Characterization of the I B-kinase NEMO Binding Domain*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 277, No. 48, Issue of November 29, pp , by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Characterization of the I B-kinase NEMO Binding Domain* Received for publication, July 1, 2002, and in revised form, August 13, 2002 Published, JBC Papers in Press, September 19, 2002, DOI /jbc.M Michael J. May, Ralf B. Marienfeld, and Sankar Ghosh From the Section of Immunobiology and Department of Molecular Biophysics and Biochemistry, Howard Hughes Medical Institute, Yale University Medical School, New Haven, Connecticut Proinflammatory activation of NF- B requires an upstream kinase complex (I B-kinase; IKK) composed of two catalytic subunits (IKK and IKK ) and a noncatalytic regulatory component named NEMO (NF- B essential modulator). NEMO interacts with a COOH-terminal sequence within both IKKs termed the NEMO-binding domain (NBD), and a cell-permeable NBD peptide blocks NEMO/IKK interactions and inhibits tumor necrosis factor- -induced NF- B. We report here that a peptide encompassing the NBD not only blocked association of both IKKs with NEMO but also disrupted preformed NEMO/IKK complexes in vitro. Furthermore, peptide blocking and alanine-scanning mutation studies revealed differences between the NBDs of IKK and IKK, and mutational analysis of the IKK NBD identified the physical properties required at each position to maintain association with NEMO. Finally, we demonstrate that loss of NEMO-binding by IKK through deletion of the NBD renders it catalytically active and that potential phosphorylation within the IKK NBD may serve as a signal to down-regulate IKK activity. Our findings therefore provide critical insight into the physical properties of the NBD that will be valuable for the design of drugs aimed at disrupting the IKK complex and also reveal potential regulatory mechanisms controlling the function of the IKK complex. Activation of the inducible transcription factor NF- B isan essential signal transduction pathway that is rapidly and transiently elicited in response to proinflammatory cytokines. Many genes that are induced or up-regulated by proinflammatory stimuli contain specific NF- B binding sites ( B sites) within their promoters (1, 2). Hence, cytokines generated during inflammatory responses such as interleukin-1 and tumor necrosis factor (TNF- ), 1 induce altered patterns of gene expression that are dependent upon the activity of NF- B. It is perhaps not surprising therefore that the NF- B activation pathway has emerged as an extremely attractive target for the development of anti-inflammatory drugs (3). * 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. Supported by the American Heart Association. Supported by a fellowship from the Emmy Noether Program of the Deutsche Forschungsgemeinschaft. Supported by the National Institutes of Health and the Howard Hughes Medical Institute. To whom correspondence should be addressed: Section of Immunobiology and Department of Molecular Biophysics and Biochemistry, Yale University Medical School, 310 Cedar St., New Haven, CT Tel.: ; Fax: ; sankar.ghosh@yale.edu. 1 The abbreviations used are: TNF-, tumor necrosis factor- ; I B, inhibitor of B; IKK, I B kinase; NBD, NEMO-binding domain; GST, glutathione S-transferase; HLH, helix-loop-helix Immense effort from many workers has revealed crucial details of the molecular mechanisms underlying proinflammatory activation of NF- B (2). In the majority of resting cells, NF- B remains sequestered in the cytoplasm through association with members of an inhibitory family of proteins named inhibitors of B(I B) that are typified by I B. Stimuli that activate NF- B induce phosphorylation of I B proteins at two specific NH 2 - terminal serine residues that tags the proteins for ubiquitination and subsequent degradation by the proteasome. Loss of bound I B reveals a nuclear localization sequence on NF- B allowing it to translocate to the nucleus, where it binds to DNA and interacts with distinct combinations of co-factors and other transcription factors to initiate or up-regulate gene expression. Various additional regulatory mechanisms such as stimulusinduced phosphorylation of the NF- B p65 subunit that renders it transcriptionally active (4 6) and repression by DNAbound NF- B complexes associated with histone deacetylases (7, 8) play critical roles in governing the NF- B activation pathway. Arguably, however, the most important point of control of proinflammatory NF- B activity is the activation and regulation of the kinase responsible for I B phosphorylation. I B-kinase (IKK) functional activity resides in a high molecular weight complex composed of a core of three separate protein subunits (1, 2). Catalytic function of the complex is provided by two of the subunits, named IKK and IKK, that exhibit striking structural similarity. These kinases are of comparable size, with IKK being slightly longer (756 amino acids compared with 745 for IKK ), and they share 52% identity throughout their entire length. IKK and IKK both contain a single NH 2 -terminal catalytic domain, a centrally positioned leucine zipper motif through which they heterodimerize, and COOH-terminal helix-loop-helix (HLH) domain that may function to regulate the catalytic activity of the IKK complex (9, 10). Despite their structural similarities, elegant genetic studies have demonstrated that IKK and IKK play distinct roles with respect to NF- B activation. Thus, IKK is responsible for proinflammatory cytokine-induced I B phosphorylation and subsequent activation of classical NF- B complexes containing the p50 and p65 subunits (11 14), whereas IKK plays a significant although poorly defined role in keratinocyte differentiation that is independent of its catalytic activity and NF- B activation (15 17). Recently, however, it has been demonstrated that IKK catalytic activity is required for RANKL (receptor activator of NF- B ligand)-induced NF- B activation in mammary epithelial cells (18) and plays a crucial role in B-lymphocyte development (19, 20). Intriguingly, however, the role of IKK in B-cell development does not involve I B phosphorylation, but it appears to function as an upstream kinase in a pathway leading to processing of the NF- B2/p100 subunit to p52 (20). The third protein within the IKK complex is named NEMO (NF- B essential modulator) but is also referred to as IKK, although it does not possess a catalytic domain (21 23). Evi- This paper is available on line at

2 Characterization of the IKK NBD dence obtained from genetically altered mice clearly demonstrates that NEMO is absolutely critical for proinflammatory activation of the IKK complex (24 26). Furthermore, mutations within NEMO that render it disfunctional have been associated with two X-linked pathologies named incontinentia pigmenti and ectodermal dysplasia associated with immunodeficiency (25, 27, 28). Thus, NEMO remains the only protein known to function in the NF- B pathway that has been linked to human genetic diseases. Nevertheless, the precise function of NEMO in IKK activation is not yet known. It has been proposed that NEMO plays a regulatory role by recruiting the IKK complex to ligated cytokine receptors, where IKKs may be activated by as yet poorly defined mechanisms that possibly include oligomerization, conformational changes, or proximityinduced transphosphorylation events (29 33). Additionally or alternatively, NEMO may facilitate the recruitment of upstream IKK activators such as kinases that specifically target the activation loops within the catalytic domains of the IKK subunits or enable the interaction of the IKKs with I B proteins (22, 34). In a previous