Recombinant IκB Kinases α and β Are Direct Kinases of IκBα*

Activation of the transcription factor NF-κB is regulated by the phosphorylation and subsequent degradation of its inhibitory subunit, IκB. A large multiprotein complex, the IκB kinase (IKK), catalyzes the phosphorylation of IκB. The two kinase components of the IKK complex, IKKα and IKKβ, were overexpressed in insect cells and purified to homogeneity. Both purified IKKα and IKKβ specifically catalyzed the phosphorylation of the regulatory serine residues of IκBα. Hence, IKKα and IKKβ were functional catalytic subunits of the IKK complex. Purified IKKα and IKKβ also preferentially phosphorylated serine as opposed to threonine residues of IκBα, consistent with the substrate preference of the IKK complex. Kinetic analysis of purified IKKα and IKKβ revealed that the kinase activity of IKKβ on IκBα is 50–60-fold higher than that of IKKα. The primary difference between the two activities is the K m for IκBα. The kinetics of both IKKα and IKKβ followed a sequential Bi Bi mechanism. No synergistic effects on IκBα phosphorylation were detected between IKKα and IKKβ. Thus, in vitro, IKKα and IKKβ are two independent kinases of IκBα.

Activation of the transcription factor NF-B is regulated by the phosphorylation and subsequent degradation of its inhibitory subunit, IB. A large multiprotein complex, the IB kinase (IKK), catalyzes the phosphorylation of IB. The two kinase components of the IKK complex, IKK␣ and IKK␤, were overexpressed in insect cells and purified to homogeneity. Both purified IKK␣ and IKK␤ specifically catalyzed the phosphorylation of the regulatory serine residues of IB␣. Hence, IKK␣ and IKK␤ were functional catalytic subunits of the IKK complex. Purified IKK␣ and IKK␤ also preferentially phosphorylated serine as opposed to threonine residues of IB␣, consistent with the substrate preference of the IKK complex. Kinetic analysis of purified IKK␣ and IKK␤ revealed that the kinase activity of IKK␤ on IB␣ is 50 -60-fold higher than that of IKK␣. The primary difference between the two activities is the K m for IB␣. The kinetics of both IKK␣ and IKK␤ followed a sequential Bi Bi mechanism. No synergistic effects on IB␣ phosphorylation were detected between IKK␣ and IKK␤. Thus, in vitro, IKK␣ and IKK␤ are two independent kinases of IB␣.
The transcription factor NF-B plays a critical role in immune and inflammatory responses. It is regulated by the signaling of receptors for inflammatory cytokines such as TNF-␣, 1 IL-1, or other external stimuli (1). In resting cells, NF-B is sequestered in the cytoplasm through its association with inhibitory proteins termed IB. Two IB proteins, IB␣ and IB␤, are rapidly phosphorylated at Ser residues in the Nterminal region upon stimulation by TNF-␣ and IL-1. The regulated phosphorylation is at Ser-32 and Ser-36 of IB␣ and, correspondingly, Ser-19 and Ser-23 of IB␤ (2,3). The more recently cloned IB isoform, IB⑀, also contains the two conserved Ser residues at the N terminus for signal-induced degradation (4). Phosphorylated IB␣ and IB␤ are subsequently ubiquitinated and undergo ubiquitin-dependent degradation by the 26 S proteasome (3,5). Degradation of IB results in the release of NF-B, which then translocates to the nucleus, where it up-regulates the transcription of its target genes (1).
