![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Previous Article | Next Article 
Molecular and Cellular Biology, June 2004, p. 4895-4908, Vol. 24, No. 11
0270-7306/04/$08.00+0 DOI: 10.1128/MCB.24.11.4895-4908.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Regulation of Constitutive p50/c-Rel Activity via Proteasome Inhibitor-Resistant I
B
Degradation in B Cells
Shelby O'Connor,1 Stuart D. Shumway,1 Ian J. Amanna,2 Colleen E. Hayes,2 and Shigeki Miyamoto1*
Program in Cellular and Molecular Biology, Department of Pharmacology,1
Department of Biochemistry, University of Wisconsin—Madison, Madison, Wisconsin 537062
Received 30 November 2003/
Returned for modification 21 January 2004/
Accepted 2 March 2004
 |
ABSTRACT
|
|---|
Constitutive NF-
B activity has emerged as an important cell survival component of physiological and pathological processes, including B-cell development. In B cells, constitutive NF-B activity includes p50/c-Rel and p52/RelB heterodimers, both of which are critical for proper B-cell development. We previously reported that WEHI-231 B cells maintain constitutive p50/c-Rel activity via selective degradation of IB
that is mediated by a proteasome inhibitor-resistant, now termed PIR, pathway. Here, we examined the mechanisms of PIR degradation by comparing it to the canonical pathway that involves IB kinase-dependent phosphorylation and ß-TrCP-dependent ubiquitylation of the N-terminal signal response domain of IB. We found a distinct consensus sequence within this domain of IB for PIR degradation. Chimeric analyses of IB and IBß further revealed that the ankyrin repeats of IB, but not IBß, contained information necessary for PIR degradation, thereby explaining IB selectivity for the PIR pathway. Moreover, we found that PIR degradation of IB and constitutive p50/c-Rel activity in primary murine B cells were maintained in a manner different from B-cell-activating-factor-dependent p52/RelB regulation. Thus, our findings suggest that nonconventional PIR degradation of IB may play a physiological role in the development of B cells in vivo.
|
INTRODUCTION
|
|---|
NF-B is a family of transcription factors that regulate diverse cellular functions in response to a wide range of stimuli (19). NF-B is typically found as a homo- or heterodimer of p50 (NFB1), p52 (NFB2), RelA (p65), RelB, or c-Rel. Regulation of NF-B is mediated by a family of inhibitor molecules called IB proteins, including IB, IBß, IB
, IB
/p105, IB
/p100, and Bcl-3. Most IB proteins associate with NF-B dimers to cause their localization in the cytoplasm. Upon activation by a wide variety of structurally and functionally distinct signals, NF-B dimers regulate the expression of a multitude of genes important for cell survival, inflammatory responses, and immune cell development (31). Thus, the NF-B system serves as an important paradigm for understanding how distinct signals activate gene expression via activation of signal transduction pathways.
Inducible NF-B activation through IB degradation has been studied extensively and is identified by several hallmark traits (reviewed in reference 19). Most cell types contain inactive NF-B/IB complexes in their cytoplasm. Upon stimulation with a variety of inducers, including tumor necrosis factor alpha (TNF-), bacterial lipopolysaccharide (LPS), and the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA), IB is phosphorylated by the IB kinase (IKK) complex on the N-terminal residues Ser32 and Ser36 (14, 38). Dually phosphorylated IB is directly recognized by the E3 ligase ß-transducin repeat-containing protein (ß-TrCP) (60, 70, 71). Polyubiquitin chain formation on Lys21 and/or Lys22 of IB is then catalyzed by ß-TrCP and leads to 26S proteasome-dependent degradation of IB and the release of free NF-B heterodimers to direct transcription in the nucleus (7). This "canonical" degradation of IB does not require its ankyrin repeats or association with NF-B because unrelated proteins (i.e., glutathione S-transferase [GST] and p53) could be efficiently targeted to this degradation pathway when the N and C termini of IB were fused to the corresponding termini of these unrelated proteins (4, 65, 68). This canonical pathway is believed to contribute to many NF-B activation pathways (19, 31).
The specificity of ß-TrCP-dependent ubiquitylation of IB involves the consensus ß-TrCP recognition motif DpSG
XpS (where pS is phosphor-Ser, is a hydrophobic amino acid, and X is any amino acid) (67). This motif is shared by a number of ß-TrCP substrates, including the IB family members, IB, IBß, and IB, and the tumor suppressor ß-catenin. A recent study of the crystal structure of ß-TrCP complexed to a ß-catenin peptide revealed that phosphorylated Ser residues corresponding to Ser32 and Ser36 of IB and the Asp residue corresponding to Asp31 of IB are absolutely necessary for proper contacts. Additionally, at the corresponding Gly33 of IB, the binding pocket of ß-TrCP appears to accommodate only Gly (67). Thus, preclusion of IB recognition by ß-TrCP by mutagenesis of IB (60) or competition peptides (70) can efficiently inhibit canonical NF-B activation pathways. Accordingly, pharmacological suppression of proteasome activity (42, 63) also prevents these pathways.
