Somatic Cell and Molecular Genetics, Vol. 16, No. 3, 1990, pp. 231-240
Sublocalization on Chromosome 21 of Human Interferon-Alpha Receptor Gene and the Gene for an Interferon-Gamma Response Protein Jerome A. Langer, 1 Abbas Rashidbaigi, 1 Li-Wen Lai, 2 David Patterson, 2'3 and Carol Jones 2'3 Received 9 January 1990
A b s t r a c t - - T h e cellular responses to alpha and beta interferons (IFN-a and -/3) are mediated through the IFN-a/B (type I) receptor, while the response to 1FN-'y is mediated through the IF]V-q/(type II) receptor. The receptors for IFN-oz//~ and IFN-~ are encoded by genes on human chromosomes" 21 and 6q, respectively. The presence of chromosome 21q confers both ligand binding and responsiveness to human IFN-c~/~, whereas chromosome 6q confers binding of Hu-IFN-% but not cellular responsiveness on somatic cell hybrids. Chromosome 6q (i.e., the Hu-IFN-3~ receptor gene) and chromosome 21q are both necessary for the cellular response of somatic cell hybrids (from fibroblasts) to Hu-IFN-% It is conceivable that the factor mediating activity through the IFN-'y receptor is, in fact, the IFN-c~ receptor, or that the two genes are distinct but part of a n "interferon response" region. Here we more precisely localize on human chromosome 21 the genes for the IFN-c~ receptor and for the factor(s) mediating the action of lFN-'y through the chromosome 6-encoded receptor. Hamster-human somatic cell hybrids containing various fragments o f human chromosome 2t were used. The presence of the human IFN-a/{3 receptor was determined by binding 3ep-labeled human IFN-ol to cells, covalently cross-linking the F2p]IFN-a-receptor complex, and analyzing it by S D S polyacrylamide gel electrophoresis. The presence of the IFN-3, receptor-related factor mediating cellular responsiveness was determined by HLA induction in hybrid cells containing the IFN-q/ receptor (chromosome 6q), a transfected copy of the human HLA-B7 gene, and various portions of chromosome 21. In all hybrids examined, the two genes eosegregate. Specifically, both genes are localized to the region of chromosome 21 containing the markers D21S58, D2tS65, and G A R T and appear to be proximal to D21S58. The implications for IFN action are discussed.
INTRODUCTION
The genetics of the interferon (IFN) system includes the study of the IFN genes themselves and of the genes governing inter-
feron responsiveness. The best localized and characterized genes are those of the tFNs themselves and some of the IFN-induced genes (reviewed in reference 1). The response to IFNs is triggered by binding to the IFN
~Department of Molecular Genetics and Microbiology, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, 675 Hoes Lane, Piscataway, New Jersey 08854-5635; 2Eleanor Roosevelt Institute for Cancer Research, 1899 Gaylord Street, Denver, Colorado 80206; and 3University of Colorado Health Sciences Center, Denver, Colorado 231
0740-7750/90/0500-0231506.00/0 © 1990 Plenum Publishing Corporation
232
receptors, of which there appear to be two, the IFN-a/fl receptor, which binds various IFNas and IFN-13, and the IFN-7 receptor, which binds IFN-7. Little is known of the intracellular signaling mechanisms involved in triggering IFN-induced responses, although the response includes the transcriptional activation of a number of genes, some of which have been identified and/or cloned (reviewed in references 1 and 2). The genes for the human and murine IFN-a/13 (type I) receptors have been localized to human chromosome 21 (3-11) and mouse chromosome 16 (12, 13), respectively (reviewed in reference 3). The localization of the human receptor was first deduced from the responsiveness to human leukocyte (a) interferon of mouse-human hybrids containing chromosome 21 (4) or similar mouse-monkey hybrids (5). This conclusion was supported by studies demonstrating chromosome 21 dosage effects for the responsiveness of human cells to various effects of IFN-a (6 and references cited in reference 3) and by the ability of mouse-human hybrid cells containing human chromosome 21 only to elicit antibodies that block several activities of IFN-a and -/3 (7). Direct binding and covalent cross-linking of [125I]Hu-IFN-a to such cells and the ability of the aforementioned antibodies to block binding or to immunoprecipitate the [125I]IFN-areceptor covalent complex provides strong confirmation that the functional effects previously seen in these hybrids reflect the expression of the human IFN-o~ receptor (8, 9). Additional evidence supported sublocalization to 21q (10) and 21q21-qter (11). Thus, human chromosome 21 is necessary and sufficient to confer binding of IFN-a or -~ to cells and to permit a cellular response. The receptor for IFN-7 is distinct from that for IFN-a/t3 (reviewed in reference 1). The human gene was localized to the long arm of chromosome 6 (6q) by analyzing the ability of various mouse-human or hamster-human somatic cell hybrids to specifically bind 32p_ labeled Hu-IFN-'y and to form a specific cova-
Langer et
al.
