Cloning of a Novel Receptor Subunit, AcPL, Required for Interleukin-18 Signaling*

We have identified a novel member of the interleukin-1 (IL-1) receptor family, which we have termed AcPL. In transient transfection assays, we were unable to demonstrate a role for AcPL in IL-1-induced activation of NFκB. Interleukin-18 (interferon-γ-inducing factor) is another member of the IL-1 family of cytokines, and it has recently been shown that IL-18 has a weak affinity for IL-1R-rp1. We examined whether AcPL might function alone or in concert with IL-1R-rp1 to mediate IL-18 signaling. We found that both IL-1R-rp1 and AcPL expression were required for induction of NFκB activity and for activation of c-Jun N-terminal kinase in response to IL-18. Furthermore, a dominant negative version of AcPL specifically inhibited IL-18 signaling. In vitroimmunoprecipitation assays demonstrated that AcPL alone was unable to bind IL-18 with any appreciable affinity. We propose that although IL-1R-rp1 binds the cytokine, IL-1R-rp1 and AcPL proteins are both required for IL-18 signaling, analogous to the requirement for both IL-1R and IL-1RAcP in IL-1-mediated responses.

Interleukin-18 (interferon-␥-inducing factor/IL-1␥) 1 has a wide range of immunoregulatory functions, including stimulation of interferon-␥ production, induction of natural killer (NK) cell cytotoxicity, potentiation of Th1 differentiation, and inhibition of osteoclast proliferation (1)(2)(3)(4). IL-18-deficient mice display a phenotype largely similar to that of IL-12 knockout mice, exhibiting reductions in interferon-␥ production, NK cell activity, and Th1 response. Interestingly, IL-18 and IL-12 doubledeficient mice displayed an even greater perturbation of NK cell activity and Th1 cell response than that seen in either of the single knockouts (5). In terms of its biological effects, IL-18 is thus closely related to and acts synergistically with IL-12. Analysis of amino acid sequence and structural motifs, however, classify IL-18 as a member of the IL-1 family of cytokines (6). In agreement with this classification, IL-18 has been shown to be processed by the interleukin-1␤-converting enzyme, which also processes IL-1␤ (7,8). Furthermore, IL-18 signaling was shown to be mediated by a member of the IL-1R family, IL-1R-rp1 (9). 2 It was also recently shown that IL-18 can activate IRAK, recruit TRAF6, and induce translocation of NFB (3,11), 2 which are all involved in IL-1 signaling (12)(13)(14).
There are several members of the IL-1R family, many of which are currently orphan receptors. The type I and type II IL-1 receptors bind IL-1-␣, IL-1␤, and IL-1ra. The type II IL-1 receptor lacks a cytoplasmic domain, thus rendering it a decoy receptor that has been shown to down-modulate IL-1 responses (15). Signaling by IL-1 is dependent not only on expression of IL-1R type I but also on expression of the IL-1RAcP (16,17). Although IL-1RAcP does not bind IL-1 directly, it does increase the affinity of IL-1R for cytokine binding. Furthermore, the IL-1RAcP in the active receptor complex is responsible for recruitment of the IL-1R-activated kinase (IRAK) (18). Two other members of the family, IL-1R-rp1 and T1/ST2, are unable to bind IL-1 but do respond to IL-1 stimulation in reporter assays when expressed in chimeric form with IL-1R extracellular and transmembrane regions fused to the cytoplasmic domains of these receptors (19 -22). IL-1R-rp2 is another orphan family member, incapable of IL-1 binding (23). Interestingly, IL-1R type I, IL-1R type II, IL-1R-rp1, and T1/ST2 all map to the same region of chromosome 2, indicating that perhaps these receptors arise from a common ancestral gene (22,24).