Recent studies have described, in addition to the canonical NF-B activation pathway, the presence of several different alternative NF-B activation mechanisms. For example, noncanonical IKK-dependent processing of NF-B2 (p100) to p52 generates active p52/RelB heterodimers by the human T-cell leukemia virus Tax protein and in developing murine B cells (53, 58, 69). Silica or pervanadate treatment of cells can also permit phosphorylation of IB on Tyr42, followed by dissociation from NF-B, thereby leading to activation of NF-B (26, 30). Additionally, calpain-dependent, but proteasome-independent, IB degradation has been reported during TNF signaling and calcium-regulated developmental processes (21, 45, 46). Moreover, the C terminus of IB contains casein kinase 2 (CK2) phosphorylation sites that regulate degradation of free IB via a proteasome degradation pathway (37, 51, 64). Thus, NF-B activation mechanisms are not limited to the canonical pathway, and distinct mechanisms may play key roles in specific physiological and pathological processes.
While knowledge regarding inducible NF-B activation pathways has been expanding extensively, as described above, less is known about the regulatory repertoires employed by constitutive NF-B activation pathways. Although NF-B activity was originally discovered in B cells as a constitutively nuclear factor that bound to the enhancer element of the light-chain gene, most of the details of this pathway remain elusive (52). Recent literature suggests that there are two major NF-B heterodimers necessary for B-cell development and survival: p50/c-Rel and p52/RelB (9, 20, 35). It has recently been shown that B-cell-activating-factor (BAFF)-dependent signaling, via the BAFF receptor, leads to the processing of NF-B2 via a "noncanonical" pathway to generate active p52/RelB heterodimers, a critical step in B-cell development (9). While it has been shown that p50/c-Rel complexes comprise the majority of the constitutively nuclear NF-B DNA binding complexes in murine splenic B cells (20, 40), the biochemical mechanism underlying activation of these complexes is not well understood.
In addition to that associated with B-cell development, constitutive NF-B activity is maintained by various cancer cell types, including Reed-Sternberg cells, diffuse large-B-cell non-Hodgkin's lymphomas, breast cancers, and head and neck squamous cell carcinomas (2, 13, 16, 59). One prominent target gene for NF-B in different cell types is the IB gene (8, 62). Therefore, it is puzzling that the augmented synthesis of IB found in cells with constitutive NF-B activity fails to inactivate these constitutively nuclear heterodimers. One way for cells to maintain constitutive NF-B activity is through continuous degradation of IB via genetic or epigenetic modifications of the canonical NF-B activation pathway. Consistently, unrelenting IKK activity is found in several cancer cell types to maintain rapid IB turnover and constitutive NF-B activity (13, 16, 18). In contrast, we have previously identified a pathway in the WEHI-231 murine B-lymphoma cell line and to some extent in normal murine splenic B cells that utilizes rapid and continuous degradation of IB in a proteasome-independent manner to maintain constitutive p50/c-Rel activity (17, 41, 56). Interestingly, even though the IKK phosphorylation sites and ß-TrCP recognition motifs are conserved among major IB family members, this pathway is selective for IB in these B cells. Moreover, calcium chelators and calmodulin inhibitors blocked this pathway without affecting LPS-inducible IB degradation. While it has been suggested that CK2-dependent phosphorylation of the C-terminal sequences can lead to calpain-mediated IB degradation in WEHI-231 B cells (54), other studies have shown that removal of the C terminus of IB does not affect this degradation (56a). Here we term this constitutive IB degradation PIR, for proteasome inhibitor-resistant IB degradation, to distinguish it from other known IB degradation pathways.
In this study, we examined the molecular requirements necessary for PIR degradation of IB through mutagenesis of the N-terminal IKK phosphorylation and ubiquitylation residues and ß-TrCP interaction sites of IB in the WEHI-231 cell background. We now provide several lines of evidence that the canonical and PIR IB degradation pathways diverge downstream of IKK activity. Additionally, we find that the ankyrin repeats of IB, but not IBß, contain additional information necessary to selectively target IB for PIR degradation. Moreover, we find that PIR IB degradation and constitutive activation of p50/c-Rel heterodimers in normal murine B cells do not require the same BAFF-dependent signaling pathway, via the BAFF receptor, necessary for p100 processing to p52. Therefore, our findings support the notion that the maintenance of constitutive p50/c-Rel activity occurs in a manner distinct from that for the recently elucidated BAFF-dependent NF-B activation pathway in developing B cells.
|
MATERIALS AND METHODS
|
|---|
Cell culture and chemicals.
WEHI-231 and W231.Bcl-XL cells and all derivatives of these cell lines were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (HyClone Laboratory, Inc.), 5 x 10–5 M ß-mercaptoethanol, and 1,250 U of penicillin G (Sigma) and 0.5 mg of streptomycin sulfate (Sigma) per ml in a 5% CO2 humidified incubator (Forma). Human embryonic kidney 293 (HEK 293) cells and all derivatives were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics as described above in a humidified incubator containing 10% CO2. Tissue culture plates were coated with 0.1% (wt/vol) gelatin for HEK 293 cells. W231.Bcl-XL cells were maintained in 300 µg of G418 (Mediatech)/ml. Puromycin-resistant cell lines were maintained in 2 µg of puromycin (Sigma)/ml. Hygromycin-resistant cell lines were maintained in 500 µg of hygromycin (Roche)/ml.
Mice.
Male and female A/J (wild-type BAFF receptor) and AW.Bcmd-2c (mutant BAFF receptor) mice were from a pathogen-free mouse colony in the Department of Biochemistry at the University of Wisconsin (10). The mice were maintained at 23°C with 40 to 60% humidity and 12-h light-dark cycles and were used at age 7 to 10 weeks for B-cell subset analysis. The protocols were approved by the Institutional Animal Care and Use Committee (protocol A00847-4-08-99).