lent complex (14). This result was confirmed in similar experiments that further localized the gene to the region 6q16-q22 (15) and, more directly, by Southern blot analysis of somatic cell hybrids with the cDNA corresponding to the IFN-7 receptor (16). Somatic cell hybrids carrying only chromosome 6 or 6q do not respond functionally to IFN-7 (14). This observation led to an elegant demonstration by genetic reconstitution in somatic celt hybrids that a gene or genes on chromosome 21q confers IFN-3, responsiveness to cells carrying chromosome 6q (i.e., the IFN-3~ receptor gene) (17, 18). This factor was putatively labeled a transducer for the IFN-7 receptor (17), although other interpretations are possible (V. Jung, personal communication). It had previously been observed that increased sensitivity of aneuploid cells to PHA-induced (i.e., immune or 30 interferon was correlated with the dosage of chromosome 21 (11, 19, 20), from which the authors had suggested that the gene for the IFN-3, receptor might be on chromosome 21 (3). While the data were not unassailable and the interpretation is now clearly incorrect, these data may relate to the second component required in somatic cell hybrids for IFN-3, responsiveness. The genetic and molecular mapping of human chromosome 21 has advanced rapidly because of a combination of improved panels of somatic cell hybrids carrying various translocated or otherwise rearranged fragments of chromosome 21, the application of pulsed-field electrophoresis of large chromosomal restriction fragments, and the mapping of a number of anonymous DNA probes and chromosome 21-specific probes for identified genes (e.g., 21-24). This intense activity stems from the tractability of the mapping effort of chromosome 21, the smallest autosome, and from its medical importance in relation to Down syndrome and familial Alzheimer disease. These recent advances have permitted us to localize with greater accuracy the gene for both the IFN-a receptor
Interferon-Alpha Receptor Gene
233
and the factor related to cellular responsiveness to IFN-7 and to investigate a possible genetic relationship between them. This localization and the relationship of these to other loci are presented here. MATERIALS AND METHODS
Interferons, Radiolabeling, and Interferon Assays. Recombinant Hu-IFN-aA and IFN-7 (specific activity of about 2 x t0 s and 1 x 107 units/ml, respectively), produced in E. coli and purified as previously described (25, 26), were supplied by Dr. Sidney Pestka. Hu-IFN-aA-PI, a derivative of IFN-aA genetically engineered to have a phosphorylation site for cAMP-dependent protein kinase at its carboxyl terminus, was produced and phosphorylated with 32p to high radiospecific activity as described (27). Hu-IFN-aA/ D(Bgl), a genetically engineered hybrid with high biological activity on hamster and mouse cells, was produced as described (28) by R. Ning, S.J. Tarnowski, C.J. Chen, and F. Khan (Hoffmann-LaRoche, Inc.). This can be used in place of hamster IFN-a, which is not available. A cytopathic effect inhibition assay was used to measure the antiviral activities of IFN-aA and IFN-aA-P1 on bovine MDBK cells and IFN-T on human WISH cells, with vesicular stomatitis virus (VSV) (29). 2FURl 21q+
P
8q-
Somatic Cell Hybrids and Cell Lines. Chinese hamster ovary (CHO) hybrid ceil lines 2Furl, 21-SAb5-23a (21q+), 8 q - , R188-1C (1:21), ACEM, 69t8-8al (6918), and 51 t- 1A-6 (5 t 1) carrying various portions of human chromosome 21 have been described (Fig. 1) (21-24). The lines 3X1S and 3X2S were produced by X irradiation of somatic cell hybrids containing chromosome 21 and were rescued by selection for phosphoribosylglycinamide synthetase, and therefore all contain fragments of chromosome 21, including GART (22). 3X1S and 3X2S were derived from parental lines 3X1 and 3X2, and have been recharacterized. These lines were described, and the chromosome 21-specific DNA in various of the cell lines was mapped relative to a number of DNA probes and the break sites of the translocations (21-23). Chromosome 21-specific DNA in the radiation-reduced cell lines and 511-1A-6 have not been mapped cytogenetically. The growth medium for the CHO-K1 (parental) cell line, 2Furl, 21-8Ab523a, 8 q - , 511-1A-6, and R188-1C is Ham's F12 with 5% fetal bovine serum (FBS). Medium for 3X1S, 3X2S, and 6918-8al is Ham's F12D (purine-free medium; Gibco, formula No. 86-5120) supplemented with 5% dialyzed FBS. The human lymphoblastoid Daudi cell line, used in many studies of the human I;21ACEM
6918
SIt BxIS
Sx2S
13 i~2"/7"/227~ 12 11.2 ii.1 II.I 11.2
m i
21 q 22.1
I
22.2 22.3 IFN ALPHA RECEPTOR IFN GfiNMIg RESPONSE
.~ +
4. ÷
-
I NO .I.
ND 4.
÷ +
4. ÷
÷ 4.
.
Fig. I. Schematic d i a g r a m o f the portion o f chromosome 21 contained in members o f the somatic ce|l h y b r i d clone
panel. The results for the I F N - a receptor and IFN-~t. response assays are summarized for each clone panel member. " + " indicates presence of receptor or response, " - " indicates their absence, N D indicates not done.
234
IFN-a receptor, was used as a control cell line for experiments involving the IFN-a receptor. The ability of human IFN-7 to induce MHC class I gene induction in cells containing both the human IFN-7 receptor (encoded on chromosome 6q) and pieces of chromosome 21 was studied using "hybrid hybrids" produced by fusing somatic cell hybrids containing all or part of chromosome 21 (see above) to a cell line (14B7-2) containing the long arm of chromosome 6 (6q) and stably transfected copies of the human HLA-B7 gene and the neomycin resistance gene (18). These hybrids were isolated and grown in F12D supplemented with 5% fetal bovine serum and the antibiotic G418 (300 #g/ml).
