Tax Deregulation of NF-κB2 p100 Processing Involves Both β-TrCP-dependent and -independent Mechanisms*

Processing of the nf-κb2 gene product p100 to generate p52 is a tightly regulated event, consistent with the fact that the processing product, p52, is hardly detected in most cell types, including T cells, although the precursor p100 is expressed abundantly in these cells. However, in T cells transformed by the human T-cell leukemia virus type I (HTLV-I), p100 processing is very active, resulting in high level expression of p52. Because overproduction of p52 is associated with lymphoid hyperplasia and transformation, deregulation of p100 processing may be part of the oncogenic mechanism of HTLV-I. We demonstrated previously that HTLV-I Tax oncoprotein is a potent inducer of p100 processing through specific targeting of IKKα via IKKγ to p100 to trigger p100 phosphorylation and ubiquitination. In this study, we further show that Tax-mediated recruitment of IKKα to p100 requires serines 866 and 870 of p100, shown to be essential for inducible processing of p100. Upon interaction with p100, activated IKKα phosphorylates both N- and C-terminal serines of p100 (serines 99, 108, 115, 123 and 872), serving as a critical step in Tax-induced p100 processing. Using a genetic approach, we find that β-transducin repeat-containing protein, a component of the SCF ubiquitin ligase complex, previously shown to be required for physiological p100 processing mediated by nuclear factor-κB-inducing kinase, is only partially involved in Tax-induced processing of p100. These results indicate that both β-transducin repeat-containing protein-dependent and -independent mechanisms contribute to Tax-deregulated p100 processing, further suggesting the involvement of different mechanisms in cellular and viral pathways of p100 processing.

The oncogenic action of human T-cell leukemia virus type I (HTLV-I) 1 Tax protein involves activation of cAMP response element-binding protein/activating transcription factor and NF-B, two major cellular transcription factors (1,2). Tax activates transcription activity of cAMP response elementbinding protein/activating transcription factor proteins by directly interacting with their basic region-leucine zipper DNAbinding domains, thereby enhancing DNA binding (3)(4)(5). Unlike cAMP response element-binding protein/activating transcription factor, which is constitutively expressed in nucleus, NF-B is normally sequestered in the cytoplasm by ankyrin repeat-containing inhibitors called IB proteins (6). The nuclear expression of NF-B can be induced by various cellular stimuli, such as T-cell mitogens, proinflammatory cytokines, and antigens. These stimuli trigger an IB kinase (IKK) complex, which consists of IKK␣, IKK␤ (two catalytic subunits), and IKK␥ (regulatory subunit, also named NEMO), to phosphorylate specific serines within the IB sequence. The phosphorylated IB is then targeted for ubiquitination and proteasome-mediated degradation, allowing the NF-B to accumulate in the nucleus and transactivate target genes (7). Under normal conditions, this classic NF-B activation is tightly controlled and usually transient (6,7). However, in Tax-expressing cells, IKK is persistently activated, resulting in constitutive nuclear expression and transcription activity of NF-B. Tax induces activation of IKK complex by directly interacting with the regulatory subunit IKK␥ (8 -12).
In mammalian cells, there are five NF-B members: RelA (p65), RelB, c-Rel, p50, and p52, which function as various homo-and heterodimers (13). Unlike the Rel proteins, p50 and p52 are synthesized as large precursors NF-B1 p105 and NF-B2 p100, respectively (13)(14)(15). It is interesting that p105 and p100 also contain ankyrin repeats at their C-terminal regions and function as IB-like inhibitors of NF-B (16,17). Different from the complete degradation of other IB proteins, the proteasome-mediated degradation of p105 and p100 leads only to loss of their C-terminal ankyrin repeat regions, leaving intact N termini, p50, and p52, respectively (18,19). Thus, the processing of p105 and p100 not only plays a role in liberating specific NF-B complexes but also serves to generate p50 and p52. Although p105 processing is constitutive, the processing of p100 is tightly regulated through its inducible phosphorylation and polyubiquitination (20). Induction of p100 phosphorylation and subsequent processing is mediated by NF-B-inducing kinase (NIK) and its downstream kinase IKK␣ (20,21). It is interesting that neither IKK␤ nor IKK␥ is required for this novel, "non-canonical" NF-B pathway, although both IKK␤ and IKK␥ are essential for canonical NF-B activation (21,22). Therefore, the NIK/IKK␣-specific NF-B pathway can not be stimulated by most of the classic NF-B inducers; rather, it responds to signals involved in B cell differentiation and lymphoid organogenesis, including those triggered by lymphotoxin ␤ (20, 23), B-cell activating factor (BAFF) (24,25) and CD40 ligand (26). Those stimuli serve as physiological stimu-lators of NIK to induce processing of p100.