Chemicals.
Cycloheximide, Bay 11-7082, dimethyl sulfoxide (DMSO), ATP, TPA, and LPS (L2880) were purchased from Sigma-Aldrich. Benzyloxycarbonyl-aspartyl-glutamyl-valyl-aspartate (z-DEVD-fmk), benzyloxycarbonyl-tyrosyl-valanyl-alanyl-aspartate (z-YVAD-fmk), and benzyloxycarbonyl-valanyl-alanyl-aspartate (z-VAD-fmk) were purchased from Alexis Biochemicals. Benzyloxycarbonyl-leucyl-leucyl-leucinal (MG132) was purchased from Peptide Institute, Inc. Benzyloxycarbonyl-leucyl-norleucinal (calpeptin), clasto-lactacystin-ß-lactone, and human recombinant TNF- were purchased from Calbiochem.
Antibodies.
Immunoglobulin G (IgG) antibodies against IKK (M-280), IB (C-21), IBß (C-20), actin (C-11), c-myc (9E10), HA (Y-11), p65 (C-20), p50 (NLS), RelB (C-19), p52 (I-18), and ICAD (FL-331) were purchased from Santa Cruz Biotechnology. A polyclonal antibody generated against p52 (06-413) was obtained from Upstate Biotechnology. A monoclonal anti-HA.11 antibody was purchased from Covance. A monoclonal antihemagglutinin (anti-HA) antibody (3F10) directly conjugated to horseradish peroxidase was purchased from Roche Molecular Biochemicals. A monoclonal anti--tubulin antibody was purchased from Oncogene Research. Affinipure F(ab')2 fragment goat anti-mouse IgM µ-chain-specific antibodies were purchased from Jackson Immunoresearch Laboratories. The 5432 antibody, generated against the N terminus of IB, was used as previously described (39). The c-Rel antibody (5071) was previously described (27). Horseradish peroxidase-conjugated protein A and horseradish peroxidase-conjugated anti-rabbit and anti-mouse antibodies were obtained from Amersham Pharmacia Biotech.
B-cell collection and flow staining analysis.
Splenocytes were dissociated by mechanical shearing, and red blood cells (RBC) were depleted by lysing them in an RBC lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM EDTA [pH 7.2 to 7.4]). When B-cell purification was performed, B cells were purified by negative selection with the magnetically activated cell sorting (MACS) B-cell purification kit (Miltenyi Biotech). For comparison of NF-B regulation in transitional B cells between the wild-type and BAFF receptor mutant backgrounds, autoreconstitution studies were initiated by subjecting A/J and AW.Bcmd-2c mice to 500 rads of sublethal irradiation in order to deplete mature splenic B cells (5). Irradiated mice were then allowed to recover for 13 days, at which point the majority of splenic B cells represented the transitional B cells, as revealed by positive surface staining of C1qRp (previously termed 493 and AA4) (1) and analyzed on a FACScan or FACScalibur using CELLQuest software (data not shown).
Mutagenesis, infections, and transfections.
Substitution mutagenesis was performed by two-step PCR using murine IB cDNA, and all mutant IB cDNAs were directly sequenced by the University of Wisconsin Biotechnology Center to verify integrity. cDNAs for murine K3R- and K5R-IB were kindly provided by Kei Tashiro (Kyoto University, Kyoto, Japan). The cDNAs encoding various forms of IB were ligated into the pLHL-CA retroviral vector (64) or pLPL-CA retroviral vector. PLPL-CA was derived from pLHL-CA and the pRetro-On vector (Clontech). PLHL-CA was digested with EcoRI and XhoI to isolate the region containing the CA promoter and the multiple cloning sites. This fragment was ligated into the pRetro-On vector (Clontech) between the long terminal repeat regions and downstream of the puromycin resistance gene. The c-myc-
F-ß-TrCP cDNA was kindly provided by Zhijian Chen (University of Texas Southwestern Medical Center). This cDNA was ligated into the pLPL-CA vector for use in the coimmunoprecipitation assays.
Transfection of HEK293 cells was performed by calcium phosphate precipitation as described previously (24). For experiments involving transient transfection of HEK293 cells, cells were treated with the appropriate inducers 40 to 45 h after transfection. To generate stable HEK 293 cells, cells were transfected with vectors and selected with 2 µg of puromycin (Sigma)/ml 24 to 48 h after transfection until drug-resistant pools grew up. Infections of WEHI-231 and W231.Bcl-XL cells were performed as previously described (41).
Analysis of IB degradation.
To measure IB degradation, cells were counted and 0.5 x 106 to 1 x 106 cells were placed in 1-ml wells in an incubator containing 5% CO2 for 2 to 4 h to reduce low-level, proteasome-dependent NF-B activation associated with medium change (our unpublished observations). Cells were then treated with the appropriate chemicals for various times. Cells were pelleted at 13,000 x g for 10 s in an Eppendorf centrifuge, washed once with phosphate-buffered saline (PBS), pelleted again at 13,000 x g for 10 s, and stored at –70°C. Pellets were thawed, resuspended in 20 µl of PBS and 20 µl of 2x Laemmli buffer, and boiled directly for 10 to 15 min. Samples were then electrophoresed in sodium dodecyl sulfate-10 or 12.5% polyacrylamide gels, electroblotted (Bio-Rad) onto a polyvinylidene fluoride membrane (Millipore), and then incubated with the appropriate antibodies as described previously (41). Western blots were analyzed by enhanced chemiluminescence as described by the manufacturer (Amersham).