Binding and Covalent Cross-linking of F2P]Hu-IFN-aA-P1 to Cells. Binding and covalent cross-linking of [32p]Hu-IFN-aA-P1 to cells were approximately as described (14). Cells were grown to near-confluence in 75-cm 2 tissue culture plates and rinsed twice with 5 ml of Dulbecco's phosphate-buffered saline (PBS) lacking calcium and magnesium. Cells were removed by incubation with 1 ml trypsin/EDTA (Gibco/lx in PBS) or by incubation in 5 ml PBS containing 5 mM Na2EDTA, to determine whether receptor proteolysis was occurring. No difference was seen in the final results for cells treated either way. Following cell detachment, the cell suspension was adjusted to 15 ml with F12 medium supplemented with 5% FBS, and the cells were counted with a hemacytometer. Cells were concentrated by centrifugation at about 500g at room temperature for 10 rain and were then resuspended in complete medium to 0.5-1.0 x 107 cells/ml. [32p]Hu-IFN-aA-P1 (1.2-2.5 x 105 cpm) was added to 0.3-0.5 ml of cells in the presence or absence of 1 ug/ml nonradioactive IFN-aA as competitor, in a 1.5-ml polypropylene tube. Reaction mixtures were incubated for 1-1.5 h at 23°C with gentle rocking. Cells were concentrated by centrifugation (11,000g, 20 sec), washed twice with 1 ml of ice-cold PBS, and suspended again in 0.5 ml cold PBS. For covalent cross-linking, 5 ~1 of a
Langer et a|.
50 mM stock solution of disuccinimidyl suberate (DSS; Pierce Chemical Co.; fleshly dissolved in dimethyl sulfoxide) was added to each reaction, for a final concentration of 0.5 mM DSS. The reaction was continued at 4°C for 20 rain and was quenched by addition of 10 Fzlof 1 M Tris HC1 (pH 7.4). After 10 min at 4°C, the cells were concentrated by centrifugation as above. Each pellet was extracted for 20 rain at 4°C with 100-150 ~l of 1% Triton X-100 in water containing a cocktail of protease inhibitors (30). The suspension was clarified by centrifugation at 11,000g for 20 rain at 4°C. Appropriate volumes of concentrated sample buffer were added to Triton X-100 supernatants for analysis by SDS-polyacrylamide gel electrophoresis. The volume of each sample was adjusted so that roughly equal radioactivity was loaded for each sample. Samples were analyzed by eleetrophoresis overnight on a 1.5-ram-thick 8% polyacrylamide gel. Gels were dried under vacuum and autoradiographed with Kodak X-Omat film at - 170°C with intensifying screens.
Induction of Human Class I Major Histocompatibility Antigens (MHC). The induction of MHC was measured by either or both of two methods. In the first method (17), cells were grown, released by trypsinization, and diluted to 25,000/ml in F12 medium containing 5% FCS. Diluted cells (0.2 ml) were added to wells of 96-well plates and were rocked gently for 15 min. Plates were incubated overnight in a CO 2 incubator. Plates were then incubated for two days in the presence or absence of IFN-7 (103 units/ml) or of IFN-a(A/D) (104 units/rot). Each experiment was done in quadruplicate. Cells were washed by addition of 200 #1 medium per well, and medium was removed. To each well was added 40 ~1 of supernatant of mouse monoclonal anti-MHC class I antibody, W6/ 32. Cells were incubated at 4°C for 1 h, washed with cold medium once with 150 #1 and twice with 200 ~zl. Plates were then incubated with l/SI-labeled anti-mouse immunoglobulin (Amersham, Inc.; 105 cpm/well in
Interferon-Alpha Receptor Gene
50 ~1 medium) for 1 h at 4°C. Plates were washed three times with 150 #1 of cold medium and two more times with 200 ~1. Cells were lysed by incubation at room temperature for 10 rain with 50 ~1 of 1 N N a O H . The tysates were absorbed with cotton swabs. The N a O H treatment and swabbing were repeated. The radioactivity in the two swabs from each well was measured in a Beckman model 300 gamma counter. Alternatively, the induction of M H C was measured by flow cytometry (31). Cells were incubated for 72 h in the absence of interferon or in the presence of Hu-IFN-3, or Hu-IFNc~A/D (1 x 104 units/ml). Cells were harvested, reacted (30 min on ice) with monoclonal antibody specific for human HLA-A, -B, -C (produced by hybridoma No. GM4280, N I G M S , Human Genetic Mutant Cell Repository), and finally reacted with a fluoresceinconjugated goat anti-mouse immunoglobulin (10 ul, 30 rain on ice; Cappel Laboratories, Malvern, Pennsylvania). Cells thus treated were analyzed by flow cytometry (Coulter Epics V model 752 cell sorter equipped with an argon laser). RESULTS
Localization of lFN-o~/fl Receptor. Binding and covalent cross-linking of [32p] Hu-IFNc~A-Pl to various hamster-human somatic cell hybrids containing different segments of human chromosome 21, as well as to the parental CHO-K1 line and the human Daudi cell line, is shown in Fig. 2. The standard 125to 150-kDa radiolabeled band is seen for Daudi cells, and its specificity is demonstrated by the failure to detect such a band when excess nonradioactive IFN-o~A is included in the binding reaction. This control for specificity, testing for the effect of inclusion of nonradioactive IFN-~A on the binding reaction, is included for each ceil line tested. The 125- to 150-kDa band was not detected on parental CHO-K1 cells. This observation is important since human IFN-c~A can elicit a weak antiviral response and the induction of
