Functional Characterization of the Human Interleukin-15 Receptor (cid:97) Chain and Close Linkage of IL15RA and IL2RA Genes*

Interleukins-2 and -15 (IL-2 and IL-15) are cytokines with overlapping but distinct biological effects. Their receptors share two subunits (the IL-2R (cid:98) and - (cid:103) chains) that are essential for signal transduction. The IL-2 re- ceptor requires an additional IL-2-specific (cid:97) subunit for high affinity IL-2 binding. Recently, a murine IL-15-spe- cific (cid:97) subunit was identified, cloned, and shown to be structurally related to IL-2R (cid:97) . However, the murine IL-15R (cid:97) alone bound IL-15 with a 1000-fold higher affinity than that seen with IL-2R (cid:97) and IL-2. We now extend these studies into the human system with the isolation of three differentially spliced human IL-15R (cid:97) allelic nucleotide 627 resulting in an amino acid change between WI-26 clone W5 (ACC/Thr) and A-172 clones A212 and A133 (AAC/ Asn). An independently isolated IL-15R (cid:97) cDNA obtained from a peripheral blood T cell cDNA library sequenced and found to encode ACC/Thr at this position. The peripheral blood T cell isolate also contained the signal peptide coding sequence of clone A212 (data not shown). Reverse transcriptase-PCR analysis was used to determine the representation of the W5 (full-length), A212 (missing exon 3), and A133 (insertion at intron 6) forms of IL-15R (cid:97) in other human mRNA populations, using oligonucleotide primers designed to amplify across the positions of either exon 3 or intron 6. Two amplified bands of the expected sizes ( (cid:54) 99 bp at exon 3; (cid:54) 120 bp at intron 6) were detected in first strand cDNAs generated from peripheral blood T cells, the NK-like YT cell line (34) and the myelogenous leukemia line K-562 (ATCC CCL 243). The same relative intensities of the different forms (A133 (cid:39) W5 (cid:46) A212) was seen in all three samples (data not shown). These results suggest that all three alternatively spliced forms of IL-15R (cid:97) mRNA are in these

The development and function of different cell lineages in multicellular organisms is controlled in large part by response to a number of secreted proteins known as cytokines. A subset of cytokines with related three-dimensional structures, the four helix bundle cytokines, mediate their biological activities through a group of structurally related receptors, the hematopoietin receptor family (1)(2)(3). This cytokine family is characterized by pleiotropic and overlapping biological functions, the latter explained in part by the sharing of common receptor subunits and downstream signal transduction components. Interleukin-2 (IL-2) 1 and interleukin-15 (IL-15) are two helical cytokines with an especially close functional relationship. IL-2, the first identified T cell growth factor, is a product of activated T cells. The IL-2 receptor (IL-2R) is a heterotrimeric complex consisting of three chains (for reviews, see Refs. 4 and 5). The ␣ subunit (IL-2R␣) (6 -8) binds IL-2 with low affinity (K a ϭ 10 8 M

Ϫ1
) but does not contribute directly to signal transduction. The IL-2R␣ is not a member of the hematopoietin receptor family but instead contains two "sushi" domains related to structures found in a number of proteins involved in the complement and coagulation cascades (9). The IL-2R␤ (10) and IL-2R␥ (11) chains are both members of the hematopoietin receptor family and are required for signaling. IL-2 cannot bind to the ␤ or ␥ chains alone but can bind to a complex of ␤ and ␥ with intermediate affinity (K a ϭ 10 9 M Ϫ1 ). Full high affinity binding (K a ϭ 10 11 M Ϫ1 ) requires the additional presence of the ␣ chain. The ␥ chain is noteworthy as a component of the receptors for IL-4, IL-7, and IL-9 and has been referred to as ␥ c , for common ␥ chain (12)(13)(14)(15)(16). It also has been shown to be defective in X-linked severe combined immunodeficiency (17).