To measure IB degradation in B cells from A/J and AW.Bcmd-2c mice, cells were counted and 0.5 x 106 to 2 x 106 cells were placed in 1-ml wells and treated similarly to WEHI-231 cells. Cells were then pelleted at 5,000 x g for 5 min in a refrigerated swinging-bucket centrifuge, washed once with PBS, pelleted again at 5,000 x g for 5 min as described above, and stored at –70°C. Pellets were thawed and subjected to Western blot analysis as described above.
Coimmunoprecipitation assays.
For ß-TrCP/IB coimmunoprecipitation experiments using HEK293 cells, cell pellets were resuspended in small amounts of PBS (10% of the final lysis buffer volume), lysed in a lysis buffer (20 mM Tris [pH 7.0], 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 0.5% NP-40, 2 mM dithiothreitol [DTT]) containing protease inhibitors (0.5 mM phenylmethylsulfonyl fluoride, 1 µg of leupeptin/ml, 16 µg of aprotinin/ml) and phosphatase inhibitors 20 mM ß-glycerolphosphate, 1 mM sodium orthovanadate, 10 mM p-nitrophenylphosphate, 10 mM sodium fluoride), and spun at 13,000 x g in an Eppendorf centrifuge at 4°C. Equal protein amounts were separated for immunoprecipitation. Twenty percent of each immunoprecipitation volume was allocated separately, 2x Laemmli buffer was added, and the samples were boiled to examine input levels by Western blot analysis. The anti-c-myc antibody was added to the lysates for 1 h at 4°C, and then protein G-Sepharose beads (Amersham Pharmacia Biotech) were added and the samples were rotated for an additional 3 to 4 h at 4°C for immunoprecipitation reactions. Immunoprecipitates were washed three times in excess immunoprecipitation buffer. Laemmli buffer (2x) was added to the beads and boiled, and the entire sample was analyzed by Western blotting as described previously (41).
To examine IB/NF-B interactions in W231.Bcl-XL cells, cell pellets were lysed as described above, equal amounts of extracts were separated for each immunoprecipitation, and 20% of the input was set aside and analyzed as described above. The appropriate antibodies and protein G-Sepharose beads were added to immunoprecipitate the NF-B/IB complexes. Immunoprecipitates were rotated overnight at 4°C, washed three times with excess lysis buffer, boiled in 2x Laemmli buffer, and analyzed by Western blotting.
EMSA and in vitro kinase assays.
W231.Bcl-XL cells and primary B cells were counted, and 2 x 106 cells were placed in 1-ml wells and allowed to rest for 2 to 4 h in a 5% CO2 incubator at 37°C. Cells were then treated with the appropriate chemicals and spun down as described above. For electrophoretic mobility shift assays (EMSA) with whole-cell extracts, cell pellets were lysed in Totex buffer (20 mM HEPES [pH 7.9], 350 mM NaCl, 20% glycerol, 1% NP-40, 1 mM MgCl2, 0.5 mM EDTA, 0.1 mM EGTA, 0.5 mM DTT) containing protease inhibitors, spun down at 13,000 x g in an Eppendorf centrifuge for 10 min at 4°C, and subjected to an EMSA as described previously (40). For EMSA with nuclear extracts, cell pellets were first lysed in buffer A (10 mM HEPES [pH 7.9], 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT) containing protease inhibitors and spun at room temperature for 10 s in an Eppendorf centrifuge at 13,000 x g. The remaining nuclear pellet was lysed in buffer C (20 mM HEPES, 25% glycerol, 0.45 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT) containing protease inhibitors for 30 min on ice and then spun at 13,000 x g for 10 min in an Eppendorf centrifuge at 4°C and the supernatant was removed and used as the nuclear extract. The Ig-B oligonucleotide probe was as described previously (40). In vitro kinase assays were performed as described previously (23).
|
RESULTS
|
|---|
PIR IB degradation requires the IKK phosphorylation sites, but not the N-terminal ubiquitylation sites, of IB.
Although the PIR pathway and canonical NF-B activation possess differential proteasome requirements for IB degradation, we wanted to determine whether both degradation pathways required the upstream IKK phosphorylation sites in IB. We have previously reported that a mutant IB with Ser-to-Ala substitutions (S32/36A-IB) at both the 32 and 36 positions underwent constitutive proteolysis in the WEHI-231 cells (41). We have found that this degradation was associated with an inadvertent apoptotic process introduced by the assay condition (see Fig. S1 in the supplemental material). To eliminate this apoptosis-associated IB degradation, we generated the W231.Bcl-XL cell line, stably expressing the antiapoptotic Bcl-XL gene, and found that S32/36A-IB was resistant to PIR degradation, similar to what was found for the LPS-inducible pathway (Fig. 1B, lanes 5 to 8). Additionally, PIR IB degradation in W231.Bcl-XL cells was also prevented by the inhibition of IKK activity (Fig. 1E to G). Degradation of similarly introduced wild-type IB and the endogenous version served as positive controls in these experiments. These results suggested that both IKK activity and the IKK phosphorylation sites were important for PIR IB degradation in these cells.
