BIRB796 inhibits all p38 MAPK isoforms in vitro and in vivo.

The compound BIRB796 inhibits the stress-activated protein kinases p38alpha and p38beta and is undergoing clinical trials for the treatment of inflammatory diseases. Here we report that BIRB796 also inhibits the activity and the activation of SAPK3/p38gamma. This occurs at higher concentrations of BIRB796 than those that inhibit p38alpha and p38beta and at lower concentrations than those that inhibit the activation of JNK isoforms. We also show that at these concentrations, BIRB796 blocks the stress-induced phosphorylation of the scaffold protein SAP97, further establishing that this is a physiological substrate of SAPK3/p38gamma. Our results demonstrate that BIRB796, in combination with SB203580, a compound that inhibits p38alpha and p38beta, but not the other p38 isoforms, can be used to identify physiological substrates of SAPK3/p38gamma as well as those of p38alpha and p38beta.

The stress-activated protein kinase (SAPK) 1 p38 isoforms are mitogen-activated protein kinase (MAPK) family members that are activated by changes in the cellular environment, such as alterations in the concentration of nutrients, cytokines, celldamaging agents, and changes in osmolarity of the surrounding medium (1). They comprise p38␣, p38␤, SAPK3/p38␥ (also known as ERK6), and SAPK4/p38␦. Each p38 isoform may have different biological functions and different physiological substrates, but they all phosphorylate substrates containing the minimal consensus sequence Ser/Thr-Pro. A major challenge of current research in this field is to identify the downstream physiological substrates and processes that each p38 MAPK regulates in the cell, as well as determining which "upstream" components regulate their activities. One of the most successful aids to the identification of physiological substrates has been the use of small cell-permeable compounds that are specific inhibitors of particular protein kinases. These compounds enter cells within minutes and act rapidly to suppress the activity of a particular kinase so that indirect effects caused, for example, by changes in gene expression or protein activity, a potential risk when cells deficient in a particular kinase are used, are excluded. Moreover, the use of protein kinase inhibitors avoids the need for transfection-based approaches, which have the potential to give misleading results since the fidelity of signaling can break down when components are overexpressed.
Identification of physiological substrates for p38␣ and p38␤ has been greatly facilitated by the availability of specific inhibitors of these enzymes, such as the cell-permeant pyridinyl imidazole SB203580 and related compounds (2,3). Substrates for p38␣ and p38␤ include other protein kinases, as well as several transcription factors and metabolic protein (4,5). However, little is known about the physiological substrates for SAPK3/p38␥ and SAPK4/p38␦ as they are not inhibited by SB203580 (6,7), and so far there are not any commercially available inhibitors for these kinases. Nevertheless, we have recently demonstrated that the synapse-associated proteins SAP90 and SAP97 are physiological substrates of SAPK3/p38␥ by using a cell-permeant peptide that blocks the interaction between SAPK3/p38␥ and these PDZ domain-containing proteins (8,9). Moreover, by using small interfering RNA technology, we have also shown that, after cellular stress, the microtubule-associated protein Tau is an in vivo substrate of SAPK4/ p38␦ in neuroblastoma cells (10).
Recently, a new class of p38 inhibitors has been described. These are diaryl urea compounds, which bear little structural similarity to the other class of well characterized p38 inhibitors, the pyridinyl-imidazoles. The compound BIRB796 is one of the most potent compounds of this series and binds to p38␣ with both slow association and dissociation rates (11). BIRB796 inhibits p38␣ by a novel mechanism, indirectly competing with the binding of ATP. Structure determination revealed that, prior to binding, the kinase undergoes a reorganization of the activation loop exposing a critical binding domain and yielding a structure incompatible with ATP binding (11). BIRB796 demonstrated efficacy in an endotoxin (lipopolysaccharide)-stimulated mouse model of tumor necrosis factor-␣ production and in a mouse model of established collagen-induced arthritis (11). BIRB796 also displayed anti-inflammatory effects in a trial of human endotoxemia and has recently been in phase IIb/III clinical trials for the treatment of rheumatoid arthritis (11)(12)(13)(14).