study, we identified the molecular mechanism through which NEMO interacts with both IKKs (35). We found that an NH 2 -terminal -helical region of NEMO associates with a hexapeptide sequence (Leu-Asp-Trp-Ser-Trp-Leu) within the extreme COOH terminus of both kinases and named this region the NEMO-binding domain (NBD). Furthermore, a short cell-permeable peptide spanning the IKK NBD disrupted the association of NEMO with IKK in vitro, blocked TNF- -induced NF- B activation in cells, and effectively ameliorated responses in distinct animal models of inflammation. Hence, we have proposed that the NBD is an attractive target for the development of anti-inflammatory drugs aimed at disrupting the IKK complex. For such drugs to be developed, it will be necessary to obtain a clear understanding of the relative contributions and importance of each amino acid within the NBD to the interaction with NEMO. Therefore, to understand the NBD in greater detail and to investigate the function of NEMO in the IKK complex, we report here the results of in vitro peptide blocking studies and an in depth mutational analysis of the IKK NBD. Conservative or nonconservative substitution mutations of critical NBD residues demonstrated the required nature of the amino acids at each position for binding function and revealed significant differences between the NBDs of IKK and IKK. Furthermore, we found that loss of NEMO binding function through deletion of the NBD rendered IKK as catalytically active as the TNF- -induced kinase and that a mutation mimicking phosphorylation within the IKK NBD reduced its ability to activate NF- B. We therefore conclude that in addition to being critical for proinflammatory IKK activation, association with NEMO may also play a significant role in maintaining the basal activity of the IKK complex. EXPERIMENTAL PROCEDURES Cell Culture and Reagents HeLa and COS cells were maintained in Dulbecco s modified Eagle s medium (Invitrogen) supplemented with 10% fetal bovine serum, 2 mm L-glutamine, penicillin (50 units/ml), and streptomycin (50 g/ml). Mouse anti-flag (M2) and anti-flag-coupled agarose beads were purchased from Sigma. Mouse Anti-Xpress was purchased from Invitrogen, and rabbit anti-nemo was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The horseradish peroxidaseconjugated secondary antibodies against either rabbit or mouse IgG were both from Amersham Biosciences. TNF- was purchased from R & D Systems. Peptides Small scale Fmoc synthesis of peptides was carried out on a Rainin Symphony instrument at the HHMI Biopolymer-Keck Foundation Biotechnology Resource Laboratory at Yale University. Peptides were characterized by matrix-assisted laser desorption ionization mass spectrometry and analytical reverse phase high pressure liquid chromatography analysis. Peptides were dissolved in Me 2 SO to stocks of between 20 and 50 mm. Plasmids and PCR Mutagenesis Full-length cdna clones of IKK and IKK were generous gifts from Dr. Michael Karin (University of California, San Diego). All subcloning and mutagenesis of the IKKs was performed by PCR using cloned Pfu DNA polymerase (Stratagene). All PCRs used primers containing restriction sites that enabled cloning into appropriate vectors (sequences available upon request). The wildtype and mutated IKK cdnas were inserted into the KpnI and NotI restriction sites of pcdna-3 or pcdna-3.1-xpress (Invitrogen), and all IKK cdnas were inserted into the EcoRI and XhoI sites of the same vectors. FLAG-tagged IKK was constructed by subcloning into pflag-cmv-2 (Sigma). A cdna encoding human NEMO was obtained as previously described (35). GST-NEMO was constructed by subcloning the full-length cdna into the EcoRI and XhoI sites of pgex-4t1 (Amersham Biosciences). The GST-NEMO protein was then made in Escherichia coli (BL21) by treating transformed bacteria with 0.4 mm isopropyl-1-thio- -D-galactopyranoside and following the protocol for protein recovery provided with the vector. Analysis of IKK-NEMO Interactions For GST pull-down analysis, pcdna-3 containing IKK or IKK was subjected to an in vitro transcription and translation reaction in the presence of [ 35 S]methionine using the TNT-T7 Quick system from Promega. Labeled proteins (1 l of reticulocyte lysate) were incubated with GST-NEMO (1 g) in 100 l of TNT (50 mm Tris, ph 7.5, 200 mm NaCl, 1% Triton X-100) at 4 C for 30 min, and then 20 l of a 50% (v/v) slurry of glutathione beads (Amersham Biosciences) was added and incubated a further 15 min. Proteins were then precipitated and washed extensively in TNT before the addition of sample buffer (20 l). Samples were then separated by SDS-PAGE (10%), and resulting gels were fixed and examined autoradiographically. In some experiments, NBD peptides were added to GST-NEMO in TNT either 15 min before or after the addition of the IKK proteins. For transient transfection studies, COS cells grown in sixwell trays were transfected with vector alone (pcdna3.1-xpress) or the IKKs together with FLAG-tagged NEMO (1 g of total DNA) using the Fugene6 transfection reagent (Roche Diagnostics). All DNA/Fugene6 incubations were performed at a ratio of 1 g of DNA to 3 l of Fugene6 according to the manufacturer s recommended protocol in OptiMEM medium (Invitrogen). After 48 h, cells were lysed in 500 l of TNT, and then complexes were immunoprecipitated using anti-flag-coupled agarose beads. A portion of each lysate (5%) was retained for analysis (preimmunoprecipitation). Precipitated proteins were analyzed by immunoblotting using epitope-specific (anti-flag or anti-xpress) antibodies that were visualized using enhanced chemiluminescence reagents from Amersham Biosciences. Luciferase Reporter Assay Luciferase reporter assays were performed as described previously (6, 7, 35). Briefly, HeLa cells grown on 12-well plates were transiently transfected using Fugene6 with the NF- B-dependent reporter construct pbiix-luciferase (0.2 g/ well) together with either vector alone (control), wild-type IKK, orthe NBD point or truncation mutants indicated (up to 1.0 g/well each). For dose-response experiments, the total concentration of transfected DNA was maintained by transfecting appropriate amounts of vector alone (pcdna-3). Cells were lysed in passive lysis buffer (Promega) h after transfection and luciferase activity was measured using a Luciferase assay kit from Promega. Immune Complex Kinase Assay For immune complex kinase assays, HeLa cells grown on six-well plates were transiently transfected with 1 g of FLAG-IKK constructs using the Fugene6 reagent as described above. Forty-eight hours after transfection, the cells were treated with TNF- (10 ng/ml) for 10 min then lysed on ice in 500 l of TNT for 15 min. Protein content in each lysate was determined using a Bio-Rad protein assay kit and then normalized between samples. Proteins in lysates were immunoprecipitated using anti-flag (M2)-coupled agarose beads for 1 h at 4 C, and then precipitates were washed extensively in TNT and then kinase buffer (20 mm HEPES, ph 7.5, 20 mm MgCl 2,1mM EDTA, 2 mm NaF, 2 mm -glycerophosphate, 1 mm dithiothreitol, 10 M ATP). Precipitates were then incubated for 15 min at 30 C in20 l of kinase buffer containing GST-I B -(1 90) and 10 Ci of - 32 P-labeled ATP (Amersham Biosciences). The substrate was then precipitated using glutathione-agarose (Amersham Biosciences) and washed extensively with TNT. Beads were then suspended in 20 l of sample buffer, and samples were separated by SDS-PAGE (10%). Kinase activity was determined by autoradiography. An equal portion of each cell lysate (10%) was precipitated using anti- FLAG (M2)-agarose beads and then processed for immunoblotting using anti-flag to ensure equal loading of proteins.