IL-15 is a recently discovered cytokine that shares biological activities with IL-2, such as the activation and proliferation of T cells and the costimulation of B cells with CD40 ligand (18,19). Our previous work demonstrated that the IL-2R␤ and -␥ chains are both essential for IL-15 signaling but that IL-2R␣ does not interact with IL-15 (18, 20 -22). Although the human ␤␥ heterodimer of the IL-2R can bind IL-15 (20), it is still unclear whether the ␤␥ heterodimer per se can transduce IL-15 signals in the absence of IL-15R␣. In contrast, we found that murine IL-2R␤ and -␥ cannot bind simian IL-15, although murine T-cell lines can proliferate in response to this cytokine, an observation which led us to the identification and molecular cloning of a murine IL-15R␣ chain (21). The murine IL-15R␣ chain is related to the IL-2R␣ chain and can reconstitute responsiveness of murine cells to simian IL-15 when transfected into cells that express murine IL-2R␤ and -␥, demonstrating that it is indeed the ␣ chain of the IL-15R. Despite the overall similarities of the structures of the receptors for IL-2 and IL-15, there remain two important differences: the IL-15R␣, unlike the IL-2R␣, has a broad distribution of expression in different cell types, and the IL-15R␣ shows a very high affinity of bind-ing to IL-15, more than a 1000-fold higher than the affinity of IL-2R␣ binding to IL-2.
In this paper, we continue the comparison of the IL-2 and IL-15 receptors. We have identified three alternately spliced forms of human IL-15R␣, all of which bind IL-15 with high affinity. Analysis of genomic clones shows that the intron-exon structures of the IL15RA and IL2RA genes are similar and that these loci are closely linked in both human and murine genomes. In addition, as seen in the IL-2/IL-2R system, the cytoplasmic domain of the IL-15R␣ is dispensable for growth signaling, and the human IL-2R ␤␥ heterodimer can transduce IL-15 signals in the absence of IL-15R␣.
Molecular Cloning-The 1.65-kb murine IL-15R␣ cDNA (21) was labeled with 32 P and used to probe a WI-26 VA4 cDNA library prepared in gt10 (24) using standard techniques (25). Briefly, plaque lifts were hybridized with a random prime-labeled full-length murine IL-15R␣ cDNA probe at 42°C in a buffer containing 50% formamide and 5 ϫ SSC and washed at 50°C in 2 ϫ SSC, 0.1% SDS. Additional human IL-15R␣ cDNA clones were subsequently isolated from an A-172 cDNA library by hybridization with pooled plasmid inserts on Southern blots, essentially as described previously (26). The A-172 library was prepared in a plasmid vector from double-stranded cDNA synthesized using random hexamer primers and a commercial cDNA synthesis kit (Pharmacia Biotech Inc.).
DNA Sequencing-DNA sequencing was performed using the Taq DyeDeoxy Terminator Cycle sequencing kit on an automated ABI DNA sequencer (model 373A, Applied Biosystems). In all cases, sequence was obtained from both strands of sample DNAs.
Generation of a Murine IL-15R␣ Cytoplasmic Domain Deletion Mutant-A murine IL-15R␣ cytoplasmic domain deletion mutant was constructed by PCR using a full-length murine IL-15R␣ cDNA template and primers 5Ј-AATAGTCGACATGGCCTCGCCGCAGCTCCG and 5Ј-ACAGCGGCCGCTTACCTTGATTTGATGTACCAGGC, resulting in an amplified fragment encoding a membrane-bound form of the murine IL-15R␣ chain lacking the last 32 amino acids of the predicted cytoplasmic domain. This fragment was subcloned into a mammalian expression vector, its sequence verified by DNA sequencing, and transfected into 32D-01 cells as described below.
Receptor Binding-Receptor binding analysis using 125 I-simian IL-15 was performed as described previously (21). Briefly, cells were incubated at 4°C for 4 -5 h with radiolabeled ligand over a concentration range extending from 0.3 to 400 pM and separated by phthalate oil. To maintain proper conditions of ligand excess, cells were diluted in the incubation mixture with a carrier cell line (B lymphoblast line Daudi (ATCC CCL 213)) that did not bind IL-15.