Here we show that in addition to p38␣, other MAPK family members activated by cellular stress, such as p38␤, SAPK3/ p38␥, and SAPK4/p38␦, are also inhibited by BIRB796. We also show that in a cell-based assay, the use of different concentrations of BIRB796, in combination with other well characterized inhibitors of p38␣/␤, such as SB203580, can be a useful tool for the identification of new substrates of SAPK3/p38␥ and thus for the elucidation of its physiological role.
Protein Expression-All protein kinases were of human origin and were expressed either as glutathione S-transferase fusion protein, as maltose-binding protein in Escherichia coli, or as hexahistidine (His 6 )tagged protein in insect Sf9 or Sf21 cells and purified as described previously (3,17).
Protein Kinase Assays-All protein kinase assays were linear with respect to time. Assays were either performed manually for 10 min at 30°C in 50-l incubations using 150 nM p38 MAPK isoforms and 10 M [␥-32 P]ATP or performed with a Biomek 2000 laboratory automation workstation in a 96-well format (Beckman Instruments) for 40 min at ambient temperature in 25-l incubations using [␥-33 P]ATP. The concentrations of magnesium acetate and ATP in the assays were 10 mM and 10 M, respectively, unless stated otherwise. Assays were initiated with MgATP. Manual assays were terminated by spotting aliquots of each incubation onto phosphocellulose paper followed by immersion in 75 mM phosphoric acid. Robotic assays were stopped by the addition of 5 l of 0.5 M orthophosphoric acid. Aliquots were then spotted onto P81 filtermats, washed four times in 75 mM phosphoric acid to remove ATP, washed once in acetone (manual assays) or methanol (robotic assays), and then dried and counted for radioactivity. All protein kinases were assayed as described previously (3,17). MAPKAP-K2 and RSK2 were immunoprecipitated from cell lysates using specific antibodies and assayed as described (3,(17)(18)(19).
Cell Culture, Transfection, and Lysis-Human embryonic kidney (HEK) 293 and HeLa cells were cultured in Dulbecco's modified Eagle's medium at 37°C, supplemented with 10% fetal calf serum, 50 units of penicillin/ml, 50 g/ml streptomycin (Invitrogen), and 2 mM glutamine (BioWhittaker). Mouse embryonic fibroblasts were cultured as described previously (9), and C2C12 myoblasts were cultured as described in Ref. 20. Cells were exposed to 0.5 M sorbitol for 30 min or 100 ng/ml EGF for 10 min and then lysed in buffer A (50 mM Tris-HCl, pH 7.5, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 50 mM sodium ␤-glycerophosphate, 5 mM pyrophosphate, 0.27 M sucrose, 0.1 mM phenylmethylsulfonyl fluoride, 1% (v/v) Triton X-100) plus 0.1% (v/v) 2-mercaptoethanol and Complete proteinase inhibitor mixture from Roche Applied Science. Lysates were centrifuged at 18,000 ϫ g for 5 min at 4°C, and the supernatants were removed, quick-frozen in liquid nitrogen, and stored at Ϫ20°C until use. When required, cells were preincubated for 1 h without or with 10 M SB 203580 or 10 M PD 184352 or with different concentrations of BIRB796 for the times indicated in the figures.
Immunoprecipitation from Cell Lysates-MAPKAP-K2 and SAP97 were immunoprecipitated from 0.1 to 1 mg of HEK293 cells, whereas RSK2 was immunoprecipitated from 0.1 mg of HeLa cell extract.
