In Vitro Activation of the IκB Kinase Complex by Human T-cell Leukemia Virus Type-1 Tax*

Human T-cell leukemia virus type-I expresses Tax, a 40-kDa oncoprotein that activates IκB kinase (IKK), resulting in constitutive activation of NFκB. Herein, we have developed an in vitro signaling assay to analyze IKK complex activation by recombinant Tax. Using this assay in combination with reporter assays, we demonstrate that Tax-mediated activation of IKK is independent of phosphatases. We show that sustained activation of the Tax-mediated activation of the NFκB pathway is dependent on an intact Hsp90-IKK complex. By acetylating and thereby preventing activation of the IKK complex by the Yersinia effector YopJ, we demonstrate that Tax-mediated activation of the IKK complex requires a phosphorylation step. Our characterization of an in vitro signaling assay system for the mechanism of Tax-mediated activation of the IKK complex with a variety of mutants and inhibitors results in a working model for the biochemical mechanism of Tax-induced activation.

Human T-cell leukemia virus type-I expresses Tax, a 40-kDa oncoprotein that activates IB kinase (IKK), resulting in constitutive activation of NFB. Herein, we have developed an in vitro signaling assay to analyze IKK complex activation by recombinant Tax. Using this assay in combination with reporter assays, we demonstrate that Tax-mediated activation of IKK is independent of phosphatases. We show that sustained activation of the Tax-mediated activation of the NFB pathway is dependent on an intact Hsp90-IKK complex. By acetylating and thereby preventing activation of the IKK complex by the Yersinia effector YopJ, we demonstrate that Tax-mediated activation of the IKK complex requires a phosphorylation step. Our characterization of an in vitro signaling assay system for the mechanism of Tax-mediated activation of the IKK complex with a variety of mutants and inhibitors results in a working model for the biochemical mechanism of Tax-induced activation.
The first identified pathogenic retrovirus, human T-cell leukemia virus type-1, the causal agent for adult T-cell leukemia, expresses Tax, a 40-kDa phospho-oncoprotein, that plays a pivotal role in the growth and transformation of T-cells (1). Tax chronically stimulates the IB kinase (IKK) 3 complex via NEMO/IKK␥, resulting in sustained phosphorylation and degradation of IB and activation of NFB (2)(3)(4). Tax, unlike all the other upstream stimuli such as NFB-inducing kinase, MEK kinase 1, and transforming growth factor ␤-activated kinase 1, is not a kinase that phosphorylates and activates the IKK complex. It activates the complex by direct interaction via a mechanism not clearly understood. It has been proposed that Tax activates the NFB pathway by inducing a conformational change onto IKK (5).
Previous studies have identified several binding partners for Tax, such as MEK kinase 1 (6) and IKK␥ in the IKK complex (3). The catalytic subunit of the serine/threonine protein phosphatase 2A (PP2A) was recently identified to bind Tax directly, and it was demonstrated that Tax forms a ternary complex together with IKK␥ and PP2A (7). The role of PP2A in inhibiting IKK activation has been established in the past (8). Based on these studies, two models have been proposed to explain the regulation of the IKK complex by Tax in conjunction with PP2A (7,9). According to one model, Tax binding to PP2A relieves negative inhibition on the IKK complex, resulting in an active IKK complex (7). In contrast, another group has demonstrated that PP2A positively regulates IB kinase signaling, i.e. IKK-PP2A complexes are essential for IKK activation (9). This gives rise to a second model wherein Tax-mediated activation of IKK requires the association of active PP2A with the IKK complex (9).
In contrast to the viral pathogen human T-cell leukemia virus type-1, the bacterial pathogen Yersinia pestis, the causal agent for bubonic plague, inhibits the NFB pathway (10,11). Yersinia species use the bacterial virulence factor YopJ to block all MAPK and NFB pathways at a common point (11,12). Recent studies revealed that YopJ functions as an acetyltransferase (13,14). It blocks the activation of all MAPK kinases, including IKK␤, by the addition of an acetyl group to the highly conserved serine and threonine residues in the activation loop of the kinase. Acetylation of these residues by YopJ prevents the activation of these kinases by inhibiting phosphorylation. Previously, overexpression studies have shown that YopJ inhibits Tax-mediated activation of the IKK complex (15).