A recent search of the expressed sequence tag data base revealed a sequence with homology to IL-1RAcP. Reasoning that this might be a new member of the IL-1R family, we utilized the corresponding IMAGE clone to screen a cDNA library to obtain full-length sequence. Analysis of the entire open reading frame verifies that this is a novel member of the IL-1R family, with the IL-1R hallmark domains showing high conservation. Based on its homology to IL-1RAcP we have termed this new protein AcPL (for accessory protein-like). AcPL was unable to mediate signaling by IL-1 but did play a role in signaling via the structurally related cytokine IL-18. Although AcPL did not bind IL-18 in vitro, coexpression of IL-1R-rp1 and AcPL was required for IL-18 responsiveness in terms of NFB induction and JNK activation. Furthermore, a mutant version of AcPL lacking the cytoplasmic domain inhibited IL-18 signaling. We therefore propose that both the previously identified IL-1R-rp1 and the novel AcPL are involved in signaling by IL-18.

EXPERIMENTAL PROCEDURES
Cloning and Mapping of AcPL-Expressed sequence tag mu27d04.r1 (GenBank TM accession number AA203986) was obtained from the IM-AGE consortium, and the insert was labeled with 32 P by random priming and used to probe an EL46.1 (mouse thymocyte) cDNA library. Hybridization was carried out at 42°C in hybridization solution containing 50% formamide. After the full-length open reading frame was defined by an EL46.1 cDNA clone, it was verified by obtaining independent isolates from 7B9 (mouse T cell) and LDA11 (mouse bone * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF077347 and AF077346.
A human cDNA clone, QQ1352, randomly sequenced at Immunex from an NK cell library had a high degree of homology to murine AcPL and was used as a probe to isolate human AcPL clones from peripheral blood lymphocyte, peripheral blood T cell, and NK cDNA libraries. The region of clone QQ1352 used as a probe was homologous to murine AcPL nucleotides 1196 -1753. A full-length clone was not obtained from any of the libraries, so vector-anchored PCR was carried out in each of the libraries to obtain the 5Ј end of the open reading frame. As for the murine sequence, the reported sequence represents sequence from at least three independent isolates over the entire open reading frame.
The chromosome map position of human AcPL was determined by radiation hybrid mapping using the Stanford G3 Radiation Hybrid Panel (Research Genetics). The primers used, by homology to IL-1R (24), corresponded to intron 4, 5Ј-CACATCATTCAGGACAAATG-TACCC-3Ј, and exon 5, 5Ј-CTAAAATCATCTTGACACAACAGGC-3Ј. Amplification was carried out under standard PCR conditions for 40 cycles.
Northern Analysis-A human multiple tissue blot was purchased from CLONTECH Laboratories, Inc. and contained 2 g of mRNA from normal human spleen, thymus, prostate, testis, ovary, small intestine, colon, and peripheral blood leukocyte. This was hybridized overnight with a 32 P-labeled antisense hAcPL riboprobe in hybridization buffer containing 50% formamide at 63°C and then washed at 68°C in 0.1ϫ SSC/0.1% SDS. After exposure, the blot was rehybridized with a random-prime labeled probe against ␤-actin for standardization.
Plasmids and Cell Culture-For expression, full-length murine and human AcPL were generated by PCR and cloned into pDC304, a variant of pDC302 (25). The human AcPL-Fc expression vector joins the extracellular portion of the receptor (amino acids 1-356) to the CH2 and CH3 domains of human IgG1 and was generated as described previously (26). Expression vectors encoding IL-1R-rp1 and IL-1R-rp1-Fc have been described previously (22). mAcPL⌬cyto contains Thr 17 to Leu 402 of mAcPL, with an immunoglobulin light chain signal peptide substituted for the native signal peptide, followed by a modified FLAG (27) epitope tag (METDTLLLWVLLWVPGSTGDYKDEGTS). mIL-1R-rp1⌬cyto contains Met 1 to Gly 368 with the native signal peptide and no epitope tag. The NFB-Luciferase plasmid has been described previously (28) and contains three NFB sites and a minimal c-Fos promoter driving expression of the luciferase gene. IL8p-Luc contains the previously described hIL8 promoter (29) subcloned into pGL2-Basic (Promega). The human IL-18 sequence was cloned into pDC206 (30) for transient overexpression.