A 500 -900-kDa protein complex that contains the TNF-␣-induced IB kinase (IKK) has been purified and characterized independently by two groups (6,7). The IKK complex phosphorylates IB␣ at the specific Ser residues that target the protein for ubiquitination and degradation. Two kinase subunits of the IKK complex, denoted IKK␣ and IKK␤, have been cloned (6 -10). IKK␣ or IKK␤ overexpressed in mammalian cells specifically phosphorylates IB␣ and IB␤ after immunoprecipitation, and their kinase activities can be induced by TNF-␣ or IL-1 (6 -10). In HeLa cells, expression of antisense IKK␣ inhibited NF-B activation by TNF-␣ or IL-1 (6). Furthermore, overexpression of dominant-negative mutants of either IKK␣ or IKK␤ blocked TNF-␣/IL-1-induced NF-B activation (8,10). Thus, both IKK␣ and IKK␤ contribute to the activity of the IKK complex and are involved in NF-B activation. An additional protein kinase, NF-B-inducing kinase (NIK), has been shown to be involved in the activation of IB phosphorylation in both the IL-1 and TNF-␣ pathways (11). IKK␣ has also been identified as an NIK-interacting protein by yeast two-hybrid analysis (8). Co-expression of NIK and IKK␣ (or IKK␤) stimulates the kinase activity of IKK␣ and IKK␤ (8,12). Moreover, NIK-dependent NF-B activation is blocked by a dominant-negative IKK␣ mutant (8). These results indicate that NIK is an upstream regulator of IKK␣ and IKK␤.
Although the results using expressed and immunoprecipitated proteins are consistent with IKK␣ and IKK␤ being the IB kinases, it has not been demonstrated that IKK␣ or IKK␤ directly phosphorylates IB␣. All phosphorylation experiments reported so far involve co-precipitation of other components of the high molecular mass complex from mammalian cells (8 -10). Additionally, it has been reported that IKK␣ or IKK␤ synthesized in wheat germ extracts was unable to phosphorylate IB (9). Thus, it is possible that IKK␣ and IKK␤ may activate an as yet unidentified IB kinase that co-immunoprecipitates with IKK␣ and IKK␤ and phosphorylates IB␣ (13). To resolve this issue, we have pursued in vitro studies with purified recombinant proteins. Recombinant IKK␣ and IKK␤ were overexpressed in insect cells and purified. Here we demonstrate that purified IKK␣ or IKK␤ alone is capable of phosphorylating specific Ser residues of IB␣ in vitro, demonstrating that both IKK␣ and IKK␤ are direct kinases of IB␣. We also show that both IKK␣ and IKK␤ display a sequential Bi Bi mechanism. Kinetic parameters for both enzymes indicate that IKK␤ is 50 -60-fold more active than IKK␣ in catalyzing the phosphorylation of IB␣. Interestingly, IKK␣ and IKK␤ showed no synergy in catalyzing the phosphorylation of IB␣ in vitro. Such mechanism studies with purified IKK␣ and IKK␤ provide new insights into the functions of these two kinase components of the multiprotein IKK complex.

EXPERIMENTAL PROCEDURES
Cloning-IKK␣ and IKK␤ were cloned from a Jurkat cDNA library (CLONTECH) by polymerase chain reaction and were expressed as N-terminal Flag-tagged fusion proteins using the baculovirus expression vector pFastBacI (Life Technologies, Inc.). IKK␣ was cloned into * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ These authors contributed equally to this work. the RsrII and XhoI sites of pFastBacI. IKK␤ was cloned into the RsrII and KpnI sites of pFastBacI. The protein sequences for the expressed IKK␣ and IKK␤ are MDYKDDDDKEF-IKK␣ (8) and MDYKDDDDK-LAAANSS-IKK␤ (10), respectively. The K44M IKK␤ mutant was constructed in the same vector as used for wild-type IKK␤. The Flag-tagged K44A IKK␣ mutant was constructed in the baculovirus expression vector pVL1393 (Invitrogen) at the BamHI and NotI sites. The above Flag-tagged recombinant proteins were expressed in Sf9 cells. IKK␣ was also expressed as an N-terminal His 6 -tagged protein in Sf9 cells using pFastBacIHTa (Life Technologies, Inc.; EcoRI and XhoI sites). IB␣ and IB␣(S32A/S36A) were expressed as a His 6 -tagged thioredoxin fusion protein (TRX-IB␣-(1-54)) in Escherichia coli BL21(DE3). The sequences of all the constructed clones were verified by DNA sequencing.
