observed to code for a 477-amino acid protein having M, = 52,276. The discrepancy between the size of the cloned product and the reported size of the receptor isolated from membranes is primarily the result of N-linked glycosylation, since incubation of the native receptor with Nglycosidase F decreases the apparent size by approximately 25 kDa (25,38).' The in vitro translation product derived from the cDNA clone was recognized by monoclonal antibody GR-20, but only when synthesized in reticulocyte lysates supplemented with microsomal membranes. The size of the recombinant receptor synthesized under such conditions is consistent with the presence of core glycosylation. Direct evidence that the cDNA codes for the mouse IFN-7 receptor was obtained by transfecting COS-7 monkey cells with a mammalian shuttle vector carrying the full-length cDNA under the direction of the SV40 promoter. The transfectants were observed to specifically bind radiolabeled mouse IFN-7, whereas the host cells showed no binding activity.
The amino acid sequence of the mouse and human IFN-r receptors as deduced from cDNA show considerable structural similarities (see Fig. 4). The two proteins show an overall similarity of 51% with some short stretches exhibiting complete homology. Differences between the two proteins were distributed throughout their entire lengths. Also, some stretches in the human receptor were missing in the mouse sequence. One striking observation concerning the mouse and human receptors is that both are rich in Ser residues with Ser comprising 12% of the residues in both proteins.
Both the human and mouse IFN--y receptors have an NH2terminal signal peptide which is cleaved from the mature protein (22,38, and this work). The mouse receptor contains an eight-amino acid extension of the signal peptide which is not present in the human receptor. A sequence of about 20 hydrophobic amino acids occurs near the center of the mouse receptor, characteristic of a transmembrane region. The sequence of this region is not identical to the corresponding region of the human receptor, although its relative position and hydrophobic nature are conserved. These observations suggest that the IFN-7 receptor is typical of transmembrane proteins which in general are characteristically oriented with the NH,-terminal portion being extracellular and the carboxyl-terminal portion being intracellular. The mouse receptor sequence shows the presence of eight potential asparagine-linked glycosylation sites (Asn-X-Ser/ Thr), five in the extracellular portion and three in the intracellular portion. Four of these potential sites are positionally conserved between the mouse and human receptors. There are also numerous potential sites for O-linked glycosylation considering that the receptors are rich in Ser and Thr residues.
Activation of protein kinase C by the addition of phorbol esters to mouse EL-4 cells is known to result in a down regulation of the IFN--y receptor (43). Additional evidence for the involvement of protein kinase C in the cellular response to IFN-y comes from the finding that addition of a phorbol ester to human macrophage-like cell lines stimulates the IFNy inducible gene 7.1, whereas protein kinase C inhibitors suppress its induction as well as the induction of HLA-DR by . Thus, protein kinase C may directly phosphorylate the IFN-7 receptor. Protein kinase C phosphorylates Ser and Thr residues in an environment in which basic residues are thought to influence the kinetics and specificity of the enzyme (45). Examination of the mouse IFN--r receptor sequence reveals a number of potential acceptor sites for this kinase. In fact, the intracellular portion of the receptor contains the majority of the Ser and Thr with adjacent basic residues. Furthermore, Ser residues in the mouse and human receptors are more highly conserved in the intracellular compared to the extracellular portions. Expression of ZFN--, Receptor mRNA in Various Cell Lines-In general, the IFN-7 receptor is expressed on cell surfaces at a very low levels, only a few thousand molecules per cell (16,(18)(19)(20).
In contrast, only minimal levels of receptor mRNA were seen in cell lines SP2/ 0, P194, BW 52, and Meth A. These observations demonstrate that the neoplastic state itself is not responsible for abnormal expression of the IFN-7 receptor and its mRNA. Nor is there a correlation between the degree of biological response to IFN-7 and the level of the receptor on the cell surface (16) or the level of its mRNA. In the case of ABPL-2 and ABPL-4, the c-myb proto-oncogene transcript is also known to be highly expressed (17,46,47).
Both the IFN-7 receptor and cmyb have been mapped to chromosome 10 (17, 48), although it is not known whether or not the two genes are closely linked. The activation of c-myb occurs as a consequence of the insertion of a defective Moloney leukemia virus into an intron of c-myb (46, 47). Viral integration therefore could be involved in enhanced expression of IFN--/ receptor in these cell lines. Alternatively, c-myb could be acting as a transactivating factor (49), the level of which modulates IFN-7 receptor transcription.
Previously, mouse cells transfected with the human receptor gene were shown to bind to human IFN--y, although in such a heterologous system, they lacked the capacity to elicit the characteristic cellular response (22). This suggests that additional components besides the receptor will need to be identified in order to elucidate the process of signal transduction. This possibility is currently being explored using the mouse system.