To decipher the requirements of Tax-induced NFB signaling, we have developed an in vitro signaling assay to analyze the activation of the IKK complex by Tax. The assay utilizes wild type and mutant recombinant Tax proteins, an S100 lysate containing unstimulated IKK complex, and a readout for the activation of the NFB pathway, phospho-IB. The mutant Tax proteins that we have used in our assays include M22, H41Q, H43Q, and K85N. Fu et al. (7) demonstrated that the ability to induce NFB activation is abrogated in the M22 mutant whereas the other three mutants are defective in binding PP2A and also fail to activate the NFB pathway. Based on their study, they proposed that binding of Tax to PP2A is essential for NFB activation. Herein, we find that activation of the NFB pathway by recombinant Tax does not require binding to or the activity of PP2A but is dependent on an intact Hsp90-IKK complex. Recombinant Tax is unable to activate the IKK complex in an S100 lysate that contains the acetyltransferase YopJ. Based on results from transcription reporter assays and our in vitro signaling system, we propose that the activation of the IKK complex by human T-cell leukemia virus type-1 Tax requires both a signaling complex containing functional chaperone Hsp90 and an activation loop on IKK␤ that is required for phosphorylation. This in vitro signaling assay will thus allow us to dissect the mechanism of Tax-dependent activation of IKK.

MATERIALS AND METHODS
Plasmids and Reagents-Tax was cloned into pET28a and pcDNA3 vectors using 5Ј-BamH1 and 3Ј-Xho1. The four Tax mutant constructs, M22, H41Q, H43Q, and K85N, in pET28a and pcDNA3 vectors were created using the Stratagene mutagenesis kit with the following pair of oligonucleotides: pSFFV YopJ-FLAG, pSFFV YopJC172A-FLAG, and 5xNFB luciferase reporter were described previously (8).
Anti-phospho-IB antibody was obtained from Cell Signaling, and anti-FLAG antibody was obtained from Sigma. Antibodies to Hsp90 and IKK␤ were purchased from Santa Cruz Biotechnology. Protein A-agarose beads were purchased from Invitrogen. Geldanamycin and okadaic acid were purchased from Alexis. pNPP was obtained from Sigma.
Expression and Purification of Recombinant Proteins-Wild type and the various mutants of Tax (M22, H41Q, H43Q, and K85N) were expressed as His-tagged proteins in Rosetta (Invitrogen) cells, grown to an O.D. of 0.6 -0.8, and induced at room temperature with 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside for 8 -12 h. The bacterial pellet was then lysed by an Emulsiflex C5 cell disruptor (Avestin) and purified using standard protocols for Ni 2ϩ -nitrilotriacetic acid purification (Qiagen). Recombinant TRAF6 was used as a positive control and was purified from Sf9 cells as described in Li et al. (16). Anti-Tax was used as previously described (17).
Tissue Culture and Lysate Preparation-HEK293 or HeLa cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% cosmic calf serum and 100 units/ml penicillin/streptomycin/glutamine (Invitrogen) in the presence of 5% CO 2 . For the in vitro assays, untransfected HeLa cells or HEK293 cells transfected with 10 g of pSFFV, pSFFV-YopJ-FLAG, or pSFFV-YopJC172A-FLAG using FuGENE (Roche Applied Science) were grown up to 100% confluency, harvested using phosphate-buffered saline-EDTA, and lysed with equal volume of HTX lysis buffer (10 mM Hepes, pH 7.4, 10 mM MgCl 2 , 1 mM MnCl 2 , 0.5% Triton X-100, 0.1 mM EGTA) containing protease inhibitor mixture tablet (Roche Applied Science) by incubating on ice for 20 -60 min. Lysate was spun sequentially at 800, 10,000, and then 100,000 ϫ g to obtain a cleared S100 lysate that was then stored at Ϫ80°C at a protein concentration of 10 mg/ml.
Luciferase and Transfection Assays-HEK293 cells were transfected with 200 ng of WT Tax or the different Tax mutants (M22, H41Q, H43Q, and K85N) and 5xNFB luciferase reporter in the absence or presence of pSFFV YopJ-FLAG or pSFFV YopJC172A-FLAG for 24 h using FuGENE transfection reagent. Each plate was also transfected with pRSV-Renilla to serve as the internal standard control. Lysates were prepared using passive lysis buffer (Promega), and luciferase assays were performed using Fluostar Optima (BMG Labtech).
pNPP Assay-A 1:50 dilution of Hela-S3 cell-free lysate was incubated in the absence or presence of 10 nM okadaic acid. The 96-well microtiter plate assay (Upstate Biotechnology) was carried out in triplicate according to the manufacturer's description. The absorbance at wavelength 410 nm was read using a FluoStar Optima (BMG Labtech).