Reporter and JNK Assays-To assess NFB activation, Cos7 cells were transiently transfected by the DEAE-dextran method as described (31), using 150 ng of each receptor and 700 ng of the reporter plasmid (unless otherwise noted)/4.5 ϫ 10 4 cells. Two days post-transfection, cells were stimulated with 10 ng/ml IL-1 or 40 ng/ml IL-18 (PeproTech, Inc.) for 4 h. Cells were lysed and luciferase activity assessed using Reporter Lysis Buffer and Luciferase Assay Reagent (Promega Corp.). S49.1 cells (1 ϫ 10 7 ) were electroporated (320 V, 960 microfarad) with 20 g of reporter DNA and 5 g of each receptor-encoding DNA. After 2 days cells were stimulated and luciferase levels were measured as described above.
Activation of JNK activity was assessed as described previously (32). Briefly, 2 days post-transfection Cos7 cells were stimulated with IL-18 for 15 min, lysed, and immunoprecipitated with a combination of two anti-JNK antibodies (C-17 and FL, Santa Cruz Biotechnology, Inc). This immunocomplex was assayed for activity by addition of glutathione S-transferase-c-Jun (Upstate Biotechnology, Inc.) and [␥-32 P]ATP in kinase buffer. The reaction was allowed to proceed for 30 min at room temperature, after which Laemmli Loading buffer was added to stop the reaction, and products were electrophoresed on a 4 -20% acrylamide gel, stained, dried, and analyzed on a STORM PhosphorImager.

RESULTS
Cloning of AcPL-A search of the expressed sequence tag data base revealed that IMAGE clone 640615 (GenBank TM Accession number AA203986) had homology to IL-1RAcP. We used this IMAGE clone to screen an EL46.1 cDNA library to isolate the full-length open reading frame. Subsequently, fulllength sequence was also isolated from two other murine T cell libraries by PCR using gene-specific primers. We have therefore verified that the same sequence was obtained from three different library sources. The murine sequence was highly homologous to a human NK cell clone, QQ1352, randomly sequenced at Immunex. Clone QQ1352 was therefore used to probe peripheral blood T cell, peripheral blood lymphocyte, and NK cell libraries, and the full-length human sequence was obtained from all three libraries. The predicted amino acid sequence of both the murine and human forms of AcPL is presented in Fig. 1. Murine and human forms of AcPL share 65% identity. Overall, AcPL shows 25% identity to IL-1R type I, 27% identity to IL-1RAcP, and 26% identity to IL-1R-rp1. A BLAST (33) search of the data base does not reveal any other proteins with significant homology to AcPL besides the noted IL-1R family members.
Similar to other members of the IL-1R family, hAcPL is predicted to contain a signal peptide (14 amino acids), a 342amino acid extracellular segment, a single transmembrane region, and a 222 amino acid cytoplasmic domain. The extracellular region classifies this receptor as a member of the immunoglobulin superfamily. In comparison with type I IL-1R, Ig domain 1 is poorly conserved, yet domains 2 and 3 are highly conserved in AcPL. There are three potential glycosylation sites that are conserved in both the murine and human sequence (as well as one in each species that is not conserved). The cytoplasmic domain of hAcPL is 32% identical to that of hIL-1R.
By Northern blot analysis, human AcPL was expressed strongly in peripheral blood leukocytes and spleen and to a lesser extent in colon. Weak expression was detected in prostate and small intestine mRNA (which may not be visible on reproduction). The predominant mRNA product was approximately 3.8 kilobases, with minor binds at approximately 2.6 and 8.0 kilobases (Fig. 2). Expression was also detected in lung mRNA by Northern analysis (data not shown). No expression was detected in thymus, prostate, testis, ovary, or small intestine (Fig. 2) nor in heart, brain, kidney, or muscle mRNA (data not shown). In agreement with these results we were unable to detect expression of AcPL in human fetal brain, murine heart, or murine brain libraries by PCR.
Radiation hybrid analysis using the Stanford G3 panel placed human AcPL on chromosome 2, most closely linked to AFM316tg5, with a logarithm of odds score of 12.72 (data not shown). This is the same region of chromosome 2 to which IL-1R type I, IL-1R type II, IL-1R-rp1, and T1/ST2 have been mapped (Stanford Human Genome Center bin 57) (22,24).