Since constitutive IB degradation is mediated in a PIR manner in both WEHI-231 and W231.Bcl-XL cells, there must be a divergence point between the canonical and PIR degradation pathways upstream of the proteasome degradation step. Thus, we next investigated the requirement for N-terminal lysine residues in IB whose ß-TrCP-dependent ubiquitylation leads to proteasome-dependent degradation in canonical pathways (19). Since we found a mutant IB with Arg substituted at both Lys21 and Lys22 to be efficiently degraded by the control canonical LPS pathway (24), we next utilized a mutant IB K3R-IB (Fig. 1A), containing Lys21, -22, and -67 mutated to Arg. Surprisingly, K3R-IB was still susceptible to both types of degradation in W231.Bcl-XL cells (Fig. 1B, lanes 9 to 12), suggesting that alternative Lys residues compensated for the lack of these three lysines. Therefore, we used K5R-IB, containing Lys21, -22, -38, -47, and -67 mutated to Arg and stably expressed it in W231.Bcl-XL cells. K5R-IB was completely resistant to LPS-inducible degradation (Fig. 1C, lane 7). However, it was still sensitive to PIR degradation (Fig. 1C, lanes 4 and 5). Moreover, on a high-percentage gel that was run for a prolonged period to separate the IB phosphorylated isoforms, the stable K5R-IB underwent signal-dependent phosphorylation, as evidenced by the characteristic migration shift found in cells treated with LPS (Fig. 1D, lane 3). This demonstrated that the lack of LPS-dependent degradation was due to the lack of ubiquitylation, not IKK-dependent phosphorylation. In contrast, K5R-IB underwent PIR degradation without the appearance of a slower-migrating phosphorylated form (Fig. 1D, lane 2). Thus, the N-terminal Lys residues required for canonical degradation were not absolutely required for PIR degradation, suggesting that these degradation pathways diverge at the ubiquitylation step.
Mutations at Asp31 of IB further reveal differences between PIR and LPS-inducible IB degradation pathways in W231.Bcl-XL cells.
Because canonical and PIR forms of IB degradation appeared to diverge primarily at the level of ubiquitylation, we wanted to determine whether or not recognition by the E3 ligase ß-TrCP was involved in the PIR degradation pathway. The recent crystal structure of ß-TrCP (67) revealed that an Asp residue corresponding to Asp31 of IB is critical for recognition by ß-TrCP. To examine the requirement for this site in PIR and LPS-inducible IB degradation, we generated the conservative mutant IB D31E-IB and stably expressed it in W231.Bcl-XL cells. This mutant IB was still susceptible to PIR (albeit less efficiently than wild-type IB) and LPS-inducible (at a level similar to that for wild-type IB) degradation (Fig. 2E), suggesting that either acidic amino acid could be tolerated at this position for both pathways. Subsequently, we mutated Asp31 to a drastically different amino acid, Phe. Surprisingly, we found that D31F-IB was completely resistant to PIR degradation but was still susceptible to LPS-inducible degradation in these cells (Fig. 2F). This result suggested that it might be possible to isolate other mutant IBs that reveal differences between PIR and LPS-inducible IB degradation pathways by amino acid replacement of the Asp31 residue. Therefore, we mutated residue 31 in IB to all of the remaining amino acids individually. We found that substitution of any other amino acid at this position rendered IB resistant to PIR degradation with the exception of partial degradation exhibited by D31V-IB (Fig. 2V). Very surprisingly, all of the mutant IBs except D31A- and D31Q-IB were still tolerated by the LPS-inducible and proteasome-dependent degradation pathways (Fig. 2). In all of these analyses, degradation of the endogenous IB protein served as a positive control for comparison.
Because the Asp31 requirement for LPS-inducible degradation discussed above appeared to conflict with the requirement for Asp at this position in the recently described ß-TrCP crystal structure (67), we next evaluated whether inducible degradation of these mutant IBs was potentially due to some unusual LPS-specific event in W231.Bcl-XL cells. In Fig. 3A, we observed that IgM cross-linking and treatment with TPA also led to inducible degradation of both D31E- and D31F-IB. Moreover, inhibition of degradation by treatment with the proteasome inhibitors MG132 and clasto-lactacystin ß-lactone, but not with the calpain inhibitor calpeptin (Fig. 3B), indicated that degradation occurred through the proteasome pathway. Furthermore, inhibition of LPS-inducible D31G-IB degradation by Bay 11-7082 (Fig. 3C, lanes 7 and 11) suggested that this pathway was also IKK dependent. Therefore, IB could be inducibly degraded in a manner dependent on both IKK activity and the proteasome with a different sequence specificity of ß-TrCP recognition in W231.Bcl-XL cells than what was suggested by the recently described crystal structure (67).