3 45994 Characterization of the IKK NBD FIG. 1.A peptide spanning the IKK NBD prevents the interaction of both IKK and IKK with NEMO. A, sequences of the wild-type (WT) and scrambled (Scr.) control peptides with the NBD residues in uppercase. The wild-type peptide corresponds to residues of IKK. B, pull-down analysis was performed using GST- NEMO and in vitro transcribed and translated IKK (upper panel). GST-NEMO was incubated for 15 min in the presence or absence (control) of either 2% Me 2 SO (vehicle) and scrambled or wild-type NBD peptides (500 and 1000 M) prior to the addition of IKK. The lower panel shows a Coomassie Blue-stained gel demonstrating that neither peptide affects the interaction of GST-NEMO with the glutathioneagarose beads used for precipitation. C, analysis was performed using GST-NEMO and [ 35 S]methionine-labeled IKK as described for B. Samples were incubated with 250, 500, and 1000 M concentrations of either scrambled or wild-type NBD peptide. D, GST-NEMO was incubated with 2% Me 2 SO (vehicle) or wild-type or scrambled peptides either 15 min before (PRE; lanes 1 3) or 15 min after (POST; lanes 4 6) the addition of [ 35 S]methionine-labeled IKK (upper panel) orikk (lower panel). RESULTS A Peptide Spanning the IKK NBD Blocks the Interaction of Both IKKs with NEMO In an earlier study, we demonstrated that a cell-permeable NBD peptide blocked the interaction of NEMO with IKK (35). To further assess the effects of the NBD peptide in the absence of a cell-permeabilizing sequence, we performed in vitro binding assays using a peptide of only 11 amino acids surrounding the NBD of IKK (Fig. 1A). We previously used a control peptide containing tryptophan (Trp) to arginine (Ala) substitutions that did not inhibit NEMO binding to IKK (35); however, to maintain the amino acid content of the wild-type peptide, we employed a scrambled control (Fig. 1A) for the present study. To test the effects of the peptides on the IKK/NEMO interaction, we performed in vitro glutathione S-transferase (GST) pull-down analysis using a bacterially expressed version of full-length NEMO fused with GST (GST- NEMO) together with in vitro transcribed and translated and 35 S-labeled IKK proteins. Consistent with our previous findings, preincubation of GST-NEMO with the wild-type NBD peptide dose-dependently inhibited its ability to interact with IKK, whereas identical concentrations of the scrambled peptide had no effect (Fig. 1B). Neither peptide blocked the association of GST-I B with 35 S-labeled IKK (data not shown); nor did they disrupt the association of GST-NEMO with the glutathione beads used for precipitation (Fig. 1B), verifying that the blocking effects are specific for the NBD. Although previous workers suggested that NEMO only interacted with IKK within the IKK complex (22, 23), we found that IKK also contains a sequence within its extreme COOH terminus identical to the IKK NBD (35). Moreover, we showed by in vitro pull-down assays that recombinant IKK could bind to NEMO and that the interaction was maintained by the NBD. We therefore wanted to determine whether the NBD peptide could inhibit the association of NEMO with IKK, and, as shown in Fig. 1C we found that the wild-type but not the scrambled peptide effectively blocked the interaction. Identical results from such in vitro analysis were obtained using the FIG. 2. Peptide inhibition of NEMO binding and mutational analysis reveals differences between IKK and IKK. A, GST- NEMO was incubated for 15 min with the concentrations of NBD-WT peptide indicated and then further incubated with [ 35 S]methioninelabeled IKK or IKK prior to pull-down analysis. B, densitometric analysis data were pooled from 11 separate experiments similar to that shown in A. The data are presented as the mean pixel density as a percentage of control (no peptide) and represent means S.E. Analysis was performed using NIH Image software. C, each residue of the IKK NBD (Leu 738 to Leu 743 ) was replaced with alanine by PCR-mutagenesis. COS cells were transiently transfected with NEMO-FLAG (lanes 1 9) or IKK -FLAG (lanes 10 18) together with either vector alone (pcdna-3.1-xpress; Control) or Xpress-tagged versions of IKK and the NBD mutants as shown. Immunoprecipitation (IP) and immunoblot (IB) analysis of the IKK -NEMO complexes was performed as described under Experimental Procedures. WT, wild type. cell-permeable NBD peptides (data not shown). Taken together, the experiments depicted in Fig. 1, A C, reinforce our conclusion that the NBD peptide specifically blocks the interaction of NEMO with the IKK complex and further demonstrate that it prevents NEMO binding to both IKK subunits in vitro. For the preceding experiments, we preincubated GST- NEMO with the NBD peptides for 15 min prior to adding the 35 S-labeled IKK proteins. However, any in vivo effects of compounds designed to target the NBD would presumably depend upon disruption of preformed IKK complexes. We therefore sought to determine whether the NBD peptide could dissociate a preformed interaction between NEMO and the IKKs in vitro. To test this, we used pull-down assays to compare the effects of preincubating the peptides with adding them after incubation of GST-NEMO with IKK or IKK. As expected, the scrambled peptide had no effect on the interaction with either kinase when tested using both incubation protocols (Fig. 1D, lanes 2 and 5). In contrast, preincubation with the wild-type peptide blocked binding of both kinases to GST-NEMO, and although less dramatic, the peptide also reduced binding when added after complex formation (Fig. 1D, compare lanes 3 and 6). These results therefore lead us to conclude that the NBD peptide can disrupt preformed NEMO/IKK interactions. During these studies, we consistently observed that concentrations of NBD peptide that completely blocked association of IKK with GST-NEMO ( M) did not fully inhibit the interaction with IKK (Fig. 1D, lane 3). To explore this further, we performed a series of dose-response experiments in which we directly compared the effects of the peptide on both kinases, and as shown in Fig. 2A, a notable difference in effective concentrations was observed. Moreover, densitometric analysis of data obtained from a series of identical experiments (n 11) clearly demonstrated that effective blockage of IKK /NEMO interactions occurred at lower concentrations of peptide