Extracts were incubated with 3 and 5 g of the specific antibodies coupled to protein G-Sepharose, respectively (Amersham Biosciences). After incubation for 2 h at 4°C, the captured proteins were centrifuged at 13,000 ϫ g, the supernatant was discarded, and the beads were washed once in buffer A containing 0.5 M NaCl and then washed twice in buffer A alone. Samples were denatured, electrophoresed in precast polyacrylamide gels (Invitrogen), and then immunoblotted. Quantification of protein phosphorylation and total protein after immunoblotting was carried out using the Odyssey TM infrared imaging system (LI-COR Biosciences).

BIRB796 Inhibits All Four SAPK/p38s Isoforms in Vitro-
There are about 500 protein kinases encoded in the human genome, most of which belong to the same superfamily. It is therefore difficult to develop compounds that inhibit one particular protein kinase without inhibiting several related enzymes. Establishing the specificity of any particular inhibitor is a critical issue. We have previously examined the specificity of more than 40 commercially available compounds against a large panel of protein kinases and found that most inhibitors target more than one kinase (3,17). There is therefore a danger that in cell-based assays, the observed effects do not result from inhibition of the kinase of interest but rather from the inhibition of another protein kinase. As the selectivity of BIRB796 against only 11 different kinases was reported previously (11), we decided to extend the specificity of this compound to a larger TABLE I Inhibition of protein kinases by BIRB796 inhibitor The inhibitor concentrations used are shown in parentheses. Results are presented as activity relative to that in control incubations where inhibitor was omitted (means of duplicate determinations). The time of pre-incubation of BIRB796 with the kinase was 5 min. ATP concentration was 5 M in SAPK3/p38␥, SAPK4/p38␦, protein kinase B (PKB), GSK3␤, CDK2 and MKK1 assays; 20 M in JNK1 to 3, PRAK, ROCK-II, p38␤, CDK2/cyclinA, CHK1, MSK1, CSK, S6K1, cAMP-dependent protein kinase (PKA), CK1, MAPKAP-K2, serum and glucocorticoid-inducible kinase (SGK), protein kinase C-␣ (PKC␣), and PDK1 assays; and 50 M for the rest of the kinases listed in the table.
panel of protein kinases. We found that, at a concentration of 10 M, BIRB796 inhibited the activity of p38␣, Lck, and JNK2 as described previously (11). However, it also inhibited the activity of the other three p38 MAPK isoforms p38␤, SAPK3/ p38␥, and SAPK4/p38␦. It had little effect on the other protein kinases in the panel (Table I).
Time-dependent Inhibition of p38 MAPKs by BIRB796 -It has been shown that BIRB796 inhibits p38␣ activity in a timedependent manner due to its slow binding behavior (11). To examine whether or not BIRB976 affects the activity of other p38 isoforms similarly, we monitored the apparent inhibitory potency of the compound as a function of the time of preincubation with the kinase (Fig. 1A). We observed firstly that BIRB796 blocked the individual p38 MAPK activities at different potencies in vitro. Thus, this compound inhibited p38␣ more potently than p38␤ and inhibited p38␤ more potently than SAPK3/p38␥, whereas SAPK4/p38␦ was inhibited the least (Fig. 1, A and B). Secondly, the apparent IC 50 value for all p38 MAPK isoforms decreased as the time of preincubation with the inhibitor increased (Fig. 1, A and B). These results are consistent with the slow binding kinetics of this compound to the p38 MAPK isoforms.
BIRB796 Is a Potent Inhibitor of p38 Isoforms in Cells-To test whether or not BIRB796 could inhibit each p38 MAPK in cells, we first investigated the effect that this compound had on the activation of one known physiological substrate of p38␣, namely the protein kinase MAPKAP-K2. We incubated HEK293 cells with different concentrations of BIRB796 for 30 min or 2 h prior to stimulation with sorbitol (an osmotic shock) and examined the activation of MAPKAP-K2 by measuring its activity ( Fig. 2A). MAPKAP-K2 activation was inhibited in these cells in a time-dependent manner with an apparent IC 50 of 30 nM after 30 min or 8 nM after 2 h of preincubation with BIRB796 ( Fig. 2A). MAPKAP-K2 activation was also blocked by preincubation of the cells with the p38 MAPK inhibitor SB203580 as described previously (2) (Fig. 2A).
