IκB Kinases α and β Show a Random Sequential Kinetic Mechanism and Are Inhibited by Staurosporine and Quercetin*

Activation of transcription factor NF-κB is regulated by phosphorylation and subsequent degradation of its inhibitory subunit IκB. The signal-induced phosphorylation of IκB involves two IκB kinases, IKKα and IKKβ. In the present study, we investigated the kinetic mechanisms of IKKα and IKKβ by substrate and product inhibition. For both IKKα and IKKβ, the product ADP was a competitive inhibitor versus ATP and a non-competitive inhibitor versus IκBα. An alternative peptide substrate, IκBα-(21–41), was a competitive inhibitorversus IκBα and a non-competitive inhibitorversus ATP for both kinases. These results rigorously eliminate the possibility of an ordered sequential mechanism and demonstrate that both kinases have a random sequential bi bi mechanism. Two natural compounds, quercetin and staurosporine, had previously been shown to inhibit the NF-κB pathway, but the molecular target(s) of these compounds in the event had not been established. Here we demonstrate that quercetin and staurosporine potently inhibit both IKKα and IKKβ. Daidzein, a quercetin analogue that does not inhibit NF-κB activation, showed no significant inhibition of either enzyme. This suggests that the inhibitory properties of quercetin and staurosporine in the NF-κB pathway are mediated in part by their inhibition of IKKα and IKKβ. Mechanism studies reveal that staurosporine is a competitive inhibitor versus ATP, whereas quercetin serves as a mixed type inhibitor versusATP. The strong inhibition of IKKβ by staurosporine (K i = 172 nm) and ADP (K i = 136 nm) provides a rationale and structural framework for designing potent ATP-site inhibitors of IKKβ, which is an attractive drug target for inflammatory diseases.

Activation of transcription factor NF-B is regulated by phosphorylation and subsequent degradation of its inhibitory subunit IB. The signal-induced phosphorylation of IB involves two IB kinases, IKK␣ and IKK␤. In the present study, we investigated the kinetic mechanisms of IKK␣ and IKK␤ by substrate and product inhibition. For both IKK␣ and IKK␤, the product ADP was a competitive inhibitor versus ATP and a non-competitive inhibitor versus IB␣. An alternative peptide substrate, IB␣-(21-41), was a competitive inhibitor versus IB␣ and a non-competitive inhibitor versus ATP for both kinases. These results rigorously eliminate the possibility of an ordered sequential mechanism and demonstrate that both kinases have a random sequential bi bi mechanism. Two natural compounds, quercetin and staurosporine, had previously been shown to inhibit the NF-B pathway, but the molecular target(s) of these compounds in the event had not been established. Here we demonstrate that quercetin and staurosporine potently inhibit both IKK␣ and IKK␤. Daidzein, a quercetin analogue that does not inhibit NF-B activation, showed no significant inhibition of either enzyme. This suggests that the inhibitory properties of quercetin and staurosporine in the NF-B pathway are mediated in part by their inhibition of IKK␣ and IKK␤. Mechanism studies reveal that staurosporine is a competitive inhibitor versus ATP, whereas quercetin serves as a mixed type inhibitor versus ATP. The strong inhibition of IKK␤ by staurosporine (K i ‫؍‬ 172 nM) and ADP (K i ‫؍‬ 136 nM) provides a rationale and structural framework for designing potent ATP-site inhibitors of IKK␤, which is an attractive drug target for inflammatory diseases.
The transcription factor NF-B is regulated by the signaling of receptors for inflammatory cytokines such as TNF␣, 1 interleukin-1, or other external stimuli (1). In resting cells, NF-B is sequestered in the cytoplasm through its association with inhibitory proteins termed IB. When cells are stimulated by TNF␣ or interleukin-1, IB proteins (IB␣ and IB␤) are rapidly phosphorylated at Ser residues in the N-terminal region (2,3). Phosphorylated IB␣ and IB␤ are subsequently ubiquitinated and undergo ubiquitin-dependent degradation by the 26 S proteasome (3,4). Degradation of IB results in the release of NF-B which then translocates to the nucleus where it up-regulates the transcription of target genes (1).