Not surprisingly, proper processing of p100 plays an essential role in the development and maturation of lymphoid organs. For example, nfb2 gene-deficient mice with no expression of both p100 and p52 show severe defects in B cell function and impairment in the formation of the proper architecture of peripheral lymphoid organs (27,28), a phenotype not observed in nfb1-deficient mice (29). Likewise, mice having other defects in this non-canonical NF-B pathway, such as the lymphotoxin ␣ or ␤ receptor knockout mice (30,31), the alymphoplasia (aly) mice (the NIK point mutant mice) (32), and the A/WySnJ mice (the BAFF-receptor deletion mutant mice) (33), which have normal p100 expression but no p52 expression (20,(23)(24)(25), also show those phenotypes, albeit to different extents. On the other hand, mice whose nfb2 gene was engineered to contain a mutation that prevents expression of the p100 C terminus leading to constitutive p52 expression but no p100 production, develop lymphoid and gastric hyperplasia (34). In humans, chromosomal translocations that cause the rearranged nfb2 gene to lose its C-terminal processing inhibitory domain, leading to constitutive processing of p100 (20), are associated with the development of various lymphomas (35)(36)(37). In addition, those constitutively processed forms of p100 have been reported to oncogenically transform fibroblasts in vitro (38). It is interesting that deregulated p100 processing is also found to be associated with T-cell transformation by HTLV-I, an etiological agent of an acute and fatal T-cell malignancy, adult T cell leukemia (22). Because p100 processing is inefficient in T cells, including activated T cells (22), overproduction of p52 in T cells is a hallmark of HTLV-I infection and transformation (39). It is noteworthy that stable expression of p100 could efficiently block HTLV-I Tax-mediated transformation (40), through specific inhibition of the function of the alternative NF-B pathway activated by HTLV-I (41). Thus, a fundamental understanding of p100 processing will not only provide important insights into the development of both human immune system and NF-B-associated diseases but may also suggest effective therapeutic strategies for HTLV-I-induced adult T cell leukemia and other p100 processing associated diseases (42)(43)(44).
Our previous studies have demonstrated that Tax functions as a potent inducer of p100 processing (22). Unlike physiological p100 processing, which involves NIK and IKK␣, but not IKK␥ (20,21), Tax induces p100 processing by specifically targeting IKK␣ indirectly via IKK␥ into p100-containing complexes, triggering p100 ubiquitination and subsequent processing (22). We have recently discovered that NIK-mediated processing of p100 also involves recruitment of IKK␣ to p100. It is interesting that purified IKK␣ could phosphorylate serines located in both N-and C-terminal regions of p100, which are involved in NIK-mediated processing of p100 (45). In the present study, we have demonstrated that Tax recruitment of IKK␣ to p100 requires serines 866 and 870 of p100, which have been shown to be essential for inducible processing of p100. Taxactivated IKK␣ also consistently phosphorylates both N-and C-terminal serines of p100, a step required for the binding of p100 to ␤-TrCP E3 ligase and Tax-induced p100 processing. However, we also find that ␤-TrCP is only partially responsible for Tax-induced processing of p100, although NIK-mediated p100 processing depends on ␤-TrCP.
Immunoblotting (IB) and Co-Immunoprecipitation (Co-IP)-293 cells were transfected with the expression vectors as indicated in the figures and lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.25% sodium deoxycholate, 1% Nonidet P-40, 1 mM dithiothreitol, and 1 mM phenylmethylsulfony fluoride) supplemented with a protease inhibitor mixture, followed by IB or Co-IP assays as described previously (20,22). The approximate amounts of cell lysates were 6 g for IB and 300 g for Co-IP assays.
Cell Labeling and Pulse-chase Assays-Transfected 293 cells were starved for 1 h in Dulbecco's modified Eagle's medium lacking methionine and cysteine, followed by metabolic labeling for 45 min with 350 Ci/ml [ 35 S]methionine/[ 35 S]cysteine. The pulse-labeled cells were chased for different times in complete Dulbecco's modified Eagle's medium supplemented with 10 mM unlabeled methionine and lysed in RIPA buffer. The radiolabeled, myc-tagged, wild-type p52 or its mutant was isolated by IP using anti-myc antibody, fractionated by SDS-PAGE, and visualized by autoradiography (20,22).
In Vitro Kinase Assay-In vitro kinase assays were performed as described previously (11,12). Cell extracts from HTLV-I-transformed cells or control Jurkat T cells were incubated at 30°C for 20 min in a kinase buffer (20 mM HEPES pH 7.6, 20 mM MgCl 2 , 20 mM ␤-glycerophosphate, 1 mM EDTA, and 2 mM dithiothreitol) containing [␥-32 P]dATP and labeled substrates. The phosphorylated proteins were fractionated by SDS-PAGE, transferred to nitrocellulose membranes, and visualized by autoradiography. The membrane was subsequently used for IB to analyze the protein level of substrates.
In Vivo Ubiquitin Conjugation Assays-293 cells were transfected with HA-tagged ubiquitin and p100 or its mutants in the presence or absence of Tax. 36 -48 h after transfection, the cells were lysed in RIPA buffer and subjected to immunoprecipitation (IP) using anti-p100. Agarose beads were washed three times with RIPA buffer followed by two additional washes with RIPA buffer supplemented with 1 M urea. The ubiquitin-conjugated p10 eluted by SDS loading buffer was analyzed by IB using anti-HA-HRP (20,22).
SiRNA-mediated Gene Silencing-The sequences of double-stranded siRNA are: for ␤-TrCP, GUG GAA UUU GUG GAA CAU CTT (sense) and GAU GUU CCA AUU CCA CTT (antisense), and for green fluorescent protein, GCU ACC UGU UCC AUG GCC ATT (sense) and UGG CCA UGG AAC AGG UAG CTT (antisense). These siRNAs were synthesized by Dharmacon Research, Inc. (Lafayette, CO). 293 cells were transfected using LipofectAMINE 2000 (Invitrogen) in accordance with the manufacturer's instructions. In brief, 293 cells were transfected in 6-well plates with 60 pmol of siRNA mixed with 800 ng of carrier DNA (pCMV4 vector). Transfection was repeated 24 h later to achieve high efficiency of siRNA transfection, except that the indicated DNA expression vectors were used in place of carrier DNA. Twenty-four hours after the second transfection, whole-cell extracts were prepared and subjected to immunoblotting or to in vivo ubiquitin conjugation assays.