We also examined the phosphorylation of p38␣ and p38␤ under the same conditions and observed that the phosphorylation of these two kinases was also blocked by pretreatment of the cells with the compound (Fig. 2, B and C). These results suggest that the binding of BIRB796 to these p38 MAPKs is also impairing their phosphorylation by the upstream kinases MKK6/MKK3 and/or enhancing their dephosphorylation. On the other hand, preincubation of cells with the SB203580 did not prevent p38␣ phosphorylation by this agonist (Fig. 2B), although it blocked p38␤ phosphorylation, as shown previously (5) (Fig. 2C).
To test whether BIRB796 could also inhibit the phosphoryl- ation, and therefore activation, of SAPK3/p38␥ and SAPK4/ p38␦, cells were preincubated with different concentrations of BIRB796 for 30 min or 2 h before stimulation with sorbitol. The phosphorylation of SAPK3/p38␥ was blocked by this compound (Fig. 3A) with an apparent IC 50 of 90 nM or 35 nM after 30 min or 2 h preincubation, respectively (Fig. 3B), which are 3 and from HEK293 treated as in panel A were immunoblotted using either an antibody that recognizes phosphorylated p38␣ (Phos-p38␣) or an antibody that recognizes both phosphorylated and unphosphorylated p38␣. C, BIRB796 inhibits the activation of p38␤. C2C12 myoblasts were treated as in panel A, and p38␤ was immunoprecipitated from 3 mg of cell lysate and immunoblotted with either the p38␣ phospho-specific antibody, which also recognizes p38␤, or an antibody that recognizes both phosphorylated and unphosphorylated p38␤.  Fig. 2A before being exposed to 0.5 M sorbitol for 30 min. Cell lysates (50 g) were immunoblotted with either the p38␣ phospho-specific antibody (Phos-p38␣), which also recognizes phosphorylated SAPK3/p38␥ (Phos-SAPK3), or an antibody that recognizes both phosphorylated and unphosphorylated SAPK3/p38␥ (SAPK3). B, the intensity of the bands in the immunoblots from panel A was quantified as described under "Materials and Methods," and the percentage of inhibition of SAPK3/p38␥ phosphorylation was calculated. Results in panel B are shown as the mean Ϯ S.E. for duplicate determinations from two experiments.
4-fold higher than the IC 50 for blockade of MAPKAP-K2 activation. Since the expression level of SAPK4/p38␦ in HEK293 cells is very low and this isoform is not activated in these cells after sorbitol treatment (9), the effect of BIRB796 on the activation of this kinase was studied in mouse embryonic fibroblasts. Surprisingly, SAPK4/p38␦ phosphorylation after osmotic shock was enhanced at low doses of the inhibitor and blocked at higher concentrations of BIRB796 (Fig. 4). The apparent IC 50 for SAPK4/p38␦ was ϳ20-fold higher than the IC 50 for SAPK3/p38␥. Phosphorylation of SAPK4/p38␦ after stimulation was also enhanced when cells were preincubated with the p38␣/p38␤ inhibitor SB203580 (Fig. 4). One possible explanation for this result is that p38␣/p38␤ negatively regulates SAPK4/p38␦ activation by an unknown mechanism. Alternatively the binding of BIRB796 and SB203580 to p38␣/p38␤ may allow MKK6 and/or MKK3 to activate SAPK4/p38␦ preferentially.