IB␣ and IB␤ are phosphorylated by a 500 -900-kDa IB kinase (IKK) (5,6). Two kinases in the IKK complex, denoted IKK␣ and IKK␤ (or IKK-1 and IKK-2), phosphorylate IB␣ at the specific Ser residues that target the protein for ubiquitination and degradation (5)(6)(7)(8)(9). Both IKK␣ and IKK␤ contribute to the activity of the IKK complex and are involved in NF-B activation (5)(6)(7)(8)(9). The physiological function of these protein kinases was recently explored by analysis of IKK␣-deficient or IKK␤-deficient mice (10 -15). Mouse embryonic fibroblast cells that were isolated from IKK␤(Ϫ/Ϫ) embryos showed a marked reduction in TNF␣-and interleukin-1-induced NF-B activity and enhanced apoptosis in response to TNF␣ (11,14,15). In contrast, IKK␣ was not required for activation of IKK and degradation of IB by pro-inflammatory stimuli (10,12). These results show that IKK␤, not IKK␣, is the target for pro-inflammatory stimuli. On the other hand, IKK␣ is essential for development of skin and skeleton during embryogenesis (10,12,13). NF-B activation is impaired in the basal layer of epidermal cells in IKK␣-deficient mice (12). Since IKK␣ and IKK␤ have distinct functions, it is informative to compare the kinetic mechanisms of both kinases. Inhibitors with selectivity between these two kinases would help to elucidate further their different functions in cells and in animal models.
IKK␣ and IKK␤ share ϳ50% overall homology, and both contain a conserved N-terminal Ser/Thr kinase domain, a leucine-zipper region, and a C-terminal helix-loop-helix (HLH) motif (6 -9). Such folding is unique among the known kinases. It has been shown that the HLH domain of IKK␤ is required for its kinase activity and the HLH domain can activate the truncated IKK␤ (HLH deletion) mutant in trans (16). This suggests a functional interaction between the HLH domain and the kinase domain of IKK␤. IKK␣ and IKK␤ also share a distinguishing feature in that they have a strong preference for Ser versus Thr on the substrates (5,6). It is important to understand the kinetic mechanisms of these two unique members of the Ser/Thr kinase family.
Several naturally existing kinase inhibitors have been reported to inhibit the NF-B pathway. Quercetin, a flavonoid that occurs in many fruits and vegetables (17), is a nonspecific inhibitor of protein kinases (18) and suppresses TNF-induced NF-B activation (19). The inhibitor blocks the degradation of IB␣ and the consequent translocation of the NF-B p65 subunit (19). Staurosporine, a microbial alkaloid that was isolated from Streptomyces staurosporeus (20), has shown potent inhibition of both tyrosine and Ser/Thr kinases (18,21). In THP-1 monocytic cells, staurosporine inhibits LPS-dependent NF-B activation, suggesting that staurosporine-sensitive kinase(s) are involved in LPS-mediated NF-B activation (22). The inhibitory effects of quercetin and staurosporine in the NF-B pathway are consistent with their anti-inflammatory responses as observed in various animal models including experimental arthritis and experimental colitis (23)(24)(25)(26). However, the molecular target(s) of staurosporine and quercetin in the NF-B * 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. signaling cascade have not been identified. Since IKK is essential for activation of NF-B by both TNF␣ and LPS (6 -9, 27), it is important to know whether quercetin and staurosporine inhibit IKK␣ and IKK␤. It was recently shown that high concentrations of the anti-inflammatory agent aspirin inhibits IKK␤ (IC 50 ϭ ϳ50 M) (28), consistent with its inhibitory effect on the NF-B pathway (29).