RESULTS
Tax Recruitment of IKK␣ into p100 Complexes Is Independent of the Kinase Activity of IKK␣-As we reported previously, Tax could specifically target IKK␣, but not IKK␤, into p100 complexes (22). To further test whether the kinase activity of IKK␣ or its active status is required for p100/IKK␣ association, both kinase constitutively active (IKK␣ SS/EE) and inactive (IKK␣ SS/AA or IKK␣ KA) forms of IKK␣ were used. As shown in Fig. 1, these IKK␣ mutants only weakly bound to p100 (top, lanes 1, 3, 5, and 7), which is consistent with their inability to induce p100 processing (bottom, lanes 1, 3, 5, and 7). However, these IKK␣ mutants, like wild-type IKK␣, could efficiently interact with p100 when Tax was coexpressed (top, lanes 2, 4, 6, and 8). These results indicated that the kinase activity of IKK␣ is not involved in its recruitment to p100 by Tax. The fact that the constitutively active form of IKK␣ (IKK␣ SS/EE) also failed to efficiently bind to p100 without Tax expression and did not induce p100 processing suggested that activation of IKK␣ is not sufficient, although required, for inducible processing of p100. Co-expression of the kinase-inactive forms of IKK␣ partially decreased Tax-induced p100 processing in a consistent fashion (bottom, lanes 6 and 8), whereas the constitutively active form marginally increased this ability of Tax (lane 4). These results are also in agreement with the fact that the expression level of endogenous IKK␣ is relatively high in 293 cells.
Deletion of the N-terminal Region of p100 Results in Constitutive p100/IKK␣ Interaction and Inducible p100 Degradation-To further define the "docking" site within p100 for IKK␣ interaction, we performed Co-IP assays using a series of systematic p100 N-and C-terminal deletion mutants (Fig 2A). In agreement with our previous report showing that Tax associates with p100 via two N-terminal ␣-helices of p100 (located between amino acids 151 and 180) (22), deletion of up to 124 amino acids from the p100 N terminus (p100 ⌬1-32, p100 ⌬1-71, p100 ⌬1-124) did not affect p100/Tax interaction ( Fig  2B, second panel from top, lanes 2, 4, 6, and 8). IKK␣ could still be recruited consistently to these p100 deletion mutants (top, lanes 2, 4, 6, and 8). Further deletion of 59 amino acids or more from this end (p100 ⌬1-183, p100 ⌬1-343) prevented p100/Tax interaction (second panel, lanes 10 and 12). It was interesting, however, that these two deletion mutants showed constitutive IKK␣ binding activities, as reflected by the fact that these p100 truncation mutants could directly interact with IKK␣ without Tax (top, lanes 9 and 11). On the other hand, co-expression of Tax actually slightly reduced the binding between IKK␣ and these two mutants of p100 (top, compare lanes 9 and 11 with lanes 10 and 12, respectively). The decreased interaction could be caused by the competition of the p100 mutants and Tax for IKK␣. These results indicated that the N-terminal sequences of p100 are not involved in Tax-induced IKK␣ binding of p100, but rather inhibit the constitutive activity of p100 in binding to IKK␣. These studies further suggested that the IKK␣ binding site of p100 is located in its C terminus.
Although p100 ⌬1-32 and p100 ⌬1-71 still responded to Tax-induced p100 processing (Fig. 2B, bottom, lanes 4 and 6), the p100 ⌬1-124 mutant was largely resistant to Tax-induced processing (lane 8), although it still retained the ability to bind to Tax and to recruit IKK␣ mediated by Tax (top and second panel, lane 8). It is noteworthy that Tax could still induce marginal processing of the p100 ⌬1-124 mutant (Fig. 2C, lane 4), suggesting that other sequences besides amino acid 1-124 of p100 might contribute to this event. Nevertheless, these results suggest that the sequences between residue 71 and 124 play an important role in Tax-mediated p100 processing. In contrast, p100 ⌬1-183 and p100 ⌬1-343 were not processed when Tax was co-expressed, but instead were degraded, as indicated by the loss of the precursors without appearance of processing products (Fig. 2B, bottom, lanes 10 and 12; Fig. 2C, lane 6). The failure to detect processing products was not a result of their instability, in that p52 ⌬1-183 (the presumptive processing product of p100 ⌬1-183) showed a half-life similar to that of p52 in our pulse-chase assays (Fig. 2D). The degradation of p100 ⌬1-183 and p100 ⌬1-343 induced by Tax seemed to be mediated by endogenous IKK, because Tax could not bind to these two p100 mutants (Fig. 2B, second panel, lanes 10 and  12). IKK␣ or IKK␤ consistently induced even more dramatic degradation of p100 ⌬1-183 (Fig. 2E). These results indicated that once it has lost the N-terminal sequences essential for inducible processing, p100 behaves similarly to typical IB proteins in undergoing complete degradation.