BIRB796 Blocks the Phosphorylation of the SAPK3/p38␥ Substrate SAP97 in Cells-The scaffold protein SAP97 is the mammalian homologue of the Drosophila discs large tumor suppressor and a physiological substrate of SAPK3/p38␥. SAP97 becomes phosphorylated in HEK293 cells at four major residues (Ser 158 , Thr 209 , Ser 431 , and Ser 442 ) in response to osmotic stress (9) (Fig. 5A). The phosphorylation of all four residues was greatly reduced when cells were pretreated with 1 M BIRB796 before stimulation but not by pretreatment with 10 M SB203580 (Fig. 5A). The apparent IC 50 for inhibition of Ser 158 phosphorylation was 150 nM or 60 nM after 30 min or 2 h preincubation, respectively (Fig. 5, B and C), similar to the IC 50 for inhibition of SAPK3/p38␥. BIRB796 Inhibits JNK Activation in Vivo-We also examined the effect of the inhibitor BIRB796 on the activation of the c-Jun N-terminal kinase (JNK), which is also a MAPK family members activated by stress. Exposure of cells to osmotic shock caused an increase in the phosphorylation of both alternative spliced isoforms (46 and 54 kDa) of JNK1 and JNK2, which was completely impaired when cells were preincubated with 10 M BIRB796 (Fig. 6A). Inhibition of p54(JNK1/2) phosphorylation was more sensitive to BIRB796 than p46(JNK1/2) phosphorylation (Fig. 6A). Thus, after a 2-h preincubation with the inhibitor, the apparent IC 50 for p46(JNK1/2) was 1 M, and for p54(JNK1/2), it was 350 nM, whereas the apparent IC 50 for p54(JNK1/2) was higher than 1 M when cells were incubated for 30 min with BIRB796 prior to stimulation. These results suggest that BIRB796 inhibits JNK by a mechanism similar to the p38 MAPK isoforms.
The phosphorylation of one physiological substrate of JNK, the transcription factor c-Jun, was also examined (Fig. 6B). After exposure of the cells to osmotic shock in the presence of increasing concentrations of BIRB796, c-Jun was immunoblotted using a phospho-specific antibody, which specifically recognizes Ser 63 . Phosphorylation of c-Jun was completely blocked at 10 M BIRB796 with an apparent IC 50 of 1 M after 2 h of preincubation with the inhibitor (Fig. 6B), similar to the IC 50 for inhibition of p46(JNK1/2) activation under these conditions (Fig. 6A). Taken together, these results show that BIRB796 blocks JNK1/2 activation and activity in HEK293 cells but at higher concentrations than those needed to block p38␣, p38␤, or SAPK3/p38␥. BIRB796 Does Not Inhibit ERK1/2 or ERK5 Activation in Vivo-To test further the specificity of BIRB796, we examined its effect on the activation and activity of other MAPK family members, ERK1/2 and ERK5. These experiments were carried out in HeLa cells in which the activation of the ERK1/2 and ERK5 pathways by EGF is well characterized (21).
Phosphorylation of ERK1/ERK2 and activation of RSK2, one of their in vivo substrates, were not prevented by preincubation for up to 2 h of the cells with even 10 M BIRB796, whereas preincubation with the classical MAPK pathway inhibitor PD184352 completely blocked ERK1/ERK2 phosphorylation FIG. 4. Effect of BIRB796 on the activation of SAPK4/p38␦ in mouse embryonic fibroblasts. Mouse embryonic fibroblasts were preincubated for 2 h with different concentrations of BIRB796 or 10 M SB203580 as indicated before being exposed to 0.5 M sorbitol for 30 min. Endogenous SAPK4/p38␦ was immunoprecipitated from 2 mg of cell lysate, and the pellets were immunoblotted using the p38␣ phosphospecific antibody that also recognizes SAPK4/p38␦ (Phos-SAPK4) or using an antibody that recognizes both phosphorylated and unphosphorylated SAPK4/p38␦ (SAPK4).

FIG. 5. Effect of BIRB796 on SAP97 phosphorylation.