Previously, we have demonstrated that purified recombinant IKK␣ and IKK␤ are direct kinases of IB␣ and function independently in vitro (30). We have also shown that both IKK␣ and IKK␤ display a sequential bi bi mechanism (30). However, our previous report did not discriminate between the possibilities of a random sequential or an ordered sequential mechanism. In the current study, we perform product and substrate inhibition experiments that demonstrate that both IKK␣ and IKK␤ proceed by a random sequential mechanism. We also demonstrate that the natural compounds quercetin and staurosporine inhibit both IKK␣ and IKK␤ with compound-specific mechanisms. Thus, the inhibitory effects of quercetin and staurosporine on the NF-B pathway are at least partially through their inhibitions of IKK␣ and IKK␤.

EXPERIMENTAL PROCEDURES
Protein Expression and Purification-IKK␣ and IKK␤ were expressed as N-terminal FLAG-tagged fusion proteins in baculovirus. The recombinant FLAG-tagged IKK␣ and IKK␤ were purified to apparent homogeneity by affinity chromatography using M2 anti-FLAG affinity gel (Sigma). The procedures for expression and purification have been described previously (30). IB␣ was expressed as a His 6 -tagged thioredoxin fusion protein (TRX-IB␣-(1-54)) in Escherichia coli and purified by a Ni 2ϩ -nitrilotriacetic acid affinity column, as described (30).
In Vitro Phosphorylation Assays-The kinase assays were performed in a plate assay format as described previously (30). Briefly, reactions (55 l) were performed at 23°C 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 dithiothreitol. The amount of substrates ATP, [␥-33 P]ATP (2000 Ci/mmol, NEN Life Science Products) and IB␣ are specified for each individual experiment. Samples were analyzed by trichloroacetic acid precipitation on a microtiter plate (Millipore), followed by liquid scintillation counting (30). 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-Initial velocity studies were performed with varying concentrations of IB␣ at a constant ATP concentration and several fixed inhibitor concentrations. Conversely, initial velocity studies were performed with varying ATP concentrations at a constant IB␣ concentration and several fixed inhibitor concentrations. All enzyme activity data are reported as the average of duplicate determinations. The initial rate v was recorded as femtomoles of phosphate transferred to IB␣ during the reaction period. Lineweaver-Burk double-reciprocal plots were generated by linear least square fits of the data. Data from inhibition experiments were fitted to either a linear competitive model (Equation 1) or a non-competitive (or mixed inhibition) model (Equation 2) (31-33).
Accordingly, secondary plots were generated by replotting the slopes, the x intercepts, and the y intercepts of the lines as a function of [inhibitor] (32). The values of K ii and K is can be determined from the secondary plots. K is is the apparent K i value that accounts for the change of the slope. K ii is the apparent K i value that accounts for the change of the y intercept.

RESULTS
Various models of kinetic mechanisms have been described for enzymes that catalyze two substrates (34,35). For IKK␣ and IKK␤, our previous study had eliminated a ping-pong mechanism and demonstrated that both enzymes followed a sequential bi bi mechanism (30). Scheme I describes the three possible sequential mechanisms: ordered sequential mechanism with ATP binding first (Model 1), ordered sequential mechanism with IB␣ binding first (Model 2), and a random sequential mechanism (Model 3). Validations of these mechanisms are described below.
Inhibition of IKK␣ and IKK␤ by the Product Inhibitor ADP-First, the kinase activities of IKK␣ and IKK␤ were determined as a function of varying concentrations of ATP at various fixed concentrations of ADP. The Lineweaver-Burk plots of the data for both IKK␣ and IKK␤ followed Michaelis-Menten kinetics ( Fig. 1A and 2A). For both IKK␣ and IKK␤, a series of doublereciprocal straight line plots intersected on the ordinate, indicating a competitive inhibition mechanism (32). Furthermore, the data were plotted as the slope of the reciprocal plot versus the concentration of the inhibitor. The replots for both IKK␣ and IKK␤ are linear (Figs. 1A and 2A, insets), and yielded K is values of 156 and 147 nM for IKK␣ and IKK␤, respectively.
We subsequently investigated the inhibition mechanism of ADP toward the substrate IB␣. The kinase activities of IKK␣ and IKK␤ were determined as a function of varying concentrations of IB␣ at various fixed concentrations of ADP. The Lineweaver-Burk plots of the data for both IKK␣ and IKK␤ yielded a series of straight lines that crossed on the abscissa, to the left side of the ordinate (Figs. 1B and 2B), indicating a non-competitive inhibition mechanism (32).