4 TABLE I Comparison of the effects of single alanine mutations within the NBD of IKK and IKK on the ability of each kinase to interact with NEMO NEMO binding ability was defined as strong ( ), weak ( / ), or absent ( ). NBD position Residue Effect of alanine substitution (NEMO binding) IKK -( ) IKK -( ) 1 Leu 2 Asp 3 Trp 4 Ser 5 Trp / 6 Leu Characterization of the IKK NBD compared with disruption of IKK /NEMO (Fig. 2B). To try to determine the molecular basis for these differences, we performed alanine-scanning mutagenesis to identify the IKK NBD residues that contribute to its association with NEMO. Similar analysis of the IKK NBD revealed that the aspartate (Asp 738 ) and two tryptophan residues (Trp 739 and Trp 741 ) were critical for NEMO association (Table I) (35). Consistent with this, truncation of IKK immediately before the NBD (1 737) along with alanine substitution of the first tryptophan residue (Trp 740 ) prevented NEMO binding (Fig. 2C, lane 6), whereas mutation of the two leucines (Leu 738 and Leu 743 ) and the serine residue (Ser 741 ) had no effect (lanes 4, 7, and 9). In contrast to IKK, however, we observed marked differences when the aspartate (Asp 739 ) and second tryptophan (Trp 742 ) residues of the IKK NBD were replaced with alanine (Table I). Thus, whereas mutations at these positions within IKK completely blocked NEMO association, alanine substitution of Asp 739 did not affect binding (Fig. 2C, lane 5), and mutation of Trp 742 reduced but did not completely block the interaction of IKK with NEMO (lane 8). The similarities and differences in NBD residue requirements for NEMO binding between IKK and IKK are summarized in Table I. It remains formally possible that in the transfection experiments described above, IKK interacted indirectly with NEMO via association with endogenous NEMO-bound IKK and that the mutations within the IKK NBD affected the formation of such complexes. We therefore tested the ability of each of the IKK NBD mutants to associate with FLAG-tagged IKK and found that they all formed complexes (Fig. 2C, lanes 10 18). Identical results were obtained using GST pull-down analysis (data not shown), and, combined with our previous findings (35), we conclude that the effects of the IKK NBD mutations were through preventing its direct interaction with NEMO. Therefore, when taken together, the data presented in Figs. 1 and 2 strongly suggest that IKK binds directly to NEMO, albeit with a lower affinity than IKK that may be due to distinct requirements for specific residues within the NBD of each kinase for NEMO binding. Mutational Analysis of the IKK NBD The results described above verify that targeting the NBD represents an attractive strategy for the development of drugs to specifically block proinflammatory activation of the IKK complex. To facilitate such drug design, it will be informative to develop a clear understanding of the relative contributions and functional importance of each amino acid within the NBD. We have therefore performed extensive mutational analysis of the IKK NBD in which we have substituted the residues that are critical for binding with appropriate conserved and nonconserved amino acids. As shown in Table I, replacing the aspartate at position 738 (Asp 738 ) with alanine prevented IKK from binding to NEMO. To investigate the required nature of the residue at position FIG. 3.Substitution analysis of the aspartate and tryptophan residues within the IKK NBD. A, the aspartate residue at position 738 of IKK was replaced with either alanine (D738A), asparagine (D738N), or glutamate (D738E), and then mutants were [ 35 S]methionine-labeled by in vitro transcription and translation and used for pull-down analysis with GST-NEMO. B D, the tryptophan residues within the IKK NBD were replaced with alanine (W739A and W741A), phenylalanine (W739F and W741F), tyrosine (W739Y and W741Y), or arginine (W739R and W741R). COS cells were transiently transfected with NEMO-FLAG together with either vector alone (Control) or Xpress-tagged versions of IKK and the NBD tryptophan mutants as shown. Immunoprecipitation (IP) and immunoblot (IB) analysis of the IKK -Xpress NEMO-FLAG complexes was performed as described under Experimental Procedures. 738, we made conservative substitutions with either asparagine (D738N) to maintain the shape or glutamate (D738E) to maintain the shape and charge of the residue. As shown in Fig. 3A, neither substitution affected the ability of IKK to associate with NEMO, whereas the original alanine substitution prevented binding. We therefore conclude that it is the shape (specifically the presence of a second carbon) and not the charge of the side chain at this position that is critical for maintaining the interaction between IKK and NEMO. Alanine scanning mutagenesis also revealed that both tryptophan residues (Trp 739 and Trp 741 ) are critical for NEMO binding of IKK (Table I). We therefore tested the effects of conservative mutations that retain the aromatic structure of the residues at these positions by replacing them with either phenylalanine or tyrosine (Fig. 3, B and D). In addition, we replaced both tryptophans with arginine (Fig. 3, C and D), since this requires only a single base change within the encoding codon (TGG 3 AGG). Although such a mutation has not been reported for IKK, it has been shown to occur naturally in other proteins and has been linked to disease conditions (36). As shown in Fig. 3B, both W739F and W739Y mutants associated with NEMO as well as IKK, whereas W739R was unable to bind (Fig. 3C). Together with the effects of alanine substitution, these findings indicate that the aromatic nature of the residue at this position is critical for the function of the NBD. Similarly, we found that replacing Trp 741 with phenylalanine (W741F) did not affect association with NEMO, whereas mutation to arginine (W741R) prevented binding (Fig. 3D). In contrast to Trp 739, substitution with tyrosine (W741Y) blocked NEMO binding, thereby demonstrating that the presence of a hydroxyl moiety within the side chain at this position is sufficient to prevent association of NEMO. Thus, it appears that whereas both of these residues are critical for binding, the presence of tryptophan at position 741 is obligatory for a functional NBD in IKK. In contrast to the effects of single alanine substitutions at Asp 738, Trp 739, and Trp 741, replacement of the serine (Ser 740 )or