235
Fig. 2. Detection of the covalent [32P]IFN-c~-receptor complex on CHO-human hybrids containing various fragments of chromosome21. The positionof the covalent complex is denoted by an arrow. For each cell line, the initial binding of [32P]IFN-aA-PI was done in the absence or presence of excess nonradioactive IFN-aA, denotedin the lanes by ( - ) and (+), respectively. M H C in hamster cells, suggesting that it can bind to the hamster IFN-a receptor, albeit less strongly than with the human receptor. Nevertheless, there is no evidence in these binding and cross-linking experiments of specific labeling of the endogenous hamster receptor. For some of the somatic cell hybrid lines (Fig. 2; cell lines 511-1A, 21q+, and 2Furl) a light band of radioactivity is seen at the same position as the [32P]IFN-a/receptor complex on human Daudi cells, thus demonstrating the presence of the human IFN-o~ receptor in these cells. The formation of the t25- to 150-kDa band is inhibited by excess nonradioactive human IFN-aA, demonstrating the specificity of labeling. Results of such experiments with a panel of hybrids are seen in Table 1. Lines 21q+ and 8 q - are essentially complementary segments of chromosome 21, and it was expected that the receptor would be expressed on one but not both of these hybrids. Indeed, 21q+ but not 8 q - demonstrated the presence of the human IFN-c~//3 receptor. The receptor gene was further localized to the region of overlap of the positive lines. Regions 21p and q22.3-qter were immediately elimi-
236
Langer et al.
Hybrids were scored as negative only if they were negative in all experiments. (3) When a hybrid appeared to be negative in two or more experiments, the number of cells for that cell Cell line Hu-IFN-a receptor line and the input [32P]IFN-o~A-P1 was CHO-K1 increased in subsequent experiments in an 21Ab5-23a (21q+) + attempt to increase sensitivity. In no case did 8q2Fuq + an apparently negative cell line change to 511-1A-6 + positive when we increased the number of cells 3X1S + or the amount of [32p]IFN-aA-P1 in the 3X2S 6918-8al + binding reaction. Finally, the consistency of the results as indicating one and only one nated as candidates, and the region for the chromosomal region suggests that false negareceptor was limited to q22.1-q22.2 (Fig. 1). tives were not scored. Localization of Chromosome 21-EnAdditional hybrids, including some produced coded Protein that Confers Responsiveness by X-irradiation reduction, were tested to to IFN-3,. A chromosome 21q-encoded facfurther delineate the chromosomal region tor(s) is required for CHO-human hybrids (Table 1). Of the three additional hybrids containing the IFN-3, receptor (on chromoshown, two (3X1S and 6918) were positive some 6q) to respond to IFN-3, by the induction and one (3X2S) was negative. of class I MHC (17). To further localize this These hybrids and those discussed above factor, hybrids containing various portions of have been mapped with a number of DNA chromosome 21 were hybridized to cells probes (24). The presence of the IFN-c~ containing the human IFN-3, receptor (i.e., receptor in hybrids 511, 6918, and 3X1S human chromosome 6q) and a human HLAsuggests its localization in the same region as D21S58, D21S65, and the GART locus. The B7 gene. The hybrids were tested for inducibilreceptor locus is clearly separated in two cell ity of the HLA-B7 gene by flow cytometry or lines (511 and 6918) from the locus for by binding of 125I-labeled anti-HLA antibodsuperoxide dismutase (SOD) on the cen- ies to cells. Examples of the induction of human tromere-proximal side, and from the ets HLA-B7 in somatic cell hybrids by Hu-IFNoncogene (found in hybrid line 8 q - ) on the % acting through the human receptor on distal side. chromosome 6, or by Hu-IFN-aA/D, acting Because of the weakness of the 125- to through the endogenous hamster IFN-a recep150-kDa radioactive band in the positive tor, are shown (Fig. 3). In two cell lines, 21q+ somatic cell hybrids (Fig. 2), the issue of and 3XIS (panels A and C), both Hu-IFN--y sensitivity is important, particularly the possiand IFN-aA/D were capable of inducing the bility of overlooking positive hybrids, i.e., expression of HLA-B7, seen as an increase in scoring false negatives. The visible, albeit the fluorescence. In contrast, in hybrids weak, bands were clearly specific, as judged derived from 8 q and 3X2S (panels B and by their absence when excess nonradioactive D), IFN-~/ was unable to induce HLA-B7, IFN-c~A was included in the reaction. The accuracy of scoring weak positive bands and suggesting the absence of the putative the problem of sensitivity was addressed in "transducer" or "coupling factor" in these several ways: (1) Each hybrid was used in a lines. In these same lines, however, the minimum of three independent experiments enhanced expression of HLA-B7 was induced over the course of several months. Only by IFN-aA/D, demonstrating the presence hybrids that were positive in all or in three of and inducibility of the human HLA reporter four experiments were scored as positive. (2) gene. The presence of the human IFN-2~ Table 1. Detectionof [32p]Hu_iFN_c~/Receptor Complexon CHO-HumanHybridswithDifferent Regionsof Chromosome21