Tax-mediated Recruitment of IKK␣ into p100 Complexes Requires Serines 866 and 870 of p100 -Consistent with the results that the N-terminal sequences of p100 are not required for Tax-mediated recruitment of IKK␣ into p100 complexes, loss of 41 or more amino acids from the C terminus of p100 prevented Tax-mediated IKK␣ recruitment (Fig. 3B, top, lanes 4 -9), although these p100 mutants still retained the ability of wild-type p100 to bind to Tax (second panel from top, lanes [3][4][5][6][7][8][9]. These results further supported our previous studies showing that Tax associates with p100 via two N-terminal ␣-helices of p100 (22), and indicated that the last 41 amino acids are required for Tax-mediated IKK␣ recruitment to p100. Because serines 866 and 870 located in this region have been shown to be essential for inducible processing of p100 (20,22), we tested the IKK␣ binding activity of the p100 S866A/S870A mutant harboring these two serine-to-alanine substitutions, in the presence of Tax expression. As a control, the p100 S872A mutant, which harbors substitution of serine 872 with alanine, was also included in these experiments. It is interesting that mutation of serines 866 and 870, but not serine 872, blocked the recruitment of IKK␣ in Tax-expressing cells (Fig. 3C, top, lanes 4 and 6, respectively). The failure of the p100 S866A/ S870A mutant to recruit IKK␣ was not caused by its inability to bind to Tax, because, like wild-type p100 and the p100 S872A mutant, p100 S866A/S870A could efficiently bind to Tax (second panel from top, compare lane 4 with lanes 2 and 6). In fact, mutation of all three of these serines did not affect p100/ Tax interaction (second panel, lane 8). Consistent with the significance of serines 866 and 870 in IKK␣ recruitment, the three-serine mutant of p100 (p100 S866A/S870A/S872A) also FIG. 1. Tax-mediated recruitment of IKK␣ into p100 complexes is independent of the kinase activity of IKK␣. 293 cells were transfected with expression vectors encoding p100 (0.4 g) and HAtagged IKK␣ (0.25 g) or its mutants, either in the absence (Ϫ) or presence (ϩ) of Tax (0.25 g). The cell lysates were subjected to immunoprecipitation using anti-p100, and the co-precipitated IKK␣ (top) and Tax (second panel from top) were detected by immunoblotting using anti-HA-HRP and anti-Tax, respectively. Aliquots of the cell lysates were subjected directly to immunoblotting to monitor the expression levels of HA-tagged IKK␣ and IKK␣ mutants (third panel) as well as the p100 and its processing product, p52 (bottom). WT and SS/EE are wild-type and constitutively active forms of IKK␣, respectively, whereas SS/AA and KA are two inactive forms. The ratio of p52 to p100 is also indicated in the figure.
failed to interact with IKK␣ when Tax was coexpressed (top, lane 8). Parallel immunoblotting assays revealed that the expression levels of all the proteins were comparable (third and bottoms). It is noteworthy that the p100 mutants 1-753, 1-665, and 1-454 are actually constitutive processing forms (Fig. 3B, bottom, lanes 5-7), because these mutants also lose processing inhibitory domains (20,50). The mechanism of constitutive processing of p100 is different from that of inducible processing and seems to be independent of both IKK␣ and ␤-TrCP (47, 50).
Nevertheless, these data clearly indicated that serines 866 and 870 are essential for Tax-mediated recruitment of IKK␣ to p100 and subsequent processing of p100.
Serine 872 of p100, an IKK␣ Phosphorylation Site, Contributes to Tax-induced Processing of p100 -Although it is not involved in binding to Tax or in Tax-mediated recruitment of IKK␣ to p100, serine 872 plays an important role in Taxinduced p100 processing, because compared with wild-type p100, p100 S872A showed a marked defect (about 50%) in FIG. 2. The N terminus of p100 contains sequences required for Tax-induced p100 processing but dispensable for Tax-mediated p100/IKK␣ interaction. A, schematic representation of p100 and its truncation mutants. The ␣-helices, the glycine-rich region (GRR), the processing site, the ankyrin repeat domain (ARD), and the C-terminal processing-required sequences (C-PRS) are indicated. p52 is equivalent to p100 (1-405). B, the N-terminal sequences are not required for Tax-mediated recruitment of IKK␣ into p100 but are essential for Tax-induced p100 processing. HA-tagged IKK␣ and myc-tagged p100 or its N-terminal deletion mutants were transfected into 293 cells either in the absence (Ϫ) or presence (ϩ) of Tax, followed by IP using anti-myc antibody. The co-precipitated IKK␣ (top) and Tax (second panel from top) were detected by immunoblotting using anti-HA-HRP and anti-Tax, respectively. The expression levels of HA-tagged IKK␣ (third panel) and Tax (fourth panel), as well as the p100 constructs and their processing products (bottom), were monitored by direct IB using cell lysates. Note, myc-tagged p52 (processing products of myc-p100) co-migrated with a background band labeled with an asterisk in the bottom panel. C, the first 124 N-terminal amino acids of p100 are essential for Tax-induced p100 processing, whereas further deletion leads to inducible degradation of p100. Five times more protein extracts than used in B were loaded for SDS-PAGE, followed by IB using higher concentration of anti-myc to detect low processing of p100 N-terminal deletion mutants. Again, a background band existed in all the lanes, indicated as NS. D, p52 and p52⌬1-183 show similar stabilities. 293 cells were transfected with p52 or its N-terminal deletion mutant p52⌬1-183. The cells were pulse-labeled for 45 min followed by chase for the indicated times. p52 and its mutant p52⌬1-183 are indicated. E, p100⌬1-183 mutant behaves like an IB, exhibiting degradation induced by IKK␣, IKK␤, or Tax. The p100⌬1-183 mutant was transfected into 293 cells with an empty vector or the indicated cDNA, followed by IB assays using the indicated antibodies.
Tax-induced processing (Fig. 3C, bottom, compare lane 6 with  lane 2), a result observed in multiple experiments ( Fig. 4B and data not shown). Thus, the three serines within the p100 C terminus are critical for Tax-induced p100 processing, although they have different roles in this event.