In Vitro Phosphorylation Assays-Kinase assays were performed in 20 mM HEPES, pH 7.5, 10 mM MgCl 2 , 2 mM MnCl 2 , 100 mM NaCl, 100 M Na 3 VO 4 , 20 mM ␤-glycerophosphate, and 1 mM DTT. The amounts of the substrates ATP, [␥-33 P]ATP (2000 Ci/mmol; NEN Life Science Products), and IB␣ are specified for each individual experiment. Reactions were performed at 23°C for 10 -30 min. Samples were analyzed by 4 -20% SDS-polyacrylamide gel electrophoresis and autoradiography or by trichloroacetic acid precipitation on a microtiter plate (Millipore Corp.). For microtiter plate assays, 55 l of reaction sample/well was quenched with 50 l of 5% pyrophosphate, precipitated with 75 l of 10% trichloroacetic acid plus 5% pyrophosphate on a Millipore plate harvester, washed twice with 200 l of 10% trichloroacetic acid plus 5% pyrophosphate, and counted in 100 l of scintillation mixture (SCINT 20, Packard Instrument Co.) using a TopCount liquid scintillation counter (Packard NXT). Autophosphorylation of IKK␣ or IKK␤ was subtracted out using a control assay without IB␣. Assay conditions were controlled so that the degree of phosphorylation of IB␣ was linear with time and concentration of enzyme. The counts represent initial velocity of IKK-catalyzed phosphorylation (Ͻ10% of total ATP conversion). All experiments were performed in duplicate.
Kinetic Analysis-All kinetic assays were performed using a microtiter plate assay (see above). For two-substrate profile analysis, initial velocity studies were performed with varying concentrations of IB␣ at several fixed ATP concentrations. Lineweaver-Burk double-reciprocal plots were generated by linear least-square fits of the data. The apparent kinetic constants were determined from the x and y intercepts of the double-reciprocal plots. Secondary plots were generated by replotting the slopes and the y intercepts of the lines as a function of 1/[ATP]. For a random sequential model as described previously (17,18), values of K ATP , K IB␣ , V max , and ␣ can be determined from the x and y intercepts of the secondary plots. K ATP and K IB␣ are the dissociation constants for ATP and IB␣, respectively. The constant ␣ is the ratio of apparent dissociation constants for binding IB␣ in the presence and absence of ATP, as described (17). The value of ␣ indicates whether binding of one substrate (ATP) affects the affinity of the enzyme for the other substrate (IB␣) (17,18).
Co-immunoprecipitation of IKK␣ and IKK␤-Sf9 cells (1.5 ϫ 10 6 cells/ml) were infected with various combinations of baculoviruses encoding His 6 -IKK␣, Flag-IKK␣, or Flag-IKK␤ at a multiplicity of infection of 6 for each virus and grown at 28°C until harvested at 72 h post-infection. A fraction of the infected cells (5 ϫ 10 6 cells) were lysed in 250 l of 50 mM HEPES, pH 7.5, 200 mM NaCl, 1% Nonidet P-40, 10% glycerol, 1 mM Na 3 VO 4 , 1 mM EDTA, 1 mM DTT, and protease inhibitor cocktail (Boehringer Mannheim and PharMingen). Cellular debris was removed by centrifugation at 16,000 ϫ g for 5 min. The lysate was mixed with 4 g each of anti-His 4 and anti-His 5 monoclonal antibodies (QIAGEN Inc.) and 30 l of protein G Plus/protein A-agarose beads (Oncogene Science Inc.) at 4°C for 1 h. The beads were washed four times with 500 l of 50 mM HEPES, pH 7.5, 300 mM NaCl, 10% glycerol, 1 mM Na 3 VO 4 , 1 mM EDTA, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride, and bound proteins were fractionated by SDS-polyacrylamide gel electrophoresis and analyzed by immunoblotting.