Interferon-Alpha Receptor Gene
iA
237
~I e----CONTROL +IFN-y
LL]
Z _J Ld
C)
~
CONTROL
LLt
Sublocalization on Chromosome 21 of Human Interferon-Alpha Receptor Gene and the Gene for an Interferon-Gamma Response Protein Jerome A. Langer, 1 Abbas Rashidbaigi, 1 Li-Wen Lai, 2 David Patterson, 2'3 and Carol Jones 2'3 Received 9 January 1990
A b s t r a c t - - T h e cellular responses to alpha and beta interferons (IFN-a and -/3) are mediated through the IFN-a/B (type I) receptor, while the response to 1FN-'y is mediated through the IF]V-q/(type II) receptor. The receptors for IFN-oz//~ and IFN-~ are encoded by genes on human chromosomes" 21 and 6q, respectively. The presence of chromosome 21q confers both ligand binding and responsiveness to human IFN-c~/~, whereas chromosome 6q confers binding of Hu-IFN-% but not cellular responsiveness on somatic cell hybrids. Chromosome 6q (i.e., the Hu-IFN-3~ receptor gene) and chromosome 21q are both necessary for the cellular response of somatic cell hybrids (from fibroblasts) to Hu-IFN-% It is conceivable that the factor mediating activity through the IFN-'y receptor is, in fact, the IFN-c~ receptor, or that the two genes are distinct but part of a n "interferon response" region. Here we more precisely localize on human chromosome 21 the genes for the IFN-c~ receptor and for the factor(s) mediating the action of lFN-'y through the chromosome 6-encoded receptor. Hamster-human somatic cell hybrids containing various fragments o f human chromosome 2t were used. The presence of the human IFN-a/{3 receptor was determined by binding 3ep-labeled human IFN-ol to cells, covalently cross-linking the F2p]IFN-a-receptor complex, and analyzing it by S D S polyacrylamide gel electrophoresis. The presence of the IFN-3, receptor-related factor mediating cellular responsiveness was determined by HLA induction in hybrid cells containing the IFN-q/ receptor (chromosome 6q), a transfected copy of the human HLA-B7 gene, and various portions of chromosome 21. In all hybrids examined, the two genes eosegregate. Specifically, both genes are localized to the region of chromosome 21 containing the markers D21S58, D2tS65, and G A R T and appear to be proximal to D21S58. The implications for IFN action are discussed.
INTRODUCTION
The genetics of the interferon (IFN) system includes the study of the IFN genes themselves and of the genes governing inter-
feron responsiveness. The best localized and characterized genes are those of the tFNs themselves and some of the IFN-induced genes (reviewed in reference 1). The response to IFNs is triggered by binding to the IFN
~Department of Molecular Genetics and Microbiology, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, 675 Hoes Lane, Piscataway, New Jersey 08854-5635; 2Eleanor Roosevelt Institute for Cancer Research, 1899 Gaylord Street, Denver, Colorado 80206; and 3University of Colorado Health Sciences Center, Denver, Colorado 231
0740-7750/90/0500-0231506.00/0 © 1990 Plenum Publishing Corporation
232
receptors, of which there appear to be two, the IFN-a/fl receptor, which binds various IFNas and IFN-13, and the IFN-7 receptor, which binds IFN-7. Little is known of the intracellular signaling mechanisms involved in triggering IFN-induced responses, although the response includes the transcriptional activation of a number of genes, some of which have been identified and/or cloned (reviewed in references 1 and 2). The genes for the human and murine IFN-a/13 (type I) receptors have been localized to human chromosome 21 (3-11) and mouse chromosome 16 (12, 13), respectively (reviewed in reference 3). The localization of the human receptor was first deduced from the responsiveness to human leukocyte (a) interferon of mouse-human hybrids containing chromosome 21 (4) or similar mouse-monkey hybrids (5). This conclusion was supported by studies demonstrating chromosome 21 dosage effects for the responsiveness of human cells to various effects of IFN-a (6 and references cited in reference 3) and by the ability of mouse-human hybrid cells containing human chromosome 21 only to elicit antibodies that block several activities of IFN-a and -/3 (7). Direct binding and covalent cross-linking of [125I]Hu-IFN-a to such cells and the ability of the aforementioned antibodies to block binding or to immunoprecipitate the [125I]IFN-areceptor covalent complex provides strong confirmation that the functional effects previously seen in these hybrids reflect the expression of the human IFN-o~ receptor (8, 9). Additional evidence supported sublocalization to 21q (10) and 21q21-qter (11). Thus, human chromosome 21 is necessary and sufficient to confer binding of IFN-a or -~ to cells and to permit a cellular response. The receptor for IFN-7 is distinct from that for IFN-a/t3 (reviewed in reference 1). The human gene was localized to the long arm of chromosome 6 (6q) by analyzing the ability of various mouse-human or hamster-human somatic cell hybrids to specifically bind 32p_ labeled Hu-IFN-'y and to form a specific cova-
Langer et
al.