RNA Analysis-Polyadenylated RNAs were prepared, electrophoresed, and blotted to nylon membranes using standard methods (25). Blots containing mRNAs from various cell lines and tissues were hybridized with radiolabeled RNA probes in a buffer containing 50% formamide at 63°C, washed in 0.1% SSC at 68°C, and autoradiographed using standard methods, as described previously (27).
IL-15R␣ Gene Structure Characterization-A genomic library prepared from human placenta DNA in the bacteriophage vector FixII (Stratagene) was screened by standard methods (25) with a random prime-labeled human IL-15R␣ cDNA probe. Inserts from candidate IL15RA genomic clones were subcloned into pBluescript (Stratagene) and analyzed by hybridization to radiolabeled oligonucleotides specific for 5Ј-non-coding, extracellular, transmembrane, cytoplasmic, and 3Ј-non-coding regions of the human IL-15R␣ cDNA to determine the identity and degree of overlap of the clones isolated. Intron positions in the IL15RA gene were determined by sequencing across intron-exon junctions using human IL-15R␣ cDNA-specific sequencing primers.
Human Chromosomal Localization of IL15RA by Metaphase Fluorescence in Situ Hybridization (FISH)-Two IL15RA genomic clones, each approximately 15 kb in size, with an overlap of approximately 2 kb and encompassing all but the first exon of IL15RA (see below), were chosen for FISH analysis. FISH was performed as described previously (28). Briefly, plasmids containing subcloned IL15RA genomic fragments were nick-translated together using digoxigenin-11-UTP and hybridized overnight at 37°C to fixed metaphase chromosomes prepared from bromodeoxyuridine-synchronized, phytohemagglutin-stimulated peripheral blood lymphocytes obtained from a normal donor. Simultaneous hybridization with a biotin-16-dUTP-labeled chromosome 10-specific ␣ satellite DNA probe (Oncor) was performed to confirm chromosome identity. Specific hybridization signals were detected by applying fluorescein-conjugated sheep antibodies to digoxigenin (Boehringer Mannheim) and avidin Texas Red (Vector Laboratories), followed by counterstaining in 4,6-diamidino-2-phenylindole (Sigma). Fluorescence microscopy was performed with a Nikon microscope equipped with a CytoVision Probe image analysis system and a fluorescence filter wheel, which sequentially captures the fluorescein, Texas Red, and 4,6-diamidino-2-phenylindole images and electronically superimposes them (Applied Imaging, Santa Clara, CA).
A description of two of the probes and restriction fragment length polymorphisms (RFLPs) for the loci linked to Il15ra, including vimentin (Vim) and ␣ spectrin 2, brain (Spna2), has been reported previously (31). Two probes have not been reported previously for our interspecific backcross. The interleukin-2 receptor ␣ (Il2ra) probe, a 1.55-kb PstI fragment of mouse cDNA, detected major TaqI fragments of 6.0 (C57BL/6J) and 4.2 (M. Spretus) kb. The B lymphoma murine Moloney leukemia virus insertion region 1 (Bmi1) probe, a 1-kb BamHI-EcoRI fragment of mouse genomic DNA, detected EcoRI fragments of 6.2 (C57BL/6J) and 12.0 (M. Spretus) kb and MspI fragments of 3.8 (C57BL/6J) and 3.1 (M. Spretus) kb. Again, the presence or absence of M. spretus-specific fragments was followed. The EcoRI and MspI data for Bmi1 were combined. Recombination distances were calculated as described (32) using the computer program Spretus Madness. Gene order was determined by minimizing the number of recombination events required to explain the allele distribution patterns.