As can be seen, the product ADP is a competitive inhibitor of IKK␣ and IKK␤ with respect to ATP and a non-competitive inhibitor with respect to IB␣. This behavior is incompatible with an ordered sequential mechanism with IB␣ binding first (Scheme I, Model 2), since otherwise ADP would have been an un-competitive inhibitor with respect to IB␣. However, the results do not exclude a random sequential mechanism or an ordered sequential mechanism with ATP binding first (Scheme I, Model 1 or 3).
Inhibition of IKK␣ and IKK␤ by a Peptide Analogue of IB␣-The peptide corresponding to amino acids 21-41 of IB␣ would compete with IB␣ for binding to the enzymes, since the peptide can be phosphorylated by both IKK␣ and IKK␤ (6,30). Thus, this peptide is an alternative substrate for IKK␣ and IKK␤ with respect to IB␣. Since the 21-amino acid peptide is not retained during trichloroacetic acid precipitation and membrane filtration in the phosphorylation assay (data not shown), the assay only monitors the appearance of the radioactive 33 P on recombinant protein Trx-IB␣. Therefore, we are able to use this peptide as an alternative substrate inhibitor to study the kinetic mechanisms of IKK␣ and IKK␤. In an effort to further elucidate the sequential mechanism (Scheme I, Model 1 or Model 3), we inhibited the phosphorylation of IB␣ with this SCHEME I. peptide using approaches similar to that employed for the ADP inhibition studies as described above. As shown in Figs. 3A and 4A, double-reciprocal plots of 1/v versus 1/[IB␣] at various fixed peptide concentrations yielded straight lines that crossed on the ordinate, confirming its being a competitive inhibitor toward the substrate IB␣ for both IKK␣ and IKK␤. The apparent K is values of 139 and 90 M for IKK␣ and IKK␤, respectively, were obtained from linear secondary plots (Figs. 3A and 4A, insets).
The kinase activities of IKK␣ and IKK␤ were also measured as a function of varying concentrations of ATP at several different fixed concentrations of peptide IB␣- . The Lineweaver-Burk plots of the data for both IKK␣ and IKK␤ yielded a series of straight lines that intersected on the abscissa, to the left side of the ordinate, indicating a non-competitive inhibition mechanism (Figs. 3B and 4B).
The different patterns of product inhibition and substrate inhibition for bi bi sequential reactions have been derived (34,35). The inhibition patterns obtained for IKK␣ and IKK␤ in this study are summarized in Table I. The fact that the product ADP was a competitive inhibitor versus ATP but a non-competitive inhibitor versus IB␣ indicates either a random sequential mechanism (Scheme I, Model 3) or an ordered sequential mechanism with ATP binding first (Scheme I, Model 1). The peptide IB␣-(21-41) behaves as a competitive inhibitor versus IB␣ but as a non-competitive inhibitor versus ATP. This eliminates the possibility of an ordered sequential mechanism with ATP binding first (Scheme I, Model 1), which would give an un-competitive inhibition pattern with respect to ATP. In conclusion, the kinetics of IKK␣ and IKK␤ follow a randomordered sequential bi bi mechanism (Scheme I, Model 3).