5 45996 Characterization of the IKK NBD FIG. 4.Effects of single and double substitution mutations of the serine and leucine residues within the IKK NBD. The leucine residues at positions 737 and 742 and the serine at position 740 of IKK were replaced with alanine (L737A, L742A, and S740A, respectively). Double mutants containing alanine substitutions at Leu 737 and Ser 740 (LS), Ser 740 and Leu 742 (SL), and both leucine residues (LL) were also constructed. COS cells were transiently transfected with vector alone (Control) or the mutants together with FLAG-tagged NEMO as shown, and immunoprecipitation (IP)/immunoblot (IB) analysis was performed as described under Experimental Procedures. FIG. 5. The NBD regulates basal IKK activity. A, HeLa cells were transiently transfected with an NF- B-dependent reporter construct (pbiix-luciferase) and either vector alone (Control), wild-type IKK, or the NBD single and double point mutants shown. Luciferase activity in cell lysates was measured as described under Experimental Procedures. The ability of the overexpressed proteins to bind ( )ornot ( ) to NEMO is indicated (right). B, HeLa cells were transfected with pbiix-luciferase together with either vector alone (Control), wild-type IKK, or a constitutively active IKK mutant in which the two serine residues within the activation loop (Ser 177 and Ser 181 ) were replaced with glutamate (SS/EE). C, HeLa cells were transfected with pbiixluciferase and either vector alone (Control), wild-type IKK or IKK,or an IKK truncation mutant lacking the NBD (residues 1 737) that does not bind to NEMO luciferase activity in cell lysates was measured 24 h after transfection. either leucine residue (Leu 737 or Leu 742 ) with alanine did not affect NEMO binding, suggesting that none of these play a critical role in maintaining the interaction (Table I). To determine whether double mutations at these positions would affect binding, we constructed a panel of mutants in which Leu 737 and Ser 740, Ser 740 and Leu 742, or both leucine residues (named LS, SL, and LL, respectively) were replaced with alanine. Following transient transfection of COS cells, we found that both the LS and SL mutants associate with NEMO to the same extent as wild-type IKK, whereas we could not detect any binding of the LL mutant (Fig. 4). Together with GST pulldown analysis that yielded identical results (data not shown), these findings demonstrate that the presence of at least one of the leucine residues is required for efficient NBD function, whereas Ser 740 does not contribute to the NEMO binding capacity of the IKK NBD. The NBD and a COOH-terminal Serine-rich Region Regulate Basal IKK Activity We demonstrated previously that transiently overexpressed versions of IKK containing single alanine mutations at the NBD residues critical for NEMO association (i.e. Asp 738, Trp 739, and Trp 741 ), thereby rendering them unable to bind to NEMO, activated NF- B to a greater extent than those capable of interacting (35). We therefore evaluated the ability of the leucine and serine single and double alanine point mutants to activate NF- B and, in agreement with our prior observations, found that the LL mutant was consistently more active than the wild-type kinase (Fig. 5A). In a series of separate experiments, this increased activity ranged between 1.5- and 2.5-fold higher than that induced by wild-type IKK. To determine how this level of activity compared with NF- B induced by activated IKK, we overexpressed a constitutively active IKK mutant in which the activation loop serines (Ser 177 and Ser 181 ) were replaced with glutamate (SS/EE) (10). As shown in Fig. 5B, IKK (SS/EE) induced NF- B activity that was 2-fold higher than that of wild-type IKK. Consequently, when taken together, these data demonstrate that in our assay system under identical transfection conditions, the NF- B inducing ability of IKK mutants unable to bind to NEMO resembles that of fully active IKK. We next tested the effects of loss of NEMO binding on IKK induced NF- B activity using a truncated version of IKK that lacked the NBD (1 737). As shown in Fig. 5C and as reported by others (37), the activity of wild-type IKK was significantly lower than that of IKK ; however, similar to IKK, loss of NEMO binding resulted in an 2-fold increase in basal IKK activity. These results therefore strongly suggest that association with NEMO suppresses the basal activity of both IKKs. Previous workers have suggested that a Ser-rich region adjacent to the NBD plays a role in down-regulating the IKK complex following proinflammatory activation (10). Furthermore, the results described in Fig. 5, A and B, indicate that the activity of overexpressed IKK lacking a functional NBD is equivalent to that of the inducibly activated kinase. Taken together, these findings strongly imply that the entire COOH terminus of IKK distal to the HLH is critical for regulating its catalytic function. We therefore wanted to determine the effects that loss of both of these regulatory domains would have on IKK activity and made FLAG epitope-tagged COOHterminal mutants of IKK (Fig. 6A) that were truncated after either the NBD (1 744) or the Ser-rich region (residues 1 733) or immediately after the HLH (residues 1 644). To determine whether these proteins could potentially interact with endogenous NEMO when overexpressed, we performed immunoprecipitation and immunoblot analysis, and, as shown in Fig. 6B, only wild-type IKK and interacted with endogenous NEMO, whereas and that lack the NBD did not associate. To investigate the effects of these truncations on the induced catalytic activity of IKK, we incubated transiently transfected HeLa cells in the absence or presence of TNF- (10 min) and then performed an immunoprecipitation-kinase assay. As

6 Characterization of the IKK NBD FIG. 6.An IKK truncation mutant lacking the NBD is constitutively active. A, depiction of the structural features of the IKK COOH terminus showing the relative positions of the HLH, the serinerich region (SER-RICH), and the NBD. The amino acid positions at which the truncation mutants were terminated are indicated at the top. B, HeLa cells were transiently transfected with FLAG-tagged versions of either wild-type IKK or the truncation mutants described in A. IKK proteins in lysates were immunoprecipitated (IP) using anti-flag and then immunoblotted (IB) using either anti-flag (upper panel) or anti- NEMO (lower panel). C, HeLa cells were transfected as described above. After 48 h, cells were treated with TNF- (10 ng/ml) for 10 min, and then an immune complex kinase assay (KA) was performed on cell lysates as described under Experimental Procedures. The positions of autophosphorylated IKK, phosphorylated NEMO, and the GST-I B substrate are indicated (right). To determine the relative expression levels of each overexpressed kinase, proteins were immunoprecipitated from equal portions of each lysate using