Cloning of Alternate Forms of Human IL-15R␣-Binding of 125
I-simian IL-15 was used to identify candidate cell lines for the cloning of a human IL-15R␣ cDNA. Simian IL-15 shares 97 and 73% amino acid identity with human and murine IL-15, respectively, and binds to receptors on both human and murine cells with equivalent affinities (18,20,21,27). Relatively large numbers of high affinity binding sites were detected on the SV40-transformed lung epithelial line WI-26 VA4 (33) (1425 Ϯ 742 sites, K a ϭ 9.8 Ϯ 4.2 ϫ 10 10 M Ϫ1 ; see Fig. 2A) and the glioblastoma line A-172 (ATCC CRL 1620) (2053 Ϯ 916 sites, K a ϭ 5.3 Ϯ 2.6 ϫ 10 10 M Ϫ1 ; data not shown). A WI-26 VA4 cDNA library prepared in the bacteriophage gt10 vector was screened by cross-hybridization with a 32 P-labeled full-length murine IL-15R␣ probe. The insert of a single hybridizing clone (W5) was subcloned, sequenced, and found to share about 65% nucleic acid sequence identity with the murine IL-15R␣ cDNA (Fig. 1A). Clone W5 lacked some of the expected 5Ј sequences (about 125 bp compared to the murine cDNA), indicating that it did not contain the initiator methionine and first portion of the putative IL-15R␣ signal peptide coding region. The only other positive clone isolated from the WI-26 library was identical to clone W5. Since both clones lacked a complete 5Ј-end, a second cDNA library prepared from A-172 cells was screened with the W5 insert to isolate additional human IL-15R␣ cDNA clones. The insert from one clone (A212) contained 129 bp of additional 5Ј sequence compared to clone W5, including a putative initiator Met and full-length signal sequence. The sequence shown in Fig. 1A is a composite of clones A212 and W5 and includes the positions of introns determined from the analysis of genomic IL15RA clones described below. The A212 cDNA was missing exon 3 (99 bp between introns 2 and 3, Fig.  1, A and B). The human IL-15R␣ shares 54% amino acid iden- tity with the murine IL-15R␣, with 82% identity in the putative sushi domain (exon 2, Fig. 1A). A second isolate from the A-172 library, clone A133, contained an incomplete cDNA insert that was missing most of the extracellular coding domain but contained a 120-bp insertion in the cytoplasmic domain predicted to encode an alternate cytoplasmic tail (at the position of intron 6, Fig. 1A). In addition, an allelic difference at nucleotide 627 resulting in an amino acid change was seen between WI-26 clone W5 (ACC/Thr) and A-172 clones A212 and A133 (AAC/ Asn). An independently isolated IL-15R␣ cDNA obtained from a peripheral blood T cell cDNA library was sequenced and found to encode ACC/Thr at this position. The peripheral blood T cell isolate also contained the signal peptide coding sequence of clone A212 (data not shown). Reverse transcriptase-PCR analysis was used to determine the representation of the W5 (full-length), A212 (missing exon 3), and A133 (insertion at intron 6) forms of IL-15R␣ in other human mRNA populations, using oligonucleotide primers designed to amplify across the positions of either exon 3 or intron 6. Two amplified bands of the expected sizes (Ϯ99 bp at exon 3; Ϯ120 bp at intron 6) were detected in first strand cDNAs generated from peripheral blood T cells, the NK-like YT cell line (34) and the myelogenous leukemia line K-562 (ATCC CCL 243). The same relative intensities of the different forms (A133 Ϸ W5 Ͼ A212) was seen in all three samples (data not shown). These results suggest that all three alternatively spliced forms of IL-15R␣ mRNA are present in these cell populations.