Serine 872 of p100 is involved in Tax-induced p100 processing, but not Tax-mediated recruitment of IKK␣ to p100, which suggests that this serine might be involved in other events mediated by Tax. One possibility is that Tax-targeted IKK␣ may phosphorylate serine 872. In this regard, our previous studies have indicated that serines 866, 870, and 872 of p100 are candidate sites for IKK␣ phosphorylation (22). More recently, we have found that purified IKK␣ protein phosphorylates serine 872, but not serines 866 and 870, of p100 (45), although it remains unknown whether Tax-activated endoge-nous IKK␣ in HTLV-I-transformed T cells also induces this phosphorylation. To further confirm that, in vitro kinase assays were performed using glutathione S-transferase (GST)/ p100 C-terminal fusion proteins (GST-p100 (860 -900)) harboring mutations of serines 866 and 870 or serine 872, converting them to alanine residues. As expected, GST-p100 (860 -900 S866A/S870A) mutant could be phosphorylated equivalently to wild-type GST-p100 (860 -900) by IKK␣ from HTLV-I-transformed T cells (Fig. 3D, top, compare lane 3 with lane 1), whereas mutation of serine 872 completely blocked the Cterminal phosphorylation of p100 by IKK␣ (lane 4). Thus, serine 872, but not serine 866 or 870, is an actual target of Tax-activated IKK␣, even though serines 866 and 870 are required for the recruitment of IKK␣ into p100 complexes and subsequent processing of p100 induced by Tax. FIG. 3. Serines 866 and 870 of p100 are essential for the recruitment of IKK␣ into p100 complexes by Tax, whereas serine 872, the C-terminal IKK␣ phosphorylation site of p100, also plays a role in Tax-induced processing of p100. A, schematic representation of p100 showing the C-terminal processing-required sequences (C-PRS), which include both IKK␣ recruitment and phosphorylation sites. The N-terminal processing-required sequences (N-PRS), the ␣-helices, the glycine-rich region (GRR), the processing site, and the ankyrin repeat domains (ARD) are also indicated. B, the C-terminal deletion mutants show defects in binding to IKK␣ induced by Tax. 293 cells were transfected with HA-tagged IKK␣ and p100 or its C-terminal deletion mutants either in the absence (Ϫ) or presence (ϩ) of Tax. Cell lysates were subjected to IP using anti-p100N for examining the p100/IKK␣ and p100/Tax interactions or to IB using the indicated antibodies for detection of the expression levels of IKK␣, p100 and its mutants as well as their processing products p52 (same as p100 (1-405)). Note that p100 (1-859) loses processing inducible domain, whereas other p100 c-terminal truncation mutants (1-753, 1-665, and 1-454) undergo constitutive processing caused by loss of the processing inhibitory domain (see text and Refs. 20, 47, and 50 for details). These p100 C-terminal deletion mutants have been described in Fig.  2A. C, serines 866 and 870 but not serine 872 of p100 are essential for Tax-induced interaction of p100/IKK␣. 293 cells were transfected with indicated expression vectors, followed by immunoprecipitation and immunoblotting assays as described in A. The ratio of p52 to p100 is indicated in the figure. S/A, SS/AA, and SSS/AAA represent the p100 S872A, p100 S866A/S870A, and p100 S866A/S870A/S872A point mutants, respectively. D, serine 872, but not serine 866 or serine 870, of p100 is an IKK␣ phosphorylation site. Tax-activated IKK␣ was isolated from the HTLV-I-  figure, respectively). The phosphorylated GST-p100 (860 -900) and its serine-to-alanine substitution mutants are indicated as P-p100C (top), and the levels of the wild-type and mutant forms of p100 substrates as well as IKK␣ were also monitored by IB analysis of the kinase assay membrane using anti-GST (middle) and anti-IKK␣ (bottom).

IKK␣ Phosphorylation Sites in the N Terminus of p100
Are Important for Tax-induced p100 Processing-The finding that the C-terminal IKK␣ phosphorylation serine 872 contributes to but is not sufficient for Tax-induced p100 processing (Fig. 3) suggested that there may be additional IKK␣ targets within p100 that are also involved in p100 processing. Because the sequences between amino acid residues 71 and 124 play a crucial role in Tax-induced p100 processing (Fig. 2B), we hypothesized that serines (serines 99, 108, 115, and 123) located in this region might also be targets of IKK␣. This hypothesis was further substantiated by our recent discovery showing that purified IKK␣ can also phosphorylate these serines in vitro (45). At first, the functional significance of these serines in Tax-induced p100 processing was investigated by using p100 mutants harboring individual or combinational serine substitutions in in vivo inducible processing assays. As shown in Fig.  4B, all the single-serine substitution mutants, such as the p100 S872A mutant, showed appreciable defects in Tax-induced processing, although to different extents. These defects were much clearer when all the four serines (p100 4S/A) were simultaneously substituted (lane 14), although mutation of these four serines together with serine 872 (p100 5S/A) completely blocked Tax-induced p100 processing (lane 16). Thus, both Nand C-terminal serines within p100 play important roles in Tax-induced processing of p100.
To further confirm that these N-terminal serines are also targets of Tax-activated IKK␣ from HTLV-I-transformed T cells, GST fusion proteins containing the N-terminal 132 amino acids of p100 (GST-p100 (1-132)) were used for in vitro kinase assays. Tax-activated IKK␣ could efficiently phosphorylate the p100 N terminus (Fig. 4C, lane 1). It is noteworthy that mutation of all four N-terminal serines blocked the phosphorylation by IKK␣ (lane 6), although single substitutions of individual serines only partially affected IKK␣-mediated phosphorylation (lanes 2, 3, 4, and 5). Thus, the serines 99, 108, 115, and 123 of p100, together with serine 872, are Tax-activated IKK␣ targets for phosphorylation.