Staurosporine Is an ATP-competitive Inhibitor of IKK␣ and IKK␤-The natural kinase inhibitor staurosporine has been implicated to inhibit the NF-B pathway since it blocks LPSstimulated NF-B activation in THP-1 monocytic cells (22). Since LPS activates NF-B through IKK in THP-1 cells (27), we decided to test whether staurosporine inhibits IKK␣ or IKK␤. Staurosporine inhibited both IKK␣ and IKK␤ in a dose-dependent manner, with an apparent IC 50 of 0.85 and 1.6 M for IKK␣ and IKK␤, respectively (Fig. 5A). The effect of staurosporine on the initial velocity patterns for IKK␣ and IKK␤ are shown in Fig. 5, B and C. Double-reciprocal plots of 1/v versus 1/[ATP] at different fixed concentrations of staurosporine intersect on the ordinate, indicating that the inhibitor is competitive with ATP for both IKK␣ and IKK␤ (Fig. 5, B and C). As represented in Fig. 5D, increased concentrations of IB␣ did not reduce the inhibition of IKK␣ and IKK␤ by staurosporine, indicating that staurosporine is non-competitive with IB␣. This is consistent with staurosporine being a competitive inhibitor with ATP (Fig. 5, B and C). Global fitting of the data in Fig. 5, B and C, to a competitive inhibition model (EnzFitter program, Biosoft) yielded K i values of 86 Ϯ 17 and 172 Ϯ 39 nM for IKK␣ and IKK␤, respectively. The potent inhibition of IKK␣ and IKK␤ by staurosporine is consistent with its potent inhibition of NF-B activation (22).
IKK␣ and IKK␤ Are Inhibited by Quercetin-Quercetin has been reported as an inhibitor of both tyrosine kinases and Ser/Thr kinases (18,36). Since quercetin inhibits TNF-induced nuclear translocation of NF-B (19), we investigated whether it acts upon IKK␣ and IKK␤. Quercetin inhibited both IKK␣ and IKK␤ (Fig. 6, A and B), with an apparent IC 50 value of 11 and 4 M, respectively. Daidzein, a structural analogue of quercetin (Scheme II), showed no significant inhibitory effects on the activities of IKK␣ and IKK␤ (Fig. 6, A and B). Since daidzein failed to block TNF-mediated NF-B activation at 80 g/ml (19), this result is consistent with IKK␣ and IKK␤ being involved as molecular targets of quercetin in the TNF pathway.
We further investigated the inhibition mechanism of quercetin on IKK␣ and IKK␤. We first examined kinase inhibition by quercetin in the presence of various amounts of ATP. Fig. 7, A and B, shows double-reciprocal plots of 1/v versus 1/[ATP] at several fixed concentrations of quercetin. The Lineweaver-Burk plots of the data for both IKK␣ and IKK␤ are linear, indicating Michaelis-Menten kinetics at each individual concentration of quercetin (Fig. 7, A and B). For both IKK␣ and IKK␤, quercetin significantly reduced the apparent V max (1/y intercept) and increased the apparent K m (1/x intercept), indicating a mixed type inhibition mechanism. However, both se-ries of double-reciprocal plots did not intersect at a single point to the left of the ordinates (Fig. 7, A and B), suggesting a more complicated mechanism than the standard linear mixed type inhibition mechanism (33). In contrast to that observed for staurosporine (Fig. 5D), the inhibition of IKK␣ and IKK␤ by quercetin was protected by increased amounts of substrate IB␣ (Fig. 7C). This result is consistent with quercetin being a non-exclusive inhibitor with respect to ATP and IB␣ as indicated by Fig. 7, A and B. These observations suggest that the binding site of quercetin may overlap with both the ATP-and IB␣-binding sites. DISCUSSION Previous kinetic studies of IKK␣ and IKK␤ did not discriminate between a random sequential or an ordered sequential mechanism (30). The results of the present inhibition studies   (37). The values of ␣, K ATP , and K IB␣ that were used in the calculations were taken from the previous report (30). The ␤ represents the ratio of K B in the presence and absence of inhibitor.
clearly demonstrate that both IKK␣ and IKK␤ proceed through a random sequential mechanism. The equilibria shown in Scheme III describe the kinetic parameters in a random sequential bi bi system. In our previous report (30), we had fitted the two-substrate profiling data of IKK␣ and IKK␤ to a random sequential model as described in Scheme III. As a result, 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. For IKK␤, values of 130 nM, 1.4 M, 0.30/min, and 1.0 were obtained for K ATP , K IB␣ , k cat , and ␣, respectively (30). Thus, as we have proven the random sequential model in this study, the kinetic mechanisms and parameters of IKK␣ and IKK␤ are now complete. Since the native 500 -900-kDa IKK complex is composed of both IKK␣ and IKK␤ (6,7), the kinetics of the IKK complex is likely to proceed through a random sequential mechanism. Consistent with this assumption, it has been shown that a multisubunit IB kinase complex isolated from HeLa cells displays a random sequential mechanism (38), although it has not been demonstrated whether it is the same IKK complex that contains IKK␣ and IKK␤.