lent complex (14). This result was confirmed in similar experiments that further localized the gene to the region 6q16-q22 (15) and, more directly, by Southern blot analysis of somatic cell hybrids with the cDNA corresponding to the IFN-7 receptor (16). Somatic cell hybrids carrying only chromosome 6 or 6q do not respond functionally to IFN-7 (14). This observation led to an elegant demonstration by genetic reconstitution in somatic celt hybrids that a gene or genes on chromosome 21q confers IFN-3, responsiveness to cells carrying chromosome 6q (i.e., the IFN-3~ receptor gene) (17, 18). This factor was putatively labeled a transducer for the IFN-7 receptor (17), although other interpretations are possible (V. Jung, personal communication). It had previously been observed that increased sensitivity of aneuploid cells to PHA-induced (i.e., immune or 30 interferon was correlated with the dosage of chromosome 21 (11, 19, 20), from which the authors had suggested that the gene for the IFN-3, receptor might be on chromosome 21 (3). While the data were not unassailable and the interpretation is now clearly incorrect, these data may relate to the second component required in somatic cell hybrids for IFN-3, responsiveness. The genetic and molecular mapping of human chromosome 21 has advanced rapidly because of a combination of improved panels of somatic cell hybrids carrying various translocated or otherwise rearranged fragments of chromosome 21, the application of pulsed-field electrophoresis of large chromosomal restriction fragments, and the mapping of a number of anonymous DNA probes and chromosome 21-specific probes for identified genes (e.g., 21-24). This intense activity stems from the tractability of the mapping effort of chromosome 21, the smallest autosome, and from its medical importance in relation to Down syndrome and familial Alzheimer disease. These recent advances have permitted us to localize with greater accuracy the gene for both the IFN-a receptor
Interferon-Alpha Receptor Gene
233
and the factor related to cellular responsiveness to IFN-7 and to investigate a possible genetic relationship between them. This localization and the relationship of these to other loci are presented here. MATERIALS AND METHODS
Interferons, Radiolabeling, and Interferon Assays. Recombinant Hu-IFN-aA and IFN-7 (specific activity of about 2 x t0 s and 1 x 107 units/ml, respectively), produced in E. coli and purified as previously described (25, 26), were supplied by Dr. Sidney Pestka. Hu-IFN-aA-PI, a derivative of IFN-aA genetically engineered to have a phosphorylation site for cAMP-dependent protein kinase at its carboxyl terminus, was produced and phosphorylated with 32p to high radiospecific activity as described (27). Hu-IFN-aA/ D(Bgl), a genetically engineered hybrid with high biological activity on hamster and mouse cells, was produced as described (28) by R. Ning, S.J. Tarnowski, C.J. Chen, and F. Khan (Hoffmann-LaRoche, Inc.). This can be used in place of hamster IFN-a, which is not available. A cytopathic effect inhibition assay was used to measure the antiviral activities of IFN-aA and IFN-aA-P1 on bovine MDBK cells and IFN-T on human WISH cells, with vesicular stomatitis virus (VSV) (29). 2FURl 21q+
P
8q-
Somatic Cell Hybrids and Cell Lines. Chinese hamster ovary (CHO) hybrid ceil lines 2Furl, 21-SAb5-23a (21q+), 8 q - , R188-1C (1:21), ACEM, 69t8-8al (6918), and 51 t- 1A-6 (5 t 1) carrying various portions of human chromosome 21 have been described (Fig. 1) (21-24). The lines 3X1S and 3X2S were produced by X irradiation of somatic cell hybrids containing chromosome 21 and were rescued by selection for phosphoribosylglycinamide synthetase, and therefore all contain fragments of chromosome 21, including GART (22). 3X1S and 3X2S were derived from parental lines 3X1 and 3X2, and have been recharacterized. These lines were described, and the chromosome 21-specific DNA in various of the cell lines was mapped relative to a number of DNA probes and the break sites of the translocations (21-23). Chromosome 21-specific DNA in the radiation-reduced cell lines and 511-1A-6 have not been mapped cytogenetically. The growth medium for the CHO-K1 (parental) cell line, 2Furl, 21-8Ab523a, 8 q - , 511-1A-6, and R188-1C is Ham's F12 with 5% fetal bovine serum (FBS). Medium for 3X1S, 3X2S, and 6918-8al is Ham's F12D (purine-free medium; Gibco, formula No. 86-5120) supplemented with 5% dialyzed FBS. The human lymphoblastoid Daudi cell line, used in many studies of the human I;21ACEM
6918
SIt BxIS
Sx2S
13 i~2"/7"/227~ 12 11.2 ii.1 II.I 11.2
m i
21 q 22.1
I
22.2 22.3 IFN ALPHA RECEPTOR IFN GfiNMIg RESPONSE
.~ +
4. ÷
-
I NO .I.
ND 4.
÷ +
4. ÷
÷ 4.
.
Fig. I. Schematic d i a g r a m o f the portion o f chromosome 21 contained in members o f the somatic ce|l h y b r i d clone
panel. The results for the I F N - a receptor and IFN-~t. response assays are summarized for each clone panel member. " + " indicates presence of receptor or response, " - " indicates their absence, N D indicates not done.
234
IFN-a receptor, was used as a control cell line for experiments involving the IFN-a receptor. The ability of human IFN-7 to induce MHC class I gene induction in cells containing both the human IFN-7 receptor (encoded on chromosome 6q) and pieces of chromosome 21 was studied using "hybrid hybrids" produced by fusing somatic cell hybrids containing all or part of chromosome 21 (see above) to a cell line (14B7-2) containing the long arm of chromosome 6 (6q) and stably transfected copies of the human HLA-B7 gene and the neomycin resistance gene (18). These hybrids were isolated and grown in F12D supplemented with 5% fetal bovine serum and the antibiotic G418 (300 #g/ml).