A, phosphorylation of SAP97 in HEK293 cells after osmotic shock. Cells were incubated for 2 h with or without 10 M SB203580 or 1 M BIRB796 and then exposed for 20 min to 0.5 M sorbitol. Endogenous SAP97 was immunoprecipitated from 0.5 to 5 mg of cell lysate, and the pellets were immunoblotted using an antibody that recognizes SAP97 phosphorylated at Ser 158 (Phos-Ser158), Thr 209 (Phos-Thr209), Ser 431 (Phos-Ser431), Ser 442 (Phos-Ser442), and an antibody that recognizes unphosphorylated and phosphorylated SAP97 equally well (SAP97). B, HEK293 cells were preincubated with different concentrations of BIRB796 or 10 M SB203580 as indicated before being exposed to 0.5 M sorbitol for 30 min. SAP97 was immunoprecipitated from 1 mg of cell lysate and immunoblotted using either anti-Phos-Ser158 or the antibody that recognizes unphosphorylated and phosphorylated SAP97. C, the intensity of the bands in immunoblots from panel B was quantified as described under "Materials and Methods," and the percentage of inhibition of SAP97 phosphorylation was calculated. Results in panel C are shown as the mean Ϯ S.E. for duplicate determinations from two experiments. and RSK2 activation (Fig. 7, A and B). These results indicated that BIRB796 does not inhibit the activation and activity of ERK1/ERK2.
The phosphorylation of ERK5 could be detected by a small decrease in its electrophoretic mobility using an antibody that recognizes the phosphorylated and the dephosphorylated forms of the protein equally well (21). Preincubation of HeLa cells with PD184352 inhibited the EGF-induced phosphorylation of ERK5, as expected from earlier studies (21,22) (Fig. 7C). However, increasing concentrations of BIRB796 did not affect the mobility of the kinase caused by its phosphorylation and therefore did not block the activation of ERK5 induced by EGF (Fig. 7C).
BIRB796 Inhibits the Phosphorylation of All Four SAPK/ p38s Isoforms in Vitro-To examine whether or not the binding of BIRB796 to p38 MAPKs or JNKs also impaired their phosphorylation by the upstream kinases MKK6 or MKK4, respectively, and/or accelerated their dephosphorylation, we first studied the in vitro phosphorylation of each p38 MAPK isoform by MKK6 and the phosphorylation of the JNK2␣2 isoform by MKK4. We found that BIRB796 blocked the individual p38 MAPK and the JNK2␣2 phosphorylation at the same potency (Fig. 8A). On the other hand, pretreatment of the cells with the phosphatase inhibitor vanadate, to prevent MAPK dephosphorylation, did not cause any effect on the inhibition of the phosphorylation of SAPK3/p38␥, p38␣, or JNK1/2 by BIRB796 after sorbitol stimulation (Fig. 8B). As a control, we checked that preincubation of cells with vanadate increased the phosphorylation of ERK1/2, indicating that the phosphatase inhibitor was working (Fig. 8B). All of these results suggest that the binding of BIRB796 to the p38 MAPKs or JNK1/2 is impairing their phosphorylation by the upstream kinase MKK6 or MKK4 rather than enhancing their dephosphorylation. DISCUSSION Recently, it has been described that the diaryl urea compound BIRB796 is a highly selective inhibitor of the protein kinase p38␣ (11). To establish the specificity of a particular inhibitor is a critical issue as a means for therapeutic intervention. Since protein kinases are a large class of enzymes, most of which belong to the same family, and the degree of homology across the entire family is relatively high, especially within the catalytic core (23), it is difficult to develop compounds that inhibit one particular protein kinase without inhibiting several related enzymes. For this reason, we decided to reexamine the specificity of the BIRB796 compound. The work described in FIG. 6. Effect of BIRB796 on the activation of JNK1/2 and the phosphorylation of c-Jun. HEK293 cells were preincubated with different concentrations of BIRB796 or 10 M SB203580 as indicated in Fig. 2A before being exposed to 0.5 M sorbitol for 30 min. A, cell lysates (100 g) were immunoblotted with either the JNK1/2 phospho-specific antibody (Phos-JNK1/2) or an antibody that recognizes both phosphorylated and unphosphorylated JNK1/2 (JNK1/2). B, c-Jun from 0.2 mg of cell lysate was immunoblotted using an antibody that recognizes c-Jun phosphorylated at Ser 63 (Phos-C-Jun) and an antibody that recognizes unphosphorylated and phosphorylated protein equally well (C-Jun). The intensity of the bands in immunoblots from panels A and B was quantified as described under "Materials and Methods," and the percentage of inhibition of JNK1/JNK2 and c-Jun phosphorylation was calculated. Results are shown as the mean Ϯ S.E. for duplicate determinations from two experiments.