Recombinant Human IL-15R␣ Expression and Analysis-The function of the three alternate human IL-15R␣ cDNAs was examined following stable transfection of full-length expression constructs into a murine myeloid cell line. To confirm that the 5Ј-ends of naturally occurring full-length forms of W5 and A133 were the same as that of the A212 isolate, PCR analysis of WI-26 and A-172 first strand cDNAs was performed using primers designed to amplify the entire coding region shown in Fig. 1A. Analysis of the amplified products from these reactions confirmed that the 5Ј-ends of all three forms are identical (data not shown). Since only clone A212 was isolated as a full-length cDNA, full-length forms of clones W5 and A133 were constructed by recombining the appropriate cDNA fragments. To generate a full-length W5 construct (designated h15R) most homologous to the murine IL-15R␣ cDNA, clone A212 sequence upstream of the BstXI site (Fig. 1A) was fused with downstream W5 sequence. An expressible, full-length form of the A133 cDNA (designated h15RaltC) was generated by substituting the A133 sequence downstream of the SphI site (Fig. 1A) with the corresponding region in h15R. The A212 cDNA (designated h15R⌬E3) could be expressed as is. These three forms, shown schematically in Fig. 1B, were subcloned into a mammalian expression vector and transfected into murine myeloid 32D-01 cells alone and in combination with murine IL-2R␤. 32D-01 cells constitutively express murine IL-2R␣ and -␥, but not IL-2R␤, and do not respond to either IL-2 or IL-15 (21). The IL-2R␤ chain has been previously shown to be required for IL-15 responsiveness (18,20,21). We have previously shown that simian IL-15 binds to murine IL-15R␣ with high affinity (K a Ն 10 11 ); however, in the absence of IL-15R␣, it is not capable of binding to murine IL-2R ␤␥ complexes (21). The murine IL-2R␤ chain was therefore used for these transfection experiments because it provides an ideal system to examine the effects of the human IL-15R␣ chains directly. Analysis of IL-15 binding and proliferation in these transfected clones is presented in Table I Fig. 2. When transfected with murine IL-2R␤, 32D-01 cells respond to IL-2 but not to IL-15 (21). All three forms of human IL-15R␣ conferred IL-15 responsiveness as measured by proliferative ability when cotransfected with murine IL-2R␤ (Table I). These results show that the presence of the IL-15R␣ chain is essential for IL-15 responsiveness in these cells. The results also show that exon 3, which is absent in h15R⌬E3, is dispensable for IL-15R␣ binding and function. This reinforces the concept that the sushi structural domain (encoded by exon 2 in IL-15R␣) is the essential element for IL-15 binding. The allelic variation resulting in the amino acid change Thr-152 (in clone h15R and h15RaltC) to Asn-152 (in clones h15R⌬E3) was neutral in regard to IL-15 binding and proliferative responses in transfected 32D-01 cells.
The Cytoplasmic Domain of the IL-15R␣ Chain Is Dispensable for IL-15 Signaling-Since we detected no difference in the signaling ability in cells transfected with h15R or h15RaltC, which encode alternate cytoplasmic domains of human IL-15R␣, we hypothesized that the IL-15R␣ cytoplasmic domain is not essential for IL-15 signaling. To test this possibility, we constructed a deletion mutant of murine IL-15R␣ (m15Rct⌬) lacking the C-terminal 32 amino acids of the predicted cytoplasmic domain and compared the proliferation of 32D-01 cells cotransfected with murine IL-2R␤ and either full-length murine IL-15R␣ or m15Rct⌬. As shown in Fig. 3, no difference in IL-15 responsiveness was detected between cells expressing full-length or truncated murine IL-15R␣, indicating that the murine IL-15R␣ cytoplasmic domain is dispensable for IL-15mediated signaling, as measured by mitogenesis.
The Human IL2R␤ and -␥ Chains Are Sufficient for Simian IL-15 Binding and Signaling-We previously showed that simian IL-15 could bind COS cells transfected with human IL-2R␤ and -␥ chains (20). To address the question of whether the ␤␥ heterodimer can transduce IL-15-induced signals in the absence of IL-15R␣, we utilized the human T lymphocytic leukemia line Jurkat, which is negative for 125 I-IL-15 binding (Table II) and therefore does not express IL-15R␣. These cells also express the IL-2R␥ chain but not the IL-2R␤ chain. Subline JP␤3, derived from JP111-6 following transfection with the human IL-2R␤ chain, showed several hundred IL-15 binding sites with an intermediate binding affinity that was approximately 50fold lower than the observed affinity of IL-15 binding to the IL-15R␣ chain (Table II). To measure the transduction of an IL-15-induced signal in these cells, we examined the phosphorylation of Jak1 kinase in response to 450 ng/ml IL-2 or IL-15 stimulation (Fig. 4). Jak1 tyrosine phosphorylation, which has been shown to be an early event in IL-2 signal transduction (16, Ϫ 32Dm␤-h15RaltC 1.5 Ϯ 0.6 ϫ 10 11 ϩ a Isolation of stably transfected 32D-01 cells, receptor binding analysis, and determination of response to IL-15 were conducted as described under "Experimental Procedures." For each cell type, Scatchard analysis of at least two sets of binding data was performed. b The three alternate forms of human IL-15R␣ were transfected either alone or with murine IL-2R␤ (m␤). [35][36][37], was seen in JP␤3 in response to IL-2 and IL-15 but not in the parental line JP111-6. This demonstrates that a high concentration of simian IL-15, like IL-2, is able to transduce a signal through a complex of human IL-2R␤ and -␥ chains in the absence of its specific receptor ␣ chain.