View larger version (49K):
[in this window]
[in a new window]
|
FIG. 3. Inducible degradation of Asp31 mutant IBs occurs by multiple inducers and is inhibited by proteasome and IKK inhibitors in W231.Bcl-XL cells. (A) W231.Bcl-XL cells stably expressing D31E- and D31F-IB were treated with 25 µg of cycloheximide (Cx)/ml in the absence or presence of MG132 (10 µM) for 3 h. To measure inducible degradation, cells were treated with 20 µg of LPS/ml, 10 µg of Affinipure F(ab')2 fragment goat anti-mouse IgM µ chain specific/ml, or 50 nM TPA for 30 min in the absence of MG132 or after pretreatment with 10 µM MG132. Whole-cell pellets were examined by Western blot analysis. OT, untreated sample. (B) W231.Bcl-XL cells stably expressing D31F-IB were left untreated or treated with 20 µg of LPS/ml in the absence of MG132 or after pretreatment with 10 µM MG132 (MG), 1 to 20 µg of calpeptin/ml, or 1 to 10 µM clasto-lactacystin ß-lactone. Whole-cell pellets were examined by Western blot analysis. (C) W231.Bcl-XL cells stably expressing D31G-IB were treated with 25 µg of cycloheximide/ml for 0 to 3 h. To measure inducible degradation, cells were treated with 20 µg of LPS/ml in the absence of MG132 or after pretreatment with 10 µM MG132, 20 µg of calpeptin (Calp)/ml, 10 µM clasto-lactacystin ß-lactone (Lact), or 25 µM Bay 11-7082 (Bay).
|
|
PIR and canonical IB degradation pathways diverge at the point of ß-TrCP recognition.
Because PIR and LPS-inducible IB degradation possessed differential sensitivities to mutagenesis of the N-terminal Lys and Asp31 residues in W231.Bcl-XL cells, we next examined whether these two pathways also displayed differential requirements for other residues within the ß-TrCP consensus sequence. We introduced a series of mutations in the ß-TrCP recognition motif of IB (DpSGXpS) that could be used to determine whether PIR and canonical IB degradation pathways diverge at the point of ß-TrCP recognition (Fig. 4A). We chose conservative and nonconservative substitutions at each amino acid position and analyzed them as described above. We found a number of mutant IBs with mutations in this motif that could still undergo PIR degradation in W231.Bcl-XL cells, including G33A-, L34I-, L34R- D35E-, and D35G-IB (Fig. 4B to D). Interestingly, G33R-IB could not support PIR degradation (Fig. 4B, bottom). Furthermore, the same mutant IBs that underwent PIR degradation were also susceptible to LPS-inducible degradation in the W231.Bcl-XL cells. In addition to the surprising data obtained from the Asp31 mutant IBs, degradation of the nonconservative mutant IB L34R-IB was still permitted by both pathways. We thus questioned whether our observations of LPS-inducible degradation in W231.Bcl-XL were representative of the canonical, ß-TrCP-dependent IB degradation.
To confirm whether these ß-TrCP recognition mutant IBs were indeed recognized by ß-TrCP in a manner dependent on phosphorylation at Ser32 and Ser36 residues, we employed a coimmunoprecipitation method that allowed us to trap phosphorylated IB associated with ß-TrCP (23). Since transient cotransfection of ß-TrCP and IB resulted in phosphorylation-independent association (data not shown), we needed to generate HEK293 cell lines stably expressing the mutant IBs; then either c-myc-F-ß-TrCP or an empty vector was transiently transfected for coimmunoprecipitation studies. Cells were then either left untreated or treated with TNF- in the presence of MG132. Under these conditions, ß-TrCP failed to interact with wild-type IB in the absence of a signal but efficiently interacted after TNF-/MG132 treatment (Fig. 5A). We next evaluated whether mutant IBs that exhibited differential degradation patterns under PIR and LPS-inducible degradation conditions in W231.Bcl-XL cells could also display differential recognition by ß-TrCP. Interestingly, we found that association of D31E-IB with ß-TrCP was markedly less than that of wild-type IB (Fig. 5B, compare lanes 2 and 4), a result that is consistent with the critical requirement of the Asp at position 31 predicted by the crystal structure (67). We next examined the association of ß-TrCP with several Asp31 mutant IBs and the other ß-TrCP consensus sequence mutant IBs described above. Consistent with the prediction from the crystal structure (67), none of the Asp31 substitutions that we tested were tolerated for inducible association with ß-TrCP (Fig. 5F). Strikingly, G33A-IB completely failed to interact with ß-TrCP (Fig. 5C, compare lanes 2 and 4), even though this substitution introduced a very modest change in the amino acid side chain. The crystal structure indicated that there is very little room in the ß-TrCP recognition pocket for any substitution at this position (67). Accordingly, G33R-IB did not interact at all (Fig. 5C). In contrast, a replacement of Leu34 with Ile or a change of the hydrophobicity by replacement with Arg did not impair association with ß-TrCP (Fig. 5D, lane 6). Both D35E- and D35G-IB also efficiently interacted (Fig. 5E). The data from these studies are summarized in Table 1. Additionally, we found that S36T permitted PIR degradation and ß-TrCP recognition but that S32T did not (data not shown). These observations together demonstrated that (i) PIR degradation of mutant IBs did not strictly correlate with their ß-TrCP interaction properties, as exemplified by the differential effects of D31E- and G33A-IB; (ii) the N-terminal consensus motif for PIR degradation is (D/E)-(p)S-(G/A)-X-X-(p)(S/T), where (p) indicates that direct observation of phosphorylation at these sites has not yet been established during this degradation process; and (iii) signal-inducible degradation of mutant IBs could occur via an alternative pathway that did not absolutely require an intact ß-TrCP recognition sequence in W231.Bcl-XL cells. These findings indicated that PIR degradation did not utilize direct interaction of the ß-TrCP E3 ubiquitin ligase with IB. This conclusion was also supported by the lack of the requirement for the N-terminal ubiquitylation Lys residues and resistance to various proteasome inhibitors of PIR degradation. Nevertheless, because of the similarities of the consensus motifs for PIR degradation, (D/E)-(p)S-(G/A)-X-X-(p)(S/T), and ß-TrCP interaction, DpSGXpS (67), these studies suggest that a protein with a distinct but somewhat related specificity of interaction may be involved in diverting IB to PIR degradation and away from the canonical pathway.