In Vitro Assay Using Recombinant Tax-Cleared lysates (10 mg/ml) were incubated with or without 2 M recombinant WT Tax and/or all the mutants or 1 M TRAF6 in the presence of an ATP-regenerating system (10ϫ stock: 10 mM ATP, 350 mM creatine phosphate, 20 mM Hepes, pH 7.2, 10 mM MgCl 2 , and 500 g/ml creatine kinase) for 10 min at 37°C. 0.5 M okadaic acid and 2 M geldanamycin were added to cleared lysates before the addition of the recombinant protein. Reactions were terminated by addition of 5ϫ SDS sample buffer. Proteins from the reaction samples were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes. IKK activation was detected by immunoblotting with anti-phospho-IB antibody. For in vitro assays using Hsp90-immunodepleted lysate, lysates were incubated with anti-Hsp90 antibody for 1 h at 4°C, followed by incubation with 30 l of Protein A-agarose beads for 30 min at 4°C. This step was performed twice to ensure complete immunodepletion of Hsp90. The lysates were finally incubated with Tax or TRAF6 recombinant proteins. In vitro kinase assays were performed on immunoprecipitated complexes as previously described using [␥-32 P]ATP and GST-IkB␣-(1-52) as the recombinant substrate (11).

In Vitro Activation of the NFB Pathway by Recombinant
Human T-cell Leukemia Virus Type-1 Tax-We previously established an assay to study the activation of MAPK and NFB Tax Activation of IKK Complex signaling pathways in vitro (13). The assay is initiated by the addition of a partially purified activator of a signaling pathway to an S100 lysate in the presence of an ATP-regenerating system, followed by termination of the reaction by addition of SDS sample buffer. The activation of the signaling pathway is assessed by analysis of a downstream indicator by immunoblotting for a phosphorylated protein. Herein, we analyzed the activation of the NFB pathway by the addition of partially purified recombinant WT His 6 -Tax protein to the unstimulated S100 lysate (Fig. 1A, lane 1). Addition of Tax protein (2 M) to cleared lysate results in the activation of the NFB pathway as indicated by robust phosphorylation of IB (Fig. 2A,  lane 2). Similarly, addition of 1 M partially purified recombinant TRAF6 (Fig. 1A, lane 6), which served as the positive control in this assay, also activates the NFB pathway ( Fig.  2A, lane 3). Using serial dilutions of WT recombinant Tax added to the in vitro signaling system, we demonstrate that the activation of the NFB pathway is detected with as little as 800 nM Tax (Fig. 2B).
To confirm that WT His 6 -Tax protein is activating the IKK complex specifically, we activated lysates with WT His 6 -Tax protein followed by immunoprecipitation of the IKK complex using antibodies for IKK␥ or Tax (Fig. 2C). The isolated signaling complexes were tested for radioactive kinase activity in an in vitro kinase assay using GST-IB␣-(1-52) as the substrate. Identical phosphorylated protein profiles were observed from signaling complexes isolated using either anti-IKK␥ or anti-Tax antibodies, demonstrating that Tax is binding to and activating the IKK complex (Fig. 2C). As expected, only complexes isolated from the Tax-activated lysates were able to phosphorylate GST-IB␣ (Fig. 2, B and C).
Tax-mediated Activation of NFB Pathway Is Independent of Binding to Phosphatases-To determine the role of phosphatases in the activation of the NFB pathway by Tax, several previously described mutants of Tax (7) were used in our in vivo and in vitro signaling assays. The effect on the activation of the NFB pathway by wild type and mutant Tax proteins was first analyzed using transfection-based luciferase assays with an NFB luciferase reporter (Fig. 3A). The activation of the NFB pathway by the various Tax mutants was also analyzed by immunoblotting cell lysates with anti-phospho-IB antibody (Fig. 3B). We observed that the M22 mutant, which has previously been described as the inactive form of Tax, fails to induce activation of the NFB pathway, and these findings are consistent with previous studies in which this mutant was shown to be unable to activate the NFB pathway (Fig. 3, A and B) (7). Tax mutants H41Q, H43Q, and  5) were expressed as His 6 -tagged proteins in BL21-DE3 cells and purified from BL21-DE3 cells using standard protocols. Recombinant TRAF6 (lane 6) was purified from Sf9 cells as previously described (16). Sample of each recombinant protein was resolved on SDS-PAGE and stained with Coomassie Blue (A) or subjected to Western blotting using anti-Tax antibody (B). Thin lines denote where lanes have been deleted from the scanned image.  K85N activated the NFB pathway to varying levels, based on both the luciferase assays and the immunoblotting for phospho-IB (Fig. 3, A and B). For the latter experiment, cells were constitutively expressed with FLAG-IB to detect phosphorylation of exogenous IB, because endogenous IB is degraded upon activation of the NFB pathway (13). The mutants H43Q and K85N were not detected using the anti-Tax antibody because of their decreased expression levels in the mammalian system (Fig. 3B) (17). Because H41Q, H43Q, and K85N, which are defective in binding PP2A, were able to activate the NF〉 pathway, the binding of PP2A to Tax does not appear to be essential for Tax-mediated activation of the IKK complex (7).