Differential Distribution of IL-15R␣, IL-2R␣, -␤, and -␥ RNA-A variety of cell lines and tissues was examined for the expression of IL-15R␣ mRNA, as well as for IL-2R␣, -␤ and -␥ mRNA, to assess the relative potential of each cell type to respond to IL-15 or IL-2. An IL-15R␣ mRNA species of about 1.7-1.8 kb was detected in most of the cell lines and all of the tissues examined by Northern blot analysis (Table III). IL-2R␣ expression was seen mainly in mRNA derived from T-cell lines and tissues containing T cells, as well as peripheral blood monocytes and the B lymphoblast line, Raji. IL-2R␤ mRNA was detected in most of the tissue RNAs examined, but among the cell lines examined was limited to T cell line HUT 102, bone marrow stromal line IMTLH, myelogenous leukemia line K-562, and the SV40-transformed lung epithelial cell line WI-26 (Table III). In some of the cell lines and tissues examined, a larger IL-2R␤ transcript was detected, the structure of which is unknown. IL-2R␥ mRNA was abundant in lymphoid and myeloid cell lines, and a low level was detected in prostate, ovary, intestine, and colon. If the expression of the specific receptor ␣ chain for IL-2 or IL-15 is a limiting factor for IL-2 or IL-15 responsiveness in cells that also express the IL-2R␤ and -␥ chains, then these results indicate that a greater variety of the cell lines and tissues examined have the potential to respond to physiological concentrations of IL-15 than to IL-2.
Analysis of IL15RA Genomic Structure-Genomic IL15RA clones were isolated from a human genomic library using a random prime-labeled human IL-15R␣ cDNA (W5) probe. Restriction mapping and probing with individual radiolabeled oligonucleotides specific for 5Ј-non-coding, coding, and 3Ј-noncoding human IL-15R␣ cDNA sequences were used to determine the overlap and extent of the genomic isolates. One isolate that contained an approximately 15-kb insert and hybridized to all but the 5Ј-non-coding region oligonucleotide probe (15Rgen19) was used to map intron positions by sequencing across intron-exon boundaries using IL-15R␣ cDNA-specific sequencing primers. Six introns in the IL15RA gene were identified (Figs. 1A and 5). None of the isolated genomic IL15RA   clones contained exon 1 sequences. We used exon 1-specific primers (dashed arrows, Fig. 1A) and PCR amplification with whole cell human genomic DNA and full-length IL-15R␣ cDNA templates to detect any additional introns upstream of intron 1. A single amplified fragment was seen in all reactions (data not shown), indicating that no additional IL15RA introns upstream of intron 1 were detected in genomic DNA. Since the 5Ј-end of the cloned cDNA may not contain the start of the IL-15R␣ mRNA, the possibility exists for additional exons upstream of the sequence presented in Fig. 1A. Analysis of the intron positions relative to the predicted amino acid sequence shows that introns 1 and 2 delineate the boundaries of the sushi coding domain of IL15RA, and introns 5 and 6 delineate the transmembrane domain (Figs. 1A and 5). The exon-intron boundaries relative to predicted structural domains of IL-15R␣ and IL-2R␣ (shown schematically in Fig. 5) show that the positions of introns 1, 2, 3, 5, and 6 of IL15RA are in close agreement with the positions of introns 1, 2, 3, 6, and 7 previously determined for IL2RA (38,39). Together with the similar roles of the IL-15R␣ and IL-2R␣ in the IL-15 and IL-2 receptor complexes, the similarities in gene organization of IL15RA and IL2RA also suggest close evolutionary relatedness between these two genes.