Purification of the IKK␣⅐IKK␤ Complex-Sf9 cells were doubly infected with baculoviruses encoding His 6 -IKK␣ and Flag-IKK␤. The cell pellet was lysed and purified using a Flag affinity column as described above, except that 1 mM ␤-mercaptoethanol instead of DTT was used in the elution buffer. One milliliter of the Flag eluate (0.5 mg/ml protein) was mixed with 50 l of 1 M sodium phosphate buffer, pH 8.0, and loaded onto a Ni 2ϩ -nitrilotriacetic acid spin column (QIAGEN Inc.). The column was washed three times with 0.6 ml of 0.1 M Na 2 HPO 4 , pH 8.0, 0.3 M NaCl, 20 mM imidazole, 10% glycerol, and 1 mM ␤-mercaptoethanol and eluted with 250 l of 0.1 M Na 2 HPO 4 , pH 8.0, 0.3 M NaCl, 0.3 M imidazole, 10% glycerol, 1 mM ␤-mercaptoethanol, and protease inhibitor cocktail (Boehringer Mannheim).

IKK␣ and IKK␤ Independently
Phosphorylate IB␣-To investigate whether IKK␣ and IKK␤ directly phosphorylate IB␣ in vitro, we expressed IKK␣ and IKK␤ in insect cells as Nterminal Flag-tagged proteins and purified them by affinity chromatography. Both IKK␣ and IKK␤ were purified to apparent homogeneity and migrated on SDS-polyacrylamide gel with molecular masses predicted by their respective DNA sequences (Fig. 1). The predicted molecular masses of recombinant IKK␣ and IKK␤ (86 kDa and 88 kDa, respectively) were confirmed by mass spectroscopic analysis (data not shown). Purified recombinant IKK␣ and IKK␤ were tested for kinase activity by phosphorylation of IB␣. Either IKK␣ or IKK␤ alone could phosphorylate IB␣ (Fig. 2). Autophosphorylation of both IKK␣ and IKK␤ was also observed. The phosphorylation of IB␣ was specific for Ser-32 and Ser-36 of IB␣ since there was no phosphorylation of the S32A/S36A mutant of IB␣ (Fig. 2). To eliminate the possibility of trace kinase contaminants from the insect cells, two catalytically inactive kinase mutants (K44A IKK␣ and K44M IKK␤) were expressed and purified. The purified recombinant kinase mutants K44A IKK␣ and K44M IKK␤ lacked activity in both the autophosphorylation and phosphorylation of IB␣ (Fig. 2). Thus, both IKK␣ and IKK␤ are direct IB␣ kinases.
One distinguishing feature of IKK as a Ser/Thr kinase is that it preferentially phosphorylates Ser versus Thr (6, 7). To test whether purified IKK␣ and IKK␤ are Ser-selective kinases, we  (Fig. 3). Both IKK␣ and IKK␤ showed significantly higher activity in the phosphorylation of Ser versus Thr residues of IB␣ (Fig. 3). Quantitation of the phosphorylation by PhosphorImager analysis indicated that IKK␣ and IKK␤ had 11-and 9-fold higher activities in the phosphorylation of Ser versus Thr residues of IB␣-(21-41), respectively. The results indicate that the Ser selectivity of IKK is, at least in part, conveyed at the kinase subunit level.