the present study shows that BIRB796 inhibits the activity in vitro, and the activation in the cell, of all p38 MAPK and JNK1/2 isoforms. Structural analysis of p38␣ has shown that BIRB796 binds to a novel site within the ATP-binding pocket of the kinase, which is created a by conformational change in the enzyme induced by the inhibitor, which yields to a structure incompatible with ATP binding (11). Moreover, solution studies have demonstrated that this class of compound has slow binding kinetics, consistent with the requirement for a conformational change (11). Our in vitro and in vivo data show that BIRB796 inhibits all p38 MAPK and JNK isoforms in a time-dependent manner, suggesting that the association of this compound with each of these kinases is slow, probably due to the requirement of the previously described conformational change. BIRB796 binding is targeted to the Phe residue in the conserved DFG motif that is buried in a hydrophobic pocket between the two major lobes of the kinase domain (11). Although the residues in this pocket are highly conserved among all of these kinases, we have observed that BIRB796 is more selective for p38␣ and p38␤ than for the SAPK3/p38␥ and SAPK4/p38␦ isoforms and other members of the MAPK family. This could be due to the fact that this diaryl urea inhibitor also utilizes the hydrophobic pocket containing Thr 106 (11) a residue unique to p38␣ and p38␤.
Moreover, we also show that BIRB796 impairs the phospho-rylation of p38 MAPKs or JNKs by the upstream kinase MKK6 or MKK4 but does not affect their dephosphorylation in vivo.
Our results suggest that the conformational change caused by the binding of the inhibitor to the MAPK may affect the structure of both its phosphorylation site and the docking site for the upstream activator, therefore impairing the phosphorylation of p38 MAPKs or JNKs. We also show that incubation of cells with BIRB796 prior to stimulation blocks phosphorylation of the physiological substrates of p38␣ and SAPK3/p38␥. In particular, BIRB796 impairs the phosphorylation induced by osmotic shock of SAP97, which is a SAPK3/p38␥ substrate.
Recently, we have shown that the phosphorylation of different PDZ domain-containing proteins by SAPK3/p38␥, such as SAP97, is dependent on the interaction of the C-terminal se- FIG. 7. BIRB796 has no effect on ERK1/2 or ERK5 activation. A, HeLa cells were preincubated either for 2 h in the presence of the indicated concentrations of BIRB796 or for 1 h in the presence of 10 M PD184352 and then for 10 min with 100 ng/ml EGF in the continued presence of the inhibitors. Lysates (50 g) were immunoblotted using either an antibody that recognizes phosphorylated ERK1/2 (Phos-ERK1/2) or an antibody that recognizes both phosphorylated and unphosphorylated ERK1/2 (ERK1/2). B, RSK2 was immunoprecipitated from 0.1 mg of protein lysate from cells treated as in panel A and assayed as described under "Materials and Methods." C, lysates (50 g) from cells treated as in panel A were immunoblotted using anti-ERK5 antibody that recognizes both phosphorylated and unphosphorylated form of the kinase (Phos-ERK5) to detect the band-shift caused by phosphorylation.