The Tax mutants were also expressed as recombinant Histagged proteins in bacteria and purified using nickel affinity chromatography (Fig. 1A). The anti-Tax antibody was able to immunoreact with all the recombinant proteins (Fig. 1B). Consistent with previous in vivo observations, addition of the recombinant inactive mutant form of Tax (M22) to the in vitro signaling system was unable to induce phosphorylation of IB (Fig. 4A, lane 5). However, addition of the three recombinant mutant forms of Tax including H41Q, H43Q, and K85N, which are unable to bind PP2A, resulted in activation of the NFB pathway as indicated by phosphorylation of IB in vitro, albeit with varying potency (Fig. 4A, lanes 3, 4, and 6). The profiles of partially purified wild type Tax and the inactive mutant M22 proteins (Fig. 1, lanes 1 and 2) appear the same, supporting the proposal that differences in the activities can be attributed to   1 and 2), Tax (lanes 3 and 4), or TRAF6 (lanes 5 and 6) in the absence (lanes 1, 3, and 5) or presence (lanes 2, 4, and 6) of 10 nM okadaic acid (OA). Activation of the IKK complex was detected by immunoblotting with anti-phospho-IB antibody. To confirm equal amounts of IB in all lanes, the same lysates were blotted with ␣-IB antibody. C, activity of PP2A in S100 lysate in the presence or absence of 10 nM OA. PP2A activity was measured using pNPP as a substrate. The data shown are representative of three independent experiments. Error bars denote mean Ϯ S.D. of triplicates; results shown are representative of four independent experiments. the mutated amino acids in the various Tax proteins. Overall, these observations do not support the first model, in that binding of PP2A to Tax is not essential for Tax activation of the IKK complex (7).
Tax-mediated Activation of the NFB Pathway Is Independent of Active Phosphatases-To further assess whether PP2A plays a positive role in the in vitro activation of the IKK complex by Tax, the in vitro activation of the NFB pathway was analyzed in the presence of okadaic acid (OA). Addition of 10 nM OA (or 500 nM, data not shown) did not alter the ability of wild type Tax to activate the IKK complex as observed by phosphorylation of IB (Fig. 4B, lane 4). At this very low concentration of OA (10 nM), PP2A activity is effectively repressed in the lysate (Fig. 4C). Therefore, when PP2A activity is inhibited, Tax is still able to activate the IKK complex. As expected, there was no change in IB phosphorylation levels upon addition of TRAF6 to the lysate in the presence of OA (Fig. 4B, lane 6). Based on these observations, Tax can activate the IKK complex independent of PP2A activity, thereby discounting the second model that predicted that the interaction between IKK and PP2A is essential for Tax to be able to activate the IKK complex (9).
Hsp90 Is Essential for Tax-mediated Activation of the IKK Complex-Studies were initiated to further understand the requirements for activation of the IKK complex by Tax. The in vitro signaling system recapitulated the activation of the NFB pathway by Tax and also demonstrated that Tax-mediated IKK activation is independent of phosphatases. It is possible that by binding to IKK␥, Tax is inducing a conformational change that results in the activation of the IKK complex (5). Hsp90, which is known to maintain the structural integrity of protein complexes (18), is an integral component of the IKK complex (19). To test whether Hsp90 plays a role in Tax-mediated activation of the NFB pathway in vitro, geldanamycin (GA), an Hsp90specific inhibitor, was used in the in vitro signaling system. Upon addition of 2 M GA to the in vitro signaling assay, it was observed that Tax could no longer activate IKK as indicated by the lack of IB phosphorylation (Fig. 5A, lane 4). GA also inhibited the TRAF6-dependent activation of the IKK complex (Fig.  5A, lane 7). This is consistent with previous observations where GA-dependent Hsp90 inhibition has been shown to interfere with IKK activation (19). To further assess the effect of Hsp90 on Tax-mediated NFB activation, we used lysate, immunodepleted for Hsp90 (Fig. 5B), in our in vitro cell-free signaling assay. Addition of recombinant Tax and TRAF6 proteins to Hsp90-immunodepleted lysate failed to activate the NFB pathway as shown by immunoblotting against phospho-IB (Fig. 5C, lanes 5 and 6). Because Hsp90 is known to maintain the structural integrity of protein complexes, these observations support a new model that a preformed native complex requiring active Hsp90 is essential for the activation of the IKK complex by Tax. This supports the proposed model that binding of Tax to IKK␥ in the IKK complex causes a conformational change that induces autoactivation of the kinases in the complex (5).