anti-flag, and then samples were immunoblotted using anti-flag (lower panel). FIG. 7.Phospho-mimicking mutation of Ser 740 reveals a potential mechanism for negative regulation of IKK activity. A, serine 740 within the IKK NBD was replaced with either alanine (S740A) or glutamic acid (S740E). Association of these point mutants or wild-type (WT) IKK with NEMO was analyzed by immunoprecipitation and immunoblotting from lysates of COS cells transiently transfected with vector alone (Control) or the IKK constructs together with FLAG-tagged NEMO. B, HeLa cells were transiently transfected with either vector alone (Control; hatched bars) or a range of concentrations (0.25, 0.5, and 1 g/ml) of wild-type IKK (solid bars), S740A (open bars; left panel), or S740E (open bars; right panel). Luciferase activity was measured in cell lysates 24 h after transfection. shown in Fig. 6C, in resting cells (lane 1), wild-type IKK exhibited low level catalytic activity against a substrate consisting of the first 90 amino acids of I B fused with GST (GST-I B ). Following TNF- treatment, however, this activity was dramatically increased and accompanied by autophosphorylation and phosphorylation of associated NEMO (upper and middle bands in lane 2, respectively). In contrast, although exhibited a similar level of basal activity as wild-type IKK, this was barely increased following TNF- treatment, and no auto- or NEMO phosphorylation was observed (lanes 3 and 4). Remarkably, however, basal activity of that contains the Ser-rich region but not the NBD was identical to that of TNF- -treated wild-type IKK and was not increased by TNF- stimulation (lanes 5 and 6 compared with lane 2). Not surprisingly, no NEMO phosphorylation associated with was detected, although, consistent with the catalytic activity, autophosphorylation of was observed in untreated as well as TNF- -stimulated samples. When the NEMO binding function of IKK was reintroduced (residues 1 744), the kinase activity observed in resting and TNF- -treated cells was identical to that of wild-type IKK (Fig. 6C, lanes 1 and 8). Serine to Glutamate Substitution of Ser 740 within the NBD Reduces the Activity of IKK The results described above suggest that the Ser-rich region and NBD may coordinately regulate the catalytic function of IKK. In a previous study, Delhase et al. (10) demonstrated that multiple phosphorylations within the Ser-rich region promote conformational changes that down-regulate IKK catalytic activity following TNF- stimulation. However, the data we present in Fig. 6C (lane 5) clearly demonstrate that phosphorylation within the Ser-rich region alone is not sufficient to down-regulate IKK, since the mutant that is heavily autophosphorylated is fully catalytically active. It therefore appears that the presence of the extreme COOH terminus including the NBD is required for basal activity. We therefore set out to test what effects, if any, a potential phosphorylation event within this region would have on IKK function. The only residue within the NBD that presents a possible phosphorylation site is Ser 740. Consequently, we constructed a single point mutant in which this was replaced with glutamate (S740E) to mimic phosphorylation and tested its ability to interact with NEMO. As shown in Fig. 7A, wild-type IKK, S740A, and S740E all bound equally well to FLAG-tagged NEMO. Identical results were obtained in vitro using GST pull-down analysis (data not shown). We next investigated the effect that this mutation had on the ability of IKK to activate NF- B by transiently transfecting HeLa cells with wild-type IKK, S740A, and S740E together with the pbiix-luciferase reporter plasmid. Consistent with the results presented in Fig.

7 45998 Characterization of the IKK NBD 5A, wild-type IKK and S740A induced comparable levels of NF- B activity, although differences were observed when we transfected lower concentrations of DNA than previously tested. Thus, transfection of 0.25 g/ml S740A induced higher levels of NF- B activity than identical amounts of wild-type IKK (Fig. 7B, left panel). Intriguingly, however, overexpressed S740E consistently induced less NF- B activity than wild-type IKK at all DNA concentrations tested (Fig. 7B, right panel). From these data, we surmise that phosphorylation at position 740 within the NBD would not affect the ability of IKK to form a complex with NEMO but may act to negatively regulate the activity of the kinase. DISCUSSION This study was performed as part of our ongoing effort to understand the molecular mechanisms that regulate the function of the IKK complex with the aim of identifying potentially novel approaches to block its activation during inflammation. We have focused our attention on trying to define the regulatory role of NEMO in the IKK complex, since a number of genetic and biochemical studies clearly demonstrate that proinflammatory IKK activation is absolutely dependent upon the presence of functional NEMO (24 26). Despite this crucial function, the molecular mechanisms underlying the action of NEMO are essentially unknown, although it has been proposed to facilitate recruitment of the IKK complex to activated receptors and/or recruitment of upstream kinases to the IKK complex (22, 29 34). Whereas its precise function remains to be determined, disrupting the interaction of NEMO with the IKKs appears to represent a reasonable strategy to specifically inhibit activation of the IKK complex. We previously reported that a cell-permeable peptide containing the NBD effectively blocked the interaction of IKK with NEMO in vitro, inhibited TNF- -induced NF- B activation in HeLa cells, and was effective in several cell and animal models of inflammation (35). Since these findings strongly supported our hypothesis that blocking NEMO function would prevent proinflammatory IKK activation, we initiated the present study to gain insight into the structure and function of the NBD that will be useful for the development of therapeutics aimed at disrupting the IKK complex. We began by performing a series of in vitro binding assays in the presence of an NBD peptide and found that it blocked the interaction of both IKK and IKK with NEMO. Moreover, the wild-type peptide but not a scrambled control was able to disrupt preformed IKK NEMO complexes in vitro. Hence we propose that drugs targeting the NBD would disrupt preformed endogenous IKK complexes and prevent de novo complex formation, thereby preventing upstream signals from reaching either of the kinases in the IKK complex. Intriguingly, our peptide studies also indicated that IKK and