IKK␣ Phosphorylation Sites within p100 Are Involved in the Tax-induced Recruitment of ␤-TrCP to p100 and Subsequent p100 Ubiquitination-Because serines 866 and 870 of p100 are required for Tax-induced IKK␣ recruitment to p100, it was interesting to test whether other IKK␣ phosphorylation serines within p100 are also involved in Tax-mediated p100/IKK␣ association, even though serine 872 is not involved (Fig. 3C). Like the p100 S872A mutant, the IKK␣ phosphorylation-deficient mutant of p100 (p100 5S/A), which harbors substitutions of the five I⌲⌲␣ phosphorylation serines by alanines, remained competent in interacting with IKK␣ in Tax-expressing cells (Fig.  4D, top, lane 4). These findings suggested that the recruitment of IKK␣ to p100 by Tax and IKK␣-mediated phosphorylation FIG. 4. N-terminal IKK␣-targeted serines 99, 108, 115, and 123 of p100 play important roles in Tax-induced p100 processing but are not involved in Tax-mediated recruitment of IKK␣. A, schematic representation of p100 showing both the N-and C-terminal IKK␣ phosphorylation sites. The processing site is also indicated. B, Tax-induced processing of p100 involves the N-terminal serines 99, 108, 115, and 123 of p100. 293 cells were transfected with the indicated p100 constructs either in the absence (Ϫ) or presence (ϩ) of Tax, followed by IB analysis of p100 and p52 (p100 processing) (top), and Tax (bottom). The new p100 mutants were designated based on the position of the mutated serine. For example, S99A harbors a substitution of serine 99 by alanine. 4S/A harbors substitutions of serines 99, 108, 115, and 123, whereas 5S/A harbors substitutions of serines 99, 108, 115, 123, and 872. SS/AA is the p100 S866A/A870A mutant. The ratio of p52 to p100 is indicated in the figure. C, serines 99, 108, 115, and 123 are targets of Tax-activated IKK␣. GST-p100 (1-132) wild type and its mutants harboring serine substitution(s), as well as immunopurified IKK␣ from HTLV-I-transformed T-cells, were used for in vitro kinase assays. Phosphorylated GST fusion mutants are shown at the top; levels of the GST fusion proteins are shown at the bottom. The GST fusion proteins were named using the same strategy as described in B. D, IKK␣ phosphorylation sites are dispensable for Tax-induced p100 binding to IKK␣. The indicated p100 constructs and HA-tagged IKK␣ were transfected alone (Ϫ) or together with Tax (ϩ), followed by Co-IP assays using anti-p100N. The co-precipitated IKK␣ (top) and Tax (second panel from top) as well as the expression levels of IKK␣ (third panel), p100, and p52 (bottom) are shown in the figure.
work sequentially and in concert to induce processing of p100.
The phosphorylation of these important serines would be hypothesized to recruit a specific E3 ligase to trigger p100 ubiquitination and subsequent processing. In this regard, we have previously demonstrated that Tax-induced p100 processing is associated with the polyubiquitination of p100 (22). Thus, we tested the possible physical interaction between the E3 ligase, ␤-TrCP, and p100 in Tax-expressing cells by Co-IP assays. As shown in Fig. 5A, in the absence of Tax, ␤-TrCP only weakly bound to p100 (top, lane 1). However, a significant interaction between ␤-TrCP and p100 was demonstrated when Tax was co-expressed (lane 2). To examine the role of IKK␣ phosphorylation sites within p100 in ␤-TrCP/p100 interactions induced by Tax, we also included the IKK␣ phosphorylationdeficient mutant of p100, p100 5S/A, in the Co-IP assays. Consistent with its resistance to Tax-induced processing, this p100 mutant failed to interact with ␤-TrCP either in the absence or presence of Tax (top, lanes 3 and 4). To further confirm the functional significance of these serines in the recruitment of ␤-TrCP, in vivo ubiquitin conjugation assays were also performed using this p100 mutant. As expected, the p100 5S/A mutant was indeed unable to become polyubiquitinated in Taxexpressing cells, although Tax could efficiently induce ubiquitination of wild-type p100, as evidenced by the formation of an ubiquitin-conjugated heterogeneous ladder (Fig. 5B, top, lanes  4 and 2, respectively). In agreement with its inability to bind to IKK␣ and be phosphorylated, p100 S866A/S870A also failed to bind to ␤-TrCP and to become ubiquitinated in Tax-expressing cells (Fig. 5, A and B, top, lane 6). These results demonstrated the significance of the IKK␣-targeted serines of p100 in the recruitment of the E3 ligase, ␤-TrCP, into p100 as part of Tax-induced p100 processing. These results also suggested that FIG. 5. Phosphorylation of p100 by IKK␣ is required for Tax-mediated ␤-TrCP recruitment into p100 complexes and subsequent ubiquitination and processing of p100. A, recruitment of IKK␣ into p100 complexes and IKK␣-induced phosphorylation of p100 are required for Tax-induced binding of p100 to ␤-TrCP. 293 cells were transfected with expression vectors encoding HA-tagged ␤-TrCP and p100 wild type, or its IKK␣ recruitment-defective (SS/AA) or phosphorylation-deficient (5S/A) mutants, together with either an empty vector (Ϫ) or Tax (ϩ). Cell lysates were subjected to IP using anti-p100, and the co-precipitated ␤-TrCP (top) and Tax (second panel from top) were detected by IB using anti-HA-HRP and anti-Tax, respectively. The cell lysates were also directly subjected to IB to detect the expression levels of ␤-TrCP (third panel) as well as p100 and its processing product p52 (bottom). B, IKK␣-induced phosphorylation of p100 is also essential for Tax-induced ubiquitination of p100. 293 cells were transfected with HA-tagged ubiquitin together with indicated constructs. The p100 wild type and its mutants were isolated by IP with anti-p100 followed by IB using anti-HA-HRP to detect the ubiquitin-conjugated p100 (top). The expression level of ubiquitin was monitored by direct IB using anti-HA-HRP to detect the total ubiquitin-conjugated proteins (bottom). The expression levels of Tax and p100 were similar to that indicted in A. C, ␤-TrCP is involved in Tax-induced p100 processing. 293 cells were transfected with siRNA either for control green fluorescent protein (G) or for ␤-TrCP (␤). 24 h after transfection, a second transfection was performed with the indicated siRNAs and cDNA expression constructs. The processing of p100 was analyzed by IB using anti-p100N 24 -36 h after the second transfection. p100 and its processing product p52 as well as their ratio are indicated. Cell lysates were also used for detecting the expression levels of Tax, NIK, and ␤-TrCP by IB. D, ␤-TrCP is required for Tax-induced ubiquitination of p100. 293 cells were transfected with siRNA as described in C, except that HA-tagged ubiquitin was also included in the second transfection. Cell lysates were subjected to Co-IP to detect ubiquitin-conjugated p100 as described in B. The expression levels of ubiquitin, Tax, NIK, p100, and ␤-TrCP were similar to that shown in C.