AcPL Is Not Involved in IL-1 Signaling-Because the extracellular domain of AcPL was homologous to IL-1R, we first tested its ability to induce signaling in response to IL-1 stimulation. Cos7 cells were transiently transfected with full-length murine AcPL and a luciferase reporter harboring three NFB sites. Cells were stimulated with 10 ng/ml IL-1␣ and assayed for luciferase activity. As shown in Fig. 3A, cells transfected with empty vector or IL-1RAcP exhibited a reporter induction in response to IL-1␣ stimulation, due to endogenous monkey IL-1R expression. Overexpression of IL-1R, however, resulted in a marked enhancement of luciferase induction over background. In contrast, transfection of AcPL did not augment the background response, suggesting that it is incapable of mediating IL-1 signaling. Similar results were obtained with IL-1␤ stimulation and with an IL-8 promoter-containing reporter.
It is possible that AcPL could function in IL-1 signaling in a role similar to IL-1RAcP, by complementing the IL-1R for IL-1 responsiveness. Due to endogenous IL-1RAcP expression in Cos7 cells, such a function for AcPL would not be uncovered in the above transfection assays. We have characterized a T cell line, S49.1, which is nonresponsive to IL-1 when electroporated with an NFB-driven reporter but is responsive to IL-1 when IL-1R and IL-1RAcP are coexpressed. Using this system, we investigated whether AcPL could replace IL-1R or IL-1RAcP to produce an NFB response to IL-1. We found that replacement of either IL-1R or IL-1RAcP with AcPL abolished responsiveness (Fig. 3B). Taken together with the results in Cos7 cells, we conclude that AcPL does not mediate IL-1 signaling.
AcPL Is Required for IL-18 Signaling-IL-18 has been classified as a member of the IL-1 family of cytokines based on amino acid sequence and structural homology. It has recently been shown that overexpression of hIL-1R-rp1 in Cos1 cells resulted in increased IL-18 binding and conferred an NFB induction in response to IL-18 (9). In our hands, however, Cos1 or Cos7 cells overexpressing IL-1R-rp1 are nonresponsive to IL-18, suggesting the requirement for another receptor chain, which is expressed in the cells used by Torigoe et al. (9), but not expressed in our Cos cells. 2 Therefore, we tested whether AcPL could function alone or as a second subunit in IL-18 responses by performing transient transfections in Cos7 cells. As shown in Fig. 4A, NFB activation by mIL-18 required the expression of both mAcPL and mIL-1R-rp1 in these cells. We observed no NFB induction in control, mIL-1R-rp1, or mAcPL-transfected cells stimulated with IL-18. In contrast, coexpression of mIL-1R-rp1 and mAcPL resulted in elevated levels of both basal and cytokine-stimulated NFB activity. A control reporter lacking the NFB sites was not activated in these assays (data not shown). Elevation of basal NFB activity has also been reported upon IL-1R and IL-1RAcP overexpression and probably reflects the formation of receptor heterodimers in the absence of ligand due to the high density of surface expression (18). This is supported by the fact that basal reporter activity is not elevated in Cos7 cells transfected with lower amounts (10 ng each as opposed to 150 ng each) of IL-1R-rp1 and AcPL (data not shown).  1. Comparison of the predicted amino acid sequences of human and murine AcPL. The alignment between human and murine versions of AcPL is presented, as determined by the UWGCG program Pileup. The predicted signal peptides are underlined, and the putative transmembrane domains are doubly underlined. Both sequences represent data from at least three independent library isolates. These sequences have been deposited in GenBank TM (accession numbers AF077347 (murine) and AF077346 (human).

FIG. 2. Northern blot analysis of the expression pattern of AcPL.
A Northern blot containing 2 g of mRNA from normal human spleen, thymus, prostate, testis, ovary, small intestine, colon, and peripheral blood leukocyte was purchased from CLONTECH. The blot was hybridized overnight against a hAcPL riboprobe. The blot was subsequently hybridized against a ␤-actin probe to standardize for integrity and amount of RNA in each lane. kb, kilobases.
We have also shown that both mIL-1R-rp1 and mAcPL receptor chains were required when S49.1 cells were electroporated, and reporter gene activity was measured in response to mIL-18 (data not shown). In these experiments, transfection efficiency was much less than in Cos7 cells, and we therefore did not observe an elevation in basal NFB activity when both receptors were coexpressed.