Human Chromosomal Localization of IL15RA by FISH-To map the IL15RA locus, we performed FISH of metaphase chromosomes prepared from peripheral blood lymphocytes of a normal male donor using IL15RA genomic clones 15Rgen19 (described above) and 15Rgen11, which also contains a 15-kb insert and overlaps about 2 kb of 15Rgen19 (5Ј-end of  15Rgen11, 3Ј-end of 15Rgen19). A representative hybridization of IL15RA clones 15Rgen11 and 15Rgen19 to a met-aphase chromosome preparation is illustrated in Fig. 6. Fluorescence signal observed with these clones was specific to the distal tip of the short arm of a group C chromosome (chromosomes 6 -12 and X). Simultaneous hybridizations of clones 15Rgen11 and 15Rgen19 with ␣ satellite DNA probes specific to the various group C chromosomes identified chromosome 10 as the chromosome containing the IL15RA locus. Specific labeling at the distal tip of the short arm was observed with the IL15RA clones on 83 of 98 chromatids (84.7%) of chromosomes hybridizing with the chromosome 10-specific ␣ satellite DNA probe (25 metaphases analyzed). No other nonrandom signals suggestive of alternative localizations or of cross-hybridization with related sequences were seen in any of the metaphase spreads hybridized. Based upon the presence of the IL15RA-specific hybridization in the extreme distal portion of the short arm of chromosome 10, we have assigned IL15RA to bands p14-p15, which is the same location determined for IL2RA (40). This finding further reinforces the relatedness of these two genes.
Mouse Chromosomal Localization of Il15ra by Interspecific Backcross Mapping-The mouse chromosomal location of Il15ra was determined by interspecific backcross analysis using progeny derived from matings of [(C57BL/6J ϫ M. spretus)F 1 ϫ C57BL/6J] mice. This interspecific backcross mapping panel has been typed for over 1900 loci that are well distributed among all the autosomes as well as the X chromosome (29). C57BL/6J and M. spretus DNAs were digested with several enzymes and analyzed by Southern blot hybridization for informative RFLPs using a mouse cDNA probe. The 20.0-and 8.0-kb SphI M. spretus RFLPs (see "Experimental Procedures") were used to follow the segregation of the Il15ra locus in  backcross mice. The mapping results indicated that Il15ra is located in the proximal region of mouse chromosome 2 linked to Il2ra, Vim, and Spna2. Although 93 mice were analyzed for every marker and are shown in the segregation analysis (Fig. 7), up to 182 mice were typed for some pairs of markers. Each locus was analyzed in pairwise combinations for recombination frequencies using the additional data. The ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analyzed for each pair of loci and the most likely gene order are centromere Il2ra-0/147-Il15ra-1/182-Vim-1/132-Bmi1-5/142-Spna2. The recombination frequencies (expressed as genetic distances in centimorgans Ϯ S.E.) are [Il2ra,Il15ra]-0.6 Ϯ 0.6 -Vim-0.8 Ϯ 0.8 -Bmi1-3.5 Ϯ 1.6 -Spna2. No recombinants were detected between Il2ra and Il15ra in 147 animals typed in common, suggesting that the two loci are within 2.0 centimorgans of each other (upper 95% confidence limit). DISCUSSION IL-15 is a member of a family of molecules with similar predicted three-dimensional structures, the helical cytokine family, whose members interact in some fashion with cell surface molecules that are members of the hematopoietin receptor superfamily (3,41). A common feature of these families is overlapping or redundant functions arising from the shared usage of receptor components and thus, intracellular signaling pathways. The IL-2 and IL-15 systems are examples of this type of relationship: both IL-2 and IL-15 have similar activities and signal through the same receptor complex (the IL-2R␤ and -␥ chains).