The product ADP is a potent inhibitor for both IKK␣ and IKK␤, with a K i value of 125 and 136 nM, respectively (Table I). These values are slightly higher than the corresponding K ATP values (85 nM for IKK␣ and 130 nM for IKK␤) (30). This suggests that the binding of ATP to IKK␣ and IKK␤ is predominantly mediated by the ADP portion of the molecule. It should be noted that, within the kinase family, a distinguishing feature for IKK␣ and IKK␤ is their low K m (ATP) value for ATP (ϳ100 nM) (30). In comparison, much higher K m (ATP) values have been reported for other Ser/Thr protein kinases, such as cAMP-dependent protein kinase (K m ϭ 10 M) (39) and p38 mitogen-activated protein kinase (K m ϭ 23 M) (40). Similarly, IKK␣ and IKK␤ have unprecedented low K i (ADP) values within the kinase family. The high affinity of IKK␣ and IKK␤ to substrate ATP would allow for the design of substrate-based inhibitors. The low K i values of ADP are in support of such a feasibility. Adenosyl-based compounds such as sulfonylbenzoyl adenosine have been previously designed and found to inhibit tyrosine kinases (41,42). In addition, based on structure homology modeling, 2Ј-thioadenosine has been successfully designed to inhibit selectively the ErbB tyrosine kinase subfamily (43). Similar rational approaches are applicable to design selective inhibitors of IKK␣ and IKK␤.
As expected, the peptide IB␣-(21-41) is a competitive inhibitor with respect to IB␣. The K i value of this peptide for IKK␤ is lower than the K i value for IKK␣, consistent with the observation that IKK␤ has a higher affinity to substrate IB␣ than IKK␣ (30) Based on the model shown in Scheme III, ␤ represents the factor by which the K i is changed by the binding of the second substrate. IKK␤ has a ␤ value of 1.0 for the inhibitor IB␣-   (Table I), indicating that the binding of the peptide inhibitor to IKK␤ has no effect on the affinity for ATP. This can be visualized in Fig. 4B as the double-reciprocal plots intersected on the abscissa, indicating that the concentration of the peptide inhibitor has no effect on the apparent K m for ATP. Similarly, a ␤ value of 1.0 was obtained for the inhibitor ADP (Table I).
These results are consistent with the ␣ value of 1.0 for IKK␤ (30). For IKK␣, ␤ values of 0.7 and 1.0 were obtained for inhibitors ADP and IB␣- , respectively (Table I). These values, with allowance for experimental error, are comparable to the 1.0 ␣ value for IKK␣ (30). Taken together, for both IKK␣ and IKK␤, the binding of one substrate has no effect on the affinity for the other substrate.
The native cytokine-inducible IKK complex contains both IKK␣ and IKK␤ (5,6). By using purified recombinant IKK␣ or IKK␤, we have previously demonstrated that IKK␣ and IKK␤ are direct kinases of IB␣ but that they have no synergistic kinase activity (30). Since these two kinases share ϳ50% homology, it is possible to inhibit both kinases with a small molecule compound. This possibility is supported by our observation that staurosporine and quercetin are potent inhibitors of both kinases. On the other hand, IKK␣ and IKK␤ have distinct physiological functions (10 -15). Specific inhibition of each individual kinase may be preferred. Inhibitors that show selectivity between these kinases would allow characterization of their physiological functions in vivo.