IKK bind to NEMO with modestly different affinities. Although we have not performed an exhaustive kinetic analysis, when we examined data accumulated from a number of separate experiments (Fig. 2B), we consistently observed that significantly higher concentrations of NBD peptide were required to inhibit the interaction with IKK than the interaction with IKK. The reason for these differences may be differential requirements for distinct NBD residues for NEMO binding, since the aspartate (Asp 739 ) and second tryptophan (Trp 742 ) within the IKK NBD do not appear to be as critical for NEMO binding as the corresponding residues in IKK. Itis therefore possible that these differences may explain why some groups failed to detect an association of NEMO with IKK (22, 23), whereas we and others readily detect IKK /NEMO interactions (29, 38). These findings also pose fascinating questions regarding the contribution of NEMO to the distinct functions of IKK and IKK. Most of the evidence accumulated to date demonstrates that proinflammatory signals to NF- B require both NEMO and IKK, whereas nothing is known about the contribution of NEMO to IKK activation. Since our data clearly demonstrate that IKK can interact with NEMO, does the possible difference in affinity or the distinct residue requirements for interaction compared with IKK underscore potentially different modes of function? Interestingly, it has been reported that replacing the entire COOH terminus of IKK with the equivalent portion of IKK results in a chimeric kinase whose catalytic activity resembles that of IKK (9). This therefore supports the notion that the NBD and hence the interaction with NEMO may directly influence the activity of the distinct kinase subunits. A more intriguing question that arises is, therefore, can the nature of the interaction with NEMO determine whether some signals (i.e. those elicited in response to TNF- and interleukin-1) activate IKK (11 14), whereas others such as those activated by LT in B-cells or RANKL in mammary epithelial cells (18 20) activate IKK? Understanding how NEMO integrates upstream signals and influences the functions of the individual IKKs may therefore begin to provide answers to these as well as the mystifying question of why the IKK complex has evolved to contain two very similar kinases with apparently disparate functions. It is also tempting to speculate that such an understanding may permit the development of subtly distinct NBD-targeted drugs that specifically block the function of one IKK subunit while maintaining the activity of the other. Since the aim of our study was to develop an understanding of the NBD that might aid the design of anti-inflammatory drugs, we focused our attention on IKK, since this subunit has been shown genetically to be critical for proinflammatory NF- B activation (11 14). We constructed a panel of conservative and nonconservative substitution mutants to expand upon our previous alanine-scanning mutagenesis (35) and provide insight into the physical requirements at the critical positions within the NBD. As we previously demonstrated, replacement of Asp 738 with alanine (D738A) blocked the association of IKK with NEMO, demonstrating that this aspartate residue is critical for maintaining the binding. In contrast, substitution with the similarly shaped amino acid glutamate (D738E) or asparagine (D738N) that is also positively charged did not affect NEMO binding. We therefore conclude that it is the shape of the residue side chain at this position that contributes to the interaction with NEMO, since differences in charge (D738A) or size (D738E) did not prevent binding. It is interesting to reiterate that substitution of Asp 739 in IKK with alanine had no effect on its ability to bind to NEMO, demonstrating that the requirement for this particular shape of residue at this position is unique to the IKK NBD. The other residues within the IKK NBD that we found previously to be critical for NEMO binding were the tryptophans at positions 739 and 741 (Table I). The results of our mutational analysis described here not only reinforce these findings but also suggest that, of the two positions, the presence of a tryptophan at position 741 is the more important. Therefore, whereas replacement of Trp 739 with either phenylalanine (W739F) or tyrosine (W739Y) maintained NEMO binding, replacement of Trp 741 with tyrosine (W741Y) prevented association with NEMO. Since the W741F mutant was able to bind to NEMO, we surmise that the presence of the hydoxyl moiety on tyrosine was sufficient to block binding. Once again, this more stringent requirement for a tryptophan at this position differentiates the IKKs, since substitution of Trp 742 in IKK with alanine did not completely prevent its association with NEMO. The fact that the two NBD tryptophans contribute

8 Characterization of the IKK NBD so significantly to the interaction with NEMO is perhaps not surprising in light of information regarding the anatomy of protein/protein interaction interfaces. Thus, studies in which data have been accumulated from a wide range of albeit larger interaction interfaces have clearly demonstrated that typtophan along with tyrosine and arginine are the three amino acids that are enriched in so-called interaction hot spots (39, 40). In addition to these three residues, aspartate has also been shown to occur approximately twice as often as glutamate in these regions, verifying at least in principle that some of the major residues known to exist within binding hot spots are present within the NBD (40). It will therefore be of extreme interest to view the results of in depth structural analysis of the NBD to determine how these residues in both IKKs interact with the NH 2 -terminal -helix of NEMO. During mutational analysis, we also tested the effects of single and double substitution mutations of the NBD serine (Ser 740 ) and leucines (Leu 737 and Leu 742 ) (Fig. 4). These studies demonstrated that whereas Ser 740 is apparently dispensable for binding, the presence of at least one of the leucines is required for NBD function. Thus, a double alanine mutant in which both leucines were replaced with alanine did not bind to NEMO. It therefore remains unclear what precise function these leucines perform in the NBD, although it is possible that they are required to occlude solvent and maintain the hydrophobic nature of the NBD (35). In addition to investigating the physical properties of the NBD, we also determined the effects of its loss on IKK function. In support of our previous findings obtained using