FIG. 8. Effect of BIRB796 on the phosphorylation of p38
MAPKs and JNKs. A, the rate of phosphorylation, of each p38 isoform and JNK2␣2, relative to control in the absence of BIRB796 (100%). MKK6 and MKK4 were activated in vitro (see "Materials and Methods"). The phosphorylation of each p38 was studied at MKK6 concentration of 10 units/ml for p38␣, 100 units/ml for p38␤, 50 units/ml for SAPK3/p38␥, or 20 units/ml for SAPK4/p38␦, and the phosphorylation of JNK2␣2 was studied at an MKK4 concentration of 10 units/ml. All four recombinant human p38 MAPKs or JNK2␣2 (1 g) were incubated at room temperature with MKK6 or MKK4, respectively, in the presence of the indicated concentration of BIRB796 for 30 min before starting the reaction with Mg-[␥ 32 P]ATP for 10 min as described under "Materials and Methods." Results are shown as the mean Ϯ S.E. for duplicate determinations from two experiments. B, HEK293 cells were preincubated for 2 h with different concentrations of BIRB796 or 10 M SB203580 as indicated and for 1 h in the presence or absence of 10 mM orthovanadate before being exposed to 0.5 M sorbitol for 30 min. Cell lysates (50 g) were immunoblotted with either the p38␣ phosphospecific antibody, which also recognizes phosphorylated SAPK3/p38␥ (Phos-p38␣), or an antibody that recognizes both phosphorylated and unphosphorylated p38␣ (p38␣). Cell lysates (100 g) were immunoblotted with either the JNK1/2 phospho-specific antibody (Phos-JNK1/2) or an antibody that recognizes both phosphorylated and unphosphorylated JNK1/2 (JNK1/2). As control, lysates (50 g) from cells incubated for 1 h with or without 10 mM orthovanadate were immunoblotted using either an antibody that recognizes phosphorylated ERK1/2 (Phos-ERK1/2) or an antibody that recognizes both phosphorylated and unphosphorylated ERK1/2 (ERK1/2). quence -ETXL of the kinase with the PDZ domain of these proteins (8,24). Exploiting this characteristic, we designed a cell-permeant peptide to identify different SAPK3/p38␥ substrates in vivo. This peptide blocks the phosphorylation of PDZ domain-containing proteins by SAPK3/p38␥ specifically, but not phosphorylation by other MAPK in intact cells, by preventing the association of the kinase with the PDZ domain of the protein substrate (8,9). The peptide contains the last nine residues of the SAPK3/p38␥ fused to the cell-membrane transduction domain of the human immunodeficiency virus-type I (HIV-1) Tat protein (8,9,25). This peptide has been very useful in the validation of PDZ domain-containing proteins as SAPK3/ p38␥ substrates such as SAP90/PSD95 (8) or SAP97 (9). However, the availability of new cell-permeant SAPK3/p38␥ inhibitors would be extremely useful in helping to elucidate the physiological role of this kinase and to find other possible substrates that do not contain PDZ domains.
The results presented in this study show that it is possible to vary the concentration of BIRB796 in the culture medium to differentially inhibit particular stress-activated protein kinases in combination with the more specific p38␣/␤ inhibitor SB203580. For example, in HEK293 cells exposed to osmotic shock that activates p38 MAPKs and JNKs, BIRB796 at 0.1 M inhibits p38␣/␤ specifically, whereas at 0.3 M, it also inhibits SAPK3/p38␥. Thus, physiological substrates for SAPK3/p38␥ can be identified by identifying proteins, the phosphorylation of which is completely blocked by preincubation of HEK293 cells with 0.3 M BIRB796 but not by 10 M SB203580. At higher concentrations than 1 M, BIRB796 also substantially blocks the JNK pathway. However, the precise concentration needed for inhibition may vary from cell to cell. For this reason, it is important to examine the minimum concentration of BIRB796 required to suppress the activity of a particular MAPK by 80 -90% by checking in parallel the phosphorylation of a validated substrate of the protein kinase studied.