In contrast to murine AcPL, overexpression of human AcPL alone was able to mediate a significant NFB induction following hIL-18 stimulation in Cos7 cells (Fig. 4B). Using primers designed against human IL-1R-rp1, we were able to verify expression of IL-1R-rp1 in our Cos7 cells (data not shown). It is therefore likely that the endogenous monkey IL-1R-rp1 is capable of cooperating with hAcPL but not with mAcPL to confer IL-18 responsiveness. Alternatively, it could be that monkey IL-1R-rp1 can bind human IL-18 but not murine IL-18.
If IL-1R-rp1 and AcPL are both absolutely required for IL-18 signaling, as suggested by the above experiments, then one would expect that mutants of either receptor in which the entire cytoplasmic domain were deleted should exhibit a dominant negative effect. To test this, we transfected Cos7 cells with limiting amounts of receptor (10 ng each) and cotransfected mutant versions of either mIL-1R-rp1 or mAcPL (100 and 1000 ng). As shown in Fig. 4C, both mIL-1R-rp1⌬cyto and mAcPL⌬cyto were capable of inhibiting IL-18 signaling. In contrast, overexpression of these mutant receptors had no effect on IL-1 signaling (data not shown).
Another downstream signaling event in the IL-1 pathway is induction of JNK activity. We have observed that AE7 (murine T) cells stimulated with IL-18 show an induction of JNK activity. 2 We therefore examined whether AcPL alone or in combination with IL-1R-rp1 was capable of mediating the induction of JNK activity by IL-18. Cos7 cells were transiently transfected with empty vector, IL-1R-rp1, AcPL, or IL-1R-rp1 and AcPL. Two days following transfection, cells were stimulated with IL-18 for 15 min, and then cells were lysed and the JNK was immunoprecipitated. JNK activity in the immunocomplexes was assayed in vitro using glutathione S-transferase-c-Jun (1-169) as a substrate. Similar to the results obtained or mAcPL lacking the cytoplasmic domain were cotransfected at 100 ng or 1 g. In all cases the total amount of DNA remained constant by addition of empty vector (pDC304). Cells were stimulated and assayed as described above. All data represent the relative light units (RLU) per sample. Experiments were performed at least three times, with one representative experiment being presented. Open bars, medium; bars with shaded grid,  regarding activation of NFB, JNK activity was only induced by IL-18 in Cos7 cells when IL-1R-Rp1 and AcPL were coexpressed (Fig. 5).
AcPL Is Unable to Bind IL-18 in Vitro-It has already been reported that overexpression of IL-1R-rp1 in Cos1 cells resulted in elevated levels of IL-18 binding (9) and that IL-1R-rp1 was able to bind IL-18 in vitro. 2 We have investigated the ability of AcPL to bind IL-18 by immunoprecipitation of mammalianexpressed IL-18 with AcPL-Fc protein. An Fc fusion protein encoding the extracellular region of hAcPL was used to precipitate 35 S-labeled IL-18 expressed in Cos7 cell supernatants. As controls, the radiolabeled IL-18 was also precipitated with Fc fusion proteins encoding the extracellular domains of IL-1R and IL-1R-rp1. As shown in Fig. 6, only IL-1R-rp1 was able to bind IL-18 in this assay. There was no detectable binding of AcPL to IL-18. Supernatants from all samples were examined for levels of hIL-18 expression prior to precipitation, and all supernatants expressed similar levels of the cytokine (data not shown). Furthermore, the above experiment has been repeated with unpurified supernatants from cells transfected with hIL-1R-rp1-Fc and hAcPL-Fc, with the same result, eliminating the possibility that purification of hAcPL-Fc ablates its cytokine binding activity. DISCUSSION We have cloned a novel member of the IL-1R family which we show is capable of mediating cell signaling in response to IL-18. That AcPL is part of the IL-18 receptor complex is not surprising given that {1} IL-18 is structurally related to IL-1 and would be expected to recognize a receptor with a structure related to IL-1R and {2} IL-18 has already been shown to bind another member of this receptor family, IL-1R-rp1 (9). Further, given that the only other characterized member of this family requires two distinct receptor chains for signaling, it would be expected that two receptor chains would be required for IL-18 signaling. In support of this, we have observed that Cos7 or Cos1 cells transfected with IL-1R-rp1 alone were not responsive to IL-18 in terms of NFB induction. 2 These results suggested that another protein was required to form an active IL-18 receptor complex. The data presented in this paper suggest that the other subunit required for IL-18 responsiveness is AcPL.