Kinetics of IKK␣ and IKK␤ in the Phosphorylation of IB␣-(1-54)-Since both IKK␣ and IKK␤ phosphorylate IB␣ independently in vitro, we determined the kinetics of both in phosphorylating IB␣. Fig. 4 shows Michaelis-Menten plots of IKK velocity as a function of the concentration of ATP at a constant concentration of IB␣. Both IKK␣ and IKK␤ displayed standard Michaelis-Menten kinetics. The apparent K m(ATP) values of IKK␣ and IKK␤ were 126 and 136 nM, respectively ( Fig. 4 and Table I). The kinase activities of IKK␣ and IKK␤ were further determined as a function of varying concentrations of IB␣ (0.7-53 M) at five fixed concentrations of ATP (30 -1200 nM). The Lineweaver-Burk plots of the data for both IKK␣ and IKK␤ followed Michaelis-Menten kinetics (Fig. 5, A and B). For both IKK␣ and IKK␤, a series of double-reciprocal straight-line plots intersected to the left side of the ordinates, indicating a sequential kinetic mechanism (18). At a saturating concentration of ATP (10-fold of K m(ATP) app ), the apparent K m value of IKK␣ for IB␣ (23 M) was 17-fold higher than that of IKK␤ (1.3 M). Under the same conditions, the apparent maximal turnover rates of IKK␣ and IKK␤ were 0.08/min and 0.27/min at 23°C, respectively (Fig. 5, A and B; and Table I). Based on the values of k cat /K m(IB␣) , IKK␤ is 58-fold more potent than IKK␣ in the phosphorylation of IB␣ (Table I). This result is consistent with the observation that when overexpressed in 293 cells, IKK␤ is more potent in activating the NF-B pathway than IKK␣ (10). One interesting observation is that the plots shown in Fig. 5 (A  and B) intersected on the abscissa, indicating that the concentration of ATP has no effect on the apparent K m for IB␣. It was recently reported by Burke et al. (17) that a multisubunit IB kinase complex purified from TNF-␣-stimulated HeLa cells showed random Bi Bi sequential kinetics with ATP and IB␣ binding in a cooperative fashion (␣ ϭ 0.11). To directly compare the kinetics of purified IKK␣ and IKK␤ with those of the multisubunit IKK complex, the plots shown in Fig. 5 (A and B) were fitted to the random sequential model as described (17,18). Accordingly, the data from Fig. 5 (A and B) were replotted as the slopes of the lines as a function of 1/[ATP] (Fig. 5C), and the y intercepts as a function of 1/[ATP] (Fig. 5D) (18). For IKK␣, values of 85 nM, 25 M, 0.09/min, and 1.0 were obtained for K ATP , K IB␣ , k cat , and ␣, respectively (Fig. 5, C and D; and Table I Table I). The 1.0 ␣ value for IKK␣ and IKK␤ demonstrates that the binding of one substrate to the enzyme has no effect on the affinity for the other substrate. The obtained K ATP and ␣ values for IKK␣ and IKK␤ are significantly different from those for the multisubunit IKK complex (K ATP ϭ  (Table I). Although the kinetics for IKK␣ and IKK␤ fit well to the random sequential mechanism, further inhibition studies are needed to distinguish between random and ordered Bi Bi mechanisms.

IKK␣ and IKK␤ Show No Synergistic Effect on the Phosphorylation of IB␣-
The native IKK complex contains both IKK␣ and IKK␤ (6,7). However, our in vitro studies showed that either IKK␣ or IKK␤ alone could phosphorylate IB␣. We then investigated whether IKK␣ and IKK␤ cooperate with each other in the phosphorylation of IB␣. When separately purified IKK␣ and IKK␤ were mixed at equimolar concentrations, no synergistic effect on kinase activity was observed (Fig. 6A). In fact, if the kinase activities shown in Fig. 6A are normalized for the amount of IKK␣ and IKK␤, the activity of the IKK␣/IKK␤ mixture was ϳ20% less than the additive activity of IKK␣ and IKK␤ (Fig. 6A). Because IKK␤ was more active than IKK␣, the activity of the IKK␣/IKK␤ mixture is probably due to the dominant activity of IKK␤. We further investigated whether the catalytically inactive mutant of IKK␣ would inhibit the kinase activity of wild-type IKK␤ in vitro. When IKK␤ was mixed with various amounts of the K44A IKK␣ mutant and then assayed for IB␣ phosphorylation, no decrease in kinase activity was observed (Fig. 6B). Conversely, IKK␣ was not inhibited by the inactive K44M IKK␤ mutant either (data not shown). These results are consistent with IKK␣ and IKK␤ being independent kinases in vitro.