Binding and Covalent Cross-linking of F2P]Hu-IFN-aA-P1 to Cells. Binding and covalent cross-linking of [32p]Hu-IFN-aA-P1 to cells were approximately as described (14). Cells were grown to near-confluence in 75-cm 2 tissue culture plates and rinsed twice with 5 ml of Dulbecco's phosphate-buffered saline (PBS) lacking calcium and magnesium. Cells were removed by incubation with 1 ml trypsin/EDTA (Gibco/lx in PBS) or by incubation in 5 ml PBS containing 5 mM Na2EDTA, to determine whether receptor proteolysis was occurring. No difference was seen in the final results for cells treated either way. Following cell detachment, the cell suspension was adjusted to 15 ml with F12 medium supplemented with 5% FBS, and the cells were counted with a hemacytometer. Cells were concentrated by centrifugation at about 500g at room temperature for 10 rain and were then resuspended in complete medium to 0.5-1.0 x 107 cells/ml. [32p]Hu-IFN-aA-P1 (1.2-2.5 x 105 cpm) was added to 0.3-0.5 ml of cells in the presence or absence of 1 ug/ml nonradioactive IFN-aA as competitor, in a 1.5-ml polypropylene tube. Reaction mixtures were incubated for 1-1.5 h at 23°C with gentle rocking. Cells were concentrated by centrifugation (11,000g, 20 sec), washed twice with 1 ml of ice-cold PBS, and suspended again in 0.5 ml cold PBS. For covalent cross-linking, 5 ~1 of a
Langer et a|.
50 mM stock solution of disuccinimidyl suberate (DSS; Pierce Chemical Co.; fleshly dissolved in dimethyl sulfoxide) was added to each reaction, for a final concentration of 0.5 mM DSS. The reaction was continued at 4°C for 20 rain and was quenched by addition of 10 Fzlof 1 M Tris HC1 (pH 7.4). After 10 min at 4°C, the cells were concentrated by centrifugation as above. Each pellet was extracted for 20 rain at 4°C with 100-150 ~l of 1% Triton X-100 in water containing a cocktail of protease inhibitors (30). The suspension was clarified by centrifugation at 11,000g for 20 rain at 4°C. Appropriate volumes of concentrated sample buffer were added to Triton X-100 supernatants for analysis by SDS-polyacrylamide gel electrophoresis. The volume of each sample was adjusted so that roughly equal radioactivity was loaded for each sample. Samples were analyzed by eleetrophoresis overnight on a 1.5-ram-thick 8% polyacrylamide gel. Gels were dried under vacuum and autoradiographed with Kodak X-Omat film at - 170°C with intensifying screens.
Induction of Human Class I Major Histocompatibility Antigens (MHC). The induction of MHC was measured by either or both of two methods. In the first method (17), cells were grown, released by trypsinization, and diluted to 25,000/ml in F12 medium containing 5% FCS. Diluted cells (0.2 ml) were added to wells of 96-well plates and were rocked gently for 15 min. Plates were incubated overnight in a CO 2 incubator. Plates were then incubated for two days in the presence or absence of IFN-7 (103 units/ml) or of IFN-a(A/D) (104 units/rot). Each experiment was done in quadruplicate. Cells were washed by addition of 200 #1 medium per well, and medium was removed. To each well was added 40 ~1 of supernatant of mouse monoclonal anti-MHC class I antibody, W6/ 32. Cells were incubated at 4°C for 1 h, washed with cold medium once with 150 #1 and twice with 200 ~zl. Plates were then incubated with l/SI-labeled anti-mouse immunoglobulin (Amersham, Inc.; 105 cpm/well in
Interferon-Alpha Receptor Gene
50 ~1 medium) for 1 h at 4°C. Plates were washed three times with 150 #1 of cold medium and two more times with 200 ~1. Cells were lysed by incubation at room temperature for 10 rain with 50 ~1 of 1 N N a O H . The tysates were absorbed with cotton swabs. The N a O H treatment and swabbing were repeated. The radioactivity in the two swabs from each well was measured in a Beckman model 300 gamma counter. Alternatively, the induction of M H C was measured by flow cytometry (31). Cells were incubated for 72 h in the absence of interferon or in the presence of Hu-IFN-3, or Hu-IFNc~A/D (1 x 104 units/ml). Cells were harvested, reacted (30 min on ice) with monoclonal antibody specific for human HLA-A, -B, -C (produced by hybridoma No. GM4280, N I G M S , Human Genetic Mutant Cell Repository), and finally reacted with a fluoresceinconjugated goat anti-mouse immunoglobulin (10 ul, 30 rain on ice; Cappel Laboratories, Malvern, Pennsylvania). Cells thus treated were analyzed by flow cytometry (Coulter Epics V model 752 cell sorter equipped with an argon laser). RESULTS
Localization of lFN-o~/fl Receptor. Binding and covalent cross-linking of [32p] Hu-IFNc~A-Pl to various hamster-human somatic cell hybrids containing different segments of human chromosome 21, as well as to the parental CHO-K1 line and the human Daudi cell line, is shown in Fig. 2. The standard 125to 150-kDa radiolabeled band is seen for Daudi cells, and its specificity is demonstrated by the failure to detect such a band when excess nonradioactive IFN-o~A is included in the binding reaction. This control for specificity, testing for the effect of inclusion of nonradioactive IFN-~A on the binding reaction, is included for each ceil line tested. The 125- to 150-kDa band was not detected on parental CHO-K1 cells. This observation is important since human IFN-c~A can elicit a weak antiviral response and the induction of
235
Fig. 2. Detection of the covalent [32P]IFN-c~-receptor complex on CHO-human hybrids containing various fragments of chromosome21. The positionof the covalent complex is denoted by an arrow. For each cell line, the initial binding of [32P]IFN-aA-PI was done in the absence or presence of excess nonradioactive IFN-aA, denotedin the lanes by ( - ) and (+), respectively. M H C in hamster cells, suggesting that it can bind to the hamster IFN-a receptor, albeit less strongly than with the human receptor. Nevertheless, there is no evidence in these binding and cross-linking experiments of specific labeling of the endogenous hamster receptor. For some of the somatic cell hybrid lines (Fig. 2; cell lines 511-1A, 21q+, and 2Furl) a light band of radioactivity is seen at the same position as the [32P]IFN-a/receptor complex on human Daudi cells, thus demonstrating the presence of the human IFN-o~ receptor in these cells. The formation of the t25- to 150-kDa band is inhibited by excess nonradioactive human IFN-aA, demonstrating the specificity of labeling. Results of such experiments with a panel of hybrids are seen in Table 1. Lines 21q+ and 8 q - are essentially complementary segments of chromosome 21, and it was expected that the receptor would be expressed on one but not both of these hybrids. Indeed, 21q+ but not 8 q - demonstrated the presence of the human IFN-c~//3 receptor. The receptor gene was further localized to the region of overlap of the positive lines. Regions 21p and q22.3-qter were immediately elimi-
236
Langer et al.