View this table:
[in this window]
[in a new window]
|
TABLE 1. Summary of stably expressed mutant IBs analyzed for PIR-mediated degradation, LPS-inducible degradation, and ß-TrCP interactiona
|
|
Chimeric IB/IBß proteins reveal that the ankyrin region of IB contains information necessary for PIR degradation.
While our data thus far supported the notion that the N terminus of IB contains a consensus motif for PIR degradation, IBß also contains this primary structure. Yet, only IB is targeted to PIR degradation, whereas IBß is spared (41). Thus, it was possible that the above consensus sequence was incomplete and that the subtle differences in the N-terminal sequences between these two proteins dictated their susceptibility to PIR degradation. To determine whether the N-terminal domain of IB contained the full information necessary for PIR degradation, we generated a chimeric protein by replacing the N terminus of IBß with the corresponding IB sequence (Fig. 6A, ßß) based on the NF-B/IB and NF-B/IBß cocrystal structures (25, 28, 36). As a control, we also generated the reciprocal mutant IB (ß). Since free IB is known to undergo proteasome-dependent degradation via a pathway distinct from PIR degradation (33, 43, 51), we initially used the ability of these C-terminally HA-tagged chimeras to stably interact with NF-B in W231.Bcl-XL cells as a measure of their ability to fold correctly. We found that both ßß and ß efficiently associated with c-Rel and p65 (see Fig. S2B in the supplemental material). The degradation assay indicated that ßß underwent rapid turnover when stably expressed in W231.Bcl-XL cells (Fig. 6B, top). However, this degradation was highly sensitive to the proteasome inhibitor MG132. In the same extracts, PIR degradation of endogenous IB remained highly resistant to MG132 treatment while degradation of endogenous IBß was sensitive. In contrast, ß degradation was slower than that of the endogenous IB but remained MG132 resistant (Fig. 6C, top). These results showed (i) that the N-terminal sequence of IB could enhance degradation in the context of the IBß ankyrin repeats and the C terminus but that this targeting was incomplete for PIR degradation and (ii) that the N-terminal IBß sequence could support PIR degradation in the context of the IB ankyrin repeats and C terminus, even though it was not as efficient as the N-terminal IB sequence. These findings indicated that the principal mechanism behind IB selectivity of PIR degradation resided outside of the relatively conserved N-terminal sequence.
To localize the second cis element required for PIR degradation of IB, further chimeric proteins from IB and IBß were generated as outlined in Fig. 6A. Once again, efficient association between c-Rel and p65 was used to assess proper folding. While ß and ßß associated with NF-B proteins efficiently, ß and ßß did not (see Fig. S2 in the supplemental material). The analysis of constitutive degradation of the former two mutant IBs (Fig. 6B and C, bottom), along with the N-terminal swap mutant IBs described above, suggested that the presence of the ankyrin repeats of IB was essential for targeting to the PIR pathway (summarized in Table 2). Moreover, the presence of the C-terminal PEST domain of IB or IBß had no impact on PIR degradation. All of the chimeric proteins were susceptible to LPS-inducible degradation, indicating that the differential effects of these mutant IBs were selective for PIR degradation. Thus, in contrast to the dispensability of the ankyrin repeats for canonical degradation (4, 65, 68), this domain of IB was found to be essential for targeting it to PIR degradation.
PIR IB degradation does not require the same BAFF receptor-dependent signals required for p100 processing.
Recent studies have suggested that BAFF receptor signaling is critical for mature B-cell development via processing of p100 to p52 to form active p52/RelB heterodimers (9). In these studies A/WySnJ mice that harbor a mutation in the BAFF receptor were utilized to demonstrate the requirement for BAFF receptor signaling in p100 processing and the generation of mature B cells, similar to what is found for NFB1/NFB2-deficient mice (9, 50). In addition to a mutation in the Bcmd-1/BAFF receptor locus, A/WySnJ mice carry an additional mutation in the Bcmd-2 locus that is responsible for the enhanced mastocytosis observed in A/WySnJ mice (10). We thus employed AW.Bcmd-2c congenic mice that are defective only at the Bcmd-1/BAFF receptor locus but wild type at the Bcmd-2 locus. These mice allowed us to examine whether constitutive p50/c-Rel activity and PIR IB degradation occurred in the absence of BAFF receptor-dependent signals required to mediate p100 processing.