IKK truncation and NBD single alanine mutants (35), we found that NBD loss rendered both IKK and IKK capable of inducing increased levels of NF- B activity. Furthermore, comparison of the levels induced by IKK lacking the NBD (residues 1 733) (Fig. 5) with a constitutively active version of the kinase suggested that loss of the NBD induced full catalytic activity of IKK, results we verified when we performed catalytic assays. Intriguingly, we observed that the mutant displayed not only enhanced catalytic activity but was also highly phosphorylated in its COOH terminus. Since previous workers have shown that multiple phosphorylations within this region cause down-regulation of IKK activity (10), our data suggest that this is also dependent upon the presence of an intact NBD and hence bound NEMO. Moreover, we found that mimicking a single phosphorylation within the NBD (Ser 740 ) significantly reduced the ability of IKK to activate NF- B. We therefore propose that a combination of phosphorylation events within the Ser-rich region and the NBD induce a NEMO-dependent signal that down-regulates the IKK complex and maintains its basal activity. Recently, such an NBD-dependent inhibitory mechanism has been disputed, and it is commonly perceived that NEMO plays only an activating role in the IKK complex (38). This reasoning is based upon the observations that NF- B cannot be activated in NEMO-deficient cells (24 26) and overexpression of NEMO can activate IKK in mammalian and yeast systems (30, 34, 38). In contrast, some workers have described NEMO as an inhibitor of IKK function (29, 41), and others have demonstrated that variations in the cellular concentration of NEMO dictate, whether it functions to activate or inhibit NF- B (32). It is therefore possible that the true function of NEMO is dual and that it contributes to both regulatory mechanisms. Identifying the mechanism of NBD-dependent downregulation of IKK activity will require a full understanding of the role of NEMO in IKK activation. Nevertheless, possible mechanisms may include phosphorylation-dependent conformational changes (in the IKKs and NEMO) that cause either disengagement from receptors, loss of activating components (such as an upstream IKK-K), or recruitment of a negative regulatory protein. Whereas none of these mechanisms have been tested, it is tempting to speculate that a possible negative regulator may be an IKK phosphatase that is recruited to NEMO to dephosphorylate and inactivate the complex. In support of this, previous workers have demonstrated that the IKK complex can be activated by treatment with specific phosphatase inhibitors (42, 43). Why then is IKK not constitutively active in NEMO / cells? Whereas the answer to this question awaits investigation, it may simply be that the endogenous IKK complex in these cells cannot be activated. In contrast, in our experiments, enforced overexpression of IKK in HeLa cells may activate the complex that, in the absence of NEMO binding, cannot be down-regulated. Further work is clearly required to define the precise role of NEMO in the IKK complex; however, our findings strongly suggest that this extends beyond a simple requirement for activation. REFERENCES 1. Karin, M., and Delhase, M. (2000) Semin. Immunol. 12, Ghosh, S., and Karin, M. (2002) Cell 109, D Acquisto, F., May, M. J., and Ghosh, S. (2002) Mol. Intervent. 2, Madrid, L. V., Wang, C. Y., Guttridge, D. C., Schottelius, A. J., Baldwin, A. S., Jr., and Mayo, M. W. (2000) Mol. Cell. Biol. 20, Wang, D., Westerheide, S. D., Hanson, J. L., and Baldwin, A. S., Jr. (2000) J. Biol. Chem. 275, Zhong, H., Voll, R. E., and Ghosh, S. (1998) Mol. Cell 1, Zhong, H., May, M. J., Jimi, E., and Ghosh, S. (2002) Mol. Cell 9, Ashburner, B. P., Westerheide, S. D., and Baldwin, A. S., Jr. (2001) Mol. Cell. Biol. 21, Kwak, Y. T., Guo, J., Shen, J., and Gaynor, R. B. (2000) J. Biol. Chem. 275, Delhase, M., Hayakawa, M., Chen, Y., and Karin, M. (1999) Science 284, Chu, W. M., Ostertag, D., Li, Z. W., Chang, L., Chen, Y., Hu, Y., Williams, B., Perrault, J., and Karin, M. (1999) Immunity 11, Li, Z. W., Chu, W., Hu, Y., Delhase, M., Deerinck, T., Ellisman, M., Johnson, R., and Karin, M. (1999) J. Exp. Med. 189, Senftleben, U., Li, Z. W., Baud, V., and Karin, M. (2001) Immunity 14, Tanaka, M., Fuentes, M. E., Yamaguchi, K., Durnin, M. H., Dalrymple, S. A., Hardy, K. L., and Goeddel, D. V. (1999) Immunity 10, Li, Q., Lu, Q., Hwang, J. Y., Buscher, D., Lee, K. F., Izpisua-Belmonte, J. C., and Verma, I. M. (1999) Genes Dev. 13, Hu, Y., Baud, V., Delhase, M., Zhang, P., Deerinck, T., Ellisman, M., Johnson, R., and Karin, M. (1999) Science 284, Takeda, K., Takeuchi, O., Tsujimura, T., Itami, S., Adachi, O., Kawai, T., Sanjo, H., Yoshikawa, K., Terada, N., and Akira, S. (1999) Science 284, Cao, Y., Bonizzi, G., Seagroves, T. N., Greten, F. R., Johnson, R., Schmidt, E. V., and Karin, M. (2001) Cell 107, Kaisho, T., Takeda, K., Tsujimura, T., Kawai, T., Nomura, F., Terada, N., and Akira, S. (2001) J. Exp. Med. 193, Senftleben, U., Cao, Y., Xiao, G., Greten, F. R., Krahn, G., Bonizzi, G., Chen, Y., Hu, Y., Fong, A., Sun, S. C., and Karin, M. (2001) Science 293, Yamaoka, S., Courtois, G., Bessia, C., Whiteside, S. T., Weil, R., Agou, F., Kirk, H. E., Kay, R. J., and Israel, A. (1998) Cell 93, Rothwarf, D. M., Zandi, E., Natoli, G., and Karin, M. (1998) Nature 395, Mercurio, F., Murray, B. W., Shevchenko, A., Bennett, B. L., Young, D. B., Li, J. W., Pascual, G., Motiwala, A., Zhu, H., Mann, M., and Manning, A. M. (1999) Mol. Cell. Biol. 19, Rudolph, D., Yeh, W. C., Wakeham, A., Rudolph, B., Nallainathan, D., Potter, J., Elia, A. J., and Mak, T. W. (2000) Genes Dev. 14, Schmidt-Supprian, M., Bloch, W., Courtois, G., Addicks, K., Israel, A., Rajewsky, K., and Pasparakis, M. (2000) Mol. Cell 5, Makris, C., Godfrey, V. L., Krahn-Senftleben, G., Takahashi, T., Roberts, J. L., Schwarz, T., Feng, L., Johnson, R. S., and Karin, M. (2000) Mol. Cell 5, Smahi, A., Courtois, G., Vabres, P., Yamaoka, S., Heuertz, S., Munnich, A., Israel, A., Heiss, N. S., Klauck, S. M., Kioschis, P., Wiemann, S., Poustka, A., Esposito, T., Bardaro, T., Gianfrancesco, F., Ciccodicola, A., D Urso, M., Woffendin, H., Jakins, T., Donnai, D., Stewart, H., Kenwrick, S. J., Aradhya, S., Yamagata, T., Levy, M., Lewis, R. A., and Nelson, D. L. (2000) Nature 405, Jain, A., Ma, C. A., Liu, S., Brown, M., Cohen, J., and Strober, W. (2001) Nat. Immunol. 2, Ye, J., Xie, X., Tarassishin, L., and Horwitz, M. S. (2000) J. Biol. Chem. 275, Li, X. H., Fang, X., and Gaynor, R. B. (2001) J. Biol. Chem. 276,