the potential role of ␤-TrCP in Tax-induced p100 processing is to mediate ubiquitination of p100.
␤-TrCP Contributes to Tax-induced Polyubiquitination and Processing of p100 -As we have shown above and before, Taxinduced p100 processing is associated with the polyubiquitination of p100 (22; see also Fig. 5B). To confirm the significance of ␤-TrCP in Tax-mediated p100 ubiquitination and processing, the siRNA-mediated gene suppression technique was used. As shown by IB, transfection of ␤-TrCP siRNA (Fig. 5, C and D, even-numbered lanes) efficiently depleted the expression of ␤-TrCP but not that of other proteins, such as p100, Tax, and NIK, whereas the control green fluorescent protein siRNA (odd-numbered lanes) did not affect the expression of these proteins (Fig. 5C, third panel from top, compare lane 11 with lane 12 for suppression of ␤-TrCP; top, lanes 1-12 for p100; bottom, lanes 3-6 for Tax; and second panel, lanes 8 -10 for NIK). A ␤-TrCP mutant (␤-TrCP RF ) harboring sense mutations in the siRNA targeting site but not affecting the protein sequence showed resistance to ␤-TrCP siRNA (third panel, lanes [5][6][7][8]. It is noteworthy that Tax-induced processing of p100 was appreciably, although not completely, abrogated in ␤-TrCPdeficient cells (top, lane 4), whereas transfection of ␤-TrCP RF could efficiently restore this ability of Tax (lane 6). Suppression of ␤-TrCP also dramatically and consistently blocked Tax-induced p100 ubiquitination (Fig. 5D, lane 4), and this defect could also be reverted by transfection of ␤-TrCP RF (lane 6). In agreement with a previous study (47), suppression of ␤-TrCP prevented NIK-mediated ubiquitination and processing of p100 (Fig. 5, C and D, lane 10). It is noteworthy that Tax, but not NIK, could still induce moderate processing of p100 in ␤-TrCPdeficient cells (Fig. 5C, top, compare lane 4 with lane 10), suggesting that a ␤-TrCP-independent mechanism also contributes to Tax-induced processing of p100. Nevertheless, these results indicated that ␤-TrCP is indeed an E3 ligase important for both ubiquitination and processing of p100 induced by Tax.

DISCUSSION
In this study, we have explored the detailed mechanisms by which the HTLV-I-encoded oncoprotein, Tax, deregulates the tightly controlled processing of p100 for its oncogenic action. We demonstrated that both N-and C-terminal serines of p100 are involved in Tax-induced p100 processing. Consistent with our previous studies showing their functional significance in inducible p100 processing (20,22), the data here further indicate that serines 866 and 870 of p100 are actually required for Tax-mediated IKK␣ recruitment into p100 complexes. Besides the C-terminal serine 872, Tax-activated IKK␣ also phosphorylates four N-terminal serines (residues 99, 108, 115, and 123) of p100, which play a critical role in Tax-induced p100 processing. We also showed that ␤-TrCP, which is required for NIKmediated ubiquitination and processing of p100, is just partially involved in Tax-mediated p100 processing. We have demonstrated recently that Tax-, but not NIK-, mediated p100 processing requires IKK␥ (22). The data presented here thus provide a second body of evidence for the involvement of different mechanisms in cellular and viral pathways of p100 processing. These findings provide the molecular basis for a possible approach to HTLV-I diseases by specifically blocking Taxderegulated p100 processing.
Unlike p105 processing, which is largely constitutive, p100 processing is highly selective and tightly controlled (7,19,44). Therefore, in most cell types, including T cells, only very small amounts of p52 are produced relative to the levels of its precursor p100 (20,22). However, in HTLV-I-transformed T cells, p100 is actively processed, resulting in p52 overproduction (22,39). Because overproduction of p52 is associated with lymphoid hyperplasia and transformation (42,44), deciphering the mech-anisms of Tax-deregulated processing of p100 will help understand the development of HTLV-I-mediated adult T cell leukemia and other NF-B-associated diseases. Like IB degradation in the canonical NF-B pathway, Tax also usurps the IKK complex to induce processing of p100 (22). However, it seems that whereas both IKK␣ and IKK␤ are involved in Tax activation of classic IB degradation (51,52), Tax specifically targets IKK␣ into p100 complexes to initiate processing of p100 (22). IKK␥, an essential adaptor for binding of Tax to IKK␣ and/or IKK␤ (8 -12), is required for both Tax-activated canonical and non-canonical NF-B pathways.