First, splenocytes were isolated from both A/J and AW.Bcmd-2c mice which have been autoreconstituted by sublethal radiation to capture mostly the transitional B cells in the spleen for direct comparison (1). When constitutive NF-B activity was analyzed by EMSAs, it was found that cells from the AW.Bcmd-2c mice had constitutive p50/c-Rel activity (Fig. 7A), which was surprisingly similar to that of the A/J counterpart, especially for the c-Rel component. As a control, we verified that p100 processing to p52 was indeed markedly reduced in B cells isolated from the AW.Bcmd-2c mice relative to that in B cells from A/J mice, while the expression levels of all other known NF-B family members were similar (Fig. 7B). Moreover, constitutive IB degradation occurred in B cells of both lines of mice at similar levels and was resistant to inhibition by the proteasome inhibitor MG132 (Fig. 7C). Finally, similar to the results for the W231.Bcl-XL cell line, we found that PIR IB degradation in B cells from both wild-type and mutant mice was sensitive to the IKK inhibitor Bay 11-7082. Our findings are consistent with the model in which neither PIR IB degradation nor constitutive p50/c-Rel activity requires the same BAFF receptor-dependent signaling events that are required for p100 processing.
|
DISCUSSION
|
|---|
We have previously identified a constitutive IB degradation pathway in WEHI-231 B cells that is proteasome inhibitor resistant (41). In this study we provide evidence that PIR IB degradation is mediated by an IKK-dependent, but ß-TrCP-independent, mechanism thereby explaining the lack of requirement for the N-terminal ubiquitylation sites and proteasome activity for IB degradation in these cells. Our analysis indicated that the PIR degradation pathway differs from the well-established canonical pathway (Fig. 8) in a significant way. While canonical IKK-dependent phosphorylation of IB and IBß leads to their recognition and ubiquitylation by ß-TrCP and subsequent degradation by the 26S proteasome (31, 55, 66), PIR degradation requires IKK activity but did not appear to require direct association with this E3 ubiquitin ligase. Consistently, a different consensus sequence for PIR degradation from that required for ß-TrCP recognition emerged through these mutagenesis studies. This suggests that a molecule other than ß-TrCP may potentially guide IB to PIR degradation. It is also important that this difference of E3 involvement may be due to the potential lack of direct phosphorylation of IB by IKK in the PIR pathway, even though the IKK activity is apparently critical. Despite the ease of its detection in the canonical pathway, we were unable to observe it in the PIR pathway.
A number of nonconventional IB degradation pathways that are distinct from PIR IB degradation have been identified. Studies have suggested that basal degradation of IB requires the C-terminal CK2-dependent phosphorylation sites and the ankyrin repeats, but this basal degradation is proteasome dependent and N terminus independent (33, 37, 51, 64). Calpain-mediated degradation has been associated with a number of NF-B activation processes, including TNF- in HepG2 cells, neoplastic cell development, and Her-2/neu regulation in breast cancer cells (21, 45, 46). It has also been observed that TNF--induced mitochondrial IB degradation is, at least partially, dependent on calpain (11). Serum deprivation can also promote phosphorylation- and ubqiuitylation-independent lysosomal degradation of IB (12). Moreover, caspase-dependent N-terminal cleavage of IB has also been observed to occur at Asp31 (3, 48). In contrast, PIR IB degradation requires both the N terminus and ankyrin repeats of IB. Furthermore, a mutant IB lacking the C-terminal PEST domain was still susceptible to PIR degradation, and the calpain inhibitor PD150606 (which targets the calcium binding domain of calpain) failed to block PIR degradation, indicating that calpain is unlikely to be directly involved in PIR degradation (56a). Moreover, D31E-IB still undergoes PIR degradation, suggesting that caspase-dependent cleavage of IB C-terminal to Asp31 is not involved in this pathway. Finally, from this study and our previous observations, PIR IB degradation is not sufficiently inhibited by proteasome, calpain, lysosomal, or caspase inhibitors, suggesting that PIR degradation of IB may occur through an alternative degradation pathway (41, 56, 56a).
While certain N-terminal residues of IB were critical for PIR degradation, these residues are conserved in IBß, a non-PIR degradation substrate in WEHI-231 and W231.Bcl-XL cells. Our chimeric studies using IB and IBß indicated that this N-terminal consensus sequence was insufficient for complete targeting to the PIR pathway. It has been reported that the canonical pathway requires the N terminus or both the N and C termini of IB but not the ankyrin repeats (4, 65, 68, 72). In contrast, PIR IB degradation was strictly dependent on its ankyrin repeats, even though it was supported by the N terminus of either IB or IBß. This dependence was so specific to the IB ankyrin repeats that even those of IBß could not substitute. Although the mechanism by which this specificity is carried out is unclear, previous studies have reported that differences in the ankyrin repeats of IB and IBß lead to their different patterns of nuclear localization and NF-B inhibition (6, 57). These studies suggested that the ankyrin repeats of IB and IBß have the capacity to perform different tasks. The major difference in ankyrin repeat structure between IB and IBß is the presence of a loop between ankyrin repeats 3 and 4 of IBß (25, 28, 36). When we analyzed the PIR degradation susceptibility of mutant IBs with the IBß loop introduced between ankyrin repeats 3 and 4 or mutant IB/IBß chimeras with the loop deleted, we were unable to observe a strict correlation between the absence of the loop domain and susceptibility to PIR degradation (S. O'Connor, unpublished observations). These results indicate that finer mapping of the ankyrin repeat domain is required to uncover the sequences required to target IB to PIR degradation. Moreover, our findings suggest that a molecular component(s) that recognizes IB likely detects both the N-terminal consensus sequence and an IB ankyrin repeat-specific motif to mediate PIR degradation (Fig. 8). Alternatively, we cannot rule out the possibility that different molecular components recognize the N terminus and the ankyrin repeats of IB for PIR degradation.
Interestingly, during these studies, we uncovered an alternative inducible IB degradation pathway in W231.Bcl-XL cells. We found that mutant IB
Top
Abstract
Introduction