Staurosporine inhibits widely divergent members of the protein kinase family (21). This suggests that staurosporine functions by binding to a region that is conserved throughout the protein kinase family. The inhibition of the mammalian small heat-shock protein (HSP25) kinase by staurosporine and its analogue K252a is competitive with respect to ATP (44). In addition, an ATP-competitive mechanism has been observed in the inhibition of protein kinase C and cAMP-dependent protein kinase by the staurosporine analogue K252a (45). The same mechanism is now shown in the inhibition of IKK␣ and IKK␤ by staurosporine. This is not surprising since both IKK␣ and IKK␤ contain a conserved catalytic kinase domain at the Nterminal region which includes the conserved ATP-binding site (5)(6)(7)(8)(9). At this time, staurosporine is the most potent compound inhibitor of IKK␣ (K i ϭ 86 nM) and IKK␤ (K i ϭ 172 nM) ever reported. Such potent inhibitions by staurosporine provide a starting point for building more selective inhibitors of IKK␣ and IKK␤. In fact, several staurosporine derivatives such as CGP 41251 (4Ј-N-benzoyl staurosporine) and Ro 318425 show significant selectivity for protein kinase C over cAMP-dependent protein kinase and epidermal growth factor receptor tyrosine kinase (26,46). The inhibition mechanism of quercetin on various kinases appears to be diverse. Quercetin inhibits pp60 Src tyrosine kinase as an ATP-competitive inhibitor (47). In contrast, the inhibition of phosphatidylinositol 3-kinase I and phosphatidylinositol 3-kinase II by quercetin is non-competitive versus ATP (48). In our studies of IKK␣ and IKK␤, quercetin showed a mixed inhibition mechanism toward ATP (Fig. 7). The binding site of quercetin is likely to overlap with both the ATP and IB␣ binding pockets.
Several tyrosine kinase inhibitors, such as quercetin, genistein, staurosporine, and herbimycin, are able to inhibit NF-B activation (19,22). Thus, it has been implicated that tyrosine kinase(s) are involved in NF-B regulation. However, there is a lack of direct evidence that tyrosine kinases participate in the NF-B pathway. We have now shown that quercetin and staurosporine inhibit IKK␣ and IKK␤, the two key regulated serine kinases in the NF-B pathway, consistent with their inhibitory effects on NF-B activation. In addition, IKK␣ and IKK␤ were not inhibited by daidzein (Fig. 6), a quercetin analogue without inhibitory effects on TNF-induced NF-B activation (19). The tyrosine kinase inhibitor genistein also inhibits IKK␤. 2 Since kinase inhibitors usually have poor selectivity, their inhibitory effects on certain signaling pathways are likely to be a combination of inhibitions of several kinase targets within multiple signaling cascades. This study suggests that the inhibitory 2 G. Peet and J. Li, unpublished data. SCHEME III. Equilibria in a random sequential mechanism. Left, without inhibitor; right, with an inhibitor that competes with substrate A. effects of staurosporine and quercetin on NF-B activation are at least partially due to the inhibition of IKK␣ and IKK␤. As NF-B is a key cellular regulator of the inflammatory response, the anti-inflammatory properties of quercetin and staurosporine (23-26) may be partially due to their inhibition of IKK␣ and IKK␤. A correlation between the anti-inflammatory effects and the inhibition of IKK␤ has been observed for aspirin and salicylate (28).
The recent in vivo knock-out studies of IKK␤ imply that IKK␤ is a valid target for inflammatory diseases (11,14,15). Thus high throughput screening for inhibitors of IKK␤ could yield small molecules of therapeutic value. Here we have demonstrated the kinetic mechanism of both IKK␣ and IKK␤ to be random sequential, with each substrate binding independently of the other. This characterized kinetic mechanism will help in the evaluation of potential drug leads. Based on the potent inhibition of IKK␤ by ADP, staurosporine, and quercetin, these compounds may be considered starting points for designing specific inhibitors. The different inhibition mechanisms of staurosporine and quercetin also indicate that potent inhibition of the enzyme can be achieved by targeting different parts of the ATP-binding site. However, it is challenging to create tight-binding inhibitors that are selective between IKK␣ and IKK␤, the two homologous kinases that have similar kinetic mechanisms. Comparison of x-ray crystal structures of both kinases will help us to accomplish this goal.