Tax Cannot Bypass Inhibition of the IKK Complex by YopJ Acetylation-Tax is one of the upstream activators of the IKK complex; however, the mechanism by which it activates IKK has not been completely deciphered. The effector protein YopJ from Yersinia was recently shown to inhibit MAPK kinase and IKK activation by acetylating the conserved serine and threonine residues in the activation loop of the kinase (13,14,20). Consistent with previous findings, YopJ inhibited IKK activation by Tax, as shown by using a NFB luciferase reporter (Fig.  6A) (15). The catalytically inactive mutant YopJC172A, however, had no effect on Tax-mediated activation of NFB (Fig.  6A). Transfection experiments with YopJ and Tax followed by immunoblotting with antibody against phospho-IB further confirmed these results because YopJ, but not YopJC172A, inhibited Tax-mediated IB phosphorylation (Fig. 6B). As before, cells constitutively expressed FLAG-IB to detect phosphorylation of exogenous IB.
YopJ Inhibits in Vitro Activation of IKK by Tax-To gain insight into the mechanism used by Tax, in vitro assay of Tax-mediated IKK activation was utilized. 10 mg/ml membrane-cleared lysate was prepared from cells transfected with vector (V), YopJ (J), or YopJC172A (C/A) plasmids (13). Recombinant WT Tax was then used to activate these lysates. Tax was able to phosphorylate endogenous IB in both V-and C/A-transfected cell lysates (Fig. 6C). By contrast and as expected, Tax had lost its ability to phosphorylate IB from cells expressing YopJ. The results from this in vitro assay confirm the above in vivo results and are in accordance with the work by Carter et al. (15). Just like all other upstream stimuli, YopJ is able to block the Tax-mediated activation of the IKK complex.

DISCUSSION
In this study, an in vitro signaling assay was used to analyze the activation of IKK by Tax. Recombinant Tax, purified from bacteria, was able to efficiently induce IB phosphorylation in cleared lysates. The phosphorylation profiles of the Tax-activated complexes isolated by immunoprecipitation with IKK␥ or Tax appear the same, supporting the accepted hypothesis that Tax, IKK␥, and the IKKs are parts of the same signaling complex. The PP2A binding-deficient mutants of Tax were also able to activate NFB signaling, although not as strongly as WT Tax. These observations do not support the model that the binding of PP2A to Tax is essential for Tax-mediated activation of the IKK complex (7). In addition, experiments with okadaic acid added to the in vitro assay did not alter the ability of WT Tax to activate the IKK complex and therefore do not support the alternative model that Tax-mediated IKK activation is positively regulated by PP2A (9). By contrast, Tax was unable to activate the NFB pathway when Hsp90, an integral compo-nent of the IKK complex, was inhibited by the addition of geldanamycin to the assay. Binding of Tax to IKK␥ in the IKK complex may cause a conformational change that induces autoactivation of the kinases in the complex. As more mutants associated with the other activities of viral Tax are discovered (21), this system can be used to diagnose their role in the activation of IKK.
Both in vivo and in vitro studies revealed that YopJ blocks the activation of IKK by Tax. YopJ leads to the acetylation of the conserved serine and threonine in the activation loop of kinases, including IKK. When the activation loop residues are acetylated, they can no longer be phosphorylated and the kinase cannot be activated (13,14,20). The acetyltransferase activity of YopJ on IKK thus competes against phosphorylation of IKK by upstream kinases. As YopJ inhibits Tax-mediated IKK activation, these studies strongly indicate that phosphorylation of IKK is a key intermediate step in the activation of IKK by Tax.
Our results support previously postulated mechanisms for Tax-mediated activation of the IKK complex, including induction of a conformational change or recruitment of an upstream kinase (5). The in vitro signaling system used in this study in combination with a number of inhibitors and activators has been useful for the elucidation of the biochemical mechanism of Tax-induced activation of the IKK complex. Based on these studies, a model is proposed whereby Tax-dependent activation of the IKK complex requires active Hsp90 and phosphorylation of IKK. Future biochemical studies that further dissect this mechanism may reveal other factors that are essential for the activation of the IKK complex by Tax.