Hybrids were scored as negative only if they were negative in all experiments. (3) When a hybrid appeared to be negative in two or more experiments, the number of cells for that cell Cell line Hu-IFN-a receptor line and the input [32P]IFN-o~A-P1 was CHO-K1 increased in subsequent experiments in an 21Ab5-23a (21q+) + attempt to increase sensitivity. In no case did 8q2Fuq + an apparently negative cell line change to 511-1A-6 + positive when we increased the number of cells 3X1S + or the amount of [32p]IFN-aA-P1 in the 3X2S 6918-8al + binding reaction. Finally, the consistency of the results as indicating one and only one nated as candidates, and the region for the chromosomal region suggests that false negareceptor was limited to q22.1-q22.2 (Fig. 1). tives were not scored. Localization of Chromosome 21-EnAdditional hybrids, including some produced coded Protein that Confers Responsiveness by X-irradiation reduction, were tested to to IFN-3,. A chromosome 21q-encoded facfurther delineate the chromosomal region tor(s) is required for CHO-human hybrids (Table 1). Of the three additional hybrids containing the IFN-3, receptor (on chromoshown, two (3X1S and 6918) were positive some 6q) to respond to IFN-3, by the induction and one (3X2S) was negative. of class I MHC (17). To further localize this These hybrids and those discussed above factor, hybrids containing various portions of have been mapped with a number of DNA chromosome 21 were hybridized to cells probes (24). The presence of the IFN-c~ containing the human IFN-3, receptor (i.e., receptor in hybrids 511, 6918, and 3X1S human chromosome 6q) and a human HLAsuggests its localization in the same region as D21S58, D21S65, and the GART locus. The B7 gene. The hybrids were tested for inducibilreceptor locus is clearly separated in two cell ity of the HLA-B7 gene by flow cytometry or lines (511 and 6918) from the locus for by binding of 125I-labeled anti-HLA antibodsuperoxide dismutase (SOD) on the cen- ies to cells. Examples of the induction of human tromere-proximal side, and from the ets HLA-B7 in somatic cell hybrids by Hu-IFNoncogene (found in hybrid line 8 q - ) on the % acting through the human receptor on distal side. chromosome 6, or by Hu-IFN-aA/D, acting Because of the weakness of the 125- to through the endogenous hamster IFN-a recep150-kDa radioactive band in the positive tor, are shown (Fig. 3). In two cell lines, 21q+ somatic cell hybrids (Fig. 2), the issue of and 3XIS (panels A and C), both Hu-IFN--y sensitivity is important, particularly the possiand IFN-aA/D were capable of inducing the bility of overlooking positive hybrids, i.e., expression of HLA-B7, seen as an increase in scoring false negatives. The visible, albeit the fluorescence. In contrast, in hybrids weak, bands were clearly specific, as judged derived from 8 q and 3X2S (panels B and by their absence when excess nonradioactive D), IFN-~/ was unable to induce HLA-B7, IFN-c~A was included in the reaction. The accuracy of scoring weak positive bands and suggesting the absence of the putative the problem of sensitivity was addressed in "transducer" or "coupling factor" in these several ways: (1) Each hybrid was used in a lines. In these same lines, however, the minimum of three independent experiments enhanced expression of HLA-B7 was induced over the course of several months. Only by IFN-aA/D, demonstrating the presence hybrids that were positive in all or in three of and inducibility of the human HLA reporter four experiments were scored as positive. (2) gene. The presence of the human IFN-2~ Table 1. Detectionof [32p]Hu_iFN_c~/Receptor Complexon CHO-HumanHybridswithDifferent Regionsof Chromosome21
Interferon-Alpha Receptor Gene
iA
237
~I e----CONTROL +IFN-y
LL]
Z _J Ld
C)
~
CONTROL
LLt