Kinase Activity of the Co-translated IKK␣⅐IKK␤ Complex Is Similar to the Additive Activity of IKK␣ and IKK␤-In mammalian cells, co-expression of IKK␣ and IKK␤ yielded an IKK␣⅐IKK␤ heterocomplex (9,10). To address whether IKK␣ and IKK␤ form a stable complex when they are co-expressed in insect cells, Sf9 cells were infected with baculoviruses expressing various combinations of His 6 -IKK␣, Flag-IKK␣, and Flag-IKK␤. The resulting cell extract was subjected to co-immunoprecipitation analysis. Flag-IKK␤ co-immunoprecipitated with His 6 -IKK␣ when they were co-expressed in Sf9 cells (Fig. 7), indicating formation of a stable IKK␣⅐IKK␤ heterocomplex. When His 6 -IKK␣, Flag-IKK␣, and Flag-IKK␤ were co-expressed in Sf9 cells, the formation of both a His 6 -IKK␣⅐Flag-IKK␣ homocomplex and a His 6 -IKK␣⅐Flag-IKK␤ heterocomplex was detected (Fig. 7). However, more Flag-IKK␤ was associated with His 6 -IKK␣ than Flag-IKK␣, despite similar expression levels of Flag-IKK␣ and Flag-IKK␤ (Fig. 7). This suggests that assembly of the IKK␣⅐IKK␤ heterocomplex may occur more readily than assembly of the IKK␣⅐IKK␣ homocomplex in insect cells. The co-expressed His 6 -IKK␣⅐Flag-IKK␤ complex was purified by anti-Flag M2 affinity chromatography, followed by a Ni 2ϩ -nitrilotriacetic acid affinity column. The kinase activity of the purified complex was only ϳ20% more than the additive   ). b The parameters were obtained by fitting the two-substrate kinetics to a random sequential model as described under "Experimental Procedures." activity of IKK␣ and IKK␤, based on normalized specific activities of each enzyme (Fig. 6A). Consistent with the data of the IKK␣/IKK␤ mixture, this result suggests that there was no significant synergy in IB␣ phosphorylation between IKK␣ and IKK␤ in the co-translated complex. Collectively, these data suggest that direct interactions between recombinant IKK␣ and IKK␤ do not enhance their kinase activity. DISCUSSION Although IKK␣ and IKK␤ are essential for IB␣ phosphorylation (6 -10), it has not been clear whether they directly phosphorylate IB␣. Here we have shown that purified recombinant IKK␣ or IKK␤ (but not inactive kinase mutants of IKK␣ or IKK␤) phosphorylates IB␣ in vitro, indicating that IKK␣ and IKK␤ are direct kinases of IB␣. It is possible that an undetected co-purifying kinase could be activated by IKKs and, in turn, phosphorylates IB␣. However, the fact that there is a significant difference in K m(IB␣) between purified IKK␣ and IKK␤ strongly suggests that the two protein preparations contain different kinases. It is highly unlikely that purified IKK␣ and IKK␤ contained two different contaminating kinases since they were expressed and purified identically. It has been pre-viously shown that IKK␣ and IKK␤ display distinct modes of regulation when overexpressed in mammalian cells (7,9). Overexpressed IKK␤ is constitutively active, whereas overexpressed IKK␣ is inactive unless cells are stimulated by TNF-␣ (7). As a consequence, overexpression of IKK␤ resulted in greater activation of an NF-B reporter gene than overexpression of IKK␣ (10). Our finding that purified recombinant IKK␤ is 50 -60-fold more active than IKK␣ toward IB␣ is in agreement with previous results.
Recently, the enzyme kinetics of a multisubunit IB kinase complex purified from TNF-␣-stimulated HeLa cells were reported by Burke et al. (17). This kinase complex showed random sequential kinetics, and the two substrates (IB␣ and ATP) bound in a cooperative fashion (␣ ϭ 0.11). Although the kinetics of recombinant IKK␣ and IKK␤ are consistent with a random sequential mechanism, both IKK␣ and IKK␤ showed an ␣ value of 1.0. Thus, unlike the IKK complex reported by Burke et al. (17), the affinity of IB␣ for either IKK␣ or IKK␤ is not affected by the binding of ATP. The reported dissociation constant of ATP for the IKK complex prepared by Burke et al. (17) was 7 M. In contrast, the ATP dissociation constants for IKK␣ and IKK␤ are 85 and 130 nM (Table I), respectively. Such notable differences in enzyme kinetics between IKK␣ (or IKK␤) and the IKK complex would probably suggest either that the kinase subunits in the IKK complex purified by Burke et al. (17) may be different from IKK␣ and IKK␤ or that the activity of IKK␣ or IKK␤ in the IKK complex was modified by other proteins in the complex. The IKK complex purified by Burke et al. (17) is apparently different from the IKK complex that contains IKK␣ and IKK␤ (6,7) since the former complex also phosphorylated the C-terminal peptide (residues 279 -303) of IB␣ (17), whereas the latter one did not phosphorylate the C-terminal region (residues 242-314) of IB␣ (7). The K m values of IKK␣ and IKK␤ for ATP (130 nM) are considerably lower than those of other protein kinases, such as cAMP-dependent protein kinase (K m ϭ 10 M) (19) and p38 mitogen-activated protein kinase (K m ϭ 23 M) (20). This low K m(ATP) feature of IKK␣ and IKK␤, in combination with their Ser versus Thr substrate selectivity, distinguishes IKK␣ and IKK␤ as unique members of the Ser kinase family.