Perhaps the most striking difference between the IL-2 and IL-15 systems is the 1000-fold higher affinity of the IL-15R␣ for IL-15 than IL-2R␣ for IL-2. As was shown for murine IL-15R␣ (21), human IL-15R␣ binds IL-15 with the same high affinity (K a Ն 10 11 M Ϫ1 ), whether IL-2R␤ is co-expressed or not ( Fig. 2 and Table I). By contrast, high affinity binding is seen in the IL-2 system only when all three IL-2 receptor chains are present (11). In addition to this difference, the widespread expression patterns of IL-15 (18) and IL-15R␣ (Ref. 21 and Table III) contrast with the narrow expression patterns of IL-2 and IL-2R␣ (reviewed in Ref. 5). The more ubiquitous distribution of IL15RA mRNA may indicate that a wider variety of cell types is responsive to IL-15 than to IL-2. However, since IL-2R␤ and -␥ have been shown to be essential for IL-15 signaling in cells of hematopoietic origin (18, 20 -22) and their expression is more limited than that of IL-15R␣, IL-15R␣ expression alone may not be a sufficient indicator of IL-15 responsiveness. However, it remains a possibility that the IL-15R␣ subunit may associate with as yet unidentified receptor subunits to signal in different cell types. In this regard, the Northern expression data from Table II may suggest that an extremely broad range of cellular targets have the capability of responding to IL-15. Alternatively, widely distributed IL-15R␣ expression in the absence of IL-2R␤ and -␥ may provide a source of cell-associated or soluble receptor to act as a decoy to remove circulating IL-15.
Despite these differences in expression pattern and binding affinity, the IL-2 and IL-15 systems share many similarities. The cytoplasmic domain of IL-15R␣, like that of IL-2R␣, appears to be dispensable for signaling. Indeed, we found that a high concentration of IL-15 can bind and transduce a signal in cells expressing only the IL-2R␤ and -␥ chains, as can IL-2 (42). As shown for the IL-2 system (43), this ability is species-specific in that simian IL-15 measurably binds human but not murine (21) IL-2R ␤␥ complexes, whereas it can clearly bind both human and murine IL-15R␣ chains. Whether concentrations of the cytokine needed for signaling through ␤␥ alone are achieved in vivo remains unknown.
The human IL15RA locus maps to chromosome 10 in bands p14-p15, the same region that contains the IL2RA locus (40). There are only a few human diseases mapped to the 10p14-p15 region that are potential candidates for IL15RA mutation. Rare cases of DiGeorge syndrome, a congenital anomaly characterized by the absence of the thymus and parathyroid glands usually associated with partial monosomy of chromosome 22, have been associated with the deletion of the distal 10p region (44,45), which could encompass the IL15RA locus. Rare but recurrent non-random chromosomal translocations involving the distal 10p region occur in acute lymphocytic and acute myeloid leukemias (t(10;11)(p14;q14-q21) and t(10;17)(p13; q12-q21)) and may be candidates for IL15RA and/or IL2RA involvement (46).
Murine Il15ra mapped to mouse chromosome 2. We have compared our interspecific map of mouse chromosome 2 with a composite mouse linkage map that reports the map location of many uncloned mouse mutations (compiled by M. T. Davisson, T. H. Roderick, A. L. Hillyard, and D. P. Doolittle and provided from GBASE, a computerized data base maintained at The Jackson Laboratory, Bar Harbor, ME). The murine Il15ra gene mapped in a region of the composite map that lacks mouse mutations with a phenotype that might be expected for an alteration in this locus (data not shown). The proximal region of mouse chromosome 2 shares regions of homology with human chromosomes 10p and 9p (summarized in Fig. 7). In particular, IL2RA maps to human 10p15-p14 and Vim to 10p13. The assignment of IL15RA to human 10p15-p14 and Il15ra to mouse 2, tightly linked to Il2ra (47), confirms and extends the region of homology between human 10p and mouse chromosome 2.
The observation that these genes are tightly linked in both mouse and humans suggests that they arose through a random gene duplication event that predated the separation of mouse and man. Consistent with this possibility is the observation that IL2RA and IL15RA are similar both with respect to gene size and organization. Additional studies to determine their physical separation on murine chromosome 2 and human chromosome 10 are warranted to further examine their relationship.