Our current study shows that Tax recruitment of IKK␣ to p100 requires serines 866 and 870 of p100, two serines essential for inducible processing of p100 (Fig. 3). It seems that recruitment of IKK␣ to p100 and phosphorylation of p100 by IKK␣ are two sequential steps in Tax-induced p100 processing, because p100 S866A/S870A defective in IKK␣ binding cannot be phosphorylated (Fig. 3D; Ref. 22), whereas a p100 phosphorylation-deficient mutant is still capable of binding to IKK␣ in Tax-expressing cells (Fig. 4). The function of p100 phosphorylation would then act to recruit the ␤-TrCP SCF complex, which mediates polyubiquitination and subsequent processing of p100 (Fig. 5). Our genetic and biochemical studies also suggest another novel function of p100 phosphorylation in inducible processing of p100. This new function seems to be ␤-TrCPindependent, because the recruitment of IKK␣ to p100 and subsequent phosphorylation of p100 are completely required for Tax-induced p100 processing, whereas in ␤-TrCP-deficient cells, Tax-induced processing of p100 is only partially inhibited (Fig. 5C). It seems that the ␤-TrCP-independent mechanism of inducible processing of p100 is Tax-specific. Unlike Tax-deregulated processing of p100, p100 processing induced by NIK, the physiological stimulator of p100 processing, is completely blocked in ␤-TrCP-deficient cells ( Fig. 5; Ref. 47). Although the mechanism of the ␤-TrCP-independent pathway for p100 processing is currently unclear, a previous study indicated that the processing of the constitutive forms of p100 is also through a ␤-TrCP-independent mechanism (47). It is noteworthy that the constitutive processing of p100 is much slower than that of inducible p100 processing mediated by NIK, whereas the dynamics of Tax-mediated p100 processing is intermediate (data not shown; Refs. 20,22), further suggesting that both ␤-TrCPdependent (rapid processing) and -independent (slow processing) mechanisms are involved in Tax-mediated processing of p100.
Although different mechanisms may exist for HTLV-I-mediated induction of the non-canonical NF-B pathway, the current study also provides some insights into the tightly controlled, physiological processing of p100. Our data demonstrate that serines 866 and 870 of p100, which have previously been shown to be essential for both NIK-and Tax-induced p100 processing (20,22), are required for the binding of p100 to IKK␣, a key step for inducible p100 processing. These results further confirm that physical recruitment of IKK␣ to p100 and subsequent phosphorylation of p100 by IKK␣ are the general mechanisms of inducible processing of p100, which is also consistent with our recent discoveries indicating that IKK␣ is required for inducible processing of p100 under both physiological and pathogenic conditions (21,22). Indeed, both the N-and C-terminal IKK␣ phosphorylation sites of p100, similar to IKK␣ docking site in the C terminus, play an essential role in p100 processing induced by NIK or Tax. Mutation of individual IKK␣ phosphorylation serines causes a partial block in Tax or NIK-induced p100 processing, whereas substitution of all of them completely prevents the processing of p100 induced by Tax or NIK (Figs. 3 and 4; Ref. 45).
There is an intriguing question of why the multiple IKK␣targeted serines in p100, unlike the IKK targeted serines within other proteins, such as IB␣, are located at distant positions in the amino acid sequence of the protein. Because the Rel homology domain of NF-B proteins can strongly associate with the ankyrin repeat domain of IBs, it seems plausible that p100 N-terminal Rel homology domain interacts with its Cterminal ankyrin repeat domain, thereby bringing the C-terminal serines close to the N-terminal serines to form a functional domain in three dimensions. One other important function of the interaction between the N and C termini of p100 seems to be to hide the access of p100 to IKK␣, because IKK␣ fails to directly associate with p100 (Fig. 1). In support of this hypothesis, the deletion of the N-terminal 183 amino acids or more results in constitutive IKK binding of p100 (Fig. 2). It is interesting that in Tax-expressing cells, IKK␣ can stably associate with p100 (Figs. [1][2][3][4]Ref. 22). It is obvious that the "adaptor" function of Tax contributes to the recruitment of IKK␣ to p100. Most probably, the interaction between Tax and p100 also changes this three-dimensional domain to facilitate exposure of the C-terminal "docking" site of p100 (serines 866 and 870). It is noteworthy that two N-terminal ␣-helices of p100 required for Tax binding to p100 are located near the N-terminal phosphorylation sites of IKK␣, suggesting these two helices may be part of this functional three-dimensional domain. Further supporting this idea, our data also indicate that loss of the p100 N-terminal region, including the two helices and N-terminal phosphorylation serines, leads to inducible degradation instead of processing of p100 (Fig. 2). This hypothesis can further explain why signal-dependent phospho-rylation of p105, a protein very homologous to p100, mainly triggers degradation but not processing, because positions of the two ␣-helices are the biggest structural difference between these two proteins at their N termini (53). This idea clearly needs to be further substantiated by more studies, particularly structural studies of full-length p100 and p105.
Based on our findings, a model of Tax-induced processing of p100 is presented in Fig. 6. In brief, in HTLV-I-infected and Tax-expressing cells, Tax directly interacts with p100 through its two N-terminal ␣-helices, probably changing the conformation of p100 to expose the two C-terminal "docking" serines 866 and 870. At the same time, Tax activates IKK␣ indirectly via interacting with IKK␥. Then Tax-activated IKK␣ binds to the exposed serines 866 and 870 of p100 with the additional help (an adaptor function) of Tax. Once stably associated with p100, IKK␣ phosphorylates both N-and C-terminal serines within p100, which triggers the interaction between p100 and ␤-TrCP and subsequent ␤-TrCP-mediated p100 ubiquitination and processing of p100. A ␤-TrCP-independent pathway also contributes to Tax-deregulated processing of p100, although the detailed mechanism remains to be elucidated. FIG. 6. A model depicting a ␤-TrCP and ubiquitination-dependent mechanism of Tax-deregulated p100 processing, contributing to its oncogenic action. A ␤-TrCP-independent mechanism may also play a role in Tax-mediated p100 processing (see text).