In addition to IB␣, both IB␤ and IB⑀ have also been reported to be phosphorylated by IKK (8,10,12). Our observation that IKK␤ displayed a greater affinity (17-fold difference in K m(IB␣) ) for IB␣ raises the possibility that the two kinases may also have different substrate preferences for phosphorylation of IB␤ and IB⑀. Thus, regulation of the degradation of different IB isoforms could be achieved by two kinases with different preferences for the substrate IB isoforms. On the other hand, IKK␣ and IKK␤ are also differentially regulated through upstream kinases; for example, MEKK1 has been shown to preferentially activate IKK␤ (12,21). Thus, differential regulation of IKK␣ and IKK␤ could also regulate the degradation of different IB proteins.
It has been reported that NIK activates IKK␣ and IKK␤ and acts as an upstream kinase of IKK␣ (8,10,21). NIK co-expressed with IKK␣ in 293 cells preferentially phosphorylated the co-expressed IKK␣ at Ser-176, located in the activation loop of IKK␣ (22). However, these experiments were performed using immunoprecipitated proteins from crude cell lysate. We have not been able to show effects of recombinant NIK or catalytically inactive NIK on the activity of recombinant IKK␣ or IKK␤ in phosphorylating IB␣ in vitro. 2 It is possible that although the baculovirus-expressed NIK showed reasonable activity in autophosphorylation, it was still less active or regulated differently than the NIK expressed in mammalian cells. Alternatively, an unidentified protein not found in insect cells may be required to mediate or regulate the action of NIK on IKK␣.
It is puzzling that IKK␣ or IKK␤ alone was active, yet they had no synergistic kinase activity. This indicates that in vitro, IKK␣ and IKK␤ are two independent kinases, despite their coexistence in the 500 -900-kDa IKK complex in cells (6,7). In insect cells, co-expression of His 6 -IKK␣ and Flag-IKK␤ yielded a stable His 6 -IKK␣⅐Flag-IKK␤ complex. Consistent with IKK␣ and IKK␤ being two independent kinases in vitro, the coexpressed His 6 -IKK␣⅐Flag-IKK␤ complex displayed activity similar to the additive kinase activity of IKK␣ and IKK␤ (Fig.  6A). It is possible that significant regulating effects on IKK␣ and IKK␤ activity could occur in cells through interactions (catalytic or protein-protein) with other subunit(s) in the 500 -900-kDa multiprotein IKK complex. Since the activation of NF-B can be induced by various signals, other subunits in the IKK complex may also serve as signaling regulators or adapter molecules to allow selectivity in responding to different stimuli.
In summary, through these studies with purified recombinant IKK␣ and IKK␤, we have demonstrated that both IKK␣ and IKK␤ are direct kinases of IB␣. It is therefore likely that IKK␣ and IKK␤ are indeed catalytic subunits of the multiprotein IKK complex. The kinetics of both enzymes followed a sequential Bi Bi mechanism. The significantly higher activity of IKK␤ in comparison with IKK␣ for IB␣ may indicate their different roles in cellular function. There was no catalytic synergy between IKK␣ and IKK␤ in the current in vitro studies, suggesting that further investigations into the functions of other components in the IKK complex are needed to understand the relative functional contributions of IKK␣ and IKK␤ in cells.