We have shown that AcPL is expressed in lung, spleen, and peripheral blood lymphocytes, as well as peripheral blood T cell and NK libraries. Expression was not detected in heart, brain, kidney, or muscle mRNA. This expression pattern closely mirrors that of IL-1R-rp1 (22), although some tissues such as heart and testis express detectable IL-1R-rp1 but not AcPL. The pattern of AcPL expression also correlates well with the cell types that have been reported to be responsive to IL-18. Previous studies have shown that IL-18 induces interferon-␥ production from T cells and NK cells. Costimulatory effects of IL-18 have also been reported in peripheral blood mononuclear cells and B cells (34).
Interestingly, we have also observed that IL-1R-rp1 and AcPL are capable of inducing IL-8 promoter activity in response to IL-18 (data not shown). The IL-8 promoter is directly activated by IL-1 (35), but it has not been shown to be activated directly by IL-18 to date. It was recently demonstrated that IL-8 was induced in IL-18 stimulated CD3 ϩ /CD4 ϩ and NK cells, but this induction was dependent on the induction of TNF␣ (36). These data suggest that in addition to activation of an artificial, NFB-containing reporter, AcPL and IL-1R-rp1 are able to mediate IL-18 signaling to a more complex promoter.
Because AcPL alone was unable to bind IL-18, we propose that it is analogous to the IL-1RAcP, which is required for IL-1 signaling but does not bind IL-1 itself. By extension of this analogy, it will be of interest to determine whether cells expressing IL-1R-rp1 and AcPL display a higher affinity toward IL-18 than do cells expressing IL-1R-rp1 alone. We have attempted to address this issue by binding recombinant IL-18, iodinated on either tyrosine or lysine residues, to Cos7 cells overexpressing IL-1R-rp1, AcPL, or both receptors together. We have been unable to demonstrate reproducible specific binding of 125 I-IL-18 in these assays, suggesting that perhaps the iodination has interfered with cytokine binding. We are currently investigating alternative approaches to address this issue.
The downstream mediators of IL-18 signaling seem to be highly related to those involved in IL-1 signaling. Both our laboratory and others have recently shown that stimulation of T cells (AE7 or EL4) with IL-18 results in the recruitment of IRAK1 and TRAF6 to IL-1R-rp1 (11). 2 In IL-1 signaling, it is IL-1RAcP that is responsible for recruitment of IRAK1 to the active receptor complex, whereas IL-1R recruits IRAK2 (10,18). It will be interesting to determine whether AcPL is able to associate with these downstream signaling molecules. Previously reported data showing that IL-1R-rp1 associates with FIG. 5. IL-18-induced JNK1 activation requires co-expression of IL-1R-rp1 and AcPL. Cos7 cells were transiently transfected with empty vector, mIL-1R-rp1, mAcPL, or mIL-1R-rp1 and mAcPL. 2 days following transfection, cells were stimulated with mIL-18 for 15 min, and then JNK1 was immunoprecipitated. Glutathione S-transferase-c-Jun (GST-cjun) was phosphorylated in vitro by the immunocomplexes. The resulting polyacrylamide gel was analyzed on a PhosphorImager to quantitate the level of activity in each sample, as presented in the graph above the autoradiogram.
FIG. 6. AcPL is unable to bind IL-18 in vitro. Cos7 cells were transfected with empty vector (pDC206) or the same vector encoding hIL-18, and 2 days post transfection the samples were labeled with [ 35 S]Cys/Met-containing medium. The indicated Fc fusion proteins and protein G-Sepharose were added to the labeled supernatants, and precipitation was carried out overnight. The immunoprecipitated proteins were separated on 4 -20% Tris-glycine gels, dried, and exposed to film. The autoradiogram following a 2-day exposure is shown.
IRAK do not rule out the possibility that IL-1R-rp1 is part of a larger IL-18 binding complex, rather than being directly involved in binding these proteins.